Proposal of Pseudorhodobacter Ferrugineus Gen. Nov., Comb. Nov

J. Gen. Appl. Microbiol., 48, 309–319 (2002)

Full Paper

Proposal of Pseudorhodobacter ferrugineus gen. nov., comb. nov., for a non-photosynthetic marine bacterium, Agrobacterium ferrugineum, related to the genus Rhodobacter

Yoshihito Uchino,* Tohru Hamada,1 and Akira Yokota

Institute of Molecular and Cellular Biosciences, The University of Tokyo, Bunkyo-ku, Tokyo 113–0032, Japan 1 Kamaishi Laboratories, Marine Biotechnology Institute, Kamaishi 026–0001, Japan

(Received August 2, 2001; Accepted October 30, 2002)

The marine gram-negative non-photosynthetic bacterium, Agrobacterium ferrugineum IAM 12616T forms one cluster with the species of the photosynthetic genus Rhodobacter in phylogenetic trees based on molecules of 16S rRNA, 23S rRNA and DNA gyrases. Agrobacterium ferrugineum and Rhodobacter species are similar in that growth occurs without NaCl in the culture medium (optimal NaCl concentration for growth of P. ferrugineus is 1%) and their major hydroxy fatty acid compositions are 3-hydroxy decanoic acids (3-OH 10:0) and 3-hydroxy tetradecanoic acids (3-OH 14:1). However, A. ferrugineum differs from Rhodobacter species in GC content (58 mol% in A. ferrugineum versus 64–73 mol% in Rhodobacter species), in having an insertion in its 16S rRNA gene sequence, and in lacking photosynthetic abilities, bacteriochlorophyll a and intracytoplasmic membrane systems. Furthermore, experiments using PCR and Southern hybridization show that A. ferrugineum does not have puhA gene and puf genes localized near the opposite ends of the photosynthesis gene cluster of Rhodobacter capsulatus. It suggests that A. ferrugineum may not have any genes for photosynthesis. We propose the transfer of A. ferrugineum IAM 12616T to the genus Pseudorhodobacter gen. nov. as Pseudorhodobacter ferrugineus comb. nov. Although Pseudorhodobacter ferrugineus disturbs the phylogenetic monophyly of the genus Rhodobacter, this taxonomic proposal seems adequate until it has been clari- ﬁed whether P. ferrugineus possesses an incomplete photosynthetic apparatus.

Key Words——Agrobacterium ferrugineum; bacterial taxonomy; DNA gyrase sequence; Pseudorhodobacter ferrugineus gen. nov., comb. nov.; Rhodobacter; 16S rRNA sequence; 23S rRNA sequence

Introduction photosynthetic bacteria, and that many photosynthetic genera were heterogeneous (Stackebrandt et al., In the past phylogenetic analyses based on various 1996). To date, the polyphyletic photosynthetic genera molecules indicated that the photosynthetic bacteria of have been divided and reorganized into several gen- the class Proteobacteria were intermingled with non- era, and consequently, almost all the present phototrophic genera are monophyletic.

