Introduction

Due to the appearance and dissemination of antibiotic-resistant bacteria, concern is increasing that certain bacteria will develop resistance to all known antibiotics, and we will enter a post-antibiotic era. Recently, phage therapy has regained interest among scientists, and some phages have been used in clinical trials with no harmful effects on the trial participants1,2,3.

While phages can be vectors for horizontal gene transfer, several studies have shown that they can also be sources for the transmission of virulence and antibiotic resistance genes among bacteria4, thereby accelerating the evolution of virulence and/or antibiotic resistance in bacteria. Phages containing such undesirable genes could be harmful to human health and must be excluded from phage therapy. One study confirmed that Stx production was phage-regulated in stx(2g)-positive strains (Shiga toxin-producing Escherichia coli (STEC))5. Report also showed that bacteriophages from poultry, cattle and pigs containing antibiotic resistance genes6. These studies suggest that bacteriophages could be environmental vectors for the horizontal transfer of virulence and antibiotic resistance genes.

Staphylococcus phages of the Myoviridae family have been isolated worldwide. Published data show that they have a wide host range and can potentially be used for phage therapy. To date, the complete genome sequences of dozens of Staphylococcus phages of the Myoviridae family have become available, including phage K, G1, Twort, A5W, Sb-1, ISP, SA5, GH15, JD007, SA11, vB_SauM_Remus, vB_SauM_Romulus, S25-3, S25-4, phiIPLA-RODI, phiIPLA-C1C, phiSA012, Team1, P108, MCE-2014, 812, SA1, Staphy1N, MSA6, 676Z, P4W, and Fi200w7,8,9,10,11,12,13,14,15. These phages can potentially be used for the treatment of infectious diseases caused by Staphylococcus aureus in animal models. Phage S25-3 has been shown to prolong life in the silkworm larval infection model and other animal infection models16. The Staphylococcus phages Romulus and Remus have infected approximately 70% of the tested S. aureus strains and display promising lytic activity against these isolates15. The Staphylococcus phage ISP was shown to be effective against 86% of tested isolates, including relevant methicillin-resistant S. aureus (MRSA) strains7. These studies demonstrate the potential use of phages for the prevention of infectious diseases caused by S. aureus. In the present report, a safety assessment based on complete genome sequences of Staphylococcus phages of the Myoviridae family was conducted to evaluate their suitability for clinical use.

Methods

Complete phage genomes

Staphylococcus phages of the Myoviridae family, including the previously reported phage JD007, had complete genome sequence data. The complete genome sequences were collected from the NCBI database (http://www.ncbi.nlm.nih.gov/genome/viruses/). All Staphylococcus phages belonged to the Myoviridae family and were verified by detailed information provided by the NCBI phylogeny system and their morphology. The published papers related to these phages were also screened. The Staphylococcus phages were confirmed to belong to the Myoviridae family by morphological characterization of the contractile sheath between the head and tail.

General genomic features of the phages

The complete genome sequences of Staphylococcus phages of the Myoviridae family were downloaded in FASTA format from NCBI along with their annotation information. The genomes of the phages were re-annotated using RAST, and each ORF and protein domain search was performed using the InterProScan program and CDD17,18. The phylogeny tree was constructed using MEGA5 with the complete genome sequences based on the neighbour-joining method19. The visual alignments of the complete genomes were constructed using the Mauve2.3.1 program20.

Lifestyles of the phages

The lifestyles of phages (virulent or temperate) were predicted using the PHACTS program (http://www.phantome.org/PHACTS/index.htm)21. The genome annotation results of each phage were analyzed using PHACTS, and their lifestyle was confirmed. PHACTS utilizes a novel algorithm and a supervised Random Forest classifier to predict whether the lifestyle of a phage is virulent or temperate. The algorithm creates a training set from phages with known lifestyles. PHACTS predictions have been shown to have a 99% precision rate, and PHACTS can also determine the lifestyle of a phage using only genomic data. A current limitation of PHACTS is that a confident lifestyle prediction cannot be made for a small number of phages. This is primarily due to the variability that arises from the random sampling during classifications. If an unknown phage does not have any similarity to phages with known lifestyles in the database, the predictions will be less certain. It is expected that as more phages with known lifestyles are added to the database, the precision rate and sensitivity of the predictions will increase21. Additionally, the experimental results of the bacteriophage one step growth curve reported in the corresponding papers were also analyzed to confirm the phage lifestyles.

