Bioinformatic Analysis

DNA sequences were downloaded from NCBI GenBank [57]. Inducible prophage t712 of NZ9000 [58], called TP712 by Roces et al. (2013) [59], was identified using the PHAge Search Tool (PHAST) [60]. BLAST, Pfam and HHpred analyses were used for functional annotations of proteins [61–64]. Putative promoter sequences of NZ9000 prophage t712 were identified using SoftBerry BPROM (http://www.softberry.com). Significant HHpred hits of TP901-1 structural module proteins are presented in Table 1, alongside Protein Data Bank (PDB) identification codes and the original source of the hit. Alignment of multiple amino acid sequences was performed using ClustalW [65]. Protein secondary structure predictions and intrinsically disordered regions were detected using Quick2D (http://toolkit.tuebingen.mpg.de/) [66–70].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Bioinformatic analysis of TP901-1’s predicted structural-encoding proteins.

https://doi.org/10.1371/journal.pone.0131676.t001

Bacteria, Phages and Growth Conditions

Bacterial strains and phages used in this study are listed in S1 Table. Growth experiments of lactococcal strains and phage propagations were performed as described previously [13, 71]. TP901-1erm and mutant derivatives were induced from corresponding lysogens of L. lactis NZ9000-Cro_t712 using the following conditions: the relevant strains were grown at 30°C to an A₆₀₀ of 0.3, at which point mitomycin C (Sigma) was added to a final concentration of 3 μg/ml with subsequent incubation at 20°C until lysis occurred. NaCl, 1 M (w/v) final concentration, was added to the resultant phage lysates followed by centrifugation in a Thermo Scientific SL16R centrifuge at 5580 × g for 15 min and stored at 4^°C.

Mutant Generation

Recombineering mutagenesis was performed as described previously [13, 72, 73]. A detailed description of all TP901-1erm-derived mutants created in this study is provided in S2 Table. Point mutations, creating a BamHI restriction site, were introduced in the deduced promoter sequence of the predicted anti-repressor-encoding gene (cro_t712) of prophage t712 of NZ9000 in order to prevent its induction during mitomycin C addition (generating strain NZ9000-Cro_t712). Mutations in targeted ORFs of the TP901-1 structural module were generated so as to insert an in-frame stop codon, thereby terminating translation (mutant nomenclature ORF_TP901-1::Ter), and where possible, introduce a restriction enzyme site. In order to prevent gpT production, an in-frame BamHI restriction site was introduced into the ‘slippery’ sequence (GGGAAAG) at the 3’ end of the TP901-1 gpG gene which is required for the sequence-specific -1 ribosomal frame-shift required for gpT translation (mutant termed gpT_TP901-1::BamHI) [48]. To further analyse the production and role of chaperone protein gpT, a single nucleotide insertion, coupled with several point mutations, was introduced into the gpG ‘slippery’ sequence resulting in a direct translational fusion between the gpG and gpT gene sequences (the resulting mutant phage was designated gpGfsT_TP901-1). Mismatch Amplification Mutation Assay (MAMA) PCRs using oligonucleotides complementary at their 3’ end to the desired nucleotide changes [73], and restriction digests, were performed to identify NZ9000-Cro_t712 derivatives (following recombineering mutagenesis) which contained the anticipated mutated TP901-1erm prophage. Oligonucleotides used for recombineering and MAMA PCR screening reactions were purchased from Integrated DNA Technologies (IDT, Belgium), and are listed in S3 Table. Mutations were confirmed by Sanger sequencing relevant PCR-amplified regions using chromosomal DNA from the mutated TP901-1erm-containing L. lactis NZ9000-Cro_t712 lysogen as a template (sequencing performed by MWG, Germany).

