JOURNAL OF BACrROLOGY, June 1973, p. 1231-1240Copyright 0 1973 American Society for Microbiology

Vol. 114, No. 3Printed in U.SA.

Glucose Fermentation Products of Ruminococcusalbus Grown in Continuous Culture with Vibriosuccinogenes: Changes Caused by Interspecies

Transfer of H2E. L. IANNOTTI,I D. KAFKEWITZ,2 M. J. WOLIN, AND M. P. BRYANT

Department of Dairy Science and Department of Microbiology, University of Illinois, Urbana, Illinois 61801

Received for publication 29 January 1973

The influence of a H2-utilizing organism, Vibrio succinogenes, on the fermen-tation of limiting amounts of glucose by a carbohydrate-fermenting, H2-produc-ing organism, Ruminococcus albus, was studied in continuous cultures. Growthof V. succinogenes depended on the production of H2 from glucose by R. albus. V.succinogenes used the H2 produced by R. albus to obtain energy for growth byreducing fumarate in the medium. Fumarate was not metabolized by R. albusalone. The only products detected in continuous cultures of R. albus alone were

acetate, ethanol, and H2. CO2 was not measured. The only products detected inthe mixed cultures were acetate and succinate. No free H2 was produced. Noformate or any other volatile fatty acid, no succinate or other dicarboxylic acids,lactate, alcohols other than ethanol, pyruvate, or other keto-acids, acetoin, or

diacetyl were detected in cultures of R. albus alone or in mixed cultures. Themoles of product per 100 mol of glucose fermented were approximately 69 forethanol, 74 for acetate, 237 for H2 for R. albus alone and 147 for acetate and 384for succinate for the mixed culture. Each mole of succinate is equivalent to theproduction of 1 mol of H2 by R. albus. Thus, in the mixed cultures, ethanolproduction by R. albus is eliminated with a corresponding increase in acetate andH2 formation. The mixed-culture pattern is consistent with the hypothesis thatnicotinamide adenine dinucleotide (reduced form), formed during glycolysis byR. albus, is reoxidized during ethanol formation when R. albus is grown alone andis reoxidized by conversion to nicotinamide adenine dinucleotide and H2 when R.albus is grown with V. succinogenes. The ecological significance of thisinterspecies transfer of H2 gas and the theoretical basis for its causing changes infermentation patterns of R. albus are discussed.

Fermentation products, particularly ethanoland lactate, of pure cultures of a number ofimportant rumen bacteria are not usuallyformed nor are they usually significant inter-mediates in the rumen fermentation (7). Themajor products of the fermentation of ingestedfood, mainly plant carbohydrates, by the mixedrumen microbial population are acetic, propi-onic, and butyric acids, and the gases CH, andCO2 (7). Hungate (7) suggested that, in therumen, the electrons derived from the oxidationof fermentation substrates by carbohydrate-fer-menting, ethanol- or lactate-producing rumen

'Present address: New York Ocean Science Laboratory,Montauk, N.Y. 11954.

' Present address: Department of Zoology and Physiology,Rutgers University, Newark, N.J. 07102.

microorganisms are somehow shunted awayfrom the reduction of the fermentation interme-diate, pyruvate, or electron acceptors derivedfrom pyruvate. He proposed that these electronsare available for use by H2-utilizing, methane-forming bacteria in the ecosystem. In pureculture, however, these same carbohydrate-fer-menting organisms have to supply their ownelectron sink in the form of a lactate- orethanol-forming system in order to satisfy theoxidation-reduction balance requirements ofany anaerobic fermentation.We began a series of experiments to attempt

to test Hungate's hypothesis. In brief, our goalwas to compare the fermentation products pro-duced by the rumen bacterium, Ruminococcusalbus, in pure culture with those produced by R.

1231

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

IANNOTTI ET AL.

albus in mixed culture with a H2-utilizingbacterium. R. albus is a cellulolytic bacteriumthat was shown to produce ethanol, acetate,formate, H2, and CO2 (some strains producesmall amounts of succinate and lactate) fromcellobiose (1, 9). We planned to mix R. albuswith Vibrio succinogenes, an organism that canobtain energy for growth by coupling the oxida-tion of H2 or formate with the reduction offumarate to succinate. If Hungate's hypothesisis correct, it would be expected that the onlyproducts of the mixed-culture fermentation ofcarbohydrate would be acetate, succinate (fromfumarate reduction by the vibrio), and CO2. Wedecided to use a continuous-culture system forthese experiments. Although there were severalreasons for this decision, a major reason was thebelief that we could synchronize the growth ofone species with respect to the other better incontinuous than in batch culture.

