Citation
Morphology and physiology of Azospirilla grown on beta-hydroxybutyrate

Material Information

Title:
Morphology and physiology of Azospirilla grown on beta-hydroxybutyrate
Creator:
Bleakley, Bruce Henry, 1956-
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Bacteria ( jstor )
Broths ( jstor )
Capsules ( jstor )
Carbon ( jstor )
Cells ( jstor )
Cultured cells ( jstor )
Cysts ( jstor )
Lawns ( jstor )
Nitrogen ( jstor )
Somatic cells ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
22306902 ( ALEPH )
13886998 ( OCLC )

Downloads

This item has the following downloads:


Full Text













MORPHOLOGY AND PHYSIOLOGY OF AZOSPIRILLA
GROWN ON BETA-HYDROXYBUTYRATE










By

BRUCE HENRY BLEAKLEY
























A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA 1986

















This dissertation is dedicated to Isabel, Stewart, and Robert Bleakley, the only three people who can say "Of course" without my doubting them.
















ACKNOWLEDGMENTS


I thank my major professor, Dr. Murray Gaskins, for

first bringing the topic of bacterial cysts to my attention and for allowing me to pursue the subject. I also thank the other members of my guidance committee, Dr. Stephan Albrecht, Dr. David Hubbell, Dr. David Mitchell, and Dr. Stephen Zam, for their patience, encouragement, and use of laboratory facilities. Thanks also go to Dr. Sylvia Coleman, whose transmission electron microscopy studies aided this work. Special thanks go to Dr. Kenneth Quesenberry and the state of Florida for my graduate assistantship.

Kelly Kirkendall Merritt taught me how to spread plate and introduced me to manipulations of azospirilla. She is a good coworker to be around, as are Stephanie Syslo, Mary Myers, and Dr. Garnet Jex, who helped me with discussions, calculations, and shared experiences in and out of the laboratory. Dr. Lakshmi Sadasivan discussed her work with me before it was published and furnished preprints of her work. Talks with Dr. Harold Sadoff helped me a great deal in understanding Azotobacter cysts. And Dr. Noel Krieg provided several cultures that were keys in the study. Thanks go to all of them.

iii










Without the Hume Library this work could not have been done. I thank Mr. William Weaver for running a fine facility.

Finally, completing the list of good coworkers, the eye and expertise of Louise L. Munro are felt throughout this study. Her scanning electron microscopy studies cleared up several uncertainties and raised new questions. She knew the right stuff when she saw it. Thanks, Louise.







































iv

















TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS. ........... .......................... 111

ABSTRACT ........................................... vi

CHAPTERS

I INTRODUCTION AND EXPERIMENTAL APPROACH.......... 1

Ecology of Azospirilla.............. ........... 1
Physiology of Azospirilla... .................. 4
Morphology of Azospirilla...................... 6
Prokaryotic Exopolysaccharides................. 8
Poly-6-Hydroxybutyrate (PHB)................... 12
Dormant Forms of Prokaryotic Cells ............. 22
Resistance of Bacteria to Drying ............... 27
Azotobacter Cysts.............................. 37
Pleomorphism of Azospirilla.................... 45
Experimental Approach.......................... 56

II PLEOMORPHISM OF AZOSPIRILLA GROWN ON BETA-HYDROXYBUTYRATE........................ 59

Materials and Methods.......................... 63
Results........................................ 71
Discussion..................................... 131

III PHYSIOLOGICAL PROPERTIES OF ENCAPSULATED FLOCS
OF AZOSPIRILLUM LIPOFERUM Sp RG6xx ............. 144

Materials and Methods .......................... 145
Results......................................... 153
Discussion........................................ 167

IV GENERAL CONCLUSIONS............................ 178

BIBLIOGRAPHY................. .................. 181

BIOGRAPHICAL SKETCH................................. 193





V
















Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

MORPHOLOGY AND PHYSIOLOGY OF AZOSPIRILLA
GROWN ON BETA-HYDROXYBUTYRATE

By

Bruce Henry Bleakley

May 1986

Chairman: Murray H. Gaskins Major Department: Agronomy

Strains of Azospirillum brasilense and Azospirillum

lipoferum were cultured with beta-hydroxybutyrate (BHB) to determine if they could be converted in high numbers to cyst-like forms, as can some strains of Azotobacter spp. Azospirillum brasilense strain JM 125A2 grew poorly on BHB but produced some nonmotile cells of cyst-like morphology. Azospirillum brasilense strain Cd grew better on BHB and often produced elongated cells as well as some nonmotile, cyst-like cells. Capsules and accumulation of poly-betahydroxybutyrate (PHB) were common features of all putative cysts. Encapsulation occurred with all A. lipoferum strains tested. Cells accumulated PHB and assumed elongated, filamentous shapes as they lost motility. Later, capsules were produced and microflocs formed. The filamentous cells eventually formed septa. Several cell shapes were present vi










in flocs, but all cells possessed intracellular PHB and capsules. Some cells within flocs appeared cyst-like. Broth studies indicated that alkaline pH does not cause these morphological changes.

Cells of Azospirillum lipoferum Sp RG6xx grown on

nitrogen-free BHB agar accumulated up to 57% of their dry weight as PHB, compared to 3.6% when grown with combined nitrogen. Neither vegetative nor encapsulated cells of this strain survived in significant numbers after 8 days of desiccation. Vegetative cells of this strain multiplied several fold and retained viability during 9 days of starvation for carbon and nitrogen, whereas encapsulated cells were reduced to 25% of their original numbers. Nonmotile, encapsulated cells produced motile vegetative cells when incubated with nitrate, ammonium, or soil extract but did not do so appreciably in nitrogen-free, buffered-salts solution with or without carbon sources. Treatment with Tris-EDTA did not result in expulsion of cells from their capsular coats, as it does for mature Azotobacter spp. cysts. Studies with chloramphenicol indicated that encapsulated cells do not possess the enzymes needed for growth and emergence from their capsules.

The studies suggested that PHB accumulation and capsule formation during unbalanced growth precede the formation of dormant cyst-like cells.




vii
















CHAPTER I
INTRODUCTION AND EXPERIMENTAL APPROACH


Ecology of Azospirilla


Bacteria of the genus Azospirillum have been isolated from soils and from the roots of cereal crops and forage grasses in several areas of the world (Dobereiner et al., 1976; Tyler et al., 1979; Lamm and Neyra, 1981). Their numbers in nonrhizosphere soil can be as high as 104 cells/g soil (Dobereiner, 1978), while their numbers in rhizosphere soil can be as high as 107 cells/g soil (Krieg and Dobereiner, 1984).

Agricultural interest in Azospirillum spp. resulted from recognition of their ability to reduce atmospheric dinitrogen. The enzyme catalyzing this reaction, nitrogenase, is inactivated in the presence of combined nitrogen or oxygen (Okon et al., 1976a). Azospirilla fix dinitrogen under microaerophilic conditions in nitrogen-free media in the laboratory (Day and Dobereiner, 1976; Okon et al., 1976a). Nonrhizosphere soil is usually too poor in available, utilizable carbon sources to enable Azospirillum spp. to fix dinitrogen, but they can do so in the more carbon-rich rhizosphere environment (Dobereiner et al., 1976). Maximum nitrogenase activity with inoculated plants

1









2

growing in soil is often found at the reproductive stage of the plant (reviewed by Patriquin et al., 1983), after plant uptake and other processes have reduced the amount of combined nitrogen in the root zone (Okon and Hardy, 1983). Sometimes low amounts of fixed nitrogen have been incorporated into plant tissue. The transfer of fixed nitrogen from bacterium to plant seems slow, probably because bacterial nitrogen is made available for plant uptake only after the mineralization of the organic nitrogen of dead bacteria (Okon et al., 1983).

Although the nitrogenase activity of Azospirillum

spp. may not directly provide quick or agriculturally significant benefits to inoculated plants, the bacteria have been found to possess other characteristics that may benefit plants. Axenic associations of grass seedlings and Azospirillum spp. have resulted in rapid proliferation of lateral roots and root hairs, probably due to bacterial production of indole-3-acetic acid and other plant growth substances (Tien et al., 1979; Umali-Garcia et al., 1980; Jain and Patriquin, 1985). It is also possible that Azospirillum spp. can enhance production of plant growth substances by the plants themselves (Kapulnik et al., 1985). In any case, associations of Azospirillum spp. with plant roots have led to significant increases of commercially valuable plant components in both axenic laboratory experiments (Kapulnik et al., 1985) and field inoculations








3

(Okon and Hardy, 1983). Short-term axenic associations have also resulted in enhanced uptake of mineral ions by grass roots (Lin et al., 1983). This effect may be due to the influence of plant growth substances, or to softening of the middle lamellae of root cells by pectolytic bacterial enzymes, which some azospirilla are known to produce (Umali-Garcia et al., 1980; Tien et al., 1981). Such effects on root morphology and activity may make inoculation with azospirilla beneficial in some agricultural situations.

The rhizosphere environment is prone to extreme chemical and physical fluctuations (Foster and Bowen, 1982). This may lead to periods when azospirilla are inactive due to environmental limitations. Cells of azospirilla can vary morphologically (Krieg and Dobereiner, 1984). Some of these cell forms may be dormant or resting stages, in which activities of possible benefit to plants are not expressed. Pleomorphic forms of azospirilla usually possess capsules, and contain large amounts of the reserve polymer poly-Bhydroxybutyrate (PHB). This study describes attempts to obtain such forms in high numbers by laboratory culture. The general topics of capsules, PHB, physiological dormancy, and desiccation resistance are directly related to this study, and will be briefly reviewed in this introduction after discussion of some key aspects of Azospirillum spp. physiology.








4

Physiology of Azospirilla


The three species in the genus Azospirillum all have a mainly respiratory type of metabolism. They fix dinitrogen in microaerophilic environments where combined nitrogen concentration is low, and utilizable carbon-and-energy sources are available. When provided with metabolizable carbon, along with ammonium, nitrate, or other combined nitrogen sources, they can grow under aerobic conditions. In either situation, they grow well on the salts of organic acids such as malate, succinate, lactate, or pyruvate (Krieg and Dobereiner, 1984). Azospirillum brasilense can use some carbohydrates, including fructose, galactose, and arabinose. Azospirillum lipoferum is also able to use these sugars, as well as glucose, mannose, and sorbose (Martinez-Drets et al., 1984). The most recently recognized species, Azospirillum amazonense, differs from the other two species in that it can grow on sucrose and other disaccharides (Martinez-Drets et al., 1985). Both A. brasilense and A. amazonense can synthesize their own biotin, whereas A. lipoferum can only grow if exogenous biotin is available (Falk et al., 1985).

Microaerophilic culture conditions for dinitrogen fixation can be established by culturing azospirilla in media containing 0.05% (wt/vol) agar. The bacteria grow and form a pellicle slightly below the agar surface, where diffusion of 02 from the culture-vessel headspace balances the








5

uptake of 02 by the bacteria, allowing both cell respiration and protection of the oxygen-sensitive nitrogenase (Okon et al., 1976a). Broth cultures can fix dinitrogen if the dissolved oxygen level is well controlled (Okon et al., 1976a). The bacteria are also able to grow on the surface of nitrogen-free, aerobically incubated agar plates (Day and Dobereiner, 1976).

The flagella of azospirilla enable them to move to

whatever sites their physiological state demands. They have been shown to exhibit aerotaxis to microaerophilic sites (Barak et al., 1982). Alternatively, cells may aggregate, thereby creating a microaerophilic environment by the respiration of many cells in a small space (Barak et al., 1982). The grass rhizosphere may contain microaerophilic sites (reviewed by Patriquin et al., 1983). Azospirilla could migrate from soil toward such sites, where nitrogenase activity could subsequently be expressed.

The respiratory metabolism of Azospirillum spp.

includes the ability of many strains to denitrify, reducing nitrate or nitrite to more reduced nitrogenous compounds under anaerobic conditions if enough metabolizable carbon source is available (Neyra and Dobereiner, 1977; Neyra and van Berkum, 1977; Nelson and Knowles, 1978). Under certain laboratory conditions, denitrification has been shown to provide enough ATP to support anaerobic growth of azospirilla (Bothe et al., 1981; Zimmer et al., 1984). The









6

ATP derived from denitrification can be used to drive nitrogenase activity (Scott et al., 1979), but it seems unlikely that dinitrogen fixation under these conditions can support growth of the bacteria (Bothe et al., 1981). Recent work by Neuer et al. (1985) has shown that, in axenic wheatAzospirillum spp. associations, both dinitrogen fixation and denitrification can occur.


Morphology of Azospirilla


Azospirilla are Gram-negative bacteria (Tarrand et al., 1978). The structural layers external to the cytoplasmic membrane of Gram-negative prokaryotes have been reviewed (Costerton et al., 1974). Depending on cultural conditions and the bacterial strain, polysaccharide or capsular layers may be present as the outermost layers of the cell.

A growing Gram-negative cell divides by binary fission to produce two daughter cells of approximately equal size. Division begins with invagination of the cytoplasmic membrane and peptidoglycan, until a complete transverse septum or cross wall is formed. When the septum is completely formed and cleaved, the two daughter cells separate (Leive and Davis, 1980). As will be discussed later, this cell division process can be disrupted, resulting in formation of filaments or chains, which accounts partially for pleomorphism of azospirilla.











Cells of Azospirillum brasilense and Azospirillum

lipoferum have a similar appearance when cultured in broth containing combined nitrogen. They are short, plump, slightly curved rods averaging 1.0 *m in diameter and 2.1 to

3.8 *m in length. They are motile in broth by means of a single, polar flagellum (Tarrand et al., 1978).

Cells of azospirilla often contain granules of the

polymer PHB (Krieg and Dobereiner, 1984). Grown with combined nitrogen, Azospirillum brasilense Sp 7 (ATCC 29145) has 0.5% to 1.0% of its dry weight as PHB. When grown in dinitrogen-fixing conditions, the PHB content rises until as much as 25.0% of its dry weight is PHB (Okon et al., 1976b). Granules of PHB are present in cells grown on combined nitrogen, but granule size and number are reduced compared to that found in dinitrogen-fixing cells (Albrecht and Okon, 1980).

Under certain cultural conditions, cells of azospirilla produce an outermost layer of capsular polysaccharide. When grown on an agar medium containing peptone, succinic acid and ammonium sulfate at 370C for 48 to 72 hours, a small proportion of cells are Gram-variable, possibly because they possess capsules. On this medium, A. brasilense exhibits more Gram-variability than does A. lipoferum. When cells of either species are cultured in the broth form of this medium, they stain uniformly Gram-negative, at least in young cultures (Krieg and Dobereiner, 1984).











Prokaryotic Exopolysaccharides


Many genera of both Gram-positive and Gram-negative bacteria include species that can produce polysaccharide layers outside their cell wall. Such layers have been referred to as capsules, exopolysaccharides (Sutherland, 1977) or glycocalyces (Costerton et al., 1981). Depending on laboratory cultural conditions, these polymers can assume different forms. Slime layers adhere loosely, if at all, to the cell and can often be separated from cells by centrifugation. Capsular layers appear to be tightly bound to the cell itself, and cannot be easily separated from cells. Microcapsules are so thin that their presence outside the cell wall cannot be observed using staining and light microscopy, while macrocapsules are of sufficient width to be so resolved (Ward and Berkeley, 1980).

Although proteins are sometimes present in bacterial capsules, most capsules are mainly polysaccharide in composition. The polysaccharides are extensively hydrated, and up to 99% by weight of the capsule is accounted for as water (Costerton et al., 1981).

The ATP needed to activate sugar residues for exopolysaccharide synthesis has been shown to comprise a significant proportion of total-cellular-ATP demand for some bacteria. Even when the carbon supply is growth limiting, some strains of bacteria produce extracellular polysaccharides (Jarman and Pace, 1984).









9

For many bacteria, a culture medium having a high

carbon to nitrogen (C/N) ratio promotes capsule formation (Sutherland, 1977; Costerton et al., 1981). Some species manufacture exopolysaccharide throughout all phases of growth, while others produce it only at certain stages of growth (Sutherland, 1977). Exopolysaccharides of more than one composition can be formed by the same bacterium under different environmental conditions (Geesey, 1982).

In laboratory culture, exopolysaccharides may be

nonessential for bacterial growth. Enzymatic removal of capsules often causes no reduction in viability of the decapsulated cells (Dudman, 1977). Nonencapsulated mutants may grow better in laboratory culture than do encapsulated cells, since they expend no energy for capsular synthesis (Costerton et al., 1981). Many nonencapsulated laboratory strains are mutants that have lost the ability of the wild type to produce exopolysaccharide. In other instances, common laboratory media have too low a C/N ratio to promote exopolysaccharide synthesis.

Attachment of bacteria to surfaces by their exopolysaccharides is the rule in nature, whether the surface is an inert mineral particle or a biological surface such as a plant root (Costerton et al., 1981). Natural environments are far different from laboratory cultural conditions, containing many more potential hazards to bacterial survival. In natural environments, the presence of








10

exopolysaccharides may aid the survival of bacteria (Dudman, 1977). Exopolysaccharides can concentrate nutrients from the surrounding solution phase. They give some bacteria increased resistance to antibiotics, surfactants, and other chemicals, as well as deterring their engulfment by phagocytic cells (Costerton et al., 1981). Other advantages of exopolysaccharides have been suggested, such as mediation of gas exchange between bacteria and their surroundings, but they have proven difficult to prove experimentally. Extracellular enzymes might also be located within or at the surface of capsules (Geesey, 1982).

Nur et al. (1980) found that A. brasilense Sp 7 and an Israeli isolate of A. brasilense both possessed small capsules discernible by electron microscopy when grown on nutrient agar. Umali-Garcia et al. (1980) found that when certain A. brasilense strains and grass seedlings were incubated together for 10 to 30 min at 300C, many bacteria adhered to the grass roots, with granular material accumulating on the surfaces of root hairs, and fibrillar material accumulating on the surfaces of older, epidermal root cells. It is known that bacterial exopolysaccharides may appear either granular or fibrous (Foster and Bowen, 1982). The A. brasilense strains seemed to rapidly produce both types of exopolysaccharides in axenic association with grass roots. After 2 to 4 days of axenic association with grass roots, from two to four cells of A. brasilense Sp 7








11

were sometimes seen to be enclosed within a common envelope or capsule. Such structures were not observed when the bacteria were grown in trypticase soy broth (Umali-Garcia et al., 1980). This is another indication that the low C/N ratio of complex broth media can repress extensive capsule formation by azospirilla, while the high C/N ratio near plant roots can promote capsule formation.

Recent work by Sadasivan and Neyra (1985) verified that the forms of carbon and nitrogen made available to azospirilla can have a profound effect on exopolysaccharide synthesis. When A. brasilense Sp 7 and A. lipoferum Sp. 59b (ATCC 29707) were cultured in broth containing 8.0 mM fructose and 0.5 mM KNO3, they grew as individual motile cells for only 6 hours and then started to clump, as exopolysaccharide production led to floc formation. Organic acids yielded fewer flocs than did sugars, and other nitrogen sources, such as ammonium, yielded fewer flocs than did nitrate. The cells in flocs appeared initially to be enmeshed in a loose, fibrillar matrix that condensed progressively over a week's time. When cells were grown, harvested by centrifugation, and resuspended in broth lacking carbon, the cells remained freely suspended. This suggests that azospirilla have a high ATP demand for exopolysaccharide synthesis. Chemical analysis showed that cellulose was a major component of the exopolysaccharide.









12

Fresh flocs were not degraded by cellulase, indicating that more than one type of exopolysaccharide was present.


Poly-B-Hydroxybutyrate (PHB)


In a constant and favorable environment where all

nutrients are present in sufficient amounts, bacteria grow for a time in a steady state, where every component of the cell culture increases by the same constant factor per unit time. This is balanced growth, and occurs during the logarithmic phase of the growth curve (Ingraham et al., 1983). If one or more nutrients become limiting, balanced growth is not maintained. When the carbon or energy supply is in excess, so that one or more other nutrients limit growth, some microorganisms respond by synthesizing and accumulating intracellular polymers having an energy-storage function (Dawes and Senior, 1973).

The cell catabolizes these polymers when the energy supply from exogenous sources is no longer sufficient to maintain processes needed for maintenance of cell viability. These processes may include osmotic regulation, maintenance of intracellular pH and transmembrane potentials, and turnover of cellular constituents such as proteins and nucleic acids. The energy required for these processes is called the energy of maintenance. Some microorganisms do not produce special energy-storage polymers. Faced with a starvation environment, they are forced to utilize their own








13

basal components, such as proteins and RNA, for energy sources. Possession of energy-storage polymers can benefit some species facing starvation, in that they degrade these polymers instead of or before they are forced to degrade such essential components as proteins (Dawes and Senior, 1973). However, different microorganisms utilize common constituents at different rates and in different sequences when starved. The possession of energy-reserve polymers does not spare degradation of protein and other basal components in some species during starvation. Most microorganisms that remain viable after prolonged starvation have a low endogenous metabolism, matched closely to their maintenance energy requirements. Starved microorganisms that rapidly metabolize polymers generally lose viability quickly (Dawes, 1976).

Three main types of microbial energy-storage compounds are known. Some species can accumulate more than one. All of these compounds have high molecular weights, and only a slight effect on the internal osmotic pressure of the cell. The amount of each compound a cell accumulates can vary widely, depending on environmental conditions.

Intracellular polyphosphates and glycogen-like polysaccharides are two types of energy-storage compounds formed by some eukaryotic and prokaryotic microorganisms. The synthesis of both types requires ATP.








14

The third microbial energy-reserve polymer is poly-Bhydroxybutyrate, a straight chain homopolymer of D(-)-3hydroxybutyrate. It is found only in prokaryotic cells, including both Gram-positive and Gram-negative species. Its synthesis requires reducing power in the forms of NADH or NADPH, but does not require the direct expenditure of ATP (Dawes and Senior, 1973).

With phase contrast microscopy, large accumulations of PHB within bacterial cells appear as light-refractile granules. A single granule may contain several thousand PHB molecules (Dawes and Senior, 1973). Each granule is bounded by a nonunit-membrane layer, which is probably formed from the cytoplasmic membrane. Presumably the enzymes for polymerization and depolymerization of PHB are present in this membrane layer (Shively, 1974).

Many of the Azotobacteraceae accumulate PHB when grown under dinitrogen-fixing conditions. There can be a wide variation in PHB content between species and between strains of the same species (Stockdale et al., 1968). The regulation of PHB levels in Azotobacter beijerinckii has been extensively studied and may provide clues to the role of PHB in-the physiology of other free-living, dinitrogen-fixing bacteria, such as azospirilla.

The route of PHB biosynthesis in A. beijerinckii has been reviewed by Dawes (1981). The synthesis and degradation of PHB in this microorganism are intimately associated








15

with intermediates and enzymes of the tricarboxylic acid (TCA) cycle, a system that azospirilla also possess (Okon et al., 1976b). When A. beijerinckii strain N.C.I.B. 9067 was cultured as a dinitrogen-fixer with 2.0% (wt/vol) glucose, PHB was deposited towards the end of exponential growth. The cells were unable to use all the available glucose, and PHB synthesis continued during the stationary phase until up to 74% of cell dry weight was PHB. Cultures grown with combined nitrogen rarely contained more than 3.0% of their dry weight as PHB (Dawes, 1981).

The initiation of PHB synthesis in the A. beijerinckii strain in batch culture coincided with the attainment of zero-oxygen concentration. Oxygen limitation was thus suspected to be a critical factor in initiating PHB synthesis. However, the nature of batch broth culture made it hard to separate oxygen effects from possible nitrogenlimitation effects (Senior and Dawes, 1971). Later experiments, using chemostat cultures having carbon, oxygen, or nitrogen limitation, clearly showed that extensive PHB accumulation only occurred under conditions of oxygen limitation (Dawes, 1981).

Before the studies reviewed by Dawes (1981), PHB was regarded as being only an endogenous, carbon-and-energy source that benefited cells during starvation. These experiments suggested that PHB could also serve other purposes. The synthesis of PHB seemed to serve as an









16

electron sink for excess reducing power (NADH and NADPH) that accumulated when the cell became oxygen limited, and electron transport to oxygen via the terminal oxidases of the electron-transport chain was restricted (Senior and Dawes, 1971). Later work revealed that the activities of certain enzymes of carbon catabolism in A. beijerinckii are inhibited by either or both NADH and NADPH. Under oxygen limitation, the concentration of these reduced coenzymes is increased, so that glucose metabolism, operation of the TCA cycle, and net biosynthesis are decreased. Growth can continue at some level, however, if PHB is synthesized and the crucial coenzymes are reoxidized (Dawes, 1981).

The synthesis of PHB under oxygen limitation may occur in other bacteria as well (Okon and Hardy, 1983). The quantity of PHB accumulated often greatly increases as the C/N ratio of the growth medium increases. Under such conditions, free-living dinitrogen-fixers may assimilate the exogenous carbon more rapidly than they can produce reduced nitrogen. As a result, the cells can accumulate large amounts of PHB (Stevenson and Socolofsky, 1966; Dawes and Senior, 1973). The metabolism of PHB is regulated such that PHB accumulates when the supply of exogenous carbon is in excess of the requirements for growth and maintenance, and it is degraded when the supply of exogenous carbon is limited or exhausted (Dawes, 1981), or when balanced growth can again occur (Nickels et al., 1979).









17

It has been shown that PHB can accumulate in cells that are not growing or proceeding toward cell division, due to limitation of available nutrients (Dawes and Senior, 1973). Nickels et al. (1979) demonstrated this in laboratory microcosms containing oak leaf detritus and estuarine water. Supplementing the nutrients in the water column with carbohydrates, especially glucose, induced a rapid accumulation of PHB without a concomitant increase in microbial biomass. When supplements were added that enabled increases in microbial biomass, PHB levels fell as the polymer was broken down to aid microbial growth.

In one study, A. brasilense Sp 7 was grown in batch

cultures for up to 14 days in microaerophilic, nitrogen-free malate broth (Papen and Werner, 1980). Both nitrogenase activity and PHB synthesis were biphasic. An initial peak of PHB content occurred at day 3, 1 day before the first peak of nitrogenase activity. During the first and maximal peak of nitrogenase activity, there was a decrease in PHB content, possibly due to accumulation of fixed nitrogen allowing use of PHB carbon skeletons for biosynthesis. A second peak of PHB accumulation occurred after the first maximum of nitrogenase activity. The results suggested that A. brasilense Sp 7, like A. beijerinckii, may accumulate PHB when it assimilates exogenous carbon faster than it can fix dinitrogen.








18

Zimmer et al. (1984) found that A. brasilense Sp 7 accumulated PHB when using nitrite as terminal electron acceptor for anaerobic growth. A maximum of 38% of cell dry weight was found to be PHB when less than 3.0 mM nitrite was present. No PHB was accumulated when in excess of 8.0 mM nitrite was made available, indicating the role of PHB as a sink for excess reducing power when other electron acceptors are scarce. It was also found that PHB-rich cells contained less protein than did PHB-poor cells.

Azospirillum lipoferum strain Br 17 (ATCC 29709) was found by Volpon et al. (1981) to accumulate nearly 24% of its dry weight as PHB near the mid-logarithmic phase of growth as a dinitrogen-fixer. Near the end of logarithmic growth, PHB synthesis seemed to stop, and the content of PHB declined to 13% of cell dry weight in stationary phase.

The PHB metabolism of A. brasilense strain Cd (ATCC 29729) has received considerable study. When this strain was grown in continuous chemostat culture with malate and ammonium chloride, a maximum PHB content of 12% of the biomass was observed under microaerophilic conditions and at intermediate growth rates (Nur et al., 1982). These growth conditions were said to approximate conditions generally encountered in the rhizosphere. The production of PHB was markedly decreased at higher levels of oxygen and higher growth rates. Once again, it was observed that cells








19

containing high amounts of PHB contained less protein than PHB-poor cells.

Recent work by Tal and Okon (1985) has further

delineated the roles PHB may play in the physiology of A. brasilense strain Cd. Grown in aerobic batch culture with malate and 2.8 mM NH4C1, the cells accumulated 40% of their dry weight as PHB after 24 hours, toward the end of exponential growth. When the level of NH4C1 was raised to 15.0 mM, the cells accumulated only 5% of their dry weight as PHB after 24 hours. In both cases, the amount of PHB decreased in stationary phase.

In chemostat continuous culture, a maximum of 30% cell dry weight accumulated as PHB when the gas atmosphere was

0.082 mM 02 (Tal and Okon, 1985). With increasing aeration, the PHB content fell to very low levels. When grown in batch culture as dinitrogen-fixers, the cells accumulated about 75% of their dry weight as PHB. Maximal PHB content was obtained in these experiments when the C/N ratio was about 70. Both the C/N ratio of the medium and the oxygen concentration were found to regulate PHB synthesis.

The forms of carbon and nitrogen made available to the cells affected the levels of PHB accumulated (Tal and Okon, 1985). Organic acids, especially pyruvate, were found to elicit PHB formation more than carbohydrates did. Sodium nitrate was found to promote PHB formation more than ammonium chloride did, possibly because nitrate does not









20

accumulate in the cytoplasm to the same extent as does ammonium.

Cells with different contents of PHB were harvested by centrifugation and resuspended in phosphate buffer to measure viability during aerobic nutrient starvation. By 140 hours, bacteria with abundant PHB reserves had given rise to more than twice as many viable cells as were present in the initial inoculum (Tal and Okon, 1985). During starvation, PHB reserves were degraded quickly but not completely. The initial inoculum contained 40% of its dry weight as PHB. This fell rapidly to about 24% of cell dry weight after 42 hours of starvation. After 130 hours of starvation, the PHB content of the cells was about 20% of cell dry weight.

In comparison, cells initially containing only 5% of their dry weight as PHB had only 7% of the original number of viable cells after 130 hours of starvation (Tal and Okon, 1985). Poly-6-hydroxybutyrate was still measurable throughout starvation of these PHB-poor cells, stabilizing at or near 3% of the dry weight of all cells present.

Starved PHB-rich cells had a higher respiration rate during starvation than the PHB-poor cells (Tal and Okon, 1985). Unlike cells having low amounts of PHB, the PHB-rich cells exhibited nitrogenase activity in the absence of exogenous carbon sources. But the PHB-rich cells were as unable to reduce nitrate anaerobically as were the PHB-poor cells in the absence of exogenous carbon.








21

This study (Tal and Okon, 1985) also suggested that

elevated PHB levels at the onset of starvation may spare the use of protein to drive endogenous metabolism. The PHB-poor cells used up two-thirds of their initial protein during the first 80 hours of starvation, whereas the protein content of starved PHB-rich cells increased slightly over 80 hours. It was also reported that PHB-rich cells were able to survive a variety of environmental stresses, including desiccation, better than PHB-poor cells (Tal and Okon, 1985).

The previous study also found that cells enriched in

PHB displayed a one hundred-fold higher aerotactic response than PHB-poor cells. This supports the claim made in an earlier study that PHB reserves could be used for aerotaxis when no exogenous carbon source was available (Barak et al., 1982).

The previous discussion has shown that both capsule and PHB synthesis can be promoted by environments with high available C/N ratios. The roles of capsules and PHB in pleomorphism in azospirilla will be discussed later. The nature of dormancy in prokaryotic cells will be discussed first, since some pleomorphic forms of azospirilla may be dormant stages. Capsular layers and PHB are often present in dormant forms of prokaryotes.








22

Dormant Forms of Prokaryotic Cells


There is general agreement that most soil bacteria spend much of their existence in soil in a state of low metabolic activity. The low respiratory rates of bulk samples of nonamended soil support this (Clark, 1967). Many soil bacteria may be metabolically dormant due to a lack of readily available carbon and energy supplies (Gray and Williams, 1971). Soil bacteria may enter into exogenous dormancy, where growth is delayed by unfavorable physical or chemical conditions (Marshall, 1980). Such bacteria probably have the same morphology as actively growing vegetative cells (Gray and Williams, 1971). These cells are probably intimately associated with the clay or organic matter of soil. The cells adsorb to these surfaces by physical or chemical interactions, or by the use of exopolysaccharides (Stotzky, 1980).

However, many bacteria may exist in soil as dormant

forms that are morphologically different from their growing, or vegetative, stages. These cells would have entered a phase of constitutive dormancy, involving the formation of spores or cysts (Marshall, 1980). Bae et al. (1972) used transmission electron microscopy to study thin sections of bacteria released from soil by centrifugation and washing. About 28% of the bacteria observed had normal vegetative morphology, of which 29% possessed capsular layers.








23

Bacterial cells resembling cysts comprised 27% of cells observed.

Sudo and Dworkin (1973) reviewed the kinds of prokaryotic resting cells that were recognized at the time. Bacterial resting cells were defined as cells in which division does not occur, and endogenous respiration is absent or greatly reduced. Usually resting cells are more resistant to environmental stresses than are vegetative cells; resting cells are often morphologically different as well. Such resting cells often differ in chemical composition from vegetative cells (Keynan, 1972). There are often either qualitative or quantitative differences between the electron transport systems of vegetative and resting cells. Many resting cells, for some period after they have germinated and resumed growth, are self-sufficient in energy sources, metabolites, and macromolecular precursors.

Perhaps the best understood bacterial resting cell

stages are the endospores of bacilli and clostridia. Cysts differ from endospores in that they are formed by the modification of an entire vegetative cell. The vegetative cell rounds up during encystment and becomes coated with one or more layers, often exopolysaccharide, external to its cell wall. No cyst forms withstand the extremely high temperatures tolerated by endospores, but they are comparably resistant to other environmental stresses (Sudo and Dworkin, 1973).








24

Certain properties are shared by all cyst-like forms of prokaryotes. They are formed when the growth rate of vegetative cells declines (a metabolic shift-down), due either to nutrient depletion or transfer of cells to an environment where balanced growth can no longer occur. Cells encountering these conditions complete their ongoing synthesis of DNA and chromosome replication but do not initiate new rounds of DNA synthesis, since growth has ceased (Sadoff, 1975). Conditions that will prohibit further growth promote the formation of dormant cells that can survive stress better than vegetative cells. These dormant cells often contain PHB or other energy-reserve polymers, and thickened cell walls or capsular layers. They have enhanced resistance to irradiation, sonic vibration, and sometimes elevated temperatures. Perhaps the most important traits for survival of dormant cells in natural environments are their resistance to starvation, low endogenous respiration rates, and desiccation resistance. Many cells entering constitutive dormancy need time to mature before they achieve maximal resistance to stress. It is important to remember that resting cells formed in natural environments may differ qualitatively and/or quantitatively in their resistance properties from those formed under laboratory conditions (Sudo and Dworkin, 1973).

Many strains of Gram-negative myxobacteria form dormant cells, called microcysts, when nutrients become limiting.








25

Nonmotile, encapsulated, mature microcysts are more resistant than are vegetative cells to environmental stresses such as ultraviolet irradiation, sonic vibration, and desiccation (Sudo and Dworkin, 1973).

Many Gram-negative, methane-oxidizing bacteria isolated from soil or mud are encapsulated and accumulate PHB when nitrogen becomes limiting for growth (Whittenbury et al., 1970b). Depending on the genus and strain, up to 90% of the cells present may form resting cells upon entering the stationary phase of growth. Lipid cysts of Methylocystis parvus accumulate large amounts of PHB and survive starvation and desiccation better than vegetative cells, but lack well-defined, capsular cyst coats. Methylomonas spp. and Methylococcus spp. form rounded, nonmotile cells that survive starvation better than do vegetative cells. These cells are called immature cysts, because they never attain desiccation resistance. Some strains of Methylobacter spp. form starvation-resistant and desiccation-resistant cysts that seem morphologically identical to Azotobacter spp. cysts (Whittenbury et al., 1970a).

Bdellovibrio sp. strain W is the only bdellovibrio that is known to encyst. Bdellocysts are larger than their vegetative counterparts and are not light-refractile. They tolerate sonic disruption, ultraviolet irradiation, and carbon starvation better than do vegetative cells. The endogenous respiration rate of bdellocysts is 80% less than








26

that of vegetative cells. When dried over silica gel desiccant under slight vacuum in glass tubes, vegetative cells of strain W die out rapidly and entirely. From 45% to 80% of bdellocysts initially present are able to survive 6 days of this desiccation treatment (Tudor and Conti, 1977). Bdellocysts possess a thickened outer layer of modified peptidoglycan, and contain inclusion bodies of an amylopectin-like polysaccharide of glucose monomers. These features are not found in vegetative cells (Tudor, 1980).

Some strains of Azotobacter spp., apparently some

azospirilla, and the bacteria described above are the only prokaryotes reported to form cysts. Why do not more bacteria possess resting stages that are morphologically differentiated into cysts? Perhaps growth media and conditions used in the laboratory discourage cyst formation (Whittenbury et al., 1970b). It is also possible that the ability to form cysts is sometimes labile and may be lost upon subculture. One Methylobacter chroococcum strain was able to form multiple-bodied cysts upon initial isolation from the environment. It ceased to do so when subcultured. Other Methylobacter spp. have retained the ability to form single- and multiple-bodied cysts over several years of subculture (Whittenbury et al., 1970b).

A mature, cyst-like cell of a prokaryote may perhaps best be characterized as follows. Mature cysts differ morphologically from vegetative cells in having thickened









27

outer layers. They are nonmotile and have low endogenous respiration rates. They only initiate growth into vegetative cells when sufficient nutrients are available. They must also possess more resistance to some environmental stresses than do vegetative cells. Enhanced resistances to starvation and desiccation are probably traits of all mature cysts. The cysts of the methane-oxidizing bacteria and of Azotobacter spp. possess these characteristics. Mature cysts of azospirilla should also have these properties.

Desiccation resistance is a critical characteristic of prokaryotic cysts. The next section will consider experiments conducted to assess the resistance of bacteria to drying.


Resistance of Bacteria to Drying


Clark (1967) stated that the majority of soil bacteria survive in air-dried soils, often for several years. When such soils are rewetted, bacterial activities including nitrification, ammonification, nonsymbiotic dinitrogen fixation, and sulfur oxidation are usually detected. The implication is that the intimate association of bacteria with clay or organic matter allows bacterial survival at hydrated microsites in a macroscopically dry soil. Later findings, reviewed by Stotzky (1980) and Marshall (1980), support this. Exopolysaccharides may help bacteria to achieve such intimate association, although capsules








28

themselves have not been found to afford any desiccation resistance in laboratory studies with pure cultures in nonsoil conditions (Dudman, 1977).

Because of the importance of desiccation as a limiting factor in legume inoculation with Rhizobium spp., several studies have been done on their resistance to drying. There are broad strain differences in resistance of rhizobia to desiccation. Many variables are present in drying experiments, and the variables may interact with one another. Rhizobium spp. withstand drying best in heaviertextured soils, where hygroscopic water can be retained by colloidal surfaces. Die off is far more rapid in drying sand. Capsules do not afford increased resistance to drying in studies with soil or other drying surfaces (Lowendorf, 1980). Often fewer rhizobia survive rapid drying procedures, such as oven drying, than survive milder desiccation over several weeks' time with controlled relative humidities (Jansen van Rensburg and Strijdom, 1980).

Robinson et al. (1965) added pure cultures of

Pseudomonas spp. or Arthrobacter spp. to sterile soils. The inoculated soils were dried by passing filtered air through them for 2 days, by which time they had reached constant weight. This forced drying resulted in rapid die off for both species. Labeda et al. (1976) found that slow evaporative drying of inoculated soil resulted in reduced death rates for both Pseudomonas spp. and Arthrobacter spp.









29

These experiments show clearly that rate of drying can profoundly affect the survival of vegetative bacteria. Vegetative cells with the capacity to become desiccation resistant may need time to alter their membrane or cytoplasmic composition before desiccation resistance is achieved. Fast-drying procedures may not allow them to do so. A differentiated resting cell, such as a cyst, may also need time, depending on how mature it is, to become desiccation resistant.

Relative humidity (RH) also has a great influence on desiccation resistance of prokaryotes. In desiccation at any RH below 90%, the free water of the cells is removed almost instantaneously. The water that remains is the bound water content of the cell, which may be necessary for continued function of essential metabolic processes and viability. Often few vegetative bacteria die when desiccated above 70% RH, but die rapidly as the RH declines to 45% (Webb, 1965). Many desiccation studies have not defined the RH at which the cells were dried, making duplication of results difficult.

Thompson and Skerman (1979) tested the desiccation

resistance of vegetative cells of many strains and genera of the Azotobacteraceae. One milliliter samples of vegetative cell cultures were added to sterile porcelain beads, positioned above silica gel in glass bottles sealed with Parafilm. These desiccation units were stored at room









30

temperature, and at different times single beads were aseptically removed and placed in broth media. The bacteria were probably in stationary phase when added to the assemblies, but it is unlikely that many cysts were present even in stationary phase broth culture (Sadoff et al., 1971). The results were surprising; the majority of strains retained viability for 1 to 2 years of desiccation. This was true even for bacteria that have never been shown to form cysts.

Mature cysts of prokaryotes survive rapid desiccation on glass surfaces far better than do their vegetative counterparts, but rarely does all the encysted inoculum survive rapid drying. Cysts of methane-oxidizers retained 60% to 90% viability after 1 week (Whittenbury et al., 1970a), and bdellocysts retained 45% to 80% viability after 6 days (Tudor and Conti, 1977). This may mean that not all the encysted cells were fully mature when exposed to drying, even if they all appeared morphologically identical. Such quick-drying assays can be valuable in determining whether morphologically differentiated cells are truly cyst-like.

Differences in the desiccation resistance of Azotobacter spp. vegetative cells and cysts are usually determined by the method of Socolofsky and Wyss (1962). They impinged suspensions of either cell form on the surfaces of membrane filters. The filters were then transferred to dry adsorbent pads in Petri dishes and placed in








31

an incubator at 330C. This method is a slow-dying procedure. At different time intervals, the cells were washed from the membranes, and viability was determined by plating. Cysts of Azotobacter vinelandii ATCC 12837 lost little viability over a 12-day-period using this drying treatment, whereas 99% of the vegetative cells were killed by the end of the first day (Socolofsky and Wyss, 1962). As a result of the rapid die off of vegetative cells with this treatment, later studies considered cells of this strain to be cysts if they could withstand 4 days of desiccation on membrane filters (Stevenson and Socolofsky, 1966; Wyss et al., 1969).

None of these membrane filter studies specified the RH at which the membranes were dried, or how many cell layers were deposited upon the membranes. Webb (1965) pointed out that if bacteria are dried on filters to test their desiccation resistance, the cells must be applied in a monolayer to achieve consistent results. If more than a cell monolayer is on the filter, most of the cells in subsurface layers will not be dried or equilibrated with the water vapor of the environment.

Most desiccation resistance experiments have given ill-defined or incomplete conditions of drying. Such experiments have proven, however, that cysts are more desiccation resistant than are their vegetative counterparts.








32

Vela (1974) tested desiccation resistance of

Azotobacter vinelandii ATCC 12837 by allowing slow drying of the agar on which the cells were grown. Vegetative cells were grown on agar plates of Burk's nitrogen-free medium, with glucose as the carbon source. Cysts were obtained by growing the cells on the same agar, except that 0.3% (vol/vol) n-butanol was employed as sole carbon source. Dried agar films were then broken with sterile forceps and placed on the surface of Burk's agar medium containing glucose. Vegetative cells borne on these agar films remained viable for nearly 2 years of desiccation, whereas cysts borne on such films remained viable for 10 years or longer.

Desiccation tolerance of azospirilla has received some attention. Lakshmi et al. (1977) recovered azospirilla from several air-dried soils stored in the laboratory. Recovery was obtained from one of four sandy soils stored air-dry for 10 years. All of these soils had less than 0.5% organic matter. Heavier-textured soils with 1.0% or more organic matter consistently yielded isolates of azospirilla. Some of these heavier-textured soils had been stored air-dry for up to 15 years. It was suggested that organic matter aids the survival of azospirilla in drying soils, and that desiccation-resistant cells may be formed by these bacteria.

Jagnow (1982) did some work with an Azospirillum lipoferum strain isolated from maize roots. In field









33

inoculations using 8 x 108 CFU/g soil, azospirilla near or on roots survived better than those in soil distant from roots. When added to pots of soil containing grass and cereal plants, populations remained at 106 to 107 CFU/g soil, even after 70 days of drought. He speculated that the presence of roots, either living or dead, enhances the drought tolerance of the associated azospirilla. In laboratory studies using nonautoclaved soil microcosms, air drying of soil was found to kill greater than 99% of the initial Azospirillum lipoferum inoculum. In comparison, the indigenous bacteria were little affected by air drying. This perhaps indicates that, unless azospirilla added to soil are able to associate quickly with plant roots, they will soon die out if drought stress occurs.

The desiccation resistance of pleomorphic encapsulated forms of azospirilla has been studied. Lamm and Neyra (1981) studied A. brasilense Sp 7 and A. lipoferum Sp 59b, in addition to several strains of azospirilla isolated from roots of various grasses in New Jersey and New York. To obtain cyst-enriched cultures, cells grown in nutrient broth were harvested by centrifugation, then washed and resuspended in sterile 0.85% (wt/vol) NaCl. A 1.0 ml sample of cells was then spread plated as a lawn onto nutrient agar plates containing 2.0% (wt/vol) agar. Plates were incubated at 300C until the agar was dried into a thin film, often requiring a month. After 15 days of incubation, cyst-like








34

cells predominated. Photographs of cyst-enriched cultures showed that many vegetative cells were still present. To obtain cyst-free cultures, cells were grown in nutrient broth, then washed and resuspended in saline. These cells were then spotted onto sterile, predried nutrient agar films so that the added cells would dry completely on the agar film in 30 min at 300C. Agar films from each treatment were then cut with sterile scissors and aseptically transferred to vials containing silica gel. To test viability, the dried agar films were removed periodically from the vials, placed on nutrient agar plates, and incubated for 1 week at 300C. Vegetative cells did not survive the initial drying process. Cyst-enriched populations that survived the initial desiccation period remained viable for up to 15 months. Interestingly, cyst-enriched cultures of two root isolates were nonviable at time zero, when they were placed into the silica-gel vials (Lamm and Neyra, 1981).

Two aspects of this study deserve special comment.

Clearly, the cyst-enriched cultures did not receive the same drying treatment as did the vegetative cells. The cystenriched agar films were obtained by a slow drying process, and the vegetative cell agar films underwent rapid drying. It does not seem valid to compare their desiccation tolerance under these different conditions. Also, two strains that contained cyst-like cells of apparently mature morphology were not desiccation resistant. Perhaps they









35

were not able to attain physiological maturity under the experimental conditions.

Papen and Werner (1982) assessed the desiccation

resistance of cyst-like forms of A. brasilense Sp 7. Cells from dinitrogen-fixing broth cultures were diluted in sterile tap water and then impinged onto the surface of sterile 0.2 4m Millipore membrane filters under vacuum. Some of the filters were immediately placed on the surface of nutrient agar plates and incubated at 280C, whereas others were placed on sterile adsorbent pads in Petri dishes and dried at 370C until they were placed on nutrient agar plates. Desiccation-resistant cells were only present after the first peak of nitrogenase activity, when nonmotile, encapsulated spheres containing PHB predominated. Cells before and during the first peak of nitrogenase activity were motile vibrioids and did not survive the desiccation treatment. As a second peak of nitrogenase activity arose, motile, dinitrogen-fixing vibrioids emerged from the spherical capsules; these vegetative cells were again not desiccation resistant. More encapsulated, spherical cells survived 2 days of desiccation than 6 days, but it was not an order of magnitude difference. This again may be an indication that morphologically mature cysts are not necessarily physiologically mature.

The recent work of Sadasivan and Neyra (1985) employed another assay for desiccation resistance of cyst-like forms








36

of azospirilla. Azospirillum brasilense Sp 7 and A. lipoferum Sp 59b were studied. Large flocs of cells enclosed in exopolysaccharide were placed on Whatman No. 1 filter paper and air-dried for 30 min. They were then placed in a closed vial, without desiccant, and incubated at 300C for up to 6 months. Small pieces of dried flocs were transferred periodically to semisolid nitrogen-free malate medium and incubated at 34'C for 2 to 4 days, and growth, pellicle formation, and nitrogenase activity were observed. Cells in dried flocs remained viable for up to 6 months of drying.

No vegetative cell controls were dried and tested for viability in the above study. Although cells remained viable in dried flocs for up to 6 months, it is not known how many cells survived in a given amount of floc. It is not known whether the cells themselves were desiccation resistant, or only physically protected from desiccation by exopolysaccharides.

Tal and Okon (1985) claimed that PHB-rich cells of A. brasilense strain Cd were 10 times more desiccation resistant than cells having little of the polymer. No details of the test used for desiccation resistance were given.

Desiccation resistance studies can be difficult to

interpret. Comparing the desiccation resistance of vegetative cells to that of cysts may be less difficult than








37

comparing that of vegetative cells of different strains. Rapid drying on a glass surface should enable most mature cysts to remain viable, but not most vegetative cells. A glass drying surface should be less hygroscopic than are membrane filters or agar films. Rapid drying on a glass surface is a severe treatment, but it should reveal the presence of physiologically modified, stress-resistant cells, such as mature cysts.


Azotobacter Cysts


Discussion of the nature of Azotobacter spp. cysts is important, because this information served as the basis for the experiments with azospirilla reported in this study.

Like azospirilla, the Azotobacteraceae are

Gram-negative aerobes, often containing PHB granules. Many are motile by flagella. They all fix dinitrogen, and some, including Azotobacter spp., do so either at atmospheric oxygen levels (unlike azospirilla), or as microaerophiles. Only one genus, Azotobacter, contains species with strains that are known to form cysts (Tchan, 1984). The isolation of Azotobacter spp. from the interior of 2,000-year-old clay bricks (Abd-El-Malek and Ishac, 1966), and their persistence in soils that had been air dried from 10 years (Vela, 1974) to 30 years (Clark, 1967), may be due largely to cyst formation.









38

When grown in nitrogen-free broth with glucose as the carbon source, young cells of Azotobacter spp. appear as rods with rounded ends, ranging from 1.3 to 2.7 pm in diameter and 3.0 to 7.0 pm in length. As cultures age, cells often accumulate PHB. Cell morphology may be altered to ellipsoids, filamentous cells, or chains of cells (Tchan, 1984).

Azotobacter spp. are commonly isolated from soil and

aquatic habitats of near-neutral pH, and are generally less acid-tolerant than azospirilla. The most common species isolated from soil is Azotobacter chroococcum, but its biochemistry and physiology have received less attention than that of Azotobacter vinelandii (Tchan, 1984). Azotobacter vinelandii ATCC 12837 forms cysts profusely under appropriate growth conditions. When this strain is cultured in Burk's nitrogen-free broth with glucose, some cysts form in stationary phase cultures, but ony 1.0% (Lin and Sadoff, 1969) to 10.0% (Reusch and Sadoff, 1981) of the population encysts under these conditions.

Early workers such as Winogradsky (1938) knew that

growing some Azotobacter spp. in nitrogen-free media, with ethanol or butanol as carbon source, led to enhanced production of nonmotile, spherical cells with double-layered coats. Socolofsky and Wyss (1961) built upon this knowledge, using A. vinelandii ATCC 12837 (which was used in all the studies that follow unless otherwise indicated).








39

When cultured as cell lawns on Burk's nitrogen-free agar with 0.3% (vol/vol) n-butanol as sole carbon source, cysts began to appear within 3 days and predominated in 5 to 7 days. Ultrastructural studies revealed that the outermost layer of the cyst, the exine, consisted of several overlapping, plate-like layers. Beneath the exine was a much thicker layer of gelatinous material, called the intine. The intine surrounded a modified resting cell, called the central body, which often contained numerous PHB granules. Cysts had no detectable endogenous respiration when suspended in buffer, but almost instantaneously began measurable respiration when exogenous carbon sources were added. In later studies, cysts were produced by growth on

0.2% (vol/vol) n-butanol (Socolofsky and Wyss, 1962), or

0.2% (wt/vol) 8-hydroxybutyrate (BHB) (Lin and Sadoff, 1968), with 90% or greater of the cells being converted to cysts in 5 to 7 days.

Eklund et al. (1966) demonstrated that the formation of capsular layers by vegetative cells was a prerequisite for cyst formation. Complete morphological encystment of cells grown on n-butanol agar with various levels of NH4NO3 only occurred in the usual 5-day-period when the NH4NO3 concentration was 0.02 M or less. The cells rounded up within 5 days into nonmotile precysts lacking exines when 0.03 M or

0.04 M NH4NO3 was initially present. By day 10, these cells had used up enough of the original combined nitrogen to








40

allow dinitrogen fixation to resume, so that capsular polysaccharide was produced, followed by formation of exines and, ultimately, morphologically mature cysts. Nonencapsulated mutants were unable to form morphologically mature cysts. The work of Pope and Wyss (1970) emphasized that cells beginning encystment first produced a capsule that acted as a structure within which the cyst coats were built, so that the exine existed inside of the capsule. The diameter of morphologically mature Azotobacter cysts measured between exine boundaries is about 2.0 pm (Reusch and Sadoff, 1983).

Abortive encystment occurs when cells round up into nonmotile precysts, but are unable to form a complete exine. This occurs in the presence of high amounts of combined nitrogen (Eklund et al., 1966), when glucose or other carbon sources are present in addition to n-butanol or BHB (Lin and Sadoff, 1968), or when calcium is unavailable. The calcium requirement is probably related to its function as a stabilizing cation that holds the cyst coats together (Page and Sadoff, 1975). Using 3.0 mM EDTA in 0.05 M Tris buffer, pH 7.8, Lin and Sadoff (1969) obtained almost instantaneous expulsion of the central body from the cyst coats, due to the chelating effect of the buffer. The empty exines had the same "horseshoe" shape seen when cysts germinate, and vegetative cells separate from the exines.








41

The role of PHB in cyst formation was examined by Stevenson and Socolofsky (1966). Cysts were defined as cells that could survive desiccation on a membrane-filter surface for 4 days at 330C. After 2 days of growth on nitrogen-free n-butanol agar, cells lost their motility, became oval-shaped, and accumulated PHB to the extent of 35% of cell dry weight. The development of mature cysts was accompanied by a reduction in PHB content. By 6 days, cultures had undergone 100% encystment, and 10% of cyst dry weight was PHB.

Lin and Sadoff (1968) developed a two-step replacement procedure for obtaining cysts in broth. Cells were grown to late exponential phase in Burk's nitrogen-free broth with glucose. After harvest by centrifugation and washing in buffer, cells were resuspended in Burk's salts broth with

0.2% (wt/vol) BHB. This procedure was used in further studies (Hitchins and Sadoff, 1970, 1973; Reusch and Sadoff, 1979; Su et al., 1981; Reusch and Sadoff, 1983), resulting in the following detailed description of BHB-induced encystment.

After 1 hour in encystment broth, cells are still

motile and flagellated but no longer possess nitrogenase activity. Within 4 to 6 hours, DNA synthesis has ceased, and soon afterward each cell divides to form two nonmotile precysts. There is rapid accumulation of PHB during this period, and the rate of phospholipid synthesis declines.








42

Simultaneously, BHB is being taken up and respired or incorporated. From the sixth to sixtieth hour, unique lipids, not found in vegetative cells, begin to be produced. These include 5-n-alkylresorcinols (ARl) and their galactoside derivatives (AR2). These lipids possess hydrophobic alkyl sidechains and hydrophilic phenolic heads. Also produced are 6-n-alkylpyrones (AP), having a similar bipolar nature. During this time, membranous vesicles migrate outward from the central body through the intine to form the exine layer. Up to 17% of the exine is composed of ARI and AR2. The central body produces ARI and AR2 in part from its PHB reserves, and exports them in the membranous vesicles to the exine region. Radio-labelled BHB accumulates in the central body and exine, whereas the intine contains almost none. This indicates that the intine is composed mainly of capsular material, formed from cell reserves that were present before encystment is triggered by BHB. Net RNA synthesis stops by the twelfth hour, and net protein synthesis continues for up to 36 hours. Lipid turnover continues beyond 60 hours, but there is no net lipid synthesis. In a mature cyst, 5.0% of the central body membranes are phospholipid, with AR and AP composing the other 95%. Molecular models suggest that AR and AP form a more rigid membrane structure at physiological temperatures than do phospholipids. The hydrophobic, viscous nature of








43

such a membrane may contribute greatly to the desiccation resistance and dormancy of cysts.

The possible contribution of the central body membranes to stress resistance of cysts was suggested in earlier studies. The cysts of Azotobacter chroococcum strain 75-I had a compact, well-defined exine layer, whereas the exine of A. chroococcum strain NTS was diffuse and fragile (Vela and Cagle, 1969). The cysts of A. chroococcum strain 75-I were much more resistant to sonic disruption than cysts of A. chroococcum strain NTS. Yet cysts of both strains were comparably resistant to desiccation on membrane filters and to ultraviolet irradiation. Kramer and Socolofsky (1970) defined cyst germination of A. vinelandii ATCC 12837 as a process whereby desiccation resistance is lost; mature cysts were defined as cell forms surviving 3 days of desiccation on membrane filters. It was found that 40.0 pg chloramphenicol/ml inhibited outgrowth of cysts in a complete medium. Many cysts lost their desiccation resistance when incubated with chloramphenicol, indicating that the antibiotic might have chemically changed some essential cyst component, perhaps the central body's membranes. Hitchins and Sadoff (1973) found that, soon after exposure to BHB, vegetative cells became resistant to 100.0 pg chloramphenicol/ml. The antibiotic had no effect on morphogenesis or rates of protein synthesis. This is another indication of rapid membrane alteration of encysting








44

cells, long before AR and AP are produced. Further support for the importance of membranes may be found in studies where mineral nutrient deficiencies lead to the production of stress-resistant cysts which lack completed cyst coats (Gonzalez-Lopez et al., 1985).

Germination of cysts has usually been defined as the emergence of a growing, motile cell from the exine layer (Socolofsky and Wyss, 1961). Loperfido and Sadoff (1973) examined the germination of cysts exposed to glucose. Cysts respired detectably within 2 min. after the addition of 1.0% (wt/vol) glucose, and soon afterwards net synthesis of RNA and protein became measurable. After 4 to 6 hours, the central body had enlarged to occupy the volume of the intine, and DNA synthesis and nitrogenase activity became measurable. After 8 hours, a vegetative cell emerged from the cyst coats, leaving behind an empty "horseshoe"-shaped exine. Germination did not occur in the absence of oxygen. Cysts also germinated in the presence of sugars other than glucose. Germination did not occur in Burk's nitrogen-free salts, indicating that the PHB reserves of the cysts could not be mobilized to initiate germination. The addition of

0.25% (wt/vol) NH4+ did not lead to germination.

When cysts are germinated on glucose, some central bodies divide within their cyst coats to form multiple central bodies. Up to six central bodies have been observed within one cyst coat (Cagle and Vela, 1974).








45

Pleomorphism of Azospirilla


Bacteria cultured in vitro can be extremely

pleomorphic. Only a few cells in a population may exhibit abnormal morphology under some cultural conditions, but sometimes the majority of a culture assumes unusual shapes. Older cultures in the stationary growth phase can be especially pleomorphic (Duguid and Wilkinson, 1961).

Hughes (1956) has reviewed the development of bacterial filaments. Filamentous cells are usually as wide as normal cells, but are several times longer and lack developed septa. They are interesting because they are often fully viable, unlike some pleomorphic or involution forms of bacteria. Under suitable cultural conditions, a filament may divide at several points along its length to produce several cells of normal length. Filaments can be induced by sublethal cell damage, interruption of balanced growth, dyes and antibiotics, extremes of pH, refrigeration, and various forms of radiation.

Slater and Schaecter (1974) emphasized how sensitive bacterial cell division is to the factors mentioned above. If sublethally stressed, rod-shaped bacteria may continue to grow and form filaments. Filaments can also form during very rapid growth in rich media, and will fragment into individual cells when growth slows, or when the environment becomes less nutritionally rich. Since cells arising from fragmentation of filaments are usually of normal length, the








46

cell's ability to control the site of cell division is not lost during filamentation. Sometimes chains of cells occur instead of filaments. The cells in chains contain septa, but final cleavage between cells has not yet occurred. It is possible that contiguous cells having incomplete septa in such chains may share continuities between their cytoplasms. In some cases, chains may be held together by very thin capsular layers common to several cells in the chain.

Jensen and Woolfolk (1985) found that several strains of Pseudomonas putida and Pseudomonas fluorescens were induced to form filaments if oxygen became limiting during the late logarithmic phase of growth in nutrient broth. Exhaustion of one or more nutrients was also a probable elicitor of filamentation. The weakly motile filaments, unlike the highly motile aerobic rods of the bacteria, migrated to microaerophilic zones. As respiration of cultures declined, the increasing levels of oxygen in the broth seemed to trigger fragmentation of the filaments into rods. Cultures containing filaments, or the progeny of fragmented filaments, retained viability longer than nonfilamentous cultures.

Morphological changes in Escherichia coli have been related to specific genes. If cellular DNA is damaged by ultraviolet irradiation or other influences, several genes are expressed in the so-called SOS response. Many of the









47

gene products are involved directly in repair of damaged DNA, but some others specifically block further cell division. Until the DNA is repaired, cell division is blocked, but cells can continue to grow into long, nonseptate filaments. Upon repair of the DNA, septa form along the filaments, and cells of normal size are produced after septum separation (Donachie et al., 1984). Certain E. coli mutants are known to produce septa, but form chains because the enzymes needed for septum cleavage are not produced (Begg and Donachie, 1985).

Thompson and Skerman (1979) showed that most members of the Azotobacteraceae are pleomorphic under certain cultural conditions. Filaments and chains of cells are produced commonly. Similar pleomorphism has been observed with azospirilla.

Becking (1982) observed that the morphology of

azospirilla varied in different culture media. On yeast extract-glucose agar, the cells were highly motile, slightly curved rods, 2.0 to 4.0 pm long and 1.0 pm wide. These cells would often become swollen with three to five PHB granules per cell. When cultured in nitrogen-deficient broth supplemented with 0.01% (wt/vol) Difco yeast extract, the cells often became long spirals of 30 to 40 pm in length. These cells had reduced motility, but were capable of rotation about their axes, and had few or no PHB granules. Peptone was found to produce similar elongated,








48

weakly motile cells containing little PHB. These cells were probably filaments, as described by Hughes (1956). Becking did not study their viability.

Eskew et al. (1977) isolated and studied the pigmented A. brasilense strain Cd. Nitrogenase activity peaked after about 2 days of growth in semisolid, nitrogen-free malate medium, and most cells were motile, curved rods of normal size, often containing PHB. After 3 days, however, nitrogenase activity declined sharply. By this time, the initial near-neutral pH of the medium had risen to pH 8.1. Most of the bacteria present then appeared as enlarged, ovoid, nonmotile cells that were resistant to Gram-staining. The decrease in nitrogenase activity, and shift to alkaline pH, coincided with the appearance of cyst-like cells.

Tarrand et al. (1978) found that A. brasilense and

A. lipoferum strains had a similar appearance after 1 day's growth in broth containing peptone, ammonium sulfate, and succinate. Most cells were short, plump, slightly-curved motile rods averaging 1.0 pm in diameter and 2.1 to 3.8 pm in length. Cell morphology changed, especially for A. lipoferum strains, when the cells were inoculated into nitrogen-free, semisolid malate medium containing 0.005% (wt/vol) yeast extract. Cells of A. lipoferum tended to increase to 1.4 to 1.7 pm in width and to 5.0 pm to over 30 pm in length. Within 1 to 2 days, many A. lipoferum cells became S-shaped or helical and retained little if any









49

motility. These long cells eventually fragmented into shorter, ovoid cells. Many of these fragments later became large, pleomorphic cells filled with light-refractile granules, probably PHB. In contrast, A. brasilense strains transferred to nitrogen-free, semisolid malate medium initially retained their normal appearance. Only after several weeks' time in this medium did they develop some S-shaped cells and some large, pleomorphic, granule-filled forms. Falk et al. (1985) found that A. amazonense strains failed to become pleomorphic under comparable conditions.

Krieg and Dobereiner (1984) maintained that alkalinization of the medium due to oxidation of malate was responsible for pleomorphism in A. lipoferum. Cultures of this species grown in semisolid, nitrogen-free glucose medium did not become alkaline, and the cells did not become pleomorphic.

Wong et al. (1980) isolated a putative Azospirillum

sp. from cellulolytic, dinitrogen-fixing mixed cultures. In semisolid, nitrogen-poor malate medium containing adequate levels of biotin, the cells were of normal size and morphology after 1 day's growth. Between the third and seventh days, the cells gradually became S-shaped and enlarged. These enlarged cells contained granules of PHB and/or polyphosphate. By 10 days, many cells had lysed and released these granules into the medium. When the initial biotin concentration of the medium was reduced to 10% of the








50

normal level, these morphological changes were accelerated, occurring within 2 to 3 days after inoculation. This strain could not fix dinitrogen with glucose as the carbon source, but otherwise its biotin requirement and pleomorphism were typical of A. lipoferum.

Lamm and Neyra (1981) found that A. lipoferum strains grown in nitrogen-free, semisolid malate medium developed many elongated cells after 2 days of culture, whereas A. brasilense strains only developed elongated cells after 10 days. In both semisolid and agar-plate, nitrogen-free malate cultures, thick-walled, optically refractile cells were present as 1.0% of the cells in day-old cultures of both species. After 4 days, the numbers of these thickwalled cells were equal to or greater than cells of normal morphology. The ovoid cells of A. lipoferum strains were about twice as large as those produced by A. brasilense strains. Such cells were never observed in nutrient broth cultures, but could be obtained in old cell lawns grown on nutrient agar. The increased desiccation resistance of these ovoid cells has already been discussed.

Papen and Werner (1982) observed apparent cysts of A. brasilense Sp 7. Such nonmotile cells were encapsulated, having a diameter of about 1.2 pm, were not fixing dinitrogen, and were desiccation resistant. They composed the majority of cells in the nitrogen-free, malate broth culture between the second and fourth days of incubation. After








51

this time, vegetative vibrioids emerged from the capsules to grow and fix dinitrogen. The authors suggested that oxygen limitation greatly affected these events. The level of PHB increased as the oxygen level of the culture decreased; nitrogenase activity ceased; and the cells encysted for a time. Their apparent reduced respiratory activity allowed the level of dissolved oxygen to be replenished in the medium, until vibrioids emerged from the cyst coats to grow and fix dinitrogen again. No encystment was observed when cultures were incubated aerobically.

The recent work of Sadasivan and Neyra (1985) stressed the roles that PHB and exopolysaccharides play in cyst formation of azospirilla. Encysting cells lost their motility and became enlarged and rounded. They accumulated PHB and synthesized capsular material. The investigators emphasized that common media, such as nutrient broth, do not promote encystment and that development of mature exine and intine layers may only be achieved under specific, welldefined cultural conditions. Sadasivan (1985) may have found the cultural conditions to promote maturation of cysts of A. brasilense Sp 7. Using phase contrast microscopy, she has observed vegetative cells emerging from cyst coats, leaving behind empty "horseshoe"-shaped capsules. She has also observed cysts containing from two to four central bodies within a single exine. In transmission electron microscopy thin sections, she has observed maturing cysts,









52

with membranous blebs migrating outward into the capsular material from central bodies containing PHB granules. She has also observed mature cysts with central bodies containing PHB and polyphosphate granules, surrounded by distinct intine and exine layers. Thus, given appropriate cultural conditions, A. brasilense Sp 7 is able to form apparently mature cysts, almost identical in appearance to those of Azotobacter spp. One unusual feature she has reported is layers of spherical, melanin-like granules outside the exines of mature A. brasilense cysts; these layers have never been observed with Azotobacter spp. cysts.

Berg et al. (1980) studied morphological and physiological changes of A. brasilense Sp 7 grown under different conditions. Encapsulated cells (C-forms) were often present on cell lawns grown on nitrogen-free succinate agar. Encapsulation was initially heaviest for cells near the lawn surface. After most cells in the surface layers were converted to nonmotile C-forms, the lower cell layers began to accumulate capsules. Such C-forms were not observed within 60 hours of growth in semisolid, nitrogen-free succinate agar. They formed rapidly on nitrogen-free agar surfaces. Most of the culture formed capsules. The appearance of the encapsulated forms varied and changed with time. Both capsule formation and PHB accumulation were inhibited by combined nitrogen. As cultures aged, enlarged vibrioid C-forms developed. The more mature forms were spheres of









53

2.0 to 4.0 pm diameter which had lost their motility. One-week-old cultures consisted mainly of spherical C-forms of 5.0 to 8.0 pm diameter, containing many PHB-rich cells within a common capsule. The authors speculated that younger encapsulated forms may be fixing dinitrogen and that older encapsulated forms may not. They suggested that the capsule may reduce oxygen flow into the cells, thereby protecting oxygen-sensitive nitrogenase activity. Azospirilla form extensive capsules only in media having a high C/N ratio (Sadasivan and Neyra, 1985). Such conditions promote nitrogenase activity. Since most capsules contain over 99% of their weight as water (Costerton et al., 1981), and oxygen diffuses through water at one ten-thousandth the rate of diffusion through air (Clark, 1967), the capsule may well help protect nitrogenase from oxygen damage. Oxidation of PHB reserves within the cell may also reduce oxygen levels near the nitrogenase (Dawes and Senior, 1973).

The description by Berg et al. (1980) of encapsulation starting at the uppermost layers of nitrogen-free agar-grown colonies and proceeding downwards is reasonable, if one assumes that encapsulated cells are metabolically active for a time, and then pass into dormancy. Initially, the uppermost encapsulated cells would be actively fixing dinitrogen. They might become dormant as a result of underlying cell layers depleting the available carbon supply, or possibly because conditions become favorable for their passage









54

into constitutive dormancy (Marshall, 1980). Cells might start to encyst when they accumulate threshold levels of capsular material and/or PHB. In any case, as dormant cells they would consume little oxygen, allowing it to diffuse to lower cell layers that previously may have been oxygenlimited, due to the actively respiring upper cell layer. These lower cell layers would become more active with the increased oxygen supply, accumulating capsules. Eventually these cell layers would also pass into dormancy.

In earlier work, Berg et al. (1979) grew A. brasilense Sp 7 in association with sugarcane callus tissue. Vegetative cells (V-forms) grew as slimy colonies on the surface of the callus, and few of these V-forms contained PHB or capsules. Encapsulated or C-forms were also observed in these conditions. This association of azospirilla with sugarcane callus exhibited nitrogenase activity, but whether the V-forms, C-forms, or both were responsible could not be ascertained, since both were present. Perhaps C-forms were able to fix dinitrogen transiently, but were poised to enter dormancy if growth became too unbalanced. The bacteria did not possess capsules near or within lysed plant cells, where the C/N ratio may have been narrow, and balanced growth may have been promoted.

An apparent contradiction in this plant callus-bacterium work is the claim by Berg et al. (1979) that C-forms of azospirilla have little similarity to Azotobacter








55

spp. cysts. Krieg and Dobereiner (1984) restated this, but the photographs of Berg et al. (1979) do not support it. The multicellular C-forms are virtually indistinguishable from Azotobacter spp. cysts having multiple central bodies (Cagle and Vela, 1974). Clearly, in the association with sugarcane callus, the azospirilla were situated in numerous sites, differing in nutrient availability and oxygen availability. It is not surprising that multiple morphologies were observed, reflecting multiple physiological states. Only a few cells resembling mature cysts were present.

Pleomorphic forms of azospirilla have been observed in a variety of axenic associations with plant roots. The work of Umali-Garcia et al. (1980) has already been discussed. Ruscoe et al. (1978) grew maize plants in sand and inoculated them with different strains of azospirilla. Enlarged, cyst-like cells, as well as cells of normal morphology, were observed in older and thicker root segments, where root tissue was often disintegrating. They also found that when two strains of azospirilla were grown in nitrogen-free, semisolid trans-aconitate agar, they often formed long chains after 4 to 5 days.

Matthews et al. (1983) used immunological techniques and transmission electron microscopy to observe strains of A. brasilense in axenic association with pearl millet roots. Both vibrioid and encapsulated cells were observed in association with the roots. The encapsulated cells








56

usually contained PHB and polyphosphate granules, and often two or more cells were enclosed by a common capsule.

Patriquin et al. (1983) observed unusual structures on the surface of wheat roots, 3 weeks of age and older, that had been axenically incubated with azospirilla in a sandvermiculite mix. They appeared as spherical "bags," within which azospirilla containing PHB granules could be seen to swim about. These structures were also found between the epidermis and outer cortex of young wheat roots.

Krieg and Dobereiner (1984) suggest that the capsule of azospirilla helps protect nitrogenase. They also support the idea that development of alkaline pH is the cause for pleomorphism in A. lipoferum and A. brasilense. This seems an incomplete explanation, implying that pleomorphic cells are poorly viable, being aberrant forms or laboratory artifacts. Pleomorphic cells of azospirilla may instead develop commonly, and perhaps transiently, when growing in natural environments of high C/N ratio, such as near plant roots. Unbalanced growth, with increased PHB and capsule formation, may be the major cause of pleomorphism.


Experimental Approach


The conversion of 90% or more of an Azotobacter

sp. cell suspension to cysts facilitates physiological studies of cysts. Growing the cells in nitrogen-free media containing n-butanol or BHB leads to this conversion.









57

Azospirilla of cyst-like morphology have been observed under various cultural conditions, but reports of conversion of 90% or more of cell populations to cyst-like forms are not found in the literature. Vegetative cells are reported as being present in high numbers, along with the cyst-like forms. This has perhaps discouraged studies on the nature of cyst-like forms of azospirilla.

All strains of A. brasilense and A. lipoferum are able to grow on BHB as sole carbon source in the presence of combined nitrogen (Tarrand et al., 1978). However, no studies have been done to see how azospirilla respond to BHB in the absence of combined nitrogen. Since such cultural conditions lead to prolific encystment of some Azotobacter strains, it was considered worthwhile to determine if strains of azospirilla might also undergo conversion in high numbers to cyst-like forms under these growth conditions.

The research reported here addresses the following questions:

1. Can high numbers of cyst-like forms of azospirilla

be obtained by growth in nitrogen-free BHB broth

or on nitrogen-free BHB agar?

2. What are the morphological differences between

azospirilla grown on BHB with or without combined

nitrogen?









58

3. Is the PHB content of azospirilla grown on BHB

without combined nitrogen higher than when they

are grown in complex broth with combined nitrogen?

4. If pleomorphism of azospirilla occurs in nitrogenfree BHB broth, is alkalinization of the medium a

prerequisite for development of pleomorphism?

5. Are azospirilla grown on nitrogen-free BHB agar

more desiccation resistant than cells grown in

complex broth with combined nitrogen?

6. Are azospirilla grown on nitrogen-free BHB agar

more resistant to starvation in carbon- and

nitrogen-free, phosphate-buffered salts solution

than cells grown in complex broth with combined

nitrogen?

7. What growth conditions favor motile azospirilla

arising from nonmotile azospirilla grown on

nitrogen-free BHB agar?

8. Is protein synthesis required before nonmotile

BHB-grown azospirilla give rise to motile

azospirilla?

9. Are BHB-grown azospirilla affected by Tris-EDTA in

a manner similar to Azotobacter cysts?

Questions 1 through 4 are considered in Chapter II, and the remaining questions are considered in Chapter III.

















CHAPTER II
PLEOMORPHISM OF AZOSPIRILLA GROWN ON BETA-HYDROXYBUTYRATE


Only a few bacterial genera contain strains known to form cysts (Sudo and Dworkin, 1973; Whittenbury et al., 1970a; Tudor and Conti, 1977). A nonmotile cyst forms when the entirety of a vegetative cell rounds up, depositing extracellular coats and often accumulating intracellular energy-reserve polymers.

The morphological changes of encystment are accompanied by a reduction in cell metabolic activities, and increased resistance to environmental stresses, such as starvation and desiccation. Cysts of Azotobacter spp. are perhaps the best understood. Like other prokaryotic resting cells, they form when vegetative cells undergo a metabolic shift-down (Sadoff, 1975).

Cysts of Azotobacter spp. do not form in media supporting good vegetative growth until stationary phase, and are present then only in low numbers (Sadoff et al., 1971). Similarly, cells of azospirilla are uniform in shape during active growth in nutritionally complete media (Umali-Garcia et al., 1980; Lamm and Neyra, 1981; Sadasivan and Neyra, 1985). As is true for Azotobacter spp., however, stationary phase cultures of azospirilla grown on complete media often 59









60

contain some rounded, nonmotile cells (Lamm and Neyra, 1981; Papen and Werner, 1982; Krieg and Dobereiner, 1984).

Azospirilla are morphologically vexing, in that different pleomorphic cell types occur under various growthlimiting conditions. Weakly motile filaments containing little PHB form in aerobic broth which is low in combined nitrogen (Becking, 1982). Under dinitrogen-fixing conditions, filamentous or S-shaped cells again may arise but contain large deposits of PHB (Tarrand et al., 1978; Wong et al., 1980; Lamm and Neyra, 1981). These elongated cells often fragment into smaller, oval cells which subsequently can assume a cyst-like morphology (Tarrand et al., 1978).

The most frequently reported pleomorphic form of azospirilla is a nonmotile cell possessing thick outer layers, probably of capsular material. These cells usually contain more extensive deposits of PHB than do vegetative cells grown with combined nitrogen. These cells have been observed in older cultures grown on combined nitrogen (Lamm and Neyra, 1981), in cultures grown as dinitrogen-fixers (Eskew et al., 1977; Berg et al., 1979; Berg et al., 1980; Papen and Werner, 1982), and in axenic associations with grass roots (Ruscoe et al., 1978; Umali-Garcia et al., 1980; Matthews et al., 1983). Recently, Sadasivan and Neyra (1985) obtained them in broth containing fructose and KNO3.

The nomenclature for describing these cells is not

standardized. Berg et al. (1979) termed them capsulated or









61

C-forms as opposed to the vegetative or V-forms, as did some later workers (Matthews et al., 1983; Krieg and Dobereiner, 1984). This terminology may be confusing, however, since capsules can also occur on azospirilla of otherwise normal morphology (Nur et al., 1980).

The presence of a capsule is usually deemed a prerequisite for cyst formation in Azotobacter spp. (Eklund et al., 1966). Azospirilla also may need to form a capsule before they can form cyst-like cells. Encapsulated azospirilla may initially be fully active vegetative cells. Upon encountering metabolic or environmental stress, such cells may mature into cyst-like cells. The change in morphology with time of some members within a C-form population (Berg et al., 1980) may reflect maturation into truly mature cysts. Two definitive traits of a mature Azospirillum spp. cyst would be greatly reduced cell metabolism and enhanced desiccation resistance. Morphologically differentiated cells of azospirilla have been called cysts when they exhibit no nitrogenase activity (Eskew et al., 1977; Papen and Werner, 1982) or exhibit enhanced desiccation resistance (Lamm and Neyra, 1981; Papen and Werner, 1982; Sadasivan and Neyra, 1985).

Another complicating factor in understanding these

forms of azospirilla is that their appearance in dinitrogenfixing cultures often coincides with alkalinization of the growth medium (Eskew et al., 1977; Krieg and Dobereiner,









62

1984). Krieg and Dobereiner (1984) suggest that these cell forms arise mainly at excessively high pH. In this case they might be only laboratory artifacts, or involution forms, that have no in situ function. The findings of Lamm and Neyra (1981), Papen and Werner (1982), and Sadasivan and Neyra (1985) argue against this viewpoint. Indeed, the ability of azospirilla to enter dormancy as cysts may help explain some of the great variability of plant responses to inoculation with these bacteria (reviewed by Patriquin et al., 1983).

Two things are presently lacking in research and understanding of cyst-like forms of azospirilla. Although cystlike forms of azospirilla have been predominant in some studies, growing cells of normal morphology (vegetative cells) have always been present in high numbers as well. Conversion of 90% or greater of a population of vegetative azospirilla to cyst-like forms (quantitative encystment) in a reproducible manner would greatly facilitate further study of these cell forms, as it did for Azotobacter spp. cysts (Socolofsky and Wyss, 1962). Also lacking is an understanding of the underlying causes of pleomorphism and cyst formation in azospirilla.

Conversion of 90% or greater of a cell population of

Azotobacter spp. to cysts often can be achieved by culturing vegetative cells in the absence of combined nitrogen on either of two precursors of PHB, n-butanol or BHB (Sadoff,









63

1975). Although all strains of A. brasilense and A. lipoferum are known to grow on BHB as sole carbon source when provided with combined nitrogen (Tarrand et al., 1978), there are no reports of the response of azospirilla to BHB in the absence of combined nitrogen. It was thought worthwhile to see if vegetative azospirilla would respond similarly to Azotobacter spp., by undergoing quantitative encystment in the presence of these carbon sources. In preliminary studies, apparent extensive PHB accumulation and capsule formation were observed in some strains of azospirilla grown with n-butanol. Since n-butanol is volatile, BHB was used for later studies.

Initial objectives of this study were to achieve

morphological encystment of high numbers of azospirilla, to document the morphology of such cells, to verify that they contained PHB, and to ascertain if alkalinization of the medium was a prerequisite for their formation.


Materials and Methods


Bacterial Strains


The Azospirillum brasilense strains used in these

studies were A. brasilense strain JM 125A2 and A. brasilense strain Cd (ATCC 29729) (both courtesy of J. Milam, Univ. of Florida, Gainesville). The Azospirillum lipoferum strains used were A. lipoferum Sp RG6xx (ATCC 29731), A. lipoferum Sp RG20a (ATCC 29708), A. lipoferum Sp RG8C, and









64

A. lipoferum Sp A3a (all courtesy of N. R. Krieg, Va. Poly. Inst., Blacksburg). All strains were maintained on slants of Tryptic Soy Agar (Difco Laboratories, Detroit, MI) at 250C with monthly transfer.


Media


Vegetative azospirilla were cultured in a modification of the complete medium of Tyler et al. (1979), denoted as trypticase-succinate salts (TSS). All components were of reagent grade and were dissolved in deionized water. The final concentrations of TSS components were (in grams per liter): (NH4 ) 2SO4, 0.5; succinic acid, 0.437; Trypticase Peptone (Baltimore Biological Laboratory, Cockeysville, MD),

1.0; d-biotin (Sigma Chemical Co., St. Louis, MO), 0.0001; NaCl, 0.1; FeCl36H20, 0.0017; Na2MoO4'2H20, 0.0002; MgSO4'7H20, 0.2; and CaCl2, 0.002. The first four components were omitted to obtain a basal salts solution. The biotin was dissolved as a 100X concentrated stock solution by heating and then filter-sterilized by passage through a

0.2 pm pore diameter Nalgene filter unit (Nalge Company, Rochester, NY). Two phosphate buffer concentrations were employed. The low phosphate (LP) buffer of Tyler et al. (1979) had a final concentration of 3.5 mM and consisted of (in grams per liter) K2HPO4, 0.1 and KH2PO4, 0.4. The high phosphate (HP) buffer of Albrecht and Okon (1980) had a final concentration of 63.8 mM and consisted of (in grams









65

per liter) K2HPO4, 6.0 and KH2PO4, 4.0. The LP buffer was prepared as a 100X concentrated stock solution, and the HP buffer as a 10X concentrated stock solution. The pH of the LP buffer was adjusted to 7.1, and that of the HP buffer to

6.7, with 10 M KOH. The buffer stock solutions were sterilized by autoclaving. All autoclavings in these studies were for 15 min at standard temperature and pressure unless otherwise stated.

The TSS components, excluding the biotin and phosphates, were dissolved and adjusted to pH 7.0 with 10 M KOH. The broth was then dispensed into 250 ml Erlenmeyer flasks, in an amount calculated to obtain a final volume of 100 ml after aseptic addition of the biotin and phosphate buffer stocks to the autoclaved TSS. The initial pH of the LP-TSS was 6.9 to 7.0, and that of HP-TSS was 6.8.

Plate counts of azospirilla were performed with a

modified succinate-nitrogen free (SNF) agar medium derived from Tyler et al. (1979) It was the same as LP-TSS, except that (NH4 )2S4 and Trypticase Peptone were omitted. It contained in addition (in grams per liter) Bacto yeast extract (Difco), 0.05 and Bacto agar (Difco), 20.0. It was prepared in the same way as TSS broth, except that agar was added after neutralization and before autoclaving. Before Petri plates were poured, biotin and LP buffer were added aseptically, as was a solution of autoclaved Congo Red








66

(Sigma), that was incorporated at a final concentration of

0.0375 grams per liter (Rodriguez Caceres, 1982).

Nitrogen-free agar plates containing n-butanol had the same composition as SNF plates, except that the agar concentration was 1.5% (wt/vol), and yeast extract and Congo Red were omitted. The n-butanol was sterilized by filtration in the same manner as the biotin and incorporated at a final concentration of 0.2% (vol/vol). Beta-hydroxybutyrate was prepared from crotonic acid (Sigma) by dissolving 23.6 g crotonic acid in 900 ml deionized water. This solution was continuously mixed with a magnetic stirrer for two to three days at 250C. Its OD235 stabilized by this time, indicating conversion to BHB (H. L. Sadoff, personal communication). It was then adjusted to pH 7.0 with 10 M KOH, the final volume made up to one liter, and sterilized by autoclaving. This served as a 10X concentrated stock solution of BHB for addition to agar or broth, to give a final concentration of

0.236% (wt/vol) BHB. Agar plates containing BHB had the same composition as n-butanol plates, except that (NH4)2SO4 or Congo Red were sometimes added at the previously described concentrations. For two-step broth replacement studies (described below), broth contained BHB, biotin, and phosphate-buffered basal salts solution. The LP and HP buffers were employed in different broth replacement studies. The initial pH, after inoculation, of LP-BHB broth was 7.2, and that of HP-BHB broth was 6.9.









67

Growth Conditions


Inocula of azospirilla were grown in screw-cap tubes

containing 10 ml of autoclaved Bacto Nutrient Broth (Difco) for 24-48 hours at 280C. One milliliter of inoculum was aseptically pipetted into 100 ml of HP-TSS broth and incubated for 20 to 22 hours at 30oC at 130 rpm on a rotary shaker. By this time, the cultures attained OD560 readings of 0.7 to 0.9, as measured with a Bausch and Lomb Spectronic 20 spectrophotometer. The pH of the cultures at harvest ranged from 7.0 to 7.2. Cultures were pelleted by centrifugation at 6,960 X g for 15 min at 200C. Cells were washed twice by resuspension and pelleting in sterile LP-basal salts solution (pH 7.3). The cells were resuspended in sterile LP-basal salts solution to give a final OD560 reading of 1.0 to 1.2.

Cell lawns were obtained by spread plating 0.1 ml of

washed cells onto agar media. Inoculated plates were sealed with Parafilm and incubated at 280C.

For two-step broth replacement studies, washed cells

were aseptically added as a 10% (vol/vol) inoculum to 250-ml Erlenmeyer flasks, containing a final volume of 100 ml BHB broth after cell addition. Duplicate flasks were incubated in the same manner as TSS flasks. These studies were modeled after the two-step replacement method of Lin and Sadoff (1968).








68

Harvest of Cell Lawns


To harvest lawns of azospirilla grown on n-butanol or BHB agar, about 7.0 ml of sterile deionized water was aseptically poured across the surface of a cell lawn, and the cells gently scraped from the agar surface with a flamed wire loop. For PHB analyses and plate counts, the suspended cells of one BHB agar plate were aseptically transferred to another plate whose cells were in turn scraped off. This was done to ensure that the cell suspension would not become too diluted.


Enumeration


Vegetative cells from TSS-broth, or cells grown on BHB agar, were diluted ten-fold in a series of dilution blanks containing LP-basal salts solution. For enumeration, 0.1 ml of cell suspension was aseptically spread on SNF-Congo Red agar plates. Four plates were spread for each dilution. Plates were incubated as described above for 5 days before counting.


Dry Weight Determination and PHB Analysis


To assay PHB content of vegetative cells of A. lipoferum Sp RG6xx, two 22-hour-old, HP-TSS cultures (OD560 = 0.6) were pooled for centrifugation and washing as described above, except that sterile deionized water was used for washing. The final cell suspension was adjusted to








69

an OD560 = 0.86. Forty plates of the same strain grown on nitrogen-free-LP-BHB agar were harvested by scraping (described above), to give a final cell volume of about 100 ml. These cells were centrifuged and washed in sterile deionized water (described above) and resuspended to give an OD560 of 0.25 to 0.28.

