ANNUAL DECLINE OF THE AQUATIC MACROPHYFE
HYDRILLA VERTICILLATA (L.F.) ROYLE
ROYAL HOWARD BERG
A DISSEPTATION PRESENTED TO THE GE'rDATF COUNCIL OF
THE UNIVERSITY OF FLURILA
IN PARTIAL FULFILL:'EFfi OF THE REQ'JIREMEINTS FOR THE
DEGREE OF rDC'CP OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The author wishes to express his sincere appreciation and gratitude
to Dr. Leon Garrard, Chairman of the Supervisory Committee, for his
generous help and guidance throughout the graduate program.
Appreciation is also expressed to Dr. William Haller for providing
the research funds (off a grant from the Florida Department of Natural
Resources), laboratory facilities, and means of conducting field re-
search much needed in completing this study. The support, both financial
and academic, of the Agronomy Department during the author's graduate
studies is deeply appreciated.
Acknowledgement is also made for the encouragement and advice
throughout this study of the other members of the candidate's Super-
visory Committee which consisted of Dr. H. L. Allen, Jr. and Dr. V. E.
Green, Jr., Agronomy, Dr. G. Bowes, Botany, and Dr. H. C. Aldrich,
Microbiology and Cell Science.
Thanks are also extended to Dr. Aldrich for the use of light and
electron microscopes, and to William Dougherty, Dr. Mike Dykstra,
Dr. Gregg Erdos, and Dr. Thai Van for their stimulating discussions.
The author wishes to thank Susan Loftin for her advice and help in
bringing this study to completion.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS............... ..................................... ii
LIST OF TABLES.......................................... ........... iv
LIST OF FIGURES.................................................... v
KEY TO SYMBOLS..................................................... vii
ABSTRACT............................. ........ .....................viii
GENERAL INTRODUCTION AND REVIEW....................................... 1
MATERIALS AND METHODS............................................. 12
Microscopy .. ................................................ 12
Bacteria Isolation Media.................................. .. 14
Plant Sterilization Methods................................... 17
Tissue Culturing ........................................... 19
Toxin Bioassays ............................................. 22
RESULTS AND DISCUSSION............................................. 23
Field Observations of Annual Decline.......................... 23
Structural Relationships of Epiphytic Flora.................. 32
Laboratory Studies of Epiphytic Bacteria..................... 77
Sunimary and Conclusions............................ ... .... 94
APPEINDiX ........................................................... 98
LIT .AT:-JE CITED................................................. 99
SIiC'-AP- h CAL S; E H. ............................. ................... 104
LIST OF TABLES
1 Media composition for callus tissue cultures.............. 21
2 Effect of treating hydrilla with varying levels
of brittle stem inoculum................................. 83
3 Characterization of brittle stem toxin using the
Avena coleoptile bioassay................................ 95
LIST OF FIGURES
1 Hydrilla mat covering 90% of the surface of Orange
Lake, Florida........................................ 25
2 Diseased hydrilla plant................................. 28
3 Hydrilla plants showing annual decline symptoms.......... 31
4 Transverse section of a hydrilla leaf showing the two
contiguous epidermal layers.............................. 34
5 Fine structure of hydrilla leaf tissue implicating it
in DOM secretion............................... ........ 36
6 Surface view of a hydrilla leaf showing thread-like
epiphytic blue-green algae.............................. 39
7 Epiphytic blue-green algae on the surface of the
adaxial side of the hydrilla leaf........................ 41
8 Point of attachment of a filamentous alga on the hydrilla
9 Filamentous algae attached to the hydrilla leaf surface.. 46
10 Epiphytic bacteria located within the depressions
demarcating cell boundaries on the leaf surface.......... 48
11 Surface staining of epiphytic bacteria using alcian
blue-ruthenium red..................................... 52
12 The extracellular coat of epiphytic algae as demon-
strated by staining with alcian blue-ruthenium red....... 54
13 Intimate association of epiphytic bacteria with the
extracellular coat of an alga ................. ............ 56
14 Types of epiphytic bacteria found on hydrilla leaves..... 59
15 Bacteria degrading hydrilla leaf cell walls.............. 61
LIST OF FIGURES Continued
16 High magnification of hydrilla leaf cell wall under-
going bacterial degradation................................. 63
17 Intracellular bacteria in hydrilla leaf tissue........... 65
18 Pathological response of the cytoplasm.................. 67
19 Association of epiphytic bacteria and algae.............. 70
20 Organisms, other than algae and bacteria, associated
with annual decline in hydrilla.......................... 72
21 Phenol-containing cells in hydrilla leaf tissue.......... 74
22 Localized production of phenol-containing cells in the
hydrilla leaf.................................. ... ....... 76
23 Distribution of phenol-containing cells in a senescent
hydrilla leaf.......................................... 76
24 Cellulytic bacteria cultured on SFP medium............... 81
25 Symptoms of hydrilla infected with a mixture of bacteria
causing brittle stem.................................... 85
26 Secondary degradation of hydrilla leaf inoculated with
the brittle-stem bacteria............................... 88
27 Ultrastructure of hydrilla plants inoculated with brittle-
stem bacteria in the laboratory.......................... 90
28 Aerobic bacteria found in the surface scum of a culture
tube after the plant it contained was killed by brittle-
stem bacteria ................................. ......... 92
KEY TO SYMBOLS
ac = air canal
cf = cellulose fiber
cpl = chloroplast
m = mitochondrion
n = nucleus
pd = plasmodesmata
pg = plastoglobuli
pm = plasmalemma
pt = wall protuberance
v = vacuole
w = wall
Abstract of Dissertation Presented to the Graduate Council of the
University of Florida in Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy
ANNUAL DECLINE OF THE AQUATIC MACROPHYTE
HYDRILLA VERTICILLATA (L.F.) ROYLE
Royal Howard Berg
Chairman: Leon A. Garrard
Major Department: Agronomy
Natural factors involved in the annual decline of the aquatic
macrophyte Hydritta vetticiUata (L.f.) Royle were investigated.
Samples of the plants undergoing annual decline in Astu were exam-
ined ultrastructurally, and studies of epiphytic bacteria were made.
Annual decline of hydrilla generally occurs in the late summer
months and exhibits several characteristics, including heavy popula-
tions of epiphytic bacteria end algae, loss of pigmentation in stems
and leaves, leaf abscission, and stem fragmentation at nodes. These
symptoms generally occur wherever hydrilla develops a heavy infes-
tation, and the decline condition may be aggravated by high plant
density and mutual shading resulting from a dense mat of hydrilla.
Annual decline symptoms are due, in part, to the occurrence of
disease in the plant.
Tissue cultures of hydrilla show the plant to be sensitive to
high salt and hormone concentrations. Axenic plants may be produced
by adventitious shoot and root formation from stem explants containing
at least one node and subjected to low concentrations of 2,4-D.
Isolations of epiphytic bacteria were made on several different
media. Isolates were grown on nutrient agar or mineral salts utilizing
cellulose as the sole carbon source. Epiphytic bacteria were able to
grow on solid and liquid infusion media made from hydrilla plants.
Isolates from these media introduced into healthy hydrilla plants were
not found to produce symptoms of annual decline.
Isolations of mixtures of epiphytic bacteria were found to be ca-
pableof producing "brittle stem" disease in hydrilla in the laboratory.
Symptoms of this disease include fragmentation of the stem at the nodes,
disintegration of stem tissue, and leaf abscission. Indications are
that the bacteria responsible for the disease are highly aerobic, grow
to the greatest extent in stem aerenchyma and air canals of the leaf,
and are capable of degrading cell walls in this tissue. Cell-free
filtrates from solutions of these bacteria were shown to produce brittle
stem symptoms in hydrilla. These filtrates were labeled "toxins" and
were found to mimic abscisic acid in the Avena coleoptile bioassay, i.e.,
they interfered with the action of indoleacetic acid (IAA). Annual
decline of hydrilla, as evidenced by symptoms of stem fragmentation
and leaf abscission, is possibly related to the presence of toxins
produced by epiphyte flora.
Ultrastructural investigations of in 5sitt hydrilla undergoing
annual decline showed the presence of cell-wall-degrading bacteria
in leaf tissue, and these bacteria appeared to cause the lysis of
cells. Lytic bacteria were also seen to attack epiphytic algae.
The close proximity of epiphytic bacteria and algae indicated the
possibility of a symbiotic relationship between them. The structural
relationships of the epiphyte flora on hydrilla during annual decline
suggested that they influence this malady. These organisms may affect
light penetration and carbon dioxide diffusion to the hydrilla plant.
