Characterisation of Methicillin-Resistant Staphylococcus aureus from Alexandria, Egypt

The present study aims to characterise clinical MRSA isolates from a tertiary care centre in Egypt’s second-largest city, Alexandria. Thirty isolates collected in 2020 were genotypically characterised by microarray to detect their resistance and virulence genes and assign them to clonal complexes (CC) and strains. Isolates belonged to 11 different CCs and 14 different strains. CC15-MRSA-[V+fus] (n = 6), CC1-MRSA-[V+fus+tir+ccrA/B-1] (PVL+) (n = 5) as well as CC1-MRSA-[V+fus+tir+ccrA/B-1] and CC1153-MRSA-[V+fus] (PVL+) (both with n = 3) were the most common strains. Most isolates (83%) harboured variant or composite SCCmec V or VI elements that included the fusidic acid resistance gene fusC. The SCCmec [V+fus+tir+ccrA/B-1] element of one of the CC1 isolates was sequenced, revealing a presence not only of fusC but also of blaZ, aacA-aphD and other resistance genes. PVL genes were also common (40%). The hospital-acquired MRSA CC239-III strain was only found twice. A comparison to data from a study on strains collected in 2015 (Montelongo et al., 2022) showed an increase in fusC and PVL carriage and a decreasing prevalence of the CC239 strain. These observations indicate a diffusion of community-acquired strains into hospital settings. The beta-lactam use in hospitals and the widespread fusidic acid consumption in the community might pose a selective pressure that favours MRSA strains with composite SCCmec elements comprising mecA and fusC. This is an unsettling trend, but more MRSA typing data from Egypt are required.


Introduction
Staphylococcus aureus (S. aureus) is a human and animal pathogen that is a global cause of morbidity and mortality. Antimicrobial resistance in S. aureus is frequently associated with mobile genetic elements, including plasmids, transposons, and staphylococcal cassette chromosome (SCC) elements that act as carrier to exchange genetic information between Staphylococcus strains. Methicillin resistance in staphylococci is based on the production of altered penicillin-binding proteins with a low affinity for beta-lactam antibiotics. These proteins are encoded by different mec genes (mecA or mecC), out of which mecA is the most common and widespread one [1]. SCCmec elements carry mec genes along with the genes that control their expression. There are three basic genetic units within SCCmec: the ccr recombinase gene complex, the mec gene complex and the joining region (J region) [2][3][4][5][6]. Other SCC elements might carry fusidic acid resistance (mediated by the fusC gene), heavy metal resistance, or certain virulence factors such as tirS or a phenol-soluble modulin (PSM-mec) [7,8].

Resistance Genes and Antibiotic Resistance, SCCmec Elements
MRSA isolates were tested against 19 antimicrobial discs ( Table 2). Apart from cefoxitin, the isolates exhibited the highest rates of resistance to gentamicin (90%), tobramycin (90%), and fusidic acid (86.7%). Table 2. Susceptibility of MRSA isolates to antibiotics using the Kirby-Bauer disk diffusion method according to CLSI guidelines or if no interpretive criteria are recommended by the CLSI, according to EUCAST (https://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Breakpoint_tables/ v_10.0_Breakpoint_Tables.pdf; as accessed 1 December 2022).  All isolates harboured mecA as part of various SCCmec elements. SCCmec elements I, II, VT, VII, VIII, IX, X and XI (including mecC) were not found. Two isolates harboured composite SCCmec III elements. Another three isolates carried "plain" SCCmec IVa elements. All others had composite elements that included SCCmec IV, V, or VI and SCCfus. Accordingly, the SCC-associated fusC gene was found in 25 isolates, i.e., in 83%.

Antibiotic
In contrast, the plasmid-borne fusidic acid resistance gene far-1 was not detected. This gene is usually associated with the PVL-positive CC80-MRSA-IV, which was conspicuously absent. The mupirocin resistance gene mupA was detected once in a CC1-MRSA-[V+fus+tir+ccrAB1] isolate that also harboured cfr and aadD. The multidrug resistance gene cfr was found once, although no linezolid resistance was observed phenotypically. Vancomycin resistance genes were not detected which was in accordance to the phenotypic glycopeptide susceptibility of all isolates. The gene fexA was detected in two CC5 isolates, and both were phenotypically resistant to chloramphenicol. SCCmec markers and resistance genes, as detected by array hybridisation, are listed in Table 3.

