The Resistome and Mobilome of Multidrug-Resistant Staphylococcus sciuri C2865 Unveil a Transferable Trimethoprim Resistance Gene, Designated dfrE, Spread Unnoticed

ABSTRACT Methicillin-resistant Staphylococcus sciuri (MRSS) strain C2865 from a stranded dog in Nigeria was trimethoprim (TMP) resistant but lacked formerly described staphylococcal TMP-resistant dihydrofolate reductase genes (dfr). Whole-genome sequencing, comparative genomics, and pan-genome analyses were pursued to unveil the molecular bases for TMP resistance via resistome and mobilome profiling. MRSS C2865 comprised a species subcluster and positioned just above the intraspecies boundary. Lack of species host tropism was observed. S. sciuri exhibited an open pan-genome, while MRSS C2865 harbored the highest number of unique genes (75% associated with mobilome). Within this fraction, we discovered a transferable TMP resistance gene, named dfrE, which confers high-level TMP resistance in Staphylococcus aureus and Escherichia coli. dfrE was located in a novel multidrug resistance mosaic plasmid (pUR2865-34) encompassing adaptive, mobilization, and segregational stability traits. dfrE was formerly denoted as dfr_like in Exiguobacterium spp. from fish farm sediment in China but escaped identification in one macrococcal and diverse staphylococcal genomes in different Asian countries. dfrE shares the highest identity with dfr of soil-related Paenibacillus anaericanus (68%). Data analysis discloses that dfrE has emerged from a single ancestor and places S. sciuri as a plausible donor. C2865 unique fraction additionally enclosed novel chromosomal mobile islands, including a multidrug-resistant pseudo-SCCmec cassette, three apparently functional prophages (Siphoviridae), and an SaPI4-related staphylococcal pathogenicity island. Since dfrE seems not yet common in staphylococcal clinical specimens, our data promote early surveillance and enable molecular diagnosis. We evidence the genome plasticity of S. sciuri and highlight its role as a resourceful reservoir for adaptive traits. IMPORTANCE The discovery and surveillance of antimicrobial resistance genes (AMRG) and their mobilization platforms are critical to understand the evolution of bacterial resistance and to restrain further expansion. Limited genomic data are available on Staphylococcus sciuri; regardless, it is considered a reservoir for critical AMRG and mobile elements. We uncover a transferable staphylococcal TMP resistance gene, named dfrE, in a novel mosaic plasmid harboring additional resistance, adaptive, and self-stabilization features. dfrE is present but evaded detection in diverse species from varied sources geographically distant. Our analyses evidence that the dfrE-carrying element has emerged from a single ancestor and position S. sciuri as the donor species for dfrE spread. We also identify novel mobilizable chromosomal islands encompassing AMRG and three unrelated prophages. We prove high intraspecies heterogenicity and genome plasticity for S. sciuri. This work highlights the importance of genome-wide ecological studies to facilitate identification, characterization, and evolution routes of bacteria adaptive features.

ABSTRACT Methicillin-resistant Staphylococcus sciuri (MRSS) strain C2865 from a stranded dog in Nigeria was trimethoprim (TMP) resistant but lacked formerly described staphylococcal TMP-resistant dihydrofolate reductase genes (dfr). Wholegenome sequencing, comparative genomics, and pan-genome analyses were pursued to unveil the molecular bases for TMP resistance via resistome and mobilome profiling. MRSS C2865 comprised a species subcluster and positioned just above the intraspecies boundary. Lack of species host tropism was observed. S. sciuri exhibited an open pan-genome, while MRSS C2865 harbored the highest number of unique genes (75% associated with mobilome). Within this fraction, we discovered a transferable TMP resistance gene, named dfrE, which confers high-level TMP resistance in Staphylococcus aureus and Escherichia coli. dfrE was located in a novel multidrug resistance mosaic plasmid (pUR2865-34) encompassing adaptive, mobilization, and segregational stability traits. dfrE was formerly denoted as dfr_like in Exiguobacterium spp. from fish farm sediment in China but escaped identification in one macrococcal and diverse staphylococcal genomes in different Asian countries. dfrE shares the highest identity with dfr of soil-related Paenibacillus anaericanus (68%). Data analysis discloses that dfrE has emerged from a single ancestor and places S. sciuri as a plausible donor. C2865 unique fraction additionally enclosed novel chromosomal mobile islands, including a multidrug-resistant pseudo-SCCmec cassette, three apparently functional prophages (Siphoviridae), and an SaPI4-related staphylococcal pathogenicity island. Since dfrE seems not yet common in staphylococcal clinical specimens, our data promote early surveillance and enable molecular diagnosis. We evidence the genome plasticity of S. sciuri and highlight its role as a resourceful reservoir for adaptive traits. IMPORTANCE The discovery and surveillance of antimicrobial resistance genes (AMRG) and their mobilization platforms are critical to understand the evolution of bacterial resistance and to restrain further expansion. Limited genomic data are available on Staphylococcus sciuri; regardless, it is considered a reservoir for critical AMRG and mobile elements. We uncover a transferable staphylococcal TMP resistance gene, named dfrE, in a novel mosaic plasmid harboring additional resistance, adaptive, and self-stabilization features. dfrE is present but evaded detection in diverse species from varied sources geographically distant. Our analyses evidence that the dfrE-carrying element has emerged from a single ancestor and position S. sciuri as the donor species for dfrE spread. We also identify novel mobilizable chromosomal islands encompassing AMRG and three unrelated prophages. We prove high intraspecies heterogenicity and genome plasticity for S. sciuri. This work highlights the importance of genome-wide mobile elements were identified; however, multiple attempts to determine their entire structure failed. PacBio sequencing retrieved one single circular chromosomal contig of 2,913,767 bp plus one circularized 40,108-bp plasmid (pUR2865-34), making a sum of 2,953,875 bp. Dot plot analysis of Illumina versus PacBio assemblies is included in the supplemental material and Fig. S1. PacBio data were used for all annotations and downstream analyses, except for both small Illumina-sequenced plasmids.
