Bacteria. Received for publication 4 January via election sinks other than H2 (6, 18, 19). The

Transcription

1 APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 1977, p Copyright 1977 American Society for Microbiology Vol. 33, No. 5 Printed in U.S.A. Growth of Desulfovibrio in Lactate or Ethanol Media Low in Sulfate in Association with H2-Utilizing Methanogenic Bacteria M. P. BRYANT,* L. LEON CAMPBELL,' C. A. REDDY,2 AND M. R. CRABILL Departments of Dairy Science and Microbiology, University of Illinois, Urbana, Illinois Received for publication 4 January 1977 In the analysis of an ethanol-co, enrichment of bacteria from an anaerobic sewage digestor, a strain tentatively identified as Desulfovibrio vulgaris and an H,-utilizing methanogen resembling Methanobacterium formicicum were isolated, and they were shown to represent a synergistic association of two bacterial species similar to that previously found between S organism and Methanobacterium strain MOH isolated from Methanobacillus omelianskii. In lowsulfate media, the desulfovibrio produced acetate and H2 from ethanol and acetate, H2, and, presumably, CO2 from lactate; but growth was slight and little of the energy source was catabolized unless the organism was combined with an H2-utilizing methanogenic bacterium. The type strains of D. vulgaris and Desulfovibrio desulfuricans carried out the same type of synergistic growth with methanogens. In mixtures of desulfovibrio and strain MOH growing on ethanol, lactate, or pyruvate, diminution of methane produced was stoichiometric with the moles of sulfate added, and the desulfovibrios grew better with sulfate addition. The energetics of the synergistic associations and of the competition between the methanogenic system and sulfate-reducing system as sinks for electrons generated in the oxidation of organic materials such as ethanol, lactate, and acetate are discussed. It is suggested that lack of availability of H2 for growth of methanogens is a major factor in suppression of methanogenesis by sulfate in natural ecosystems. The results with these known mixtures of bacteria suggest that hydrogenase-forming, sulfate-reducing bacteria could be active in some methanogenic ecosystems that are low in sulfate. Methanobacillus omelianskii (1) was described as a strictly anaerobic bacterium which obtained its energy for growth by oxidizing ethanol and a number of alcohols other than methanol to the corresponding acid or ketone. The electrons generated were utilized to reduce CO, to methane. The fermentation of ethanol is represented by equation A + B (Table 1). It was later shown that M. omelianskii was actually a synergistic association of two bacterial species (6). One of these, the S organism, is a gramnegative, anaerobic, peritrichous rod which obtains energy for growth by fermenting ethanol with production of acetate and H, according to equation B (Table 1). However, it is unable to grow well as a pure culture in media containing ethanol or other utilizable alcohols because accumulation of H2 inhibits its growth; and it does not possess the capability to dispose of I Present address, 104 Hullihen Hall, University of Delaware, Newark, DE Present address, Department of Microbiology, Michigan State University, East Lansing, MI electrons genevated in the oxidation of ethanol via election sinks other than H2 (6, 18, 19). The other organism, Methanobacterium strain MOH, is a methanogenic bacterium that utilizes the H2 produced by S organism to reduce CO2 to methane according to equation A (Table 1) and, thus, serves as a hydrogen sink and allows good growth of the S organism. Methanobacterium strain MOH is similar to Methanobacterium formicicum except that it does not utilize formate (5). It is unable to utilize alcohols or many other compounds other than H2 and CO2 as energy sources for growth. S organism also grows well on ethanol when combined with Methanobacterium ruminantium and, presumably, with any bacterium that effectively utilizes H2 and thus maintains a low partial pressure of H2 in the medium (6, 18). Based on the above work, it was suggested that methanogenic bacteria in nature do not utilize alcohols other than methanol (6); however, since only one culture was analyzed, the present study was initiated to determine if