Virulence Factors
Regarding virulence factors (Table 4), 12 isolates (40%) were Panton-Valentine leucocidin (PVL)-positive, of which five belonged to CC1, three to CC 1153, two to CC 152, one to CC 121 and one to CC 30. The enterotoxin gene cluster (egc, consisting of seg, sei, selm, seln, selo and selu) was found in all CC5, CC22, CC30, and CC121 isolates, representing 20% of all the tested isolates. The toxic shock toxin gene (tst1) was found in all CC22 isolates (6.7%). Enterotoxin genes sec and sel as well as the exfoliative toxins genes etB and etD were not detected. However, etD2=etE yielded signals in two isolates belonging to CC152 (6.7%). The epidermal cell differentiation inhibitor gene edinB was also found in these two isolates. The edinA gene was not identified in any of the isolates, and edinC was present in only one isolate belonging to CC5. Various combinations of the immune evasion complex (IEC) genes (sea, sak, chp and scn) were also detected in most isolates. Amongst the IEC-positive isolates, IEC type D (sea, sak, and scn) predominated (36.6%). Amongst the individual IEC genes, the most predominant gene was scn (in all but one CC1 isolate). The gene encoding surface-anchored protein X, sasX=sesI, was not found, neither in the two CC239 isolates (which represent the lineage from which it was initially described) nor elsewhere.

Strain Affiliations
This study identified 11 different CCs and 14 strains (as defined by CC affiliation, toxin gene carriage, and SCCmec subtype; Table 5). A recently published sequencing described MRSA from Alexandria [44] collected in 2015. Forty-six genomes from this study were analysed (see Section 4) and assigned to strains allowing a direct comparison of strain prevalence in 2015 and 2020. Table 5. Affiliations to CCs and strains, comparison to the data from the 2015 study [44].

