Vol. 56, No. 12

Evidence for a Third Uptake Hydrogenase Phenotype among

the Soybean BradyrhizobiaPETER VAN BERKUM

Nitrogen Fixation and Soybean Genetics Laboratory, Agricultural Research Service,U.S. Department ofAgriculture, HH-19, Building 011, BARC-West,

Beltsville, Maryland 20705

Received 11 July 1990/Accepted 2 October 1990

The existence of a hydrogen uptake host-regulated (Hup-hr) phenotype was established among the soybeanbradyrhizobia. The Hup-hr phenotype is characterized by the expression of uptake hydrogenase activity insymbiosis with cowpea but not soybean. Uptake hydrogenase induction is not possible under free-living culturalconditions by using techniques developed for uptake hydrogenase-positive (Hup+) Bradyrhizobium japonicum.Hydrogen oxidation by Hup-hr phenotype USDA 61 in cowpea symbioses was significant because hydrogenevolution from nitrogen-fixing nodules was not detected. An examination for uptake hydrogenase activity insoybean and cowpea with 123 strains diverse in origin and serology identified 16 Hup+ and 28 Hup-hrphenotype strains; the remainder appeared to be Hup-. The Hup-hr phenotype was associated with serogroups

31, 76, and 94, while strains belonging to serogroups 6, 31, 110, 122, 123, and 38/115 were Hup+. Hup+ strainsof the 123 serogroup typed positive with USDA 129-specific antiserum. The presence of the uptake hydrogenaseprotein in cowpea bacteroids of Hup+ strains was demonstrated with immunoblot analyses by using antibodiesagainst the 65-kDa subunit of uptake hydrogenase purified from strain SR470. However, the hydrogenaseprotein of Hup-hr strains was not detected. Results of Southern hybridization analyses with pHUl showed theregion ofDNA with hydrogenase genes among Hup+ strains to be similar. Hybridization was also obtained withHup-hr strains by using a variety of cloned DNA as probes including hydrogenase structural genes. Bothhydrogenase structural genes also hybridized with the DNA of four Hup- strains.

Biological nitrogen fixation may significantly enhance pro-

ductivity of soybean under conditions when combined Naccretion from the soil pool limits yield. The efficiency ofbiological nitrogen fixation in turn may affect productivity ofthe crop dependent upon N2 for growth. Many factorsinfluence the efficiency of biological nitrogen fixation. Hy-drogen is produced during the reduction of N2 to ammonia(9, 25). In most symbioses formed by soybean withBradyrhizobium japonicum, the energy and reducing powerlost through the formation of H2 cannot be recovered be-cause the H2 escapes to the soil atmosphere. Although mostsoybean symbioses evolve H2, a few strains of B. japonicumexpress uptake hydrogenase activity and oxidize the H2produced during N2 fixation (9, 25). Oxidation of H2 in B.japonicum increases biosynthesis of ATP (10), and theresulting increase in efficiency of symbiotic N2 fixationstimulates soybean yield (11).Bradyrhizobium strains capable of hydrogen oxidation

have the Hup+ phenotype. Investigations of uptake hydrog-enase have been predominantly with the Hup+ strain SR,which is a kanamycin- and streptomycin-resistant clone of asmall colony isolate of USDA 122. Studies of hydrogenoxidation regulation have predominantly relied upon theability of the Hup+ strains to express uptake hydrogenaseactivity in free-living culture either under chemoautotrophic(23, 26) or heterotrophic (38) growth conditions. Conse-quently, little is known about the regulation of uptakehydrogenase activity by soybean bradyrhizobia in symbio-sis. In contrast to B. japonicum, the expression of uptakehydrogenase activity in Rhizobium and Bradyrhizobium spp.of the cowpea miscellany group has been investigated undersymbiotic conditions, and the host plant was reported toinfluence hydrogen oxidation (8, 14).Keyser et al. (19) reported that two Hup- strains of B.

japonicum expressed uptake hydrogenase activity in symbi-osis with cowpea. These two strains (USDA 61 and USDA74) evolved H2 in soybean symbioses, and H2 oxidation bybacteroid preparations was not detected. However, theincidence of differential expression of uptake hydrogenaseactivity in B. japonicum and the genetic characterization ofthese strains has not been reported.My objectives were to investigate the relative incidence of

the uptake hydrogenase character among a diverse group ofsoybean bradyrhizobia. An examination for hydrogen oxida-tion in symbioses with cowpea as well as with soybean as

host plants was included to identify strains with differentialexpression of uptake hydrogenase activity because of hostinfluence. Selected strains were further characterized byidentifying regions of DNA coding for uptake hydrogenaseby using Southern hybridization analysis with cloned DNAcontaining hydrogenase genes of the Hup+ strain SR. Sincecontamination of a Hup- culture with a Hup+ strain able tonodulate one host but not the other may erroneously indicatedifferential expression of uptake hydrogenase activity, sev-

eral different methods were used to evaluate strain purity ofthe USDA 61 culture.

