Res. Microbiol. 152 (2001) 95103 2001 ditions scientiques et mdicales Elsevier SAS.

All rights reserved S0923-2508(00)01172-4/FLA

Improvement in the RFLP procedure for studying the diversity of nifH genes in communities of nitrogen xers in soil
Franck Poly , Lucile Jocteur Monrozier, Ren Bally
Laboratoire dcologie microbienne, UMR-CNRS 5557, UCB Lyon 1, 43, boulevard du 11 novembre 1918, 69622 Villeurbanne cedex, France Received 3 April 2000; accepted 8 August 2000

Abstract Several specic primers for the nifH gene were tested with different pure telluric N2 -xing strains. A PolF/PolR primer set provided successful amplication of 19 representative N2 -xing strains. Three restriction enzymes, HaeIII, NdeII and MnlI, chosen for restriction fragment length polymorphism (RFLP) analyses, were the most discriminating for the study of nifH gene diversity as they resulted in differences between strains at the species level. Amplication by selected primers and RFLP were applied to assess the genetic diversity of the nifH gene pool in soil. Pair soils, one under cultivation, the second under permanent pasture, were found to harbor a contrasting diversity of nifH genes. Pure strain proles could not be recognized in the nifH soil patterns. Using the simple procedure described, it was shown that the structure of nitrogen xers in soil was inuenced by soil functioning. 2001 ditions scientiques et mdicales Elsevier SAS PCR-RFLP / nifH / gene / soil / functional genetic diversity

1. Introduction Nitrogen xation is a process that enables reduction of the atmospheric nitrogen N2 in ammonium (NH+ 4 ) by nitrogenase, a universal enzyme. This process introduces nitrogen into the biosphere. The natural xing process is responsible for 65% of annual xation, while industrial processes represent only 25% [16]. Natural xation is accomplished by xing microorganisms only; these microorganisms belong to the Archaea and bacteria. Among the bacteria, the ability to x N2 has been shown in organisms with various metabolisms such as anaerobes and aerobes, cyanobacteria and actinomycetes [28]. Such diversity in physiology and ecology renders impossible the use of a universal selective culture medium for all nitrogen-xing microorganisms. These various microorganisms share the same operon in which the nifH gene encodes for the Fe protein subunit (= nitrogenase reductase) of the nitrogenase. It has been shown by Young [28] that many features of a nifH -based phylogenic tree are entirely consistent

Correspondence and reprints.

E-mail address: poly@biomserv.univ-lyon1.fr (F. Poly).

with the 16S rRNA-based phylogeny of the nitrogenxing bacteria. Therefore, the diversity of the nifH gene permits bulk representation of the taxonomic diversity of xing bacteria [24], and can be used to study the diversity of the bacterial community that can x nitrogen. Through molecular biology techniques, the diversity of nitrogen xers in a complex environment such as the soil can be estimated, avoiding the bias of culturing the organisms on synthetic media, a procedure known to select a low percentage of soil populations [26]. No recent work has been reported on the different nifH primers, their specicity on pure cultures and their use for describing the diversity of nitrogen xers in soils. However, Chelius and Lepo [8] report on approaching nifH diversity of the plant rhizosphere community, using primers dened for cyanobacteria. The objective of our study was to test several primers on different telluric bacterial genera well known to x nitrogen (positive test) or on microorganisms that fail to x nitrogen (negative test). We suggest a simple protocol for polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP), limiting the number of cycles to minimize the possible bias

96

F. Poly et al. / Res. Microbiol. 152 (2001) 95103

of PCR [25], so as to estimate the diversity of nitrogen xers through the diversity of nifH genes. 2. Materials and methods
2.1. PCR primers

