APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 1992, p. 1711-17180099-2240/92/051711-08$02.00/0Copyright C) 1992, American Society for Microbiology

Excretion of Ammonium by a nifL Mutant ofAzotobacter vinelandiiFixing Nitrogen

ANIL BALI,t GONZALO BLANCO,4 SUSAN HILL, AND CHRISTINA KENNEDY§*

Nitrogen Fixation Laboratory, AFRC Institute ofPlant Science Research, University of Sussex,Brighton BN1 9RQ, United Kingdom

Received 13 September 1991/Accepted 2 February 1992

A mutation in the gene upstream of nifA in Azotobacter vinelandii was introduced into the chromosome toreplace the corresponding wild-type region. The resulting mutant, MV376, produced nitrogenase constitutivelyin the presence of 15 mM ammonium. When introduced into a niJH-lacZ fusion strain, the mutation permitted1-galactosidase production in the presence of ammonium. The gene upstream of nifA is therefore designatednifL because of its similarity to the Klebsiella pneumoniae nifL gene in proximity to nifA, in mutant phenotype,and in amino acid sequence of the gene product. The A. vinelandii nifL mutant MV376 excreted significantquantities of ammonium (-10 mM) during diazotrophic growth. In contrast, ammonium excretion duringdiazotrophy was much lower in a K. pneumoniae niJL deletion mutant (maximum, 0.15 mM) but significantlyhigher than in NifL+ K. pneumoniae. The expression of the A. vinelandii nifA gene, unlike that of K.pneumoniae, was not repressed by ammonium.

Free-living diazotrophs fix dinitrogen sufficient for theirown needs and do not generally excrete significant amountsof ammonium into their environment; fixed nitrogen isreleased after death and lysis of bacteria. Attempts to induceammonium excretion have up to now centered on physio-logical suppression or genetic manipulation of the enzymesinvolved in ammonium assimilation. Treatment of cyanobac-teria with L-methionine-DL-sulfoximine, an inhibitor of glu-tamine synthetase (GS), resulted in the excretion of 0.3 to 7mM NH4' into the growth medium (32, 33, 36). Anabaenamutants resistant to L-methionine-DL-sulfoximine or theNH4' analog ethylenediamine excreted up to 1.6 mM NH4'(35, 43, 45). Among eubacteria, ammonium excretion was

reported to occur in mutants of Klebsiella pneumoniaedefective in ammonium assimilation or transport (1, 8, 41), inmutants of Rhodobacter capsulatus and Azospirillum brasil-iense altered in the production of GS (27, 48), in methyl-amine-resistant mutants of Azotobacter vinelandii (17), andin ethylenediamine-resistant mutants of A. brasiliense (9,27).Although little is known about the physiology and genetics

of ammonium repression of nitrogenase synthesis in Ana-baena, Rhodobacter, or Azospirillum species, these path-ways and mechanisms have been extensively characterizedin the free-living diazotroph K pneumoniae and, to a lesserextent, in A. vinelandii. Of central importance in bothorganisms (and other gram-negative nitrogen-fixing bacteria)is that a positively acting regulatory protein, NifA, is re-

quired to activate transcription of the other nif genes neces-

sary for nitrogenase structure and activity. In K pneumo-niae, ammonium represses nitrogenase synthesis bypreventing either the activity or the synthesis of NifA by twoseparate mechanisms. The nifA gene is adjacent to and

* Corresponding author.t Present address: Department of Biology, Washington Univer-

sity, St. Louis, MO 63130.i Present address: Department of Microbiology, Faculty of Biol-

ogy, University of Seville, Seville, Spain.§ Present address: Department of Plant Pathology, College of

Agriculture, University of Arizona, Tucson, AZ 85721.

downstream of nifL, and these two genes are coexpressed(see reference 13 for a review). The NifL protein binds toand inactivates NifA when ammonium is present even atrelatively low levels (greater than -5 ,uM). At higher levelsof ammonium (greater than -200 ,uM), the expression of thenifL4 operon does not occur, so the NifA protein is notsynthesized (29). Transcription of nifLA requires the phos-phorylated NtrC protein, which is dephosphorylated andhence inactive in cells grown with ammonium. Thus, therepression of nitrogenase synthesis by ammonium occurs attwo levels in K pneumoniae, inactivation of NifA by NifLand prevention of nifA expression by dephosphorylatedNtrC.

Nitrogen fixation in A. vinelandii is complicated by thepresence of three biochemically and genetically distinctnitrogenase enzymes, each of which is synthesized underdifferent conditions of metal supply (5). The work presentedhere concerns the regulation of the conventional molybde-num nitrogenase, whose subunits are encoded by the nif-HDK genes and which is similar to the enzyme purified froma number of other nitrogen-fixing organisms. The nifHDKgenes are located in a large cluster of nif genes whichincludes, in order, nipIDKTYENXUSVWZMF, somewhatsimilar to the nif gene cluster ofK pneumoniae (19). As inK pneumoniae and several other diazotrophs, the expres-sion of niplDK requires the NifA protein as an activator (2,38). The nifA gene ofA. vinelandii is located upstream of thenifQB genes in a cluster which is not linked to the major nifgenes (2, 20). This organization is different from that in Kpneumoniae, in which the nifQBAL genes are adjacent tonifF. Although ntrC is present in A. vinelandii as part of a

glnA-ntrBC gene cluster, as in K pneumoniae, its product,NtrC, is not required for the expression of nitrogen fixationgenes (46, 47). Therefore, the mechanism of ammoniumrepression of nif gene expression in A. vinelandii is unlikelyto include dephosphorylation of NtrC as one aspect.DNA sequencing of the nifA region of A. vinelandii

