Genomics Insight into cfr-Mediated Linezolid-Resistant LA-MRSA in Italian Pig Holdings

The cfr genes encode for a 23S rRNA methyltransferase, conferring a multiresistance phenotype to phenicol, lincosamide, oxazolidinone, pleuromutilin, and streptogramin A antibiotics. These genes have been described in staphylococci, including methicillin-resistant Staphylococcus aureus (MRSA). In this study, we retrospectively performed an in-depth genomic characterisation of three cfr-positive, multidrug-resistant (MDR) livestock-associated (LA) MRSA clonal complexes (CCs) 1 and 398 detected in different Italian pig holdings (2008–2011) during population studies on Italian livestock (2008–2014). We used a combined Illumina and Oxford Nanopore Technologies (ONT) whole genome sequencing (WGS) approach on two isolates (the 2008 CC1 and the 2010 CC398 isolates, but not the 2011 CC1 isolate). Interestingly, the three isolates presented different cfr variants, with only one displaying a linezolid-resistant phenotype. In isolate 2008 CC1, the cfr gene was identified within a Tn558 composite transposon-like structure flanked by IS elements located on a novel 44,826 bp plasmid. This represents the first report of CC1 LA-MRSA harbouring the cfr gene in its functional variant. Differently, cfr was chromosomally located in isolate 2010 CC398. Our findings have significant public health implications, confirm the need for the continuous genomic surveillance of cfr-positive zoonotic LA-MRSA, and backdate cfr presence in LA-MRSA from Italian pigs to at least 2008.


Introduction
Linezolid was the first oxazolidinone antimicrobial available exclusively for human use from April 2000, and it represents one of the few therapeutic options to treat methicillinresistant Staphylococcus aureus (MRSA) skin and soft-tissue infections, osteomyelitis, and pneumonia [1]. Seven mobilisable oxazolidinone resistance genes, including cfr, cfr(B), cfr(C), cfr(D), cfr(E), optrA, and poxtA, have been described so far [2]. In particular, after being identified for the first time about 20 years ago in Staphylococcus sciuri of bovine origin in Germany [3], cfr genes have been found in 19 staphylococcal species from both human and animal sources as well as in other bacterial species [2]. These genes encode for a 23S rRNA methyltransferase, which confers a multiresistance phenotype (phenotype PhLOPS A ), including resistance to phenicol, lincosamide, oxazolidinone, pleuromutilin, and streptogramin A antibiotic classes [4]. The cfr genes have been detected in both methicillin-susceptible S. aureus (MSSA) and MRSA isolates from human clinical samples worldwide [2], including Italy [5]. The cfr(B) gene has recently been described for the first time in MRSA isolates from French clinical human samples [6]. In livestock, cfr was first described in S. aureus of porcine origin in Germany in 2006 [7]. Since then, it has sporadically been detected in pig samples from Belgium [8], Spain [9], Portugal [10], the Netherlands [11], China [12], and South Korea [13], and in bovine samples from Egypt [14]. These genes are most frequently located on mobile genetic elements (MGEs), especially plasmids, which were proven to harbour cfr-carrying transposons and translocatable units (TUs), playing a crucial role in their dissemination [15]. Additionally, TUs and integrative and conjugative elements (ICEs) have also been detected at a chromosomal level [2].
In this study, we retrospectively report the in-depth genomic characterisation of three cfr-positive, multidrug-resistant (MDR) livestock-associated (LA) MRSA isolates belonging to clonal complexes (CCs) 1 (two isolates) and 398 (one isolate), isolated between 2008 and 2011 from Italian pig holdings, combining Illumina short-read and Oxford Nanopore Technologies (ONT) long-read sequencing approaches.

WGS and Bioinformatics Analysis (Illumina Short Reads)
The previous results of multilocus sequence typing (MLST), spa typing, and staphylococcal cassette chromosome mec (SCCmec) typing were confirmed by the results obtained by in silico typing for all the three de novo assembled genomes. Isolates IDs 21554/1 and 55864/23 belonged to spa type t127, sequence type (ST) 1, CC1, and isolate ID 76669/3 belonged to spa type t034, ST398.
