Comparative Phenotypic and Genomic Features of Staphylococci from Sonication Fluid of Orthopedic Implant-Associated Infections with Poor Outcome

Staphylococcus spp. remain the leading biofilm-forming agents causing orthopedic implant-associated infections (OIAI). This is a descriptive study of phenotypic and genomic features identified in clinical isolates of S. aureus and coagulase-negative Staphylococcus (CoNS) recovered from OIAIs patients that progressed to treatment failure. Ten isolates were identified by matrix-time-of-flight laser-assisted desorption mass spectrometry (MALDI-TOF-MS) and tested for antibiotic susceptibility and biofilm formation. Genotypic characteristics, including, MLST (Multi Locus Sequence Typing), SCCmec typing, virulence and resistance genes were assessed by whole-genome sequencing (WGS). All S. aureus harbored mecA, blaZ, and multiple resistance genes for aminoglycosides and quinolones. All MRSA were strong biofilm producers harboring the complete icaADBC and icaR operon. Seven CoNS isolates comprising five species (S. epidermidis, S. haemolyticus, S. sciuri, S. capitis and S. lugdunensis) were analyzed, with mecA gene detected in five isolates. S. haemolitycus (isolate 95), and S. lugdunensis were unable to form biofilm and did not harbor the complete icaADBCR operon. High variability of adhesion genes was detected, with atl, ebp, icaADBC operon, and IS256 being the most common. In conclusion, MRSA and CoNS isolates carrying genes for biofilm production, and resistance to β-lactam and aminoglycosides are associated with treatment failure in OIAIs.


Introduction
Staphylococci are the leading pathogens of orthopedic implant-associated infections (OIAIs), including those related to biofilm formation. S. aureus and S. epidermidis are the most commonly isolated species from OIAI reaching rates above 50% [1]. Methicillinresistant Staphylococcus aureus (MRSA) poses a threat to the management OIAI due to its therapeutic limitations. Likewise, the coagulase-negative Staphylococcus (CoNS) group, including methicillin-resistant S. epidermidis (MRSE) also plays an important role in OIAIs [2]. Other pathogens of this group with clinical importance are emerging, such as S. capitis, S. haemolyticus, S. hominis, S. warneri, and S. lugdunensis [3,4].
Most OIAIs start with the introduction of pathogens colonizing the skin and mucous membranes during the surgical procedure. According to the "race to the surface" theory,

Study Population and Sample Collection
At a tertiary university center specializing in musculoskeletal infections, clinical, microbiological and antibiotic therapy data were collected from 10 patients with a confirmed diagnosis of OIAI who evolved to poor outcome (failure) after undertaking surgical and antibiotic treatment. Orthopedic surgeries included arthroplasties, osteosynthesis (plate, screws, and intramedullary nail) for stabilization of open and closed fractures of long bones, and spinal osteosynthesis for correction of deformities or degenerative spine disease. The clinical and microbiological diagnosis of OIAI was fulfilled according to the definition of prosthetic joint infection (PJI) and fracture-related infection (FRI) published elsewhere [18,19]. Briefly, the microbiological diagnosis of FRI and PJI requires at least two positive tissue cultures out of at least 4 samples of tissue collected aseptically, with the same low virulence microorganisms (coagulase-negative Staphylococci) identified phenotypically. Therefore, the identified CoNS is considered pathogenic when the organism is found in at least 2 different culture samples, including sonication fluid cultures [17]. All patients underwent surgical revision with the removal of the orthopedic implants, which were then submitted to the sonication technique for microbial diagnosis. After surgical removal of the implants in the operating room, each device was immediately placed in a sterile polypropylene container, hermetically closed, identified, and sent to the microbiology laboratory of the Institution. This study was reviewed and approved by the local ethics committee (n. 3.622.166, on 4 October 2019).

