Citation
Characterization of Agrobacterium tumefaciens adsorption to potato tissue

Material Information

Title:
Characterization of Agrobacterium tumefaciens adsorption to potato tissue
Creator:
Kluepfel, Daniel Albert, 1955- ( Dissertant )
Peuppke, Steven G. ( Thesis advisor )
Stall, Robert E. ( Reviewer )
Mitchell, David J. ( Reviewer )
Gurley, William B. ( Reviewer )
Moudgil, Brij M. ( Reviewer )
Fry, Jack L. ( Degree grantor )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1984
Language:
English
Physical Description:
vi, 134 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Adsorption ( jstor )
Agrobacterium ( jstor )
Agrobacterium tumefaciens ( jstor )
Bacteria ( jstor )
Cells ( jstor )
Inoculation ( jstor )
Isotherms ( jstor )
Plant cells ( jstor )
Tubers ( jstor )
Tumors ( jstor )
Agrobacterium tumefaciens ( lcsh )
Dissertations, Academic -- Plant Pathology -- UF
Phytopathogenic bacteria ( lcsh )
Plant Pathology thesis Ph. D
Potatoes -- Diseases and pests ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
The Freudlich adsorption isotherm accurately defines the adsorption of Aqrobacterium tumefaciens to potato tuber discs. This isotherm describes a system where the available binding sites are not saturated. The_A. tumefaciens isotherm was developed for inoculum densities that ranged from 10 ' to 10^ bacteria/ml of buffer. Adsorption is rapid with more than 2.8 X 10^ bacteria adsorbing per potato disc within 0.5 min. Aqrobacterium tumefaciens dissociates from the disc and reaches equilibrium within 3 to 4 hr at 6°C. The adsorption isotherm was examined over a more limited initial inoculum density at three different temperatures: 6°C, 28°C, 38°C. Analysis of these isotherm data with the Clausius-Clapeyron equation shows that the heat of adsorption is close to zero. These combined data suggest that the adsorption of _A. tumefaciens to potato tuber tissue is governed by physical adsorption. Optimum adsorption of A_. tumefaciens to potato tissue occurred at pH 7.2, and adsorption was reduced at either pH 4.5 or 9.5. Adsorption to hydrophobic plastic was insensitive to changes in pH. Conversely, the adsorption to hydrophilic plastic surfaces was optimum at pH 4.5 and decreased with increasing pH. The isoelectric point of A^. tumefaciens suspended in water is approximately pH 2.9, as determined by zeta potential measurements. Adsorption of _A. tumefaciens cells to potato tissue was greatly reduced by detergents at concentrations as low as 0.05% (v/v). Anionic, cationic and neutral detergents all were effective. The effect of agents that influence both hydrophobic and ionic forces is consistent with the proposed physisorption mechanism for A^. tumefaciens. Calcium chloride treatment of the bacteria or EGTA treatment of the potato tuber discs did not, however, significantly affect bacterial adsorption. When viewed with the scanning electron microscope, bacterial colonization of the potato surface was the same for both the virulent A. tumefaciens strain ACH5 and its heat-cured derivative ACH5C3. The only ultrastructural feature that distinguished tissues inoculated with the virulent strain from those inoculated with the avirulent strain was the appearance of tumors five days after inoculation with the former.
Thesis:
Thesis (Ph. D.)--University of Florida, 1984.
Bibliography:
Bibliography: leaves 120-133.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Daniel Albert Kluepfel.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
030404576 ( alephbibnum )
11494285 ( oclc )
ACM1116 ( notis )

Downloads

This item has the following downloads:


Full Text










CHARACTERIZATION OF AGROBACTERIUM TUMEFACIENS
ADSORPTION TO POTATO TISSUE















BY

DANIEL ALBERT KLUEPFEL


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

UNIVERSITY OF FLORIDA

1984














ACKNOWLEDGEMENTS


The work completed here at the University of Florida was aided greatly

by the support and guidance of many people. Heading the list to receive thanks

is my committee chairman Dr. Steven G. Pueppke. His seemingly boundless

patience, accurate guidance, and constant availability resulted in the

cultivation and maturation of my ability to develop a truly scholarly approach

to the art of scientific investigation. I thank you. I am also indebted for the

help given by Dr. H. Aldrich, Dr. G. Erdos, and especially Dr. H. Berg in

distinguishing fact from fiction when peering through the electron

microscope. The time spent by John Ransdell and Steve Linda sharing their

expertise with me is greatly appreciated. I also appreciate the critiques given

my dissertation by the other members of my committee Dr. R. Stall, Dr. D.

Mitchell, Dr. B. Gurley, and Dr. B. Moudgil. Finally, innumerable thanks go to

my wife, Marjan, for her help with the statistical analysis, valuable discussions

of the data, love and understanding. This work would not have been possible

without the financial support of the University of Florida.















TABLE OF CONTENTS


PAGE
ACKNOWLEDGEMENTS .......................................... ii

ABSTRACT ......................................................v

CHAPTER ONE INTRODUCTION ................................. I

CHAPTER TWO LITERATURE REVIEW ...........................3
General Adsorption Chemistry ................................... 3
Adsorption of Bacterial Invadors................................8
Adsorption of Rhizobium spp. to Root Surfaces .............. 10
Adsorption of Agrobocterium tumefaciens to
Plant Surfaces ........................................... 14

CHAPTER THREE ADSORPTION OF AGROBACTERIUM TUMEFACIENS
TO POTATO (SOLANUM TUBEROSUM) TISSUE .....31
Introduction .................................................. 31
Materials and Methods..........................................33
Results ................................................... 35
Discussion ................................. ................... 44

CHAPTER FOUR THERMODYNAMICS OF AGROBACTERIUM
TUMEFACIENS ADSORPTION TO PLANT
SURFACES................................... 49
Introduction ................................................ 49
Materials and Methods ...................................... 50
Results ................................................... 53
Discussion........................ ........................... 61

CHAPTER FIVE EFFECT OF AGENTS THAT ALTER IONIC AND
HYDROPHOBIC INTERACTIONS ON
AGROBACTERIUM TUMEFACIENS ADSORPTION ...67
Introduction ................................................... 67
Materials and Methods ........................................68
Results .............................................. ........76
Discussion....................................................87

CHAPTER SIX SCANNING ELECTRON MICROSCOPY OF
POTATO TUBER TISSUE INOCULATED
WITH AGROBACTERIUM TUMEFACIENS ..........93
Introduction ................................................. 93
Materials and Methods ...... ................................ 94
Results ............................................... .......96
Discussion ..................................... .............. I13

Ill











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



CHARACTERIZATION OF AGROBACTERIUM TUMEFACIENS
ADSORPTION TO POTATO TISSUE

By

Daniel Albert Kluepfel

August 1984


Chairman: Steven G. Pueppke
Major Department: Plant Pathology

The Freudlich adsorption isotherm accurately defines the adsorption of

Agrobacterium tumefaciens to potato tuber discs. This isotherm describes a

system where the available binding sites are not saturated. The A. tumefaciens

isotherm was developed for inoculum densities that ranged from 101 to 109

bacteria/ml of buffer. Adsorption is rapid with more than 2.8 X 104 bacteria

adsorbing per potato disc within 0.5 min. Agrobacterium tumefaciens

dissociates from the disc and reaches equilibrium within 3 to 4 hr at 60C. The

adsorption isotherm was examined over a more limited initial inoculum density

at three different temperatures: 60C, 280C, 380C. Analysis of these isotherm

data with the Clausius-Clapeyron equation shows that the heat of adsorption is

close to zero. These combined data suggest that the adsorption of A.

tumefaciens to potato tuber tissue is governed by physical adsorption.

Optimum adsorption of A. tumefaciens to potato tissue occurred at pH 7.2, and

adsorption was reduced at either pH 4.5 or 9.5. Adsorption to hydrophobic











plastic was insensitive to changes in pH. Conversely, the adsorption to

hydrophilic plastic surfaces was optimum at pH 4.5 and decreased with

increasing pH. The isoelectric point of A. tumefaciens suspended in water is

approximately pH 2.9, as determined by zeta potential measurements.

Adsorption of A. tumefaciens cells to potato tissue was greatly reduced by

detergents at concentrations as low as 0.05% (v/v). Anionic, cationic and

neutral detergents all were effective. The effect of agents that influence both

hydrophobic and ionic forces is consistent with the proposed physisorption

mechanism for A. tumefaciens. Calcium chloride treatment of the bacteria or

EGTA treatment of the potato tuber discs did not, however, significantly affect

bacterial adsorption. When viewed with the scanning electron microscope,

bacterial colonization of the potato surface was the same for both the virulent

A. tumefaciens strain ACH5 and its heat-cured derivative ACHSC3. The only

ultrastructural feature that distinguished tissues inoculated with the virulent

strain from those inoculated with the avirulent strain was the appearance of

tumors five days after inoculation with the former.















CHAPTER ONE
INTRODUCTION


Microorganisms readily accumulate on both natural and man-made

structures. They quickly colonize structures in aqueous environments and the

surfaces of both plant and animal cells. There is evidence that in some

mammalian systems adsorption of bacteria to the host cell surface is an

important factor in pathogenicity (Arbuthnott and Smyth 1979, Berkeley et al.

1980, Bitton and Marshall 1980, Ofek and Beachey 1980). The importance of

adsorption in plant-bacterial interactions also has been receiving considerable

attention (Dazzo 1980a, 1980b, Pueppke 1984a, 1984b).

The surfaces of most plants are colonized by a large number and variety

of bacteria (Blakeman 1982, Foster and Bowen 1982, Suslow 1982). The

presence of these bacteria, however, rarely leads to necrotic plant disease.

Even more unusual is the establishment of a beneficial symbiotic relationship.

However, when plant disease or beneficial symbiotic interactions do occur, the

role of bacterial adsorption is not clear.

Bacteria have the ability to adsorb to plant, animal, and inanimate

surfaces quite tenaciously and, at times, in very large numbers. This

phenomenon is of major ecological significance for bacteria. Microbial

adsorption and colonization of surfaces provide bacteria with a favorable

ecological niche where nutrient levels are high and protection occurs against

unfavorable environmental factors (Bitton and Marshall 1980, Costerton et al.






2



1981). The following literature review will examine some of the features

associated with bacterial adsorption to solid surfaces. Salient examples from

plant-bacterial systems will be discussed in terms of specificity, adsorption

mechanisms, and the relationship between adsorption and the pathogenic or

symbiotic responses.















CHAPTER TWO
LITERATURE REVIEW


General Adsorption Chemistry


A colloid is a suspension that is neither a true solution nor a complete two

phase system. Colloidal particles are small enough to remain widely dispersed

and in constant motion due to bombardment by solvent molecules. Particles in

collodial systems characteristically remain suspended and settle very slowly,

whereas they sediment very rapidly under centrifugal force. Colloids also

cause light to scatter, a useful tool for particle size and concentration

determinations. The stability of a colloid, i.e. the tendency of the particles to

remain in a highly dispersed state, is partially governed by the electrical charge

carried by the particles. Particles within a colloid often contain a large

homogeneous charge. This charge causes the particles to repel one another,

thereby helping to maintain the stability of the colloid (Saunders 1971, Lips and

Jessup 1979).

Bacteria fall within the size range of colloid particles and often exhibit

the properties of a colloidal suspension (Marshall 1976). The adsorption of

bacteria to solid surfaces involves many of the same forces of attraction and

repulsion that operate in the adsorption of inert colloidal particles to solid

surfaces (Daniels 1972, Marshall 1976, Berkeley et al. 1980, Bitton and Marshall

1980). Consequently, bacterial suspensions can be considered to be living

colloidal systems. Because the bacterial surface is highly charged, the use of










physical nonbiologicall) colloidal adsorption systems as models can provide

useful information concerning bacterial adsorption. However, it is important to

remember that bacteria are living organisms capable of movement and altered

activity in response to physiological stimuli.

There are two basic types of adsorption of particles from a liquid phase

onto a solid. These two processes, termed physisorption and chemisorption,

differ in several fundamental respects (Brunauer 1943, Saunders 1971, Tompkins

1978). The most basic difference is the type of molecular force mediating the

adsorption process. Van der Waals forces are primarily responsible for

mediating physisorption. Chemisorption requires chemical reactions, i.e.

formation of covalent bonds. Chemical bonds in chemisorption systems are

much stranger than the forces that account for physisorption. The heat of

adsorption, a measure of the bond strength, is on the order of 10 kcal in

physisorption and greater than 20 kcal in chemisorption. The basic difference

in the bond type is also thought to influence the specificity of adsorption. At

low temperatures van der Waals mediated adsorption occurs between any two

competent surfaces, but chemisorption depends on the affinity of the two

adsorbing components for each other.

The rate of adsorption also is a useful indication of the adsorption type.

The rate of physisorption is extremely rapid, whereas chemisorption requires

energy of activation for bond formation and, in general, is much slower. In

some nonbiological systems, however, the heat of activation can be extremely

small, and thus chemisorption can occur relatively rapidly. Physisorption onto

a solid surface may be multilayered or monolayered, but chemisorption always

occurs in a monolayer. The number of particles adsorbed in a physisorption

system is a function of the particle concentration. In chemisorption the










adsorption is a function of both the area of the solid surface and particle

concentration. A final contrasting feature is that of desorption. The

desorption of particles occurs readily in physisorption, but chemisorbed

particles are only removed with some difficulty.

A useful method of adsorption data analysis is the development of an

adsorption isotherm. Adsorption isotherms are plots of the number of particles

bound versus the number of particles remaining in the suspension at

equilibrium. There are four or five basic types of adsorption isotherms

(Brunauer 1943, Tompkins 1978). Although adsorption isotherms do not provide

data on adsorption mechanisms, other analyses can be used to differentiate

chemisorption from physisorption. Analysis of adsorption at various

temperatures within a 200C to 300C range may suggest a possible mechanism.

Because van der Waals forces are not drastically affected over a 30 degree

temperature range spanning physiological temperatures, physisorption is largely

independent of temperature. On the other hand, because chemisorption

requires heat of activation, chemisorption will vary significantly as a function

of temperature. Heats of adsorption also may be measured directly or with the

use of the thermodynamic law expressed as the Clausius-Clapeyron equation

(Brunauer 1943). These measurements in conjunction with the adsorption

isotherm analysis provide a good description of the possible mechanism for the

adsorption system under examination.



Adsorption of Bacteria onto Solid Surfaces



The adsorption of bacteria onto the solid surfaces of plant cells can be

divided into three temporal stages (Fletcher et al. 1980). The first stage,










simply termed adsorption, refers to events encompassing the initial approach

and contact of the bacterium with the plant cell surface. The second step is

one of anchoring the bacterium to a surface in a firm manner. This may be

accomplished by the production of cellulose microfibrils or other

polysaccharide matrices collectively known as the glycocalyx (Costerton et al.

1981). The third step, termed colonization, is a secondary phenomenon that

describes the long term adsorbed state of the microorganism on the host

surface. This includes continued bacterial growth and multiplication on the

solid surface.

When adsorption of the microbe to living surfaces is measured,

consideration should be given to the time frame under examination. For

example, the adsorption characteristics after 60 min may be quite different

than those measured after 2 to 4 hr, at which time microbial multiplication or

metabolism of extracellular polysaccharides may greatly complicate the

analysis. Therefore, if determination of the mechanism of initial adsorption is

desired, incubation times should be short, i.e. well below the first doubling

time, and the physiological conditions should be held constant.

The outer surface of both Gram-negative and Gram-positive

microorganisms contains a complex assortment of potential attachment

structures. These include outer membrane proteins, lipopolysaccharides (LPS),

lipoteichoic acids, pili, and flagella. An additional structure that is less firmly

associated with the bacterial surface is the capsule. Each of these components

has been implicated in mediating specific attachment in any number of

mammalian or marine adsorption systems (Berkeley et al. 1980, Bitton and

Marshall 1980). It is of interest to note that few of these bacterial cell surface

components have been studied with respect to adsorption to plant surfaces.











Exceptions include reports suggesting the involvement of LPS in both the

attachment of Agrobacterium sp. to wound sites (Whatley et al. 1976) and the

adsorption of Pseudomonas solanacearum to mesophyll cells (Sequeira and

Graham 1977). Capsular polysaccharide-mediated adsorption of rhizobia to

root hairs also has been reported (Dazzo and Hubbell 1975).

Like most colloidal particles, bacteria usually acquire a surface charge

when immersed in an aqueous environment. This is due to the ionization of

surface chemical groups. Due to the presence of carboxyl, amino, or phosphate

groups, the charge of the bacterial surface depends upon its environment. The

surface charge of the bacterium can be measured as a function of pH, providing

a quantitative measure of the behavior of a particle in on electrical field

(Harden and Harris 1953). Once a charge has been established on the surface, it

will attract a layer of oppositely charged counter ions and generate what is

known as the Stern layer (Bohn et al. 1979). Adjacent to the adsorbed Stern

layer is the thick diffuse double layer. This layer contains a higher

concentration of counter ions than does the rest of the aqueous phase.

The ionic characteristics of the bacterial surface, which vary as a

function of ionic strength, have been shown to influence adsorption onto

surfaces (Nissen 1971, Marshall 1976, Bitton and Marshall 1980, Gordon and

Millero 1984, Pueppke and Benny 1984). It is now apparent that many Gram-

positive and Gram-negative bacteria also exhibit some degree of hydrophobicity

on their outer surface (Marshall et al. 1975, Norkrans 1981, Rosenberg et al.

1982). This characteristic appears to be an important factor in bacterial

adsorption to host and inert surfaces alike (Marshall 1976, Gibbon et al. 1983,

Rosenberg et al. 1983). Bacteria have been shown to adsorb to oil-water

interfaces and several different inert plastic-water interfaces (Marshall et al.










1975, Marshall 1976). The contribution of hydrophobic interactions in

mediating bacterial adsorption, however, remains unclear.



Adsorption of Bacterial Invaders


Plant-bacteria interactions can be separated into three groups, based

primarily on the host response. In the first group no visible host response is

elicited by the presence of the bacteria. This is termed a saprophytic

relationship. In the second group, the interaction of a virulent bacterial strain

with a susceptible host results in plant disease. This condition, which is the

exception rather than the rule in nature, is termed a compatible response.

Group three consists of the incompatible reaction, which is defined by the

interaction of a phytopathagenic bacterium with a resistant plant. An avirulent

form of the bacterium interacting with a normally susceptible host also results

in an incompatible reaction.

When plant cells are disturbed by incompatible bacteria, normal host

physiological functions often are altered, and affected cells rapidly collapse

and die. These responses collectively are termed the hypersensitive reaction

(HR). The term hypersensitivity was borrowed from medical terminology,

where it describes an organism that is abnormally sensitive to a pathogenic

agent (Muller 1959). The HR in plants has been known for many years (Ward

1902), and although plants exhibit the HR to a large group of fungi (Kiraly 1980)

and viruses (Holmes 1929), bacterial elicitation of the HR was not documented

until the 1960's (Klement and Lovrekovich 1961, 1962, Klement et al. 1964).

The basic test for the HR involved injecting or vacuum infiltrating a suspension

of viable bacteria (106 to 1010 cells/ml) into the intercellular spaces of a leaf.

Injection is accomplished by placing the beveled end of a hypodermic syringe










containing the bacterial suspension against the leaf surface and applying

pressure (Klement 1963). The intercellular spaces usually dry within 30 min,

and hypersensitive flecks appear within 8 to 24 hr.

Bacterial adsorption onto the plant cell surface is associated with the

initiation of the HR, and bacterial adsorption thus may trigger this plant

response. The HR is only triggered by metabolically active incompatible cells

(Klement and Goodman 1967). The initial step of the HR, termed the induction

period, defines the time required for the irreversible activation of the HR.

Plant cell-bacterial cell contact leading to bacterial adsorption is thought to be

required (Cook and Stall 1977, Klement 1977, Stall and Cook 1979).

The mechanisms by which plants differentiate compatible from

incompatible bacteria and initiate the HR against the latter are matters of

great interest. One attractive hypothesis is that compatible bacteria multiply

because they are allowed to remain free in the intercellular spaces of the leaf

tissues. Incompatible bacteria, on the other hand, are thought to be adsorbed

onto mesophyll cell wall surfaces and then enveloped by wall materials. These

events are considered to immobilize the bacteria and to initiate the HR through

an unknown series of biochemical events that culminate in bacteriostasis

(Kiraly 1980, Sequeira 1980, Klement 1982). Early adsorption and

immobilization events initially were described in tobacco (Nicotiana tabacum)

leaves infiltrated with incompatible Pseudomonos syringae pv. pisi (Goodman

and Plurod 1971, Goodman et al. 1977). During the induction period a portion

of the outer cell wall separates from the mesophyll cells, apparently in

response to adsorbed incompatible bacterial cells. Eventually the separated

cell wall encompasses the bacteria, immobilizing them on the plant cell

surface. A similar response was observed after injection of compatible and










incompatible strains of Pseudomonas solanacearum into tobacco leaves

(Sequeira et al. 1977).

Data from other systems do not support the theory that adsorption and

envelopment of incompatible bacteria are required to initiate the HR. For

example, envelopment of the saprophyte P. putida occurs in bean, but an

incompatible, HR-inducing strain of P. syringae pv. tomato is not enveloped

(Sing and Schroth 1977). In addition, although compatible P. syringe pv.

phaseolicola cells are not enveloped by bean, numerous bacterial cells are

adsorbed onto the plant cell walls (Sing and Schroth 1977). Sequeira et al.

(1977) have reported the adsorption and envelopment of non HR-inducing

organisms such as Escherichia coli, Bacillus subtilis, and heat-killed

incompatible Pseudomonas solanacearum by tobacco. This demonstrates that

adsorption is not sufficient for induction of the HR in these systems.

Though adsorption and envelopment are of questionable significance in

terms of the HR, the above data are consistent with several provisional

conclusions: i) metabolically active incompatible cells are required to trigger

the HR (Klement and Goodman 1967); ii) close contact and possibly bacterial

adsorption onto the plant cell surface are prerequisities for induction of the HR

(Klement 1977, Stall and Cook 1979); iii) because envelopment occurs in some

incompatible interactions and not in others, this postulated defense mechanism

does not appear to be a generalized plant response to incompatible invaders.



Adsorption of Rhizobium spp. to Root Surfaces


The ability of bacteria of the genus Rhizabium to infect the roots of

legumes and produce nitrogen-fixing nodules is well documented (Fred et al.

1932). The symbiosis between legumes and members of the genus Rhizobium










generates nearly half of the 175 million metric tons of dinitrogen fixed each

year by microorganisms (Hardy and Havelka 1975). The legume-Rhizobium

symbiosis is controlled by both the plant (Nutman 1956, Hall and Larue 1976,

Lie et al. 1981) and the bacterium (Ljunggren 1961, Maier and Brill 1976,

Beringer 1980, Hobbs and Mahon 1983). This symbiosis is highly specific in that

only certain Rhizobium spp. are able to nodulate any given legume species. It

is this specificity which provides the basis for taxonomic groupings within the

genus Rhizobium (Fred et al. 1932). Initial control of this specificity in the

legume-Rhizobium sp. interaction has been postulated to occur at four to five

different levels starting with regulation of growth of rhizobia in the

rhizosphere and culminating with infection thread formation (Broughton 1978,

Vincent 1980). Host-symbiont specificity exhibited in the legume-Rhizobium

system is hypothesized to be initiated by the specific attachment of infective

bacteria to the host (for reviews, see Broughton 1978, Dazzo 1980a, Solheim

and Paxton 1981, Pueppke 1984a).

Substantial numbers of rhizabia adsorb, in many orientations, to root

surfaces (Dazzo 1980a, 1980b, Higashi and Abe 1980, Stacey et al. 1980,

Turgeon and Bauer 1982). Bal et al. (1978), however, found that only 5% of

1,798 examined root hairs contained adsorbed bacteria, and individual root hairs

adsorbed only I to 16 bacteria. When adsorption is quantified and analyzed

statistically, rhizobia are found adsorbed to both host and nonhost tissue in

equivalent numbers (Chen and Phillips 1976, Bauer 1982, Law et al. 1982). In

addition, both nodulating and non-nodulating strains of a given Rhizobium sp.

bind to roots of the host legume (Broughton et al. 1980, Solheim 1983, Pueppke

1984b). Adsorption of four slow-growing strains of rhizobia to soybean and

cowpea roots also was shown to be unrelated to the ability of the bacteria to

infect and nodulate (Pueppke 1984b).










A positive correlation between bacterial adsorption and legume infection

has been suggested by other workers using alternate systems. Dazzo et al.

(1976) demonstrated a direct relationship between the ability of R. trifolii to

infect white clover and its ability to adsorb to white clover root hairs. Specific

adsorption of R. trifolii to red clover roots hairs has been correlated with the

ability to infect and the presence of a symbiosis plasmid (Zurkowski 1980).

Strains of R. trifolii containing the plasmid are adsorbed in much greater

numbers than are strains lacking the plasmid. Several strains which have lost

or reduced ability to nodulate soybean were found to have correspondingly

reduced ability to adsorb to plant root tissue (Jansen van Rensburg and Strijdom

1982, Stacey et al. 1982).

The inconsistency of these experimental results may be partially

explained by the widely divergent techniques and systems used to examine

bacteria-root interactions. For example, bacterial adsorption to root epidermal

cells may mask specific adsorption to incipient root hairs, the site of

infection. Even when adsorption to root hairs is monitored directly, care should

be taken to quantify adsorption only to susceptible tissues (Bhuvaneswari, 1981,

Bhuvaneswari et al. 1981).

The frequently observed polar mode of attachment of the bacterium to

the host's root surface was hypothesized to be of significance to infection

(Sahlman and Fahraeus 1963, Dazzo 1980a, Gotz and Hess 1980, Gotz 1980).

This polar attachment, however, appears to be a nonspecific phenomenon; polar

binding of Rhizobium spp. occurs at oil-water and inert plastic-water interfaces

(Marshall et al. 1975, Marshall 1976). The bacteria also attach in a polar

fashion to roots of nonhost species such as wheat and petunia (Shimshick and

Hebert 1978, Hess et al. 1982). Polar binding may simply reflect the fact that











the poles of the bacterium are more hydrophobic, favoring this type of

adsorption. Charge characteristics of both the bacterium and the plant cell

surface also may function in making polar attachment energetically favorable.

Fibrillar extracellular structures have been observed to encompass

microcolonies of Rhizobium spp. on the surface of root cells (Reporter et al.

1975, Napoli et al. 1975). Deinema and Zevenhuizen (1971) provided evidence

that some such structures are cellulose fibrils of bacterial origin. The

participation of cellulose microfibrils in the initial adsorption events is

probably limited, though they may serve to securely anchor the bacterium to

the host in later stages of the interaction (Napoli et al. 1975, Pueppke 1984).

Other substances with suspected adhesive properties also are observed in

electron micrographs of the bacterium-plant interface. Marshall et al. (1975)

illustrated that some type of bridging polymers bound polarly attached bacteria

to the root surface. Others have seen similar structures in electron

micrographs studies (Menzel et al. 1972, Dazzo and Hubbell 1975). These

structures, however, also bind bacteria to inert plastic substrates. Thus, they

are of bacterial origin and are induced nonspecifically (Marshall et al. 1975,

Marshall 1976).

Currently there is no clearcut consensus in the literature on the

relationship between host-specific adsorption and nodulation specificity. This

is understandable given the fact that there exist a large number of potential

control points available to both the bacterium and host from the time of

bacterial root contact to nodule formation. I believe that each potential

regulation site deserves careful examination, which when accomplished, will

most likely suggest a concerted multi-level control with both partners

participating.









Adsorption of Agrobacterium tumefaciens to Plant Surfaces


Agrobacterium tumefaciens is a Gram-negative, soil-borne, rod-shaped

bacterium which causes the plant disease known as crown gall. The crown gall

is a neoplastic undifferentiated growth of the plant resulting from the stable

incorporation and expression of a piece of bacterial DNA (T-DNA) in the

nuclear DNA of the host (Chilton et al. 1977, Gurley et al. 1979, Thomashow et

al. 1980). Roughly 38% of the tested dicot families and 86% of the tested

gymnosperm families, and a few monocotyledonous species, are susceptible (De

Cleene and De Ley 1976). Agrobacterium tumefaciens is thought to adsorb to a

specific chemical moiety in a wound site of the host's cell wall (Lippincott et

al. 1977b, Rao et al. 1982). If specific adsorption does not occur, the bacterium

is thought to be incapable of tumor induction. The specific adsorption

hypothesis, developed by Lippincott and Lippincott (1969), provides the basis

for what is known as the site-attachment hypothesis.

Most of the data in support of the site-attachment hypothesis comes from

tumor formation assays employing sequential inoculations. All inoculations in

these experiments are carried out by dusting pinto bean leaves with

carborundum, applying 0.1 ml of bacterial suspension, and thoroughly rubbing

the surface of the leaf with a glass rod (Lippincott and Heberlein 1965a,

1965b). Like many biological assays, however, the pinto bean leaf assay (PBLA)

suffers from a large degree of variability from experiment to experiment. For

example, the mean number of tumors per leaf after inoculation with 2.8 to 4.7

X 108 cells/ml varied from 4.6 to 71 to 136 in three successive experiments.

This 27-fold difference in tumor number was observed, even though the

inoculum densities fell within the linear range of the dose-response relationship

between tumor numbers and inoculum density (Lippincott and Heberlein










1965b). Changes in plant age, greenhouse conditions, and the amount of

pressure applied with the glass rod during inoculation result in a considerable

difference in the number of tumors formed, and thus may account for some of

the variability. Lippincott and Heberlein (1965b) suggested that an all or none

assay could reduce the amount of variability associated with carborundum

inoculations. Leaves could be wounded with a multiple needle holder,

inoculated, and the wounds in which tumors developed scored.

The relationship between tumor numbers and initial bacterial

concentration in the PBLA is linear over approximately two log units of

bacterial concentration, from approximately 3 X 106 to 8 X 108 bacteria/ml.

The number of tumors remains relatively constant at bacterial concentrations

greater than 8 X 108 cells/ml. In this narrow range of linearity ca. 3 X 106

bacteria are required for induction of a single tumor. Additionally, the PBLA is

insensitive (i.e. no tumor formation) to bacterial concentrations less than ca. 3

X 106 cells/ml (Lippincott and Heberlein 1965b). Theoretical curves for one-

and two-particle events suggest that individual tumors are initiated by a single

bacterium acting at a susceptible site. This would be in agreement with

Hildebrand's (1942) earlier observation that a single bacterium placed on a

competent plant cell can initiate tumor formation. However, this is

inconsistent with the observation that tumors are not formed when inoculum

densities are below ca. 3 X 106 bacteria/ml.

The extrapolation of tumor data to bacterial adsorption poses pitfalls for

those attempting to understand bacterial adsorption. This difficulty has not

escaped the attention of those working with the PBLA. "When more complex

developments are used to measure the interaction such as disease symptom

formation or the induction of host defense systems, many factors in addition to









adherence may function to determine if the specific response sought will

occur. In such cases, adherence may constitute only one small part of the

specificity-recognition system, or, in the extreme and even though essential, an

adherence mechanism could prove so nonspecific that it would ordinarily have

no role in specificity-recognition phenomena." (Lippincott and Lippincott 1980,

pp. 377).

In the PBLA, virulent A. tumefaciens strains form tumors that can be

easily tallied 7 days after inoculation. When pinto bean leaves were inoculated

with the avirulent strain IIBNV6 prior to or in combination with, virulent strain

B6, the number of tumors was reduced relative to that in controls inoculated

with strain B6 alone. Preinoculation and coinoculation with strain IIBNV6

caused similar levels of inhibition. To test the possibility that such inhibition is

due to simple nonspecific physical blockage by strain IIBNV6, other bacteria

also were tested. Preinoculation with E. coli, Bacillus sp., and Rhizobium sp.

did not reduce the number of tumors. Additionally, when the process was

reversed, and the plant was inoculated with B6 prior to IIBNV6, no reduction in

tumor formation was observed. These results eliminated the possibility that

strain IIBNV6 nonspecifically blocked the adsorption of strain B6. In addition to

IIBNV6, several virulent (heat-treated) and avirulent A. tumefaciens strains

inhibit tumor formation by strain B6. Such strains are termed site-binding

strains (Lippincott and Lippincott 1975). The above data provided the basis for

the site-attachment hypothesis which, simply stated, suggests that there exists

a single specific attachment site within a wound to which a single bacterium

can attach and initiate tumor formation. Tumor inhibition, however, results

from the exclusion of a single virulent cell from a specific site by a single

virulent cell. It should be emphasized that the A. tumefaciens site-


-- `IL C' i.









attachment hypothesis is based on tumor formation data and is not directly

supported by bacterial adsorption data from the PBLA.

What appears to be a similar phenomenon, i.e. inhibition of tumor

formation by inactivated cells, was noted 46 years previously (Brown 1923). No

tumors formed on daisy plants puncture-inoculated with heat-killed A.

tumefaciens and later inoculated in the same site with viable, virulent A.

tumefaciens cells.

Enhanced tumor formation by mixtures of virulent and virulent A.

tumefaciens cells has led to some interesting, though contradictory, concepts

dealing with A. tumefaciens attachment (Lippincott and Lippincott 1970,

1977a, 1978b, Lippincott et al. 1978). When pinto bean leaves were inoculated

with mixtures containing virulent and virulent cells of Agrobacterium sp., a 10

to 20-fold increase in tumor numbers was noted. The same enhancement was

not observed if virulent cells were substituted for virulent cells, i. e. if the

inoculum density was increased. In these enhancement complementationn)

experiments, the virulent strain is termed the donor strain, and the virulent

strain is designated the receptor strain (Lippincott and Lippincott 1970). Heat

treatment of either the donor or receptor strain totally abolishes the enhancing

activity. Although independent of the ratio of virulent to virulent cells,

complementation is affected by the absolute numbers of bacteria in the

inoculum. For example, a 2.3:1 ratio (5 X 108 IIBNV6 cells and 2.2 X 108 B6

cells/ml) caused a 33% increase in tumor formation, but another 2.3:1 ratio (I

X 109 IIBNV6 cells and 4.4 X 108 B6 cells/ml) caused a 44% reduction in tumor

formation. A ratio of 68:1 (6.4 X 109 IIBNV6 cells and 9.4 X 107 B6 cells/ml)

resulted in a 63% reduction in tumor formation, yet cells in a similar ratio, 54:1

(4.5 X 109 IIBNV6 cells and 8.4 X 107 B6 cells/mi), resulted in a 40% increase in

tumor formation.


- II










Lippincott and Lippincott (1970) believe that complementation results

from the separate action of a helper strain at one site and an avirulent strain

at a second site. In their view, "the enhancement apparently depends on one or

more substances moving between wounds occupied by virulent bacteria and

wounds where IIBNV6 [receptor strain] cells are attached" (Lippincott et al.

1977a, pp. 828). Although either inhibition or enhancement can occur under

conditions where ratios of virulent and virulent cells are similar, tumor

enhancement requires that the absolute number of virulent cells be relatively

low, and tumor inhibition requires that the number of virulent cells be

relatively high. When the concentration of the virulent strain is held constant

and the concentration of the complementary avirulent receptor strain is

increased, the degree of tumor enhancement increases.

The explanation for tumor enhancement is radically different from the

site-attachment hypothesis proposed to explain the tumor inhibition data.

Inoculation procedures are identical in both tumor enhancement and inhibition

studies. When tumor inhibition is observed, competition for available

attachment sites is suggested to account for the reduction in tumor

formation. When tumor enhancement is observed, it is postulated that

different bacteria communicate between attachment sites in separate wounds.

It would be equally plausible to hypothesize that an increased level of

adsorption leads to the increased tumor formation noted in the complementa-

tion experiments, i.e. that virulent cells facilitate the adsorption of virulent

cells. This has been reported to occur when animal tissue was coinoculated

with two Streptococcus mutants defective in adsorption (Larrimore et al.

1983). Although the timed inoculation data are difficult to explain, the

available evidence also is consistent with the postulate that tumor inhibition is










a result of the influence of inhibitory compounds from avirulent cells on

virulent cells bound to sites competent for transformation. Taken together, the

results suggest that the interactions responsible for the altered tumor

formation are very complex. Consequently, I believe that the evidence

provides little justification to support the extrapolation to bacterial

attachment in the case of tumor inhibition, or the postulation of an additional

mechanism when conditions result in enhancement of tumor formation.



Proposed Molecular Components Involved in Site-attachment


The LPS on the bacterial cell surface are thought to be involved in the

site-attachment process. The LPS isolated from site-binding A. tumefaciens

strains by the hot phenol-water method of Westphal and Jann (1965) inhibit

tumor formation. When LPS were applied to the wound sites prior to

inoculation with B6, a reduction in tumor formation was observed (Whatley et

al. 1976). This was not the case with LPS from non-site-binding strains. When

the LPS were hydrolyzed into two components, the o-antigens and lipid A

region, the o-antigens were shown to be responsible for the inhibitory activity

(Whatley et al. 1976, Lippincott and Lippincott 1977).

Tumorigenicity is suppressed by introduction of plasmid pSa into A.

tumefaciens (Farrand et al. 1981). More recently New et al. (1983) examined

tumor inhibitory activity of strains with and without the pSa plasmid. Strains

harboring pSa were no longer tumorigenic. LPS isolated from this strain also

were unable to inhibit tumor formation by the virulent strain. When these

strains were cured of the pSa plasmid, tumorigenicity and the ability of the

isolated LPS to inhibit tumor formation were restored. Isolated LPS from

strains containing and lacking pSa were compared with a variety of










techniques. Strains with and without the pSa plasmid were not qualitatively

different. Carbohydrate content of the LPS was similar, and the relative

amount of 2-keto-3-deoxyoctonic acid to o-antigenic polysaccharides was not

significantly different (New et al. 1983).

Purity of any of these LPS fractions, however, was not established, and

their compositions were not determined. The LPS isolated using the Westphal

and Jann (1965) technique commonly contain 2-5% nucleic acid by weight.

