Proteomic Characterization of Bacteriophage Peptides from the Mastitis Producer Staphylococcus aureus by LC-ESI-MS/MS and the Bacteriophage Phylogenomic Analysis

The present work describes LC-ESI-MS/MS MS (liquid chromatography-electrospray ionization-tandem mass spectrometry) analyses of tryptic digestion peptides from phages that infect mastitis-causing Staphylococcus aureus isolated from dairy products. A total of 1933 nonredundant peptides belonging to 1282 proteins were identified and analyzed. Among them, 79 staphylococcal peptides from phages were confirmed. These peptides belong to proteins such as phage repressors, structural phage proteins, uncharacterized phage proteins and complement inhibitors. Moreover, eighteen of the phage origin peptides found were specific to S. aureus strains. These diagnostic peptides could be useful for the identification and characterization of S. aureus strains that cause mastitis. Furthermore, a study of bacteriophage phylogeny and the relationship among the identified phage peptides and the bacteria they infect was also performed. The results show the specific peptides that are present in closely related phages and the existing links between bacteriophage phylogeny and the respective Staphylococcus spp. infected.


Introduction
The vast majority of mastitis cases are due to an intramammary infection caused by a microorganism belonging to either the Staphylococcus or Streptococcus genus [1,2]. Staphylococcus aureus is considered one of the major foodborne pathogens that can cause serious food intoxication in humans due to the production of endotoxins; this pathogen remains a major issue in the dairy industry due to its persistence in cows, its pathogenicity, its contagiousness and its ease of colonization of the skin and mucosal epithelia [3][4][5].
It is well-known that S. aureus bacteriophages encode genes for staphylococcal virulence factors, such as Panton-Valentine leucocidin, staphylokinase, enterotoxins, chemotaxisinhibitory proteins or exfoliative toxins [6]. These phages are usually integrated into bacterial chromosomes as prophages, wherein they encode new properties in the host, or vice versa, as transcriptions may hardly be affected by gene disruptions [7]. Phage-encoded recombinases, rather than the host recombinase, RecA, are involved in bacterial genome excisions and integrations [8,9]. These integrations may occur at specific bacterial genome sites that are identical to those present in the DNA of the phage, or, as in the case of phage

Bacteria
In this study, a total of 20 different S. aureus strains obtained from different sources were analyzed (Table S1 in Supplemental Data 2). These strains were previously characterized by MALDI-TOF mass spectrometry [1] after being obtained from the Institute of Science of Food Production of the National Research Council of Italy (Italy) and from the Spanish Type Culture Collection (Spain). The majority of the strains are from food origins, except for strain U17, which is a human clinical strain. Strains ATCC (American Type Culture Collection) 9144 and ATCC 29213 are classified as S. aureus subsp. aureus, while strain ATCC 35845 is categorized as S. aureus subsp. anaerobius. In previous works, the species identification of S. aureus and the presence of enterotoxins were evaluated by multiplex polymerase chain reactions (multiplex PCRs) [3,34,35]. The strains were reactivated in a brain-heart infusion medium (BHI, Oxoid Ltd., Hampshire, UK) and incubated at 31 • C for 24 h. Bacterial cultures were then grown on plate count agar (PCA, Oxoid) at 31 • C for 24 h [1,3,36]. Tubes of broth were inoculated under aerobic conditions.

Protein Extraction and Peptide Sample Preparation
Protein extraction was prepared as described previously [37]. All analyses were performed in triplicate. Protein extracts were subjected to in-solution tryptic digestion [38].

Shotgun LC-MS/MS Analysis
Peptide digests were acidified with formic acid (FA), cleaned on a C18 MicroSpin™ column (The Nest Group, South-borough, MA, USA) and analyzed by LC-ESI-MS/MS using a Proxeon EASY-nLC II Nanoflow system (Thermo Fisher Scientific, San Jose, CA, USA) coupled to an LTQ-Orbitrap XL mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) [3]. Peptide separation (2 µg) was performed on a reverse-phase (RP) column (EASY-Spray column, 50 cm × 75 µm ID, PepMap C18, 2-µm particles, 100-Å pore size, Thermo Fisher Scientific, San Jose, CA, USA) with a 10-mm precolumn (Accucore XL C18, Thermo Fisher Scientific, San Jose, CA, USA) using a linear 120-min gradient from 5% to 35% solvent B (solvent A: 98% water, 2% ACN (Acetonitrile) and 0.1% FA and solvent B: 98% ACN, 2% water and 0.1% FA) at a flow rate of 300 nL/min. For ionization, a spray voltage of 1.95 kV and a capillary temperature of 230 • C were used. Peptides were analyzed in the positive mode from 400 to 1600 amu (1 µscan), which was followed by 10 data-dependent collision-induced dissociation (CID) MS/MS scans (1 µscan) using an isolation width of 3 amu and a normalized collision energy of 35%. Fragmented masses were set in dynamic exclusion for 30 s after the second fragmentation event, and unassigned charged ions were excluded from the MS/MS analysis.