Sonication and Microbiological Methods
The implants were sonicated according to the technique of Trampuz et al. [20] and modified by Yano et al. [17]. Briefly, the sonication technique consisted of vortexing the sterilized solid polyethylene containers with the implants and 50 to 250 mL (depending upon the device width) of Ringer solution for 30 s using a Vortex-Genie 2 (Scientific Industries, Inc., Bohemia, NY, USA), and then treat it in an ultrasonic bath (BactoSonic; Bandelin GmbH, Berlin, Germany) for 5 min at a frequency of 40 ± 2 kHz and power density of 0.22 ± 0.04 W/cm 2 , followed by another 30 s of vortexing. Subsequently, a centrifugation step at 600× g (2500 rpm) for 5 min was carried out, to concentrate the sonication fluid. The supernatant was aspirated, leaving 0.5 mL (100-fold concentration), and aliquots of 0.1 mL of concentrated sonicate fluid (SF) were then plated on blood agar (Probac do Brasil, São Paulo, SP, Brazil) and incubated for 18-24 h at 37 • C. Colonies of isolated microorganisms cultured on plates were quantified (number of colony-forming units [CFU]/mL of sonication fluid). Due to the addition of a sonicate fluid (SF) concentration step, counts above the cut-off point of 50 CFU/mL were considered positive and used for optimal sensitivity and specificity analyses. Staphylococcus spp. were identified by colonial morphology, Gram staining features, the catalase test, and the coagulase test (rabbit plasma), and species were confirmed using matrix-assisted laser desorption/ionization time-of-flight mass spectroscopy (MALDI-TOF MS) (Bruker, Billerica, MA, USA).

Species Identification by MALDI-TOF MS
Ten non-duplicated clinically important isolates, phenotypically characterized as Staphylococcus spp., were identified at the species level by MALDI-TOF MS. The extraction of total proteins from each isolate was performed according to the manufacturer's standards (Bruker Daltonics, Billerica, MA, USA). Spectra were obtained in triplicate for each pathogen, using the Microflex LT mass spectrometer (Bruker, Billerica, MA, USA)., and bacterial identification was based on spectra comparison with those present in the MALDI Biotyper 3.3 software database (Bruker, Billerica, MA, USA ). According to this software, the value ≥ 2.3 indicates that genus and species identification is reliable; a value between 2.0 and 2.29 indicates that the identification of the genus is reliable and the identification of the species is probable, and between 1.7 and 1.9 indicates that gender identification is likely. Values lower than 1.69 indicate that the identification is unreliable and must be repeated.

Antibiotic Susceptibility Tests
The antibiotic susceptibility profiles of the ten isolates of Staphylococcus spp. were evaluated by the Kirby-Bauer disk diffusion technique and for the evaluation of MICs, microdilution was performed in broth and E-test based on the criteria and recommendations of the Brazilian Committee on Antibiotic Susceptibility Testing-BrCAST and the European Committee on Antibiotic Susceptibility Testing-EUCAST (BrCAST (http://brcast.org.br/ Accessed in February/March 2020)/EUCAST (https://www.eucast.org/ Accessed in February/March 2020)). Quality control was performed with the standard strains S. aureus ATCC ™ ,25923, S. aureus ATCC ™ 29213 and Enterococcus faecalis ATCC ™ 29212.

Inoculum Preparation
After being seeded on Blood agar for about 18 h to ensure the purity of the samples, with the aid of the 10µL seeding loop, around 3 to 5 colonies isolated from each sample were transferred to tubes containing 3 mL of saline solution at 0.85%. The bacterial suspension was homogenized and the turbidity measured in a digital turbidimeter (Baxter ® , Sacramento, CA, USA), to obtain a bacterial concentration of around 1.5 × 10 8 of colony-forming units (CFU)/mL corresponding to 0.5 of the McFarland scale.

Broth Microdilution Test
Minimum inhibitory concentration (MIC) determination for vancomycin was performed by broth microdilution (BrCAST/EUCAST, 2020). Solutions were prepared in Müeller-Hinton (MH) broth with adjusted concentrations of calcium (Ca 2+ ) and magnesium (Mg 2+ ) cations (Oxoid). Then, a final volume of 100 µL of each dilution was dispensed into 96-well polystyrene microplates. In each microdilution plate, column 11 was used as a bacterial growth control and column 12 was used as a sterility control for the medium. After the inoculum preparation, an additional dilution was performed to obtain a 10 × 10 8 CFU/mL inoculum. A volume of 100 µL of this bacterial suspension was added to the microdilution plates already containing 100 µL of the antibiotic solution (dilution 1:2). Additionally, the test Sensititre TM Gram Positive MIC Plate (Thermo Scientific TM , Delaware, USA) was used, which contains 16 antimicrobials, including oxacillin, levofloxacin, tigecycline, linezolid and gentamicin. A 10 µL aliquot of the 0.5 McFarland suspension was transferred to a tube containing 11 mL of MH cation broth mixed and slowly vortexed to obtain a final bacterial concentration around 2.5 × 10 8 CFU/ ml. After homogenization, 100 µL of this bacterial suspension was inoculated into each well of the microdilution plate in broth containing the lyophilized antimicrobials, according to the manufacturer's recommendations. The plates were incubated in an incubator at 35 • C ± 2 • C for 18-24 h; the MICs were determined as the lowest concentration of antimicrobial capable of inhibiting bacterial growth.

