Abstract:
37 Hydrogen production and transfer in the rumen R.S. Hegarty1 and R. Gerdes2 1 2 NSW Agriculture Beef Industry Centre, Armidale NSW 2351 c/_ Division of Molecular and Cellular Biology, University of New England,Armidale NSW 2351 Summary Hydrogen is a central metabolite in rumen fermentation and its partial pressure is an important determinant of rumen methanogenesis. The balance of hydrogen ion (H+) and dissolved hydrogen gas (H2) concentrations directly determine the redox potential of the rumen and therefore the possible extent of oxidation of feedstuffs. Differential permeability of the bacterial membrane to H+ and H2 means the proportions of these metabolites in the extracellular fluid will differ from their proportions within the bacterium. Interpretation of extracellular H2 concentration in the bulk fluid is also complicated by interspecies hydrogen transfer among H 2 producing and utilising rumen organisms which live in close physical (syntrophic) associations. The role of H2 concentration in determining the array of fermentation products and the efficiency of energy capture by rumen organisms is discussed. Developing a greater understanding of rumen hydrogen metabolism may present new opportunities for the control of rumen methanogenesis. Introduction Because ruminant livestock contribute approximately 16% of global methane emissions (Moss 1993), rumen methanogenesis is increasingly a focus of nutritional research (Van Nevel and Demeyer, 1996; Mathison et al. 1998). Early studies demonstrated the rate of methane production was directly proportional to the rate of hydrogen uptake by rumen fluid (Czerkawski et al. 1972). The direct relationship between the partial pressure of hydrogen (p H2 ) and methanogenesis suggests some research emphasis should rightly move from methanogenesis per se to factors affecting the availability of hydrogen within the rumen. Hydrogen for methane synthesis occurs in three key states in the rumen, these being hydrogen gas (H 2), reduced cofactors (such as NADH and NADPH), and as free protons. While hydrogen gas rarely accumulates to concentrations greater than 1% in headspace gas (Table 1), hydrogen in all its forms is a central regulator of rumen fermentation and for this reason it has been referred to as the currency of rumen fermentation Table 1 Species Cattle2 Sheep1 Reported composition of the rumen headspace gas. Nu trition Unspecified H a y _prefeed _feeding _2hr postfeed CO 2 CH4 26.8 36.2 12.0 33.0 26.1 16.3 24.8 0.07 16.9 H2 0.18 0.033 0.046 0.062 0.023 0.319 0.135 23.82 0.18 CO _ nd nd nd <0.001 0.001 0.00 1_0.01 0.15 0.25 O 2 N2 7 14.9 51.4 17.1 16.8 37.5 5.3 17.41 18.14 H2S 0.01 nd nd nd nd nd nd nd nd 63.4 47.1 24.5 47.5 54.3 35.7 68.4 56.4 60.9 0.56 2.1 10.2 2.4 4.5 10.5 1.5 1.92 3.61 Sheep1 Concentrate _prefeed _feeding _2hr postfeed Cattle3 Non_bloated Bloated Source: (1) Barry et al. 1977 (2) McArthur and Miltimore 1961 (3) Olson 1940; data for bloated cattle averaged for animals consuming sweet clover and lucerne; data for non_bloated animals averaged for cattle consuming sweet clover, lucerne, sudan grass and sorghum; nd, not determined. Recent Advances in Animal Nutrition in Australia, Volume 12 (1999) 38 Hegarty, R.S. and Gerdes, R. (Czerkawski 1986). Additionally, the H+/H2 couple is used as the reference for microbial energetics and checking hydrogen balance has been the accepted way of validating results for in vitro incubations for over 20 years (Demeyer and Van Nevel 1975). Despite its importance, the kinetics of distribution and transfer of rumen hydrogen are poorly understood. This paper links basic chemical data and microbial biochemistry of hydrogen with its significance in the rumen and explores potential opportunities for improving animal production as a result of this understanding. Distribution of hydrogen in the rumen Examination of the distribution of hydrogen atoms in the rumen shows that over 90% of