Vol. 54, No. 2APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Feb. 1988, p. 358-3630099-2240/88/020358-06$02.00/0Copyright © 1988, American Society for Microbiology

Transposon Tn5-Generated Bradyrhizobium japonicum MutantsUnable To Grow Chemoautotrophically with H2SHERMAN SIU MING HOM,t PATRICIA D. NOVAK, AND ROBERT J. MAIER*

Department of Biology and McCollum-Pratt Institute, The Johns Hopkins University, Baltimore, Maryland 21218

Received 10 August 1987/Accepted 9 November 1987

Twelve TnS-induced mutants of Bradyrhizobiumjaponicum unable to grow chemoautotrophically with CO2and H2 (Aut-) were isolated. Five Aut- mutants lacked hydrogen uptake activity (Hup-). The other sevenAut- mutants possessed wild-type levels of hydrogen uptake activity (Hup+), both in free-living culture andsymbiotically. Three of the Hup- mutants lacked hydrogenase activity both in free-living culture and as nodulebacteroids. The other two mutants were Hup- only in free-living culture. The latter two mutants appeared tobe hypersensitive to repression by oxygen, since Hup activity could be derepressed under 0.4% 02. All fiveHup- mutants expressed both ex planta and symbiotic nitrogenase activities. Two of the seven Aut- Hup+mutants expressed no free-living nitrogenase activity, but they did express it symbiotically. These two strains,plus one other Aut- Hup+ mutant, had CO2 fixation activities 20 to 32% of the wild-type level. The cosmidpSH22, which was shown previously to contain hydrogenase-related genes of B.japonicum, was conjugated intoeach Aut- mutant. The Aut- Hup- mutants that were Hup- both in free-living culture and symbiotically werecomplemented by the cosmid. None of the other mutants was complemented by pSH22. Individual subclonedfragments of pSH22 were used to complement two of the Hup- mutants.

Bradyrhizobium japonicum and soybeans (Glycine max)form an effective nitrogen-fixing symbiosis. H2 evolvedduring the ATP-dependent reduction of N2 by nitrogenasedecreases the efficiency of nitrogen fixation, apparently as aresult of energy loss (24, 29). Some strains of B. japonicumpossess a membrane-bound uptake hydrogenase enzyme(Hup) as nodule bacteroids which can recycle the H2 pro-duced by nitrogenase to produce ATP (6, 7).

B. japonicum Hup+ strains also can oxidize H2 in free-living cultures when growing chemoautotrophically undermicroaerobic conditions (10, 15). The inability of Hup-mutants to grow chemoautotrophically has been used effec-tively to isolate many mutants with defects in the Hupsystem (16, 18). Physiological and biochemical analyses ofthese Hup- mutants, generated by chemical mutagenesis,have provided information concerning both the componentsinvolved in the hydrogen uptake system and the way inwhich it is regulated. The many types of Hup mutantsinclude some unable to oxidize H2 with either 02 or meth-ylene blue as electron acceptor (16, 18), others simulta-neously lacking both hydrogenase and nitrogenase enzymeactivities (23), and still others that are hypersensitive torepression of H2 oxidation by 02 (21).

In this study, we have isolated and characterized someHup- mutants generated by transposon Tn5 (insertion)mutagenesis in an attempt to elucidate further the biochem-ical and genetic components involved in the hydrogen up-take system of B. japonicum. In addition, genes involved inH2 oxidation were used to complement some of the Hup-mutants.

MATERIALS AND METHODS

Bacterial strains and plasmids. The bacterial strains andplasmids used in this study are listed in Table 1. B. japoni-

* Corresponding author.t Present address: Henkel Research Corp., Santa Rosa, CA

95407.

cum stock cultures were maintained on mannitol-yeast ex-tract (MSY) agar plates (13), supplemented with the follow-ing antibiotics when necessary: rifampin (100 jig/ml),kanamycin (Km; 100 pLgIml), streptomycin (Sm; 100 ,ug/ml),and tetracycline (Tc; 80 ,ug/ml). B. japonicum was grown inBM (2) containing 5 ,uM NiCl2 or in MSY. When the cellscontained Tn5 elements, 50 pug of kanamycin per ml was alsopresent. Escherichia coli cultures were maintained on LBagar plates supplemented with the appropriate antibiotic(tetracycline, 20 ,ug/ml; streptomycin, 50 jig/ml; chloram-phenicol, 25 ,ug/ml; kanamycin, 50 ,ug/ml). E. coli culturescontaining cloned plasmids were grown in LB supplementedwith 20 ,ug of tetracycline per ml. E. coli HB101 containingpRK2013 was grown in LB containing 20 ,ug of kanamycinper ml. For filter matings, B. japonicum and E. coli cellswere grown to mid-log phase (optical density of 0.5 at 540and 550 nm, respectively).Tn5 mutagenesis of B. japonicum. B. japonicum SU was