Clonal Complex
Strain n in 2020 % in 2020 n in 2015 % in 2015 Eight isolates harboured SCCmec elements described according to the array profiles as SCCmec [V+fus+tir+ccrA/B-1]. All these isolates belonged to CC1. To the best of our knowledge, no contiguous sequence of such an element was yet available, as related MUM475 (GenBank AZSG01000015.1 plus AZSG01000034) and MRSA1_ST20130096 (FSRY01000032.1) carry SCCmec [VT+fus+tir+ccrA/B-1] elements. Thus, one representative isolate (the PVL-negative Alexandria_2020-19) was fully sequenced using nanopore technology. This confirmed affiliation to CC1, Sequence Type 1 and the absence of PVL from that particular isolate (although another prophage inhabited the integration site usually occupied by PVL phages).
Its SCCmec element spanned 72,298 bp. Its sequence and a detailed list of all identified genes are provided as supplemental files S2/S3, and a schematic representation is provided in Figure 1.
It comprised the markers of a typical SCCmec V element (mecA, ugpQ, ccrAA, ccrC, mvaS-SCC), but it lacked the additional ccrAA/ccrC and D1GU38 genes that define SCCmec VT as they are present in the previously released CC1 sequence MUM475 (GenBank AZSG01000015.1 plus AZSG01000034).
isolate (the PVL-negative Alexandria_2020-19) was fully sequenced using nanopore technology. This confirmed affiliation to CC1, Sequence Type 1 and the absence of PVL from that particular isolate (although another prophage inhabited the integration site usually occupied by PVL phages).
Its SCCmec element spanned 72,298 bp. Its sequence and a detailed list of all identified genes are provided as supplemental files S2/S3, and a schematic representation is provided in Figure 1. It comprised the markers of a typical SCCmec V element (mecA, ugpQ, ccrAA, ccrC, mvaS-SCC), but it lacked the additional ccrAA/ccrC and D1GU38 genes that define SCCmec VT as they are present in the previously released CC1 sequence MUM475 (GenBank AZSG01000015.1 plus AZSG01000034).
The SCCmec V genes were combined with several transposase genes and integrated mobile genetic elements that, among various "putative proteins", also included heavy metal resistance genes cadD+cadX. The aminoglycoside and streptothricin resistance genes aphA3 and sat were also present. The gene aadE (that typically accompanies these two genes) was found to be truncated from 909 bp to 624 bp. This truncation was also present in other Egyptian CC1 sequences (JAEOUR010000048.1 and JAEOWJ010000028.1 [44]), but it can also be found in entirely unrelated strains such as TCH1516 (CP000731.1, pos. 15,611 to 16,252). Furthermore, there was the tetracycline resistance gene tet(L), the penicillinase operon blaI+blaR+blaZ and the gentamicin/tobramycin resistance gene aacA-aphD.
Then, ccrA/B-1 recombinase genes and a truncated gene cluster encoding an incomplete type I restriction-modification system followed. The hsdM gene was still present there, hsdS was truncated and hsdR has been replaced by yet another transposase The SCCmec V genes were combined with several transposase genes and integrated mobile genetic elements that, among various "putative proteins", also included heavy metal resistance genes cadD+cadX. The aminoglycoside and streptothricin resistance genes aphA3 and sat were also present. The gene aadE (that typically accompanies these two genes) was found to be truncated from 909 bp to 624 bp. This truncation was also present in other Egyptian CC1 sequences (JAEOUR010000048.1 and JAEOWJ010000028.1 [44]), but it can also be found in entirely unrelated strains such as TCH1516 (CP000731.1, pos. 15,611 to 16,252). Furthermore, there was the tetracycline resistance gene tet(L), the penicillinase operon blaI+blaR+blaZ and the gentamicin/tobramycin resistance gene aacA-aphD.
Then, ccrA/B-1 recombinase genes and a truncated gene cluster encoding an incomplete type I restriction-modification system followed. The hsdM gene was still present there, hsdS was truncated and hsdR has been replaced by yet another transposase copy which also removed the first gene (for a "putative protein" Q6GD54) from the actual fusC-associated complex. This complex included, among other genes, tirS, fusC and yobV; it was with regard to gene content as well as to allelic variants most closely related to the one in other CC1 strains (MSSA476, BX571857 and KT/314250, AOCP01000013) and thus it could be assigned to the previously defined [35] fusC-complex class "A". The association of a fusCcomplex class "A" with ccrA/B-1 recombinase genes is typical for, but not restricted to, CC1 strains [35].

