Enrichment and Isolation of Rumen Bacteria That Reduce trans-Aconitic Acid to Tricarballylic Acid
JAMES B. RUSSELL
Agricultural Research Service, U.S. Department of Agriculture, and Department ofAnimal Science, Cornell University,Ithaca, New York 14853
Received 2 August 1984/Accepted 8 October 1984
Bacteria from the bovine rumen capable of reducing trans-aconitate to tricarballylate were enriched in an
anaerobic chemostat containing rumen fluid medium and aconitate. After 9 days at a dilution rate of 0.07 h-',the medium was diluted and plated in an anaerobic glove box. Three types of isolates were obtained from theplates (a crescent-shaped organism, a pleomorphic rod, and a spiral-shaped organism), and all three producedtricarballylate in batch cultures that contained glucose and trans-aconitate. In glucose-limited chemostats (0.10h-1), trans-aconitate reduction was associated with a decrease in the amount of reduced products formed fromglucose. The crescent-shaped organism produced less propionate, the pleomorphic rod produced less ethanol,and the spiral made less succinate and possibly H2. Aconitate reduction by the pleomorphic rod and the spiralorganism was associated with a significant increase in cellular dry matter. Experiments with stock cultures ofpredominant rumen bacteria indicated that Selenomonas ruminantium, a species taxonomically similar to thecrescent-shaped isolate, was an active reducer of trans-aconitate. Strains of Bacteroides ruminicola, Butyrivibriofibrisolvens, and Megasphaera elsdenii produced little if any tricarballylate. Wolinella succinogenes producedsome tricarballylate. Based on its stability constant for magnesium (Keq = 115), tricarballylate could be a factorin the hypomagnesemia that leads to grass tetany.
When plants are grown under cloudy conditions, with highlevels of nitrogen fertilization, organic acids and, in partic-ular, trans-aconitic acid can accumulate (9, 12, 19). Trans-aconitic acid, in severe conditions, can account for morethan 4% of the total dry matter (23), and high levels of thisacid were associated with toxic responses in ruminant ani-mals (2, 3, 6, 11, 12). Stout et al. suggested that trans-aconi-tic acid could form chelates with magnesium and decreasethe availability of dietary magnesium (23). Bohman et al.subsequently showed that oral administration of trans-aconi-tic acid could induce symptoms of hypomagnesemia, com-monly termed grass tetany (3).Recent experiments indicated that trans-aconitic acid was
rapidly fermented by mixed rumen bacteria in vitro, and itseemed unlikely that trans-aconitic acid would be present inrumen fluid for a long enough time to decrease the availabil-ity of dietary magnesium (22). Trans-aconitic acid fermenta-tion, however, led to an accumulation of acetic acid and anunknown compound that was subsequently identified astricarballylic acid (Fig. 1 [22]). Tricarballylic acid was fer-mented very slowly by mixed rumen bacteria. Based on itsthree exposed carboxyl groups, it appeared that tricarbal-lylic acid might be an important factor in the hypomagnes-emia that leads to grass tetany.The following experiments describe the enrichment, iso-
lation, and characteristics of trans-aconitic acid-reducingrumen bacteria. Several common strains of rumen bacteriawere also tested for the capacity to reduce trans-aconiticacid to tricarballylic acid.
