INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

In the rumen, methanogenic bacteria utilize hydrogen (H₂) to reduce carbon dioxide (CO₂) and/or formate to methane (CH₄) as follows: CO₂ + 4H₂  CH₄ + 2H₂O. Eructation of methane constitutes a 3 to 12% loss of gross energy intake for ruminants (4, 5, 17, 27) and contributes to atmospheric methane concentrations implicated in global warming (24). Johnson and Johnson (29) recently estimated that beef cattle account for 67% of methane emissions by the U.S. cattle herd. Beef cows account for 21% of the herd and 40% of the herd's methane emissions. Feedlot cattle are principally steers and account for 15% of the herd and 6% of methane emissions. These two classes of cattle are widely divergent in their dietary management and tractability for fermentative modification. Previous efforts to minimize digestible energy loss by suppressing ruminal methanogenesis have included the use of antibiotics, ionophores, or halogenated methane analogs (15, 43). Short-duration experiments with these analogs succeeded in suppressing CH₄ production, but H₂ and formate accumulated (16) and food consumption by sheep was adversely affected (13). The accumulation of H₂ indicated that halogenated CH₄ analogs disrupted interspecies H₂ transfer and thus were not sufficiently selective in suppression of methanogenesis. As a consequence, it is now realized that minimization of ruminal methane production needs to involve a strategy whereby electron disposal via interspecies H₂ transfer is not disrupted, and it would be advantageous if reducing equivalents were deposited in a metabolite(s) which serves as a substrate for ruminant tissue metabolism. One novel approach is the involvement of reductive acetogenesis. Reductive acetogenic bacteria reduce 2 mol of CO₂ to acetate by oxidation of H₂ as follows: 2CO₂ + 4H₂  CH₃COOH + 2H₂O. Diversion of energy from eructated CH₄ to acetate by H₂/CO₂-consuming acetogenic bacteria could potentially enhance the energetic efficiency of ruminants and decrease methane emissions (34).

The potential importance of H₂/CO₂-utilizing acetogenic bacteria in the rumen has been described; however, their capacity to effect a total synthesis of acetate from H₂ and CO₂ in ruminal contents and the factors influencing the magnitude of this activity have not been investigated. The presence of H₂/CO₂-utilizing acetogenic bacteria in the rumen has been shown (22, 25). Acetitomaculum ruminis produces acetate via heterotrophic growth on, for example, glucose and ferulic acid and via autotrophic growth on formate, carbon monoxide (CO), and H₂/CO₂ (30). Acetate produced via autotrophic growth would constitute a competition with methanogenesis for hydrogen. While reductive acetogenesis is quantitatively important in the termite hindgut (6), there are apparently no reports of nonmethanogenic ruminants. On the basis of data for nonruminal isolates, it has been hypothesized that ruminal acetogenic bacteria are not able to compete with ruminal methanogenic bacteria due to a less effective H₂-scavenging ability, yet there is a paucity of evidence to support this hypothesis.

Recent efforts to study ruminal acetogens and acetogenesis have hinged on the use of 2-bromoethanesulfonic acid (BES), an analog for coenzyme M (21). This coenzyme is essential for the growth of some of the ruminal Methanobrevibacter species (35). Whereas chloroform was a nonselective inhibitor of methanogenesis and other metabolisms dependent on transmethylation (23), BES has been used as a selective inhibitor of ruminal methanogenic bacteria because it is a methylreductase inhibitor (44).