Discussion
With regard to antimicrobial resistance, the most remarkable observation of the present study was a presence of fusC in diverse lineages of S. aureus and its extremely high prevalence. It was as high as 83%%, while in 2015 [44], the rate of genotypically fusidic acid (FA)-resistant strains (e.g., positive either for fusC or far1) was 30,4%. This could indicate an alarming trend, although several caveats (low numbers, absence of data from other hospitals, other towns and provinces, and predominance of skin/soft tissue infections among study samples) apply. As many genotypically different strains were involved, a local outbreak situation as a cause for this observation can likely be ruled out. A high and possibly increasing rate of genotypic FA resistance is likely to be related to a high rate of consumption of that drug, as observed in other countries such as New Zealand, where an increase in FA consumption led to a parallel increase in MRSA with composite SCC[mec+fus] Antibiotics 2023, 12, 78 9 of 19 elements [54]. Indeed, FA is over-the-counter available in Egyptian pharmacies, without prescription, as ointment, cream, or eye-drops. It is extensively misused and/or overused as monotherapy, even for non-infectious skin conditions or for prophylaxis. It is inexpensive, with prices as low as 15 to 25 Egyptian Pounds (ca. 0.75 to 1.30 Euro) for 15 g crème with 2% FA content. MRSA with composite SCC[mec+fus] elements can be expected to have a clear evolutionary advantage under such conditions in both ecological niches, in hospitals and in the community. In hospitals, they thrive because of their beta-lactam resistance (with about half of all antibiotics used being beta-lactams; see [55]). The high consumption of FA in the community poses a selective pressure favouring fusC-positives. When mecA and fusC are located on the same mobile genetic element, outpatient use of FA promotes MRSA in the same way as an in-hospital use of beta-lactams favours FA resistance. This might be a reason for a blurring of the distinction between CA-and HA-MRSA. It also means that the excessive use of FA in the community eventually endangers the lives of Egyptian hospital patients. Therefore, the use of FA should be curtailed, e.g., by requiring a prescription by a physician, as has been recently done in the U.A.E.
Another interesting observation was the high rate of carriage of PVL genes. Similar observations were made at other study sites in Egypt [56] and the Middle East [57,58]. A previous study [56] from another Egyptian city (Cairo) observed, in 20202, an even higher PVL rate of 75% (29% in hospital-and 92% in community-acquired infection) clearly indicating that the rate observed here was no outlier. In general, Middle Eastern and Northern African studies indicate that PVL-MRSA are no longer restricted to the "community" (if they ever were in this part of the world) but also thoroughly infiltrated hospital settings. The high PVL prevalence in our study, as well as in others from the region, cannot be attributed to a single outbreak strain, simply because of a high genotypic diversity of PVL-positive strains. However, the high proportion of wound and pus samples could in our case have caused a bias towards PVL-positive strains. The PVL rate for MRSA collected in Alexandria in 2015 was much lower 17% [44], and it cannot be ruled out that the lower rate observed in this study could be related to an outbreak of a PVL-negative strain (CC239).
The recent publication of genomes of MRSA isolates also from Alexandria [44] allowed to look at temporal changes affecting population structure, as defined by affiliations to CCs. CC1 strains with SCCmec V or VT (as in MUM475, GenBank AZGS) elements that additionally harbour fusC and the virulence factor tirS have frequently been observed in various countries. However, in most cases, whether they originated from SCCmec V or VT elements was not determined. Some previously described isolates [52,59] with Middle Eastern or Eastern African provenance indeed harboured composite elements based on SCCmec V rather than on SCCmec VT, as all Egyptian isolates described herein or in the earlier study [44] did. As discussed above, one isolate was sequenced to characterise its SCCmec element. It was found to harbour mecA, fusC, and several other resistance genes (including aacA-aphD), fitting into a broader trend of increasing "multi-resistance" in supposedly community-associated SCCmec IV/V strains. Another aminoglycoside resistance gene in this strain's SCCmec element, aadE, appeared to be truncated in our sequence. This was not a sequencing artifact, as corresponding contigs of previously sequenced CC1 strains [44] and strains from unrelated lineages (e.g., TCH1516, CC8 or TW20, CC239) showed the same. This likely indicated that this truncation predated the acquisition of the mobile genetic element carrying aphA3/sat/aadE by diverse MRSA strains.
Not much is known about CC6-MRSA-[V+fus], although a similar or related strain was observed in Kuwait [33]. In general, CC6 MRSA are common and widespread in the Middle East, but previously described strains usually differ in having SCCmec type IV [61][62][63][64].
CC15-MRSA-[V+fus] is a remarkable strain given that it is nearly the only MRSA strain that emerged from the globally spread and common lineage CC15 [65][66][67]. It has been found in humans in Saudi Arabia and other Gulf countries and in livestock and camel meat [68][69][70][71]. It was also observed in chickens from Egypt (unpubl. communication with Dr. Hotzel, Jena, Germany), and it was detected in a farmer from the Nile Delta region in Egypt [42]. In 2015, this MRSA strain was not found in Alexandria [44].
CC22-MRSA-IVa carrying the tst1 gene has been frequently observed around the Mediterranean Sea [72], in Middle Eastern counties [69,[73][74][75][76], and among refugees from the Middle East after it was first from Gaza [77][78][79]. It was also found in Egypt, in livestock, and in farm personnel [42], as well as in an Alexandrian hospital [44]. Thus, its detection is not surprising, but it was remarkable for being one of three lineages that did not harbour SCC-encoded fusidic acid resistance.
PVL-positive CC30-MRSA-IV (PVL+) has been dubbed the "WSPP/Southwest Pacific Clone" after an initial outbreak among New Zealanders and Samoans [80,81]. Meanwhile, such strains can be found globally, but different SCCmec subtypes might indicate a polyphyletic emergence. SCCmec IVa, as in the present isolate, has been observed in WSPP-like isolates from Europe and the Middle East [52,62] and the U.S. (GenBank CP026066).
CC97-MRSA- [V+fus] has been found in Europe and the Middle East [69], and it was detected some years ago in chicken meat brought from Egypt to Germany [82], indicating both a prolonged presence in Egypt as well as a possible livestock connection. Other, fusC-negative, CC97-MRSA have also been observed in Egypt [44].
CC152-MRSA-[V+fus] belongs to a lineage known to be common in Africa, but these isolates are usually MSSA. The particular MRSA strain was observed in Egyptian livestock and contact persons [42], and again, the presence of PVL could be seen as an indication of transmission from humans to animals. Other observations came from the Arabian Gulf [62] and Egypt [44]. CC152 was recently shown to carry a novel etD/E homologue, etE2 [53]. The observation of weak and/or irregular signals for etD2 likely can be attributed to a presence of this gene.
CC239-MRSA is a comparatively ancient, truly pandemic lineage of hospital-acquired MRSA with a core genome that can be described as a chimera of CC8 and CC30 [87] harbouring the large and distinct SCCmec III element [3]. Many variants of that strain largely correlate with geographic regions of origin [88][89][90][91][92]. The two isolates found in our study were not identical. One isolate belonged to a clade previously identified in various Middle Eastern countries or people from there [88]. The other isolate matched a group of strains and sequences from Western Europe (Portugal), the U.S. (ATCC33592), and Russia [88], as well as from Egypt [44]. Since it differed only in the absence of ccrAA/C recombinase genes from the abovementioned clade, this entire group, or single specimens out of it, might be a mere deletion variant of the Middle Eastern clade. CC239-MRSA was previously found in Alexandria [44], and then (in 2015) it was the most common strain, comprising nearly half of the MRSA isolates characterised, i.e., 23 out of 47 isolates (49%) that belonged to various variants of the ST239-MRSA-III strain.
The issue of the receding CC239 clone might, although the numbers of typed strains are low, suggest a profound change in the MRSA population structure, blurring the distinction between HA-and CA-MRSA. As mentioned in the Introduction, there was an idea of distinguishing CA-and HA-MRSA by molecular means, e.g., based on PVL status, type of SCCmec elements, and affiliation to "unusual" clonal complexes. This concept might still apply in countries such as China (where CC239-MRSA-III at least until recently predominated in hospitals and CC59-MRSA-IV or -V in the community [93,94]) or the USA (where CC5-MRSA-II used to be common in hospitals, while PVL-positive CC8-MRSA-IV prevail in the community). However, among the Egyptian sample analysed herein, only two out of 30 hospital isolates could be assigned to the CC239 strain traditionally associated with a hospital-acquired infection. In contrast, all others belonged to various strains with features associated with CA-MRSA (PVL, SCCmec IV/V/VI elements). Similar observations were also made in Middle Eastern countries [37]. No molecular marker can be used in these settings anymore as a surrogate for the assignment to hospital-versus community-acquired infections. In order to discern these, one must thoroughly interview the patient and assess the case history.
CC1153-MRSA-[V+fus] (PVL+) is a strain that we previously found in a patient of Egyptian origin in Germany [35] and Dubai [58]. Furthermore, isolates for which it was not determined whether they harboured SCC[mec V+fus] or SCC[mec VT+fus] elements have been observed in the United Arab Emirates, Saudi Arabia, and Kuwait [35]. PVL-positive MSSA from this lineage has also been found in Egypt's livestock [40]. Given the pathogenetic role of PVL in humans, this might be attributed to an anthropozoonotic transition. However, this observation raises the question whether ancestral, susceptible CC1153 strains might already have circulated in Egypt prior to the emergence of MRSA from this lineage. In 2015, no CC153 MRSA was found [44], possibly indicating a recent emergence.
Finally, the PVL-positive CC80-MRSA-IV strain, widespread in the Mediterranean and the Middle East [30,73,[95][96][97][98][99] and previously found in Alexandria [44], was not observed. Whether this is due to the small sample or a recent decline still needs to be established.
A limitation of the study is of course the small sample size, resulting from opportunistic sampling at a single location. The predominance of isolates from swab/pus specimens might have caused a bias towards fusCand/or PVL-positive isolates that might be less common in other types of samples.
More comprehensive surveys into S. aureus/MRSA populations in Egypt and elsewhere in Africa and the Greater Middle East are urgently needed as well as studies on a possible impact of new antibiotics such as daptomycin and fifth generation cephalosporins. A two-pronged approach of array-based typing followed by genome sequencing of "interesting" or conspicuous strains, as described herein, might help to gather more typing data for these parts of the world.