S. sciuri exhibits an open pan-genome while MRSS C2865 shows the highest unique genome content and constitutes a subgroup within the species. Pan-genome analysis of 21 S. sciuri species genomes, which originated from human, animal, food, or environmental samples, including clinical isolates, revealed a pan-genome of 5,721 genes, comprising a core genome of 1.3 Mb with 1,547 shared proteins at 95% of identity (27% of genes) and a flexible genome of 4,174 (73%). The pan-genome curve did not level off, as the addition of each new genome increased the total gene pool. Instead, the core genome appeared to have reached a plateau (,1,600 genes) (Fig. 1A). This ability to acquire exogenous DNA indicates an open pan-genome for S. sciuri (Fig. 1B). This is further supported by a prevalence of "cloud genes" (genes found in up to 15% of the strains), which corresponds with approximately 43.2% of the pangenome (Fig. 1A). S. sciuri C2865 harbored the highest number of genes (3,063) and the highest number of unique genes (521), whereas the average gene numbers were 2,707 and 95, respectively, considering the 21 strains (Fig. 1C). This was not correlated with the genome size or overall genome coding density (89%, median 89); however, MRSS C2865 encompassed the shortest median intergenic distance (Fig. 1C), mainly included within the flexible unique gene set ( Table 2). Several novel chromosomal and extrachromosomal elements were detected (see below), which represented 75% of C2865 unique genes (390/521). Three plasmids were identified within the unique fraction: two small single resistance plasmids (pUR2865-1 and pUR2865-2) (Fig. S4, plus Text S1 for details) and a novel MDR plasmid, pUR2865-34 (see section below). Table 2 summarizes the most relevant characteristics of the MGEs identified.
A refined S. sciuri phylogenomic tree revealed that MRSS C2865 formed a separate subgroup, together with MDR S. sciuri Z8, isolated from a human skin wound infection in China (38), MDR S. sciuri SNUDS-18, from a duckling with tremor in South Korea (39), and S. sciuri CCUG39509, from a sliced veal leg in Sweden ( Fig. 2A). To obtain the most accurate phylogeny possible, we removed from the alignment all regions affected by recombination (see Materials and Methods). Average nucleotide identity (ANI) analysis evidenced that these four genomes shared $98% identity, while they displayed an ANI of 96% with respect to the major species cluster, positioning just above the species demarcation threshold (.95%) (Fig. S2). Genomic comparison of MRSS C2865 chromosome against its three closest relatives (Fig. 2C) evidenced the MRSS C2865 unique chromosomal islands. They corresponded to novel site-specific recombinase carrying MGEs (SCCmec, three prophages, staphylococcal pathogenicity island, transposon c Tn554) described in the following sections or in the supplemental material (for c Tn554, see also Fig. S3).
FIG 1 Pan-genome analyses for S. sciuri genomes at the species level (n = 21, 95% identity). (A) Pie chart showing the proportions of coding DNA sequence (CDS) in the core, soft core, shell, and cloud genomes. The parameters were defined as follows. Core genes: $99% of analyzed genomes, accessory genes: 1 to 99% (soft core 95 to 99%; shell 15 to 95%; cloud #15%). (B) Number of core genes (green) and total number of genes (pan genome) (red) curve for 21 S. sciuri strains. The upper and lower edges of the boxes indicate the first quartile (25th percentile of the data) and third quartile (75th percentile), respectively, of 1,000 random different input orders of the genomes. The central horizontal line indicates the sample median (50th percentile). (C) Bar chart of the total number of genes per genome indicating the number of non-unique and unique genes per genome (gray and dark blue, respectively) ordered by the genome size. Secondary axis displays the median intergenic distance per genome in base pairs. Mosaic mobile adaptive elements within the novel plasmid pUR2865-34. Plasmid pUR2865-34 was 40,108 bp in size and harbored 41 coding sequences (CDSs) (Fig. 3A). It showed a mosaic IS6-like-delimited modular organization (Fig. 3A), encompassing novel adaptive and backbone modules. pUR2865-34 cargo region enclosed a copy of the tetracycline efflux major facilitator superfamily (MFS) transporter gene tet(K), located within the small rolling-circle replication (RCR)-integrated plasmid, designated pUR2865-int. Sequence identity of RepC (83.1%) and composition of the double-(nick site: 59-AAAACCGGaTACTCT/AATAGCCGGTT-39, where capital letters denote conserved bases with respect to pT181, and slash the site of cleavage) and single-strand origin of replication (dso and sso, respectively), typical of RCR plasmids, classified this plasmid as a member of the pT181 family ( Fig. 3A; Fig. S4). In addition, pUR2865-34 harbored a macrolide/lincosamide/streptogramin B resistance gene erm(B), coding for a 23S rRNA adenine N-6-methyltransferase and an aminoglycoside modifying aacA-aphD gene, encoding the bifunctional enzyme 69-aminoglycoside N-acetyltransferase AAC(69)-Ie aminoglycoside O-phosphotransferase APH(20)-Ia, located immediately downstream of an IS257/IS1216E copy in the same orientation.