Detergent treatment will precipitate nucleic acids, but this procedure was not

performed on the LPS used in the tumor inhibition experiments. Glucose-rich

polysaccharides contaminate LPS isolated from Rhizobium sp. (Carlson et al.

1978), and LPS, regardless of the isolation procedure, always contain inorganic

cations and low molecular weight basic amines, e.g. spermine, spermidine and

ethanolamine (Burton and Carter 1964, Luderitz et al. 1968, Galanos and

Luderitz 1975). The LPS fractions are comprised of an unpredictable mixture

of different salt forms. This results in a heterogeneous LPS fraction having

variable physiochemical properties that may affect biological activity (Galanos

and Luderitz, 1975).

Johnson and Perry (1976) suggested five criteria for establishing LPS

purity: i) elution of a single symmetrical peak at the void volume of a

Sepharose 4B or 6B column, ii) the absence of ribose or deoxyribose, iii) lack of

adsorption between 210 and 300 nm, iv) symmetrical elution peaks of o-antigen

and core-polysaccharides from Sephadex G-50 columns, and v) a single

absorbance maximum at 460 nm with the carbocyanine dye assay for LPS. The

LPS material used in the tumor inhibition studies apparently was not examined

by these criteria. Attributing the observed activity to the LPS component thus,

unfortunately, may be misleading.










Plant surface molecular components also were analyzed for their possible

involvement in the site-attachment process. An isolated pinto bean leaf cell

wall fraction inhibited tumor formation, but the plant cell membrane fraction

was without activity (Lippincott et al. 1977b, Lippincott and Lippincott

1978a). Treatment of the cell wall fraction with either hot water or acid

destroyed its inhibitory activity (Lippincott and Lippincott 1977, Rao et al.

1982), whereas treatment with ethylenediaminetetraacetic acid (EDTA),

detergent, pectinase, or cellulose was without effect (Lippincott and Lippincott

1977). However, subsequent work from the same laboratory suggests that

pectinase destroys the ability of the cell wall fraction to inhibit tumor

formation (Lippincott et al. 1983).

Considerable caution, however, should be taken with results obtained

using isolated crude cell wall fractions. These fractions were obtained by

grinding leaves in buffer and removing the particulate wall material by low

speed centrifugation. The material was then suspended in acetone,

homogenized, filtered, washed three times with acetone and air dried. Similar

preparations have been shown to contain 6% (w/v) protein (Slusarenko and Wood

1983). Although purity was not demonstrated for wall fractions used in the

tumor inhibition assays, the authors suggest that the pectic portion of the

fraction is the active component responsible for tumor inhibition. This was

verified by examining several off-the-shelf commercial preparations of cell

wall origin. Citrus pectin and the sodium salt of its unesterified form,

polygalacturonic acid, were found to be inhibitory; cellulose and several other

polysaccharides were without activity. Sodium polygalacturonate exhibited the

greatest degree of tumor inhibition and was active to a concentration of I

ng/ml. The observation that polygalacturanate is much more effective than










pectin led to the conclusion that the degree of methylation of the galacturonan

residues may be responsible for the observed differences in activity (Lippincott

and Lippincott 1980, Rao et al. 1983). The influence of methylation on tumor

formation was examined by enzymatically removing methyl groups with a

commercial preparation of pectin methylesterase. The product of this reaction

had enhanced ability to inhibit tumor formation relative to pectin. On the

other hand, enzymatic methylation with pectin methyltransferase resulted in

reduced tumor inhibition.

In most cases cell wall fragments isolated from monocot tissue exhibited

little inhibitory activity (Lippincott and Lippincott 1978a). However, upon

treatment with pectin methylesterase, they became inhibitory (Lippincott and

Lippincott 1978a). This suggested a possible mechanism for the well-known

resistance of most monocotyledonous plants to infection by Agrobacterium

spp. Thus, the apparent lack of attachment sites in monocots was hypothesized

to be due to highly methylated galacturonan residues that may present a

formidable barrier to tumor formation (Lippincott and Lippincott 1977,

1978a). These results led Lippincott and Lippincott (1977) to state, "The

simplest direct conclusion from these data is that the cell walls of monocots,

tumors and embroyonic bean tissues are sheathed with pectic substances which

are sufficiently methylated that Agrobacterium does not adhere." This

mechanism suggests that treatment of wound sites in monocotyledonous plants

with pectinmethylesterase would expose attachment sites and render the

monocot susceptible to tumor formation. Conversely, treatment of wounds in

dicot tissue with pectin methyl transferase should confer some degree of

resistance to Agrobacterium spp. These experiments, which would provide a

direct test of the attachment site hypothesis, have not been reported.










Tumor inhibition by pectin, polygalacturonate, LPS, and virulent site-

binding A. tumefaciens strains also have been shown by others (Brown 1923,

Schilperoort 1969, Glogowsky and Galsky 1978, Cooksey and Moore 1981,

Pueppke and Benny 1981). Not all these reports, however, totally support the

site-attachment hypothesis. For example, Glogowski and Galsky (1978) found

that both E. coli and Pseudomonas fluorescens inhibited tumor formation on

potato. In addition, pretreatment of Datura stramonium wounds with avirulent

Agrabacterium sp. cells had no effect on tumor formation (Beaud et al. 1963).

Site-binding, virulent A. radiobactor strain K84 reduced tumor formation by

70% when applied 3 hours after inoculation with a virulent, agrocin resistant

strain (EI-Kady and Sule 1981).

Furthermore, others have found, in a somewhat different assay system,

that only living, site-binding strains of A. tumefaciens were able to inhibit

tumor formation (Schilperoort 1969, Bogers 1972, Douglas et al. 1982). These

data may not necessarily conflict with previous reports, because Lippincott and

Lippincott (1969) reported that the heat treated (600 C, 20 min) bacterial

suspensions (1010 to 1011 cells/ml) still contained 104 viable cells/mI. Though

present in low numbers, a population of viable cells could influence the degree

of tumor formation recorded 6 to 7 days later. The number of viable cells

remaining after heat treatment was not reported by Bogers (1972) or

Schilperoort (1969).

It has been reported that a process similar to the proposed site

attachment step in tumor formation is essential for the induction of moss

gametophore formation that occurs in the presence of A. tumefaciens (Spiess et

al. 1971, 1976). In the absence of a wound, gametophore induction requires

physical contact with A. tumefaciens. The extent of gametophore induction is










related linearly to the concentration of bacteria and is inhibited by heat-killed

virulent cells (Spiess et al. 1976). Whatley and Spiess (1977) have identified the

polysaccharide portion of the LPS as the active component that mediates the

attachment of A. tumefaciens to moss. Moss gametophore induction, however,

is not uniquely induced by A. tumefaciens, because five different Rhizobium

spp. also were active (Spiess et al. 1977a). It would prove interesting to

examine the ability of a wide variety of bacteria to induce gametophore

formation. Such a test would directly address the question of whether

gametophore induction by A. tumefaciens is a useful model system to study

early attachment events. When adsorption of non-site binding A. radiobactor

stain 6467 and site-binding A. tumefaciens strain B6 onto moss protonemal

filaments was monitored by scanning electron microscopy, few differences

were found (Spiess et al. 1977b). On the average, the number of B6 cells bound

per spore was six more than that of A. radiobactor 6467. This appears to be the

strongest evidence for site-specific adsorption mediating gametophore

induction.



Adsorption of Agrobacterium spp. to plant tissue culture cells


It should be emphasized that the entire preceding discussion deals with

the use of tumor formation and inhibition data to define the adsorption of

Agrobacterium sp. onto plant tissue. These events are separated by 7 days and

may not be directly related to one another. Several workers have overcome

this problem by directly measuring the attachment of A. tumefaciens onto

tissue culture or mechanically separated plant cells (Matthysse et al. 1978,

Smith and Hindley 1978, Ohyama et al. 1979, Douglas et al. 1982, Draper et al.

1983). In only one case, however, have workers using tissue culture cells









directly monitored bacterial adsorption under conditions where transformation

is known to occur (Smith and Hindley 1978). Using cultured tobacco cells,

Smith and Hindley (1978) found that both virulent C-58 and avirulent NT-I

strains bound, though C-58 was adsorbed in larger numbers. Transformation in

this system was substantiated by measuring the production of nopaline in the

infected tobacco cells. It was subsequently shown that wall-regenerating

tobacco protoplasts can be transformed with both A. tumefaciens cells and

purified Ti plasmid and then regenerated into transformed plants (Marton et al.

1979, Wullems et al. 1981, Krens et al. 1982).

Matthysse et al. (1978, 1981) examined the attachment of various A.

tumefaciens strains to both carrot and tobacco tissue culture cells. Using a

ratio of approximately two bacteria per carrot cell (104 bacterial cells/ml:5 X

103 plant cells/ml) Matthysse et al. (1978) observed that a larger percentage of

the virulent cells bound after incubation for 2 hours than did the avirulent

cells. The initial bacterial inoculum varied from 5 X 103 cells/ml to 1.5 X 104

cells/ml. If 10 cells/ml are assumed to be the initial inoculum, 0.5 to 0.9 cells

of each of five virulent Ti plasmid-containing strains bound per plant cell. On

the other hand, approximately 0.2 bacteria bound per plant cell for each of the

Five tested virulent ogrobacteria. Therefore, in all cases examined, more than

half the plant cells would be expected to be devoid of adsorbed bacteria. Even

more interesting is the fact that when the number of plant cells is increased

10-fold, there is no change in the number of bacteria adsorbed from the

bacterial suspension (Matthysse and Gurlitz 1982).

Close examination of the absolute values of the adsorption data using the

carrot tissue culture cells leads to some interesting observations. Plant cells

adsorbed 24% of an initial inoculum containing 5 X 103 cells of A. tumefaciens











strain A178/ml. Thus 1.2 X 103 bacterial cells were removed from the

suspension. In similar adsorption experiments with A. tumefaciens NT-I, 8-9%

of the initial inoculum (1.5 X 104 cells/ml) attached to the plant surface. Thus

1.2 X 103 NT-I cells were removed from the suspension. Therefore the actual

number of virulent (A178) or virulent (NT-I) bacterial cells bound was

identical and independent of virulence.

Matthysse et al. (1978) also reported that the adsorption of site-binding

strains IIBNV6 and NT-I to tobacco was small compared to that of A.

tumefaciens B6. This deviates from what would be expected if the site-

attachment process as described by Lippincott and Lippincott (1969) is

operative. Since prior application of IIBNV6 or NT-I cells reduced tumor

formation by strain B6, similar levels of adsorption would be postulated. The

differences, however, may reflect differences in the two attachment systems.

On the other hand, they may simply call into question the general applicability

of the site-attachment hypothesis to systems other than pinto bean.

Matthysse et al. (1982) provided additional evidence that questions the

validity of the site-attachment hypothesis for suspension culture systems. The

kinetics of attachment of A. tumefaciens B6 to both glutaraldehyde-fixed

carrot pratoplasts and intact carrot cells from tissue culture were similar.

Protoplasts were determined to be free of cell wall material by calcofluor

white ST staining. Therefore, bacterial attachment to plant tissue occurred in

the absence of the postulated host receptor-site, pectic wall material.

However, partial synthesis of the cell wall is required for the transformation of

tobacco protoplasts by intact A. tumefaciens cells (Marton et al. 1979),

suggesting that A. tumefaciens attachment to carrot protoplasts is not the

prelude to transformation.










Cellulose microfibril production by A. tumefaciens cells is associated

with anchoring the bacterium to the surface of tissue culture plant cells and

protoplasts (Matthysse et al. 1981). The ability of Aqrobacterium spp. to

produce micrafibrils first was described by Deinema and Zevenhuizen (1971)

and later suggested by Spiess et al. (1977b) to play a possible role in bacterial

attachment to moss tissue. Subsequently, however, it has been shown that the

ability to synthesize cellulose microfibrils is not a prerequistie for bacterial

virulence or the ability to attach to plant cells (Matthysse, 1983). Since the

microfibrils are not observed until approximately 60 min after inoculation, they

are postulated to function in anchoring large numbers of bacterial cells, thus

creating a more favorable microenvironment for the subsequent transfer of the

T-DNA.



Adsorption of Agrobacterium spp. to Culture Monocot and Dicot Cells.


Comparative data on the number of bacteria adsorbed to cultured

monocot and dicot cells vary from laboratory to laboratory. Matthysse and

Gurlitz (1982) reported that dicot cells adsorb larger numbers of A.

tumefaciens cells, but the distinction was not apparent in other studies

(Ohyama et al. 1979, Draper et al. 1983). When a ratio of approximately one

bacterial cell to every 100 plant cells is used, a larger percentage of the

inoculum is adsorbed to carrot (Daucus carota) suspension culture cells (48%)

than to corn (Zea mays) suspension culture cells (3%). However, when actual

numbers are calculated from the data, 2 X 105 carrot cells bound 960 bacteria

and, if it is assumed that individual plant cells bind no more than I bacterial

cell, 99.5% of the plant cells have no bacteria adsorbed. Approximately 60

bacteria bound to the 2 X 105 corn cells, leaving 99.9% of the plant cells










without a single bacterium attached. When viewed in this manner, A.

tumefaciens adsorption to monocot versus dicot tissue becomes much less

distinct. This distinction was further blurred by information provided by

Ohyama et al. (1979), who measured the adsorption of radioisotope-labeled A.

tumefaciens B6 to a variety of monocot and dicot suspension culture cells.

Corn, bromegrass (Bromus inermis), rice (Oryza sativa), and tobacco cells bound

equivalent numbers of 86 cells. In each case, 1-4 mg dry wt of plant cell

suspension/mi adsorbed approximately 2.3 X 106 bacteria from an initial

inoculum of 3.6 X 107 cells. Soybean cells bound significantly larger numbers

of B6 cells than did tobacco. It is of interest to note that intact soybean plants

are resistant to infection by most Aqrobacterium spp. Recently however A.

tumefaciens strain A208 from Monsanto Co., St. Louis, Missouri, has been

shown to induce tumor formation on soybean plants (Jill Winter, personal

communication).



Adsorption of A. tumefaciens to Susceptible Host Tissue


When A. tumefaciens adsorption onto susceptible host tissue is measured

directly, neither LPS nor pectic material participates in the adsorption process,

as predicted by the site-attachment hypothesis (Pueppke and Benny 1984).

Pueppke and Benny (1981, 1983) reported that tumor formation on potato tubers

was inhibited by added galacturonans in a similar manner as seen with the

PBLA. In contrast, both the degree of galacturonan methylation and the

presence of LPS in the inoculum was without influence on tumor inhibition in

the potato system (Pueppke and Benny 1983). A single exception was noted

with the virulent, heat-cured strain NT-I, the LPS of which was inhibitory to

tumor formation. Adsorption assays consisted of measuring the level of










attachment of 35S-methionine-labeled bacteria to potato tuber discs (Pueppke

and Benny 1984). Four different galacturonans, which varied from 0 to 11.7%

methoxy by weight, were examined for their influence on adsorption. When

adsorption was measured in the presence of citrus pectin (11.7% methoxy),

polygalacturonic acid (0% methoxy) and demethylpectin (2.6% methoxy), the

number of adsorbed bacteria was statistically greater than that in the absence

of each compound. The presence of methylpolygalacturonic acid (8.2%

methoxy) in the inoculum caused no statistically significant change. Adsorption

in the presence of LPS isolated from virulent and avirulent strains was not

statistically different from adsorption measured in the absence of LPS. LPS

from a single strain, NT-I, were enigmatic and inhibited bacterial adsorption.

LPS from strain NT-I were the only additive that inhibited both adsorption and

tumor formation.



Summary of Literature Review


According to the specific site-attachment hypothesis there exists a single

site within a wound to which a single virulent bacterium can attach and cause

tumor formation. According to the hypothesis, tumor inhibition results from

the exclusion of a single virulent cell from a specific site by a single virulent,

site-binding cell. Adsorption is hypothesized to be mediated by the binding of

bacterial cell surface LPS to sparingly methylated galacturonans in wounded

dicot plant cell walls. The hypothesis has several corollaries. The first is that

monocotyledonous plants are resistant to A. tumefaciens because the bacteria

are unable to adsorb to the plant cell surface (Lippincott and Lippincott 1977,










1978a, Rao et al. 1982). The second is that there exist two different types of

virulent A. tumefaciens cells, which can be distinguished by their abilities to

bind to plant surfaces.

The site-attachment hypothesis and its corollaries have been discussed

and for the most part accepted by many reviewers of A. tumefaciens

tumorigenesis (Merlo 1978, Dazzo 1980a, Dazzo 1980b, Moore and Cooksey

1981, Nester and Kosuge 1981). However, the hypothesis is entirely based on

indirect data from pinto bean. In addition, although adsorption has been

measured directly and quantified, the experimental conditions used in these

studies do not lead to transformation. Thus, there is no assurance that the

measured adsorption events are relevant to T-DNA transfer.

The initial overall objective of this study was to test the site-attachment

hypothesis and its corollaries. This was done (I) by measuring the adsorption of

virulent, avirulent site-binding, and avirulent nonsite-binding strains to

susceptible host tissue under conditions known to permit T-DNA transfer and

(2) by comparing the attachment of A. tumefaciens cells to monocotyledonous

and dicotyledonous plant tissues. In addition, I performed a series of

experiments designed to define the mechanism of A. tumefaciens adsorption to

potato tissue.














CHAPTER THREE
ADSORPTION OF AGROBACTERIUM TUMEFACIENS TO
POTATO (SOLANUM TUBEROSUM) TISSUE

Introduction

Aqrobacterium tumefaciens is a plant pathogen that causes a tumorigenic

disease known as crown gall. This pathogen exhibits an extremely wide host

range that includes 86% of the tested gymnosperm families, 38% of the tested

dicot families and several monocotyledonous plants (De Cleene and De Ley

1976). Crown gall tumors are neoplastic growths that result from the transfer

and stable incorporation of a piece of bacterial plasmid (T-DNA) into the plant

nuclear DNA (Chilton et al. 1977, Yadau et al. 1980).

Adsorption of A. tumefaciens to the surface of the plant host is thought

to be one of the earliest events leading to tumor formation (Lippincott and

Lippincott 1969, 1977, 1980). Little is known of the precise adsorption

mechanism, although adsorption most likely results from the interaction of both

attractive and repulsive forces acting at the plant-bacterium interface.

Hydrophobicity, cell surface charges, and surface polymers all may contribute

to the formation of a stable host-parasite cell interface. Several authors have

suggested that adsorption in this system is site-specific (Lippincott and

Lippincott 1969, Lippincott et al. 1977b). Most of the supporting data are

from experiments that measure the effects of various isolated plant or

bacterial compounds on the ability of A. tumefaciens strain B6 to form tumors

on wounded pinto bean (Phaseolus vulgaris) leaves (Whatley et al. 1976,

Lippincott et al. 1977b, Rao et al. 1982). The compounds are added to the

bacterial inoculum or applied prior to inoculation. After 7 days, the number of










tumors is counted. Substances that reduce the number of tumors are

considered to do so because they interfere with the initial bacterial adsorption

step.

The bacterial component responsible for mediating this site-specific

adsorption is believed to be the o-antigen portion of the lipopolysaccharides

(LPS) (Whatley et al. 1976, Whatley and Spiess 1977). Exogenously supplied LPS

interfere with tumor formation and are thought to recognize some

galacturonan-rich pectic portion of the wounded plant cell wall (Whatley et al.

1976, Lippincott et al. 1977a, 1980). Exogenously supplied pectic compounds

also inhibit tumor formation. It has been suggested that the degree of

methylation of the pectic material influences the ability of A. tumefaciens to

adsorb to plant cell surfaces. Thus, the resistance of very young dicotyledonous

plants and most monocotyledonous species to A. tumefaciens, is hypothesized

to reflect the extent of methylation of pectic polysaccharides of the cell wall

(Lippincott and Lippincott 1978a).

In a system where bacterial adsorption to potato tuber tissue is measured

directly, rather than inferred from tumor experiments, most LPS and various

pectic compounds have no inhibitory effect on adsorption (Pueppke and Benny

1984). However, citrus pectin actually enhanced bacterial adsorption, and LPS

isolated from A. tumefaciens strain NT-I, the single exception, reduced the

number of bacteria adsorbed to the potato disc surface. In this system, pectic

compounds dramatically reduce tumor formation, and LPS have little or no

effect (Pueppke and Benny 1981, 1983). In addition, when attachment of A.

tumefaciens to monocot tissue is measured directly, the numbers of bound

bacteria are comparable to those bound to potato tissue (Chapter Five this

dissertation). These observations suggest that adsorption of A. tumefaciens to

plant cells is less specific than previously thought.











In this investigation my objective is to provide a quantitative description

of the initial binding of A. tumefaciens to potato tissue.



Materials and Methods



Adsorption Experiments


Tumorigenic A. tumefaciens strain B6, from J. A. Lippincott,

Northwestern University, was maintained at 40C on slants of the defined

gluconate-mannitol medium of Bhuvaneswari et al. (1977). After incubation for

2 to 3 days A. tumefaciens cells were washed from the slants using sterile

phosphate-buffered saline (PBS: 7.2 g NaCI, 2.79 g Na2HPO4 71-120, 0.43 g

KH2PO4 per liter of deionized water; final pH 7.2), centrifuged for 10 min at

7700 X G, and washed twice with PBS. The washed bacteria were resuspended

in PBS and the bacterial concentration determined turbidimetrically.

The effect of initial bacterial concentration on bacterial adsorption to

plant tissue was assayed by a modification of the tumorigenesis assay described

previously (Pueppke and Benny 1981). Tuber discs of potato (Solanum

tuberosum L. cv. Red LaSoda) (8 mm diam X 2 mm thick) were prepared

aseptically. With the use of a wire basket, batches of discs (30 discs/basket)

were lowered into a 100-ml beaker containing 30 ml of bacterial inoculum (101

to 109 cells/ml of PBS). The discs were agitated gently and allowed to remain

undisturbed for 60 min. The basket containing the discs then was removed and

immersed in two separate rinses, each containing 30 ml of sterile PBS. Each of

the two rinses consisted of vigorous agitation of 4 to 5 seconds to remove any

loosely attached bacterial cells. The discs were removed from the basket,

blotted dry on sterile paper towels, and three randomly chosen discs were











transferred to 2 ml of PBS contained in a hand-held, ground-glass tissue

homogenizer. The discs then were homogenized. Aliquots of 100 ul from each

of three dilutions, previously determined to give less than 500 colonies/plate,

were spread on each of five petri plates containing yest extract-mannitol agar

(0.5 g K2HPO4, 0.2 g MgSO4 7H20, 0.1 g NaCI, 10 g mannitol, 0.4 g yeast

extract per liter of deionized water). Viable colonies were counted after

incubation of the plates for 36 to 48 hr at 280C. The reported number of

bacteria adsorbed at each inoculum density is the average from two to five

experiments, each of which was replicated three times. Therefore, the

adsorption value at each inoculum density is derived from measurements of

colonies on 6 to 15 individual plates.

As a control, noninoculated discs were homogenized, and the homogenates

spread on plates. Although a few colonies formed infrequently, none resembled

the white, mucoid A. tumefaciens colonies. To determine the effect of

homogenization on bacterial viability, a known number of bacterial cells was

added to the buffer prior to homogenization. Recovery of these cells was not

influenced by the homogenization step.



Desorption Experiments


Potato tuber discs were incubated in a bacterial suspension of 109

cells/ml of PBS for 40 min. Three discs were removed, washed as described

above, and the number of attached bacteria determined. The remaining

bacterial suspension, which then contained 27 tuber discs, was diluted with PBS

such that the bacterial concentration was reduced by one log unit, i.e. to 108

cells/ml. The initial I:1 ratio of potato discs to ml of bacterial suspension was

always maintained. After 40 min another three discs were removed and the










number of adsorbed bacterial cells determined as above. Two additional

desorption experiments were done in an identical manner using a stepwise

reduction in bacterial density from 108 to 107 cells/ml of PBS, and 106 to 104

cells/ml of PBS.



Scanning Electron Microscopy


Potato discs for scanning electron microscopy were obtained immediately

after inoculation with 2 X 107 cells of strain B6/ml of PBS. Discs were fixed

overnight at 40C in 4% glutaraldehyde and rinsed with 0.05 M cacodylate buffer

at pH 7.2. Samples then were fixed in 0s04 for 2 hr at room temperature,

washed with cacodylate buffer, and dehydrated through an ascending series of

ethanol solutions. Samples were critical-point dried by substitution of ethanol

with liquid CO2, sputter coated with gold, and viewed with a Hitachi 450

Scanning electron microscope at 20 KV.



Results

Bacterial Adsorption and Desorption


Depending on initial bacterial concentration, 0.23 X 101 to 1.2 X 107

bacteria adsorbed per disc (Table I). The percentage of bacteria adsorbed from

suspensions containing from 10 to 104 cells/mi of PBS varied from 23 to 5.

Approximately 1% of the inoculum was adsorbed at bacterial concentrations

greater than 105 cells/ml of PBS (Table I). Saturation was never observed in

these experiments. Adsorption studies also were carried out using potato discs










Table I. Adsorption of Agrobacterium tumefaciens to Potato Tissue.

Initial Bacterial No. of Bacteria Adsorbed/Disc
Concentration
(cells/ml) Measured" (%) Predictedb

101 0.23 X 101(23.0) 0.13 X 101
5X 101 0.55 X 101(11.0) 0.57 X 10I
102 0.85 X 101 (9.0) 1.00 X 101
5X 102 7.50 X 101(15.0) 3.60 X 10I
103 7.80 X 10 (8.0) 6.70 X 101
5X 103 3.30 X 102 (7.0) 2.60 X 102
104 5.10 X 102 (5.0) 4.70 X 102
105 1.60 X 103 (1.6) 3.20 X 103
106 1.10X 104 (1.1) 2.00X 104
107 1.30 X 105 (1.3) 1.40X 105
5X 107 5.80 X 105 (1.2) 5.30 X 105
108 1.30 X 106 (1.3) 9.40 X 105
5 X 108 6.30 X 106 (1.3) 3.50 X 106
109 1.20X 107 (1.2) 6.20 X 106

a The data are from four separate experiments. Three discs were homogenized at
each sampling time, and each homoginate was spread on five plates.
b Predictions are based on the Freundlich adsorption isotherm, Cb = 0.25 1Cs 2.










with a 43% smaller surface area than those used in most experiments. When

these smaller discs were exposed to bacterial densities up to 109 cells/ml of

PBS, saturation of the available adsorption sites was not reached.

When bacterial suspensions in equilibrium with potato discs were diluted

from 109 to 108 cells/ml, there was coincident reduction in the number of

bacterial cells adsorbed to the discs (Table 2). The number of bacteria

adsorbed per disc also decreased after the 10-fold dilution from 108 to 107

cells/ml. A 100-fold dilution from 106 to 104 cells/ml similarly resulted in a

reduction in the number of cells bound per disc. These data indicate that

binding is reversible.


Theoretical Considerations


Before experimental data can be analyzed with respect to adsorption

isotherms, several preconditions about the adsorption system under analysis

must be met (Brunauer 1943, Saunders 1971, Tompkins 1978). First, the total

number of adsorption sites should be constant under all experimental

conditions. This condition is met because discs of the same size were used in

each experiment. Second, adsorption must be restricted to a monolayer. This

was confirmed by examination of inoculated potato tuber surfaces by scanning

electron microscopy (Fig. I). The existence of monolayer adsorption eliminates

the possible use of the Brunauer-Emmett-Teller adsorption isotherm, which

describes multilayer adsorption phenomena (Brunauer 1943). Third, every

adsorption site should be equivalent, and the ability of the bacteria to adsorb

must be independent of whether or not adjacent sites are occupied. These

conditions also appear to be fulfilled. Potato tuber tissue is relatively uniform,







38




Table 2. Desorption of Agrobacterium tumefaciens Cells from Potato Tuber
Discs.a


Initial No. Bacteria Final No. Bacteria

Per ml of Inoculum Adsorbed/Disc Per ml of Inoculum Adsorbed/Disc


109 2.8 + 0.3 X 107 108 5.9 + 1.3 X 106
108 2.0 + 0.5 X 106 107 9.7 + 1.3 X 105
106 1.4 + 0.2 X 104 102 6.0 + 0.3 X 103

a The data are from separate experiments. Three discs were homogenized at
each sample time, and each homogenate was spread on plates.
b Standard Deviation.













































Fig. I. Scanning electron micrograph of Agrobacterium tumefacbens strain B6
on potato tuber tissue. The initial inoculum density was 2 X 10' bacteria/ml
and the disc was fixed immediately after inoculation. The field of view
presented here is 870 urm and contains 12 rod-shaped bacteria. X 5200.










and it would be expected that adsorption sites would be reasonably equivalent.

Moreover, bacterial adsorption to the discs did not reach saturation, even when

small discs were exposed to very high bacterial densities. This makes it

unlikely that the plant cell surfaces were packed with.bacteria and that

occupation of adjacent sites was of importance. These observations provide

evidence for the equivalency of adsorption sites. Fourth, the rate of adsorption

to unoccupied sites must equal the rate of desorption from occupied sites at

equilibrium, i.e. equilibrium must be reversible. This is illustrated in Table 2.

The binding of A. tumefaciens strain B6 to potato tuber tissue con be

described best in terms of the Freundlich adsorption isotherm (Brunauer 1943,

Tompkins, 1978), which is expressed as

Cb= k X Csl/k2
where Cb = the number of bacterial bound per disc,

Cs = the concentration of the bacteria in the suspension at

equilibrium,

kI and k2 are both constants.

When a linear regression is performed on the plot of log Cs vs log Cb, a

correlation coefficient of r = 0.987 is obtained for the entire range of inoculum

densities studied. Demonstration of the linear relationship between log C, vs

log Cb is in accordance with the criteria set forth for the Freudlich isotherm

(Brunanuer 1943, Tompkins 1978).

When the data are fitted to three other isotherms (Langmuir, Temkin, or

Henry's law), the curves fall outside acceptable levels of confidence. The

Temkin isotherm requires the relationship between log Cs and Cb to be linear.

My data yield a correlation coefficient of r = 0.60. The Langmuir isotherm

requires the plot of Cs/Cb vs Cs to be linear. My data yield a correlation










coefficient of r = 0.39. Henry's law defines a linear relationship over the entire

isotherm. The potato-A. tumefaciens adsorption isotherm is nonlinear in the

range of 101 to 105 cells/mi of PBS (Table I). Therefore, none of these three

alternative isotherms adequately describes the experimental data for the entire

inoculum range studied.

When theoretical isotherm plots of bound versus free bacteria for each of

the isotherms in Fig. 2 are compared to the experimental data presented in

Table I, the empirical correspondence of the data to the Freundlich isotherm

plot is very good. Similar to Fig. 2a, the experimental data show a rapid

increase in bacterial adsorption per unit increase in initial inoculum density

from 101 to 104. At higher initial bacterial densities, in the range of 105 to

109 cells/mi of PBS, the numbers of bacteria adsorbed are linearly related to

the initial inoculum density. At all high bacterial densities, approximately

1.2% of the bacteria adsorb to the potato discs. This simple linear part of the

adsorption isotherm is described best by Henry's law: Cb = ki X Cs (Fig. 2c).

Saturation as predicted by the Langmuir isotherm (Fig. 2a) is not reached.

Linear regression analysis of log Cs vs log Cb was used to determine the slope

and y-intercept. These values then were used to determine values of the

constants kI and k2 of the Freundlich isotherm over the entire range of

bacterial concentrations (101 to 109 cells/ml of PBS):

Cb=kl XCs 1/k2

When log Cb = log kI + 1/k2 log Cs

Slope 1/k2 = 0.822

y-intercept log k = -0.601

k = 0.251










Thus the Freundlich adsorption isotherm for the adsorption of A.

tumefaciens to potato tuber discs is

Cb= 0.251 X Cs 1/1.217 = 0.251 X Cs 0.822 (a)

Using equation (a) a predicted curve of log C. vs log Cb was plotted, and the

experimental data points superimposed (Fig. 3). The predicted number of

adsorbed bacteria per disc at each inoculum density is listed in Table I. A

curve fit to the experimental data points using the Statistical Analysis System

(SAS) nonlinear curve fitting procedures (NLIN) generated

Cb = (0.098) X Cs 0.897

for parameters kI = 0.098 and I/k2 = 0.897, asymptotic 95% confidence

intervals are -0.062 to 0.257 and 0.984 to 0.823 respectively. Clearly, the

parameters kI and 1/k2 generated from the log Cs vs log Cb plot fall within

these confidence intervals.


Discussion


The binding of cells of A. tumefaciens strain B6 to potato tuber tissue is a

reversible equilibrium process. This binding can be described accurately in

terms of the Freundlich adsorption isotherm, which is a non-ideal variant of the

Langmuir adsorption isotherm (Brunauer 1943). Over the wide range of initial

bacterial concentrations examined, non-saturation of available binding sites is

in accordance with the Freundlich isotherm.

Adsorption of cells of Pseudomonas lachrymans onto leaf surfaces is

nonspecific and nonsaturating over inoculum densities of 104 to 1010 cells/leaf

(Haas and Rotem 1976). Leben and Whitmoyer (1979) also observed that the

number of adsorbed P. lachrymans cells is related linearly to the number of

cells applied. Fletcher (1977) suggested that the initial adsorption of marine
















10

LOG Cb=0.82 LOG Cs-0.60

8



6



4


y-
2




0 2 4 6 8 10
LOG Cs


Fig. 3. The Freudlich adsorption isotherm for Agrobacterium tumefaciens
strain B6 on potato tuber tissues. The line is the predicted adsorption curve
on log Cs vs log Cb generated from equation (a). The experimental data
points are from Table 3.











pseudomonads to nonbiological surfaces is driven by nonbiological,

physiochemical adsorption forces as described by a modified Langmuir-like

adsorption isotherm. In addition, the adsorption of Rhizobium spp. to cereal

roots was described by the Langmuir isotherm and hypothesized to be

influenced by nonspecific, as opposed to specific binding events (Shimshick and

Hebert 1978, 1979). My data similarly point to the importance of nonspecific

forces in the adsorption of A. tumefaciens to potato tissue.

The experimental data presented here do not lend support to the specific

site attachment hypothesis advanced by Lippincott and Lippincott (1969, 1977,

1980). According to these authors, there exist in wounded pinto bean leaves a

small, finite number of specific binding sites for tumorigenic A. tumefaciens

cells. Heat-treated, tumorigenic A. tumefaciens cells and cells of certain

nontumorigenic Agrobacterium sp. strains are hypothesized to be capable of

occupying these sites and excluding pathogenic cells from them. Coinoculation

studies using unlabeled bacteria in competition with radioisotope-labeled

bacteria indicated that cells are not competing for a limited number of specific

sites on the plant surface (Pueppke, unpublished observations). The data also

indicate that cut surfaces of potato tuber discs contain an extremely large

number of nonspecific adsorption sites for A. tumefaciens. Additionally, if a

specific site-attachment phenomenon is operative, one would expect bacterial

adsorption to be saturable as described by the Langmuir isotherm. This clearly

is not the case with adsorption of A. tumefaciens to potato tuber tissue.

There are three possible explanations for the discrepancies between the

data presented here and those of Lippincott and associates. The first is simply

that the adsorption of A. tumefaciens to potato discs is much different than

that to pinto bean leaves. Although this is possible, A. tumefaciens infects and










produces tumors on both plants, and similar infection mechanisms are to be

expected. The second explanation is that Lippincott and associates may have

been misled by their extrapolation of tumor data to the adsorption

phenomenon. Many events intervene between initial bacterial adsorption,

which occurs rapidly after inoculation, and the appearance of tumors, which

occurs 7 to 14 days after inoculation. It is to be expected that some

treatments would influence tumor development without having an effect on

adsorption. Pueppke and Benny (1984) directly monitored adsorption of A.

tumefaciens to potato tissue in the presence of either LPS or various pectic

components. Pectic components reduced the numbers of tumors formed on

potato and all but one LPS preparation was without effect (Pueppke and Benny

1983). In nearly every case, however, these substances did not reduce bacterial

adsorption, suggesting that pectic materials influence some step subsequent to

adsorption. The existence of two types of binding sites is a third possible

explanation. A small number of highly specific irreversible adsorption sites

that lead to tumor formation may coexist with a large number of nonspecific,

reversible adsorption sites.

The dose-response relationship between inoculum concentration and

tumor number is linear at bacterial concentrations from 105 to 108 bacteria/ml

but the response saturates at higher bacterial concentrations (Pueppke and

Benny 1981). In contrast, the relationship between inoculum concentration and

adsorption is linear at bacterial concentrations from 105 to 109 bacteria/ml.

Thus, there is a lower threshold below which tumors do not form and an upper

threshold above which the tumor response does not keep pace with increases in

adsorption. These differences are not surprising, given the complex nature of

tumorigenesis relative to that of adsorption and the fact that tumors coalesce









and become unresolvable at bacterial concentrations above 108 bacteria/ml

(Pueppke and Benny 1981). The effect of inoculum concentration on the

adsorption of rhizobia to soybean roots (Pueppke 1984b) and the subsequent

nodulation response (Perkins 1925) appears to be similar.

It should be noted that adsorption isotherms are only empirical

expressions that reasonably approximate the behavior of experimental binding

systems over restricted concentration ranges. Although the Freundlich

isotherm was derived for nonbiological systems, the compliance of my data

with the necessary assumptions and principles of this model underscores the

importance of nonspecific physical forces in A. tumefaciens adsorption.

Although it is conceivable that the good fit of the experimental data to the

Freundlich adsorption isotherm is fortuitous, the data justify further

examination of bacterial adsorption in terms of nonspecific physical forces.

Adherence to a particular adsorption isotherm does not predicate a particular

type of binding far the system. Adsorption described by the Freundlich

adsorption isotherm, for example, does not imply that a physisorption or a

chemisorption process is operative. Rather, the isotherm should permit useful

inferences to be drawn and parameters to be set for further testing of bacterial

adsorption to plant surfaces.