MATERIALS AND METHODSBacterial strains, media, and growth conditions. The strains

of soybean bradyrhizobia were from the U.S. Department ofAgriculture Agricultural Research Service National Rhizo-bium Culture Collection or from a survey of serogroupidentity and hydrogenase phenotype distribution in theUnited States (20). Bacteria were grown in MAG, which isAG (34, 39) modified by substituting potassium phosphatefor sodium phosphate (1.6 mM) and the addition of

Na2MoO4 .2H20 (0.04 mM) and NiCl2 (0.005 mM). Forserology, the bacteria were grown in yeast-salts arabinose

3835

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Dec. 1990, p. 3835-38410099-2240/90/123835-07$02.00/0Copyright © 1990, American Society for Microbiology

on June 16, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

broth (42) until early stationary phase. Stock cultures weremaintained on MAG slants after recovery of the bacteriafrom lyophilization and plating on MAG. Spontaneous anti-biotic-resistant clones were isolated by plating 0.1 ml of aturbid MAG broth culture (approximately 108 cells) on MAGcontaining 100 ,g of nalidixic acid per ml. Cultures withintrinsic resistance to tetracycline and rifampin were grownin the presence of one or the other antibiotic at 100 ,ug/ml.

Serology. Bacteria were grown to a density of approxi-mately 109 cells per ml, and the heat-labile flagellar antigenswere destroyed by heating the cultures to 100°C for 30 min.Fluorescent antibodies were used to determine the sero-group of each culture and were prepared with rabbit antiseraagainst serotype strains USDA 4, USDA 6, USDA 31,USDA 38, USDA 46, USDA 76, USDA 94, USDA 110,USDA 122, USDA 123, and USDA 135. Antisera againstUSDA 123 also cross-reacts with strains of serogroups 127and 129. Therefore, cultures belonging to serogroup 123were further classified with cross-absorbed antisera specificagainst USDA 127 and USDA 129.

Plant culture. Seeds of soybean (Glycine max L. Merr.)cv. Williams unless otherwise indicated and cowpea (Vignaunguiculata L.) cv. Black-eye Five were surface sterilizedwith 0.1% (wt/vol) acidified HgCl2 (42) for 3 min and washedfive times with tap water, which was previously determinedwith plant tests to be free of bradyrhizobia and rhizobia. Theseeds were sown in vermiculite moistened with 300 ml ofN-free nutrient solution (30) in modified Leonard jars (22)sterilized by autoclaving for 4 h. Each jar was inoculatedwith 5 ml of approximately 109 cells per ml of B. japonicumgrown in MAG. Each jar contained two soybean and twocowpea plants, which were grown for 6 weeks in a growthchamber set at day and night cycles of 25 and 20°C and 16and 8 h, respectively. Cowpea plants were also grown in thegreenhouse for 6 weeks in sterile vermiculite by using25-cm-diameter pots to obtain bacteroids for use in immu-noblotting analyses.

Acetylene reduction, hydrogen evolution, and plant nitro-gen determinations. Plants were removed from the Leonardjars or pots, the vermiculite was shaken off the roots, and thetops were excised at the cotyledonary node. The roots wereplaced in 50-ml stoppered syringes adjusted to 25- or 50-mlvolumes and incubated for 30 min before removing gassamples for H2 determinations. Subsequently, C2H2 wasinjected to 10% (vol/vol), and samples were incubated for 10min for the determination of nitrogenase activities. Concen-trations of C2H4 and H2 in the gas samples were determinedby gas chromatography as described previously (19, 40). Theplant tops were dried for 2 days at 60°C, weighed, milled,and analyzed for nitrogen concentration by using an Auto-mated Nitrogen Analyzer (Carlo Erba Instruments, Milan,Italy). Nitrogen fixation was determined from a comparisonof total nitrogen accumulation between inoculated and unin-oculated control plants.

Bacteroid preparation and measurement of uptake hydrog-enase activity. The roots were washed in tap water to removeadhering vermiculite, and total (fresh weight) nodule tissuewas determined. Approximately 200 mg (fresh weight) ofnodules was homogenized aerobically in 10 ml of phosphate-MgCl2 buffer (31). The homogenates were passed throughfour layers of cheesecloth, centrifuged at 4°C for 10 min at10,000 x g, and resuspended in 10 ml of buffer. Uptakehydrogenase activity of air-sparged bacteroid suspensionswas determined in the presence of 37 nmol of H2 at roomtemperature by using a 3-ml amperometric chamber (16).The absorbance at 540 nm of bacteroid suspensions were

used to determine dry weights from a standard curve.

