Vol. 175, No. 24JOURNAL OF BACTERIOLOGY, Dec. 1993, p. 8024-80290021-9193/93/248024-06$02.00/0Copyright © 1993, American Society for Microbiology

Isolation of a Glutamate Synthase (GOGAT)-Negative,Pleiotropically N Utilization-Defective Mutant of Azospirillumbrasilense: Cloning and Partial Characterization of GOGAT

Structural GeneASIM KUMAR MANDAL AND SUDHAMOY GHOSH*

Department of Biochemistry and Centre for Plant Molecular Biology, Bose Institute,AJCB Centenary Building, Calcutta 700 054, India

Received 6 July 1993/Accepted 5 October 1993

An Azospirillum brasilense mutant (N12) pleiotropically defective in the assimilation of nitrogenouscompounds (Asm-) was isolated and found lacking in the glutamate synthase (GOGAT-). The glt (GOGAT)locus of A. brasilense was identified by isolating a broad-host-range pLAFR1 cosmid clone from a gene libraryof the bacterium that rectified Asm - and GOGAT - defects (full recovery of activities of the nitrogenase, theassimilatory nitrate and nitrite reductases, and the glutamate synthase). A 7.5-kb EcoRI fragment of thecosmid clone that also complemented N12 was partially sequenced to identify the open reading frame for theot-subunit of GOGAT. The amino acid sequences deduced from the partial nucleotide sequences of the glt locusof A. brasilense showed considerable homology with that of the oa-subunit of GOGAT coded by the gitB gene ofEscherichia coli. The genetic lesion of N12 was found within the gltB gene ofA. brasilense. The gltB promoter ofA. brasilense showed the presence of a consensus sigma-70-like recognition site (as in E. coli) in addition topotential NtrA-RNA polymerase, IHF, and NifA binding sites.

In Azospirillum brasilense, a nitrogen-fixing aerobic soilbacterium well known for its ability to colonize roots of grassplants of economic importance and to increase plant produc-tivity (8, 22), two enzymatic pathways exist for ammoniaassimilation: the glutamate dehydrogenase (GDH; EC 1.4.1.3)pathway in which ammonia is available at a high level (16, 35)and the glutamine synthetase (GS; EC 6.3.1.2)-glutamatesynthase (GOGAT [glutamine 2-oxoglutarate amidotrans-ferase]; EC 1.4.1.13) pathway in which ammonia availability islow, for example, during N2 fixation (2, 4, 7, 8, 35). TheGS-GOGAT pathway is particularly important for all free-living N2-fixing bacteria for growth under ammonia-deficientconditions (3, 25), but for symbiotic bacteria, e.g., Rhizobiumspp. within root nodules, the pathway may remain repressed tohelp ammonia export (11, 24). The importance of the GS-GOGAT pathway was first demonstrated by Nagatani et al.(19), who observed that GOGAT-deficient mutants of Kleb-siella pneumoniae showed an ammonia assimilation (Asm-)defect; they failed to fix N2 (Nif-) and utilize a number ofamino acids as N sources. GOGAT-deficient mutants ofN2-fixing bacteria showing the Asm- phenotype have beendescribed for A. brasilense (2), Bradyrhizobium japonicum (21),Rhizobium meliloti (13, 15, 24), and Azorhizobium sesbaniaeORS571 (11). It is not yet very clear why GOGAT deficiencywould lead to pleiotropic N-assimilation defects (Asm -) in themutants. An abnormally high concentration of glutamine maycause adenylylation of GS, and repression of sensitive operonsin GOGAT - mutants has been suggested (2). The GOGATstructural gene of Escherichia coli has been cloned (5) andsequenced (23), and it showed that the cx- and (-subunits ofGOGAT are determined by the gltB and gltD genes of thegltBDF operon (5). Recent studies with the glt operon of E. coli(5) and that ofK pneumoniae (14) indicate that the pleiotropicAsm - defect may be due to the gltF product, which plays an

* Corresponding author.

undefined role in ntr regulation. Any role of a gitF-like gene innonenteric bacteria remains unknown. GOGAT of A.brasilense has already been well characterized (26, 34), and itscx- and (-subunits are very similar to those of E. coli in size.Isolated a-subunits of GOGAT from A. brasilense and E. colialso showed a high degree of homology in their N-terminalsequences (34), but the same deduced sequences from theirstructural genes could not be compared, as the sequence of theglt gene of A. brasilense is not known. In our study with A.brasilense RG, for an understanding of regulation of N2fixation and N metabolism, we isolated a number of Nif- andAsm - mutants. One of these mutants (N12) showing a Nif-Asm - phenotype also lacked GOGAT activity. We report herethe cloning of the GOGAT structural gene (glt) ofA. brasilenseRG and partial sequencing of the gltB gene to determine itspromoter structure, deduced N-terminal amino acid sequence,and the locus of the mutation in N12.A. brasilense RG, a wild-type strain (sp. 81 from the stock of

