Vol. 169, No. 2

Involvement of Glutamate in the Respiratory Metabolism ofBradyrhizobium japonicum Bacteroidst

SEPPO 0. SALMINEN AND JOHN G. STREETER*Department ofAgronomy, Ohio State University, and Ohio Agricultural Research and Development Center,

Wooster, Ohio 44691

Received 20 August 1986/Accepted 12 November 1986

Bradyrhizobium japonicum bacteroids were isolated anaerobically and supplied with 14C-labeled succinate,malate, aspartate, or glutamate for periods of up to 60 min in the presence of myoglobin to control the 02

concentration. Succinate and malate were absorbed about twice as rapidly as glutamate and aspartate.Conversion of substrate to CO2 was most rapid for malate, followed by succinate, glutamate, and aspartate.When CO2 production was expressed as a proportion of total carbon taken up, malate was still the most rapidlyrespired substrate, with 68% of the label absorbed converted to CO2. The comparable values for succinate,glutamate, and aspartate were 37, 50, and 38%, respectively. Considering the fate of labeled substrate notrespired, >95% of absorbed glutamate remained as glutamate in the bacteroids. In contrast, from 39 to 66%of the absorbed succinate, malate, or aspartate was converted to glutamate. An increase in the rate of CO2formation from labeled substrates after 20 min appeared to coincide with a maximum accumulation of label inglutamate. The results indicate the presence of a substantial glutamate pool in bacteroids and the involvementof glutamate in the respiratory metabolism of bacteroids.

A question about symbiotic nitrogen fixation which re-mains unresolved is what compounds are supplied by thehost to the microsymbiont to support the high demand forreducing equivalents in the conversion of N2 to NH4+. Mostof the available evidence supports the idea that thetricarboxylic acid cycle acids succinate and malate are thecompounds supplied to N2-fixing bacteroids in legume nod-ules (23). Much of the recent support for this propositioncomes from studies showing that bacterial mutants lackingthe ability to absorb dicarboxylic acids form nodules that areincapable of N2 fixation (2, 7, 18).Other groups have studied the kinetics of uptake of

various compounds by bacteroids and have identified or-

ganic acids as compounds absorbed by active mechanisms(16, 19). Still other groups have studied the ability of variouscompounds to support respiration or N2 fixation (acetylenereduction) by isolated bacteroids (15, 20, 28). The results ofall of these studies support an important role for dicarboxylicacids in bacteroid metabolism and legume nodule function.However, neither of these approaches has involved a com-plete analysis of the metabolic fate of compounds supplied tobacteroids. We report here experiments in which the goalwas to inventory metabolic products. In an early experimentwe discovered that [2,3-14C]succinate is very rapidly con-

verted to glutamate in bacteroids, and this led us to study themetabolic fate of malate, glutamate, and aspartate as well.

MATERIALS AND METHODS

Isolation of bacteroids. Soybean plants [Glycine max (L.)Merr cv. Beeson 80] were inoculated with Bradyrhizobiumjaponicum 61A76 (Nitragin Co., Milwaukee, Wis.) or USDA110 (Nitrogen Fixation and Soybean Genetics Laboratory,U.S. Department of Agriculture, Agricultural Research Ser-vice, Beltsville, Md.) and grown in a greenhouse. Nodulesfrom 37- to 47-day-old plants were chilled (2°C), placed in a

glove box under N2 flow, and ground in a mortar with a

* Corresponding author.t OARDC journal article no. 107/86.

pestle, with 0.15 M sodium phosphate buffer, pH 7.5, addedin 2-ml portions to a final volume of 2 ml of buffer per g ofnodules. The homogenate was filtered through twoMiracloth disks in a 50-ml syringe into a 30-ml centrifugetube. The tube was capped tightly and centrifuged at 6,000 xg for 10 min. The tube was opened in the glove box, thesupernatant was discarded, and the pellet was washed twiceby pipetting 1 ml of the buffer over it. The pellet was

resuspended, and the mixture was again filtered throughMiracloth disks to remove any unsuspended bacteroids andadjusted to a final concentration of 2 ml of buffer per g ofnodules. This suspension was used in the studies on uptakeand metabolism. The protein concentration of the bacteroidsuspension was determined by the method of Lowry et al.(10).Uptake and metabolism. All substrates were purchased

from Amersham Corp. and were uniformly labeled exceptsuccinate, which was labeled in the 2 and 3 or the 1 position.[2,3-14C]- and [1-14C]succinate were mixed to give equalradioactivity in each carbon, providing a substrate whichwas statistically equivalent to [U-14C]succinate.