mutagenized with transposon Tn5 by conjugating the suicideplasmid pSUP1011 from E. coli to B. japonicum as previ-ously described (13), with the minor modifications describedby Maier and Hom (20). Use of Tn5 to obtain the Hup-mutants has been described, but the mutants were notcharacterized (20). Because Tn5 expresses resistance to bothkanamycin and streptomycin in B. japonicum (9), bothantibiotics were used to select Tn5 recipients. B. japonicumstrains containing Tn5 insertions were detected at a fre-quency of 2.4 x 10-6 per recipient.Hydrogen uptake assays. Hydrogen uptake activities in

free-living cultures of B. japonicum strains were measuredby the derepression procedure of Hom et al. (12) with thefollowing minor modifications. The growth medium usedwas BM supplemented with 5 ,uM NiCl2. When transposonTn5 mutants were cultured, the growth medium was supple-mented with kanamycin (20 to 50 ,ug/ml). All H2 uptakeassays were performed by using as the electron acceptoreither 02 (100 to 180 ,uM) or methylene blue (200 ,uM) inTriton X-100 as described previously (27). These concentra-tions of electron acceptors were saturating for all H2 uptake

358

on July 8, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

TnS-GENERATED B. JAPONICUM MUTANTS 359

TABLE 1. Bacterial strains and plasmids

Strain or Relevant genotype or Referenceplasmid phenotypea

B. japonicumSU Rif Hup' Nif' 13SU59 Rif' Kmr Smr Hup- This studySU59(pSH22) Rif' Tcr Kmr Smr Hup+ This studySU21 Rif Kmr Smr Hup+ This studySU27 Rift Kmr Smr Hup- This studySU27(pSH22) Rift Tcr Kmr Smr Hup+ This studySU47 Rif' Kmr Smr Hup- 12SU47(pSH22) Rif' Tcr Kmr Smr Hup+ 12SU17 Rif' Kmr Smr Hup- This studySU44 Rifr Kmr Smr Aut- This studySU55 Rif' Kmr Smr Aut- Nif This studySU28 Rif' Kmr Smr Aut- This studySU46 Rif' Kmr Smr Aut- This studySU14 Rif' Kmr Smr Aut- This studySU24 Rif Kmr Smr Aut- This studySU19 Rif' Kmr Smr Aut- Nif This studySU15 Two Tn5 insertionsSU39 Two Tn5 insertions

E. coliSM1o thi thr leu terA lacY supE 25

recA [(RP4.2.Tc::Ma) ApsTc' Mu c+ Kmr Tra+]

HB101 pro leu thi lacY Smr endA 3

PlasmidpSUPlOll Cmr Kmr Nmr oriTRp4 3pSH22 Tcr niflhup hup 12pRK2013 Kmr 5

a Rif, Rifampin; Km, kanamycin; Sm, streptomycin; Tc, tetracycline; Cm,chloramphenicol; Nm, neomycin; Hup, oxygen-dependent hydrogen uptakeactivity; Nif, nitrogenase activity; Aut, chemoautotrophic growth; Tra, trans-fer functions; oriTRP4, origin of transfer from RP4; r, resistance; s, sensitivity.

assays. H2 uptake values are the average of at least twoindependent determinations.

Bacteroids were isolated from soybean nodules, and sym-biotic hydrogen uptake activities were determined as de-scribed before (21). The protein concentrations of bacteroidsamples were assayed by the procedure of Lowry et al. (17),with bovine serum albumin (Sigma Chemical Co.) as astandard, after disruption of the bacteroids in NaOH (21).