N. R. Krieg, Blacksburg, Va.) was used for this study. Thebacterium was grown, as described previously (16), in succi-nate-salt medium with no N source (SSM), with 0.1% NH4Cl(SSMN), or with 0.2% of another fixed N source whennecessary (adding 1.6% agar for plates). Mutagenesis of thebacterium by N-methyl-N-nitro-N'-nitrosoguanidine was car-ried out essentially by the procedure of Miller (17). Forenrichment of the Nif- mutants, aliquots from the stocks ofmutants were added to SSM (optical density, 0.3), the cellswere allowed to grow microaerobically (optical density, 0.6),and ampicillin and cycloserine were added to the growingculture at final concentrations of 2 mg and 50 ,ug ml-',respectively. After lysis of the wild-type Nif+ cells (in about 4h of further microaerobic incubation), enriched Nif- cellswere collected from the lysed culture, washed, resuspended insaline, and spread on SSM plates containing 0.003% yeastextracts (SSM-0.003% YE). Tiny colonies (Nif-) were se-lected from large colonies (Nif+) and replica plated on SSMN

8024

on Septem

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

.org/D

ownloaded from

on S

eptember 23, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

on Septem

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

.org/D

ownloaded from

NOTES 80)25

TABLE 1. Specific activities of ammonia, nitrate, and dinitrogen assimilatory pathway enzymes in A. brasilense RG, its mutants, andcomplemented N 12

Sp act (nmol min mg of protein )A.

brasilense N source" Nitrogenase" GSstrain" GOGAT GDHI NAR NIR

Mn2+ Mn2+ + Mg2+

RG NH4C1 0.01 1,621 814 113 246 35.5 t).1Glutamine 6.4 2,626 1,696 60 18 114 9.3Glutamate 169 3,0)46 1,918 38 16 NT' NTNo fixed N 175

N 12 NH4Cl NT 837 20)0 5.21 147 46 NTNo fixed N 0).23

N12Glu I NH4C1 NT NT NT 5.2 NT 37 NTGlutamine NT NT NT NT NT 37.7 0).2No fixed N 5.5

N12(pMGI-2) NH4C1 NT 1,643 817 108 203 36.3 NTGlutamine NT NT NT NT NT 113 8.t)No fixed N 189

N12(pMG3-3) NH4C1 NT 1,662 831 1M1 242 35.6 NTGlutamine NT NT NT NT NT 112 8.1No fixed N 187

"The strains N12(pMGI-2) and N12(pMG3-3) are described in Fig. 1; N12GluI (GOGAT nitrate ) is an isolate of N12 growing faster on SSM-0.2% glutamateplates and showing nearly normal-size colonies like those of N12 (see the text).

" Cells were grown aerobically in SSM supplemented with nitrogen sources (weight/volume): NH4Cl (t).101), glutamine (0.2%), and glutamic acid (0.2%). No fixedN, cells grown microaerobically for nitrogenase assays only, as described in footnote c.

' For nitrogenase assay, cells were pregrown aerobically with 0.1 NH4C1 as the N source followed by microaerobic growth for 18 h from an initial optical densityof ().5 in 20) ml of SSM containing 0.05% agar in a 150-ml conical flask (27). Results are given in nanomoles of C2H4 formed per hour per milligram of cell protein.

" GS (3), GOGAT (18), GDH (16), nitrate reductase (NAR) (32), nitrite reductase (NIR) (29), and nitrogenase (27) assays were performed, following the publishedmethods. For NAR and NIR induction, 11) mM NaNO3 was added to exponentially growing cells at an optical density of about 1.0 for the former and ().5 mM NaNO,was added for the latter and further incubation was continued (aerobically) for 3 h at 32°C before collection of cells for assay. When RG (wild-type) cells were grownaerobically in SSM plus 0.01% NH4Cl, (low ammonium), the specific activity of GOGAT was reduced to 25; when grown in SSM plus 0). 1% KNO3, the specific activitiesof nitrate and nitrite reductases were 16.7 and 2.1, respectively.