revealed a gene whose product was similar to NifA of Kpneumoniae (2). The DNA sequence of 300 bp upstream ofnifA contained a partial open reading frame encoding a

protein with homology to the C-terminal regions of the nifL

1711

Vol. 58, No. 5

on June 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

and ntrB gene products of K. pneumoniae. We now describethe cloning of the entire gene upstream of nifA in A.vinelandii and the phenotype resulting from mutations in thisgene. High nitrogenase activity was found in mutants grownin the presence of ammonium, as was high ,B-galactosidaseactivity in nifL mutant derivatives carrying nifll-lacZ fu-sions and grown in the presence of ammonium (15 mM). Thisgene has been designated nifL because of its proximity tonifA and the phenotype of mutants which are somewhatsimilar to nonpolar nifL mutants of K. pneumoniae withrespect to ammonium repression of nif gene expression.However, while K. pneumoniae nifL mutants are not re-

pressed at low to medium levels of a fixed nitrogen supply(29), A. vinelandii nifL mutants express nifHDK and synthe-size nitrogenase in high levels of ammonium. Furthermore, a

nifL mutant of A. vinelandii excreted up to 10 mM ammo-nium during nitrogen fixation. In contrast, a nifL mutant ofK. pneumoniae excreted very little ammonium during diaz-otrophic growth.

MATERIALS AND METHODS

Strains, plasmids, and bacteriophage. The A. vinelandii,Escherichia coli, and K. pneumoniae strains, the plasmids,and the bacteriophage used in this work and their origins arelisted in Table 1.Media and growth conditions. Strains ofA. vinelandii were

grown aerobically at 30°C in Burk's sucrose medium asdescribed previously (46). Liquid 25-ml cultures, containedin 125-ml flasks, were incubated on a rotary shaker (180rpm). Competence medium (CM) was Burk's sucrose me-dium prepared without the addition of Fe and Mo salts.Luria-Bertani medium was used for growing E. coli. Antibi-otics for selection of resistance genes on plasmids or ingenomic transformants were added at the concentrationspreviously reported (38).

K. pneumoniae strains were grown or derepressed fornitrogenase at 28°C. Anaerobiosis was maintained by bub-bling either the growth medium with 1% CO2 in N2 (vol/vol)or the derepression medium with N2 or Ar. Diazotrophicgrowth cultures were contained in 15 ml of NFDM (3) in25-ml bottles. Derepression cultures were set up with cellsobtained from an 18-h culture grown in NFDM supple-mented with excess (NH4)2SO4 (15 mM) and placed in 45 mlof NFDM in 100-ml bottles. Cultures were centrifuged, andthe bacteria were resuspended (to an optical density at 600nm of about 1.5) in 8-ml aliquots of either N2- or Ar-bubbledNFDM. Glucose, the carbon and energy source in NFDM,was replaced by glycerol (60 mM) plus sodium fumarate (100mM) in some derepression experiments.

Isolation of an A. vinelandii genomic fragment carrying theregion upstream of nifA. About 300 bp of the 3' end of a geneencoding a protein with partial sequence homology to the Ctermini of the nifL and ntrB gene products of K. pneumoniaehad been cloned in pDB150 and sequenced by Bennett et al.(2). To isolate and clone the entire upstream gene, we

screened plaques from a lambda library of A. vinelandiigenomic DNA for hybridization to a DNA probe labelledwith [32P]dCTP. This probe was a 1.2-kb SalI-KpnI fragmentfrom pDB150 which contained the 3' end of the nifL- or

ntrB-like gene and the 5' end of nifA (Fig. 1) (2). Anapproximately 12.5-kb EcoRI fragment was identified in theinsert DNA prepared from the progeny of one hybridizingplaque; it was subcloned into pBR325 to yield pAB21. (This12.5-kb fragment corresponded in size to a genomic fragmentpreviously reported to hybridize to a nifA probe; a nifA::TnS

mutant, MV3, contained a new nifA-hybridizing fragment ofabout 19 kb [38].) A restriction map of the insert in pAB21showed that one end of the 12.5-kb EcoRI fragment hadrestriction sites corresponding to those reported for pDB150(2). Part of this fragment and the nifA-nifB region andsubclones derived for this work are shown in Fig. 1.Methods for blotting and screening of the lambda library

and for cloning ofDNA fragments were standard proceduresdescribed by Sambrook et al. (37) and done in accordancewith instructions provided by suppliers of restriction en-zymes.

Construction of a nifA-lacZ fusion. Transposon TnS-B21carrying lacZ and a gene encoding tetracycline resistance(Tcr) was introduced into E. coli carrying pAB8 (Fig. 1 andTable 1) by infection with A::TnS-B21 (42). Plasmid deriva-tives carrying Tn5-B21 insertions were isolated and charac-terized as described previously (49).

