Whole genome sequence and comparative genomic analysis of multidrug-resistant Staphylococcus capitis subsp. urealyticus strain LNZR-1

The blood samples were collected from a patient with sigmoid colon cancer from the Fourth Affiliated Hospital of Harbin Medical University, in March 2013. S. capitis subsp. urealyticus strain LNZR-1 was isolated after cultivation. It is a Gram positive, coccus-shaped bacterium growing on 5% sheep blood enriched Columbia agar (BioMérieux, Marcyl’Etoile, France) at 37°C. Cell morphology, motility and sporulation were examined by using scanning electron microscopy.

Late log-phase cells were harvested and lysed with EDTA and lysozyme, followed by proteinase K and RNase digestion. Genomic DNA was extracted using a DNeasy Blood & Tissue Kit (Qiagen, Germany) according to the manufacturer’s recommended protocol. Agarose gel (0.7%) electrophoresis was used to evaluate the genomic DNA purity and the concentration was measured using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific, USA). The genomic DNA was stored at -20°C. Strain LNZR-1 was identified by 16S rRNA gene sequencing as described earlier [6]. PCR amplification was performed by using primers 27 F (5′-AGAGTTTGATCCTG GCTCAG-3′) and 1500R (5′-AGAAAGGAGGTGATCCAGGC-3′). Agarose gel (1%) electrophoresis was used to separate amplified PCR fragments which were subjected to sequencing of the 16 s rRNA gene. Phylogenetic analysis was conducted based on the 16S rRNA nucleotide sequence. The representative 16S rRNA nucleotide sequence of strain LNZR-1 was compared against the most recent release of the EzTaxon-e database [10]. Phylogenetic inferences were made using Neighbor-joining method based on Tamura-Nei model within the MEGA 6.06 [11].

Antimicrobial susceptibility was determined by the disk diffusion method on Mueller-Hinton agar recommended by the Clinical and Laboratory Standards Institute guidelines [12]. The following antimicrobial agents were tested: sulfamethoxazole, gentamicin, oxacillin, tetracycline, linezolid, clindamycin, ciprofloxacin, cefoxitin, cefazolin, cefuroxime and vancomycin. The other reference strains used for this study were Escherichia coli ATCC 25922 and Klebsiella pneumoniae ATCC 700603.

The genome of S. capitis subsp. urealyticus strain LNZR-1 was sequenced using a standard run of Illumina HiSeq 2000 sequencing technology which generated paired-end libraries (500-bp insert size) according to the manufacturer’s instructions. Clean reads were assembled into scaffolds using Velvet version 1.2.07 [13], and Post-Assembly Genome Improvement Toolkit (PAGIT) was used to extend the initial contiguous sequences (contigs) and to correct sequencing errors [14]. Open reading frames (ORFs) were identified using Glimmer version 3.0 [15]. Transfer RNAs and ribosomal RNA genes rRNAs were detected by tRNAscan-SE [16] and RNAmmer 1.2 software [17], respectively. The genome was annotated using the RAST (Rapid Annotation using Subsystem Technology) server [18]. The classification of some predicted genes and pathways was analyzed using the Clusters of Orthologous Groups of proteins (COGs) [19] and Kyoto Encyclopedia of Genes and Genomes (KEGG) [20] databases. Functional annotation was also performed by using public database of National Centre for Biotechnology Information (NCBI).

For comparative analysis, we downloaded the reference genome sequences of the closest genetic relatives of strain LNZR-1 and representative strains from the NCBI database: S. capitis C87 (ACRH00000000), S. capitis SK14 (ACFR00000000), S. capitis CR01 (CBUB000000000), S. capitis VCU116 (AFTX00000000) and S. capitis QN1 (AJTH00000000). Mauve in the progressive mode was used for whole genome comparison [21].

Orthology identification was performed by using NCBI blastp 2.2.25+ with default parameters. Then, bidirectional best hits (BBHs) among proteins from different strains were identified. Furthermore, an identity threshold over a given alignment length to define orthologous genes was applied. Score Ratio Values (SRVs) of BBHs between two genes was calculated with the following formula:

Bits (AB) means bits score when using gene A as query while B as database; Bits (BA) means bits score when using gene B as query while A as database; Bits (AA) or Bits (BB) are bits score when using gene A or B to align with itself, respectively. BBHs with SRVs no less than 0.3 was considered as one candidate orthology group (SRVs Table in Additional file 1: Table S1).