" NT, not tested.

and SSM-0.003% YE agar plates. Nif- mutants yieldednormal colonies on SSMN but only tiny colonies on SSM-0.003% YE. All Nif- mutants were further screened for theirability to grow on a mixture of arginine plus proline as the Nsource (i.e., on SSM-ArgPro plates). Mutants pleiotropicallydefective in N assimilation were selected by their normal-sizecolonies on SSMN but by very tiny colonies on SSM-ArgProand SSM-0.003% YE plates.A. brasilense RG can assimilate various inorganic and or-

ganic nitrogenous compounds as the sole source of N forgrowth (31). By our selection procedure, a pleiotropic Nassimilation-defective mutant, N12, that failed to utilize any ofthese compounds (e.g., glutamine, asparagine, glutamate, as-partate, arginine, proline, alanine, urea, nitrate, dinitrogen, ora low concentration of ammonium [<0.01% NH4Cl]) as thesole N source for growth was isolated. This mutant (N12)showed a phenotype very similar to that of Asm- mutants ofK pneumoniae (19), E. coli (25), and A. brasilense (2) describedpreviously that had very little or no GOGAT activity. There-fore, ammonia assimilation pathway enzymes in A. brasilenseRG (wild type) and N12 were assayed. As shown in Table 1,the GOGAT activity of N12 was 22-fold less than that of thewild type. The specific activity of GOGAT in RG was maximalwhen grown with 0.1% NH4Cl; lower activities were observedwhen it was grown with 0.01% NH4Cl, glutamate, or glutamineas the sole nitrogen source (Table 1). Compared with activitiesin the wild type, some reductions of the GS and cold-labileGDH activities were observed in N12; lower activities of thesetwo enzymes were also noticed with other Asm - mutants (2).

GOGAT mutants ofA. brasilense described earlier by Bani etal. (2) showed N assimilation defects (Asm- phenotype), butunlike N12, they could assimilate glutamine and glutamate asan N source for growth. In fact, GOGAT mutants ofnitrogen-fixing bacteria do show diverse abilities to utilize Nsources for growth. GOGAT- mutants of R. meliloti (13, 15,24) and B. japonicum (21) obtained by chemical or UVmutagenesis could use histidine, arginine, or proline but notammonia at any concentration for growth. On the other hand,glt insertion mutants of R. meliloti (33) could grow with prolineor asparagine but not with ammonium, aspartate, arginine, orhistidine. In the case of stem-nodulating symbiotic A. sesbaniaeORS571, well-defined Tn5-induced GOGAT- mutantsshowed a Nif- Asm- phenotype with the ability to grow on

glutamine, asparagine, or aspartate as the N source (11).Therefore, one would tend to believe that the physical locationof the mutation in the glt locus might be responsible for thisvariable Asm - phenotype.

Although N12 failed to assimilate many N sources forgrowth, we tested only the enzymes needed for the metabolismof two of these compounds: dinitrogen and nitrate. No nitro-genase activity could be detected in N12 under conditions ofderepression that expressed the enzyme in the wild type (Table1). Reduction of nitrate in A. brasilense RG can be carried outby two enzymes-assimilatory and dissimilatory nitrate reduc-tases (20). Growth of A. brasilense with nitrate as the sole Nsource depends on both the assimilatory nitrate and nitritereductases. Presence of the dissimilatory nitrite reductase in A.brasilense is strain dependent; the enzyme is absent in A.

VOL. 175, 1993

on Septem

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

.org/D

ownloaded from

8026 NOTES

DNA Frc gments Plosmi;ds: Complementation:

B B EI

1.3 3.1 3.4 1.0

EI_

EL

E

P PB

1.14 0.8 0.9

P-I

BP 'E"I

2.5 0.2 0.8

B P

L.JP8 0.9

0.28

FIG. 1. Physical map of A. brasilense RG glt locus, cloned glt regions, and their abilities to complement N12. The topmost line indicates the27.1-kb EcoRI insert of cosmid clone pMG1-198 in which various restriction sites and approximate distances (in kilobases) are shown. The linesbelow indicate the various regions of the 27.1-kb insert cloned. On the right-hand side of each line, the name of the plasmid clone and the abilityof the cloned DNA fragment to complement N12 are indicated. Restriction fragments were subcloned in pLAFRI (pMG1 series) or in pLAFR3(pMG3 series) with E. coli S17.1 as the host for their conjugal transfer to N12 for complementation analysis. Subcloning in pUC19 vector (pMG19series) was done for nucleotide sequencing. The small arrows under PB and ES DNA fragments indicate the direction and extent of the sequencedetermined. Sequencing was done in both directions by the dideoxy chain termination method using a Sequenase version 2.0 kit supplied by U.S.Biochemical Corp., Cleveland, Ohio. Ordered deletion subclones were made by using a double-stranded nested deletion kit supplied by PharmaciaP-L Biochemicals, Inc., Milwaukee, Wis., following the manufacturer's protocol. The approximate limits of the analogous gltB gene of E. coli (23)are shown (hatched arrow). Restriction sites: E, EcoRI; B, BamHI; P, PstI; S, Sall. +a, complementation by some recombinational events (see thetext); - , no complementation when transferred as an insert in pLAFR3; NT, not tested.

