Clonal Complexes Distribution of Staphylococcus aureus Isolates from Clinical Samples from the Caribbean Islands

The aim of this study was to comprehensively characterise S. aureus from the Caribbean Islands of Trinidad and Tobago, and Jamaica. A total of 101 S. aureus/argenteus isolates were collected in 2020, mainly from patients with skin and soft tissue infections. They were characterised by DNA microarray allowing the detection of ca. 170 target genes and assignment to clonal complexes (CC)s and strains. In addition, the in vitro production of Panton–Valentine leukocidin (PVL) was examined by an experimental lateral flow assay. Two isolates were identified as S. argenteus, CC2596. The remaining S. aureus isolates were assigned to 21 CCs. The PVL rate among methicillin-susceptible S. aureus (MSSA) isolates was high (38/101), and 37 of the 38 genotypically positive isolates also yielded positive lateral flow results. The isolate that did not produce PVL was genome-sequenced, and it was shown to have a frameshift mutation in agrC. The high rate of PVL genes can be attributed to the presence of a known local CC8–MSSA clone in Trinidad and Tobago (n = 12) and to CC152–MSSA (n = 15). In contrast to earlier surveys, the USA300 clone was not found, although one MSSA isolate carried the ACME element, probably being a mecA-deficient derivative of this strain. Ten isolates, all from Trinidad and Tobago, were identified as MRSA. The pandemic ST239–MRSA–III strain was still common (n = 7), but five isolates showed a composite SCCmec element not observed elsewhere. Three isolates were sequenced. That showed a group of genes (among others, speG, crzC, and ccrA/B-4) to be linked to its SCC element, as previously found in some CC5– and CC8–MRSA, as well as in S. epidermidis. The other three MRSA belonged to CC22, CC72, and CC88, indicating epidemiological connections to Africa and the Middle East.


Introduction
Staphylococcus aureus, a ubiquitous bacterium, is a common cause of hospital-and community-associated infections in humans around the world [1][2][3][4][5]. Bacterial infections caused by methicillin-resistant S. aureus (MRSA) have become a serious global healthcare problem [6,7]. Aside from resistance, which limits options for treatment, another cause for concern is its continuous evolution and global spread of emerging and/or virulent clones [1]. Indeed, MRSA strains are associated with a number of widespread or even pandemic lineages. To describe these lineages, clonal complexes (CCs) based on multilocus sequence typing (MLST; [8]) are identified. The classification of CCs is based on the sequencing of seven ubiquitous housekeeping genes and the analysis of their sequence In Jamaica, the prevalence of MRSA has continued to increase since the identification of the first case at the University Hospital of the West Indies in 1988 [48]. Reports emanating from Jamaica show that MRSA has fairly remained very low in prevalence, with 4% (2004), 5% (2005), and 7% in 2008 [44,45]. This is even lower at the teaching hospital where of 7304 clinical isolates analysed, 689 were identified as S. aureus, with only 31 (4.5%) proved to be methicillin-resistant in 2017 (unpublished data). This is comparable to the 3% prevalence found in 2013. In Jamaica, commonly identified CCs include ST8-MRSA-IV, USA300, and ST5/ST225-MRSA-II, New York-Japan clone [31].
In general, knowledge of the population structure of S. aureus/MRSA on the Caribbean islands is poor, and changes to it might remain unnoticed. In order to address this, DNA microarray-based analysis was used in this study to classify and assign clinical S. aureus isolates from Trinidad and Tobago, and Jamaica into clonal complexes, strains, and SCCmec types. In addition, the presence and expression of PVL were studied.

Analysis of Clinical Isolates
Just over half of the specimens analysed from Trinidad and Tobago (48/85; 57%) were recovered from wound swabs. The remaining isolates were recovered from blood (12/85; 14%), pus (7/85; 8%), urine (8/85; 9%), ear swabs, vulvar swabs, peritoneal swabs, joint fluid, eye swabs, and, in one case, from an undocumented specimen (10/85; 12%). Most of the Jamaican isolates (13/16; 81%) were from wounds, while the remaining isolates were recovered from a throat swab (1/16; 6%) and blood (2/16; 13%). Tables 1 and 2 show the specimen source distribution of MSSA, MRSA, and S. argenteus isolates from both islands. Table 3 shows the age distribution of patients infected with MSSA, MRSA, and S. argenteus strains from Trinidad and Tobago. Patients aged 30 to 39 years had the highest rate of S. aureus infections (22%; 19/85). Of the 19 patients, 95% of infections were caused by MSSA strains. MRSA accounted for one infection (5%) in this group. The second group of patients that frequently experienced infections were those in the paediatric group (0 to 9 years) and also those between 10 and 19 years of age. Both groups had 18% of patients being infected. In these groups, MRSA accounted for 13% (2/15) and 7% (1/15) of infections, while MSSA accounted for 80% (12/15) and 93% (14/15) of infections, respectively. MSSA infections were found in 92% of patients over the age of 60, while the remaining 8% were related to MRSA. S. argenteus infections were found in two patients, ages 6 and 45 years.   Data from Table 4 showed that males (n = 46) more commonly presented with staphylococcal infections than women. MSSA infections were also most prevalent in both males and females as opposed to MRSA, with 89% (41/46) occurring in males and 81% (30/37) in females. MRSA infections were found in 9% (4/46) of male patients and 16% (6/37) of female patients. S. argenteus isolates accounted for 2% (1/46) of infections in males and 3% (1/37) in females.