In this study, a series of experiments were conducted to assess the presence, endogenous activity, and competitiveness of reductive acetogenesis in the bovine rumen. When cattle were the source of inocula, their nutritional management was intended to simulate the beef cow and feedlot sectors of the U.S. beef cattle industry. The principal energy source in beef cow diets is forage (10), whereas feedlot cattle are fed a high-grain diet containing an ionophore (11). Specifically, the population size of reductive acetogenic bacteria was determined for cattle in these two scenarios, and endogenous reductive acetogenic activity was quantified in rumen-like incubations by using mass spectroscopy (MS). We also investigated the ability of A. ruminis 190A4 to compete with rumen methanogens by altering the concentration of this acetogen or H₂ in ruminal contents and measurement of its threshold for H₂.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Media and growth conditions. For enumeration studies, AC-11 medium, used for the cultivation of acetogenic bacteria from the termite hindgut (6), was modified for the cultivation of acetogenic bacteria from the rumen. AC-B1 contained (per liter) the following: KH₂PO₄, 0.28 g; K₂HPO₄, 0.94 g; NaCl, 0.14 g; KCl, 0.16 g; MgSO₄ · 7H₂O, 0.02 g; NH₄Cl, 0.5 g; CaCl₂ · 2H₂O, 0.001 g; trace mineral solution (25), 10 ml; vitamin solution (25), 10 ml; yeast extract, 0.5 g; NaHCO₃, 6.0 g; reducing agent (2.5% [wt/vol] each cysteine hydrochloride · H₂O and Na₂S · 9H₂O, pH 10.0), 10 ml; clarified bovine rumen fluid, 100 ml; BES (sodium salt; filter sterilized) as indicated, 20 ml; and resazurin, 0.001 g. Broth media were boiled and cooled under a flow of 80% N₂-20% CO₂ (vol%) prior to addition of reducing agent and NaHCO₃. The final volume per tube was 5 ml. The final pH of the medium was approximately 6.8, with a headspace gas of 304 kPa of 80% N₂-20% CO₂ (vol%). Serum tubes were incubated horizontally with 304 kPa of 80% H₂-20% CO₂ (vol%) or 80% N₂-20% CO₂ (vol%) at 39°C and 200 rpm.

Animal diets and rumen samples. For enumeration studies, a rumen sample was obtained via stomach tube 1.5 h postfeeding from each of four beef steers fed once daily a typical finishing, i.e., high-grain (90 corn and supplement:10 corn silage [dry matter basis]) diet containing 0.03 g of monensin/kg of dry matter. The average rumen sample pH for animals on the high-grain diet was 6.2 ± 0.2 (standard error of the mean [SEM]). A high-forage rumen sample was obtained 2 to 4 h postfeeding from each of four beef cows fed alfalfa-grass hay once daily. Average rumen sample pH for animals on the high-forage diet was 7.1 ± 0.1 (SEM). For in vitro competition studies, beef cows described above were subsequently fistulated and fed alfalfa-grass hay. A ventral rumen sample from each animal was obtained 2 h postfeeding and passed through a 2-mm-mesh screen in an anaerobic glove box. Samples were pooled across three cattle to incorporate animal variation into the rumen sample. The pooled rumen sample pH was adjusted to 6.1 since addition of NaH¹³CO₃ for in vitro incubations excessively increased the pH of unadjusted samples.

Enrichment and enumeration studies. H₂/CO₂-supported acetogenic bacteria were enumerated by the three-tube MPN method. Rumen fluid was transferred to an anaerobic chamber (gas phase, 78% N₂-17% CO₂-5% H₂ [vol%]) for preparation of serial dilutions. Six tubes of AC-B1 medium with 2.5 mM BES were inoculated with each dilution. The tubes were evacuated and pressurized to 304 kPa with a gassing manifold (1), three tubes with 80% H₂-20% CO₂ (vol%) and three with 80% N₂-20% CO₂ (vol%) for controls. Tubes were incubated for 8 to 12 days at 39°C and 200 rpm. Optical density of the enrichment cultures was monitored every other day, with periodic repressurization with the appropriate gas mixtures. H₂/CO₂-incubated cultures were considered positive for acetogenic bacteria when the optical density at 600 nm was 0.3 over that in the N₂/CO₂-incubated tubes, methane was not detected in the headspace gas, and the pH was 6.1.