Nitrogenase assays. Both symbiotic nitrogenase activitiesof root sections containing soybean nodules and ex plantanitrogenase activity of free-living cultures of B. japonicumwere determined as described previously (12). The proce-dure for determining ex planta activity is a modification (8) ofthat described by Agarwal and Keister (1). Gas chromato-graphic conditions for assaying ethylene production havealso been described (8). Kanamycin (20 ,.g/ml) was includedin the TnS-containing cultures of B. japonicum mutantswhich were to be induced for nitrogenase and hydrogenaseactivity.Soybean nodulation experiments. B. japonicum strains

were assayed for nitrogen fixation and hydrogen uptakeactivities with soybeans [Gl. max (L.) Merr. cv. Essex].Seeds were surface sterilized with 95% ethanol and 0.2%HgCl2 as previously described (28). Germinated soybeanseeds were inoculated with approximately 109 B. japonicumcells and grown for 6 weeks in sterile Leonard jar assembliesas described before (21). Surface-sterilized nodules (28) werecrushed onto MSY agar plates containing rifampin (100,ug/ml), kanamycin (100 ,ug/ml), streptomycin (100 jig/ml),and cycloheximide (150 ,ug/ml).

Subcloning and mapping. Individual EcoRI fragments ofpSH22 were isolated and cloned into the EcoRI site of thebroad-host-range vector pRK290 (5). Restriction endonucle-ase sites of pSH22 and subclones were mapped by usingstandard mapping techniques (22).

Filter matings. Subcloned plasmids were conjugated fromE. coli to B. japonicum TnS mutants, using the triparentalfilter mating technique (4) described by Hom et al. (12) withthe following modifications. The ratio of donor/recipientcells was 1:5, and cells were mixed and placed on filtersimmediately as they reached the desired cell density. Afterincubation of filters on yeast extract medium for 5 days (12),cells were suspended with 0.01% Tween 20 and plated onmethylene blue agar containing 80 ,ug of tetracycline per ml.Recipient B.japonicum cells, which were Tcr, were streakedon no-carbon agar medium and incubated in a microaerophi-lic environment as previously described (18) to test forcomplementation to autotrophy. A total of 500 colonies fromeach mating were streaked.

Whole-cell CO2 uptake. CO2 uptake rates were assayed asdescribed previously (18), with the following adjustments.NaH14CO3 of specific activity 42.5 mCi/mmol was obtainedfrom New England Nuclear Corp., Boston, Mass. Cellsderepressed for H2 uptake (3 x 109 cells in 6-ml volume)were added to a 32-ml serum vial, the bottle was flushed withN2 and sealed, and gases were injected to obtain an atmo-sphere of 89% N2-10% H2-1.0% 02. NaHCO3 was added toa concentration of 20 mM. The final reaction solution con-tained 0.1 pCi of NaH14CO3 per ,umol. The sealed vials wereshaken at 30°C for 30 min, and five 1-ml samples wereremoved for assays of radioactivity. The 1-ml samples wereadded to scintillation vials containing 0.3 ml of 60% tri-chloroacetic acid. The vials were shaken and left in a fumehood for 48 h. This procedure was sufficient to dissipatemost of the 14CO2 that was not fixed. Counting efficiency inAquasol was 85%, and boiled cells of each correspondingstrain were used to determine background. The reportedvalues are the mean + standard deviation for the fivereplicate samples.

Southern blot analysis. Physical verification of the TnSinsertion in the mutant strains was carried out by DNAhybridization analysis as described previously (19).

RESULTS

Isolation of TnS-induced mutants of B. japonicum withdefects in hydrogen uptake activity. After TnS-induced muta-genesis of B. japonicum SU, approximately 6,000 transcon-jugant colonies resistant to both kanamycin and streptomy-cin were obtained. Of these 6,000, 14 mutants were unable togrow chemoautotrophically. The TnS-generated Aut- mu-tants were then screened for their ability to oxidize hydrogen(Hup) in free-living cultures (Table 2). Of the 14 Aut-mutants, five strains, designated class 1, lacked Hup activitywith either 02 or methylene blue provided as an electronacceptor. Since methylene blue is an acceptor for the puri-fied hydrogenase enzyme, class 1 mutants must lack activityfor that enzyme (Hup-). Seven Aut- strains, class 2, pos-sessed Hup activities that ranged from 50 to 70% of that ofthe parent strain when 02 was used as the terminal electronacceptor (Aut- Hup+). These Aut- Hup+ strains also pos-sessed substantial hydrogenase activity when methyleneblue was provided as the acceptor. Previously reported Aut-mutants of B. japonicum SR with levels of Hup activitysimilar to those of the class 2 mutants were shown to possessdefects in CO2 fixation activity (18).

