Algoriphagus machipongonensis sp. nov., co-isolated with a colonial choanoflagellate

The Algoriphagus genus belongs to the phylum Bacteroidetes. The type species, Algoriphagus ratkowskyi IC025^T, was first described by Bowman et al. (2003). At the time of writing, the genus consists of 23 species predominantly isolated from aquatic habitats (Bowman et al., 2003; Copa-Patiño et al., 2008; Lee et al., 2012; Li et al., 2011; Liu et al., 2009; Nedashkovskaya et al., 2004, 2007; Oh et al., 2012; Park et al., 2010; Rau et al., 2012; Tao et al., 2006; Tiago et al., 2006; Van Trappen et al., 2004; Yoon et al., 2005a, b, 2006; Young et al., 2009). In this study, we report the taxonomic characterization of an Algoriphagus-like bacterial strain, PR1^T, which was co-isolated with the choanoflagellate Salpingoeca rosetta ATCC 50818 in May 2000 from a mud core sample near Hog Island, Virginia, USA, part of the coastal barrier system of the Virginia Coast Reserve (Dayel et al., 2011).

Strain PR1^T was isolated by the standard dilution-plating technique on modified ZoBell medium at 25 °C (Carlucci & Pramer, 1957). Algoriphagus halophilus JC 2051^T, obtained from the Korean Collection for Type Cultures, was used as a reference strain for phenotypic characterization and fatty acid analysis.

The morphological characteristics of strain PR1^T were investigated after cultivation on seawater complete media (SWC; Atlas, 2004) at 25 °C for 5 days. Cell morphology was examined by light microscopy (DMIL; Leica). The Gram reaction and absorption maximum of crude extracts were determined as described by Tindall et al. (2007). Organic cell extracts of strain PR1^T showed an absorption peak maximum at 487 nm, which indicated the presence of carotenoids. Flexirubin-type pigments were not produced, as shown by a negative KOH test result (Reichenbach, 1989). Growth with 0, 0.5, 1.0, 2.0 and 4.0 % (w/v) NaCl was examined in trypticase soy broth (Difco) that had been prepared according to the manufacturer’s instructions except that NaCl was added to the desired final concentration and 0.45 % (w/v) MgCl₂ . 6H₂0 or 0.06 % (w/v) KCl were used as supplements. Growth with 2–10 % (w/v) NaCl (in increments of 1 %) was investigated in marine broth 2216 (MB; Difco), containing a base of 2 % NaCl and supplemented with additional NaCl. Growth at pH 4.5–9.5 (in increments of 0.5 pH units) was investigated in MB by addition of HCl or Na₂CO₃. Growth was detected by changes in OD₆₀₀ for 3 days. Growth at 4, 10, 20, 25, 28, 30, 35, 37, 40 and 45 °C was measured on marine agar 2216 (MA; Difco). Carbon source assimilation was determined using the GN2 MicroPlate system (Biolog), according to the manufacturer’s instructions. Acid production from carbohydrates was determined using API 50 CH test trips and CH B/E medium (bioMérieux) according to the manufacturer’s recommendations. Evaluation of growth was performed after 2 and 5 days. Susceptibility to antibiotics was determined by streaking strain PR1^T on modified ZoBell agar containing the following (μg ml⁻¹, unless otherwise stated): polymyxin B (100 U), streptomycin (50), chloramphenicol (100), ampicillin (100), gentamicin (30), neomycin (50), tetracycline (30) or kanamycin (30). Other physiological tests were performed with the API ZYM system (bioMérieux).

Cell biomass of strain PR1^T for analysis of cellular fatty acids was obtained from cultures grown for 1 day in SWC medium at 30 °C. A. halophilus JC 2051^T was used as a reference strain. Cellular fatty acid methyl ester content was determined using the MIDI Sherlock Microbial Identification System (Microbial ID, MidiLabs; Sasser, 1990). In addition, GC-MS analysis was performed to resolve ambiguities in fatty acid identification (Jahnke et al., 2001). For isoprenoid quinone and polar lipid analysis, strain PR1^T was grown for 1 day in SWC medium at 30 °C, harvested, lyophilized and analysed by the Identification Service and Dr Brian Tindall of the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ) Braunschweig, Germany.

DNA was extracted from cell biomass of strain PR1^T grown in modified ZoBell medium using a Bacterial Genomic DNA Mini-prep kit (Bay Gene, Burlingame, CA, USA) according to the manufacturer’s specifications. The 16S rRNA gene was amplified using universal primers 8F (5′- AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-ACCTTGTTACGRCTT-3′; Weisburg et al., 1991). A global sequence alignment of members of the genus Algoriphagus with validly published names and closely related taxa was computed using iterative pairwise methods by the map software program (Huang, 1994). Poorly aligned regions were removed by the software program Gblocks (version 0.91b) using default block parameters (Fig. S1 available in IJSEM Online; Castresana, 2000; Talavera & Castresana, 2007). Sequence alignments have also been deposited online at Treebase (). A distance-matrix method (distance options according to the Kimura two-parameter model), including clustering with the neighbour-joining, maximum-likelihood and discrete character-based maximum-parsimony algorithms, was applied using the phylip version 3.67 software package (Felsenstein, 1989) based on comparison of 1244 base pairs. In each case, the stability of the groups was estimated by bootstrap analysis based on 1000 replications.