The thin leaf cuticle and the presence of transfer cells in the abaxial
cell layers of the leaf, as well as large vacuoles in the cells com-
prising the adaxial cell layers of the leaf, suggests that the leaf is
capable of secreting dissolved organic matter. It is postulated that
dissolved organic matter is secreted by the hydrilla leaf in increasing
amounts over the growing season, and that this is an important factor
influencing epiphyte populations and, subsequently, annual decline.
GENERAL INTRODUCTION AND REVIEW
Hydci1ela ve-tictitLcat (L.f.) Royle is a submersed vascular aquatic
macrophyte belonging to the monocotyledonous family Hydrocharitaceae.
Subsequent to its introduction into Florida in 1960, it has spread
throughout the state and has been reported in Georgia, Alabama, Missis-
sippi, Louisiana, Texas, Iowa, and California (W. T. Haller, personal
communication). Hydrilla is closely related to Etodea canacde is
(Michx.) and Eget.t denisa (Planch), its distinguishing characteristic
being serrations both on the leaf margins and midvein (Blackburn et al.,
The hydrilla plant is rooted, with long branching stems. It is
made bLioianrt by the presence of aerench);a tissue within the stem and
air spaces in leaf tissue (Pendland, 1976). The lower leaves are
opposite, the median and upper leaves are in whorls of four to eight.
Leaf dimensions are 1-2 cm long by 1.5-2 mm wide.
Pendland (1976) found the hydrilla leaf, in transverse section,
to consist of two contiguous epidermal cell layers with a single
midvein composed of three to four layers of cells. She produced
evidence for the presence of sieve elements and companion cells in
the midvein, both indicative of phloem-type tissue. The two cell
layers are distinct in that the cells of the adaxial epidermal layer
are larger, containing a fibrous ]ayer beneath the outer wall, while
the cells of the abaxial epidermal layer are considerably smaller and
specialized into transfer cells (Gunning and Pate, 1969). These trans-
fer cells presumably enhance apoplastic exchange of nutrients and/or
metabolites (Gunning and Pate, 1974), eventually altering the cell's
exchange with its external environment. The thin cuticle of the
hydrilla leaf would give minimal resistance to interactions of this
type. Pendland proposed the lack of stomata in hydrilla leaf tissue
to be compensated for by the thin cuticle, allowing gas exchange.
Since only the pistillate plant has been introduced into the
United States, reproduction is by vegetative means. Fragmented tissue,
containing at least one node from stems, stolons, or rhizomes is ca-
pable of producing plants. Propagules formed in leaf axils (turions)
are modified buds resistant to environmental stress and are capable of
vegetative propogation (Mitra, 1964). Tubers are formed 5-10 cm beneath
the surface of the hydrosoil at the tips of stolons. They are equiv-
alent to turions, in function, and are highly resistant to eradication
(Miller, 1975). In Florida, turions and tubers are generally formed
from September to April, as the plant goes through its annual decline,
and sprout during the spring or summer.
Hydrilla forms extensive monotypic stands, often dominating large
bodies of water in a short period of time. In north Florida, Rodman
Reservoir had one ha of hydrilla in 1971. This had increased to a
sizable stand of 1200 ha by 1975 (Haller, 1976). In a nearby lake,
O:.aij Lake, Florida, infestation followed a similar pattern. This
lake is a relatively shallow (2-4 m), warm lake covering 5000 ha.
Over a period of three years hydrilla spread dramatically from 1 ha
growth in 1973 to an infestation of over 4000 ha in 1976. During the
summer of that year 90 percent of the lake's surface was covered by
a mat of hydrilla, causing economic disaster for people dependent on
the lake for their livelihood. It appears the lake, once well known
for its beauty and aquatic life, will eventually become hypereutrophic,
viz, an aquatic desert.
Hydrilla forms a dense canopy at the surface of a body of water,
limiting light penetration to less than 5 percent 30 cm below the
surface (Haller, 1974). Van et al. (1976) found, in studies comparing
the photosynthetic characteristics of hydrilla to those of Cetatophy jum
demetum and Mytiophyllum spicatun, that hydrilla had the lowest light
compensation point of the three plants. They postulated that this
characteristic gives hydrilla a distinct competitive advantage in the
field, especially during early morning hours when irradiance is low.
Submersed weeds are perhaps the most serious of all aquatic weed
problems; they are not easily treated with herbicides and do not lend
themselves to clearance by machines (Holm et al., 1969). Herbicidal
control involves treating the whole body of water, raising the possi-
bility of harmful effects on beneficial plant species and animal life.
Diquat, in combination with copper sulfate, and endothall are two
herbicides commonly used on hydrilla (Burkhalter et al.). Because
of its submersed habitat and its high content of water (over 95 percent),
mechanical removal as a control measure for hydrilla is impractical in
The use of biological controls (biocontrols) in aquatic weed
management programs has enjoyed recent success (alligatorweed beetle;
Maddox et al., 1971), and will probably become the method of choice
in future control programs. Ideally, organisms chosen as biocontrols
preserve species diversity, minimizing effects on ecosystems into which
they are introduced. Natural agents, such as insects, fish, snails,
and pathogenic microorganisms have been used as biocontrols for aquatic
weeds, reducing their populations to non-nuisance levels (Blackburn et al.,
1971). The white amur or Chinese grass carp (Ctenopharyngodon idefLa
Val.), though the source of much controversy, is being tested as a bio-
control agent for hydrilla (Michewicz et al., 1972).
Bacterial and fungal diseases have received little attention as
biocontrols of aquatic weeds. This is unfortunate as there are approx-
imately 100,000 plant diseases (McNew, 1966) from which selections of
biocontrol agents most certainly could be made. Zettler and Freeman
(1972) offered the following advantages in using plant pathogens as bio-
controls for aquatic weeds: a) control applications would presumably
require minimal technology and, if successfully established, the path-
ogen in theory would be self-maintaining, b) the overwhelming number of
different plant-pathogenic species from which to choose offers an
unmatched versatility in selecting a specific biological control,
c) virtually none can attack man or his animals, therefore providing
an important advantage over the use of various animals such as snails,
which may harbor chordate pathogens, d) plant pathogens, although often
killing individuals in a given population, would not be expected to
cause the extermination of a species" (p. 459). Aquatic bacteria and
fungi are common (Breed et al., 1957; Sparrow, 1960); however, care
should be taken in testing plant pathogens. The marine eelgrass
(Zostera marina) underwent a dramatic decline along the northeast
coast of the United States in the 1930's, showing the destruction
possible with an aquatic disease (Tutin, 1938).
Conway (1976a) isolated an aquatic fungus indigenous to Florida,
Cercosporta todmanii Conway, from waterhyacinth (Eichhornia crassipes
(Mart.) Solms) growing in Rodman Reservoir. He associated the fungus
with decline symptoms of the plant and is currently conducting field
trials in an evaluation of its use as a biocontrol of waterhyacinth
There is a scarcity of reports on diseases of submersed plants.
Undoubtedly, this is due to a lack of research in this area.
Charudattan (1973) evaluated the pathogenicity (to hydrilla and
waterhyacinth) of several bacteria and fungi from India. Several of
these, including a gram-negative rod bacterium, showed a detrimental
effect on hydrilla. Freeman et al., (1976), conducting an extensive
investigation of plant pathogens as biocontrols of aquatic weeds,
could find no pathogen to be effective as a biocontrol of hydrilla.
To date, no known plant pathogen for biocontrol of hydrilla has
been found, though a recently isolated fungus is being studied
(R. Charudattan, personal communication).
An interesting review of the spread of Etodea canadensi, a
close relative of hydrilla, may be found in Sculthorpe (1967; pp.
360-364). According to his account, elodea was first noted in Ireland
in 1836. It rapidly spread throughout Great Britain and, by the end
of the nineteenth century, was well established over the continent of
Europe. The interesting point to be made here is the subsequent
decline of the weed, which occurred in virtually all of the areas
it had infested. This natural decline would take place after 5 to
7 years of colonization and was rapid in some cases, gradual in
others. Elodea ceased to be a problem in these areas, though the
plant never entirely disappeared. Several theories have been offered
to explain this decline. A frequent argument was that, since repro-
duction was strictly vegetative, the plant lacked a source of renewal
of genetic material. This is refuted by the observation that the
plant retained the capability to infest new areas rapidly. Salisbury
(1961) proposed that the plant is held in check by some sort of
nutrient deficiency. This would not explain the ability of the plant
to grow, though in low populations, in these areas. Hutchinson (1975)
feels that the decline is due to a "yet unrecognized biotic factor"
(p. 249). Indeed, the decline could have been the result of infes-
tation by a plant pathogen. These events observed in elodea population
trends may be significant when considering the problem of hydrilla's
rapid infestation in the United States. Hydrilla has disappeared from
a few scattered, formerly heavily-infested areas in Florida (W. T.
