J. Gen. Appl. Microbiol., 50, 129–135 (2004)

Full Paper

Phylogenetic analyses of the nitrogen-fixing genus

Cheng-Hui Xie* and Akira Yokota

Institute of Molecular and Cellular Biosciences, The University of Tokyo, Bunkyo-ku, Tokyo 113–0032, Japan

(Received August 27, 2003; Accepted May 21, 2004)

Phylogenetic analyses of the 16S rRNA gene sequence indicate that the genus Derxia forms a distinct lineage in the b-. On the NJ tree Derxia has a low bootstrap value (30.9%) with Alcaligeneceae, and on the ML tree it shows an independent cluster separated from other families. Moreover, there is below 93.4% 16S rDNA sequence similarity between genus Derxia and the genera of the b-Proteobacteria. These facts reveal that Derxia is not grouped with any known family of b-Proteobacteria and should be placed as a separate genus of b-Proteobacteria. The data on high GC content (71 mol%), the cellular fatty acid composition, and the physiologi- cal characteristics of facultative hydrogen autotrophy and nitrogen fixation are unique for Derxia. The nifH gene sequence was found in this genus and phylogenetically compared among nitrogen-fixing to indicate that Derxia is clustered with the diazotrophs of b-Proteobac- teria.

Key Words——cellular fatty acid; Derxia; nifH; phylogeny; b-Proteobacteria; 16S rDNA

Introduction thinobacterium, Alcaligenes, and a few other taxa of b- Proteobacteria; the genus was, however, placed in the The genus Derxia Jensen et al. 1960 consists of a family (a-Proteobacteria) in the latest single species, . Cells are Gram-neg- edition of “Bergey’s Manual of Systematic Bacteriol- ative, rod-shaped with rounded ends, motile by means ogy” without any comments regarding this placement of a short polar flagellum, and catalase-negative. Mo- (Garrity and Holt, 2000). The phylogenetic position of lecular nitrogen can be fixed under both aerobic condi- Derxia has not previously been studied, making the tions and decreased oxygen pressures, and can grow of Derxia a source of confusion. In the pres- as a facultative hydrogen autotroph (Derxia Jensen et ent study, we investigated the phylogenetic position of al. 1960). Derxia is usually found in tropical soils. the species of the genus Derxia based on 16S rDNA, The genus Derxia has been considered to have a nifH gene and chemotaxonomic analyses. relationship with the diazotrophic genera, , Azomonas, , and Pseudomonas, based on Materials and Methods morphological, physiological, and chemotaxonomic characteristics (Becking, 1984, 1991); based on the Bacterial strains. The two strains of D. gummosa method of rRNA cistron similarity, De Smedt et al. were obtained from the IAM Culture Collection (Japan) (1980) have also grouped the genus Derxia with Jan- and LMG Culture Collection (Belgium). The studied strains, the type strain of D. gummosa, IAM 13946T T T * Address reprint requests to: Dr. Cheng-Hui Xie, Institute of ( ATCC 15594 LMG 3977 ) and the reference Molecular and Cellular Biosciences, The University of Tokyo, strain, IAM 14990 ( LMG 3975) were isolated from 1–1–1 Hongo, Bunkyo-ku, Tokyo 113–0032, Japan. the soil in West Bengal, India, in 1960 by Jensen et al. E-mail: [email protected] (1960). Bacterial strains were grown in the medium 130 XIE and YOKOTA Vol. 50

IAM B-1 (peptone 5.0 g, beef extract 3.0 g, NaCl 3.0 g, cataway, NJ, USA). distilled water 1.0 L, pH 7.0) or in LMG Medium-10 DNA sequencing. Sequencing of the complete 16S