The morphological, physiological and biochemical characteristics of strain PR1^T are given in the species description and Table 1. Strain PR1^T was distinguishable from A. halophilus JC 2051^T and Algoiphagus lutimaris S1-3^T by differences in several phenotypic characteristics, most of which were determined under the same conditions and methods (Table 1; Park et al., 2010). The distinctive characteristics of strain PR1^T were as follows: d-galactose, mannitol and glycerol were positive in assays with the Biolog system, but cellobiose, d-fructose, d-mannose, salicin, sucrose, trehalose, l-arabinose, d-glucose, lactose and maltose were negative; alkaline phosphatase, esterase (C4) (weak), valine arylamidase (weak), β-galactosidase (weak), α-glucosidase and β-glucosidase (weak) were present in the API ZYM test, whereas N-acetyl-β-glucosaminidase was absent; acids were not produced from l-arabinose, d-xylose, d-galactose, d-glucose, d-fructose, d-mannose, l-rhamnose, cellobiose, maltose, lactose, melibiose, sucrose, trehalose, melezitose and raffinose in the API 50 CH kit; and sensitivity to gentamicin and tetracycline. The chemotaxonomic analysis was in agreement with the result of phylogenetic classification of strain PR1^T as a member of the genus Algoriphagus (Bowman et al., 2003; Copa-Patiño et al., 2008; Liu et al., 2009; Nedashkovskaya et al., 2004, 2007; Park et al., 2010; Tiago et al., 2006; Van Trappen et al., 2004; Yoon et al., 2005a, b, 2006; Young et al., 2009).

Strain PR1^T had a cellular fatty acid composition similar to its closest phylogenetic neighbours but showed minor differences (Table 2). The major fatty acids (>10 % of total) in strain PR1^T were iso-C_15 : 0 (29.4 %) and summed feature 3 (C_16 : 1ω6c and/or C_16 : 1ω7c; 11.3 %), which are also the dominant fatty acids in other members of the genus Algoriphagus (Table 2; Park et al., 2010), and a third fatty acid, summed feature 9 (C_16 : 0 10-methyl and/or iso-C_17 : 1ω9c; 18.5 %), which was predominant only in strain PR1^T. GC-MS analysis, together with comparison with an authentic standard, showed that 10-methyl fatty acids were below the detection limit and summed feature 9 was, therefore, inferred to be iso-C_17 : 1ω9c. The polar lipid profile consisted of phosphatidylethanolamine, one unknown phospholipid, one unknown aminolipid, one unknown aminophospholipid and two unknown lipids (Fig. S2).

A nearly full-length 16S rRNA gene sequence of strain PR1^T, comprising 1420 nt, was determined in this study. Strain PR1^T was most closely related to A. halophilus JC 2051^T (95.4 % 16S rRNA gene sequence similarity) and A. lutimaris S1-3^T (95.3 %). Lower levels of 16S rRNA gene sequence similarity (91–95 %) were found between strain PR1^T and the type strains of all other species of the genus Algoriphagus, and <88 % sequence similarity was found with the type strains of the other species used in the phylogenetic analysis. A 16S rRNA gene sequence similarity threshold range of 98.7–99 % is recommended as the point above which DNA–DNA hybridization experiments should be mandatory (Stackebrandt & Ebers, 2006). Since strain PR1^T displayed 16S rRNA gene sequence similarity to A. halophilus JC 2051^T and A. lutimaris S1-3^T below the threshold, DNA–DNA hybridization experiments were not performed. The DNA G+C content of strain PR1^T was 38.7 mol%, as determined by genomic sequencing by the J. Craig Venter Institute and Broad Institute (Alegado et al., 2011).

In the neighbour-joining tree based on 16S rRNA gene sequences (Fig. 1), strain PR1^T fell within the cluster comprising the genus Algoriphagus and formed a coherent subcluster with A. halophilus JC 2051^T and A. lutimaris S1-3^T with a bootstrap resampling value of 77.6 %. The close relationship of strain PR1^T, A. halophilus JC 2051^T and A. lutimaris S1-3^T was also found when the maximum-likelihood algorithm was used (Fig. S3).

The phylogenetic distinctiveness of strain PR1^T together with differential phenotypic properties are sufficient to demonstrate that this strain is distinct from previously recognized Algoriphagus species, including A. halophilus and A. lutimaris (Stackebrandt & Goebel, 1994). Therefore, on the basis of phenotypic, chemotaxonomic and phylogenetic data, strain PR1^T is considered to represent a novel species of the genus Algoriphagus, for which the name Algoriphagus machipongonensis sp. nov. is proposed.