Glutamate Biosynthesis in Bacillus axotofixans 15 N NMR AND ENZYMATIC STUDIES

Pathways of ammonia assimilation into glutamic acid in Bacillus azotofixans, a recently characterized nitrogen-fixing species of Bacillus, were investigated through observation by NMR spectroscopy of in vivo incorporation of “N into glutamine and glutamic acid in the absence and presence of inhibitors of ammoniaassimilating enzymes, in combination with measurements of the specific activities of glutamate dehydrogenase, glutamine synthetase, glutamate synthase, and alanine dehydrogenase. In ammonia-grown cells, both the glutamine synthetase/glutamate synthase and the glutamate dehydrogenase pathways contribute to the assimilation of ammonia into glutamic acid. In nitrategrown and nitrogen-fixing cells, the glutamine synthetase/glutamate synthase pathway was found to be predominant. NADPH-dependent glutamate dehydrogenase activity was detectable at low levels only in ammonia-grown and glutamate-grown cells. Thus, B. azotofixans differs from Bacillus polymyxa and Bacillus macerans, but resembles other N2-fixing prokaryotes studied previously, as to the pathway of ammonia assimilation during ammonia limitation. Implications of the results for an emerging pattern of ammonia assimilation by alternative pathways among nitrogenfixing prokaryotes are discussed, as well as the utility of ’‘N NMR for measuring in vivo glutamate synthase activity in the cell.

Pathways of ammonia assimilation into glutamic acid in Bacillus azotofixans, a recently characterized nitrogen-fixing species of Bacillus, were investigated through observation by NMR spectroscopy of in vivo incorporation of "N into glutamine and glutamic acid in the absence and presence of inhibitors of ammoniaassimilating enzymes, in combination with measurements of the specific activities of glutamate dehydrogenase, glutamine synthetase, glutamate synthase, and alanine dehydrogenase. In ammonia-grown cells, both the glutamine synthetase/glutamate synthase and the glutamate dehydrogenase pathways contribute to the assimilation of ammonia into glutamic acid. In nitrategrown and nitrogen-fixing cells, the glutamine synthetase/glutamate synthase pathway was found to be predominant. NADPH-dependent glutamate dehydrogenase activity was detectable at low levels only in ammonia-grown and glutamate-grown cells. Thus, B. azotofixans differs from Bacillus polymyxa and Bacillus macerans, but resembles other N2-fixing prokaryotes studied previously, as to the pathway of ammonia assimilation during ammonia limitation. Implications of the results for an emerging pattern of ammonia assimilation by alternative pathways among nitrogenfixing prokaryotes are discussed, as well as the utility of ''N NMR for measuring in vivo glutamate synthase activity in the cell.
Our recent studies of the pathways of ammonia assimilation into glutamic acid in N,-fixing Bacillus showed that the glutamate dehydrogenase (GDH) pathway: NH; + a-ketoglutarate + N.AD(P)H + H' e L-glutamate + NAD(P)' + H,O is the predominant pathway in Nz-fixing cells of Bacillus polymyxa (1) and a major pathway in those of Bacillus macera m (2). This is in marked contrast to all other NP-fixing prokaryotes studied previously which have been shown to assimilate ammonia by the glutamine synthetase (GS)/glutamate synthase (GOGAT) pathway: GDH NH, + L-glutamate + ATP "* L-glutamine + ADP + Pi GS * This work was supported by National Science Foundation Grant DMB85-01617 and is Contribution 7645 from the Gates and Crellin Laboratories of the California Institute of Technology. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "oduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed Dept. of Chemistry and Biochemistry, University of California, 405 Hilgard Ave. Los Angeles, CA 90024. a-ketoglutarate + L-glutamine + NAD(P)H + H+ -2 L-glutamate + NAD(P)+ The coupled pathway is efficient for assimilating ammonia at low concentrations by virtue of the low K,,, of glutamine synthetase for ammonia (3). B. polymyxa and B. macerans have glutamate dehydrogenases with K , for NH: of 2.9 and 2.2 mM, respectively, whereas all other N,-fixing prokaryotes studied previously (4) have either a glutamate dehydrogenase with unusually high K , (>11 mM) for NH: or barely detectable levels of glutamate dehydrogenase even in ammonia-rich media (see Ref. 2). These findings raised the possibility that, for prokaryotes having a glutamate dehydrogenase with K , for NH: in the common range of 1-5 mM, the glutamate dehydrogenase pathway which does not consume ATP may be more advantageous than the ATP-requiring glutamine synthetase/glutamate synthase pathway during the energydemanding process of N P fixation, particularly for anaerobic NP fixers that must generate ATP through the inefficient process of fermentation. By contrast, N,-fixing prokaryotes lacking glutamate dehydrogenase or having a glutamate dehydrogenase with a very high K , for NH: must by necessity assimilate ammonia by the glutamine synthetase/glutamate synthase pathway.