Distribution of S. aureus Clonal Complexes and Strains
High clonal diversity was reported from Trinidad and Tobago, as S. aureus isolates were assigned to twenty different clonal complexes and 83 strains comprising both MRSA and MSSA (Table 5). Two additional isolates were identified as S. argenteus, CC2596. The most prevalent CC identified was CC8 (15 isolates), followed by CC152 (14 isolates). Other CCs included CC97, CC6, and CC239, which comprised the majority of MRSA. The remaining CCs were rare, comprising five or fewer isolates each. All 16 isolates (100%) from Jamaica were MSSA and distributed across eight different CCs. CC1 (four isolates) and CC6 (three isolates) were the most preventable among these isolates; CC152 was also found, but CC8 was not identified. However, the numbers are too low to assess the population structure of S. aureus/MSSA on this island.   Table 6 for further explanation.
The overall distribution of CCs from both islands, the various strains identified, as well as results for the resistance and virulence markers, SCCmec elements, and toxins examined can be found in the Supplemental File S1 attached, and summarized in Table 5.

Observations Regarding Individual Clonal Complexes of S. aureus
Four isolates from Jamaica and one from Trinidad and Tobago belonged to CC1. All were MSSA, all lacked the SCC-borne fusidic-acid resistance gene fusC, and all were PVLnegative. All carried the enterotoxin genes seh, sek + seq, and, in addition, all Jamaican isolates were also positive for the toxic shock syndrome toxin gene tst1, as well as for the enterotoxin genes sec and sel.
Three MSSA isolates from Trinidad and Tobago were assigned to CC5. Two of them carried, in addition to PVL, the edinA gene encoding an epidermal cell differentiation inhibitor.
PVL-positive as well as PVL-negative CC6-MSSA were found in Trinidad and Tobago as well as in Jamaica. Six isolates from this CC harboured PVL genes; five were positive for the enterotoxin A gene sea.
CC7 was represented by five MSSA isolates from Trinidad and Tobago. All were PVL-negative but harboured a sea allele as known from N315 (BA000018.3; pos. 2,011,380 to 2,012,153), also known as enterotoxin P or sep.
All CC8 isolates originated from Trinidad and Tobago; not a single one was recovered from Jamaica. Twelve isolates belonged to a previously noticed CC8-MSSA that harboured PVL genes, enterotoxin genes sek and seq, as well as (in 11/12) also sed, sej and ser. Both sequenced isolates showed sed, sej, and ser to be located on the same contig as blaZ + blaI + blaR together with plasmid genes and a cadmium resistance operon. A single isolate out of these twelve genotypically PVL-positive isolates was negative for PVL production, as discussed below. Another two CC8-MSSA isolates were negative for PVL and enterotoxin genes.
CC8-MRSA were not noted, neither "North American" (ACME-positive) nor "South American (ACME-negative/mer-operon-positive) USA300 strains. However, one additional PVL-positive isolate carried ACME genes, the SCC-associated copper resistance gene copA2-SCC, opp3B, opp3C, adhC, and speG, but lacked mecA as well as all genes associated with SCCmec IV. This suggests that it might have been a mecA-deficient derivative of the North American USA300 strain.
A Clonal Complex 22 MRSA isolate was retrieved from a patient presenting with cellulitis/trauma to the left foot. It was also the only isolate assigned to this common and widespread lineage and the only PVL-and tst1-(toxic shock syndrome toxin 1) positive MRSA identified. It harboured various genes associated with SCCmec, which encodes methicillin resistance. These include ugpQ, mecA, the truncated methicillin resistance operon repressor 1, ccrA/B-2, and Q9XB68dcs. Subtyping by an additional array [49] identified the SCCmec element as SCCmec IVa.
A total of three CC72 isolates were identified; however, only one was methicillinresistant. This was a PVL-negative strain with a composite SCCmec VT element comprising ugpQ, mecA, and ccrC genes that also included the fusidic acid resistance gene fusC (=Q6GD50). The probe for D1GU38, a marker that accompanies the second copy of ccrC in SCCmec VT, yielded a signal indicating that the composite element derived from SCCmec VT rather than V. The other two CC72 isolates were MSSA that lacked PVL but carried enterotoxin genes sec and sel.
A CC88-MRSA-IV was the only isolate of this lineage collected from a patient with a vulval abscess. As it carried a SCCmec IVa (MW2-like) element, this observation suggested an African connection. Except for the beta-lactamase operon blaZ/I/R, no other resistance genes were detected. Neither PVL nor enterotoxin or exfoliative toxin genes were detected.
Sixteen isolates (including two from Jamaica) belonged to CC152. All were MSSA, but 13 harboured the penicillinase operon. A total of 15 out of 16 were positive for PVL genes, and these 15 also yielded phenotypically detectable PVL. In addition, all were positive for the edinB gene.
CC239 comprised seven CC239-MRSA-III isolates. The majority (n = 5) of CC239 isolates carried a complex SCC [mec III + speG + Cd/czrC + ccrAB4 + ccrC] element. Three of these isolates were sequenced (see below). Four isolates from this cluster also harboured a mercury resistance operon that, however, in two sequenced isolates, was plasmid-borne rather than SCC-associated.
The other two CC239 isolates carried SCC [III + SCCmer + ccrC] elements but lacked czrC, speG, and the additional recombinase genes. Four of the seven isolates (representing both strains) were tested for sasX = sesI, and all were positive in accordance with an affiliation to the "Southeast Asian Clade" of CC239-MRSA-III.