For dry weight determinations, 10.0 ml of the final cell suspension were pipetted into previously weighed and desiccated aluminum pans. Five replicate pans were prepared for each cell type. The pans containing cells were dried to constant weight at 1000C. Pans were kept in a glass desiccator over anhydrous CaSO4 (Drierite) after removal from the oven and before weighing.

For PHB determination, 10.0 ml of washed cells were

added to 15 ml Corex centrifuge tubes (Corning Glass Works, Corning, NY) and pelleted by centrifugation at 7,080 X g for 20 min at 40C. Three replicate tubes were prepared for each cell type. The supernatant was poured off, and subsequent steps were performed by the method of Law and Slepecky (1961). Digestion of cell pellets was begun with the addition of 10 ml of Clorox bleach (5% (wt/vol) hypochlorite). Cells were suspended in the bleach with Pasteur pipettes; then the tubes were capped with glass marbles and incubated in a 370C water bath. Digestion to constant OD650 was monitored with a Spectronic 20 spectrophotometer and was judged to be complete after 18 hours. The insoluble cell








70

material was pelleted by centrifugation as above, then washed once in 10 ml of sterile deionized water, and pelleted again. The volume for all subsequent washings and digestions was maintained at 10.0 ml, and all chemicals were of reagent grade. The OD235 of the samples in the final digestion of concentrated H2SO4 was measured in quartz cuvettes (1.0 cm light path), using a Carl Zeiss M4QIII spectrophotometer. For the standard curve, the sodium salt of DL--hydroxybutyric acid (Sigma) was dissolved directly in concentrated H2SO4. The standard curve was linear up to 8.0 pg BHB/ml. The PHB content of cell digests was related back to dry weight values, to determine what percentage of cell dry weight was present as PHB. Scanning Electron Microscopy (SEM)


Samples of 0.4 ml from either LP-BHB agar plates or

two-step, broth-replacement cultures were employed for SEM studies. Cells were removed aseptically from the two-step, broth-replacement cultures at the same time that culture pH was measured. Cells were aseptically impinged upon autoclaved 25-mm-diameter, 0.45-pm-pore-size Nucleporeo polycarbonate filters (Nuclepore Corporation, Pleasanton, CA), housed in a filter chimney attached to a vacuum source. About 10.0 ml of sterile, deionized water was added to the chimney after cell addition, to help distribute the cells evenly over the membrane surface, then a vacuum not









71

exceeding 33.8 kPa was applied. Filter membranes were then removed and placed into Karnovsky's fixative (1965) for 1 hour. Filter membranes were subsequently rinsed twice for 10 min in cacodylate buffer and then dehydrated in a graded series of ethanol concentrations (10, 20, 30, 50, 70, 90, 95, 100, and 100%) for 10 min at each concentration. The samples were then air dried. Sections of filters were excised, placed onto aluminum stubs with double-stick tape, and gold coated with an Eiko IB-2 coater. Specimens were examined with a Hitachi S450 scanning electron microscope at 20 kilovolts. Photographs were taken with Polaroid Type 55, positive/negative, 4X5 Land film. Light Microscopy


Cells were routinely observed by phase-contrast

microscopy using a Wild M20 or a Nikon Labophot microscope. Cell dimensions were measured with an ocular micrometer. Photographs of cells viewed with the latter microscope were taken with a Microflex AFX camera attachment, using Ilford FP4 black and white film. All photos were taken using phase- contrast optics, unless otherwise indicated.


Results


Quantitative Morphological Change


In the initial phase of these studies only three strains of azospirilla were used, A. brasilense strain








72

JM 125A2, A. brasilense strain Cd, and A. lipoferum Sp RG6xx. Slime developed at the bottom of stationary phase LP-TSS broth cultures of all three strains. Phase-contrastmicroscopy examination of A. lipoferum Sp RG6xx slime often revealed numerous, nonmotile masses of cells similar to zooglea, surrounded by nonmotile vegetative cells (Figure 2-la, b). These masses were notable for their symmetrical but varied shapes. They were darker than most of the surrounding vegetative cells, perhaps indicating greater viability than that of the surrounding pale vegetative cells. These zoogleal masses retained their shape and did not fragment into individual cells when disassociated from the larger masses of cells. Similar zoogleal forms were sometimes observed in the slime of A. brasilinese strain Cd but not in that of A. brasilense strain JM 125A2. These zoogleal forms of azospirilla may be referred to as microscopic flocs, or microflocs, that are kept intact by exopolysaccharide. Although microflocs were numerous, individual normal cells were also present in large numbers under these cultural conditions.

Azospirilla were cultured as cell lawns on agar containing precursors of PHB to see if high numbers of pleomorphic forms would arise. The A. brasilense strains produced some pleomorphic forms, but cells of normal shape and size predominated, even on old lawns.
































Figure 2-1. Zoogleal masses in stationary phase
40-day-old, low phosphate-trypticasesuccinate-salts broth culture of
Azospirillum lipoferum Sp RG6xx. a) Cells
at 600X magnification. Bar equals 6.0 pm. b) Detail from same mass of cells
viewed at 1,500X magnification. Bar
equals 3.0 pm.







74


























a




























b








75

After 63 hours of growth, A. brasilense strain JM 125A2 lawns grown on BHB contained ovoids, vibrioids, and chains of cells. Many cells contained phase-bright, putative PHB granules. More cells were present at this time on agar containing combined nitrogen. The several cell types present on 63-hour-old, LP-BHB agar with combined nitrogen are shown in Figure 2-2a. Some cells appeared at this time to be undergoing plasmolysis on this medium, as well as on HP-BHB agar with combined nitrogen. By 96 hours, the lawns on LP-BHB agar with and without combined nitrogen contained more chains of cells and microflocs than the HP-BHB lawns, which consisted mostly of individual ovoids or pairs of ovoids.

Cells from month-old, nitrogen-free, HP-BHB lawns of A. brasilense strain JM 125A2 are shown in Figure 2-2b. Individual, nonmotile vibrioids and ovoids were still predominant, as were pairs of cells. Enlarged, nonmotile, spherical cells were present, but not numerous. A few nonmotile filaments appeared to be undergoing septation.

After 79 days, cells from lawns of this strain grown with combined nitrogen had the appearance of stationary phase cells from TSS broth cultures grown with combined nitrogen, and spheroplasts and cell ghosts predominated. Nitrogen-free cultures at both phosphate buffer concentrations contained numerous pleomorphic forms. Figure 2-3a shows nonmotile, enlarged, rounded individual cells from
































Figure 2-2. Cell types of Azospirillum brasilense
strain JM 125A2, grown on B-hydroxybutyrate (BHB) agar. a) 63-hour-old cells
from low phosphate-BHB agar with combined
nitrogen. 1,500X magnification. Bar
equals 3.0 pm. b) Month-old cells from nitrogen-free, high phosphate-BHB agar.
1,000X magnification. Bar equals 4.0 pm.







77























































b
































Figure 2-3. Call types of Azospirillum brasilense
strain JM 125A2, from 79-day-old nitrogenfree, high phosphate-6-hydroxybutyrate agar cultures. a) Individual rounded
cells. 1,500X magnification. Bar equals
3.0 pm. b) Microfloc focused so that
capsules are visible around cells on
right side of floc. 1,500X magnification.
Bar equals 3.0 pm. c) Same cells as (b),
but focused so that capsules are no longer
evident. Note empty capsule at bottom of floc. 1,500X magnification. Bar equals
3.0 pm.








79









80

nitrogen-free HP-BHB agar. Their phase-bright inclusions are probably PHB granules; some contain dark bodies, probably polyphosphate granules. Also shown is a microfloc of nonmotile, enlarged, PHB-rich cells (Figure 2-3b, c). By adjusting the distance of the objective lens from the specimen, many of these cells were observed to be encapsulated (Figure 2-3b). The encapsulated cells fitted together closely, as did those observed by Sadasivan and Neyra (1985). The thickness of the capsule was about

0.5 pm. Such encapsulated microflocs were also observed on nitrogen-free, LP-BHB agar at this time.

Azospirillum brasilense strain JM 125A2 may have lacked an efficient mechanism for BHB uptake, compared to the other strains of azospirilla used. Unlike the other strains, few motile cells were observed on any BHB agar medium, even in young cultures. It also differed from the other strains by having many phase-dark cells that contained little or no PHB. It eventually grew well on BHB agar when combined nitrogen was available, however. A final difference between this strain and the others was that its cells always resuspended in water to give uniform turbidity, with no macroscopic flocs, or macroflocs, being present. This indicates that, with or without combined nitrogen, cells of this strain produce little capsular material when cultured on BHB.








81

The best growth of A. brasilense strain JM 125A2 on

agar occurred on SNF-Congo Red agar. Cells from 6-day-old lawns grown on this agar medium were often seen as encapsulated microflocs (Figure 2-4). The microfloc in Figure 2-4a and b appears to have arisen mainly from one or more filamentous cells that underwent septation. This may also have occurred for many of the cells in Figure 2-4c. The capsules were of thickness comparable to those observed on BHB agar, about 0.5 4m. The lawns on SNF-Congo Red agar had a scarlet or blood-red appearance, unlike lawns of this strain growing on nitrogen-free BHB-Congo Red agar, which were pale orange.

The other A. brasilense strain, A. brasilense strain Cd, also failed to convert in high numbers to pleomorphic forms, but it grew far better on BHB. After 63 hours of growth, lawns of this strain on each BHB agar medium contained many motile vibrioids possessing large granules of putative PHB. Elongated, filamentous cells were also present in high numbers. These cells had about the same width (1.5 pm) as normal dinitrogen-fixing cells but were much longer, some being 9 to 13 pm in length (Figure 2-5a). The filaments were sometimes observed to undulate slowly and were much slower than motile vibrioids. In the presence of combined nitrogen, these filaments were seen to septate and fragment. This fragmentation was observed at 63 to 96 hours, and sometimes was complete within a population of
































Figure 2-4. Cells of Azospirillum brasilense strain JM
125A2, from 6-day-old lawns grown on
succinate-nitrogen-free-Congo Red agar.
a) Microfloc showing capsules and
filamentous cell patterns. 1,500X magnification. Bar equals 3.0 4m. b) Same floc as (a), but focused so that capsules
and filamentous cell outlines are no
longer evident. 1,500X magnification.
Bar equals 3.0 pm. c) Different mass of
encapsulated cells. 1,500X magnification. Bar equals 3.0 pm.







83


















b































C
































Figure 2-5. Cell types of Azospirillum brasilense
strain Cd, from lawns on B-hydroxybutyrate
(BHB) agar. a) Filaments from 63-hourold, high phosphate-BHB agar with combined
nitrogen. 1,500X magnification. Bar
equals 3.0 pm. b) Microfloc from
11-day-old nitrogen-free, low phosphateBHB agar, focused to show capsules and
filamentous cell outline. 1,000X magnification. Bar equals 4.0 pm. c) Same floc
as (b), but focused so that capsules and
filamentous cell outline are no longer
evident. 1,000X magnification. Bar
equals 4.0 pm.







85


























a






















b c









86

filaments soon after 96 hours. In nitrogen-free cultures, such elongated filaments persisted, some being weakly motile even after 79 days on nitrogen-free LP-BHB agar.

After 96 hours, lawns of A. brasilense strain Cd on all BHB agar media contained mixtures of vibrioids, ovoids, filaments, and chains. Sometimes the cell material from LP-BHB agar lawns with or without combined nitrogen did not resuspend uniformly in water, but as macroflocs, due to extensive encapsulation. A microfloc from an 11-day-old, nitrogen-free, LP-BHB agar plate is shown in Figure 2-5b,c. By adjusting the objective lens, the capsule is made evident. The entire microfloc may have arisen from one elongated filament that underwent septation, as suggested by the apparent linear continuities between cytoplasmic contents.

After 79 days of culture, lawns of this strain grown with combined nitrogen contained mainly spheroplasts and cell ghosts, appearing to have entered stationary phase. Cultures grown on nitrogen-free agar at both phosphate levels contained numerous pleomorphic forms at this time, in addition to cells of normal morphology (Figure 2-6).

A microfloc of A. brasilense strain Cd from a 6-day-old lawn on SNF-Congo Red agar is shown in Figure 2-7. All the cells are encapsulated, and empty capsules are evident. The lawn was scarlet in color, unlike the pale orange lawns of the same age grown on nitrogen-free BHB-Congo Red agar.
































Figure 2-6. Cell types of Azospirillum brasilense
strain Cd, from 79-day-old lawns on
nitrogen-free, 6-hydroxybutyrate (BHB)
agar. a) Multicellular packets, a chain, and individual ovoids from low phosphateBHB agar. 1,500X magnification. Bar
equals 3.0 pm. b) Microfloc from high
phosphate-BHB agar, focused to show capsules around several cells. Air
bubble is above floc. 1,500X magnification. Bar equals 3.0 pm.








88



























a


























b







89


































Figure 2-7. Microfloc of Azospirillum brasilense
strain Cd, from 6-day-old lawn on
succinate-nitrogen-free-Congo Red agar.
Note capsules around cells and empty
capsules. 1,500X magnification. Bar
equals 3.0 um.








90

The two A. brasilense strains failed to produce morphologically uniform populations on BHB-agar. A uniform response was observed for A. lipoferum Sp RG6xx. Good growth usually occurred within 18 to 24 hours. Figure 2-8 shows 18-hour-old cells grown on nitrogen-free, HP-BHB agar. Filaments and chains were present, which were sometimes as swiftly motile as vibrioids. On LP-BHB agar at both phosphate levels, with or without combined nitrogen, septation of filaments was almost complete between 48 and 72 hours, although new filaments would arise and septate for up to the fifth day. Figure 2-9a, b shows such completely septated microflocs on 63-hour-old, nitrogen-free, LP-BHB agar. The flocs are encapsulated, and most seemed to arise from one filament that underwent complete septation. Although microflocs were present at this time, the cells resuspended from agar as uniformly turbid suspensions without macroflocs.

Many filaments were also completely septated by 63

hours on HP-BHB agar containing combined nitrogen, but some filaments still lacked completed septa (Figure 2-9 c). Nitrogen-free, HP-BHB agar at this time contained few if any microflocs. As for all other media, very motile vibrioids and ovoids, as well as filaments of varying motility, were present at 63 hours.

Encapsulated microflocs sometimes formed on nitrogenfree, HP-BHB agar (Figure 2-10a), but cells from young or







91

































Figure 2-8. Cells of Azospirillum lipoferum Sp RG6xx,
from 18-hour-old lawn on nitrogen-free,
high phosphate- 8-hydroxybutyrate agar.
Note individual cells and filaments
at various stages of septum formation.
1,000X magnification. Bar equals 4.0 Fm.
































Figure 2-9. Cell types of Azospirillum lipoferum
Sp RG6xx, from 63-hour-old lawns on
B-hydroxybutyrate (BHB) agar. a) Microflocs from nitrogen-free, low phosphateBHB agar, focused to show capsules and
filamentous cell outlines. 1,500X magnification. Bar equals 3.0 pm.
b) Same microflocs as (a) but focused so
that capsules and filamentous cell
outlines are no longer evident.
1,500X magnification. Bar equals 3.0 pm.
c) Filament from high phosphate-BHB agar with combined nitrogen. 1,500X magnification. Bar equals 3.0 pm.








93 a b























c




Full Text
154
Table 3-1. Desiccation resistance
RG6xx.
of Azospirillum lipof
erum Sp
Experi- Cell
ment type
Number
Initial.
CFU ml 1
(X 10b)
drieda
Beaker CFU ml ^
1 2
final %
survivors0
3
1 Vege-
0.72
0.0
22.5
5.0
0.03
tative
(0.03)
(0.0)C
(13.0)
(5.0)
Encapsu-
0.42
92.5
55.0
e
0.35
latedd
(0.06)
(37.7)
(28.7)
2
Vege-
0.14
10.0
0.0
0.0
0.05
tative
(0.03)
( 7.1)
(0.0)
(0.0)
Encapsu-
1.34
2.5
25.0
40.0
0.03
lated£
(0.10)
(4.3)
(43.0)
(17.3)
aAverages of colony forming units (CFU) of four spread plates.
Values in parentheses are standard deviations.
Average of all beakers available.
c
Averages of four spread plates. Values in parentheses are
standard deviations.
^75-day-old cells.
0
Beaker was lost.
f53-day-old cells.


41
The role of PHB in cyst formation was examined by
Stevenson and Socolofsky (1966). Cysts were defined as
cells that could survive desiccation on a membrane-filter
surface for 4 days at 33C. After 2 days of growth on
nitrogen-free n-butanol agar, cells lost their motility,
became oval-shaped, and accumulated PHB to the extent of 35%
of cell dry weight. The development of mature cysts was
accompanied by a reduction in PHB content. By 6 days, cul
tures had undergone 100% encystment, and 10% of cyst dry
weight was PHB.
Lin and Sadoff (1968) developed a two-step replacement
procedure for obtaining cysts in broth. Cells were grown to
late exponential phase in Burk's nitrogen-free broth with
glucose. After harvest by centrifugation and washing in
buffer, cells were resuspended in Burk's salts broth with
0.2% (wt/vol) BHB. This procedure was used in further
studies (Hitchins and Sadoff, 1970, 1973; Reusch and Sadoff,
1979; Su et al., 1981; Reusch and Sadoff, 1983), resulting
in the following detailed description of BHB-induced
encystment.
After 1 hour in encystment broth, cells are still
motile and flagellated but no longer possess nitrogenase
activity. Within 4 to 6 hours, DNA synthesis has ceased,
and soon afterward each cell divides to form two nonmotile
precysts. There is rapid accumulation of PHB during this
period, and the rate of phospholipid synthesis declines.


15
with intermediates and enzymes of the tricarboxylic acid
(TCA) cycle, a system that azospirilla also possess (Okon et
al., 1976b). When A. beijerinckii strain N.C.I.B. 9067 was
cultured as a dinitrogen-fixer with 2.0% (wt/vol) glucose,
PHB was deposited towards the end of exponential growth.
The cells were unable to use all the available glucose, and
PHB synthesis continued during the stationary phase until up
to 74% of cell dry weight was PHB. Cultures grown with
combined nitrogen rarely contained more than 3.0% of their
dry weight as PHB (Dawes, 1981).
The initiation of PHB synthesis in the A. beijerinckii
strain in batch culture coincided with the attainment of
zero-oxygen concentration. Oxygen limitation was thus
suspected to be a critical factor in initiating PHB synthe
sis. However, the nature of batch broth culture made it
hard to separate oxygen effects from possible nitrogen-
limitation effects (Senior and Dawes, 1971). Later
experiments, using chemostat cultures having carbon, oxygen,
or nitrogen limitation, clearly showed that extensive PHB
accumulation only occurred under conditions of oxygen
limitation (Dawes, 1981).
Before the studies reviewed by Dawes (1981), PHB was
regarded as being only an endogenous, carbon-and-energy
source that benefited cells during starvation. These
experiments suggested that PHB could also serve other
purposes. The synthesis of PHB seemed to serve as an


169
cells in microflocs had a cyst-like appearance. However,
cells in microflocs survived desiccation no better than
nonencapsulated vegetative cells. Encysting bacteria may
gradually mature, with their resistance to environmental
stress increasing with time although their morphology
appears cyst-like throughout. Modification of the cell
membrane may be the true key to stress-resistance of these
forms, as may also be true of Azotobacter spp. cysts (Reusch
and Sadoff, 1983). Immature cysts of azospirilla may mature
best if they are removed from the growth medium and/or
slowly dried. Sadasivan and Neyra (1985) removed their
encapsulated azospirilla from broth and dried the floes
slowly, and possibly obtained desicccation resistant forms.
Similarly, Lamm and Neyra (1981) obtained desiccation-
resistant azospirilla by allowing the agar of lawns to dry
slowly.
There was a significant difference in response to
carbon and nitrogen starvation between vegetative and encap
sulated cells of A. lipoferum Sp RG6xx. After 9 days, only
about 25% of the original encapsulated inoculum retained
viability. The cell densities used in these starvation
6
experiments were less than 10 CFU/ml, so microscope
observations were not performed. However, the microscope
observations of extended germination incubations with higher
cell densities indicated that the majority of PHB-rich cells
within floes will eventually deplete their PHB reserves and


128
Figure 2-24. Cells of Azospirillum lipoferum Sp RG20a
from a 33-day-old, nitrogen-free, high
phosphate-8-hydroxybutyrate broth
culture, pH 7.2, viewed by scanning
electron microscopy. Note thorough
encapsulation of lower cell layer and
presence of filaments and chains.
3,000X magnification. Bar equals 5.0
pm.


191
Tal, S., and Y. Okon. 1985. Production of the reserve
material poly-B-hydroxybutyrate and its function in
Azospirillum brasilense Cd. Can. J. Microbiol.
31:608-613.
Tarrand, J. J., N. R. Krieg, and J. Dobereiner. 1978. A
taxonomic study of the Spirillum lipoferum group, with
descriptions of a new genus, Azospirillum gen. nov. and
two species, Azospirillum lipoferum (Beijerinck)
comb. nov. and Azospirillum brasilense sp. nov.
Can. J. Microbiol. 24:967-980.
Tchan, Y.-T. 1984. Family II. Azotobacteraceae, p. 219-
229. In N. R. Krieg and J. G. Holt (ed.), Bergey's
manual of systematic bacteriology, vol. I. The
Williams and Wilkins Co., Baltimore.
Thompson, J. P., and V. B. D. Skerman. 1979. Azotobac
teraceae : The taxonomy and ecology of the aerobic
nitrogen-fixing bacteria. Academic Press, New York.
Tien, T. M., H. G. Diem, M. H. Gaskins, and D. H. Hubbell.
1981. Polygalacturonic acid transeliminase production
by Azospirillum species. Can. J. Microbiol.
27 .-426-431.
Tien, T. M., M. H. Gaskins, and D. H. Hubbell. 1979. Plant
growth substances produced by Azospirillum brasilense
and their effect on the growth of pearl millet
(Pennisetum americanum L.). Appl. Environ. Microbiol.
37:1016-1024.
Tudor, J. J. 1980. Chemical analysis of the outer cyst
wall and inclusion material of Bdellovibrio bdello-
cysts. Curr. Microbiol. 4:251-256.
Tudor, J. J., and S. F. Conti. 1977. Characterization of
bdellocysts of Bdellovibrio sp. J. Bacteriol. 131:
314-322.
Tyler, M. E., J. R. Milam, R. L. Smith, S. C. Schank, and
D. A. Zuberer. 1979. Isolation of Azospirillum from
diverse geographic regions. Can. J. Microbiol.
25:693-697.
Umali-Garcia, M., D. H. Hubbell, M. H. Gaskins, and
F. B. Dazzo. 1980. Association of Azospirillum with
grass roots. Appl. Environ. Microbiol. 39:219-226.
Vela, G. R. 1974. Survival of Azotobacter in dry soil.
Appl. Microbiol. 28:77-79.


186
Krieg, N. R., and J. Dobereiner. 1984. The genus
Azospirillum, p. 94-104. In N. R. Krieg and J. G. Holt
(ed.), Bergey's manual of systematic bacteriology,
vol. 1. The Williams & Wilkins Co., Baltimore.
Labeda, D. P., K.-C. Liu, and L. E. Casida, Jr. 1976.
Colonization of soil by Arthrobacter and Pseudomonas
under varying conditions of water and nutrient
availability as studied by plate counts and trans
mission electron microscopy. Appl. Environ.
Microbiol. 31:551-561.
Lakshmi, V. A. Satyanarayana Rao, K. Vijayalakshmi, M.
Lakshmi-Kumari, K. V. B. R. Tilak, and N. S. Subba
Rao. 1977. Establishment and survival of Spirillum
lipoferum. Proc. Indian Acad. Sci. 86B:397-404.
Lamm, R. B., and C. A. Neyra. 1981. Characterization and
cyst production of azospirilla isolated from selected
grasses growing in New Jersey and New York. Can. J.
Microbiol. 27:1320-1325.
Law, J. H., and R. A. Slepecky. 1961. Assay of poly-8-
hydroxybutyric acid. J. Bacteriol. 82:33-36.
Leive, L. L., and B. D. Davis. 1980. Cell envelope;
spores, p. 73-110. In B. D. Davis, R. Dulbecco, H. N.
Eisen, and H. S. Ginsberg (ed.), Microbiology, 3rd ed.
Harper & Row, Publishers, Philadelphia.
Lin, L. P., and H. L. Sadoff. 1968. Encystment and polymer
production by Azotobacter vinelandii in the presence of
-hydroxybutyrate. J. Bacteriol. 95:2336-2343.
Lin, L. P., and H. L. Sadoff. 1969. Preparation and ultra
structure of the outer coats of Azotobacter vinelandii
cysts. J. Bacteriol. 98:1335-1341.
Lin, W., Y. Okon, and R. W. F. Hardy. 1983. Enhanced
mineral uptake by Zea mays and Sorghum bicolor roots
inoculated with Azospirillum brasilense. Appl.
Environ. Microbiol. 45:1775-1779.
Loperfido, B., and H. L. Sadoff. 1973. Germination of
Azotobacter vinelandii cysts: Sequence of macromolecu-
lar synthesis and nitrogen fixation. J. Bacteriol.
113:841-846.
Lowendorf, H. S. 1980. Factors affecting survival of
Rhizobium in soil, p. 87-124. In M. Alexander (ed.),
Advances in microbial ecology, vol. 4. Plenum Press,
New York.


146
LP-BHB agar, as described in Chapter II. Plate counts were
performed on the modified SNF agar medium using LP buffer
and Congo Red, as described in Chapter II.
Harvest of Cells
Harvest and washing of vegetative broth cultures always
employed sterile LP-basal salts solution (pH 7.3), as
described in Chapter II. Lawns on LP-BHB agar were
harvested in two ways. When cells were to be added to
semisolid agar or assayed for desiccation and starvation
resistance, they were harvested and washed in sterile LP-TSS
salts solution. For every other assay that was performed,
the cells were harvested and washed in sterile deionized
water.
Desiccation Resistance
Vegetative HP-TSS cultures were grown for 17-22 hours,
attaining an OD^q of 0.5 in each experiment. The cells
were then centrifuged and washed twice in sterile LP-basal
salts solution, as described in Chapter II. The cells were
resuspended in a third volume of LP-basal salts solution to
attain an OD,.^ of 0.3 in one experiment and 0.61 in another
experiment. Then ten-fold dilution series in sterile
LP-basal salts solution were prepared aseptically from the
_ o
cells. A 0.1-ml portion of the 10 dilution was asepti
cally added to each of three autoclaved 10-ml glass beakers,


177
Figure 3-4. Rounded, possibly cyst-like cells of
Azospirillum lipoferum Sp RG6xx, from a
29-hour incubation in low phosphate-basal
salts solution with glucose. 1,000X
magnification. Bar equals 4.0 pm.


25
Nonmotile, encapsulated, mature microcysts are more
resistant than are vegetative cells to environmental
stresses such as ultraviolet irradiation, sonic vibration,
and desiccation (Sudo1 and Dworkin, 1973).
Many Gram-negative, methane-oxidizing bacteria isolated
from soil or mud are encapsulated and accumulate PHB when
nitrogen becomes limiting for growth (Whittenbury et al.,
1970b). Depending on the genus and strain, up to 90% of the
cells present may form resting cells upon entering the sta
tionary phase of growth. Lipid cysts of Methylocystis
parvus accumulate large amounts of PHB and survive starva
tion and desiccation better than vegetative cells, but lack
well-defined, capsular cyst coats. Methylomonas spp. and
Methylococcus spp. form rounded, nonmotile cells that
survive starvation better than do vegetative cells. These
cells are called immature cysts, because they never attain
desiccation resistance. Some strains of Methylobacter
spp. form starvation-resistant and desiccation-resistant
cysts that seem morphologically identical to Azotobacter
spp. cysts (Whittenbury et al., 1970a).
Bdellovibrio sp. strain W is the only bdellovibrio that
is known to encyst. Bdellocysts are larger than their
vegetative counterparts and are not light-refractile. They
tolerate sonic disruption, ultraviolet irradiation, and
carbon starvation better than do vegetative cells. The
endogenous respiration rate of bdellocysts is 80% less than


104
2-16 show successive magnifications of cells scraped from
a 75-day-old, nitrogen-free, LP-BHB agar lawn. Cells from
19-, 57-, and 66-day-old lawns on the same medium were of
identical appearance to this older culture. This supported
the phase-contrast microscopy studies, in that the morphol
ogy of microflocs from nitrogen-free, LP-BHB agar did not
change once they were formed. The 500X magnification photo
(Figure 2-14) shows the great variability in numbers of
cells per floe and that individual, encapsulated cells are
often present. The 1,500X-magnification photo (Figure 2-15)
reveals, in agreement with the phase-contrast observations,
the frequent tight fit between adjacent capsules. Inter
cellular gaps are often observed within floes. The 7,000X-
magnification photo (Figure 2-16) reveals some variability
in the surfaces of encapsulated cells, possibly indicating a
difference in exopolysaccharide composition. It is clear at
this higher magnification that the cells of agar-grown floes
range from monolayers to trilayers. Empty capsules are also
visible.
Similar results were obtained with this strain on LP-n-
butanol agar. Figure 2-17 shows 178-day-old cells from this
agar medium after the agar had dried into a thin film. When
the agar surface was rehydrated and the cells scraped from
it, many empty capsules were seen, but many capsules still
contained cells. The capsular material was thus observed to
retain its outline, even if it contained no cell.


Figure 2-18.
Microflocs of Azospirillum lipoferum
Sp RG8c, grown on nitrogen-free, low
phosphate-6-hydroxybutyrate agar. a)
Microfloc from 69-day-old lawn, focused
to show capsules and filamentous cell
outline. 1,500X magnification. Bar
equals 3.0 pm. b) Same floe as (a) but
focused so that capsules and filamentous
cell outline are no longer evident.
1,500X magnification. Bar equals 3.0
pm. c) Microflocs from 136-day-old
lawn. Note empty capsules. 1,000X
magnification. Bar equals 4.0 pm.


150
additions to the LP-basal salts solution were added as 10X-
concentrated, sterile stock solutions that were sterilized
by autoclaving.
The sugars employed were D-glucose (Difco), sucrose
(Difco), and D-fructose (Calbiochem, San Diego, CA). All
were prepared as separate 4.37% (wt/vol) stocks in deionized
water. The organic acids were succinic acid (Fisher) and
DL-malic acid (Sigma). Each acid was prepared as a separate
4.37% (wt/vol) salt stock that was neutralized to pH 7.0
with 10 M KOH before being brought to final volume. The
nitrogen sources were reagent grade KNO^ and (NH^^SO^. The
KNO^ was prepared as a separate 0.765% (wt/vol) stock, and
the (NH^)2^0^ was prepared as a separate 0.5% (wt/vol)
stock.
The same type of screw cap tubes used in carbon starva
tion studies were used in these studies. Each empty,
sterile tube had aseptically added to it 8.0 ml of sterile,
LP-basal salts solution plus biotin and a single carbon or
nitrogen source. Sometimes cells were incubated in LP-basal
salts solution with biotin, but without carbon or nitrogen
sources. Then 2.0 ml of water-washed cells were aseptically
added to each tube. Treatments for each carbon or nitrogen
source were done in triplicate. Tubes were incubated
horizontally on a 130 rpm rotary shaker at 30C. The pH of
amended buffer incubations ranged from 7.1 to 7.2.


53
2.0 to 4.0 pm diameter which had lost their motility.
One-week-old cultures consisted mainly of spherical C-forms
of 5.0 to 8.0 pm diameter, containing many PHB-rich cells
within a common capsule. The authors speculated that
younger encapsulated forms may be fixing dinitrogen and that
older encapsulated forms may not. They suggested that the
capsule may reduce oxygen flow into the cells, thereby
protecting oxygen-sensitive nitrogenase activity. Azo
spirilla form extensive capsules only in media having a high
C/N ratio (Sadasivan and Neyra, 1985). Such conditions
promote nitrogenase activity. Since most capsules contain
over 99% of their weight as water (Costerton et al., 1981),
and oxygen diffuses through water at one ten-thousandth the
rate of diffusion through air (Clark, 1967), the capsule may
well help protect nitrogenase from oxygen damage. Oxidation
of PHB reserves within the cell may also reduce oxygen
levels near the nitrogenase (Dawes and Senior, 1973).
The description by Berg et al. (1980) of encapsulation
starting at the uppermost layers of nitrogen-free agar-grown
colonies and proceeding downwards is reasonable, if one
assumes that encapsulated cells are metabolically active for
a time, and then pass into dormancy. Initially, the upper
most encapsulated cells would be actively fixing dinitro
gen. They might become dormant as a result of underlying
cell layers depleting the available carbon supply, or pos
sibly because conditions become favorable for their passage


31
an incubator at 33C. This method is a slow-dying pro
cedure. At different time intervals, the cells were washed
from the membranes, and viability was determined by plat
ing. Cysts of Azotobacter vinelandii ATCC 12837 lost little
viability over a 12-day-period using this drying treatment,
whereas 99% of the vegetative cells were killed by the end
of the first day (Socolofsky and Wyss, 1962). As a result
of the rapid die off of vegetative cells with this treat
ment, later studies considered cells of this strain to be
cysts if they could withstand 4 days of desiccation on
membrane filters (Stevenson and Socolofsky, 1966; Wyss et
al., 1969).
None of these membrane filter studies specified the RH
at which the membranes were dried, or how many cell layers
were deposited upon the membranes. Webb (1965) pointed out
that if bacteria are dried on filters to test their desic
cation resistance, the cells must be applied in a monolayer
to achieve consistent results. If more than a cell mono-
layer is on the filter, most of the cells in subsurface
layers will not be dried or equilibrated with the water
vapor of the environment.
Most desiccation resistance experiments have given
ill-defined or incomplete conditions of drying. Such
experiments have proven, however, that cysts are more
desiccation resistant than are their vegetative
counterparts.


176
during starvation and which finally lost their viability.
The remaining cells within the microfloc may include cells
that eventually mature into cysts. Finally, the similar
"horseshoe" appearance of some empty capsules of azospirilla
and of germinated cysts of Azotobacter spp. may have some
importance. More than one type of capsule may exist within
an encapsulated microfloc of azospirilla, and some capsules
may have proceeded further toward a cyst-coat composition
than others.
A few encapsulated cells in floes survived the desicca
tion treatment. A few cells within floes also had the
appearance of rounded, possibly mature cysts. Sometimes
they broke free of floes (Figure 3-4). They were never
observed to be motile. If these are truly mature cysts, the
problem remaining is how to convert most of the cells in a
vegetative inoculum quantitatively into this form.


140
(Jensen and Woolfolk, 1985). Studies reported here indicate
that azospirilla may form them under both conditions.
Strains of A. lipoferum grown in broth containing
combined nitrogen have a greater tendency to clump than do
strains of A. brasilense (Krieg and Dobereiner, 1984). In
this study, the SEM photographs of stationary phase HP-TSS
broth cultures of A. lipoferum Sp RG6xx indicated there is
some structural regularity in clumps. The cells were often
arranged so that the floes contained spaces. The existence
of spaces was more regular and pronounced in floes of cells
cultured with BHB in the absence of combined nitrogen,
probably because the exopolysaccharides of these cells were
more rigid than the exopolysaccharides of stationary phase
cultures grown with combined nitrogen, which tended to be
slimy.
The formation of such microflocs may provide some
advantages for azospirilla in nature. Bergersen (1984) dis
cussed the strategies that microaerophilic, dinitrogen
fixing bacteria such as azospirilla may have to protect
their oxygen-sensitive nitrogenase. His suggestions are
incorporated below into some of the advantages that
azospirilla may find in growing as encapsulated microflocs.
1. The capsules may help to regulate the availability
of oxygen to dinitrogen-fixing cells. Assuming
the capsules to be highly hydrated (Costerton et
al., 1981) and assuming that water reduces the


184
Geesey, G. G. 1982. Microbial exopolymers: Ecological and
economic considerations. Am. Soc. Microbiol. News
48:9-14.
Goldschmidt, M. C., and 0. Wyss. 1966. Chelation effects
on Azotobacter cells and cysts. J. Bacteriol.
91:120-124.
Gonzalez-Lopez, J., M. V. Martinez-Toledo, J. Moreno,
F. Ballesteros, and A. Ramos-Cormezana. 1985.
Resistant properties of Azotobacter cysts induced in
phosphate-limited media. Microbios Letters 28:151-161.
Gonzalez-Lopez, J., and G. R. Vela. 1981. True morphology
of the Azotobacteraceae--Filterable bacteria. Nature
289:588-590.
Gray, T. R. G., and S. T. Williams. 1971. Microbial pro
ductivity in soil, p. 255-287. In D. E. Hughes and
A. H. Rose (ed.), Microbes and biological productiv
ity. Twenty-first Symposium of the Society for General
Microbiology. Cambridge University Press, Cambridge.
Griffin, G. J., M. G. Hale, and F. J. Shay. 1976. Nature
and quantity of sloughed organic matter produced by
roots of axenic peanut plants. Soil Biol. Biochem.
8:29-32.
Halsall, D. M., G. L. Turner, and A. H. Gibson. 1985.
Straw and xylan utilization by pure cultures of
nitrogen-fixing Azospirillum spp. Appl. Environ.
Microbiol. 49:423-428.
Hitchins, V. M., and H. L. Sadoff. 1970. Morphogenesis of
cysts in Azotobacter vinelandii. J. Bacteriol. 104:
492-498.
Hitchins, V. M., and H. L. Sadoff. 1973. Sequential
metabolic events during encystment of Azotobacter
vinelandii. J. Bacteriol. 113:1273-1279.
Hughes, W. H. 1956. The structure and development of the
induced long forms of bacteria, p. 341-362. In
E. T. C. Spooner and B. A. D. Stocker (ed.), Bacterial
anatomy. Sixth Symposium of the Society for General
Microbiology. Cambridge University Press, Cambridge.
Ingraham, J. L., 0. Maaloe, and F. C. Neidhardt. 1983.
Growth of the bacterial cell. Sinauer Associates,
Inc. Publishers, Sunderland, MA.


Figure 2-5. Cell types of Azospirillum brasilense
strain Cd, from lawns on B-hydroxybutyrate
(BHB) agar. a) Filaments from 63-hour-
old, high phosphate-BHB agar with combined
nitrogen. 1,500X magnification. Bar
equals 3.0 pm. b) Microfloc from
11-day-old nitrogen-free, low phosphate-
BHB agar, focused to show capsules and
filamentous cell outline. 1,000X magnifi
cation. Bar equals 4.0 pm. c) Same floe
as (b), but focused so that capsules and
filamentous cell outline are no longer
evident. 1,000X magnification. Bar
equals 4.0 pm.


93
c


28
themselves have not been found to afford any desiccation
resistance in laboratory studies with pure cultures in
nonsoil conditions (Dudman, 1977).
Because of the importance of desiccation as a limiting
factor in legume inoculation with Rhizobium spp., several
studies have been done on their resistance to drying. There
are broad strain differences in resistance of rhizobia to
desiccation. Many variables are present in drying
experiments, and the variables may interact with one
another. Rhizobium spp. withstand drying best in heavier-
textured soils, where hygroscopic water can be retained by
colloidal surfaces. Die off is far more rapid in drying
sand. Capsules do not afford increased resistance to drying
in studies with soil or other drying surfaces (Lowendorf,
1980). Often fewer rhizobia survive rapid drying
procedures, such as oven drying, than survive milder
desiccation over several weeks' time with controlled rela
tive humidities (Jansen van Rensburg and Strijdom, 1980).
Robinson et al. (1965) added pure cultures of
Pseudomonas spp. or Arthrobacter spp. to sterile soils. The
inoculated soils were dried by passing filtered air through
them for 2 days, by which time they had reached constant
weight. This forced drying resulted in rapid die off for
both species. Labeda et al. (1976) found that slow evapora
tive drying of inoculated soil resulted in reduced death
rates for both Pseudomonas spp. and Arthrobacter spp.


173
contained enough cells in the initial inoculum to be visibly
turbid, but the cells remained dispersed throughout the agar
and formed no pellicle. As a result, these treatments are
listed as giving no germination (Table 3-3). A few cells
seemed able to slowly mobilize their PHB reserves and become
motile under these conditions, but the majority of cells
remained in the floe, or once free from the floe, remained
nonmotile. Motile cells continued to retain extensive
visible deposits of PHB.
Undeniable germination occurred when the floes of
encapsulated cells were added to soil dialysis flasks, or to
buffered-salts solution containing nitrate or ammonium. The
uniformity of response among these treatments indicates that
combined nitrogen in the soil dialysate was responsible for
its germination effect. It also indicates that most of the
cells in the floes were not similar to mature Azotobacter
spp. cysts, which do not germinate in the presence of
ammonium (Loperfido and Sadoff, 1973). The availability of
combined nitrogen apparently prompted most of the cells in
encapsulated floes to mobilize their PHB reserves and return
to an actively motile, vegetative state.
An interesting feature of these positive germination
treatments was the persistence of nonmotile, PHB-rich cells
within floes even after 10 days of incubation. Some of
these cells no longer possessed a plump appearance, and
their PHB granules were dispersed irregularly within the


71
exceeding 33.8 kPa was applied. Filter membranes were then
removed and placed into Karnovsky's fixative (1965) for 1
hour. Filter membranes were subsequently rinsed twice for
10 min in cacodylate buffer and then dehydrated in a graded
series of ethanol concentrations (10, 20, 30, 50, 70, 90,
95, 100, and 100%) for 10 min at each concentration. The
samples were then air dried. Sections of filters were
excised, placed onto aluminum stubs with double-stick tape,
and gold coated with an Eiko IB-2 coater. Specimens were
examined with a Hitachi S450 scanning electron microscope at
20 kilovolts. Photographs were taken with Polaroid Type 55,
positive/negative, 4X5 Land film.
Light Microscopy
Cells were routinely observed by phase-contrast
microscopy using a Wild M20 or a Nikon Labophot microscope.
Cell dimensions were measured with an ocular micrometer.
Photographs of cells viewed with the latter microscope were
taken with a Microflex AFX camera attachment, using Ilford
FP4 black and white film. All photos were taken using
phase- contrast optics, unless otherwise indicated.
Results
Quantitative Morphological Change
In the initial phase of these studies only three
strains of azospirilla were used, A. brasilense strain


M
130
a
b


34
cells predominated. Photographs of cyst-enriched cultures
showed that many vegetative cells were still present. To
obtain cyst-free cultures, cells were grown in nutrient
broth, then washed and resuspended in saline. These cells
were then spotted onto sterile, predried nutrient agar films
so that the added cells would dry completely on the agar
film in 30 min at 30C. Agar films from each treatment were
then cut with sterile scissors and aseptically transferred
to vials containing silica gel. To test viability, the
dried agar films were removed periodically from the vials,
placed on nutrient agar plates, and incubated for 1 week at
30C. Vegetative cells did not survive the initial drying
process. Cyst-enriched populations that survived the
initial desiccation period remained viable for up to 15
months. Interestingly, cyst-enriched cultures of two root
isolates were nonviable at time zero, when they were placed
into the silica-gel vials (Lamm and Neyra, 1981).
Two aspects of this study deserve special comment.
Clearly, the cyst-enriched cultures did not receive the same
drying treatment as did the vegetative cells. The cyst-
enriched agar films were obtained by a slow drying process,
and the vegetative cell agar films underwent rapid drying.
It does not seem valid to compare their desiccation toler
ance under these different conditions. Also, two strains
that contained cyst-like cells of apparently mature
morphology were not desiccation resistant. Perhaps they


Figure 3-2. Encapsulated cells of Azospirillum
lipoferum Sp RG6xx that have undergone
germination in soil dialysis flasks. a)
Cells from a 29-hour-old incubation. Note
empty horseshoe-shaped capsules and phase-
dark vegetative cells. 1,000X magnifi
cation. Bar equals 4.0 pm. b) Other
cells from a 29-hour incubation. Note
empty horseshoe-shaped capsules, phase-
dark vegetative cells and poly-6-hydroxy-
butyrate-rich cells remaining in the
floe. 1,000X magnification. Bar equals
4.0 pm.


Figure 2-10-
Cell types of Azospirillum lipoferum
Sp RG6xx from lawns on B-hydroxybutyrate
(BHB) agar. a) Microfloc from 13-day-
old cell lawn on nitrogen-free HP-BHB
agar. Note empty capsules. 1,500X
magnification. Bar equals 3.0 pm. b)
Individual cells and septating filaments
from 13-day-old, high phosphate-BHB agar
with combined nitrogen. 1,500X magnifi
cation. Bar equals 3.0 pm. c) Micro
floc from 79-day-old, low phosphate-BHB
agar with combined nitrogen. 1,500X
magnification. Bar equals 3.0 pm.


162
a


101
b


38
When grown in nitrogen-free broth with glucose as the
carbon source, young cells of Azotobacter spp. appear as
rods with rounded ends, ranging from 1.3 to 2.7 pm in
diameter and 3.0 to 7.0 pm in length. As cultures age,
cells often accumulate PHB. Cell morphology may be altered
to ellipsoids, filamentous cells, or chains of cells (Tchan,
1984 ) .
Azotobacter spp. are commonly isolated from soil and
aquatic habitats of near-neutral pH, and are generally less
acid-tolerant than azospirilla. The most common species
isolated from soil is Azotobacter chroococcum, but its
biochemistry and physiology have received less attention
than that of Azotobacter vinelandii (Tchan, 1984).
Azotobacter vinelandii ATCC 12837 forms cysts profusely
under appropriate growth conditions. When this strain is
cultured in Burk's nitrogen-free broth with glucose, some
cysts form in stationary phase cultures, but ony 1.0% (Lin
and Sadoff, 1969) to 10.0% (Reusch and Sadoff, 1981) of the
population encysts under these conditions.
Early workers such as Winogradsky (1938) knew that
growing some Azotobacter spp. in nitrogen-free media, with
ethanol or butanol as carbon source, led to enhanced produc
tion of nonmotile, spherical cells with double-layered
coats. Socolofsky and Wyss (1961) built upon this
knowledge, using A. vinelandii ATCC 12837 (which' was used in
all the studies that follow unless otherwise indicated).


110
Growth of this A. lipoferum strain on nitrogen-free,
LP-BHB agar resulted in homogeneous encapsulation and fila
ment formation. But, as is evident from the photographs,
several cell shapes and sizes were present within any one
floe. Despite their morphological heterogeneity, cells in
these floes generally appeared to be more rounded and
swollen than cells grown on SNF-Congo Red agar, although
capsules were of equal width (0.5 pm) under both cultural
conditions.
The homogeneous encapsulation and filament formation of
A. lipoferum Sp RG6xx prompted a search for similar response
in other strains of this species. Three other strains of
this species, cultured on LP-BHB nitrogen-free agar,
responded about as well as A. lipoferum Sp RG6xx (Figures
2-18 to 2-20). Cells of all strains did not resuspend
uniformly in water, due to macroflocs. As was true for
A. lipoferum Sp RG6xx, the appearance of the other three
strains did not change noticeably with time, and several
cell sizes and shapes were usually present within any one
microfloc. Azospirillum lipoferum Sp A3a differed from all
the other A. lipoferum strains in consistently having large
numbers of individual, PHB-rich, nonmotile, nonencapsulated,
ovoid cells in its resuspended lawns. Possibly many of
these individual cells were initially present within
capsules, but were released from capsules upon the addition
of water. The other strains had only a few free,


51
this time, vegetative vibrioids emerged from the capsules to
grow and fix dinitrogen. The authors suggested that oxygen
limitation greatly affected these events. The level of PHB
increased as the oxygen level of the culture decreased;
nitrogenase activity ceased; and the cells encysted for a
time. Their apparent reduced respiratory activity allowed
the level of dissolved oxygen to be replenished in the
medium, until vibrioids emerged from the cyst coats to grow
and fix dinitrogen again. No encystment was observed when
cultures were incubated aerobically.
The recent work of Sadasivan and Neyra (1985) stressed
the roles that PHB and exopolysaccharides play in cyst
formation of azospirilla. Encysting cells lost their
motility and became enlarged and rounded. They accumulated
PHB and synthesized capsular material. The investigators
emphasized that common media, such as nutrient broth, do not
promote encystment and that development of mature exine and
intine layers may only be achieved under specific, well-
defined cultural conditions. Sadasivan (1985) may have
found the cultural conditions to promote maturation of cysts
of A. brasilense Sp 7. Using phase contrast microscopy, she
has observed vegetative cells emerging from cyst coats,
leaving behind empty "horseshoe"-shaped capsules. She has
also observed cysts containing from two to four central
bodies within a single exine. In transmission electron
microscopy thin sections, she has observed maturing cysts,


171
motile. There are numerous starvation conditions that
bacteria can be exposed to in vitro, and the response of
different strains can vary widely under different condi
tions. It would be interesting to follow up on these
initial studies of starvation resistance of azospirilla, to
gain further insight into how they might survive in soil in
the absence of plant material. It is possible that some
azospirilla are able to enter into two types of dormancy
(Marshall, 1980) in unfavorable soil conditions. If the
cells have been experiencing balanced growth before they are
starved of exogenous carbon, they might enter into exogenous
dormancy. Such cells would be poor in PHB and might have no
different morphology than growing vegetative cells, but
their metabolism would be greatly reduced. If the cells
have accumulated large amounts of PHB through dinitrogen
fixation or extremely rapid uptake of carbon sources during
growth on combined nitrogen, they might be prone to enter
constitutive dormancy or an encysted state when faced with
starvation.
Papen and Werner (1982) suggested that depletion of
available oxygen was responsible in part for encystment of
azospirilla in their studies. The low oxygen consumption of
encysted cells may have allowed oxygen to diffuse back into
the medium from the headspace, whereupon vegetative cells
emerged from the capsular coats and resumed dinitrogen
fixation. Because of their suggestion, in this study


55
spp. cysts. Krieg and Dobereiner (1984) restated this, but
the photographs of Berg et al. (1979) do not support it.
The multicellular C-forms are virtually indistinguishable
from Azotobacter spp. cysts having multiple central bodies
(Cagle and Vela, 1974). Clearly, in the association with
sugarcane callus, the azospirilla were situated in numerous
sites, differing in nutrient availability and oxygen avail
ability. It is not surprising that multiple morphologies
were observed, reflecting multiple physiological states.
Only a few cells resembling mature cysts were present.
Pleomorphic forms of azospirilla have been observed in
a variety of axenic associations with plant roots. The work
of Umali-Garcia et al. (1980) has already been discussed.
Ruscoe et al. (1978) grew maize plants in sand and inocu
lated them with different strains of azospirilla. Enlarged,
cyst-like cells, as well as cells of normal morphology, were
observed in older and thicker root segments, where root
tissue was often disintegrating. They also found that when
two strains of azospirilla were grown in nitrogen-free,
semisolid trans-aconitate agar, they often formed long
chains after 4 to 5 days.
Matthews et al. (1983) used immunological techniques
and transmission electron microscopy to observe strains of
A. brasilense in axenic association with pearl millet
roots. Both vibrioid and encapsulated cells were observed
in association with the roots. The encapsulated cells


Figure 3-1. Encapsulated cells of Azospirillum
lipoferum Sp RG6xx that have undergone
germination in low phosphate-basal salts
solution with combined nitrogen.
a) Cells from a 29-hour nitrate
incubation. Note germinated vegetative
cell at left of an empty capsule.
1,000X magnification. Bar equals 4.0 pm
b) Cells from a 29-hour ammonium
incubation. Note germinated vegetative
cells, empty capsules and poly-6-hydroxy
butyrate-rich cells remaining in the
floe. 1,000X magnification. Bar equals
4.0 pm.


61
C-forms as opposed to the vegetative or V-forms, as did some
later workers (Matthews et al., 1983; Krieg and Dobereiner,
1984). This terminology may be confusing, however, since
capsules can also occur on azospirilla of otherwise normal
morphology (Nur et al., 1980).
The presence of a capsule is usually deemed a pre
requisite for cyst formation in Azotobacter spp. (Eklund et
al., 1966). Azospirilla also may need to form a capsule
before they can form cyst-like cells. Encapsulated azo
spirilla may initially be fully active vegetative cells.
Upon encountering metabolic or environmental stress, such
cells may mature into cyst-like cells. The change in
morphology with time of some members within a C-form
population (Berg et al., 1980) may reflect maturation into
truly mature cysts. Two definitive traits of a mature
Azospirillum spp. cyst would be greatly reduced cell
metabolism and enhanced desiccation resistance. Morphologi
cally differentiated cells of azospirilla have been called
cysts when they exhibit no nitrogenase activity (Eskew et
al., 1977; Papen and Werner, 1982) or exhibit enhanced
desiccation resistance (Lamm and Neyra, 1981; Papen and
Werner, 1982; Sadasivan and Neyra, 1985).
Another complicating factor in understanding these
forms of azospirilla is that their appearance in dinitrogen
fixing cultures often coincides with alkalinization of the
growth medium (Eskew et al., 1977; Krieg and Dobereiner,


192
Vela, G. R-, and G. Cagle. 1969. Formation of fragile
cysts by a strain of Azotobacter chroococcum. J. Gen.
Microbiol. 57:365-368.
Volpon, A. G. T., H. De-Polli, and J. Dobereiner. 1981.
Physiology of nitrogen fixation in Azospirillum
lipoferum Br 17 (ATCC 29709). Arch. Microbiol. 128:
371-375.
Ward, J. B., and R. C. W. Berkeley. 1980. The microbial
cell surface and adhesion, p. 47-66. In R. C. W.
Berkeley, J. M. Lynch, J. Melling, P. R. Rutter, and
B. Vincent (ed.), Microbial adhesion to surfaces.
Ellis Horwood Limited Publishers, Chicester.
Webb, S. J. 1965. Bound water in biological integrity,
p. 146-171. Charles C. Thomas, Springfield, IL.
Whittenbury, R., S. L. Davies, and J. F. Davey. 1970a.
Exospores and cysts formed by methane-utilizing
bacteria. J. Gen. Microbiol. 61:219-226.
Whittenbury, R., K. C. Phillips, and J. F. Wilkinson.
1970b. Enrichment, isolation and some properties of
methane-utilizing bacteria. J. Gen. Microbiol.
61:205-218.
Winogradsky, S. 1938. Sur la morphologie et l'ecologie des
Azotobacter. Ann. Inst. Pasteur 60:351-400.
Wong, P. P., N. E. Stenberg, and L. Edgar. 1980. Charac
terization of a bacterium of the genus Azospirillum
from cellulolytic nitrogen-fixing mixed cultures.
Can. J. Microbiol. 26:291-296.
Wyss, O., D. D. Smith, L. M. Pope, and K. E. Olson. 1969.
Endogenous encystment of Azotobacter vinelandii.
J. Bacteriol. 100:475-479.
Zimmer, W., M. Penteado Stephan, and H. Bothe. 1984.
Denitrification by Azospirillum brasilense Sp 7 I.
Growth with nitrite as respiratory electron acceptor.
Arch. Microbiol. 138:206-211.


167
PHB-rich cells were observed. Each was rotating about its
own long axis without moving off on runs.
Discussion
The desiccation resistance assay used in these studies
involved rapid drying of the cells. Mature, encysted forms
of some prokaryotes are much better able to withstand drying
on glass surfaces than their vegetative counterparts
(Whittenbury et al., 1970a; Tudor and Conti, 1977). Filter
membranes are often used as a surface upon which cells are
slowly dried in desiccation resistance assays (Socolofsky
and Wyss, 1962). In such experiments (data not shown),
vegetative cells of azospirilla sometimes survive slow
drying on membrane filters without appreciable die off, as
do encapsulated cells. Webb (1965) has pointed out the dif
ficulties of using membrane filters in such assays. It was
thought that glass surfaces would be easier to use, with
less inherent hydrophilic behavior than membrane filters.
Rapid drying of cells within a day's time or less usually
causes a rapid and nearly complete die off of vegetative
cells for many genera (Robinson et al., 1965; Whittenbury et
al. 1970a; Tudor and Conti, 1977 ). The results of the work
reported here show no apparent significant difference in
response to rapid drying between vegetative or encapsulated
cells of A. lipoferum Sp RG6xx. This is in agreement with
studies where capsules have not enhanced the ability of


45
Pleomorphism of Azospirilla
Bacteria cultured in vitro can be extremely
pleomorphic. Only a few cells in a population may exhibit
abnormal morphology under some cultural conditions, but
sometimes the majority of a culture assumes unusual shapes.
Older cultures in the stationary growth phase can be
especially pleomorphic (Duguid and Wilkinson, 1961).
Hughes (1956) has reviewed the development of bacterial
filaments. Filamentous cells are usually as wide as normal
cells, but are several times longer and lack developed
septa. They are interesting because they are often fully
viable, unlike some pleomorphic or involution forms of
bacteria. Under suitable cultural conditions, a filament
may divide at several points along its length to produce
several cells of normal length. Filaments can be induced by
sublethal cell damage, interruption of balanced growth, dyes
and antibiotics, extremes of pH, refrigeration, and various
forms of radiation.
Slater and Schaecter (1974) emphasized how sensitive
bacterial cell division is to the factors mentioned above.
If sublethally stressed, rod-shaped bacteria may continue to
grow and form filaments. Filaments can also form during
very rapid growth in rich media, and will fragment into
individual cells when growth slows, or when the environment
becomes less nutritionally rich. Since cells arising from
fragmentation of filaments are usually of normal length, the


68
Harvest of Cell Lawns
To harvest lawns of azospirilla grown on n-butanol or
BHB agar, about 7.0 ml of sterile deionized water was asep-
tically poured across the surface of a cell lawn, and the
cells gently scraped from the agar surface with a flamed
wire loop. For PHB analyses and plate counts, the suspended
cells of one BHB agar plate were aseptically transferred to
another plate whose cells were in turn scraped off. This
was done to ensure that the cell suspension would not become
too diluted.
Enumeration
Vegetative cells from TSS-broth, or cells grown on BHB
agar, were diluted ten-fold in a series of dilution blanks
containing LP-basal salts solution. For enumeration, 0.1 ml
of cell suspension was aseptically spread on SNF-Congo Red
agar plates. Four plates were spread for each dilution.
Plates were incubated as described above for 5 days before
counting.
Dry Weight Determination and PHB Analysis
To assay PHB content of vegetative cells of A.
lipoferum Sp RG6xx, two 22-hour-old, HP-TSS cultures
(OD560 = 0.6) were pooled for centrifugation and washing as
described above, except that sterile deionized water was
used for washing. The final cell suspension was adjusted to


170
apparently lose viability due to starvation for exogenous
carbon. Most encapsulated cells within floes do not become
motile when confronted with starvation for exogenous
carbon. From 5 to 25% of cells within floes retained their
PHB after extended incubation and often underwent reduction
in size within their capsules. One interpretation of this
size reduction is that the cells were undergoing maturation
into physiologically mature cysts. In contrast to the
encapsulated cells, two different densities of vegetative
inoculum increased several fold during the same 9 days of
£
starvation to give about 10 CFU/ml. This apparent multi
plication to a certain cell density and continued viability
of vegetative cells faced with starvation have a precedent
in studies of other bacteria, such as Rhizobium japonicum
(Crist et al., 1984).
In a recent study, Tal and Okon (1985) reported that
vegetative, PHB-poor cells of A. brasilense strain Cd died
off to about 7% of their initial numbers after 130 hours of
starvation in sterile, 0.06 M potassium phosphate buffer.
In comparison, PHB-rich cells proliferated 2.3-fold over the
same time span. Their experimental conditions differed from
conditions reported here not only in using a different
bacterial species but also in incubating the cells in
phosphate buffer alone, without other salts. Both of their
cell types were apparently nonencapsulated and actively
motile, while the encapsulated cells in this study were not


44
cells, long before AR and AP are produced. Further support
for the importance of membranes may be found in studies
where mineral nutrient deficiencies lead to the production
of stress-resistant cysts which lack completed cyst coats
(Gonzalez-Lopez et al., 1985).
Germination of cysts has usually been defined as the
emergence of a growing, motile cell from the exine layer
(Socolofsky and Wyss, 1961). Loperfido and Sadoff (1973)
examined the germination of cysts exposed to glucose. Cysts
respired detectably within 2 min. after the addition of 1.0%
(wt/vol) glucose, and soon afterwards net synthesis of RNA
and protein became measurable. After 4 to 6 hours, the
central body had enlarged to occupy the volume of the
intine, and DNA synthesis and nitrogenase activity became
measurable. After 8 hours, a vegetative cell emerged from
the cyst coats, leaving behind an empty "horseshoe"-shaped
exine. Germination did not occur in the absence of oxygen.
Cysts also germinated in the presence of sugars other than
glucose. Germination did not occur in Burk's nitrogen-free
salts, indicating that the PHB reserves of the cysts could
not be mobilized to initiate germination. The addition of
0.25% (wt/vol) NH^+ did not lead to germination.
When cysts are germinated on glucose, some central
bodies divide within their cyst coats to form multiple
central bodies. Up to six central bodies have been observed
within one cyst coat (Cagle and Vela, 1974).


CHAPTER III
PHYSIOLOGICAL PROPERTIES OF ENCAPSULATED FLOCS OF
AZOSPIRILLUM LIPOFERUM Sp RG6xx
In Chapter II, a method was described for quantita
tively converting vegetative, cell-lawn inocula of
A. lipoferum strains into nonmotile, encapsulated cells
having extensive, intracellular PHB deposits. Although many
of the cells within encapsulated floes had vibrioid or ovoid
morphologies, some were rounded and cyst-like in appear
ance. Mature cysts have a lower endogenous metabolic rate
and greater ability than do vegetative cells to survive
carbon starvation (Sudo and Dworkin, 1973). Mature cysts of
Azotobacter spp. are known to be more resistant than
vegetative cells to environmental stresses, including
desiccation (Socolofsy and Wyss, 1962). Such mature cysts
germinate in phosphate buffer containing certain carbon
sources, but not in carbon-free phosphate buffer containing
ammonium, or in unamended phosphate buffer (Loperfido and
Sadoff, 1973). The central bodies of mature cysts are
violently and almost immediately expelled from their cyst
coats when cysts are suspended in Tris-EDTA (Socolofsky and
Wyss, 1961; Goldschmidt and Wyss, 1966; Lin and Sadoff,
1969; Page and Sadoff, 1975).
144


Figure 2-23. Cells of Azospirillum lipoferum Sp A3a
from a 33-day-old, nitrogen-free, high
phosphate-6-hydroxybutyrate broth cul
ture, pH 7.2, viewed by scanning
electron microscopy. a) Lower cell
layer is thoroughly encapsulated.
3,000X magnification. Bar equals
5.0 pm. b) Details of cell surfaces.
Note strands of material joining some
cells. 17,000X magnification. Bar
equals 0.5 pm.


183
Day, J. M., and J. Dobereiner. 1976. Physiological aspects
of ^-fixation by a Spirillum from Diqitaria roots.
Soil Biol. Biochem. 8:45-40.
Dobereiner, J. 1978. Influence of environmental factors on
the occurrence of Spirillum lipoferum in soils and
roots. In U. Granhall (ed.), Environmental Role of
Nitrogen-fixing Blue-green Algae and Asymbiotic
Bacteria. Ecol. Bull. (Stockholm) 26:343-352.
Dobereiner, J., I. E. Marriel, and M. Nery. 1976. Ecologi
cal distribution of Spirillum lipoferum Beijerinck.
Can. J. Microbiol. 22:1464-1473.
Donachie, W. D., K. J. Begg, and N. F. Sullivan. 1984.
Morphogenes of Escherichia coli, p. 27-62. In
R. Losick and L. Shapiro (ed.), Microbial development.
Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
Dudman, W. F. 1977. The role of surface polysaccharides in
natural environments, p. 357-414. In I. W. Sutherland
(ed.), Surface carbohydrates of the prokaryotic cell.
Academic Press, New York.
Duguid, J. P., and J. F. Wilkinson. 1961. Environmentally
induced changes in bacterial morphology, p. 69-99. In
G. G. Meynell and H. Gooder (ed.), Microbial reaction
to environment. Eleventh Symposium of the Society for
General Microbiology. Cambridge University Press,
Cambridge.
Eklund, C., L. M. Pope, and 0. Wyss. 1966. Relationship of
encapsulation and encystment in Azotobacter. J.
Bacteriol. 92:1828-1830.
Eskew, D. L., D. D. Focht, and I. P. Ting. 1977. Nitrogen
fixation, denitrification, and pleomorphic growth in a
highly pigmented Spirillum lipoferum. Appl. Environ.
Microbiol. 34:582-585.
Falk, E. C., J. Dobereiner, J. L. Johnson, and N. R. Krieg.
1985. Deoxyribonucleic acid homology of Azospirillum
amazonense Magalhaes et al. 1984 and emendation of the
description of the genus Azospirillum. Int. J. Syst.
Bacteriol. 35:117-118.
Foster, R. C., and G. D. Bowen. 1982. Plant surfaces and
bacterial growth: The rhizosphere and rhizoplane, p.
159-185. In M. S. Mount and G. H. Lacy (ed.), The
phytopathogenic prokaryotes, vol. I. Academic Press,
Inc., New York.


96
old lawns grown on it were always resuspended uniformly,
without macroflocs. Older cultures on this medium consisted
mainly of PHB-rich, nonmotile ovoids or peanut-shaped cells,
with few if any microflocs. The high phosphate level did
not appear to inhibit PHB accumulation, but did inhibit
extensive capsule formation. As was true for the
A. brasilense strains, combined nitrogen led to eventual
good growth and passage into stationary phase. Figure 2-10b
shows cells of a 13-day-old culture from HP-BHB agar
containing combined nitrogen. Chains of cells and
individual ovoids are present.
The LP-BHB lawns grown with or without combined nitro
gen had the same appearance by 7 days. The lawns consisted
almost entirely of floes that broke into various sizes when
resuspended in water. The cells would not suspend evenly in
water, due to the presence of many macroflocs. Very few
motile cells were present at this time. Eventually the
LP-BHB lawns grown with combined nitrogen resumed vegetative
growth and passed into stationary phase, but the floes
persisted even in stationary phase cultures (Figure 2-10c).
Figure 2-11 shows cells from 17-day-old, nitrogen-free,
LP-BHB-Congo Red lawns. Figure 2-lla was taken with
bright-field optics, showing the clearly outlined capsules
and enlarged PHB-rich cells. Figure 2-llb was taken with
phase-contrast optics, and the capsules enclosing all of the
microfloc are again evident. It was interesting to find


75
After 63 hours of growth, A. brasilense strain JM 125A2
lawns grown on BHB contained ovoids, vibrioids, and chains
of cells. Many cells contained phase-bright, putative PHB
granules. More cells were present at this time on agar
containing combined nitrogen. The several cell types
present on 63-hour-old, LP-BHB agar with combined nitrogen
are shown in Figure 2-2a. Some cells appeared at this time
to be undergoing plasmolysis on this medium, as well as on
HP-BHB agar with combined nitrogen. By 96 hours, the lawns
on LP-BHB agar with and without combined nitrogen contained
more chains of cells and microflocs than the HP-BHB lawns,
which consisted mostly of individual ovoids or pairs of
ovoids.
Cells from month-old, nitrogen-free, HP-BHB lawns of
A. brasilense strain JM 125A2 are shown in Figure 2-2b.
Individual, nonmotile vibrioids and ovoids were still pre
dominant, as were pairs of cells. Enlarged, nonmotile,
spherical cells were present, but not numerous. A few non
motile filaments appeared to be undergoing septation.
After 79 days, cells from lawns of this strain grown
with combined nitrogen had the appearance of stationary
phase cells from TSS broth cultures grown with combined
nitrogen, and spheroplasts and cell ghosts predominated.
Nitrogen-free cultures at both phosphate buffer concentra
tions contained numerous pleomorphic forms. Figure 2-3a
shows nonmotile, enlarged, rounded individual cells from


CHAPTER II
PLEOMORPHISM OF AZOSPIRILLA GROWN ON
BETA-HYDROXYBUTYRATE
Only a few bacterial genera contain strains known to
form cysts (Sudo and Dworkin, 1973; Whittenbury et al.,
1970a; Tudor and Conti, 1977). A nonmotile cyst forms when
the entirety of a vegetative cell rounds up, depositing
extracellular coats and often accumulating intracellular
energy-reserve polymers.
The morphological changes of encystment are accompanied
by a reduction in cell metabolic activities, and increased
resistance to environmental stresses, such as starvation and
desiccation. Cysts of Azotobacter spp. are perhaps the best
understood. Like other prokaryotic resting cells, they form
when vegetative cells undergo a metabolic shift-down
(Sadoff, 1975).
Cysts of Azotobacter spp. do not form in media support
ing good vegetative growth until stationary phase, and are
present then only in low numbers (Sadoff et al., 1971).
Similarly, cells of azospirilla are uniform in shape during
active growth in nutritionally complete media (Umali-Garcia
et al., 1980; Lamm and Neyra, 1981; Sadasivan and Neyra,
1985). As is true for Azotobacter spp., however, stationary
phase cultures of azospirilla grown on complete media often
59


6
ATP derived from denitrification can be used to drive nitro-
genase activity (Scott et al., 1979), but it seems unlikely
that dinitrogen fixation under these conditions can support
growth of the bacteria (Bothe et al., 1981). Recent work by
Neuer et al. (1985) has shown that, in axenic wheat-
Azospirillum spp. associations, both dinitrogen fixation and
denitrification can occur.
Morphology of Azospirilla
Azospirilla are Gram-negative bacteria (Tarrand et al.,
1978). The structural layers external to the cytoplasmic
membrane of Gram-negative prokaryotes have been reviewed
(Costerton et al., 1974). Depending on cultural conditions
and the bacterial strain, polysaccharide or capsular layers
may be present as the outermost layers of the cell.
A growing Gram-negative cell divides by binary fission
to produce two daughter cells of approximately equal size.
Division begins with invagination of the cytoplasmic mem
brane and peptidoglycan, until a complete transverse septum
or cross wall is formed. When the septum is completely
formed and cleaved, the two daughter cells separate (Leive
and Davis, 1980). As will be discussed later, this cell
division process can be disrupted, resulting in formation of
filaments or chains, which accounts partially for pleo-
morphism of azospirilla.


47
gene products are involved directly in repair of damaged
DNA, but some others specifically block further cell divi
sion. Until the DNA is repaired, cell division is blocked,
but cells can continue to grow into long, nonseptate fila
ments. Upon repair of the DNA, septa form along the fila
ments, and cells of normal size are produced after septum
separation (Donachie et al., 1984). Certain E. coli mutants
are known to produce septa, but form chains because the
enzymes needed for septum cleavage are not produced (Begg
and Donachie, 1985).
Thompson and Skerman (1979) showed that most members of
the Azotobacteraceae are pleomorphic under certain cultural
conditions. Filaments and chains of cells are produced
commonly. Similar pleomorphism has been observed with
azospirilla.
Becking (1982) observed that the morphology of
azospirilla varied in different culture media. On yeast
extract-glucose agar, the cells were highly motile, slightly
curved rods, 2.0 to 4.0 pm long and 1.0 pm wide. These
cells would often become swollen with three to five PHB
granules per cell. When cultured in nitrogen-deficient
broth supplemented with 0.01% (wt/vol) Difco yeast extract,
the cells often became long spirals of 30 to 40 pm in
length. These cells had reduced motility, but were capable
of rotation about their axes, and had few or no PHB
granules. Peptone was found to produce similar elongated,


62
1984). Krieg and Dobereiner (1984) suggest that these cell
forms arise mainly at excessively high pH. In this case
they might be only laboratory artifacts, or involution
forms, that have no in situ function. The findings of Lamm
and Neyra (1981), Papen and Werner (1982), and Sadasivan and
Neyra (1985) argue against this viewpoint. Indeed, the
ability of azospirilla to enter dormancy as cysts may help
explain some of the great variability of plant responses to
inoculation with these bacteria (reviewed by Patriquin et
al., 1983).
Two things are presently lacking in research and under
standing of cyst-like forms of azospirilla. Although cyst
like forms of azospirilla have been predominant in some
studies, growing cells of normal morphology (vegetative
cells) have always been present in high numbers as well.
Conversion of 90% or greater of a population of vegetative
azospirilla to cyst-like forms (quantitative encystment) in
a reproducible manner would greatly facilitate further study
of these cell forms, as it did for Azotobacter spp. cysts
(Socolofsky and Wyss, 1962). Also lacking is an understand
ing of the underlying causes of pleomorphism and cyst
formation in azospirilla.
Conversion of 90% or greater of a cell population of
Azotobacter spp. to cysts often can be achieved by culturing
vegetative cells in the absence of combined nitrogen on
either of two precursors of PHB, n-butanol or BHB (Sadoff,


39
When cultured as cell lawns on Burk's nitrogen-free agar
with 0.3% (vol/vol) n-butanol as sole carbon source, cysts
began to appear within 3 days and predominated in 5 to 7
days. Ultrastructural studies revealed that the outermost
layer of the cyst, the exine, consisted of several over
lapping, plate-like layers. Beneath the exine was a much
thicker layer of gelatinous material, called the intine.
The intine surrounded a modified resting cell, called the
central body, which often contained numerous PHB granules.
Cysts had no detectable endogenous respiration when
suspended in buffer, but almost instantaneously began
measurable respiration when exogenous carbon sources were
added. In later studies, cysts were produced by growth on
0.2% (vol/vol) n-butanol (Socolofsky and Wyss, 1962), or
0.2% (wt/vol) 8-hydroxybutyrate (BHB) (Lin and Sadoff,
1968), with 90% or greater of the cells being converted to
cysts in 5 to 7 days.
Eklund et al. (1966) demonstrated that the formation of
capsular layers by vegetative cells was a prerequisite for
cyst formation. Complete morphological encystment of cells
grown on n-butanol agar with various levels of NH^NO^ only
occurred in the usual 5-day-period when the NH^NO^ concen
tration was 0.02 M or less. The cells rounded up within 5
days into nonmotile precysts lacking exines when 0.03 M or
0.04 M NH^NO^ was initially present. By day 10, these cells
had used up enough of the original combined nitrogen to


Figure 2-1. Zoogleal masses in stationary phase
40-dayold, low phosphate-trypticase-
succinate-salts broth culture of
Azospirillum lipoferum Sp RG6xx. a) Cells
at 600X magnification. Bar equals 6.0
pm. b) Detail from same mass of cells
viewed at 1,500X magnification. Bar
equals 3.0 pm.


149
Microaerobic Incubation
The same washed BHB-grown cells were used for these
experiments as for the previous experiments. The incubation
medium consisted of LP-basal salts solution containing
Bacto-Agar (Difco). The agar was added to give a final
concentration of 0.05% (wt/vol) per flask after cell
addition. The basal salts solution and agar were dissolved
by boiling, then 23.5 ml were added per 125-ml Erlenmeyer
flask. The flasks were autoclaved, and concentrated sterile
LP buffer and biotin were added aseptically soon after
autoclaving and before cell addition. A volume of 6.0 ml of
encapsulated cell suspension was added to each flask.
Flasks were prepared in triplicate and incubated in
stationary position at 30C.
Aerobic Incubation
The BHB-grown, encapsulated cell inoculum for this and
all following experiments was harvested, washed twice in
sterile deionized water, and resuspended in a third volume
of sterile deionized water to give a final ODr, of 0.23 to
0.25. Lawns of 58 to 66 days of age were used as inocula.
For the incubation solution, basal salts were dissolvsed in
concentrated amounts to give their final, correct concentra
tions after aseptic additions of LP buffer, biotin, carbon,
or nitrogen sources and cells. The biotin was aseptically
added as a 100X concentrated stock solution, and all other


106
Encapsulated cells of Azospirillum
lipoferum Sp RG6xx from a 75-day-old
lawn on nitrogen-free, low phosphate-B-
hydroxybutyrate agar. Cells are viewed
at 1,500X magnification by scanning
electron microscopy. Bar equals 5.0 pm.
Figure 2-15.


Figure 2-9. Cell types of Azospirillum lipoferum
Sp RG6xx, from 63-hour-old lawns on
B-hydroxybutyrate (BHB) agar. a) Micro-
flocs from nitrogen-free, low phosphate-
BHB agar, focused to show capsules and
filamentous cell outlines. 1,500X
magnification. Bar equals 3.0 pm.
b) Same microflocs as (a) but focused so
that capsules and filamentous cell
outlines are no longer evident.
1,500X magnification. Bar equals 3.0 pm.
c) Filament from high phosphate-BHB agar
with combined nitrogen. 1,500X magnifi
cation. Bar equals 3.0 pm.


103
b


3
(Okon and Hardy, 1983). Short-term axenic associations have
also resulted in enhanced uptake of mineral ions by grass
roots (Lin et al., 1983). This effect may be due to the
influence of plant growth substances, or to softening of the
middle lamellae of root cells by pectolytic bacterial
enzymes, which some azospirilla are known to produce
(Umali-Garcia et al., 1980; Tien et al., 1981). Such
effects on root morphology and activity may make inoculation
with azospirilla beneficial in some agricultural situations.
The rhizosphere environment is prone to extreme chemi
cal and physical fluctuations (Foster and Bowen, 1982).
This may lead to periods when azospirilla are inactive due
to environmental limitations. Cells of azospirilla can vary
morphologically (Krieg and Dobereiner, 1984). Some of these
cell forms may be dormant or resting stages, in which
activities of possible benefit to plants are not expressed.
Pleomorphic forms of azospirilla usually possess capsules,
and contain large amounts of the reserve polymer poly-B-
hydroxybutyrate (PHB). This study describes attempts to
obtain such forms in high numbers by laboratory culture.
The general topics of capsules, PHB, physiological dormancy,
and desiccation resistance are directly related to this
study, and will be briefly reviewed in this introduction
after discussion of some key aspects of Azospirillum spp.
physiology.


64
A. lipoferum Sp A3a (all courtesy of N. R. Krieg, Va. Poly.
Inst., Blacksburg). All strains were maintained on slants
of Tryptic Soy Agar (Difco Laboratories, Detroit, MI) at
25C with monthly transfer.
Media
Vegetative azospirilla were cultured in a modification
of the complete medium of Tyler et al. (1979), denoted as
trypticase-succinate salts (TSS). All components were of
reagent grade and were dissolved in deionized water. The
final concentrations of TSS components were (in grams per
liter): (NH^J^SO^, 0.5; succinic acid, 0.437; Trypticase
Peptone (Baltimore Biological Laboratory, Cockeysville, MD),
1.0; d-biotin (Sigma Chemical Co., St. Louis, MO), 0.0001;
NaCl, 0.1; FeCl3'6H20, 0.0017; Na2Mo04'2H20, 0.0002;
MgS04'7H20, 0.2; and CaCl2, 0.002. The first four com
ponents were omitted to obtain a basal salts solution. The
biotin was dissolved as a 100X concentrated stock solution
by heating and then filter-sterilized by passage through a
0.2 pm pore diameter Nalgene filter unit (Nalge Company,
Rochester, NY). Two phosphate buffer concentrations were
employed. The low phosphate (LP) buffer of Tyler et
al. (1979) had a final concentration of 3.5 mM and consisted
of (in grams per liter) K2HPC>4, 0.1 and KH2P04, 0.4. The
high phosphate (HP) buffer of Albrecht and Okon (1980) had a
final concentration of 63.8 mM and consisted of (in grams


77
a
b


141
diffusion of oxygen by a factor of 10,000 (Clark,
1967), such a role for capsules is not improb
able. If the capsular material remains pliable,
the cells within an encapsulated microfloc may be
able to move closer together or further apart as
the situation requires. When oxygen is in excess,
they may move closer together, reducing the oxygen
tension in one spot and thus allowing continued
nitrogenase activity at that localized site. When
oxygen is limiting, the cells may move further
apart, the separation allowing each cell's nitro
genase to remain functional.
The spaces between encapsulated cells may provide
sites for other bacteria to enter into intimate
association with the microfloc, to act in
cross-feeding, or to help reduce local oxygen
tension.
Encapsulation would provide the general benefits
to azospirilla that most bacteria seem to derive
from encapsulation (Costerton et al., 1981).
These benefits include protection from predation
and enhanced nutrient accumulation and uptake.
The sustained rigid structure of microflocs sug
gests that nutrients within the encapsulated
microflocs may be sequestered from the surrounding
environment, giving the azospirilla a storehouse


105
Encapsulated cells of Azospirillum
lipoferum Sp RG6xx from a 75-day-old
lawn on nitrogen-free, low phosphate-8-
hydroxybutyrate agar. Cells are viewed
at 500X magnification by scanning elec
tron microscopy. Bar equals 50.0 pm.
Figure 2-14.


89
Figure 2-7. Microfloc of Azospirillum brasilense
strain Cd, from 6-day-old lawn on
succinate-nitrogen-free-Congo Red agar.
Note capsules around cells and empty
capsules. 1,500X magnification. Bar
equals 3.0 pm.


11
were sometimes seen to be enclosed within a common envelope
or capsule. Such structures were not observed when the
bacteria were grown in trypticase soy broth (Umali-Garcia et
al., 1980). This is another indication that the low C/N
ratio of complex broth media can repress extensive capsule
formation by azospirilla, while the high C/N ratio near
plant roots can promote capsule formation.
Recent work by Sadasivan and Neyra (1985) verified that
the forms of carbon and nitrogen made available to azo
spirilla can have a profound effect on exopolysaccharide
synthesis. When A. brasilense Sp 7 and A. lipoferum Sp. 59b
(ATCC 29707 ) were cultured in broth containing 8.0 mM
fructose and 0.5 mM KNO^ they grew as individual motile
cells for only 6 hours and then started to clump, as exo
polysaccharide production led to floe formation. Organic
acids yielded fewer floes than did sugars, and other nitro
gen sources, such as ammonium, yielded fewer floes than did
nitrate. The cells in floes appeared initially to be
enmeshed in a loose, fibrillar matrix that condensed
progressively over a week's time. When cells were grown,
harvested by centrifugation, and resuspended in broth lack
ing carbon, the cells remained freely suspended. This
suggests that azospirilla have a high ATP demand for exo
polysaccharide synthesis. Chemical analysis showed that
cellulose was a major component of the exopolysaccharide.


10
exopolysaccharides may aid the survival of bacteria (Dudman,
1977). Exopolysaccharides can concentrate nutrients from
the surrounding solution phase. They give some bacteria
increased resistance to antibiotics, surfactants, and other
chemicals, as well as deterring their engulfment by
phagocytic cells (Costerton et al., 1981). Other advantages
of exopolysaccharides have been suggested, such as mediation
of gas exchange between bacteria and their surroundings, but
they have proven difficult to prove experimentally. Extra
cellular enzymes might also be located within or at the
surface of capsules (Geesey, 1982).
Nur et al. (1980) found that A. brasilense Sp 7 and an
Israeli isolate of A. brasilense both possessed small cap
sules discernible by electron microscopy when grown on
nutrient agar. Umali-Garcia et al. (1980) found that when
certain A. brasilense strains and grass seedlings were
incubated together for 10 to 30 min at 30C, many bacteria
adhered to the grass roots, with granular material
accumulating on the surfaces of root hairs, and fibrillar
material accumulating on the surfaces of older, epidermal
root cells. It is known that bacterial exopolysaccharides
may appear either granular or fibrous (Foster and Bowen,
1982). The A. brasilense strains seemed to rapidly produce
both types of exopolysaccharides in axenic association with
grass roots. After 2 to 4 days of axenic association with
grass roots, from two to four cells of A. brasilense Sp 7


153
Chloramphenicol Treatment
A 0.2% (wt/vol) solution of chloramphenicol (Sigma) in
deionized water was sterilized by passage through a 0.2-pm
pore size, Nalgene filter unit. Various volumes of this
stock solution were added to water-washed, encapsulated
cells in sterile Nutrient Broth (Difco), having a final
concentration after all additions of 0.8% (wt/vol). Usually
2.0 ml of cells were added to these tubes. Incubation was
in 50-ml screw cap tubes, under the same conditions for
tubes as described above.
Phase-Contrast Microscopy and Photographs
These were the same as in Chapter II.
Results
Desiccation Resistance
Neither the vegetative nor the encapsulated cells
displayed significant resistance to the drying method
employed. Virtually all cells lost viability during the 8
days of desiccation. Table 3-1 gives the results of two
separate desiccation experiments. No statistical analyses
were performed, because even if some difference could be
revealed between vegetative and encapsulated cells, the
survival of either was so poor as to be negligible.


121
Cells of Azospirillum lipoferum Sp
RG6xx, from 14-day-old, stationary
phase, high phosphate-trypticase-suc-
cinate-salts broth culture of pH 9.1.
Cells are viewed at 1,500X magnification
by scanning electron microscopy. Bar
equals 5.0 pm.
Figure 2-21.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZTWB5LV5_CJLM9P INGEST_TIME 2014-12-18T23:10:26Z PACKAGE AA00026412_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


156
Table 3-2
. Starvation resistance
Sp RG6xx.
of Azospirillum
lipoferum
Experi
ment
Number
Cell
type
Initial.
CFU ml 1
(X 106)a
Final .
CFU ml 1
(X 106)b
% Initial
CFU ml 1
1
Vege
tative
0.72
(0.03)
1.02 (0.26)
142.04
Encapsu
lated0
0.42
(0.06)
0.11 (0.01)
26.61
2
Vege
tative
0.14
(0.03)
1.02 (0.22)
751.23
Encapsu
lated0
1.34
(0.10)
0.33 (0.18)
24.22
aAverages
plates.
of colony
Values in
forming units
parentheses are
(CFU) of four spread
standard deviations.
Averages of three starvation tubes, with four spread plate
counts per tube. Values in parentheses are standard
deviations.
C75-day-old cells.
53-day-old cells.