VOL. 54, 1988

on July 8, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

TABLE 2. Hydrogen oxidation activities ofTnS-induced Aut- mutantsa

Hydrogen uptake activity(nmol of H2 oxidized/h per 108 cells)

Strain02 as electron MB as electron

acceptor acceptorb

SU (parent) 99 73

Class 1SU59 0 <0.5SU21 0 <0.5SU27 0 <0.5SU47 0 <0.5SU17 0 <0.5

Class 2SU44 68 NDcSU55 50 NDSU28 56 40SU46 49 43SU14 53 39SU24 47 36SU19 56 44

" 6,000 Smr Kmr colonies were screened in chemoautotrophic conditions,and 14 Smr Kmr strains unable to grow autotrophically were detected.

b Methylene blue (MB) was added at 200 pLM after cells were permeabilizedwith Triton X-100, as described by Stults et al. (27).

c ND, Not determined.

Southern blot analysis of TnS-induced mutants of B. japo-nicum. Southern hybridizations (not shown) verified thepresence of TnS in the genome of the B. japonicum Smr KmrAut- mutants. EcoRI digests of total genomic DNA fromthese mutants were hybridized to the BglI-SalI fragment ofTnS DNA. Single bands of hybridization were detected in allclass 1 and class 2 mutants. Since Tn5 rarely inserts intandem, this result indicated that the Tn5 was inserted at asingle site within some gene involved in the expression ofHup activity for each mutant. Multiple bands hybridizing tothe probe were detected in strains SU15 and SU39, andtherefore they were not characterized further.

Whole-cell CO2 fixation activities. An inability to growchemoautotrophically in an atmosphere containing H2 andCO2 could be due to deficiencies in H2 oxidation or CO2fixation. CO2 uptake rates were measured in cultures whichhad been derepressed for H2 uptake (Table 3). All Hup-mutant strains tested (SU59, SU47, SU21, and SU17) weredeficient in CO2 uptake; cells which are Hup- may well beunable to provide ATP for carboxylation enzymes, thusexhibiting lower CO2 fixation rates. In addition, some Aut-Hup+ mutants were also deficient in whole-cell CO2 fixation(Table 3, class 2). Mutant strains SU19, SU44, and SU55 hadsignificantly less CO2 fixation activity than did the wild type.These mutants may lack activity for the key C02-fixingenzyme ribulose bis phosphate carboxylase.

Symbiotic H2 uptake activities of TnS-induced Hup- mu-tants. The five Hup- mutants (class 1) were screened forHup activity as bacteroids from soybean nodules. Bacteroidpreparations of Hup- strains SU59, SU27, and SU47 pos-sessed no detectable hydrogen uptake activity (Table 4). Onthe other hand, mutant strains SU21 and SU17 possessedsymbiotic hydrogenase activities comparable to that of thewild-type strain SU. All strains reisolated from soybeannodules were resistant to rifampin, streptomycin, and ka-namycin, indicating that they retained the transposon inser-tion.

TABLE 3. H2 and CO2 uptake rates of whole cells of strainSU and various mutant strains

Strain H2 uptake (nmol/h CO2 uptake (nmol/minStrain per 108 cells)a per mg of protein)b

SU 105 11.0 ± 0.8

Class 1SU59 <0.5 4.9 ± 1.1SU47 <0.5 4.7 ± 1.2SU21 <0.5 6.4 ± 0.7SU17 <0.5 4.6 ± 1.1

Class 2SU44 83 2.4 1.2SU55 69 2.2 ± 0.9SU14 64 6.5 ± 1.7SU28 71 4.2 ± 1.0SU46 55 7.2 ± 1.3SU14 59 5.9 ± 0.9SU19 70 3.6 ± 1.0a Average value of two independent determinations.bMean ± standard deviation for five replicate samples.

To determine whether derepression of H2 uptake activityin strains SU21 and SU17 could be hypersensitive to repres-sion by 02, both strains were derepressed under both 2.0 and0.4% partial pressures of oxygen (Table 4). Mutants SU21and SU17 had low levels of Hup activity when derepressedunder 2.0% 02, but much higher levels of activity under0.4% 02 (Table 4). These mutants are therefore similar topreviously isolated mutants of strain SR classified as hyper-sensitive to 02 (21). The low 02 tension found in the rootnodule is probably the reason that strains SU21 and SU17have hydrogenase activity as bacteroids.