Reaction mixtures (total volume, 2.0 ml) were made up in15-ml Corex centrifuge tubes (15 [inner diameter] by 100mm) and contained 1 ml of bacteroids, 0.2 ml of 0.4 mMmyoglobin (Sigma Chemical Co.) oxygenated as describedpreviously (16), and 0.3 ml of 0.15 M sodium phosphatebuffer, pH 7.5. The tubes were flushed with N2 and sealedwith a serum stopper with an attached cup containing a filterpaper wick impregnated with 70 p.l of 10% KOH. A sample(1.4 ml) was removed from the gas phase and replaced with1.4 ml of Ar-02 (80:20), giving a final concentration of 2% 02

in the gas phase. Buffered substrate (0.5 ml; 1 ,umol at 1,uCi/,umol) was then injected into the reaction mixture, andthe tubes were shaken in a rotary shaker at 250 rpm.Incubation was terminated by placing the tubes in an icebath. The stoppers were quickly removed, and the tubeswere centrifuged at 25,000 x g for 10 min. The supernatantwas discarded, and the pellet was washed twice by pipetting1 ml of the cold buffer over it and discarding the washes. The

495

JOURNAL OF BACTERIOLOGY, Feb. 1987, p. 495-4990021-9193/87/020495-050$02.00/0Copyright X3 1987, American Society for Microbiology

on Septem

ber 30, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

496 SALMINEN AND STREETER

pellets were then extracted two times with 80% ethanol, andthe ethanolic extracts were combined and dried in a flashevaporator. The residues were taken up in 2.0 ml of H20.The aqueous samples were applied to Dowex 50 H+ andDowex 1-formate columns (0.7 cm diameter by 2 cm bedheight) in tandem. The neutral fraction was eluted with H2O.The columns were then separated, and amino acids wereeluted from the Dowex 50 H+ column with 2 N HCI and theorganic acids from the Dowex 1-formate column with 4 MHCOOH. The eluates were evaporated to dryness and takenup in 1 ml of 80% ethanol. Portions of the samples werecounted in a cocktail containing 24 g of 2,5-diphenyloxazole(PPO) and 1 g of 1,4-bis(5-phenyloxazolyl)benzene in 2 litersof toluene plus 2 liters of Triton X-100. The KOH-impregnated filter paper was assayed for radioactivity with aPPO-toluene scintillation cocktail consisting of 6 g of PPOper liter of a mixture of toluene and methoxyethanol (2:1).The amino acids were separated by thin-layer chromatog-

raphy. Each plate was spotted with 5 ,ul of amino acidstandard containing 9.1 ,umol of each of 15 amino acids perml and 20 ,ul of concentrated sample from the amino acidfraction. Chromatography and analysis of the spots werecarried out as described previously (22).

Zero-time values were obtained and subtracted from allvalues to give the results reported in figures and tables. Incalculations, we assumed that 1 mol of succinate yielded 4mol of CO2. Nanomoles of substrate in bacteroids and inrespired CO2 were calculated by assuming a constant spe-cific radioactivity of 1 tXCi/,xmol in endogenous substrate.Although the endogenous substrate levels were probablyquite small, any endogenous substrate, e.g., succinate,would have lowered the specific activity of [14C]succinateand led to underestimation of the actual rates. Thus, thereported rates represent minimum values.

Incorporation of radioactivity into the ethanol-insolublefraction of bacteroids was evaluated by incubating insolubleresidue with 0.3 N KOH at 37°C for 24 h. Radioactivitysolubilized by KOH after bacteroids were fed with labeledmalate or glutamate for 60 min amounted to 20% of the totalradioactivity taken up by the bacteroids. Characterization ofthe ethanol-insoluble radioactivity was not pursued in thesestudies, and this fraction is not considered in the resultspresented here. Thus, total uptake is defined as the sum ofthe amount of substrate converted to CO2 and 80% ethanol-soluble compounds.

