The Population Structure and Mobile Genetic Elements Harboured by Hospital-Adapted Vancomycin-Resistant and Linezolid-Resistant Enterococcus faecium and Enterococcus faecalis in Ireland Investigated by Whole-Genome Sequencing

Vancomycin-resistant enterococci (VRE) are a major cause of healthcare-associated infection worldwide. For over a decade, Ireland has consistently had some of the highest rates of invasive VRE infections in Europe. VRE harbouring the vanA operon encoding vancomycin resistance, which is often located on plasmids and associated with the Tn1546 transposon, are the primary causative genotype. Linezolid is an antibiotic used to treat infections caused by multi-drug resistant (MDR) Gram-positive bacteria, such as VRE. Linezolid-resistant enterococci (LRE) have been reported with increasing frequency in recent years, with a recent rise in LRE harbouring transferable linezolid resistance, encoded by the optrA, poxtA and/or cfr genes. Little is known about the population structure of hospital-adapted VRE in Ireland or about the prevalence and genetic organisation of mobile genetic elements encoding linezolid resistance. The overall objectives of this study were to investigate the population structure of hospital-adapted E. faecium screening and bloodstream infection isolates (BSIs) from patients in a large acute hospital in Dublin, Ireland and to investigate the prevalence and genetic organisation of transferable genetic elements encoding linezolid resistance genes in LRE from patients in multiple Irish hospitals. A combination of short-read (Illumina) and long-read (Oxford Nanopore Technologies [ONT]) whole-genome sequencing (WGS) technology was used as the principal approach in this study. The first aim was to examine the population structure of hospital-adapted E. faecium in a large Dublin hospital. Between June 2017-July 2019, 365 E. faecium isolates were collected, including vancomycin-resistant E. faecium (VREfm) screening isolates (n=286) and all E. faecium bloodstream infection isolates (VREfm, n=45, vancomycin-susceptible E. faecium [VSEfm], n=34) recovered during the study period. All isolates underwent Illumina WGS and were typed using conventional multilocus sequence typing (MLST) and high-resolution core-genome (cg) MLST. The vast majority of isolates (360/365, 98.6%) belonged to the hospital-adapted clade A1 and 57.5% of clade A1 E. faecium isolates belonged to ST80 (207/360) based on MLST. However, cgMLST divided the 360 clade A1 isolates into 33 clusters and 63 singletons, with an inter-cluster allelic difference range of 25-1201. The overall population was polyclonal, with highly related isolates (≤ 20 allelic differences) identified in multiple hospital wards and persisting over extended periods of time (up to 21 months). The identification of highly related VREfm and VSEfm in the isolate collection indicated that VSEfm can readily acquire vanA, resulting in new VREfm. The vanA transposon regions of three selected VREfm isolates from representative STs, were resolved using hybrid assembly from isolates SJ10 (ST789, SJ10vanA), SJ11 (non-typeable) and BSI_SJ40 (ST80). In addition, two vanA-encoding plasmids were also resolved using hybrid assembly from SJ82 (ST203; 48,934 bp) and SJ245 (ST117; 40,559 bp). The vanA region in Irish isolates differed from the prototype Tn1546 vanA transposon by multiple insertions of IS1216E, differing orientations of the vanA operon genes. The SJ10vanA sequence was used as a reference against which all VREfm (n=331) isolates were compared; 95.5% (316/331) of isolates harboured a vanA region with >90% sequence identity to the SJ10vanA reference. The repetitive insertion of IS1216E in the vanA operon observed in Irish VREfm is also a likely source of instability within the operon. The second aim was to investigate the prevalence of transferable linezolid resistance genes in LRE recovered from patients in Irish hospitals. One-hundred and fifty-four LRE recovered from patients in 14 Irish hospitals between June 2016-August 2019 were screened for the optrA, poxtA and cfr genes by PCR. All isolates harbouring at least one of these genes, and 20 without, underwent Illumina WGS. The optrA and/or poxtA genes were identified in 35/154 (22.7%) isolates, the highest prevalence of transferable linezolid resistance reported to date. Fifteen isolates with diverse STs harboured optrA only; one E. faecium isolate harboured optrA (chromosome) and poxtA (plasmid). Seven E. faecalis and one E. faecium harboured optrA on a 36,331 bp plasmid with 100% identity to the previously described optrA-encoding conjugative plasmid pE349. Variations around optrA were also observed, with optrA located on plasmids in five isolates and within the chromosome in three isolates. Nine E. faecium and 10 E. faecalis harboured poxtA, flanked by IS1216E, within an identical 4,001 bp region found on plasmids exhibiting 72.9% 100% sequence coverage to a 21,849 bp conjugative plasmid (pM16/0594) encoding poxtA. Enterococcus faecalis isolates primarily belonged to ST480, whereas E. faecium isolates belonged to diverse STs. Twenty representative isolates without linezolid resistance genes investigated all harboured the G2576T 23S RNA gene mutation associated with linezolid resistance. High prevalence of optrA and poxtA in diverse enterococcal lineages found in Irish hospitals indicated the presence of significant selective pressure(s) for maintenance. The third aim was to investigate a hospital outbreak of linezolid-resistant VREfm (LVREfm) using WGS. The outbreak occurred in a second Dublin hospital in October 2019. Thirty-nine VREfm from patient screening (19 isolates, 17 patients) and environmental sites (20 isolates) were investigated. Isolates were screened for optrA, poxtA and cfr by PCR and underwent Illumina WGS. One LVREfm underwent hybrid assembly to resolve an optrA-encoding plasmid. Twenty isolates (51.3%) were LVREfm and optrA-positive, including the LVREfm from the index patient. A closely related cluster of 28 ST80 isolates was identified using cgMLST, including all 20 LVREfm and eight linezolid-susceptible VREfm (including a VREfm from the index patient recovered in October 2018 a year before the outbreak), with an average allelic difference of two, indicating an outbreak. Nineteen (95%) LVREfm harboured a 56,684 bp conjugative plasmid, termed pEfmO_03. Presence of pEfmO_03 in LVREfm and the close relatedness of the outbreak cluster isolates indicated the spread of a single clone. The inclusion of the October 2018 VREfm, provided evidence this clone had persisted within the index patient s gastrointestinal tract for at least one year prior to the acquisition of pEfmO_03 and the subsequent hospital outbreak. The outbreak was terminated by rapid implementation of enhanced infection prevention and control (IPC).WGS was central in confirming/understanding an outbreak of ST80 LVREfm. VRE, and more recently LRE, have become a persistent and increasing problem in hospitals in Ireland and urgent action is required to understand their true burden. Ireland is disadvantaged by its sub-optimal hospital infrastructure, especially a paucity of single rooms. Updating current IPC guidelines should be considered to reduce VRE/LRE rates. Implementing admission screening of all patients for VRE/LRE, increasing the frequency of and improving current cleaning regimes, improving hand hygiene compliance in healthcare and improved antimicrobial stewardship both in healthcare and agriculture would contribute significantly to managing this problem. The main goal should be to halt the further spread of VRE/LRE in Ireland. Additionally, national and international surveillance of VRE/LRE is critical for controlling their spread. WGS for typing has been vital in understanding VRE/LRE population structure. Finally, without action increasing levels of both VRE and LRE could lead to an increase in the transfer of antimicrobial resistance genes to other Gram-positive organisms, such as S. aureus and/or MRSA