2 VOL. 33, 1977 DESULFOVIBRIO GROWTH IN LOW-SULFATE MEDIA 1163 TABLE 1. Equations and free-energy changes for reactions involving anaerobic oxidation of ethanol, lactate, and pyruvate to acetate or acetate and CO2 with protons or sulfate serving as electron acceptor by pure cultures of desulfovibrios or mixtures of desulfovibrios with H2-utilizing methanogens Equation AG.' tion)a (kcal/reac- A 4 H2 + HCO3- + H+ =CH4 + 3 H B 2 CH3CH20H + 2 H20 2 CH3COO-+ 2 H+ + 4 H A + B 2 CH3CH20H + HCO3-2 CH3COO- + H+ + CH4 + H C 2 CH3CH20H + S042-2 CH3COO- + H2S + 2 H D 2 CH3CHOHCOO- + 4 H20 2 CH3COO- + 2 HC H+ + 4 H2-1.9 A + D 2 CH3CHOHCOO- + H20 2 CH3COO- + HCO3- + H+ + CH E 2 CH3CHOHCOO- + S042-2 CH3COO- + 2 HCO3- + H2S F 4 CH3COCOO- + 8 H20 4 CH3COO- + 4 HC H+ + 4 H A + F 4 CH3COCOO- + 5 H20 4 CH3COO- + 3 HCO H+ + CH G 4 CH3COCOO- + S H20 4 CH3COO- + 4 HCO H+ + H2S a H2 and CH4 in the gaseous state; all other substances in aqueous solution at 1 mol of activity per kg. The free-energy data have been calculated using the free energies of formation from the elements compiled by Thauer et al. (26). other kinds of bacteria may be concerned in methane formation from ethanol and CO2 (equation A + B, Table 1) in an elective enrichment from sewage sludge. It is shown that strains of the genus Desulfovibrio grow with energy sources such as ethanol and lactate in media deficient in sulfate if combined with H2- utilizing methanogenic bacteria (equations A + B and A + D, Table 1). A brief communication was previously published (M. P. Bryant, Am. Chem. Soc. Abstr. Microbiol. Sect., p. 18, 1969). MATERIALS AND METHODS Methanobacterium strain MOH (6), Desulfovibrio desulfuricans type strain Essex 6, and Desulfovibrio vulgaris type strain Hildenborough were from culture collections at the University of Illinois. Anaerobic culture techniques were modifications of those of Hungate as used by Bryant (4), and media and methods were those used in the study of M. omelianskii (6) except as indicated below. The ethanol enrichment was carried out in a medium containing sulfur-free mineral solution 3 (27), 5 mm NH4Cl, and 36,uM FeSO4. Na2CO3, resazurin, sulfide-cysteine reducing system, 0.5% ethanol, and 100% CO2 gas phase were the same as in the rumen fluid media indicated below. The medium was tubed in 5-ml amounts in 18- by 150-mm tubes. The anaerobic dilution solution, rumen fluid agar roll-tube medium with a 1:1 H2-CO2 gas phase, rumen fluidethanol-agar roll-tube medium with 100% CO2 gas phase, and rumen fluid slants were prepared and used as previously indicated (6). The Trypticaseyeast extract (TY) medium contained the same ingredients as the enrichment medium plus 0.1% each Trypticase and yeast extract (Difco). Ethanol or other energy sources were added in amounts as indicated in Results. When used as a liquid medium, the TY medium was tubed in 5-ml amounts in 18- by 150- mm tubes. Ethanol-TY slant medium contained 0.5% ethanol and 1% agar and was tubed in 3.5-ml amounts in 13- by 100-mm tubes. Methods for analyses of fermentation products included gas chromatographic and volume measurements for H2 and CH4 (6), gas chromatographic methods for acetate, other volatile acids, and ethanol (18), and colorimetric methods for lactate (2) and pyruvate (9). All incubations were at 37 C and static except for the liquid cultures of methanogens growing on CO2-H2, which were shaken (6). Growth was estimated as optical density at 600 nm in the 18- by 150-mm culture tubes and using a Bausch and Lomb Spectromic 20 spectrophotometer. Cells of Desulfovibrio strain EC1 used for detection of desulfoviridin and cytochrome and for isolation of deoxyribonucleic acid were grown in 3-liter amounts in the TY medium with addition of 21 mm Na2SO4 and 53 mm sodium lactate. The methods for difference spectra described by White et al. (29) were used for detection of cytochrome. Desulfoviridin was demonstrated using the method of Postgate (16). Deoxyribonucleic acid was isolated using the method of Saito and Miura (20), and the percent guanine plus cytosine was determined by buoyant density measurements in a CsCl gradient (21) using the equations of Sueoka (24). RESULTS Enrichment. The ethanol enrichment was carried for several months, with transfers (10% inoculum) being made every 3 to 5 days. Growth equal to an optical density of about 0.4 to 0.5 occurred, and analyses of tubes for methane just before transfer indicated production of 35 to 45,umol of methane per ml of medium, with little or no H2 production. Based on the stoichiometry of equation A + B (Table 1), this methane was equivalent to utilization of 70 to 90 mm ethanol. Acetate formation was not quantitated, but qualitative analysis showed

