Comparative Genomics Identifies Novel Genetic Changes Associated with Oxacillin, Vancomycin and Daptomycin Susceptibility in ST100 Methicillin-Resistant Staphylococcus aureus

Infections due to vancomycin-intermediate S. aureus (VISA) and heterogeneous VISA (hVISA) represent a serious concern due to their association with vancomycin treatment failure. However, the underlying molecular mechanism responsible for the hVISA/VISA phenotype is complex and not yet fully understood. We have previously characterized two ST100-MRSA-hVISA clinical isolates recovered before and after 40 days of vancomycin treatment (D1 and D2, respectively) and two in vitro VISA derivatives (D23C9 and D2P11), selected independently from D2 in the presence of vancomycin. This follow-up study was aimed at further characterizing these isogenic strains and obtaining their whole genome sequences to unravel changes associated with antibiotic resistance. It is interesting to note that none of these isogenic strains carry SNPs in the regulatory operons vraUTSR, walKR and/or graXRS. Nonetheless, genetic changes including SNPs, INDELs and IS256 genomic insertions/rearrangements were found both in in vivo and in vitro vancomycin-selected strains. Some were found in the downstream target genes of the aforementioned regulatory operons, which are involved in cell wall and phosphate metabolism, staphylococcal growth and biofilm formation. Some of the genetic changes reported herein have not been previously associated with vancomycin, daptomycin and/or oxacillin resistance in S. aureus.


Introduction
Staphylococcus aureus is a relevant pathogen with an extraordinary ability to evolve and acquire resistance to several antibiotics. Over the last decades, the large spread of antimicrobial-resistant (AMR) strains, including methicillin-resistant S. aureus (MRSA), vancomycin-intermediate S. aureus (VISA) and heterogeneous VISA (hVISA), have raised an alarm worldwide as declared by the World Health Organization in 2017 [1]. hVISA/VISA isolates are associated with persistent infections, vancomycin treatment failure and poor clinical outcomes [2].
Although the prevalence of hVISA and VISA is relatively low worldwide [3][4][5], a recent review and meta-analysis revealed that it has been increasing since 2010 (especially in Asia and America) [3]. This highlights the importance of understanding the resistance mechanism in order to define adequate control measures.
It has been twenty-five years since the first hVISA/VISA strains were reported [6,7]. However, the underlying molecular mechanism responsible for the hVISA/VISA phenotype is not yet fully understood. Moreover, the reduced susceptibility to vancomycin is often accompanied by concomitant changes in the susceptibility to oxacillin [8,9] and other last-resort antibiotics such as daptomycin [10], making it more difficult to establish a correct treatment for infections caused by these strains. Whole genome sequencing (WGS) of hVISA/VISA strains has been essential for detecting genetic changes associated with their phenotype. Despite the evidence showing modifications in peptidoglycan metabolism in hVISA/VISA, associated genetic changes seem to implicate a diverse set of mutations and chromosomal rearrangements. Non-synonymous single nucleotide polymorphisms (SNPs) in vraSR, yvqF/vraT, walKR, graXRS (involved in peptidoglycan metabolism and cell wall stress stimulon) or rpoB were amongst the first and most frequently reported genetic changes in hVISA/VISA [11][12][13]. Mutations in those genes have been experimentally tested to be responsible for promoting vancomycin resistance in VISA. In addition, some researchers reported IS256 insertions disrupting different genes implied in cell wall synthesis (tcaA, walKR) that lead to the VISA phenotype [14][15][16][17] and daptomycin resistance [18].
IS256 is an insertion sequence that has been detected in multiple copies in the genome of Staphylococcus spp. strains recovered from humans and animals [19,20]. IS256 can be found flanking the ends of transposon Tn4001 [21], and it has been prevalently described in MRSA clones belonging to CC8 (ST239, ST247, ST8) and CC5 (ST5, ST100, ST228) [16,18,[22][23][24]. The transposition of IS256 in S. aureus is a copy-and-paste mechanism [25] and can result in a variety of genetic modifications that affect the expression of genes involved in virulence and antimicrobial resistance [26][27][28][29]. Our results highlighted that vancomycin treatment increased IS256 transposition and showed that different VISA phenotypes could be selected from the same parental ST100-hVISA strain [24,30].
In this follow-up study, our aim is to further characterize these isogenic strains and to obtain their whole genome sequences in order to unravel genetic changes associated with antibiotic resistance.
We herein describe novel mutations and genetic rearrangements developed after vancomycin pressure (in vivo treatment and in vitro selection), some of which (as far as we know) have not been previously reported in the literature as associated with vancomycin, daptomycin and/or oxacillin resistance in S. aureus.

