Molecular Typing of Invasive Staphylococcus aureus from the Emerging Infections Program (EIP) Using Whole-Genome Sequencing

Background: The CDC has performed surveillance for invasive Staphylococcus aureus (iSA) infections through the Emerging Infections Program (EIP) since 2004. SCC mec and spa typing for clonal complex (CC) assignment and genomic markers have been used to characterize isolates. In 2019, whole-genome sequencing (WGS) of isolates began, allowing for high-resolution assessment of genomic diversity. Here, we evaluate the reliability of SCC mec typing, spa typing, and CC assignment using WGS data compared to traditional methods to ensure that backwards compatibility is maintained. Methods: S. aureus isolates were obtained from a convenience sample of iSA cases reported through the EIP surveillance system. Overall, 78 iSA isolates with diverse s pa repeat patterns, CCs, SCC mec types, and antimicrobial susceptibility profiles were sequenced (MiSeq, Illumina). Real-time PCR and Sanger sequencing were used as the SCC mec and spa typing reference methods, respectively. spa- MLST mapping (Ridom SpaServer) served as the reference method for CC assignment. WGS assembly and multilocus sequence typing (MLST) were performed using the CDC QuAISAR-H pipeline. WGS-based MLST CCs were assigned using eBURST and SCC mec types using SCC mec Finder. spa types were assigned from WGS assemblies using BioNumerics. For isolate subtyping, previously published and validated canonical single-nucleotide polymorphisms (canSNPs) as well as the presence of the Panton-Valentine leukocidin (PVL) toxin and arginine catabolic mobile element (ACME) virulence factor were assessed for all genome assemblies. Results: All isolates were assigned WGS-based s pa types, which were 100% concordant (78 of 78) with Sanger-based spa typing. SCC mec Finder assigned 91% of isolates (71 of 78) SCC mec types, which were 100% concordant with reference method results. Also, 7 isolates had multiple cassettes predicted or an incomplete SCC mec region assembly. Using WGS data, 96% (75 of 78) of isolates were assigned CCs; 3 isolates had unknown sequence types that were single-locus variants of established sequence types. Overall, 70 isolates had CCs assigned by the reference method; 100% (70 of 70) concordance was observed with WGS-based CCs. Analysis of canSNPs placed 42% (33 of 78) of isolates into CC8, with 17 (52%) of these isolates classified as USA300. PVL and ACME were not accurate markers for inferring the USA300 subtype as 24% (4 of 17) of isolates did not contain these markers. Conclusions: S. aureus CCs, SC