Transducing Particles of Staphylococcus aureus Pathogenicity Island SaPI1 Are Comprised of Helper Phage-Encoded Proteins

Abstract

The relationship between the composition of SaPI1 transducing particles and those of helper phage 80α was investigated by direct comparison of virion proteins. Twelve virion proteins were identified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and mass spectrometry; all were present in both 80α and SaPI1 virions, and all were encoded by 80α. No SaPI1-encoded proteins were detected. This confirms the prediction that SaPI1 is encapsidated in a virion assembled from helper phage-encoded proteins.

Staphylococcus aureus pathogenicity islands (SaPIs) are a family of 15- to 27-kb genetic elements that generally carry one or more superantigen genes, encoding a variety of enterotoxins and/or toxic shock syndrome toxin 1. SaPIs are stably integrated at specific chromosomal sites but can be mobilized following infection by certain staphylococcal bacteriophages (10) or by induction of endogenous prophages (12, 23). Specific high-frequency transfer of SaPI1 has been shown to occur in transducing particles that have a morphology similar to that of its helper phage, 80α, but with capsids that are about one-third the volume of those of the helper phage, commensurate with the smaller size of the SaPI1 DNA (19). Integration and excision of SaPIs are phage-like, occurring by site-specific recombination between a specific att sequence on the pathogenicity island and a corresponding chromosomal att site, which leads to the generation of short direct repeats flanking the integrated element. SaPI1 encodes an integrase that resembles phage integrases and is sufficient for the site-specific integration of the pathogenicity island but cannot promote SaPI1 excision in the absence of the helper phage (19). A second phage-like SaPI1 gene encodes a protein homologous to the small subunit of phage terminase, an enzyme involved in DNA encapsidation. However, there are no SaPI1 genes encoding recognizable virion structural proteins (10). This has led to the prediction that SaPI1 particles are comprised of phage-encoded proteins and the suggestion that generation of the smaller SaPI1 virions is similar to the interaction between satellite phage P4 and its helper, enterobacteriophage P2 (reviewed in reference 9). The work reported here presents a direct comparison of the proteins in SaPI1 transducing particles and 80α virions and confirms that both particles are comprised of the same 80α-encoded proteins.

Phage 80α and SaPI1 transducing lysates were prepared by 80α infection of SaPI1-negative S. aureus strain RN450 (16) and SaPI1-positive strain RN7045 (21), respectively, following standard procedures (17). Phage and transducing particles were concentrated by precipitation with polyethylene glycol 8000 (1% [wt/vol]) and NaCl (0.5 M) and purified by equilibrium sedimentation in CsCl. Although Ruzin et al. (19) did not obtain resolution of 80α and SaPI1 particles when a lysate was sedimented in a CsCl gradient, in these highly concentrated phage preparations we were able to observe two close but distinct bands at an approximate density (ρ) of 1.45 in the gradient from the SaPI1 transducing lysate. After a second equilibrium banding of the upper of these two bands, we obtained a preparation that was enriched at least sevenfold for SaPI1 transducing particles, as assayed by biological activity (plaque-forming titer, 3 × 10⁸/ml; transducing titer, 2 × 10⁹/ml). Direct examination of negatively stained phage particles using the electron microscope indicated that enrichment of this sample for the smaller SaPI1 transducing particles was actually significantly greater than this; fewer than 4% of the virions had large heads. A representative field is shown in Fig. 1B. The discrepancy between the two methods for determining the relative numbers of phage and transducing particles suggests that many of the SaPI1 particles may not be infective or that introduction of SaPI1 DNA into a recipient cell does not always result in formation of a stable transductant. Regardless, enrichment of this sample is clearly sufficient to allow identification of any novel proteins in SaPI1-specific virions.

FIG. 1.

Electron micrographs of 80α and SaPI1 preparations analyzed in this study. CsCl was removed by dialysis in 10 mM Tris-HCl, pH 7.8, 20 mM NaCl, and 1 mM MgCl_2. Samples were applied to a 200 mesh, formvar-coated copper grid, stained with 1% phosphotungstic acid, and magnified 10,000× using a JEOL-JEM-1230 transmission electron microscope equipped with an UltraScan 4000SP charge-coupled-device camera. (A) Phage 80α particles. (B) SaPI1-enriched virions. (C) 80α fraction containing apparent procapsids.

An additional band with an approximate density (ρ) of 1.4 was observed in the gradient from the 80α lysate; this band had a greatly reduced phage titer and contained mainly empty capsid-like particles (Fig. 1C). An equivalent capsid band was not observed in the gradient from the transducing lysate, but since the total yield of phage particles following infection of a SaPI-positive strain is reduced by at least 2 orders of magnitude (10), it could have easily been below the limits of visual detection.

