research paper

document3.pdf

Home >Biology homework help >research paper

The targeted recognition of Lactococcus lactis phages to their polysaccharide receptors

Orla McCabe,1† Silvia Spinelli,2,3† Carine Farenc,2,3

Myriam Labbé,4,5 Denise Tremblay,4

Stéphanie Blangy,2,3 Stefan Oscarson,1* Sylvain Moineau4,5 and Christian Cambillau2,3* 1Centre for Molecular Innovation and Drug Discovery, School of Chemistry and Chemical Biology, University College Dublin, Belfield, Dublin, Ireland. 2Architecture et Fonction des Macromolécules Biologiques, CNRS, Marseille, UMR 7257, France. 3Aix-Marseille University, Campus de Luminy, Case 932, Marseille, 13288 France. 4Groupe de recherche en écologie buccale & Félix d’Hérelle Reference Center for Bacterial Viruses, Faculté de médecine dentaire, Université Laval, Québec, G1V 0A6, Canada. 5Département de biochimie, de microbiologie et de bio-informatique, Faculté des sciences et de génie, Université Laval, Québec, G1V 0A6, Canada.

Summary

Each phage infects a limited number of bacterial strains through highly specific interactions of the receptor-binding protein (RBP) at the tip of phage tail and the receptor at the bacterial surface. Lactococcus lactis is covered with a thin polysaccharide pellicle (hexasaccharide repeating units), which is used by a subgroup of phages as a receptor. Using L. lactis and phage 1358 as a model, we investigated the interaction between the phage RBP and the pellicle hexasaccha- ride of the host strain. A core trisaccharide (TriS), derived from the pellicle hexasaccharide repeating unit, was chemically synthesised, and the crystal structure of the RBP/TriS complex was determined. This provided unprecedented structural details of RBP/receptor site-specific binding. The complete hexasaccharide repeating unit was modelled and found to aptly fit the extended binding site. The speci- ficity observed in in vivo phage adhesion assays could be interpreted in view of the reported structure. There- fore, by combining synthetic carbohydrate chemistry,

X-ray crystallography and phage plaquing assays, we suggest that phage adsorption results from distinct recognition of the RBP towards the core TriS or the remaining residues of the hexasacchride receptor. This study provides a novel insight into the adsorption process of phages targeting saccharides as their receptors.

Introduction

The infection process of viruses is initiated by intermolecu- lar interactions between the viral host recognition device and a receptor usually located at the surface of the host cell. This receptor can be a protein, a polysaccharide or both. For example, using reversible attachment to cell wall saccharides, bacterial viruses (bacteriophages or phages) can scout the host cell surface to locate and irreversibly bind to a specific receptor (Parent et al., 2014). Typical examples include phage T5 (Plancon et al., 2002), which infects Gram-negative Escherichia coli, and phage SPP1 (Alonso et al., 2006), which infects Gram-positive Bacillus subtilis. Phage T5 uses the FhuA porin, an iron-importing membrane protein, as a receptor, binding with sub- nanomolar Kd (Breyton et al., 2013). Phage SPP1 uses the protein YueB, a trans-membrane component of the type VII secretion system, as a receptor and binds to it with sub- nanomolar affinity (Sao-Jose et al., 2006).

On the other hand, most virulent phages infecting the Gram-positive Lactococcus lactis utilise saccharides as specific receptors (Chapot-Chartier et al., 2010; Bebeacua et al., 2013; Ainsworth et al., 2014; Farenc et al., 2014; Spinelli et al., 2014), including the predominant 936 group and the rare 1358 group (Deveau et al., 2006). L. lactis is the most important bacterial species used for cheese manufacture, but the presence of ubiquitous virulent phages in milk may lead to the lysis of lactococcal cells thereby delaying the fermentation process. Accordingly, hundreds of different lactococcal phages have been iso- lated worldwide but surprisingly each replicate within a specific set of L. lactis strains (Mahony and van Sinderen, 2012). The remarkable diversity of lactococcal phage host ranges has been poorly studied but explanations are beginning to emerge.

Lactococcal phages belong to the Caudovirales order and, accordingly, these double-stranded DNA genome-

Accepted 19 February, 2015. *For correspondence. E-mail cambillau @afmb.univ-mrs.fr; [email protected]; Tel. 0033491825590; Fax +35317162318. †These authors contributed equally to the work.