Episillometric Test (E-Test)
To evaluate the MICs of rifampicin and tetracycline, the episillometric test with gradient tape (E-test ® , BioMérieux, France) was performed, due to the lack of these salts in our laboratory during the COVID-19 pandemic period. With the aid of a swab, the sample containing 1 to 2 × 10 8 CFU/mL was seeded on the surface of the agar plate. After an average of 15 min, the Etest ® strips were dispensed onto the MHagar, in 150 mm diameter plates. After the incubation period, the MIC determination was read as the intersection point between the Etest ® strip and the zone of inhibition of the growth of the microorganism.

Quantitative Assay of Biofilm Formation on Abiotic Surfaces
The assay for quantification of biofilm mass was performed as described by Stepanović et al. [21] with minor modifications. Cultures were grown for 24 h in Trypticase Soy Broth (TSB) and then bacterial suspensions were sub-cultivated in fresh TSB containing 1% glucose (TSBG) to approximately 1 × 10 8 CFU/mL (0.5 McFarland scale) in a 96-well polystyrene plate with a flat bottom, where they were incubated for 24 h at 37 • C. The plates were washed with 0.9% saline solution three times and fixed with methanol PA, being later stained with 2% Hucker's Crystal Violet solution. After this step, the plates were washed with running water and allowed to dry at room temperature. The absorbance reading was performed using the Synergy H1 Hybrid Multi-Mode Microplate Reader spectrophotometer (Biotek, Winooski, VT, USA). The optical density (OD) was measured at 550 nm. All isolates were tested in triplicate in three independent experiments. The results were evaluated using the scale described by Stepanovic et al. [21], by which the isolates can fit into four categories: non-biofilm producer, weak biofilm producer, intermediate biofilm producer, and strong biofilm producer. Based on the values of OD and the cut-off value (OD c), defined as three standard deviations (SD) above the mean OD of the negative control: OD c = mean OD of the negative control + (3 × SD of the negative control). The strength of biofilm production of each isolate was calculated as follows: OD ≤ OD c = non-biofilm producing, OD c < OD ≤ 2 × OD c = weak biofilm producer, 2 × OD c < OD ≤ 4 × OD c = intermediate biofilm producer and 4 × OD c < OD = strong biofilm producer.

Whole-Genome Sequencing and Assembly
Ten isolates of Staphylococcus spp. were chosen (from OIAI patients with treatment failure) for the execution of complete genome sequencing. DNA from the isolates was extracted used using the QIAamp DNA minikit (Qiagen, Hilden, Germany), according to the manufacturer's instructions, and quantified in a Qubit Ò 3.0 fluorometer (Ther-moFisher Scientific, Waltham, MA, USA). DNA libraries were prepared with the Nextera XT library prep kit (Illumina, San Diego, CA, USA) and sequenced on the Illumina HiSeq 2500 platform. The assembly and annotation of the genomes were carried out on the PATRIC (https://www.patricbrc.org/ Accessed in June 2020) platform version 3.6.12 using Unicycler v0.4.9 as a pipeline, the SPAdes v3.11.1 program optimizer and the RAST tool program [22]. The genomes were analyzed following pipelines available on the Center for Genomic Epidemiology (CGE (https://www.genomicepidemiology.org/ Acessed in June 2020)) research website, which includes a cassette of resistance (ResFinder) and virulence (VirulenceFinder) genes, allowing identification of the staphylococcal chromosome mec (Sccmec) (SccmecFinder), spa typing in S. aureus (SpaTyper) and Multi Locus Sequence Typing (MLST). In addition, the CARD RGI program was used for the analysis of resistance genes. Likewise, virulence genes were also searched in the Virulence Factors of Pathogenic Bacteria (VFDB) platforms. Manual curation was also carried out using SnapGene and NCBI BLAST tool to confirm the results indicated by the pipelines. Sequence Type Confirmation was done in the Public Molecular Typing Database (PubMLST), and taxonomic confirmation in the Ribosomal Multilocus Sequence Types Database (rMLST/ Institut Pauster) [23].