Prior to the mixed-culture experiment, it wasessential to accurately establish a fermentationbalance for R. albus in continuous culture.Hungate studied the growth of R. albus incontinuous culture, but he did not characterizethe fermentation products (6). Fermentationproduct analyses for R. albus were done inbatch cultures with cellobiose as the substrate(1, 9), but these were not complete balances. Inaddition, there is no guarantee that batch andcontinuous-culture fermentations will be thesame. An anaerobic continuous culture appara-tus was designed which permits the cultivationof nonsporeforming anaerobes such as R. albusand also permits gas accumulation and meas-urement for complete product analysis (8). Weused this apparatus for the determination of afermentation balance for R. albus in a glucose-limited continuous culture at various growthrates and for similar experiments with mixedcultures of R. albus and V. succinogenes. In thelatter case, H2 production by R. albus andfumarate in the medium supported the growthof V. succinogenes. V. succinogenes can obtainenergy for growth by reducing fumarate withH2, but it will not use glucose, ethanol, oracetate as energy sources in the presence orabsence of fumarate (15). We will show that R.albus ferments glucose to ethanol, acetate,hydrogen, probably CO2, and unidentified prod-ucts when glucose limits growth in a chemostat.No formate is produced. In mixed culture,all of the H2 produced by R. albus is used by V.succinogenes to reduce fumarate to succinate.In addition, no ethanol is produced in themixed-culture system, but acetate productionincreases relative to the monoculture fermenta-tion of R. albus. The increased acetate carbon

corresponds to the decrease in ethanol carbonrelative to the monoculture, and the electronsfound in ethanol in the monoculture are di-verted to succinate formation by the vibriopresumably by interspecies transfer of H2.

MATERIALS AND METHODSOrganism, media, and growth conditions. R.

albus strain 7 was used. It was grown in continuousculture in a medium of the following composition (perliter): KH2PO4, 0.48 g; K2HPO4, 0.48 g; (NH4)2SO4,0.48 g; NaCl, 0.96 g; Trypticase, 5.0 g; yeast extract,1.0 g; isobutyric acid, isovaleric acid, and DL-2-methylbutyric acid, 0.1 ml of each; cysteine hydro-chloride, 0.5 g; CaCl2.2H20, 0.13 g; MgSO4.7H20,0.2 g; Na2CO3, 4.0 g; sodium fumarate, 3.9 g; glucose,1.0 g, resazurin, 1.0 mg. The fumarate, cysteine,carbonate, fatty acids (neutralized with NaOH), glu-cose, magnesium sulfate, and calcium chloride wereeach made up as separate solutions, autoclaved,cooled, and individually added to the other ingredi-ents which were autoclaved as a single solution (13.5liters) in a 19-liter carboy. The total volume of themedium was 15 liters; the final pH was 6.7 to 6.8 aftersaturation with 100% CO2. Batch cultures for mainte-nance and inoculation of the continuous-culture ves-sel were grown in this same medium with or withoutfumarate.The continuous-culture apparatus and operating

procedures are described in detail in a separatepublication (8). In brief, the culture vessel held 345 mlof medium up to an overflow tube. The entire systemwas initially made anaerobic by flushing with 02-freeCO2. After inoculation, gas flow through the cultureand collecting vessels was discontinued, and thegrowth and collection portion of the apparatus weremaintained as a closed, anaerobic system. Mediumwas pumped into the culture vessel from the anaero-bic reservoir after the inoculum grew up as a batchculture. Produced gases collected within the cultureand collection vessels, and pressure build-up wasprevented by allowing gases to expand into a col-lapsed football bladder. The culture overflow wascollected in a flask, held at 0 C, that was emptied byapplying a positive CO2 pressure to the system afterapproximately 2,000 ml of culture was collected. Thecontinuous-culture system was never in a steady statewith respect to gas atmosphere composition becauseof the production of H2 by R. albus. After collection ofthe overflow, there was no H2 in the atmosphere. TheH2 concentration increased with time between empty-ings of the collection flask.

Continuous cultures were checked periodically forcontamination by microscopic examination and inoc-ulation into trypticase soy broth (Baltimore Biologi-cal Laboratories) with incubation in air withoutshaking. R. albus does not grow in trypticase soybroth aerobically. A positive check for R. albus wascarried out by inoculation into the continuous-culturemedium with cellulose or xylan substituted for glu-cose and verifying the ability of the organism to growon these energy sources.

Fermentation analyses. After emptying the col-

1232 J. BACTERIOL.

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

R. ALBUS FERMENTATION PRODUCTS

lection flask contents, cells were immediately re-moved by centrifugation at 13,200 x g for 30 min at 0C in a Servall RC2B centrifuge. The supernatantsolutions were frozen and kept at - 20 C untilanalyzed.

Glucose was determined by the colorimetric glucoseoxidase procedure of the Sigma Chemical Co., exceptthat the reagent was dissolved in 0.1 M potassiumphosphate buffer, pH 7.0.

Organic acids were initially determined by themethod of Ramsey (10). After finding that the onlyorganic acid produced was acetic acid, the methodwas modified for a rapid batch separation of acetic,formic, fumaric, and succinic acids. Formic acid wasof concern because of its reported production in batchcultures (1, 9), and fumaric and succinic acids were ofinterest because the reduction of fumaric to succinicacid by V. succinogenes was used to remove H, fromthe system. All solvents were saturated with 0.5 NH,SO4. A 10-g amount of silicic acid was mixed with0.5 NH,SO4 until a free-flowing powder was formed.This was slurried with chloroform and poured into a12-mm glass chromatographic column plugged withglass wool. The silicic acid was packed under 2 to 3psi of N2. A 1-ml amount of the sample was pipettedinto a 30-ml beaker and mixed with a few drops of 50%(v/v) H.S04 and 1.6 g of silicic acid until a free-flowingpowder was formed. This was poured on top of thesilicic acid column through a small layer of chloro-form. Finally, a piece of glass wool was added. Thebranched, volatile fatty acids added to the continuousculture medium were eluted with 80 ml of chloroform,acetic acid with 80 ml of 2% t-butanol in chloroform,followed by 140 ml of the same solvent to removeformic acid, fumaric acid with 125 ml of 6% t-butanolin chloroform, and succinic acid with 175 ml of10% t-butanol in chloroform. The eluted organicacids were quantitated by titration with 0.01 Nethanolic potassium hydroxide while bubbling astream of nitrogen through the sample. Solvent blankvalues were subtracted from the sample values.