LC-MS/MS Mass Spectrometry Data Processing
LC-ESI-MS/MS spectra were searched using SEQUEST-HT (Proteome Discoverer 2.4, Thermo Fisher Scientific, San Jose, CA, USA) against the S. aureus UniProt/TrEMBL database (208,158 protein sequence entries in July 2020). The following parameters were used: semi-tryptic cleavage with up to two missed cleavage sites and tolerance windows set at 10 ppm for the precursor ions and 0.06 Da for the MS/MS fragment ions. These additional identified semi-tryptic peptides increased the sequence coverage and confidence in protein assignments. The variable modifications that were allowed were as follows: (M*) methionine oxidation (+15.99 Da), (C*) carbamidomethylation of Cys (+57.02 Da) and acetylation of the N-terminus of the protein (+42.0106 Da). To validate the peptide assignments, the results were subjected to a statistical analysis with the Percolator algorithm [39]. The false discovery rate (FDR) was kept below 1%. The mass spectrometric data were deposited into the public database PRIDE (Proteomics Identification Database), with the dataset identifier PXD023530.

Selection of Potential Peptide Biomarkers
For each peptide identified by LC-ESI-MS/MS, we used the BLASTp program to determine the homologies and exclusiveness with protein sequences registered in the NCBI (National Center for Biotechnology Information) database [40]. For the BLASTp search, the Staphylococcus taxon was included and excluded with the aim of finding the peptides that belonged to the Staphylococcus phages, Staphylococcus spp. and only to S. aureus.

Phage Genome Comparison and Relatedness
Genomes of all studied Staphylococcus spp. phages were downloaded from the Gen-Bank database, analyzed and compared using the Web server VICTOR (Virus Classification and Tree Building Online Resource, http://ggdc.dsmz.de/victor.php, accessed on 27 November 2020) for the calculation of the intergenomic distances and the construction of the phylogenomic tree [41].

S. aureus Proteome Repository
Protein mixtures from each of the 20 different S. aureus strains (Table S1 in Supplemental Data 2) were digested with trypsin and analyzed by LC-ESI-MS/MS.
A total of 1933 nonredundant peptides corresponding to 1282 nonredundant annotated proteins were identified for all S. aureus strains (see the Excel dataset in Supplemental Data 1). Among them, 79 phage peptides were identified. These peptides belong to proteins such as phage repressors, structural phage proteins, uncharacterized phage proteins and complement inhibitors. Figure 1 shows a comparative representation of the different types of phage proteins identified in this study. These phage peptides were selected and analyzed using the BLASTp algorithm. For the BLASTp search, Staphylococcus was included and excluded with the aim of finding peptides belonging to Staphylococcus bacteriophages.  The obtained staphylococcal phage-specific peptides shared homology with the Staphylococcus phages and Staphylococcus spp. in the NCBI database. Among them, all shared homology with S. aureus; however, eighteen peptides were specific to S. aureus (IRLPYYDVK, LYVGVFNPEATK, SIINGKLDSQWTVPNEHK, M*NDSNQGLQANPQYTIHYLSQEITR, PCPALM*NKRNSIATIHR, SQDSNLTPELSTKAPK, ESINANTYINQNLEK, VAVLSTPLVTS-FESK, KDGEILFDAIDIYLRNK, MPVYKDGNTGKWYFSI, KTTSEALKEVLSDT, EPKPV-DATGADDPLKPDDRM*ITNFHANLVDQKVSY, MSHNALTTGIGIGAGAG, VQHPGK-LVNKVM*SGLNINFGGGANATAK, QM*MEGLSGVMDLAAVSGEDLGAVSDIVTDGLTA FGLKAKDSG, KSNVEAFSNAVK, GMVASMQMQVVQVNVLTM*ELAQQNAMLTQQLTELK and DIITVYC*PENGTATDEY). Figure S1 shows the MS/MS spectra for these S. aureusspecific peptide biomarkers. Table 1 summarizes the list of 79 specific staphylococcal bacteriophage peptides, bacterial peptides with putative phage origins and bacteria and phages with 100% homology with respect to the NCBI protein database.
All staphylococcal phage peptides with 100% homology were found to belong to the Siphoviridae family: 52 staphylococcal phages belong to the Phietavirus genus, 37 belong to the Biseptimavirus genus, 30 are Triavirus, two are phieta-like viruses and one is a SPbeta-like virus, and the others are nonclassified Siphoviridae viruses (Table S2 in Supplemental Data 2). Siphoviridae genomes are usually organized into functional modules, such as lysogeny, DNA replication, packaging, morphogenesis and lysis modules [6,42].