* Address reprint requests to: Dr. Yoshihito Uchino, NITE Bio- Rhodobacter is one of the photosynthetic genera logical Resource Center, National Institute of Technology and which was newly proposed by Imhoff et al. (1984), and Evaluation, 2–5–8 Kazusakamatari, Kisarazu, Chiba 292–0812, later reorganized by Hiraishi and Ueda (1994). Hiraishi Japan. E-mail: [email protected] and Ueda (1994) divided the genus Rhodobacter; the 310 UCHINO, HAMADA, and YOKOTA Vol. 48 species whose habitats were seawater were trans- medium containing 1% polypeptone, 0.2% yeast ex- ferred to the new genus Rhodovulum, and, as a result, tract and 0.1% MgSO4 (PY medium). Rhodobacter the genus Rhodobacter involved only species whose veldkampii DSM 11550T was grown phototrophically at habitats were freshwater. At present Rhodobacter is a 27°C in SA medium (Kawasaki et al., 1992) to which a monophyletic genus in the Proteobacteria a-3 sub- sulfide solution had been added, and Rhodovulum group. strictum JCM 9220T was grown phototrophically at As described in previous reports, the marine non- 27°C in MMYS-II (Hiraishi and Ueda, 1995). Tests to photosynthetic bacterium Agrobacterium ferrugineum determine NaCl ranges for growth were performed IAM 12616T formed one cluster with the species of the with PY medium containing different concentrations of genus Rhodobacter in phylogenetic trees based on nu- NaCl (0–20%). cleotide sequences of 16S rRNA (Uchino et al., 1997, 23S rDNA sequencing. Fragments including the 1998). Agrobacterium ferrugineum did not grow pho- total 23S rRNA genes of A. ferrugineum IAM 12616T, totrophically under anoxic conditions, and the bacteri- Rba. azotoformans IAM 14814T, Rba. blasticus DSM ochlorophyll a (BChl a), intracytoplasmic membrane 2131T and Rba. veldkampii DSM 11550T were ampli- systems and genes for photosynthesis, e.g., puf L and fied by PCR using conserved primers 16SF01 (5-AC- puf M coding subunits L and M of the photosynthetic CGCCCGTCACACC-3) and 5SR01 (5-SYGTTCG- reaction center could not be detected. Hence, the GRAWGGGA-3) (Ludwig et al., 1992). The PCR tem- monophyly of the phototrophic members of Rhodobac- perature control was done by using the following ther- ter was disturbed by non-photosynthetic Agrobac- mal profile: after initial denaturation at 94°C for 2 min, terium ferrugineum, and the genus Rhodobacter be- a total of 35 cycles of amplification were performed came a paraphyletic group. In a previous report, we with template DNA denaturation at 94°C for 1.5 min, deferred an official nomenclatural proposal for A. fer- primer annealing at 50°C for 1 min and primer exten- rugineum until more phenotypic information was avail- sion at 72°C for 2.5 min, and final extension at 72°C able. for 10 min. The PCR products were purified using Ul- In this study, we tried to detect genes for photosyn- trafree-MC centrifugal filter units (Millipore Corp., Bed- thesis in A. ferrugineum using PCR and Southern ford, MA, USA). Sequencing was carried out using an hybridization methods. The target genes were puf L ABI PRISMTM Big-Dye terminator Cycle Sequencing and puf M genes, and puhA gene coding subunit H Ready Reaction Kit (Perkin-Elmer Co., Foster City, of the reaction center. If these genes can be found in CA, USA) and a model ABI 310 Genetic Analyzer A. ferrugineum, it may suggest that A. ferrugineum is (Perkin-Elmer Co.) according to the manufacturer’s a variant of the photosynthetic bacterium, and we instructions. The sequencing primers used are listed can propose A. ferrugineum as species of genus in Table 1. The primers in this study included degener- Rhodobacter. ated primers (the mixture of some kinds of oligonu- In this study, we performed the multigene phyloge- cleotides). netic analyses based on nucleotide sequences of 23S gyrB sequencing. The protocol for determining the rRNA and amino acid sequences of subunit B of DNA gyrB sequences was almost the same as that gyrase adding the information based on the 16S rRNA described by Yamamoto and Harayama (1995). The to clarify the phylogeny of the organisms. gyrB fragments of A. ferrugineum IAM 12616T, P. aminophilus IAM 14245T, Rba. capsulatus ATCC T T Materials and Methods 11166 , Rba. sphaeroides ATCC 11167 , Rdv. strictum JCM 9220T, and Rsb. litoralis IFO 15278T were ampli- Bacterial culture conditions. Agrobacterium ferru- fied by PCR using primers UP-1E (Yamamoto et al., gineum IAM 12616T and Roseobacter litoralis IFO 1999) and UP-2r (Yamamoto and Harayama, 1995). 15278T were cultured aerobically at 27°C in Difco Ma- The gyrB fragments of Rba. azotoformans IAM rine broth 2216 medium. Paracoccus aminophilus IAM 14814T, Rba. blasticus DSM 2131T and Rba. veld- 14245T, Rhodobacter azotoformans IAM 14814T, kampii DSM 11550T were amplified by PCR using Rhodobacter blasticus DSM 2131T, Rhodobacter cap- primers gy21A (5-CAGGAAACAGCTATGACCAAR- sulatus ATCC 11166T and Rhodobacter sphaeroides MGICCNGSIATGTAYATHGG-3) and gyQTKr (5-AC- ATCC 11167T were cultured aerobically at 27°C in a SAGCTTGTCCTTGGTYTG-3). The primer gy21A 2002 Pseudorhodobacter ferrugineus gen. nov., comb. nov. 311