Multi Locus Sequence Typing (MLST) Analysis
The GC content and genome size of the 10 Staphylococcus spp. isolates ranged from 32% to 33.7%, 2,407,407 bp to 2,894,360 bp, respectively (Table S1). The classification of S. aureus isolates according to the description of ST (sequence typing) is based on the nucleotide sequence at seven loci (arcC, aroE, GlpF, gmk, pta, pi and yqiL). The three S. aureus isolates sequenced belonged to two distinct STs, ST5 (215 and 260) and ST105 (371), all grouped into the same clonal complex (CC)-CC5, revealing a common origin. The spa type t002 was identified in two S. aureus isolates (260 and 371). The two S. epidermidis isolates showed two different types of STs, the 216-isolate belonging to ST2/CC2 and the 403-isolate belonging to a new ST similar to ST183. S. haemolyticus 53 and the 95 belonged to the ST9 and ST3, respectively. S. lugdunensis belonged to ST2/CC2. It was not possible to establish the STs of the isolates belonging to S. sciuri (226) and S. capitis (160), due to the unavailability of an MLST scheme in the databases used (PubMLST and CGE) ( Table 2).

Phenotypic Results
Antibiotic sensitivity patterns, phenotypic and genotypic characteristics of MRSA and coagulase-negative Staphylococci isolates are described in Tables 3 and 4, respectively. Briefly, the three strains of S. aureus were resistant to the cefoxitin disk (MRSA) and harbored mecA gene as confirmed by WGS. The MRSA (215) was resistant to clindamycin, levofloxacin, and gentamicin, whereas 260 expressed higher MICs for levofloxacin and gentamicin and S. aureus 371 for clindamycin and levofloxacin. All strains were sensitive to sulfamethoxazole + trimethoprim, linezolid and rifampicin. Vancomycin MICs (broth microdilution) of S. aureus ranged from 0.5 to 1.0 µg/mL. Additional MICs values for oxacillin, clindamycin, erythromycin, levofloxacin, linezolid, gentamicin, tigecycline, tetracycline and rifampicin are summarized in Table 3.   Five of the seven isolates in the CoNS group (S. epidermidis (216), S. capitis (160) and S. sciuri (226), and two isolates of S. haemolitycus (53 and 95) showed disc resistance to cefoxitin and also harbored the mecA gene. Importantly, the S. epidermidis 216 was a multidrug-resistant strain (MDR) with higher MICs for clindamycin, levofloxacin, gentamicin, trimethoprim, rifampicin and vancomycin (MIC = 4.0 µg/mL). Conversely, the S. epidermidis 403 was only resistant to gentamicin. While S. haemolitycus 53 was resistant to tetracyclines, and gentamicin, isolate 95 was a multidrug-resistant strain (MRSH) with higher MICs also for clindamycin, levofloxacin, tetracyclines, gentamicin, and trimethoprim. S. capitis (160) was another MDR strain showing resistance to clindamycin, erythromycin, levofloxacin, tigecycline, and gentamicin. Resistance to clindamycin, and levofloxacin was observed in S. sciuri (226) and S. lugdunensis (167) was sensitive to all antibiotics tested. The MICs values for CoNS are summarized in Table 3.