Transformation of A. vinelandii. Competent cells wereprepared by a simplification of a method described previ-ously (34): instead of being grown in liquid CM prior totransformation, cells were resuspended directly from thesecond round of growth on CM agar into 1 ml of liquid CMplus 16 mM MgSO4 to a density of about 108 cells ml-'.Approximately 50 ,u of resuspended cells was spotted ontoa CM agar plate, and 0.1 to 1 ,ug of plasmid DNA was mixedwith the cells. After incubation at 30°C for 2 days, approxi-mately 5 x 107 cells were transferred to a selective mediumcontaining appropriate antibiotics. Cells not mixed withDNA were plated as controls. Transformants arose at afrequency of 104 to 105 per ,ug of DNA.Enzyme assays. Nitrogenase and 3-galactosidase activities

were measured as described previously (38, 49).Ammonium estimation. Samples of cultures were taken at

different times and centrifuged for A. vinelandii strains orfiltered (through cellulose acetate membranes; pore size,0.45 ,um) forK pneumoniae strains. Appropriate amounts ofsupernatant or filtrate were tested for the presence of am-monium by the indophenol method (3). This consisted of theaddition, in order, of 0.5 ml of phenol-sodium nitroprussidesolution (phenol, 50 g liter-'; sodium nitroprusside, 0.25 gliter-'), 0.5 ml of sodium hypochlorite solution (0.1 M), and2 ml of distilled water. The mixture was incubated for 30 minat room temperature. The A625 was measured, and theammonium concentration was estimated from a standardcurve obtained with ammonium solutions at various concen-trations assayed with the same reagent solutions.

Nucleotide sequence accession number. The EMBL acces-sion number for the A. vinelandii nifL nucleotide sequence isX64832.

RESULTS

Mutations in the gene upstream of nifA in A. vinelandiirelieve ammonium repression of nitrogen fixation. The regionupstream of nifA inA. vinelandii was identified and cloned asdescribed in Materials and Methods. The KIXX cassette,containing the kanamycin resistance (Kmr) gene (aph) fromTn5, was isolated as a 1.5-kb SmaI fragment from pUC4-KIXX. This fragment was ligated with pAB26 which hadbeen digested with SalI and SmaI; the SalI overhang wasblunt ended with Klenow polymerase and deoxynucleotides(Fig. 1). The resulting plasmids, pAB29 and pAB30, with theKIXX cassette inserted in opposite orientations, were trans-formed into two strains of A. vinelandii: wild-type UW136and the nifH-lacZ fusion strain MV101. KMr transformantswere screened for resistance to carbenicillin (Cbr); CbS

1712 BALI ET AL.

on June 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

AMMONIUM EXCRETION IN A. VINELANDII nifL MUTANTS 1713

TABLE 1. Bacterial strains, plasmids, and bacteriophage

Strain, plasmid, or Relevant phenotype, genotype, or feature Source or

bacteriophage reference(s)

StrainsA. vinelandiiUwiUW136MV101MV300MV376MV378MV380

K pneumoniaeUN4357MSal

nifAlrif-1 (Nif+)/if-1 ni/Hi:lacZ-KSSrif-1 ni/A3::TnS-B21rif-i ni/LJ:KIXXrif-1 ni/Hi:lacZ-KSS ni/Ll:KIXXrif-i nifHi:lacZ-KSS nifL2:KIXX

ni/L5037::Mu nif2401 rfb-4002Wild type

2, 40449This work (Fig. 1)This work (Fig. 1)This work (Fig. 1)This work (Fig. 1)

16, 28Nitrogen Fixation

Laboratory col-lection

Sall-Sall fragment of pDB150 and SalI-SphI fragment of pDB160 cloned togetherin pTZ18

SmaI-SphI (Hindlll site from pTZ18) fragment of pAB5 in pACYC177pAB8 cointegrated with pLAFR3 at BamHI sitesnifA::TnS-B21 in pAB812.5-kb EcoRI fragment including the nifL region from the A. vinelandii genomic

library cloned into pBR3251.4-kb KpnI-SphI fragment in pTZ19Like pAB25, but with the fragment in the opposite orientationKIXX cassette inserted between the Sall and SphI sites in the nifL gene in pAB26Like pAB29, but with KIXX in the opposite orientationBamHI-HindIII fragment of pAB5 in pJRD2153-kb SalI fragment ofK pneumoniae nifA in pKT2303-kb SalI fragment ofK pneumoniae nifA in pRK2901.7-kb Sall ni/LA fragment in pUC71.7-kb Sall nifA or nifB fragment in pUC7IncP TCr cosmidK pneumoniae nifA gene cloned in pACYC177Tra+; mobilizing plasmidCloning and sequencing vector

aph (KMr) cartridge (KIXX)

1; This work (Fig.1)

This work (Fig. 1)This work (Fig. 1)This work (Fig. 1)This work (Fig. 1)

This work (Fig. 1)This work (Fig. 1)This work (Fig. 1)This work (Fig. 1)This work23222244715Pharmacia Ltd.,London, UnitedKingdom

Pharmacia Ltd.

Bacteriophage X::Tn5-B21 lacZ Tcr derivative of TnS on Xb221cI857 Pam80

derivatives were assumed to have arisen from a doublecrossover recombination event in which the wild-type nifL-or ntrB-like gene was replaced by the KIXX-containingDNA. Km' CbS transformants of MV101, the nifH-lacZfusion strain, were obtained with both pAB29 and pAB30,yielding strains MV378 and MV380, respectively. Kmr CbStransformants of UW136 were obtained with pAB29 but notwith pAB30 (see Discussion). The pAB29-derived strain,MV376, was Nif+. Southern blots of DNA from MV376,MV378, and MV380 digested with Sall were hybridized to a

32P-labelled probe from the mutated region (the 0.3-kbSalI-SphlI fragment from pAB31; Fig. 1). Southern hybrid-ization experiments showed that in all three mutants, the3.5-kb wild-type Sall fragment was absent and was replacedby a 6.2-kb hybridizing fragment (data not shown).The data in Table 2 show that MV376 was Nif+ and

expressed similarly high acetylene reduction activities bothin N-free medium and in medium with ammonium at 15 mM,a concentration that repressed nitrogenase activity in wild-type strain UW136. Both MV378 and MV380 expressed