Primers were chosen taking into account three parameters considered to be essential for the study of diversity by RFLP: i) nifH amplication on the widest range of N2 -xing species, ii) specicity of amplication; and iii) an amplication size large enough so as to obtain restriction fragment proles representative of the nitrogen xer diversity. Few regions of the nifH gene have been conserved among species of nitrogen-xing bacteria (data not shown). This explains why, in previous studies [13, 16, 18, 21, 23], all primers used were dened from only three main regions. The forward primers were chosen between nucleotides 19 and 47 (referring to Azotobacter vinelandii accession number M20568 [13]) or between nucleotides 112 and 134, with the corresponding amino acid sequences being A(I/F)YGKGGIGK and GCDPKADS, respectively. The reverse primers were dened between nucleotides 457 and 494 corresponding to the SGEMMA(M/L/T)YAANNI amino acid sequence. All primers had a similar sequence, with change in the number and type of degeneracies only (table I). First, we tried to amplify the selected nitrogenxing strains Azospirillum brasilense strain Sp7, Azospirillum lipoferum strain CRT1 [10], Rhizobium leguminosarum ATCC strain 14480, Sinorhizobium meliloti ATCC strain 9930 and Frankia alni strain ARI3 [7]) with already published primers (table I). These primers were modied (table I) to increase their specicity by reducing the degeneracy number but maintaining successful amplication over the widest range of bacterial genera of the cluster I branch of nifH phylogeny [6]. PCR amplications from pure strains (1 L of cells in glycerol/water 40% or 50 ng of DNA) were performed in a total volume of 50 L. The nal concentrations were 0.5 M of each selected primer, and 4 M for highly degenerate primers dened by Zehr and Mac Reynolds [29] (Zf and Zr) (table I) and 469/470 (this study). Other reagents were 200 mM of each dNTP, 2 U of Expand High Fidelity DNA polymerase (Boehringer Mannheim), 1 PCR buffer

as specied by the manufacturer and 1 g T4 gene 32 protein (Boehringer Mannheim) [15]. PCR conditions consisted of 30 cycles at 94 C (1 min), 1 min for the annealing step (temperature is presented in table II) and 72 C (2 min), with a 5-min extension at 72 C for the last cycle. All amplications were performed in a GeneAmp PCR System 2400 thermocycler (Perkin-Elmer, Cetus). Amplication products were submitted to electrophoresis in a 2% agarose gel and stained with ethidium bromide. Results of PCR amplication experiments are shown in table II. Depending on the primer combinations used, we obtained either: i) no PCR product; ii) no expected PCR product sizes; iii) expected product sizes only; or iv) expected product size with additional unexpected bands. Amplication was considered successful when only one PCR product with the correct size was detected. This was obtained with the PolF/PolR primer set (table I) successfully on DNA from all ve tested strains. To evaluate primer specicity and universality, this preliminary experiment was complemented with other reference nitrogen-xing strains representative for diverse bacterial sub-groups (table III) and nonnitrogen-xing strains: Agrobacterium rhizogenes ATCC 11325; Bacillus azotoformans ATCC 29788; Burkholderia cepacia ATCC 17759; Geodermatophilus obscurus ATCC 25080; Pseudomonas syringae var. pisi strain 895A (Horticulture Research International, Wellesbourne, UK, 1975); Pseudomonas viridiava ATCC 13222; Streptomyces lividans strain TK24 [12]; Escherichia coli strain DH5 (Life Biotechnologies, France); and Xanthomonas populi strain Spm9 (ATCC 27642) were also assayed. The PolF/ PolR primer set was able to amplify the nifH fragment out of the totality (19/19) of the tested xing strains. As expected, no amplication was obtained with negative control strains. The selected primers PolF/PolR were tested on bacterial DNA extracted from soil. Soil samples were collected from the upper layer (020 cm) of paired soils, i.e. soils of the same type under permanent pasture (LCSA-p), and under maize cultivation (LCSAc) in close eld situation. One hundred nanograms of DNA, extracted and puried as described [20], were used for soil DNA amplication. A single band of 360 bp was obtained from soil DNA (data not shown). The specicity of this band was veried by hybridization of the PCR product with a probe for the nifH gene. For this the nifH gene of Azospirillum

Table I. Sequences of various primers tested in previous work or in this study for nifH gene amplication.