Antimicrobial Susceptibility
ST100 strains D1, D2, D23C9 and D2P11 differed not only in their susceptibility to vancomycin, but also to oxacillin (Table 1). A deeper analysis revealed that D2, recovered after arthrotomy and surgical cleaning after 40 days of vancomycin treatment, significantly increases its oxacillin susceptibility when compared to D1, rendering a phenotype that resembles a heteroresistant behavior ( Figure 1). In addition, VISA derivatives D23C9 and D2P11 presented a 4-16-fold and 1.31-2.63-fold increase in oxacillin and daptomycin MIC relative to parental strain D2, respectively (Table 1).

General Genomic Features
The assembled draft genomes of the four strains yielded the following results: total genome length of 2,785,355-2,795,159 pb; GC content of 32.79-32.81%; 31-40 contigs > 1 kb in length; and N50 of 148,220-337,832 bp. (Table 2). All strains harbor the mecA gene within the class B mec gene complex of a truncated SCCmec with no ccr genes (Table 3). An overview of mobile genetic elements and AMR determinants found is summarized in Supplementary Table S2. Table 1. Antimicrobial susceptibility profile of the bacterial strains used in this study. MICs (µg/mL), inhibition zone diameter obtained after daptomycin (DAP) pre-diffusion method (mm), and vancomycin (VAN) population analysis profile and area under the curve (PAP-AUC) ratio. The AUC was measured for each sample and the ratio of test isolate AUC/mean Mu3 AUC was calculated. Mu3 was used as hVISA control strain. The criteria to define hVISA and VSSA were a PAP-AUC ratio ≥ 0.90 and a PAP/AUC ratio < 0.90, respectively.

General Genomic Features
The assembled draft genomes of the four strains yielded the following results: total genome length of 2,785,355-2,795,159 pb; GC content of 32.79-32.81%; 31-40 contigs > 1 kb in length; and N50 of 148,220-337,832 bp. (Table 2). All strains harbor the mecA gene within the class B mec gene complex of a truncated SCCmec with no ccr genes (Table 3). An overview of mobile genetic elements and AMR determinants found is summarized in Supplementary Table S2.   Table 2. Accession numbers, genome coverage and assembly metrics of genomes included in this study. # Contigs is the total number of contigs in the assembly. # Contigs (≥x bp) is the total number of contigs of length ≥ x bp. Largest contig is the length of the longest contig in the assembly. Total length is the total number of bases in the assembly. Total length (≥x bp) is the total number of bases in contigs of length ≥ x bp. GC (%) is the total number of G and C nucleotides in the assembly, divided by the total length of the assembly. N50 is the length for which the collection of all contigs of that length or longer covers at least half (50%) the total base content of the assembly. N90 is used for the same purpose but the length is set at 90% of total base content instead of 50%. L50 is the  Table 2. Accession numbers, genome coverage and assembly metrics of genomes included in this study. # Contigs is the total number of contigs in the assembly. # Contigs (≥x bp) is the total number of contigs of length ≥ x bp. Largest contig is the length of the longest contig in the assembly. Total length is the total number of bases in the assembly. Total length (≥x bp) is the total number of bases in contigs of length ≥ x bp. GC (%) is the total number of G and C nucleotides in the assembly, divided by the total length of the assembly. N50 is the length for which the collection of all contigs of that length or longer covers at least half (50%) the total base content of the assembly. N90 is used for the same purpose but the length is set at 90% of total base content instead of 50%. L50 is the number of contigs equal to or longer than the N50 length. L90 is used for the same purpose in reference to the N90 length. # N's per 100 kbp is the number of ambiguous bases (Ns) per 100 kbp.