Phage Peptides Determined from the Analyzed S. aureus Strains
For strains S2 and S3, six and three phage peptides were determined, respectively. For strain S4, seventeen phage peptides were determined, and three phage peptides were determined for strain S5. For strains S6 and S7, three and one phage peptides were determined, respectively. Moreover, for strains S8 and S9, two phage peptides and seven phage peptides were determined. For strains S10 and S11, five and three phage peptides were determined, respectively. For strains S12 and S13, five phage peptides and six phage peptides were determined, respectively. For strains S14 and S15, four and two phage peptides were determined, respectively. For strain S16, three phage peptides were determined, and one phage peptide was determined for strain S17. For strains S18 and S19, one phage peptide each was determined. Finally, for strain S20, seven phage peptides were determined.
A large number of phage peptides from structural proteins were identified (Table 1). Peptides from proteins such as the major capsid protein, major tail protein, minor structural protein, phage head morphogenesis protein, tail tape measure protein and phage tail fiber protein were determined. Moreover, different phage peptides from the major capsid protein and tail protein were determined (Table 1). Identifying these phage peptides is reasonable, as the major capsid protein and major tail protein are the most abundant proteins in mature virions [6].
There are a large number of uncharacterized protein sequences in databases, and more than 20% of all protein domains are annotated as "domains of unknown function" (DUFs). Several uncharacterized phage proteins and DUFs from Staphylococcus bacteriophages were identified for the analyzed strains (Table 1) [43,44].
Different peptides from repressor-type Cro/CI were determined. For strains S11 and S20 (both potential enterotoxin C producers), the same phage peptides of repressor-type Cro/CI were identified (Table 1). CI and Cro are encoded in the lysogeny module of lambdoid bacteriophages, particularly λ bacteriophages. Together, CII and CIII (that are formed through the anti-terminator role of protein N) act as an inducer that favors the first expression of the cI gene from the appropriate promoter; if the CI repressor predominates, the phage remains in the lysogenic state, but if the Cro predominates, the phage transitions into the lytic cycle, helped by the late Q regulator. The xenobiotic XRE regulator is extended in bacteria and has similarity to the Croλ repressor, exhibiting a helix-turn-helix (HTH) conformation [45]. Peptides of the CI/Cro-repressor types are usually named XRE family proteins in the NCBI database for bacteria.
Three phage peptides of the complement inhibitor were identified (Table 1). Staphylococcal complement inhibitors are involved in the evasion of human phagocytosis by blocking C3 convertases, and a study reported that complement inhibitor genes were also found in staphylococcal phages [46]. Another autolysin was determined in the present results, an N-acetylmuramoyl-L-alanine amidase that plays a role in bacterial adherence to eukaryotic cells [19]. The phage protein NrdI, which is a type of ribonucleotide reductase (RNR), was also identified. Several peptides of transposases, integrases and terminases were identified along with a DNA primase phage associated protein and a DNA phage binding protein. Moreover, peptides of other proteins, such as GNAT family N-acetyltransferase, holin, peptidase, methylase, anti-repressor protein (Ant), phage-resistant protein, phage-encoded lipoprotein, phage infection protein, phage portal protein, toxin phage proteins associated with pathogenicity islands and a protein involved in fibrinogen-binding proteins, were identified. A PBSX family phage terminase peptide was determined, and this protein is involved in double-stranded DNA binding, DNA packaging and endonuclease and ATPase activities [47].
As shown in Table 1, the vast majority of phage-specific peptides are not specific to S. aureus and can be found in other species of Staphylococcus. As an exception, the same peptides, such as peptide LLHALPTGNDSGGDKLLPK from a major capsid protein, were also found in Streptococcus pneumoniae, and peptide AYINITGLGFAK from a major tail protein was also found in Pararheinheimera mesophila; whether these examples represent direct recombinations between bacteria belonging to different families or whether phagemediated recombination occurs remains to be elucidated. Furthermore, as mentioned before, eighteen identified peptides were very specific for S. aureus based on the NCBI database (see Figure S1).

