Plasmid-Coded Linezolid Resistance in Methicillin-Resistant Staphylococcus aureus from Food and Livestock in Germany

Resistance of methicillin-resistant Staphylococcus aureus (MRSA) from food and livestock to last resort antibiotics such as linezolid is highly concerning, since treatment options for infections in humans might be diminished. Known mechanisms of linezolid resistance include point mutations in the 23S rRNA gene and in the ribosomal proteins L3, L4 and L22 as well as an acquisition of the cfr, optrA or poxtA gene. The objective of our study was to characterize antimicrobial resistance (AMR) determinants and phylogenetic relationships among linezolid-resistant (LR-) MRSA from food and livestock. In total, from more than 4000 incoming isolates in the years 2012 to 2021, only two strains from 2015 originating from pig samples exhibited linezolid resistance in the antimicrobial susceptibility testing with MICs of ≥8 mg/L. These LR-MRSA were characterized in detail by whole-genome sequencing and phylogenetic analyses using cgMLST. The LR-MRSA strains showed resistances to ten and eight different antibiotics, respectively. Both strains harbored plasmid-coded cfr genes mediating the linezolid resistance. The cfr genes showed identical sequences in both strains. In addition to the cfr gene, genes for phenicol and clindamycin resistance were detected on the respective plasmids, opening the possibility for a co-selection. The LR-MRSA differed distantly in the phylogenetic analyses and also to other MRSA from pig samples in the year 2015. In conclusion, the occurrence of LR-MRSA in food and livestock seems to be very rare in Germany. However, carriage of plasmids with linezolid resistance determinants could lead to further linezolid-resistant strains by horizontal gene transfer.


Introduction
Methicillin-resistant Staphylococcus aureus (MRSA) are frequently detected in livestock and food [1,2]. In addition to their resistance to virtually all beta-lactam antibiotics, resistance to critically important or last resort antibiotics such as linezolid is highly concerning for animal and human health due to the further reduction of treatment options for severe MRSA infections. Linezolid belongs to the oxazolidinone antibiotic group. Its mode of action is the inhibition of protein synthesis by binding to the ribosomal RNA [3]. Linezolid resistance may be mediated by several genes (cfr, optrA and poxtA) as well as by point mutations in the genes for the 23S rRNA and ribosomal proteins L3, L4 and L22 [4][5][6][7][8]. The cfr, optrA and poxtA genes are often located on plasmids and may be transmitted between staphylococcal species or even between different genera [9]. As well as mediating linezolid resistance, the cfr gene, for which the variants cfr, cfr(B), cfr(C), cfr(D) and cfr(E) were detected [9], may also cause resistance to phenicols, lincosamides, pleuromutilins, and streptogramin A [5]. poxtA is a phenicol-oxazolidinone-tetracycline resistance gene [7], whereas the optrA gene confers resistance to oxazolidinone and tetracycline antibiotics [6]. In addition, for the optrA gene, different variants have been discovered in enterococci [10]. Several studies have reported linezolid-resistant (LR-) MRSA strains in humans, livestock and food; however, the prevalence was rather low compared to resistance to other antibiotics [11][12][13][14]. In the One Health context, handling livestock or food has been shown to be a risk factor for the acquisition of livestock-associated MRSA from animals [15][16][17][18]. Therefore, it has to be considered that acquisition of LR-MRSA from animals might lead to difficult-to-treat infections in people working in the livestock or food sector. As a consequence, monitoring of antimicrobial resistance in MRSA from food and livestock is of high importance.
Knowledge about AMR genes and their detection and characterization by wholegenome sequencing (WGS) is of high value for medical treatment of MRSA infections. The prediction of AMR by sequencing may support faster treatment decisions for MRSA infections in comparison to cultural approaches. Moreover, phylogenetic relationships and transmission of resistance genes may be unraveled by analyzing and comparing whole genome sequences of MRSA. For instance, MRSA spread in livestock farms may be tracked genomically by the application of WGS and subsequent sequence analyses [19][20][21].
The aim of our study was to characterize LR-MRSA from food and livestock in Germany by WGS and phenotypic antimicrobial susceptibility testing. The genomic causes for linezolid resistance were discovered. Moreover, phylogenetic relationships among the strains were analyzed to evaluate the relationships of LR-MRSA from livestock and food.