Nitrogenase activities of Tn5-induced Aut- mutants. Bothex planta and symbiotic nitrogenase activities of the TnS-induced Aut- strains were determined. All class 1 mutants(Aut- Hup-) possessed both ex planta and symbiotic nitro-genase activities which ranged from 35 to 93% (ex planta)and 76 to 100% (symbiotic) of that of the parent strain SU(Table 5). In addition, all class 2 (Aut- Hup+) mutants testedhad symbiotic nitrogenase activities. However, in the free-living state, only five of the seven were inducible fornitrogenase activity. The free-living nitrogenase activitiesranged from 42 to 113% of the parent strain's activity. Theother two Aut- Hup+ strains, SU19 and SU55, totallylacked ex planta nitrogenase activity (Table 5). These twostrains were also deficient in CO2 fixation activity, althougha third strain (SU44), which took up CO2 poorly, could beinduced to express ex planta nitrogenase activity.

TABLE 4. Effect of 02 on Hup activity of TnS-inducedHup- mutants (class 1)

Hydrogenase activity

Strain Free living (nmol of Symbiotic(pemol ofH2/h per 108 cells) H2/h per mg of

2.0% 02 0.4% 02 protein)

SU (wild type) 142 158 3.77SU21 <2.0 9.5 3.80SU17 <2.0 15.2 3.42SU47 <1.0 <1.0 <0.1SU59 <1.0 <1.0 <0.1SU27 <1.0 <1.0 <0.1

360 HOM ET AL.

on July 8, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

Tn5-GENERATED B. JAPONICUM MUTANTS 361

TABLE 5. Symbiotic and ex planta nitrogenase activities ofTn5-induced Hup- and Aut- mutants

Nitrogenase activity

Strain Ex planta (nmol Symbiotic (,umolof C2H4/h per of C2H4/h per g

108 cells) of nodule wt)

SU (Hup+ Nif+) 2.3 5.0

Class 1 (Hup-)SU59 1.5 4.4SU21 0.8 4.4SU27 1.2 5.0SU47 1.7 4.8SU17 2.1 3.8

Class 2 (Aut-)SU46 2.6 NDaSU24 1.5 NDSU19 0.0 3.9SU44 1.2 5.0SU28 1.1 4.3SU14 1.0 NDSU55 0.0 2.7- ND, Not determined.

Transfer of the cosmid pSH22 into Tn5-induced Hup- andAut- mutants of B. japonicum. Since the cosmid pSH22possessed a gene involved with both hydrogenase and nitro-genase activities as well as at least one gene involvedexclusively with hydrogenase activity (12), it was tested forits ability to complement the TnS-induced Aut- Hup- andAut- Hup+ mutants. Cosmid pSH22 was mobilized from E.coli to the 12 TnS-induced mutants (Table 6), and Tcr, Smr,and Kmr transconjugants were isolated and tested for theirability to grow chemoautotrophically. Aut+ Tcr transconju-gants of SU47 and SU59 were detected at a frequency of 1.0per cosmid transfer (Table 6). Strain SU27 was also comple-mented at a frequency of 1.0, but that transconjugant grewmore slowly than either of the other transconjugants or theparent strain. It is not clear why this strain grew moreslowly, but amperometric assays of Hup activity (Table 7)

TABLE 6. Complementation of Tn5-induced Hup- andAut- mutants with cosmid pSH22

Strain TcW/recipient Aut+/Tcr

Class 1 (Hup-)SU59 8.4 x 10-4 1.0SU47 3.0 x 10-4 1.0aSU27 3.2 x 10-41.nbSU21 5.5 x 10-3 <2 x 10-3SU17 3.7 x 10-4 <2 x 10-3

Class 2 (Aut-)SU46 5.0 x 10-6 <1.4 x 10-2SU24 5.4 x 10- <1.4 x 10-SU19 (Aut- Nif)c 3.2 x 10-6 <1.4 x 10-2SU44 6.4 x 10-6 <1.0 x 10-2SU28 5.0 x 10-6 <1.2 x 10-2SU14 4.8 x 10-6 <1.7 x 10-2SU55 (Aut- Nif-)c 2.9 x 10-' <1.0 X 10-2a Previously reported.6 Although functionally complemented, cells grew at least twice as slowly

as did wild type or other complemented mutants.c Not able to reduce acetylene in free-living culture.