Observations Regarding S. argenteus
Two isolates were assigned to S. argenteus. They yielded signals for agr III and capsule type 5 alleles, harboured cna but lacked the egc enterotoxin gene cluster as the genome sequence of H115100079; GenBank CCEP/SAMEA1557135 does. Thus, they were assigned to CC2596. Both isolates lacked PVL genes, and any resistance or enterotoxin genes were covered by the array.

Detection of PVL
Among isolates from Trinidad and Tobago (Figure 1), PVL genes were common, albeit no MRSA and no S. argenteus with these genes were identified. Thirty-two out of 73 (44%) MSSA isolates harboured lukF/S-PV genes. PVL genes were found in CC5, CC6, CC8, and CC152, with CC8 and CC152 being the dominant lineages among PVL positives (thirteen isolates each). Of these isolates, CC8 and CC152 were the largest contributors of PVL, with thirteen PVL-positive strains each. However, the CC8 strain commonly found in Trinidad and Tobago was not found among the Jamaica isolates. PVL-positive isolates from that island belonged to CC6 and CC152. these isolates were sequenced (see below). Four isolates from this cluster also harboured a mercury resistance operon that, however, in two sequenced isolates, was plasmid-borne rather than SCC-associated.
The other two CC239 isolates carried SCC [III + SCCmer + ccrC] elements but lacked czrC, speG, and the additional recombinase genes. Four of the seven isolates (representing both strains) were tested for sasX = sesI, and all were positive in accordance with an affiliation to the "Southeast Asian Clade" of CC239-MRSA-III.

Observations Regarding S. argenteus
Two isolates were assigned to S. argenteus. They yielded signals for agr III and capsule type 5 alleles, harboured cna but lacked the egc enterotoxin gene cluster as the genome sequence of H115100079; GenBank CCEP/SAMEA1557135 does. Thus, they were assigned to CC2596. Both isolates lacked PVL genes, and any resistance or enterotoxin genes were covered by the array.