Genotypic Analyses and AMR
The strain collection of the National Reference Laboratory for coagulase-positive Staphylococci, including Staphylococcus aureus, at the German Federal Institute for Risk Assessment was screened for MRSA strains that were phenotypically linezolid-resistant. From more than 4000 incoming MRSA strains in the years 2012 to 2021, only two MRSA strains showed phenotypic resistance to linezolid, exhibiting a minimum inhibitory concentration of ≥8 mg/L. Both LR-MRSA strains harbored a SCCmec type V element, but carried different spa types t011 and t034. Accordingly, the strains were associated to the sequence type (ST) 398 that is highly prevalent in livestock in Germany. The LR-MRSA strains showed a multidrug-resistant phenotype with resistance to ten or eight of the included antimicrobials, respectively. Both LR-MRSA strains exhibited resistance to chloramphenicol, clindamycin, cefoxitin, linezolid, penicillin, quinupristin-dalfopristin, tetracycline and tiamulin. One of the two strains (LR-MRSA-1) was additionally resistant to streptomycin and trimethoprim.
Point mutations in the 23S rRNA gene as well as in the ribosomal proteins L3, L4 and/or L22 genes were not discovered in the LR-MRSA strains. In both strains, a cfr gene was detected, whereas optrA or poxtA genes were not found. Gene sequences of the cfr genes were identical and clustered closely together with cfr gene sequences from other staphylococcal species in the NCBI database ( Figure 1). In addition to the cfr gene, the LR-MRSA-1 strain harbored the AMR genes ant(9)-Ia, dfrG, fexA, lnu(B), lsa(B), lsa(E), mecA, str, tet (38), tet(K) and tet(M) whereas the LR-MRSA-2 strain additionally harbored the AMR gene blaZ but lacked lsa(B) and str.
Analyses of the flanking regions of the cfr genes revealed that the cfr genes were located on plasmids ( Figure 2). The LR-MRSA-1 strain harbored sequences that were highly similar to the plasmid CP028164.1 (sequence identity 99.99%, query coverage 100%), whereas the LR-MRSA-2 strain showed sequences with a high identity to the plasmid CP065195.1 (sequence identity 98.73%, query coverage 95%). Both plasmids additionally harbored a fexA gene. Moreover, plasmid CP028164.1 carried an lsa gene.

Phylogenetic Analysis
Core genome multilocus sequence typing (cgMLST) with concurrent minimum spanning tree (MST) analysis was carried out with the LR-MRSA strains in comparison to 13 other ST9-and ST398-MRSA strains with a pig origin and from the year 2015 to evaluate possible phylogenetic relationships and transmission events between strains from the same year and origin. The MST illustrates the high diversity of the MRSA strains in the year 2015 with distantly related strains ( Figure 3). The two detected LR-MRSA strains were also distantly related.

Phylogenetic Analysis
Core genome multilocus sequence typing (cgMLST) with concurrent minimum spanning tree (MST) analysis was carried out with the LR-MRSA strains in comparison to 13 other ST9-and ST398-MRSA strains with a pig origin and from the year 2015 to evaluate possible phylogenetic relationships and transmission events between strains from the same year and origin. The MST illustrates the high diversity of the MRSA strains in the year 2015 with distantly related strains ( Figure 3). The two detected LR-MRSA strains were also distantly related.

