PLACE II RETURN BOX to roman this checkout from your mood.
TO AVOID FINEB rgtumpn or baton duo duo. '

DATE DpE "DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

I |

MSU In An Affirmative Action/Emu Oppommlly lmthlon
W

 

 

 

”3-9.1

 

HYDROGENOTROPHIC ARCHAEA AND BACTERIA ASSOCIATED WITH
THE HINDGUTS OF TERMITES: IN VIT R0 AND IN VIVO STUDIES

By

Jared Renton Leadbetter

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Microbiology

1997

ABSTRACT

HYDROGENOTROPHIC ARCH/{EA AND BACTERIA ASSOCIATED WITH
THE HINDGUTS OF TERMITES: IN VIT R0 AND IN VIVO STUDIES

By

Jared Renton Leadbetter

Three morphologically distinct, HZ/Coz-utilizing methanogens were isolated from
gut homogenates of the subterranean termite, Reticulitermesflavipes (Kollar)
(Rhinotermitidae) collected in Michigan and Massachusetts. Their cell wall ultrastructure,
Gram-positive staining reaction, resistance to cell lysis by chemical agents, and narrow
range of utilizable substrates were typical of species belonging to the
Methanobacteriaceae. Analysis of the nearly complete sequences of their small subunit
rRNA-encoding genes confirmed this affiliation and supported their recognition as 3 new
species of Methanobrevibacter: M. cuticularis; M. curvatus; and M. filiformis. UV
epifluorescence- and electron microscopy confirmed that cells of similar morphologies
were the dominant methanogens in the specimens of R. flavipes studied, and that they
colonized the peripheral, microoxic region of the hindgut, i.e. residing on or near the
hindgut epithelium. This is the first detailed description of an important and oft-cited, but
poorly understood, component of the termite gut microbiota.

Acetogenesis dominates methanogenesis as an H2 sink in the gut R. flavipes and
of most other wood-feeding termites, but the critical factors affecting the outcome of this

competition have yet to be identified. The mean H2 threshold for 7 termite gut acetogens

(3 that had been characterized previously; 4 that were isolated during the course of this
study) was determined to be 294 ppmv (parts per million volume) Hz; for the three
termite gut methanogens it was 42 ppmv H2. Live termites belonging to three different
species both emitted and consumed H2, and exhibited a mean H2 compensation point
(c.p.) of 815 ppmv. In comparison, a H2 c.p. exhibited by bovine rumen contents was
much lower, 28 ppmv H2. Thus, it appears that intrinsic affinity for H2 may not be

relevant to the competition for H2 by termite gut hydrogenotrophs.

To JoEllyn, Chloe, Mom, Dad, and the rest of my family (real and microbial): I
dedicate this thesis with love --and great appreciation for your roles in having made my

life such a wonderful experience.

ACKNOWLEDGEMENTS

As I have reached this, the end of my formal education, I have reflected back upon
a time prior to its start 22 years ago. It is remarkable how in many ways little has changed
in thattime, inasmuch as I am as eager to study insects now as I was then --or at any time

in the meantime.

I have mentioned this as a prelude to saying that I have not reached this milestone
having done it alone; rather, I have been aided by the considerable effort a large number
of individuals over those years. This is especially true regarding the maintenance of my
interests in biology during what would have otherwise been 12 years of scientific
monotony endured during my pre-collegiate education. For maintaining this interest I

have almost entirely my family to thank:

My sister Aletha, for rearing Milkweed bugs from egg to egg and crickets with
me, as well as for taking me to her college entomology lecture on one occasion [which
was quite a big deal for that first grader]; to my brother Garth, who showed me how to
maintain and investigate the Apis melifera-hive in our backyard; to my sister Briana,
whose insect collecting during her entomology summer course helped to spark my own

interest in insects, and whose beautiful biological illustrations I am still in envious awe.

My mother has always been one of my closest friends. Throughout my childhood,
she made every possible effort to provide me with access to a large variety of disciplines,
especially those in the area of biology. We were constantly perusing “field guides” to
identify and learn about this flower, or that bird, or whatever interesting item that we
could “discover”. Those experiences have proven to be invaluable. My mother has always
been a voracious reader and encouraged us to do the same; to these combined activities I

am sure I owe my profound enjoyment in biology and in reading the scientific literature.

I think that it genuinely surprised my father when I became fascinated with
microbes during my first college microbiology course, since prior to that I had primarily
been interested in insects. But then again, I did spend a childhood sailing in a boat named
Ignis F atuus; this, in retrospect, may have guided me towards one of my thesis research
topics. Since making the decision to become a microbiologist, I have had an absolutley
tremendous time discussing all sorts of bacteriological issues with my father, who has
also given me dozens of related books and reprints that have been used throughout my

graduate studies.

To Joan Wilce and her father, the late Prof. T. H. Hubbell, I thank for their
extending an invitation to join them on an insect collecting trip (which included 5000
miles of driving throughout the wilds of Mexico) needed as part of a revision of the
“camel cricket” genus Pristoceuthophilus. I especially would like to thank my mother,
who was most enthusiastic in agreeing to take them up on their kind offer «and ended up
with all of that hard, hard driving spent in close confines with an angst filled-teenager. It

was on that trip that I enjoyed my first thrill of “scientific discovery”, the collection of a

vi

small iridescent-black Pristoceuthophilus specimen that I had noticed under a shrub while
out strolling one evening in the forest near Uruapan, Michoacan. I cannot describe the
emotion felt by that eigth-grader when, on viewing the specimen for the first time, Prof.
Hubbell said: “You know, I have not ever seen anything even remotely like this one
before” --this coming from a man in his mid-eighties who had spent his adult life

studying such “crickets”. What an amazing time that was. Thank you.

After graduating high school, I had the good fortune to attend Goucher College,
and especially to receive such outstanding tutelage and career advice from Prof. Leleng
To and Prof. Ann M. Lacy. After graduating from Goucher, I immensely enjoyed and
benefited from summer spent in the laboratory of Paul Dunlap at the Woods Hole

Oceanographic Institutution studying luminous bacteria.

During my first year at here at Michigan State University, I was very fortunate to
meet and work with Jim Smith, whom since has provided me with excellent advice on a

number of occasions, and for which I am very grateful.

During my tenure in John Breznak’s laboratory I have enjoyed working with a
number of outstanding individuals: Sheridan Haack, who provided such a fine example of
everything that a senior laboratory colleague should be (not to mention furniturel); Stefan
Wagener, who provided me with some important words of wisdom, especially regarding
the highly desired cultivation of not-yet-cultivated bacterial species; Colleen Sweeney;
Edouard Miambi; Kwi Suk Kim, who provided me with the personal advice “be a flower

and let the bee come to you”; Dave Emerson, who told me to go take more micrographs

vii

until I could actually get one in focus; Tim Lilbum, to whom I should be continually
apologising for always eating the last bagel; Andreas Brune, whose critical mind set has
changed the way we all think about termites; and the late Rich Behmlander, who I
continue to rrriss and who seems to have personified the saying “all good things must
come to an end”: the world of the myxobacteria will never quite be the same without your
“myxo-roadmaps”. I am absolutely indebted to Mike Renner for having always stood by
in support without apparent reservation during the number of times of stress that have
defined good portions of my graduate academic experience: I will never be able to think

about a Cot curve without thinking about your help after our first graduate school exam.

During the early years of graduate school I shared apartment living and many fine
times with Houman Dehghani and Deanne Lehmann, two extraordinary individuals
indeed. Graduate school would have been much more difficult without their excellent

company.

I remember vividly that first phone conversation regarding my graduate school
application to this department: the calm and informative individual on the other end of
the line made quite an impression that has never become diminished. JoEllyn has had to
endure my extreme and seemingly around the clock “termito”—ecentricities, while I have
had the benefit of noticing that Michigan skies are more beautiful than I had originally
thought and that the air does indeed have a nice aroma ...... Michigan has become a more
pleasant place to live under her expert tutelage. The preparation of the thesis manuscript

has taken several stress filled months during which Jo and Chloe have had to endure the

viii

behavior that is unfortunately typical of everyone who has chosen to get such a degree:

thank you for your caring and understanding.

My graduate committee members (Prof. Bush, Prof. Hausinger, Prof. Oriel, and
Prof. Jackson) have provided much needed input and advice throughout my graduate
studies, for which I am grateful. I also appreciate all the time that Dr. Schmidt has taken

to “get me up to speed” in the world of phylogenetic analysis.

Completing graduate work on an insect subject fulfills a longheld goal of mine,
and it has been as exciting as I had ever expected it might be. From the moment I first
viewed the microbial world within the termite gut I knew I had found a “home”: I am
very appreciative to have been given the opportunity to work on such research in the
laboratory of Prof. John A. Breznak. My first rotation in his lab was typified by my ~-
cryptic, to some- rallying cry “By Christmas or Bust!”. I hope to finally make good on
that promise someday, as one way to show him my deep gratitude for the years of
mentoring that I have received since I was first his student in the 1990 summer MBL

Microbial Diversity course.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS .......................................................................................... v
TABLE OF CONTENTS .............................................................................................. x
LIST OF TABLES ......................................................................................................... xii
LIST OF FIGURES ....................................................................................................... xiii
CHAPTER 1. BACKGROUND AND INTRODUCTION ........................................... l
Termite Distribution and Phylogeny .......................................................................... 2
Social Behavior of Termites ...................................................................................... 3
Termite Digestion of Wood ....................................................................................... 5

CHAPTER 2. PHYSIOLOGICAL ECOLOGY OF Methanobrevibacter cuticularis SP.
NOV. AND Methanobrevibacter curvatus SP. NOV. ISOLATED FROM THE

HINDGUT OF THE TERMITE Reticulitermesflavipes ................................................. 1.1
ABSTRACT ............................................................................................................... 12
INTRODUCTION ..................................................................................................... 14
MATERIALS AND METHODS ............................................................................... 16
RESULTS .................................................................................................................. 27
DISCUSSION ............................................................................................................ 36
ACKNOWLEDGMENTS ......................................................................................... 46
REFERENCES .......................................................................................................... 47

CHAPTER 3 Methanobrevibacterfiliformis SP. NOV., AN UNUSUAL
FILAMENTOUS METHANOGEN ISOLATED FROM THE HINDGUT OF THE
TERMITE Reticulitermesflavipes ................................................................................... 73

ABSTRACT ............................................................................................................... 74

INTRODUCTION ..................................................................................................... 75

MATERIALS AND METHODS ............................................................................... 76
RESULTS .................................................................................................................. 78
DISCUSSION ............................................................................................................ 87
ACKNOWLEDGMENTS ......................................................................................... 90
CHAPTER 4. HYDROGEN METABOLISM BY TERMITES AND THEIR
ASSOCIATED MICROBES ........................................................................................... 91
ABSTRACT ............................................................................................................... 92
INTRODUCTION ..................................................................................................... 93
MATERIALS AND METHODS ............................................................................... 103
RESULTS .................................................................................................................. 109
DISCUSSION ............................................................................................................ 131
ACKNOWLEDGMENTS ......................................................................................... 142
REFERENCES ............................................................................................................... 143

xi

LIST OF TABLES

TABLE 1. PROPERTIES OF METHANOGENS ISOLATED FROM
RETICULITERMES FLA VIPES TERMITES ..................................................... 52

TABLE 2. DISTANCE MATRIX FROM THE COMPARISON OF 168 RRNA
SEQUENCES OF STRAIN RFM-l, RPM-2 AND OTHER SELECTED
ARCHAEA BELONGING TO THE FAMILY METHANOBACTERIACEAE...53

TABLE 3. DISTANCE MATRIX COMPARING THE 168 RRNA SEQUENCE OF
STRAIN RPM-3 WITH OTHER SELECTED MEMBERS OF THE FAMILY
MET HANOBAC TERIACEAE. ............................................................................ 82

TABLE 4. COMPARISON OF BASE PAIRED POSITIONS WITHIN THE DEDUCED
SSU RRNA SECONDARY STRUCTURE THAT ARE INVARIANT IN
EACH OF TWO METHANOGEN GENERA. .................................................. 86

TABLE 5. RELATIONSHIP BETWEEN THE REDOX POTENTIALS OF ELECTRON
ACCEPTORS, GIBBS FREE ENERGY CHANGES, H2 THRESHOLDS, AND
ENVIRONMENTAL H2 CONCENTRATIONS FOR VARIOUS ANAEROBIC
PROCESSES. ..................................................................................................... 95

TABLE 6. CHARACTERISTICS OF ACETOGENIC STRAINS ISOLATED FROM
CONGOLESE “HIGHER-TERMITES” ......................................................... 112

TABLE 7. HYDROGEN THRESHOLDS OF ACETOGENS AND METHANOGENS
ISOLATED FROM TERMITE GUTSA. .......................................................... 113

TABLE 8. H2 EMISSIONS AND COMPENSATION POINTS EXHIBITED BY 3
SPECIES OF TERMITES FED DIFFERENT SUBSTRATES ....................... 119

TABLE 9. EFFECT OF PROKARYOTIC INHIBITORS ON H2 COMPENSATION
POINT AND GAS EMISSIONS BY CELLULOSE-F ED R. FLA VIPES
TERMITES. ...................................................................................................... 127

xii

LIST OF FIGURES

FIGURE 1. MODEL FOR THE FERMENTATION OF WOOD IN THE TERMITE
RE TIC ULIT ERMES FLA VIPES. .......................................................................... 6

FIGURE 2. MORPHOLOGY OF STRAIN RFM-l BY F420 EPIFLUORESCENCE
MICROSCOPY. ................................................................................................. 55

FIGURE 3. MORPHOLOGY OF STRAIN RPM-2 BY F 420 EPIFLUORESCENCE
MICROSCOPY. ................................................................................................. 57

FIGURE 4. GROWTH OF, AND METHANOGENESIS BY, STRAIN RFM-l WITH
H2/C02- .............................................................................................................. 59

FIGURE 5. PHYLOGENETIC POSITION OF STRAIN RFM-l AND STRAIN RPM-2.61

FIGURE 6. IN SITU MORPHOLOGY OF RFM-l-TYPE CELLS AND RFM-2-TYPE
CELLS. ............................................................................................................... 64

FIGURE 7. RFM-l-TYPE CELLS ATTACHED TO FILAMENTOUS
PROKARYOTES. .............................................................................................. 66

FIGURE 8. TEM OF PUTATIVE METHANOGENS IN TRANSVERSE SECTIONS
OF THE HINDGUT. .......................................................................................... 68

FIGURE 9. F420 EPIFLUORESCENCE MICROGRAPH OF THE HINDGUT
EPITHELIUM OF R. FLA VIPES COLLECTED IN WOODS HOLE, MA. ..... 70

FIGURE 10. OXYGEN PROFILES IN MEDIUM INOCULATED WITH STRAIN
RFM-l AND UNINOCULATED MEDIUM. .................................................... 72

FIGURE 11. MORPHOLOGY OF STRAIN RPM-3 BY TEM AND BY PHASE
CONTRAST MICROSCOPY. ........................................................................... 80

FIGURE 12. PHYLOGENETIC POSITION OF STR. RPM-3 WITHIN THE
METHANOBACTERIACEAE. ............................................................................ 84

FIGURE 13. PHASE CONTRAST MICROGRAPHS OF 4 NEWLY ISOLATED
ACETOGENIC BACTERIA ............................................................................ 1 11

FIGURE 14. H2 CONSUMPTION BY A SPOROMUSA-METHANOBRE VIBAC T ER CO-
CULTURE. ....................................................................................................... l 16

xiii

FIGURE 15. H2 CONSUMPTION, EMISSION, AND THE COMPENSATION POINT
EXHIBITED BY FRESHLY COLLECTED R. FLA VIPES WORKER
TERMITES. ...................................................................................................... 121

xiv

CHAPTER 1.

BACKGROUND AND INTRODUCTION

Termite Distribution and Phylogeny

Termites belong to the order Isoptera, and thus are insects typified by their
possession of 2 sets of wings which are both of equal size, a trait that makes them
notoriously poor fliers (Krishna, 1969). Approximately 2000 species of termites exist,
and they are distributed amongst all the continents, save for Antarctica. Termites are
primarily confined to regions between the latitudes 45 N° and 45 8°, and are most
abundant (both in terms of sheer number of individuals and in overall species diversity)
in tropical rainforests and savannahs. In those regions, numbers of individual foraging
specimens may reach as high as 6000 - m2 (Edwards and Mill, 1986). Globally, it has

been estimated that there are between 10" to 10'7 termites.

Taxonomically, these insects have been subdivided into 7 families (the Masto-,
Kalo-, Hodo-, Rhino-, and Serri- termitidae; the Termitidae; and the Termopsidae)
(Kambhampati et al. , 1996). The number of species in each of these families ranges from
1 -i.e. Mastotermes darwiniensis, the sole species and genus belonging to the
Mastotermitidae, and which is only found in Australia-- to well over half the extant
species, which belong to the Temritidae, a family whose members are often referred to as

the “higher” termites.

Social Behavior of Termites

In addition to their great speciation and broad geographical distribution, termites
are well noted for their true sociality, a behavior shared with many bees, wasps and ants;
the latter three belong to the Hymenoptera, an insect order not closely related to the

Isoptera (Krishna, 1969).

Termites form colonies or nests wherein individuals can reach numbers that range
from the hundreds to the hundreds of thousands. These nests can be enormous in size «up
to several meters in height above the ground. Nests can also be subterranean or arboreal

(Edwards and Mill, 1986).

Labor within nests is divided amongst morphologically specialized members,
which are most commonly referred to as castes. Essentially, this morphological and
behavioral division of labor begins with the mass exodus (from a pre-existing nest) of
large numbers of winged reproductives, called alates. After landing and shedding their
wings, male and female alate meet and court each other. If this courtship is successful, the
queen and king pair for life, and go about selecting a site to build their nest. After doing
so they mate and become the king and queen of a nascent colony (Edwards and Mill,

1986).

The young termites that hatch fi'om eggs produced by the queen develop into
several different castes. One caste consists of so-called soldiers that are charged with
defending the colony; they tend to have a head-capsule that has become differentiated;

such differentiation can take diverse forms that depends on the species of termite.

Examples of such differentiation are the development of large defensive pincers or
nozzle-like extensions from which defensive secretions can be rapidly ejected onto

intruders entering the nest.

The royal pair, soldiers, young, and pre-alates in a mature colony depend on the
activities of the so called worker caste (which make up the vast majority of the colony in
terms of numbers). They are charged with repairing and increasing the size of the nest,

with the feeding the other castes, and with foraging for food.

It is in regards to the foraging for food that termites exhibit some of their most
fascinating and diverse behaviors. Every termite species tends to have a specialized diet
but typically involves the consumption of a refractory item such as wood, which will be
discussed below. Roughly one half of the “higher” (T ermitidae) termite genera are so-
called humus soil-feeders, the components of which that are of nutritional value remains
to be elucidated. Termite species are also known to specialize on a diet of grass, dung,
grain, the nest material of other termites, or even on the hide of carrion. Another
fascinating dietary behavior involves the culturing of fungal-gardens within the termite
nest. Often times the basidomycetes cultured are specific to termite-nests and fail to
flourish in the event that the nest fails; this is an overtly obvious example of a dietary,
mutualistic symbiosis: the termite feeds and cultivates the fungus, in return the termite

consumes a part of the fungus, its enzymes, and its partially digested foodstuff.

F ungal-termite relationships are not the only example of microbe-animal

interactions in these insects: all termite species examined house a dense microbial

community in their gut. Certain members of this community have been shown to engage
in a dietary mutualism with their host. The following sections primarily focus on such

interactions in wood-feeding termites.

Termite Digestion of Wood

Wood is a refractory material and often nitrogen poor. The mobilization of wood
is a limiting factor in its digestion by termites: it must be triturated prior to its ingestion,
and this is accomplished by the mouthparts and in the foregut gizzard of the termite. Even

so, wood-feeding termites are often voracious (Breznak, 1975).

Some of that carbon turned over by termites is considered to be of economic and
social importance to our own species, and estimates suggest that the cost of termite
control and subsequent repair is well over one billion dollars annually in the US. alone.
However, despite having earned such a bad reputation, only ca. 2% of all termite species

have been identified as being economic pests (Edwards and Mill, 1986).

That only a small percentage of termites are pests serves to illustrate an important
point: that what may be true for one species of termite may not be so in the case of
another; these insects are so diverse in there speciation and dietary behavior that it is
difficult to make any broad reaching statements about “termites”, regardless of the topic,

without being naive.

.mee 165on EB 28.02% 88m 333$

              

 

£35533 .
a 5.32.6 L
.8 8:58 , 32.28822.
3 noncomn< : I . vac;
@ma ,
2.803 2.5382 ecwwuwhmmm
«003.. E9: noeficm

8.5.8 3 v aoa:u..<
8.55. Al :0 0.59.5522

.umEEKEESEESEK 85:8 05 5 week we 53858.5“ 8: com 382 A 95$...—

All examined species of the 6 “lower” termite families are wood-feeders and contain an
abundance of protozoal flagellates in their hindgut (Honigberg, 1969; Yamin, 1979).
Research on such protozoa over the last 70 years has shown that most are anaerobic and
cellulolytic [(Odelson and Breznak, 1985) and references therein]. The only non-gaseous
fermentation product formed by these microbes during their digestion of wood particles is
acetate (Odelson and Breznak, 1985), which is subsequently absorbed by the termite and
used as its primary carbon and energy source (see Figure 1). Thus, the termite and its
resident protozoa engage in one of the most cited mutualistic symbioses: the termite
delivers small particles of wood to the hindgut wherein the protozoa phagocytose and
ferment the polysaccharide components of it to a product that provides the majority of the
carbon and energy needs of the termite. The fermentation of each glucosyl unit in

cellulose follows a stoichiometry such as this (Odelson and Breznak, 1985):

c6H,,o, + 214,0 -+ 2CH3COOH + 4H, + 200, [eq. 1]

In R. flavipes and in most wood-feeding “lower” termites, H2 and CO2 formed
during the cellulose fermentation is consumed by HZ/CO2 acetogenic bacteria in the
hindgut, producing a third acetate per glucosyl unit fermented by the protozoa; this is
another example of a nutritional mutualism between termite and microbe, and one that
can account for up to 1/3 of the insect’s carbon and energy needs (Breznak and Switzer,

1986). The acetogenic reaction is summarized as follows.

4H2 + 2co2 —> CH3COOH + 2 H20 [eq. 2]

Wood-feeding, “higher termites” (Termitidae) do not contain cellulolytic protozoa
in their hindgut (Honigberg, 1969); the likely sources of cellulases therein are now
believed to be of insect origin (Slaytor, 1993; Veivers et al., 1991). However, the
fermentation of the wood polysaccharides in higher termites also involves activities of
acetogenic bacteria and results in the production of acetate fi'om this Hz/CO2 metabolism

(Brauman et al., 1992; Breznak and Switzer, 1986).

Not all H2 and CO2 produced during the fermentation of wood polysaccharides is
used by the microbiota to form acetate, as some is converted to methane which is emitted
by the termite (Brauman et al., 1992; Breznak, 1975). The reaction for this is summarized

as follows:

4H2 + co2 —) CH, + 2 H20 [eq. 3]

The grand mean ratio of H2 that flows to acetate rather than to methane in wood-
feeding termites is 11 to l (Breznak, 1994). This trend towards acetogenesis makes a lot
of sense from the standpoint of insect nutrition, inasmuch as methane can not be used by
the termite, and therefore represents a carbon and energy loss to the termite. However, in
most other sulfate poor environments, methanogenesis is dominant, and this is also true
in soil feeding termites, wherein the mean ratio of H2 that flows to acetate as compared to
that which flows to methane is 1 to 5.6 (Breznak, 1994). This competition for H2 by

acetogens and methanogens will be discussed in detail in Chapter 4.

Before it was known that there was such large variability in methane emission by

termites, one researcher in the 1980’s concluded that 40% of the worldwide emissions of

this greenhouse gas originated from these insects (Zimmerman, 1982). This theory was
never in danger of becoming established: currently, < 4% of such emissions are

considered to be of termite origin (Fraser er al., 1986).

Nitrogen relationships. Wood is typically nitrogen poor, with a C:N ratio
typically ca. 400:1, and so the conservation of nitrogen is a critical factor in animals that
use utilize wood in their diet (Potrikus and Breznak, 1981). Breznak and Benemann
independently, but simultaneously, demonstrated the potential for the termite microbiota
to fix atmospheric N2 (Benemann, 1973; Breznak et al., 1973). Although this was
demonstrated indirectly, by using the acetylene reduction method for estimating
nitrogenase activity, their studies indicated that in some (but certainly not in all) termites
Nz-fixation was likely occurring at biologically relevant rates. This was later confirmed

by Bentley, who performed 15"5N2 labeling studies (Bentley, 1984).