MATERIALS AND METHODS
Enrichment and isolation. An anaerobic chemostat (360-mlculture vessel, O2-free CO2 gas phase, 0.07 h-1 dilution rate)was inoculated with mixed rumen bacteria from a cow fedtimothy hay. The medium reservoir contained (in milligramsper liter): K2HPO4, 292; KH2PO4, 292; (NH4)2SO4, 480;NaCl, 480; MgSO4 - 7H20, 100; CaC12 * 2H20, 64; NaCO3,
1,000; cysteine hydrochloride, 600; hemin, 1; pyridoridor-amine dihydrochloride, 2; riboflavin, 2; thiamine hydrochlo-ride, 2; nicotinamide, 2; calcium pantothenate, 2; lipoic acid,1; para-aminobenzoic acid, 0.1; folic acid, 0.05; biotin, 0.05;coenzyme B12, 0.05; valeric acid, 100; isovaleric acid, 100;isobutyric acid, 100; 2-methylbutyric acid, 100; trans-aconi-tate, 13.5 mmol; and clarified rumen fluid, 10% (vol/vol) (seereferences 20 and 21 for method of preparation). The transisomer of aconitic acid was obtained from Sigma ChemicalCo., St. Louis, Mo., and hereafter is referred to simply asaconitic acid.Each day samples were taken from the culture vessel, and
optical density, pH, and organic acids were analyzed (meth-ods described below). On day 9, the culture was seriallydiluted (10-fold increments) with sterile medium (same aschemostat medium) that contained 2% molten agar (47°C)and 0.2% carbohydrates (equal parts glucose, maltose, andcellobiose) or no added carbohydrates. After 48 h of incu-bation (39°C) in an anaerobic glove box (Coy LaboratoryProducts, Ann Arbor, Mich.), the plates were examined forgrowth. When carbohydrates were provided, colonies werepresent up to the 108 dilution. Plates not containing carbo-hydrates also had colonies, but these colonies were muchsmaller. Fifteen isolated colonies from each set of plateswere picked and inoculated into medium plus 0.2% glucose.All 30 isolates grew on glucose. Broth cultures were replat-ed, picked, and examined microscopically (x1,250) for pu-rity. Three cell types were observed (pleomorphic rods,crescent-shaped cells, and spiral organisms). Each of thestrains was then inoculated in medium with different carbo-hydrates, lactate, and mannitol (Table 1). All of the strainswithin a cell type exhibited the same pattern of growth, andone strain from each cell type was retained.
Fermentation. The pleomorphic rod (strain D), the cres-cent-shaped cells (strain A), and the spiral organism (strainM) were then grown in rumen fluid medium (see above)which was supplemented with glucose (2 g/liter). Strains D
120
trans-ACONITIC ACID REDUCTION 121
ACONITIC ACID
H CC- H
11H OO C C H2 C OO H
[2 H]
H H COOO OH
H O O c/\ CH2C O O HH
TRICARBALLYLIC ACID
FIG. 1. Structure and likely pathway of aconitic acid conversionto tricarballylic acid (22).
and M grew slowly in the medium and were provided with anadditional 4.0 g of Trypticase (BBL Microbiology Systems,Cockeysville, Md.) and 4.0 g of yeast extract per liter. Afterthese additions, all three strains grew well in the medium.The strains were also grown in continuous culture (seeabove). In this case, the medium reservoir contained rumenfluid medium (see above), 4.0 g of Trypticase per liter, and4.0 g of yeast extract per liter. Each strain was grown withand without aconitic acid (approximately 7 mM) to ascertainthe effect of this acid on fermentation products.
Other rumen bacteria. A variety of stock cultures were
likewise tested for the ability to reduce aconitate to tricarbal-lyate. Bacteroides ruminicola B14, Butyrivibrio fibrisolvensA38 and 49, Megasphaera elsdenii T81 and B159, andSelenomonas ruminantium HD4 were obtained from M. P.Bryant, University of Illinois, Urbana. B. ruminicola 23 andS. ruminantium D were provided by K. A. Dawson, UIniver-sity of Kentucky, Lexington. Streptococcus bovis 45S1 was
obtained from C. S. Stewart, Rowett Research Institute,Scotland, and T. L. Miller, New York State Department ofHealth, Albany, provided Wolinella succinogenes. All strainswere grown in basal medium that contained aconitate (10mM), Trypticase (4.0 g/liter), and yeast extract (4.0 g/liter).
Analyses. Optical density was measured with a Gilfordmodel 260 spectrophotometer (600 nm and cuvettes of 1-cmlight path). Volatile fatty acids, glucose, lactate, succinate,aconitate, and tricarballylate were analyzed by high-pres-sure liquid chromatography (Beckman model 334 liquidchromatograph, model 156 refractive index detector, model421 CRT data controller, CRIA integrator, 50-,ul loop, 50°C),using a Bio-Rad HPX-87H organic acid column (see refer-ences 10 and 22). Ethanol was determined by an enzymaticprocedure (4), and ammonia was assayed by the method ofChaney and Marbach (8). Cysteine interference with ammo-nia color formation was minimized by using six times as
much reagent. All analyses were performed in duplicate, andthe variation between replicates was always <7%.