References [24] [27] [21] This study [17] [27] This study [29] This study This study

Name of primer 19Fb For A FGPH19b IGK IGK 1b For B 469b Zfb Kadinob PolFb

Sequence positiona 19 19 25 25 31 112 112 115 115 115

Number of degeneracies 16 128 24 24 384 96 128 128 8 24

Forward primers GCI GCI WTY WTI TAY TAY TAC TAC GAY GAY GAY GAY GAY GGI GGN GGC GGY CCN CCN CCN CCI CCS AAR AAR AAR AAR AAR AAV AAR AAR AAR AAR GGI GGN GGT GGB GGN GCN GC GCN GCI GCB GG GG GGN GGY GGN GA

F. Poly et al. / Res. Microbiol. 152 (2001) 95103

ATH ATC ATH

G GG GGN

AA

GGI GGN

TGT TGY TGY TGY TGC

GA GA GAC TC Reverse primers

279 16 CTC CGG GCC RCC NGA YTC [21] FGPH273b 407 8 AAI CCR CCR CAI ACI ACR TC [24] 407Rb 436 2 G ACG ATG TAG ATY TCC TG This study AQEb [29] Zrb 460 96 ADN GCC ATC ATY TCN CC b 457 8 ATS GCC ATC ATY TCR CCG GA This study PolR 460 512 TAN ANN GCC ATC ATY TCN CC This study 470b 463 4 GCR TAI AII GCC ATC ATY TC This study Eminob [27] Rev 463 48 GCR TAI ABN GCC ATC ATY TC [17] YAA 478 256 YAA ATR TTR TTN GCN GCR TA a Sequence position with reference to the A. vinelandii nifH coding sequence (Genbank accession number M20568). b Primers tested in this study. The International Union of Pure and Applied Chemistry Conventions was used to described DNA sequence degeneracies: Y = C/T; S = G/C; R = A/G; B = C/G/T; D = G/A/T; H = T/C/A; N = A/G/C/T; W = A/T; I = inosine.

97

98

F. Poly et al. / Res. Microbiol. 152 (2001) 95103

Table II. Results of nifH amplications with different primer sets using ve reference strains of nitrogen-xing bacteria.

Positive controls

Final concentration

Annealing temperature ( C)

A. brasilense ATCC 29145

A. lipoferum [10]

R. leguminosarum ATCC 14480

S. meliloti ATCC 9930

F. alni ARI3 [7]

Primer sets IGK1/FGPH273 0.5 M 55 ND +m ND IGK1/AQE 0.5 M 48 + + w + IGK1/PolR 0.5 M 55 + ND w +m ND 19F/407R 0.5 M 50 + +m ND ND ND FGPH19/FGPH273 0.5 M 55 + + + Kadino/Emino 0.5 M 50 ND ND ND 469/470 0.5 M 45 + ND 469/470 4 M 55 + +m w w Zf/Zr 0.5 M 57 w +m w Zf/Zr 4 M 57 +m ND +m ND ND PolF/AQE 0.5 M 48 + ND +m ND ND PolF/PolR 0.5 M 55 + + + + + ND, not determined; +, PCR product of the expected size; , no PCR product; w, no expected PCR products, only products of unexpected and nonspecic size; +m, expected PCR products plus other products of unexpected and nonspecic size.

Table III. Bacterial strains.

