Receptor-Binding Protein of Lactococcus lactis Phages: Identification and Characterization of the Saccharide Receptor-Binding Site
http://www.100md.com
《细菌学杂志》
Architecture et Fonction des Macromolecules Biologiques, UMR 6098 CNRS and Universites d'Aix-Marseille I & II, Campus de Luminy, case 932, 13288 Marseille CEDEX 09, France,Groupe de Recherche en ecologie Buccale (GREB), Faculte de Medecine Dentaire,Felix d'Herelle Reference Center for Bacterial Viruses,Departement de Biochimie et de Microbiologie, Faculte des Sciences et de Genie, Universite Laval, Quebec City, Quebec, Canada, G1K 7P4
ABSTRACT
Phage p2, a member of the lactococcal 936 phage species, infects Lactococcus lactis strains by binding initially to specific carbohydrate receptors using its receptor-binding protein (RBP). The structures of p2 RBP, a homotrimeric protein composed of three domains, and of its complex with a neutralizing llama VH domain (VHH5) have been determined (S. Spinelli, A. Desmyter, C. T. Verrips, H. J. de Haard, S. Moineau, and C. Cambillau, Nat. Struct. Mol. Biol. 13:85-89, 2006). Here, we show that VHH5 was able to neutralize 12 of 50 lactococcal phages belonging to the 936 species. Moreover, escape phage mutants no longer neutralized by VHH5 were isolated from 11 of these phages. All of the mutations (but one) cluster in the RBP/VHH5 interaction surface that delineates the receptor-binding area. A glycerol molecule, observed in the 1.7- resolution structure of RBP, was found to bind tightly (Kd = 0.26 μM) in a crevice located in this area. Other saccharides bind RBP with comparable high affinity. These data prove the saccharidic nature of the bacterial receptor recognized by phage p2 and identify the position of its binding site in the RBP head domain.
INTRODUCTION
Bacteriophage infection is a major problem impairing any industrial fermentation that relies on bacteria to transform a substrate into fermented products or specific molecules. Several strains of the gram-positive bacterium Lactococcus lactis are used quite extensively worldwide for the manufacture of fermented milk products. For decades, the dairy industry has been dealing with phage infections of their L. lactis strains because virulent phages are ubiquitous within their environment (28).
Phages of L. lactis are classified within several groups (21), although those commonly found in dairy plants belong to three different species, i.e., 936, P335, and c2 (18, 27, 29). From a taxonomical standpoint, these phages are all members of the Siphoviridae family (order Caudovirales) that also includes the coliphage lambda. Siphophages are characterized by a double-stranded DNA genome and a long noncontractile tail. The phages belonging to the species 936 and P335 have a small isometric capsid, while those from the c2 species have a prolate capsid.
The bacteriophage infection process starts by the specific recognition between the phage receptor-binding protein (RBP) located at the tip of the tail and the receptor distributed over the host cell surface (14, 35). Bacterial receptors have been well studied in gram-negative bacteria, particularly in Escherichia coli, while similar information lags behind in gram-positive bacteria (12). It has been previously shown that phages infecting some L. lactis strains adsorb initially to the cell wall surface and likely to various carbohydrates containing rhamnose, glucose, or galactose (for a review, see Forde and Fitzgerald [15]). For many phages, this binding step is reversible, and phages of the c2 species, for example, require a second irreversible binding step to a predicted membrane-attached protein (PIP) of 901 amino acids. However, phages of the species 936 and P335 do not use PIP as a secondary receptor (14). A better understanding at a molecular level of these phage-host interactions in gram-positive bacteria should lead to better tools for the control of phage infection in various biotechnological processes as well as reveal insights into the molecular basis of host specificity.
A novel antiphage strategy was recently designed based on the use of a phage-neutralizing heavy chain antibody fragment (VHH5) obtained from llama (31). VHH5 produced at large scale in the food microorganism Saccharomyces cerevisiae prevented the infection of an L. lactis strain by phage p2 (936 species) during the manufacture of cheese (23). VHH5 recognized the RBP of phage p2, a 30-kDa protein (ORF18) located at the distal part of the phage tail, with a high affinity (Kd = 1.4 nM) (8).
We previously determined the crystal structure of the p2 RBP trimer at 2.3- resolution and that of its complex with three molecules of VHH5 at 2.7- resolution (34). The RBP ternary complex is formed of three subunits, each assembling three symmetry-related domains: the shoulders, the interlaced neck, and the heads, from N to C terminus (Fig. 1A). The shoulder domains, encompassing residues 1 to 141, have a -sandwich fold formed of two -sheets of four antiparallel -strands each. The neck (residues 142 to 163) folds as a triple-stranded interlaced -helix that forms an equilateral triangular prism around the threefold axis (Fig. 1A). The receptor recognition domain (residues 164 to 264), designated as the head, is a -barrel formed of seven antiparallel -strands. Each domain is parallel to the threefold axis and interacts with the two others, forming a very compact structure (Fig. 1A). In the structure of the head trimer complexed with VHH5, each of the three VHH5 domains binds the RBP head at the top half of the trimer and forms a regular trimeric complex, sharing the same threefold axis as the isolated RBP (Fig. 1B).
Here, we further investigated the molecular interactions between various RBPs of Lactococcus lactis phages and VHH5. First, the broadness of the neutralization conferred by VHH5 was tested against 50 distinct lactococcal phages of the 936 species. Then, phage mutants escaping the VHH5 neutralization were isolated, and their mutated RBPs were analyzed by surface plasmon resonance (SPR). We extended the resolution of the p2 RBP structure to 1.7- resolution and localized glycerol molecules in the RBP head, within the VHH5 binding area. We performed fluorescence quenching experiments with saccharides to demonstrate the existence of a high-affinity saccharide binding site overlapping with the structural glycerol attachment site.
MATERIALS AND METHODS
Bacterial strains, phages, and host range. The phages and hosts used in this study are listed in Table 1. L. lactis strains were grown at 30°C in M17 broth (Quelab) supplemented with 0.5% glucose (GM17). For phage propagation, GM17 was supplemented with 10 mM CaCl2 as described previously (27). The neutralization assays were performed in BCP broth (2% tryptone, 0.5% yeast extract, 0.5% glucose, 0.4% NaCl, 0.15% Na-acetate, 40 mg of bromocresol purple per liter [2]).
Native and mutant RBPs, VHH5, and head domain production. Phage p2 wild-type RBP and llama VHH5 were cloned, expressed, and purified as described previously (8, 23). Single-amino-acid mutations of RBP were generated using a QuickChange site-directed mutagenesis kit (Stratagene). Briefly, the pET28a/ORF18 (8, 23) was entirely amplified by PfuTurbo DNA polymerase using two complementary mutated primers: the Thr 165 codon was mutated to encode Met using the primers T165M-F (5'-TCGATGTTCCAGTTCAAATGTTGACAGTTGAAGCTGG-3') and T165M-R (5'-CCAGCTTCAACTGTCAACATTTGAACTGGAACATCGA-3'); Trp 201 was mutated to encode Arg using the primers W201R-F (5'-GTGTCAAATATACAAAAAGGCCGGAATATGTCTGGAACATGGG-3') and W201R-R (5'-CCCATGTTCCAGACATATTCCGGCC TTTTTGTATATTTGACAC-3'); Val 217 was mutated to encode Ile using the primers V217I-F (5'-GACCATTTCGTCCAGCTGCTATTCAAAGTCTTGTTGGTCATTTTG-3') and V217I-R (5'-CAAAATGACCAACAAGACTTTGAATAGCAGCTGGACGAAATGGTC-3'); and Asn 236 was mutated to encode Lys using the primers N236K-F (5'-GATACTTCTTTCCATATTGATATAAAGCCAAATGGTAGTATTACTTGGT-3') and N236K-R (5'-ACCAAGTAATACTACCATTTGGCTTTATATCAATATGGAAAGAAGTATC-3') (bold nucleotides mark mutational changes). Parental plasmid was digested by DpnI, and the vector containing the mutation was transformed in supercompetent XL1-Blue cells. Transformants were selected onto LB ampicillin (100 μg/ml) agar plates. Mutations were verified by automated DNA sequencing (Genome Express, France). The DNA fragments encoding the phage p2 RBP head segment (residues 164 to 264) were amplified by PCR from the pET28a/ORF18 plasmid (8, 23) using the following primers: forward primer, 5'-ACAAGTTTGTACAAAAAGCAGGCTtagaaggagatagaaccatgaaagttcaaacgttgacagttgaagctgg-3' (attB1 sequence in uppercase, initial atg codon in bold, and the ribosome binding site and the first part of the open reading frame in lowercase); reverse primer, 5'-ACCACTTTGTACAAGAAAGCTGGGTcTTAGTGATGGTGATGGTGATGtttaatgaagtaacttccgttacc-3' (attB2 sequence in uppercase, stop codon in bold, six-His tag in italic, and the last part of the open reading frame in lowercase). The PCR product was subcloned in the pDest14 vector by recombination (36) using Gateway technology (Invitrogen). Expression of mutant proteins and the RBP head was carried out using the BL21(DE3) and Rosetta(DE3)pLysS E. coli strains in Terrific broth (Gibco) and 2YT supplemented with antibiotics, respectively. When the optical density at 600 nm reached 0.5, expression was induced with 0.5 mM IPTG (isopropyl--D-thiogalactopyranoside), and the temperature was decreased to 15°C and 17°C for mutants and head, respectively. All proteins produced were soluble and were purified using a two-step procedure on a Ni column, taking advantage of the presence of a six-His tag at the C terminus, followed by gel filtration on a Superdex 200 column in 1.8 mM KH2PO4, 10.1 mM Na2HPO4, pH 7.2, 2.7 mM KCl, 137 mM NaCl. For each protein, productions yielded between 20 and 50 mg of pure protein per liter of culture.
Neutralization assay. The efficacy of the VHH5 antibody in neutralizing various lactococcal phages was determined by a lactic acid production assay. Briefly, 104 phages and 1 μg/ml of VHH5 antibody were added to 3 ml of BCP medium supplemented with 10 mM CaCl2. The mixture was incubated for 1 h at 30°C. Then, the appropriate L. lactis host was inoculated at 1% and incubated for an additional 5 h at 30°C. This neutralization assay also included a control that consisted of the BCP medium inoculated with only the L. lactis host and another control in which only the host and the phages were added into the BCP medium. After the incubation period, the lactic acid production by L. lactis strains (without phages) resulted in a color change (from purple to yellow) of the medium containing a pH indicator (bromocresol purple). Conversely, the presence of phages led to cell lysis and limited acid production, and thus the color of the medium remained purple. If the antibody properly neutralized the phage, L. lactis was able to produce lactic acid and the pH indicator turned yellow. This assay was performed in two independent experiments for all of the 50 lactococcal phages tested.
Isolation of nonneutralized phage mutants. For all lactococcal phages neutralized by VHH5, the isolation of nonneutralized phage mutants was carried out. Following the standard neutralization assay, the cultures were centrifuged, and the supernatant was filtered on a 0.45-μm Acrodisc. The phage titer was determined (27), and a second neutralization assay was performed using 104 phages from the above filtrate. Three to five neutralization assays were needed to obtain nonneutralized phages in which no color change occurred in the presence of VHH5. Phages were then plated, and individual plaques were isolated and retested for neutralization by VHH5. Phage adsorption tests were carried out for wild-type and mutant phages as reported elsewhere (12). Adsorption assays were performed in three independent experiments.
DNA manipulation and fingerprinting analysis. Phage DNA was isolated using a Lambda mini kit (QIAGEN) with the following modifications. The phage pellet was resuspended in buffer L3 containing 1 mg/ml of proteinase K and then incubated for 30 min at 65°C. In the final step, the DNA pellet was dissolved in 50 μl of 10 mM Tris (pH 8.0). DNA was digested with EcoRI and EcoRV as recommended by the manufacturer (Roche Diagnostics). Restriction fragments were heat-treated at 65°C for 10 min, electrophoresed on agarose gel (0.7%) in Tris-acetate-EDTA buffer, stained with ethidium bromide, and visualized under UV light. Phage restriction fragment length polymorphism (RFLP) patterns were analyzed with Molecular Analyst Fingerprinting Plus version 1.6 software (Bio-Rad), using the unweighted-pair group method using average linkages.
DNA sequencing and analysis. The genes coding for the receptor-binding protein of wild-type lactococcal phages and nonneutralized phage mutants were sequenced on both strands by primer walking using synthetic oligonucleotide primers (Invitrogen) and phage genomic DNA as a template. DNA sequencing was performed using an ABI Prism 3100 apparatus from the service center at the Universite Laval. Computer-assisted DNA and protein analyses were performed using the Genetics Computer Group Sequence Analysis software package version 10.3 (11). The multiple sequence alignment was achieved with ClustalW, available at EMBL-EBI (http://www.ebi.ac.uk/clustalw/). For a few phages (712, bIL170, P008, and sk1), the RBP gene was already available in GenBank but was still confirmed by sequencing. Phylogenetic and molecular evolutionary analyses were conducted using MEGA version 3.0 (20). The bootstrap test for phylogeny was used and replicated 2,000 times.
Crystallographic study of RBP. Crystallization of the RBP was described elsewhere (34). The 1.7- data set was collected at beam line ID14-4 (ESRF, Grenoble). Data were integrated and reduced using MOSFLM and SCALA (6). Refinement was performed with REFMAC5 (30) alternated with rebuilding using Turbo-Frodo (33), using the 2.3- RBP structure as a starting point (2BSD). Statistics are presented in Table 2. The high resolution and the low R and R-free values indicate the excellent quality of the structure.
Surface plasmon resonance. All surface plasmon resonance experiments were performed using a Biacore 1000 instrument (Biacore Inc., Piscataway, NJ) at 25°C. The buffer for the continuous-flow pump was HBS-P (10 mM HEPES, pH 7.4, 150 mM NaCl, 0.005% [vol/vol] surfactant P-20) supplemented with 0.05 mM EDTA; the buffer for the sample pump was HBS-EP (10 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.005% [vol/vol] surfactant P-20). The buffers were degassed before use. The general strategy followed in the Biacore experiment was to take advantage of the presence of a His tag in the native and mutant RBPs and the absence of a His tag in the VHH5 molecule. The chip Ni-nitrilotriacetic acid (Ni-NTA) was first saturated with Ni2+ by washing it with 0.5 μM NiCl2 (20 μl at 20 μl/min). Then, the RBP solution (at 190 nM, 40 μl at 10 μl/min) was passed, followed by the VHH5 solution (100 nM, 40 μl at 10 μl/min). Regeneration was achieved by washing the flow cell with 350 mM EDTA (20 μl at 20 μl/min). Step 1 was run with the command Inject, followed by Extraclean; steps 2 and 3 were run with the command Kinject; step 4 was run with the command Quickinject, followed by Extraclean. The four-step procedure described above was performed identically for the wild-type RBP, the mutants (Thr165Met, Trp201Arg, Val217Ile, and Asn236Lys), and the control injection of running buffer instead of VHH5.
Fluorescence quenching experiments. Fluorescence experiments were carried out on a Varian Eclipse spectrofluorimeter using a quartz cuvette in a right-angle configuration; the light path was 0.4 and 1 cm for the excitation and emission, respectively. The interaction of RBP with saccharides was monitored by recording the quenching of the intrinsic protein fluorescence upon addition of ligand aliquots. The excitation wavelength was 290 nm, and emission spectra were recorded in the range of 320 to 400 nm. The excitation slit was 5 nm while the emission slit was 10 nm for a protein concentration of 1 μM. A moving-average smoothing procedure was applied, with a window of 3 nm. Titrations were carried out at room temperature with 1 μM protein in 10 mM phosphate buffer Na/Na2, 50 mM NaCl, pH 7.5. The fluorescence intensities at the maximum of emission (346 nm) for different concentrations of quencher were corrected for the buffer contribution before plotting and further analysis. The affinity was estimated by plotting the decrease of fluorescence intensity at the emission maximum as 100 – (Ii – Imin)/(I0 – Imin) x 100 against the quencher concentration; I0 is the maximum of fluorescence intensity of the protein alone, Ii is the fluorescence intensity after the addition of quencher (i), and Imin is the fluorescence intensity at saturating concentration of quencher. The Kd values were estimated using Prism 3.02 (GraphPad Software, Inc.) by nonlinear regression for a single binding site with the equation Y = Bmax x [X/(Kd + X)], where Bmax is the maximal binding and Kd is the concentration of ligand required to reach half-maximal binding.
Nucleotide sequence accession numbers. All RBP gene sequences were deposited in the GenBank database (see Table 1 for accession numbers).
Protein structure accession number. The coordinates and structure factors for the RBP have been deposited in the Protein Data Bank of the Research Collaboratory for Structural Bioinformatics (http://www.rcsb.org/pdb/) as entry 1ZRU.
RESULTS
Neutralization of 936-related phages by VHH5. It was previously shown that VHH5 binds to the RBP of the lactococcal phage p2, a member of the 936 group (8, 23). The addition of VHH5 to milk also prevents phage infection and fermentation failure during the manufacture of cheese (8, 23). In order to investigate the neutralizing potential of the VHH5 antibody, 50 lactococcal phages of the 936 group were selected and tested in a neutralization assay (Table 1). All of these phages were isolated from various geographic areas and are distinct, based on the analysis of their host range as well as the EcoRV and EcoRI restriction profiles (RFLP) of their genomes (data not shown). Of the 50 phages tested, 12 phages were neutralized by VHH5 (Table 1). The 38 other lactococcal phages were not neutralized by the antibody and were still able to infect their respective host cells (Table 1). These results indicate a significant diversity of the RBP epitope and suggest a limited efficiency of VHH5 for phage control in an industrial framework.
Isolation of nonneutralized phages. From an evolutionary standpoint, phages are known to adapt to their local environment. Thus, the isolation of phage mutants that were no longer neutralized by VHH5 was carried out. Repeated infection experiments of the 12 phages in the presence of VHH5 were performed and led to isolation of nonneutralized mutants derived for 11 of these 12 phages (Table 3). No nonneutralized mutants were isolated for phage HD12, while two different mutants were identified for phage p2. Three to five challenges were needed to obtain high-titer mutant phages. The escape phage mutants were no longer neutralized by VHH5 but had the same RFLP patterns as the neutralized wild-type phages (data not shown). These results suggest point mutations within the epitope specificity of VHH5 to the receptor-binding protein. Adsorption of the escape phage mutants to their hosts was also measured and found to be similar to the adsorption of the wild-type phages (data not shown). Altogether, these data show that most phages can rapidly evolve to evade recognition by VHH5 without affecting their host range.
Analysis of the RBP of lactococcal phages. In order to investigate the epitope recognition of VHH5, the RBP gene of each of the 50 wild-type lactococcal phages and the nonneutralized phage mutants was sequenced and analyzed (Fig. 1 and Table 3). A single mutation was found to be sufficient to avoid neutralization by the antibody fragment. In all mutants, the point mutation led to an amino acid change within the C-terminal region of the RBP. Overall, four different mutations were observed (Table 3).
The deduced RBPs from 50 phages range in size from 252 to 308 amino acid residues. The alignment of the multiple RBP sequences revealed that the N terminus is highly conserved, as the 50 shoulder domains (residues 1 to 141) shared 50.4% identity (71/141 amino acids). In fact, the first 120 amino acids were particularly conserved (56.7% identity), while the next 21 residues were less conserved. It was previously predicted that this conserved N-terminal part of the protein may be involved in protein-protein interactions with other phage tail proteins (12, 17). The last 21 residues are probably slightly less conserved to allow connection to the neck domain.
The neck (residues 142 to 163) and head (C-terminal) domains were more diverse, possibly reflecting the differences observed in host recognition (data not shown). However, in most cases, a specific neck domain was associated with a particular head domain. One could argue that the rigid neck structure serves as a linker between the larger shoulder and head domains. Nonetheless, the 50 RBPs could be classified in at least 10 groups based on their overall identities (Fig. 2). The 12 phages neutralized by VHH5 were classified in one large group of 18 RBPs sharing almost 80% identity. Comparative analysis of the 12 RBPs recognized by VHH5 and the 6 nonrecognized RBPs allowed the identification of a few amino acid discrepancies between both groups of RBP (data not shown). Finally, the other 32 phages have a distinct C-terminal RBP compared to the above group, explaining their nonneutralization by VHH5 and leading to the identification of 9 other groups of RBPs. Recently, Dupont et al. (13) reported six different types of RBP genes of lactococcal 936 species bacteriophages and developed a PCR assay to detect them. They suggested that the PCR assay could be used in cheese plants for phage detection. Here, we showed that there are many more RBP groups within this lactococcal phage species, and perhaps this variability may limit the usefulness of such a PCR assay.
Complex of the RBP with neutralizing llama VHH5. We have already reported that, overall, the water-accessible surface area of the RBP of phage p2 covered in the interaction with VHH5 is 825 2, with 647 2 and 188 2 for monomers A and B, respectively. The interaction surface involves many charged or semipolar residues, but nonpolar residues are also observed (Table 4). The VHH5 domain interacts with the RBP in a semilateral way, as observed in several camelid VHH complexes with haptens or proteins (9, 31). Facing the RBP surface, the VHH5 interacting surface is also rather polar and formed by a mixture of charged or semipolar residues (Table 4).
The isolation of p2 phage escape mutants reported above can be rationalized in view of the interaction surfaces observed in the RBP/VHH5 complex. The four observed mutations were localized in the RBP structure and spotted on the interaction surface (Fig. 3A). The first mutation appearing in sequence (Thr165Met) is located outside the interaction area, whatever monomer is taken into account (Fig. 3A). In contrast, the three other mutations involve residues belonging to the interacting area. The Trp201Arg mutation is very drastic, as Trp 201 establishes three optimal van der Waals contacts with VHH5 residues Tyr 59, Phe 103, and Ser 105. Visual inspection of models of this mutant with molecular graphics clearly indicates that the mutation Trp201Arg is no longer compatible with binding to VHH5, due to the severe clashes it introduces between both partners. Val 217 is in van der Waals contact with VHH5 residue Ser 30. The Val217Ile mutation leads to a clash between both partners that does not seem severe, since only an extra methyl group is introduced. Asn 236 establishes two strong hydrogen bonds with VHH5 residues Ser 30 and Asn 31. Therefore, the replacement of Asn 236 with a lysine destroys this arrangement and introduces a severe steric conflict between the lysine and VHH5 residues facing it.
Binding of RBP escape mutants to llama VHH5. In order to correlate and quantify the viral and structural molecular data obtained on the escape phage mutants, we expressed the four mutated RBPs (Table 3) as well as the wild-type RBP from p2 and tested their affinity for VHH5 by plasmon resonance. In this technique, direct or indirect attachment of a protein to the chip produces an increase of the plasmon resonance (y axis in Fig. 4). The six-His-tagged RBP molecules, wild type or mutant, were attached to a Ni-NTA chip, and the tagless VHH5 was injected in a second step (Fig. 4). The difference resonance units (DRU) observed after injection of p2 RBP varied between 4,600 and 4,800 for wild-type p2 and mutant Val217Ile, while they were around 5,300 for mutant Thr165Met, 9,800 for Trp201Arg, and 10,200 for Asn236Lys. As the concentrations of the different RBPs were identical, the DRU differences after RBP injections are due to different adsorptions on the chips. The DRU observed after injection of VHH5 was between 2,000 and 2,350 for the wild type and the mutants Val217Ile and Thr165Met and 160 and 40 for the mutants Trp201Arg and Asn236Lys, respectively. Control injection of buffer instead of VHH5, repeated two times, resulted in a DRU of 117. Control injection of VHH5 on the Ni-NTA chip saturated with Ni2+ gave a negligible DRU (17); the RU level after regeneration with 350 mM EDTA was reproducible (11,600 to 11,700). The wild-type molecule is able to bind VHH5, as a sharp increase of the DRU that was followed by a very slow decrease was observed (Fig. 4). The Thr165Met mutant has the same behavior as the wild-type protein and binds to VHH5. In contrast, both Asn236Lys and Trp201Arg do not exhibit a DRU increase upon VHH5 injection and therefore do not bind VHH5. The Val217Ile mutant exhibits a large DRU increase upon VHH5 injection. However, the DRU decrease is faster than those observed with the wild-type or Thr165Met proteins, indicating a larger kinetic off rate of the ligand and thus lower affinity of the former mutant.
Structure of the RBP in complex with glycerol. In this study, the resolution of the RBP structure was extended from 2.3 to 1.7 . In addition to the better quality of the map and the model (Table 2), six glycerol molecules originating from the cryocooling liquor (two per monomer) were clearly identified in the electron density map (Fig. 3C and D). One glycerol molecule is located in the head domain, with an average B-factor of 17.8 2, while the other is located at the upper face of the shoulders, with an average B-factor of 27.6 2. The head domain tightly binds glycerol, as three hydrogen bonds were established between the His 232 and Asp 234 side chains and the glycerol O-1 atom and between the Arg 256 (from the other monomer) guanidyl group and the glycerol O-2 atom. Furthermore, the hydrophobic face of glycerol packs nicely against the Trp 244 side chain, an interaction often observed with sugars (Fig. 3C). The glycerol binding is looser at the shoulders site, as a unique hydrogen bond was established between the Met 33 carbonyl group and the glycerol atom O-1, although a water molecule network also participates in binding. The Trp 43 B side chain is located at 3.8 , but no stacking occurred (Fig. 3D).
The interaction of p2 RBP with VHH5 covers a large area. However, a very specific interaction is observed within the glycerol-binding site. The Tyr 55 from the VHH CDR2 penetrates deeply in the glycerol-binding site (Fig. 3E and F) at the exact position of glycerol, stacked between Trp 244a and Arg 256b. The OH group from Tyr 55 superimposes with the OH-1 of glycerol and establishes similar hydrogen bonds with His 232 and Asp 234 (Fig. 3F).
RBP in complex with glycerol and saccharides in solution. The binding of glycerol to RBP in solution was also analyzed. Considering the close vicinity of Trp residues to glycerol molecules (Trp 244 and 43), we performed fluorescence quenching experiments. Increasing quantities of glycerol were added to a 1 μM solution of RBP. A decrease of Trp fluorescence was observed and registered. A good fit between experimental data and the theoretical curve could be obtained when a unique binding site was taken into account. The same experiment was repeated with 1-phospho-glycerol (D/L racemic) and glucopyranosyl saccharides, galactose, and muramyl dipeptide (MurNAc-D-Ala-D-iGln). The Kd constants observed with the four saccharides (Table 5) range from 0.26 to 0.13 μM.
In order to discriminate between the two glycerol-binding sites, we cloned and expressed the head domain (residues 164 to 264) and repeated the fluorescence quenching experiments with this construct. We also observed a fluorescence quenching curve that could be interpreted by taking into account a unique binding site. Furthermore, the Kd constants, ranging from 0.17 to 0.14 μM, are identical (within experimental error) to the native RBP Kd constants (Table 5). We conclude that the titrated site is located in the head domain. Surprisingly, the Kd values are very close for all sugars, including glycerol. This suggests a lack of specificity and that the extra hydroxyl groups introduced by the glucopyranosyl saccharides compared to glycerol have no or little effect on binding. Visual inspection of models of their complexes with RBP confirms that the extra hydroxyl groups are facing the bulk solvent and do not interact with the RBP molecule (data not shown). In contrast, the phosphate group of the glycerol phosphate is well placed to establish an ionic bond with Arg 256b. Surprisingly, the Kd value of glycerol phosphate is only slightly smaller than that of glycerol.
Soaking with these various saccharides followed by cryocooling and data collection yielded again the glycerol in the electron density maps. Indeed, the large excess of glycerol used in cryocooling was able to displace any bound sugar. No diffraction was observed, unfortunately, when no or other types of cryoprotectants were used.
DISCUSSION
Because phages are omnipresent in the dairy environment, considerable efforts in the last decades were aimed at controlling their proliferation rather than trying to eradicate them (26). In many countries, lactococcal phages of the 936 species are the most predominant. Therefore, several targeted anti-936 strategies have been implemented over the years, while others, such as the use of a phage-neutralizing heavy chain antibody fragment, have been envisioned. In spite of these extensive efforts, "phage attacks" remain today the most common cause of slow milk fermentation, and this is mainly due to phage diversity and evolution. Here, we have convincingly shown that VHH5, which binds to the RBP and prevents phage infection, confers only a limited protection against 936-like phages, as only 12 out of the 50 phages tested could be neutralized. Moreover, escape phage mutants could readily be isolated from phage lysates. The latter observation was somewhat unforeseen, as it was assumed that a mutation in the RBP (to avoid neutralization by VHH5) could lead to phage mutants no longer able to infect the host strain. Obviously, this was not the case. It remains to be seen if other heavy chain antibody fragments, capable of neutralizing both the wild-type phages and the escape mutants, can be selected.
The competitive advantage for the lactococcal 936-like phage mutants to escape neutralization by VHH5 is profitable only if the mutations do not alter significantly their binding to the host receptors. Hence, the mutations should not involve residues located in the receptor-binding site. This was observed here, since none of the four essential residues of the binding site were mutated (Fig. 5). The observation of a genetically stable receptor-binding site raises the possibility of using small molecules targeting specifically this site with an antiviral therapy.
The structure of the complex of phage p2 RBP with neutralizing VHH5 has made it possible to identify the interaction area. Since VHH5 prevents phage infection, the interaction area should contain the receptor-binding site or at least part of it. Analysis of the deduced amino acid sequences of the 12 RBPs neutralized by VHH5 indicated that all of the binding site residues are conserved, with two exceptions (Fig. 5). First, the phage Q50 has a histidine at position 217 instead of a valine. Second, the phages HD8 and HD20 have a valine instead of an alanine at position 225. The former difference might introduce a clash in the contact area of Q50 with VHH5. When isoleucine was replaced by valine in other phages, they could escape to VHH5 neutralization. However, SPR data have indicated that this residue is less sensitive, in molecular terms, than the two other neutralizing mutants. The differences observed with phages HD8 and HD20 seem compatible with VHH5 binding. Ala 225 is buried in a crevice between two subunits (Fig. 3B). It faces Tyr 55 from VHH5, whose side chain can slightly rotate to release enough space for the extra methyl groups to fit (data not shown). SPR experiments with mutated RBP confirm the structural analysis. Surprisingly, the Thr165Met mutant does not belong to the RBP interaction area with VHH5 and exhibits a native-like behavior. Perhaps this mutation affects the binding of the RBP with other phage structural proteins, somehow leading to nonrecognition by VHH5 in vivo. In contrast, the three other mutant RBPs do not bind to VHH5 at all or bind with lower affinity.
It has been shown recently that lactococcal phages belonging to the 936 species, such as bIL170 and 645, bind to cell wall polysaccharides found at the surface of L. lactis (14). Moreover, Geller et al. (16) suggested that the 936-like phages sk1, jj50, and Q64 might not require a host membrane protein or lipoteichoic acid and that cell wall components are sufficient for the initial steps of phage infection. Here, a glycerol molecule was located in the X-ray structure, bound to His 232, Asp 234, Arg 256b, and Trp 244, residues conserved in the list of phages under study (Fig. 3C). Fluorescence quenching experiments confirm this binding site with high affinity to saccharides. The nature of the receptor-binding site is in agreement with polysaccharides acting as receptors at the L. lactis surface. It is reminiscent of sugar-binding sites, such as those observed in lectins (3), which often associate hydrophobic patches of aromatic residues, to accommodate the saccharide's nonpolar face, and polar residues functioning as hydrogen bond donors or acceptors, bringing specificity (4). All saccharides tested here yielded lower Kd values than what is generally observed with monosaccharides. However, these Kd values are restricted within a narrow range and are all close to that of glycerol, making it difficult to identify the natural saccharide receptor. Nonetheless, phosphoglycerol, a component of teichoic or lipoteichoic acids, exhibits an excellent affinity for RBP, a fact that points to teichoic or lipoteichoic acid as a candidate receptor.
In conclusion, we showed that VHH5 has a limited efficacy, since it could neutralize only 12 of the 50 lactococcal phages tested and phages can easily mutate to avoid recognition by the heavy chain antibody fragment. Moreover, we have extended the resolution of the RBP structure to 1.7 and proven the saccharidic nature of the bacterial receptor recognized by lactococcal phages of the 936 species. We propose that teichoic or lipoteichoic acid might be the phage receptor in the L. lactis cell wall. Experiments are under way to validate this hypothesis.
ACKNOWLEDGMENTS
This work was supported, in part, by the Genopole of Marseille-Nice, Unilever, and the Natural Sciences and Engineering Research Council (NSERC) of Canada.
Marie-Therèse Guidici-Orticoni is acknowledged for her help with the SPR experiment, Geneviève Drolet for the neutralization assays, and Hans de Haard, C. Theo Verrips, and Aat M. Ledeboer for providing us with VHH5 and for their interest and support.
These authors contributed equally to the work.
REFERENCES
Bissonnette, F., S. Labrie, H. Deveau, M. Lamoureux, and S. Moineau. 2000. Characterization of mesophilic mixed starter cultures used for the manufacture of aged cheddar cheese. J. Dairy Sci. 83:620-627.
Boucher, I., M. Parrot, H. Gaudreau, C. P. Champagne, C. Vadeboncoeur, and S. Moineau. 2002. Novel food-grade plasmid vector based on melibiose fermentation for the genetic engineering of Lactococcus lactis. Appl. Environ. Microbiol. 68:6152-6161.
Bourne, Y., J. Mazurier, D. Legrand, P. Rouge, J. Montreuil, G. Spik, and C. Cambillau. 1994. Structures of a legume lectin complexed with the human lactotransferrin N2 fragment, and with an isolated biantennary glycopeptide: role of the fucose moiety. Structure 2:209-219.
Bourne, Y., A. Roussel, M. Frey, P. Rouge, J. C. Fontecilla-Camps, and C. Cambillau. 1990. Three-dimensional structures of complexes of Lathyrus ochrus isolectin I with glucose and mannose: fine specificity of the monosaccharide-binding site. Proteins 8:365-376.
Chandry, P. S., S. C. Moore, J. D. Boyce, B. E. Davidson, and A. J. Hillier. 1997. Analysis of the DNA sequence, gene expression, origin of replication and modular structure of the Lactococcus lactis lytic bacteriophage sk1. Mol. Microbiol. 26:49-64.
Collaborative Computational Project, Number 4. 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D 50:760-763.
Crutz-Le Coq, A. M., B. Cesselin, J. Commissaire, and J. Anba. 2002. Sequence analysis of the lactococcal bacteriophage bIL170: insights into structural proteins and HNH endonucleases in dairy phages. Microbiology 148:985-1001.
De Haard, H. J., S. Bezemer, A. M. Ledeboer, W. H. Muller, P. J. Boender, S. Moineau, M. C. Coppelmans, A. J. Verkleij, L. G. Frenken, and C. T. Verrips. 2005. Llama antibodies against a lactococcal protein located at the tip of the phage tail prevent phage infection. J. Bacteriol. 187:4531-4541.
Desmyter, A., S. Spinelli, F. Payan, M. Lauwereys, L. Wyns, S. Muyldermans, and C. Cambillau. 2002. Three camelid VHH domains in complex with porcine pancreatic alpha-amylase. Inhibition and versatility of binding topology. J. Biol. Chem. 277:23645-23650.
Deveau, H., M. R. Van Calsteren, and S. Moineau. 2002. Effect of exopolysaccharides on phage-host interactions in Lactococcus lactis. Appl. Environ. Microbiol. 68:4364-4369.
Devereux, J., P. Haeberli, and O. Smithies. 1984. A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12:387-395.
Duplessis, M., and S. Moineau. 2001. Identification of a genetic determinant responsible for host specificity in Streptococcus thermophilus bacteriophages. Mol. Microbiol. 41:325-336.
Dupont, K., F. K. Vogensen, and J. Josephsen. 2005. Detection of lactococcal 936-species bacteriophages in whey by magnetic capture hybridization PCR targeting a variable region of receptor-binding protein genes. J. Appl. Microbiol. 98:1001-1009.
Dupont, K., F. K. Vogensen, H. Neve, J. Bresciani, and J. Josephsen. 2004. Identification of the receptor-binding protein in 936-species lactococcal bacteriophages. Appl. Environ. Microbiol. 70:5818-5824.
Forde, A., and G. F. Fitzgerald. 1999. Bacteriophage defence systems in lactic acid bacteria. Antonie Leeuwenhoek 76:89-113.
Geller, B. L., H. T. Ngo, D. T. Mooney, P. Su, and N. Dunn. 2005. Lactococcal 936-species phage attachment to surface of Lactococcus lactis. J. Dairy Sci. 88:900-907.
Haggard-Ljungquist, E., C. Halling, and R. Calendar. 1992. DNA sequences of the tail fiber genes of bacteriophage P2: evidence for horizontal transfer of tail fiber genes among unrelated bacteriophages. J. Bacteriol. 174:1462-1477.
Jarvis, A. W., G. F. Fitzgerald, M. Mata, A. Mercenier, H. Neve, I. B. Powell, C. Ronda, M. Saxelin, and M. Teuber. 1991. Species and type phages of lactococcal bacteriophages. Intervirology 32:2-9.
Josephsen, J., and F. K. Vogensen. 1989. Identification of three different plasmid-encoded restriction/modification systems in Streptococcus lactis subsp. cremoris W56. FEMS Microbiol. Lett. 59:161-166.
Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment. Brief Bioinform. 5:150-163.
Labrie, S., and S. Moineau. 2002. Complete genomic sequence of bacteriophage ul36: demonstration of phage heterogeneity within the P335 quasi-species of lactococcal phages. Virology 296:308-320.
Labrie, S., and S. Moineau. 2000. Multiplex PCR for detection and identification of lactococcal bacteriophages. Appl. Environ. Microbiol. 66:987-994.
Ledeboer, A. M., S. Bezemer, J. J. W. de Haard, I. M. Schaffers, C. T. Verrips, C. van Vliet, E.-M. Dusterhoft, P. Zoon, S. Moineau, and L. G. Frenken. 2002. Preventing phage lysis of Lactococcus lactis in cheese production using a neutralizing heavy-chain antibody fragment from llama. J. Dairy Sci. 85:1376-1382.
Loof, M., J. Lembke, and M. Teuber. 1983. Characterisation of the genome of the Streptococcus lactis subsp. diacetylactis bacteriophages P008 widespread in German cheese factories. Syst. Appl. Microbiol. 4:413-423.
McKay, L. L., K. A. Baldwin, and E. A. Zottola. 1972. Loss of lactose metabolism in lactic streptococci. Appl. Microbiol. 23:1090-1096.
Moineau, S. 1999. Applications of phage resistance in lactic acid bacteria. Antonie Leeuwenhoek 76:377-382.
Moineau, S., J. Fortier, and H.-W. Ackermann. 1992. Characterization of lactococcal bacteriophages from Quebec cheese plants. Can. J. Microbiol. 38:875-882.
Moineau, S., D. Tremblay, and S. Labrie. 2002. Phages of lactic acid bacteria: from genomics to industrial applications. ASM News 68:388-393.
Moineau, S., M. Borkaev, B. J. Holler, S. A. Walker, J. K. Kondo, E. R. Vedamuthu, and P. A. Vandenbergh. 1996. Isolation and characterization of lactococcal bacteriophages from cultured buttermilk plants in the United States. J. Dairy Sci. 79:2104-2111.
Murshudov, G., A. A. Vagin, and E. J. Dodson. 1997. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D 53:240-255.
Muyldermans, S., C. Cambillau, and L. Wyns. 2001. Recognition of antigens by single-domain antibody fragments: the superfluous luxury of paired domains. Trends Biochem. Sci. 26:230-235.
Rince, A., M. Tangney, and G. F. Fitzgerald. 2000. Identification of a DNA region from lactococcal phage sk1 protecting phage 712 from the abortive infection mechanism AbiF. FEMS Microbiol. Lett. 182:185-191.
Roussel, A., and C. Cambillau. 1991. The TURBO-FRODO graphics package, vol. 81. Silicon Graphics Geometry Partners Directory, Mountain View, Calif.
Spinelli, S., A. Desmyter, C. T. Verrips, H. J. de Haard, S. Moineau, and C. Cambillau. 2006. Lactococcal bacteriophage p2 receptor-binding protein structure suggests a common ancestor gene with bacterial and mammalian viruses. Nat. Struct. Mol. Biol. 13:85-89.
Valyasevi, R., W. E. Sandine, and B. L. Geller. 1994. Lactococcus lactis ssp. lactis C2 bacteriophage sk1 receptor involving rhamnose and glucose moieties in the cell wall. J. Dairy Sci. 77:1-6.
Walhout, A. J., G. F. Temple, M. A. Brasch, J. L. Hartley, M. A. Lorson, S. van den Heuvel, and M. Vidal. 2000. GATEWAY recombinational cloning: application to the cloning of large numbers of open reading frames or ORFeomes. Methods Enzymol. 328:575-592.(Denise M. Tremblay, Marie)
ABSTRACT
Phage p2, a member of the lactococcal 936 phage species, infects Lactococcus lactis strains by binding initially to specific carbohydrate receptors using its receptor-binding protein (RBP). The structures of p2 RBP, a homotrimeric protein composed of three domains, and of its complex with a neutralizing llama VH domain (VHH5) have been determined (S. Spinelli, A. Desmyter, C. T. Verrips, H. J. de Haard, S. Moineau, and C. Cambillau, Nat. Struct. Mol. Biol. 13:85-89, 2006). Here, we show that VHH5 was able to neutralize 12 of 50 lactococcal phages belonging to the 936 species. Moreover, escape phage mutants no longer neutralized by VHH5 were isolated from 11 of these phages. All of the mutations (but one) cluster in the RBP/VHH5 interaction surface that delineates the receptor-binding area. A glycerol molecule, observed in the 1.7- resolution structure of RBP, was found to bind tightly (Kd = 0.26 μM) in a crevice located in this area. Other saccharides bind RBP with comparable high affinity. These data prove the saccharidic nature of the bacterial receptor recognized by phage p2 and identify the position of its binding site in the RBP head domain.
INTRODUCTION
Bacteriophage infection is a major problem impairing any industrial fermentation that relies on bacteria to transform a substrate into fermented products or specific molecules. Several strains of the gram-positive bacterium Lactococcus lactis are used quite extensively worldwide for the manufacture of fermented milk products. For decades, the dairy industry has been dealing with phage infections of their L. lactis strains because virulent phages are ubiquitous within their environment (28).
Phages of L. lactis are classified within several groups (21), although those commonly found in dairy plants belong to three different species, i.e., 936, P335, and c2 (18, 27, 29). From a taxonomical standpoint, these phages are all members of the Siphoviridae family (order Caudovirales) that also includes the coliphage lambda. Siphophages are characterized by a double-stranded DNA genome and a long noncontractile tail. The phages belonging to the species 936 and P335 have a small isometric capsid, while those from the c2 species have a prolate capsid.
The bacteriophage infection process starts by the specific recognition between the phage receptor-binding protein (RBP) located at the tip of the tail and the receptor distributed over the host cell surface (14, 35). Bacterial receptors have been well studied in gram-negative bacteria, particularly in Escherichia coli, while similar information lags behind in gram-positive bacteria (12). It has been previously shown that phages infecting some L. lactis strains adsorb initially to the cell wall surface and likely to various carbohydrates containing rhamnose, glucose, or galactose (for a review, see Forde and Fitzgerald [15]). For many phages, this binding step is reversible, and phages of the c2 species, for example, require a second irreversible binding step to a predicted membrane-attached protein (PIP) of 901 amino acids. However, phages of the species 936 and P335 do not use PIP as a secondary receptor (14). A better understanding at a molecular level of these phage-host interactions in gram-positive bacteria should lead to better tools for the control of phage infection in various biotechnological processes as well as reveal insights into the molecular basis of host specificity.
A novel antiphage strategy was recently designed based on the use of a phage-neutralizing heavy chain antibody fragment (VHH5) obtained from llama (31). VHH5 produced at large scale in the food microorganism Saccharomyces cerevisiae prevented the infection of an L. lactis strain by phage p2 (936 species) during the manufacture of cheese (23). VHH5 recognized the RBP of phage p2, a 30-kDa protein (ORF18) located at the distal part of the phage tail, with a high affinity (Kd = 1.4 nM) (8).
We previously determined the crystal structure of the p2 RBP trimer at 2.3- resolution and that of its complex with three molecules of VHH5 at 2.7- resolution (34). The RBP ternary complex is formed of three subunits, each assembling three symmetry-related domains: the shoulders, the interlaced neck, and the heads, from N to C terminus (Fig. 1A). The shoulder domains, encompassing residues 1 to 141, have a -sandwich fold formed of two -sheets of four antiparallel -strands each. The neck (residues 142 to 163) folds as a triple-stranded interlaced -helix that forms an equilateral triangular prism around the threefold axis (Fig. 1A). The receptor recognition domain (residues 164 to 264), designated as the head, is a -barrel formed of seven antiparallel -strands. Each domain is parallel to the threefold axis and interacts with the two others, forming a very compact structure (Fig. 1A). In the structure of the head trimer complexed with VHH5, each of the three VHH5 domains binds the RBP head at the top half of the trimer and forms a regular trimeric complex, sharing the same threefold axis as the isolated RBP (Fig. 1B).
Here, we further investigated the molecular interactions between various RBPs of Lactococcus lactis phages and VHH5. First, the broadness of the neutralization conferred by VHH5 was tested against 50 distinct lactococcal phages of the 936 species. Then, phage mutants escaping the VHH5 neutralization were isolated, and their mutated RBPs were analyzed by surface plasmon resonance (SPR). We extended the resolution of the p2 RBP structure to 1.7- resolution and localized glycerol molecules in the RBP head, within the VHH5 binding area. We performed fluorescence quenching experiments with saccharides to demonstrate the existence of a high-affinity saccharide binding site overlapping with the structural glycerol attachment site.
MATERIALS AND METHODS
Bacterial strains, phages, and host range. The phages and hosts used in this study are listed in Table 1. L. lactis strains were grown at 30°C in M17 broth (Quelab) supplemented with 0.5% glucose (GM17). For phage propagation, GM17 was supplemented with 10 mM CaCl2 as described previously (27). The neutralization assays were performed in BCP broth (2% tryptone, 0.5% yeast extract, 0.5% glucose, 0.4% NaCl, 0.15% Na-acetate, 40 mg of bromocresol purple per liter [2]).
Native and mutant RBPs, VHH5, and head domain production. Phage p2 wild-type RBP and llama VHH5 were cloned, expressed, and purified as described previously (8, 23). Single-amino-acid mutations of RBP were generated using a QuickChange site-directed mutagenesis kit (Stratagene). Briefly, the pET28a/ORF18 (8, 23) was entirely amplified by PfuTurbo DNA polymerase using two complementary mutated primers: the Thr 165 codon was mutated to encode Met using the primers T165M-F (5'-TCGATGTTCCAGTTCAAATGTTGACAGTTGAAGCTGG-3') and T165M-R (5'-CCAGCTTCAACTGTCAACATTTGAACTGGAACATCGA-3'); Trp 201 was mutated to encode Arg using the primers W201R-F (5'-GTGTCAAATATACAAAAAGGCCGGAATATGTCTGGAACATGGG-3') and W201R-R (5'-CCCATGTTCCAGACATATTCCGGCC TTTTTGTATATTTGACAC-3'); Val 217 was mutated to encode Ile using the primers V217I-F (5'-GACCATTTCGTCCAGCTGCTATTCAAAGTCTTGTTGGTCATTTTG-3') and V217I-R (5'-CAAAATGACCAACAAGACTTTGAATAGCAGCTGGACGAAATGGTC-3'); and Asn 236 was mutated to encode Lys using the primers N236K-F (5'-GATACTTCTTTCCATATTGATATAAAGCCAAATGGTAGTATTACTTGGT-3') and N236K-R (5'-ACCAAGTAATACTACCATTTGGCTTTATATCAATATGGAAAGAAGTATC-3') (bold nucleotides mark mutational changes). Parental plasmid was digested by DpnI, and the vector containing the mutation was transformed in supercompetent XL1-Blue cells. Transformants were selected onto LB ampicillin (100 μg/ml) agar plates. Mutations were verified by automated DNA sequencing (Genome Express, France). The DNA fragments encoding the phage p2 RBP head segment (residues 164 to 264) were amplified by PCR from the pET28a/ORF18 plasmid (8, 23) using the following primers: forward primer, 5'-ACAAGTTTGTACAAAAAGCAGGCTtagaaggagatagaaccatgaaagttcaaacgttgacagttgaagctgg-3' (attB1 sequence in uppercase, initial atg codon in bold, and the ribosome binding site and the first part of the open reading frame in lowercase); reverse primer, 5'-ACCACTTTGTACAAGAAAGCTGGGTcTTAGTGATGGTGATGGTGATGtttaatgaagtaacttccgttacc-3' (attB2 sequence in uppercase, stop codon in bold, six-His tag in italic, and the last part of the open reading frame in lowercase). The PCR product was subcloned in the pDest14 vector by recombination (36) using Gateway technology (Invitrogen). Expression of mutant proteins and the RBP head was carried out using the BL21(DE3) and Rosetta(DE3)pLysS E. coli strains in Terrific broth (Gibco) and 2YT supplemented with antibiotics, respectively. When the optical density at 600 nm reached 0.5, expression was induced with 0.5 mM IPTG (isopropyl--D-thiogalactopyranoside), and the temperature was decreased to 15°C and 17°C for mutants and head, respectively. All proteins produced were soluble and were purified using a two-step procedure on a Ni column, taking advantage of the presence of a six-His tag at the C terminus, followed by gel filtration on a Superdex 200 column in 1.8 mM KH2PO4, 10.1 mM Na2HPO4, pH 7.2, 2.7 mM KCl, 137 mM NaCl. For each protein, productions yielded between 20 and 50 mg of pure protein per liter of culture.
Neutralization assay. The efficacy of the VHH5 antibody in neutralizing various lactococcal phages was determined by a lactic acid production assay. Briefly, 104 phages and 1 μg/ml of VHH5 antibody were added to 3 ml of BCP medium supplemented with 10 mM CaCl2. The mixture was incubated for 1 h at 30°C. Then, the appropriate L. lactis host was inoculated at 1% and incubated for an additional 5 h at 30°C. This neutralization assay also included a control that consisted of the BCP medium inoculated with only the L. lactis host and another control in which only the host and the phages were added into the BCP medium. After the incubation period, the lactic acid production by L. lactis strains (without phages) resulted in a color change (from purple to yellow) of the medium containing a pH indicator (bromocresol purple). Conversely, the presence of phages led to cell lysis and limited acid production, and thus the color of the medium remained purple. If the antibody properly neutralized the phage, L. lactis was able to produce lactic acid and the pH indicator turned yellow. This assay was performed in two independent experiments for all of the 50 lactococcal phages tested.
Isolation of nonneutralized phage mutants. For all lactococcal phages neutralized by VHH5, the isolation of nonneutralized phage mutants was carried out. Following the standard neutralization assay, the cultures were centrifuged, and the supernatant was filtered on a 0.45-μm Acrodisc. The phage titer was determined (27), and a second neutralization assay was performed using 104 phages from the above filtrate. Three to five neutralization assays were needed to obtain nonneutralized phages in which no color change occurred in the presence of VHH5. Phages were then plated, and individual plaques were isolated and retested for neutralization by VHH5. Phage adsorption tests were carried out for wild-type and mutant phages as reported elsewhere (12). Adsorption assays were performed in three independent experiments.
DNA manipulation and fingerprinting analysis. Phage DNA was isolated using a Lambda mini kit (QIAGEN) with the following modifications. The phage pellet was resuspended in buffer L3 containing 1 mg/ml of proteinase K and then incubated for 30 min at 65°C. In the final step, the DNA pellet was dissolved in 50 μl of 10 mM Tris (pH 8.0). DNA was digested with EcoRI and EcoRV as recommended by the manufacturer (Roche Diagnostics). Restriction fragments were heat-treated at 65°C for 10 min, electrophoresed on agarose gel (0.7%) in Tris-acetate-EDTA buffer, stained with ethidium bromide, and visualized under UV light. Phage restriction fragment length polymorphism (RFLP) patterns were analyzed with Molecular Analyst Fingerprinting Plus version 1.6 software (Bio-Rad), using the unweighted-pair group method using average linkages.
DNA sequencing and analysis. The genes coding for the receptor-binding protein of wild-type lactococcal phages and nonneutralized phage mutants were sequenced on both strands by primer walking using synthetic oligonucleotide primers (Invitrogen) and phage genomic DNA as a template. DNA sequencing was performed using an ABI Prism 3100 apparatus from the service center at the Universite Laval. Computer-assisted DNA and protein analyses were performed using the Genetics Computer Group Sequence Analysis software package version 10.3 (11). The multiple sequence alignment was achieved with ClustalW, available at EMBL-EBI (http://www.ebi.ac.uk/clustalw/). For a few phages (712, bIL170, P008, and sk1), the RBP gene was already available in GenBank but was still confirmed by sequencing. Phylogenetic and molecular evolutionary analyses were conducted using MEGA version 3.0 (20). The bootstrap test for phylogeny was used and replicated 2,000 times.
Crystallographic study of RBP. Crystallization of the RBP was described elsewhere (34). The 1.7- data set was collected at beam line ID14-4 (ESRF, Grenoble). Data were integrated and reduced using MOSFLM and SCALA (6). Refinement was performed with REFMAC5 (30) alternated with rebuilding using Turbo-Frodo (33), using the 2.3- RBP structure as a starting point (2BSD). Statistics are presented in Table 2. The high resolution and the low R and R-free values indicate the excellent quality of the structure.
Surface plasmon resonance. All surface plasmon resonance experiments were performed using a Biacore 1000 instrument (Biacore Inc., Piscataway, NJ) at 25°C. The buffer for the continuous-flow pump was HBS-P (10 mM HEPES, pH 7.4, 150 mM NaCl, 0.005% [vol/vol] surfactant P-20) supplemented with 0.05 mM EDTA; the buffer for the sample pump was HBS-EP (10 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.005% [vol/vol] surfactant P-20). The buffers were degassed before use. The general strategy followed in the Biacore experiment was to take advantage of the presence of a His tag in the native and mutant RBPs and the absence of a His tag in the VHH5 molecule. The chip Ni-nitrilotriacetic acid (Ni-NTA) was first saturated with Ni2+ by washing it with 0.5 μM NiCl2 (20 μl at 20 μl/min). Then, the RBP solution (at 190 nM, 40 μl at 10 μl/min) was passed, followed by the VHH5 solution (100 nM, 40 μl at 10 μl/min). Regeneration was achieved by washing the flow cell with 350 mM EDTA (20 μl at 20 μl/min). Step 1 was run with the command Inject, followed by Extraclean; steps 2 and 3 were run with the command Kinject; step 4 was run with the command Quickinject, followed by Extraclean. The four-step procedure described above was performed identically for the wild-type RBP, the mutants (Thr165Met, Trp201Arg, Val217Ile, and Asn236Lys), and the control injection of running buffer instead of VHH5.
Fluorescence quenching experiments. Fluorescence experiments were carried out on a Varian Eclipse spectrofluorimeter using a quartz cuvette in a right-angle configuration; the light path was 0.4 and 1 cm for the excitation and emission, respectively. The interaction of RBP with saccharides was monitored by recording the quenching of the intrinsic protein fluorescence upon addition of ligand aliquots. The excitation wavelength was 290 nm, and emission spectra were recorded in the range of 320 to 400 nm. The excitation slit was 5 nm while the emission slit was 10 nm for a protein concentration of 1 μM. A moving-average smoothing procedure was applied, with a window of 3 nm. Titrations were carried out at room temperature with 1 μM protein in 10 mM phosphate buffer Na/Na2, 50 mM NaCl, pH 7.5. The fluorescence intensities at the maximum of emission (346 nm) for different concentrations of quencher were corrected for the buffer contribution before plotting and further analysis. The affinity was estimated by plotting the decrease of fluorescence intensity at the emission maximum as 100 – (Ii – Imin)/(I0 – Imin) x 100 against the quencher concentration; I0 is the maximum of fluorescence intensity of the protein alone, Ii is the fluorescence intensity after the addition of quencher (i), and Imin is the fluorescence intensity at saturating concentration of quencher. The Kd values were estimated using Prism 3.02 (GraphPad Software, Inc.) by nonlinear regression for a single binding site with the equation Y = Bmax x [X/(Kd + X)], where Bmax is the maximal binding and Kd is the concentration of ligand required to reach half-maximal binding.
Nucleotide sequence accession numbers. All RBP gene sequences were deposited in the GenBank database (see Table 1 for accession numbers).
Protein structure accession number. The coordinates and structure factors for the RBP have been deposited in the Protein Data Bank of the Research Collaboratory for Structural Bioinformatics (http://www.rcsb.org/pdb/) as entry 1ZRU.
RESULTS
Neutralization of 936-related phages by VHH5. It was previously shown that VHH5 binds to the RBP of the lactococcal phage p2, a member of the 936 group (8, 23). The addition of VHH5 to milk also prevents phage infection and fermentation failure during the manufacture of cheese (8, 23). In order to investigate the neutralizing potential of the VHH5 antibody, 50 lactococcal phages of the 936 group were selected and tested in a neutralization assay (Table 1). All of these phages were isolated from various geographic areas and are distinct, based on the analysis of their host range as well as the EcoRV and EcoRI restriction profiles (RFLP) of their genomes (data not shown). Of the 50 phages tested, 12 phages were neutralized by VHH5 (Table 1). The 38 other lactococcal phages were not neutralized by the antibody and were still able to infect their respective host cells (Table 1). These results indicate a significant diversity of the RBP epitope and suggest a limited efficiency of VHH5 for phage control in an industrial framework.
Isolation of nonneutralized phages. From an evolutionary standpoint, phages are known to adapt to their local environment. Thus, the isolation of phage mutants that were no longer neutralized by VHH5 was carried out. Repeated infection experiments of the 12 phages in the presence of VHH5 were performed and led to isolation of nonneutralized mutants derived for 11 of these 12 phages (Table 3). No nonneutralized mutants were isolated for phage HD12, while two different mutants were identified for phage p2. Three to five challenges were needed to obtain high-titer mutant phages. The escape phage mutants were no longer neutralized by VHH5 but had the same RFLP patterns as the neutralized wild-type phages (data not shown). These results suggest point mutations within the epitope specificity of VHH5 to the receptor-binding protein. Adsorption of the escape phage mutants to their hosts was also measured and found to be similar to the adsorption of the wild-type phages (data not shown). Altogether, these data show that most phages can rapidly evolve to evade recognition by VHH5 without affecting their host range.
Analysis of the RBP of lactococcal phages. In order to investigate the epitope recognition of VHH5, the RBP gene of each of the 50 wild-type lactococcal phages and the nonneutralized phage mutants was sequenced and analyzed (Fig. 1 and Table 3). A single mutation was found to be sufficient to avoid neutralization by the antibody fragment. In all mutants, the point mutation led to an amino acid change within the C-terminal region of the RBP. Overall, four different mutations were observed (Table 3).
The deduced RBPs from 50 phages range in size from 252 to 308 amino acid residues. The alignment of the multiple RBP sequences revealed that the N terminus is highly conserved, as the 50 shoulder domains (residues 1 to 141) shared 50.4% identity (71/141 amino acids). In fact, the first 120 amino acids were particularly conserved (56.7% identity), while the next 21 residues were less conserved. It was previously predicted that this conserved N-terminal part of the protein may be involved in protein-protein interactions with other phage tail proteins (12, 17). The last 21 residues are probably slightly less conserved to allow connection to the neck domain.
The neck (residues 142 to 163) and head (C-terminal) domains were more diverse, possibly reflecting the differences observed in host recognition (data not shown). However, in most cases, a specific neck domain was associated with a particular head domain. One could argue that the rigid neck structure serves as a linker between the larger shoulder and head domains. Nonetheless, the 50 RBPs could be classified in at least 10 groups based on their overall identities (Fig. 2). The 12 phages neutralized by VHH5 were classified in one large group of 18 RBPs sharing almost 80% identity. Comparative analysis of the 12 RBPs recognized by VHH5 and the 6 nonrecognized RBPs allowed the identification of a few amino acid discrepancies between both groups of RBP (data not shown). Finally, the other 32 phages have a distinct C-terminal RBP compared to the above group, explaining their nonneutralization by VHH5 and leading to the identification of 9 other groups of RBPs. Recently, Dupont et al. (13) reported six different types of RBP genes of lactococcal 936 species bacteriophages and developed a PCR assay to detect them. They suggested that the PCR assay could be used in cheese plants for phage detection. Here, we showed that there are many more RBP groups within this lactococcal phage species, and perhaps this variability may limit the usefulness of such a PCR assay.
Complex of the RBP with neutralizing llama VHH5. We have already reported that, overall, the water-accessible surface area of the RBP of phage p2 covered in the interaction with VHH5 is 825 2, with 647 2 and 188 2 for monomers A and B, respectively. The interaction surface involves many charged or semipolar residues, but nonpolar residues are also observed (Table 4). The VHH5 domain interacts with the RBP in a semilateral way, as observed in several camelid VHH complexes with haptens or proteins (9, 31). Facing the RBP surface, the VHH5 interacting surface is also rather polar and formed by a mixture of charged or semipolar residues (Table 4).
The isolation of p2 phage escape mutants reported above can be rationalized in view of the interaction surfaces observed in the RBP/VHH5 complex. The four observed mutations were localized in the RBP structure and spotted on the interaction surface (Fig. 3A). The first mutation appearing in sequence (Thr165Met) is located outside the interaction area, whatever monomer is taken into account (Fig. 3A). In contrast, the three other mutations involve residues belonging to the interacting area. The Trp201Arg mutation is very drastic, as Trp 201 establishes three optimal van der Waals contacts with VHH5 residues Tyr 59, Phe 103, and Ser 105. Visual inspection of models of this mutant with molecular graphics clearly indicates that the mutation Trp201Arg is no longer compatible with binding to VHH5, due to the severe clashes it introduces between both partners. Val 217 is in van der Waals contact with VHH5 residue Ser 30. The Val217Ile mutation leads to a clash between both partners that does not seem severe, since only an extra methyl group is introduced. Asn 236 establishes two strong hydrogen bonds with VHH5 residues Ser 30 and Asn 31. Therefore, the replacement of Asn 236 with a lysine destroys this arrangement and introduces a severe steric conflict between the lysine and VHH5 residues facing it.
Binding of RBP escape mutants to llama VHH5. In order to correlate and quantify the viral and structural molecular data obtained on the escape phage mutants, we expressed the four mutated RBPs (Table 3) as well as the wild-type RBP from p2 and tested their affinity for VHH5 by plasmon resonance. In this technique, direct or indirect attachment of a protein to the chip produces an increase of the plasmon resonance (y axis in Fig. 4). The six-His-tagged RBP molecules, wild type or mutant, were attached to a Ni-NTA chip, and the tagless VHH5 was injected in a second step (Fig. 4). The difference resonance units (DRU) observed after injection of p2 RBP varied between 4,600 and 4,800 for wild-type p2 and mutant Val217Ile, while they were around 5,300 for mutant Thr165Met, 9,800 for Trp201Arg, and 10,200 for Asn236Lys. As the concentrations of the different RBPs were identical, the DRU differences after RBP injections are due to different adsorptions on the chips. The DRU observed after injection of VHH5 was between 2,000 and 2,350 for the wild type and the mutants Val217Ile and Thr165Met and 160 and 40 for the mutants Trp201Arg and Asn236Lys, respectively. Control injection of buffer instead of VHH5, repeated two times, resulted in a DRU of 117. Control injection of VHH5 on the Ni-NTA chip saturated with Ni2+ gave a negligible DRU (17); the RU level after regeneration with 350 mM EDTA was reproducible (11,600 to 11,700). The wild-type molecule is able to bind VHH5, as a sharp increase of the DRU that was followed by a very slow decrease was observed (Fig. 4). The Thr165Met mutant has the same behavior as the wild-type protein and binds to VHH5. In contrast, both Asn236Lys and Trp201Arg do not exhibit a DRU increase upon VHH5 injection and therefore do not bind VHH5. The Val217Ile mutant exhibits a large DRU increase upon VHH5 injection. However, the DRU decrease is faster than those observed with the wild-type or Thr165Met proteins, indicating a larger kinetic off rate of the ligand and thus lower affinity of the former mutant.
Structure of the RBP in complex with glycerol. In this study, the resolution of the RBP structure was extended from 2.3 to 1.7 . In addition to the better quality of the map and the model (Table 2), six glycerol molecules originating from the cryocooling liquor (two per monomer) were clearly identified in the electron density map (Fig. 3C and D). One glycerol molecule is located in the head domain, with an average B-factor of 17.8 2, while the other is located at the upper face of the shoulders, with an average B-factor of 27.6 2. The head domain tightly binds glycerol, as three hydrogen bonds were established between the His 232 and Asp 234 side chains and the glycerol O-1 atom and between the Arg 256 (from the other monomer) guanidyl group and the glycerol O-2 atom. Furthermore, the hydrophobic face of glycerol packs nicely against the Trp 244 side chain, an interaction often observed with sugars (Fig. 3C). The glycerol binding is looser at the shoulders site, as a unique hydrogen bond was established between the Met 33 carbonyl group and the glycerol atom O-1, although a water molecule network also participates in binding. The Trp 43 B side chain is located at 3.8 , but no stacking occurred (Fig. 3D).
The interaction of p2 RBP with VHH5 covers a large area. However, a very specific interaction is observed within the glycerol-binding site. The Tyr 55 from the VHH CDR2 penetrates deeply in the glycerol-binding site (Fig. 3E and F) at the exact position of glycerol, stacked between Trp 244a and Arg 256b. The OH group from Tyr 55 superimposes with the OH-1 of glycerol and establishes similar hydrogen bonds with His 232 and Asp 234 (Fig. 3F).
RBP in complex with glycerol and saccharides in solution. The binding of glycerol to RBP in solution was also analyzed. Considering the close vicinity of Trp residues to glycerol molecules (Trp 244 and 43), we performed fluorescence quenching experiments. Increasing quantities of glycerol were added to a 1 μM solution of RBP. A decrease of Trp fluorescence was observed and registered. A good fit between experimental data and the theoretical curve could be obtained when a unique binding site was taken into account. The same experiment was repeated with 1-phospho-glycerol (D/L racemic) and glucopyranosyl saccharides, galactose, and muramyl dipeptide (MurNAc-D-Ala-D-iGln). The Kd constants observed with the four saccharides (Table 5) range from 0.26 to 0.13 μM.
In order to discriminate between the two glycerol-binding sites, we cloned and expressed the head domain (residues 164 to 264) and repeated the fluorescence quenching experiments with this construct. We also observed a fluorescence quenching curve that could be interpreted by taking into account a unique binding site. Furthermore, the Kd constants, ranging from 0.17 to 0.14 μM, are identical (within experimental error) to the native RBP Kd constants (Table 5). We conclude that the titrated site is located in the head domain. Surprisingly, the Kd values are very close for all sugars, including glycerol. This suggests a lack of specificity and that the extra hydroxyl groups introduced by the glucopyranosyl saccharides compared to glycerol have no or little effect on binding. Visual inspection of models of their complexes with RBP confirms that the extra hydroxyl groups are facing the bulk solvent and do not interact with the RBP molecule (data not shown). In contrast, the phosphate group of the glycerol phosphate is well placed to establish an ionic bond with Arg 256b. Surprisingly, the Kd value of glycerol phosphate is only slightly smaller than that of glycerol.
Soaking with these various saccharides followed by cryocooling and data collection yielded again the glycerol in the electron density maps. Indeed, the large excess of glycerol used in cryocooling was able to displace any bound sugar. No diffraction was observed, unfortunately, when no or other types of cryoprotectants were used.
DISCUSSION
Because phages are omnipresent in the dairy environment, considerable efforts in the last decades were aimed at controlling their proliferation rather than trying to eradicate them (26). In many countries, lactococcal phages of the 936 species are the most predominant. Therefore, several targeted anti-936 strategies have been implemented over the years, while others, such as the use of a phage-neutralizing heavy chain antibody fragment, have been envisioned. In spite of these extensive efforts, "phage attacks" remain today the most common cause of slow milk fermentation, and this is mainly due to phage diversity and evolution. Here, we have convincingly shown that VHH5, which binds to the RBP and prevents phage infection, confers only a limited protection against 936-like phages, as only 12 out of the 50 phages tested could be neutralized. Moreover, escape phage mutants could readily be isolated from phage lysates. The latter observation was somewhat unforeseen, as it was assumed that a mutation in the RBP (to avoid neutralization by VHH5) could lead to phage mutants no longer able to infect the host strain. Obviously, this was not the case. It remains to be seen if other heavy chain antibody fragments, capable of neutralizing both the wild-type phages and the escape mutants, can be selected.
The competitive advantage for the lactococcal 936-like phage mutants to escape neutralization by VHH5 is profitable only if the mutations do not alter significantly their binding to the host receptors. Hence, the mutations should not involve residues located in the receptor-binding site. This was observed here, since none of the four essential residues of the binding site were mutated (Fig. 5). The observation of a genetically stable receptor-binding site raises the possibility of using small molecules targeting specifically this site with an antiviral therapy.
The structure of the complex of phage p2 RBP with neutralizing VHH5 has made it possible to identify the interaction area. Since VHH5 prevents phage infection, the interaction area should contain the receptor-binding site or at least part of it. Analysis of the deduced amino acid sequences of the 12 RBPs neutralized by VHH5 indicated that all of the binding site residues are conserved, with two exceptions (Fig. 5). First, the phage Q50 has a histidine at position 217 instead of a valine. Second, the phages HD8 and HD20 have a valine instead of an alanine at position 225. The former difference might introduce a clash in the contact area of Q50 with VHH5. When isoleucine was replaced by valine in other phages, they could escape to VHH5 neutralization. However, SPR data have indicated that this residue is less sensitive, in molecular terms, than the two other neutralizing mutants. The differences observed with phages HD8 and HD20 seem compatible with VHH5 binding. Ala 225 is buried in a crevice between two subunits (Fig. 3B). It faces Tyr 55 from VHH5, whose side chain can slightly rotate to release enough space for the extra methyl groups to fit (data not shown). SPR experiments with mutated RBP confirm the structural analysis. Surprisingly, the Thr165Met mutant does not belong to the RBP interaction area with VHH5 and exhibits a native-like behavior. Perhaps this mutation affects the binding of the RBP with other phage structural proteins, somehow leading to nonrecognition by VHH5 in vivo. In contrast, the three other mutant RBPs do not bind to VHH5 at all or bind with lower affinity.
It has been shown recently that lactococcal phages belonging to the 936 species, such as bIL170 and 645, bind to cell wall polysaccharides found at the surface of L. lactis (14). Moreover, Geller et al. (16) suggested that the 936-like phages sk1, jj50, and Q64 might not require a host membrane protein or lipoteichoic acid and that cell wall components are sufficient for the initial steps of phage infection. Here, a glycerol molecule was located in the X-ray structure, bound to His 232, Asp 234, Arg 256b, and Trp 244, residues conserved in the list of phages under study (Fig. 3C). Fluorescence quenching experiments confirm this binding site with high affinity to saccharides. The nature of the receptor-binding site is in agreement with polysaccharides acting as receptors at the L. lactis surface. It is reminiscent of sugar-binding sites, such as those observed in lectins (3), which often associate hydrophobic patches of aromatic residues, to accommodate the saccharide's nonpolar face, and polar residues functioning as hydrogen bond donors or acceptors, bringing specificity (4). All saccharides tested here yielded lower Kd values than what is generally observed with monosaccharides. However, these Kd values are restricted within a narrow range and are all close to that of glycerol, making it difficult to identify the natural saccharide receptor. Nonetheless, phosphoglycerol, a component of teichoic or lipoteichoic acids, exhibits an excellent affinity for RBP, a fact that points to teichoic or lipoteichoic acid as a candidate receptor.
In conclusion, we showed that VHH5 has a limited efficacy, since it could neutralize only 12 of the 50 lactococcal phages tested and phages can easily mutate to avoid recognition by the heavy chain antibody fragment. Moreover, we have extended the resolution of the RBP structure to 1.7 and proven the saccharidic nature of the bacterial receptor recognized by lactococcal phages of the 936 species. We propose that teichoic or lipoteichoic acid might be the phage receptor in the L. lactis cell wall. Experiments are under way to validate this hypothesis.
ACKNOWLEDGMENTS
This work was supported, in part, by the Genopole of Marseille-Nice, Unilever, and the Natural Sciences and Engineering Research Council (NSERC) of Canada.
Marie-Therèse Guidici-Orticoni is acknowledged for her help with the SPR experiment, Geneviève Drolet for the neutralization assays, and Hans de Haard, C. Theo Verrips, and Aat M. Ledeboer for providing us with VHH5 and for their interest and support.
These authors contributed equally to the work.
REFERENCES
Bissonnette, F., S. Labrie, H. Deveau, M. Lamoureux, and S. Moineau. 2000. Characterization of mesophilic mixed starter cultures used for the manufacture of aged cheddar cheese. J. Dairy Sci. 83:620-627.
Boucher, I., M. Parrot, H. Gaudreau, C. P. Champagne, C. Vadeboncoeur, and S. Moineau. 2002. Novel food-grade plasmid vector based on melibiose fermentation for the genetic engineering of Lactococcus lactis. Appl. Environ. Microbiol. 68:6152-6161.
Bourne, Y., J. Mazurier, D. Legrand, P. Rouge, J. Montreuil, G. Spik, and C. Cambillau. 1994. Structures of a legume lectin complexed with the human lactotransferrin N2 fragment, and with an isolated biantennary glycopeptide: role of the fucose moiety. Structure 2:209-219.
Bourne, Y., A. Roussel, M. Frey, P. Rouge, J. C. Fontecilla-Camps, and C. Cambillau. 1990. Three-dimensional structures of complexes of Lathyrus ochrus isolectin I with glucose and mannose: fine specificity of the monosaccharide-binding site. Proteins 8:365-376.
Chandry, P. S., S. C. Moore, J. D. Boyce, B. E. Davidson, and A. J. Hillier. 1997. Analysis of the DNA sequence, gene expression, origin of replication and modular structure of the Lactococcus lactis lytic bacteriophage sk1. Mol. Microbiol. 26:49-64.
Collaborative Computational Project, Number 4. 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D 50:760-763.
Crutz-Le Coq, A. M., B. Cesselin, J. Commissaire, and J. Anba. 2002. Sequence analysis of the lactococcal bacteriophage bIL170: insights into structural proteins and HNH endonucleases in dairy phages. Microbiology 148:985-1001.
De Haard, H. J., S. Bezemer, A. M. Ledeboer, W. H. Muller, P. J. Boender, S. Moineau, M. C. Coppelmans, A. J. Verkleij, L. G. Frenken, and C. T. Verrips. 2005. Llama antibodies against a lactococcal protein located at the tip of the phage tail prevent phage infection. J. Bacteriol. 187:4531-4541.
Desmyter, A., S. Spinelli, F. Payan, M. Lauwereys, L. Wyns, S. Muyldermans, and C. Cambillau. 2002. Three camelid VHH domains in complex with porcine pancreatic alpha-amylase. Inhibition and versatility of binding topology. J. Biol. Chem. 277:23645-23650.
Deveau, H., M. R. Van Calsteren, and S. Moineau. 2002. Effect of exopolysaccharides on phage-host interactions in Lactococcus lactis. Appl. Environ. Microbiol. 68:4364-4369.
Devereux, J., P. Haeberli, and O. Smithies. 1984. A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12:387-395.
Duplessis, M., and S. Moineau. 2001. Identification of a genetic determinant responsible for host specificity in Streptococcus thermophilus bacteriophages. Mol. Microbiol. 41:325-336.
Dupont, K., F. K. Vogensen, and J. Josephsen. 2005. Detection of lactococcal 936-species bacteriophages in whey by magnetic capture hybridization PCR targeting a variable region of receptor-binding protein genes. J. Appl. Microbiol. 98:1001-1009.
Dupont, K., F. K. Vogensen, H. Neve, J. Bresciani, and J. Josephsen. 2004. Identification of the receptor-binding protein in 936-species lactococcal bacteriophages. Appl. Environ. Microbiol. 70:5818-5824.
Forde, A., and G. F. Fitzgerald. 1999. Bacteriophage defence systems in lactic acid bacteria. Antonie Leeuwenhoek 76:89-113.
Geller, B. L., H. T. Ngo, D. T. Mooney, P. Su, and N. Dunn. 2005. Lactococcal 936-species phage attachment to surface of Lactococcus lactis. J. Dairy Sci. 88:900-907.
Haggard-Ljungquist, E., C. Halling, and R. Calendar. 1992. DNA sequences of the tail fiber genes of bacteriophage P2: evidence for horizontal transfer of tail fiber genes among unrelated bacteriophages. J. Bacteriol. 174:1462-1477.
Jarvis, A. W., G. F. Fitzgerald, M. Mata, A. Mercenier, H. Neve, I. B. Powell, C. Ronda, M. Saxelin, and M. Teuber. 1991. Species and type phages of lactococcal bacteriophages. Intervirology 32:2-9.
Josephsen, J., and F. K. Vogensen. 1989. Identification of three different plasmid-encoded restriction/modification systems in Streptococcus lactis subsp. cremoris W56. FEMS Microbiol. Lett. 59:161-166.
Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment. Brief Bioinform. 5:150-163.
Labrie, S., and S. Moineau. 2002. Complete genomic sequence of bacteriophage ul36: demonstration of phage heterogeneity within the P335 quasi-species of lactococcal phages. Virology 296:308-320.
Labrie, S., and S. Moineau. 2000. Multiplex PCR for detection and identification of lactococcal bacteriophages. Appl. Environ. Microbiol. 66:987-994.
Ledeboer, A. M., S. Bezemer, J. J. W. de Haard, I. M. Schaffers, C. T. Verrips, C. van Vliet, E.-M. Dusterhoft, P. Zoon, S. Moineau, and L. G. Frenken. 2002. Preventing phage lysis of Lactococcus lactis in cheese production using a neutralizing heavy-chain antibody fragment from llama. J. Dairy Sci. 85:1376-1382.
Loof, M., J. Lembke, and M. Teuber. 1983. Characterisation of the genome of the Streptococcus lactis subsp. diacetylactis bacteriophages P008 widespread in German cheese factories. Syst. Appl. Microbiol. 4:413-423.
McKay, L. L., K. A. Baldwin, and E. A. Zottola. 1972. Loss of lactose metabolism in lactic streptococci. Appl. Microbiol. 23:1090-1096.
Moineau, S. 1999. Applications of phage resistance in lactic acid bacteria. Antonie Leeuwenhoek 76:377-382.
Moineau, S., J. Fortier, and H.-W. Ackermann. 1992. Characterization of lactococcal bacteriophages from Quebec cheese plants. Can. J. Microbiol. 38:875-882.
Moineau, S., D. Tremblay, and S. Labrie. 2002. Phages of lactic acid bacteria: from genomics to industrial applications. ASM News 68:388-393.
Moineau, S., M. Borkaev, B. J. Holler, S. A. Walker, J. K. Kondo, E. R. Vedamuthu, and P. A. Vandenbergh. 1996. Isolation and characterization of lactococcal bacteriophages from cultured buttermilk plants in the United States. J. Dairy Sci. 79:2104-2111.
Murshudov, G., A. A. Vagin, and E. J. Dodson. 1997. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D 53:240-255.
Muyldermans, S., C. Cambillau, and L. Wyns. 2001. Recognition of antigens by single-domain antibody fragments: the superfluous luxury of paired domains. Trends Biochem. Sci. 26:230-235.
Rince, A., M. Tangney, and G. F. Fitzgerald. 2000. Identification of a DNA region from lactococcal phage sk1 protecting phage 712 from the abortive infection mechanism AbiF. FEMS Microbiol. Lett. 182:185-191.
Roussel, A., and C. Cambillau. 1991. The TURBO-FRODO graphics package, vol. 81. Silicon Graphics Geometry Partners Directory, Mountain View, Calif.
Spinelli, S., A. Desmyter, C. T. Verrips, H. J. de Haard, S. Moineau, and C. Cambillau. 2006. Lactococcal bacteriophage p2 receptor-binding protein structure suggests a common ancestor gene with bacterial and mammalian viruses. Nat. Struct. Mol. Biol. 13:85-89.
Valyasevi, R., W. E. Sandine, and B. L. Geller. 1994. Lactococcus lactis ssp. lactis C2 bacteriophage sk1 receptor involving rhamnose and glucose moieties in the cell wall. J. Dairy Sci. 77:1-6.
Walhout, A. J., G. F. Temple, M. A. Brasch, J. L. Hartley, M. A. Lorson, S. van den Heuvel, and M. Vidal. 2000. GATEWAY recombinational cloning: application to the cloning of large numbers of open reading frames or ORFeomes. Methods Enzymol. 328:575-592.(Denise M. Tremblay, Marie)