RESULTSWhen mixed rumen bacteria were inoculated into a che-
mostat that contained rumen fluid medium and aconitate,there was a marked decrease in optical density between days0 and 4 (Fig. 2). Soon after inoculation virtually all of theavailable aconitate was degraded and acetate and propionatewere the primary fermentation products. Between days 2and 5, aconitate utilization declined and there was an in-crease in tricarballylate. From day 5 to 9, optical density andthe rate of aconitate utilization once again increased. Tri-carballylate also increased during this time period and therewas little increase in acetate and propionate until day 8. pHwas nearly neutral throughout the incubation, and recovery
of carbon from aconitic acid as acetic, propionic, or tricarbal-lylic acids ranged from 53 to 86% on days 1 to 4 and from 87to 102% on days 5 to 9.On day 9 of continuous culture, samples were withdrawn
from the culture vessel, diluted, and plated on rumen fluidmedia that contained either aconitate or aconitate and car-bohydrates. After 48 h of incubation, both plate types hadcolonies at the 108 dilution, but colonies from plates withonly aconitate were much smaller. Isolated colonies were
transferred to rumen fluid media that contained glucose, andgrowth was observed in all cases. Isolates were then exam-ined microscopically and screened for energy source utiliza-tion and fermentation products (Table 1). Three types of
TABLE 1. Characteristics of aconitic acid-reducing isolates
a C, Crescent shaped; R, rods; S, spiral organisms.b +, High turbidity; -, no turbidity; +, small amount of turbidity after 24 or
48 h of growth with 2.0 g of the energy source per liter.c Fermentation products formed in batch culture with 2.0 of glucose per
liter.d Tricarballyate was only a product when aconitic acid was added to the
medium.
VOL. 49, 1985
APPL. ENVIRON. MICROBIOL.
cnzw
-i
1-
0.0
z0
N-J
-
0(r
w
i -
() E 12,<-
8.0
pH Ip7.0
a
OD
-6.0
16
12
£E 8
241
p.---., ~ ~ 0--4
1 se
I "/,Is ,
A
~~-E~~~--~~~-14O ~~P
0 2 4 6 8D A Y S
. 010
FIG. 2. Growth of mixed rumen bacteria in continuous culture(dilution rate, 0.07 h-') with aconitate as the only added carbonsource. Symbols: A, optical density; *, pH; *, aconitate; utiliza-tion; A, tricarballylate; 0, acetate; 0, propionate.
organisms were identified and duplicate strains were dis-carded.The crescent-shaped organism (Fig. 3a) was able to fer-
ment a variety of hexoses, pentoses, disaccharides, starch,and mannitol (Table 1). It was unable to grow on cellulose,lactate, or malate. When grown on glucose in batch culture,it was motile and produced lactate, acetate, and propionate.The pleomorphic rod (Fig. 3b) was also able to grow onglucose, fructose, arabinose, sucrose, and cellobiose (Table1). When mannose, xylose, lactose, and starch were pro-vided, final turbidity was low (<0.3 optical density unit)even after 48 h of incubation. This motile rod producedformate, acetate, and ethanol in batch culture. The spiral-shaped organism (Fig. 3c) was only able to grow on glucose,fructose, mannose,, and mannitol and produced succinateand acetate. All three organisms produced tricarballylatewhen aconitate was added to glucose rumen fluid medium.Tricarballylate was not formed in the absence of aconitate.When the crescent-shaped organism (strain A) was grown
in batch culture with glucose and aconitate, glucose wasrapidly fermented and lactate was the primary product (Fig.4a). Aconitate was also degraded and this degradation wasassociated with a nearly equal molar increase in tricarbal-lylate. After glucose depletion at 5 h, the rate of aconitatedisappearance declined, and a significant amount of aconi-tate was left in the medium even after 50 h.The pleomorphic rod (strain D) also grew rapidly on
glucose, and aconitate was completely metabolized in lessthan 10 h lFig. 4b). Once again, aconitate utilization wasaccompanied by an equal molar increase in tricarballylate.Acetate and formate were also formed during the period ofglucose fermentation, and the molar ratio of these two acidswas nearly 1:1.