SCCmec Analysis and Detection of Resistance Genes by WGS
The presence of the mecA gene and characterization of SCCmec were evaluated by genome sequencing. CGE pipelines were defined for a lower limit of 90% identity and a lower bound of 60% gene sequence length. Overall, the mecA gene was absent only in the S. epidermidis (403) and S. lugdunensis (167) isolates. The S. aureus isolates 215, 260 and 371 contained the SCCmec type I (1B), II (2A) and V(5C2), respectively. Based on gene prediction for CoNS isolates, the SCCmecs detected with the highest homology were SCCmec type III for the S. epidermidis isolate (416), SCCmec type I and V for S. haemolyticus isolates (53 and 95), SCCmec type III for S. sciuri (226) and SCCmec type V for S. capitis (160) (Tables 2 and 4).
In general, a higher frequency of antibiotic resistance genes was identified for S. aureus and S. epidermidis isolates. Regarding the genotypic profile of antibiotic resistance, all S. aureus isolates (215, 260 and 371) had the mecA and blaZ genes, which confer phenotypic resistance to β-lactam antibiotics. Additionally, several genes of aminoglycoside modifying enzymes (AMEs) were found, such as aph (3 )-III, aac (6 )-aph (2"), ant (9)-Ia, ant (4)-Ib, and ant (6)-Ia, and for quinolones, 260 also carried resistance genes for tetracycline, quinolones and trimethoprim (dfrC). The ermA gene that confers cross-resistance to macrolides, lincosamides and streptogramin B was also detected. Several mutations related to the quinolone resistance profile were observed in the S. aureus isolates (CC5). Six mutations in the gyrA gene, and two in the parC gene.

Detection of Biofilm-Forming Genes and Adhesins (MSCRAMMs)
As for the biofilm formation capacity, the complete icaADBCR operon (intracellular adhesins) was found in all isolates, while IS256 was found in the 215 and 260 isolates.
Multiple toxin genes were found in S. aureus isolates, as summarized in Table S2. No genes encoding pvl and tsst-1 were found.
The genomic analysis of CoNS isolates (216, 403, 53, 95, 160, 226, and 167) allowed the detection of high variability in adherence genes, the most common genes being atl, ebp, and sdr. Except for S. haemolitycus, which does not harbor the ica locus the S. sciuri, S. epidermidis (216, 403) and S. capitis isolates carried the icaADBCR operon. Only the S. epidermidis isolates (216 and 403) contained the sdrF, sdrG and sdrH (Ser-Asp-rich fibrinogen binding proteins) genes. The S. haemolitycus isolates (53 and 95) contained only two adhesion genes, atl and ebp. In S. capitis (160) it was possible to identify the atl, ebh, and ebp genes, in addition to the ica ADBCR operon, the IS256 and, sdrH genes, even though this isolate was a weak biofilm former. Interestingly, in S. sciuri (226) was a strong biofilm former, but lack the ica operon and the presence of any adhesive gene. In S. lugdunensis (SS167) the genes atl, ica A, icaB, icaC and icaD, were identified. Few virulence and enzymes producing genes were identified in CoNS isolates, the most common being nuc (thermonuclease), hlb (beta-hemolysin), lip (lipase) and sspA (serine V8 protease). The S. epidermidis isolates (216 and 403) contained the sspA, sspB, geh, lip, nuc and hlb genes. In S. capitis (160) and S. lugdunensis (167) only the nuc, hlb genes were identified. In the S. haemolitycus isolates (53 and 95), only the lip, and nuc genes were identified. In the S. sciuri isolate (226), only the sspA gene was identified. Table S3 summarizes the genes of virulence detected in the sequenced CoNS.