CHAPTER I
INTRODUCTION AND EXPERIMENTAL APPROACH
Ecology of Azospirilla
Bacteria of the genus Azospirillum have been isolated
from soils and from the roots of cereal crops and forage
grasses in several areas of the world (Dobereiner et al.,
1976; Tyler et al., 1979; Lamm and Neyra, 1981). Their
4
numbers in nonrhizosphere soil can be as high as 10 cells/g
soil (Dobereiner, 1978), while their numbers in rhizosphere
soil can be as high as 10^ cells/g soil (Krieg and
Dobereiner, 1984).
Agricultural interest in Azospirillum spp. resulted
from recognition of their ability to reduce atmospheric
dinitrogen. The enzyme catalyzing this reaction, nitro-
genase, is inactivated in the presence of combined nitrogen
or oxygen (Okon et al., 1976a). Azospirilla fix dinitrogen
under microaerophilic conditions in nitrogen-free media in
the laboratory (Day and Dobereiner, 1976; Okon et al.,
1976a). Nonrhizosphere soil is usually too poor in avail
able, utilizable carbon sources to enable Azospirillum
spp. to fix dinitrogen, but they can do so in the more
carbon-rich rhizosphere environment (Dobereiner et al.,
1976). Maximum nitrogenase activity with inoculated plants
1


48
weakly motile cells containing little PHB. These cells were
probably filaments, as described by Hughes (1956). Becking
did not study their viability.
Eskew et al. (1977) isolated and studied the pigmented
A. brasilense strain Cd. Nitrogenase activity peaked after
about 2 days of growth in semisolid, nitrogen-free malate
medium, and most cells were motile, curved rods of normal
size, often containing PHB. After 3 days, however, nitro
genase activity declined sharply. By this time, the initial
near-neutral pH of the medium had risen to pH 8.1. Most of
the bacteria present then appeared as enlarged, ovoid, non-
motile cells that were resistant to Gram-staining. The
decrease in nitrogenase activity, and shift to alkaline pH,
coincided with the appearance of cyst-like cells.
Tarrand et al. (1978) found that A. brasilense and
A. lipoferum strains had a similar appearance after 1 day's
growth in broth containing peptone, ammonium sulfate, and
succinate. Most cells were short, plump, slightly-curved
motile rods averaging 1.0 pm in diameter and 2.1 to 3.8 pm
in length. Cell morphology changed, especially for
A. lipoferum strains, when the cells were inoculated into
nitrogen-free, semisolid malate medium containing 0.005%
(wt/vol) yeast extract. Cells of A. lipoferum tended to
increase to 1.4 to 1.7 pm in width and to 5.0 pm to over 30
pm in length. Within 1 to 2 days, many A. lipoferum cells
became S-shaped or helical and retained little if any




136
azospirilla do not become pleomorphic during balanced growth
with combined nitrogen. Pleomorphic forms are only observed
in such cultures in stationary phase (Lamm and Neyra, 1981;
Papen and Werner, 1982; Krieg and Dobereiner, 1984), after a
nutritional down-shift has occurred. Media with high C/N
ratios have more often resulted in pleomorphism of
azospirilla (Tarrand et al., 1978; Papen and Werner, 1982;
Sadasivan and Neyra, 1985). Such conditions generally
promote the formation of PHB (Dawes and Senior, 1973) and
exopolysaccharides (Sutherland, 1977; Costerton et al.,
1981) by bacteria. High C/N ratios may also foster unbal
anced growth of azospirilla, leading to pleomorphism. The
formation of PHB and capsules by azospirilla seems to lead
to a pleomorphic cell type that may reach maturity as a
cyst.
In these experiments, azospirilla inocula were grown in
HP-TSS broth containing combined nitrogen. It is probable
that, in this rich medium, the cells had most of their
biosynthetic operons repressed, so their biosynthetic
enzymes were present only at low levels. After harvest and
washing, these cells were exposed to a new carbon source
(BHB) and deprived of combined nitrogen, forcing the cells
into a severe metabolic shift-down requiring synthesis of
biosynthetic enzymes (Ingraham et al., 1983). The
azospirilla in two-step, replacement cultures were starved
for nitrogen, since azospirilla cannot fix dinitrogen


27
outer layers. They are nonmotile and have low endogenous
respiration rates. They only initiate growth into vegeta
tive cells when sufficient nutrients are available. They
must also possess more resistance to some environmental
stresses than do vegetative cells. Enhanced resistances to
starvation and desiccation are probably traits of all mature
cysts. The cysts of the methane-oxidizing bacteria and of
Azotobacter spp. possess these characteristics. Mature
cysts of azospirilla should also have these properties.
Desiccation resistance is a critical characteristic of
prokaryotic cysts. The next section will consider experi
ments conducted to assess the resistance of bacteria to
drying.
Resistance of Bacteria to Drying
Clark (1967) stated that the majority of soil bacteria
survive in air-dried soils, often for several years. When
such soils are rewetted, bacterial activities including
nitrification, ammonification, nonsymbiotic dinitrogen
fixation, and sulfur oxidation are usually detected. The
implication is that the intimate association of bacteria
with clay or organic matter allows bacterial survival at
hydrated microsites in a macroscopically dry soil. Later
findings, reviewed by Stotzky (1980) and Marshall (1980),
support this. Exopolysaccharides may help bacteria to
achieve such intimate association, although capsules


95
a
c


125
pipette to the filter membrane. It appears that large,
encapsulated cell floes settled first onto the filter,
followed by free, nonflocculated cells. The lower layers of
cells have the same close-fitting appearance as was often
observed on BHB agar surfaces.
Similar results were obtained for the other three
strains of A. lipoferum. Figure 2-23a shows cells from a
33-day-old culture of A. lipoferum Sp A3a of pH 7.2. Some
encapsulated cells in the lower cell layer are fitted
together snugly, while others are joined by strands of
putative exopolysaccharide. Figure 2-23b is a higher
magnification of cells from this culture, again showing
the strands joining cells. The lumpy appearance of the
cells is probably due to large, intracellular accumulations
of PHB.
Figure 2-24 shows cells from a 33-day-old culture of
A. lipoferum Sp RG20a of pH 7.2. The formation of filaments
and eventually chains was very pronounced in this strain.
Cells from a 33-day-old culture of A. lipoferum Sp RG8c
of pH 7.1 are shown in Figure 2-25a. This strain was some
times observed to form intricately structured clumps.
Figure 2-25b shows such a clump from a 9-day-old culture of
pH 7.0. Filamentous, septate cells are present, and again
empty spaces occur within the floe.
The two A. brasilense strains used in this study
responded poorly to two-step broth replacement. The pH of


98


107
Figure 2-16. Encapsulated cells of Azospirillum
lipoferum Sp RG6xx from a 75-day-old
lawn on nitrogen-free, low phosphate-B-
hydroxybutyrate agar. Cells are viewed
at 7,000X magnification by scanning
electron microscopy. Bar equals 5.0 pm.


5
uptake of 02 by the bacteria, allowing both cell respiration
and protection of the oxygen-sensitive nitrogenase (Okon et
al., 1976a). Broth cultures can fix dinitrogen if the
dissolved oxygen level is well controlled (Okon et al.,
1976a). The bacteria are also able to grow on the surface
of nitrogen-free, aerobically incubated agar plates (Day and
Dobereiner, 1976).
The flagella of azospirilla enable them to move to
whatever sites their physiological state demands. They have
been shown to exhibit aerotaxis to microaerophilic sites
(Barak et al., 1982). Alternatively, cells may aggregate,
thereby creating a microaerophilic environment by the
respiration of many cells in a small space (Barak et al.,
1982). The grass rhizosphere may contain microaerophilic
sites (reviewed by Patriquin et al., 1983). Azospirilla
could migrate from soil toward such sites, where nitrogenase
activity could subsequently be expressed.
The respiratory metabolism of Azospirillum spp.
includes the ability of many strains to denitrify, reducing
nitrate or nitrite to more reduced nitrogenous compounds
under anaerobic conditions if enough metabolizable carbon
source is available (Neyra and Dobereiner, 1977; Neyra and
van Berkum, 1977; Nelson and Knowles, 1978). Under certain
laboratory conditions, denitrification has been shown to
provide enough ATP to support anaerobic growth of
azospirilla (Bothe et al., 1981; Zimmer et al., 1984). The


21
This study (Tal and Okon, 1985) also suggested that
elevated PHB levels at the onset of starvation may spare the
use of protein to drive endogenous metabolism. The PHB-poor
cells used up two-thirds of their initial protein during the
first 80 hours of starvation, whereas the protein content of
starved PHB-rich cells increased slightly over 80 hours. It
was also reported that PHB-rich cells were able to survive a
variety of environmental stresses, including desiccation,
better than PHB-poor cells (Tal and Okon, 1985).
The previous study also found that cells enriched in
PHB displayed a one hundred-fold higher aerotactic response
than PHB-poor cells. This supports the claim made in an
earlier study that PHB reserves could be used for aerotaxis
when no exogenous carbon source was available (Barak et al.,
1982 ) .
The previous discussion has shown that both capsule and
PHB synthesis can be promoted by environments with high
available C/N ratios. The roles of capsules and PHB in
pleomorphism in azospirilla will be discussed later. The
nature of dormancy in prokaryotic cells will be discussed
first, since some pleomorphic forms of azospirilla may be
dormant stages. Capsular layers and PHB are often present
in dormant forms of prokaryotes.


188
Nur, I., Y. Okon, and Y. Henis. 1980. Comparative studies
of nitrogen-fixing bacteria associated with grasses in
Israel with Azospirillum brasilense. Can. J.
Microbiol. 26:714-718.
Nur, I., Y. Okon, and Y. Henis. 1982. Effect of dissolved
oxygen tension on production of carotenoids, poly-g-
hydroxybutyrate, succinate oxidase and superoxide
dismutase by Azospirillum brasilense Cd grown in
continuous culture. J. Gen. Microbiol. 128:2937-
2943.
Okon, Y., S. L. Albrecht, and R. H. Burris. 1976a. Factors
affecting growth and nitrogen fixation of Spirillum
lipoferum. J. Bacteriol. 127:1248-1254.
Okon, Y., S. L. Albrecht, and R. H. Burris. 1976b. Carbon
and ammonia metabolism of Spirillum lipoferum.
J. Bacteriol. 128:592-597.
Okon, Y., and R. W. F. Hardy. 1983. Developments in basic
and applied biological nitrogen fixation, p. 5-54. Iji
F. C. Steward and R. G. S. Bidwell (ed.), Plant
physiology, a treatise, vol. VIII. Academic Press,
Inc., Orlando.
Okon, Y., P. G. Heytler, and R. W. F. Hardy. 1983. N^
fixation by Azospirillum brasilense and its
incorporation into host Setaria itlica. Appl.
Environ. Microbiol. 46:694-697.
Page, W. J., and H. L. Sadoff. 1975. Relationship between
calcium and uronic acids in the encystment of
Azotobacter vinelandii. J. Bacteriol. 122:145-151.
Papen, H., and D. Werner. 1980. Biphasic nitrogenase
activity in Azospirillum brasilense in long lasting
batch cultures. Arch. Microbiol. 128:209-214.
Papen, H., and D. Werner. 1982. Organic acid utilization,
succinate excretion, encystation and oscillating nitro
genase activity in Azospirillum brasilense under micro-
aerobic conditions. Arch. Microbiol. 132:57-61.
Patriquin, D. G., J. Dobereiner, and D. K. Jain. 1983.
Sites and processes of association between diazotrophs
and grasses. Can. J. Microbiol. 29:900-915.
Pope, L. M., and 0. Wyss. 1970. Outer layers of the
Azotobacter vinelandii cyst. J. Bacteriol. 102:
234-239.


4
Physiology of Azospirilla
The three species in the genus Azospirillum all have a
mainly respiratory type of metabolism. They fix dinitrogen
in microaerophilic environments where combined nitrogen
concentration is low, and utilizable carbon-and-energy
sources are available. When provided with metabolizable
carbon, along with ammonium, nitrate, or other combined
nitrogen sources, they can grow under aerobic conditions.
In either situation, they grow well on the salts of organic
acids such as malate, succinate, lactate, or pyruvate (Krieg
and Dobereiner, 1984). Azospirillum brasilense can use some
carbohydrates, including fructose, galactose, and
arabinose. Azospirillum lipoferum is also able to use
these sugars, as well as glucose, mannose, and sorbose
(Martinez-Drets et al., 1984). The most recently recognized
species, Azospirillum amazonense, differs from the other two
species in that it can grow on sucrose and other disacchar
ides (Martinez-Drets et al., 1985). Both A. brasilense and
A. amazonense can synthesize their own biotin, whereas
A. lipoferum can only grow if exogenous biotin is available
(Falk et al., 1985).
Microaerophilic culture conditions for dinitrogen
fixation can be established by culturing azospirilla in
media containing 0.05% (wt/vol) agar. The bacteria grow and
form a pellicle slightly below the agar surface, where dif
fusion of 02 from the culture-vessel headspace balances the


86
filaments soon after 96 hours. In nitrogen-free cultures,
such elongated filaments persisted, some being weakly motile
even after 79 days on nitrogen-free LP-BHB agar.
After 96 hours, lawns of A. brasilense strain Cd on all
BHB agar media contained mixtures of vibrioids, ovoids,
filaments, and chains. Sometimes the cell material from
LP-BHB agar lawns with or without combined nitrogen did not
resuspend uniformly in water, but as macroflocs, due to
extensive encapsulation. A microfloc from an 11-day-old,
nitrogen-free, LP-BHB agar plate is shown in Figure 2-5b,c.
By adjusting the objective lens, the capsule is made
evident. The entire microfloc may have arisen from one
elongated filament that underwent septation, as suggested by
the apparent linear continuities between cytoplasmic
contents.
After 79 days of culture, lawns of this strain grown
with combined nitrogen contained mainly spheroplasts and
cell ghosts, appearing to have entered stationary phase.
Cultures grown on nitrogen-free agar at both phosphate
levels contained numerous pleomorphic forms at this time, in
addition to cells of normal morphology (Figure 2-6).
A microfloc of A. brasilense strain Cd from a 6-day-old
lawn on SNF-Congo Red agar is shown in Figure 2-7. All the
cells are encapsulated, and empty capsules are evident. The
lawn was scarlet in color, unlike the pale orange lawns of
the same age grown on nitrogen-free BHB-Congo Red agar.


49
motility. These long cells eventually fragmented into
shorter, ovoid cells. Many of these fragments later became
large, pleomorphic cells filled with light-refractile
granules, probably PHB. In contrast, A. brasilense strains
transferred to nitrogen-free, semisolid malate medium ini
tially retained their normal appearance. Only after several
weeks' time in this medium did they develop some S-shaped
cells and some large, pleomorphic, granule-filled forms.
Falk et al. (1985) found that A. amazonense strains failed
to become pleomorphic under comparable conditions.
Krieg and Dobereiner (1984) maintained that alkalini-
zation of the medium due to oxidation of malate was
responsible for pleomorphism in A. lipoferum. Cultures of
this species grown in semisolid, nitrogen-free glucose
medium did not become alkaline, and the cells did not become
pleomorphic.
Wong et al. (1980) isolated a putative Azospirillum
sp. from cellulolytic, dinitrogen-fixing mixed cultures. In
semisolid, nitrogen-poor malate medium containing adequate
levels of biotin, the cells were of normal size and
morphology after 1 day's growth. Between the third and
seventh days, the cells gradually became S-shaped and
enlarged. These enlarged cells contained granules of PHB
and/or polyphosphate. By 10 days, many cells had lysed and
released these granules into the medium. When the initial
biotin concentration of the medium was reduced to 10% of the


ACKNOWLEDGMENTS
I thank my major professor, Dr. Murray Gaskins, for
first bringing the topic of bacterial cysts to my attention
and for allowing me to pursue the subject. I also thank the
other members of my guidance committee, Dr. Stephan
Albrecht, Dr. David Hubbell, Dr. David Mitchell, and
Dr. Stephen Zam, for their patience, encouragement, and use
of laboratory facilities. Thanks also go to Dr. Sylvia
Coleman, whose transmission electron microscopy studies
aided this work. Special thanks go to Dr. Kenneth
Quesenberry and the state of Florida for my graduate
assistantship.
Kelly Kirkendall Merritt taught me how to spread plate
and introduced me to manipulations of azospirilla. She is a
good coworker to be around, as are Stephanie Syslo, Mary
Myers, and Dr. Garnet Jex, who helped me with discussions,
calculations, and shared experiences in and out of the
laboratory. Dr. Lakshmi Sadasivan discussed her work with
me before it was published and furnished preprints of her
work. Talks with Dr. Harold Sadoff helped me a great deal
in understanding Azotobacter cysts. And Dr. Noel Krieg
provided several cultures that were keys in the study.
Thanks go to all of them.
iii


81
The best growth of A. brasilense strain JM 125A2 on
agar occurred on SNF-Congo Red agar. Cells from 6-day-old
lawns grown on this agar medium were often seen as encap
sulated microflocs (Figure 2-4). The microfloc in Figure
2-4a and b appears to have arisen mainly from one or more
filamentous cells that underwent septation. This may also
have occurred for many of the cells in Figure 2-4c. The
capsules were of thickness comparable to those observed on
BHB agar, about 0.5 pm. The lawns on SNF-Congo Red agar had
a scarlet or blood-red appearance, unlike lawns of this
strain growing on nitrogen-free BHB-Congo Red agar, which
were pale orange.
The other A. brasilense strain, A. brasilense strain
Cd, also failed to convert in high numbers to pleomorphic
forms, but it grew far better on BHB. After 63 hours of
growth, lawns of this strain on each BHB agar medium con
tained many motile vibrioids possessing large granules of
putative PHB. Elongated, filamentous cells were also
present in high numbers. These cells had about the same
width (1.5 pm) as normal dinitrogen-fixing cells but were
much longer, some being 9 to 13 pm in length (Figure 2-5a).
The filaments were sometimes observed to undulate slowly and
were much slower than motile vibrioids. In the presence of
combined nitrogen, these filaments were seen to septate and
fragment. This fragmentation was observed at 63 to 96
hours, and sometimes was complete within a population of


23
Bacterial cells resembling cysts comprised 27% of cells
observed.
Sudo and Dworkin (1973) reviewed the kinds of pro
karyotic resting cells that were recognized at the time.
Bacterial resting cells were defined as cells in which
division does not occur, and endogenous respiration is
absent or greatly reduced. Usually resting cells are more
resistant to environmental stresses than are vegetative
cells; resting cells are often morphologically different as
well. Such resting cells often differ in chemical composi
tion from vegetative cells (Keynan, 1972). There are often
either qualitative or quantitative differences between the
electron transport systems of vegetative and resting cells.
Many resting cells, for some period after they have
germinated and resumed growth, are self-sufficient in energy
sources, metabolites, and macromolecular precursors.
Perhaps the best understood bacterial resting cell
stages are the endospores of bacilli and Clostridia. Cysts
differ from endospores in that they are formed by the
modification of an entire vegetative cell. The vegetative
cell rounds up during encystment and becomes coated with one
or more layers, often exopolysaccharide, external to its
cell wall. No cyst forms withstand the extremely high
temperatures tolerated by endospores, but they are com
parably resistant to other environmental stresses (Sudo and
Dworkin, 1973).


159
^ r
* #
-M,v2V
A -jJ*
*
*
v: **
!&* **%
vWk^ CLf /*
< % y> + A
*£ *'*# *
* * *


Cells of Azospirillum brasilense and Azospirillum
lipoferum have a similar appearance when cultured in broth
containing combined nitrogen. They are short, plump,
slightly curved rods averaging 1.0 *m in diameter and 2.1 to
3.8 m in length. They are motile in broth by means of a
single, polar flagellum (Tarrand et al., 1978).
Cells of azospirilla often contain granules of the
polymer PHB (Krieg and Dobereiner, 1984). Grown with com
bined nitrogen, Azospirillum brasilense Sp 7 (ATCC 29145)
has 0.5% to 1.0% of its dry weight as PHB. When grown in
dinitrogen-fixing conditions, the PHB content rises until as
much as 25.0% of its dry weight is PHB (Okon et al.,
1976b). Granules of PHB are present in cells grown on
combined nitrogen, but granule size and number are reduced
compared to that found in dinitrogen-fixing cells (Albrecht
and Okon, 1980).
Under certain cultural conditions, cells of azospirilla
produce an outermost layer of capsular polysaccharide. When
grown on an agar medium containing peptone, succinic acid
and ammonium sulfate at 37C for 48 to 72 hours, a small
proportion of cells are Gram-variable, possibly because they
possess capsules. On this medium, A. brasilense exhibits
more Gram-variability than does A. lipoferum. When cells
of either species are cultured in the broth form of this
medium, they stain uniformly Gram-negative, at least in
young cultures (Krieg and Dobereiner, 1984).


175
their PHB reserves unless exogenous, combined nitrogen
becomes available. If these cells are found to have a very
low endogenous respiratory rate, it might further indicate
their state as nascent cysts.
Living cells may form structures that prove immediately
useful for some functions. By chance these structures may
also prove beneficial to the cells in other ways (Cairns-
Smith, 1982). It is suggested that the microflocs of azo
spirilla are such structures. Their possible benefits were
suggested in Chapter II. Four observations in this chapter
deserve further comment. One is the great size difference
between motile, vegetative cells of A. lipoferum Sp RG6xx
and nonmotile, encapsulated cells. The encapsulated cells
occupy much more volume. Secondly, Costerton et al. (1981)
suggested that most bacterial cells in nature assume two
forms. Sessile forms surrounded by a capsule maintain a
population on a surface and give rise to motile swarmer
cells which colonize new surfaces. This is a good descrip
tion of the conversion of encapsulated to motile forms of
A. lipoferum Sp RG6xx. Thirdly, the ability of only 5 to
20% of cells within encapsulated floes to retain their
visible PHB deposits over 65 days of aerobic nitrogen
starvation may indicate the physiological diversity of cells
within an encapsulated PHB-rich microfloc. Most cells may
be poised to become motile or resume vegetative growth and
may represent the cells that depleted their PHB reserves


117
nonencapsulated cells in their suspensions. Whether these
free cells never merged with microflocs, or whether they
were ejected from their capsules upon wetting and suspension
for microscopy, is not clear. The width of the capsules on
these three strains was again observed to average about 0.5
pm, and most floes arose from one or more filaments that
eventually underwent septation.
Enumeration of Encapsulated Cells
It was first suspected that cells of A. lipoferum Sp
RG6xx harvested from nitrogen-free, LP-BHB agar would not be
quantifiable by plate counting. Plate counts demand that
the inoculum be uniformly suspended and diluted, and floccu
lation makes these difficult to accomplish. But the resus
pended cells gave consistent CFU counts for a given OD^q
range (Table 2-1). After resuspension in water and washes
in buffer, the cell suspensions appeared silvery, and floe
size was reduced to the lower limits of visibility to the
naked eye. Macroflocs broke into smaller microfloc domains,
which had formed within the first few days on the agar
medium. These small domains retained their integrity even
after repeated shaking and washing steps. The similar CFU
counts for cells of different ages suggested that the cells
were in a sort of stasis, where they were no longer
multiplying or dying off appreciably.


20
accumulate in the cytoplasm to the same extent as does
ammonium.
Cells with different contents of PHB were harvested by
centrifugation and resuspended in phosphate buffer to meas
ure viability during aerobic nutrient starvation. By 140
hours, bacteria with abundant PHB reserves had given rise to
more than twice as many viable cells as were present in the
initial inoculum (Tal and Okon, 1985). During starvation,
PHB reserves were degraded quickly but not completely. The
initial inoculum contained 40% of its dry weight as PHB.
This fell rapidly to about 24% of cell dry weight after 42
hours of starvation. After 130 hours of starvation, the PHB
content of the cells was about 20% of cell dry weight.
In comparison, cells initially containing only 5% of
their dry weight as PHB had only 7% of the original number
of viable cells after 130 hours of starvation (Tal and Okon,
1985). Poly-S-hydroxybutyrate was still measurable through
out starvation of these PHB-poor cells, stabilizing at or
near 3% of the dry weight of all cells present.
Starved PHB-rich cells had a higher respiration rate
during starvation than the PHB-poor cells (Tal and Okon,
1985). Unlike cells having low amounts of PHB, the PHB-rich
cells exhibited nitrogenase activity in the absence of
exogenous carbon sources. But the PHB-rich cells were as
unable to reduce nitrate anaerobically as were the PHB-poor
cells in the absence of exogenous carbon.


72
JM 125A2, A. brasilense strain Cd, and A. lipoferum Sp
RG6xx. Slime developed at the bottom of stationary phase
LP-TSS broth cultures of all three strains. Phase-contrast-
microscopy examination of A. lipoferum Sp RG6xx slime often
revealed numerous, nonmotile masses of cells similar to
zooglea, surrounded by nonmotile vegetative cells (Figure
2-la, b). These masses were notable for their symmetrical
but varied shapes. They were darker than most of the sur
rounding vegetative cells, perhaps indicating greater
viability than that of the surrounding pale vegetative
cells. These zoogleal masses retained their shape and did
not fragment into individual cells when disassociated from
the larger masses of cells. Similar zoogleal forms were
sometimes observed in the slime of A. brasilinese strain Cd
but not in that of A. brasilense strain JM 125A2. These
zoogleal forms of azospirilla may be referred to as micro
scopic floes, or microflocs, that are kept intact by
exopolysaccharide. Although microflocs were numerous,
individual normal cells were also present in large numbers
under these cultural conditions.
Azospirilla were cultured as cell lawns on agar con
taining precursors of PHB to see if high numbers of
pleomorphic forms would arise. The A. brasilense strains
produced some pleomorphic forms, but cells of normal shape
and size predominated, even on old lawns.


9TI


Figure 2-11.
Cell types of Azospirillum lipoferum Sp
RG6xx, from 17-day-old lawns on nitro
gen-free low phosphate-6-hydroxybuty-
rate-Congo Red agar. a) Microfloc
viewed with bright-field optics. 1,500X
magnification. Bar equals 3.0 pm. b)
Another microfloc, viewed with phase
contrast optics. Note that capsules of
the two bottom-left cells are apparently
undergoing division with their cells.
1,500X magnification. Bar equals 3.0
pm.


132
Encapsulated cells of Azospirillum
lipoferum Sp RG6xx from a 50-day-old
nitrogen-free, high phosphate-B-hydroxy-
butyrate broth culture. Note apparent
continuities between cytoplasms of
several cells that appear to have under
gone plasmolysis. 1,500X magnifica
tion. Bar equals 3.0 pm.
Figure 2-26.


Figure 2-4. Cells of Azospirillum brasilense strain JM
125A2, from 6-day-old lawns grown' on
succinate-nitrogen-free-Congo Red agar,
a) Microfloc showing capsules and
filamentous cell patterns. 1,500X mag
nification. Bar equals 3.0 pm. b) Same
floe as (a), but focused so that capsules
and filamentous cell outlines are no
longer evident. 1,500X magnification.
Bar equals 3.0 pm. c) Different mass of
encapsulated cells. 1,500X magnifica
tion. Bar equals 3.0 pm.


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
:ry H.
Professor
Ga/skins, Chairman
or Agronomy
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Stephan L. Albrecht
Assistant Professor of Agronomy
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Sbbpheh G. zam
Associate Professor of Microbiology
Cell Science
an'


137
aerobically. The A. lipoferum strains remained quite
healthy in appearance in SEM studies under these condi
tions. Their ready uptake of BHB leading to PHB accumula
tion may have allowed them to retain their cellular
integrity under conditions of nitrogen starvation. The
A. brasilense strains in two-step, broth-replacement
cultures did not produce visible exopolysaccharides in SEM
studies. The collapsed and shrunken appearance of the
A. brasilense strains in SEM studies also suggests that they
were not extensively accumulating PHB.
Following completion of these studies, it was learned
that the high level of phosphates used in HP-BHB, two-step,
broth-replacement studies can inhibit growth of azospirilla
in the absence of combined nitrogen (Scott et al., 1979; Das
and Mishra, 1984). However, the A. lipoferum strains used
in these studies still accumulated PHB and capsules in the
two-step, broth-replacement studies where this buffer was
used. These studies also showed that encapsulation and
pleomorphism can occur at near-neutral pH for some A.
lipoferum strains. The accumulation of PHB occurred on both
HP-BHB and LP-BHB agar. Extensive encapsulation only
occurred on nitrogen-free, LP-BHB agar, where the pH may
have become alkaline. Alkaline pH may prevent pleomorphic
cells from resuming vegetative growth.
The sequence of events for A. lipoferum strains grown
in BHB broth or on BHB agar was PHB accumulation, followed


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of
the Requirements for the Degree of
Doctor of Philosophy
MORPHOLOGY AND PHYSIOLOGY OF AZOSPIRILLA
GROWN ON BETA-HYDROXYBUTYRATE
By
Bruce Henry Bleakley
May 1986
Chairman: Murray H. Gaskins
Major Department: Agronomy
Strains of Azospirillum brasilense and Azospirillum
lipoferum were cultured with beta-hydroxybutyrate (BHB) to
determine if they could be converted in high numbers to
cyst-like forms, as can some strains of Azotobacter spp.
Azospirillum brasilense strain JM 125A2 grew poorly on BHB
but produced some nonmotile cells of cyst-like morphology.
Azospirillum brasilense strain Cd grew better on BHB and
often produced elongated cells as well as some nonmotile,
cyst-like cells. Capsules and accumulation of poly-beta-
hydroxybutyrate (PHB) were common features of all putative
cysts. Encapsulation occurred with all A. lipoferum strains
tested. Cells accumulated PHB and assumed elongated,
filamentous shapes as they lost motility. Later, capsules
were produced and microflocs formed. The filamentous cells
eventually formed septa. Several cell shapes were present
vi


Figure 2-20. Microflocs of Azospirillum lipoferum Sp
A3a from nitrogen-free, low phosphate-8-
hydroxybutyrate agar. a) Microfloc from
69-day-old lawn. 1,500X magnification.
Bar equals 3.0 pm. b) Microfloc from
136-day-old lawn. Note empty capsules.
1,000X magnification. Bar equals
4.0 pm.


Figure 2-8.
Cells of Azospirillum lipoferum Sp RG6xx
from 18-hour-old lawn on nitrogen-free,
high phosphate-8-hydroxybutyrate agar.
Note individual cells and filaments
at various stages of septum formation.
1,000X magnification. Bar equals 4.0 pm


135
the cells were not actively dividing after perhaps a week's
time.
Patriquin et al. (1983) observed spherical, bag-like
structures on the surfaces and interiors of 3-week-old and
older wheat roots in axenic association with azospirilla.
Azospirilla containing PHB granules moved actively within
these structures. These structures were similar to the
zoogleal-type microflocs observed in LP-TSS broth cultures
that had passed into stationary phase. It is probable that
these zoogleal forms arose mainly through filamentation,
followed by septation. Also reported in the previous study
were sharply defined, small colonies of azospirilla of
apparently determinate size on the surfaces of wheat roots.
This is a good description of the microflocs of A. lipoferum
from nitrogen-free, LP-BHB lawns. The formation of
filaments (Tarrand et al., 1978) and chains (Ruscoe
et al., 1978) of azospirilla have also been previously
reported.
Most laboratory studies of bacteria use cultures in a
state of balanced growth, where every component of the cell
culture increases at the same rate. This is done for
reproducibility of results and standarization of condi
tions. Cultures that have passed into stationary phase have
experienced a metabolic shift-down, growing more slowly and
with more widely variable characteristics than log phase
cells (Ingraham et al., 1983). It would appear that


MORPHOLOGY AND PHYSIOLOGY OF AZOSPIRILLA
GROWN ON BETA-HYDROXYBUTYRATE
By
BRUCE HENRY BLEAKLEY
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA

This dissertation is dedicated to Isabel, Stewart, and
Robert Bleakley, the only three people who can say "Of
course" without my doubting them.

ACKNOWLEDGMENTS
I thank my major professor, Dr. Murray Gaskins, for
first bringing the topic of bacterial cysts to my attention
and for allowing me to pursue the subject. I also thank the
other members of my guidance committee, Dr. Stephan
Albrecht, Dr. David Hubbell, Dr. David Mitchell, and
Dr. Stephen Zam, for their patience, encouragement, and use
of laboratory facilities. Thanks also go to Dr. Sylvia
Coleman, whose transmission electron microscopy studies
aided this work. Special thanks go to Dr. Kenneth
Quesenberry and the state of Florida for my graduate
assistantship.
Kelly Kirkendall Merritt taught me how to spread plate
and introduced me to manipulations of azospirilla. She is a
good coworker to be around, as are Stephanie Syslo, Mary
Myers, and Dr. Garnet Jex, who helped me with discussions,
calculations, and shared experiences in and out of the
laboratory. Dr. Lakshmi Sadasivan discussed her work with
me before it was published and furnished preprints of her
work. Talks with Dr. Harold Sadoff helped me a great deal
in understanding Azotobacter cysts. And Dr. Noel Krieg
provided several cultures that were keys in the study.
Thanks go to all of them.
iii

Without the Hume Library this work could not have been
done. I thank Mr. William Weaver for running a fine
facility.
Finally, completing the list of good coworkers, the eye
and expertise of Louise L. Munro are felt throughout this
study. Her scanning electron microscopy studies cleared up
several uncertainties and raised new questions. She knew
the right stuff when she saw it. Thanks, Louise.
IV

TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS iii
ABSTRACT vi
CHAPTERS
I INTRODUCTION AND EXPERIMENTAL APPROACH 1
Ecology of Azospirilla 1
Physiology of Azospirilla 4
Morphology of Azospirilla 6
Prokaryotic Exopolysaccharides 8
Poly-8-Hydroxybutyrate ( PHB ) 12
Dormant Forms of Prokaryotic Cells 22
Resistance of Bacteria to Drying 27
Azotobacter Cysts 37
Pleomorphism of Azospirilla 45
Experimental Approach 5 6
II PLEOMORPHISM OF AZOSPIRILLA GROWN ON
BETA-HYDROXYBUTYRATE 5 9
Materials and Methods 63
Results 71
Discussion 131
III PHYSIOLOGICAL PROPERTIES OF ENCAPSULATED FLOCS
OF AZOSPIRILLUM LIPOFERUM Sp RG6xx 144
Materials and Methods 145
Results 153
Discussion 167
IVGENERAL CONCLUSIONS 178
BIBLIOGRAPHY 181
BIOGRAPHICAL SKETCH 193
V

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of
the Requirements for the Degree of
Doctor of Philosophy
MORPHOLOGY AND PHYSIOLOGY OF AZOSPIRILLA
GROWN ON BETA-HYDROXYBUTYRATE
By
Bruce Henry Bleakley
May 1986
Chairman: Murray H. Gaskins
Major Department: Agronomy
Strains of Azospirillum brasilense and Azospirillum
lipoferum were cultured with beta-hydroxybutyrate (BHB) to
determine if they could be converted in high numbers to
cyst-like forms, as can some strains of Azotobacter spp.
Azospirillum brasilense strain JM 125A2 grew poorly on BHB
but produced some nonmotile cells of cyst-like morphology.
Azospirillum brasilense strain Cd grew better on BHB and
often produced elongated cells as well as some nonmotile,
cyst-like cells. Capsules and accumulation of poly-beta-
hydroxybutyrate (PHB) were common features of all putative
cysts. Encapsulation occurred with all A. lipoferum strains
tested. Cells accumulated PHB and assumed elongated,
filamentous shapes as they lost motility. Later, capsules
were produced and microflocs formed. The filamentous cells
eventually formed septa. Several cell shapes were present
vi

in floes, but all cells possessed intracellular PHB and
capsules. Some cells within floes appeared cyst-like.
Broth studies indicated that alkaline pH does not cause
these morphological changes.
Cells of Azospirillum lipoferum Sp RG6xx grown on
nitrogen-free BHB agar accumulated up to 57% of their dry
weight as PHB, compared to 3.6% when grown with combined
nitrogen. Neither vegetative nor encapsulated cells of this
strain survived in significant numbers after 8 days of
desiccation. Vegetative cells of this strain multiplied
several fold and retained viability during 9 days of starva
tion for carbon and nitrogen, whereas encapsulated cells
were reduced to 25% of their original numbers. Nonmotile,
encapsulated cells produced motile vegetative cells when
incubated with nitrate, ammonium, or soil extract but did
not do so appreciably in nitrogen-free, buffered-salts
solution with or without carbon sources. Treatment with
Tris-EDTA did not result in expulsion of cells from their
capsular coats, as it does for mature Azotobacter spp.
cysts. Studies with chloramphenicol indicated that
encapsulated cells do not possess the enzymes needed for
growth and emergence from their capsules.
The studies suggested that PHB accumulation and capsule
formation during unbalanced growth precede the formation of
dormant cyst-like cells.
Vll

CHAPTER I
INTRODUCTION AND EXPERIMENTAL APPROACH
Ecology of Azospirilla
Bacteria of the genus Azospirillum have been isolated
from soils and from the roots of cereal crops and forage
grasses in several areas of the world (Dobereiner et al.,
1976; Tyler et al., 1979; Lamm and Neyra, 1981). Their
4
numbers in nonrhizosphere soil can be as high as 10 cells/g
soil (Dobereiner, 1978), while their numbers in rhizosphere
soil can be as high as 10^ cells/g soil (Krieg and
Dobereiner, 1984).
Agricultural interest in Azospirillum spp. resulted
from recognition of their ability to reduce atmospheric
dinitrogen. The enzyme catalyzing this reaction, nitro-
genase, is inactivated in the presence of combined nitrogen
or oxygen (Okon et al., 1976a). Azospirilla fix dinitrogen
under microaerophilic conditions in nitrogen-free media in
the laboratory (Day and Dobereiner, 1976; Okon et al.,
1976a). Nonrhizosphere soil is usually too poor in avail
able, utilizable carbon sources to enable Azospirillum
spp. to fix dinitrogen, but they can do so in the more
carbon-rich rhizosphere environment (Dobereiner et al.,
1976). Maximum nitrogenase activity with inoculated plants
1

2
growing in soil is often found at the reproductive stage of
the plant (reviewed by Patriquin et al., 1983), after plant
uptake and other processes have reduced the amount of com
bined nitrogen in the root zone (Okon and Hardy, 1983).
Sometimes low amounts of fixed nitrogen have been incor
porated into plant tissue. The transfer of fixed nitrogen
from bacterium to plant seems slow, probably because
bacterial nitrogen is made available for plant uptake only
after the mineralization of the organic nitrogen of dead
bacteria (Okon et al., 1983).
Although the nitrogenase activity of Azospirillum
spp. may not directly provide quick or agriculturally sig
nificant benefits to inoculated plants, the bacteria have
been found to possess other characteristics that may benefit
plants. Axenic associations of grass seedlings and
Azospirillum spp. have resulted in rapid proliferation of
lateral roots and root hairs, probably due to bacterial
production of indole-3-acetic acid and other plant growth
substances (Tien et al., 1979; Umali-Garcia et al., 1980 ;
Jain and Patriquin, 1985). It is also possible that
Azospirillum spp. can enhance production of plant growth
substances by the plants themselves (Kapulnik et al.,
1985). In any case, associations of Azospirillum spp. with
plant roots have led to significant increases of commer
cially valuable plant components in both axenic laboratory
experiments (Kapulnik et al.,
1985) and field inoculations

3
(Okon and Hardy, 1983). Short-term axenic associations have
also resulted in enhanced uptake of mineral ions by grass
roots (Lin et al., 1983). This effect may be due to the
influence of plant growth substances, or to softening of the
middle lamellae of root cells by pectolytic bacterial
enzymes, which some azospirilla are known to produce
(Umali-Garcia et al., 1980; Tien et al., 1981). Such
effects on root morphology and activity may make inoculation
with azospirilla beneficial in some agricultural situations.
The rhizosphere environment is prone to extreme chemi
cal and physical fluctuations (Foster and Bowen, 1982).
This may lead to periods when azospirilla are inactive due
to environmental limitations. Cells of azospirilla can vary
morphologically (Krieg and Dobereiner, 1984). Some of these
cell forms may be dormant or resting stages, in which
activities of possible benefit to plants are not expressed.
Pleomorphic forms of azospirilla usually possess capsules,
and contain large amounts of the reserve polymer poly-B-
hydroxybutyrate (PHB). This study describes attempts to
obtain such forms in high numbers by laboratory culture.
The general topics of capsules, PHB, physiological dormancy,
and desiccation resistance are directly related to this
study, and will be briefly reviewed in this introduction
after discussion of some key aspects of Azospirillum spp.
physiology.

4
Physiology of Azospirilla
The three species in the genus Azospirillum all have a
mainly respiratory type of metabolism. They fix dinitrogen
in microaerophilic environments where combined nitrogen
concentration is low, and utilizable carbon-and-energy
sources are available. When provided with metabolizable
carbon, along with ammonium, nitrate, or other combined
nitrogen sources, they can grow under aerobic conditions.
In either situation, they grow well on the salts of organic
acids such as malate, succinate, lactate, or pyruvate (Krieg
and Dobereiner, 1984). Azospirillum brasilense can use some
carbohydrates, including fructose, galactose, and
arabinose. Azospirillum lipoferum is also able to use
these sugars, as well as glucose, mannose, and sorbose
(Martinez-Drets et al., 1984). The most recently recognized
species, Azospirillum amazonense, differs from the other two
species in that it can grow on sucrose and other disacchar
ides (Martinez-Drets et al., 1985). Both A. brasilense and
A. amazonense can synthesize their own biotin, whereas
A. lipoferum can only grow if exogenous biotin is available
(Falk et al., 1985).
Microaerophilic culture conditions for dinitrogen
fixation can be established by culturing azospirilla in
media containing 0.05% (wt/vol) agar. The bacteria grow and
form a pellicle slightly below the agar surface, where dif
fusion of 02 from the culture-vessel headspace balances the

5
uptake of 02 by the bacteria, allowing both cell respiration
and protection of the oxygen-sensitive nitrogenase (Okon et
al., 1976a). Broth cultures can fix dinitrogen if the
dissolved oxygen level is well controlled (Okon et al.,
1976a). The bacteria are also able to grow on the surface
of nitrogen-free, aerobically incubated agar plates (Day and
Dobereiner, 1976).
The flagella of azospirilla enable them to move to
whatever sites their physiological state demands. They have
been shown to exhibit aerotaxis to microaerophilic sites
(Barak et al., 1982). Alternatively, cells may aggregate,
thereby creating a microaerophilic environment by the
respiration of many cells in a small space (Barak et al.,
1982). The grass rhizosphere may contain microaerophilic
sites (reviewed by Patriquin et al., 1983). Azospirilla
could migrate from soil toward such sites, where nitrogenase
activity could subsequently be expressed.
The respiratory metabolism of Azospirillum spp.
includes the ability of many strains to denitrify, reducing
nitrate or nitrite to more reduced nitrogenous compounds
under anaerobic conditions if enough metabolizable carbon
source is available (Neyra and Dobereiner, 1977; Neyra and
van Berkum, 1977; Nelson and Knowles, 1978). Under certain
laboratory conditions, denitrification has been shown to
provide enough ATP to support anaerobic growth of
azospirilla (Bothe et al., 1981; Zimmer et al., 1984). The

6
ATP derived from denitrification can be used to drive nitro-
genase activity (Scott et al., 1979), but it seems unlikely
that dinitrogen fixation under these conditions can support
growth of the bacteria (Bothe et al., 1981). Recent work by
Neuer et al. (1985) has shown that, in axenic wheat-
Azospirillum spp. associations, both dinitrogen fixation and
denitrification can occur.
Morphology of Azospirilla
Azospirilla are Gram-negative bacteria (Tarrand et al.,
1978). The structural layers external to the cytoplasmic
membrane of Gram-negative prokaryotes have been reviewed
(Costerton et al., 1974). Depending on cultural conditions
and the bacterial strain, polysaccharide or capsular layers
may be present as the outermost layers of the cell.
A growing Gram-negative cell divides by binary fission
to produce two daughter cells of approximately equal size.
Division begins with invagination of the cytoplasmic mem
brane and peptidoglycan, until a complete transverse septum
or cross wall is formed. When the septum is completely
formed and cleaved, the two daughter cells separate (Leive
and Davis, 1980). As will be discussed later, this cell
division process can be disrupted, resulting in formation of
filaments or chains, which accounts partially for pleo-
morphism of azospirilla.

Cells of Azospirillum brasilense and Azospirillum
lipoferum have a similar appearance when cultured in broth
containing combined nitrogen. They are short, plump,
slightly curved rods averaging 1.0 *m in diameter and 2.1 to
3.8 m in length. They are motile in broth by means of a
single, polar flagellum (Tarrand et al., 1978).
Cells of azospirilla often contain granules of the
polymer PHB (Krieg and Dobereiner, 1984). Grown with com
bined nitrogen, Azospirillum brasilense Sp 7 (ATCC 29145)
has 0.5% to 1.0% of its dry weight as PHB. When grown in
dinitrogen-fixing conditions, the PHB content rises until as
much as 25.0% of its dry weight is PHB (Okon et al.,
1976b). Granules of PHB are present in cells grown on
combined nitrogen, but granule size and number are reduced
compared to that found in dinitrogen-fixing cells (Albrecht
and Okon, 1980).
Under certain cultural conditions, cells of azospirilla
produce an outermost layer of capsular polysaccharide. When
grown on an agar medium containing peptone, succinic acid
and ammonium sulfate at 37C for 48 to 72 hours, a small
proportion of cells are Gram-variable, possibly because they
possess capsules. On this medium, A. brasilense exhibits
more Gram-variability than does A. lipoferum. When cells
of either species are cultured in the broth form of this
medium, they stain uniformly Gram-negative, at least in
young cultures (Krieg and Dobereiner, 1984).

8
Prokaryotic Exopolysaccharides
Many genera of both Gram-positive and Gram-negative
bacteria include species that can produce polysaccharide
layers outside their cell wall. Such layers have been
referred to as capsules, exopolysaccharides (Sutherland,
1977) or glycocalyces (Costerton et al., 1981). Depending
on laboratory cultural conditions, these polymers can assume
different forms. Slime layers adhere loosely, if at all, to
the cell and can often be separated from cells by centri
fugation. Capsular layers appear to be tightly bound to the
cell itself, and cannot be easily separated from cells.
Microcapsules are so thin that their presence outside the
cell wall cannot be observed using staining and light micro
scopy, while macrocapsules are of sufficient width to be so
resolved (Ward and Berkeley, 1980).
Although proteins are sometimes present in bacterial
capsules, most capsules are mainly polysaccharide in com
position. The polysaccharides are extensively hydrated, and
up to 99% by weight of the capsule is accounted for as water
(Costerton et al., 1981).
The ATP needed to activate sugar residues for exo
polysaccharide synthesis has been shown to comprise a
significant proportion of total-cellular-ATP demand for some
bacteria. Even when the carbon supply is growth limiting,
some strains of bacteria produce extracellular polysacchar
ides (Jarman and Pace, 1984).

9
For many bacteria, a culture medium having a high
carbon to nitrogen (C/N) ratio promotes capsule formation
(Sutherland, 1977; Costerton et al., 1981). Some species
manufacture exopolysaccharide throughout all phases of
growth, while others produce it only at certain stages of
growth (Sutherland, 1977). Exopolysaccharides of more than
one composition can be formed by the same bacterium under
different environmental conditions (Geesey, 1982).
In laboratory culture, exopolysaccharides may be
nonessential for bacterial growth. Enzymatic removal of
capsules often causes no reduction in viability of the
decapsulated cells (Dudman, 1977). Nonencapsulated mutants
may grow better in laboratory culture than do encapsulated
cells, since they expend no energy for capsular synthesis
(Costerton et al., 1981). Many nonencapsulated laboratory
strains are mutants that have lost the ability of the wild
type to produce exopolysaccharide. In other instances,
common laboratory media have too low a C/N ratio to promote
exopolysaccharide synthesis.
Attachment of bacteria to surfaces by their exopoly
saccharides is the rule in nature, whether the surface is an
inert mineral particle or a biological surface such as a
plant root (Costerton et al., 1981). Natural environments
are far different from laboratory cultural conditions,
containing many more potential hazards to bacterial
survival. In natural environments, the presence of

10
exopolysaccharides may aid the survival of bacteria (Dudman,
1977). Exopolysaccharides can concentrate nutrients from
the surrounding solution phase. They give some bacteria
increased resistance to antibiotics, surfactants, and other
chemicals, as well as deterring their engulfment by
phagocytic cells (Costerton et al., 1981). Other advantages
of exopolysaccharides have been suggested, such as mediation
of gas exchange between bacteria and their surroundings, but
they have proven difficult to prove experimentally. Extra
cellular enzymes might also be located within or at the
surface of capsules (Geesey, 1982).
Nur et al. (1980) found that A. brasilense Sp 7 and an
Israeli isolate of A. brasilense both possessed small cap
sules discernible by electron microscopy when grown on
nutrient agar. Umali-Garcia et al. (1980) found that when
certain A. brasilense strains and grass seedlings were
incubated together for 10 to 30 min at 30C, many bacteria
adhered to the grass roots, with granular material
accumulating on the surfaces of root hairs, and fibrillar
material accumulating on the surfaces of older, epidermal
root cells. It is known that bacterial exopolysaccharides
may appear either granular or fibrous (Foster and Bowen,
1982). The A. brasilense strains seemed to rapidly produce
both types of exopolysaccharides in axenic association with
grass roots. After 2 to 4 days of axenic association with
grass roots, from two to four cells of A. brasilense Sp 7

11
were sometimes seen to be enclosed within a common envelope
or capsule. Such structures were not observed when the
bacteria were grown in trypticase soy broth (Umali-Garcia et
al., 1980). This is another indication that the low C/N
ratio of complex broth media can repress extensive capsule
formation by azospirilla, while the high C/N ratio near
plant roots can promote capsule formation.
Recent work by Sadasivan and Neyra (1985) verified that
the forms of carbon and nitrogen made available to azo
spirilla can have a profound effect on exopolysaccharide
synthesis. When A. brasilense Sp 7 and A. lipoferum Sp. 59b
(ATCC 29707 ) were cultured in broth containing 8.0 mM
fructose and 0.5 mM KNO^ they grew as individual motile
cells for only 6 hours and then started to clump, as exo
polysaccharide production led to floe formation. Organic
acids yielded fewer floes than did sugars, and other nitro
gen sources, such as ammonium, yielded fewer floes than did
nitrate. The cells in floes appeared initially to be
enmeshed in a loose, fibrillar matrix that condensed
progressively over a week's time. When cells were grown,
harvested by centrifugation, and resuspended in broth lack
ing carbon, the cells remained freely suspended. This
suggests that azospirilla have a high ATP demand for exo
polysaccharide synthesis. Chemical analysis showed that
cellulose was a major component of the exopolysaccharide.

12
Fresh floes were not degraded by cellulase, indicating that
more than one type of exopolysaccharide was present.
Poly-B-Hydroxybutyrate (PHB)
In a constant and favorable environment where all
nutrients are present in sufficient amounts, bacteria grow
for a time in a steady state, where every component of the
cell culture increases by the same constant factor per unit
time. This is balanced growth, and occurs during the
logarithmic phase of the growth curve (Ingraham et al.,
1983). If one or more nutrients become limiting, balanced
growth is not maintained. When the carbon or energy supply
is in excess, so that one or more other nutrients limit
growth, some microorganisms respond by synthesizing and
accumulating intracellular polymers having an energy-storage
function (Dawes and Senior, 1973).
The cell catabolizes these polymers when the energy
supply from exogenous sources is no longer sufficient to
maintain processes needed for maintenance of cell viabil
ity. These processes may include osmotic regulation, main
tenance of intracellular pH and transmembrane potentials,
and turnover of cellular constituents such as proteins and
nucleic acids. The energy required for these processes is
called the energy of maintenance. Some microorganisms do
not produce special energy-storage polymers. Faced with a
starvation environment, they are forced to utilize their own

13
basal components, such as proteins and RNA, for energy
sources. Possession of energy-storage polymers can benefit
some species facing starvation, in that they degrade these
polymers instead of or before they are forced to degrade
such essential components as proteins (Dawes and Senior,
1973). However, different microorganisms utilize common
constituents at different rates and in different sequences
when starved. The possession of energy-reserve polymers
does not spare degradation of protein and other basal com
ponents in some species during starvation. Most micro
organisms that remain viable after prolonged starvation have
a low endogenous metabolism, matched closely to their
maintenance energy requirements. Starved microorganisms
that rapidly metabolize polymers generally lose viability
quickly (Dawes, 1976).
Three main types of microbial energy-storage compounds
are known. Some species can accumulate more than one. All
of these compounds have high molecular weights, and only a
slight effect on the internal osmotic pressure of the cell.
The amount of each compound a cell accumulates can vary
widely, depending on environmental conditions.
Intracellular polyphosphates and glycogen-like poly
saccharides are two types of energy-storage compounds formed
by some eukaryotic and prokaryotic microorganisms. The
synthesis of both types requires ATP.

14
The third microbial energy-reserve polymer is poly-8-
hydroxybutyrate, a straight chain homopolymer of D(-)-3-
hydroxybutyrate. It is found only in prokaryotic cells,
including both Gram-positive and Gram-negative species. Its
synthesis requires reducing power in the forms of NADH or
NADPH, but does not require the direct expenditure of ATP
(Dawes and Senior, 1973).
With phase contrast microscopy, large accumulations of
PHB within bacterial cells appear as light-refractile gran
ules. A single granule may contain several thousand PHB
molecules (Dawes and Senior, 1973). Each granule is bounded
by a nonunit-membrane layer, which is probably formed from
the cytoplasmic membrane. Presumably the enzymes for
polymerization and depolymerization of PHB are present in
this membrane layer (Shively, 1974).
Many of the Azotobacteraceae accumulate PHB when grown
under dinitrogen-fixing conditions. There can be a wide
variation in PHB content between species and between strains
of the same species (Stockdale et al., 1968). The regula
tion of PHB levels in Azotobacter beijerinckii has been
extensively studied and may provide clues to the role of PHB
in-the physiology of other free-living, dinitrogen-fixing
bacteria, such as azospirilla.
The route of PHB biosynthesis in A. beijerinckii has
been reviewed by Dawes (1981). The synthesis and degrada
tion of PHB in this microorganism are intimately associated

15
with intermediates and enzymes of the tricarboxylic acid
(TCA) cycle, a system that azospirilla also possess (Okon et
al., 1976b). When A. beijerinckii strain N.C.I.B. 9067 was
cultured as a dinitrogen-fixer with 2.0% (wt/vol) glucose,
PHB was deposited towards the end of exponential growth.
The cells were unable to use all the available glucose, and
PHB synthesis continued during the stationary phase until up
to 74% of cell dry weight was PHB. Cultures grown with
combined nitrogen rarely contained more than 3.0% of their
dry weight as PHB (Dawes, 1981).
The initiation of PHB synthesis in the A. beijerinckii
strain in batch culture coincided with the attainment of
zero-oxygen concentration. Oxygen limitation was thus
suspected to be a critical factor in initiating PHB synthe
sis. However, the nature of batch broth culture made it
hard to separate oxygen effects from possible nitrogen-
limitation effects (Senior and Dawes, 1971). Later
experiments, using chemostat cultures having carbon, oxygen,
or nitrogen limitation, clearly showed that extensive PHB
accumulation only occurred under conditions of oxygen
limitation (Dawes, 1981).
Before the studies reviewed by Dawes (1981), PHB was
regarded as being only an endogenous, carbon-and-energy
source that benefited cells during starvation. These
experiments suggested that PHB could also serve other
purposes. The synthesis of PHB seemed to serve as an

16
electron sink for excess reducing power (NADH and NADPH)
that accumulated when the cell became oxygen limited, and
electron transport to oxygen via the terminal oxidases of
the electron-transport chain was restricted (Senior and
Dawes, 1971). Later work revealed that the activities of
certain enzymes of carbon catabolism in A. beijerinckii are
inhibited by either or both NADH and NADPH. Under oxygen
limitation, the concentration of these reduced coenzymes is
increased, so that glucose metabolism, operation of the TCA
cycle, and net biosynthesis are decreased. Growth can
continue at some level, however, if PHB is synthesized and
the crucial coenzymes are reoxidized (Dawes, 1981).
The synthesis of PHB under oxygen limitation may occur
in other bacteria as well (Okon and Hardy, 1983). The
quantity of PHB accumulated often greatly increases as the
C/N ratio of the growth medium increases. Under such con
ditions, free-living dinitrogen-fixers may assimilate the
exogenous carbon more rapidly than they can produce reduced
nitrogen. As a result, the cells can accumulate large
amounts of PHB (Stevenson and Socolofsky, 1966; Dawes and
Senior, 1973). The metabolism of PHB is regulated such that
PHB accumulates when the supply of exogenous carbon is in
excess of the requirements for growth and maintenance, and
it is degraded when the supply of exogenous carbon is
limited or exhausted (Dawes, 1981), or when balanced growth
can again occur (Nickels et al., 1979).

17
It has been shown that PHB can accumulate in cells that
are not growing or proceeding toward cell division, due to
limitation of available nutrients (Dawes and Senior, 1973).
Nickels et al. (1979) demonstrated this in laboratory micro
cosms containing oak leaf detritus and estuarine water.
Supplementing the nutrients in the water column with carbo
hydrates, especially glucose, induced a rapid accumulation
of PHB without a concomitant increase in microbial biomass.
When supplements were added that enabled increases in
microbial biomass, PHB levels fell as the polymer was broken
down to aid microbial growth.
In one study, A. brasilense Sp 7 was grown in batch
cultures for up to 14 days in microaerophilic, nitrogen-free
malate broth (Papen and Werner, 1980). Both nitrogenase
activity and PHB synthesis were biphasic. An initial peak
of PHB content occurred at day 3, 1 day before the first
peak of nitrogenase activity. During the first and maximal
peak of nitrogenase activity, there was a decrease in PHB
content, possibly due to accumulation of fixed nitrogen
allowing use of PHB carbon skeletons for biosynthesis. A
second peak of PHB accumulation occurred after the first
maximum of nitrogenase activity. The results suggested that
A. brasilense Sp 7, like A. beijerinckii, may accumulate PHB
when it assimilates exogenous carbon faster than it can fix
dinitrogen.

18
Zimmer et al. (1984) found that A. brasilense Sp 7
accumulated PHB when using nitrite as terminal electron
acceptor for anaerobic growth. A maximum of 38% of cell dry
weight was found to be PHB when less than 3.0 mM nitrite was
present. No PHB was accumulated when in excess of 8.0 mM
nitrite was made available, indicating the role of PHB as a
sink for excess reducing power when other electron acceptors
are scarce. It was also found that PHB-rich cells contained
less protein than did PHB-poor cells.
Azospirillum lipoferum strain Br 17 (ATCC 29709) was
found by Volpon et al. (1981) to accumulate nearly 24% of
its dry weight as PHB near the mid-logarithmic phase of
growth as a dinitrogen-fixer. Near the end of logarithmic
growth, PHB synthesis seemed to stop, and the content of PHB
declined to 13% of cell dry weight in stationary phase.
The PHB metabolism of A. brasilense strain Cd (ATCC
29729) has received considerable study. When this strain
was grown in continuous chemostat culture with malate and
ammonium chloride, a maximum PHB content of 12% of the
biomass was observed under microaerophilic conditions and at
intermediate growth rates (Nur et al., 1982). These growth
conditions were said to approximate conditions generally
encountered in the rhizosphere. The production of PHB was
markedly decreased at higher levels of oxygen and higher
growth rates. Once again, it was observed that cells

containing high amounts of PHB contained less protein than
PHB-poor cells.
Recent work by Tal and Okon (1985) has further
delineated the roles PHB may play in the physiology of A.
brasilense strain Cd. Grown in aerobic batch culture with
malate and 2.8 mM NH.C1, the cells accumulated 40% of their
4
dry weight as PHB after 24 hours, toward the end of exponen
tial growth. When the level of NH^Cl was raised to 15.0 mM
the cells accumulated only 5% of their dry weight as PHB
after 24 hours. In both cases, the amount of PHB decreased
in stationary phase.
In chemostat continuous culture, a maximum of 30% cell
dry weight accumulated as PHB when the gas atmosphere was
0.082 mM C>2 (Tal and Okon, 1985 ). With increasing aeration
the PHB content fell to very low levels. When grown in
batch culture as dinitrogen-fixers, the cells accumulated
about 75% of their dry weight as PHB. Maximal PHB content
was obtained in these experiments when the C/N ratio was
about 70. Both the C/N ratio of the medium and the oxygen
concentration were found to regulate PHB synthesis.
The forms of carbon and nitrogen made available to the
cells affected the levels of PHB accumulated (Tal and Okon,
1985). Organic acids, especially pyruvate, were found to
elicit PHB formation more than carbohydrates did. Sodium
nitrate was found to promote PHB formation more than
ammonium chloride did, possibly because nitrate does not

20
accumulate in the cytoplasm to the same extent as does
ammonium.
Cells with different contents of PHB were harvested by
centrifugation and resuspended in phosphate buffer to meas
ure viability during aerobic nutrient starvation. By 140
hours, bacteria with abundant PHB reserves had given rise to
more than twice as many viable cells as were present in the
initial inoculum (Tal and Okon, 1985). During starvation,
PHB reserves were degraded quickly but not completely. The
initial inoculum contained 40% of its dry weight as PHB.
This fell rapidly to about 24% of cell dry weight after 42
hours of starvation. After 130 hours of starvation, the PHB
content of the cells was about 20% of cell dry weight.
In comparison, cells initially containing only 5% of
their dry weight as PHB had only 7% of the original number
of viable cells after 130 hours of starvation (Tal and Okon,
1985). Poly-S-hydroxybutyrate was still measurable through
out starvation of these PHB-poor cells, stabilizing at or
near 3% of the dry weight of all cells present.
Starved PHB-rich cells had a higher respiration rate
during starvation than the PHB-poor cells (Tal and Okon,
1985). Unlike cells having low amounts of PHB, the PHB-rich
cells exhibited nitrogenase activity in the absence of
exogenous carbon sources. But the PHB-rich cells were as
unable to reduce nitrate anaerobically as were the PHB-poor
cells in the absence of exogenous carbon.

21
This study (Tal and Okon, 1985) also suggested that
elevated PHB levels at the onset of starvation may spare the
use of protein to drive endogenous metabolism. The PHB-poor
cells used up two-thirds of their initial protein during the
first 80 hours of starvation, whereas the protein content of
starved PHB-rich cells increased slightly over 80 hours. It
was also reported that PHB-rich cells were able to survive a
variety of environmental stresses, including desiccation,
better than PHB-poor cells (Tal and Okon, 1985).
The previous study also found that cells enriched in
PHB displayed a one hundred-fold higher aerotactic response
than PHB-poor cells. This supports the claim made in an
earlier study that PHB reserves could be used for aerotaxis
when no exogenous carbon source was available (Barak et al.,
1982 ) .
The previous discussion has shown that both capsule and
PHB synthesis can be promoted by environments with high
available C/N ratios. The roles of capsules and PHB in
pleomorphism in azospirilla will be discussed later. The
nature of dormancy in prokaryotic cells will be discussed
first, since some pleomorphic forms of azospirilla may be
dormant stages. Capsular layers and PHB are often present
in dormant forms of prokaryotes.

22
Dormant Forms of Prokaryotic Cells
There is general agreement that most soil bacteria
spend much of their existence in soil in a state of low
metabolic activity. The low respiratory rates of bulk
samples of nonamended soil support this (Clark, 1967). Many
soil bacteria may be metabolically dormant due to a lack of
readily available carbon and energy supplies (Gray and
Williams, 1971). Soil bacteria may enter into exogenous
dormancy, where growth is delayed by unfavorable physical or
chemical conditions (Marshall, 1980). Such bacteria
probably have the same morphology as actively growing
vegetative cells (Gray and Williams, 1971). These cells are
probably intimately associated with the clay or organic
matter of soil. The cells adsorb to these surfaces by
physical or chemical interactions, or by the use of
exopolysaccharides (Stotzky, 1980).
However, many bacteria may exist in soil as dormant
forms that are morphologically different from their growing,
or vegetative, stages. These cells would have entered a
phase of constitutive dormancy, involving the formation of
spores or cysts (Marshall, 1980). Bae et al. (1972) used
transmission electron microscopy to study thin sections of
bacteria released from soil by centrifugation and washing.
About 28% of the bacteria observed had normal vegetative
morphology, of which 29% possessed capsular layers.

23
Bacterial cells resembling cysts comprised 27% of cells
observed.
Sudo and Dworkin (1973) reviewed the kinds of pro
karyotic resting cells that were recognized at the time.
Bacterial resting cells were defined as cells in which
division does not occur, and endogenous respiration is
absent or greatly reduced. Usually resting cells are more
resistant to environmental stresses than are vegetative
cells; resting cells are often morphologically different as
well. Such resting cells often differ in chemical composi
tion from vegetative cells (Keynan, 1972). There are often
either qualitative or quantitative differences between the
electron transport systems of vegetative and resting cells.
Many resting cells, for some period after they have
germinated and resumed growth, are self-sufficient in energy
sources, metabolites, and macromolecular precursors.
Perhaps the best understood bacterial resting cell
stages are the endospores of bacilli and Clostridia. Cysts
differ from endospores in that they are formed by the
modification of an entire vegetative cell. The vegetative
cell rounds up during encystment and becomes coated with one
or more layers, often exopolysaccharide, external to its
cell wall. No cyst forms withstand the extremely high
temperatures tolerated by endospores, but they are com
parably resistant to other environmental stresses (Sudo and
Dworkin, 1973).

24
Certain properties are shared by all cyst-like forms of
prokaryotes. They are formed when the growth rate of vege
tative cells declines (a metabolic shift-down), due either
to nutrient depletion or transfer of cells to an environment
where balanced growth can no longer occur. Cells encounter
ing these conditions complete their ongoing synthesis of DNA
and chromosome replication but do not initiate new rounds of
DNA synthesis, since growth has ceased (Sadoff, 1975).
Conditions that will prohibit further growth promote the
formation of dormant cells that can survive stress better
than vegetative cells. These dormant cells often contain
PHB or other energy-reserve polymers, and thickened cell
walls or capsular layers. They have enhanced resistance to
irradiation, sonic vibration, and sometimes elevated
temperatures. Perhaps the most important traits for
survival of dormant cells in natural environments are their
resistance to starvation, low endogenous respiration rates,
and desiccation resistance. Many cells entering constitu
tive dormancy need time to mature before they achieve
maximal resistance to stress. It is important to remember
that resting cells formed in natural environments may differ
qualitatively and/or quantitatively in their resistance
properties from those formed under laboratory conditions
(Sudo and Dworkin, 1973).
Many strains of Gram-negative myxobacteria form dormant
cells, called microcysts, when nutrients become limiting.

25
Nonmotile, encapsulated, mature microcysts are more
resistant than are vegetative cells to environmental
stresses such as ultraviolet irradiation, sonic vibration,
and desiccation (Sudo1 and Dworkin, 1973).
Many Gram-negative, methane-oxidizing bacteria isolated
from soil or mud are encapsulated and accumulate PHB when
nitrogen becomes limiting for growth (Whittenbury et al.,
1970b). Depending on the genus and strain, up to 90% of the
cells present may form resting cells upon entering the sta
tionary phase of growth. Lipid cysts of Methylocystis
parvus accumulate large amounts of PHB and survive starva
tion and desiccation better than vegetative cells, but lack
well-defined, capsular cyst coats. Methylomonas spp. and
Methylococcus spp. form rounded, nonmotile cells that
survive starvation better than do vegetative cells. These
cells are called immature cysts, because they never attain
desiccation resistance. Some strains of Methylobacter
spp. form starvation-resistant and desiccation-resistant
cysts that seem morphologically identical to Azotobacter
spp. cysts (Whittenbury et al., 1970a).
Bdellovibrio sp. strain W is the only bdellovibrio that
is known to encyst. Bdellocysts are larger than their
vegetative counterparts and are not light-refractile. They
tolerate sonic disruption, ultraviolet irradiation, and
carbon starvation better than do vegetative cells. The
endogenous respiration rate of bdellocysts is 80% less than

26
that of vegetative cells. When dried over silica gel
desiccant under slight vacuum in glass tubes, vegetative
cells of strain W die out rapidly and entirely. From 45% to
80% of bdellocysts initially present are able to survive 6
days of this desiccation treatment (Tudor and Conti, 1977).
Bdellocysts possess a thickened outer layer of modified
peptidoglycan, and contain inclusion bodies of an
amylopectin-like polysaccharide of glucose monomers. These
features are not found in vegetative cells (Tudor, 1980).
Some strains of Azotobacter spp., apparently some
azospirilla, and the bacteria described above are the only
prokaryotes reported to form cysts. Why do not more bac
teria possess resting stages that are morphologically dif
ferentiated into cysts? Perhaps growth media and conditions
used in the laboratory discourage cyst formation
(Whittenbury et al., 1970b). It is also possible that the
ability to form cysts is sometimes labile and may be lost
upon subculture. One Methylobacter chroococcum strain was
able to form multiple-bodied cysts upon initial isolation
from the environment. It ceased to do so when subcultured.
Other Methylobacter spp. have retained the ability to form
single- and multiple-bodied cysts over several years of
subculture (Whittenbury et al., 1970b).
A mature, cyst-like cell of a prokaryote may perhaps
best be characterized as follows. Mature cysts differ
morphologically from vegetative cells in having thickened

27
outer layers. They are nonmotile and have low endogenous
respiration rates. They only initiate growth into vegeta
tive cells when sufficient nutrients are available. They
must also possess more resistance to some environmental
stresses than do vegetative cells. Enhanced resistances to
starvation and desiccation are probably traits of all mature
cysts. The cysts of the methane-oxidizing bacteria and of
Azotobacter spp. possess these characteristics. Mature
cysts of azospirilla should also have these properties.
Desiccation resistance is a critical characteristic of
prokaryotic cysts. The next section will consider experi
ments conducted to assess the resistance of bacteria to
drying.
Resistance of Bacteria to Drying
Clark (1967) stated that the majority of soil bacteria
survive in air-dried soils, often for several years. When
such soils are rewetted, bacterial activities including
nitrification, ammonification, nonsymbiotic dinitrogen
fixation, and sulfur oxidation are usually detected. The
implication is that the intimate association of bacteria
with clay or organic matter allows bacterial survival at
hydrated microsites in a macroscopically dry soil. Later
findings, reviewed by Stotzky (1980) and Marshall (1980),
support this. Exopolysaccharides may help bacteria to
achieve such intimate association, although capsules

28
themselves have not been found to afford any desiccation
resistance in laboratory studies with pure cultures in
nonsoil conditions (Dudman, 1977).
Because of the importance of desiccation as a limiting
factor in legume inoculation with Rhizobium spp., several
studies have been done on their resistance to drying. There
are broad strain differences in resistance of rhizobia to
desiccation. Many variables are present in drying
experiments, and the variables may interact with one
another. Rhizobium spp. withstand drying best in heavier-
textured soils, where hygroscopic water can be retained by
colloidal surfaces. Die off is far more rapid in drying
sand. Capsules do not afford increased resistance to drying
in studies with soil or other drying surfaces (Lowendorf,
1980). Often fewer rhizobia survive rapid drying
procedures, such as oven drying, than survive milder
desiccation over several weeks' time with controlled rela
tive humidities (Jansen van Rensburg and Strijdom, 1980).
Robinson et al. (1965) added pure cultures of
Pseudomonas spp. or Arthrobacter spp. to sterile soils. The
inoculated soils were dried by passing filtered air through
them for 2 days, by which time they had reached constant
weight. This forced drying resulted in rapid die off for
both species. Labeda et al. (1976) found that slow evapora
tive drying of inoculated soil resulted in reduced death
rates for both Pseudomonas spp. and Arthrobacter spp.

29
These experiments show clearly that rate of drying can
profoundly affect the survival of vegetative bacteria.
Vegetative cells with the capacity to become desiccation
resistant may need time to alter their membrane or cyto
plasmic composition before desiccation resistance is
achieved. Fast-drying procedures may not allow them to do
so. A differentiated resting cell, such as a cyst, may also
need time, depending on how mature it is, to become desicca
tion resistant.
Relative humidity (RH) also has a great influence on
desiccation resistance of prokaryotes. In desiccation at
any RH below 90%, the free water of the cells is removed
almost instantaneously. The water that remains is the bound
water content of the cell, which may be necessary for
continued function of essential metabolic processes and
viability. Often few vegetative bacteria die when desic
cated above 70% RH, but die rapidly as the RH declines to
45% (Webb, 1965). Many desiccation studies have not defined
the RH at which the cells were dried, making duplication of
results difficult.
Thompson and Skerman (1979) tested the desiccation
resistance of vegetative cells of many strains and genera of
the Azotobacteraceae. One milliliter samples of vegetative
cell cultures were added to sterile porcelain beads,
positioned above silica gel in glass bottles sealed with
Parafilm. These desiccation units were stored at room

30
temperature, and at different times single beads were
aseptically removed and placed in broth media. The bacteria
were probably in stationary phase when added to the assem
blies, but it is unlikely that many cysts were present even
in stationary phase broth culture (Sadoff et al., 1971).
The results were surprising; the majority of strains
retained viability for 1 to 2 years of desiccation. This
was true even for bacteria that have never been shown to
form cysts.
Mature cysts of prokaryotes survive rapid desiccation
on glass surfaces far better than do their vegetative coun
terparts, but rarely does all the encysted inoculum survive
rapid drying. Cysts of methane-oxidizers retained 60% to
90% viability after 1 week (Whittenbury et al., 1970a), and
bdellocysts retained 45% to 80% viability after 6 days
(Tudor and Conti, 1977). This may mean that not all the
encysted cells were fully mature when exposed to drying,
even if they all appeared morphologically identical. Such
quick-drying assays can be valuable in determining whether
morphologically differentiated cells are truly cyst-like.
Differences in the desiccation resistance of Azoto-
bacter spp. vegetative cells and cysts are usually
determined by the method of Socolofsky and Wyss (1962).
They impinged suspensions of either cell form on the
surfaces of membrane filters. The filters were then trans
ferred to dry adsorbent pads in Petri dishes and placed in

31
an incubator at 33C. This method is a slow-dying pro
cedure. At different time intervals, the cells were washed
from the membranes, and viability was determined by plat
ing. Cysts of Azotobacter vinelandii ATCC 12837 lost little
viability over a 12-day-period using this drying treatment,
whereas 99% of the vegetative cells were killed by the end
of the first day (Socolofsky and Wyss, 1962). As a result
of the rapid die off of vegetative cells with this treat
ment, later studies considered cells of this strain to be
cysts if they could withstand 4 days of desiccation on
membrane filters (Stevenson and Socolofsky, 1966; Wyss et
al., 1969).
None of these membrane filter studies specified the RH
at which the membranes were dried, or how many cell layers
were deposited upon the membranes. Webb (1965) pointed out
that if bacteria are dried on filters to test their desic
cation resistance, the cells must be applied in a monolayer
to achieve consistent results. If more than a cell mono-
layer is on the filter, most of the cells in subsurface
layers will not be dried or equilibrated with the water
vapor of the environment.
Most desiccation resistance experiments have given
ill-defined or incomplete conditions of drying. Such
experiments have proven, however, that cysts are more
desiccation resistant than are their vegetative
counterparts.

32
Vela (1974) tested desiccation resistance of
Azotobacter vinelandii ATCC 12837 by allowing slow drying of
the agar on which the cells were grown. Vegetative cells
were grown on agar plates of Burk's nitrogen-free medium,
with glucose as the carbon source. Cysts were obtained by
growing the cells on the same agar, except that 0.3%
(vol/vol) n-butanol was employed as sole carbon source.
Dried agar films were then broken with sterile forceps and
placed on the surface of Burk's agar medium containing
glucose. Vegetative cells borne on these agar films
remained viable for nearly 2 years of desiccation, whereas
cysts borne on such films remained viable for 10 years or
longer.
Desiccation tolerance of azospirilla has received some
attention. Lakshmi et al. (1977) recovered azospirilla from
several air-dried soils stored in the laboratory. Recovery
was obtained from one of four sandy soils stored air-dry for
10 years. All of these soils had less than 0.5% organic
matter. Heavier-textured soils with 1.0% or more organic
matter consistently yielded isolates of azospirilla. Some
of these heavier-textured soils had been stored air-dry for
up to 15 years. It was suggested that organic matter aids
the survival of azospirilla in drying soils, and that
desiccation-resistant cells may be formed by these bacteria.
Jagnow (1982) did some work with an Azospirillum
lipoferum strain isolated from maize roots. In field

33
g
inoculations using 8 x 10 CFU/g soil, azospirilla near or
on roots survived better than those in soil distant from
roots. When added to pots of soil containing grass and
6 7
cereal plants, populations remained at 10 to 10 CFU/g
soil, even after 70 days of drought. He speculated that the
presence of roots, either living or dead, enhances the
drought tolerance of the associated azospirilla. In labora
tory studies using nonautoclaved soil microcosms, air drying
of soil was found to kill greater than 99% of the initial
Azospirillum lipoferum inoculum. In comparison, the
indigenous bacteria were little affected by air drying.
This perhaps indicates that, unless azospirilla added to
soil are able to associate quickly with plant roots, they
will soon die out if drought stress occurs.
The desiccation resistance of pleomorphic encapsulated
forms of azospirilla has been studied. Lamm and Neyra
(1981) studied A. brasilense Sp 7 and A. lipoferum Sp 59b,
in addition to several strains of azospirilla isolated from
roots of various grasses in New Jersey and New York. To
obtain cyst-enriched cultures, cells grown in nutrient broth
were harvested by centrifugation, then washed and resus
pended in sterile 0.85% (wt/vol) NaCl. A 1.0 ml sample of
cells was then spread plated as a lawn onto nutrient agar
plates containing 2.0% (wt/vol) agar. Plates were incubated
at 30C until the agar was dried into a thin film, often
requiring a month. After 15 days of incubation, cyst-like

34
cells predominated. Photographs of cyst-enriched cultures
showed that many vegetative cells were still present. To
obtain cyst-free cultures, cells were grown in nutrient
broth, then washed and resuspended in saline. These cells
were then spotted onto sterile, predried nutrient agar films
so that the added cells would dry completely on the agar
film in 30 min at 30C. Agar films from each treatment were
then cut with sterile scissors and aseptically transferred
to vials containing silica gel. To test viability, the
dried agar films were removed periodically from the vials,
placed on nutrient agar plates, and incubated for 1 week at
30C. Vegetative cells did not survive the initial drying
process. Cyst-enriched populations that survived the
initial desiccation period remained viable for up to 15
months. Interestingly, cyst-enriched cultures of two root
isolates were nonviable at time zero, when they were placed
into the silica-gel vials (Lamm and Neyra, 1981).
Two aspects of this study deserve special comment.
Clearly, the cyst-enriched cultures did not receive the same
drying treatment as did the vegetative cells. The cyst-
enriched agar films were obtained by a slow drying process,
and the vegetative cell agar films underwent rapid drying.
It does not seem valid to compare their desiccation toler
ance under these different conditions. Also, two strains
that contained cyst-like cells of apparently mature
morphology were not desiccation resistant. Perhaps they

35
were not able to attain physiological maturity under the
experimental conditions.
Papen and Werner (1982) assessed the desiccation
resistance of cyst-like forms of A. brasilense Sp 7. Cells
from dinitrogen-fixing broth cultures were diluted in
sterile tap water and then impinged onto the surface of
sterile 0.2 pm Millipore membrane filters under vacuum.
Some of the filters were immediately placed on the surface
of nutrient agar plates and incubated at 28C, whereas
others were placed on sterile adsorbent pads in Petri dishes
and dried at 37C until they were placed on nutrient agar
plates. Desiccation-resistant cells were only present after
the first peak of nitrogenase activity, when nonmotile,
encapsulated spheres containing PHB predominated. Cells
before and during the first peak of nitrogenase activity
were motile vibrioids and did not survive the desiccation
treatment. As a second peak of nitrogenase activity arose,
motile, dinitrogen-fixing vibrioids emerged from the
spherical capsules; these vegetative cells were again not
desiccation resistant. More encapsulated, spherical cells
survived 2 days of desiccation than 6 days, but it was not
an order of magnitude difference. This again may be an
indication that morphologically mature cysts are not
necessarily physiologically mature.
The recent work of Sadasivan and Neyra (1985) employed
another assay for desiccation resistance of cyst-like forms

36
of azospirilla. Azospirillum brasilense Sp 7 and
A. lipoferum Sp 59b were studied. Large floes of cells
enclosed in exopolysaccharide were placed on Whatman No. 1
filter paper and air-dried for 30 min. They were then
placed in a closed vial, without desiccant, and incubated at
30C for up to 6 months. Small pieces of dried floes were
transferred periodically to semisolid nitrogen-free malate
medium and incubated at 34C for 2 to 4 days, and growth,
pellicle formation, and nitrogenase activity were observed.
Cells in dried floes remained viable for up to 6 months of
drying.
No vegetative cell controls were dried and tested for
viability in the above study. Although cells remained
viable in dried floes for up to 6 months, it is not known
how many cells survived in a given amount of floe. It is
not known whether the cells themselves were desiccation
resistant, or only physically protected from desiccation by
exopolysaccharides.
Tal and Okon (1985) claimed that PHB-rich cells of
A. brasilense strain Cd were 10 times more desiccation
resistant than cells having little of the polymer. No
details of the test used for desiccation resistance were
given.
Desiccation resistance studies can be difficult to
interpret. Comparing the desiccation resistance of vegeta
tive cells to that of cysts may be less difficult than

37
comparing that of vegetative cells of different strains.
Rapid drying on a glass surface should enable most mature
cysts to remain viable, but not most vegetative cells. A
glass drying surface should be less hygroscopic than are
membrane filters or agar films. Rapid drying on a glass
surface is a severe treatment, but it should reveal the
presence of physiologically modified, stress-resistant
cells, such as mature cysts.
Azotobacter Cysts
Discussion of the nature of Azotobacter spp. cysts is
important, because this information served as the basis for
the experiments with azospirilla reported in this study.
Like azospirilla, the Azotobacteraceae are
Gram-negative aerobes, often containing PHB granules. Many
are motile by flagella. They all fix dinitrogen, and some,
including Azotobacter spp., do so either at atmospheric
oxygen levels (unlike azospirilla), or as microaerophiles.
Only one genus, Azotobacter, contains species with strains
that are known to form cysts (Tchan, 1984). The isolation
of Azotobacter spp. from the interior of 2,000-year-old clay
bricks (Abd-El-Malek and Ishac, 1966), and their persistence
in soils that had been air dried from 10 years (Vela, 1974)
to 30 years (Clark, 1967), may be due largely to cyst forma
tion .

38
When grown in nitrogen-free broth with glucose as the
carbon source, young cells of Azotobacter spp. appear as
rods with rounded ends, ranging from 1.3 to 2.7 pm in
diameter and 3.0 to 7.0 pm in length. As cultures age,
cells often accumulate PHB. Cell morphology may be altered
to ellipsoids, filamentous cells, or chains of cells (Tchan,
1984 ) .
Azotobacter spp. are commonly isolated from soil and
aquatic habitats of near-neutral pH, and are generally less
acid-tolerant than azospirilla. The most common species
isolated from soil is Azotobacter chroococcum, but its
biochemistry and physiology have received less attention
than that of Azotobacter vinelandii (Tchan, 1984).
Azotobacter vinelandii ATCC 12837 forms cysts profusely
under appropriate growth conditions. When this strain is
cultured in Burk's nitrogen-free broth with glucose, some
cysts form in stationary phase cultures, but ony 1.0% (Lin
and Sadoff, 1969) to 10.0% (Reusch and Sadoff, 1981) of the
population encysts under these conditions.
Early workers such as Winogradsky (1938) knew that
growing some Azotobacter spp. in nitrogen-free media, with
ethanol or butanol as carbon source, led to enhanced produc
tion of nonmotile, spherical cells with double-layered
coats. Socolofsky and Wyss (1961) built upon this
knowledge, using A. vinelandii ATCC 12837 (which' was used in
all the studies that follow unless otherwise indicated).

39
When cultured as cell lawns on Burk's nitrogen-free agar
with 0.3% (vol/vol) n-butanol as sole carbon source, cysts
began to appear within 3 days and predominated in 5 to 7
days. Ultrastructural studies revealed that the outermost
layer of the cyst, the exine, consisted of several over
lapping, plate-like layers. Beneath the exine was a much
thicker layer of gelatinous material, called the intine.
The intine surrounded a modified resting cell, called the
central body, which often contained numerous PHB granules.
Cysts had no detectable endogenous respiration when
suspended in buffer, but almost instantaneously began
measurable respiration when exogenous carbon sources were
added. In later studies, cysts were produced by growth on
0.2% (vol/vol) n-butanol (Socolofsky and Wyss, 1962), or
0.2% (wt/vol) 8-hydroxybutyrate (BHB) (Lin and Sadoff,
1968), with 90% or greater of the cells being converted to
cysts in 5 to 7 days.
Eklund et al. (1966) demonstrated that the formation of
capsular layers by vegetative cells was a prerequisite for
cyst formation. Complete morphological encystment of cells
grown on n-butanol agar with various levels of NH^NO^ only
occurred in the usual 5-day-period when the NH^NO^ concen
tration was 0.02 M or less. The cells rounded up within 5
days into nonmotile precysts lacking exines when 0.03 M or
0.04 M NH^NO^ was initially present. By day 10, these cells
had used up enough of the original combined nitrogen to

40
allow dinitrogen fixation to resume, so that capsular poly
saccharide was produced, followed by formation of exines
and, ultimately, morphologically mature cysts. Nonencap-
sulated mutants were unable to form morphologically mature
cysts. The work of Pope and Wyss (1970) emphasized that
cells beginning encystment first produced a capsule that
acted as a structure within which the cyst coats were built,
so that the exine existed inside of the capsule. The
diameter of morphologically mature Azotobacter cysts
measured between exine boundaries is about 2.0 pm (Reusch
and Sadoff, 1983).
Abortive encystment occurs when cells round up into
nonmotile precysts, but are unable to form a complete
exine. This occurs in the presence of high amounts of
combined nitrogen (Eklund et al., 1966), when glucose or
other carbon sources are present in addition to n-butanol or
BHB (Lin and Sadoff, 1968), or when calcium is unavailable.
The calcium requirement is probably related to its function
as a stabilizing cation that holds the cyst coats together
(Page and Sadoff, 1975). Using 3.0 mM EDTA in 0.05 M Tris
buffer, pH 7.8, Lin and Sadoff (1969) obtained almost
instantaneous expulsion of the central body from the cyst
coats, due to the chelating effect of the buffer. The empty
exines had the same "horseshoe" shape seen when cysts
germinate, and vegetative cells separate from the exines.

41
The role of PHB in cyst formation was examined by
Stevenson and Socolofsky (1966). Cysts were defined as
cells that could survive desiccation on a membrane-filter
surface for 4 days at 33C. After 2 days of growth on
nitrogen-free n-butanol agar, cells lost their motility,
became oval-shaped, and accumulated PHB to the extent of 35%
of cell dry weight. The development of mature cysts was
accompanied by a reduction in PHB content. By 6 days, cul
tures had undergone 100% encystment, and 10% of cyst dry
weight was PHB.
Lin and Sadoff (1968) developed a two-step replacement
procedure for obtaining cysts in broth. Cells were grown to
late exponential phase in Burk's nitrogen-free broth with
glucose. After harvest by centrifugation and washing in
buffer, cells were resuspended in Burk's salts broth with
0.2% (wt/vol) BHB. This procedure was used in further
studies (Hitchins and Sadoff, 1970, 1973; Reusch and Sadoff,
1979; Su et al., 1981; Reusch and Sadoff, 1983), resulting
in the following detailed description of BHB-induced
encystment.
After 1 hour in encystment broth, cells are still
motile and flagellated but no longer possess nitrogenase
activity. Within 4 to 6 hours, DNA synthesis has ceased,
and soon afterward each cell divides to form two nonmotile
precysts. There is rapid accumulation of PHB during this
period, and the rate of phospholipid synthesis declines.

42
Simultaneously, BHB is being taken up and respired or
incorporated. From the sixth to sixtieth hour, unique
lipids, not found in vegetative cells, begin to be
produced. These include 5-n-alkylresorcinols (ARl) and
their galactoside derivatives (AR2). These lipids possess
hydrophobic alkyl sidechains and hydrophilic phenolic
heads. Also produced are 6-n-alkylpyrones (AP), having a
similar bipolar nature. During this time, membranous
vesicles migrate outward from the central body through the
intine to form the exine layer. Up to 17% of the exine is
composed of ARl and AR2. The central body produces ARl and
AR2 in part from its PHB reserves, and exports them in the
membranous vesicles to the exine region. Radio-labelled BHB
accumulates in the central body and exine, whereas the
intine contains almost none. This indicates that the intine
is composed mainly of capsular material, formed from cell
reserves that were present before encystment is triggered by
BHB. Net RNA synthesis stops by the twelfth hour, and net
protein synthesis continues for up to 36 hours. Lipid
turnover continues beyond 60 hours, but there is no net
lipid synthesis. In a mature cyst, 5.0% of the central body
membranes are phospholipid, with AR and AP composing the
other 95%. Molecular models suggest that AR and AP form a
more rigid membrane structure at physiological temperatures
than do phospholipids. The hydrophobic, viscous nature of

43
such a membrane may contribute greatly to the desiccation
resistance and dormancy of cysts.
The possible contribution of the central body membranes
to stress resistance of cysts was suggested in earlier
studies. The cysts of Azotobacter chroococcum strain 75-1
had a compact, well-defined exine layer, whereas the exine
of A. chroococcum strain NTS was diffuse and fragile (Vela
and Cagle, 1969). The cysts of A. chroococcum strain 75-1
were much more resistant to sonic disruption than cysts of
A. chroococcum strain NTS. Yet cysts of both strains were
comparably resistant to desiccation on membrane filters and
to ultraviolet irradiation. Kramer and Socolofsky (1970)
defined cyst germination of A. vinelandii ATCC 12837 as a
process whereby desiccation resistance is lost; mature cysts
were defined as cell forms surviving 3 days of desiccation
on membrane filters. It was found that 40.0 pg
chloramphenicol/ml inhibited outgrowth of cysts in a
complete medium. Many cysts lost their desiccation
resistance when incubated with chloramphenicol, indicating
that the antibiotic might have chemically changed some
essential cyst component, perhaps the central body's mem
branes. Hitchins and Sadoff (1973) found that, soon after
exposure to BHB, vegetative cells became resistant to
100.0 pg chloramphenicol/ml. The antibiotic had no effect
on morphogenesis or rates of protein synthesis. This is
another indication of rapid membrane alteration of encysting

44
cells, long before AR and AP are produced. Further support
for the importance of membranes may be found in studies
where mineral nutrient deficiencies lead to the production
of stress-resistant cysts which lack completed cyst coats
(Gonzalez-Lopez et al., 1985).
Germination of cysts has usually been defined as the
emergence of a growing, motile cell from the exine layer
(Socolofsky and Wyss, 1961). Loperfido and Sadoff (1973)
examined the germination of cysts exposed to glucose. Cysts
respired detectably within 2 min. after the addition of 1.0%
(wt/vol) glucose, and soon afterwards net synthesis of RNA
and protein became measurable. After 4 to 6 hours, the
central body had enlarged to occupy the volume of the
intine, and DNA synthesis and nitrogenase activity became
measurable. After 8 hours, a vegetative cell emerged from
the cyst coats, leaving behind an empty "horseshoe"-shaped
exine. Germination did not occur in the absence of oxygen.
Cysts also germinated in the presence of sugars other than
glucose. Germination did not occur in Burk's nitrogen-free
salts, indicating that the PHB reserves of the cysts could
not be mobilized to initiate germination. The addition of
0.25% (wt/vol) NH^+ did not lead to germination.
When cysts are germinated on glucose, some central
bodies divide within their cyst coats to form multiple
central bodies. Up to six central bodies have been observed
within one cyst coat (Cagle and Vela, 1974).

