A Transducing Bacteriophage Infecting Staphylococcus epidermidis Contributes to the Expansion of a Novel Siphovirus Genus and Implies the Genus Is Inappropriate for Phage Therapy

ABSTRACT The effort to discover novel phages infecting Staphylococcus epidermidis contributes to both the development of phage therapy and the expansion of genome-based phage phylogeny. Here, we report the genome of an S. epidermidis-infecting phage, Lacachita, and compare its genome with those of five other phages with high sequence identity. These phages represent a novel siphovirus genus, which was recently reported in the literature. The published member of this group was favorably evaluated as a phage therapeutic agent, but Lacachita is capable of transducing antibiotic resistance and conferring phage resistance to transduced cells. Members of this genus may be maintained within their host as extrachromosomal plasmid prophages, through stable lysogeny or pseudolysogeny. Therefore, we conclude that Lacachita may be temperate and members of this novel genus are not suitable for phage therapy. IMPORTANCE This project describes the discovery of a culturable bacteriophage infecting Staphylococcus epidermidis that is a member of a rapidly growing novel siphovirus genus. A member of this genus was recently characterized and proposed for phage therapy, as there are few phages currently available to treat S. epidermidis infections. Our data contradict this, as we show Lacachita is capable of moving DNA from one bacterium to another, and it may be capable of maintaining itself in a plasmid-like state in infected cells. These phages’ putative plasmid-like extrachromosomal state appears to be due to a simplified maintenance mechanism found in true plasmids of Staphylococcus and related hosts. We suggest Lacachita and other identified members of this novel genus are not suitable for phage therapy.

In recent years, there has been increasing research and effort to isolate S. epidermidis phages (7)(8)(9)(10)(11)(12). As part of this S. epidermidis phage prospecting, members of a novel genus were isolated in multiple parts of the world in 2021. Fanaei Pirlar et al. (13) isolated, characterized, and sequenced a double-stranded DNA (dsDNA) siphovirus, CUB-EPI_14 (GenBank accession no. ON325435.2), that is ;43 kb and has a narrow host range within S. epidermidis. Despite the genome of CUB-EPI_14 being labeled as likely temperate by PhageAI, which is normally disqualifying for a potential phage for therapy (14), CUB-EPI_14 lacks an integrase, and the authors suggest it could be a potential candidate for phage therapy (13).
Simultaneous with the work of Fanaei Pirlar et al., we also isolated a related phage from this novel genus by culturing wastewater on S. epidermidis, as did a third group, who has deposited their phage's genome in GenBank without an accompanying paper (GenBank accession no. ON550478.1). We sequenced the genome of our representative (Lacachita) and conducted additional host range and transduction assays. We found three additional representatives of this novel genus in GenBank (including one from an S. epidermidis shotgun sequencing project) and analyzed these with the three cultured phage genomes. We identified in all genomes a common phage resistance gene and a partitioning protein that is associated with being maintained in a plasmid state. Combined with our observation that Lacachita can transduce antibiotic resistance genes, we propose that members of this novel genus are likely temperate and therefore inappropriate for phage therapy.

RESULTS
Isolation of Lacachita. Phages capable of forming plaques on S. epidermidis 1457 were successfully isolated from samples of wastewater influent from a treatment plant in the mid-Atlantic United States. Concentrated samples of the unenriched wastewater did produce plaques on S. epidermidis 1457, but the enriched wastewater produced orders of magnitude more plaques. Plating the host alone (without wastewater) did not produce any plaques. During the isolation procedure, a total of 11 plaques were harvested for potential further work. Of these isolated plaques, two were chosen for sequencing. Upon sequencing and assembly of the genomes, it was discovered that the two were 100% identical, and so only one (Lacachita) was further characterized.
Lacachita genome and annotation. Lacachita has a 46,473-bp dsDNA genome that is likely circular (GenBank accession no. OP142323). It contains 72 predicted open reading frames (ORFs), 19 putative promoters, 1 putative noncoding RNA (which encodes a group I catalytic intron), 19 putative rho-independent terminators, and no predicted tRNAs. Several similar phage genomes were identified by BLAST (.95% identity over $93% of the genome), and the Lacachita genome was linearized and oriented to mimic the genomes of its close relatives, which were also isolated on S. epidermidis (accession no. ON550478.1 and ON325435.2) ( Table 1).
Lacachita's putative protein products contain an expected assortment of phage proteins and some hypothetical proteins (Fig. 1). Nine structural proteins were identified, which were similar, by blastp, to those of siphoviruses with long, noncontractile tails. Lacachita has both a holin and an endolysin, and 14 proteins involved in DNA replication and metabolism were identified. Two ORFs are associated with a plasmid prophage lifestyle: a ParB-like protein and a potential phage resistance protein. The remaining 43 ORFs in the Lacachita genome are either hypothetical (34 ORFs) or are identified only with a protein family or as including a known domain (9 ORFs).