Three serine recombinase (S-rec) genes of the resolvase/invertase subfamily (Text S1), named res/rec1, res/rec2, and res/rec3, were identified. Res/Rec2 and Res/Rec3 shared the highest amino acid identity to few staphylococcal isolates (four S. aureus, two S. sciuri, and one Staphylococcus arlettae strains) (.93% identity), while most entries corresponded to resolvases found in other Gram-positive bacteria (Macrococcus caseolyticus, Exiguobacterium sp., Bacillus sp., Enterococcus sp., Lactococcus sp., Streptococcus sp., Lysinibacillus sp., Paenibacillus sp., Clostridium sp.). A dfr gene, designated dfrE, and a thymidylate synthase gene (thy) were detected immediately downstream of res/rec3. DfrE shared 31.5 and 15.5% identity to two additional intrinsic Dhfr detected in S. sciuri C2865 genome, both located in conserved chromosomal regions lacking insertion sequences (ISs) or other mobile elements. DfrE was present in a limited number of strains at 100% . Phylogenomic view of S. sciuri group species other than S. sciuri is collapsed as an outgroup to facilitate relatedness analysis of S. sciuri strains. Sample source is indicated as human, animal, food, and environment as well as whether the sample was isolated from a clinical infection. Note: strains S. sciuri DSM 20345 and S. sciuri NCTC12103 correspond to the same type strain (also known as ATCC 29062). (B) Comparative diagram of S. sciuri C2865 (pink ring) (reference genome) against its three closest genomes (S. sciuri Z8 [green ring], S. sciuri SNUD18 [blue], and S. sciuri CCUG39509 [gray]) as a set of concentric rings, where color indicates a BLAST match using BLAST Ring Image Generator (BRIG) (83). GC content and GC skew of reference genome is also displayed. Several features of interest are depicted, highlighting the unique presence of remarkable mobile genetic elements (SCCmec, C2865-pp3, c Tn554, C2865-pp1, C2865-pp2, SscPIC2865) in S. sciuri C2865. In addition, the genome location of the methicillin-susceptible mecA1 gene, two intrinsic dhfr genes (dhfr1 and dhfr2, the former next to a thy gene), a bacterial chromosomal hicA-antitoxin hicB (type II TA system), and the most relevant adaptive genes from the novel chromosomally located MGEs are displayed.

FIG 3
Comparative analysis of novel pUR2865-34 with the closest sequences in NCBI. Arrows denote the genes, length, and orientation. Gene colors other than gray represent the following genes of interest (color): antimicrobial resistance genes (pink), genes involve in metal resistance or transport (bright blue), intercellular adhesion gene cluster (ica) genes (navy blue), genes involved in transcription regulation (bright green), plasmid backbone genes (dark green), genes involved in transposition or recombination (yellow), and insertion sequences with defined imperfect inverted repeats (boxed and black). Areas of nucleotide similarity (nblastn, .100 bp match, .80% identity) between strains/structures are indicated in gray. For plasmids pSA-01, pAFS11, and S. sciuri FDAARGOS plasmid unnamed, only the area of interest is represented. (A) S. sciuri C2865 pUR2865-34 underlining its modular organization and indicating strain/plasmid sequences detected in NCBI with $50% coverage and $95% identity carrying those modules. The novel predicted secondary (Continued on next page) Novel dfrE and Mobile Elements in Distinct S. sciuri sequence identity ( Table 3). The dfrE-carrying strains corresponded to all the staphylococcal isolates harboring the res/rec3 plus M. caseolyticus strain JCSC5402 and Exiguobacterium sp. strain S3-2, the latter harboring the dfrE gene in two different coexisting plasmids (Table 3). All dfrE-carrying staphylococcal genomes corresponded to WGS projects with direct submission to NCBI. In all these cases, the dfrE gene was unnoticed and included within MDR plasmids or plasmid-associated elements hosting additional AMR genes (Table 3). Of note, all strains harboring the overlooked dfrE were reported in Asian countries from animal-related or clinical human samples (Table 3). Only Yang et al. (40) denoted an identical gene, designated "dfr_like," in Exiguobacterium sp. S3-2 from a fish farm sediment in China and proved its ability to confer TMP resistance in Gram-negative Escherichia coli DH5a. Hence, the role of this gene, here renamed dfrE (for Exiguobacterium spp), in TMP resistance in staphylococci remained open. DfrE revealed phylogenetically distant from the intrinsic Dhfr (Fig. S5) and was closest to the Dhfr of soil-related Paenibacillus anaericanus (68% identity) (Fig. 4). Within the staphylococcal TMP resistance genes, DfrE shared closer phylogenetic identity with TMP resistance DfrF, typical of enterococci and streptococci (Fig. S5).
The pUR2865-34 backbone region contains a type Ib partitioning system and a repA_N replication initiation gene (repA), characteristic of theta-replicating plasmids (42). It also harbors a single relaxase gene (mob) with an origin-of-transfer sequence (oriT) located immediately upstream as part of the integrated pUR2865-int (Fig. 3). In addition, an oriT mimic sequence (positions 6,077 to 6,132) similar to one recently described by Bukowski et al. (43) was observed proximal to the replication and partitioning genes (Fig. 3A). It is important to note that although pUR2865-34 does not carry any conjugative genes, these oriT regions can mediate the transfer of plasmid-carried phenotypes, such as resistance to trimethoprim, in the presence of conjugative elements.
Novel dfrE and Mobile Elements in Distinct S. sciuri C2853, C2854, and C2855 harbored a similar pUR2865-34 plasmid and enclosed the dfrE and ica-gene cluster (see Text S1 for details).
The novel dfrE gene confers high-level resistance to trimethoprim. All S. aureus RN4220 transformants carrying the different types of dfrE-carrying constructs exhibited a 2,048-fold increase in TMP resistance with respect to the empty S. aureus RN4220, irrespective of the presence of the natural or constitutive promoter located upstream of dfrE (Table 4). These results report the activity of the dfrE in S. aureus for the first time in literature. Likewise, all E. coli dfrE-carrying DH5aa transformants exhibited a 2,046-or 4,096fold increase in TMP resistance with respect to the control ( Table 4). Lack of synergistic activity was observed when dfrE and thy were cloned together into E. coli DH5a and S. aureus RN4220. Two S. aureus RN4220 transformants carrying entire pUR2865-34, named S319 and S320, exhibited a 2,048-fold increase in TMP resistance with respect to control strains and displayed additional resistance to tetracycline, erythromycin, clindamycin, gentamicin, kanamycin, and tobramycin, as confirmed by disc-diffusion agar tests.  The tetracycline resistance tet(S) was only detected after WGS in this study.