45
Pleomorphism of Azospirilla
Bacteria cultured in vitro can be extremely
pleomorphic. Only a few cells in a population may exhibit
abnormal morphology under some cultural conditions, but
sometimes the majority of a culture assumes unusual shapes.
Older cultures in the stationary growth phase can be
especially pleomorphic (Duguid and Wilkinson, 1961).
Hughes (1956) has reviewed the development of bacterial
filaments. Filamentous cells are usually as wide as normal
cells, but are several times longer and lack developed
septa. They are interesting because they are often fully
viable, unlike some pleomorphic or involution forms of
bacteria. Under suitable cultural conditions, a filament
may divide at several points along its length to produce
several cells of normal length. Filaments can be induced by
sublethal cell damage, interruption of balanced growth, dyes
and antibiotics, extremes of pH, refrigeration, and various
forms of radiation.
Slater and Schaecter (1974) emphasized how sensitive
bacterial cell division is to the factors mentioned above.
If sublethally stressed, rod-shaped bacteria may continue to
grow and form filaments. Filaments can also form during
very rapid growth in rich media, and will fragment into
individual cells when growth slows, or when the environment
becomes less nutritionally rich. Since cells arising from
fragmentation of filaments are usually of normal length, the

46
cell's ability to control the site of cell division is not
lost during filamentation. Sometimes chains of cells occur
instead of filaments. The cells in chains contain septa,
but final cleavage between cells has not yet occurred. It
is possible that contiguous cells having incomplete septa in
such chains may share continuities between their
cytoplasms. In some cases, chains may be held together by
very thin capsular layers common to several cells in the
chain.
Jensen and Woolfoik (1985) found that several strains
of Pseudomonas putida and Pseudomonas fluorescens were
induced to form filaments if oxygen became limiting during
the late logarithmic phase of growth in nutrient broth.
Exhaustion of one or more nutrients was also a probable
elicitor of filamentation. The weakly motile filaments,
unlike the highly motile aerobic rods of the bacteria,
migrated to microaerophilic zones. As respiration of cul
tures declined, the increasing levels of oxygen in the broth
seemed to trigger fragmentation of the filaments into rods.
Cultures containing filaments, or the progeny of fragmented
filaments, retained viability longer than nonfilamentous
cultures.
Morphological changes in Escherichia coli have been
related to specific genes. If cellular DNA is damaged by
ultraviolet irradiation or other influences, several genes
are expressed in the so-called SOS response. Many of the

47
gene products are involved directly in repair of damaged
DNA, but some others specifically block further cell divi
sion. Until the DNA is repaired, cell division is blocked,
but cells can continue to grow into long, nonseptate fila
ments. Upon repair of the DNA, septa form along the fila
ments, and cells of normal size are produced after septum
separation (Donachie et al., 1984). Certain E. coli mutants
are known to produce septa, but form chains because the
enzymes needed for septum cleavage are not produced (Begg
and Donachie, 1985).
Thompson and Skerman (1979) showed that most members of
the Azotobacteraceae are pleomorphic under certain cultural
conditions. Filaments and chains of cells are produced
commonly. Similar pleomorphism has been observed with
azospirilla.
Becking (1982) observed that the morphology of
azospirilla varied in different culture media. On yeast
extract-glucose agar, the cells were highly motile, slightly
curved rods, 2.0 to 4.0 pm long and 1.0 pm wide. These
cells would often become swollen with three to five PHB
granules per cell. When cultured in nitrogen-deficient
broth supplemented with 0.01% (wt/vol) Difco yeast extract,
the cells often became long spirals of 30 to 40 pm in
length. These cells had reduced motility, but were capable
of rotation about their axes, and had few or no PHB
granules. Peptone was found to produce similar elongated,

48
weakly motile cells containing little PHB. These cells were
probably filaments, as described by Hughes (1956). Becking
did not study their viability.
Eskew et al. (1977) isolated and studied the pigmented
A. brasilense strain Cd. Nitrogenase activity peaked after
about 2 days of growth in semisolid, nitrogen-free malate
medium, and most cells were motile, curved rods of normal
size, often containing PHB. After 3 days, however, nitro
genase activity declined sharply. By this time, the initial
near-neutral pH of the medium had risen to pH 8.1. Most of
the bacteria present then appeared as enlarged, ovoid, non-
motile cells that were resistant to Gram-staining. The
decrease in nitrogenase activity, and shift to alkaline pH,
coincided with the appearance of cyst-like cells.
Tarrand et al. (1978) found that A. brasilense and
A. lipoferum strains had a similar appearance after 1 day's
growth in broth containing peptone, ammonium sulfate, and
succinate. Most cells were short, plump, slightly-curved
motile rods averaging 1.0 pm in diameter and 2.1 to 3.8 pm
in length. Cell morphology changed, especially for
A. lipoferum strains, when the cells were inoculated into
nitrogen-free, semisolid malate medium containing 0.005%
(wt/vol) yeast extract. Cells of A. lipoferum tended to
increase to 1.4 to 1.7 pm in width and to 5.0 pm to over 30
pm in length. Within 1 to 2 days, many A. lipoferum cells
became S-shaped or helical and retained little if any

49
motility. These long cells eventually fragmented into
shorter, ovoid cells. Many of these fragments later became
large, pleomorphic cells filled with light-refractile
granules, probably PHB. In contrast, A. brasilense strains
transferred to nitrogen-free, semisolid malate medium ini
tially retained their normal appearance. Only after several
weeks' time in this medium did they develop some S-shaped
cells and some large, pleomorphic, granule-filled forms.
Falk et al. (1985) found that A. amazonense strains failed
to become pleomorphic under comparable conditions.
Krieg and Dobereiner (1984) maintained that alkalini-
zation of the medium due to oxidation of malate was
responsible for pleomorphism in A. lipoferum. Cultures of
this species grown in semisolid, nitrogen-free glucose
medium did not become alkaline, and the cells did not become
pleomorphic.
Wong et al. (1980) isolated a putative Azospirillum
sp. from cellulolytic, dinitrogen-fixing mixed cultures. In
semisolid, nitrogen-poor malate medium containing adequate
levels of biotin, the cells were of normal size and
morphology after 1 day's growth. Between the third and
seventh days, the cells gradually became S-shaped and
enlarged. These enlarged cells contained granules of PHB
and/or polyphosphate. By 10 days, many cells had lysed and
released these granules into the medium. When the initial
biotin concentration of the medium was reduced to 10% of the

50
normal level, these morphological changes were accelerated,
occurring within 2 to 3 days after inoculation. This strain
could not fix dinitrogen with glucose as the carbon source,
but otherwise its biotin requirement and pleomorphism were
typical of A. lipoferum.
Lamm and Neyra (1981) found that A. lipoferum strains
grown in nitrogen-free, semisolid malate medium developed
many elongated cells after 2 days of culture, whereas A.
brasilense strains only developed elongated cells after 10
days. In both semisolid and agar-plate, nitrogen-free
malate cultures, thick-walled, optically refractile cells
were present as 1.0% of the cells in day-old cultures of
both species. After 4 days, the numbers of these thick-
walled cells were equal to or greater than cells of normal
morphology. The ovoid cells of A. lipoferum strains were
about twice as large as those produced by A. brasilense
strains. Such cells were never observed in nutrient broth
cultures, but could be obtained in old cell lawns grown on
nutrient agar. The increased desiccation resistance of
these ovoid cells has already been discussed.
Papen and Werner (1982) observed apparent cysts of A.
brasilense Sp 7. Such nonmotile cells were encapsulated,
having a diameter of about 1.2 pm, were not fixing dinitro
gen, and were desiccation resistant. They composed the
majority of cells in the nitrogen-free, malate broth culture
between the second and fourth days of incubation. After

51
this time, vegetative vibrioids emerged from the capsules to
grow and fix dinitrogen. The authors suggested that oxygen
limitation greatly affected these events. The level of PHB
increased as the oxygen level of the culture decreased;
nitrogenase activity ceased; and the cells encysted for a
time. Their apparent reduced respiratory activity allowed
the level of dissolved oxygen to be replenished in the
medium, until vibrioids emerged from the cyst coats to grow
and fix dinitrogen again. No encystment was observed when
cultures were incubated aerobically.
The recent work of Sadasivan and Neyra (1985) stressed
the roles that PHB and exopolysaccharides play in cyst
formation of azospirilla. Encysting cells lost their
motility and became enlarged and rounded. They accumulated
PHB and synthesized capsular material. The investigators
emphasized that common media, such as nutrient broth, do not
promote encystment and that development of mature exine and
intine layers may only be achieved under specific, well-
defined cultural conditions. Sadasivan (1985) may have
found the cultural conditions to promote maturation of cysts
of A. brasilense Sp 7. Using phase contrast microscopy, she
has observed vegetative cells emerging from cyst coats,
leaving behind empty "horseshoe"-shaped capsules. She has
also observed cysts containing from two to four central
bodies within a single exine. In transmission electron
microscopy thin sections, she has observed maturing cysts,

52
with membranous blebs migrating outward into the capsular
material from central bodies containing PHB granules. She
has also observed mature cysts with central bodies contain
ing PHB and polyphosphate granules, surrounded by distinct
intine and exine layers. Thus, given appropriate cultural
conditions, A. brasilense Sp 7 is able to form apparently
mature cysts, almost identical in appearance to those of
Azotobacter spp. One unusual feature she has reported is
layers of spherical, melanin-like granules outside the
exines of mature A. brasilense cysts; these layers have
never been observed with Azotobacter spp. cysts.
Berg et al. (1980) studied morphological and physio
logical changes of A. brasilense Sp 7 grown under different
conditions. Encapsulated cells (C-forms) were often present
on cell lawns grown on nitrogen-free succinate agar.
Encapsulation was initially heaviest for cells near the lawn
surface. After most cells in the surface layers were con
verted to nonmotile C-forms, the lower cell layers began to
accumulate capsules. Such C-forms were not observed within
60 hours of growth in semisolid, nitrogen-free succinate
agar. They formed rapidly on nitrogen-free agar surfaces.
Most of the culture formed capsules. The appearance of the
encapsulated forms varied and changed with time. Both
capsule formation and PHB accumulation were inhibited by
combined nitrogen. As cultures aged, enlarged vibrioid
C-forms developed. The more mature forms were spheres of

53
2.0 to 4.0 pm diameter which had lost their motility.
One-week-old cultures consisted mainly of spherical C-forms
of 5.0 to 8.0 pm diameter, containing many PHB-rich cells
within a common capsule. The authors speculated that
younger encapsulated forms may be fixing dinitrogen and that
older encapsulated forms may not. They suggested that the
capsule may reduce oxygen flow into the cells, thereby
protecting oxygen-sensitive nitrogenase activity. Azo
spirilla form extensive capsules only in media having a high
C/N ratio (Sadasivan and Neyra, 1985). Such conditions
promote nitrogenase activity. Since most capsules contain
over 99% of their weight as water (Costerton et al., 1981),
and oxygen diffuses through water at one ten-thousandth the
rate of diffusion through air (Clark, 1967), the capsule may
well help protect nitrogenase from oxygen damage. Oxidation
of PHB reserves within the cell may also reduce oxygen
levels near the nitrogenase (Dawes and Senior, 1973).
The description by Berg et al. (1980) of encapsulation
starting at the uppermost layers of nitrogen-free agar-grown
colonies and proceeding downwards is reasonable, if one
assumes that encapsulated cells are metabolically active for
a time, and then pass into dormancy. Initially, the upper
most encapsulated cells would be actively fixing dinitro
gen. They might become dormant as a result of underlying
cell layers depleting the available carbon supply, or pos
sibly because conditions become favorable for their passage

54
into constitutive dormancy (Marshall, 1980). Cells might
start to encyst when they accumulate threshold levels of
capsular material and/or PHB. In any case, as dormant cells
they would consume little oxygen, allowing it to diffuse to
lower cell layers that previously may have been oxygen-
limited, due to the actively respiring upper cell layer.
These lower cell layers would become more active with the
increased oxygen supply, accumulating capsules. Eventually
these cell layers would also pass into dormancy.
In earlier work, Berg et al. (1979) grew A. brasilense
Sp 7 in association with sugarcane callus tissue. Vegeta
tive cells (V-forms) grew as slimy colonies on the surface
of the callus, and few of these V-forms contained PHB or
capsules. Encapsulated or C-forms were also observed in
these conditions. This association of azospirilla with
sugarcane callus exhibited nitrogenase activity, but whether
the V-forms, C-forms, or both were responsible could not be
ascertained, since both were present. Perhaps C-forms were
able to fix dinitrogen transiently, but were poised to enter
dormancy if growth became too unbalanced. The bacteria did
not possess capsules near or within lysed plant cells, where
the C/N ratio may have been narrow, and balanced growth may
have been promoted.
An apparent contradiction in this plant callus-bacter
ium work is the claim by Berg et al. (1979) that C-forms of
azospirilla have little similarity to Azotobacter

55
spp. cysts. Krieg and Dobereiner (1984) restated this, but
the photographs of Berg et al. (1979) do not support it.
The multicellular C-forms are virtually indistinguishable
from Azotobacter spp. cysts having multiple central bodies
(Cagle and Vela, 1974). Clearly, in the association with
sugarcane callus, the azospirilla were situated in numerous
sites, differing in nutrient availability and oxygen avail
ability. It is not surprising that multiple morphologies
were observed, reflecting multiple physiological states.
Only a few cells resembling mature cysts were present.
Pleomorphic forms of azospirilla have been observed in
a variety of axenic associations with plant roots. The work
of Umali-Garcia et al. (1980) has already been discussed.
Ruscoe et al. (1978) grew maize plants in sand and inocu
lated them with different strains of azospirilla. Enlarged,
cyst-like cells, as well as cells of normal morphology, were
observed in older and thicker root segments, where root
tissue was often disintegrating. They also found that when
two strains of azospirilla were grown in nitrogen-free,
semisolid trans-aconitate agar, they often formed long
chains after 4 to 5 days.
Matthews et al. (1983) used immunological techniques
and transmission electron microscopy to observe strains of
A. brasilense in axenic association with pearl millet
roots. Both vibrioid and encapsulated cells were observed
in association with the roots. The encapsulated cells

56
usually contained PHB and polyphosphate granules, and often
two or more cells were enclosed by a common capsule.
Patriquin et al. (1983) observed unusual structures on
the surface of wheat roots, 3 weeks of age and older, that
had been axenically incubated with azospirilla in a sand-
vermiculite mix. They appeared as spherical "bags," within
which azospirilla containing PHB granules could be seen to
swim about. These structures were also found between the
epidermis and outer cortex of young wheat roots.
Krieg and Dobereiner (1984) suggest that the capsule of
azospirilla helps protect nitrogenase. They also support
the idea that development of alkaline pH is the cause for
pleomorphism in A. lipoferum and A. brasilense. This seems
an incomplete explanation, implying that pleomorphic cells
are poorly viable, being aberrant forms or laboratory arti
facts. Pleomorphic cells of azospirilla may instead develop
commonly, and perhaps transiently, when growing in natural
environments of high C/N ratio, such as near plant roots.
Unbalanced growth, with increased PHB and capsule formation,
may be the major cause of pleomorphism.
Experimental Approach
The conversion of 90% or more of an Azotobacter
sp. cell suspension to cysts facilitates physiological
studies of cysts. Growing the cells in nitrogen-free media
containing n-butanol or BHB leads to this conversion.

57
Azospirilla of cyst-like morphology have been observed under
various cultural conditions, but reports of conversion of
90% or more of cell populations to cyst-like forms are not
found in the literature. Vegetative cells are reported as
being present in high numbers, along with the cyst-like
forms. This has perhaps discouraged studies on the nature
of cyst-like forms of azospirilla.
All strains of A. brasilense and A. lipoferum are able
to grow on BHB as sole carbon source in the presence of
combined nitrogen (Tarrand et al., 1978). However, no
studies have been done to see how azospirilla respond to BHB
in the absence of combined nitrogen. Since such cultural
conditions lead to prolific encystment of some Azotobacter
strains, it was considered worthwhile to determine if
strains of azospirilla might also undergo conversion in high
numbers to cyst-like forms under these growth conditions.
The research reported here addresses the following
questions:
1. Can high numbers of cyst-like forms of azospirilla
be obtained by growth in nitrogen-free BHB broth
or on nitrogen-free BHB agar?
2. What are the morphological differences between
azospirilla grown on BHB with or without combined
nitrogen?

58
3. Is the PHB content of azospirilla grown on BHB
without combined nitrogen higher than when they
are grown in complex broth with combined nitrogen?
4. If pleomorphism of azospirilla occurs in nitrogen-
free BHB broth, is alkalinization of the medium a
prerequisite for development of pleomorphism?
5. Are azospirilla grown on nitrogen-free BHB agar
more desiccation resistant than cells grown in
complex broth with combined nitrogen?
6. Are azospirilla grown on nitrogen-free BHB agar
more resistant to starvation in carbon- and
nitrogen-free, phosphate-buffered salts solution
than cells grown in complex broth with combined
nitrogen?
7. What growth conditions favor motile azospirilla
arising from nonmotile azospirilla grown on
nitrogen-free BHB agar?
8. Is protein synthesis required before nonmotile
BHB-grown azospirilla give rise to motile
azospirilla?
9. Are BHB-grown azospirilla affected by Tris-EDTA in
a manner similar to Azotobacter cysts?
Questions 1 through 4 are considered in Chapter II, and
the remaining questions are considered in Chapter III.

CHAPTER II
PLEOMORPHISM OF AZOSPIRILLA GROWN ON
BETA-HYDROXYBUTYRATE
Only a few bacterial genera contain strains known to
form cysts (Sudo and Dworkin, 1973; Whittenbury et al.,
1970a; Tudor and Conti, 1977). A nonmotile cyst forms when
the entirety of a vegetative cell rounds up, depositing
extracellular coats and often accumulating intracellular
energy-reserve polymers.
The morphological changes of encystment are accompanied
by a reduction in cell metabolic activities, and increased
resistance to environmental stresses, such as starvation and
desiccation. Cysts of Azotobacter spp. are perhaps the best
understood. Like other prokaryotic resting cells, they form
when vegetative cells undergo a metabolic shift-down
(Sadoff, 1975).
Cysts of Azotobacter spp. do not form in media support
ing good vegetative growth until stationary phase, and are
present then only in low numbers (Sadoff et al., 1971).
Similarly, cells of azospirilla are uniform in shape during
active growth in nutritionally complete media (Umali-Garcia
et al., 1980; Lamm and Neyra, 1981; Sadasivan and Neyra,
1985). As is true for Azotobacter spp., however, stationary
phase cultures of azospirilla grown on complete media often
59

60
contain some rounded, nonmotile cells (Lamm and Neyra, 1981;
Papen and Werner, 1982; Krieg and Dobereiner, 1984).
Azospirilla are morphologically vexing, in that differ
ent pleomorphic cell types occur under various growth-
limiting conditions. Weakly motile filaments containing
little PHB form in aerobic broth which is low in combined
nitrogen (Becking, 1982). Under dinitrogen-fixing condi
tions, filamentous or S-shaped cells again may arise but
contain large deposits of PHB (Tarrand et al., 1978; Wong et
al., 1980; Lamm and Neyra, 1981). These elongated cells
often fragment into smaller, oval cells which subsequently
can assume a cyst-like morphology (Tarrand et al., 1978).
The most frequently reported pleomorphic form of
azospirilla is a nonmotile cell possessing thick outer
layers, probably of capsular material. These cells usually
contain more extensive deposits of PHB than do vegetative
cells grown with combined nitrogen. These cells have been
observed in older cultures grown on combined nitrogen (Lamm
and Neyra, 1981), in cultures grown as dinitrogen-fixers
(Eskew et al., 1977; Berg et al., 1979; Berg et al., 1980;
Papen and Werner, 1982), and in axenic associations with
grass roots (Ruscoe et al., 1978; Umali-Garcia et al., 1980;
Matthews et al., 1983). Recently, Sadasivan and Neyra
(1985) obtained them in broth containing fructose and KNO^.
The nomenclature for describing these cells is not
standardized. Berg et al. (1979) termed them capsulated or

61
C-forms as opposed to the vegetative or V-forms, as did some
later workers (Matthews et al., 1983; Krieg and Dobereiner,
1984). This terminology may be confusing, however, since
capsules can also occur on azospirilla of otherwise normal
morphology (Nur et al., 1980).
The presence of a capsule is usually deemed a pre
requisite for cyst formation in Azotobacter spp. (Eklund et
al., 1966). Azospirilla also may need to form a capsule
before they can form cyst-like cells. Encapsulated azo
spirilla may initially be fully active vegetative cells.
Upon encountering metabolic or environmental stress, such
cells may mature into cyst-like cells. The change in
morphology with time of some members within a C-form
population (Berg et al., 1980) may reflect maturation into
truly mature cysts. Two definitive traits of a mature
Azospirillum spp. cyst would be greatly reduced cell
metabolism and enhanced desiccation resistance. Morphologi
cally differentiated cells of azospirilla have been called
cysts when they exhibit no nitrogenase activity (Eskew et
al., 1977; Papen and Werner, 1982) or exhibit enhanced
desiccation resistance (Lamm and Neyra, 1981; Papen and
Werner, 1982; Sadasivan and Neyra, 1985).
Another complicating factor in understanding these
forms of azospirilla is that their appearance in dinitrogen
fixing cultures often coincides with alkalinization of the
growth medium (Eskew et al., 1977; Krieg and Dobereiner,

62
1984). Krieg and Dobereiner (1984) suggest that these cell
forms arise mainly at excessively high pH. In this case
they might be only laboratory artifacts, or involution
forms, that have no in situ function. The findings of Lamm
and Neyra (1981), Papen and Werner (1982), and Sadasivan and
Neyra (1985) argue against this viewpoint. Indeed, the
ability of azospirilla to enter dormancy as cysts may help
explain some of the great variability of plant responses to
inoculation with these bacteria (reviewed by Patriquin et
al., 1983).
Two things are presently lacking in research and under
standing of cyst-like forms of azospirilla. Although cyst
like forms of azospirilla have been predominant in some
studies, growing cells of normal morphology (vegetative
cells) have always been present in high numbers as well.
Conversion of 90% or greater of a population of vegetative
azospirilla to cyst-like forms (quantitative encystment) in
a reproducible manner would greatly facilitate further study
of these cell forms, as it did for Azotobacter spp. cysts
(Socolofsky and Wyss, 1962). Also lacking is an understand
ing of the underlying causes of pleomorphism and cyst
formation in azospirilla.
Conversion of 90% or greater of a cell population of
Azotobacter spp. to cysts often can be achieved by culturing
vegetative cells in the absence of combined nitrogen on
either of two precursors of PHB, n-butanol or BHB (Sadoff,

63
1975). Although all strains of A. brasilense and A. lipo-
ferum are known to grow on BHB as sole carbon source when
provided with combined nitrogen (Tarrand et al., 1978),
there are no reports of the response of azospirilla to BHB
in the absence of combined nitrogen. It was thought worth
while to see if vegetative azospirilla would respond simi
larly to Azotobacter spp., by undergoing quantitative
encystment in the presence of these carbon sources. In pre
liminary studies, apparent extensive PHB accumulation and
capsule formation were observed in some strains of
azospirilla grown with n-butanol. Since n-butanol is
volatile, BHB was used for later studies.
Initial objectives of this study were to achieve
morphological encystment of high numbers of azospirilla, to
document the morphology of such cells, to verify that they
contained PHB, and to ascertain if alkalinization of the
medium was a prerequisite for their formation.
Materials and Methods
Bacterial Strains
The Azospirillum brasilense strains used in these
studies were A. brasilense strain JM 125A2 and A. brasilense
strain Cd (ATCC 29729) (both courtesy of J. Milam, Univ. of
Florida, Gainesville). The Azospirillum lipoferum strains
used were A. lipoferum Sp RG6xx (ATCC 29731), A. lipoferum
Sp RG20a (ATCC 29708), A. lipoferum Sp RG8C, and

64
A. lipoferum Sp A3a (all courtesy of N. R. Krieg, Va. Poly.
Inst., Blacksburg). All strains were maintained on slants
of Tryptic Soy Agar (Difco Laboratories, Detroit, MI) at
25C with monthly transfer.
Media
Vegetative azospirilla were cultured in a modification
of the complete medium of Tyler et al. (1979), denoted as
trypticase-succinate salts (TSS). All components were of
reagent grade and were dissolved in deionized water. The
final concentrations of TSS components were (in grams per
liter): (NH^J^SO^, 0.5; succinic acid, 0.437; Trypticase
Peptone (Baltimore Biological Laboratory, Cockeysville, MD),
1.0; d-biotin (Sigma Chemical Co., St. Louis, MO), 0.0001;
NaCl, 0.1; FeCl3'6H20, 0.0017; Na2Mo04'2H20, 0.0002;
MgS04'7H20, 0.2; and CaCl2, 0.002. The first four com
ponents were omitted to obtain a basal salts solution. The
biotin was dissolved as a 100X concentrated stock solution
by heating and then filter-sterilized by passage through a
0.2 pm pore diameter Nalgene filter unit (Nalge Company,
Rochester, NY). Two phosphate buffer concentrations were
employed. The low phosphate (LP) buffer of Tyler et
al. (1979) had a final concentration of 3.5 mM and consisted
of (in grams per liter) K2HPC>4, 0.1 and KH2P04, 0.4. The
high phosphate (HP) buffer of Albrecht and Okon (1980) had a
final concentration of 63.8 mM and consisted of (in grams

65
per liter) I^HPO^, 6.0 and KI^PO^, 4.0. The LP buffer was
prepared as a 100X concentrated stock solution, and the HP
buffer as a 10X concentrated stock solution. The pH of the
LP buffer was adjusted to 7.1, and that of the HP buffer to
6.7, with 10 M KOH. The buffer stock solutions were
sterilized by autoclaving. All autoclavings in these
studies were for 15 min at standard temperature and pressure
unless otherwise stated.
The TSS components, excluding the biotin and phos
phates, were dissolved and adjusted to pH 7.0 with 10 M
KOH. The broth was then dispensed into 250 ml Erlenmeyer
flasks, in an amount calculated to obtain a final volume of
100 ml after aseptic addition of the biotin and phosphate
buffer stocks to the autoclaved TSS. The initial pH of the
LP-TSS was 6.9 to 7.0, and that of HP-TSS was 6.8.
Plate counts of azospirilla were performed with a
modified succinate-nitrogen free (SNF) agar medium derived
from Tyler et al. (1979) It was the same as LP-TSS, except
that (NH^)2^0^ and Trypticase Peptone were omitted. It
contained in addition (in grams per liter) Bacto yeast
extract (Difco), 0.05 and Bacto agar (Difco), 20.0. It was
prepared in the same way as TSS broth, except that agar was
added after neutralization and before autoclaving. Before
Petri plates were poured, biotin and LP buffer were added
aseptically, as was a solution of autoclaved Congo Red

66
(Sigma), that was incorporated at a final concentration of
0.0375 grams per liter (Rodriguez Caceres, 1982).
Nitrogen-free agar plates containing n-butanol had the
same composition as SNF plates, except that the agar concen
tration was 1.5% (wt/vol), and yeast extract and Congo Red
were omitted. The n-butanol was sterilized by filtration in
the same manner as the biotin and incorporated at a final
concentration of 0.2% (vol/vol). Beta-hydroxybutyrate was
prepared from crotonic acid (Sigma) by dissolving 23.6 g
crotonic acid in 900 ml deionized water. This solution was
continuously mixed with a magnetic stirrer for two to three
days at 25C. Its 0^235 stabilized by this time, indicating
conversion to BHB (H. L. Sadoff, personal communication).
It was then adjusted to pH 7.0 with 10 M KOH, the final
volume made up to one liter, and sterilized by autoclaving.
This served as a 10X concentrated stock solution of BHB for
addition to agar or broth, to give a final concentration of
0.236% (wt/vol) BHB. Agar plates containing BHB had the
same composition as n-butanol plates, except that (NH^^SO.
or Congo Red were sometimes added at the previously
described concentrations. For two-step broth replacement
studies (described below), broth contained BHB, biotin, and
phosphate-buffered basal salts solution. The LP and HP
buffers were employed in different broth replacement
studies. The initial pH, after inoculation, of LP-BHB broth
was 7.2, and that of HP-BHB broth was 6.9.

67
Growth Conditions
Inocula of azospirilla were grown in screw-cap tubes
containing 10 ml of autoclaved Bacto Nutrient Broth (Difco)
for 24-48 hours at 28C. One milliliter of inoculum was
aseptically pipetted into 100 ml of HP-TSS broth and
incubated for 20 to 22 hours at 30C at 130 rpm on a rotary
shaker. By this time, the cultures attained readings
of 0.7 to 0.9, as measured with a Bausch and Lomb Spectronic
20 spectrophotometer. The pH of the cultures at harvest
ranged from 7.0 to 7.2. Cultures were pelleted by centri
fugation at 6,960 X g for 15 min at 20C. Cells were washed
twice by resuspension and pelleting in sterile LP-basal
salts solution (pH 7.3). The cells were resuspended in
sterile LP-basal salts solution to give a final OD^q
reading of 1.0 to 1.2.
Cell lawns were obtained by spread plating 0.1 ml of
washed cells onto agar media. Inoculated plates were sealed
with Parafilm and incubated at 28C.
For two-step broth replacement studies, washed cells
were aseptically added as a 10% (vol/vol) inoculum to 250-ml
Erlenmeyer flasks, containing a final volume of 100 ml BHB
broth after cell addition. Duplicate flasks were incubated
in the same manner as TSS flasks. These studies were mod
eled after the two-step replacement method of Lin and Sadoff
( 1968 ) .

68
Harvest of Cell Lawns
To harvest lawns of azospirilla grown on n-butanol or
BHB agar, about 7.0 ml of sterile deionized water was asep-
tically poured across the surface of a cell lawn, and the
cells gently scraped from the agar surface with a flamed
wire loop. For PHB analyses and plate counts, the suspended
cells of one BHB agar plate were aseptically transferred to
another plate whose cells were in turn scraped off. This
was done to ensure that the cell suspension would not become
too diluted.
Enumeration
Vegetative cells from TSS-broth, or cells grown on BHB
agar, were diluted ten-fold in a series of dilution blanks
containing LP-basal salts solution. For enumeration, 0.1 ml
of cell suspension was aseptically spread on SNF-Congo Red
agar plates. Four plates were spread for each dilution.
Plates were incubated as described above for 5 days before
counting.
Dry Weight Determination and PHB Analysis
To assay PHB content of vegetative cells of A.
lipoferum Sp RG6xx, two 22-hour-old, HP-TSS cultures
(OD560 = 0.6) were pooled for centrifugation and washing as
described above, except that sterile deionized water was
used for washing. The final cell suspension was adjusted to

69
an OD560 = 0.86. Forty plates of the same strain grown on
nitrogen-free-LP-BHB agar were harvested by scraping
(described above), to give a final cell volume of about 100
ml. These cells were centrifuged and washed in sterile
deionized water (described above) and resuspended to give an
OD of 0.25 to 0.28.
560
For dry weight determinations, 10.0 ml of the final
cell suspension were pipetted into previously weighed and
desiccated aluminum pans. Five replicate pans were prepared
for each cell type. The pans containing cells were dried to
constant weight at 100C. Pans were kept in a glass
desiccator over anhydrous CaSO^ (Drierite) after removal
from the oven and before weighing.
For PHB determination, 10.0 ml of washed cells were
added to 15 ml Corex centrifuge tubes (Corning Glass Works,
Corning, NY) and pelleted by centrifugation at 7,080 X g for
20 min at 4C. Three replicate tubes were prepared for each
cell type. The supernatant was poured off, and subsequent
steps were performed by the method of Law and Slepecky
(1961). Digestion of cell pellets was begun with the addi
tion of 10 ml of Clorox bleach (5% (wt/vol) hypochlorite).
Cells were suspended in the bleach with Pasteur pipettes;
then the tubes were capped with glass marbles and incubated
in a 37C water bath. Digestion to constant OD^r. was
monitored with a Spectronic 20 spectrophotometer and was
judged to be complete after 18 hours. The insoluble cell

70
material was pelleted by centrifugation as above, then
washed once in 10 ml of sterile deionized water, and pel
leted again. The volume for all subsequent washings and
digestions was maintained at 10.0 ml, and all chemicals were
of reagent grade. The OD235 t^ie samPles i-n the final
digestion of concentrated H^SO^ was measured in quartz
cuvettes (1.0 cm light path), using a Carl Zeiss M4QIII
spectrophotometer. For the standard curve, the sodium salt
of DL-6-hydroxybutyric acid (Sigma) was dissolved directly
in concentrated H2SC>4. The standard curve was linear up to
8.0 gg BHB/ml. The PHB content of cell digests was related
back to dry weight values, to determine what percentage of
cell dry weight was present as PHB.
Scanning Electron Microscopy (SEM)
Samples of 0.4 ml from either LP-BHB agar plates or
two-step, broth-replacement cultures were employed for SEM
studies. Cells were removed aseptically from the two-step,
broth-replacement cultures at the same time that culture pH
was measured. Cells were aseptically impinged upon auto
claved 25-mm-diameter, 0.45-gm-pore-size Nuclepore poly
carbonate filters (Nuclepore Corporation, Pleasanton, CA),
housed in a filter chimney attached to a vacuum source.
About 10.0 ml of sterile, deionized water was added to the
chimney after cell addition, to help distribute the cells
evenly over the membrane surface, then a vacuum not

71
exceeding 33.8 kPa was applied. Filter membranes were then
removed and placed into Karnovsky's fixative (1965) for 1
hour. Filter membranes were subsequently rinsed twice for
10 min in cacodylate buffer and then dehydrated in a graded
series of ethanol concentrations (10, 20, 30, 50, 70, 90,
95, 100, and 100%) for 10 min at each concentration. The
samples were then air dried. Sections of filters were
excised, placed onto aluminum stubs with double-stick tape,
and gold coated with an Eiko IB-2 coater. Specimens were
examined with a Hitachi S450 scanning electron microscope at
20 kilovolts. Photographs were taken with Polaroid Type 55,
positive/negative, 4X5 Land film.
Light Microscopy
Cells were routinely observed by phase-contrast
microscopy using a Wild M20 or a Nikon Labophot microscope.
Cell dimensions were measured with an ocular micrometer.
Photographs of cells viewed with the latter microscope were
taken with a Microflex AFX camera attachment, using Ilford
FP4 black and white film. All photos were taken using
phase- contrast optics, unless otherwise indicated.
Results
Quantitative Morphological Change
In the initial phase of these studies only three
strains of azospirilla were used, A. brasilense strain

72
JM 125A2, A. brasilense strain Cd, and A. lipoferum Sp
RG6xx. Slime developed at the bottom of stationary phase
LP-TSS broth cultures of all three strains. Phase-contrast-
microscopy examination of A. lipoferum Sp RG6xx slime often
revealed numerous, nonmotile masses of cells similar to
zooglea, surrounded by nonmotile vegetative cells (Figure
2-la, b). These masses were notable for their symmetrical
but varied shapes. They were darker than most of the sur
rounding vegetative cells, perhaps indicating greater
viability than that of the surrounding pale vegetative
cells. These zoogleal masses retained their shape and did
not fragment into individual cells when disassociated from
the larger masses of cells. Similar zoogleal forms were
sometimes observed in the slime of A. brasilinese strain Cd
but not in that of A. brasilense strain JM 125A2. These
zoogleal forms of azospirilla may be referred to as micro
scopic floes, or microflocs, that are kept intact by
exopolysaccharide. Although microflocs were numerous,
individual normal cells were also present in large numbers
under these cultural conditions.
Azospirilla were cultured as cell lawns on agar con
taining precursors of PHB to see if high numbers of
pleomorphic forms would arise. The A. brasilense strains
produced some pleomorphic forms, but cells of normal shape
and size predominated, even on old lawns.

Figure 2-1. Zoogleal masses in stationary phase
40-dayold, low phosphate-trypticase-
succinate-salts broth culture of
Azospirillum lipoferum Sp RG6xx. a) Cells
at 600X magnification. Bar equals 6.0
pm. b) Detail from same mass of cells
viewed at 1,500X magnification. Bar
equals 3.0 pm.

74

75
After 63 hours of growth, A. brasilense strain JM 125A2
lawns grown on BHB contained ovoids, vibrioids, and chains
of cells. Many cells contained phase-bright, putative PHB
granules. More cells were present at this time on agar
containing combined nitrogen. The several cell types
present on 63-hour-old, LP-BHB agar with combined nitrogen
are shown in Figure 2-2a. Some cells appeared at this time
to be undergoing plasmolysis on this medium, as well as on
HP-BHB agar with combined nitrogen. By 96 hours, the lawns
on LP-BHB agar with and without combined nitrogen contained
more chains of cells and microflocs than the HP-BHB lawns,
which consisted mostly of individual ovoids or pairs of
ovoids.
Cells from month-old, nitrogen-free, HP-BHB lawns of
A. brasilense strain JM 125A2 are shown in Figure 2-2b.
Individual, nonmotile vibrioids and ovoids were still pre
dominant, as were pairs of cells. Enlarged, nonmotile,
spherical cells were present, but not numerous. A few non
motile filaments appeared to be undergoing septation.
After 79 days, cells from lawns of this strain grown
with combined nitrogen had the appearance of stationary
phase cells from TSS broth cultures grown with combined
nitrogen, and spheroplasts and cell ghosts predominated.
Nitrogen-free cultures at both phosphate buffer concentra
tions contained numerous pleomorphic forms. Figure 2-3a
shows nonmotile, enlarged, rounded individual cells from

Figure 2-2.
Cell types of Azospirillum brasilense
strain JM 125A2, grown on g-hydroxybuty-
rate (BHB) agar. a) 63-hour-old cells
from low phosphate-BHB agar with combined
nitrogen. 1,500X magnification. Bar
equals 3.0 pm. b) Month-old cells from
nitrogen-free, high phosphate-BHB agar.
1,000X magnification. Bar equals 4.0 pm.

77
a
b

Figure 2-3. Cell types of Azospirillum brasilense
strain JM 125A2, from 79-day-old nitrogen-
free, high phosphate-B_hydroxybutyrate
agar cultures. a) Individual rounded
cells. 1,500X magnification. Bar equals
3.0 pm. b) Microfloc focused so that
capsules are visible around cells on
right side of floe. 1,500X magnification.
Bar equals 3.0 pm. c) Same cells as (b),
but focused so that capsules are no longer
evident. Note empty capsule at bottom of
floe. 1,500X magnification. Bar equals
3.0 pm.

79
a
c

80
nitrogen-free HP-BHB agar. Their phase-bright inclusions
are probably PHB granules; some contain dark bodies,
probably polyphosphate granules. Also shown is a microfloc
of nonmotile, enlarged, PHB-rich cells (Figure 2-3b, c). By
adjusting the distance of the objective lens from the
specimen, many of these cells were observed to be encap
sulated (Figure 2-3b). The encapsulated cells fitted
together closely, as did those observed by Sadasivan and
Neyra (1985). The thickness of the capsule was about
0.5 pm. Such encapsulated microflocs were also observed on
nitrogen-free, LP-BHB agar at this time.
Azospirillum brasilense strain JM 125A2 may have lacked
an efficient mechanism for BHB uptake, compared to the other
strains of azospirilla used. Unlike the other strains, few
motile cells were observed on any BHB agar medium, even in
young cultures. It also differed from the other strains by
having many phase-dark cells that contained little or no
PHB. It eventually grew well on BHB agar when combined
nitrogen was available, however. A final difference between
this strain and the others was that its cells always resus
pended in water to give uniform turbidity, with no macro
scopic floes, or macroflocs, being present. This indicates
that, with or without combined nitrogen, cells of this
strain produce little capsular material when cultured on
BHB.

81
The best growth of A. brasilense strain JM 125A2 on
agar occurred on SNF-Congo Red agar. Cells from 6-day-old
lawns grown on this agar medium were often seen as encap
sulated microflocs (Figure 2-4). The microfloc in Figure
2-4a and b appears to have arisen mainly from one or more
filamentous cells that underwent septation. This may also
have occurred for many of the cells in Figure 2-4c. The
capsules were of thickness comparable to those observed on
BHB agar, about 0.5 pm. The lawns on SNF-Congo Red agar had
a scarlet or blood-red appearance, unlike lawns of this
strain growing on nitrogen-free BHB-Congo Red agar, which
were pale orange.
The other A. brasilense strain, A. brasilense strain
Cd, also failed to convert in high numbers to pleomorphic
forms, but it grew far better on BHB. After 63 hours of
growth, lawns of this strain on each BHB agar medium con
tained many motile vibrioids possessing large granules of
putative PHB. Elongated, filamentous cells were also
present in high numbers. These cells had about the same
width (1.5 pm) as normal dinitrogen-fixing cells but were
much longer, some being 9 to 13 pm in length (Figure 2-5a).
The filaments were sometimes observed to undulate slowly and
were much slower than motile vibrioids. In the presence of
combined nitrogen, these filaments were seen to septate and
fragment. This fragmentation was observed at 63 to 96
hours, and sometimes was complete within a population of

Figure 2-4. Cells of Azospirillum brasilense strain JM
125A2, from 6-day-old lawns grown' on
succinate-nitrogen-free-Congo Red agar,
a) Microfloc showing capsules and
filamentous cell patterns. 1,500X mag
nification. Bar equals 3.0 pm. b) Same
floe as (a), but focused so that capsules
and filamentous cell outlines are no
longer evident. 1,500X magnification.
Bar equals 3.0 pm. c) Different mass of
encapsulated cells. 1,500X magnifica
tion. Bar equals 3.0 pm.

* '
% M
r#,
vf /
* df
s'* *

Figure 2-5. Cell types of Azospirillum brasilense
strain Cd, from lawns on B-hydroxybutyrate
(BHB) agar. a) Filaments from 63-hour-
old, high phosphate-BHB agar with combined
nitrogen. 1,500X magnification. Bar
equals 3.0 pm. b) Microfloc from
11-day-old nitrogen-free, low phosphate-
BHB agar, focused to show capsules and
filamentous cell outline. 1,000X magnifi
cation. Bar equals 4.0 pm. c) Same floe
as (b), but focused so that capsules and
filamentous cell outline are no longer
evident. 1,000X magnification. Bar
equals 4.0 pm.

85
a
I
b
c

86
filaments soon after 96 hours. In nitrogen-free cultures,
such elongated filaments persisted, some being weakly motile
even after 79 days on nitrogen-free LP-BHB agar.
After 96 hours, lawns of A. brasilense strain Cd on all
BHB agar media contained mixtures of vibrioids, ovoids,
filaments, and chains. Sometimes the cell material from
LP-BHB agar lawns with or without combined nitrogen did not
resuspend uniformly in water, but as macroflocs, due to
extensive encapsulation. A microfloc from an 11-day-old,
nitrogen-free, LP-BHB agar plate is shown in Figure 2-5b,c.
By adjusting the objective lens, the capsule is made
evident. The entire microfloc may have arisen from one
elongated filament that underwent septation, as suggested by
the apparent linear continuities between cytoplasmic
contents.
After 79 days of culture, lawns of this strain grown
with combined nitrogen contained mainly spheroplasts and
cell ghosts, appearing to have entered stationary phase.
Cultures grown on nitrogen-free agar at both phosphate
levels contained numerous pleomorphic forms at this time, in
addition to cells of normal morphology (Figure 2-6).
A microfloc of A. brasilense strain Cd from a 6-day-old
lawn on SNF-Congo Red agar is shown in Figure 2-7. All the
cells are encapsulated, and empty capsules are evident. The
lawn was scarlet in color, unlike the pale orange lawns of
the same age grown on nitrogen-free BHB-Congo Red agar.

Figure 2-6.
Cell types of Azospirillum brasilense
strain Cd, from 79-day-old lawns on
nitrogen-free, 3-hydroxybutyrate (BHB)
agar. a) Multicellular packets, a chain
and individual ovoids from low phosphate
BHB agar. 1,500X magnification. Bar
equals 3.0 pm. b) Microfloc from high
phosphate-BHB agar, focused to show
capsules around several cells. Air
bubble is above floe. 1,500X magnifica
tion. Bar equals 3.0 pm.

88
b

89
Figure 2-7. Microfloc of Azospirillum brasilense
strain Cd, from 6-day-old lawn on
succinate-nitrogen-free-Congo Red agar.
Note capsules around cells and empty
capsules. 1,500X magnification. Bar
equals 3.0 pm.

90
The two A. brasilense strains failed to produce morpho
logically uniform populations on BHB-agar. A uniform
response was observed for A. lipoferum Sp RG6xx. Good
growth usually occurred within 18 to 24 hours. Figure 2-8
shows 18-hour-old cells grown on nitrogen-free, HP-BHB
agar. Filaments and chains were present, which were some
times as swiftly motile as vibrioids. On LP-BHB agar at
both phosphate levels, with or without combined nitrogen,
septation of filaments was almost complete between 48 and 72
hours, although new filaments would arise and septate for up
to the fifth day. Figure 2-9a, b shows such completely
septated microflocs on 63-hour-old, nitrogen-free, LP-BHB
agar. The floes are encapsulated, and most seemed to arise
from one filament that underwent complete septation.
Although microflocs were present at this time, the cells
resuspended from agar as uniformly turbid suspensions with
out macroflocs.
Many filaments were also completely septated by 63
hours on HP-BHB agar containing combined nitrogen, but some
filaments still lacked completed septa (Figure 2-9 c).
Nitrogen-free, HP-BHB agar at this time contained few if any
microflocs. As for all other media, very motile vibrioids
and ovoids, as well as filaments of varying motility, were
present at 63 hours.
Encapsulated microflocs sometimes formed on nitrogen-
free, HP-BHB agar (Figure 2-10a), but cells from young or

Figure 2-8.
Cells of Azospirillum lipoferum Sp RG6xx
from 18-hour-old lawn on nitrogen-free,
high phosphate-8-hydroxybutyrate agar.
Note individual cells and filaments
at various stages of septum formation.
1,000X magnification. Bar equals 4.0 pm

Figure 2-9. Cell types of Azospirillum lipoferum
Sp RG6xx, from 63-hour-old lawns on
B-hydroxybutyrate (BHB) agar. a) Micro-
flocs from nitrogen-free, low phosphate-
BHB agar, focused to show capsules and
filamentous cell outlines. 1,500X
magnification. Bar equals 3.0 pm.
b) Same microflocs as (a) but focused so
that capsules and filamentous cell
outlines are no longer evident.
1,500X magnification. Bar equals 3.0 pm.
c) Filament from high phosphate-BHB agar
with combined nitrogen. 1,500X magnifi
cation. Bar equals 3.0 pm.

93
c

Figure 2-10-
Cell types of Azospirillum lipoferum
Sp RG6xx from lawns on B-hydroxybutyrate
(BHB) agar. a) Microfloc from 13-day-
old cell lawn on nitrogen-free HP-BHB
agar. Note empty capsules. 1,500X
magnification. Bar equals 3.0 pm. b)
Individual cells and septating filaments
from 13-day-old, high phosphate-BHB agar
with combined nitrogen. 1,500X magnifi
cation. Bar equals 3.0 pm. c) Micro
floc from 79-day-old, low phosphate-BHB
agar with combined nitrogen. 1,500X
magnification. Bar equals 3.0 pm.

95
a
c

96
old lawns grown on it were always resuspended uniformly,
without macroflocs. Older cultures on this medium consisted
mainly of PHB-rich, nonmotile ovoids or peanut-shaped cells,
with few if any microflocs. The high phosphate level did
not appear to inhibit PHB accumulation, but did inhibit
extensive capsule formation. As was true for the
A. brasilense strains, combined nitrogen led to eventual
good growth and passage into stationary phase. Figure 2-10b
shows cells of a 13-day-old culture from HP-BHB agar
containing combined nitrogen. Chains of cells and
individual ovoids are present.
The LP-BHB lawns grown with or without combined nitro
gen had the same appearance by 7 days. The lawns consisted
almost entirely of floes that broke into various sizes when
resuspended in water. The cells would not suspend evenly in
water, due to the presence of many macroflocs. Very few
motile cells were present at this time. Eventually the
LP-BHB lawns grown with combined nitrogen resumed vegetative
growth and passed into stationary phase, but the floes
persisted even in stationary phase cultures (Figure 2-10c).
Figure 2-11 shows cells from 17-day-old, nitrogen-free,
LP-BHB-Congo Red lawns. Figure 2-lla was taken with
bright-field optics, showing the clearly outlined capsules
and enlarged PHB-rich cells. Figure 2-llb was taken with
phase-contrast optics, and the capsules enclosing all of the
microfloc are again evident. It was interesting to find

Figure 2-11.
Cell types of Azospirillum lipoferum Sp
RG6xx, from 17-day-old lawns on nitro
gen-free low phosphate-6-hydroxybuty-
rate-Congo Red agar. a) Microfloc
viewed with bright-field optics. 1,500X
magnification. Bar equals 3.0 pm. b)
Another microfloc, viewed with phase
contrast optics. Note that capsules of
the two bottom-left cells are apparently
undergoing division with their cells.
1,500X magnification. Bar equals 3.0
pm.

98

99
that some cells that were undergoing division in one floe
appeared to have their capsules dividing as well at the site
of septum formation (Figure 2-llb).
Cells from 17-day-old, SNF-Congo Red lawns are shown in
Figure 2-12. Abundant capsules are again evident. These
cells appeared more vibrioid in shape than most A.
brasilense cells cultured in the same manner. Cells of
A. lipoferum Sp RG6xx grown on SNF-Congo Red agar appeared
less swollen and rounded than their counterparts on LP-BHB
Congo Red agar. However, unlike the A. brasilense strains,
lawns of A. lipoferum Sp RG6xx were scarlet on both
SNF-Congo Red and LP-BHB-Congo Red agar.
Nitrogen-free, LP-BHB cultures of A. lipoferum SP RG6xx
did not change in appearance from the seventh day onward,
even after months had passed. Figure 2-13 shows a microfloc
of this strain, with the objective lens adjusted to show the
capsules (Figure 2-13a), and then readjusted to show the
capsules and the apparent continuities between cytoplasms
(Figure 2-13b). Cells from this medium of 7-days-age or
older were consistently resuspended as macroflocs and
microflocs. Individual, septated filaments apparently
consolidated into floes, and individual, motile cells may
have attached to septated filaments to give rise to large
macrofIocs.
More details of floe structure of A. lipoferum Sp
RG6xx were obtained from SEM photographs. Figures 2-14 to

Figure 2-12.
Cells of Azospirillum lipoferum Sp
RG6xx, from 17-day-old lawns on
succinate-nitrogen-free-Congo Red agar,
a) Microfloc with empty capsules, as
well as capsules retaining their cells.
1,500X magnification. Bar equals
3.0 pm. b) Another mass of encapsulated
cells. Note that some capsules seem to
contain inclusion granules but no
cytoplasm. 1,500X magnification. Bar
equals 3.0 pm.

101
b

Figure 2-13.
Microflocs of Azospirillum lipoferum Sp
RG6xx, from 36-day-old nitrogen-free,
low phosphate-3-hydroxybutyrate agar,
a) Microflocs focused to show capsules.
1,000X magnification. Bar equals 4.0
pm. b) Same floes as (a) but focused to
show both capsules and filamentous cell
outlines. 1,000X magnification. Bar
equals 4.0 pm.

103
b

104
2-16 show successive magnifications of cells scraped from
a 75-day-old, nitrogen-free, LP-BHB agar lawn. Cells from
19-, 57-, and 66-day-old lawns on the same medium were of
identical appearance to this older culture. This supported
the phase-contrast microscopy studies, in that the morphol
ogy of microflocs from nitrogen-free, LP-BHB agar did not
change once they were formed. The 500X magnification photo
(Figure 2-14) shows the great variability in numbers of
cells per floe and that individual, encapsulated cells are
often present. The 1,500X-magnification photo (Figure 2-15)
reveals, in agreement with the phase-contrast observations,
the frequent tight fit between adjacent capsules. Inter
cellular gaps are often observed within floes. The 7,000X-
magnification photo (Figure 2-16) reveals some variability
in the surfaces of encapsulated cells, possibly indicating a
difference in exopolysaccharide composition. It is clear at
this higher magnification that the cells of agar-grown floes
range from monolayers to trilayers. Empty capsules are also
visible.
Similar results were obtained with this strain on LP-n-
butanol agar. Figure 2-17 shows 178-day-old cells from this
agar medium after the agar had dried into a thin film. When
the agar surface was rehydrated and the cells scraped from
it, many empty capsules were seen, but many capsules still
contained cells. The capsular material was thus observed to
retain its outline, even if it contained no cell.

105
Encapsulated cells of Azospirillum
lipoferum Sp RG6xx from a 75-day-old
lawn on nitrogen-free, low phosphate-8-
hydroxybutyrate agar. Cells are viewed
at 500X magnification by scanning elec
tron microscopy. Bar equals 50.0 pm.
Figure 2-14.

106
Encapsulated cells of Azospirillum
lipoferum Sp RG6xx from a 75-day-old
lawn on nitrogen-free, low phosphate-B-
hydroxybutyrate agar. Cells are viewed
at 1,500X magnification by scanning
electron microscopy. Bar equals 5.0 pm.
Figure 2-15.

107
Figure 2-16. Encapsulated cells of Azospirillum
lipoferum Sp RG6xx from a 75-day-old
lawn on nitrogen-free, low phosphate-B-
hydroxybutyrate agar. Cells are viewed
at 7,000X magnification by scanning
electron microscopy. Bar equals 5.0 pm.

Figure 2-17.
Microflocs of Azospirillum lipoferum
Sp RG6xx from 178-day-old lawns grown on
nitrogen-free, low phosphate-n-butanol
agar. a) Microfloc showing numerous
cells within capsules as well as empty
capsules. 1,000X magnification. Bar
equals 4.0 pm. b) Microflocs with few
cells remaining within capsules. 1,000X
magnification. Bar equals 4.0 pm.

109
a
b

110
Growth of this A. lipoferum strain on nitrogen-free,
LP-BHB agar resulted in homogeneous encapsulation and fila
ment formation. But, as is evident from the photographs,
several cell shapes and sizes were present within any one
floe. Despite their morphological heterogeneity, cells in
these floes generally appeared to be more rounded and
swollen than cells grown on SNF-Congo Red agar, although
capsules were of equal width (0.5 pm) under both cultural
conditions.
The homogeneous encapsulation and filament formation of
A. lipoferum Sp RG6xx prompted a search for similar response
in other strains of this species. Three other strains of
this species, cultured on LP-BHB nitrogen-free agar,
responded about as well as A. lipoferum Sp RG6xx (Figures
2-18 to 2-20). Cells of all strains did not resuspend
uniformly in water, due to macroflocs. As was true for
A. lipoferum Sp RG6xx, the appearance of the other three
strains did not change noticeably with time, and several
cell sizes and shapes were usually present within any one
microfloc. Azospirillum lipoferum Sp A3a differed from all
the other A. lipoferum strains in consistently having large
numbers of individual, PHB-rich, nonmotile, nonencapsulated,
ovoid cells in its resuspended lawns. Possibly many of
these individual cells were initially present within
capsules, but were released from capsules upon the addition
of water. The other strains had only a few free,

Figure 2-18.
Microflocs of Azospirillum lipoferum
Sp RG8c, grown on nitrogen-free, low
phosphate-6-hydroxybutyrate agar. a)
Microfloc from 69-day-old lawn, focused
to show capsules and filamentous cell
outline. 1,500X magnification. Bar
equals 3.0 pm. b) Same floe as (a) but
focused so that capsules and filamentous
cell outline are no longer evident.
1,500X magnification. Bar equals 3.0
pm. c) Microflocs from 136-day-old
lawn. Note empty capsules. 1,000X
magnification. Bar equals 4.0 pm.

112
c

Figure 2-19. Microflocs of Azospirillum lipoferum
Sp RG20a, from 69-day-old nitrogen-free,
low phosphate-B-hydroxybutyrate agar,
a) Microflocs focused so capsules and
filamentous cell outline are evident in
upper left part of left floe. 1,500X
magnification. Bar equals 3.0 pm. b)
Microflocs with cells having varied
morphologies. 1,500X magnification.
Bar equals 3.0 pm.


Figure 2-20. Microflocs of Azospirillum lipoferum Sp
A3a from nitrogen-free, low phosphate-8-
hydroxybutyrate agar. a) Microfloc from
69-day-old lawn. 1,500X magnification.
Bar equals 3.0 pm. b) Microfloc from
136-day-old lawn. Note empty capsules.
1,000X magnification. Bar equals
4.0 pm.

9TI

117
nonencapsulated cells in their suspensions. Whether these
free cells never merged with microflocs, or whether they
were ejected from their capsules upon wetting and suspension
for microscopy, is not clear. The width of the capsules on
these three strains was again observed to average about 0.5
pm, and most floes arose from one or more filaments that
eventually underwent septation.
Enumeration of Encapsulated Cells
It was first suspected that cells of A. lipoferum Sp
RG6xx harvested from nitrogen-free, LP-BHB agar would not be
quantifiable by plate counting. Plate counts demand that
the inoculum be uniformly suspended and diluted, and floccu
lation makes these difficult to accomplish. But the resus
pended cells gave consistent CFU counts for a given OD^q
range (Table 2-1). After resuspension in water and washes
in buffer, the cell suspensions appeared silvery, and floe
size was reduced to the lower limits of visibility to the
naked eye. Macroflocs broke into smaller microfloc domains,
which had formed within the first few days on the agar
medium. These small domains retained their integrity even
after repeated shaking and washing steps. The similar CFU
counts for cells of different ages suggested that the cells
were in a sort of stasis, where they were no longer
multiplying or dying off appreciably.

118
Table 2-1. Optical density (ODc- n ) and colony forming units
ml- (CFU ml 1) of encapsulated cells of
Azospirillum lipoferum Sp RG6xx.
Culture age
(days)a
OD560
CFU ml 1
o
r1
X
11
0.27
2.67
(0.12 )b
21
0.29
5.08
(0.79)
22
0.25
6.40
(1.00)
58
0.24
1.94
(0.19)
75
0.25
2.11
(0.32)
80
0.30
6.65
(0.49)
aCells were harvested from nitrogen-free, low phosphate-B
-hydroxybutyrate agar plates.
Values are averages of four spread plates. Values in
parentheses are standard deviations.

119
Poly-B-Hydroxybutyrate Content
The phase-bright inclusion bodies in the cells of
A. lipoferum Sp RG6xx grown on nitrogen-free, LP-BHB agar
were confirmed to be PHB using the method of Law and
Slepecky (Table 2-2). This method is subject to error due
to repeated centrifugations and pipettings, so it is not
certain whether the differences in PHB content between cell
lawns of different ages were real or artifacts. The purpose
was to verify that PHB existed in the encapsulated cells in
greater amounts than in vegetative cells grown in HP-TSS
broth. The assay gave evidence of this.
Two-step Broth Replacement Studies
Strains of A. lipoferum cultured in LP-TSS broth
clumped and flocculated in under 24 hours. Clumping of
these strains was delayed in HP-TSS broth. The pH of these
cultures at 20 to 22 hours ranged from 7.0 to 7.2, having
risen from an initial pH of 6.8. Cells of each A. lipoferum
strain in this study, grown in HP-TSS broth, usually started
to clump and flocculate by about 24 hours after inocula
tion. Figure 2-21 shows a SEM photo of cells from a 14-day-
old, HP-TSS culture of A. lipoferum Sp RG6xx. The pH of the
culture at this time was 9.1. The floe has a similar
arrangement to microflocs of the same strain grown on BHB
agar, with filamentous cells and frequent spaces in the

120
Table 2-2. Poly-8-hydroxybutyrate (PHB) content of
Azospirillum lipoferum Sp RG6xx.
Cultural
conditions
Age
Dry
weight
(mg ml 1)a
Volume
of CHC13
extract"3
used (ml)
,-1
ug i
PHB
% dry
weight
as PHB
1C
11
days
0.17 (0.02)
0.5d
3.39 (0.52)
39.9
1
21
days
0.18 (0.05)
0.5e
4.56 (0.33)
50.7
1
22
days
0.12 (0.04)
0.5f
3.45 (0.03)
57.5
2?
22
hours
0.49 (0.03)
3.0e
5.28 (0.08)
3.6
aValues are averages of five replicates. Values in parentheses
are standard deviations.
j_
Two H^SO^ replicates of each CHCl^ replicate were used. Values
are averages of all J^SO^ replicates. Values in parentheses are
standard deviations.
1 = Nitrogen-free, low phosphate-8-hydroxybutyrate agar.
Four replicate CHCl^ extracts were used.
eThree replicate CHCl^ extracts were used.
^Two replicate CHCl^ extracts were used.
g2 = High phosphate-trypticase-succinate-salts broth.

121
Cells of Azospirillum lipoferum Sp
RG6xx, from 14-day-old, stationary
phase, high phosphate-trypticase-suc-
cinate-salts broth culture of pH 9.1.
Cells are viewed at 1,500X magnification
by scanning electron microscopy. Bar
equals 5.0 pm.
Figure 2-21.

122
clumps. Sometimes floes from HP-TSS broth cultures had no
such empty spaces.
Two-step, broth-replacement studies were conducted to
see if quantitative pleomorphism could be induced in broth
and if there was any connection between pH and pleomor
phism. Sometimes such cultures of A. lipoferum strains
would clump extensively within 24 hours, so that the broth
appeared clear to the naked eye except for the clumps, but
this phenomenon was not consistently reproducible. It
seemed that clumping occurred, sooner and persisted longer in
two-step, broth-replacement studies with A. lipoferum Sp
RG6xx using LP buffer. When HP buffer was used, clear broth
columns often became turbid eventually.
Figure 2-22a shows cells of A. lipoferum Sp RG6xx from
a 43-day-old culture in LP-BHB broth of pH 8.4. The cells
appear healthier than stationary phase cells in TSS broth.
Filamentation and complete septation of cells occurred in
most cases. The elevated pH of the culture again indicated
the low buffering capacity of the LP buffer.
The results from HP-BHB, two-step, broth-replacement
studies indicated that pleomorphism and encapsulation could
occur at near-neutral pH for the A. lipoferum strains. All
the following studies were conducted in HP-BHB broth.
Figure 2-22b shows cells from a 43-day-old culture of
A. lipoferum Sp RG6xx of pH 7.2. In interpreting this
photo, it should be remembered that the cells were added by

Figure 2-22.
Cells of Azospirillum lipoferum Sp RG6xx
from 43-day-old, nitrogen-free, g-
hydroxybutyrate (BHB) broth cultures,
viewed by scanning electron microscopy,
a) Cells from low phosphate-BHB broth,
pH 8.4. 7,000X magnification. Bar
equals 5.0 pm. b) Cells from high
phosphate-BHB broth, pH 7.2. Note
thorough encapsulation of lower cell
layer.. 3,000X magnification. Bar
equals 5.0 pm.

124
a
b

125
pipette to the filter membrane. It appears that large,
encapsulated cell floes settled first onto the filter,
followed by free, nonflocculated cells. The lower layers of
cells have the same close-fitting appearance as was often
observed on BHB agar surfaces.
Similar results were obtained for the other three
strains of A. lipoferum. Figure 2-23a shows cells from a
33-day-old culture of A. lipoferum Sp A3a of pH 7.2. Some
encapsulated cells in the lower cell layer are fitted
together snugly, while others are joined by strands of
putative exopolysaccharide. Figure 2-23b is a higher
magnification of cells from this culture, again showing
the strands joining cells. The lumpy appearance of the
cells is probably due to large, intracellular accumulations
of PHB.
Figure 2-24 shows cells from a 33-day-old culture of
A. lipoferum Sp RG20a of pH 7.2. The formation of filaments
and eventually chains was very pronounced in this strain.
Cells from a 33-day-old culture of A. lipoferum Sp RG8c
of pH 7.1 are shown in Figure 2-25a. This strain was some
times observed to form intricately structured clumps.
Figure 2-25b shows such a clump from a 9-day-old culture of
pH 7.0. Filamentous, septate cells are present, and again
empty spaces occur within the floe.
The two A. brasilense strains used in this study
responded poorly to two-step broth replacement. The pH of

Figure 2-23. Cells of Azospirillum lipoferum Sp A3a
from a 33-day-old, nitrogen-free, high
phosphate-6-hydroxybutyrate broth cul
ture, pH 7.2, viewed by scanning
electron microscopy. a) Lower cell
layer is thoroughly encapsulated.
3,000X magnification. Bar equals
5.0 pm. b) Details of cell surfaces.
Note strands of material joining some
cells. 17,000X magnification. Bar
equals 0.5 pm.

127
b

128
Figure 2-24. Cells of Azospirillum lipoferum Sp RG20a
from a 33-day-old, nitrogen-free, high
phosphate-8-hydroxybutyrate broth
culture, pH 7.2, viewed by scanning
electron microscopy. Note thorough
encapsulation of lower cell layer and
presence of filaments and chains.
3,000X magnification. Bar equals 5.0
pm.

Figure 2-25.
Cells of Azospirillum lipoferum Sp RG8c
from nitrogen-free, high phosphate-B-
hydroxybutyrate broth, viewed by
scanning electron microscopy. a) Cells
from 33-day-old culture, pH 7.1. Note
thorough encapsulation of lower cell
layer. 3,000X magnification. Bar
equals 5.0 pm. b) Cells from 9-day-old
culture, pH 7.0. Note holes within the
floe. 1,700X magnification. Bar equals
5.0 pm.

M
130
a
b

131
these cultures at 9 days was 6.9 to 7.1. When viewed by
SEM, cells of A. brasilense strain Cd were usually rods
which had collapsed or shrunken during fixation and dehydra
tion. Cells of A. brasilense strain JM 125A2 usually
appeared as very small ovoids, having the appearance of
starved or stationary phase cells. A small number of cells
of the latter strain were enlarged ovoids and possibly were
encapsulated. Unlike all strains of A. lipoferum tested,
the two A. brasilense strains showed little or no tendency
to clump in HP-BHB, two-step, broth-replacement studies.
One more observation made during HP-BHB, broth-replace
ment studies deserves mention. It was suggested earlier
that some filamentous cells that had undergone septation and
encapsulation might retain cytoplasmic connections between
adjacent cells. Figure 2-26 shows a clump of cells from a
58-day-old, HP-BHB broth culture of A. lipoferum Sp RG6xx.
Several of the cells in the clump appear to be undergoing
plasmolysis, but there appear to be continuities of cyto
plasm between some neighboring, plasmolyzed cells.
Discussion
Only two A. brasilense strains were employed in this
study, and both responded far less uniformly to growth
on BHB than did the A. lipoferum strains. There may have
been poor uptake of BHB by A. brasilense strain JM 125A2.
Cells of this strain that were able to grow on BHB

132
Encapsulated cells of Azospirillum
lipoferum Sp RG6xx from a 50-day-old
nitrogen-free, high phosphate-B-hydroxy-
butyrate broth culture. Note apparent
continuities between cytoplasms of
several cells that appear to have under
gone plasmolysis. 1,500X magnifica
tion. Bar equals 3.0 pm.
Figure 2-26.

133
accumulated PHB and capsular material, and sometimes became
rounded and nonmotile. Better uptake and growth on BHB
occurred with A. brasilense strain Cd. This strain was
notable for its formation of numerous weakly motile fila
ments that sometimes underwent septation and fragmentation
into individual cells. The presence of combined nitrogen
still allowed PHB accumulation and filament formation of
A. brasilense strain Cd, as was also true for A. lipoferum
Sp RG6xx. Both A. brasilense strains showed some tendency
to form encapsulated microflocs arising from one or more
filaments that had undergone septation.
The A. brasilense strains and A. lipoferum Sp RG6xx
both developed scarlet coloration when cultured as lawns on
SNF-Congo Red agar. Sadasivan and Neyra (1985) have shown
that azospirilla can produce cellulose as one component of
their exopolysaccharides. Congo Red is known to stain many
polysaccharides, including cellulose (Nakanishi et al.,
1974), and colonies of azospirilla grown as
dinitrogen-fixers on agar surfaces take up the dye more
avidly than other free-living, dinitrogen-fixing prokaryotes
(Rodriguez Caceres, 1982). It is suggested that Congo Red
may profitably be used more often in cultural studies with
azospirilla to examine conditions promoting capsule forma
tion. This study indicated that nitrogen-free lawns of
A. lipoferum Sp RG6xx produced capsules extensively when
cultured on succinate or BHB, whereas both A. brasilense

134
strains seemed to produce capsules extensively only on
succinate.
All the strains of A. lipoferum used in this study
responded uniformly when cultured as lawns on nitrogen-free,
LP-BHB agar. They appeared to accumulate PHB and grow as
filaments that gradually lost motility. Within 5-7 days,
these filaments had accumulated capsular material and become
septated. By this time the cells sometimes stuck tena
ciously to the agar surface and to glass surfaces as well,
and macroflocs were produced. The response could be said to
be homogeneous in that over 90% of the cells present were
encapsulated microflocs, but the cells themselves were not
morphologically homogeneous. All contained large
accumulations of PHB, but the size and shape of cells
varied.
The most extensive studies were done with A. lipoferum
Sp RG6xx, but its behavior was typical of other A. lipoferum
strains. The appearance of the microflocs did not change
noticeably with time when viewed by SEM or phase-contrast
microscopy. The PHB content of the cells also remained
about the same over time. It would be interesting to see if
these microflocs have lower respiration than comparably
cultured cells of the same age on SNF agar. The fact that
lawns from nitrogen-free, LP-BHB agar consistently gave
CFU counts of the same order of magnitude at different
culture ages for equivalent OD,-60
readings indicates that

135
the cells were not actively dividing after perhaps a week's
time.
Patriquin et al. (1983) observed spherical, bag-like
structures on the surfaces and interiors of 3-week-old and
older wheat roots in axenic association with azospirilla.
Azospirilla containing PHB granules moved actively within
these structures. These structures were similar to the
zoogleal-type microflocs observed in LP-TSS broth cultures
that had passed into stationary phase. It is probable that
these zoogleal forms arose mainly through filamentation,
followed by septation. Also reported in the previous study
were sharply defined, small colonies of azospirilla of
apparently determinate size on the surfaces of wheat roots.
This is a good description of the microflocs of A. lipoferum
from nitrogen-free, LP-BHB lawns. The formation of
filaments (Tarrand et al., 1978) and chains (Ruscoe
et al., 1978) of azospirilla have also been previously
reported.
Most laboratory studies of bacteria use cultures in a
state of balanced growth, where every component of the cell
culture increases at the same rate. This is done for
reproducibility of results and standarization of condi
tions. Cultures that have passed into stationary phase have
experienced a metabolic shift-down, growing more slowly and
with more widely variable characteristics than log phase
cells (Ingraham et al., 1983). It would appear that

136
azospirilla do not become pleomorphic during balanced growth
with combined nitrogen. Pleomorphic forms are only observed
in such cultures in stationary phase (Lamm and Neyra, 1981;
Papen and Werner, 1982; Krieg and Dobereiner, 1984), after a
nutritional down-shift has occurred. Media with high C/N
ratios have more often resulted in pleomorphism of
azospirilla (Tarrand et al., 1978; Papen and Werner, 1982;
Sadasivan and Neyra, 1985). Such conditions generally
promote the formation of PHB (Dawes and Senior, 1973) and
exopolysaccharides (Sutherland, 1977; Costerton et al.,
1981) by bacteria. High C/N ratios may also foster unbal
anced growth of azospirilla, leading to pleomorphism. The
formation of PHB and capsules by azospirilla seems to lead
to a pleomorphic cell type that may reach maturity as a
cyst.
In these experiments, azospirilla inocula were grown in
HP-TSS broth containing combined nitrogen. It is probable
that, in this rich medium, the cells had most of their
biosynthetic operons repressed, so their biosynthetic
enzymes were present only at low levels. After harvest and
washing, these cells were exposed to a new carbon source
(BHB) and deprived of combined nitrogen, forcing the cells
into a severe metabolic shift-down requiring synthesis of
biosynthetic enzymes (Ingraham et al., 1983). The
azospirilla in two-step, replacement cultures were starved
for nitrogen, since azospirilla cannot fix dinitrogen

137
aerobically. The A. lipoferum strains remained quite
healthy in appearance in SEM studies under these condi
tions. Their ready uptake of BHB leading to PHB accumula
tion may have allowed them to retain their cellular
integrity under conditions of nitrogen starvation. The
A. brasilense strains in two-step, broth-replacement
cultures did not produce visible exopolysaccharides in SEM
studies. The collapsed and shrunken appearance of the
A. brasilense strains in SEM studies also suggests that they
were not extensively accumulating PHB.
Following completion of these studies, it was learned
that the high level of phosphates used in HP-BHB, two-step,
broth-replacement studies can inhibit growth of azospirilla
in the absence of combined nitrogen (Scott et al., 1979; Das
and Mishra, 1984). However, the A. lipoferum strains used
in these studies still accumulated PHB and capsules in the
two-step, broth-replacement studies where this buffer was
used. These studies also showed that encapsulation and
pleomorphism can occur at near-neutral pH for some A.
lipoferum strains. The accumulation of PHB occurred on both
HP-BHB and LP-BHB agar. Extensive encapsulation only
occurred on nitrogen-free, LP-BHB agar, where the pH may
have become alkaline. Alkaline pH may prevent pleomorphic
cells from resuming vegetative growth.
The sequence of events for A. lipoferum strains grown
in BHB broth or on BHB agar was PHB accumulation, followed

138
by filamentation and septation. Cells gradually lost their
motility during this time. The septating filaments later
produced extensive capsular material on nitrogen-free,
LP-BHB agar, and sometimes in HP-BHB broth. The HP buffer
repressed extensive capsule formation on nitrogen-free,
HP-BHB agar.
These and previous studies indicate that pleomorphic
cells of azospirilla arise under two different conditions.
After cells have experienced balanced growth with combined
nitrogen, some may become pleomorphic when a nutrient
essential to growth becomes limiting. The production of
both filamentous cells and exopolysaccharides seemed to
occur extensively in HP-TSS broth in stationary phase, but
PHB accumulation was not extensive.
Azospirilla may also become pleomorphic during growth
where the C/N ratio of nutrients available to the cells is
high. Dinitrogen-fixing cells may be poised to become
pleomorphic. It is also evident that azospirilla may take
up some carbon sources more rapidly than they can utilize
combined nitrogen, resulting in extensive PHB accumulation
and capsule formation even when combined nitrogen is avail
able. Cells under these conditions may experience a
temporary shift-down, until enough combined nitrogen can be
assimilated to mobilize their PHB deposits and enable
further growth. This would explain the morphological
changes observed in this study with A. brasilense strain Cd

139
and A. lipoferum Sp RG6xx when they were cultured with BHB
and combined nitrogen.
The environment near plant roots, where azospirilla are
most often found, can be expected to provide available
nutrients having a high C/N ratio (Griffin et al., 1976;
Beck and Gilmour, 1983; Kraffczyk et al., 1984). Depletion
of nitrate from the root zone by physical factors and plant
uptake (Okon and Hardy, 1983), in combination with deni
trification (Smith and Tiedje, 1979), can be expected to
further elevate the C/N ratio of nutrients available to
azospirilla. These bacteria have recently been found to
grow and fix dinitrogen with straw, which again has a high
C/N ratio (Halsall et al., 1985). Azospirilla associated
with plant material can be expected to possess capsules and
PHB, and sometimes to assume pleomorphic cell shapes.
As mentioned before, the most often-reported, pleo
morphic form of azospirilla has been a rounded, nonmotile
cell possessing a capsule and PHB granules. Individual
vegetative cells can likely assume this form without being a
member of a microfloc or microcolony. Such rounded cells
were sometimes observed within microflocs or after they had
broken free of such floes. But the pleomorphic cell types
most often observed were filaments or chains of cells that
eventually formed the microflocs. Bacterial filaments can
arise when cells are growing very rapidly (Slater and
Schaechter, 1974), or when the growth rate shifts down

140
(Jensen and Woolfolk, 1985). Studies reported here indicate
that azospirilla may form them under both conditions.
Strains of A. lipoferum grown in broth containing
combined nitrogen have a greater tendency to clump than do
strains of A. brasilense (Krieg and Dobereiner, 1984). In
this study, the SEM photographs of stationary phase HP-TSS
broth cultures of A. lipoferum Sp RG6xx indicated there is
some structural regularity in clumps. The cells were often
arranged so that the floes contained spaces. The existence
of spaces was more regular and pronounced in floes of cells
cultured with BHB in the absence of combined nitrogen,
probably because the exopolysaccharides of these cells were
more rigid than the exopolysaccharides of stationary phase
cultures grown with combined nitrogen, which tended to be
slimy.
The formation of such microflocs may provide some
advantages for azospirilla in nature. Bergersen (1984) dis
cussed the strategies that microaerophilic, dinitrogen
fixing bacteria such as azospirilla may have to protect
their oxygen-sensitive nitrogenase. His suggestions are
incorporated below into some of the advantages that
azospirilla may find in growing as encapsulated microflocs.
1. The capsules may help to regulate the availability
of oxygen to dinitrogen-fixing cells. Assuming
the capsules to be highly hydrated (Costerton et
al., 1981) and assuming that water reduces the

141
diffusion of oxygen by a factor of 10,000 (Clark,
1967), such a role for capsules is not improb
able. If the capsular material remains pliable,
the cells within an encapsulated microfloc may be
able to move closer together or further apart as
the situation requires. When oxygen is in excess,
they may move closer together, reducing the oxygen
tension in one spot and thus allowing continued
nitrogenase activity at that localized site. When
oxygen is limiting, the cells may move further
apart, the separation allowing each cell's nitro
genase to remain functional.
The spaces between encapsulated cells may provide
sites for other bacteria to enter into intimate
association with the microfloc, to act in
cross-feeding, or to help reduce local oxygen
tension.
Encapsulation would provide the general benefits
to azospirilla that most bacteria seem to derive
from encapsulation (Costerton et al., 1981).
These benefits include protection from predation
and enhanced nutrient accumulation and uptake.
The sustained rigid structure of microflocs sug
gests that nutrients within the encapsulated
microflocs may be sequestered from the surrounding
environment, giving the azospirilla a storehouse

142
of nutrients that may not be readily available to
microbial competitors.
4. The encapsulated microfloc may become a fixed site
where azospirilla are sometimes able to outcompete
motile bacteria for carbon sources. For example,
if such a microfloc became established on a root
surface, it might be able to continually deplete
the carbon supply from that area of root by
assimilating it into PHB, without need for further
cell division to occur immediately. Competing
bacteria without the ability to accumulate PHB
might be limited to growing at sites on the root
where only balanced growth could occur.
5. If a microfloc is faced with starvation for
exogenous carbon sources, some of its members may
serve to feed others. If all septa between cells
in a microfloc are completed, it could be that
these septa are lysed during starvation, so that
the substance of dead cells becomes available to
healthier cells within the microfloc.
6. Most microflocs contained several different cell
morphologies. Septation almost always resulted in
cells of significantly different sizes and
shapes. Thus, cells within a microfloc formed
from a single filament may be destined to attain
different physiological states after septation is

143
completed. Some may be poised to become actively
motile as soon as exogenous nutrients enabling
balanced growth become available. Others may be
primed for continued dinitrogen-fixation, and
still others might enter a truly dormant, cyst
like state. Most microflocs observed contained at
least one rounded cell that may have been
cyst-like. Such a diversity of physiological
states within a microcolony may help ensure the
persistence of the colony at that site, providing
a multiplicity of possible rapid cell responses to
environmental conditions.
Although these studies failed to produce apparent
quantitative, morphological encystment of azospirilla, quan
titative encapsulation and microfloc formation of the
A. lipoferum strains were obtained. Encapsulation seems to
be a prerequisite for encystment in Azotobacter spp. (Eklund
et al., 1966). It may be that encysted forms of azospirilla
could be obtained in quantity by further nutritional
manipulation of the encapsulated cells.

CHAPTER III
PHYSIOLOGICAL PROPERTIES OF ENCAPSULATED FLOCS OF
AZOSPIRILLUM LIPOFERUM Sp RG6xx
In Chapter II, a method was described for quantita
tively converting vegetative, cell-lawn inocula of
A. lipoferum strains into nonmotile, encapsulated cells
having extensive, intracellular PHB deposits. Although many
of the cells within encapsulated floes had vibrioid or ovoid
morphologies, some were rounded and cyst-like in appear
ance. Mature cysts have a lower endogenous metabolic rate
and greater ability than do vegetative cells to survive
carbon starvation (Sudo and Dworkin, 1973). Mature cysts of
Azotobacter spp. are known to be more resistant than
vegetative cells to environmental stresses, including
desiccation (Socolofsy and Wyss, 1962). Such mature cysts
germinate in phosphate buffer containing certain carbon
sources, but not in carbon-free phosphate buffer containing
ammonium, or in unamended phosphate buffer (Loperfido and
Sadoff, 1973). The central bodies of mature cysts are
violently and almost immediately expelled from their cyst
coats when cysts are suspended in Tris-EDTA (Socolofsky and
Wyss, 1961; Goldschmidt and Wyss, 1966; Lin and Sadoff,
1969; Page and Sadoff, 1975).
144

145
It was of interest to determine if the cells of
encapsulated microflocs of A. lipoferum Sp RG6xx were more
resistant to desiccation and carbon starvation than were
motile, vegetative cells. It was also of interest to define
the conditions under which germination occurred, defined as
motile cells arising from nonmotile, encapsulated
microflocs. Microflocs were exposed to treatment that
results in the rupture of capsular coats of mature
Azotobacter spp. cysts. Finally, whether chloramphenicol
inhibited production of motile cells from a nonmotile
encapsulated inoculum was studied. All of these assays
represented attempts to determine if floes of A. lipoferum
Sp RG6xx contained significant numbers of physiologically
cyst-like cells.
Materials and Methods
Bacterial Strain
The only strain used in these studies was Azospirillum
lipoferum Sp RG6xx. Its maintenance and subculture were as
described in Chapter II.
Growth Media and Enumeration
Vegetative cells were cultured as previously described
in the modified complete medium of Tyler et al. (1979),
using the HP buffer of Albrecht and Okon (1980). Encapsu
lated lawns of the bacterium were cultured on nitrogen-free,

146
LP-BHB agar, as described in Chapter II. Plate counts were
performed on the modified SNF agar medium using LP buffer
and Congo Red, as described in Chapter II.
Harvest of Cells
Harvest and washing of vegetative broth cultures always
employed sterile LP-basal salts solution (pH 7.3), as
described in Chapter II. Lawns on LP-BHB agar were
harvested in two ways. When cells were to be added to
semisolid agar or assayed for desiccation and starvation
resistance, they were harvested and washed in sterile LP-TSS
salts solution. For every other assay that was performed,
the cells were harvested and washed in sterile deionized
water.
Desiccation Resistance
Vegetative HP-TSS cultures were grown for 17-22 hours,
attaining an OD^q of 0.5 in each experiment. The cells
were then centrifuged and washed twice in sterile LP-basal
salts solution, as described in Chapter II. The cells were
resuspended in a third volume of LP-basal salts solution to
attain an OD,.^ of 0.3 in one experiment and 0.61 in another
experiment. Then ten-fold dilution series in sterile
LP-basal salts solution were prepared aseptically from the
_ o
cells. A 0.1-ml portion of the 10 dilution was asepti
cally added to each of three autoclaved 10-ml glass beakers,

147
contained within an autoclaved 80-mm X 10-mm Pyrex storage
dish (Fisher Scientific Company, Fair Lawn, NJ). The
beakers had been washed in 3.7% (vol/vol) HC1 for 24 hours,
then rinsed in several changes of deionized water before
autoclaving.
Encapsulated cells grown on nitrogen-free, LP-BHB agar
were harvested and washed in sterile, LP-basal salts solu
tion as described in Chapter II. Cells of 75 days of age
were resuspended to give a final OD,.^ of 0.25 in one
experiment, and cells of 53 days of age were resuspended to
give a final of 0.3 in another experiment. A 0.1-ml
portion of cell suspension was aseptically added to each of
three, 10-ml beakers housed in a storage dish as described
above. Dilution series for plate counts were also prepared
as for the vegetative cells. The storage dishes containing
cells were placed in a glass desiccator over Drierite at
25C for 24 hours, by which time the cell suspensions had
dried onto the glass surfaces. The desiccator was then
placed in a 30C incubator for 8 days.
Initial cell numbers before drying treatment were
enumerated by spread plating. For enumeration of cells
surviving the desiccation treatment, the beakers were
removed from their storage dishes in a laminar flow hood.
The dried cell films were outlined with an ink marker to
help ensure that they would be resuspended. A volume of 2.0
ml of sterile, LP-basal salts solution was aseptically added

148
to each beaker, the cells were resuspended by scraping with
a flamed wire loop, then mixed with a sterile 1.0 ml
pipette. One milliliter from each beaker was used for
ten-fold dilution series in sterile, LP-basal salts solution
for the purpose of spread-plate counts. Of the remaining
volume, 0.1 ml was aseptically pipetted and spread plated
onto agar plates. Four plates were spread for each
dilution.
Carbon Starvation
The same washed LP-TSS cell cultures were used for
these experiments as for the desiccation experiments. Cells
were starved in 50.0-ml, Kimax screw-cap test tubes. The
tubes were autoclaved empty, and 8.0 ml of sterile, LP-basal
salts solution were later added aseptically to each. The pH
of the starvation solution was 7.2. A 2.0-ml volume of each
cell type was then added to each of three tubes. The tubes
were incubated horizontally on a 130 rpm rotary shaker at
30C for 9 days. For enumeration of cells surviving the
starvation treatment, 0.1 ml from each tube was aseptically
spread plated onto each of four SNF-Congo Red plates.
Ten-fold dilution series in sterile LP-TSS salts were also
prepared and enumerated by spread plating.

149
Microaerobic Incubation
The same washed BHB-grown cells were used for these
experiments as for the previous experiments. The incubation
medium consisted of LP-basal salts solution containing
Bacto-Agar (Difco). The agar was added to give a final
concentration of 0.05% (wt/vol) per flask after cell
addition. The basal salts solution and agar were dissolved
by boiling, then 23.5 ml were added per 125-ml Erlenmeyer
flask. The flasks were autoclaved, and concentrated sterile
LP buffer and biotin were added aseptically soon after
autoclaving and before cell addition. A volume of 6.0 ml of
encapsulated cell suspension was added to each flask.
Flasks were prepared in triplicate and incubated in
stationary position at 30C.
Aerobic Incubation
The BHB-grown, encapsulated cell inoculum for this and
all following experiments was harvested, washed twice in
sterile deionized water, and resuspended in a third volume
of sterile deionized water to give a final ODr, of 0.23 to
0.25. Lawns of 58 to 66 days of age were used as inocula.
For the incubation solution, basal salts were dissolvsed in
concentrated amounts to give their final, correct concentra
tions after aseptic additions of LP buffer, biotin, carbon,
or nitrogen sources and cells. The biotin was aseptically
added as a 100X concentrated stock solution, and all other

150
additions to the LP-basal salts solution were added as 10X-
concentrated, sterile stock solutions that were sterilized
by autoclaving.
The sugars employed were D-glucose (Difco), sucrose
(Difco), and D-fructose (Calbiochem, San Diego, CA). All
were prepared as separate 4.37% (wt/vol) stocks in deionized
water. The organic acids were succinic acid (Fisher) and
DL-malic acid (Sigma). Each acid was prepared as a separate
4.37% (wt/vol) salt stock that was neutralized to pH 7.0
with 10 M KOH before being brought to final volume. The
nitrogen sources were reagent grade KNO^ and (NH^^SO^. The
KNO^ was prepared as a separate 0.765% (wt/vol) stock, and
the (NH^)2^0^ was prepared as a separate 0.5% (wt/vol)
stock.
The same type of screw cap tubes used in carbon starva
tion studies were used in these studies. Each empty,
sterile tube had aseptically added to it 8.0 ml of sterile,
LP-basal salts solution plus biotin and a single carbon or
nitrogen source. Sometimes cells were incubated in LP-basal
salts solution with biotin, but without carbon or nitrogen
sources. Then 2.0 ml of water-washed cells were aseptically
added to each tube. Treatments for each carbon or nitrogen
source were done in triplicate. Tubes were incubated
horizontally on a 130 rpm rotary shaker at 30C. The pH of
amended buffer incubations ranged from 7.1 to 7.2.