Recently, a new N,-fixing Bacillus species, B. azotofixans, has been isolated from Brazilian soil and characterized (5). B. azotofixans is identical with the Bacillus species that had been isolated by Hino and Wilson (6) and tentatively classified as the Hino strain of B. polymyxa although it differed from other B. polymyxa strains in its inability to ferment lactose, arabinose, and glycerol. B. azotofixans and the Hino strain have now been conclusively shown to be a separate species from B. polymyxa and to grow on Nz much more efficiently than B. polymyxa or B. macerans (5). This raises the possibility that the mode of assimilation of ammonia derived from N, may be quite different in B. azotofixam and merits investigation in view of the unusual characteristics observed for the other two N,-fixing Bacillus species.
15N nuclear magnetic resonance (NMR) spectroscopy is useful for determining whether 15NH: is assimilated into glutamic acid by the glutamate dehydrogenase or the glutamine synthetase/glutamate synthase pathway through observation of time-dependent assimilation of 15N into y-N of glutamine and glutamic acid N in cells incubated with 15Nlabeled precursor (1,7). It also permits measurement of in vivo rates of biosynthesis of these amino acids. This paper reports a study of the pathways of ammonia assimilation in ammonia-, nitrate-, and N2-grown B. azotofixans by 15N NMR in combination with measurements of the specific activities of ammonia-assimilating enzymes. 13 Klett/ml for N2-fixing cells), washing with the specified buffers and disrupting the cells by sonication as described previously (1). All enzyme assays were performed at 21 ? 1 "C within 1 h of harvesting the cells. Protein was measured by the method of Lowry et al. (8).
Glutamate dehydrogenase, glutamate synthase, and alanine dehydrogenase activities were determined spectrophotometrically by modifications of the standard procedures (9-11) as described previously (1) with the following exception. The glutamate dehydrogenase assay solution contained 5 mM a-ketoglutarate, 80 mM NH,CI, and 0.3 mM NADPH in 50 mM KH2P0,. K,HPO, buffer, pH 7.8. Specific activities are reported as milliunits, i.e. nmol of NADPH (glutamate dehydrogenase and glutamate synthase) or NADH (alanine dehydrogenase) oxidized per minute, per milligram of protein. In the cell-free extracts of N2-fixing cells, where the oxidation of NADPH (at 0.3 mM concentration) by NADPH oxidase and other enzymes in the extracts was too rapid for accurate measurement of glutamate-synthase activity, the activity was assayed by measuring the rate of formation of ["N]glutamate from [y-"N]glutamine by "N NMR spectroscopy as described previously (2). NADH-dependent glutamate dehydrogenase activity in ammonia-grown cells, where the NADH oxidase activity was high, was measured by addition of cellfree extracts to an assay solution containing 50 mM KH2P04. K2HP04 buffer, pH 7.8, 5 mM a-ketoglutarate, 80 mM "NH4CI, and 17.5 mM NADH. At 11, 20, and 40 min, the reaction was terminated by withdrawing a 2-ml aliquot of the reaction mixture and acidifying to pH 2.0. The extent of formation of [''N]glutamic acid was measured by "N NMR.
Glutamine-synthetase activity was measured by a modification of the radiochemical method of Prusiner and Milner (12) as described previously (1). The K,,, value of alanine dehydrogenase for NH: was determined by the method of Lineweaver and Burk (see Ref. 13).
The intracellular NH: concentration in N2-fixing cells was determined on a duplicate culture as described previously (1). For ammonia-grown cells, the intracellular NH: concentration could not be determined because the amount of NH: trapped in the residual slime in the unwashed cell pellet was large relative to intracellular NH:.