brasilense RG (15a) but present in A. brasilense Sp7 (36). Theassimilatory nitrate and nitrite reductases are repressed by thegrowth of the bacterium in high levels of ammonia, but thedissimilatory nitrate reductase is not susceptible to ammoniarepression; assimilatory and dissimilatory enzymes are allinduced to high levels by nitrate in the presence of glutamineinA. brasilense RG (Table 1). N12 and A. brasilense RG grownon SSMN with nitrate as the inducer showed nearly the samelow levels of synthesis of dissimilatory nitrate reductase (Table1), and both of them also showed the same degree of chloratesensitivity (data not shown). In the wild-type A. brasilense RG,growth in the presence of glutamine makes it possible for thenitrate to induce fully the assimilatory nitrate reductase and forthe nitrite to induce the assimilatory nitrite reductase (Table1); the inability of N12 to grow with glutamine as an N sourcemakes such induction experiments unfeasible. It seems that theinability of the mutant N12 to grow with nitrate as the N sourceis mainly due to failure of the nitrate or nitrite to induce anyassimilatory nitrate or nitrite reductase that converts nitrate toammonia, on which cell growth depends. To confirm the statusof the nitrite reductase in N12, we isolated a few colonies ofN12 with the ability to grow faster on SSM-glutamate. Afterpurifying a number of these colonies (N12Glu1), they werealso found to grow well in glutamine but were GOGAT- and

nitrate negative. When the nitrate and nitrite reductase activ-ities of N12Glul were measured after the bacteria had grownin SSM-glutamine in the presence of appropriate inducers, theformer was found in the cell extracts at the same level as thedissimilatory nitrate reductase and the latter at the uninducedbasal level of the assimilatory nitrite reductase in the wild-typebacteria (Table 1).

Cloning of the GOGAT structural gene of A. brasilense RGand its partial sequencing were carried out, following thestandard DNA recombinant techniques as described in refer-ence 28. For this purpose, a genomic library of A. brasilenseRG DNA was constructed with tetracycline-resistant broad-host-range cosmid pLAFRI (1) at the EcoRI site with E. coliS17.1 (30) as the host that can directly transfer the recombi-nant pLAFRI cosmid clone into A. brasilense by mating. Acosmid clone, pMGI-198, containing a 27.1-kb insert of DNAthat could rectify Asm - and Nif- defects when conjugallytransferred into N12 (Fig. 1) was identified from the library.The complemented N12(pMG1-198) strain also showedGOGAT activity like that of the wild type (Table 1). Arestriction site map of the 27.1-kb DNA insert is shown in Fig.1. Cloning of the EcoRI-digested DNA fragments (inpLAFRI) followed by complementation analysis with N12identified the locus of the GOGAT structural gene (git) within

E

8.5

E EI

E

E

9.8pMGl-198

E

+

EI pMGI- 1

pMGI- 2 +

pMGl- 3ES S

0.86

ES S

0.863

-4-4-

pMG3-1

pMG 3-2

pMG3-3

pMGI 9-4

pMGI 9-3

+a

+a

_b

NT

-

177771 1 1-1 x i I I ff 7 r --Z i ff , ;b',rI

p 8 pI I

J. BACTERIOL.

on Septem

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

.org/D

ownloaded from

NOTES 8027

A1 GAAT TCACGGT CCAGGCCCATCTGtGTCATCAAGCGCCTCGGTCACCCGAGGscrsACCacTGCCCGsc TCCCCCCTGTGCA ACGGACTAAGGTCACCcGcT

1 01 GGGGCACGCCTGCTCGTCGACCAeCCGAGCAAGATGCACCAACCTGGCAACGCGTGCT TCGCCGCCGCGCCACAT* CGT GCGCGCCCCCCCTCCGCTGGCT CCGTGT GGGCC"-35.

201 CATCCGCGACGGCCGCGACGCGCGGAGGGCATCCACGCCTACGCCAGGGCGAAGGCCGAGGCTCCGGTCGCCGT TGCGGCTGAATAAGCCAAGCAATCCA

'-10'. "-24' "-12"301 TAAATCCCCGGGTTCCAGGGGTCCATTTCGGCCTGGCGGGGGTGAAGGGGCGAAGTCCCCCTTCCGAACAGACTGCCCCCGTCAAACACACTGACTGAAA