Novel dfrE and Mobile Elements in Distinct S. sciuri Novel SCCmec C2865 element lacking formerly described chromosomal cassette recombinases. A novel MDR SCCmec cassette, denominated SCCmec C2865 , was identified at the 39-end region of 23S rRNA [pseudouridine(1915)-N(3)]-methyltransferase RlmH gene (rmlH). SCCmec C2865 was 55,137 bp in size and contained a class A mec gene complex (IS431-mecA-mecR1-mecI) (Fig. 5A). None of the so-far-described staphylococcal or macrococcal chromosome cassette recombinase genes (ccr), the resultant large serine recombinases (LSR) of which are responsible for excision and integration of the cassette, were detected. SCCmec C2865 was delimited at both ends by characteristic SCCmec-flanking direct repeats (DRs) with typical insertion site sequences (ISS). Characteristic imperfect inverted repeats (IRs), required for CcrAB or CcrC recognition of the att sites, were located at the internal boundaries of the complete cassette ( Fig. 5A; Text. S1) (44). The SCCmec C2865 rlmH-proximal region (attR-attL1) carried a class A mec gene complex and one transposon-like structure consisting of one tetracycline resistance gene tet(S), coding for a translation elongation factor G (EF-G) involved in tetracycline ribosomal protection, flanked by two IS1216E copies in the same orientation. No DRs were present, which could evidence its mobilization as a composite transposon (Tn) or translocatable unit (Fig. 5A). A streptomycin resistance gene aadE, coding for a 6-aminoglycoside adenylyltransferase, was detected immediately downstream of the 39-end IS1216E copy. In addition, an arsenic resistance operon (arsCBR) and a copper resistance gene (copB), coding for a copper-translocating P-type ATPase, were located (Fig. 5A). SCCmecTXG24 from S. sciuri TXG24 and SCCmecGVGS2 from S. sciuri GVGS2 were identified as closest relatives to SCCmec C2865 , sharing 57% and 40% coverage, respectively (Fig. 5A) (20,45).
Protein domain analysis revealed a phage-related LSR, designated LSR SCCmec , which encompassed the typical PF00239, PF07508, and PF13408 domains, also present in Ccrs. As Ccrs can still transpose SCCmec elements when located elsewhere in the genome (46)
C2865-pp3 was 126,192 bp and consisted of 166 CDSs, of which only 34 (20.5%) had predicted functions. C2865-pp3 had a GC content of 30.8% and was integrated truncating a yeeE/yedE gene, coding for an inner membrane protein, positioned 819,965 bp downstream of dnaA. C2865-pp3 genome was organized into three proposed modules delimited by three different integrases, one LSR and two Y-Int/Rec. The combined modules contained genes involved in lysogeny, DNA replication and metabolism, virion packaging, phage morphogenesis, cell lysis, and adaptation ( Fig. 7A; see Text S1 for details). Two putative elements involved in phage adaption to the host were detected at its left and central regions, hicA-hicB and ardA, respectively. The putative HicA and HicB belong to type II toxin-antitoxin systems, where the toxin (HicA) acts as mRNA interferase and the antitoxin (HicB) as neutralizer. The ardA gene coded for a putative antirestriction protein ArdA with an N-terminal domain (PF07275) involved in evasion of the bacterial type I restrictionmodification system. Moreover, C2865-pp3 harbored a tRNA, which shared 86.7% nucleotide identity with the corresponding tRNA of several Listeria phages (see supplemental material for analysis of tRNA codon and amino acid usage; Fig. S9). The integration of C2865-pp3 generated two 8-bp perfect DRs (59-GTACTTGG-39) at its boundaries (Fig. 7A). C2865-pp3 shared the closest identity to the prophage-like element of mouse Staphylococcus lentus HT5 and to Staphylococcus phage SPbeta-like, obtained from a clinical Staphylococcus epidermidis 36-1, both carrying three different integrases ( Fig. 7B; Text S1). Importantly, the left-arm region of SPbeta-like phage enclosed a composite transposon-like element (IS257-flanked) carrying the TMP resistance gene dfrA and transposon Tn4001 (IS256-aacA/aphD-IS256), which harbors the aminoglycoside resistance gene aacA-aphD (Fig. 7A).
Sequencing analyses of potential circular intermediates (CIs), as well as restorage of respective phage chromosomal integration genes, revealed that the three prophages could excise the bacterial genome. Phylogenetic analysis of terminase TerL plus the presence of an HNH endonuclease gene in front of terminase terS, involved in DNA packaging of cos phages (49), predicted a cohesive end packaging strategy (cos) for C2865-pp1 and C2865-pp2 (Text S1 ; Fig. S8).
The novel S. sciuri pathogenicity island (SscPIC2865) belongs to the SaPI4 family, and related elements are disclosed to be common among S. sciuri genomes. A novel phage-inducible chromosomal island (PICI) of 13,101 bp, named SscPIC2865, was detected 20,705 bp upstream of dnaA gene. SscPIC2865 was classified as an S. sciuri pathogenicity island (SscPI), as it contains homologues of the entire core set of genes characteristic of Comparative analysis of the novel staphylococcal prophages C2865-pp1 and C2865-pp2 integrated in the chromosome of S. sciuri strain C2865 and its closest relatives based on (i) all staphylococcal phages deposited in the Viral RefSeq database and (ii) genetic elements enclosing the closest integrases in the NCBI NR database. Arrows denote the genes, length, and orientation. Gene colors other than gray (hypothetical proteins) and gray-turquoise (others) represent genes involved in the following processes (color): chromosomal integration (yellow), regulation (green), DNA metabolism (orange), DNA packaging (dark green), phage morphogenesis (bright blue), cell lysis (purple), host adaptation (navy blue). Areas of identity (tblastx, .50 aa match, .50% identity) between (pro)phages are indicated in the color scale. (B) Maximum likelihood phylogenetic tree of C2865-pp1 and C2865-pp2 integrases (Int), as well as all integrases sharing .80% amino acid identity with respect to either C2865-pp1 and C2865-pp2 integrases and integrase of Enterobacteria lambda phage, used as outgroup (colored in blue). C2865-pp1 and C2865-pp2 integrases are highlighted in faint yellow. Percentage of identity is indicated, with that sharing highest identity to respective integrase highlighted in faint green.
