5;; IlllllllllllllllllllfllllllllllllllL 31293 00998 5635 This is to certify that the dissertation entitled Role of Protozoa in the Hindgut Fermentation of Wood-Eating Termites: lfl_V1VO and lg_vitro Studies. presented by David Arthur Odelson has been accepted towards fulfillment of the requirements for Ph.D degreeinMicrobiologyJ Public Health ! MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771 MSU RETURNING MATERIALS: Place in book drop to remove this checkout from LIBRARIES . .—:l—-. your record. FINES W1” be charged if book is returned after the date' stamped below. lf 2:: yrs»- £85 919% ROLE or momma IN THE HINDGUT ramnnou or noon-EATING rmms: a 31179 AND in mg STUDIES BY David Arthur Odelaon A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Microbiology and Public Health 1986 ABSTRACT ROLE OF PROTOZOA IN THE HINDGUT FERHENTATION OF HOOD-EATING TERHITES: _I_q vxvo AND 115 vrrao snmnzs By David Arthur Odelson Studies on the in zigg dynamics of the hindgut fermentation of wood—eating termites revealed that acetate dominated the extracellular pool of volatile fatty acids (VEAs) in 6 species of lower termites (geticulitermes flavipes. Coptotermes formosanus. Incisitermes schwarzi, Prorhinotermes simplex, Schedorhinotermes lamaniances, and Zootermopsis ‘gggusticollis.) as well as the higher termite Nasutitermes corniger. Small amounts of C3 to C5 VFAs were also observed. Estimates of in situ acetogenesis by the hindgut microbiota of‘g, flavipes (20.2 to 43.3 nmol x termite.1 x hr-l) revealed that this activity could support 77 to 1002 of the respiratory requirements of the termite (51.6 to 63.6 nmol of 02 x termite.l x hr-l). This conclusion was buttressed by the presence of acetate in g, flavipes hemolymph as well as by the ability of termite tissues to readily oxidize acetate to CO About 802 of the acetate 2. produced in the hindgut was derived from cellulose C; the remainder was derived from hemicellulose C. Selective removal of microbes from the hindgut indicated that protozoa were primarily responsible for acetogenesis, but that bacteria also functioned in this capacity. David Arthur Odelson _I_n 11133 studies with Trichomitopsis termogidis. an anaerobic, cellulolytic protozoan from Zootermopsis termites. revealed that putatively axenic cultures (see H. Yamin, 1978. J. Protozool. §:535-538) contained a methanogenic consort. However. I. termopsidis could be cured of methanogenic activity by incubation in the presence of bromoethanesulfonate. Growth of 1. termopsidis was markedly improved by substituting heat-killed Bacteroides sp. strain JWZO (a termite gut isolate) for heat-killed rumen bacteria as a source of bacterial cell material, and JWZO was the best of a number of bacteria tested. Growth was further improved by co-cultivation in the presence of Hethanospirillum hunggtii. H. hungatii was the best of a number of H -consuming bacteria tested, and under these conditions CHI. was 2 produced, indicating interspecies hydrogen transfer between the protozoa and 31. hungatii . Crude extracts of I. termopsidis possessed endo-B 1,4-glucanase and cellobiase activity as well as hydrolytic activity against wood, xylan, starch and protein. Cell extracts also hydrolyzed microcrystalline cellulose albeit in limited amounts; suggesting that cells may lack a "complete cellulase" system. Hydrolysis of cellulose displayed an optima of pH 5.0 and 30°C, and glucose was the sole product liberated from this activity. With love To my father who told me "things take time". To my mother who said "be happy". And to my brother who told me I could do this. ii ACKNOHLEDGEMENTS I would like to thank the members of my guidance committee, Dr. R. Brubaker. Dr. J. Jensen, Dr. M. Klug, and Dr. C. Reddy for their assistance and encouragement throughout the course of these studies. My special thanks to Dr. J. Miller who provided additional support during the initial phase of this research. I offer my most sincere appreciation to my parents, Bob. and Sue for their support and advice, and most importantly, I offer them my love. To my friend Larry I present my hand in both eternal friendship and appreciation; someday we may agree on something. To Bob Hespell, who introduced me to research, I express my fondness and my respect. And to Kathy, who endured the wrath of the final phase, I present a note of indebtedness; earth would not be the same without you. Last, but not least, I express a warm and hearty thank you to Dr. John Breznak: my mentor, my friend, and. perhaps one day, my colleague. I hope I can maintain your standards and your enthusiasm for scientific research during my future career. Thanks... and cheers. iii Table of Contents Page List of Tables ........................................... v List of Figures .......................................... vii Introduction ............................................. l Termite Biology ..................................... l Termite Nutrition ................................... 6 Lower Termites ...................................... 7 Higher Termites ..................................... 11 List of References .................................. 14 Article I ................................................ 21 Volatile fatty acid production by the hindgut microbiota of xylophagous termites AttiCle II O...OOOOOOOOOOOOOCOOOOOOOOOOOOOOOIOIOOOOOO0.... 34 Nutrition and growth characteristics of Trichomitopsis termopsidis, a cellulolytic protozoan from termites AttiCle III 00.0.00...IOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 72 Cellulolytic and other polymer hydrolyzing activities of Trichomitopsis termgpsidis, a symbiotic protozoan from termites iv List of Tables Table Page Article I l Extracellular pool size of VEAs in termite hindgut fluid .0...OO0..OOOOOOOOOOOOOOOOOOOOOOOO 26 2 Acetate content of gut homogenates of worker termites 0.0.0.0...00......OOOOOOOOOOOOOOOOOOOO. 27 14 14 3 Origin of C02 from g. flavipes fed C-labelled VFAS 0.0...O..0.0..OOOOOOOOOOOOOOOOOOOOOOOOOO... 28 4 Origin of acetate in hindgut fluid of g, flavipgs termites O...0.......0.0000000000000000000000000 29 5 Possible steps in symbiotic dissimilation of g. f1av1228 0......O0......OOOOOOOOOOCOOOOOOOOOO 31 Article II 1 Effect of medium components on growth of T. termopsidis strain 6057C ........................ 52 2 Growth of I. termopsidis strain 6057C on insoluble palysaCCharides 0.00.0.0...OOOOOOOOOOOOOOOOOOOOOO 53 3 Growth of I. termopsidis strain 6057C on heat-killed cells Of-s-O ventr1CUIi OOOOOOOOOOOOOOOOOOOOO0.0.. 59 Table Page Article III 1 Cellulolytic enzyme activities of Trichomitopsis tempsidis .0.00.0.00.0...OOOOOOOOOOOOOOOOOO0.0. 82 2 Hydrolytic activities of Trichomitopsis tempsidis COO...OOOOOOOOOOCIOOOOOOOOOO000...... 91 vi Figure Page Article I 1 Gas chromatograms of a standard mixture of VEAs and a diluted sample of hindgut fluid from g. flavipes termites OOOOOOOOOOOOIOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 25 2 Mass spectrum of acetic acid from hindgut fluid of . _R-O flav1eesten1tes COOOOOOOOOOOOOOOOOOOOOOOO00...... 26 3 Rates of in situ acetogenesis by the hindgut microbiota and O2 consumption by intact worker larvae of.§. flav12e8 OOOOOOIOOOOOOO00.0.00...OOOOOOOOOOOOOOOOOOOOO 27 4 Effect of defaunation or removal of hindgut bacteria on acetogenic activity of hindguts of g. flavipes termites 0.00.0...OOOOOOOOOOOOOOOOOOOOOOOCOOOOOOOOOOOO 28 14 5 Evolution of CO2 by g. flavipes termites feeding on (U-I‘C) acetate, (1-14C) propionate. or (2-14C) butyrate 0.000000000000000...0.0000000000000000IOOOOOO 28 6 Proposed working model for symbiotic wood polysaccharide dissimilation in 5. flavipes ......................... 32 Article II 1 Growth of I. termopsidis 6057 in modified Yamin's medium List of Figures at 25.C 0..OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO. ‘5 vii Figure Page Article II continued 2 Growth of.I. termopsidis 6057 in modified Yamin's medium with heat-killed cells of Bacteroides sp. JUZO in place Of heat-killed mun bBCteria eeeeeeeeaoeaeeeeeeeeeeoe 48 3 Growth of'I. termopsidis 6057 in co-culture with Hethanospirillum hungatii ............................ 55 4 Phase contrast photomicrograph of a cell of.T. termopsidis 6057C in the presence of heat-killed.§. ventriculi .0.0....O...O000......OOOOOOOOOOOOOOOOOOOOO 60 Article III 1 Effect of pH on the hydrolysis of microcrystalline cellulose at 30°C by soluble extracts of I. termopsidis 84 2 Effect of temperature on the hydrolysis of micro- crystalline cellulose at pH 5.0 by soluble extracts of I. tempsidis O0.000000000000000000000000......COO... 86 3 Production of glucose and reducing sugar from micro- crystalline cellulose by soluble extracts of I. temp81d18 O...00....I.00....OOOOOOOOOOOOOOOOOOOOOOOO 89 viii INTRODUCTION TERMITB BIOLOGY Termites are social insects belonging to the order Isoptera, and are represented by about 2000 different living and fossilized species (48). They are ubiquitous in nature, and can be found on approximately two-thirds of the world's land surface (80). All termites examined to date contain a dense population of microorganisms in their hindgut. In the phylogenetically "lower" termites (families Hasto-, Kato-, Hodo-, Rhino-, and Serritermitidae) the microbiota includes unique genera and species of anaerobic protozoa, as well as bacteria. In the "higher" termites (family Termitidae), hindgut microbial populations consist almost exclusively of bacteria, although protozoa have been observed in a few species (39). Although often noted for their economic damage and cost of control (>$750 Million/yr in the 0.8. (56)), termites are ecologically important in decomposition processes and nutrient cycling (80), and possibly in the global production of methane (87). A termite society or colony is comprised of several castes, including workers, soldiers, and reproductives, which are morphologically and functionally different (27). In fact, the termite colony has been used to illustrate evolutionary principles by drawing analogies between the termite society, as a fundamental unit, and an individual organism (48). It is the worker larvae which dominate the colony's nutritive activities in both sheer numbers and in responsibilities, feeding nutrient-dependent castes, such as soldiers, l newly hatched larvae, and the queen. In both higher and lower termites nutrient dependent castes solicit stomodeal food (a mixture of saliva and regurgitated raw food) from workers (60). Lower termites also rely on solicitation of proctodeal food (liquid excretions from the hindgut) from workers, which also serves as a means of transferring the hindgut microbiota to newly hatched larvae, as well as reinfecting recently molted larvae. Additionally, workers are responsible for building the nest and foraging for food. For these reasons most biochemical and microbiological studies of termites have been carried out on the worker caste. Although frequently cited for their ability to utilize wood as a food source, the diet of termites (depending on the species) can consist of fungal decayed wood, soil, humus, leaves, or even dung (50). while these materials are quite diverse, they are generally rich in cellulose, hemicellulose, and lignin and relatively poor in soluble substrates, including combined nitrogen. Both cellulose and hemicellulose are efficiently dissimilated by termites. Digestion and assimilation (65-992 and 54-931, respectively) of these polymers are quite high in comparison to other invertebrates (28,79). The extent of lignin decomposition (0-832) varies widely among species and has been discussed, as well as questioned, by several authors (10.50.61.73). The most convincing evidence for lignin degradation by termites was obtained with the higher termite Nasutitermes exitiosus (13), as determined by ll‘COZ evolution from ll'C--lignin. Unfortunately, it is not known whether the termite itself, or its gut microorganisms, or both mediate this activity. The alimentary tract of termites is comprised of three main divisions: the foregut, including the crap and gizzard; the midgut; and the hindgut, which includes the paunch, colon, and rectum (60). Both the crop and the gizzard, as well as the mandibles (i.e. the mouth parts) are structurally conserved throughout the entire order of termites (60), suggesting that they are well suited to the needs of the animal. Some higher termites also have a mixed segment at the juncture of the mid- and hindgut, a section bounded on one side of the lumen by midgut tissue and on the other side by hindgut tissue. The hindgut of most termites has a pH of 6-7.S and is anaerobic (Eo'--230 to -270 mV) (61) while the foregut and midgut are both aerobic (Eo' > 100 mV) and found to have a pH of 2.0 to 6.8 and 5.5 to 8.0, respectively. In most insects the main site of nutrient absorption is the midgut (60), and it is likely that termites also depend on this activity to absorb soluble material from ingested food and/or stomodeal and proctodeal solicitations. However, the bulk of the microbiota, in both higher and lower termites, is contained in the bulbous paunch, a region analagous to an anaerobic fermentation chamber. Mereover, Kovoor (44), using termites fed BaSO4 and detecting its location by using x-rays, has shown that in the higher termite Hicrocerotermes edentatus particulate material takes about 24 hr to move completely through the gut; transit thru the fore- and midgut was rapid (4 hr) and the material was retained the longest in the paunch region. Inasmuch as the hindgut also has absorptive features (60), and both higher (45) and lower (42) termites contain acetate and other volatile fatty acids in this region it is not unreasonable to suggest that termites may represent an exception to the rule, utilizing the hindgut as the main site of nutrient absorption. TERHITE NUTRITION As mentioned above cellulose and hemicellulose are both efficiently dissimilated by termites, and represent the primary, ultimate carbon and energy source of these insects. However, very little is known about the mechanism of hemicellulose degradation in termites. Host researchers have focussed their attention on the hydrolysis of cellulose, inasmuch as this polymer comprises the bulk (about 502) of the dry weight of wood (28). As one might expect, the principle question that has been addressed by investigators has been the contribution of specific gut microorganisms, versus the termite itself, to the overall digestion of cellulose. At this point a brief discussion is warranted on the structure of cellulose, as well as the enzymatic hydrolysis of this polymer in nature. Native cellulose is water insoluble and its structure is complex, containing both amorphous and crystalline regions (25,31,34). In wood this structure is further complicated by association with other polymers such as lignin and hemicellulose (25). Cellulase activity is responsible for recycling of much of the carbon fixed annually into biomass (34). 'lgqgiggg the rate of hydrolysis of cellulose appears to be related to its degree of crystallinity as well as its association with lignin (31). Decreasing the crystallinity index (31), by mechanical grinding or by ball-milling, or extracting the lignin, increases the rate of enzymatic hydrolysis. Most work on cellulase activity has focused on the complete cellulase system of fungi, and, to a lesser extent, on bacterial cellulase. At least three major enzymes are associated with the fungal cellulase system: 1,4-B-D-glucan- 4-glucanohydrolase (Cx cellulase or endoglucanase, EC 3.2.1.4), 1,4-B-D—glucan cellobiohydrolase (Cl cellulase or cellobiohydrolase, EC 3.2.1.91), and B-glucosidase (cellobiase, EC 3.2.1.21). A recent review by Erikkson (26) details the biological mechanisms of cellulose hydrolysis. It is important to note that the efficient and complete hydrolysis of cellulose is dependent on the combined and synergistic action of all 3 enzymes. If any one of these enzymes is missing digestion of native cellulose will be severely limited. To demonstrate the "complete cellulase" system (i.e. C1 cellulase, endoglucanase, and cellobiase) in 33552, workers often use insoluble substrates, such as filter paper, cotton fibers or microcrystalline cel- lulose, and measure the production of reducing sugar. However, it is ‘225 the rate of reducing sugar release from these insoluble substrates that is important, but the extent of hydrolysis (i.e. the percent hydro- lyzed). As mentioned above, cellulose is a mixture of amorphous and crystalline regions, and it is the latter regions which are most resis- tant to enzymatic attack. Ghose gt a; (34) point out that at least 42 (reducing sugar equivalent) of a 50 mg piece of filter paper must be hydrolyzed to show the presence of a "complete cellulase" system. C1 cellulase activity can be specifically measured by quantitating the release of cellobiase from crystalline cellulose. Endoglucanase is deménstrated by using carboxymethyl cellulose, a soluble derivative of cellulose, and determining the rate of reducing sugar release. Cello- biase is assayed by measuring the production of glucose from cellobiase, although some worker utilize p-nitrophenylglucoside for measuring this activity (i.e. B-glucosidase). The reader is referred to an excellent review by Chose gt a; (34) for a more detailed description of methodologies. ‘ rt" CW"; 9' . Detailed knowledge of the mechanisms of digestion of cellulose by termites has been hampered by a number of factors, including the lack of an available system to rear germ-free termites, as well as the inability to cultivate many of the gut microorganisms lgflgiggg. Nevertheless, information which has accumulated over the years has provided some insight into the biochemistry of symbiosis between the termites and their microbiota, and has been recently reviewed by Breznak (9-11) and O'Brien and Slaytor (61). 