Mutations Associated with hVISA/VISA
All 4 strains shared 20 non-synonymous SNPs in genes related to cell wall metabolism and/or the hVISA/VISA phenotype (Table 4) when compared to S. aureus N315 (listed in Supplementary Table S3). It is worth noting, none of the strains carried SNPs in vraTSR, walKR and/or graXRS operons. Hence, comparative genomic analyses between D1 and D2, and between D2 and its in vitro-derived mutants were performed to unravel the genetic differences associated with their AMR profiles.
Seven SNPs and four INDELs distinguish D2 and its derived mutants (D23C9, D2P11) from D1 (recovered before vancomycin treatment) ( Table 5). INDELs affected genes stp1, braS, sagB and era, related to cell stress and envelope metabolism [31][32][33][34][35] and possibly related to the different oxacillin and vancomycin phenotypes observed (Table 1, Figure 1). Table 5. Genetic changes differing between strains analyzed in this study after mapping the reads to the S. aureus D1 genome. All changes and positions are expressed in reference to the S. aureus N315 reference genome except otherwise stated. Genes without an annotated name are in reference to the CDS of N315 (or D1 genome if not present in the latter) by their locus tag. In addition, in vitro-derived VISA strains D23C9 and D2P11, harbor mutations leading to premature stop codons in two genes linked to staphylococcal growth (phoR and era, respectively) (Table 5) [35,36]. Along with these genetic changes, both VISA strains showed a lower median cell diameter, longer latency growth phase and slower growth (slopes in the logarithmic exponential growth phase were significantly different, p < 0.0001), when compared to parental strain D2 ( Figure 2).

IS256-Mediated Genomic Rearrangements
Changes in IS256 transposition after vancomycin selective pressure for this set of strains [24] were also detected by WGS bioinformatic analysis performed in this study (Tables 6 and 7).
Comparative genomic analysis of paired S. aureus strains showed evidence of genetic rearrangements after the IS256 transposition-mediated antibiotic treatment. In total, 11 different IS256 insertions sites (4/11 disrupting genes) were shared by the 4 strains (Table 7), while both in vitro-derived mutants (D23C9 and D2P11) showed modifications in the IS256 copy number and position (Table 5). We did not find genetic changes in the sigB and rsbU genes, known global regulators of IS256 transposition [26,37]. In addition, no change in the copy number and/or location of other staphylococcal IS elements was evident.

IS256-Mediated Genomic Rearrangements
Changes in IS256 transposition after vancomycin selective pressure for this set of strains [24] were also detected by WGS bioinformatic analysis performed in this study (Tables 6 and 7).    -Absence of IS256 in the genomic location. * IS256 insertion site was detected but one of the genes in the flanking region is not present in the analyzed assembly, ** Genomic region encompassing pitR and sarX genes, absent in D23C9 and D2P11 (8 kb deletion), *** The annotated msaC gene was found deleted in all genomes analyzed in this study.
It is worth noting that both VISA derivatives are characterized by the absence of a ≈ 8 kb region encompassing genes pitR (phosphate uptake regulator), pitA (low affinity inorganic phosphate transporter), SA0620 (secretory antigen ssA-like protein-CHAP domain peptidoglycan hydrolase), SA0621 (integral membrane protein interacts with ftsHlike protein), rbf (araC type transcriptional regulator) and sarX (transcriptional regulator). This was confirmed by the lack of raw reads mapping to the corresponding region in the assembled genome of parental strain D2, and by PCR-Sanger sequencing (Figure 3, Supplementary Table S1).
The presence of IS256 insertion sites between vraG-pitR, SA0621-rbf and sarX-SA0624, in the D1 and D2 genomes, suggests that genetic rearrangements between neighboring IS256 elements might be responsible for this region deletion in the in vitro-derived VISA strains.
We further evaluated if the ≈8 kb deletion and/or IS256 insertion sites present in its genetic environment were shared by other strains. No other public genomes were found with the ≈8 kb deletion carried by D23C9 and D2P11 (BLAST searches against the NCBI Refseq complete S. aureus genomes database, last accessed 20 January 2023).
In line with previous results [24], IS256 transposition was higher in D23C9. The genome of this strain contains four additional IS256 insertion sites: one interrupting the agrB gene (already reported), and three newly reported (Table 7).