References (source) Proteobacteria A. lipoferum CRT1 [10] A. brasilense sp. 7 ATCC 29145 A. brasilense L4 [5] A. brasilense sp. 245 [2] A. irakense KAC5 [14] A. irakense KBC1 [14] A. caulinodans ATCC 43989 S. paucimobilis 5AJ [3] Rhizobium legumirosarum ATCC 14480 M. loti LMG 6125 S. meliloti ATCC 9930 Proteobacteria A. tolulyticus ATCC 51758 B. vietnamiensis TVV75 [11] Proteobacteria K. oxytoca 1ABI [3] E. cloacae 7ATR [4] Gram + high GC% F. alni ACN14 [9] F. alni ARI3 [7] Gram + low GC% P. polymyxa CF43 [1] P. polymyxa PMD230 [1]

Subgroup

Organism

PCR product, 32 P-labelled by random priming (Boehringer Mannheim kit), was used as the probe. Previous work had already demonstrated probe specicity and reported that it did not hybridize with negative controls, chromosomal DNA of non-nitrogen-xing bacteria and with non-nifH PCR products (data not shown). All PCR products of 360 bp obtained with PolF/PolR from DNA of reference strains or from DNA of soil bacteria hybridized with the nifH gene probe (data not shown).
2.2. Enzymes for nifH RFLP

brasilense cloned in pAb1 plasmid, as described [19], was amplied with the primer set PolF/PolR. The

To use PCR-RFLP analysis on nifH pool genes from nitrogen-xing populations as a tool to study the community structure, it was necessary to choose endonucleases providing fragments in the range of 20 360 bp (fragment sizes smaller than 20 bp are difcult to detect). For adequate separation, the resulting fragments must be distributed throughout the size range. Enzymes that generated the same size band from various nitrogen xing species must be rejected as this may hide the complexity of the nifH gene pool. Thus, the frequency of apparition for each fragment must be low (<10%) (gure 1). Six different enzymes were used for RFLP assays. Ten microliters of each PCR product were directly used for restriction enzyme cleavage by

F. Poly et al. / Res. Microbiol. 152 (2001) 95103

99

Figure 1. Frequencies (% of occurrence in tested strains or sequences) of nifH gene restriction fragments obtained with four enzymes (A, HinfI; B, HaeIII; C, NdeII; D, Mnl I) from pure strain DNAs together with fragments obtained from theoretical digestions of database sequences.

MnlI, Fnu4HI, HaeIII, TaqI, NdeII, HinfI (Boehringer Mannheim). Digestions were performed overnight to ensure that complete fragmentation was achieved. Digested DNAs were analyzed in a 5% polyacrylamide gel (19:1) (Bio-Rad, Ivry-sur-Seine, France) using manufacturers recommendations and staining with 1 Syber Green (FMC BioProducts). This experiment was completed with theoretical digestions simulated from 14 sequences of the nifH gene found in the GenBank database: Anabaena sp. strain PCC 7120 (X05595 accession number); Alcaligenes faecalis pBZ1 (X96609); Azoarcus sp BH72 (Y12545); Azospirillum brasilense Sp7 (X51500); Azotobacter chroococcum ATCC 4412 (M73020); Azotobacter vinelandii M20568 [13]; Bradyrhizobium japonicum USDA 110 (K01620); Fischerella sp UTEX 1931 (U49514); Frankia sp FaC1 (X73983); Klebsiella pneumoniae (J01740) [22]; Rhizobium meliloti

(J01781) [23]; Rhodobacter sphaeroides 16PHC (AF031817); Thiobacillus ferrooxidans ATCC 33020 (M15238); Trichodesmium thiebautii (U23507). The main features of the occurrence (% of tested strains and sequences) of the restriction fragments obtained by four of the six enzymes are shown in gure 1. 3. Results and discussion All six enzymes discriminated pure strains, except for HinfI, which gave similar patterns with the various reference strains (gure 1A). Obviously, HinfI restriction targeted two sites highly conserved in nifH gene sequences and, in most cases, produced three fragments ( 15 bp, 130 bp and 210 bp) from amplicons (gure 1A). This endonuclease could be used to conrm the nifH origin of a PCR product, but not to discriminate between strains. Three enzymes, HaeIII,

100

F. Poly et al. / Res. Microbiol. 152 (2001) 95103

Figure 2. HaeIII RFLP of nifH PCR products from pure strains and soil DNA.