TABLE 7. Hydrogen uptake activities of Tn5-inducedHup- mutants harboring cosmid pSH22a

Hydrogen uptakeStrain Phenotype activity (nmol of H,

oxidized/h per 108 cells)

SU Hup+ 104

SU59 Aut- Hup- <1.0SU59-A(pSH22) Aut+ 79SU59-B(pSH22) Aut+ 99

SU27 Aut- Hup- <1.0SU27A(pSH22) Aut+ 18

SU47 Aut- Hup- <1.0SU47A(pSH22) Aut+ 100SU47B(pSH22) Aut+ 66

a Transconjugants were derepressed for hydrogenase activity as describedpreviously (12) and assayed amperometrically for H, uptake activity withoxygen as electron acceptor. Each value is the average of data from twoindependent derepression vials.

show that the transconjugant SU27(pSH22) expressed onlyabout 18% of the level of Hup activity of the parent strain.Aut+ Tcr transconjugants of strains SU47 and SU59 whichcontained pSH22 were also tested for hydrogen uptakeactivity in free-living culture (Table 7). Both transconjugantstrains possessed substantial levels of hydrogen uptakeactivity; thus, the cosmid appears to fully complement thesestrains to the wild-type phenotype. In contrast to the resultsobtained with strains SU27, SU59, and SU47, pSH22 com-plemented neither the two remaining class 1 (Hup-) mutants(strains SU21 and SU17) nor any of the seven class 2 Aut-Hup+ mutants.Complementation by subclones of pSH22. Southern hybrid-

ization of TnS to EcoRI digests of genomic DNA showedthat the TnS insertions in SU59, SU27, and SU47 were indifferent EcoRI fragments (data not shown). EcoRI frag-ments of pSH22 (Fig. 1) were subcloned into plasmidpRK290, and each subclone was mated into SU59 and SU47:Because of its slow rate of growth even when comple-mented, SU27 was not tested. The two mutants were com-plemented by different EcoRI subclones of pSH22, and eachmutant could be complemented by only one subclone of thefive tested (Table 8). SU47 was complemented to autotrophyat a frequency of 1.0 per plasmid transfer by the 13.2-kilobase (kb) fragment (pDN11). In contrast, SU59 was

R RR B B B R R B R

H H HH

R = EcoRI IKbH = Hind IIIB = BomHI

R RR- R pDN28

R R:. pDN4IO

R R-_ pDNII

R R- 4 pDN211

R Ri-4 pDN38

FIG. 1. Restriction map and subclones of cosmid pSH22 (12).Procedures for mapping and subcloning are described in Materialsand Methods.

VOL. 54, 1988

on July 8, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

TABLE 8. Complementation of selected mutantsby subclones of pSH22a

Subclone SU47 (Hup-) SU59 (Hup-)

pDN11 +pDN211 - +pDN38 -

pDN28 -

pDN410 -

a +, Ratio of 1.0 Aut+/Tcr, when 500 tetracycline-resistant colonies werescreened; -, <5.0 x 10'3 Aut+/Tcr.

complemented to autotrophy by the 3.0-kb subcloned frag-ment (pDN211), again at a frequency of 1.0 per plasmidtransfer.

DISCUSSION

B. japonicum mutants defective in the hydrogen uptakesystem (Hup-) have been isolated by transposon Tn5 muta-genesis. TnS-induced mutants were identified initially bytheir inability to grow chemoautotrophically (Aut-) and thenwere assayed for the ability to oxidize H2 or to fix CO2 infree-living culture. Of 12 Aut- mutants, 5 possessed no Hupactivity with either 02 or methylene blue provided as anelectron acceptor. This Hup- phenotype, completely lack-ing hydrogen uptake activity with either natural or artificialelectron acceptors, has been isolated previously by chemicalmutagenesis (18).Three of these five Tn5-induced Aut- Hup- mutant

strains (SU59, SU27, and SU47) were also Hup- as bacte-roids from soybean nodules and as such were similar to thepreviously described mutants B. japonicum SR140 (18) andB. japonicum PJ17 (16). In this study, the nifihup cosmidpSH22 (12) complemented each mutant strain (SU59, SU47,and SU27) to autotrophy at a frequency of 1.0 per cosmidtransfer. Furthermore, different EcoRI subclones of pSH22were required to complement mutants SU47 and SU59 toautotrophy. pDN11 (13.2 kb) and pDN211 (3.0 kb) comple-mented SU47 and SU59, respectively, at a frequency of 1.0per plasmid transfer. Thus, pSH22 must encode at least twoseparate diffusible products involved in the Hup+ pheno-type, one contained within the 3.0-kb EcoRI fragment andone contained within the 13.2-kb EcoRI fragment. Thesegenes must be different from the niflhup gene located onpSH22 as described by Hom et al. (12), since SU47 andSU59 are Hup- but Nif+ in phenotype.Lambert et al. (14) have independently isolated a clone,