151
Soil Dialysis Incubation
These cultures were prepared by a modification of the
method of Gonzalez-Lopez and Vela (1981). The soil employed
was Arredondo fine sand (Grossarenic Paleudult, loamy,
siliceous, hyperthemic), obtained from the top 20 cm of a
soil profile. The soil was air dried, then the fraction
that would pass through a 0.250 mm sieve was recovered for
further use.
Spectrapor membrane tubing (Spectrum Medical Indus
tries, Inc., Los Angeles, CA) of 40 mm diameter and 6,000-
8,000 molecular weight cutoff was used. Tubing was cut into
approximately 25 cm long pieces and soaked in a solution of
1.0% (wt/vol) NaHCO^ in which was dissolved 0.005% (wt/vol)
EDTA. The pH of this solution was reduced to 7.5 with 1 M
HC1 before making it up to final volume. The pieces of
dialysis tubing were soaked and then boiled in this solution
for at least 10 min to remove substances that might harm
bacterial cells. The tubing was then rinsed five times in
changes of deionized water and resuspended in a sixth wash.
One end of each piece of tubing was then knotted once.
Five grams of sieved soil were added to each piece of
tubing by inserting a glass funnel in the open end of each
piece and pouring the soil in. The soil was washed down and
evenly distributed in each piece of tubing by adding about
10 ml of deionized water to the open end with a squirt
bottle. The open end of each tube was then knotted once,

152
allowing empty space in each tube to help reduce stress due
to expansion of the tubing upon autoclaving. Each piece of
tubing was then rocked gently and agitated while holding one
end in each hand, to make sure the soil had become
thoroughly wetted. Each piece of tubing was then added to
45 ml of deionized water in a 250 ml Erlenmeyer flask and
autoclaved for 25 min. The pH of the sterile, equilibrated
soil solution surrounding the sterile, intact tubing was
6.1. For cell incubations, 0.5 ml of 100X sterile biotin
was aseptically added, followed by 5 ml of water-washed cell
suspension. Triplicate flasks were incubated at 130 rpm at
30C for 24 to 48 hours. The friction between fine soil
particles and dialysis tubing caused bags of soil to break
during longer shaken incubations.
Tris-EDTA Treatment
A solution of 30 mM EDTA dissolved in 0.05 M Tris-HCl
was prepared, and its pH adjusted to 8.4 with 10 M KOH. A
solution of 0.05 M Tris-HCl was also prepared and similarly
adjusted to the same pH. Both solutions were sterilized by
autoclaving. For lysis experiments, 4 ml of either solution
were aseptically added to 4 ml of water-washed, encapsulated
cells in 50-ml screw cap tubes, to give a final concentra
tion of 0.025 M Tris-HCl alone or in combination with 15 mM
EDTA. Tubes were prepared in triplicate and incubated as
for the other tube assays.

153
Chloramphenicol Treatment
A 0.2% (wt/vol) solution of chloramphenicol (Sigma) in
deionized water was sterilized by passage through a 0.2-pm
pore size, Nalgene filter unit. Various volumes of this
stock solution were added to water-washed, encapsulated
cells in sterile Nutrient Broth (Difco), having a final
concentration after all additions of 0.8% (wt/vol). Usually
2.0 ml of cells were added to these tubes. Incubation was
in 50-ml screw cap tubes, under the same conditions for
tubes as described above.
Phase-Contrast Microscopy and Photographs
These were the same as in Chapter II.
Results
Desiccation Resistance
Neither the vegetative nor the encapsulated cells
displayed significant resistance to the drying method
employed. Virtually all cells lost viability during the 8
days of desiccation. Table 3-1 gives the results of two
separate desiccation experiments. No statistical analyses
were performed, because even if some difference could be
revealed between vegetative and encapsulated cells, the
survival of either was so poor as to be negligible.

154
Table 3-1. Desiccation resistance
RG6xx.
of Azospirillum lipof
erum Sp
Experi- Cell
ment type
Number
Initial.
CFU ml 1
(X 10b)
drieda
Beaker CFU ml ^
1 2
final %
survivors0
3
1 Vege-
0.72
0.0
22.5
5.0
0.03
tative
(0.03)
(0.0)C
(13.0)
(5.0)
Encapsu-
0.42
92.5
55.0
e
0.35
latedd
(0.06)
(37.7)
(28.7)
2
Vege-
0.14
10.0
0.0
0.0
0.05
tative
(0.03)
( 7.1)
(0.0)
(0.0)
Encapsu-
1.34
2.5
25.0
40.0
0.03
lated£
(0.10)
(4.3)
(43.0)
(17.3)
aAverages of colony forming units (CFU) of four spread plates.
Values in parentheses are standard deviations.
Average of all beakers available.
c
Averages of four spread plates. Values in parentheses are
standard deviations.
^75-day-old cells.
0
Beaker was lost.
f53-day-old cells.

155
Starvation Resistance
After 9 days of starvation in carbon-and-nitrogen-free,
LP-basal salts solution lacking biotin, the encapsulated
\
cells retained 24 to 27% viability. On the other hand,
vegetative cells multiplied several-fold and retained
viability (Table 3-2). It was interesting that, although
two different initial densities of vegetative inocula were
used, each seemed to stabilize at about 10 CFU/ml after 9
days of starvation.
Germination Experiments
Table 3-3 gives a summary of the germination
experiments involving semisolid agar, buffered salts
solution containing single carbon or nitrogen sources, and
soil dialysis flasks. The soil dialysis flasks and
combined-nitrogen incubations all resulted in germination of
encapsulated cells. Within 18 hours, the majority of cells
in these treatments were extremely motile, vegetative
cells. By this time the motile cells in soil dialysis
flasks had lost most or all visible deposits of phase-bright
PHB granules. The PHB granules were usually still visible
in the nitrate and ammonium treatments after 24 hours, but
were markedly reduced in size from those in the initial
inoculum. Within 72 hours the extremely motile cells in the
combined-nitrogen treatments assumed the totally phase-dark
appearance of motile cells in soil dialysate. Figure 3-1

156
Table 3-2
. Starvation resistance
Sp RG6xx.
of Azospirillum
lipoferum
Experi
ment
Number
Cell
type
Initial.
CFU ml 1
(X 106)a
Final .
CFU ml 1
(X 106)b
% Initial
CFU ml 1
1
Vege
tative
0.72
(0.03)
1.02 (0.26)
142.04
Encapsu
lated0
0.42
(0.06)
0.11 (0.01)
26.61
2
Vege
tative
0.14
(0.03)
1.02 (0.22)
751.23
Encapsu
lated0
1.34
(0.10)
0.33 (0.18)
24.22
aAverages
plates.
of colony
Values in
forming units
parentheses are
(CFU) of four spread
standard deviations.
Averages of three starvation tubes, with four spread plate
counts per tube. Values in parentheses are standard
deviations.
C75-day-old cells.
53-day-old cells.

157
Table 3-3. Response of encapsulated cells of Azospirillum
lipoferum Sp RG6xx to various incubations.
Treatment
Germination3
Soil dialysate
+
N3~
+
nh4+
+
Glucose
-
Fructose
-
Sucrose
-
Malate
-
Succinate
-
Aerobic, low phosphate, basal salts
solution
-
Microaerobic, low phosphate, basal
salts solution
-
+ denotes the majority of cells present becoming motile
and depleting their visible poly-8-hydroxybutyrate reserves
within 29 hours.

Figure 3-1. Encapsulated cells of Azospirillum
lipoferum Sp RG6xx that have undergone
germination in low phosphate-basal salts
solution with combined nitrogen.
a) Cells from a 29-hour nitrate
incubation. Note germinated vegetative
cell at left of an empty capsule.
1,000X magnification. Bar equals 4.0 pm
b) Cells from a 29-hour ammonium
incubation. Note germinated vegetative
cells, empty capsules and poly-6-hydroxy
butyrate-rich cells remaining in the
floe. 1,000X magnification. Bar equals
4.0 pm.

159
^ r
* #
-M,v2V
A -jJ*
*
*
v: **
!&* **%
vWk^ CLf /*
< % y> + A
*£ *'*# *
* * *

160
shows cells from 29-hour incubations in nitrate (Figure
3-la) and ammonium (Figure 3-lb), and Figure 3-2 shows cells
from a 29-hour incubation in soil dialysate. As is apparent
in these photographs, most encapsulated floes had
germinated, retaining their general shape after most of
their cells had left them. The pH of inoculated soil
dialysis flasks rose to 6.4 to 6.5 within 24 hours after
inoculation.
Encapsulated floes suspended in buffered-salts solution
produced a few motile, ovoid to peanut-shaped cells that
remained swollen with intracellular PHB deposits. This was
also true in the semisolid agar flasks and in the buffered-
salts solution tubes containing single carbon sources.
Sometimes these PHB-rich cells were as rapidly motile as the
cells that germinated with combined nitrogen and soil
dialysate. Usually, however, they moved slowly and were
prone to long periods of twiddling before they actively
moved off on a run. Although these treatments often
resulted in large numbers of free, nonencapsulated cells,
most of these individual cells were not motile. The origi
nal inocula did not contain even weakly motile cells and
contained few individual cells. Figure 3-3 shows cells from
a 29-hour incubation in buffered-salts solution containing
glucose, an incubation where the floes remained largely
occupied with cells.

Figure 3-2. Encapsulated cells of Azospirillum
lipoferum Sp RG6xx that have undergone
germination in soil dialysis flasks. a)
Cells from a 29-hour-old incubation. Note
empty horseshoe-shaped capsules and phase-
dark vegetative cells. 1,000X magnifi
cation. Bar equals 4.0 pm. b) Other
cells from a 29-hour incubation. Note
empty horseshoe-shaped capsules, phase-
dark vegetative cells and poly-6-hydroxy-
butyrate-rich cells remaining in the
floe. 1,000X magnification. Bar equals
4.0 pm.

162
a

Figure 3-3.
Encapsulated cells of Azospirillum
lipoferum Sp RG6xx that have not undergone
widespread germination in low phosphate-
basal salts solution with glucose,
a) Cells froma 29-hour incubation. A few
empty capsules are present, but most
capsules in the microfloc still retain
their cells. 1,000X magnification.
Bar equals 4.0 pm. b) Microfloc
from a 29-hour incubation with several
empty capsules. Note rounded cyst-like
appearance of some cells remaining in
the floe. 1,000X magnification. Bar
equals 4.0 pm.

164
b

165
After several weeks of incubation, some morphological
changes were observed for cells in the treatments that were
scored as negative at 29 hours. Floes that were incubated
in aerobic, LP-basal-salts solution for 60-65 days had
somewhat the same appearance as germinated floes, in that
most capsular spaces appeared empty or contained cells of
unhearthy appearance. Less than 5% of the capsules in this
extended incubation contained PHB-rich cells. The cells
apparently had not germinated, since no individual vegeta
tive cells were observed outside of the floes. Cells
apparently had depleted their visible PHB reserves and
starved in place within their capsules. The addition of
carbon sources seemed to reduce the extensive loss of
visible PHB deposits. The number of cells within a floe
that retained visible PHB deposits and an overall viable
appearance after 47 days of incubation ranged from about 5
to 20% in aerobic, carbon-amended incubations. Often these
cells appeared somewhat reduced in size, having contracted
from the capsule boundary while retaining their rounded
appearance. A few individual, motile, PHB-rich cells were
also present in carbon-amended incubations at this time.
The nitrate and ammonium incubations at 47 days consisted of
empty capsules and nonmotile, phase-dark vegetative cells
which had the appearance of stationary phase cells. After
60 to 65 days, the microaerobic incubations were similar in
appearance to the aerobic, carbon-amended incubations, with

166
many PHB-rich cells still contained within capsules. Unlike
all the other extended incubations, only the microaerobic
treatment contained numerous, motile, phase-dark vegetative
cells.
Tris-EDTA Treatment
When encapsulated floes were incubated in Tris-EDTA
solution, the only observable effect was dissolution of
macroscopic floes within 24 hours. To the naked eye, the
cell suspensions appeared more evenly turbid and nonfloc-
culated than those from any other treatment. When viewed by
microscopy, however, microflocs were still present. They
were of about the same size range as microflocs in other
treatments, and had about as many cells remaining in their
capsules. Few motile cells were observed in this treat
ment. There were not noticeably fewer cells within capsules
than in the other treatments where overall germination did
not occur. The same results were obtained when encapsulated
floes were incubated in Tris alone.
Effect of Chloramphenicol
The addition of 50 yg chloramphenicol/ml to nutrient
broth prevented growth for at least 36 hours, whereas
nutrient broth tubes without the antibiotic became turbid
within 18 hours with the same inoculum level. Even in the
presence of chloramphenicol, however, a few weakly motile,

167
PHB-rich cells were observed. Each was rotating about its
own long axis without moving off on runs.
Discussion
The desiccation resistance assay used in these studies
involved rapid drying of the cells. Mature, encysted forms
of some prokaryotes are much better able to withstand drying
on glass surfaces than their vegetative counterparts
(Whittenbury et al., 1970a; Tudor and Conti, 1977). Filter
membranes are often used as a surface upon which cells are
slowly dried in desiccation resistance assays (Socolofsky
and Wyss, 1962). In such experiments (data not shown),
vegetative cells of azospirilla sometimes survive slow
drying on membrane filters without appreciable die off, as
do encapsulated cells. Webb (1965) has pointed out the dif
ficulties of using membrane filters in such assays. It was
thought that glass surfaces would be easier to use, with
less inherent hydrophilic behavior than membrane filters.
Rapid drying of cells within a day's time or less usually
causes a rapid and nearly complete die off of vegetative
cells for many genera (Robinson et al., 1965; Whittenbury et
al. 1970a; Tudor and Conti, 1977 ). The results of the work
reported here show no apparent significant difference in
response to rapid drying between vegetative or encapsulated
cells of A. lipoferum Sp RG6xx. This is in agreement with
studies where capsules have not enhanced the ability of

168
cells to survive desiccation (Dudman, 1977; Lowendorf,
1980).
Other studies have found enhanced desiccation resis
tance for some encapsulated forms of azospirilla. Such
forms have been called cysts (Lamm and Neyra, 1981; Papen
and Werner, 1982; Sadasivan and Neyra, 1985). In the first
study, however, cyst-like forms were slowly dried while
vegetative cells were rapidly dried, so a true comparison of
desiccation resistance between forms seems questionable. In
the last study, floes of cells were stored without desiccant
for long time periods, and some cells within pieces of floes
were shown to remain viable for up to six months. But
whether this reflects a true desiccation resistance of
single cells, or only an ability to withstand drought in the
presence of hydrophilic polymers (Jagnow, 1982), is
unclear. Perhaps the best studies were done by Papen and
Werner (1982), where vegetative cells were found to almost
completely die out rapidly on dried membrane filters,
whereas encapsulated, nonmotile cells survived this treat
ment in high numbers.
Azospirilla may undergo extensive morphological
changes, including accumulation of PHB and capsular
material, as only a first step towards becoming mature
cysts. Morphologically cyst-like azospirilla have exhibited
variable responses to desiccation (Lamm and Neyra, 1981;
Papen and Werner, 1982). In the studies reported here, many

169
cells in microflocs had a cyst-like appearance. However,
cells in microflocs survived desiccation no better than
nonencapsulated vegetative cells. Encysting bacteria may
gradually mature, with their resistance to environmental
stress increasing with time although their morphology
appears cyst-like throughout. Modification of the cell
membrane may be the true key to stress-resistance of these
forms, as may also be true of Azotobacter spp. cysts (Reusch
and Sadoff, 1983). Immature cysts of azospirilla may mature
best if they are removed from the growth medium and/or
slowly dried. Sadasivan and Neyra (1985) removed their
encapsulated azospirilla from broth and dried the floes
slowly, and possibly obtained desicccation resistant forms.
Similarly, Lamm and Neyra (1981) obtained desiccation-
resistant azospirilla by allowing the agar of lawns to dry
slowly.
There was a significant difference in response to
carbon and nitrogen starvation between vegetative and encap
sulated cells of A. lipoferum Sp RG6xx. After 9 days, only
about 25% of the original encapsulated inoculum retained
viability. The cell densities used in these starvation
6
experiments were less than 10 CFU/ml, so microscope
observations were not performed. However, the microscope
observations of extended germination incubations with higher
cell densities indicated that the majority of PHB-rich cells
within floes will eventually deplete their PHB reserves and

170
apparently lose viability due to starvation for exogenous
carbon. Most encapsulated cells within floes do not become
motile when confronted with starvation for exogenous
carbon. From 5 to 25% of cells within floes retained their
PHB after extended incubation and often underwent reduction
in size within their capsules. One interpretation of this
size reduction is that the cells were undergoing maturation
into physiologically mature cysts. In contrast to the
encapsulated cells, two different densities of vegetative
inoculum increased several fold during the same 9 days of
£
starvation to give about 10 CFU/ml. This apparent multi
plication to a certain cell density and continued viability
of vegetative cells faced with starvation have a precedent
in studies of other bacteria, such as Rhizobium japonicum
(Crist et al., 1984).
In a recent study, Tal and Okon (1985) reported that
vegetative, PHB-poor cells of A. brasilense strain Cd died
off to about 7% of their initial numbers after 130 hours of
starvation in sterile, 0.06 M potassium phosphate buffer.
In comparison, PHB-rich cells proliferated 2.3-fold over the
same time span. Their experimental conditions differed from
conditions reported here not only in using a different
bacterial species but also in incubating the cells in
phosphate buffer alone, without other salts. Both of their
cell types were apparently nonencapsulated and actively
motile, while the encapsulated cells in this study were not

171
motile. There are numerous starvation conditions that
bacteria can be exposed to in vitro, and the response of
different strains can vary widely under different condi
tions. It would be interesting to follow up on these
initial studies of starvation resistance of azospirilla, to
gain further insight into how they might survive in soil in
the absence of plant material. It is possible that some
azospirilla are able to enter into two types of dormancy
(Marshall, 1980) in unfavorable soil conditions. If the
cells have been experiencing balanced growth before they are
starved of exogenous carbon, they might enter into exogenous
dormancy. Such cells would be poor in PHB and might have no
different morphology than growing vegetative cells, but
their metabolism would be greatly reduced. If the cells
have accumulated large amounts of PHB through dinitrogen
fixation or extremely rapid uptake of carbon sources during
growth on combined nitrogen, they might be prone to enter
constitutive dormancy or an encysted state when faced with
starvation.
Papen and Werner (1982) suggested that depletion of
available oxygen was responsible in part for encystment of
azospirilla in their studies. The low oxygen consumption of
encysted cells may have allowed oxygen to diffuse back into
the medium from the headspace, whereupon vegetative cells
emerged from the capsular coats and resumed dinitrogen
fixation. Because of their suggestion, in this study

172
encapsulated floes were incubated microaerobically in semi
solid agar, as well as aerobically in shaken broth tubes, to
give the cells different oxygen regimes. There was no
observable difference in response between floes incubated
microaerobically or aerobically in buffered salts solution,
with or without single carbon sources. In each of these
treatments, a few weakly motile, PHB-rich cells were
observed within 18 hours, and they persisted for up to 10
days. Half of the capsular spaces in some floes were empty,
but most floes retained the majority of their cells within
capsules. Unlike mature Azotobacter spp. cysts, encapsu
lated cells of this strain did not become synchronously
motile when exposed to carbon sources (Loperfido and Sadoff,
1973). Phosphate-buffered-salts solutions and buffered-
salts solutions containing sucrose, a nonmetabolizable
carbon source for this species (Tarrand et al., 1978),
produced about as many motile cells as did metabolizable
carbon sources. Like mature Azotobacter spp. cysts, encap
sulated cells of this strain seemed unable to mobilize their
PHB reserves in unamended buffered salts solution to enable
germination and widespread motility (Loperfido and Sadoff,
1973). Either wetting released cells from floes and these
cells became motile, or else the cells became motile and
actively left floes after wetting. Perhaps both events
occurred. In any case, most individual cells were not
motile in these incubations. The semisolid agar flasks

173
contained enough cells in the initial inoculum to be visibly
turbid, but the cells remained dispersed throughout the agar
and formed no pellicle. As a result, these treatments are
listed as giving no germination (Table 3-3). A few cells
seemed able to slowly mobilize their PHB reserves and become
motile under these conditions, but the majority of cells
remained in the floe, or once free from the floe, remained
nonmotile. Motile cells continued to retain extensive
visible deposits of PHB.
Undeniable germination occurred when the floes of
encapsulated cells were added to soil dialysis flasks, or to
buffered-salts solution containing nitrate or ammonium. The
uniformity of response among these treatments indicates that
combined nitrogen in the soil dialysate was responsible for
its germination effect. It also indicates that most of the
cells in the floes were not similar to mature Azotobacter
spp. cysts, which do not germinate in the presence of
ammonium (Loperfido and Sadoff, 1973). The availability of
combined nitrogen apparently prompted most of the cells in
encapsulated floes to mobilize their PHB reserves and return
to an actively motile, vegetative state.
An interesting feature of these positive germination
treatments was the persistence of nonmotile, PHB-rich cells
within floes even after 10 days of incubation. Some of
these cells no longer possessed a plump appearance, and
their PHB granules were dispersed irregularly within the

174
cytoplasm. Sometimes capsules seemed to contain PHB and
polyphosphate granules without any cytoplasm. Empty
capsules, however, often possessed the "horseshoe" shape
typical of empty exines of mature Azotobacter spp. cysts
(Lin and Sadoff, 1969).
The Tris-EDTA incubations did not produce any obvious
expulsion of cells from capsules. The concentration of EDTA
was about five times that which produces prompt expulsion of
central bodies from mature Azotobacter spp. cysts (Lin and
Sadoff, 1969). As mentioned earlier, however, the
macroscopic appearance of the inocula was rendered more
evenly turbid by this treatment. Incubation in Tris buffer
alone had the same effect as Tris-EDTA. The high pH of the
treatments (pH 8.4) may have been related to the dispersive
effect, along with the chelating effects of the Tris and the
EDTA.
It seems that protein synthesis is necessary before
encapsulated cells are able to become motile in large
numbers. The chloramphenicol treatment did not prevent some
free cells from spinning about their own long axes, however.
Based on these tests, there appears to be little
physiological similarity between mature cysts of Azotobacter
spp. and most cells in the encapsulated floes of A.
lipoferum Sp RG6xx. Most of the cells in encapsulated floes
represent immature cysts, lacking desiccation resistance,
but being largely nonmotile and unable to readily mobilize

175
their PHB reserves unless exogenous, combined nitrogen
becomes available. If these cells are found to have a very
low endogenous respiratory rate, it might further indicate
their state as nascent cysts.
Living cells may form structures that prove immediately
useful for some functions. By chance these structures may
also prove beneficial to the cells in other ways (Cairns-
Smith, 1982). It is suggested that the microflocs of azo
spirilla are such structures. Their possible benefits were
suggested in Chapter II. Four observations in this chapter
deserve further comment. One is the great size difference
between motile, vegetative cells of A. lipoferum Sp RG6xx
and nonmotile, encapsulated cells. The encapsulated cells
occupy much more volume. Secondly, Costerton et al. (1981)
suggested that most bacterial cells in nature assume two
forms. Sessile forms surrounded by a capsule maintain a
population on a surface and give rise to motile swarmer
cells which colonize new surfaces. This is a good descrip
tion of the conversion of encapsulated to motile forms of
A. lipoferum Sp RG6xx. Thirdly, the ability of only 5 to
20% of cells within encapsulated floes to retain their
visible PHB deposits over 65 days of aerobic nitrogen
starvation may indicate the physiological diversity of cells
within an encapsulated PHB-rich microfloc. Most cells may
be poised to become motile or resume vegetative growth and
may represent the cells that depleted their PHB reserves

176
during starvation and which finally lost their viability.
The remaining cells within the microfloc may include cells
that eventually mature into cysts. Finally, the similar
"horseshoe" appearance of some empty capsules of azospirilla
and of germinated cysts of Azotobacter spp. may have some
importance. More than one type of capsule may exist within
an encapsulated microfloc of azospirilla, and some capsules
may have proceeded further toward a cyst-coat composition
than others.
A few encapsulated cells in floes survived the desicca
tion treatment. A few cells within floes also had the
appearance of rounded, possibly mature cysts. Sometimes
they broke free of floes (Figure 3-4). They were never
observed to be motile. If these are truly mature cysts, the
problem remaining is how to convert most of the cells in a
vegetative inoculum quantitatively into this form.

177
Figure 3-4. Rounded, possibly cyst-like cells of
Azospirillum lipoferum Sp RG6xx, from a
29-hour incubation in low phosphate-basal
salts solution with glucose. 1,000X
magnification. Bar equals 4.0 pm.

CHAPTER IV
GENERAL CONCLUSIONS
The strains of azospirilla used in this study did not
undergo quantitative morphological encystment when grown on
nitrogen-free BHB agar. The strains of Azospirillum
lipoferum synthesized exopolysaccharides more extensively
than did the strains of Azospirillum brasilense. The
A. lipoferum strains experienced unbalanced growth under
these cultural conditions. They accumulated PHB and experi
enced unbalanced cell wall synthesis, as evidenced by the
common formation of filaments and chains. Eventually the
filaments or chains lost motility and accumulated capsular
material. The final outcome was the formation of microflocs
of encapsulated, PHB-rich cells that often arose from only a
few elongated cells. Some of the cells within these floes
had a cyst-like morphology. Environments with available
nutrients having a high C/N ratio, such as the rhizosphere,
may promote the formation of PHB and capsules by
azospirilla. Some cells having these features may
eventually form cysts. Cells in encapsulated microflocs may
have some survival advantages that individual cells of
azospirilla lack.
178

179
Encapsulated cells of Azospirillum lipoferum Sp RG6xx
grown on nitrogen-free BHB agar were found to have far more
PHB than cells grown in broth with combined nitrogen.
Neither cell type displayed significant desiccation
resistance. When faced with aerobic starvation for exo
genous carbon and nitrogen, encapsulated cells of this
strain died off after 9 days to about 25% of their original
numbers. These survivors may have represented cells that
were maturing into cysts. Vegetative cells grown with
combined nitrogen multiplied several fold over the same
period of starvation. This indicates that cells of this
strain may not need to form cysts in nature to survive
prolonged periods of starvation. Vegetative cells having
reduced metabolic activity may survive such periods.
Combined nitrogen promoted germination of nonmotile,
encapsulated cells of this strain.
Although these studies failed to obtain cysts of
azospirilla in high quantity, they may have provided some
information of practical importance. The starvation studies
suggested that encapsulated, PHB-rich cells of azospirilla
are less active physiologically than motile, vegetative
cells. Further studies might measure nitrogenase activity
and plant growth substance production by azospirilla in
relation to PHB deposition and capsule formation. Such
studies should lead to an understanding of what physiologi
cal form of azospirilla is most beneficial to plant growth.

180
Inoculum production may then be designed to introduce and
sustain the most beneficial physiological form in the root
zone.

BIBLIOGRAPHY
Abd-El-Malek, Y., and Y. Z. Ishac. 1966. Longevity of
Azotobacter. Plant and Soil 24:325-327.
Albrecht, S. L., and Y. Okon. 1980. Cultures of
Azospirillum, p. 740-479. Iri Methods in enzymology,
vol. 69. Academic Press, Inc., New York.
Bae, H. C., E. H. Cota-Robles, and L. E. Casida, Jr. 1972.
Microflora of soil as viewed by transmission electron
microscopy. Appl. Microbiol. 23:637-648.
Barak, R., I. Nur, Y. Okon, and Y. Henis. 1982. Aerotactic
response of Azospirillum brasilense. J. Bacteriol.
152:643-649.
Beck, S. M., and C. M. Gilmour. 1983. Role of wheat root
exudates in associative nitrogen fixation. Soil
Biol. Biochem. 15:33-38.
Becking, J. H. 1982. Azospirillum lipoferum--A
reappraisal, p. 127-146. In W. Klingmuller (ed.),
Azospirillum: Genetics, physiology, ecology.
Experientia Supplementum, vol. 42. Birkhauser Verlag,
Basel, Switzerland.
Begg, K. J., and W. D. Donachie. 1985. Cell shape and
division in Escherichia coli: Experiments with shape
and division mutants. J. Bacteriol. 163:615-622.
Berg, R. H., V. Vasil, and I. K. Vasil. 1979. The biology
of Azospirillum-sugarcane association II. Ultrastruc
ture. Protoplasma 101:143-163.
Berg, R. H., M. E. Tyler, N. J. Novick, V. Vasil, and
I. K. Vasil. 1980. Biology of AzospiriHum-sugarcane
association: Enhancement of nitrogenase activity.
Appl. Environ. Microbiol. 39:642-649.
Bergersen, F. J. 1984. Oxygen and the physiology of
diazotrophic microorganisms, p. 171-180. In C. Veeger
and W. E. Newton (ed.), Advances in nitrogen fixation
research. Nijhoff/Junk, Pudoc, Wageningen, The
Netherlands.
181

182
Bothe, H., B. Klein, M. P. Stephan, and J. Dobereiner.
1981. Transformations of inorganic nitrogen by
Azospirillum spp. Arch. Microbiol. 130:96-100.
Cagle, G. D., and G. R. Vela. 1974. Giant cysts and cysts
with multiple central bodies in Azotobacter
vinelandii. J. Bacteriol. 107:315-319.
Cairns-Smith, A. G. 1982. Genetic takeover and the mineral
origins of life, p. 107-109. Cambridge University
Press, Cambridge.
Clark, F. E. 1967. Bacteria in soil, p. 15-49. In
A. Burgess and F. Raw (ed.), Soil biology. Academic
Press, New York.
Costerton, J. W., J. M. Ingram, and K.-J. Cheng. 1974.
Structure and function of the cell envelope of Gram
negative bacteria. Bacteriol. Rev. 38:87-110.
Costerton, J. W., R. T. Irvin, and K.-J. Cheng. 1981. The
bacterial glycocalyx in nature and disease. Ann. Rev.
Microbiol. 35:299-324.
Crist, D. K., R. E. Wyza, K. M. Mills, W. D. Bauer, and
W. R. Evans. 1984. Preservation of Rhizobium
viability and symbiotic infectivity by suspension in
water. Appl. Environ. Microbiol. 47:895-900.
Das, A., and A. K. Mishra. 1984. Aerotolerant growth in
Azospirillum brasilense induced by dihydroxyphenyl
iron-binding compound. Curr. Microbiol. 11:313-136.
Dawes, E. A. 1976. Endogenous metabolism and the survival
of starved prokaryotes, p. 19-53. In T. R. G. Gray and
J. R. Postgate (ed.), The survival of vegetative
microbes. Twenty-sixth Symposium of the Society for
General Microbiology. Cambridge University Press,
Cambridge, London.
Dawes, E. A. 1981. Carbon metabolism, p. 1-38. I_n P. H.
Calcott (Ed.), Continuous cultures of cells, vol. II.
CRC Press, Inc., Boca Raton, FL.
Dawes, E. A., and P. J. Senior. 1973. The role and regula
tion of energy reserve polymers in microorganisms,
P- 135-266. In A. H. Rose and D. W. Tempest (ed.),
Advances in microbial physiology, vol. 10. Academic
Press, New York.

183
Day, J. M., and J. Dobereiner. 1976. Physiological aspects
of ^-fixation by a Spirillum from Diqitaria roots.
Soil Biol. Biochem. 8:45-40.
Dobereiner, J. 1978. Influence of environmental factors on
the occurrence of Spirillum lipoferum in soils and
roots. In U. Granhall (ed.), Environmental Role of
Nitrogen-fixing Blue-green Algae and Asymbiotic
Bacteria. Ecol. Bull. (Stockholm) 26:343-352.
Dobereiner, J., I. E. Marriel, and M. Nery. 1976. Ecologi
cal distribution of Spirillum lipoferum Beijerinck.
Can. J. Microbiol. 22:1464-1473.
Donachie, W. D., K. J. Begg, and N. F. Sullivan. 1984.
Morphogenes of Escherichia coli, p. 27-62. In
R. Losick and L. Shapiro (ed.), Microbial development.
Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
Dudman, W. F. 1977. The role of surface polysaccharides in
natural environments, p. 357-414. In I. W. Sutherland
(ed.), Surface carbohydrates of the prokaryotic cell.
Academic Press, New York.
Duguid, J. P., and J. F. Wilkinson. 1961. Environmentally
induced changes in bacterial morphology, p. 69-99. In
G. G. Meynell and H. Gooder (ed.), Microbial reaction
to environment. Eleventh Symposium of the Society for
General Microbiology. Cambridge University Press,
Cambridge.
Eklund, C., L. M. Pope, and 0. Wyss. 1966. Relationship of
encapsulation and encystment in Azotobacter. J.
Bacteriol. 92:1828-1830.
Eskew, D. L., D. D. Focht, and I. P. Ting. 1977. Nitrogen
fixation, denitrification, and pleomorphic growth in a
highly pigmented Spirillum lipoferum. Appl. Environ.
Microbiol. 34:582-585.
Falk, E. C., J. Dobereiner, J. L. Johnson, and N. R. Krieg.
1985. Deoxyribonucleic acid homology of Azospirillum
amazonense Magalhaes et al. 1984 and emendation of the
description of the genus Azospirillum. Int. J. Syst.
Bacteriol. 35:117-118.
Foster, R. C., and G. D. Bowen. 1982. Plant surfaces and
bacterial growth: The rhizosphere and rhizoplane, p.
159-185. In M. S. Mount and G. H. Lacy (ed.), The
phytopathogenic prokaryotes, vol. I. Academic Press,
Inc., New York.

184
Geesey, G. G. 1982. Microbial exopolymers: Ecological and
economic considerations. Am. Soc. Microbiol. News
48:9-14.
Goldschmidt, M. C., and 0. Wyss. 1966. Chelation effects
on Azotobacter cells and cysts. J. Bacteriol.
91:120-124.
Gonzalez-Lopez, J., M. V. Martinez-Toledo, J. Moreno,
F. Ballesteros, and A. Ramos-Cormezana. 1985.
Resistant properties of Azotobacter cysts induced in
phosphate-limited media. Microbios Letters 28:151-161.
Gonzalez-Lopez, J., and G. R. Vela. 1981. True morphology
of the Azotobacteraceae--Filterable bacteria. Nature
289:588-590.
Gray, T. R. G., and S. T. Williams. 1971. Microbial pro
ductivity in soil, p. 255-287. In D. E. Hughes and
A. H. Rose (ed.), Microbes and biological productiv
ity. Twenty-first Symposium of the Society for General
Microbiology. Cambridge University Press, Cambridge.
Griffin, G. J., M. G. Hale, and F. J. Shay. 1976. Nature
and quantity of sloughed organic matter produced by
roots of axenic peanut plants. Soil Biol. Biochem.
8:29-32.
Halsall, D. M., G. L. Turner, and A. H. Gibson. 1985.
Straw and xylan utilization by pure cultures of
nitrogen-fixing Azospirillum spp. Appl. Environ.
Microbiol. 49:423-428.
Hitchins, V. M., and H. L. Sadoff. 1970. Morphogenesis of
cysts in Azotobacter vinelandii. J. Bacteriol. 104:
492-498.
Hitchins, V. M., and H. L. Sadoff. 1973. Sequential
metabolic events during encystment of Azotobacter
vinelandii. J. Bacteriol. 113:1273-1279.
Hughes, W. H. 1956. The structure and development of the
induced long forms of bacteria, p. 341-362. In
E. T. C. Spooner and B. A. D. Stocker (ed.), Bacterial
anatomy. Sixth Symposium of the Society for General
Microbiology. Cambridge University Press, Cambridge.
Ingraham, J. L., 0. Maaloe, and F. C. Neidhardt. 1983.
Growth of the bacterial cell. Sinauer Associates,
Inc. Publishers, Sunderland, MA.

185
Jagnow, G. 1982. Growth and survival of Azospirillum
lipoferum in soil and rhizospheres as influenced by
ecological stress conditions, p. 97-104. In
W. Klingmuller (ed.), Azospirillum: Genetics,
physiology, ecology. Experientia Supplementum,
vol. 42. Birkhauser Verlag, Basel, Switzerland.
Jain, D. K., and D. G. Patriquin. 1985. Characterization
of a substance produced by Azospirillum which causes
branching of wheat root hairs. Can. J. Microbiol.
31:206-210.
Jansen van Rensburg, H., and B. W. Strijdom. 1980.
Survival of fast- and slow-growing Rhizobium spp. under
conditions of relatively mild desiccation. Soil Biol.
Biochem. 12:353-356.
Jarman, T. R., and G. W. Pace. 1984. Energy requirements
for microbial exopolysaccharide synthesis. Arch.
Microbiol. 137:231-235.
Jensen, R. H., and C. A. Woolfolk. 1985. Formation of
filaments by Pseudomonas putida. Appl. Environ.
Microbiol. 50:364-372.
Kapulnik, Y., R. Gafny, and Y. Okon. 1985. Effect of
Azospirillum spp. inoculation on root development and
NO, uptake in wheat (Triticum aestivum cv. Miriam) in
hydroponic systems. Can. J. Bot. 63:627-631.
Karnovsky, M. J. 1965. A formaldehyde-glutaraldehyde
fixative of high osmolarity for use in electron
microscopy. J. Cell. Biol. 27; (2, Pt. 2):137-138a.
Keynan, A. 1972. Cryptobiosis: A review of the mechanisms
of the ametabolic state in bacterial spores,
p. 355-362. Iji H. O. Halvorson, R. Hanson, and
L. 0. Campbell (ed.), Spores V. American Society for
Microbiology, Washington, DC.
Kraffczyk, I., G. Trolldenier, and H. Beringer. 1984.
Soluble root exudates of maize: Influence of potassium
supply and rhizosphere microorganisms. Soil. Biol.
Biochem. 16:315-322.
Kramer, M. J., and M. D. Socolofsky. 1970. Effect of
chloramphenicol on the conversion of dormant
Azotobacter cysts into vegetative cells. Texas Rep.
Biol. Med. 1 & 2:43-48.

186
Krieg, N. R., and J. Dobereiner. 1984. The genus
Azospirillum, p. 94-104. In N. R. Krieg and J. G. Holt
(ed.), Bergey's manual of systematic bacteriology,
vol. 1. The Williams & Wilkins Co., Baltimore.
Labeda, D. P., K.-C. Liu, and L. E. Casida, Jr. 1976.
Colonization of soil by Arthrobacter and Pseudomonas
under varying conditions of water and nutrient
availability as studied by plate counts and trans
mission electron microscopy. Appl. Environ.
Microbiol. 31:551-561.
Lakshmi, V. A. Satyanarayana Rao, K. Vijayalakshmi, M.
Lakshmi-Kumari, K. V. B. R. Tilak, and N. S. Subba
Rao. 1977. Establishment and survival of Spirillum
lipoferum. Proc. Indian Acad. Sci. 86B:397-404.
Lamm, R. B., and C. A. Neyra. 1981. Characterization and
cyst production of azospirilla isolated from selected
grasses growing in New Jersey and New York. Can. J.
Microbiol. 27:1320-1325.
Law, J. H., and R. A. Slepecky. 1961. Assay of poly-8-
hydroxybutyric acid. J. Bacteriol. 82:33-36.
Leive, L. L., and B. D. Davis. 1980. Cell envelope;
spores, p. 73-110. In B. D. Davis, R. Dulbecco, H. N.
Eisen, and H. S. Ginsberg (ed.), Microbiology, 3rd ed.
Harper & Row, Publishers, Philadelphia.
Lin, L. P., and H. L. Sadoff. 1968. Encystment and polymer
production by Azotobacter vinelandii in the presence of
-hydroxybutyrate. J. Bacteriol. 95:2336-2343.
Lin, L. P., and H. L. Sadoff. 1969. Preparation and ultra
structure of the outer coats of Azotobacter vinelandii
cysts. J. Bacteriol. 98:1335-1341.
Lin, W., Y. Okon, and R. W. F. Hardy. 1983. Enhanced
mineral uptake by Zea mays and Sorghum bicolor roots
inoculated with Azospirillum brasilense. Appl.
Environ. Microbiol. 45:1775-1779.
Loperfido, B., and H. L. Sadoff. 1973. Germination of
Azotobacter vinelandii cysts: Sequence of macromolecu-
lar synthesis and nitrogen fixation. J. Bacteriol.
113:841-846.
Lowendorf, H. S. 1980. Factors affecting survival of
Rhizobium in soil, p. 87-124. In M. Alexander (ed.),
Advances in microbial ecology, vol. 4. Plenum Press,
New York.

187
Marshall, K. C. 1980. Bacterial adhesion in natural
environments, p. 187-196. In R. C. W. Berkeley, J. M.
Lynch, J. Melling, P. R. Rutter, and B. Vincent (ed.),
Microbial adhesion to surfaces. Ellis Horwood Limited
Publishers, Chicester.
Martinez-Drets, G., M. Del Gallo, C. Burpee, and R. H.
Burris. 1984. Catabolism of carbohydrates and organic
acids and expression of nitrogenase by azospirilla.
J. Bacteriol. 159:80-85.
Martinez-Drets, G., E. Fabiano, and A. Cardona. 1985.
Carbohydrate catabolism in Azospirillum amazonense.
Appl. Environ. Microbiol. 50:183-185.
Matthews, S. W., S. C. Schank, H. C. Aldrich, and R. L.
Smith. 1983. Peroxidase-antiperoxidase labeling of
Azospirillum brasilense in field-grown pearl millet.
Soil Biol. Biochem. 15:699-703.
Nakanishi, I., K. Kimura, S. Kusui, and E. Yamazaki.
1974. Complex formation of gel-forming bacterial
(1^3)-6-D-glucans (curdlan-type polysaccharides) with
dyes in aqueous solution. Carbohydrate Research
32:47-52.
Nelson, L. M., and R. Knowles. 1978. Effect of oxygen and
nitrate on nitrogen fixation and denitrification by
Azospirillum brasilense grown in continuous culture.
Can. J. Microbiol. 24:1395-1403.
Neuer, G., A. Kronenberg, and H. Bothe. 1985. Denitrifica
tion and nitrogen fixation by Azospirillum III.
Properties of a wheat-Azospirillum association.
Arch. Microbiol. 141:364-370.
Neyra, C. A., and J. Dobereiner. 1977. Denitrification by
N^-fixing Spirillum lipoferum. Can. J. Microbiol.
23:300-305.
Neyra, C. A., and P. van Berkum. 1977. Nitrate reduction
and nitrogenase activity in Spirillum lipoferum.
Can. J. Microbiol. 23:306-310.
Nickels, J. S., J. D. King, and D. C. White. 1979. Poly-B-
hydroxybutyrate accumulation as a measure of unbalanced
growth of the estuarine detrital microbiota. Appl.
Environ. Microbiol. 37:459-465.

188
Nur, I., Y. Okon, and Y. Henis. 1980. Comparative studies
of nitrogen-fixing bacteria associated with grasses in
Israel with Azospirillum brasilense. Can. J.
Microbiol. 26:714-718.
Nur, I., Y. Okon, and Y. Henis. 1982. Effect of dissolved
oxygen tension on production of carotenoids, poly-g-
hydroxybutyrate, succinate oxidase and superoxide
dismutase by Azospirillum brasilense Cd grown in
continuous culture. J. Gen. Microbiol. 128:2937-
2943.
Okon, Y., S. L. Albrecht, and R. H. Burris. 1976a. Factors
affecting growth and nitrogen fixation of Spirillum
lipoferum. J. Bacteriol. 127:1248-1254.
Okon, Y., S. L. Albrecht, and R. H. Burris. 1976b. Carbon
and ammonia metabolism of Spirillum lipoferum.
J. Bacteriol. 128:592-597.
Okon, Y., and R. W. F. Hardy. 1983. Developments in basic
and applied biological nitrogen fixation, p. 5-54. Iji
F. C. Steward and R. G. S. Bidwell (ed.), Plant
physiology, a treatise, vol. VIII. Academic Press,
Inc., Orlando.
Okon, Y., P. G. Heytler, and R. W. F. Hardy. 1983. N^
fixation by Azospirillum brasilense and its
incorporation into host Setaria itlica. Appl.
Environ. Microbiol. 46:694-697.
Page, W. J., and H. L. Sadoff. 1975. Relationship between
calcium and uronic acids in the encystment of
Azotobacter vinelandii. J. Bacteriol. 122:145-151.
Papen, H., and D. Werner. 1980. Biphasic nitrogenase
activity in Azospirillum brasilense in long lasting
batch cultures. Arch. Microbiol. 128:209-214.
Papen, H., and D. Werner. 1982. Organic acid utilization,
succinate excretion, encystation and oscillating nitro
genase activity in Azospirillum brasilense under micro-
aerobic conditions. Arch. Microbiol. 132:57-61.
Patriquin, D. G., J. Dobereiner, and D. K. Jain. 1983.
Sites and processes of association between diazotrophs
and grasses. Can. J. Microbiol. 29:900-915.
Pope, L. M., and 0. Wyss. 1970. Outer layers of the
Azotobacter vinelandii cyst. J. Bacteriol. 102:
234-239.

189
Reusch, R. N., and H. L. Sadoff. 1979. 5-n-Alkyl-
resorcinols from encysting Azotobacter vinelandii:
Isolation and characterization. J. Bacteriol. 139:
448-453.
Reusch, R. N., and H. L. Sadoff. 1981. Lipid metabolism
during encystment of Azotobacter vinelandii.
J. Bacteriol. 145:889-895.
Reusch, R. N;, and H. L. Sadoff. 1983. Novel lipid
components of the Azotobacter vinelandii cyst mem
brane. Nature 302:268-270.
Robinson, J. B., P. 0. Salonius, and F. E. Chase. 1965. A
note on the differential response of Arthrobacter
spp. and Pseudomonas spp. to drying in soil.
Can. J. Microbiol. 11:746-748.
Rodriguez Caceres, E. A. 1982. Improved medium for isola
tion of Azospirillum spp. Appl. Environ. Microbiol.
44 .-990-991.
Ruscoe, A. W. E. H. Newcomb, and R. H. Burris. 1978.
Pleomorphic forms in strains of Spirillum lipoferum by
light and electron microscopy, Abstract C-60, p. 30.
In Proceedings of the Steenbock-Kettering International
Symposium on Nitrogen Fixation. University of
Wisconsin, Madison.
Sadasivan, L. 1985. In Proceedings of the Third Workshop
on Azospirillum: Physiology, genetics, ecology.
Bayreuth, West Germany (in press).
Sadasivan, L., and C. A. Neyra. 1985. Flocculation in
Azospirillum brasilense and Azospirillum lipoferum:
Exopolysaccharides and cyst formation. J. Bacteriol.
163:716-723.
Sadoff, H. L. 1975. Encystment and germination in
Azotobacter vinelandii. Bacteriol. Rev. 39:516-539.
Sadoff, H. L., E. Berke, and B. Loperfido. 1971.
Physiological studies of encystment in Azotobacter
vinelandii. J. Bacteriol. 105:185-189.
Scott, D. B., C. A. Scott, and J. Dobereiner. 1979. Nitro-
genase activity and nitrate respiration in Azospirillum
spp. Arch. Microbiol. 121:141-145.

190
Senior, P. J., and E. A. Dawes. 1971. Poly-6-
hydroxybutyrate biosynthesis and the regulation of
glucose metabolism in Azotobacter beijerinckii.
Biochem. J. 125:55-66.
Shively, J. M. 1974. Inclusion bodies of prokaryotes.
Ann. Rev. Microbiol. 28:167-187.
Slater, M., and M. Schaechter. 1974. Control of cell
division in bacteria. Bacteriol. Rev. 38:199-221.
Smith, M. S., and J. M. Tiedje. 1979. The effects of roots
on soil denitrification. Soil Sci. Soc. Am. J.
43:951-955.
Socolofsky, M. D., and 0. Wyss. 1961. Cysts of
Azotobacter. J. Bacteriol. 81:946-954.
Socolofsky, M. D., and O. Wyss. 1962. Resistance of the
Azotobacter cyst. J. Bacteriol. 84:119-124.
Stevenson, L. H., and M. D. Socolofsky. 1966. Cyst forma
tion and poly-B-hydroxybutyric acid accumulation in
Azotobacter. J. Bacteriol. 91:304-310.
Stockdale, H., D. W. Ribbons, and E. A. Dawes. 1968.
Occurrence of poly- -hydroxybutyrate in the
Azotobacteriaceae. J. Bacteriol. 95:1798-1803.
Stotzky, G. 1980. Surface interactions between clay
minerals and microbes, viruses and soluble organics,
and the probable importance of these interactions to
the ecology of microbes in soil, p. 231-247. In
R. C. W. Berkeley, J. M. Lynch, J. Melling,
P. R. Rutter, and B. Vincent (ed.), Microbial adhesion
to surfaces. Ellis Horwood Limited Publishers,
Chicester.
Su, C.-J., R. N. Reusch, and H. L. Sadoff. 1981. Isolation
and characterization of several unique lipids from
Azotobacter vinelandii cysts. J. Bacteriol. 147:80-90.
Sudo, S. Z., and M. Dworkin. 1973. Comparative biology of
prokaryotic resting cells, p. 153-224. In A. H. Rose
and D. W. Tempest (ed.), Advances in microbial
physiology, vol. 9. Academic Press, New York.
Sutherland, I. W. 1977. Bacterial exopolysaccharides--
Their nature and production, p. 27-96. In I. W.
Sutherland (ed.), Surface carbohydrates of the
prokaryotic cell. Academic Press, New York.

191
Tal, S., and Y. Okon. 1985. Production of the reserve
material poly-B-hydroxybutyrate and its function in
Azospirillum brasilense Cd. Can. J. Microbiol.
31:608-613.
Tarrand, J. J., N. R. Krieg, and J. Dobereiner. 1978. A
taxonomic study of the Spirillum lipoferum group, with
descriptions of a new genus, Azospirillum gen. nov. and
two species, Azospirillum lipoferum (Beijerinck)
comb. nov. and Azospirillum brasilense sp. nov.
Can. J. Microbiol. 24:967-980.
Tchan, Y.-T. 1984. Family II. Azotobacteraceae, p. 219-
229. In N. R. Krieg and J. G. Holt (ed.), Bergey's
manual of systematic bacteriology, vol. I. The
Williams and Wilkins Co., Baltimore.
Thompson, J. P., and V. B. D. Skerman. 1979. Azotobac
teraceae : The taxonomy and ecology of the aerobic
nitrogen-fixing bacteria. Academic Press, New York.
Tien, T. M., H. G. Diem, M. H. Gaskins, and D. H. Hubbell.
1981. Polygalacturonic acid transeliminase production
by Azospirillum species. Can. J. Microbiol.
27 .-426-431.
Tien, T. M., M. H. Gaskins, and D. H. Hubbell. 1979. Plant
growth substances produced by Azospirillum brasilense
and their effect on the growth of pearl millet
(Pennisetum americanum L.). Appl. Environ. Microbiol.
37:1016-1024.
Tudor, J. J. 1980. Chemical analysis of the outer cyst
wall and inclusion material of Bdellovibrio bdello-
cysts. Curr. Microbiol. 4:251-256.
Tudor, J. J., and S. F. Conti. 1977. Characterization of
bdellocysts of Bdellovibrio sp. J. Bacteriol. 131:
314-322.
Tyler, M. E., J. R. Milam, R. L. Smith, S. C. Schank, and
D. A. Zuberer. 1979. Isolation of Azospirillum from
diverse geographic regions. Can. J. Microbiol.
25:693-697.
Umali-Garcia, M., D. H. Hubbell, M. H. Gaskins, and
F. B. Dazzo. 1980. Association of Azospirillum with
grass roots. Appl. Environ. Microbiol. 39:219-226.
Vela, G. R. 1974. Survival of Azotobacter in dry soil.
Appl. Microbiol. 28:77-79.

192
Vela, G. R-, and G. Cagle. 1969. Formation of fragile
cysts by a strain of Azotobacter chroococcum. J. Gen.
Microbiol. 57:365-368.
Volpon, A. G. T., H. De-Polli, and J. Dobereiner. 1981.
Physiology of nitrogen fixation in Azospirillum
lipoferum Br 17 (ATCC 29709). Arch. Microbiol. 128:
371-375.
Ward, J. B., and R. C. W. Berkeley. 1980. The microbial
cell surface and adhesion, p. 47-66. In R. C. W.
Berkeley, J. M. Lynch, J. Melling, P. R. Rutter, and
B. Vincent (ed.), Microbial adhesion to surfaces.
Ellis Horwood Limited Publishers, Chicester.
Webb, S. J. 1965. Bound water in biological integrity,
p. 146-171. Charles C. Thomas, Springfield, IL.
Whittenbury, R., S. L. Davies, and J. F. Davey. 1970a.
Exospores and cysts formed by methane-utilizing
bacteria. J. Gen. Microbiol. 61:219-226.
Whittenbury, R., K. C. Phillips, and J. F. Wilkinson.
1970b. Enrichment, isolation and some properties of
methane-utilizing bacteria. J. Gen. Microbiol.
61:205-218.
Winogradsky, S. 1938. Sur la morphologie et l'ecologie des
Azotobacter. Ann. Inst. Pasteur 60:351-400.
Wong, P. P., N. E. Stenberg, and L. Edgar. 1980. Charac
terization of a bacterium of the genus Azospirillum
from cellulolytic nitrogen-fixing mixed cultures.
Can. J. Microbiol. 26:291-296.
Wyss, O., D. D. Smith, L. M. Pope, and K. E. Olson. 1969.
Endogenous encystment of Azotobacter vinelandii.
J. Bacteriol. 100:475-479.
Zimmer, W., M. Penteado Stephan, and H. Bothe. 1984.
Denitrification by Azospirillum brasilense Sp 7 I.
Growth with nitrite as respiratory electron acceptor.
Arch. Microbiol. 138:206-211.

BIOGRAPHICAL SKETCH
Bruce Henry Bleakley was born in Pontiac, Michigan, in
February of 1956. In June 1974 he graduated from New Haven
High School in New Haven, Michigan. He entered college in
September 1974 at Michigan State University, East Lansing,
Michigan. In September 1978 he graduated with a Bachelor of
Science degree in agronomy. He entered the Graduate School
at Michigan State in January 1979 and received the degree of
Master of Science in agronomy in March 1982. He entered the
Graduate School of the University of Florida in January 1982
to work on the degree of Doctor of Philosophy in agronomy.
193

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
:ry H.
Professor
Ga/skins, Chairman
or Agronomy
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Stephan L. Albrecht
Assistant Professor of Agronomy
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Sbbpheh G. zam
Associate Professor of Microbiology
Cell Science
an'

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
David J. Mitchell
Professor of Pathology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
David H. Hubpell
Professor of Soil Science
This dissertation was submitted to the Graduate Faculty of
the College of Agriculture and the Graduate School and was
accepted as partial fulfillment of the requirements for the
degree of Doctor of Philosophy.
May 1986
Dean,
aJt 7. 3.
ege of Agriculture
a
Dean, Graduate School



in floes, but all cells possessed intracellular PHB and
capsules. Some cells within floes appeared cyst-like.
Broth studies indicated that alkaline pH does not cause
these morphological changes.
Cells of Azospirillum lipoferum Sp RG6xx grown on
nitrogen-free BHB agar accumulated up to 57% of their dry
weight as PHB, compared to 3.6% when grown with combined
nitrogen. Neither vegetative nor encapsulated cells of this
strain survived in significant numbers after 8 days of
desiccation. Vegetative cells of this strain multiplied
several fold and retained viability during 9 days of starva
tion for carbon and nitrogen, whereas encapsulated cells
were reduced to 25% of their original numbers. Nonmotile,
encapsulated cells produced motile vegetative cells when
incubated with nitrate, ammonium, or soil extract but did
not do so appreciably in nitrogen-free, buffered-salts
solution with or without carbon sources. Treatment with
Tris-EDTA did not result in expulsion of cells from their
capsular coats, as it does for mature Azotobacter spp.
cysts. Studies with chloramphenicol indicated that
encapsulated cells do not possess the enzymes needed for
growth and emergence from their capsules.
The studies suggested that PHB accumulation and capsule
formation during unbalanced growth precede the formation of
dormant cyst-like cells.
Vll


50
normal level, these morphological changes were accelerated,
occurring within 2 to 3 days after inoculation. This strain
could not fix dinitrogen with glucose as the carbon source,
but otherwise its biotin requirement and pleomorphism were
typical of A. lipoferum.
Lamm and Neyra (1981) found that A. lipoferum strains
grown in nitrogen-free, semisolid malate medium developed
many elongated cells after 2 days of culture, whereas A.
brasilense strains only developed elongated cells after 10
days. In both semisolid and agar-plate, nitrogen-free
malate cultures, thick-walled, optically refractile cells
were present as 1.0% of the cells in day-old cultures of
both species. After 4 days, the numbers of these thick-
walled cells were equal to or greater than cells of normal
morphology. The ovoid cells of A. lipoferum strains were
about twice as large as those produced by A. brasilense
strains. Such cells were never observed in nutrient broth
cultures, but could be obtained in old cell lawns grown on
nutrient agar. The increased desiccation resistance of
these ovoid cells has already been discussed.
Papen and Werner (1982) observed apparent cysts of A.
brasilense Sp 7. Such nonmotile cells were encapsulated,
having a diameter of about 1.2 pm, were not fixing dinitro
gen, and were desiccation resistant. They composed the
majority of cells in the nitrogen-free, malate broth culture
between the second and fourth days of incubation. After


138
by filamentation and septation. Cells gradually lost their
motility during this time. The septating filaments later
produced extensive capsular material on nitrogen-free,
LP-BHB agar, and sometimes in HP-BHB broth. The HP buffer
repressed extensive capsule formation on nitrogen-free,
HP-BHB agar.
These and previous studies indicate that pleomorphic
cells of azospirilla arise under two different conditions.
After cells have experienced balanced growth with combined
nitrogen, some may become pleomorphic when a nutrient
essential to growth becomes limiting. The production of
both filamentous cells and exopolysaccharides seemed to
occur extensively in HP-TSS broth in stationary phase, but
PHB accumulation was not extensive.
Azospirilla may also become pleomorphic during growth
where the C/N ratio of nutrients available to the cells is
high. Dinitrogen-fixing cells may be poised to become
pleomorphic. It is also evident that azospirilla may take
up some carbon sources more rapidly than they can utilize
combined nitrogen, resulting in extensive PHB accumulation
and capsule formation even when combined nitrogen is avail
able. Cells under these conditions may experience a
temporary shift-down, until enough combined nitrogen can be
assimilated to mobilize their PHB deposits and enable
further growth. This would explain the morphological
changes observed in this study with A. brasilense strain Cd


66
(Sigma), that was incorporated at a final concentration of
0.0375 grams per liter (Rodriguez Caceres, 1982).
Nitrogen-free agar plates containing n-butanol had the
same composition as SNF plates, except that the agar concen
tration was 1.5% (wt/vol), and yeast extract and Congo Red
were omitted. The n-butanol was sterilized by filtration in
the same manner as the biotin and incorporated at a final
concentration of 0.2% (vol/vol). Beta-hydroxybutyrate was
prepared from crotonic acid (Sigma) by dissolving 23.6 g
crotonic acid in 900 ml deionized water. This solution was
continuously mixed with a magnetic stirrer for two to three
days at 25C. Its 0^235 stabilized by this time, indicating
conversion to BHB (H. L. Sadoff, personal communication).
It was then adjusted to pH 7.0 with 10 M KOH, the final
volume made up to one liter, and sterilized by autoclaving.
This served as a 10X concentrated stock solution of BHB for
addition to agar or broth, to give a final concentration of
0.236% (wt/vol) BHB. Agar plates containing BHB had the
same composition as n-butanol plates, except that (NH^^SO.
or Congo Red were sometimes added at the previously
described concentrations. For two-step broth replacement
studies (described below), broth contained BHB, biotin, and
phosphate-buffered basal salts solution. The LP and HP
buffers were employed in different broth replacement
studies. The initial pH, after inoculation, of LP-BHB broth
was 7.2, and that of HP-BHB broth was 6.9.


180
Inoculum production may then be designed to introduce and
sustain the most beneficial physiological form in the root
zone.


MORPHOLOGY AND PHYSIOLOGY OF AZOSPIRILLA
GROWN ON BETA-HYDROXYBUTYRATE
By
BRUCE HENRY BLEAKLEY
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA



PAGE 1

MORPHOLOGY AND PHYSIOLOGY OF AZOSPIRILLA GROWN ON BETA-HYDROXYBUTYRATE By BRUCE HENRY BLEAKLEY DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1986

PAGE 2

This dissertation is dedicated to Isabel, Stewart, and Robert Bleakley, the only three people who can say "Of course" without my doubting them. I i

PAGE 3

ACKNOWLEDGMENTS I thank my major professor. Dr. Murray Gaskins, for first bringing the topic of bacterial cysts to my attention and for allowing me to pursue the subject. I also thank the other members of my guidance committee, Dr. Stephan Albrecht, Dr. David Hubbell, Dr. David Mitchell, and Dr. Stephen Zam, for their patience, encouragement, and use of laboratory facilities. Thanks also go to Dr. Sylvia Coleman, whose transmission electron microscopy studies aided this work. Special thanks go to Dr. Kenneth Quesenberry and the state of Florida for my graduate assistant ship. Kelly Kirkendall Merritt taught me how to spread plate and introduced me to manipulations of azospirilla. She is a good coworker to be around, as are Stephanie Syslo, Mary Myers, and Dr. Garnet Jex, who helped me with discussions, calculations, and shared experiences in and out of the laboratory. Dr. Lakshmi Sadasivan discussed her work with me before it was published and furnished preprints of her work. Talks with Dr. Harold Sadoff helped me a great deal in understanding Azotobacter cysts. And Dr. Noel Krieg provided several cultures that were keys in the study. Thanks go to all of them. iii

PAGE 4

Without the Hume Library this work could not have been done. I thank Mr. William Weaver for running a fine facility. Finally, completing the list of good coworkers, the eye and expertise of Louise L. Munro are felt throughout this study. Her scanning electron microscopy studies cleared up several uncertainties and raised new questions. She knew the right stuff when she saw it. Thanks, Louise. iv

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGMENTS iii ABSTRACT vi CHAPTERS I INTRODUCTION AND EXPERIMENTAL APPROACH 1 Ecology of Azospirilla 1 Physiology of Azospirilla 4 Morphology of Azospirilla 6 Prokaryotic Exopolysaccharides 8 Pbly-6-Hydroxybutyrate ( PHB ) 12 Dormant Forms of Prokaryotic Cells 22 Resistance of Bacteria to Drying 27 Azotobacter Cysts 37 Pleomorphism of Azospirilla 45 Experimental Approach 56 II PLEOMORPHISM OF AZOSPIRILLA GROWN ON BETA-HYDROXYBUTYRATE 5 9 Materials and Methods 63 Results 71 Discussion 131 III PHYSIOLOGICAL PROPERTIES OF ENCAPSULATED FLOCS OF AZOSPIRILLUM LIPOFERUM Sp RG6xx 144 Materials and Methods 145 Results 153 Discussion 167 IV GENERAL CONCLUSIONS 178 BIBLIOGRAPHY 181 BIOGRAPHICAL SKETCH 193 V

PAGE 6

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MORPHOLOGY AND PHYSIOLOGY OF AZOSPIRILLA GROWN ON BETA-HYDROXYBUTYRATE By Bruce Henry Bleakley May 1986 Chairman: Murray H. Gaskins Major Department: Agronomy Strains of Azospirillum brasilense and Azospirillum lipof erum were cultured with beta-hydroxybutyrate ( BHB ) to determine if they could be converted in high numbers to cyst-like forms, as can some strains of Azotobacter spp. Azospirillum brasilense strain JM 125A2 grew poorly on BHB but produced some nonmotile cells of cyst-like morphology. Azospirillum brasilense strain Cd grew better on BHB and often produced elongated cells as well as some nonmotile, cyst-like cells. Capsules and accumulation of poly-betahydroxybutyrate (PHB) were common features of all putative cysts. Encapsulation occurred with all A. lipof erum strains tested. Cells accumulated PHB and assumed elongated, filamentous shapes as they lost motility. Later, capsules were produced and microflocs formed. The filamentous cells eventually formed septa. Several cell shapes were present vi

PAGE 7

in floes, but all cells possessed intracellular PHB and capsulesSome cells within floes appeared cyst-like. Broth studies indicated that alkaline pH does not cause these morphological changes. Cells of Azospirillum lipof erum Sp RG6xx grown on nitrogen-free BHB agar accumulated up to 57% of their dry weight as PHB, compared to 3.6% when grown with combined nitrogen. Neither vegetative nor encapsulated cells of this strain survived in significant numbers after 8 days of desiccation. Vegetative cells of this strain multiplied several fold and retained viability during 9 days of starvation for carbon and nitrogen, whereas encapsulated cells were reduced to 25% of their original numbers. Nonmotile, encapsulated cells produced motile vegetative cells when incubated with nitrate, ammonium, or soil extract but did not do so appreciably in nitrogen-free, buf f ered-salts solution with or without carbon sources. Treatment with Tris-EDTA did not result in expulsion of cells from their capsular coats, as it does for mature Azotobacter spp. cysts. Studies with chloramphenicol indicated that encapsulated cells do not possess the enzymes needed for growth and emergence from their capsules. The studies suggested that PHB accumulation and capsule formation during unbalanced growth precede the formation of dormant cyst-like cells. vii

PAGE 8

CHAPTER I INTRODUCTION AND EXPERIMENTAL APPROACH Ecology of Azospirilla Bacteria of the genus Azospirillum have been isolated from soils and from the roots of cereal crops and forage grasses in several areas of the world (Dobereiner et al., 1976; Tyler et al., 1979; Lamm and Neyra, 1981). Their 4 numbers in nonrhizosphere soil can be as high as 10 cells/g soil (Dobereiner, 1978), while their numbers in rhizosphere soil can be as high as 10^ cells/g soil (Krieg and Dobereiner, 1984). Agricultural interest in Azospirillum spp. resulted from recognition of their ability to reduce atmospheric dinitrogen. The enzyme catalyzing this reaction, nitrogenase, is inactivated in the presence of combined nitrogen or oxygen (Okon et al., 1976a). Azospirilla fix dinitrogen under microaerophilic conditions in nitrogen-free media in the laboratory (Day and Dobereiner, 1976; Okon et al. 1976a) Nonrhizosphere soil is usually too poor in available, utilizable carbon sources to enable Azospirillum spp. to fix dinitrogen, but they can do so in the more carbon-rich rhizosphere environment (Dobereiner et al., 1976). Maximum nitrogenase activity with inoculated plants

PAGE 9

2 growing in soil is often found at the reproductive stage of the plant (reviewed by Patriquin et al., 1983), after plant uptake and other processes have reduced the amount of combined nitrogen in the root zone (Okon and Hardy, 1983). Sometimes low amounts of fixed nitrogen have been incorporated into plant tissue. The transfer of fixed nitrogen from bacterium to plant seems slow, probably because bacterial nitrogen is made available for plant uptake only after the mineralization of the organic nitrogen of dead bacteria (Okon et al., 1983). Although the nitrogenase activity of Azospirillum spp. may not directly provide quick or agriculturally significant benefits to inoculated plants, the bacteria have been found to possess other characteristics that may benefit plants. Axenic associations of grass seedlings and Azospirillum spp. have resulted in rapid proliferation of lateral roots and root hairs, probably due to bacterial production of indole-3-acetic acid and other plant growth substances (Tien et al., 1979; Umali-Garcia et al., 1980; Jain and Patriquin, 1985). It is also possible that Azospirillum spp. can enhance production of plant growth substances by the plants themselves (Kapulnik et al., 1985). In any case, associations of Azospirillum spp. with plant roots have led to significant increases of commercially valuable plant components in both axenic laboratory experiments (Kapulnik et al., 1985) and field inoculations

PAGE 10

(Okon and Hardy, 1983). Short-term axenic associations have also resulted in enhanced uptake of mineral ions by grass roots (Lin et al., 1983). This effect may be due to the influence of plant growth substances, or to softening of the middle lamellae of root cells by pectolytic bacterial enzymes, which some azospirilla are known to produce (Umali-Garcia et al. 1980; Tien et al. 1981). Such effects on root morphology and activity may make inoculation with azospirilla beneficial in some agricultural situations. The rhizosphere environment is prone to extreme chemical and physical fluctuations (Foster and Bowen, 1982). This may lead to periods when azospirilla are inactive due to environmental limitations. Cells of azospirilla can vary morphologically (Krieg and Dobereiner, 1984). Some of these cell forms may be dormant or resting stages, in which activities of possible benefit to plants are not expressed. Pleomorphic forms of azospirilla usually possess capsules, and contain large amounts of the reserve polymer poly-6hydroxybutyrate ( PHB ) This study describes attempts to obtain such forms in high numbers by laboratory culture. The general topics of capsules, PHB, physiological dormancy, and desiccation resistance are directly related to this study, and will be briefly reviewed in this introduction after discussion of some key aspects of Azospirillum spp. physiology.

PAGE 11

4 Physiology of Azospirilla The three species in the genus Azospirillum all have a mainly respiratory type of metabolism. They fix dinitrogen in microaerophilic environments where combined nitrogen concentration is low, and utilizable carbon-and-energy sources are available. When provided with metabolizable carbon, along with ammonium, nitrate, or other combined nitrogen sources, they can grow under aerobic conditions. In either situation, they grow well on the salts of organic acids such as malate, succinate, lactate, or pyruvate (Krieg and Dobereiner, 1984). Azospirillum brasilense can use some carbohydrates, including fructose, galactose, and arabinose. Azospirillum lipof erum is also able to use these sugars, as well as glucose, mannose, and sorbose (Martinez-Drets et al., 1984). The most recently recognized species, Azospirillum amazonense differs from the other two species in that it can grow on sucrose and other disaccharides (Martinez-Drets et al., 1985). Both A. brasilense and A. amazonense can synthesize their own biotin, whereas A. lipof erum can only grow if exogenous biotin is available (Falk et al. 1985 ) Microaerophilic culture conditions for dinitrogen fixation can be established by culturing azospirilla in media containing 0.05% (wt/vol) agar. The bacteria grow and form a pellicle slightly below the agar surface, where diffusion of 0^ from the culture-vessel headspace balances the

PAGE 12

5 uptake of 0^ by the bacteria, allowing both cell respiration and protection of the oxygen-sensitive nitrogenase (Okon et al., 1976a). Broth cultures can fix dinitrogen if the dissolved oxygen level is well controlled (Okon et al., 1976a). The bacteria are also able to grow on the surface of nitrogen-free, aerobically incubated agar plates (Day and Dobereiner, 1976). The flagella of azospirilla enable them to move to whatever sites their physiological state demands. They have been shown to exhibit aerotaxis to microaerophilic sites (Barak et al., 1982). Alternatively, cells may aggregate, thereby creating a microaerophilic environment by the respiration of many cells in a small space (Barak et al., 1982). The grass rhizosphere may contain microaerophilic sites (reviewed by Patriquin et al., 1983). Azospirilla could migrate from soil toward such sites, where nitrogenase activity could subsequently be expressed. The respiratory metabolism of Azospirillum spp. includes the ability of many strains to denitrify, reducing nitrate or nitrite to more reduced nitrogenous compounds under anaerobic conditions if enough metabolizable carbon source is available (Neyra and Dobereiner, 1977; Neyra and van Berkum, 1977; Nelson and Knowles, 1978). Under certain laboratory conditions, denitrif ication has been shown to provide enough ATP to support anaerobic growth of azospirilla (Bothe et al., 1981; Zimmer et al., 1984). The

PAGE 13

6 ATP derived from denitrif ication can be used to drive nitrogenase activity (Scott et al., 1979), but it seems unlikely that dinitrogen fixation under these conditions can support growth of the bacteria (Bothe et al., 1981). Recent work by Neuer et al. (1985) has shown that, in axenic wheatAzospirillum spp. associations, both dinitrogen fixation and denitrif ication can occur. Morphology of Azospirilla Azospirilla are Gram-negative bacteria (Tarrand et al., 1978). The structural layers external to the cytoplasmic membrane of Gram-negative prokaryotes have been reviewed (Costerton et al., 1974). Depending on cultural conditions and the bacterial strain, polysaccharide or capsular layers may be present as the outermost layers of the cell. A growing Gram-negative cell divides by binary fission to produce two daughter cells of approximately equal size. Division begins with invagination of the cytoplasmic membrane and peptidoglycan until a complete transverse septum or cross wall is formed. When the septum is completely formed and cleaved, the two daughter cells separate (Leive and Davis, 1980). As will be discussed later, this cell division process can be disrupted, resulting in formation of filaments or chains, which accounts partially for pleomorphism of azospirilla.

PAGE 14

7 Cells of Azospirillum brasilense and Azospirillum lipof erum have a similar appearance when cultured in broth containing combined nitrogen. They are short, plump, slightly curved rods averaging 1.0 'm in diameter and 2.1 to 3.8 -m in length. They are motile in broth by means of a single, polar flagellum (Tarrand et al., 1978). Cells of azospirilla often contain granules of the polymer PHB (Krieg and Dobereiner, 1984). Grown with combined nitrogen, Azospirillum brasilense Sp 7 (ATCC 29145) has 0.5% to 1.0% of its dry weight as PHB. When grown in dinitrogen-f ixing conditions, the PHB content rises until as much as 25.0% of its dry weight is PHB (Okon et al., 1976b) Granules of PHB are present in cells grown on combined nitrogen, but granule size and number are reduced compared to that found in dinitrogen-f ixing cells (Albrecht and Okon, 1980 ) Under certain cultural conditions, cells of azospirilla produce an outermost layer of capsular polysaccharide. When grown on an agar medium containing peptone, succinic acid and ammonium sulfate at 37C for 48 to 72 hours, a small proportion of cells are Gram-variable, possibly because they possess capsules. On this medium, A. brasilense exhibits more Gram-variability than does A. lipof erum When cells of either species are cultured in the broth form of this medium, they stain uniformly Gram-negative, at least in young cultures (Krieg and Dobereiner, 1984).

PAGE 15

8 Prokaryotic Exopolysaccharides Many genera of both Gram-positive and Gram-negative bacteria include species that can produce polysaccharide layers outside their cell wall. Such layers have been referred to as capsules, exopolysaccharides (Sutherland, 1977) or glycocalyces (Costerton et al., 1981). Depending on laboratory cultural conditions, these polymers can assume different forms. Slime layers adhere loosely, if at all, to the cell and can often be separated from cells by centrifugation. Capsular layers appear to be tightly bound to the cell itself, and cannot be easily separated from cells. Microcapsules are so thin that their presence outside the cell wall cannot be observed using staining and light microscopy, while macrocapsules are of sufficient width to be so resolved (Ward and Berkeley, 1980). Although proteins are sometimes present in bacterial capsules, most capsules are mainly polysaccharide in composition. The polysaccharides are extensively hydrated, and up to 99% by weight of the capsule is accounted for as water (Costerton et al., 1981). The ATP needed to activate sugar residues for exopolysaccharide synthesis has been shown to comprise a significant proportion of total-cellular-ATP demand for some bacteria. Even when the carbon supply is growth limiting, some strains of bacteria produce extracellular polysaccharides (Jarman and Pace, 1984).

PAGE 16

9 For many bacteria, a culture medium having a high carbon to nitrogen (C/N) ratio promotes capsule formation (Sutherland, 1977; Costerton et al., 1981). Some species manufacture exopolysaccharide throughout all phases of growth, while others produce it only at certain stages of growth (Sutherland, 1977). Exopolysaccharides of more than one composition can be formed by the same bacterium under different environmental conditions (Geesey, 1982). In laboratory culture, exopolysaccharides may be nonessential for bacterial growth. Enzymatic removal of capsules often causes no reduction in viability of the decapsulated cells (Dudman, 1977). Nonencapsulated mutants may grow better in laboratory culture than do encapsulated cells, since they expend no energy for capsular synthesis (Costerton et al., 1981). Many nonencapsulated laboratory strains are mutants that have lost the ability of the wild type to produce exopolysaccharide. In other instances, common laboratory media have too low a C/N ratio to promote exopolysaccharide synthesis. Attachment of bacteria to surfaces by their exopolysaccharides is the rule in nature, whether the surface is an inert mineral particle or a biological surface such as a plant root (Costerton et al., 1981). Natural environments are far different from laboratory cultural conditions, containing many more potential hazards to bacterial survival. In natural environments, the presence of

PAGE 17

10 exopolysaccharides may aid the survival of bacteria (Dudman, 1977). Exopolysaccharides can concentrate nutrients from the surrounding solution phase. They give some bacteria increased resistance to antibiotics, surfactants, and other chemicals, as well as deterring their engulfment by phagocytic cells (Costerton et al., 1981). Other advantages of exopolysaccharides have been suggested, such as mediation of gas exchange between bacteria and their surroundings, but they have proven difficult to prove experimentally. Extracellular enzymes might also be located within or at the surface of capsules (Geesey, 1982). Nur et al. (1980) found that A. brasilense Sp 7 and an Israeli isolate of A. brasilense both possessed small capsules discernible by electron microscopy when grown on nutrient agar. Umali-Garcia et al. (1980) found that when certain A. brasilense strains and grass seedlings were incubated together for 10 to 30 min at 30C, many bacteria adhered to the grass roots, with granular material accumulating on the surfaces of root hairs, and fibrillar material accumulating on the surfaces of older, epidermal root cells. It is known that bacterial exopolysaccharides may appear either granular or fibrous (Foster and Bowen, 1982). The A. brasilense strains seemed to rapidly produce both types of exopolysaccharides in axenic association with grass roots. After 2 to 4 days of axenic association with grass roots, from two to four cells of A. brasilense Sp 7

PAGE 18

11 were sometimes seen to be enclosed within a common envelope or capsule. Such structures were not observed when the bacteria were grown in trypticase soy broth ( Umali-Garcia et al., 1980). This is another indication that the low C/N ratio of complex broth media can repress extensive capsule formation by azospirilla, while the high C/N ratio near plant roots can promote capsule formation. Recent work by Sadasivan and Neyra (1985) verified that the forms of carbon and nitrogen made available to azospirilla can have a profound effect on exopolysaccharide synthesis. When A. brasilense Sp 7 and A. lipof erum Sp. 59b (ATCC 29707) were cultured in broth containing 8 0 mM fructose and 0.5 mM KNO^, they grew as individual motile cells for only 6 hours and then started to clump, as exopolysaccharide production led to floe formation. Organic acids yielded fewer floes than did sugars, and other nitrogen sources, such as ammonium, yielded fewer floes than did nitrate. The cells in floes appeared initially to be enmeshed in a loose, fibrillar matrix that condensed progressively over a week's time. When cells were grown, harvested by centrif ugation and resuspended in broth lacking carbon, the cells remained freely suspended. This suggests that azospirilla have a high ATP demand for exopolysaccharide synthesis. Chemical analysis showed that cellulose was a major component of the exopolysaccharide.

PAGE 19

12 Fresh floes were not degraded by cellulase, indicating that more than one type of exopolysaccharide was present. Poly-B-Hydroxybutyrate (PHB) ^ In a constant and favorable environment where all nutrients are present in sufficient amounts, bacteria grow for a time in a steady state, where every component of the cell culture increases by the same constant factor per unit time. This is balanced growth, and occurs during the logarithmic phase of the growth curve ( Ingraham et al., 1983). If one or more nutrients become limiting, balanced growth is not maintained. When the carbon or energy supply is in excess, so that one or more other nutrients limit growth, some microorganisms respond by synthesizing and accumulating intracellular polymers having an energy-storage function (Dawes and Senior, 1973). The cell catabolizes these polymers when the energy supply from exogenous sources is no longer sufficient to maintain processes needed for maintenance of cell viability. These processes may include osmotic regulation, maintenance of intracellular pH and transmembrane potentials, and turnover of cellular constituents such as proteins and nucleic acids. The energy required for these processes is called the energy of maintenance. Some microorganisms do not produce special energy-storage polymers. Faced with a starvation environment, they are forced to utilize their own

PAGE 20

basal components, such as proteins and RNA, for energy sources. Possession of energy-storage polymers can benefit some species facing starvation, in that they degrade these polymers instead of or before they are forced to degrade such essential components as proteins (Dawes and Senior, 1973). However, different microorganisms utilize common constituents at different rates and in different sequences when starved. The possession of energy-reserve polymers does not spare degradation of protein and other basal components in some species during starvation. Most microorganisms that remain viable after prolonged starvation have a low endogenous metabolism, matched closely to their maintenance energy requirements. Starved microorganisms that rapidly metabolize polymers generally lose viability quickly (Dawes, 1976). Three main types of microbial energy-storage compounds are known. Some species can accumulate more than one. All of these compounds have high molecular weights, and only a slight effect on the internal osmotic pressure of the cell. The amount of each compound a cell accumulates can vary widely, depending on environmental conditions. Intracellular polyphosphates and glycogen-like polysaccharides are two types of energy-storage compounds formed by some eukaryotic and prokaryotic microorganisms. The synthesis of both types requires ATP.

PAGE 21

14 The third microbial energy-reserve polymer is poly-6hydroxybutyrate a straight chain homopolymer of D(-)-3hydroxybutyrate It is found only in prokaryotic cells, including both Gram-positive and Gram-negative species. Its synthesis requires reducing power in the forms of NADH or NADPH, but does not require the direct expenditure of ATP (Dawes and Senior, 1973). With phase contrast microscopy, large accumulations of PHB within bacterial cells appear as light-ref ractile granules. A single granule may contain several thousand PHB molecules (Dawes and Senior, 1973). Each granule is bounded by a nonunit-membrane layer, which is probably formed from the cytoplasmic membrane. Presumably the enzymes for polymerization and depolymerization of PHB are present in this membrane layer (Shively, 1974). Many of the Azotobacteraceae accumulate PHB when grown under dinitrogen-f ixing conditions. There can be a wide variation in PHB content between species and between strains of the same species (Stockdale et al., 1968). The regulation of PHB levels in Azotobacter bei jerinckii has been extensively studied and may provide clues to the role of PHB in-the physiology of other free-living, dinitrogen-f ixing bacteria, such as azospirilla. The route of PHB biosynthesis in A. bei jerinckii has been reviewed by Dawes (1981). The synthesis and degradation of PHB in this microorganism are intimately associated

PAGE 22

15 with intermediates and enzymes of the tricarboxylic acid (TCA) cycle, a system that azospirilla also possess (Okon et al., 1976b). When A. bei jerinckii strain N.C.I.B. 9067 was cultured as a dinitrogen-f ixer with 2.0% (wt/vol) glucose, PHB was deposited towards the end of exponential growth. The cells were unable to use all the available glucose, and PHB synthesis continued during the stationary phase until up to 74% of cell dry weight was PHB. Cultures grown with combined nitrogen rarely contained more than 3.0% of their dry weight as PHB (Dawes, 1981). The initiation of PHB synthesis in the A. bei jerinckii strain in batch culture coincided with the attainment of zero-oxygen concentration. Oxygen limitation was thus suspected to be a critical factor in initiating PHB synthesis. However, the nature of batch broth culture made it hard to separate oxygen effects from possible nitrogenlimitation effects (Senior and Dawes, 1971). Later experiments, using chemostat cultures having carbon, oxygen, or nitrogen limitation, clearly showed that extensive PHB accumulation only occurred under conditions of oxygen limitation (Dawes, 1981). Before the studies reviewed by Dawes (1981), PHB was regarded as being only an endogenous, carbon-and-energy source that benefited cells during starvation. These experiments suggested that PHB could also serve other purposes. The synthesis of PHB seemed to serve as an

PAGE 23

electron sink for excess reducing power (NADH and NADPH ) that accumulated when the cell became oxygen limited, and electron transport to oxygen via the terminal oxidases of the electron-transport chain was restricted (Senior and Dawes, 1971). Later work revealed that the activities of certain enzymes of carbon catabolism in A. bei jerinckii are inhibited by either or both NADH and NADPH. Under oxygen limitation, the concentration of these reduced coenzymes is increased, so that glucose metabolism, operation of the TCA cycle, and net biosynthesis are decreased. Growth can continue at some level, however, if PHB is synthesized and the crucial coenzymes are reoxidized (Dawes, 1981). The synthesis of PHB under oxygen limitation may occur in other bacteria as well (Okon and Hardy, 1983). The quantity of PHB accumulated often greatly increases as the C/N ratio of the growth medium increases. Under such conditions, free-living dinitrogen-f ixers may assimilate the exogenous carbon more rapidly than they can produce reduced nitrogen. As a result, the cells can accumulate large amounts of PHB (Stevenson and Socolofsky, 1966; Dawes and Senior, 1973). The metabolism of PHB is regulated such that PHB accumulates when the supply of exogenous carbon is in excess of the requirements for growth and maintenance, and it is degraded when the supply of exogenous carbon is limited or exhausted (Dawes, 1981), or when balanced growth can again occur (Nickels et al., 1979).

PAGE 24

17 It has been shown that PHB can accumulate in cells that are not growing or proceeding toward cell division, due to limitation of available nutrients (Dawes and Senior, 1973). Nickels et al. (1979) demonstrated this in laboratory microcosms containing oak leaf detritus and estuarine water. Supplementing the nutrients in the water column with carbohydrates, especially glucose, induced a rapid accumulation of PHB without a concomitant increase in microbial biomass. When supplements were added that enabled increases in microbial biomass, PHB levels fell as the polymer was broken down to aid microbial growth. In one study, A. brasilense Sp 7 was grown in batch cultures for up to 14 days in microaerophilic nitrogen-free malate broth (Papen and Werner, 1980). Both nitrogenase activity and PHB synthesis were biphasic. An initial peak of PHB content occurred at day 3, 1 day before the first peak of nitrogenase activity. During the first and maximal peak of nitrogenase activity, there was a decrease in PHB content, possibly due to accumulation of fixed nitrogen allowing use of PHB carbon skeletons for biosynthesis. A second peak of PHB accumulation occurred after the first maximum of nitrogenase activity. The results suggested that A. brasilense Sp 7, like A. bei jerinckii may accumulate PHB when it assimilates exogenous carbon faster than it can fix dinitrogen

PAGE 25

"1 18 Zimmer et al. (1984) found that A. brasilense Sp 7 accumulated PHB when using nitrite as terminal electron acceptor for anaerobic growth. A maximum of 38% of cell dry weight was found to be PHB when less than 3 0 mM nitrite was present. No PHB was accumulated when in excess of 8.0 mM nitrite was made available, indicating the role of PHB as a sink for excess reducing power when other electron acceptors are scarce. It was also found that PHB-rich cells contained less protein than did PHB-poor cells. Azospirillum lipof erum strain Br 17 (ATCC 29709) was found by Volpon et al. (1981) to accumulate nearly 24% of its dry weight as PHB near the mid-logarithmic phase of growth as a dinitrogen-f ixer Near the end of logarithmic growth, PHB synthesis seemed to stop, and the content of PHB declined to 13% of cell dry weight in stationary phase. The PHB metabolism of A, brasilense strain Cd (ATCC 29729) has received considerable study. When this strain was grown in continuous chemostat culture with malate and ammonium chloride, a maximum PHB content of 12% of the biomass was observed under microaerophilic conditions and at intermediate growth rates (Nur et al., 1982). These growth conditions were said to approximate conditions generally encountered in the rhizosphere. The production of PHB was markedly decreased at higher levels of oxygen and higher growth rates. Once again, it was observed that cells

PAGE 26

19 containing high amounts of PHB contained less protein than PHB-poor cells. Recent work by Tal and Okon (1985) has further delineated the roles PHB may play in the physiology of A. brasilense strain Cd. Grown in aerobic batch culture with malate and 2.8 mM NH^Cl, the cells accumulated 40% of their dry weight as PHB after 24 hours, toward the end of exponential growth. When the level of NH^Cl was raised to 15.0 mM, the cells accumulated only 5% of their dry weight as PHB after 24 hours. In both cases, the amount of PHB decreased in stationary phase. In chemostat continuous culture, a maximum of 30% cell dry weight accumulated as PHB when the gas atmosphere was 0.082 mM 0^ (Tal and Okon, 1985). With increasing aeration, the PHB content fell to very low levels. When grown in batch culture as dinitrogen-f ixers the cells accumulated about 75% of their dry weight as PHB. Maximal PHB content was obtained in these experiments when the C/N ratio was about 70. Both the C/N ratio of the medium and the oxygen concentration were found to regulate PHB synthesis. The forms of carbon and nitrogen made available to the cells affected the levels of PHB accumulated (Tal and Okon, 1985). Organic acids, especially pyruvate, were found to elicit PHB formation more than carbohydrates did. Sodium nitrate was found to promote PHB formation more than ammonium chloride did, possibly because nitrate does not

PAGE 27

20 accumulate in the cytoplasm to the same extent as does ammonium. Cells with different contents of PHB were harvested by centrif ugation and resuspended in phosphate buffer to measure viability during aerobic nutrient starvation. By 140 hours, bacteria with abundant PHB reserves had given rise to more than twice as many viable cells as were present in the initial inoculum (Tal and Okon, 1985). During starvation, PHB reserves were degraded quickly but not completely. The initial inoculum contained 40% of its dry weight as PHB. This fell rapidly to about 24% of cell dry weight after 42 hours of starvation. After 130 hours of starvation, the PHB content of the cells was about 20% of cell dry weight. In comparison, cells initially containing only 5% of their dry weight as PHB had only 7% of the original number of viable cells after 130 hours of starvation (Tal and Okon, 1985). Poly-6-hydroxybutyrate was still measurable throughout starvation of these PHB-poor cells, stabilizing at or near 3% of the dry weight of all cells present. Starved PHB-rich cells had a higher respiration rate during starvation than the PHB-poor cells (Tal and Okon, 1985). Unlike cells having low amounts of PHB, the PHB-rich cells exhibited nitrogenase activity in the absence of exogenous carbon sources. But the PHB-rich cells were as unable to reduce nitrate anaerobically as were the PHB-poor cells in the absence of exogenous carbon.