NdeII and MnlI (gure 1B, C, D) were selected on the basis of the above arguments to study nifH pool diversity: HaeIII was used in their study on the nifH gene pool by Widmer et al. [27] and by Chelius and Lepo [8], while NdeII and MnlI were applied here for the rst time to the nifH study. Pure strains from N2 xers from various genera were discriminated by RFLP resulting from the three selected enzymes HaeIII (gure 2), NdeII (gure 3) and MnlI (gure 4). Slight differences between nifH proles of (i) Azospirillum irakense and Azoarcus tolulyticus, and (ii) Rhizobium leguminosarum and Mesorhizobium loti resulted from the presence of small fragments (<40 bp). Taking into account the <40-bp size fragments, HaeIII, MnlI and NdeII RFLP showed differences in proles not only between closely related genera such as Rhizobium, Sinorhizobium and Mesorhizobium, but also between species of the same genera such as Azospirillum brasilense,

A. lipoferum and A. irakense (gures 24). Depending on the enzymes used and the species considered, strains of the same species could be discriminated as well. HaeIII proles differentiated two strains of Frankia alni (ARI3 and ACN14A) and two strains of Paenibacillus polymixa (CF43 and PMD 230) (gure 2). HaeIII digestion did not distinguish between Sp7 and L4 strains of A. brasilense, whereas these two strains were distinguished by MnlI and NdeII digestion. In contrast, MnlI and NdeII presented identical proles of the two strains of P. polymixa (gures 3 and 4). Complementary use of HaeIII, MnlI and NdeII restrictions was needed to access the differences in nifH sequences. Two strains of A. irakense (KbC1 and KaC5) showed the same proles for three of the assayed enzymes. The common origin of these two strains isolated from rice roots in the same eld could explain such a similarity in their nifH gene.

F. Poly et al. / Res. Microbiol. 152 (2001) 95103

101

Figure 3. NdeIII RFLP of nifH PCR products from pure strains and soil DNA.

The sum of the molecular weights of all fragments higher than the PCR products was obtained in Sphingomonas paucimobilis and Enterobacter cloacae RFLP proles. The presence of only one band in the PCR product and the positive hybridization with nifH probes excluded a nonspecic amplication. Such a discrepancy would indicate a substantial contamination of DNA from S. paucimobilis and E. cloacae by DNA of other nitrogen xers.
3.1. RFLP results on soil DNA

RFLP data concerning the amplication products from soil DNA digested by either endonuclease HaeIII, MnlI or NdeII were reproducible (data not shown). Each enzyme provided a specic prole for each soil DNA (gures 24). Comparison of soil and strain proles showed that some strain proles were found in soils such as A. brasilense Sp7 and A. tolulyticus with MnlI restriction in LCSA-c soil

or S. meliloti and Paenibacillus polymixa (CF43 and PMD230) with NdeII restriction in LCSA-p soil. However, soil DNA RFLP never resulted in proles identical to a pure strain on the three restriction enzymes. For instance, A. brasilense Sp 245 proles were found in LCSA-c soil with HaeIII and NdeII but not with MnlI. Moreover, several bands in soil DNA proles did not correspond with any band of the pure strains. These observations could lead to the conclusion that the pure strains used were not dominant in the environment and that there existed a number of nitrogen xers still unknown (nonculturable) [24, 27]. The protocol used revealed that the nifH gene pool, and probably the nitrogenxing community, were different under cultivation and under permanent pasture, even in the same soil type. A simple protocol with specic primers of the nifH gene and with three discriminating endonucle-

102

F. Poly et al. / Res. Microbiol. 152 (2001) 95103

Figure 4. Mnl I RFLP of nifH PCR products from pure strains and soil DNA.