Chemicak-"'NH4C1 (99% "N) and K"N0, (98% "N) were purchased from Cambridge Isotope Laboratories, and ~-[y-''N]glutamine (95% "N) from MSD Isotopes. All other chemicals were reagent grade. Table I shows the doubling times of B. azotofixans in various nitrogen and carbon sources. B. azotofixans, when grown with D-glUCOSe as the carbon source, produced heavy slime consisting of viscous extracellular polysaccharides which are difficult to separate from the cells. Slime production was effectively reduced when D-mannitol was used as a carbon source (14). B. azotofixans grew with very similar doubling times when D-mannitol was substituted for D-glucose in am- monia-grown cells or for sucrose in N,-fixing cells (Table I).

Growth-
Thus, its growth is not limited when D-mannitol is the carbon source. Substitution of D-mannitol for D-glucose was found to have no significant effect on the specific activities of glutamate dehydrogenase in ammonia-grown B. polymyxa ATCC 8519 (990 milliunits.mg" protein (mannitol) uersus 798 milliunits. mg" protein (glucose)) or in B. macerans ATCC 8515 (195 milliunits . mg" protein (mannitol) uersus 276 milliunits .
mg" protein (glucose)). Thus, it is reasonable to assume that the substitution of D-mannitOl for D-glucose has no effect on the pathways of ammonia assimilation in the Bacillus species. D-mannitol was used as the carbon source throughout this study.
In ammonia-grown cells, the initial NH: concentration in the medium was 22 mM; increasing the concentration to 60 or 100 mM slowed the growth to doubling times of 4.6 and 6.6 h, respectively, a t 30 "C. B. azotofixans, although characterized as nitrate-reductase negative (5) on the complex media of Gordon et al. (15), was found to grow aerobically with nitrate as the sole nitrogen source as described earlier for the Hino strain (6). The intracellular NH: concentration in Nzfixing cells was found to be 0.4 f 0.18 mM. Although the intracellular NH: concentration in nitrate-and ammoniagrown B. azotofixans could not be determined due to residual slime in the cell pellet, it is reasonable to assume that the cells are growing under ammonia-limited and ammonia-rich conditions, respectively, as was found for B. polymyxa (1).  medium. The I5N peaks were assigned on the basis of previous work (1,16). The I5NO;, taken up by the cell, is converted by nitrate reductase to 15NO; which, in turn, is reduced by nitrite reductase to 15NH:. After 3 min, as a result of 15NH: assimilation, 54-68% of the glutamine pool was "N-labeled in the 7-N (peak at 263.6 ppm) whereas only 6% of the glutamic acid was "N-labeled relative to the ["N]glutamic acid pool observed after 30 min. The observed f 1 2 % variation in the [y-15N]glutamine peak intensities between the 18-and 30min spectra probably arises from experimental error in the extraction of I5N metabolites. The average of the two peak intensities was taken to represent fully I5N-labeled glutamine y N , because in a similar experiment performed at 30 "C (instead of 21 f 1 "C) where metabolic processes are expected to proceed approximately twice as fast, [y-'5N]glutamine peak intensities were found to show no increase between 8, 18, and 30 min. The glutamic acid pool, on the other hand, gradually becomes "N-labeled over a period of 30 min (peak at 335.06 ppm). By 18-30 min, a part of the ["N]glutamic acid pool had been recycled by the glutamine-synthetase reaction to form [a,?-"N]glutamine whose a-I5N peak (334.93 ppm) was resolved from the a-amino 15N peak of glutamic acid as shown in the expanded-scale spectrum. The extensive "N-labeling of the glutamine y-N prior to that of the glutamic acid N strongly suggests that "NH: is assimilated mainly by the glutamine synthetase/glutamate synthase pathway in nitrategrown cells. Fig. 1B shows the effect of inhibitors of glutamine synthetase and glutamate synthase on the biosynthesis of [a,y-15N] glutamine and ["N]glutamic acid in nitrate-grown cells. L-Methionine DL-sulfoximine and azaserine (a structural analog of glutamine) are irreversible inhibitors of glutamine synthetase and glutamate synthase, respectively, but have no inhibitory effect on glutamate dehydrogenase (17)(18)(19). Preincubation of cells with 22 mM DL-methionine DL-sulfoximine + 0.5 mM azaserine prior to the addition of 15NO; resulted in 93% inhibition of the biosynthesis of ['5N]glutamic acid as indicated by the decrease in its peak intensity compared to that in the control (Fig. 1B). DL-Methionine DL-sulfoximine was added at the high concentration because at 4 mM concentration (in the absence of azaserine), 15NH: incorporation into the 7-N of glutamine was inhibited by only 12%. The extensive inhibition observed in the presence of the two inhibitors clearly shows that glutamic acid is formed predominantly by the glutamine synthetase/glutamate synthase pathway in nitrate-grown cells. Fig. 2 shows the time-dependent formation of ['5N]glutamic acid (in nmol.mg" protein) calculated