pHU52, which also encodes multiple hup-related genes.Recently, two adjacent restriction fragments of this plasmidhave been subcloned and shown to contain the genes whichencode the two subunits of hydrogenase (30). Although theywere isolated independently, restriction maps of pSH22 andpHU52 indicate that extensive similarity exists between thetwo. Interestingly, the fragments of pSH22 which comple-ment SU47 and SU59 (pDN11 and pDN211) seem to encom-pass the regions of pHU52 found to encode the subunits ofhydrogenase. Because both clones, particularly pDN11, arelarge, the mutations they complement in SU47 and SU59 donot necessarily lie within the hydrogenase structural genes.When clone pHU1, which overlaps pHU52, was mutagen-ized with Tn5, Hup-related sequences were shown to span atleast 15 kb (11). Unfortunately, attempts to assign specifichydrogenase subunit deficiencies to strains SU47 and SU59by using subunit-specific antibodies (26) have not beensuccessful (W. A. Sray and R. J. Maier, unpublished data),

so that the specific characteristics complemented by thesubclones are unknown.The other two TnS-induced Hup- mutants (SU21 and

SU17) had wild-type levels of Hup activity as bacteroids.Thus, they resemble the B. japonicum SR Hup- mutantswhich are hypersensitive to repression of H2 oxidation by 02(21). In addition, like those hypersensitive mutants, SU17and SU21 possessed significant levels of Hup activity whenthey were derepressed for hydrogenase in free-living cultureunder reduced oxygen tensions. These mutants were notcomplemented to Hup+ by the cosmid pSH22.Both types of TnS-induced Hup- mutants described above

possessed wild-type levels of both ex planta and symbioticnitrogenase activities. Thus, in these five Hup- mutants, theTn5 insertion events disrupted hup genes exclusively in-volved in the expression of Hup activity.Two of the seven Aut- Hup+ mutants (SU19 and SU55)

totally lacked ex planta nitrogenase activity. These twomutants were deficient in CO2 fixation activity. An addi-tional mutant (SU44), although deficient in CO2 fixationactivity, expressed ex planta nitrogenase activity. Furthercharacterization of these three mutants may provide valu-able information concerning the biochemical or geneticrelationship between anabolic carbon metabolism and nitro-gen fixation. The remaining Aut- Hup+ mutants expressednitrogenase activities both ex planta and symbiotically. Inaddition, they possessed CO2 fixation activity. Further char-acterization is necessary to pinpoint the defect which causestheir Aut- phenotype. The TnS insertion mutations de-scribed here may also be useful for the isolation of genesinvolved in H2 metabolism and N2 fixation.

ACKNOWLEDGMENTS

This work was supported by a grant from the Allied ChemicalCorp.We are grateful to Robert Keefe for his Southern analysis of the

mutants and to Mark O'Brian for critical reading of the manuscript.

LITERATURE CITED

1. Agarwal, A. K., and D. L. Keister. 1983. Physiology of ex plantanitrogenase activity in Rhizobium japonicum. AppI. Environ.Microbiol. 45:1592-1601.

2. Bishop, P. E., J. G. Guevarra, J. A. Engelke, and H. J. Evans.1976. Relation between the glutamine synthetase and nitroge-nase activities in the symbiotic association between Rhizobiumjaponicum and Glycine max. Plant Physiol. 57:542-546.

3. Boyer, H. W., and D. Roulland-Dussoix. 1969. A complementa-tion analysis of the restriction and modification of DNA inEscherichia coli. J. Mol. Biol. 41:459-472.

4. Ditta, G. 1985. Tn5 mapping of Rhizobium nitrogen fixationgenes. Methods Enzymol. 118:519-528.

5. Ditta, G., S. Stanfield, D. Corbin, and D. R. Helinski. 1980.Broad host range DNA cloning system for gram negativebacteria: construction of a gene bank of Rhizobium meliloti.Proc. NatI. Acad. Sci. USA 77:7347-7351.

6. Dixon, R. 0. D. 1972. Hydrogenase in legume root nodulebacteroids: occurrence and properties. Arch. Mikrobiol. 85:193-201.

7. Emerich, D. W., T. Ruiz-Argueso, T.-M. Ching, and H. J.Evans. 1979. Hydrogen-dependent nitrogenase activity and ATPformation in Rhizobium japonicum bacteroids. J. Bacteriol.137:153-160.

8. Graham, L. A., L. W. Stults, and R. J. Maier. 1984. Nitroge-nase-hydrogenase relationships in Rhizobium japonicum. Arch.Microbiol. 140:243-246.

9. Hahn, M., L. Meyer, D. Studer, B. Regensburger, and H.Hennecke. 1984. Insertion and deletion mutations within the nifregion of Rhizobium japonicum. Plant Mol. Biol. 3:159-168.