Strain M (the spiral organism) grew at a slower rate, andglucose fermentation closely paralleled the degradation ofaconitate (Fig. 4c). At approximately 12 h, both glucose andaconitate were depleted, and after this time there was littleincrease in acetate, succinate, or tricarhallylate. Comparisonof aconitate utilization and tricarballylate formation alsoindicated a one-to-one relationship.
FIG. 3. Phase-contrast micrographs of isolates that reduced aconitate to tricarballylate (bar, 10 ,um). (a) Crescent-shaped organism (strainA); (b) pleomorphic rods (strain D); (c) spiral-shaped organisms (strain M).
122 RUSSELL
trans-ACONITIC ACID REDUCTION 123
14 a 20 b
E F
12A
a~~~~~~~~~~~~~~~~~o 16
w 14ui ~~~~~~~~~~~~~E
co Co.P
0 6 c_j D--------_s__--------jj T
21t~~~~~~~~------T < 10 1
0 10 20 30 40 50 0
TIME (h) <f14 A
E A0S 10O R 20 30 40 50
_ f' ~~~~~~~~~TI ME (h )12 c
a-7,ff ,~~~~~~~~~___~~FIG. 4. Growth of strain A (crescent-shaped organism) (a), strains\MC_ -____4-' ~~~~-- T D (pleomorphic rod) (b), and strain M (spiral organism) (c) in
cJ\ ~~~~~~~~~~~~~batchculture with glucose and aconitate. Symbols: *, glucose; 0,<
10 S ~~~~~~~~~~~~~aconitate; A, tricarballylate; 0, acetate; 0, lactate; O, formate;§lb i so 40, succinate.
4wI/ ~~~~~~~~~Tosee what effect aconitate reduction to tricarballylate/8\,8d/ ~~~~~~~~washaving on fermentation patterns, each of the three
o;1l ~~~~~~~~~~~isolateswas grown in continuous culture, with and withoutolllll ~~~~~~~~~~aconitate(Table 2). When strain A was grown in continuous= t | ~~~~~~~~~~~~culturewith glucose and no aconitate, propionate and ace-
>J 6 tate were the primary products, and less lactate was ob-0 served. Addition of aconitate caused a marked decline in[ ~~~~~~~~~~~propionateand an increase in lactate, acetate, and tricarbal-
41 lylate. In both cases, the oxidation-reduction state (OIR) of]/ ~~~~~~~~~~thesubstrates and that of the products were roughly equal.t> ~~~~~~~~~~Carbonrecovery ranged from 95 to 103%b. Strain A was not,8/\ ~~~~~~~~~~greatlystimulated by the addition of yeast extract or Trypti-24/Jt ~~~~~~~~~case(see above), and it seems likely that glucose was the,Zza ~~~~~~~~~~primarysource of cell carbon.>WG ~~~~~~~~~~Whenstrain D was grown in continuous culture, formate,gu ~~~~~~~~~~~acetate,and ethanol were formed from glucose (Table 2).
OS _ a , , _ _ ~~~~~Aconitate addition caused an equal molar increase in tri-
0 10 20 30 40 50 carballylate and a reduction in ethanol formation. OIRT M,.E (h- , balances were consistent but carbon recoveries were >100%E.
VOL. 49, 1985
APPL. ENVIRON. MICROBIOL.
TABLE 2. Growth of aconitate-reducing isolates in continuous culture with and without aconitate'
Dilution rate of 0.10 h-'.b Assumes that bacterial cells are 50% carbon.c Theoretical amount of CO, needed to form succinate.dH = -½/ and 0 = +1.e Theoretical amount of CO, produced from acetate formation if formate was not a product.
Strain D was greatly stimulated by added yeast extract andTrypticase, and these values indicated that much of the cellcarbon was derived from these sources and not from glu-cose.