RESULTSFreedom of the bacteroid preparations from contaminat-

ing mitochondria was established in two ways. First, malatedehydrogenase activity of sonicated bacteroids was notinhibited by inhibitors of mitochondrial malate dehydroge-nase activity (29), indicating that the preparation was virtu-ally free of mitochondria (S. 0. Salminen and J. G. Streeter,Plant Physiol., in press). Second, bacteroid cytochromeoxidase is insensitive to CO, whereas mitochondrial cyto-chrome oxidase is highly sensitive (1). A cytochrome oxi-dase assay (5) of the same sonicated bacteroid preparationgave no evidence of CO sensitivity.

In a preliminary experiment we also compared the rate ofCO2 formation from succinate with 0.04 mM myoglobin(Sigma Chemical Co.), 0.05 mM soybean leghemoglobin, orno 02 carrier (4). Both myoglobin and leghemoglobin treat-ment resulted in significantly greater CO2 evolution than thecontrol, and there was essentially no difference between thehemoproteins; myoglobin (0.04 mM) was used in all subse-quent experiments.

,30 /E

0

10a.

0

0 20 40 60TIME ( minutes)

FIG. 1. Uptake of 14C-labeled succinate (O), malate (/A), gluta-mate (O), and aspartate (M) by B. japonicum bacteroids. Allsubstrates were uniformly labeled except succinate, which waslabeled in the 2 and 3 carbons. Results for a 60-min incubation with[U_14C]succinate are indicated (a). Nanomoles of substrate inethanol extracts of bacteroids a'nd in respired C02 were calculatedby assuming a constant specific radioactivity of 1 ,uCi/,umol inendogenous substrate. In the figure, total uptake represents the sumof these two components; the vertical bars represent two standarderrors, and each curve represents a different bacteroid preparation.The total bacteroid protein in each reaction mixture was 7.6 mg forsuccinate, 7.1 mg for glutamate, 9.3 mg for malate, and 12.4 mg foraspartate feeding.

Bacteroids prepared and incubated as despribed abovewere checked for nitrogenase activity by an acetylene reduc-tion assay (11) in a reaction mixture that was the same as thatused with radioactive substrates. The 2%0°2 concentrationwas not optimal (11), but the rates, after an initial lag period,were linear for at least 90 min. The rate in the absence ofexogenous substrates was 74 -+- 1 mol/g of nodule (freshweight) per h, comparable to the rates reported elsewherefor this 02 concentration (11). The addition of 1 ,uwmol ofmalate or glutamate resulted in a 25%o increase in the rate ofC2H2 reduction. This level of stimulation of nitrogenase islow, but the substrate concentration used here (0.5 mM) isless than that generally used (S mM). No effort was made inthe present study to establish the optimal conditions fornitrogenase activity.The uptake of aspartate was linear for 1 h, whereas the

rate of succinate, glutamate, and malate uptake appeared todecline after 20 min (Fig. 1). The dicarboxylic acids weretaken up three to five times faster than the amino acids. Thetotal uptake in 60 min represented 23% of the succinate inthe reaction mixture. The corresponding percentages formalate, glutamate, and aspartate were 20, 11, and 8%,respectively. Carbon dioxide was evolved from malate muchfaster than from any other substrate (Fig. 2). The rate of C02evolution increased with longer incubation times, the rateafter 20 min being clearly greater than the rate before 20 min

J. BACTERIOL.

on Septem

ber 30, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

GLUTAMATE IN BACTEROID METABOLISM 497

60 -

0a 50_

E

240U)

0

E0

C 30

z

0

j20

-J

0

w

cN 100

0 20 40 60

TIME (minutes)

FIG. 2. CO2 production from labeled substrate. Nanomoles ofCO2 were calculated by assuming no change in the specific activityof endogenous substrate. See the legend to Fig. 1 for symbols andexperimental conditions.