Phage Purification

Phage lysates of TP901-1erm wild type and the TP901-1erm-derived mutants Portal_TP901-1::Ter, MCP1_TP901-1::Ter, MHP_TP901-1::Ter, Ttp_TP901-1::Ter, MTP_TP901-1::Ter, gpG_TP901-1::Ter, gpT_TP901-1::BamHI and gpGfsT_TP901-1 were purified by sucrose gradient ultracentrifugation as follows: polyethylene glycol 8000 (PEG8000; 10% w/v) was added to NaCl-treated (as described above) phage lysates and mixed gently at Room Temperature (RT) until completely dissolved. Phage samples were placed on ice for 3–4 hr before centrifugation in a 4°C pre-chilled Thermo Scientific SL16R centrifuge at 5580 × g for 20 min. Pellets were air-dried, before resuspending in 1:50 initial volume of SM buffer (50 mM Tris-HCl, 100 mM NaCl, 10 mM MgSO₄; pH 7.5). PEG8000 was removed from phage samples by extraction with an equal volume of chloroform, followed by vortexing for 30 sec and centrifugation at 1660 × g for 15 min. The aqueous phase was extracted and applied to a sucrose gradient. Linear sucrose gradients, 20–70% (w/v), were prepared using a Masterflex peristaltic pump (Thermo Scientific) in Beckman Thinwall Ultra-Clear centrifuge tubes (14 x 89 mm). Sucrose was added to all phage samples, at a final concentration of 14% (w/v), before layering them on top of the prepared linear sucrose gradients. Samples were centrifuged for 1.5 hr at 250,000 × g in a Beckman SW40Ti ultracentrifuge rotor at 4°C. Bands were extracted from the sucrose gradient and immediately dialysed against SM buffer (12–14 kDa MW cut-off; Medicell International Ltd., London, UK). For TP901-1erm-derived mutants MCP2_TP901-1::Ter, Sfp_TP901-1::Ter, HTC1_TP901-1::Ter, HTC2_TP901-1::Ter and Tap_TP901-1::Ter, CsCl purification was performed instead of sucrose gradient purification as this permitted higher quality visualization of the produced virion structures. Chloroform-extracts of PEG-precipitates of mutant phage lysates (as described above for sucrose gradient purification) were prepared and then used for CsCl gradient purification of phage particles as previously described [74].

Phage Assays

Spot and plaque assays of TP901-1erm and its mutant derivatives were performed as described previously [75]. In vitro assembly of TP901-1erm-derived structural module mutants was performed using sucrose gradient-purified bands of TP901-1erm and mutant derivatives of this phage (in order to standardise the results obtained, only sucrose gradient-purified bands, and not CsCl purified bands, were tested). Sucrose gradient-purified fractions of mutant MHP_TP901-1::Ter and mutant MTP_TP901-1::Ter were used as either a phage tail or capsid donor, respectively. Briefly, 10 μl of purified TP901-1erm capsids or tails were mixed with 10 μl of purified TP901-1erm structural module mutant derivative fractions, and incubated for 16 hr at 30°C. The phage mix was then tested for its ability to infect L. lactis 3107 by assessing the frequency of lysogeny, which was determined through acquisition of the adenine methylase-encoding gene present on the TP901-1erm phage genome (or its mutant derivatives), conferring resistance to 5 μg/ml erythromycin [11, 76]. This was performed by adding 50 μl of the phage mix to 450 μl of L. lactis 3107 at an A₆₀₀ of 0.3 and incubated at 30°C for 1 hr before plating. Three technical replicates calculating the frequency of lysogeny were performed and results averaged.

Protein Expression, Purification and Antibody Generation

Antibodies against Tuc2009 tail proteins were previously generated and proved successful in detecting the homo-immune structural proteins of Tuc2009 and TP901-1 [10]. Therefore, the following Tuc2009 capsid proteins were overproduced and purified in L. lactis NZ9000: Portal_Tuc2009, MCP1_Tuc2009, MCP2_Tuc2009, Sfp_Tuc2009, MHP_Tuc2009, MCP3_Tuc2009, HTC1_Tuc2009 and Tap_Tuc2009. Attempts to express recombinant proteins of HTC2_Tuc2009 and Ttp_Tuc2009 were unsuccessful (data not shown). Recombinant protein-expression constructs created during this study, and the oligonucleotides used to create them, are outlined in S1 and S3 Tables, respectively. Primers used for cloning were purchased from Eurofins MWG (Germany). KOD high-fidelity DNA polymerase (Novagen, UK) was used for PCR amplifications. For cloning reactions, restriction enzymes were supplied by Roche (Germany) and ligations were performed with T4 DNA ligase (Promega, USA). For genes cloned using pNZ8048, NcoI and SpeI were used, while for pTX8049, BamHI and SpeI were employed. In both cases, primers were used so as to facilitate the incorporation of a hexahistidine-tag at the C-terminus of the encoded protein, while pTX8049-generated proteins also incorporate an N-terminally fused thioredoxin tag [77]. Proteins were expressed and purified as outlined previously [77]. Rabbit antibody production was carried out by Harlan Laboratories (Leicester, UK). Immunization was initially carried out with individual proteins supplemented with Freund’s adjuvant at a concentration of ~200 μg/ml; this was followed by five booster injections over the 112 day protocol.