Another mutualism affecting the N-economy of termites involves the microbial
recycling of nitrogenous waste products produced by the termite. Potrikus and Breznak
have demonstrated that uric acid (that has been voided by R. flavipes into its gut fluid) is
fermented by anaerobic bacteria to acetate and NH], which can then be reabsorbed by the
termite (Potrikus and Breznak, 1981). The termite itself does not contain a uricase, so the
bacteria aid the termite by helping retain, recycle, and concentrate nitrogen obtained from

the nitrogen poor diet.

10

Microbes of unknown function. From the first demonstration of the dense
microbial colonization of the termite hindgut in the late 19‘h century (Leidy, 1877), a
number of morphologically conspicuous prokaryotes have confounded every attempt at
cultivation. These spirochetes --spira1 or screw-shaped bacteria that possess an
internalized, periplasmic flagellum-- are amongst the most abundant prokaryotes in the
hindgut fluid of wood-feeding termites. On the basis of their sheer numbers alone, it is
attractive to postulate that they too may be involved in an activity that is mutually
beneficial to host and microbe, but until such bacteria can be cultivated, this is without
experimental support. Recently, it has been shown that many or all of these spirochetes
may be members of the genus T reponema, as judged by phylogenetic analyses performed
on their SSU rDNA genes obtained after PCR amplification from preparations obtained

from the guts of several termite species (Berchtold et al., 1994; Paster et a1. , 1996).

REFERENCES

Works cited in Chapters 1, 3 and 4 are referenced together in a Bibliography
provided at the end of the thesis. Chapter 2 maintains the citation numbers and

bibliography (provided within the chapter) used for its publication.

11

CHAPTER 2.
PHYSIOLOGICAL ECOLOGY OF Methanobrevibacter cuticularis SP. NOV. AND
Methanobrevibacter curvatus SP. NOV. ISOLATED F ROM THE HINDGUT OF THE

TERMITE Reticulitermesflavipes

This chapter was published in 1996 and appeared in Applied and Environmental

Microbiology 62(10):3620-3631 .

12

ABSTRACT

Two morphologically distinct, Hz/COZ-utilizing methanogens were isolated from
gut homogenates of the subterranean termite, Reticulitermesflavipes (Kollar)
(Rhinotermitidae). Strain RFM-l was a short straight rod (0.4 x 1.2 pm), whereas strain
RPM-2 was a slightly curved rod (0.34 x 1.6 pm) that possessed polar fibers. Their
morphology, Gram-positive staining reaction, resistance to cell lysis by chemical agents,
and narrow range of utilizable substrates were typical of species belonging to the
Methanobacteriaceae. Analysis of the nearly complete sequences of the small subunit
rRNA-encoding genes confirmed this affiliation and supported their recognition as new
species of Methanobrevibacter: M. cuticularis (RF M-1) and M. curvatus (RF M-2). The
per-cell rates of methanogenesis by strains RFM-l and RPM-2 in vitro, taken together
with their in situ population densities [ca. 106 cells-gut]; equiv. 109 cells-(ml gut fluid)-
l], could fully account for the rate of methane emission by the live termites. UV
epifluorescence- and electron microscopy confirmed that RFM-l- and RFM-2-type cells
were the dominant methanogens in R. flavr’pes collected in Michigan (but they were not
the only methanogens associated with this species) and that they colonized the peripheral,
microoxic region of the hindgut, i.e. residing on or near the hindgut epithelium and also
attached to filamentous prokaryotes associated with the gut wall. An examination of their
oxygen tolerance revealed that both strains possessed catalase-like activity. Moreover,

when dispersed in tubes of agar medium under H2/C02/02 (75/18.8/6.2, v/v), both strains

l3

grew to form a thin plate about 6 mm below the meniscus, just beneath the oxic-anoxic
interface. Such growth plates were capable of mediating a net consumption of 02 that
otherwise penetrated much deeper into uninoculated control tubes. Similar results were
obtained with an authentic strain of Methanobrevibacter arboriphilicus. This is the first
detailed description of an important and oft-cited, but poorly understood, component of

the termite gut microbiota.

14

INTRODUCTION

Termites emit methane, and they are one of the few terrestrial arthropods to do so
(1, 6, 24, 46). The methane emitted [$1.30 umol-(g fresh wt-h)"; (4)] arises from
methanogenic Archaea, which reside in the gut and appear to be one of the terminal "H2
sink" organisms of the hindgut fermentation, i.e., catalyzing the reaction 4 H2 + C02 —>
CH4 + 2 H20 (10, 44). Owing to this property and to their large biomass densities,
particularly in tropical habitats, termites have been cited as a potentially significant
source of atmospheric methane. However, their precise contribution to global methane
emissions has been hotly debated (4; and references therein). Among the uncertainties in
global estimates of methane emission by termites are knowledge of the exact number of
termites on earth and of the extent of intra- and interspecific variation in emission rates
among the 2000 or so known species, as well as of environmental factors which may

affect these rates.

Microbial HZ/COZ acetogenesis (4 H2 + 2 C02 -) CH3COOH + 2 H20) also
occurs in the hindgut of termites, and rates range from 0.01 to 5.96 umol acetate formed-
(g fresh wt-h)'l [4]. Considering the overall reaction for acetogenesis, it would appear that
acetogens are in competition with methanogens for the same reductant, i.e. H2. Curiously,
however, the extent to which H2 flows to COz-reducing methanogenesis versus
acetogenesis varies with the feeding guild to which termites belong. In soil-feeding and
fungus-cultivating termites, methanogenesis dominates acetogenesis as an H2 sink;

however, the reverse is true for wood- and grass-feeding termites (4). This in itself is

15

enigmatic, because in most anoxic habitats in which C02 reduction is the is the primary

electron sink reaction, methanogenesis almost always outprocesses acetogenesis (7).

We seek a better understanding of factors that affect competition for H2 between
termite gut methanogens and acetogens. A key step toward that goal is the isolation of the
relevant organisms for study under controlled conditions in the laboratory. Several
species of H2/C02 acetogens have already been isolated flom termite guts and
characterized, including Sporomusa termitida (11), Acetonema longum (29) and
Clostridium mayombei (28). However, aside flom the visualization of methanogens by
F420 epifluorescence microscopy of termite gut contents (39, 44) and brief descriptions of
the enrichment and isolation of such forms [unpublished result of D. A. Odelson & J. A.
Breznak (cited in ref. [6]); (5 8)], our understanding of the physiological ecology of
termite gut methanogens is virtually nonexistent (8). Accordingly, the present effort was

initiated.

Herein we describe the isolation, characteristics and in situ location of two new
methanogens flom Reticulitermesflavipes (Kollar), the common eastern subterranean
termite. In this wood-feeding termite, H2/C02 acetogenesis [0.93 umol acetate formed -(g
flesh wt-h)'1] typically outcompetes methanogenesis [0.10 umol-(g flesh wt-h)"] as the
primary electron sink of the hindgut fermentation (4). [A preliminary report of these

findings has been presented (37)].

16

MATERIALS AND METHODS

Termites. Workers (i.e. externally undifferentiated larvae beyond the 3rd instar)
of Reticulitermesflavipes (Kollar) (Rhinotermitidae) were used for all experiments.
Unless indicated otherwise, they were collected in a wooded area in Dansville, MI, USA
and used within 48 h, or after various periods of maintenance in the laboratory as

previously described (46).

Media and cultivation methods. Anoxic cultivation was routinely done by using
COZ/bicarbonate-buffered, dithiothreitol (DT'I')-reduced media under an Oz-flee
atmosphere containing 80% H2 and 20% C02 (11). Medium JM—l contained (g°1‘1):
NaCl, 1.0; KCl, 0.5; MgC12'6HzO, 0.4; CaClz-ZHZO, 0.1; NH4C1, 0.3; KH2P04, 0.2;
NaZSO4, 0.15; NaHCO3, 5.8; trace element solution SL10 and selenite-tungstate solution
(56); and 7—vitamin solution (5 7). These latter 4 components, and any further
supplements (where specifically noted), were added to the autoclaved medium flom
sterile stock solutions as described by Widdel and Pfennig (57). The pH was adjusted to
7.4, when necessary, with sterile 1M solutions of either HCl or NazCO3. Prior to
inoculation, DTT (1 mM final cone.) was added to the medium as a reductant. Medium
JM-2 was identical to JM-l , but also contained 0.05% w/v each of Casamino Acids
(Difco) and yeast extract. Medium JM-3 was identical to JM-2, but also contained bovine
rumen fluid (2%, v/v) prepared as described below. Medium JM-4 was identical to JM-l ,
but also contained Nutrient Broth (0.2 %; Difco) and clarified rumen fluid (40%, v/v).
For solid media, agar (Difco; water-washed before use) was incorporated at a final

concentration of 0.8%.

17

Cells were grown in: 16 mm tubes containing 4.5 ml liquid medium [dilution-to-
extinction series; see below]; 18 mm anaerobe tubes (no. 2048-18150; Bellco) containing
4.5 ml of liquid medium or 8 ml agar medium (for routine culture or agar dilution series,
respectively); or serum bottles containing 1/4 to 1/3 their total volume of liquid medium.
Some of the latter vessels were custom-fitted with sampling ports and lateral arms (for
spectrophotometric determination of culture turbidity) that were derived flom the mouth
and bottom portion, respectively, of anaerobe tubes (above). All culture vessels were
sealed with butyl rubber stoppers, and all incubations were at 30°C in the dark unless
indicated otherwise. Liquid cultures were held static until visible turbidity developed, at
which time they were placed horizontally (tubes) or vertically (bottles) on a reciprocal

shaker operating at 130 cycles-min".

The anti-bacterial drugs Rifamycin SV and Cephalothin (each 10 ug/ml, final
cone.) were included in media in two instances to help achieve or ensm'e culture purity
(45): during the first agar dilution series for isolation of pure cultures; and in cultures to
be harvested for enzyme assays. The methanogens were naturally resistant to these
antibiotics. In all instances, culture purity was routinely checked by phase contrast and

UV epifluorescence microscopy (below).

Enumeration and isolation of methanogens. Enumeration of methanogens was
done by using a dilution-to—extinction enrichment culture method. Termites were
degutted while held in an anoxic glove box (9) to yield the entire hindgut, along with a
nearly full length of attached midgut. Nine extracted guts were pooled in 4.5 ml of DTT-

reduced buffered salts solution (BSS; containing 10.8 mM KZHPO4, 6.9 mM KHZPO4,

18

21.5 mM KCl, 24.5 mMNaCl, and 1 mM DTT; pH 7.2) at a concentration of 2 guts-ml'l
and homogenized (10). Six independent replicates of gut homogenate were prepared, and
each was subject to a serial lO-fold dilution in tubes of medium JM-2, JM-3, or JM-4
such that the final tube contained 10'3 gut equivalents. Tubes were scored as positive for
the presence of methanogens when, after 8 weeks of incubation: a negative pressure had
developed in the tube; methane accumulated in the headspace gas; and F420

autofluorescent cells were observed by UV epifluorescence microscopy.

For differential enumeration of gut epithelium-attached versus nonattached (or
loosely attached) methanogens, a similar approach was used, but with the following
modifications. Extracted guts were placed individually on a small square of Parafilm M
(American National Can, Greenwich, CT, USA) and the hindgut region was sliced
longitudinally by using a razor blade. The sliced gut was then flooded with a 100 pl dr0p
of BSS, in which it was agitated while being held with fine-tipped forceps. The liquid
portion, now containing liberated gut contents, was taken up with a syringe and
transferred to a small test tube. The gut was rinsed in another drop of BSS, which was
subsequently pooled with the first, and the rinsed gut was transferred to a separate tube
containing 1.0 ml of BSS. This procedure was repeated with 8 more guts, combining all
like fractions. The tube containing the 9 pooled guts was then held for 30 sec on a vortex
mixer, after which the gut pieces were allowed to settle and the liquid phase was
transferred to the tube containing expressed gut contents. This step was repeated once
more after adding 1.0 ml of flesh BSS to the pooled guts. At this point the guts were

somewhat translucent, most of the contents having been washed flom them. They were

19

then transferred to a homogenizer tube with 4.5 ml BSS, homogenized, and used as an
inoculum for one serial lO-fold dilution series in medium JM-3 (as described above) to
estimate the number of " gut wall-attached" methanogens. The tube containing the pooled
gut contents was also made up to 4.5 ml, homogenized, and used for a separate dilution
series to estimate the number of "nonattached or loosely attached" methanogens. The

entire dissection, washing, and dilution procedure was repeated in triplicate.

From the highest dilution tubes containing methanogens, pure cultures were
isolated by preparing an agar dilution series in medium JM-2 for strain RPM-1 and
medium JM-3 for strain RPM-2. Purity of strains was assumed when, afier passage
through 3 successive agar dilution series: all colonies flom the last tube were uniform in
size, color, and morphology and exhibited F420 autofluorescence; all cells in the colony
(and liquid cultures established flom such colonies) were of similar morphology and
exhibited F420 autofluorescence; and liquid cultures exhibited no growth when inoculated
into tubes of medium JM-2 under a headspace of NZ/COZ (80/20, v/v) nor Fluid
Thioglycolate Broth USP (BBL Microbiology Systems, Cockeysville NH), USA) under

100% N2.

Growth of cells in, and measurement of, Oz gradients. Growth of cells in 02
gradients was examined by inoculating and mixing ca. 107 cells/m1 (final cone.) in DTT-
reduced, molten JM-4 agar medium in anaerobe tubes (11 ml per tube) held at 45°C under
a headspace of N2/C02 (80/20, v/v). Immediately after inoculation, tubes were allowed
to solidify in a vertical position and the composition of the headspace gas (ca. 14 ml) was

changed to H2/C02/02 (75: 18.8:6.2, v/v). When growth plates were apparent in the agar

20

(see Results), the Oz gradient existing in the agar was determined with microelectrodes
(12) immediately after removing the butyl rubber stopper. Such measurements were
completed within 2 min, well before any significant change occurred in the 02 gradient

preexisting throughout most of the agar.

Localization of methanogens in situ. Visual inspection of termite guts for the
presence and location of putative methanogens was done by using F 420 and F350 (i.e.
methanopterinepifluorescence microscopy (1 9). A Nikon Optiphot microscope was used
and was equipped with a mercury vapor UV light source and excitation and emission
filter sets analogous to those described by the authors. This same microscope was used

for phase contrast microscopy.

Inspection of the flee (i.e. unattached to the wall) gut contents was done by
removing guts as above and placing them in a drop of 100 pl BSS or medium JM-3 on a
glass microscope slide, whereupon the gut was punctured with the tip of forceps to

liberate the contents. The preparation was then covered with a coverslip for viewing.

To inspect the luminal (i.e. inner facing) surface of the hindgut epithelium for
attached methanogens, sliced pieces of hindgut, prepared in a manner similar to those
used for differential enumeration (above), were laid onto a microscope slide and teased
fully open with fine needles such that the luminal surface was facing up, i.e. toward the
objective lens of the light microscope. This manipulation was done with the aid of a
stereomicroscope and was facilitated by two things. First, the luminal surface was

recognizable, because it was covered with an array of regularly-spaced, circular

21

depressions or "cups" (9) that created a dotted pattern on it. Second, the luminal surface
was somewhat hydrophobic such that pieces of hindgut floating in a small pool of buffer
on the slide were usually presented in the "up" orientation. Once properly oriented on the

microscope slide, the hindgut pieces were covered with a coverslip for viewing.

For examining the radial distribution of methanogens in termite guts, flozen thin-
sections of hindguts were made perpendicular to their long axis. Extracted guts were first
placed onto the surface of a 1 cm3-block of agar (3%, w/v; in BSS), whereupon they were
then totally encased by pouring over them an additional layer of molten (47°C) agar,
which quickly solidified. Afier trimming the block containing the guts, it was affixed to a
cork stopper and flash flozen by immersion in supercooled (with liquid N2) isopentane.
Excess isopentane was washed flom the block by brief immersion in the liquid N2 itself
and, after properly orienting the block on a cryostat, 3-5 pm thick sections were then cut.
Appropriate sections were placed on a microscope slide, quickly covered with a coverslip
to minimize evaporation, and then viewed immediately by using phase contrast and UV

epifluorescence microscopy.

Electron Microscopy. Samples for transmission electron microscopy (TEM)
were fixed with glutaraldehyde, postfixed with 0304, and embedded in an ERL 4206-
Quetol 653-NSA resin mixture, as described by Spurr (53) and Kushida (34). Thin
sections were then made and post-stained with uranyl acetate and lead citrate, as
previously described (9). Non-sectioned whole-cell preparations for TEM were

negatively stained by using equal parts of flesh cells in BSS and a saturated (ca. 5% w/v)

22

uranyl acetate solution. Preparations were examined by using a Philips model CM10 or

EM3 00 electron microscope.

Nucleotide sequence analysis of SSU rDNA. Nearly complete nucleotide
sequences of the small subunit (SSU) rRNA of strains RFM-l and RPM-2 were inferred
flom their corresponding rDNA genes, which were amplified by the polymerase chain
reaction (PCR). To prepare cell extracts of strain RPM-1, dense cell suspensions (10 mg
dry mass equivalent per ml H20) were sonic disrupted (5 minutes, constant, at setting no.
6 of a Branson Sonifier Model 450; Branson Ultrasonics, Danbury CT, USA). Cell
extracts of strain RPM-2 were made by French pressure cell treatment (3 succesive
treatments, each at 100 MPa) of 10 mg dry mass equiv. in 10 ml TE (Tris HCl, 10 mM;
EDTA, 1 mM; pH 8.0). Nucleic acid present in this latter extract was precipitated with 20
ml ethanol (100%) and 3 ml NH4CH3COOH (3.0 M) then sedimented, air dried, and
resuspended in 1 ml H20 (42). Both RFM-l and RF M-2 preparations were then treated
with RNAse (10 ug/ml; 42) and used as template DNA for PCR under the following
conditions: 1 pl DNA template (ca. 50 ng) was added to 24 ul PCR reaction solution
(33). The Archaea-specific PCR primers ARCH 21BF [5'-TTC CGC TTG ATC C(C/T)G
CC(A/G) G-3'; modfied flom (16)] or ARCH69F [5'-TAA GCC ATG C(A/G)A GTC
GAA (C/T)G-3'; (this study)] were used with either of two "universal" reverse primers,
1391R [5'-GAC GGG CGG TGT GT(A/G) CA-3'; modified flom (35)] or 1492R [5'-
GGT TAC CTT GTT ACG ACT T-3'; modified flom (3 5)] both of which were gifts flom
Dr. T. M. Schmidt. Prior to use they were purified by denaturing polyacrylamide gel

electrophoresis (42) and then passed through a TSK-DEAE (Supelco, Bellafonte PA)

23

column with 1 M NH4CH3COOH as the eluent (T.M. Schmidt, unpublished protocol).
PCR amplification consisted of the following schedule: (1) 95 °C for 5 minutes; (2) 94°C
for 1 min; (3) 42°C for 1 min; (4) 72°C for 1 minute; (5) repeat steps 2-4 34 times; (6) 72
°C for 5 minutes. PCR products were purified by electrophoresing the desired DNA
bands out of an agarose gel essentially as described by Girvitz et. a1. (22) and then ligated
into the pCRWII cloning vector using the TA® cloning kit (#K2000-01; Invitrogen, San
Diego CA). Insert-containing plasmids were then obtained flom transformed E. coli (INV
orF') cells using a QIAGEN Midi Plasmid kit (#12145; Chatsworth, CA) and quantified
by using a DyNA Quant 200 fluorometer and Hoechst dye #33258, as described by the
manufacturer (Hoefer Pharmacia Inc., San Francisco CA). Nucleotide sequencing was
done flom the plasmid by the staff of the Nucleic Acid Sequencing Facility of Michigan
State University using an ABI Prism sequencer (Applied Biosystems). The following
sequencing primers were used: 519R [5'-G(A/T)A TIA CCG CGG C(G/T)G CTG-3';
(35)]; 533F [5'-GTG CCA GC(A/C) GCC GCG GTA A-3'; modified flom (35)]; 922F
[5'-GAA ACT TAA A(G/I')G AAT TG-3'; modified flom (3 5)]; 958R [5'-(C/T')C CGG
CGT TGA (A/C)T CCA ATT—3'; (16)]; and standard M13 forward and M13 reverse
primers. An edited, contiguous sequence was constructed flom the data obtained flom the
sequencing of both DNA strands, using Sequencher 3.0 software for Power Macintosh
(Gene Codes, Ann Arbor, MI). Phylogenetic analysis of the deduced 5' to 3' rRNA
sequence was initiated with its submission to the "Similarity_Rank" routine (41) at the
Ribosomal Database Project (RDP; University of Illinois, Urbana-Champaign, IL, USA).

This was done in order to build a list of known sequences that were most similar to the

24

ones submitted, so that the best intra- and inter-specific alignments could be made.
Sequences were then manually aligned while using the Genome Database Environment
version 2.2 (GDE; available flom the RDP) operating on a Sun SPARC station.
Similarity matrices were constructed with the Jukes and Cantor (27) correction for base
changes, using unambiguously aligned data. Phylogenetic trees were constructed flom
these same alignments using distance [DeSoete algorithm (17) with Jukes and Cantor
(27) correction for base changes], maximum parsimony, and maximum likelihood
methods [the latter two were bootstrapped using SEQBOOT]. These analyses were run
using PHYLIP version 3.55c (J. Felsenstein and the University of Washington, Seattle,

WA; public domain) as incorporated into the GDE program.

Enzyme Assays. Enzyme activities were measured in crude cell-flee extracts
prepared by French pressure cell treatment (3 times at 100 MPa). Catalase (EC 1.11.1.6)
was assayed by measuring the rate of decrease in absorbance of H202 at 240 nm (2).
Peroxidase (EC 1.11.1.7) was assayed by measuring the rate of increase in absorbance at
414 nm of the radical formed by reduction of 2,2'-azino-di-(3-ethyl-benzthiazoline-6-
sulphonic acid) [15]. NADH oxidase (EC 1.6.99.3) was determined by following the
decrease in absorbance at 340 nm owing to the oxidation of NADH (54). Superoxide
dismutase (EC 1.15.1.1) was assayed by using the xanthine/xanthine oxidase-cytochrome
c reduction method (43). Catalase, peroxidase, and superoxide dismutase preparations

(for calibrations and positive controls) were obtained flom Sigma.

Analytical methods. H2 and CH4 were analyzed by gas chromatography using

TCD or FID detectors, respectively (46) and organic acids were determined by HPLC

25

using RI detection (10). Turbidity measurements of cell suspensions or growing cultures
were made at 600 nm by using a Spectronic 20 colorimeter or a Gilson Model 2451-A
spectrophotometer. Protein was determined by the Bradford method, using an assay kit
(#500-0006) purchased flom Bio Rad (Richmond, CA, USA) with bovine serum albumin

as standard.

Other procedures and materials. Direct cell counts were made by using a
Petroff-Hausser counting chamber (32). Per-cell rates of methanogenesis were
determined by removing a 50 ml sample of known cell density flom a late log phase
culture to a 120 ml, butyl rubber stoppered serum vial under Hz/COZ (80/20, v/v) and
measuring the rate of methanogenesis over a short time interval (34 hours). Rumen fluid
was prepared by straining fleshly-collected rumen contents (obtained flom a fistulated,
forage fed dairy cow at the MSU Dairy Facility) through cheesecloth and then incubating
the fluid at 37°C for ca. 18 hours. It was then neutralized with NaOH, clarified by
centrifugation (3 cycles; 20,000 x g for 20 min each), dispensed into serum bottles under
N2, and autoclaved. Soluble hot water extracts of strain RFM-l were prepared by
suspending ca. 5 mg dry mass equivalent of fleshly harvested cells in 10 ml BSS,
autoclaving the suspension under N2, and then removing insoluble material by
centrifugation. Liver infusion consisted of the supernatant fluid recovered, by
centrifugation, after autoclaving a 40% aqueous suspension of dried liver (Difco) for 20
min. Termite extract was prepared by grinding 1 g (flesh wt) live termites in 10 ml BSS

with a mortar and pestle, autoclaving the mixture for 15 min under N2, and harvesting the

26

supernatant liquid after centrifugation. All other chemicals were of reagent grade and

were obtained from commercial sources.