,-galactosidase activity from the nifH-lacZ transcriptionalfusion in the presence of 15 mM ammonium. This activitywas 30 to 40% that obtained in the absence of ammonium,while in the NifL+ strain, activity in the presence of ammo-nium was only 3% that in N-free cultures. On the other hand,nitrogenase activity in the nifL mutants grown in the pres-ence of ammonium was equal to that in N-free medium.These results suggest that in the absence of fixed nitrogen,levels of nifHDK mRNA do not limit nitrogenase activity.The threefold difference in ,-galactosidase activity in themutants in the absence and presence of ammonium mayhave been due to the much greater stability of nif mRNAunder N-limited conditions, as occurs in K pneumoniae (10,21).These results demonstrate that the gene upstream of nifA

in A. vinelandii mediates ammonium repression of nif geneexpression. This phenotype is similar to that of the nifLmutants of K pneumoniae, in which nifL is located imme-diately upstream of nifA. Also, translation of the entire A.vinelandii upstream gene sequence (approximately 1.4 kb)

PlasmidspABS

pAB8pAB10pAB12pAB21

pAB25pAB26pAB29pAB30pAB39pCK1pCK3pDB150pDB160pLAFR3pMC71apRK2013pTZ18R

pUC4-KIXX

42

VOL. 58, 1992

on June 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

I 1 D

nifL nif A) ( )....

>(nifB)

a

I 11 I

pAB31

KIXXp aph

o PAB30 (MV380)3 - pAB29 (MV376.

pDB 50

I pAB25. pAB26

MV378)

J pDBI60

' pAB5

pAB8. pABIO

lacZ pAB12 (MV300)

FIG. 1. A. vinelandii nifL and nifA regions, subclones, and insertion derivatives. The DNA spanning this entire region was obtained froma lambda gene library and from pDB150 and pDB160 (see Materials and Methods). For clarity, the restriction sites shown do not representa complete restriction map of sites for these enzymes. The vector plasmids carrying the subclones and insertion derivatives are described inTable 1.

revealed a protein with greater homology toK pneumoniaeNifL than to NtrB (31 versus 24% identity) (6). We thereforenamed the gene upstream of nifA in A. vinelandii nifL.Ammonium excretion by nifL mutant MV376. Culture

supernatants of wild-type strain UW136 and of MV376grown on N2 were tested for the presence of ammonium. Incontrast to UW136, MV376 excreted ammonium rathersuddenly toward the end of exponential growth (Fig. 2). Themean level of ammonium excreted in six MV376 stationary-phase cultures was 9.3 + 2.7 mM. Also, the pH increased asammonium was excreted, reaching a final value of 8.5 (in

contrast to a final pH of about 6.5 in the wild-type cultures).The final optical density of wild-type stationary-phase cul-tures was slightly but consistently higher than that of thenifL mutant stationary-phase cultures (Fig. 2). The release ofammonium was not associated with the lysis of nifL mutantcells, since the count of viable MV376 cells was approxi-mately the same as that of viable wild-type UW136 cellswhen assayed after the cessation of ammonium excretion inMV376 (both cultures contained 2 x 109 to 3 x 109 CFUml-'). When a stationary-phase culture of MV376 that hadceased to excrete ammonium was centrifuged and resus-

TABLE 2. Nitrogenase and 1-galactosidase activities in the wild type and nifL mutants of A. vinelandii

Activity' in:Enzyme Strain Genotype N-free Medium containing Activity"

medium 15 mM NH4'

Nitrogenasec UW136 Wild type 46 0 0MV376 nifLI:KIXX 48 46 96

,B Galactosidased MV101 nifH1:lacZ-KSS 15,841 432 3MV378 nifHI:lacZ-KSS nifLl:KIXX 11,967 3,772 32MV380 nifHI1:lacZ-KSS nifL2:KIXX 15,384 6,474 42

a Each value is the mean from two or three independent determinations; variation was 10% from the mean.h In NH4+-containing medium relative to N-free medium.c Measured as acetylene reduction (nanomoles of ethylene produced minute-' mg of protein-').d Measured as Miller units (30).

pAB21

1714 BALI ET AL.

on June 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

AMMONIUM EXCRETION IN A. VINELANDII nifL MUTANTS 1715

5 10 15 20 25 3

0

Time (Hours)

FIG. 2. Growth (in Klett units), pH, and ammonium concentra-tions in cultures of A. vinelandii wild-type UW136 (open symbols)and nifL mutant MV376 (closed symbols). Cultures were grown asdescribed in Materials and Methods. Samples (1 ml) were removedfor analysis at the times shown.

pended in fresh medium to the same density, ammonium was

again released after 1 to 2 h of incubation, indicating thatexcretion was limited by nutrient supply or a high pH or hadbeen prevented by the accumulation of an inhibitor in themedium.The sucrose concentration in MV376 supernatants con-

taining high levels of ammonium was half that contained inthe starting culture (1 versus 2%), indicating that the carbonsource was probably not the limiting factor preventingfurther ammonium excretion. More likely, the high pH incultures of MV376 explains the cessation of ammoniumexcretion, because when KOH was added to late-logarith-mic-phase cultures of wild-type UW136, nitrogenase (acety-lene reduction) activity was inhibited at pH 8.5 but not at pH8.0 or a lower pH. Equivalent additions of KCI had no effecton nitrogenase activity in cultures of UW136.