rumen hydrogen atoms reside in water (Table 2). The protons in water are not, however, free to exchange with other protons except when liberated or utilised in hydrogenation/hydration reactions. The same is true for other covalently bound protons such as those of the methyl group of acetate (Leng and Leonard 1965) which do not exchange with water in an aqueous solution. In rumen fluid however, substantial proton transfer occurs during microbial fermentation such that a substantial proportion of the hydrogen atoms in methane and hydrogen gas are derived from water (Figure 1). The difference in flux of substrate hydrogen and substrate carbon flux due to H + transfer means that tritiated and deuterated compounds are not useful as tracers of organic acid production in rumen fermentation studies (Leng and Leonard 1965; Hungate 1966). Being a nonpolar gas, H2 is poorly soluble in water and its solubility is unaffected by pH since it does not ionise. Based on the solubility of hydrogen in water (IUPAC 1981) and on typical H2 concentrations in the rumen headspace (Table 1) it can be estimated that at 39oC, the concentration of dissolved hydrogen in bulk rumen fluid would vary between about 90 mM and 250 mM. The effects of headspace H2 pressure on fermentation are equivocal. Czerkawski et al. (1972) changed the H2 concentration in the headspace of rumen fluid incubations and showed H2 to be utilised as fast as it was able to dissolve. However Nelson et al. (1960), reported no change in the CO2: CH4 ratio in rumen gas during sparging of the rumen with H 2, suggesting methane production did not respond to the concentration of H2 in the rumen gas. The rate at which H2 dissolves in solution may be an important difference in these studies. It is likely that cultures grown in closed atmospheres for long periods permit the dissolved and headspace gases to equilibrate which is not possible in the rumen due to H2 being slow to dissolve. For this reason, fermentation patterns change in laboratory cultures in response to a higher H2 pressure in the headspace (Van Nevel et al. 1974; Hillman et al. 1991) but not in situ (Nelson et al. 1960). The mass transfer of dissolved hydrogen may be limiting to growth of methanogens if reliant on H2 dissolved in the bulk medium, and so intimate (syntrophic) associations exist between many hydrogen utilising microbes such as methanogens, and those which produce hydrogen (Conrad et al. 1985). Direct physical associations to facilitate syntrophy exist in the rumen such as those between rumen protozoa and endo and ectosymbiotic methanogens (Newbold et al. 1995; Ushida et al. 1997; Hegarty 1999). In sludge ecosystems, less than 10% of hydrogen transfer may Figure 1 Changes in the specific activity (dpm per �g atom H) of H2 (�), water (n) and methane (p) when rumen contents were incubated under unlabelled H2 gas with 3H2O (Czerkawski 1975). Table 2 Calculated distribution of hydrogen atoms in the rumen of a sheep assuming typical pool sizes of rumen digesta and metabolites. Pool size 4000 mL 450g DM 163g DM 40mL <1mL H (g) 444.0 30.2 10.4 0.007 <0.0001 % 91.6 6.2 2.1 <0.1 <0.1 Constituent Water Feed Microbes Methane Hydrogen gas Hydrogen production and transfer in the rumen 39 occur via H2 dissolved in the bulk fluid (Conrad et al. 1985) and it should not be expected that hydrogen in the rumen gas space will equilibrate with, or be indicative of, the availability of dissolved hydrogen used by many rumen bacteria. The importance of close physical association in syntrophy was shown by Boone et al. (1989) using coculture of a methanogen and a (H2 liberating) butyrate degrader. By their calculations Boone and coworkers showed that unless microbes were within 10mm of each other, the presence of a H2 producing microbe did not provide an advantage to the methanogen, other than by increasing the H 2 concentration in the bulk solution (Figure 2) The partial pressure of H2 has