Whole Genome Sequencing and Molecular Epidemiology of Clinical Isolates of Staphylococcus aureus from Algeria

Staphylococcus aureus is an important pathogen responsible for various healthcare- and community-acquired infections. In this study, whole genome sequencing (WGS) was used to genotype S. aureus clinical isolates from two hospitals in Algeria and to characterize their genetic determinants of antimicrobial resistance. Seventeen S. aureus isolates were included in this study. WGS, single-nucleotide polymorphism (SNP)-based phylogenetic analysis, in silico multilocus sequence typing (MLST), spa and staphylococcal cassette chromosome mec (SCCmec) typing and in silico antimicrobial resistance profiling were performed. Phenotypic antibiotic susceptibility testing was performed using the Vitek 2 system and the disk diffusion method. The isolates were separated into sequence types (STs), with ST80 being predominant; five clonal complexes (CCs); four spa types (t044, t127, t368, t386); and two SCCmec types (IVc and IVa). Whole genome analysis revealed the presence of the resistance genes mecA, blaZ, ermC, fusB, fusC, tetK, aph(3′)-IIIa and aad(6) and mutations conferring resistance in the genes parC and fusA. The rate of multidrug resistance (MDR) was 64%. This work provides a high-resolution characterization of methicillin-resistant S. aureus (MRSA) and methicillin-susceptible S. aureus (MSSA) isolates and emphasizes the importance of continuous surveillance to monitor the spread of S. aureus in healthcare settings in the country.


Introduction
S. aureus is the leading cause of nosocomial infections worldwide [1]. Although it is a normal resident of the skin and mucous membranes of humans and animals, S. aureus can become an opportunistic pathogen by deploying a plethora of virulence factors to cause a variety of infections, ranging from mild skin and soft tissue infections to severe and life-threatening diseases, such as toxic shock syndrome and sepsis [2]. This strong potential of S. aureus to cause diseases is aggravated by its propensity to acquire resistance to multiple antibiotics, limiting the therapeutic options against this pathogen [3].
MRSA is a group of S. aureus strains that have developed resistance to methicillin and to the majority of the β-lactam antibiotics following the acquisition of a mecA gene. This gene which resides on a complex mobile genetic element known as "Staphylococcal Cassette Chromosome" mec element (SCCmec) encodes a 78 kDa penicillin-binding protein (PBP2a) that has a low affinity for β-lactams [3]. MRSA is recognized as one of the leading pathogens responsible for nosocomial and community-associated infections worldwide [4]. Effective MRSA control and prevention strategies in the healthcare system as well as in the community depend on accurate characterization of circulating MRSA clones and identification of their reservoirs and sources of transmission.
Although the prevalence and molecular epidemiology of MRSA in Europe and North America have been extensively documented [5], comparatively, data available from North

Figure 1.
Core genome maximum likelihood phylogenetic tree of the 17 S. aureus isolates and the reference genome of the S. aureus strain MSSA476, visualized using Phandango [27]. The phylogeny tree on the left is linked to the metadata in the middle: STs, CCs, SCCmec types, Spa types (represented by various colors as indicated in the legend on the right); presence (dark blue) and absence (grey) of PVL, TSST and antibiotic resistance gene/mutation. Core genome maximum likelihood phylogenetic tree of the 17 S. aureus isolates and the reference genome of the S. aureus strain MSSA476, visualized using Phandango [27]. The phylogeny tree on the left is linked to the metadata in the middle: STs, CCs, SCCmec types, Spa types (represented by various colors as indicated in the legend on the right); presence (dark blue) and absence (grey) of PVL, TSST and antibiotic resistance gene/mutation.

Phylogenetic Analysis
Core genome SNP-based phylogeny revealed that the SNP differences among the 17 S. aureus isolates ranged from 0 to 60210. Phylogenetic reconstruction grouped the isolates into five separate clusters which matched their respective STs (Figure 1).

Phenotypic Antimicrobial Susceptibility Testing
The 17 isolates were tested for susceptibility against a panel of 21 antibiotics belonging to 13 classes. All S. aureus isolates were susceptible to VAN, QD, SXT, LZD, TEC, TGC and NIT.
Seven S. aureus isolates (41%) carried the methicillin resistance gene mecA and were therefore considered as MRSA. The ermC gene encoding a 23S rRNA adenine-N-6 methyltransferase that confers resistance to the macrolide-lincosamide-streptogramin B (MLSB) class of antibiotics was the most frequent antibiotic resistance genetic determinant identified in our collection of isolates, being present in eight isolates (47%). The blaZ gene encoding a type C β-lactamase was present in seven (41%) S. aureus isolates. The fusC or fusB genes encoding proteins that block the binding of fusidic acid to EF-G responsible for fusidic acid resistance were identified in four isolates (23.5%), whereas fusidic acid resistance conferred by a mutation H457Q in fusA was detected in one isolate (5.8%), 15RN (ST80) (Figure 1).