Uptake hydrogenase activities by cultured bacteria were

determined with cells induced for activity while suspendedin buffer or grown heterotrophically as described by van

Berkum (38).Sodium dodecyl sulfate-polyacrylamide gel electrophoresis,

immunoblotting, and antigen detection. Cowpea bacteroids,washed and suspended in 50 mM phosphate buffer (pH 7.0)containing 1 mM phenylmethylsulfonyl fluoride were passedthree times through a French pressure cell at 2,700 kg/cm2.The supernatant fractions, after centrifugation at 10,000 x gfor 10 min at 4°C, were used either directly for immunoblotanalyses or were used to prepare membrane fractions.Membranes were sedimented at 40,000 x g for 2 h andresuspended in buffer containing phenylmethylsulfonyl flu-oride. Protein concentrations of the extracts were deter-mined by using the biuret method (35). Electrophoresis,immunoblotting, and the detection of the 65-kDa subunit ofuptake hydrogenase were as previously described (38). Thepresence or absence of the 33-kDa subunit of uptake hydrog-enase on the blots was not determined. Purified uptakehydrogenase was included in the analysis as a control andduplicate gels of the proteins resolved by sodium dodecylsulfate-polyacrylamide gel electrophoresis were stained withCoomassie Blue R-250.

Purification of uptake hydrogenase. Uptake hydrogenasewas purified from the mutant SR470 (28) by using the aerobicaffinity protocol described by Stults et al. (36).Genomic DNA isolation and Southern hybridization. The

bradyrhizobia were grown in a gyratory incubator at 150 rpmin 30 ml ofMAG from an inoculum of approximately 5 x 108cells for 3 days at 30°C. The cultures were centrifuged at10,000 x g for 10 min at 4°C, washed in TEN buffer (50 mMTris [pH 8.0], 20 mM disodium EDTA, 50 mM NaCI), andsuspended in 8 ml of TEN buffer. Lysozyme (5 mg) wasadded to each of the cell suspensions, which were incubatedat 37°C for 30 min before adding 1.0 ml of protease (Sigmatype XIV [5 mg/ml] predigested in TEN buffer for 60 min at37°C) to each. The mixtures were incubated at 37°C for 30min before adding 1.0 ml of 20% (wt/vol) Sarkosyl in TENbuffer. CsCl (11 g) and 0.6 ml of a 10-mg/ml (wt/vol) aqueoussolution of ethidium bromide were added to the cell lysatesafter allowing them to clear at 37°C for 1 h. The sampleswere centrifuged at 40,000 rpm for 48 h at 20°C in a 50 Tifixed-angle rotor (Beckman Instruments, Inc.). The high-molecular-weight DNA bands were collected, and ethidiumbromide was removed by using NaCl-saturated 1-butanol,followed by 24 h of dialysis against TE buffer (10 mM Tris[pH 7.5], 1 mM NaCl).Genomic DNA was digested with the restriction endonu-

clease EcoRI (Bethesda Research Laboratories, Gaithers-burg, Md.) as specified by the manufacturer. Restrictionfragments were separated by horizontal electrophoresis in0.7% (wt/vol) agarose gels in Tris-EDTA-borate buffer (27),stained with 0.5 jig of ethidium bromide per ml and photo-graphed under UV light.DNA transfer to Nytran (Schleicher and Schuell, Keene,

N.H.) and hybridization protocols were done as describedby Maniatis et al. (27). The clones with hydrogenase genes ofB. japonicum used as probes in the Southern hybridizationswere pHU1 (7); the 12.9- and 2.9-kb EcoRI subclones ofpSH22 (17), pDN11 and pDN211 (18), respectively; and thecloned hydrogenase structural genes (hupA and hupB)pKM15 or pKM16 and pKM20 (33).

3836 VAN BERKUM

on June 16, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

UPTAKE HYDROGENASE AMONG SOYBEAN BRADYRHIZOBIA

RESULTS

Verification of differential H2 oxidation in USDA 61. Anin-depth examination of USDA 61 for purity was necessary

to eliminate the possibility that differential H2 oxidation was

due to a Hup+ contaminant in the inoculum able to form a

symbiosis with cowpea but not with soybean. A nalidixicacid-resistant clone of USDA 61 was used for this purpose,

and isolations from the nodules of both host plants were

made on selective medium. Single colonies from the isola-tions on medium containing nalidixic acid were used toprepare inoculum for soybean and cowpea. Cowpea was

inoculated with isolates from soybean, and cultures obtainedfrom cowpea nodules were used for soybean. The coloniesisolated from soybean formed Hup+ symbioses with cow-

pea, and the inoculum originating from the cowpea noduleswere Hup- with soybean. The bacteria isolated from boththe cowpea and soybean nodules were resistant to nalidixicacid. Therefore, the cowpea host formed nodules with a

similar population of USDA 61 cells as the soybean host,which indicated that the nalidixic acid-resistant culture hadformed both symbioses. These results also indicated that theUSDA 61 culture was not contaminated with Hup+ soybeanstrains.The cultures and bacteroid preparations were also exam-

ined for their serology using fluorescent antibodies preparedagainst somatic cells of the nine major serogroups of soy-