If the nifL-like gene in A. vinelandii inactivates the nifAgene product (NifA) in the presence of ammonium, as it doesin K pneumoniae, then the overproduction of NifA mightalso result in ammonium excretion in this organism. Theconstitutive expression ofK pneumoniae nifA from the aphpromoter in plasmids pCK1 and pCK3 (high- and low-copy-number plasmids, respectively) complemented the Nif- phe-notype of nifA mutants UWI and MV300 and also resulted innitrogenase expression in the transconjugants of both mutant

and wild-type A. vinelandii strains grown in high concentra-tions of ammonium (22, 23). pAB10, a low-copy-numberIncP plasmid (Fig. 1) which expresses the nifA gene of A.vinelandii from the same constitutive promoter as that inpCK1 and pCK3, complemented the Nif- phenotype of bothUW1 and MV300. Cultures of transconjugants of UW136and UW1 carrying pCK1, pCK3, or pAB10 neither excretedammonium (<0.01 mM) nor showed an increase in pHduring the late logarithmic or stationary phases. Thus, theammonium excretion phenotype in MV376 is apparently duenot to an excess of NifA but to the absence of NifL.Ammonium excretion by a K. pneumoniae nifL mutant.

Ammonium excretion by the A. vinelandii nifL mutantprompted a measurement of ammonium levels in the culturemedium of a nifL deletion mutant ofK pneumoniae, strainUN4357. Approximately five times more ammonium wasexcreted by the mutant than by the wild type, M5al, after 48h of diazotrophic growth or after 8 h of derepression inN-free medium with either glucose or glycerol plus fumarateas the carbon and energy source following an NH4' down-shift (see Materials and Methods). Excretion occurred dur-ing derepression in cultures incubated under N2 but notunder Ar. The final concentrations of ammonium measuredin the N2-sparged nifL mutant cultures were 0.04 to 0.15 mMand thus were only about 1% those in the A. vinelandii nifLmutant cultures. In contrast to the pH of the A. vinelandiimutant cultures, the pH of the N2-sparged cultures of Kpneumoniae UN4357 was not higher but was somewhatlower than that of the wild type. As with A. vinelandii, theintroduction of a plasmid expressing K. pneumoniae nifAfrom a constitutive promoter (pMC71a) did not result inammonium excretion by wild-type K pneumoniae; ammo-nium levels were <0.01 mM in cultures of both M5al andM5al (pMC71a).Ammonium does not control nifA expression in A. vinelan-

dii. A significant difference between nif gene expression inK pneumoniae and A. vinelandii is that ntrC mutants of thelatter are Nif+ but those of the former are Nif-. In Kpneumoniae, NtrC activates the expression of the nifLAoperon in the absence of significant fixed nitrogen (<200 ,uMNH4+). Since A. vinelandii ntrC mutants are Nif+, theexpression of nifA is independent of NtrC and therefore maynot be regulated by NH4+.To examine nifA expression in A. vinelandii, we con-

structed a nifA-lacZ fusion by inserting a TnS-lacZ deriva-tive, TnS-B21, into pAB8 (see Materials and Methods andFig. 1). The insertion site and orientation of Tn5-B21 in thenifA gene in the resulting plasmid, pAB12, were determinedby restriction analysis (Fig. 1). pAB12 was transformed intoUW136; the replacement of the wild-type nifA region withnifA::TnS-B21 was confirmed by Southern hybridizationanalysis of DNA from Tcr CbS UW136 transformants (datanot shown). 1-Galactosidase measurements in transformantMV300 revealed that nifA expression was not repressed byammonium; in cultures grown with urea at 2 mM (nonre-pressing N source) or with NH4' at 15 mM (repressing Nsource), ,B-galactosidase activities were similar (896 and1,031 Miller units, respectively).

DISCUSSION

Up to now, significant ammonium excretion from nitro-gen-fixing free-living diazotrophs, including A. vinelandii,has been achieved only through chemical or mutationalblockage of the pathway(s) of ammonium assimilation. Suchmanipulated organisms treated with the GS inhibitor L-me-

350

300

250

200.04, 150

100

50J

99r

8

7

/bY-.YL V

m

F.-

6 I.I i0 5 10 15 20 25 34

7 r6 -

5

3-~~~~~~~~~~~~~~~~~~~~~~~-

t- I ___

0 5 10 15 20 25 3

~o

VOL. 58, 1992

I I I I I

on June 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

thionine-DL-sulfoximine or mutants which are deficient inGS do not grow unless glutamine is provided. In addition,ethylenediamine-resistant mutants ofA. brasiliense excretedup to 5 mM ammonium during diazotrophic growth, con-tained highly adenylylated (hence partially inactive) GS, andgrew slowly in the absence of glutamine (27).The ability of nifL mutants of A. vinelandii to excrete

ammonium is not related to defective ammonium assimila-tion; these mutants grow as well as wild-type organisms ondinitrogen or other N sources, such as ammonium, nitrate,or urea. In A. vinelandii, the GS-glutamate synthase path-way is the only route for ammonium assimilation and, unlikethe situation in enteric and other bacteria, glutamate dehy-drogenase is not present (24). The gene encoding GS, glnA,is apparently an essential gene, since replacement of thewild-type glnA gene with DNA carrying insertion mutationsin glnA was not possible (47).The nifL mutants of A. vinelandii excrete ammonium