Phylogenetic Analysis
Core genome multilocus sequence typing (cgMLST) with concurrent minimum spanning tree (MST) analysis was carried out with the LR-MRSA strains in comparison to 13 other ST9-and ST398-MRSA strains with a pig origin and from the year 2015 to evaluate possible phylogenetic relationships and transmission events between strains from the same year and origin. The MST illustrates the high diversity of the MRSA strains in the year 2015 with distantly related strains ( Figure 3). The two detected LR-MRSA strains were also distantly related.

Discussion
The risk of a lack in treatment options for infections caused by MRSA is ever-present, since resistance to a high number of antimicrobials, including critically important antibiotics, is occasionally detected in MRSA strains. Linezolid is one of the last resort antibiotics. Although it is not licensed for use in veterinary medicine, LR-MRSA have been found in livestock and food samples. As a consequence, transmission of these strains to humans via animal contact or food may occur and subsequent infections caused by these strains might be difficult-to-treat.
In our study, LR-MRSA obtained from food and livestock samples in Germany collected between 2012 and 2021, were analyzed in detail by WGS and phenotypic antimicrobial susceptibility testing. Only two LR-MRSA isolates were recovered from the strain collection. Both strains originated from pigs sampled in 2015. The two strains belonged to the livestock-associated ST398. In relation to the total number of MRSA, which were analyzed by our laboratory in this time period (>4000), the LR-MRSA only comprise a very minor fraction of MRSA. This is in line with recent data about a low proportion of LR-MRSA-caused human bloodstream infections in Europe [22] and the low global prevalence of LR-MRSA in clinical samples [11].
Linezolid resistance may be associated with point mutations in the 23S rRNA gene, mutations in the genes of the ribosomal proteins L3, L4 and L22 and/or with an acquisition of the cfr, optrA or poxtA gene, which are often harbored on mobile genetic elements [8]. In our study, point mutations in the 23S rRNA gene or in the proteins L3, L4 or L22 genes were not found in the detected LR-MRSA strains. Accordingly, a recent study from Belgium reported, in particular, the presence of cfr, optrA and poxtA genes in linezolid-re-