SDS-PAGE and Western Blotting

SDS-PAGE separation of proteins was performed using 12% or 15% (v/v) acrylamide gels following standard procedures [78]. After separation of ~10 μg protein by SDS-PAGE, proteins were transferred to a nitrocellulose membrane by electro-blotting using 10 mM CAPS [3-(cyclohexylamino)-1-propanesulfonic acid; pH 11], 10% methanol transfer buffer. Membranes were blocked for 1 hr at RT with 50% (v/v) Odyssey blocking buffer in PBS (Licor, USA). Optimal conditions for each rabbit polyclonal antibody were determined empirically and the reproducibility of the results verified by repeated Western blot analyses during these optimisations (data not shown). Membranes were incubated with primary rabbit polyclonal antibodies, in a range from 1:500–1:15000, in blocking buffer containing 0.12–0.25% (v/v) Tween-20 for 1 hr at RT or overnight at 4°C. Following 4 × 5 min washes in 0.1% (v/v) Tween-20, the membranes were incubated in 1:7500 fluorescently labelled goat anti-rabbit IgG secondary antibody (Licor, USA) in blocking buffer containing 0.12–0.25% Tween-20. Membranes were washed again and visualised on a Licor Odyssey imaging system (USA).

Electron Microscopy

For negative staining, 5 μl of each sample were applied onto glow-discharged carbon-coated grids (Agar Scientific, Stansted, UK) and incubated for 1 min. The grids were washed with 5 μl of deionized water before incubating for 30 sec in 1% (w/v) of Uranyl Acetate (Agar Scientific, Stansted, UK). CCD images were collected using a Tecnai Spirit operated at 120 Kv and a 2Kx2K CCD camera.

Optimizing TP901-1erm host background

Several improvements were made to optimize the host background for analysing mutations made in TP901-1erm, a derivative of TP901-1 that is marked with an erythromycin-resistance (Em^r) cassette[76], which had been used to select a TP901-1erm lysogenized derivative of strain L. lactis NZ9000 [13]. This lysogenized NZ9000 derivative offered an advantage over the original TP901-1erm host, L. lactis 901–1, as the latter strain contains a second prophage similar in sequence to TP901-1 which may supply proteins in trans (Finn Vogensen, personal communication). In addition, NZ9000 contains an integrated copy of the NisRK cassette for protein expression using the nisin-inducible promoter system [79], which was used for the overproduction of RecT for recombineering mutagenesis [73]. In order to further optimize NZ9000 as a background for analysing mutations in TP901-1erm, a BamHI restriction site was introduced into the deduced -10 promoter sequence of the lysogenic/lytic switch of prophage t712 (thereby generating strain NZ9000-Cro_t712). Causing ablation of this promoter was aimed at preventing t712 prophage induction and virion production. Induction of prophage t712 was identified as a potential problem as; (i) t712 particle production would interfere with TP901-1 phage particle study, (ii) t712 induction would increase the metabolic burden on the host when only TP901-1 induction was desired, and (iii) it was deemed possible that, due to the conserved nature of Siphoviridae proteins, t712 may supply TP901-1 with structural proteins being analysed during this study in trans [59]. This expectation was indeed verified, as mutagenesis of the deduced promoter sequence for t712 Cro protein caused an approximate 100-fold increase in the number of produced infective TP901-1erm phage particles obtained following mitomycin C induction from the NZ9000-Cro_t712 background as compared to that obtained when strain NZ9000 was used (data not shown; indicator strain used for TP901-1erm phage titre determination was L. lactis 3107).