substantial effects on the energetics of methanogens themselves and extremely low or high pH2 may cause the ATP yield to vary from 0.33 to 2 moles of ATP per mole of methane produced respectively (Keltjens and Vogels 1996). Intracellular hydrogen ions Since H2 is nonpolar it can be assumed to pass freely through microbial membranes, allowing intra and extra cellular H 2 concentrations to move towards, but probably not achieve, equilibrium. In contrast to H2, the hydrogen ion concentration in the cytoplasm is a condition over which microbial cells exert tight control. The intracellular pH for all living cells is by necessity more alkaline than the surrounding environment. Extensive studies of Streptococcus bovis (Russell 1987) and Streptococcus faecalis (Harold et al. 1970) have found the internal pH for these organisms to be 0.41 unit higher than the medium. Intracellular pH is not a constant however, and acidtolerant rumen bacteria allow their intracellular pH to decline in proportional to extracellular pH (Russell 1991). Cytoplasmic pH is a function of transmembrane proton migration (controlled by ATPase and symport systems) and production of acids within the cytoplasm (Booth 1985). Bacterial membranes have controlled permeability to protons, so passive inward diffusion of protons is minimal. The fact that bacteria import nutrients by either iondriven or ATP driven transport (Russell and Strobel 1993) means maintenance of the pH differential across the membrane is critical for growth. The action of monensin in decreasing the trans membrane Na+/K+ gradients can cause internal proton accumulation and decrease intracellular pH (Russell 1987). This change in proton gradient is associated with a reduced microbial growth efficiency which is most severe at low extracellular pH (Chow and Russell 1990). Production of fermentation acids in the cytoplasm and their flux across the bacterial membrane play some part in intracellular pH regulation (Thauer et al. 1977). The inability of ionised acids to diffuse out of cells can cause accumulation and feedback inhibition of VFA production. Also, undissociated acids are lipophilic and can pass from the relatively acidic external environment into the cell carrying a proton and then, at the higher intracellular pH, dissociate releasing the proton into the cytosol. Because the ionised VFA cannot leave the cell, VFA entry can only serve as a means of introducing protons, reducing rather than improving the cells capacity to maintain a proton gradient across the membrane. Unless the cells do decrease their intracellular pH as external pH declines, extreme levels of intracellular VFA could accumulate (Russell and Strobel 1993). Build up of these acids intracellularly is thought to be one way by which the balance of end products changes with external pH although the role of ADP/ATP and NADH/NAD+ ratios in determining VFA balance must also be considered, as discussed later. Transmembrane VFA entry may therefore serve more as a means of changing the pattern of VFA a cell produces via endproduct inhibition, than as a means to control internal pH or uncoupling fermentation and growth. Sources of rumen hydrogen The atmosphere contains approximately 0.0005% H2 and there is no H2 carrier system in blood so inhaled or engulfed air will contribute little to H 2 in the rumen. A daily production of approximately 100 L of hydrogen gas may be anticipated in the rumen of a sheep on grounds that 4 moles of hydrogen are required per mole of methane produced and 25 L/d is a typical emission of methane by sheep (Pelchen and Peters 1998). Rumen H2 is liberated by hydrogenase enzymes acting on reduced ferredoxin arising from either phosphoroclastic reactions (which do not involve NADH) or as part of a redox couple which allows reoxidation of reducing equivalents, such as NADH. The bacterial hydrogenases which reduce protons to H2 are single polypeptide chains (Van Dijk et al. 1979) or multisubunit enzymes (Sawers 1994) which may contain iron/sulphur clusters and nickel. In rumen protozoa (Paul et al. 1990) and fungi (Yarlett et al. 1986) hydrogenase enzymes are principally located within the hydrogenosomes. The pyruvate synthase, hydrogenase Figure 2 Diffusion of H2 to a coccus 1 �m in diameter. Diffusion assumes a steady_state concentration profile of H2 near a H2 consuming cell with a cell surface concentration of 0, 17.5 or 35 nM and a bulk phase concentration of 63 nM (Boone et al. 1989). 