Detection of PVL
Among isolates from Trinidad and Tobago (Figure 1), PVL genes were common, albeit no MRSA and no S. argenteus with these genes were identified. Thirty-two out of 73 (44%) MSSA isolates harboured lukF/S-PV genes. PVL genes were found in CC5, CC6, CC8, and CC152, with CC8 and CC152 being the dominant lineages among PVL positives (thirteen isolates each). Of these isolates, CC8 and CC152 were the largest contributors of PVL, with thirteen PVL-positive strains each. However, the CC8 strain commonly found in Trinidad and Tobago was not found among the Jamaica isolates. PVL-positive isolates from that island belonged to CC6 and CC152. While thirty-two isolates carried lukF/S-PV, only thirty-one were found to be phenotypically positive for LukF-PV production by the lateral flow. Thus, the lateral flow assay-compared to genotyping-yielded a sensitivity of 96.9%, a specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 98.6%. The one genotypically positive but phenotypically negative isolate (a CC8-MSSA; 2020-042_7641M) was re-cultured and re-tested, but the results remained unchanged. It was then subjected to genome sequencing, in parallel to a pheno-and genotypically PVL-positive isolate of the same strain, in order to find a reason for the discrepant results. It showed a frameshift mutation in agrC because of a deletion of a single nucleotide (see Supplemental File S2, 2020-042_7641M chromosome, positions 2,055,453 to 2,056,744; whereas agrC of a phenotypically PVL positive control can be found at Supplemental File S2, 2020-043_7352M, chromosome, pos. 2,099,143 to 2,100,435).
The other agr genes (agrA, agrB, and agrD) and hld were inconspicuous (for an alignment to the sequences of reference strain NCTC8325 and PVL-producing 2020-043_7352M, see Supplemental File S3), as were the lukF-PV and lukS-PV genes.

Sequencing the Composite SCCmec Element in Clonal Complex 239 Isolates
Three CC239 isolates (2020-021_7037M, 2020-048_8421A, and 2020-009_371M) were sequenced (Supplemental File S2) to characterise the novel composite SCCmec element. It encompassed approximately 67,500 base pairs. Its gene content is summarised in Table 6, and a graphical overview is provided in Figure 2.
It comprised, directly adjacent to orfX, a gene encoding a type I restriction-modification system site-specificity determinate, hsdS, followed by a truncated transposase gene, the spermidine N-acetyltransferase gene speG, five genes encoding "putative proteins" (one of which was present in two copies), and recombinase genes ccrA/B-4. These are followed by a truncated ccr-associated cassette chromosome helicase gene (cch), yet another gene encoding a "putative protein," as well as by yozA (HTH-type transcriptional repressor) and czrC (cadmium and zinc resistance gene C, formerly known as cadA or copA; [50]).
The remaining part of the composite SCCmec III element of the Trinidad and Tobago strain was essentially identical to the corresponding part of the SCCmec element of the Southeast Asian Clade CC239 strain TW20, FN433596.1:(34140 to 48481) and in CMRSA-6 (CP027788.1). Besides the mec element (including mecA, mecI, mecR2 and mecR1, psmMEC, ydeM, ugpQ, and the dru region) and recombinase genes ccrA/B-3 as well as ccrAA/C, it harboured the aminoglycoside resistance gene ant9 and the macrolide/clindamycin resistance gene erm(A) as well as a cadmium-resistance operon (cadC, cadA, cadD).
Contrarily to TW20 and CMRSA-6, where the mercury resistance operon is localised on the SCCmec element, it was plasmid-borne in two of the three sequenced strains, and absent in the third one. When present, it was accompanied by yet another set of cadmium resistance genes and by quaternary ammonium compound resistance markers (qacA/R).
In one of the three isolates sequenced (2020-009_371M), a region comprising roughly 27,000 base pairs flanked on both sides by multiple transposase genes was inverted in order and orientation. It contained the mec element, the cadmium, and cstR/A/B-SCC operons, as well as ant9 and erm(A).