Lacachita is part of novel genus, along with other proposed members. Four other phage genomes were identified by blastn to be relatives of Lacachita, and another was identified by blastp (Table 1). Of these five relatives, only CUB-EPI_14 (accession no. ON325435.2) has yet been thoroughly described (13). The authors of that article noted that CUB-EPI_14 appears to represent a novel genus and identified two other potential members of the genus via calculation of intergenomic distance-Caudovirales phage clone 9S_3 (accession no. MF417888.1) and third party annotation (TPA) Myoviridae sp. isolate ct5pN1 (accession no. BK030923.1). These two phages were also independently identified as relatives of Lacachita during our searches, and so our analysis complements and bolsters the evidence for these phages representing a new genus. The genome of another cultured phage, Sazerac (accession no. ON550478.1), was deposited in GenBank after the manuscript about CUB-EPI_14 was submitted for publication, and we propose that Sazerac is also part of this novel genus. The final relative, Sep_B35_CVC_2019 (accession no. NZ_CAJUVG010000006), was identified due to its consistent protein sequence identity to Lacachita protein products. Although Sep_B35 is catalogued in NCBI as a contig of a S. epidermidis whole-genome shotgun sequence, we argue that this contig represents a full phage genome from an infected S. epidermidis strain. Furthermore, since the sample of S. epidermidis was sequenced as a bacterial shotgun sequencing project, not labeled as a study in phage infection, we suggest that the Sep_B35 genome represents a prophage that was being maintained within the S. epidermidis isolate at the time it was sequenced.
Taxonomic assignment of Lacachita and its relatives confirmed that these phages represent a novel genus within the family Siphoviridae (Fig. 2). The six genomes form a monophyletic clade, clustered near the Sextaecvirus infecting other staphylococci, among other siphoviruses. There was strong support for this group forming a novel genus (symmetrical Theil's uncertainty correlation of 0.863). Genomic maps of the six members of the putative genus reveal some observable regions of synteny (Fig. 3).
Following the example of Fanaei Pirlar et al. (13), we ran the six genomes through the PhageAI lifestyle classifier. We found that Lacachita and all members of this putative novel genus were predicted to be temperate with at least 99.95% confidence ( Table 1). Three of the five proposed relatives of Lacachita are uncultured, putative phages. Based on the length and query coverage of these genomes compared to those of cultured isolates, it is likely that these represent essentially complete genomes.
Host range results. Of the 7 S. epidermidis strains tested in this project, 5 mL of Lacachita lysate was found to be capable of lysing S. epidermidis strains 1457, NRS101 (RP62a), B72-22, 158-22, and B138-22. Of the other Staphylococcus species strains tested, 5 mL of Lacachita   (Table 2). Lacachita is capable of transduction and lysogeny. Transduction assays were conducted three separate times, and in 5/6 cases, Lacachita was capable of transducing plasmid-encoded erythromycin resistance to erythromycin-sensitive S. epidermidis 1457.
Spotting of serial dilutions of Lacachita lysate on bacterial lawns of transduced S. epidermidis 1457 revealed that these transduced cells were 100-fold less susceptible to lysis than S. epidermidis 1457 that had not undergone transduction. While the highest dilution of the lysate able to produce lysis on S. epidermidis 1457 was 10 26 , the highest dilution of the lysate able to lyse transduced S. epidermidis 1457 was 10 24 . We were unable to identify a putative integrase gene in the genomes of Lacachita or its close relatives. ORF analysis of Lacachita and its close relatives revealed the presence of putative parB and common phage resistance gene ( Fig. 1) in all six genomes (Table 3), which is partial evidence that these phages are temperate and suggests the prophages are maintained extrachromosomally. We elected to classify Lacachita ORF70 (UVD33307.1) as "ParB protein" because the highest blastp result (99.8% identity) was annotated as a ParB protein (MAG TPA ParB protein [Myoviridae sp.] accession no. DAI53229.1). We conducted a phylogenetic analysis of the ParB-like proteins from this putative genus and related sequences (identified by blastp) and annotated ParB proteins from Staphylococcus genomes (Fig. 4). The sequences from Lacachita's putative genus form a robust clade (99.7% bootstrap support),    (15,16).