40 Hegarty, R.S. and Gerdes, R. and lactate dehydrogenase enzymes (Yarlett et al. 1981) make the hydrogenosome the principal site of H2 and CO 2 generation. The activity of protozoal and Neocallimastix hydrogenases are inhibited by carbon monoxide and a high pCO2 drives fermentation towards H2 consuming reactions such as butyrate and lactate production (MarvinSikkema et al. 1993; Ellis et al. 1991). Effects of pH and Eh on the availability of hydrogen gas to microbes Considering the differences in intracellular and extracellular H+ concentrations caused by the cell membrane, it is of some interest to understand what linkages, if any, exist between extracellular pH and intra cellular H2 availability. The microbial and biochemical mechanisms by which consumption of high levels of readily fermentable carbohydrates reduce rumen pH (increase extracellular proton concentration) are well understood (Owens et al. 1998). The partial pressure of H2 is directly related to both H+ ion concentration (pH) as well as the redox potential of the rumen according to the Nernst Equation (equations 1, 2) which may be simplified for rumen conditions to equation 3 (Sauer and Teather 1987). Oxidation potential (Eo) is referenced against a standard hydrogen electrode and describes the capacity of a system to donate or accept electrons when the concentration of the oxidised and reduced forms are equal. Because these concentrations are unequal in biological systems, the redox potential (Eh) is calculated according to equation 1. Equation 3 demonstrates that redox potential will change in direct proportion to hydrogen ion concentration. Eh = Eo RT/F ln ([oxidised form]/[reduced form]) (1) Oxidation of reducing equivalents The principal means of regenerating NAD+ from NADH is by the enzyme NADH ferredoxin oxidoreductase, coupled to a hydrogenase. The activity of NADH ferredoxin oxidoreductase is primarily controlled by concentration of dissolved H2, being suppressed by a high pH2 (Gottschalk 1986). NADH + H + Oxidised ferredoxin H 2 NADH ferredoxin oxidoreductase NAD + Hydrogenase Reduced ferredoxin Decarboxylation of pyruvate Anaerobic decarboxylation of pyruvate in the rumen is achieved by two enzyme systems. Both the pyruvate ferredoxin oxidoreductase (pyruvate synthase) and pyruvate formate lyase enzymes liberate ATP without use of NADH as an intermediate electron carrier (Doelle 1975; Figure 3). The formate liberated by pyruvate formate lyase is however rapidly degraded in the rumen to CO2 and H2 and formate rarely accumulates (Hungate 1970). Formate may also be directly reduced by some rumen methanogens (Hungate et al. 1970). In ecosystems where these occur, formate may be a more important vehicle for interspecies hydrogen (electron) transfer for syntropic rumen organisms than is H2 (Boon et al. 1989). Additional to decarboxylation of pyruvate, hydrogen is also produced by oxidation of butyrate back to acetate (McInerny et al. 1981). This fermentation is conducted by a number of organisms (e.g. Syntrophomonas wolfei) and cycling of carbon between acetate and butyrate may provide a futile cycle to uncouple ATP production from cell synthesis. There is almost no secondary fermentation of propionate in the rumen (Rowe et al. 1985) and this is possibly due to the high energy cost of producing acetate from propionate (DGo� = 71.7 kJ/mole ) compared to producing acetate from butyrate (DGo� = 48.3 