Table 1. Primer sequences of 23S ribosomal RNA gene (Perkin-Elmer Co.) and a model ABI 377 DNA Se- for PCR and sequence reaction. quencer (Perkin-Elmer Co.) according to the manufacturer’s instructions. The sequencing primers used 16SF01 5 -ACCGCCCGTCACACC OH-3 were UP-1E, UP-2r, gyQTKr and M13R (5-CAGGAA- 23SF01 5-CCGAATGGGGAAACCC OH-3 ACAGCTATGACC-3) –21M13 (5-TGTAAAACGACG- 23SF02 5-AGTAGTGGCGAGCGAA OH-3 23SF03 5-AGTACCGTGAGGGAAAG OH-3 GCCAGT-3 ) universal primer. 23SF04 5-AGCTGGTTCTCCGCGAAA OH-3 Phylogenetic analysis. The phylogenetic analysis 23SF05 5-GCGTAACAGCTCACT OH-3 based on nucleotide sequences of 16S rRNA and 23S 23SF06 5-GTAGCGAAATTCCTTGTCG OH-3 rRNA and amino acid sequences of DNA gyrase B 23SF07 5-CCTCGATGTCGGCTC OH-3 subunit were performed. The 16S rDNA sequences of 23SR01 5-CTTTCCCTCACGGTACT OH-3 the following 33 species were newly obtained from the 23SR02 5 -TTTCGCGGAGAACCAGCT OH-3 DNA data bank, and used with the previous data 23SR03 5-TCAGGGTTGTTTCCCT OH-3 (Uchino et al., 1998). 23SR04 5-CCACCTGTGTCGGTTT OH-3 Antarctobacter heliothermus (accession number: 23SR05 5-CTTAGATGCCTTCAGC OH-3 23SR06 5-ACTWAGATGTTTCAGTTC OH-3 Y11552), Brucella melitensis (AF220147), Caulobacter 23SR07 5-CCTTCTCCCGAAGTTACGG OH-3 crescentus (NC_002696), Hyphomonas adhaerens 5SR01 5-SYGTTCGGRAWGGGA OH-3 (AF082790), Hyphomonas hirschiana (AF082794), Hy- phomonas johnsonii (AF082791), Hyphomonas neptu- The abbreviation R meant adenine or guanine; S, cytosine or nium (AF082798), Hyphomonas oceanitis (AF082797), guanine; W, adenine or thymine; and Y, cytosine or thymine, re- Hyphomonas polymorpha (AF082796), Hyphomonas spectively. rosenbergii (AF082795), “Marinosulfonomonas methyl- otropha” (U62894), Mesorhizobium loti (AP003001), was designed from the phylogenetically conserved N- Methylarcula marina (AF030436), Methylarcula terri- terminal region (KRPGMYIG and KRPAMYIG) of the cola (AF030437), Octadecabacter antarcticus amino acid sequences of the subunit B protein of DNA (U14583), Octadecabacter arcticus (U73725), Para- gyrases from Escherichia coli K-12 (P06982), Bacillus coccus alkenifer (Y13827), Paracoccus carotinifaciens subtilis 168 (P37525), Haloferax alicantei Aa 2.2 (AB006899), Paracoccus marcusii (Y12703), Paracoc- (P21558), Mycoplasma pneumoniae SGC3 (P22447), cus pantotrophus (Y16933), Rhodovulum iodosum Synechocystis sp. PCC 6803 (P77966) and Thermo- (Y15011), Rhodovulum robiginosum (Y15012), Ro- toga maritima MSB8 (JC4960). The primer gyQTKr seinatronobacter thiooxidans (AF249749), Roseivivax was designed from the phylogenetically conserved re- halodurans (D85829), Roseivivax halotolerans gion (QTKDKLV) of the amino acid sequences of the (D85831), Roseobacter gallaeciensis (Y13244), subunit B protein of DNA gyrases of 40 strains in Pro- Roseovarius tolerans (Y11551), Rubrimonas cliftonen- teobacteria a, b and g subclasses obtained from the sis (D85834), Silicibacter lacuscaerulensis (U77644), ICB database of the Marine Biotechnology Institute Staleya guttiformis (Y16427), Sulfitobacter brevis http://seaweed.mbio.co.jp/icb/index.php. The PCR (Y16425), Sulfitobacter mediterraneus (Y17387), and temperature control was done by using the following Zymomonas mobilis (AF088897). thermal profile: after initial denaturation at 94°C for 10 The 23S rDNA sequences of the following 9 species min, a total of 35 cycles of amplification were per- were obtained from the DNA databank and used with formed with template DNA denaturation at 94°C for the data determined in this study: Brucella melitensis 1 min, primer annealing at 59°C for 1 min and primer (AF220147), Caulobacter crescentus (NC_002696), extension at 72°C for 2 min, and final extension at Escherichia coli (NC_000913), Mesorhizobium loti 72°C for 10 min. The amplified products were purified (AP003001), Paracoccus denitrificans (X87287), by gel electrophoresis on 1% (w/v) low-melting-point Rhodobacter capsulatus (X06485), Rhodobacter agarose. The purified fragments were recovered from sphaeroides (X53853), Sinorhizobium meliloti the agarose by a QIAEX II gel extraction kit (Qiagen, (NC_003047), and Zymomonas mobilis (AF088897). Hilden, Germany) and used for sequencing. Sequenc- The DNA gyrase B subunit sequences of the following was carried out using an ABI PRISMTM Big-Dye ing 6 species were obtained from the DNA databank terminator Cycle Sequencing Ready Reaction Kit and used with the data determined in this study: Bru- 312 UCHINO, HAMADA, and YOKOTA Vol. 48 cella melitensis (NC_003317), Caulobacter crescentus Table 2. Primer sequences of puf L, puf M, and puhA genes (NC_002696), Escherichia coli (NC_000913), Mesorhi- for PCR and sequence reaction. zobium loti (AP003005), Sinorhizobium meliloti (NC_003047), and Zymomonas mobilis (AF088897). pufL1F 5 -TTCGACTTCTGGGT OH-3 pufL2F 5-TATGTCGGCTTCTTCGG OH-3 The sequences were aligned using Clustal W ver- pufL3R 5-CCGATCGAATAGCC OH-3 sion 1.8 (Macintosh) (Thompson et al., 1994). Finally, pufL4R 5-CCACCAGTTCCACCA OH-3 the alignment was refined manually, and the positions pufM1F 5-ATGGCTGAGTATCA OH-3 with gaps and undetermined and ambiguous se- pufM2F 5-CAGATCGGGCCGATCTA OH-3 quences were removed for subsequent phylogenetic pufM3R 5-AAGCCCATCGTCCAGCGCCAGAA OH-3 analyses. The phylogenetic analyses were performed pufM4R 5-CCAGACGTACCAGTTGTC OH-3 by the neighbor-joining (NJ) method and the maxi- pufL1R 5-ACCCAGAAGTCGAA OH-3 mum-likelihood (ML) method (Felsenstein, 1981; pufL3F 5 -GGCTATTCGATCGG OH-3 pufM2R 5-TAGATCGGCCCGATCTG OH-3 Saitou and Nei, 1987). The NJ analysis was performed puhA1F 5-GAGRAYMKNCGCGARGGCTAYCC-3 using Clustal W version 1.8 (Mac) and PAUP* version puhA1R 5-TAGSCNSANAYCTTGTCYTCTT-3 4.0b8 (Mac) (Swofford, 2000). The distance was calculated by the method of Jukes-Cantor (1969) and The abbreviation K meant guanine or thymine; M, adenine or Kimura 2-parameter (Kimura, 1980). The statistical cytosine; R, adenine or guanine; S, cytosine or guanine; W, significance of the tree branches was assessed by adenine or thymine; and Y, cytosine or thymine, respectively. bootstrap analysis, involving the construction of 1,000 trees from resampled data (Felsenstein, 1985). ML genes (e.g., puf L, puf M and puhA) were detected by analysis was performed by MOLPHY version 2.3b amplifying their partial DNA fragments by PCR using (UNIX) (Adachi and Hasegawa, 1996). The local boot- chromosomal DNA, TaKaRa Ex Taq (TaKaRa Shuzo, strap probabilities (LBPs) were calculated as the sta- Kyoto, Japan), and the primer sets (Table 2). The PCR tistical confidences of the tree branches (Hasegawa temperature control was done by using the following and Kishino, 1994). thermal profile: after initial denaturation at 98°C for Southern hybridization. Southern hybridization was 2min, a total of 35 cycles of amplification were per- performed for the detection of photosynthetic genes formed with template DNA denaturation at 98°C for (e.g., puf L) in A. ferrugineum IAM 12616T. DIG-labeled 45 s, primer annealing at 45, 50 or 60°C for 45 s and hybridization probe was generated with the PCR DIG primer extension at 72°C for 1 min, and final extension Probe Synthesis Kit (Roche Molecular Biochemicals, at 72°C for 10 min. Mannheim, Germany) from Rba. sphaeroides ATCC DNA base composition. Chromosomal DNA was 11167T chromosomal DNA template by using primers extracted and purified by the method of Marmur pufL1F (TTCGACTTCTGGGT) and pufL3R (CCGATC- (1961). The DNA was hydrolyzed with P1 nuclease GAATAGCC). Total genomic DNA from A. ferrugineum and the nucleotides were dephosphorylated with alka- and Rba. sphaeroides was extracted by using the line phosphatase. The resulting deoxyribonucleosides lysozyme, proteinase K, phenol/chloroform/isoamyl al- were analyzed by HPLC (Mesbah and Whitman, cohol, SDS, and RNase A (Marmur, 1961). The ge- 1989). nomic DNA was digested with the restriction enzyme EcoR I. The digested DNA and l-EcoT14 I DNA stan- Results dard marker were electrophoresed on 1% agarose gel and transferred onto positively charged nylon mem- 16S rDNA phylogenetic analysis branes (Roche Molecular Biochemicals, Mannheim, Sequences were aligned, and the positions with Germany). Hybridization and detection of the probe gaps and the undetermined and ambiguous se- were performed using a DIG Luminescent Detection quences were removed, and finally, 1,063 sites were Kit (Roche Molecular Biochemicals) according to its used for this phylogenetic analyses. Escherichia coli manufacturer’s protocol. K12 was included as an outgroup. Figure 1 shows the Detection of genes related to photosynthesis by NJ tree from distances estimated by a Kimura 2-param- PCR. Chromosomal DNA was extracted and purified eter model. Agrobacterium ferrugineum formed a by the method of Marmur (1961). The photosynthetic cluster with the species of the genus Rhodobacter, 2002 Pseudorhodobacter ferrugineus gen. nov., comb. nov. 313