Haller, personal communication) without plausible explanation. This
will be discussed in Chapter 3.
Biotic factors important to hydrilla include the epiphytic algae
and bacteria on leaf surfaces. Godward's (1934) study of algal
epiphytes has shown the number of epiphytes to be least on younger
leaves of aquatic plants, increasing to a maximum on older living
leaves. The number decreases thereafter until the leaves die. Algal
epiphytes are younger towards the apex of the plant. There are more
epiphytes on the upper surface than the lower surface of the leaves.
Godward found that most of the algae settled at the depressions
demarcating the cell boundaries of a leaf.
Macrophytes influence epiphytic populations on their surfaces
in many ways. Fitzgerald (1969) studied conditions influencing
epiphyte colonization of macrophytes (Cladophota), as well as barley
seedling roots in the field. He found epiphyte growth to be low if
the plant was limited by nitrogen (nitrate or ammonia). Conversely,
if combined nitrogen was present in excess of the needs of the plant,
epiphyte growth was heavy. As nitrogen demands of the plants decreased
in late summer, epiphytic blue-green algae appeared on the leaves.
In an important series of papers, Wetzel and Allen reported
studies of interactions of aquatic macrophytes, including the submersed
Najas &texixAi, with their epiphytic bacteria and algae. Wetzel (1969)
found that a major fraction of recently synthesized organic carbon and
nitrogen in dissolved form were lost from macrophytes through secre-
tion into ambient waters. Under optimum conditions for the plant,
this caused a significant reduction in photosynthetic efficiency
(at least 10 percent of photosynthetically-fixed carbon). Plants
placed under stress lost considerably greater amounts of dissolved
organic matter (DOM). Wetzel postulated this loss to be due to an
incomplete macrophyte adaption to a totally aqueous medium.
That this loss of DOM caused development of a heavy epiphyte
community on these plants was shown by Allen (1971), working with
pure cultures of epiphytic bacteria (Cautobeate; P seudomonas) and
algae (Gomphonema, CtorLelta, and CycZoteUla). Studies of the
kinetics (Michaelis-Menten plots), using 14C-labeled glucose and
acetate, showed the bacterial uptake to follow strictly first order
kinetics, indicative of active (enzymatic) uptake. Algal uptake
followed zero-order kinetics, indicative of uptake by diffusion.
In mixed cultures of bacteria and algae kinetics were first order
at low (field) concentrations of substrate and zero order at higher
concentrations. Bacterial uptake was unaffected by culturing with
algae, while algal uptake increased substantially in mixed cultures.
These single and mixed cultures were subjected to 1C-labeled DOM
secretions from the submersed Najctai Lexii Axenic cultures of
bacteria took up two times as much label as pure cultures of algae.
Mixtures of bacteria and algae took up higher amounts of DOM than in
their combined monocultures. Pretreatment of the secretary solutions
by growing epiphytic bacteria in them caused subsequent growth of
algae in these solutions to be significantly increased. It was sug-
gested bacterial growth provided CO2, vitamin B12, or some other
metabolite that stimulated algal growth. In itua measurements showed
that epiphyte growth was greatest in summer, when the standing crop
and temperatures were both at a maximum. The epiphytic flora was
found to metabolize glucose, fructose, galactose, acetate, glycolate,
succinate, glycine, alanine, and serine. Acetate was most readily
used. Following patterns of 14C-labeled DOM release, Allen showed
that plants with their epiphytes removed exuded DOM into the ambient
water, whereas plants with epiphytes released little DOM into the
ambient water, the DOM being retained by the epiphytes. Some evidence
was found for stimulation of DOM secretion by macrophytes when epiphytic
flora were present. Allen proposed that epiphytes were dependent on
DOM lost from their macrophyte host for full development.
Scanning electron micrographs of epiphytes on Potamogeton natan.s
and Chzta sp. show that the spatial proximity needed for epiphyte
utilization of macrophyte exudations exists (Allanson, 1973).
Algae are not considered to be plant pathogens, though they may
have some allelopathic effect in aquatic systems (Rice, 1974). Sand-
Jensen (1977) studied the effect of algal epiphytes (mainly diatoms)
on photosynthesis in eelgrass. He found that photosynthetic rates
were reduced because epiphytes interfered with CO2 diffusion and light
Growth of the marine alga Utva tactuca was shown by Waite and
Mitchell (1976) to be influenced by epiphytic bacteria occurring on
the surface of the macrophyte. Photosynthesis was increased by 50
percent with a mixture of epiphytic bacteria, presumably due to
production of growth factors by the bacteria. A few isolates caused
a significant decline in photosynthesis. Electron micrographs showed
the presence of bacteria in UZva cell walls and within cells. It was
concluded that individual types of bacteria could either stimulate or
inhibit the growth of UWva.
The great bacteriologist Beijerinck had two rules for the study
of microorganisms in their environment: a) everything is everywhere
and b) the milieu selects (Bass-Becking, 1959). Recognizing the
ubiquity of plant pathogenic bacteria, and the milieu favorable to
epiphytic bacterial'growth on macrophytes, it is feasible that a
bacterial biocontrol could be developed for hydrilla.
One of the most severe limitations in studying aquatic bacteria
is determining microbial kinds and numbers occurring in a body of
water (Brock, 1971). The total number of bacteria occurring in lake
waters is often 1,000-20,000 times the numbers found by inoculating
media (Kuznetsov, 1970; p. 122). This is most likely due to limi-
tations involved in making bacterial growth media, i.e., environmental
factors occurring in situ can rarely be duplicated in the laboratory
(Brock, 1971). Direct microscope counting methods are commonly used
for the determination of bacterial numbers (Collins and Kipling, 1957),
though other methods may be more accurate and convenient (Brock, 1971).
The more successful cultures of aquatic bacteria have been made using
enrichment techniques (Stanier et al., 1970; pp. 88-96).
The density of planktonic bacteria per ml is about 150,000 in
oligotrophic lakes, from 500,000 to 1,500,000 in mesotrophic lakes,
and from 2,000,000 to 10,000,000 in eutrophic lakes (Kuznetsov, 1970;
p. 122). Little is known of bacterial numbers and types growing as
epiphytes on aquatic macrophytes. It has already been mentioned that
epiphytic bacteria growth is at a maximum in the summer; this is true
for planktonic bacteria also (Kuznetsov, 1970; p. 125). Free-floating
bacteria often become attached to solid surfaces, be they planktonic
organisms, sessile algae, rocks, or aquatic plants (Van Niel and
Stanier, 1959). This allows exposure to nutrients from the substrate
and in the water flow. Planktonic bacteria in lake waters concentrate
at or near the surface (aerobic bacteria and bacteria dispersed through
the air) and near the bottom anaerobess) (Kuznetsov, 1970; pp. 124-125).
Hydrilla may be a suitable substrate for aquatic bacteria. Indeed,
by late summer, hydrilla forms a dense mat over the surface of the water
and during this time epiphyte growth is commonly observed on the plants.
The interactions occurring in this situation will be discussed in Chapter
As the summer passes into fall hydrilla undergoes what is known as
annual decline. Little research has been done on what causes this
annual decline. The plant dies and sinks to the hydrosoil as turions
and tubers are formed. These will germinate under favorable conditions,
usually in the spring and summer. Subsequent growth over the next
growing season results in the formation of a dense mat during the summer
months. During the growing season photosynthesis probably increases,
increasing oxygen production in the mat as well as photosynthate pro-
duction in the plant. One explanation given for annual (Haller, 1976)
decline is that, during the rapid summer growth, the hydrilla mat sur-
passes optimal plant density and mutual shading of leaves occurs. This,
coupled with the shortening of daylength in late summer, would cause a
decrease in net photosynthesis and relative increase in respiration,
resulting in plant starvation and death (physiological decline). Decline
has also been attributed to photo-oxidation of leaf tissue subjected to
the high intensities of light occurring in summer. In terms of senes-
cence, it is not known if the plant itself undergoes self-directed
metabolic changes causing an annual cyclic growth habit.
To be cognizant of what is occurring in the annual decline of
hydrilla, examination of the plant in situ is required. The complex-
ities of limnological systems make this a difficult task. Since factors
applicable to biocontrol could be involved in annual decline, the inquiry
presented here concerns itself with biological interactions between the
hydrilla leaf and its epiphytes, especially as they pertain to annual
decline of hydrilla.
MATERIALS AND METHODS
Light micrographs were made using either whole or hand sectioned
leaves. Micrographs were made with a Zeiss light microscope using
differential contrast interference optics.