kJ/mole; Boone et al. 1989) Eh = Eo RT/F ln ([H+]/[H2]) Eh = 0.062 * log [H+]/pH2 (2) (3) Equation 3 also shows the direct mathematical basis for why a rise in H2 partial pressure is associated with a decrease in Eh when pH is held constant, as observed in continuous fermenters (Sauer and Teather 1987). In the ruminant animal, feeding causes a sharp rise in p H2 and this is associated with a rise in headspace H2 concentration (mL/L) and a fall in Eh (Barry et al. 1977). The postfeeding rise in H2 concentration occurs in spite of dilution through introduction of atmospheric N2 and O2 with feeding (these gases being up to 50% and 9% of rumen gas respectively). The strong positive association between hydrogen ion concentration in the rumen and pH2 observed by Barry et al. (1977) is unlikely to result from a direct chemical mechanism. Since production of H2 is achieved through the action of intracellular hydrogenase enzymes, it is difficult to see how their activity could be altered by extracellular H+ ion concentration. Instead it may reflect specific inhibition of methanogenesis as pH declines (Lana et al. 1998; Russell 1998) which then causes an accumulation of unused H2 in the rumen. The antimethanogenic compounds which target the methanogen itself and cause H 2 accumulation reduce Eh, while those which inhibit methanogenesis Hydrogen production and transfer in the rumen 41 only by reducing H2 availability (e.g. monensin) do not decrease Eh (Sauer and Teather 1987). The impacts of pH2 on liberation of H2 from the phosphoroclastic reactions and reduced cofactors are as follows. H2 accumulation itself does not greatly inhibit hydrogenase enzymes and the phosphoroclastic production of acetate from pyruvate by pyruvate ferredoxin oxidoreductase is a fully reversible process. A high pH2 can therefore cause a direct feedback inhibition of the main acetate generating pathway in the rumen. A high pH2 also has a direct effect on the release of H2 from NADH. The action of NADH oxidoreductase is subject to endproduct inhibition by H2. At low H2 pressures (pH2< 1x 103 atm), H2 liberation becomes thermodynamically feasible (DGo� < 0 mV) but at high pH2, the reaction becomes endergonic and does not proceed (Gottschalk 1986). Because the action of NADH oxidoreductase is not reversible, NADH accumulates in times of high H 2 potential and so fermentation is stopped by lack of NAD+. Hydrogen availability as a control of rumen fermentation Just as methanogenesis is directly proportional to pH2 in the rumen, so the balance of VFA produced may be expected to change with hydrogen availability both in vitro and in vivo . Within the cell, hydrogen concentration moderates the NADH/NAD+ ratio which feeds back on fermentation pathways (Hino and Russell 1985). High substrate pressures which accelerate the glycolytic pathway induce a high NADH/NAD+ ratio which stimulates propionate synthesis so oxidising NADH. If the ATP/ADP ratio is also high (as may occur due to a sudden influx of immediately fermentable substrate such as molasses), propionate synthesis is suppressed and NADH is instead utilised in butyrate production. Either butyrate or propionate synthesis will reduce the NADH ratio and allow accelerated acetate production to continue (Sutherland 1977). Sutherland (1977) concluded that it was the free energy of interconversion which determined the flow of electrons and therefore the endproducts produced by adaptable organisms. This is in keeping with the view that the relative rates of component processes are optimised for overall ATP gain or thermodynamic efficiency (Thauer et al. 1977), as well as to maintain internal homeostasis in the cell. Applying the Nernst equation, the energetically favourable product could be determined by the redox state of the cell as reflected in NADH/NAD+ and ATP/ADP ratios as follows. The example of energetic optimisation of the acetate propionate balance (Gap) is shown below. G ap = G�apo + RT ln ([propionate]/[acetate]) *([ATP]/[ADP]*[PI])/([NADH]*[H + ][NAD + ]) 4 .