from the observed peak intensities in Fig. 1A. The nmoles of ['"NJglutamic acid. mg" protein formed a t 18 and 30 min were calculated from the combined peak intensities of a-amino nitrogens of glutamic acid and glutamine because the latter is recycled from the former. The rate of biosynthesis of ['5N]glutamic acid calculated from Fig. 2 is 11.7 k 0.3 nmol. min" . mg" protein ( Table I). The average rate of utilization of ["N]glutamic acid for protein synthesis during the 30-min interval, calculated from the doubling time at 21 f 1 "C (Table I) by the method described previously (1) is less than 0.2 nmol [15N]gl~tamic acid. min".mg" protein which rate is negligibly small compared with the observed rate of its biosynthesis. Because the study with the inhibitors has shown that glutamate is biosynthesized predominantly by the glutamine synthetase/glutamine synthase pathway, the rate of ['5N]glutamic acid biosynthesis, when its substrate glutamine has become fully I5N-  Table I This is understandable because the intracellular concentration of glutamine is 11 mM (Table I)  In N,-fixing cells, the glutamate dehydrogenase activity was undetectable whereas glutamine synthetase and glutamate synthase activities were 38 and 11.4 milliunits. mg-' protein, respectively. The result strongly suggests that in N,-fixing cells of B. azotofixans, ammonia is assimilated predominantly by the glutamine synthetase/glutamate synthase pathway.

Ammonia Assimilation in Ammonia-grown
Cells-The time-dependent incorporation of 15NH: into glutamine and glutamic acid in ammonia-grown cells, as observed by NMR at 3, 8, 16, and 24 min after transfer to 15NH: medium, is shown in Fig. 3. At 3 min, approximately 70% of the glutamine pool has been 15N-labeled in the r-N, whereas for glutamic acid, only 18% was 15N-labeled relative to the pool at 24 min. The extensive 15N-labeling of glutamine r -N , combined with the slightly faster "N-labeling of glutamic acid compared to that in nitrate-grown cells (Fig. lA) suggests that while ammonia may be assimilated mainly by the glutamine synthetase/glutamate synthase pathway, direct assimilation by the glutamate dehydrogenase pathway also contributes to the formation of ['5N]glutamic acid. The result rules out glutamate dehydrogenase as the predominant pathway because, in such case, the glutamic acid pool would have been rapidly saturated with 15N, prior to the glutamine pool, as observed for ammonia-grown B. polymyxa (1). The rate of biosynthesis of ['5N]glutamic acid in ammonia-grown cells was found to be 12.5 & 0.9 nmol.min-l.mg-' protein ( Fig. 2 and Table I).  Table I). The results suggest that, while assimilation occurs through the glutamine synthetase/ glutamate synthase pathway and to a lesser extent through the glutamate dehydrogenase pathway, assimilation via alanine by the alanine dehydrogenase/alanine-glutamic transaminase pathway may also occur in view of the observed high level of alanine dehydrogenase relative to other ammoniaassimilating enzymes.
To investigate the relative contributions of the three pathways to glutamate biosynthesis, the incorporation of 15NH: into glutamic acid and alanine was studied in the presence of the following inhibitors: L-methionine DL-sulfoximine (glutamine synthetase inhibitor) + azaserine (glutamate synthase inhibitor), aminooxyacetate (alanine-glutamic transaminase inhibitor (20)), and glutarate (an inhibitor of glutamate dehydrogenase (21) and, to a lesser extent, of glutamate synthase). Through in vitro assays, glutarate, a structural analog of a-ketoglutarate, added at 0.05 M concentration to the assay solution was shown to inhibit glutamate dehydrogenase by 64% and glutamic-alanine transaminase by less than 20% in B. polymyxa (l), and glutamate synthase by 36% in B. azotofixans. Fig. 4 shows 15N NMR spectra of ammonia-grown B. azotofixans incubated with 15NH: without (A) or with ( B ) 20 min of preincubation in 11 mM L-methionine DL-sulfoximine + 0.5 mM azaserine. In the presence of the inhibitors, the synthesis of ["N]glutamic acid decreased by 30-40%. The inhibition of ['5N]glutamic acid formation was calculated from the ratio of the peak intensity in the inhibited cells to that in the control cells using the combined peak intensities (in integrated areas as shown in the expanded-scale spectra) for the a-amino 15N of glutamic acid and that of glutamine because the latter is recycled from the former. The result is consistent with a substantial contribution of the glutamine  Fig. 4C and D show the effect of preincubation with 0.3 M glutarate. In the presence of the inhibitor, the synthesis of [15N]glutamic acid decreased by 59%. The result is consistent with participation of the glutamate dehydrogenase pathway to ammonia assimilation. The more extensive inhibition by glutarate than by L-methionine DL-sulfoximine + azaserine is understandable because glutarate has inhibitory effects on both glutamate dehydrogenase and glutamate synthase.