Volatile fatty acids (except formate) were alsodetermined by gas chromatography. A 2-ml amount ofculture supernatant fluid was acidified with 0.4 ml of2 N HCl. This solution was directly analyzed on a 6' x1/8" stainless-steel column (6 ft by '/8 inch [approx. 1.8m by 0.32 cm]) containing 60% FFAP (5) on 80- to100-mesh Porapack T. The instrument was a Beck-man GC 5 gas chromatograph (Beckman Instru-ments, Inc.) with a column temperature of 190 C, ahelium carrier gas flow rate of 120 ml/min, and ahydrogen flame detector. Volatile fatty acids wereidentified and quantified by comparison of retentiontimes and peak heights with those of known stand-ards.

Alcohols were determined by gas chromatographyof the acidified solutions used for gas chromato-graphic analysis of volatile fatty acids. The proce-dures used were identical to those used for gaschromatographic analysis of volatile fatty acids ex-cept for a reduction of the column T to 120 C.Gas analysis was by gas chromatography. De-

tails of gas sampling, volume measurements, and gaschromatographic procedures are given in a separate

publication (8). It was necessary to sample fromdifferent sites in the culture-collection system toobtain H,-production values because of poor mixing ofH, in the gas atmosphere (8).

Qualitative analysis for acetoin or diacetyl (orboth) was by the Voges-Proskauer reaction. A 0.6-mlamount of 5% a-naphthol (in 40% KOH) was added to1 ml of culture supernatant fluid, and the mixture wasincubated at 25 C for 1 h with intermittent vigorousshaking.

Incorporation of [U-_4C]glucose into cells. [U-"4C]glucose (190 mCi/mmol) was injected into theculture vessel as described in the previous publication(8). A rubber tube between the overflow side arm andthe collection flask was clamped to collect a smallvolume of culture overflow. After removing a samplewith a hypodermic needle and syringe for analysis at aparticular time, the clamp was removed to clear theoverflow tube. The clamp was replaced, and theprocess was repeated for the next sampling period.The sample was immediately cooled in a mixture ofice and salt, and 1 ml was centrifuged at 27,000 x g at0 C for 15 min in a Sorvall RC2B centrifuge. Portionsof the sample before centrifugation and the superna-tant fluid after centrifugation were transferred toscintillation fluid which contained (per liter of diox-ane): naphthalene, 100 g; Cab-O-Sil (Beckman In-struments, Inc.), 20 g; and PPO, 5 g. Details of themethod of calculation of "4C incorporation into cellsare given below.

RESULTSIncorporation of glucose into cells. In order

to determine fermentation balances for R. albusand the mixed cultures, it was necessary tomeasure the amount of glucose incorporatedinto cells. This was accomplished by determin-ing the incorporation of [U-"C ]glucose into cellsfrom cultures of R. albus alone and cells fromthe mixed culture. If labeled glucose were addedto a usual culture system in which the satura-tion constant for glucose uptake is very small,the label would be rapidly taken up by the cellsand incorporated into the cells or degraded toend products. The counts found in the cellswould then decrease with a rate constant equalto the dilution rate. A significant amount ofglucose, however, remained in the supernatantfluids, especially at faster dilution rates. This isindicative of a high saturation constant (KJ) forglucose metabolism by R. albus (14). Because ofthe high saturation constant, complete metabo-lism of labeled glucose was not achieved until asignificant time after the addition of label (Fig.1 and 2). The percentage of glucose incor-porated into cells was determined by comparingthe counts per minute found in the cells to thetotal counts per minute in cells per supernatantfluid, i.e., nonmetabolized [4C ]glucose plusfermentation products, at the tinie of sampling.

VOL. 114, 1973 1233

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

IANNOTfI ET AL.

2 3 4 5 6 7 8 9 10 11 12 2 3 4 5 6Time (hr) Time (hr)

FIG. 1. Incorporation of ['4Cjglucose into R. albus cells. Symbols: Theoretical washout (no loss of 14C)(-. . ); 14C in liquid portion (cells included) (-O-); 14C in cells (-0-0); a calculated curve for disappearance of[14C] glucose from the fermentation vessel (- - -). A, Dilution rate 0.16/h; B, dilution rate 0.35/h.

The experimental total counts per minute at thesampling times do not include all productsbecause of loss of 14CO2 from the medium. Inorder to calculate percent incorporation, it wasnecessary to calculate the theoretically ex-pected total counts for the times of sampling.The theoretical (no loss of "4C) total counts

per minute expected at time t were calculatedby first extrapolating the total counts per min-ute in the initial samples back to time zero(to). Separate experiments showed that theextrapolated to values were the same as the tovalues obtained by diluting the 14C label to thesame extent as the label actually added to thefermentors. The theoretical counts per minuteat time t were then calculated from the equa-tion: ln (x/xo) = -D(t), where xo is the ex-trapolated counts per minute at to, x is thecalculated counts per minute at t, and D is thedilution rate. Figures 1A and 1B show thecalculated theoretical washout of total labelfrom the R. albus fermentors at dilution rates of0.16 and 0.35/h along with the actual washout oflabel in the liquid portion (cells included), andthe amount of label incorporated into the cells.Figures 2A and 2B show analogous data for themixed culture at dilution rates of 0.12 and0.37/h. Figures 1 and 2 also show calculatedcurves for the rate of disappearance of[14C ]glucose from the fermentor. It was useful to

calculate these values to determine when essen-tially all of the labeled glucose was converted tofermentation products or cells. The subsequentrate constants for washout of label in cellsshould then be identical to the calculated rateconstants for washout of total label. Calculationof the percent incorporation into cells frommeasured values obtained after labeled sub-strate is completely used eliminates errors thatmay be introduced when sampling is carried outwhen cells are actively metabolizing the labeledsubstrate. Since labeled glucose disappearsfrom the culture vessels as a function of washoutand metabolism, the decrease in concentrationcan be evaluated from the following equations:

-dSc/dt = D(Sr - S) (Sc) + D(Sc) (1)

where Sc is the 114C ]glucose concentration in thefermentor, S, is the glucose concentration in thereservoir, and S is the steady-state glucoseconcentration in the fermentor. D(Sr - S) is therate constant for metabolism of [14C]glucosebecause it is the rate of metabolism of glucose inthe fermentor. D(Sc) evaluates disappearance of[14C ]glucose due to washout. Rearrangementgives,

-dSc/dt = D[(Sr - S) + 1]Sc (2)D[(Sr - S) + 1] is a rate constant (K,) for the

1234 J. BACTERIOL.

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

R. ALBUS FERMENTATION PRODUCTS

disappearance of added ["IC ]glucose from a

particular fermentor. K, can be evaluated at a

particular D by determining S, and equation 2is equal to

-dSc/dt = K,Sc (3)

Integration of (3) gives

Sc = 1/Aekt

Equation 3 was used to calculate the valuesshown in Fig. 1 and 2 for [4C ]glucose disap-pearance from the fermentors. The values for Sused to calculate the respective K, values were

from determinations of glucose in culture super-

natant fluids. The label in cells reaches a peakat approximately the time that the labeledglucose in the fermentor is less than 5% of theinitial label. Subsequent to the attainment ofthe peak, cell label washes out with the samerate constant as the theoretical or actual labelin the effluent from the culture vessel. Since thetheoretical value is the theoretical total label inthe effluent at any time t, the decreasing celllabel at the points of constant, negative slopedivided by the theoretical total value gives theproportion of label incorporated into the cells.The slopes for the washout of labeled cells (Fig.1 and 2) subsequent to the point at whichlabeled glucose in the culture vessel was cal-

culated to be 1% of the original label werecalculated by the method of least squares.Table 1 shows the calculated slopes, the slopesfor theoretical washout, and the calculatedpercent incorporation into cells for each fermen-tor. Since the values for R. albus alone were notsignificantly different, the value of 15% wasused to calculate percent incorporation of glu-cose into cells in all subsequent experiments.For the mixed cultures, we used 21% incorpora-tion into cells for dilution rates of 0.2/h or lessand 25% for dilution rates of 0.3/h or more.

Analysis of R. albus fermentation endproducts. A sample from a culture (dilutionrate of 0.35/h) was analyzed with the completesolvent system of Ramsey with fraction collec-tion. The only acid produced was acetic acid.No propionic or longer-carbon-chain volatileacid, lactic, formic, pyruvic, or succinic acidswere produced. No keto acid or neutral ketone,such as acetone, could be detected as thehydrazone. Gas chromatographic analysis ofvolatile fatty acids showed that acetic acid was

the only volatile acid produced. No alcoholsother than ethanol could be found by gaschromatographic analysis. The sample was

Voges-Proskauer negative. The methods util-ized would detect most common microbial fer-mentation end products except glycerol.Samples from cultures run at different dilu-

2 3 4 5 6 7

Time (hr)8 9 10 11 12

Time (hr)

FIG. 2. Incorporation of [14C]glucose into cells. Symbols: Theoretical washout (no loss of 14C) (--- -); 14C inliquid portion (cells included) (-O-); 14C in cells (-0-); calculated curve for disappearance of [14C]glucose fromthe fermentation vessel (-- -). A, Dilution rate 0.12/h; B, dilution rate 0.35/h.