(glucose 10.0 g, CaCl2 ·2H2O 0.1 g, MgSO4 ·7H2O rDNA gene was performed on the coding and comple- 0.1 g, K2HPO4 0.9 g, KH2PO4 0.1 g, CaCO3 5g, mentary strands by using four primer pairs: 520F: 5 - FeSO4 ·7H2O 10 mg, Na2MoO4 ·2H2O 5 mg, distilled CAGCAGCCGCGGTAATAC-3 (520–537)/1100R: 5 - water 1.0 L, pH 7.3) at 25°C. GGGTTGCGCTCGTTTG-3 (1100–1114); 926F: 5-A- Biochemical analyses. API 20NE and 50CHL (bio- AACTCAAAGGAATTGACGG-3 (926–945)/1510R: 5- Mérieux, S. A., Marcy-l’Etoile, France) were used to GGCTACCTTGTTACGTA-3 (1510–1527). 8F: 5-AG- determine the physiological and biochemical character- AGTTTGATCCTGGCTCAG-3 (8–27)/700R: 5-TCTA- istics. The API strips were incubated for 2 days at 25°C. CGCATTTCACC-3 (700–714). Sequencing reactions Chemotaxonomic investigations. Respiratory qui- were performed using a BigDye Terminator Cycle Se- nones were extracted with chloroform/methanol (2 : 1, quencing Ready Reaction Kit (Applied Biosystems, v/v), and were purified by TLC on silica-gel F254 plates Foster City, CA, USA) according to the manufacturer’s (Merck, Darmstadt, Germany) with hexane/diethyl protocol. Sequences were obtained with the ABI ether (85 : 15, v/v) being used as the solvent. The PRISMTM 310 Genetic Analyzer (Applied Biosystems). ubiquinone fraction was extracted with acetone, dried Phylogenetic analyses. The DNA sequences of D. under a stream of nitrogen, and subsequently ana- gummosa were compared with the sequences ob- lyzed by high-performance liquid chromatography tained from the DNA database. The sequences were (Model LC-10A apparatus, Shimadzu, Kyoto, Japan). aligned with the CLUSTAL W software package Fatty acid methyl esters were prepared from cells (Thompson et al., 1994), and evolutionary distances grown on TSA (Trypticase soy agar, Becton Dickinson and Knuc values (Kimura, 1980) were generated. Align- and Co., Sparks, MD, USA) for 48 h. The fatty acid ment gaps and ambiguous bases were not taken into methylesters (FAMES) were obtained from the cells by consideration and comparison. Phylogenetic trees saponification, methylation, and extraction according were constructed using the neighbor-joining method to the manual for the MIDI System (Microbial ID, Inc., (Saitou and Nei, 1987) and the maximum likelihood Newark, MD, USA). Analysis by gas chromatography method in PHYLIP Package (Felsenstein, 1989). The was controlled by MIS software. Following the stan- topology of the phylogenetic tree was evaluated by the dard protocol of the MIDI/Hewlett Packard Microbial bootstrap resampling method of Felsenstein with 1,000 Identification System, fatty acid methyl ester extracts replicates. The similarity values were calculated using were analyzed in a Hewlett Packard (model HP PAUP 4.068 PPC (Swofford, 1998). 6890A) GC equipped with a flame-ionization detector, Nucleotide sequence accession numbers. The an automatic sampler, and a computer. EMBL/GenBank accession numbers for the 16S rDNA PCR and sequencing. Genomic DNA was pre- sequences in this study are: D. gummosa (IAM pared from bacterial cells suspended in TE buffer by 13946T) AB089482 and D. gummosa (IAM 14990) heating at 95°C for 5 min, followed by cooling and cen- AB089481. The accession number of the nifH se- trifugation to collect the lysate. An approximately quences are: D. gummosa (IAM 13946T) AB089483 1,500-bp fragment of the 16S rDNA was amplified and D. gummosa (IAM 14990) AB089485. from the extracted DNA by using eubacterial universal primers specific to the 16S rDNA gene: 8F: 5 - Results and Discussion AGAGTTTGATCCTGGCTCAG-3 [8–27, the E. coli numbering system of Brosius et al. (1978)] and 1510R: The phylogenetic trees were constructed by the 5-GGCTACCTTGTTACGTA-3 (1510–1527). The 360- comparison of the 16S rDNA sequence of D. gum- base fragment of the nifH gene was amplified from the mosa and related genera from a database using the extracted DNA with the primers following forward and neighbor-joining (NJ) algorithm and the maximum like- backward: TGCGAYCCSAARGCBGACTC and ATS- lihood (ML) method, which are shown in Figs. 1 and 2, GCCATCATYTCRCCGGA (YC/T; SG/C; RA/G; respectively. They revealed that the genus Derxia be- BC/G/T) (Poly et al., 2001). The amplified fragments longs to the b-Proteobacteria. On the NJ phylogenetic were purified by GFXTM PCR DNA and a Gel Band Pu- tree Derxia has a low bootstrap value (30.9%) with the rification Kit (Amersham Pharmacia Biotech, Inc., Pis- Alcaligeneceae group, which strongly supports that is 2004 Phylogenetic analyses of the nitrogen-fixing genus Derxia 131