during the late logarithmic and early stationary growthphases to yield a final concentration of about 10 mM in thegrowth medium. The ammonium produced from nitrogenfixation during the exponential phase is probably just suffi-cient to support the growth of the nifL mutant cultures, asindicated by the equal nitrogenase activities in the mutantand wild-type strains during exponential growth. Whengrowth ceases, continued nitrogenase activity in the mutantsresults in excretion of the ammonium which is no longerassimilated. Another factor affecting excretion may be theoxygen-limiting conditions that probably occur in the latelogarithmic growth phase of A. vinelandii batch cultures.Kuhla et al. (26) showed that the excretion of nitrogen-containing compounds, including ammonium, was muchgreater in chemostat cultures grown under a low oxygensupply. Although their wild-type organism excreted no morethan about 0.2 mM ammonium under the most limitingoxygenation conditions, this effect might enhance theamount of ammonium excreted in the A. vinelandii nifLmutants reported here. Excretion was not limited by carbonsource availability, because half the supplied sucrose re-mained at the time when ammonium excretion stopped. Aconcomitant rise in the pH to approximately 8.5 occurredduring the period when ammonium was released. This rise inpH probably resulted in a loss of nitrogenase activity in themutants because (i) ammonium excretion recommencedwhen the organisms were resuspended to the same station-ary phase density in fresh medium and (ii) a loss of nitroge-nase activity occurred in wild-type strain UW136 when thepH of a late-logarithmic-phase culture was increased to 8.5.The failure of NH4' to repress nitrogenase synthesis in

the nifL mutants ofA. vinelandii, as well as the ability of themutants to excrete ammonium produced from nitrogen fix-ation, indicates that NifL has a role in the regulation of nifgene expression by fixed N in this organism. Further evi-dence that the nifL gene product responds to levels of fixednitrogen is the ability of nifL mutations to suppress the Nif-phenotype of nfrX mutants (12). The nfrX gene of A.vinelandii is homologous to the glnD gene of E. coli; theglnD gene product, uridylyl transferase, initiates a cascadeof nitrogen regulatory responses in enteric bacteria. In Kpneumoniae, NifL is thought to inactivate the nifA geneproduct, NifA, in the presence of >5 mM ammonium (14, 18,31); NifL may also influence the stability of nif mRNAtranscripts (11). A nipL mutant of K pneumoniae excreted100-fold less ammonium into the medium than an A. vine-landii nifL mutant. An explanation for this difference may liein the control of the expression of the nifA gene. In K

pneumoniae, transcription of the nifLA operon requiresNtrC, and ntrC mutants are Nif-. The expression of ntrC isrepressed by ammonium (>200 mM), as is nifLA expression.In contrast, ntrC mutants ofA. vinelandii are Nif+ (45), andthe work presented here shows that the expression of anifA-lacZ fusion is not repressed by ammonium even at highconcentrations (15 mM).Although the KIXX insertion in both orientations in the

nifL gene of A. vinelandii resulted in a Nif+ phenotype,insertions of Tn5 or the fl (transcription terminator) cassettein A. vinelandii nifL resulted in a Nif- phenotype (6).Therefore, the nifLA genes are probably cotranscribed in A.vinelandii, as they are in K pneumoniae. Although theKIXX aph cartridge inserted upstream to oppose transcrip-tion can inhibit the expression of downstream genes in A.vinelandii (39), the nifA gene in the nifL (KIXX) mutantsMV376 and MV378 was obviously expressed, possibly fromunexpected promoter activity in the cassette or from pro-moterlike sequences generated by the KIXX insertion in thenifL region.

Insertion of KIXX in the orientation in which the expres-sion of aph was in the same direction as nifA resulted insomewhat higher constitutive levels of P-galactosidase ac-tivity (MV380) than those obtained when KIXX was insertedin the opposite orientation (MV378). However, it was notpossible to construct a Nif+ nifL2:KIXX derivative in thewild type equivalent to MV380 which, because it carries anifHl-lacZ fusion, is Nif-. Therefore, excessive expressionof nifA driven by the very strong aph promoter is probablylethal in a Nif+ A. vinelandii background, possibly becauseof the diversion of too much ATP toward nitrogenasesynthesis and activity or to NH4' toxicity.

Constitutive nifA expression in the A. vinelandii wild typeor a nifA mutant did not result in ammonium excretion,whether NifA was from A. vinelandii or from K pneumo-niae. In the latter case, A. vinelandii NifL probably cannotinactivate K pneumoniae NifA, since K pneumoniae NifLdoes not inhibitA. vinelandii NifA activity (22). A. vinelandiiNifL may therefore influence NH4' excretion in ways otherthan simply through its effect on NifA. One possibility is thatNifL directly or indirectly inhibits the transport of ammo-nium or of ammonia by an active or a passive process,respectively (25).

Optimization of ammonium excretion by, for example,alteration of the C source, 02 supply, or buffering capacity ofthe medium was not investigated. Therefore, the potentialfor ammonium excretion by nifL mutants ofA. vinelandii isprobably greater than that reported here. The direct use ofA. vinelandii NifL- mutants in agriculture to provide asource of fixed N for plant growth is unlikely to be effective,because C sources are relatively scarce in the soil and mostAzotobacter species are generally present at a low density.On the other hand, if the ability of NifL- mutants to excreteammonium is widespread among other nitrogen-fixing bac-teria, then there is a potential strategy for the manipulationof rhizosphere diazotrophs with strong plant associations toenhance fixed N supply to plants. Ethylenediamine-resistantmutants ofA. brasiliense, which are presumably deficient insome aspect of NH4' assimilation and which excrete muchless ammonium than the A. vinelandii nifL mutants, wererecently reported to enhance nitrogen supply to a wheat host(9). As yet, a nifL gene has not been reported (or possiblyhas not been looked for) in diazotrophs other than Kpneumoniae and A. vinelandii. Thus, a search for nifL andits mutagenesis in plant-associated nitrogen-fixing bacteriacould be rewarding.

1716 BALI ET AL.

on June 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

AMMONIUM EXCRETION IN A. VINELANDIH nifL MUTANTS 1717

ACKNOWLEDGMENTS

We thank Dennis Dean for providing nifA plasmids, Richard Paufor the A. vinelandii X library, Martin Drummond and Barry Smithfor critical and constructive comments on the manuscript, andRosemary Foote for typing.

