Streptomyces asenjonii sp. nov., isolated from hyper-arid Atacama Desert soils and emended description of Streptomyces viridosporus Pridham et al. 1958

A polyphasic study was undertaken to establish the taxonomic status of Streptomyces strains isolated from hyper-arid Atacama Desert soils. Analysis of the 16S rRNA gene sequences of the isolates showed that they formed a well-defined lineage that was loosely associated with the type strains of several Streptomyces species. Multi-locus sequence analysis based on five housekeeping gene alleles showed that the strains form a homogeneous taxon that is closely related to the type strains of Streptomyces ghanaensis and Streptomyces viridosporus. Representative isolates were shown to have chemotaxonomic and morphological properties consistent with their classification in the genus Streptomyces. The isolates have many phenotypic features in common, some of which distinguish them from S. ghanaensis NRRL B-12104T, their near phylogenetic neighbour. On the basis of these genotypic and phenotypic data it is proposed that the isolates be recognised as a new species within the genus Streptomyces, named Streptomyces asenjonii sp. nov. The type strain of the species is KNN35.1bT (NCIMB 15082T = NRRL B-65050T). Some of the isolates, including the type strain, showed antibacterial activity in standard plug assays. In addition, MLSA, average nucleotide identity and phenotypic data show that the type strains of S. ghanaensis and S. viridosporus belong to the same species. Consequently, it is proposed that the former be recognised as a heterotypic synonym of the latter and an emended description is given for S. viridosporus. Electronic supplementary material The online version of this article (doi:10.1007/s10482-017-0886-7) contains supplementary material, which is available to authorized users.


Introduction
The prospect of isolating novel filamentous actinobacteria that synthesise new specialised metabolites is enhanced when bioprospecting strategies are focused on neglected and unexplored habitats (Hong et al. 2009;Tiwari and Gupta 2012;Guo et al. 2015), including desert soils (Meklat et al. 2011;Boubetra et al. 2013). The most extensive surveys of culturable actinobacterial diversity in desert biomes have been concentrated on sites in the Atacama Desert in northern Chile, the driest non-polar desert on the planet (Bull and Asenjo 2013;Bull et al. 2016). The application of a taxonomic approach to drug discovery (Goodfellow and Fiedler 2010) has been effective in the isolation of putatively novel filamentous actinobacteria from Atacama Desert habitats, some of which produce novel natural products Wichner et al. 2016). Indeed, polyphasic taxonomic studies on dereplicated actinobacteria isolated from hyper-arid and extreme hyper-arid Atacama Desert soils have led to the description of novel species of Lechevalieria (Okoro et al. 2010), Lentzea (Idris et al. 2017a) and Modestobacter (Busarakam et al. 2016a) and to the detection of rare thermophilic Amycolatopsis species (Busarakam et al. 2016b). In addition, several new Streptomyces species have been described (Santhanam et al. 2012a(Santhanam et al. , b, 2013Idris et al. 2017b), one of which, Streptomyces leeuwenhoekii (Busarakam et al. 2014), encompasses strains that synthesise novel antibiotics (Nachtigall et al. 2011;Rateb et al. 2011a, b) and chaxapeptin, a new lasso peptide (Elsayed et al. 2015).
The present study was designed to establish the taxonomic position of several closely related Atacama Desert streptomycetes. These strains were the subject of a polyphasic taxonomic study which showed that they belong to a new species, Streptomyces asenjonii sp. nov.