Annotation of TP901-1erm Structural Module and Mutants

The structural module of TP901-1 is represented by a polycistronic operon, approximately 20 kb in length, transcribed from a promoter located upstream of the gene encoding the predicted small terminase subunit [7, 8, 12]. This structural module of TP901-1 is predicted to encode 22 proteins (Fig 1) [7]. Functional analyses have previously been performed for a number of TP901-1 structural proteins (and by inference those of the closely related phage Tuc2009) [6, 9, 11, 13, 16, 19, 56, 80, 81]. Due to the conserved nature of phage structural proteins, functions of several TP901-1 proteins can be predicted based on a number of well characterized phage prototypes [42, 55, 82–84]. However, some TP901-1 proteins do not exhibit significant similarity to known proteins and thus their functional assignment is based on literature and bioinformatic analyses.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schematic representation of the 22 annotated genes of the structural module of phage TP901-1.

Mutagenesis of individual genes of the TP901-1 structural module showed gene products to be essential (+) or non-essential (-) for creating infectious phage virions. The efficiency of plaquing is displayed for mutants of non-essential gene products. The genomic regions encoding capsid and tail morphogenesis proteins are indicated and such assignments are based on data described in the text.

https://doi.org/10.1371/journal.pone.0131676.g001

Stop codon mutations in TP901-1erm structural module mutants were created, via recombineering, in the identified genes from orf32 through to orf43, inclusive, which encode the portal protein and the TAC gpG, respectively. Two additional mutations were created to investigate the production and role of orf44, encoding gpT (S2 Table). The nomenclature assigned to TP901-1erm structural module mutants is based on the abbreviated name assigned to the phage open reading frames (Table 1), followed by a designation signifying the insertion of a translation terminating stop codon or restriction site (e.g. MCP1_TP901-1::Ter or gpT_TP901-1::BamHI, respectively); the TP901-1erm mutant that carries a translational fusion mutation of ORF gpG and gpT was termed gpGfsT_TP901-1 (see also Materials and Methods section).

Complementation of TP901-1erm Structural Mutants

Individual TP901-1erm-derived mutants were induced from their lysogenic host by mitomycin C (see Materials and Methods). Obtained lysates were then tested for their ability to infect host strain L. lactis 3107 by spot and plaque assays to determine which genes/proteins are essential for TP901-1 phage assembly and/or infection. Of the 14 mutants created in the TP901-1erm structural module, only mutants carrying a non-sense mutation in the genes for MCP2_TP901-1 or MCP3_TP901-1 retained sufficient residual infectious ability to form plaques, while at the same time displaying decreased Efficiencies of Plaquing (EOPs) relative to wild-type TP901-1erm control (p-values < 0.01; Fig 1). Mutant phages were then tested for the ability to lysogenize their host bacterium. Lysogeny assays, which reflect a phage’s ability to adsorb to its host and to perform DNA injection followed by integration, rather than its ability to form a plaque, demonstrated that sucrose gradient-purified preparations of mutants MCP1_TP901-1::Ter, Sfp_TP901-1::Ter, HTC1_TP901-1::Ter, HTC2_TP901-1::Ter and Tap_TP901-1::Ter were able to generate Em^r lactococcal colonies (Table 2). While lysogeny frequencies of these mutants were substantially below that of the TP901-1erm control, the results were noted in subsequent analyses.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Lysogeny by TP901-1erm and structural module purified mutants mixed with phage capsids and tails.

https://doi.org/10.1371/journal.pone.0131676.t002

Previous studies of phage λ have shown that capsids and tails can be produced independently, and, following purification and subsequent mixing, can assemble spontaneously in vitro [85]. In order to determine if capsid and tail structures of lactococcal phage TP901-1 are similarly capable of in vitro assembly, purified capsids and tails of TP901-1 were complemented with the generated mutants and tested for the ability to lysogenize host strain L. lactis 3107. Unable to establish lysogeny on their own, following incubation of purified MHP_TP901-1::Ter and MTP_TP901-1::Ter, which act as intact phage tail and capsid donors, respectively, lysogenization of L. lactis 3107 did occur as was deduced from the appearance of Em^r-colonies (Table 2). This demonstrates that lactococcal phage TP901-1erm capsids and tails can indeed assemble in vitro to form functional, DNA-injecting virions.