bean bradyrhizobia. The cell and bacteroid suspensionswere positive with antisera prepared with USDA 31, whichis the serotype strain for the 31 or c3 serogroup. USDA 61belongs to serogroup 31, confirming its identity antigeni-cally. No cross-reaction with the other fluorescent antibod-ies was observed, indicating the culture was not contam-inated with strains belonging to the other serogroups tested.The identity and purity of the USDA 61 culture was also

confirmed by symbiotic performance with soybean hostshaving the dominant Rj4 gene. USDA 61 poorly nodulatesand forms ineffective symbioses with cultivars of soybeanhaving the variety Dunfield in their ancestry (41). Most otherstrains nodulate these soybean varieties normally and are

moderately or fully effective. The nalidixic acid-resistantsingle-colony isolates obtained from the nodules of soybeanand cowpea were tested for symbiotic performance with G.max cv. Hill, which has the Rj4 gene from Dunfield. Thesymbioses with Hill had few ineffective nodules, whilecowpea produced effective Hup+ symbioses. These resultsindicated that the differential expression of uptake hydroge-nase activity was due to host influence and could not beexplained by the contamination of the culture with Hup+bradyrhizobial strains.A time course analysis showed significant H2 evolution

and no uptake hydrogenase activity by USDA 61 nal inWilliams soybean (Fig. 1B and D). In contrast, the same

culture expressed uptake hydrogenase activity in the cowpeasymbiosis and H2 evolution was not detected. The degree ofuptake hydrogenase activity in the cowpea symbiosis was

significant because rates were sufficient to recycle the H2produced during N2 reduction. Nitrogen fixation and nitro-genase activity were similar in soybean and cowpea (Fig. 1Aand C), indicating similar rates of electron flow throughnitrogenase in both symbioses.

Incidence of differential H2 oxidation among soybeanstrains. Measurement of H2 oxidation in bacteroids formedby a diverse collection of strains with soybean and cowpeaestablished that of 123 strains tested, 16 oxidized H2 withboth host plants, 28 oxidized H2 with cowpea only, and 79

o1.15 73.

20 40 20 40

Days Days

FIG. 1. Nitrogenase activity (A), nitrogen fixation (C), hydrogenevolution (B), and uptake hydrogenase activity (D) by USDA 61 nalin symbioses with Williams soybean (0) and Black-eye Five cowpea

(0). Nitrogenase and hydrogen evolution measurements were madewith root and whole nodule samples. Nitrogen fixation was deter-mined from the difference in total nitrogen accumulation in the planttops between inoculated and uninoculated controls. Uptake hydrog-enase activity was determined with bacteroid preparations.

did not oxidize H2 with either host plant (Table 1). Differ-ential expression of uptake hydrogenase activity with cow-

pea but not with soybean was apparent in strains belongingto serogroups 31, 76, and 94; strains of serogroups 6, 110,122, 123, and 38/115 were Hup+. The Hup+ phenotype was

also present in several strains that did not cross-react withthe antisera used. One serogroup 31 strain was Hup+, andthe Hup+ strains within serogroup 123 typed positive withserogroup 129-specific antisera. Under free-living condi-tions, cultures of Hup+ strains were able to induce uptakehydrogenase activity by standard methods. The Hup-hrstrains failed to induce uptake hydrogenase activity as

free-living cultures.Immunoblot analysis of cowpea bacteroid proteins. Immu-

TABLE 1. Hydrogen oxidation by soybean bradyrhizobia insymbiosis with Glycine max cv. Williams and Vigna unguiculata

cv. Black-eye Five

No. of strains showingSerogroup No. of strains hydrogen oxidation on:tested

Soybean Cowpea

4 4 0 06 18 2 231 17 1 1546 3 0 076 14 0 894 9 0 6110 8 2 2122 3 2 2123 23 6 6135 4 0 038/115 4 1 1Nega 12 2 2a These strains did not cross-react with the antisera used in this study.

VOL. 56, 1990 3837

on June 16, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

FIG. 2. Southern hybridization analysis of B. japonicum with pHUl. (A) Hup+ strains. Lane 1, AK13 lc; lane 2, AK8 4a; lane 3, USDA110; lane 4, USDA 137; lane 5, USDA 423; lane 6, USDA 134; lane 7, SR. (B) Hup-hr strains. Lane 1, USDA 31; lane 2, USDA 61; lane 3,LA4c 2a; lane 4, AK7 la; lane 5, PA3 6c; lane 6, NC5 2a. (C) Hup- strains. Lane 1, USDA 16; lane 2, DE1 3a; lane 3, DE1 5a; lane 4, USDA185; lane 5, USDA 219; lane 6, USDA 38. Molecular sizes (in kilobases) are indicated.