Rba. sphaeroides, Rba. azotoformans, Rba. capsula- sphaeroides ATCC 11167T (AB014944), Rba. veld- tus, Rba. blasticus and Rba. veldkampii at 34% boot- kampii DSM 11550T (AB065366), Rdv. strictum JCM strap value. Agrobacterium ferrugineum, Rba. capsula- 9220T (AB014945), and Rsb. litoralis IFO 15278T tus, Rba. blasticus, Rba. sphaeroides and Rba. azoto- (AB014941). The amino acid sequences of the first do- formans formed a cluster at 51% bootstrap value. And, main of an N-terminal fragment of DNA gyrase B sub- A. ferrugineum, Rba. sphaeroides and Rba. azotofor- unit were used for the phylogenetic analysis. The posi- mans formed a cluster at 100% bootstrap value. The tions with gaps and the positions for which sequences topography of the NJ tree using a Jukes-Cantor model could not be aligned were removed, and finally, 101 was similar to one using a Kimura 2-parameter model. residues were used for this analysis. Figure 3 shows The 16S rDNA sequence similarity values between A. the ML tree based on amino acid sequences of DNA ferrugineum and Rba. azotoformans, Rba. blasticus, gyrase B subunit for a Poisson model. Agrobacterium Rba. capsulatus, Rba. sphaeroides and Rba. veld- ferrugineum formed a cluster with the species of the kampii were 95.7, 93.3, 94.2, 95.3 and 93.8%. genus Rhodobacter, Rba. capsulatus, Rba. blasticus, Rba. sphaeroides and Rba. azotoformans at 64% 23S rDNA sequencing and phylogenetic analysis bootstrap value. Rhodobacter capsulatus clustered Sequences of the 23S rDNA involving a part of 16S with Rba. blasticus at 68% bootstrap value. Agrobac- rDNA and 5S rDNA of A. ferrugineum IAM 12616T, terium ferrugineum clustered with this group at 100% Rba. azotoformans, Rba. blasticus and Rba. veld- bootstrap value. Rhodobacter sphaeroides and Rba. kampii determined in this study were deposited in the azotoformans clustered with each other at 99% boot- DNA Data Bank of Japan (DDBJ) under serial acces- strap value. Rhodobacter veldkampii clustered with P. sion numbers AB050736, AB050733, AB050734 and aminophilus at 68% bootstrap value. AB050735, respectively. The positions with gaps and the undetermined and ambiguous sequences were re- Southern hybridization moved, and finally, 2,658 sites were used for this phy- The result of the Southern hybridization is shown in logenetic analysis. Figure 2 shows the NJ tree from Fig. 4. In lane 4 (Rba. sphaeroides), one strong signal distances estimated by a Kimura 2-parameter model. (A) appeared at the position of about 10 to 15 kb, one Agrobacterium ferrugineum formed a cluster with the weak signal (B) appeared below 10 kb and one weak species of the genus Rhodobacter, Rba. blasticus, signal (C) appeared at about 1.5 kb. The length of the Rba. sphaeroides and Rba. azotoformans at 100% Rba. sphaeroides genome fragment digested by bootstrap value, and Rba. blasticus was nearer to A. EcoRI containing puf L is 11,845 bp to which the signal ferrugineum. These two species formed a cluster at (A) is equivalent. Signals (B) and (C) were non-spe- 99% bootstrap value. Rhodobacter capsulatus and cific. In lane 5 (A. ferrugineum), no signals appeared. Rba. veldkampii were not involved in the cluster con- These results suggested that Rba. sphaeroides had sisting of the other species of Rhodobacter and A. fer- one copy of the puf L gene and A. ferrugineum did not rugineum. This species clustered with P. denitrificans have the puf L gene. at 57% bootstrap value. The topography of the NJ tree using a Jukes-Cantor model was similar to the topog- PCR-detection of genes related to photosynthesis raphy using a Kimura 2-parameter model. From the results of a pre-experiment using Rba. capsulatus, we decided to use the primer sets pufL1F- gyrB sequencing and phylogenetic analysis pufL3R (annealing: 50°C), pufM2F-pufM4R (50°C) and The sequences determined in this study were parts puhAF1-puhAR1 (60°C). The presence of puf L, puf M coding two crystallographic domains of an N-terminal and puhA genes was confirmed in all species of fragment of DNA gyrase B subunit (Reece and Rhodobacter, but not in A. ferrugineum (Fig. 5, A–C). Maxwell, 1991). The sequences of 9 species were deposited in DDBJ under accession numbers, A. ferru- DNA base composition gineum IAM 12616T (AB014904), P. aminophilus IAM The DNA GC content of A. ferrugineum IAM 14245T (AB014972), Rba. azotoformans IAM 14814T 12616T was 58 mol% which is the same as in a previ- (AB065365), Rba. blasticus DSM 2131T (AB065367), ous report (Rüger and Höfle, 1992). Rba. capsulatus ATCC 11166T (AB014940), Rba. 314 UCHINO, HAMADA, and YOKOTA Vol. 48