¹³CO₂ fixation studies. Serum vials (70 ml) containing 60 mM potassium phosphate buffer were maintained in an anaerobic chamber (gas phase, 78% N₂-17% CO₂-5% H₂ [vol%]) for at least 1 h prior to autoclaving. After cooling, sterile BES (5.0 mM, final concentration; sparged with 100% N₂) was added to vials in the anaerobic chamber followed by 9 ml of a freshly strained, pooled rumen sample. Vials were evacuated for 5 min while being shaken at 200 rpm and were then pressurized with 124 kPa of 100% H₂ or 100 kPa of 100% N₂ with a gassing manifold (1). Residual CO₂ remaining after evacuation was less than 0.01 mmol as measured by the BaCO₃ method (31). Gas volumes were measured with a pressure transducer (model PX126-015DV; Omega Engineering, Stamford, Conn.) (20). The volume of gas was calculated from the measured internal pressure and with reference to a standard curve. Initial H₂ concentration in the headspace gas was approximately 6.0 mmol per vial. NaH¹³CO₃ solution (75 mM [final concentration] as determined by the BaCO₃ method [31]) was added to all vials followed immediately by addition of an A. ruminis 190A4 inoculum as indicated. NaH¹³CO₃ solution was prepared by dissolving NaH¹³CO₃ (greater than 98% ¹³C enriched; Isotec Inc., Miamisburg, Ohio) in sterile CO₂-free water containing 1% (vol/vol) reducing agent (described above). The final volume in the vials was 10.7 ml, and the initial and final pH of the reaction mixture were approximately 7.1 and 6.9, respectively. Vials were incubated at 39°C with shaking at 200 rpm for approximately 48 h. Reactions were terminated by addition of 0.4 ml of concentrated HCl. After mixing and allowing CO₂ equilibration for 15 min, final gas volumes were measured. Headspace gas and liquid phase were sampled for gas chromatographic (GC), liquid chromatographic, and GC-mass spectometry (MS) analyses.

Pure-culture H₂ threshold studies. Bacteria were cultured in 125-ml serum bottles each containing 25 ml of medium under 232 kPa of 80% H₂-20% CO₂ (vol%). After incubation, the bottles were flushed with sterile 80% N₂-20% CO₂ (vol%) and evacuated three times, and the contents were pooled. To the pooled cultures, an equal volume of sterile medium was added under a flow of 80% N₂-20% CO₂ (vol%). Aliquots of 5 ml of suspension were distributed into sterile tubes under a flow of 80% N₂-20% CO₂ (vol%). Tubes were sealed with butyl rubber stoppers held in place with aluminum seals. Ten milliliters of 80% N₂-20% CO₂ (vol%) was added to some of the tubes to serve as controls for endogenous H₂ production. Four of these tubes immediately were analyzed for H₂ so that the initial background concentration of H₂ could be estimated. Finally, 10 ml of 1.2% H₂-75.2% N₂-23.6% CO₂ (vol%) was added to another set of tubes to provide 6,000 ppm of H₂ as the substrate for estimation of the H₂ thresholds. The initial H₂ concentration was in excess of the H₂ threshold values measured for all selected microorganisms. Additional tubes of medium only (uninoculated) were prepared as described above to serve as controls for abiological consumption and production of H₂.

Liquid and gas chromatographic analyses. The concentration of VFAs (formate, acetate, propionate, and butyrate) was measured by liquid chromatography. Samples were prepared and analyzed as described by Barlaz et al. (2). The gas phase of the enrichment cultures and in vitro incubations was analyzed for H₂, CH₄, and CO₂ by GC by injecting the sample into a Packard 438 gas chromatograph (Chrompack, Raritan, N.J.) equipped with a model 914 thermal conductivity detector and HP 3390A integrator (Hewlett-Packard, Avondale, Pa.) for data acquisition. The column used was 120/140 Carbosieve S2 (2.7 m by 3.2 mm [outside diameter]; Supelco, Bellefonte, Pa.). Operating conditions for the gas chromatograph were as follows: N₂ carrier gas, 30 ml/min; column temperature, 200°C; injector and detector temperature, 210°C; and injection volume, 0.4 ml.

