Bacteriophage diversity of Panton-Valentine leukocidine positive clinical Staphylococcus aureus isolates identified in South Africa and Nigeria

Abstract

The Panton-Valentine leukocidine (PVL) gene, encoding for a cytolytic toxin was previously considered unique to community-associated methicillin-resistant Staphylococcus aureus (CAMRSA) strains. This bi-component toxin is conferred by the lukS-PV and lukF-PV genes, residing in the genome of icosahedral-or-elongated head shaped bacteriophage types. Through horizontal gene transfer (HGT), the PVL gene has been spread among healthcare-associated methicillin-resistant Staphylococcus aureus (HA-MRSA) and methicillin-susceptible S. aureus strains (MSSA). Thirteen PVL-encoding bacteriophage types have been reported, however, only ten have been described using an extended III Part PCR-based typing scheme. Panton-Valentine leukocidine-positive S. aureus strains have been reported in South Africa and Nigeria, however, there is limited data on the bacteriophage types circulating in these countries. The PVL-encoding bacteriophage types were investigated, following the PVL-positive screening multiplex-PCR (M-PCR) assay. Part I of the PCR-based typing scheme grouped the bacteriophage types into one of three morphology head groups. All three morphology groups were detected with the icosahedral head group II morphology, predominating among the isolates from South Africa and Nigeria. In some isolates more than one bacteriophage morphology head group was observed. The second part of the PCR-based typing scheme linked the specific morphology tail genes to the lukS/F-PV genes, to discriminate between PVL-encoding and non-PVL encoding bacteriophage types. However, this assay was not able to distinguish between the PVL-encoding and non-PVL encoding bacteriophage types. This was a limitation of the second part (Part II) of the PCR-based typing scheme. Part III of the PCR-based typing scheme identified the specific PVL-encoding bacteriophage types and detected seven of the ten bacteriophage types. Three bacteriophage types; ΦSa2USA-like, ΦPVL-like and Φ108PVL-like were predominant in South Africa and ΦSa2USA-like, Φ108PVL-like and ΦPVL-like were predominant in Nigeria. Nineteen percent (13/70) of the S. aureus isolates were non-typeable with this scheme. Transmission electron microscopy (TEM) showed bacteriophage types belonging to icosahedral head group I and the elongated group. No icosahedral head group II bacteriophage types were observed. The TEM showed multiple bacteriophage types and bacteriophage tails in one of the representative isolates, confirming that S. aureus strains may harbour more than one bacteriophage type. Whole-genome sequencing (WGS) identified seven novel bacteriophage types (including four PVL-encoding and three non-PVL encoding) and partial sequences of the lukS/F-PV genes in all the isolates. Non-PVL-encoding bacteriophage types carried the lukE/D genes, with 91% and 94% sequence homology to the lukS/F-PV genes, indicating that the screening M-PCR assay also detected the lukE/D genes. The clinical significance of these genes has not yet been reported but may explain the debate regarding the role of the PVL toxin in S. aureus infections. The detection of partial lukS/F-PV sequences may explain why patients with these PVL-positive S. aureus infections show a less severe clinical picture compared to what has been reported in the literature. The PCR-assays used in this study were valuable screening and typing tools but showed several limitations. Whole-genome sequencing provided important and new data regarding novel bacteriophages and the truncated sequences of the PVL gene. The current PCR-based typing scheme can be improved with the inclusion of this new data.