Virion proteins from the three samples described above were resolved by electrophoresis on sodium dodecyl sulfate (SDS)-polyacrylamide gels. The banding pattern of 80α was similar to that reported by Stewart et al. (22), although they did not report the low-abundance, high-molecular-weight polypeptides (>75 kDa) shown in Fig. 2 and there are differences in resolution of low-molecular-weight polypeptides (<25 kDa).

FIG. 2.

Identification of proteins in 80α and SaPI1-enriched virion preparations. (A) Negative image of a Sypro-Ruby-stained 10% Criterion XT SDS-PAGE bis-Tris polyacrylamide gel (Bio-Rad, Hercules, CA). Phage samples were denatured for 15 min at 95°C in the loading buffer provided by the manufacturer and vortexed for 20 min prior to gel loading. Lanes a and e, Precision Plus Protein standard (10 kDa to 250 kDa); lane b, 80α; lane c, SaPI1; lane d, 80α procapsid fraction. Numbers indicate those bands excised (1 to 8 from both 80α and SaPI1 samples) and analyzed as described in the text. (B) Identification of virion proteins from the corresponding bands shown in panel A. While bands 2 to 4 could be assigned collectively to ORFs 59, 61, and 62, specific assignment of each of these three bands to the gene encoding it was not possible, since the three proteins were not resolved from each other. This uncertainty is indicated by the brackets in Fig. 2A and B.

The polypeptides detected in the 80α and smaller SaPI1 virions were indistinguishable by SDS-polyacrylamide gel electrophoresis (PAGE) (Fig. 2, lanes b and c). No additional SaPI1-specific bands were observed. The absence of SaPI1-encoded proteins in the mature virions is consistent with the expectation that SaPI1 DNA is encapsidated in a particle appropriated from the helper phage. The particles from the 80α empty capsid fraction contained three predominant proteins, two of which were found in mature virions as well and one of which was unique (Fig. 2, lane d). Based on the similarity of the novel protein in this fraction to phage scaffolding proteins and the identification of the other two proteins as the portal and major capsid proteins (see below), we conclude that this fraction corresponds to procapsids.

In order to confirm the identities of 80α and SaPI1 proteins of the same electrophoretic mobility and to allow assignment of virion proteins to the genes that encoded them, individual protein bands were excised and subjected to trypsin in-gel digestion. Peptides were then extracted and analyzed using nano-high-performance liquid chromatography electrospray ionization ion trap mass spectrometry (described in detail in method S1 in the supplemental material). Nine common virion proteins were identified in bands excised from gels of various percentages of acrylamide and are described in detail below. Further analysis was performed to detect possible additional virion proteins that were not sufficiently abundant for visualization on the polyacrylamide gels by disrupting whole 80α and SaPI1 virions with urea and analyzing tryptic peptides by mass spectrometry as described by da Poian et al. (3) with some modifications (see method S1 the supplemental material). In addition to the proteins identified on polyacrylamide gels, three additional 80α-encoded proteins of low abundance were identified in the analysis of entire virions, and like the other proteins these were present in both the 80α and SaPI1 particles. The individual peptides identified in each of these proteins are listed in Table S2 in the supplemental material.

Sequenced staphylococcal phages have been grouped into three classes based on genome size and organization; phages within each class also share virion morphology (7). Class II phages, with genomes of about 40 kb, belong to the family Siphoviridae. This class contains most of the characterized staphylococcal phages, including the commonly used generalized transducing phages and the prophages resident in sequenced S. aureus genomes. The sequence of 80α (GenBank accession no. DQ517338) indicates that it belongs to this class as well. The genomes of these phages are largely syntenic and, as is common for most characterized phages, exhibit extensive genetic mosaicism (2). With the exception of several phage genes involved in integration/excision and its regulation and in host cell lysis, gene function for these phages has largely been assigned based on amino acid similarities of predicted gene products to those encoded by known phage genes in the database. Some proteins had been classified as components of the head or tail based on previous electrophoretic analysis of disrupted virions (8). The analysis of virion proteins carried out in this study has allowed unambiguous assignment of 13 phage structural proteins to the genes that encode them (Fig. 3) and provided evidence supporting the functions of these gene products as predicted by BLAST and matches in the CDD database (13). Each of the specific individual proteins is discussed below.

FIG. 3.