Fig. 1. Neighbor-joining tree showing phylogenetic relationships among members of b-Proteobacteria and the genus Derxia based on 1,138 nucleotide positions of the 16S rDNA sequence after excluding the positions with gaps and ambiguous bases. ATCC 9093T (M59060) was used as the outgroup. Only bootstrap values over 80% are shown. not grouped with any known family such as the Burk- b-Proteobacteria, and it is not at all related to other holderiaceae, Alcaligeneceae, Oxalobacteraceae, or free-living nitrogen-fixing bacteria such as Beijerinkia, Comamonadaceae. The ML phylogenetic analyses Azospirillum, Azomonas, or Pseudomonas. confirms that Derxia is a distinct lineage and not close We also analyzed the chemotaxonomic characteris- to any other taxa. Moreover, there is less than 93.4% tics of D. gummosa (Table 1). The major cellular fatty 16S rDNA sequence similarity between the genus acid content was 18 : 1 w7c, 16 : 1 w7c, and 16 : 0, while Derxia and the genera of b-Proteobacteria. So far, 3OH-12 : 0, 3OH-14 : 0, and 2OH-14 : 0 were the major Derxia shows no close to phylogenetic relationship to hydroxy fatty acids. The high GC content of DNA 132 XIE and YOKOTA Vol. 50

Fig. 2. Maximum likelihood tree showing phylogenetic relationships among members of b-Proteobacteria and the genus Derxia based on 1,138 nucleotide positions of the 16S rDNA sequence after excluding the positions with gaps and ambiguous bases. Only bootstrap values over 80% are shown.

(71 mol%), the facultative hydrogen autotrophy, and ase, shows a high degree of conservation of structure, the catalase-negative characteristics of this genus are function, and amino acid sequences across wide phy- different from those of other bacteria of b-Proteobacte- logenetic ranges. It is known that Mo-nitrogenase con- ria. The major quinone systems contained ubiquinone sists of two components, component I (also called dini-

Q-8, the same as for the other members of b-Pro- trogenase or Fe-Mo protein), an a2b2 tetramer en- teobacteria (Table 2). coded by the nifD and nifK genes, and component II Nitrogen fixation, the biological conversion of atmo- (dinitrogenase reductase or Fe protein), a homodimer spheric nitrogen to ammonia, is known to be wide- encoded by the nifH gene. The nifH is being closely spread in both bacteria and archaea (Young, 1992). examined as the part of nitrogen-fixation gene group The enzyme responsible for nitrogen fixation, nitrogen- (nifHDK) showing strong conservation and being avail- 2004 Phylogenetic analyses of the nitrogen-fixing genus Derxia 133 able for use as a molecular marker in phylogenetic phyletic group with the genera Burkholderia and analysis for the diazotrophs. The nifH phylogenetic Herbaspirillum of b-Proteobacteria and separates tree had been established, and is largely consistent from those of g-Proteobacteria (Azomonas and Azoto- with the 16S rRNA gene phylogeny, except for discrep- bacter) and a-Proteobacteria (Beijerinkia and Azospir- ancies with a few taxa (Moulin et al., 2001; Rosado et illum). This result is consistent with the 16S rRNA al., 1998; Young, 1992). The determined nifH se- gene phylogeny, indicating that Derxia is not included quence (360 bases) of the two strains of D. gummosa within other diazotrophic genera. (IAM 13946T and IAM 14990) showed only one nu- As a conclusion: phylogenetic analyses of 16S rDNA cleotide difference (C-T) at position 300 in the number- sequence strongly support that the genus Derxia is a ing system of the Azotobacter vinelandii M20568 se- member of b-Proteobacteria and forms a separate quence (Jacobson et al., 1989). A phylogenetic tree cluster not included within any other family, and in the was constructed for these strains for comparison with nifH phylogenetic tree it forms a monophyletic group the other sequences of nifH from the database (shown with the genera Burkholderia and Herbaspirillum. The in Fig. 3). The highest degree similarity for the nifH se- chemotaxonomic data also support that Derxia is a quence of Derxia is 92% against Burkholderia fungo- distinct lineage within b-Proteobacteria. rum (NZ_AAAC01000309). Derxia forms a mono-