For electron microscopy, several methods of fixation were at-
tempted, none of which proved to be completely satisfactory. The
most commonly used method consisted of submersing whole leaves in a
mixture of 3% glutaraldehyde, 3% acrolein, and water in which the
tissue was growing. Tissue was left in the fixative for 24 h at room
temperature, rinsed in the water, and then post-fixed in 2% osmium
tetroxide for 3 h. The tissue was then dehydrated in ethanol and
embedded in three types of plastic: low fiscosity plastic (Spurr,
1969), and epon-araldite mixture (Mollenhauer, 1964), and modified
plastic mixture. The first two plastics were infiltrated for 24 h,
including 12 h in 70% plastic; the plastic mixture consisted of
infiltration with 70% Spurr's plastic (12 h), followed by transfer
to 100% Mollenhauer's plastic, infiltrating for 1 h, and then poly-
merization for 3 days in a 70 oven. Other fixation methods attempted
are described in the Appendix. The reader is referred to Pendland
(1976) for additional discussion on preparing hydrilla leaf tissue
for electron microscopy.
Ultrathin sections made with a Sorvall Microtome, Model MT-2,
were stained 20 min in uranyl acetate and 15 min in lead citrate
(Reynolds, 1963), and examined in a Hitachi HU11-E electron microscope
(Biological Ultrastructure Laboratory, University of Florida).
Alcian blue-ruthenium red cytochemical methods (Dykstra and
Aldrich, 1976) were used to elucidate cell surface interactions of
hydrilla leaf epiphytes. The primary fixative was 0.5% alcian blue
in 2.5% glutaraldehyde made by mixing 2.0 ml alcian blue stock solution
(0.1 g alcian blue in 10.0 ml distilled water) with 1.0 ml 10% glu-
taraldehyde and 1.0 ml of the water in which the tissue was growing.
Primary fixation was for 1 h at room temperature. This was followed
by three 5-min rinses in water, after which the tissue was placed in
the secondary fixative for 2 h. The secondary fixative was 0.05%
ruthenium red in 2.0% osmium tetroxide made by mixing 1.0 ml of
ruthenium red stock (0.01 g ruthenium red in 10.0 ml distilled water)
with 1.0 ml of 4% osmium tetroxide. After secondary fixation the tissue
was rinsed in water and processed through dehydration, plastic, and
post-staining as described.
In cases where isolated strains of bacteria were prepared for
electron microscopy, liquid suspensions of the bacteria were used.
Solutions were changed by centrifuging the bacteria in a table-top
centrifuge and resuspending the pellet. After post-fixation the
bacteria were dehydrated and embedded in the form of an agar-suspended
pellet. Preparation of bacteria required much less time than hydrilla
Bacteria Isolation Media
Media referred to in Chapter 3 (Results and Discussion) are
described here. In cases where media contain agar, the agar was first
melted by boiling for 1.0 min, then autoclaved, and subsequently
cooled to 370 before being poured into plastic, sterile petri dishes.
Autoclaving was at 121 for 15 min. All media were stored at 4.
The following media were used:
Trypticase soy agar (BBL), a general purpose nutrient agar.
Trypticase soy broth (BBL) supplemented with 2.5 g/1 dextrose,
a general purpose nutrient broth.
Hutchinson Salt Media
(From Rodina, 1972, pg 202) these media are designed for enrich-
ment in celluyltic bacteria by limiting the carbon source to a form of
cellulose. All the other nutrients in these media are inorganic salts:
NaNO3, 2.5 g/l; K2HPO4, 1.0 g/l; MgSO47H20, 0.3 g/l; CaC12'6H20, 0.1
g/1; NaC1, 0.1 g/l; and FeC13'6H20, 0.01 g/l. The pH was adjusted to
7.2-7.3 with 1 N HC1 or NaOH.
A crude cell wall extract of hydrilla stems and leaves was prepared.
This was made by placing 100 g fresh hydrilla apical stem segments
(8-15 cm in length) in a blender, adding 300 ml deionized water, and
blending at high speed for 1.0 min. The resulting homogenate was
centrifuged at 500 x g for 10 min, and the pellet was resuspended in
deionized water. This suspension was boiled for 15 min, and filtered
through' Whatman No. 1 filter paper in a Buchner filter. The filter
paper was washed with 100 ml 95 ethanol, and the cell walls were
placed in a 700 oven for 12 h. The dried walls were ground in a Wiley
mill (40 mesh screen) and stored in a dessicator.
Three media were made using the Hutchinson salts.
HAS median contained 2.0 g of hydrilla cell wall extract in 1.0 1
deionized water with the described salts added and agar to 1.5% (vw/v).
Carbon source for this medium was 5.0 g powdered cellulose (Sigma
Chemical Co.) in 1.0 1 deionized water. Hutchinson salts were added
in the amounts described, as was agar to 1.5% (w/v).
This medium was identical to SPCA but with the addition of a Whatman
No. 1 filter paper to the surface of the medium after it had been poured
into the petri dishes. The filter paper was autoclaved separately.
Two other supplementary media were made as comparison media:
2.0 g of the cell wall extract in 1.0 1 of deionized water supp-
lemented with agar to 1.5% (w/v). The pH was adjusted as described
for the salts.
TSA medium as described, with the addition of a Whatman No. 1
filter paper to the surface of the medium after it had been poured
into the petri dishes.
Berg et al. Salt Media
(From Berg, et al. 1972) These media were designed for enrichment
in cellulytic bacteria by limiting the carbon source to a form of
cellulose. The salts differ slightly from the Hutchinson salts and are
at lower concentrations. These media are liquid media utilized in
enrichment cultures and contain the following inorganic salts: NaNO3,
2.0 g/l; K2HPO4, 0.5 g/l; MgSO4'7H20, 0.2 g/l; CaCl2'2H20, 0.02 g/l;
FeSO4'7H20, 0.02 g/l; and MnSO4'H20, 0.02 g/l. The pH was adjusted
to 7.2-7.5 with 1 N HC1 or NaOH.
Three media were made using these salts.
Salts with 1.0 g microcrystalline cellulose (Sigma Chemical Co.)
and deionized water to form 1.0 1 of medium.
Salts with 1.0 g hydrilla cell wall extract (previously described)
and deionized water to form 1.0 1 of medium.
Salts in deionized water to form 1.0 1 of medium; after the medium
was dispensed into flasks, strips of Whatman No. 1 filter paper (1 cm x
4 cm) were added as the carbon source. The filter paper had been pre-
Made as a comparison medium SGSH contained 1.0 g hydrilla cell wall
extract in 1.0 1 of spring water (Silver Glen Springs, Florida).
Since these were liquid media, they were placed in 25-ml Erlenmeyer
flasks, 10.1 ml per flask, to give a large surface area for oxygen
The infusion media described relies solely on nutrients contained
within the hydrilla plant for growth of bacteria.
Reasonably clean hydrilla plants obtained from the mouth of a spring
(Silver Glen Springs, Florida) were rinsed thoroughly in deionized water.
Samples of 100 g of fresh apical stem segments (8-15 cm) were placed
in a blender with 300 ml deionized water and blended at high speed for
1.0 min. The homogenate was allowed to stand at room temperature for
20 min, after which it was filtered through four layers of cheesecloth
and centrifuged at 500 x g for 10 min. The tar-colored, viscous super-
natant was aseptically filtered through a 0.8 u Millipore filter. The
filter retained the green material, but the solution remained viscous
and tar colored. Aliquots of 10.0 ml were aseptically transferred to
sterile (autoclaved) 25-ml Erlenmeyer flasks.
This medium was made similarly to Infusion I with the exception
that, after centrifugation, sterilization was accomplished by auto-
claving. The infusion solution was then dispensed into 25-ml Erlenmeyers.
This infusion medium was prepared like Infusion I, except filtration
was only with four layers of cheesecloth, after which 250-mi aliquots of
the solution were dispensed in 500-ml Erlenmeyer flasks and then auto-
This medium was made like LI with the exception- that the medium
was solidified with agar to 1.5% (w/v).
Plant Sterilization Methods
Due to the aquatic habitat of hydrilla, obtaining plants completely
free of epiphytic organisms was difficult. Indeed, the wet milieu of
the hydrilla leaf is an ideal environment for microflora. Realizing
that essentially all bacteria are aquatic, and considering the nature
of these studies, it seemed appropriate that plants subjected to
inoculation with bacterial cultures should best be as free as possible
of any contaminating epiphytic bacteria.