I.-

a~

- ~~~Mixed D= 0.12 hr7'

\Cells plus Mediu105 ^

Cells

IGlucose in Fermentor (Calculated)

103 l

1235VOL. 114, 1973

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

IANNOTTI ET AL.

TABLE 1. Calculation of percent of [U-'4Cjglucoseincorporated into cells

Dilution rate/h

Experimental values R. albus R. albus + V.succinogenes

0.16 0.35 0.12 0.35

Washout slopes (log countsper min per h)Theoretical total 0.069 0.152 0.052 0.152

radioactivityCell radioactivity 0.070 0.141 0.053 0.142

104 counts per min per ml at yinterceptTheoretical total 14.45 14.15 18.10 24.20

radioactivityCell radioactivity 2.13 2.07 3.81 5.98

Glucose incorporated into 14.7 14.6 21.0 24.7cells (%)

tion rates were then analyzed for acids with theRamsey column modified for batch collection,for ethanol by gas chromatography, for glucosewith glucose oxidase, and for H2 by volumemeasurements and gas chromatography. Table2 shows the data collected for all analyses at alldilution rates. The acetate values in Table 2 arethe gas chromatography values. The titrationvalues obtained from the modified Ramseyprocedure were slightly higher but were consid-ered to be less accurate because of the poortitration end point obtained with the low con-

centrations of acetate in large volumes of sol-vent. Formate, lactate, or succinate productionwere not detected in any of the samples by themodified Ramsey procedure. The glucose valuesin Table 2 are corrected for incorporation ofglucose into cells. Table 3 shows the same

product data expressed as moles of product per100 mol of glucose along with the means andstandard deviations. The CO2 values were cal-culated by assuming that the moles of CO2produced were equal to the sum of the moles ofethanol and acetate produced. Carbon recover-ies and the oxidized products/reduced products(O/R) index, calculated from the means, were72% and 0.76, respectively.As can be seen in Table 3, the data are

scattered. These are most likely due to inherentexperimental difficulties. H2 does not equili-brate in the gas phase of the fermentationapparatus, and it is necessary to use gas sam-ples from several points in the apparatus for gaschromatography to obtain best values for the H2concentration in a particular portion of the gasphase in the system (8). The commercial Tryp-ticase in the medium is contaminated with

acetate at a level which gives a blank concentra-tion of 5.6 to 6.5 mM in the medium, or about 1to 4 times the concentrations of acetate pro-duced in the culture vessels.Analysis of mixed-culture fermentation

products. Separate batch-culture experimentsshowed that V. succinogenes did not grow on themedium used in the chemostat unless R. albuswas present. The concentration of fumarate inthe medium remains unchanged in batch orcontinuous monocultures of R. albus. Examina-tion of selected, continuous, mixed cultures bygas chromatography and liquid chromatogra-phy on silicic acid with the Ramsey columnshowed that acetic and succinic acids were theonly acid products. Gas chromatographic analy-sis showed that no alcohols, including ethanol,propanol, isopropanol, or butanol were pro-duced in the mixed culture. Acetoin or diacetylwere not produced. No H2 ever accumulated inthe mixed-culture system.Samples from cultures run at different dilu-

tion rates were then analyzed for acids with theRamsey column modified for batch collection,for ethanol and acetate by gas chromatography,and for glucose with glucose oxidase. Table 4shows the data collected for all analyses at all

TABLE 2. Glucose fermented and products producedby R. albus at different dilution ratesa

Dilution Glucose Productsrate (h) fer-dmentedic Ethanolc Acetatec H,c H,d0.17 3.9 3.2 2.7 9.7 .210.17 4.1 3.4 2.7 8.6 .180.25 4.2 2.6 4.9 11.1 .210.25 4.1 3.4 3.7 7.2 .150.25 3.5 2.5 2.3 9.6 .190.33 3.0 2.2 2.3 9.2 .190.33 3.4 2.0 2.3 9.4 .200.49 3.5 2.0 2.1 5.9 .140.59 2.7 1.4 1.9 4.6 .070.59 2.7 1.8 1.6 7.2 .16

a An amount of medium equivalent to approxi-mately four to five times the fermentor volume (345ml) was pumped through the fermentor and removedfrom the collection flask at each dilution rate; approx-imately 2.5 liters was collected for analysis at eachdilution rate. The time for collecting 2 liters rangedfrom approximately 64 h for dilution rate 0.09/h to 12h for dilution rate 0.59/h.

° Glucose fermented was calculated by correctingthe difference between the medium reservoir andculture supernatant fluids for the amount incor-porated into cells.

c Micromoles used or produced per milliliter ofculture.

dPartial pressure at time of sampling in atmos-phere.

1236 J. BACTERIOL.

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

R. ALBUS FERMENTATION PRODUCTS

TABLE 3. Products produced by R. albus per 100 molglucose

Products (mol/100 molDilution Ethanol of glucose)rate (h)

Acetate H, co2a

0.17 82 69 249 1510.17 83 66 211 1490.25 62 117 266 1790.25 83 90 176 1730.25 71 67 275 1380.33 74 76 310 1500.33 59 66 278 1250.49 57 60 170 1170.59 52 71 172 1230.59 65 59 264 124

Mean 69 74 237 143SDb 11 17 51 21SEC 4 5 16 7

a Calculated as equal to ethanol plus acetate.'Standard deviation.C Standard error.

TABLE 4. Glucose fermented, fumarate used, andproducts produced at different dilution rates in mixed

culturesa

Glucose Fumarate Products (mM)Dilution fer- Fumaaedrate (h) mented used Acetate Succinate

(MM) (m) Aeae (H2)b