MS analysis. Mass spectra of VFAs were determined by GC-MS to calculate the ratio of ¹³C- to ¹²C-labeled acetate, propionate, and butyrate. Samples for GC-MS were derivatized to butyl esters and extracted from an aqueous phase into hexane as described by Salanitro and Muirhead (40). The mass spectrum fragmentation pattern for butyl acetate compared well with published spectra and the fragment ion peak for [U-¹²C]acetate. Derivatization yielded an average 98.5% recovery for VFA compounds. GC-MS samples were carried onto a DB-1 column (0.25 mm by 12 m; J&W Scientific, Folsom, Calif.) fitted in an HP 5890 gas chromatograph operating in split mode (1:50). Helium was the carrier gas, with a mean linear velocity of 0.8 ml/min. The temperature was held at 40°C for 2 min, programmed to increase to 70°C at 5°C/min, then incremented to a final temperature of 120°C at 20°C/min and held at that temperature for 1 min, for a total run time of 11.5 min. Eluting compounds were volatilized into an HP 5970 mass selective detector (70-eV ionization) controlled by an HP UNIX data station, and total ion chromatograms were reconstructed. From abundance ratios of fragment ions, e.g., O¹²C¹²CH₃ (mass = 43), O¹³C¹²CH₃ or O¹²C¹³CH₃ (mass = 44), and O¹³C¹³CH₃ (mass = 45) for butyl acetate species, and quantitation of total VFAs by liquid chromatography, the concentration of a ¹³C isotope was calculated. Recovery of [U-¹³C]acetate was 101.0%, which indicated that the method was sufficiently accurate to use in quantitative fermentation studies.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Enumeration of H₂/CO₂-supported acetogenic bacteria from the rumen. H₂/CO₂-supported acetogenic bacteria in the rumen were enumerated by using a selective inhibitor of methanogenesis, BES. For cattle receiving a high-forage diet, H₂/ CO₂-supported acetogenic bacteria ranged in population density from 3.5 × 10¹ to 2.6 × 10⁵ cells/ml of rumen fluid (Fig. 1A). The total viable anaerobe concentration was between 1.5 × 10⁸ and 4.8 × 10⁸ CFU/ml of rumen fluid for all high-forage enrichments. For cattle receiving a typical finishing (high-grain) diet, H₂/CO₂-supported acetogenic bacteria were less numerous (P < 0.05) than for the high-forage diet and ranged in population density from 2 to 75 cells/ml of rumen fluid (Fig. 1B). Total viable anaerobic population for this diet was 1.4 × 10⁹ to 4.7 × 10⁹ CFU/ml of rumen fluid. Additional enrichments from the high-grain diet rumen samples were done on AC-B1 medium at pH 6.1. The medium at pH 6.1 was expected to be more habitat simulating (for high-grain diet) than the medium at pH 6.8; however, no H₂/CO₂-supported acetogenic bacteria were detected (data not shown). H₂/CO₂-supported acetogenic cultures were confirmed to be acetogenic. Acetate accumulated in the H₂/CO₂-incubated tubes (58.6 mM ± 6.3) over that in the N₂/CO₂-incubated tubes (16.7 mM ± 0.6) for all 45 positive MPN cultures analyzed in duplicate (mean ± SEM, P < 0.001). Repressurization was done during the enumeration protocol, which precluded the evaluation of the reaction stoichiometry. Cell morphologies observed in enrichment broths were similar to that of A. ruminis 190A4 in addition to a long rod and a coccus.

View larger version (43K): [in this window] [in a new window]   FIG. 1.   Enumeration of acetogenic and total viable anaerobic bacteria from the bovine rumen. Acetogenic bacteria were enriched on a rumen fluid-based medium containing 2.5 mM BES and enumerated by the MPN method. (A) Enumeration from ruminal contents of cattle fed a high-forage diet. Bars of the same pattern represent duplicate enumeration from the same animal done within 1 to 5 months. (B) Enumeration from ruminal contents of cattle fed a high-grain diet. Each bar represents one enumeration from one animal.