Except for linezolid resistance, all MDR phenotypes were consistent with the genetic analysis. All three isolates tested cfr-positive, but only one CC1 isolate (ID 21554/1) was phenotypically linezolid-resistant (Table 1).
The variant calling analysis indicated that both linezolid-susceptible isolates, ID 55864/23 (CC1) and 76669/3 (CC398), presented a cfr gene with a single nucleotide deletion corresponding with position 10,692 bp in the Tn558 transposon of plasmid pSCFS6. S. warneri (AM408573) was used as a reference sequence. This single base-pair deletion caused a reading frameshift and, consequently, a truncated protein. The linezolid-resistant CC1 LA-MRSA isolate (ID 21554/1) harboured a cfr sequence identical to AM408573 from S. warneri ( Figure 1). phenotypically linezolid-resistant ( Table 1).
The variant calling analysis indicated that both linezolid-susceptible isolates, ID 55864/23 (CC1) and 76669/3 (CC398), presented a cfr gene with a single nucleotide deletion corresponding with position 10,692 bp in the Tn558 transposon of plasmid pSCFS6. S. warneri (AM408573) was used as a reference sequence. This single base-pair deletion caused a reading frameshift and, consequently, a truncated protein. The linezolidresistant CC1 LA-MRSA isolate (ID 21554/1) harboured a cfr sequence identical to AM408573 from S. warneri ( Figure 1).

WGS and Bioinformatics Analysis (Hybrid Assembly)
The complete genomes of isolates ID 76669/3 (CC398) and 21554/1 (CC1) subjected to a hybrid (Illumina-ONT) assembly were obtained and analysed. The results of the Bakta annotation and MobileElementFinder analyses are reported in detail in Tables S1 and S2. In particular, the annotation of the assembly of isolate ID 76669/3 (CC398) identified 2649 coding sequences (CDSs), and cfr was located on contig 9 (160,836 bp; 30227..31275 nt) together with the fexA gene. These AMR genes were surrounded by IS elements within a Tn558 composite transposon-like structure, including truncated tnpA (ΔtnpA), tnp-IS21, and istB (IS21-558 element). A BLAST analysis revealed that contig 9 of isolate ID 76669/3 had the BLAST best match (according to Max Score) with the chromosome sequence of LA-MRSA ST398 (strain 17Gst354; CP073065.1) isolated from the nasal swab of a healthy Swiss farmer in 2017, showing a 96.00% coverage and a 99.78% identity. However, this strain did not harbour the cfr gene. The BLAST best match (98.00% coverage and 99.72%

WGS and Bioinformatics Analysis (Hybrid Assembly)
The complete genomes of isolates ID 76669/3 (CC398) and 21554/1 (CC1) subjected to a hybrid (Illumina-ONT) assembly were obtained and analysed. The results of the Bakta annotation and MobileElementFinder analyses are reported in detail in Tables S1 and S2. In particular, the annotation of the assembly of isolate ID 76669/3 (CC398) identified 2649 coding sequences (CDSs), and cfr was located on contig 9 (160,836 bp; 30227..31275 nt) together with the fexA gene. These AMR genes were surrounded by IS elements within a Tn558 composite transposon-like structure, including truncated tnpA (∆tnpA), tnp-IS21, and istB (IS21-558 element). A BLAST analysis revealed that contig 9 of isolate ID 76669/3 had the BLAST best match (according to Max Score) with the chromosome sequence of LA-MRSA ST398 (strain 17Gst354; CP073065.1) isolated from the nasal swab of a healthy Swiss farmer in 2017, showing a 96.00% coverage and a 99.78% identity. However, this strain did not harbour the cfr gene. The BLAST best match (98.00% coverage and 99.72% identity) with cfr-positive MRSA was with the chromosome sequence of the LA-MRSA ST398 strain (RIVM_M047065; CP096539) isolated from a human nasal swab in the Netherlands in 2019 [11].