Fig. 1 2002 Pseudorhodobacter ferrugineus gen. nov., comb. nov. 315

NaCl requirement Rhodobacter species have a close relationship. In the Agrobacterium ferrugineum grew in the presence of trees based on 23S rRNA and DNA gyrases, Rba. 0 to 3% NaCl and required 1% NaCl for its optimum veldkampii and Rba. capsulatus were separated from growth. the other species of the genus Rhodobacter. If the number of Operational Taxonomic Units is increased, Discussion we think that Rhodobacter species may form one cluster in these trees. In this report as well as in our previous report The conclusion in the former report that (1) A. fer- (Uchino et al., 1998), the absence of puf genes in A. rugineum does not have photosynthetic activities and ferrugineum was proven from the result of searching that (2) A. ferrugineum and the Rhodobacter species by PCR using other primer sets and Southern hy- are closely related was not changed by data reported bridization. Furthermore, the presence of the puhA here. gene was denied by PCR with primers defined from At first, we thought that A. ferrugineum might pos- puhA gene, too. Thus, A. ferrugineum does not seem sess part of the photosynthetic apparatus. Rhodobac- to possess genes localized near the opposite ends of ter species can grow not only photosynthetically but the photosynthesis gene cluster of Rba. capsulatus also heterotrophically. Even if a Rhodobacter strain (Fig. 6). These data suggest that A. ferrugineum may has a defect in the elements for photosynthesis and have lost the whole photosynthetic gene cluster. This loses photosynthesis, it can survive. In such a case, hypothesis provides the explanation why we were un- probably, parts of the photosynthesis equipment will able to detect BChl a in A. ferrugineum because the remain. If these genes could be found in A. ferrugi- genes coding the enzymes for the synthesis of BChl a neum, it might suggest that this strain was a mutant of generally exist on the photosynthetic gene cluster of photosynthetic bacterium, and A. ferrugineum could be the related photosynthetic bacteria. No evidence was proposed to be transferred to the genus Rhodobacter. found for photosynthetic abilities of A. ferrugineum. However, this does not seem to be the case for A. In this report as well as in the previous report ferrugineum. Our experiments indicate that genes for (Uchino et al., 1998), the close phylogenetic relation- photosynthesis seem to be absent in A. ferrugineum ships between A. ferrugineum and the photosynthetic and the GC content of this species greatly differs species of the genus Rhodobacter were strongly sup- from that of the Rhodobacter species (A. ferrugineum ported by the phylogenies based on various molecules IAM 12616T: 58%, Rhodobacter: 64–73%). It has been (Figs. 1–3). The topologies of the trees based on three suggested that large-scale rearrangement events hap- different molecules (16S rRNA, 23S rRNA and DNA pened in the genome of this species. Possibly the pho- gyrases) are different from each other. But the stability tosynthesis gene cluster was lost at the time of such of the clusters consisting of A. ferrugineum and some an event. species of the genus Rhodobacter in these trees was Imhoff et al. (1984) stated that within a determinative supported by high bootstrap values. In the tree based taxonomic system the phototrophic bacteria should be on 16S rRNA, A. ferrugineum, Rba. azotoformans and treated separately from their non-phototrophic “rela- Rba. sphaeroides formed a cluster at 100% bootstrap tives.” At present, since there is no evidence showing value (Fig. 1). In the tree based on 23S rRNA, A. fer- the relation between A. ferrugineum and photosynthe- rugineum, Rba. azotoformans, Rba. blasticus and sis, we propose therefore the transfer of A. ferrugineum Rba. sphaeroides formed a cluster at 100% bootstrap IAM 12616T to the new genus Pseudorhodobacter as value (Fig. 2). In the tree based on DNA gyrase, A. fer- Pseudorhodobacter ferrugineus. rugineum, Rba. blasticus and Rba. capsulatus formed Pseudorhodobacter ferrugineus has characteristics a cluster at 100% bootstrap value (Fig. 3). These fig- similar to the Rhodobacter species on the following ures reflect that the genomes of A. ferrugineum and point: growth occurs without NaCl in the culture