Discussion
The risk of a lack in treatment options for infections caused by MRSA is ever-present, since resistance to a high number of antimicrobials, including critically important antibiotics, is occasionally detected in MRSA strains. Linezolid is one of the last resort antibiotics. Although it is not licensed for use in veterinary medicine, LR-MRSA have been found in livestock and food samples. As a consequence, transmission of these strains to humans via animal contact or food may occur and subsequent infections caused by these strains might be difficult-to-treat.
In our study, LR-MRSA obtained from food and livestock samples in Germany collected between 2012 and 2021, were analyzed in detail by WGS and phenotypic antimicrobial susceptibility testing. Only two LR-MRSA isolates were recovered from the strain collection. Both strains originated from pigs sampled in 2015. The two strains belonged to the livestock-associated ST398. In relation to the total number of MRSA, which were analyzed by our laboratory in this time period (>4000), the LR-MRSA only comprise a very minor fraction of MRSA. This is in line with recent data about a low proportion of LR-MRSA-caused human bloodstream infections in Europe [22] and the low global prevalence of LR-MRSA in clinical samples [11].
Linezolid resistance may be associated with point mutations in the 23S rRNA gene, mutations in the genes of the ribosomal proteins L3, L4 and L22 and/or with an acquisition of the cfr, optrA or poxtA gene, which are often harbored on mobile genetic elements [8]. In our study, point mutations in the 23S rRNA gene or in the proteins L3, L4 or L22 genes were not found in the detected LR-MRSA strains. Accordingly, a recent study from Belgium reported, in particular, the presence of cfr, optrA and poxtA genes in linezolid-resistant staphylococcal and enterococcal strains from food-producing animals [14]. Moreover, LR-MRSA harboring a cfr gene were recently detected in pigs in Spain [13] and in retail food in China [23]. In addition, a cfr gene was found in staphylococci from veal calves and pigs from farms in Germany and Luxembourg [24]. In our study, both LR-MRSA strains harbored a cfr gene. The cfr gene may be located on plasmids and thus be transmittable between staphylococci [9]. Plasmid-coded cfr genes were also found in the two strains of our study. In contrast, the genes optrA and poxtA, which may also mediate linezolid resistance and are often found on mobile genetic elements, were not detected in the strains of our study. Accordingly, in other studies, linezolid resistance in staphylococci from several food-producing animals or pigs was indeed mediated by a cfr gene, whereas optrA and poxtA genes were not detected [14,25]. Both cfr genes carrying LR-MRSA strains in our study exhibited a rather high MIC for chloramphenicol (>64 mg/L). This was most probably caused by two different resistance mechanisms that are transmitted by the cfr and fexA genes present in both isolates. As well as linezolid resistance, the cfr gene was described to confer resistance to phenicols, lincosamides, pleuromutilins, and streptogramin A antibiotics [5]. Moreover, the fexA gene mediates phenicol resistance as well. Both genes, cfr and fexA, were co-located in close proximity on a plasmid. In addition, the strains exhibited resistance to clindamycin, which belongs to the lincosamides, and tiamulin, a pleuromutilin. Both resistances may also be conferred by the cfr gene.
Linezolid is not licensed for use in livestock. Therefore, the question about the selective pressure for the acquisition and persistence of linezolid resistance determinants in livestock arises. Likely, LR-MRSA harboring resistance to other antibiotics that are used in livestock farming are selected by antibiotic use on livestock farms. Florfenicol may be used in pig farming for acute respiratory illnesses and this antibiotic was also used on German pig farms in 2015 [26]. Since the cfr and fexA gene confer a rather high MIC for florfenicol, use of florfenicol might have co-selected for linezolid resistance. Moreover, tiamulin may be used in pig farming in the treatment of dysentery. Tiamulin resistance is also transmitted by the cfr gene, so that a possible co-selection also might have appeared due to the use of tiamulin. In addition, the lincosamide lincomycin was used in pig farming in Germany in 2015 [26]. Both LR-MRSA strains harbored lsa and lnu genes, which may confer resistance to lincosamides. At least in the case of one LR-MRSA strain, the lsa gene was located on the same plasmid as the cfr gene. Therefore, use of lincomycin might have been another possibility for co-selection of the respective LR-MRSA strain. However, since resistance to florfenicol is coded on the same plasmid as resistance to linezolid, it is most reasonable to assume that florfenicol use was the main driver of co-selection after acquisition of the plasmids by the two LR-MRSA strains in our study. Moreover, it has to be considered that the detected LR-MRSA strains originated from retail and slaughterhouses. Therefore, a transmission of the strains by humans during animal and meat processing cannot be completely excluded. Hypothetically, workers carrying a linezolid-resistant strain in their noses or on their skin might have passed the strain to the meat.
In particular, the multidrug resistance phenotype of the two LR-MRSA strains, exhibiting resistance to ≥8 antimicrobial substances, is concerning with regard to treatment options for severe infections in animals and humans. This underlines the need to monitor AMR in MRSA from livestock and food. The introduction of such strains into hospitals would further reduce treatment options for severe infections in humans. Moreover, MRSA infections in animals, such as mastitis in cattle, might become difficult to treat, causing high costs for farm operators.
The application of WGS also allows for analyses of phylogenetic relationships among bacteria and typing methods based on WGS were shown to be suitable for tracking foodborne outbreaks [27]. The diversity of MRSA with a pig origin in our study was high and both LR-MRSA strains were only distantly related. MRSA belonging to ST398 or ST9 were clearly separated by cgMLST analyses. Interestingly, large allelic differences between the analyzed MRSA ST398 strains were observed in the MST although some strains harbored the same spa type, such as t011 or t034. Therefore, when investigating transmission of multidrug-resistant MRSA, spa typing alone might be misleading due to an insufficient discriminatory power. The improved discriminatory power of cgMLST analyses improves the traceability. The suitability of core genome analyses was also shown in other studies such as those on transmission of MRSA between pigs and humans in Australia [19], phylogenetic tracking of MRSA at German dairy farms [21] and the prevalence and diversity of S. aureus in the dairy value chain in Zambia [28].
To date, monitoring of MRSA in food and livestock is voluntary in Europe. Therefore, only a few European countries report on the prevalence of MRSA and their resistance phenotypes in livestock and food [29]. Our study illustrates that surveillance of AMR in MRSA from food and livestock in particular with regard to antimicrobials of great importance to the medical sector is an important contribution to the one health assessment of the resistance situation in S. aureus and for an early detection of MRSA with a high public health importance. Molecular surveillance based on sequencing may identify possible transmission events and the distribution of specific clones across matrices including animals, food and humans, such as farm staff.