Molecular Microbiology (2015) 96(4), 875–886 ■ doi:10.1111/mmi.12978 First published online 16 March 2015

mailto:[email protected]

containing bacterial viruses use the receptor-binding pro- teins (RBP) located at tip of their tail to recognise receptors on the host (Spinelli et al., 2014). The structure of several RBPs from various lactococcal phages has been solved (Ricagno et al., 2006; Spinelli et al., 2006a,b; Farenc et al., 2014). Among others, it was found that lactococcal RBPs are highly diversified; however, RBPs share a modular structure and a saccharide-binding site located at the C-terminal end that is critical for host binding. This assort- ment of phage RBPs is likely related to the diversity of the phage receptors at the surface of L. lactis cells.

The exact nature of these cell receptors, however, has long been a matter of debate. It was proposed that they may be phosphosaccharides, such as cell wall lipotei- choic acids (Tremblay et al., 2006). These polymers have been found to serve as receptors for various Staphylococ- cus aureus phages (Xia et al., 2010; 2011) but are not diverse enough structurally to promote finely tuned host specificity.

In 2010, a new phosphosaccharide pellicle enveloping L. lactis cells was discovered, and its structure was determined for strain MG1363 (Andre et al., 2010; Chapot-Chartier et al., 2010). The MG1363 pellicle is made of a repeat of a phosphohexasaccharide, possibly linked covalently to the cell wall. Very recently, two other pellicles were analysed, those of L. lactis strain 3107 (Ainsworth et al., 2014) and strain SMQ-388 (Farenc et al., 2014). The differences in their structures (Fig. 1) were ascribed as the trigger of the fine specificity of lactococcal phages for their host (Ainsworth et al., 2014). Comparative analyses led to the hypothesis that a trisaccharide motif (GlcNAc-Galf-GlcNAc-1P or GlcNAc-Galf-Glc-1P) within the pellicle is the core of these phage receptors, whereas strain specificity may be defined by the remaining compo- nents of the pellicle hexasaccharide (Farenc et al., 2014) (All the saccharide residues in this report are pyranosides, with the exception of the furanoside Galf).

We reported recently the X-ray structures of phage 1358 RBP in complex with glycerol, GlcNAcp and Glcp-1P (Farenc et al., 2014). These structures identified two sites

at the RBP surface, ∼ 8 Å apart, one accommodating a GlcNAc monosaccharide, the other a GlcNAc or a Glc-1P monosaccharide. Given that GlcNAc and Glc-1P are com- ponents of the polysaccharide pellicle of L. lactis SMQ- 388 (the host of phage 1358), a Galf sugar bridging the two GlcNAc was modelled, reasoning that the trisaccha- ride motif GlcNAc-Galf-GlcNAc-1P (or Glc-1P) may be common to receptors of genetically distinct lactococcal phages p2, TP901-1 and 1358. Here, we experimentally support this hypothesis by first synthesising the con- served core trisaccharide, and then, demonstrating that it binds to the phage RBP at the same locations as the monosaccharide complexes and in a mode comparable, but slightly distinct, to the previous in silico model. We also propose a molecular model of the complete hexasac- charide binding.

Results

Chemical synthesis of a conserved core trisaccharide

A trisaccharide commonly expressed in the polysaccha- ride pellicle repeating units of L. lactis strains (Fig. 1, bottom) was chemically synthesised from three monosac- charides, D-GlcNAc, D-Galf, and D-Glc, each possessing varied protecting group patterns. As the glucose moiety is α-linked in the phosphodiester bridges, the trisaccharide was synthesised as its α-methyl glucoside in order to maintain the natural stereochemical configuration at the reducing end.

Following a previously reported strategy (Sato et al., 2003), the galactofuranosyl intermediate 4 was syn- thesised from commercially available 1,2:5,6-di-O- isopropylidene-α-D-glucofuranose (1) (Fig. 2A). Com- pound 1 was first treated with triflic anhydride to form the 3-O-triflate derivative 2, which was subsequently heated in DBU/DMSO to give the elimination compound 3 with a 91% yield. The 1,2:5,6-di-O-isopropylidene-α-D- galactofuranose (4) was then afforded following the regio- and stereoselective hydroboration-oxidation of 3. The

Fig. 1. Structure of the phosphosaccharide repeating motifs of the pellicle of the L. lactis strains MG1363 (Chapot-Chartier et al., 2010), 3107 (Ainsworth et al., 2014) and SMQ-388 (Farenc et al., 2014) (from top to bottom) and of the TriS molecule. The three saccharides forming the core of the pellicle are boxed.