Accession numbers of microbial strains and nucleotide sequences. Cultures of
M. cuticularis strain RFM-l (DSM 11139) and M. curvatus strain RPM-2 (DSM 1 ll 1 1)
have been deposited in the Deutsche Sarnmlung von Microorganismen, Gottingen, FRG.
Methanobrevibacter arboriphilicus strain DH] (ATCC #33747) was obtained flom the
American Type Culture Collection, Rockville, MD. The SSU rDNA sequences for strains
RFM-l and RPM-2 have been submitted to GenBank (accession numbers U41095 and
U62533, respectively). Other sequences used in the analysis were obtained flom the RDP
(M. ruminantium and M arboriphilicus, for which no publications or authors are
credited) or GenBank (M. stadtmanae, M59139 [51]; M. formicicum, M36508 [38]; M.
thermoautotrophicum, X15364 [48]; and the R. speratus clone, D64027 [47]). The rRNA
sequence for M. smithir' is unpublished and was obtained flom Dr. David Stahl (see

Acknowledgments).

27

RESULTS

Enumeration and isolation of methanogens. Enumeration of HZ/COZ-
consuming methanogens in R. flavipes by using medium JM-2 implied, after 8 weeks of
incubation, a population of ca. 106 viable cellsgut'1 (i.e. 3/3 tubes inoculated with 10:5
gut equivalents yielded methanogens, as did 2/3 of those inoculated with 10'6 gut
equivalents). No higher population densities were inferred by extending the incubation
time to 6 months. Considering that the methanogens are located in the hindgut region of
R. flavipes (see below), whose cell-flee fluid volume is only about 0.27 1.11 (46), this value
translates to an in situ population density of about 3.5 x 109 methanogen cells-(ml hindgut
fluid)". Similar population densities were inferred when gut homogenates were serially
diluted in medium JM-3 (1 independent gut homogenate) and JM-4 (2 independent gut
homogenates), whereupon 3/ 3 tubes inoculated with 10'5 gut equivalents yielded
methanogens, as did 2/3 of those inoculated with 10‘6 gut equivalents (one tube each of

JM-3 and M4).

Microscopic examination of dilution tubes scored positive for methanogenesis
usually revealed a uniform population of F420 fluorescent, nonmotile, short straight rods
approximately 0.4 x 0.9 to 1.8 pm in size, regardless of whether medium JM-2 or JM-4
was used. However, flom the one dilution series made with medium JM-3, the tube
inoculated with 10'6 homogenized gut equivalents contained F420 fluorescent rods that
were curved and morphologically distinct flom the short straight rods. From the highest
dilution tubes containing each methanogen morphotype, pure cultures were isolated by

using agar dilution series.

28

No HZ/COZ-consuming acetogenesis occurred in any of the dilution tubes, as
measured by acetate production, even when 5 mM bromoethanesulfonate (BES) was
included in the medium as a methanogenesis inhibitor and tubes were left to incubate for

up to 6 months.

General properties of strains RFM-l and REM-2. A methanogen isolate
representing the short, straight rod morphotype was designated strain RFM-l (Fig. 2), and
one representing the curved rod morphotype was designated strain RPM-2 (Fig. 3). TEM
of thin sections revealed that the cell wall of both strains lacked an outer membrane and
resembled that of gram-positive Bacteria (Fig. 2B; 33, C). TEM also revealed that,
although both strains had slightly tapered ends, RFM-2 was distinguished by its
possession of polar fibers, each measuring approximately 3 x 300 nm (Fig. 3B-D). No
such fibers were observed on cells of strain RFM-l (Fig. ZB). Additionally, cells of
RFM-2 occasionally remained together after cell division giving rise to loosely-coiled
helical chains up to 50 um long (not shown). This cell arrangement was common in

colonies in agar dilution tubes. Strain RFM-l rarely formed chains more than two cells in

length.

The general characteristics of strain RFM-l and RPM-2 (which are considered to
be two new species of Methanobrevibacter, see below) are summarized in Table 1. Both
strains were essentially limited to H2 + C02 as energy source. Strain RFM-l also used
formate, but poorly so, requiring several months to achieve visible turbidity in broth
cultures. Strain RFM-l was capable of chemoautotrophic growth, although the doubling

time under these conditions was rather long (>200 hours). The growth rate of strain RFM-

29

l was markedly stimulated by inclusion of 0.05% yeast extract, 0.05% Casamino Acids,
or 2% (v/v) bovine rumen fluid in media (doubling time <50 h). With all of these
supplements together (i.e. medium JM-3), cells grew at 37° C and pH 7.7 with a doubling
time of 35 h and, with periodic replenishment of H2 + C02, attained cell yields in excess

of 350 pg dry mass/ml (OD600nm 2 1.0).

Strain RPM-2 required complex supplements for growth. Among the best
supplements was bovine rumen fluid, which gave linear cell yield increases up to a
concentration of 40% (v/v), the highest concentration tested. Hence, 40% rumen fluid
was used in media for routine cultivation. Rumen fluid could not be replaced by
coenzyme M (mercaptoethanesulfonate), nor by a soluble, hot water extract of strain
RFM-l cells. Growth of RPM-2 in rumen fluid-containing media was further stimulated
by inclusion of 0.2% Nutrient Broth (Difco), but not by an equal amount of yeast extract
(Difco), Casamino Acids (Difco), Trypticase Soy Broth (Difco), liver infusion (Difco), or
termite extract (see Methods). With rumen fluid and Nutrient Broth as supplements (i.e.
medium JM-4), strain RPM-2 grew at 30°C and pH 7.2 with a doubling time of 40 h, but,
even with periodic replenishment of H2 + C02, attained cell yields no greater than 79 ug

dry mass-ml'l (ODWMm S 0.2).

A representative growth curve of strain RFM-l with H2 + C02 as energy source is
shown in Figure 4. Under such conditions, recovery of Hyderived electrons as CH4 was
consistent with the equation, 4 H2 + C02 -—> CH4 + 2 H20 (Table l; footnote (1), and
molar growth yields were 1.55 and 1.28 g dry mass per mol CH4 for strain RFM-l and

RPM-2, respectively (Table l). Per-cell rates of methanogenesis just prior to entering

30

stationary phase were about 2.85 x 10'‘0 umol CH4-(cell-h)'I for strain RFM-l and

7.55 x 10'10 umol CH4o(cell-h)'l for strain RF M-2. Given the total in situ population
density of RFM-l - and RFM-2-type methanogens estimated by the dilution-to-extinction
method (ca. 106 cells of each per gut; above), then both strains together could produce
about 1.04 x 10'3 mol CH4.(termite-h)'l [equiv. 0.30 umol CH4-(g flesh wt-h)'l for R.
flavr’pes worker larvae weighing 3.5 mg]. This would more than account for the actual
rate of CH, emission by live R. flavipes workers, which is typically 0.07 to 0.10 umol

CH4-(g flesh wt-h)'l [4, 5], and suggests that the methanogens may be Hz-limited in situ.

SSU rRNA sequence analysis. Nearly complete sequences of the SSU rDNAs of
both strains RFM-l and RPM-2 were obtained. The sequence for RFM-l corresponded to
E. coli SSU rRNA nucleotide positions 2 through 1408. Initially, we were unable to
obtain a PCR product flom strain RPM-2 by using the ARCH21BF primer. This may
have been due to mismatches occurring between the primer and the target sequence.
However, PCR amplification of the rDNA gene flom RPM-2 was obtained by using a
newly designed primer, ARCH69F. The productcorresponded to E. coli SSU rRNA
nucleotide positions 49 through 1511. Phylogenetic analysis of these sequences indicated
that both RFM-l and RPM-2 belong to the genus Methanobrevibacter. Strain RFM-l
shares 96.9%, 93.8%, and 93.2% sequence similarities with Methanobrevibacter
arboriphilicus, M. smithii and M. ruminantium, respectively. Strain RPM-2 has 95.3%,
93.3%, and 93.1% sequence similarities with these same species. Both isolates share
93.5% sequence similarity with a partial 16S rDNA clone flom the gut of the Japanese

termite, Reticulitermes speratus, and 95.4% similarity with each other (Table 2). The

31

topology of the unrooted phylogenetic trees constructed flom the data was identical using
both maximum likelihood (ML; Fig. 5) and maximum parsimony analyses (MP; tree not
shown). The distance method also grouped strains RFM-l and RPM-2 within the genus
Methanobrevibacter, however the topology of the tree depicting the phylogeny within the
genus differed slightly from that obtained using the other methods, in that strain RPM—2
clustered with strain RFM-l and M arboriphilicus, whereas M ruminantium formed a
separate, deep branch flom the other methanobrevibacters. The grouping of strains RFM-
1 and RF M-2 within the genus Methanobrevibacter was supported by bootstrap values of
96% (MP; not shown) and 99% (ML; Fig. 5) for the node flom which these strains -and
the other members of the genus -radiate, and the possession, by both strains, of a
signature sequence (5'-TGT GAG (A/C)AA TCG CG-3'; corresponding to E. coli
positions 37 5-3 88) shared only with other members of the genus. The SSU rDNA of
RFM—l had 4 dual-compensatory differences (8 nucleotide changes corresponding to E.
coli base-paired positions 451:493, 4532491, 607:646, and 835:851) flom its closest
relative, M. arborr'philr’cus, and the SSU rDNA gene of strain RFM-2 encoded for a
unique nucleotide bulge (5'-TTC TTA TGT T-3'; corresponding to a stem loop structure
at E. coli positions 200-218) not shared with either RF M-l or the other members of the
genus, which instead shared a different sequence (5'-Tn-3'; n = 6 or 8). This latter region
was not used in the phylogenetic analysis owing to our uncertainty about its correct
alignment, yet its presence is consistent with the results of the MP and ML analyses,

which place strain RPM-2 as a separate, deep branch within the genus (Fig. 5).

32

In situ morphology and localization of methanogens. We were surprised by the
relatively large number of methanogens in gut homogenates of R. flavipes (i.e. z 106 per
gut), because F420 fluorescent cells were actually quite scarce in contents expressed flom
punctured hindguts. Neither were methanogens seen as flee cells among the numerous
other prokaryotes in hindgut contents, nor were they associated in significant numbers
with flagellate protozoa, which are also abundant in the hindgut of R. flavr‘pes (9).
However, an abundance of F420 fluorescent cells was associated with the hindgut
epithelium: either attached to the cuticle surface directly or mixed among other
prokaryotes that were attached to it (Fig. 6A, B); or attached as epibionts to larger (up to
1.5 um dia.) filamentous prokaryotes, which themselves were associated with the hindgut

wall (Fig. 7A-D). Some of the latter filaments appeared to possess endospores (Fig. 7C).

The morphologies of such F420 fluorescent cells in situ were indistinguishable
flom those of RF M-l (Fig. 6A, 7B & D; compare to Fig. 2A) and RPM-2 (Fig. 6B;
compare to Fig. 3A) and the two morphologies were seen with comparable flequency.
TEM of thin sections of hindguts also revealed cells whose size and ultrastructure were
similar to that of RPM-1, and which were often seen surrounding thicker (presumably
filamentous) cells of diameter 0.8 um (Fig. 8A; compare to Fig. 2B). Such arrangements
may represent one of the epibiontic associations of methanogens seen by UV
epifluorescence microscopy (Fig. 7A-D). Likewise, TEM revealed RFM-2-like cells
(0.35 x 1.1 um) with polar fibers that, in situ, may facilitate their attachment to the
hindgut cuticle (Fig. 8B; compare to Fig. 3B & C). No F420 fluorescent methanogens

were ever observed in the midgut region of extracted guts.

33

Additional evidence that strains RPM-1 and RFM-2 were among the dominant
methanogens in guts of R. flavipes collected in Michigan and were attached onto or near
the hindgut epithelium in situ was as follows: (i) the density of F420 fluorescent cells on
randomly viewed portions of the hindgut wall (8-80 per 100 umz), when multiplied by
the total surface area of the hindgut cuticle (ca. 6 m2), was approximately equal to the
viable cell count of methanogens as determined by the dilution-to-extinction method, i.e.
106 per gut (above); (ii) flozen transverse sections of the hindgut revealed that F420
fluorescent RFM-l-like and RF M-2-like cells were located within the 10-20 pm zone
adjacent to the cuticular surface of the hindgut epithelium (not shown); and (iii)
differential enumeration indicated that over half of the cultivable methanogens were
tightly attached to the hindgut wall, at least to the extent that they were resistant to
detachment by vigorous vortex mixing. The combined recovery of methanogens inferred
flom such enumerations was in the same order of magnitude as that recovered when

homogenates of entire guts were used as inocula (above).

Although strains RFM-l and RPM-2 appeared to be the only methanogens present
in R. flavr‘pes collected in Dansville, MI, they are not the only methanogens associated
with this termite species. Specimens of R. flavipes recently collected in Woods Hole,
Massachusetts, USA had not only RFM-l and RFM-2-like methanogens associated with
their hindgut wall, but a thin filamentous form as well (Fig. 9). Cells of the latter
measured 0.3 pm in diameter and 2100 um in length. This filamentous methanogen
morphotype, which we have designated RPM-3, grew slowly and produced methane flom

H2 + C02 in all dilution tubes of JM-4 medium up to and including a 5 x 10'6 dilution of

34

gut homogenate. We have recently isolated a representative strain of the RF M-3

morphotype and have begun a characterization of it.

Tolerance to oxygen. The in situ location of strains RFM-l and RPM-2 (i.e. on
or within 10-20 pm of the cuticular surface of the hindgut epithelium) was entirely
unexpected, because the concentration of 02 near the cuticular surface may be as much as
25-50 M (corresponding to about 10% air saturation), diminishing to anoxia 100-200 u
m below the epithelial surface, in the central region of the hindgut (12). Consequently,

the tolerance of the isolates to oxygen was investigated.

When inoculated into anoxic, DTT-reduced, molten agar medium whose gas
phase was changed to H2/C02/02 (75/18.8/6.2, v/v) after solidification of the agar, cells
of RFM-l and RPM-2 grew as a plate approximately 200 pm thick and located about 6
mm below the meniscus of the medium. The position of such plates in the agar was just
beneath the point at which 02 could no longer be detected with microelectrodes (Oz <200
nM) and coincided with the zone of transition between the oxidized (pink) <—> reduced
(colorless) forms of resorufin, included in the medium as a visual redox indicator (Fig.
10). By contrast, penetration of 02 in uninoculated control tubes was about 3 times
deeper into the agar, as was the resorufin redox transition zone (Fig. 10. Note: readings
obtained flom the top 2 mm of each tube should be disregarded, as they are largely due to
the intrusion of 02 when the stoppers were removed flom tubes in order to perform the
measurements). These results indicated that RFM-l and RPM-2 cells within such plates
were somehow mediating a small net consumption of 02. However, neither strain grew in

such tubes unless a reducing agent (e.g. 1 mM DTT) was included in the medium. This

35

was true whether the headspace gas was HZ/C02/Oz (75/18.8/6.2, v/v) or Oz-flee Hz/COZ
(80/20, v/v). Virtually identical results were obtained with Methanobrevibacter

arboriphilicus strain DHl (ATCC 33747).

Strains RFM-l and RPM-2, as well as Methanobrevibacter arboriphilicus DHl,
also possessed catalase activity, as judged by the evolution of gas bubbles when a drop of
3% H202 was added to cells on a microscope slide (Table 1; footnote a). The specific
activity of catalase-like enzyme in crude cell-flee extracts of RFM-l was 54 umol H202
decomposed-(min-mg protein)“. However, such extracts did not exhibit NADH oxidase,
peroxidase, or superoxide dismutase activities. Exposure of RFM-l cells to oxygen for 18

hours had no measurable affect on any of the aforementioned enzyme activities.

36

DISCUSSION

Results presented herein show that HZ/COZ-utilizing methanogens are present in
hindguts of R. flavipes termites in relatively large numbers (ca. 106 viable cells-gut];
equiv. 109 viable cells-ml gut fluid") and are represented by strains RFM-l and RPM-2,
and by the morphotype RPM-3. The overall density of methanogens in hindguts of R.
flavipes is as great as that in many other methanogenic habitats, including the bovine
rumenand nonruminant large bowel (40, 45). Although most dilution tubes used to
enumerate methanogens flom Michigan-collected R. flavipes gave rise to RFM-l-type
cells, both RFM-l- and RFM-2-type cells appear to be present in roughly equal numbers
in situ. Cells of both strains were seen with comparable flequency by F420
epifluorescence microscopy and TEM of gut preparations, and development of RFM-2-
type cells in a (albeit single) dilution tube of JM-3 medium inoculated with 10'6 gut
equivalents suggested that its cell density was at least of the same order of magnitude as
that of RFM-l. RFM-l-type cells may simply tend to outgrow RFM-2-types in dilution
tubes, because the latter are more fastidious, and they have longer doubling times and
lower growth yields than RF M-l-type cells. It may also be that attachment of the RPM-2-
type cells to the hindgut cuticle, via their polar fibers (Fig. 8B), may be tighter than that
of RF M-l making it more difficult to liberate individual cells of the former by

homogenization of extracted guts.

F 420 epifluorescence microscopy revealed that methanogens in R. flavipes were
associated primarily with the hindgut wall -- either attached to it directly or existing

among other prokaryotes attached to it, or attached as epibionts to prokaryotic filaments,

37

which were also associated with the hindgut wall. To our knowledge, the attachment of
methanogens to filamentous prokaryotes as described herein has not been previously
documented. The basis for such attachment is still obscure, inasmuch as the filamentous
organism(s) have not yet been obtained in pure culture. However, this physical
association may reflect a syntrophic interaction between the cells based on interspecies
H2 transfer (14). Given their abundance, it was not difficult to find cells resembling
RFM-l and RFM-2 in thin sections of hindguts examined by TEM also (Fig. 8A and 8B,
respectively) and which were referred to as morphotypes l and R-2, respectively, in an
earlier study (see Fig. 3, 9, 10, 18 & 19 in ref. [9]). That previous study also described
short rods attached, as epibionts, to endospore-forming prokaryotic filaments. However
those epibionts, whose ultrastructure was that of gram-negative cells (see Fig. 14-17 in
ref. [9]), bear no resemblance to strain RFM-l (Fig. 2B, 7, 8A). Obviously, such

attachments between prokaryotes are not restricted to methanogens.

Methanogens are also known to occur on and within flee-living, as well as host-
associated, anaerobic protozoa (21, 24, and references therein). However, such
associations were not apparent or rare in hindguts of R. flavipes. This finding was similar
to the observations of Hackstein and Stumm (24) for Reticulitermes santonensis and
several other termites, but in contrast to those of Lee et al. (39) and Messer and Lee (44)
for certain hindgut protozoa flom Zootermopsis angusticollis termites. There may be
particular physiological and/or morphological properties of protozoa that affect their
ability to harbor methanogenic ecto- or endosymbionts, as even in Z angusticollis it

appeared that only some species of flagellates did so.

38

Association of methanogens with the hindgut wall of R. flavipes almost certainly
facilitates colonization and discourages washout. Hackstein and Stumm (24) described
analogous attachments of F420 fluorescent methanogens to bristle- and brush-like
structures that protrude into the hindgut lumen of Blattidae (cockroaches) and Cetoniidae
(rose chafers). However, R. flavipes possesses no such protrusions, and the peripheral
region of the hindgut near the wall appears to be microoxic (12), suggesting that some
mechanisms exist that enable strains RFM-l and RPM-2 to avoid toxicity of 02 and/or
by-products of 02 metabolism. One passive mechanism may be the consumption of 02 by
some members of the dense and diverse flora of nonmethanogens that also colonize the
hindgut epithelium, and which are known to constitute an "oxygen sink" (12, 13), a role
analogous to that played by facultative bacteria in conferring Oz-tolerance on
methanogens present in sludge granules (30). On the other hand, one adaptive mechanism
appears to be the possession by strains RFM-l and RPM-2 of a catalase-like activity,
which would help to detoxify any H202 that may inadvertently be formed by them, or by
their neighbors, flom a partial reduction of 02. In fact, the specific activity of catalase-
like enzyme in strain RFM-l is similar to that in Escherichia coli (50). To our

knowledge, this is the first report of a catalase-like activity in methanogens.

Another, more cryptic, adaptation may be represented by the growth plates of
RFM-l and RPM-2 cells that develop in agar tubes under a gas phase of H2 + C02 + 02
(Fig. 10). Presumably plate formation occurred several mm below the meniscus, because
H2 (i.e. the growth-limiting energy source) was in a headspace gas mixture that also

contained 6% 02. Hence, although cells were initially distributed throughout the tube, the

39

only ones capable of significant grth were those as close to the source of the H2
gradient as permitted by their tolerance to 02, which also diffused down flom the same
headspace. In similar tubes under HZ/COZ (80/20, v/v), cells grew primarily at the
meniscus, i.e. at the agar-gas phase interface (not shown). The ability of cells to initiate
growth in oxygen gradient tubes in the first place was probably facilitated by the fact that
H2, which was present in roughly ten-fold greater concentration than 02, but which is
sixteen times smaller in molecular mass, would diffuse more rapidly through the agar
than would 02, which was also being scavenged by the reducing agent present (DT'T).
However, it was also apparent that the growth plates effected a net consumption of 02
that otherwise penetrated much deeper in uninoculated control tubes. We do not yet know
what such "consumption" of oxygen means in biochemical terms, although it does not
necessarily mean that cells were performing aerobic respiration. One explanation may be
that anaerobic metabolism of the methanogens in the growth plate serves to re-reduce
some Oz-oxidized reducing agent or the redox dye resorufin, which then diffuses away
flom the cells to scavenge more oxygen. In this regard, we wish to reemphasize that cells
did not initiate growth in any medium unless a reducing agent was present, regardless of
whether the headspace gas contained 02 or was completely anoxic. Whatever its basis,
one wonders whether such Oz-consuming activity observed in vitro might have an in situ
correlate. We do not know the nature, concentration, or rate of turnover of natural
reducing agents that may exist in the hindgut of R. flavipes. However, Ritter (49)

obtained preliminary evidence that glutathione (or a similar compound) might be a

40

natural reducing agent in the gut fluid of the wood-eating cockroach, Cryptocercus

punctulatus.

Oxygen tolerance may also be important for the recolonization of guts by RFM-l
and RPM-2 following molting of R. flavipes, which includes the expulsion of the hindgut
contents. Reinoculation is then achieved, in part, by transfer of hindgut contents solicited
flom colony mates -- a process which exposes cells in the inoculum to air. However,
tolerance to oxygen is not restricted to methanogens flom termite guts. Virtually identical
results were obtained with Methanobrevibacter arboriphilicus strain DHl , a methanogen
originally isolated flom wetwood disease of trees (59). Although methanogens are
generally thought of as "strict anaerobes", their metabolic responses to the presence of
oxygen and their sensitivity to it vary with the species (20, 23, 25, 30, 31, 60; and
references therein), but this is an aspect of methanogen physiology that has not been
studied extensively. It seems likely that various mechanisms conferring oxygen-tolerance
may be present in methanogenic Archaea and are of adaptive significance for those

species conflonting periodic, or constant, exposure to Oz.

The fact that strains RFM-l and RPM-2 were virtually restricted to H2 + C02 as
energy source implies that they are in direct competition with HZ/COZ-acetogens for this
same resource in situ. This interpretation is also consistent with the ability of gut
homogenates of R. flavipes to form l4CH4 flom l4C02 + H2, but not flom 14C-UL-acetate
(10). However, it is still puzzling to us that methanogens are the only HZ/COz-utilizing
anaerobes we have ever been able to culture flom R flavr'pes, even though HZ/COZ-

acetogenesis outprocesses methanogenesis as the principal H2 sink of the hindgut

41

fermentation (10). Numerous, specific attempts to enrich or isolate HZ/COz-acetogens
from R. flavipes have met with frustration and failure, even if methanogenesis inhibitors
(e. g. BBS) were included in media (this study) or enrichments were tried by using
noncompetitive substrates, e. g. methoxylated aromatic compounds (unpublished results).
This has made us question whether methanogens such as RFM-l and RF M-2, which
perform a quantitative conversion of H2 + C02 to CH4 in vitro (Table l, footnote (1; Fig.
4), might actually be responsible for H2/C02 acetogenesis in situ. Many methanogens
possess key enzymes of the acetyl-CoA pathway for assimilation of C02 or for
aceticlastic methanogenesis (52; and references therein) and some strains of
Methanosarcina excrete acetate (albeit in small amounts) during growth on H2 + C02
(55). Conceivably, there may be some condition(s) in the gut of R. flavipes that
suppresses COz-reductive methanogenesis and provokes the synthesis and excretion of
acetate by such cells. Although we have no direct evidence to support this notion, the
possession of pure cultures of termite gut methanogens now enables us to explore this
bizarre possibility, as well as their seemingly inferior ability to compete with termite gut

acetogens for H2. Such studies have already been initiated in our laboratory

Taxonomy and description of new species. The following phenotypic properties
of strains RF M-1 and RPM-2 support their assignment to the genus Methanobrevibacter
within the family Methanobacteriaceae (3, 26): their rod shape; gram-positive staining
reaction and cell wall morphology, which was similar to that of gram-positive Bacteria
(Fig. 2B; 3B, C); their resistance to lysis when exposed to distilled H20, SDS, or NaOH;

their mesothermal temperature optima for growth (3 7° C and 30° C, respectively); and

42

their narrow spectrum of utilizable energy sources, which was essentially limited to H2 +

C02 (Table 1).