Analysis of virulence and antibiotic resistance genes

All annotated phage ORFs were used to search an antibiotic resistance gene database (ARDB, http://ardb.cbcb.umd.edu/)22 and a virulence factor database (VFDB, http://www.mgc.ac.cn/VFs/main.htm)23. Hits with more than 70% coverage and 30% identity were considered as positive results. Lihong et al. improved the infrastructural dataset of VFDB: (i) removed the redundancy introduced by previous releases and generated two hierarchical datasets–one core dataset of experimentally verified virulence factors (VFs) only and another full dataset including all known and predicted VFs. Their efforts enhanced the data quality of the VFDB and promoted the usability of the database in the big data era for the bioinformatic mining of the explosively growing data regarding bacterial VFs23. Antibiotic Resistance Genes Database (ARDB)—unifying most of the publicly available information on antibiotic resistance. Each gene and resistance type is annotated with rich information, including resistance profile, mechanism of action, ontology, COG and CDD annotations, as well as external links to sequence and protein databases. ARDB database also supports sequence similarity searches and implements an initial version of a tool for characterizing common mutations that confer antibiotic resistance. The information they provided can be used as compendium of antibiotic resistance factors as well as to identify the resistance genes of newly sequenced genes, genomes, or metagenomes22.

Phylogenetic relationship of the phages

To identify the phylogenetic relationship among Staphylococcus phages of the Myoviridae family, other S. aureus phages, including Podoviridae (including Staphylococcus phages 44AHJD, 66, GRCS, P68, PT1028, S24-1, and SAP-2) and Siphoviridae (including Staphylococcus phages 3MRA, 13, 29, 52A, 55, 69, 77, 92, 187, 2638A, DW2, IME-SA4, Ipla7, P954, phi5967PVL, phiSa119, phiSauS-IPLA88, PVL, SA13, SA97, SAP-26, SpaA1, StauST-398-3, StB20, StB20-like, StB27, 3A, 11, 23MRA, 37, 71, 88, 96, B166, B236, CNPH82, Ipla5, JS01, phiBU01, phiETA, phiETA2, phiETA3, phiJB, phiMR11, phiNM1, and phiNM2) (phages’ nucleic acid access numbers refer additional file 1), were chosen to evaluate the phylogenetic relationship using complete genome sequences (http://www.ncbi.nlm.nih.gov/genome/viruses/). The neighbour-joining tree was constructed using MEGA519.

Results

General genomic features of Staphylococcus phages of the Myoviridae family

A total of 22 Staphylococcus phages with complete genome sequence data were examined: K, G1, Twort, A5W, Sb-1, ISP, SA5, GH15, JD007, SA11, vB_SauM_Remus, vB_SauM_Romulus, S25-3, S25-4, phiIPLA-RODI, phiIPLA-C1C, phiSA012, Team1, P108, MCE-2014, 812, and SA17,8,9,10,11,12,13,14,15. The Staphylococcus phages Staphy1N, MSA6, 676Z, P4W, and Fi200w, for which only partial genome sequences were available, were excluded from this study. The phages have been isolated worldwide, and all belong to the Myoviridae family. As shown in Table 1, the genomes were approximately 127 kb–148 kb in length, contained between 170 and 220 ORFs, and had a GC content of 27–31%. However, the GC content of phage SA1 was 45.83%, which was much higher than other Staphylococcus phages of the Myoviridae family.

Table 1 The Staphylococcus phages of the Myoviridae family with published complete genome sequences (through June, 2016).
Full size table

Lytic or lysogenic analysis

PHACTS, a computational approach used to classify the lifestyle of bacteriophages, was used after the complete genome sequences were annotated and the proteins were converted into the FASTA format. The results showed that the Staphylococcus phages K, G1, Twort, A5W, Sb-1, ISP, SA5, GH15, JD007, SA11, vB_SauM_Remus, vB_SauM_Romulus, S25-3, S25-4, phiIPLA-RODI, phiIPLA-C1C, phiSA012, Team1, P108, MCE-2014, 812, and SA1 belong to the Myoviridae family and are all lytic phages7,8,9,10,11,12,13,14,15. The previous report shows that phage phiIPLA-RODI and phiIPLA-C1C are lytic phages11, our results were consistent with the studies reported previously.