PAGE 28

This study (Tal and Okon, 1985) also suggested that elevated PHB levels at the onset of starvation may spare the use of protein to drive endogenous metabolism. The PHB-poor cells used up two-thirds of their initial protein during the first 80 hours of starvation, whereas the protein content of starved PHB-rich cells increased slightly over 80 hours. It was also reported that PHB-rich cells were able to survive a variety of environmental stresses, including desiccation, better than PHB-poor cells (Tal and Okon, 1985). The previous study also found that cells enriched in PHB displayed a one hundred-fold higher aerotactic response than PHB-poor cells. This supports the claim made in an earlier study that PHB reserves could be used for aerotaxis when no exogenous carbon source was available (Barak et al., 1982 ) The previous discussion has shown that both capsule and PHB synthesis can be promoted by environments with high available C/N ratios. The roles of capsules and PHB in pleomorphism in azospirilla will be discussed later. The nature of dormancy in prokaryotic cells will be discussed first, since some pleomorphic forms of azospirilla may be dormant stages. Capsular layers and PHB are often present in dormant forms of prokaryotes.

PAGE 29

22 Dormant Forms of Prokaryotic Cells There is general agreement that most soil bacteria spend much of their existence in soil in a state of low metabolic activity. The low respiratory rates of bulk samples of nonamended soil support this (Clark, 1967). Many soil bacteria may be metabolically dormant due to a lack of readily available carbon and energy supplies (Gray and Williams, 1971). Soil bacteria may enter into exogenous dormancy, where growth is delayed by unfavorable physical or chemical conditions (Marshall, 1980). Such bacteria probably have the same morphology as actively growing vegetative cells (Gray and Williams, 1971). These cells are probably intimately associated with the clay or organic matter of soil. The cells adsorb to these surfaces by physical or chemical interactions, or by the use of exopolysaccharides (Stotzky, 1980). However, many bacteria may exist in soil as dormant forms that are morphologically different from their growing, or vegetative, stages. These cells would have entered a phase of constitutive dormancy, involving the formation of spores or cysts (Marshall, 1980). Bae et al. (1972) used transmission electron microscopy to study thin sections of bacteria released from soil by centrif ugation and washing. About 28% of the bacteria observed had normal vegetative morphology, of which 29% possessed capsular layers.

PAGE 30

23 Bacterial cells resembling cysts comprised 27% of cells observed Sudo and Dworkin (1973) reviewed the kinds of prokaryotic resting cells that were recognized at the time. Bacterial resting cells were defined as cells in which division does not occur, and endogenous respiration is absent or greatly reduced. Usually resting cells are more resistant to environmental stresses than are vegetative cells; resting cells are often morphologically different as well. Such resting cells often differ in chemical composition from vegetative cells (Keynan, 1972). There are often either qualitative or quantitative differences between the electron transport systems of vegetative and resting cells. Many resting cells, for some period after they have germinated and resumed growth, are self-sufficient in energy sources, metabolites, and macromolecular precursors. Perhaps the best understood bacterial resting cell stages are the endospores of bacilli and Clostridia. Cysts differ from endospores in that they are formed by the modification of an entire vegetative cell. The vegetative cell rounds up during encystment and becomes coated with one or more layers, often exopolysaccharide external to its cell wall. No cyst forms withstand the extremely high temperatures tolerated by endospores, but they are comparably resistant to other environmental stresses (Sudo and Dworkin, 1973).

PAGE 31

24 Certain properties are shared by all cyst-like forms of prokaryotes. They are formed when the growth rate of vegetative cells declines (a metabolic shift-down), due either to nutrient depletion or transfer of cells to an environment where balanced growth can no longer occur. Cells encountering these conditions complete their ongoing synthesis of DNA and chromosome replication but do not initiate new rounds of DNA synthesis, since growth has ceased (Sadoff, 1975). Conditions that will prohibit further growth promote the formation of dormant cells that can survive stress better than vegetative cells. These dormant cells often contain PHB or other energy-reserve polymers, and thickened cell walls or capsular layers. They have enhanced resistance to irradiation, sonic vibration, and sometimes elevated temperatures. Perhaps the most important traits for survival of dormant cells in natural environments are their resistance to starvation, low endogenous respiration rates, and desiccation resistanceMany cells entering constitutive dormancy need time to mature before they achieve maximal resistance to stress. It is important to remember that resting cells formed in natural environments may differ qualitatively and/or quantitatively in their resistance properties from those formed under laboratory conditions (Sudo and Dworkin, 1973). Many strains of Gram-negative myxobacteria form dormant cells, called microcysts, when nutrients become limiting.

PAGE 32

25 Nonmotile, encapsulated, mature microcysts are more resistant than are vegetative cells to environmental stresses such as ultraviolet irradiation, sonic vibration, and desiccation ( Sudo> and Dworkin, 1973). Many Gram-negative, methane-oxidizing bacteria isolated from soil or mud are encapsulated and accumulate PHB when nitrogen becomes limiting for growth (Whittenbury et al., 1970b). Depending on the genus and strain, up to 90% of the cells present may form resting cells upon entering the stationary phase of growth. Lipid cysts of Methylocystis parvus accumulate large amounts of PHB and survive starvation and desiccation better than vegetative cells, but lack well-defined, capsular cyst coats. Methylomonas spp. and Methylococcus spp. form rounded, nonmotile cells that survive starvation better than do vegetative cells. These cells are called immature cysts, because they never attain desiccation resistance. Some strains of Methylobacter spp. form starvation-resistant and desiccation-resistant cysts that seem morphologically identical to Azotobacter spp. cysts (Whittenbury et al., 1970a). Bdellovibrio sp. strain W is the only bdellovibrio that is known to encyst. Bdellocysts are larger than their vegetative counterparts and are not light-ref ractile They tolerate sonic disruption, ultraviolet irradiation, and carbon starvation better than do vegetative cells. The endogenous respiration rate of bdellocysts is 80% less than

PAGE 33

26 that of vegetative cells. When dried over silica gel desiccant under slight vacuum in glass tubes, vegetative cells of strain W die out rapidly and entirely. From 45% to 80% of bdellocysts initially present are able to survive 6 days of this desiccation treatment (Tudor and Conti, 1977). Bdellocysts possess a thickened outer layer of modified peptidoglycan, and contain inclusion bodies of an amylopectin-like polysaccharide of glucose monomers. These features are not found in vegetative cells (Tudor, 1980). Some strains of Azotobacter spp. apparently some azospirilla, and the bacteria described above are the only prokaryotes reported to form cysts. Why do not more bacteria possess resting stages that are morphologically differentiated into cysts? Perhaps growth media and conditions used in the laboratory discourage cyst formation (Whittenbury et al., 1970b). It is also possible that the ability to form cysts is sometimes labile and may be lost upon subculture. One Methylobacter chroococcum strain was able to form multiple-bodied cysts upon initial isolation from the environment. It ceased to do so when subcultured. Other Methylobacter spp. have retained the ability to form singleand multiple-bodied cysts over several years of subculture (Whittenbury et al., 1970b). A mature, cyst-like cell of a prokaryote may perhaps best be characterized as follows. Mature cysts differ morphologically from vegetative cells in having thickened

PAGE 34

27 outer layers. They are nonmotile and have low endogenous respiration rates. They only initiate growth into vegetative cells when sufficient nutrients are available. They must also possess more resistance to some environmental stresses than do vegetative cells. Enhanced resistances to starvation and desiccation are probably traits of all mature cysts. The cysts of the methane-oxidizing bacteria and of Azotobacter spp. possess these characteristics. Mature cysts of azospirilla should also have these properties. Desiccation resistance is a critical characteristic of prokaryotic cysts. The next section will consider experiments conducted to assess the resistance of bacteria to drying Resistance of Bacteria to Drying Clark (1967) stated that the majority of soil bacteria survive in air-dried soils, often for several years. When such soils are rewetted, bacterial activities including nitrification, ammonif ication nonsymbiotic dinitrogen fixation, and sulfur oxidation are usually detected. The implication is that the intimate association of bacteria with clay or organic matter allows bacterial survival at hydrated microsites in a macroscopically dry soil. Later findings, reviewed by Stotzky (1980) and Marshall (1980), support this. Exopolysaccharides may help bacteria to achieve such intimate association, although capsules

PAGE 35

28 themselves have not been found to afford any desiccation resistance in laboratory studies with pure cultures in nonsoil conditions (Dudman, 1977). Because of the importance of desiccation as a limiting factor in legume inoculation with Rhizobium spp. several studies have been done on their resistance to drying. There are broad strain differences in resistance of rhizobia to desiccation. Many variables are present in drying experiments, and the variables may interact with one another. Rhizobium spp. withstand drying best in heaviertextured soils, where hygroscopic water can be retained by colloidal surfaces. Die off is far more rapid in drying sand. Capsules do not afford increased resistance to drying in studies with soil or other drying surfaces (Lowendorf, 1980). Often fewer rhizobia survive rapid drying procedures, such as oven drying, than survive milder desiccation over several weeks' time with controlled relative humidities (Jansen van Rensburg and Strijdom, 1980). Robinson et al. (1965) added pure cultures of Pseudomonas spp. or Arthrobacter spp. to sterile soils. The inoculated soils were dried by passing filtered air through them for 2 days, by which time they had reached constant weight. This forced drying resulted in rapid die off for both species. Labeda et al. (1976) found that slow evaporative drying of inoculated soil resulted in reduced death rates for both Pseudomonas spp. and Arthrobacter spp.

PAGE 36

29 These experiments show clearly that rate of drying can profoundly affect the survival of vegetative bacteria. Vegetative cells with the capacity to become desiccation resistant may need time to alter their membrane or cytoplasmic composition before desiccation resistance is achieved. Fast-drying procedures may not allow them to do so. A differentiated resting cell, such as a cyst, may also need time, depending on how mature it is, to become desiccation resistant. Relative humidity (RH) also has a great influence on desiccation resistance of prokaryotes. In desiccation at any RH below 90%, the free water of the cells is removed almost instantaneously. The water that remains is the bound water content of the cell, which may be necessary for continued function of essential metabolic processes and viability. Often few vegetative bacteria die when desiccated above 70% RH, but die rapidly as the RH declines to 45% (Webb, 1965). Many desiccation studies have not defined the RH at which the cells were dried, making duplication of results difficult. Thompson and Skerman (1979) tested the desiccation resistance of vegetative cells of many strains and genera of the Azotobacteraceae. One milliliter samples of vegetative cell cultures were added to sterile porcelain beads, positioned above silica gel in glass bottles sealed with Parafilm. These desiccation units were stored at room

PAGE 37

30 temperature, and at different times single beads were aseptically removed and placed in broth media. The bacteria were probably in stationary phase when added to the assemblies, but it is unlikely that many cysts were present even in stationary phase broth culture (Sadoff et al., 1971). The results were surprising; the majority of strains retained viability for 1 to 2 years of desiccation. This was true even for bacteria that have never been shown to form cysts. Mature cysts of prokaryotes survive rapid desiccation on glass surfaces far better than do their vegetative counterparts, but rarely does all the encysted inoculum survive rapid drying. Cysts of methane-oxidizers retained 60% to 90% viability after 1 week (Whittenbury et al. 1970a), and bdellocysts retained 45% to 80% viability after 6 days (Tudor and Conti, 1977). This may mean that not all the encysted cells were fully mature when exposed to drying, even if they all appeared morphologically identical. Such quick-drying assays can be valuable in determining whether morphologically differentiated cells are truly cyst-like. Differences in the desiccation resistance of Azotobacter spp. vegetative cells and cysts are usually determined by the method of Socolofsky and Wyss (1962). They impinged suspensions of either cell form on the surfaces of membrane filters. The filters were then transferred to dry adsorbent pads in Petri dishes and placed in

PAGE 38

31 an incubator at 33C. This method is a slow-dying procedure. At different time intervals, the cells were washed from the membranes, and viability was determined by plating. Cysts of Azotobacter vinelandii ATCC 12837 lost little viability over a 12-day-period using this drying treatment, whereas 99% of the vegetative cells were killed by the end of the first day (Socolofsky and Wyss, 1962). As a result of the rapid die off of vegetative cells with this treatment, later studies considered cells of this strain to be cysts if they could withstand 4 days of desiccation on membrane filters (Stevenson and Socolofsky, 1966; Wyss et al., 1969). None of these membrane filter studies specified the RH at which the membranes were dried, or how many cell layers were deposited upon the membranes. Webb (1965) pointed out that if bacteria are dried on filters to test their desiccation resistance, the cells must be applied in a monolayer to achieve consistent results. If more than a cell monolayer is on the filter, most of the cells in subsurface layers will not be dried or equilibrated with the water vapor of the environment. Most desiccation resistance experiments have given ill-defined or incomplete conditions of drying. Such experiments have proven, however, that cysts are more desiccation resistant than are their vegetative counterparts

PAGE 39

32 Vela (1974) tested desiccation resistance of Azotobacter vinelandii ATCC 12837 by allowing slow drying of the agar on which the cells were grown. Vegetative cells were grown on agar plates of Burk's nitrogen-free medium, with glucose as the carbon source. Cysts were obtained by growing the cells on the same agar, except that 0.3% (vol/vol) n-butanol was employed as sole carbon source. Dried agar films were then broken with sterile forceps and placed on the surface of Burk's agar medium containing glucose. Vegetative cells borne on these agar films remained viable for nearly 2 years of desiccation, whereas cysts borne on such films remained viable for 10 years or longer. Desiccation tolerance of azospirilla has received some attention. Lakshmi et al. (1977) recovered azospirilla from several air-dried soils stored in the laboratory. Recovery was obtained from one of four sandy soils stored air-dry for 10 years. All of these soils had less than 0.5% organic matter. Heavier-textured soils with 1.0% or more organic matter consistently yielded isolates of azospirilla. Some of these heavier-textured soils had been stored air-dry for up to 15 years. It was suggested that organic matter aids the survival of azospirilla in drying soils, and that desiccation-resistant cells may be formed by these bacteria. Jagnow (1982) did some work with an Azospirillum lipof erum strain isolated from maize roots. In field

PAGE 40

33 g inoculations using 8 x 10 CFU/g soil, azospirilla near or on roots survived better than those in soil distant from roots. When added to pots of soil containing grass and 6 7 cereal plants, populations remained at 10 to 10 CFU/g soil, even after 70 days of drought. He speculated that the presence of roots, either living or dead, enhances the drought tolerance of the associated azospirilla. In laboratory studies using nonautoclaved soil microcosms, air drying of soil was found to kill greater than 99% of the initial Azospirillum lipof erum inoculum. In comparison, the indigenous bacteria were little affected by air drying. This perhaps indicates that, unless azospirilla added to soil are able to associate quickly with plant roots, they will soon die out if drought stress occurs. The desiccation resistance of pleomorphic encapsulated forms of azospirilla has been studied. Lamm and Neyra (1981) studied A. brasilense Sp 7 and A. lipof erum Sp 59b, in addition to several strains of azospirilla isolated from roots of various grasses in New Jersey and New York. To obtain cyst-enriched cultures, cells grown in nutrient broth were harvested by centrif ugation then washed and resuspended in sterile 0.85% (wt/vol) NaCl. A 1.0 ml sample of cells was then spread plated as a lawn onto nutrient agar plates containing 2.0% (wt/vol) agar. Plates were incubated at 30C until the agar was dried into a thin film, often requiring a month. After 15 days of incubation, cyst-like

PAGE 41

34 cells predominated. Photographs of cyst-enriched cultures showed that many vegetative cells were still present. To obtain cyst-free cultures, cells were grown in nutrient broth, then washed and resuspended in saline. These cells were then spotted onto sterile, predried nutrient agar films so that the added cells would dry completely on the agar film in 30 min at 30C. Agar films from each treatment were then cut with sterile scissors and aseptically transferred to vials containing silica gel. To test viability, the dried agar films were removed periodically from the vials, placed on nutrient agar plates, and incubated for 1 week at 30C. Vegetative cells did not survive the initial drying process. Cyst-enriched populations that survived the initial desiccation period remained viable for up to 15 months. Interestingly, cyst-enriched cultures of two root isolates were nonviable at time zero, when they were placed into the silica-gel vials (Lamm and Neyra, 1981). Two aspects of this study deserve special comment. Clearly, the cyst-enriched cultures did not receive the same drying treatment as did the vegetative cells. The cystenriched agar films were obtained by a slow drying process, and the vegetative cell agar films underwent rapid drying. It does not seem valid to compare their desiccation tolerance under these different conditions. Also, two strains that contained cyst-like cells of apparently mature morphology were not desiccation resistant. Perhaps they

PAGE 42

35 were not able to attain physiological maturity under the experimental conditions. Papen and Werner (1982) assessed the desiccation resistance of cyst-like forms of A. brasilense Sp 7. Cells from dinitrogen-f ixing broth cultures were diluted in sterile tap water and then impinged onto the surface of sterile 0.2 jam Millipore membrane filters under vacuum. Some of the filters were immediately placed on the surface of nutrient agar plates and incubated at 28 C, whereas others were placed on sterile adsorbent pads in Petri dishes and dried at 37 C until they were placed on nutrient agar plates. Desiccation-resistant cells were only present after the first peak of nitrogenase activity, when nonmotile, encapsulated spheres containing PHB predominated. Cells before and during the first peak of nitrogenase activity were motile vibrioids and did not survive the desiccation treatment. As a second peak of nitrogenase activity arose, motile, dinitrogen-f ixing vibrioids emerged from the spherical capsules; these vegetative cells were again not desiccation resistant. More encapsulated, spherical cells survived 2 days of desiccation than 6 days, but it was not an order of magnitude difference. This again may be an indication that morphologically mature cysts are not necessarily physiologically mature. The recent work of Sadasivan and Neyra (1985) employed another assay for desiccation resistance of cyst-like forms

PAGE 43

36 of azospirilla. Azospirillum brasilense Sp 7 and A. lipof erum Sp 59b were studied. Large floes of cells enclosed in exopoly saccharide were placed on Whatman No. 1 filter paper and air-dried for 30 min. They were then placed in a closed vial, without desiccant, and incubated at 30C for up to 6 months. Small pieces of dried floes were transferred periodically to semisolid nitrogen-free malate medium and incubated at 3 4C for 2 to 4 days, and growth, pellicle formation, and nitrogenase activity were observed. Cells in dried floes remained viable for up to 6 months of drying. No vegetative cell controls were dried and tested for viability in the above study. Although cells remained viable in dried floes for up to 6 months, it is not known how many cells survived in a given amount of floe. It is not known whether the cells themselves were desiccation resistant, or only physically protected from desiccation by exopolysaccharides Tal and Okon (1985) claimed that PHB-rich cells of A. brasilense strain Cd were 10 times more desiccation resistant than cells having little of the polymer. No details of the test used for desiccation resistance were given. Desiccation resistance studies can be difficult to interpret. Comparing the desiccation resistance of vegetative cells to that of cysts may be less difficult than

PAGE 44

37 comparing that of vegetative cells of different strains. Rapid drying on a glass surface should enable most mature cysts to remain viable, but not most vegetative cells. A glass drying surface should be less hygroscopic than are membrane filters or agar films. Rapid drying on a glass surface is a severe treatment, but it should reveal the presence of physiologically modified, stress-resistant cells, such as mature cysts. Azotobacter Cysts Discussion of the nature of Azotobacter spp. cysts is important, because this information served as the basis for the experiments with azospirilla reported in this study. Like azospirilla, the Azotobacteraceae are Gram-negative aerobes, often containing PHB granules. Many are motile by flagella. They all fix dinitrogen, and some, including Azotobacter spp. do so either at atmospheric oxygen levels (unlike azospirilla), or as microaerophiles Only one genus, Azotobacter contains species with strains that are known to form cysts (Tchan, 1984). The isolation of Azotobacter spp. from the interior of 2,000-year-old clay bricks ( Abd-El-Malek and Ishac, 1966), and their persistence in soils that had been air dried from 10 years (Vela, 1974) to 30 years (Clark, 1967), may be due largely to cyst formation.

PAGE 45

38 When grown in nitrogen-free broth with glucose as the carbon source, young cells of Azotobacter spp. appear as rods with rounded ends, ranging from 1.3 to 2.7 um in diameter and 3.0 to 7.0 pun in length. As cultures age, cells often accumulate PHB. Cell morphology may be altered to ellipsoids, filamentous cells, or chains of cells (Tchan, 1984 ) Azotobacter spp. are commonly isolated from soil and aquatic habitats of near-neutral pH, and are generally less acid-tolerant than azospirilla. The most common species isolated from soil is Azotobacter chroococcum but its biochemistry and physiology have received less attention than that of Azotobacter vinelandii (Tchan, 1984). Azotobacter vinelandii ATCC 12837 forms cysts profusely under appropriate growth conditions. When this strain is cultured in Burk s nitrogen-free broth with glucose, some cysts form in stationary phase cultures, but ony 1.0% (Lin and Sadoff, 1969) to 10.0% (Reusch and Sadoff, 1981) of the population encysts under these conditions. Early workers such as Winogradsky (1938) knew that growing some Azotobacter spp. in nitrogen-free media, with ethanol or butanol as carbon source, led to enhanced production of nonmotile, spherical cells with double-layered coats. Socolofsky and Wyss (1961) built upon this knowledge, using A. vinelandii ATCC 12837 (which' was used in all the studies that follow unless otherwise indicated).

PAGE 46

when cultured as cell lawns on Burk s nitrogen-free agar with 0.3% (vol/vol) n-butanol as sole carbon source, cysts began to appear within 3 days and predominated in 5 to 7 days. Ultrastructural studies revealed that the outermost layer of the cyst, the exine, consisted of several overlapping, plate-like layers. Beneath the exine was a much thicker layer of gelatinous material, called the intine. The intine surrounded a modified resting cell, called the central body, which often contained numerous PHB granules. Cysts had no detectable endogenous respiration when suspended in buffer, but almost instantaneously began measurable respiration when exogenous carbon sources were added. In later studies, cysts were produced by growth on 0.2% (vol/vol) n-butanol (Socolofsky and Wyss, 1962), or 0.2% (wt/vol) B-hydroxybutyrate (BHB) (Lin and Sadoff, 1968), with 90% or greater of the cells being converted to cysts in 5 to 7 days. Eklund et al. (1966) demonstrated that the formation of capsular layers by vegetative cells was a prerequisite for cyst formation. Complete morphological encystment of cells grown on n-butanol agar with various levels of NH^NO^ only occurred in the usual 5-day-period when the NH^NO^ concentration was 0.02 M or less. The cells rounded up within 5 days into nonmotile precysts lacking exines when 0.03 M or 0.04 M NH^NO^ was initially present. By day 10, these cells had used up enough of the original combined nitrogen to

PAGE 47

40 allow dinitrogen fixation to resume, so that capsular polysaccharide was produced, followed by formation of exines and, ultimately, morphologically mature cysts. Nonencapsulated mutants were unable to form morphologically mature cysts. The work of Pope and Wyss (1970) emphasized that cells beginning encystment first produced a capsule that acted as a structure within which the cyst coats were built, so that the exine existed inside of the capsule. The diameter of morphologically mature Azotobacter cysts measured between exine boundaries is about 2.0 urn (Reusch and Sadoff, 1983). Abortive encystment occurs when cells round up into nonmotile precysts, but are unable to form a complete exine. This occurs in the presence of high amounts of combined nitrogen (Eklund et al. 1966), when glucose or other carbon sources are present in addition to n-butanol or BHB (Lin and Sadoff, 1968), or when calcium is unavailable. The calcium requirement is probably related to its function as a stabilizing cation that holds the cyst coats together (Page and Sadoff, 1975). Using 3 0 mM EDTA in 0.05 M Tris buffer, pH 7.8, Lin and Sadoff (1969) obtained almost instantaneous expulsion of the central body from the cyst coats, due to the chelating effect of the buffer. The empty exines had the same "horseshoe" shape seen when cysts germinate, and vegetative cells separate from the exines.

PAGE 48

41 The role of PHB in cyst formation was examined by Stevenson and Socolofsky (1966). Cysts were defined as cells that could survive desiccation on a membrane-filter surface for 4 days at 33 C. After 2 days of growth on nitrogen-free n-butanol agar, cells lost their motility, became oval-shaped, and accumulated PHB to the extent of 35% of cell dry weight. The development of mature cysts was accompanied by a reduction in PHB content. By 6 days, cultures had undergone 100% encystment, and 10% of cyst dry weight was PHB. Lin and Sadoff (1968) developed a two-step replacement procedure for obtaining cysts in broth. Cells were grown to late exponential phase in Burk s nitrogen-free broth with glucose. After harvest by centrif ugation and washing in buffer, cells were resuspended in Burk's salts broth with 0.2% (wt/vol) BHB. This procedure was used in further studies (Hitchins and Sadoff, 1970, 1973; Reusch and Sadoff, 1979; Su et al. 1981; Reusch and Sadoff, 1983), resulting in the following detailed description of BHB-induced encystment. After 1 hour in encystment broth, cells are still motile and flagellated but no longer possess nitrogenase activity. Within 4 to 6 hours, DNA synthesis has ceased, and soon afterward each cell divides to form two nonmotile precysts. There is rapid accumulation of PHB during this period, and the rate of phospholipid synthesis declines.

PAGE 49

42 Simultaneously, BHB is being taken up and respired or incorporated. From the sixth to sixtieth hour, unique lipids, not found in vegetative cells, begin to be produced. These include 5-n-alkylresorcinols (ARl) and their galactoside derivatives (AR2). These lipids possess hydrophobic alkyl sidechains and hydrophilic phenolic heads. Also produced are 6-n-alkylpyrones (AP), having a similar bipolar nature. During this time, membranous vesicles migrate outward from the central body through the intine to form the exine layer. Up to 17% of the exine is composed of ARl and AR2 The central body produces ARl and AR2 in part from its PHB reserves, and exports them in the membranous vesicles to the exine region. Radio-labelled BHB accumulates in the central body and exine, whereas the intine contains almost none. This indicates that the intine is composed mainly of capsular material, formed from cell reserves that were present before encystment is triggered by BHB. Net RNA synthesis stops by the twelfth hour, and net protein synthesis continues for up to 36 hours. Lipid turnover continues beyond 60 hours, but there is no net lipid synthesis. In a mature cyst, 5.0% of the central body membranes are phospholipid, with AR and AP composing the other 95%. Molecular models suggest that AR and AP form a more rigid membrane structure at physiological temperatures than do phospholipids. The hydrophobic, viscous nature of

PAGE 50

43 such a membrane may contribute greatly to the desiccation resistance and dormancy of cysts. The possible contribution of the central body membranes to stress resistance of cysts was suggested in earlier studies. The cysts of Azotobacter chroococcum strain 75-1 had a compact, well-defined exine layer, whereas the exine of A. chroococcum strain NTS was diffuse and fragile (Vela and Cagle, 1969). The cysts of A. chroococcum strain 75-1 were much more resistant to sonic disruption than cysts of A. chroococcum strain NTS. Yet cysts of both strains were comparably resistant to desiccation on membrane filters and to ultraviolet irradiation. Kramer and Socolofsky (1970) defined cyst germination of A. vinelandii ATCC 12837 as a process whereby desiccation resistance is lost; mature cysts were defined as cell forms surviving 3 days of desiccation on membrane filters. It was found that 40.0 ug chloramphenicol/ml inhibited outgrowth of cysts in a complete medium. Many cysts lost their desiccation resistance when incubated with chloramphenicol, indicating that the antibiotic might have chemically changed some essential cyst component, perhaps the central body's membranes. Hitchins and Sadoff (1973) found that, soon after exposure to BHB, vegetative cells became resistant to 100.0 ug chloramphenicol/ml. The antibiotic had no effect on morphogenesis or rates of protein synthesis. This is another indication of rapid membrane alteration of encysting

PAGE 51

44 cells, long before AR and AP are produced. Further support for the importance of membranes may be found in studies where mineral nutrient deficiencies lead to the production of stress-resistant cysts which lack completed cyst coats (Gonzalez-Lopez et al. 1985). Germination of cysts has usually been defined as the emergence of a growing, motile cell from the exine layer (Socolofsky and Wyss, 1961). Loperfido and Sadoff (1973) examined the germination of cysts exposed to glucose. Cysts respired detectably within 2 min. after the addition of 1.0% (wt/vol) glucose, and soon afterwards net synthesis of RNA and protein became measurable. After 4 to 6 hours, the central body had enlarged to occupy the volume of the intine, and DNA synthesis and nitrogenase activity became measurable. After 8 hours, a vegetative cell emerged from the cyst coats, leaving behind an empty "horseshoe"-shaped exine. Germination did not occur in the absence of oxygen. Cysts also germinated in the presence of sugars other than glucose. Germination did not occur in Burk's nitrogen-free salts, indicating that the PHB reserves of the cysts could not be mobilized to initiate germination. The addition of 0.25% (wt/vol) NH^"*" did not lead to germination. When cysts are germinated on glucose, some central bodies divide within their cyst coats to form multiple central bodies. Up to six central bodies have been observed within one cyst coat (Cagle and Vela, 1974).

PAGE 52

45 Pleomorphism of Azospirilla Bacteria cultured in vitro can be extremely pleomorphic. Only a few cells in a population may exhibit abnormal morphology under some cultural conditions, but sometimes the majority of a culture assumes unusual shapes. Older cultures in the stationary growth phase can be especially pleomorphic (Duguid and Wilkinson, 1961). Hughes (1956) has reviewed the development of bacterial filaments. Filamentous cells are usually as wide as normal cells, but are several times longer and lack developed septa. They are interesting because they are often fully viable, unlike some pleomorphic or involution forms of bacteria. Under suitable cultural conditions, a filament may divide at several points along its length to produce several cells of normal length. Filaments can be induced by sublethal cell damage, interruption of balanced growth, dyes and antibiotics, extremes of pH, refrigeration, and various forms of radiation. Slater and Schaecter (1974) emphasized how sensitive bacterial cell division is to the factors mentioned above. If sublethally stressed, rod-shaped bacteria may continue to grow and form filaments. Filaments can also form during very rapid growth in rich media, and will fragment into individual cells when growth slows, or when the environment becomes less nutritionally rich. Since cells arising from fragmentation of filaments are usually of normal length, the

PAGE 53

46 cell's ability to control the site of cell division is not lost during f ilamentation Sometimes chains of cells occur instead of filaments. The cells in chains contain septa, but final cleavage between cells has not yet occurred. It is possible that contiguous cells having incomplete septa in such chains may share continuities between their cytoplasms. In some cases, chains may be held together by very thin capsular layers common to several cells in the chain. Jensen and Woolfolk (1985) found that several strains of Pseudomonas putida and Pseudomonas f luorescens were induced to form filaments if oxygen became limiting during the late logarithmic phase of growth in nutrient broth. Exhaustion of one or more nutrients was also a probable elicitor of f ilamentation. The weakly motile filaments, unlike the highly motile aerobic rods of the bacteria, migrated to microaerophilic zones. As respiration of cultures declined, the increasing levels of oxygen in the broth seemed to trigger fragmentation of the filaments into rods. Cultures containing filaments, or the progeny of fragmented filaments, retained viability longer than nonf ilamentous cultures Morphological changes in Escherichia coli have been related to specific genes. If cellular DNA is damaged by ultraviolet irradiation or other influences, several genes are expressed in the so-called SOS response. Many of the

PAGE 54

47 gene products are involved directly in repair of damaged DNA, but some others specifically block further cell division. Until the DNA is repaired, cell division is blocked, but cells can continue to grow into long, nonseptate filaments. Upon repair of the DNA, septa form along the filaments, and cells of normal size are produced after septum separation (Donachie et al. 1984). Certain E. coli mutants are known to produce septa, but form chains because the enzymes needed for septum cleavage are not produced ( Begg and Donachie, 1985). Thompson and Skerman (1979) showed that most members of the Azotobacteraceae are pleomorphic under certain cultural conditions. Filaments and chains of cells are produced commonly. Similar pleomorphism has been observed with azospirilla Becking (1982) observed that the morphology of azospirilla varied in different culture media. On yeast extract-glucose agar, the cells were highly motile, slightly curved rods, 2.0 to 4.0 um long and 1.0 \m wide. These cells would often become swollen with three to five PHB granules per cell. When cultured in nitrogen-deficient broth supplemented with 0.01% (wt/vol) Difco yeast extract, the cells often became long spirals of 30 to 40 um in length. These cells had reduced motility, but were capable of rotation about their axes, and had few or no PHB granules. Peptone was found to produce similar elongated.

PAGE 55

48 weakly motile cells containing little PHB. These cells were probably filaments, as described by Hughes (1956). Becking did not study their viability. Eskew et al. (1977) isolated and studied the pigmented A. brasilense strain Cd. Nitrogenase activity peaked after about 2 days of growth in semisolid, nitrogen-free malate medium, and most cells were motile, curved rods of normal size, often containing PHB. After 3 days, however, nitrogenase activity declined sharply. By this time, the initial near-neutral pH of the medium had risen to pH 8.1. Most of the bacteria present then appeared as enlarged, ovoid, nonmotile cells that were resistant to Gram-staining. The decrease in nitrogenase activity, and shift to alkaline pH, coincided with the appearance of cyst-like cells. Tarrand et al. (1978) found that A. brasilense and A. lipof erum strains had a similar appearance after 1 day's growth in broth containing peptone, ammonium sulfate, and succinate. Most cells were short, plump, slightly-curved motile rods averaging 1.0 um in diameter and 2.1 to 3.8 um in length. Cell morphology changed, especially for A. lipof erum strains, when the cells were inoculated into nitrogen-free, semisolid malate medium containing 0.005% (wt/vol) yeast extract. Cells of A. lipoferum tended to increase to 1.4 to 1.7 um in width and to 5.0 ym to over 30 ym in length. Within 1 to 2 days, many A. lipoferum cells became S-shaped or helical and retained little if any

PAGE 56

49 motility. These long cells eventually fragmented into shorter, ovoid cells. Many of these fragments later became large, pleomorphic cells filled with light-ref ractile granules, probably PHB. In contrast, A. brasilense strains transferred to nitrogen-free, semisolid malate medium initially retained their normal appearance. Only after several weeks' time in this medium did they develop some S-shaped cells and some large, pleomorphic, granule-filled forms. Falk et al. (1985) found that A. amazonense strains failed to become pleomorphic under comparable conditions. Krieg and Dobereiner (1984) maintained that alkalinization of the medium due to oxidation of malate was responsible for pleomorphism in A. lipof erum Cultures of this species grown in semisolid, nitrogen-free glucose medium did not become alkaline, and the cells did not become pleomorphic Wong et al. (1980) isolated a putative Azospirillum sp. from cellulolytic, dinitrogen-f ixing mixed cultures. In semisolid, nitrogen-poor malate medium containing adequate levels of biotin, the cells were of normal size and morphology after 1 day's growth. Between the third and seventh days, the cells gradually became S-shaped and enlarged. These enlarged cells contained granules of PHB and/or polyphosphate. By 10 days, many cells had lysed and released these granules into the medium. When the initial biotin concentration of the medium was reduced to 10% of the

PAGE 57

50 normal level, these morphological changes were accelerated, occurring within 2 to 3 days after inoculation. This strain could not fix dinitrogen with glucose as the carbon source, but otherwise its biotin requirement and pleomorphism were typical of A. lipof erum Lamm and Neyra (1981) found that A. lipof erum strains grown in nitrogen-free, semisolid malate medium developed many elongated cells after 2 days of culture, whereas A. brasilense strains only developed elongated cells after 10 days. In both semisolid and agar-plate, nitrogen-free malate cultures, thick-walled, optically refractile cells were present as 1.0% of the cells in day-old cultures of both species. After 4 days, the numbers of these thickwalled cells were equal to or greater than cells of normal morphology. The ovoid cells of A. lipof erum strains were about twice as large as those produced by A. brasilense strains. Such cells were never observed in nutrient broth cultures, but could be obtained in old cell lawns grown on nutrient agar. The increased desiccation resistance of these ovoid cells has already been discussed. Papen and Werner (1982) observed apparent cysts of A. brasilense Sp 7. Such nonmotile cells were encapsulated, having a diameter of about 1.2 um, were not fixing dinitrogen, and were desiccation resistant. They composed the majority of cells in the nitrogen-free, malate broth culture between the second and fourth days of incubation. After

PAGE 58

51 this time, vegetative vibrioids emerged from the capsules to grow and fix dinitrogen. The authors suggested that oxygen limitation greatly affected these events. The level of PHB increased as the oxygen level of the culture decreased; nitrogenase activity ceased; and the cells encysted for a time. Their apparent reduced respiratory activity allowed the level of dissolved oxygen to be replenished in the medium, until vibrioids emerged from the cyst coats to grow and fix dinitrogen again. No encystment was observed when cultures were incubated aerobically. The recent work of Sadasivan and Neyra (1985) stressed the roles that PHB and exopolysaccharides play in cyst formation of azospirilla. Encysting cells lost their motility and became enlarged and rounded. They accumulated PHB and synthesized capsular material. The investigators emphasized that common media, such as nutrient broth, do not promote encystment and that development of mature exine and intine layers may only be achieved under specific, welldefined cultural conditions. Sadasivan (1985) may have found the cultural conditions to promote maturation of cysts f ^brasilense Sp 7. Using phase contrast microscopy, she has observed vegetative cells emerging from cyst coats, leaving behind empty "horseshoe"-shaped capsules. She has also observed cysts containing from two to four central bodies within a single exine. In transmission electron microscopy thin sections, she has observed maturing cysts.

PAGE 59

52 with membranous blebs migrating outward into the capsular material from central bodies containing PHB granules. She has also observed mature cysts with central bodies containing PHB and polyphosphate granules, surrounded by distinct intine and exine layers. Thus, given appropriate cultural conditions, A. brasilense Sp 7 is able to form apparently mature cysts, almost identical in appearance to those of Azotobacter spp. One unusual feature she has reported is layers of spherical, melanin-like granules outside the exines of mature A. brasilense cysts; these layers have never been observed with Azotobacter spp. cysts. Berg et al. (1980) studied morphological and physiological changes of A. brasilense Sp 7 grown under different conditions. Encapsulated cells (C-forms) were often present on cell lawns grown on nitrogen-free succinate agar. Encapsulation was initially heaviest for cells near the lawn surface. After most cells in the surface layers were converted to nonmotile C-forms, the lower cell layers began to accumulate capsules. Such C-forms were not observed within 60 hours of growth in semisolid, nitrogen-free succinate agar. They formed rapidly on nitrogen-free agar surfaces. Most of the culture formed capsules. The appearance of the encapsulated forms varied and changed with time. Both capsule formation and PHB accumulation were inhibited by combined nitrogen. As cultures aged, enlarged vibrioid C-forms developed. The more mature forms were spheres of

PAGE 60

53 2.0 to 4.0 \im diameter which had lost their motility. One-week-old cultures consisted mainly of spherical C-forms of 5.0 to 8.0 \m diameter, containing many PHB-rich cells within a common capsule. The authors speculated that younger encapsulated forms may be fixing dinitrogen and that older encapsulated forms may not. They suggested that the capsule may reduce oxygen flow into the cells, thereby protecting oxygen-sensitive nitrogenase activity. Azospirilla form extensive capsules only in media having a high C/N ratio (Sadasivan and Neyra, 1985). Such conditions promote nitrogenase activity. Since most capsules contain over 99% of their weight as water (Costerton et al. 1981), and oxygen diffuses through water at one ten-thousandth the rate of diffusion through air (Clark, 1967), the capsule may well help protect nitrogenase from oxygen damage. Oxidation of PHB reserves within the cell may also reduce oxygen levels near the nitrogenase (Dawes and Senior, 1973). The description by Berg et al. (1980) of encapsulation starting at the uppermost layers of nitrogen-free agar-grown colonies and proceeding downwards is reasonable, if one assumes that encapsulated cells are metabolically active for a time, and then pass into dormancy. Initially, the uppermost encapsulated cells would be actively fixing dinitrogen. They might become dormant as a result of underlying cell layers depleting the available carbon supply, or possibly because conditions become favorable for their passage

PAGE 61

into constitutive dormancy (Marshall, 1980). Cells might start to encyst when they accumulate threshold levels of capsular material and/or PHB. In any case, as dormant cells they would consume little oxygen, allowing it to diffuse to lower cell layers that previously may have been oxygenlimited, due to the actively respiring upper cell layer. These lower cell layers would become more active with the increased oxygen supply, accumulating capsules. Eventually these cell layers would also pass into dormancy. In earlier work. Berg et al. (1979) grew A. brasilense Sp 7 in association with sugarcane callus tissue. Vegetative cells (V-forms) grew as slimy colonies on the surface of the callus, and few of these V-forms contained PHB or capsules. Encapsulated or C-forms were also observed in these conditions. This association of azospirilla with sugarcane callus exhibited nitrogenase activity, but whether the V-forms, C-forms, or both were responsible could not be ascertained, since both were present. Perhaps C-forms were able to fix dinitrogen transiently, but were poised to enter dormancy if growth became too unbalanced. The bacteria did not possess capsules near or within lysed plant cells, where the C/N ratio may have been narrow, and balanced growth may have been promoted. An apparent contradiction in this plant callus-bacterium work is the claim by Berg et al. (1979) that C-forms of azospirilla have little similarity to Azotobacter

PAGE 62

55 spp. cysts. Krieg and Dobereiner (1984) restated this, but the photographs of Berg et al. (1979) do not support it. The multicellular C-forms are virtually indistinguishable from Azotobacter spp. cysts having multiple central bodies (Cagle and Vela, 1974). Clearly, in the association with sugarcane callus, the azospirilla were situated in numerous sites, differing in nutrient availability and oxygen availability. It is not surprising that multiple morphologies were observed, reflecting multiple physiological states. Only a few cells resembling mature cysts were present. Pleomorphic forms of azospirilla have been observed in a variety of axenic associations with plant roots. The work of Umali-Garcia et al. (1980) has already been discussed. Ruscoe et al. (1978) grew maize plants in sand and inoculated them with different strains of azospirilla. Enlarged, cyst-like cells, as well as cells of normal morphology, were observed in older and thicker root segments, where root tissue was often disintegrating. They also found that when two strains of azospirilla were grown in nitrogen-free, semisolid trans-aconitate agar, they often formed long chains after 4 to 5 days. Matthews et al. (1983) used immunological techniques and transmission electron microscopy to observe strains of A. brasilense in axenic association with pearl millet roots. Both vibrioid and encapsulated cells were observed in association with the roots. The encapsulated cells

PAGE 63

56 usually contained PHB and polyphosphate granules, and often two or more cells were enclosed by a common capsule. Patriquin et al. (1983) observed unusual structures on the surface of wheat roots, 3 weeks of age and older, that had been axenically incubated with azospirilla in a sandvermiculite mix. They appeared as spherical "bags," within which azospirilla containing PHB granules could be seen to swim about. These structures were also found between the epidermis and outer cortex of young wheat roots. Krieg and Dobereiner (1984) suggest that the capsule of azospirilla helps protect nitrogenase. They also support the idea that development of alkaline pH is the cause for pleomorphism in A. lipof erum and A. brasilense This seems an incomplete explanation, implying that pleomorphic cells are poorly viable, being aberrant forms or laboratory artifacts. Pleomorphic cells of azospirilla may instead develop commonly, and perhaps transiently, when growing in natural environments of high C/N ratio, such as near plant roots. Unbalanced growth, with increased PHB and capsule formation, may be the major cause of pleomorphism. Experimental Approach The conversion of 90% or more of an Azotobacter sp. cell suspension to cysts facilitates physiological studies of cysts. Growing the cells in nitrogen-free media containing n-butanol or BHB leads to this conversion.

PAGE 64

57 Azospirilla of cyst-like morphology have been observed under various cultural conditions, but reports of conversion of 90% or more of cell populations to cyst-like forms are not found in the literature. Vegetative cells are reported as being present in high numbers, along with the cyst-like forms. This has perhaps discouraged studies on the nature of cyst-like forms of azospirilla. All strains of A. brasilense and A. lipof erum are able to grow on BHB as sole carbon source in the presence of combined nitrogen (Tarrand et al. 1978). However, no studies have been done to see how azospirilla respond to BHB in the absence of combined nitrogen. Since such cultural conditions lead to prolific encystment of some Azotobacter strains, it was considered worthwhile to determine if strains of azospirilla might also undergo conversion in high numbers to cyst-like forms under these growth conditions. The research reported here addresses the following questions : 1. Can high numbers of cyst-like forms of azospirilla be obtained by growth in nitrogen-free BHB broth or on nitrogen-free BHB agar? 2. What are the morphological differences between azospirilla grown on BHB with or without combined nitrogen?

PAGE 65

58 3. Is the PHB content of azospirilla grown on BHB without combined nitrogen higher than when they are grown in complex broth with combined nitrogen? 4. If pleomorphism of azospirilla occurs in nitrogenfree BHB broth, is alkalinization of the medium a prerequisite for development of pleomorphism? 5. Are azospirilla grown on nitrogen-free BHB agar more desiccation resistant than cells grown in complex broth with combined nitrogen? 6. Are azospirilla grown on nitrogen-free BHB agar more resistant to starvation in carbonand nitrogen-free, phosphate-buffered salts solution than cells grown in complex broth with combined nitrogen? 7. What growth conditions favor motile azospirilla arising from nonmotile azospirilla grown on nitrogen-free BHB agar? 8. Is protein synthesis required before nonmotile BHB-grown azospirilla give rise to motile azospirilla? 9. Are BHB-grown azospirilla affected by Tris-EDTA in a manner similar to Azotobacter cysts? Questions 1 through 4 are considered in Chapter II, and e remaining questions are considered in Chapter III.

PAGE 66

CHAPTER II PLEOMORPHISM OF AZOSPIRILLA GROWN ON BETA-HYDROXYBUTYRATE Only a few bacterial genera contain strains known to form cysts (Sudo and Dworkin, 1973; Whittenbury et al., 1970a; Tudor and Conti, 1977). A nonmotile cyst forms when the entirety of a vegetative cell rounds up, depositing extracellular coats and often accumulating intracellular energy-reserve polymers. The morphological changes of encystment are accompanied by a reduction in cell metabolic activities, and increased resistance to environmental stresses, such as starvation and desiccation. Cysts of Azotobacter spp. are perhaps the best understood. Like other prokaryotic resting cells, they form when vegetative cells undergo a metabolic shift-down (Sadoff, 1975). Cysts of Azotobacter spp. do not form in media supporting good vegetative growth until stationary phase, and are present then only in low numbers (Sadoff et al., 1971). Similarly, cells of azospirilla are uniform in shape during active growth in nutritionally complete media ( Umali-Garcia et al., 1980; Lamm and Neyra, 1981; Sadasivan and Neyra, 1985). As is true for Azotobacter spp., however, stationary phase cultures of azospirilla grown on complete media often 59

PAGE 67

60 contain some rounded, nonmotile cells (Lamm and Neyra, 1981; Papen and Werner, 1982; Krieg and Dobereiner, 1984). Azospirilla are morphologically vexing, in that different pleomorphic cell types occur under various growthlimiting conditions. Weakly motile filaments containing little PHB form in aerobic broth which is low in combined nitrogen (Becking, 1982). Under dinitrogen-f ixing conditions, filamentous or S-shaped cells again may arise but contain large deposits of PHB ( Tarrand et al., 1978; Wong et al., 1980; Lamm and Neyra, 1981). These elongated cells often fragment into smaller, oval cells which subsequently can assume a cyst-like morphology (Tarrand et al., 1978). The most frequently reported pleomorphic form of azospirilla is a nonmotile cell possessing thick outer layers, probably of capsular material. These cells usually contain more extensive deposits of PHB than do vegetative cells grown with combined nitrogen. These cells have been observed in older cultures grown on combined nitrogen (Lamm and Neyra, 1981), in cultures grown as dinitrogen-f ixers (Eskew et al., 1977; Berg et al., 1979; Berg et al., 1980; Papen and Werner, 1982), and in axenic associations with grass roots (Ruscoe et al., 1978; Umali-Garcia et al., 1980; Matthews et al., 1983). Recently, Sadasivan and Neyra (1985) obtained them in broth containing fructose and KNO^ The nomenclature for describing these cells is not standardized. Berg et al. (1979) termed them capsulated or

PAGE 68

61 C-forms as opposed to the vegetative or V-forms, as did some later workers (Matthews et al. 1983; Krieg and Dobereiner, 1984). This terminology may be confusing, however, since capsules can also occur on azospirilla of otherwise normal morphology (Nur et al., 1980). The presence of a capsule is usually deemed a prerequisite for cyst formation in Azotobacter spp. (Eklund et al., 1966). Azospirilla also may need to form a capsule before they can form cyst-like cells. Encapsulated azospirilla may initially be fully active vegetative cells. Upon encountering metabolic or environmental stress, such cells may mature into cyst-like cells. The change in morphology with time of some members within a C-form population (Berg et al., 1980) may reflect maturation into truly mature cysts. Two definitive traits of a mature Azospirillum spp. cyst would be greatly reduced cell metabolism and enhanced desiccation resistance. Morphologically differentiated cells of azospirilla have been called cysts when they exhibit no nitrogenase activity (Eskew et al., 1977; Papen and Werner, 1982) or exhibit enhanced desiccation resistance (Lamm and Neyra, 1981; Papen and Werner, 1982; Sadasivan and Neyra, 1985). Another complicating factor in understanding these forms of azospirilla is that their appearance in dinitrogenfixing cultures often coincides with alkalinization of the growth medium (Eskew et al., 1977; Krieg and Dobereiner,

PAGE 69

62 1984). Krieg and Dobereiner (1984) suggest that these cell forms arise mainly at excessively high pH. In this case they might be only laboratory artifacts, or involution forms, that have no in situ function. The findings of Lamm and Neyra (1981), Papen and Werner (1982), and Sadasivan and Neyra (1985) argue against this viewpoint. Indeed, the ability of azospirilla to enter dormancy as cysts may help explain some of the great variability of plant responses to inoculation with these bacteria (reviewed by Patriquin et al., 1983). Two things are presently lacking in research and understanding of cyst-like forms of azospirilla. Although cystlike forms of azospirilla have been predominant in some studies, growing cells of normal morphology (vegetative cells) have always been present in high numbers as well. Conversion of 90% or greater of a population of vegetative azospirilla to cyst-like forms (quantitative encystment) in a reproducible manner would greatly facilitate further study of these cell forms, as it did for Azotobacter spp. cysts (Socolofsky and Wyss, 1962). Also lacking is an understanding of the underlying causes of pleomorphism and cyst formation in azospirilla. Conversion of 90% or greater of a cell population of Azotobacter spp. to cysts often can be achieved by culturing vegetative cells in the absence of combined nitrogen on either of two precursors of PHB, n-butanol or BHB (Sadoff,

PAGE 70

63 1975). Although all strains of A. brasilense and A. lipof erum are known to grow on BHB as sole carbon source when provided with combined nitrogen (Tarrand et al., 1978), there are no reports of the response of azospirilla to BHB in the absence of combined nitrogen. It was thought worthwhile to see if vegetative azospirilla would respond similarly to Azotobacter spp. by undergoing quantitative encystment in the presence of these carbon sources. In preliminary studies, apparent extensive PHB accumulation and capsule formation were observed in some strains of azospirilla grown with n-butanol. Since n-butanol is volatile, BHB was used for later studies. Initial objectives of this study were to achieve morphological encystment of high numbers of azospirilla, to document the morphology of such cells, to verify that they contained PHB, and to ascertain if alkalinization of the medium was a prerequisite for their formation. Materials and Methods Bacterial Strains The Azospirillum brasilense strains used in these studies were A. brasilense strain JM 125A2 and A. brasilense strain Cd (ATCC 29729) (both courtesy of J. Milam, Univ. of Florida, Gainesville). The Azospirillum lipof erum strains used were A. lipof erum Sp RG6xx (ATCC 29731), A. lipoferum Sp RG20a (ATCC 29708), A. lipoferum Sp RG8C, and

PAGE 71

A. lipof erum Sp A3a (all courtesy of N. R. Krieg, Va. Poly. Inst,, Blacksburg). All strains were maintained on slants of Tryptic Soy Agar (Difco Laboratories, Detroit, MI) at 25C with monthly transfer. Media Vegetative azospirilla were cultured in a modification of the complete medium of Tyler et al. (1979), denoted as trypticase-succinate salts (TSS). All components were of reagent grade and were dissolved in deionized water. The final concentrations of TSS components were (in grams per liter): (NH^)2S0^, 0.5; succinic acid, 0.437; Trypticase Peptone (Baltimore Biological Laboratory, Cockeysville MD ) 1.0; d-biotin (Sigma Chemical Co., St. Louis, MO), 0.0001; NaCl, 0.1; FeCl^-GH^O, 0.0017; Na^MoO^ 2H2O 0.0002; MgSO^-7H20, 0.2; and CaCl2, 0.002. The first four components were omitted to obtain a basal salts solution. The biotin was dissolved as a lOOX concentrated stock solution by heating and then filter-sterilized by passage through a 0.2 urn pore diameter Nalgene filter unit (Nalge Company, Rochester, NY). Two phosphate buffer concentrations were employed. The low phosphate (LP) buffer of Tyler et al. (1979) had a final concentration of 3.5 mM and consisted of (in grams per liter) K2HP0^, 0.1 and KH^PO^, 0.4. The high phosphate (HP) buffer of Albrecht and Okon (1980) had a final concentration of 63.8 mM and consisted of (in grams

PAGE 72

65 per liter) K^HPO^, 6.0 and KH2P0^, 4.0. The LP buffer was prepared as a 10 OX concentrated stock solution, and the HP buffer as a IGX concentrated stock solution. The pH of the LP buffer was adjusted to 7.1, and that of the HP buffer to 6.7, with 10 M KOH. The buffer stock solutions were sterilized by autoclaving. All autoclavings in these studies were for 15 min at standard temperature and pressure unless otherwise stated. The TSS components, excluding the biotin and phosphates, were dissolved and adjusted to pH 7.0 with 10 M KOH. The broth was then dispensed into 250 ml Erlenmeyer flasks, in an amount calculated to obtain a final volume of 100 ml after aseptic addition of the biotin and phosphate buffer stocks to the autoclaved TSS. The initial pH of the LP-TSS was 6.9 to 7.0, and that of HP-TSS was 6.8. Plate counts of azospirilla were performed with a modified succinate-nitrogen free ( SNF ) agar medium derived from Tyler et al. (1979) It was the same as LP-TSS, except that (NH_^)2S0^ and Trypticase Peptone were omitted. It contained in addition (in grams per liter) Bacto yeast extract (Difco), 0.05 and Bacto agar (Difco), 20.0. It was prepared in the same way as TSS broth, except that agar was added after neutralization and before autoclaving. Before Petri plates were poured, biotin and LP buffer were added aseptically, as was a solution of autoclaved Congo Red

PAGE 73

66 (Sigma), that was incorporated at a final concentration of 0.0375 grams per liter (Rodriguez Caceres, 1982). Nitrogen-free agar plates containing n-butanol had the same composition as SNF plates, except that the agar concentration was 1.5% (wt/vol), and yeast extract and Congo Red were omitted. The n-butanol was sterilized by filtration in the same manner as the biotin and incorporated at a final concentration of 0.2% (vol/vol). Beta-hydroxybutyrate was prepared from crotonic acid (Sigma) by dissolving 23.6 g crotonic acid in 900 ml deionized water. This solution was continuously mixed with a magnetic stirrer for two to three days at 25C. Its OD^^^ stabilized by this time, indicating conversion to BHB (H. L. Sadoff, personal communication). It was then adjusted to pH 7.0 with 10 M KOH, the final volume made up to one liter, and sterilized by autoclaving. This served as a lOX concentrated stock solution of BHB for addition to agar or broth, to give a final concentration of 0.236% (wt/vol) BHB. Agar plates containing BHB had the same composition as n-butanol plates, except that (NH^)2S0^ or Congo Red were sometimes added at the previously described concentrations. For two-step broth replacement studies (described below), broth contained BHB, biotin, and phosphate-buffered basal salts solution. The LP and HP buffers were employed in different broth replacement studies. The initial pH, after inoculation, of LP-BHB broth was 7.2, and that of HP-BHB broth was 6.9.

PAGE 74

67 Growth Conditions Inocula of azospirilla were grown in screw-cap tubes containing 10 ml of autoclaved Bacto Nutrient Broth (Difco) for 24-48 hours at 28C. One milliliter of inoculum was aseptically pipetted into 100 ml of HP-TSS broth and incubated for 20 to 22 hours at 30 C at 130 rpm on a rotaryshaker. By this time, the cultures attained OD^g^ readings of 0.7 to 0.9, as measured with a Bausch and Lomb Spectronic 20 spectrophotometer. The pH of the cultures at harvest ranged from 7.0 to 7.2. Cultures were pelleted by centrifugation at 6,960 X g for 15 min at 20C. Cells were washed twice by resuspension and pelleting in sterile LP-basal salts solution (pH 7.3). The cells were resuspended in sterile LP-basal salts solution to give a final OD^gg reading of 1.0 to 1.2. Cell lawns were obtained by spread plating 0.1 ml of washed cells onto agar media. Inoculated plates were sealed with Parafilm and incubated at 28 C. For two-step broth replacement studies, washed cells were aseptically added as a 10% (vol/vol) inoculum to 250-ml Erlenmeyer flasks, containing a final volume of 100 ml BHB broth after cell addition. Duplicate flasks were incubated in the same manner as TSS flasks. These studies were modeled after the two-step replacement method of Lin and Sadoff (1968 )

PAGE 75

68 Harvest of Cell Lawns To harvest lawns of azospirilla grown on n-butanol or BHB agar, about 7.0 ml of sterile deionized water was aseptically poured across the surface of a cell lawn, and the cells gently scraped from the agar surface with a flamed wire loop. For PHB analyses and plate counts, the suspended cells of one BHB agar plate were aseptically transferred to another plate whose cells were in turn scraped off. This was done to ensure that the cell suspension would not become too diluted. Enumeration Vegetative cells from TSS-broth, or cells grown on BHB agar, were diluted ten-fold in a series of dilution blanks containing LP-basal salts solution. For enumeration, 0 1 ml of cell suspension was aseptically spread on SNF-Congo Red agar plates. Four plates were spread for each dilution. Plates were incubated as described above for 5 days before counting. Dry Weight Determination and PHB Analysis To assay PHB content of vegetative cells of A. lipof erum Sp RG6xx, two 22-hour-old, HP-TSS cultures (560 ^ ^'^^ vjere pooled for centrif ugation and washing as described above, except that sterile deionized water was used for washing. The final cell suspension was adjusted to

PAGE 76

69 an OD^gQ = 0.86. Forty plates of the same strain grown on nitrogen-f ree-LP-BHB agar were harvested by scraping (described above), to give a final cell volume of about 100 ml. These cells were centrifuged and washed in sterile deionized water (described above) and resuspended to give an OD^^Q of 0.25 to 0.28. For dry weight determinations, 10.0 ml of the final cell suspension were pipetted into previously weighed and desiccated aluminum pans. Five replicate pans were prepared for each cell type. The pans containing cells were dried to constant weight at 100C. Pans were kept in a glass desiccator over anhydrous CaSO^ (Drierite) after removal from the oven and before weighing. For PHB determination, 10.0 ml of washed cells were added to 15 ml Corex centrifuge tubes (Corning Glass Works, Corning, NY) and pelleted by centrif ugation at 7,080 X g for 20 min at 4C. Three replicate tubes were prepared for each cell type. The supernatant was poured off, and subsequent steps were performed by the method of Law and Slepecky (1961). Digestion of cell pellets was begun with the addition of 10 ml of Clorox bleach (5% (wt/vol) hypochlorite). Cells were suspended in the bleach with Pasteur pipettes; then the tubes were capped with glass marbles and incubated in a 37 C water bath. Digestion to constant 0D,^„ was 5 5 0 monitored with a Spectronic 20 spectrophotometer and was judged to be complete after 18 hours. The insoluble cell

PAGE 77

70 material was pelleted by centrif ugation as above, then washed once in 10 ml of sterile deionized water, and pelleted again. The volume for all subsequent washings and digestions was maintained at 10.0 ml, and all chemicals were of reagent grade. The OD^^^ of the samples in the final digestion of concentrated H2S0^ was measured in quartz cuvettes (1.0 cm light path), using a Carl Zeiss M4QIII spectrophotometer. For the standard curve, the sodium salt of DL-6-hydroxybutyric acid (Sigma) was dissolved directly in concentrated H^SO^. The standard curve was linear up to 8.0 yg BHB/ml. The PHB content of cell digests was related back to dry weight values, to determine what percentage of cell dry weight was present as PHB. Scanning Electron Microscopy (SEM) Samples of 0.4 ml from either LP-BHB agar plates or two-step, broth-replacement cultures were employed for SEM studies. Cells were removed aseptically from the two-step, broth-replacement cultures at the same time that culture pH was measured. Cells were aseptically impinged upon autoclaved 2 5-mm-diameter 0 4 S-ym-pore-size Nuclepore polycarbonate filters (Nuclepore Corporation, Pleasanton, CA), housed in a filter chimney attached to a vacuum source. About 10.0 ml of sterile, deionized water was added to the chimney after cell addition, to help distribute the cells evenly over the membrane surface, then a vacuum not

PAGE 78

71 exceeding 33.8 kPa was applied. Filter membranes were then removed and placed into Karnovsky's fixative (1965) for 1 hour. Filter membranes were subsequently rinsed twice for 10 min in cacodylate buffer and then dehydrated in a graded series of ethanol concentrations (10, 20, 30, 50, 70, 90, 95, 100, and 100%) for 10 min at each concentration. The samples were then air dried. Sections of filters were excised, placed onto aluminum stubs with double-stick tape, and gold coated with an Eiko IB-2 coater. Specimens were examined with a Hitachi S450 scanning electron microscope at 20 kilovolts. Photographs were taken with Polaroid Type 55, positive/negative, 4X5 Land film. Light Microscopy Cells were routinely observed by phase-contrast microscopy using a Wild M20 or a Nikon Labophot microscope. Cell dimensions were measured with an ocular micrometer. Photographs of cells viewed with the latter microscope were taken with a Microflex AFX camera attachment, using Ilford FP4 black and white film. All photos were taken using phasecontrast optics, unless otherwise indicated. Results Quantitative Morphological Change In the initial phase of these studies only three strains of azospirilla were used, A. brasilense strain

PAGE 79

JM 125A2, A. brasilense strain Cd, and A. lipof erum Sp RG6xx. Slime developed at the bottom of stationary phase LP-TSS broth cultures of all three strains. Phase-contrastmicroscopy examination of A. lipof erum Sp RG6xx slime often revealed numerous, nonmotile masses of cells similar to zooglea, surrounded by nonmotile vegetative cells (Figure 2-la, b). These masses were notable for their symmetrical but varied shapes. They were darker than most of the surrounding vegetative cells, perhaps indicating greater viability than that of the surrounding pale vegetative cells. These zoogleal masses retained their shape and did not fragment into individual cells when disassociated from the larger masses of cells. Similar zoogleal forms were sometimes observed in the slime of A. brasilinese strain Cd but not in that of A. brasilense strain JM 125A2. These zoogleal forms of azospirilla may be referred to as microscopic floes, or microflocs, that are kept intact by exopolysaccharide. Although microflocs were numerous, individual normal cells were also present in large numbers under these cultural conditions. Azospirilla were cultured as cell lawns on agar containing precursors of PHB to see if high numbers of pleomorphic forms would arise. The A. brasilense strains produced some pleomorphic forms, but cells of normal shape and size predominated, even on old lawns.

PAGE 80

Figure 2-1. Zoogleal masses in stationary phase 40-day-old, low phosphate-trypticasesuccinate-salts broth culture of Azospirillum lipoferum Sp RG6xx. a) Cells at 600X magnification. Bar equals 6.0 urn. b) Detail from same mass of cells viewed at 1,500X magnification. Bar equals 3.0 \xm.

PAGE 81

b

PAGE 82

75 After 63 hours of growth, A. brasilense strain JM 125A2 lawns grown on BHB contained ovoids, vibrioids, and chains of cells. Many cells contained phase-bright, putative PHB granules. More cells were present at this time on agar containing combined nitrogen. The several cell types present on 63-hour-old, LP-BHB agar with combined nitrogen are shown in Figure 2-2a. Some cells appeared at this time to be undergoing plasmolysis on this medium, as well as on HP-BHB agar with combined nitrogen. By 96 hours, the lawns on LP-BHB agar with and without combined nitrogen contained more chains of cells and microflocs than the HP-BHB lawns, which consisted mostly of individual ovoids or pairs of ovoids Cells from month-old, nitrogen-free, HP-BHB lawns of A. brasilense strain JM 125A2 are shown in Figure 2-2b. Individual, nonmotile vibrioids and ovoids were still predominant, as were pairs of cells. Enlarged, nonmotile, spherical cells were present, but not numerous. A few nonmotile filaments appeared to be undergoing septation. After 79 days, cells from lawns of this strain grown with combined nitrogen had the appearance of stationary phase cells from TSS broth cultures grown with combined nitrogen, and spheroplasts and cell ghosts predominated. Nitrogen-free cultures at both phosphate buffer concentrations contained numerous pleomorphic forms. Figure 2-3a shows nonmotile, enlarged, rounded individual cells from

PAGE 83

Figure 2-2. Cell types of Azospirillum brasilense strain JM 125A2, grown on g-hydroxybutyrate (BHB) agar. a) 63-hour-old cells from low phosphate-BHB agar with combined nitrogen. 1,500X magnification. Bar equals 3.0 urn. b) Month-old cells from nitrogen-free, high phosphate-BHB agar. 1,OOOX magnification. Bar equals 4.0 ym.

PAGE 84

77 b

PAGE 85

Figure 2-3. Call types of Azospirillum brasilense strain JM 125A2, from 79-day-old nitrogenfree, high phosphateg-hydroxybutyrate agar cultures. a) Individual rounded cells. 1,500X magnification. Bar equals 3.0 um. b) Microfloc focused so that capsules are visible around cells on right side of floe. 1,500X magnification. Bar equals 3.0 jam. c) Same cells as (b), but focused so that capsules are no longer evident. Note empty capsule at bottom of floe. 1,500X magnification. Bar equals 3.0 ym

PAGE 86

79 a c

PAGE 87

80 nitrogen-free HP-BHB agar. Their phase-bright inclusions are probably PHB granules; some contain dark bodies, probably polyphosphate granules. Also shown is a microfloc of nonmotile, enlarged, PHB-rich cells (Figure 2-3b, c). By adjusting the distance of the objective lens from the specimen, many of these cells were observed to be encapsulated (Figure 2-3b), The encapsulated cells fitted together closely, as did those observed by Sadasivan and Neyra (1985). The thickness of the capsule was about 0.5 urn. Such encapsulated microflocs were also observed on nitrogen-free, LP-BHB agar at this time. Azospirillum brasilense strain JM 125A2 may have lacked an efficient mechanism for BHB uptake, compared to the other strains of azospirilla used. Unlike the other strains, few motile cells were observed on any BHB agar medium, even in young cultures. It also differed from the other strains by having many phase-dark cells that contained little or no PHB. It eventually grew well on BHB agar when combined nitrogen was available, however. A final difference between this strain and the others was that its cells always resuspended in water to give uniform turbidity, with no macroscopic floes, or macroflocs, being present. This indicates that, with or without combined nitrogen, cells of this strain produce little capsular material when cultured on BHB.

PAGE 88

81 The best growth of A. brasilense strain JM 125A2 on agar occurred on SNF-Congo Red agar. Cells from 6-day-old lawns grown on this agar medium were often seen as encapsulated microflocs (Figure 2-4). The microfloc in Figure 2-4a and b appears to have arisen mainly from one or more filamentous cells that underwent septation. This may also have occurred for many of the cells in Figure 2-4c. The capsules were of thickness comparable to those observed on BHB agar, about 0.5 \xm. The lawns on SNF-Congo Red agar had a scarlet or blood-red appearance, unlike lawns of this strain growing on nitrogen-free BHB-Congo Red agar, which were pale orange. The other A. brasilense strain, A. brasilense strain Cd, also failed to convert in high numbers to pleomorphic forms, but it grew far better on BHB. After 63 hours of growth, lawns of this strain on each BHB agar medium contained many motile vibrioids possessing large granules of putative PHB. Elongated, filamentous cells were also present in high numbers. These cells had about the same width (1.5 um) as normal dinitrogen-f ixing cells but were much longer, some being 9 to 13 nm in length (Figure 2-5a). The filaments were sometimes observed to undulate slowly and were much slower than motile vibrioids. In the presence of combined nitrogen, these filaments were seen to septate and fragment. This fragmentation was observed at 63 to 96 hours, and sometimes was complete within a population of

PAGE 89

Figure 2-4. Cells of Azospirillum brasilense strain JM 125A2, from 6-day-old lawns grown" on succinate-nitrogen-f ree-Congo Red agar, a) Microfloc showing capsules and filamentous cell patterns. 1,500X magnification. Bar equals 3.0 jam. b) Same floe as (a), but focused so that capsules and filamentous cell outlines are no longer evident. 1,500X magnification. Bar equals 3.0 nm. c) Different mass of encapsulated cells. 1,500X magnification. Bar equals 3.0 nm.

PAGE 90

83

PAGE 91

Figure 2-5. Cell types of Azospirillum brasilense strain Cd, from lawns on 6-hydroxybutyrate (BHB) agar. a) Filaments from 63-hourold, high phosphate-BHB agar with combined nitrogen. 1,500X magnification. Bar equals 3.0 \xm. b) Microfloc from 11-day-old nitrogen-free, low phosphateBHB agar, focused to show capsules and filamentous cell outline. 1,000X magnification. Bar equals 4.0 nm. c) Same floe as (b), but focused so that capsules and filamentous cell outline are no longer evident. 1,000X magnification. Bar equals 4.0 um.

PAGE 92

J

PAGE 93

86 filaments soon after 96 hours. In nitrogen-free cultures, such elongated filaments persisted, some being weakly motile even after 79 days on nitrogen-free LP-BHB agar. After 96 hours, lawns of A. brasilense strain Cd on all BHB agar media contained mixtures of vibrioids, ovoids, filaments, and chains. Sometimes the cell material from LP-BHB agar lawns with or without combined nitrogen did not resuspend uniformly in water, but as macroflocs, due to extensive encapsulation. A microfloc from an 11-day-old, nitrogen-free, LP-BHB agar plate is shown in Figure 2-5b,c. By adjusting the objective lens, the capsule is made evident. The entire microfloc may have arisen from one elongated filament that underwent septation, as suggested by the apparent linear continuities between cytoplasmic contents After 79 days of culture, lawns of this strain grown with combined nitrogen contained mainly spheroplasts and cell ghosts, appearing to have entered stationary phase. Cultures grown on nitrogen-free agar at both phosphate levels contained numerous pleomorphic forms at this time, in addition to cells of normal morphology (Figure 2-6). A microfloc of A. brasilense strain Cd from a 6-day-old lawn on SNF-Congo Red agar is shown in Figure 2-7. All the cells are encapsulated, and empty capsules are evident. The lawn was scarlet in color, unlike the pale orange lawns of the same age grown on nitrogen-free BHB-Congo Red agar.

PAGE 94

I I i Figure 2-6. Cell types of Azospirillum brasilense strain Cd, from 79-day-old lawns on nitrogen-free, B-hydroxybutyrate (BHB) agar. a) Multicellular packets, a chain, and individual ovoids from low phosphateBHB agar. 1,500X magnification. Bar equals 3.0 jam. b) Microfloc from high phosphate-BHB agar, focused to show capsules around several cells. Air bubble is above floe. 1,500X magnification. Bar equals 3.0 \im.

PAGE 95

b

PAGE 96

89 Figure 2-7. Microfloc of Azospirillum brasilense strain Cd, from 6-day-old lawn on succinate-nitrogen-f ree-Congo Red agar. Note capsules around cells and empty capsules. 1,500X magnification. Bar equals 3.0 um.

PAGE 97

The two A. brasilense strains failed to produce morpho logically uniform populations on BHB-agar. A uniform response was observed for A. lipof erum Sp RG6xx. Good growth usually occurred within 18 to 24 hours. Figure 2-8 shows 18-hour-old cells grown on nitrogen-free, HP-BHB agar. Filaments and chains were present, which were sometimes as swiftly motile as vibrioids. On LP-BHB agar at both phosphate levels, with or without combined nitrogen, septation of filaments was almost complete between 4 8 and 7 hours, although new filaments would arise and septate for u to the fifth day. Figure 2-9a, b shows such completely septated microflocs on 63-hour-old, nitrogen-free, LP-BHB agar. The floes are encapsulated, and most seemed to arise from one filament that underwent complete septation. Although microflocs were present at this time, the cells resuspended from agar as uniformly turbid suspensions without macroflocs. Many filaments were also completely septated by 63 hours on HP-BHB agar containing combined nitrogen, but some filaments still lacked completed septa (Figure 2-9 c). Nitrogen-free, HP-BHB agar at this time contained few if an^ microflocs. As for all other media, very motile vibrioids and ovoids, as well as filaments of varying motility, were present at 63 hours. Encapsulated microflocs sometimes formed on nitrogenfree, HP-BHB agar (Figure 2-lOa), but cells from young or

PAGE 98

Figure 2-8. Cells of Azospirillum lipof erum Sp RG6xx, from 18-hour-old lawn on nitrogen-free, high phosphateg-hydroxybutyrate agar. Note individual cells and filaments at various stages of septum formation. 1,000X magnification. Bar equals 4.0 \m.

PAGE 99

Figure 2-9. Cell types of Azospirillum lipof erum Sp RG6xx, from 63-hour-old lawns on 6-hydroxybutyrate (BHB) agar. a) Microflocs from nitrogen-free, low phosphateBHB agar, focused to show capsules and filamentous cell outlines. 1,500X magnification. Bar equals 3.0 ]im. b) Same microflocs as (a) but focused so that capsules and filamentous cell outlines are no longer evident. 1, 500X magnification. Bar equals 3.0 \xm. c) Filament from high phosphate-BHB agar with combined nitrogen. 1,500X magnification. Bar equals 3.0 jam.

PAGE 100

93

PAGE 101

Figure 2-10. Cell types of Azospirillum lipof erum Sp RG6xx from lawns on 6-hydroxybutyrate (BHB) agar. a) Microfloc from 13-dayold cell lawn on nitrogen-free HP-BHB agar. Note empty capsules. 1,500X magnification. Bar equals 3.0 um. b) Individual cells and septating filaments from 13-day-old, high phosphate-BHB agar with combined nitrogen. 1,500X magnification. Bar equals 3.0 um. c) Microfloc from 79-day-old, low phosphate-BHB agar with combined nitrogen. 1,500X magnification. Bar equals 3.0 um.

PAGE 103

old lawns grown on it were always resuspended uniformly, without macroflocs. Older cultures on this medium consisted mainly of PHB-rich, nonmotile ovoids or peanut-shaped cells, with few if any microflocs. The high phosphate level did not appear to inhibit PHB accumulation, but did inhibit extensive capsule formation. As was true for the A. brasilense strains, combined nitrogen led to eventual good growth and passage into stationary phase. Figure 2-lOb shows cells of a 13-day-old culture from HP-BHB agar containing combined nitrogen. Chains of cells and individual ovoids are present. The LP-BHB lawns grown with or without combined nitrogen had the same appearance by 7 days. The lawns consisted almost entirely of floes that broke into various sizes when resuspended in water. The cells would not suspend evenly in water, due to the presence of many macroflocs. Very few motile cells were present at this time. Eventually the LP-BHB lawns grown with combined nitrogen resumed vegetative growth and passed into stationary phase, but the floes persisted even in stationary phase cultures (Figure 2-lOc). Figure 2-11 shows cells from 17-day-old, nitrogen-free, LP-BHB-Congo Red lawns. Figure 2-lla was taken with bright-field optics, showing the clearly outlined capsules and enlarged PHB-rich cells. Figure 2-llb was taken with phase-contrast optics, and the capsules enclosing all of the microfloc are again evident. It was interesting to find

PAGE 104

Figure 2-11. Cell types of Azospirillum lipof erum Sp RG6xx, from 17-day-old lawns on nitrogen-free low phosphateB-hydroxybutyrate-Congo Red agar. a) Microfloc viewed with bright-field optics. 1,500X magnification. Bar equals 3.0 urn. b) Another microfloc, viewed with phase contrast optics. Note that capsules of the two bottom-left cells are apparently undergoing division with their cells. 1,500X magnification. Bar equals 3.0

PAGE 105

98 b

PAGE 106

99 that some cells that were undergoing division in one floe appeared to have their capsules dividing as well at the site of septum formation (Figure 2-llb). Cells from 17-day-old, SNF-Congo Red lawns are shown in Figure 2-12. Abundant capsules are again evident. These cells appeared more vibrioid in shape than most A. brasilense cells cultured in the same manner. Cells of A. lipof erum Sp RG6xx grown on SNF-Congo Red agar appeared less swollen and rounded than their counterparts on LP-BHB Congo Red agar. However, unlike the A. brasilense strains, lawns of A. lipof erum Sp RG6xx were scarlet on both SNF-Congo Red and LP-BHB-Congo Red agar. Nitrogen-free, LP-BHB cultures of A. lipoferum SP RG6xx did not change in appearance from the seventh day onward, even after months had passed. Figure 2-13 shows a microfloc of this strain, with the objective lens adjusted to show the capsules (Figure 2-13a), and then readjusted to show the capsules and the apparent continuities between cytoplasms (Figure 2-13b). Cells from this medium of 7-days-age or older were consistently resuspended as macroflocs and microflocs. Individual, septated filaments apparently consolidated into floes, and individual, motile cells may have attached to septated filaments to give rise to large macroflocs More details of floe structure of A. lipoferum Sp RG6xx were obtained from SEM photographs. Figures 2-14 to

PAGE 107

Figure 2-12. Cells of Azospirillum lipof erum Sp RG6xx, from 17-day-old lawns on succinate-nitrogen-f ree-Congo Red agar, a) Microfloc with empty capsules, as well as capsules retaining their cells. 1,500X magnification. Bar equals 3.0 \im. b) Another mass of encapsulated cells. Note that some capsules seem to contain inclusion granules but no cytoplasm. 1,500X magnification. Bar equals 3.0 \xm.

PAGE 109

Figure 2-13. Microflocs of Azospirillum lipof erum Sp RG6xx, from 36-day-old nitrogen-free, low phosphate3-hydroxybutyrate agar, a) Microflocs focused to show capsules. 1,000X magnification. Bar equals 4.0 ym. b) Same floes as (a) but focused to show both capsules and filamentous cell outlines. 1, 000X magnif i ^^ation Bar equals 4.0 \im.

PAGE 110

103

PAGE 111

104 2-16 show successive magnifications of cells scraped from a 75-day-old, nitrogen-free, LP-BHB agar lawn. Cells from 19-, 57-, and 66-day-old lawns on the same medium were of identical appearance to this older culture. This supported the phase-contrast microscopy studies, in that the morphology of microflocs from nitrogen-free, LP-BHB agar did not change once they were formed. The 500X magnification photo (Figure 2-14) shows the great variability in numbers of cells per floe and that individual, encapsulated cells are often present. The 1 500X-magnif ication photo (Figure 2-15) reveals, in agreement with the phase-contrast observations, the frequent tight fit between adjacent capsules. Intercellular gaps are often observed within floes. The 7,000Xmagnif ication photo (Figure 2-16) reveals some variability in the surfaces of encapsulated cells, possibly indicating a difference in exopolysaccharide composition. It is clear at this higher magnification that the cells of agar-grown floes range from monolayers to trilayers. Empty capsules are also visible Similar results were obtained with this strain on LP-nbutanol agar. Figure 2-17 shows 178-day-old cells from this agar medium after the agar had dried into a thin film. When the agar surface was rehydrated and the cells scraped from it, many empty capsules were seen, but many capsules still contained cells. The capsular material was thus observed to retain its outline, even if it contained no cell.

PAGE 112

Figure 2-14. Encapsulated cells of Azospirillum lipof ertim Sp RG6xx from a 75-day-old lawn on nitrogen-free, low phosphateBhydroxybutyrate agar. Cells are viewed at 500X magnification by scanning electron microscopy. Bar equals 50.0 lom.

PAGE 113

106 Figure 2-15. Encapsulated cells of Azospirillum lipof erum Sp RG6xx from a 75-day-old lawn on nitrogen-free, low phosphate6hydroxybutyrate agar. Cells are viewed at 1,500X magnification by scanning electron microscopy. Bar equals 5.0 ym.

PAGE 114

107 Figure 2-16. Encapsulated cells of Azospirillum lipof erum Sp RG6xx from a 75-day-old lawn on nitrogen-free, low phosphate6hydroxybutyrate agar. Cells are viewed at 7,000X magnification by scanning electron microscopy. Bar equals 5.0 pun.

PAGE 115

Figure 2-17. Microflocs of Azospirillum lipof erum Sp RG6xx from 17 8-day-old lawns grown on nitrogen-free, low phosphate-n-butanol agar. a) Microfloc showing numerous cells within capsules as well as empty capsules. 1,000X magnification. Bar equals 4.0 \im. b) Microflocs with few cells remaining within capsules. 1,000X magnification. Bar equals 4.0 um.

PAGE 116

109 b

PAGE 117

110 Growth of this A. lipof erum strain on nitrogen-free, LP-BHB agar resulted in homogeneous encapsulation and filament formation. But, as is evident from the photographs, several cell shapes and sizes were present within any one floe. Despite their morphological heterogeneity, cells in these floes generally appeared to be more rounded and swollen than cells grown on SNF-Congo Red agar, although capsules were of equal width (0.5 jam) under both cultural conditions The homogeneous encapsulation and filament formation of A. lipof erum Sp RG6xx prompted a search for similar response in other strains of this species. Three other strains of this species, cultured on LP-BHB nitrogen-free agar, responded about as well as A. lipof erum Sp RG6xx (Figures 2-18 to 2-20). Cells of all strains did not resuspend uniformly in water, due to macroflocs. As was true for A. lipof erum Sp RG6xx, the appearance of the other three strains did not change noticeably with time, and several cell sizes and shapes were usually present within any one microfloc. Azospirillum lipof erum Sp A3a differed from all the other A. lipof erum strains in consistently having large numbers of individual, PHB-rich, nonmotile, nonencapsulated, ovoid cells in its resuspended lawns. Possibly many of these individual cells were initially present within capsules, but were released from capsules upon the addition of water. The other strains had only a few free.

PAGE 118

Figure 2-18. Microflocs of Azospirillum lipof erum Sp RG8c, grown on nitrogen-free, low phosphate6-hydroxybutyrate agar. a) Microfloc from 69-day-old lawn, focused to show capsules and filamentous cell outline. 1,500X magnification. Bar equals 3.0 \im. b) Same floe as (a) but focused so that capsules and filamentous cell outline are no longer evident. 1,500X magnification. Bar equals 3.0 Um. c) Microflocs from 136-day-old lawn. Note empty capsules. 1,000X magnification. Bar equals 4.0 \im.

PAGE 120

Figure 2-19. Microflocs of Azospirillum lipof erum Sp RG20a, from 69-day-old nitrogen-free, low phosphate6-hydroxybutyrate agar, a) Microflocs focused so capsules and filamentous cell outline are evident in upper left part of left floe. 1,500X magnification. Bar equals 3.0 urn. b) Microflocs with cells having varied morphologies. 1,500X magnification. Bar equals 3.0 ^m.

PAGE 121

114

PAGE 122

Figure 2-20. Microflocs of Azospirillum lipof erum Sjj A3a from nitrogen-free, low phosphateBhydroxybutyrate agar. a) Microfloc from 69-day-old lawn. 1,500X magnification. Bar equals 3.0 nm. b) Microfloc from 136-day-old lawn. Note empty capsules. 1,000X magnification. Bar equals 4.0 urn

PAGE 123

b

PAGE 124

nonencapsulated cells in their suspensions. Whether these free cells never merged with microflocs, or whether they were ejected from their capsules upon wetting and suspension for microscopy, is not clear. The width of the capsules on these three strains was again observed to average about 0.5 urn, and most floes arose from one or more filaments that eventually underwent septation. Enumeration of Encapsulated Cells It was first suspected that cells of A. lipof erum Sp RG6xx harvested from nitrogen-free, LP-BHB agar would not be quantifiable by plate counting. Plate counts demand that the inoculum be uniformly suspended and diluted, and flocculation makes these difficult to accomplish. But the resuspended cells gave consistent CFU counts for a given OD^^q range (Table 2-1). After resuspension in water and washes in buffer, the cell suspensions appeared silvery, and floe size was reduced to the lower limits of visibility to the naked eye. Macroflocs broke into smaller microfloc domains, which had formed within the first few days on the agar medium. These small domains retained their integrity even after repeated shaking and washing steps. The similar CFU counts for cells of different ages suggested that the cells were in a sort of stasis, where they were no longer multiplying or dying off appreciably.

PAGE 125

118 Table 2-1. Optical density (OD ) and colony forming units ml(CFU ml ^) of encapsulated cells of Azospirillum lipof erum Sp RG6xx. Culture aae (days)^ on "560 L,r U mi ( A 1 U ) 11 0.27 2.67 (0.12)^ 21 0.29 5.08 (0.79) 22 0.25 6.40 (1.00) 58 0.24 1.94 (0.19) 75 0.25 2.11 (0.32) 80 0.30 6.65 (0.49) Cells were harvested from nitrogen-free, low phosphate-6 -hydroxybutyrate agar plates. '^Values are averages of four spread plates. Values in parentheses are standard deviations.