Based on the nucleotide sequence of the SSU rDNA of strain RF M-l (which
differs flom that of M arboriphilicus by >3%; Table 2), it is considered to be a new
species for which the name M cuticularis is herein proposed (see below). Likewise, the
curved morphology and polar fibers of strain RPM-2 (Figures 3B-D) are properties not
shared by any other species of Methanobrevibacter, although the latter were similar to
those fibers observed on certain strains of Methanobacterium (18, 36). Nevertheless, the
SSU rDNA sequence of strain RPM-2 indicates it should also be regarded as a new
species within the genus Methanobrevibacter (Table 2; Fig. 5), for which we pr0pose the

epithet M curvatus (see below).

It is interesting that an abstract by Yang et al. (58) reported the presence of
Methanobrevibacter arboriphilicus and Methanobacterium bryantii in guts of wood-
eating "higher" termites (Nasutitermes costaIis and N. nigriceps), a group of termites
(family Terrnitidae) considered to be phylogenetically remote flom "lower" termites such
as R. flavipes. More recently, a partial rDNA sequence of an uncultivated methanogen
was obtained after PCR amplification of DNA flom gut contents of Reticulitermes
speratus. This clone was affiliated with the Methanobacteriales and was stated to
represent a novel lineage within the order (47). However, the authors did not report any
comparisons of their clone with SSU rRNA sequences obtained flom
methanobrevibacters, to which their sequence seems related by our analysis (Table 2). In

any case, it is tempting to speculate that termite hindguts may be a rich reservoir of novel

43

methanogen diversity, as reflected by the two new species isolated in this study and

whose formal description follows.

Description of Methanobrevibacter cuticularis sp. nov.

Methanobrevibacter cuticularis sp. nov. [cu.tic’ul.ar’ is. L. cuticula, dim. skin; L. adj.
cuticularis, referring to the cuticular surface of the termite hindgut epithelium which is

colonized by this organism].

Straight short rods with slightly tapered ends, 0.4 x 1.2 pm in size, occurring
singly, in pairs, or in short chains. Non-motile. Gram positive-like by staining and cell

wall ultrastructure. No endospores formed.

Strict anaerobe. Catalase positive, oxidase negative. Metabolizes H2 + C02 and
forrnate (the latter very poorly) yielding CH4 as the sole product. Methanol, methanol +
H2, CO, acetate, ethanol, isopropanol, trimethylamine, dimethylamine, tlreobromine,

theophylline, trimethoxybenzoate, lactate, pyruvate, and glucose not metabolized.

pH optimum, 7.7 (range 6.5-8.5 ); temperature optimum 37° C (range 10°-37° C).
Chemolithotrophic growth occurs very slowly. Yeast extract, a source of amino acids
(e. g. Casamino acids), and ca. 2% clarified rmnen fluid are markedly stimulatory to

growth.

44
Source. Hindgut contents of the termite Reticulitermesflavipes (Kollar)

(Rhinotermitidae).

Type strain: RFM-l. Deposited in the Deutsche Samrnlung von

Mikroorganismen, Gottingen, FRG (DSM 11139).

45

Description of Methanobrevibacter curvatus sp. nov.

Methanobrevibacter curvatus sp. nov. [cur.va'tus. L. curva, bent; L. adj. curvatus,

referring to the curved shape of the cell].

Curved rods with slightly tapered ends, 0.34 x 1.6 pm in size, occurring singly or
in pairs. Non-motile. Gram positive-like by staining and cell wall ultrastructure. Cells

have polar fibers 3 x 300 nm in dimension. No endospores formed.

Strict anaerobe. Catalase positive, oxidase negative. Metabolizes H2 + C02
yielding CH4 as the sole product. Methanol, methanol + H2, CO, acetate, ethanol,
isopropanol, trimethylamine, dimethylamine, theobromine, theophylline,

trimethoxybenzoate, lactate, pyruvate, glucose and formate are not metabolized.

pH optimum, 7.1 -7.2 (range 6.5-8.5 ); temperature optimum 30° C (range 10°-
30° C). Complex nutritional supplements, e. g. 40% (v/v) clarified rumen fluid and

Nutrient Broth (Difco) required for growth.

Source. Hindgut contents of the termite Reticulitermesflavipes (Kollar)

(Rhinotermitidae).

Type strain: RPM-2. Deposited in the Deutsche Sarnmlung von

Mikroorganismen, Gottingen, F RG (DSM 11111).

46

ACKNOWLEDGMENTS

This research was funded by National Science Foundation grants IBN91-06636

(to J AB) and BIR91-20006 (to the Center for Microbial Ecology).

Some of the electron micrographs (Fig. 8A, B) were part of an unpublished
collection obtained during an earlier study (9). Thin sections and negative stains of pure
cultures for electron microscopy were prepared by the Electron Microscopy Laboratory of
the MSU Pesticide Research Center. Frozen thin sections were prepared at the
Histotechnology Laboratory of the MSU Department of Pathology. We are extremely
grateful to Drs. Tom Schmidt and Randall Hicks for help and advice on those aspects
dealing with molecular phylogeny; and to Dr. David Emerson for his gift of oxygen
microelectrodes and for his many helpful discussions. We thank Dr. David Stahl for
kindly providing the unpublished SSU rRNA sequence of M smithii, a result of work
supported by NSF grant DEB-9408243. We also thank Dr. Andreas Brune for helpful

discussions and criticisms.

47

REFERENCES

. Bayon, C., and P. Etiévant. 1980. Methanic fermentation in the digestive tract of a
xylophagous insect: Oryctes nasicornis L. larva (Coleoptera; Scarabaeidae).
Experientia 36: 154-1 55.

. Beers, R. F., and I. W. Sizer. 1952. Spectrophotometric method for measuring the
breakdown of hydrogen peroxide by catalase. J. Biol. Chem. 195: 133-140.

. Boone, D. R., and R. A. Mah. 1989. Family I. Methanobacteriaceae Barker 1956,
15,AL" emend. Balch and Wolfe in Balch, Fox, Magrum, Woese and Wolfe 1979,
267, p. 2174-2183. In J .T. Staley, M.P. Bryant, N. Pfennig, J .G. Holt (ed.), Bergey's
Manual of Systematic Bacteriology, Vol. 3. The Williams & Wilkins Co., Baltimore.

. Brauman, A., M. D. Kane, M. Labat, and J. A. Breznak. 1992. Genesis of acetate
and methane by gut bacteria of nutritionally diverse termites. Science 257: 1384-1387.

. Breznak, J. A. 1975. Symbiotic relationships between termites and their intestinal
microbiota, p. 559-580. In D. H. Jennings (ed.), Symbiosis [Soc. Exper. Biol. Symp.
Ser. no. 29]. Cambridge Univ. Press, Cambridge, UK.

. Breznak, J. A. 1984. Biochemical aspects of symbiosis between termites and their
intestinal microbiota, p. 173-203. In J. M. Anderson, A. D. M. Rayner and D. W. H.
Walton (ed.), Invertebrate-microbial interactions. Cambridge Univ. Press, Cambridge,
UK.

. Breznak, J. A. 1994. Acetogenesis flom carbon dioxide in termite guts, p. 303-330.
In H. L. Drake (ed.), Acetogenesis. Chapman and Hall, New York.

. Breznak, J. A., and A. Brune. 1994. Role of microorganisms in the digestion of
lignocellulose in termites. Annu. Rev. Entomol. 39:453-487.

. Brenak, J. A., and H. S. Pankratz. 1977. In situ morphology of the gut nricrobiota
of wood-eating termites [Reticulitermesflavipes (Kollar) and Coptotermes
formosanus Shiraki]. Appl. Environ. Microbiol. 33:406-426.

10. Breznak, J. A., and Switzer, J. M. 1986. Acetate synthesis flom H2 plus C02 by

termite gut microbes. Appl. Environ. Microbiol. 52:623—630.

11. Breznak, J. A., J. M. Switzer, and H.-J. Seitz. 1988. Sporomusa termitida sp. nov.,

an H2/COz-utilizing acetogen isolated flom termites. Arch. Microbiol. 150:282-288.

12. Brune, A., D. Emerson, and J. A. Bro-ak. 1995. The termite microflora as an

oxygen sink: nricroelectrode determination of oxygen and pH gradients in guts of
lower and higher termites. Appl. Environ. Microbiol. 61:2681-2687.

l3. Brune, A., E. Miambi, and J. A. Bro-ak. 1995. Roles of oxygen and the intestinal

microflora in the metabolism of ligrrin-derived phenylpropanoids and other
monoaromatic compounds by termites. Appl. Environ. Microbiol. 61:2688-2695.

48

14. Bryant, M. P., E. A. Wolin, M. J. Wolin, and R. S. Wolfe. 1967. Methanobacillus
omelianskii, a symbiotic association of two species of bacteria. Arch. Mikrobiol.
59:20-31.

15. Childs, R. E., and W. G. Bardsley. 1975. The steady-state kinetics of peroxidase
with 2,2'-azino-di-(3-ethyl-benzthiazoline-6-sulphonic acid) as chromogen. Biochem.
J. 145:93-103.

16. DeLong, E. F. 1992. Archaea in coastal marine environments. Proc. Nat. Acad. Sci.
89:5685-5689.

17. DeSoete, G. 1983. A least squares algorithm for fitting additive trees to proximity
data Psychometrika 48:621-626.

18. Doddema, H. J., J. W. M. Derksen, and G. D. Vogels. 1979. Fimbriae and flagella
of methanogenic bacteria FEMS Microbiol. Letters 5: 135-138.

19. Doddema, H. J., and G. D. Vogels. 1978. Improved identification of methanogenic
bacteria by fluorescence microscopy. Appl. Environ. Microbiol. 36:752-754.

20. Gerritse, J., and J. C. Gottschal. 1993. Two-membered mixed cultures of
methanogenic and aerobic bacteria in Oz-limited chemostats. J. Gen. Microbiol.
139:1853-1860.

21. Gijzen, H. J., C. A. M. Broers, M. Barughare, and C. K. Stumm. 1991.
Methanogenic bacteria as endosymbionts of the ciliate Nyctotherus ovalis in the
cockroach hindgut. Appl. Environ. Microbiol. 57:1630-1634.

22. Girvitz, C., S. Bacchetti, A. J. Rainbow and F. L. Graham. 1980. A rapid and
efficient procedure for the purification of DNA flom agarose gels. Anal. Biochem.
106: 492-6.

23. Gloss, L. M., and R. P. Hausinger. 1988. Methanogen factor 390 formation: species
distribution, reversibility and effects of non-oxidative cellular stresses. BioFactors
1:237-240.

24. Hackstein, J. H. P., and C. K. Stumm. 1994. Methane production in terrestrial
arthropods. Proc. Nat. Acad. Sci. 91:5441-5445.

25. Hausinger, R. P., W. H. Ormc-Johnson and C. Walsh. 1985. Factor 390
chromophores: phosphodiester between AMP or GMP and methanogen Factor 420.
Biochemistry 24: 1629-1633.

26. Holt, J. G., N. R. Krieg, P. H. A. Sneath, J. T. Staley, and S. T. Williams. 1994.
Group 31, the methanogens, p. 719-736. In Bergey's Manual of Determinative
Bacteriology, 9th ed. The Williams & Wilkins Co., Baltimore.

27. Jukes, T. H., and C. R. Cantor. 1969. Evolution of protein molecules, p. 21-132. In
H. N. Munro (ed.), Mammalian protein metabolism. Academic Press, New York.

49

28. Kane, M. D., A. Brauman, and J. A. Breznak. 1991. Clostridium mayombei sp.
nov., an H2/C02 acetogenic bacterium flom the gut of the Aflican soil-feeding
termite, Cubitermes speciosus. Arch. Microbiol. 156:99-104.

29. Kane, M. D., and J. A. Breznak. 1991. Acetonema Iongum gen. nov. sp. nov., an
H2/COz acetogenic bacterium flom the termite, Pterotermes occidentis. Arch.
Microbiol. 156:91-98.

30. Kato, M. T., J. A. Field, and G. Lettinga. 1993. High tolerance of methanogens in
granular sludge to oxygen. Biotech. Bioeng. 32: 1360-1366.

31. Kiener, A., and T. Leisinger. 1983. Oxygen sensitivity of methanogenic bacteria.
Syst. Appl. Microbiol. 4:305-312.

32. Koch, A. L. 1994. Growth measurement, p. 248-277. In P. Gerhardt, R. G. E.
Murray, W. A. Wood, and N. R. Krieg (ed.), Methods for General and Molecular
Bacteriology. American Society for Microbiology, Washington, DC.

33. Kogan, S. C., M. Doherty, and J. Gitschier. 1987. An improved method for
prenatal diagnosis of genetic diseases by analysis of amplified DNA sequences.
Application to hemophilia A. N. Engl. J. Med. 317 :985-990.

34. Kushida, H. 1974. New method for embedding with a low viscosity epoxy resin
"Quetol 651". J. Electron Microscopy 23: 197.

35. Lane, D. J. 1991. 16S/23S rRNA Sequencing, p.115-175. In E. Stackebrandt and M.
Goodfellow (ed.), Nucleic Acid Techniques in Bacterial Systematics. John Wiley &
Sons Ltd., New York.

36. Langenberg, K. F., M. P. Bryant, and R. S. Wolfe. 1968. Hydrogen-oxidizing
methane bacteria: 11. electron microscopy. J. Bacteriol. 95:1124-1129.

37. Leadbetter, J. R., and J. A. Breznak. 1995. Isolation, characterization, and in situ
localization of termite gut methanogens, p. 86-87. In J. P. van Dijken and W. A.
Scheffers (ed.), Book of Abstracts of the Beijerinck Centennial, Microbial Physiology
and Gene Regulation: Emerging Principles and Applications, 10-14 December 1995,
The Hague, The Netherlands. Delft University Press, Delft, The Netherlands.

38. Lechner, K., G. Wich, and A. Blick. 1985. The nucleotide sequence of the 16S
rRNA gene and flanking regions flom Methanobacteriumformicicum: the

phylogenetic relationship between methanogenic and halophilic archaebacteria. Syst.
Appl. Microbiol. 6: 157-163.

39. Lee, M. J., P. J. Schreurs, A. C. Messer, and S. H. Zinder. 1987. Association of
methanogenic bacteria with flagellated protozoa flom a termite hindgut. Curr.
Microbiol. 15:337-341.

40. Lovley, D. R., R. C. Greening, and J. G. Ferry. 1984. Rapidly growing rumen
methanogenic organism that synthesizes Coenzyme M and has a high affinity for
formate. Appl. Environ. Microbiol. 48:81-87.

50

41. Maidak, B. L., N. Larsen, M. J. McCaughey, R. Overbeek, G. J. Olsen, K. Fogel,
J. Blandy, and C. R. Woese. 1994. The Ribosomal Database Project. Nucleic Acids
Res. 22:3485-3487.

42. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a

laboratory manual, p. 184-185, 468. Cold Spring Harbor Press, Cold Spring Harbor,
NY.

43. McCord, J. M., and I. Fridovich. 1969. Superoxide dismutase. An enzyme function
for erythrocuprein (hemocuprein). J. Biol. Chem. 244:6049-6055.

44. Messer, A. C., and M. J. Lee. 1989. Effect of chemical treatments on methane
emission by the hindgut microbiota in the termite Zootermopsis angusticollis.
Microb. Ecol. 18:275-284.

45. Miller, T. L., and M. J. Wolin. 1986. Methanogens in human and animal intestinal
tracts. System. Appl. Microbiol. 7:223-229.

46. Odelson, D. A., and J. A. Breznak. 1983. Volatile fatty acid production by the

hindgut microbiota of xylophagous termites. Appl. Environ. Microbiol. 45: 1602-
1 6 1 3 .

47. Ohkuma, M., S. Noda, K. Horikoshi, and T. Kudo. 1995. Phylogeny of symbiotic

methanogens in the gut of the termite Reticulitermes speratus. FEMS Microbiol Lett
134:45-50.

48. Ostergaard, L., N. Larsen, H. Leffers, J. Kjems, and R. Garrett. 1987. A
ribosomal RNA operon and its flanking region florn the Archaebacterium
Methanobacterium thermoautotrophicum, Marburg strain: transcription signals,
structure and evolutionary implications. Syst. Appl. Microbiol. 9:199-209.

49. Ritter, H. 1961. Glutathione-controlled anaerobiosis in Cryptocercus, and its
detection by polarography. Biol. Bull. 121:330-346.

50. Rolfe, R. D., D. J. Hentges, B. J. Campbell, and J. T. Barrett. 1978. Factors

related to the oxygen tolerance of anaerobic bacteria. Appl. Environ. Microbiol.
36: 306-3 13 .

51. Rouviere, P., L. Mandelco, S. Winker, and C. R. Woese. 1992. A detailed
phylogeny for the Methanomicrobiales. Syst. Appl. Microbiol. 15:363-371.

52. Simpson, P. G., and W. B. Whitman. 1993. Anabolic pathways in methanogens, p.
445-472. In J. G. Ferry (ed.), Methanogenesis. Chapman and Hall, New York.

53. Spurr, A. R. 1969. A low viscosity epoxy resin embedding medium for electron
microscopy. J. Ultrastruct. Res. 37 :146

54. Stanton, T. B., and N. S. Jensen. 1993. Purification and characterization of NADH
oxidase flom Serpulina (T reponema) hyodysenteriae. J. Bacteriol. 175:2980-2987.

55. Westermann, P., B. K. Ahring, and R. A. Mah. 1989. Acetate production by
methanogenic bacteria. Appl. Environ. Microbiol. 55:2257-2261.

51

56. Widdel, F., G.-W. Kohring, and F. Mayer. 1983. Studies on dissimilatory sulfate-
reducing bacteria that decompose fatty acids 111. Characterization of the filamentous

gliding Desulfonema Iimicola gen. nov. sp. nov., and Desulfonema magnum sp. nov.
Arch. Microbiol. 134:286—294.

57. Widdel, F., and Pfennig, N. 1981. Studies on dissimilatory sulfate-reducing bacteria
that decompose fatty acids 1. Isolation of new sulfate-reducing bacteria enriched with
acetate from saline environments. Description of Desulfobacter postgatei gen. nov.,
sp. nov. Arch. Microbiol. 129:395—400.

58. Yang, J., F. M. Bordeaux, and P. H. Smith. 1985. Isolation of methanogenic
bacteria flom termites, abstr. I-83, p. 160. In Abstracts of the 85th General Meeting of
the American Society of Microbiology 1985. American Society for Microbiology,
Washington, DC.

59. Zeikus, J. G., and D. L. Henning. 1975. Methanobacterium arbophilr’cum sp. nov.
an obligate anaerobe isolated flom wetwood of living trees. Antonie van
Leeuwenhoek 41:543-552.

60. Zinder, S. H. 1993. Physiological ecology of methanogens, p. 128-206. In J. G. Ferry
(ed.), Methanogenesis. Chapman and Hall, New York.

52

.0... N + .5 T .8 + N: v a. 3288 .3: 5.88 .2255

$8.8. Ba .3: 8:52.. 2-2.55 $2.8. a... 532m 8.5.. £0 a .88....» 352.-.: .o 888: .c
83:5 .Néfim

mo 88 05 E .55 882w .8933 .8808 8853235888 6:53:85 6889585 658825855
68882505.: 8:559:88 8285.. .888: .00 am + 3:8.qu €2.58: .58: 6: :5 .5888 «8.535 0
2.3.8.3. .o .2 Ba 0 .8. a so.» .2. e... 3,5. 2... 72...: as: a

10:2

.2 m6 8 .mom 8: :o 6.: 58.585 8 05898 .3 58...: 8.3 05 803 :2. .852: 83 :55... .8562 .8588:
8888 88> 5:: SESSEVS: 2mm 5:: 24.» 5.5.55.0 55:53... .3 o>5§9EEw 28» 25.5 510m:r

 

 

555 :85:
52m 58.2.2 .555 :08... .8585 839A .m5_o< 058530 8:825“... 5380
.5 555...... be m 2.. .5 5:585 5 m 2.. moot: 8 2.... 5325
.09.... 583 385.. gout: .553... 5385
6 .23: u .2 6 .52.: u .3 .353 58...... 23.88....
53.8 2-. s .368 E 355 858...... m.
E: 53. x E: m - 88E 8:.—

w:o_ :5 cmA 2550 58:8. 5.5

:05: .81 5.. x end .58 53.50 E: N8 x 5.5 .58 28:8... 380592: :00

N82":— 58... 72.2 52%
2.53.5: a885§aee§5ms< 8.8383: 5.655.652.5585 Ea...—

 

.m8::..c.w 3855.85553832 :85 588:8 503.8502 mo 85305 A 85;.

53

85:58). 5:: 8:58:55 8: 80:868. :85 .0 88:8 :5 50:5

.858. :. :35 8: 809885 8:25.08 5853 .5
5:: 85.888: 5o:m..: 582.8588: 9... :o 5:85 83 5853 m 3:868. ..o. 2.88 855082.:
588.: 582.8583: mm ... .888: 80:88.55 .822. :5 :o 5:85 :83 82.8:5 :

 

 

 

- m... «an. 5.... I md. 56. m... 5... 88.3535... ukuupxmnotusug a
- - .... .... I 2: ...: .... 2: 58.8.5.8.53...»...385533 n
5385850385.»...
- - - 2: I a... ...: S. .... 53...».85523582 ..
- - - - I .5 Q5 ..5 ”.5 E:...=Q=..EE ..NBQQEMLQQEQEE 0
0:80
a .2 . ... 2. ...: - I I I I 8.88. 8.38:. 88.8.3885 :m
- - - - ...: - ...: N: :5 55......» 58:55.8..5555583 v
- - - - m5 - - S. 5... N525. 5:...m m
- - - - 5 w. - - - ..m 88.38.8555 ..8...:5..>:..5:=§.:V< N
- - - - ...: - - - - .-.).b. :.:...m .
a a ... 5 :m 5 m N .
328:5 5.80.5.3”... 885.30

 

.. .. 5:85.5855523583 5.85 :5 8 w:.w:o.o5

82.9.... 588.8 85: 5:: N555. ..-.25. 5:5: .0 838.8: <28. 8. .0 58.82.80 :5 :8... 58:... 8:889 .N 85: ...

54

Figure 2. Morphology of strain RF M-l by F420 epifluorescencc microscopy (A)

and by TEM of a thin section (B). Bars = 5 pm (A) and 0.4 pm (B).

55

Figure 2. Morphology of strain RFM-l by F420 epifluorescence microscopy.

 

56

Figure 3. Morphology of strain RFM-2 by F420 epifluorescence microscopy (A)
and by TEM (B, C, D). Polar fibers are visible on thin-sectioned cells (B, C), as well as
on negatively stained, unsectioned cells (D). Bars = 5 pm (A); 0.5 pm (B); 0.2 pm (C and

D).

57

Figure 3. Morphology of strain RPM-2 by F420 epifluorescence microscopy.

 

 

 

 

 

 

58

Figure 4. Growth of, and methanogenesis by, strain RFM-l with H2/C02 (80/20,
v/v; 101 kPa) as energy source. Cells were grown in medium JM-Z at 30°C with shaking,
and the decrease in headspace gas pressure was periodically compensated for by the
addition of N2/C02 (80/20, v/v). Initial CH4 and final H2 concentrations were not plotted,

as they were far below the lowest ordinate value.