3 1164 BRYANT ET AL. that it accumulated in the enrichment, and no organisms resembling known acetate-utilizing methane bacteria were seen in microscopic observations. One milliliter of enrichment was serially diluted through seven tubes containing 9 ml of anaerobic dilution solution, and 0.5 ml of dilution in tubes 3 through 7 was used to inoculate, in triplicate, roll tubes of rumen fluid agar with a 1:1 H2-CO gas phase and the same medium modified to contain CO2 gas phase and 1% ethanol. Colony counts after 25 days of incubation averaged 3.6 x 108/ml for the H2-CO2 medium and 4.2 x 108/ml for the ethanol-co2 medium. Gas analyses of one tube of each medium inoculated with each dilution indicated that H2, but no methane, was produced by some of the most numerous colonies developing in the medium with CO2 gas phase and methane was formed by some of the most numerous colonies in the medium with H2-CO2 gas phase. Methanogenic strains. Colonies seen in the medium with H2-CO2 gas phase inoculated with 5 x 10-7 or 5 x 10-8 ml of enrichment were mainly of one type. These colonies became visible after about 1 week of incubation and increased in size until some were about 2 mm in diameter after 20 days. They were greyish white, roughly round, and filamentous. Fourteen of these colonies were picked and stabbed into the H2-CO2 slant medium. Eight of these showed visible growth within 2 weeks and were morphologically identical. Examination of wet mounts prepared from the water of syneresis of 4- to 5-day-old slant cultures with the phasecontrast microscope and of Gram strains showed all to be slender, cylindrical rods with bluntly rounded ends. They were nonmotile, gram positive to gram variable, 0.4 to 0.6 /im wide, and usually 3 to 10,um long, but some long filaments were often present. Chains were also common, and many cells were more or less curved. All of these cultures utilized all of the H2 in the slant cultures within 10 days and produced methane. In further studies of representative strain E5, formate and H2 were the only electron donors found to be utilized for methane formation. When the strain was incubated for 1 month in the H2-CO2 liquid medium with 0.5% of various compounds added, the methane formed was not greater than that due to H2 utilization (116,umol of CH4), except in the case of sodium formate (171,umol of CH4). Compounds not utilized included acetate, butyrate, methanol, ethanol, butanol, glucose, cellobiose, maltose, xylose, arabinose, serine, glycine, alanine, glutamate, aspartate, and succinate. The colony type, morphology, and APPL. ENVIRON. MICROBIOL. electron donors utilized for methane formation indicate that strain E5 is probably M. formicicum (5). Non-methanogenic strains. Twelve colonies, representative of all colony types found in the rumen fluid-ethanol agar roll tubes (CO, gas phase) inoculated with 5 x 10-7 or 5 x 10-8 ml of enrichment, were picked to 0.5% ethanol-ty slants, and, based on colony type, cell shape, and motility, three types of bacteria were isolated. All were gram-negative rods. Three cultures were relatively large curved rods, 0.8 to 1.0,um wide and 3 to 5,um long. They were actively motile, and electron micrographs showed monotrichous and polar flagellation. Colonies in the rumen fluid-ethanol agar roll tubes were lens shaped, dark brown, and 0.5 to 1.2 mm in diameter. They were strict anaerobes and produced a small amount of H2 and no methane during growth in the ethanol-ty slant medium. Strain ECi, representative of the group, was maintained for further study. Seven cultures of which strain EC3 was representative were smaller motile vibrios, and two strains of which EC5 was representative were small nonmotile rods. The small vibrios produced traces of H2 in the ethanol-ty slant medium but no methane, and the nonmotile rods produced neither gas. Recombination experiments. The four kinds of bacteria isolated appeared to represent the main morphological groups seen in wet mounts and Gram stains of the enrichment; however, none of them produced methane in ethanol medium with CO2 gas phase. Therefore, each of the non-methanogenic strains was combined with the methanogenic strain E5 in ethanol-ty slants. Results in Table 2 show that good methane formation occurred only when strain EC1 was combined with an H2-utilizing methane bacterium. Either methanogenic strain E5 isolated from the enrichment or strain MOH, a methanogen that utilizes only H2-CO2 in methane formation, was active. Note that strain ECi produced from 57 to 65 times as much H2 when combined with a methanogen (H2 equals methane times 4) as when grown in pure culture. Identification of strain ECL. The morphological features, fermentation of ethanol, and the dark colonies of strain ECi suggested that it was a member of the genus Desulfovibrio, and this was subsequently proven. Results in Table 3 show that it requires sulfate for significant growth with ethanol or lactate as energy source, but grows well with pyruvate in the presence or absence of sulfate and does not utilize glucose as energy source. It is of interest