Staphylococcus spp. Phage Genome Comparisons and Their Relatedness
A phylogenomic tree of Staphylococcus spp. phages from the NCBI database (accession numbers in Table S2 in Supplemental Data 2) with 100% similarity to those found in this study was built (Figure 2). The phages identified in this study were classified in the order Caudovirales and the family Siphoviridae. Many of these bacteriophages were classified into the genera Phietavirus, Biseptimavirus, Triavirus phieta-like virus, SPbeta-like virus and unclassified genera. Genomes of well-known phages of the families Siphoviridae, Myoviridae and Podoviridae, such as phage Lambda, T4 and T7, respectively, were added for comparison purposes. The genome analysis showed three well-defined clusters that mainly divided the phylogenomic tree into different phage genera (Phietavirus, Biseptimavirus and Triavirus). Two principal branches separated Clusters A, B and C from D. Cluster A was formed by Staphylococcus Phietavirus, two phieta-like viruses and two unclassified Staphylococcus phages. Cluster B was formed by Staphylococcus phages classified as Biseptimavirus and by one unclassified Staphylococcus phage. Cluster C was formed by enterobacterial bacteriophages and one SPbeta-like virus. Finally, cluster D was formed by Triavirus Staphylococcus phages and two unclassified Staphylococcus phages. To the best of our knowledge, this is the first time that phages from mastitis-causing staphylococci were grouped in a phylogenomic tree.
Specific peptides were found in related Staphylococcus spp. phages (Table 2) located closely in the phylogenomic tree ( Figure 2). Peptides HAGYVRC*KLF and MPVYKDGNTGKWYFSI were found in phages of cluster A. Furthermore, peptides IYDRNSDTLDGLPVVNLK, QKNVLNYANEQLDEQNKV, EVPNEPDYIVIDVC*EDYSASK, KSNVEAFSNAVK and KLYIIEEYVKQGM were found in Staphylococcus phages of the A.1 subbranch in cluster A. Additionally, peptide AVAELLKEINR was found in phages of the A.2 branch. The peptide AYINITGLGFAK was found in phages of cluster B.1, and TSIELITGFTK was found in phages of cluster B.2. Peptides VSYTLDDDDFITDVETAK and LLHALPTGNDSGGD-KLLPK, which belong to the phage major capsid protein, were found in the same 14 Staphylococcus phages of cluster D. Peptides ELAEAIGVSQPTVSNWIQQTK and IQQLA-DYFNVPK, which belong to the phage-repressor Cro/CI family of proteins, were found in the same bacteriophages of cluster D. Moreover, peptides LYVGVFNPEATK, RVSYTLD-DDDFITDVETAKELKL LYVGVFNPEATK, VLEMIFLGEDPK, KAMIKASPK, EFRNKL-NELGADK and GMPTGTNVYAVKGGIADK were also found in phages of cluster D. Peptides IHDKELDDPSEEESKLTQEEENSI, IIINHDEIDLL, KDRYSSVSY and AEEAGVTVKQL are specific to Staphylococcus phage SPbeta-like.   In addition, a correlation relating bacterial species for each cluster with all peptides found in the bacteriophages with 100% similarity was found. The results showed that clustered phages were related to specific species of Staphylococcus. All studied phages were found to be related to S. aureus; however, most of them were also found to be related to additional Staphylococcus species. S. argenteus was found to be related in all clusters of the phylogenomic tree. Cluster A phage peptides were found to be mainly related to S. simiae. However, different Staphylococcus species (S. xylosus, S. muscae, S. haemolyticus, S. simiae, S. sciuri, S. pseudintermedius, S. devriesei, S. warneri and S. capitis) were found to be related to phages of cluster D.