(Fig. 2). When the radioactivity in bacteroids incubated with['4C]succinate or [14C]malate was analyzed, very little radio-activity was found in the organic acid fraction at anyincubation time. Virtually all of the recovered radioactivitywas in the amino acid fraction, with about 72% of the label inglutamate (data not shown). The fate of aspartate andglutamate label in bacteroids was similar to that of malateand succinate.A minor concern with these initial experiments was that

different bacteroid preparations were used for each sub-strate. Also, [2,3-14C]succinate had been compared withother substrates which were uniformly labeled. Therefore,

TABLE 2. Labeled substrate found in individual amino acids"

Radioactive Amino acid accumulatedb (nmol/mg of protein per h,substrate mean + SE)supplied Glutamate Alanine Aspartate

Succinate 6.45 + (0.1) 2.92 ± (0.15) 0.22 ± (0.02)Malate 3.82 ± (0.24) 0.75 ± (0.46) 0.03 ± (0)Glutamate 3.69 ± (0.04) 0 0.03 ± (0.01)Aspartate 2.90 ± (0.19) <0.01 0

a Data are from the same 1-h experiment reported in Table 1. Means andstandard errors for duplicate samples are given.

b Amounts recovered from thin-layer chromatography plates. Correctionsfor recoveries were made for the percentages reported in the text. Labeling ofother amino acids was essentially nil (<0.01 nmol).

all four substrates were compared with the same bacteroidpreparation, with a single 60-min incubation time and [U-'4C]succinate (Table 1). The results agreed with the data inFig. 1 and 2. Succinate was taken up most rapidly, whereasmalate was the most rapidly respired substrate. Relative tototal uptake, 68% of the malate was converted to C02,whereas the corresponding value for succinate was 37%.Most of the difference was accounted for by the amino acidfraction; 48% (12.8 nmol/mg of protein per h) of the suc-cinate taken up was in this fraction, whereas only 26% of themalate was converted into amino acids. Glutamate andaspartate were taken up more slowly than succinate, butglutamate had a greater relative rate of conversion to CO2than succinate. In fact, the major fate of glutamate appearedto be conversion to CO2

Thin-layer chromatography of the amino acid fractionindicated that glutamate was consistently the most highlylabeled amino acid (Table 2). With succinate feeding, theradioactivity in the glutamate spot [6.45 nmol/mg of proteinper h) represented 67% of the radioactivity recovered fromthe plate (total recoveries were ca. 80 to 90% of the radio-activity spotted); for malate, glutamate represented 80% ofradioactivity, for glutamate 97%, and for aspartate 88%. Thelabeling of glutamate was quite rapid; there was little furtherincrease in glutamate radioactivity after the first 20 min (Fig.3). Only when aspartate was fed was there a continuedincrease in 14C-labeled glutamate. The marked differencesamong substrates in the labeling of glutamate (Fig. 3) seem toindicate that the bacteroids were depleted of glutamate whenfeedings were initiated and that the capacity for accumula-tion of glutamate is substantial. Alanine and, to a lesserdegree, aspartate were the only other amino acids showingsignificant labeling (Table 2). In time course studies, supply-

TABLE 1. Conversion of '4C-labeled substrates to CO2 and other metabolites by B. japonicum bacteroidsa

Distribution of label (nmol/mg of protein per h, mean t SE)Radioactive Ratio, converted to

substrate suppliedb Amino acid Organic acid Total uptaked Converted to C02/total uptakefractionc fractionc CO2

Succinate 12.8 ± 0.4 1.1 ± 0.1 26.5 ± 0.4 9.8 + 0.3 0.37Malate 5.7 ± 0.7 0.4 ± 0.1 21.8 ± 0 14.8 + 0.9 0.68Glutamate 5.8 ± 0.3 0 12.2 ± 0.6 6.1 ± 0.1 0.50Aspartate 3.8 + 0.2 0.1 + 0 7.2 ± 0.5 2.8 ± 0.1 0.38

a Subsamples of the same bacteroid preparation were used for the comparisons of all four substrates in a 1-h incubation. The total protein per reaction mixturewas 8.6 mg.

b All four substrates were uniformly labeled, [U-_4C]succinate consisting of a mixture of [2,3-_4C]- and [1-'4C]succinate, and all substrates were adjusted to aspecific ratioactivity of 1 p.Ci/j±mol with unlabeled substrate.

c The nanomoles of substrate converted into each fraction were calculated by using the ratio of 1 p.Ci/p.mol in the applied substrate. Means and standard errorsfor duplicate samples are given.

d For all substrates, labeling of the neutral fraction was essentially nil and is not shown. Total uptake does not include 14C incorporated into ethanol-insolublecompounds.