0.11 4.4 18.8 6.2 16.60.12 4.4 18.7 6.2 15.90.19 3.9 18.5 6.0 15.60.19 4.0 18.2 5.5 15.60.33 3.3 13.9 5.3 12.70.35 3.3 15.6 4.5 15.20.65 1.2 4.8 2.0 4.60.65 1.3 5.0 2.0 4.2

a See Table 2, footnotes a and b, for experimentaldetails.

" H2 is considered to be equivalent to the amount ofsuccinate produced.

dilution rates. The glucose values are correctedfor incorporation of glucose into cells. Table 5shows the same product data expressed as molesof product per 100 mol of glucose fermentedalong with the means and standard deviations.Succinate is considered to be equivalent to thenumber of moles of H2 produced by R. albusand used by V. succinogenes (H2 + fumarate -

succinate). The CO2 values were calculated byassuming that the moles of CO2 produced wereequal to the moles of acetate produced. Carbonrecoveries and the oxidized/reduced products(O/R) index, calculated from the means were74% and 0.77, respectively.

Acetate incorporation into cells. One possi-ble reason for the low carbon recovery andexcess of reduced products could be the netuptake of significant amounts of acetate, pres-ent in the uninoculated medium, into cellmaterial. If the missing carbon is calculated asacetate and equivalent CO,, the O/R balancebecomes 1.04. We examined the amount of[I4C ]acetate incorporated into cell material in abatch culture of R. albus grown with 5.4 mMglucose, in the same medium used in thechemostat experiments. The [4C ]acetate (12.3mCi/mmol) was added to the uninoculatedmedium (50 nCi/ml) which contained 5.3 mMacetate. The medium acetate was a contami-nant in the commercial Trypticase and wasbetween 5.5 and 6.5 mM in all of the media usedin these studies. When 5.5 mM glucose wasused, the final level of acetate in the mediumwas 8.7 mM and the cells had incorporated 5%of the ["C]acetate. Under these same condi-tions, the cells incorporated 9% of added["C]glucose. Assuming a similar relationshipbetween glucose and acetate incorporation inthe chemostat cultures, the total acetate valuescould be low by about 8.3% because of theapproximately 15% incorporation of glucose intocells in the chemostat. For the mean values ofTable 4, this acetate and equivalent CO2 wouldaccount for approximately 4% of the missingcarbon. Although an acetate-incorporation ex-periment was not done with a mixed culture,the similarities between C recoveries and O/Rindices for single and mixed continuous culturesindicate that it is unlikely that acetate incorpo-ration can account for a significant amount ofthe missing C in either case.

TABLE 5. Fumarate used and products produced per100 mol of glucose in mixed cultures

Products

Dilution Fumarate (mol x 100 mol of glucose)rate (h) used Succinate Co,a

Acetate (H2) GO?

0.11 426 139 376 1390.12 423 139 359 1390.19 473 153 398 1530.19 457 137 391 1370.33 419 158 383 1580.35 471 137 459 1370.65 406 168 389 1680.65 373 147 313 147

Mean 431 147 384 147SD 34 11 41 11SE 12 4 20 4

a Calculated as equal to acetate.

1237VOL. 114, 1973

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

IANNO1TI ET AL.

DISCUSSIONTable 6 shows a comparison of the fermenta-

tion balance data of the monoculture of R. albuswith the data for the mixed R. albus-V.succinogenes system. Although ethanol is amajor product of the R. albus monoculturefermentation, ethanol is not produced in themixed system. Instead, acetate production in-creased in the mixed system by an amountalmost equivalent to the amount of ethanolproduced by the monoculture of R. albus. Theamount of succinate produced (H2 equivalents)indicates an increase in H2 production in anamount almost equivalent to the electron deficitcaused by the disappearance of ethanol and theincrease in acetate in the monoculture to mixedculture change. These results essentially provethe Hungate hypothesis (7), i.e., a H2-utilizingorganism can cause electrons to be shifted awayfrom the production of a typical fermentationproduct, e.g., ethanol, to give the more oxidizedproduct, acetate. The shifted electrons probablyare converted to H2 (see below) and used by theH2-utilizer for reduction of its oxidized sub-strate. Thus, the explanation offered by Hun-gate (7) for the usual lack of production orintermediate formation of ethanol in the rumenecosystem appears to be correct, although, inthe rumen ecosystem, methane-forming bacte-ria would be the major H2-utilizing organisms.It is likely that the same explanation offered byHungate (7) for the usual lack of production or

TABLE 6. Comparison of R. albus monoculturefermentation with R. albus-V. succinogenes mixed

fermentation

(Mol x 100 mol of glucose)

Substrate R. albus R. albus plusused and V. succinogenesproducts

Uncor- Correcteda Uncor- Correctedarected rected Cretd

SubstrateFumarate 0 0 431 431

ProductsEthanol 69 69 0 0Acetate 74 131 147 200H2b 237 237 384 384CO2C 143 200 147 200

C recovery 73 100 74 100

O/R 0.76 1.09 0.77 1.04

a Assuming that the missing carbon is acetate plusCo2.

Succinate equivalents in the case of the mixedculture.

c Calculated as equal to 2 carbon compounds.

intermediate formation of lactate in the rumenis also correct.Although the lack of complete carbon recov-