4 VOL. 33, 1977 TABLE 2. Effect of combination of methanogenic bacteria with three strains of non-methanogenic bacteria on growth and gas formation in ethanol-ty agar slant medium with CO2 gas phasea (strain)b Visible H2 (Omol) CH4 (j,mol) EC E5-0 0 EC + E EC EC3 + E EC EC5 + E MOH EC1 + MOH a Slants were stab-inoculated into the base with a loop. Growth was maximal after about 5 days of incubation. Gas analyses were done after 2 weeks. All determinations are means of analyses on duplicate tubes in which results were essentially identical ḃ EC1, EC3, and EC5 are the non-methanogenic strains typical of the large motile curved-rod group, small motile curved-rod group, and nonmotile rod group, respectively, as described in the text. Strain E5 is the methanogenic bacterium isolated in the present study and MOH is Methanobacterium strain MOH isolated from M. omelianskii (6). TABLE 3. Some energy sources of strain ECI Additions to basal mediuma Growth (OD) None 0.04 (12)b S (12) Ethanol 0.03 (12) Ethanol + SO (2) Lactate 0.04 (12) Lactate + S (12) Pyruvate 0.65 (7) Pyruvate + S (2) Glucose 0.03 (12) Glucose + S (12) a The basal medium was the low-sulfate TY medium with about 60 mm ethanol, sodium lactate, sodium pyruvate, 22 mm glucose, and 20 mm Na2SO4 added as indicated. b Values are means of duplicate tubes, which were quite similar; number in parentheses refers to days of incubation required to reach maximal optical density (OD). that after maintenance in ethanol media, it grew very slowly when inoculated into lactate medium. However, after it was maintained in lactate medium for several transfers, growth was somewhat more rapid. Results shown in Table 4 indicate that as expected with a member of the genus Desulfovibrio, EC1 produces acetate from ethanol when it is grown in either the medium containing sulfate or in the me- DESULFOVIBRIO GROWTH IN LOW-SULFATE MEDIA 1165 dium without sulfate but in association with Methanobacterium strain MOH. A difference spectrum of air-oxidized versus hydrosulfite-reduced cells indicated the presence of a c-type cytochrome with absorption maxima at about 415, 524, and um. Cells contained desulfoviridin. The guanine plus cytosine content of the deoxyribonucleic acid was and mol% for two determinations on separate batches of cells. The percent guanine plus cytosine and most other features suggest that in the scheme of Postgate and Campbell (13, 16) EC1 belongs in the species D. vulgaris; however, its ability to grow well with pyruvate as energy source in the medium low in sulfate suggests that it is not a typical member of this species. Other strains of Desulfovibrio. Results in Table 5 show that Essex 6 (NCIB 8307), the neotype strain of the species D. desulfuricans, as well as strain EC1, grow well in the lactate medium with low sulfate in the presence of the methanogenic bacterium. Essex 6 ferments lactate much more rapidly than EC1, and the reverse is true for ethanol. These results (Table 5) also show that methane formation, in the case of both strains in lactate medium and in the case of strain EC1 in ethanol medium, was lowered when sulfate was included by an amount approximately equivalent to the amount of sulfate added, i.e., 20,umol/ml of medium. This indicates that the H,-producing system does not compete with the sulfate-reducing system as electron sink in these bacteria even when the H,-utilizing methanogenic bacterium is present in sufficient numbers to rapidly utilize the H. produced by the sulfate reducer in the absence of sulfate. Results in Table 5 also indicate that electrons TABLE 4. Fermentation products produced by strain ECI in ethanol-ty medium containing sulfate and in ethanol-ty medium without sulfate when combined with Methanobacterium strain MOHa,umol/ml of medium Parameter EC1 + SO42- EC + MOH Ethanol utilized Product Acetate Hydrogen Methane athe medium was the low-sulfate TY medium with 60 mm ethanol added for the combination of strains and the same medium plus 20 mm Na2SO4 for strain EC1 alone. Values are means of duplicate tubes after 1 week of incubation; values for products produced in the medium minus ethanol were subtracted from those of the medium with ethanol.