noblot analyses using antisera prepared with hydrogenase ofstrain SR470 indicated the presence of the 65-kDa subunit incowpea bacteroid membrane preparations of Hup+ strains.However, uptake hydrogenase bands were not detected incowpea bacteroid membrane preparations of the Hup-hrstrains (data not shown). Failure to detect the hydrogenaseprotein with the immunoblot analysis in samples of theHup-hr strains was not due to the absence of uptake hydrog-enase. The uptake hydrogenase activities by whole bac-teroids of USDA 31, USDA 61, LA4c 2a, AK7 la, PA3 6c,and NC5 2a were 13.6, 22.8, 6.9, 4.5, 6.3, and 6.4 nmol of H2mg 1 (dry weight) h-1, respectively. Membrane prepara-tions were used in the immunoblotting analyses because theuptake hydrogenase activity of the USDA 61 cowpea bac-teroids was predominantly associated with this fraction. Thehydrogenase protein was not detected either in whole-cellextracts of USDA 31 and USDA 61 cowpea bacteroids,which had uptake hydrogenase activities of 70.9 and 30.9nmol of H2 mg-' (dry weight) h-1, respectively.

Southern hybridization analyses. Southern hybridizationanalyses were used to determine whether the hydrogenasegenes encoded in the genomes of the Hup-hr and Hup+strains share homology. The analyses also included Hup-strains to determine whether hydrogenase determinantswere present in some of these strains, even though noevidence for uptake hydrogenase capability was available.DNA containing hydrogenase genes cloned in pHU1 hybrid-ized with EcoRI-digested genomic DNA of Hup+ soybeanbradyrhizobia identifying six fragments (Fig. 2A). ProbespDN11, pKM15, and pKM20 hybridized with a 12.9-kbEcoRI fragment of Hup+ strains SR, USDA 110, and USDA137 (data not shown). These results show that the region ofDNA with hydrogenase genes among the Hup+ strains issimilar irrespective of serogroup. Exceptions were sero-group 129 strains USDA 422 (data not shown) and USDA

423, which had seven EcoRI fragments hybridizing withpHUL.Cosmid pHU1 also identified three EcoRI fragments with

homology in the genomes of strains expressing uptake hy-drogenase activity with cowpea only (Fig. 2B). It is unclearwhether these results represent the identification of hydrog-enase genes, because the DNA contained in pHU1 is large;only the 12.9-kb EcoRI fragment has been identified toencode two hydrogenase structural genes (12). The 12.9-kbEcoRI subclone ofpSH22 identical to that ofpHU1, pDN11,also hybridized with DNA of these strains (Fig. 3A). Hybrid-ization was also observed with cloned hydrogenase struc-tural genes in pKM16 (hupA) and in pKM20 (hupB) withpatterns which, when combined, resembled those obtainedwith pDN11 (Fig. 3C and D). The region of DNA sharinghomology with DNA cloned in pKM16 also hybridized withthe 2.9-kb EcoRI subclone of pSH22, pDN211 (Fig. 3B).

Positive results were also obtained when pHU1 was usedin Southern hybridization analyses with five of six Hup-phenotype strains (Fig. 2C). The cloned hupA gene inpKM15 hybridized with a 6.0- and a 5.0-kb EcoRI fragmentin the DNA of USDA 16, DE1 3a, and USDA 38, while onlya 6.0-kb EcoRI fragment was evident with USDA 185 (datanot shown). The cloned hupB gene in pKM20 hybridizedwith a 12.9- and 6.0-kb EcoRI fragment of the DNA inUSDA 16, USDA 38, and DE1 3a. The latter two strains alsoshowed a hybridization band of 5.0 kb with pKM20. InUSDA 185, pKM20 identified a 15.5- and 6.0-kb EcoRIfragment showing homology with hupB (data not shown).

DISCUSSION

Some strains of B. japonicum oxidize hydrogen in symbi-osis and are inducible for uptake hydrogenase activity asfree-living bacteria when grown under appropriate cultural

3838 VAN BERKUM

on June 16, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

UPTAKE HYDROGENASE AMONG SOYBEAN BRADYRHIZOBIA

A 1 2 3 4 5 6 B * 2 3 4 6 6S~~20.0

10.0

5.0-

4

10.0-

FIG. 3. Southern hybridization analysis of Hup-hr strains of B.japonicum with pDN11 (A), pDN211 (B), pKM16 (C), and pKM20(D). Lane identification is as for Fig. 2B. Molecular sizes (inkilobases) are indicated.

conditions (23, 26, 38). However, most soybean bradyrhizo-bia were considered incapable of hydrogen oxidation insymbiosis or as free-living bacteria (12). The existence ofthese two uptake hydrogenase phenotypes (Hup+ andHup-) among the soybean bradyrhizobia was identified byusing whole cells and analyses of hydrogen oxidation. Thiswork has shown the existence of an additional uptakehydrogenase phenotype among the soybean bradyrhizobia.The strains belonging to this third phenotype have hydrogenoxidation capability only in symbiosis with specific hostplants and are not capable of having hydrogen oxidationinduced as free-living bacteria by using standard methods.These characteristics are distinctly different from those usedto describe the Hup+ phenotype. Therefore, the term hydro-gen uptake host-regulated (Hup-hr) phenotype is suggestedto describe strains belonging to this third class of hydrogen-oxidizing bradyrhizobia.The host plant influences hydrogen oxidation by bac-

