Yoshiaki Kawamura, Department of Microbiology, School of Pharmacy, Aichi Gakuin University, 1-100 Kusumoto-cho, Nagoya 464-8650, Japan. Tel: +81 52 757 6780; fax: +81 52 757 6799; email: email@example.com
The genus Enhydrobacter, first reported as a member of the family Vibrionaceae, has been placed in the family Moraxellaceae, but as a genus incertae sedis in Bergey's Manual of Systematic Bacteriology 2nd edition. During our taxonomic investigation of Enhydrobacter-like organisms, we observed that the 16S rRNA sequences of E. aerosaccus-type strain versions NCIMB 12535T, ATCC 27094 T and CCUG 58314T were very different from the accessible data (accession no. AJ550856). Phylogenetic analysis of our 16S rRNA sequence data revealed that these organisms were located within the family Rhodospirillaceae. The genera Inquilinus, Oceanibaculum, Skermanella and Nisaea were closely related (sequence similarities were 88.3∼87.0%), but Enhydrobacter could be distinguished from these genera by growth characteristics, fatty acid profiles (C19:0 cyclo ω8c; 38.4% C18:1ω7c; 32.2%, and C16:0; 8.9% were major components), in being non-flagellated, and differing in enzymatic activities, including trypsin and β-glucosidase. From these data, we conclude that the genus Enhydrobacter should be recognized as an independent genus of the family Rhodospirillaceae within the class Alphaproteobacteria.
Culture Collection, University of Goteborg, Sweden
high-performance liquid chromatography
Laboratorium voor Microbiologie, Universiteit Gent, Belgium
National Collections of Industrial Food and Marine Bacteria, UK
Department of Microbiology, School of Pharmacy, Aichi Gakuin University, Japan
transmission electron microscope
The genus Enhydrobacter with its single species Enhydrobacter aerosaccus had been proposed by Staley et al. (1) in 1987 as a gas-vacuolated facultative anaerobic, heterotrophic rod. The genus Enhydrobacter was allocated tentatively to the family Vibrionaceae based on phenotypic features most closely resembling members of the genus Aeromonas in growing on a defined medium with D-glucose as the sole carbon source and using ammonium salts as the sole nitrogen source for growth. Enhydrobacter could use some amino acids such as L-alanine, L-serine, and L-arginine as carbon sources for growth, a feature that was also found in some Aeromonas sp. However, E. aerosaccus differs from Aeromonas sp. because of the high G+C mol% of DNA (66.3% buoyant density for E. aerosaccus versus 57–63% for Aeromonas sp.). E. aerosaccus did not grow on some routine media, including nutrient agar, King A and B media, methyl red-Voges Proskauer medium (Clark Lubs), and blood agar. Gas vacuoles were a characteristic of E. aerosaccus, a feature not observed in any Aeromonas sp.
In 2003, the 16S rRNA sequence of the type strain of E. aerosaccus LMG 21877 was determined (DDBJ/EMBL accession no. AJ550856). Phylogenetic analysis indicated that this organism was not closely related to the family Vibrionaceae, but was highly related to the type strain of Moraxella osloensis CCUG350T (FR726160), with almost 100% sequence homology (2). However, many of the biochemical features of E. aerosaccus, described above differed from those of the genus Moraxella. In Bergey's Manual of Systematic Bacteriology 2nd edition, this genus was described as ‘Family III incertae sedis, Genus I Enhydrobacter’ in the family Moraxellaceae (3). Consequently, the taxonomic position of the genus Enhydrobacter has not yet been resolved.
In the present study, we conducted a polyphasic approach in order to determine the taxonomic position of E. aerosaccus.
MATERIALS AND METHODS
Strains used and culture conditions
We compared four versions of the type strain of Enhydrobacter aerosaccus; PAGU 1623T (= NCIMB 12535T), PAGU 1624T (= ATCC27094T), PAGU1700T (= CCUG 58314T), and PAGU 1622T (= LMG 21877T). We also examined the type strain of Moraxella osloensis PAGU 1621T (= CCUG 350T), Inquilinus limosus PAGU 1644T (= LMG 20952T), and Inquilinus ginsengisoli PAGU 1645T (= LMG 23638T) all purchased directly from the culture collections.
Casamino acids medium (ATCC medium 590) was used for growing E. aerosaccus strains at 37°C under aerobic or microaerobic conditions (6% O2, 6% CO2, 88% N2). We used 5% defibrinated sheep blood agar, trypticase soy agar (TSA; Difco, Detroit, MI, USA), CLED medium (Oxoid, Basingstoke, UK) and TGE medium (0.5% Trypton, 0.3% beef extract, 0.1% glucose, 1.5% agar) for the growth tests.