59

Figure 4. Growth of, and methanogenesis by, strain RPM-l with Hz/COZ.

0.1 -

0 OD 600 nm

0.01

 

125
7100

T ‘,Y v

I
_L
O

[L-(amuno Iw) seloum] eueunew [:1 uafimpAH I

I V [—
—L

 

 

V I Y 1 ‘

50 100 150 200 250

Hours

 

h 0.1

60

Figure 5. Phylogenetic position of strain RFM-l and strain RPM-2 within the
Methanobacteriaceae, based on 1,183 unambiguous nucleotide positions in rDNA used in
a maximum likelihood analysis. The bar represents a 5% difference in evolutionary
distance as determined by measuring the lengths of the horizontal lines connecting the
species. Bootstrap values are placed to the immediate left of each node. M. stadtmanae

was used as the outgroup in the construction of the tree.

61

Figure 5. Phylogenetic position of strain RFM-l and strain RPM-2.

885:5 en

 

 

oncogene: SegmaezcfiuE
Eauwomgtexsxtmtuaegamfi

ansmoheuaaezt at Eatfiogocgte: ,
«.25— Eahm _

 

 

 

 

 

 

 

as?” sagassgéac: _:
SEEQSEE Luguufigsgafiu: on 3
T25. 58% a
uaufiaitofiu gououfizmgaoaafim: co

62

63

Figure 6. In situ morphology of RFM-l-type cells (A) and RFM-Z-type cells (B)
on the cuticular surface of the hindgut epithelium, as seen by F420 epifluorescence

microscopy. Bar = 5 um.

64

Figure 6. In situ morphology of RFM-l-type cells and RFM-2-type cells.

 

65

Figure 7. RFM-l-type cells attached to filamentous prokaryotes associated with
the hindgut wall. A and C are phase contrast micrographs; B and D are F420
epifluorescence micrographs of the same respective fields. Note the presence of

endospores in the filament shown in C (arrow). Marker bar = 5 pm.

66

Figure 7. RFM-l-type cells attached to filamentous prokaryotes.

 

67

Figure 8. TEM of putative methanogens in transverse sections of the hindgut of
R. flavipes. (A) RFM-l-type cells (arrow) attached to a thicker, central prokaryotic cell.
(B) RFM-Z-type cell (straight arrow) apparently attached to the cuticle of the hindgut

epithelium (e) by means of polar fibers (curved arrow). Bars = 1 pm.

68

Figure 8. TEM of putative methanogens in transverse sections of the hindgut.

 

69

Figure 9. F 420 epifluorescence micrograph of the hindgut epithelium of R.
flavipes collected in Woods Hole, MA. Note the presence of filamentous methanogens

(RF M-3—type) intermixed with the much shorter RFM-l- and RFM-Z—type methanogens.

Bar = 5 pm.

70

Figure 9. F420 epifluorescence micrograph of the hindgut epithelium of R.
flavipes collected in Woods Hole, MA.

 

Figure 10. Oxygen profile in JM-4 agar median inoculated with strain RF M-l
(Tube A; and inset A) and in an uninoculated eonu'ol (Tube B; and inset B) after 18 days
under a headspace of H2/C02/02 (75/18.8/6.2, v/v). Incubation was at 30°C. Dashed
horizontal line represents the position of the thin growth plate of RFM-l cells, which
develops about 6-7 mm below the agar surface in tube A (also shown in inset A, at arrow)
within the redox transition zone of resomfin. Arrow near tube B (inset) indicates the

redox transition zone of resorufin in the uninoculated control. Inset bar = 15 mm.

72

Figure 10. Oxygen profiles in medium inoculated with strain RF M-1 and

uninoculated medium.

 

uMolarOxygen
O 50 100 150 200 250
Imz.ml ..1.-iLiUi $__
‘TubeA on 39.9? 9 ' '

   
  

Tube 3 (no cells)

 

 

I Growth plate

Depth From Agar Surface (mm)
1
I

 

 

 

73

CHAPTER 3
Methanobrevibacterfiliformis SP. NOV., AN UNUSUAL FILAMENTOUS
METHANOGEN ISOLATED F ROM THE HINDGUT OF THE TERMI'I'E

Reticulitermesflavipes

74

ABSTRACT

A morphologically distinct, filamentous methanogen was isolated from hindguts
of the subterranean termite, Reticulitermesflavipes (Kollar) (Rhinotermitidae). Strain
RPM-3 measured 0.23-0.28 pm x up tolOO um, and aggregated into bundles and flocs
that were often > 0.1 mm in diameter. Its morphology, Gram-positive staining reaction,
resistance to cell lysis by chemical agents, narrow range of utilizable substrates (entirely
restricted to H2 + C02), and TEM cell-wall ultrastructure were typical of species
belonging to the Methanobacteriaceae. Analysis of the nearly complete sequence of the
small subunit rRNA-encoding gene of strain RPM-3 confirmed this affiliation and also
supported its assignment to the genus Methanobrevibacter as a new species for which the

epithet “filiformis” is proposed.

75

INTRODUCTION

We seek a better understanding of factors that affect competition for H2 between
termite gut methanogens and acetogens. A key step toward that goal was the isolation of
the relevant organisms for study under controlled conditions in the laboratory. Several
species of H2/C02 acetogens have already been isolated from termite guts and
characterized, including Sporomusa termitida (Breznak et al., 1988), Acetonema Iongum
(Kane and Breznak, 1991), and Clostridium mayombei (Kane et al., 1991). More recently,
2 novel methanogenic species (Methanobrevibacter cuticularis and M. curvatus) have
been isolated and described (Leadbetter and Breznak, 1996). During the latter study, an
unusual filamentous F420 fluorescent methanogen was observed in Woods Hole, MA
collected specimens of R. flavipes. This chapter describes the isolation and the

characterization of a strain representing this filamentous morphotype.

76

MATERIALS AND METHODS

Termites. Workers (i.e., externally undifferentiated larvae beyond the 3rd instar)
of Reticulitermesflavipes (Kollar) (Rhinotermitidae) were collected in a wooded
residential area in Woods Hole, Massachusetts (USA) and used within 48 h of their direct

collection from decaying logs.

Media and cultivation methods. Isolation of methanogens was done by dilution-
to-extinction enrichments, followed by three successive single colony picks from agar
dilution series (Leadbetter and Breznak, 1996). Cultivation media were COzlbicarbonate-
buffered and dithiothreitol (BTU-reduced under an Oz-fi'ee atmosphere containing 80%

H2 and 20% C02 (Leadbetter and Breznak, 1996).

Electron Microscopy. Samples for transmission electron microscopy (TEM)
were prepared as described previously (Breznak and Pankratz, 1977), and examined using

a Philips model CMlO electron microscope.

Nucleotide sequence analysis of SSU rDNA. An edited, contiguous sequence
was constructed from the data obtained after sequencing both DNA strands of the cloned
rDNA (cloned by L. D. Crosby, unpublished results) using Sequencher 3.0 software for
Power Macintosh (Gene Codes, Ann Arbor, MI). The sequence was manually aligned
with others sequences using the Genome Database Environment (GDE; version 2.2)

operating on a Sun SPARC station. Similarity matrices were constructed using only

77

unambiguously aligned data. Phylogenetic trees were constructed from these same
alignments using distance, maximum parsimony, and maximum likelihood methods.
These analyses were run using PHYLIP version 3.55c (J. Felsenstein and the University

of Washington, Seattle, WA; public domain) as incorporated into the GDE program.

Analytical methods. H2 and CH4 were analyzed by gas chromatography using
thermal conductivity or flame ionization detectors, respectively. Organic acids were
determined by a high performance liquid chromatograph equipped with an on-line UV212

m detector.

Accession numbers of microbial strains and nucleotide sequences. Cultures of
M. filiformis strain RPM-3 have been deposited (as DSM 11501) in the Deutsche
Sammlung von Microorganismen, Gottingen, Germany. The SSU rDNA sequence for
strain RPM-3 has been submitted to GenBank (U 82322). Other sequences [referenced in
(Leadbetter and Breznak, 1996)] used in the analysis were either obtained from the RDP
(M. ruminantium and M. arboriphilicus); GenBank (M. cuticularis, U41095 ; M
curvatus, U62533; M. stadtmanae, M59139; M. formicicum, M36508; M.
thermoautotrophicum, X15364; and the R speratus clone, D64027); or from David Stahl

(who provided the unpublished M. smithii sequence).

78

RESULTS

General properties of strain RFM-3. A filamentous methanogen observed in
hindguts of R. flavipes collected in Woods Hole, MA, was successfully isolated from a
dilution-to-extinction enrichment inoculated with 5 - 10'6 of gut equivalents from a gut
homogenate made from these same termites. The isolated strain formed filaments that
were up to 100 um long, and generally formed bundles and flocs when grown in liquid
culture (Figure l 1, panel c): these flocs were often macroscopic in size and were difficult
to disperse completely, even by vigorous agitation. The isolate was previously designated

as methanogen strain RPM—3 (Leadbetter and Breznak, 1996).

TEM of thin sections revealed that the cell wall of str. RPM-3 lacked an outer
membrane and resembled that of gram-positive Bacteria (see Figure 11, panel b). Cells
comprising the filaments were 0.23-0.28 um in diameter, and filament septation generally
occurred in 4 um increments (Figure 11a, arrows). TEM did not suggest that RPM-3 cells
possessed an outer sheath, nor was septation between such cells in the filaments

accompanied by plugs, plates, or sheaths as is sometimes observed in species of

Methanothrix and Methanospirillum (Holt et al. , 1994)]

Strain RPM-3 grew in media that was reduced with dithiothreitol (D'IT); however,

further growth was inhibited by the addition of cysteine (1 mM) or sulfide (1 mM) to

79

Figure 11. Morphology of strain RPM-3 by TEM (A, B) and by phase contrast
microscopy (C). Filament septation is visible on thin sectioned cells (A, arrows, and B).

Bars, 1 pm (A), 0.2 pm (B), 50 um (C).

80

Figure 11. Morphology of strain RPM-3 by TEM and by phase contrast

microscopy.

 

81

cultures pre-grown in DTT-reduced media. A small amount of yeast extract (0.01% w/v)
was required for growth; however, increasing the concentration of yeast extract 2.0% did
not further stimulate such growth. Acetate (1 mM), casamino acids (0.1%), or rumen
fluid (40%) alone could not replace the requirement for yeast extract, nor did the addition
of these components (to medium containing yeast extract) further stimulate growth. The

optimum pH for growth of strain RFM-l was 7.0-7.2 (range, 6.0-7.5).

SSU rRNA sequence analysis. The nearly complete sequence of the SSU rDNA of strain
RPM-3 (cloned by L. D. Crosby, unpublished results) corresponded to E. coli SSU rRNA
nucleotide positions 49 through 1511. Phylogenetic analysis of this sequence indicated
that strain RPM-3 belonged to the genus Methanobrevibacter. Strain RPM-3 shared
93.4% to 95.5% sequence similarities with the five other members of that genus (i.e.
Methanobrevibacter arboriphilicus, M smithii, M. ruminantium, M. cuticularis, and M.
curvatus) and also shared 94.9% sequence similarity with a partial l6S rDNA gene
cloned from the gut of the Japanese termite, Reticulitermes speratus (see Table 3). The
topology of the unrooted phylogenetic trees constructed from the data [using maximum
likelihood (ML; Figure 12), maximum parsimony, and distance methods (latter 2 trees
not shown)] always resulted in strain RPM-3 grouping within the genus
Methanobrevibacter. However, the topology of the tree depicting the phylogeny within
the genus differed slightly from method to method, and such variation was dependent on
which positions were considered as being ambiguous and excluded from the analyses.
Nevertheless, the grouping of strain RFM-3 within the genus Methanobrevibacter was

supported by l) bootstrap values of 99% (ML; Figure 12) for the node from which this

82

Table 3. Distance matrix comparing the 16S rRNA sequence of strain RPM-3 with other

selected members of the family Methanobacteriaceae“.

 

 

 

Organism Tvolutionary distance (%)"
I j 3 4 5 6 7 8 9 10—

] StrafiRFM—Ti 5f
2 M. curvatus str. RPM-2 4.5 5. I
3 M. arborr'philicus 4.5 4.3 4.8
4 M. cuticularis str. RFM-l 4.9 4.2 3.1 5.5
5 R. speratus (termite) clone 8.2 7. 6 8. 6 8.2 10. 8
6 M. ruminantium 5.8 6.4 5.7 6.4
7 M. smithii 6.6 6.2 5.2 6.1 5.7
8 M. formicicum 8.8 8.2 8.3 9.2 8.4 9.5
9 M. thermoautotrophicum 9.3 8.0 8.1 8.5 9.8 9.2 7.8
10 M. stadtmanae 10.3 9.9 10.8 11.2 10.2 11.7 9.5 11.8

 

a For the sources of these sequences please see Materials and Methods.

b Distances are based on the percent differences among 1164 unambiguously aligned
nucleotides, except for sequence 5, which is based on 843 unambiguously aligned
nucleotides and for which percent differences are given in italics.

83

Figure 12. Phylogenetic position of strain RPM-3 within the Methanobacteriaceae,
based on 1,164 unambiguous characters in rDNA used in a maximum-likelihood analysis.
Bootstrap values are placed to the immediate left of each node. The bar represents a 5%
difference in evolutionary distance as determined by measuring the lengths of the

horizontal lines connecting the species.

84

Figure 12. Phylogenetic position of str. RPM-3 within the Methanobacteriaceae.

338% o\on

 

wetness»? Eogmuezcfimg

 

Eaaoafiafifiatmaogeagamg

 

 

5838838585 EarmaoeaoeefimE _ o2

méémd .5 “3.9.36 aoaocfiaoaaosgfim:
_ mm
n-2,:— 525 ll— _

 

 

 

 

.zSE» amauefiamanezgae:

 

£53553 Luaoenagacozefio: he

 

wig .bm MIGNaUtaQ kmuuUngkQOKQu‘um: 5N

macaiktefe amaoefiamSQUSo: mm

85

strain "and the other members of the genus-- radiated; 2) by the possession of a signature
sequence (5'-TGT GAG (A/C)AA TCG CG-3'; corresponding to E. coli positions 375-
388) shared only with other members of the genus; 3) by its nucleotide composition at
specific base-paired positions (see Table 4; compare with the genus Methanobacterium);
and 4) by a nucleotide bulge (5'-Tn-3'; n = 6 or 8; corresponding to a stem loop structure
at E. coli positions 200-218) it shared with all other members of the genus except M.
curvatus (which instead possessed the sequence; 5'-'ITC 'ITA TGT T-3'). In support of its
overall sequence dissimilarity from the other members of the genus, the SSU rDNA of
RPM-3 had at least 3 dual-compensatory differences (i.e., 6 nucleotide changes
corresponding to E. coli base-paired positions 155:167, 248:276, and 680:710) when

compared to the other members of the genus Methanobrevibacter.

86

Table 4. Comparison of base paired positions within the deduced SSU rRNA

secondary structure that are invariant in each of two methanogen genera.

 

Nucleotide Methanobrevibacter Methanobacterium

 

position“ and strain RF M-3

378:385 G-C A-U
369:384 A-U - C-G
410:432 U-A C-G
599:650 U-A C-G
613:627 G-C U-A
658:748 U-A c-G
682:708 A-U G-C

 

‘ Positions correspond to E. coli l6s rRNA numbering.

87

DISCUSSION

The guts of Massachusetts-collected specimens of R. flavipes contained a
filamentous F420 fluorescent cell morphotype that did not appear to be present in
Michigan-collected termites, and a representative of this morphotype was isolated and
designated as strain RPM-3. The reason(s) are unclear for this apparent variation in gut
methanogen composition between R. flavipes collected in the two locations, but may
reflect their geographical separation, or “subtle” differences in their dietary habits, or
merely that they were collected from different nests. Whatever the case may be, it would
be interesting in the future to investigate in detail such intra-species variation: R. flavipes
has a wide-spread distribution (ranging from Kansas to the Atlantic; from Michigan and
Maine to Tennessee and the Carolina’s; and recent infestations of this species have also
occurred in Europe); this species has also been shown to occasionally divert from their
typically wood diet: in the Great Plains R flavipes have been noted to feed on grass and
the dung of herbivores (W eesner, 1969). The effect of such factors on the microbial
composition of termites is not well studied. Towards that end, targeted archaeal
rRNA/rDNA primers [such as those used during this, and an earlier study; (Leadbetter
and Breznak, 1996)] could be used to amplify methanobrevibacter genes from gut DNA
isolated from termites collected from disparate regions or exhibiting different feeding
behaviors. Determining any further species diversity within the methanobrevibacters

found in R. flavipes would also aid in predicting the potential diversity of

88

Methanobrevibacter species within all termite species: it is conceivable that the order
Isoptera contains literally thousands of species of Methanobrevibacter. Whatever the case
may be, the methodology and the the initial database of sequences needed to do such

studies is now available.

Taxonomy of strain RFM-3. The following phenotypic properties of strain
RF M-3 support its assignment to the genus Methanobrevibacter within the family
Methanobacteriaceae (Boone and Mah, 1989; Holt et al. , 1994): its Gram-positive
staining reaction and cell wall morphology, which was similar to that of Gram-positive
Bacteria (Figure 11, panels a and b); its resistance to lysis when exposed to distilled H20,
SDS, or NaOH (L.D. Crosby, unpublished results); and its narrow spectrum of utilizable
energy sources, which was essentially limited to H2 + C02 (L.D. Crosby, unpublished

results).

Based on the analysis of the nucleotide sequence of its SSU rDNA (cloned by L.
D. Crosby), strain RPM-3 is considered to be a new species belonging to the genus
Methanobrevibacter and for which the name M. filiformis is herein proposed (see below).
This conclusion is supported by the filamentous morphology (other methanobrevibacters
are typically less than 2 pm in length) and flocculent nature of strain RPM-3, properties
which are not shared by other species belonging to the genus Methanobrevibacter.
However, before the following taxonomy can be formally proposed and validly

published, the optimum growth temperature for this strain will have to be determined.

89

Description of Methanobrevibacterflliformis sp. nov.

Methanobrevibacterfiliformr's sp. nov. [fil.i.'form.is. L. n. filum -i, thread; L. v. formo
-are, shaped; L. adj. filiformis, referring to the filamentous morphology of the cells,

which draws contrast to that inferred by the generic epithet].

Filament forming rods with slightly tapered ends, 0.23-28 pm in width, by up to several
hundred 100 um. Filament septation typically occurs every 4 pm, rarely occurring singly.
Non-motile. Gram positive-like by staining and cell wall ultrastructure. No endospores

formed.

Strict anaerobe. Catalase-positive by H202 spot test. Metabolizes H2 + C02 and yielding
CH4 as the sole product. Methanol, methanol + H2, CO, acetate, ethanol, isopropanol,
trimethylamine, dimethylamine, theobromine, theophylline, trimethoxybenzoate, lactate,

pyruvate, and glucose not metabolized.

pH optimum, 7.0-7.2 (range 6.0-7.5); Yeast extract required for growth. Growth inhibited

by use of cysteine or sulfide but not by dithiothreitol as a reducing agent.

Source. Hindgut contents of the termite Reticulitermesflavipes (Kollar)

(Rhinotermitidae) collected in Woods Hole, Massachusetts, USA.

Type strain: RPM-3. Deposited in the Deutsche Sammlung von Mikroorganismen,

Gottingen, Germany (DSM 11501).

90

ACKNOWLEDGMENTS

This research was funded by the National Science Foundation.

Electron microscopy was performed by H. S. Pankratz, to whom we are very grateful. We
also thank Dr. David Stahl for kindly providing the unpublished SSU rRNA sequence of

M. smithii, a result of work supported by NSF grant DEB-9408243.

REFERENCES

Works cited in Chapters 1, 3 and 4 are referenced together in a Bibliography
provided at the end of the thesis. Chapter 2 maintains the citation numbers and

bibliography (provided within the chapter) used for its publication.

91

CHAPTER 4.

HYDROGEN METABOLISM BY TERMITES AND THEIR ASSOCIATED

MICROBES

92

ABSTRACT

Acetogenesis dominates (but usually does not entirely replace) methanogenesis as
an H2 sink in the guts of most wood-feeding termites, but the basis for this unexpected
outcome is not yet clear. H2 thresholds determined for 7 strains of termite gut acetogens
ranged from 251 to 380 ppmv (parts per million volume) Hz; for three termite gut
methanogens they ranged from 36 to 45 ppmv H2. Three different wood feeding termite
species both emitted and consumed H2 and exhibited a mean H2 compensation point (c.p.)
of 815 ppmv. Experiments (that included the study of protozoan-defaunated termites, as
well as those fed prokaryotic inhibitors) suggested that the H2 c.p. observed in R. flavipes
was dependent on the activity of anaerobic hydrogenotrophic (acetogenic) Bacteria, but
not on the activity of methanogenic Archaea or cellulolytic (protozoan) Eukarya. These
results imply that factors other than uncharacteristically efficient H2 scavenging by

acetogens are more relevant to the processing of H2 by termite gut hydrogenotrophs.

93

INTRODUCTION

Termites are one of the few terrestrial arthropods to emit methane, which arises
from methanogenic Archaea that reside in the hindgut. Methanogens constitute one of the
terminal "H2 sink" organisms of the hindgut fermentation (Brauman et al., 1992; Breznak,

1975; Leadbetter and Breznak, 1996)

Microbial acetogenesis from CO2 also occurs in the hindgut of termites, and the
acetogens appear to be in direct competition with methanogens for the same reductant,
i.e. H2 (Brauman et al., 1992; Breznak, 1994; Breznak and Switzer, 1986). Curiously, the
extent to which H2 flows to COz-reducing methanogenesis versus acetogenesis varies
with the feeding guild to which termites belong. In soil-feeding and fungus-cultivating
termites, methanogenesis dominates acetogenesis (Brauman et al., 1992). The reverse is
true for wood- and grass-feeding termites. This dominant role of acetogenesis is
enigmatic: in most anoxic habitats in which C02 reduction is the primary electron sink

reaction, methanogenesis almost always outprocesses acetogenesis.

That COz-reducing methanogens usually outcompete acetogens in most anoxic,
sulfate-poor (i.e., non-marine) habitats appears to be related to the lower H2 thresholds
typically displayed by the former, where “H2 threshold” refers to the lowest
concentration of H2 that can still be used in consumptive processes by cells. The
importance of such H2 thresholds in predicting the flow of H2 in anoxic habitats is
embodied in a more general hypothesis that is referred to as the “competition concept”,

which states that when two or more hydrogenotrophs compete for Hz, the organism

94

performing the most exergonic reaction should be able to maintain a steady state H2
concentration that is below the H2 threshold of its competitors, thus excluding their
potential activity. Hence, in a homogeneous system with limiting H2, most of the H2
would be processed and consumed by the hydrogenotroph performing the energy-yielding
reaction with the greatest free energy change (i.e., the more negative AG' value).
Inasmuch as the magnitude of the AG°' is directly proportional to the AE°' [i.e., the
difference in redox potential between the electron-acceptor and the electron-donor (in this
case H2)], by AG°' = -nFAE°' [wherein: n = no. of electrons transferred; and F = Faraday’s
constant, 96.48 kJN)], this hypothesis predicts that the Hz-consuming organism using the
electron acceptor having the most positive E°' should possess the lowest H2 threshold and
process most of the H2 by that overall reaction (see Table 5). This prediction is supported
by the H2 thresholds exhibited by pure cultures of hydrogenotrophs and by the H2
concentrations in many natural and man-made environments, both of which are lower for
the electron accepting processes that yield more fiee energy and have the more positive
E°' (Table 5). In this regard, it is easy to see why CO2 reduction to methane by
methanogens (COleH4; E°' = -0.244 V) is usually dominant over CO2 reduction to

acetate (COzlacetate; E°' = -0.290 V).