362 HOM ET AL.

on July 8, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

VnS-GENERATED B. JAPONICUM MUTANTS 363

10. Hanus, F. J., R. J. Maier, and H. J. Evans. 1979. Autotrophicgrowth of H2-uptake positive strains of Rhizobiumjaponicum inan atmosphere supplied with hydrogen gas. Proc. Natl. Acad.Sci. USA 76:1788-1792.

11. Haugland, R. A., M. A. Cantrell, J. S. Beaty, F. J. Hanus, S. A.Russell, and H. J. Evans. 1984. Characterization of Rhizobiumjaponicum hydrogen uptake genes. J. Bacteriol. 159:1006-1012.

12. Hom, S. S. M., L. A. Graham, and R. J. Maier. 1985. Isolationof genes (niflhup cosmids) involved in hydrogenase and nitro-genase activities in Rhizobium japonicum. J. Bacteriol. 161:882-887.

13. Hom, S. S. M., S. L. Uratsu, and F. Hoang. 1984. TransposonTnS-induced mutagenesis of Rhizobium japonicum yielding a

wide variety of mutants. J. Bacteriol. 159:335-340.14. Lambert, G. R., M. A. Cantrell, F. J. Hanus, S. A. Russell,

K. R. Haddad, and H. J. Evans. 1985. Intra- and interspeciestransfer and expression of Rhizobium japonicum hydrogenuptake genes and autotrophic growth capability. Proc. Natl.Acad. Sci. USA 82:3232-3236.

15. Lepo, J. E., F. J. Hanus, and H. J. Evans. 1980. Chemoautotro-phic growth of H2 uptake-positive strains of Rhizobium japoni-cum. J. Bacteriol. 141:664-670.

16. Lepo, J. E., R. E. Hickok, M. A. Cantrell, S. A. Russell, andH. J. Evans. 1981. Revertible hydrogen uptake-deficient mu-tants of Rhizobium japonicum. J. Bacteriol. 146:614-620.

17. 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.

18. Maier, R. J. 1981. Rhizobiumjaponicum mutant strains unableto grow chemoautotrophically with H2. J. Bacteriol. 145:533-540.

19. Maier, R. J., L. Graham, R. G. Keefe, T. PihI, and E. Smith.1987. Bradyrhizobium japonicum mutants defective in nitrogenfixation and molybdenum metabolism. J. Bacteriol. 169:2548-2554.

20. Maier, R. J., and S. S. M. Hom. 1985. H2 uptake negative

(Hup-) mutants of Rhizobium and their use in the isolation ofhup genes. Methods Enzymol. 118:528-536.

21. Maier, R. J., and D. M. Merberg. 1982. Rhizobium japonicummutants that are hypersensitive to repression of H2 uptake byoxygen. J. Bacteriol. 150:161-167.

22. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

23. Moshiri, F., L. Stults, P. Novak, and R. J. Maier. 1983. Nif-Hup- mutants of Rhizobium japonicum. J. Bacteriol. 155:926-929.

24. Schubert, K. R., and H. J. Evans. 1976. Hydrogen evolution: amajor factor affecting the efficiency of nitrogen fixation innodulated symbionts. Proc. Natl. Acad. Sci. USA 73:1207-1211.

25. 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.

26. Stults, L. W., F. Moshiri, and R. J. Maier. 1986. Aerobicpurification of hydrogenase from Rhizobium japonicum by af-finity chromatography. J. Bacteriol. 166:795-800.

27. Stults, L. W., E. B. O'Hara, and R. J. Maier. 1984. Nickel is acomponent of hydrogenase in Rhizobium japonicum. J. Bacte-riol. 159:153-158.

28. Vincent, J. M. 1970. A manual for the practical study of theroot-nodule bacteria. International Biological Program, Black-well Scientific Publications, Oxford.

29. Zablotowicz, R. M., S. A. Russell, and H. J. Evans. 1980. Effectof the hydrogenase system in Rhizobiurn japonicum on thenitrogen fixation and growth of soybeans at different stages ofdevelopment. Agron. J. 72:555-559.

30. Zuber, M., A. R. Harker, M. A. Sultana, and H. J. Evans. 1986.Cloning and expression of Bradyrhizobium japonicum uptakehydrogenase structural genes in Escherichia coli. Proc. Natl.Acad. Sci. USA 83:7668-7672.

VOL. 54, 1988

on July 8, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from