VOL. 169, 1987

on Septem

ber 30, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

498 SALMINEN AND STREETER

E

E0

z

0

w

00 20 40 60

TIME (minutes)

FIG. 3. Accumulation of label in glutamate from the substratessupplied to B. japonicum bacteroids. Nanomoles of glutamate werecalculated by assuming a specific radioactivity the same as thesubstrate supplied. See the legend to Fig. 1 for symbols andexperimental conditions.

ing bacteroids with malate, aspartate, or glutamate resultedin a transient peak in the labeling of alanine, giving maximumlevels in the 20-min samples of 1.33 + 0.13 nmol/mg ofprotein (mean ± standard error) for malate feeding, andvalues of 0.84 ± 0.01 nmol for aspartate feeding and a tracepeak for glutamate feeding. [2,3-'4C]succinate in contrasthas 1.91 ± 0.49 nmol of substrate converted to alanine after20 min. After 60 min the value was 2.80 ± 0.16 nmol.

DISCUSSION

Very few previous studies are comparable to those re-

ported here. The metabolism of 14C-labeled pyruvate, suc-cinate, and acetate by B. japonicum bacteroids has beenreported for periods of 1 to 24 h (21). However, labeling ofindividual compounds in bacteroids was not reported. Glu-tamate was included as a substrate in one experiment butwas only poorly absorbed by bacteroids (21). A problem incomparing our work with the older literature is that the 02

concentration was not controlled in some previous studies(21, 28). It is now clear that bacteroids, which are condi-tioned to a low-02 environment, are markedly influenced by02 concentration in their metabolism of carbon compounds(4, 15, 25, 26). In a recent study with low 02 concentrations,[2,3-14C]succinate was incorporated mainly into tricar-boxylic acid cycle acids (24), with little or no 14C in aminoacids. We suspect that the discrepancy between their resultsand ours may be explainable on the basis that they used an

Ar-02 mixture, whereas we used N2-02. Provided withN2-02, bacteroids should be actively fixing N2 (4), which

would not be the case with Ar-O2. The C2H2 reduction in oursamples continued for the duration of the experiments (60min), and the rates were in agreement with previous data(11).The comparative experiment with the same bacteroid

preparation (Tables 1 and 2) confirmed the features of thetime course experiments (Fig. 1, 2, and 3). (i) Succinate wasmore readily converted to amino acids than was malate. Infact, about half of the succinate metabolized was convertedto amino acids. (ii) Malate was more readily converted toCO2 than was succinate. This result suggests that malateentering the bacteroid is metabolized via a pathway differentfrom that involved in the metabolism of malate formed in thecell. There is increasing evidence for the aggregation ofenzymes in mitochondria (6, 17), as well as in bacteria (3).The resulting "channeling" of metabolites can lead to spe-cific pathways being favored; perhaps such an arrangementexists in Rhizobium bacteroids. (iii) There was essentially noradioactivity in the organic acid fraction. This result indi-cates that organic acid pools in bacteroids are very small.Organic acids taken up are quickly converted either to CO2or to amino acids. (iv) Glutamate and aspartate were takenup slowly relative to organic acids, but both were readilyconverted to CO2. In fact, essentially the only fate ofglutamate was conversion to CO2, and glutamate was morereadily converted to CO2 than was succinate relative to thetotal uptake (Table 1).Our suggestion of the importance of glutamate in bacteroid

metabolism is supported by two other lines of evidence.When 15N2 was supplied to intact soybean nodules and thedistribution of 15N in bacteroids was examined after 5 or 10min, most of the label was in glutamate (14). Although thetotal quantity of '5N in bacteroids was small, the level ofenrichment of glutamate in bacteroids was as great as that inthe cytosol (14). Secondly, a mutant of B. japonicum whichlacks glutamate synthase activity could be induced to fix N2in vitro but had a Fix- phenotype in nodules (12). It hasrecently been suggested that this result may indicate arequirement for glutamate synthesis in bacteroids (9), but itis not clear why transamination could not supply glutamatein these bacteroids.