Strain M fermented glucose to acetate and succinate(Table 2). Aconitate reduction to tricarballylate was associ-ated with a small decline in succinate and an increase inacetate. O/R calculations indicated that the produicts weremore oxidized than the substrates, and this would indicatethat a highly reduced product (i.e., H2) was not detected. Tomaintain a strictly anaerobic environment in the chemostat,the culture vessel was continuously purged with 02-freeCO2. This continuous gas flow confounded gas measure-ments, and H2 production was not determined. Aconitatereduction may have been associated with a decrease in H2production. Trace amounts of H2 were detected in batchculture. The organism was greatly stimulated by yeastextract and Trypticase, and carbon recoveries indicated thatmuch of the cell carbon could have been derived from thesesources. Variations in H2 production would not have aneffect on carbon recovery.None of the isolates was able to grow on aconitate alone,
and it is probable that the presence of bacteria in the
TABLE 3. Utilization of aconitic acid by other strains of rumenbacteria'
Aconitate TricarballylateOrganism used (mM) produced (mM)
Selenomonas ruminantium HD4 9.5 9.0S. ruminantium D 4.7 4.6
a Glucose was provided at 11.11 mM and aconitate was provided at 10 mM.The incubation time was 24 h.
b W. succinogenes was grown with 65 mM formate and 26 mM fumarate.
enrichment chemostat (Fig. 2) was at least partially depend-ent on trace amounts of energy source in sterile rumen fluid.When aconitate was added to glucose-limited chemostats,however, there was an increase in cell dry matter (Table 2).These increases were most significant with strains D and Mand indicated that aconitate reduction was having a positiveinfluence on the energetics of growth.To ascertain whether other rumen bacteria were able to
reduce aconitate to tricarballylate, various strains of rumenbacteria were grown in batch culture with glucose andaconitate (Table 3). S. ruminantium HD4 was able to de-grade 9.5 mM aconitate, and 9.0 mM tricarballylate wasformed. S. ruminantium D used less aconitate but theutilization was once again accompanied by an equal molarincrease in tricarballylate. Strains of B. ruminicola, Butyri-vibrio fibrisolvens, M. elsdenii, and Streptococcus bovisused little aconitate and produced trace or insignificantamounts of tricarballylate. W. succinogenes converted 2 mMaconitate to tricarballylate, which was not nearly as much asS. ruminantium.
DISCUSSIONBatch cultures of mixed rumen bacteria from cows fed
either timothy hay or 60% concentrate were found to con-vert approximately 40% of the added aconitate to tricarbal-lylate (22). When a chemostat containing rumen fluid me-dium and aconitate was inoculated with mixed rumen bac-teria, all aconitate entering the culture vessel disappeared byday 1, but tricarballylate was not a major product (Fig. 2).As the medium entry continued, however, tricarballylateproduction increased, accounting for 45% of the aconitateutilization on day 8.
Isolates from day 9 of continuous culture all producedtricarballylate in batch culture when aconitate was added tothe medium (Table 1), and the ratio of aconitateutilized/tricarballylate produced was approximately 1 (Fig.4). Acetate was also a significant product in continuousculture, but none of the isolates produced significant amountsof acetate from aconitate (Table 2). The absence of organ-isms that fermented aconitate to acetate may have beenrelated to the 5% H2 content of the anaerobic glove box. Apossible pathway of acetate conversion to aconitate mightproceed via citrate, oxaloacetate, pyruvate, and acetyl co-enzyme A. In this case, H2 could be a product, and theinhibition of H2-producing species by high partial pressuresof H2 is well documented (25).