Cells of strain LNZR-1 are cocci, 0.7 to 1.2 μm in diameter, occurring predominantly singly or in pairs (Additional file 2: Figure S1). To assess the purity of strain LNZR-1, the 16S rRNA gene sequence of strain LNZR-1 was aligned with sequences of other members of the genus Staphylococcus retrieved from the EzTaxon database. Phylogenetic tree indicated the taxonomic status of strain LNZR-1 clearly classified into the same branch with species S. capitis subsp. urealyticus GTC 727^T (Figure 1).

A total of 2,214,438 available reads were filtered from 5,619,015 raw reads. Quality control was performed with following criteria: reads that contained more than one N bases were removed. Reads that contained more than 50 bases with low quality (Q30) were removed. Reads with more than 10 bases with low quality (Q30) or N bases in the tail of the reads were trimmed. For sequences which lost their mated reads were considered as single reads and were not used in the downstream analysis (Additional file 1: Table S1). 177 initial contigs (best kmer length 57) was assembled by Velvet; then 90 prolonged contigs were assembled based on PAGIT flow (PAGIT is just a flow and it actually does not provide any practically available scripts or programs). The assembled genome of S. capitis subsp. urealyticus revealed a genome size of 2,595,865 bp and a G + C content of 32.67% (90 scaffolds). The largest contig consisted of 319,806 bp and the length of N50 contig was 66,677 bp. These scaffolds contain 2430 coding sequences (CDSs), 8 tRNAs (excluding 1 Pseudo tRNA) and 2 incomplete rRNA operons (1 small subunit rRNA and 1 large subunit rRNA). The properties and the statistics of the genome are summarized in Table 1. RAST server based annotation of the whole genome describes the subsystem distribution of strain LNZR-1 (Figure 2). Genes responsible for amino acids and derivatives (264 ORFs), carbohydrates (209 ORFs), and protein metabolism (175 ORFs) were abundant among the subsystem categories. 1934 genes were categorized into COGs functional groups (including putative or hypothetical genes, Figure 3). For COGs distribution, R (general function prediction only; 427 ORFs), E (amino acid transport and metabolism; 300 ORFs), P (inorganic ion transport and metabolism; 220 ORFs), S (function unknown; 207 ORFs), and G (carbohydrate metabolism and transport; 191 ORFs) were abundant categories (>10% of total COGs matched counts).

The in vitro antibiotic sensitivity tests demonstrated that this strain is susceptible to vancomycin and sulfamethoxazole, and resistant to tetracycline, gentamicin, ampicillin, methicillin, linezolid, clindamycin, cefoxitin, cefazolin, cefuroxime and ciprofloxacin. To gain insights into the genomic basis for the observed antibiotic resistance traits, the genome was searched for specific genes known to confer antibiotic resistance. The results revealed the presence of antibiotic resistance genes conferring resistance against some of the tested antibiotics as well as non-tested antibiotics (Table 2). MDR-type ABC transporters, multidrug and toxin extrusion (MATE) family efflux pumps and multidrug major facilitator superfamily (MFS) transporters were also detected in the genome. Furthermore, 10 putative MarR family transcriptional regulators were found in the genome, which are recognized as a widely conserved group of multiple antibiotic resistance regulators that respond to diverse antibiotics [22].

Whole genome comparison of the LNZR-1 with S. capitis C87, S. capitis SK14, S. capitis CR01, S. capitis VCU116 and S. capitis QN1 showed that some functional regions are highly homologous between the six assemblies (Figure 4). The LNZR-1 genome has high similarity with VCU116 and CR01, although some short stretches present in the genome of VCU116 and CR01 were absent in LNZR-1. Furthermore, LNZR-1, C87, CR01, VCU116 and QN1 revealed a large number of orthologous genes (Figure 5). Venn diagram indicates the presence of a large core-genome. These five S. capitis strains shared 2042 CDS in the genome. A particular overlap between C87 and QN1 became evident; these two chromosomes shared 121 orthologous CDS exclusively. The chromosome of LNZR-1 overlapped less with the C87 and QN1, which shared 5 and 4 exclusively orthologous CDS, respectively. In addition, 244 CDS from the LNZR-1 genome were classified as unique.

This research was approved by the Research Ethics Committee of the Fourth affiliated Hospital of Harbin Medical University, and informed consent was obtained from the patient.