PAGE 126

119 PolyB-Hydroxybutyrate Content The phase-bright inclusion bodies in the cells of A. lipof erum Sp RG6xx grown on nitrogen-free, LP-BHB agar were confirmed to be PHB using the method of Law and Slepecky (Table 2-2). This method is subject to error due to repeated centrif ugations and pipettings, so it is not certain whether the differences in PHB content between cell lawns of different ages were real or artifacts. The purpose was to verify that PHB existed in the encapsulated cells in greater amounts than in vegetative cells grown in HP-TSS broth. The assay gave evidence of this. Two-step Broth Replacement Studies Strains of A. lipof erum cultured in LP-TSS broth clumped and flocculated in under 24 hours. Clumping of these strains was delayed in HP-TSS broth. The pH of these cultures at 20 to 22 hours ranged from 7.0 to 7.2, having risen from an initial pH of 6.8. Cells of each A. lipof erum strain in this study, grown in HP-TSS broth, usually started to clump and flocculate by about 24 hours after inoculation. Figure 2-21 shows a SEM photo of cells from a 14-dayold, HP-TSS culture of A. lipof erum Sp RG6xx. The pH of the culture at this time was 9.1. The floe has a similar arrangement to microflocs of the same strain grown on BHB agar, with filamentous cells and frequent spaces in the

PAGE 127

120 Table 2-2, Poly6-hydroxybutyrate (PHB) content of Azospirillum lipof erum Sp RG6xx. Cultural conditions Age Dry weight (mg ml ^ )^ Volume of CHCl^ extract used (ml) ,-1 ml PHB % dry weight as PHB 1^ 11 days 0.17 (0.02) 0.5^ 3 .39 (0.52) 39.9 1 21 days 0.18 (0.05) 0.5^ 4 .56 (0.33) 50.7 1 22 days 0.12 (0.04) 0.5^ 3 .45 (0.03) 57.5 2^ 22 hours 0.49 (0.03) 3.0^ 5 .28 (0.08) 3.6 a Values are averages of five replicates. Values in parentheses are standard deviations. Two H^SO^ replicates of each CHCl^ replicate were used. Values are averages of all H SO. replicates. Values in parentheses are standard deviations. c 1 = Nitrogen-free, low phosphate6-hydroxybutyrate agar. *^Four replicate CHCl^ extracts were used. ^Three replicate CHCl^ extracts were used. ^Two replicate CHCl^ extracts were used. ^2 = High phosphate-trypticase-succinate-salts broth.

PAGE 128

121 Figure 2-21. Cells of Azospirillum lipof erum Sp RG6xx, from 14-day-old, stationary phase, high phosphate-trypticase-succinate-salts broth culture of pH 9.1. Cells are viewed at 1,500X magnification by scanning electron microscopy. Bar equals 5.0 ym.

PAGE 129

122 clumps. Sometimes floes from HP-TSS broth cultures had no such empty spaces. Two-step, broth-replacement studies were conducted to see if quantitative pleomorphism could be induced in broth and if there was any connection between pH and pleomorphism. Sometimes such cultures of A. lipof erum strains would clump extensively within 24 hours, so that the broth appeared clear to the naked eye except for the clumps, but this phenomenon was not consistently reproducible. It seemed that clumping occurred sooner and persisted longer in two-step, broth-replacement studies with A. lipof erum Sp RG6xx using LP buffer. When HP buffer was used, clear broth columns often became turbid eventually. Figure 2-22a shows cells of A. lipof erum Sp RG6xx from a 43-day-old culture in LP-BHB broth of pH 8.4. The cells appear healthier than stationary phase cells in TSS broth. Filamentation and complete septation of cells occurred in most cases. The elevated pH of the culture again indicated the low buffering capacity of the LP buffer. The results from HP-BHB, two-step, broth-replacement studies indicated that pleomorphism and encapsulation could occur at near-neutral pH for the A. lipof erum strains. All the following studies were conducted in HP-BHB broth. Figure 2-22b shows cells from a 43-day-old culture of A. lipoferum Sp RG6xx of pH 7 2 In interpreting this photo, it should be remembered that the cells were added by

PAGE 130

Figure 2-22. Cells of Azospirillum lipof erum Sp RG6xx from 43-day-old, nitrogen-free, Qhydroxybutyrate ( BHB ) broth cultures, viewed by scanning electron microscopy, a) Cells from low phosphate-BHB broth, pH 8.4. 7,000X magnification. Bar equals 5.0 ym. b) Cells from high phosphate-BHB broth, pH 7.2. Note thorough encapsulation of lower cell layer.. 3, 000X magnification. Bar equals 5.0 pm.

PAGE 131

b

PAGE 132

125 pipette to the filter membrane. It appears that large, encapsulated cell floes settled first onto the filter, followed by free, nonf locculated cells. The lower layers of cells have the same close-fitting appearance as was often observed on BHB agar surfaces. Similar results were obtained for the other three strains of A. lipof erum Figure 2-23a shows cells from a 33-day-old culture of A. lipof erum Sp A3a of pH 7-2. Some encapsulated cells in the lower cell layer are fitted together snugly, while others are joined by strands of putative exopolysaccharide Figure 2-23b is a higher magnification of cells from this culture, again showing the strands joining cells. The lumpy appearance of the cells is probably due to large, intracellular accumulations of PHB. Figure 2-24 shows cells from a 33-day-old culture of A. lipoferum Sp RG20a of pH 7.2. The formation of filaments and eventually chains was very pronounced in this strain. Cells from a 33-day-old culture of A. lipoferum Sp RG8c of pH 7.1 are shown in Figure 2-25a. This strain was sometimes observed to form intricately structured clumps. Figure 2-25b shows such a clump from a 9-day-old culture of pH 7.0. Filamentous, septate cells are present, and again empty spaces occur within the floe. The two A. brasilense strains used in this study responded poorly to two-step broth replacement. The pH of

PAGE 133

Figure 2-23. Cells of Azospirillum lipof erum Sp A3a from a 33-day-old, nitrogen-free, high phosphate6-hydroxybutyrate broth culture, pH 7.2, viewed by scanning electron microscopy. a) Lower cell layer is thoroughly encapsulated. 3,000X magnification. Bar equals 5.0 urn. b) Details of cell surfaces. Note strands of material joining some cells. 17,000X magnification. Bar equals 0.5 um.

PAGE 134

127

PAGE 135

128 Figure 2-24. Cells of Azospirillum lipof erum Sp RG20a from a 33-day-old, nitrogen-free, high phosphate6-hydroxybutyrate broth culture, pH 7.2, viewed by scanning electron microscopy. Note thorough encapsulation of lower cell layer and presence of filaments and chains. 3,000X magnification. Bar equals 5.0 Vim.

PAGE 136

Figure 2-25. Cells of Azospirillum lipof erum Sp RG8c from nitrogen-free, high phosphate-Bhydroxybutyrate broth, viewed by scanning electron microscopy. a) Cells from 33-day-old culture, pH 7.1. Note thorough encapsulation of lower cell layer. 3,000X magnification. Bar equals 5.0 um. b) Cells from 9-day-old culture, pH 7-0. Note holes within the floe. 1,700X magnification. Bar equals 5.0 ]im

PAGE 138

131 these cultures at 9 days was 6.9 to 7.1. When viewed by SEM, cells of A. brasilense strain Cd were usually rods which had collapsed or shrunken during fixation and dehydration. Cells of A. brasilense strain JM 125A2 usually appeared as very small ovoids, having the appearance of starved or stationary phase cells. A small number of cells of the latter strain were enlarged ovoids and possibly were encapsulated. Unlike all strains of A. lipof erum tested, the two A. brasilense strains showed little or no tendency to clump in HP-BHB, two-step, broth-replacement studies. One more observation made during HP-BHB, broth-replacement studies deserves mention. It was suggested earlier that some filamentous cells that had undergone septation and encapsulation might retain cytoplasmic connections between adjacent cells. Figure 2-26 shows a clump of cells from a 58-day-old, HP-BHB broth culture of A. lipof erum Sp RG6xx. Several of the cells in the clump appear to be undergoing plasmolysis, but there appear to be continuities of cytoplasm between some neighboring, plasmolyzed cells. Discussion Only two A. brasilense strains were employed in this study, and both responded far less uniformly to growth on BHB than did the A. lipof erum strains. There may have been poor uptake of BHB by A. brasilense strain JM 125A2. Cells of this strain that were able to grow on BHB

PAGE 139

132 o Figure 2-26. Encapsulated cells of Azospirillum lipof erum Sp RG6xx from a 50-day-old nitrogen-free, high phosphate6-hydroxybutyrate broth culture. Note apparent continuities between cytoplasms of several cells that appear to have undergone plasmolysis. 1,500X magnification. Bar equals 3.0 \m.

PAGE 140

133 accumulated PHB and capsular material, and sometimes became rounded and nonmotile. Better uptake and growth on BHB occurred with A. brasilense strain Cd. This strain was notable for its formation of numerous weakly motile filaments that sometimes underwent septation and fragmentation into individual cells. The presence of combined nitrogen still allowed PHB accumulation and filament formation of A. brasilense strain Cd, as was also true for A. lipof erum Sp RG6xx. Both A. brasilense strains showed some tendency to form encapsulated microflocs arising from one or more filaments that had undergone septation. The A. brasilense strains and A. lipof erum Sp RG6xx both developed scarlet coloration when cultured as lawns on SNF-Congo Red agar. Sadasivan and Neyra (1985) have shown that azospirilla can produce cellulose as one component of their exopoly saccharides Congo Red is known to stain many polysaccharides, including cellulose (Nakanishi et al., 1974), and colonies of azospirilla grown as dinitrogen-f ixers on agar surfaces take up the dye more avidly than other free-living, dinitrogen-f ixing prokaryotes (Rodriguez Caceres, 1982). It is suggested that Congo Red may profitably be used more often in cultural studies with azospirilla to examine conditions promoting capsule formation. This study indicated that nitrogen-free lawns of A. lipof erum Sp RG6xx produced capsules extensively when cultured on succinate or BHB, whereas both A. brasilense

PAGE 141

134 strains seemed to produce capsules extensively only on succinate All the strains of A. lipof erum used in this study responded uniformly when cultured as lawns on nitrogen-free, LP-BHB agar. They appeared to accumulate PHB and grow as filaments that gradually lost motility. Within 5-7 days, these filaments had accumulated capsular material and become septated. By this time the cells sometimes stuck tenaciously to the agar surface and to glass surfaces as well, and macroflocs were produced. The response could be said to be homogeneous in that over 90% of the cells present were encapsulated microflocs, but the cells themselves were not morphologically homogeneous. All contained large accumulations of PHB, but the size and shape of cells varied. The most extensive studies were done with A. lipof erum Sp RG6xx, but its behavior was typical of other A. lipof erum strains. The appearance of the microflocs did not change noticeably with time when viewed by SEM or phase-contrast microscopy. The PHB content of the cells also remained about the same over time. It would be interesting to see if these microflocs have lower respiration than comparably cultured cells of the same age on SNF agar. The fact that lawns from nitrogen-free, LP-BHB agar consistently gave CFU counts of the same order of magnitude at different culture ages for equivalent OD readings indicates that

PAGE 142

135 the cells were not actively dividing after perhaps a week's time Patriquin et al. (1983) observed spherical, bag-like structures on the surfaces and interiors of 3-week-old and older wheat roots in axenic association with azospirilla. Azospirilla containing PHB granules moved actively within these structures. These structures were similar to the zoogleal-type microflocs observed in LP-TSS broth cultures that had passed into stationary phase. It is probable that these zoogleal forms arose mainly through f ilamentation, followed by septation. Also reported in the previous study were sharply defined, small colonies of azospirilla of apparently determinate size on the surfaces of wheat roots. This is a good description of the microflocs of A. lipof erum from nitrogen-free, LP-BHB lawns. The formation of filaments (Tarrand et al. 1978) and chains (Ruscoe et al., 1978) of azospirilla have also been previously reported. Most laboratory studies of bacteria use cultures in a state of balanced growth, where every component of the cell culture increases at the same rate. This is done for reproducibility of results and standarization of conditions. Cultures that have passed into stationary phase have experienced a metabolic shift-down, growing more slowly and with more widely variable characteristics than log phase cells (Ingraham et al., 1983). It would appear that

PAGE 143

136 azospirilla do not become pleomorphic during balanced growth with combined nitrogen. Pleomorphic forms are only observed in such cultures in stationary phase (Lamm and Neyra, 1981; Papen and Werner, 1982; Krieg and Dobereiner, 1984), after a nutritional down-shift has occurred. Media with high C/N ratios have more often resulted in pleomorphism of azospirilla (Tarrand et al., 1978; Papen and Werner, 1982; Sadasivan and Neyra, 1985). Such conditions generally promote the formation of PHB (Dawes and Senior, 1973) and exopolysaccharides (Sutherland, 1977; Costerton et al., 1981) by bacteria. High C/N ratios may also foster unbalanced growth of azospirilla, leading to pleomorphism. The formation of PHB and capsules by azospirilla seems to lead to a pleomorphic cell type that may reach maturity as a cyst. In these experiments, azospirilla inocula were grown in HP-TSS broth containing combined nitrogen. It is probable that, in this rich medium, the cells had most of their biosynthetic operons repressed, so their biosynthetic enzymes were present only at low levels. After harvest and washing, these cells were exposed to a new carbon source (BHB) and deprived of combined nitrogen, forcing the cells into a severe metabolic shift-down requiring synthesis of biosynthetic enzymes (Ingraham et al., 1983). The azospirilla in two-step, replacement cultures were starved for nitrogen, since azospirilla cannot fix dinitrogen

PAGE 144

137 aerobically. The A. lipof erum strains remained quite healthy in appearance in SEM studies under these conditions. Their ready uptake of BHB leading to PHB accumulation may have allowed them to retain their cellular integrity under conditions of nitrogen starvation. The A. brasilense strains in two-step, broth-replacement cultures did not produce visible exopolysaccharides in SEM studies. The collapsed and shrunken appearance of the A. brasilense strains in SEM studies also suggests that they were not extensively accumulating PHB. Following completion of these studies, it was learned that the high level of phosphates used in HP-BHB, two-step, broth-replacement studies can inhibit growth of azospirilla in the absence of combined nitrogen (Scott et al., 1979; Das and Mishra, 1984). However, the A. lipof erum strains used in these studies still accumulated PHB and capsules in the two-step, broth-replacement studies where this buffer was used. These studies also showed that encapsulation and pleomorphism can occur at near-neutral pH for some A. lipof erum strains. The accumulation of PHB occurred on both HP-BHB and LP-BHB agar. Extensive encapsulation only occurred on nitrogen-free, LP-BHB agar, where the pH may have become alkaline. Alkaline pH may prevent pleomorphic cells from resuming vegetative growth. The sequence of events for A. lipof erum strains grown in BHB broth or on BHB agar was PHB accumulation, followed

PAGE 145

138 by f ilamentation and septation. Cells gradually lost their motility during this time. The septating filaments later produced extensive capsular material on nitrogen-free, LP-BHB agar, and sometimes in HP-BHB broth. The HP buffer repressed extensive capsule formation on nitrogen-free, HP-BHB agar. These and previous studies indicate that pleomorphic cells of azospirilla arise under two different conditions. After cells have experienced balanced growth with combined nitrogen, some may become pleomorphic when a nutrient essential to growth becomes limiting. The production of both filamentous cells and exopolysaccharides seemed to occur extensively in HP-TSS broth in stationary phase, but PHB accumulation was not extensive. Azospirilla may also become pleomorphic during growth where the C/N ratio of nutrients available to the cells is high. Dinitrogen-f ixing cells may be poised to become pleomorphic. It is also evident that azospirilla may take up some carbon sources more rapidly than they can utilize combined nitrogen, resulting in extensive PHB accumulation and capsule formation even when combined nitrogen is available. Cells under these conditions may experience a temporary shift-down, until enough combined nitrogen can be assimilated to mobilize their PHB deposits and enable further growth. This would explain the morphological changes observed in this study with A. brasilense strain Cd

PAGE 146

139 and A. lipof erum Sp RG6xx when they were cultured with BHB and combined nitrogen. The environment near plant roots, where azospirilla are most often found, can be expected to provide available nutrients having a high C/N ratio (Griffin et al., 1976; Beck and Gilmour, 1983; Kraffczyk et al., 1984). Depletion of nitrate from the root zone by physical factors and plant uptake (Okon and Hardy, 1983), in combination with denitrification (Smith and Tiedje, 1979), can be expected to further elevate the C/N ratio of nutrients available to azospirilla. These bacteria have recently been found to grow and fix dinitrogen with straw, which again has a high C/N ratio (Halsall et al., 1985). Azospirilla associated with plant material can be expected to possess capsules and PHB, and sometimes to assume pleomorphic cell shapes. As mentioned before, the most often-reported, pleomorphic form of azospirilla has been a rounded, nonmotile cell possessing a capsule and PHB granules. Individual vegetative cells can likely assume this form without being a member of a microfloc or microcolony. Such rounded cells were sometimes observed within microflocs or after they had broken free of such floes. But the pleomorphic cell types most often observed were filaments or chains of cells that eventually formed the microflocs. Bacterial filaments can arise when cells are growing very rapidly (Slater and Schaechter, 1974), or when the growth rate shifts down

PAGE 147

140 (Jensen and Woolfolk, 1985). Studies reported here indicate that azospirilla may form them under both conditions. Strains of A. lipof erum grown in broth containing combined nitrogen have a greater tendency to clump than do strains of A. brasilense (Krieg and Dobereiner, 1984). In this study, the SEM photographs of stationary phase HP-TSS broth cultures of A. lipof erum Sp RG6xx indicated there is some structural regularity in clumps. The cells were often arranged so that the floes contained spaces. The existence of spaces was more regular and pronounced in floes of cells cultured with BHB in the absence of combined nitrogen, probably because the exopolysaccharides of these cells were more rigid than the exopolysaccharides of stationary phase cultures grown with combined nitrogen, which tended to be slimy. The formation of such microflocs may provide some advantages for azospirilla in nature. Bergersen (1984) discussed the strategies that microaerophilic dinitrogenfixing bacteria such as azospirilla may have to protect their oxygen-sensitive nitrogenase. His suggestions are incorporated below into some of the advantages that azospirilla may find in growing as encapsulated microflocs. 1. The capsules may help to regulate the availability of oxygen to dinitrogen-f ixing cells. Assuming the capsules to be highly hydrated (Costerton et al., 1981) and assuming that water reduces the

PAGE 148

141 diffusion of oxygen by a factor of 10,000 (Clark, 1967), such a role for capsules is not improbable. If the capsular material remains pliable, the cells within an encapsulated microfloc may be able to move closer together or further apart as the situation requiresWhen oxygen is in excess, they may move closer together, reducing the oxygen tension in one spot and thus allowing continued nitrogenase activity at that localized site. When oxygen is limiting, the cells may move further apart, the separation allowing each cell's nitrogenase to remain functional. 2. The spaces between encapsulated cells may provide sites for other bacteria to enter into intimate association with the microfloc, to act in cross-feeding, or to help reduce local oxygen tension. 3. Encapsulation would provide the general benefits to azospirilla that most bacteria seem to derive from encapsulation (Costerton et al., 1981). These benefits include protection from predation and enhanced nutrient accumulation and uptake. The sustained rigid structure of microflocs suggests that nutrients within the encapsulated microflocs may be sequestered from the surrounding environment, giving the azospirilla a storehouse

PAGE 149

of nutrients that may not be readily available to microbial competitors. The encapsulated microfloc may become a fixed site where azospirilla are sometimes able to outcompete motile bacteria for carbon sources. For example, if such a microfloc became established on a root surface, it might be able to continually deplete the carbon supply from that area of root by assimilating it into PHB, without need for further cell division to occur immediately. Competing bacteria without the ability to accumulate PHB might be limited to growing at sites on the root where only balanced growth could occur. If a microfloc is faced with starvation for exogenous carbon sources, some of its members may serve to feed others. If all septa between cells in a microfloc are completed, it could be that these septa are lysed during starvation, so that the substance of dead cells becomes available to healthier cells within the microfloc. Most microflocs contained several different cell morphologies. Septation almost always resulted in cells of significantly different sizes and shapes. Thus, cells within a microfloc formed from a single filament may be destined to attain different physiological states after septation is

PAGE 150

143 completed. Some may be poised to become actively motile as soon as exogenous nutrients enabling balanced growth become available. Others may be primed for continued dinitrogen-f ixation, and still others might enter a truly dormant, cystlike state. Most microflocs observed contained at least one rounded cell that may have been cyst-like. Such a diversity of physiological states within a microcolony may help ensure the persistence of the colony at that site, providing a multiplicity of possible rapid cell responses to environmental conditions. Although these studies failed to produce apparent quantitative, morphological encystment of azospirilla, quantitative encapsulation and microfloc formation of the A. lipof erum strains were obtained. Encapsulation seems to be a prerequisite for encystment in Azotobacter spp. (Eklund et al., 1966). It may be that encysted forms of azospirilla could be obtained in quantity by further nutritional manipulation of the encapsulated cells.

PAGE 151

CHAPTER III PHYSIOLOGICAL PROPERTIES OF ENCAPSULATED FLOCS OF AZOSPIRILLUM LIPOFERUM Sp RG6xx In Chapter II, a method was described for quantitatively converting vegetative, cell-lawn inocula of A. lipof erum strains into nonmotile, encapsulated cells having extensive, intracellular PHB deposits. Although many of the cells within encapsulated floes had vibrioid or ovoid morphologies, some were rounded and cyst-like in appearance. Mature cysts have a lower endogenous metabolic rate and greater ability than do vegetative cells to survive carbon starvation (Sudo and Dworkin, 1973). Mature cysts of Azotobacter spp, are known to be more resistant than vegetative cells to environmental stresses, including desiccation (Socolofsy and Wyss, 1962), Such mature cysts germinate in phosphate buffer containing certain carbon sources, but not in carbon-free phosphate buffer containing ammonium, or in unamended phosphate buffer (Loperfido and Sadoff, 1973). The central bodies of mature cysts are violently and almost immediately expelled from their cyst coats when cysts are suspended in Tris-EDTA (Socolofsky and Wyss, 1961; Goldschmidt and Wyss, 1966; Lin and Sadoff, 1969; Page and Sadoff, 1975). 144

PAGE 152

145 It was of interest to determine if the cells of encapsulated microflocs of A. lipof erum Sp RG6xx were more resistant to desiccation and carbon starvation than were motile, vegetative cells. It was also of interest to define the conditions under which germination occurred, defined as motile cells arising from nonmotile, encapsulated microflocs. Microflocs were exposed to treatment that results in the rupture of capsular coats of mature Azotobacter spp. cysts. Finally, whether chloramphenicol inhibited production of motile cells from a nonmotile encapsulated inoculum was studied. All of these assays represented attempts to determine if floes of A. lipof erum Sp RG6xx contained significant numbers of physiologically cyst-like cells. Materials and Methods Bacterial Strain The only strain used in these studies was Azospirillum lipof erum Sp RG6xx. Its maintenance and subculture were as described in Chapter II. Growth Media and Enumeration Vegetative cells were cultured as previously described in the modified complete medium of Tyler et al. (1979), using the HP buffer of Albrecht and Okon (1980). Encapsulated lawns of the bacterium were cultured on nitrogen-free.

PAGE 153

146 LP-BHB agar, as described in Chapter II. Plate counts were performed on the modified SNF agar medium using LP buffer and Congo Red, as described in Chapter II. Harvest of Cells Harvest and washing of vegetative broth cultures always employed sterile LP-basal salts solution (pH 7.3), as described in Chapter II. Lawns on LP-BHB agar were harvested in two ways. When cells were to be added to semisolid agar or assayed for desiccation and starvation resistance, they were harvested and washed in sterile LP-TSS salts solution. For every other assay that was performed, the cells were harvested and washed in sterile deionized water. Desiccation Resistance Vegetative HP-TSS cultures were grown for 17-22 hours, attaining an OD^^^ of 0.5 in each experiment. The cells were then centrifuged and washed twice in sterile LP-basal salts solution, as described in Chapter II. The cells were resuspended in a third volume of LP-basal salts solution to attain an OD^^q of 0.3 in one experiment and 0.61 in another experiment. Then ten-fold dilution series in sterile LP-basal salts solution were prepared aseptically from the cells. A 0.1-ml portion of the lO"^ dilution was aseptically added to each of three autoclaved 10-ml glass beakers.

PAGE 154

147 contained within an autoclaved 80-min X 10-mm Pyrex storage dish (Fisher Scientific Company, Fair Lawn, NJ), The beakers had been washed in 3.7% (vol/vol) HCl for 24 hours, then rinsed in several changes of deionized water before autoclaving. Encapsulated cells grown on nitrogen-free, LP-BHB agar were harvested and washed in sterile, LP-basal salts solution as described in Chapter II. Cells of 75 days of age were resuspended to give a final 00^^^ of 0.25 in one experiment, and cells of 53 days of age were resuspended to give a final 00^^^ of 0.3 in another experiment. A 0.1-ml portion of cell suspension was aseptically added to each of three, 10-ml beakers housed in a storage dish as described above. Dilution series for plate counts were also prepared as for the vegetative cells. The storage dishes containing cells were placed in a glass desiccator over Drierite at 25 C for 24 hours, by which time the cell suspensions had dried onto the glass surfaces. The desiccator was then placed in a 30 C incubator for 8 days. Initial cell numbers before drying treatment were enumerated by spread plating. For enumeration of cells surviving the desiccation treatment, the beakers were removed from their storage dishes in a laminar flow hood. The dried cell films were outlined with an ink marker to help ensure that they would be resuspended. A volume of 2.0 ml of sterile, LP-basal salts solution was aseptically added

PAGE 155

148 to each beaker, the cells were resuspended by scraping with a flamed wire loop, then mixed with a sterile 1.0 ml pipette. One milliliter from each beaker was used for ten-fold dilution series in sterile, LP-basal salts solution for the purpose of spread-plate counts. Of the remaining volume, 0.1 ml was aseptically pipetted and spread plated onto agar plates. Four plates were spread for each dilution. Carbon Starvation The same washed LP-TSS cell cultures were used for these experiments as for the desiccation experiments. Cells were starved in 50.0-ml, Kimax screw-cap test tubes. The tubes were autoclaved empty, and 8.0 ml of sterile, LP-basal salts solution were later added aseptically to each. The pH of the starvation solution was 7.2. A 2.0-ml volume of each cell type was then added to each of three tubes. The tubes were incubated horizontally on a 130 rpm rotary shaker at 30C for 9 days. For enumeration of cells surviving the starvation treatment, 0.1 ml from each tube was aseptically spread plated onto each of four SNF-Congo Red plates. Ten-fold dilution series in sterile LP-TSS salts were also prepared and enumerated by spread plating.

PAGE 156

149 Microaerobic Incubation The same washed BHB-grown cells were used for these experiments as for the previous experiments. The incubation medium consisted of LP-basal salts solution containing Bacto-Agar (Difco). The agar was added to give a final concentration of 0.05% (wt/vol) per flask after cell addition. The basal salts solution and agar were dissolved by boiling, then 23.5 ml were added per 125-ml Erlenmeyer flask. The flasks were autoclaved, and concentrated sterile LP buffer and biotin were added aseptically soon after autoclaving and before cell addition. A volume of 6.0 ml of encapsulated cell suspension was added to each flask. Flasks were prepared in triplicate and incubated in stationary position at 30C. Aerobic Incubation The BHB-grown, encapsulated cell inoculum for this and all following experiments was harvested, washed twice in sterile deionized water, and resuspended in a third volume of sterile deionized water to give a final OD^^q of 0.23 to 0.25. Lawns of 58 to 66 days of age were used as inocula. For the incubation solution, basal salts were dissolvsed in concentrated amounts to give their final, correct concentrations after aseptic additions of LP buffer, biotin, carbon, or nitrogen sources and cells. The biotin was aseptically added as a lOOX concentrated stock solution, and all other

PAGE 157

150 additions to the LP-basal salts solution were added as lOXconcentrated sterile stock solutions that were sterilized by autoclaving. The sugars employed were D-glucose (Difco), sucrose (Difco), and D-fructose ( Calbiochem, San Diego, CA). All were prepared as separate 4.37% (wt/vol) stocks in deionized water. The organic acids were succinic acid (Fisher) and DL-malic acid (Sigma). Each acid was prepared as a separate 4.37% (wt/vol) salt stock that was neutralized to pH 7.0 with 10 M KOH before being brought to final volume. The nitrogen sources were reagent grade KNO^ and (NH^)2S0^. The KNO^ was prepared as a separate 0.765% (wt/vol) stock, and the (NH^)2S0^ was prepared as a separate 0.5% (wt/vol) stock The same type of screw cap tubes used in carbon starvation studies were used in these studies. Each empty, sterile tube had aseptically added to it 8.0 ml of sterile, LP-basal salts solution plus biotin and a single carbon or nitrogen source. Sometimes cells were incubated in LP-basal salts solution with biotin, but without carbon or nitrogen sources. Then 2.0 ml of water-washed cells were aseptically added to each tube. Treatments for each carbon or nitrogen source were done in triplicate. Tubes were incubated horizontally on a 130 rpm rotary shaker at 30 C. The pH of amended buffer incubations ranged from 7.1 to 7.2.

PAGE 158

151 Soil Dialysis Incubation These cultures were prepared by a modification of the method of Gonzalez-Lopez and Vela (1981). The soil employed was Arredondo fine sand (Grossarenic Paleudult, loamy, siliceous, hyperthemic ) obtained from the top 20 cm of a soil profile. The soil was air dried, then the fraction that would pass through a 0.250 mm sieve was recovered for further use. Spectrapor membrane tubing (Spectrum Medical Industries, Inc., Los Angeles, CA) of 40 mm diameter and 6,0008,000 molecular weight cutoff was used. Tubing was cut into approximately 25 cm long pieces and soaked in a solution of 1.0% (wt/vol) NaHCO^ in which was dissolved 0.005% (wt/vol) EDTA. The pH of this solution was reduced to 7.5 with 1 M HCl before making it up to final volume. The pieces of dialysis tubing were soaked and then boiled in this solution for at least 10 min to remove substances that might harm bacterial cells. The tubing was then rinsed five times in changes of deionized water and resuspended in a sixth wash. One end of each piece of tubing was then knotted once. Five grams of sieved soil were added to each piece of tubing by inserting a glass funnel in the open end of each piece and pouring the soil in. The soil was washed down and evenly distributed in each piece of tubing by adding about 10 ml of deionized water to the open end with a squirt bottle. The open end of each tube was then knotted once.

PAGE 159

152 allowing empty space in each tube to help reduce stress due to expansion of the tubing upon autoclaving. Each piece of tubing was then rocked gently and agitated while holding one end in each hand, to make sure the soil had become thoroughly wetted. Each piece of tubing was then added to 45 ml of deionized water in a 250 ml Erlenmeyer flask and autoclaved for 25 min. The pH of the sterile, equilibrated soil solution surrounding the sterile, intact tubing was 6.1. For cell incubations, 0.5 ml of lOOX sterile biotin was aseptically added, followed by 5 ml of water-washed cell suspension. Triplicate flasks were incubated at 130 rpm at 30C for 24 to 48 hours. The friction between fine soil particles and dialysis tubing caused bags of soil to break during longer shaken incubations. Tris-EDTA Treatment A solution of 30 mM EDTA dissolved in 0.05 M Tris-HCl was prepared, and its pH adjusted to 8.4 with 10 M KOH. A solution of 0.05 M Tris-HCl was also prepared and similarly adjusted to the same pH. Both solutions were sterilized by autoclaving. For lysis experiments, 4 ml of either solution were aseptically added to 4 ml of water-washed, encapsulated cells in 50-ml screw cap tubes, to give a final concentration of 0.025 M Tris-HCl alone or in combination with 15 mM EDTA. Tubes were prepared in triplicate and incubated as for the other tube assays.

PAGE 160

153 Chloramphenicol Treatment A 0.2% (wt/vol) solution of chloramphenicol (Sigma) in deionized water was sterilized by passage through a 0.2-um pore size, Nalgene filter unit. Various volumes of this stock solution were added to water-washed, encapsulated cells in sterile Nutrient Broth (Difco), having a final concentration after all additions of 0.8% (wt/vol). Usually 2.0 ml of cells were added to these tubes. Incubation was in 50-ml screw cap tubes, under the same conditions for tubes as described above. Phase-Contrast Microscopy and Photographs These were the same as in Chapter II. Results Desiccation Resistance Neither the vegetative nor the encapsulated cells displayed significant resistance to the drying method employed. Virtually all cells lost viability during the 8 days of desiccation. Table 3-1 gives the results of two separate desiccation experiments. No statistical analyses were performed, because even if some difference could be revealed between vegetative and encapsulated cells, the survival of either was so poor as to be negligible.

PAGE 161

154 Table 3-1. Desiccation resistance of Azospirillum lipof erum Sp RG6xx. Experiment Cell type Initial, CFU ml Beaker CFU ml f inal ^ b survivors Number (X 10^) dried 1 2 3 1 Vegetative 0.72 I U U i } 0.0 (0.0) 22.5 (13.0) 5.0 (5.0) 0 03 Encapsulated 0.42 (0.06) 92.5 (37.7) 55.0 (28.7) e 0.35 2 Vegetative 0.14 (0.03) 10.0 ( 7.1) 0.0 (0.0) 0.0 (0.0) 0.05 Encapsulated 1.34 (0.10) 2.5 (4.3) 25.0 (43.0) 40.0 (17.3) 0.03 Averages of colony forming units (CFU) of four spread plates Values in parentheses are standard deviations. Average of all beakers available. Averages of four spread plates. Values in parentheses are standard deviations. 1 75-day-old cells. 'Beaker was lost. "53-day-old cells.

PAGE 162

155 Starvation Resistance After 9 days of starvation in carbon-and-nitrogen-f ree LP-basal salts solution lacking biotin, the encapsulated cells retained 24 to 27% viability. On the other hand, vegetative cells multiplied several-fold and retained viability (Table 3-2). It was interesting that, although two different initial densities of vegetative inocula were used, each seemed to stabilize at about 10^ CFU/ml after 9 days of starvation. Germination Experiments Table 3-3 gives a summary of the germination experiments involving semisolid agar, buffered salts solution containing single carbon or nitrogen sources, and soil dialysis flasks. The soil dialysis flasks and combined-nitrogen incubations all resulted in germination of encapsulated cells. Within 18 hours, the majority of cells in these treatments were extremely motile, vegetative cells. By this time the motile cells in soil dialysis flasks had lost most or all visible deposits of phase-bright PHB granules. The PHB granules were usually still visible in the nitrate and ammonium treatments after 24 hours, but were markedly reduced in size from those in the initial inoculum. Within 72 hours the extremely motile cells in the combined-nitrogen treatments assumed the totally phase-dark appearance of motile cells in soil dialysate. Figure 3-1

PAGE 163

156 Table 3-2. Starvation resistance of Azospirillum lipof erum Sp RG6xx. Experiment Number Cell type Initial, CFU ml (X 10^)^ Final CFU ml (X 10^)^ % Initial CFU ml 1 Vegetative 0.72 (0.03) 1.02 (0.26) 142.04 Encapsulated^ 0.42 (0.06) 0.11 (0.01) 26.61 2 Vegetative 0.14 (0.03) 1.02 (0.22) 751.23 Encapsulated 1.34 (0.10) 0.33 (0.18) 24.22 Averages of colony forming units (CFU) of four spread plates. Values in parentheses are standard deviations. Averages of three starvation tubes, with four spread plate counts per tube. Values in parentheses are standard deviations 75-day-old cells. 53-day-old cells.

PAGE 164

157 Table 3-3. Response of encapsulated cells of Azospirillum lipof erum Sp RG6xx to various incubations. a Treatment Germination Soil dialysate + NO3" + nh/ + Glucose Fructose Sucrose Malate Succinate Aerobic, low phosphate, basal salts solution Microaerobic, low phosphate, basal salts solution ^+ denotes the majority of cells present becoming motile and depleting their visible poly6-hydroxybutyrate reserves within 29 hours.

PAGE 165

Figure 3-1. Encapsulated cells of Azospirillum lipof erum Sp RG6xx that have undergone germination in low phosphate-basal salts solution with combined nitrogen. a) Cells from a 29-hour nitrate incubation. Note germinated vegetative cell at left of an empty capsule. 1,000X magnification. Bar equals 4.0 um. b) Cells from a 29-hour ammonium incubation. Note germinated vegetative cells, empty capsules and poly6-hydroxybutyrate-rich cells remaining in the floe. 1,000X magnification. Bar equals 4 0 um.

PAGE 166

160 shows cells from 29-hour incubations in nitrate (Figure 3-la) and ammonium (Figure 3-lb), and Figure 3-2 shows cells from a 29-hour incubation in soil dialysate. As is apparent in these photographs, most encapsulated floes had germinated, retaining their general shape after most of their cells had left them. The pH of inoculated soil dialysis flasks rose to 6.4 to 6.5 within 24 hours after inoculation. Encapsulated floes suspended in buf f ered-salts solution produced a few motile, ovoid to peanut-shaped cells that remained swollen with intracellular PHB deposits. This was also true in the semisolid agar flasks and in the bufferedsalts solution tubes containing single carbon sources. Sometimes these PHB-rich cells were as rapidly motile as the cells that germinated with combined nitrogen and soil dialysate. Usually, however, they moved slowly and were prone to long periods of twiddling before they actively moved off on a run. Although these treatments often resulted in large numbers of free, nonencapsulated cells, most of these individual cells were not motile. The original inocula did not contain even weakly motile cells and contained few individual cells. Figure 3-3 shows cells from a 29-hour incubation in buf f ered-salts solution containing glucose, an incubation where the floes remained largely occupied with cells.

PAGE 167

Figure 3-2. Encapsulated cells of Azospirillum lipof erum Sp RG6xx that have undergone germination in soil dialysis flasks. a) Cells from a 29-hour-old incubation. Note empty horseshoe-shaped capsules and phasedark vegetative cells. 1,000X magnification. Bar equals 4.0 urn. b) Other cells from a 29-hour incubation. Note empty horseshoe-shaped capsules, phasedark vegetative cells and poly6-hydroxybutyrate-rich cells remaining in the floe. 1,000X magnification. Bar equals 4.0 lam

PAGE 168

Figure 3-3. Encapsulated cells of Azospirillum lipof erum Sp RG6xx that have not undergone widespread germination in low phosphatebasal salts solution with glucose, a) Cells froma 29-hour incubation. A few empty capsules are present, but most capsules in the microfloc still retain their cells. 1,000X magnification. Bar equals 4.0 ym. b) Microfloc from a 29-hour incubation with several empty capsules. Note rounded cyst-like appearance of some cells remaining in the floe. 1,000X magnification. Bar equals 4.0 jam.

PAGE 170

165 After several weeks of incubation, some morphological changes were observed for cells in the treatments that were scored as negative at 29 hours. Floes that were incubated in aerobic, LP-basal-salts solution for 60-65 days had somewhat the same appearance as germinated floes, in that most capsular spaces appeared empty or contained cells of unhealthy appearance. Less than 5% of the capsules in this extended incubation contained PHB-rich cells. The cells apparently had not germinated, since no individual vegetative cells were observed outside of the floes. Cells apparently had depleted their visible PHB reserves and starved in place within their capsules. The addition of carbon sources seemed to reduce the extensive loss of visible PHB deposits. The number of cells within a floe that retained visible PHB deposits and an overall viable appearance after 47 days of incubation ranged from about 5 to 20% in aerobic, carbon-amended incubations. Often these cells appeared somewhat reduced in size, having contracted from the capsule boundary while retaining their rounded appearance. A few individual, motile, PHB-rich cells were also present in carbon-amended incubations at this time. The nitrate and ammonium incubations at 4 7 days consisted of empty capsules and nonmotile, phase-dark vegetative cells which had the appearance of stationary phase cells. After 60 to 65 days, the microaerobic incubations were similar in appearance to the aerobic, carbon-amended incubations, with

PAGE 171

many PHB-rich cells still contained within capsules. Unlike all the other extended incubations, only the microaerobic treatment contained numerous, motile, phase-dark vegetative cells Tris-EDTA Treatment When encapsulated floes were incubated in Tris-EDTA solution, the only observable effect was dissolution of macroscopic floes within 24 hours. To the naked eye, the cell suspensions appeared more evenly turbid and nonflocculated than those from any other treatment. When viewed by microscopy, however, microflocs were still present. They were of about the same size range as microflocs in other treatments, and had about as many cells remaining in their capsules. Few motile cells were observed in this treatment. There were not noticeably fewer cells within capsules than in the other treatments where overall germination did not occur. The same results were obtained when encapsulated floes were incubated in Tris alone. Effect of Chloramphenicol The addition of 50 \ig chloramphenicol/ml to nutrient broth prevented growth for at least 36 hours, whereas nutrient broth tubes without the antibiotic became turbid within 18 hours with the same inoculum level. Even in the presence of chloramphenicol, however, a few weakly motile.

PAGE 172

PHB-rich cells were observed. Each was rotating about its own long axis without moving off on runs. Discussion The desiccation resistance assay used in these studies involved rapid drying of the cells. Mature, encysted forms of some prokaryotes are much better able to withstand drying on glass surfaces than their vegetative counterparts (Whittenbury et al., 1970a; Tudor and Conti, 1977). Filter membranes are often used as a surface upon which cells are slowly dried in desiccation resistance assays (Socolofsky and Wyss, 1962). In such experiments (data not shown), vegetative cells of azospirilla sometimes survive slow drying on membrane filters without appreciable die off, as do encapsulated cells. Webb (1965) has pointed out the difficulties of using membrane filters in such assays. It was thought that glass surfaces would be easier to use, with less inherent hydrophilic behavior than membrane filters. Rapid drying of cells within a day's time or less usually causes a rapid and nearly complete die off of vegetative cells for many genera (Robinson et al., 1965; Whittenbury et al., 1970a; Tudor and Conti, 1977). The results of the work reported here show no apparent significant difference in response to rapid drying between vegetative or encapsulated cells of A. lipof erum Sp RG6xx. This is in agreement with studies where capsules have not enhanced the ability of

PAGE 173

168 cells to survive desiccation (Dudman, 1977; Lowendorf, 1980) Other studies have found enhanced desiccation resistance for some encapsulated forms of azospirilla. Such forms have been called cysts (Lamm and Neyra, 1981; Papen and Werner, 1982; Sadasivan and Neyra, 1985). In the first study, however, cyst-like forms were slowly dried while vegetative cells were rapidly dried, so a true comparison of desiccation resistance between forms seems questionable. In the last study, floes of cells were stored without desiccant for long time periods, and some cells within pieces of floes were shown to remain viable for up to six months. But whether this reflects a true desiccation resistance of single cells, or only an ability to withstand drought in the presence of hydrophilic polymers (Jagnow, 1982), is unclear. Perhaps the best studies were done by Papen and Werner (1982), where vegetative cells were found to almost completely die out rapidly on dried membrane filters, whereas encapsulated, nonmotile cells survived this treatment in high numbers. Azospirilla may undergo extensive morphological changes, including accumulation of PHB and capsular material, as only a first step towards becoming mature cysts. Morphologically cyst-like azospirilla have exhibited variable responses to desiccation (Lamm and Neyra, 1981; Papen and Werner, 1982). In the studies reported here, many

PAGE 174

cells in microflocs had a cyst-like appearance. However, cells in microflocs survived desiccation no better than nonencapsulated vegetative cells. Encysting bacteria may gradually mature, with their resistance to environmental stress increasing with time although their morphology appears cyst-like throughout. Modification of the cell membrane may be the true key to stress-resistance of these forms, as may also be true of Azotobacter spp. cysts (Reusch and Sadoff, 1983). Immature cysts of azospirilla may mature best if they are removed from the growth medium and/or slowly dried. Sadasivan and Neyra (1985) removed their encapsulated azospirilla from broth and dried the floes slowly, and possibly obtained desicccation resistant forms. Similarly, Lamm and Neyra (1981) obtained desiccationresistant azospirilla by allowing the agar of lawns to dry slowly. There was a significant difference in response to carbon and nitrogen starvation between vegetative and encapsulated cells of A. lipof erum Sp RG6xx. After 9 days, only about 25% of the original encapsulated inoculum retained viability. The cell densities used in these starvation experiments were less than 10^ CFU/ml so microscope observations were not performed. However, the microscope observations of extended germination incubations with higher cell densities indicated that the majority of PHB-rich cells within floes will eventually deplete their PHB reserves and

PAGE 175

170 apparently lose viability due to starvation for exogenous carbon. Most encapsulated cells within floes do not become motile when confronted with starvation for exogenous carbon. From 5 to 25% of cells within floes retained their PHB after extended incubation and often underwent reduction in size within their capsules. One interpretation of this size reduction is that the cells were undergoing maturation into physiologically mature cysts. In contrast to the encapsulated cells, two different densities of vegetative inoculum increased several fold during the same 9 days of starvation to give about 10^ CFU/ml. This apparent multiplication to a certain cell density and continued viability of vegetative cells faced with starvation have a precedent in studies of other bacteria, such as Rhizobium japonicum (Crist et al. 1984 ) In a recent study, Tal and Okon (1985) reported that vegetative, PHB-poor cells of A. brasilense strain Cd died off to about 7% of their initial numbers after 130 hours of starvation in sterile, 0.06 M potassium phosphate buffer. In comparison, PHB-rich cells proliferated 2.3-fold over the same time span. Their experimental conditions differed from conditions reported here not only in using a different bacterial species but also in incubating the cells in phosphate buffer alone, without other salts. Both of their cell types were apparently nonencapsulated and actively motile, while the encapsulated cells in this study were not

PAGE 176

motile. There are numerous starvation conditions that bacteria can be exposed to in vitro, and the response of different strains can vary widely under different conditions. It would be interesting to follow up on these initial studies of starvation resistance of azospirilla, to gain further insight into how they might survive in soil in the absence of plant material. It is possible that some azospirilla are able to enter into two types of dormancy (Marshall, 1980) in unfavorable soil conditions. If the cells have been experiencing balanced growth before they are starved of exogenous carbon, they might enter into exogenous dormancy. Such cells would be poor in PHB and might have no different morphology than growing vegetative cells, but their metabolism would be greatly reduced. If the cells have accumulated large amounts of PHB through dinitrogen fixation or extremely rapid uptake of carbon sources during growth on combined nitrogen, they might be prone to enter constitutive dormancy or an encysted state when faced with starvation. Papen and Werner (1982) suggested that depletion of available oxygen was responsible in part for encystment of azospirilla in their studies. The low oxygen consumption of encysted cells may have allowed oxygen to diffuse back into the medium from the headspace, whereupon vegetative cells emerged from the capsular coats and resumed dinitrogen fixation. Because of their suggestion, in this study

PAGE 177

encapsulated floes were incubated microaerobically in semisolid agar, as well as aerobically in shaken broth tubes, to give the cells different oxygen regimes. There was no observable difference in response between floes incubated microaerobically or aerobically in buffered salts solution, with or without single carbon sources. In each of these treatments, a few weakly motile, PHB-rich cells were observed within 18 hours, and they persisted for up to 10 days. Half of the capsular spaces in some floes were empty, but most floes retained the majority of their cells within capsules. Unlike mature Azotobaeter spp. cysts, encapsulated cells of this strain did not become synchronously motile when exposed to carbon sources (Loperfido and Sadoff, 1973). Phosphate-buff ered-salts solutions and bufferedsalts solutions containing sucrose, a nonmetabolizable carbon source for this species (Tarrand et al., 1978), produced about as many motile cells as did metabolizable carbon sources. Like mature Azotobaeter spp. cysts, encapsulated cells of this strain seemed unable to mobilize their PHB reserves in unamended buffered salts solution to enable germination and widespread motility (Loperfido and Sadoff, 1973). Either wetting released cells from floes and these cells became motile, or else the cells became motile and actively left floes after wetting. Perhaps both events occurred. In any case, most individual cells were not motile in these incubations. The semisolid agar flasks

PAGE 178

contained enough cells in the initial inoculum to be visibly turbid, but the cells remained dispersed throughout the agar and formed no pellicle. As a result, these treatments are listed as giving no germination (Table 3-3). A few cells seemed able to slowly mobilize their PHB reserves and become motile under these conditions, but the majority of cells remained in the floe, or once free from the floe, remained nonmotile. Motile cells continued to retain extensive visible deposits of PHB. Undeniable germination occurred when the floes of encapsulated cells were added to soil dialysis flasks, or to buf f ered-salts solution containing nitrate or ammonium. The uniformity of response among these treatments indicates that combined nitrogen in the soil dialysate was responsible for its germination effect. It also indicates that most of the cells in the floes were not similar to mature Azotobacter spp. cysts, which do not germinate in the presence of ammonium (Loperfido and Sadoff, 1973). The availability of combined nitrogen apparently prompted most of the cells in encapsulated floes to mobilize their PHB reserves and return to an actively motile, vegetative state. An interesting feature of these positive germination treatments was the persistence of nonmotile, PHB-rich cells within floes even after 10 days of incubation. Some of these cells no longer possessed a plump appearance, and their PHB granules were dispersed irregularly within the

PAGE 179

cytoplasm. Sometimes capsules seemed to contain PHB and polyphosphate granules without any cytoplasm. Empty capsules, however, often possessed the "horseshoe" shape typical of empty exines of mature Azotobacter spp. cysts (Lin and Sadoff, 1969). The Tris-EDTA incubations did not produce any obvious expulsion of cells from capsules. The concentration of EDTA was about five times that which produces prompt expulsion of central bodies from mature Azotobacter spp. cysts (Lin and Sadoff, 1969). As mentioned earlier, however, the macroscopic appearance of the inocula was rendered more evenly turbid by this treatment. Incubation in Tris buffer alone had the same effect as Tris-EDTA. The high pH of the treatments (pH 8.4) may have been related to the dispersive effect, along with the chelating effects of the Tris and the EDTA. It seems that protein synthesis is necessary before encapsulated cells are able to become motile in large numbers. The chloramphenicol treatment did not prevent some free cells from spinning about their own long axes, however. Based on these tests, there appears to be little physiological similarity between mature cysts of Azotobacter spp. and most cells in the encapsulated floes of A. lipof erum Sp RG6xx. Most of the cells in encapsulated floes represent immature cysts, lacking desiccation resistance, but being largely nonmotile and unable to readily mobilize

PAGE 180

their PHB reserves unless exogenous, combined nitrogen becomes available. If these cells are found to have a very low endogenous respiratory rate, it might further indicate their state as nascent cysts. Living cells may form structures that prove immediately useful for some functions. By chance these structures may also prove beneficial to the cells in other ways (CairnsSmith, 1982). It is suggested that the microflocs of azospirilla are such structures. Their possible benefits were suggested in Chapter II. Four observations in this chapter deserve further comment. One is the great size difference between motile, vegetative cells of A. lipof erum Sp RG6xx and nonmotile, encapsulated cells. The encapsulated cells occupy much more volume. Secondly, Costerton et al. (1981) suggested that most bacterial cells in nature assume two forms. Sessile forms surrounded by a capsule maintain a population on a surface and give rise to motile swarmer cells which colonize new surfaces. This is a good description of the conversion of encapsulated to motile forms of A. lipof erum Sp RG6xx. Thirdly, the ability of only 5 to 20% of cells within encapsulated floes to retain their visible PHB deposits over 65 days of aerobic nitrogen starvation may indicate the physiological diversity of cells within an encapsulated PHB-rich microfloc. Most cells may be poised to become motile or resume vegetative growth and may represent the cells that depleted their PHB reserves

PAGE 181

176 during starvation and which finally lost their viability. The remaining cells within the microfloc may include cells that eventually mature into cysts. Finally, the similar "horseshoe" appearance of some empty capsules of azospirilla and of germinated cysts of Azotobacter spp. may have some importance. More than one type of capsule may exist within an encapsulated microfloc of azospirilla, and some capsules may have proceeded further toward a cyst-coat composition than others. A few encapsulated cells in floes survived the desiccation treatment. A few cells within floes also had the appearance of rounded, possibly mature cysts. Sometimes they broke free of floes (Figure 3-4). They were never observed to be motile. If these are truly mature cysts, the problem remaining is how to convert most of the cells in a vegetative inoculum quantitatively into this form.

PAGE 182

Figure 3-4. Rounded, possibly cyst-like cells of Azospirillum lipof erum Sp RG6xx, from a 2 9-hour incubation in low phosphate-basal salts solution with glucose. 1,000X magnification. Bar equals 4.0 um.

PAGE 183

CHAPTER IV GENERAL CONCLUSIONS The strains of azospirilla used in this study did not undergo quantitative morphological encystment when grown on nitrogen-free BHB agar. The strains of Azospirillum lipof erum synthesized exopolysaccharides more extensively than did the strains of Azospirillum brasilense The A. lipof erum strains experienced unbalanced growth under these cultural conditions. They accumulated PHB and experienced unbalanced cell wall synthesis, as evidenced by the common formation of filaments and chains. Eventually the filaments or chains lost motility and accumulated capsular material. The final outcome was the formation of microflocs of encapsulated, PHB-rich cells that often arose from only a few elongated cells. Some of the cells within these floes had a cyst-like morphology. Environments with available nutrients having a high C/N ratio, such as the rhizosphere, may promote the formation of PHB and capsules by azospirilla. Some cells having these features may eventually form cysts. Cells in encapsulated microflocs may have some survival advantages that individual cells of azospirilla lack. 178

PAGE 184

Encapsulated cells of Azospirillum lipof erum Sp RG6xx grown on nitrogen-free BHB agar were found to have far more PHB than cells grown in broth with combined nitrogen. Neither cell type displayed significant desiccation resistanceWhen faced with aerobic starvation for exogenous carbon and nitrogen, encapsulated cells of this strain died off after 9 days to about 25% of their original numbers. These survivors may have represented cells that were maturing into cysts. Vegetative cells grown with combined nitrogen multiplied several fold over the same period of starvation. This indicates that cells of this strain may not need to form cysts in nature to survive prolonged periods of starvation. Vegetative cells having reduced metabolic activity may survive such periods. Combined nitrogen promoted germination of nonmotile, encapsulated cells of this strain. Although these studies failed to obtain cysts of azospirilla in high quantity, they may have provided some information of practical importance. The starvation studies suggested that encapsulated, PHB-rich cells of azospirilla are less active physiologically than motile, vegetative cells. Further studies might measure nitrogenase activity and plant growth substance production by azospirilla in relation to PHB deposition and capsule formation. Such studies should lead to an understanding of what physiological form of azospirilla is most beneficial to plant growth.

PAGE 185

180 Inoculum production may then be designed to introduce and sustain the most beneficial physiological form in the root zone.

PAGE 186

BIBLIOGRAPHY Abd-El-Malek, Y. and Y. Z. Ishac. 1966. Longevity of Azotobacter Plant and Soil 24:325-327. Albrecht, S. L. and Y. Okon. 1980. Cultures of Azospirillum p. 740-479. In Methods in enzymology, vol. 69. Academic Press, Inc., New York. Bae, H. C, E. H. Cota-Robles, and L. E. Casida, Jr. 1972 Microflora of soil as viewed by transmission electron microscopy. Appl. Microbiol. 23:637-648. Barak, R. I. Nur, Y. Okon, and Y. Henis. 1982. Aerotactic response of Azospirillum brasilense J. Bacterid. 152 :643-649. Beck, S. M. and C. M. Gilmour. 1983. Role of wheat root exudates in associative nitrogen fixation. Soil Biol. Biochem. 15:33-38. Becking, J. H. 1982. Azospirillum lipof erum — A reappraisal, p. 127-146. In W. Klingmuller (ed.), Azospirillum : Genetics, physiology, ecology. Experientia Supplementum, vol. 42. Birkhauser Verlag, Basel, Switzerland. Begg, K. J., and W. D. Donachie. 1985. Cell shape and division in Escherichia coli: Experiments with shape and division mutants. J. Bacterid. 163:615-622. Berg, R. H. V. Vasil, and I. K. Vasil. 1979. The biology of Azospirillum -sugarcane association II. Ultrastructure. Protoplasma 101:143-163. Berg, R. H. M. E. Tyler, N. J. Novick, V. Vasil, and I. K. Vasil. 1980. Biology of Azospirillum -sugarcane association: Enhancement of nitrogenase activity. Appl. Environ. Microbiol. 39:642-649. Bergersen, F. J. 1984. Oxygen and the physiology of diazotrophic microorganisms, p. 171-180. In C. Veeger and W. E. Newton (ed.). Advances in nitrogen fixation research. Ni jhof f /Junk, Pudoc, Wageningen, The Netherlands 181

PAGE 187

182 Bothe, H. B. Klein, M. P. Stephan, and J. Dobereiner. 1981. Transformations of inorganic nitrogen by Azospirillum spp. Arch. Microbiol. 130:96-100. Cagle, G. D. and G. R. Vela. 1974. Giant cysts and cysts with multiple central bodies in Azotobacter vinelandii J. Bacterid. 107:315-319. Cairns-Smith, A. G. 1982. Genetic takeover and the mineral origins of life, p. 107-109. Cambridge University Press, Cambridge. Clark, F. E. 1967. Bacteria in soil, p. 15-49. In A. Burgess and F. Raw (ed.). Soil biology. Academic Press, New York. Costerton, J. W. J. M. Ingram, and K.-J. Cheng. 1974. Structure and function of the cell envelope of Gramnegative bacteria. Bacterid. Rev. 38 :87-110. Costerton, J. W. R. T. Irvin, and K.-J. Cheng. 1981. The bacterial glycocalyx in nature and disease. Ann. Rev. Microbiol. 35:299-324. Crist, D. K., R. E. Wyza, K. M. Mills, W. D. Bauer, and W. R. Evans. 1984. Preservation of Rhizobium viability and symbiotic infectivity by suspension in water. Appl. Environ. Microbiol. 47:895-900. Das, A., and A. K. Mishra. 1984. Aerotolerant growth in Azospirillum brasilense induced by dihydroxyphenyl iron-binding compound. Curr. Microbiol. 11:313-136. Dawes, E. A. 1976. Endogenous metabolism and the survival of starved prokaryotes, p. 19-53. In T. R. G. Gray and J. R. Postgate (ed.). The survival of vegetative microbes. Twenty-sixth Symposium of the Society for General Microbiology. Cambridge University Press, Cambridge, London. Dawes, E. A. 1981. Carbon metabolism, p. 1-38. In P. H. Calcott (Ed.), Continuous cultures of cells, vol. II. CRC Press, Inc., Boca Raton, FL. Dawes, E. A., and P. J. Senior. 1973. The role and regulation of energy reserve polymers in microorganisms, p. 135-266. In A. H. Rose and D. W. Tempest (ed.). Advances in microbial physiology, vol. 10. Academic Press, New York.

PAGE 188

183 Day, J. M. and J. Dobereiner. 1976. Physiological aspects of N^-fixation by a Spirillum from Digitaria roots. Soil Biol. Biochem. 8:45-40. Dobereiner, J. 1978. Influence of environmental factors on the occurrence of Spirillum lipof erum in soils and roots. In U. Granhall (ed.). Environmental Role of Nitrogen-fixing Blue-green Algae and Asymbiotic Bacteria. Ecol. Bull. (Stockholm) 26:343-352. Dobereiner, J., I. E. Marriel, and M. Nery. 1976. Ecological distribution of Spirillum lipof erum Beijerinck, Can. J. Microbiol. 22:1464-1473. Donachie, W. D. K. J. Begg, and N. F. Sullivan. 1984. Morphogenes of Escherichia coli p. 27-62. In R. Losick and L. Shapiro (ed.). Microbial development. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Dudman, W. F. 1977. The role of surface polysaccharides in natural environments, p. 357-414. In I. W. Sutherland (ed.). Surface carbohydrates of the prokaryotic cell. Academic Press, New York. Duguid, J. P., and J. F. Wilkinson. 1961. Environmentally induced changes in bacterial morphology, p. 69-99. In G. G. Meynell and H. Gooder (ed.). Microbial reaction" to environment. Eleventh Symposium of the Society for General Microbiology. Cambridge University Press, Cambridge Eklund, C, L. M. Pope, and 0. Wyss. 1966. Relationship of encapsulation and encystment in Azotobacter J. Bacterid. 92 :1828-1830. Eskew, D. L. D. D. Focht, and I. P. Ting. 1977. Nitrogen fixation, denitrif ication, and pleomorphic growth in a highly pigmented Spirillum lipof erum Appl. Environ. Microbiol. 34:582-585. Falk, E. C. J. Dobereiner, J. L. Johnson, and N. R. Krieg. 1985. Deoxyribonucleic acid homology of Azospirillum amazonense Magalhaes et al. 1984 and emendation of the description of the genus Azospirillum Int. J. Syst. Bacterid. 35:117-118. Foster, R. C. and G. D. Bowen. 1982. Plant surfaces and bacterial growth: The rhizosphere and rhizoplane, p. 159-185. In M. S. Mount and G. H, Lacy (ed.). The phytopathogenic prokaryotes, vol. I. Academic Press, Inc. New York.

PAGE 189

184 Geesey, G. G. 1982. Microbial exopolymers : Ecological and economic considerations. Am. Soc. Microbiol. News 48:9-14. Goldschmidt, M. C. and 0. Wyss. 1966. Chelation effects on Azotobacter cells and cysts. J. Bacterid. 91:120-124. Gonzalez-Lopez, J., M. V. Martinez-Toledo, J. Moreno, F. Ballesteros, and A. Ramos-Cormezana 1985. Resistant properties of Azotobacter cysts induced in phosphate-limited media. Microbios Letters 28:151-161. Gonzalez-Lopez, J., and G. R. Vela. 1981. True morphology of the Azotobacteraceae -Filter able bacteria. Nature 289:588-590. Gray, T. R. G. and S. T. Williams. 1971. Microbial productivity in soil, p. 255-287. In D. E. Hughes and A. H. Rose (ed.). Microbes and biological productivity. Twenty-first Symposium of the Society for General Microbiology. Cambridge University Press, Cambridge. Griffin, G. J., M. G. Hale, and F. J. Shay. 1976. Nature and quantity of sloughed organic matter produced by roots of axenic peanut plants. Soil Biol. Biochem. 8 :29-32. Halsall, D. M. G. L. Turner, and A. H. Gibson. 1985. Straw and xylan utilization by pure cultures of nitrogen-fixing Azospirillum spp. Appl. Environ. Microbiol. 49:423-428. Hitchins, V. M. and H. L. Sadoff. 1970. Morphogenesis of cysts in Azotobacter vinelandii. J. Bacterid. 104: 492-498. Hitchins, V. M. and H. L. Sadoff. 1973. Sequential metabolic events during encystment of Azotobacter vinelandii J. Bacterid. 113 :1273-1279. Hughes, W. H. 1956. The structure and development of the induced long forms of bacteria, p. 341-362. In E. T. C. Spooner and B. A. D. Stocker (ed.). Bacterial anatomy. Sixth Symposium of the Society for General Microbiology. Cambridge University Press, Cambridge. Ingraham, J. L. 0. Maaloe, and F. C. Neidhardt. 1983. Growth of the bacterial cell. Sinauer Associates, Inc. Publishers, Sunderland, MA.

PAGE 190

185 Jagnow, G. 1982. Growth and survival of Azospirillum lipof erum in soil and rhizospheres as influenced by ecological stress conditions, p. 97-104. In W. Klingmuller (ed.), Azospirillum : Genetics, physiology, ecology. Experientia Supplementum, vol. 42. Birkhauser Verlag, Basel, Switzerland. Jain, D. K. and D. G. Patriquin. 1985. Characterization of a substance produced by Azospirillum which causes branching of wheat root hairs. Can. J. Microbiol. 31:206-210. Jansen van Rensburg, H. and B. W. Strijdom. 1980. Survival of fastand slow-growing Rhizobium spp. under conditions of relatively mild desiccation. Soil Biol. Biochem. 12:353-356. Jarman, T. R. and G. W. Pace. 1984. Energy requirements for microbial exopolysaccharide synthesis. Arch. Microbiol. 137:231-235. Jensen, R. H. and C. A. Woolfolk. 1985. Formation of filaments by Pseudomonas putida Appl. Environ. Microbiol. 50:364-372. Kapulnik, Y. R. Gafny, and Y. Okon. 1985. Effect of Azospirillum spp. inoculation on root development and NO, uptake in wheat ( Triticum aestivum cv. Miriam) in hydroponic systems. Can. J. Bot. 63:627-631. Karnovsky, M. J. 1965. A f ormaldehyde-glutaraldehyde fixative of high osmolarity for use in electron microscopy. J. Cell. Biol. 27; (2, Pt. 2):137-138a. Keynan, A. 1972. Cryptobiosis : A review of the mechanisms of the ametabolic state in bacterial spores, p. 355-362. Tn H. 0. Halvorson, R. Hanson, and L. 0. Campbell (ed.). Spores V. American Society for Microbiology, Washington, DC. Kraffczyk, I., G. Trolldenier, and H. Beringer. 1984. Soluble root exudates of maize: Influence of potassium supply and rhizosphere microorganisms. Soil. Biol. Biochem. 16:315-322. Kramer, M. J., and M. D. Socolofsky. 1970. Effect of chloramphenicol on the conversion of dormant Azotobacter cysts into vegetative cells. Texas Rep. Biol. Med. 1 & 2:43-48.

PAGE 191

186 Krieg, N. R. and J. Dobereiner. 1984. The genus Azospirillum p. 94-104. In N. R. Krieg and J. G. Holt (ed.), Bergey's manual of systematic bacteriology, vol. 1. The Williams & Wilkins Co., Baltimore. Labeda, D. P., K.-C. Liu, and L. E. Casida, Jr. 1976. Colonization of soil by Arthrobacter and Pseudomonas under varying conditions of water and nutrient availability as studied by plate counts and transmission electron microscopy. Appl. Environ. Microbiol. 31:551-561. Lakshmi, V., A. Satyanarayana Rao, K. Vi jayalakshmi M. Lakshmi-Kumari K. V. B. R. Tilak, and N. S. Subba Rao. 1977. Establishment and survival of Spirillum lipof erum Proc. Indian Acad. Sci. 86B:397-404. Lamm, R. B. and C. A. Neyra. 1981. Characterization and cyst production of azospirilla isolated from selected grasses growing in New Jersey and New York. Can. J. Microbiol. 27:1320-1325. Law, J. H. and R. A. Slepecky. 1961. Assay of poly-6hydroxybutyric acid. J. Bacterid. 82 :33-36. Leive, L. L., and B. D. Davis. 1980. Cell envelope; spores, p. 73-110. In B. D. Davis, R. Dulbecco, H. N. Eisen, and H. S. Ginsberg (ed.). Microbiology, 3rd ed. Harper & Row, Publishers, Philadelphia. Lin, L. P., and H. L. Sadoff. 1968. Encystment and polymer production by Azotobacter vinelandii in the presence of -hydroxybutyrate. J. Bacterid. 95:2336-2343. Lin, L. P., and H. L. Sadoff. 1969. Preparation and ultrastructure of the outer coats of Azotobacter vinelandii cysts. J. Bacterid. 98 :1335-1341. Lin, W. Y. Okon, and R. W. F. Hardy. 1983. Enhanced mineral uptake by Zea mays and Sorghum bicolor roots inoculated with Azospirillum brasilense Appl. Environ. Microbiol. 45:1775-1779. Loperfido, B. and H. L. Sadoff. 1973. Germination of Azotobacter vinelandii cysts: Sequence of macromolecular synthesis and nitrogen fixation. J. Bacterid. 113 :841-846. Lowendorf, H. S. 1980. Factors affecting survival of Rhizobium in soil, p. 87-124. In M. Alexander (ed.). Advances in microbial ecology, vol. 4. Plenum Press, New York.

PAGE 192

187 Marshall, K. C. 1980. Bacterial adhesion in natural environments, p. 187-196. In R. C. W. Berkeley, J. M. Lynch, J. Melling, P. R. Rutter, and B. Vincent (ed.). Microbial adhesion to surfaces. Ellis Horwood Limited Publishers, Chicester. Martinez-Drets G., M. Del Gallo, C. Burpee, and R. H. Burris. 1984. Catabolism of carbohydrates and organic acids and expression of nitrogenase by azospirilla. J. Bacterid. 159 :80-85. Martinez-Drets, G. E. Fabiano, and A. Cardona. 1985. Carbohydrate catabolism in Azospirillum amazonense Appl. Environ. Microbiol. 50:183-185. Matthews, S. W. S. C. Schank, H. C. Aldrich, and R. L. Smith. 1983. Peroxidase-antiperoxidase labeling of Azospirillum brasilense in field-grown pearl millet. Soil Biol. Biochem. 15:699-703. Nakanishi, I., K. Kimura, S. Kusui, and E. Yamazaki 1974. Complex formation of gel-forming bacterial ( 1^3 ) S-D-glucans ( curdlan-type polysaccharides) with dyes in aqueous solution. Carbohydrate Research 32:47-52. Nelson, L. M. and R. Knowles. 1978. Effect of oxygen and nitrate on nitrogen fixation and denitrif ication by Azospirillum brasilense grown in continuous culture. Can. J. Microbiol. 24:1395-1403. Neuer, G. A. Kronenberg, and H. Bothe. 1985. Denitrification and nitrogen fixation by Azospirillum III. Properties of a wheatAzospirillum association. Arch. Microbiol. 141:364-370. Neyra, C. A., and J. Dobereiner. 1977. Denitrif ication by N -fixing Spirillum lipof erum Can. J. Microbiol. 23:300-305. Neyra, C. A., and P. van Berkum. 1977. Nitrate reduction and nitrogenase activity in Spirillum lipof erum Can. J. Microbiol. 23:306-310. Nickels, J. S., J. D. King, and D. C. White. 1979. Poly-6hydroxybutyrate accumulation as a measure of unbalanced growth of the estuarine detrital microbiota. Appl. Environ. Microbiol. 37:459-465.

PAGE 193

188 Nur, I., Y. Okon, and Y. Henis. 1980. Comparative studies of nitrogen-fixing bacteria associated with grasses in Israel with Azospirillum brasilense Can. J. Microbiol. 26:714-718. Nur, I., Y. Okon, and Y. Henis. 1982. Effect of dissolved oxygen tension on production of carotenoids, poly-6~ hydroxybutyrate succinate oxidase and superoxide dismutase by Azospirillum brasilense Cd grown in continuous culture. J. Gen. Microbiol. 128:29372943. Okon, Y., S. L. Albrecht, and R. H. Burris. 1976a. Factors affecting growth and nitrogen fixation of Spirillum lipof erum J. Bacterid. 127 :1248-1254. Okon, Y., S. L. Albrecht, and R. H. Burris. 1976b. Carbon and ammonia metabolism of Spirillum lipof erum J. Bacterid. 128 :592-597. Okon, Y., and R. W. F, Hardy. 1983. Developments in basic and applied biological nitrogen fixation, p. 5-54. In^ F. C. Steward and R. G. S. Bidwell (ed.). Plant physiology, a treatise, vol. VIII. Academic Press, Inc Orlando Okon, Y., P. G. Heytler, and R. W. F. Hardy. 1983. fixation by Azospirillum brasilense and its incorporation into host Setaria italica Appl. Environ. Microbiol. 46:694-697. Page, W. J., and H. L. Sadoff. 1975. Relationship between calcium and uronic acids in the encystment of Azotobacter vinelandii J. Bacterid. 122:145-151. Papen, H. and D. Werner. 1980. Biphasic nitrogenase activity in Azospirillum brasilense in long lasting batch cultures. Arch. Microbiol. 128:209-214. Papen, H. and D. Werner. 1982. Organic acid utilization, succinate excretion, encystation and oscillating nitrogenase activity in Azospirillum brasilense under microaerobic conditions. Arch. Microbiol. 132:57-61. Patriquin, D. G. J. Dobereiner, and D. K. Jain. 1983. Sites and processes of association between diazotrophs and grasses. Can. J. Microbiol. 29:900-915. Pope, L. M. and 0. Wyss. 1970. Outer layers of the Azotobacter vinelandii cyst. J. Bacterid. 102: 234-239

PAGE 194

189 Reusch, R. N., and H. L. Sadoff. 1979. 5-n-Alkylresorcinols from encysting Azotobacter vinelandii ; Isolation and characterization. J. Bacteriol. 139: 448-453. Reusch, R. N. and H. L. Sadoff. 1981. Lipid metabolism during encystment of Azotobacter vinelandii J. Bacteriol. 145:889-895. Reusch, R. N: and H. L. Sadoff. 1983. Novel lipid components of the Azotobacter vinelandii cyst membrane. Nature 302:268-270. Robinson, J. B. P. 0. Salonius, and F. E. Chase. 1965. A note on the differential response of Arthrobacter spp. and Pseudomonas spp. to drying in soil. Can. J. Microbiol. 11:746-748. Rodriguez Caceres, E. A. 1982. Improved medium for isolation of Azospirillum spp. Appl. Environ. Microbiol. 44 :990-991. Ruscoe, A. W., E. H. Newcomb, and R. H. Burris. 1978. Pleomorphic forms in strains of Spirillum lipof erum by light and electron microscopy. Abstract C-60, p. 30. In Proceedings of the Steenbock-Kettering International Symposium on Nitrogen Fixation. University of Wisconsin, Madison. Sadasivan, L. 1985. In Proceedings of the Third Workshop on Azospirillum : Physiology, genetics, ecology. Bayreuth, West Germany (in press). Sadasivan, L. and C. A. Neyra. 1985. Flocculation in Azospirillum brasilense and Azospirillum lipof erum : Exopolysaccharides and cyst formation. J. Bacteriol. 163:716-723. Sadoff, H. L. 1975. Encystment and germination in Azotobacter vinelandii Bacteriol. Rev. 39:516-539. Sadoff, H. L. E. Berke, and B. Loperfido. 1971. Physiological studies of encystment in Azotobacter vinelandii. J. Bacteriol. 105:185-189. Scott, D. B., C. A. Scott, and J. Dobereiner. 1979. Nitrogenase activity and nitrate respiration in Azospirillum spp. Arch. Microbiol. 121:141-145.

PAGE 195

190 Senior, P. J., and E. A. Dawes. 1971. Poly-6hydroxybutyrate biosynthesis and the regulation of glucose metabolism in Azotobacter bei jerinckii Biochem. J. 125:55-66. Shively, J. M. 1974. Inclusion bodies of prokaryotes. Ann. Rev. Microbiol. 28:167-187. Slater, M. and M. Schaechter. 1974. Control of cell division in bacteria. Bacterid. Rev. 38:199-221. Smith, M. S., and J. M. Tiedje. 1979. The effects of roots on soil denitrif ication. Soil Sci. Soc. Am. J. 43:951-955. Socolofsky, M. D. and 0. Wyss. 1961. Cysts of Azotobacter J. Bacterid. 81:946-954, Socolofsky, M. D. and 0. Wyss. 1962. Resistance of the Azotobacter cyst. J. Bacterid. 84 :119-124 Stevenson, L. H. and M. D. Socolofsky. 1966. Cyst formation and poly-6-hydroxybutyric acid accumulation in Azotobacter J. Bacterid. 91 :304-310. Stockdale, H. D. W. Ribbons, and E. A. Dawes. 1968. Occurrence of poly-hydroxybutyrate in the Azotobacteriaceae J. Bacterid. 95:1798-1803. Stotzky, G. 1980. Surface interactions between clay minerals and microbes, viruses and soluble organics, and the probable importance of these interactions to the ecology of microbes in soil, p. 231-247. In R. C. W. Berkeley, J. M. Lynch, J. Melling, P. R. Rutter, and B. Vincent (ed.). Microbial adhesion to surfaces. Ellis Horwood Limited Publishers, Chicester Su, C.-J., R. N. Reusch, and H. L. Sadoff. 1981. Isolation and characterization of several unique lipids from Azotobacter vinelandii cysts. J. Bacterid. 147 :80-90. Sudo, S. Z., and M. Dworkin. 1973. Comparative biology of prokaryotic resting cells, p. 153-224. In A. H. Rose and D. W. Tempest (ed.). Advances in microbial physiology, vol. 9. Academic Press, New York. Sutherland, I. W. 1977. Bacterial exopolysaccharides — Their nature and production, p. 27-96. In I. W. Sutherland (ed.). Surface carbohydrates of the prokaryotic cell. Academic Press, New York.

PAGE 196

191 Tal, S., and Y. Okon. 1985. Production of the reserve material poly6-hydroxybutyrate and its function in Azospirillum brasilense Cd. Can. J. Microbiol. 31:608-613. Tarrand, J. J., N. R. Krieg, and J. Dobereiner. 1978. A taxonomic study of the Spirillum lipof erum group, with descriptions of a new genus, Azospirillum gen. nov. and two species, Azospirillum lipof erum (Beijerinck) comb. nov. and Azospirillum brasilense sp. nov. Can. J. Microbiol. 24:967-980. Tchan, Y.-T. 1984. Family II. Azotobacteraceae p. 219229. In N. R. Krieg and J. G. Holt (ed.), Bergey's manual of systematic bacteriology, vol. I. The Williams and Wilkins Co. Baltimore. Thompson, J. P., and V. B. D. Skerman. 1979. Azotobacteraceae : The taxonomy and ecology of the aerobic nitrogen-fixing bacteria. Academic Press, New York. Tien, T. M. H. G. Diem, M. H. Gaskins, and D. H. Hubbell. 1981. Polygalacturonic acid transeliminase production by Azospirillum species. Can. J. Microbiol. 27 :426-431. Tien, T. M. M. H. Gaskins, and D. H. Hubbell. 1979. Plant growth substances produced by Azospirillum brasilense and their effect on the growth of pearl millet ( Pennisetum americanum L. ) Appl. Environ. Microbiol. 37:1016-1024. Tudor, J. J. 1980. Chemical analysis of the outer cyst wall and inclusion material of Bdellovibrio bdellocysts. Curr. Microbiol. 4:251-256. Tudor, J. J., and S. F. Conti. 1977. Characterization of bdellocysts of Bdellovibrio sp. J. Bacterid. 131: 314-322. Tyler, M. E. J. R. Milam, R. L. Smith, S. C. Schank, and D. A. Zuberer. 1979. Isolation of Azospirillum from diverse geographic regions. Can. J. Microbiol. 25:693-697. Umali-Garcia M. D. H. Hubbell, M. H. Gaskins, and F. B. Dazzo. 1980. Association of Azospirillum with grass roots. Appl. Environ. Microbiol. 39:219-226. Vela, G. R. 1974. Survival of Azotobacter in dry soil. Appl. Microbiol. 28:77-79.

PAGE 197

192 Vela, G. R., and G. Cagle. 1969. Formation of fragile cysts by a strain of Azotobacter chroococcum J, Gen. Microbiol. 57:365-368. Volpon, A. G. T., H. De-Polli, and J. Dobereiner. 1981. Physiology of nitrogen fixation in Azospirillum lipof erum Br 17 (ATCC 29709). Arch. Microbiol. 128: 371-375. Ward, J. B., and R. C. W. Berkeley. 1980. The microbial cell surface and adhesion, p. 47-66. In R. C. W. Berkeley, J. M. Lynch, J. Mailing, P. R. Rutter, and B. Vincent (ed.). Microbial adhesion to surfaces. Ellis Horwood Limited Publishers, Chicester. Webb, S. J. 1965. Bound water in biological integrity, p. 146-171. Charles C. Thomas, Springfield, IL. Whittenbury, R. S. L. Davies, and J. F. Davey. 1970a. Exospores and cysts formed by methane-utilizing bacteria. J. Gen. Microbiol. 61:219-226. Whittenbury, R. K. C. Phillips, and J. F. Wilkinson, 1970b. Enrichment, isolation and some properties of methane-utilizing bacteria. J. Gen. Microbiol. 61:205-218. Winogradsky, S. 1938. Sur la morphologie et I'ecologie des Azotobacter. Ann. Inst. Pasteur 60:351-400. Wong, P. P., N. E. Stenberg, and L. Edgar. 1980. Characterization of a bacterium of the genus Azospirillum from cellulolytic nitrogen-fixing mixed cultures. Can. J. Microbiol. 26:291-296. Wyss, 0., D. D. Smith, L. M. Pope, and K. E. Olson. 1969. Endogenous encystment of Azotobacter vinelandii J. Bacterid. 100 :475-479. Zimmer, W. M. Penteado Stephan, and H. Bothe. 1984. Denitrif ication by Azospirillum brasilense Sp 7 I. Growth with nitrite as respiratory electron acceptor. Arch. Microbiol. 138:206-211.

PAGE 198

BIOGRAPHICAL SKETCH Bruce Henry Bleakley was born in Pontiac, Michigan, in February of 1956. In June 1974 he graduated from New Haven High School in New Haven, Michigan. He entered college in September 1974 at Michigan State University, East Lansing, Michigan. In September 1978 he graduated with a Bachelor of Science degree in agronomy. He entered the Graduate School at Michigan State in January 1979 and received the degree of Master of Science in agronomy in March 1982. He entered the Graduate School of the University of Florida in January 1982 to work on the degree of Doctor of Philosophy in agronomy. 193

PAGE 199

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Wur^y H. G|/skins, Chairman Professor or Agronomy I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Stephan L. Albrecht Assistant Professor of Agronomy I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. G zam Assoji^ate Professor of Microbiology Cell Science

PAGE 200

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. This dissertation was submitted to the Graduate Faculty of the College of Agriculture and the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May 1986 David J. Mitchell Professor of Pathology Professor of Soil Science Dean, Graduate School


14
The third microbial energy-reserve polymer is poly-8-
hydroxybutyrate, a straight chain homopolymer of D(-)-3-
hydroxybutyrate. It is found only in prokaryotic cells,
including both Gram-positive and Gram-negative species. Its
synthesis requires reducing power in the forms of NADH or
NADPH, but does not require the direct expenditure of ATP
(Dawes and Senior, 1973).
With phase contrast microscopy, large accumulations of
PHB within bacterial cells appear as light-refractile gran
ules. A single granule may contain several thousand PHB
molecules (Dawes and Senior, 1973). Each granule is bounded
by a nonunit-membrane layer, which is probably formed from
the cytoplasmic membrane. Presumably the enzymes for
polymerization and depolymerization of PHB are present in
this membrane layer (Shively, 1974).
Many of the Azotobacteraceae accumulate PHB when grown
under dinitrogen-fixing conditions. There can be a wide
variation in PHB content between species and between strains
of the same species (Stockdale et al., 1968). The regula
tion of PHB levels in Azotobacter beijerinckii has been
extensively studied and may provide clues to the role of PHB
in-the physiology of other free-living, dinitrogen-fixing
bacteria, such as azospirilla.
The route of PHB biosynthesis in A. beijerinckii has
been reviewed by Dawes (1981). The synthesis and degrada
tion of PHB in this microorganism are intimately associated


CHAPTER IV
GENERAL CONCLUSIONS
The strains of azospirilla used in this study did not
undergo quantitative morphological encystment when grown on
nitrogen-free BHB agar. The strains of Azospirillum
lipoferum synthesized exopolysaccharides more extensively
than did the strains of Azospirillum brasilense. The
A. lipoferum strains experienced unbalanced growth under
these cultural conditions. They accumulated PHB and experi
enced unbalanced cell wall synthesis, as evidenced by the
common formation of filaments and chains. Eventually the
filaments or chains lost motility and accumulated capsular
material. The final outcome was the formation of microflocs
of encapsulated, PHB-rich cells that often arose from only a
few elongated cells. Some of the cells within these floes
had a cyst-like morphology. Environments with available
nutrients having a high C/N ratio, such as the rhizosphere,
may promote the formation of PHB and capsules by
azospirilla. Some cells having these features may
eventually form cysts. Cells in encapsulated microflocs may
have some survival advantages that individual cells of
azospirilla lack.
178


54
into constitutive dormancy (Marshall, 1980). Cells might
start to encyst when they accumulate threshold levels of
capsular material and/or PHB. In any case, as dormant cells
they would consume little oxygen, allowing it to diffuse to
lower cell layers that previously may have been oxygen-
limited, due to the actively respiring upper cell layer.
These lower cell layers would become more active with the
increased oxygen supply, accumulating capsules. Eventually
these cell layers would also pass into dormancy.
In earlier work, Berg et al. (1979) grew A. brasilense
Sp 7 in association with sugarcane callus tissue. Vegeta
tive cells (V-forms) grew as slimy colonies on the surface
of the callus, and few of these V-forms contained PHB or
capsules. Encapsulated or C-forms were also observed in
these conditions. This association of azospirilla with
sugarcane callus exhibited nitrogenase activity, but whether
the V-forms, C-forms, or both were responsible could not be
ascertained, since both were present. Perhaps C-forms were
able to fix dinitrogen transiently, but were poised to enter
dormancy if growth became too unbalanced. The bacteria did
not possess capsules near or within lysed plant cells, where
the C/N ratio may have been narrow, and balanced growth may
have been promoted.
An apparent contradiction in this plant callus-bacter
ium work is the claim by Berg et al. (1979) that C-forms of
azospirilla have little similarity to Azotobacter


60
contain some rounded, nonmotile cells (Lamm and Neyra, 1981;
Papen and Werner, 1982; Krieg and Dobereiner, 1984).
Azospirilla are morphologically vexing, in that differ
ent pleomorphic cell types occur under various growth-
limiting conditions. Weakly motile filaments containing
little PHB form in aerobic broth which is low in combined
nitrogen (Becking, 1982). Under dinitrogen-fixing condi
tions, filamentous or S-shaped cells again may arise but
contain large deposits of PHB (Tarrand et al., 1978; Wong et
al., 1980; Lamm and Neyra, 1981). These elongated cells
often fragment into smaller, oval cells which subsequently
can assume a cyst-like morphology (Tarrand et al., 1978).
The most frequently reported pleomorphic form of
azospirilla is a nonmotile cell possessing thick outer
layers, probably of capsular material. These cells usually
contain more extensive deposits of PHB than do vegetative
cells grown with combined nitrogen. These cells have been
observed in older cultures grown on combined nitrogen (Lamm
and Neyra, 1981), in cultures grown as dinitrogen-fixers
(Eskew et al., 1977; Berg et al., 1979; Berg et al., 1980;
Papen and Werner, 1982), and in axenic associations with
grass roots (Ruscoe et al., 1978; Umali-Garcia et al., 1980;
Matthews et al., 1983). Recently, Sadasivan and Neyra
(1985) obtained them in broth containing fructose and KNO^.
The nomenclature for describing these cells is not
standardized. Berg et al. (1979) termed them capsulated or


112
c


33
g
inoculations using 8 x 10 CFU/g soil, azospirilla near or
on roots survived better than those in soil distant from
roots. When added to pots of soil containing grass and
6 7
cereal plants, populations remained at 10 to 10 CFU/g
soil, even after 70 days of drought. He speculated that the
presence of roots, either living or dead, enhances the
drought tolerance of the associated azospirilla. In labora
tory studies using nonautoclaved soil microcosms, air drying
of soil was found to kill greater than 99% of the initial
Azospirillum lipoferum inoculum. In comparison, the
indigenous bacteria were little affected by air drying.
This perhaps indicates that, unless azospirilla added to
soil are able to associate quickly with plant roots, they
will soon die out if drought stress occurs.
The desiccation resistance of pleomorphic encapsulated
forms of azospirilla has been studied. Lamm and Neyra
(1981) studied A. brasilense Sp 7 and A. lipoferum Sp 59b,
in addition to several strains of azospirilla isolated from
roots of various grasses in New Jersey and New York. To
obtain cyst-enriched cultures, cells grown in nutrient broth
were harvested by centrifugation, then washed and resus
pended in sterile 0.85% (wt/vol) NaCl. A 1.0 ml sample of
cells was then spread plated as a lawn onto nutrient agar
plates containing 2.0% (wt/vol) agar. Plates were incubated
at 30C until the agar was dried into a thin film, often
requiring a month. After 15 days of incubation, cyst-like


2
growing in soil is often found at the reproductive stage of
the plant (reviewed by Patriquin et al., 1983), after plant
uptake and other processes have reduced the amount of com
bined nitrogen in the root zone (Okon and Hardy, 1983).
Sometimes low amounts of fixed nitrogen have been incor
porated into plant tissue. The transfer of fixed nitrogen
from bacterium to plant seems slow, probably because
bacterial nitrogen is made available for plant uptake only
after the mineralization of the organic nitrogen of dead
bacteria (Okon et al., 1983).
Although the nitrogenase activity of Azospirillum
spp. may not directly provide quick or agriculturally sig
nificant benefits to inoculated plants, the bacteria have
been found to possess other characteristics that may benefit
plants. Axenic associations of grass seedlings and
Azospirillum spp. have resulted in rapid proliferation of
lateral roots and root hairs, probably due to bacterial
production of indole-3-acetic acid and other plant growth
substances (Tien et al., 1979; Umali-Garcia et al., 1980 ;
Jain and Patriquin, 1985). It is also possible that
Azospirillum spp. can enhance production of plant growth
substances by the plants themselves (Kapulnik et al.,
1985). In any case, associations of Azospirillum spp. with
plant roots have led to significant increases of commer
cially valuable plant components in both axenic laboratory
experiments (Kapulnik et al.,
1985) and field inoculations


17
It has been shown that PHB can accumulate in cells that
are not growing or proceeding toward cell division, due to
limitation of available nutrients (Dawes and Senior, 1973).
Nickels et al. (1979) demonstrated this in laboratory micro
cosms containing oak leaf detritus and estuarine water.
Supplementing the nutrients in the water column with carbo
hydrates, especially glucose, induced a rapid accumulation
of PHB without a concomitant increase in microbial biomass.
When supplements were added that enabled increases in
microbial biomass, PHB levels fell as the polymer was broken
down to aid microbial growth.
In one study, A. brasilense Sp 7 was grown in batch
cultures for up to 14 days in microaerophilic, nitrogen-free
malate broth (Papen and Werner, 1980). Both nitrogenase
activity and PHB synthesis were biphasic. An initial peak
of PHB content occurred at day 3, 1 day before the first
peak of nitrogenase activity. During the first and maximal
peak of nitrogenase activity, there was a decrease in PHB
content, possibly due to accumulation of fixed nitrogen
allowing use of PHB carbon skeletons for biosynthesis. A
second peak of PHB accumulation occurred after the first
maximum of nitrogenase activity. The results suggested that
A. brasilense Sp 7, like A. beijerinckii, may accumulate PHB
when it assimilates exogenous carbon faster than it can fix
dinitrogen.