Reproduction in hydrilla is primarily through vegetative means;
however, stem fragments are not resistant to sterilizing solutions
and turions are damaged somewhat by these solutions. Tubers are the
most resistant part of hydrilla tissue to damage from sterilization
procedures. Therefore, tubers obtained from the hydrosoil in shallow
areas of Rodman Reservoir (Kenwood, Florida) were used as a source for
the initial production of sterile plants. Various concentrations,
combinations, and times of exposure with the sterilizing solutions
were tested. The method that best preserved tuber viability and
achieved the highest degree of sterilization is described. The tubers
were rinsed thoroughly in distilled water and placed in 95% ethanol
for 1 min. They were then soaked for 20 min in 2% sodium hypochlorite
to which a few drops of Micro detergent had been added as a surfactant.
This was followed by three aseptic rinses in autoclaved, distilled water.
The tubers were then soaked for 20 min in 0.5% HgC12. This was followed
by six aseptic rinses in autoclaved distilled water. If the tissue was
not used immediately it was stored aseptically at 4.
Germination tests in sterile petri dishes with moistened filter
paper gave a germination of 80 with this treatment.
On v.an/ occasions a method was needed that could produce many
sterile plants in a short period of time; this could not be done with
tubers. Waite and Mitchell (1976) used antibiotics to sterilize the
marine alga Utva -ac tuca. This method was tested on 8-12 cm apical
stem segments of hydrilla.
These segments were obtained from hydrilla plants found growing
at the mouth of Silver Glen Springs. In all of the studies presented
here, the source of relatively epiphyte-free plants was this spring.
Presumably, the high rate of flow of water which had percolated through
many meters of soil, combined with constant temperature conditions,
helped to keep these plants clean (Odum, 1957).
Selected stem segments were thoroughly rinsed in autoclaved dis-
tilled water, placed in previously autoclaved 30 x 100 mm culture
tubes, and subjected to the following antibiotic solution for 2 h
or 12 h: penicillin, 1000 mg/l; neomycin, 500 mg/l; polymyxin B,
300 mg/l; and streptomycin, 500 mg/l. These were dissolved in auto-
claved distilled water. After treatment the plants were rinsed
aseptically with autoclaved distilled water. Alternative methods
of obtaining aseptic plants included using plant tissue cultures.
The use of plant tissue cultures as a method to produce aseptic
plants and to study morphogenesis is well documented (Gamborg et al.,
1976). These methods were used in studies on hydrilla.
In all cases the salt medium developed by Murashige and Skoog
(1962) was used, but variations were made in salt concentrations,
amount of sucrose, and amounts and kinds of hormones used.
Initially, stem or root organogenesis was attempted. The explant
used was tuber material, sterilized as described, in which the unde-
veloped shoot within the tuber was excised under aseptic conditions.
These were placed in 18 x 150 mm culture tubes and capped with stain-
less steel closures. The liquid mediumwas as follows: Murashige and
Skoog salts, 30 g/l sucrose, 100 mg/l myo-inositol, and either hormones
to stimulate shoot development (2.0 mg/l kinetin and 0.1 mg/l NAA),
or hormones to stimulate root development (0.1 mg/1 kinetin and 2.0
mg/l NAA). The pH was adjusted to 5.7-5.9 with IN HC1 or NaOH, and
aliquots of 10.0 ml were dispensed into the tubes and then autoclaved.
Later experiments were conducted using solid medium consisting of:
Murashige and Skoog salts, 30 g/l sucrose, 100 mg/l myo-inositol, 1%
agar, 10.0 mg/l IAA, and 0.04 mg/l kinetin. Explants were tubers,
sterilized as described, tuber parenchyma (0.5 cm3 pieces removed
aseptically from the middle of sterile tubers), and immature shoot
tips excised from sterile tubers.
It was decided to conduct experiments to produce callus tissue,
an indirect source of aseptic plants. The media consisted of:
Murashige and Skoog salts with 30 g/l sucrose and 100 mg/l myo-
inositol, used at different percentages; this was supplemented with
varying concentrations of the callus-inducing hormone 2,4-D. In
addition, the lower percentages of the salt media were supplemented,
in some cases, with Fe:EDTA. Basiouny et al. (1977) found optimal
growth of hydrilla with 8 ppm iron and that this element could be a
limiting nutrient. The specific media compositions are shown in
Table 1. The pH was adjusted to 5.7-5.9 with IN HC1 or NaOH, and
15.0 ml aliquots were dispensed into 20 x 100 mm culture tubes and
25-mi Erlenmeyer flasks. These were then plugged with cotton wrapped
in cheesecloth and autoclaved. Explants were tubers, sterilized as
described, tuber parenchyma (0.5 cm3 pieces removed aseptically from
the middle of sterile tubers) and sterile stem segments. These segments
were obtained from 12-day-old germinated tubers having stem lengths of
Table 1. Media composition for callus tissue cultures.
1/ Standard salts from Murashige and Skoog (1962) with 30 g/1
sucrose and 100 mg/l myo-inositol.
2/ The ppm iron is in the form of Fe:EDTA.
3-5 cm; cut segments were 1-2 cm and included at least one node. In
a few cases only leaves from these stems were used as explants.
All tissues were cultured at room temperature (24-270) 14 h of
light (fluorescent, Cool White) at approximately 1/4 full sunlight.
All liquid cultures were shaken at 40 rpm.
Toxin was extracted from liquid cultures of bacteria by centrifuging
at 500 x g for 10 min and then filtering through Millipore filters of
pore size 0.45 U. This cell-free extract was reduced to dryness under
a vacuum. The residue was extracted in a small amount of methanol,
which was again brought to dryness under a vacuum. The residue was
resuspended in a small amount of distilled water. Using this method
148 ml of toxin was extracted and reduced to 8.55 ml in water. This
solution was used in several bioassay experiments.
Bioassays of the toxin included its affect on lettuce seed germ-
ination (Miller, 1967) and the Avena coleoptile bioassay for auxins
and ABA (Morre'and Key, 1967).
RESULTS AND DISCUSSION
Field Observations of Annual Decline
Annual decline of Hydria ve&ticilUata (L.f.) Royle is a subject
of much interest because of the growing economic importance of this
aquatic weed. During this decline a rapidly growing, high competi-
tive aquatic plant shows an abrupt change in growth characteristics
ultimately leading to death. What could cause such a change? The
establishment of the plant in a given area causes permanent ecological
changes that often have an adverse economic impact on man. Investi-
gation of the cause of annual decline of hydrilla is justified when
one realizes the knowledge gained may be utilized in management programs.
The author first observed annual decline in hydrilla growing in
Orange Lake, Florida in July, 1976. This lake is in north central
Florida and covers 5000 ha. As may be seen in Figure 1, a dense mat
of hydrilla was found to cover over 90% of the lake surface. Annual
decline occurs every year in the late summer months after the plants
have grown to the surface of the lake and formed an extensive mat.
The exact time of year during which annual decline occurs varies in
Orange Lake. This has been found also to be the case where annual
decline occurs in other bodies of water. Another lake in north
central Florida, Rodman Reservoir (Kenwood, Florida), was observed
by the author to contain large areas of hydrilla undergoing annual
Hydrilla mat covering 90% of the surface of Orange
Lake, Florida. Photograph was taken in July, 1976.
decline. This is a man-made lake with water quality considerably
different from that found in Orange Lake (W. T. Haller, personal
communication). As in the case of Orange Lake, the hydrilla mat
formed in Rodman Reservoir was extensive and was found to go through
annual decline in the late summer months, the exact time varying from
year-to-year. In Florida (and probably elsewhere as well) annual
decline in hydrilla occurs in all regions in which the plant grows.
Hydrilla found in lakes and canals of central Florida undergoes annual
decline at roughly the same time of year and growing season as that of
south Florida. Similar reports of annual decline in hydrilla-infested
areas outside the state of Florida indicate that annual decline occurs
during the same time of year in widely varying regions. This suggests
decline is not primarily a temperature response.
What, then, are the conditions requisite to annual decline? The
hydrilla stand invariably forms a dense mat over the surface of the
water before annual decline occurs, shading a significant amount of
its own leaf surface and reducing photosynthesis in all but the upper
regions of the plant canopy. The plant also harbors well-developed
populations of epiphytes, including bacteria, fungi, and large amounts
of algae. At this time of year the weather is the warmest of the
growing season and light intensities are at a maximum. Thus, the
growing conditions immediately preceding annual decline onset in late
summer appear to be condusive to maximum plant productivity.
Symptoms of annual decline may offer some insight into its cause.