Several attractive arguments suggesting that glutamatemay serve as an important source of carbon for bacteroidshave recently been put forward (8). The hypothesis issupported by the observation that when glutamate is sup-plied to rhizobia induced to fix N2 in vitro, NH4' assimila-tion is inhibited and the excretion of NH4 is promoted (13,27). However, this hypothesis is not supported by studiesshowing that Rhizobiumn mutants unable to absorbdicarboxylic acids form Fix- nodules (2, 7, 18), and this isespecially true when the mutant is able to grow in culturewith glutamate as the source of carbon (2).We suggest that both dicarboxylic acids and glutamate are

respiratory substrates of Rhizobium bacteroids. Several ofour observations are consistent with some involvement ofglutamate in bacteroid respiration. For example, the in-creased rate of 14Co2 production (Fig. 2) appeared to coin-cide with the leveling off of the radioactivity in the glutamatepool (Fig. 3) when succinate, malate, and glutamate weresupplied. The changes occurred in the absence of anymarked change in uptake rate, at least for malate andglutamate (Fig. 1). Thus, the overall time course resultsindicate a rapid funneling of carbon into the glutamate pooland concomitant conversion of glutamate to C02, the latterprocess reaching equilibrium with the former after the label-ing of the glutamate pool has reached a certain level. The

J. BACTERIOL.

on Septem

ber 30, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

GLUTAMATE IN BACTEROID METABOLISM 499

fact that another amino acid, aspartate, is rapidly convertedto glutamate emphasizes the importance of glutamate inbacteroid metabolism.

ACKNOWLEDGMENTS

We thank William Fuchsman of Oberlin College for the gift ofpurified soybean nodule leghemoglobin and Mary Kilpatrick fortechnical assistance.

This work was supported in part by the U.S. Department ofAgriculture under agreement 83-CRCR-1-1306. Salaries and re-

search support were provided by state and federal funds appropri-ated to the Ohio Agricultural Research and Development Center andthe Ohio State University.

LITERATURE CITED1. Appleby, C. A. 1969. Electron transport systems of Rhizobium

japonicum. I. Haemoprotein P-450, other CO-reactive pig-ments, cytochromes and oxidases in bacteroids from N2-fixingroot nodules. Biochim. Biophys. Acta 172:71-87.

2. Arwas, R., I. A. McKay, F. R. P. Rowney, M. J. Dilworth, andA. R. Glenn. 1985. Properties of organic acid utilization mutantsof Rhizobium leguminosarum strain 300. J. Gen. Microbiol.131:2059-2066.

3. Barnes, S. J., and P. D. J. Weitzman. 1986. Organization ofcitric acid cycle enzymes into a multienzyme cluster. FEBSLett. 201:267-270.

4. Bergersen, F., and G. L. Turner. 1979. Systems utilizing oxy-genated leghemoglobin and myoglobin as sources of free dis-solved 02 at low concentrations for experiments with bacteria.Anal. Biochem. 96:165-174.

5. Cooperstein, S. J., and A. Lazarow. 1950. A microspec-trophotometric method for the determination of cytochromeoxidase. J. Biol. Chem. 189:665-670.

6. Datta, A., J. M. Merz, and H. 0. Spivey. 1985. Substratechanneling of oxaloacetate in solid-state complexes of malatedehydrogenase and citrate synthase. J. Biol. Chem. 260:15008-15012.

7. Finan, T. M., J. M. Wood, and D. C. Jordan. 1983. Symbioticproperties of C4-dicarboxylic acid transport mutants of Rhizo-bium leguminosarum. J. Bacteriol. 154:1403-1413.

8. Kahn, M. L., J. Kraus, and J. E. Somerville. 1985. A model ofnutrient exchange in the Rhizobium-legume symbiosis, p.

193-199. In H. J. Evans, P. J. Bottomley, and W. J. Newton(ed.), Nitrogen fixation research progress. Martinus Nijhoff,Dordrecht.

9. Lane, M., J. Meade, S. S. Manian, and F. O'Gara. 1086.Expression atid regulation of the Escherichia coli glutamatedehydrogenase gene (gdh) in Rhizobium japonicum. Arch. Mi-crobiol. 144:29-34.

10. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. kandall.1951. Protein measurement with the Folin phenol reagent. J.Biol. Chem. 193:265-275.