16SrRNA gene sequence determination and phylogenetic analysis
16S rRNA gene sequencing of the purified products of PCR amplification was carried out (4). The sequence was determined using a BigDye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, Tokyo, Japan). 16S rRNA gene (>1300 bases) sequences of both strands of the gene were determined using the 3130 Genetic Analyzer (Applied Biosystems). Other sequences used for alignment and for calculating levels of homology were obtained from the DDBJ/EMBL database. Multiple sequence alignments of DNA sequences were carried out using Clustal X software (5). Phylogenetic distances were calculated using the neighbor-joining method (6). The phylogenetic tree was constructed using TreeView software (7) and NJplot software (8).
The organisms were biochemically characterized using the API-ZYM and API-20NE (bioMérieux, Tokyo, Japan) systems, according to the manufacturers’ instructions, with the results observed after 24 hrs for API-ZYM and after 72 hrs for API-20NE. Catalase activity was determined by bubble production in 3% (v/v) H2O2. Oxidase activity was observed by using a commercially available oxidase test paper (Eiken Chemical Co., Tokyo, Japan).
For inducing gas vacuolation, the strain was incubated on the casamino acids medium with 0.1% pyruvate added instead of glucose as a carbon source (1). Gas vacuolation was observed using a confocal laser scanning microscope (LSM 510META; Carl Zeiss, Jena, Germany) and a transmission electron microscope (TEM; Model H7650; Hitachi, Tokyo, Japan). Sample preparation for TEM was according to Nakayama et al. (9), but used LR white acrylic resin (London Resin Company, London UK).
Quinones were extracted as described previously (10, 11). The purified quinones were analyzed by high-performance liquid chromatography (HPLC) with a solvent containing methanol and 2-propanol (2:1, v/v) using a model L-2400 HPLC system (Hitachi). Cellular lipids and fatty acids were analyzed as described previously (11–13). Methyl ester derivatives of extracted fatty acids were made by treatment with 10% trimethylsilyldiazomethane in n-hexane (Nacalai Tesque Inc., Kyoto, Japan), and analyzed by gas chromatography-mass spectrometry (GC/MS).
DNA G+C content
The guanine-plus-cytosine (G+C) mol% content was determined by HPLC (14). A total of 5 μg denatured DNA was hydrolyzed with P1 nuclease (Yamasa Syoyu, Chiba, Japan) for 1 hr at 50°C. Alkaline phosphatase (Sigma, St Louis, MO, USA) was then added, and the mixture was incubated at 37°C for 30 min for nucleotide dephosphorylation. The nucleosides were quantified with a GC analysis standard (Yamasa Syoyu) using a model L-2400 HPLC system (Hitachi) and an Inertsil ODS-3 HPLC Column (GL Sciences, Tokyo, Japan). The nucleosides were eluted with a solvent containing 0.2 M NH4H2PO4 and acetonitrile (20:1, v/v). G+C mol% was determined using the mean values of three experiments.
Strain growth and morphology
Three versions of the type strain of E. aerosaccus, PAGU 1623T (= NCIMB 12535T), PAGU 1624T (= ATCC27094T), and PAGU1700T (= CCUG 58314T), could grow on casamino acids medium at 37°C under aerobic or microaerobic conditions. Microaerobic incubation was better for growing these strains although visible colonies took 5–7 days to grow. These type strain versions could not grow on blood agar or on TSA. On CLED medium, these three versions could grow only under microaerobic conditions. In contrast, PAGU 1622 (= LMG 21877), which was also designated as a type strain of E. aerosaccus, could grow on blood agar, TSA, casamino acids medium and CLED medium and made visible colonies within 2 days. The colony morphology of PAGU 1622 was very similar to M. osloensis PAGU 1621T (= CCUG 350T). We observed gas vacuolation from PAGU 1623, but not from PAGU 1622 (Fig. 1).
Phylogenetic position of E. aerosaccus
The 16S rRNA sequences of ‘E. aerosaccus’ PAGU 1622 and M. osloensis PAGU 1621T were in agreement with published data (AJ550856 and FR726160) in being almost identical to each other, as previously described (2).
