68-2

57/110

sequencing using a MiSeq Next generation sequencer (Illumina, Inc. San Diego, CA, USA). The genome of IDSA1 was also subjected to PacBio RS II (Pacific Biosciences, Menlo Park, CA, USA) to determine the complete genome. Total reads of 883.6 Mbp (285x coverage) were assembled with HGAP 3.08). Following complete genome sequencing, the 16S ribosomal RNA gene sequence of IDSA1 was compared with identical sequences in the data-bases. Average nucleotide identity (ANI) analysis was performed9, 10) using an ANI calculator11) on the EZbiocloud homepage (https:// www.ezbio-cloud.net/tools/ani), which confirmed that the IDSA strains were S. aureus. Coding sequences (CDS) and ribosomal and transfer RNA genes were extracted, along with their initial functional assignments, using the RASTtk algorithm12) avail-able on the RAST annotation server13). The CDS annotations were confirmed by visual comparisons with known gene products on In Silico Molecular Cloning (IMC) software (In Silico Biology, Inc., Kanagawa, Japan), which assists in evaluating the prevalence of the annotated sequence by compar-ison of each CDS with its homologues registered in databases. IMC software was also used for circular genome display and comparative analyses of the IDSA1 genome with the genome of S. aureus strain TW20. The sequence and annotation data of IDSA1 have been registered in the databases under acces-sion number AP025249 .Determination of minimum inhibitory concentra-tions of IDSA strains against antibiotics. Minimum inhibitory concentrations (MICs) of the MRSA strains against various antimicrobial reagents were determined by the microbroth dilution method according to Clinical and Laboratory Standards Institute (CLSI) guidelines14).Computer-based molecular epidemiology and other genome analyses. MLST3) was determined by depositing the whole genome sequences of IDSA strains in the Center for Genomic Epidemiology (CGE) website (http://www.genomicepidemiology.org). S. aureus virulence genes were identified using VirulenceFinder 2.015) of the CGE server with thresholds of 90% nucleotide sequence identity and 60% minimum length. Drug resistance genes were initially identified by ResFinder 3.216, 17) on the CGE server, followed by one-to-one visual inspection of annotated genes. The phylogenetic relationships among sequenced S. aureus strains were analyzed by CSI Phylogeny 1.418) on the CGE server, a method based on single nucleotide polymorphisms (SNPs) among genomes, allowing a maximum-likelihood phylogenetic tree to be drawn18). Based on the results of analysis of CSI phylogeny, and using a NEWICK-format file, a tree was re-drawn as a radial layout with a root centered by FigTree ver. 1.4.4 software (http://tree.bio.ed.ac.uk/, kindly gifted by Professor Andrew Rambaut of the Insti-tute of Evolutionary Biology at the University of Edinburgh). The IMC software described above was used to assess GC contents, GC-skewing and dot plots to identify homologous regions in two genomes. Types of SCCmec were determined by SCCmecFinder 1.219, 20) on the CGE server.Drug resistance, virulence and other genetic features of the IDSA strains. Analysis of the MICs of the 10 selected IDSA strains showed that these strains were multidrug-resistant (Table 1). Whole genome sequencing with a next-generation sequencer showed that these strains consisted of several tens of contigs, allowing molecular epidemiological anal-yses. Sequencing of their 16S ribosomal genes and determination of their average nucleotide identity (ANI) showed that all of these strains were S. aureus. These strains varied in their drug resis-tance determinants (Table 2). The presence of genes confirming resistance to β-lactams and aminoglycosides was consistent with the deter-mined MICs tested; whereas others were not (Table 1). This may have been due to the involve-ment of functionally-uncharacterized genes that confer antibiotic resistance to these strains. The strains also showed variations in their possessions of virulence-related genes (Table 3). Variations in drug resistance genes as well as in virulence factors do not always reflect whole genome-wide similarities, because drug-resistance and virulence-related genes are often encoded on mobile genetic elements and may therefore migrate among strains, in contrast to their chromosomes. These findings could not exclude the possibility that the MRSA strains were closely related based on their genetic background. The whole-genome wide diversities of the MRSA strains were assessed by analyzing their MLST. This analysis allows 149Results