[CO 2 ] 2 Figure 3 Reaction pathways which liberate H2 from pyruvate in the rumen ecosystem. 42 Hegarty, R.S. and Gerdes, R. Hydrogen effects on energy yield from fermentation Ruminant nutritionists often lament that the rumen is an inefficient means of extracting energy for the animal from its diet, focusing on the fact that the anaerobic environment reduces the yield of ATP per mole of glucose from 38 to 2 ATP. When the complete animal/ rumen unit is considered, however, there is little basis for such despondency. The efficiency with which released energy is captured in ATP from anaerobic fermentation (as far as it proceeds) is comparable to that of aerobes, being typically 2550% compared to 4459 % for aerobes (Thauer et al. 1977). A further similarity is that aerobes and anaerobes do not differ greatly in the efficiency of ATP use for cell growth and YATP may be higher for some rumen bacteria than for aerobes (Russell and Wallace 1997). What does differ greatly is how far oxidation proceeds in the anaerobic compared to the aerobic environment. In the aerobic bacterium or mammalian cell, sugars are largely oxidised completely through to CO2 while in the rumen, fermentation ceases at VFA. When the VFA are absorbed, the residual energy present in the fermentation products is made available to the host via aerobic respiration. Were rumen fermentation as extensive as in sewage fermenters, there would be no high energy compounds leaving the fermentation except methane, and no possibility of life for the ruminant. An important metabolic difference between aerobes and anaerobes is the failure of the TCA cycle to fully operate in anaerobes. Possible explanations for why the TCA cycle does not operate in anaerobic environments are: The TCA cycle does not proceed because the oxidation of succinate to fumarate has an Eo� of 33 mV and only electron acceptors with a more positive redox potential can support the reaction (e.g. O2, NO, NO3, ubiquinone) and these are absent (Thauer et al. 1977) There is a deficiency of NAD+ available, diminishing the NADH liberating steps in the TCA cycle The TCA cycle contains oxygen induced enzymes (Doelle 1975) one guanosine diphosphate from succinylCoA to succinate conversion and methanogens would capture some of the energy in hydrogen as ATP while reducing CO2. Conclusions The partial pressure of hydrogen is a significant determinant of the rate of rumen methanogenesis and the array of VFA produced in the rumen. The presence of significant syntrophic relationships between a range of H2 producing organisms and H2 utilisers means the pH2 in the headspace is unlikely to be an accurate reflection of intracellular H2 availability. The balance of hydrogen ions (oxidised form of hydrogen) and H2 (reduced form of hydrogen) determine the redox potential of the rumen and therefore the extent and products of fermentation. High H2 concentrations induce a low redox potential allowing production of highly reduced end products (e.g. propionate, methane). When pH2 is low, the fermentation balance swings away from these to an acetate fermentation. Strict control of the entry of hydrogen ions into bacteria at the membrane means extracellular pH does not have a direct chemical effect on H2 generation within the cell. The increase in pH2 associated with low ruminal pH is considered to be, instead, a consequence of low pH being inhibitory to H2 use by methanogens. Further understanding of factors which regulate H2 availability in the rumen, particularly those reactions which utilise H2, is likely to lead to recognition of new techniques to decrease methanogenesis in the rumen. References Barry, T.N., Thompson, A. and Armstrong, D.G. (1977). Rumen fermentation studies on two contrasting diets. 