Discussion
The DNA microarray analysis approach, along with MLST and SCCmec typing, provided a comprehensive characterization and assignment of all isolates into clonal complexes and strains. Some clinically useful data with respect to antimicrobial resistance, virulence, and toxin gene carriage were also revealed. Results confirmed a high diversity of S. aureus clonal complexes from Trinidad and Tobago, with CC8 being the most prominent. While most of the infections in Trinidad and Tobago were caused by MSSA (73/85; 86%), all were caused by MSSA in Jamaica. Although the sample size from both islands was relatively small, most of the findings were consistent when compared to previous studies conducted on S. aureus isolates from both islands on much larger sample sizes. In the present study, most of the S. aureus infections occurred in patients with wound infections, mainly post-surgical in nature. Similar findings have been reported previously by several authors from Trinidad and Tobago [42,43,46,47,53,54] and Jamaica [45]. The exact reason for this occurrence is unknown; some authors have outlined possible explanations in reports from Trinidad and Tobago. Akpaka et al. [47] attributed the high levels of infection to excessive wound swabbing following medical procedures, and, in another report, noted that some surgical procedures can result in nosocomial infections with S. aureus as the main aetiological agent [54]. According to Orrett [43], the hands of health care providers may contribute to infection during dressing exercises, whereas Swanston [46] attributed the high usage of antibiotics to the increase in MRSA isolates from the surgical wards at the Port of Spain General Hospital. While any of the reasons outlined may be plausible as a contributory factor, it re-emphasizes the importance of implementing a strict infection control policy and applying guidelines for the prevention of surgical site infections in hospitals. This would include some form of multifaceted approach involving all team members to provide quality service before, during, and after surgical procedures.
Of the studies emanating from Trinidad and Tobago, Akpaka et al. [54] discovered a high prevalence of S. aureus infections in children aged 0 to 9. This high infection rate was attributed to an underdeveloped immune system and its inability to recognize staphylococcal components during infection at that age. Usually, older people tend to be more vulnerable to infections because aging is associated with immune dysfunction, multiple comorbidities, increased hospitalizations, and increased antibiotic use. Ramdass et al. [55] reported, at the Port of Spain General Hospital Trinidad, a peak age range of 60 to 69 years, in which 93% (14/15) of the MRSA-positive patients had used antibiotics (mainly beta-lactams) prior to admission into the hospital, had comorbidities (diabetes and hypertension), had long hospital stays (>1 week), were previously hospitalized and had previous surgery. In Trinidad and Tobago, most infections occurred among patients aged 30 to 39. This was followed by 0 to 9, 10 to 19 and finally patients over 60 years of age. Similar to the risk factors associated with infants and elderly, who are most often confined to nurseries and nursing homes, overcrowding may be a contributing factor to the high infection rate among the 10 to 39 age group. Persons in this age group frequently congregate in crowded areas of the community where the risk and exposure to infection are increased, such as correctional facilities, military/army camps, schools, and sports club settings. In Trinidad and Tobago, Orrett [53] outlined other major contributory factors that could be applied; these include the inappropriate use of antimicrobials and patients discontinuing therapy after being discharged due to the cost of prescriptions at local pharmacies. Most S. aureus infections in this study occurred in males from Trinidad and Tobago. Though not statistically significant, high S. aureus infection rates among males have been reported previously in other studies from the island [54,55]. According to Akpaka et al. [47], no obvious reason for the impact of gender on the prevalence of MRSA or MSSA in community or hospital settings has been reported in the literature. However, similar observations also have been made in other parts of the world.
Reports on the molecular characteristics and prevalence of MRSA in hospitals and within the community are constantly expanding in Trinidad and Tobago. The 12% MRSA prevalence observed herein was consistent with previously reported rates ranging from 9.8% to 15.3% [39,42,47,53,54,56], although, in 2018, Vire et al. [57] recorded the highest reported rate for the island, 44.4%. The most common MRSA strain was CC239-MRSA-III (8.3% of all S. aureus isolates, and 70% of MRSA). It has already been identified as the most prevalent clone in Trinidad and Tobago [41,54,56]. Until 2004, one variant of this strain (CMRSA-6) was the only MRSA clone circulating in Trinidad and Tobago [42]. According to previous work [58,59], CC239 with SCCmec III is a major dominant hospital-associated MLST type that has been described as the oldest truly pandemic MRSA strain in circulation since the 1970s and as the most successful international epidemic clone of MRSA. Its clade harbouring the virulence factor sasX/sesI is known to be widespread in Southeast Asia [51], but it was also, surprisingly, present in Trinidad and Tobago [59]. In this study, isolates were also tested positive for this gene. Sequencing revealed the localisation of sasX/sesI on a large mobile element, together with aminoglycoside genes (aacA-aphD and aphA3) and genes for phage enzymes. However, genes encoding phage structural proteins, such as those forming the capsid or tail, were not noted. Essentially the same complex can be found in TW20 (FN433596.1) and CMRSA-6 (CP027788.1), confirming the previous observation on the close relationship of the Trinidad and Tobago CC239 strains to the "South East Asian" Clade of CC239-MRSA-III and CMRSA-6 [42].
The Trinidad and Tobago variant of the "Southeast Asian" CC230-MRSA-III lineage, as described herein, harbours a composite SCCmec element including speG, czrC, and ccrA/B-4 recombinase genes in addition to its SCCmec III element. This variant has, to the best of our knowledge, exclusively been described from Trinidad and Tobago. Apparently, it has been extant there for at least 10 years (see [41]; when six out of seventy-six CC239 isolates collected in 2010 yielded array hybridisation signals for ccrA/B-4 genes).
The speG, czrC, and ccrA/B-4 cluster or "SCC-M1" element might have been transferred into S. aureus from Staphylococcus epidermidis. To the best of our knowledge, the oldest strain in which that element can be found is ATCC12228 (see [63] and GenBank AE015929.1), which was isolated many decades ago [64]. In ATCC12228, it differs only in an apparently random integration of a transposase gene, and it is not linked to a SCCmec element but to an SCC element carrying, among other genes, heavy metal resistance genes and a second copy of speG. The speG, czrC, and ccrA/B-4 cluster is also present in several other sequences of S. epidermidis, including the one of strain ATCC14990 (GenBank CP035288.1), which was isolated in 1963 [65].