Discussion
In this study, we analyzed the complete genome sequencing of ten Staphylococci isolates from the SF of implants removed from patients with poor outcome (failure) following surgical and antibiotic treatment of OIAI (PJI and FRI). The main focus was to provide insights into the association of clinical S. aureus and CoNS isolates causing OIAIs with poor outcomes and specific phenotypic and genomic signatures. A growing body of evidence has pointed towards specific bacterial resistance, biofilm formation, adhesins and toxins genes of S. aureus and CoNS isolates [24].
The three MRSA isolates showed similar phenotypic and genomic characteristics concerning the pattern and presence of many antibiotic resistance (AR), biofilm formation and virulence genes, and genomic similarity (CC5) revealing a probable common origin. Interestingly, our results confirmed a previous study that identified a trend toward a higher incidence of MRSA, gentamicin resistance and hemolysis activity associated with OIAI cases [25].
The species of CoNS (S. epidermidis, S. haemolyticus, S. lugdunensis, S. sciuri, S. capitis) showed an expected diversity in resistance and virulence patterns, indicating the complexity of a growing number of Staphylococci causing OIAIs. S. aureus and CoNS share the skin and mucous membranes colonization niches, enabling horizontal gene transmission (HGT) of several genes along with mobile genetic elements encoding for AR, biofilm formation and virulence [26,27]. Therefore, this study adds to the importance of evaluating the phenotypic and genomic characteristics of clinical isolates of CoNS together with S. aureus, since there is a paucity of genomic studies on emerging species of CoNS likely to cause OIAI [28]. Indeed, despite the remaining predominance of S. aureus and S. epidermidis in the classical etiology of OIAIs, recent studies depict the increasing identification of other Staphylococci, including S. haemolyticus, S. saprophyticus, S. lugdunensis, S. sciuri, S. capitis, and S. hominis [27,28]. In a study by Lourtet-Hascoët et al. [29], of 215 CoNS PJI, 40% of the species were not S. epidermidis, with S. capitis and S. lugdunensis being the commonest.
Methicillin resistance associated with the expression of the mecA gene, which encodes the altered penicillin-binding protein PBP2a, may be an important factor in treatment failure in OIAI [30,31]. In the study by Post et al. [32], methicillin resistance was associated with OIAI poorer outcomes. In the current study, all S. aureus isolates were MRSA and belonged to CC5. The prospective cohort study published by Muñoz-Gallego et al. [33] shows a high frequency of MRSA, with 80% CC5 in PJI associated with treatment failure. In fact, MRSA isolates of hospital origin belonging to CC5 and CC8 are closed related, and spreads globally including within Latin American countries [17,34,35]. Likewise, in the CoNS group, only one strain of S. epidermidis (403) and S. lugdunensis did not harbor mecA gene. Despite its known pathogenicity resembling S. aureus life-threatening infections such as bacteremia endocarditis and PJI, carriage mecA gene by S. lugdunensis is unlikely as they remain remarkably sensitive to most antibiotics, particularly β-lactams, contrary to other members of CoNS group [36]. Although unlikely to draw any strong conclusion, our preliminary results corroborate studies that mecA gene carriage may be used as a reliable marker for Staphylococci OIAIs poor outcome [27].
Beyond methicillin resistance, MDR was a common feature of all of our S. aureus and S. epidermidis isolates. Except for S. epidermidis (403) which carried resistance genes for quinolones, rifampicin, trimethoprim, and fosfomycin, but phenotypic resistance was observed only against gentamycin. All S. aureus and S. epidermidis (216) isolates were resistant to at least three different classes of antibiotics, and resistance to gentamicin was observed in 70%, which was associated with the presence of aac (6 )-Ie-aph (2")-Ia, aph (3 )-IIIa, aadD, ant (4 )-Ib, ant (9 )-Ia genes (aminoglycoside-modifying enzymes). Studies with genomic analysis of Staphylococcus spp. causing OIAIs have identified a strong relationship between the presence of aminoglycoside resistance genes, biofilm formation genes, and treatment failure of OIAIs [25,32]. According to Arciola et al. [11], up to 40% of S. epidermidis and 32% of S. aureus isolated from postoperative OIAIs were resistant to gentamicin. The direct relationship between aminoglycoside resistance and treatment failure in Staphylococci IOAI is becoming increasingly clear in recent literature, and the present study pinpoints this feature [37]. Notably, HGT occurring among staphylococcal species has been frequently evidenced, in which CoNS act as reservoirs of genes with many resistant phenotypes related to multi-resistant genes located on mobile genetic elements [27].