CHAPTER FOUR

THERMODYNAMICS OF AGROBACTERIUM TUMEFACIENS ADSORPTION
TO PLANT SURFACES


Introduction


The adsorption of bacteria to plant cell surfaces is acknowledged to play

an important role in the ecology of prokaryote-plant interactions (Dazzo 1980a,

Pueppke 1984a). Before legume nodulation can occur, Rhizobium spp. must

adsorb to the root hair surface to initiate the nodulation process (Schmidt 1979,

Bhuvaneswari 1981). Bacteria also adsorb quite tenaciously to leaf surfaces

(Haas and Rotem 1976, Leben and Whitmoyer 1979). The differential

adsorption of Xanthomonas spp. or Pseudomonas spp. to mesophyll cell surfaces

may determine whether a disease or hypersensitive response will occur

(Klement 1982).

Additionally, tumor formation studies imply that site-specific adsorption

of A. tumefaciens to plant wounds is required for plant cell transformation

(Lippincott and Lippincott 1969). In view of the fact that a piece of the

bacterial T-DNA is mobilized and incorporated into the plant nuclear genome,

adsorption to the plant cell surface appears to be a logical precondition for

transformation. It has been shown that A. tumefaciens adsorbs to potato tissue

in very large numbers (Chapter Three this dissertation, Pueppke and Benny

1984) and that adsorption can be described in terms of the Freundlich

adsorption isotherm (Chapter Three this dissertation). However, the basic

mechanism governing the adsorption of A. tumefaciens is poorly understood.










The objective here was to describe the mechanism of A. tumefaciens adsorption

in terms of physi- or chemisorption.



Materials and Methods



Maintenance of Bacteria and Preparation of Inocula


Agrobacterium tumefaciens strain B6 was maintained at 60C on

glutamate-mannitol agar slants. For all adsorption experiments bacteria were

washed from 4- to 5-day-old slants with phosphate-buffered saline (PBS),

centrifuged at 8,000 X G, and resuspended in PBS to the desired bacterial

concentration. Bacterial concentration was determined turbidimetrically.



Effect of Temperature on Adsorption


Adsorption experiments at elevated temperatures were done in a 380C

water bath. Before the adsorption experiment, the bacterial suspension and

PBS solutions to be used for rinsing were equilibrated to 380C in the water bath

for 25 min. Potato discs were placed in a covered plastic petri dish, which was

floated in the 380C water bath for approximately 3 min to equilibrate to

380C. The 6C adsorption experiments were carried out in a refrigerator,

where all washes, and bacterial suspensions were equilibrated to 60C for 25 min

prior to the adsorption assay. Potato discs were placed in a covered plastic

petri dish and stored at 60C for 3 min prior to the adsorption assay. The room

temperature experiments were done at 280C. The adsorption isotherm

experiments were repeated three times at each of the three temperatures.









The entire adsorption assay used here has been described in detail

elsewhere (Chapter Three this dissertation). After a 60 min adsorption period,

discs were washed three times in consecutive 30-ml portions of PBS, ground in

PBS, serially diluted and plated onto yeast extract mannitol (YEM) agar

plates. Viable colonies were counted after incubation for 36 to 48 hr at 280C.

At each inoculum density the value reported is the average of three

replications. The number of viable colonies in four plates at three different

dilutions was tallied to determine the number of bacterial cells adsorbed per

potato disc.



Kinetic Binding Experiments


These experiments employed the adsorption assay described above.

However, the binding periods were interrupted at predetermined intervals (0.5,

1, 5, 10, 30, and 60 min) and the number of adsorbed bacteria determined.

Baskets containing the discs immersed in the bacterial suspension (2 X 10

bacteria/mi) were removed after the prescribed binding period and rinsed as

described above. The number of bacteria bound per disc then was determined

as above.



Desorption Experiments


After incubation of the discs with 2 X 107 bacteria/ml for 30 min the

discs were removed and quickly washed in sterile PBS for 2 to 3 sec. Each disc

then was placed in a 1.5-cm X 12.5-cm test tube containing 2 ml of sterile PBS

at 6C. Aliquots of 100 ul were withdrawn at specific intervals over a 10.5-hr

period. At each sample time the disc was gently swirled before the sample was









withdrawn. These aliquots were either plated directly onto YEM agar or

serially diluted prior to plating. The number of bacteria adsorbed per disc prior

to and after the desorption period was determined as described previously

(Chapter Three this dissertation). The experiment was conducted twice with

three replicates taken at each sample time.



Sequential Adsorption


The standard adsorption assay using A. tumefaciens strain B6 was used

with the following modifications. A bacterial suspension containing 2 X 10

bacteria/ml was used to inoculate one set of discs at the ratio of one disc/ml of

bacterial suspension. After removal of the discs, another set of discs was

placed into the inoculum. After removal of the second set of discs, the

bacterial suspension was used to inoculate a third set of freshly prepared potato

tuber discs. Individual discs from all three sets were homogenized, and the

number of bacteria adsorbed per disc was determined. The adsorption period

was 30 min, and the temperature was 280C. The sequential adsorption

experiment was repeated twice with four replications in each repitition.



Zeta Potential Measurements


Bacteria were washed three times with sterile, deionized water,

centrifuged at 8000 X G for 12 min and resuspended in sterile, deionized

water. The bacteria were adjusted to 5 X 107 cells/ml of sterile deionized

water, and then titrated to the desired pH with KOH or HNO3 just prior to

measurement of the zeta potential. Each suspension was injected into the

electrode chamber of a Laser Zee Meter Model 502, and the focal plane above










the inside bottom of the electrode cell adjusted to 873 um. All zeta potential

measurements were taken in this focal plane, the upper stationary phase of the

cell. The zeta potential of the bacteria was measured using an electrical field

of 75 volts. Two to three samples were measured at-each pH, and three to five

measurements of each sample were recorded. After all measurements were

made, the sample was decanted and the pH remeasured. Deviation of greater

than 0.1 pH unit was not observed. Due to the relative transparency of the

bacteria to the laser beam across the cell, the bacteria were visualized with

substage tungsten illumination. A video camera projected the image onto a

video monitor where bacterial mobility in the electrical field was observed.



Results



The adsorption of A. tumefaciens was measured at 60C, 280C, and 380C.

The adsorption isotherms obtained at each of these three temperatures, using

bacterial concentrations spanning three orders of magnitude, are illustrated in

Fig. 4. Least squares regression analysis using the General Linear Model of the

Statistical Analysis System was used to compare slopes and y-intercepts of the

isotherms. Log-transformed data of both the initial inoculum and number

adsorbed were used in the regression analysis, and temperature was considered

to be a class variable. Temperature did not significantly alter the y-intercepts

of the three isotherms at P = 0.05. The slopes, however, were significantly

different at P = 0.05, but not at P = 0.045. When the same analysis was carried































10-








z
2 104













15 106 107 108

INOCULUM DENSITY


Fig. 4. Isotherms for the adsorption of Agrobacterium tumefaciens cells to
potato tissue at three different temperatures. Each point is the mean of
measurements taken in three separate experiments with three replications in
each experiment. Bars represent the standard errors of the means.










out using the 280C and 380C isotherms, neither the slopes nor the y-intercepts

were significantly different (Zarr 1974). Comparison of these isotherms using

simple linear regression is valid, because at these bacterial concentrations the

adsorption isotherm is described by a linear relationship, Henry's law: Cb = kCs

(Brunauer 1943, Chapter Three this dissertation).

Isosteric heats of adsorption were determined using the Clausius-

Clapeyron equation expressed as

AH/R X d(I/T) = din C

where AH = change in heat content of the system;

R = the gas constant, 0.082 I-atm/mole K;

T = degrees Kelvin;

Cs = [free bacteria] eq, the concentration of unbound

bacteria at equilibrium.


In the form of a straight line the expression becomes

In Cs=AH/R X I/T.


The concentration of free bacteria at equilibrium at a given level of bacterial

adsorption for each of the three temperatures examined is plotted against I/T

in Fig. 5. In plots of log Cs against I/T, at a constant level of bacterial

adsorption, H/R is the slope of a line that allows the heat of adsorption to be

derived. The correlation coefficient for each of the four lines plotted in Fig. 5

is greater than r = 0.91. Analysis of variance shows the slopes of this family of

lines not to be different from one another and not to be statistically different

from zero. This indicates that the heat of adsorption is exceedingly small.












































































Vt31V8a 338J 'ON 301


U,





0




07

-I


Lo










Kinetics of Adsorption


The number of bacteria adsorbed per disc was measured as a function of

the adsorption period, which is defined as the period of time during which the

potato discs are immersed in the bacterial suspension. More than 2.8 X 104

cells bound per disc after incubation for 30 sec (Table 3). The number of

bacteria adsorbed to the surface continued to increase with time. However,

there is no significant difference between the number of cells bound during a 30

min or a 60 min adsorption period.



Desorption Experiments


The adsorption of A. tumefaciens to potato tuber tissue can be described

by the Freundlich adsorption isotherm (Chapter Three this dissertation). Such a

dynamic equilibrium model implies that the adsorbed bacteria must desorb and

undergo a reversible binding. Bacterial desorption was measured over a period

of 10 hr at 60C to reduce interference due to bacterial multiplication (Fig. 6).

Bacterial desorption is not linearly related to time. After 4 to 5 hr the

desorption curve flattens as equilibrium is approached, i.e. the rates of

desorption and adsorption are equal (Fig. 6). By 10.5 hr approximately 31% of

the bound bacteria (5.8 X 104 of 1.9 X 105 bacteria) had desorbed from each

disc.



Sequential Adsorption


A single bacterial suspension was used to sequentially inoculate three

groups of freshly cut potato tuber discs. The mean number of bacteria






59


Table 3. Time Dependence of the Adsorption of Aqrobocterium tumefociens B6
to Potato Tuber Tissue.


Adsorption Perioda Number of Bacteria (X 105)
(Min.) Adsorbed per Discb


0.5 0.29 + 0.06 a
1.0 0.46 + 0.09 b
10.0 1.44 + 0.73 c
30.0 2.60 + 0.28 d
60.0 2.50 + 0.40 d


a All adsorption experiments were carried out with an initial inoculum of 2 X
10 cells/ml.
b Numbers followed by the same letter are not significantly different (p = 0.05)
according to Student's t-test.






60













10


8


6

6"

4






2









1 I I I I
0 2 4 6 8 10
HOURS


Fig. 6. Desorption of Agrobacterium tumefaciens from potato tuber disc
surfaces. The log of the total number of bacterial cells desorbed from the
surface is plotted against the time in hours. Bars represent + I standard
deviation from the mean.
deviation from the mean.










adsorbed per disc (+ standard deviation) for groups one, two, and three was 9.8

+ 4.1 X 104, 1.1 + 0.43 X 105, and 1.5 +0.66 X 105, respectively. There were

no significant differences between any of the three groups at the 1% level,

when analyzed using Student's t-test (Fig. 7).



Zeta Potential Measurements


Zeta potential measurements of bacteria suspended in deionized water

indicate an isoelectric point (pl) of approximately 2.9 (Fig. 8). Above pH 2.9

the bacterial cell surface was strongly negative. A zeta potential of greater

than -30 mV was observed for values above pH 5.0. Below pH 2.9 the bacterial

cell surface was positively charged. However, the magnitude of the positive

charge was not as great as that of the negative charge at an equivalent

distance above the pl.



Discussion


Although Agrobacterium tumefaciens adsorption can be described by the

Freundlich adsorption isotherm (Chapter Three this dissertation), this does not

imply an adsorption mechanism for the system. The basic effort in this

communication was to define the mechanism of adsorption of A. tumefaciens to

potato tuber tissue by distinguishing between physisorption and chemisorption.

I have attempted to do this by comparing and contrasting my experimental

results with the salient features of each of these two basic adsorption types.

Bacterial adsorption to potato discs occurs very rapidly and in large numbers at






















14



12 -


!210


8-


6


4


C2-

0

1 2 3

GROUP


Fig. 7. Sequential adsorption assay. Three discs were sampled in each
sequential group, and the experiment was done twice. Bars represent + I
standard deviation from the mean.







63




















10
10
*








1 2 4 5
pH




-10



o



-20







-30







Fig. 8. Zeta potential measurements. Each point represents the mean of four
measurements taken for each of the three samples.










bound per disc. Unless the heat of activation for bacterial adsorption is

extremely small, these data suggest that physisorption is operative.

Most chemisorbed particles are irreversibly bound and exceedingly

difficult to remove (Brunauer 1943). Adsorption of A. tumefaciens, however, is

reversible and approaches an equilibrium. Adsorption assays done at 60C, 280,

and 380C indicate that temperature does not have a significant effect on the

number of bacteria bound. The heat of adsorption as determined with the

Clausius-Clapeyron equation is nearly zero, well below the 10-20 kcal value

arbitrarily established as the point below which physisorption most likely

predominates (Brunauer 1943, Saunders 1971). These data all indicate that the

binding of A. tumefaciens can be described in terms of physisorption.

If the adsorption process is to be favored, the free energy of the system

will decrease as particles are adsorbed onto the surface. This concept is

related by the expression;
.Fadh
Fadh = S -yBL ySL

where A Fadh is the free energy of adsorption, yBS is the bacterium-substratum

interfacial tension, 1L is the bacterium-liquid interfacial tension, and YSL is

the substratum-liquid interfacial tension. This simple thermodynamic model

for the adsorption of small particles has been shown to describe the qualitative

features of bacterial adsorption (Absolom et al. 1983). This further illustrates

that bacterial adsorption can be viewed in thermodynamic terms and related to

colloidal adsorption systems.

Statistical analysis of data taken over a 320C temperature range suggests

little or no effect of temperature on the actual number of bacteria bound.

Ohyama et al. (1979) also found little effect of temperature (240 to 350C) on

the number of A. tumefaciens cells bound to plant suspension culture cells.










Similarly, agglutination of asparagus cells by A. tumefaciens appeared to be

independent of temperature (Draper et al. 1983). This would suggest that an

ionic interaction is one of the forces involved in the adsorption process, a

characteristic of physisorption.

Since physisorption appears to be operative in this system,

characterization of the ionic nature of the bacterial surface indicate a pl of 2.9

in distilled water. Harden and Harris (1953) gathered data on the isoelectric

point of 21 species of Gram-negative bacteria and found pl values ranged from

2.07 to 3.65. At pH values below the pl, the A. tumefaciens surface exhibits a

small net positive charge. This may be due to the fact that few ionizable

amino groups are present on the bacterial surface. At pH values above the pl, a

large number of carboxyl and/or phosphate groups may be ionized and account

for the large net negative charge. These data would then suggest that a very

high ratio of carboxyl or phosphate to amino groups is present on the bacterial

surface.

Since the pH of the sap in potato tuber wounds is 5 (unpublished results),

which is above the isoelectric point of A. tumefaciens, the bacterial cell is

most likely negatively charged within the wound site. However, there are a

large number of components within the wound which, when adsorbed onto the

bacterial surface, may alter the bacterial charge characteristics. The possible

role of these surface charges in adsorption remains to be defined.

At inoculum densities that fall within the linear range of the A.

tumefaciens adsorption isotherm, 105 to 109 bacteria/mi, a relatively constant

1.2% of the inoculum is adsorbed from the inoculum (Chapter Three this

dissertation). This suggests that only a small percentage of the bacterial

population may be competent for adsorption. The sequential adsorption










experiments represented in Fig. 7 do not lend support to this hypothesis because

similar numbers of A. tumefaciens cells bound to all three groups of discs.

Mathysse and Gurlitz (1982) similarly found no change in the number of

bacteria from the bacterial suspension when a 10-fold increase in plant cells

was used.

Lippincott and Lippincott (1969) have proposed a specific site-attachment

hypothesis to describe the adsorption of A. tumefaciens to cells of pinto bean

leaves. When the virulent A. tumefaciens strain B6 was added to the wound

prior to inoculation with the avirulent IIBNV6 strain, tumor number was not

reduced. However, when the order of inoculation was reversed, a reduction in

tumor number was noted. This hypothesis suggests a very specific irreversible

adsorption of A. tumefaciens strain B6 to pinto bean leaf tissue. The data

presented here suggest that A. tumefaciens adsorption is a nonspecific

phenomenon governed by physisorption forces. If adsorption was site-specific

and irreversible, equilibrium binding as is shown in Fig. 6 would not be

expected. Reduced or elevated temperatures also would be expected to

strongly influence an irreversable specific site-attachment.

However, there may be two separate types of binding sites. One could be

site-specific and adsorb bacteria as a prelude to tumor formation. Other sites

may nonspecifically adsorb bacteria and not lead to tumorigenesis. Since

complete desorption of all bacteria from the potato surface during the 6 hr

desorption period at 60C was not obtained, the possible existence of two types

of adsorption sites on potato tuber tissue varying in their affinity for

Agrobacterium spp. cannot be eliminated. However, I did not obtain any direct

evidence to support the existence of more than one type of adsorption site.














CHAPTER FIVE
EFFECTS OF AGENTS THAT ALTER IONIC AND HYDROPHOBIC
INTERACTIONS ON AGROBACTERIUM TUMEFACIENS ADSORPTION


Introduction



The induction of tumors by A. tumefociens is thought to require the site-

specific attachment of the bacterium to a plant wound site (Lippincott and

Lippincott 1969). Specific attachment was inferred from studies in which

plants were inoculated with an virulent A. tumefaciens strain prior to or

coincident with a virulent A. tumefociens strain. Changes in tumor number

indicated that a single virulent bacterium attached to a single host site, and

that such attachment resulted in a single tumor (Lippincott and Lippincott

1969, Whatley et al. 1977). If an virulent cell occupied the site, the virulent

cell was excluded, and a tumor did not form. Unfortunately, attachment was

not quantified in these studies.

Various host and bacterial molecular components have been examined for

their possible role in site-attachment. When bacterial suspensions were mixed

with isolated plant cell wall fractions, tumor number was reduced (Lippincott

and Lippincott 1977). Several off-the-shelf compounds were tested, and

galacturonans were found to inhibit tumorigenesis dramatically. Galacturonans

are thought to inhibit tumor formation by binding to the bacterial receptor that

mediates attachment to the host. These results led to the hypothesis that a

pectic component in the host's cell wall contains the specific attachment site

for Agrobacterium sp. Crude lipopolysaccharides (LPS) and bacterial cell










envelope preparations from A. tumefaciens have been examined for their

participation in specific-site attachment (Whatley et al. 1976). Plant tissue

pretreated with the o-antigen region of the LPS reduces the ability of virulent

bacteria to form tumors and thus is thought to mediate bacterial attachment to

some pectic component in the plant cell wall.

Pueppke and Benny (1984) directly measured adsorption of A. tumefaciens

to potato tissue. Adsorption was not reduced by galacturonans and in fact was

enhanced by citrus pectin. Cell-free LPS from all but one strain did not

influence bacterial adsorption. Additionally, LPS did not influence tumor

formation in the potato tuber disc assay system (Pueppke and Benny 1983).

The heat of adsorption for A. tumefaciens attachment to plant cells is

very small, an observation consistent with physisorption (Chapter Four this

dissertation). Physisorption implies that adsorption is nonspecific and mediated

by ionic and/or hydrophobic forces. Here I provided evidence to support the

hypothesized nonspecific nature of A. tumefaciens adsorption. The

involvement of ionic and hydrophobic forces in the adsorption process also was

tested directly. The ionic nature of the bacterial surface was affected by

changes in pH and divalent cations. The possible role of hydrophobic forces and

bacterial proteins in adsorption also was tested.



Materials and Methods



Maintenance of Bacterial Strains


The sources of the Agrobacterium spp. and other bacteria used in this

study are listed in Table 4. All of the strains were maintained on slants of the

defined gluconate-mannitol medium of Bhuvaneswari et al. (1977)















Table 4. Sources of Characteristics of Bacterial Strains.


Dathogenic
Strain Source0 Type Ti-plosmid on Potato


A. tumefaciens

B2F Anand Wild type
B6 Matthysse Wild type
B6M Anand Laboratory substrain
of B6
CV3132 De Block Tn7 insertion mutant
of pTi C58
178 Anand Wild type
ATCC 15955 Curley Wild type
B6-95 Gurley Avirulent B6 mutant +
B2R Pueppke Spontaneous variant
of 82F
NTI-86 Gurley NTI transformed with
pTi 66
ACH5 Lippincott Wild type
C58 Lippincott Wild type
ACHSC3 Lippincott Heat-cured ACMS
NT I Lippincott Heat-cured C58
IIBNV6 Anand Avirulent deviation of
11BV7
NTI-95 Gurley NT transformed with
pTi B6-95
Rhizobium jaoonicum

USDA 201 Keyser Fast grower
nodulates soybean

Xanthomonas vesicatoria

71-21 Stail WiTd type

A. radiobactor

ATCC 4718 Anand Wild type
51005 Gurley Wild type
K84 Lippincott Wild type, Australia

A. rhizogenes

15834 Lippincatt 'ild type Ri-plasmid

A rubi

ATCC 13335 Anond Wild type


a Sources of strains: Dr. V. K. Anond, Department of Biology, University of Missouri, St. Louis, MO
63121 U.S.A.: Dr. N. S. Gurley, Deornment of Microbiology and Cell Science, LUnversitv of Flori:.
Gainesivlle, Fi 32611 U.S.A.; Dr. J. A. Lippincott, Deoarfment of Biological Sciences, Northwestern
University, Evanston, IL 60201. U.S.A.; Dr. Ann G. Matthysse, Department of Botany, University of
North Carolina, Chapel Hill. NC 27514 U.S.D.A.; Dr. S. G. Pueppke, and Dr. R. E. Stall Depar-ment of
Plant Pathology, University of Florida, Gainesville, FI 32611 U.S.A.; Dr. M. De Bloc,
Rijksuniversitelt Gent, Ledegancstraat 3S, B-900 Gent, Belgium,










Bacterial Attachment Assay


One ml of buffered gluconate-mannitol medium (Meyer and Pueppke

1980), containing 50 uCi of 35S-L-methionine, was inoculated with I to 2 X 107

bacterial cells. The culture was incubated for 15 to 16 hr on a rotary shaker at

125 rpm and 270C. The bacteria were harvested by centrifugation at 10,000 X

G for 10 min. The bacterial pellet then was washed three times with PBS.

Specific activities ranged from 20 to 150 bacteria per cpm. The final bacterial

concentration was adjusted turbidimetrically to 2 X 107 cells/10 ul of PBS.

Adsorption was measured by the potato tuber disc assay (Pueppke and

Benny 1984) which is described here briefly. Potato tubers (Solanum tuberosum

L. cultivar Red LaSoda or Red LaRouge) were washed, peeled, surface

sterilized in a 20% solution of Clorox for 15 min and rinsed copiously with

sterile water. A 8-mm diameter cork borer was used to extract a core of

tissue, which was cut into 2-to 3-mm discs. The discs were blotted dry on filter

paper and placed on a 10-ul droplet containing 2 X 107 radioisotope-labeled

bacterial cells. After a 10 min adsorption period, the discs were washed with

three 5-ml portions of PBS. Individual discs were transferred to glass

scintillation vials, dried overnight, and digested with perchloric acid and

hydrogen peroxide. Five ml of Amersham ACS II Scintillation fluid were added

to each vial, and the radioactivity was determined with a Beckman LS-50 liquid

scintillation spectrometer. The number of bound bacteria was estimated from

the measured radioactivity.










Adsorption of A. tumefaciens to Monocot Tissue


Three monocot tissues were examined: I) rhizomes of Smilax bona-nox L.

(Liliaceae), 2) aerial tubers of Dioscorea bulbifora L. (Dioscoreaceae) and 3)

corm tissue of Gladiolus tristis L. (Iridaceae). Each of these monocot tissues

was prepared for the adsorption assay as described above. Dioscorea bulbifera

and the Smilax bona-nox were collected near Gainesville, Fl. Gladiolus tristis

corm tissue was kindly provided by Beth Logan, University of Florida.



Zeta Potential Measurements


Bacteria were washed twice with water, centrifuged at 7,700 X G for 10

min, and resuspended in one of the following buffers: 50 mM acetate buffer

containing 7.2 g of NaCI/I (final pH 4.5), PBS (pH 7.2), or 50 mM borate buffer

containing 7.2 g of NaCI/I (final pH 9.5). Measurements were made as

described previously (Chapter Four this dissertation) with a Laser Zee Meter

Model 502, set at 75 V. The zeta potential of the bacteria in the upper

stationary phase was recorded.



Negative Staining for Electron Microscopy


Formvar-coated grids were sputter coated with carbon and placed on a

bacterial colony growing on agar or floated on a bacterial suspension. After 30

min the grids were removed and washed thoroughly with water. Grids with

adsorbed bacteria were stained with 1% phosphotungstic acid for 10 min and

excess stain was removed by washing with water. The grids were dried and

then viewed with a Hitachi-600 electron microscope.










Proteolytic Treatment of Bacteria


Radioisotope-labeled cells of A. tumefaciens strain ACH5 were adjusted

to 1.6 X 109 cells/ml of PBS. A 12.5-ul droplet containing 2 X 107 cells was

placed on a hydrophobic petri plate (Fisher No. 8-757-12), to which was added

12.5 ul of protease (Streptomyces griseus, Sigma Chem. Co.) at I mg/ml or

trypsin (Bovine pancreas, Sigma Chem. Co.) at 250 mg/ml and incubated at

room temperature for a predetermined time period. A potato disc (9-mm diam,

3-mm thick) then was placed onto the 25-ul droplet for a 10 min adsorption

period. The number of bacterial cells bound to the surface was determined as

described above. In replicate experiments, the disc size was reduced to 8 mm

in diameter and the volume of the droplet was reduced to 20 ul.



Influence of CaCI2 on Adsorption


A suspension of radioisotope-labeled bacterial cells was divided into two

equal aliquots, each of which was washed separately. One aliquot was washed

three times with 0.15 M NaCI (pH 7.2) and the other with a solution of 0.15 M

NaCI (pH 7.2) containing 7.8 mM or 15.6 mM CaCI2. The CaCI2 treated

suspension was adjusted to 2 X 107 cells/10 ul in the CaCI2-NaCI solution and

used for inoculation as described above. The control suspension was adjusted to

2 X 107 cells/10 ul in the NaCI solution. After the 10 min adsorption period,

the discs were washed three times in the buffer originally used to wash the

bacteria. Seven separate experiments were conducted with 10 replications in

each.










Influence of a Calcium Chelating Agent on Adsorption


Potato discs were incubated in 3 mM aqueous ethylene-

bis(oxyethylenenitrile) tetraacetic acid (EGTA) for 15 min at room

temperature, after which the discs were removed and blotted dry on filter

paper. Control potato discs were incubated in water for 15 min. The potato

discs then were placed on a 10-ul droplet containing 2 X 107 radioisotope-

labeled bacteria that had been washed five times with deionized water. After

10 min, the discs were treated as described above, and the number of bound

bacteria determined. The experiment was repeated five times, with each

experiment containing 10 replications.



Agglutination of Bacteria by Divalent Cations


Bacteria from 2-day-old cultures on glutamate-mannitol slants were

washed twice, suspended in PBS, and adjusted to a concentration of 4.4 X 109

cells/ml. Fifty ul of a I M solution of FeSO4, CaCI2, or MgSO4 were serially

diluted in a two-fold series to a titer of 12 in a polystyrene V-type microtiter

plate (Dyntech Laboratories, Alexandria, Va.). Twenty-five ul of PBS

containing I.I X 108 cells were added to each well such that the final volume

was 75 ul. The final concentration of Fe2+, Ca2+, or Mg2+ ranged from 670

mM to 0.33 mM. The microtiter plate was agitated gently and then left

undisturbed for 3 hr at room temperature. The degree of agglutination was

determined visually. The experiment was repeated twice, each with two

replications.









Influence of Detergents on Adsorption


Bacteria were radioisotope-labeled as described above and washed four

times with half-strength PBS. The bacterial suspension then was divided into

four equal portions and centrifuged at 10,000 X G for 10 min. One of the four

portions was resuspended in half-strength PBS (control) and the other portions

resuspended in a half-strength PBS solution of the detergent to be tested.

Three detergents were examined; sodium dodecyl sulfate (SDS) (anionic),

hexadecyltrimethylammonium bromide cationicc), and Tween-20 (nonionic).

The bacteria were incubated at room temperature in the detergent for 15 min,

after which the detergent solution containing the bacteria was adjusted to a

bacterial concentration of 2 X 107 cells/10 ul. The bacteria then were used as

inoculum in the adsorption assay as described above. This experiment was

conducted twice with 10 replications within each experiment.



Adsorption of A. tumefaciens to Hydrophobic Polystyrene


Bacterial adsorption to polystyrene was measured by three techniques.

The first technique has been described previously (Rosenberg 1981), and is

briefly outlined here. The bacteria to be examined were streaked onto

glutamate-mannitol agar and incubated for 2 to 3 days at 270C. A flat

hydrophobic polystyrene disc, cut from the lid of a petri plate, was pressed

against the surface of the bacterial colonies. After 5 to 6 min the plastic disc

was removed and vigorously washed under a stream of tap water for

approximately 2 min. The bacteria on the plastic surface were fixed by

immersing the discs in methanol for 30 sec, after which they were air-dried,

and stained with Gentian violet. A hair dryer then was used to dry the stained









plastic surface, containing the adsorbed bacteria. The plastic disc was

immersed in tap water to selectively remove the stain from the plastic surface

leaving the stained adsorbed bacteria. The intensity of staining was examined

visually and the discs given a qualitative rating of ++ = high level of adsorption,

+ = medium level of adsorption, and = no observable adsorption.

In the second technique, bacterial cells were radioisotope-labeled, washed

as described previously, and adjusted to a final concentration of 2 X 107

bacteria/10-ul droplet of PBS. Discs (8 mm diam) were cut from the bottom of

hydrophobic petri plates, and the discs were marked to ensure that the

bacterial suspension would be applied to the untouched inside surface of the

plastic disc. Ten ul of the bacterial suspension were placed on each plastic

disc. After a 10 min adsorption period, discs were transferred using forceps to a

beaker of water. The discs were agitated for 10 sec to remove both non- and

loosely-adsorbed cells. Individual plastic discs were placed in scintillation

vials, which contained 0.5 ml of a 4% aqueous solution of SDS. The vials were

shaken to remove the adsorbed bacteria. Five ml of scintillation fluid were

added to the vials, and radioactivity was determined.

To monitor the effect of pH on the adsorption of radioisotope-labeled

bacterial cells onto hydrophobic polystyrene surfaces, the bacteria were washed

and resuspended in either 50 mM acetate buffer containing 7.2 g of NaCI/I, (pH

4.5), PBS (pH 7.2), or 50 mM carbonate/bicarbonate buffer containing 7.2 g of

NaCI/I (pH 9.5). The adsorption assay was carried out as described above. The

experiment was repeated twice with 10 replications in each experiment.


The Effect of pH on Adsorption to Hydrophobic and Hydrophilic Plastic


A third adsorption technique was used to study the effect of pH on

adsorption to both hydrophobic and hydrophilic polystyrene (Fletcher 1977).










Bacteria were washed three times with water, centrifuged at 8,000 X G for 10

min, resuspended in either acetate buffer (pH 4.5) PBS, (pH 7.2), or borate

buffer (pH 9.5), and adjusted to a concentration of 5 X 108 bacteria/ml.

Twenty ml of the bacterial suspension were poured into either a hydrophobic

(Fisher No. 8-757-12) or hydrophilic (Corning No. 25020) 9-cm polystyrene petri

plate. After 2 hr at room temperature the suspension was decanted, and the

plates were gently washed with 30 ml of distilled water. The adsorbed bacteria

then were fixed to the plastic surface by drying the surface thoroughly with a

hair dryer. A freshly filtered solution of crystal violet solution (Conn 1940) was

added until the bottom of the petri dish was covered. After 5 min the stain was

decanted and the plates were slowly passed through a stream of running tap

water five times to remove the excess stain. The plates were dried thoroughly

with a hair dryer. Each petri plate was divided into eight equal sections by

marking the edge with a felt pen. Absorbance values for each of the eight

sections on the petri plate were obtained at 590 nm with Beckman ACTA CII

spectrophotometer.



Results



Both monocot and dicot tissues are capable of adsorbing large numbers of

cells of A. tumefaciens strain ACH5 (Table 5). The number of bacteria

attached to S. bona-nox tissue after the 10 min adsorption period was

significantly greater (P = 0.01) than that to D. bulfifera, G. tristis or S.

tuberosum tissue. There was no significant difference in bacterial adsorption

to D. bulbifera, G. tristis or S. tuberosum tissue (Table 5).










There was no significant difference in the capacity of ACH5 cells and any

of the seven other bacterial species to bind to potato tissue (Table 6). However

a significant difference does exist between two A. tumefaciens strains

(ACH5C3 and IIBNV6) and A. radiobacter strain 4718 (Table 6). Included in the

list is a fast-growing Rhizobium japonicum strain and Xanthomonas

vesicatoria. Neither of these organisms is symbiotic or pathogenic on potato.

A. tumefaciens GV3132, A. radiobacter 4718, and A. rhizogenes 15834 have

been designated as non site-binding virulent strains (Lippincott and Lippincott

1969). Their attachment to potato tuber tissue, however, was similar to that of

the virulent site-binding strain ACHS.



Influence of Proteolytic Treatment of Bacteria on Adsorption


Preincubation of bacteria with protease for 30 min resulted in a

statistically significant increase in the number of ACH5 cells bound relative to

controls (Table 7). Trypsin treatment had no significant effect. A.

tumefaciens B6 cells pretreated with either trypsin or protease bound in

significantly greater numbers than did cells in the control.



Influence of Detergents on Adsorption


All three detergents influenced bacterial adsorption to potato tissue, but

the ionic character of the detergent does not appear to be an important

factor. Tween-20 at concentrations of 7%, 0.4%, and 0.04% (v/v) inhibited














Table 5. Adsorption of Aqrobacterium tumefaciens ACH5 to Monocot and
Dicot Tissue.


Number of Bactpria Bound
Plant Species per Disc (X 10')*


Smilax bona-nox 86.0 + 18.2 a Monocot
Dioscorea bulbifera 18.3 + 09.0 b Monocot
Gladiolus tristis 16.0 + 06.7 b Monocot
Solanum tuberosum 8.3 + 02.3 b Dicot


The values are means (+1 standard deviation) of three experiments, each of
which contained 10 replications.
*Numbers followed by the same letter are not significantly different (P = 0.05)
according to Duncan's Multiple Range Test.











Table 6. Adsorption of Bacteria to Potato Tuber Tissue


Bacterial Species


Average Number of Eacteria
Bound per Disc (X 10J)


Agrobacterium tumefaciens ACH5 5.10 + 2.8 ab*
Aqrobacterium tumefaciens ACHSC3 9.76 + 1.7 a
Agrobacterium tumefaciens IIBNV6 8.98 + 1.6 a
Agrobacterium tumefaciens GV3132 6.41 + 1.0 ab
Agrobacterium rhizogenes 15834 4.40 + 2.1 ab
Aqrobacterium radiobacter 4718 1.70 +0.4 b
Rhizabium japonicum USDA 201 7.85+ 1.5 ab
Xanthomonas vesicatoria 71-21 6.88 + 2.6 ab



The values are means (+1 standard deviation) of two experiments, each of which
contained 10 replications.
*Numbers followed by the same letter are not significantly different (P = 0.05)
according to Duncan's Multiple Range Test.

























o 0 --
o C
(N













0 .0 m




























0 .0 .0
Co a














oa 0 -
00 .5
















a) 0,
O C >-
0- 1


C)
c

0
C

0
c




0(
C

o
0
01
c
a)







n





0
o
u
o























.0
C
L-0






















-E0
o*

a.


0,





01

Sac
Q-











11 .
LC
^1





LU* Q









bacterial adsorption by 56%, 41% and 20%, respectively. In the presence of

0.05% Tween-20, adsorption was reduced from 2.8 + 0.4 X 106 to 1.87 + 0.4 X

106 cells/disc, a 35% reduction. Adsorption was reduced from 2.80 + 0.5 X 106

cells bound/disc in the control to 0.95 + 0.3 X 106 cells bound/disc by 0.05%

SDS. This is a 66% reduction in the number of cells bound. Adsorption was

reduced by 35%, from 2.8 + 0.5 X 106 to 1.82 + 0.27 X 106 bacteria per disc, in

the presence of 0.5% hexadecyltrimethylammonium bromide as compared to

the control. Tween-20 and SDS did not reduce bacterial colony forming units

(CFU) at the concentrations studied, whereas 0.05% hexadecyltrimethyl-

ammonium bromide reduced bacterial CFU.



Adsorption of Agrobacterium spp. to Hydrophobic Polystyrene


Ten strains of A. tumefaciens, one A. rhizogenes strain, two A.

radiobacter strains, and one A. rubi strain were screened semiquantitatively for

their ability to adsorb to hydrophobic polystyrene (Rosenberg 1981). Both

virulent and avirulent Agrobacterium spp. adsorb to hydrophobic polystyrene

(Table 8). Significantly, strain B2R, a spontaneous mutant of B2F, is able to

adsorb to hydrophobic plastic, but B2F is not.


Influence of pH on Bacterial Adsorption to Plastic


There was no significant effect of pH on adsorption of 35S-labeled A.

tumefaciens ACH5 to hydrophobic plastic. The numbers of cells bound per disc

were 1.3 + 0.45 X 106, 1.2 + 0.6 X 106, 0.95 + 0.3 X 106 at pH 4.5, 7.2, 9.5,

respectively. When measured photometrically, the adsorption to hydrophobic












Table 8. Adsorption of Agrobocterium spp. to Hydrophobic Polystyrene.


Agrobocterium spp.