1. Some characteristics of the invivo fermentation with special reference to the composition of the gas phase, oxidation/reduction state and volatile fatty acid proportions. Journal of Agricultural Science, Cambridge 89, 183195. Boone, D.R., Johnson, R.L. and Liu, Y. (1989). Diffusion of interspecies electron carriers H2 and formate in methanogenic ecosystems and its implications in the measurement of Km for H2 or formate uptake. Applied and Environmental Microbiology 55, 17351741. Booth, I.R. (1985). Regulation of cytoplasmic pH in bacteria. Microbiological Reviews 49, 359378. Chow, J.M. and Russell, J.B. (1990). Effect of ionophores and pH on growth of Streptococcus bovis in batch and continuous culture. Applied and Environmental Microbiology 56, 15881593. Conrad, R., Phelps, T.J. and Zeikus, J.G. (1985). Gas metabolism evidence in support of the juxtaposition of hydrogenproducing and methanogenic bacteria in sewage sludge and lake sediments. Applied and Environmental Microbiology 50, 595601. Of these possibilities, blocking of the cycle by accumulation of NADH and a concomitant lack of NAD+ is considered to be the most important. Even if pH2 was reduced to minimal levels and regeneration of NAD+ from NADH was rapid and complete, only methanogens, sulphate reducers and a few carbohydrate fermenting rumen bacteria have electron transport chains which would enable them to derive ATP from NADH generated by the TCA cycle (Russell and Wallace 1997). An active TCA cycle would then not itself increase ATP yield to most primary fermenters, although they would obtain Hydrogen production and transfer in the rumen 43 Czerkawski, J.W. (1975). Some stoichiometric relations and their value in rumen studies. In: Tracer Studies on Nonprotein Nitrogen for Ruminants II, pp. 5563. IAEA, Vienna. Czerkawski, J.W. (1986). An Introduction to Rumen Studies. Pergamon Press, Oxford. Czerkawski, J.W., Harfoot, C.G. and Breckenridge, G. (1972). The relationship between methane production and concentrations of hydrogen in the aqueous and gaseous phases during rumen fermentation in vitro. Journal of Applied Bacteriology 35, 537551. Demeyer, D.I. and Van Nevel, C.J. (1975). Methanogenesis, an integrated part of carbohydrate fermentation, and its control. In: Digestion and Metabolism in the Ruminant, pp. 366382 (eds. I.W. McDonald and A.C.I. Warner). University of New England, Armidale NSW. Doelle, H.W. (1975). Bacterial Metabolism (2nd edition.). Academic Press, New York. Ellis, J.E., McIntyre, P.S., Saleh, M., Williams, A.G. and Lloyd, D. (1991). Influence of CO2 and low concentrations of O2 on fermentative metabolism of the ruminal ciliate Polyplastron multivesiculatum. Applied and Environmental Microbiology 57, 14001407. Gottschalk, G. (1986) Bacterial Metabolism (2 SpringerVerlag New York Inc. nd IUPAC (1981). Solubility Data Series Vol 5/6 (ed. C.L.Young). Pergamon Press, Sydney. Keltjens, J.T. and Vogels, G.D. (1996). Metabolic regulation in methanogenic Archaea during growth on hydrogen and CO2. Environmental Monitoring and Assessment 42, 1937. Lana, R.P., Russell, J.B. and Van Amburgh, M.E. (1998). The role of pH in regulating ruminal methane and ammonia production. Journal of Animal Science 76, 21902196. Leng, R.A. and Leonard, G.J. (1965). Loss of methyl tritium from [3H] acetate in rumen fluid. Nature 207, 760761. Mathison, G.W., Okine, E.K., McAllister, T.A., Dong, Y., Galbraith, J. and Dmytruk, O. (1998). Reducing methane emissions from ruminant animals. Journal of Applied Animal Research 14, 128. MarvinSikkema, F.D., Rees, E., Kraak, M.N., Gottschal, J.C. and Prins, R.A.. (1993). Influence of metronidazole, CO, CO2 ane methanogens on the fermentative metabolism of the anaerobic fungus Neocallimastix sp. Strain L2. Applied and Environmental Microbiology 59, 26782683. McArthur, J.M. and Miltimore, J.E. (1961). Rumen gas analysis by gassolid chromatography. Canadian Journal of Animal Science 41, 187196. McInerney, M.J., Mackie, R.I. and bryant, M.P. (1981). Syntrophic association of butyrate