Analysis of SCCmec types in S. aureus isolates identified types I, II, and V. SCCmec I and II are large genetic elements capable of conferring resistance to many antibiotics and are found in hospital-sourced MRSA isolates limiting antibiotic therapy options. On the other hand, SCCmec V are minor genetic elements, harboring few resistant genes and are generally associated with MRSA isolates of community origin [35,38]. Interestingly, the three MRSA isolates were recovered from sonication fluid of osteosynthesis from patients with treatment failure associated with FRI (260 and 371) and spinal infection (215) in which ciprofloxacin had been employed empirically, while cefazolin was used to patients with S. epidermidis (216) and S. lugdunensis infection, and vancomycin to the others. We argue that the antibiotic selective pressure may have influenced the antibiotic resistant patterns and also in the patient's poor outcomes.
The isolate 260 with SCCmec V was sensitive to clindamycin and quinolone, unlike the isolates with SCCmec I and II (215 and 371). The reported cases of communityacquired MRSA (CA-MRSA) bone infections are infrequent and predominantly described in hematogenous osteomyelitis of long bones in children younger than two years of age [39].
In the sequencing data of our S. aureus, S. epidermidis, S. lugdunensis and S. capitis isolates, the icaADBC operon was found. Besides, IS256 gene was also identified in the S. aureus (215, 260), S. epidermidis (216), S. haemolyticus (95) and S. capitis. ST2 clonal type always bears the insertion sequence IS256 and ica genes, which are implicated in biofilm production [24]. We hypothesize that the co-presence of both virulence factors in addition to mecA gene may be labeled as possible genomic makers also for non-epidermidis CoNS causing OIAI with poor outcome. S. epidermidis and S. lugdunensis harboring, respectively, ica and IS256 genes, have been identified in pathogenic strains associated with severe infections such as endocarditis and OIAIs [36]. Moreover, the presence of these three markers (icaA, mecA and IS256) has been significantly associated with CoNS invasive nosocomial strains [40].
In our study, the S. epidermidis (216) isolate belonging to the ST2 (CC2), showed a strong biofilm production, harbored multiple genes associated with adherence, to host matrix binding proteins (atl, ebh, clfA), to biofilm formation (icaADBC locus and IS256 gene), resistance to rifampicin, and had a MIC of 4 µg/mL for vancomycin. Interestingly, it was isolated from a patient with a history of diabetes and a bone tumor, with a closed fracture of the femur that progressed to FRI who had previously received vancomycin and rifampicin as chronic suppression therapy for 12 weeks. Epidemic isolates of CoNS ST2 and ST5 have been isolated worldwide, including in Brazil, and are likely related to the presence of the cfr gene that confers resistance to linezolid, are associated with musculoskeletal infections (PJI, FRI) [41,42]. Sanchéz et al. [1] reported that pathogenic isolates of S. epidermidis causing PJI often belong to the ST2 clone and carry genes such as the ica operon, IS256, sdrF, bhp and mecA. Indeed, the presence of the ica operon in S. epidermidis isolates has long been associated with biofilm production [42].
Furthermore, in the genomic analysis of the non-epidermidis CoNS isolates (S. lugdunensis, S. capitis, S. lugdunensis, and S. sciuri) we aimed at searching for possible virulence markers in the OIAI setting due to the scarcity of literature. Despite the great variability in the presence of genes that express adherence proteins (MSCRAMMs) and other virulence traits, the most frequently identified genes were atl, ebp, hlb and nuc. However, in the S. sciuri isolate we were unable to find any genomic marker towards adhesion or biofilmforming gene, despite harboring mecA gene and having shown a strong capacity for biofilm formation. Interestingly, the animal origin S. sciuri species group have been proposed as the origin and/or reservoir of the S. aureus mecA gene, which has been increasingly identified as a OIAI etiology [3,43]. This isolate was identified in a patient with bone tumor who underwent endoprosthetic reconstructions for lower limb savage and evolved to infected endoprostheses. Treatment consisted of multiple surgeries of debridement, antibiotics as suppressive therapy and implant retention. Future genomic studies are needed to better understand this pathogen in OIAIs.
S. haemolyticus isolates carried few adherence genes (atl and ebp). They have been frequently associated with MDR strain outbreaks in the hospital setting including neonatal units, but currently are the second most frequent CoNS in implant-associated infections. They are reported to have the highest level of antibiotic resistance, which seems to be the main genomic marker for this CoNS [27,43]. In fact, our S. haemolyticus strains carried multiple AR traits, including resistant genes for β-lactam, aminoglycosides, tetracyclines, quinolones, and clindamycin. Nevertheless, few publications have completely assessed the virulence traits in this species, warranting future studies [24].
