Whole-genome analysis showed the promotion of genetic diversity and coevolution in Staphylococcus aureus lytic bacteriophages and their hosts mediated by prophages via worldwide recombination events

Prophages as a part of Staphylococcus aureus genome contribute to the genetic diversity as well as survival strategies of their host. Some S. aureus prophages also have an imminent risk of host cell lysis and become a lytic phage. Nonetheless, interactions among S. aureus prophages, lytic phages, and their hosts, as well as the genetic diversity of S. aureus prophages, remain unclear. We identified 579 intact and 1,389 incomplete prophages in the genomes of 493 S. aureus isolates obtained from the NCBI database. The structural diversity and gene content of intact and incomplete prophages were investigated and compared with 188 lytic phages. Mosaic structure comparison, ortholog group clustering, phylogenetic analysis, and recombination network analysis were performed to estimate genetic relatedness among S. aureus intact prophages, incomplete prophages, and lytic phages. The intact and incomplete prophages harbored 148 and 522 distinct mosaic structures, respectively. The major difference between lytic phages and prophages was the lack of functional modules and genes. Compared to the lytic phages, both the S. aureus intact and incomplete prophages harbored multiple antimicrobial resistance (AMR) and virulence factor (VF) genes. Several functional modules of lytic phages 3_AJ_2017 and 23MRA shared more than 99% nucleotide sequence identity with S. aureus intact (ST20130943_p1 and UTSW_ MRSA_55_ip3) and incomplete prophages (SA3_LAU_ip3 and MRSA_FKTN_ip4); other modules showed little nucleotide sequence similarity. Ortholog and phylogenetic analyses revealed a common gene pool shared between the prophages and lytic Siphoviridae phages. Moreover, most shared sequences existed within intact (43428/137294, 31.6%) and incomplete prophages (41248/137294, 30.0%). Therefore, the maintenance or loss of functional modules in intact and incomplete prophages is key to balance the costs and benefits of large prophages harboring various AMR and VF genes in the bacterial host. The shared identical functional modules between S. aureus lytic phages and prophages are likely to result in the exchange, acquisition, and loss of functional modules, and therefore contribute to their genetic diversity. Moreover, constant recombination events within prophages globally were responsible for the coevolution of lytic phages and their bacterial hosts.


Introduction
The Staphylococcus aureus pathogen causes nosocomial and community-acquired infections in humans and animals owing to several immune resistance and evasion factors, toxins, and invasive mechanisms (Zhou et al., 2018;Guerillot et al., 2019). The threat of S. aureus can be attributed to a highly variable genome associated with phenotypic diversity and varying epidemiologic factors in different strains (Moon et al., 2015), with prophage genetic material accounting for 10-20% of the host genome (Johnson et al., 2022). Moreover, S. aureus can rapidly acquire antimicrobial resistance (AMR) and virulence factor (VF) genes via prophage transduction (Kondo et al., 2021). Studies on S. aureus prophage diversity at the nucleotide, structural, and genomic levels are necessary to understand the genomic plasticity and potential threats of S. aureus isolates.
Prophages originate from temperate bacteriophages (phages) that undergo a lysogenic cycle and integrate into the host chromosome, acting as a genetic reservoir and promoting diversity of their bacterial host (Ramisetty and Sudhakari, 2019;Johnson et al., 2022). A previous study indicated that distinct distribution of AMR and VF genes has been observed in intact and incomplete prophages in the genomes of seven pathogens, including 424 S. aureus isolates, derived from a public database (Kondo et al., 2021). Intact prophages can be characterized as genetic elements that can transmit VF genes and other genetic traits to S. aureus isolates (Moon et al., 2016). Despite the deletion of most phage genes, incomplete prophages or prophagelike elements occur stably in the genome of Lactococcus lactis IL1403 (Aucouturier et al., 2018). Nonetheless, the genetic relationship among lytic phages, intact prophages and incomplete prophages that is vital for elucidating the evolutionary relationships between S. aureus prophages and their bacterial host remains unclear.
Previously, the association of S. aureus prophages with increased bacterial pathogenicity and fitness has been examined (Kondo et al., 2021). The presence of different prophage-encoded VF and AMR genes among methicillin-resistant S. aureus (MRSA) strains enables them to produce a broad range of diseases (Dini et al., 2019). Furthermore, as genetic reservoirs, prophages promote diversity in hosts via enhanced recombination (Nadeem and Wahl, 2017). Besides, prophages mediate active horizontal gene transfer (HGT) via transduction, resulting in a large common gene pool shared by lytic phage genomes (Ramisetty and Sudhakari, 2019;Dion et al., 2020). Therefore, S. aureus prophages can remarkably influence the diversity and evolution of lytic phages and their hosts.
In this study, we identified 1968 prophages in 493 S. aureus isolate genomes and compared their sequences to those of 188 previously published lytic phages to (i) better elucidate structural diversity and gene content of S. aureus prophages, (ii) explore the genetic relationship between S. aureus intact and incomplete prophages, and (iii) understand the coevolutionary strategy of lytic phages and their hosts involving interaction with prophages.

Collection of Staphylococcus aureus prophage metadata
The genomic sequences of 500 S. aureus isolates including 83 complete genome sequences and 417 incomplete genomes generated by high-throughput sequencing technologies were downloaded from the NCBI Genome database (April 2022). PATRIC was used to access the genome quality, as previously described (Johnson et al., 2022). In the assessment, 493 sequences derived from 58 countries globally (North America, South America, Asia, Europe, Africa, and Oceania) were selected (Supplementary Table S1). These strains were isolated from various samples (humans, n = 346; bovine, n = 39; porcine, n = 20; poultry, n = 12; sheep, n = 5; canine, n = 5; other origins, n = 27; unknown origin, n = 36). The cohort comprised genomes of 188 methicillin-sensitive S. aureus and 302 MRSA isolates. Phage sequences were predicted using the PHASTER software (Arndt et al., 2016) and confirmed by CheckV (Nayfach et al., 2021). The prophage regions were classified as intact (Supplementary Table S2) or incomplete (Supplementary Table S3) using the PHASTER criteria.

Phylogenetic analysis
Phylogenetic trees were constructed based on single-nucleotide polymorphisms (SNPs) to assess evolutionary relationships of S. aureus intact and incomplete prophages, respectively. Another phylogenetic analysis was performed to estimate the genetic relationship among the representative 148 intact prophages (Supplementary Table S4), 522 incomplete prophages (Supplementary Table S5), and 188 previously published lytic phages. Phylogenetic trees of S. aureus prophage and lytic phage genomes were generated using the Erwinia phage phiEa2809 as an outgroup. SNPs were determined for all genomes using the kSNP3 (version 3.1) software package (Gardner et al., 2015). The k-mer size estimated using the Kchooser software was set to 15 (Gardner et al., 2015). The SNP-based phylogenetic trees were constructed based on the Maximum Likelihood methods by RAxML (v 8.2.12) with 1,000 bootstraps and the GTRCAT nucleotide substitution model (Stamatakis, 2014). The phylogenetic trees were rooted using the outgroup and annotated using iTOL (Letunic and Bork, 2021). The mosaic structure of the S. aureus prophages was aligned using progressive MAUVE (Zhou et al., 2022).

Recombination network
The recombination events between the 188 lytic phages and 1968 prophages were estimated as previously described (Belcaid et al., 2010). Each sequence was queried against each other in the database using local BLAST. Homologous results with a nucleotide sequences identity of >95% were considered as the shared segments (Owen et al., 2017). The shared segments among intact prophages, incomplete prophages and lytic phages from different geographic were used to establish a worldwide recombination network as previously described (Göller et al., 2021).

Statistical analyses
Pearson's chi-squared test (two-tailed) was performed to analyze the differences in the distribution of genes encoding for phage morphogenesis, DNA metabolism, host cell lysis, DNA packaging, lysogeny, AMR genes and VF among intact, incomplete, and lytic phages using SPSS software (version 26).
We determined two phylogenies for 579 S. aureus intact and 1,389 incomplete prophages, respectively. Except for 8 singletons, the phylogenetic tree based on the 49,189 SNPs from genome sequences of 579 S. aureus intact prophages and the Erwinia phage phiEa2809 revealed 2 major genetic lineages (I-II; Supplementary Figure S1). Lineage I consisted of 35 intact prophages and lineage II concluded 536 intact prophages. However, statistical analyses based on the Pearson's chi-square test showed no significant differences in the distribution of intact prophages from distinct geographic origin, host and isolation source between the two lineages (p > 0.05). Another phylogenetic tree based on the 292,450 SNPs from genome sequences of 1,389 S. aureus incomplete prophages and the Erwinia phage phiEa2809 revealed 51 singletons and 2 main lineages (I-II, Supplementary Figure S2). Lineage I consisted of 46 incomplete prophages and lineage II conclude 1,292 incomplete prophages. Consistently, there was no significant differences in the distribution of incomplete prophages from distinct geographic origin, host and isolation source between the two lineages (p > 0.05). The mosaic structure and gene content were compared across phylogenetic groups. Colinary and MARVE analyses revealed 148 mosaic structures among intact S. aureus prophages (Supplementary Table S4 All five intact prophages (CP1-CP5) typically consisted of genes associated with four functional modules: phage morphogenesis, host cell lysis, DNA packaging, and DNA metabolism ( Figure 1). CP1 was consist of three functional regions: host cell lysis and phage morphogenesis, DNA packaging, and DNA metabolism. Despite CP2 and CP3 comprised only one module, this module included genes Frontiers in Microbiology 04 frontiersin.org associated all four functions. CP4 comprised four functional regions including 2 DNA metabolism modules, 1 DNA packaging module, and 1 phage morphogenesis and host cell lysis module. CP5 comprised two functional regions: DNA metabolism, and phage morphogenesis and host cell lysis. However, incomplete prophages IP1-IP5 comprised only genes associated with phage morphogenesis, DNA metabolism, lysogeny, and virulence, but genes associated with host cell lysis and DNA packaging ( Figure 2).
A total of 493 S. aureus isolates were screened for VF genes (Supplementary Mosaic structure of the main structure of incomplete prophages IP1-5.  Heat map showing gene distribution associated with phage morphogenesis, host cell lysis, DNA metabolism, DNA packaging, and lysogeny of S. aureus 148 intact prophages, 522 incomplete prophages, and 16 lytic phages. Gene presence (colored blocks) or absence (white) is shown.

Staphylococcus aureus lytic phages and prophages shared identical functional modules
The mosaic structures of intact prophage ST20130943_p1, lytic phage 3_AJ_2017, and incomplete prophage SA3_LAU_ip3 were analyzed (Figure 4). The prophage ST20130943_p1 genome comprised 47,507 bp and three functional modules (host cell lysis module, phage morphogenesis and DNA packaging module, and DNA metabolism module). The host cell lysis module comprised eight ORFs with four virulence genes (hlb, scn, chp, and sak) and three host cell lysis genes (lys and hol). The phage morphogenesis and DNA packaging module comprised eight genes encoding phage morphogenesis proteins and two DNA packaging genes. The DNA metabolism module comprised four virulence genes (dut, hlb, lukG, and lukH) and one lysogeny gene (int). The lytic phage-3_AJ_2017 genome comprised 43,922 bp and three functional modules (host cell lysis module, phage morphogenesis and DNA packaging module, and DNA metabolism module). The DNA metabolism module comprised one lysogeny (int), one AMR gene (bla), and two virulence (dut and hlb) genes. The prophage SA3_ LAU_ip3 genome comprised only 28,267 bp and two functional modules (phage morphogenesis and DNA packaging module, and host cell lysis module). Notably, region A of the intact prophage ST20130943_p1 comprised 27,929 bp and 35 ORFs, including two functional modules (host cell lysis module, and phage morphogenesis and DNA packaging module). This region showed more than 99.0% nucleotide sequence identity with region B in the lytic phage 3_ AJ_2017 and region C in the incomplete prophage SA3_LAU_ip3.
The mosaic structures of intact prophage UTSW_MRSA_55_p3, lytic phage 23MRA, and incomplete prophage MRSA_FKTN_ip4 were also compared ( Figure 5). The incomplete prophage-UTSW_ MRSA_55_p3 comprised 70,093 bp and four functional modules including three DNA metabolism modules, and one host cell lysis and phage morphogenesis module. The first DNA metabolism module comprised one virulence gene (map). The host cell lysis and phage morphogenesis module comprised one lysogeny gene (clp), two host cell lysis genes (lys and hol), seven phage morphogenesis genes, four virulence genes (hlb, scn, chp, and sak) and one DNA packaging gene (terL). The second DNA metabolism module comprised three virulence genes (dut, pvl and hlb), and one lysogeny gene (int). The third DNA metabolism module comprised two virulence genes (lukG, and lukH), and one DNA packaging gene (terS). The lytic phage 23MRA genome comprised two modules (DNA metabolism module, and host cell lysis and phage morphogenesis module). The DNA metabolism module comprised two virulence genes (dut, and pvl), and one lysogeny gene (int). The host cell lysis and phage morphogenesis module comprised two lysin genes (lys), one holin gene (hol), one lysogeny gene (clp), four virulence genes (hlb, scn, chp, and sak) and one DNA packaging gene (terL). The incomplete prophage-MRSA_ FKTN_ip4 genome was 27,240 bp in size and comprised one DNA packaging gene (terL), one lysogeny gene (clp), two host cell lysis genes (hol and lys), and four virulence genes (sak, chp, scn, and hlb). Notably, region A in the UTSW_MRSA_55_p3 genome comprised one functional module (host cell lysis and phage morphogenesis module), which showed more than 99.0% nucleotide sequence identity with region D of the lytic phage 23MRA and region E of the incomplete prophage MRSA_FKTN_ip4. Moreover, region B in the UTSW_ MRSA_55_p3 genome comprised another functional module (DNA metabolism module) and showed more than 99.0% nucleotide sequence identity with region C of the lytic phage 23MRA.

Discussion
Prophages play potent roles in the pathogenicity, fitness and diversity of their S. aureus hosts via HGT. However, the genetic diversity of prophages at the nucleotide, structural, and genomic levels remains unknown. The both phylogenies for S. aureus intact and incomplete prophages revealed that prophages from distinct geographic origin and host were clustered together, indicating the genetic diversity of the S. aureus prophages. Besides, our previous study analyzed the genetic diversity of 188 S. aureus lytic phages and observed 16 mosaic structures (Zhou et al., 2022), while in this study, we identified as many as 670 distinct mosaic structures (148 intact and Comparative structural analyses of lytic phage-23MRA with intact prophage UTSW_MRSA_55_p3 and incomplete prophage MRSA_FKTN_ip4. Frontiers in Microbiology 09 frontiersin.org 522 incomplete prophages) in the genomes of prophages, indicating the extraordinary structural diversity of S. aureus prophages. The transduction of temperate bacteriophage phiSaBov is accompanied by mobilization of the prophages vSaα, vSaβ, and vSaγ (Moon et al., 2015(Moon et al., , 2016. Therefore, distinct mosaic structures of intact and incomplete prophages of S. aureus promote genomic plasticity in hosts and genetic diversity in S. aureus lytic phages. Intact and incomplete prophages exhibited distinct mosaic structures, which was consistent with those of a previous study indicating that mosaicism of phage genomes is perhaps one of the few common features (Dion et al., 2020). Our previous study revealed that the mosaic structure that harbored the most VF genes among the 16 main structures of S. aureus lytic phages was the clade IIIa (hlg, pvl, dut, and virE), clade IIIb (dut, pvl, scn, and sak) and clade IIIc (hlb, sak, and dut) phages (Zhou et al., 2022). However, higher prevalence of the VF genes was observed in the genomes of CP1 (dut, hlb, lukG, lukH, and hld), CP3 (virE,dut,lukH,lukG,hlb,scn and sak), CP4 (lukH,lukG,hlb,scn,sak,and dut), and IP3 (splA, splB, splC, lukD, lukE, seg, seln, yent1, yent2, sei, selm and selo). Therefore, the S. aureus prophages probably contributes to the transmission of VF genes in their hosts. The major difference of the major mosaic structures between lytic phages, and the intact and incomplete prophages was the functional modules as well as genes. The 16 major mosaic structures of lytic phages normally harbored four functional modules and other genes associated with the VF and lysogeny (Zhou et al., 2022). However, the intact and incomplete prophages were lack of functional genes associated with typical phage modules including phage morphogenesis, host cell lysis, DNA metabolism, and DNA packaging. These results were inconsistent with those of previous studies, suggesting that large-scale deletions occurred in the accessory genes Comparative structural analyses of lytic phage-3_AJ_2017 with intact prophage ST20130943_p1 and incomplete prophage SA3_LAU_ip3. Venn diagrams showing the number of orthologous groups among 579 intact prophage sequences, 1,389 incomplete prophages, and 188 lytic phages.
Frontiers in Microbiology 10 frontiersin.org of prophages such as integrase and cargo genes, rather than conserved genes involved in lytic gene repression, replication, expression of capsid proteins, and packaging (Ramisetty and Sudhakari, 2019). S. aureus isolates harboring large prophages and related genes impose an extra metabolic burden but may promote bacterial adaptation and habitat colonization (Fillol-Salom et al., 2019;Ramisetty and Sudhakari, 2019). Therefore, we hypothesized that the maintenance or loss of functional genes is key to balancing the costs and benefits of large prophages, thus expediting bacterial genome evolution. Besides, phage-induced lysis involves the destruction of the inner membrane, peptidoglycan layer, and outer membrane of the host cell through the phage encoded holin and lysin (Young, 2014). Double-stranded DNA bacteriophages package their genome at high pressure inside a procapsid through the DNA packaging proteins (Cuervo et al., 2019). These results indicated the loss of functional modules in prophages is the possible mechanism to ground the phages and, thus protect the S. aureus isolates from the lysis cycle of phages, which is consistent with a previous study (Bernheim and Sorek, 2020). AMR in S. aureus is a serious growing problem associated with enormous human and medical costs, and resistance transmission is promoted by mobile genetic elements (Allen et al., 2010). Furthermore, S. aureus can survive robust activation of the host's innate immune system as a result of the expression of a wide array of VFs, disabling the acquired immune response (Goldmann and Medina, 2018). Compared to the only 1 AMR gene (bla) and 5 VF genes observed in the genomes of 188 lytic phages (Zhou et al., 2022), this study revealed various AMR and VF genes identified in the genomes of S. aureus prophages. These results indicates that prophages act as a genetic reservoir of AMR and VF for S. aureus and that transduction is the major means of HGT for AMR and VF, consistent with the results of a recent study (Kondo et al., 2021). Over the past 75 years, beta-lactams have been the most important antibiotics used in treating S. aureus infections (Vestergaard et al., 2019). A previous study surveyed global antibiotic consumption from 2000 to 2010 and suggested that antimicrobial drugs, such as aminoglycosides, fluoroquinolones, trimethoprim, and tetracycline, were the most abundantly consumed antibiotics globally (Van Boeckel et al., 2014). S. aureus strains that secrete exotoxins and exoenzymes can disrupt endothelial and epithelial barriers via cell lysis and junction protein cleavage (Tam and Torres, 2019). Adhesins allow S. aureus to attach to host cells and the extracellular matrix during the early colonization phase (Negron et al., 2022). Nutritional/metabolic factors are critical for the fitness of S. aureus and its adaptation to nutritionally diverse environments (Balasubramanian et al., 2017). The formation of S. aureus biofilms in vivo renders it highly resistant to chemotherapeutics, leading to chronic diseases (Paharik and Horswill, 2016). Hence, intact and incomplete prophages of S. aureus contribute to adaptation to the host environment and the severity of S. aureus infections. Phylogeny of 188 lytic phages, 148 representative intact prophages and 522 representative incomplete prophages analyzed using Erwinia phage phiEa2809 as the outgroup based on 439,973 single-nucleotide polymorphisms. The inner ring is colored according to type of the sequence and the outer ring according to the geographic region.
Frontiers in Microbiology 11 frontiersin.org In addition, our analysis revealed that the prevalence of several AMR and VF genes in the intact prophages were significantly higher than those in incomplete prophages and lytic phages. These results suggested the easily spreading of AMR and virulence via prophage transduction, which was consistent with previous studies (Moon et al., 2015(Moon et al., , 2016. Besides, the numbers of some AMR and VF genes in lysogeny genes (rec, tnp, and int) harboring phages were significantly higher than those in phages without lysogeny genes, which is inconsistent with a previous study revealing that prophages containing VF-encoding genes are not likely to possess recombination-related genes near to VF genes (Kondo et al., 2021). These results indicated that the AMR genes and VF in the prophages may be acquired by various mechanism as well as phage transduction.
Phage 3-AJ-2017, isolated in Colombia, belongs to the Siphoviridae family. The host cell lysis module, and phage morphogenesis and DNA packaging module of phage 3-AJ-2017 shared high nucleotide sequence identity with an intact prophage in S. aureus strain ST20130943 isolated from humans in Brazil and an incomplete prophage in S. aureus strain SA3_LAU isolated from humans in Lebanon. However, the DNA metabolism module of phage-3-AJ-2017 shared little nucleotide sequence identity with ST20130943_p1. Consistently, the host cell lysis and phage morphogenesis module of phage 23MRA were similar to those of an intact prophage in S. aureus strain UTSW_MRSA_55 and an incomplete prophage in S. aureus strain MRSA_FKTN isolated from humans in the United States. The DNA metabolism module of phage 23MRA shared high nucleotide sequence identity with UTSW_ MRSA_55_p3. However, the genome of UTSW_MRSA_55_p3 harbored other genes associated with DNA metabolism. These results revealed the identical functional modules between the lytic phages and prophages, which is likely to result in the easily homologous recombination events between the lytic phages and prophages (Dion et al., 2020). A previous study indicated that extensive mosaicism with genes organized into functional modules that are frequently exchanged between phages because they coexist in a common host (Deghorain and Van Melderen, 2012). Therefore, exchange, acquisition, and loss of functional modules resulting from homologous recombination may augment the genetic diversity of the lytic phages, prophages and their hosts. However, a limitation of this study can be attributed to the numbers of S. aureus lytic phage genomes analyzed, therefore a more accurate determination of the genetic relationship between the prophage and lytic phage is required.
Consistently, ortholog and phylogenetic analyses indicated that prophages and lytic phages shared a common core gene pool and therefore could easily exchange, acquire, and lose genetic material in the host genomes. In the phylogenetic tree, the prophages were exclusively clustered with the S. aureus lytic Siphoviridae phages, indicating that the Siphoviridae lytic phage and prophages shared a common ancestor. These results suggest that a broad recombination events probably occurred between prophages and S. aureus Siphoviridae phages, consistent with the previous study (Deghorain and Van Melderen, 2012). The phylogenetic tree obtained via wholegenome phylogenetic analysis revealed that the S. aureus prophages and lytic Siphoviridae phages from different regions belonged to the same lineage, indicating that the recombination events between S. aureus prophages and lytic Siphoviridae phages occurred globally. However, future research should elucidate the taxonomic classification of these prophages and their evolutionary relationship with Siphoviridae phages based on their genetic features.
The worldwide genome interaction network revealed the majority of shared sequences in S. aureus prophages, representing the complex recombination events may easily occur in these prophages (Belcaid et al., 2010). Recombination events, called HGT, are major mediators of phage evolution and render the understanding of evolutionary relationships difficult (Dion et al., 2020). Therefore, frequent recombination events within S. aureus prophages resulting from homologous recombination promoted the genetic diversity of S. aureus isolates as well as prophages. Consistent with these results, region A in the genome of 09_01244_p4 comprised 43,514 bp and shared 99.0% nucleotide sequence identity with region B in the S. aureus prophage BSAR202_p1 (Supplementary Figure S13). Moreover, region B in the MAL9_ip4 genome comprised 29,477 bp and shared 98.0% nucleotide sequence identity with region D in the prophage MRGR3_ip9 (Supplementary Figure S14). Incomplete prophages that were considered genetic junks confer advantageous phenotypes to bacterial hosts with respect to virulence, stress resistance, or mutation rate (Ramisetty and Sudhakari, 2019). The present study illustrates that incomplete prophages can also be reservoirs of genetic material for lytic phages. These results identified that the constant recombination between S. aureus prophages in host chromosomes and lytic phages tightly links their evolution and indirectly accelerates bacterial evolution. However, future studies should elucidate the exact mechanism underlying Genetic relationship network with shared sequences as coupling links. S. aureus intact prophage, incomplete prophages, and lytic phages isolated from different geographic regions are represented as nodes. The area of each node directly correlates with the average number of shared sequences connected to a species. The number of shared sequences between nodes is represented by colored edges.
Frontiers in Microbiology 12 frontiersin.org recombination events among S. aureus intact prophages, incomplete prophages and lytic phages. The CRISPR-Cas system is an important constraint for HGT, and ensuring the maintenance or loss of prophages is key to the environmental adaption of this pathogen (Wheatley and MacLean, 2021). The non-specific DNase activity of the staphylococcal type III-A CRISPR-Cas system contributes to mutation rates in bacteria and AMR in S. aureus and S. epidermidis (Mo et al., 2021). Therefore, our study suggested that type III-A CRISPR-Cas proteins may mediate the genetic diversity of S. aureus prophages as well as their hosts. Considering the low detection rate of CRISPR-Cas genes in S. aureus genomes, recombination within S. aureus prophages via HGT was responsible for the genetic diversity and coevolution in S. aureus prophages, lytic phages, and their hosts.

Conclusion
In summary, our analysis suggests that intact and incomplete S. aureus prophages exhibited distinct mosaic structures, and the major difference between lytic phages, and intact and incomplete prophages was the loss of functional modules and genes. Compared to the lytic phages, both the S. aureus intact and incomplete prophages harbored multiple AMR and VF genes and contributed to the environmental adaption of its hosts and infection severity. Hence, the maintenance or loss of functional modules in intact and incomplete prophages is key to balance the costs and benefits of large prophages harboring various AMR and VF genes in the bacterial host. Besides, exchange, acquisition, and loss of functional modules between lytic phages and prophages resulting from homologous recombination contribute to the genetic diversity of S. aureus prophages. Constant recombination within prophages globally were responsible for the genetic diversity of S. aureus lytic phages and their host. Through these evolutionary strategies, the S. aureus prophage accelerates the evolution itself and the coevolution of lytic phages and their bacterial hosts.

Data availability statement
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary material.