degrading bacterium and Methanosarcina enriched from bovine rumen fluid. Moss, A.R. (1993). Methane: Global Warming and Production by Animals (Page 23). Chalcombe Publications, Canterbury. Nelson, W.O., Brown, R.E. and Kingwill, R.G. (1960). Factors affecting ratios of CO2: CH4 in bovine rumen gas. Journal of Dairy Science 43, 16541655. Newbold,C.J., Lassalas, B. and Jouany, J.P. (1995). The importance of methanogens associated with ciliate protozoa in ruminal methane production in vitro. Letters in Applied Microbiology 21, 23034. Olson, T.M. (1940). Bloat in dairy cattle. Journal of Dairy Science 23, 343353. Owens, F.N., Secrist, D.S., Hill, W.J. and Gill, D.R. (1998). Acidosis in cattle: a review. Journal of Animal Science 76, 275286. Paul, R.G., Williams, A.G. and Butler, R.D. (1990). Hydrogenosomes in the rumen entodiniomorphid ciliate Polyplastron multivesiculatum. Journal of General Microbiology. 136,198189. Pelchen, A. and Peters, K.J. (1998). Methane emissions from sheep. Small Ruminant Research 27, 13750. Rowe, J.B., Davies, A. and. Broome, A.W.J (1985). Quantitative effects of defaunation on rumen fermentation and digestion in sheep. British Journal of Nutrition 54, 10519. Russell, J.B. (1987). A proposed mechanism of monensin action in inhibiting ruminal bacterial growth: effects of edition). Harold, F.M., Pavlasova, E. and Baarda, J.R. (1970). A transmembrane pH gradient in Streptococcus faecalis: origin, and dissipation by proton conductors and N,Ndicyclohexylcarbodiimide. Biochimica et Biophysica Acta 196, 235244. Hegarty, R.S. (1999). Reducing rumen methane emissions through elimination of rumen protozoa. Australian Journal of Agricultural Research (in press). Hillman, K., Lloyd, D. and Williams, A.G. (1988). Interactions between the methanogen Methanosarcina barkeri and rumen holotrich ciliate protozoa. Letters in Applied Microbiology 7, 4953. Hillman, K., Williams, A.G. and Lloyd, D. (1991). The influence of headspace gas composition on the production of VFA by rumen ciliate protozoa. Letters in Applied Microbiology 13, 1115. Hino, T. and Russell, J.B. (1985). Effect of reducing equivalent disposal and NADH/NAD on deamination of amino acids by intact rumen microorganisms and their cell extracts. Applied and Environmental Microbiology 50, 13681374. Hungate, R.E. (1966). The Rumen and its Microbes. Academic Press, London. Hungate, R.E. (1970). Interrelationships in the rumen microbiota. In: Physiology of Digestion and Metabolism in the Ruminant, pp. 292 305 (ed. A.T. Phillipson). Oriel Press Ltd, England. Hungate, R.E., Smith, W., Bauchop, T., Yu, I. and Rabinowitz, J.C. (1970). Formate as an intermediate in the bovine rumen. Journal of Bacteriology 102, 389397. 44 Hegarty, R.S. and Gerdes, R. ion flux and protonmotive force. Journal of Animal Science 64, 15191525. Russell, J.B. (1991). Intracellular pH of acidtolerant ruminal bacteria. Applied and Environmental Microbiology 57, 33833384. Russell, J.B. (1998). The importance of pH in the regulation of ruminal acetate to propionate ratio and methane production in vitro. Journal of Dairy Science 81, 322230. Russell, J.B. and Strobel, H.J. (1993). Microbial energetics. In: Quantitative Aspects of Ruminant Digestion and Metabolism, pp. 165186 (eds. J.M. Forbes and J. France). CAB International, Wallingford. Russell, J.B. and Wilson, D.B. (1996). Why are cellulolytic bacteria unable to digest cellulose at low pH? Journal of Dairy Science 79, 15031509. Russell, J. B. and Wallace, R.J. (1997). Energy yielding and energyconsuming reactions. In: The Rumen Microbial Ecosystem, pp. 246282 (eds. P.N. Hobson and C.S. Stewart). Blackie, Oxford. Sauer, F.D. and reduction by rumen inhibited. Teather, R.M. (1987). Changes in oxidation potentials and volatile fatty acid production bacteria when methane synthesis is Journal of Dairy Science 70, 183540. Thauer, R.K., Jungerman, K. and Decker, K. (1977). Energy conservation in chemotrophic an