S. aureus pathogenicity islands (SaPIs) and displays synteny with previously characterized SaPIs (Text S1). SscPIC2865 consisted of 21 CDSs, of which 9 (42.9%) lacked any predicted function (Fig. 8A). SscPIC2865 enclosed a tyrosine recombinase (Y-Int/Rec) directing chromosomal integration at the end of the 39-end of the 30S ribosomal protein S18 gene (rpsR). Two SscPIC2865-flanking 15-bp DRs (59-AAAGAAGAACAATAA-39) constituted its integrase attachment core sites (Fig. 8A). None of the genes involved in interference of the helper phage reproduction in favor of SscPIC2865 packaging were identified. However, this 39 region carried a putative transcriptional regulator, an RNA polymerase sigma factor, and a putative helicase, which might be involved in the replication of the element (Fig. 8A). Interestingly, a putative mazF toxin gene, coding for an endoribonuclease (mRNA interferase) of a type II toxin-antitoxin system (MazF/MazE), was identified in the accessory region of SscPIc2865.
Several integrases sharing .80% identity to SscPIc2865 integrase were identified, all truncating rpsR (Fig. 8B). Of these, nine belong to S. sciuri and the rest to Staphylococcus sciuri strain C2865 and its closest relatives based on (i) all staphylococcal phages deposited in the Viral RefSeq database as well as on (ii) genetic elements enclosing the closest integrases in the NCBI NR database. Arrows denote the genes, length, and orientation. Gene colors other than gray (hypothetical proteins) and gray-turquoise (others) represent genes involved in the following processes (color): chromosomal integration (yellow), regulation (green), DNA metabolism (orange), DNA packaging (dark green), phage morphogenesis (bright blue), cell lysis (purple), host adaptation (navy blue), antimicrobial resistance (pink). Areas of identity (tblastx, .50 aa match, .50% identity) between (pro)phages are indicated in the color scale. Staphylococcus phage SPbeta-like was linearized at position of interest. Area with green background denotes the region enclosing an independent mobile genetic region (flanked by two IS257 copies in the same orientation), which encompasses the transposon Tn4001 (IS256-aacA/aphD-IS256) as well as the trimethoprim resistance gene dfrA. (B) Maximum likelihood phylogenetic tree of the three integrases (Int) detected in C2865-pp3, as well as those present in its two closest (pro)phages and closest integrases (.80% amino acid identity) and integrase of Enterobacteria lambda phage, used as outgroup (colored in blue). Integrases from C2865-pp3 are highlighted in faint yellow. Percentage of identity is indicated, with that sharing highest identity to respective integrase highlighted in faint green. schleiferi, S. lentus, and Staphylococcus fleurettii. Their integrase-surrounding regions share the characteristic structure of SaPIs (data not shown). We denoted that these elements belong to the SaPI4 family (integrase group I), according to the integrase sequence homology, integration site, and conserved core site 59-AAAGAAGAACAATAA-39 (Text S1; Fig. 8C). Sequencing analyses of potential circular intermediates (CIs) and restorage of SscPIC2865 chromosomal integration site revealed that SscPIC2865 could excise the bacterial chromosome.

DISCUSSION
The combination of Illumina followed by deep PacBio sequencing using large DNA fragmentation and size selection allowed the resolution of the complete genome of ISrich MDR S. sciuri C2865 at a minimum error rate. This enabled the discovery of transferable TMP resistance dfrE within a novel mosaic plasmid and additional novel MGEs. MRSS C2865 revealed high genome plasticity with respect to the S. sciuri genomes analyzed, as it was enriched in novel unique chromosomal and extrachromosomal elements.
We display for the first time that S. sciuri has an open pan-genome, with an everincreasing flexible genome as new strains are added. Similar patterns were observed in other clinically relevant staphylococcal species formerly analyzed, such as S. aureus (50), S. epidermidis (51)(52)(53), Staphylococcus haemolyticus (54), Staphylococcus capitis, and Staphylococcus caprae (52). This result could be anticipated, as an open pan-genome is characteristic of species living in multiple environments and/or mixed bacterial communities. This feature facilitates multiple ways of exchanging genetic material, enabling them to unceasingly expand their total gene repertoire. Our data set revealed MRSS C2865 as the major acceptor of adaptive mobile traits, as it exhibited the highest number of unique genes, most of them (75%) corresponding to MGEs. However, this was not correlated with phylogeny, as C2865 closest relatives (SNUDS-18 and Z8) did not disclose this sharp profile. Based on a range of phenotypic, biochemical, physiological, and genetic analyses, Svec et al. (55) recently showed high S. sciuri intraspecies heterogeneity with no clear differentiation into different subspecies. While agreeing with those statements, our whole-genome-based phylogenomic analysis revealed a clear distinction of an intraspecies subcluster, which included MRSS C2865 (96% ANI with the rest of the S. sciuri genomes). This ANI value falls just above the threshold to be considered a different species (%95%) (56,57). Hence, we suggest that there indeed may be subspecies discrimination among the S. sciuri species, but such distinction needs to be addressed by WGS comparisons and might not correspond to the outcomes retrieved by the above-mentioned traditional methods. Alternatively, the available metadata of compared genomes evidence the absence of clear phylogeny delineation depending on the origin, source, or host. This corroborates the low host tropism suggested for this species (3).