A. tumefaciens











A. rhizogenes

A. radiobacter

A. rubi


Adsorption*


B6
B6M
C58
ACH5
B2R
1400
15955
B6 95
NTI
B2F
NT 1(95)


15834

S1005

13335


*Adsorption is based on a visual, semi-quantitative rating:
++ = high level of adsorption, + = medium level, and = no visually
observable adsorption







83















.125 r 0- HYDROPHILIC

-- HYOROPHOBIC


.10 -




.075




.05 -




.025




0/"
4-5 7-2 9-5
pH



Fig. 9. Adsorption of Agrobacterium tumefaciens strain ACH5 to
hydrophilic or hydrophobic polystyrene as a function of pH. Bars
represent +1 standard deviation.









plastic also was not significantly influenced by pH (Fig. 9). Adsorption to

hydrophilic plastic, however, was significantly greater at pH 4.5 than at pH 7.2

or pH 9.5 (Fig. 9).

Adsorption of A. tumefaciens GV 3132 to potato tuber discs was optimal

at pH 7.2 (6.5 + 0.95 X 105 bacteria/disc) and significantly reduced at both pH

4.5 (1.72 + 0.38 X 105 bacteria/disc) and pH 8.5 (2.1 + 0.75 X 105 bacteria/disc)

according to the Student's t-test.



Effect of EGTA Treatment


Pretreatment of potato discs with 3 mM EGTA did not influence their

ability to adsorb radioisotope-labeled bacteria. Discs treated with EGTA were

compared with PBS control discs and analyzed using the Student's t-test. One-

sided tests at the P = 0.05 level indicate that treatment of the potato discs

with EGTA does not have a significant effect on bacterial adsorption. In these

experiments pretreated EGTA discs bound 1.41 + 0.56 X 106 bacteria per disc,

and control discs bound 1.54 + 0.80 X 106 bacteria disc (+1 standard deviation).



Effect of Divalent Cations on Agglutination


A. tumefaciens ACH5 cells are agglutinated by CaCI2. The dilution end

point is 5.2 mM. The greatest agglutination of ACH5 occurred at a CaCI2

concentration of 83 mM. FeSO4 exhibited a tremendous ability to agglutinate

A. tumefaciens. Agglutination was evident in the last well, which contained

0.33 mM FeSO4. The greatest degree of agglutination was found in the

presence of 5.2 mM FeSO4.









Even after 6 hr in the presence of I M MgCI2, agglutination of ACH5 was

slight. However, in the presence of either CoCI2 or FeSO4 agglutination was

very rapid and nearly complete within I hr, and stable overnight at room

temperature. Control bacteria suspended in either water or PBS did not

agglutinate.



Effect of CaCI2 on Adsorption to Plant Tissue


Binding of strain ACH5 to potato tissue was not appreciably affected by

7.8 mM or 15.6 mM CaCI2 (Table 9). Comparison of CaCI2-treated to non-

treated controls by the Student's t-test showed no significant difference in the

number of bacteria bound. However, in six of the seven experiments, potato

discs consistently bound fewer CoCI2-treated bacteria. The CaCI2

concentrations used in these experiments fall near the dilution end point of

agglutination. Therefore, agglutination of the bacteria during washing and the

subsequent binding period should have been minimal.



Zeta Potential Measurements


The zeta potential of bacteria suspended in the borate buffer (pH 9.5) is

-I I mV. Bacteria suspended in PBS have a zeta potential of approximately -2.5

mV. In acetate buffer (pH 4.5) the zeta potential is slightly positive with a

value of +0.1 mV. It should be emphasized that zeta potential determinations

are more variable at or near the isoelectric point of the particle. Therefore











Table 9. Effect of CaCI2 on Agrobacterium tumefaciens Adsorption.


Cells bound (X 106) Statistical

Experiment CaCI2 (mM) CaCI2 Control Significance


I 7.8 1.40 + 0.63 1.80 + 0.49 NS
2 7.8 1.15 + 0.25 1.30 + 0.74 NS
3 7.8 1.60 + 0.80 1.98 + 0.40 NS
4 15.6 0.97 + 0.10 0.80+ 0.20 NS
5 15.6 1.20 + 0.38 1.50 + 0.39 NS
6 15.6 0.70+ 0.20 1.00+ 0.19 NS
7 15.6 0.86 + 0.26 0.90 + 0.22 NS


All adsorption experiments were carried out with an initial inoculum of 2 X 107
bacteria/ml.
Statistical significance was determined with the Student's t-test. (P = 0.05).









the values given for measurements at pH 4.5 and pH 7.2 are not necessarily the

absolute zeta potentials of the bacteria suspended in either of those buffers.



Negative Staining


The result of negative staining cells of strain ACH5 with phosphotungstic

acid is shown in Fig. 10. Bacteria from either colonies on agar plates or

suspension cultures gave similar results (Fig. 10).



Discussion


The transfer of the T-DNA from Agrobacterium spp. into the host cell

most likely requires the attachment of the bacterium to the plant cell

surface. The bacterial o-antigen and the pectic materials in the host cell wall

have been hypothesized to mediate such bacterial adsorption (Lippincott and

Lippincott 1969, Whatley et al. 1976, Lippincott et al. 1977b, Raa et al. 1982).

These data, however, should be viewed with caution, because bacterial

adsorption was not measured directly (Lippincott and Lippincott 1980).

Pueppke and Benny (1981, 1983) have shown that extrapolation from tumor

formation to bacterial adsorption may not be justified. Additionally, Pueppke

and Benny (1983) used the potato tuber disc assay to show that neither tumor

formation nor A. tumefaciens adsorption was inhibited by cell-free LPS. In

contrast, inoculation of pinto bean leaves in the presence of cell-free LPS

substantially reduced tumor formation (Whatley et al. 1976).

The site attachment hypothesis predicts that the relative resistance of

most monocotyledonous plants to Agrobacterium spp. is a consequence of their












































'A1







Fig. 10. Negatively stained Aqrobacterium tumefaciens strain ACH5.
X 10,000.










inability to adsorb Agrobacterium spp. (Lippincott and Lippincott 1969,

Lippincott et al. 1977b, Lippincott and Lippincott 1978a). However, adsorption

of virulent, site-binding A. tumefaciens cells to S. tuberosum, was not

significantly different (P = 0.05) than that to two monocots, D. bulbifera, and

G. tristis. However, the number of bacteria bound to the S. bona-nox tissue

was significantly greater than that bound to any of the other species. Ohyama

et al. (1979) have shown that suspension culture cells of a variety of monocots

adsorbed large numbers of A. tumefaciens cells. In addition, suspension culture

cells of soybean (Glycine max) bound a large number of A. tumefaciens cells,

even though soybean apparently is resistant to A. tumefaciens infection (De

Cleene and De Ley 1976). Draper et al. (1983) have shown that suspensions of

cells of another monocot, asparagus, are agglutinated by A. tumefaciens cells.

A series of virulent and avirulent Agrobacterium spp. adsorb in

statistically similar numbers to potato tuber tissue. The adsorption of all of

these species is similar to that of A. tumefaciens ACH5 (Table 6).

Additionally, the adsorption of Rhizobium japonicum USDA 201 and

Xanthomonas vesicatoria 71-21 were not significantly different from that of

strain ACH5 (Table 6).

Divalent calcium did not statistically alter the number of bacteria bound

to wounded tissues. Additionally, when discs were treated with the calcium

chelating agent EGTA prior to bacterial inoculation, no alteration of bacterial

adsorption was observed. This suggests that the availability of Ca2+ is not

important to bacterial adsorption. Ohyama et al. (1979) reported somewhat

similar results using suspension cultured cells.

In light of a report suggesting that LPS and galacturonans do not

participate in A. tumefaciens adsorption to susceptible potato tissue (Pueppke










and Benny 1984), the actual molecular components of A. tumefaciens

adsorption remain unclear. A. tumefaciens adsorption, however, may be

described in terms of physical adsorption, which is governed by ionic and

hydrophobic forces (Chapters Three and Four this dissertation). The adsorption

of A. tumefaciens ACH5 to hydrophobic plastic was not altered as a function of

pH. However, adsorption to hydrophilic plastic was optimal at pH 4.5, and

decreased with increasing pH (Fig. 9). When measured as a function of pH,

adsorption of A. tumefaciens GV3132 to potato tuber tissue was optimal at pH

7.2, and significantly decreased at either acidic or basic pH. Similarly,

Pueppke and Benny (1984) observed that the adsorption of A. tumefaciens

ACH5 to potato tuber tissue was optimal at pH 7.2 and reduced at either pH 4.5

or 9.5. Zeta potential measurements of the bacteria suspended in the same

buffers, show that the surfaces of the bacterial cells are highly negatively

charged at pH 9.5 (borate), are neutral to slightly negative at pH 7.2 (PBS), and

are slightly positive at pH 4.5 (acetate buffer).

As the pH approaches the pl of A. tumefaciens cells, a reduced surface

charge may facilitate a closer approach to the host's surface. Attachment then

may be facilitated by cellular appendages, such as the flagellae shown in Fig.

10. These could help overcome the remaining electrostatic repulsion barrier

between the bacterium and the plant cell (Ottow 1975). Matthysse et al. (1982)

illustrated the adsorption of A. tumefaciens in large numbers to the cell

membranes of tobacco and carrot protoplasts that were devoid of cell wall

material. The importance of hydrophobic interactions in the adsorption process

that leads to tumor formation remains to be determined. It is clear, however,

that not all Agrobacterium spp. adsorb to hydrophobic plastic similarly, as

measured according to Rosenberg (1981) (Table 8), and yet are able to adsorb to

potato tuber tissue (Table 6).









The zeta potential of bacteria suspended in deionized water at pH 4 or

greater is highly negative (Chapter Four this dissertation), yet in acetate buffer

at pH 4.5, the zeta potential is slightly positive. The descrepancy can be

explained as follows. Bacteria suspended in the acetate buffer pH 4.5 adsorb

cations from the solution onto their surfaces. This alters their behavior in an

electrical potential and results in a less negative zeta potential. The effect is

even more dramatic with bacteria suspended in PBS at pH 7.2. In water (low

ionic strength) the bacteria normally would be negatively charged at pH 7.2.

However, in PBS the negative bacterial surfaces adsorb cations from the

solution, and this substantially reduces the negative zeta potential.

Outer membrane proteins exposed to the environment mediate

hydrophobic interactions in several mammalian systems (Berkeley et al. 1980,

Bitton and Marshall 1980). Pretreatment of A. tumefaciens with either trypsin

or protease to digest exposed protein components resulted in increased numbers

of bacteria bound. Consequently, the presence of intact outer proteins may in

some way inhibit adsorption. Treatment of the bacteria with proteolytic

enzymes presumably degrades external protein components, outer membrane

proteins, pili, and flagellae, all of which may allow a closer approach of the

bacterium to the host's cell surface. At this point the much shorter range of

ionic forces comes into play, perhaps accounting for the increased adsorption

observed after proteolytic digestion (Table 7). Alternatively, the enzymes

themselves may adsorb tenaciously to the bacterial surface, enhancing the

adsorption capacity of the bacteria.

At a concentration of 0.05%, all three detergents substantially reduced

bacterial adsorption. The sensitivity of adsorption to the presence of

detergents suggests that hydrophobic interactions play a role in A. tumefaciens










adsorption to potato tissues. Pili are a likely external component of the

bacterium that could be involved in hydrophobic interactions. For example,

gonococcal pili have been shown to be very hydrophobic with approximately

46% of the amino acid resides non-polar. This causes the pili to exist in the

aggregated form in vitro. Characterization of Agrobacterium spp. pili and

examination of their involvement in the adsorption process will certainly

enhance our understanding of the attachment process.

These results, in addition to the fact that the genus Agrobacterium has

the largest host range of any known phytopathogenic prokaryote, do not lend

support to the site-attachment hypothesis, which describes a highly specific

adsorption of Agrobacterium spp. to the surface of plant tissue. Rather, a

nonspecific adsorption mediated by physical adsorption forces appears to be

operative in the potato tuber disc assay system. Adsorption to plant and plastic

surfaces is pH-sensitive. Anionic, cationic, and neutral detergents significantly

reduce the number of A. tumefaciens adsorbed to plant cell surfaces. These

results strongly indicate that both ionic and hydrophobic forces are involved in

the non-specific adsorption of A. tumefaciens to plant cell surfaces.














CHAPTER SIX
SCANNING ELECTRON MICROSCOPY OF POTATO TUBER TISSUE
INOCULATED WITH AGROBACTERIUM TUMEFACIENS


Introduction



The etiological agent of crown gall disease Agrobacterium tumefaciens,

causes a tumor-like growth on many dicotyledonous plants (De Cleene and De

Ley 1976). Once in a wound site on the plant, A. tumefaciens cells are thought

to adsorb to a surface component of the host cell. A piece of the bacterial Ti-

plasmid, known as the T-DNA, is subsequently transferred to the host, where it

is stably incorporated and expressed (Watson et al. 1975, Chilton et al. 1977,

Gurley et al. 1979, Thomashow et al. 1980). The method of DNA transfer,

however, is unknown.

Plant protoplasts have been transformed by both intact A. tumefaciens

cells and isolated Ti-plasmid (Marton et al. 1979, Wullems et al. 1981, Krens et

al. 1982, Hanold 1983). It is possible that a similar situation occurs in a plant

wound site, where the partial removal of cell wall material could expose the

intact cytoplasmic membrane. This may facilitate the transfer to T-DNA out

of the bacterium into the wounded host cell. A second possibility is that of

bacterial-plant conjugation (Roberts and Kerr 1974, Kerr 1975). A third

alternative is that intact bacteria penetrate viable plant cells and then transfer

T-DNA. This idea has been rejected by some (Stonier 1956, Schilperoort 1969)

and proposed as the essential step in tumorigenesis by others (Sigee et al. 1982).










The purpose of this research was to examine the infection of potato tissue

by A. tumefaciens at the ultrastructural level. An virulent, heat-cured strain

was included for comparison.



Materials and Methods



Bacterial Culture, Maintenance, and Inoculum Preparation


Agrobacterium tumefaciens strains ACH5, ACH5C3, and B6 were

maintained at 40C on slants of the defined gluconate-mannitol medium of

Bhuvaneswari et al. (1977). Inocula were prepared by streaking A. tumefaciens

cells onto gluconate-mannitol slants. After 2 to 3 days at 270C, the cells were

washed from the slants with sterile PBS, centrifuged for 10 min at 7700 X G,

and washed twice with PBS. The bacteria then were resuspended in PBS and

adjusted turbidimetrically to a concentration of 2 X 107 cells/ml.



Preparation of Protoplasts


Suspension culture cells derived from potato (Solanum tuberosum cv Red

LaSoda) tubers were maintained in Murashige and Skoog media (Gamborg and

Wetter 1975). Cells were harvested and suspended in an equilibration medium

consisting of B5 salts (Gamborg and Wetter 1975), 30 mM MES (2 N-

morpholino ethanesulfonic acid), and 300 mM sorbitol, adjusted to pH 5.6. This

suspension was incubated at 150 rpm at room temperature. After equilibration

for I hr, the cells were harvested by centrifugation at 1000 X G for 10 min and

resuspended in an enzyme mixture, where they were incubated for 4-6 hours at

room temperature at 150 rpm. The enzyme solution contained 250 mg of










cellulose (Calbiochem, La Jolla, CA), 50 mg of Macerase (Calbiochem), 250 mg

of Rhozyme (Rohm and Haas), and 50 mg of Driselase (Sigma) dissolved in 10 ml

of the equilibration medium used above. After enzyme digestion the cell

suspension was centrifuged at 30 X C rpm for 10 min and the protoplasts

removed from the surface of the enzyme solution. The protoplasts were then

washed three times with the equilibration medium.



Inoculation Procedure



Potato tuber discs were inoculated with A. tumefaciens as described by

Pueppke and Benny (1981). Peeled potato tubers are surface sterilized for 10

min, and 8-mm-diameter cores of tuber tissue are removed. Discs, 2 to 3 mm

thick, are sliced from the cores and immersed in a suspension containing 2 X

107 cells of strain ACH5 or ACH5C3/ml of PBS. A ratio of I ml of bacterial

suspension per disc was maintained. After 10 min the discs were removed,

rinsed for 2 or 3 seconds in sterile PBS, blotted dry, and placed on 1.5% water

agar plates and stored in the dark at 280C. At specified times after

inoculation, discs were removed for scanning electron microscopy or the

determination of bacterial colony forming units. The number of bacteria

colonizing the potato disc was determined as described previously (Chapter

Three this dissertation).

Isolated protoplasts were inoculated with A. tumefaciens strain B6 at a

concentration of 2 X 107 cells/ml. After incubation for 3 hr the inoculation

mixture was filtered through Whatman No. I filter paper, and the protoplasts

rinsed with two volumes of equilibration buffer. The filter paper, which then

contained the protoplasts and bound bacteria, was removed and prepared for

scanning electron microscopy.




Full Text

PAGE 1

CHARACTERIZATION OF AGROBACTERIUM TUMEFACIENS ADSORPTION TO POTATO TISSUE BY DANIEL ALBERT KLUEPFEL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS OF THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1984

PAGE 2

ACKNOWLEDGEMENTS The work completed here at the University of Florida was aided greatly by the support and guidance of many people. Heading the list to receive thanks is my committee chairman Dr. Steven G. Pueppke. His seemingly boundless patience, accurate guidance, and constant availability resulted in the cultivation and maturation of my ability to develop a truly scholarly approach to the art of scientific investigation. I thank you. I am also indebted for the help given by Dr. H. Aldrich, Dr. G. Erdos, and especially Dr. H. Berg in distinguishing fact from fiction when peering through the electron microscope. The time spent by John Ransdell and Steve Linda sharing their expertise with me is greatly appreciated. I also appreciate the critiques given my dissertation by the other members of my committee Dr. R. Stall, Dr. D. Mitchell, Dr. B. Gurley, and Dr. B. Moudgil. Finally, innumerable thanks go to my wife, Marjan, for her help with the statistical analysis, valuable discussions of the data, love and understanding. This work would not have been possible without the financial support of the University of Florida.

PAGE 3

TABLE OF CONTENTS PAGE ACKNOWLEDGEMENTS ii ABSTRACT v CHAPTER ONE INTRODUCTION I CHAPTER TWO LITERATURE REVIEW 3 General Adsorption Chemistry 3 Adsorption of Bacterial Invadors 8 Adsorption of Rhizobium spp. to Root Surfaces 10 Adsorption of Aqrobacterium tumefaciens to Plant Surfaces 14 CHAPTER THREE ADSORPTION OF AGROBACTERIUM TUMEFACIENS TO POTATO (SOLANUM TUBEROSUM) TISSUE 31 Introduction 31 Materials and Methods 33 Results 35 Discussion 44 CHAPTER FOUR THERMODYNAMICS OF AGROBACTERIUM TUMEFACIENS ADSORPTION TO PLANT SURFACES.... 49 Introduction 49 Materials and Methods 50 Results 53 Discussion 61 CHAPTER FIVE EFFECT OF AGENTS THAT ALTER IONIC AND HYDROPHOBIC INTERACTIONS ON AGROBACTERIUM TUMEFACIENS ADSORPTION ...67 Introduction 67 Materials and Methods 68 Results 76 Discussion 87 CHAPTER SIX SCANNING ELECTRON MICROSCOPY OF POTATO TUBER TISSUE INOCULATED WITH AGROBACTERIUM TUMEFACIENS 93 Introduction 93 Materials and Methods 94 Results 96 Discussion 113

PAGE 4

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 CHARACTERIZATION OF AGROBACTERIUM TUMEFACIENS ADSORPTION TO POTATO TISSUE By Daniel Albert Kluepfel August 1984 Chairman: Steven G. Pueppke Major Department: Plant Pathology The Freudlich adsorption isotherm accurately defines the adsorption of Aqrobacterium tumefaciens to potato tuber discs. This isotherm describes a system where the available binding sites are not saturated. The_A. tumefaciens isotherm was developed for inoculum densities that ranged from 10 ' to 10^ bacteria/ml of buffer. Adsorption is rapid with more than 2.8 X 10^ bacteria adsorbing per potato disc within 0.5 min. Aqrobacterium tumefaciens dissociates from the disc and reaches equilibrium within 3 to 4 hr at 6°C. The adsorption isotherm was examined over a more limited initial inoculum density at three different temperatures: 6°C, 28°C, 38°C. Analysis of these isotherm data with the Clausius-Clapeyron equation shows that the heat of adsorption is close to zero. These combined data suggest that the adsorption of _A. tumefaciens to potato tuber tissue is governed by physical adsorption. Optimum adsorption of A_. tumefaciens to potato tissue occurred at pH 7.2, and adsorption was reduced at either pH 4.5 or 9.5. Adsorption to hydrophobic

PAGE 5

plastic was insensitive to changes in pH. Conversely, the adsorption to hydrophilic plastic surfaces was optimum at pH 4.5 and decreased with increasing pH. The isoelectric point of A^. tumefaciens suspended in water is approximately pH 2.9, as determined by zeta potential measurements. Adsorption of _A. tumefaciens cells to potato tissue was greatly reduced by detergents at concentrations as low as 0.05% (v/v). Anionic, cationic and neutral detergents all were effective. The effect of agents that influence both hydrophobic and ionic forces is consistent with the proposed physisorption mechanism for A^. tumefaciens. Calcium chloride treatment of the bacteria or EGTA treatment of the potato tuber discs did not, however, significantly affect bacterial adsorption. When viewed with the scanning electron microscope, bacterial colonization of the potato surface was the same for both the virulent A. tumefaciens strain ACH5 and its heat-cured derivative ACH5C3. The only ultrastructural feature that distinguished tissues inoculated with the virulent strain from those inoculated with the avirulent strain was the appearance of tumors five days after inoculation with the former.

PAGE 6

CHAPTER ONE INTRODUCTION Microorganisms readily accumulate on both natural and man-made structures. They quickly colonize structures in aqueous environments and the surfaces of both plant and animal cells. There is evidence that in some mammalian systems adsorption of bacteria to the host cell surface is an important factor in pathogenicity (Arbuthnott and Smyth 1979, Berkeley et al. 1980, Bitton and Marshall 1980, Ofek and Beachey 1980). The importance of adsorption in plant-bacterial interactions also has been receiving considerable attention (Dazzo 1 980a, 1 980b, Pueppke 1 984a, 1 984b). The surfaces of most plants are colonized by a large number and variety of bacteria (Blakeman 1982, Foster and Bowen 1982, Suslow 1982). The presence of these bacteria, however, rarely leads to necrotic plant disease. Even more unusual is the establishment of a beneficial symbiotic relationship. However, when plant disease or beneficial symbiotic interactions do occur, the role of bacterial adsorption is not clear. Bacteria have the ability to adsorb to plant, animal, and inanimate surfaces quite tenaciously and, at times, in very large numbers. This phenomenon is of major ecological significance for bacteria. Microbial adsorption and colonization of surfaces provide bacteria with a favorable ecological niche where nutrient levels are high and protection occurs against unfavorable environmental factors (Bitton and Marshall 1980, Costerton et al.

PAGE 7

1981). The following literature review will examine some of the features associated with bacterial adsorption to solid surfaces. Salient examples from plant-bacterial systems will be discussed in terms of specificity, adsorption mechanisms, and the relationship between adsorption and the pathogenic or symbiotic responses.

PAGE 8

CHAPTER TWO LITERATURE REVIEW General Adsorption Chemistry A colloid is a suspension that is neither a true solution nor a complete two phase system. Colloidal particles are small enough to remain widely dispersed and in constant motion due to bombardment by solvent molecules. Particles in collodial systems characteristically remain suspended and settle very slowly, whereas they sediment very rapidly under centrifugal force. Colloids also cause light to scatter, a useful tool for particle size and concentration determinations. The stability of a colloid, i.e. the tendency of the particles to remain in a highly dispersed state, is partially governed by the electrical charge carried by the particles. Particles within a colloid often contain a large homogeneous charge. This charge causes the particles to repel one another, thereby helping to maintain the stability of the colloid (Saunders 1971, Lips and Jessup 1979). Bacteria fall within the size range of colloid particles and often exhibit the properties of a colloidal suspension (Marshall 1976). The adsorption of bacteria to solid surfaces involves many of the same forces of attraction and repulsion that operate in the adsorption of inert colloidal particles to solid surfaces (Daniels 1972, Marshall 1976, Berkeley et al. 1980, Bitton and Marshall 1980). Consequently, bacterial suspensions can be considered to be living colloidal systems. Because the bacterial surface is highly charged, the use of

PAGE 9

physical (nonbiological) colloidal adsorption systems as models can provide useful information concerning bacterial adsorption. However, it is important to remember that bacteria are living organisms capable of movement and altered activity in response to physiological stimuli. There are two basic types of adsorption of particles from a liquid phase onto a solid. These two processes, termed physisorption and chemisorption, differ in several fundamental respects (Brunauer 1943, Saunders 1971, Tompkins 1978). The most basic difference is the type of molecular force mediating the adsorption process. Van der Waals forces are primarily responsible for mediating physisorption. Chemisorption requires chemical reactions, i.e. formation of covalent bonds. Chemical bonds in chemisorption systems are much stronger than the forces that account for physisorption. The heat of adsorption, a measure of the bond strength, is on the order of 10 kcal in physisorption and greater than 20 kcal in chemisorption. The basic difference in the bond type is also thought to influence the specificity of adsorption. At low temperatures van der Waals mediated adsorption occurs between any two competent surfaces, but chemisorption depends on the affinity of the two adsorbing components for each other. The rate of adsorption also is a useful indication of the adsorption type. The rate of physisorption is extremely rapid, whereas chemisorption requires energy of activation for bond formation and, in general, is much slower. In some nonbiological systems, however, the heat of activation can be extremely small, and thus chemisorption can occur relatively rapidly. Physisorption onto a solid surface may be multi layered or monolayered, but chemisorption always occurs in a monolayer. The number of particles adsorbed in a physisorption system is a function of the particle concentration. In chemisorption the

PAGE 10

adsorption is a function of both the area of the solid surface and particle concentration. A final contrasting feature is that of desorption. The desorption of particles occurs readily in physisorption, but chemisorbed particles are only removed with some difficulty. A useful method of adsorption data analysis is the development of an adsorption isotherm. Adsorption isotherms are plots of the number of particles bound versus the number of particles remaining in the suspension at equilibrium. There are four or five basic types of adsorption isotherms (Brunauer 1943, Tompkins 1978). Although adsorption isotherms do not provide data on adsorption mechanisms, other analyses can be used to differentiate chemisorption from physisorption. Analysis of adsorption at various temperatures within a 20°C to 30°C range may suggest a possible mechanism. Because van der Waals forces are not drastically affected over a 30 degree temperature range spanning physiological temperatures, physisorption is largely independent of temperature. On the other hand, because chemisorption requires heat of activation, chemisorption will vary significantly as a function of temperature. Heats of adsorption also may be measured directly or with the use of the thermodynamic law expressed as the Clausius-Clapeyron equation (Brunauer 1943). These measurements in conjunction with the adsorption isotherm analysis provide a good description of the possible mechanism for the adsorption system under examination. Adsorption of Bacteria onto Solid Surfaces The adsorption of bacteria onto the solid surfaces of plant cells can be divided into three temporal stages (Fletcher et al. 1980). The first stage,

PAGE 11

simply termed adsorption, refers to events encompassing the initial approach and contact of the bacterium with the plant cell surface. The second step is one of anchoring the bacterium to a surface in a firm manner. This may be accomplished by the production of cellulose microfibrils or other polysaccharide matrices collectively known as the glycocalyx (Costerton et al. 1981). The third step, termed colonization, is a secondary phenomenon that describes the long term adsorbed state of the microorganism on the host surface. This includes continued bacterial growth and multiplication on the solid surface. When adsorption of the microbe to living surfaces is measured, consideration should be given to the time frame under examination. For example, the adsorption characteristics after 60 min may be quite different than those measured after 2 to 4 hr, at which time microbial multiplication or metabolism of extracellular polysaccharides may greatly complicate the analysis. Therefore, if determination of the mechanism of initial adsorption is desired, incubation times should be short, i.e. well below the first doubling time, and the physiological conditions should be held constant. The outer surface of both Gram-negative and Gram-positive microorganisms contains a complex assortment of potential attachment structures. These include outer membrane proteins, lipopolysaccharides (LPS), lipoteichoic acids, pili, and flagella. An additional structure that is less firmly associated with the bacterial surface is the capsule. Each of these components has been implicated in mediating specific attachment in any number of mammalian or marine adsorption systems (Berkeley et al. 1980, Bitton and Marshall 1980). It is of interest to note that few of these bacterial cell surface components have been studied with respect to adsorption to plant surfaces.

PAGE 12

Exceptions include reports suggesting the involvement of LPS in both the attachment of Aqrobacterium sp. to wound sites (Whatley et al. 1976) and the adsorption of Pseudomonas solanacearum to mesophyll cells (Sequeira and Graham 1977). Capsular polysaccharide-mediated adsorption of rhizobia to root hairs also has been reported (Dazzo and Hubbell 1975). Like most colloidal particles, bacteria usually acquire a surface charge when immersed in an aqueous environment. This is due to the ionization of surface chemical groups. Due to the presence of carboxyl, amino, or phosphate groups, the charge of the bacterial surface depends upon its environment. The surface charge of the bacterium can be measured as a function of pH, providing a quantitative measure of the behavior of a particle in an electrical field (Harden and Harris 1953). Once a charge has been established on the surface, it will attract a layer of oppositely charged counter ions and generate what is known as the Stern layer (Bohn et al. 1979). Adjacent to the adsorbed Stern layer is the thick diffuse double layer. This layer contains a higher concentration of counter ions than does the rest of the aqueous phase. The ionic characteristics of the bacterial surface, which vary as a function of ionic strength, have been shown to influence adsorption onto surfaces (Nissen 1971, Marshall 1976, Bitton and Marshall 1980, Gordon and Millero 1984, Pueppke and Benny 1984). It is now apparent that many Grampositive and Gram-negative bacteria also exhibit some degree of hydrophobicity on their outer surface (Marshall et al. 1975, Norkrans 1981, Rosenberg et al. 1982). This characteristic appears to be an important factor in bacterial adsorption to host and inert surfaces alike (Marshall 1976, Gibbon et al. 1983, Rosenberg et al. 1983). Bacteria have been shown to adsorb to oil-water interfaces and several different inert plastic-water interfaces (Marshall et al.

PAGE 13

1975, Marshall 1976). The contribution of hydrophobic interactions in mediating bacterial adsorption, however, remains unclear. Adsorption of Bacterial Invaders Plant-bacteria interactions can be separated into three groups, based primarily on the host response. In the first group no visible host response is elicited by the presence of the bacteria. This is termed a saprophytic relationship. In the second group, the interaction of a virulent bacterial strain with a susceptible host results in plant disease. This condition, which is the exception rather than the rule in nature, is termed a compatible response. Group three consists of the incompatible reaction, which is defined by the interaction of a phytopathogenic bacterium with a resistant plant. An avirulent form of the bacterium interacting with a normally susceptible host also results in an incompatible reaction. When plant cells are disturbed by incompatible bacteria, normal host physiological functions often are altered, and affected cells rapidly collapse and die. These responses collectively are termed the hypersensitive reaction (HR). The term hypersensitivity was borrowed from medical terminology, where it describes an organism that is abnormally sensitive to a pathogenic agent (Muller 1959). The HR in plants has been known for many years (Ward 1902), and although plants exhibit the HR to a large group of fungi (Kiraly 1980) and viruses (Holmes 1929), bacterial elicitation of the HR was not documented until the 1960's (Klement and Lovrekovich 1961, 1962, Klement et al. 1964). The basic test for the HR involved injecting or vacuum infiltrating a suspension of viable bacteria (I0 6 to I0 10 cells/ml) into the intercellular spaces of a leaf. Injection is accomplished by placing the beveled end of a hypodermic syringe

PAGE 14

containing the bacterial suspension against the leaf surface and applying pressure (Klement 1963). The intercellular spaces usually dry within 30 min, and hypersensitive flecks appear within 8 to 24 hr. Bacterial adsorption onto the plant cell surface is associated with the initiation of the HR, and bacterial adsorption thus may trigger this plant response. The HR is only triggered by metabolically active incompatible cells (Klement and Goodman 1967). The initial step of the HR, termed the induction period, defines the time reguired for the irreversible activation of the HR. Plant cell-bacterial cell contact leading to bacterial adsorption is thought to be reguired (Cook and Stall 1977, Klement 1977, Stall and Cook 1979). The mechanisms by which plants differentiate compatible from incompatible bacteria and initiate the HR against the latter are matters of great interest. One attractive hypothesis is that compatible bacteria multiply because they are allowed to remain free in the intercellular spaces of the leaf tissues. Incompatible bacteria, on the other hand, are thought to be adsorbed onto mesophyll cell wall surfaces and then enveloped by wall materials. These events are considered to immobilize the bacteria and to initiate the HR through an unknown series of biochemical events that culminate in bacteriostasis (Kiraly 1980, Segueira 1980, Klement 1982). Early adsorption and immobilization events initially were described in tobacco (Nicotiana tabacum ) leaves infiltrated with incompatible Pseudomonas syringae pv. p_is[ (Goodman and Plurad 1971, Goodman et al. 1977). During the induction period a portion of the outer cell wall separates from the mesophyll cells, apparently in response to adsorbed incompatible bacterial cells. Eventually the separated cell wall encompasses the bacteria, immobilizing them on the plant cell surface. A similar response was observed after injection of compatible and

PAGE 15

incompatible strains of Pseudomonas solanacearum into tobacco leaves (Sequeira et al. 1977). Data from other systems do not support the theory that adsorption and envelopment of incompatible bacteria are required to initiate the HR. For example, envelopment of the saprophyte P. put i da occurs in bean, but an incompatible, HR-inducing strain of P. syrinqae pv. tomato is not enveloped (Sing and Schroth 1977). In addition, although compatible P. syrinqae pv. phaseolicola cells are not enveloped by bean, numerous bacterial cells are adsorbed onto the plant cell walls (Sing and Schroth 1977). Sequeira et al. (1977) have reported the adsorption and envelopment of non HR-inducing organisms such as Escherichia coli , Bacillus subtilis , and heat-killed incompatible Pseudomonas solanacearum by tobacco. This demonstrates that adsorption is not sufficient for induction of the HR in these systems. Though adsorption and envelopment are of questionable significance in terms of the HR, the above data are consistent with several provisional conclusions: i) metabolically active incompatible cells are required to trigger the HR (Klement and Goodman 1967); ii) close contact and possibly bacterial adsorption onto the plant cell surface are prerequisities for induction of the HR (Klement 1977, Stall and Cook 1979); iii) because envelopment occurs in some incompatible interactions and not in others, this postulated defense mechanism does not appear to be a generalized plant response to incompatible invaders. Adsorption of Rhizobium spp. to Root Surfaces The ability of bacteria of the genus Rhizobium to infect the roots of legumes and produce nitrogen-fixing nodules is well documented (Fred et al. 1932). The symbiosis between legumes and members of the genus Rhizobium

PAGE 16

generates nearly half of the 175 million metric tons of dinitrogen fixed each year by microoganisms (Hardy and Havelka 1975). The legume -Rhizobium symbiosis is controlled by both the plant (Nutman 1956, Holl and Larue 1976, Lie et al. 1981) and the bacterium (Ljunggren 1961, Maier and Brill 1976, Beringer 1980, Hobbs and Mahon 1983). This symbiosis is highly specific in that only certain Rhizobium spp. are able to nodulate any given legume species. It is this specificity which provides the basis for taxonomic groupings within the genus Rhizobium (Fred et al. 1932). Initial control of this specificity in the legume -Rhizobium sp. interaction has been postulated to occur at four to five different levels starting with regulation of growth of rhizobia in the rhizosphere and culminating with infection thread formation (Broughton 1978, Vincent 1980). Host-symbiont specificity exhibited in the legume -Rhizobium system is hypothesized to be initiated by the specific attachment of infective bacteria to the host (for reviews, see Broughton 1978, Dazzo 1980a, Solheim andPaxton l98l,Pueppke 1984a). Substantial numbers of rhizobia adsorb, in many orientations, to root surfaces (Dazzo 1980a, 1980b, Higashi and Abe 1980, Stacey et al. 1980, Turgeon and Bauer 1982). Bal et al. (1978), however, found that only 5% of 1,798 examined root hairs contained adsorbed bacteria, and individual root hairs adsorbed only I to 16 bacteria. When adsorption is quantified and analyzed statistically, rhizobia are found adsorbed to both host and nonhost tissue in equivalent numbers (Chen and Phillips 1976, Bauer 1982, Law et al. 1982). In addition, both nodulating and non-nodulating strains of a given Rhizobium sp. bind to roots of the host legume (Broughton et al. 1980, Solheim 1983, Pueppke 1984b). Adsorption of four slow-growing strains of rhizobia to soybean and cowpea roots also was shown to be unrelated to the ability of the bacteria to infect and nodulate (Pueppke 1984b).