Isolates
Clinical samples (see Table 1) were cultured routinely on blood agar (Oxoid Ltd., Basingstoke, Hampshire, UK) and MacConkey agar and were incubated at 37 • C under aerobic conditions for 16-24 h. Subculturing was aided by microscopy; Gram-positive cocci were further identified by biochemical tests, including Catalase test and Coagulase test (Remel-Oxoid, Basingstoke, Hampshire, UK) and they were subjected to antibiotic susceptibility testing by disc diffusion (Oxoid Ltd., Basingstoke, Hampshire, UK) using CLSI methodology and breakpoints (CLSI; https://clsi.org/media/3481/m100ed30_sample.pdf; accessed on 1 December 2022). In addition, identification as S. aureus/MRSA was confirmed by PCR (see below). For long-term storage at the laboratory in Egypt, one ml of fresh saturated bacterial culture grown on Luria Bertani (LB) broth was added to one ml of sterile glycerol solution in screw capped glass tubes. The tubes were stored at −20 • C. Strains at the German laboratory were stored at −80 • C using microbank tubes (Fisher Scientific GmbH, Schwerte, Germany/Pro-Lab Diagnostics, Richmond Hill, ON, Canada) according to the manufacturer's instructions. For re-culturing, one loop of bacterial material was streaked over blood agar and incubated overnight at 37 • C.