For two reasons, it might be speculated that the acquisition of this element might confer an evolutionary advantage. First, it recently appeared in several unrelated S. aureus strains in distant parts of the world. Second, in Trinidad and Tobago, it was noted in 6 out of 76 isolates collected in 2010, but in 5 out of 7 in 2020, indicating an increasing prevalence (with the caveat of low numbers). The gene crzC causes cadmium and zinc resistance [50], and it is frequently observed in livestock-associated MRSA [66], which might have benefited from resistance to formerly widely used zinc supplements to animal fodder. Whether the dermatological use of zinc-containing cremes and ointments might pose a selective pressure favouring this strain remains to be clarified. The speG gene, encoding a spermidine N-acetyltransferase, can also be found as part of ACME elements, such as in the USA300 strain and in composite SCCmec elements (see above and [67]). It has been associated with resistance to exogenous polyamines [68].
Similar to the CC239-MRSA-III, the rapidly emerging "USA300" CC8 (ST8-MRSA-IV) clone has been isolated previously in Trinidad and Tobago and appeared to be quite common [31,39,41,54,56]. This well-known pandemic CA-MRSA strain is common in various regions of the globe. A direct transfer of this strain from North American visitors was proposed as a reason for its high local prevalence [54], along with its emergence from the Caribbean/Latin American region [69]. However, in this study, no USA300 isolate was found. Only one CC8-MSSA isolates harboured both PVL and ACME in this study and was assumed to be a variant of "USA300" that lost its mecA/SCCmec IV element. Thus, it is tempting to speculate that the prevalence of this hypervirulent, multi-resistant clone in Trinidad and Tobago may be declining. The related but distinct "Spanish/Latin American USA300" CC8-MRSA-IV ( [70], as represented by the genomes with GenBank accession numbers CP007670; CP007672), was also not detected despite geographical proximity.
Frequent domestic and international travel to Trinidad and Tobago has resulted in the discovery of several strains with global and regional connections. This study identified one CC88-MRSA-IVa strain, which has previously been reported in Africa and Australia [71]. The CC22-MRSA-IV (PVL+/tst1+) strain with a SCCmec IVa (MW2-like) element might indicate a Middle Eastern/Persian Gulf connection. It was observed in the Middle Eastern/Persian Gulf region and in Central Asia (see [72][73][74][75] as well as GenBank CP038850.1, from Pakistan). Another strain, CC72-MRSA-[VT + fusC], could also have been brought from there as composite SCCmec elements harbouring fusC are abundant in the Greater Middle East, and, indeed, similar or identical isolates have previously been observed from the UAE [76]. Findings of this nature should initiate a thorough investigation into a patient's travel history, which should provide a clear indication of the strains' possible origin. Moreover, a possible emergence of strains with composite SCCmec elements harbouring fusC should be closely monitored, as fusidic acid is commonly used in Trinidad and Tobago.
Another interesting observation is the high prevalence of PVL-positive S. aureus, reported consistently over the last decade with rates from 48% or 50% [39,77] to up to 62% [57]. In this study, the overall PVL prevalence for Trinidad and Tobago was lower, at 39%. CC8-and CC152-MSSA strains were the most common PVL-positive strains accounting for 15% and 14% of all Trinidad and Tobago isolates tested, with the former one apparently decreasing and the latter increasing compared to earlier studies [39,77]. The authors had already proposed an endemicity of the PVL-positive CC8-MSSA strains based on earlier observations [39], including the case of a severe and fatal infection in a previously healthy child [40]. Another common, endemic strain is PVL-positive CC152-MSSA. Interestingly, CC152 is an abundant lineage in Africa [78][79][80][81][82], and it was also observed in Haiti and Martinique [32,38]. Thus, it might be speculated that it was brought to the Caribbean islands together with people of African descent sometime in the history of colonisation. This might also apply to CC72. Since PVL-positive CC8-and CC152-MSSA has dominated Trinidad and Tobago for at least a decade [41], the rate of infection and possible patterns of evolution of these two clonal complex lineages will need to be closely monitored in the future.
While little information on PVL and MSSA is available for Jamaica, the 25% of PVLpositive MSSA strains identified should be monitored.
In Trinidad and Tobago's literature, there are currently limited reports of S. argenteus strains. A low prevalence of S. argenteus strains to belong to CC2596 (2/85; 2.35%) was reported in this study, which is comparable with previous findings by Monecke et al. [39], where seven (7/294; 2.38%) S. argenteus strains were identified. When re-assessing hybridisation patterns of S. argenteus isolates from all Trinidad and Tobago studies by the authors, three can be assigned to CC2596 (23%, including the two from this study), three to CC1223 (23%) and seven to CC2250 (54%), but none to CC1850 ("ST75", the longest known S. argenteus lineage). No S. argenteus isolate was found positive for mecA or PVL genes. Despite a low prevalence in both studies, S. argenteus should not be overlooked as this non-staphyloxanthin-producing strain [83], previously thought to be less virulent than S. aureus, is said to be becoming clinically important, with significant global prevalence and a virulence comparable to the one of S. aureus [84,85]. In the future, it is recommended that these strains should closely be monitored and reported [86]. The creation of a database within laboratories for the epidemiology, clinical impact, and implications for infection control of such isolates would be beneficial to combat potential outbreaks or increases in prevalence or virulence, given especially the observations of emerging PVL-positive [87] or methicillin-resistant S. argenteus [88,89] Isolates in this study harboured a slew of resistance genes, virulence factors, and toxins. Despite this, all were vancomycin susceptible. This indicates that vancomycin remains the first-line treatment for (severe and/or bloodstream) infections with MRSA or with any S. aureus as long as susceptibility tests are pending, as previously reported from Trinidad and Tobago and Jamaica [44,47].
Unfortunately, with a strong local presence of virulent, PVL-positive strains in the community and the ongoing evolution of multidrug-resistant MRSA in healthcare settings, a need for vaccine development, ongoing surveillance, proper infection control, and reinforcement of preventative measures in both hospital and community settings, remains.