PAGE 17

12 A positive correlation between bacterial adsorption and legume infection has been suggested by other workers using alternate systems. Dazzo et al. (1976) demonstrated a direct relationship between the ability of R_. trifolii to infect white clover and its ability to adsorb to white clover root hairs. Specific adsorption of R. trifolii to red clover roots hairs has been correlated with the ability to infect and the presence of a symbiosis plasmid (Zurkowski 1980). Strains of R. trifolii containing the plasmid are adsorbed in much greater numbers than are strains lacking the plasmid. Several strains which have lost or reduced ability to nodulate soybean were found to have correspondingly reduced ability to adsorb to plant root tissue (Jansen van Rensburg and Strijdom 1982, Stacey et al. 1982). The inconsistency of these experimental results may be partially explained by the widely divergent techniques and systems used to examine bacteria-root interactions. For example, bacterial adsorption to root epidermal cells may mask specific adsorption to incipient root hairs, the site of infection. Even when adsorption to root hairs is monitored directly, care should be taken to quantify adsorption only to susceptible tissues (Bhuvaneswari, 1981, Bhuvaneswari et al. 1981). The frequently observed polar mode of attachment of the bacterium to the host's root surface was hypothesized to be of significance to infection (Sahlman and Fahraeus 1963, Dazzo 1980a, Gotz and Hess 1980, Gotz 1980). This polar attachment, however, appears to be a nonspecific phenomenon; polar binding of Rhizobium spp. occurs at oil-water and inert plastic-water interfaces (Marshall et al. 1975, Marshall 1976). The bacteria also attach in a polar fashion to roots of nonhost species such as wheat and petunia (Shimshick and Hebert 1978, Hess et al. 1982). Polar binding may simply reflect the fact that

PAGE 18

13 the poles of the bacterium are more hydrophobic, favoring this type of adsorption. Charge characteristics of both the bacterium and the plant cell surface also may function in making polar attachment energetically favorable. Fibrillar extracellular structures have been observed to encompass microcolonies of Rhizobium spp. on the surface of root cells (Reporter et al. 1975, Napoli et al. 1975). Deinema and Zevenhuizen (1971) provided evidence that some such structures are cellulose fibrils of bacterial origin. The participation of cellulose microfibrils in the initial adsorption events is probably limited, though they may serve to securely anchor the bacterium to the host in later stages of the interaction (Napoli et al. 1975, Pueppke 1984). Other substances with suspected adhesive properties also are observed in electron micrographs of the bacterium-plant interface. Marshall et al. (1975) illustrated that some type of bridging polymers bound polarly attached bacteria to the root surface. Others have seen similar structures in electron micrographs studies (Menzel et al. 1972, Dazzo and Hubbell 1975). These structures, however, also bind bacteria to inert plastic substrates. Thus, they are of bacterial origin and are induced nonspecifically (Marshall et al. 1975, Marshall 1976). Currently there is no clearcut consensus in the literature on the relationship between host-specific adsorption and nodulation specificity. This is understandable given the fact that there exist a large number of potential control points available to both the bacterium and host from the time of bacterial root contact to nodule formation. I believe that each potential regulation site deserves careful examination, which when accomplished, will most likely suggest a concerted multi-level control with both partners participating.

PAGE 19

14 Adsorption of Agrobacterium tumefaciens to Plant Surfaces Agrobocterium tumefaciens is a Gram-negative, soil-borne, rod-shaped bacterium which causes the plant disease known as crown gall. The crown gall is a neoplastic undifferentiated growth of the plant resulting from the stable incorporation and expression of a piece of bacterial DNA (T-DNA) in the nuclear DNA of the host (Chilton et al. 1977, Gurley et al. 1979, Thomashow et al. 1980). Roughly 38% of the tested dicot families and 86% of the tested gymnosperm families, and a few monocotyledonous species, are susceptible (De Cleene and De Ley 1976). Aqrobacterium tumefaciens is thought to adsorb to a specific chemical moiety in a wound site of the host's cell wall (Lippincott et al. 1977b, Rao et al. 1982). If specific adsorption does not occur, the bacterium is thought to be incapable of tumor induction. The specific adsorption hypothesis, developed by Lippincott and Lippincott (1969), provides the basis for what is known as the site-attachment hypothesis. Most of the data in support of the site-attachment hypothesis comes from tumor formation assays employing sequential inoculations. All inoculations in these experiments are carried out by dusting pinto bean leaves with carborundum, applying 0.1 ml of bacterial suspension, and thoroughly rubbing the surface of the leaf with a glass rod (Lippincott and Heberlein 1965a, 1965b). Like many biological assays, however, the pinto bean leaf assay (PBLA) suffers from a large degree of variability from experiment to experiment. For example, the mean number of tumors per leaf after inoculation with 2.8 to 4.7 X 10 cells/ml varied from 4.6 to 71 to 136 in three successive experiments. This 27-fold difference in tumor number was observed, even though the inoculum densities fell within the linear range of the dose-response relationship between tumor numbers and inoculum density (Lippincott and Heberlein

PAGE 20

1965b). Changes in plant age, greenhouse conditions, and the amount of pressure applied with the glass rod during inoculation result in a considerable difference in the number of tumors formed, and thus may account for some of the variability. Lippincott and Heberlein (1965b) suggested that an all or none assay could reduce the amount of variability associated with carborundum inoculations. Leaves could be wounded with a multiple needle holder, inoculated, and the wounds in which tumors developed scored. The relationship between tumor numbers and initial bacterial concentration in the PBLA is linear over approximately two log units of bacterial concentration, from approximately 3 X 1 6 to 8 X I0 8 bacteria/ml. The number of tumors remains relatively constant at bacterial concentrations greater than 8 X 10 8 cells/ml. In this narrow range of linearity ca. 3 X I0 6 bacteria are required for induction of a single tumor. Additionally, the PBLA is insensitive (i.e. no tumor formation) to bacterial concentrations less than ca. 3 X 10 cells/ml (Lippincott and Heberlein 1965b). Theoretical curves for oneand two-particle events suggest that individual tumors are initiated by a single bacterium acting at a susceptible site. This would be in agreement with Hildebrand's (1942) earlier observation that a single bacterium placed on a competent plant cell can initiate tumor formation. However, this is inconsistent with the observation that tumors are not formed when inoculum densities are below ca. 3X10° bacteria/ml. The extrapolation of tumor data to bacterial adsorption poses pitfalls for those attempting to understand bacterial adsorption. This difficulty has not escaped the attention of those working with the PBLA. "When more complex developments are used to measure the interaction such as disease symptom formation or the induction of host defense systems, many factors in addition to

PAGE 21

16 adherence may function to determine if the specific response sought will occur. In such cases, adherence may constitute only one small part of the specificity-recognition system, or, in the extreme and even though essential, an adherence mechanism could prove so nonspecific that it would ordinarily have no role in specificity-recognition phenomena." (Lippincott and Lippincott 1980, pp. 377). In the PBLA, virulent _A. tumefaciens strains form tumors that can be easily tallied 7 days after inoculation. When pinto bean leaves were inoculated with the avirulent strain IIBNV6 prior to or in combination with, virulent strain B6, the number of tumors was reduced relative to that in controls inoculated with strain B6 alone. Preinoculation and coinoculation with strain IIBNV6 caused similar levels of inhibition. To test the possibility that such inhibition is due to simple nonspecific physical blockage by strain IIBNV6, other bacteria also were tested. Preinoculation with E. coli, Bacillus sp., and Rhizobium sp. did not reduce the number of tumors. Additionally, when the process was reversed, and the plant was inoculated with 36 prior to IIBNV6, no reduction in tumor formation was observed. These results eliminated the possibility that strain 11BNV6 nonspecifically blocked the adsorption of strain B6. In addition to IIBNV6, several virulent (heat-treated) and avirulent _A. tumefaciens strains inhibit tumor formation by strain B6. Such strains are termed site-binding strains (Lippincott and Lippincott 1975). The above data provided the basis for the site-attachment hypothesis which, simply stated, suggests that there exists a single specific attachment site within a wound to which a single bacterium can attach and initiate tumor formation. Tumor inhibition, however, results from the exclusion of a single virulent cell from a specific site by a single avirulent cell. It should be emphasized that the A. tumefaciens site-

PAGE 22

17 attachment hypothesis is based on tumor formation data and is not directly supported by bacterial adsorption data from the PBLA. What appears to be a similar phenomenon, i.e. inhibition of tumor formation by inactivated cells, was noted 46 years previously (Brown 1923). No tumors formed on daisy plants puncture-inoculated with heat-killed A. tumefaciens and later inoculated in the same site with viable, virulent A. tumefaciens cells. Enhanced tumor formation by mixtures of virulent and avirulent A. tumefaciens cells has led to some interesting, though contradictory, concepts dealing with _A. tumefaciens attachment (Lippincott and Lippincott 1970, 1977a, 1978b, Lippincott et al. 1978). When pinto bean leaves were inoculated with mixtures containing virulent and avirulent cells of Agrobacterium sp., a 10 to 20-fold increase in tumor numbers was noted. The same enhancement was not observed if virulent cells were substituted for avirulent cells, i. e. if the inoculum density was increased. In these enhancement (complementation) experiments, the virulent strain is termed the donor strain, and the avirulent strain is designated the receptor strain (Lippincott and Lippincott 1970). Heat treatment of either the donor or receptor strain totally abolishes the enhancing activity. Although independent of the ratio of virulent to avirulent cells, complementation is affected by the absolute numbers of bacteria in the inoculum. For example, a 2.3:1 ratio (5 X I0 8 IIBNV6 cells and 2.2 X I0 8 B6 cells/ml) caused a 33% increase in tumor formation, but another 2.3:1 ratio (I X I0 9 11BNV6 cells and 4.4 X 1 8 B6 cells/ml) caused a 44% reduction in tumor formation. A ratio of 68:1 (6.4 X I0 9 I1BNV6 cells and 9.4 X I0 7 B6 cells/ml) resulted in a 63% reduction in tumor formation, yet cells in a similar ratio, 54:1 (4.5 X I0 9 1IBNV6 cells and 8.4 X 10 7 B6 cells/ml), resulted in a 40% increase in tumor formation.

PAGE 23

Lippincott and Lippincott (1970) believe that complementation results from the separate action of a helper strain at one site and an avirulent strain at a second site. In their view, "the enhancement apparently depends on one or more substances moving between wounds occupied by virulent bacteria and wounds where I1BNV6 [receptor strain] cells are attached" (Lippincott et al. 1977a, pp. 828). Although either inhibition or enhancement can occur under conditions where ratios of virulent and avirulent cells are similar, tumor enhancement requires that the absolute number of virulent cells be relatively low, and tumor inhibition requires that the number of virulent cells be relatively high. When the concentration of the virulent strain is held constant and the concentration of the complementary avirulent receptor strain is increased, the degree of tumor enhancement increases. The explanation for tumor enhancement is radically different from the site-attachment hypothesis proposed to explain the tumor inhibition data. Inoculation procedures are identical in both tumor enhancement and inhibition studies. When tumor inhibition is observed, competition for available attachment sites is suggested to account for the reduction in tumor formation. When tumor enhancement is observed, it is postulated that different bacteria communicate between attachment sites in separate wounds. It would be equally plausible to hypothesize that an increased level of adsorption leads to the increased tumor formation noted in the complementation experiments, i.e. that avirulent cells facilitate the adsorption of virulent cells. This has been reported to occur when animal tissue was coinoculated with two Streptococcus mutants defective in adsorption (Larrimore et al. 1983). Although the timed inoculation data are difficult to explain, the available evidence also is consistent with the postulate that tumor inhibition is

PAGE 24

19 a result of the influence of inhibitory compounds from avirulent cells on virulent cells bound to sites competent for transformation. Taken together, the results suggest that the interactions responsible for the altered tumor formation are very complex. Consequently, i believe that the evidence provides little justification to support the extrapolation to bacterial attachment in the case of tumor inhibition, or the postulation of an additional mechanism when conditions result in enhancement of tumor formation. Proposed Molecular Components Involved in Site-attachment The LPS on the bacterial cell surface are thought to be involved in the site-attachment process. The LPS isolated from site-binding A. tumefaciens strains by the hot phenol-water method of Westphal and Jann (1965) inhibit tumor formation. When LPS were applied to the wound sites prior to inoculation with B6, a reduction in tumor formation was observed (Whatley et al. 1976). This was not the case with LPS from non-site-binding strains. When the LPS were hydroiyzed into two components, the o-antigens and lipid A region, the o-antigens were shown to be responsible for the inhibitory activity (Whatley et al. 1976, Lippincott and Lippincott 1977). Tumorigenicity is suppressed by introduction of plasmid pSa into A_. tumefaciens (Farrand et al. 1981). More recently New et al. (1983) examined tumor inhibitory activity of strains with and without the pSa plasmid. Strains harboring pSa were no longer tumorigenic. LPS isolated from this strain also were unable to inhibit tumor formation by the virulent strain. When these strains were cured of the pSa plasmid, tumorigenicity and the ability of the isolated LPS to inhibit tumor formation were restored. Isolated LPS from strains containing and lacking pSa were compared with a variety of

PAGE 25

20 techniques. Strains with and without the pSa plasmid were not qualitatively different. Carbohydrate content of the LPS was similar, and the relative amount of 2-keto-3-deoxyoctonic acid to o-antigenic polysaccharides was not significantly different (New et al. 1983). Purity of any of these LPS fractions, however, was not established, and their compositions were not determined. The LPS isolated using the Westphal and Jann (1965) technique commonly contain 2-5% nucleic acid by weight. Detergent treatment will precipitate nucleic acids, but this procedure was not performed on the LPS used in the tumor inhibition experiments. Glucose-rich polysaccharides contaminate LPS isolated from Rhizobium sp. (Carlson et al. 1978), and LPS, regardless of the isolation procedure, always contain inorganic cations and low molecular weight basic amines, e.g. spermine, spermidine and ethanolamine (Burton and Carter 1964, Luderitz et al. 1968, Galanos and Luderitz 1975). The LPS fractions are comprised of an unpredictable mixture of different salt forms. This results in a heterogeneous LPS fraction having variable physiochemical properties that may affect biological activity (Galanos and Luderitz, 1975). Johnson and Perry (1976) suggested five criteria for establishing LPS purity: i) elution of a single symmetrical peak at the void volume of a Sepharose 4B or 6B column, ii) the absence of ribose or deoxyribose, iii) lack of adsorption between 210 and 300 nm, iv) symmetrical elution peaks of o-antigen and core-polysaccharides from Sephadex G-50 columns, and v) a single absorbance maximum at 460 nm with the carbocyanine dye assay for LPS. The LPS material used in the tumor inhibition studies apparantly was not examined by these criteria. Attributing the observed activity to the LPS component thus, unfortunately, may be misleading.

PAGE 26

21 Plant surface molecular components also were analyzed for their possible involvement in the site-attachment process. An isolated pinto bean leaf cell wall fraction inhibited tumor formation, but the plant cell membrane fraction was without activity (Lippincott et al. 1977b, Lippincott and Lippincott 1978a). Treatment of the cell wall fraction with either hot water or acid destroyed its inhibitory activity (Lippincott and Lippincott 1977, Rao et al. 1982), whereas treatment with ethylenediaminetetraacetic acid (EDTA), detergent, pectinase, or cellulose was without effect (Lippincott and Lippincott 1977). However, subsequent work from the same laboratory suggests that pectinase destroys the ability of the cell wall fraction to inhibit tumor formation (Lippincott et al. 1983). Considerable caution, however, should be taken with results obtained using isolated crude cell wall fractions. These fractions were obtained by grinding leaves in buffer and removing the particulate wall material by low speed centrifugation. The material was then suspended in acetone, homogenized, filtered, washed three times with acetone and air dried. Similar preparations have been shown to contain 6% (w/v) protein (Slusarenko and Wood 1983). Although purity was not demonstrated for wall fractions used in the tumor inhibition assays, the authors suggest that the pectic portion of the fraction is the active component responsible for tumor inhibition. This was verified by examining several off-the-shelf commercial preparations of cell wall origin. Citrus pectin and the sodium salt of its unesterified form, polygalacturonic acid, were found to be inhibitory; cellulose and several other polysaccharides were without activity. Sodium polygalacturonate exhibited the greatest degree of tumor inhibition and was active to a concentration of I ng/ml. The observation that polygalacturonate is much more effective than

PAGE 27

22 pectin led to the conclusion that the degree of methylation of the galacturonan residues may be responsible for the observed differences in activity (Lippincott and Lippincott 1980, Rao et al. 1983). The influence of methylation on tumor formation was examined by enzymatically removing methyl groups with a commercial preparation of pectin methylesterase. The product of this reaction had enhanced ability to inhibit tumor formation relative to pectin. On the other hand, enzymatic methylation with pectin methyltransferase resulted in reduced tumor inhibition. In most cases cell wall fragments isolated from monocot tissue exhibited little inhibitory activity (Lippincott and Lippincott 1978a). However, upon treatment with pectin methylesterase, they became inhibitory (Lippincott and Lippincott 1978a). This suggested a possible mechanism for the well-known resistance of most monocotyledonous plants to infection by Agrobacterium spp. Thus, the apparent lack of attachment sites in monocots was hypothesized to be due to highly methylated galacturonan residues that may present a formidable barrier to tumor formation (Lippincott and Lippincott 1977, 1978a). These results led Lippincott and Lippincott (1977) to state, "The simplest direct conclusion from these data is that the cell walls of monocots, tumors and embroyonic bean tissues are sheathed with pectic substances which are sufficiently methylated that Agrobacterium does not adhere." This mechanism suggests that treatment of wound sites in monocotyledonous plants with pectinmethylesterase would expose attachment sites and render the monocot susceptible to tumor formation. Conversely, treatment of wounds in dicot tissue with pectin methyl transferase should confer some degree of resistance to Agrobacterium spp. These experiments, which would provide a direct test of the attachment site hypothesis, have not been reported.

PAGE 28

23 Tumor inhibition by pectin, polygalacturonase, LPS, and avirulent sitebinding A. tumefaciens strains also have been shown by others (Brown 1923, Schilperoort 1969, Glogowsky and Galsky 1978, Cooksey and Moore 1981, Pueppke and Benny 1981). Not all these reports, however, totally support the site-attachment hypothesis. For example, Glogowski and Galsky (1978) found that both E. coli and Pseudomonas fluorescens inhibited tumor formation on potato. In addition, pretreatment of Datura stramonium wounds with avirulent Agrobacterium sp. cells had no effect on tumor formation (Beaud et al. 1963). Site-binding, avirulent A. radiobactor strain K84 reduced tumor formation by 70% when applied 3 hours after inoculation with a virulent, agrocin resistant strain (El-Kady and Sule 1981). Furthermore, others have found, in a somewhat different assay system, that only living, site-binding strains of A. tumefaciens were able to inhibit tumor formation (Schilperoort 1969, Bogers 1972, Douglas et al. 1982). These data may not necessarily conflict with previous reports, because Lippincott and Lippincott (1969) reported that the heat treated (60° C, 20 min) bacterial suspensions (I0 10 to I0 1 ' cells/ml) still contained 10^ viable cells/ml. Though present in low numbers, a population of viable cells could influence the degree of tumor formation recorded 6 to 7 days later. The number of viable cells remaining after heat treatment was not reported by Bogers (1972) or Schilperoort (1969). It has been reported that a process similar to the proposed site attachment step in tumor formation is essential for the induction of moss gametophore formation that occurs in the presence of A. tumefaciens (Spiess et al. 1971, 1976). In the absence of a wound, gametophore induction requires physical contact with A,, tumefaciens. The extent of gametophore induction is

PAGE 29

24 related linearly to the concentration of bacteria and is inhibited by heat-killed virulent cells (Spiess et al. 1976). Whatley and Spiess (1977) have identified the polysaccharide portion of the LPS as the active component that mediates the attachment of A. tumefaciens to moss. Moss gametophore induction, however, is not uniquely induced by A. tumefaciens, because five different Rhizobium spp. also were active (Spiess et al. 1977a). It would prove interesting to examine the ability of a wide variety of bacteria to induce gametophore formation. Such a test would directly address the question of whether gametophore induction by A. tumefaciens is a useful model system to study early attachment events. When adsorption of non-site binding _A. radiobactor stain 6467 and site-binding A. tumefaciens strain B6 onto moss protonemal filaments was monitored by scanning electron microscopy, few differences were found (Spiess et al. 1977b). On the average, the number of B6 cells bound per spore was six more than that of A. radiobactor 6467. This appears to be the strongest evidence for site-specific adsorption mediating gametophore induction. Adsorption of Agrobacterium spp. to plant tissue culture cells It should be emphasized that the entire preceding discussion deals with the use of tumor formation and inhibition data to define the adsorption of Agrobacterium sp. onto plant tissue. These events are separated by 7 days and may not be directly related to one another. Several workers have overcome this problem by directly measuring the attachment of _A. tumefaciens onto tissue culture or mechanically separated plant cells (Matthysse et al. 1978, Smith and Hindley 1978, Ohyama et al. 1979, Douglas et al. 1982, Draper et al. 1983). In only one case, however, have workers using tissue culture cells

PAGE 30

25 directly monitored bacterial adsorption under conditions where transformation is known to occur (Smith and Hindley 1978). Using cultured tobacco cells, Smith and Hindley (1978) found that both virulent C-58 and avirulent NT1 strains bound, though C-58 was adsorbed in larger numbers. Transformation in this system was substantiated by measuring the production of nopaline in the infected tobacco cells. It was subsequently shown that wall-regenerating tobacco protoplasts can be transformed with both _A. tumefaciens cells and purified Ti plasmid and then regenerated into transformed plants (Marton et al. 1979, Wullems et al. 1981, Krens et al. 1982). Matthysse et al. (1978, 1981) examined the attachment of various A. tumefaciens strains to both carrot and tobacco tissue culture cells. Using a ratio of approximately two bacteria per carrot cell (10 bacterial cells/ml:5 X 10 plant cells/ml) Matthysse et al. (1978) observed that a larger percentage of the virulent cells bound after incubation for 2 hours than did the avirulent cells. The initial bacterial inoculum varied from 5X10 cells/ml to 1.5 X I0 4 cells/ml. If 10 cells/ml are assumed to be the initial inoculum, 0.5 to 0.9 cells of each of five virulent Ti plasmid-containing strains bound per plant cell. On the other hand, approximately 0.2 bacteria bound per plant cell for each of the five tested avirulent agrobacteria. Therefore, in all cases examined, more than half the plant cells would be expected to be devoid of adsorbed bacteria. Even more interesting is the fact that when the number of plant cells is increased 10-fold, there is no change in the number of bacteria adsorbed from the bacterial suspension (Matthysse and Gurlitz 1982). Close examination of the absolute values of the adsorption data using the carrot tissue culture cells leads to some interesting observations. Plant cells adsorbed 24% of an initial inoculum containing 5X10 cells of A. tumefaciens

PAGE 31

26 strain AI78/ml. Thus 1.2 X I0 3 bacterial cells were removed from the suspension. In similar adsorption experiments with A. tumefaciens NT1, 8-9% of the initial inoculum (1.5 X lO 4 cells/ml) attached to the plant surface. Thus 1.2 X 10 NT1 cells were removed from the suspension. Therefore the actual number of virulent (A 1 78) or avirulent (NT1) bacterial cells bound was identical and independent of virulence. Matthysse et al. (1978) also reported that the adsorption of site-binding strains IIBNV6 and NT1 to tobacco was small compared to that of A. tumefaciens B6. This deviates from what would be expected if the siteattachment process as described by Lippincott and Lippincott (1969) is operative. Since prior application of IIBNV6 or NT1 cells reduced tumor formation by strain B6, similar levels of adsorption would be postulated. The differences, however, may reflect differences in the two attachment systems. On the other hand, they may simply call into question the general applicability of the site-attachment hypothesis to systems other than pinto bean. Matthysse et al. (1982) provided additional evidence that questions the validity of the site-attachment hypothesis for suspension culture systems. The kinetics of attachment of A. tumefaciens B6 to both glutaraldehyde-fixed carrot protoplasts and intact carrot cells from tissue culture were similar. Protoplasts were determined to be free of cell wall material by calcofluor white ST staining. Therefore, bacterial attachment to plant tissue occurred in the absence of the postulated host receptor-site, pectic wall material. However, partial synthesis of the cell wall is required for the transformation of tobacco protoplasts by intact A. tumefaciens cells (Marton et al. 1979), suggesting that A_. tumefaciens attachment to carrot protoplasts is not the prelude to transformation.

PAGE 32

27 Cellulose microfibril production by A. tumefaciens cells is associated with anchoring the bacterium to the surface of tissue culture plant cells and protoplasts (Matthysse et al. 1981). The ability of Aqrobacterium spp. to produce microfibrils first was described by Deinema and Zevenhuizen (1971) and later suggested by Spiess et al. (1977b) to play a possible role in bacterial attachment to moss tissue. Subsequently, however, it has been shown that the ability to synthesize cellulose microfibrils is not a prerequistie for bacterial virulence or the ability to attach to plant cells (Matthysse, 1983). Since the microfibrils are not observed until approximately 60 min after inoculation, they are postulated to function in anchoring large numbers of bacterial cells, thus creating a more favorable microenvironment for the subsequent transfer of the T-DNA. Adsorption of Agrobacterium spp. to Culture Monocot and Dicot Cells. Comparative data on the number of bacteria adsorbed to cultured monocot and dicot cells vary from laboratory to laboratory. Matthysse and Gurlitz (1982) reported that dicot cells adsorb larger numbers of A_. tumefaciens cells, but the distinction was not apparent in other studies (Ohyama et al. 1979, Draper et al. 1983). When a ratio of approximately one bacterial cell to every 100 plant cells is used, a larger percentage of the inoculum is adsorbed to carrot (Daucus carota ) suspension culture cells (48%) than to corn (Zea mays) suspension culture cells (3%). However, when actual numbers are calculated from the data, 2X10 carrot cells bound 960 bacteria and, if it is assumed that individual plant cells bind no more than I bacterial cell, 99.5% of the plant cells have no bacteria adsorbed. Approximately 60 bacteria bound to the 2 X 10 corn cells, leaving 99.9% of the plant cells

PAGE 33

28 without a single bacterium attached. When viewed in this manner, A. tumefaciens adsorption to monocot versus dicot tissue becomes much less distinct. This distinction was further blurred by information provided by Ohyama et al. (1979), who measured the adsorption of radioisotope-labeled A. tumefaciens B6 to a variety of monocot and dicot suspension culture cells. Corn, bromegrass (Bromus inermis) , rice (Qryza sativa ), and tobacco cells bound equivalent numbers of B6 cells. In each case, 1-4 mg dry wt of plant cell suspension/ml adsorbed approximately 2.3 X I0 6 bacteria from an initial inoculum of 3.6 X 10 cells. Soybean cells bound significantly larger numbers of B6 cells than did tobacco. It is of interest to note that intact soybean plants are resistant to infection by most Agrobacterium spp. Recently however A. tumefaciens strain A208 from Monsanto Co., St. Louis, Missouri, has been shown to induce tumor formation on soybean plants (Jill Winter, personal communication). Adsorption of A. tumefaciens to Susceptible Host Tissue When A. tumefaciens adsorption onto susceptible host tissue is measured directly, neither LPS nor pectic material participates in the adsorption process, as predicted by the site-attachment hypothesis (Pueppke and Benny 1984). Pueppke and Benny (1981, 1983) reported that tumor formation on potato tubers was inhibited by added galacturonans in a similar manner as seen with the PBLA. In contrast, both the degree of galacturonan methylation and the presence of LPS in the inoculum was without influence on tumor inhibition in the potato system (Pueppke and Benny 1983). A single exception was noted with the avirulent, heat-cured strain NT1, the LPS of which was inhibitory to tumor formation. Adsorption assays consisted of measuring the level of

PAGE 34

29 attachment of S-methionine-labeled bacteria to potato tuber discs (Pueppke and Benny 1984). Four different galacturonans, which varied from to I 1.7% methoxy by weight, were examined for their influence on adsorption. When adsorption was measured in the presence of citrus pectin (I 1.7% methoxy), polygalacturonic acid (0% methoxy) and demethylpectin (2.6% methoxy), the number of adsorbed bacteria was statistically greater than that in the absence of each compound. The presence of methylpolygalacturonic acid (8.2% methoxy) in the inoculum caused no statistically significant change. Adsorption in the presence of LPS isolated from virulent and avirulent strains was not statistically different from adsorption measured in the absence of LPS. LPS from a single strain, NT1, were enigmatic and inhibited bacterial adsorption. LPS from strain NT1 were the only additive that inhibited both adsorption and tumor formation. Summary of Literature Review According to the specific site-attachment hypothesis there exists a single site within a wound to which a single virulent bacterium can attach and cause tumor formation. According to the hypothesis, tumor inhibition results from the exclusion of a single virulent cell from a specific site by a single avirulent, site-binding cell. Adsorption is hypothesized to be mediated by the binding of bacterial cell surface LPS to sparingly methylated galacturonans in wounded dicot plant cell walls. The hypothesis has several corollaries. The first is that monocotyledonous plants are resistant to A. tumefaciens because the bacteria are unable to adsorb to the plant cell surface (Lippincott and Lippincott 1977,

PAGE 35

30 1978a, Rao et al. 1982). The second is that there exist two different types of avirulent A. tumefaciens cells, which can be distinguished by their abilities to bind to plant surfaces. The site-attachment hypothesis and its corollaries have been discussed and for the most part accepted by many reviewers of A. tumefaciens tumorigenesis (Merlo 1978, Dazzo 1980a, Dazzo 1980b, Moore and Cooksey 1981, Nester and Kosuge 1981). However, the hypothesis is entirely based on indirect data from pinto bean. In addition, although adsorption has been measured directly and quantified, the experimental conditions used in these studies do not lead to transformation. Thus, there is no assurance that the measured adsorption events are relevant to T-DNA transfer. The initial overall objective of this study was to test the site-attachment hypothesis and its corollaries. This was done (I) by measuring the adsorption of virulent, avirulent site-binding, and avirulent nonsite-binding strains to susceptible host tissue under conditions known to permit T-DNA transfer and (2) by comparing the attachment of A. tumefaciens cells to monocotyledonous and dicotyledonous plant tissues. In addition, I performed a series of experiments designed to define the mechanism of A. tumefaciens adsorption to potato tissue.

PAGE 36

CHAPTER THREE ADSORPTION OF AGROBACTERIUM TUMEFACIENS TO POTATO (SOLANUM TUBEROSUM ) TISSUE Introduction Aqrobacterium tumefaciens is a plant pathogen that causes a tumorigenic disease known as crown gall. This pathogen exhibits an extremely wide host range that includes 86% of the tested gymnosperm families, 38% of the tested dicot families and several monocotyledonous plants (De Cleene and De Ley 1976). Crown gall tumors are neoplastic growths that result from the transfer and stable incorporation of a piece of bacterial plasmid (T-DNA) into the plant nuclear DNA (Chilton et al. 1977, Yadau et al. 1980). Adsorption of A_. tumefaciens to the surface of the plant host is thought to be one of the earliest events leading to tumor formation (Lippincott and Lippincott 1969, 1977, 1980). Little is known of the precise adsorption mechanism, although adsorption most likely results from the interaction of both attractive and repulsive forces acting at the plant-bacterium interface. Hydrophobicity, cell surface charges, and surface polymers all may contribute to the formation of a stable host-parasite cell interface. Several authors have suggested that adsorption in this system is site-specific (Lippincott and Lippincott 1969, Lippincott et al. 1977b). Most of the supporting data are from experiments that measure the effects of various isolated plant or bacterial compounds on the ability of A. tumefaciens strain B6 to form tumors on wounded pinto bean (Phaseolus vulgaris) leaves (Whatley et al. 1976, Lippincott et al. 1977b, Rao et al. 1982). The compounds are added to the bacterial inoculum or applied prior to inoculation. After 7 days, the number of 31

PAGE 37

32 tumors is counted. Substances that reduce the number of tumors are considered to do so because they interfere with the initial bacterial adsorption step. The bacterial component responsible for mediating this site-specific adsorption is believed to be the o-antigen portion of the lipopolysaccharides (LPS) (Whatley et al. 1976, Whatley and Spiess 1977). Exogenously supplied LPS interfere with tumor formation and are thought to recognize some galacturonan-rich pectic portion of the wounded plant cell wall (Whatley et al. 1976, Lippincott et al. 1977a, 1980). Exogenously supplied pectic compounds also inhibit tumor formation. It has been suggested that the degree of methylation of the pectic material influences the ability of A. tumefaciens to adsorb to plant cell surfaces. Thus, the resistance of very young dicotyledonous plants and most monocotyledonous species to A. tumefaciens, is hypothesized to reflect the extent of methylation of pectic polysaccharides of the cell wall (Lippincott and Lippincott 1978a). In a system where bacterial adsorption to potato tuber tissue is measured directly, rather than inferred from tumor experiments, most LPS and various pectic compounds have no inhibitory effect on adsorption (Pueopke and Benny 1984). However, citrus pectin actually enhanced bacterial adsorption, and LPS isolated from A. tumefaciens strain NT1, the single exception, reduced the number of bacteria adsorbed to the potato disc surface. In this system, pectic compounds dramatically reduce tumor formation, and LPS have little or no effect (Pueppke and Benny 1981, 1983). In addition, when attachment of _A. tumefaciens to monocot tissue is measured directly, the numbers of bound bacteria are comparable to those bound to potato tissue (Chapter Five this dissertation). These observations suggest that adsorption of _A. tumefaciens to plant cells is less specific than previously thought.

PAGE 38

33 In this investigation my objective is to provide a quantitative description of the initial binding of A. tumefaciens to potato tissue. Materials and Methods Adsorption Experiments Tumorigenic A_. tumefaciens strain B6, from J. A. Lippincott, Northwestern University, was maintained at 4°C on slants of the defined gluconate-mannitol medium of Bhuvaneswari et al. (1977). After incubation for 2 to 3 days A. tumefaciens cells were washed from the slants using sterile phosphate-buffered saline (PBS: 7.2 g NaCI, 2.79 g Na 2 HP0 4 • 7H 2 0, 0.43 g Kh^PO^ per liter of deionized water; final pH 7.2), centrifuged for 10 min at 7700 X G, and washed twice with PBS. The washed bacteria were resuspended in PBS and the bacterial concentration determined turbidimetrically. The effect of initial bacterial concentration on bacterial adsorption to plant tissue was assayed by a modification of the tumorigenesis assay described previously (Pueppke and Benny 1981). Tuber discs of potato (Solanum tuberosum L. cv. Red LaSoda) (8 mm diam X 2 mm thick) were prepared aseptically. With the use of a wire basket, batches of discs (30 discs/basket) were lowered into a 100-ml beaker containing 30 ml of bacterial inoculum (10' to 10 cells/ml of PBS). The discs were agitated gently and allowed to remain undisturbed for 60 min. The basket containing the discs then was removed and immersed in two separate rinses, each containing 30 ml of sterile PBS. Each of the two rinses consisted of vigorous agitation of 4 to 5 seconds to remove any loosely attached bacterial cells. The discs were removed from the basket, blotted dry on sterile paper towels, and three randomly chosen discs were

PAGE 39

34 transferred to 2 ml of PBS contained in a hand-held, ground-glass tissue homogenizer. The discs then were homogenized. Aliquots of 100 ul from each of three dilutions, previously determined to give less than 500 colonies/plate, were spread on each of five petri plates containing yest extract-mannitol agar (0.5 g K 2 HP0 4 , 0.2 g MgS0 4 7H 2 0, 0.1 g NaCI, 10 g mannitol, 0.4 g yeast extract per liter of deionized water). Viable colonies were counted after incubation of the plates for 36 to 48 hr at 28°C. The reported number of bacteria adsorbed at each inoculum density is the average from two to five experiments, each of which was replicated three times. Therefore, the adsorption value at each inoculum density is derived from measurements of colonies on 6 to 15 individual plates. As a control, noninoculated discs were homogenized, and the homogenates spread on plates. Although a few colonies formed infrequently, none resembled the white, mucoid A. tumefaciens colonies. To determine the effect of homogenization on bacterial viability, a known number of bacterial cells was added to the buffer prior to homogenization. Recovery of these cells was not influenced by the homogenization step. Desorption Experiments Potato tuber discs were incubated in a bacterial suspension of 10 cells/ml of PBS for 40 min. Three discs were removed, washed as described above, and the number of attached bacteria determined. The remaining bacterial suspension, which then contained 27 tuber discs, was diluted with PBS such that the bacterial concentration was reduced by one log unit, i.e. to 10^ cells/ml. The initial 1:1 ratio of potato discs to ml of bacterial suspension was always maintained. After 40 min another three discs were removed and the

PAGE 40

35 number of adsorbed bacterial cells determined as above. Two additional desorption experiments were done in an identical manner using a stepwise reduction in bacterial density from I0 8 to I0 7 cells/ml of PBS, and \0 6 to I0 4 cells/ml of PBS. Scanning Electron Microscopy Potato discs for scanning electron microscopy were obtained immediately after inoculation with 2 X I0 7 cells of strain B6/ml of PBS. Discs were fixed overnight at 4°C in 4% glutaraldehyde and rinsed with 0.05 M cacodylate buffer at pH 7.2. Samples then were fixed in OsO^ for 2 hr at room temperature, washed with cacodylate buffer, and dehydrated through an ascending series of ethanol solutions. Samples were critical-point dried by substitution of ethanol with liquid C0 2 , sputter coated with gold, and viewed with a Hitachi 450 Scanning electron microscope at 20 KV. Results Bacterial Adsorption and Desorption Depending on initial bacterial concentration, 0.23 X I0 1 to 1.2 X I0 7 bacteria adsorbed per disc (Table I ). The percentage of bacteria adsorbed from suspensions containing from I0 1 to I0 4 cells/ml of PBS varied from 23 to 5. Approximately 1% of the inoculum was adsorbed at bacterial concentrations greater than I0 5 cells/ml of PBS (Table I). Saturation was never observed in these experiments. Adsorption studies also were carried out using potato discs

PAGE 41

36 Table I. Adsorption of Aqrobacterium tumefaciens to Potato Tissue. Initial Bacterial Concentration (cells/ml) No. of Bacteria Adsorbed/Disc Measured

PAGE 42

37 with a 43% smaller surface area than those used in most experiments. When these smaller discs were exposed to bacterial densities up to I0 9 cells/ml of PBS, saturation of the available adsorption sites was not reached. When bacterial suspensions in equilibrium with potato discs were diluted 9 8 from 10/ to 10° cells/ml, there was coincident reduction in the number of bacterial cells adsorbed to the discs (Table 2). The number of bacteria adsorbed per disc also decreased after the 10-fold dilution from I0 8 to I0 7 cells/ml. A 100-fold dilution from I0 6 to I0 4 cells/ml similarly resulted in a reduction in the number of cells bound per disc. These data indicate that binding is reversible. Theoretical Considerations Before experimental data can be analyzed with respect to adsorption isotherms, several preconditions about the adsorption system under analysis must be met (Brunauer 1943, Saunders 1971, Tompkins 1978). First, the total number of adsorption sites should be constant under all experimental conditions. This condition is met because discs of the same size were used in each experiment. Second, adsorption must be restricted to a monolayer. This was confirmed by examination of inoculated potato tuber surfaces by scanning electron microscopy (Fig. I). The existence of monolayer adsorption eliminates the possible use of the Brunauer-Emmett-Teller adsorption isotherm, which describes multilayer adsorption phenomena (Brunauer 1943). Third, every adsorption site should be equivalent, and the ability of the bacteria to adsorb must be independent of whether or not adjacent sites are occupied. These conditions also appear to be fulfilled. Potato tuber tissue is relatively uniform,

PAGE 43

38 Table 2. Desorption of Agrobacterium tumefaciens Cells from Potato Tuber Discs. a Initial No. Bacteria Final No. Bacteria Per ml of Inoculum Adsorbed/Disc Per ml of Inoculum Adsorbed/Disc I0 9 2.8 + 0.3XI0 7 10 8 5.9+I.3XI0 6 I0 8 2.0 + 0.5XI0 6 I0 7 9.7+I.3XI0 5 I0 6 I.4 + 0.2XI0 4 I0 2 6.0 + 0.3X10 3 a The data are from separate experiments. Three discs were homogenized at each sample time, and each homogenate was spread on plates. Standard Deviation.