The results of the average nucleotide identity (ANI) for the alignment coverage and identity of the complete sequence of isolate ID 76669/3 (2,843,032 bp) compared with the complete chromosomes of the LA-MRSA ST398 strain 17Gst354 (2,783,931 bp) and LA-MRSA ST398 strain RIVM_M047065 (2,806,671 bp) showed that they shared an identity of 99.76% and 99.86% with a 93.24% and 96.00% coverage, respectively.
From the NCBI database interrogation, we also observed that the same single basepair deletion (c.381delA) found in the cfr gene of the CC398 LA-MRSA isolate ID 76669/3 was recently described in a cfr gene located in a 18,240 bp chromosomal region (cfr gene cluster; MW298531) of a CC398 LA-MRSA sample isolated in 2019 from a single pig nasal swab in Central Italy [18]. Multiple alignment confirmed this deletion by comparing our sequence with the cfr reference sequences of S. sciuri (NG_047631) and S. warneri (AM408573) (Figure 2). ST398 strain (RIVM_M047065; CP096539) isolated from a human nasal swab in the Netherlands in 2019 [11].
The results of the average nucleotide identity (ANI) for the alignment coverage and identity of the complete sequence of isolate ID 76669/3 (2,843,032 bp) compared with the complete chromosomes of the LA-MRSA ST398 strain 17Gst354 (2,783,931 bp) and LA-MRSA ST398 strain RIVM_M047065 (2,806,671 bp) showed that they shared an identity of 99.76% and 99.86% with a 93.24% and 96.00% coverage, respectively.
From the NCBI database interrogation, we also observed that the same single basepair deletion (c.381delA) found in the cfr gene of the CC398 LA-MRSA isolate ID 76669/3 was recently described in a cfr gene located in a 18,240 bp chromosomal region (cfr gene cluster; MW298531) of a CC398 LA-MRSA sample isolated in 2019 from a single pig nasal swab in Central Italy [18]. Multiple alignment confirmed this deletion by comparing our sequence with the cfr reference sequences of S. sciuri (NG_047631) and S. warneri (AM408573) (Figure 2). An almost (99.99%) identical cfr gene cluster of MRSA G322 (MW298531), described by Fioriti and colleagues in 2021 [18], was identified within the transposon harboured on contig 9 of our isolate ID 76669/3 (21276..39515 nt). Both cfr gene clusters differed by only one nucleotide in the fexA gene encoding a phenicol resistance. In our isolate, this gene was identical to the fexA1 reference sequence in the ResFinder database (AJ549214) whereas the fexA gene within the cfr gene cluster of isolate MRSA G322 presented a singlepoint mutation (G to C) at position 1192 (G1192C). MRSA G322 remained phenotypically resistant to chloramphenicol and florfenicol [18].
Resolved plasmids were obtained only for isolate ID 21554/1 (CC1) (circular contigs 3, 4, 5, and 6; Tables S1 and S2). The cfr gene was located on a novel 44,826 bp plasmid (contig 3; 23694..24743 nt) named pMOL21554 ( Figure 3). Similar to contig 9 of isolate ID 76669/3, cfr was flanked by IS elements in this plasmid within a Tn558 composite transposon-like structure (~13,774 bp). This structure presented an identity of 99.70% with a coverage of 100% with the reference sequence of IS21-558 (AM086211), composed by truncated tnpA (ΔtnpA), tnp-IS21, istB (IS21-558 element), cfr, and fexA. The major difference with the reference sequence was an insertion in our IS21-558-like structure, which included additional mobile elements (25432..30683 nt) represented by two transposases, an ATP-binding protein, and two recombinases ( Figure 3). An almost (99.99%) identical cfr gene cluster of MRSA G322 (MW298531), described by Fioriti and colleagues in 2021 [18], was identified within the transposon harboured on contig 9 of our isolate ID 76669/3 (21276..39515 nt). Both cfr gene clusters differed by only one nucleotide in the fexA gene encoding a phenicol resistance. In our isolate, this gene was identical to the fexA1 reference sequence in the ResFinder database (AJ549214) whereas the fexA gene within the cfr gene cluster of isolate MRSA G322 presented a single-point mutation (G to C) at position 1192 (G1192C). MRSA G322 remained phenotypically resistant to chloramphenicol and florfenicol [18].