S/DN T T E L N1 G E Q F V A D F R A M A AA T T A N A Y H P

401 C TCCACCGATGACCACCGAGTTGAACCAGGGTGAGCAGTTCGTCGCCC ACTTCCGCGCCAACGCCGCGCGCCCTGACCACTGCCAACGCCCTACAACCCG

E D E N D A C G V Af I A A I D C K P fR S V V E K G I E A L K A V501 GAGGATGAGCACGACGCCTGCGGCGTCGGCTTCATCCCCGATCCACGGCAACCCCGCCGCTCTGGTCAGAAAGCATCGACGCTGAACCG

U HR G YD DA GK T G D G A G I O t V P Q K F F K D H V K V601 TCTGGCACCCG GCG CCCGACGGCAAGACGGGCGACGGTGCGGGCATCCACGTCGGGTGCCGCAGAAGTTCTTCAAGGCCATGTGAAGGT

I G H R A P D H K L A V G O V F L P R I S L D A O E A C R C I V E701 CATCGGCCACCCCGCCGCACAACAAGCTGCCCGGCCAGCTTCCTGCCCGCATCAGCCTGGACGCGCAGGAACCTGCCGCTGCCGTCGAG

T E I L R F G Y Y I Y U RW R V P I N V D801 ACCGAGATCCTGCGCTTCGGCTACTACATCTACGGCTGGCGCCAGGTGCCGATCAACGTCGAC

B 1Ab HTTEL NGEQFVADFRANAA ----- ALTTANA--YNPEDEDACGYVGFIAAIDGKPRRSVVEKGIEALKAVWHNGAVDAD 73

II I I I I II I II I I I II I II 1111 IIEc NTRKPRR-HALSVPVRSGSEVCFPQSLGEVHDNLYDKSLIEEDIGFGL IAN I EGEPSHKVVRTAIHA LARQHRGA I LAD 79

Ab GKTGDGAGIQRRVPOKFFKDNVKVIHREAPDNKLAVGQVFLPRISLDAQEACRCIVETEILRFGYYIYGRYOVPIHVD 151111111 I I 11 I III 11 1 I I III I I III 11 I

Ec GKTGDGCGLLLQKPDRFFRIIVAERLAFRK-NY-AVGNLFLNKDPELAAAARR-IVEEELQRNTL-SIYVGRDVPTHEG 154

FIG. 2. (A) Nucleotide sequence of the 0.863-kb EcoRI-SalI fragment of the gltB gene of A. brasilense RG. The coding strand of the DNA ispresented in a 5'-to-3' direction. Amino acids deduced from the nucleotide sequence and specified by standard one-letter abbreviations areindicated above the first base of each codon. The putative Shine-Dalgarno sequence upstream of the ATG codon is underlined. The potentialNtrA-dependent promoter (-12 and -24 sequences) and sigma-70 promoter (- 10 and -35 sequences) are overlined. ***, potential NifAbinding site (TGT-N12-ACA); , IHF binding site. (B) N-terminal amino acid sequence homology between gltB of A. brasilense RG (Ab) andthat of E. coli (Ec), as deduced from their nucleotide sequences. The N-terminal amino acid sequences (residues 1 to 151 of A. brasilense RG and1 to 154 of E. coli) are aligned; the residues connected by vertical lines are identical, and dashes in the sequences represent gaps introduced toimprove sequence homology. The predicted cleavage site for generating the mature ox-subunit of GOGAT (34) is indicated by vertical arrows.There is about 43.5% identity in the N-terminal amino acids of the two mature a-subunits of GOGAT.

the 7.5-kb insert DNA of pMG1-2 (Fig. 1). As shown in Table1, N12(pMG1-2) was GOGAT+ and Nif+ like the wild type,and it could also use all normal N sources for growth (Asm+).The 7.5-kb DNA fragment was recloned in tetracycline-

resistant IncP1 cosmid pLAFR3 (obtained from N. Keen,Riverside, Calif.) containing multiple cloning sites, and a new

restriction site map of the DNA fragment was constructed. Asshown in Fig. 1, the two small PstI-PstI 1.18-kb and BamHI-PstI 0.9-kb DNA fragments subcloned in pLAFR3 also couldcomplement the mutant N12 by mating although at a lowfrequency. Only a few among the tetracycline-resistanttransconjugants obtained after mating of S17.1(pMG3-2) or