PAGE 44

39 Fig. I. Scanning electron micrograph of Agrobacterium tumefaciens strain B6 on potato tuber tissue. The initial inoculum density was 2 X 10 bacteria/ml and the disc was fixed immediately after inoculation. The field of view presented here is 870 unrr and contains 12 rod-shaped bacteria. X 5200.

PAGE 45

40 and it would be expected that adsorption sites would be reasonably equivalent. Moreover, bacterial adsorption to the discs did not reach saturation, even when small discs were exposed to very high bacterial densities. This makes it unlikely that the plant cell surfaces were packed with, bacteria and that occupation of adjacent sites was of importance. These observations provide evidence for the equivalency of adsorption sites. Fourth, the rate of adsorption to unoccupied sites must equal the rate of desorption from occupied sites at equilibrium, i.e. equilibrium must be reversible. This is illustrated in Table 2. The binding of A_. tumefaciens strain B6 to potato tuber tissue can be described best in terms of the Freundlich adsorption isotherm (Brunauer 1943, Tompkins, 1978), which is expressed as C b = k| XCs l/k 2 where C^ = the number of bacterial bound per disc, C $ = the concentration of the bacteria in the suspension at equilibrium, k | and \<2 are both constants. When a linear regression is performed on the plot of log C vs log C^, a correlation coefficient of r = 0.987 is obtained for the entire range of inoculum densities studied. Demonstration of the linear relationship between log C s vs log C^ is in accordance with the criteria set forth for the Freudlich isotherm (Brunanuer 1943, Tompkins 1978). When the data are fitted to three other isotherms (Langmuir, Temkin, or Henry's law), the curves fall outside acceptable levels of confidence. The Temkin isotherm requires the relationship between log C $ and Cu to be linear. My data yield a correlation coefficient of r = 0.60. The Langmuir isotherm requires the plot of Cg/C^ vs C to be linear. My data yield a correlation

PAGE 46

41 coefficient of r = 0.39. Henry's law defines a linear relationship over the entire isotherm. The potato-A. tumefaciens adsorption isotherm is nonlinear in the range of 10 ' to I0 5 cells/ml of PBS (Table I). Therefore, none of these three alternative isotherms adequately describes the experimental data for the entire inoculum range studied. When theoretical isotherm plots of bound versus free bacteria for each of the isotherms in Fig. 2 are compared to the experimental data presented in Table I, the empirical correspondence of the data to the Freundlich isotherm plot is very good. Similar to Fig. 2a, the experimental data show a rapid increase in bacterial adsorption per unit increase in initial inoculum density from 10 to 10 . At higher initial bacterial densities, in the range of 10^ to Q 10 cells/ml of PBS, the numbers of bacteria adsorbed are linearly related to the initial inoculum density. At all high bacterial densities, approximately 1.2% of the bacteria adsorb to the potato discs. This simple linear part of the adsorption isotherm is described best by Henry's law: C b = ki X C (Fig. 2c). Saturation as predicted by the Langmuir isotherm (Fig. 2a) is not reached. Linear regression analysis of log C $ vs log C b was used to determine the slope and y-intercept. These values then were used to determine values of the constants k| and k 2 of the Freundlich isotherm over the entire range of bacterial concentrations (10* to I0 9 cells/ml of PBS): C b = k,XC s ^2 When log C b = log k ( + l/k 2 log C s Slope |/k 2 = 0.822 y-intercept log ki =-0.601 kj =0.251

PAGE 47

E E LU i
PAGE 48

43

PAGE 49

44 Thus the Freundlich adsorption isotherm for the adsorption of A. tumefaciens to potato tuber discs is C b = 0.251 XC S l/L2l7 = 0.251 XC S ' 822 . (a) Using equation (a) a predicted curve of log C $ vs log C^ was plotted, and the experimental data points superimposed (Fig. 3). The predicted number of adsorbed bacteria per disc at each inoculum density is listed in Table I. A curve fit to the experimental data points using the Statistical Analysis System (SAS) nonlinear curve fitting procedures (NLIN) generated C b = (0.098) X C s °' 897 . for parameters ki = 0.098 and l/l<2 = 0.897, asymptotic 95% confidence intervals are -0.062 to 0.257 and 0.984 to 0.823 respectively. Clearly, the parameters k| and l/k2 generated from the log C $ vs log C^ plot fall within these confidence intervals. Discussion The binding of cells of A. tumefaciens strain B6 to potato tuber tissue is a reversible equilibrium process. This binding can be described accurately in terms of the Freundlich adsorption isotherm, which is a non-ideal variant of the Langmuir adsorption isotherm (Brunauer 1943). Over the wide range of initial bacterial concentrations examined, non-saturation of available binding sites is in accordance with the Freundlich isotherm. Adsorption of cells of Pseudomonas lachrymans onto leaf surfaces is nonspecific and nonsaturating over inoculum densities of 10 to 10 cells/leaf (Haas and Rotem 1976). Leben and Whitmoyer (1979) also observed that the number of adsorbed P_. lachrymans cells is related linearly to the number of cells applied. Fletcher (1977) suggested that the initial adsorption of marine

PAGE 50

45 LOG C c Fig. 3. The Freudlich adsorption isotherm for Agrobacterium tumefaciens strain B6 on potato tuber tissues. The line is the predicted adsorption curve on log C s vs log C^ generated from equation (a). The experimental data points are from Table 3.

PAGE 51

46 pseudomonads to nonbiological surfaces is driven by nonbiological, physiochemical adsorption forces as described by a modified Langmuir-like adsorption isotherm. In addition, the adsorption of Rhizobium spp. to cereal roots was described by the Langmuir isotherm and hypothesized to be influenced by nonspecific, as opposed to specific binding events (Shimshick and Hebert 1978, 1979). My data similarly point to the importance of nonspecific forces in the adsorption of A. tumefaciens to potato tissue. The experimental data presented here do not lend support to the specific site attachment hypothesis advanced by Lippincott and Lippincott (1969, 1977, 1980). According to these authors, there exist in wounded pinto bean leaves a small, finite number of specific binding sites for tumorigenic A. tumefaciens cells. Heat-treated, tumorigenic A. tumefaciens cells and cells of certain nontumorigenic Agrobacterium sp. strains are hypothesized to be capable of occupying these sites and excluding pathogenic cells from them. Coinoculation studies using unlabeled bacteria in competition with radioisotope-labeled bacteria indicated that cells are not competing for a limited number of specific sites on the plant surface (Pueppke, unpublished observations). The data also indicate that cut surfaces of potato tuber discs contain an extremely large number of nonspecific adsorption sites for A. tumefaciens. Additionally, if a specific site-attachment phenomenon is operative, one would expect bacterial adsorption to be saturable as described by the Langmuir isotherm. This clearly is not the case with adsorption of A. tumefaciens to potato tuber tissue. There are three possible explanations for the discrepancies between the data presented here and those of Lippincott and associates. The first is simply that the adsorption of A. tumefaciens to potato discs is much different than that to pinto bean leaves. Although this is possible, A. tumefaciens infects and

PAGE 52

47 produces tumors on both plants, and similar infection mechanisms are to be expected. The second explanation is that Lippincott and associates may have been misled by their extrapolation of tumor data to the adsorption phenomenon. Many events intervene between initial bacterial adsorption, which occurs rapidly after inoculation, and the appearance of tumors, which occurs 7 to 14 days after inoculation. It is to be expected that some treatments would influence tumor development without having an effect on adsorption. Pueppke and Benny (1984) directly monitored adsorption of A. tumefaciens to potato tissue in the presence of either LPS or various pectic components. Pectic components reduced the numbers of tumors formed on potato and all but one LPS preparation was without effect (Pueppke and Benny 1983). In nearly every case, however, these substances did not reduce bacterial adsorption, suggesting that pectic materials influence some step subsequent to adsorption. The existence of two types of binding sites is a third possible explanation. A small number of highly specific irreversible adsorption sites that lead to tumor formation may coexist with a large number of nonspecific, reversible adsorption sites. The dose-response relationship between inoculum concentration and tumor number is linear at bacterial concentrations from ICr to 10° bacteria/ml but the response saturates at higher bacterial concentrations (Pueppke and Benny 1981). In contrast, the relationship between inoculum concentration and adsorption is linear at bacterial concentrations from ICr to 10° bacteria/ml. Thus, there is a lower threshold below which tumors do not form and an upper threshold above which the tumor response does not keep pace with increases in adsorption. These differences are not surprising, given the complex nature of tumorigenesis relative to that of adsorption and the fact that tumors coalesce

PAGE 53

48 and become unresolvable at bacterial concentrations above 10 bacteria/ml (Pueppke and Benny 1981). The effect of inoculum concentration on the adsorption of rhizobia to soybean roots (Pueppke 1984b) and the subsequent nodulation response (Perkins 1925) appears to be similar. It should be noted that adsorption isotherms are only empirical expressions that reasonably approximate the behavior of experimental binding systems over restricted concentration ranges. Although the Freundlich isotherm was derived for nonbiological systems, the compliance of my data with the necessary assumptions and principles of this model underscores the importance of nonspecific physical forces in A_. tumefaciens adsorption. Although it is conceivable that the good fit of the experimental data to the Freundlich adsorption isotherm is fortuitous, the data justify further examination of bacterial adsorption in terms of nonspecific physical forces. Adherence to a particular adsorption isotherm does not predicate a particular type of binding for the system. Adsorption described by the Freundlich adsorption isotherm, for example, does not imply that a physisorption or a chemisorption process is operative. Rather, the isotherm should permit useful inferences to be drawn and parameters to be set for further testing of bacterial adsorption to plant surfaces.

PAGE 54

CHAPTER FOUR THERMODYNAMICS OF AGROBACTERIUM TUMEFACIENS ADSORPTION TO PLANT SURFACES Introduction The adsorption of bacteria to plant cell surfaces is acknowledged to play an important role in the ecology of prokaryote-plant interactions (Dazzo 1980a, Pueppke 1984a). Before legume nodulation can occur, Rhizobium spp. must adsorb to the root hair surface to initiate the nodulation process (Schmidt 1979, Bhuvaneswari 1981). Bacteria also adsorb quite tenaciously to leaf surfaces (Haas and Rotem 1976, Leben and Whitmoyer 1979). The differential adsorption of Xanthomonas spp. or Pseudomonas spp. to mesophyll cell surfaces may determine whether a disease or hypersensitive response will occur (Klement 1982). Additionally, tumor formation studies imply that site-specific adsorption of A. tumefaciens to plant wounds is required for plant cell transformation (Lippincott and Lippincott 1969). In view of the fact that a piece of the bacterial T-DNA is mobilized and incorporated into the plant nuclear genome, adsorption to the plant cell surface appears to be a logical precondition for transformation. It has been shown that A. tumefaciens adsorbs to potato tissue in very large numbers (Chapter Three this dissertation, Pueppke and Benny 1984) and that adsorption can be described in terms of the Freundlich adsorption isotherm (Chapter Three this dissertation). However, the basic mechanism governing the adsorption of A. tumefaciens is poorly understood. 49

PAGE 55

50 The objective here was to describe the mechanism of A. tumefaciens adsorption in terms of physior chemisorption. Materials and Methods Maintenance of Bacteria and Preparation of Inocula Agrobacterium tumefaciens strain B6 was maintained at 6°C on glutamate-mannitol agar slants. For all adsorption experiments bacteria were washed from 4to 5-day-old slants with phosphate-buffered saline (PBS), centrifuged at 8,000 X G, and resuspended in PBS to the desired bacterial concentration. Bacterial concentration was determined turbidimetrically. Effect of Temperature on Adsorption Adsorption experiments at elevated temperatures were done in a 38°C water bath. Before the adsorption experiment, the bacterial suspension and PBS solutions to be used for rinsing were equilibrated to 38°C in the water bath for 25 min. Potato discs were placed in a covered plastic petri dish, which was floated in the 38°C water bath for approximately 3 min to equilibrate to 38°C. The 6°C adsorption experiments were carried out in a refrigerator, where all washes, and bacterial suspensions were equilibrated to 6°C for 25 min prior to the adsorption assay. Potato discs were placed in a covered plastic petri dish and stored at 6°C for 3 min prior to the adsorption assay. The room temperature experiments were done at 28°C. The adsorption isotherm experiments were repeated three times at each of the three temperatures.

PAGE 56

The entire adsorption assay used here has been described in detail elsewhere (Chapter Three this dissertation). After a 60 min adsorption period, discs were washed three times in consecutive 30-ml portions of PBS, ground in PBS, serially diluted and plated onto yeast extract mannitol (YEM) agar plates. Viable colonies were counted after incubation for 36 to 48 hr at 28°C. At each inoculum density the value reported is the average of three replications. The number of viable colonies in four plates at three different dilutions was tallied to determine the number of bacterial cells adsorbed per potato disc. Kinetic Binding Experiments These experiments employed the adsorption assay described above. However, the binding periods were interrupted at predetermined intervals (0.5, 1,5, 10, 30, and 60 min) and the number of adsorbed bacteria determined. 3askets containing the discs immersed in the bacterial suspension (2 X 10 bacteria/ml) were removed after the prescribed binding period and rinsed as described above. The number of bacteria bound per disc then was determined as above. Desorption Experiments After incubation of the discs with 2 X 10 bacteria/ml for 30 min the discs were removed and quickly washed in sterile PBS for 2 to 3 sec. Each disc then was placed in a 1.5-cm X 12.5-cm test tube containing 2 ml of sterile PBS at 6°C. Aliquots of 100 ul were withdrawn at specific intervals over a 10.5-hr period. At each sample time the disc was gently swirled before the sample was

PAGE 57

52 withdrawn. These aliquots were either plated directly onto YEM agar or serially diluted prior to plating. The number of bacteria adsorbed per disc prior to and after the desorption period was determined as described previously (Chapter Three this dissertation). The experiment was conducted twice with three replicates taken at each sample time. Sequential Adsorption The standard adsorption assay using A^. tumefaciens strain B6 was used with the following modifications. A bacterial suspension containing 2X10' bacteria/ml was used to inoculate one set of discs at the ratio of one disc/ml of bacterial suspension. After removal of the discs, another set of discs was placed into the inoculum. After removal of the second set of discs, the bacterial suspension was used to inoculate a third set of freshly prepared potato tuber discs. Individual discs from all three sets were homogenized, and the number of bacteria adsorbed per disc was determined. The adsorption period was 30 min, and the temperature was 28°C. The sequential adsorption experiment was repeated twice with four replications in each repitition. Zeta Potential Measurements Bacteria were washed three times with sterile, deionized water, centrifuged at 8000 X G for 12 min and resuspended in sterile, deionized water. The bacteria were adjusted to 5 X I0 7 cells/ml of sterile deionized water, and then titrated to the desired pH with KOH or HNOo just prior to measurement of the zeta potential. Each suspension was injected into the electrode chamber of a Laser Zee Meter Model 502, and the focal plane above

PAGE 58

53 the inside bottom of the electrode cell adjusted to 873 um. All zeta potential measurements were taken in this focal plane, the upper stationary phase of the cell. The zeta potential of the bacteria was measured using an electrical field of 75 volts. Two to three samples were measured at-each pH, and three to five measurements of each sample were recorded. After all measurements were made, the sample was decanted and the pH remeasured. Deviation of greater than 0. 1 pH unit was not observed. Due to the relative transparency of the bacteria to the laser beam across the cell, the bacteria were visualized with substage tungsten illumination. A video camera projected the image onto a video monitor where bacterial mobility in the electrical field was observed. Results The adsorption of A. tumefaciens was measured at 6°C, 28°C, and 38°C. The adsorption isotherms obtained at each of these three temperatures, using bacterial concentrations spanning three orders of magnitude, are illustrated in Fig. 4. Least squares regression analysis using the General Linear Model of the Statistical Analysis System was used to compare slopes and y-intercepts of the isotherms. Log-transformed data of both the initial inoculum and number adsorbed were used in the regression analysis, and temperature was considered to be a class variable. Temperature did not significantly alter the y-intercepts of the three isotherms at P 0.05. The slopes, however, were significantly different at P 0.05, but not at P = 0.045. When the same analysis was carried

PAGE 59

54 INOCULUM DENSITY Fig. 4. Isotherms for the adsorption of Agrobacterium tumefaciens cells to potato tissue at three different temperatures. Each point is the mean of measurements taken in three separate experiments with three replications in each experiment. Bars represent the standard errors of the means.

PAGE 60

55 out using the 28°C and 38°C isotherms, neither the slopes nor the y-intercepts were significantly different (Zarr 1974). Comparison of these isotherms using simple linear regression is valid, because at these bacterial concentrations the adsorption isotherm is described by a linear relationship, Henry's law: Cu = kC (Brunauer 1943, Chapter Three this dissertation). Isosteric heats of adsorption were determined using the ClausiusClapeyron equation expressed as AH/RXd(l/T) = dlnC s where AH change in heat content of the system; R = the gas constant, 0.082 l-atm/mole K; T = degrees Kelvin; C $ = [free bacteriaj eq, the concentration of unbound bacteria at equilibrium. In the form of a straight line the expression becomes lnC s = AH/RX l/T. The concentration of free bacteria at equilibrium at a given level of bacterial adsorption for each of the three temperatures examined is plotted against l/T in Fig. 5. In plots of log C $ against l/T, at a constant level of bacterial adsorption, H/R is the slope of a line that allows the heat of adsorption to be derived. The correlation coefficient for each of the four lines plotted in Fig. 5 is greater than r = 0.91. Analysis of variance shows the slopes of this family of lines not to be different from one another and not to be statistically different from zero. This indicates that the heat of adsorption is exceedingly small.

PAGE 61

u u Li_

PAGE 62

57 O VIH313VS ]3HJ "ON 001

PAGE 63

58 Kinetics of Adsorption The number of bacteria adsorbed per disc was measured as a function of the adsorption period, which is defined as the period of time during which the potato discs are immersed in the bacterial suspension. More than 2.8 X 10^ cells bound per disc after incubation for 30 sec (Table 3). The number of bacteria adsorbed to the surface continued to increase with time. However, there is no significant difference between the number of cells bound during a 30 min or a 60 min adsorption period. Desorption Experiments The adsorption of A. tumefaciens to potato tuber tissue can be described by the Freundlich adsorption isotherm (Chapter Three this dissertation). Such a dynamic eguilibrium model implies that the adsorbed bacteria must desorb and undergo a reversible binding. Bacterial desorption was measured over a period of 10 hr at 6°C to reduce interference due to bacterial multiplication (Fig. 6). Bacterial desorption is not linearly related to time. After 4 to 5 hr the desorption curve flattens as eguilibrium is approached, i.e. the rates of desorption and adsorption are equal (Fig. 6). By 10.5 hr approximately 31% of the bound bacteria (5.8 X I0 4 of 1.9 X I0 5 bacteria) had desorbed from each disc. Seguential Adsorption A single bacterial suspension was used to sequentially inoculate three groups of freshly cut potato tuber discs. The mean number of bacteria

PAGE 64

59 Table 3. Time Dependence of the Adsorption of Aqrobacterium tumefaciens B6 to Potato Tuber Tissue. Adsorption Period Number of Bacteria (X I0 5 ) (Min.) Adsorbed per Disc b 0.5 0.29 + 0.06 a 1.0 0.46 + 0.09 b 10.0 1.44 + 0.73 c 30.0 2.60 + 0.28 d 60.0 2.50 + 0.40 d AIL adsorption experiments were carried out with an initial inoculum of 2 X I0 7 cells/ml. Numbers followed by the same letter are not significantly different (p = 0.05) according to Student's t-test.

PAGE 65

60 10 4 ^ 2 -

PAGE 66

61 adsorbed per disc (+ standard deviation) for groups one, two, and three was 9.8 + 4.1 X I0 4 , I.I +0.43 X I0 5 , and 1.5 +0.66 X I0 5 , respectively. There were no significant differences between any of the three groups at the 1% level, when analyzed using Student's t^-test (Fig. 7). Zeta Potential Measurements Zeta potential measurements of bacteria suspended in deionized water indicate an isoelectric point (pi) of approximately 2.9 (Fig. 8). Above pH 2.9 the bacterial cell surface was strongly negative. A zeta potential of greater than -30 mV was observed for values above pH 5.0. Below pH 2.9 the bacterial cell surface was positively charged. However, the magnitude of the positive charge was not as great as that of the negative charge at an equivalent distance above the pi. Discussion Although Aqrobacterium tumefaciens adsorption can be described by the Freundlich adsorption isotherm (Chapter Three this dissertation), this does not imply an adsorption mechanism for the system. The basic effort in this communication was to define the mechanism of adsorption of A. tumefaciens to potato tuber tissue by distinguishing between physisorption and chemisorption. I have attempted to do this by comparing and contrasting my experimental results with the salient features of each of these two basic adsorption types. Bacterial adsorption to potato discs occurs very rapidly and in large numbers at

PAGE 67

62 14

PAGE 68

63 _ -10-20 Fig. 8. Zeta potential measurements. Each point represents the mean of four measurements taken for each of the three samples.

PAGE 69

64 bound per disc. Unless the heat of activation for bacterial adsorption is extremely small, these data suggest that physisorption is operative. Most chemisorbed particles are irreversibly bound and exceedingly difficult to remove (Brunauer 1943). Adsorption of A. tumefaciens, however, is reversible and approaches an equilibrium. Adsorption assays done at 6°C, 28°, and 38°C indicate that temperature does not have a significant effect on the number of bacteria bound. The heat of adsorption as determined with the Clausius-Clapeyron equation is nearly zero, well below the 10-20 kcal value arbitrarily established as the point below which physisorption most likely predominates (Brunauer 1943, Saunders 1971). These data all indicate that the binding of A. tumefaciens can be described in terms of physisorption. If the adsorption process is to be favored, the free energy of the system will decrease as particles are adsorbed onto the surface. This concept is related by the expression; A padh _ y3c yBL ySL JL where a F is the free energy of adsorption, yqs ' s tne bacterium-substratum interfacial tension, ^[_ is the bacterium-liquid interfacial tension, and Yci is the substratum-liquid interfacial tension. This simple thermodynamic model for the adsorption of small particles has been shown to describe the qualitative features of bacterial adsorption (Absolom et al. 1983). This further illustrates that bacterial adsorption can be viewed in thermodynamic terms and related to colloidal adsorption systems. Statistical analysis of data taken over a 32°C temperature range suggests little or no effect of temperature on the actual number of bacteria bound. Ohyama et al. (1979) also found little effect of temperature (24° to 35°C) on the number of A. tumefaciens cells bound to plant suspension culture cells.

PAGE 70

65 Similarly, agglutination of asparagus cells by A. tumefaciens appeared to be independent of temperature (Draper et al. 1983). This would suggest that an ionic interaction is one of the forces involved in the adsorption process, a characteristic of physisorption. Since physisorption appears to be operative in this system, characterization of the ionic nature of the bacterial surface indicate a pi of 2.9 in distilled water. Harden and Harris (1953) gathered data on the isoelectric point of 21 species of Gram-negative bacteria and found pi values ranged from 2.07 to 3.65. At pH values below the pi, the A. tumefaciens surface exhibits a small net positive charge. This may be due to the fact that few ionizable amino groups are present on the bacterial surface. At pH values above the pi, a large number of carboxyl and/or phosphate groups may be ionized and account for the large net negative charge. These data would then suggest that a very high ratio of carboxyl or phosphate to amino groups is present on the bacterial surface. Since the pH of the sap in potato tuber wounds is 5 (unpublished results), which is above the isoelectric point of A. tumefaciens, the bacterial cell is most likely negatively charged within the wound site. However, there are a large number of components within the wound which, when adsorbed onto the bacterial surface, may alter the bacterial charge characteristics. The possible role of these surface charges in adsorption remains to be defined. At inoculum densities that fall within the linear range of the _A. tumefaciens adsorption isotherm, I0 5 to 10° bacteria/ml, a relatively constant 1.2% of the inoculum is adsorbed from the inoculum (Chapter Three this dissertation). This suggests that only a small percentage of the bacterial population may be competent for adsorption. The sequential adsorption

PAGE 71

66 experiments represented in Fig. 7 do not lend support to this hypothesis because similar numbers of _A. tumefaciens cells bound to all three groups of discs. Mathysse and Gurlitz (1982) similarly found no change in the number of bacteria from the bacterial suspension when a 10-fold increase in plant cells was used. Lippincott and Lippincott (1969) have proposed a specific site-attachment hypothesis to describe the adsorption of A. tumefaciens to cells of pinto bean leaves. When the virulent A. tumefaciens strain B6 was added to the wound prior to inoculation with the avirulent IIBNV6 strain, tumor number was not reduced. However, when the order of inoculation was reversed, a reduction in tumor number was noted. This hypothesis suggests a very specific irreversible adsorption of A. tumefaciens strain B6 to pinto bean leaf tissue. The data presented here suggest that A. tumefaciens adsorption is a nonspecific phenomenon governed by physisorption forces. If adsorption was site-specific and irreversible, equilibrium binding as is shown in Fig. 6 would not be expected. Reduced or elevated temperatures also would be expected to strongly influence an irreversable specific site-attachment. However, there may be two separate types of binding sites. One could be site-specific and adsorb bacteria as a prelude to tumor formation. Other sites may nonspecifically adsorb bacteria and not lead to tumorigenesis. Since complete desorption of all bacteria from the potato surface during the 6 hr desorption period at 6°C was not obtained, the possible existence of two types of adsorption sites on potato tuber tissue varying in their affinity for Agrobacterium spp. cannot be eliminated. However, I did not obtain any direct evidence to support the existence of more than one type of adsorption site.

PAGE 72

CHAPTER FIVE EFFECTS OF AGENTS THAT ALTER IONIC AND HYDROPHOBIC INTERACTIONS ON AGROBACTERIUM TUMEFACIENS ADSORPTION Introduction The induction of tumors by A. tumefaciens is thought to require the sitespecific attachment of the bacterium to a plant wound site (Lippincott and Lippincott 1969). Specific attachment was inferred from studies in which plants were inoculated with an avirulent _A. tumefaciens strain prior to or coincident with a virulent A. tumefaciens strain. Changes in tumor number indicated that a single virulent bacterium attached to a single host site, and that such attachment resulted in a single tumor (Lippincott and Lippincott 1969, Whatley et al. 1977). If an avirulent cell occupied the site, the virulent cell was excluded, and a tumor did not form. Unfortunately, attachment was not quantified in these studies. Various host and bacterial molecular components have been examined for their possible role in site-attachment. When bacterial suspensions were mixed with isolated plant cell wall fractions, tumor number was reduced (Lippincott and Lippincott 1977). Several off-the-shelf compounds were tested, and galacturonans were found to inhibit tumorigenesis dramatically. Galacturonans are thought to inhibit tumor formation by binding to the bacterial receptor that mediates attachment to the host. These results led to the hypothesis that a pectic component in the host's cell wall contains the specific attachment site for Agrobacterium sp. Crude lipopolysaccharides (LPS) and bacterial cell 67

PAGE 73

68 envelope preparations from A. tumefaciens have been examined for their participation in specific-site attachment (Whatley et al. 1976). Plant tissue pretreated with the o-antigen region of the LPS reduces the ability of virulent bacteria to form tumors and thus is thought to mediate bacterial attachment to some pectic component in the plant cell wall. Pueppke and Benny (1984) directly measured adsorption of A. tumefaciens to potato tissue. Adsorption was not reduced by galacturonans and in fact was enhanced by citrus pectin. Cell-free LPS from all but one strain did not influence bacterial adsorption. Additionally, LPS did not influence tumor formation in the potato tuber disc assay system (Pueppke and Benny 1983). The heat of adsorption for A. tumefaciens attachment to plant cells is very small, an observation consistent with physisorption (Chapter Four this dissertation). Physisorption implies that adsorption is nonspecific and mediated by ionic and/or hydrophobic forces. Here I provided evidence to support the hypothesized nonspecific nature of A. tumefaciens adsorption. The involvement of ionic and hydrophobic forces in the adsorption process also was tested directly. The ionic nature of the bacterial surface was affected by changes in pH and divalent cations. The possible role of hydrophobic forces and bacterial proteins in adsorption also was tested. Materials and Methods Maintenance of Bacterial Strains The sources of the Agrobacterium spp. and other bacteria used in this study are listed in Table 4. All of the strains were maintained on slants of the defined gluconate-mannitol medium of Bhuvaneswari et al. (1977)

PAGE 74

69 Table 4. Sources of Characteristics of Bacterial Strains. Source Pathogenic on Potato A. tumefaciens 32F

PAGE 75

70 Bacterial Attachment Assay One ml of buffered gluconate-mannitol medium (Meyer and Pueppke 1980), containing 50 uCi of "S-L-methionine, was inoculated with I to 2 X 10 bacterial cells. The culture was incubated for 15 to 16 hr on a rotary shaker at 125 rpm and 27°C. The bacteria were harvested by centrifugation at 10,000 X G for 10 min. The bacterial pellet then was washed three times with PBS. Specific activities ranged from 20 to 150 bacteria per cpm. The final bacterial concentration was adjusted turbidimetrically to 2 X 10 cells/ 10 ul of PBS. Adsorption was measured by the potato tuber disc assay (Pueppke and Benny 1984) which is described here briefly. Potato tubers (Solanum tuberosum L. cultivar Red LaSoda or Red LaRouge) were washed, peeled, surface sterilized in a 20% solution of Clorox for 15 min and rinsed copiously with sterile water. A 8-mm diameter cork borer was used to extract a core of tissue, which was cut into 2-to 3-mm discs. The discs were blotted dry on filter paper and placed on a 10-ul droplet containing 2 X 10 radioisotope-labeled bacterial cells. After a 10 min adsorption period, the discs were washed with three 5-ml portions of PBS. Individual discs were transferred to glass scintillation vials, dried overnight, and digested with perchloric acid and hydrogen peroxide. Five ml of Amersham ACS II Scintillation fluid were added to each vial, and the radioactivity was determined with a Beckman LS-50 liquid scintillation spectrometer. The number of bound bacteria was estimated from the measured radioactivity.

PAGE 76

71 Adsorption of A. tumefaciens to Monocot Tissue Three monocot tissues were examined: I) rhizomes of Smilax bona-nox L. (Liliaceae), 2) aerial tubers of Dioscorea bulbifora L. (Dioscoreaceae) and 3) corm tissue of Gladiolus tristis L. (Iridaceae). Each of these monocot tissues was prepared for the adsorption assay as described above. Dioscorea bulbifera and the Smilax bona-nox were collected near Gainesville, Fl. Gladiolus tristis corm tissue was kindly provided by Beth Logan, University of Florida. Zeta Potential Measurements Bacteria were washed twice with water, centrifuged at 7,700 X G for 10 min, and resuspended in one of the following buffers: 50 mM acetate buffer containing 7.2 g of NaCI/l (final pH 4.5), PBS (pH 7.2), or 50 mM borate buffer containing 7.2 g of NaCI/l (final pH 9.5). Measurements were made as described previously (Chapter Four this dissertation) with a Laser Zee Meter Model 502, set at 75 V. The zeta potential of the bacteria in the upper stationary phase was recorded. Negative Staining for Electron Microscopy Formvar-coated grids were sputter coated with carbon and placed on a bacterial colony growing on agar or floated on a bacterial suspension. After 30 min the grids were removed and washed thoroughly with water. Grids with adsorbed bacteria were stained with 1% phosphotungstic acid for 10 min and excess stain was removed by washing with water. The grids were dried and then viewed with a Hitachi-600 electron microscope.

PAGE 77

72 Proteolytic Treatment of Bacteria Radioisotope-labeled cells of A. tumefaciens strain ACH5 were adjusted to 1.6 X I0 9 cells/ml of PBS. A 12.5-ul droplet containing 2 X I0 7 cells was placed on a hydrophobic petri plate (Fisher No. 8-757-12), to which was added 12.5 ul of protease (Streptomyces qriseus, Sigma Chem. Co.) at I mg/ml or trypsin (Bovine pancreas, Sigma Chem. Co.) at 250 mg/ml and incubated at room temperature for a predetermined time period. A potato disc (9-mm diam, 3-mm thick) then was placed onto the 25-ul droplet for a 10 min adsorption period. The number of bacterial cells bound to the surface was determined as described above. In replicate experiments, the disc size was reduced to 8 mm in diameter and the volume of the droplet was reduced to 20 ul. Influence of CaCl? on Adsorption A suspension of radioisotope-labeled bacterial cells was divided into two equal aliquots, each of which was washed separately. One aliquot was washed three times with 0.15 M NaCI (pH 7.2) and the other with a solution of 0.15 M NaCI (pH 7.2) containing 7.8 mM or 15.6 mM CaCI 2 The CaCI 2 treated suspension was adjusted to 2 X 10' cells/ 10 ul in the CaC^-NaCI solution and used for inoculation as described above. The control suspension was adjusted to 2X10' cells/ 10 ul in the NaCI solution. After the 10 min adsorption period, the discs were washed three times in the buffer originally used to wash the bacteria. Seven separate experiments were conducted with 10 replications in each.

PAGE 78

73 Influence of a Calcium Chelating Agent on Adsorption Potato discs were incubated in 3 mM aqueous ethylenebis(oxyethylenenitrile) tetraacetic acid (EGTA) for 15 min at room temperature, after which the discs were removed and blotted dry on filter paper. Control potato discs were incubated in water for 15 min. The potato discs then were placed on a 10-ul droplet containing 2 X 10 radioisotopelabeled bacteria that had been washed five times with deionized water. After 10 min, the discs were treated as described above, and the number of bound bacteria determined. The experiment was repeated five times, with each experiment containing 10 replications. Agglutination of Bacteria by Divalent Cations Bacteria from 2-day-old cultures on glutamate-mannitol slants were Q washed twice, suspended in PBS, and adjusted to a concentration of 4.4 X 10 cells/ml. Fifty ul of a I M solution of FeSO^, CaC^, or MgSO^ were serially diluted in a two-fold series to a titer of 12 in a polystyrene V-type microtiter plate (Dyntech Laboratories, Alexandria, Va.). Twenty-five ul of PBS Q containing I.I X 10 cells were added to each well such that the final volume was 75 ul. The final concentration of Fe , Ca , or Mg ranged from 670 mM to 0.33 mM. The microtiter plate was agitated gently and then left undisturbed for 3 hr at room temperature. The degree of agglutination was determined visually. The experiment was repeated twice, each with two replications.

PAGE 79

74 Influence of Detergents on Adsorption Bacteria were radioisotope-labeled as described above and washed four times with half-strength PBS. The bacterial suspension then was divided into four equal portions and centrifuged at 10,000 X G for 10 min. One of the four portions was resuspended in half-strength PBS (control) and the other portions resuspended in a half-strength PBS solution of the detergent to be tested. Three detergents were examined; sodium dodecyl sulfate (SDS) (anionic), hexadecyltrimethylammonium bromide (cationic), and Tween-20 (nonionic). The bacteria were incubated at room temperature in the detergent for 15 min, after which the detergent solution containing the bacteria was adjusted to a bacterial concentration of 2 X 10 eel Is/ 10 ul. The bacteria then were used as inoculum in the adsorption assay as described above. This experiment was conducted twice with 10 replications within each experiment. Adsorption of A. tumefaciens to Hydrophobic Polystyrene Bacterial adsorption to polystyrene was measured by three techniques. The first technique has been described previously (Rosenberg 1981), and is briefly outlined here. The bacteria to be examined were streaked onto glutamate-mannitol agar and incubated for 2 to 3 days at 27°C. A flat hydrophobic polystyrene disc, cut from the lid of a petri plate, was pressed against the surface of the bacterial colonies. After 5 to 6 min the plastic disc was removed and vigorously washed under a stream of tap water for approximately 2 min. The bacteria on the plastic surface were fixed by immersing the discs in methanol for 30 sec, after which they were air-dried, and stained with Gentian violet. A hair dryer then was used to dry the stained

PAGE 80

75 plastic surface, containing the adsorbed bacteria. The plastic disc was immersed in tap water to selectively remove the stain from the plastic surface leaving the stained adsorbed bacteria. The intensity of staining was examined visually and the discs given a qualitative rating of ++ = high level of adsorption, + = medium level of adsorption, and = no observable adsorption. In the second technique, bacterial cells were radioisotope-labeled, washed as described previously, and adjusted to a final concentration of 2 X 10 bacteria/ 1 0-ul droplet of PBS. Discs (8 mm diam) were cut from the bottom of hydrophobic petri plates, and the discs were marked to ensure that the bacterial suspension would be applied to the untouched inside surface of the plastic disc. Ten ul of the bacterial suspension were placed on each plastic disc. After a 10 min adsorption period, discs were transfered using forceps to a beaker of water. The discs were agitated for 10 sec to remove both nonand loosely-adsorbed cells. Individual plastic discs were placed in scintillation vials, which contained 0.5 ml of a 4% aqueous solution of SDS. The vials were shaken to remove the adsorbed bacteria. Five ml of scintillation fluid were added to the vials, and radioactivity was determined. To monitor the effect of pH on the adsorption of radioisotope-labeled bacterial cells onto hydrophobic polystyrene surfaces, the bacteria were washed and resuspended in either 50 mM acetate buffer containing 7.2 g of NaCI/l, (pH 4.5), PBS (pH 7.2), or 50 mM carbonate/bicarbonate buffer containing 7.2 g of NaCI/l (pH 9.5). The adsorption assay was carried out as described above. The experiment was repeated twice with 10 replications in each experiment. The Effect of pH on Adsorption to Hydrophobic and Hydrophilic Plastic A third adsorption technique was used to study the effect of pH on adsorption to both hydrophobic and hydrophilic polystyrene (Fletcher 1977).

