Dactylosporangium luridum sp. nov., Dactylosporangium luteum sp. nov. and Dactylosporangium salmoneum sp. nov., nom. rev., isolated from soil

The genus Dactylosporangium was proposed by Thiemann et al. (1967) to accommodate aerobic, filamentous actinomycetes that released motile spores formed in sporangia borne on short sporangiophores carried on the substrate mycelium. At the time of writing, the genus comprises six recognized species: Dactylosporangium aurantiacum (the type species; Thiemann et al., 1967), D. fulvum (Shomura et al., 1986), D. matsuzakiense (Shomura et al., 1980), D. roseum (Shomura et al., 1985), D. thailandense (Thiemann et al., 1967) and D. vinaceum (Shomura et al., 1983); two additional taxa, ‘Dactylosporangium salmoneum’ and ‘D. variesporum’, have been cited as species incertae sedis (Vobis, 1989). Dactylosporangiae form a distinct phyletic branch in the Micromonosporaceae 16S rRNA gene sequence tree (Koch et al., 1996; Thawai et al., 2008), and can be distinguished from other genera classified in the family Micromonosporaceae by using a combination of chemotaxonomic, morphological and phylogenetic data (Vobis, 2006; Ara et al., 2008). In contrast, members of Dactylosporangium species show minimal differences in chemical, morphological and physiological properties and are distinguished mainly on the basis of diffusible pigment and substrate mycelial colours (Vobis, 1989, 2006).

The primary aim of the present investigation was to determine the taxonomic status of a group of filamentous soil actinomycetes that had colonial and morphological properties characteristic of members of the genus Dactylosporangium. To this end, the isolates were compared with the type strains of recognized Dactylosporangium species and with ‘D. salmoneum’ NRRL B-16294 in a polyphasic taxonomic study. The resultant genotypic and phenotypic data showed that the isolates represented two novel Dactylosporangium species. It is also proposed that ‘D. salmoneum’ NRRL B-16294 be recognized as representing a novel species.

Actinomycetes were isolated on plates of Streptomyces isolation medium (per litre distilled water: 0.4 g casein, 1.0 g starch, 0.1 g CaCO₃, 0.2 g KH₂PO₄, 0.5 g KNO₃, 0.1 g MgSO₄, 15 g agar), supplemented with cycloheximide (20 μg ml⁻¹) and oxytetracycline (20 μg ml⁻¹), following inoculation with suspensions of soil taken from Palace Leas meadow hay plot 6 (Atalan et al., 2000) at Cockle Park Experimental Farm, Northumberland, UK (National Grid Reference NZ 200913). Forty isolates that produced orange substrate mycelium characteristic of Dactylosporangium strains were purified and maintained on oatmeal agar (ISP medium 3; Shirling & Gottlieb, 1966) at room temperature and as suspensions of mycelial fragments in glycerol (20 %, v/v) at −20 °C. rep-PCR, based on the use of BOX-, ERIC- and REP-PCR primers, was used to group and select isolates for the subsequent polyphasic study. Isolate BK63^T, ‘D. salmoneum’ NRRL B-16294 and each of the Dactylosporangium marker strains were clearly separated from one another and from the remaining isolates on the basis of their rep-PCR profiles.

Thirteen strains were taken to represent the taxonomic variation shown by the isolates based on rep-PCR profiles. These isolates, together with the type strains of recognized Dactylosporangium species and ‘D. salmoneum’ NRRL B-16194, were grown in shake flasks of modified Bennett’s broth (Jones, 1949) at 28 °C for 14 days at 150 r.p.m. to obtain biomass for molecular systematic studies. The cultures were checked for purity and harvested by centrifugation. Biomass preparations were washed in NaCl/EDTA buffer (0.1 M EDTA, pH 8.0, 0.1 M NaCl) and stored at −20 °C until needed.

The phylogenetic positions of the strains were determined by 16S rRNA gene sequence analysis. Genomic DNA, PCR and direct sequencing of the purified products were carried out as described by Tan et al. (2006). The resultant, almost-complete 16S rRNA gene sequences (1430–1439 nt) were aligned manually, by using the jphydit program (Jeon et al., 2005), against corresponding sequences of representatives of genera classified in the family Micromonosporaceae, retrieved from the GenBank database. Phylogenetic trees were inferred by using the neighbour-joining (Saitou & Nei, 1987) and maximum-parsimony (Fitch, 1971) algorithms from the mega version 3 program (Kumar et al., 2004) and the maximum-likelihood method (Felsenstein, 1981) from the phylip suite of programs (Felsenstein, 1993). The evolutionary distance model of Jukes & Cantor (1969) was used to generate evolutionary distance matrices for the neighbour-joining algorithm. The topologies of the resultant trees were evaluated in a bootstrap analysis (Felsenstein, 1985) based on 1000 resamplings of the neighbour-joining dataset by using the consense and seqboot options from the phylip package.