Materials and methods
Isolation, maintenance and cultivation of strains Isolates KNN6.11a, KNN35.1b T , KNN35.2b, KNN48.3e and KNN83.e were recovered from a hyper-arid soil collected in 2012 by one of us (ATB) from the Chaxa de Laguna, Salar de Atacama near Tocanão (23°17 0 33 00 S, 68°10 0 99 00 W at 2219 m above sea level), using the dilution plate procedure described by Okoro et al. (2009). The strains were isolated on Gauze's No.1 agar (KNN6.11a) (Zakharova et al. 2003), humic acid-vitamin agar (KNN35.1b T , KNN35.2b) (Hayakawa and Nonomura 1987) and SM1 agar (KNN48.3e, KNN83.e) (Tan et al. 2006) following incubation for 14 days at 28°C. Similarly, the final strain, KNN42.f, was isolated from a starchcasein agar plate (Küster and Williams 1964) following inoculation with a suspension of an extreme hyperarid soil collected by ATB in 2010 from the Yungay core region of the Atacama Desert (24°06 0 18.6 00 S, 70°01 0 55.6 00 W at 1016 m asl). These strains, together with Streptomyces ghanaensis NRRL B12104 T (Wallhäuser et al. 1965), were maintained on yeast extractmalt extract agar (International Streptomyces Project [ISP2] medium., Shirling and Gottlieb 1966) and as suspensions of spores and hyphal fragments in 20%, v/v glycerol at -20 and -80°C. Biomass samples for most of the chemotaxonomic analyses and for the 16S rRNA gene sequencing studies were prepared in shake flasks (180 revolutions per minute) of ISP 2 broth after incubation at 28°C for 14 days and washed twice in distilled water. Cells for the chemotaxonomic analyses were freeze-dried and those for the sequencing studies stored at room temperature. Biomass preparations for the fatty acid analyses were harvested from shake flasks of Tryptic Soy broth (Difco) following incubation at 28°C for 7 days.
Phylogenetic analysis 16S rRNA gene sequencing. Genomic DNA extraction, PCR-mediated amplification of 16S rRNA genes and purification of the resultant products were carried out on all of the isolates using the procedures described by Kim and Goodfellow (2002). Identification of phylogenetic neighbours and calculation of pairwise 16S rRNA gene sequence similarities were achieved using the EzTaxon-e server (http://www. ezbiocloud.net/taxonomy; Yoon et al. 2017) and the resultant sequences aligned using the CLUSTAL W algorithm from the MEGA 6 software package (Tamura et al. 2013). Phylogenetic analyses using the maximum-likelihood (ML) (Felsenstein 1981) and maximum-parsimony (MP) algorithms (Fitch 1971) were also realised using the GGDC web server (Meier-Kolthoff et al. 2013a) of the DSMZ phylogenomics pipeline (Meier-Kolthoff et al. 2014) adapted to single genes available at http://ggdc.dsmz.de/. ML and MP trees were inferred from the alignment with RAxML (Stamatakis 2014) and TNT (Goloboff et al. 2008), respectively. The topologies of the resultant trees were evaluated by bootstrap analyses (Felsenstein 1985) based on 1000 replicates used in conjunction with treebisection-and-reconnection branch swapping and ten additional random sequence replicates for MP and rapid bootstrapping in conjunction with the auto MRE bootstopping criterion (Pattengale et al. 2010) for ML. The trees were rooted using the 16S rRNA gene sequence of Streptomyces albus subspecies albus DSM 40317 T (GenBank accession number AJ621602). The V 2 test implemented in PAUP* (Swofford 2002) was used to check for compositional bias. Pairwise sequence similarities were calculated using the method recommended by Meier-Kolthoff et al. (2013b) for 16S rRNA genes and a multiple sequence alignment was created with MUSCLE (Edgar 2004).
Multi-locus sequence analysis. Genomic DNA extracted from each of the isolates following growth in ISP2 broth at 28°C was purified, as described by Idris et al. (2017a). The housekeeping genes used in previous analyses on streptomycetes (Busarakam et al. 2014;Labeda et al. 2017;Idris et al. 2017b;Labeda 2016), namely atpD (ATP synthase F1, beta subunit), gyrB (DNA gyrB subunit), rpoB (RNA polymerase beta subunit), recA (recombinase A) and trpB (tryptophane B, beta subunit), were amplified, sequenced, purified, deposited in the GenBank database and organised using the Bacterial Isolate Genome Sequence Database BIGSdb version 1.15.4 on the ARS Microbial Genome Sequence Database server (http: //199.133.98.43). The sequences of the protein loci of the strains were aligned with one another and with those of their close neighbours and phylogenetic relationships established using the ML algorithm after Idris et al. (2017a). Pairwise distances between the sequences of each locus were established using the Kimura two-parameter model (Kimura 1980). Strain pairs having MLSA evolutionary distances B0.007 were considered conspecific based on the cut-off point empirically determined by Huang (2012, 2014), a value that corresponds to the 70% DNA:DNA threshold recommended for the delineation of prokaryotic species by Wayne et al. (1987).