As reported for contig 9 of isolate ID 76669/3, this plasmid only harboured fexA and cfr as AMR resistance genes. Isolate ID 21554/1 harboured other AMR genes such as blaZ, bleO, aadD located in a rep20 plasmid and the tet(K) gene in a rep7a plasmid.
The BLAST analysis revealed that pMOL21554 of isolate ID 21554/1 had the BLAST best match with: (i) a 39,287 bp plasmid (88% coverage and 100% identity) named pSA737 (KC206006) from ST239 MSSA (strain 004-737X) isolated in 2007 from a human clinical sample in the United States [19]; and (ii) a 39,212 bp plasmid (88% coverage and 99.99% identity) named pY96A (CP065516.1) from ST1 MSSA (strain GDY8P96A) isolated in 2019 from a pig nasal swab in China. The main differences between the 3 plasmid sequences were related to the presence in our plasmid of a 5249 bp (25430..30679 nt) region harbouring several transposable elements represented by a transposase, a recombinase, and an ATP-binding protein encoding gene. This region was absent in the two previously deposited plasmid sequences (Figure 4). As reported for contig 9 of isolate ID 76669/3, this plasmid only harboured fexA and cfr as AMR resistance genes. Isolate ID 21554/1 harboured other AMR genes such as blaZ bleO, aadD located in a rep20 plasmid and the tet(K) gene in a rep7a plasmid.
The BLAST analysis revealed that pMOL21554 of isolate ID 21554/1 had the BLAST best match with: (i) a 39,287 bp plasmid (88% coverage and 100% identity) named pSA737 (KC206006) from ST239 MSSA (strain 004-737X) isolated in 2007 from a human clinica sample in the United States [19]; and (ii) a 39,212 bp plasmid (88% coverage and 99.99% identity) named pY96A (CP065516.1) from ST1 MSSA (strain GDY8P96A) isolated in 2019 from a pig nasal swab in China. The main differences between the 3 plasmid sequence were related to the presence in our plasmid of a 5249 bp (25430..30679 nt) region harbouring several transposable elements represented by a transposase, a recombinase and an ATP-binding protein encoding gene. This region was absent in the two previously deposited plasmid sequences (Figure 4).  As reported for contig 9 of isolate ID 76669/3, this plasmid only harboured fexA and cfr as AMR resistance genes. Isolate ID 21554/1 harboured other AMR genes such as blaZ, bleO, aadD located in a rep20 plasmid and the tet(K) gene in a rep7a plasmid.
The BLAST analysis revealed that pMOL21554 of isolate ID 21554/1 had the BLAST best match with: (i) a 39,287 bp plasmid (88% coverage and 100% identity) named pSA737 (KC206006) from ST239 MSSA (strain 004-737X) isolated in 2007 from a human clinical sample in the United States [19]; and (ii) a 39,212 bp plasmid (88% coverage and 99.99% identity) named pY96A (CP065516.1) from ST1 MSSA (strain GDY8P96A) isolated in 2019 from a pig nasal swab in China. The main differences between the 3 plasmid sequences were related to the presence in our plasmid of a 5249 bp (25430..30679 nt) region harbouring several transposable elements represented by a transposase, a recombinase, and an ATP-binding protein encoding gene. This region was absent in the two previously deposited plasmid sequences (Figure 4).   The complete sequences of the cfr-positive isolates and the resolved cfr-carrying plasmid were deposited in the European Nucleotide Archive (ENA) at EMBL-EBI under project accession number PRJEB59270.