1. Lower Termites One of the most frequently cite examples of a biological mutualism is the interaction between lower termites and their cellulolytic hindgut protozoa which enable the insects to thrive on a diet of wood. The obligatory dependence of lower termites upon the hindgut protozoa was first recognized by Cleveland (14-20). Cleveland showed that defaunation (i.e. removal of protozoa) of termites could be achieved by incubating the insects at elevated temperatures (14) or under hyperbaric oxygen (19), and that such termites continued to feed on wood, but starved to death. He also showed that defaunation by either of these methods did not harm the termite itself; defaunated termites could be "reinfected" with their normal protozoan symbionts and would survive on a diet of wood. A similar mmtalistic relationship was also shown for the wood-feeding cockroach Cryptocercus punctulatus (22). Trager (75) was the first to show that extracts of hindgut contents of Zootermopsis termites could degrade crystalline cellulose to glucose, and that little or no cellulase activity was found in other parts of the digestive tract. Hungate (40) confirmed and extended this work by showing that cellulase activity was absent in the hindgut of defaunted termites. More recent work with Mastotermes darwiniensis (77) and Coptotermes lacteus (62) has also found that the hydrolysis of cellulose was principally associated with the hindgut. As hindgut contents contain both protozoa and bacteria, the definitive role of the protozoa in wood digestion could not be ascertained until the protozoa were obtained in axenic culture. This major achievement was recently accomplished by Yamin (92,83) who expanded the work of Trager (76) and successfully axenically cultivated 2 anaerobic, cellulolytic protozoa from Zootermopsis sp. termites. One of the isolates, Trichomitopsis termopsidis, was shown to require cellulose for growth; starch, glucose, or a variety of other substrates would not support growth. If defaunted Zootermopsis termites (fed a diet of cellulose) were reinoculated with I. termopsidis, the insects survived as well as those insects reinfected with hindgut contents of normal termites, and significantly longer than non-reinfected controls (85), implying _i__n_ gig cellulose hydrolysis by T. termopsidis. Additionally, extracts of I. termopsidis contained both endoglucanase and B-glucosidase activity. Other workers (49,57,58,62,78,81,86) have also presented evidence that the termite itself secretes cellulase components (endoglucanase and cellobiase). Veivers e_t g_l_ (78) found that in g; darwiniensis endoglucanase and cellobiase activity was principally associated with the salivary glands and midgut of the termite. Nevertheless, defaunated E. darwiniensis could not survive on a diet of wood. These workers also found that g. darwiniensis could be defaunated by feeding termites starch, but survive as well as normal termites that were fed wood. The ability to survive on starch was attributed to termite enzymes and hindgut bacteria. However, g. darwiniensis may be a unique example of a lower termite; aside from wood these termites are also known to feed on sugar cane and vegetables (38). The end-products of cellulose dissimilation _ig 333.2 were first determined by Hungate (41,42). Suspensions of hindgut contents from Zootermopsis termites were mixed with powdered cellulose and found to produce principally acetate, carbon dioxide, and hydrogen. By extrapolating the igflziggg rates of hydrogen production to hydrogen production by the intact termite he proposed that acetate served as the major oxidizable energy source for termites. However, attempts to rear defaunated termites on a diet of acetate (24,43) were unsuccessful. Hungate also qualitatively demonstrated acetate in hindgut fluid obtained directly from the termite, and showed that acetate could transverse the hindgut tissue. This work was recently buttressed by Yamin (83.84) who showed that the major products of cellulose hydrolysis by axenic cultures of protozoa were hydrogen, carbon dioxide and acetate. Although no quantitative in situ studies have been done on the end-products of wood dissimilation in lower (as well as higher) termites, acetate has been implicated as a major end product inasmuch as various studies have shown ll‘C-acetate fed to termites is readily incorporated into termite long chain fatty acids (8,55), terpenes (70), and cuticular hydrocarbons (7). ‘Hhile bacterial populations are quite high (109 -1010/ml termite) (11 8 references therein) in the hindgut of lower termites there is currently no compelling evidence that bacteria are quantitatively significant in cellulose hydrolysis. MicroscOpic studies have revealed that the bacterial papulation is morphologically diverse and predominantly found in close association with the paunch epithelium. Bacterial isolates from the hindgut are principally facultative or strict anaerobes and include Streptococcus sp., Bacteroides sp., and Enterobacteriaceae. Various activities have been ascribed to the bacterial component of the paunch region including nitrogen fixation (64), uric acid nitrogen cycling (65), which appears to be an important 10 symbiotic strategy for nitrogen conservation, and methane production (9,11). Bacteria are also found as endo- and ectosymbionts of hindgut protozoa (see ref. 12), and, in some cases, they provide motility for the eukaryotes (74). 11 II. Higher Termites Considering that higher termites constitute about two-thirds of the total known species of termites, it is ironic that less is known about cellulose digestion in this group of termites than in lower termites. High termites may secrete their own cellulase, although it is not known if, in general, they possess the full complement of enzymes required to digest native cellulose. Potts and Hewitt (66) found that about 702 of the endoglucanase activity of Trinervitermes trinervoides was associated with the midgut; about 1/2 of this activity was localized within the gut tissue, an area devoid of microorganisms. Subsequent purification of this enzyme also revealed activity against microcrystalline cellulose, but not to the extent one might expect with a "complete cellulase" system (67). Hicrocerotermes edentatus also contained most of its endoglucanase activity in the midgut (47). O'Brien gt gl_(62) found that in Nasutitermes exitiosus about 802 of the total endoglucanase activity of the termite gut was associated with the fore- and midgut. These workers further showed that removal of the hindgut bacteria, by feeding termites tetracycline, did not alter this activity. Cellobiase activity was also predominantly associated with the midgut of N. exitiosus, as well as in N. walkeri (57). Other workers have found cellobiase activity either mostly associated with the hindgut (in 355955 (Cyclotermes) obesus, 59) or evenly distributed between the mid- and hindgut (Trinervitermes trinervoides, 66). One of the most informative studies on cellulase activity in higher termites was done with the fungus growing termite, Macrotermes natalensis (52,53). 5. natalensis maintains fungal colonies in their 12 nests on structures derived from chewed, but undigested, plant material (23), and the termites eat both fungal material (i.e. mycelia and conidia) (2) and plant fragments (35). These termites can not survive if the fungal material is first sterilized. Martin and Martin (52,53) showed that in M. natalensis microcrystalline cellulose hydrolysis, endoglucanase, and B-glycosidase activity was predominately associated with midgut homogenates. Isoelectric focusing of these enzyme activities further showed that all of the cellulose hydrolysis activity and part of the endoglucanase activity was acquired from consumption of fungal conidia. B-glucosidase activity appeared to originate from salivary glands, midgut tissue, and fungal conidia. Abu-Khatwa (1) arrived at similar conclusions with studies of M. subhyalinus. Higher termites generally feed on food material that is in active decay (21), and it is not unreasonable to assume they might also ingest some cellulolytic enzymes as well. As pointed out by Martin (54) the importance of acquired enzymes (i.e. ingested with the food) to higher termite nutrition requires further investigations. The end-products of wood dissimilation in higher termites have as yet not been quantitated, although in Microcerotermes edantatus (45) a qualitative examination revealed the presence of acetate, propionate, and butyrate. Although it appears hindgut bacteria may not play an initial role in the digestion of cellulose by higher termites their presence is critical for the termite's survival. Eutick gt 5; (29) showed that feeding N. exitiosus termites the antibiotic metronidazole was lethal to the termite. Limited microbiological studies have shown the hindgut to contain morphologically diverse bacteria, including spirochetes l3 (3-6,63,66). In a detailed microscopic study Bignell‘gtugl (3-6) found that in the higher termites Procubitermes aburiensis and Cubitermes severus bacteria are also contained within the crop, midgut, and mixed segment. Euticklggugl (30) found Staphlococcus sp. to be the predominate isolate (107 bacteria/m1 x termite) from Nasutitermes sp. termites. Pseudomonas sp. (31) and Achromobacter sp. (32) have also been isolated from Nasutitermes, although both studies failed to quantitate the relative numbers of these isolates. One important role of bacteria in higher termite nutrition may be in nitrogen economy (10). Nitrogen fixation, a prokaryotic phenomenon, has been demonstrated by acetylene reduction (37) in Nasutitermes sp. termites (33.68.69) and in Macrotermes ukuzii and Trinervitermes trinervoides (71). French‘s; a; (33) isolated nitrogen fixing Citrobacter sp. from N. exitiosus, however, the bacteria were not quantitated. Unfortunately no other attempts have been made to isolate the causative agent of the nitrogen fixation activity in higher termites. 10. ll. 12. LIST OF REFERENCES Abu-Khatwa, N. 1978. Cellulase of fungus-growing termites: A new hypothesis on its origin. Experientia.'§4:559-560. Batra, S.N. 1975. Termites (Isoptera) eat and manipulate symbiotic fungi. J. Kansas Entomol. Soc. '4§:89-92. Bignell, D.E., H. Oskarsson, J.M. Anderson. 1979. Association of actinomycete-like bacteria with soil feeding termites. (Termitidae, Termitinae). Appl. Environ. Microbiol. ‘§Z:339-342. Bignell, D.E., H. Oskarsson, and J.M. Anderson. 1980. Distribution and abundance of bacteria in the gut of a soil-feeding termite Procubitermes aburiensis (Termitidae, Termiinae). J. Gen. Microbiol. .1l1:393-403. Bignell, D.E., H. Oskarsson, and J.M. Anderson. 1980. Specialization of the hindgut wall for the attachment of symbiotic ‘microorganisms in a termite Procubitermes aburiensis (Isoptera, Termitidae, Termitinae) Zoomorphology. .26:103-112. Bignell, D.E., H. Oskarsson, and J.M. Anderson. 1980. Colonization of the peritrophic membrane and the ectoperitrophic space by actinomycetes in a soil-feeding termite. J. Invert. Path. 36:426-8. Blomquist, G.J., R.N. Howard, and C.A. McDaniel. 1979. Biosynthesis of the cuticular hydrocarbons of the termite Zootermopsis angusticollis (Hagen). Incorporation of propionate into dimethylalkanes. Insect. Biochem. .2:371-374. Blomquist, G.J., L.A. Dwyer, A.J. Dwyer, A.J. Chu, R.O. Ryan, and M. de Reobales. 1982. Biosynthesis of linoleic acid in a termite, cockroach, and cricket. Insect Biochem.‘lg:349-353. Breznak, J.A. 1975. Symbiotic relationships between termites and their intestinal microbiota. p. 559-580. In D.H. Jennings and D.H. Lee (eds.), Symbiosis (29th Symposium of the Society for Experimental Biology), 213559-580. Breznak, J.A. 1982. Intestinal microbiota of termites and other xylophagous insects. Ann. Rev. Microbiol. .36:323-343. Breznak, J.A. 1983. Biochemical aspects of symbiosis between termites and their intestinal microbiota. p. 173-203, In J.M. Anderson, A.D.M. Rayner, and D. Nalton (eds.), Invertebrate-Microbial Interactions, Cambridge University Press, Cambridge, England. Breznak, J.A. 1984. Hindgut spirochetes of termites and Cryptocercus punctulatus, p. 67-70, In N.R. Krieg and J.G. Holt (ed.), Bergey's Manual of Systematic Bacteriology, Williams and Hilkins, Baltimore, MD. 14 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 15 Butler, J.H.A., and J.C. Buckerfield. 1979. Digestion of lignin by termites. Soil Biol. Biochem. .LL:507-511. Cleveland, L.R. 1923. Symbiosis between termites and their intestinal protozoa. Proc. Natl. Acad. Sci. U.S.A. .2:424-428. Cleveland, L.R. 1924. The physiological and symbiotic relationship between the intestinal protozoa of termites and their host, with special reference to Reticulitermes flavipes Kollar. Biol. Bull. ‘4§:l77-225. Cleveland, L.R. 1925. The method by which Trichonympha campanula, a protozoan in the intestine of termites, ingests solid particles of wood for food. Biol. Bull. .4§:282-289. Cleveland, L.R. 1925. The ability of termites perhaps to live indefinitely on a diet of pure cellulose. Biol. Bull. 148:289-293. Cleveland, L.R. 1925. The feeding habit of termite castes and its relation to the intestinal flagellates. Biol. Bull. .§§:295-306. Cleveland, L.R. 1925. The effects of oxygenation and starvation on the symbiosis between the termite, Termopsis, and its intestinal flagellates. Biol. Bull. 48:309-327. Cleveland, L.R. 1925. Toxicity of oxygen for protozoalin.vivo and 'ig_vitro: animals defaunated without injury. Bio. Bull. 48:455-468. Cleveland, L.R. 1923. Correlation between food and morphology of termites and the presence of intestinal protozoa. Amer. J. Hygiene. .3:444-46l. Cleveland, L.R., S.R. Hall, E.P. Sanders, and J. Collier. 1934. The wood-feeding roach Cryptocercus, its protozoa, and the symbiosis between protozoa and the roach. Memoirs of the Amer. Acad. of Arts and Sci. ‘12:185-342. Cmelik, S.H., and C.C. Douglas. 1970. Chemical composition of ”fungus gardens" from two species of termites. Comp. Biochem. Physiol. “36:493-502. Cook, S.F. 1943. Nonsymbiotic utilization of carbohydrate by the termite Zootermopsis aggusticollis. Physiol. 2001. {16:123-8. Cowling, E.B., and T.R. Kirk. 1976. Properties of cellulose and lignocellulosic materials as substrates for enzymatic conversion processes. Biotechnol. Bioeng. Symp. .QF95‘135- Erikksson, K.E. 1981. Cellulases of fungi. p. 19-32. In A. Hollender (ed.), Trends in the Biology of Fermentations for Fuels and Chemicals. Plenum Press. New York, NY. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 16 Esenther, G.R. 1969. Termites in Wisconsin. Ann. Entomol. Sac. Amer.'§2:1274-1284. Esenther, G.R., and T.K. Kirk. 1974. Catabolism of Aspen sapwood in Reticulitermes flavipes (Isaptera:Rhinotermitidae). An. Entomol. Soc. Amer. .61:989-991. Eutick, M.L., P.C. Veiver, N.R. O'Brien, and M. Slaytor. 1978. Dependence of the higher termite Nasutitermes exitiosus and the lower termite Captatermes lacteus on their gut flora. J. Insect Physiol. '24:363-8. Eutick, M5L., R.N. O'Brien, and M. Slaytor. 1978. Bacteria from the gut of Australian termites. Appl. Environ. Microbial. 35:823-8. Fan, L.T., Y. Lee, and D.H. Beardmare. 1980. Mechanism of the enzymatic hydrolysis of cellulose: effects of major structural features of cellulose an enzymatic hydrolysis. Biotech. Bioeng. 22:177-199. French, J.R.J. 1975. The role of termite hindgut bacteria in wood decomposition. Mater. Org. 12:1-13. French, J.H.J., G.L. Turner, and J.F. Bradbury. 1976. Nitrogen fixation by bacteria from the hindguts of termites. J. Gen. Micro. .2§:202-206. Chase, T., B.S. Mountenecourt, and D.H. Eveliegh. 1981. Measure of cellulase activity. IUPAC, Commission on Biotechnology. Crasse, P.P., and C. Nairat. 1957. Le significant meules a champignons des Macratermitinae (Ins., Isaptera). Campt. Rend. Hebdimadaires Seances Acad. Sci. 244:1845-50. Grasse, P.P., and C. Nairat. 1959. L'evalution de la symbiase chez les Isapteres. Experientia. 12:365-372. Hardy, R.D.F., R.C. Burns, and R.D. Halstan. 1973. Applications of the acetylene-ethylene assay for measurement of nitrogen fixation. Sail Biol. Biochem. ‘§:47-81. Hill, G.F. 1942. Termites from the Australian region. Council for Scientific and Industrial Research. Melbourne, Australia. Honigberg, B.M. 1970. Protozoa associated with termites and their role in digestion. p. 1-36. In K. Krishna and F.M; Heesner (eds.), Biology of Termites, II, Academic Press, New York, NY. Hungate, R.E. 1938. Studies on the nutrition of Zaatermapsis. II. The relative importance of the termite and the protozoa in wood digestion. Ecol. 12:1-25. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 17 Hungate, R.E. 1939. Experiments on the nutrition of Zootermopsis. III. The anerabic carbohydrate dissimilation by the intestinal protozoa. Ecol.‘2Q:230-245. Hungate, R.E. 1943. Quantitative analysis an the cellulose fermentation by termite protozoa. Ann. Entomol. Soc. Amer. 26: 730-739. Hungate, R.E. 1946. The symbiotic utilization of cellulose. J. Elisha Mitchell Sci. Soc. 22:9-24. Kavaar, J. 1967. Etude radiagraphique du transit intestinal chez un termite superieur. Experientia. ‘2§:820-1. Kovaor, J. 1967. Presence d'acides gras volatiles dans la panse d'un termite superieur (Microcerotermes edentatus was., Amitermitidae) C.R. Hebd. Seances Acad. Sci. Ser. 264:486-488. Kovoar, J. 1968. L'intestin d'un termite superieur (Microcerotermes edentatus, Nas., Amitermitidae). Histophysiologie et flare bacterienne symbiotique. Bull. Biol. Fr. Belg. .222:45-84. Kavaor, J. 1970. Presence d'enzymes cellulytiques dans l'intestin d'un termite superieur, Microcerotermes edentatus (was.). An.. Krishna, K. 1969. Introduction. p.1-17. In K. Krishna and F.M. Heesner (eds.), Biology of Termites, VI, Academic Press, New York, NY. Lafage, J.P., and W.L. Nutting. 1978. Nutrient dynamics in termites. p. 165-232. In M.V. Brian (ed.), Production Ecology of Ants and Termites, Cambridge University Press, Cambridge, England. Lee, K.E., and T.G. Hood. 1971. Termites and Soils. Academic Press, Inc. New York, NY. Mannesmann, R. 1977. Untersuchungen des glucose-staffwechsels van Carkammer bakterien der termite Nasutitermes nigriceps (Haldemann). .Martin, M.M., and J.S. Martin. 1978. Cellulose digestion in the midgut of the fungus-growing termite Macrotermes nasalensis: The role of acquired digestive enzymes. Science 199:1453-5. 'Martin, M.M., and J.S. Martin. 1979. The distribution and origin of the cellulolytic enzymes of the higher termite, Mastotermes natalensis. Physiol. 2001. 22:11-21. Martin, M. 1983. The role of ingested enzymes in the digestive processes of insects. In J.M. Anderson, A.D.M. Rayner, and D. Walton (eds.), Invertebrate-Microbial Interactions. Cambridge University Press, Cambridge, England. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 18 {Mauldin, J.K. 1982. Lipid synthesis from ll'C-acetate by two subterranean termites. Reticulitermes flavipes and Coptatermes formosanus. Insect Biochem. ‘22:193-199. iMauldin, J.K. 1982. The economic important of termites in N. America. p. 138-141. In M.D. Boyd, C.D. Michena, and H.E. Evans (eds.), The Biology of Social Insects, Nestview Press, Boulder, CO. McEwen, S.E., M; Slaytor, and R.N. O'Brien. 1980. Cellobiase activity in 3 species of Australian termites. Insect Biochem. 22: 563-567 . Mishra, 8.6. 1980. Carbohydrases in Neotermes basei Snyder (Isaptera:Kalatermitidae). Mater. Org. ‘22:253-261. Misra, J.N., and V. Ranganathan. 1954. Digestion of cellulose by the mound building termite Termes (cyclotermes) obesus (Rambor). Proc. Ind. Acad. Sci. (Serb). ‘22:100-113. Noirot, C., and C. Noirot-Timothee. 1969. The digestive system. p. 49-88. In K. Krishna and E.M. Neesner (ed.), Biology of Termites, VI, Academic Press, New York, NY. O'Brien, R.N., and M. Slaytor. 1982. Role of microorganisms in the metabolism of termites. Aust. J. Biol. Sci. 22:239-262. O'Brien, S.H., P.C. Veivers, S.E. MeEwen, M. Slaytor, and R.N. O'Brien. 1979. The origin and distribution of cellulase in the termites, Nasutitermes exitiosus and Coptatermes lacteus. Insect Biochem. .2:6l9-625. Pochon, J., H. deBarjac, and A. Roche. 1959. Recherches sur la digestion de la cellulose chez le termite Sphaerotermes sphaerotharax. Ann. Inst. Pasteur. Paris. ‘2§:352-5. Potrikus, G.J., and J.A. Breznak. 1977. Nitrogen-fixing Enterabacter agglomerans isolated from guts of wood-eating termites. Appl. Environ. Microbiol. 22:392-399. Potrikus, C.J., and J.A., Breznak. 1981. Cut bacteria recycle uric acid nitrogen in termites: A strategy for nutrient conservation. Proc. Natl. Acad. Sci. U.S.A. ‘12:4601-5. Potts, P.C., and P.H. Hewitt. 1973. The distribution of intestinal bacteria and cellulase activity in the harvester termite Trinervitermes trinervoides (Nasutitermitinae). Insect Sociaux. 32:215-220. Potts, R.C., and P.H. Hewitt. 1974. The partial purification of the cellulase from the termite Trinervitermes trinervoides (Nasutitermitinae). Comp. Biochem. Biophys. 478:317-326. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 19 Prestwich, C.D., B.L. Bentley, and E.J. Carpenter. 1980. Nitrogen sources for neotrapical nasute termites: fixation and selective foraging. Oecologia. .§§:397-401. Prestwich, C.D., and B.L. Bentley. 1981. Nitrogen fixation by intact colonies of the termite Nasutitermes corniger. Oecologia. 52: 249-251 . Prestwich, C.D., R.N. Jones, and M.S. Collins. 1981. Terpene biosynthesis by nasute termite soldiers (Isoptera:Nasutitermitinae). Insect Biochem. I21:331-336. Rohrmann, G.F., and A. Rossman. 