80
nitrogen-free HP-BHB agar. Their phase-bright inclusions
are probably PHB granules; some contain dark bodies,
probably polyphosphate granules. Also shown is a microfloc
of nonmotile, enlarged, PHB-rich cells (Figure 2-3b, c). By
adjusting the distance of the objective lens from the
specimen, many of these cells were observed to be encap
sulated (Figure 2-3b). The encapsulated cells fitted
together closely, as did those observed by Sadasivan and
Neyra (1985). The thickness of the capsule was about
0.5 pm. Such encapsulated microflocs were also observed on
nitrogen-free, LP-BHB agar at this time.
Azospirillum brasilense strain JM 125A2 may have lacked
an efficient mechanism for BHB uptake, compared to the other
strains of azospirilla used. Unlike the other strains, few
motile cells were observed on any BHB agar medium, even in
young cultures. It also differed from the other strains by
having many phase-dark cells that contained little or no
PHB. It eventually grew well on BHB agar when combined
nitrogen was available, however. A final difference between
this strain and the others was that its cells always resus
pended in water to give uniform turbidity, with no macro
scopic floes, or macroflocs, being present. This indicates
that, with or without combined nitrogen, cells of this
strain produce little capsular material when cultured on
BHB.


189
Reusch, R. N., and H. L. Sadoff. 1979. 5-n-Alkyl-
resorcinols from encysting Azotobacter vinelandii:
Isolation and characterization. J. Bacteriol. 139:
448-453.
Reusch, R. N., and H. L. Sadoff. 1981. Lipid metabolism
during encystment of Azotobacter vinelandii.
J. Bacteriol. 145:889-895.
Reusch, R. N;, and H. L. Sadoff. 1983. Novel lipid
components of the Azotobacter vinelandii cyst mem
brane. Nature 302:268-270.
Robinson, J. B., P. 0. Salonius, and F. E. Chase. 1965. A
note on the differential response of Arthrobacter
spp. and Pseudomonas spp. to drying in soil.
Can. J. Microbiol. 11:746-748.
Rodriguez Caceres, E. A. 1982. Improved medium for isola
tion of Azospirillum spp. Appl. Environ. Microbiol.
44 .-990-991.
Ruscoe, A. W. E. H. Newcomb, and R. H. Burris. 1978.
Pleomorphic forms in strains of Spirillum lipoferum by
light and electron microscopy, Abstract C-60, p. 30.
In Proceedings of the Steenbock-Kettering International
Symposium on Nitrogen Fixation. University of
Wisconsin, Madison.
Sadasivan, L. 1985. In Proceedings of the Third Workshop
on Azospirillum: Physiology, genetics, ecology.
Bayreuth, West Germany (in press).
Sadasivan, L., and C. A. Neyra. 1985. Flocculation in
Azospirillum brasilense and Azospirillum lipoferum:
Exopolysaccharides and cyst formation. J. Bacteriol.
163:716-723.
Sadoff, H. L. 1975. Encystment and germination in
Azotobacter vinelandii. Bacteriol. Rev. 39:516-539.
Sadoff, H. L., E. Berke, and B. Loperfido. 1971.
Physiological studies of encystment in Azotobacter
vinelandii. J. Bacteriol. 105:185-189.
Scott, D. B., C. A. Scott, and J. Dobereiner. 1979. Nitro-
genase activity and nitrate respiration in Azospirillum
spp. Arch. Microbiol. 121:141-145.


Without the Hume Library this work could not have been
done. I thank Mr. William Weaver for running a fine
facility.
Finally, completing the list of good coworkers, the eye
and expertise of Louise L. Munro are felt throughout this
study. Her scanning electron microscopy studies cleared up
several uncertainties and raised new questions. She knew
the right stuff when she saw it. Thanks, Louise.
IV


13
basal components, such as proteins and RNA, for energy
sources. Possession of energy-storage polymers can benefit
some species facing starvation, in that they degrade these
polymers instead of or before they are forced to degrade
such essential components as proteins (Dawes and Senior,
1973). However, different microorganisms utilize common
constituents at different rates and in different sequences
when starved. The possession of energy-reserve polymers
does not spare degradation of protein and other basal com
ponents in some species during starvation. Most micro
organisms that remain viable after prolonged starvation have
a low endogenous metabolism, matched closely to their
maintenance energy requirements. Starved microorganisms
that rapidly metabolize polymers generally lose viability
quickly (Dawes, 1976).
Three main types of microbial energy-storage compounds
are known. Some species can accumulate more than one. All
of these compounds have high molecular weights, and only a
slight effect on the internal osmotic pressure of the cell.
The amount of each compound a cell accumulates can vary
widely, depending on environmental conditions.
Intracellular polyphosphates and glycogen-like poly
saccharides are two types of energy-storage compounds formed
by some eukaryotic and prokaryotic microorganisms. The
synthesis of both types requires ATP.


58
3. Is the PHB content of azospirilla grown on BHB
without combined nitrogen higher than when they
are grown in complex broth with combined nitrogen?
4. If pleomorphism of azospirilla occurs in nitrogen-
free BHB broth, is alkalinization of the medium a
prerequisite for development of pleomorphism?
5. Are azospirilla grown on nitrogen-free BHB agar
more desiccation resistant than cells grown in
complex broth with combined nitrogen?
6. Are azospirilla grown on nitrogen-free BHB agar
more resistant to starvation in carbon- and
nitrogen-free, phosphate-buffered salts solution
than cells grown in complex broth with combined
nitrogen?
7. What growth conditions favor motile azospirilla
arising from nonmotile azospirilla grown on
nitrogen-free BHB agar?
8. Is protein synthesis required before nonmotile
BHB-grown azospirilla give rise to motile
azospirilla?
9. Are BHB-grown azospirilla affected by Tris-EDTA in
a manner similar to Azotobacter cysts?
Questions 1 through 4 are considered in Chapter II, and
the remaining questions are considered in Chapter III.


122
clumps. Sometimes floes from HP-TSS broth cultures had no
such empty spaces.
Two-step, broth-replacement studies were conducted to
see if quantitative pleomorphism could be induced in broth
and if there was any connection between pH and pleomor
phism. Sometimes such cultures of A. lipoferum strains
would clump extensively within 24 hours, so that the broth
appeared clear to the naked eye except for the clumps, but
this phenomenon was not consistently reproducible. It
seemed that clumping occurred, sooner and persisted longer in
two-step, broth-replacement studies with A. lipoferum Sp
RG6xx using LP buffer. When HP buffer was used, clear broth
columns often became turbid eventually.
Figure 2-22a shows cells of A. lipoferum Sp RG6xx from
a 43-day-old culture in LP-BHB broth of pH 8.4. The cells
appear healthier than stationary phase cells in TSS broth.
Filamentation and complete septation of cells occurred in
most cases. The elevated pH of the culture again indicated
the low buffering capacity of the LP buffer.
The results from HP-BHB, two-step, broth-replacement
studies indicated that pleomorphism and encapsulation could
occur at near-neutral pH for the A. lipoferum strains. All
the following studies were conducted in HP-BHB broth.
Figure 2-22b shows cells from a 43-day-old culture of
A. lipoferum Sp RG6xx of pH 7.2. In interpreting this
photo, it should be remembered that the cells were added by


109
a
b


Figure 2-3. Cell types of Azospirillum brasilense
strain JM 125A2, from 79-day-old nitrogen-
free, high phosphate-B_hydroxybutyrate
agar cultures. a) Individual rounded
cells. 1,500X magnification. Bar equals
3.0 pm. b) Microfloc focused so that
capsules are visible around cells on
right side of floe. 1,500X magnification.
Bar equals 3.0 pm. c) Same cells as (b),
but focused so that capsules are no longer
evident. Note empty capsule at bottom of
floe. 1,500X magnification. Bar equals
3.0 pm.


TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS iii
ABSTRACT vi
CHAPTERS
I INTRODUCTION AND EXPERIMENTAL APPROACH 1
Ecology of Azospirilla 1
Physiology of Azospirilla 4
Morphology of Azospirilla 6
Prokaryotic Exopolysaccharides 8
Poly-8-Hydroxybutyrate ( PHB ) 12
Dormant Forms of Prokaryotic Cells 22
Resistance of Bacteria to Drying 27
Azotobacter Cysts 37
Pleomorphism of Azospirilla 45
Experimental Approach 5 6
II PLEOMORPHISM OF AZOSPIRILLA GROWN ON
BETA-HYDROXYBUTYRATE 5 9
Materials and Methods 63
Results 71
Discussion 131
III PHYSIOLOGICAL PROPERTIES OF ENCAPSULATED FLOCS
OF AZOSPIRILLUM LIPOFERUM Sp RG6xx 144
Materials and Methods 145
Results 153
Discussion 167
IVGENERAL CONCLUSIONS 178
BIBLIOGRAPHY 181
BIOGRAPHICAL SKETCH 193
V


8
Prokaryotic Exopolysaccharides
Many genera of both Gram-positive and Gram-negative
bacteria include species that can produce polysaccharide
layers outside their cell wall. Such layers have been
referred to as capsules, exopolysaccharides (Sutherland,
1977) or glycocalyces (Costerton et al., 1981). Depending
on laboratory cultural conditions, these polymers can assume
different forms. Slime layers adhere loosely, if at all, to
the cell and can often be separated from cells by centri
fugation. Capsular layers appear to be tightly bound to the
cell itself, and cannot be easily separated from cells.
Microcapsules are so thin that their presence outside the
cell wall cannot be observed using staining and light micro
scopy, while macrocapsules are of sufficient width to be so
resolved (Ward and Berkeley, 1980).
Although proteins are sometimes present in bacterial
capsules, most capsules are mainly polysaccharide in com
position. The polysaccharides are extensively hydrated, and
up to 99% by weight of the capsule is accounted for as water
(Costerton et al., 1981).
The ATP needed to activate sugar residues for exo
polysaccharide synthesis has been shown to comprise a
significant proportion of total-cellular-ATP demand for some
bacteria. Even when the carbon supply is growth limiting,
some strains of bacteria produce extracellular polysacchar
ides (Jarman and Pace, 1984).


67
Growth Conditions
Inocula of azospirilla were grown in screw-cap tubes
containing 10 ml of autoclaved Bacto Nutrient Broth (Difco)
for 24-48 hours at 28C. One milliliter of inoculum was
aseptically pipetted into 100 ml of HP-TSS broth and
incubated for 20 to 22 hours at 30C at 130 rpm on a rotary
shaker. By this time, the cultures attained readings
of 0.7 to 0.9, as measured with a Bausch and Lomb Spectronic
20 spectrophotometer. The pH of the cultures at harvest
ranged from 7.0 to 7.2. Cultures were pelleted by centri
fugation at 6,960 X g for 15 min at 20C. Cells were washed
twice by resuspension and pelleting in sterile LP-basal
salts solution (pH 7.3). The cells were resuspended in
sterile LP-basal salts solution to give a final OD^q
reading of 1.0 to 1.2.
Cell lawns were obtained by spread plating 0.1 ml of
washed cells onto agar media. Inoculated plates were sealed
with Parafilm and incubated at 28C.
For two-step broth replacement studies, washed cells
were aseptically added as a 10% (vol/vol) inoculum to 250-ml
Erlenmeyer flasks, containing a final volume of 100 ml BHB
broth after cell addition. Duplicate flasks were incubated
in the same manner as TSS flasks. These studies were mod
eled after the two-step replacement method of Lin and Sadoff
( 1968 ) .


85
a
I
b
c


151
Soil Dialysis Incubation
These cultures were prepared by a modification of the
method of Gonzalez-Lopez and Vela (1981). The soil employed
was Arredondo fine sand (Grossarenic Paleudult, loamy,
siliceous, hyperthemic), obtained from the top 20 cm of a
soil profile. The soil was air dried, then the fraction
that would pass through a 0.250 mm sieve was recovered for
further use.
Spectrapor membrane tubing (Spectrum Medical Indus
tries, Inc., Los Angeles, CA) of 40 mm diameter and 6,000-
8,000 molecular weight cutoff was used. Tubing was cut into
approximately 25 cm long pieces and soaked in a solution of
1.0% (wt/vol) NaHCO^ in which was dissolved 0.005% (wt/vol)
EDTA. The pH of this solution was reduced to 7.5 with 1 M
HC1 before making it up to final volume. The pieces of
dialysis tubing were soaked and then boiled in this solution
for at least 10 min to remove substances that might harm
bacterial cells. The tubing was then rinsed five times in
changes of deionized water and resuspended in a sixth wash.
One end of each piece of tubing was then knotted once.
Five grams of sieved soil were added to each piece of
tubing by inserting a glass funnel in the open end of each
piece and pouring the soil in. The soil was washed down and
evenly distributed in each piece of tubing by adding about
10 ml of deionized water to the open end with a squirt
bottle. The open end of each tube was then knotted once,


133
accumulated PHB and capsular material, and sometimes became
rounded and nonmotile. Better uptake and growth on BHB
occurred with A. brasilense strain Cd. This strain was
notable for its formation of numerous weakly motile fila
ments that sometimes underwent septation and fragmentation
into individual cells. The presence of combined nitrogen
still allowed PHB accumulation and filament formation of
A. brasilense strain Cd, as was also true for A. lipoferum
Sp RG6xx. Both A. brasilense strains showed some tendency
to form encapsulated microflocs arising from one or more
filaments that had undergone septation.
The A. brasilense strains and A. lipoferum Sp RG6xx
both developed scarlet coloration when cultured as lawns on
SNF-Congo Red agar. Sadasivan and Neyra (1985) have shown
that azospirilla can produce cellulose as one component of
their exopolysaccharides. Congo Red is known to stain many
polysaccharides, including cellulose (Nakanishi et al.,
1974), and colonies of azospirilla grown as
dinitrogen-fixers on agar surfaces take up the dye more
avidly than other free-living, dinitrogen-fixing prokaryotes
(Rodriguez Caceres, 1982). It is suggested that Congo Red
may profitably be used more often in cultural studies with
azospirilla to examine conditions promoting capsule forma
tion. This study indicated that nitrogen-free lawns of
A. lipoferum Sp RG6xx produced capsules extensively when
cultured on succinate or BHB, whereas both A. brasilense


187
Marshall, K. C. 1980. Bacterial adhesion in natural
environments, p. 187-196. In R. C. W. Berkeley, J. M.
Lynch, J. Melling, P. R. Rutter, and B. Vincent (ed.),
Microbial adhesion to surfaces. Ellis Horwood Limited
Publishers, Chicester.
Martinez-Drets, G., M. Del Gallo, C. Burpee, and R. H.
Burris. 1984. Catabolism of carbohydrates and organic
acids and expression of nitrogenase by azospirilla.
J. Bacteriol. 159:80-85.
Martinez-Drets, G., E. Fabiano, and A. Cardona. 1985.
Carbohydrate catabolism in Azospirillum amazonense.
Appl. Environ. Microbiol. 50:183-185.
Matthews, S. W., S. C. Schank, H. C. Aldrich, and R. L.
Smith. 1983. Peroxidase-antiperoxidase labeling of
Azospirillum brasilense in field-grown pearl millet.
Soil Biol. Biochem. 15:699-703.
Nakanishi, I., K. Kimura, S. Kusui, and E. Yamazaki.
1974. Complex formation of gel-forming bacterial
(1^3)-6-D-glucans (curdlan-type polysaccharides) with
dyes in aqueous solution. Carbohydrate Research
32:47-52.
Nelson, L. M., and R. Knowles. 1978. Effect of oxygen and
nitrate on nitrogen fixation and denitrification by
Azospirillum brasilense grown in continuous culture.
Can. J. Microbiol. 24:1395-1403.
Neuer, G., A. Kronenberg, and H. Bothe. 1985. Denitrifica
tion and nitrogen fixation by Azospirillum III.
Properties of a wheat-Azospirillum association.
Arch. Microbiol. 141:364-370.
Neyra, C. A., and J. Dobereiner. 1977. Denitrification by
N^-fixing Spirillum lipoferum. Can. J. Microbiol.
23:300-305.
Neyra, C. A., and P. van Berkum. 1977. Nitrate reduction
and nitrogenase activity in Spirillum lipoferum.
Can. J. Microbiol. 23:306-310.
Nickels, J. S., J. D. King, and D. C. White. 1979. Poly-B-
hydroxybutyrate accumulation as a measure of unbalanced
growth of the estuarine detrital microbiota. Appl.
Environ. Microbiol. 37:459-465.


119
Poly-B-Hydroxybutyrate Content
The phase-bright inclusion bodies in the cells of
A. lipoferum Sp RG6xx grown on nitrogen-free, LP-BHB agar
were confirmed to be PHB using the method of Law and
Slepecky (Table 2-2). This method is subject to error due
to repeated centrifugations and pipettings, so it is not
certain whether the differences in PHB content between cell
lawns of different ages were real or artifacts. The purpose
was to verify that PHB existed in the encapsulated cells in
greater amounts than in vegetative cells grown in HP-TSS
broth. The assay gave evidence of this.
Two-step Broth Replacement Studies
Strains of A. lipoferum cultured in LP-TSS broth
clumped and flocculated in under 24 hours. Clumping of
these strains was delayed in HP-TSS broth. The pH of these
cultures at 20 to 22 hours ranged from 7.0 to 7.2, having
risen from an initial pH of 6.8. Cells of each A. lipoferum
strain in this study, grown in HP-TSS broth, usually started
to clump and flocculate by about 24 hours after inocula
tion. Figure 2-21 shows a SEM photo of cells from a 14-day-
old, HP-TSS culture of A. lipoferum Sp RG6xx. The pH of the
culture at this time was 9.1. The floe has a similar
arrangement to microflocs of the same strain grown on BHB
agar, with filamentous cells and frequent spaces in the


52
with membranous blebs migrating outward into the capsular
material from central bodies containing PHB granules. She
has also observed mature cysts with central bodies contain
ing PHB and polyphosphate granules, surrounded by distinct
intine and exine layers. Thus, given appropriate cultural
conditions, A. brasilense Sp 7 is able to form apparently
mature cysts, almost identical in appearance to those of
Azotobacter spp. One unusual feature she has reported is
layers of spherical, melanin-like granules outside the
exines of mature A. brasilense cysts; these layers have
never been observed with Azotobacter spp. cysts.
Berg et al. (1980) studied morphological and physio
logical changes of A. brasilense Sp 7 grown under different
conditions. Encapsulated cells (C-forms) were often present
on cell lawns grown on nitrogen-free succinate agar.
Encapsulation was initially heaviest for cells near the lawn
surface. After most cells in the surface layers were con
verted to nonmotile C-forms, the lower cell layers began to
accumulate capsules. Such C-forms were not observed within
60 hours of growth in semisolid, nitrogen-free succinate
agar. They formed rapidly on nitrogen-free agar surfaces.
Most of the culture formed capsules. The appearance of the
encapsulated forms varied and changed with time. Both
capsule formation and PHB accumulation were inhibited by
combined nitrogen. As cultures aged, enlarged vibrioid
C-forms developed. The more mature forms were spheres of


142
of nutrients that may not be readily available to
microbial competitors.
4. The encapsulated microfloc may become a fixed site
where azospirilla are sometimes able to outcompete
motile bacteria for carbon sources. For example,
if such a microfloc became established on a root
surface, it might be able to continually deplete
the carbon supply from that area of root by
assimilating it into PHB, without need for further
cell division to occur immediately. Competing
bacteria without the ability to accumulate PHB
might be limited to growing at sites on the root
where only balanced growth could occur.
5. If a microfloc is faced with starvation for
exogenous carbon sources, some of its members may
serve to feed others. If all septa between cells
in a microfloc are completed, it could be that
these septa are lysed during starvation, so that
the substance of dead cells becomes available to
healthier cells within the microfloc.
6. Most microflocs contained several different cell
morphologies. Septation almost always resulted in
cells of significantly different sizes and
shapes. Thus, cells within a microfloc formed
from a single filament may be destined to attain
different physiological states after septation is


190
Senior, P. J., and E. A. Dawes. 1971. Poly-6-
hydroxybutyrate biosynthesis and the regulation of
glucose metabolism in Azotobacter beijerinckii.
Biochem. J. 125:55-66.
Shively, J. M. 1974. Inclusion bodies of prokaryotes.
Ann. Rev. Microbiol. 28:167-187.
Slater, M., and M. Schaechter. 1974. Control of cell
division in bacteria. Bacteriol. Rev. 38:199-221.
Smith, M. S., and J. M. Tiedje. 1979. The effects of roots
on soil denitrification. Soil Sci. Soc. Am. J.
43:951-955.
Socolofsky, M. D., and 0. Wyss. 1961. Cysts of
Azotobacter. J. Bacteriol. 81:946-954.
Socolofsky, M. D., and O. Wyss. 1962. Resistance of the
Azotobacter cyst. J. Bacteriol. 84:119-124.
Stevenson, L. H., and M. D. Socolofsky. 1966. Cyst forma
tion and poly-B-hydroxybutyric acid accumulation in
Azotobacter. J. Bacteriol. 91:304-310.
Stockdale, H., D. W. Ribbons, and E. A. Dawes. 1968.
Occurrence of poly- -hydroxybutyrate in the
Azotobacteriaceae. J. Bacteriol. 95:1798-1803.
Stotzky, G. 1980. Surface interactions between clay
minerals and microbes, viruses and soluble organics,
and the probable importance of these interactions to
the ecology of microbes in soil, p. 231-247. In
R. C. W. Berkeley, J. M. Lynch, J. Melling,
P. R. Rutter, and B. Vincent (ed.), Microbial adhesion
to surfaces. Ellis Horwood Limited Publishers,
Chicester.
Su, C.-J., R. N. Reusch, and H. L. Sadoff. 1981. Isolation
and characterization of several unique lipids from
Azotobacter vinelandii cysts. J. Bacteriol. 147:80-90.
Sudo, S. Z., and M. Dworkin. 1973. Comparative biology of
prokaryotic resting cells, p. 153-224. In A. H. Rose
and D. W. Tempest (ed.), Advances in microbial
physiology, vol. 9. Academic Press, New York.
Sutherland, I. W. 1977. Bacterial exopolysaccharides--
Their nature and production, p. 27-96. In I. W.
Sutherland (ed.), Surface carbohydrates of the
prokaryotic cell. Academic Press, New York.


46
cell's ability to control the site of cell division is not
lost during filamentation. Sometimes chains of cells occur
instead of filaments. The cells in chains contain septa,
but final cleavage between cells has not yet occurred. It
is possible that contiguous cells having incomplete septa in
such chains may share continuities between their
cytoplasms. In some cases, chains may be held together by
very thin capsular layers common to several cells in the
chain.
Jensen and Woolfoik (1985) found that several strains
of Pseudomonas putida and Pseudomonas fluorescens were
induced to form filaments if oxygen became limiting during
the late logarithmic phase of growth in nutrient broth.
Exhaustion of one or more nutrients was also a probable
elicitor of filamentation. The weakly motile filaments,
unlike the highly motile aerobic rods of the bacteria,
migrated to microaerophilic zones. As respiration of cul
tures declined, the increasing levels of oxygen in the broth
seemed to trigger fragmentation of the filaments into rods.
Cultures containing filaments, or the progeny of fragmented
filaments, retained viability longer than nonfilamentous
cultures.
Morphological changes in Escherichia coli have been
related to specific genes. If cellular DNA is damaged by
ultraviolet irradiation or other influences, several genes
are expressed in the so-called SOS response. Many of the


12
Fresh floes were not degraded by cellulase, indicating that
more than one type of exopolysaccharide was present.
Poly-B-Hydroxybutyrate (PHB)
In a constant and favorable environment where all
nutrients are present in sufficient amounts, bacteria grow
for a time in a steady state, where every component of the
cell culture increases by the same constant factor per unit
time. This is balanced growth, and occurs during the
logarithmic phase of the growth curve (Ingraham et al.,
1983). If one or more nutrients become limiting, balanced
growth is not maintained. When the carbon or energy supply
is in excess, so that one or more other nutrients limit
growth, some microorganisms respond by synthesizing and
accumulating intracellular polymers having an energy-storage
function (Dawes and Senior, 1973).
The cell catabolizes these polymers when the energy
supply from exogenous sources is no longer sufficient to
maintain processes needed for maintenance of cell viabil
ity. These processes may include osmotic regulation, main
tenance of intracellular pH and transmembrane potentials,
and turnover of cellular constituents such as proteins and
nucleic acids. The energy required for these processes is
called the energy of maintenance. Some microorganisms do
not produce special energy-storage polymers. Faced with a
starvation environment, they are forced to utilize their own


26
that of vegetative cells. When dried over silica gel
desiccant under slight vacuum in glass tubes, vegetative
cells of strain W die out rapidly and entirely. From 45% to
80% of bdellocysts initially present are able to survive 6
days of this desiccation treatment (Tudor and Conti, 1977).
Bdellocysts possess a thickened outer layer of modified
peptidoglycan, and contain inclusion bodies of an
amylopectin-like polysaccharide of glucose monomers. These
features are not found in vegetative cells (Tudor, 1980).
Some strains of Azotobacter spp., apparently some
azospirilla, and the bacteria described above are the only
prokaryotes reported to form cysts. Why do not more bac
teria possess resting stages that are morphologically dif
ferentiated into cysts? Perhaps growth media and conditions
used in the laboratory discourage cyst formation
(Whittenbury et al., 1970b). It is also possible that the
ability to form cysts is sometimes labile and may be lost
upon subculture. One Methylobacter chroococcum strain was
able to form multiple-bodied cysts upon initial isolation
from the environment. It ceased to do so when subcultured.
Other Methylobacter spp. have retained the ability to form
single- and multiple-bodied cysts over several years of
subculture (Whittenbury et al., 1970b).
A mature, cyst-like cell of a prokaryote may perhaps
best be characterized as follows. Mature cysts differ
morphologically from vegetative cells in having thickened


185
Jagnow, G. 1982. Growth and survival of Azospirillum
lipoferum in soil and rhizospheres as influenced by
ecological stress conditions, p. 97-104. In
W. Klingmuller (ed.), Azospirillum: Genetics,
physiology, ecology. Experientia Supplementum,
vol. 42. Birkhauser Verlag, Basel, Switzerland.
Jain, D. K., and D. G. Patriquin. 1985. Characterization
of a substance produced by Azospirillum which causes
branching of wheat root hairs. Can. J. Microbiol.
31:206-210.
Jansen van Rensburg, H., and B. W. Strijdom. 1980.
Survival of fast- and slow-growing Rhizobium spp. under
conditions of relatively mild desiccation. Soil Biol.
Biochem. 12:353-356.
Jarman, T. R., and G. W. Pace. 1984. Energy requirements
for microbial exopolysaccharide synthesis. Arch.
Microbiol. 137:231-235.
Jensen, R. H., and C. A. Woolfolk. 1985. Formation of
filaments by Pseudomonas putida. Appl. Environ.
Microbiol. 50:364-372.
Kapulnik, Y., R. Gafny, and Y. Okon. 1985. Effect of
Azospirillum spp. inoculation on root development and
NO, uptake in wheat (Triticum aestivum cv. Miriam) in
hydroponic systems. Can. J. Bot. 63:627-631.
Karnovsky, M. J. 1965. A formaldehyde-glutaraldehyde
fixative of high osmolarity for use in electron
microscopy. J. Cell. Biol. 27; (2, Pt. 2):137-138a.
Keynan, A. 1972. Cryptobiosis: A review of the mechanisms
of the ametabolic state in bacterial spores,
p. 355-362. Iji H. O. Halvorson, R. Hanson, and
L. 0. Campbell (ed.), Spores V. American Society for
Microbiology, Washington, DC.
Kraffczyk, I., G. Trolldenier, and H. Beringer. 1984.
Soluble root exudates of maize: Influence of potassium
supply and rhizosphere microorganisms. Soil. Biol.
Biochem. 16:315-322.
Kramer, M. J., and M. D. Socolofsky. 1970. Effect of
chloramphenicol on the conversion of dormant
Azotobacter cysts into vegetative cells. Texas Rep.
Biol. Med. 1 & 2:43-48.


* '
% M
r#,
vf /
* df
s'* *


145
It was of interest to determine if the cells of
encapsulated microflocs of A. lipoferum Sp RG6xx were more
resistant to desiccation and carbon starvation than were
motile, vegetative cells. It was also of interest to define
the conditions under which germination occurred, defined as
motile cells arising from nonmotile, encapsulated
microflocs. Microflocs were exposed to treatment that
results in the rupture of capsular coats of mature
Azotobacter spp. cysts. Finally, whether chloramphenicol
inhibited production of motile cells from a nonmotile
encapsulated inoculum was studied. All of these assays
represented attempts to determine if floes of A. lipoferum
Sp RG6xx contained significant numbers of physiologically
cyst-like cells.
Materials and Methods
Bacterial Strain
The only strain used in these studies was Azospirillum
lipoferum Sp RG6xx. Its maintenance and subculture were as
described in Chapter II.
Growth Media and Enumeration
Vegetative cells were cultured as previously described
in the modified complete medium of Tyler et al. (1979),
using the HP buffer of Albrecht and Okon (1980). Encapsu
lated lawns of the bacterium were cultured on nitrogen-free,


35
were not able to attain physiological maturity under the
experimental conditions.
Papen and Werner (1982) assessed the desiccation
resistance of cyst-like forms of A. brasilense Sp 7. Cells
from dinitrogen-fixing broth cultures were diluted in
sterile tap water and then impinged onto the surface of
sterile 0.2 pm Millipore membrane filters under vacuum.
Some of the filters were immediately placed on the surface
of nutrient agar plates and incubated at 28C, whereas
others were placed on sterile adsorbent pads in Petri dishes
and dried at 37C until they were placed on nutrient agar
plates. Desiccation-resistant cells were only present after
the first peak of nitrogenase activity, when nonmotile,
encapsulated spheres containing PHB predominated. Cells
before and during the first peak of nitrogenase activity
were motile vibrioids and did not survive the desiccation
treatment. As a second peak of nitrogenase activity arose,
motile, dinitrogen-fixing vibrioids emerged from the
spherical capsules; these vegetative cells were again not
desiccation resistant. More encapsulated, spherical cells
survived 2 days of desiccation than 6 days, but it was not
an order of magnitude difference. This again may be an
indication that morphologically mature cysts are not
necessarily physiologically mature.
The recent work of Sadasivan and Neyra (1985) employed
another assay for desiccation resistance of cyst-like forms


120
Table 2-2. Poly-8-hydroxybutyrate (PHB) content of
Azospirillum lipoferum Sp RG6xx.
Cultural
conditions
Age
Dry
weight
(mg ml 1)a
Volume
of CHC13
extract"3
used (ml)
,-1
ug i
PHB
% dry
weight
as PHB
1C
11
days
0.17 (0.02)
0.5d
3.39 (0.52)
39.9
1
21
days
0.18 (0.05)
0.5e
4.56 (0.33)
50.7
1
22
days
0.12 (0.04)
0.5f
3.45 (0.03)
57.5
2?
22
hours
0.49 (0.03)
3.0e
5.28 (0.08)
3.6
aValues are averages of five replicates. Values in parentheses
are standard deviations.
j_
Two H^SO^ replicates of each CHCl^ replicate were used. Values
are averages of all J^SO^ replicates. Values in parentheses are
standard deviations.
1 = Nitrogen-free, low phosphate-8-hydroxybutyrate agar.
Four replicate CHCl^ extracts were used.
eThree replicate CHCl^ extracts were used.
^Two replicate CHCl^ extracts were used.
g2 = High phosphate-trypticase-succinate-salts broth.


36
of azospirilla. Azospirillum brasilense Sp 7 and
A. lipoferum Sp 59b were studied. Large floes of cells
enclosed in exopolysaccharide were placed on Whatman No. 1
filter paper and air-dried for 30 min. They were then
placed in a closed vial, without desiccant, and incubated at
30C for up to 6 months. Small pieces of dried floes were
transferred periodically to semisolid nitrogen-free malate
medium and incubated at 34C for 2 to 4 days, and growth,
pellicle formation, and nitrogenase activity were observed.
Cells in dried floes remained viable for up to 6 months of
drying.
No vegetative cell controls were dried and tested for
viability in the above study. Although cells remained
viable in dried floes for up to 6 months, it is not known
how many cells survived in a given amount of floe. It is
not known whether the cells themselves were desiccation
resistant, or only physically protected from desiccation by
exopolysaccharides.
Tal and Okon (1985) claimed that PHB-rich cells of
A. brasilense strain Cd were 10 times more desiccation
resistant than cells having little of the polymer. No
details of the test used for desiccation resistance were
given.
Desiccation resistance studies can be difficult to
interpret. Comparing the desiccation resistance of vegeta
tive cells to that of cysts may be less difficult than


88
b


Figure 2-2.
Cell types of Azospirillum brasilense
strain JM 125A2, grown on g-hydroxybuty-
rate (BHB) agar. a) 63-hour-old cells
from low phosphate-BHB agar with combined
nitrogen. 1,500X magnification. Bar
equals 3.0 pm. b) Month-old cells from
nitrogen-free, high phosphate-BHB agar.
1,000X magnification. Bar equals 4.0 pm.


43
such a membrane may contribute greatly to the desiccation
resistance and dormancy of cysts.
The possible contribution of the central body membranes
to stress resistance of cysts was suggested in earlier
studies. The cysts of Azotobacter chroococcum strain 75-1
had a compact, well-defined exine layer, whereas the exine
of A. chroococcum strain NTS was diffuse and fragile (Vela
and Cagle, 1969). The cysts of A. chroococcum strain 75-1
were much more resistant to sonic disruption than cysts of
A. chroococcum strain NTS. Yet cysts of both strains were
comparably resistant to desiccation on membrane filters and
to ultraviolet irradiation. Kramer and Socolofsky (1970)
defined cyst germination of A. vinelandii ATCC 12837 as a
process whereby desiccation resistance is lost; mature cysts
were defined as cell forms surviving 3 days of desiccation
on membrane filters. It was found that 40.0 pg
chloramphenicol/ml inhibited outgrowth of cysts in a
complete medium. Many cysts lost their desiccation
resistance when incubated with chloramphenicol, indicating
that the antibiotic might have chemically changed some
essential cyst component, perhaps the central body's mem
branes. Hitchins and Sadoff (1973) found that, soon after
exposure to BHB, vegetative cells became resistant to
100.0 pg chloramphenicol/ml. The antibiotic had no effect
on morphogenesis or rates of protein synthesis. This is
another indication of rapid membrane alteration of encysting


containing high amounts of PHB contained less protein than
PHB-poor cells.
Recent work by Tal and Okon (1985) has further
delineated the roles PHB may play in the physiology of A.
brasilense strain Cd. Grown in aerobic batch culture with
malate and 2.8 mM NH.C1, the cells accumulated 40% of their
4
dry weight as PHB after 24 hours, toward the end of exponen
tial growth. When the level of NH^Cl was raised to 15.0 mM
the cells accumulated only 5% of their dry weight as PHB
after 24 hours. In both cases, the amount of PHB decreased
in stationary phase.
In chemostat continuous culture, a maximum of 30% cell
dry weight accumulated as PHB when the gas atmosphere was
0.082 mM C>2 (Tal and Okon, 1985 ). With increasing aeration
the PHB content fell to very low levels. When grown in
batch culture as dinitrogen-fixers, the cells accumulated
about 75% of their dry weight as PHB. Maximal PHB content
was obtained in these experiments when the C/N ratio was
about 70. Both the C/N ratio of the medium and the oxygen
concentration were found to regulate PHB synthesis.
The forms of carbon and nitrogen made available to the
cells affected the levels of PHB accumulated (Tal and Okon,
1985). Organic acids, especially pyruvate, were found to
elicit PHB formation more than carbohydrates did. Sodium
nitrate was found to promote PHB formation more than
ammonium chloride did, possibly because nitrate does not


BIOGRAPHICAL SKETCH
Bruce Henry Bleakley was born in Pontiac, Michigan, in
February of 1956. In June 1974 he graduated from New Haven
High School in New Haven, Michigan. He entered college in
September 1974 at Michigan State University, East Lansing,
Michigan. In September 1978 he graduated with a Bachelor of
Science degree in agronomy. He entered the Graduate School
at Michigan State in January 1979 and received the degree of
Master of Science in agronomy in March 1982. He entered the
Graduate School of the University of Florida in January 1982
to work on the degree of Doctor of Philosophy in agronomy.
193


32
Vela (1974) tested desiccation resistance of
Azotobacter vinelandii ATCC 12837 by allowing slow drying of
the agar on which the cells were grown. Vegetative cells
were grown on agar plates of Burk's nitrogen-free medium,
with glucose as the carbon source. Cysts were obtained by
growing the cells on the same agar, except that 0.3%
(vol/vol) n-butanol was employed as sole carbon source.
Dried agar films were then broken with sterile forceps and
placed on the surface of Burk's agar medium containing
glucose. Vegetative cells borne on these agar films
remained viable for nearly 2 years of desiccation, whereas
cysts borne on such films remained viable for 10 years or
longer.
Desiccation tolerance of azospirilla has received some
attention. Lakshmi et al. (1977) recovered azospirilla from
several air-dried soils stored in the laboratory. Recovery
was obtained from one of four sandy soils stored air-dry for
10 years. All of these soils had less than 0.5% organic
matter. Heavier-textured soils with 1.0% or more organic
matter consistently yielded isolates of azospirilla. Some
of these heavier-textured soils had been stored air-dry for
up to 15 years. It was suggested that organic matter aids
the survival of azospirilla in drying soils, and that
desiccation-resistant cells may be formed by these bacteria.
Jagnow (1982) did some work with an Azospirillum
lipoferum strain isolated from maize roots. In field


This dissertation is dedicated to Isabel, Stewart, and
Robert Bleakley, the only three people who can say "Of
course" without my doubting them.


172
encapsulated floes were incubated microaerobically in semi
solid agar, as well as aerobically in shaken broth tubes, to
give the cells different oxygen regimes. There was no
observable difference in response between floes incubated
microaerobically or aerobically in buffered salts solution,
with or without single carbon sources. In each of these
treatments, a few weakly motile, PHB-rich cells were
observed within 18 hours, and they persisted for up to 10
days. Half of the capsular spaces in some floes were empty,
but most floes retained the majority of their cells within
capsules. Unlike mature Azotobacter spp. cysts, encapsu
lated cells of this strain did not become synchronously
motile when exposed to carbon sources (Loperfido and Sadoff,
1973). Phosphate-buffered-salts solutions and buffered-
salts solutions containing sucrose, a nonmetabolizable
carbon source for this species (Tarrand et al., 1978),
produced about as many motile cells as did metabolizable
carbon sources. Like mature Azotobacter spp. cysts, encap
sulated cells of this strain seemed unable to mobilize their
PHB reserves in unamended buffered salts solution to enable
germination and widespread motility (Loperfido and Sadoff,
1973). Either wetting released cells from floes and these
cells became motile, or else the cells became motile and
actively left floes after wetting. Perhaps both events
occurred. In any case, most individual cells were not
motile in these incubations. The semisolid agar flasks


56
usually contained PHB and polyphosphate granules, and often
two or more cells were enclosed by a common capsule.
Patriquin et al. (1983) observed unusual structures on
the surface of wheat roots, 3 weeks of age and older, that
had been axenically incubated with azospirilla in a sand-
vermiculite mix. They appeared as spherical "bags," within
which azospirilla containing PHB granules could be seen to
swim about. These structures were also found between the
epidermis and outer cortex of young wheat roots.
Krieg and Dobereiner (1984) suggest that the capsule of
azospirilla helps protect nitrogenase. They also support
the idea that development of alkaline pH is the cause for
pleomorphism in A. lipoferum and A. brasilense. This seems
an incomplete explanation, implying that pleomorphic cells
are poorly viable, being aberrant forms or laboratory arti
facts. Pleomorphic cells of azospirilla may instead develop
commonly, and perhaps transiently, when growing in natural
environments of high C/N ratio, such as near plant roots.
Unbalanced growth, with increased PHB and capsule formation,
may be the major cause of pleomorphism.
Experimental Approach
The conversion of 90% or more of an Azotobacter
sp. cell suspension to cysts facilitates physiological
studies of cysts. Growing the cells in nitrogen-free media
containing n-butanol or BHB leads to this conversion.


166
many PHB-rich cells still contained within capsules. Unlike
all the other extended incubations, only the microaerobic
treatment contained numerous, motile, phase-dark vegetative
cells.
Tris-EDTA Treatment
When encapsulated floes were incubated in Tris-EDTA
solution, the only observable effect was dissolution of
macroscopic floes within 24 hours. To the naked eye, the
cell suspensions appeared more evenly turbid and nonfloc-
culated than those from any other treatment. When viewed by
microscopy, however, microflocs were still present. They
were of about the same size range as microflocs in other
treatments, and had about as many cells remaining in their
capsules. Few motile cells were observed in this treat
ment. There were not noticeably fewer cells within capsules
than in the other treatments where overall germination did
not occur. The same results were obtained when encapsulated
floes were incubated in Tris alone.
Effect of Chloramphenicol
The addition of 50 yg chloramphenicol/ml to nutrient
broth prevented growth for at least 36 hours, whereas
nutrient broth tubes without the antibiotic became turbid
within 18 hours with the same inoculum level. Even in the
presence of chloramphenicol, however, a few weakly motile,


165
After several weeks of incubation, some morphological
changes were observed for cells in the treatments that were
scored as negative at 29 hours. Floes that were incubated
in aerobic, LP-basal-salts solution for 60-65 days had
somewhat the same appearance as germinated floes, in that
most capsular spaces appeared empty or contained cells of
unhearthy appearance. Less than 5% of the capsules in this
extended incubation contained PHB-rich cells. The cells
apparently had not germinated, since no individual vegeta
tive cells were observed outside of the floes. Cells
apparently had depleted their visible PHB reserves and
starved in place within their capsules. The addition of
carbon sources seemed to reduce the extensive loss of
visible PHB deposits. The number of cells within a floe
that retained visible PHB deposits and an overall viable
appearance after 47 days of incubation ranged from about 5
to 20% in aerobic, carbon-amended incubations. Often these
cells appeared somewhat reduced in size, having contracted
from the capsule boundary while retaining their rounded
appearance. A few individual, motile, PHB-rich cells were
also present in carbon-amended incubations at this time.
The nitrate and ammonium incubations at 47 days consisted of
empty capsules and nonmotile, phase-dark vegetative cells
which had the appearance of stationary phase cells. After
60 to 65 days, the microaerobic incubations were similar in
appearance to the aerobic, carbon-amended incubations, with


148
to each beaker, the cells were resuspended by scraping with
a flamed wire loop, then mixed with a sterile 1.0 ml
pipette. One milliliter from each beaker was used for
ten-fold dilution series in sterile, LP-basal salts solution
for the purpose of spread-plate counts. Of the remaining
volume, 0.1 ml was aseptically pipetted and spread plated
onto agar plates. Four plates were spread for each
dilution.
Carbon Starvation
The same washed LP-TSS cell cultures were used for
these experiments as for the desiccation experiments. Cells
were starved in 50.0-ml, Kimax screw-cap test tubes. The
tubes were autoclaved empty, and 8.0 ml of sterile, LP-basal
salts solution were later added aseptically to each. The pH
of the starvation solution was 7.2. A 2.0-ml volume of each
cell type was then added to each of three tubes. The tubes
were incubated horizontally on a 130 rpm rotary shaker at
30C for 9 days. For enumeration of cells surviving the
starvation treatment, 0.1 ml from each tube was aseptically
spread plated onto each of four SNF-Congo Red plates.
Ten-fold dilution series in sterile LP-TSS salts were also
prepared and enumerated by spread plating.


24
Certain properties are shared by all cyst-like forms of
prokaryotes. They are formed when the growth rate of vege
tative cells declines (a metabolic shift-down), due either
to nutrient depletion or transfer of cells to an environment
where balanced growth can no longer occur. Cells encounter
ing these conditions complete their ongoing synthesis of DNA
and chromosome replication but do not initiate new rounds of
DNA synthesis, since growth has ceased (Sadoff, 1975).
Conditions that will prohibit further growth promote the
formation of dormant cells that can survive stress better
than vegetative cells. These dormant cells often contain
PHB or other energy-reserve polymers, and thickened cell
walls or capsular layers. They have enhanced resistance to
irradiation, sonic vibration, and sometimes elevated
temperatures. Perhaps the most important traits for
survival of dormant cells in natural environments are their
resistance to starvation, low endogenous respiration rates,
and desiccation resistance. Many cells entering constitu
tive dormancy need time to mature before they achieve
maximal resistance to stress. It is important to remember
that resting cells formed in natural environments may differ
qualitatively and/or quantitatively in their resistance
properties from those formed under laboratory conditions
(Sudo and Dworkin, 1973).
Many strains of Gram-negative myxobacteria form dormant
cells, called microcysts, when nutrients become limiting.


139
and A. lipoferum Sp RG6xx when they were cultured with BHB
and combined nitrogen.
The environment near plant roots, where azospirilla are
most often found, can be expected to provide available
nutrients having a high C/N ratio (Griffin et al., 1976;
Beck and Gilmour, 1983; Kraffczyk et al., 1984). Depletion
of nitrate from the root zone by physical factors and plant
uptake (Okon and Hardy, 1983), in combination with deni
trification (Smith and Tiedje, 1979), can be expected to
further elevate the C/N ratio of nutrients available to
azospirilla. These bacteria have recently been found to
grow and fix dinitrogen with straw, which again has a high
C/N ratio (Halsall et al., 1985). Azospirilla associated
with plant material can be expected to possess capsules and
PHB, and sometimes to assume pleomorphic cell shapes.
As mentioned before, the most often-reported, pleo
morphic form of azospirilla has been a rounded, nonmotile
cell possessing a capsule and PHB granules. Individual
vegetative cells can likely assume this form without being a
member of a microfloc or microcolony. Such rounded cells
were sometimes observed within microflocs or after they had
broken free of such floes. But the pleomorphic cell types
most often observed were filaments or chains of cells that
eventually formed the microflocs. Bacterial filaments can
arise when cells are growing very rapidly (Slater and
Schaechter, 1974), or when the growth rate shifts down


Figure 2-25.
Cells of Azospirillum lipoferum Sp RG8c
from nitrogen-free, high phosphate-B-
hydroxybutyrate broth, viewed by
scanning electron microscopy. a) Cells
from 33-day-old culture, pH 7.1. Note
thorough encapsulation of lower cell
layer. 3,000X magnification. Bar
equals 5.0 pm. b) Cells from 9-day-old
culture, pH 7.0. Note holes within the
floe. 1,700X magnification. Bar equals
5.0 pm.


118
Table 2-1. Optical density (ODc- n ) and colony forming units
ml- (CFU ml 1) of encapsulated cells of
Azospirillum lipoferum Sp RG6xx.
Culture age
(days)a
OD560
CFU ml 1
o
r1
X
11
0.27
2.67
(0.12 )b
21
0.29
5.08
(0.79)
22
0.25
6.40
(1.00)
58
0.24
1.94
(0.19)
75
0.25
2.11
(0.32)
80
0.30
6.65
(0.49)
aCells were harvested from nitrogen-free, low phosphate-B
-hydroxybutyrate agar plates.
Values are averages of four spread plates. Values in
parentheses are standard deviations.


131
these cultures at 9 days was 6.9 to 7.1. When viewed by
SEM, cells of A. brasilense strain Cd were usually rods
which had collapsed or shrunken during fixation and dehydra
tion. Cells of A. brasilense strain JM 125A2 usually
appeared as very small ovoids, having the appearance of
starved or stationary phase cells. A small number of cells
of the latter strain were enlarged ovoids and possibly were
encapsulated. Unlike all strains of A. lipoferum tested,
the two A. brasilense strains showed little or no tendency
to clump in HP-BHB, two-step, broth-replacement studies.
One more observation made during HP-BHB, broth-replace
ment studies deserves mention. It was suggested earlier
that some filamentous cells that had undergone septation and
encapsulation might retain cytoplasmic connections between
adjacent cells. Figure 2-26 shows a clump of cells from a
58-day-old, HP-BHB broth culture of A. lipoferum Sp RG6xx.
Several of the cells in the clump appear to be undergoing
plasmolysis, but there appear to be continuities of cyto
plasm between some neighboring, plasmolyzed cells.
Discussion
Only two A. brasilense strains were employed in this
study, and both responded far less uniformly to growth
on BHB than did the A. lipoferum strains. There may have
been poor uptake of BHB by A. brasilense strain JM 125A2.
Cells of this strain that were able to grow on BHB


179
Encapsulated cells of Azospirillum lipoferum Sp RG6xx
grown on nitrogen-free BHB agar were found to have far more
PHB than cells grown in broth with combined nitrogen.
Neither cell type displayed significant desiccation
resistance. When faced with aerobic starvation for exo
genous carbon and nitrogen, encapsulated cells of this
strain died off after 9 days to about 25% of their original
numbers. These survivors may have represented cells that
were maturing into cysts. Vegetative cells grown with
combined nitrogen multiplied several fold over the same
period of starvation. This indicates that cells of this
strain may not need to form cysts in nature to survive
prolonged periods of starvation. Vegetative cells having
reduced metabolic activity may survive such periods.
Combined nitrogen promoted germination of nonmotile,
encapsulated cells of this strain.
Although these studies failed to obtain cysts of
azospirilla in high quantity, they may have provided some
information of practical importance. The starvation studies
suggested that encapsulated, PHB-rich cells of azospirilla
are less active physiologically than motile, vegetative
cells. Further studies might measure nitrogenase activity
and plant growth substance production by azospirilla in
relation to PHB deposition and capsule formation. Such
studies should lead to an understanding of what physiologi
cal form of azospirilla is most beneficial to plant growth.


37
comparing that of vegetative cells of different strains.
Rapid drying on a glass surface should enable most mature
cysts to remain viable, but not most vegetative cells. A
glass drying surface should be less hygroscopic than are
membrane filters or agar films. Rapid drying on a glass
surface is a severe treatment, but it should reveal the
presence of physiologically modified, stress-resistant
cells, such as mature cysts.
Azotobacter Cysts
Discussion of the nature of Azotobacter spp. cysts is
important, because this information served as the basis for
the experiments with azospirilla reported in this study.
Like azospirilla, the Azotobacteraceae are
Gram-negative aerobes, often containing PHB granules. Many
are motile by flagella. They all fix dinitrogen, and some,
including Azotobacter spp., do so either at atmospheric
oxygen levels (unlike azospirilla), or as microaerophiles.
Only one genus, Azotobacter, contains species with strains
that are known to form cysts (Tchan, 1984). The isolation
of Azotobacter spp. from the interior of 2,000-year-old clay
bricks (Abd-El-Malek and Ishac, 1966), and their persistence
in soils that had been air dried from 10 years (Vela, 1974)
to 30 years (Clark, 1967), may be due largely to cyst forma
tion .


79
a
c


Figure 2-22.
Cells of Azospirillum lipoferum Sp RG6xx
from 43-day-old, nitrogen-free, g-
hydroxybutyrate (BHB) broth cultures,
viewed by scanning electron microscopy,
a) Cells from low phosphate-BHB broth,
pH 8.4. 7,000X magnification. Bar
equals 5.0 pm. b) Cells from high
phosphate-BHB broth, pH 7.2. Note
thorough encapsulation of lower cell
layer.. 3,000X magnification. Bar
equals 5.0 pm.


143
completed. Some may be poised to become actively
motile as soon as exogenous nutrients enabling
balanced growth become available. Others may be
primed for continued dinitrogen-fixation, and
still others might enter a truly dormant, cyst
like state. Most microflocs observed contained at
least one rounded cell that may have been
cyst-like. Such a diversity of physiological
states within a microcolony may help ensure the
persistence of the colony at that site, providing
a multiplicity of possible rapid cell responses to
environmental conditions.
Although these studies failed to produce apparent
quantitative, morphological encystment of azospirilla, quan
titative encapsulation and microfloc formation of the
A. lipoferum strains were obtained. Encapsulation seems to
be a prerequisite for encystment in Azotobacter spp. (Eklund
et al., 1966). It may be that encysted forms of azospirilla
could be obtained in quantity by further nutritional
manipulation of the encapsulated cells.


168
cells to survive desiccation (Dudman, 1977; Lowendorf,
1980).
Other studies have found enhanced desiccation resis
tance for some encapsulated forms of azospirilla. Such
forms have been called cysts (Lamm and Neyra, 1981; Papen
and Werner, 1982; Sadasivan and Neyra, 1985). In the first
study, however, cyst-like forms were slowly dried while
vegetative cells were rapidly dried, so a true comparison of
desiccation resistance between forms seems questionable. In
the last study, floes of cells were stored without desiccant
for long time periods, and some cells within pieces of floes
were shown to remain viable for up to six months. But
whether this reflects a true desiccation resistance of
single cells, or only an ability to withstand drought in the
presence of hydrophilic polymers (Jagnow, 1982), is
unclear. Perhaps the best studies were done by Papen and
Werner (1982), where vegetative cells were found to almost
completely die out rapidly on dried membrane filters,
whereas encapsulated, nonmotile cells survived this treat
ment in high numbers.
Azospirilla may undergo extensive morphological
changes, including accumulation of PHB and capsular
material, as only a first step towards becoming mature
cysts. Morphologically cyst-like azospirilla have exhibited
variable responses to desiccation (Lamm and Neyra, 1981;
Papen and Werner, 1982). In the studies reported here, many


57
Azospirilla of cyst-like morphology have been observed under
various cultural conditions, but reports of conversion of
90% or more of cell populations to cyst-like forms are not
found in the literature. Vegetative cells are reported as
being present in high numbers, along with the cyst-like
forms. This has perhaps discouraged studies on the nature
of cyst-like forms of azospirilla.
All strains of A. brasilense and A. lipoferum are able
to grow on BHB as sole carbon source in the presence of
combined nitrogen (Tarrand et al., 1978). However, no
studies have been done to see how azospirilla respond to BHB
in the absence of combined nitrogen. Since such cultural
conditions lead to prolific encystment of some Azotobacter
strains, it was considered worthwhile to determine if
strains of azospirilla might also undergo conversion in high
numbers to cyst-like forms under these growth conditions.
The research reported here addresses the following
questions:
1. Can high numbers of cyst-like forms of azospirilla
be obtained by growth in nitrogen-free BHB broth
or on nitrogen-free BHB agar?
2. What are the morphological differences between
azospirilla grown on BHB with or without combined
nitrogen?


90
The two A. brasilense strains failed to produce morpho
logically uniform populations on BHB-agar. A uniform
response was observed for A. lipoferum Sp RG6xx. Good
growth usually occurred within 18 to 24 hours. Figure 2-8
shows 18-hour-old cells grown on nitrogen-free, HP-BHB
agar. Filaments and chains were present, which were some
times as swiftly motile as vibrioids. On LP-BHB agar at
both phosphate levels, with or without combined nitrogen,
septation of filaments was almost complete between 48 and 72
hours, although new filaments would arise and septate for up
to the fifth day. Figure 2-9a, b shows such completely
septated microflocs on 63-hour-old, nitrogen-free, LP-BHB
agar. The floes are encapsulated, and most seemed to arise
from one filament that underwent complete septation.
Although microflocs were present at this time, the cells
resuspended from agar as uniformly turbid suspensions with
out macroflocs.
Many filaments were also completely septated by 63
hours on HP-BHB agar containing combined nitrogen, but some
filaments still lacked completed septa (Figure 2-9 c).
Nitrogen-free, HP-BHB agar at this time contained few if any
microflocs. As for all other media, very motile vibrioids
and ovoids, as well as filaments of varying motility, were
present at 63 hours.
Encapsulated microflocs sometimes formed on nitrogen-
free, HP-BHB agar (Figure 2-10a), but cells from young or


Figure 2-6.
Cell types of Azospirillum brasilense
strain Cd, from 79-day-old lawns on
nitrogen-free, 3-hydroxybutyrate (BHB)
agar. a) Multicellular packets, a chain
and individual ovoids from low phosphate
BHB agar. 1,500X magnification. Bar
equals 3.0 pm. b) Microfloc from high
phosphate-BHB agar, focused to show
capsules around several cells. Air
bubble is above floe. 1,500X magnifica
tion. Bar equals 3.0 pm.


69
an OD560 = 0.86. Forty plates of the same strain grown on
nitrogen-free-LP-BHB agar were harvested by scraping
(described above), to give a final cell volume of about 100
ml. These cells were centrifuged and washed in sterile
deionized water (described above) and resuspended to give an
OD of 0.25 to 0.28.
560
For dry weight determinations, 10.0 ml of the final
cell suspension were pipetted into previously weighed and
desiccated aluminum pans. Five replicate pans were prepared
for each cell type. The pans containing cells were dried to
constant weight at 100C. Pans were kept in a glass
desiccator over anhydrous CaSO^ (Drierite) after removal
from the oven and before weighing.
For PHB determination, 10.0 ml of washed cells were
added to 15 ml Corex centrifuge tubes (Corning Glass Works,
Corning, NY) and pelleted by centrifugation at 7,080 X g for
20 min at 4C. Three replicate tubes were prepared for each
cell type. The supernatant was poured off, and subsequent
steps were performed by the method of Law and Slepecky
(1961). Digestion of cell pellets was begun with the addi
tion of 10 ml of Clorox bleach (5% (wt/vol) hypochlorite).
Cells were suspended in the bleach with Pasteur pipettes;
then the tubes were capped with glass marbles and incubated
in a 37C water bath. Digestion to constant OD^r. was
monitored with a Spectronic 20 spectrophotometer and was
judged to be complete after 18 hours. The insoluble cell


155
Starvation Resistance
After 9 days of starvation in carbon-and-nitrogen-free,
LP-basal salts solution lacking biotin, the encapsulated
\
cells retained 24 to 27% viability. On the other hand,
vegetative cells multiplied several-fold and retained
viability (Table 3-2). It was interesting that, although
two different initial densities of vegetative inocula were
used, each seemed to stabilize at about 10 CFU/ml after 9
days of starvation.
Germination Experiments
Table 3-3 gives a summary of the germination
experiments involving semisolid agar, buffered salts
solution containing single carbon or nitrogen sources, and
soil dialysis flasks. The soil dialysis flasks and
combined-nitrogen incubations all resulted in germination of
encapsulated cells. Within 18 hours, the majority of cells
in these treatments were extremely motile, vegetative
cells. By this time the motile cells in soil dialysis
flasks had lost most or all visible deposits of phase-bright
PHB granules. The PHB granules were usually still visible
in the nitrate and ammonium treatments after 24 hours, but
were markedly reduced in size from those in the initial
inoculum. Within 72 hours the extremely motile cells in the
combined-nitrogen treatments assumed the totally phase-dark
appearance of motile cells in soil dialysate. Figure 3-1


Figure 3-3.
Encapsulated cells of Azospirillum
lipoferum Sp RG6xx that have not undergone
widespread germination in low phosphate-
basal salts solution with glucose,
a) Cells froma 29-hour incubation. A few
empty capsules are present, but most
capsules in the microfloc still retain
their cells. 1,000X magnification.
Bar equals 4.0 pm. b) Microfloc
from a 29-hour incubation with several
empty capsules. Note rounded cyst-like
appearance of some cells remaining in
the floe. 1,000X magnification. Bar
equals 4.0 pm.


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
David J. Mitchell
Professor of Pathology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
David H. Hubpell
Professor of Soil Science
This dissertation was submitted to the Graduate Faculty of
the College of Agriculture and the Graduate School and was
accepted as partial fulfillment of the requirements for the
degree of Doctor of Philosophy.
May 1986
Dean,
aJt 7. 3.
ege of Agriculture
a
Dean, Graduate School


Figure 2-13.
Microflocs of Azospirillum lipoferum Sp
RG6xx, from 36-day-old nitrogen-free,
low phosphate-3-hydroxybutyrate agar,
a) Microflocs focused to show capsules.
1,000X magnification. Bar equals 4.0
pm. b) Same floes as (a) but focused to
show both capsules and filamentous cell
outlines. 1,000X magnification. Bar
equals 4.0 pm.


63
1975). Although all strains of A. brasilense and A. lipo-
ferum are known to grow on BHB as sole carbon source when
provided with combined nitrogen (Tarrand et al., 1978),
there are no reports of the response of azospirilla to BHB
in the absence of combined nitrogen. It was thought worth
while to see if vegetative azospirilla would respond simi
larly to Azotobacter spp., by undergoing quantitative
encystment in the presence of these carbon sources. In pre
liminary studies, apparent extensive PHB accumulation and
capsule formation were observed in some strains of
azospirilla grown with n-butanol. Since n-butanol is
volatile, BHB was used for later studies.
Initial objectives of this study were to achieve
morphological encystment of high numbers of azospirilla, to
document the morphology of such cells, to verify that they
contained PHB, and to ascertain if alkalinization of the
medium was a prerequisite for their formation.
Materials and Methods
Bacterial Strains
The Azospirillum brasilense strains used in these
studies were A. brasilense strain JM 125A2 and A. brasilense
strain Cd (ATCC 29729) (both courtesy of J. Milam, Univ. of
Florida, Gainesville). The Azospirillum lipoferum strains
used were A. lipoferum Sp RG6xx (ATCC 29731), A. lipoferum
Sp RG20a (ATCC 29708), A. lipoferum Sp RG8C, and


127
b


29
These experiments show clearly that rate of drying can
profoundly affect the survival of vegetative bacteria.
Vegetative cells with the capacity to become desiccation
resistant may need time to alter their membrane or cyto
plasmic composition before desiccation resistance is
achieved. Fast-drying procedures may not allow them to do
so. A differentiated resting cell, such as a cyst, may also
need time, depending on how mature it is, to become desicca
tion resistant.
Relative humidity (RH) also has a great influence on
desiccation resistance of prokaryotes. In desiccation at
any RH below 90%, the free water of the cells is removed
almost instantaneously. The water that remains is the bound
water content of the cell, which may be necessary for
continued function of essential metabolic processes and
viability. Often few vegetative bacteria die when desic
cated above 70% RH, but die rapidly as the RH declines to
45% (Webb, 1965). Many desiccation studies have not defined
the RH at which the cells were dried, making duplication of
results difficult.
Thompson and Skerman (1979) tested the desiccation
resistance of vegetative cells of many strains and genera of
the Azotobacteraceae. One milliliter samples of vegetative
cell cultures were added to sterile porcelain beads,
positioned above silica gel in glass bottles sealed with
Parafilm. These desiccation units were stored at room


Figure 2-19. Microflocs of Azospirillum lipoferum
Sp RG20a, from 69-day-old nitrogen-free,
low phosphate-B-hydroxybutyrate agar,
a) Microflocs focused so capsules and
filamentous cell outline are evident in
upper left part of left floe. 1,500X
magnification. Bar equals 3.0 pm. b)
Microflocs with cells having varied
morphologies. 1,500X magnification.
Bar equals 3.0 pm.


74


134
strains seemed to produce capsules extensively only on
succinate.
All the strains of A. lipoferum used in this study
responded uniformly when cultured as lawns on nitrogen-free,
LP-BHB agar. They appeared to accumulate PHB and grow as
filaments that gradually lost motility. Within 5-7 days,
these filaments had accumulated capsular material and become
septated. By this time the cells sometimes stuck tena
ciously to the agar surface and to glass surfaces as well,
and macroflocs were produced. The response could be said to
be homogeneous in that over 90% of the cells present were
encapsulated microflocs, but the cells themselves were not
morphologically homogeneous. All contained large
accumulations of PHB, but the size and shape of cells
varied.
The most extensive studies were done with A. lipoferum
Sp RG6xx, but its behavior was typical of other A. lipoferum
strains. The appearance of the microflocs did not change
noticeably with time when viewed by SEM or phase-contrast
microscopy. The PHB content of the cells also remained
about the same over time. It would be interesting to see if
these microflocs have lower respiration than comparably
cultured cells of the same age on SNF agar. The fact that
lawns from nitrogen-free, LP-BHB agar consistently gave
CFU counts of the same order of magnitude at different
culture ages for equivalent OD,-60
readings indicates that


147
contained within an autoclaved 80-mm X 10-mm Pyrex storage
dish (Fisher Scientific Company, Fair Lawn, NJ). The
beakers had been washed in 3.7% (vol/vol) HC1 for 24 hours,
then rinsed in several changes of deionized water before
autoclaving.
Encapsulated cells grown on nitrogen-free, LP-BHB agar
were harvested and washed in sterile, LP-basal salts solu
tion as described in Chapter II. Cells of 75 days of age
were resuspended to give a final OD,.^ of 0.25 in one
experiment, and cells of 53 days of age were resuspended to
give a final of 0.3 in another experiment. A 0.1-ml
portion of cell suspension was aseptically added to each of
three, 10-ml beakers housed in a storage dish as described
above. Dilution series for plate counts were also prepared
as for the vegetative cells. The storage dishes containing
cells were placed in a glass desiccator over Drierite at
25C for 24 hours, by which time the cell suspensions had
dried onto the glass surfaces. The desiccator was then
placed in a 30C incubator for 8 days.
Initial cell numbers before drying treatment were
enumerated by spread plating. For enumeration of cells
surviving the desiccation treatment, the beakers were
removed from their storage dishes in a laminar flow hood.
The dried cell films were outlined with an ink marker to
help ensure that they would be resuspended. A volume of 2.0
ml of sterile, LP-basal salts solution was aseptically added


Figure 2-12.
Cells of Azospirillum lipoferum Sp
RG6xx, from 17-day-old lawns on
succinate-nitrogen-free-Congo Red agar,
a) Microfloc with empty capsules, as
well as capsules retaining their cells.
1,500X magnification. Bar equals
3.0 pm. b) Another mass of encapsulated
cells. Note that some capsules seem to
contain inclusion granules but no
cytoplasm. 1,500X magnification. Bar
equals 3.0 pm.


124
a
b


18
Zimmer et al. (1984) found that A. brasilense Sp 7
accumulated PHB when using nitrite as terminal electron
acceptor for anaerobic growth. A maximum of 38% of cell dry
weight was found to be PHB when less than 3.0 mM nitrite was
present. No PHB was accumulated when in excess of 8.0 mM
nitrite was made available, indicating the role of PHB as a
sink for excess reducing power when other electron acceptors
are scarce. It was also found that PHB-rich cells contained
less protein than did PHB-poor cells.
Azospirillum lipoferum strain Br 17 (ATCC 29709) was
found by Volpon et al. (1981) to accumulate nearly 24% of
its dry weight as PHB near the mid-logarithmic phase of
growth as a dinitrogen-fixer. Near the end of logarithmic
growth, PHB synthesis seemed to stop, and the content of PHB
declined to 13% of cell dry weight in stationary phase.
The PHB metabolism of A. brasilense strain Cd (ATCC
29729) has received considerable study. When this strain
was grown in continuous chemostat culture with malate and
ammonium chloride, a maximum PHB content of 12% of the
biomass was observed under microaerophilic conditions and at
intermediate growth rates (Nur et al., 1982). These growth
conditions were said to approximate conditions generally
encountered in the rhizosphere. The production of PHB was
markedly decreased at higher levels of oxygen and higher
growth rates. Once again, it was observed that cells


42
Simultaneously, BHB is being taken up and respired or
incorporated. From the sixth to sixtieth hour, unique
lipids, not found in vegetative cells, begin to be
produced. These include 5-n-alkylresorcinols (ARl) and
their galactoside derivatives (AR2). These lipids possess
hydrophobic alkyl sidechains and hydrophilic phenolic
heads. Also produced are 6-n-alkylpyrones (AP), having a
similar bipolar nature. During this time, membranous
vesicles migrate outward from the central body through the
intine to form the exine layer. Up to 17% of the exine is
composed of ARl and AR2. The central body produces ARl and
AR2 in part from its PHB reserves, and exports them in the
membranous vesicles to the exine region. Radio-labelled BHB
accumulates in the central body and exine, whereas the
intine contains almost none. This indicates that the intine
is composed mainly of capsular material, formed from cell
reserves that were present before encystment is triggered by
BHB. Net RNA synthesis stops by the twelfth hour, and net
protein synthesis continues for up to 36 hours. Lipid
turnover continues beyond 60 hours, but there is no net
lipid synthesis. In a mature cyst, 5.0% of the central body
membranes are phospholipid, with AR and AP composing the
other 95%. Molecular models suggest that AR and AP form a
more rigid membrane structure at physiological temperatures
than do phospholipids. The hydrophobic, viscous nature of


182
Bothe, H., B. Klein, M. P. Stephan, and J. Dobereiner.
1981. Transformations of inorganic nitrogen by
Azospirillum spp. Arch. Microbiol. 130:96-100.
Cagle, G. D., and G. R. Vela. 1974. Giant cysts and cysts
with multiple central bodies in Azotobacter
vinelandii. J. Bacteriol. 107:315-319.
Cairns-Smith, A. G. 1982. Genetic takeover and the mineral
origins of life, p. 107-109. Cambridge University
Press, Cambridge.
Clark, F. E. 1967. Bacteria in soil, p. 15-49. In
A. Burgess and F. Raw (ed.), Soil biology. Academic
Press, New York.
Costerton, J. W., J. M. Ingram, and K.-J. Cheng. 1974.
Structure and function of the cell envelope of Gram
negative bacteria. Bacteriol. Rev. 38:87-110.
Costerton, J. W., R. T. Irvin, and K.-J. Cheng. 1981. The
bacterial glycocalyx in nature and disease. Ann. Rev.
Microbiol. 35:299-324.
Crist, D. K., R. E. Wyza, K. M. Mills, W. D. Bauer, and
W. R. Evans. 1984. Preservation of Rhizobium
viability and symbiotic infectivity by suspension in
water. Appl. Environ. Microbiol. 47:895-900.
Das, A., and A. K. Mishra. 1984. Aerotolerant growth in
Azospirillum brasilense induced by dihydroxyphenyl
iron-binding compound. Curr. Microbiol. 11:313-136.
Dawes, E. A. 1976. Endogenous metabolism and the survival
of starved prokaryotes, p. 19-53. In T. R. G. Gray and
J. R. Postgate (ed.), The survival of vegetative
microbes. Twenty-sixth Symposium of the Society for
General Microbiology. Cambridge University Press,
Cambridge, London.
Dawes, E. A. 1981. Carbon metabolism, p. 1-38. I_n P. H.
Calcott (Ed.), Continuous cultures of cells, vol. II.
CRC Press, Inc., Boca Raton, FL.
Dawes, E. A., and P. J. Senior. 1973. The role and regula
tion of energy reserve polymers in microorganisms,
P- 135-266. In A. H. Rose and D. W. Tempest (ed.),
Advances in microbial physiology, vol. 10. Academic
Press, New York.


157
Table 3-3. Response of encapsulated cells of Azospirillum
lipoferum Sp RG6xx to various incubations.
Treatment
Germination3
Soil dialysate
+
N3~
+
nh4+
+
Glucose
-
Fructose
-
Sucrose
-
Malate
-
Succinate
-
Aerobic, low phosphate, basal salts
solution
-
Microaerobic, low phosphate, basal
salts solution
-
+ denotes the majority of cells present becoming motile
and depleting their visible poly-8-hydroxybutyrate reserves
within 29 hours.


160
shows cells from 29-hour incubations in nitrate (Figure
3-la) and ammonium (Figure 3-lb), and Figure 3-2 shows cells
from a 29-hour incubation in soil dialysate. As is apparent
in these photographs, most encapsulated floes had
germinated, retaining their general shape after most of
their cells had left them. The pH of inoculated soil
dialysis flasks rose to 6.4 to 6.5 within 24 hours after
inoculation.
Encapsulated floes suspended in buffered-salts solution
produced a few motile, ovoid to peanut-shaped cells that
remained swollen with intracellular PHB deposits. This was
also true in the semisolid agar flasks and in the buffered-
salts solution tubes containing single carbon sources.
Sometimes these PHB-rich cells were as rapidly motile as the
cells that germinated with combined nitrogen and soil
dialysate. Usually, however, they moved slowly and were
prone to long periods of twiddling before they actively
moved off on a run. Although these treatments often
resulted in large numbers of free, nonencapsulated cells,
most of these individual cells were not motile. The origi
nal inocula did not contain even weakly motile cells and
contained few individual cells. Figure 3-3 shows cells from
a 29-hour incubation in buffered-salts solution containing
glucose, an incubation where the floes remained largely
occupied with cells.


164
b


22
Dormant Forms of Prokaryotic Cells
There is general agreement that most soil bacteria
spend much of their existence in soil in a state of low
metabolic activity. The low respiratory rates of bulk
samples of nonamended soil support this (Clark, 1967). Many
soil bacteria may be metabolically dormant due to a lack of
readily available carbon and energy supplies (Gray and
Williams, 1971). Soil bacteria may enter into exogenous
dormancy, where growth is delayed by unfavorable physical or
chemical conditions (Marshall, 1980). Such bacteria
probably have the same morphology as actively growing
vegetative cells (Gray and Williams, 1971). These cells are
probably intimately associated with the clay or organic
matter of soil. The cells adsorb to these surfaces by
physical or chemical interactions, or by the use of
exopolysaccharides (Stotzky, 1980).
However, many bacteria may exist in soil as dormant
forms that are morphologically different from their growing,
or vegetative, stages. These cells would have entered a
phase of constitutive dormancy, involving the formation of
spores or cysts (Marshall, 1980). Bae et al. (1972) used
transmission electron microscopy to study thin sections of
bacteria released from soil by centrifugation and washing.
About 28% of the bacteria observed had normal vegetative
morphology, of which 29% possessed capsular layers.


70
material was pelleted by centrifugation as above, then
washed once in 10 ml of sterile deionized water, and pel
leted again. The volume for all subsequent washings and
digestions was maintained at 10.0 ml, and all chemicals were
of reagent grade. The OD235 t^ie samPles i-n the final
digestion of concentrated H^SO^ was measured in quartz
cuvettes (1.0 cm light path), using a Carl Zeiss M4QIII
spectrophotometer. For the standard curve, the sodium salt
of DL-6-hydroxybutyric acid (Sigma) was dissolved directly
in concentrated H2SC>4. The standard curve was linear up to
8.0 gg BHB/ml. The PHB content of cell digests was related
back to dry weight values, to determine what percentage of
cell dry weight was present as PHB.
Scanning Electron Microscopy (SEM)
Samples of 0.4 ml from either LP-BHB agar plates or
two-step, broth-replacement cultures were employed for SEM
studies. Cells were removed aseptically from the two-step,
broth-replacement cultures at the same time that culture pH
was measured. Cells were aseptically impinged upon auto
claved 25-mm-diameter, 0.45-gm-pore-size Nuclepore poly
carbonate filters (Nuclepore Corporation, Pleasanton, CA),
housed in a filter chimney attached to a vacuum source.
About 10.0 ml of sterile, deionized water was added to the
chimney after cell addition, to help distribute the cells
evenly over the membrane surface, then a vacuum not


40
allow dinitrogen fixation to resume, so that capsular poly
saccharide was produced, followed by formation of exines
and, ultimately, morphologically mature cysts. Nonencap-
sulated mutants were unable to form morphologically mature
cysts. The work of Pope and Wyss (1970) emphasized that
cells beginning encystment first produced a capsule that
acted as a structure within which the cyst coats were built,
so that the exine existed inside of the capsule. The
diameter of morphologically mature Azotobacter cysts
measured between exine boundaries is about 2.0 pm (Reusch
and Sadoff, 1983).
Abortive encystment occurs when cells round up into
nonmotile precysts, but are unable to form a complete
exine. This occurs in the presence of high amounts of
combined nitrogen (Eklund et al., 1966), when glucose or
other carbon sources are present in addition to n-butanol or
BHB (Lin and Sadoff, 1968), or when calcium is unavailable.
The calcium requirement is probably related to its function
as a stabilizing cation that holds the cyst coats together
(Page and Sadoff, 1975). Using 3.0 mM EDTA in 0.05 M Tris
buffer, pH 7.8, Lin and Sadoff (1969) obtained almost
instantaneous expulsion of the central body from the cyst
coats, due to the chelating effect of the buffer. The empty
exines had the same "horseshoe" shape seen when cysts
germinate, and vegetative cells separate from the exines.


65
per liter) I^HPO^, 6.0 and KI^PO^, 4.0. The LP buffer was
prepared as a 100X concentrated stock solution, and the HP
buffer as a 10X concentrated stock solution. The pH of the
LP buffer was adjusted to 7.1, and that of the HP buffer to
6.7, with 10 M KOH. The buffer stock solutions were
sterilized by autoclaving. All autoclavings in these
studies were for 15 min at standard temperature and pressure
unless otherwise stated.
The TSS components, excluding the biotin and phos
phates, were dissolved and adjusted to pH 7.0 with 10 M
KOH. The broth was then dispensed into 250 ml Erlenmeyer
flasks, in an amount calculated to obtain a final volume of
100 ml after aseptic addition of the biotin and phosphate
buffer stocks to the autoclaved TSS. The initial pH of the
LP-TSS was 6.9 to 7.0, and that of HP-TSS was 6.8.
Plate counts of azospirilla were performed with a
modified succinate-nitrogen free (SNF) agar medium derived
from Tyler et al. (1979) It was the same as LP-TSS, except
that (NH^)2^0^ and Trypticase Peptone were omitted. It
contained in addition (in grams per liter) Bacto yeast
extract (Difco), 0.05 and Bacto agar (Difco), 20.0. It was
prepared in the same way as TSS broth, except that agar was
added after neutralization and before autoclaving. Before
Petri plates were poured, biotin and LP buffer were added
aseptically, as was a solution of autoclaved Congo Red


152
allowing empty space in each tube to help reduce stress due
to expansion of the tubing upon autoclaving. Each piece of
tubing was then rocked gently and agitated while holding one
end in each hand, to make sure the soil had become
thoroughly wetted. Each piece of tubing was then added to
45 ml of deionized water in a 250 ml Erlenmeyer flask and
autoclaved for 25 min. The pH of the sterile, equilibrated
soil solution surrounding the sterile, intact tubing was
6.1. For cell incubations, 0.5 ml of 100X sterile biotin
was aseptically added, followed by 5 ml of water-washed cell
suspension. Triplicate flasks were incubated at 130 rpm at
30C for 24 to 48 hours. The friction between fine soil
particles and dialysis tubing caused bags of soil to break
during longer shaken incubations.
Tris-EDTA Treatment
A solution of 30 mM EDTA dissolved in 0.05 M Tris-HCl
was prepared, and its pH adjusted to 8.4 with 10 M KOH. A
solution of 0.05 M Tris-HCl was also prepared and similarly
adjusted to the same pH. Both solutions were sterilized by
autoclaving. For lysis experiments, 4 ml of either solution
were aseptically added to 4 ml of water-washed, encapsulated
cells in 50-ml screw cap tubes, to give a final concentra
tion of 0.025 M Tris-HCl alone or in combination with 15 mM
EDTA. Tubes were prepared in triplicate and incubated as
for the other tube assays.


16
electron sink for excess reducing power (NADH and NADPH)
that accumulated when the cell became oxygen limited, and
electron transport to oxygen via the terminal oxidases of
the electron-transport chain was restricted (Senior and
Dawes, 1971). Later work revealed that the activities of
certain enzymes of carbon catabolism in A. beijerinckii are
inhibited by either or both NADH and NADPH. Under oxygen
limitation, the concentration of these reduced coenzymes is
increased, so that glucose metabolism, operation of the TCA
cycle, and net biosynthesis are decreased. Growth can
continue at some level, however, if PHB is synthesized and
the crucial coenzymes are reoxidized (Dawes, 1981).
The synthesis of PHB under oxygen limitation may occur
in other bacteria as well (Okon and Hardy, 1983). The
quantity of PHB accumulated often greatly increases as the
C/N ratio of the growth medium increases. Under such con
ditions, free-living dinitrogen-fixers may assimilate the
exogenous carbon more rapidly than they can produce reduced
nitrogen. As a result, the cells can accumulate large
amounts of PHB (Stevenson and Socolofsky, 1966; Dawes and
Senior, 1973). The metabolism of PHB is regulated such that
PHB accumulates when the supply of exogenous carbon is in
excess of the requirements for growth and maintenance, and
it is degraded when the supply of exogenous carbon is
limited or exhausted (Dawes, 1981), or when balanced growth
can again occur (Nickels et al., 1979).


9
For many bacteria, a culture medium having a high
carbon to nitrogen (C/N) ratio promotes capsule formation
(Sutherland, 1977; Costerton et al., 1981). Some species
manufacture exopolysaccharide throughout all phases of
growth, while others produce it only at certain stages of
growth (Sutherland, 1977). Exopolysaccharides of more than
one composition can be formed by the same bacterium under
different environmental conditions (Geesey, 1982).
In laboratory culture, exopolysaccharides may be
nonessential for bacterial growth. Enzymatic removal of
capsules often causes no reduction in viability of the
decapsulated cells (Dudman, 1977). Nonencapsulated mutants
may grow better in laboratory culture than do encapsulated
cells, since they expend no energy for capsular synthesis
(Costerton et al., 1981). Many nonencapsulated laboratory
strains are mutants that have lost the ability of the wild
type to produce exopolysaccharide. In other instances,
common laboratory media have too low a C/N ratio to promote
exopolysaccharide synthesis.
Attachment of bacteria to surfaces by their exopoly
saccharides is the rule in nature, whether the surface is an
inert mineral particle or a biological surface such as a
plant root (Costerton et al., 1981). Natural environments
are far different from laboratory cultural conditions,
containing many more potential hazards to bacterial
survival. In natural environments, the presence of


30
temperature, and at different times single beads were
aseptically removed and placed in broth media. The bacteria
were probably in stationary phase when added to the assem
blies, but it is unlikely that many cysts were present even
in stationary phase broth culture (Sadoff et al., 1971).
The results were surprising; the majority of strains
retained viability for 1 to 2 years of desiccation. This
was true even for bacteria that have never been shown to
form cysts.
Mature cysts of prokaryotes survive rapid desiccation
on glass surfaces far better than do their vegetative coun
terparts, but rarely does all the encysted inoculum survive
rapid drying. Cysts of methane-oxidizers retained 60% to
90% viability after 1 week (Whittenbury et al., 1970a), and
bdellocysts retained 45% to 80% viability after 6 days
(Tudor and Conti, 1977). This may mean that not all the
encysted cells were fully mature when exposed to drying,
even if they all appeared morphologically identical. Such
quick-drying assays can be valuable in determining whether
morphologically differentiated cells are truly cyst-like.
Differences in the desiccation resistance of Azoto-
bacter spp. vegetative cells and cysts are usually
determined by the method of Socolofsky and Wyss (1962).
They impinged suspensions of either cell form on the
surfaces of membrane filters. The filters were then trans
ferred to dry adsorbent pads in Petri dishes and placed in


174
cytoplasm. Sometimes capsules seemed to contain PHB and
polyphosphate granules without any cytoplasm. Empty
capsules, however, often possessed the "horseshoe" shape
typical of empty exines of mature Azotobacter spp. cysts
(Lin and Sadoff, 1969).
The Tris-EDTA incubations did not produce any obvious
expulsion of cells from capsules. The concentration of EDTA
was about five times that which produces prompt expulsion of
central bodies from mature Azotobacter spp. cysts (Lin and
Sadoff, 1969). As mentioned earlier, however, the
macroscopic appearance of the inocula was rendered more
evenly turbid by this treatment. Incubation in Tris buffer
alone had the same effect as Tris-EDTA. The high pH of the
treatments (pH 8.4) may have been related to the dispersive
effect, along with the chelating effects of the Tris and the
EDTA.
It seems that protein synthesis is necessary before
encapsulated cells are able to become motile in large
numbers. The chloramphenicol treatment did not prevent some
free cells from spinning about their own long axes, however.
Based on these tests, there appears to be little
physiological similarity between mature cysts of Azotobacter
spp. and most cells in the encapsulated floes of A.
lipoferum Sp RG6xx. Most of the cells in encapsulated floes
represent immature cysts, lacking desiccation resistance,
but being largely nonmotile and unable to readily mobilize


Figure 2-17.
Microflocs of Azospirillum lipoferum
Sp RG6xx from 178-day-old lawns grown on
nitrogen-free, low phosphate-n-butanol
agar. a) Microfloc showing numerous
cells within capsules as well as empty
capsules. 1,000X magnification. Bar
equals 4.0 pm. b) Microflocs with few
cells remaining within capsules. 1,000X
magnification. Bar equals 4.0 pm.


99
that some cells that were undergoing division in one floe
appeared to have their capsules dividing as well at the site
of septum formation (Figure 2-llb).
Cells from 17-day-old, SNF-Congo Red lawns are shown in
Figure 2-12. Abundant capsules are again evident. These
cells appeared more vibrioid in shape than most A.
brasilense cells cultured in the same manner. Cells of
A. lipoferum Sp RG6xx grown on SNF-Congo Red agar appeared
less swollen and rounded than their counterparts on LP-BHB
Congo Red agar. However, unlike the A. brasilense strains,
lawns of A. lipoferum Sp RG6xx were scarlet on both
SNF-Congo Red and LP-BHB-Congo Red agar.
Nitrogen-free, LP-BHB cultures of A. lipoferum SP RG6xx
did not change in appearance from the seventh day onward,
even after months had passed. Figure 2-13 shows a microfloc
of this strain, with the objective lens adjusted to show the
capsules (Figure 2-13a), and then readjusted to show the
capsules and the apparent continuities between cytoplasms
(Figure 2-13b). Cells from this medium of 7-days-age or
older were consistently resuspended as macroflocs and
microflocs. Individual, septated filaments apparently
consolidated into floes, and individual, motile cells may
have attached to septated filaments to give rise to large
macrofIocs.
More details of floe structure of A. lipoferum Sp
RG6xx were obtained from SEM photographs. Figures 2-14 to


BIBLIOGRAPHY
Abd-El-Malek, Y., and Y. Z. Ishac. 1966. Longevity of
Azotobacter. Plant and Soil 24:325-327.
Albrecht, S. L., and Y. Okon. 1980. Cultures of
Azospirillum, p. 740-479. Iri Methods in enzymology,
vol. 69. Academic Press, Inc., New York.
Bae, H. C., E. H. Cota-Robles, and L. E. Casida, Jr. 1972.
Microflora of soil as viewed by transmission electron
microscopy. Appl. Microbiol. 23:637-648.
Barak, R., I. Nur, Y. Okon, and Y. Henis. 1982. Aerotactic
response of Azospirillum brasilense. J. Bacteriol.
152:643-649.
Beck, S. M., and C. M. Gilmour. 1983. Role of wheat root
exudates in associative nitrogen fixation. Soil
Biol. Biochem. 15:33-38.
Becking, J. H. 1982. Azospirillum lipoferum--A
reappraisal, p. 127-146. In W. Klingmuller (ed.),
Azospirillum: Genetics, physiology, ecology.
Experientia Supplementum, vol. 42. Birkhauser Verlag,
Basel, Switzerland.
Begg, K. J., and W. D. Donachie. 1985. Cell shape and
division in Escherichia coli: Experiments with shape
and division mutants. J. Bacteriol. 163:615-622.
Berg, R. H., V. Vasil, and I. K. Vasil. 1979. The biology
of Azospirillum-sugarcane association II. Ultrastruc
ture. Protoplasma 101:143-163.
Berg, R. H., M. E. Tyler, N. J. Novick, V. Vasil, and
I. K. Vasil. 1980. Biology of AzospiriHum-sugarcane
association: Enhancement of nitrogenase activity.
Appl. Environ. Microbiol. 39:642-649.
Bergersen, F. J. 1984. Oxygen and the physiology of
diazotrophic microorganisms, p. 171-180. In C. Veeger
and W. E. Newton (ed.), Advances in nitrogen fixation
research. Nijhoff/Junk, Pudoc, Wageningen, The
Netherlands.
181