In the 16S rRNA gene sequence tree of representative members of the Micromonosporaceae, the new isolates, ‘D. salmoneum’ NRRL B-16294 and the type strains of recognized Dactylosporangium species formed a well-delineated clade (data not shown). The members of this taxon were most closely related to representatives of the genus Virgisporangium, a relationship that was supported by all of the tree-making algorithms (Fig. 1⇓). All but one of the isolates had identical 16S rRNA gene sequences and formed a distinct subclade together with the type strain of D. aurantiacum; they shared 99.4 % 16S rRNA gene sequence similarity with the latter, a value that corresponded to 9 nucleotide differences over 1435 locations. Relatively high 16S rRNA gene sequence similarities were shown with the type strains of D. fulvum (98.1 %), D. matsuzakiense (98.3 %), D. roseum (98.2 %), D. thailandense (98.5 %) and D. vinaceum (98.2 %), as well as with ‘D. salmoneum’ NRRL B-16294 (98.2 %). The remaining isolate, strain BK63^T, also shared highest 16S rRNA gene sequence similarity with D. aurantiacum NRRL B-8018^T (98.9 %; a value equivalent to 16 nucleotide differences across 1436 sites). The new isolates together with the type strain of D. aurantiacum formed a subclade that was supported by all of the tree-making algorithms.

A second subclade in the 16S rRNA gene sequence tree contained ‘D. salmoneum’ NRRL B-16294 and the type strains of D. matsuzakiense and D. vinaceum. The close relationship between this taxon, the D. matsuzakiense subclade and the D. aurantiacum subclade was underpinned by all of the tree-making algorithms. ‘D. salmoneum’ NRRL B-16294 and the type strains of D. matsuzakiense and D. vinaceum shared 16S rRNA gene sequence similarities of 98.9 and 98.8 %, respectively, values that corresponded to 16 and 17 nucleotide differences over 1432 locations. D. matsuzakiense DSM 43810^T and D. vinaceum DSM 43823^T shared 99.6 % 16S rRNA gene similarity, a value equivalent to 6 nucleotide differences over 1468 locations. D. fulvum DSM 43917^T and D. roseum DSM 43916^T, the two most closely related type strains in the genus, exhibited 99.7 % 16S rRNA gene sequence similarity, equivalent to 5 nucleotide differences across 1471 sites. These strains, together with D. thailandense DSM 43158^T, formed a well-delineated taxon, the D. thailandense subclade, the integrity of which was underscored by all of the tree-making algorithms and by a bootstrap value of 94 %.

Phylogenetic analyses based on partial RNA polymerase β-subunit (rpoB) gene sequences have provided valuable data in polyphasic studies designed to clarify relationships within and between genera of actinomycetes (Kim et al., 1999, 2004; Goodfellow et al., 2007). In the present study, partial rpoB gene sequences generated for four representative isolates, strains BK51^T, BK53, BK63^T and BK69, ‘D. salmoneum’ NRRL B-16294 and the type strains of recognized Dactylosporangium species were compared with corresponding genes of members of genera classified in the family Micromonosporaceae, drawn from GenBank, by using the same tree-making algorithms as above, but with Streptomyces coelicolor A3(2) as the outgroup.

Isolates BK51^T, BK53 and BK69 had identical rpoB gene sequences and formed a subclade with D. aurantiacum NRRL B-8018^T (Fig. 2⇓); the integrity of this taxon was underpinned by all of the tree-making algorithms and by a bootstrap value of 95 %. The remaining isolate, strain BK63^T, formed a distinct, single-membered subclade at the periphery of the Dactylosporangium rpoB gene sequence tree. ‘D. salmoneum’ NRRL B-16294 also formed a distinct subclade that was associated with D. roseum DSM 43916^T. These results are in good agreement with those from the 16S rRNA gene sequence analyses as they show that the isolates fall into one distinct multi-membered and one single-membered clade, although discrepancies are apparent between the topologies of the two trees (see Figs 1⇑ and 2⇓).