ases was described here which allows rapid assessment of the structure of nitrogen xers in an ecosystem through the diversity of nifH genes. This molecular tool revealed that long-term cultivation with tillage, fertilization, pesticide treatments or specic plant cover would result in a structure of the nifH gene pool different from permanent pasture. This approach could be used to compare the diversity of nitrogen xers according to soil characteristics such as texture, structure, plant cover, etc. It should be a powerful tool for monitoring the community of nitrogen xers in soil in order to investigate seasonal changes, variations due to modication in the plant cover or effects of the input of chemical products in soils. Acknowledgements This investigation was supported by the Ecocompatibility of Solid Wastes program (grant 9674056/

DIMT/mfb) from the Agence de lEnvironnemment et de la Matrise de lEnergie (ADEME). We are grateful to Thierry Heulin and Wafa Achouak for kindly providing P. polymixa strains.

References
[1] Achouak W., Normand P., Heulin T., Comparative phylogeny of rrs and nifH genes in the Bacillaceae, Int. J. Syst. Bacteriol. 49 (1999) 961967. [2] Baldani V.L.D., Baldani J.I., Dbereiner J., Effects of Azospirillum inoculation on root infection and incorporation in wheat, Can. J. Microbiol. 29 (1983) 924929. [3] Bally R., Givaudan A., Bernillon J., Heulin T., Balandreau J., Bardin R., Numerical taxonomic study of three N2 xing yellow-pigmented bacteria related to Pseudomonas paucimobilis , Can. J. Microbiol. 36 (1990) 850855. [4] Bally R., Thomas-Bauzon D., Heulin T., Balandreau J., Richard C., De Ley J., Determination of the most frequent N2 -xing bacteria in a rice rhizosphere, Can. J. Microbiol. 29 (1983) 881887.

F. Poly et al. / Res. Microbiol. 152 (2001) 95103

[5] Bouillant M.L., Mich L., Ouedraogo O., Alexandre G., Jacoud C., Sall G., Bally R., Inhibition of striga seed germination associated with sorghum growth promotion by soil bacteria, C.R. Acad. Sci. Paris/Life Sci. 320 (1997) 159 162. [6] Braun S.T., Proctor L.M., Zani S., Mellon M.T., Zehr J.P., Molecular evidence for zooplankton-associated nitrogenxing anaerobes based on amplication of nifH gene, FEMS Microbiol. Ecol. 28 (1999) 273279. [7] Berry A., Torrey J.G., Isolation and characterization in vivo and in vitro of actinomycetous endophyte from Alnus rubre Bong, in: Gordon J.C., Wheeler C.T., Perry D.A. (Eds.), Symbiotic Nitrogen Fixation in the Management of Temperate Forests, Forests Research Laboratory, Oregon State University, Corvallis, 1979, pp. 6983. [8] Chelius M.K., Lepo J.E., Restriction fragment length polymorphism analysis of PCR-amplied nifH sequences from wetland plant rizosphere communities, Envir. Technol. 20 (1999) 883889. [9] Cournoyer B., Normand P., Characterization of a spontaneous thiostrepton resistant Frankia alni infective isolate using PCR-RFLP of nif and glnII genes, Soil Biol. Biochem. 26 (1994) 553559. [10] Fages J., Mulard D., Isolement de bactries rhizosphriques et effef de leur inoculation en pots chez Zea mays , Agronomie 8 (1988) 309312. [11] Gillis M., Tran Van V., Bardin R., Goor M., Hebbar P., Willems A., Segers P., Kerster K., Heulin T., Fernandez M.P., Polyphasic taxonomy in the genus Burkholderia leading to an emended description of the genus and the proposition of Burkholderia vietnamiensis sp. nov. for N2 -xing isolates from rice in Vietnam, Int. J. Syst. Bacteriol. 45 (1995) 274 289. [12] Hahn D., Lechevalier M.P., Fisher A., Stackebrant E., Evidence for a close phylogenic relationship between members of the genera Frankia, Geodermatophilus, and Blastococcus and emendation of the family Frankiaceae, Syst. Appl. Microbiol. 11 (1989) 236242. [13] Jacobson M.R., Brigle K.E., Bennett L.T., Setteerquist R.A., Wilson M.S., Cash V.L., Beynon J., Newton W.E., Dean D.R., Physical and genetic map of the major nif gene cluster from Azotobacter vinelandii , J. Bacteriol. 171 (1989) 10171027. [14] Khammas K.M., Ageron E., Grimont P.A.D., Kaiser P., Azospirillum irakense sp. nov., a nitrogen-xing bacterium associated with rice roots and rhizosphere soil, Res. Microbiol. 140 (1989) 679693. [15] Kreader C.A., Relief of amplication in PCR with bovine serum albumin or T4 gene 32 protein, Appl. Environ. Microbiol. 62 (1996) 11021106. [16] Newton W.E., Kish-othmer Encyclopedia of Chemical Technology, 4th Ed., John Wiley & Sons Inc., New York, 1996, pp. 172204.