Draft genome preparation and ANI calculations
The draft genome sequence of Streptomyces viridosporus NRRL 2414 T was prepared following the protocol outlined in Labeda et al. (2016) with the exception that CLCbio Genomic Workbench Version 9.5.3 (CLCbio; Boston, MA) was used for contig trimming and de novo assembly. This Whole Genome Shotgun project has been deposited at DDBJ/EMBL/ GenBank under the accession MSGP00000000.

Chemotaxonomy and morphology
Isolates KNN35.1b T and KNN35.2b were examined for spore chain arrangement and spore-surface ornamentation following growth on oatmeal agar (ISP 3 medium; Shirling and Gottlieb 1966) for 14 days at 28°C, by scanning electron microscopy (Cambridge 240 instrument), using the protocol described by O'Donnell et al. (1993). Key chemotaxonomic markers were sought using standard chromatographic procedures. All of the isolates were examined for isomers of diaminopimelic acid (A 2 pm) after Staneck and Roberts (1974). Strains KNN35.1b T and KNN35.2b were analysed for menaquinone, whole cell sugar and polar lipid composition using the procedures described by Collins et al. (1985), Lechevalier and Lechevalier (1970) and Minnikin et al. (1984), respectively. S. ghanaensis NRRL B-12104 T , the close phylogenetic neighbour of the isolates, was included in the sugar and polar lipid analyses. Fatty acids of representative isolates, namely strains KNN35.1b T , KNN35.2b and KNN42.f, and the S. ghanaensis type strain, were extracted, methylated and analysed using the established Sherlock Microbial Identification (MIDI) system and the ACTIN version 6 database (Sasser 1990).

Cultural characteristics
The cultural properties of the isolates were recorded on tryptone-yeast extract, yeast extract-malt extract, inorganic salts-starch, glycerol-asparagine, peptoneyeast extract-iron and tyrosine agar plates (ISP media 1-7, Shirling and Gottlieb 1966) after 14 days at 28°C. Aerial and substrate mycelial colours and those of diffusible pigments were determined by comparison against chips from the Inter-Society Colour Council-National Bureau of Standard Colour charts (Kelly 1964).

Phenotypic tests
The isolates and S. ghanaensis NRRL B-12104 T were examined for standard biochemical, degradative and physiological characteristics after Williams et al. (1983) and enzyme profiles determined using API-ZYM kits (BioMerieux) employing a standardised inoculum corresponding to 5 on the McFarland scale (Murray et al. 1999) and the protocol provided by the manufacturer. The oxidation of carbon sources and resistance to inhibitory compounds were determined using GENIII microplates in an Omnilog device (Biolog Inc., Haywood, USA). The microplates were inoculated with cell suspensions made in a 'gelling' inoculating fluid at a cell density of 98% transmittance with a run time of 7 days in phenotypic microarray mode at 28°C. The exported data were analysed using the opm package for R version 1.0.6. (Vaas et al. 2012(Vaas et al. , 2013. The Biolog tests were carried out in duplicate.

Antibacterial sensitivity assays
Four of the isolates, strains KNN6.11a, KNN35.1b T , KNN35.2b and KNN83.e, were examined for their ability to inhibit the growth of wild type strains of Bacillus subtilis, Escherichia coli, Pseudomonas fluorescens and Staphylococcus aureus using a standard plug assay (Fiedler 2004). The isolates were grown on yeast extract-malt extract sloppy agar (0.8%, w/v agar) for 14 days at 30°C and then plugs were transferred to nutrient agar plates which had been inoculated with 100 ll of the wild type strains grown overnight in lysogeny broth. The inoculated plates were incubated overnight and then examined for the presence of inhibition zones around the agar plugs.

Results and discussion
The six strains isolated from the hyper-arid and extreme hyper-arid Atacama Desert soils were shown to form a well delineated subclade in the Streptomyces 16S rRNA gene tree, a relationship that was supported by all of the tree-making algorithms and by a 78% bootstrap value (Fig. 1). The isolates were found to share 16S rRNA gene sequence similarities within the range 99.85-100%, which corresponds to up to 3 nucleotide (nt) differences at 1373 locations. The strains were seen to be closely related to the type strains of . These data suggest that the Atacama Desert isolates are not particularly closely related to any of their near phylogenetic neighbours in the Streptomyces 16S rRNA gene tree.
The isolates were found to belong to a distinct and homogeneous lineage in the Streptomyces MLSA gene tree based on concatenated partial sequences of the five housekeeping genes, a result supported by a 100% bootstrap value (Fig. 2). The MLSA evolutionary distances between the isolates ranged from \0.000 to 0.001 (Table 1), that is, well within the species level threshold of B0.007 proposed by Huang (2012, 2014). Members of this well delineated taxon were found to be closely related to the type strains of S. ghanaensis DSM 40746 T and S. viridosporus DSM 40243 T (Pridham et al. 1958), albeit with MLSA distances well above the species cut-off point (Table 1). These results provide further evidence of the value of MLSA sequence analyses in clarifying the subgeneric relationships of Streptomyces (Guo et al. 2008;Rong and Huang 2010, 2014Busarakam et al. 2014;Idris et al. 2017b;Labeda et al. 2014Labeda et al. , 2017Labeda 2016). The S. ghanaensis and S. viridosporus strains formed a well-supported subclade in the 16S rRNA gene tree ( Fig. 1) but were not particularly closely related (99.5% sequence similarity, 19 nt differences), results that are clearly more apparent than real given the corresponding MLSA data.
The isolates were shown to form extensively branched substrate mycelia bearing aerial hyphae, to contain LL-A 2 pm as the wall diamino acid and exhibited good growth on all of the ISP media, notably on oatmeal and yeast extract-malt extract agar (Table 2). In general, the substrate mycelia were observed to be grey to yellowish white and the aerial spore mass greyish yellow or bright orange yellow, as were the diffusible pigments. Isolates KNN35.1b T and KNN35.2b were seen to form open spirals of hairy ornamented surfaced spores, as shown in Fig. 3. These isolates and S. ghanaensis NRRL B-12104 T , their close phylogenetic neighbour, were found to have glucose, mannose, ribose and xylose in whole organism hydrolysates, whilst the S. ghanaensis strain was also found to contain galactose. The polar lipid patterns of these strains showed the presence of diphosphatidylglycerol, glycophospholipid, phosphatidylethanolamine, phosphatidylglycerol, phosphatidylinositol and a number of unidentified components, as shown in Figure S1. The predominant  (Table 3). The predominant components in all of these organisms were found to be anteiso-C 15:0 (11.5-17.8%) and iso-C 16:0 (21.3-32.6%); quantitative differences were seen in these and other components while some of the minor fatty acids were discontinuously distributed, as exemplified by the presence of anteiso-C 17:1 and C 17:1 cis9 in the S. ghanaensis type strain and anteiso-C 18:0 amongst the isolates.
Identical results were obtained for nearly all of the duplicated strains included in the phenotypic tests, whilst the exceptions were a few of the carbon source features recorded from the GENIII microplates (Table 4). It can also be seen from Table 4 that the isolates can be distinguished from one another showing that they are not clones. In addition, several properties distinguished all of the isolates from the type strain of S. ghanaensis (Table 4). Thus, only the Atacama isolates produced N-acetyl-b-glucosaminidase, oxidised L-arginine, butyric acid, L-keto-butryric acid, citric acid, D-and L-fucose and D-sorbitol and grew in the presence of 4%, w/v sodium chloride, potassium tellurite and rapamycin SV and at 10°C. In contrast, only S. ghanaensis NRRL B-12104 T oxidised N-acetyl-b-Dmannosamine, N-acetyl-neuraminic acid and D-glucuronic acid. It is also apparent from Table 4 that all of the strains have many phenotypic properties in common.
Isolates KNN35.1b T and KNN35.2b were found to inhibit the growth of the wild type strains of B. subtilis, Fig. 2 Subtree from the Streptomyces phylogenetic tree inferred from concatenated partial sequences of the housekeeping genes atpD, gyrB, recA, rpoB and trpB in IQ-Tree version 1.4.2 (Nguyen et al. 2015) as described by Labeda et al. (2017). Bootstrap values less than 95% were omitted as suggested by the IQ-Tree developers. Bar scale reflects number of substitutions per site It can be concluded that the six isolates from the hyper-arid Atacama Desert soils have identical or almost identical 16S rRNA and MLSA gene sequences and share many phenotypic features in common, some of which distinguish them from the type strain of S. ghanaensis, their close phylogenetic neighbour in the Streptomyces MLSA gene tree generated from the five housekeeping genes. It is, therefore, proposed that the isolates be recognised as a new species within the genus Streptomyces, named Streptomyces asenjonii sp. nov. It seems likely that S. asenjonii strains are common in hyper-arid Atacama Desert soils as additional isolates from the Salar de Atacama sampling site show the same aerial/substrate mycelial and diffusible pigment colours as the current isolates when grown on oatmeal agar. Colour groups such as these have been shown to be reliable indicators of Streptomyces species identity (Antony-Babu et al. 2010;Goodfellow and Fiedler 2010).
It is evident from Table 1 that the type strains of S. ghanaensis and S. viridosporus have a low MLSA distance consistent with their assignment to a single genomic species. Indeed, in their extensive MLSA study of type strains of the family Streptomycetaceae, Labeda et al. (2017) noted that S. ghanaensis NRRL B-12104 T (also ATCC 14672 T ) is a later synonym of S. viridosporus NRRL ISP-5243 T . This observation was confirmed by determination of the ANIm and ANIb average-nucleotide identity values between draft genome sequences of the type strains of these species using the calculate_ani.py script (https://github.com/ widdowquinn/scripts/blog/master/bioinformatics/calc ulate_ani.py), as shown in Table S1. Note that the ANIm percentages between the genome sequences of S. viridosporus NRRL 2414 T , S. viridosporus T7A and S. ghanaensis ATCC 14672 T are[96% and the ANIb percentages between these genomes are [97% which is indicative of species level relatedness (Richter and Rosselló-Móra 2009). Thus, according to Rule 38 of the Bacteriological Code of Nomenclature of Bacteria (Lapage et al. 1992;Parker et al. 2015), S. viridosporus Pridham et al. 1958 has priority over S. ghanaensis Wallhäuser et al. 1965. The type strains of these taxa form spiral chains of spiny to hairy spores (Kämpfer 2012), properties known to be predictive in Streptomyces systematics (Labeda et al. 2012) and have many physiological features in common (Kämpfer et al. 1991). Consequently, on the basis of these observations an emended description is given of Streptomyces viridosporus Pridham et al. (1958).
Description of Streptomyces asenjonii sp. nov.
Streptomyces asenjonii (a.sen.jo'ni.i. N.L. gen. n., asenjonii, named after Juan A. Asenjo in recognition of his promotion of work on Atacama Desert actinobacteria). Aerobic, Gram-positive actinobacteria which form an extensively branched substrate mycelium which carry aerial hyphae that differentiate into open spirals of hairy ornamented spores. Grows from 10 to 50°C, optimally 37°C; from pH 5 to 11, optimally 7.5; and in the presence of up to 5% w/v sodium chloride.
The source of the type strain NRRL ISP-5243 T -= NRRL 2414 T is not known; 'S. ghanaensis' NRRL B-12104 was isolated from soil from Ghana.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.