1980. Nutrient strategies of Macratermes ukuzii. Pedobiologia. .2Q:60-73. Schultz, J.E., and J.A. Breznak. 1978. Heterotrophic bacteria present in hindguts of wood-eating termites (Reticulitermes flavipes (Kollar)). Appl. Environ. Microbiol. '22:930-936. Seifert, K., and G. Becker. 1965. Der chemische abbau van laub-und nadelholzarten durch verschriedene termiten. Halzfor schung. ‘22:105-111. Tamm, S.L. 1982. Flagellated ectosymbiotic bacteria propel a eucaryatic cell. J. Cell. Biol. 23:697-709. Trager, N. 1932. A cellulase from the symbiotic intestinal flagellates of termites and of the roach Cryptocercus punctulatus. Trager, N. 1934. The cultivation of a cellulose-digesting flagellate, Trichomonas termopsidis and of certain other termite protozoa. Biol. Bull. ‘2§:182-190. Veivers, P.C., A.M. Musca, R.W. O'Brien, and M. Slaytor. 1982. Digestive enzymes of the salivary glands and guts of Mastotermes darwiniensis. Insect Biochem. .22:35-40. Veivers, P.C., R.H. O'Brien, and M. Slaytor. 1983. Selective defaunation of Mastotermes darwiniensis and its effect on cellulose and starch metabolism. Insect Biochem. .22:95-101. Hood, T.G. 1978. Food and feeding habits of termites. p. 55-80. In M.V. Brian (ed.), Production Ecology of Ants and Termites, Cambridge University Press, Cambridge, England. ‘Nbod, T.G., and H. Sand. 1978. The role of termites in ecosystems. p. 245-292. In M.V. Brian (ed.), Production Ecology of Ants and Termites. Cambridge University Press, Cambridge, England. 81. 82. 83. 84. 85. 86. 87. 20 Yamaoka, I., and Y. Nagatani. 1975. Cellulose digestion system in the termite, Reticulitermes speratus (Kolbe). I. Producing sites and physiological significance of two kinds of cellulase in the worker. Zool. Mag. 83323-9. Yamin, M.A. 1978. Axenic cultivation of the cellulolytic flagellate Trichomitopsis termopsidis (Cleveland) from the termite Zootermopsis. J. Protozool. 3§;535-538. Yamin,‘M.A. 1980. Cellulose metabolism by the flagellate Trichonympha from a termite is independent of endosymbiotic bacteria. Science 211:58-9. Yamin,lM.A. 1980. Cellulose metabolism.by the termite flagellate Trichomitopsis termopsidis. Appl. Environ. Microbiol. ‘22;859-863. Yandn, M.A., and W. Trager. 1979. Cellulolytic activity of an axenically cultivated termite flagellate, Trichomitopsis termopsidis. J. Gen. Microbiol. 113:417-420. Yokoe, Y. 1964. Cellulase activity in the termite, Leucotermes speratus, with new evidence in support of a cellulase produced by the termite itself. Scientific Papers of the College of General Education, U. Tokyo. ‘lfl:115-120. Zimmerman, P.R., J.P. Greenberg, 5.0. wandigo, and P.J. Crutzen. 1982. Termites: a potentially large source of atmospheric methane, carbon dioxide, and molecular hydrogen. Science. ‘lg§;563-565. ARTICLE I VOLATILE FATTY ACID PRODUCTION BY THE HINDGUT MICROBIOTA OF XYLOPHAGOUS TERMITES By D.A. Odelson and J.A. Breznak 21 22 Annex: awn Euvtaouuemnt Dirac-tower. May I”). p. 1602-1613 Vol. 45. No. 5 mnmsmtsoz-tzsoz.m Volatile Fatty Acid Production by the Hindgut Microbiota of Xylophagous Termitesi DAVID A. ODELSON AND JOHN A. BREZNAK’ Department of Microbiology and Public Health, Michigan State University, East Lansing. Michigan 48824 Received 31 January 1%3/Accepted 2 March 1983 Acetate dominated the extracellular pool of volatile fatty acids (VFAs) in the hindgut fluid of Reticuiitennes flavipes, Zootermopsis angusticollis, and Incisi- termes schwarzi, where it occurred at concentrations of 57.9 to 80.6 mM and accounted for 94 to 98 mol% of all VFAs. Small amounts of C3 to C5 VFAs were also observed. Acetate' was also the major VFA in hindgut homogenates of Schedorhinotermes lamanianus, Prorhinotermes simplex, Coptatermes forma- sanus, and Nasutitermes corniger. Estimates of in situ acetogenesis by the hindgut microbiota of R. flavipes (20.2 to 43.3 nmol - termite" ' h") revealed that this activity could support 77 to 100% of the respiratory requirements of the termite (51.6 to 63.6 nmol of 02 - termite" - h"). This conclusion was buttressed by the demonstration of acetate in R. flavipes hemolymph (at 9.0 to 11.6 mM), but not in feces, and by the ability of termite tissues to readily oxidize acetate to C0,. About 85% of the acetate produced by the hindgut microbiota was derived from cellulose C ; the remainder was derived from hemicellulose C. Selective removal of major groups of microbes from the hindgut of R. flavipes indicated that protozoa were primarily responsible for acetogenesis but that bacteria also functioned in this capacity. Hi and CH, were evolved by R. flavipes (usually about 0.4 nmol - termite" - h' ), but these compounds represented a minor fate of electrons derived from wood dissimilation within R. flavipes. A working model is proposed for symbiotic wood polysaccharide degradation in R. flavipes, and the possible roles of individual gut microbes, including Correducing acetogenic bacteria, are discussed. A classical example of nutritional symbiosis is that which occurs between phylogenetically “lower“ termites (families Masto-, K3103 Hodo-, and Rhinotermitidae) and their intestinal microbiota, an interaction that enables such termites to thrive by xylophagy. Our under- standing of this symbiosis has been presented in several reviews (8—10, 20, 23, 30, 38). To sum- marize. the carbon and energy nutrition of lower termites is centered on wood polysaccharides (cellulose and hemicelluloses). which constitute about 70% of the dry weight of wood and which undergo up to 99% degradation on passage through the gut of the insect. Most of this degradation occurs in the hindgut, a region anal- ogous to an anaerobic fermentation chamber, and the hindgut microbiota appears to be the driving force of dissimilatory activity. The hind- gut microbiota of lower termites includes a het- erogeneous population of bacteria, as well as unique genera and species of flagellate protozoa. t Journal article 10752 from the Michigan Agricultural Ex- periment Station. Although the overall hindgut fermentation prob- ably reflects a concerted interaction between protozoa and bacteria, protozoa appear to domi- nate this activity inasmuch as they are abundant and are key agents of cellulose hydrolysis. Their presence in the gut is critical to the survival of lower termites. A scheme for symbiotic cellulose utilization in lower termites was proposed many years ago by Hungate (22), who studied cellulose fermenta- tion by crude suspensions of mixed protozoa obtained from Zootermopsis species. According to his model, protozoa ferment wood cellulose to C02, H2. and acetate, and the acetate is subsequently absorbed from the hindgut and oxidized by the termites for energy. The validity of Hungate‘s model was recently buttressed by Yarnin, who found that axenic cultures of hind- gut protozoa also formed C02, H2, and acetate from cellulose (56. 57) and that survival of defaunated Zootermopsis termites (i.e., with protozoa removed) on a diet of cellulose could be achieved by refaunating the termites with axenic cultures of Trichomitopsis termopsidis (58). 1602 VOL. 45. 1’3 Although acetate appears to be an important oxidizable substrate for termites. as well as an important precursor for the biosynthesis of ter- mite fatty acids (4, 33), terpenes (45). and (along with propionate) cuticular hydrocarbons (5). lit- tle is known of the actual concentration of acetate and other volatile fatty acids (VFAs) in termite hindgut fluid. Moreover, no direct mea~ surements have been made of VFA production and oxidation in situ. This is probably due to the small size of most termites, which hampers such analyses. Nevertheless, we felt such information was critical to a better understanding of the termite hindgut ecosystem and the importance of hindgut microbes to termite nutrition. Ac- cordingly. it was to these issues that the present study was directed. Reticulitermes flavipes (Kollar). the common eastern subterranean ter- mite, was used for the major portion of this study because (i) it is one of the most abundant termites in the United States (54), (ii) specimens were available locaII)’. (iii) wood polysaccha- rides are extensively degraded in its gut (18), and (iv) it is a species around which we have devel- oped a substantial data base (8, 11, 12, 40—44, 48, 49). (Portions of this work were presented at the Blst Annual Meeting of the American Society for Microbiology, 1—6 March 1981, Dallas, Tex., and at the 16th Conference on Rumen Function, 11-13 November 1981, Chicago, Ill.) MATERIALSANDME‘I'HODS Tet-hes .d runen lid. R. flavipes (Kollar) (Rhi- notermitidae) was collected from Janesville, Wis, and from Dansville and Spring Arbor, Mich. Termites were used immediately or were maintained in the laboratory for periods up to a year by incubation in covered. polycarbonate boxes containing slices of commercial Douglas fir lumber interspersed with moist brown paper towels (Scott Singlefold, no. 175). Incubation was at 22 to 26°C, and the paper towels were periodically remoistened with distilled water. Zootennopsis angusticollis (Hagen) (Hodotermitidae) was obtained from Dahl Biological Supplies, Berkeley, Calif; Coptatermes formosanus Shiraki (Rhinotermi- tidae) was kindly supplied by G. R. Esenther (Forest Products Laboratory. US Dept. of Agriculture, Mad- ison. Wis): Prorhinotermes simplex (Hagen) and Schedorhinotermes lamanianus (Sjostedt) (Rhinoter- mitidae). Incisitermes schwarzi (Banks) (Kalotermiti- dae). and the “higher" termite Nasutitermes corniger (Motschulsky) (T errnitidae) were kindly supplied by B. L. Bentley and G. D. Prestwich (State University of New York, Stony Brook). Worker termites (i.e., externally undifferentiated larvae beyond the third instar) were used for all experiments. Rumen fluid was obtained by aspiration from a fistulated dairy cow and passed through three layers of cheesecloth before use. Sampling of VFAs. A “Micro" method was used to sample termite hindgut fluid for VI-‘A analysis. Ter- 23 TERMITE HINDGUT FERMENTATION 1603 mites were first chilled to 4‘C to immobilize them and then degutted by using fine-tipped forceps (12). a procedure which yielded the entire hindgut along vn‘th a short piece of attached midgut. Such preparations were referred to as extracted guts. Extracted guts were dipped into 2 mM potassium phosphate bufl’er (pH 7.6) to rinse ofl’ hemolymph, blotted with paper tissue. and transferred to a slab of dental wax. The bulbous. paunch region of the hindgut (12) was then pierced with a dissecting needle, and the liquidy hindgut contents that issued from the puncture site were aspirated into a glass capillary tube (l-ul capaci- ty; Dade Div. American Hospital Supply Corp.. Mi- ami, Fla). One capillary tube was used per termite, and all sampling was done as rapidly as possible at 2 to 4'C. The height of the column of hindgut contents within each capillary tube was measured by using a dissecting microscope equipped with an ocular mi- cnometer. From this measurement the extracellular fluid volume was inferred (see below). The contents of 2 to 31 capillary tubes were then pooled by quantita- tive transfer into a small polypropylene centrifuge tube (Brinkmann Instruments, Westbury, N.Y.) containing 10 to 50 ul of BIS (BIS is 2 mM potassium phosphate bufler [pH 7.6] containing 1.0 mM a-methylbutyrate as an 'mternal standard). This resulted in a 20- to 40-fold dilution of hindgut fluid. but yielded volumes that could be more easily manipulated for further process- ing. Mixtures were then centrifuged at 13,” X g for 20 min at 4°C, and supernatant fluids were used for quantitation of VFAs by gas chromatography. To determine the extracellular concentration of VFAs in hindgut fluid, it was first necessary to estab- lish a conversion factor for the fraction of extracellular fluid in a given volume of hindgut contents. To do this, we plugged capillary tubes containing various amounts of hindgut contents at the base with paraffin. capped them with a small piece of Patafilm M (American Can Co., Greenwich, Conn.), and centrifuged them at 13.!!!) x g for 60 min with a hematocrit centrifuge. After 60 min. no further compaction of particulate material was observed. The height of the supernatant fluid was then measured as described above and taken to be the extracellular fluid volume. For various termites, the extracellular fluid volume of hindgut contents was (mean a standard error of the mean): R. flavipes, 38.6 2 7.0% (n = 9); l. schwarzi, 60.7 t 3.0% (n = 4); and Z. angusticollis, 63.2 s 7.5% (n -= 5). These conversion factors were used to infer the exua- cellular fluid volume during routine analyses. All capil- lary tubes used in this study were first cleaned by immersion in Chromerge (Manostat, New York. N.Y.)—H,SO, solution for 24 h. followed by rinsing with tap and deionized water. The capillary tubes were then oven dried and coated with Sigmacote silicone reagent (Sigma Chemical Co., St. Louis. Mo.). Capil- lary tubes were calibrated by using a 311,0 standard (specific activity. 2.96 x 10" dpm/rnl). Termite hemolymph was obtained by gently piercing the cuticle between the third and fourth abdominal tergites and aspirating the clear fluid into a capillary tube. After volume determination, hemolymph sam- ples from nine termites were pooled in BIS for subse- quent analysis. To sample termite feces, we placed 30 termites in a tared polypropylene centrifuge tube in which they were allowed to defecate for 24 h. The termites were 21+ 1604 ODELSON AND BREZNAK then removed, and the centrifuge tube was reweighed to estimate the fecal mass. afler which 50 pl of BIS was added to extract VFAs. Particulate material was removed by centrifugation, and the supernatant fluid was used for analysis. To estimate the VFA content of different hindgut regions. extracted guts were removed to a slab of dry ice covered with a thin sheet of plastic, whereupon they immediately froze. A razor blade was then used to liberate the posterior portion (rectum) of each hindgut. The bulbous anterior portion (paunch plus colon) of each hindgut was then separated from the midgut by a razor slice just anterior to the enteric valve. Generally, 6 anterior or 30 posterior hindgut regions were pooled in 20 ul of B18 and homogenized with a glass rod. Homogenates were then centrifuged as described above. and supernatant fluids were used for analysis of VFAs. Rumen fluid was sampled by the Micro method described above. or handled by a “Macro" method as follows. A 200-ml quantity of fluid was clarified by Zn(OH)2 precipitation (37). and 15-ml samples of the clarified liquor were subjected to steam distillation (37). Distillates were neutralized with NaOH and brought to dryness by heating at 80°C. The dry sodium salts of VFAs were dissolved in 2.0 ml of water. acidified with H,SO,. and extracted into diethyl ether (40) for subsequent quantitation. Analysis of VFAs. Quantitation of VFAs in termite preparations. as well as in rumen fluid sampled by the Micro method. was done with a Varian model 2440 gas chromatograph equipped with an H, flame ionization detector. Temperature settings were ('0: injector, 175; column. 125; detector. 175. Flow rates were (ml/min): N, carrier gas, 30; air, 300; Hz, 30. The glass column (183 by 0.2 cm) was coated with Sigmacote and packed with Carbopack C impregnated with 0.3% Carbowax 20 M and 0.1% H,PO, (Supelco, Inc.. Bellefonte. Pa.). Before use, the packed columns were preconditioned with water or H,PO, according to the manufacturer's recommendations (bulletin no. 7518; Supelco. Bellefonte. Pa.). Chromatograms were re- corded with a Hewlett-Packard model 3390A reporting integrator. which was also used to calculate quantities of VFAs by reference to the internal n-methylbutyrate standard. This procedure afforded excellent separation and quantitation of C; to C, VFAs in the range encountered with diluted hindgut contents (0.2 to 5.0 mM; see Fig. 1). To estimate formate. we homogenized 150 extracted guts in 1.0 ml of 50 mM potassium phosphate bufler (pH 7.6) and removed particulate material by centrifu- gation. The supernatant fluid was then acidified with H,SO., extracted with diethyl ether, and analyzed by gas chromatography on a column of SPo1220 (Supelco, Inc.. Bellefonte. Pa.) (40). VFA analysis of rumen fluid sampled by the Macro method was done by using gas chromatographic condi- tions previously described (40). Mass spectra of acetate were determined by divert- ing a portion of the compound separated by gas chromatography to a Finnigan model MS/GC mass spectrometer operating at 6 x 10‘2 Pa and 70 eV. Feeding experiments. Incubation vessels and condi- tions were similar to those described previously (44). However. when "C-labeled substrates were fed to termites. vessels were modified to be gas tight and contained a piece of tissue wetted with 0.5% H,PO, below the screen platform supporting the termites. For Ant. ENVtaos'. Micaoat0t. such vessels. the headspace was periodically flushed with air. and "C0; in the exit air was trapped in phenethylamine (44). Douglas fir powder (DFP) was the main constituent of food tablets. To prepare DFP. sawdust was first extracted with hot water (12). oven dried. mid then ball milled for 120 h. Particles small enough to pass through a standard ISO-um sieve were then pooled as DFP and used for compaction into till-n. food tab- lets. DFP was assumed to contain 47% cellulose and 23% hemicelluloses (dry weight basis; reference 18). To determine the contribution of cellulose versus hemicellulose carbon to acetogenesis in termite hind- guts, we amended DFP with 20% (wt/wt) lU—“Clcellu- lose or 10.5% (wt/wt) [U-"Clhemicellulose before compaction into tablet form. The specific activity of cellulose or hemicellulose in such tablets was calculat- ed from the amount of radioactivity in weighed por- tions of tablets and the amount of specific polysaccha- ride in those portions. The former was determined by complete combustion of tablet material with an ele- mental analyzer (C. Erba model 1104; Sanda, Inc.. Philadelphia. Pa). followed by measurement of radio- activity liberated as "CO; (44). The latter was estimat- ed from the amount of "C-labeled polysaccharide incorporated into tablets and the amount d unlabeled component present as part of the DPP material itself (see above). Termites which had fed on such tablets for 4 days were degutted. and 4 to 6 extracted guts were placed in a small centrifuge tube containing 20 ul of BIS. quickly frozen in dry ice. and then homoge- nized with a glass rod while thawing. Preparations were centrifuged as described above. and supernatant fluids were used for analysis of l"C)acetate. The legitimacy of this approach was based on the fact that virtually all of the acetate associated with extracted guts was present in the hindgut fluid (see below). For determination of the specific activity of ["Clacetate. samples were first injected into a Varian model 3700 gas chromatograph to quantitate the amount of acetate present. The glass column (183 by 0.4 cm) was packed with Chromosorb W impregnated with 15% SP-1220 and 1% H,P0.. Temperatures were (‘0: injector. 200; column, 175; flame ionization detector. 200. Flow rates were (ml/min): N2 carrier gas. 30; air, 300; Hz. 30. "(30,. liberated from [“Clacetate in the flame ionization tower. was then collected by bubbling the gas through 7.0 ml of ethanolamine-methanol. 3:4 (vol/vol). The trapping efficiency of this system was determined to be 88%. The specific activity of ("Cla- cetate was then calculated from the amount of acetate of acetate present and its radioactivity. The ability of termites to oxidize VFAs was evaluat- ed by measuring “CO, evolution from termites feed- ing on I‘C-labeled VFAs. DFP tablets used for such experiments were first moistened lightly with 2 mM potassium phosphate buffer (pH 7.6) followed by 20 pl of bufler containing (nmol): sodium lU-"Clacetate, 29.2; sodium [1-"Clpropionate, 179.0; or sodium [2- “Clbutyrate. 142.0. Carrier-free radioactive com- pounds were used, and the amounts added were such that each food tablet contained 0.2 to 2.0 uCi. To determine the origin of "CO; evolution in such experi- ments, we subsequently removed termites from the incubation vessels. and intact specimens. as well as degutted bodies and extracted guts. were reincubated separately in S—ml stoppered serum vials containing a filter paper disk moistened with water. “CO; was collected as previously described (42). 25 Va. 45. 1983 The ability of defaunated or normally faunated ter- mites to survive on a diet of acetate was evaluated by usin food tablets consisting of 2 M sodium acetate bcaporated into a gel d 10% (wt/vol) agarose. Corr- trol tablets consisted of agarose alone and agarose containing 2 M NaCl. Anti-genie aetlvly h situ. In situ rates of acetogene- sis by the termite hindgut microbiota were estimated by a modification of the zero-time-rate method (24). Termites were introduced into an anaerobic glove box (Coy Manufacturing Co., Ann Arbor. Mich.) contain. in an atmosphere of 90% Nr10% H, and quickly degutted. Groups of six extracted guts were incubated at 23'C in centrifuge tubes containing 45 pl of 2 mM potassium phosphate buffer (pH 7.6). The buffer solu- tion had been boiled and cooled under 100% N, before use. At zero time. and at periodic intervals thereafter. tripbcate groups of guts were quick frozen in dry ice and processed for VFA analysis as described above. except that the a-methylbutyrate standard was added after the homogenization step. This approach permit- ted the isolation of acetogenic activity from the respi- ratay activity of the host (via gut removal and anaero- bic incubation). and it also allowed estimation of acetate present within extracted guts as well as that exported to the incubation buffer. linination of gut microbes. Defaunation was ac- complished by treating termites with more 0, at 172 kPa (25 lb/in ) for 2 h at ambient temperature (14). Termites were then reincubated in isolation for 4 days before use. Microscopic examination of gut contents in-ediately after hyperbaric 0, treatment revealed thl all protozoa were immobile; after 3 days. no prmozoa were present. Iacteria were eliminated from hindguts by feeding termites antibacterial drugs for 4 days. a treatment with reduces specific heterotrophic populations 1m- fol (44). lesplratory gas exchage. Respiratory activity of termites was measured at 23’C by conventional mano- metric methods (52) and was corrected for H2 and CH. cited by the termites (see below). Generally. groups of 50 to 80 termites were held in individual reaction vessels. and each determination was made in tripli- c&. The volume occupied by the termites in the rention vessels was estimated separately by liquid displacement. CH. and H; emission by termites was measured by inabating 30 to 40 termites in 5-ml stoppered serum vids containing air and periodically sampling the head- spce gas. CH. analysis was made by using gas cllomatographic conditions similar to those used for masuring QB. (40). H, analysis was done as de- scrired previously by Ufl‘en (51). «her procedures. “C0, evolution from lU-“Clace- m by termite homogenates was measured by using 5~ ml stoppered serum vials as reaction vessels. General- Iy. 10 extracted guts or degutted bodies were hmogenized in 0.9 ml of 10 mM potassium phosphate b‘er (pH 7.6) and dispensed into the serum vials. and rertions were initiated by the addition of 0.1 ml of 2.5 pl lU-“Clacetate (ca. 0.1 pCi). Reactions were terminated by the addition of0.1 ml ofI N IICI, and ‘CO; in the headspace was trapped as described above. Measurements of radioactivity and quench correc- tit-s were made as previously described (49). Chemicals. All chemicals used were reagent grade ad purchased from commercial sources. Radioactive TERMITE HINDGUT FERMENTATION I605 IJIAJLJJJJJAlfiljlflllllljl A DETECTOR RESPONSE IAJIIAAAJIAJIIJL'AAAJJAAJ O 4 8 12 16 20 24 RETENTION TIME (min) FIG. 1. Gas chromatograms of a standard mixture of VFAs (4 mM each in BIS. top) and a diluted sample of hindgut fluid from R. flavr’pes termites (bottom). Symbols: A. acetate; P, propionate; iB. isobutyrate; B. butyrate; MB, a-methylbutyrate (internal stan- dard); iV. isovalerate; V. valerate; and U. unknown compounds. chemicals were obtained from New England Nuclear C .. Boston. Mass., except for [U-"Clcellulose and [U- ‘Clhemicellulose, which were obtained from ICN Pharmaceuticals. Inc.. Irvine. Calif. The “C-labeled cellulose and hemicellulose were repurified before use by the methods of Rapson (47) and Myhre and Smith (36). respectively. RESULTS VFAs h termite hindguts and feces. Acetate was the major VFA present in the hindgut fluid of R. flavr'pes. The compound was readily de- tected by gas chromatography (Fig. 1). and its mass spectrum (Fig. 2) was virtually identical to that of authentic acetic acid (16). The mean extracellular acetate pool size was 80.6 mM. which accounted for 94 mol% of all C, to C5 VFAs (Table 1). Small amounts of propionate and butyrate were present, but their concentra- tion frequently fell below the limits of reliable quantitation. Consequently. the values for pro- pionate and butyrate reported in Table 1 were only from those analyses done on more concen- trated samples of hindgut fluid (i.e.. pooled from >8 termites) which gave detector responses within quantifiable limits. Trace amounts of iso- butyrate, valerate. and isovalerate were some- times observed; however, formate was not de- tected. No significant differences in VFA 26 16% ODELSON AND BREZNAK 100 r- 9 r . 3‘ an 'a' C O E .0 P 4 3. ~ 2 O " .l‘r 11: 0 Q Q FIG. 2. Massspectrumofaceticacidfromhindgut fluid ofR.flavipes termites. content were observed between freshly collect- ed and laboratory-maintained R. flavipes, so the data in Table 1 represent a pooled estimate from all determinations. Limited experiments with Z. angusticollt's and l. schwarzi revealed that ace- tate dominated the VFA pool in the hindgut fluid of these termites as well (Table 1). Since the Micro method of sampling for VFA analysis entailed manipulation of minute amounts of hindgut fluid. a critical concern was possible errors incurred through sample evapo- ration or VFA volatilization. However. when bovine rumen fluid was used as a control. results obtained by the Micro sampling method com- pared favorably with those obtained by a Macro method (Table 1). These in turn agreed well with published values for the VFA content of rumen fluid (24). As expected, acetate was the major VFA present in homogenates of extracted guts of R. flavr'pes and occurred at a concentration of 18 nmol per gut equivalent (Table 2). Since the Am. Emacs. Mrcaoaros. volume of a hindgut is about 0.7 pl (48) and consisted of 39% fluid (see above) containing acetate at 81 mM (Table 1). it could be calculat- ed that essentially all of the acetate in extracted guts was present in the hindgut fluid; little existed in intracellular pools of the hindgut mi- crobiota or gut tissue. When extracted guts were sectioned. the anterior region of the hindgut (i.e., paunch plus colon) was found to contain 19 nmol of acetate per gut equivalent, whereas the rectum contained only 0.4 nmol per gut equiva- lent (T able 2). Acetate was also the major VFA in extracted guts of three other rhinotermitids examined and in N. corniger. and when dissections were per- formed, the compound occurred mainly in the anterior portion of the hindgut (Table 2). How- ever, when the acetate content of extracted guts was normalized to body weight, a threefold variation was observed. ranging from 1.5 nmol/mg (N. comiger) to 4.6 nmol/mg (R. fla— vipes) (Table 2). In some analyses of hindgut fluid or extracted guts. gas chromatograms revealed minor peaks which did not correspond to those of standard VFAs (Fig. 1. bottom). The compounds respon- sible for such peaks remain to be identified. although separate experiments indicated they were not acetoacetate. B-hydroxybutyrate. lac- tate. or ethanol. Pooled feces, voided by 30 R. flavr‘pes work- ers over a 24-h period, contained only a trace amount of acetate which could not be accurately quantitated. These data indicated that: (i) acetate was associated with. and probably produced in. the bulbous. microbe-packed anterior region of the hindgut; (ii) virtually all of the acetate present in the gut was metabolized within R. flavipes and not voided with feces; and (iii) the short segment of midgut usually attached to extracted guts contained little or no acetate. TABLE I. Extracellular pool size of VFAs in termite hindgut fluid VFA Termite Method of concn (mM)‘ ”“9"" Acetate Propionate Butyrate 3.1m"! Micro no.6 : 31.5 (n . 16) 2.8 2 1.0 (n a s) 2.0 2 1.1 (n =- 5) Z. angusri- Micro 66.2 a 0.7 (n - 5) 0.9 t 0.2 (n - 3) Trace collis l. schwarzi Micro (11 = 3) 57.9 t 3.1 1.4 t 0.4 Not detected Control (bovine Micro (11 = 3) 44.5 t 3.9 16.7 t. 1.1 6.1 t 0.6 rumen fluid) Macro (n = 3) 48.8 1 1.7 14.3 t 0.3 5.2 1 0.1 ‘ Values are means 2 standard error of the means. 11. Number of independent determinations. ‘ Trace amounts of isobutyrate, valerate. and isovalerate were also detected. Formate was not observed. ' 27 VOL. 45. 1fl3 TABLE 2. Acetate content of gut homogenates of worker termites —— 1mg 3.... Acetate (nmol) pert' Tar-'1: M '1 *- l‘ . I‘ d (as) Gm W M, m 3.71» 4.0 so 11 13.3 2 3.0 as a as viper Ali 3 19.0 a 2.6 4.8 2 0.7 R 3 0.4 t 0.2 0.1 2 0.1 S. laman- 3.9 56 2 7.6 2.0 inns All 2 8.2 2.1 P. sim- 2.8 E13 2 6.5 2.3 plea All 3 5.8 t 3.4 2.1 t 1.2 C. [armo- 3.0 56 4 7.9 a 0.6 2.6 t 0.2 ”M” N. can 2.2 E0 2 3.2 1.5 niger ‘ £6. Extracted gut; AH. anteriorhindgut;R. rec- turn. ‘ a. Number of independent determinations. ‘Meanz standarderrorofthemean.ormeanof two determinations as indicated. Acetate and atlintioa h R. fla- rr‘pes. Because acetate dominated the VFA pool in R. flavipes hindguts. it was of interest to estimate its rate of production in situ. to evalu- ate the importance of the hindgut microbiota to acetogenesis, and to appraise the termite as a potential user of this metabolite. By using a modified zero-time-rate method. in situ acetogenesis was estimated to be 20.2 .t 3.4 nmol - termite" - h" (equivalent to 5.8 t 0.1 nmol - mg of fresh weight" - h") (n - 3) for laboratory-maintained termites and 43.3 a 6.9 nmol - termite" ° h" (equivalent to 12.4 1 2.0 nmol ° mg of fresh weight" - h") (n = 3) for specimens freshly collected from the field. Re- sults of a representative experiment are depicted in Fig. 3. Rates of acetogenesis appeared linear for the first 2 h of incubation, and regression analyses yielded correlation coefficients ranging from 0.904 to 0.972 for all determinations. Propi- onate and butyrate were also observed to in- crease during the incubation period (data not shovm); however. their rates of production could not be accurately measured because of their low concentrations. Measurements of 0; consumption by R. fla- vipes revealed rates of 51.6 t 10.7 nmol - ter- mite" - h" (equivalent to 0.357 :t 0.074 pl - mg offrcsh weight" ° h") (n = 6) and 63.6 1 7.7 nmol - termite" - h" (equivalent to 0.440 :2 0.053 pl - mg" - h") (n == 3) for laboratory- maintained and fresh field specimens. respec- tively (Fig. 3). However. respiratory quotients TERMITE HleUT FERMENTATION 1”? (R05) were 1.1!) to 1.05 regardless of the origin of the insects. Defaunation. by treatment of termites with hyperbaric 01. caused a marked decrease in both the steady-state level of acetate in hindgut contents (1.1 t 0.4 nmol - termite") as well as in its rate of production (3.9 nmol - termite" - h") (Fig. 4). Removal of bacteria with antibac- terial drugs also reduced the steady-state level of acetate (8.4 t 0.7 nmol ° termite") and its rate of production (6.8 nmol - termite" - h"). but to a lesser extent (Fig. 4). The ability of R. flavipes to respire acetate was assessed by feeding them [U-"Clacetate and measuring the "C02 evolved. "C02 evolu- tion commenced immediately with an initial rate of 0.013 nmol of “C02 - termite" - h" (Fig. 5). “CO; was also readily evolved from [1-“Clpro- pionate and [2-“Clbutyrate. although rates of ‘CO; evolution from all VFAs began to decline 120 nMolaa / Tarmlta s Hours FIG. 3. Rates of in situ acetogenesis by the hindgut microbiota and 0; consumption by intact worker larvae of R. flavr'pes. Termites were freshly collected from the field before assay. 28 I“ ODELSON AND BREZNAK O 2 g . g. \ O a I a 8 ‘ 2° Antibiotics . / , s // s . (”my ‘ri a a a 00 15 so 45 Minutes FIG. 4. Eflect of defaunation (hyperbaric 05) or removal of hindgut bacteria (antibiotics) on acetogenic activity of hindguts of R. flavi'pes termites. between 6 and 24 h of incubation (Fig. 5). This suggests that “C-labeled VFAs may not have been unifome distributed throughout the food tablets or that the feeding rate of termites de- creased during incubation. In the absence of termites. negligible amounts of "CO, were evolved from food pellets (Fig. 5). After 52 h. 1.8 r 0‘ . V E Propionate 0/ .2 t 0 . \ / 1 0“ 3° / ' Iutyrata J . 0.5 f g i x g ‘Aeatata Control 0 _- 12 18 24 82 Hours FIG. 5. Evolution of "CO; bch. flavipes termites feeding on lU-"Clacetate. [I-' lpropionate. or [2- I‘Clbutyrate. Control vessels contained lU—“Clace- tate incubated in the absence of termites. Am. ENVIRON. MicaoaiOL. TABLE 3. Origin of “CO; from R. flovi'pes fed "C- libeled VFAs VFA fed to Incubation "gm“ .91"??? termit s‘ atmosphere c Intact lDeguttedtltncted termite body gut [U-“Clacetate Air 0.82 0.76 0.20 100% N; 0.34 0.32 0.10 [1-“Clpropionate Air 1.86 1.63 0.30 100% N; 0.87 0.80 0.24 lZ-"Clbutyrate Air 3.41 3.03 0.71 100% N; 1.65 1.26 0.47 ‘ Termites were fed “C-labeled VFAs for 52 h before assay. The specific activities of the substrates were (dpm ° nmol"): [U-"Clacetate. 60.050; [1- "Clpropionate. 31.1!)6; and [2-"Clbutyrate. 3.419. ‘ Groups of 7 to 10 termite equivalents were used per determination. termites were removed from the incubation ves- sels and dissected to determine the origin of "coy-evolving activity (Table 3). Most of the "C02 evolution was associated with degutted bodies (i.e., termite tissues only) and was signifi- cantly retarded by incubation under N2. The observation that some “CO; was evolved under N2 suggests that dissolved 02 may still have been present in termite tissues. Only small amounts of “CO; were evolved from extracted guts and presumably arose from respiratory ac- tivity of gut tissue. In separate ex riments, termites previously unexposed to [ ‘Clacetate were degutted, homogenized. and incubated in vitro with [U—“Clacetate. Under aerobic condi— tions. 83% of the "C02 evolved arose from degutted bodies. whereas 17% arose from gut homogenates. Analysis of hemolymph. The ability of R. fla- viper tissues to oxidize acetate suggested that acetate produced in the hindgut was absorbed from that site and transported to tissues via the hemolymph. Accordingly. hemolymph was as- sayed for the presence of this compound. In two separate experiments. acetate was found to be present at concentrations of 9.0 and 11.6 mM. respectively. Propionate and butyrate were not observed. but could have been present in con- centrations below the limits of detection (<4 mM in hemolymph). Acetogenesis from wood polysaccharides. To estimate the contribution of cellulose versus hemicellulose carbon to acetogenesis. we fed DFP amended with [U-"Cl-labeled polymer to R. flavr’pes and determined the specific activity of acetate in the hindgut fluid (T able 4). Compar- ison of the specific activity of the food compo- 29 VOL. 45. 1N3 TABLE 4. Origin of acetate in hindgut fluid of R. fiai'ipes termites Spacttdpm-agofC"? in food . % Contri- Food Acetate in bution “bl“ component hindgut fluid lU-“Clcellu- 0.86 1' 0.12 0.75 3. 0.16 87.2 lose lU-“Clhemicel- 11.5 11.2 2.66:0.59 23.1 hilose ‘Mean:standarderrorofthemean(nc3). nent with that of acetate revealed that about 87% of the acetate was derived from cellulose. whereas 23% was derived from hemicellulose. The sum of these values strongly suggests that acetate was produced only from wood polysac- charides and not from lignin or other wood components. That the sum was greater than 100% suggested that the labeled polymers added to food tablets were more readily converted to acetate than were the corresponding unlabeled polymers in wood. which are complexed with lignin (17). Survival of R. flovipes on acetate. Attempts to prolong the survival of R. flavipes on a cellulose- free diet with acetate were unsuccessful. This was true whether defaunated or normally faunat- ed termites were used. In fact, termites feeding on sodium acetate or NaCl (control) generally died faster than did those feeding on unamended agarose. B; and CH. emission by R. fiovipes. Live specimens of R. fiavr’pes usually emitted H2 at rates of 0.4 to 0.9 nmol - termite" - h" (equiv- alent to 0.003 to 0.006 pl - mg of fresh weight" - h"). although occasional groups evolved up to 4.2 nmol - termite" - h" (equiv- alent to 0.029 pl - mg" - h"). Rates of CH. emission were 0.38 t 0.2 nmol - termite" - h" (equivalent to 0.(l)3 : 0.001 pl . mg" - h") (n == 20). No significant difference in the CH. emission rate was observed between laboratory- maintained and freshly collected termites. Thus. on a molar basis. the rate of H2 or CH. evolution was always 57% (and usually 50.8%) that of 0; consumption or C0; evolution. Interestingly. however, when a group of termites emitting trace amounts of Hz were fed antibacterial drugs (chloramphenicol. penicillin. and tetracycline), the rates of H2 evolution increased to 1.7 to 7.0 nmol - termite" - h" within 48 h of drug treat- ment, and CH. emission increased from 0.25 to 1.50 nmol - termite" - h". DISCUSSION VFAs present in the hindguts of xylophagous termites included C2 to C5 representatives (T a— bles 1 and 2) and were qualitatively similar to TERMITE I'IINDGUT FERMENTATION 1609 those found in other gastrointestinal ecosystems harboring a dense microbiota. such as the large bowel and cecum of vertebrates (1. 13. 35. 50). the rumen (24, 25). and the hindgut of cock- roaches (6) and scarabaeid beetles (3). In addi- tion. our results were consistent with those of Kovoor (27). who qualitatively identified ace- tate, propionate. and butyrate in the hindgut of the higher termite Microcerotermes edentatus by using paper chromatography. and those of Hungate (21. 22), who identified acetate in the hindgut fluid of Z. ongusticolli's. Quantitativeiy. however. the VFA profile in termite hindgut fluid was quite different from that of most other intestinal ecosystems in that acetate accounted for an unusually large fraction (94 to 98 mol%) of all VFAs (Table 1). This was not due to prefer- ential absorption and oxidation by termites of C3 to C5 VFAs, because acetate was still the major VFA produced when the oxidative activity of host tissue was circumvented (i.e.. by the zero— time-rate method of analysis). The striking dominance of acetate in the hind- gut fluid of lower termites is probably due to several factors. including the relativelyinsoluble nature of the food of the termites and the rather unique composition of the hindgut microbiota. The major components of wood are insoluble polysaccharides (cellulose and hemicelluloses) complexed with 18 to 35% lignin (17). Although it appears that some lower termites might se- crete their own cellulases (53). and may in fact initiate wood glucan decomposition in the fore. and midgut. the majority of cellulolytic activity occurs in the hindgut (where the bulk of wood decomposition occurs) and is of protozoan ori- gin (8—10, 20. 23. 38, and references therein). Moreover, the transit of food from the foregut to the hindgut is rapid (<3 h; 28. 29; Odelson and Breznak, manuscript in preparation). Conse- quently. most of the food remains in an insolu- ble, particulate form during passage to the hind- gut. Since the protozoan population in termite hindguts is quite large (the hindgut of R. flavi'pes contains 4 x 10‘ protozoa [34] and 3 x 10" bacteria [48]. and the ratio of protozoa to bacte- ria is about 1.000 times greater than that of the bovine rumen [25, 55]; in lootermopsi's termites the hindgut protozoa account for about one-third of the body weight of the insect [26]). and since most of the termite hindgut flagellates endocy- tose (and thereby sequester) wood particles as they enter the hindgut, it follows that VFA production in the hindgut should reflect mainly a protozoan fermentation of wood polysaccha- rides and be dominated by acetate. Acetate is the only detectable VFA produced during cellu- lose fermentation by mixed suspensions of hind- gut protozoa from Zoorermopsi’s termites (21 . 22), as well as by axenic cultures of these forms 30 1610 ODELSON AND BREZNAK (56. 57). and a similar situation probably holds for cellulolytic protozoa from R. flavi'pes. Con- sistent with this interpretation is the drastic inhibition of acetogenesis in the hindgut of R. flat-ipes after defaunation (Fig. 4). Nevertheless. bacteria undoubtedly also produce acetate in situ. This inference is based on the moderate suppression of hindgut acetogenesis after R. flavi'pes termites were fed antibacterial drugs (Fig. 4). as well as the recognized ability of heterotrophic bacterial isolates to produce ace- tate (and C, and C3 to C, VFAs) in pure culture (42. 43. 48) and in two-species cocultures (49). However. it is impossible at this time to ascribe the exact quantitative contribution of protozoa or bacteria to acetogenesis in situ, because the sum of the acetogenic activity of defaunated R. flavi'pes and bacteria-free R. flavt’pes is signifi- cantly less than the acetogenic activity of con- trol termites (Fig. 4). Either one or both of the following explanations for this observation are possible. (i) The treatment used to remove pro- tozoa or bacteria (exposure of termites to hyper- baric 02 or to antibacterial drugs, respectively) has some deleterious eflect on nontarget orga- nisms; or (ii) acetogenesis in normal R. flavipes involves a synergistic interaction between hind- gut protozoa and bacteria. Notwithstanding. it seems safe to conclude that protozoa dominate acetogenesis in R. flavi'pes hindguts. whereas bacteria are of secondary importance in this particular activity. Substrates for bacterial pro- duction of VFAs could include the small amount of soluble carbohydrate present in the wood itself (30). soluble intermediates secreted by the protozoa (9. 56) or liberated from wood by termite enzymes. or possibly CO; and H; (see below). True cellulolytic bacteria. i.e.. bacteria capable of degrading crystalline cellulose, ap- pear to be quantitatively insignificant in the hindgut of R. flavr’pes (48). Results of our present studies with R. flavi’pes (family Rhinotennitidae) are consistent with Hungate‘s (22) model for mutualistic cellulose utilization which was derived from his studies with Zooremiopsi's species (family Hodotermiti- dae). First. protozoa appear to be primarily responsible for acetogenesis in the hindgut; sec- ond, rates of 0; consumption by R. flavi'pes (52 to 64 nmol ° termite" - h") were approximate- ly twice that of hind t acetogenesis (20 to 43 nmol - termite" - h‘ for laboratory-main- tained and freshly collected termites. respective- ly). Since 2 mol of O; is required for complete oxidation of acetate to 2C0; and 2H20. it ap- peared that 77 to 100% of the energy require- ments of the termites could be met by oxidation of the acetate produced by the hindgut micro- biota. In support of this interpretation was the demonstration of significant amounts of acetate Ant. Erwraon. 1410011101.. in R. [lavi'pes hemolymph. as well as the ability of termite tissues to readily respire acetate and other VFAs (Fig. 5; Table 3). We do not know why the rates of acetogenesis in the hindguts of freshly collected R. flavipes were consistently greater than those of laboratory-maintained specimens. However, it seems likely that the food on which the former were feeding before assay was more readily convertible to acetate (perhaps because it was partially decayed by fungi) than was the sound wood and paper towel mixture fed to laboratory specimens. Neverthe- less. the ability of both cellulose and hemicellu- lose to serve as substrates for acetogenesis by the hindgut microbiota (Table 4) was in line with the high digestibility of these compounds, but not lignin. for R. flavipes (18). Rates of 0; consumption by R. flavi'pes re- ported herein (0.357 to 0.440 pl - ing of fresh weight" - h") were similar to those of various other termites. as summarized by Peakin and Josens (39). as well as that determined by La- Fage and Nutting (31) for Marginilermes hub- bardi'. By contrast, present values were consid- erably lower than most of those determined for R. flavipes by Damaschke and Becker (summa- rized in reference 39). The reasons for this discrepancy are not known. Rates of H; and CH. emissions by R. flavipes were also similar to those previously reported for various termites including R. flavipes and R. tibiali's (8. 15. 31. 59). Although emission of such gases by termites might have a significant impact on our atmo- sphere globally (59). rates of H2 and CH. emis- sion by R. flavi'pes were only about 0.7% that of 0; consumption. Consequently. overall carbo hydrate oxidation in R. flavi'pes closely approxi- mated the classical scheme: 100 (CH20) + 1000; -v 100C02 + 100H20. Assuming an oxy- calorific equivalent of 5.05 meal/pl of 02 con- sumed (39). our respirometric data indicate that energy flow through normally faunated. feeding workers of R. flow)»: would be 1.80 to 2.22 meal - mg" - h" at 23°C. If Hungate’s model (22) for symbiotic cellu- lose degradation in lower termites is fundamen- tally valid. it must be amplified to accommodate those termite species that evolve relatively little H2 and CH.. For example. if symbiotic wood utilization in R. flavipes is envisioned to consist mainly of an anaerobic fermentation of glucan (I'ngHnOg) to acetate. C02, and H; by protozoa (T able 5, reaction A), followed by termite oxida- tion of acetate (Table 5. reaction B). then an appreciable amount of reducing equivalents, as 11;. should be evolved by termites (Table 5. reaction A + B). In fact, according to reaction A + B, rates of H2 evolution should be equal to that of 0; consumption and 66% that of C02 evolution, but they are almost always <1% of 31 VOL. 45. 1N3 TABLE 5. Possible steps in symbiotic dissimilation of glucan (nCJ-InO.) by R. flovi'pes‘ Designation Reaction A ............. nCJ‘luO. + 2nH10-o 2nCH,COOH + 2nCO; + 4nH; B ............. 2nCH3COOH 4» mo, -9 4nC0, + 4nH30 A + B ......... nC.HnO. + 4n0; -9 6nCO, + 4nH; + 2nHzO C ............. 411113 + “CO; ‘P ”CH4 + ”"10 A + B + C....nC.H,,0.+ 4nO,-'5nCO; + “CH. ‘1" ‘DHIO D ............. 4nH; + 2n0, -i 4nH10 E ............. I'ICH‘ '1’ 21102 "P ”CO; + 20820 F ............. 4nH, + 2nCO, -o nCH3COOH + 2an B + D or ............. 2nCH,COOl-I + (mo, + 4nH, —. B+C+E 4nCO,+8nl~l,O A + B + D a ............. “1206 + 8102 '9 “C02 + A + F + 1.58 6nH10 ‘ See text for details. those values. Furthermore. reaction A + B is inconsistent with the R0 of R. flavi'pes. which was 1.00 to 1.05. Several possibilities exist re- garding alternate fates of reducing equivalents produced during glucan decomposition by R. flavipes. Interspecies transfer of Hz to methano- genic bacteria might occur (Table 5. reaction C). and it probably does inasmuch as R. flavipes emits CH.. but the combined reaction A + B + C (T able 5) predicts that CH. emission rates should be 20 to 25% that of CO; evolution and 0, consumption. respectively. However, as for H2 emission. CH. emission was <1% of those values, and the RQ predicted by reaction A + B + C is inconsistent with that observed. Con- sumption of H; by aerobic hydrogenotrophic bacteria (T able 5, reaction D) would yield an RQ of 1.“) (Table 5. reaction A + B + D). but implies that 0; consumption should be threefold greater than hindgut acetogenesis (T able 5, reac- tion B + D) instead of the observed 1.5- to 2.4- fold. A similar argument can be leveled against the possibility of rapid formation (reaction C) and subsequent oxidation (T able 5, reaction E) of CH. (Table 5. reaction B + C + E). The most likely fate of H2 is depicted by reaction F (Table 5), i.e., the use of Hz for the TERMITE HINDGUT FERMENTATION I611 reduction of CO; to acetate. The overall dissimi- lation of glucan by R. flavi'pes is then envisioned to consist of anaerobic fermentation by hindgut protozoa (reaction A). coupled to anaerobic acetogenesis from C0; and H; by some mem- ber(s) of the hindgut microbiota (Table 5. reac- tion F). followed by aerobic oxidation of acetate by R. flavi'pes tissues (1.5 X reaction B). The sum of these reactions (Table 5. reaction A + F + 1.58) would be consistent with the R0 and relative rates of 0; consumption and hindgut acetogenesis of R. flavi'pes. Reaction F is not known to occur in eucaryotes. but is recognized in a few bacterial species such as Acerobacrer- firm (2) and Acerogerii'um (32). and in certain species of C losrri'di'um (7) and Eubacleri'imi (19). Moreover, it is of interest that the cecal micro- biota of rats. rabbits. and guinea pigs can effect a total synthesis of acetate from C0; and H2 (46). although the specific bacteria have not yet been identified. Although conclusive proof for the existence of such reactions or organisms in R. flavipes hindguts is not yet available, the ob- served increase in H2 and CH. emission by termites fed antibacterial drugs (see above) sug- gests that some type of procaryotes. other than methanogens. constitute important “electron- sink“ organisms in the hindgut food web. A working model for symbiotic utilization of wood polysaccharides by R. flavi'pes is depicted in Fig. 6. The central elements of Fig. 6 include a graphic representation of reactions A. F. and B (Table 5). Also depicted in this figure is the participation (to a lesser degree) of other hetero- trophic bacteria in acetogenesis. either directly or indirectly through H2 and CO; production (42, 43. 48. 49). As with Hungate‘s original model (22). the present model (Fig. 6) implies that defaunated termites should be able to sur- vive on a cellulose-free diet if fed acetate. Al- though this has been tried previously (15. 23). as well as in the present study. negative results have been obtained. Perhaps the feeding of acetate salts to termites does not adequately mimic the in vivo situation in which acetic acid is continuously produced and absorbed from the hindgut. On the other hand. perhaps the cation component of the acetate salts has a deleterious effect on termite survival. as indicated in the present study. However. Fig. 6 suggesrs a novel way to circumvent the use of acetate salts in such survival experiments, i.e., the present model predicts that defaunated (or normally faunated) termites should be able to survive on a cellulose-free diet longer under an air atmo sphere enriched with H; and CO; than under an atmosphere of air alone. A test of this predic- tion, as well as a search for acetogenic C02- reducing bacteria in R. flavipes hindguts, is currently under way in our laboratory. MW. ~,'.. ‘3 '1'! 32 1612 ODELSON AND BREZNAK . I -\ . \ APPL. ENVIRON. Micaoaim. cu. ME 1 HANOGENIC sacrum 00! act rosemc co: morozoa H2 ’ ‘ .“ZI‘SEWCERS 9 5 ‘ L Pousacciiaaiocs acerarr: L > ABSORBED BY kTERIITES FOR SOLUBLE MON0-. OXIDATION O 0' -. ANDROL'GO' BIOSYN‘I’I'IESIS S‘CCH‘ ”’55 ACETATE c::> FERMENTWE PROPIONATE ————. aacnms BUT YRATE ————* eq Srrcproeaerus (LACTATE) ACE TAT E Barnum!“ _J “intent Tim (:0 IACETOGENIC coz Fusooocnnum 3 l RE DUCE RS ? tocloboaI/us H z b " ’ " acruanosemc (FORMATEI ““5“" cu. FIG. 6. Proposed working model for symbiotic wood polysaccharide dissimilation in R. flavipes. Thickness of arrows represents approximate relative contribution of the respective reactions to the overalldissimilatory pathway. Major products of the hindgut fermentation are indicated in boldfaced type; probable intermediates. which do not accumulate to detectable levels. are indicated in parentheses. (Reprinted from reference 10 by permission from the British Mycological Society.) ACKNOWLEDGMENTS We thank E. Oliver for help with mass spectrometry and M. l. Klug for use of the elemental analyzer and for help with determination of acetate specific activity. This work was supported in part by National Science Foundation pants PCM 77415732 and PCM “12026 and by the Michigan Agricultural Experiment Station and College of Osteopathic Medicine of Michigan State University. LITERATURE CITED 1. Argeazio. R. A.. .11 M. Southwarth. 1975. Sites oforgan- ic acid production and absorption in gastrointestinal tract ofthe pig. Am. J. Physiol. 38:454-460. 2. Batch. W. E.. S. Schaberth. R. S. Tanner. .4 R. S. Wolfe. 1977. Arerobarrerr‘am. a new genus (1’ hydrogen. oxidizing. carbon dioxide-reducing. anaerobic bacteria. Int. J. Syst. Bacteriol. 27:355-361. 3. Bayou. C.. and J. Matheh. 1%0. Carbohydrate fermenta- tion and by-product absorption studied with labelled cellu- lose in Oryrrrs nasi'rorni's larvae (Coleoptera: Scarabaei— dae). J. Insect Physiol. 3:833-840. 4. Bloraqulst, G. 1.. L. A. Dwyer. A. J. Chu, R. 0. Ryan. .d M. de Reaobdes. 1932. Biosynthesis of linoleic acid in a termite. cockroach and cricket. Insect Biochem. 12:349- 353. 5. Blomqulst. G. J., R. W. W, .d C. A. hit-Mel. 1979. Biosynthesis of the cuticular hydrocarbons of the termite Zooremiopsi's aagusri'eolli‘s (Hagen). Incorpo- ration of propionate 'mto dimethylalkanes. Insect Bio- chem. 9:371—374. 6. Braehe. J. W., .d A. J. Mar-hm 19110. Transport of bacterial end products from the colon of Periplaneto amerr‘cana. J. Insect Physiol. 26:85—89. 7. Iran. M.. F. Mayer. ad 6. Gottsehalk. 1%1. Closrridi- rim aren't-um (Wieringa). a microorganism producing ace- tic acid from molecular hydrogen and carbon dioxide. Arch. Microbiol. 128:288-293. 8. Breznak. J. A. 1975. Symbiotic relationships between 10. 11. 12. 13. I4. 15. 16. 18. 19. termites and their intestinal microbiota. p. 559-580. Iii D. H. Jennings and D. L. Lee (ed.). Symbiosis (Society for Experimental Biology Symposium Series number 29). Cambridge University Press. Cambridge. England. . Breznak. J. A. 1%2. Intestinal microbiota of termites and other xylophagous insects. Annu. Rev. Microbiol. 3:323- 343. Ireand. .l. A. 1Q3. Biochemical aspects of symbiosis between termites and their intestinal microbiota. In I. M. Anderson. A. D. M. Rayner. and D. Walton (ed.), Inver- tebrate-microbial interactions. Cambridge University Press. Cambridge. England. Bran-k. J.A.. W. J. Bel, J. W. Maths. .d II. C. Cappel. 1973. Nitrogen fixation in termites. Nature (Lon- don) 244:577-580. Breznak. J. A.. .d II. S. Pdratz. 1977. In situ morphol- ogy of the gut mic robiota of wood-eating termites [Reticu- li'reririrs fiavipes (Kollar) and C optorrmtes [amorous Shiraki). Appl. Environ. Microbiol. 33:406-426. Byrae. B. M.. and J. Dankert. 1979. Volatile fatty acids and aerobic flora in the gastrointestinal tract of mice under various conditions. Infect. Immun. 23:559-563. Clevehad. L. R. 1925. Toxicity of oxygen for protozoa in vivo and in vitro: animals defaunated without injury. Biol. Bull. 48:455-468. Cask. S. I". 1943. Nonsymbiotic utilization d carbohy- thates by the termite. lootermopsi's ongiisricolfi's. Physi- ol. 2001. 16:123-128. Car-air. A.. and It. Mamet. 1966. Compilation of mass spectral data. Heyden A. Son. Ltd.. London. . Cawh. E. R.. and T. K. Kit. 1976. Properties ofcellu- lose and hgnocellulosic materials as substrates for my matic conversion processes. Biotechnol. Bioerg Symp. 6:95-123. heather. G. R.. all T. 1‘. Kit. 1974. Catabolism of aspen sapwood in Rericali'rernies flavi'pes (lsoptera: Rhi- notermitidae). Ann. Entomol. Soc. Am. 67:989-991. Geathaer. B. R. 8., C. L. Davis, and M. P. Dry-t.1%1. VOL. 45. 1Q3 21. 31. 32. 33. 35. 37. 39. Features of rumen and sewage sludge strains of cumin. lain itinerant. a methanol- and HrCOyutilizing species. Appl. Environ. Microbiol. 42:12-19. . ”berg. B. M. 1970. Protozoa associated with termites ad their role in digestion. p. 1-36. In K. Krishna and F. M. Weesner (ed.). Biology of termites. Academic Press. Inc.. New York. llqate. R. E. 1939. Experiments on the nutrition of Zoorrrniopsi's. Ill. The anaerobic carbohydrate dissimila- tion by the intestinal protozoa. Ecology ”:230-245. . llqate. R. E. 1943. Quantitative analyses on the cellu- lose fermentation by termite protozoa. Ann. Entomol. Soc. Am. 36:730-739. . lime. R. E. 1946. The symbiotic utilization of celluo lose. J. Elisha Mitchell Sci. Soc. 62:9-24. . lime. R. E. 1966. The rumenand its microbes. Aca- demic Press. Inc.. New York. . We. R. E. 1975. The rumen microbial ecosystem. Annu. Rev. Ecol. Syst. 6:39-66. . m. L. 1.. and H. Rh'by. Jr. 1939. The relative weights of termites and their protozoa. J. Parasitol. 25:444-445. . Raven. J. 1967. Presence d'acides gras volatils dans la panse d'un termite superieur (Microcerotermes edentatus Was. Amitermitinae). C.R. Acad. Sci. (Paris) 264:486- 488 . an}... r. 1967. Etude radiagraphique du transit intesti- rml chez un termite superieur. Experientia 23:820-821. . I“. S. S.. .d N. B. Sigh. 1968. SugaHlye move. men! through the alimentary canal of Odontotcrmes obe- sus (Isoptera: Termitidae). Ann. Entomol. Soc. Am. 61:230 . tar-é. r. r.. as w. L. Mating. ms. Nutrient dynam- ics oftermites. p. 165-232. In M. V. Brian (ed). Produc- tion ecology of ants and termites. Cambrithe University Press. Cambridge. England. 1..-PHI. J. P.. .d W. L. m. 1979. Respiratory gas exchange in the dry-wood termite. Margiiii’remes liab- Mrdi' (Banks) (Isoptera: Kaiotermitidae). Sociobiology 4:257—267. w. J. A.. F. Mayer. .d R. S. Wife. 1”]. Accrogeni- rim kit-iii. a new thermophilic hydrogen-oxidizing. aceto- ynic bacterium. Arch. Microbiol. 129:275-280. Mafia, J. It. 1%2. Lipid synthesis from ("Cl-acetate by two subterranean termites. Reticuli'iermes flow)»: and Coptatermes formosanus. Insect Biochem. 12:193-199. . Mafia. J. K., lad N. M. Rich. 1980. Eflect of chlortetn- cycline and other antibiotics on protozoan numbers in the eastern subterranean termite. J. Econ. Entomol. 73:123- 128. MeNel. N. 1.. J. H. C‘gs. .d W. P. 1‘. Jam. 1978. Short chain fatty acid absorption by the human large intestine. Gut 19:819-822. Myhre. D. V.. .d F. Sflh. 1960. Constitution of the hemicellulose of alfalfa (Medicago satire). Hydrolysis of hemicellulose and identification of neutral and acidic components. J. Agric. Food Chem. 8:359-364. Nehb. A. C. 1952. Analytical methods for bacterial fer- mentations. report no. 46-8-3 (2nd revision). National Research Council of Canada. Saskatoon. Saskatchewan. O'Brien. R. W.. m M. Slaytor. 1%2. Role of microorga- nisms in the metabolism of termites. Aust. J. Biol. Sci. 35:239—262. Peakh. G. J.. ad 6. Jeaeas. 1978. Respiration and energy flow. p. 111-163. In M. V. Brian (edJ. Production ecolo- gy of ants and termites. Cambridge University Press. Cambl'idec. England. 33 TERMITE HINDGUT FERMENTATION 41. 42. 43. 45. 47. 51. 52. 53. 55. 57. 59. 1613 Petrlas. C. J.. .d J. A. Breaaak. 1977. Nitrogen-fixing Eatrrobocrcr ogglomcmns isolated from guts of wood- eating termites. Appl. Environ. Microbiol. 33:392-399. Mas. C. J.. .0 J. A. Bree-alt. IMO. Uric acid in wood-eating termites. Insect Biochem. 10:19-27. Patties. C. J... all J. A. Brand. IMO. Uric acid-de- gradiru bacteria in guts of termites chtt'cuIi'trmres flo- v'pcs (Kollar)). Appl. Environ. Microbiol. “:117-124. Petr-Ros. C. J., .d J. A. Brenak. 1%0. Anaerobic deg- radation of uric acid by gut bacten'a of termites. Appl. Environ. Microbiol. 0:125-132. . Patric. C. J., .d J. A. Breznak. 1%]. Gut bacteria recycle uric acid nitrogen in termites: a strategy for nutrient conservation. Proc. Natl. Acad. Sci. U.S.A. 18:4601—4605. Piveatwich. c. D.. R. W. Jones. and M. S. CeIas. Ifll. Terpene biosynthesis by nasute termite soldiers (lsoptera: Nasutiterrnitinae). Insect Biochem. 11:331-336. . Prior. R. A.. and A. W. 1977. Synthesis of acetate from C0; in the cecum of some rodents. FEMS Microbiol. Lett. 1:255-258. Rapaon, W. 11. 1963. Cellulose from bleached wood pulp. p. 22-24. in R. L. Whistler (ed.). Methods in carbohy- drate chemistry. vol. 3. Academic Press. Inc.. New York. . Sehdtz. J. E.. .d J. A. Breznak. 1978. Heterotrophic bacteria present in hindguts of wood-eating termites [Reti- culi'tcrnies flavr'pes (Kollar)]. Appl. Environ. Microbiol. 35:930—936. . Schultz. J. E.. nil J. A. Breznak. 1979. Cross-feeding of lactate between Streptococcus lactis and Bacteroides sp. isolated from termite hindguts. Appl. Environ. Microbiol. 37:1206-1210. . Sada. S. 2.. and G. E. Die. IMO. Kinetics ofabsorption of volatile fatty acids from the ceca of domestic turkeys. Comp. Biochem. Physiol. 67Az231—237. Ulea. R. L. 1976. Anaerobic growth of a Rhodopseiido- mono: species in the dark with carbon monoxide as sole carbon and energy substrate. Proc. Natl. Acad. Sci. U.S.A. 73:3298-3302. Umbreit. W. W., R. [1. Bank. .d J. P. Sta-lee. 1964. Manometric techniques. Burgess Publishiru Co., Minne- apolis. Veivers, P. C.. A. M. Masca. R. W. O'Brien. ad M. flaytar. 1982. Digestive enzymes of the salivary glands and gut of Mastotermes darwiniensis. Insect Biochem. 12:35-40. . Weesaer. F. M. 1965. The termites of the United States. a handbook. The National Put Control Association. Eliza- beth. NJ. Walk. M. J. 1979. The rumen fermentation: a model for microbial interactions in anaerobic ecosystems. Adv. Mi- crob. Ecol. 3:49-77. . Y‘a. M. A. 1900. Cellulose metabolism by the termite Iagellate Trichomitopsis termopsidis. Appl. Environ. Mi- crobiol. 39:859-863. Vain. M. A. 1981. Cellulose metabolism by the flagellate Trichonympha from a termite is independent of endosym- biotic bacteria. Science 211:58-59. . Vamhi. M. A.. and W. Trager. 1979. Cellulolytic activity of an axenically-cultivated termite flagellate, Trichomitop- si's termopsidis. 1. Gen. Microbiol. 113:417-420. Zimmerman. P. R.. J. P. Greenberg. S. O. Wandiga. .d P. J. Crataen. 1982. Termites: a potentially large source of atmospheric methane. carbon dioxide. and molecular hy- drogen. Science 218:563-565. 8. {J O ARTICLE II NUTRITION AND GRONTN CHARACTERISTICS OF TRICHOHITOPSIS TERHOPSIDIS, A CELLULOLYTIC PROTOZOAN FROM TERHITES BY D.A. Odelson and J.A. Breznak 31+ 35 ABSTRACT Putatively axenic cultures of Trichomitogsis termopsidis 6057. isolated by H.A. Iamin (J. Protozool.‘g§:535-538. (1978)) from the hindgut of Zocterm0psis termites, apparently contained methanogenic bacteria, inasmuch as small amounts of Cl!“ were produced during growth. However, I. termopsidis could be 'cured' of methanogenic activity by incubation in the presence of bromoethanesulfonate. Both the cured derivative (6057C) and the parent strain (6057) required llallCO3 and fetal bovine serum for good growth; the presence of yeast extract in media was stimulatory. Growth of both strains was markedly improved by substituting heat-killed cells of Bacteroides sp. JHZO (a termite gut isolate) for heat-killed rumen bacteria in media as a source of bacterial cell material. Beat-killed Bacteroides JHZO was the best of a number of bacteria tested, and under these conditions Hz was a major protozoan fermentation product. Growth of 1. termopsidis strains was further improved by co-cultivation in the presence of Hethanospirillum hungatii. g. hungatii was the best of a number of B2 -consuming bacteria tested, and under these conditions 684, but not 32, was produced, indicating interspecies transfer of 82 between the protozoa and g. hungatii. Both strains of I. termopsidis used powdered, particulate forms of cellulose (eg. pure cellulose, corncob, cereal leaves) as fermentable energy sources, although powdered wood, chitin, or xylan supported little or no growth. Cells of the cellulose-forming coccus Sarcina ventriculi also served as a fermentable energy source, but these were used poorly as a source of bacterial cell material. The only substantial difference between I. 36 termogsidis 6057 and 60570 was that the latter grew poorly or not at all with rumen bacteria as a source of bacterial cell material. The improved growth of I. termopsidis in vitro, as described herein, should facilitate further studies on the cell biology and biochemistry of these symbiotic, anaerobic protozoa. .37 turnonucrron One of the most fascinating examples of nutritional symbiosis is the association between phylogenetically lower termites and their hindgut microbiota, which includes dense and diverse pepulations of flagellated protozoa and bacteria. Early studies by Cleveland (to-13), Hungate (21-23), Trager (no,n1), and others (19) indicated that many of the protozoa were cellulolytic and that their presence in . the hindgut was critical to termite survival on a diet of sound wood or cellulose. It now seems certain that protozoa are the major (if not sole) agents of wood cellulose hydrolysis in the hindgut of lower termites, and our current understanding of this symbiosis, including the roles of bacteria in termite nutrition, has been presented in several recent reviews (2-h,30). Despite the central role of protozoa in the nutrition of lower termites, our knowledge of their cell biology and biochemistry has remained meager. This is because for many years pure cultures have not been available. Consequently, many studies have been restricted to the morphology of these intriguing eukaryotes and their physical association with bacteria (5.33). The relatively few biochemical investigations which have been done have either employed mixed species of protozoa harvested directly from the termite hindgut, and which were undoubtedly contaminated to some degree with living bacteria (23), or have employed in gitgg cultures in which living bacteria were known to be present (”1). Recently, however, Yamin succeeded in obtaining axenic cultures of two cellulolytic species: Trichomitopsis termopsidis (“7) and Trichonympha sphaerica ("9). He showed that both 38 of these flagellates grew anaerobically and fermented cellulose to acetate, 82, and 002, although their growth in £11.32 was generally poor. To facilitate further biochemical studies on these protozoa we examined the nutrition and growth characteristics of one of Tamin's isolates, Trichomitopsis termopsidis, in an effort to improve its growth rate and yield. The results of that endeavor constitute the substance of this paper. A companion paper reports on cellulolytic and other hydrolytic enzyme activities oflg. termopsidis (32). (Portions of this work were presented at the 83rd Annual Meeting of the American Society for Microbiology, 6-11 March 1983, New Orleans, LA, and at the Third International Symposium on Microbial Ecologv, 7-12 August 1983, East Lansing, MI.) 39 MATERIALS urn muons Organism . Trichomitopsis termopsidis strain 6057 was obtained from M. A. Tamin (Rockefeller University, New York, NY). Bacteroides sp strain JH20, Bacteroides termitidis strain DAD-50, and Streptococcus lggtig strain JH1 were previously isolated from the hindgut of Reticulitermes flavipgs termites in our laboratory (35, 37). Bacteroides amylophilus strain 818, Bacteroides ruminicola strain 23. Eubacterium limosum strain RF, and Desulfovibrio sp. strain 611 were obtained from M.P.Bryant (0. of Illinois, Urbana, IL). Methanosarcina barkeri strain MS and Methanospirillum hungatii strain JPt were obtained from 8.8. Holfe (U. of Illinois, Urbana, IL), and Bacteroides fragilis strain’Vfl79, Bacteroides ovatus strain V211, and Sarcina ventriculi strain AL-z were obtained from R.B. Hespell (U. of Illinois, Urbana, IL). Media and cultivation conditions. Strict anaerobic techniques (20) were employed for preparation of media and for cultivation and manipulation of cells. I. termopsidis was routinely grown under 100 S nitrogen (initial gas phase) in 18 x 150 mm serum tubes equipped with black rubber stoppers and crimped aluminum caps (1) and containing 10.0 ml of medium. The culture medium was slightly modified from that of Yamin (h?) and contained (mM. unless indicated otherwise): KZHan, 10.8; KBZPQA’ 6.9; KCl, 21.5; NaCl, 2u.5; CaClz, 0.5; H3504, 5.3; Pfennig's metal solution (26), 0.1 3 (v/v); cellulose powder (Type 20, Sigma Chemical Co., St. Louis, M0), 0.1 1 (w/v); to glutathione (reduced form), 3.2; RaRCO3, 9.5; yeast extract (ION Biochemicals, Cleveland, Ohio). 0.2-0.11 3 (w/v); heat inactivated fetal bovine serum (Gibco Laboratories, Grand Island, NY), 2.5 1 (v/v); and heat-killed mixed rumen bacteria (RB; see below). 1 x 10.4 3 (w/v). To prepare this medium, a basal solution containing the inorganic salts, cellulose powder, and glutathione was tubed under N2 in 9.25 ml amounts and heat-sterilized. Remaining components were added to the sterile, cooled basal solution from filter-sterilized or heat-sterilized (i.e. RB) stock solutions. The pit of the medium prior to inoculation was 6.7-6.8, and inoculated tubes were incubated vertically and unshaken at zit-26' C. Additions to, or samples from. the tubes were made by using llz-flushed, sterile syringes equipped with 22 g or 25 g hypodermic needles (25). Stock cultures of I. termopsidis were maintained by transferring 5 1 (v/v) inoculum to fresh medium every 20-30 days. Culture purity was periodically verified by microscopic examination of wet mount preparations and by inoculation of cultures into BRIG broth (see below) amended with 0.3 3 (w/v) cellobiase, under which conditions I. termopsidis does not grow. RB were prepared according to the method of Tamin (ll?) and were stored under 100 1 R2 at -20’ C until needed. The source of RB was a hay-fed, fistulated dairy cow. Stock suspensions of RB contained approximately 1.2 x 10 1° cells/ml (equiv. 3.63 mg dry wt./ml). Individual bacterial strains tested as a replacement for RB were prepared in a similar manner, except that cell suspensions of §. ventriculi were also incubated for 5 min in an ultrasonic cleaner bath (Mettler Electronics) to disrupt large aggregates of the cocci. hi 2. amylophilus and _B_. ruminicola were grown at 37 'C in medium 0 10 as described by Caldwell and Bryant (6). Mew—s sp. JNZO, _B_. fragilis, _B_. M, _B. termitidis, and _s. m were grown at 30' C in brain heart infusion broth (Difco) supplemented with 0.3 1 (w/v) glucose and 0.05 I cysteine-8C1 (BRIG, 37). S. ventriculi was grown at 30 C in MIA medium as described by Canals-Parole (7). _E_. limosum, _l_l_. hunggti , and _tg. barkeri were grown at 30 C in the basal medium as described by McInerney gt a; (26) under a gas phase of 82:002, 80:20; Desulfovibrio sp. G11 was grown under the same conditions except that the medium was supplemented with 0.3 i (w/v) Na2804. Enumeration of bacteria was done microscopically by using a Petroff-Rauser counting chamber . Nutrition I Growth studies. Nutritional and growth characteristics of I. termopsidis were evaluated by determining their growth rate and/or cell yield, as well as gas production, in response to changes in medium composition or incubation conditions. Three to four replicates were done for each culture condition tested. Cell densities were determined by direct microscopic counts on samples drawn into 50 a long rectangular glass capillary tubes (i.e. microslides; Vitro Dynamics Inc., Rockaway, NJ). The volume of each microslide was approximately 1.25 pl (Cat.# 5005) or 5 fl (CatJ 5010), but was determined precisely for each lot by filling representative samples with a standard solution of l4C—toluene and determining the radioactivity present. n2 Growth offlz. termogsidis was also monitored semi-quantitatively by examining cells, directly in the culture tubes, with an inverted microscope. The method was less time-consuming than preforming direct cell counts (see above) and helped insure that accurate cell yield determinations were centered about the stationary phase of growth. This strategy thereby minimized unnecessary sampling of cultures and the risk of exposure of cells to oxygen. Co-cultivation of I. termopsidis with viable bacteria was done by inoculating modified Tamin's medium with a 3 to 5 1 (v/v) inoculum of each of 1. termopsidis (eg. a 20-30 day culture containing about 8000 cells/ml) and the bacteria (eg. for M. hungatii a 3-5 day broth culture displaying an C.D. (600 nm) of 0.1). Analysis of metabolic products. Hz and CR4 present in the headspace of T. termopsidis cultures were analyzed by gas chromatography. For 82 a column of Molecular Sieve 5A (Raters Associates Inc.) was used with thermal conductivity detection (”2). For CE“ a column of Porapak N (Haters Associates Inc.) was used with R2 flame ionization detection (3n). Acetate present in cell-free culture fluids was determined enzymatically by measuring the oxidation of NADB in the presence of the following enzymes: acetleoA sythetase; pyruvate kinase; myokinase; and lactate dehydrogenase (1?). Reducing sugar was determined by the colorimetric ferricyanide assay with glucose as the standard (33). Absorbance readings were determined by using a Gilford spectrophotometer. #3 Cell free culture fluid was screened fer the presence of organic acids other than acetate by using the method of Salanitro and Muirhead (36), and for the presence of ethanol by distillation of clarified culture fluid (29) followed by gas chromatographic analysis of the distillate on a column of Porapak Q with 82 flame ionization detection (3h). Clarified culture fluid was also screened for products of cellulose hydrolysis by using high pressure liquid chromatography as previously described (29). Other procedures. Methanogenic bacteria were examined microscopically by using a Leitz epifluorescense microscope according to the procedure of Doddema and Vogels (16). Photomicrographs of I. termopsidis were taken by using a Zeiss GEL phase-contrast microscope equipped with a C-35 camera attachment and Tri-x (Kodak Co., ASA 900) film. Substrates and chemicals. Cereal leaves (Sigma Cat.# c.71u1), chitin (Sigma Cat.# C-3387), and Riley-milled corncob (obtained from R.B. Respell) were further dry ball-milled for 2h-72 hours at ambient temperature to produce fine powder. Douglas fir wood powder was prepared as previously described (31). A crude mixture of cellodextrins (containing cellobiose to cellohexaose) was prepared by acid hydrolysis of cellulose powder followed by column chromatography (2?). Coenzyme M was isolated from Mbthanobacterium bryantii and was a gift from R. L. Uffen. All other chemicals were of analytical rea- gent grade and were obtained from commercial sources. 14C-toluene calibration standard was obtained from New England Nuclear, Boston,MA. RESULTS General growth characteristics and nutrition of T. termopsidis 6051. I. termOJsidis 6057 grew in modified Tamin's medium with a doubling time of 9.5 to 7.5 days and achieved final yields of 3800 to 7600 cells/ml after 30 days incubation. Growth of cells was accompanied by the production of acetate and 82 (Fig. 1), and presumably 002(n8). No other organic acids or ethanol were found. Surprisingly, cultures of T. termopsidis 6057 also evolved CR“. Evolution of CR accompanied that of B during exponential growth, but 4 2 subsided as cells approached stationary phase. By contrast, R2 evolution continued during stationary phase (Fig. 1). Consequently, the final concentration of B2 was generally 7 to 10-fold greater than that of CEa. The amount of CRa evolved was not increased by supplementing the initial gas phase of cultures with 80 1 82 or by ~..-__---, ~ -_ incorporation of up to 5 pH coenzyme M in the medium. The probable origin of'CR4 in cultures offil. termopsidis 6057 is discussed below. Nith the culture conditions employed (static, vertical culture tubes), 1. termopsidis settled to the bottom and grew on and amongst a soft pellet of cellulose particles and RB. Incubation of cultures horizontally, with or without agitation, was detrimental to growth. Consequently, since agitation was required to resuspend cells uniformly for making direct cell counts, each point on the curves in Fig. 1, 2, and 3 was the mean value of 3-9 independent, and different, cultures. All values were within 5-10 1 of the mean. Figure 1 . Growth of T. termopsidis 6057 in modified Yamin's medium at 25’ C. Each point is the mean value of 3-ll independ- ant, and different, cultures. “5 /ML CELLS _o_n nMOL ES YA MIN '8 MED. Figure 1. DAYS u? A significant improvement in growth of strain 6057 was achieved by replacing RB with heat-killed cells of Bacteroides strain JNZO, a strain originally isolated from termite guts (37). By using ‘gacteroides JNZO cells at a concentration of 0.1-0.2 mg dry wt./ml of medium, I. termopsidis grew with a doubling time of 9.0 to 9.5 days and reached final yields of 10,900 to 31,900 cells/ml (Fig. 2). Cell-free filtrates of heat-killed Bacteroides JN20 could not replace the cells themselves. However, when Bacteroides JN20 was used in place of RB little or no 0114 ((1 nmole/ml culture) was evolved by protozoan cultures, and Bzwas the only reduced gas formed. Even more striking was the observation that I. termopsidis could be passed through more than 10 transfers in this medium (approx. 1 year) without producing cut, but would resume methanogenesis immediately if shifted back to a medium containing RB. Qgigin of methanogenesis in cultures of T. termopsidis 6057. Methane emision by 1. termopsidis 6057 was entirely unexpected, inasmuch as methanogenesis is a bacterial phenomenon (95) and Yamin (97) reported the original cultures of I. termopsidis to be free of living bacteria. Nevertheless, our results suggested that cultures of strain 6057 harbored a methanogenic consort, since uninoculated media, as well as heat-sterilized cultures, did not produce methane. However, epifluorescence microscopic examination of living or glutaraldehyde—fixed cultures failed to reveal F920 fluorescent cells (i.e. putative methane bacteria) associated with the protozoa or free in the culture fluid. By contrast, F920 fluorescent cells were readily observed in bovine rumen contents and by examination of pure Figure 2. Growth of T. termopsidis 6057 in modified Yamin's medium with heat-killed cells of Bacteroides sp. JH20 in place of heat-killed rumen bacteria. Other conditions were as for Figure 1. 