Discussion
In this study, WGS was used to determine the phylogeny and molecular characteristics of 10 MSSA and 7 MRSA isolated from clinical samples. The percentage of MRSA among S. aureus isolates in our study (41.1%, 7/17) was in line with that reported in North African countries, 31% in Libya, 45% in Algeria and Tunisia and 52% in Egypt [7]. The prevalence of MRSA in Europe varied from 2% in the Netherlands to 58% in Italy [28,29], and that in Asia varied from 22.6% (India) to 86.5% (Sri Lanka) [28], with an average of 25% in both continents [28,30].
The results presented in this study revealed diversity among our collection of S. aureus isolates with the isolates falling to one of five clonal complexes. The MRSA clone ST80spa-t044-SCCmec-IVc(2b)-PVL+ which was predominant in this study was also reported as dominant in many parts of the world as a cause of infections in both hospital and community settings [31]. This clone was also previously identified as dominant in different ecological niches in Algeria [32,33].
The second most abundant clone, ST1-t127, was identified in the 1990s as the first CA-MRSA clone [34] and then emerged in diverse healthcare settings [35][36][37], and it was also recovered from companion animals, livestock and livestock products in several countries [38][39][40]. Interestingly, the ST1-t127 clone, which was mainly found as MRSA in many studies [41], was both MRSA and MSSA in this study.
The other second most predominant clone in our study, ST22, was represented by three MSSA and one MRSA (SCCmec type IVa). MRSA belonging to ST22 (EMRSA-15) were reported as the most frequently responsible for nosocomial infections in Europe, are becoming widespread in Asia, Australia, Europe and North America [42][43][44][45], and have also been sporadically isolated in Algeria, Tunisia and South Africa [46]. However, the ST22 MRSA identified in this study does not belong to the EMRSA-15 clone as it contains a type IVa SCCmec element rather than a type IVh and is also ciprofloxacin-sensitive, whereas EMRSA-15 is ciprofloxacin-resistant [42].
Our least frequent clone, ST45, which was commonly reported in North America, Australia and Europe as both MSSA and MRSA in both healthcare and community settings, was also less frequently isolated in South America, Asia and Africa [47].
The SNP-based analysis revealed that the ST1 isolates (10RN and 20RN) were indistinguishable (0 SNP differences). Interestingly, these isolates were recovered from two patients who were admitted 19 days apart to the same hospital and ward (internal medicine), suggesting ward contamination and intra-ward transmission between patients.
Similarly, all the isolates of the ST80 lineage were isolated from the same hospital, two of which (1RN and 16RN) differed by six SNPs only and were from two different patients admitted to the same ward (trauma), but 30 days apart. Isolates 2RN and 15RN, also having a six-SNP difference, were, isolated from patients admitted 45 days apart to separate wards in the same hospital. Isolates 2RN and 4RN, which differed by five SNPs, were isolated from two children admitted 15 days apart to the same ward. The identification of these genomic clusters potentially suggests a persistent contamination of the hospital by the above three clones.
On the other hand, the ST22 isolates (21RN and 25RN) differed by two SNPs; however, they were recovered 51 days apart from two different patients admitted to different hospitals. Considering that (i) the two hospitals are in close proximity (within the same catchment area), (ii) a patient's choice of being re-admitted to a particular hospital is not restricted, (iii) patients' health records are decentralized and (iv) inter-hospital patient transfers (for diagnostic procedure or for extra medical care) are frequent, it is therefore possible that any of the above types of patients' movements and/or referrals between hospitals may have contributed to the dissemination of the S. aureus clones 21RN and 25RN between the two hospitals. Indeed, patient movements between healthcare facilities have been recognized as an important route of pathogen transmission [48].
The three CC15 isolates 13RN, 17RN and 18RN were also closely related, with 7-SNP differences between 13RN and 17RN and 17RN and 18RN and an 11-SNP difference between 13RN and 18RN. These three isolates were recovered from different patients admitted to different wards in the same hospital, suggesting an inter-ward transmission in the same hospital.
Irrespective of the exact mode of intra-or inter-hospital transmissions highlighted above, all the possible scenarios should be considered as breaches in infection control and prevention.
The antibiotic sensitivity testing revealed that all the S. aureus isolates exhibited susceptibility to VAN, QD (a valuable alternative to vancomycin for the treatment of MRSA infections), SXT, LZD, TEC, TGC and NIT. A similar antibiotic susceptibility profile was also reported in a Kenyan study [49].
The prevalence of vancomycin resistance varied globally from 1% in Europe, 3% in South America and 4% in North America to 5% in Asia and 16% in Africa. Similarly, the susceptibility of all our isolates to SXT was interesting, as a high resistance level to this antibiotic was reported among MRSA isolates in Latin America (up to 100%), Taiwan (89%), China (21%) and Africa (from 55% to 72%) [50].
Overall, the rate of MDR in the present study was relatively high (64%), which is consistent with previous reports from Algeria [11,51,52].
The global prevalence of PVL among MRSA strains varied remarkably between geographical regions and populations [53]. The prevalence of PVL among our MRSA isolates (29.4%) falls within the range of other studies [54]. While PVL is generally considered as a potential genetic marker for the differentiation of CA-MRSA and HA-MRSA [55,56], it was, however, equally present in CA-MRSA and HA-MRSA isolates in the present investigation, which was also consistent with a study from Uganda [57].

Conclusions
This study suffers from some limitations, mainly the small sample size; despite this, it provides preliminary insights into the genetic diversity and antibiotic resistance of S. aureus strains circulating in hospital settings in Algeria. To the best of our knowledge, this is the first WGS-based study that included a relatively large collection of clinical S. aureus isolates from Algeria, a country where surveillance of S. aureus has been limited thus far.
Our findings stress the need for effective MRSA control and prevention strategies in the Algerian healthcare system. In addition, this work highlights the importance of WGS as a useful approach in clinical settings, as it provides high-resolution analyses of pathogens, allowing the determination of the relatedness between epidemic strains and the tracing of their transmission events.