Leaves lose their green pigmentation, most becoming yellowish-brown,
though some become transparent (Figure 2). Often the leaf margins
deteriorate, giving the leaf a jagged appearance. Many leaves contain
Diseased hydrilla plant. Photograph was taken at
Orange Lake, Florida in July, 1976.
large amounts of epiphytes, algae being the most obvious (although
bacteria and fungi are also present). Abscision of leaves to varying
degrees may also occur. Often the occurrence of annual decline symp-
toms within a hydrilla mat is confined to individual stems, which are
interspersed among apparently healthy stems (Figure 3). Stems undergo
the same pigmentation changes as the leaves and also become brittle,
breaking at nodes. The fragments of chlorotic stems float to the
surface of the water (Figure 3). In the mat as a whole, annual decline
is manifested as the disappearance of plants in localized areas. The
spottiness noted is comparable to disease development in fields of
terrestrial crops. As the growing season comes to an end, the majority
of the mat has disappeared. The plants, breaking up as described,
eventually sink to the hydrosoil.
There are several possible explanations of annual decline and these
were investigated. It is possible that hydrilla loses photosynthate
through its leaves and that epiphytes utilize this material in their
growth (Allen, 1971). The epiphytes, in turn, may interfere with CO2
diffusion or light transmittance to the hydrilla leaf (Sand-Jensen,
1977), reducing its photosynthetic rates. Epiphytic flora may affect
the plant by the production of growth regulators (Waite and Mitchell,
1976). Alternatively, plant pathogens may be present in the hydrilla
mat, and conditions condusive to their pathogenicity could develop
within the hydrilla mat during the growing season for hydrilla.
Similarly, epiphytes or planktonic microflora may produce toxins that
are detrimental to the hydrilla plant.
Hydrilla plants showing annual decline symptoms.
Yellowish-brown stem fragments are floating in the
Structural Relationships of Epiphytic Flora
The leaf tissue of hydrilla was found to support large populations
of epiphytes. A transverse section of the leaf is presented in Figure
4. The upper adaxiall) cell layer is highly vacuolate, a structural
feature commonly known to cause problems when the tissue is fixed for
electron microscopy. The lower abaxiall) cell layer is composed of
cells less than one quarter the size of those found in the adaxial
layer. The abaxial cell layer is composed of transfer cells, which
are formed as the leaf matures. The smaller size of the abaxial cells
is due to a much reduced vacuole, and the cells of this layer have many
organelles associated with high metabolic activity (chloroplasts, mito-
chondria, endoplasmic reticulum).
Various aspects of the fine structure of hydrilla leaves that may
be involved in DOM secretion by this macrophyte are shown in Figure 5.
The close proximity of chloroplasts to the wall labyrinth of the transfer
cell, as seen in Figure 5(a), suggests DOM is available to these wall
protuberances. These are also shown in Figure 5(b). This micrograph
shows a close proximity of the plasmalemma to a wall protuberance.
This particular tissue was embedded in the Spurr's-Mollenhauer plastic
mixture. Other plastics used were found to leave an electron-transparent
gap between the plasmalemma and wall protuberances. The active or
passive loss of DOM through the plasmalemma would allow apoplastic
movement of this material. This could result in DOM loss to the
external environment because the cuticle of the hydrilla leaf, shown
in Figure 5(c), is very thin and presumably does not serve as a barrier
to movement of materials. Symplastic movement of cellular material in
hydrilla is probably an important transport mechanism. Great numbers
Transverse section of a hydrilla leaf showing the two
contiguous epidermal layers. The upper adaxiall) layer
is highly vacuolate and much larger than the lower
abaxiall) layer, which is composed of transfer cells.
Fine structure of hydrilla leaf tissue implicating it
in DOM secretion.
a. Close proximity of chloroplasts to wall protuberances
of the transfer cell. x 20,000.
b. Plasmalemma in close association with a wall protube-
rance. x 49,000.
c. The thin cuticle of the hydrilla leaf, distance between
the arrows is 0.06 p. x 215,000.
d. Plasmadesmatal connections between two transfer cells.
~ ..,1 ....
of plasmodesmata (Figure 5(d))were found to connect cells within and
between the two contiguous epidermal layers. Because of the presence
of many large vacuoles (temporary storage sites of photosynthate) in
the adaxial cell layer and extensive systems of symplastic and apo-
plastic movement, it is proposed that a certain fraction of DOM produced
by the hydrilla leaf is lost to the external environment, in accordance
with the model Wetzel (1969) developed for the submersed Najab leexZiu6.
It was found that hydrilla leaves develop an extensive epiphyte
population, and that this population appears to exhibit characteristics
described by Allen (1971) for epiphytes influenced by macrophyte DOM
secretions. Hydrilla with epiphytic flora shown in the following
discussion was obtained during the summer months of 1976 (July-August)
from experimental ponds at Biven's Arm Lake (University of Florida
campus), Orange Lake, Florida, or Rodman Reservoir (Kenwood, Florida).
Populations of epiphytic flora on hydrilla leaves were found to be
greatest during the summer months as hydrilla reached its maximum
biomass. Little is known of in situ photosynthetic rates of hydrilla,
but is probable that, as conditions of temperature and light become
progressively more favorable, photosynthetic rates in the hydrilla mat
as a whole increase to a maximum in the early summer. The author pro-
poses that this is accompanied by increasing rates of DOM secretion from
hydrilla leaf tissue and a concommitant increase in epiphyte populations
of the hydrilla leaf.
Figures 6-10 show some of the typical epiphytes occurring on hydrilla
leaves during rapid summer growth. Abundant amounts of blue-green algae
occur on leaf surfaces in the summer (Figure 6, surface view; Figure 7,
ultrastructural view, adaxial surface). The filamentous green alga shown
Surface view of a hydrilla leaf showing thread-like
epiphytic blue-green algae. x 1250.
Epiphytic blue-green algae on the surface of the adaxial
side of the hydrilla leaf. x 52,000.
.' k^ .. *. -._' -",,
;^' .r -, ,
" ;- I ':S .jB
.. ., .- -- ,
in Figure 8 is attached to the leaf surface with what appears to be
thread-like appendages. No algae were found to penetrate hydrilla leaf
tissue. Figure 9 shows two green algae attached to the surface by what
appears to be a mucopolysaccaride slime (Figure 9(c); also see Figures
11-13). The alga in Figure 9(a) has associated epiphytic bacteria near
its point of attachment to the hydrilla leaf. Godward's finding (1934)
that algal epiphytes attach at depressions demarcating the cell bound-
aries of a leaf were verified in this study. In addition, it was found
that bacterial epiphytes generally settled at the same areas (Figure 10).
The bacteria commonly formed clusters in these regions (Figure 10(b),
(c)). As was pointed out by Godward, more algae were found on the
adaxial leaf surfaces. No such trend was found with epiphytic bacteria
per se, though they were often in close association with epiphytic algae.
These micrographs support the requirement for a close spacial relation-
ship between these organisms for macrophyte-epiphyte interactions
The gradual increase in both macrophyte photosynthesis and epiphyte
populations cause a cyclic phenomenon that increasingly stresses the
hydrilla plant, ultimately causing a reduction in net photosynthesis
as epiphytes cover leaf surfaces reducing gas exchange and actinic
radiation. This is a turning point in the growing season of the
hydrilla plant, the point where annual decline begins to become
The cyclic phenomenon referred to is that caused by an epiphyte-
macrcphyte feedback mechanism: 1) increasing DOM secretion by the
macrophyte over the growing season causes 2) increasing epiphyte
populations which, in turn, 3) stimulate increased amounts of DOM
Point of attachment of a filamentous alga on the hydrilla
leaf. x 1250.
; ,I o
Filamentous algae attached to the hydrilla leaf surface.
a. x 3,100., b. x 8,000.
c. High magnification of algal attachment shown in (b).
' jI. b
q.. l. 1'."
*~7 .. r ?+
*- 0 *-
4- S- 0
+ 0 .0
4-3 CO -
^ Jf'sS ... '
~t -y ". ,,'
~, f ^ a.
to be secreted by the macrophyte. This cycle increases to the point
where the macrophyte is weakened and additional stresses come to bear
in the system. The additional stresses include light attenuation by
the epiphyte stand, which also interferes with CO2 diffusion to mac-
rophyte leaf cells (Sand-Jensen, 1977). It is also possible that algal
epiphytes compete with the macrophyte for CO2. This is difficult to
ascertain because bacteria undoubtedly are producing considerable CO2
at this time. Allen (1971) suggested that bacteria produce growth
regulators, which may have an effect on the macrophyte. During these
stresses macrophyte demand for combined nitrogen probably falls, causing
an increase in epiphytes (Fitzgerald, 1969).
How epiphytic bacteria and algae interact is an important consider-
ation. Allen (1971) found epiphytic bacteria to actively uptake DOM
at low concentrations, while epiphytic algae uptake DOM, at higher
concentrations by a diffusion process. During maximum photosynthesis
oxygen levels are high in the hydrilla mat. Therefore, conditions for
growth of epiphytic bacteria are optimum and colonization occurs. Allen
also found epiphytic bacteria to stimulate the growth of epiphytic algae,
perhaps through the production of growth factors such as vitamin B12.