11. Nelson, L. M., and S. 0. Salminen. 1982. Uptake hydrogenaseactivity and ATP formation in Rhizobium leguminosarumbacteroids. J. Bacteriol. 151:989-995.

12. O'Gara, F., S. Manian, and J. Meade. 1984. Isolation of anAsm- mutant of Rhizobium japonicum defective in symbioticN2-fixation. FEMS Microbiol. Lett. 24:241-245.

13. O'Gara, F., and K. T. Shanmugam. 1976. Regulation of nitrogenfixation by rhizobia: export of fixed N2 as NH4'. Biochim.Biophys. Acta 437:313-321.

14. Ohyama, T., and K. Kumazawa. 1980. Nitrogen assimilation insoybean nodules. II. 5N2 assimilation in bacteroid and cytosolfractions of soybean nodules. Soil Sci. Plant Nutr. 26:205-213.

15. Peterson, J. B., and T. A. LaRue. 1981. Utilization of aldehydesand alcohols by soybean bacteroids. Plant Physiol. 68:489-493.

16. Reibach, P. H., and J. G. Streeter. 1984. Evaluation of activeversus passive uptake of metabolites by Rhizobium japonicumbacteroids. J. Bacteriol. 159:47-52.

17. Robinson, J. B., Jr., and P. A. Srere. 1985. Organization ofKrebs tricarboxylic acid cycle enzymes in mitochondria. J.Biol. Chem. 260:10800-10805.

18. Ronson, C. W., P. Lyttleton, and J. G. Robertson. 1981. C4-dicarboxylate transport mutants of Rhizobium trifolii form inef-fective nodules on Trifolium repens. Proc. Natl. Acad. Sci.USA 78:4284-4288.

19. San Francisco, M. J. D., and G. R. Jacobson. 1985. Uptake ofsuccinate and malate in cultured cells and bacteroids of twoslow-growing species of Rhizobium. J. Gen. Microbiol.131:765-773.

20. Saroso, S., A. R. Glenn, and M. J. Dilworth. 1984. Carbonutilization by free-living and bacteroid forms of cowpea Rhizo-bium strain NGR234. J. Gen. Microbiol. 130:1809-1814.

21. Stovall, I., and M. Cole. 1978. Organic acid metabolism byisolated Rhizobium japonicum bacteroids. Plant Physiol.61:787-790.

22. Streeter, J. G. 1973. In vivo and in vitro studies on asparaginebiosynthesis in soybean seedlings. Arch. Biochem. Biophys.157:613-624.

23. Streeter, J. G., and S. 0. Salminen. 1985. Carbon metabolism inlegume nodules, p. 277-283. In H. J. Evans, P. J. Bottomley,and W. J. Newton (ed.), Nitrogen fixation research progress.Martinus Nijhoff, Dordrecht.

24. Tajima, K., I. Kimura, and H. Sasahara. 1986. Succinatemetabolism of isolated soybean nodule bacteroids at low oxygenconcentration. Agric. Biol. Chem. 50:1009-1014.

25. Tajima, K., H. Sasahara, H. Kouchi, T. Yoneyama, and J.Ishizuka. 1985. Effects of oxygen concentration andleghemoglobin in organic acid degradation by isolated soybeannodule bacteroids. Agric. Biol. Chem. 49:3473-3479.

26. Trinchant, J. C., A. M. Birot, and J. Rigaud. 1981. Oxygensupply and energy-yielding substrates for nitrogen fixation(acetylene reduction) by bacteroid preparations. J. Gen. Micro-biol. 125:159-165.

27. Tubb, R. S. 1976. Regulation of nitrogen fixation in Rhizobiumsp. Appl. Environ. Microbiol. 32:483-488.

28. Tuzimurg, K., and H. Meguro. 1960. Respiration substrate ofrhizobium in the nodules. J. Biochem. 47:391-397.

29. Waters, J. K., D. B. Karr, and D. W. Emerich. 1985. Malatedehydrogehase from Rhizobium japonicum 31lb-143 bacteroidsand Glycine max root-nodule mitochondria. Biochemistry 24:6479-6486.

VOL. 169, 1987

on Septem

ber 30, 2017 by guesthttp://jb.asm

.org/D

ownloaded from