103

[17] Ohkuma M., Noda S., Usami R., Horikoshi K., Kudo T., Diversity of nitrogen xation genes in the symbiotic intestinal microora of the termite Reticulitermes speratus , Appl. Environ. Microbiol. 62 (1996) 27472752. [18] Ohkuma M., Noda S., Kudo T., Phylogenetic diversity nitrogen xation genes in the symbiotic microbial community in the gut of diverse termites, Appl. Environ. Microbiol. 65 (1999) 49264934. [19] Quiviger B., Franche C., Lutfalla G., Rice D., Haselkorn R., Elmerich C., Cloning of nitrogen xation (nif ) gene cluster of Azospirillum brasilense, Biochem. 64 (1982) 495502. [20] Ranjard L., Poly F., Combrisson J., Richaume A., Nazaret S., A single procedure to recover DNA from the surface or inside aggregates and in various size fraction of soil suitable for PCR based assays of bacteria, Eur. J. Soil Biol. 34 (1998) 8997. [21] Simonet P., Grosjean M.C., Misra A.K., Nazaret S., Cournoyer B., Normand P., Frankia genus-specic characterization by polymerase chain reaction, Appl. Environ. Microbiol. 57 (1991) 32783286. [22] Sundaresan V., Ausubel F.M., Nucleotide sequence of the gene coding for the nitrogenase in iron protein from Klebseilla pneumoniae, J. Biol. Chem. 256 (1981) 28082812. [23] Torok I., Kondorosi A., Nucleotide sequence of the R. meliloti nitrogenase reductase (nifH ) gene, Nucleic Acids Res. 9 (1981) 57115723. [24] Ueda T., Suga Y., Yahiro N., Matsuguchi T., Remarkable N2 xing bacterial diversity detected in rice roots by molecular evolutionary analysis of nifH gene sequences, J. Bacteriol. 177 (1995) 14141417. [25] Wang G.C.Y., Wang Y., Frequency of formation of chimeric molecules as a consequence of PCR coamplication of 16S rRNA genes from mixed bacterial genomes, Appl. Environ. Microbiol. 63 (1997) 46454630. [26] Ward D.M., Weller R., Bateson M.M., 16S rRNA sequences reveal numerous uncultured micro-organisms in natural community, Nature 345 (1990) 6365. [27] Widmer F., Shaffer B.T., Porteous L.A., Seidler R.J., nifH gene pool complexity in soil and litter at a douglas r forest site in the Oregon Cascade Mountain Range, Appl. Environ. Microbiol. 65 (1999) 374380. [28] Young J.P.W., in: Stacy G., Burris R.H., Evans H.J. (Eds.), Biological Nitrogen Fixation, Phylogenic Classication of Nitrogen-Fixing Organisms, Chapman and Hall, New York, 1992, pp. 4386. [29] Zehr J.P., MacReynolds L.A., Use of degenerate oligonucleotides for amplication of the nifH gene from the marine Cyanobacterium Trichodesmium thiebautii , Appl. Environ. Microbiol. 55 (1989) 25262552.