DISCUSSION
Lacachita is part of a novel genus. Our GRAViTy analysis strongly suggests that Lacachita and its relatives belong to a siphovirus genus recently described by Fanaei Pirlar et al. (13), who characterized CUB-EPI_14 as belonging to a novel genus along with 9S_3 and ct5pN1. We have also expanded this putative genus by two more phages: Sazerac and the Sep_B35_CVC_2019 putative prophage. Based on the high genetic identity between CUB-EPI_14 and the other genomes, we can assume all members of this genus have long, noncontractile tails (13). Members of this genus have already been found on three continents, and we anticipate further isolates will be characterized in the upcoming years. Additional hosts may be identified for these phages as well, as we have expanded the potential hosts for members of this genus to include other Staphylococcus species than S. epidermidis (relative to Fanaei Pirlar et al. [13]).
Lacachita and its relatives are not suitable for phage therapy. However, unlike Fanaei Pirlar et al. (13), we do not think members of this genus should be used for phage therapy. There are several characteristics that are typically screened for when assessing whether a phage could be used for phage therapy, including host range, phage virulence, transduction potential, stability against environmental pressures, and the presence of toxin genes (17). Bioinformatic analysis suggests members of this genus are temperate, which is contraindicated for phage therapy. Temperate phages capable of transduction have the potential to increase the pathogenicity of lysogenized bacteria by carrying virulence factors between hosts (18). This is observed in temperate phages of S. epidermidis that can mobilize antibiotic resistance plasmids (19). In the interest of self-preservation, prophages also typically cause lysogenized bacteria to become immune to lytic infection by other phages that share similar repression systems (17). None of the members of this putative genus has an integrase gene, which is a key indicator of a temperate lifestyle because it allows stable integration of the phage genome into that of its host (14). Instead, the signal that PhageAI is picking up on in the phage genomes may be the presence of the parB gene (typically found as a parA-parB pair, implying a ParAB-parS system for chromosome segregation [20]) and the putative phage resistance gene, which is not required in phages that only rapidly lyse their host cells (21). Some temperate phages are known to be maintained in their bacterial host cells as extrachromosomal circular plasmids and maintain their presence in their hosts with similar mechanisms to plasmids (21)(22)(23). Some phage prospecting projects anticipate that these plasmid-like prophages may be isolated (24) but require that genomes have both parA and parB partitioning protein ORFs identified to be considered temperate (https://seaphages.org/forums/topic/4367/). To our knowledge, there are no characterized phages, temperate or otherwise, that contain only a gene for ParB, which binds to specific DNA sequences (parS sequences, which vary among bacteria [25]). A blastp search with the ParB of Lacachita found only members of its genus and 50% coverage to other phage proteins (typically the N terminus of ParB [data not shown]). Nonetheless, we did not find a ParA homolog, which is an ATPase that assists with localization of ParB (26,27) in these six phage genomes. However, members of the genera Staphylococcus and Streptococcus are known to use only a ParB-parS system to ensure chromosome segregation without ParA (25); Staphylococcus aureus plasmid SK1 has a ParB-parS system as well (28). Therefore, extrachromosomally maintained prophages of these hosts may also not need ParA in order to stably vertically transmit to daughter cells.
There is no perfect test for whether a phage is temperate and capable of creating a lysogen. We observed intermittent turbidity of plaques on S. epidermidis 1457 and in the spot Novel Transducing S. epidermidis Phage Genus mSphere plating experiments on other Staphylococcus strains, which is often considered an important phenotype of temperate phages (21). However, turbidity can be affected by many factors (21). Importantly, we have repeated evidence of Lacachita's capacity to transduce erythromycin resistance to a previously susceptible strain of S. epidermidis. Transduction is a phenomenon typically associated with temperate phage, although it can be due to "pseudolysogeny," or the formation of a carrier state (29)(30)(31). Our experimental evidence of Lacachita's ability to confer phage resistance to transduced cells is more indicative of lysogeny. Many temperate phages do not confer complete resistance to infection with that phage, and a 100-fold reduction in efficiency of plating is consistent with the behavior of other temperate phages (32,33). Regardless of the durability of lysogeny with Lacachita, any transduction ability is empirical evidence that Lacachita and its close relatives should not be used in phage therapy. Phage therapy remains a promising avenue of research for treating S. epidermidis infections, but members of this genus are not appropriate therapeutic agents. Genetic engineering is one way to modify temperate phages to be more appropriate therapeutic candidates (34), but that approach is controversial (35). Additional isolation of S. epidermidis phages is needed to find obligately lytic phage.

