J. Biochem., 82, 95-103 (1977)

Pathway for Isoleucine Formation from Pyruvate by Leucine Biosynthetic Enzymes in Leucine-Accumulating Isoleucine Revertants of Serratia marcescens

Masahiko KISUMI, Saburo KOMATSUBARA, and Ichiro CHIBATA

Research Laboratory of Applied Biochemistry, Tanabe Seiyaku Co., Ltd., Kashima, Yodogawa-ku, Osaka, Osaka 532

Received for publication, January 19, 1977

Leaky revertants isolated from isoleucine auxotrophs of Serratia marcescens mutant resistant to a-aminobutyric acid were previously reported to accumulate leucine in the medium, due to the absence of both feedback inhibition and repression of leucine biosynthesis. Growth of the revertant was accelerated by pyruvate, n(-)-citramalate, citraconate, and a-ketobutyrate, but not by threonine. Extracts of the revertant exhibited high activities of pyruvate-dependent coenzyme A liberation from acetyl-coenzyme A, hydration of citraconate, and conversion of citraconate to a-ketobutyrate, but showed no threonine-deaminating activity. In the leucine accumulating revertants the above three activities were not affected by leucine, but in the wild strain and other revertants accumulating no leucine all or one of these activities was controlled by leucine. A leucine auxotroph isolated from the leucine-accumulating revertant showed isoleucine auxotrophy as well. From these data, it is concluded that, in leucine-accumulating revertants of S. marcescens, isoleucine is synthesized from a-ketobutyrate via citramalate formed from pyruvate and acetyl-coenzyme A by leucine biosynthetic enzymes, as a result of desensitization of a-isopropylmalate synthetase to feedback inhibition.

Generally isoleucine is synthesized from aspartate viously isolated leaky revertants from isoleucine via threonine in microorganisms. L-Threonine auxotrophs of S. marcescens mutants resistant to dehydratase [EC 4.2.1.16] is the rate-limiting enzyme a-aminobutyric acid. Most of the isoleucine re for isoleucine biosynthesis. Nevertheless, some vertants accumulated large amounts of leucine in microorganisms are known to form isoleucine from the medium (6). Leucine accumulation was precursors other than threonine by other pathways. ascribed to both constitutive synthesis of leucine We have been studying biosynthetic regulation biosynthetic enzymes and desensitization of a and accumulation of branched-chain amino acids isopropylmalate synthetase [EC 4.1.3.12], the first using Serratia marcescens mutants (1-6). In S. enzyme of leucine biosynthesis, to feedback inhibi marcescens as well as other enteric bacteria, iso tion. The constitutive synthesis and desensitiza leucine is synthesized from threonine by five iso tion resulted from a-aminobutyric acid resistance leucine-valine biosynthetic enzymes. We pre and reversion of isoleucine auxotrophy, respectively.

Vol. 82, No. 1, 1977 95 96 M. KISUMI, S. KOMATSUBARA, and I. CHIBATA

Interestingly, the partial reversion of isoleucine Enzyme Assays-The bacteria were cultured auxotrophy in leucine-accumulating revertants did in minimal medium under the conditions described not depend on the restoration of L-threonine de in the previous paper (4). The cells were harvested hydratase activity. Evidence for the absence of the from exponentially growing culture by centrifuga activity in vivo will be given in a separate paper tion and extracts were prepared by the method (in preparation). It is also suggested that a single described previously (4, 6). Protein was de mutational event influences both isoleucine auxo termined as described previously (4). Pyruvate trophy and the sensitivity of a-isopropylmalate dependent coenzyme A liberation from acetyl synthetase to feedback inhibition. It appears that coenzyme A was measured under the conditions these revertants might not synthesize isoleucine via described previously (6) using pyruvate instead of threonine; i.e., might not require the participation a-ketoisovalerate as a substrate. To assay citra of L-threonine dehydratase. conate hydration, the method for a-isopropyl This report describes a pathway for isoleucine malate isomerase [EC 4.2.1.33] (8) was modified as formation from pyruvate and acetyl-coenzyme A, follows. The reaction mixture contained citra and suppression of isoleucine auxotrophy by conate (1.5 pmol), potassium phosphate buffer (pH desensitization of the first enzyme of leucine bio 7.0, 300 pmol), and cell-free extracts (0.3-0.5 mg of synthesis. protein) in a total volume of 3.0 ml. Incubation was carried out at room temperature in a cuvette

MATERIALS AND METHODS with 1 cm light path. The rate of decrease in ab sorbance was followed with a Varian Techtron

Bacteria-S. marcescens strain I and its mu spectrophotometer (model 635). a-Ketobutyrate tants listed in Table I were used. formation from citraconate was assayed by a mo Isolation of Revertants from Strain 1656-Cells dification of the (ƒÀ-isopropylmalate dehydrogenase of strain 1656 were spread on agar plates of Davis [EC 1.1.1] (9) as follows. The reaction mixture Mingioli minimal medium (7) modified by omitting contained citraconate (20 ƒÊmol), KCl (50 ƒÊmol), citrate and increasing glucose to 0.5 %. Incubation MgCl2 (10 ƒÊmol), nicotinamide dinucleotide (5 was carried out at 30°C for 3-5 days. Colonies ƒÊ mol), potassium phosphate buffer (pH 8.0, 100 were selected and purified by single colony isolaƒÊƒÊ mol), and cell-free extracts (0.5-1.0 mg of protein) tion. in a total volume of 1.0 ml. Incubation was carried Growth Experiments-Growth experiments out at 30•Ž and stopped by the addition of 10 were performed in minimal medium as described trichloroacetic acid. a-Ketobutyrate formed was previously (3); the medium was inoculated with 106 determined by the method of Friedemann and cells per ml. Haugen (10). Bioassay of a-ketobutyrate was also

TABLE I. Strains used.

a Nitrosoguanidine. b a-Aminobutyric acid-resistant.

J. Biochem. PATHWAY FOR ISOLEUCINE FORMATION FROM PYRUVATE 97 performed using strain 621 (L-threonine dehydrat Moreover, a-ketobutyrate-forming activities from ase-deficient strain which exhibits growth response these amino acids were very low in the revertants to either a-ketobutyrate or isoleucine, but not to as well as in the wild strain. citraconate or other isoleucine precursors). Speci Charon et al. presented indirect evidence for fic activities were expressed as Icmol of products isoleucine biosynthesis from pyruvate in aerobic formed or substrate reacted per mg of protein per spirochetes (15). Sai et al. reported that citra min. malate accumulated in the medium is formed from Chemicals-D(-) and L(-}-)-Citramalates were pyruvate in yeast mutants (16). Sasaki et al. stated obtained from Sigma Chemical Co. (St. Louis, Mo). that citraconate is degraded to a-ketobutyrate by Unless otherwise noted, the n(-)-form was used. extracts of a Bacillus strain grown on n-methyl malate as a sole carbon source (17). We found

RESULTS that various a-keto acids are converted to a-keto acids possessing one more carbon atom and to Effects of Various Compounds on the Growth of analogs of branched-chain amino acids by S. Strains 149, 149-V17, and S-11-To identify the marcescens mutants (18). Therefore, we tested isoleucine biosynthetic pathway in leucine-accumu whether isoleucine is formed from pyruvate via lating revertants, growth experiments were carried citramalate and citraconate. out. Several pathways for isoleucine formation Pyruvate accelerated the growth of strain S-11, are known in aerobic microorganisms (Fig. 1). a representative leucine-accumulating revertant, in Phillips et alL suggested a pathway from glutamate which a-isopropylmalate synthetase is insensitive via ƒÀ-methylaspartate in Escherichia coli (11, 12). to feedback inhibition by leucine (Table II). On The growth of revertants and isoleucine auxotrophs the other hand, pyruvate had no effect on the of S. marcescens was stimulated by (ƒÀ-methylas growth of strains 149 and 149-V 17, in which the

partate, but not by glutamate. Glutamate mutase synthetases are sensitive. n(-)-Citramalate and [EC 5.4.99.1] activity, forming ,B-methylaspartate, citraconate had stimulating effects on the growth was not found in the cell-free extracts by the method of these three strains, though at high concentrations. of Phillips et al. (11). Therefore, it was concluded L(+)-Citramalate and mesaconate, stereoisomers that S. marcescens mutants lack the glutamate route of the two compounds, had no effect on growth. for isoleucine formation. These data strongly suggest that, in leucine-accumu Homoserine is directly converted to a-keto lating revertants, isoleucine is formed from pyruvate butyrate by extracts of Neurospora crassa (13). by the leucine biosynthetic enzymes via citramalate Homocysteine and methionine may also be de and citraconate as intermediates of a-ketobutyrate graded to a-ketobutyrate in some bacteria (14). formation. These amino acids did not stimulate the growth of Pyruvate-Dependent Coenzyme A Liberation leaky isoleucine revertants of S. marcescens. from Acetyl-Coenzyme A and Its Control by Leucine

Fig. 1. Possible pathways of a-ketobutyrate and isoleucine formation in microorganisms.

Vol. 82, No. 1, 1977 98 M. KISUMI, S. KOMATSUBARA, and I. CHIBATA

TABLE II. Effects of various compounds on the growth of strains 149, S-11, and 149-Vl7a.

a The medium was inoculated with 106 cells per ml (optical density, 0.002). b a-Ketobutyrate and the other com

pounds were added to the minimal medium at concentrations of 10-1 M and 5 •~ 10-1 M, respectively.

-To identify any alternative isoleucine pathway, tase, we examined whether the liberation is con enzymatic studies were performed. It was con trolled by leucine or not (Table III). The coenzyme sidered that citramalate might be formed from A-liberating activities of strains 149 and S-11, which pyruvate and acetyl-coenzyme A by a-isopropyl are derepressed for a-isopropylmalate synthetase, malate synthetase. Therefore, the pyruvate-de were high with or without the addition of leucine pendent liberation of coenzyme A was examined by to minimal medium, while that of the wild strain the method of Kohlhaw et al. (19). As expected, was low in the minimal medium and decreased on marked liberation of coenzyme A was observed addition of leucine. The activity of strain S-11, in with extracts of strain S-11 when pyruvate was which a-isopropylmalate synthetase is desensitized, added to the reaction mixture, indicating citramalate was not inhibited by the addition of leucine to the formation from pyruvate and acetyl-coenzyme A reaction mixture. On the other hand, the activity (Fig. 2). The substrate specificity of a-isopropyl of strain 149, in which the synthetase is feedback malate synthetase of S. marcescens and biosynthesis sensitive, was inhibited by leucine. From these of several end-products, analogs of branched-chain results, it was concluded that pyruvate-dependent amino acids, are described in a separate paper (20). coenzyme A liberation from acetyl-coenzyme A, To verify that pyruvate-dependent coenzyme A that is, citramalate formation from pyruvate, is liberation is catalyzed by a-isopropylmalate synthe catalyzed by a-isopropylmalate synthetase.

J. Biochem. PATHWAY FOR ISOLEUCINE FORMATION FROM PYRUVATE 99

Fig. 2. Pyruvate-dependent coenzyme A liberation from acetyl-coenzyme A by extracts of strain S-11. The reaction mixture contained acetyl-coenzyme A (0.2 temol), KCI (20 ƒÊmol), Tris-HCI buffer (pH 8.5, 50 Fig. 3. Hydration of citraconate by extracts of strain ƒÊ mol), pyruvate (2.5 ƒÊmol), and extracts (0.34 mg of S-11. The reaction mixture contained citraconate or protein) in a total volume of 0.25 ml. Incubation was mesaconate (1.5 pmol), potassium phosphate buffer (pH carried out at 37•Ž. 7.0, 300 temol), and extracts of strain S-11 (0.36 mg of protein) in a total volume of 3.0 ml. Incubation was TABLE III. Control of pyruvate-dependent coenzyme carried out at room temperature in a cuvette with 1 cm A liberation from acetyl-coenzyme A by leucine in light path. strains 1, 149, and S-11.

conate as a substrate (Fig. 3). Extracts of strain S-11 caused a rapid decrease of absorbance at 235

nm of citraconate. On the other hand, the ab

sorbance of mesaconate, a stereoisomer of citra conate, which might be formed from L(+)-citra

malate, was not changed. These data suggest that

n(-)-citramalate is isomerized to erythro j3-methyl

malate via citraconate.

a-Ketobutyrate Formation from Citraconate

In S. marcescens ƒÀ-methylmalate converted from citraconate may be changed to ƒ¿-ketobutyrate by a L-Isoleucine and L-leucine were added at a concentra the third leucine biosynthetic enzyme, ƒÀ-isopropyl tion of 10-8 M. malate dehydrogenase. Therefore, ƒ¿-ketobutyrate

formation from citraconate was examined (Fig. 4).

Hydration of Citraconate-The second enzyme Marked formation of a-keto acid was observed of leucine biosynthesis is isopropylmalate iso with extracts of strain S-11 by the 2, 4-dinitro

merase, which catalyzes the interconversion of s phenylhydrazone method (10). This keto acid was and (3-isopropylmalates via dimethylcitraconate. identified as ƒ¿-ketobutyrate by paper chromatog This activity is generally determined by measuring raphy of the hydrazone and by bioassay of the re the decrease of absorbance at 235 nm caused by action mixture using strain 621, which shows a hydration of the substrate, dimethylcitraconate (8). growth response to a-ketobutyrate. This reaction Therefore, S. marcescens was tested using citra required nicotinamide dinucleotide as a coenzyme,

Vol. 82, No. 1, 1977 100 M. KISUMI, S. KOMATSUBARA, and I. CHIBATA

synthetic enzymes are constitutive, exhibited higher activities with or without the addition of leucine. These results also suggest that citraconate hydra tion, that is, citramalate isomerization and a-keto butyrate formation, are catalyzed by leucine bio synthetic enzymes. Properties of Strain 1656, Requiring Isoleucine and Leucine-To verify genetically that isoleucine is formed by leucine biosynthetic enzymes in leucine accumulating revertants, a mutant requiring leucine was isolated from strain S-11 using minimal agar plates containing isoleucine. Isoleucine was added because a leucine auxotroph derived from the re Fig. 4. ƒ¿-Ketobutyrate formation from citraconate by vertants was considered to exhibit isoleucine auxo extracts of strain S-11. The reaction mixture contained trophy as well. As expected, the isolated strain citraconate or mesaconate (20 ƒÊmol), KCI (50 ƒÊmol), 1656 showed both auxotrophies. Leucine plus a MnSO4 (1 pmol), nicotinamide dinucleotide (5 ƒÊmol), ketobutyrate supported the growth of this auxo potassium phosphate buffer (100 ƒÊmol, pH 8.0), and troph, indicating that isoleucine biosynthetic extracts of strain S-11 (1 mg of protein) in a total volume enzymes other than L-threonine dehydratase are of 1.0 ml. Incubation was carried out at 37•Ž. not deficient in this auxotroph. On the other hand, no growth was observed on the addition of both TABLE IV. Control of citraconate hydration and a leucine and citraconate, and the activity of a-keto ketobutyrate formation from citraconate by leucine in butyrate formation from citraconate was very low strains 1, 149, and S-11. as compared with that of strain S-11. pontaneous prototrophic revertants were isolated from strain 1656 to determine whether a single mutation might have caused both auxotro phies. Revertants isolated in the presence of either isoleucine or leucine were tested for isoleucine and leucine requirements. As shown in Table V, the reversion of isoleucine auxotrophy was accom panied by loss of leucine auxotrophy and the re version of leucine auxotrophy was accompanied by loss of isoleucine auxotrophy. Accordingly, both

a L-Isoleucine and L-leucine were added at a concentra TABLE V. Simultaneous reversion of isoleucine and tion of 10-3 M. leucine auxotrophies in strain 1656.

as does that of ƒÀ-isopropylmalate dehydrogenase

(9). Mesaconate was not converted to a-keto butyrate under the same conditions. Control of Activities of Citraconate Hydration and a-Ketobutyrate Formation-To examine whe ther citramalate isomerization and a-ketobutyrate formation from citraconate are catalyzed by leucine biosynthetic enzymes, these activities were com pared in various strains (Table IV). The wild a The value is uncertain because the leucine auxotrophs strain showed lower activities in the presence of grew by utilizing leucine formed by the revertants. leucine. Strains 149 and S-11, whose leucine bio b Number of colonies found/number of colonies tested .

J. Biochem. PATHWAY FOR ISOLEUCINE FORMATION FROM PYRUVATE 101

auxotrophies in strain 1656 are considered to be due to a single mutation.

Reversion of Leucine-Accumulating Activity to

Both Valine-Accumulating Activity and Isoleucine

Auxotrophy-Leucine-accumulating activity in S.

marcescens mutants resistant to ƒ¿-aminobutyric acid reflects desensitization of ƒ¿-isopropylmalate

synthetase to feedback inhibition (6). Isoleucine

auxotrophs were spontaneously isolated from strain

S-11 at high frequency. These strains grew in the

presence of either isoleucine or ƒ¿-ketobutyrate, and accumulated valine but not leucine, as did strain L

149, indicating the reversion of desensitized a

isopropylmalate synthetase to the sensitive enzyme.

These results also indirectly confirm that isoleucine synthesis in strain S-11 depends on leucine bio

synthetic enzymes.

DISCUSSION

In leucine-accumulating revertants derived from Fig. 5. Pathway for isoleucine formation from pyru isoleucine auxotrophs of S. marcescens mutants vate in leucine-accumulating isoleucine revertants. Kbu, resistant to a-aminobutyric acid, a-ketobutyrate a-ketobutyrate; KIV, a-ketoisovalerate; KIC, a-keto (which is converted to isoleucine) was found to be isocaproate; IPM, isopropylmalate; DMC, dimethyl formed from pyruvate via citramalate, based on the citraconate. following evidence (Fig. 5): 1. Absence of L-threo

nine dehydratase activity. 2. Growth stimulation dependent coenzyme A liberation from acetyl-co

by pyruvate, citramalate, and citraconate. 3. enzyme A and a-ketobutyrate formation from Pyruvate-dependent coenzyme A liberation from citraconate by cell-free extracts. 2. Leucine auxo

acetyl-coenzyme A, suggesting citramalate forma trophy accompanied by isoleucine requirement in

tion. 4. Hydration of citraconate, suggesting iso a leucine-accumulating revertant. 3. Leaky rever

merization of citramalate of S-methylmalate. 5. sion of isoleucine auxotrophy accompanied by

a-Ketobutyrate formation from citraconate, sug desensitization of a-isopropylmalate synthetase. Thus, the features of a-ketobutyrate formation gesting conversion of p-methylmalate to a-keto butyrate. from pyruvate are consistent with those of leucine

The reactions in the leucine-specific pathway formation. Therefore, the participation of the

are analogous to those in the Krebs cycle forming other pathways may be excluded. ƒ¿-ketoglutarate from oxaloacetate and acetyl In yeast mutants, citramalate was reported to

coenzyme A (21). Other analogous reactions are be formed from pyruvate by a-isopropylmalate

those of oxaloglutarate formation from ƒ¿-keto synthetase (16). In Salmonella typhimurium, a isopropylmalate synthetase was also found to have glutarate and acetyl-coenzyme A in lysine synthesis by yeast and fungi (22). These pathways involve an affinity for pyruvate (19). It is of interest to

condensing enzyme, isomerase, and dehydrogenase determine whether citramalate can be converted to in this order. It is possible that, in leucine isoleucine by leucine biosynthetic enzymes in these

accumulating revertants of S. marcescens, a-keto strains. Km values of yeast and Salmonella synthe

butyrate might be synthesized by a pathway similar tases for pyruvate are over 100 times those for a

to the leucine biosynthetic pathway. The evidence ketoisovalerate, a substrate for leucine synthesis. Therefore, their citramalate formation may be for participation of leucine biosynthetic enzymes in inhibited competitively by ƒ¿-ketoisovalerate, pre a-ketobutyrate formation from pyruvate is as follows: 1. Controls by leucine of pyruvate venting the formation of ƒ¿-ketobutyrate necessary

Vol. 82, No. 1, 1977 102 M. KISUMI, S. KOMATSUBARA, and I. CHIBATA for their growth. In contrast with the above O,We are grateful to Mr. T. Takayanagi, former Senior strains, S. marcescens has an ƒ¿-isopropylmalate Manager of the Research and Development Division of synthetase with a Km value for pyruvate only 5 this company for his encouragement during the course of this investigation. We also thank Mr. F. Murakami times higher than that for ƒ¿-ketoisovalerate (20). and Mrs. M. Ashibe for technical assistance. For this reason, citramalate formation from pyruvate is considered to be caused by genetic loss of feedback controls of leucine biosynthesis in S. REFERENCES marcescens. High affinity of ƒ¿-isopropylmalate synthetase for pyruvate is also the primary cause of 1. Kisumi, M. (1962) J. Biochem. 52, 390-399 2. Kisumi, M., Komatsubara, S., & Chibata, I. (1971) the suppression of isoleucine auxotrophy resulting J. Bacteriol. 107, 824-827 from desensitization of the first enzyme of leucine 3. Kisumi, M., Komatsubara, S., & Chibata, 1. (1971) biosynthesis to feedback inhibition. J. Bacteriol. 106, 493-499 The term " indirect suppression " has been 4. Kisumi, M., Komatsubara, S., Sugiura, M., & used to describe suppression which is based on Chibata, 1. (1971) J. Bacterial. 107, 741-745 alteration of the expression of a gene distinct from 5. Kisumi, M., Komatsubara, S., Sugiura, M., & the original mutation (23). From the viewpoint Chibata, I. (1972) J. Bacteriol. 110, 761-763 of metabolic regulation, we are interested in the 6. Kisumi, M., Komatsubara, S., & Chibata, I. (1973) indirect suppression caused by changes of enzymes J. Biochem. 73, 107-115 different from the primarily altered enzyme. Speci 7. Davis, B.D. & Mingioli, E.S. (1950) J. Bacteriol. 60, 17-28 fic proline auxotrophy was shown to be suppressed 8. Cho-Chung, Y.S. & Umbarger, H.E. (1971) in by genetic loss of the arginine biosynthetic enzyme Methods in Enzymology (Tabor, H. & Tabor, C.W., in E. coli and S. typhimurium (24, 25). That is, eds.) Vol. 17A, pp. 782-785, Academic Press, New this suppression is based on glutamate-r-semial York dehyde (a proline precursor) synthesis from N 9. Parsons, S.J. & Burns, R.O. (1971) in Methods in acetylglutamate-r-semialdehyde (an arginine pre Enzymology (Tabor, H. & Tabor, C.W., eds.) Vol. cursor). Another example was reported for the 17A, pp. 793-799, Academic Press, New York isoleucine revertant of Bacillus subtilis (26, 27). In 10. Friedemann, T.E. & Haugen, G.E. (1943) J. Biol. this revertant, a-ketobutyrate (an isoleucine pre Chem. 147, 415-442 cursor) was directly synthesized from phospho 11. Phillips, A.T., Nuss, J.L, Moosic, J., & Foshay, C. homoserine (a threonine precursor) by derepressed (1972) J. Bacteriol. 109, 714-719 12. Abramsky, T. & Shemin, D. (1965) J. Biol. Chem. threonine synthetase [EC 4. 22. 99. 2] (28). The 240,2971-2975 reversion of isoleucine auxotrophy in S. marcescens 13. Flavin, M. & Segal, A. (1964) J. Biol. Chem. 239, can be taken as a new example of indirect suppres 2220-2227 sion. Certain enzymes are known to have affinities 14. Meister, A. (1965) in Biochemistry of the Amino for several substrates and to have catalytic versati Acids Vol. 2, pp. 757-818, Academic Press, New lity. The study of indirect suppression is therefore York important in elucidating the significance of various 15. Charon, N.W., Johnson, R.C., & Peterson, D. enzymes in metabolism. Such phenomena have (1974) J. Bacteriol. 117, 203-211 been observed in the deacylation of N-acetylgluta 16. Sai, T., Aida, K., & Uemura, T. (1969) J. Gen. Appl. mate-r-semialdehyde catalyzed by acetylornithinase Microbiol. 15, 345-363 17. Sasaki, K., Nakano, H., & Katsuki, H. (1971) [EC 3.5.1.16], the deamination of phosphohomo J. Biochem. 70, 441-449 serine catalyzed by threonine synthetase, and the 18. Kisumi, M., Sugiura, M., Kato, J. & Chibata, I. synthesis of citramalate catalyzed by a-isopro (1976) J. Biochem. 79, 1021-1028 pylmalate synthetase. Furthermore, the study of 19. Kohlhaw, G., Leary, T.R., & Umbarger, H.E. (1969) indirect suppression is also likely to be important J. Biol. Chem. 244, 2218-2225 in the selection of feedback-insensitive and consti 20. Kisumi, M., Sugiura, M., & Chibata, I. (1976) tutive mutants, when suitable amino acid analogs J. Biochem. 80, 333-339 are not available as antagonists. 21. Krampitz, L.O. (1961) in The Bacteria (Gunsalus, I.C. & Stainier, R.Y., eds.) Vol. 2, pp. 205-256, Academic Press, New York

J. Biochem. PATHWAY FOR ISOLEUCINE FORMATION FROM PYRUVATE 103

22. Meister, A. (1965) in Biochemistry of the Amino 26. Vapnek, D. & Greer, S. (1971) J. Bacteriol. 106, Acids Vol. 2, pp. 928-951, Academic Press, New 615-625 York 27. Skarstedt, M.T. & Greer, S.B. (1973) J. Biol. Chem. 23. Gorini, L., & Beckwith, J.R. (1966) Ann. Rev. 248,1032-1044 Microbiol. 20, 401-422 28. Schildkraut, I. & Greer, S. (1973) J. Bacteriol. 115, 24. Itikawa, H., Baumberg, S., & Vogel, H.J. (1968) 777-785 Biochim. Biophys. Acta 159, 547-550 25. Berg, C.M. & Rossi, J.J. (1974) J. Bacteriol. 118, 928-939

Vol. 82, No. 1, 1977