Fig. 1. NJ tree of Proteobacteria a-3 subgroup species based on 16S rDNA sequences. The tree was constructed by the NJ method from the data sets aligned on the 1,063 sites in 16S rDNA using CLUSTAL W. The percentage of bootstraps was derived from 1,000 resamplings. Bold lines indicated branches whose bootstrap values were more than 95%. 316 UCHINO, HAMADA, and YOKOTA Vol. 48

Fig. 2. NJ tree based on 23S rDNA sequences. The tree was constructed by the NJ method from the data sets aligned on the 2,658 sites in 23S rDNA using CLUSTAL W. The percentage of bootstraps was derived from 1,000 resamplings. Bold lines indicated branches whose bootstrap values were more than 95%.

Fig. 3. ML trees based on gyrB amino acid sequences. The tree was constructed by the ML method employing MOLPHY version 2.3b (UNIX) from the data sets aligned on the 101 sites in gyrB using CLUSTAL W. The local bootstrap probabilities (LBPs) were derived using ProtML. Bold lines indicated branches whose bootstrap values were more than 95%. 2002 Pseudorhodobacter ferrugineus gen. nov., comb. nov. 317

Fig. 6. The photosynthesis gene cluster of Rhodobacter capsulatus (accession number: Z11165). Agrobacterium ferrugineum does not have puf or puhA genes localized near the opposite ends of the photosynthesis gene cluster.

Descriptions

Description of Pseudorhodobacter gen. nov. Fig. 4. Detection of puf L by Southern hybridization. (Pseudorhodobacter Pseudo.rho.do.bac’ter; Gr. adj. Lane 1, l-EcoT14 I DNA standard marker; lane 2, total pseudes false; Gr. n. rhodon rose; N.L. masc. n. bacter DNA of Rhodobacter sphaeroides digested by EcoRI; lane 3, equivalent of Gr. neut. n. bakterion rod; N.L. masc. n. total DNA of Agrobacterium ferrugineum digested by EcoRI; Pseudorhodobacter, false Rhodobacter red-colored lane 4, the result of Southern hybridization of Rhodobacter rod) sphaeroides; lane 5, the result of Southern hybridization of The cells are gram-negative rods, 0.6 to 1.6 mm Agrobacterium ferrugineum. wide and 1.0 to 4.0 mm long, and do not form spores. Non-motile. Aerobic chemoorganotrophic bacterium with strictly respiratory type of metabolism with oxygen as terminal electron acceptor. No photosynthetic growth occurs under both oxic and anoxic conditions. Bacteriochlorophyll is absent. Oxidase and catalase are produced. The major quinone is ubiquinone 10.

The major fatty acid is C18:1. The 3-hydroxy fatty acid is

C10:0 3-OH and C14:1 3-OH. The 2-hydroxy fatty acids are absent. The GC content of the DNA is 58 mol%. The type species is Pseudorhodobacter ferrugineus.