Endogenous rumen acetogenic activity. Activity of endogenous rumen acetogenic bacteria was investigated by incubating ruminal contents with 100% H₂ gas phase and NaH¹³CO₃. When methanogenic bacteria were inhibited by BES, 0.18 mmol of net [1- or 2-¹³C]acetate-carbon and 0.2 mmol of net [U-¹³C]acetate-carbon accumulated, indicating ruminal acetogenic activity (Table 1). H₂ and CO₂ were in excess during the entire incubation period, and methane was not detected in the headspace gas (data not shown). The concentration of residual gas was not determined. There was a specific enhancement of acetogenesis, as shown by ¹³C incorporation into acetate under H₂ compared to N₂, presumably due to limiting endogenous reducing equivalents (i.e., H₂ or formate). Only trace amounts of ¹³C were associated with propionate and butyrate (<0.02 mmol).

Role of population density in reductive ruminal acetogenic activity: in vitro incubations with nonlimiting concentration of H₂. Whether endogenous ruminal acetogenesis could be limited by the population density of acetogenic bacteria was addressed in the following experiments. Ruminal contents were incubated with 100% H₂ gas phase, NaH¹³CO₃, BES, and an A. ruminis 190A4 inoculum. When methanogenic bacteria were inhibited by BES, accumulation of acetate-carbon was positively correlated with the population density of added A. ruminis 190A4 from 10³ to 10⁶ CFU/ml (Fig. 2). Maximum levels of ¹³CO₂ fixed into acetate, as indicated by the final concentration of carbon in singly and doubly labeled acetate, were observed when the final concentration of A. ruminis 190A4 added to ruminal contents was 10⁵ to 10⁷ CFU/ml of ruminal contents. An average of 0.39 mmol of total acetate (equivalent to 0.78 mmol of acetate-carbon) accumulated from the utilization of approximately 1.7 mmol of H₂. Based on accumulation of doubly and singly labeled acetate, levels of acetogenic activity for these levels of inoculation were approximately 50% greater than that found for H₂-supported, net endogenous acetogenic activity in the presence of BES (Table 1). The concentration of A. ruminis 190A4 required for enhancement of rumen acetogenic activity was equal to or greater than the acetogen population enumerated in animals fed the high-forage diet (Fig. 1A). Growth of A. ruminis 190A4 was not negatively affected by a BES concentration of 5 mM (data not shown). Methane was not detected in these vials which contained BES (data not shown). Low levels of propionate and butyrate accumulated; however, only trace amounts (<0.02 mmol) of ¹³C were associated with propionate and butyrate. All values in Fig. 2 represent strictly H₂-dependent fixation products since the amount of CO₂ incorporated into a particular product under H₂ was corrected for the accumulation of the same product under an N₂ atmosphere.

View larger version (44K): [in this window] [in a new window]   FIG. 2.   H2-dependent [13C]acetate accumulation in rumen-like incubations amended with A. ruminis 190A4 and BES. Reaction mixtures contained 9 ml of ruminal contents, 5 mM BES, 0.8 mmol of NaH13CO3, A. ruminis 190A4 inoculum, 100% H2 gas phase (6.0 mmol), and other reagents as indicated in Materials and Methods. Vials were incubated for 48 h at 200 rpm and 39°C prior to sample collection for [13C]acetate analysis by MS. Means of the same bar pattern with different letters are different (P < 0.025). Data are presented as means + SEMs for three determinations.

View larger version (20K): [in this window] [in a new window]   FIG. 3.   H2-dependent [13C]acetate and methane accumulation in rumen-like incubations with A. ruminis 190A4 and active methanogenic bacteria (without BES). Reaction mixtures contained 9 ml of ruminal contents, 0.8 mmol of NaH13CO3, A. ruminis 190A4 inoculum, 100% H2 gas phase (6.0 mmol), and other reagents as indicated in Materials and Methods. Vials were incubated for 48 h at 200 rpm and 39°C prior to sample collection for methane and [13C]acetate analysis by GC and MS, respectively. [13C]acetate-carbon is the summation of [1- or 2-13C]- and [U-13C]acetate-carbon. Each point represents the mean ± SEM for three determinations.