This research was supported in part by grants from the Associa-tion of Commonwealth Universities to A. Bali and from the Minis-terio de Educaion y Ciencia (Spain) to G. Blanco.

REFERENCES

1. Andersen, K., and K. T. Shanmugam. 1977. Energetics ofbiological nitrogen fixation: determination of the ratio of forma-tion of H2 to NH4' catalysed by nitrogenase of Klebsiellapneumoniae in vivo. J. Gen. Microbiol. 103:107-122.

2. Bennett, L. T., F. C. Cannon, and D. Dean. 1988. Nucleotidesequence and mutagenesis of the nifA gene from Azotobactervinelandii. Mol. Microbiol. 2:315-321.

3. Bergersen, F. J. 1980. Methods for evaluating biological nitro-gen fixation. John Wiley & Sons, Ltd., London.

4. Bishop, P. E., and W. J. Brill. 1977. Genetic analysis ofAzotobacter vinelandii mutant strains unable to fix nitrogen. J.Bacteriol. 130:954-956.

5. Bishop, P. E., and R. D. Joerger. 1990. Genetics and molecularbiology of alternative nitrogen fixation systems. Annu. Rev.Plant Physiol. Plant Mol. Biol. 41:109-125.

6. Blanco, G., A. Bali, C. Kennedy, and P. Woodley. Unpublisheddata.

7. Buchanan-Wollaston, V., M. C. Cannon, J. L. Beynon, and F. C.Cannon. 1981. Role of the nifA gene product in the regulation ofnif expression in Klebsiella pneumoniae. Nature (London)294:776-778.

8. Castorph, H., and D. Kleiner. 1984. Some properties of aKlebsiella pneumoniae ammonium transport negative mutant(amt-). Arch. Microbiol. 139:245-247.

9. Christiansen-Weniger, C., and J. A. VanVeen. 1991. NH4+-excretingAzospirillum brasiliense mutants enhance the nitrogensupply of a wheat host. Appl. Environ. Microbiol. 57:3006-3012.

10. Collins, J. J., and W. J. Brill. 1985. Control of Klebsiellapneumoniae nif mRNA synthesis. J. Bacteriol. 162:1186-1190.

11. Collins, J. J., G. P. Roberts, and W. J. Brill. 1986. Posttran-scriptional control of Klebsiella pneumoniae nifmRNA stabilityby the nifL product. J. Bacteriol. 168:173-178.

12. Contreras, C., M. Drummond, A. Bali, G. Blanco, E. Garcia, G.Bush, C. Kennedy, and M. Merrick. 1991. The product of thenitrogen fixation regulatory gene nfrX ofAzotobacter vinelandiiis functionally and structurally homologous to the uridylyltrans-ferase encoded by ginD in enteric bacteria. J. Bacteriol. 173:7741-7749.

13. Dixon, R. 1988. Genetic regulation of nitrogen fixation, p. 417.In J. A. Cole and S. J. Ferguson (ed.), The nitrogen and sulphurcycles. Cambridge University Press, Cambridge.

14. Drummond, M. H., A. Contreras, and L. A. Mitchenall. 1990.The function of isolated domains and chimaeric proteins con-structed from the transcriptional activators NifA and NtrC ofKlebsiella pneumoniae. Mol. Microbiol. 4:29-37.

15. Figurski, D. H., and D. R. Helinski. 1979. Replication of anorigin-containing derivative of plasmid RK2 dependent on aplasmid function provided in trans. Proc. Natl. Acad. Sci. USA73:1648-1652.

16. Filser, M., M. Merrick, and F. Cannon. 1983. Cloning andcharacterisation of nifLA regulatory mutations from Klebsiellapneumoniae. Mol. Gen. Genet. 191:485-491.

17. Gordon, J. K., and M. R. Jacobson. 1983. Isolation and charac-terization of Azotobacter vinelandii mutant strains with poten-tial as bacterial fertilizer. Can. J. Microbiol. 29:973-978.

18. Henderson, N., S. A. Austin, and R. A. Dixon. 1989. Role ofmetal ions in negative regulation of nitrogen fixation by the nifLgene product from Klebsiella pneumoniae. Mol. Gen. Genet.216:484-491.

19. Jacobson, M. R., K. E. Brigle, L. Bennett, R. A. Setterquist, R. A.Wilson, V. L. Cash, J. Beynon, W. E. Newton, and D. R. Dean.

1989. Physical and genetic map of the major nif gene cluster fromAzotobacter vinelandii. J. Bacteriol. 171:1017-1027.

20. Joerger, R. D., and P. E. Bishop. 1988. Nucleotide sequence andgenetic analysis of the nifB-nifQ region from Azotobacter vine-landii. J. Bacteriol. 170:1475-1487.

21. Kahn, D., M. Hawkins, and R. R. Eady. 1982. Metabolic controlof Klebsiellapneumoniae mRNA degradation by the availabilityof fixed nitrogen. J. Gen. Microbiol. 128:3011-3018.

22. Kennedy, C., and M. H. Drummond. 1985. The use of cloned nifregulatory elements from Klebsiella pneumoniae to examine nifregulation in Azotobacter vinelandii. J. Gen. Microbiol. 131:1787-1795.

23. Kennedy, C., and R. Robson. 1983. Activation of nif geneexpression in Azotobacter by the nifA gene product of Klebsi-ella pneumoniae. Nature (London) 301:626-628.