S17.1(pMG3-3) with N12 could grow on SSM-0.003% YE or

SSM-ArgPro plates, and they showed an Asm+ GOGAT+phenotype. [The enzyme activities of only N12(pMG3-3) are

shown in Table 1.] The N12(pMG3-2) and N12(pMG3-3)transconjugants showing the wild-type characteristics are ex-

tremely unlikely to be wild-type revertants of N12, since theAsm+ GOGAT+ reversion frequency is very low (<10-7) andwe failed to detect any such revertants in control plates.Moreover, the Asm+ GOGAT+ transconjugants were very

specifically obtained only when the donor's plasmid DNAcontained the 0.9-kb BamHI-PstI sequence. Since GOGATsubunits of A. brasilense are known to be 140 and 56.5 kDa(34), the observed complementation by the small DNA frag-ment (0.9-kb PstI-BamHI fragment) could only be due to somerecombinational rectification of the mutant gene in N12. This

was confirmed by curing of the resident plasmid of theGOGAT+ Asm+ N12(pMG3-2) and N12(pMG3-3) transcon-jugants by the RP4 plasmid (a kanamycin- and tetracycline-resistant IncPl, broad-host-range plasmid). The cured kana-mycin-resistant N12 strains did show the expected Asm+GOGAT+ phenotype where RP4 was the only detectableresident plasmid. These results most probably mean that thelocus of the genetic lesion that caused the GOGAT- Asm-phenotype in N12 is within the 0.9-kb BamHI-PstI fragment.To confirm that the GOGAT structural gene is in the 7.5-kb

DNA fragment, the small 0.28-kb PstI-BamHI fragment wassequenced and compared with the homologous gltB region ofE. coli. The nucleotide sequences and the deduced amino acidsequences of the two gltB genes have very high homology inthis region. A 50% identity at the level of amino acids and56.7% identity at the level of nucleotides were observed inthese two partial gene sequences. This homology enabled us toalign the promoter and the translation start site of the gltBgene of A. brasilense RG (in the 0.863-kb EcoRI-SalI DNAfragment) with the corresponding gltB region of E. coli (23).The deduced N-terminal amino acid sequences of the gltB genefrom A. brasilense RG and E. coli (Fig. 2A and B) showed that,starting from the cysteine residue at position 37 in the formerand position 43 in the latter, there exists a strong homologybetween the two. This is expected because it is already knownthat the first 20 N-terminal amino acid sequences of thea-subunit of GOGAT of the two bacteria are quite similar, and

VOL. 175, 1993

on Septem

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

.org/D

ownloaded from

8028 NOTES

both the sequences start with cysteine (10, 34). In fact, the20-amino-acid N-terminal sequence of the o-subunit of A.brasilense Sp7 GOGAT reported by Vanoni et al. (34) exactlymatches the deduced N-terminal amino acid sequence of A.brasilense RG GOGAT, starting with the cysteine residue atposition 37, as shown in Fig. 2B.The promoter sequences of the gltB gene ofA. brasilense and

E. coli differed considerably. Whereas the A. brasilense pro-moter appears to have sigma-70 as well as sigma-54 recognitionsites (Fig. 2A), E. coli has only the former. Surprisingly, thegltB and nif promoters of A. brasilense show considerablesimilarity. The NtrA recognition domain CTGGCA-CG5-ATGCAG in the nifH promoter of A. brasilense (7) is similar tothe corresponding domain, CTGGCG-G3-GTGAAG, in thegltB promoter. In addition, the presence of TGT-NI2-ACA(NifA consensus: TGT-N,l -ACA) and AAGCAATCCATAA(IHF consensus: WATCAA-N4-TTR; W = A or T; and R = Aor G) sequences in the gltB promoter are suggestive ofpotential NifA and IHF binding sites, respectively (12). Al-though it is not very unusual to find multiple promoters inhighly regulated genes (6), it is not clear why this should beencountered in the glt operon of A. brasilense, which is notknown to be highly regulated. Obviously, we have to learnmore about the glt operon of A. brasilense, particularly itsinterrelationship with other regulatory genes, e.g., lrp (9) orgltF of enterobacteria, that may also control N metabolism inthe bacterium. Study in this direction should also take intoaccount our results showing that the Asm- phenotype iscorrelated with a mutation within the gltB gene ofA. brasilense.

Nucleotide sequence accession numbers. The nucleotidesequences for the 5'-terminal and internal coding regions havebeen given the accession numbers X71090 and X71632, respec-tively, by the EMBL Data Bank.

We thank the Council of Scientific and Industrial Research, Gov-ernment of India, for supporting us with an Emeritus ScientistFellowship to S.G. and a Senior Research Fellowship to A.K.M.We are indebted to Sudip Chattopadhyay for supplying us his

cosmid pLAFRI library of A. brasilense RG. We thank N. Keen(Riverside, Calif.) for giving us cosmid pLAFR3. Expert technicalassistance of Asim Banerjee and A. Ali Molla is gratefully acknowl-edged. We also thank M. K. Maiti, N. Das, and A. Basu for theirhelpful suggestions and critical comments.

REFERENCES

1. Ausubel, F. M. 1982. Construction of a broad host range cosmidcloning vector and its use in the genetic analysis of Rhizobiummutants. Gene 18:289-296.