Sampling and MRSA Strain Selection
From 2012 to 2021, the National Reference Laboratory for coagulase-positive Staphylococci, including Staphylococcus aureus, at the German Federal Institute for Risk Assessment obtained more than 4000 MRSA isolates. Two LR-MRSA strains (MIC ≥ 8 mg/L) were detected in the strain collection. Both LR-MRSA strains were collected in the year 2015 and originated from pigs (pig meat at retail or carcass from a slaughterhouse).

WGS and Bioinformatics Analyses
MRSA strains were inoculated in 5 mL brain-heart-infusion broth and incubated at 37 • C for 24 h. DNA of 1 mL culture was extracted using the Qiagen DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol modified by adding 10 µL lysostaphin (Sigma Aldrich, St. Louis, MO, USA) to the lysis buffer. The DNA library was prepared using an Illumina Nextera DNA Prep kit (Illumina Inc., San Diego, CA, USA) and the 150 bp paired-end sequencing run was performed on an Illumina NextSeq 500 instrument. Raw Illumina reads were trimmed and de novo assembled with the in-house developed AQUAMIS pipeline [30]. Bacterial characterization was carried out with the in-house developed Bakcharak pipeline (https://gitlab.com/bfr_bioinformatics/bakcharak, accessed on 1 October 2022) using the NCBI AMRfinder database [31] for screening of AMR genes. Moreover, software tools from the Center for Genomic Epidemiology (https://cge.food.dtu.dk/services/, accessed on 1 October 2022) were applied. ResFinder 4.1 [32][33][34] was used for point mutation investigation, SCCmecFinder 1.2 [34][35][36] was used for analyses of SCCmec types and spaTyper 1.0 [37] was used for determination of spa types. Analyses by cgMLST as well as MLST determination were conducted using Ridom SeqSphere+ version 7.0.4 according to S. aureus cgMLST scheme comparing 1861 alleles. Determination of point mutations in the ribosomal protein L3, L4 and L22 genes was conducted by alignment to reference sequences from the strains JP080 (AP017922.1) and M12/0145 (KU510541.1) using blastn suite (NCBI BLAST). Comparison and alignment of cfr gene sequences was conducted in MEGA X version 10.1.7. The cfr gene sequences of different bacterial species were retrieved in NCBI Genbank. Identification of plasmids was carried out in NCBI BLAST by using contigs with cfr gene carriage and sequence analyses of flanking regions.

Conclusions
The study illustrates that linezolid-resistant MRSA are rarely found in livestock and food samples from Germany. Due to the multiple resistance to antimicrobials these strains might cause difficult-to-treat infections in animals or humans. Plasmid-coded linezolid resistance might be transmitted to other staphylococcal species or even to other genera such as Enterococcus. Since linezolid is not licensed for use in livestock farming, co-selection by other antibiotics such as florfenicol might take place. LR-MRSA from livestock might be transmitted to humans by direct or indirect contact. Therefore, even if rarely found at the moment, it is important for human health to monitor the possible reservoirs for bacteria with resistance to critically important antimicrobials.