teroids in several host-strain combinations. The Hup+ strainof Rhizobium leguminosarum bv. viceae ONA 311 is Hup-in symbiosis with Vicia faba (8). Similarly, the pea hostFeltham First significantly suppressed rhizobial hydrogenoxidation in the Hup+ strains of R. leguminosarum 128C53and 3960 (2). Transmissible shoot factor(s) in the pea hostinfluenced expression of uptake hydrogenase activity by thebacteroids (3). A single dominant plant gene was reported toinfluence hydrogen oxidation by bacteroids in pea symbioses(32). Uptake hydrogenase activity by strains CB756 and32H1 of Bradyrhizobium sp. is also controlled by the hostplant (14), but in this case the genotype of the root is theprincipal factor determining expression of uptake hydroge-

nase activity. This work has shown that at least 28 soybeanstrains express uptake hydrogenase activity in symbioseswith cowpea and not with soybean, but host factors influ-encing uptake hydrogenase activity are unknown.The Hup-hr phenotype is associated with serogroups 31,

76 and 94; 75% of these strains had the capability ofexpressing uptake hydrogenase activity. Strains belonging tothese serogroups predominate in the soils of the southeast-ern USA (5, 13, 20, 43). Bradyrhizobia indigenous to soil andadapted to their environment may be less efficient fornitrogen fixation with soybean than inoculum-quality strains(13). Often it is impossible to introduce more efficient strainswith inoculation because the adapted soil population ishighly competitive for nodulation (4, 6, 15). Therefore, theapproach of increasing hydrogen oxidation in soybean byinoculating seed at sowing with Hup+ strains is probablyineffective. The identification of the Hup-hr phenotypeamong the soybean bradyrhizobia and the observation thatthis phenotype is associated with most of the adapted soilpopulation in Southern soils indicate that there is potentiallyan alternative approach for increasing hydrogen oxidation insoybean. This approach depends upon the identification ofsoybean germplasm permitting uptake hydrogenase expres-sion by the Hup-hr phenotype strains and backcrossing thistrait into agronomically adapted soybean cultivars. Theresulting progeny may have the potential of enhancinghydrogen oxidation with the indigenous Hup-hr strains.

Southern hybridization analyses identified DNA of Hup-hrstrains sharing homology with hydrogenase determinants ofHup+ strain SR. This observation is not unusual becausehydrogenase sequences in R. leguminosarum bv. viceae(24), Bradyrhizobium sp. (29), Rhodobacter capsulatus (21),Azotobacter chroococcum (37), and Pyrodictium brockii (33)have been reported to share homology with hydrogenasegenes of strain SR. A high degree of variation in restrictionfragment length was evident among the Hup-hr strains in theregion of the DNA encoding the hydrogenase determinants.This variation was also observed among strains within thesame serogroup. These results may indicate the existence ofgenetic variability among strains of serogroups 31, 76, and94.

Considering the DNA homology among hydrogenase se-quences of the two uptake hydrogenase phenotypes, it iscurious that immunoblotting results for the 65-kDa subunitof bradyrhizobial uptake hydrogenase were negative withcowpea bacteroid proteins of the Hup-hr strains. This resultwas unexpected because of the reported high degree ofhomology among the hydrogenases of B. japonicum, Alcali-genes eutrophus, Alcaligenes latus, Azotobacter vinelandii(1), and P. brockii (33). Failure to detect the hydrogenaseprotein of the Hup-hr strains cannot be explained by theabsence of the protein in bacteroids at the time of harvestbecause uptake hydrogenase activity was detected beforethe samples were prepared for electrophoresis. Also, it isunlikely that there was a problem with the immunoblottingtechnique because the hydrogenase protein was detected incontrol lanes, which included purified uptake hydrogenaseand membrane preparations of Hup+ strains of B. japoni-cum. Possible explanations for the failure to detect thehydrogenase protein of the Hup-hr strains include an uptakehydrogenase protein concentration below the level of detec-tion using enzyme-linked immunosorbent assay, the rapidbiodegradation of the uptake hydrogenase protein afterrupturing bacteroids of Hup-hr strains, or lack of homologybetween the portion of the SR470 protein against which the

VOL. 56, 1990 3839

on June 16, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

antibody was formed and a similar portion of the Hup-hrprotein.The survey of strains for hydrogen oxidation capability

indicated that most soybean bradyrhizobia are Hup- insymbiosis with soybean or cowpea. This agrees with thegeneral viewpoint that most soybean bradyrhizobia lackhydrogen oxidation capability (12). However, Southern hy-bridization analysis showed that at least four Hup- strainshave DNA with homology to both hydrogenase structuralgenes. It is possible that the Hup- phenotype is due to rapiddegradation of the gene products or the loss of regulatorysequences to activate transcription through the hydrogenasegenes. Alternatively, some Hup- strains may coincidentallycontain non-hydrogenase-related DNA with sequences ho-mologous to the cloned DNA which contains hydrogenasegenes.