In the presence of aminooxyacetate, which inhibits the reversible glutamic-alanine transaminase as well as other transaminases (20), the biosynthesis of [15N]alanine decreased to a barely detectable level (Fig. 3E (control) versus Fig. 3F (inhibited)). This suggests that [15N]alanine is synthesized predominantly via ['5N]glutamic acid by transamination and not directly from 15NH: and pyruvate by alanine dehydrogenase. In the latter case, an increase in [15N]alanine would occur in the inhibited cells due to inhibition of transamination to [15N]glutamic acid. (The observed decrease in ["N]glutamic acid synthesis in the inhibited cells compared to the control is probably due to accumulation of unlabeled glutamic acid during preincubation with aminooxyacetate which slowed down the glutamate dehydrogenase-catalyzed assimilation of 15NH: and channeled it into the glutaminesynthetase pathway.) Further evidence supporting this pathway for alanine biosynthesis is the 53% decrease in [15N] alanine biosynthesis, comparable to the 59% decrease in [15N] glutamic acid biosynthesis, observed in cells incubated with glutarate (Fig. 3, C and D , insets). Thus, the alanine dehydrogenase/alanine-glutamic transaminase pathway makes little, if any, contribution to 15NH: assimilation into ['5N]glutamic acid, despite the observed high level of alanine dehydrogenase. Alanine dehydrogenase was found to have an apparent K , of 7.6-18 mM for NH:; a more precise determination was precluded by nonlinearity of the Lineweaver-Burk plot of l/uo versus l/[NH:]. The low affinity for NH: probably precludes participation of alanine dehydrogenase in ammonia assimilation. The 18-fold induction of alanine dehydrogenase observed in alanine-grown cells ( Table I) suggests that its physiological function is mainly catabolic.  Fig. 40, inset) increased substantially compared to the control. The cause of this increase is unknown at present, but direct incorporation of 15NH: into aspartic acid by the aspartate dehydrogenase-catalyzed reaction with oxaloacetate or by the aspartase-catalyzed reaction with fumarate is unlikely because (i) the occurrence of aspartate dehydrogenase has not been confirmed in bacteria (22), (ii) succinate dehydrogenase, which catalyzes the synthesis of fumarate, is lacking in the Hino strain (23), and (iii) in the cells incubated with aminooxyacetate, an accumulation of [15N]aspartic acid due to inhibition of transamination to [15N] glutamic acid which is expected to occur if either one of the direct pathways were operative, is not observed (Fig. 4, E and

F).
Properties and Regulation of the Enzymes-Coenzyme specificities of glutamate dehydrogenase and glutamate synthase were investigated with the following results. Only NADPHdependent glutamate dehydrogenase activities were detected in B. azotofixans. No NADH-dependent glutamate dehydrogenase could be detected in ammonia-grown cells by spectrophotometric or NMR methods ("Experimental Procedures") or in glutamate-grown cells in which NAD+-dependent glutamate dehydrogenase, if present, is expected to be maximally induced. The coenzyme specificity of glutamate synthase was investigated through observation of the time-dependent formation of [15N]glutamic acid from [y-'5N]glutamine with NADPH ( Fig. 5A) or NADH (Fig. 5B) on addition of cell-free extracts of Np-fixing B. azotofixans. NADPH-and NADHdependent activities of 11.4 and 7.8 2 3.4 milliunits.mg-' protein, respectively, were observed. The results suggest that the glutamate synthase of B. azotofixans can utilize NADPH or NADH. In this respect, the glutamate synthase of B. azotofixans appears to resemble that of B. subtilis PC1 219 which has been shown to utilize NADPH or NADH with relative activities of 100 and 21, respectively (24).