124 RUSSELL
trans-ACONITIC ACID REDUCTION 125
Based on motility, crescent-shaped morphology, Gramstain, energy source utilization, and fermentation products(Table 1), it appeared that strain A closely fit the taxonomyof the rumen bacterium S. ruminantium (13, 15). The pleo-morphic rod (strain D) resembled a bacteroides but absenceof either succinate or propionate indicated that it was not anormal rumen type (i.e., B. ruminicola, B. succinogenes, orB. amylophilus). Based on absence of either succinate orpropionate, abundant gas formation, and motility, strain Dresembled Clostridum clostridiiforme, a species previouslydesignated as Bacteroides clostriiformis subsp. girans (15).Taxonomic classification of the spiral organism, strain M, isnot certain. The spiral-shaped morphology indicated that itis most likely a spirochete, but other tests are needed beforea more precise identification can be made.Many rumen bacteria produce lactate, ethanol, and H2 in
pure culture, but these products are usually not detected inrumen fluid. An explanation of the apparent discrepancybetween pure- and mixed-culture studies was first proposedby Hungate (15). He stated that the removal of hydrogen bymethanogens would allow rumen bacteria to reoxidize re-duced nucleotides and produce more acetate. Since thistime, Wolin and his colleagues (25) have demonstrated thatinterspecies hydrogen transfer to methanogens can decreasethe formation of reduced end products (i.e., lactate, propi-onate, succinate, ethanol, and H2).
Aconitate reduction to tricarballylate in chemostat culturewas associated with a change in fermentation products(Table 2). With strain A, there was a decrease in propionateand an increase in acetate. Acetate production was notaffected in strain D, but there was a decline in ethanol. Onlysmall changes in the fermentation pattern were noted withcultures of strain M in the presence of aconitate. Differencesin O/R indicated that H2 may have been a product, andaconitate reduction could have been at the expense of H2production. These changes in fermentation pathways in-creased the efficiency of glucose utilization and cell growth.Because tricarballylate formation represents an alternativemeans of reducing-equivalent disposal, there could be anantagonism between tricarballylate and methane productionin vivo.
Formation of tricarballylate from aconitate occurs by asimple reduction (Fig. 1), and similar reactions are known tooccur in the rumen (1, 14, 16, 18, 24, 26). When rumenbacteria capable of reducing unsaturated bonds were testedfor their ability to reduce aconitate, only S. ruminantiumwas highly active (Table 3). W. succinogenes was also activebut low numbers in rumen contents (26) suggest that thisorganism would be less important.The importance of tricarballylate formation in the bovine
rumen is related to its slow rate of metabolism by mixedrumen bacteria and its potential as a magnesium chelator(22). Literature values indicate that the stability constant oftricarballylate for magnesium is 115 (17). Based on thisbinding capacity, it is conceivable that tricarballylate couldcomplex a significant portion of dietary magnesium, espe-cially when the magnesium status of the animal is low.Toxicity is obviously related to the relative rates of aconitateconversion to tricarballylate versus acetate. Experimentsare currently being conducted to isolate and characterizerumen bacteria that produce acetate from aconitate.
ACKNOWLEDGMENTS
This research was supported by the U.S. Dairy Forage ResearchCenter, Madison, Wis.
Thanks are extended to W. C. Ghiorse for developing the phase-contrast micrographs.
LITERATURE CITED1. Baldwin, R. L., W. A. Wood, and R. S. Emery. 1965. Lactate
metabolism by Peptostreptococcus elsdenii: evidence for lactylcoenzyme A dehydrase. Biochim. Biophy. Acta 97:202-213.
2. Bohman, V. R., F. P. Horn, B. A. Steward, A. C. Mathers, andD. L. Grunes. 1983. Wheat pasture poisoning. I. An evaluationof cereal pastures as related to tetany in beef cows. J. Anim.Sci. 57:1352-1363.
3. Bohman, V. R., A. L. Lesperance, G. D. Harding, and D. L.Grunes. 1969. Induction of experimental tetany in cattle. J.Anim. Sci. 29:99-102.
4. Bonnichsen, R. 1963. Ethanol: determination with alcohol dehy-drogenase and DPN, p. 285-289. In H. U. Bergmeyer (ed.),Methods of enzymatic analysis. Academic Press, Inc., NewYork.
5. Bryant, M. P. 1974. Genus Selenomonas, p. 424-426. In R. E.Buchanan and N. E. Gibbons (ed.), Bergey's manual of deter-minative biology, 8th ed. The Williams & Wilkins Co., Balti-more.
6. Burau, R. G., and P. R. Stout. 1965. Trans-aconitic acid in rangegrasses in early spring. Science 150:766-767.
7. Canale-Parola, E. 1974. The spirochetes, p. 38-70. In N. R.Krieg, and J. R. Holt (ed.), Bergey's manual of systemicbacteriology, 8th ed. The Williams & Wilkins Co., Baltimore.