ACKNOWLEDGMENTS

I thank Harold J. Evans for pHU1. I am also grateful to Robert J.Maier for antibodies to bradyrhizobial uptake hydrogenase, mutantSR470, and plasmids pDN11, pDN211, pKM15, pKM16, andpKM20. The technical assistance of Marian Pezzano is sincerelyappreciated.

LITERATURE CITED1. Arp, D. A., L. C. McCollum, and L. C. Seefeldt. 1985. Molecular

and immunological comparison of membrane-bound, H2-oxidiz-ing hydrogenases of Bradyrhizobium japonicum, Alcaligeneseutrophus, Alcaligenes latus, and Azotobacter vinelandii. J.Bacteriol. 163:15-20.

2. Bedmar, E. J., S. A. Edie, and D. A. Phillips. 1983. Host plantcultivar effects on hydrogen evolution by Rhizobium legminosa-rum. Plant Physiol. 72:1011-1015.

3. Bedmar, E. J., and D. A. Phillips. 1984. A transmissible plantshoot factor promotes uptake hydrogenase activity in Rhizo-bium symbionts. Plant Physiol. 75:629-633.

4. Boonkerd, N., D. F. Weber, and D. F. Bezdicek. 1978. Influenceof Rhizobium japonicum strains and inoculation methods onsoybeans grown in rhizobia-populated soils. Agron. J. 70:547-549.

5. Caldwell, B. E., and E. E. Hartwig. 1970. Serological distribu-tion of soybean root nodule bacteria in soils of southeasternUSA. Agron. J. 62:621-622.

6. Caldwell, B. E., and G. Vest. 1970. Effects of Rhizobiumjaponicum strains on soybean yields. Crop Sci. 10:19-21.

7. Cantrell, M. A., R. A. Haugland, and H. J. Evans. 1983.Construction of a Rhizobium japonicum gene bank and use inisolation of a hydrogen uptake gene. Proc. Natl. Acad. Sci.USA 80:181-185.

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

9. Eisbrenner, G., and H. J. Evans. 1983. Aspects of hydrogenmetabolism in nitrogen-fixing legumes and other plant-microbeassociations. Annu. Rev. Plant Physiol. 34:105-136.

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

11. Evans, H. J., F. J. Hanus, R. A. Haugland, M. A. Cantrell, L. S.Xu, S. A. Russell, G. R. Lambert, and A. R. Harker. 1985.Hydrogen recycling in nodules affects nitrogen fixation andgrowth of soybeans, p. 935-942. In R. Shibles (ed.), WorldSoybean Research Conference III: Proceedings. WestviewPress, Boulder, Colo.

12. Evans, H. J., A. R. Harker, H. Papen, S. A. Russell, F. J. Hanus,and M. Zuber. 1987. Physiology, biochemistry, and genetics ofuptake hydrogenase in rhizobia. Annu. Rev. Microbiol. 41:335-361.

13. Fuhrmann, J. 1990. Symbiotic effectiveness of indigenous soy-

bean bradyrhizobia as related to serological, morphological,rhizobitoxine, and hydrogenase phenotypes. Appl. Environ.Microbiol. 56:224-229.

14. Gibson, A. H., B. L. Dreyfus, R. J. Lawn, J. I. Sprent, and G. L.Turner. 1981. Host and environmental factors affecting hydro-gen evolution and uptake, p. 373. In A. H. Gibson and W. E.Newton (ed.), Current perspectives in nitrogen fixation. Austra-lian Academy of Sciences, Canberra, Australia.

15. Ham, G. E., V. B. Cardwell, and H. W. Johnson. 1971. Evalu-ation of Rhizobium japonicum inoculants in soils containingnaturalized populations of rhizobia. Agron. J. 63:301-303.

16. Hanus, F. J., K. R. Carter, and H. J. Evans. 1980. Techniquesfor measurement of hydrogen evolution by nodules. MethodsEnzymol. 69:731-739.

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

18. Hom, S. S. M., P. D. Novak, and R. J. Maier. 1988. TransposonTn5-generated Bradyrhizobium japonicum nmutants unable togrow chemoautotrophically with H2. Appl. Environ. Microbiol.54:358-363.

19. Keyser, H. H., P. van Berkum, and D. F. Weber. 1982. Acomparative study of the physiology of symbioses formed byRhizobium japonicum with Glycine max, Vigna unguiculata,and Macroptilium atropurpureum. Plant Physiol. 70:1626-1630.

20. Keyser, H. H., D. F. Weber, and S. L. Uratsu. 1984. Rhizobiumjaponicum serogroup and hydrogenase distribution in 12 states.Appl. Environ. Microbiol. 47:613-615.