24. Kennedy, C., and A. Toukdarian. 1987. Genetics of azotobac-ters: applications to nitrogen fixation and related aspects ofmetabolism. Annu. Rev. Microbiol. 41:227-248.

25. Kleiner, D. 1985. Bacterial ammonium transport. FEMS Micro-biol. Rev. 32:87-100.

26. Kuhla, J., C. Dingier, and J. Oelze. 1991. Production of extra-cellular nitrogen-containing components by Azotobacter vine-landii fixing dinitrogen in oxygen-controlled continuous culture.Arch. Microbiol. 141:297-302.

27. Machado, H. B., S. Funayama, and L. U. Rigo. 1991. Excretionof ammonia by Azospinillum brasiliense, p. 558. In P. M.Greshoff, L. E. Roth, and G. Stacey (ed.), Nitrogen fixation:achievements and objectives. Chapman and Hall, New York.

28. MacNeil, T., D. MacNeil, G. P. Roberts, M. A. Supiano, andW. J. Brill. 1978. Fine-structure mapping and complementationanalysis of nif (nitrogen fixation) genes in Klebsiella pneumo-niae. J. Bacteriol. 136:253-266.

29. Merrick, M., S. Hill, H. Hennecke, M. Hahn, R. Dixon, and C.Kennedy. 1982. Repressor properties of the nifL gene product ofKlebsiella pneumoniae. Mol. Gen. Genet. 185:75-81.

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

31. Morett, E., R. Kreutzer, W. Cannon, and M. Buck. 1990. Theinfluence of the Klebsiella pneumoniae regulatory gene nifLupon the transcriptional activator protein NifA. Mol. Microbiol.4:1253-1258.

32. Musgrave, S. C., N. W. Kerby, G. A. Codd, and W. D. P.Stewart. 1982. Sustained ammonia production by immobilizedfilaments of the nitrogen-fixing cyanobacterium Anabaena27893. Biotechnol. Lett. 4:647-652.

33. Newton, J. W., and J. F. Cavins. 1985. Liberation of ammoniaduring nitrogen fixation by a facultatively heterotrophic cyano-bacterium. Biochim. Biophys. Acta 809:44-50.

34. Page, W. J., and M. von Tigerstrom. 1978. Induction of trans-formation competence in Azotobacter vinelandii iron-limitedcultures. Can. J. Microbiol. 24:1590-1594.

35. Polukhina, L. E., G. N. Sakhurieva, and S. V. Shestakov. 1982.Ethylenediamine-resistant Anabaena variabilis mutants withderepressed nitrogen-fixing system. Microbiology 51:90-95.

36. Ramos, J. L., M. G. Guerrero, and M. Losada. 1984. Sustainedphotoproduction of ammonia from dinitrogen and water by thenitrogen-fixing cyanobacterium Anabaena sp. strain ATCC33047. Appl. Environ. Microbiol. 48:114-118.

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

38. Santero, E., A. Toukdarian, R. Humphrey, and C. Kennedy.1988. Identification and characterisation of two nitrogen fixationregulatory regions, nifA and nfirX, in Azotobacter vinelandii andAzotobacter chroococcum. Mol. Microbiol. 2:303-314.

39. Selman Sosa, G., and C. Kennedy. Unpublished data.40. Shah, V. K., L. C. Davis, J. K. Gordon, W. H. Orme-Johnson,

and W. J. Brill. 1973. Nitrogenase. III. Nitrogenaseless mutantsof Azotobacter vinelandii: activities, cross-reactions and EPRspectra. Biochim. Biophys. Acta 292:246-255.

41. Shanmugam, K. T., and R. C. Valentine. 1975. Microbialproduction of ammonium ion from nitrogen. Proc. Natl. Acad.Sci. USA 72:136-139.

VOL. 58, 1992

on June 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

42. Simon, R., J. Qyandt, and W. Klipp. 1989. New derivatives oftransposon Tn5 suitable for mobilization of replicons, genera-tion of operon fusions and induction of genes in gram-negativebacteria. Gene 80:161-169.

43. Spiller, H., C. Latorre, M. E. Hassan, and K. T. Shanmugam.1986. Isolation and characterization of nitrogenase-derepressedmutant strains of the cyanobacterium Anabaena variabilis. J.Bacteriol. 165:412-419.

44. Staskawicz, B., D. Dahlbeck, N. Keen, and C. Napoli. 1987.Molecular characterization of cloned avirulence genes from race0 and race 1 of Pseudomonas syringae pv. glycinea. J. Bacte-riol. 169:5789-5794.

45. Thomas, S. P., A. Zaritsky, and S. Boussiba. 1990. Ammoniumexcretion by an L-methionine-DL-sulfoximine-resistant mutantof the rice field cyanobacterium Anabaena siamensis. Appl.

Environ. Microbiol. 56:3499-3504.46. Toukdarian, A., and C. Kennedy. 1986. Regulation of nitrogen

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

47. Toukdarian, A., G. Saunders, G. Selman-Sosa, E. Santero, P.Woodley, and C. Kennedy. 1990. Molecular analysis of theAzotobacter vinelandii glnA gene encoding glutamine syn-thetase. J. Bacteriol. 172:6529-6539.

48. Wall, J. D., and H. Gest. 1979. Derepression of nitrogenaseactivity in glutamine auxotrophs of Rhodopseudomonas capsu-lata. J. Bacteriol. 137:1459-1463.

49. Walmsley, J., and C. Kennedy. 1991. Temperature-dependentregulation by molybdenum and vanadium of expression of thestructural genes encoding three nitrogenases in Azotobactervinelandii. Appl. Environ. Microbiol. 57:522-524.

1718 BALI ET AL.

on June 1, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from