2. Bani, D., C. Berberio, M. Bazzicalupo, F. Favilli, E. Gallori, andM. Polsinelli. 1980. Isolation and characterization of glutamatesynthase mutants of Azospirillum brasilense. J. Gen. Microbiol.119:239-244.

3. Bender, R. A., K. A. Janssen, A. D. Resnick, M. Blumenberg, F.Foor, and B. Magasanik. 1977. Biochemical parameters of glu-tamine synthetase from Klebsiella aerogenes. J. Bacteriol. 129:1001-1009.

4. Bozouklian, H., C. Fogher, and C. Elmerich. 1986. Cloning andcharacterization of ginA gene of Azospirillum brasilense Sp7. Ann.Inst. Pasteur/Microbiol. (Paris) 137B:3-18.

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

6. Collado-Vides, J., B. Magasanik, and J. D. Gralla. 1991. Controlsite location and transcriptional regulation in Escherichia coli.Microbiol. Rev. 55:371-394.

7. de Zamaroczy, M., F. Delorme, and C. Elmerich. 1989. Regulationof transcription and promoter mapping of the structural genes fornitrogenase (nifHDK) of Azospirillum brasilense Sp7. Mol. Gen.Genet. 220:88-94.

8. Elmerich, C. 1984. Molecular biology and ecology of diazotrophsassociated with non-leguminous plants. Biotechnology 2:967-978.

9. Ernsting, B. R., M. R. Atkinson, A. J. Ninfa, and R. G. Matthews.1992. Characterization of the regulon controlled by leucine-responsive regulatory protein in Escherichia coli. J. Bacteriol.174:1109-1118.

10. Gosset, G., E. Merino, F. Recillas, G. Oliver, B. Becerril, and F.Bolivar. 1989. Amino acid sequence analysis of the glutamatesynthase enzyme from Escherichia coli K-12. Protein SequencesData Anal. 2:9-16.

11. Hilgert, U., J. Schell, and F. J. de Bruijn. 1987. Isolation andcharacterization of Tn5-induced NADPH-glutamate synthase(GOGAT-) mutants of Azorhizobium sesbaniae ORS571 andcloning of the corresponding g/t locus. Mol. Gen. Genet. 210:195-202.

12. Hoover, T. R., E. Santero, S. Porter, and S. Kustu. 1990. Theintegration host factor stimulates interaction of RNA polymerasewith NIFA, the transcriptional activator for nitrogen fixationoperons. Cell 63:11-22.

13. Kondorosi, A., Z. Svab, G. B. Kiss, and R. A. Dixon. 1977.Ammonium assimilation and nitrogen fixation in Rhizobiuni me-liloti. Mol. Gen. Genet. 151:221-226.

14. Kuczius, T., T. Eitinger, R. D'Ari, H. Castroph, and D. Kleiner.1991. The g/tF gene of Klebsiella pneumoniae: cloning and initialcharacterization. Mol. Gen. Genet. 229:479-482.

15. Lewis, T. A., R. Gonzalez, and J. L. Bostford. 1990. Rlhizobiuimmeliloti glutamate synthase: cloning and initial characterization ofthe glt locus. J. Bacteriol. 172:2413-2420.

15a.Mandal, A. K., and S. Ghosh. Unpublished data.16. Maulik, P., and S. Ghosh. 1986. NADPH/NADH dependent

cold-labile glutamate dehydrogenase in Azospirillum brasilense:purification and properties. Eur. J. Biochem. 155:595-60)2.

17. Miller, J. H. 1972. Experiments in molecular genetics, p. 431. ColdSpring Harbor Laboratory, Cold Spring Harbor, N.Y.

18. Miller, R. E., and E. R. Stadman. 1972. Glutamate synthase fromE. coli: an iron-sulfide flavoprotein. J. Biol. Chem. 247:7409-7419.

19. Nagatani, H., M. Shimizu, and R. C. Valentine. 1971. Themechanism of ammonia assimilation in nitrogen fixing bacteria.Arch. Microbiol. 79:164-175.

20. Neyra, C. A., and P. VanBerkum. 1977. Nitrate reduction andnitrogenase activity in Spirillum lipoferumn. Can. J. Microbiol.23:306-310.

21. O'Gara, F., S. Manian, and J. Meade. 1984. Isolation of an Asmmutant of Rhizobium japoniicum defective in symbiotic nitrogenfixation. FEMS Microbiol. Lett. 24:241-245.

22. Okon, Y. 1985. Azospirillumn as a potential inoculant for agricul-ture. Trends Biotechnol. 3:223-228.

23. Oliver, G., G. Gosset, R. Sanchez-Pescador, E. Lozoya, L. M. Ku,N. Flores, B. Becerril, F. Valle, and F. Boliver. 1987. Determina-tion of the nucleotide sequence for the glutamate synthase struc-tural genes of Escherichia coli K-12. Gene 60:1-11.