5 1166 BRYANT ET AL. produced in the oxidation of pyruvate, presumably to acetate and CO2 (equations F and A + F, Table 1), were essentially all recovered in methane; i.e., 57 mm pyruvate should yield about one-quarter as much or 14,umol of methane per ml of medium. The actual values were 13.7 and 12.0,umol. Similar results were previously obtained when S organism was grown on pyruvate with M. ruminantium (18). D. desulfuricans, in the absence of sulfate, utilizes pyruvate, but whole cells of strain MOH do not metabolize pyruvate. Data in Table 6 show that the type strains of both of the species D. desulfuricans and D. vulgaris actively ferment lactate in low-sulfate medium when in combination with the methanogenic bacterium, and it is concluded that the electrons generated in the oxidation of lactate to acetate are given off as H2, which is utilized for growth and methane formation by the methanogenic bacterium. Neither of the strains metabolize more than traces of lactate in this medium in the absence of sulfate or the methanogenic bacterium. CO, is presumed to be another product but was not estimated in these experiments. DISCUSSION The present results support the previous tentative conclusion (6) that methanogenic bacteria in nature do not utilize normal and isoalcohols other than methanol as electron donors TABLE 5. Methane formation and growth of Desulfovibrio strains ECI and Essex 6 in combination with Methanobacterium strain MOH in media with and without sulfatea EC1 + MOH Essex 6 + MOH Substrates CH, CH, (.Umol/ OD mol/ ( OD ml) ml) None (9)b (6)b s (7) (6) Lactate (9) (6) Lactate + S (7) (2) Pyruvate (7) (3) Pyruvate (6) (2) S04 Ethanol (5) (10) Ethanol + S (3) (6) APPL. ENVIRON. MICROBIOL. a The basal medium was the low-sulfate TY medium. Where indicated, 20 mm Na,SO4, 80 mm sodium lactate, 57 mm sodium pyruvate, or 60 mm ethanol was added. Inoculum was 0.1 ml of mixed culture (carried in the medium with lactate added) per 5 ml of experimental medium in 18- by 150-mm tubes. Gas was analyzed from duplicate tubes after 12 days of incubation, and H2 was 0.2,umol/ml of medium or less. " Number in parentheses indicates days of incubation for maximal optical density (OD). TABLE 6. Fermentation products produced from lactate by types strains ofd. desulfuricans (Essex 6) and D. vulgaris (Hildenborough) grown in combination with Methanobacterium strain MOH in low-sulfate mediuma,mol/ml of medium Parameter Hildenbor- Essex 6 ough Lactate utilized Product Ethanol Acetate Methane Hydrogen a The medium was the TY medium with 80 mm sodium lactate. Values are means of duplicate culture tubes after 1 week of incubation. Values for small amounts of products produced in medium minus lactate were subtracted from those of the medium with lactate. in the reduction of CO2 to methane. Ethanol was shown to be catabolized to acetate and H., (equation B) by nonmethanogenic bacteria, i.e., S organism in the study of M. omelianskii and members of the genus Desulfovibrio in the present study. Methane production was via reduction of CO, with H., (equation A, Table 1) by Methanobacterium strain MOH in the earlier study (6) and by a strain tentatively identified as M. formicicum in the present study. In unpublished work done in the laboratory of Bryant, Thomas Glass isolated five different strains of bacteria that grew well in propanol or ethanol medium only in the presence of an H.,- utilizing methanogen. Although detailed studies were not carried out, some strains were morphologically different from either desulfovibrios or S organism. It thus appears that a number of nonmethanogenic species are capable of anaerobic oxidation of ethanol. The ethanol-oxidizing desulfovibrios differ from S organism (6, 18) in that they have the ability to very effectively utilize sulfate reduction to sulfide (equations C, E, and G, Table 1) as the sink for disposal of electrons generated in the oxidation of ethanol, lactate, or pyruvate. In addition, desulfovibrios utilize lactate and other energy sources (13) not utilized by S organism (18). Our results show that in the absence of sulfate and in the presence of methanogens, the desulfovibrio strains catabolize ethanol and lactate according to equations B and D (Table 1). These reactions are not thermodynamically favorable; i.e., the free-energy change is not negative enough to allow the fermentation to proceed unless it is coupled with another reaction