Biofilm Formation
Knowing that rbf, sarX and agrB are involved in biofilm regulation [38,39], we studied the biofilm phenotype in these strains. D23C9 showed a significantly higher biofilm formation compared to parental strain D2 (Figure 4), and we speculated this is possibly due to both agrB disruption and extracellular DNA (product of D23C9 increased autolysis [24,30]. Moreover, D2P11 did not differ significantly from D2 in its ability to form biofilm despite the deletion of rbf and sarX genes but tends to display lower OD 570 nm values. However, small differences are not always detected on polystyrene microplates [40]; hence, microscopic changes affecting the three-dimensional biofilm structure should not be disregarded. organic phosphate transporter), SA0620 (secretory antigen ssA-like protein-CHA main peptidoglycan hydrolase), SA0621 (integral membrane protein interacts wit like protein), rbf (araC type transcriptional regulator) and sarX (transcriptional regu This was confirmed by the lack of raw reads mapping to the corresponding region assembled genome of parental strain D2, and by PCR-Sanger sequencing (Figure  plementary Table S1). The presence of IS256 insertion sites between vraG-pitR, SA0621-rbf and sarX-S to both agrB disruption and extracellular DNA (product of D23C9 increased autolysis [24,30]. Moreover, D2P11 did not differ significantly from D2 in its ability to form biofilm despite the deletion of rbf and sarX genes but tends to display lower OD 570 nm values. However, small differences are not always detected on polystyrene microplates [40]; hence, microscopic changes affecting the three-dimensional biofilm structure should not be disregarded.

Discussion
Several single nucleotide polymorphisms (SNPs) have been described in hVISA/VISA strains since their first report, and new studies are still trying to understand their genetic basis [41,42]. Our findings reinforce the diversity of the genetic patterns The horizontal bars show significant differences between individual groups detected by Dunn's multiple comparison test. * p < 0.05; *** p < 0.001. NRS101 (S. epidermidis NRS101), strong biofilm producer; ∆ica (S. aureus Newman ∆ica), ica independent-biofilm producer strain. The bottom picture shows a representative microtiter plate of the biofilm assay for each strain.