Table 1. Cellular fatty acid profile of strains of The emended description of Derxia gummosa (Jensen AL Derxia gummosa. et al., 1960, 193 ), based on the data compiled by Becking (1984) and this study Fatty acid IAM 13946T IAM 14990 Cells are rod-shaped, 1.0–1.2 mm in width, catalase- 12 : 0 5.8 4.3 negative, oxidase-positive, and motile by means of a 12 : 0 3OH 5.3 5.0 short polar flagellum; molecular nitrogen can be fixed 14 : 0 1.1 0.6 under both aerobic conditions and decreased oxygen 14 : 0 2OH 4.0 3.1 pressures; it can grow as a facultative hydrogen au- 14 : 0 3OH 3.4 2.7 totroph; and it can grow on glucose and sugars but 16:1 w7c 29.3 35.2 can not grow on malic acid used as a carbon source. 16:0 16.3 22.3 18 : 0 0.6 0.5 The optimum temperature for growth is 25–35°C, and 18:1 w7c 32.9 24.8 it can grow at 40°C. Growth occurs between pH 5.5– 9.0; no growth at pH 4.4. Utilizes methane or methanol Fatty acid composition below 0.5% is not shown. as the sole carbon source. Growth on glucose, fruc-

Table 2. Differential characteristics among Derxia and the genera in b-Proteobacteria.

Quinoneb Major 3OHb Major 2OHb GC contentb Hydrogena Genus Catalasea Oxidasea N -fixation system fatty acids fatty acids (mol%) autotrophy 2

Derxia Q-8 12 : 0, 14 : 0 14 : 0 69–72 Alcaligenes Q-8 14 : 0 12 : 0 56–70 Bordetella Q-8 14 : 0 12 : 0, 14 : 0 66–70 ND / Brackiella Q-8 16 : 0 3 — ND Burkholderia Q-8 14 : 0, 16 : 0 16 : 0 64–69 / Herbaspirillum Q-8 12 : 0 14 : 0, 12 : 0 60–65 / Janthinobacterium Q-8 10 : 0 12 : 0 64–65 Oxalobacter Q-8 12 : 0 14:0 50 ND ND Pelistega Q-8 16 : 0 — 42–43 Ralstonia Q-8 14 : 0 16 : 0 63–69 Taylorella Q-8 16 : 0 — 36–37

a From Becking (1984); b from Coenye et al. (1999, 2001), Goris et al. (2001), Jendrossek (2001), Lincoln et al. (1999), Willems et al. (1989), Urakami et al. (1994), Vandamme et al. (1996). 134 XIE and YOKOTA Vol. 50