98 4 1 0 -RB; +JW20 /' /ML -A O 0) MY . //- V‘CETATf/ -— / CELLS 95 nMOLES 8” \ O '10 20 30 Figure 2. DAYS 99 SC) cultures of methane bacteria (g. hu_nggtii and N. barkeri). Moreover, we were unable to successfully enrich for, or isolate, methanogens from cultures of T. termopsidis 6057 by using a variety of substrates (ll.Z + C02; acetate; formats; or methanol), although methanogens were readily isolated from termite gut homogenates and from putatively axenic cultures of Trichonygpha sphaerica (another cellulolytic protozoan isolated from termite hindguts, (99)) by these same procedures (Odelson and Breznak, unpublished). Methanogenesis by I. termopsidis 6057 could be permanently abolished, however, by growth of cells in the presence of 0.5 pH bromoethanesulfonate (BES), a potent and specific inhibitor of methanogenesis (50). As little as one transfer of cells in medium containing BES was sufficient to accomplish this. The resulting culture was then designated 6057C, the letter 'C' denoting the fact that cells were “cured“ of their putative methanogenic consort. Strain 6057C grew poorly, however, in culture medium with RB as the source of heat-killed bacteria; cell yields reached only 500-1000 cells/ml. Replacement of RB with.heat-killed Bacteroides JN20 improved the growth of strain 60570 in much the same way as it did for the uncured parent strain (6057). Nith Bacteroides JN20, strain 6057C exhibited a doubling time of 2.6 to 5.2 days and achieved yields of 5,600 to 22,000 cells/ml. Having established suitable growth conditions for 1. termopsidis 6057C, we sought to extend the nutritional studies initiated by Yamin (97) and to compare, where possible, the nutritional characteristics of strain 6057C to those of strain 6057. 51 Evaluation of soluble medium constituents and heat-killed bacteria. 2. termopsidis 6057C required fetal bovine serum and wancoa for good growth (Table 1). lo significant difference in cell yields (at p‘ 0.05) was observed by omission of yeast extract. However, cell yields in the absence of yeast extract were significantly lower than that of the control at p‘0.1 level (Student's ‘1' test). Therefore, yeast extract was Judged to be stimulatory. Strain 6057C grew poorly or not at all if heat-killed cells of Bacteroides JHZO were omitted from the medium, and heat-killed cells ofjg. termitidis, g. amy10philus, g. fragilis, g. m, g. ruminicola, or §. la_ctis_, could not replace Bacteroides JH20. Hydrogen production by strain 6057C closely paralleled cell yields (Table 1). Similar results were obtained with strain 6057. Evaluation of insoluble polysaccarides. Insoluble polysaccharides, in both purified and natural forms, were tested for their ability to support growth of I. termopsidis 60570 (Table 2). Sigma type 20 cellulose powder supported the best growth and hydrogen production . Hhatman microgranular cellulose, powdered corncob, and powdered cereal leaves also supported growth, but to a lesser extent. Powdered chitin and powdered wood supported little or no growth of the protozoa, and xylan appeared detrimental. Soluble intermediates or products of cellulose hydrolysis (i.e. glucose, cellobiose, cellodextrins), or 10 mg discs of cellulose filter paper (Hhatman #1), could not replace powdered particulate forms of cellulose as a fermentable growth substrate (data not shown). Similar results were obtained with I. termopsidis 6057. 52 Table 1 Effect of medium components on growth of I. termopsidis strain 6057C. Yield 2 Omitted from Hediuml Cells/ml nHoles Hzlml culture lo Omission (Control) 5600 g 1100 "67 z’23 least Extract 2700 :_1h00 2&9 3 102 NaBCO3 1300 3 I£00 9'3 1 38 Fetal Bovine Serum (200 0 Bacteroides JU20 (200 0 1Growth medium was modified Iahin's medium containing heat-killed Bacteroides .1320 in place of heat-killed rumen bacteria. The initial pH of all media was 6.7 3 0.2. 2Determined 30 days after the second transfer in the medium. Values are the mean t SE)! (mu). 53 Table 2 Growth of I. termopsidis strain 60570 on insoluble polysaccharides. Tieldz Polysaccharide Testedl Cells/ml you: emu Culture Sigma Type 20 Cellulose 6100 g 1600 601 :_91 Hhatman Hicrogranular Cellulose 2900 g 700 396 1 109 Corncob 3900 3 900 939 1.90 Cereal Leaves 1700 :_100 208 :_73 Hood 900 _+_ 300 251 g; 50 Chitin 700 3 200 83 z 50 Xylan - 0 0 No Substrate 600 g 600 69 3,93 1 Growth medium was modified Yamin's medium containing heat-killed Bacteroides JH20 in place of heat-killed rumen bacteria. Polysaccharides were incorporated in the medium at a final concentration of 0.1 9 (w/v). 2 Determined 30 days following inoculation into test medium. Values are the mean 3 SEN (nz9). 59 Co-culture of T. termopsidis with known Hz-consuming bacteria. A curious feature of I. termopsidis 6057C, and of strain 6057 growing with heat-killed Eacteroides “20 in place of RE, was the fact that H 2 appeared to be the only reduced end product formed (eg. see Fig. 2). The final molar ratio of szicetate was approximately. 2:1, and no other volatile fatty acids, lactate, succinate, or ethanol were detected in culture fluid. It seemed likely that some of this Hz was derived from the oxidation of reduced pyridine nucleotides (eg. HAD(P)H + H"), a reaction thermodynamically unfavorable at PHZ’IO'"3 atm (96). Therefore, on the notion that H2 accumulation might be suppressing the growth of I. termopsidis, cells were grown in the presence of a variety of potential electron acceptors. However, no stimulation of growth was observed by including in the medium 1‘. ,bxaloacetate, fumarate, K1103 .fructose, triphenyltetrazolium chloride, or neutral red (singly at 9 at! final concentration). Moreover, growth of cells was completely inhibited by as little as 0.01 atm 02 . Attempts were then made to keep the p82 low by co-cultivation of I. termopsidis with anaerobic, Hz-consuming bacteria. Co-cultivation of I. termopsidis 6057C (in modified Yamin's medium containing heat killed Bacteroides JH20) with viable g. hunggtii increased protozoan yields to 25,000 to 96,000 cells/m1, but gave little change in dou- bling time (3.6 to 9.2 days). is expected, CH4, but not Hz, was evolved, and the final molar ratio of C84zacetate was approximately 1:2. Similar results-were obtained with strain 6057, which displayed final yields of 58,000 to 75,000 cells/ml and a doubling time of 2.2- 5.3 days (Fig. 3). Growth of, and methanogenesis by, g. hungatii was Figure 3. Growth of 1. termopsidis 6057 in co-culture with Hethanospirillum hunjatii (Hsp). Growth medium and conditions were as for Figure 2. 55 ascrmm o:a \ z... P S M + o 2 w J + B, R - ._s_\m:.<»mo< 3.5:: 3. 3.50 DAYS Figure 3. 56 57 dependent on growth of the I. W strains; and heat-killed cells of _IQ. hmtii did not improve growth of the protozoa. Co-cultivation with g. hmtii did not alter the nutritional requirements of I. termopsidis strains for soluble medium consti- tuents, heat-killed bacteria, or fermentable polysaccharides as described above. Co-cultivation with Hethanosarcina barkeri also improved growth of I. termcmsidis strains, but not to the same extent as did 33. hmti , probably because )1. barkeri did not scavenge all of the H2 produced by the protozoa. By contrast, co-culture of I. termopsidis with Eubacterium limosm (an Hz—COZ-utilizing acetogen (18)) did not improve protozoan growth, and co-culture with Desulfovibrio G11 appeared to suppress growth, possibly because of accumulation of toxic levels of 828. Co-cultures of T. termopsidis with _11. hu_nga_tii grew best with the reducing agent normally present in modified Tamin's medium (i.e. glutathione). Replacement of glutathione with cysteine-HCl (0.05 1, w/v), a combination of cysteine-HCl and Hazs (0.025 1 each, w/v), or dithiothreitol (1.5 ‘4) inhibited growth of 1. W1: in co-culture. Glucose, cellobiase, or cellodextrins (or an increase in total reducing sugars) was not detected in extracellular fluid during growth of I. termopsidis. 58 Growth of T. termopsidis on Sarcina ventriculi. In an effort to identify a single particulate substrate that would serve as both a fermentable energy source (i.e. cellulose) and a source of bacterial cell material, attempts were made to grow 1. termopsidis on heat-killed cells of the packet-forming coccus Sarcina ventriculi. _S_. ventriculi was chosen because it synthesizes cellulose as a uniform, outer layer that accounts for as much as 19 1 of the cells' dry weight (8, 9). In co-culture with viable g. hunggtii, 1. termopsidis 6057C used _S_. ventriculi simultaneously as a fermentable energy source and as a source of bacterial cell material (Table 3). However, final yields of I. termopsidis under these conditions (1500 cells/ml) were too low to be of practical use. This appeared due mainly to the fact that §. ventriculi was inferior to Bacteroides JH20 as a source of bacterial cell material (Table 3). By contrast, S. ventriculi was entirely effective as a replacement for powdered cellulose alone (in the presence of heat-killed Bacteroides JH20) and significantly (p‘0.05, Student's T Test) boosted cell yields of _1_'. termopsidis 60570 when added to the complete medium (Table 3). The data also suggested that cell carbon from S. ventriculi could ultimately be used as a substrate for methane formation by g. hungatii. Phase contrast microscopy revealed that I. termopsidis endocytosed packets of the sarcinae; appeared to hydrolyze the outer cellulose layer; then apparently expelled §. ventriculi as single or doublet, denuded (phase dark) cells which accumulated in the medium during growth of the protozoa (Fig. 9). 59 Table 3 Growth of'z. termopsidis 60576 on heat-killed cells of §. ventriculi. Iieldz Omitted from Added to nMoles CH4 Mediuml’ Medium Cells/ml ml culture -- - (Control) 36,900 1.10,9OO 1982 1 139 Cellulose SV (1.0 mg dry wt. 39,200 1,7900 1010 g 139 x ml'l) JHZO SV (0.1 mg dry wt. 9300 1.1200 182 :_36 x ml'l) Cellulose 37 (1.0 mg dry wt. and szo x ml'l) 1500 1 1500 109 1 16 -- sv (0.1 mg dry w’t. 51,800 1 2500 1963 1 86 x ml’l) Cellulose -- 900 :_200 157 3,16 szo -- 200 1 200 33 1 12 1 Protozoa were grown in co-culture with M. hunggtii in modified Yamin's medium containing heat-killed Bacteroides JH20 in place of heat-killed rumen bacteria. Symbols: SV, Sarcina ventriculi; JH20, Bacteroides JH20. 2 Determined after 31 days incubation.7alues are the mean 1 SEM (n=9). Figure 9. Phase contrast photomicrograph.of a cell of I. termopsidis 6057C growing in the presence of heat-killed1§. ventriculi. Note the cellulose-containing (refractile) sarcinae (black arrows) outside, as well as within, the protozoan, and the phase-dark (presumably cellulose-depleted) sarcinae (white arrows). The latter accumulate in the medium during growth of I. termopsidis. 6C) Figure 9. 62: DISCUSSION In li_t_:_‘9_ growth of I. termopsidis strains in modified Tamin's medium was markedly improved by replacement of the heat-killed rumen bacteria component with heat-killed cells of Bacteroides sp. strain JH20 (a termite gut isolate (37)). and by co-cultivation in the presence of an HZ-consuming methanogen, _M. hmtii. Under these conditions, the growth yield of‘g. termonsidis was increased from about 7600 cells/ml to 75,000 cells/ml, and the doubling time was decreased from 9.5 days to 2.2 days. Fetal bovine serum and HaH003 were found to be required for good growth, and yeast extract was Judged to be stimulatory. Nutritional and growth characteristics of I, termopsidis 6057 (the parent strain) were similar to those of the strain 60570 (the cured derivative), except that the latter showed little or no growth with mixed rumen bacteria as a source of bacterial cell material. Our results are consistent with those of Yamin (97) who showed that I. termopsidis requires bacteria in media as a nutrient. However, like Tamin found, not all bacteria were nutritionally equivalent. In fact, of a number of Bacteroides strains tested in this study, only strain JH20 was capable of supporting growth of 1. termopsidis. A decided preference for certain bacteria has also been observed fer aerobic (39) as well as other anaerobic protozoa (15). The precise role of bacterial cell material in the nutrition of 1, termopsidis is not yet known. Our results suggest that even bacteria of the same genus (i.e Bacteroides) differ significantly in their intrinsic nutritive value for I. termopsidis. However, it may 63 also be that some bacterial strains are simply more easily endocytosed than others, and are therefore more readily available for intracellular’metabolism. Nevertheless, while endocytosis of bacteria ‘may be a necessary prerequisite for their utilization by‘z. termopsidis, it is not alone sufficient: Sarcina ventriculi cells were readily endocytosed by the protozoa (Fig. 9), and were used as a fermentable source of cellulose, but §. ventriculi was a rather poor replacement for Bacteroides JH20 cell material (Table 3). Our results confirm and extend Iamin's (97) observation that a particulate source of cellulose is required as a fermentable energy source by I. termopsidis. Powdered cellulases, corncob, or cereal leaves supported growth, whereas chitin or xylan alone did not. Soluble intermediates of cellulose hydrolysis (i.e. cellodextrins) were also ineffective. Surprisingly, wood, the presumed natural substrate for I. termopsidis in situ, was used poorly or not at all. Either one or a number of the following explanations for this observation are possible: (i) heat-sterilization of wood particles adversely affects their utilization by I. termopsidis; (ii) the suitability of wood as a substrate for I. termopsidis depends on physical and/or chemical (i.e. enzymatic) pre-treatment, such as might be afforded by passage through the termite gut; or (iii) most of the wood particles obtained by ball-milling were still too large for efficient endocytosis by 1. termopsidis. In this regard it is pertinent to note that mixed cultures of rumen protozoa cannot grow on wheat grain that has been sterilized by heat or by treatment with 69 ethylene oxide (19). Moreover, it should be emphasized that I- termopsidis must endocytose oellulosic materials to digest them: no growth was obtained if pure cellulose was included in media as 1 cm discs of filter paper. The apparent presence of'methanogenic bacteria in Tamin's original culture of I. termopsidis (strain 6057) was unexpected, but perhaps not surprising. Methane bacteria (identified by their fluorescence when illuminated with ultraviolet light) are commonly observed on and within cells of rumen protozoa (99) and sapropelic protozoa (93), and their presence in termite guts is indicated by the fact that numerous termite species evolve methane (2) and such bacteria can be isolated from termite guts (Reticulitermes flavi s; Odelson and Breznak, unpublished). Moreover, the treatment used by Iamin (97) to render I. termopsidis 6057 putatively "bacteria-free“ involved incubation of protozoa in the presence of penicillin and streptomycin. These drugs, while effective against many eubacteria, are ineffective against methanogens which are archaebacteria (1). That we failed to successfully enrich for or isolate methanogens from I. termopsidis cultures, and also failed to observe methanogens by fluorescence microscopy of protozoan cultures, suggests that the- methanogens are present in small numbers in cultures of I. termopsidis 6057 and may also require for growth a nutrient provided by the protozoa. Indeed, the long term persistence of methanogens in cultures of 1. termopsidis 6057 grown on Bacteroides JHZO, during which incubation no methane is detected, leads one to speculate that the endogenous methanogens may be capable of functioning in part as intracellular "energy parasites" similar to chlamydiae (28). 65 Ievertheless, I. termopsidis 6057 could be completely and permanently cured of’methanogenic activity (and presumably of the methanogenic cells as well) by cultivation in the presence of bromoethanesulfonate, a potent and specific inhibitor of’methanogenesis (50). At the present time, we do not know what substrate is used by the endogenous methanogen of I. termopsidis 6057 for methanogenesis, but it is probably H2'+ 005, or acetate, or a mixture of these, since these are the only significant products formed by I. termopsidis from cellulose fermentation. In any case, hydrogen was the major or sole reduced end product of strain 6057 or strain 6057C, respectively. This observation suggested that some of the hydrogen produced by‘z. term0psidis was derived from reduced pyridine nucleotide via HAD(P)H + H+ -r HAD(P)+ + H2 , a reaction thermodynamically unfavorable at 982 > 10"3 atm (96). Accordingly, this prompted an attempt to improve the growth of I. termopsidis by making cellulose fermentation more favorable energetically (i.e. by keeping the sz low through co-cultivation with H2-consuming bacteria). This was successful, and _M. hungatii proved to be the most effective of all Hz-consumers tested. Under these conditions CH“, but no Hz, was produced. To our knowledge, this is the first in vitro demonstration of H2 transfer between protozoa and methanogens, and it suggests that interspecies transfer of hydrogen between protozoa and bacteria is important in "pulling“ anaerobic decomposition of cellulose in the hindgut of lower termites. 66 It is difficult to predict whether I. termopsidis might ultimately prove useful in anaerobic bioconversion schemes designed to produce food, fuels or chemical feedstocks from lignocellulosic substrates. Nevertheless, the improved growth of I. termopsidis in .11552, as described herein, should facilitate further studies on the cell biology and biochemistry of these unique eukaryotes. This is exemplified in a companion paper (32) which reports on studies of cellulase and other hydrolytic enzyme activities of this symbiotic protozoan. 67 ACKNOWLEDGEMENTS He are grateful to M.A. Tamin for providing the original cultures of I. termopsidis and I. sphaerica and for many helpful discussions. He also thank M.J. Klug for help with analysis of butylated volatile fatty acids and R.H. O'Brien for preparation of cellodextrins. This work was supported 'in part by National Science Foundation grants PCM 80-12026 and PCM 83-09367; by the College of Osteopathic Medicine of Michigan State University; and by the Michigan Agricultur- al Experiment Station of Michigan State University. 1. 2. 10. 11. 12. 13. 68 LIST OF REFERENCES Belch, H.E., G. E. Fox, L. J. Magrum, C. R. Hoese, and R.S. Rolfe. 1979. Methanogenszreevaluation of a unique biological group. Microbial. Rev.‘§3:260-296. Breznak, J. A. 1975. Symbiotic relationships between termites and their intestinal microbiota, p. 559-580. In D. H. Jennings and D. L. Lee (eds.), Symbiosis (Society for Experimental Biology Symposium Series #29). Cambridge Univ. Press, Cambridge, England. Breznak, J. A. 1982. Intestinal microbiota of termites and other xylophagous insects. Annu. Rev. Microbiol. 36:323-393. Breznak, J. A. 1983. Biochemical aspects of symbiosis between termites and their intestinal microbiota, p.173-203. In J. M. Anderson, A. D. M. Rayner, and D. Halton (eds.), Invertebrate-Microbial Interactions. Cambridge Univ. Press, Cambridge, England. Breznak, J. A. 1989. Hindgut spirochetes of termites and Cryptocercus punctulatus, p. 67-70. In H. R. Krieg and J. G. Holt (eds.), Bergey's Manual of Systematic Bacteriology. Hilliams and Hilkins, Baltimore, MD. Caldwell, D.R., and M. P. Bryant. 1966. Medium without rumen fluid for non-selective enumeration and isolation of rumen bacteria. Appl. Microbiol. 19:799-801. Canale-Parola, E. 1970. Biology of the sugar-fermenting sarcinae. 880te Rev. 35: 82-970 Canale-Parola, E.. R. Borasky, and R. S. Wolfe. 1961. Studies on Sarcina ventriculi: III. Localization of cellulose. J. Bact. _B_1: 31 1-31‘6. Canale-Parola, E.. and R. S. Wolfe. 1969. Synthesis of cellulose by Sarcina ventriculi. Biochim. Biophys. Acta..§g:903-5. Cleveland, L. R. 1923. Symbiosis between termites and their intestinal protozoa. Proc. Natl. Acad. Sci. U.S.A.‘2:929-8. Cleveland, L. R. 1929. The physiological and symbiotic relationships between the intestinal protozoa of termites and their host, with special reference to Reticulitermes flavipes Kollar. Biol. Bull. 96:178-227. Cleveland, L. R. 1925. The ability of termites to live perhaps indefinitely on a diet of cellulose. Biol. Bull. 98:289-293. Cleveland, L. R. 1925. The effects of oxygenation and starvation on the symbiosis between the termite, Termo sis, and its intestinal flagellates. Biol. Bull. 98:309-32 . 