Discussion
In this study, we report the in-depth molecular characterisation of CC1 (two isolates) and CC398 (one isolate) cfr-positive MDR LA-MRSA isolates of porcine origin detected for the first time in Italian animal production in 2008-2011 [16,17] by using the combined approach of short-and long-read sequencing. Based on previous population studies (2008-2014) conducted on Italian animal production and according to other European-wide studies on cfr-positive MRSA isolated from pig holdings, the occurrence of cfr-positive LA-MSSA and -MRSA in animal production is emerging, but seems to be sporadic [8][9][10][11]. Accordingly, in a recent study (2019-2020) conducted on dairy cattle farms located in Italian regions representing around 80% of Italian milk production, 22 LA-MRSA samples isolated from 316 bulk tank milk samples and belonging to different CCs tested negative for the presence of cfr [20].
The isolates characterised in this study presented different cfr variants/locations and linezolid-resistant phenotypes. The cfr gene was demonstrated to be located on a novel 44,826 bp plasmid in the CC1 isolate (ID 21554/1) and, to the best of our knowledge, this represents the first report of CC1 LA-MRSA harbouring the cfr gene in its functional variant; i.e., displaying both a microbiological and a clinical resistance. The presence of the cfr gene was previously detected only in a 39,212 bp plasmid (CP065516.1) from one CC1 MSSA sample isolated in 2019 from a pig nasal swab in China. Differently, in our CC398 LA-MRSA isolate (ID 76669/3), cfr seemed to be chromosomally located. This result was supported by a comparison with a previously deposited complete chromosome sequence of cfr-positive CC398 MRSA of human origin (strain RIVM_M047065; CP096539), considering the obtained coverage (96.00%) and identity (99.86%). This isolate, together with the CC1 isolate ID 55864/23, presented a frameshift mutation caused by a single base-pair deletion in the cfr gene and, consequently, a truncated protein, corroborating their phenotypic susceptibility to linezolid (MIC value = 1 mg/L). Prior to our findings, linezolid-susceptible MSSA and MRSA associated with cfr variants with a single base-pair deletion or mutation were isolated from two human clinical cases (one in Brazil [21] and one in the Netherlands [11]), and also from two nasal swabs of porcine origin (one in Italy [18] and the other in Australia (CP029172)). All of them belonged to CC398.
As for the genetic environment of cfr insertion sequences and mobile elements found close to cfr, those of the IS21 family (e.g., IS21-558) in particular have been reported to have an important role in its mobilisation [15]. The importance of the IS21 family appears to be consistent, irrespective of whether the cfr gene is found in plasmids or in chromosomal structures. IS21-558, also called ISSau9, was first identified on plasmid pSCFS3 from S. aureus of porcine origin in Germany [7], and a cfr-harbouring segment bound by two directly oriented copies of IS21-558 was first identified in plasmid pSCFS6 from both S. warneri and S. simulans of porcine origin isolated in Denmark [22], suggesting that this structure could be transferred and spread among different staphylococcal species. In this regard, we identified in both isolates subjected to the hybrid assembly approach (IDs 76669/3-CC398 and 21554/1-CC1) a IS21-558-like structure near to the fexA and cfr genes. In plasmid pMOL21554 from isolate ID 21554/1, this structure was very similar (99.7% identity with 100% coverage) to the reference sequence of IS21-558 (AM086211), suggesting that the cfr gene may have been transmitted through the Tn558 transposition. Considering the potential transmissibility of cfr and its emergence in animal production in Europe, these results have significant public health implications and backdate the cfr presence in LA-MRSA from Italian pigs at least to 2008. However, we could speculate that LA-MRSA of swine origin has had, so far, a mild attitude to acquire and maintain such resistance genes along the production cycles. This may reflect the low frequency of detection of cfr-positive MRSA observed so far in humans [11], even in case they may represent an important reservoir of exposure for animal production.