Three versions of the type strain of E. aerosaccus, PAGU1623T (= NCIMB12535T, Acc. No. AB641398) and PAGU 1624T (= ATCC27094T, Acc. No. AB41399), and PAGU1700T (= CCUG58314T. Acc. No. AB41400)) shared identical 16S rRNA sequence homology, but were very different from AJ550856. BLAST search for AB41398, carried out on the DDBJ homepage, revealed that the sequence was close with ‘Candidatus Reyranella massiliensis strain’ or ‘Alpha proteobacterium strains’.
To determine the phylogenetic position of the genus Enhydrobacter, we collected representative sequences of the class Proteobacteria (119 sequences, including the type genera of 27 families in 10 orders), and carried out the phylogenetic analysis. As a result, the genus Enhydrobacter was included in the family Rhodospirillaceae within the order Rhodospirillales (data not shown).
Detailed phylogenetic analysis was carried out using the sequence data of the type species of all genera within the family Rhodospirillaceae (Fig. 2) and we reconfirmed that the genus Enhydrobacter was located within the family Rhodospirillaceae, and was remotely related to the genera Inquilinus, Oceanibaculum, Skermanella and Nisaea with homology values of 88.3∼87.0%.
Differential chemotaxonomic and phenotypic characters
From the phylogenetic analysis, E. aerosaccus was clearly located within the family Rhodospirillaceae. Several chemotaxonomic and phenotypic features could differentiate this organism from phylogenetically related genera Inquilinus, Oceanibaculum, Skermanella and Nisaea as shown in Table 1. E. aerosaccus are distinguished in having no flagellae, very low salt tolerance and a relatively high content (more than 38%) of C19:0 cyclo ω8c as a major fatty acid.
Table 1. Selected characteristics that differentiate genus Enhydrobacter from other phylogenetically related genera in the family Rhodospirillaceae
Colorless to pale white
Genus: 1, Enhydrobacter (our data, strain PAGU1623T and (1); 2, Inquilinus (our data, strain PAGU 1644Tand PAGU 1645T, and taken from 19, 20); 3, Skermanella (data taken from 21); 4, Oceanibaculum (data taken from 22, 23); 5, Nisaea (data taken from 24). +, positive; w, weakly positive; −, negative; V, variable; MP, monopolar; BP, bipolar; NA, not available.
Cell size (width × length; μm)
0.5–0.7 × 1.0–5.0
0.6–0.8 × 2.5–4.0
1–1.5 × 2–3
0.6–1.5 × 2.3–2.5
0.7–1.1 × 1.9–3.1
Coccobacillary to rod shaped
Temperature range (°C)
Temperature optimum (°C)
Salt tolerance (%, w/v)
Major fatty acid (%)
C19:0 cyclo ω8c
DNA G+C content (mol%)
Water, eutrophic lake
Cystic fibrosis patients
Lake water, air, sand
Coastal surface seawater
We used four versions of the type strain of Enhydrobacter aerosaccus, each from different culture collections (NCIMB, ATCC, CCUG, and LMG). The phenotypic and chemotaxonomic features of one version PAGU 1623T (= NCIMB 12535T), are according to the original description of E. aerosaccus (1) and we therefore consider this strain to be the type strain of E. aerosaccus. Versions PAGU 1624T (= ATCC27094T) and PAGU 1700T (= CCUG 58314T) also share identical 16S rRNA sequences with version PAGU 1623T (= NCIMB 12535T) and are probably the same strain. However, LMG 21877 is clearly not a strain of E. aerosaccus, differing in growth properties, G+C mol% (43%), 16S rRNA sequence and other features. Unfortunately, the 16S rRNA sequence of this strain (Acc. No. AJ550856) was first published as a representative strain of E. aerosaccus and, as a consequence, confusion may easily arise. Several studies relating to descriptions of the genus Moraxella have linked these sequence data to closely related organisms (15–18). In Bergey's Manual of Systematic Bacteriology 2nd edition, this organism was described within the family Moraxellaceae, but as a genus incertae sedis (3). In addition, the whole genome sequence has been determined for strain SK60 and submitted as an example of E. aerosaccus (GenBank, Acc. No. ACYI01000000); however, the total G+C mol% was 43%, indicating that this strain might belong to the genus Moraxella.
These data demonstrate that strain LMG 21877 is not a strain of E. aerosaccus, and is possibly related to M. osloensis. However, some biochemical reactions of LMG 21877 differ from those of M. osloensis, such as being positive for nitrate reduction (by API 20NE), and being negative for valine arylamidase and cystine arylamidase (by API-ZYM). Clearly, further investigations are needed to fully clarify the situation.