PCR for S. aureus/MRSA
Multiplex PCR was used for genotypically identifying S. aureus and methicillin resistance by amplification of femA and mecA genes, respectively. Nucleotide sequences of primers (Biosearch Technologies, Inc., Petaluma, CA, USA) used in this investigation are shown in Table 6. DNA was extracted from MRSA isolates by boiling method [103]. Multiplex PCR was performed using 12 µL as total volume, consisting of 6 µL of mastermix (MyTaq HS Red Mix 2X; BioLine, London, UK), 0.5 µL of each of the diluted DNA extract, femA, and mecA primers, and 3.5 µL of PCR grade water. A negative control was prepared by adding the same contents to the tubes with water placed instead of the DNA extract. PCR cycling conditions were as follows: 4 min of initial denaturation at 95 • C, followed by 35 cycles of denaturation at 95 • C for 15 s, annealing at 52 • C for 15 s, and extension at 72 • C for 1 min, followed by a final extension at 72 • C for 5 min. PCR was performed using a Veriti Thermal Cycler (Applied Biosystems, Foster, CA, USA). The PCR products were loaded on 1.5% agarose gel and analysed by gel electrophoresis (Mupid-exU, ADVANCE Co., Ltd., Tokyo, Japan).

Array Procedures
For this study, a new experimental microarray was used. It was based on a previously described system [23,47]. However, it has been modified by adding probes for detecting some recently described markers and for a more detailed typing of SCC elements. These probes have already been used in earlier work when they were, among others, localised on another second array [52]. Experimental procedures were performed as described for earlier versions [23,47]; primer and probe sequences have been disclosed thence ( [23,47,51,52]; see also Tables 1-3 and Supplemental Table S1 for individual target genes).
Isolates were cultured overnight at 37 • C on Columbia blood agar. Harvested cells were digested enzymatically [104]. DNA was purified using Qiagen columns (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The assay relied on a linear multiplex primer elongation using one primer per target. During amplification, biotin-16-dUTP was incorporated into the amplicons which were then hybridised to the array. After washing and blocking, horseradish-peroxidase-streptavidin was added binding to the biotin and causing local precipitation of a dye in case of a positive reaction. Finally, an array image was recorded and analysed using a designated reader and software (Arraymate, Iconoclust, both by Alere Technologies/Abbott, Jena, Germany).