Moreover, it has to be considered that even if oxazolidinones and streptogramin A antibiotic classes are not used in food-producing animals, the selective pressure exerted by the non-prudent use of PhLOPSA antibiotic classes in both human and veterinary sectors (e.g., in humans, lincosamides and oxazolidinones; in veterinary medicine, phenicols, lincosamides, and pleuromutilins) may lead to the further dissemination of the cfr gene.
Although our identified plasmid harboured only fexA and cfr, the potential coexistence of cfr and other acquired AMR genes in the same plasmids is worrying as co-selection mechanisms and the persistence of the cfr gene may occur under selective pressure induced by the use of non-PhLOPSA antibiotic classes [15].
Notably, after the first European survey conducted in 2008 [23], a new European Union-wide baseline survey is to be carried out in the next years in order to provide updated and representative information on the hazard of LA-MRSA from pigs [24]. This type of harmonised study will help provide data at both European and national levels on linezolid-resistant LA-MRSA.

Conclusions
In conclusion, our findings confirm the need for the continuous genomic surveillance of cfr-positive zoonotic LA-MRSA and other oxazolidinone-resistant bacteria in livestock. This approach will help the understanding of the genetic environment and the transmission patterns of MGEs involved in the dissemination of powerful genetic determinants, mediating multiresistance to last-resort drugs for human medicine.

Library Preparation and Whole Genome Sequencing (WGS)
DNA extraction and library preparation were performed as previously reported [30]. Briefly, genomic DNA was extracted using a QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's protocol. The libraries for short-read pair-end sequencing were prepared for the three isolates using a NexteraXT DNA library preparation kit (Illumina, Inc., San Diego, CA, USA) and sequenced on an Illumina platform (MiSeq sequencer, Illumina, Inc., San Diego, CA, USA)). In parallel, the libraries of two isolates (IDs 21554/1 and 76669/3) were prepared with a ligation sequencing kit (SQK-LSK109) and sequenced using a nanopore-based MinION device (ONT) [31].

Bioinformatics Analysis (Hybrid Assembly)
For the two isolates subjected to both short-and long-read sequencing, high accuracy base-calling was performed on the long reads obtained from the MinION device according to the ONT workflow. A hybrid (Illumina-ONT) assembly was implemented using the Unicycler pipeline [46] with default parameters [47]. The obtained assemblies were also annotated using the online version of the Bakta tool (https://bakta.computational.bio/ (accessed on 19 July 2022) [48]. Additionally, a manual curation for the obtained annotation was performed. The identification of the mobile genetic elements and their relation to the AMR genes and virulence factors was performed with the online version of the CGE tool, MobileElementFinder v1.0.3 (last accessed 19 July 2022) [49] using default parameters. From the hybrid assemblies, the obtained contigs containing cfr were also compared with the BLAST algorithm using the online tool against the nr/nt database with default parameters. The complete sequence of isolate ID 76669/3 was also compared with selected complete chromosome sequences of cfr-positive S. aureus obtained from publicly available databases showing the BLAST best match and calculating the ANI for the alignment coverage and identity using the BLAST algorithm [50].
The presence of mutations of the cfr genes was confirmed, and their genetic environment was investigated by comparing the multiple alignment of the contig-harbouring cfr of isolate ID 76669/3 with publicly available selected sequences harbouring cfr from: (i) S. warneri (AM408573) [22]; (ii) S. sciuri (NG_047631) [51]; and (iii) S. aureus G322 (MW29853) [18]. This was achieved using Geneious Prime 2023.0.1 software.
The graphical representation of the general structures and of the genetic regions of the cfr-carrying plasmid of isolate ID 21554/1 as well as the comparison with selected already-deposited plasmid sequences were performed using Geneious Prime 2023.0.1 and the Mauve algorithm for the alignment [52].