Discussion
Several single nucleotide polymorphisms (SNPs) have been described in hVISA/VISA strains since their first report, and new studies are still trying to understand their genetic basis [41,42]. Our findings reinforce the diversity of the genetic patterns observed [11,12,41], but also highlight the important role of INDELs and genomic rearrangements mediated by insertion sequences, particularly IS256 in the emergence of these complex phenotypes.
The genetic changes found here could have resulted from the antibiotic selective pressure (in vivo and in vitro), but not all of them may necessarily have a direct correlation with the observed phenotypes. Evidence showed modifications in the peptidoglycan metabolism for hVISA/VISA strains and those analyzed in this study [11,30]. Mutations in genes rpoB, rpoC, rpoD, pbp2, stk1 and tcaA, shared by the four strains, were also reported in clinical strains with vancomycin reduced susceptibility [11,12,43]. Nonetheless, we identified novel mutations in genes related to peptidoglycan metabolism (Tables 4 and 5) in clinical hVISA strains (D1, D2). These novel mutations or other additional mutations related to different cellular processes (Supplementary Table S2) could have cumulative effects that contribute to hVISA/VISA, and/or oxacillin and daptomycin resistances, and should be experimentally investigated in future studies. However, mutations shared by all strains do not explain the phenotypic differences observed among them (Table 1, Figure 5) [30].
It is worth noting that two INDELs were exclusively found in hVISA strain D1. The stp1 gene was associated with the reduced susceptibility to vancomycin, and the braS gene (alternatively named nsaS or bceS) is part of the braRS two-component system responding to cell envelope stress, which is referred to as graS ortholog (involved in hVISA/VISA phenotype) [34,44]. The in-frame deletion found in D1 (Gln63del) is located on the BraS cytoplasmic domain, next to the HisKA domain in charge of signal transduction and gene expression. We assume the hypothesis that these two genetic changes (Table 3), together with non-synonymous SNPs in genes related to cell wall metabolism (Table 2), could be involved in the vancomycin heteroresistance phenotype of D1. Nonetheless, these two INDELs are absent in the isogenic strain D2, possibly reflecting that different hVISA populations can be selected after vancomycin treatment. However, we recognize that as we sequenced just one colony from each clinical sample, we may have found a small proportion of all possible genetic changes, considering that hVISA strains are nonhomogeneous populations. Nonetheless, our results show genetic changes that might contribute to antimicrobial resistance.
observed [11,12,41], but also highlight the important role of INDELs and genomic rearrangements mediated by insertion sequences, particularly IS256 in the emergence of these complex phenotypes.
The genetic changes found here could have resulted from the antibiotic selective pressure (in vivo and in vitro), but not all of them may necessarily have a direct correlation with the observed phenotypes. Evidence showed modifications in the peptidoglycan metabolism for hVISA/VISA strains and those analyzed in this study [11,30]. Mutations in genes rpoB, rpoC, rpoD, pbp2, stk1 and tcaA, shared by the four strains, were also reported in clinical strains with vancomycin reduced susceptibility [11,12,43]. Nonetheless, we identified novel mutations in genes related to peptidoglycan metabolism (Tables 4 and 5) in clinical hVISA strains (D1, D2). These novel mutations or other additional mutations related to different cellular processes (Supplementary Table S2) could have cumulative effects that contribute to hVISA/VISA, and/or oxacillin and daptomycin resistances, and should be experimentally investigated in future studies. However, mutations shared by all strains do not explain the phenotypic differences observed among them (Table 1  It is worth noting that two INDELs were exclusively found in hVISA strain D1. The stp1 gene was associated with the reduced susceptibility to vancomycin, and the braS gene (alternatively named nsaS or bceS) is part of the braRS two-component system responding to cell envelope stress, which is referred to as graS ortholog (involved in hVISA/VISA phenotype) [34,44]. The in-frame deletion found in D1 (Gln63del) is located on the BraS cytoplasmic domain, next to the HisKA domain in charge of signal transduction and gene expression. We assume the hypothesis that these two genetic changes (Table 3), together with non-synonymous SNPs in genes related to cell wall metabolism (Table 2), could be involved in the vancomycin heteroresistance phenotype of D1. Nonetheless, these two INDELs are absent in the isogenic strain