8. Chaney, A. L., and E. P. Marbach. 1962. Modified reagents fordetermination of urea and ammonia. Clin. Chem. 8:130-132.
9. Dijkshoorn, W. 1973. Organic acids, and their role in ion uptake,p. 163-188. In G. W. Butler and R. W. Bailey (ed.), Chemistryand biochemistry of herbage. Academic Press, London.
10. Erlich, G. G., D. F. Goerlitz, J. H. Bourell, G. V. Eisen, andE. M. Godsy. 1981. Liquid chromatographic procedure forfermentation product analysis in the identification of anaerobicbacteria. Appl. Environ. Microbiol. 42:878-885.
11. Grunes, D. L. 1967. Grass tetany of cattle as affected by plantcomposition and organic acids, p. 105-110. In Cornell NutritionConference for Feed Manufacturers, Buffalo, N.Y. CornellAgricultural Experiment Station, Ithaca, N.Y.
12. Grunes, D. L., P. R. Stout, and J. R. Brownell. 1970. Grasstetany of ruminants. Adv. Agron. 22:331-374.
13. Holdeman, L. V., R. W. Kelly, and W. E. C. Moore. 1984.Anaerobic Gram negative straight, curved and helical rods, p.605-662. In N. R. Krieg and J. G. Holt (ed.), Bergey's manualof systemic bacteriology. The Williams & Wilkins Co., Balti-more.
14. Hughes, P. E., and S. B. Tove. 1980. Identification of anendogenous electron donor for biohydrogenation as alpha-to-copherolquinol. J. Biol. Chem. 255:4447-4452.
15. Hungate, R. E. 1966. The rumen and its microbes, p. 245-280.Academic Press, Inc., New York.
16. Kroger, A., and E. Winkler. 1981. Phosphorylative fumaratereduction in Vibrio succinogenes: stoichiometry of ATP synthe-sis. Arch. Microbiol. 129:100-104.
17. Martell, A. E., and R. M. Smith (ed.). 1977. Critical stabilityconstants, p. 158, vol. 3. Plenum Press, New York.
18. Paynter, M. J. B., and S. R. Elsden. 1970. Mechanism ofpropionate formation by Selenomonas ruminantium, a rumenmicroorganism. J. Gen. Microbiol. 61:1-7.
19. Prior, R. L., D. L. Grunes, R. P. Patterson, F. W. Smith, H. F.Mayland, and W. J. Visek. 1973. Partition column chromatog-raphy for quantitating effects of fertilization on plant acids. J.Agric. Food Chem. 21:73-77.
20. Russell, J. B., and R. L. Baldwin. 1978. Substrate preferences inrumen bacteria: evidence of catabolite regulatory mechanisms.Appl. Environ. Microbiol. 36:319-329.
21. Russell, J. B., and R. L. Baldwin. 1979. Comparison of substrateaffinities among several rumen bacteria: a possible determinantof rumen bacterial competition. Appl. Environ. Microbiol.37:531-536.
22. Russell, J. B., and P. J. Van Soest. 1984. In vitro ruminal
VOL. 49, 1985
126 RUSSELL APPL. ENVIRON. MICROBIOL.
fermentation of organic acids common in forage. Appl. Environ.Microbiol. 47:155-159.
23. Stout, P. R., J. R. Brownell, and R. G. Burau. 1967. Occur-rences of trans-aconitate in range forage species. Agron. J.59:21-24.
24. Wallnofer, P., and R. L. Baldwin. 1967. Pathway of propionateformation in Bacteroides ruminicola. J. Bacteriol. 93:504-505.
25. Wolin, M. J. 1975. Interactions between bacterial species of therumen, p. 134-148. In I. W. McDonald and A. C. I. Warner(ed.), Digestion and metabolism in the ruminant. AcademicPress, Inc., New York.
26. Wolin, M. J., E. A. Wolin, and N. J. Jacobs. 1961. Cytochrome-producing anaerobic vibrio, Vibrio succinogenes, sp. n. J.Bacteriol. 81:911-917.