Role of H₂ concentration in ruminal reductive acetogenic activity: in vitro incubations with limiting concentration of H₂. To assess the competitiveness of acetogenic bacteria in rumen-like conditions, in vitro incubations were conducted with an ecologically common source of H₂. The accumulation of H₂ in the headspace gas over time was measured in vials containing ruminal contents incubated with 100% N₂ gas phase, NaH¹³CO₃, and alfalfa in the presence or absence of BES and/or A. ruminis 190A4. Alfalfa was added as a slowly degradable carbohydrate source for the production of a rumen-like H₂ concentration. Under conditions with active methanogenesis (without BES), H₂ was generated from the alfalfa fermentation and utilized for methanogenesis (Fig. 4). These vials which supported methanogenic activity contained approximately 0.002 mmol of gaseous H₂, which was equivalent to an H₂ equilibrium concentration in the headspace gas of less than 400 ppm during the entire 43-h incubation. This H₂ concentration was independent of added A. ruminis 190A4 (Fig. 4A; P > 0.95). The H₂ equilibrium concentration was considered to be the balance between H₂ production and utilization. Methanogenic activity was evidenced by the accumulation of headspace methane and also found to be similar for vials with or without 4.1 × 10⁸ A. ruminis 190A4 CFU/ml (Fig. 4B; P > 0.9), suggesting minimal H₂/CO₂-supported activity of A. ruminis 190A4 in methanogenesis-supportive incubations. Furthermore, singly or doubly labeled acetate-carbon did not accumulate in vials with active methanogenesis and added A. ruminis 190A4 to levels above those measured in control vials without the addition of A. ruminis 190A4 (Table 2; P > 0.2).

View larger version (12K): [in this window] [in a new window]   FIG. 4.   Hydrogen and methane concentrations over time in vials with ruminal contents incubated with alfalfa and 100% N2. Reaction mixtures contained 9 ml of ruminal contents, with or without 5 mM BES, 0.82 mmol of NaH13CO3, A. ruminis 190A4 inoculum, 0.38 g of alfalfa, and other reagents as indicated in Materials and Methods. When added, the final concentration of A. ruminis was 4.1 × 108 CFU/ml. Symbols: open circles, with BES; closed circles, with BES and added A. ruminis 190A4; open triangles, without BES and without addition of A. ruminis 190A4; closed triangles, without BES and with added A. ruminis 190A4. Each point represents the mean ± SEM for three determinations. Standard errors were plotted but are too small to be visible on the graph.

Pure-culture H₂ threshold studies. Competition for H₂ can be partially explained by the threshold model, which states that the successful organism keeps the H₂ partial pressure below the level necessary to allow H₂ oxidation by competitors. For this reason, hydrogen threshold values for selected acetogens and a methanogen were determined (Table 3). The H₂ threshold concentrations for A. ruminis 190A4 and the methanogen were consistent with the equilibrium concentrations found in in vitro incubations with an alfalfa fermentation providing H₂ in the presence and absence of BES, respectively (Fig. 4A). When S. termitida was incubated under N₂-CO₂, H₂ was produced to a level of 1,200 ppm. It appears that S. termitida produces H₂ to a level near its threshold. As with S. termitida, A. woodii can produce H₂ (328 ppm) up to its threshold level. Approximately 40 ppm of H₂ was found in all inoculated tubes analyzed just after initial pressurization with N₂-CO₂. This value is at most only 35% of the lowest threshold value presented. In incubated, uninoculated tubes pressurized with N₂-CO₂, there was no abiological production of H₂ (data not shown).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this study, we have demonstrated that reductive acetogenic bacteria are inhabitants of the rumen ecosystem yet have negligible endogenous H₂/CO₂-consuming activity. Reductive acetogenic activity was enhanced in rumen-like incubations when (i) an axenic culture of the rumen acetogen A. ruminis 190A4 was added under conditions of H₂ excess or (ii) methanogenesis was selectively inhibited by BES and A. ruminis 190A4 was added to incubations with a limiting concentration of H₂. These data, in addition to those from pure-culture H₂ threshold studies, confirm that ruminal methanogenic bacteria limit reductive acetogenesis by lowering the H₂ partial pressure below the minimum level necessary for H₂ consumption by a ruminal acetogen, A. ruminis 190A4.