PAGE 81

76 Bacteria were washed three times with water, centrifuged at 8,000 X G for 10 min, resuspended in either acetate buffer (pH 4.5) PBS, (pH 7.2), or borate buffer (pH 9.5), and adjusted to a concentration of 5 X 10° bacteria/ml. Twenty ml of the bacterial suspension were poured into either a hydrophobic (Fisher No. 8-757-12) or hydrophilic (Corning No. 25020) 9-cm polystyrene petri plate. After 2 hr at room temperature the suspension was decanted, and the plates were gently washed with 30 ml of distilled water. The adsorbed bacteria then were fixed to the plastic surface by drying the surface thoroughly with a hair dryer. A freshly filtered solution of crystal violet solution (Conn 1940) was added until the bottom of the petri dish was covered. After 5 min the stain was decanted and the plates were slowly passed through a stream of running tap water five times to remove the excess stain. The plates were dried thoroughly with a hair dryer. Each petri plate was divided into eight equal sections by marking the edge with a felt pen. Absorbance values for each of the eight sections on the petri plate were obtained at 590 nm with Beckman ACTA CI I spectrophotometer. Results Both monocot and dicot tissues are capable of adsorbing large numbers of cells of A. tumefaciens strain ACH5 (Table 5). The number of bacteria attached to S_. bona-nox tissue after the 10 min adsorption period was significantly greater (P = 0.01) than that to D. bulfifera, G_. tristis or 5. tuberosum tissue. There was no significant difference in bacterial adsorption to D. bulbifera, G_. tristis or S_. tuberosum tissue (Table 5).

PAGE 82

77 There was no significant difference in the capacity of ACH5 cells and any of the seven other bacterial species to bind to potato tissue (Table 6). However a significant difference does exist between two A. tumefaciens strains (ACH5C3 and IIBNV6) and A. radiobacter strain 4718 (Table 6). Included in the list is a fast-growing Rhizobium japonicum strain and Xanthomonas vesicatoria. Neither of these organisms is symbiotic or pathogenic on potato. A. tumefaciens GV3 1 32, A. radiobacter 47 1 8, and A. rhizogenes 1 5834 have been designated as non site-binding avirulent strains (Lippincott and Lippincott 1969). Their attachment to potato tuber tissue, however, was similar to that of the virulent site-binding strain ACH5. Influence of Proteolytic Treatment of Bacteria on Adsorption Preincubation of bacteria with protease for 30 min resulted in a statistically significant increase in the number of ACH5 cells bound relative to controls (Table 7). Trypsin treatment had no significant effect. _A. tumefaciens B6 cells pretreated with either trypsin or protease bound in significantly greater numbers than did cells in the control. Influence of Detergents on Adsorption All three detergents influenced bacterial adsorption to potato tissue, but the ionic character of the detergent does not appear to be an important factor. Tween-20 at concentrations of 7%, 0.4%, and 0.04% (v/v) inhibited

PAGE 83

78 Table 5. Adsorption of Agrobacterium tumefaciens ACH5 to Monocot and Dicot Tissue. Plant Species Number of Bacteria Bound per Disc (X 1 5 )* Smilax bona-nox Dioscorea bulbifera Gladiolus tristis Solanum tuberosum 86.0+ 18.2 a Monocot 18.3 + 09.0 b Monocot 16.0 + 06.7 b Monocot 8.3 + 02.3 b Dicot The values are means (+1 standard deviation) of three experiments, each of which contained 10 replications. Numbers followed by the same letter are not significantly different (P = 0.05) according to Duncan's Multiple Range Test.

PAGE 84

79 Table 6. Adsorption of Bacteria to Potato Tuber Tissue Bacterial Species Average Number of Bacteria Bound per Disc (X I0 5 ) Agrobacterium tumefaciens ACH5 5. 10 +_ 2.8 ab* Agrobacterium tumefaciens ACH5C3 9.76 _+ 1.7 a Agrobacterium tumefaciens 11BNV6 8.98 +_ 1.6 a Agrobacterium tumefaciens GV3I32 6.41 +_ 1.0 ab Agrobacterium rhizogenes 15834 4.40^2.1 ab Agrobacterium radiobacter 4718 1.70^0.4 b Rhizobium japonicum USDA 201 7.85 +_ 1.5 ab Xanthomonas vesicatoria 71-21 6.88 + 2.6 ab The values are means (+ 1 standard deviation) of two experiments, each of which contained 10 replications. Numbers followed by the same letter are not significantly different (P = 0.05) according to Duncan's Multiple Range Test.

PAGE 85

80 < 02 O 3 c « § .2 c £ re 'jZ

PAGE 86

bacterial adsorption by 56%, 41% and 20%, respectively. In the presence of 0.05% Tween-20, adsorption was reduced from 2.8 + 0.4 X I0 6 to 1.87 + 0.4 X I0 6 cells/disc, a 35% reduction. Adsorption was reduced from 2.80 + 0.5 X I0 6 cells bound/disc in the control to 0.95 + 0.3 X I0 6 cells bound/disc by 0.05% SDS. This is a 66% reduction in the number of cells bound. Adsorption was reduced by 35%, from 2.8 + 0.5 X I0 6 to 1.82 +_ 0.27 X I0 6 bacteria per disc, in the presence of 0.5% hexadecyltrimethylammonium bromide as compared to the control. Tween-20 and SDS did not reduce bacterial colony forming units (CFU) at the concentrations studied, whereas 0.05% hexadecyltrimethylammonium bromide reduced bacterial CFU. Adsorption of Agrobacterium spp. to Hydrophobic Polystyrene Ten strains of A. tumefaciens, one A. rhizogenes strain, two A. radiobacter strains, and one A. rubi strain were screened semiguantitatively for their ability to adsorb to hydrophobic polystyrene (Rosenberg 1981). Both virulent and avirulent Agrobacterium spp. adsorb to hydrophobic polystyrene (Table 8). Significantly, strain B2R, a spontaneous mutant of B2F, is able to adsorb to hydrophobic plastic, but B2F is not. Influence of pH on Bacterial Adsorption to Plastic There was no significant effect of pH on adsorption of -"S-labeled A_. tumefaciens ACH5 to hydrophobic plastic. The numbers of cells bound per disc were 1.3 + 0.45 X I0 6 , 1.2 + 0.6 X I0 6 , 0.95 + 0.3 X I0 6 at pH 4.5, 7.2, 9.5, respectively. When measured photometrically, the adsorption to hydrophobic

PAGE 87

82 Table 8. Adsorption of Aqrobacterium spp. to Hydrophobic Polystyrene. Aqrobocterium spp. Adsorption* ++ ++ ++ ++ + + A.

PAGE 88

33 .125 .10 .075 .05 .025 O o HYOROPHILIC • m HYDROPHOBIC 4-5 PH 9-5 Fig. 9. Adsorption of Agrobacterium tumefaciens strain ACH5 to hydrophilic or hydrophobic polystyrene as a function of pH. Bars represent +l standard deviation.

PAGE 89

84 plastic also was not significantly influenced by pH (Fig. 9). Adsorption to hydrophilic plastic, however, was significantly greater at pH 4.5 than at pH 7.2 or pH 9.5 (Fig. 9). Adsorption of A. tumefaciens GV 3132 to potato tuber discs was optimal at pH 7.2 (6.5 +_ 0.95 X 10 bacteria/disc) and significantly reduced at both pH 4.5 (1.72 + 0.38 X I0 5 bacteria/disc) and pH 8.5 (2.1 +0.75 X I0 5 bacteria/disc) according to the Student's Jj-test. Effect of EGTA Treatment Pretreatment of potato discs with 3 mM EGTA did not influence their ability to adsorb radioisotope-labeled bacteria. Discs treated with EGTA were compared with PBS control discs and analyzed using the Student's t-test. Onesided tests at the P = 0.05 level indicate that treatment of the potato discs with EGTA does not have a significant effect on bacterial adsorption. In these experiments pretreated EGTA discs bound 1.41 _+ 0.56 X 10 bacteria per disc, and control discs bound 1.54 +_ 0.80 X 10 bacteria disc (+1 standard deviation). Effect of Divalent Cations on Agglutination A. tumefaciens ACH5 cells are agglutinated by CaC^. The dilution end point is 5.2 mM. The greatest agglutination of ACH5 occurred at a CaCU concentration of 83 mM. FeSO^ exhibited a tremendous ability to agglutinate A_. tumefaciens. Agglutination was evident in the last well, which contained 0.33 mM FeSO^. The greatest degree of agglutination was found in the presence of 5.2 mM FeSGv.

PAGE 90

85 Even after 6 hr in the presence of I M MgCU, agglutination of ACH5 was slight. However, in the presence of either CaCIo or FeSO^ agglutination was very rapid and nearly complete within I hr, and stable overnight at room temperature. Control bacteria suspended in either water or PBS did not agglutinate. Effect of CaC^ on Adsorption to Plant Tissue Binding of strain ACH5 to potato tissue was not appreciably affected by 7.8 mM or 15.6 mM CaC^ (Table 9). Comparison of CaC^-treated to nontreated controls by the Student's t-test showed no significant difference in the number of bacteria bound. However, in six of the seven experiments, potato discs consistently bound fewer CaCU-treated bacteria. The CaC^ concentrations used in these experiments fall near the dilution end point of agglutination. Therefore, agglutination of the bacteria during washing and the subsequent binding period should have been minimal. Zeta Potential Measurements The zeta potential of bacteria suspended in the borate buffer (pH 9.5) is -I I mV. Bacteria suspended in PBS have a zeta potential of approximately -2.5 mV. In acetate buffer (pH 4.5) the zeta potential is slightly positive with a value of +0.1 mV. It should be emphasized that zeta potential determinations are more variable at or near the isoelectric point of the particle. Therefore

PAGE 91

86 Table 9. Effect of CaC^ on Agrobacterium tumefaciens Adsorption. Cells bound (X 1 6 ) Statistical Experiment CaC^ (mM) CaC^ Control Significance 7.8

PAGE 92

87 the values given for measurements at pH 4.5 and pH 7.2 are not necessarily the absolute zeta potentials of the bacteria suspended in either of those buffers. Negative Staining The result of negative staining cells of strain ACH5 with phosphotungstic acid is shown in Fig. 10. Bacteria from either colonies on agar plates or suspension cultures gave similar results (Fig. 10). Discussion The transfer of the T-DNA from Agrobacterium spp. into the host cell most likely requires the attachment of the bacterium to the plant cell surface. The bacterial o-antigen and the pectic materials in the host cell wall have been hypothesized to mediate such bacterial adsorption (Lippincott and Lippincott 1969, Whatley et al. 1976, Lippincott et al. 1977b, Rao et al. 1982). These data, however, should be viewed with caution, because bacterial adsorption was not measured directly (Lippincott and Lippincott 1980). Pueppke and Benny (1981, 1983) have shown that extrapolation from tumor formation to bacterial adsorption may not be justified. Additionally, Pueppke and Benny (1983) used the potato tuber disc assay to show that neither tumor formation nor A_. tumefaciens adsorption was inhibited by cell-free LPS. In contrast, inoculation of pinto bean leaves in the presence of cell-free LPS substantially reduced tumor formation (Whatley et al. 1976). The site attachment hypothesis predicts that the relative resistance of most monocotyledonous plants to Agrobacterium spp. is a conseguence of their

PAGE 93

. 38 Fig. 10. Negatively stained Aqrobacterium tumefaciens strain ACH5. X 10,000.

PAGE 94

89 inability to adsorb Aqrobacterium spp. (Lippincott and Lippincott 1969, Lippincott et al. 1977b, Lippincott and Lippincott 1978a). However, adsorption of virulent, site-binding A. tumefaciens cells to S. tuberosum, was not significantly different (P = 0.05) than that to two monocots, D_. bulbifera, and G. tristis. However, the number of bacteria bound to the S_. bona-nox tissue was significantly greater than that bound to any of the other species. Ohyama et al. (1979) have shown that suspension culture cells of a variety of monocots adsorbed large numbers of A. tumefaciens cells. In addition, suspension culture cells of soybean (Glycine max ) bound a large number of A. tumefaciens cells, even though soybean apparently is resistant to A. tumefaciens infection (De Cleene and De Ley 1976). Draper et al. (1983) have shown that suspensions of cells of another monocot, asparagus, are agglutinated by A. tumefaciens cells. A series of virulent and avirulent Aqrobacterium spp. adsorb in statistically similar numbers to potato tuber tissue. The adsorption of all of these species is similar to that of A. tumefaciens ACH5 (Table 6). Additionally, the adsorption of Rhizobium japonicum USDA 201 and Xanthomonas vesicatoria 71-21 were not significantly different from that of strain ACH5 (Table 6). Divalent calcium did not statistically alter the number of bacteria bound to wounded tissues. Additionally, when discs were treated with the calcium chelating agent EGTA prior to bacterial inoculation, no alteration of bacterial adsorption was observed. This suggests that the availability of Ca + is not important to bacterial adsorption. Ohyama et al. ( i 979) reported somewhat similar results using suspension cultured cells. In light of a report suggesting that LPS and galacturonans do not participate in A. tumefaciens adsorption to susceptible potato tissue (Pueppke

PAGE 95

90 and Benny 1984), the actual molecular components of A_. tumefaciens adsorption remain unclear. A. tumefaciens adsorption, however, may be described in terms of physical adsorption, which is governed by ionic and hydrophobic forces (Chapters Three and Four this dissertation). The adsorption of A. tumefaciens ACH5 to hydrophobic plastic was not altered as a function of pH. However, adsorption to hydrophilic plastic was optimal at pH 4.5, and decreased with increasing pH (Fig. 9). When measured as a function of pH, adsorption of A. tumefaciens GV3I32 to potato tuber tissue was optimal at pH 7.2, and significantly decreased at either acidic or basic pH. Similarly, Pueppke and Benny (1984) observed that the adsorption of _A. tumefaciens ACH5 to potato tuber tissue was optimal at pH 7.2 and reduced at either pH 4.5 or 9.5. Zeta potential measurements of the bacteria suspended in the same buffers, show that the surfaces of the bacterial cells are highly negatively charged at pH 9.5 (borate), are neutral to slightly negative at pH 7.2 (PBS), and are slightly positive at pH 4.5 (acetate buffer). As the pH approaches the pi of A. tumefaciens cells, a reduced surface charge may facilitate a closer approach to the host's surface. Attachment then may be facilitated by cellular appendages, such as the flagellae shown in Fig. 10. These could help overcome the remaining electrostatic repulsion barrier between the bacterium and the plant cell (Ottow 1975). Matthysse et al. (1982) illustrated the adsorption of A_. tumefaciens in large numbers to the cell membranes of tobacco and carrot protoplasts that were devoid of cell wall material. The importance of hydrophobic interactions in the adsorption process that leads to tumor formation remains to be determined. It is clear, however, that not all Aqrobacterium spp. adsorb to hydrophobic plastic similarly, as measured according to Rosenberg (1981) (Table 8), and yet are able to adsorb to potato tuber tissue (Table 6).

PAGE 96

The zeta potential of bacteria suspended in deionized water at pH 4 or greater is highly negative (Chapter Four this dissertation), yet in acetate buffer at pH 4.5, the zeta potential is slightly positive. The descrepancy can be explained as follows. Bacteria suspended in the acetate buffer pH 4.5 adsorb cations from the solution onto their surfaces. This alters their behavior in an electrical potential and results in a less negative zeta potential. The effect is even more dramatic with bacteria suspended in PBS at pH 7.2. In water (low ionic strength) the bacteria normally would be negatively charged at pH 7.2. However, in PBS the negative bacterial surfaces adsorb cations from the solution, and this substantially reduces the negative zeta potential. Outer membrane proteins exposed to the environment mediate hydrophobic interactions in several mammalian systems (Berkeley et al. 1980, Bitton and Marshall 1980). Pretreatment of A. tumefaciens with either trypsin or protease to digest exposed protein components resulted in increased numbers of bacteria bound. Consequently, the presence of intact outer proteins may in some way inhibit adsorption. Treatment of the bacteria with proteolytic enzymes presumably degrades external protein components, outer membrane proteins, pili, and flagellae, all of which may allow a closer approach of the bacterium to the host's cell surface. At this point the much shorter range of ionic forces comes into play, perhaps accounting for the increased adsorption observed after proteolytic digestion (Table 7). Alternatively, the enzymes themselves may adsorb tenaciously to the bacterial surface, enhancing the adsorption capacity of the bacteria. At a concentration of 0.05%, all three detergents substantially reduced bacterial adsorption. The sensitivity of adsorption to the presence of detergents suggests that hydrophobic interactions play a role in A. tumefaciens

PAGE 97

92 adsorption to potato tissues. Pili are a likely external component of the bacterium that could be involved in hydrophobic interactions. For example, gonococcal pili have been shown to be very hydrophobic with approximately 46% of the amino acid resides non-polar. This causes the pili to exist in the aggregated form ]n_ vitro. Characterization of Agrobacterium spp. pili and examination of their involvement in the adsorption process will certainly enhance our understanding of the attachment process. These results, in addition to the fact that the genus Agrobacterium has the largest host range of any known phytopathogenic prokaryote, do not lend support to the site-attachment hypothesis, which describes a highly specific adsorption of Agrobacterium spp. to the surface of plant tissue. Rather, a nonspecific adsorption mediated by physical adsorption forces appears to be operative in the potato tuber disc assay system. Adsorption to plant and plastic surfaces is pH-sensitive. Anionic, cationic, and neutral detergents significantly reduce the number of A. tumefaciens adsorbed to plant cell surfaces. These results strongly indicate that both ionic and hydrophobic forces are involved in the non-specific adsorption of A. tumefaciens to plant cell surfaces.

PAGE 98

CHAPTER SIX SCANNING ELECTRON MICROSCOPY OF POTATO TUBER TISSUE INOCULATED WITH AGROBACTERIUM TUMEFACIENS Introduction The etiological agent of crown gall disease Agrobacterium tumefaciens , causes a tumor-like growth on many dicotyledonous plants (De Cleene and De Ley 1 976). Once in a wound site on the plant, A. tumefaciens cells are thought to adsorb to a surface component of the host cell. A piece of the bacterial Tiplasmid, known as the T-DNA, is subsequently transferred to the host, where it is stably incorporated and expressed (Watson et al. 1975, Chilton et al. 1977, Gurley et al. 1979, Thomashow et al. 1980). The method of DNA transfer, however, is unknown. Plant protoplasts have been transformed by both intact A_. tumefaciens cells and isolated Ti-plasmid (Marton et al. 1979, Wullems et al. 1981, Krens et al. 1982, Hanold 1983). It is possible that a similar situation occurs in a plant wound site, where the partial removal of cell wall material could expose the intact cytoplasmic membrane. This may facilitate the transfer to T-DNA out of the bacterium into the wounded host cell. A second possibility is that of bacterial-plant conjugation (Roberts and Kerr 1974, Kerr 1975). A third alternative is that intact bacteria penetrate viable plant cells and then transfer T-DNA. This idea has been rejected by some (Stonier 1956, Schilperoort 1969) and proposed as the essential step in tumorigenesis by others (Sigee et al. 1982). 93

PAGE 99

9h The purpose of this research was to examine the infection of potato tissue by A. tumefaciens at the ultrastructural level. An avirulent, heat-cured strain was included for comparison. Materials and Methods Bacterial Culture, Maintenance, and Inoculum Preparation Agrobacterium tumefaciens strains ACH5, ACH5C3, and B6 were maintained at 4°C on slants of the defined gluconate-mannitol medium of Bhuvaneswari et al. (1977). Inocula were prepared by streaking A. tumefaciens cells onto gluconate-mannitol slants. After 2 to 3 days at 27°C, the cells were washed from the slants with sterile PBS, centrifuged for 10 min at 7700 X G, and washed twice with PBS. The bacteria then were resuspended in PBS and adjusted turbidimetrically to a concentration of 2 X 10 cells/ml. Preparation of Protoplasts Suspension culture cells derived from potato (Solanum tuberosum cv Red LaSoda) tubers were maintained in Murashige and Skoog media (Gamborg and Wetter 1975). Cells were harvested and suspended in an equilibration medium consisting of B5 salts (Gamborg and Wetter 1975), 30 mM MES (2 Nmorpholino ethanesulfonic acid), and 300 mM sorbitol, adjusted to pH 5.6. This suspension was incubated at 150 rpm at room temperature. After equilibration for I hr, the cells were harvested by centrifugation at 1000 X G for 10 min and resuspended in an enzyme mixture, where they were incubated for 4-6 hours at room temperature at 150 rpm. The enzyme solution contained 250 mg of

PAGE 100

95 cellulose (Calbiochem, La Jolla, CA), 50 mg of Macerase (Calbiochem), 250 mg of Rhozyme (Rohm and Haas), and 50 mg of Driselase (Sigma) dissolved in 10 ml of the eguilibration medium used above. After enzyme digestion the cell suspension was centrifuged at 30 X G rpm for 10 min and the protoplasts removed from the surface of the enzyme solution. The protoplasts were then washed three times with the equilibration medium. Inoculation Procedure Potato tuber discs were inoculated with A^. tumefaciens as described by Pueppke and Benny (1981). Peeled potato tubers are surface sterilized for 10 min, and 8-mm-diameter cores of tuber tissue are removed. Discs, 2 to 3 mm thick, are sliced from the cores and immersed in a suspension containing 2 X I0 7 cells of strain ACH5 or ACH5C3/ml of PBS. A ratio of I ml of bacterial suspension per disc was maintained. After 10 min the discs were removed, rinsed for 2 or 3 seconds in sterile PBS, blotted dry, and placed on 1.5% water agar plates and stored in the dark at 28°C. At specified times after inoculation, discs were removed for scanning electron microscopy or the determination of bacterial colony forming units. The number of bacteria colonizing the potato disc was determined as described previously (Chapter Three this dissertation). Isolated protoplasts were inoculated with A. tumefaciens strain B6 at a concentration of 2 X 10 cells/ml. After incubation for 3 hr the inoculation mixture was filtered through Whatman No. I filter paper, and the protoplasts rinsed with two volumes of equilibration buffer. The filter paper, which then contained the protoplasts and bound bacteria, was removed and prepared for scanning electron microscopy.

PAGE 101

96 Fixation for Scanning Electron Microscopy Discs were placed in 0.1 M sodium cacodylate buffer (pH 7.2) containing 2% glutaraldehyde and fixed overnight at 6°C. Samples then were rinsed three times with 0.1 M cacodylate buffer and post-fixed overnight at 6°C in 1% OsOv The discs then were rinsed and sequentially dehydrated in solutions of 25%, 50%, 75%, 95%, and 100% ethanol. Samples were transferred into fresh 100% ethanol, which was changed twice, and then dried in a CPD 010 critical point drying apparatus using CC>2 as the exchange fluid. Specimens then were sputter coated with gold and viewed with a Hitachi S-450 scanning electron microscope using an accelerating voltage of 20 KV. Discs were fixed for scanning electron microscopy at time intervals of 0, I, 2, 6, 8, 19, and 24 hr post-inoculation and then every 24 hr up to 12 days after inoculation. To prevent lysis of the protoplasts during fixation, 2% glutaraldehyde and 1% 0s0a were prepared in the equilibration medium. Three extra steps in the ethanol dehydration series (35%, 65%, and 85%) were included to minimize artifacts caused by rapid dehydration. Results The cut surface of potato tubers provides a very heterogenous substrate for adsorption of A. tumefaciens cells. Cut cell wall fragments often create cavities, the bottoms of which may be covered with cytoplasmic membranes or tonoplasts of the ruptured cells or the outer walls of the cells beneath. Granules of various sizes are found dotting the surface. A. tumefaciens cells

PAGE 102

97 are bound to broken cell wall fragments, membranous components, starch granules, etc., with no apparent affinity for any particular component (Fig. 15, 16, 17). The adsorption of A. tumefaciens cells to the potato disc surface was examined during two different time courses in these experiments. The first time course extended from inoculation to 2k hr post-inoculation and is termed the early events time period. Although the bacterial inoculum contained only normal rod-shaped bacteria of approximately 2 to 3 urn X 0.7 urn in size (data not shown), bacterial rods are rarely found on the surfaces of inoculated potato discs until 6 to 8 hr post-inoculation. Instead, spherical structures, 0.5 to 1.5 urn in diameter, are present on the disc surface (Fig. I I, Fig. 12). A field of view approximately 900 unrr was maintained when scanning the disc surface. Assuming random distribution of the 2 X I0 5 bacteria that adsorbed per disc (8 mm dia X 2.2 mm thick), 1.3 bacteria should be in each field. Twenty to thirty fields, however, were examined before a single rod-shaped bacterium was observed in samples harvested before 6 hr post-inoculation. The bacteria! rods were apparent in nearly every 900-um 2 field of samples taken 6-8 hr after inoculation (Fig 13). The second time course, termed the late events time period, compared the behavior of pathogenic and nonpathogenic A. tumefaciens strains on inoculated potato tuber tissue I to 12 days after inoculation. Few differences are observed in the surface ultrastructure of discs inoculated with virulent A. tumefaciens ACH5 or the heat-cured, avirulent derivative, ACH5C3. Most noteworthy is the appearance of tumors at day 5 on tissues inoculated with ACH5 (Fig 22), but not on those inoculated with ACH5C3. Both strains formed

PAGE 103

98 microcolonies. Incipient microcolonies are first noted 8 hr after inoculation (Fig. 13) and develop into dome-shaped structures within 19 hr (Fig. 14). The microcolonies initially are small and appear to continually increase in size. New microcolonies appear to be initiated over a long time period, because small domed colonies are present in samples harvested on day 12. Based on data from plating experiments bacterial multiplication on each disc is roughly equivalent for both strains and increases during an 8-day time period (Table 10). These colonies consist of densely packed bacteria apparently held together by some type of matrix (Fig. 15, 16, 17). For example, in Fig. 15 approximately 300 bacteria are tightly packed into a dome-shaped structure 6.5 urn in diameter. Unique microcolony shapes are not associated with any particular stage of colony development. Microcolonies often are dome-shaped, but they also can be elongate and span 40-50 um (Fig. 15, 16, and 17). Fig. 18 illustrates another point of interest. Rod-like structures appear to originate from the surfaces of both virulent and avirulent A. tumefaciens cells and to anchor individual bacterial cells to plant cell surfaces. These structures are seen as early as 24 hr and as late as 12 days post-inoculation. However, cellulose microfibrils were never observed in this study. The adsorption of A. tumefaciens strain B6 to protoplasts from potato suspension cultures also was examined (Fig. 19, 20, 21a, 21b). Although agglutination of the plant cells had occurred at the time of fixation, 3 hr postinoculation (Fig. 21a), cellulose microfibrils are not evident, and microcolony formation is not observed at this early time after inoculation. Most noteworthy, however, is the predominance of polar attachment, which is a rare event on potato tuber tissue (Fig. 19, Fig. 20).

PAGE 104

99 Table 10. Bacterial Colonization of Potato Tuber Discs. Time After

PAGE 105

100 Fig. II. Potato disc inoculated with Agrobacterium tumefaciens strain ACH5 (10 min after inoculation). Note the presence of a large number of spherical structures approximately 0.9-1.5 um in diameter. No rod-shaped bacteria are present in this field of /ew. Three large starch particles are illustrated, and the central particle ii ruptured. X 4,000.

PAGE 106

Fig. 12. Potato disc inoculated with Agrobacterium tumefaciens strain ACH5 (1.5 hr after inoculation). The plant cell surface appears to be rippled and contains material that may be remnants of a membrane (MR). One rod-shaped bacterial cell is present (single arrow). Double arrows identify short rods and spherical structures that appear to be pleomorphic bacteria. X 7,600.

PAGE 107

102 Fig. 13. Potato disc inoculated with Agrobacterium tumefaciens strain ACH5 (8 hr after inoculation). Three rod-shaped bacteria are in the initial stages of microcolony formation. Note the ring of material around and on the bacteria. This material probably consists of dehydrated polysaccharide. X 34,000.

PAGE 108

103 Fig. 14. Potato disc inoculated with Agrobacterium tumefaciens strain ACH5 (19 hr after inoculation). Ten to twelve bacteria are forming an incipient microcolony (upper center of photograph). X 8,500.

PAGE 109

104 Fig. 15. Potato disc inoculated with Aqrobacterium tumefaciens strain ACH5C3 (3 days after inoculation). Note the different morphology of the microcolonies: an elongated colony on the right and a tightly packed dome-shaped colony on the left of the photograph. X 5,300.

PAGE 110

I OS Fig. 16. Potato disc inoculated with Aqrobacterium tumefaciens strain ACH5 (8 days after inoculation). Both elongated and dome-shaped colonies are present. The elongated microcolony in the upper right corner (arrow) appears to be ruptured. The bacteria are also colonizing the area around the two starch grains (SG) located below the ruptured colony. Colonies of the size and shape seen in this field of view dot the entire surface of the disc. X 1,800.

PAGE 111

106 Fig. 17. Potato disc inoculated with Aqrobacterium tumefaciens strain ACH5C3 (4 days after inoculation). Note the torn wall fragment (WF) in the upper right and the adsorbed bacteria (arrow). Microcolonies and many single adsorbed bacteria are scattered across the surface. X 3,200.

PAGE 112

107 Fig. 18. Potato disc inoculated with Aqrobacterium tumefaciens strain ACH5 (7 days after inoculation). Note the rod-shaped structures (arrows) connecting the bacteria to the potato disc surface and to each other. X 7,600.

PAGE 113

108 Fig. 19. Suspension culture potato protoplast inoculated with Agrobacterium tumefaciens strain B6 (3 hr after inoculation). The polar attachment viewed here was common. Recumbent attachment of the bacterium to protoplasts was observed infrequently. Note the absence of cellulose microfibrils. X 8,800.

PAGE 114

109 Fig. 20. Suspension culture potato protoplasts inoculated with Agrobacterium tumefaciens strain B6 (3 hr after inoculation). Note polar attachment of bacteria to the protoplast surface and the absence of cellulose microfibrils. X 6,400.

PAGE 115

Fig. 21. Suspension culture potato protoplasts inoculated with Agrobacterium tumefaciens strain B6, 3 hr after inoculation, a) Strings of bacteria have tied the protoplasts together resulting in agglutination. Note the polar attachment of the bacteria to the protoplast. X 2,500. b) Enlargement of the bacterial string shown in Fig 21a (boxed area). Note the absence of cellose microfibrils. X 6,200.

PAGE 117

112 Fig. 22. Potato disc inoculated with Agrobacterium tumefa ciens strain ACH5 (5 days after inoculation). Lateral view of tumor fills the photograph. X 400.

PAGE 118

13 Discussion The initial step in tumor formation is thought to be the adsorption of _A. tumefaciens cells to the .host cell surface (Lippincott and Lippincott 1969). Such adsorption may facilitate the transfer of the T-DNA (Lippincott and Lippincott 1969, Matthysse et al. 1981, Matthysse 1983), and thus adsorption may be a regulatory step in tumorigenesis. The basis of the resistance of monocots to A_. tumefaciens is thought to lie in the inability of A. tumefaciens cells to adsorb, thereby preventing the transfer of the T-DNA (Lippincott and Lippincott 1978, 1980). Subseguently, however, workers have shown by direct measurement that both tissue and suspension culture cells from monocotyledonous plants are able to adsorb Agrobacterium spp. in substantial numbers (Ohyama et al. 1979, Chapter Three this dissertation). Additionally, equivalent numbers of both avirulent and virulent Agrobacterium spp. bind to potato tissue (Chapter Three this dissertation). Therefore, it was of interest to examine not only adsorption, but also post-adsorption events that might differentiate avirulent from virulent A. tumefaciens strains. Parallel time-course experiments with the virulent strain ACH5 and its heat-cured avirulent derivative gave similar results. Cellulose microfibrils, which have been reported to anchor cells of A_. tumefaciens strain A6 to suspension culture cells and protoplasts of carrot (Matthysse et al. 1981, Matthysse 1983) were not seen. The cellulose microfibrils appear within 2 hr of coincubation with carrot protoplasts and are suggested to participate in the firm attachment of large numbers of bacteria to the host cell (Matthysse 1983). Similarly, Spiess et al. (1977a, 1977b) observed a dense fibrillar matrix, apparently of bacterial origin, that entrapped A. tumefaciens cells bound on the surface of moss (Pylaisiella selwynii) .

PAGE 119

114 After a 3-hr coincubation of potato tuber protoplasts with A. tumefaciens strain B6, massive agglutination of the protoplasts had occurred, even though microfibrils were not observed. Two features may account for the absence of microfibrils in this system, and their presence on carrot protoplasts. The host tissue and A_. tumefaciens strains are different in each of these experiments, as is the inoculation procedure. The inoculation period for potato tuber tissue was 10 min, after which the discs were rinsed, blotted dry, and placed on water agar plates. In both the carrot suspension cell and moss cell systems, coincubation lasted 2 to 19 hr. Most A^. tumefaciens cells were oriented perpendicularly on the protoplast surface, but not on the potato tuber surface. Habituated tobacco callus cells also adsorbed A. tumefaciens in a polar orientation (Smith and Hindley 1978). Such an orientation, however, is of questionable significance. Members of the closely related genus, Rhizobium, adsorb perpendicularly to all sorts of surfaces (Pueppke 1984). Marshall and Cruickshank (1973) and Marshall et al. (1975) observed that rhizobia orient themselves at right angles to oil-water interfaces, to blocks of Spurr's medium, and to legume root hair surfaces. This suggests that polar attachment of bacteria to surfaces may be nonspecific. During the early time course, from time to 6 hr after inoculation, normal bacterial rods were extremely difficult to find on the surfaces of potato discs. Spherical structures with dimensions similar to those of bacteria were common at this time. By 6 hr post-inoculation normal rod-shaped bacteria were observed in nearly every field of view. This suggests that the bacteria may be pleomorphic after introduction into a plant wound site. Fixation artifacts are not likely to be responsible for the pleomorphic structures, because discs harvested 12 hr after inoculation and at time were fixed and

PAGE 120

dried for SEM simultaneously and under identical conditions; normal rod-shaped bacteria were abundant in the former, but very rare in the later. Identification of the spherical particles is complicated by the presence of many spherical starch grains in the preparations. A^. tumefaciens cells are, however, capable of forming spherical structures under certain conditions. Cells of A. tumefaciens strain C58 became spherical after incubation for 2 hr in the presence of Agrocin 84, a bacteriocin thought to alter the integrity of the bacterial cell wall (Smith and Hindley 1978). Pleomorphism is well documented for the closely related bacterial genus, Rhizobium (Tu and Braybrook 1981, Kaneshiro et al. 1983). Recently, sphere-shaped pleomorphs of the normally rod-shaped Azotobacter vinelandii also were found in soil (Lopez and Vela I 98 1 ). The late time course experiments revealed the presence of microcolonies on the surface of the potato disc. The bacterial microcolonies were quite variable in size throughout the entire 12-day time course, and appeared to be initiated at different times during the 12-day period. This is consistent with the observation that both ACH5 and ACH5C3 colonize the disc at equivalent rates until well past the 5th to 6th day after inoculation, when tumors are visible. Each colony appeared to be surrounded by an electron-lucent coating. Although the exact nature of this coating is unknown, it is morphologically similar to polysaccharide material termed the glycocalyx (Costerton et al. 1981). This is of interest because the glycocalyx coating is thought to function as an ion exchange medium, possibly adsorbing materials such as antibiotics from the surrounding environment and affording the bacterium protection (Costerton et al. 1981). Tumor formation on pinto bean leaves is dramatically recuced when rifampin is introduced into the wound site up to 5 hr after

PAGE 121

and Heberlein 1975). Sensitivity to rifampin is progressively diminished when more than 5 hr elapse between inoculation and antibiotic application. Microcolonies with glycocalyx-like coatings appear 4 to 8 hr after inoculation of potato discs. A similar phenomenon could explain the change in sensitivity of tumor formation on pinto bean leaves to rifampin. Whereas microcolonies appeared to be anchored to the surface of potato tissue by the amorphous polysaccharide-like coating, small rod-like structures were observed bridging individual bacteria to the plant surface. If fixation artifacts were responsible, widespread occurrence of these structures would be expected. However, they were observed on a relatively small proportion of the bacteria on the surface. Patriquin et al. (1983) observed similar structures and suggested that they were responsible for anchoring Azospi rill urn spp. cells to sugarcane suspension culture cells. What role, if any, these rod-like structures play in bacterial adsorption or infection is not known.

