Glutamate Synthesis in coelicolorwith good sources of glutamate. S. coelicolor mutants deficient in GOGAT(Glt-) required glutamate for growth with L-alanine, asparagine, arginine,

JOURNAL OF BACTERIOLOGY, May 1989, p. 2372-2377 Vol. 171, No. 50021-9193/89/052372-06$02.00/0Copyright C) 1989, American Society for Microbiology

Glutamate Synthesis in Streptomyces coelicolorSUSAN H. FISHER

Department of Microbiology, Boston University School of Medicine, Boston, Massachusetts 02118

Received 1 November 1988/Accepted 26 January 1989

Both glutamate synthase (GOGAT) and glutamate dehydrogenase (GDH) are involved in glutamate synthesisin Streptomyces coelicolor. The highest levels of GDH were seen in extracts of cells grown with high levels ofammonium as the nitrogen source. GOGAT activity was reduced two- to threefold in extracts of cells grownwith good sources of glutamate. S. coelicolor mutants deficient in GOGAT (Glt-) required glutamate forgrowth with L-alanine, asparagine, arginine, or histidine as the nitrogen source but grew like wild-type cellswhen ammonium, glutamine, or aspartate was the nitrogen source. The glt mutations were tightly linked tohisAl. Mutants deficient in both GOGAT and GDH (Gdh-) required glutamate for growth in all media. Thegdh-5 mutation was mapped to the left region of the S. coelicolor chromosomal map, between proAI and uraAI.

Members of the genus Streptomyces are gram-positivesporulating soil bacteria. These organisms synthesize nu-merous unique compounds, secondary metabolites, thatoften possess antibacterial, antitumor, or antiparasitic activ-ity (5). Although primary metabolites serve as substrates forthe biosynthetic enzymes for secondary metabolites, sur-prisingly little is known about the basic physiology ofStreptomyces spp. (5). Enzymes and pathways responsiblefor the synthesis of many primary metabolites have not beendescribed, and the mechanisms regulating carbon and nitro-gen metabolism are not understood.

In bacteria, glutamate serves as a major source of nitrogenfor cellular metabolites because its amino group providesnitrogen for the synthesis of most amino acids (14). Toincrease our understanding of nitrogen metabolism in Strep-tomyces spp., enzymes responsible for glutamate synthesiswere investigated. Streptomyces coelicolor was chosen forthese studies because previous work with this organism ledto a system for genetic transfer by conjugation, constructionof a chromosomal map, and development of methods forDNA cloning (8).

Genetic and biochemical studies in members of the familyEnterobacteriaceae have shown that two enzymes, gluta-mate dehydrogenase (GDH) and glutamate synthase(GOGAT), are involved in glutamate synthesis (14). Duringgrowth in media containing high levels of ammonium(greater than 1 mM), ammonium can be assimilated directlyinto glutamate by GDH. When the level of ammonium in themedium is low (less than 0.1 mM), GDH is unable tosynthesize adequate levels of glutamate because its K,n forammonium is relatively high. Instead, glutamate is synthe-sized from glutamine and 2-ketoglutarate by GOGAT. Thus,both glutamine synthetase and GOGAT are required forglutamate synthesis during growth on nitrogen sources thatare poor sources of ammonium. In Bacilluis subtilis, agram-positive bacterium, only GOGAT is involved in gluta-mate synthesis. B. suibtilis mutants deficient in GOGAT areglutamate auxotrophs on all nitrogen sources (6).The enzymes involved in glutamate synthesis have not

been extensively characterized in Streptomyces spp. BothGDH and GOGAT activities can be detected in cell extractsof S. venezuelae (18) and S. nolursei (7), whereas onlyGOGAT activity is present in S. clavuligerus extracts (1, 2).GOGAT appears to be the only enzyme involved in gluta-mate synthesis in S. cla'uligerus because S. clavuligerusmutants deficient in GOGAT require glutamate for growth in

all media (2). GOGAT purified from the actinomycete No-cardia mediterranei is similar to other bacterial GOGATenzymes, i.e., a complex of iron-sulfur flavoproteins withtwo dissimilar subunits (13). However, neither GOGAT norGDH from Streptomyces spp. has been described.

MATERIALS AND METHODSBacterial strains. All bacterial strains used were deriva-

tives of S. coelicolor A3(2) (8) (Table 1).Media and culture techniques. R2 medium (8), a complex

buffered medium containing 10% sucrose, 1% glucose, 0.5%yeast extract, and 0.01% Casamino Acids (Difco Laborato-ries, Detroit, Mich.), was used for preparation of spores andgenetic crosses. In some experiments, modified R2 mediumlacking glucose and buffered with 0.1 M MOPS (morpholinepropanesulfonic acid; pH 7.0) was used. SMS minimalmedium was used to analyze growth of mutants on variousnitrogen sources and to score for nutritional growth require-ments of recombinants from genetic crosses. The basal SMSminimal solution contained 50 mM NaH2PO4-K2HPO4 (pH7) and 0.1% of the trace element solution (8). After autoclav-ing, 2% glucose and 0.02% MgSO4 7H20 was added. Thefinal concentration of all nitrogen sources except glutaminewas 0.2%; glutamine was added at 0.4%. Glutamine andhistidine were prepared fresh for each experiment and filtersterilized. When growth of mutants on various nitrogensources was tested, plates were prepared with 2% Nobleagar (Difco), and nutritional supplements were added at 20%of the recommended concentrations (8). Streptomycin wasused at final concentrations of 20 ,ug/ml in minimal plates and50 ,ug/ml in R2 plates.

Since streptomycetes are nonfragmenting mycelial organ-isms, special procedures are necessary to obtain reasonablydispersed growth of cells in liquid culture. SMS minimalmedium was supplemented with 5% polyethylene glycol8000 (Sigma Chemical Co., St. Louis, Mo.) and 0.2 mMCaCl, 2H20. Nutritional supplements were added at twicethe recommended amounts (8). YEME (8), a complex me-dium containing yeast extract and Bacto-Peptone (Difco),was buffered with 0.1 M MOPS (pH 7). Cultures were grownin 500-ml sidearm flasks baffled with steel springs (1.3-cm-diameter coil 19 gauge wire), siliconized with 0.1% Prosilto reduce cell growth on flask walls, and closed with eithercotton plugs or aluminum foil. Sidearm flasks containing 50to 100 ml of culture were shaken at 300 rpm in an orbitalshaker (New Brunswick Scientific Co., Inc., Edison, N.J.) at

2372

on Septem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

GLUTAMATE SYNTHESIS IN S. COELICOLOR 2373

TABLE 1. Strains of S. coelicolor used

Strain Relevant genotypea Source or derivation

A3(2) Wild type, SCP1+ SCP2+ D. A. Hopwood380 glt_3b D. A. Hopwood; muta-

genesis of A3(2)547 glt_5b D. A. Hopwood; muta-

genesis of A3(2)1258 argAl cysC3 hisC9 proAl K. F. Chater

strAl uraAl NFS31 glt-3 argAl cysC3 strAl This work; 1258 x 380

NF2612 proAl argAl cysC3 NF D. A. Hopwood

SCP2+J712 proAl hisAl NF K. F. ChaterJ1508 hisAl uraAl strAl NF K. F. Chater

SCP2-Sil hisAl uraAl strAl glt-JI This work; mutagenesis

NF SCP2- of J1508S80 proAl hisAl glt-80 NF This work; mutagenesis

of J712S1o0 proAl hisAl glt-101 NF This work; mutagenesis

of J712S221 proAl hisAl glt-221 NF This work; mutagenesis

of J712S275 proAl hisAl glt-275 NF This work; mutagenesis

of J712S55 glt-5 gdh-5b This work; mutagenesis

of 547S58 glt-5 gdh-8b This work; mutagenesis

of 547S59 glt-5 gdh-9b This work; mutagenesis

of 547a Abbreviations: SCP1 and SCP2, S. coelicolor plasmids 1 and 2 (8); NF,

SCP1 integrated into the chromosome; glt, glutamate requirement for growthon medium containing asparagine or alanine as the nitrogen source; gdh,glutamate requirement for growth on medium containing ammonium as thenitrogen source.

b Strain is likely to be SCP1+ SCP2+, but the plasmid state has not beendetennined.

32°C. Growth was monitored with a Klett-Summerson col-orimeter, using the green filter.

Liquid cultures were inoculated by a modification of theprocedure of Hopwood et al. (8). A dense suspension offrozen spores was pelleted, suspended in 0.5 ml of 5.73%TES [N-tris(hydroxymethyl) methyl-2-aminoethane sulfonicacid], and incubated at 45°C for 15 min. Heat-shocked sporeswere cooled to room temperature, pelleted, and inoculatedinto germination medium (1% yeast extract, 1% CasaminoAcids, 0.001% Triton X-100, 0.01% MgSO4 7H20, 0.1 MMOPS [pH 7.0], 0.1 M Tris [pH 7.3], 10 mM CaC12- 2H20).After 4 to 5 h at 32°C, 5 to 25% of the spores had microscop-ically visible germ tubes. Germinated spores were pelleted,washed once with SMS lacking a carbon or nitrogen source,and inoculated into minimal medium so that the initialturbidity of the culture was less than 1 Kl'ett unit. Sincecultures inoculated with aggregated germinating spores growas mycelial balls, spore clumps were removed either byfiltration through sterile cotton (8) or by a brief (1 to 2 s) spinin a microfuge immediately before inoculation. Spores wereprepared as described by Hopwood et al. (8). Typically,spores from a well-sporulated R2 plate were used to inocu-late four to six cultures.

Mutagenesis and mutant isolation. Mutagenesis of sporeswith N-methyl-N'-nitro-N-nitrosoguanidine was performedas described by Hopwood et al. (8). Mutagenized sporeswere plated on R2 plates so that each plate contained 50 to150 colonies. When the colonies were densely covered with

spores, growth on various nitrogen sources was examined byreplica plating.

Genetic mapping techniques. Genetic crosses and dataanalysis were carried out as described by Hopwood et al. (8).In all crosses, parental genomes were excluded from recom-binants by selective plating. Recombinants were purified bystreaking on selective medium, and single colonies werepatched onto R2 medium. Nutritional growth requirementsof the recombinants were determined by replica plating ontoappropriate minimal medium. X2 values were calculated asdescribed by Hopwood et al. (8).Enzyme assays. Extracts for assays were prepared from

cells grown to mid-logarithmic growth phase (Klett readingof 70 to 90). Harvested cells were washed once with buffer A(50 mM imidazole [pH 7.5], 150 mM NaCl, 1 mM MnCl2, 0.5mM dithiothreitol) and stored at -20°C. Thawed cells weresuspended in the appropriate assay buffer and sonicatedbriefly at 4°C; cell debris was removed by centrifugation for15 min at 4°C in an Eppendorf microfuge.

Histidase was assayed by the method of Chasin andMagasanik (4) in cell extracts prepared with either buffer Aor buffer B (0.05 M Tris [pH 7.5], 150 mM NaCi). Extractsfor GOGAT, GDH, alanine dehydrogenase (ADH), andtransaminase assays were prepared with buffer B. GOGATand GDH were measured by the method of Meers et al. (12)except that NADH was used instead of NADPH in theGOGAT assay. ADH was assayed as described by Brana etal. (2). The method of lijima et al. (9) was used to assayglutamate:oxaloacetate transaminase (GOT). Alanine:2-oxo-glutarate transaminase activity was measured as describedby Segal and Matsuzawa (17). Alanine:2-isovalerate trans-aminase was assayed by substituting 10 mM 2-isovalerateacid for 2-ketoglutarate in the alanine:2-oxoglutarate trans-aminase assay. Protein concentrations were determined bythe method of Lowry et al. (11), with bovine serum albuminas the standard.

RESULTS

Expression of glutamate biosynthetic enzymes in wild-typecells. To identify enzymes involved in glutamate synthesis inS. coelicolor, GOGAT and GDH were assayed in extracts ofJ1508 grown with various nitrogen sources (Table 2). Bothenzyme activities were present. As in S. clavuligerus (2) andS. venezuelae (18), GOGAT activity was detected in S.coelicolor extracts only when NADH, not NADPH, wasused in the GOGAT assay mixture,. GOGAT expression wasreprissed severalfold by growth in medium containing com-pounds that can be metabolized to glutamate. The lowestGOGAT levels were seen in extracts of cells grown either inYEME, a rich medium, or in aspartate-containing medium.Aspartate could be metabolized to glutamate in S. coelicolorbecause GOT activity was present in cell extracts (Table 2).Similarly, GOGAT levels were threefold lower in extracts ofcells grown in medium containing ammonium, aspartate, andglutamine than in extracts of cells grown in medium contain-ing only ammonium (Table 2).

In contrast, GDH levels were 8- to 10-fold higher inextracts of cells grown with high levels of ammonium as thesole nitrogen source than in cells grown in medium contain-ing either asparagine or aspartate or in complex mediumcontaining ammonium and glutamine (Table 2). However,GDH synthesis was not completely derepressed in cellsgrown with both high levels of ammonium and a good sourceof glutamate. GDH levels were 2.5-fold lower in extracts ofcells grown in medium with ammonium, aspartate, and

VOL. 171, 1989

on Septem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

TABLE 2. Enzyme activities in J1508 cells grown on various nitrogen sources

Sp act"Nitrogen source

GOGAT GDH GOT ADH

YEME + NH4' + glutamine 20 ± 14 (3) 27 + 19 (3) 265 + 64 (2) 88 ± 16 (2)NH4+ 90 + 17 (3) 273 ± 36 (3) 313 ± 80 (2) 35NH4+ + aspartate + glutamine 24 9 (4) 96 32 (3) ND NDAspartate 46 28 321 20Asparagine 62 + 32 (2) 25 ± 15 (3) 337 21Glutamine 82 ± 27 (2) 101 ± 4 (2) 209 ± 40 (2) 50 ± 5 (2)Histidine 112 80 321 27L-Alanine 135 71 ND 2,138

a Expressed as nanomoles per minute per milligram of protein. Numbers in parentheses are numbers of independent determinations. Values are averages ofall measurements; error is the standard deviation. ND, Not done.

glutamine than in cells grown in medium containing onlyammonium (Table 2).ADH activity is known to be present in S. clavuligerus, S.

noursei, S. erythreus, S. venezuelae, and S. coelicolor (2, 7,15', 18; Table 2). This enzyme could participate in ammoniumassimilation. ADH levels were 100-fold lower in extracts ofwild-type S. coelicolor cells grown with aspartate or ammo-nium than in cells grown with L-alanine (Table 2). Thisfinding suggests that ADH activity is important for theutilization of L- alanine as a nitrogen source but not forNH4+ assimilation in S. coelicolor. If ADH does play asignificant role in ammonium assimilation in S. coelicolor,then enzymes that transfer the amino group from L-alanineto other metabolites should be present in S. coelicolorextracts. Neither alanine:2-ketoglutarate transaminase noralanine:2-isovalerate transaminase activity could be de-tected in extracts of S. coelicolor cells grown with ammo-nium as the nitrogen source (data not shown).Mutants deficient in GOGAT. The observation that GDH

synthesis is derepressed in extracts of cells grown with highlevels ofammonium suggests that GDH is the major providerof glutamate durtng growth on this itiedium. If so, twohypotheses can be put forth: (i) GOGAT is the enzymeresponsible for glutamate synthesis during growth of S.coelicolor on compounds that provide low levels of ammo-nium, and (ii) S. coelicolor mutants deficient in GOGATrequire glutamate for growth on nitrogen compounds that arepoor sources of ammonium but grow like wild-type cells onhigh levels of ammonium (since they still have GDH activ-ity).

TABLE 3. Growth phenotypes of wild-type and mutant cellsa

PhenotypeNitrogen source

Wild type Glt- Glt- Gdh-

NH4+ ++ ++ -NH4+ + glutamate ++ ++ ++Glutamine i+ ++ ++Glutamate + + +Aspartate +-+ +i+ ++Asparagine ++Alanine +Proline + + +Histidine +Arginine +Urea + +

Wild-type and mutant cells were allowed to sporulate on modified R2medium and were replica plated onto minimal plates containing glucose andthe indicated nitrogen sources. Plates were scored after incubation at 32°C for2 days. + +, Good growth; +, growth; ±. poor growth; -. no growth.

These hypotheses were confirmed by the growth pheno-type of mutants deficient in GOGAT. Seven mutants defi-cient in GOGAT (Glt) and whose glutamate auxotrophycould not be satisfied by asparagine, alanine, histidine, orarginine were isolated. These mutants grew like wild-typecells on high levels of either ammonium or nitrogen com-pounds that could be rapidly metabolized to glutamate(Table 3). Measurements of GOGAT and GDH activities inextracts of Glt- and wild-type cells demonstrated thatGOGAT activity, but not GDH activity, was deficient in theGlt- mutants (Table 4).The bacterial GOGAT enzyme consists of two dissimilar

subunits (6, 13, 14, 16). If the S. coelicolor GOGAT has ananalogous structure, it is possible that extracts of mutantshaving lesions in different GOGAT subunits might comple-ment each other. No increase in GOGAT activity was seenin any combination of the Glt- extracts, although wild-typeGOGAT activity was not inhibited by the addition of Glt-extracts (data not shown).Mapping of the glt mutations. The glt mutations were

mapped by plasmid-mediated chromosomal transfer as de-scribed by Hopwood et al. (8). In the cross between S31 andJ712, Strr and His' recombinants were selected and ana-lyzed (Fig. 1). Since the glt-3 mutation segregated 100% withhisA in this cross, the glt-3 mutation is tightly linked to hisA.The glt-5 and glt-JJ mutations are also tightly linked to hisA.In a cross between strains 547 (glt-5) and 1258, 97% of allHis' recombinants were Glt- (data not shown). All His'recombinants obtained from a cross between S11 (glt-JJ) and2612 were Glt- (data not shown).

Histidase expression in Glt- mutants. In Escherichia coli,the gltBDF operon contains the two GOGAT structural

TABLE 4. GOGAT and GDH activities in wild-type and mutantcells grown with ammonium as the nitrogen source

Sp actaStrain Relevant genotype

GOGAT GDH

A3(2) glt+ 61 241380 glt-3 <1.2 128547 glt-5 <1.6 154J1508 glt+ 72 257Sil git-lI <2.4 209J712 glt+ 83 NDS80 glt-80 <0.96 NDS1o0 glt-101 <1.2 NDS221 glt-221 <1.2 NDS275 glt-275 <2.0 ND

" Expressed as nanomoles per minute per milligram of protein. ND, Notdone.

2374 FISHER J. BACTERIOL.

on Septem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

GLUTAMATE SYNTHESIS IN S. COELICOLOR 2375

St rai n J7 12

q I t I

Sir S

0

hI st hIsAl

0 48qit-3 20 0

p < 0.001FIG. 1. Mapping of glt-3 by crossing S31 (inner circle) with J712

(outer circle). Selection for strAl and Arg+ (A) excluded parentalgenomes. Numbers around the circle indicate allele frequenciesamong the 68 recombinants scored. Segregation of the glt-3 allelewith respect to the hisAl allele is tabulated.

TABLE 5. Histidase activity in wild-type and Glt- cells grownwith glutamine and histidine as nitrogen sources

Strain Relevant genotype Histidase activitya

A3(2) glt+ 75 ± 5 (2)380 glt-3 47547 glt-5 69 ± 10 (2)J1508 glt+ 90 ± 10 (2)S11 glt-lI <1 ± 0.1 (3)J712 glt+ 58S80 glt-80 <2S101 glt-101 53S275 glt-275 58

a Expressed as nanomoles per minute per milligram of protein. Numbers inparentheses are numbers of independent determinations. Values are averagesof all measurements; error is the standard deviation. Histidase levels inextracts of A3(2) grown in the absence of histidine were <1.5.

from the Glt- strain 547. Measurement of GDH levels inextracts of the single and double mutants showed that GDHlevels were 2.5- to 9-fold lower in extracts of S55, S58, andS59 cells than in extracts of 547 cells (Table 6). GOT levelsin extracts of S55 were similar to those seen in the geneticparents, A3(2) and 547, whereas ADH levels were 30-foldhigher in extracts of S55 cells than in A3(2) and 547 extracts(Table 6).Mapping of the gdh-5 mutation. The gdh-S mutation was

mapped by crossing S55 with 1258 (Fig. 2). On the basis ofthe allele frequencies (Table 7), gdh-5 is located betweenuraAl and proAl. The gradient of allele frequencies in theleft region of the chromosome is uneven in this cross,possibly because of the growth phenotype of the gdh-5recombinants. Unlike the donor Gdh-5- strain S55, all gdh-5recombinants grew as small colonies on solid media, did notmake the pigmented antibiotic actinorhodin, and sporulatedpoorly. Thus, the gdh-5 recombinants may be underrepre-sented among recombinants because of their lower frequen-cies of sporulation.

genes and a third gene, gltF, which is involved in theexpression of the histidine-degrading enzymes (hut) (3).Since mutations in the gltB and gltD genes affect expressionof gitF, hut expression is defective in E. coli Glt- mutants.To determine whether hut synthesis was altered in the S.coelicolor Glt- strains, histidase, the first enzyme in thehistidine-degradative pathway, was assayed in extracts ofmutant and wild-type cells. Histidase was present at wild-type levels in extracts of the Glt- strains 380, 547, S101, andS275, whereas uninduced levels of histidase were seen inextracts of the Glt- mutants Sli and S80 (Table 5).The defect in histidase synthesis in the Glt- strain S11

could have resulted from a second mutation in hut ratherthan from the git-JI mutation. If so, Hut- Glt+ recombinantsshould be obtained in a cross between S11 and a wild-typestrain because the chromosomal map position of hut, atapproximately 9 o'clock on the S. coelicolor chromosomalmap, is well separated from the glt locus (10; Fig. 1). Pro'and Ura+ recombinants from a cross between S11 and 2612were scored for the Glt- and Hut- phenotypes by examininggrowth on L-alanine and histidine, respectively, as nitrogensources. Since all Hut- recombinants were Glt- and all Glt-recombinants were Hut-, the glt-1l mutation was likelyresponsible for the defect in histidase synthesis in S11.Mutants deficient in GDH. Glt- mutants that require

glutamate for growth on high levels of ammonium should bedeficient in GDH. Three such mutant strains were isolated

TABLE 6. Enzyme activities in wild-type and mutant cells

Sp actbStraina Derivation

GOGAT GDH GOT ADH

Expt 1A3(2) Wild type 31 48 ND ND547 Mutagenized <0.8 36 ND ND

A3(2)S55 Mutagenized <1.1 4 ND ND

547S58 Mutagenized <1.0 10 ND ND

547S59 Mutagenized <1.2 14 ND ND

547

Expt 2A3(2) Wild type 26 + 4 (2) 86 176 28547 Mutagenized <0.9 106 176 33

A3(2)S55 Mutagenized <1.0 11 + 3 (2) 273 1,053 + 16 (2)

547

"Cells in experiment 1 were grown in minimal medium containing ammo-nium and glutamate as nitrogen sources. Cells in experiment 2 were grownwith glutamine as the nitrogen source.

b Expressed as nanomoles per minute per milligram of protein. Numbers inparentheses are numbers of independent determinations. Values are averagesof all measurements; error is the standard deviation. ND, Not done.

VOL. 171, 1989

on Septem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

2376 FISHER

219 219hi st 1gt-5

gdh-51 8

S t r S0

Strain S55FIG. 2. Mapping of gdh-5 by crossing S55 (inner circle) with 1258

(outer circle). Selection for His' and strAl (A) excluded parentalgenomes. Numbers around the circle indicate allele frequenciesamong the 219 recombinants scored.

DISCUSSION

Both physiological and genetic evidence indicates thatglutamate is synthesized by GDH when S. coelicolor cellsare grown on medium containing high levels of ammonium asthe nitrogen source. GDH synthesis is derepressed 8- to10-fold when high levels of ammonium are available. Glt-strains that are completely deficient in GOGAT expressGDH at wild-type levels and grow like wild-type cells onhigh levels of ammonium. In contrast, mutants deficient inboth GOGAT and GDH require glutamate for growth evenwith high levels of ammonium.GOGAT also appears to be involved in glutamate synthe-

sis in S. coelicolor. GOGAT synthesis is repressed several-

TABLE 7. Segregation of gdh-5 with auxotrophic markers"

Probability ofAllele No. No. X2 independentgdh+ gdhi-S segregation

arg+ 189 17argA l 12 1 0.0058 >0.99

cVs+ 100 13cysC3 101 5 3.3 >0.3

pro+ 22 18proAI 179 0 87.7 <0.001

ira+ 44 17iraA l 157 1 43.2 <0.001

" Genotypes of recombinants obtained from S55 x 1258 (Fig. 2).

fold in cells grown in the presence of a good source ofglutamate, such as aspartate. In addition, S. coelicolormutants deficient in GOGAT are unable to utilize arginine,histidine, asparagine, or L-alanine as a nitrogen source andgrow poorly compared with wild-type cells on urea as anitrogen source. Degradation of urea and L-alanine is un-likely to provide glutamate in S. coelicolor. Urea is metab-olized to ammonium and CO2 by urease (S. Fisher, unpub-lished observations), and L-alanine appears to be degradedto ammonium and pyruvate by ADH. Since our Glt- mu-tants cannot grow or grow poorly on at least two nitrogensources whose degradation does not provide glutamate,GOGAT must be required for glutamate synthesis in S.coelicolor under some growth conditions.

In contrast, glutamate may be produced by histidine,arginine, and asparagine degradation in S. coelicolor. Histi-dine is degraded to formyglutamic acid in S. coelicolor (10).It is not known whether formyglutamic acid is furthermetabolized to glutamate (10). The pathways for arginineand asparagine catabolism have not been described for S.coelicolor. Degradation of asparagine to aspartate and am-monium by asparaginase could result in glutamate synthesisif the amide group of aspartate were subsequently trans-ferred to 2-ketoglutarate by GOT. If histidine, arginine, andasparagine are degraded to glutamate, the inability of S.coelicolor Glt- mutants to grow on these compounds sug-gests that additional defects in nitrogen metabolism may bepresent in the Glt- mutants.

Unexpectedly, a defect in the synthesis of at least one ofthe hut enzymes was found in several of the Glt- mutants.Since a second mutation could not be separated duringconjugation from the glt-lI mutation, the defect in hutexpression appears to be due to the glt mutation. The gltmutation in this mutant maps in the same chromosomalregion as do the glt-3 and glt-5 mutations. It is possible thatthe Hut' Glt- mutants contain mutations in the glt structuralgenes, whereas the glt mutation in the Hut- Glt- mutantslies in a gene required for expression of both the glt and thehut genes. Alternatively, the glt operon in S. coelicolor maybe similar to the E. coli gltBDF operon (3) and contain a genewhich is involved in hut expression. If so, the hut regulatorygene would be expressed in the Hut' Glt- mutants, butexpression of this gene would be reduced in the Hut- Glt-mutants, perhaps because of polarity.There are at least two possible explanations for the

observation that all gdh-5 recombinants grow and sporulatepoorly compared with the donor strains. First, the gdh-5mutation may cause the growth defect. If so, the donorstrain, S55, must contain a second mutation that compen-sates for the altered growth phenotype of the gdh-S muta-tion. The gdh-5 recombinants would grow poorly if thissuppressor mutation did not segregate with the gdh-S muta-tion during conjugation. Alternatively, since 1258 and S55are not nonisogenic strains, 1258 may contain a mutationthat causes a growth defect in the presence of the gdh-5mutation. Thus, the gdh-S recombinants grow poorly be-cause of transfer of the gdh-S mutation into the 1258 geneticbackground. If gdh mutations do cause defects in growth andsporulation, the inability to isolate mutants completely defi-cient in GDH may have been due to the poor viability of suchmutants.ADH levels were 50-fold higher in extracts of the Gdh-

mutant S55 grown with glutamine as the nitrogen source thanthey were in analogous wild-type extracts (Table 6). Highlevels of ammonium caused a 50-fold increase in ADHexpression in S. clavuligerus (1, 2) but not in wild-type S.

J. BACTERIOL.

on Septem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

GLUTAMATE SYNTHESIS IN S. COELICOLOR 2377

coelicolor (Table 2) or S. venezulae (18). It is possible thatthe reduced GDH levels in mutant S55 caused an elevation inthe intracellular ammonium pool, which in turn inducedADH expression. These metabolic perturbations may beinvolved in the altered growth and sporulation phenotypes ofthe Gdh- recombinants.

ACKNOWLEDGMENTS

I am grateful to D. A. Hopwood and K. F. Chater for sharing S.coelicolor strains, for helpful advice on mapping mutations in S.coelicolor, and for providing laboratory space for the initial exper-iments. I also thank L. V. Wray for helpful discussions, SusanCongers and Mari Atkinson for technical support, and L. V. Wrayand A. L. Sonenshein for reading the manuscript.

This work was supported by Public Health Service research grantRO1-AI23168 and biomedical research support grant RR05380-26from the National Institutes of Health and by an Eli Lilly life sciencegrant.

LITERATURE CITED1. Aharonowitz, Y. 1979. Regulatory interrelationships of nitrogen

metabolism and cephalosporin biosynthesis, p. 210-217. In 0.Sebek and A. Laskin (ed.), Genetics of industrial microorgan-isms. American Society for Microbiology, Washington, D.C.

2. Brana, A. F., N. Paiva, and A. L. Demain. 1986. Pathways andregulation of ammonium assimilation in Streptomyces cla-vuligerus. J. Gen. Microbiol. 132:1305-1317.

3. Castano, I., F. Bastarrachea, and A. A. Covarrubias. 1988.gltBDF operon of Escherichia coli. J. Bacteriol. 170:821-827.

4. Chasin, L. A., and B. Magasanik. 1968. Induction and repres-sion of the histidine-degrading enzymes of Bacillus subtilis. J.Biol. Chem. 243:5165-5178.

5. Demain, A. L., Y. Aharonowitz, and J. F. Martin. 1983. Meta-bolic control of secondary biosynthetic pathways. p. 49-72. InL. Vining (ed.), Biochemistry and genetic regulation of com-mercially important antibiotics. Addison-Wesley PublishingCo., Reading, Mass.

6. Deshpande, K. L., and J. F. Kane. 1980. Glutamate synthasefrom Bacillus subtilis: in vitro reconstitution of an active ami-dotransferase. Biochem. Biophys. Res. Commun. 93:308-314.

7. Grafe, U., H. Bocker, and H. Thrum. 1977. Regulative influence

of o-aminobenzoic acid on the biosynthesis of nourseothricin incultures of Streptomyces noursei JA3890b. II. Regulation ofglutamine synthetase and the role of the glutamine synthetase/glutamate synthase pathway. Z. Allg. Mikrobiol. 17:201-209.

8. Hopwood, D. A., M. J. Bibb, K. F. Chater, T. Kieser, C. J.Bruton, H. M. Kieser, D. J. Lydiate, C. P. Smith, J. M. Ward,and H. Schrempf. 1985. Genetic manipulations of Streptomyces,a laboratory manual. The John Innes Foundation, Norwich,United Kingdom.

9. Iijima, T., M. D. Diesterhaft, and E. Freese. 1977. Sodium effectof growth on aspartate and genetic analysis of a Bacillus subtilismutant with high aspartase activity. J. Bacteriol. 129:1440-1447.

10. Kendrick, K. E., and M. L. Wheelis. 1982. Histidine dissimila-tion in Streptomyces coelicolor. J. Gen. Microbiol. 128:2029-2040.

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

12. Meers, J. L., D. W. Tempest, and C. M. Brown. 1970. 'Glu-tamine(amide):2-oxoglutarate amino transferase oxidoreductase(NADP),' an enzyme involved in the synthesis of glutamate bysome bacteria. J. Gen. Microbiol. 64:187-194.

13. Mei, B., and R. Jiao. 1988. Purification and properties ofglutamate synthase from Nocardia mediterranei. J. Bacteriol.170:1940-1944.

14. Reitzer, L. J., and B. Magasanik. 1987. Ammonium assimilationand the biosynthesis of glutamine, glutamate, aspartate, aspar-agine, L-alanine, and D-alanine, p. 302-320. In F. C. Neidhart,J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaecter, andH. E. Umbarger (ed.), Escherichia coli and Salmonella typhi-murium: cellular and molecular biology. American Society forMicrobiology, Washington, D.C.

15. Roszkowski, J., A. Rafalski, and K. Raczynska-Bojanowska.1969. Alanine and alanine dehydrogenase in Streptomyceserythreus. Acta Microbiol. Pol. 1:59-68.

16. Schrier, H. J., and R. W. Bernlohr. 1984. Purification andproperties of glutamate synthase from Bacillus licheniformis. J.Bacteriol. 160:591-599.

17. Segal, H. L., and T. Matsuzawa. 1960. L-Alanine aminotrans-ferase (rat liver). Methods Enzymol. 17A:153-159.

18. Shapiro, S., and L. C. Vining. 1983. Nitrogen metabolism andchloramphenicol production in Streptomyces venezuelae. Can.J. Microbiol. 29:1706-1714.

VOL. 171, 1989

on Septem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from