19. 15. 16. 17. 18. 19. 20. 21. 22. 23. 29. 25. 26. 69 Coleman, G. 8., 1980. Rumen ciliate protozoa. Adv. Parasitol. ‘1§:121-173. Coleman, G. 8., and D. C. Sandford. 1979. The engulfment and digestion of mixed rumen bacteria and individual bacterial species by single and mixed species of rumen ciliated protozoa grown in vivo. J. Agric. Sci. 22:729-792. Doddema, H. J., and G. D. Vogels. 1978. Improved identification of methanogenic bacteria by fluorescence microscOpy. Appl. Environ. Microbiol.I3§:752-759. Dorn, M., J. R. Andreesen, and G. Gottschalk. 1978. Fermentation of fumarate and L-malate by Clostridium formicoaceticum. J. Bact. 33:26-32. Genthner, B. R. S., C. L. Davis, and M. P. Bryant. 1981. Features of rumen and sewage sludge strains of Eubacterium limosum, a methanol- and H —CO2 utilizing species. Appl. Environ. Microbiol. 92:1 -19. Gutierrez, J. 1956. The metabolism of cellulose-digesting, symbiotic flagellates of the genus Trichonygpha from the termite Zootermopsis. J. Protozool.'3:39-92. Holdeman, L. V., and H. E. C. Moore (ed.). 1975. Anaerobe Laboratory Manual. Virginia Polytechnic Institute and State University, Blacksburg, VA. Hungate, R. E. 1938. Studies on the nutrition of Zootermopsis. II. The relative importance of the termite and the protozoa in wood digestion. Ecology. 12:1-25 Hungate, R. E. 1939. Experiments on the nutrition of Zootermopsis. III. The anaerobic carbohydrate dissimilation by the intestinal protozoa. Ecology. 29:230-295. Hungate, R. E. 1993. Quantitative analysis on the cellulose fermentation by termite protozoa. Annls. Ent. Soc. Amer. 31:730-739. Ladish, M. R., A. L. Huebner, and G. T. Tsao. 1978. High-speed liquid chromatography of cellodextrins and other saccharide mixtures using water as the eluent. J. Chromatog. 191:185-193. Macy, J. M., J. E. Snellen, and R. E. Hungate. 1972. Use of syringe methods for anaerobiosis. Am. J. Clin. Nutr. 25:1318-1323. McInerney, M. J., M. P. Bryant, and N. Pfennig. 1979. Anaerobic bacterium that degrades fatty acids in syntrophic association with methanogens. Arch Microbiol. 122:129-135. 27. 28. 29. 31. 32. 33- 3n. 35- 36. 37- 38. 39. 70 Miller, G. L., J. Dean, and R. Blum, 1960. A study of methods for preparing oligosaccharides from cellulose. Arch. Biochem. Moulder, J. N. 1989. Chlamydiales, p 729-739. In N. R. Krieg and J. G. Holt (eds.), Bergey's Manual of Systematic Bacteriology, V. I. Uilliams and Wilkins, Baltimore, MD. Heish, A. C.. 1952. Analytical methods for bacterial fermentations, report no.96-8-3 (2nd Revision) National Research Council of Canada, Saskatoon, Saskatchewan. O'Brien, R. R., and M. Slayter. 1982. Role of’microorganisms in the metabolism of termites. Aust. J. Biol. Sci. 35:239-262. Odelson, D. A.. and J. A. Breznak. 1983. Volatile fatty acid production by the hindgut microbiota of xylophagous termites. Appl. Environ. Microbiol. 95:1602-1613. Odelson, D. A.. and J. A. Breznak. 198_. Cellulolytic and other polymer—hydrolyzing activities of Trichomitopsis termopsidis, a symbiotic protozoan from termites. Appl. Environ. Micobiol. Park, J. T., and M. J. Johnson. 1999. A submicrodetermination of glucose. J. Biol. Chem. 181:199-151. Potrikus, C. J., and J. A. Breznak. 1977. Nitrogen-fixing Enterobacter agglomerans isolated from guts of wood-eating termites. Appl. Environ. Microbiol. 33:392-399. Potrikus, C. J., and J. A. Breznak. 1980. Uric acid-degrading bacteria in guts of termites (Reticulitermes flavaes (Kollar)). Appl. Environ. Microbiol. 99:117-129. Salanitro, J. P., and P. A. Muirhead. 1975. Quantitative method for the gas chromatographic analysis of short-chain monocarboxylic and dicarboxylic acids in fermentation media. Schultz, J. E., and J. A. Breznak. 1978. Cross-feeding of lactate between Streptococcus lactis and Bacteroides sp. isolated from termite hindguts. Appl. Environ. Microbiol. 31:1206-1210. Tamm, S. L. 1982. Flagellated ectosymbiotic bacteria propel a eucaryotic cell. J. Cell Biol. 29:697-709. Taylor, H. D., and J. Berger. 1976. Growth responses of cohabiting ciliate protozoa to various prey bacteria. Can. J. 2001. 51:1111-1119. 90. 91. 92. 93. 99. 95. 96. 97. 98. 99. 50. 71 Trager, H. 1932. A cellulase from the symbiotic intestinal flagellates of termites and of the roach, Cryptocercus punctulatus. Biochem. J.‘g§:1762-71. Trager, U. 1939. The cultivation of a cellulose-digesting flagellate Trichomonas termopsidis and of certain other termite protozoa. Biol. Bull. 66:182-190. Uffen, R. L., 1976. Anaerobic growth of Rhodopseudomonas species in the dark with carbon monoxide as sole carbon and energy source. Proc. Natl. Acad. Sci. U.S.A. 13:3298-3302. van Bruggen, J. J. A., C. I. Stumm, and G. D. Vogels. 1983. Symbiosis of methanogenic bacteria and sapropelic protozoa. Arch. Microbiol. 136:89-95. Vogels. G. D., H. F. Hoppe, and C. K. Stumm. 1980. Association of methanogenic bacteria with rumen ciliates. Appl. Environ. Wolfe, R. S. 1979. Methanogenesis. p. 270-300. In J. R. Quayle (ed.), International Review of Biochemistry, Microbial Biochemistry, V. 21, University Park Press, Baltimore, MD. Holin, M. J., 1982. Hydrogen transfer in microbial communities. p.323-356. in A. T. Bull and J. H. Slater (eds.). Microbial Interactions and Communities, Vol I. Academic Press. NY NY. Iamin, M. A.. 1978. Axenic cultivation of the cellulolytic flagellate Trichomitopsis termopsidis (Cleveland) from the termite Zootermopsis. J. Protozool.‘g§:535-538. Iamin, M. A.. 1980. Cellulose metabolism by the termite flagellate, Trichomito is termopsidis. Appl. Environ. Microbiol. 32:859-863. Tamin, M. A.. 1980. Cellulose metabolism by the flagellate Trichonygpha from a termite is independent of endosymbiotic bacteria. Science. 211:58-59. Zinder, S. H., T. Anguish, and S. C. Cardwell. 1989. Selective inhibition by 2-bromoethanesulfonate of methanogenesis from acetate in a thermophilic anaerobic digestor. Appl. Environ. Microbiol. 91: 1393-5 . ARTICLE III CELLULOLITIC AND OTHER POLIMER-EYDROLIZING ACTIVITIES OE TRICHOMITOPSIS TERMOPSIDIS, ‘A SYHBIOTIC PROTOZOAN FROM TERMITES. By D. A. Odelson and J. A. Breznak 72 73 ABSTRACT Crude extracts of the anaerobic, cellulolytic protozoan Trichomitopsis termopsidis possessed endo-p-1,9-glucanase and cello- biase activities, as evidenced by hydrolytic action on carboxymethyl cellulose and cellobiase, respectively. Cell extracts also hydrolyzed microcrystalline cellulose, although the extent of saccharification of this substrate was only 19.9 9 of the theoretical maximum, suggesting that cells may lack a complete cellulase enzyme system. Hydrolysis of microcrystalline cellulose displayed aptima at pH 5 and at 30"C, and glucose was the sole product liberated. Cellulolytic activities of T. termopsidis appeared to be entirely cell-associated. Hydralytic activity was also detected against Douglas fir wood powder, xylan, starch and protein, but not chitin. The importance of these enzymes in the nutrition afjg. termopsidis is discussed in terms of their natural habitat (the hindgut of wood-eating termites). 79’ INTRODUCTION Phylogenetically lower termites depend on dense and diverse populations of hindgut protozoa-and bacteria in order to thrive by xylophagy, and our current understanding of this nutritional symbiosis has been recently reviewed (9-6,19). Among this heterogeneous microbial community it appears that anaerobic, flagellate protozoa are the major, if not sole, agents of wood cellulose hydrolysis: they are not only abundant in the hindgut, but they also have the ability to endocytose, and thereby sequester, wood particles for intracellular degradation. Major products of cellulose fermentation by mixed and axenic suspensions of the protozoa include H2, C02, and acetate (12,13,21,29,30), the latter compound serving as a major oxidizable energy source for termites (20), as well as constituting an important biosynthetic precursor for the insect (2,3,17,29). Our knowledge of the cell biology and biochemistry of termite gut protozoa is still in its infancy, having been hampered for many years by the lack of pure cultures for detailed study. However, a major breakthrough was recently made by Iamin, who reported the axenic cultivation of two species (Trichomitopsis termopsidis (28) and Trichonympha sphaerica (30)) and who studied some of their metabolic activities (31). In a companion paper (21) we reported our studies on the nutrition and growth characteristics of one of Iamin's isolates, I. termopsidis, and our methods for increasing the cell yields of I. termopsidis in vitro. In the present study, we capitalized on the 75 improved growth characteristics of _‘l_'. termopsidis to obtain cells for biochemical studies. Herein we report on cellulase and other polymer- hydrolyzing enzyme activities in cell extracts of 1. termopsidis. ‘76 MATERIALS AND METHODS Organisms. '1. termopsidis strain 6057C was used throughout this study and was obtained by bromoethanesulfonate treatment of strain 6057, as previously described (21). The sources of, and cultivation methods for, Bacteroides JH20 and Methanospirillum hunggtii JF1 are given in a companion paper (21). Cultivation of_protozoa. ,1. termopsidis was grown under anaerobic conditions in modified Iamin's medium containing heat-killed cells of Bacteroides sp. JH20 (0.1 mg dry wt./ml, final concentration) in place of heat-killed rumen bacteria, and in co-cultivation with M. hungatii JF1 as previously described (21). For large volumes, bottles (Hheaton borosilicate glass, Type 900) were used containing 500 ml of medium and were sealed with a butyl rubber stopper. The medium for battle cultures was prepared in a similar fashion to that used for tube cultures (21), except that glutathione was added to the sterile medium from a separate heat-sterilized stock solution. Growth of co-cultures of I. termopsidis and M. hungatii was monitored by period- ic analysis of headspace gas for methane by using gas chromatography (21). Agitation of bottle cultures was completely avoided, or kept to a minimum, as shaking retards the growth of I. termopsidis (21). 77 Propagation of cell extracts. Bottle cultures of‘g. termopsidis were harvested when protozoan densities reached 2-3 x 104 cells/ml. Generally, the contents of 3 bottles were pooled (ca. 1.5 1 total volume), and cells were collected by centrifugation in air-tight centrifuge bottles (Sorvall, cat. no. 03256; equipped with caps, cat. no. 03278) at 1200 x g for 20 min at 10° C. The resulting pellets, consisting mainly of I. termopsidis and unused cellulose particles, were resuspended in a small volume of phosphate-buffered salts solution (see below), pooled, and then layered onto 100 ml of a 20 1 (w/v) solution of’Ficoll prepared in phosphate-buffered salts. After 30 min, most of the undigested cellulose particles descended into the Ficoll layer. The layer above.the Ficoll was then collected by aspiration and re-centrifuged at 1200 x g for 20 min at 10°C. The resulting pellet was again resuspended in phosphate-buffered salts solution, centrifuged as described above, and the final pellet was resuspended in 1-2 ml phosphate-buffered salts solution at 9‘ C. Cell densities were then determined by direct microscopic counts (21). By using this procedure, recovery osz. termopsidis was approximately 30-50 1. However, the final suspension was highly enriched in I. termopsidis, with only minor contamination from cellulose particles or bacterial cells. Cell extracts were prepared by either addition of Triton x-1oo (2 1, v/v, final concentration) to the cell suspension, or by freezing and thawing. The latter was done by immersion of the cell suspension in a dry ice-acetone bath for 3 min, followed by incubation at room temperature until completely thawed. Both procedures resulted in vir- tually 100 S lysis of protozoa, as judged microsc0pically. Cell 78 extracts were then centrifuged at 12000 x g at 9 'C for 20 min and separated into a soluble (i.e. supernatant fluid) and particulate (i.e. pellet) fraction. All manipulations of cells and cell extracts were performed under 02-free l2, by using strict anaerobic techniques (11). The phosphate-buffered salts solution used in the above procedures was prepared under anaerobic conditions and contained (mM): EZHP04, 10.8; EHZPOA, 6.9; HaCl, 29.5; KCl, 21.5; and dithiothreitol, 1.0. The final pH of this solution was 7.0. Enzyme assays. Unless indicated otherwise, all assays were performed at 30' C, and determinations of cellulase activities generally followed the recommendations of Chase et al (10). 'Cellulase' (i.e. the complete enzyme system (10.16)) and endo-p-1,9-glucanase (EC 3.2.1.9) were assayed by measuring release of reducing sugar from microcrystalline cellulose (Sigma Chemical Co.; Type 20, 5.0 1, w/v, final concentration) and carboxymethylcellulose (Hercules Inc., Type 7MP; 0.5 5, w/v, final concentration), respectively, in 0.2 M acetate buffer (pH 5). Reaction mixtures (0.1 ml final volume) were run in triplicate in Eppendorf conical centrifuge tubes (1.5 m1 cap.) and were terminated by the addition of 0.025 ml 1 5 H01. Following neutralization with 0.025 ml of 1 g HaOH, samples were centrifuged at 11,310 x g for 5 min and the supernatant fluids were analyzed for reducing sugar by the method of Park and Johnson (22) or Bernfeld (1). In addition, reaction mixtures for the complete cellulase enzyme system were screened for specific products 79 and intermediates of cellulose hydrolysis (i.e glucose, cellobiase, and cellodextrins) by subjecting supernatant fluids to high pressure liquid chromatography (15). The extent of saccharification of cellulose was estimated by using the prarocedures described above, but with less microcrystalline cellulose in reaction mixtures (0.09 1, w/v, final concentration ) as suggested by Johnson 33 g; (19). Cellobiase (EC 3.2.1.21) was assayed by using reaction mixtures similar to those described above, but by measuring release of glucose. from cellobiase (1 1, w/v, final concentration) with glucose oxidase (26). In general, reaction mixtures were incubated for (hr): the complete cellulase enzyme system, 29; carboxymethyl cellulase, 0.5-1; and cellobiase, 1-3. One unit of activity was defined as the release of 1 mole of glucose equivalent per minute. To determine other hydrolytic activities of‘I. termopsidis, ball-milled Douglas fir wood powder (20), xylan (Sigma cat. no. X-3875), or ball-milled chitin (21) was substituted for microcrystalline cellulose, and reaction mixtures were assayed for reducing sugar as described above. Protease was assayed by using azocasein as substrate as described by Brook 11 _a_1_ (7). and P-amylase was assayed by using potato starch as substrate according to the method of Bernfeld (1). Estimation of protozoan cell protein. It was virtually impossible to separate 1, termopsidis cells completely from bacterial cells (see above). yet still recover enough protozoa to prepare a suitable cell extract. Consequently, enzyme activities are reported 80 herein as mUnits/106 I. termogsidis cell equivalents, determined by direct cell counts of suspensions prior to extract preparation. How- ever, a numerical factor could be calculated for converting protozoan cell equivalents to mg protozoan specific protein by the knowledge tint I. termogidis cells were approximately spheres 90 1 5 )II in diameter (see Figure 9 of Ref. #21), and assuming: their density was 1.0; their water content was 80 S; and protein constituted 50 S of the cell's dry weight. By using these assumptions, it was estimated that 106 cells of 1. term0psidis were equivalent to 3.9 mg protein. Chemicals. Cellodextrins (cellobiase through cellohexaose) were prepared by acid hydrolysis of cellulose followed by column ,chromatography of the hydrolysate (18). All other chemicals were of reagent grade and were obtained from comercial sources. 81 RESULTS Cell extracts osz, termopsidis possessed enzyme activity against carboxymethylcellulose and cellobiase, which indicated the presence of endojp-1,9-glucanase and cellobiase, respectively (Table 1). Compared to freeze-thaw preparations, extracts prepared by Triton 1-100 treatment had greater total amounts of endoglucanase and cellobiase activities, as well as a greater percentage of these activities associated with the soluble fraction of extracts, although both procedures resulted in lysis of protozoan cells. This observation suggested that an appreciable amount of enzyme activity in freeze-thaw preparations remained masked, perhaps because it was associated with membranous subcellular organelles and was inaccessible to the substrates. Cellobiase activity was linear for at least 29 hr (the maximum time tested), however endoglucanase activity decreased significantly after 1 hr, at which time about 2 I degradation of carboxymethylcellulose had occured. Cell extracts affil. termopsidis also possessed hydrolytic activity against microcrystalline cellulose, although the rate of hydrolysis of cellulose was 3 to 10-fold less than that of carboxy- methylcellulose or cellobiase (Table 1). The rate of hydrolysis of microcrystalline cellulose was linear for up to 29 hr, but declined slowly from 29 to 72 hr. By using the saccharification assay system ) 82! Table 1 Cellulolytic enzyme activities of Trichomitopsis termopsidi . Enz Activit A inst Substrate)1 Extract Extract ‘Microcrystalline Carboxymethyl Preparation Fraction Cellulose cellulose Cellobiase TRITON SOLUBLE 1.60 1 0.73 22.8 1 10.5 17.5 1 1.1 X-1OO (ns9) (n=2) PARTICULATE 0.88 1 0.22 5.8 1 0.8 2.9 1 0.6 FREEZE SOLUBLE 1.71 1 0.22 3.7 1 0.6 3.5 1 0.8 nuw (11:3) PARTICULATE 1.59 1 0.28 5.3 1 1.9 8.7 1 0.1 1 Activity expressed as mUnits/ 106 cell equivalents. Minimum level of activity detectable (mU): microcrystalline cellulose, 0.02; carboxymethylcellulose, 0.2; and cellobiase, 0.2. Values are the mean t S.E.M. for 3-6 determinations on a single extract (n=1), except where indicated otherwise. 83 a maximum of 19.9 I saccharification of microcrystalline cellulose was observed during a 69 hr incubation period. However, most of this saccharification took place during the first 29 hr. All cellulolytic enzyme activities measured were proportional to the amount of cell extract used in the reaction mixture; were dependent on the presence of substrate; and were abolished by heating the extract to 100 ‘C for 10 min. Moreover, control experiments showed that cellulolytic activity was associated only with the protozoa: no cellulolytic activities were detected in separate extracts of heat-killed Bacteroides JH20 or viable M. hunggtii which were used in media to culture I, termopsidis. Inasmuch as protozoan extracts contained small, but significant and variable amounts of bacterial cell protein (and probably fetal bovine serum protein as well), the values reported in Table 1 were normalized to 106 protozoan cell equivalents. However, the total (soluble plus particulate) activity of any of the enzymes could be normalized to protozoan cell protein by assuming that 106 T. termopsidis cells were equivalent to 3.9 mg protein (see Materials and Methods). Thus, for extracts prepared by Triton X-100 treatment, the specific rates of substrate hydrolysis were approximately (mUnits/mg protein): microcrystalline cellulose, 0.73; carboxymethylcellulose, 8.9; and cellobiase, 6.0 . Closer examination of the hydrolysis of microcrystalline cellulose by I. termopsidsis extracts revealed apparent optima at pH 5 (Figure 1) and at a temperature of 30' C (Figure 2). Moreover, Figure 1. Effect of pH on the hydrolysis of microcrystalline cellulose at 30° C by soluble (Triton x-100) extracts 6 of I. termopsidis (equivalent to 7.3 x 10 cells/ml extract). an 1O h0