D2, possibly reflecting that different hVISA While the increase in oxacillin susceptibility was already reported in VISA strains [8,9], changes observed in this study are not due to mecA/blaZ mutations or deletions as described before. The sagB gene codes for the major glucosaminidase in charge of glycan chain processing in S. aureus (248aa) [33,45]. Its function, non-redundant despite the presence of other autolysins, is critical for cellular enlargement [33]. The frame-shift insertion shared by D2-D23C9-D2P11 generates a premature stop codon, and the predicted translated protein (148aa) lacks most of the glucosaminidase domain. SagB in vitro-selected mutants were described to display diminished resistance to oxacillin and increased resistance to vancomycin [45,46] as observed in D2 when compared to D1 (Table 1) [24].
No other mutations related to peptidoglycan metabolism/regulation were found between the D1 and D2 (mecA, pbps, blaZ operon and/or cell wall stimulon). Hence, it is most likely that genetic changes in cap5D, stp1, braS and/or sagB genes (related to peptidoglycan metabolism, Table 5) might play a role in the modification of cell wall thickness, pbp2 expression, oxacillin and vancomycin susceptibility [24,30] between these two isogenic strains ( Figure 5). As far as we know, this would be the first report on the acquisition of genetic changes in braS and sagB after the in vivo treatment with vancomycin, and its association with the hVISA phenotype and changes in oxacillin susceptibility in clinical strains. Furthermore, one genetic change seems to be linked to the hVISA to VISA conversion. Both VISA strains selected in independent in vitro assays share not only the increase in oxacillin, vancomycin and daptomycin resistance, slower growth rate and reduced cell diameter (Table 1, Figures 1 and 2), but also the IS256-mediated deletion of a ≈ 8 kb chromosomal region including regulatory genes related to metabolism (pitRA, SA0620, SA0621) and virulence (rbf, sarX) (Figures 3 and 5). Genes related to the inorganic phosphate (Pi) metabolism (including pitRA and phoR) were previously associated with or reported to play a role in the development of vancomycin and daptomycin resistance [47][48][49][50], as changes in the intracellular Pi concentrations can affect the metabolism of DNA, phospholipids, cell envelope (including net positive surface charge), intracellular signalling and stress response [47,51]. Moreover, the expression of pitRA, SA0620, SA0621 and sarX is regulated by walKR and/or graSR operons involved in cell wall metabolism [43,[52][53][54][55], and the development of vancomycin resistance in S. aureus [11]. The rbf gene was also frequently mutated in daptomycin-resistant S. aureus [56]. Together, all these findings highlight a potential relevance of the deleted genomic region for the development of oxacillin, vancomycin and daptomycin resistance in the ST100 genetic background.
Nonetheless, other SNPs or IS256 rearrangements distinguishing D23C9 and D2P11 may also possibly contribute to their vancomycin resistance phenotype. In particular, mutations in genes associated with staphylococcal growth (phoR [36,47] and era [35], Table 5, Figure 5) may also impact on their growth rate and fitness cost (Figure 2), an already described feature of VISA isolates [11]. Interestingly, an INDEL in era was already described as a rare genetic change observed in a laboratory-derived VISA strain belonging to CC5 [50].
WGS is a powerful tool for the detection and surveillance of new AMR genetic determinants, but it has not been widely distributed in clinical laboratories yet, especially in low-middle income countries. Moreover, those who have access to this technology cannot depend solely on molecular assays to reliably detect all hVISA/VISA. Because of the multiplicity of genes involved, genomic approaches trying to establish only a few genetic markers as predictors of vancomycin heteroresistance will lead to an underestimation of the real prevalence, leaving behind new, unexplored hVISA phenotypes.
Nevertheless, studies supplementing whole genome sequences, the gold standard PAP-AUC method and MIC determinations are essential for detecting hVISA/VISA and other AMR phenotypes [50,57,58] until new approaches are developed.
Comparative omics analyses will help clarify the molecular mechanisms involved in the emergence of the hVISA/VISA phenotype in future research. Furthermore, the role of mutations (SNPs and INDELs) and the transposition of insertion sequences on the adaptation of S. aureus under antibiotic selective pressure should be explored. This work provides new evidence of the genetic rearrangements mediated by IS256 transposition after antibiotic treatment, with the potential to impact the AMR and virulence of S. aureus strains.