Genome organization and annotation

The complete genomic sequences of Staphylococcus phages of the Myoviridae family were compared. As showed in Fig. 1, the genomic structure was chimeric and linear. Several functional modules shown in Table 2 consistently exhibited mosaicism in the genomes, including those involved in packaging, lysis, structure, phage and host interactions, DNA manipulation, and some ORFs with additional functions. High degrees of sequence similarity were evident across the phage sequences, which is unique among Staphylococcus phages of the Myoviridae family. While, recombinase gene was widely existed in the complete genomes of Staphylococcus phages of the Myoviridae family, what suggest that genomic sequences combination may occur between the homologue sequences during phage infection. The difficulties to knock out recombinase gene in these phages may block experimental validation the fully inactivity of Recombinase or confirmation there were no homologue sequences combination occur in the recombinase knock-out phages.

Figure 1
figure 1

Comparative genomic analysis of Staphylococcus phages of the Myoviridae family.

Full size image
Table 2 The functional modules of the ORFs in the genomes of Staphylococcus phages of the Myoviridae family.
Full size table

Analysis of virulence and antibiotic resistance genes

All of the predicted ORFs were compared with the sequences in the Antibiotic Resistance Genes Database (ARDB) and the virulence factors database VFDB using the BLASTn algorithm, with the criteria that genes with more than 70% coverage and 30% identity were considered to exhibit positive results. No significant hits based on the ORFs predicted in the complete genomes of the Staphylococcus phages of the Myoviridae family were obtained.

As shown in Fig. 2, the SceD-like transglycosylase, which is a biomarker for vancomycin-intermediate strains24, was encoded in the genome of Staphylococcus phages K, G1, Twort, A5W, Sb-1, ISP, SA5, GH15, JD007, SA11, vB_SauM_Remus, vB_SauM_Romulus, S25-3, S25-4, phiIPLA-RODI, phiIPLA-C1C, phiSA012, Team1, P108, MCE-2014, and 8127,8,9,10,11,12,13,14,15. Furthermore, approximately 70% of the ORFs were predicted to have unknown functions. It suggested that these ORFs encoding proteins with no similarity functional domains of proteins in InterProScan and CDD database.

Figure 2
figure 2

The SceD-like transglycosylase gene is widely distributed in the genome of Staphylococcus phages of the Myoviridae family.

Full size image

Clustering of Staphylococcus phages of the Myoviridae family

The phylogeny tree showed that Staphylococcus phages of the Myoviridae family clustered in the same branch, based on the alignments of the complete genome sequences. As shown in Fig. 3, the phages K, G1, Twort, A5W, Sb-1, ISP, SA5, GH15, JD007, SA11, vB_SauM_Remus, vB_SauM_Romulus, S25-3, S25-4, phiIPLA-RODI, phiIPLA-C1C, phiSA012, Team1, P108, MCE-2014, 812, and SA17,8,9,10,11,12,13,14,15 clustered together on the same sub-branch with other Staphylococcus phages of the Myoviridae family. These results are consistent with the morphology of the phages. It is well known that the proteins such as Helicase, DNA Polymerase, Primase, Large terminase and Major Capsid Protein encoded in the genome of phages are usually chosen to draw phylogenic trees25. However, such genes are not always simultaneously existed in the genomes of phages our study enrolled, so it is difficult to cluster phages using these genes when some phages have while others are not. With the development of genome sequencing recently, the complete genome sequences of viruses with high identity were chosen to draw the phylogenic trees with high accuracy.

Figure 3: The phylogenetic tree of Staphylococcus phages of the Myoviridae family.
figure 3

The diagram represents the phylogeny constructed using the complete phage genome sequences via the neighbour-joining method.

