Robertkochia solimangrovi sp. nov., isolated from mangrove soil, and emended description of the genus Robertkochia

To date, there is sparse information for the genus Robertkochia with Robertkochia marina CC-AMO-30DT as the only described member. We report here a new species isolated from mangrove soil collected at Malaysia Tanjung Piai National Park and perform polyphasic characterization to determine its taxonomic position. Strain CL23T is a Gram-negative, yellow-pigmented, strictly aerobic, catalase-positive and oxidase-positive bacterium. The optimal growth conditions were determined to be at pH 7.0, 30–37 °C and in 1–2 % (w/v) NaCl. The major respiratory quinone was menaquinone-6 (MK-6) and the highly abundant polar lipids were four unidentified lipids, a phosphatidylethanolamine and two unidentified aminolipids. The 16S rRNA gene similarity between strain CL23T and R. marina CC-AMO-30DT is 96.67 %. Strain CL23T and R. marina CC-AMO-30DT clustered together and were distinguished from taxa of closely related genera in 16S rRNA gene phylogenetic analysis. Genome sequencing revealed that strain CL23T has a genome size of 4.4 Mbp and a G+C content of 40.72 mol%. Overall genome related indexes including digital DNA–DNA hybridization value and average nucleotide identity are 17.70 % and approximately 70%, below the cutoffs of 70 and 95%, respectively, indicated that strain CL23T is a distinct species from R. marina CC-AMO-30DT. Collectively, based on the phenotypic, chemotaxonomic, phylogenetic and genomic evidences presented here, strain CL23T is proposed to represent a new species with the name Robertkochia solimangrovi sp. nov. (KCTC 72252T=LMG 31418T). An emended description of the genus Robertkochia is also proposed.

Flavobacteriaceae is one of the widely spread bacterial families and is composed of 158 genera at the time of writing [1]. The genus Robertkochia was introduced by Hameed et al. in 2014 [2] as one of the new genera in the family Flavobacteriaceae. Until now, the genus consisted of a single species Robertkochia marina CC-AMO-30D T , which was isolated from surface seawater Collected at Taichung harbour, Taiwan [2]. The species was described as Gram-negative, strictly aerobic, orange-pigmented and with iso-C 15 : 0 , iso-C 15 : 1 G and iso-C 17 : 0 3-OH as predominant fatty acids. The report for Robertkochia is scarce as the previous study only focused on taxonomic assignment with one species reported so far [2]. Furthermore, the genome of this genus and its prospective applications have not been studied or reported.
Robertkochia and many other members of the Flavobacteriaceae are halophilic or halotolerant bacteria that reside in diverse saline environments such as seawater, mangrove forest and marine sediment [3][4][5]. Mangroves are inter-tidal wetlands that connect terrestrial and marine ecosystems [6]. Due to periodic tidal flats, drastic changes in salinity and nutrient availability of the mangrove environment make it a unique ecosystem [7]. Free-living and symbiotic bacteria in such environments were found to play essential roles in maintaining mangrove ecosystem by, for example, recycling organic matter and biotransformation of minerals [8][9][10]. It was estimated that less than 5 % of species in mangrove environments have been described so far [11]. Therefore, it could be considered as an interesting area to be explored.
In the present study, strain CL23 T was isolated from soil obtained from a mangrove forest located at Tanjung Piai National Park, Johor, Malaysia. This strain was characterized using a polyphasic approach (phenotypic, chemotaxonomic and genomic aspects) following the recommended guidelines [12,13] and new criteria for classification [14] to elucidate its taxonomic position. The results indicated that strain CL23 T represents a new species within the genus Robertkochia and the name Robertkochia solimangrovi sp. nov. is proposed.

ISOlATION AND hOME hAbITAT
Soil from the mangrove forest was sampled at Tanjung
Phylogenetic trees of the 16S rRNA genes were reconstructed by using the neighbour-joining (NJ) [16] and maximumlikelihood (ML) [17] algorithms using mega 7.0 software [18] based on 1000 bootstrap replications [19] and Kimura's two-parameter model. The results of the 16S rRNA gene phylogenetic analysis (Fig. 1) demonstrated that strain CL23 T and R. marina CC-AMO-30D T formed a clade in the NJ and ML trees, confirming the placement of strain CL23 T within the genus Robertkochia. The high bootstrap value at the node separating the branch containing strain CL23 T and R. marina CC-AMO-30D T in 16S rRNA gene phylogenetic tree supported that these two strains are distinct from each other.

PhENOTyPIC AND ChEMOTAXONOMIC ChARACTERIzATION
Colony morphology was observed on MA at 30 °C after 48 h incubation. Gram-staining was performed according to the protocol described previously [20]. Malachite green staining was used to assess the presence of endospore in 7 days old cultures [21]. Gram-stain reaction and endospore formation were examined under light microscope (Nikon Eclipse E200). Cell morphology was examined under scanning electron microscope (SEM; JSM-IT300LV, jeol). Bacterial motility was investigated by following the hanging-drop approach [22]. The presence of flexirubin-type pigment was determined by flooding the cells with 20 % (w/v) KOH [12].
Catalase activity was detected by effervescence using 3 % (v/v) H 2 O 2 while oxidase activity was determined by oxidation of tetramethyl-p-phenylenediamine. Hydrolysis of starch, casein, ʟ-tyrosine, hypoxanthine, xanthine, Tween 20, Tween 40, Tween 60, Tween 80, carboxymethyl-cellulose (CMC) and xylan were tested according to Smibert and Krieg [21]. Bile aesculin hydrolysis was investigated using the method of Facklam and Moody [23]. Other biochemical characteristics were revealed by using API 20 E and API 20 NE kits (bioMérieux). Carbohydrate utilization and enzyme activity profiles of both strains were investigated by using API 50 CHB and API ZYM kits (bioMérieux), respectively. All API assays were carried out by following the manufacturer's instructions with the slight modification that inoculation was supplemented up to 2 % (w/v) NaCl.
Strain CL23 T was determined as a Gram-negative, rodshaped, non-spore-forming, oxidase-positive and catalasepositive bacterium with motile ability by gliding. The colony was in a circular form with 0.5-1.0 mm diameter, a smooth surface, convex elevation, entire margin and had translucent property on MA after 48 h incubation. Under SEM, cells of strain CL23 T were 0.2-0.4 µm wide and 2.3-3.2 µm long. The notable distinctive features that differentiate strain CL23 T from R. marina CC-AMO-30D T are shown in Table 1. In terms of morphology, strain CL23 T is yellow-pigmented while R. marina CC-AMO-30D T was found to be orangepigmented. Strain CL23 T grew well in 15-42 °C, pH 5-9 and 0-9 % (w/v) NaCl, and in general strain CL23 T demonstrated a broader growth range compared to R. marina CC-AMO-30D T ( Table 1). The optimal growth conditions of strain CL23 T were observed at 30-37 °C, pH 7 and 1-2 % (w/v) NaCl. Strain CL23 T was also able to produce acetoin, β-galactosidase and test results were weakly positive toward amygdalin according to the API 20 E assay, but not for R. marina CC-AMO-30D T . Based on API ZYM assay results, strain CL23 T was able to produce α-galactosidase, β-galactosidase and α-mannosidase, which were absent in R. marina CC-AMO-30D T . Both strains were further distinguished by the hydrolysis capability of gelatin, Tween 20, Tween 40 and Tween 60, and exhibiting resistance towards ampicillin, penicillin G, piperacillin and bacitracin (Table 1).
For the chemotaxonomic analysis, cellular fatty acids were extracted following the protocol of Microbial Identification System (midi, version 6.1) [26]. Biomass of strain CL23 T and its reference strain R. marina CC-AMO-30D T were harvested from MA after 48 h of incubation at 30 °C. The cells were saponified with a methanolic base, then the resulting sodium salts of fatty acids were methylated. In the final step, methyl esters were transferred to the organic phase and washed. Fatty acid methyl esters were analysed on Agilent 6890 apparatus equipped with an Ultra-2 capillary column and subsequently identified in the RTSBA6 library. As exhibited in Table 2, the predominant cellular fatty acids of strain CL23 T and R. marina CC-AMO-30D T were found to be iso-C 15 : 0 , iso-C 15 : 1 G and iso-C 17 : 0 3-OH (>10 %). Nonetheless, some fatty acid patterns and abundance of strain CL23 T varied when compared to R. marina CC-AMO-30D T , such as summed features 3 (3.64 %) and 9 (5.24 %) were detected in strain CL23 T but none for R. marina CC-AMO-30D T . In addition, the amounts of iso-C 16 : 0 , anteiso-C 15 : 0 and iso-C 16 : 0 3-OH in strain CL23 T were remarkably lower than those in R. marina CC-AMO-30D T ( Table 2).

Colony pigmentation Yellow Orange
Oxidase activity + -  the respiratory quinones were extracted by solvent methanol : hexane (2 : 1 v/v), separated by TLC and HPLC following the standard method by Tindall [27]. The polar lipids were extracted using chloroform : methanol solvent and separated by two-dimensional silica gel TLC [28]. Total lipid material was identified using molybdatophosphoric acid and specific functional groups were determined using spray reagents specific for defined functional groups.
The major respiratory quinone of strain CL23 T was identified to be menaquinone-6 (MK-6), which matched R. marina [2] and other members in the family Flavobacteriaceae [12].

gENOMIC ChARACTERIzATION
The genome of reference strain R. marina CC-AMO-30D T was not available at the time of study, therefore, both the genomes of strain CL23 T (NCBI accession: QKWN00000000) and R. marina CC-AMO-30D T (NCBI accession: QXMP00000000) were sequenced in this study. Whole genome sequencing of strain CL23 T was accomplished on an Illumina HiSeq 2500 platform (2×150 bp). The raw reads were filtered, and the quality data was de novo assembled using SOAPdenovo 2.04 [29]. The resulting genome was annotated using the ncbi Prokaryotic Genome Annotation Pipeline (PGAP) [30].
The assembled genome of strain CL23 T , consisting of 23 contigs with 322× depth of sequencing coverage (average), made up the size of genome with 4 407 290 bp and a G+C content of 40.72 mol%. The genome size of strain CL23 T is significantly larger than that of R. marina CC-AMO-30D T (3 571 649 bp). The G+C content of strain CL23 T is slightly lower than that of R. marina CC-AMO-30D T (43.67 mol%). Based on pgap annotation, a total of 3669 protein-coding genes was found in the genome of strain CL23 T . The genes responsible for phosphatase activity were found in the genome of strain CL23 T and R. marina CC-AMO-30D T with a total of 12 and 7 phosphatases encoded, respectively (Table S1). This correlated to the API ZYM results in which both strains were positive to acidic and alkali phosphatases. Notably, the number of phosphatases annotated is higher in strain CL23 T as compared to R. marina CC-AMO-30D T . On the other hand, strain CL23 T consists of a series of genes for assimilatory sulfate reduction into sulfite (sulfate adenylyltransferase subunit CysN and CysD, adenylylsulfate kinase and phosphoadenylylsulfate reductase) and then sulfite reduction into sulfide (FAD-binding oxidoreductase and LLM class flavindependent oxidoreductase) (Table S1). Nevertheless, the genes responsible for reduction of sulfite to sulfide are absent in R. marina CC-AMO-30D T (Table S1). Furthermore, strain CL23 T also encodes a set of genes for reduction of nitrate to ammonia (NirBD and NrfAH) in which NirBD genes were not found in genome of R. marina CC-AMO-30D T (Table S1). *Summed features are groups of two or three fatty acids that cannot be separated by GLC with the midi system. †Summed feature 3 consisted of iso-C 15 : 0 2-OH, C 16 : 1 ω6c and/ or C 16 : 1 ω7c and annotated here as iso-C 15 : 0 2-OH based on the equivalent chain length (ECL). ‡Summed feature 9 consisted of iso-C 17 : 1 ω9c and/or C 16 : 0 10-methyl.
These genes suggest that strain CL23 T participates in nutrient recycling in mangrove environments.
Multilocus sequence analysis (MLSA) was conducted on five housekeeping genes of strain CL23 T , R. marina CC-AMO-30D T and related genera, which the sequences were retrieved from genome data. The sequences of housekeeping genes were aligned individually and then concatenated in the following order: rpoB-gyrB-recA-mutL-atpD. The phylogenetic tree of concatenated housekeeping genes was reconstructed using mega 7.0 similarly as described above. In this tree (Fig. 2), strain CL23 T and R. marina CC-AMO-30D T are clustered together but well distinguished from each other with high level of support (>90 % bootstrap value). Likewise, the phylogenetic tree based on whole genome sequences that was built using realphy 1.12 [31] also supported the finding that both strain CL23 T and R. marina CC-AMO-30D T are grouped within the same clade (Fig. S2).
To further underpin the classification of strain CL23 T as representing a new species, the overall genome related indexes (OGRIs) were determined. Average nucleotide identity based on blast (ANIb) was calculated using JSpeciesWS [32]. ANI based on usearch (OrthoANIu) was determined by using ChunLab's online ANI calculator [33]. The digital DNA-DNA hybridization (dDDH) value was calculated by using the Genome-to-Genome Distance Calculator [34].
The ANIb and OrthoANIu values between strain CL23 T and R. marina CC-AMO-30D T were 69.35 and 70.47% respectively. These ANI values are below the recommended threshold of 95-96 % for species delineation [35]. Similarly, the dDDH value between two strains was found to be 17.70%, lower than 70%, the cut-off for species boundaries [34]. Combining the interpretation of ANI and dDDH values, the result revealed the identity of strain CL23 T as a distinct species within the same genus as R. marina CC-AMO-30D T .

EMENDED DESCRIPTION Of ThE gENuS RobeRtkochia hAMEED et al. 2014
The characteristics of the genus Robertkochia are described according to Hameed et al. 2014 [2] with following amendments and additional information. Oxidase is either positive or negative and catalase is positive. The DNA G+C content of the type strain of type species is 43.67 mol% based on genome data. The Whole Genome Shotgun project of type strain of type species is available at EMBL/DDBJ/GenBank under accession QXMP00000000. The version described in this paper is QXMP01000000.