Molecular epidemiology of MRSA isolates
The distribution of SCCmec types and spa types among various STs of MRSA isolates is shown in Table 1. SCCmec V (42%) was the predominant gene followed by SCCmec III (27%), SCCmec IV (13%), SCCmec I (7%) and SCCmec II (3%). Notably, eighteen isolates had multiple SCCmec types and the most common combination was SCCmec III with SCCmec V. MLST analysis revealed ten different CC and three singletons (ST616, ST1598, ST1947) (Table 1). Three major CCs were CC1 (ST1, ST772), CC22 (ST22, ST217, ST636, ST896, ST1037, ST2371, ST3976) and CC8 (ST8, ST239, ST368, ST630, ST1803, ST3324) which accounted for 82% (n=192) of the isolates (Table 1). Thirty two distinct STs were identified, ST772 was the most common followed by ST22, ST239 and ST368 (Additional file 1: Figure S1). Although, STs were observed to be diverse, most of them were recognised as a SLV or DLV of the dominant STs (Additional file 1: Table S1). Thirty distinct spa types were identified (Table 1), of them t657 (34%) was predominant followed by t425 (18%), t037 (12%), t852 (10%), and t030 (8%). The remaining 25 different spa types were found in fewer than five isolates. Furthermore, PVL gene was identified in 66% (n=153) of the isolates and majority of them belongs to ST772 (34%, n=52) and ST22 (27%, n=42).
Genomic analysis of S. aureus ST772-SCCmec V
A maximum likelihood phylogenetic tree was constructed using 32 MRSA genome (belongs to ST772-SCCmec V) sequenced in this study and 273 S. aureus genomes belongs to ST772 from previous studies (Additional file 1: Table S2). This collection comprised of MRSA (n=270) and MSSA (n=34), which also includes ten basal strains (4 MSSA, 6 MRSA). A genomic comparison was made between the sequenced ST772-SCCmec V MRSA (n=32) from Vellore and a collection of 273 S. aureus belongs to ST772 from 14 countries (including India) (Additional file 1: Figure S2). The earliest ST772 S. aureus isolates (n=7) in the entire global collection were from India and Bangladesh (reported in 2004). The phylogenetic comparison of ST772 S. aureus genome from India against the genomes available from other countries revealed that Indian isolates were distributed across all the previously defined subgroups (Basal strains, ST772-A1, ST772-A2 and ST772-A3) (Additional file 1: Figure S2). As expected, all the ST772-SCCmec V (n=32) MRSA genome sequenced in this study (isolated from the year 2013 onwards) belongs to the successful ST772-A clade (Fig 1). Nearly, half of the Indian isolates are clustered in the dominant subgroup ST772-A2 (56%) followed by ST772-A1 (23%) and ST772-A3 (15%).
Most of the isolates, 71% (n=61) carried SCCmec V (5C2) variant, while the composite SCCmec V (5C2 and 5) was seen in only two isolates (Fig 1). A transposon Tn4001 carrying aminoglycoside resistance gene (aacA-aphD) in SCCmec V was noticed in 73% (n=63) of the genomes. Furthermore, IRP was found in all the isolates, except four basal strains.
Virulence and antimicrobial resistance
There is no significant variation in the distribution of virulence genes among subgroups (Fig 2). A 59.4 kb Φ IND772 prophage carrying PVL-operon (lukS and lukF) and a staphylococcal enterotoxin A (sea) was noticed in all isolates, except two. In addition, three different genomic islands were identified in all isolates (including basal strains); a 41 kb, vSa-alpha genomic island containing an exotoxin cluster (set6, set7, set8, set10, set11, set12, set13, set14, set15), a 25.2 kb vSa-beta genomic island containing an enterotoxin gene cluster (sec, seg, sei, sem, sen, seo), and a 24.6 kb, vSa-gamma genomic island contains an extracellular fibrinogen binding protein (ecb), a fibrinogen binding protein (fbp), an alpha hemolysin (hla) and an exfoliative toxin A (eta).
Screening of AMR genes revealed variable gene distribution and QRDR mutations within and across subgroups (Fig 2). A pattern of increasing antimicrobial resistance was noticed in the dominant and emerging subgroups. Among AMR genes, the penicillin resistance encoding blaZ gene was identified in all the isolates. Furthermore, methicillin resistance encoding mecA gene was distributed across subgroups with the exception of basal strains (3 isolates) and ST772-A1 (14 isolates). In addition, aminoglycoside resistant are predominantly seen in isolates of dominant and emerging subgroups. The macrolide resistance encoding genes mph (C) and msr(A) were seen across the subgroups (ST772-A1, A2, A3), except basal strains. Analysis of quinolone resistance mechanism revealed S84L in gyrA with S80Y in grlA were the most frequent double serine mutations (Fig 2). In the QRDR region, other mutations such as S84L in gyrA , S80Y in grlA and S84L in gyrA with S80F in grlA were also noticed.
Bayesian time scaled phylogenetic analysis
The root to tip analysis revealed a strong correlation (correlation coefficient 0.6862 and R squared 0.4709) between the time of isolation and distance from root suggesting the temporal clock like evolution in the lineage (Additional file 1: Figure S3). Moreover, several demographic models were tested on the Indian ST772 S. aureus genomes (n=86) and compared their marginal likelihood estimation. This revealed that skyline model with strict clock was the best supported model (Additional file 1: Table S3). The mean substitution rate was estimated to be 1.22 X10-6 substitution per site per year. Based on this rate, the most recent common ancestor (MRCA) of ST772 S. aureus was estimated as 1964 (95% highest posterior density [HPD], 1956-1970). Further, we then dated the acqusition and loss of various mobile genetic elements associated with antimicrobial resistance as well as QRDR mutations. This analysis suggests that IRP and SCCmec V (5C2) were gained in 1991 (95% HPD, 1988-1993). Bayesian hierarchical clustering using core SNPs segregated S. aureus ST772 isolates that had IRP and a variant of SCCmec V (5C2) into separate subgroups (Fig 3). The double serine mutation (S84L in gyrA with S80F in grlA) was predicted to be replaced by S84L with S80Y in ~1998. This infers that the ST772 S. aureus multi-resistant subgroups are likely expanded rapidly after 1991 and has since sustained.