Strains and Culture Conditions
S. aureus strains D1 and D2 were isolated from a patient with bone and joint infection, before and after 40 days of vancomycin treatment. The in vitro selection of D2-derived mutants (D23C9, D2P11) was performed in two independent assays by serial passage in BHI broth (Britania, Argentina) with increasing concentrations of vancomycin. D23C9 and D2P11 were selected at 9 and 11 µg/mL of vancomycin, respectively. The detailed medical record and the in vitro selection of D2-derived mutants (D23C9, D2P11) was previously described [24]. Strains were grown aerobically on Brain Heart Infusion (BHI) broth and BHI agar (Britania, Argentina) at 37 • C.

DNA Extraction and Whole Genome Sequencing
Genomic DNA was extracted from overnight BHI cultures using the Epicentre Master-Pure Complete DNA and RNA Purification Kit according to the manufacturer's instructions, with the addition of lysostaphin (0.03 µg/µL) in the lysis step with an incubation time of at least half an hour at 37 • C. Shotgun gDNA libraries were prepared and whole genome sequencing (WGS) was performed using the Illumina MiSeq platform (paired end, 250 bp).

Whole Genome Sequencing Analysis
Reads were quality assessed with FASTQC [63], and de novo assembled using SPAdes (v3.9.0) [64]. Contigs less than 500 bp and 70× coverage were discarded. Remaining contigs were annotated using Prokka (1.14.5) [65] and a genus-specific database from RefSeq [66], and they were manually inspected. Mapping and variant calling was carried out using Snippy v3.2 [67] with the following parameters: minimum quality of 30, minimum coverage of 15, minimum proportion of reads which must differ from the reference of 0.75. The genome of S. aureus N315 (CC5, Genebank Accession number BA000018.3) was used as a reference. Alternatively, the assembled genome of the first clinical isolate, D1 (ST100) was used as a reference sequence. All variants (SNPs, INDELs) were manually inspected and visualized with Artemis [68].
ISseeker [79] was used with default parameters to explore the genome in order to detect differences in insertion sequence (IS) content between parental and mutant strains, and also to annotate the flanking edges of IS elements in draft genomes. ISseeker identifies the termini of IS (>97% of identity) at contig edges and annotate flanking regions based on alignment of IS flanks with a reference genome. IS256-mediated insertions/deletions were confirmed by PCR with primers designed for that purpose (Supplementary Table S1). The 8 kb deletions amplified by PCR in VISA strains were sequenced by the Sanger method and analyzed with SnapGene v6.1.2 ® . Genomic comparisons were performed using Clinker [80]. The sequence of the 8 kb deletion was searched against the NCBI Refseq S. aureus complete genomes database using BLAST. All genomes, MGE and genome comparisons with reference sequences of interest were additionally visualized in Artemis and/or ACT [68,81].

Growth Curves
Growth curves were plotted to determine whether the genetic changes were associated with a fitness cost. These assays were performed by triplicate. Fresh culture of each strain (dilution 1/1000) was grown in BHI broth (Britania, Argentina) and incubated at 37 • C and 180 rpm, and OD 620 nm was measured. A growth curve was constructed plotting the OD 620 nm over time.

Transmission Electron Microscopy (TEM)
TEM of exponential phase S. aureus cultures was performed as already described [30]. Cell diameter was measured (30 cells for each strain) at the equatorial plane of each cell Antibiotics 2023, 12, 372 13 of 17 using a 50,000× magnification and images analyzed with ImageJ 1.46r [82]. The results for each strain were expressed as median and interquartile range.

Biofilm Production Assay
Biofilm development was assessed by measuring the accumulation of biomass on the surface of sterile 96-well flat-bottom polystyrene plates (Extragene) following Stepanovic et al. recommendations [83]. Briefly, 200 µL of a 1/100 dilution of a bacterial suspension adjusted to an OD 620 nm = 0.2 (≈108 CFU/mL) in TSB supplemented with sterile 1% glucose was added to wells (6 replicates per strain). Following 24 h incubation at 37 • C, the plate was washed twice with 0.9% NaCl and air-dried for 2 h. The remaining attached bacteria were fixed with 200 µL of methanol 99% (v/v) per well and after 15 min the plates were emptied and air-dried. Afterward the plates were stained for 20 min with 200 µL per well of 0.5% crystal violet. Finally, wells were washed with water, air-dried, the dye was solubilized with 33% acetic acid solution and the OD 570 nm for each well was measured. S. aureus Newman ∆ica (non-ica-dependent biofilm producer) and S. epidermidis NRS101 (prototype biofilm producer) were included in the assay as control strains. Biofilm production was calculated as Final OD 570 nm of a tested strain = average OD 570 nm value of the strain-ODc. ODc = average OD 570 nm of negative control ∆ica + 3SD of negative control.

Statistical Analysis
Growth rates were compared by slope analysis using linear regression. Biofilm production (Final OD 570 nm) and cell diameter values were compared using the Kruskal-Wallis's test, and differences between individual groups were detected by Dunn's multiple comparison test. Analyses were performed using GraphPad prism 5.0 software with a significance level set at p < 0.05 in all cases.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/antibiotics12020372/s1, Table S1: Oligonucleotides used to confirm chromosomal deletion and IS256 insertions. Region amplified is referred to the annotation in S. aureus N315 reference genome. * amplicon was purified and sequenced; Table S2: Genomic features of the genomes analyzed in this study; Table S3: Core mutations present in the strains analyzed in this study found by Snippy core (not including INDELs) after mapping the reads to the S. aureus N315 reference genome.