Conversely, our S. capitis is an MDR isolate (resistance to β-lactam, AME, MLS b and quinolones) and had several biofilm-forming (complete ica operon and IS256), adherence (atl, ebh, ebp, sdrE) and other virulence (hlb and nuc) genes. The genomic analysis of this isolate adds importance due to the scarcity of clinical reports on OIAIs caused by S. capitis. Recently, genomic sequencing analysis of a large amount of S. capitis collection including OIAI strains was carried out by Swedish researchers. They identified a robust biofilm-forming ability and MDR traits and made speculations regarding the in-hospital dissemination of this pathogen that has been classically associated with neonatal intensive care unit sepsis [44]. MDR is likely a crucial issue in S. capitis infection, in which our isolate expressed an MDR phenotype and a vancomycin MIC = 2.0 µg/mL.
Despite harboring no resistance genes, our S. lugdunensis isolate (ST2/CC2) had icaABC operon and atl, sdrC, hld and nuc genes. S. lugdunensis produces a fibrinogenbinding protein linked to the bacterial cell wall that has been involved with endocarditis and persistent bacteremia in vitro studies [36]. The presence of ica operon and atl gene (autolysin) in biofilm formation may play a role in the OIAI poor outcomes in patients.
We acknowledge that the analysis of very small number of isolates is a strong limitation, and few conclusions can be drawn. The importance of this type of research would be at assessing specific traits including bacterial resistance, biofilm formation, adhesins and toxins genes of S. aureus and CoNS isolates that would trigger preventive measures to be taken at the bed site before carrying out elective surgeries. Moreover, the molecular epidemiology of non-epidermidis CoNS implant-associate infection isolates need more attention. The literature regarding this type of research has been focusing on S. aureus and S. epidermidis, but scarce among other CoNS. Besides, back no more than a decade ago, new microbiological technologies including the use of MALD-TOF MS was a tool to be applied only to few universities' microbiological laboratories. However, it is currently widespread available, including in developing countries' public hospitals. We also envision that metagenomics and WGS technology may assume the same importance in the near future, which may change the way we apply microbiology at the bed site to prevent, diagnosis and treat bacterial infections. The ordinary medical strategies that have been carried out today in the daily basis may have little impact to the genomic analysis of pathogens storage at the microbiological laboratory. However, in many implant-associated and orthopaedic infections, the phenotypic bacterial resistance identification does not help either in the management of this infection. An important number of these infections is due to the presence of pathogens growing into biofilms, in which the traditional approach of prescribing antibiotics based upon phenotypical susceptibility testing to determine the adequate therapeutic approach is unhelpful. Understanding the bacterial behavior of emerging pathogens in this medical situation seems to be crucial. Moreover, the retrospective analysis of patients' records accounts for the unknown exact number of tissue samples collected per patients. However, the isolates came from sonication fluid from implants taken from patients with PJI and FRI, conferring a broad spectrum of IOAIs. Finally, we were unable to carry out rifampicin and tetracycline sensitivity using the gold standard technique of broth microdilution due to the lack of these salts in our laboratory during the COVID-19 pandemic period. To overcome this situation, the episillometric test with gradient tape (E-test®) was carried out. Genomic evaluation of S. aureus isolates together with CoNS may seem confusing, but these comparisons with the clinical outcomes can draw attention to common bacterial sources and the transfer of antibiotic resistance genes between these species.

Conclusions
In conclusion, the genomic analysis of Staphylococci allowed elucidation of MRSA and CoNS features that are associated with treatment failure in OIAIs. Internationally spreading isolates of MRSA and S. epidermidis are associated with OIAI in Brazil. The MRSA isolates showed genomic similarity, revealing a probable common origin, and harboring multiple resistance, biofilm formation, and virulence genes. Our findings corroborate a probable association between isolates harboring resistance genes to β-lactam and aminoglycosides with treatment failure in OIAIs. In addition to S. epidermidis, there is variability in CoNS isolates with distinct genomic features that require further attention.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/microorganisms10061149/s1: title; Table S1: Genome assembly metrics of 10 Staphylococcus genomes; Table S2: Virulence genes of S. aureus isolates sequenced by WGS; Table S3: Virulence genes of CoNS isolates sequenced by WGS.  Informed Consent Statement: Patient consent was waived due the research involves no more than minimal risk to the subject because is a retrospect observational study.

Data Availability Statement:
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https: //www.ncbi.nlm.nih.gov/ Accessed on 15 October 2021, BioProject number PRJNA736948.