Full size image

Discussions

Phage therapy has become a subject of renewed interest recently, and some phages have been used in clinical trials, including the Staphylococcus phages of the Myoviridae family. Several studies have evaluated the safety of certain phages in clinical trials. Rhoads et al. reported a study in which ulcers were treated for 12 weeks with bacteriophages targeting Pseudomonas aeruginosa, S. aureus, and Escherichia coli or a saline control3. Follow-up continued until week 24, and no adverse events were attributed to the phage therapy. In addition, no significant difference was determined between the test and control groups regarding the frequency of adverse events, rate of healing, or frequency of healing3. Unlike normal antibiotics, a bacteriophage is a type of virus that contains genetic material. Therefore, potentially useful phages require full evaluation prior to use in phage therapy. A bacteriophage may be a potential mechanism for the transmission of antibiotic resistance genes or virulence genes among bacteria. A prophage may contribute the pathogenic traits of Enterococcus feaclis26. Antibiotic resistance genes, including blaTEM, qnrA, blaCTX-M-1, and toxin genes, have been found in the DNA of bacteriophages isolated from human faecal samples27,28. Therefore, safety assessment of phages at the genome level is critically important. First, an analysis should determine whether the bacteriophage carries genes that may accelerate the virulence and antibiotic resistance of bacteria when these genes integrate into their genomes. Second, the ability of the bacteriophage genome that will integrate into the genome of bacteria should be assessed. In this evaluation, it is important and necessary to assess the safety of bacteriophages at the genome level. McCallin et al. evaluated the safety of a phage cocktail from one company, and a small volunteer trial did not report an association between adverse effects and oral phage exposure29. During that study, the genome sequences of phages were obtained, and bioinformatics analysis revealed that no undesirable genes were present29. A safety assessment of the phage at the genome level is an essential step during the evaluation of the suitability of phages for therapeutic applications.

Staphylococcus phages of the Myoviridae family are potentially useful for phage therapy. Dozens of phages have been shown to protect mice from death after infection with S. aureus. Some phages have been used in clinical trials, and no adverse effects have been reported. In this report, our results showed that none of these bacteriophages carried any antibiotic resistance or virulence genes. In addition, all of the studied bacteriophages were virulent and were therefore unable to integrate into the genome of S. aureus. Our studies show that Staphylococcus phages of the Myoviridae family are safe at the genome level. While, as showed in Table 2, recombinase was widely distributed in phage GH15, JD007, MCE-2014, phiIPLA-C1C, phiSA012, SA11, Twort. Team1, vB_SauM_Romulus and vB_SauM_Remus, it was reported that Recombinase could recombine homologous genome sequences between bacteriophages or/and its host30,31; and the SceD-like transglycosylase, which is a biomarker for vancomycin-intermediate strains24 is encoded in the genome of Staphylococcus phages K, G1, Twort, A5W, Sb-1, ISP, SA5, GH15, JD007, SA11, vB_SauM_Remus, vB_SauM_Romulus, S25-3, S25-4, phiIPLA-RODI, phiIPLA-C1C, phiSA012, Team1, P108, MCE-2014, and 8127,8,9,10,11,12,13,14,15. The expression of this gene during S. aureus infection may increase the minimum inhibitory concentration of vancomycin used for the treatment of infections caused by MRSA; furthermore, approximately 70% of the ORFs encoded in the phage genomes have unknown functions, and their potential roles in S. aureus antibiotic resistance and virulence are unknown, what’s more, with the development of genome sequencing, acquisition of a complete genome sequence of organism become much more easily, and lots of ORFs encoded proteins with unknown functions32. All of the above factors represent a potential genome-level risk for phage therapy. Therefore, the clinical use of these phages should be evaluated comprehensively.

In summary, Staphylococcus phages of the Myoviridae family have been isolated worldwide. Morphology studies have shown that these phages are contractile between the head and tail. In addition, they have a wide host range and can prevent infectious diseases in animal models, and pre-clinical and clinical trials have not reported any adverse effects. Our results show that Staphylococcus phages of the Myoviridae family and are lytic and therefore cannot integrate into the genomes of their host. In addition, no antibiotic resistance or virulence genes were present in their genomes, indicating that they are safe at the genome level. In addition, the phages analyzed in this study clustered in the same branch of an evolutionary tree, suggesting that they share a common ancestor.

Additional Information

How to cite this article: Cui, Z. et al. Safety assessment of Staphylococcus phages of the family Myoviridae based on complete genome sequences. Sci. Rep. 7, 41259; doi: 10.1038/srep41259 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.