Identification of Peptides of Virulence Factors
In this work, 405 peptides from S. aureus were determined to be related to virulence factors (Excel dataset Supplemental Data). Among these peptides, proteins such as staphopain, beta-lactamase, elastin-binding protein peptides and a multidrug ATP-binding cassette (ABC) transporter were identified.  Table S2 in Supplemental Data 2. Genomes of the lambda (NC_001416.1), T4 (NC_000866.4) and T7 (NC_001604.1) phages were added for comparison purposes. The VICTOR phylogenetic tree construction was based on an intergenic distance analysis with the GBDP tool (Genome BLAST Distance Phylogeny). The significance of each branch is indicated by a pseudo-bootstrap value calculated as a percentage for 1000 subsets. Bar

Discussion
LC-MS/MS-based methods for bacteriophage identification offer several advantages compared with other approaches, since bacteriophages can be directly identified with this method without using genomic tools, which provides a new strategy for drawing the appropriate conclusions. In addition, the method proposed here may be applied for further analyses without the requirement of growing bacteria, since the samples can be collected directly from foodstuffs. The study of noninduced prophages provides a fast analysis and can detect specific temperate phage proteins produced by S. aureus while integrated in the bacterial genome or by phages that are infecting the bacteria. Both cases provide the identification of specific S. aureus species or strains-in this case, an S. aureus mastitis producer. In the proteomic repository of the 20 different S. aureus strains analyzed, 79 peptides from staphylococcal bacteriophages were identified. Among them, eighteen of these phage peptides were S. aureus-specific. As bacteriophages are host-specific, these putative diagnostic peptides could be good diagnostic biomarkers for the detection and characterization of S. aureus and S. aureus phages.
The results show that a given specific peptide is present in closely related phages ( Table 2). These bacteriophage peptides can be used as specific markers to establish S. aureus bacteriophage relationships ( Figure 2). Additionally, phages that show the same peptides and are specific to Staphylococcus spp. are located close to one another in the phylogenomic tree, suggesting that a link does exist between phage phylogeny and bacteriophages that can infect the same bacterial species.
The study shown here exemplifies how phylogenomic trees based on the genome analysis provide useful information, and the study corroborates previous investigations, which suggested that viral genomic or subgenomic region analyses provide the best tool for reconstructing viral evolutionary histories [48]. Nevertheless, the lack of knowledge of the phage genomic content [49] makes a phage analysis more difficult. The first priority must be the contribution of new large amounts of data for phages infecting bacteria [12].
In addition, there is an urgent need for novel therapies to treat and prevent mastitis [50]. Bacteriophage therapy is an alternative to the antibiotic treatment of bovine mastitis [51], with a high specificity and a low probability for bacterial resistance development [52]. Many studies have demonstrated the effectiveness of bacteriophages in a variety of animal models to fight several mastitis-causing pathogenic bacteria. Some studies have shown how virulent phages such as SPW and SA phages are active against bovine mastitis-associated S. aureus. Moreover, SAJK-IND and MSP phages have specific lytic activity against several strains of S. aureus isolated from mastitis milk samples [53]. Indeed, mouse-induced mastitis models decreased their bacterial counts after treatment with a vBSM-A1 and vBSP-A2 phage cocktail [54]. Finally, several temperate phage mixtures have been shown to be more effective than using a single temperate phage for inhibiting S. aureus. According to the data obtained for the different models of mastitis, phage therapy using bacteriophages in this study can be considered an innovative alternative to antibiotics for the treatment of mastitis caused by S. aureus.
Finally, the proteomic analysis by LC-ESI-MS/MS performed in this study provides relevant insights into the search for potential phage origin diagnostic peptide biomarkers for mastitis-causing S. aureus. In addition, this method may be useful for searching peptide biomarkers for the identification and characterization of mastitis-causing species and for finding new S. aureus phages useful as possible therapies for mastitis.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/foods10040799/s1: Figure S1: MS/MS spectrums for S. aureus-specific peptide biomarkers. The corresponding peptides were tested for specificity using the BLASTp algorithm. Excel Dataset Supplemental Data 1: Complete nonredundant peptide dataset. Supplemental Data 2: Table S1: Staphylococcus aureus (SA) strains used in this study. Table S2: Linage, authors and accession number of studied bacteriophages .