Fig. 3. Phylogenetic relationships among nifH gene products (309 bases) of the diazotrophs of Proteobacteria by the neighbor-joining method. Cyanobacteria were used as the outgroup. tose, ethanol, glycerol, mannitol and sorbitol is good to 15594T). excellent. No growth or only a trace of growth occurs on lactose, galactose, maltose, sucrose, formate, ac- References etate, propionate, pyruvate, succinate, malate, fu- marate, dulcitol and starch. Nitrate is not reduced to ni- Becking, J. H. (1984) Genus Derxia Jensen, Petersen, De and AL trite or N2, and indole is not produced from tryptophan. Bhattacharya 1960, 193 . In Bergey’s Manual of System- It is usually found in tropical soils. The major cellular atic Bacteriology, Vol. 1, ed. by Krieg, N. R. and Holt, J. G., fatty acids are 18 : 1 w7c, 16 : 1 w7c, 16 : 0, and the Williams & Wilkins, Baltimore, MD, USA, pp. 321–325. major hydroxy fatty acids are 3OH-12 : 0, 3OH-14 : 0, Becking, J. H. (1991) The genus Derxia. In The Prokaryotes, Vol. 3, ed by Balows, A., Trüper, H. G., Dworkin, M., and 2OH-14 : 0. The GC content of the DNA is Harder, W., and Schlegel, H. G., Springer-Verlag, Berlin, 71 mol%. The quinone system is ubiquinone-8. Based pp. 2605–2611. on the 16S rDNA sequence analysis, Derxia belongs Brosius, J., Palmer, J. L., Kennedy, J. P., and Noller, H. F. to b-Proteobacteria. (1978) Complete nucleotide sequence of a 16S ribosomal The type strain is IAM 13946T (LMG 3977TATCC RNA gene from Escherichia coli. Proc. Natl. Acad. Sci. 2004 Phylogenetic analyses of the nitrogen-fixing genus Derxia 135