By complementing TP901-1erm structural module mutants through the addition of either functional capsids or tails, it was possible to distinguish gene products required for functional capsid formation from those that are crucial for tail production. There is a clear distinction between TP901-1erm mutants that establish lysogeny at higher levels following in vitro assembly with either phage capsids or phage tails. TP901-1erm derivatives carrying individual mutations targeting the portal protein-encoding gene (orf32) through to the gene specifying the HTC2 protein (orf39) were able to lysogenize L. lactis 3107 following incubation with purified capsids (donated from TP901-1erm MTP mutant; see Materials and Methods section) to a significantly higher level as compared to the mutant on its own or when mixed with purified tails (Table 2). An exception to the in vitro assembly and lysogeny assays is MCP1_TP901-1::Ter, where only a slight increase in lysogeny frequency is observed when purified capsids are added to the mutant preparation. This suggests that capsids and tails of the MCP1_TP901-1::Ter mutant do connect during virion assembly but that the assembled structure is largely defective in achieving infection (see Discussion section). TP901-1erm mutations in genes tap (orf40) to gpT (orf44) were significantly increased in their frequency of lysogeny following the addition of purified tails compared to lysogeny by the purified samples on their own or mixed with purified capsids. In contrast, the frequency of lysogeny of mutant gpGfsT_TP901-1 did not increase following the addition of purified capsids or tails (discussed below).

Western Blot and Electron Microscopic Analysis of TP901-1erm Structural Mutants

In order to further assess the functionality of the mutant derivatives of TP901-1, and to ascertain if negative polar downstream effects were the cause of loss/reduction of infectivity of these mutants, a dual approach of Western blotting (employing antibodies raised against capsid and tail-associated proteins) and electron microscopic (EM) analysis of sucrose or CsCl gradient purified particles was applied (Figs 2 and 3, respectively). EM analysis of purified phage lysates of TP901-1erm mutants was performed to visualize structural assemblages. For each phage mutant, approximately 50 particles were observed to establish the reproducibility of the analysis. Through these observations, complete phage particles or separated head (DNA-filled or empty) and tail structures were discerned, depending on the mutant.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Immunological detection of TP901-1erm structural proteins precipitated from phage lysates by PEG8000.

Negative control represents PEG8000 precipitation of mitomycin C induced strain NZ9000-Cro_t712. The protein detected by the primary antibody is depicted at the right hand side of the image, with black-filled triangles highlighting the TP901-1 protein sizes and a black-unfilled triangle indicating the MCP1_Tuc2009 protein in the Tuc2009 control lane that is ~25 kDa smaller than MCP1_TP901-1. Western blots against MCP1_TP901-1 resulted in two non-specific bands, at approximately 50 and 40 kDa, that could not be removed during optimisation.

https://doi.org/10.1371/journal.pone.0131676.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. EM images of (A-N) TP901-1erm structural module mutants and (O) TP901-1erm control.

(A) Few tails, but no capsids, produced by mutant Portal_TP901-1::Ter. (B) Capsids, without attached tails, visualised for mutant MCP1_TP901-1::Ter. (C) Full, but scarce, phage particles produced following mutation MCP2_TP901-1::Ter. (D) Tails, but no capsids, produced for mutant Sfp_TP901-1::Ter. (E) Tails-only produced for mutant MHP_TP901-1::Ter. (F) Mature-resembling phages produced for mutant MCP3_TP901-1::Ter. (G) Unconnected capsids and tails observed in HTC1_TP901-1::Ter. (H) Distinct capsids and tails observed in HTC2_TP901-1::Ter. (I) Separated capsids and tails imaged from mutant Tap_TP901-1::Ter. (J) Capsids with long polytails isolated for mutant Ttp_TP901-1::Ter. (K) Capsids only are observed for mutant MTP_TP901-1::Ter. (L) Capsids and insoluble aggregates, proposed to be the TMP, observed in preparations of mutant gpG_TP901-1::Ter. (M) Capsids, but no tails, observed in preparations of mutant gpT_TP901-1::BamHI. (N) Mature-resembling phage produced by mutant gpGfsT_TP901-1. (O) Control TP901-1erm mature infectious phages. Refer to text for a more detailed description characterizing each mutant. Scale bars are 50 nm long.