H₂/CO₂-supported acetogenic bacteria in the rumen were enumerated (Fig. 1A and B) and found to be present at concentrations greater than those found in the feed and water (31a). Bryant suggested that the number of a given species present in the rumen compared to its numbers in the feed and water consumed is probably the best measure of whether an organism is a true rumen microorganism (8). A large variability was observed within duplicate enrichments from animals 05, 134, 6305, and 69, which was believed to be due to sample variation instead of the MPN technique, since good repeatability was associated with the latter (31a). Leedle and Greening found H₂/CO₂-utilizing acidogenic bacteria in the rumens of steers fed a high-forage (3.9 × 10⁸ cells per g of ruminal contents) or a high-grain (8.7 × 10⁸ cells per g of ruminal contents) diet (30). In contrast, we found H₂/CO₂-supported ruminal acetogenic bacteria at a concentration at least 1,000-fold lower than that found by Leedle and Greening (30). This discrepancy could be explained by sample variation due to a rumen sample obtained via stomach tube versus through a cannula (30) or an overestimation of H₂/CO₂-consuming acidogens due to enumeration via bromocresol green-staining colonies. We found that beef cows fed a hay diet supported a greater population density of ruminal reductive acetogenic bacteria than did steers fed a finishing diet. Some of the determinants affecting the population density of H₂/CO₂-consuming ruminal acetogenic bacteria may be the presence of alternate energy sources in the hay diet, inhibition by monensin in the high-grain diet, and/or lower rumen pH with the high-grain diet. The rumen pH of animals fed the high-grain diet (<6.0) was lower than that found for animals fed the high-forage diet (<6.5) because starch is a readily fermentable substrate (3). H₂/CO₂-consuming acetogenic bacteria have pH optima close to 7.0. Optimal pHs for A. ruminis and Eubacterium limosum are pH 6.8 and 7.2, respectively; therefore, a lower rumen pH would likely be inhibitory to the growth of acetogenic bacteria (22, 30). The lack of H₂/CO₂-supported acetogenic bacteria in enrichments from high-grain diet rumen samples done in pH 6.1 medium support this hypothesis. Also, monensin disrupts K⁺ and Na⁺ gradients, which are usually associated with inhibition of ruminal gram-positive bacteria (39). A. ruminis tends to be gram variable but stains gram positive in 24-h cultures (25).

In in vitro incubations with a nonlimiting concentration of H₂, endogenous acetogenic activity was not detected in rumen-like incubations unless methanogenesis was selectively inhibited by BES (Table 1). This activity in the presence of BES could be further enhanced by the addition of an axenic culture of the rumen acetogen A. ruminis 190A4 (Fig. 2). Reductive acetogenic activity in rumen-like incubations without BES was dependent on the concentration of added A. ruminis 190A4 under conditions of H₂ excess (Fig. 3). Simultaneous activity of acetogenic and methanogenic bacteria was indicated only when the population density of A. ruminis 190A4 was 10⁷ CFU/ml, nearly approaching the rumen concentration of methanogenic bacteria (10⁸ cells/ml). Hence, the population density of A. ruminis 190A4 dramatically influenced the accumulation of acetate resulting from H₂/CO₂-utilizing activity. These results are similar to those found by Nollet et al. upon addition of the acetogen Peptostreptococcus productus to ruminal contents (36).