The relationship between H2 threshold concentrations and E°' becomes more
apparent when the former is used to derive a AG', i.e., the free energy change calculated
by using non-standard concentrations of reactants and products. For this discussion, the
only non-standard value being used in the free energy calculations is the H2 threshold

concentration itself. Using such concentrations, the calculated free energy

95

.55c .355 RomwoBB 8:55 2: £05558— N. T we .04 a S 3:82 55 a: no 5555280 05. a

.52 .588 Ba 3:3 ӣ2
. .3 5 5835520an 5655 c0553 395389 558 88m 5E5.” m_ 038 $5 we 555.550 2.: 5
5m: 85 nous—H585 3mm 5 v5 20:55; demwan—Eoo B.“ 05: 59:2: mm 0325 3 55:5.— ofim—sm a

 

 

0.5 + -m:

__ >23 55 . Ea. e; 253 t._ 52. E. t .m + ...0m + .5.
"ammoaoweussm

0.3 +

. Ea 3-3 . Ema 872 Sam 26 52- as- .5 t .m + .65 + .5.
"anonymous—.52

0.5 +

........ _. 22% 838 255 m: 3.2- as. 390% T .62: + .5
fiE5w83<

3.5355950 22?.th a: ...—canoaah a: cu: 98v 555..
a: 5mm. 5 =25< .5325...—. 5:95:55 am 93358555.»:

 

.amommoooa 05885 33,—? SM 53555050 am 55:80:50

v5 £20525 f .5w55 .355 02% 3me £8338 5555 .«o massaged x88 05 5035p manners—om .m «San.

96

88805.54—

nm- 8 m _ - mo .O< 0 8 80:5 855835 9558-8055 8 55800 85585050 055. ..

8880582 $1 8 mm. mo .O< a 8 80:5 05585050 055. m

.5558?— 2- 8 m T .«o .O< 0 8 80:5 8585850 08055508 5 58580.0 85585050 05.5. a

s.

8880085— 31 8 2. mo .O< a 8 8055 05585050 055. 9
8055.50— 3. 8 mm- mo .O< 0 8 80:5 05585050 0005. 0

AN: 328%
532: 00 585509 053 523V 85 333.8 05 88% >85 8 585250 503 0.6: 05: 5:88 0023/ o

.05.— 28303 05 80.: 5.3350 £80508 m 020,—.

97

(Table 5, footnotes d, e, and g) tends to be fairly close to the minimum free energy or
minimum quantum required to translocate 1 mol of protons or other ions through a fully
energized membrane, ca. 17-20 M - mol'l product formed (Conrad, 1996; Seitz etal.,
1990). At H2 concentrations below such thresholds, the potential free energy made
available by carrying out any given reaction would decrease (i.e., be increasingly positive
or even endergonic) and, in biological terms, would effectively be endergonic. Cells
would have no means to harness or couple energy released from the reaction; they might
even have to input energy to perform it. As can be seen in Table 5, there is a correlation
between the theoretical H2 threshold [i.e., the H2 concentration that would yield a AG' = -
17 kJ - mol'I reaction] and H2 thresholds of pure cultures and environmental samples. At
an H2 mixing ratio of 25 ppmv, a typical methanogenic threshold, acetogenesis could
only yield a meager 136' of -5 1d - mol“l reaction; at the even lower H2 concentrations
typically found in methanogenic environments (see Table 5), the AG' for acetogenesis
would be endergonic, requiring the expenditure of 12 to 19 kJ - mol'I reaction. Owing to
these energetic limitations, it not surprising that acetogenic activity is only marginal in
many such environments. In summary, the “competition” and “threshold” concepts
appear have sound theoretical and experimental foundations. These concepts function
well in explaining the preferential use of certain electron acceptors over others in a

variety of environments.

However, acetogens outprocess methanogens in certain habitats, such as in the

hindguts of most wood feeding termites. Several hypotheses can be to explain the reasons

98

for high acetogenic activities in such termites, and these will be briefly outlined in the

following sections.

Hypothesis #1: Mixotrophy may make acetogens more competitive for H2 in
situ (Breznak and Switzer Blum, 1991). Known acetogens are not restricted to H2 + CO2
as an energy source; they are typically capable of fermenting a fairly wide range of
saccharides, organic acids, certain amino acids, alcohols, and the O-methyl and N-methyl
constituents of a variety of parent compounds. In comparison, methanogens tend to be
restricted to H2, acetate, selected alcohols, and a few CI-compounds. The growth rate of
acetogens on organic substrates is often greater than when growing on H2. For example,
the termite gut acetogen Acetonema Iongum has a doubling time of 8 hrs when growing
on glucose, but 36 hours when growing on H2 + CO2 (Kane and Breznak, 1991). Such
increased growth rates could aid in the maintenance of acetogen population levels,
especially if washout from an environment is a possibility, as it may be in a gut
environment. Of particular relevance was the demonstration that the termite gut acetogen
Sporomusa termitida was always primed for hydrogenotrophy, regardless of its growth
substrate. Moreover, this species was able to grow mixotrophically, that is, concurrently
on both H2 + CO2 and a fennentable organic substrate (Breznak and Switzer Blum, 1991).
Hypothetically, mixotrophy might lower the H2 threshold compared to cells growing on
H2 + CO2 alone. For example, the metabolic summation of the autotrophic reaction,

4H2 + 2HC03' + l-I+ -> CH3COO' + 4 H20 [AG°' = -104.4 kJ - mol'l reaction], with
the fermentation of methanol, 1.33CH3OH + 0.67HCO3' —) CH3COO' + 0.33H+ + 1.33H2

[AG°' = -73.3 k] - mol‘l reaction], would result in the following stoichiometry: CH3OH +

99

H2 + HCO,’ —> CH3COO' + 2H20 [AG°' = -81.1 kl - mol" reaction]. If an acetogen
normally exhibited a H2 threshold of 950 ppmv (equating to a AG' of -36 kJ - mol'l
reaction; see Table 5, footnote d), the theoretical mixotrophic H2 threshold would
decrease dramatically, to O. 12 ppmv H2 if the same AG' of -36 kJ - mol'l reaction were
maintained. Therefore, when any of a diverse pool of organic substrate(s) (such as
methanol) is abundant, an acetogen might competitively exclude not only methanogens,
but sulfate reducing bacteria as well (see Table 5; compare in situ H2 concentrations with
0.12 ppmv H2). The hypothesis that mixotrophy may affect the H2 threshold of acetogens
is attractive. It fits well, albeit with a new twist, into the competition concept outlined

earlier.

Hypothesis #2: Methanogenesis itself may be inhibited in termites. It has just
been suggested that acetogens may be superior competitors for H2 under certain
environmental conditions. However, if methanogenesis itself was inhibited in the guts of
termites, then acetogens might only be filling an open niche. The relationship between
termites and acetogenic microbes may be viewed as mutualistic. It is clearly of nutritional
benefit to the termite, as up to 1/3 of the carbon and energy needs of the insect can be met
(Breznak and Switzer, 1986). Thus, it is not implausible that termites might have evolved
a way to promote acetogenic activity. This could be achieved if the insect were to
produce antimicrobial compounds directed against methanogenic Archaea. While such a
notion has no experimental support in this case, it is not completely without precedent in
insects per se. Insects are well known to produce a wide variety of antimicrobial peptides

such as the cecropin family of compounds. If such was proven to be basis for the apparent

100

acetogenic dominance in the termite hindgut fermentation, then the in situ concentrations
of H2 would fall within a range that is typical to that of acetogenic H2 thresholds, ca. 10-
fold higher than that of typical methanogens. Thus, reaching a better understanding of the
processing of H2 in termite guts depends not only on the study of pure cultures, but also

on studies of intact termites and their resident, hydrogenotrophic microbiota.

Hypothesis #3: Microscale stratification within the gut may influence
processing of H2. Recent studies have shown that both axial (lengthwise) and radial
gradients of pH and O2 exist within the termite hindgut (Brune et al., 1995; Brune and
Kuhl, 1996). As a result of these findings, our concept of the intestinal microecology of
termite guts has been evolving rapidly. It is no longer considered to be “a simple,
homogenous, anoxic fermentor, but one that is heterogeneous, structured, and
characterized by steep gradients of metabolites” (Andreas Brune, personal
communication). This structure is reflected in the spatial distribution of at least some of
the microbes in the gut. As was demonstrated in Chapter 2, methanogens R flavipes are
localized in the peripheral, microoxic regions of the hindgut. Additionally, Yoshirnura
and co-workers have demonstrated that protozoa in Coptotermesformosanus are
distributed axially in the hindgut: the largest species occupying the most anterior region
and the smallest species occupying the most posterior region (Yoshimura et al., 1992).
Interestingly, they discovered that the small protozoan Pseudotrichonympha grassi,
located in the posterior-hindgut, contained F420 fluorescent cells. Such putative
methanogen cells were not observed elsewhere in the gut of Coptotermes (Tsunoda et aI. ,

1993; Tsunoda et a1. , 1993). This heterogeneous, axial distribution of methanogens

101

contrasts to their radial distribution in R. flavipes. In the dampwood termite Zootermopsis
angusticollis, Kidder demonstrated that the protozoan Streblomastix strix was unevenly
distributed within the hindgut, being attached by holdfasts to the epithelium of the most

anterior regions of the hindgut paunch (Kidder, 1929).

As we continue to learn more about the microscale architecture of the hindgut, we
will undoubtedly become aware of additional unique or unusual facets of this symbiosis.
Related to the issues of interest in this thesis, it would likely be very informative to
elucidate the location of the acetogenic community in termites. At this time, nothing is
known about the location of acetogens in situ, but it seems reasonable to assume that they
might preferentially occupy the central (anoxic) regions of the R. flavipes hindgut. By
doing so, they would be closer to their H2 source (the cellulolytic protozoa) than the
methanogens, which are attached to the gut wall, and would better protected from
inwardly diffusing 02. Acetogens located in, on, or amongst the protozoa should exhibit
more favorable reaction rates for acetogenesis when compared to methanogenesis. In a
study by Goodwin et al. (1991), it was demonstrated that the diffusion of H2 significantly
affected the kinetic parameters governing its consumption. By moving the H2 emitting
point source extremely close to hydrogenotrophic cells, the apparenth for H2 uptake for
decreased by an order of magnitude. Since the Kn,s for H2 uptake by acetogens and
methanogens are similar (Breznak et al., 1988), the physiological group that is closer to

the H2 point source will have a distinct kinetic advantage.

The competitive advantages that accrue fi'om physical juxtapositioning have also

been demonstrated for hydrogenotrophs that are located within flocs in anaerobic reactors

102

(Conrad et aI. , 1995). The rate of H2 production and its conversion to methane within
flocs was significantly greater than that of H2 emitted from the flocs, thus limiting those

hydrogenotrophs in the surrounding medium.

Physical juxtapositioning between acetogens and their H2 source not only has
kinetic consequences, it has thermodynamic ones as well. Dolfing showed that the
distance of consumers from their H2 point source can effect their potential free energy
yields: the more distant an organism from their source of H2 along a steady state gradient,
the smaller would be the H2 pool at that location, and so the potential free energy from
can'ying out any given hydrogenotrophic reaction would decrease. In this regard, spatial
distribution would have a direct relationship to the “competition concept” introduced
earlier, and it would be important to elucidate factors that influence the distribution of
physiological groups along such gradients. For instance, it is not clear why methanogens

do not occupy the central regions of the R flavipes hindgut

In an effort to determine whether any of the aforementioned factors might affect
the competition for H2 between termite gut methanogens and acetogens, studies were
done on H2 metabolism by both pure and defined co—cultures of acetogens and

methanogens, as well on live termites and their smrounding nest material.

103

MATERIALS AND METHODS

Termites. Species of so-called "higher termites" (family Termitidae) were
collected in the Mayombe rain forest, Republic of Congo, and included: T rinervitermes
trinervoides (S j ostedt) (N asutitermitinae); Microcerotermes parvus (Haviland)
(Amitermitinae); and Nasutitermes arborum (Smeathman) (N asutitermitinae). Most of
these were used within 48 h of collection. "Lower termites" used in this study were
Reticulitermesflavipes (Kollar) (Rhinotermitidae), collected near Dansville, lvfl, USA;
Zootermopsis angusticollis (Hagen) (T ermopsidae), collected in the Golden Gate Park,
San Francisco, CA and kindly provided by Janet Shellman-Reeve; and Coptotermes

formosanus (Shiraki) (Rhinotermitidae), collected in Ft. Lauderdale, FL, kindly provided
by N.~Y. Su. These latter three species were used within 48 h of collection or receipt, or
after various periods of maintenance in the laboratory. All “higher” and “lower” termite
species used in this study belong to the “wood-feeding” guild, except for T. trinervoides,
which is a “grass-feeding” species. Workers (i.e., externally undifferentiated larvae

beyond the 3rd instar) were used for all experiments.

Availability of microbial strains used in this study. All strains used were
isolated previously to or during this study, are available fi'om the culture collection at the
Deutsche Sammlung von Microorganismen und Zellkulturen GmbH, Braunschweig,
Germany. Their accession numbers are: Methanobrevibacter cuticularr's str. RPM-1(DSM
11139; (Leadbetter and Breznak, 1996)); Methanobrevibacter curvatus str. RPM-2 (DSM
1 l 11 1; (Leadbetter and Breznak, 1996)); Methanobrevibacterfiliformis str. RPM-3

(DSM 11501); Sporomusa termitida str. J SN-2 (DSM 4440; (Breznak et al., 1988));

104

Clostridium mayombei str. SF C-S (DSM 6539; (Kane et al., 1991)); and Acetonema
Iongum str. APO-1 (DSM 6540; (Kane and Breznak, 1991)). Acetogenic bacteria isolated
and used during this study, but whose taxonomic position was not determined, were:
strain ATR-l (DSM 11379); strain NAG-2 (DSM 11380); strain GMP-2 (DSM 11381);

and strain GMP-3 (DSM 11382).

Microbiological media and routine cultivation of strains. Media JM-3 and JM-
4 (Leadbetter and Breznak, 1996) were used for the cultivation of methanogens and were
amended with final cone. of 10 mM 3-N-[morpholino]propanesulfonic acid (MOPS) pH
7.8 or pH 7.2, respectively. Acetogenic strains were isolated by using medium JM-2
(Leadbetter and Breznak, 1996) containing 5 mM syringate (3,5-dimethoxy 4-hydroxy
benzoate) and incubated under N2/CO2 80/20. Afier isolation, they were routinely
cultivated under an Hz/CO2 (80/20 v/v) headspace in JM-2 medium without syringate. All
media were reduced with dithiothreitol (DTT; 1 mM, final cone). Enrichments for
homoacetogenic bacteria were initiated by using 2, 1, 1/10, 1/20, 1/ 100, and 1/200
homogenized gut equivalents from each of the 3 “higher termites” species (see above).
Supernatant fluids from enrichment cultures were monitored, by reverse phase HPLC
(RP-HPLC), for the presence or absence of syringate «which was the substrate for
selective enrichment-- or its demethylation products. Isolation of strains from enrichment
cultures that demethylated syringate was achieved by picking single colonies fiom each
of three sequential passages through agar (1% v/v) dilution series of medium JM-2

containing syringate and incubated under an H2/CO2 (80/20 v/v) headspace. Since all

105

initial colony picks were dominated by cells containing putative endospores, a

pasteurization step (15 minutes at 80° C) was included prior to subsequent dilution series.

Determination of H2 Thresholds. H2 thresholds were determined during growth
of cells in medium JM—2 (for the acetogen strains and methanogen str. RPM-3), JM-3 (for
Methanobrevibacter cuticularis), or JM-4 (for M. curvatus). Triplicate 10 ml cultures of
each strain were incubated under initial 1% H2 headspaces (balance N2/COZ; 80/20 v/v) in
70 ml capacity serum stoppered glass vials. H2 consumption was followed until a distinct

plateau (i.e. the H2 threshold) was reached (Conrad and Wetter, 1990).

During and afier mixotrophic growth of S. termitida J SN-2, H2 concentrations
were monitored in a similar manner except that: 1) the headspace was initially adjusted to
contain 250 ppmv H2, instead of 10,000 ppmv H2; and 2) the culture medium was
amended with one the following substrates (in mM, final conc.): mannitol (5); methanol
(15); lactate (10); succinate (25); or glycine (5). Substrate disappearance and/or product

formation was confirmed afier performing HPLC analysis on culture supernatants.

Co-cultures of M. cuticularis RFM-l and S. termitida J SN-2 were initiated by
inoculating each strain into 120 ml stoppered serum bottles containing 10 ml of JM-3
medium and incubated under H2/CO2 (80/20 v/v). After incubation for 8 days, the
headspace was then replaced with N2/CO2 (80/20 v/v). The density of RFM-l and J SN-2
cells were determined to be 0.4 - 108 and 1.2 - 108 cells per ml, respectively (by using a
Petroff-Hausser counting chamber). S. termitida cells were vigorously motile and M.

cuticularis cells F420 fluorescent. H2 (ca. 0.1 ml) was injected into the culture headspace to

106

achieve a final concentration of ca. 1500 ppmv, and H2 disappearance and CH4
production were then monitored by gas chromatography (GC) until rates of H2
consumption were near zero, whereupon another addition of H2 was made and the process
was repeated. After 5 such additions, bromoethanesulfonate (BES; 50 mM) was added to

the medium, and the headspace was again spiked with H2.

‘ Regimented termite diets. All regimented diets included water agar as a source
of hydration for the termites. For this, 20 ml of molten agar (2% w/v) was poured into
petri dishes and allowed to solidify, after which the agar was cut into 4 pieces that were
individually distributed into sterile petri dishes used to contain the termites. To each dish,
0.3-0.4 g fresh wt termites were added, along with an additional food source (Whatrnan
#1 cellulose filter paper; corn starch pellets; or no further additions, if the termites were to
be starved). For feeding antimetabolites, the agar was amended prior to solidification with
either BES (50 mM final conc.); or a combination of rifamycin SV, ampicillin, and
tetracycline (200, 500, and 100 ug/ml w/v, final conc., respectively). Parafilm (American
National Can, Greenwich, CT) was used to seal the edges of the petri dishes in order to
avoid desiccation. Plates were prepared in triplicate for each feeding regimen. Termites
were kept on these diets for up to 10 weeks, during which time mortality was less than
10%. Mortality primarily occurred with specimens of Zootermopsis, which did not

consume the water agar or cellulose filter paper as vigorously as the others species.

Viable bacterial counts. Triplicate gut homogenates were made from control,

BES fed, and rif-tet-amp fed specimens. For each homogenate, 9 guts were extracted and

107

homogenized in 5 ml Brain Heart Infusion broth (BI-II; Difco), after which the volume
was adjusted to 9 ml. Each homogenate was subjected to serial 10 fold dilutions; 100 pl
of the third dilution (i.e., containing 10‘2 gut equiv. - ml) of each series was then plated
onto BI-II agar. After incubation of these plates at 37° C for 48 hours, the colony forming
units of the (indicator) bacteria were counted. The dilution tubes of 8H] broth were also

incubated.

Measurement of gas emission, consumption, and compensation points by
termites and their nest material. Live termites (ca. 0.3-0.8 g flesh wt) were added to 8
ml vials stoppered with butyl rubber septa. The vials were lined with a layer of H20-
moistened paper towel, so that the termites could distribute themselves more evenly over
the internal surface of the vial. In the case of the starch- and agar-fed termites, paper was
not used to line the vial so that the cellulose-flee feeding regimen would be preserved
Samples of headspace gas (100-200 pl) were periodically withdrawn for gas
chromatographic analyses. After each sampling, a wetted, air-filled 2 ml ground glass
syringe was used to equilibrate the vial to atmospheric pressure. In a similar manner, nest
material (0.5 g flesh wt) flom R flavipes and C. formosanus laboratory colonies was also
examined. Gas consumption or production was monitored in the absence or presence of

exogenously added H2. Most such experiments were performed within a 6 hour time span.

Analytical methods. Trace H2 gas analysis (i.e. 0.1 to 20,000 ppmv Hz) was
performed via gas chromatography (GC) with a HgO reduction detector (GC model

RGA2, detector model RGD2; both flom Trace Analytical, Menlo Park, Calif, USA).

108

These units were modified to include a lO-port valve injector (V alco Instruments Co.,

Houston, TX). Several sample loops (V alco) were used for different H2 ranges: 50 pl, for
ca. 1.0-1000 ppmv H2; 50 ul, for ca. 10.0-10,000 ppmv H2; or 500 pl, for ca. SOD-20,000
ppmv H2. Gas standards were either obtained flom Scotty Specialty Gas or were prepared

by mixing H2 with N2.

Quantification of CH4 was performed by conventional gas chromatography with
flame ionization detection (Odelson and Breznak, 1983); aromatic acids were determined
by reverse phase high performance liquid chromatography (RP-HPLC) (Brune et al. ,
1995); organic acids were also determined by HPLC as described in (Breznak and

vaitzer, 1986).

Microscopy. Phase contrast and F,20 epifluorescence microscopy of cells, gut
contents, and gut epithelia were performed as previously described (Leadbetter and
Breznak, 1996). Protozoan counts were determined by using a Petroff-Hausser counting
chamber using cells (released flom minced hindguts) fixed in a glutaraldehyde (4% v/v)

containing buffered salts solution [BSS; described in (Leadbetter and Breznak, 1996)].

109

RESULTS

Isolation of 4 new acetogen strains. Four new strains of acetogenic bacteria were
enriched and isolated --based on their ability to demethylate syringic acid» flom gut
homogenates of three different “higher termite” species (Figure 13 and Table 6). These 4
strains were found to be IIZ/CO2 acetogenic; i.e., they consumed gas when incubated
under H2/C02, but not under NZ/COZ, and formed acetate, assessed by comparing the

respective culture supematants via HPLC analysis.

H2 thresholds of hydrogenotrophic strains. H2 threshold values for all termite
gut acetogen and methanogen strains are given in Table 7. The mean H2 threshold for 7

acetogen strains was 295 i 48 ppmv; the mean for the 3 methanogen strains was 42 i 5

ppmv.

Effect of mixotrophy on the H2 threshold of S. termitida. The apparently
constitutive nature of the H2/CO2 acetogenic system of S. termitida, as well as its ability
to grow mixotrophically (Breznak and Switzer Blum, 1991), prompted us to examine the
effect of organic growth substrates on its H2 threshold. S. termitida cells were grown on
each of several organic substrates [e.g. methanol, 15 mM; succinate, 25 mM; glycine, 5
mM; lactate, 10 mM; mannitol, 5 mM; or in the basal medium alone, which contained
nutrient broth, 0.04% w/v ] under a headspace of 250 ppmv H2 (balance N2/CO2 80/20
v/v). During 15 days incubation during which cell densities reached ca. 107 cells to 108
cells - ml, H2 in each culture never decreased below its initial concentration. Rather, H2

transiently accumulated in several of these cultures. Concentrations of H2 peaked by day

110

Figure 13. Phase contrast micrographs of 4 new acetogenic isolates florn termites.
Arrows indicate endospores. A, str. NAG-2; B, str. GMP-2; C, str. GMP-3 (arrow a,
released mature spore; arrow b, immature spore); D, str. ATR-l. Bar = 10 pm for all

panels.

 

112

Table 6. Characteristics of acetogenic strains isolated flom Congolese “higher-termites“.

 

Strain Source Enrichment Cell Cell Dimension
Dilution Tube Morphology length (um) x width (um)

 

Termite gut equiv. Vegetative Sporulentb
NAG-2 N. arborum 2 - 10’2 Curved rod 5.4 x 0.7 5.4 x 0.9-2.4
GMP-2 M. parvus 1 - 10'2 Curved rod 2.9 x 0.7 2.9 x 1.2
GMP-3° M parvus 2 . 10'2 Flexible >10 x 0.3 10.0 x 0.3-1.5
filament
ATR-l T. trinervoides l Curved rod 3.1 x 0.6 3.8 x 1.0-1.6

 

a All strains were strictly anaerobic; Hz/CO2 acetogenic; demethylated syringic acid; formed
endospores; and were resistant to pasteurization at 80° for 15 minutes.

b The width of cells of all strains tended to increase when they contained endospores.

° Strain GMP-3 differed flom the other strains «which all exhibited tumbling motility when
observed on microscopic wetrnounts-- in that the only evidence for its motility was
during its growth in deep agar “shake” tubes, wherein colonies that formed spread out
through the agar as spheres. It also differed flom the other strains in that it formed 1,2,3
hydroxybenzene afier both decarboxylating and demethylating syringate, whereas the
other strains did not carry out this decarboxylation, thus they formed 3,4,5 trihydroxy
benzoic acid as the final product.

113

Table 7. Hydrogen thresholds of acetogens and methanogens isolated flom termite guts“.

