Escherichia coli Rosetta™ (DE3) strain (Novagen) and TG1 strain (Gibson, 1984) were grown in Luria-Bertani (LB) medium at 37°C with shaking or on LB plates supplemented with 1.5% agar. L. lactis IL1403 strain (Chopin et al., 1984) was grown at 30°C in M17 broth (Oxoid) supplemented with 0.5% glucose (GM17) without shaking or on GM17-agar plates. Phage F13 (GenBank accession no. MG253653) propagation was performed on L. lactis strain IL1403 in liquid GM17 medium supplemented with 10 mM calcium chloride (GM17-CaCl₂) as previously described (Chmielewska-Jeznach et al., 2018).

The overproduction of the L. lactis phage F13 GP46 protein (GenBak accession no. ATW69819.1) was primarily attempted in E. coli using the IMPACT™ expression system (NEB). Yet, with this approach we were able to obtain only very low protein recovery levels, irrespectively of conditions assayed (induction temperature, IPTG and salt concentration, glycerol addition, etc.). To circumvent these problems, we cloned gene 46 (gp46) in the pET28a-SUMO-PIN vector (courtesy of M. Pastor). For this, gp46 was amplified from bacteriophage F13 lysate by PCR reaction using primers SF13for/SF13rev listed in Table 1. The resultant PCR product was digested with BamHI and XhoI (Fermentas) enzymes and ligated into BamHI and XhoI sites of the pET28a-SUMO-PIN vector creating the N-terminal fusion plasmid for the overproduction of the GP46_F13 protein (pET28:SUMO:gp46). Ligated molecules were transformed into the E. coli TG1 cells. Transformants were selected on LB-agar medium containing 50 μg/ml kanamycin at 37°C. The presence of the cloned fragment was verified by PCR using SUMOF13_F and SUMOF13_R primers (Table 1) and sequencing.

The recombinant plasmid pET28:SUMO:gp46 was introduced into E. coli Rosetta™ (DE3) cells. Freshly grown colonies were inoculated into 1 L Luria broth (LB) medium containing 50 μg/ml kanamycin and 10 μg/ml chloramphenicol at 28°C until optical density at 600 nm (OD₆₀₀) reached 0.7. Expression of the target protein was induced by adding isopropyl-b-D-thiogalactopyranoside (IPTG) to a final concentration of 0.4 mM and the culture was incubated further on an orbital shaker at 150 rpm for 24 h at 16°C. Next, cells were harvested by centrifugation at 5,000 rpm for 15 min at 4°C. All subsequent purification steps were carried out at 4°C. Bacterial pellet was resuspended in 30 ml of optimized suspension buffer (50 mM Tris–HCl pH 8, 500 mM NaCl, 20 mM imidazole, 2% Triton-X100) and lysed by sonication on ultrasonic cell disruptor XL (MISONIX) 10 times for 15 s with 15 s intervals on ice. Without addition of Triton-X100 the protein after sonication was in the membrane fraction. With low salt in the suspension buffer, the protein recovery was very low. After sonication, the cell debris was then removed by centrifugation at 11,000 rpm for 1 h at 4°C. The supernatant was applied on a HisTrap FF crude column pre-packed with the Ni-Sepharose 6 Fast Flow affinity medium (GE Healthcare) and equilibrated with column buffer (50 mM Tris–HCl pH 8, 500 mM NaCl, 20 mM imidazole). The resin-bound protein was rinsed with 50 ml of column buffer and washed out with elution buffer (50 mM Tris–HCl pH 8, 500 mM NaCl, 300 mM imidazole). The eluate was digested by SUMO protease at 4°C overnight to remove SUMO-tag and then analyzed for the presence of GP46_F13 by SDS-PAGE. Protein concentration was determined using the Bradford assay on a UV spectrophotometer (Shimadzu).

Further protein purification was performed by FPLC (Fast Protein Liquid Chromatography) on the ÄKTA Purifier using the HiTrap^® Heparin HP column (GE Healthcare) equilibrated with 50 mM Tris-HCl buffer (pH 8). Before applying on the column, the elution buffer was exchanged to a buffer containing 50 mM Tris–HCl pH 8, 100 mM NaCl, by dialysis. The GP46_F13 protein sample was loaded on the column at a flow rate of 0.4 ml/min. Protein elution was carried out using a growing gradient concentration of NaCl (from 100 mM to 1 M) in 50 mM Tris–HCl pH 8 buffer. The peak fractions were analyzed by SDS-PAGE and stored after adding 50% glycerol (v/v) at −20°C for further experiments.

The DNA-binding activity of GP46_F13 was evaluated by electrophoretic mobility-shift assay (EMSA) with fluorescently labeled oligonucleotides (Cy5 on 5′ end) differing in size (22, 45, 68, 110 nt). The dsDNA 110 nt substrate was prepared by PCR reaction using L. lactis c2 phage genome as a template and a pair of complementary primers (ori110_F/ori110_R). The binding reactions were carried out for 30 min at 30°C in binding buffer (20 mM Tris–HCl pH 8, 50 mM NaCl, 1 mM DTT, 1 mM EDTA, 0.1 mg/ml of BSA, 2.5% glycerol) in 20 μl reaction volumes containing 1 nM of appropriate DNA substrate and the indicated amounts of analyzed protein (0.1–0.75 μM). Reaction samples were loaded onto 7% non-denaturing polyacrylamide gels and resolved for 40 min at 100 V in 0.5 × TBE buffer at 4°C with 1-h pre-run. Gels were developed using FluorChemQ MultiImageIII ChemiImager (Alpha Innotech, San Leandro, CA, United States) and the images were captured using Alpha View software (Alpha Innotech, San Leandro, CA, United States).

Helicase activity of the GP46_F13 protein was assessed on a variety of DNA substrates (Table 1). Asymmetric or forked DNA duplexes were prepared by annealing a 5′ Cy5-labeled oligonucleotides to a series of partially complementary unlabeled ssDNAs (as described by Curti et al., 2007). The helicase activity was assayed by mixing the protein with 1.5 nM of Cy5-labeled DNA substrate in a standard reaction buffer containing (if not stated otherwise) 50 mM Tris-HCl, pH 8, 50 mM NaCl, 5 mM MgCl₂ at 20 μl of final volume. Reactions were initiated by the addition of 5 mM ATP (or other NTP/dNTP), and the mixtures were incubated at 30°C for 30 min (if not otherwise stated). Reactions were terminated by the addition of 6 μl of stop solution (10% glycerol, 0.4% SDS, 50 mM EDTA). To minimize reannealing of the unwound oligonucleotides, a 10-fold molar excess of unlabeled DNA trap, corresponding to the Cy5-labeled strand was added after 10 min of the reaction. As a negative control, the substrates were incubated in the reaction mix in the absence of the protein. Samples were resolved in a 15% polyacrylamide gel in 0.5 × TBE buffer for 40 min at 18°C with a 30-min pre-run. For data analysis, gels were developed in FluorChemQ MultiImageIII ChemiImager (Alpha Innotech, San Leandro, CA, United States) and the images were captured using Alpha View software (Alpha Innotech, San Leandro, CA, United States). The unwound DNA was quantified using ImageJ (Image Processing and Analyzing In Java).

Size exclusion chromatography (SEC) coupled with multi-angle light-scattering (MALS) was done using a high-performance liquid chromatography (HPLC) equipment (1260 Infinity LC, Agilent Technologies Inc., Santa Clara, CA, United States) connected to a UV detector, a MALS detector (DAWN HELEOS II, Wyatt Technology Santa Barbara, CA, United States), and a differential refractometer (Optilab T-rEX, Wyatt Technology, Santa Barbara, CA, United States). The GP46_F13 sample (100 μl) purified as described earlier was injected at 1.5 mg/ml final concentration onto a Superdex 200 Increase 10/300 column (GE Healthcare, Milwaukee, WI, United States) and run at room temperature and a 0.5 ml/min flow rate in buffer containing 50 mM Tris–HCl pH 8, 500 mM NaCl, 20 mM imidazole. Output data were analyzed using ASTRA v. 6 software (Wyatt Technology, Santa Barbara, CA, United States).

Homologs of the GP46_F13 protein were collected using PSI-BLAST (Altschul et al., 1997) search (three iterations, inclusion threshold 0.05) against non-redundant (nr) database with GP46_F13 protein sequence (GenBank accession no. ATW69819.1) as a query. Sequence similarity to proteins with known structures and other Pfam families was scanned with HHpred at MPI Bioinformatics Toolkit (Zimmermann et al., 2018). All multiple sequence alignments were calculated using Mafft (L-INS-i) (Katoh et al., 2002). GP46_F13 structure modeling was done using AlphaFold2 (Jumper et al., 2021). Multiple structure alignments between identified homologs of known structure and GP46_F13 model were prepared manually.

Sequences of proteins belonging to other helicase families identified by HHpred were collected using PSI-BLAST searches starting with one representative sequence for each family. All the sequences, including GP46_F13 homologs, were altogether clustered to 65% of sequence identity using CD-HIT (Li et al., 2001) to reduce redundancy. The dataset was further clustered using CLANS (Frickey and Lupas, 2004).

The GP46 protein of L. lactis phage F13 (GP46_F13) according to HHpred mappings displays high sequence similarity (all hits with scores > 90 and estimated probability of being true positive > 99%) to several P-loop NTPase protein families, including RepA, RadA, RecA, GvpD, DnaB, RAD51, KaiC, and AAA_25 (Figure 1 and Supplementary Table 1). RepA and DnaB are replicative helicases (Niedenzu et al., 2001; Bailey et al., 2007), RadA is a DnaB-type helicase interacting with RecA in homologous recombination (Marie et al., 2017), RAD51 and RecA are recombinases forming helical filaments seeking for homology in homologous recombination (Chen et al., 2008; Short et al., 2016), GvpD regulates gas vesicle formation in Archaea, KaiC is an ATPase controlling cyanobacterial circadian clock (Abe et al., 2015) and AAA_25 proteins have no known function described yet. Due to the functional diversity of homologous protein families and taking into consideration sequence and structural similarities between them we performed a more detailed sequence clustering in order to provide a more reliable classification and annotation. Obtained results suggest that GP46_F13 is closely related to the RepA family. It retains all sequence motifs characteristic for NTP hydrolysis, including residues Lys49 from Walker A (P-loop) motif, Glu82 (motif 1a), Asp136 (Walker B), and His177 (motif 3) (Figure 2A). It also has an arginine finger (Arg233) which might suggest a similar catalytic mechanism. Interestingly, the AAA_25 family identified among GP46_F13 homologs also displays RecA-like features and will probably not function as an AAA ATPase (Figure 2A).

Molecular modeling of GP46_F13 confirmed conservation of all structural elements essential for helicase function. The model computed by AlphaFold2 (pLDDT score 83.9; scores above 70 are regarded as confident) demonstrates no major differences compared to the RepA structure (Figure 2B) which suggest that GP46_F13 may also function as hexameric helicase. Despite the similarity of GP46_F13 to RepA, it retained the arginine finger in its canonical position at strand 5b, while in RepA it is located on the proximal strand 5 from where it executes its catalytic function.

To obtain sufficient quantities of GP46_F13 protein for in vitro studies the pET28a:SUMO-PIN expression system was used. For this, the phage F13 gp46 open reading frame encoding the 298-AA product was amplified by PCR and fused at its N-terminal end with 6xHis tag and SUMO protein-encoding sequences of the pET28a:SUMO-PIN vector in E. coli. The GP46_F13 6xHis- and SUMO-tagged protein was purified by nickel-affinity chromatography, and after removal of the SUMO tag, further purification was done using the heparin-sepharose column (Supplementary Figure 1). SDS-PAGE analysis of the final purification protein revealed a major protein band with a molecular weight of approximately 34 kDa, which corresponds to the molecular weight calculated from the amino acid sequence (Supplementary Figure 1).

In order to elucidate the binding of GP46_F13 to different DNA substrates in vitro EMSA assays were performed. For this purpose 1 nM of 5′ fluorescently labeled DNA substrates (Table 1) were incubated with varying GP46_F13 protein concentrations.

As shown in Figure 3, GP46_F13 interacts with ssDNA substrates of different length (22, 45, 68, and 110 nt) with nanomolar affinity and the relative amounts of unshifted ssDNA decrease proportionally to protein concentrations added. Clear retardation complexes could be seen for the 110-nt ssDNA (ss110). For shorter (ss22, ss45, ss68) substrates, a formation of discrete protein-DNA aggregates in the wells of the gel were observed. Although we see a fraction of the ssDNA substrate stuck in the well, most probably associated with the tendency of the protein to aggregate in vitro, our results indicate that GP46_F13 can interact with DNA. At the same time, GP46_F13 did not bind to dsDNA; no nucleoprotein complexes were observed with the 110-bp dsDNA (ds110) vs. ss110 substrate (Figure 3). To examine whether GP46_F13 has affinity for substrates mimicking the replication forks, we performed EMSA assays using a branched 30-bp duplex with a short 15-nt 3′ tail and an extended, 35-nt 5′ ssDNA overhang (30duplex_5′ext). We show that the GP46_F13 gives clear protein-DNA complexes with 30duplex_5’ext, suggesting that GP46_F13 can bind also to heteroduplex structures.

Results of our comparative analyses in silico suggested that GP46_F13 is conserved in key amino acids and has similar fold structure to DNA helicases, including RepA from plasmid RSF1010 and phage T7 gp4 protein. To evaluate whether GP46_F13 exhibits also a similar DNA unwinding ability, we performed helicase activity assays under various conditions. For these we used a series of asymmetric substrates with an ssDNA overhang from either 5′ or 3′ end or forked substrates with two free ssDNA ends of different length (Table 1). We determined that the GP46_F13 protein was not active on blunt-ended dsDNA substrates (results not shown). Instead, we observed a preference for an asymmetric DNA substrate with 5′ tail (30duplex_5′ss), but not with a 3′ overhang (30duplex_3′ss), indicative of a 5′–3′ directionality of the protein’s unwinding activity (Figure 4).

As DNA helicases can act at replication forks by breaking down the hydrogen bonds and separating the strands of the duplex DNA, we examined whether GP46_F13 is capable of unwinding branched duplexes that mimic such structure. For this we used three types of substrates varying in the length of the dsDNA region and the free 3′ and 5′ ss ends (Table 1). Results showed clearly that extension (from 15 nt to 35 nt) of the 5′ end increases the rate of the separation of duplex DNA strands (Figures 5A–C). A 15-nt long 5′ ssDNA overhang was sufficient enough for GP46_F13 unwinding of the forked substrate. We also showed that the protein is capable of displacing longer (53 bp) stretches of duplex DNA with 5′ tail (Figure 5C). Yet, liberation of the labeled oligonucleotide product occurred at slightly higher protein:DNA molar ratio for the 53duplex_5′ext vs. 30duplex_5′ext substrate (Figures 5A,C). To this end, the 30duplex_5′ext was recognized as the most preferable substrate and was used in further assays.

To determine the efficiency of the reaction of GP46_F13 on the substrate DNA, we conducted a time-dependent DNA unwinding assay (Figure 6A). Results showed that the extent of displaced oligonucleotide increased in time and after 30 min it was almost completely liberated from the duplex DNA. Optimal temperature and pH conditions for DNA unwinding were determined to be at 30°C and pH 8 (Figure 6B). Under these conditions the strands of the duplex DNA were completely separated. As the unwinding activity of DNA helicase proteins is known to be fueled by various nucleotide substrates, we tested the specific requirement of GP46_F13 for NTPs or dNTPs (Figure 6C). Results of the assay demonstrated efficient separation of duplex strands in a 30-min reaction in the presence of 5 mM ATP. Moderate displacement of the labeled oligonucleotide was detected for dATP and dTTP at the same concentrations, while no unwinding activity was observed for the remaining nucleotide substrates. We thus concluded that the unwinding activity of GP46_F13 is ATP-dependent, while the binding to DNA can occur in the absence of ATP. Next, we examined the influence of rising ATP concentrations on the unwinding activity of GP46_F13 (Figure 6D). Complete unwinding of the forked DNA substrate was observed at 5 mM ATP under the tested conditions, indicating that this is the optimal concentration of the supplied nucleotides. Lower and higher ATP concentrations allowed only for partial separation of duplex DNA strands. The activity of ATP-dependent helicases relies strongly on the presence of metal cofactors, particularly Mg²⁺ ions. The ratio of ATP to Mg²⁺ was documented previously to influence the efficiency of helicase activity. Therefore, we also investigated the effects of different concentrations of Mg²⁺ under constant ATP level on the unwinding activity of GP46_F13. The most efficient separation of duplex DNA was determined when Mg²⁺ and ATP were supplied at an equimolar (5 mM) concentration (Figure 6E).

DNA helicases generally function as multimeric forms (dimers or hexamers) (West, 1996). In order to infer the oligomerization status of GP46_F13, we performed SEC-MALS analysis. For purification of GP46_F13, we used in all the buffers high salt concentration (500 mM NaCl) to increase solubility and stability of the protein (see Section “Materials and Methods”). In turn, formation of oligomeric forms is generally promoted at low salt concentrations (50–200 mM). Thus, we initially performed several SEC runs at low salt conditions (50 and 200 mM). Unfortunately, in these cases the protein was eluted as a large multimeric form in the void volume of the SEC column (Supplementary Figure 2). To stabilize the protein, we increased the salt concentration to 500 mM and obtained a major single peak. By DLS assays we checked that addition of cofactors (DNA, ATP, and Mg²⁺) does not influence GP46_F13 oligomerization (data not shown). In the light of these data, SEC-MALS analysis was performed at restored higher salt conditions without cofactors and revealed several GP46_F13 populations–a monomer and two oligomeric forms (Figure 7). The main fraction had a mass value of 65 kDa, the second peak corresponded to 39 kDa fraction, while the smallest peak was dispersed with a mean mass of 410 kDa; yet, the signal detecting the largest peak was too weak for accurate size determination. The values detected were similar to the theoretical molecular weight (MW) of a dimer (2 × 34.25 = 68.5 kDa), a monomer (34.25 kDa), and possibly a 12-mer (12 × 34.25 = 411 kDa). Results of the assay indicate that GP46_F13 can self-associate in solution. The presence of the monomeric fraction indicates that under the tested conditions the multimeric protein structure is not stable since high salt concentrations prompt possible dissociation. On the other hand, it is surprising to observe multimeric forms in a buffer containing 500 mM NaCl. This proves a strong GP46_F13 tendency for oligomerization and suggests assembly of GP46_F13 into higher multimeric forms.

To probe for potential phage or host protein partners, we performed pull-down assays with tagged GP46_F13 as bait, followed by LC-MS and computational analysis. We detected high distribution of peptides corresponding to GP46_F13 and to three other non-structural phage proteins, i.e., GP23 (GenBank accession no.: ATW69796.1), GP28 (GenBank accession no.: ATW69801.1), and GP32 (GenBank accession no.: ATW69805.1), suggesting their interaction with the affinity-bound GP46_F13 (Supplementary Table 2). By BlastP and HHPred searches we found that GP28 carries a Zn²⁺ binding domain, GP32 contains a conserved DUF3310 (domain of unknown function), and GP23 has no significant hits. Thus, the explicit function of all three detected proteins and their role in phage replication remains hypothetical. By LC-MS we also detected significant amounts of peptides representing host L. lactis IL1403 proteins—DnaK (GenBank accession no.: AAK05052.1), DnaJ (GenBank accession no.: AAK06322.1), topoisomerase I (GenBank accession no.: AAK05328.1), and SSB (GenBank accession no.: AAK06288.1). All of these proteins have previously been shown to be engaged in DNA replication. Affinity interactions of these proteins with the tagged GP46_F13 in our pull down assay are a strong suggestion that they are partners in a common biological process. Their specific interaction with GP46_F13 demands further experimental confirmation.