ery in the fermentation balances cannot beexplained on the basis of definitive experimen-tal evidence, it seems highly unlikely that themissing carbon can be implicated in the fermen-tation product shift that occurs in the transitionfrom monoculture to mixed culture. It wouldhave to be argued that unknown oxidized prod-uct, produced in the monoculture, becomesreduced in the mixed culture, at the expense ofethanol production, to an unknown product.This would be inconsistent with the observationthat the increased succinate production (H2equivalents) in the mixed culture is almostequivalent to the electron deficit caused byethanol disappearance in the change from mon-oculture to mixed culture. Since the carbonrecovery and O/R index are almost identical inthe monoculture and mixed-culture experi-ments (Table 6), it seems reasonable to assumethat V. succinogenes does not influence theproduction of the as yet unidentified product(s)produced by R. albus.We will now discuss the probable biochemical

explanation of the shift of electrons fromethanol production in the monoculture to succi-nate formation in the mixed system. For thispurpose, it is convenient to start with theprobable fermentation equations for both sys-tems. Table 6 shows the experimental fermenta-tion values and the same values corrected on theassumption that the missing carbon is in theform of acetate and CO2. As was previouslyindicated, a compound or compounds equiva-lent to acetate plus CO2 would provide thecarbon and O/R state necessary to satisfy abalanced equation for glucose fermentation.Table 7 shows likely partial reactions for themonoculture and mixed-culture fermentations,based on the corrected values of Table 6, andthe sums of the partial reactions. Although mostof the reactions have not been directly demon-strated in R. albus, they represent commonbacterial enzyme reactions. The most uniqueaspect of the partial sequence is the postulationof a nicotinamide adenine dinucleotide (re-duced form) (NADH)-linked hydrogenase in R.albus. The presence of the hydrogenase wouldpermit two alternative electron pathways forthe reoxidation of the NADH generated byoxidation of glyceraldehyde-3-phosphate (G3P).Either acetyl-coenzyme A could be reduced,through acetaldehyde, to ethanol, or the NADHcould be directly oxidized to H2 and nicotin-amide adenine dinucleotide (NAD). The freeenergy change at pH 7 and 25 C (AG',) for

1238 J. BACTERIOL.

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

R. ALBUS FERMENTATION PRODUCTS

TABLE 7. Partial reactions for monoculture and mixed culture fermentations

Type of culture

Monoculture ormixed culture

Monoculture

Mixed culture

Reactiona

100 CsH,20, + 200 ATP - 200 glyceraldehyde-3-P(G3P) + 200 ADP200 G3P + 200 NAD+ + 200 Pi _ 200 1,3-diphosphoglyceric acid + 200 NADH + 200 H+200 DPGA + 400 ADP - 200 CH2COCO2H + 400 ATP200 CH.COCO2H + 200 CoASH - 200 CHsCOSCoA + 200 H2 + 200 CO2

70 CH3COSCoA + 140 NADH + 140 H+ - 70 C2H,OH + 140 NAD+ + 70 CoASH130 CHsCOSCoA + 130 ADP + 130 Pi _ 130 CH8CO2H + 130 ATP + 130 CoASH60 NADH + 60 H+ -.60 H2 + 60 NAD+

Sum: 100 CsHl20. + 330 ADP + 330 Pi - 70 C2H,OH + 130 CH8C02H + 200 CO2 + 260 H2 +330 ATP

200 CH,COSCoA + 200 ADP + 200 Pi - 200 CH2CO2H + 200 ATP + 200 CoASH200NADH + 200H+ 200H2 + 200NAD+400 H2 + 400 fumarate - 400 succinate

Sum: 100 C6HI20. + 400 ADP + 400 Pi + 400 fumarate -.200 CH2C02H + 200 CO2+ 400 succinate+ 400 ATP

a Abbreviations: ATP, adenosine triphosphate; ADP, adenosine diphosphate; G3P, glyceraldehyde-3-phos-phate; NAD, nicotinamide adenine dinucleotide; NADH, reduced NAD; DPGA, diphosphoglyceric acid;CoASH, coenzyme A.

the oxidation of G3P to H, and 1, 3-diphospho-glyceric acid (DPGA) is +5.83 kcal (3). Theequilibrium constant is 5.28 x 10-6. This ener-getically unfavorable reaction can be coupled tothe exergonic production of ethanol in themonoculture or the exergonic formation of suc-cinate in the mixed system. The formation of H,from the conversion of pyruvate to acetate andCO2 is a thermodynamically favorable reaction,i.e., AG'o is -13.69 kcal (3). Although theoxidation of G3P to DPGA and H, is not athermodynamically favorable reaction, the bal-ance of intermediate reactions for the monocul-ture (table 7) suggests that the specific condi-tions in the cell and its environment permitsome flow of electrons from NADH to H2. Theequations of Table 7 suggest that approximately30% of the NADH generated by G3P oxidationis oxidized to H2 and NAD. In the mixedcultures, however, it would seem that 100% ofthe NADH generated in glycolysis is oxidized toH2, which, in turn, is used for succinate forma-tion.The situation is analogous to the formation of

CH, and acetate from ethanol or pyruvate andCO2 by a mixed-culture system (2, 11-13). Inthis case, the pulling of the NAD-dependentoxidation of ethanol to acetaldehyde and H2 bya H2-producing bacterium, called S organism, isobligatorily linked to CH, formation by theH2-utilizing methane-forming bacterium, i.e.,no alternative electron sink exists for the mono-

culture of the ethanol-oxidizing S organism.Interspecies H2 transfer is obligatory for goodgrowth of the ethanol oxidizer. S organism canuse pyruvate as an energy source in the absenceof a methanogenic bacterium. The major prod-ucts of pyruvate fermentation are ethanol, ace-tate, and CO2 (11). In the presence of a H2-uti-lizing, methanogenic bacterium, ethanol pro-duction by S organism is almost completelyeliminated, and acetate production increases inan amount equivalent to ethanol produced by Sorganism alone (11). In the case of R. albus,interspecies electron transfer is not obligatoryfor growth on glucose, but it does cause a similarshift from ethanol to acetate formation andprobably also in the amount of energy (ATP)available to the carbohydrate-fermenting orga-nism for biosynthetic purposes.