6 VOL. 33, 1977 such as that occurring when they are associated with H2-utilizing methane bacteria (equations A + B and A + D). Calculations of Wolin (32) show that the free-energy change for equation B (Table 1) becomes progressively more negative and the reaction proceeds more effectively when the partial pressure of H, is maintained at a much lower level than the standard conditions shown in Table 1. [The same would be true for lactate and pyruvate [equations D and F]; however, the oxidation of pyruvate to acetate, CO, and H2 is thermodynamically quite favorable even with 1-atm pressure of H,. In the case of S organism, growth on pyruvate was much better when it was combined with a methanogen [18]; we did not compare growth of desulfovibrios alone and in combination with a methanogen on pyruvate low-sulfate medium in the present study.] The methane bacteria have a very great affinity for H2 and the partial pressure of H2 in methanogenic ecosystems is maintained at a very low level. For example, Hungate et al. (10, 11) found the K,, for H, use in CH4 production either by M. ruminantium or by rumen contents to be 10-fi M, and the concentration of H2 in the rumen ecosystem was also about 10-6 M. This is equivalent to about 1.5 x 10-3 atm of H2 in the gas phase. For detailed discussions of the concepts of electron disposal via H2 production by non-methanogenic bacteria during their oxidation of metabolites such as ethanol, lactate, pyruvate, and reduced pyridine nucleotides in association with methanogens and other H2-utilizing species, the reader is referred to the excellent discussion by Wolin (32). Our results show that growth of desulfovibrios on ethanol, lactate, or pyruvate was faster and yields were higher when sulfate was added as electron acceptor as compared to H2 production and subsequent use by the methanogenic system (Table 5); and that flow of electrons via H2 and C02 to methane probably does not compete with their flow into reduction of sulfate. The latter is indicated by the fact that in the mixed culture of desulfovibrio plus methanogen, methane production was reduced by an amount about stoichiometric with the amount of sulfate added, and no methane was produced if excess sulfate was added; e.g., see the data on pyruvate catabolism in Table 5. The lack of methane in cultures containing sulfate is probably due to lack of H2 production and not due to toxicity of sulfate, since sulfate is regularly used in culture media for growth of pure cultures of H2-utilizing methanogens. It is also probable that toxicity of sulfide was not a major factor because, if this was the case, a small DESULFOVIBRIO GROWTH IN LOW-SULFATE MEDIA 1167 amount of H2 would have accummulated after sulfate was used and this did not occur; i.e., methane was produced from lactate and ethanol after 20 mm sulfate was reduced (Table 5). The free-energy change for the oxidation of ethanol, lactate, and pyruvate is only slightly more favorable when sulfate is the electron acceptor as compared to the mixed methanogenic system; in each case the AG,,' is approximately 4 kcal more negative for the sulfate system (Table 1). Thus, we attempted to find information favoring sulfate as electron acceptor from literature available on adenosine triphosphate (ATP) yields. Although it is evident that 1 mol of ATP would be gained for each mole of acetate generated in equations B, C, D, E, F, and G (Table 1) and that ATP is probably not generated in electron transfer to H2, the net ATP produced in reduction of sulfate is still not authenticated (7, 13, 28). It would be of interest to compare growth yields of desulfovibrios with lactate and pyruvate as limiting energy sources and sulfate or methanogen as electron acceptor system in the chemostat set at comparable dilution rates. Also, since Sorokin (23) has shown growth of D. desulfuricans using H2-sulfate or formate-sulfate as energy source when acetate and CO2 are available as carbon sources, growth yields when these substrates are used would be of particular interest. An increased ATP gain when sulfate was the electron acceptor would explain why growth yields were higher in the two-species culture in the presence of sulfate than in its absence. The reason why the flow of electrons via H2 and CO2 to methane does not appear to compete with the flow into reduction of sulfate may be a spatial rather than a thermodynamic one. A factor that appears to favor sulfate as electron acceptor is that the electrons are accepted within the same cell (or membranes?) that produces them; thus, time required for movement from the generation sites to the sites of utilization should be less than in the two-organism system where H2 must move out of one cell and into another. However, Bell et al. (3) have made the interesting observation that the hydrogenase in Desulfovibrio gigas has a periplasmic location, and they have suggested that this location near the cell surface may serve a function in H2 production under thermodynamically unfavorable conditions as in interspecies H2 transfer (32). Whether hydrogenases of other species have a similar location is not yet known. It is now well known that biological methanogenesis in natural ecosystems does not occur