Locations of genes encoding the proteins identified in this study. The late region of the 80α prophage genomic map is shown, from nucleotide 16407 to the end. Arrows indicate predicted open reading frames encoding proteins of at least 50 amino acids, as annotated in the database entry (accession no. DQ517338). Genes for which functional assignments have been made based on sequence similarity are indicated above the map. Shaded arrows represent genes for which the assigned function is supported by the protein analysis reported here, the black arrows indicate genes encoding proteins identified by SDS-PAGE and mass spectrometry, and the gray arrows indicate genes encoding proteins identified only in the mass spectrometry analysis of complete virions. The ORF number for each gene encoding an identified virion protein is indicated.

Protein 1, the largest protein encoded in the 80α phage genome, corresponds to accession number ABF71627 in the GenBank protein database, the product of open reading frame (ORF) 56. This protein is predicted to be the tail tape measure protein. It is present in relatively low abundance, as would be expected, and is absent from the procapsid fraction. Migration of this polypeptide is extremely aberrant; it has a predicted mass of 126 kDa but migrates with an apparent M_r in excess of 250,000. We cannot account for this discrepancy at present; it may be an artifact due to protein conformation, or it may represent a dimer that is cross-linked or otherwise resistant to dissociation by SDS. The gene that encodes this protein is in a location characteristic of tape measure genes, i.e., downstream of the major tail subunit gene and separated from it by two small, overlapping reading frames containing a potential −1 translational frameshift (26). The predicted polypeptide contains two conserved domains. The N terminus shows weak similarity to a variety of motifs that form coiled-coil structures, including those found in eukaryotic SMC proteins and the myosin tail. The C terminus contains several regions of similarity to COG5412, a phage-related protein domain of unknown function. This protein is 94% identical to the presumptive tape measure protein of staphylococcal phage φ11 and is highly conserved among staphylococcal phages.

Proteins 2, 3, and 4 were not sufficiently resolved by SDS-PAGE for mass spectrometry on individual bands. Analysis of tryptic peptides from the mixture of proteins did allow assignment of these proteins to predicted genes. These three proteins correspond to three predicted minor tail proteins, accession no. ABF71630, ABF71632, and ABF71633, encoded by ORF 59; ORF 61, and ORF 62, respectively. These proteins were much less abundant in the 80α procapsid fraction, consistent with the prediction that these proteins are components of the phage tail. They are similar to hypothetical proteins in other staphylococcal phages. Two of the proteins, encoded by ORF 61 and ORF 62, contain no known motifs. The protein encoded by ORF 59 contains two previously described domains. The N terminus shows weak similarity (4e−06) to DUF1142, a domain of unknown function found in phage and bacterial proteins. The C terminus contains a domain with similarity (3e−12) to the conserved domain of the SGNH hydrolase subfamily of lipases and esterases. Given the diversity of enzymatic activities among members of this family, the precise function of this domain is unclear, but we speculate that this activity plays a role in the degradation of cell wall components to allow infection.

Protein 5 corresponds to accession no. ABF71613, the predicted portal protein encoded by ORF 42. It is present in virions as well as in the 80α procapsid fraction. This protein shows significant similarity (9e−71) to the SPP1 Gp6-like family of phage portal proteins. It is highly conserved among other staphylococcal phages.

Protein 6 corresponds to accession no. ABF71639, the product of ORF 68. This protein is a component of the phage tail, based on its absence from the 80α procapsid fraction, and is similar to a large number of putative tail fiber proteins in other staphylococcal phages. The predicted protein sequence contains no known motifs.

Protein 7 corresponds to accession no. ABF71618, the predicted protein encoded by ORF 47. This protein was identified as the major capsid protein based on its high abundance in both the virion and procapsid fractions. While the protein contains no identified domains, it is similar to putative major capsid proteins from phages of a number of gram-positive bacteria.

Protein 8 corresponds to accession no. ABF71624, the protein encoded by ORF 53. This protein is predicted to be the major tail protein, based upon its high abundance in virions and its absence from the procapsid fraction. It shows similarity to other staphylococcal phages and to putative major tail proteins characterized in phages isolated from other gram-positive bacteria.

Protein 9 corresponds to accession no. ABF71617, the predicted protein encoded by ORF 46. This protein is abundant in the procapsid fraction but absent from mature virions. This suggests that it is the capsid scaffolding protein. It contains no known conserved domains. A BLAST search reveals high conservation of this protein among a number of staphylococcal phages.