 

 

Organism Source termite Threshold
(ppmv H2)
Acetogens:
Spor‘omusa termitida J SN-2b Nasutitermes nigriceps 267 i 18
Acetonema longum APO-1° Pterotermes occidentis 251 i 3
Clostridium mayombei SFC-Sd C ubitermes speciosus 314 i 12
Acetogen str. NAG-2° Nasutitermes arborum 263 i 14
Acetogen str. GMP-2c Microcerotermes parvus 331 i 9
Acetogen str. GMP-3° Microcerotermes parvus 380 i 30
Acetogen str. ATR-l‘ Trinervitermes lrinervoides 257 i 6
Methanogens:
Methanobrevibacter Reticulitermesflavipes 36 i 3

cuticularis RFM-lf
M. curvatus RPM-2f Reticulitermesflavipes 45 i 7

M. filiformis RPM-3 Reticulitermesflavipes 44 :l: 5

 

‘ Mean values :t SD of triplicate measurements.

b’ °’ d Isolated previously Breznak et al. (1988); Kane et al. (1991); and Kane and
Breznak, (1991); respectively.

° Isolated during this study.

f Characterized in a previous study (Leadbetter and Breznak, 1996).

114

6 at 1800, 1340, and 1400 ppmv for succinate, glycine and lactate grown cultures,
respectively, then fell to 480, 900, 720, 1500 and 490 ppmv H, in the succinate, glycine,
lactate, mannitol, and methanol grown cultures, respectively. HPLC analysis of culture
supematants confirmed that lactate and succinate were completely consumed by the end
of the experiment, producing acetate (98.2 % recovery) and propionate (94.4.0 %
recovery), respectively, as observed in previous studies (Breznak and Switzer Blum,
1991; Breznak et al. , 1988). Although substrate disappearance was not monitored in the
other cultures, acetate production occurred in amounts implying the fermentation of
90.0%, 10%, and 97%, respectively, of the original methanol, glycine, and mannitol

present.

H, threshold in an acetogen + methanogen co-culture. In a co-culture of M. cuticularis
RFM-l and S. termitida JSN-2 initially grown under a headspace of H,/CO, (80/20 v/v),
the headspace was adjusted to a mixing ratio of ca. 2000 ppmv H, (balance N,/CO2 80/20
v/v). H, was then periodically adjusted back to 800-2500 ppmv when it was consumed.
After each of 5 such adjustments over an 8 day period (Figure 14, open arrows), H, was
consumed to levels that ranged flom 40-70 ppmv H,. The recovery of the consumed H, as
CH4 during this 8 day period was determined to be 97% , according to the “classical” 4H,
+ CO, -) CH4 + 2H,O stoichiometry. After H, was added with BES at day 8 (Figure 14,
closed arrow), no further production of CH4 occurred. After a 2 week lag, H, was
consumed down to a mixing ratio of 250 ppmv (data not shown on figure). S. termitida

cells were observed to be motile throughout the course of the experiment.

115

Figure 14. H, consumption by a Sporomusa termitida-Methanobrevibacter cuticularis
co-culture. Arrows indicate the additions of H,. The filled arrow also indicates the
addition of H, and BES (10 mM final cone.) to the culture. Recovery of net H, as CH4

prior to the addition of BES was 97.7%, after the addition of BBS no CH4 was formed.

116

Figure 14. H, consumption by a Sporomusa-Methanobrevibacter co-culture.

1000‘.

H2 (PPMV)

 

 

 

 

 

 

 

 

 

117

H, compensation points of live termites. When various wood-feeding termites
(R. flavipes, C. formosanus, and Z. angusticollis) were either fleshly collected or removed
from laboratory nests and placed in sealed vials under an air headspace, the initial rates of
H, emission were linear. Within ca. 1-2 hours, however, net emission of H, ceased and
the H, concentration in the headspace stabilized. Conversely, if H, was initially present in
the vials in concentrations above this steady state value, H, was consumed, following
pseudo-first order kinetics, until similar steady state H, concentrations were achieved.
Such plateaus occurred at headspace H, concentrations of ca 150-1600 ppmv, regardless
of the species and the food on which the termites were feeding prior to being tested
(Table 8). Such steady state, “plateau” concentrations are hereafter referred to as
compensation concentrations or compensation points (c.p.) and represent the
concentrations of headspace H, which there was zero net flux of H, into, or flom, the
termites in the vial. An example of such H, emission and consumption exhibited by

fleshly-collected R. flavipes workers, as well as the resulting c.p., is shown in Figure 15.

In order to compare the H, c.p. values exhibited by termites with an intestinal
environment wherein methanogenesis dominates, rumen fluid was collected and was
allowed to incubate at 37° C for 3 days, whereupon it exhibited a c.p. of 28 i 4 ppmv [n =
8]. Freshly-collected R. flavipes emitted H, at an initial rate of 0.13 :t: 0.06
umoles - h‘l - (g flesh wt)‘1 [11 = 3] and CH4 at a rate of 0.12 i 0.01 umoles . h" - (g flesh
wt)‘l [n = 2], and exhibited a H, c.p. of 463 :t 121 ppmv H, [n = 5]. Pine-fed, laboratory-

maintained specimens exhibited a somewhat higher H, c.p., 1425 ppmv. H, emission by

118

Table 8. H, emissions and compensation points exhibited by 3 species of termites fed

 

 

 

different substrates".
Termite H, c.p. Gas emission (initial rate)
Food Substrate (ppmv) (umoles - [gram flesh wt - h]")

H, CH,

Reticulitermesflavipes:
Freshly collectedb 463 :1: 121 (5) 0.13 i 0.06 0.12 :1: 0.01 (2)
Pine-fedc 1425 i 177 (2) 0.18 (1) 0.12 :t: 0.03
Birch-fed“ 584 :1: 27 (2) 0.06 i 0.00 (2) ----- °
Starchf 450 i 218 0.02 i 0.00 (2) ----
Cellulose-fed“ 1500 :t 436 0.31 :l: 0.1 ----
Starvedh 164 :t: 38 0.02 :1: 0.01 ---
Starved (anoxic incubation)”i 467 :t 207 0.03 :1: 0.01 -----
Coptotermesformosanus:
Freshb 574 i 60 0.23 (1) <0.01
Pine-fedc 1567 i 52 0.63 :t 0.10 -----
Starved“ 274 i 65 (6) 0.07 (1) -----
Zootermopsis angusticollis:
Pine-fedj 1167 :t 290 (2) 0.29 :t: 0.03 0.70 i 0.09

Starved“ 266 i 101 (9) 0.02 i- 0.00

 

119

Table 8 footnotes, continued from previous page.

° Measurements performed in triplicate, except when the number is reported within
parentheses that occur after a given value.

b Experiments performed within 1) 6 hours of collection of R. flavipes flom cardboard
traps set in Dansville, MI, or 2) 3 days of receipt [flom Ft. Lauderdale, FL] of
cardboard traps containing freshly collected C. formosanus.

° Termites were laboratory maintained on pine for 3 months.
d Termites were laboratory maintained on birch for 2 months.
6 Measurement not performed.

fMaintained on a starch diet for 10 weeks after collection, in order to defaunate the
termites of protozoa. Protozoan counts after such treatment were
15,000 :1: 10,800 per gut equivalent, with those protozoan cells remaining being small,
Trichomonas—type flagellates which are not considered to be cellulolytic. In
comparison, protozoal counts were 62,000 :1: 8,500 for termites fed cellulose for the
same period of time.

9’ Maintained on a filter paper diet for 7 days after being collected.

h Maintained on water agar for 2 days (after collection flom the field or after receipt, see
Methods) in order to starve specimens of fermentable substrates.

i Termites were incubated under 100% N, during such measurements.

j Termites were laboratory maintained on pine for ca. 8 years prior to receipt.

120

Figure 15. Emission of endogenous H, (B) and consumption of exogenous H, (p) by 1
gram of fleshly collected R. flavipes incubated under air in a stoppered, 8 ml vial. Note

that both curves converge at a compensation point of ~ 470 ppmv H,.

121

Figure 15. H, consumption, emission, and the compensation point exhibited by fleshly

collected R. flavipes worker termites.

 

 

 

 

2000'-
1000-. °“<> .......
I O‘O-Q ..... (>-
9 .. “0 -------
E I— I— I - I ”'0“ —-<>
a I
a: _ _/ '
I? - F
1001',F
a;
50 IIIIIITIIIIITIlllllllllllllllllllllllllll
100 200 300

Minutes

122

these specimens was slightly greater than that of fleshly-collected specimens. The higher
H, c.p. and H, emission rates of pine-fed R. flavipes may reflect a greater digestibility of
this wood. By contrast, laboratory maintained, birch-fed R. flavipes exhibited a c.p. more
similar to that of the fleshly-collected specimens, but emitted H, at ca. one half the rate.
Similar results were observed for C. formosanus, another subterranean termite in the

same Family (Rhinotermtidae) as R. flavipes.

To investigate the importance of protozoa in H, emission and in the modulation of
the H, c.p., R. flavipes was fed starch for 8-10 weeks. Such termites became completely
defaunated of their larger protozoa (i.e. Dinenympha spp., Holomastigotes spp.,
Pyrsonympha spp., Trichonympha spp., and Spirotrichonympha spp.), although smaller
Trichomonas-type protozoa «which do not appear to be cellulolytic« were still present.
The overall protozoan count in starch-fed termites was 1.5 - 104 i 1.1 ' 104 [n = 3], a
nearly 4-fold decrease in comparison to specimens fed cellulose for the same period of
time (6.2 - 10‘ :I: 0.9 - 10‘ [n = 3]). Spirochetal diversity also appeared to decrease in
starch-fed termites, as judged qualitatively by phase contrast microscopy. Here, two
morphotypes dominated rather then the typical 10-20 different morphotypes. The
observation that the spirochetal population had changed was supported by an analysis of
rDNA PCR amplified flom hindgut DNA (W .-T. Liu and T. Marsh, unpublished results).
By contrast, the density of F4,0 fluorescent methanogen cells attached to the gut
epithelium did not appear to change in starch fed termites. Despite this alteration of the

protozoan and Spirochetal population of the hindgut upon starch feeding, the H, c.p. of

123

such R. flavipes termites, 450 :t 218 ppmv [n = 3], remained similar to that exhibited by

fleshly collected specimens. However, their H, emissions decreased by ca. 80%.

By contrast, fleshly-collected R. flavipes fed cellulose for 7 days exhibited
increases in both H, c.p. and rate of H, emission (Table 8). Specimens maintained on
cellulose for longer periods of time, or specimens that had been laboratory-maintained on
pine prior to cellulose feeding (detailed in a later section), failed to exhibit a c.p.; rather,
they continued to emit H, to over 5000 ppmv. This was also observed during prokaryotic-

inhibitor studies (below).

When R. flavipes were fed only water agar for 24-48 hours in order to starve them
of fermentable substrates, their H, c.p. decreased to 164 :l: 38 ppmv, as did the rate of H,
emission (Table 8). When starved termites were incubated under N, during the
measurements, their H, c.p. (467 ppmv) was similar to that of fleshly collected-specimens

and the rate of H, emission was also slightly.

Coptotermesformosanus emitted H, at an initial rate of 0.23 umoles H, - h" ~ (g
flesh wt)", but emitted virtually no methane, i.e. $0.01 umoles CH4 - h" - (g flesh wt)"
[n=3] (T able 8). When maintained on pine for 3 months, their rate of H, emission and

their H, c.p. increased substantially; when they were starved, both decreased (Table 8).

Enrichments for H,-consuming anaerobes flom gut homogenates of Coptotermes
(using medium JM-4) were not successful, nor were Fm fluorescent cells ever observed
during microscopic examination of the gut fluid, gut epithelium, or gut protozoa of this

termite. The absence of significant methane production and F 4,0 fluorescent cells in C.

124

formosanus is in contrast to the results of Tsunoda (1993), who demonstrated methane
emission as high as ca. 0.375 umoles CH4 - h" ' (g flesh wt)" and Pseudotrichonympha—
associated F 4,0 fluorescent cells in Japan-collected specimens of C. formosanus (T sunoda
et al., 1993; Tsunoda et al., 1993). The absence of methanogens and methane emission
flom Florida-collected specimens used here probably reflects an oft-observed intra-

specific variation, the basis for which is not understood (Wheeler et al. , 1996).

The dampwood termite Zootermopsis angusticollis emitted 0.29 i 0.03 umoles H,
. h" ~ (g flesh wt)" [n=3] and 0.70 i 0.09 umoles CH4 . h" . (g flesh wt)" [n=3] (Table 8).
These specimens had been maintained on pine for ca 8 years prior to their receipt. Their
H, c.p. (1167 ppmv) was similar to that of pine-fed R. flavipes and C. formosanus (1425
and 1567 ppmv, respectively). Zootermopsis also behaved similarly to these other two
other termite genera in that when starved, H, emission and the H, c.p. decreased

substantially (Table 8).

Microscopic examination of Zootermopsis gut contents reconfirmed the
demonstration by Lee et al (1987) that F4,o fluorescent cells were abundant on or within
the smaller protozoa T ricercomitus sp. and Hexamastix sp. which are not considered to be
cellulolytic. Fluorescent cells were less abundant in the medium sized protozoan
Trichomitopsis, and completeley absent in flom the large Trichonympha spp., all of
which are cellulolytic. This study extends those observations by noting that the greatest
density of Tricercomitus protozoa (and their associated F4,0 fluorescent, methanogen

cells) was in a small region of the hindgut just anterior to a constriction found in the

125

anterior region in the hindgut paunch, just posterior to the midgut-hindgut junction. This
protozoan coexisted with dense populations of the protozoan Streblomastix strix,
reconfirming Kidder’s localization of this latter eukaryote (Kidder, 1929). Larger
protozoa such as T richonympha spp. and the medium sized T richomitopsis were not

observed to be present in this anterior location.

Throughout the hindgut, P.,,o fluorescent short rods were occasionally observed to
be attached to the epithelium, although not at the high cell densities seen on the gut wall

of R. flavipes temrites or in the region just described.

Acetogens and methanogens were readily enriched flom gut fluid exuded flom the
proctadeum of Zootermopsis angusticollis using medium JM-4. However, no was made

to isolate pure cultures flom these enrichments.

Effect of prokaryotic inhibitors on gaseous emissions of R. flavipes. Results of

these studies are summarized in Table 9.

When a mixture of anti-eubacterial drugs (rifamycin SV, tetracycline, and
arnpicillin) was fed to R. flavipes for 8 days, the bacterial population in the hindgut
decreased dramatically. No colonies grew on BI-II plates or in BI-II Broth inoculated with
gut homogenates containing 8 gut equivalents. Moreover, the only prokaryotic cells
observed in microscopic preparations of rif-tet-amp fed termites were F 4,0 fluorescent

(methanogenic) cells associated with the gut wall. By contrast, in control and BBS—fed

126

termites bacteria were present at 5.15 . 104 i 2.69 - 10" and 5.70 . 104 i- 0.55 - 104 CPU
per termite gut equivalent, respectively. H, emissions by these latter two 100-fold by Day
8, probably reflecting the more ready digestibility of the non-lignified, pure cellulose

(filter paper) food source.

The rate of H, emission in rif-tet-amp treated termites increased flom 0.18 umoles
- h" ' (g flesh wt)" at day 0 to 3.26 i 0.9 umoles - h" ° (g flesh wt)" [n = 3] by day 4 and
12.61 i 1.60 umoles - h" - (gram flesh wt live termite)" [n = 3] by day 8, at which time
they no longer exhibited an H, c.p. Instead, H, continued to accumulate in the vials to
concentrations in excess of 5000 ppmv. The rate of H, emission by day 8 flom the control
(no inhibitor) and BBS-fed groups also increased, but 5- to 7-fold. They also ceased to
exhibit an H, c.p. by day 8, although both had exhibited a c.p. on day 4 (Table 9). On day
28, at which time the three groups had been off of their cellulose diet starved for 14 days,
rates of H, emission had decreased roughly 20-fold for the untreated and BES treated
specimens and l40-fold for the rif-tet-amp treated specimens, compared to day 8.
Interestingly, the control and BBS-fed groups regained their ability to exhibit an h2 c.p.,
although the H, c.p.’s (460 ppmv and 490 ppmv, respectively) were somewhat lower than
at day 0. In contrast, rif-tet-amp treated termites did not regain their ability to exhibit an
H, c.p., even though their rate of H, emission had decreased to below that emitted at day

0 (Table 9).

127

Table 9. Effect of prokaryotic inhibitors on H, compensation point and gas emissions by

cellulose-fed R. flavipes termites'.

 

Inhibitor

Hydrogen c.p. Gas emission (initial rate)
(ppmv H,) (umoles - [gram flesh wt - h]")

 

H, CH,

 

Day 0: (pine-fed)”

Day 4: (cellulose fed)
None
Rif-Tet-Ampc

BES‘

Day 8: (cellulose fed)
None ’
Rif-Tet-Amp‘

BESh

Day 28:i (starved since day 14)
None
Rif-Tet-Amp

BES

1425 :1: 177(2) 0.18 (1) 0.12 :t: 0.01 (2)

1300(1) 022(1) 0.04 (1)
>5000 d 3.26 i 0.90 0.11 a: 0.00

1800(1) 0.51 (1) 0(1)
>5000 0.97 :t 0.25 0.07 :t 0.01
>5000 12.61 1: 1.60 0.11 i 0.06
>5000 1.20 :t 0.68 0

460 :1 61 0.04 i 0.01 0
>5000 0.09 i 0.01 0

490 i 160 0.05 i 0.02 0

 

128

Table 9 footnotes, continued from previous page.

8 Measurements performed in triplicate, except when the number is reported within
parentheses that occur after a given value.

b Day 0 termites had viable bacterial cell counts in BI-II of 5.08 ° 10‘ :1: 0.88 - 10‘ colony
forming units per gut equivalent.

° Rifamycin SV, tetracycline, and ampicillin are inhibitory to most Bacteria, but
methanogenic Archaea are naturally resistant to these drugs.

d H, emission did not plateau in such specimens, thus they did not exhibit a compensation
point.

‘ BES (bromoethanesulfonate) is a potent inhibitor of methanogenesis in Archaea, but is
not known to inhibit any activities of Bacteria. Its use in the elimination of
methanogens flom termites has previously been demonstrated (Messer and Lee, 1989).

f. g. h Viable cell counts (as colony forming units per gut equivalent) at day 8 were 5.15 -

10‘ :t 2.69 - 10‘, 0, and 5.7 - 104 :t 0.55 ~ 10‘ for the untreated, rif-tet-amp treated, and
the BES treated termites, respectively.

i After 14 days of feeding on cellulose, termites were starved by feeding them water agar
only. However, treatment with inhibitor-containing water agar continued during this
time.

129

Methane emission remained essentially unchanged in rif-tet-amp treated termites
while feeding on cellulose, decreasing slightly flom 0.12 umoles CH4 - h" - (g flesh wt)"
at day 0 to 0.11 :l: 0.06 umoles CH4 - h" - (g flesh wt)" at days 4 and 8 (Table 9).
Moreover, F4,0 fluorescent cell density did not appear to fluctuate in rif-tet-amp treated
termites during the course of the cellulose feeding. This was also largely true of the
control cellulose-fed specimens, although by day 8 their rate of emission of methane had

decreased to 40% of what it had been at Day 0 (Table 9).

BBS-treatment completely eliminated CH4 emission and the presence of P.,,0
fluorescent cells in the guts of R. flavipes by day 4. This was accompanied by a small
increase in the rate of H, relative to the control group (Table 9). The small increase in H,
emission by the BBS-fed termites at days 4 and 8 was close to that expected flom their
decrease in CH4 emission, assuming a 4 to 1 ration between H2 consumption and CH4

production.

When termites were assayed on Day 28, after 14 days starvation, CH4 emission
did not occur in any of the three test groups (Table 9), nor were F4,0 fluorescent cells

observed to be present in their guts.

Gas Consumption by Nest Material. Laboratory-maintained specimens of R.
flavipes distributed a gray carton-like material as enclosed runways over and around the
surfaces of the colony. This material was 82% w/w H,O, and samples of it consumed H,

at an initial rate of 3.44 :t 0.28 umoles - h" - (gram dry wt)" [n = 3] when incubated under

130

a 1% H, headspace. Such material also exhibited a H, threshold of 1.3 i 0.2 ppmv [n =
3]. No evidence was obtained for CH4 consumption by this material, even after several

days of incubation.

C. formosanus termites also distributed a tan carton material as lace-like arrays
extending flom the infested squares of pine within the laboratory nest. This material was
84 % w/w H,O, and samples of it consumed H, at an initial rate of 9.14 d: 2.27 umoles - h'
‘ - (gram dry wt)" [11 = 4 ] at 1% initial H,. It also exhibited a H, threshold close to that of
atmospheric H, concentrations, which were near the detection limits of the gas
chromatograph, i.e. ca. 0.5 i 0.2 ppmv [n = 3], but did not exhibit any CH4 consuming

activity.

131

DISCUSSION

H, thresholds exhibited by termite gut acetogens, isolated either previously or
during this study, ranged flom 251 - 380 ppmv and were comparable to those determined
previously for 2 flee living species, A. woodii and A. carbinolicum [260 and 300 ppmv,
respectively; (Conrad and Wetter, 1990)]. In an earlier study, the H, threshold of S.
termitida was determined to be 830 ppmv :t 415; those of A. woodii and A. carbinolicum
were 520 ppmv :t 260 and 950 ppmv :t 475, respectively (Cord-Ruwisch et al., 1988).
These higher values may reflect the differences in the way the cultures were grown, i.e.
under elevated initial H, concentrations (80% v/v) rather than lower ones (1% v/v). This
possibility is supported by the observation that higher threshold values (836-3590 ppmv
H,) were obtained with our 7 acetogenic strains when they were grown initially at 80 %
v/v H, (data not presented in Results). Inasmuch as H, concentrations at such magnitudes
(80% v/v) are not typically found in natural habitats, the thresholds attained after

incubation under the lower H, concentrations likely represent more meaningful values.

By contrast. the H, thresholds exhibited by the termite gut methanobrevibacters
used in this study were about 7-fold lower than that of the termite gut acetogens, ranging
flom 36 - 45 ppmv. Such values were similar to those determined previously for the flee-
living methanogens Methanospirillum hungatei, Methanobacterium formicicum, and M.
bryantii [28, 30 and 30 ppmv, respectively; (Conrad and Wetter, 1990; Cord-Ruwisch et
al. , 1988)]. Two other methanobrevibacters, the human gut isolate M. smithii and the

“wetwood diseased” cottonwood isolate, M. arboriphilicus, have been reported to exhibit

132

slightly higher H, thresholds of 100 and 90 ppmv H,, respectively (Cord-Ruwisch et al.,
1988). From these results, it seems clear that the enigmatic dominance of acetogens over

methanogens as an H, sink in guts of wood-feeding termites cannot be explained by

unusually low H, thresholds.

Neither can the competitiveness of termite gut acetogens for H, be attributed to
mixotrophy, as the headspace H, concentration in cultures of S. termitida growing on five
different organic substrates [four of which have previously been shown to support
mixotrophic growth by this strain (Breznak and Switzer Blum, 1991)] actually increased
flom the initial level of 250 ppmv to 450 - 1500 ppmv. These results are consistent with
those of an earlier study (Breznak et al., 1988), in which it was observed that H, was a
product of the fermentation of mannitol, glycine, lactate, and methanol, accumulating to
concentrations in excess of 2000 ppmv in the gas phase. In another study, S. acidovorans
was shown to display a H, threshold of 430 ppmv when growing on methanol + CO,
alone (Cord-Ruwisch et al. , 1988), and thus it too must have emitted H, until a steady
state was reached that was comparable to its H, consumption threshold. Likewise,
methanogens such as Methanosarcina produce H, during acetoclastic methanogenesis,
attaining a H, threshold similar to that of cells consuming exogenous H, alone (Lovley
and Ferry, 1985). Although Methanosarcina may grow mixotrophically on H, and
acetate, this is likely a function of having high concentrations of available H,. It appears
that acetogens and methanogens may consume H, while growing on an organic substrate
when the ambient H, concentration is significantly above their H, threshold (i.e. they

grow mixotrophically), but they may also produce H, to concentrations at or above their

133

H, threshold during the fermentation of these same organic substrates. Thus, although
mixotrophic growth by acetogens may increase their overall energy yield and can result
in higher overall growth rates in vitro (Breznak and Switzer Blum, 1991), it is unlikely
that such broad metabolic capabilities contribute to their competition for low

concentrations of H, per se.