7 1168 BRYANT ET AL. when sulfate is present (see Martens [15] for many references). Cappenberg (7, 8) suggests that this is due to the sensitivity of methanogenic bacteria to H.S produced by sulfate-reducing bacteria during their oxidation of organic matter. The main evidence for this was from experiments in which H.S (ps2- of 11 or lower) was shown to be toxic for growth of a methanogenic bacterium growing with acetate as sole energy source. The present study suggests that Methanobacterium strain MOH is quite tolerant of high total sulfide levels (about 20 mm at ph 6.6 to 6.8 in experiments shown in Table 5). We suggest that lack of availability of H2 for growth of methanogens is a factor, possibly the major factor, in suppression of methanogenesis by sulfate in natural ecosystems. H2-CO2 is the main energy source for methanogens in ecosystems such as the rumen and large bowel where volatile fatty acids are also major end products (10, 11), and, whereas the methyl group of acetate is a major precursor of methane in ecosystems such as sewage digestors (12, 14, 22) or equatic sediments (8) where complete anaerobic degradation occurs, i.e., where volatile fatty acids are not major end products, it is still not certain whether H2 or acetate is the main energy source in nature for acetate-utilizing methanogens (14, 25, 33). The presence of sulfate would probably eliminate the production of H2 by desulfovibrios from organic precursors such as ethanol and lactate. It is possible that desulfovibrios would also preferentially use H2 produced by other organisms for reduction of sulfate and, thereby, make H2 unavailable to methane-producing species. Both of these points need further experimental verification, but the results in this paper support the suggestion that H2 is not produced from organic substrates by desulfovibrios until sulfate is reduced to sulfide. Even if acetate proves to be the major energy source for the bacteria that use it in methane production, the recent discovery of Widdel and Pfennig (30) suggests that acetate would be completely oxidized to C02 rather than to methane and C02 in the presence of sulfate in anaerobic microbial ecosystems. Desulfotomaculum acetoxidans, a newly discovered bacterial species that oxidizes acetate to C00 and reduces sulfate to sulfide, was found in several different anaerobic microbial habitats including piggery wastes, bovine feces, and aquatic sediments. This species grows in medium with acetatesulfate as the energy source at a much faster rate (30) than do methane bacteria using acetate (7, 12, 14); and D. acetoxidans grows much APPL. ENVIRON. MICROBIOL. faster with ethanol or butyrate as the electron donor as compared to acetate (F. Widdel, personal communication). As with H2, ethanol, lactate, and pyruvate discussed above, conversion of acetate to CO with sulfate reduction to sulfide is thermodynamically more favorable than acetate conversion to C02 and CH4 (26). Winfrey and Zeikus (31) have independently proposed that competition for available H2 and acetate is the mechanism by which sulfate inhibits methanogenesis in aquatic systems. Finally, the present results suggest an additional function or ecological niche for sulfatereducing bacteria that can produce H2 from organic materials. In association with H2-utilizing methanogenic bacteria, they could obtain enough energy to be maintained and could be of importance in degradation of some organic materials in anaerobic ecosystems containing little or no sulfate. This possibility should be kept in mind in future studies on anaerobic degradation and biological methane formation. ACKNOWLEDGMENTS We are indebted to M. Kasprzycki for demonstration of the monotrichous cells with the electron microscope, to R. D. DeMoss for use of the Beckman model E analytical ultracentrifuge, and to M. E. Reichman for use ofthe microdensitometer. R. Thauer offered constructive criticisms of the discussions concerning thermodynamics and calculated free-energy changes from material in press (26). G. Zeikus also allowed us to see pertinent materials in press (31). This work was supported by Public Health Service grant EP00259, by U. S. Department of Agriculture grant , and by the Agricultural Experiment Station of the University of Illinois. LITERATURE CITED 1. Barker, H. A Bacterial fermentations. John Wiley & Sons, New York. 2. Barker, S. B., and W. H. Summerson The colorimetric determination of lactic acid in biological materials. J. Biol. Chem. 138: Bell, G. R., J. LeGall, and H. D. Peck Evidence fpr the periplasmic location of hydrogenase in Desulfovibrio gigas. J. Bacteriol. 