https://doi.org/10.1371/journal.pone.0131676.g003

In the majority of cases, both Western blotting and EM analysis indicated that the introduction of capsid mutations did not affect the production of many downstream capsid-related functions, and also that tails were produced for the vast majority of cases. This would indicate that in most instances, polar effects were not the cause of loss/reduced infectivity of these mutants. However, there are some exceptions, which are discussed below, and in these cases, other capsid-related proteins appear not to be produced, while downstream-encoded tail structures are still being produced and assembled. It is possible that the capsid-related proteins are absent because they are required for complex formation with the mutated element and are therefore lost during the purification process,or their expression is negatively affected due to polar effects of the introduced mutation, such as those associated with translational coupling or translational fusion.

Immunoblot detection was performed on chloroform-treated PEG8000-mediated precipitates of induced TP901-1erm mutants to determine the role of the different proteins during virion assembly (Fig 2). Analysis of PEG-precipitated protein fractions was performed as; (i) direct analysis of phage lysates was not sensitive enough, (ii) analysis of PEG-precipitated phage lysates was predicted to be more accurate and reproducible than samples prepared by ultracentrifugation, (iii) PEG is anticipated to precipitate only large proteins or proteinaceous complexes such as those which form during phage assembly, and (iv) it was anticipated that some phage proteins and (small) protein assemblies would be lost during the preparation and centrifugal forces generated during purification procedures. However, while several potential protein-protein interactions are inferred during the analysis of PEG-precipitated lysates, due to the complex nature of phage assembly, these observations will require further experimental scrutiny to validate their roles in our proposed phage assembly model. Immunoblot detection of the portal protein of TP901-1 demonstrated it is detectable in all TP901-1erm mutants, and the purified phage controls, except for the lysate obtained from the TP901-1erm mutant that carries a non-sense mutation in the gene encoding the portal protein (Fig 2). This result validates the recombineering mutagenesis strategy, confirming that the insertion of an in-frame stop codon successfully prevents its production. As expected, EM images of purified fractions of Portal_TP901-1::Ter showed only tail structures produced, and no detectable capsid structures (Fig 3A). In addition to the absence of the portal protein in PEG-mediated precipitates of mutant Portal_TP901-1::Ter, proteins MCP1_TP901-1 and HTC1_TP901-1 are also undetectable in Western blots of Portal_TP901-1::Ter. Also, while not quantified, there appears to be a decreased amount of proteins Sfp_TP901-1 and MHP_TP901-1 in Western blots of mutant Portal_TP901-1::Ter. The decreased detection/loss of these proteins in immunoblots of mutant Portal_TP901-1::Ter indicates that they associate by specific protein-protein interactions during TP901-1 virion assembly, although this presumption will require experimental validation (see Discussion).