USA, 75, 4801–4805. cleic acid from micro-organisms. J. Mol. Biol., 3, 208–218. Coenye, T., Falsen, E., Vancanneyt, M., Hoste, B., Govan, J. R. Moulin, L., Munive, A., Dreyfus, B., and Boivinmasson, C. W., Kersters, K., and Vandamme, P. (1999) Classification of (2001) Nodulation of legumes by members of the b-sub- Alcaligenes faecalis-like isolates from the environment and class of Proteobacteria. Nature, 411, 948–950. human clinical samples as Ralstonia gilardii sp. nov. Int. J. Poly, F., Monrozier, L. J., and Bally, R. (2001) Improvement in Syst. Bacteriol., 49, 405–413. the RFLP procedure for studying the diversity of nifH genes Coenye, T., Laevens, S., Willems, A., Ohlén, M., Hannant, W., in communities of nitrogen fixers in soil. Res. Microbiol., Govan, J. R. W., Gillis, M., Falsen, E., and Vandamme, P. 152, 95–103. (2001) Burkholderia fungorum sp. nov. and Burkholderia Rosado, A. S., Duarte, G. F., Seldin, L., and Van Elsas, L. D. caledonica sp. nov., two new species isolated from the en- (1998) Genetic diversity of nifH gene sequences in Paeni- vironment, animals and human clinical samples. Int. J. bacillus azotofixans strains and soil samples analyzed by Syst. Bacteriol., 51, 1099–1107. denaturing gradient gel electrophoresis of PCR-amplified De Smedt, J., Bauwens, M., Tytgat, R., and De Ley, J. (1980) gene fragments. Appl. Environ. Microbiol., 64, 2770–2779. Intra- and intergeneric similarities of ribosomal ribonucleic Saitou, N. and Nei, M. (1987) The neighbor-joining method: A acid cistrons of free-living, nitrogen-fixing bacteria. Int. J. new method for reconstructing phylogenetic trees. Mol. Syst. Bacteriol., 30, 106–122. Biol. Evol., 4, 406–425. Felsenstein, J. (1989) PHYLIP—Phylogeny Inference Package Swofford, D. L. (1998) PAUP* Phylogenetic analysis using pari- (Version 3.2). Cladistics, 5, 164–166. mony (* and other methods), version 4, Sinauer Associ- Garrity, G. M. and Holt, J. G. (2000) The road map to the man- ates, Sunderland, MA. ual. In Bergey’s Manual of Systematic Bacteriology, 2nd Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994) ed., Vol. 1, ed. by Boone, D. R. and Castenholz, R. W., CLUSTAL W: Improving the sensitivity of progressive multi- Springer-Verlag, Heidelberg, New York, Berlin, pp. 119– ple sequence alignment through sequence weighting, posi- 166. tion specific gap penalties and weight matrix choice. Nu- Goris, J., De Vos, P., Coenye, T., Hoste, B., Janssens, D., Brim, cleic Acids Res., 22, 4673–4680. H., Diels, L., Mergeay, M., Kersters, K., and Vandamme, P. Urakami, T., Ito-Yoshida, C., Araki, H., Kijima, T., Suzuki, K., (2001) Classification of metal-resistant bacteria from indus- and Komagata, K. (1994) Transfer of Pseudomonas plan- trial biotopes as Ralstonia campinensis sp. nov., Ralstonia tarii and Pseudomonas glumae to Burkholderia as Burk- metallidurans sp. nov. and Ralstonia basilensis Steinle et holderia spp. and description of Burkholderia vandii sp. al. 1998 emend. Int. J. Syst. Evol. Microbiol., 51, 1773– nov. Int. J. Syst. Bacteriol., 44, 235–245. 1782. Vandamme, P., Heyndrickx, M., Vancanneyt, M., Hoste, B., De Jacobson, M. R., Brigle, K. E., Bennett, L. T., Setterquist, R. A., Vos, P., Falsen, E., Kersters, K., and Hinz, K. H. (1996) Wilson, M. S., Cash, V. L., Beynon, J., Newton, W. E., and Bordetella trematum sp. nov., isolated from wounds ear in- Dean, D. R. (1989) Physical and genetic map of the major fections in humans, and reassessment of Alcaligenes deni- nif gene cluster from Azotobacter vinelandii. J. Bacteriol., trificans Rüger and Tan 1983. Int. J. Syst. Bacteriol., 46, 171, 1017–1027. 849–858. Jendrossek, D. (2001) Transfer of [Pseudomonas] lemoignei, a Willems, A., Busse, J., Goor, M., Pot, B., Falsen, E., Jantzen, Gram-negative rod with restricted catabolic capacity, to E., Hose, B., Gillis, M., Kersters, K., Auling, G., and De Ley, Paucimonas gen. nov. with one species, Paucimonas J. (1989) Hydrogenophaga, a new genus of hydrogen-oxi- lemoignei comb. nov. Int. J. Syst. Bacteriol., 51, 905–908. dizing bacteria that includes Hydrogenophaga flava comb. Jensen, H. L., Peterson, E. J., De, P. K., and Bhattacharya, R. nov. (formerly Pseudomonas flaava), Hydrogenophaga (1960) A new nitrogen fixing bacterium: Derxia gummosa palleronii (formerly Pseudomonas palleronii), Hy- nov. gen. nov. spec. Arch. Mikrobiol., 36, 182–195. drogenophaga pseudoflava (formerly Pseudomonas pseu- Kimura, M. (1980) A simple method for estimating evolutionary doflava and “Pseudomonas carboxydoflava”), and Hy- rates of base substitutions through comparative studies of drogenophaga taeniospiralis (formerly Pseudomonas tae- nucleotide sequences. J. Mol. Evol., 16, 111–120. niospiralis). Int. J. Syst. Bacteriol., 39, 319–333. Lincoln, S. L., Fermor, T. R., and Tindall, B. J. (1999) Janthi- Young, J. P. W. (1992) Phylogenetic classification of nitrogen- nobacterium agaricidamnosum sp. nov., a soft rot pathogen fixing organisms. In Biological Nitrogen Fixation, ed. by of Agaricus bisporus. Int. J. Syst. Bacteriol., 49, 1577– Stacey, G., Burris, R. H., and Evans, H. J., Chapman and 1589. Hall, New York, USA, pp. 43–86. Marmur, J. (1961) A procedure of the isolation of deoxyribonu-