120: Bryant, M. P Commentary on the Hungate technique for culture of anaerobic bacteria. Am. J. Clin.,//utr. 25: lryant, M. P Part 13. The methanogenic bacteria, p In R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 6. Bryant, M. P., E. A. Wolin, M. J. Wolin, and R. S. Wolfe Methanobacillus omelianskii, a symbiotic association of two species of bacteria. Arch. Mikrobiol. 59: Cappenberg, T. E A study of mixed continuous cultures of sulfate-reducing and methane-producing bacteria. Microb. Ecol. 2: Cappenberg, T. E., and R. A. Pins Inter-relationships between sulfate-reducing and methane-producing bacteria in bottom sediments of a fresh-water lake. III. Experiments with 14C-labeled substrates. Antonie van Leeuwenhoek J. Microbiol. Serol. 40:

8 VOL. 33, Friedemann, T. E., and G. E. Hagen The determination of keto acids in blood and urine. J. Biol. Chem. 147: Hungate, R. E Hydrogen as an intermediate in the rumen fermentation. Arch. Mikrobiol. 59: Hungate, R. E., W. Smith, T. Bauchop, I. Yu, and J. C. Rabinowitz Formate as an intermediate in the bovine rumen fermentation. J. Bacteriol. 102: Lawrence, A. W Application of process kinetics to design of anaerobic processes, p In F. G. Pohland (ed.), Anaerobic biological treatment processes, Advances in Chemistry Series 105. American Chemical Society, Washington, D.C. 13. LeGall, J., and J. R. Postgate The physiology of sulphate-reducing bacteria, p In A. H. Rose and D. W. Tempest (eds.), Advances in microbial physiology, vol. 10. Academic Press Inc., New York. 14. McCarty, P. L Energetics and kinetics ofanaerobic treatment, p In F. G. Pohland (ed.), Anaerobic biological treatment processes, Advances in Chemistry Series 105. American Chemical Society, Washington, D.C. 15. Martens, C. S Control of methane sediment-water bubble transport by macrofaunal irrigation in Cape Lookout Bight, North Carolina, Science 192: Postgate, J. R A diagnostic reaction of Desulphovibrio desulphuricans. Nature (London) 182: Postgate, J. R., and L. L. Campbell Classification ofdesulfovibrio species, the nonsporulating sulfate-reducing bacteria. Bacteriol. Rev. 30: Reddy, C. A., M. P. Bryant, and M. J. Wolin Characteristics of S organism isolated from Methanobacillus omelianskii. J. Bacteriol. 109: Reddy, C. A., M. P. Bryant, and M. J. Wolin Ferredoxin and nicotinamide adenine dinucleotide dependent H2 production from ethanol in extracts of S organism isolated from Methanobacillus omelianskii. J. Bacteriol. 110: Saito, H., and K. Miura Preparation of transforming deoxyribonucleic acid by phenol treatment. Biochim. Biophys. Acta 72: Schildkraut, C. L., J. Mannur, and P. Doty DESULFOVIBRIO GROWTH IN LOW-SULFATE MEDIA 1169 Determination of the base composition of deoxyribonucleic acid from its buoyant density in CsCl. J. Mol. Biol. 4: Smith, P. H., and R. A. Mah Kinetics of acetate metabolism during sludge digestion. Appl. Microbiol. 14: Sorokin, Y. I Role of carbon dioxide and acetate in biosynthesis by sulphate-reducing bacteria. Nature (London) 210: Sueoka, N Variation and heterogeneity of base composition of deoxyribonucleic acids: a compilation of old and new data. J. Mol. Biol. 3: Toenen, D. E., P. G. Thiel, and W. A. Pretorius Substrate flow in anaerobic digestion. Adv. Water Pollut. Res. 11:11 29/1-29/ Thauer, R. K., K. Jungermann, and K. Decker Energy conservation in chemotrophic anaerobic bacteria. Bacteriol. Rev. 41: Varel, V. H., and M. P. Bryant Nutritional features ofbacteroides fragilis. Appl. Microbiol. 28: Vosjan, J. H Respiration and fermentation of the sulfate-reducing bacterium Desulfovibrio desulfuricans in a continuous culture. Plant Soil Sci. 43: White, D. C., M. P. Bryant, and D. R. Caldwell Cytochrome-linked fermentation in Bacteroides ruminicola. J. Bacteriol. 83: Widdel, F., and N. Pfennig A new anaerobic sporing acetate-oxidizing, sulfate-reducing bacterium: Desulfotomaculum (emend.) acetoxidans. Arch. Microbiol. 112: Winfrey, M. R., and J. G. Zeikus Effect of sulfate on carbon and electron flow during microbial methanogenesis in fresh-water sediments. Appl. Environ. Microbiol. 33: Wolin, M. J Interactions between H2-producing and methane-producing species. In H. G. Schlegel, G. Gottschalk, and N. Pfennig (ed.), Microbial formation and utilization of gases (H2, CH4, CO). Goltze, Gottingen. In press. 33. Zeikus, J. G., P. J. Weimer, D. R. Nelson, and L. Daniels Bacterial methanogenesis: acetate as a methane precursor in pure culture. Arch. Microbiol. 104: