Complete genome sequence of Kocuria rhizophila BT304, isolated from the small intestine of castrated beef cattle

The whole genome sequencing of K. rhizophila BT304 was performed using PacBio RS II (20-kb SMRTbell™ templates) and Illumina HiSeq4000 (TruSeq DNA PCR-Free 350-bp library) strategies. For obtaining 20-kb library, genomic DNA was sheared with g-TUBE (Covaris) and purified by AMPure PB magnetic beads (Beckman Coulter). The sequencing library for Illumina HiSeq4000 was prepared by random fragmentation of the DNA sample, followed by 5′ and 3′ adapter ligation. After filtering, we obtained 165,979 sub-reads with mean length of 8036-bp from the PacBio system. From the Illumina data set (34,363,358 raw reads), we obtained 12,131,322 filtered reads (quality score over 20 (Q20) = 99.12%). The reads were assembled using RS HGAP Assembly version 3.0 and polished with Quiver. The assembly of K. rhizophila BT304 was annotated using the RAST prokaryotic genome annotation server (http://rast.nmpdr.org/) [8], and the Position-Specific Iterative BLAST (PSI-BLAST) against the eggNOG version 4.5 database [9]. RNAmmer 1.2 [10] and tRNAscan-SE 1.21 [11] were used to identify rRNA and tRNA sequences, respectively. Prophage insert regions were detected with an on-line phage search tool, PHASTER [12].

We constructed a phylogenetic tree based on 16S rRNA gene sequences. Using the multiple sequence alignment program CLUSTAL W [13], the 16S rRNA gene sequences of the BT304 and related species were aligned. Phylogenetic trees were constructed using the following algorithms: neighbor-joining [14], maximum-parsimony [15] and maximum-likelihood [16] algorithms based on 1000 bootstrap replications in MEGA version 7 [17].

We performed comparative genomic analysis on nine K. rhizophila strains: DC2201, FDAARGOS 302, G2, D2, 14ASP, P7-4, TPW45, RF, and UMD0131. The reference genome sequences were downloaded from the NCBI genome database (http://www.ncbi.nlm.nih.gov/genome/). Average nucleotide identity (ANI) based on USEARCH (OrthoANIu) was used to assess overall similarity between two genome sequences [18]. Functional genes in each genome were predicted and annotated using the SEED subsystem in the RAST server. For whole genome comparison, the genome of K. rhizophila BT304 and other K. rhizophila genomes were aligned using the progressive MAUVE algorithm in the MAUVE multiple genome alignment software 2.4.0 [19].

A single colony of the K. rhizophila strain BT304 was repeatedly transferred to fresh BHI medium to obtain pure cultures. Before DNA extraction, the identity of the strain was verified through 16S rRNA gene sequencing. After the genome sequence was obtained, we confirmed the identity of the strain through a BLAST search of the 16S rRNA gene found in the genome.

The genome coverage of the K. rhizophila strain BT304 was 329 fold, and the complete genome sequence consisted of a single circular chromosome of 2,763,150 bp. The genome contained 2359 coding sequences, 51 tRNA genes, and 9 rRNA genes. The circular map of the genome is displayed in Fig. 1, and detailed genomic features are listed in Table 1. GC content of the complete genome was 71.2%. The GC content and size of the BT304 chromosome was similar to those of other K. rhizophila strains, including K. rhizophila DC2201 (2,697,540 bp, 71.2% GC) [2], K. rhizophila P7-4 (2,820,331 bp, 70.5% GC) [4], and K. rhizophila TPW45 (2,701,701 bp, 70.6% GC) [3] (Table 2). The results of RAST annotation showed that the genome contained 2427 coding sequences with 55 RNA genes. In the SEED subsystem distribution, “amino acids and derivatives” (345 ORFs), “carbohydrates” (252 ORFs), “protein metabolism” (206 ORFs), and “cofactors, vitamins, prosthetic groups, pigments” (206 ORFs) were the most abundant categories (Fig. 2). Similar to the RAST annotation results, annotation of the BT304 genome using the eggNOG database revealed “amino acid transport and metabolism” (203 ORFs) as the most abundant category, followed by “replication, recombination and repair” (148 ORFs), and “translation, ribosomal structure and biogenesis” (145 ORFs) (Additional file 1: Table S1).

Bild vergrößern

Bild vergrößern

Based on 16S rRNA gene sequences, we constructed a phylogenetic tree to identify the phylogeny of the K. rhizophila strain BT304 within the family Micrococcaceae (Additional file 1: Figure S1). We compared the genome of K. rhizophila BT304 with those of nine other K. rhizophila strains. Analysis of the OrthoANI values revealed that the genome of strain BT304 is closest to that of K. rhizophila DC2201 (98.92% OrthoANI), followed by FDAARGOS 302 (98.80%), and G2 (97.89%) (Table 2). Between two genomes, an ANI value higher than 95–96% is regarded as the same species [20]. Accordingly, the strain BT304 was confirmed to be a species of K. rhizophila. Interestingly, however, the genome of BT304 showed much less genomic relatedness with several other K. rhizophila strains: D2 (88.14%), 14ASP (87.97%), P7-4 (87.96%), TPW45 (87.92%), RF (87.90%), and UMD0131 (87.83%). We next conducted whole genome comparison of the genome of K. rhizophila BT304 with other publically available K. rhizophila genomes based on the MAUVE multiple genome alignment. The locally collinear blocks between the BT304 and DC2201 genomes were highly homologous, whereas considerable differences in genomic regions were observed between the BT304 and other K. rhizophila genomes (Fig. 3). These results imply that each K. rhizophila strain can adapt to its respective ecological niche by gaining and/or losing genomic elements.

Bild vergrößern

Because several strains belonging to the species K. rhizophila have been isolated from clinical specimens (i.e., human blood) [6, 7], we evaluated virulence-related factors from the genome of strain BT304 and other publically available K. rhizophila genomes. Based on comparison with the SEED database, we found 28–46 CDSs annotated as the “virulence, disease and defense” category from the K. rhizophila genomes (Table 2). The BT304 strain possessed the lowest number of virulence-related genes (28 CDSs), whereas the 14ASP strain (isolated from soil) contained the highest number of virulence-related genes (46 CDSs). To determine whether a degree of virulence-related factors observed from the strain BT304 is suitable for probiotic applications, we accessed the number of virulence-related genes in genomes of eight commercially available probiotic strains [21]. The probiotic strains mainly composed of the genera Bifidobacterium and Lactobacillus contained 22–47 CDSs assigned to the “virulence, disease and defense” category (Additional file 1: Table S2).

We next assessed antibiotic resistant potential in the K. rhizophila genomes. Genes related to the sub-category “resistance to fluoroquinolones” were found in the genomes of all K. rhizophila strains. Several strains (e.g., BT304, DC2201, FDAARGOS 302, G2, and RF) contained genes related to “resistance to vancomycin”. Only the UMB0131 strain isolated from human urine possessed genes related to “multidrug resistance efflux pumps”. Collectively, our comparative analysis suggested that the antibiotic properties of K. rhizophila vary among the strains.

To identify prophage contamination in the genome of K. rhizophila BT304, we conducted PHAST phage search analysis. The genome contained four incomplete prophages at positions 876,510–884,942, 1,118,199–1,125,394, 1,367,325–1,389,936, and 2,021,003–2,029,604 bp (Additional file 1: Table S3). PHAST analysis in other K. rhizophila strains further revealed that the number of prophages varies among the strains. Strains with an ANI of over 97% with the strain BT304 were predicted to have 1–3 incomplete prophages, whereas the remaining strains contained no prophage. The strain P7-4, isolated from the fish gut, was the only exception, containing one incomplete prophage.

The BT304 strain was isolated from the bovine small intestine where dietary nutrients (e.g., amino acids, lipids, and carbohydrates) are metabolized by abundant microbiota [22, 23]. As described above, the BT304 strain possessed a relatively small genome, but many genes related to amino acid metabolism (Fig. 2 and Additional file 1: Table S1). Based on comparison with the SEED database, we found a relatively similar number of genes (318–356 CDSs) annotated as the “amino acid and derivatives” category across the K. rhizophila genomes (Table 2), suggesting that K. rhizophila can utilize dietary amino acids, and might be an indigenous member of the small intestinal gut microbiota in beef cattle.

In mammals, including human and mouse, branched chain amino acids (BCAAs) are responsible for lipogenesis in adipocytes [24] and lipid accumulation in skeletal muscles [25]. Therefore, accumulation of BCAAs is regarded as a key risk factor leading to obesity. However, these metabolic aspects of BCAAs would be beneficial in economical feeding of breeding animals. Annotating the genome of BT304 using the SEED database revealed genes related to both BCAA biosynthesis (20 CDSs) and degradation (52 CDSs) (Table 2). Comparative genomic analysis showed that other reference K. rhizophila genomes also contain more than 70 genes related to BCAA metabolism.