Study Design and Eligibility
This cross-sectional study was conducted using 101 clinical samples recovered from Trinidad and Tobago and Jamaica. Samples were collected from August to September 2020. All eligible subjects, regardless of ethnic group, gender, social status, or educational level, who agreed to participate by means of written consent and assent, were included in this study. This includes patients with varying clinical presentations who were diagnosed with various S. aurei infections such as furunculosis or carbuncles, as well as with cutaneous abscesses, mastitis, and other prominent and severe skin and soft tissue infections (SSTI; necrotizing fasciitis, chronically purulent and painful "spider bites", particularly in cases with travel history) and recurrent or chronic SSTIs and necrotizing community-acquired pneumonia, including cases associated with influenza.

Laboratory Identification of Isolates
Samples were subjected to routine screening tests for the detection and isolation of S. aureus/argenteus. The clinical samples were cultured on mannitol fermentation salt agar at 37 • C for 24-48 h. Presumptive S. aureus was identified as gram-positive cocci isolates that produced yellow colonies with a yellow halo on mannitol salt agar. Colonies that were mannitol-salt-positive were conventionally identified as S. aureus based on colony morphology, Gram stain, haemolysis on sheep blood agar, catalase, and coagulase/latex agglutination tests.
Antibiotics susceptibility tests for sixteen antibiotics (including ampicillin, ampicillin + clavulanic acid, cefoxitin, ceftriaxone, cefuroxime, ceftazidime, ciprofloxacin, clindamycin, chloramphenicol, ertapenem, erythromycin, gentamicin, norfloxacin, penicillin, vancomycin, tetracycline) were routinely tested using the Kirby Bauer disc diffusion method on Mueller-Hinton agar in accordance with the Clinical and Laboratory Standards Institute (CLSI) guidelines. Cefoxitin 1 µg and oxacillin 30 µg antibiotic discs were used to screen for methicillin resistance. S. aureus strains ATCC 29213 and ATCC 25923 were used as controls.
All suspected S. aureus/argenteus samples were subjected to molecular testing using DNA microarray analysis as described below. This led to the exclusion of nine coagulasenegative Staphylococcus spec. isolates and the identification of two S. argenteus isolates, which were included in the analysis. An experimental lateral flow test was applied to all of these isolates to detect PVL expression (see below).