This interpretation is supported by experiments in which S. termitida was co-
cultured with M. cuticularis (Figure 14). H, was repeatedly consumed to levels below 70
ppmv (i.e., close to the methanogenic threshold) after each of five H, additions; the H,
being almost completely recovered (97%) as CH4. Thus, even when provided with H, at
concentrations 5 to 10 times above its acetogenic H, threshold, S. termitida (even at cell
densities three times greater than M. cuticularis) was not able to outcompete the
methanogens for H,. Even after BES had been added to the medium, the S. termitida

exhibited a lag time of 2 weeks before consuming H, down to its threshold.

H, relationships exhibited by live termites. H, is emitted by termites in small
quantities. This was first suspected by the early studies of Cook (1932) and confirmed
later by La Page and Nutting (1979) as well as Odelson and Breznak (1983). The current
study extends our understanding of H, emission in termites by showing that they not only
emit H,, but that they also are able to consume it and display a “compensation point”, i.e.
a concentration of ambient H, at which there is a zero net flux of H, into or out of the
insect [discussed in (Conrad, 1994; Conrad, 1995; Conrad, 1996)]. To our knowledge,
this is the first demonstration of an H, compensation point for an animal. Such results,

obtained by using a relatively noninvasive method of analysis, allow one to conclude that

134

the magnitude and direction of H, flux between termites and their surroundings will
depend on the ambient H, concentration (discussed below). Moreover the H, c.p. is likely
to be indicative of the lowest concentration of dissolved H, in the hindgut. This is likely
to occur at the periphery of the gut, since this would be the site of most rapid gas
exchange with the atmosphere. The estimated in situ concentration at the periphery can
then be estimated flom the c.p. by considering that 1 ppmv H, gas mixing ratio is
equivalent to 0.78 nMolar dissolved H,. [One liter of H, gas at 25° C is equivalent to 40.9
mmoles H,. The Otswald coefficient for H, gas dissolved in H,O is 0.0191 ml per ml H,O

at 1 atrn at 25° C.]

The H, c.p. values of fleshly-collected R. flavipes and C. formosanus termites
were 463 and 574 ppmv, respectively. When these two species were laboratory-
maintained on pine, their rate of H, emission increased, and this coincided with a ca. 3-
fold increase in their H, c.p. (Table 8). The increase in the c.p. presumably reflects an
inability of the hydrogenotrophic community to completely keep pace with increased H,
production. Consistant with this interpretation was the drop in the c.p. to levels more
typical of fleshly collected termites when pine- or birch-fed termites were starved (Table
8). The somewhat lower H, c.p. exhibited by fleshly collected or birch-fed termites
versus pine- or cellulose-fed termites may reflect a more ready degradability of the latter

food sources (Table 8).

The coincidence between the rate of H, emission and the magnitude of the c.p.

was readily apparent when R. flavipes, C. formosanus, and Z. angusticollis termites were

135

starved and they all exhibited lower H, c.p. values than when they were wood-fed and
their rates of H, emission decreased (Table 8). However, in none of the three species did
c.p. values ever fall below 125 ppmv H, when they were starved. Additionally, the
exhibition of a H, c.p. did not appear to be O,-dependent, inasmuch as the H, c.p. of
starved R. flavipes incubated under N, (467 ppmv H,) remained close to that of fleshly-

collected specimens (463 ppmv).

These last two points «that the lowest H, c.p. observed was ca. 125 ppmv, and
that the exhibition of a H, c.p. was O,-independent« are significant. They suggest that the
lower limit may be related to the energetics of the primary H,-consuming anaerobic
activity occurring in the hindgut community, i.e. acetogenesis, an activity well
documented in this termite species (Brauman et al., 1992; Breznak and Switzer, 1986). It
is also important to recognize that an acetogenic H, consuming population, no matter how
active it might be at elevated H, concentrations, would not be expected to contribute any
more effectively in the exhibition of a H, c.p. First, 125 ppmv H, is already ca. 50%
below the lowest H, threshold that has been determined for pure cultures of acetogens
(Table 7). Secondly, 125 ppmv H, is strikingly close to the theoretical acetogenic H,
threshold, i.e. acetogenesis could only yield ca. 15 k] of flee energy per reaction at such
H, concentrations (Table 5; compare the in situ H, concentrations of methanogenic or
sulfate-reducing environments and the energetics of those processes at such H,
concentrations). Methanogens could theoretically obtain ca. 3-fold more energy at 125
ppmv H,, i.e. 50 kJ per reaction, and would be predicted to lower the H, c.p. to anywhere

between ca. 5 to 100 ppmv in starved termites, if their population density and overall

136

activity allowed this. For reasons that remain uncertain, methanogens have not achieved

in the termite what they have in vitro and in many other environments.

Therefore, it appears that 1) the bulk of the H, consumptive activity associated
with the exhibition of a H, c.p. in these termites is provided by anaerobic, acetogenic
bacteria; and 2) hindgut acetogens are not operating in a fashion that is far superior to that
which is predicted by theory and which is exhibited by pure cultures. Consistant with this
interpretation were the results of inhibitor studies (Table 9) , which indicated that the
consmnption of exogenous H, and the exhibition of a H, c.p. by R. flavipes was
dependent upon the activity of rif-tet-amp sensitive microbiota, i.e. Bacteria. When
Bacteria were eliminated flom R flavipes, H, emissions increased dramatically, and the
ability of the termite to exhibit a H, was lost. In contrast, the elimination of most of the
protozoa or methanogens flom R. flavipes hindguts did not effect the exhibition of a H,
c.p., since the H, c.p.’s of starch-fed (Table 8) or BBS-fed (Table 9; day 4 and 28)
termites were comparable to that of the control group (compare the starch H, c.p.

presented in Table 8 with the BBS-fed and control groups.

Comparison of the rates at which H, was emitted by R. flavipes during this study
allows several other conclusions to be made. For example, by comparing the rate emitted
of H, emission by the rif-tet-amp treated group and the control group at day 4 (when the
control and BBS-fed specimens still exhibited a c.p.; Table 9), it can be inferred that 3.04
umoles H, - [gram flesh wt - h]" and 2.75 umoles H, - [gram flesh wt - h]" are normally

being consumed by rif-tet-amp-sensitive microbes. This value is close to that inferred for

137

H, consumption by hindgut homoacetogens: 3.36 umoles H, - [g flesh wt - h]"; based on
rates of H,-dependent reduction of 1“CO, to acetate by gut homogenates of R. flavipes
(Brauman et al. , 1992). Likewise, rates of H, emission by the BBS-treated termites at day
4 imply that 0.29 umoles H, - [g flesh wt - h]" are normally consumed by hindgut
methanogens of R. flavipes. this is in good agreement with the amount of H, that would
support a CH4 emission rate of 0.04 umoles - [gram flesh wt - h]" (Table 9), assuming

that 4 mol H, is consumed per mol CH4 emitted.

Comparison of H, emission by these groups at days 4 and 8 (Table 8) also
indicates that it is unlikely that the absence of an H, c.p. in the BES-fed and control
groups at day 8 is due to an inhibition of hydrogenotrophic bacterial activity. Considering
that the rate of H, emission by the rif-tet-amp fed termites had increased to 12.6 umoles
H, - [g flesh wt - h]" by day 8 (see Table 9), it is inferred that ca 11.6 and 11.4 umoles
H, - [g flesh wt - h]" were actually being consumed in the guts of the control and BBS fed
group, respectively « 3.8- and 4.1-fold increases in activity flom day 4. Clearly, the
acetogenic community was responding to a dramatic increase in the internal production of
H, (which was almost certainly due to a more ready digestibility of cellulose filter),
although not to the extent of being able to maintain or create a H, compensation point. On
a lighter side, the rate of H, production by rif-tet-amp fed termites at day 8 equates to a
remarkable production of gas: more than 1 ul of H, was emitted every hour by each
individual gut of R. flavipes (whose guts are typically less than 1 ul in volume); imagine
what their discomfort might have been were it not for the excellent gas exchange likely

mediated by their tracheated gut epithelium!

138

Results of inhibitor studies also indicate that Archaeal methanogens are not likely
to be involved in acetogenic activity in situ, as was hypothesized at the end of Chapter 2
(Leadbetter and Breznak, 1996). If they were, then the effect of rif-tet-amp treatment on
H, consumption would have been minimal, since Archaea [specifically those

methanogenic isolates described in Chapters 2 and 3] are resistant to such drugs.

It is still curious that methanogenic cells did not take advantage of the dramatic
increase in H, availability in rif-tet-amp treated termites, as judged by their meager
increase in methane emission (Table 9). It is also curious that the absence of
methanogenesis in BBS-fed R. flavipes was associated with a marginal increase in H,
emissions, even though (as mentioned above) the bacterial component was capable of
dramatically increasing its H,-consuming activity overall. It appears that the two groups
of organisms were unable to fill each others niche. Inasmuch as both groups are
hydrogenotrophic, it seems unlikely that the differences in their niches are defined by
their common energy source (H,) alone, although it may be defined by the cells that
produce that substrate. Another explanation is becoming ever more relevant to consider:
the axial and radial spatial distribution of acetogens and methanogens [especially in
relation to chemical, physical, and biological (protozoal) stratification] in the gut of
wood-feeding termites may explain the dominance of one group over the other. This has

recently been proposed by a Andreas Brune who, in a soon to be published study, stated:

“It is no longer a question of [understanding the] direct competition for a

mutual resource [i.e. H2], but rather [of understanding why and how it is that] a

139

resource partitioning is [being] effected by the spatial distribution [i.e. separation] of

the different hydrogen-consuming populations within the gut.”

Towards that end, Brune docmnented precisely the nature of H, profiles in guts
extracted from the termite R. flavipes by using H, microelectrodes. Although caution
must be exercised in the interpretation of such results (because of the extraction
procedure and the subsequent restriction of gas flow between the gut and the atmosphere
by the agar embedding the gut), this elegant study has profound implications. His results
indicated that their were 2 distinct zones of H, consumption occurring in the hindgut of
this termite: the first was within the central region, and the second was at the epithelial
surface. Although O, availability and toxicity was shown to affect such H, profiles, Brune
concluded that the consumption of H, remained essentially an O,-independent process (in
agreement with one of the conclusions made during this study), suggesting that the two
sinks reflect the differential location of the acetogenic community (i.e. in the central
regions) and the methanogenic community (i.e. in the peripheral regions, which was

documented in Chapter 2).

In future studies, it will be important to determine the physical location of
acetogen cells in hindguts of termites such as R. flavipes, especially relative to 1) the
radial and/or axial H, gradients therein, and 2) the location of methanogen cells, which
are located on the epithelium. Such studies could make use of nucleic acid techniques:
perhaps localization could be achieved by using an in situ PCR/hybridization techniques
targeting relevant acetogenic genes [such as the gene encoding the formyl-

tetrahydrofolate synthetase (F THFS), for which a DNA probe has already been developed

140

(Lovell and Hui, 1991); or those genes encoding the carbon monoxide dehydrogenase

complex (CODH)].

It would also important to determine the H, c.p. a soil-feeding and/or fungus
cultivating “higher termite” species, wherein the hindgut “electron sink” is dominated by
methanogenesis (Brauman et al., 1992), as well in wood-feeding “higher termites” which
contain no cellulolytic protozoa. Comparison of such H, c.p. values with those presented
here would indicate whether any correlation can be drawn between the c.p. value
exhibited and the feeding strategies or dominant hindgut terminal electron sink reactions
in these diverse termites. It wouldn’t be surprising to find that in soil-feeding termites,

whose H, sink is dominated by methanogenesis, the H, c.p.

Gas consumption by nest material. That termite nest material may constitute an
H, sink was previously proposed (Khalil et al., 1990), wherein it was shown that H,
concentrations in termite mounds were low or even depleted in respect to atmospheric H,.
Preliminary measurements reported here suggest that H, consumption by nest material is
significant enough to consume H, that is emitted by termites, and it is unlikely that H,
accumulates in the galleries and runways of these termites. Most likely, H, mixing ratios
around termites will be close to that of atmospheric, ca. 0.5 ppmv. Thus, the direction of
flow of H, is always expected to be from the termite to its surroundings. As a result, a
steady-state gradient of H, is likely to form between the sources of H, within the central

regions of the termite gut and the external sink. Such a steady-state would contrast with

141

those achieved under artificial laboratory conditions such as: 1) at H, c.p. concentrations
in a vial; and 2) with extracted guts embedded in agar (such as those used by for
microelectrode measurements). In the future, it is the steady state reached between the
live termite and the low ambient H, concentrations of its surroundings (such as in the nest
or at the foraging site) that must be taken into account when making more refined

predictions about H, gas flux flom, and concentrations within, termite guts.

142
ACKNOWLEDGMENTS
I would like to thank Bob Hickey (for suggesting how best to initiate trace gas

analyses) and Andreas Brune, for years of helpful discussions and for providing access to

his unpublished study on H, profiles in R. flavipes termites.

REFERENCES

Works cited in Chapters 1, 3 and 4 are referenced together in a Bibliography
provided at the end of the thesis. Chapter 2 maintains the citation numbers and

bibliography (provided within the chapter) used for its publication.

10.

11.

12.

143

REFERENCES

Benemann, J. R. 1973. Nitrogen fixation in termites. Science 181:164-165.

Bentley, B. L. 1984. Nitrogen fixation in termites: fate of newly fixed nitrogen. J.
Insect Physiol. 30:653-655.

Berchtold, M., W. Ludwig, and H. Konig. 1994. 16S rDNA sequence and
phylogenetic position of an uncultivated spirochete flom the hindgut of the
termite Mastotermes darwiniensis. FEMS Microbiol. Lett. 123:269-274.

Boone, D. R., and R. A. Mah. 1989. Family I. Methanobacteriaceae Barker

1956, 15,AL* emend. Balch and Wolfe in Balch, Fox, Magrum, Woese and Wolfe
1979, 267, p. 2174-2183. In J. T. Staley, M. P. Bryant, N. Pfennig, and J. G. Holt
(ed.), Bergey's Manual of Systematic Bacteriology, vol. 3. The Williams &
Wilkins Co., Baltimore.

Brauman, A., M. D. Kane, M. Labat, and J. A. Breznak. 1992. Genesis of
acetate and methane by gut bacteria of nutritionally diverse termites. Science
257:1384-1387.

Breznak, J. A. 1994. Acetogenesis flom carbon dioxide in termite guts, p. 303-
330. In H. L. Drake (ed.), Acetogenesis. Chapman & Hall, New York.

Breznak, J. A. 1975. Symbiotic relationships between termites and their
intestinal microbiota Soc Exp Biol Symp Ser 29:559-580.

Bra-ak, J. A., W. J. Brill, J. W. Mertins, and H. C. Coppel. 1973. Nitrogen
fixation in termites. Nature 244:577-580.

Bra-ak, J. A., and H. S. Pankratz. 1977. In situ morphology of the gut
microbiota of wood-eating termites [Reticulitermes flavipes (Kollar) and
C optotermes formosanus Shiraki]. Appl Environ Microbiol 33:406-426.

Breznak, J. A., and J. Switzer Blum. 1991. Mixotrophy in the termite gut
acetogen, Sporomusa termitida. Arch Microbiol 156:105-110.

Breznak, J. A., and J. M. Switzer. 1986. Acetate synthesis flom H2 plus C02 by
termite gut microbes. Appl Environ Microbiol 52:623-630.

Breznak, J. A., J. M. Switzer, and H. J. Seitz. 1988. Sporomusa termitida sp.

nov., an H2/C02-utilizing acetogen isolated flom termites. Arch Microbiol
150:282-288.

13.

14.

15.

16.

17.

l8.

19.

20.

21.

22.

23.

24.

144

Brune, A., D. Emerson, and J. A. Breznak. 1995. The termite gut microflora as
an oxygen sink: microelectrode determination of oxygen and pH gradients in guts
of lower and higher termites. Appl Environ Microbiol 61:2681-2687.

Brune, A., and M. Kuhl. 1996. pH profiles of the extremely alkaline hindguts of
soil-feeding termites (Isoptera: Terrnitidae) determined with microelectrodes. J.
Insect. Physiol. 42:1 121-1 127.

Brune, A., E. Miambi, and J. A. Breznak. 1995. Roles of oxygen and the
intestinal microflora in the metabolism of lignin-derived phenylpropanoids and

other monoaromatic compounds by termites. Appl Environ Microbiol 61:2688-
2695.

Conrad, R. 1994. Compensation concentration as critical variable for regulating
the flux of trace gases between soil and atmosphere. Biogeochemistry 27:155-
1 70.

Conrad, R. 1995. Soil microbial processes involved in production and
consumption of atmospheric trace gases. Adv Microb Ecol 14:207-250.

Conrad, R. 1996. Soil microorganisms as controllers of atmospheric trace gases
(H2, CO, CH4, N20, NO). Microbiol Rev 60:609-640.

Conrad, R., T. J. Phelps, and J. G. Zeikus. 1995. Gas metabolism evidence in
support of the juxtaposition of hydrogen-producing and methanogenic bacteria in
sewage sludge and lake sediments. Appl Environ Microbiol 50:595-561.

Conrad, R., and B. Wetter. 1990. Influence of temperature on energetics of
hydrogen metabolism in homoacetogenic, methanogenic, and other anaerobic
bacteria. Arch Microbiol 155:94-98.

Cook, S. F. 1932. The respiratory gas exchange in T ermopsis nevadensis. Biol
Bulletin 63:246-257.

Cord-Ruwisch, R., H. J. Seitz, and R. Conrad. 1988. The capacity of
hydrogenotrophic anaerobic bacteria to compete for traces of hydrogen depends
on the redox potential of the terminal electron acceptor. Arch Microbiol 149:350-
357.

Edwards, R., and A. E. Mill. 1986. Termites in buildings - Their biology and
control. Rentokil Limited, Esat Grinstead.

Fraser, P. J., R. A. Rasmussen, J. W. Creffield, J. R. French, and M. A. K.
Khalil. 1986. Termites and global methane - another assessment. J Atmospheric
Chemistry 4:295-310.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

145

Goodwin, S., E. Giraldo-Gomez, B. Mobarry, and M. S. Switzenbaum. 1991.

Comparison of diffusion and reaction rates in anaerobic microbial aggregates.
Microb Ecol 22:161-174.

Holt, J. G., N. R. Krieg, P. H. A. Sneath, J. T. Staley, and S. T. Williams.
1994. Group 31, the methanogens, 9 ed. The Williams & Wilkins Co., Baltimore.

Honigberg, B. M. 1969. Protozoa associated with termites and their roles in
digestion, p. 1-36. In K. Krishna and F. M. Weesner (ed.), Biology of Termites,
vol. 2. Academic Press, New York.

Kambhampati, 8., K. M. Kjer, and B. L. Thorne. 1996. Phylogenetic
relationship among termite families based on DNA sequence of mitochondrial
16S ribosomal RNA gene. Insect Mol. Biol. 5:229-238.

Kane, M. D., A. Brauman, and J. A. Breznak. 1991. Clostridium mayombei sp.
nov., an H2/C02 acetogenic bacterium flom the gut of the Aflican soil-feeding
termite, Cubitermes speciosus. Arch Microbiol 156:99-104.

Kane, M. D., and J. A. Breznak. 1991. Acetonema longurn gen. nov. sp. nov., an
H,/C02 acetogenic bacterium flom the termite, Pterotermes occidentis. Arch
Microbiol 156:91-98.

Khalil, M. A. K., R. A. Rasmussen, J. R. J. French, and J. A. Holt. 1990. The
influence of termites on atmospheric trace gases: CH4, C02, CHC13, N20, CO,
H2, and light hydrocarbons. J Geophysical Res 95:3619-3634.

Kidder, G. W. 1929. Steblomastix strix, morphology and mitosis. Univ. Calif.
Publ. 2001. 33:109-124.

Krishna, K. 1969. Introduction, p. 1-17, Biology of Termites, vol. 1. Academic
Press, New York.

La Fage, J. P., and W. L. Nutting. 1979. Respiratory gas exchange in the dry-
wood termite, Marginitermes hubbardi (Banks) (Isoptera: Kalotermitidae)‘.
Sociobiology 4:257-267.

Leadbetter, J. R., and J. A. Breznak. 1996. Physiological ecology of
Methanobrevibacter cuticularis sp. nov. and Methanobrevibacter curvatus sp.

nov., isolated flom the hindgut of the termite Reticulitermesflavipes. Appl
Environ Microbiol 62:3620-3631.

Lee, M. J., P. J. Scheurs, A. C. Messer, and A. C. Zinder. 1987. Association of
methanogenic bacteria with flagellated bacteria flom a termite hindgut. Curr
Microbiol 15:337-341.

 

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

48.

146

Leidy, J. 1877. On the intestinal parasites of Termesflavipes. Proc Natl Acad Sci
USA 29: 146-149.

Lovell, C. R., and Y. Hui. 1991. Design and testing of a functional group-
specific DNA probe for the study of natural populations of acetogenic bacteria.
Appl. Environ. Microbiol. 57:2602-2609.

Lovley, D. R., and J. G. Ferry. 1985. Production and consumption of H2 during
growth of Methanosarcina spp. on acetate. Appl Environ Microbiol 49:247-249.

Lovley, D. R., and S. Goodwin. 1988. Hydrogen concentrations as an indicator
of the terminal electron-accepting reactions in aquatic sediments. Geochim
Cosmochim Acta 52:2993-3003.

Messer, A. C., and M. J. Lee. 1989. Effect of chemical treatments on methane
emission by the hindgut microbiota in the termite Zootermopsis angusticollis.
Microb Ecol 18:275-284.

Odelson, D. A., and J. A. Breznak. 1985. Nutrition and growth characteristics of
Trichomitopsis termopsidis, a cellulolytic protozoan flom termites. Appl Environ
Microbiol 49:614-621.

Odelson, D. A., and J. A. Breznak. 1983. Volatile fatty acid production by the
hindgut microbiota of xylophagous termites. Appl Environ Microbiol 45:1602-
1613.

Paster, B. J., F. E. Dewhirst, S. M. Cooke, V. Fussing, L. K. Poulsen, and J.
A. Breznak. 1996. Phylogeny of not-yet-cultivated spirochetes flom termite guts.
Appl. Environ. Microbiol. 62:347-352.

Potrikus, C. J., and J. A. Breznak. 1981. Gut bacteria recycle uric acid nitrogen
in termites: a strategy for nutrient conservation. Proc. Natl. Acad. Sci. USA
78:4601-4605.

Seitz, H.-J., B. Schink, N. Pfennig, and R. Conrad. 1990. Energetics of
syntrophic ethanol oxidation in defined chemostat cocultures. 1. Energy
requirement for H2 production and H2 oxidation. Arch Microbiol 155:82-88.

Slaytor, M. 1993. Cellulose digestion in termites and cockroaches: what role do
the symbionts play? Comp. Biochem. Physiol. 103B:775-784.

Tsunoda, K., W. Ohmura, and T. Yoshimura. 1993. Methane emission by the
termite, Coptotemesformosanus Shiraki (Isoptera : Rhinoterrnitidae) (II) Presence
of methanogenic bacteria and effect of food on methane emission rate. Jpn. J.
Environ. Entomol. 2001. 5:166-174.

49.

50.

51.

52.

53.

54.

55.

147

Tsunoda, K., W. Ohmura, T. Yoshimura, and M. Tokoro. 1993. Methane
emission by the termite, Coptotemesformosanus Shiraki (Isoptera :
Rhinotermitidae) 1. effects of termite caste, population size and volume of test
containers. Wood Res. 79:34-40.

Veivers, P. C., R. Muhlemann, M. Slaytor, R. H. Leuthold, and D. E. Bignell.
1991. Digestion, diet, and polyethism in two fungus-growing termites:

Macrotermes subhyalinus Rambur and M. michaelseni Sjostedt. J. Insect. Physiol.
37 :675-682.

Weesner, F. M. 1969. Termites of the Nearctic region, p. 477-525. In K. Krishna
and F. M. Weesner (ed.), Biology of termites, vol. 2. Academic Press, New York.

Wheeler, G., M. Tokoro, R. Schefl'rahn, and N.-Y. Su. 1996. Comparative
respiration and methane production rates in nearctic termites. J. Insect. Physiol.
42:799-806.

Yamin, M. A. 1979. Flagellata with lower termites. Sociobiology 4: 1-120.

Yoshimura, T., K. Tsunoda, and M. Takahashi. 1992. Distribution of hindgut
protozoa in the hindgut of Coptotemesformosanus Shiraki (Isoptera ;
Rhinotermitidae). Jpn. J. Environ. Entomol. Zool. 4:115-120.

Zimmerman, P. R. 1982. Termites: a potentially large source of atmospheric
methane, carbon dioxide, and molecular hydrogen. Science 218:563-565.