MATERIALS AND METHODS
Wastewater sample screening. Aliquots of wastewater influent from a treatment plant in the mid-Atlantic United States were obtained in March 2021 and were screened for lytic phages effective against S. epidermidis 1457 (36). Five milliliters of the wastewater samples was combined with 0.15 g powdered tryptic soy broth (TSB) medium, 25 mL of 1 M CaCl 2 , and 50 mL bacterial broth culture (S. epidermidis 1457) and then incubated overnight at 37°C. A second sample of wastewater underwent the same procedure without the addition of host bacteria. After incubation, the mixtures were centrifuged at 3,220 Â g for 15 min and the supernatant was passed through a 0.22-mm-pore filter. One hundred microliters of the filtrate was cultured with 100 mL of 10 21 bacterial dilution of an overnight culture using the pour plate method (in 3 mL of 0.3% molten tryptic soy agar [TSA] combined with 25 mL of 1 M CaCl 2 , vortexed, and poured onto TSB-1% agar plates). The plates were then incubated overnight at 37°C and examined for the presence of plaques.
Phage isolation. In order to purify phages identified during the screening process, isolated plaques were picked up using a sterile glass pipette tip, and the agar was deposited into a culture tube containing 2 mL TSB, 50 mL bacterial broth culture, and 25 mL of 1 M CaCl 2 and incubated overnight at 37°C. This liquid culture was then centrifuged at 3,220 Â g for 15 min, and the supernatant was passed through a 0.22-mmpore filter. The filtrate was diluted, and 100 mL of this diluted filtered supernatant was combined with 100 mL 10 21 bacterial dilution, 25 mL of 1 M CaCl 2 , and 3 mL of molten TSA, vortexed, and poured onto TSA plates. The plates were then incubated overnight at 37°C and examined for plaques. This subculturing procedure was performed a total of three times to yield a purified, enriched phage stock.
DNA isolation and sequencing. The DNA genome of one isolated phage was extracted using a Qiagen QIAamp MinElute virus spin kit. Paired-end Illumina sequencing was performed at MiGS (Microbial Genome Sequencing Center [now SeqCenter]). Reads were analyzed using CPT Galaxy Phage genome assembler v.2021.01 Workflow (37), which uses SPAdes Galaxy v.3.12.0 (38). This yielded three assembled contigs. These contigs were aligned manually using Aliview (39) to verify that they were identical (except for short regions of duplication due to the likely circular genomes having been assembled linearly) and to produce a complete genome without such duplications. The Lacachita genome was reoriented to mimic the linearization of relatives found using NCBI Standard Nucleotide BLAST.
Genome annotation. The Lacachita genome was annotated using Prokka (v.1.14.6; Galaxy) with the parameters Kingdom: Viruses (40). Predicted ORFs were annotated further using NCBI Standard Protein BLAST, and the sequences producing significant alignments were analyzed to determine functional gene annotations for Lacachita. When the BLAST search produced multiple identical hits, we chose the annotation that was most relevant to a phage lifestyle (e.g., the name given in another phage genome). Phylogenetic analysis of the Lacachita genome and other dsDNA phage genomes was performed with GRAViTy v.1.1.0 (Genome Relationships Applied to Virus Taxonomy; http://gravity.cvr.gla.ac.uk/) (41).
Further functional annotation was performed. Promoter sequences were predicted by inputting the Lacachita genome into the Genome2D Prokaryote Promoter Prediction tool (42). Rho-independent termination sites were predicted using the ARNold web tool (43). Noncoding RNAs were found using Rfam (44). TRNAscan-SE was used to search the Lacachita genome for transfer RNAs (45).
Comparison to related phage genomes. To identify close relatives of Lacachita, the assembled Lacachita genome was used to query NCBI Standard Nucleotide BLAST. Other phage relatives were identified by searching predicted Lacachita ORFs using NCBI Standard Protein BLAST and making note of organisms with consistent protein homology to Lacachita ORFs, whose genomes were then compared to Lacachita directly using NCBI Align Sequences Nucleotide BLAST. Lacachita and identified relatives were analyzed with GRAViTy v.1.1.0 to determine their taxonomy (41). GRAViTy results were visualized using FigTree v.1.4.4 (http:// tree.bio.ed.ac.uk/software/figtree/). The protein products of Lacachita predicted ORFs were analyzed to identify potential indicators of phage lifestyle, and the genomes of Lacachita and its relatives were also analyzed using the PhageAI lifestyle classifier algorithm (46).