Microarray Analysis
The detection of virulence genes and resistance markers, as well as the assignment to clonal complexes, epidemic strains, and SCCmec types, was performed by microarray analysis using the StaphyType DNA microarray (Abbott [Alere Technologies GmbH], Jena, Germany) and the INTER-ARRAY Genotyping Kit S.aureus (Inter-Array GmbH, Bad Langensalza, Germany). The probes, primers, and procedures were previously described [2,49,90]. SCCmec subtyping and detection of sasX = sesI, using a second experimental array, was carried out on selected isolates as previously described [49]. In short, isolates were cultured overnight at 37 • C on Columbia blood agar. DNA was purified using Qiagen columns (Qiagen, Hilden, Germany) after enzymatic lysis of staphylococcal cells. The assay was based on a linear primer elongation that used one primer per target but amplified all targets simultaneously. During amplification, biotin-16-dUTP was incorporated into the single-stranded amplicons. After hybridisation to the DNA probes immobilised on the surface of the microarray, washing, and blocking, horseradish-peroxidase-streptavidin was conjugated. In case of a positive reaction, this conjugate bound to the biotin labels incorporated into the amplicons, causing in the next step a localised dye precipitation resulting in a visible and detectable formation. A transmission image of the microarray was recorded and analysed using a designated reader and software. This allowed the detection of individual target genes, as well as an automated comparison to a reference database facilitating assignment to CCs, strains, and SCCmec types [2,49,90].

Nanopore Sequencing
The Oxford Nanopore MinION platform was used to sequence the genomes of two MSSA isolates (2020-042_7641M, 2020-043_7352M) in order to investigate the lack of PVL production in the former. In addition, three CC239-MRSA (2020-009_371M, 2020-021_7037M, and 2020-048_8421A) were sequenced with the aim of describing their SCCmec elements.
The library preparation was performed using the 1D genomic DNA ligation kit (SQK-LSK109 with barcoding kit EXP-NBD104, and SQK-NBD114.24; Oxford Nanopore Technologies, Oxford, UK) following the manufacturer's instructions for flow cells (FLO-MIN106 containing an R9.4.1 pore, and FLO-MIN114 containing an R10.4.1 pore). Prior to library preparation, a size selection was performed using AMPure-beads (Beckman Coulter, Krefeld, Germany) in a ratio of 1:1 (v/v) with the isolated DNA sample. The flow cell was loaded with a total of approximately 600 ng/µL DNA (according to Qubit4 Fluorometer; Thermo Fisher Scientific, Waltham, MA, USA). The sequencing ran for 72 h using MinKNOW software version 22.12.5, and 22.12.7 starting with a total of 1200-1600 active pores.

PVL Lateral Flow Test
For the detection of PVL production, an experimental lateral flow assay (Senova, Weimar, Germany) was utilised as previously described [91]. In short, the assay relied on monoclonal antibodies targeting the LukF-PV protein. In order to test them, subcultured S. aureus isolates were incubated overnight or maximally up to 24 h at 37 ± 2 • C on Columbia Blood Agar. One inoculation loop (~10 µL) full of culture material was inoculated into 300 µL of the kit buffer and vortexed for 15-30 s in order to produce a homogenous suspension of cells. Then, the buffer with the inoculum was centrifuged at 2000× g for 30 s. A total of 100 µL of the supernatant was pipetted onto the sample well of the test device, and it was then incubated for 10 min at room temperature. The appearance of test and/or control lines was assessed visually, and the lateral flow device was photographed. The image of the test result was independently reviewed by two authors.

Conclusions
This study was conducted on 101 clinical S. aureus/argenteus isolates from Trinidad and Tobago and Jamaica. Samples were collected from August to September 2020. Ten isolates, all from Trinidad and Tobago, were identified as MRSA. The pandemic ST239-MRSA-III strain was still common (n = 7), but five of these isolates showed a novel composite SCCmec element in which a group of genes including speG, crzC, and ccrA/B-4 was linked to SCCmec III. The prevalence of PVL genes was high (in 38/101 isolates), although lower than in previous studies from Trinidad and Tobago. The USA300 PVL-MRSA strain was not found anymore, while the predominant PVL-positive strains were CC6-, CC8-, and CC152-MSSA. Funding: There was no external funding for the Caribbean groups. The Jena group acknowledges support by the German Federal Ministry of Education and Research and the German Federal Ministry for Economic Affairs and Energy, within the framework of four projects, "ADA" (13GW0456) and "ResiCheck" (13GW0422) aiming to develop rapid assays for the detection and characterisation of resistance genes and virulence factors in S. aureus, "LPI-BT5" aiming on novel technologies for rapid diagnostic assays (13N15717) and "Development of a leukocidin rapid assay" (49MF180153).

Institutional Review Board Statement:
This study was carried out in accordance with guidelines approved by the University of the West Indies Campus Ethics Committee, as well as the ethical review boards of the Trinidad and Tobago Regional Health Authorities (RHAs).

Informed Consent Statement: Not applicable.
Data Availability Statement: All relevant data are provided as Supplementary Files. The sequences of the genomes discussed, including the SCCmec element discussed, can be accessed under BioProject accession number PRJNA978032, and GenBank accession numbers CP127017 to CP127027.