We reveal that the novel dfrE confers high-level resistance to TMP in both staphylococci and E. coli. DfrE is phylogenetically distant from all staphylococcal TMP resistance genes and displays a common ancestor with Dfrs from soil-associated P. anaericanus. Scarce data are available on P. anaericanus, but bacteria belonging to this genus are ubiquitous in nature, and closest species have been reported in different environmental sources, such as soils and rhizosphere from different crops (58)(59)(60). Importantly, dfrE evaded discovery in all available dfrE-carrying genomes except Exiguobacterium sp. (dfr_like), where it proved to confer TMP resistance in E. coli. Exiguobacterium spp. are extremophiles adapted to a wide range of habitats, including cold environments (40,(61)(62)(63)(64). We also identify here that dfrE is already distributed in Asia, as it is present in at least three different staphylococcal species (S. sciuri, S. aureus, and S. arlettae) of human and animal origin (including clinical samples) as well as in animal-associated M. caseolyticus strain JCSC5402 (65) from several countries. Importantly, all dfrE-carrying genomes harbored it in MDR plasmids or plasmid-associated elements. This colocalization is of concern, because it enables the transfer of diverse adaptive traits via a single HGT event. Arrows denote the genes, length, and orientation. The core integration site sequence detected (59-AAAGAAGAACAATAA-39) as well as position for integration (attC for putative integration site in S. sciuri SNUDS-18 chromosome; attL and attR at both extremities of SsPIC2865) are indicated. Areas of nucleotide similarity (nblastn, .100 bp match, .99% identity) between both structures are indicated in gray. (B) Maximum likelihood phylogenetic tree of (left) all sitespecific integrases (Int) deposited in the protein NCBI database and (right) all terminases small subunit (TerS) sharing .80% amino acid identity with respect to those present in SscPIC2865, in addition to those of prototype SaPI4 from S. aureus MRSA252 (69) (displayed in blue). Strains with Int but either lack of or divergent (,50% ID) TerS, and vice versa, are indicated in yellow or green color, respectively. Percentages of amino acid identity with respect to Int or TerS of SscPIC2865 are displayed on the right of the respective tree. TerS-carrying strains corresponded to those harboring the SscPIC2865-conserved integrase in all cases but one (S. sciuri SNUC70), which was integrated at a different position of the bacterial chromosome (at tRNA, ssrA gene). (C) Sequence comparison analysis of the phage-related chromosomal island att sequence core site region (attL and attR) of SscPIC2865 and corresponding region of all strains carrying a similar integrase (.80% identity), in addition to that of S. aureus MRSA252 (harboring prototype SaPI4), which also integrates at the 39-end of the S18 ribosomal gene rpsR. Underlined area represents the proposed integration core sequence of SscPIC2865 and related elements. Bases differing from the att region of S. sciuri C2865 SscPIC2865 are indicated in red. Original att core site length of SaPI4 is indicated in dark blue (69), and the additional direct repeat is shown in blue.
Our data reflect the transferability of TMP resistance via dfrE across diverse bacteria from different taxonomic families (order Bacillales), diverse environments, and geographically distant regions. Of note, in Exiguobacterium sp. strain S3-2, dfrE was enclosed within a conserved Tn3-like transposon (40). MRSS C2865 (pUR2865-34) and all additional dfrEcarrying strains harbored a truncated version of this element. In particular, the dfrE-carrying region of S. sciuri GN5-1 (pSS-04) seems to have evolved afterwards by losing an additional recombinase-carrying segment from the Tn3-like remnant. This indicates that Exiguobacterium sp. harbored an ancestral dfrE transposable element that has recombined and jumped to different species from a single ancestor. The fact that the dfrE element is already detected in several unrelated animal S. sciuri strains from China and Nigeria, and that S. sciuri GN5-1 has an evolved dfrE segment, points toward the assumption that S. sciuri is the ancestral species within the genus, as several recombination events have already occurred within this region. Unfortunately, no flanking regions of the dfr gene of P. anaericanus are available, so potential homology with its surrounding area could not be addressed. The backbone of mosaic pUR2865-34 denoted a theta replication mechanism, characteristic of larger staphylococcal plasmids. The putative type Ib partitioning system detected is well studied in Gram-negative bacteria, but little to nothing is known about its functions in Gram-positive cocci. These systems contribute to the prevalence and spread of these plasmids, ensuring stable inheritance and effectively maintaining resistance in the absence of selection. pUR2865-34 could only move by conjugation via the detected oriT mimic sequence in the presence of a pWBG749-like plasmid or via the integrated mob gene and associated oriT of pUR2865-int in the presence of a conjugative plasmid (42). However, MRSS C2865 did not harbor any member of the three known staphylococcal conjugative plasmid families (pSK41, pWBG4, and pWBG749) or any recognizable transfer gene cluster that could act in trans.
We identified a novel ccr-lacking SCCmec C2865 encompassing additional AMR genes, including an IS1216E-flanked region carrying the tetracycline resistance gene tet(S). This gene has been found among Firmicutes and Gammaproteobacteria from diverse ecological sources since the 1950s (66). However, the tet(S) has been detected only twice before in staphylococci: (i) among methicillin-resistant Staphylococcus aureus (MRSA) isolates from animal carcasses (67) and (ii) within the S. sciuri SCCmec C2865related SCCmecTG24 cassette, from ready-to-eat meat (45). The additional closest cassette corresponded to mecA-mecC hybrid SCCmec-mecC in S. sciuri GVGS2 from a bovine infection (20). These strain sources highlight that S. sciuri isolates from animals behave as reservoirs for mosaic MGEs carrying AMR genes.
MRSS C2865 resulted in a polylysogen with three novel unrelated prophages, two of them harboring several putative adaptive features that may promote bacteria and/ or phage survival. Only these prophages constituted over 7% of MRSS C2865 genome, evidencing their role as drivers of bacterial evolution and genome modulation. As recently observed by Oliveira et al. (68), a low rate of staphylococcal phages may exhibit more than one site-specific recombinase, as was the case for phage C2865-pp3. Phage integrases are required for the establishment of the lysogeny, but their maintenance is also dependent on the regulatory proteins next to them (68). At least two of its three integrases were surrounded by transcriptional regulators and/or antirepressor proteins. This may enable the integration at additional chromosomal sites depending on the integrase used. In fact, the closest putative prophage, integrated into S. lentus HT5, was incorporated at a different location. Importantly, C2865-pp3-related S. epidermidis SPbeta-like phage harbored an AMR gene cluster enclosed within a composite transposon-like element, harboring resistance determinants for aminoglycosides and TMP. This is a central feature demonstrating that C2865-pp3-related functional phages can indeed harbor mobilizable AMR genes.