The epiphytic algae subsequently develop on the macrophyte leaves.
Allen also showed that epiphytic bacteria could metabolize substrates
furnished by epiphytic algae. Perhaps both types of epiphytes create
a temporary symbiosis, each benefitting from the other's metabolic
Cell surfaces of epiphyte flora were stained using the alcian
blue/ruthenium red cytochemical method of Dykstra and Aldrich (1976).
These stains do not penetrate cellular material but cause a staining
of surface-associated molecules containing glucose (or other sugar)
moieties, including glycoproteins and mucopolysaccarides. That epiphyte
populations adhere to the macrophyte leaf by secreting these types of
molecules is shown in Figures 11-13. Epiphytic bacteria exhibit exten-
sive surface staining with alcian blue/ruthenium red, as shown in
Figure 11. The stain revealed a "polarity" of the extracellular coat,
as seen in the region distal to the point of leaf attachment. The
stain shows the extracellular material to be radiating outward on
this region (Figure 11(a), (c), (d)). Bacteria have been shown to be
capable of secreting extracellular glycoproteins (Glenn, 1976). It is
probable that these epiphytic bacteria excrete a variety of compounds
within the epiphyte complex, compounds which could affect macrophyte
or epiphytic algal growth (Allen,1971). The extracellular coat of
epiphytic algae stained with alcian blue/ruthenium red is shown in
Figure 12. Figure 12(a) shows the characteristic secretion of a
filamentous alga and its contact with the adaxial surface of a hydrilla
leaf. In Figure 12(b) a blue-green algae adheres to this substance.
As seen in Figure 12(c), the extracellular secretions of blue-green
algae were much less, in comparison with the larger green algae. In
Figure 13, epiphytic bacteria are seen to be in intimate association
with the algal extracellular coat. It has already been proposed that
a significant interaction occurs between epiphytic bacteria and algae.
A variety of epiphytic bacterial types are associated with hydrilla
(Figure 14). Gram-negative rods (Figure 13(a)) were the most predom-
inant types. Rod bacteria were often attached to the leaf surface in
the manner shown in Figures 14(b), (c); this allowed for a higher
*r- S- C
Figure 12. The extracellular coat of epiphytic algae as demonstrated
by staining with alcian blue-ruthenium red.
a. x 52,000., b. x 30,000., c x 33,000.
-^, t :i
population density of epiphytic bacteria. The other commonly found
type of epiphytic bacteria were cocci (Figures 14(d), (e)).
The discovery that the hydrilla leaf contained both cell-wall-
degrading and intracellular bacteria is significant to the understanding
of events leading to annual decline of hydrilla. Cell-wall-degrading
bacteria are shown in Figure 15. Bacteria were found to be capable of
degrading any type of cell wall in the hydrilla leaf; however, they are
unable to degrade the cuticle (Figure 15(a)). The bacteria, both rod
and cocci types, were found intimately associated with cell wall com-
ponents (Figure 15(b), (c)). The rod bacterium shown in Figure 15(d)
has a distinctive cell wall, perhaps helpful in the secretion of exo-
proteins used in cell wall lysis (Glenn, 1976). The ultrastructure
of degraded cell wall components is shown in Figure 16. It is apparent
that cell wall polysaccarides are broken down into smaller units which
are presumably used by the bacteria. Intracellular bacteria are shown
in Figure 17. Both rod and cocci types were found within hydrilla leaf
cells. The cell of figure 17(a) has become lysed, probably due to the
intracellular and wall-degrading bacteria present. In Figure 17(b)
it may be seen that bacteria are capable of degrading the fibrous
(partly proteinaceous) layer found in the adaxial cell layer of the
hydrilla leaf (Pendland, 1976).
Cytologically, the hydrilla leaf cell reacts to the presence of
these bacteria by undergoing pathological responses typical for cells
under bacterial attack (Fox, 1972), as seen in Figure 18. Cytoplasmic
or; ,,elles clump together and membrane lysis occurs (Figure 18(a), (b)).
Chloroplasts react to bacterial-induced stresses by forming many plast-
oglobuli (Figure 18(c)) and are the first organelles to become completely
Figure 14. Types of epiphytic bacteria found on hydrilla leaves.
a. x 33,000., b. x 49,000., c. x 23,000., d. x 32,000.,
e. x 25,000.
4 .1 A
Figure 15. Bacteria degrading hydrilla leaf cell walls.
a. x 16,000., b. x 49,000., c. x 23,000., d. x 55,000.
* .< .
s^ ~ *; M' /'' -'*;'
Figure 16. High magnification of hydrilla leaf cell wall undergoing
bacterial degradation. x 50,000.
; -- I- -
S ,-** S > *
2 r .- .' ,,. '
"" ..., "- ;
: ..* ',,: ,,,, I .-
.. +. ,, + +. ,+,
a t. *.
+ %' 1
+ +r r. ".. r, '+ ', "
4 ,t~ ,, : ."+
.,. I eaf:, \
i 'U .. I. -
:;; -" i t: t
, t. -* -.
,, .+. ..
t ,, at .j-r
++ I. .." ,:f 1 1
. .. ,
. + ,. tA *. .
,', t .. -" .
Figure 17. Intracellular bacteria in hydrilla leaf tissue.
a. x 3,100., b. x 16,000.
0, W ~- .'
^ *^ .***/r
crtA ? r
: .^ ', ,,": ,.
i: 1, ~rf
4-- S- x D
- C C*
CD C -
0 0 0-
aI C -.
S .r- OO
disintegrated (Figure 18(b)). Other cytoplasmic organelles are lost,
eventually leaving only the cell wall "skeleton".
Epiphytic algae are often found in close association with cells
undergoing bacterial lysis (Figure 19(a)), perhaps using breakdown
products for their nutrition. However, lytic bacteria were found
associated with epiphytic algae, too (Figure 19(b)). This indicates
that a general decline of autotrophs occurs in s6Zt as late summer
conditions encourage a high activity of heterotrophs. Occasionally,
fungi (Figure 20(a)) and other organisms (Figure 20(b)) were found
to be implicated in this heterotrophic activity.
On a limited scale, the hydrilla leaf produces phenol-containing
cells, known to play a role in disease resistance in aquatic macrophytes
(D. Samuelson, personal communication). These phenolic compounds are
located within the vacuole (Figure 21(a), (b)). The phenol-containing
cells may be formed in localized areas as a group (Figure 22) or may be
distributed singularly throughout the leaf, as is shown in the senescent
leaf of Figure 23. Phenol-containing cells are resistant to attack by
It is evident, on a structural basis, that bacteria are capable of
lysing cells of the hydrilla leaf, perhaps after these cells have been
weakened by the stresses previously described. In a search for poten-
tial biocontrols of hydrilla, an attempt was made to isolate the lyso-
genic bacteria seen in ultrastructural studies of in situ hydrilla
Association of epiphytic bacteria and algae.
a. Blue-green alga in association with cell-wall-
degrading bacteria, possibly using break-down
products. x 12,000.
b. Epiphytic alga being degraded by bacteria. x 6,300.
Organisms, other than algae and bacteria, associated
with annual decline in hydrilla.
a. Fungi. x 6,300.
b. Unidentified organism. x 16,000.
~J~.r sr .3 F
-- 'o w
tn 0- 7
U "" U
,--- C C-
~tl.li~r'4 r ~b~
*I / I"
'I f ~
L. 3~' L' Z
Fijigr' 22. Localized production of phenol-containing cells in the
hydrilla leaf. x 200.
Figure 23. Distribution of phenol-containing cells in a senescent
hydrilla leaf. x 200.
Laboratory Studies of Epiphytic Bacteria
Isolations of epiphytic bacteria and inoculations of epiphyte-free
plants with these isolates were made in order to investigate in situ
relationships of these organisms during annual decline of hydrilla.