As expected, MCP1_TP901-1 was not observed in Western blots using purified preparations of mutant MCP1_TP901-1::Ter. In addition, no noticeable effect was caused by this mutation in any of the additional proteins tested by Western blot, indicating that downstream polar effects were not interfering with the translation of assayed proteins of other functions. However, EM images of MCP1_TP901-1::Ter showed DNA-packaged capsids only with no tails attached, suggesting that virion assembly or stability is affected by the absence of MCP1_TP901-1 (Fig 3B). Attempts to assign a function to MCP1_TP901-1 through bioinformatic analysis highlighted that the N-terminal portion of MCP1_TP901-1 and the complete MCP1_Tuc2009 protein share a very high level (94.3%) of aa similarity. However, MCP1_TP901-1 possesses an approximate 28 kDa C-terminal extension, or an additional 248 aa, which is lacking in MCP1_Tuc2009. The initial ~100 aa of this 248 aa extension of MCP1_TP901-1 appears to be due to a genetic duplication of a 5’-part of the MCP1_TP901-1-encoding gene (data not shown). The remaining C-terminally located 145 amino acids of MCP1_TP901-1 are lacking in its equivalent in Tuc2009, and no significant function can be ascribed to this C-terminal region based on database searches. The genes specifying MCP1_TP901-1 and MCP1_Tuc2009, as well as SPP1 capsid protein gp7 (gp7_SPP1) all occur immediately downstream of the gene encoding their corresponding portal protein, and have a strong positive charge (pI values of ~9.0), and are largely α-helical (data not shown). Furthermore, ClustalW alignment of MCP1_TP901-1, MCP1_Tuc2009 and gp7_SPP1 highlights sequence conservation, with the exception of the C-terminal extension of MCP1_TP901-1 (S1 Fig).

Intact virions were observed during EM analysis of mutant MCP2_TP901-1::Ter (Fig 3C); this was expected as the MCP2_TP901-1::Ter mutant phage was shown to be still capable of forming plaques and establishing lysogeny of host L. lactis strain 3107 (Fig 1 and Table 2, respectively). ORF35 of TP901-1 (designated here as Sfp_TP901-1) is annotated in this study as specifying a putative scaffolding protein (Sfp) based on the localization within the capsid-encoding structural module region and the predicted secondary structure of its corresponding protein product. Scaffolding proteins are involved in assisting the correct assembly of viral capsid structures by ensuring specific protein-protein interactions, nucleating the capsid assembly within cells, and preventing/excluding non-specific interactions between proteins during the molecular construction of phage capsids [35]. Sfp_TP901-1 is predominantly α-helical and part of its structure is predicted to be intrinsically disordered (S2 Fig), features commonly associated with phage scaffolding proteins [35]. The translation terminating mutation in Sfp_TP901-1 was shown to result in a decreased amount of the upstream-situated Portal_TP901-1 protein_, and very little MHP_TP901-1, detected in Western blots (Fig 2). This supports the prediction that Sfp_TP901-1 acts as a scaffolding protein, interacting with essential capsid components to coordinate correct assembly.

Immunoblot examination of mutant MHP_TP901-1 also revealed a decrease in the level of detected Portal_TP901-1 and a complete absence of Sfp_TP901-1, both of whose corresponding genes are situated upstream. The absence/reduction in capsid-related proteins in the PEG-precipitates of these mutants suggests that the proteins are produced, but are not assembled into the appropriate complexes. Consistent with this notion it was shown that tail structures, but no capsids, are present in Western blot and EM images of mutants Sfp_TP901-1::Ter and MHP_TP901-1::Ter (Figs 2, 3D and 3E, respectively).

Popovic et al. (2012) identified sequence identity between MCP3_TP901-1 and phage λ capsid chaperone protein gpFI, which facilitates the interaction of DNA-terminase holoenzyme complexes with procapsids [86]. While Popovic and colleagues proposed MCP3_TP901-1 is associated with phage capsids through a -1 ribosomal frameshift at the 3’ end of MHP_TP901-1, a feature observed for several Listeria-infecting phages [86, 87], Western blots against MHP_TP901-1 did not reveal a second band. HHpred analysis of MCP3 of TP901-1, and the analogous MCP3 of Tuc2009, revealed homology to GP10 of a T7-like Podoviridae phage infecting Prochlorococcus (e-value of 1.5e^-12), and also to the major capsid protein gp5 of the well characterized phage HK97 (e-value of 5.8e^-05). MCP3_Tuc2009 exhibits similarity to a 52 aa region located at the C-terminus of HK97 gp5, corresponding to residues 332–384 of this viral procapsid protein. Amino acid Asn356 of gp5, conserved in MCP3_Tuc2009, has been shown to be involved in covalently bonding neighbouring HK97 capsid subunits via an isopeptide bond [88]. Covalent cross-linkage of capsid protein subunits has previously been observed for lactococcal phage r1t [89], though it has not been documented to date for phages TP901-1 or Tuc2009. Consistent with the literature and bioinformatic analyses, MCP3_TP901-1 is a non-essential protein that aids in efficient phage infectivity. This conclusion is supported by the results demonstrating mutants of this gene produce infective phages with a lower EOP that are still capable of establishing lysogeny (Fig 1 and Table 2, respectively), and full phages were observed in EM images (Fig 3F).