PICIs are characterized by a specific set of phage-related functions that enable them to hijack the phage lytic reproduction cycle of helper phages for their own highly efficient transduction. These elements often carry critical staphylococcal virulence genes (69). Several staphylococcal strains harbored SscPIC2865-related PICIs at the same integration site, disclosing that these elements are common among S. sciuri and contribute to genome plasticity and potential toxigenicity. Hence, the ribosomal protein S18 (rpsR gene) appears as a hub for the integration of mobile islands and should be considered when addressing staphylococcal chromosomal-integrated elements.
The conservation of the integrase-generated DRs flanking the three prophage genomes, together with their ability to excise the bacterial chromosomal DNA and circularize, as well as the high genome synteny observed among functional phages, strongly suggest that the novel prophages are functional. If this is the case, along with progeny generation, also plasmids or chromosomal DNA of the bacterial host may be mistakenly encapsidated (26). Hence, it is conceivable that any of these prophages may package and mobilize the MGEs discovered here, highlighting any of the three nonconjugative plasmids, the novel ccr-lacking SCCmec, and/or SscPIC2865 of the SaPI4 family. In fact, PICIs of the SaPI4 family are only induced by endogenous prophages (69). Upcoming studies are warranted to explore the transduction ability of these prophages.
We unveil a transferable TMP resistance gene that has so far evaded identification; although dfrE is already present in diverse species from varied sources geographically distant, dfrE seems to have emerged from a single ancestor. Molecular data analysis suggests S. sciuri as the donor for transfer within staphylococci. Since dfrE seems not yet common in staphylococcal clinical specimens, the data presented here enable early surveillance and facilitate molecular diagnosis, which could promote spread mitigation. Additional experimental analyses may elucidate the spread vehicle and routes for this evolving element. We highlight S. sciuri as a resourceful hub for diverse mobilizable adaptive traits.

MATERIALS AND METHODS
Bacterial strain selection and characteristics. MRSS strains C2865, C2853, C2854, and C2855 were obtained in a former study on the occurrence of methicillin-resistant coagulase negative staphylococci from canine samples and were recovered from the groin area of unrelated stranded dogs in Nsukka, Nigeria (12). These strains exhibited TMP resistance, while none of the known staphylococcal TMP resistance genes (dfrS1, dfrD, dfrG, dfrK) (32,33,35,36) nor the typical streptococcal/enterococcal dfrF gene (34,37) was detected. All strains were MDR and exhibited related AMR patterns. Strain C2865 was selected for WGS to unveil the genetic basis for TMP resistance. MRSS C2865 was chosen for WGS as it harbored the highest number of detected AMR genes.
DNA extraction, whole-genome sequencing, assembly, and annotation. Detailed DNA extraction procedures are described in the supplemental material. Briefly, genomic DNA was extracted by two different methods: for Illumina sequencing, the Wizard Genomic DNA purification kit was used including both lysozyme and lysostaphin (10 mg/ml each) (A1120, Promega Corporation, Spain), and for PacBio sequencing, DNA was isolated using a phenol-chloroform method with some modifications for improved cell lysis. Plasmid DNA was obtained using the GenElute Plasmid Miniprep kit (PLN350, Sigma) also including a lysis step with lysozyme (2.5 mg/ml) and lysostaphin (0.25 mg/ml) for 20 min at 37°C after the resuspension solution step.
High-throughput WGS of S. sciuri C2865 DNA was performed with Illumina Miseq (2 Â 300 bp), with NEBNext Ultra kit. For long-read sequencing, PacBio RSII was used after DNA fragmentation of 15 kb followed by a mild size selection. Illumina read quality was checked by Fastqc (103), and reads were trimmed using Trimmomatic v0.36 (70). Good-quality Illumina reads were de novo assembled using SPAdes (71). PacBio RSII raw reads were assembled using Canu (72). Protein-coding genes, tRNAs, and rRNA operons were predicted using Prodigal (73), tRNAscan-SE, and RNAmmer on both data sets (74). Predicted protein sequences were compared against the NCBI NR database (NCBI nonredundant database) using DIAMOND (75) and against COG (76) and TIGFRAM (77) using HMMscan (78) for taxonomic and functional annotation. Genomic alignment dot plots between Illumina-and PacBio-resultant contigs were generated with D-GENIES software to evaluate consistency and reliability of both sequencing and assembly approaches (79). Amino acid and codon usages were determined for C2865 chromosome and prophages using the compareM package (https://github.com/dparks1134/CompareM).
Pan-genome examination, S. sciuri phylogenomic analysis, and comparative genomics. All genomes available belonging to the S. sciuri group species were downloaded from the NCBI database (accessed in April 2018). In total, 30 strains (29 from NCBI, MRSS C2865) were included, which belonged to the following species (no. of strains): S. sciuri (21), S. lentus (5), Staphylococcus vitulinus (2), S. fleuretti (1), and Staphylococcus stepanovicii (1) (see the supplemental material for strain characteristics). In order to measure the probability of two genomes belonging to the same species, an ANI of the 30 S. sciuri group genomes was calculated using JSpecies as indicated before (56). A heat map was generated using the ANI matrix output table with R (80). In parallel, a maximum likelihood tree for all the S. sciuri group genomes was generated using RAxML (version 7.2.6) (81) using core alignment obtained with Parsnp software within Harvest Suite package (82). Before phylogenomic analysis, all the genomic regions where recombination was Novel dfrE and Mobile Elements in Distinct S. sciuri detected were removed from the alignment. These regions were determined using the Gingr software, also included in the Harvest Suite package (82), which uses as input the output obtained directly from the Pasnp software. The results were visualized using iTOL v6 (https://itol.embl.de/). All S. sciuri group species other than S. sciuri were considered an outgroup for graphical representation. BLAST Ring Image Generator (BRIG) was used to evaluate and visualize comparisons between MRSS strain C2865 and its closest genomes (.98% ANI), using C2865 as reference (83). Pan-genome analysis (core plus accessory genome) for the 21 S. sciuri genomes was carried out using Roary with a 95% identity cutoff value (84).
Detection and analysis of resistome and mobilome from S. sciuri C2865. Antimicrobial resistance (AMR) genes. AMR genes formerly detected in MRSS C2865 (see Table S1 in reference 85) were blasted against the WGS of strain C2865. Contig(s) were manually checked for redundant dhfr genes and for additional resistance genes of interest.
Plasmids. Plasmid contig identification and plasmid reconstruction were achieved by contig coverage, sequence similarity with plasmid backbone genes, gene composition and organization, and contig boundaries redundancy (circularity). Putative oriT and oriT mimics on the mobilizable elements were searched using the core sequence of those from conjugative plasmids (86). For RCR plasmids, dso and sso regions, involved in the initiation of replication of the leading and lagging strand, respectively, were searched using core sequences of an RCR representative per plasmid family (87). The secondary structures of dso, sso, oriT, and oriT mimic were generated using Mfold web server for single-stranded linear DNA at default parameters (88).
Prophages. Detailed analyses are described in the supplemental material. Manual inspection of phage-associated genes (morphology/structure, lysogeny, cell lysis, DNA metabolism) and characteristic functional modular organization was implemented for phage confirmation and integrity. Integrase motif and domain analysis of the translated candidate CDSs was performed against ScanProsite database (89), Pfam database (90), and NCBI conserved domain database (CDD) (91). Integrase-directed generation of DRs as a result of genome integration was investigated manually by sequence comparison of bacterial chromosome-prophage boundaries using a S. sciuri strain prototype (SNUDS-18, GenBank accession no. CP020377) lacking those prophages. ANI pairwise comparison between C2865-enclosed prophages and all staphylococcal phage genomes available in the Viral RefSeq database (accessed until November 2018, n = 187) were calculated using the JSpecies with default parameters (56). A heat map was generated using the ANI matrix output table with R (80). Others. Additional recombinases associated with MGEs were investigated as described above for motif and domain analysis (supplemental material). SscPI chromosomal integration site was compared with similar integrase-carrying genomic islands (supplemental material). Integrase-directed generation of DRs as a result of SscPI integration was analyzed by sequence comparison of bacterial chromosome-SscPI boundaries with corresponding regions of S. sciuri strain SNUDS-18, which lacks any insertion in that region. In addition, ISsaga2 web tool was used for IS identification and quantification (92).
Excision ability of chromosomally located MGEs. Potential excision and circularization of selected chromosomally located mobile elements, in addition to detection of the resultant chromosomal region after excision, were tested by specific inverse and conventional PCR, respectively (supplemental material; see Table S1 in reference 85).
Phylogenetic analyses of proteins of interest. Phylogenetic analyses for Dhfr/Dfr and Ccr of interest were investigated by the construction of unrooted phylogenetic networks with SplitsTree v4 (93) using a neighbor-net with default parameters. Evidence for phylogenetic heterogeneity due to recombination was conducted with SplitsTree v4 using the Phi test for recombination option.
A maximum likelihood phylogenetic tree for MRSS C2865 prophage integrases and all related integrases was built upon amino acid sequence alignment as indicated above (using CLC Genomics Workbench 11) (supplemental material). To estimate the packaging mechanism of identified prophages, a circular phylogenetic tree of the terminase large subunit (TerL) from MRSS C2865 prophages and those identified in the 187 staphylococcal phages deposited in the Viral RefSeq database was likewise created upon amino acid sequence alignment using CLUSTALW (CLC Genomics Workbench 11). TerL of staphylococcal phages with known packaging mechanism was indicated (95).
Phylogenetic analyses for all site-specific tyrosine recombinases and terminases small subunit (TerS) related to those of SscPI in C2865 were performed by the construction of a maximum likelihood tree using the MEGA 7.0.21 program (supplemental material) (94).
Construction of recombinant plasmids to analyze DfrE functionality, transfer assays, and analysis of native plasmid integrity. To address the functionality of the candidate TMP resistance gene and for potential synergistic activity by its immediate 39-end thymidylate synthase gene (thy), three different dfrE-containing regions were amplified and cloned into pBUS-Pcap-HC (constitutive promoter) or pBUS-HC vectors using the Gibson assembly workflow: (i) dfrE gene alone, (ii) dfrE gene plus flanking regions, and (iii) dfrE gene, thy gene, and flanking regions (Table S2; supplemental material) (96).
Constructs of interest were transformed into Escherichia coli DH5a and subsequently into S. aureus RN4220 (supplemental material). In addition, native dfrE-containing plasmid was transformed into RN4220.
Phenotypic characterization of biofilm formation. Biofilm formation ability of the four S. sciuri strains as well as S. aureus RN4220/pUR2865-34 transformants was tested by a modified Congo red agar (CRAmod) assay and by crystal violet microtiter plate assay (see supplemental material for detailed procedures) (98,99). S. aureus SA113 and S. aureus ATCC 25923 were used as positive controls for their strong biofilm-forming potential, while S. aureus RN4220 (DSM 26309) was selected as negative control for biofilm production.
PCR detection of the TMP resistance gene dfrE and the ica-locus variant. Primers were designed for the detection of the TMP resistance dfrE gene, the biofilm formation ica-locus (icaADBC) variant genes, and the ica-locus repressor (icaR) (see Table S1 in reference 85). For the dfrE gene, primers designed enclosed the Dhfr superfamily domain region (Cd-Search PF00186). All dhfr genes with $45% nucleotide similarity were included to search for specificity of dfrE gene primer set. Original strains (C2853, C2854, C2855), RN4220/pUR2865-34 transformants, and several control strains were tested for sensitivity and specificity (supplemental material).
Data availability. The whole-genome shotgun sequence data supporting this work are available via NCBI SRA database under BioProject and Biosample accession numbers PRJNA663854 and SAMN16182282, respectively.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, PDF file, 0.5 MB. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors have declared that no competing interests exist.

ADDENDUM IN PROOF
During the time this paper was prepared for publication, the taxonomy of the Staphylococcaceae family has been revisited. Following phylogenomic analyses, Madhaiyan et al. (104) propose species belonging to the S. sciuri species group (S. sciuri, S. fleurettii, S. lentus, S. stepanovicii, and S. vitulinus) be reassigned to the novel genus Mammaliicoccus, with Mammaliicoccus sciuri as the type species.