Novel Transducing S. epidermidis Phage Genus mSphere
ParB protein maximum likelihood tree. Two sets of ParB-like protein sequences were collected for phylogenetic analysis: Lacachita ParB blastp hits and annotated ParB sequences from Staphylococcus genomes. Sequences were aligned by MUSCLE (47), and the alignment was checked by eye. The aligned sequences were used to build a maximum likelihood tree with PhyML (48) (B21-22), and S. aureus (LAC WT, SH1000, MW2, N315). With the exception of S. epidermidis strains 1457, ATCC 12228, and NRS101 (RP62a) and the S. aureus strains, the isolates are deidentified clinical isolates that were collected at the University of Nebraska Medical Center Clinical Microbiology Laboratory. Pour plates of each strain were prepared by combining 3 mL of 0.7% molten TSA, 25 mL of 1 M CaCl 2 , and 10 mL bacterial overnight culture, vortexing the mixture, and pouring it onto TSA-1% agar plates. Once the top agar solidified, 5 mL of a dilution series (1 to 10 29 ) of high-titer Lacachita lysate was spotted onto the surface. As a control, 5 mL of TSB was also spotted onto the plates. The plates were then incubated overnight at 37 ³C and examined for evidence of lysis. Experiments were conducted in triplicate.
Transduction. In order to determine whether Lacachita possesses transduction abilities, a plasmid transduction experiment was performed. A modified S. epidermidis strain (1457 saeR/pNF155) carrying a 9-kb plasmid that is marked with an erythromycin resistance gene served as a donor strain, and erythromycin-sensitive S. epidermidis 1457 served as a recipient strain (49). Phage-bacterial cocultures were prepared with 2 mL of S. epidermidis 1457 saeR/pNF155 overnight culture (grown in TSB with 10 mg/mL erythromycin), which was combined with 5 mL TSB, 100 mL 1 M CaCl 2 , and 100 mL Lacachita purified phage stock. These bacterial-phage cocultures were incubated overnight at 37°C. The following day, Lacachita phages were harvested from the cocultures by centrifugation (13,000 Â g for 3 min) and the resulting supernatant was then filtered using sterile 0.22-mm-pore filters to remove bacterial cells. This filtered supernatant was then combined with overnight cultures of S. epidermidis 1457 recipient strain: 500 mL S. epidermidis 1457 culture, 500 mL TSB, 100 mL 1 M CaCl 2 , and 100 mL of the harvested Lacachita donor phage preparation. These cocultures were incubated at 37°C for 1 h. Following incubation, 400 mL of 1 M sodium citrate was added to each, each tube was vortexed to mix, and each coculture was transferred to a microcentrifuge tube. Cells were pelleted by centrifugation at 13,000 Â g for 2 min, and the supernatant was discarded. The cells were resuspended in 1 mL TSB and centrifuged again at 13,000 Â g for 2 min. The cells were then resuspended in 200 mL TSB and plated on TSA plates containing 10 mg/mL erythromycin and 2 mM sodium citrate. As negative controls, erythromycin-sensitive S. epidermidis strain 1457 was plated on erythromycin-containing TSA plates and the Lacachita phage stock was spotted onto erythromycin-containing TSA plates. Inoculated plates were incubated overnight at 37°C and were then examined for the presence of bacterial growth. The transduction experiments were performed six times.
Assessment of phage resistance of lysogenized bacterial cells. To determine whether transduced bacteria were resistant to new lytic Lacachita infection, we spotted serial dilutions of Lacachita lysate onto lawns of transduced S. epidermidis 1457 cells. We obtained transduced cells using a modification of the transduction protocol described above (excluding sodium citrate from the agar plates), and representative resulting colonies were streaked onto erythromycin-containing TSA plates. Liquid cultures were inoculated with isolated colonies in 10 mL TSB containing 10 mg/mL erythromycin and were incubated overnight at 37°C. Pour plates of lysogenized cells were prepared by combining 3 mL of 0.7% molten TSA, 25 mL of 1 M CaCl 2 , 10 mL bacterial overnight culture, and 3 mL 10 mg/mL erythromycin, vortexing the mixture, and pouring it onto TSA plates. Once the top agar solidified, 5 mL of a dilution series (from 1 to 10 29 ) of high-titer Lacachita lysates was spotted onto the surface. As a control, 5 mL of TSB was also spotted onto the plates. For comparison, this was also performed on S. epidermidis 1457 that had not undergone transduction and potential lysogeny. The plates were then incubated overnight at 37°C and examined for evidence of lysis. Experiments were conducted in triplicate.

ACKNOWLEDGMENTS
This work was supported by NIAID 1R01AI139100-01, NSF 1750624 and USDA MRF project NE-1028 to J.M.B. and NSF 1453241 to S.D. TPA was supported by NIGMS NIH 1T32GM139804-01.
We thank the utility partner (who wished to remain anonymous) for providing the wastewater influent. We also thank two anonymous reviewers for constructive feedback, which improved the manuscript.