In addition to HTC1 being absent in mutant HTC1_TP901-1::Ter, as expected, this protein is also absent in PEG-precipitated lysates of mutants Portal_TP901-1::Ter, Sfp_TP901-1::Ter and MHP_TP901-1::Ter. EM images of HTC1_TP901-1::Ter produced loose capsids and tails (Fig 3G). Analysis of the CsCl gradient-purified band of mutant HTC2_TP901-1::Ter revealed, similar to what was observed for HTC1_TP901-1::Ter, unconnected capsids and tails (Fig 3H), confirming the role of these proteins being required for head-tail connection.

Although no significant similarities to other proteins were detected by means of database searches, ORF40 of TP901-1 is annotated in this study as a Tail-activator protein (Tap_TP901-1; see Discussion). DNA-packaged capsids and tail structures were visualized in the EM analysis of CsCl gradient-purified mutant Tap_TP901-1::Ter, however, these structures were not connected (Fig 3I). Interestingly, EM images of the mutant Ttp_TP901-1::Ter (ORF41), with a stop codon introduced into the gene putatively encoding the Tail terminator protein (Ttp), showed DNA-packaged capsids with long polytail structures (Fig 3J), possibly due to uncontrolled tail-tube polymerization, consistent with the absence of a Ttp.

Analysing the TP901-1erm mutants for the production of MTP showed that lysates of all presumed capsid mutants contained tail proteins, a clear indication that these mutations do not cause a negative polar effect on the RNA stability or translation of genes that were located downstream of such mutations (Fig 2). No MTP was detected in mutants of MTP_TP901-1, gpG_TP901-1 and gpT_TP901-1. However, MTP was detected in the TP901-1 mutant permanently fusing gpG and gpT production together (gpGfsT_TP901-1). EM images of mutant MTP_TP901-1::Ter resulted, as expected, in tail-less capsids (Fig 3K).

In addition to detecting MTP_TP901-1, Western blots were performed to detect the distal tail protein Dit_TP901-1 in mutant lysates, as it is known to be crucial for the formation of functional phage tails [9]. Interestingly, Dit_TP901-1 was undetectable in mutants MTP_TP901-1, gpG_TP901-1 and gpT_TP901-1, where MTP_TP901-1 was also absent (Fig 2), indicating that tail assembly was affected in these mutants.

The TAC proteins gpG and gpT of TP901-1, are expected to fulfil a similar role as phage λ gpG and gpT TAC proteins based on genetic synteny with lambdoid-like phages and the presence of a conserved ‘slippery’ sequence that is associated with gpGT production [49, 50]. Furthermore, a recent study focusing on the structural conservation and identification of TACs supported the annotation of gpG_TP901-1 since the conserved spiral structure that typifies these TACs was maintained in the TP901-1 gpG equivalent [90]. EM visualization of the TP901-1erm carrying the gpG_TP901-1::Ter revealed DNA-containing capsids and large disordered aggregates (Fig 3L). We propose that these aggregates, not present in any of the other analyzed phage preparations, are hydrophobic TMP molecules incapable of remaining in solution in the absence of the chaperone protein gpG. Analysis of mutant gpT_TP901-1::BamHI revealed only the presence of phage capsids (Fig 3M). Interestingly, apparently complete phage particles were observed for mutant gpGfsT_TP901-1 (Fig 3N). Despite having the appearance of wild-type TP901-1erm, mutant gpGfsT_TP901-1 is incapable of infection (as determined by its inability to form plaques and lysogens) indicating aberrant tail assembly. Understanding the role(s) of gpG and gpGT may require higher resolution images of this tail assembly mutant and a better understanding of the chaperone-TMP interaction.