ORF 44 encodes an additional virion protein of low abundance that was identified on an 18% SDS-PAGE gel (data not shown) as well as after digestion of whole phage particles. The predicted protein (331 amino acids; 38 kDa) corresponds to accession no. ABF71615, a putative minor head protein with close homologues in a number of staphylococcal phages. The C terminus of this protein has similarity (8e−09) to the Phage_Mu_F domain. Other proteins in this family include the minor head protein Gp7 from Bacillus subtilis phage SPP1, which has been implicated in routing viral DNA into the host bacterium during infection (25).

The product of ORF 49, corresponding to accession no. ABF71620, was identified in the whole-phage tryptic digest. The predicted protein (110 amino acids; 13 kDa) is highly conserved among staphylococcal phages and has 30% identity to a putative DNA packaging protein from Lactobacillus phage ΦJL-1.

The product of ORF 67, corresponding to accession no. ABF71638, was also identified only in the trypsin-digested whole-phage preparation. This protein (632 amino acids; 72 kDa) is highly conserved among staphylococcal phages and is predicted to be a peptidoglycan hydrolase. Like many of the phage-encoded endolysins involved in host cell lysis, this protein appears to be modular (11); it contains two putative enzymatic domains. The N terminus is similar (2e−13) to the CHAP domain, which has amidase activity, which is found in many gram-positive autolysins and phage endolysins and which catalyzes cleavage of the amide bond between N-acetylmuramic acid and l-alanine in bacterial cell wall peptidoglycan (1, 18). This domain, encoded by ORF 67, shares 21% amino acid identity with the CHAP domain of the well-characterized phage φ11 endolysin, which has been shown to hydrolyze staphylococcal cell walls (4, 20). The C terminus is similar (1e−50) to the LytD domain, which has β-N-acetylglucosaminidase activity and cleaves the glycosidic bond between the long glycan strands of N-acetylglucosamine and N-acetyl-muramic acid. The LytD domain is commonly found in staphylococcal autolysins, and this domain, encoded by ORF 67, shares 43% amino acid identity with the corresponding domain in the major autolysin from S. aureus NCTC8325 (5). The ORF 67 gene product protein is presumed to function as a virion-associated hydrolase that plays a role in genome penetration of the infected cell. This region of the 80α genome is highly homologous to that of staphylococcal phage φ11, which has been shown to have a virion-associated protein of approximately 47 kDa that has peptidoglycan hydrolytic activity (15). The φ11 gene encoding the protein with this activity has not been identified. φ11 ORF 49 does encode a protein that is 98% identical to the product of 80α ORF 67, but these proteins are predicted to be 72 kDa. If this conserved gene product is responsible for the observed hydrolytic activity, the protein must be subject to proteolytic processing, with only one of the two domains retaining activity independent of virion association. An alternative possibility is that the observed hydrolytic activity could be associated with the product of the adjacent gene (80α ORF 68/φ11 ORF 50), which encodes a 44-kDa virion-associated protein (band 6 in Fig. 2) that is also 98% identical between these two phages. If this is the case, it would represent a novel motif for peptidoglycan hydrolysis, since this protein contains no identified domains.

The product of ORF 69, corresponding to accession no. ABF71640, was another protein of low abundance identified only in the whole-phage tryptic digest. This protein (131 amino acids; 14 kDa) is conserved among staphylococcal phages but has no known function. Based on the location of the gene in the 80α genome, we would predict that this protein is a component of the phage tail.

The additional faint bands on the gel of >75 kDa, which were also present for both 80α and SaPI1 virions, were not analyzed individually by mass spectrometry. They were larger than any of the predicted gene products except for the 80α tape measure protein and may represent variants of that protein. The doublet running between 20 and 25 kDa was analyzed; the only phage peptides identified corresponded to the major capsid protein. It is likely that these small proteins represent processed forms of the major capsid protein. These two minor bands were not evident in the procapsid fraction.

All proteins of sufficient abundance to be visualized by SDS-PAGE or identified by mass spectrometry of whole virions were identical in the 80α and SaPI1 preparations. This demonstrates that SaPI1 transducing particles are comprised of virion proteins encoded by 80α. Furthermore, SaPI1 does not contribute any detectable components to the mature virion. In contrast, satellite phage P4 encodes a nonessential protein, Psu, which is present as a decoration protein on P4 virions and helps to stabilize the smaller P4 capsids (6). SaPI1 appears to encode three proteins involved in directing assembly of 80α-encoded capsid precursors into the smaller SaPI1 capsids (24), but their roles have not been defined. Any that are incorporated during assembly must be removed during maturation, as is the case with the external scaffolding protein P4 Sid (14). A comparison of 80α and SaPI1 procapsids will be required to identify SaPI1-encoded components that play a structural role during capsid assembly.