Nanopore Sequencing
The Oxford Nanopore MinION platform was used to sequence the genome of one MRSA isolate (Alexandria_2020-19). Library preparation was done using the 1D genomic DNA by ligation kit (SQK-LSK109, version GDE_9063_v109_revX_14Aug2019; ONT) following the manufacturer's instruction for flongles (FLO-FLG001 containing an R9.4.1 pore). Before library preparation, size selection was performed using AMPure-beads (Beckman Coulter) in a ratio of 1:1 (v/v) with the isolated DNA sample. The flongle flow cell was loaded with ca. 200 ng DNA (measured by Qubit4 Fluorometer; Thermo Fisher Scientific, Waltham, WA, USA). The sequencing ran for 48 h using the MinKNOW software version 20.10.3 starting with a total of 65 active pores.
The Guppy basecaller (version 4.4.2+9623c1626, Oxford Nanopore Technologies, Oxford, UK) translated the MinION raw reads (FAST5) into quality tagged sequence reads (4000 reads per FASTQ-file) using the barcode trimming option. Flye (version 2.8.3-b1695) was used to assemble each strain's quality tagged sequence reads into one big circular contig. The polishing of assemblies was divided into two steps. At first, racon (v1.4.17) was iteratively used four times with the following parameter: match 8; mismatch 6; gap 8, and windows-lengths 500. Afterwards, medaka (version 1.4.3) ran on the last racon (version 1.4.21) polished assembly using the model r941_min_high_g360. This corrected assembly was used for further analysis.

Analysis of Previously Published Genome Sequences
Published genome sequences from the previously published paper on MRSA from Alexandria [44]  . Sequences were analysed for the presence of the known probe sequences and their reverse complement sequences. A perfect match was assigned a score of 0.9. In case of one or two mismatches, a score of 0.4 was assigned. Probes with no hits, or with more than two mismatches, got a score of 0.0. Thus, a list of probes with corresponding scores was generated for each sequence and this list was analysed in the same way as the measurements from array experiments. This allowed us to assign previously sequenced isolates to strains using the same nomenclature and criteria for array experiments and directly compare results from both approaches. Two isolates were described as SCCmec un-typable (JAEOUU and JAEOWG) but were identified as CC97 and CC1 MRSA, respectively. This contradiction cannot be resolved here, but results for these sequences were included in Table 4. Another sequence (JAEOWU) included mecA (as also mentioned in the supplemental file to [44]) but lacked any other SCC-associated genes; it was excluded.

Conclusions
In conclusion, the population structure of MRSA from Alexandria, 2020, was characterised by the presence of many genotypically diverse strains. Most (83%) harboured SCCmec elements that included a fusidic acid resistance gene, and PVL was also common (40%). A comparison to an earlier study [44] from the same city in 2015 suggested a dramatic increase in the prevalence of fusidic acid resistance and of PVL carriage while a long-known strictly hospital-associated strain (CC239-MRSA-III) was receding. These observations indicate a diffusion of community-acquired strains into hospital settings and a selective pressure by beta-lactam use in hospitals and fusidic acid consumption in the community that favours MRSA strains with composite SCCmec elements comprising mecA and fusC. This could indicate an unsettling trend, but more data are needed to assess the current situation.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/antibiotics12010078/s1, Table S1: Array hybridisation profiles (pdf); File S2a: Genome sequence of strain Alexandria_2020-19 (fasta); File S2b: Genes identified in the SCCmec element of strain Alexandria_2020-19 (fasta). Funding: There was no external funding for the Alexandria group. The Jena group acknowledges support from the German Federal Ministry of Education and Research within the framework of two projects, ADA (13GW0456), aiming to develop rapid assays for the detection and characterisation of resistance genes and virulence factors in zoonotic S. aureus, and Resicheck (13GW0422), aiming to develop rapid tests for MRSA.
Institutional Review Board Statement: The study was approved by the Alexandria University Ethics Committee (IORG0008812; Protocol E/C. S/N. T44/2021).
Informed Consent Statement: Not applicable, as no human samples or personal data were used. Isolates were not purposefully obtained for this study but derived from routine diagnostics aiming to monitor antibiotic resistance.
Data Availability Statement: All relevant data are provided as supplementary files. The genome sequence, including the SCCmec element discussed, can be accessed under BioSample accession number SAMN31868372 and GenBank accession number CP113244.1.