DISCUSSION
The results show that, in B. azotofixans, the glutamine synthetase/glutamate synthase pathway is the predominant pathway of ammonia assimilation in ammonia-limited cells and a major pathway in ammonia-grown cells. The NADPHdependent glutamate dehydrogenase activity is detectable at very low levels only in ammonia-and glutamate-grown cells. This suggests that the enzyme has limited physiological function, to participate in assimilating ammonia in the early and midexponential phase of growth when the NH: concentration in the medium is relatively high (14-22 mM), and in oxidative deamination of glutamate to provide NH: when glutamate is the sole nitrogen source. Whether B. azotofixans is incapable of synthesizing high levels of glutamate dehydrogenase, or does not do so because a high K,,, for NH: limits its utility, cannot be determined at present because its low activity in cell-free extracts precludes measurement of its K,,,, and the cells grow poorly when the medium NH: concentration is very high (60-100 mM). Thus, B. azotofixans differs from B. polymyxa and B. macerans, but resembles other prokaryotes ' The abbreviation used is: Hepes, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid. studied previously in the pathway of ammonia assimilation during ammonia limitation. It is significant that Nz-fixing prokaryotes studied to date fall into two groups with respect to the pathway of ammonia assimilation. Thus, they are: (i) B. polymyxa and B. macerans which have glutamate dehydrogenases with moderate affinity for NH: and utilize the glutamate dehydrogenase pathway during Nz fixation (1,2); and (ii) B. azotofixans and other prokaryotes, such as Clostridium pasteurianum and Klebsiella pneumoniae (4), which utilize the glutamine synthetase/glutamate synthase pathway because they have either barely detectable levels of glutamate dehydrogenase even in ammonia-rich medium, or else glutamate dehydrogenase with an unusually high K,,, for NH: (see Ref. 2). The question raised by our studies on B. polymyxa and B. macerans (1, Z)-whether, among Nz-fixing prokaryotes that are capable of synthesizing glutamate dehydrogenase with moderate affinity for NH:, the glutamate dehydrogenase pathway is more advantageous than the ATP-requiring glutamine synthetase/glutamate synthase pathway for assimilating ammonia during the energy-demanding process of nitrogen fixation-is an interesting question that requires further investigation among N,-fixing prokaryotes that possess both pathways.
It is interesting that, whereas the glutamine synthetase activity is derepressed in the ammonia-limited cells, the glutamate synthase activity shows little variation except in glutamate-grown cells where it is repressed. Such apparent lack of derepression of glutamate synthase has been observed in many microorganisms (25). In B. azotofixans which has a high intracellular concentration of glutamine (Table I), derepression of glutamate synthase may not be necessary if the enzyme has very high affinities for substrates. Purified glutamate synthases from other Bacillus species have K,,, values of 0.1-0.18 mM for glutamine, 0.05-0.09 mM for a-ketoglutarate, and 0.007 mM for NADPH (24,26,27), whereas purified glutamine synthetases from Bacillus species are remarkably similar in having K,,, values of 0.3-0.4 mM for NH:, 0.8-3.6 mM for glutamic acid and 0.2-0.9 mM for Mn. ATP (28-30). Thus, while derepression of glutamine synthetase may be necessary to optimize the assimilation of low concentrations of NH: into glutamine in nitrate-and Nz-grown cells, the utilization of the glutamine nitrogen for glutamic acid biosynthesis may require only a basal level of glutamate synthase because of the high affinities for substrates and the observed high intracellular concentration of glutamine.
The reliability of the 15N NMR method for distinguishing between the glutamate dehydrogenase and the glutamine synthetase/glutamate synthase pathways is clearly demonstrated by the contrasting kinetic patterns of 15N incorporation into glutamine 7-N and glutamic acid N observed in nitrate-grown B. azotofixans (Fig. 1) and B. polymyxa (1). The direct in vivo method should prove useful for determining the pathway in those microorganisms for which in vitro enzyme assays are difficult because of enzyme instability, high background NAD(P)H oxidation, or the heavy slime production observed in many free-living Nn-fixing prokaryotes (31). For organisms in which glutamic acid is formed predominantly by the glutamine synthetase/glutamate synthase pathway, the rate of ['5N]glutamic acid biosynthesis observed by NMR represents the in vivo glutamate synthase activity. The in vivo activity measurement should be particularly useful for ferredoxindependent glutamate synthase in algae and plants, whose in vitro measurement is quite difficult because it requires the isolation of species-specific ferredoxin, as well as separation of the product glutamate (32). helpful discussion and careful reading of the manuscript.