24. Osbourne, M. S., and E. R. Singer. 1980. Ammonia assimilation inRhizobium meliloti. J. Bacteriol. 143:1234-1240.

25. Pahel, G., A. D. Zelenetz, and B. M. Tyler. 1978. g/tB gene andregulation of nitrogen metabolism by glutamine synthetase inEscherichia coli. J. Bacteriol. 133:139-148.

26. Ratti, S., B. Curti, G. Zanetti, and E. Galli. 1985. Purification andcharacterization of glutamate synthase from Azospirillumbrasilense. J. Bacteriol. 163:724-729.

27. Rauth, S., and S. Ghosh. 1981. Effect of dimethyl sulfoxide onderepression of nitrogenase in Spirillum lipoferum. FEBS Lett.126:77-80.

28. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring Harbor Labo-ratory, Cold Spring Harbor, N.Y.

29. Santero, E., F. Luque, J. R. Medina, and M. Tortolero. 1986.Isolation of ntrA-like mutants of Azotobacter vinclandii. J. Bacte-riol. 166:541-544.

30. Simon, R., U. Priefer, and A. Puhler. 1983. A broad host rangemobilization system for in vivo genetic engineering: transposonmutagenesis in gram negative bacteria. Biotechnology 1:784-791.

31. Tarrand, J. J., N. R. Krieg, and J. Dobereiner. 1978. A taxonomic

J BACVI I Kl()OL.

on Septem

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

.org/D

ownloaded from

VOL. 175, 1993 NOTES 8029

study of the Spirillum lipoferum group, with descriptions of a new

genus, Azospirillum gen. nov. and two species, Azospirillum lipof-erum (Beijerinck) comb. nov. and Azospirillum brasilense sp. nov.

Can. J. Microbiol. 24:967-980.32. Taukdarian, A., and C. Kennedy. 1986. Regulation of nitrogen

metabolism in Azotobacter vinelandii: isolation of ntr and ginAgenes and construction of ntr mutants. EMBO J. 5:399-407.

33. Thomas, A. L., R. Gonzalez, and J. L. Bostford. 1990. Rhizobiummeliloti glutamate synthase: cloning and initial characterization ofthe glt locus. J. Bacteriol. 172:2413-2420.

34. Vanoni, M. A., A. Negri, G. Zanetti, S. Ronchi, and B. Curti. 1990.Structural studies on the subunits of glutamate synthase fromAzospirillum brasilense. Biochim. Biophys. Acta 1039:374-377.

35. Westby, C. A., C. S. Enderlin, N. A. Steinberg, C. M. Joseph, andJ. C. Meeks. 1987. Assimilation of NH4 by Azospirillum brasilensegrown under nitrogen limitation and excess. J. Bacteriol. 169:4211-4214.

36. Zimmer, W., M. P. Stephan, and H. Bothe. 1984. Denitrification byAzospirillum brasilense Sp7. I. Growth with nitrite as respiratoryelectron acceptor. Arch. Microbiol. 138:206-211.

on Septem

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

.org/D

ownloaded from

AUTHOR'S CORRECTION

Isolation of a Glutamate Synthase (GOGAT)-Negative, Pleiotropically NUtilization-Defective Mutant of Azospirnllum brasilense: Cloning and

Partial Characterization of GOGAT Structural GeneASIM KUMAR MANDAL AND SUDHAMOY GHOSH

Department of Biochemistry and Centre for Plant Molecular Biology, Bose Institute,AJCB Centenary Building, Calcutta 700 054, India

Volume 175, no. 24, p. 8024-8029: We reported a partial nucleotide sequence of the Azospirillum brasilense RG gltB locus thatincluded 453 bp of the gltB coding region and 410 bp of the 5'-flanking DNA. Unfortunately, we were unaware of a publicationby Pelanda et al. that should have been cited (R. Pelanda, M. A. Vanoni, M. Perego, L. Piubelli, A. Galizzi, B. Curti, and G.Zanetti, J. Biol. Chem. 268:3099-3106, 1993). In this publication, the entire sequence of the glt locus from another strain of A.brasilense, Sp7, was reported. This sequence included 852 bp of DNA 5' to the gltD coding sequence, the 1443-bp gltD sequence,and a 141-bp intergenic region followed by the 4545-bp coding sequence for gltB and 295 bp of 3'-flanking DNA. Within thealigned sequences of the corresponding DNA from the t-wo A. brasilense strains there is 98.5% identity in the gltB open readingframes and 84.6% identity in the DNA upstream of gltB.

3092