Description of Pseudorhodobacter ferrugineus comb. Fig. 5. Detection of puf L (A), puf M (B) and puhA (C) genes nov. by amplifying their partial DNA fragments by PCR. Lane 1, 100 bp DNA ladder marker; lane 2, Rhodobacter cap- (ferrugineus fer.ru.gi’ne.us; L. adj. ferrugineus, re- sulatus; lane 3, Rhodobacter sphaeroides; lane 4, Rhodobacter sembling iron rust, dark red) azotoformans; lane 5, Rhodobacter blasticus; lane 6, Rhodobac- Basonym: Agrobacterium ferrugineum Rüger and ter veldkampii; lane 7, Agrobacterium ferrugineum. Höﬂe 1992. The following description is based on our own ob- medium (optimal NaCl concentration for growth of P. servations and the previous description of the species ferrugineus is 1%) and their major hydroxy fatty acid by Rüger and Höﬂe (1992). Colonies were circular; ini- compositions are 3-hydroxy decanoic acids (3-OH tially translucent, colorless or light brown; later dark 10:0) and 3-hydroxy tetradecanoic acids (3-OH 14:1). brown with brown center. Nitrate is not reduced to ni- However, P. ferrugineus is clearly distinguishable from trite or gas. Acids are produced from fructose and xy- Rhodobacter species by the low GC content, an in- lose after 4 to 6 weeks of incubation. Isolated from sertion in its 16S rRNA gene sequence, and the lack of seawater of the Baltic Sea. Growth occurs in the pres- photosynthetic abilities, bacteriochlorophyll a, intracy- ence of NaCl concentrations ranging from 0 to 3% (op- toplasmic membrane systems and the photosynthetic timum NaCl concentration, 1%). The major quinone is genes (Table 3). ubiquinone 10. The major fatty acid is C18:1. The 3-hy- 318 UCHINO, HAMADA, and YOKOTA Vol. 48

droxy fatty acid is C10:0 3-OH and C14:1 3-OH. The 2- hydroxy fatty acids are absent. The GC content of the DNA is 58 mol%. rRNA

Insertion T T in the 16S

TGGTGGG The type strain is IAM 12616 ( ATCC 25652 ). C

Acknowledgments G (mol%) We thank Mr. Kengo Tsuji for technical support woth Southern hybridization. Quinone content References

ND Q-10 62–69 — Adachi, J. and Hasegawa, M. (1996) Computer Science Mono- graphs, No. 28. MOLPHY Version 2.3.—Programs for Mo- e

fatty acids lecular Phylogenetics Based on Maximum Likelihood, Insti- ND ND Q-10 51–63 — ND ND Q-10 64–67 — ND Major hydroxy 2-OH 3-OH (190–204) None 10:0, 14:1 Q-10 58 CACTGTAG- None 10:0, 14:1 Q-10 64–73 — tute of Statistical Mathematics, Tokyo. Doronina, N. V., Trotsenko, Y. A., Kuznetsov, B. B., and a Tourova, T. P. (2002) Emended description of Paracoccus kondratievae. Int. J. Syst. Evol. Microbiol., 52, 679–682. Felsenstein, J. (1981) Evolutionary trees from DNA sequences: and related genera. a maximum likelihood approach. J. Mol. Evol., 17, 368– 376.

growth Felsenstein, J. (1985) Conﬁdence limits on phylogenies; an approach using the bootstrap. Evolution, 39, 783–791. Photosynthetic Hasegawa, M. and Kishino, H. (1994) Accuracies of the simple methods for estimating the bootstrap probability of a maxi-

Pseudorhodobacter mum-likelihood tree. Mol. Biol. Evol., 11, 142–145. Hiraishi, A., Muramatsu, K., and Ueda, Y. (1996) Molecular genetic analyses of Rhodobacter azotoformans sp. nov. and related species of phototrophic bacteria. Syst. Appl. Micro- biol., 19, 168–177.

for optimal growth Hiraishi, A. and Ueda, Y. (1994) Intrageneric structure of the Requirement Aerobic genus Rhodobacter: Transfer of Rhodobacter sulﬁdophilus and related marine species to the genus Rhodovulum gen. nov. Int. J. Syst. Bacteriol., 44, 15–23. Hiraishi, A. and Ueda, Y. (1995) Isolation and characterization of Rhodovulum strictum sp. nov. and some other purple environments environments nonsulfur bacteria from colored blooms in tidal and seawater pools. Int. J. Syst. Bacteriol., 45, 319–326. able 3. Comparison among the genus

T Imhoff, J. F., Trüper, H. G., and Pfennig, N. (1984) Rearrange- ments of the species and genera of the phototrophic purple nonsulfur bacteria. Int. J. Syst. Bacteriol., 34, 340–343. Jukes, T. H. and Cantor, C. R. (1969) Evolution of protein molecules. In Mammalian Protein Metabolism, Vol III, ed. by Munro, H. N., Academic Press, New York., pp. 21–132. Rod Non-motile Marine Cocci Non-motile Activated shaped non-motile hypersaline shaped non-motile and terrestrial tetrads) Katayama, Y., Hiraishi, A., and Kuraishi, H. (1995) Paracoccus short rod non-motile Coccus to Motile and Soil, sewage Cell shape Motility Habitat of NaCl dark BChl (arranged in sludge, sewage Ovoid to rod Motile andOvoid to rod Freshwater Motile and Marine or thiocyanatus sp. nov., a new species of thiocyanate-utiliz- ing facultative chemolithotroph and transfer of Thiobacillus versutus to the genus Paracoccus as Paracoccus versutus a d b comb. nov., with emendation of the genus. Microbiology, c 141, 1469–1477. Kawasaki, H., Hoshino, Y., Kuraishi, H., and Yamasato, K. (1992) Rhodocista centenaria gen. nov., sp. nov., a cyst- Imhoff et al. (1984), and Hiraishi (1996). Imhoff Hiraishi and Ueda (1994, 1995). Maszenan et al. (1997). No data. Doronina et al. (2002), Katayama et al. (1995), Rainey et al. (1999), and Tsubokura et al. (1999). Doronina et al. (2002), Katayama (1995), Rainey (1999), and Tsubokura a b c d e Rhodobacter Amaricoccus Paracoccus Pseudorhodobacter Rhodovulum forming anoxygenic phototrophic bacterium and its phylo- 2002 Pseudorhodobacter ferrugineus gen. nov., comb. nov. 319