21. Leclerc, M., A. Colbeau, B. Cauvin, and P. M. Vignais. 1988.Cloning and sequencing of the genes encoding the large and thesmall subunits of the H2 uptake hydrogenase (hup) ofRhodobacter capsulatus. Mol. Gen. Genet. 214:97-107.

22. Leonard, L. T. 1943. A simple assembly for use in the testing ofcultures of rhizobia. J. Bacteriol. 45:523-527.

23. Lepo, J. E., F. J. Hanus, and H. J. Evans. 1980. Chemoau-totrophic growth of hydrogen-uptake-positive strains of Rhizo-bium japonicum. J. Bacteriol. 141:664-670.

24. Leyva, A., J. M. Palacios, and T. Ruiz-Argueso. 1987. Con-served plasmid hydrogen-uptake (hup)-specific sequenceswithin Hup+ Rhizobium leguminosarum strains. Appl. Environ.Microbiol. 53:2539-2543.

25. Maier, R. J. 1986. Biochemistry, regulation, and genetics ofhydrogen oxidation in Rhizobium. Crit. Rev. Biotechnol. 3:17-38.

26. Maier, R. J., N. E. R. Campbell, F. J. Hanus, F. B. Simpson,S. A. Russell, and H. J. Evans. 1978. Expression of hydrogenaseactivity in free-living Rhizobium japonicum. Proc. Natl. Acad.Sci. USA 75:3258-3262.

27. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual, p. 109. Cold Spring HarborLaboratory, Cold Spring Harbor, N.Y.

28. Merberg, D., and R. J. Maier. 1983. Mutants of Rhizobiumjaponicum with increased hydrogenase activity. Science 220:1064-1065.

29. Murillo, J., A. Villa, M. Chamber, and T. Ruiz-Argueso. 1989.Occurrence of H2-uptake hydrogenases in Bradyrhizobium sp.(Lupinus) and their expression in nodules of Lupinus spp. andOrnithopus compressus. Plant Physiol. 89:78-85.

30. Norris, D. 0. 1964. Techniques used in work with Rhizobium.Commonw. Bur. Pastures Field Crops Hurley Berkshire Bull.47:186-198.

31. O'Brian, M. R., and R. J. Maier. 1982. Electron transportcomponents involved in hydrogen oxidation in free-living Rhizo-bium japonicum. J. Bacteriol. 152:422-430.

32. Phillips, D. A., Y. Kapulnik, E. J. Bedmar, and C. M. Joseph.1990. Development and partial characterization of nearly iso-genic pea lines (Pisum sativum L.) that alter uptake hydroge-nase activity in symbiotic Rhizobium. Plant Physiol. 92:983-989.

33. Pihl, T., R. N. Schicho, R. M. Kelly, and R. J. Maier. 1989.Characterization of hydrogen-uptake activity in the hyperther-mophile Pyrodictium brockii. Proc. Natl. Acad. Sci. USA

3840 VAN BERKUM

on June 16, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

UPTAKE HYDROGENASE AMONG SOYBEAN BRADYRHIZOBIA

86:138-141.34. Sadowsky, M. J., R. E. Tully, P. B. Cregan, and H. H. Keyser.

1987. Genetic diversity in Bradyrhizobiumjaponicum serogroup

123 and its relation to genotype-specific nodulation of soybean.Appl. Environ. Microbiol. 53:2624-2630.

35. Somasegaran, P., and H. J. Hoben. 1985. Methods in legume-Rhizobium technology, p. 115-116. U.S. Agency for Interna-tional Development document, University of Hawaii, NifTALProject, Paia, Hawaii.

36. 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.

37. Tibelius, K. H., R. L. Robson, and M. G. Yates. 1987. Cloningand characterization of hydrogenase genes from Azotobacterchroococcum. Mol. Gen. Genet. 206:285-290.

38. van Berkum, P. 1987. Expression of uptake hydrogenase and

hydrogen oxidation during heterotrophic growth of Bradyrhizo-bium japonicum. J. Bacteriol. 169:4565-4569.

39. van Berkum, P., and R. J. Maier. 1988. Lack of carbon substraterepression of uptake hydrogenase activity in Bradyrhizobiumjaponicum SR. J. Bacteriol. 170:1962-1964.

40. van Berkum, P., and C. Sloger. 1979. Immediate acetylenereduction by excised grass roots not previously preincubated atlow oxygen tensions. Plant Physiol. 64:739-743.

41. Vest, G., and B. E. Caldwell. 1972. Rj4-a gene conditioningineffective modulation in soybeans. Crop Sci. 12:692-694.

42. Vincent, J. M. 1970. A manual for the practical study ofroot-nodule bacteria. International Biological Programme hand-book no. 15. Blackwell Scientific Publications, Oxford.

43. Weber, D. F., H. H. Keyser, and S. L. Uratsu. 1989. Serologicaldistribution of Bradyrhizobium japonicum from U.S. soybeanproduction areas. Agron. J. 81:786-789.

VOL. 56, 1990 3841

on June 16, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from