DNA–DNA relatedness studies were carried out to determine the finer taxonomic relationships between isolates BK51^T and BK63^T and ‘D. salmoneum’ NRRL B-16294 and their closest phylogenetic neighbours (Table 1⇓). Levels of DNA–DNA relatedness between the tested strains were established by measuring the divergence between the thermal denaturation midpoints of homologous and heterologous DNA (ΔT_m) following the procedure developed by Gonzalez & Saiz-Jimenez (2005). Confidence can be placed in the resultant data as comparable results have been obtained when the same strains have featured in thermal denaturation and spectrophotometric studies (Jurado et al., 2005; Goodfellow et al., 2007). ΔT_m values of 5.0 and 6.0 °C correspond to DNA–DNA relatedness levels of 60 and 70 %, respectively (Rosselló-Mora & Amann, 2001). DNA G+C contents were determined for all of the strains included in the DNA–DNA experiments by using the procedure described by Gonzalez & Saiz-Jimenez (2005). The strains had DNA G+C contents within the range 70–74 mol% (see Table 4⇓).

Isolate BK51^T showed very low ΔT_m values (0.2–1.6 °C) with isolates BK53 and BK69 (Table 1⇑), values well below the cut-off point recommended for the delineation of genomic species (ΔT_m >5.0 °C; Wayne et al., 1987). In contrast, isolate BK51^T showed ΔT_m levels well above this cut-off point with isolate BK63^T, ‘D. salmoneum’ NRRL B-16294 and the type strains of D. aurantiacum, D. matsuzakiense and D. vinaceum. Isolate BK63^T also exhibited relatively high ΔT_m values (5.8–6.2 °C) with the type strains of related Dactylosporangium species and with ‘D. salmoneum’ NRRL B-16294, indicating that it belongs to a distinct genomic species. ‘D. salmoneum’ NRRL B-16294 could also be assigned to a novel genomic species as it exhibited high ΔT_m values (5.8–7.5 °C) with its nearest neighbours, namely D. matsuzakiense NRRL B-16293^T and D. vinaceum NRRL B-16297^T.

Biomass for chemotaxonomic studies was prepared by growing representative isolates BK51^T and BK63^T, ‘D. salmoneum’ NRRL B-16294 and the type strains of recognized Dactylosporangium species in modified Bennett’s broth at 150 r.p.m. for 21 days at 28 °C; cells were harvested by centrifugation, washed in distilled water, recentrifuged and freeze-dried. Standard procedures were used to extract and analyse the isomers of diaminopimelic acid (Staneck & Roberts, 1974), isoprenoid quinones (Collins, 1994), muramic acid type (Uchida et al., 1999), polar lipids (Minnikin et al., 1984) and whole-organism sugars (Schaal, 1985), with appropriate controls. Cellular fatty acids were extracted, methylated and analysed by GC by using the standard Sherlock MIDI (Microbial Identification) system (Sasser, 1990) and mycolic acids were identified by using the TLC procedure introduced by Minnikin et al. (1975).

All of the organisms contained mixtures of 3-hydroxy and meso-diaminopimelic acid, N-glycolylmuramic acid, hexahydrogenated and octahydrogenated menaquinones with nine isoprene units [MK-9(H₆) and MK-9(H₈)] as predominant menaquinones and arabinose, galactose, glucose, mannose and xylose in whole-organism hydrolysates; mycolic acids were not found. The isolates contained iso-C_15 : 0 and iso-C_16 : 0 as predominant fatty acids with varying kinds and proportions of minor compounds and diphosphatidylglycerol, phosphatidylethanolamine, phosphatidylinositol and phosphatidylglycerol as major polar lipids (phospholipid pattern type 2 sensu Lechevalier et al., 1977) with a discontinuous distribution of phosphatidylinositol mannosides, an unknown aminolipid and an unknown phospholipid (Table 2⇓). These results are in good agreement with those of members of the genus Dactylosporangium (Kroppenstedt, 1985; Vobis, 2006).

Isolates BK51^T and BK63^T and ‘D. salmoneum’ NRRL B-16294 were grown on tryptone-yeast extract, yeast extract-malt extract, oatmeal, inorganic salts-starch, glycerol-asparagine, peptone-yeast extract-iron and tyrosine agars (ISP media 1–7, respectively; Shirling & Gottlieb, 1966) for 21 days at 28 °C. Colonies growing on these media were examined by eye to determine substrate mycelial pigmentation and the colour of any diffusible pigments; colours were recorded by using National Bureau of Standards (NBS) Colour Name Charts (Kelly, 1958; NBS, 1964). Peptone-yeast extract-iron and tyrosine agar plates were examined to determine whether the strains produced melanin pigments. The strains were also examined to establish whether they grew on Luria–Bertani, nutrient and trypticase soy agars (Sambrook et al., 1989) after incubation at 28 °C for 3 weeks.