genetic position in the Proteobacteria alpha group. J. Gen. Saitou, N. and Nei, M. (1987) The neighbor-joining method: a Appl. Microbiol., 38, 541–551. new method for reconstructing phylogenetic trees. Mol. Kimura, M. (1980) A simple method for estimating evolutionary Biol. Evol., 4, 406–425. rate if base substitutions through comparative studies of Stackebrandt, E., Rainey, F. A., and Ward-Rainey, N. (1996) nucleotide sequences. J. Mol. Evol., 16, 111–120. Anoxygenic phototrophy across the phylogenetic spectrum: Ludwig, W., Kirchof, G., Klugbauer, N., Weizenegger, M., Betzl, Current understanding and future perspectives. Arch. Mi- D., Ehrmann, M., Hertel, C., Jilg, S., Tatzel, R., Zitzels- crobiol., 166, 211–223. berger, H., Liebl, S., Hochberger, M., Shah, J., Lane, D., Swofford, D. L. (2000) PAUP*. Phylogenetic Analysis Using and Wallnoef, P. R. (1992) Complete 23S ribosomal RNA Parsimony (*and Other Methods), Version 4, Sinauer Asso- sequences of Gram-positive bacteria with a low DNA GC ciates, Sunderland, Massachusetts. content. Syst. Appl. Microbiol., 15, 487–501. Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994) Marmur, J. (1961) A procedure for the isolation of deoxyribonu- CLUSTAL W: Improving the sensitivity of progressive multi- cleic acid from microorganisms. J. Mol. Biol., 3, 208–218. ple sequence alignment through sequence weighting, posi- Maszenan, A. M., Seviour, R. J., Patel, B. K. C., Rees, G. N., tion specific gap penalties and weight matrix choice. Nu- and McDougall, B. M. (1997) Amaricoccus gen. nov., a cleic Acids Res., 22, 4673–4680. Gram-negative coccus occurring in regular packages or Tsubokura, A., Yoneda, H., and Mizuta, H. (1999) Paracoccus tetrads, isolated from activated sludge biomass, and de- carotinifaciens sp. nov., a new aerobic Gram-negative as- scriptions of Amaricoccus veronensis sp. nov., Amaricoc- taxanthin-producing bacterium. Int. J. Syst. Bacteriol., 49, cus tamworthensis sp. nov., Amaricoccus macauensis sp. 277–282. nov., and Amaricoccus kaplicensis sp. nov. Int. J. Syst. Uchino, Y., Hirata, A., Yokota, A., and Sugiyama, J. (1998) Re- Bacteriol., 47, 727–734. classification of marine Agrobacterium species: Proposals Mesbah, M. and Whitman, W. B. (1989) Measurement of de- of Stappia stellulata gen. nov., comb. nov., Stappia aggre- oxyguanosine/thymidine ratios in complex mixtures by gata sp. nov., nom. rev., Ruegeria atlantica gen. nov., high-performance liquid chromatography for determination comb. nov., Ruegeria gelatinovora comb. nov., Ruegeria al- of the mole percentage guaninecytosine of DNA. J. Chro- gicola comb. nov., and Ahrensia kieliense gen. nov., sp. matogr., 479, 297–306. nov., nom. rev. J. Gen. Appl. Microbiol., 44, 201–210. Rainey, F. A., Kelly, D. P., Stackebrandt, E., Burghardt, J., Hi- Uchino, Y., Yokota, A., and Sugiyama, J. (1997) Phylogenetic raishi, A., Katayama, Y., and Wood, A. P. (1999) A re-eval- position of the marine subdivision of Agrobacterium species uation of the taxonomy of Paracoccus denitrificans and a based on 16S rRNA sequence analysis. J. Gen. Appl. Mi- proposal for the combination Paracoccus pantotrophus crobiol., 43, 243–247. comb. nov. Int. J. Syst. Bacteriol., 49, 645–651. Yamamoto, S., Bouvet, P. J., and Harayama, S. (1999) Phylo- Reece, R. J. and Maxwell, A. (1991) DNA gyrase: Structure and genetic structures of the genus Acinetobacter based on function. CRC Crit. Rev. Biochem. Mol. Biol., 26, 335–375. gyrB sequences: Comparison with the grouping by DNA- Rüger, H.-J. and Höfle, M. G. (1992) Marine star-shaped-aggre- DNA. Int. J. Syst. Bacteriol., 49, 87–95. gate-forming bacteria: Agrobacterium atlanticum sp. nov.; Yamamoto, S. and Harayama, S. (1995) PCR amplification and Agrobacterium meteori sp. nov.; Agrobacterium ferrugi- direct sequencing of gyrB genes with universal primers and neum sp. nov., nom. rev.; Agrobacterium gelatinovorum sp. their application to the detection and axonomic analysis of nov., nom. rev.; and Agrobacterium stellulatum sp. nov., Pseudomonas putida strains. Appl. Environ. Microbiol., 61, nom. rev. Int. J. Syst. Bacteriol., 42, 133–143. 1104–1109.