Plants which were inoculated with bacteria cultures had to be as
free as possible of bacteria initially. Several methods of obtaining
sterile plants were investigated. Hydrilla plants obtained from ster-
ilized tubers were found to be comparatively clean. However, this
method could not be used to obtain sufficient numbers of plants over
long periods of time due to difficulties in obtaining tubers. Rela-
tively clean plants from Orange Lake, Florida were treated with anti-
biotic solutions described in Chapter 2. This proved to be unsatisfactory
because the antibiotics caused a complete chlorosis of the plants, even
with treatment limited to 2 h. The initial tissue culturing experiments,
using combinations of kinetin and NAA in liquid Murashige and Skoog salts
(1962), gave sprouting of the shoots excised from tubers (hormones were
2.0 mg/l kinetin and 0.1 mg/l NAA). However, the growth of shoots was
limited, and the plants displayed abnormalities that may have been due
to the high concentrations of salts. Culturing hydrilla tissue on solid
media resulted in sprouting of tubers and shoots excised from tubers,
but these became desiccated after a few days. Cultures in which the
hormone 2,4-D was used (see Table 1) did not produce callus tissue
after 6 weeks. However, medium A, containing 2,4-D at 10-5 mM and
full strength Murashige and Skoog salts, caused the production of
over 30 shoots with roots from a single stem segment explant after
5 weeks. This tissue was healthy in appearance and, in future studies,
this method may be the best way to produce aseptic plants.
The source of relatively epiphyte-free hydrilla plants was Silver
Glen Springs, Florida. Plants were removed from the mouth of the
springs and immediately placed in 38 x 300-mm autoclaved test tubes
filled with 250-ml spring water and capped with plastic closures.
Plants obtained in this manner remained comparatively free of epiphytes
for more than 3 months.
Hydrilla plants undergoing annual decline were obtained from Orange
Lake, Florida. These plants had large epiphyte populations. Inocula
from selected leaves were streaked on TSA plates and on plates of all
the Hutchinson salt media (Rodina, 1972; p. 202) listed in Chapter 2.
Identification of the resulting colonies was by colony morphology.
More extensive characterizations were to be made if pathogenic types
were found. Growth on TSA was predominately of three colony types:
1) a pinpoint white colony, 2) a smooth white colony, and 3) a wrinkled
white colony. No growth was observed on any of the Hutchinson media.
Larger amounts of inoculum were obtained from isolated colonies of
these three types by incubating each in 125 ml of TSB contained in
250-ml Erlencleyer flasks. Prolific growth occurred after 24 h.
Bacteria from 125 ml of TSB were washed and resuspended in 250 ml
sterile water. This solution was placed in a sterile 38 x 300-test
tube and a 15-cm apical sprig of clean hydrilla added. After 4 weeks
no detrimental effects were noted for all three isolates. These ex-
periments were repeated for two isolates from plants growing in exper-
imental ponds at Biven's Arm Lake (University of Florida campus) and
for three isolates from plants growing in Rodman Reservoir (Kenwood,
Florida). In no case was pathological stress noted with the inoculated
, drilla plants. Experiments were continued; however, inoculation was
with the various media described as Berg et al. salt media (1972) in
Chapter 2. The rationale being that selection, using enrichment
cultures such as these, for cellulytic bacteria present in hydrilla
leaf cell walls would lead to isolation of a pathogen. Inoculation
was with plants from the same three areas described above. Three to
four leaves with heavy epiphyte populations were used in every inoc-
ulation. Growth only occurred on SFP medium, which contained nutrient
salts and cellulose (as a carbon source) in the form of filter paper.
After 6 days growth, the filter paper had disintegrated at the liquid-
air interface. The culture stained the filter paper bright orange where
it touched the inorganic salt solution. Berg and his colleagues (Hofsten
et al., 1971) found similarly colored colonies in their cultures of
Sporocygtophaga, a cellulose-degrading saprophyte. They found that
mixed cultures of bacteria were more effective than pure cultures in
degrading complex biological materials such as cellulose. Pieces of
the orange filter paper were examined ultrastructurally. As seen in
Figure 24, a mixture of rod and cocci bacteria were found to be in
close association with the cellulose fibers of the filter papers.
Hydrilla plants were inoculated with these cultures of bacteria.
.*ain, the plants failed to develop pathological symptoms. The
difficulties found in showing pathogenic bacteria to be present could
be attributed to the selection of isolation media. It is difficult to
simulate in s.Ctu conditions with laboratory culture media (Brock, 1971),
and therefore, isolations reported thus far could have been selecting
for bacteria other than those pathological to hydrilla in the field.
That culture media can misrepresent field conditions was shown by
Figure 24. Cellulytic bacteria cultured on SFP medium.
a. x 20,000., b. x 4,600., c. x 23,000.
' *- .*
NA, **< P'
'b 't." i
- *. 4
-NA. 7 I.
Overbeck (1972). He found the vertical distribution of aquatic bacterial
biomass to be different from the distribution of bacteria grown on nutrient
agar and no mineral agar.
Preparations were then made to inoculate clean hydrilla plants
directly with epiphyte-containing plants from the field. In order to
get the large amounts of inoculum needed in these experiments, field
samples of epiphyte-containing plants were placed in 5-liter containers
with 10 times their number of clean plants (gathered from Silver Glen
Springs, Florida). The plants were covered with autoclaved distilled
water and incubated under the same conditions as other plants used in
inoculation experiments (approximately 1/2 full sunlight from Cool White
lamps, 16 h light-8 h dark). After 6 days, 10.0 g samples were used as
inoculum for healthy hydrilla plants, set up as described for bacterial
monoculture inoculations. In addition, 25 ml of the liquid from the
5-liter containers were added to the same plants. In 5 days the inoc-
ulated plants began to exhibit pathological symptoms of what has come
to be known as "brittle stem". Inoculations from all three sources
produced the symptoms. These results are shown in Table 2.
Inoculating healthy hydrilla plants with a mixture of bacteria
capable of producing brittle stem produces distinctive symptomology.
After the incubation periods (3-5 days) the stem becomes chlorotic and
breaks apart at the nodes (Figure 25 (a), plant in the tube on the right;
Figure 25 (b)). With agitation the stem disintegrates, and the leaves
fall to the bottom of the test tube (Figure 25 (c)). Remnants of the
stem, on close examination, are found to consist of minute pieces of
stem material. Leaves remain green and intact, with the exception of
the region adjacent to where the leaf was formerly attached to the
LO Lo LO L D o
CON C\ rN CM ui
LO LO LC) LT CD
CD C\j r-O -LO
tl- ::I- -Zl -It lz
Symptoms of hydrilla infected with a mixture of bacteria
causing brittle stem.
a. The tube on the left was a control, the tube on the
right was inoculated with 12.5% brittle stem inoculum.
b. Breakage of nodes on the stem.
c. Pile of leaves in the bottom of the culture tube,
resulting from agitation of the inoculated tube
shown in (a).
d. Abscised leaves.
stem (Figure 25(d)). If the leaves are left in the inoculation solution
the midrib eventually decomposes to the pointthat the leaf divides into
two pieces. Subsequent deterioration of the leaf is secondary, the leaf
serrations and other phenol-containing cells being most resistant to
decomposition (Figure 26).
Ultrastructural studies of brittle stem show that bacteria are the
associative organisms (Figure 27). Bacteria are found in the walls of
stem tissue fragments (Figure 27(a)). Material in the region of the
leaf that was formerly attached to the stem show the bacteria to be
present within cell walls (Figure 27(b)) and air spaces (Figure 27(c)).
Hydrilla stem tissue contains aerenchyma. The rapid disintegration of
the stem indicates that the mixture of bacteria is highly aerobic.
This is also shown by the presence of large numbers of bacteria in the
air spaces of the leaf. As brittle stem develops, a thick scum of
aerobic bacteria forms on the top of the liquid in the test tube.
Figure 28 shows the kinds of bacteria found in this scum. Brittle
stem appears to be caused by a mixture of bacteria.
Attempts were made to isolate a single pathogenic strain of bacteria
from this mixture. Portions of the inoculum fluid, stem fragments, and
leaves were used to inoculate the various media described under "Infusion
media" in Chapter 2. Infusion I was not used, as contaminating bacteria
were not removed by the 0.8 p Millipore filter. Infusion II and LI gave
the same results in respect to culturing bacteria. The methodology
included inoculating the liquid infusion media, which was used in
enrichment cultures. Of all media tested, infusion media were considered
to be the closest approximation of i' i'La conditions, as the sole source
of nutrients was liquified hydrilla plants. Over 2-day periods, loopfuls
Figure 26. Secondary degradation of hydrilla leaf inoculated with the
brittle-stem bacteria. The phenol-containing cell of the
serration is resistant to break-down. Diatoms are seen
to invade the dead leaf cells. x 500.
Ultrastructure of hydrilla plants inoculated with brittle-
stem bacteria in the laboratory.
a. Cellulytic bacteria found in stem fragments. x 12,000.
b. Bacteria within cell walls in the region of the leaf
closest to the point of abscission. x 2,000.
c. Aerobic bacteria located within an air canal of the
leaf. x 4,600.
d. High magnification of bacteria found localized within
air canals. x 20,000.
*1 i,~ -
* '.: : .. ,
-S.t .. ..
+ **c ',*.
* eo, ac "*
' ^ ^
I t- '