All of the strains grew well on inorganic salts-starch, oatmeal and tyrosine agar plates, producing a range of substrate mycelial colours but relatively few diffusible pigments (Table 3⇓). The type strains of D. roseum, D. thailandense and D. vinaceum together with ‘D. salmoneum’ NRRL B-16294 grew well on all of the media except glycerol-asparagine agar. In contrast, isolates BK51^T and BK63^T grew poorly on glycerol-asparagine, peptone-yeast extract-iron, tryptone-yeast extract and yeast extract-malt extract agar plates. In general, few of the organisms produced soluble pigments on the agar media, although D. vinaceum NRRL B-16297^T formed soluble pigments on all of these media. It is also interesting that isolates BK51^T and BK63^T could be distinguished by the colour of the substrate mycelium that they produced on the various media.

Isolates BK51^T and BK63^T together with ‘D. salmoneum’ NRRL B-16294 were examined for micromorphological properties on 3-week-old glycerol-asparagine, oatmeal, tyrosine and yeast extract-malt extract agar plates by using a Nikon Optiphot light microscope fitted with a long-working-distance objective. In addition, gold-coated, dehydrated preparations from the plates were examined by using a Cambridge Stereoscan 240 scanning electron microscope following the procedure described by O’Donnell et al. (1993). Preparations from oatmeal agar plates were flooded with sterile water, left for 60 min and then examined for the presence of motile spores with a light microscope. All of the strains formed irregular branched substrate hyphae (0.3–0.7 μm in diameter), which penetrated the agar. Similarly, all three strains formed sporangia, which released motile spores. Motile spores were formed in finger-like, short, narrow sporangia produced directly on the substrate mycelium (Fig. 3⇓). They also formed globose bodies with smooth surfaces at the top of short sporophores on substrate hyphae.

Isolates BK51^T, BK53, BK63^T and BK69, ‘D. salmoneum’ NRRL B-16294 and the type strains of recognized Dactylosporangium species were examined for a broad range of phenotypic properties by using the media and methods described by Goodfellow et al. (1997). All of the strains were positive for catalase, degraded aesculin and xylan, utilized glycogen, (−)-d-glucose and (−)-sucrose as sole carbon sources, and were sensitive to novobiocin (8 μg ml⁻¹) and rifampicin (16 μg ml⁻¹). All were negative for nitrite reduction, did not degrade adenine, cellulose, chitin, guanine, hypoxanthine, pectin, tributyrin, Tween 20, uric acid or xanthine, did not use (+)-l-lactic acid, (−)-l-sorbose, oxalic acid or (−)-d-ribose as sole carbon sources and did not grow in the presence of 3 % (w/v) NaCl or 0.05 % (w/v) lysozyme.

Isolates BK51^T, BK53 and BK69 had identical phenotypic profiles that served to distinguish them from isolate BK63^T, ‘D. salmoneum’ NRRL B-16294 and the type strains of recognized Dactylosporangium species (Table 4⇓). Indeed, these three isolates were the only strains to use (+)-d-arabitol as a sole carbon source and to be unable to produce H₂S or degrade Tween 80. Similarly, unlike most of the other strains, they utilized adonitol, (+)-l-arabinose, dextrin, (−)-d-fructose and (+)-melibiose as sole carbon sources, but were unable to hydrolyse gelatin or to grow at 37 °C. Isolate BK63^T and ‘D. salmoneum’ NRRL B-16294 could also readily be separated from one another and from the Dactylosporangium type strains, notably by the inability of strain BK63^T to metabolize most of the sole carbon sources and by the activity of ‘D. salmoneum’ NRRL B-16294 in the degradation tests.

The genotypic and phenotypic data clearly indicate that isolates BK51^T, BK53 and BK69, isolate BK63^T and ‘D. salmoneum’ NRRL B-16294 represent three new centres of taxonomic variation in the genus Dactylosporangium. It is therefore proposed that these strains be classified in the genus Dactylosporangium as representing Dactylosporangium luteum sp. nov., Dactylosporangium luridum sp. nov. and Dactylosporangium salmoneum, respectively. The genotypic data acquired on isolates BK50, BK52, BK54–BK57, BK60, BK67 and BK68 are consistent with their assignment to D. luteum.