PAGE 122

CHAPTER SEVEN SUMMARY 1 characterized A. tumefaciens adsorption to potato tuber tissue that is competent to be transformed. It has been hypothesized from tumor formation studies that the adsorption of A. tumefaciens to plant tissue is specific and mediated by complementary molecular constituents, i.e. that o-antigenic region of the bacterial LPS 'recognizes' and binds io a specific galacturonan present in the plant cell wall. Subsequently, it has been shown by measuring adsorption directly that neither the LPS nor galacturonans of the plant mediate specific adsorption of A. tumefaciens to potato tuber tissue. The purpose of my work was: i) to characterize the adsorption of A. tumefaciens to tissue that is competent for transformation (This was done by considering the bacterial suspension as a colloid and describing adsorption out of the liquid phase onto a solid in physical terms); ii) to provide a thermodynamic treatment of the adsorption data to distinguish between physical and chemical adsorption, and iii) to examine the adsorption sites, bacterial colonization, and general surface morphology during the infection process at the ultrastructural level. The adsorption isotherm for A. tumefaciens was accurately defined by the Freundlich adsorption isotherm, that describes a system in which the available binding sites are not saturated. The adsorption isotherm was examined over a more limited initial inoculum density at three different temperatures (6°C, 17

PAGE 123

28°C, and 38°C). Analysis of these isotherm data with the Clausius-Clapeyron equation shows that the heat of adsorption is close to zero. These combined data suggest that the adsorption of A. tumefaciens to potato tuber tissue is governed by physical processes. Physical adsorption processes are partially governed by ionic van der Waals forces, and thus characterization of the ionic nature of the bacterial cell surface is important. In water the bacterium has an isoelectric point of 3, which indicates that the ratio of carboxyl or phosphate to amino groups of the bacterial cell surface is high. If A_. tumefaciens cells bind to potato tissues by physisorption, changes in the ionic or hydrophobic nature of the binding medium should alter bacterial adsorption. Optimum adsorption occurred at pH 7.2, and adsorption was reduced at either pH 4.5 or 9.5. Adsorption to hydrophobic plastic surfaces was insensitive to changes in pH. Conversely, the adsorption to hydrophilic plastic surfaces was optimal at pH 4.5 and decreased with increasing pH. Adsorption of A* tumefaciens cells to potato tissues was greatly reduced by detergents at concentrations as low as 0.05% (v/v). Anionic, cationic, and neutral detergents were found to be effective. The effect of agents that influence both hydrophobic and ionic forces is consistent with the proposed physisorption mechanism for _Atumefaciens. The molecular components involved in A. tumefaciens adsorption to plant tissues are unknown. Proteolytic enzymes either enhanced or had no effect on adsorption. Hydrophobic pili participate in microbial attachment to several substrates and thus pili warrant further examination for their possible involvement in adsorption.

PAGE 124

119 Surface polysaccharides would be likely candidates to possess ionizable groups, which would be affected by pH changes. Thus, these data do not rule out the possibility of the o-antigen reacting with the galacturonan material of the plant cell wall through some type of ionic interaction. However, it would seem unlikely that two acidic polysaccharides that possess a similar charge at a given pH will be bound to each other. It is more likely that each would react with a different molecular component possessing a more complementary charge. In conclusion, adsorption of A. tumefaciens to potato tuber tissue is a reversible, equilibrium, physisorption process described by the Freundlich adsorption isotherm.

PAGE 125

LITERATURE CITED Absolom, D. R., F. V. Lamberti, Z. Policova, W. Zingg, C. J. van Oss, and W. Neumann. 1983. Surface thermodynamics of bacterial adhesion. Appl. Environ. Microbiol. 46:90-97. Anand, V. K., and G. T. Heberlein. 1975. Effect of rifampin application time on crown-gall tumor induction by Agrobacterium tumefaciens. Can. J. Bot. 53:2581-2588. Arbuthnott, J. P., and C. J. Smyth. 1979. Bacterial adhesion in host-pathogen interactions in animals, pp. 165-198. in D. C. Ellwood, J. Melling, and P. Rutter (eds.), Adhesion of microorganisms to surfaces. Academic Press, New York. Bal, A. K., S. Shantharam, and S. Ratnam. 1978. Ultrastructure of Rhizobium japonicum in relation to its attachment to root hairs. J. Bacteriol. 133:1393-1400. Bauer, W. D. 1982. Attachment of Rhizobia to soybean roots. Plant Physiol. 69:Suppl. 143. Beaud, G., P. Manigault, and G. Stoll. 1963. Observations sur des tumeurs vegetales incompletes. Plant Soil 47:25-37. Beringer, J. E. 1980. The development of Rhizobium genetics. The fourth Fleming lecture. J. Gen. Microbiol. I 16:1-7. Berkeley, R. C. W., J. M. Lynch, J. Melling, P. R. Rutter, and B. Vincent. 1980. Microbial adhesion to surfaces. Ellis Harwood Ltd., Chichester, England, pp. 559. Bhuvaneswari, T. V. 1981. Recognition mechanisms and infection process in legumes. Econ. Bot. 35:204-223. Bhuvaneswari, T. V., A. A. Bhagwat, and W. D. Bauer. 1981. Transient susceptibility of root cells in four common legumes to nodulation by rhizobia. Plant Physiol. 68:1144-1149. Bhuvaneswari, T. V., S. G. Pueppke, and W. D. Bauer. 1977. Role of lectins in plant-microorganism interactions. I. Binding of soybean lectin to rhizobia. Plant Physiol. 60:486-491. 120

PAGE 126

121 Bitton, G., and K. C. Marshall. 1980. Adsorption of microorganisms to surfaces. John Wiley & Sons, New York, pp. 439. Blakeman, J. P. 1982. Phylloplane interactions, pp. 308-333. in M. S. Mount and G. H. Lacy (eds.), Phytopathogenic prokaryotes, Vol. I. Academic Press, New York. Bogers, R. J. 1972. On the interaction of Aqrobacterium tumefaciens with cells of Kalanchoe diagremontiana, pp. 239-250. _ln H. P. Maas Gaesteranus (ed), Proc. 34rd. Int. Congr. Plant Pathog. Bact. Cent. Agric. Publ. & Doc., Wageningen, Netherlands. Bohn, H. L., B. L. McNeal, and G. A. O'Conner. 1979. Soil Chemistry. John Wiley and Sons, New York, 329 pp. Broughton, W. J. 1978. Control of specificity in legumeRhizobium associations. J. Appl. Bacteriol. 45:165-194. Broughton, W. J., A. W. S. M. van Egeraat, and T. A. Lie. 1980. Dynamics of Rhizobium competition for nodulation of Pisum sativum cv. Afghanistan. Can. J. Microbiol. 26:562-565. Brown, N. A. 1923. Experiments with Paris daisy and rose to produce resistance to crown gall. Phytopathol. 13:87-99. Brunauer, S. 1943. The adsorption of gases and vapors. Princeton University Press, Princeton, New Jersey. 51 I pp. Burton, A. J., and H. E. Carter. 1964. Purification and characterization of the lipid A component of the lipopolysaccharides from Escherichia coli. Biochem. 3:41 1-418. Carlson, R. W., R. E. Sanders, C. A. Napoli, and P. Albersheim. 1978. Hostsymbiont interactions, 111. Purification and partial characterization of Rhizobium lipopolysaccharides. Plant Physiol. 62:912-917. Chen, A. P. T., and D. A. Phillips. 1 976. Attachment of Rhizobium to legume roots as the basis for specific interactions. Physiol. Plant. 38:83-88. Chilton, M. D., M. H. Drummond, D. J. Merlo, D. Sciaky, A. L. Montoya, M. P. Gordon, and E. W. Nester. 1977. Stable incorporation of plasmid DNA into higher plant cells: the molecular basis of crown gall tumorigenesis. Cell 3:263-271. Conn, H. J. 1940. Biological stains. Biotech. Publ., Geneva, New York, 308 pp. Cook, A. A., and R. E. Stall. 1977. Effect of watersoaking on response to Xanthomonas vesicatoria in pepper leaves. Phytopathol. 67:1101-1103. Cooksey, D. A., and L. W. Moore. 1982. Biological control of crown gall with an aqrocin mutant of Aqrobacterium tumefaciens. Phytopathol. 72:919921.

PAGE 127

: 22 Costerton, J. W., R. T. lrvin, and K. J. Cheng. 1981. The bacterial glycocalyx in nature and disease. Annu. Rev. Microbiol. 35:299-324. Daniels, S. L. 1972. The adsorption of microorganisms onto surfaces: A review. Dev. lnd. Microbiol. 13:21 1-253. Dazzo, F. B. 1980a. Adsorption of microorganisms to roots and other plant surfaces, pp. 253-316. hi G. Bitton and K. C. Marshall (eds.), Adsorption of microorganisms to surfaces. Wiley-lnterscience, New York. Dazzo, F. B. 1980b. Microbial adhesion to plant surfaces, pp. 31 1-328. In R. C. W. Berkeley, J. M. Lynch, J. Melling, P. R. Rutter, and B. Vincent (eds.), Microbial adhesion to surfaces. Ellis Horwood Ltd, Chichester, England. Dazzo, F. B., and D. H. Hubbell. 1975. Cross-reactive antigens and lectin as determinants of symbiotic specificity in the Rhizobiumclover association. Appl. Microbiol. 30:1017-1033. Dazzo, F. B., C. A. Napoli, and D. H. Hubbell. 1976. Adsorption of bacteria to roots as related to host specificity in the Rhizobium -clover symbiosis. Appl. Environ. Microbiol. 32: 1 66171. De Cleene, M., and J. De Ley. 1976. The host range of crown gall. Bot. Rev. 42:389-466. Deinema, M. H., and L. P. T. M. Zevenhuizen. 1971. Formation of cellulose fibrils by Gram-negative bacteria and their role in bacterial flocculation. Arch. Mikrobiol. 78:42-57. Douglas, C. J., W. Halperin, and E. W. Nester. 1982. Agrobacterium tumefaciens mutants affected in attachment to plant cells. J. Bacterid. 152:1265-1275. Draper, J., 1. A. Mackenzie, M. R. Davey, and J. P. Freeman. 1983. Attachment of Agrobacterium tumefaciens to mechanically isolated asparagus cells. Plant Sci. Lett. 29:227-236. El-Kady, S., and S. Sule. 1981. Biological control of crown-gall tested on bean leaves. Acta Phytopathol. Acad. Sci. Hung. 16:307-313. Farrand, S. K., C. 1. Kado, and C. R. Ireland. 1981. Suppression of tumorigenicity by the IncW R plasmid pSa in Agrobacterium tumefaciens . Mol. Gen. Genet. 181:44-51. Fletcher, M. 1977. The effects of culture concentration and age, time, and temperature on bacterial attachment to polystyrene. Can. J. Microbiol. 23:1-6.

PAGE 128

123 Fletcher, M., M. J. Latham, J. M. Lynch, and P. R. Rutter. 1980. The characteristics of interfaces and their role in microbial attachment, pp. 67-78. in R. C. W. Berkeley, J. M. Lynch, J. Mel ling, P. R. Rutter, and B. Vincent (eds.), Microbial adhesion to surfaces. Ellis Harwood Ltd. Chichester, England. Fred, E. B., I. L. Baldwin, and E. McCoy. 1932. Root nodule bacteria and leguminous plants. University of Wisconsin Press, Madison, 788pp. Foster, R. C, and G. D. Bowen. 1982. Plant surfaces and bacterial growth: the rhizosphere and rhizoplane, pp. 159-185. ]n M. S. Mount and G. H. Lacy (eds.), Phytopathogenic prokaryotes, Vol. I. Academic Press, New York. Galanos, C, and 0. Luderitz. 1975. Electrodialysis of lipopolysaccharides and their conversion to uniform salt forms. Eur. J. Biochem. 54:603-610. Gamborg, 0. L., and L. R. Wetter. 1975. Plant tissue culture methods. Nat. Res. Couns. Canada, Saskatoon, Saskatchewan. I 1 pp. Gibbons, R. J., I. Etherden, and Z. Skobe. 1983. Association of fimbriae with the hydrophobicity of Streptococcus sanguis FC-I and adherence to salivary pellicles. Infect. Immun. 41:414-417. Glogowsky, W., and A. G. Galsky. 1978. Agrobacterium tumefaciens site attachment as a necessary prerequisite for crown gall tumor formation on potato discs. Plant Physiol. 61:1031-1033. Goodman, R. N., P.-Y. Huang, and J. A. White. 1976. Ultrastructural evidence for immobilization of an incompatible bacterium, Pseudomonas pisi , in tobacco leaf tissue. Phytopathol. 66:754-764. Goodman, R. N., and S. B. Plurad. 1971. Ultrastructural changes in tobacco undergoing the hypersensitive reaction caused by plant pathogenic bacteria. Physiol. Plant Pathol. 1:11-15. Goodman, R. N., D. J. Politis, and J. A. White. 1977. Ultrastructural evidence of an 'active' immobilization process of incompatible bacteria in tobacco leaf tissue: A resistance reaction, pp. 423-437. \n B. Solheim, and J. Raa, (eds.), Cell wall biochemistry related to specificity in host-plant pathogen interactions. Universitetsforlaget, Oslo, Norway. Gordon, A., S., and F. J. Millero. 1984. Electrolyte effects on attachment of an estuarine bacterium. Appl. Environ. Microbiol. 47:495-499. Gotz, E. -M. 1980. Attachment to plant root surface and nitrogenase activity of Rhizobia associated with Petunia plants. Z. Pflanzenphysiol. 98:464470.

PAGE 129

24 Gotz, E. M., and D. Hess. 1980. Nitrogenase activity induced by wheat plants. Z. Pflanzenphysiol. 98:453-458. Gurley, W. B., J. D. Kemp, M. J. Albert, D. W. Sutton, and J. Callis. 1979. Transcription of Ti plasmid-derived sequences in three octopine-type crown gall tumor lines. Proc. Nat. Acad. Sci. U.S.A. 76:2828-2832. Haas, J. H., and J. Rotem. 1976. Pseudomonas lachrymans adsorption, survival, and infectivity following precision inoculation of leaves. Phytopathol. 66:992-997. Hanold, D. 1983. In vitro transformation of protoplast-derived Hyoscyamus muticus cells by Aqrobacterium tumefaciens. Plant Sci. Lett. 30:177-183. Harden, V. P., and J. O. Harris. 1953. The isoelectric point of bacterial cells. J. Bacteriol. 65:198-202. Hardy, R. W. F., and U. D. Havelka. 1975. Nitrogen fixation research a key to world food. Science 188:633-643. Hess, D., E. -M. Gotz, M. Harfold. 1982. Rhizobium sp. 32HI associated with wheat and petunia: invasion of plant root tissue. Z. Pflanzenphysiol. 107:81-84. Hildebrand, E. M. 1942. A micrurgical study of crown gall infection in tomato. Jour. Agric. Res. 65:45-59. Higashi, S., and M. Abe. 1980. Scanning electron microscopy of Rhizobium trifolii infection sites on root hairs of white clover. Appl. Environ. Microbiol. 40:1094-1099. Hobbs, S. L. A., and J. D. Matton. 1983. Variability and interaction in the Pisum sativum L. Rhizobium leguminosarum symbiosis. Can. J. Plant Sci. 63:591-599. Holl, F. B., and T. A. La Rue. 1976. Genetics of legume plant hosts, pp. 391399. Jn W. E. Newton and C. J. Nyman (eds.), Proc. Int. Sym. on Nitrogen Fixation. Vol. II. Washington State Univ. Press, Pullman, Washington. Holmes, F. 0. 1929. Local lesions in tobacco mosaic. Bot. Gaz. 87:39-55. Jansen van Rensburg, H. J., and B. A. Strijdom. 1982. Root surface association in relation to nodulation of Medicago sativa. Appl. Environ. Microbiol. 44:93-97. Johnson, K. G., and M. B. Perry. 1976. Improved techniques for the preparation of bacterial lipopolysaccharides. Can. J. Microbiol. 22:29-34.

PAGE 130

!25 Kaneshiro, T., F. L. Baker, and D. E. Johnson. 1983. Pleomorphism and acetylene-reducing activity of free-living rhizobia. J. Bacteriol. 153:10451050. Kerr, A. 1975. A genetic model for pathogenicity in Agrobacterium and for tumor induction in plants. J. Theor. Biol. 51:409-417. Kiraly, Z. 1980. Defenses triggered by the invader: hypersensitivity, pp. 201224. ]n J. G. Horsfall, and E. B. Cowling (eds.), Plant disease an advanced treatise, Volume 5. Academic Press, New York. Klement, Z. 1963. Rapid detection of the pathogencity of phytopathogenic pseudomonads. Nature 199:299-300. Klement, Z. 1977. Cell contact recognition versus toxin action in induction of bacterial hypersensitive reaction. Acta Phytopathol. Acad. Sci. Hung. 6: 1 1 51 1 8. Klement, Z. 1982. Hypersensitivity, pp. 149-177. In M. S. Mount, and G. H. Lacy, (eds.), The phytopathogenic prokaryotes, Volume 2. Academic Press, New York. Klement, Z., G. L. Farkas, and L. Lovrekovich. 1964. Hypersensitive reaction induced by phytopathogenic bacteria in the tobacco leaf. Phytopathol. 54:474-477. Klement, Z., and R. N. Goodman. 1967. The role of the living bacterial cell and induction time in the hypersensitive reaction of the tobacco plant. Phytopathol. 57:322-323. Klement, Z., and L. Lovrekovich. 1961. Defense reactions induced by phytopathogenic bacteria in bean pods. Phytopathol. Z. 41:217-227. Klement, Z., and L. Lovrekovich. 1962. Studies on host-parasite relations in bean pods infected with bacteria. Phytopathol. Z. 45:81-88. Krens, F. A., L. Molendijk, G. J. Wullems, and R. A. Schilperoort. 1982. ]n vitro transformation of plant protoplasts with Ti-plasmid DNA. Nature 296:72-74. Larrimore, S., H. Murchison, T. Shiota, S. M. Michalek, and R. Curtiss III. 1 983. In vitro and in vivo complementation of Streptococcus mutans mutants defective in adherence. Inf. Immun. 42:558-566. Law, i. J., A. J. Mort, and W. D. Bauer. 1982. Nodulation of soybean by Rhizobium japonicum mutants with altered capsule synthesis. Planta 154:100-109. Leben, C, and R. E. Whitmoyer. 1979. Adherence of bacteria to leaves. Can. J. Microbiol. 25:896-901.

PAGE 131

126 Lie, T. A., P. C. J. M. Timmermans, and G. Ladzinsky. 1981. Host controlled nitrogen fixation in Pisum sativum ecotype fulvum, p. 419. jn_A. H. Gibson and W. E. Newton (eds.), Current perspectives in nitrogen fixation. Aust. Acad, of Sci., Canberra, Australia. Lippincott, B. B., and J. A. Lippincott. 1969. Bacterial attachment to a specific wound site as an essential stage in tumor initiation by Agrobacterium tumefaciens. J. Bacteriol. 97:620-628. Lippincott, B. B., J. B. Margot, and J. A. Lippincott. 1977a. Plasmid content and tumor initiation complementation by Agrobacterium tumefaciens IIBNV6. J. Bacteriol. 132:824-831. Lippincott, B. B., M. H. Whatley, and J. A. Lippincott. 1977b. Tumor induction by Agrobacterium involves attachment of the bacterium to a site on the host plant cell wall. Plant Physiol. 59:388-390. Lippincott, J. A., C. C. Chang, V. R. Creaser-Pence, P. R. Birnberg, S. S. Rao, J. B. Margot, M. H. Whatley, and B. B. Lippincott. 1978. Genetic determinants governing enhancement of tumor initiation by avirulent agrobacteria, Agrobacteri umhost adherence and octopine synthesis, pp. 189-197. ]n Proc. 4th Int. Conf. Plant. Path. Bact. Gibert-Clarey, Tours, France Lippincott, J. A., and G. T. Heberlein. 1965a. The induction of leaf tumors by Agrobacterium tumefaciens. Am. J. Bot. 52:396-403. Lippincott, J. A., and G. T. Heberlein. 1965b. The quantitative determination of the infectivity of Agrobacterium tumefaciens. Am. J. Bot. 52:856-863. Lippincott, J. A., and B. B. Lippincott. 1970. Enhanced tumor initiation by mixtures of tumorigenic and nontumorigenic strains of Agrobacterium. Inf. Immun. 2:623-630. Lippincott, J. A., and B. B. Lippincott. 1975. The genus Agrobacterium and plant tumorigenesis. Annu. Rev. Microbiol. 29:377-405. Lippincott, J. A., and B. B. Lippincott. 1977. Nature and specificity of the bacterium-host attachment in Agrobacterium infection, pp. 438-451. In B. Solheim and J. Raa (eds.), Cell wall biochemistry related to specificity in host-plant pathogen interaction. Universitetsforlaget, Oslo, Norway. Lippincott, J. A., and B. B. Lippincott. 1978a. Cell walls of crown-gall tumors and embryonic plant tissues lack Agrobacterium adherence sites. Science 199:1075-1077. Lippincott, J. A., and B. B. Lippincott. 1978b. Tumor initiation complementation on bean leaves by mixtures of tumorigenic and nontumorigenic Agrobacterium rhizogenes . Phytopathol. 68:365-370.

PAGE 132

127 Lippincott, J. A., and B. B. Lippincott. 1980. Microbial adherence in plants, pp. 377-397. In W. H. Beachey (ed.), Bacterial adherence. Chapman and Hall, London, England. Lippincott, J. A., and B. B. Lippincott. 1983. Adherence and host recognition in Agrobacterium infection. Third Int. Symp. Microbiol. Ecol., Michigan (abstr.). Lips, A., and N. E. Jessup, 1979. Colloidal aspects of bacterial adhesion, pp. 527. ]n D. C. Ellwood, J. Melting, and R. Rutter (eds.) Adhesion of microorganisms to surfaces. Academic Press, New York. Ljunggren, H. 1961. Transfer of virulence in Rhizobium trifolii. Nature 191:623. Lopez, J. G., and G. R. Vela. 1981. True morphology of the Azotobacteraceaefilterable bacteria. Nature 289:588-590. Luderitz, 0., K. Jann, and R. Wheat. 1968. Somatic and capsular antigens of Gram-negative bacteria. Compr. Biochem. 26A: 1 05-228. Maier, R. J., and W. J. Brill. 1976. Ineffective and non-nodulating mutant strains of Rhizobium japonicum. J. Bacteriol. 127:763-769. Marshall, K. C. 1976. Interfaces in microbial ecology. Harvard Univ. Press, Cambridge, Massachusetts. 156 pp. Marshall, K. C, and R. H. Cruickshank. 1973. Cell surface hydrophobicity and the orientation of certain bacteria at interfaces. Arch. Microbiol. 91:2940. Marshall, K. C, R. H. Cruickshank, and H. V. A. Bushley. 1975. The orientation of certain root-nodule bacteria at interfaces, including legume root-hair surfaces. J. Gen. Microbiol. 91:198-200. Marton, L., G. L. Wullems, L. Molendijk, and R. A. Schilperoort. 1979. In-vitro transformation of cultured cells from Nicotiana tabacum by Agrobacterium tumefaciens. Nature 277: 1 29131. Matthysse, A. G. 1983. Role of bacterial cellulose fibrils in Agrobacterium tumefaciens infection. J. Bacteriol. 154:906-915. Matthysse, A. G., and R. H. G. Gurlitz. 1982. Plant cell range for attachment of Agrobacterium tumefaciens to tissue culture cells. Physiol. Plant Pathol. 21:381-387. Matthysse, A. G., K. V. Holmes, and R. H. G. Gurlitz. 1981. Elaboration of cellulose fibrils by Agrobacterium tumefaciens during attachment to carrot cells. J. Bacteriol. 145:583-995. Matthysse, A. G., V. K. Holmes, and R. H. G. Gurlitz. 1982. Binding of Agrobacterium tumefaciens to carrot protoplasts. Physiol. Plant Pathol. 20:27-33.

PAGE 133

28 Matthysse, A. G., P. M. Wyman, and K. V. Holmes. 1978. Plasmid-dependent attachment of Agrobacterium tumefaciens to plant tissue culture cells. Inf. Immun. 22:516-522. Menzel, G., H. Uhlig, and G. Weichsel. 1972. Uber die besiedlung der wurzeln einiger leguminosen and nichtleguminosen mit rhizobien and anderen bodenbakterien. Zentralbl. Bakteriol., Parasitenk., D. Infektionskr. Hyg. Abt. Microbiol. Landwirtsch. Tech. Mikrobiol. 127-248-358. Merlo, D. S. 1978. Crown gall— A unique disease, pp. 201-213. Jn J. G. Horsfall and b. B. Cowling (eds.) Plant disease an advanced treatise, Vol III. Academic Press, New York. Meyer, M. C, and S. G. Pueppke. 1980. Differentiation of Rhizobium japonicum strain derivatives by antibiotic sensitivity patterns, lectin binding, and utilization of biochemicals. Can. J. Microbiol. 26:606-617. Moore, L. W., and D. A. Cooksey. 1981. Biology of Agrobacterium tumefaciens : Plant interactions, pp. 15-46. J£ K. L. Giles and A. G. Atherly (eds.), Int. Rev. Cyt. Suppl. 13. Academic Press, New York. Muller, K. 0. 1959. Hypersensitivity, pp. 469-519. jn J. G. Horsfall, and A. E. Dimond (eds.), Plant pathology an advanced treatise. Academic Press, New York. Napoli, C. A., F. Dazzo, and D. Hubbell. 1975. Production of cellulose microfibrils by Rhizobium. Appl. Microbiol. 30:123-131. Nester, E. W., and T. Kosuge. 1981. Plasmids specifying plant hyperplasias. Annu. Rev. Microbiol. 35:531-564. New, P. B., J. J. Scott, C. R. Ireland, S. K. Farrand, B. B. Lippincott, and J. A. Lippincott. 1983. Plasmid pSa causes loss of LPS-mediated adherence in Agrobacterium . J. Gen. Microbiol. 129:3657-3660. Nissen, P. 1971. Choline sulfate permease: transfer of information from bacteria to higher plants II. Induction processes, pp. 201-212. In L. Ledoux (ed.), Information molecules in biological systems. North Holland Publ. Co. Amsterdam, Netherlands. Norkrans, 3. 1981. The hydrophobicity of bacteria an important factor in their adhesion at the air-water interface. Arch. Microbiol. 128:267-270. Nutman, P. S. 1981. Hereditary host factors affecting nodulation and nitrogen fixation, pp. 1 94-204. hn A. H. Gibson and W. E. Newton (eds.), Current perspectives in nitrogen fixation, Proc. 4th Int. Sym. Nitrogen Fixation. Australian Acad. Sci., Canberra, Australia. Ofek, I., and E. H. Beachey. 1980. Bacterial adherence. Adv. Int. Med. 25:503-532.

PAGE 134

129 Ohyama, K., L. E. Pelcher, and A. Schaefer. 1979. In vitro binding of Aqrobacterium tumefaciens to plant cells from suspension culture. Plant Physiol. 63:382-387. Otto, J. C. G. 1975. Ecology, physiology and genetics of fimbrae and pili. Ann. Rev. Microbiol. 29:79-108. Patriquin, D. G., J. Dobereiner, and D. K. Jain. 1983. Sites and processes of association between diazotrophs and grasses. Can J. Microbiol. 29:900915. Perkins, A. T. 1925. The effect of bacterial numbers on the nodulation of Virginia soybeans. J. Agric. Res. 30:95-96. Pueppke, S. G. 1984a. Adsorption of bacteria to plant surfaces. In T. Kosuge and E. W. Nester (eds.), Plant-microbe interactions, molecular and genetic perspectives. Macmillan Publishing Co., New York (in press). Pueppke, S. G. 1984b. Adsorption of slow and fast-growing Rhizobia to soybean and cowpea roots. Plant Physiol, (in press). Pueppke, S. G., and U. K. Benny. 1981. Induction of tumors on Solanum tuberosum L. by Aqrobacterium : quantitative analysis, inhibition by carbohydrates, and virulence of selected strains. Physiol. Plant Pathol. 18:169-179. Pueppke, S. G., and U. K. Benny. 1983. Aqrobacterium tumefaciens in potato: effect of added Aqrobacterium lipopolysaccharides and the degree of methylation of added plant galacturonans. Physiol. Plant Pathol. 23:439-446. Pueppke, S. G., and U. K. Benny. 1984. Adsorption of tumorigenic Agrobacterium tumefaciens cells to susceptible potato tuber tissue. Can. J. Microbiol. 30: (in press). Rao, S. S., B. B. Lippincott, and J. A. Lippincott. 1982. Agrobacterium adherence involves the pectic portion of the host cell wall and is sensitive to the degree of pectin methylation. Physiol. Plant Pathol. 56:374-380. Reporter, M., D. Raveed, and G. Norris. 1975. Binding of Rhizobium japonicum to cultured soybean root cells: morphological evidence. Plant Sci. Lett. 5:73-76. Roberts, W. P., and A. Kerr. 1974. Crown gall induction: serological reactions, isozyme patterns, sensitivity to mitomycin C and to bacteriocin, of pathogenic and non-pathogenic strains of Agrobacterium radiobacter . Physiol. Plant Pathol. 4:81-91. Rosenberg, M. 1981. Bacterial adherence to polystyrene, a replica method of screening for bacterial hydrophobicity. Appl. Environ. Microbiol. 42:375377.

PAGE 135

130 Rosenberg, M., S. Rottem, and E. Rosenberg. 1982. Cell surface hydrophobicity of smooth and rough Proteus mirabilis strains as determined by adherence to hydrocarbons. FEMS Microbiol. Lett. 13:167-169. Rosenberg, E., A. Gottlieb, and M. Rosenberg. 1983. Inhibition of bacterial adherence to hydrocarbons and epithelial cells by Emulsan. Inf. Immun. 39:1024-1028. Sahlman, K., and G. Fahraeus. 1963. An electron microscope study of roothair infection by RNzo^ium. J. Gen. Microbiol. 33:425-427. Saunders, L. 1971. Principles of physical chemistry for biology and pharmacy. Oxford Univ. Press, London, England. 420 pp. Schilperoort, R. A. 1969. Investigations on plant tumors— crown gall. PhD. Thesis, Univ. of Leiden, Netherlands. Schmidt, E. L. 1979. Initiation of plant root-microbe interactions. Annu. Rev. Microbiol. 33:355-376. Sequeira, L. 1980. Defenses triggered by the invader: recognition and compatibility phenomena, pp. 1 79-200. ]n J. G. Horsfall, and E. B. Cowling, (eds.), Plant disease an advanced treatise, Volume 5. Academic Press, New York. Sequeira, L., G. Gaard, and G. A. DeZoeten. 1977. Attachment of bacteria to host cells walls: Its relation to mechanisms of induced resistance. Physiol. Plant Pathol. 10:43-50. Sequeira, L, and T. L. Graham. 1977. Agglutination of avirulent strains of Pseudomonas solanacearum by potato lectin. Physiol. Plant Pathol. 11:43-54. Shimshick, E. J. and R. R. Hebert. 1978. Adsorption of Rhizobia to cereal roots. Biochem. Biophys. Res. Comm. 84:736-742. Shimshick, E. J. and R. R. Hebert. 1979. Binding characteristics of N 2 -fixing bacteria to cereal roots. Appl. Env. Microbiol. 38:447-453. Sigee, D. C, V. A. Smith, and J. Hindley. 1982. Passage of bacterial DNA into host cells during in vitro transformation of Nicotiana tabacum by Agrobacterium tumefaciens. Microbios 34:1 13-132. Sing, V. O., and M. N. Schroth. 1977. Bacteria-plant cell surface interaction: Active immobilization of saprophytic bacteria in plant leaves. Science 197:759-761. Slusarenko, A. J., and R. K. S. Wood. 1983. Agglutination of Pseudomonas phaseolicola by pectic polysaccharide from leaves of Phaseolus vulgaris Physiol. Plant Pathol. 23:217-227.

PAGE 136

131 Smith, V. A., and J. Hindley. 1978. Effect of agrocin 84 on attachment of Agrobacterium tumefaciens to cultured tobacco cells. Nature 276:498500. Solheim, B. 1983. Possible role of lectins in binding rhizobia to host roots, pp. 539-547. hnT. C. Bog-Hansen and G. Spengler (eds.), Lectins-Biology, biochemistry, clinical biochemistry, Volume 3. Walter de Gruyter, Berlin, Germany. Solheim, B., and J. Paxton. 1981. Recognition in Rhizobiumlegume systems, pp. 71-83. hnR. C. Staples and G. H. Toenniessen (eds.), Plant disease control: Resistance and susceptibility. John Wiley, New York. Spiess, L. D., B. B. Lippincott, and J. A. Lippincott. 1971. Development and gametophore induction in the moss Pylaisiella selwynii as influenced by Agrobacterium tumefaciens. Am. J. Bot. 58:726-731. Spiess, L. D., B. B. Lippincott, and J. A. Lippincott. 1976. The requirement of physical contact for moss gametophore induction by Agrobacterium tumefaciens . Am. J. Bot. 58:726-731. Spiess, L. D., B. B. Lippincott, and J. A. Lippincott. 1977a. Comparative response of Pylaisiella selwynii to Agrobacterium and Rhizobium species. Bot. Gaz. 138:35-40. Spiess, L. D., J. C. Turner, P. G. Mahlberg, B. B. Lippincott, and J. A. Lippincott. 1977b. Adherence of Agrobacteria to moss protonema and gametophores viewed by scanning electron microscopy. Am. J. Bot. 64: 1 2001 204. Stacey, G., A. S. Paau, and W. J. Brill. 1980. Host recognition in the Rhizobiumsoybean symbiosis. Plant Physiol. 66:609-614. Stacey, G., A. S. Paau, K. D. Noel, R. J. Maier, L. E. Silver, and W. J. Brill. I 982. Mutants of Rhizobium japonicum defective in nodulation. Arch. Microbiol. 132:219-224. Stall, R. E., and A. A. Cook. 1979. Evidence that bacterial contact with the plant cell is necessary for the hypersensitive reaction but not the susceptible reaction. Physiol. Plant Pathol. 14:77-84. Stonier, T. 1956. Radioautographic evidence for the intercellular location of crown gall bacteria. Am. J. Bot. 43:647-655. Suslow, T. V. 1982. Role of root-colonizing bacteria in plant growth, pp. 187223. ]n M. S. Mount, and G. H. Lacy (eds.), Phytopathogenic prokarytotes, Volume I. Academic Press, New York. Thomashow, M. F., R. Nutter, A. L. Montoya, M. P. Gordon, and E. W. Nester. 1980. Integration and organization of Ti plasmid sequences in crown gall tumors. Cell 19:729-739.

PAGE 137

32 Tompkins, F. C. 1978. Chemisorption of gases and metals. Academic Press, New York, 250 pp. Tu, J. C, and G. Braybrook, 1981. Structural organization of the rhizobial root nodule of clover. Microbios 32:77-88. Turakhia, M. H., K. E. Cooksey, and W. G. Characklis. 1983. Influence of a calcium-specific chelant on biofilm removal. Appl. Environ. Microbiol. 46:1236-1238. Turgeon, B. G., and W. D. Bauer. 1982. Early events in the infection of soybean by Rhizobium japonicum . Time course and cytology of the initial infection process. Can J. Bot. 60: 1 52161. Vincent, J. M. 1980. Factors controlling the legumeRhizobium symbiosis, pp. 103-129. In W. E. Newton and W. H. Orme-Johnson (eds.), Nitrogen fixation II. Univ. Park Press, Baltimore. Ward, H. M. 1902. On the relations between host and parasite in the Bromes and their brown rust, Puccinia dispersa (Erikss.). Ann. Bot. 16:233-315. Watson, B., T. C. Currier, M. P. Gordon, M. D. Chilton, and E. W. Nester. 1975. Plasmid required for virulence of Agrobacterium tumefaciens . J. Bacteriol. 123:255-264. Westphal, 0., and K. Jann. 1965. Bacterial lipopolysaccharides. Extraction with phenol-water and further applications of the procedure, pp 83-9 I. In R. L. Whistler (ed.), Methods in carbohydrate chemistry, Vol. 5. Academic Press, New York. Whatley, M. H., J. S. Bodwin, B. B. Lippincott, and J. A. Lippincott. 1976. Role of Agrobacterium cell envelope lipopolysaccharide in infection site attachment. Inf. Immun. 13:1080-1083. Whatley, M. H., J. B. Margot, J. Schell, B. B. Lippincott, and J. A. Lippincott. 1978. Plasmid and chromosomal determination of Agrobacterium adherence specificity. J. Gen. Microbiol. 107:395-398. Whatley, M. H., and L. D. Spiess. 1977. Role of bacterial lipopolysaccharides in attachment of Agrobacterium to moss. Plant Physiol. 60:765-766. Wullums, G. J., L. Molendijk, G. Ooms, and R. A. Schilperoort. 1981. Differential expression of crown gall tumor markers in transformants obtained after in vitro Agrobacteriuminduced transformation of cell wall regenerating protoplasts derived from Nicotiana tabacum . Proc. Nat. Acad. Sci. USA 78:4344-4348. Yadau, N. S., K. Postle, R. K. Saiki, M. F. Thomashow, and M. D. Chilton. 1980. T-DNA of a crown gall teratoma is covalently joined to the host plant DNA. Nature 287:458-461.

PAGE 138

33 Zarr, J. H. 1974. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, New Jersey. 620 pp. Zurkowski, W. 1980. Specific adsorption of bacteria to clover root hairs, related to the presence of the plasmid pWZ2 in cells of Rhizobium trifolii. Microbios 27:27-32.

PAGE 139

BIOGRAPHICAL SKETCH Daniel Albert Kluepfel was born in St. Louis, Missouri on January 29, 1955, to Marion and Albert Kluepfel. He attended school in the St. Louis area till his family moved to Jefferson County, Missouri. Upon completion of high school in 1973 he took a position as a computer-programmer operator for three years. During this time he also enrolled in the University of Missouri at St. Louis, from which he received his Bachelor of Arts in biology December of 1978. He came to the University of Florida in the fall of 1979. He received his Ph.D. in plant pathology with a minor in microbiology in 1984. While a graduate student in Florida Dan married Marjan van Nuffelen. In March of 1984 he was awarded a Fulbright fellowship to do research at the Agricultural University of Wageningen in the Netherlands upon completion of his Ph.D. 134

PAGE 140

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. Steven G. Pueppl{ i*David J. Mitchell Professor of Plant 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. J UJtiUatu ^6&t William B. Gurley Assistant Professor of Microbiology and Cell Science 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. M. Brij My Moudgil Associate Professor of Material Science

PAGE 141

This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August 1984 Jack •/ i/ l V Dean, College of Agriculture Dean, Graduate School

PAGE 142

UNIVERSITY OF FLORIDA IIIIIIIIIIIJI 3 1262 08553 2058


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 E6CYIG8QY_EMP5JD INGEST_TIME 2017-07-14T22:24:39Z PACKAGE UF00099488_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES