Analysis host-recognition mechanism of staphylococcal kayvirus ɸSA039 reveals a novel strategy that protects Staphylococcus aureus against infection by Staphylococcus pseudintermedius Siphoviridae phages

Following the emergence of antibiotic-resistant bacteria such as methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-resistant Staphylococcus pseudintermedius (MRSP), phage therapy has attracted significant attention as an alternative to antibiotic treatment. Bacteriophages belonging to kayvirus (previously known as Twort-like phages) have broad host range and are strictly lytic in Staphylococcus spp. Previous work revealed that kayvirus ɸSA039 has a host-recognition mechanism distinct from those of other known kayviruses: most of kayviruses use the backbone of wall teichoic acid (WTA) as their receptor; by contrast, ɸSA039 uses the β-N-acetylglucosamine (β-GlcNAc) residue in WTA. In this study, we found that ɸSA039 could switch its receptor to be able to infect S. aureus lacking the β-GlcNAc residue by acquiring a spontaneous mutation in open reading frame (ORF) 100 and ORF102. Moreover, ɸSA039 could infect S. pseudintermedius, which has a different WTA structure than S. aureus. By comparison, with newly isolated S. pseudintermedius–specific phage (SP phages), we determined that glycosylation in WTA of S. pseudintermedius is essential for adsorption of SP phages, but not ɸSA039. Finally, we describe a novel strategy of S. aureus which protects the bacteria from infection of SP phages. Notably, glycosylation of ribitol phosphate (RboP) WTA by TarM or/and TarS prevents infection of S. aureus by SP phages. These findings could help to establish a new strategy for the treatment of S. aureus and S. pseudintermedius infection, as well as provide valuable insights into the biology of phage–host interactions.


Introduction
analyzed by Sanger sequencing. Phage plaques were touched with a toothpick and mixed into the PCR mixtures.

156
Generation and characterization of chimeric phage 157 Chimeric phage was generated by homologous recombination using plasmid pNL9164 158 (Sigma-Aldrich, MO, USA) in S. aureus SA003. Mutated DNA fragments of ORF100 and 159 ORF102 were amplified from spontaneous mutant phages by PCR using KOD-plus Neo enzyme 160 (Toyobo, Shiga, Japan). PCR fragments and the plasmid were digested with appropriate 161 restriction enzymes. The DNA fragment was inserted into plasmid pNL9164 using T4 DNA 162 Ligase (New England BioLabs, Ipswich, MA, USA). The constructed plasmid was cloned into 163 Escherichia coli JM109 competent cells (TAKARA, Shiga, Japan) and pre-introduced into 164 restriction-deficient S. aureus RN4220 before being transformed into S. aureus SA003. 165 To perform homologous recombination, transformant SA003 harboring plasmid with 166 homologous region was infected with phage (MOI = 1). The mixture was incubated at 37°C 167 overnight. Recombinant phages were isolated from the supernatant fraction of the mixture after 168 centrifugation (8000 × g, 3 minutes). SA003∆TarS was used to screen recombinant phages from 169 homologous recombination of ORF100, and RN4220 was used to screen recombinant phages 170 from homologous recombination of ORF102. Mutated ORF100 from spontaneous mutant phage 171 ɸM1 was introduced into wild-type ɸSA039, yielding chimeric phage ɸrM1/r-100. Mutated 172 ORF102 of ɸM1 was introduced into ɸrM1/r-100, yielding chimeric phage ɸrM1/r-100&102.

173
Deletion of oatA gene in RN4220∆tarM∆tarS 174 The gene was deleted by pCasSA plasmid with clustered regularly interspaced short 175 palindromic repeats (CRISPR)-Cas9 system (Chen et al. 2017). Plasmid construction was performed as previously described (Chen et al. 2017;Azam et al. 2018). Spacers were manually selected by searching the protospacer adjacent motif (PAM) region. Two oligos were designed as  receptor (Azam and Tanji 2019b), we sought to determine whether our SP phages also require 231 WTA. To this end, we generated WTA-free Staphylococcus by inhibiting WTA synthesis in the cell using tunicamycin (Zhu et al. 2018). ɸSA039, ɸSA012, and three SP phages (ɸDP001, 233 ɸSP120, and ɸSP197) had no ability to infect WTA-free S. pseudintermedius and S. epidermidis 234 (data not shown). The phages could not form a plaque and failed to adsorb onto WTA-free 235 isolates. By contrast, ɸSP276 did not completely lose its infectivity toward WTA-free isolates, 236 indicating that phage can use another component as a receptor. These findings indicated that all 237 phages in this study utilize WTA in S. pseudintermedius and S. epidermidis as their receptor, but 238 that the WTA of these bacteria is likely distinct from that of S. aureus.

239
Accumulation of point mutations in ORF100 enables ɸSA039 to infect TarS-null S. aureus 240 ɸSA039 requires β-GlcNAc glycosylation of WTA by TarS (Azam et al. 2018). In this 241 study, we found that ɸSA039 could generate spontaneous mutants capable of infecting TarS-null 242 S. aureus. Mutants of ɸSA039 that could infect TarS-null S. aureus were obtained from the fifth 243 batch of two co-cultures. One spontaneous mutant phage (ɸM1and ɸM2) was purified from each 244 of co-culture and further characterized. In SA003∆tarS, adsorption of ɸM1and ɸM2 was around 245 8-fold than that of wild-type ɸSA039 (Fig 1a). ɸM1and ɸM2 exhibited similar adsorption toward 246 SA003∆tarS. Because the phage tail fiber and baseplate region are thought to be involved in 247 phages' adsorption specificity, we amplified the genomic region encoding the tail and baseplate 248 proteins (ORF94 until ORF102) using primers described in Supplemental Table S1. Spontaneous

258
Because all ɸSA039 mutants had spontaneous base changes in ORF100, we speculated 259 that these mutations enabled the phages to infect SA003∆tarS. We hypothesized that by 260 introducing the point mutation in ORF100 into wild type ɸSA039, we should be able to construct 261 chimeric ɸSA039 capable of infecting SA003∆tarS. Hence, we introduced ORF100 of ɸM1 into 262 wild-type ɸSA039 by homologous recombination, yielding the chimeric phage ɸrM1/r-100.

263
Spot tests revealed that ɸrM1/r-100 could infect SA003∆tarS (data not shown). We then 264 performed an adsorption assay to determine whether the ability of the recombinant phages to 265 infect SA003∆tarS was due to an effect of adsorption as a result of replacement of ORF100.

285
As with ɸM2, the chimeric phage ɸrM1/r-100 was unable to adsorb onto RN4220 286 (6.06%), but was infectious toward RN4220∆TarM (14.60% adsorbed phage). Based on this 287 observation, we hypothesized that the absence of a point mutation in ORF102 in ɸM2 and 288 ɸrM1/r-100 may make these phages unable to infect RN4220. To test this idea, we introduced a 289 point mutation in ORF102 into ɸrM1/r-100 by homologous recombination, yielding the chimeric 290 phage ɸrM1/r-100&102. Indeed, adsorption of the chimeric phage ɸrM1/r-100&102 improved 291 three times higher than that of ɸrM1/r-100 ( fig 1b).

292
Whole-genome sequencing of SP phages 293 Because ɸSP120, ɸSP197, and ɸSP276 had a broad host range in S. pseudintermedius, 294 we further investigated their potential as antimicrobial agents. To this end, we first sequenced the 295 entire genomes of all three phages. In doing so, our primary goal was to investigate the presence 296 of toxin, virulence, and antibiotic-resistance genes that could make them inappropriate for phage 297 therapy.

298
The whole-genome sequencing analysis revealed that ɸSP120, ɸSP197, and ɸSP276 299 belong to Siphoviridae, with genome sizes of 40530, 41149, and 40711 bp, respectively. The GC 300 content of ɸSP120, ɸSP197, and ɸSP276 was similar to that of S. pseudintermedius, but 301 somewhat higher that of S. aureus (36%, 35%, and 36%, respectively). The genomes of the 302 phages had a low degree of identity relative to one another. Their genomes were organized into 303 six functional modules; lysogeny, DNA replication, packaging, head, tail, and lysis. The identical. Toxin, virulence, and antibiotic-resistance genes were absent from all three genomes, 307 according to the PHASTER server.

346
Thus, SP phages can utilize non-glycosylated WTA as their recognition site on S. aureus cells.  Reduced adsorption of three SP phages was observed in phage-resistant SP015 R1 (Fig   361   3). Using primers targeting the WTA gene cluster in SP015, we identified a point mutation in 362 tagN of R1 that caused a premature stop codon at amino acid 629 (Fig 3a). Complementation of 363 tagN using a wild-type allele restored adsorption of SP phages around 50%. Notable, mutation in R1 improved adsorption of ɸSA039 (56.20%), whereas complementation of tagN in R1 365 decreased adsorption of ɸSA039 (31.70%). In SP015 and its mutant derivatives, adsorption did 366 not differ significantly between mutant ɸSA039 (ɸM1, ɸM2, ɸrM1/r-100, and ɸrM1/r-100&102) 367 and wild-type ɸSA039 (Supplemental Figure S2). to adsorb onto RN4220. Phage ɸM1, which has mutations in ORF100 and ORF102, was able to 391 bind WTA of RN4220. Our hypothesis is consistent with the ability of a chimeric phage 392 harboring point mutations in ORF100 and ORF102 (ɸrM1/r-100&102) to infect RN4220.

393
A similar phenomenon has been documented in phages infecting Escherichia coli: 394 mutation in gp38, which encodes the tail protein of coliphage PP01, enables PP01 to infect E.

410
The same study also showed that Siphoviridae phage ɸ11 and ɸ80α (S. aureus-specific phages) 411 can infect S. aureus but not Staphylococcus species with GroP WTA. Therefore, the host range 412 of phages belong to this group likely depends on the type of WTA.

413
In this study, we found that ɸSA039 could infect different Staphylococcus species with  Phage host range is an important criterion when considering a candidate phage for 429 therapeutic application. According to our results, staphylococcal Myoviridae, especially those of kayvirus group, are the best candidates in terms of host range; however, to establish efficient 431 treatment, it will be necessary to precisely determine the bacteria responsible for each infection.