Singly labeled acetate most likely represents a total synthesis of acetate from CO₂ arising from the fixation of 1 mol of ¹²CO₂ and 1 mol of ¹³CO₂ into [1- or 2-¹³C]acetate (41). Attempts were made to remove soluble ¹²CO₂ from ruminal contents; however, complete removal could not be attained (<0.01 mmol of residual CO₂). In addition, there was presumably production of ¹²CO₂ from residual substrates in ruminal contents. The [1- or 2-¹³C]acetate could also arise from an exchange reaction between ¹²CO₂ and the carboxyl or methyl group of [U-¹³C]acetate or [U-¹²C]acetate (46). We found that <15% ¹²CO₂ is exchanged with [U-¹³]acetate in ruminal contents, and thus only a small portion of [¹³C]acetate may not represent a total synthesis of acetate. For this study, doubly labeled acetate was used to confirm total synthesis of acetate from CO₂ and the summation of singly and doubly labeled acetate was used to quantitate H₂/CO₂-utilizing acetogenic activity. Our ability to detect CO₂ fixation activity by MS was verified by analysis of culture supernatants of A. ruminis 190A4 incubated in pure culture with H₂ and NaH¹³CO₃. Mass spectra revealed explicitly the accumulation of [U-¹³C]acetate, which was further confirmed by nuclear magnetic resonance analysis (31a).

It has been hypothesized, on the basis of nonruminal acetogenic isolates, that ruminal acetogenic bacteria cannot compete with ruminal methanogenic bacteria because they have less effective H₂-scavenging ability. We have shown that two representative ruminal acetogens, A. ruminis 139B and A. ruminis 190A4, have H₂ threshold concentrations which are 30- to 37-fold greater than that of the methanogen 10-16B (Table 3) and 3-fold higher than the median ruminal in situ H₂ concentration (1.0 µM in the aqueous phase, which is equivalent to 1,360 ppm in the atmospheric phase) (18, 27, 38, 42, 45). The ruminal acetogens had 4- to 13-fold-higher threshold values compared to nonruminal H₂-utilizing acetogenic bacteria. This may be related to the fact that ruminal acetogens have not been selected in vivo on the basis of H₂-scavenging ability. For example, S. termitida had a lower H₂ threshold value than A. ruminis and was isolated from a wood-eating termite hindgut, where H₂-dependent acetogenesis is the dominant H₂ sink reaction (7). H₂ threshold values for S. termitida and A. woodii were in close agreement with the values of 830 and 520 ppm, respectively, reported by Cord-Ruwisch et al. (14). The pure-culture H₂ threshold data indicate that the two rumen acetogens have poorer H₂-scavenging ability than methanogens in pure culture. Supporting this conclusion, in in vitro incubations with a limiting concentration of H₂, methanogenic bacteria and A. ruminis 190A4 maintained H₂ concentrations of approximately 400 and 4,800 ppm, respectively (Fig. 4), which are similar to their H₂ thresholds. This finding demonstrated that ruminal methanogenic bacteria limited acetogenesis by lowering the H₂ partial pressure to a level insufficient for H₂ utilization by A. ruminis 190A4. This observation agrees well with our understanding of methanogenesis as the predominant H₂ sink in the rumen (26, 27). Redirection of ruminal H₂ disposal seems to require a strategy for compromising H₂ consumption by ruminal methanogens and selection of a ruminal acetogen with a H₂-scavenging ability approaching that of ruminal methanogens. Failure of an alternative H₂ disposal strategy to achieve equally low ruminal in situ H₂ concentrations could be deleterious to the thermodynamics of interspecies H₂ transfer and, hence, degradative activity of ruminal fermentative bacteria.

In summary, the coexistence of methanogenic and reductive acetogenic bacteria in the rumen suggests that the acetogens grow on substrates other than H₂ and CO₂ (i.e., noncompetitive substrates such as organic compounds) in situ or there is an abundance of competitive substrates (i.e., H₂) in the rumen (33, 37) due to diurnal fluctuations of H₂ and/or juxtapositioning between H₂-producing and H₂-consuming microorganisms (3, 18, 38, 42).