It remains to be proven, however, that R.albus does contain an NAD-linked hydrogenaseactivity. Preliminary experiments (unpublisheddata) show that NAD or nicotinamide adeninedinucleotide phosphate (oxidized form) can bereduced by H2 with cell-free extracts of R.albus, but only when methyl or benzyl viologenis present. We will continue our efforts todemonstrate a physiological system for thereduction of pyridine nucleotides by hydrogen.We feel, however, that the results in this paper,together with the results of Reddy et al. (11-13)on the association of S organism with H2-utiliz-ing, methanogenic bacteria, provide strong evi-

1239VOL. 114, 1973

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

IANNOTTI ET AL.

dence for a type of interspecies electron transfermediated by H2 gas which can be of greatimportance in the determination of productformation in anaerobic ecosystems.There appears to be little significant influ-

ence of growth rate on the kinds and amounts ofproducts of glucose fermentation in the single ormixed cultures. Significant amounts of formatewere not detected at any growth rate. Otherinvestigators reported that formate was a majorproduct of cellobiose fermentation in batchcultures of R. albus (1, 4, 9). Dehority, however,reported that R. albus strain 7 produced for-mate from cellobiose when cultivated in anatmosphere containing 100% CO2, but no for-mate was produced when growth was limited bythe presence of only 5% CO2 in the gas atmos-phere (4). Dehority suggested that the CO2concentration regulated the production of for-mate. This would not be the case in our continu-ous-culture experiments which were carried outin the presence of 100% CO2. Preliminary exper-iments carried out by T. L. Miller in ourlaboratory show that R. albus strain 7 producesformate from glucose only during the late logand stationary phases of growth in batch cul-ture in a medium identical to that used in thecontinuous-culture experiments except for theconcentration of glucose. The amount of formateproduced in batch culture depends on theamount of glucose added, i.e., how much resid-ual glucose remains to be fermented in the latelog and stationary phases. It is highly probablethat steady-state growth on limiting glucose isthe explanation for the lack of formate produc-tion by the continuous cultures of R. albus.

ACKNOWLEDGMENTSThis investigation was supported by Public Health Service

research grant AI 09467 from the National Institute of Allergyand Infectious Diseases, and by a Department of AgricultureGrant Hatch 35-325. D.K. was supported by a Public HealthService postdoctoral fellowship GM 46647 from the NationalInstitute of General Medical Sciences.We thank Jack Althaus, Sue Hensley, and Melvin Crabill

for technical assistance.

LITERATURE CITED1. Bryant, M. P., N. Small, C. Bouma, and I. M. Robinson.

1958. Characteristics of ruminal anaerobic cellulolyticcocci and Cillobacterium cellulosolvens n. sp. J. Bac-teriol. 76:529-537.

2. Bryant, M. P., E. A. Wolin, M. J. Wolin, and R. S. Wolfe.1967. Methanobacillus omelianskii, a symbiotic as-sociation of two species of bacteria. Arch. Microbiol.59:20-31.

3. Burton, K. 1957. Free energy data of biological interest, p.275-285. In H. A. Krebs and H. L. Kornberg (ed.),Energy transformations in living matter. Springer-Ver-lag, Berlin.

4. Dehority, B. A. 1971. Carbon dioxide requirement ofvarious species of rumen bacteria. J. Bacteriol.105:70-76.

5. Hammarstrand, K. 1966. Analysis of free fatty acids, p.9-10. In K. Hammarstrand (ed.), Chromatographicanalysis of fatty acids. Varian Aerograph, WalnutCreek, Calif.

6. Hungate, R. E. 1963. Polysaccharide storage and growthefficiency in Ruminococcus albus. J. Bacteriol.86:848-854.

7. Hungate, R. E. 1966. Quantities of carbohydrate fermen-tation products, p. 245-280. In The rumen and itsmicrobes. Academic Press Inc., New York.

8. Kafkewitz, D., E. L. Iannotti, M. J. Wolin, and M. P.Bryant. 1973. Anaerobic chemostat that permits thecollection and measurement of fermentation gases.Appl. Microbiol. 25:612-627.

9. Kistner, A. and L. Gouws. 1964. Cellulolytic cocci occur-ring in the rumen of sheep conditioned to lucerne hay.J. Gen. Microbiol. 34:447-458.

10. Ramsey, H. A. 1963. Separation of organic acids in bloodby partition chromatography. J. Dairy Sci. 46:480-483.

11. Reddy, C. A., M. P. Bryant, and M. J. Wolin. 1972.Characteristics of S organism isolated from Methanol-bacillus omelianskii. J. Bacteriol. 109:539-545.

12. Reddy, C. A., M. P. Bryant, and M. J. Wolin. 1972.Ferredoxin- and nicotinamide adenine dinucleotide-dependent H, production from ethanol and formate inextracts of S organism isolated from "Methanobacillusomelianskii." J. Bacteriol. 110:126-132.

13. Reddy, C. A., M. P. Bryant, and M. J. Wolin. 1972.Ferredoxin-dependent conversion of acetaldehyde toacetate and H, in extracts of S organism. J. Bacteriol.110:133-138.

14. Tempest, D. W. 1970. The continuous cultivation ofmicro-organisms. I. The theory of the chemostat, p.259-276. In Norris, J. R. and D. W. Ribbons (ed.),Methods in microbiology, vol 2, Academic Press Inc.,New York.

15. Wolin, M. J., E. A. Wolin, and N. J. Jacobs. 1961.Cytochrome-producing anaerobic vibrio, Vibrio suc-cinogenes, Sp. N. J. Bacteriol. 81:911-917.

1240 J. BACTERIOL.

on March 5, 2020 by guest

http://jb.asm.org/

Dow

nloaded from