876 O. McCabe et al. ■

5,6-O-isopropylidene acetal was selectively cleaved before treatment with NaH and benzyl bromide in DMF to obtain compound 6. Hydrolysis of the 1,2-isopropylidene acetal in 6, followed by benzoylation with BzCl in pyridine, gave compound 7 in an 89% yield. The corresponding thioglycoside was formed by treating 7 with ethanethiol and BF3·OEt2, yielding ethyl 2-O-benzoyl-3,5,6-tri-O-benzyl-1- thio-D-galactofuranoside (8) (Fig. 2B) as an anomeric mixture (α:β, 1:5). The highly reactive nature of compound 7 often resulted in di-thiolation at the anomeric carbon during thioglycoside formation.

The 2-O-benzoyl ester in compound 8 has two functions: to control by means of neighbouring group participation the stereochemistry of newly formed glycosidic bonds in gly-

cosylation using 8 as a glycosyl donor, and to allow for selective deprotection in the presence of benzyl ethers to form a 2-OH glycosyl acceptor (Fig. 2B). The NIS/AgOTf- promoted glycosylation in CH2Cl2 (Konradsson et al., 1990) of methyl glucoside 9 (Zissis and Fletcher, 1970; Garegg et al., 1990) with donor 8 gave the β-linked disac- charide 10 with a 92% yield. Disaccharide acceptor 11 was acquired following debenzoylation using standard Zemplén conditions. Acceptor 11 was then coupled with donor 12 (Cirla et al., 2004) in a subsequent NIS/AgOTf- promoted glycosylation affording the fully protected trisac- charide 13 with an 82% yield.

The phthalimido group was removed using ethylenedi- amine in ethanol at an elevated temperature (→14) before

Fig. 2. Chemical synthesis of a conserved core trisaccharide. A. (i) Tf2O, dry pyridine, dry CH2Cl2, −20°C, 1 h; (ii) DBU, dry DMSO, 85°C, 1.5 h; (iii) 1. BH3·THF, dry THF, 40°C, 4 h, 2. H2O, NaOH, H2O2, rt, 30 min; (iv) 70% AcOH, rt, 6 h; (v) BnBr, NaH dry DMF, rt, 24 h; (vi) (a) 85% AcOH, 70°C, 24 h, (b) BzCl, dry pyridine, rt, 24 h; (vii) EtSH, BF3·OEt2, dry CH2Cl2, 0°C, 15 min. B. (i) NIS, AgOTf, 4Å MS, dry CH2Cl2, 0°C, 45 min; (ii) NaOMe, dry MeOH, rt, 24 h; (iii) NIS, AgOTf, 4Å MS, dry CH2Cl2, −15°C, 40 min. C. (i) (a) EDA, EtOH, 70°C, 24 h, (b) Ac2O, pyridine, rt, 3 h; (ii) Pd/C, H2 (20 bar), EtOAc/EtOH/H2O (6:4:1), rt, 48 h. The subsequent steps of the synthesis are indicated by letters (i), (ii) and (iii).

Receptor binding to lactococcal phage 1358 877

N-acetylation was carried out to produce trisaccharide 15 in an 83% yield over the two steps (Fig. 2C). Complete deprotection was achieved via hydrogenolysis of the benzyl ethers using Pd/C under a hydrogen atmosphere (20 bar). This afforded the desired trisaccharide methyl (2-acetamido-2-deoxy-β-D-glucopyranosyl)-(1→2)-(β-D- galactofuranosyl)-(1→6)-α-D-glucopyranoside (16, TriS) in a 95% yield.

Taken altogether, a trisaccharide (TriS) commonly found in the polysaccharide pellicle repeating units of L. lactis strains was chemically synthesised as its α-methyl glyco- side in six steps from the three monosaccharide interme- diates 8, 9 and 12 in an overall yield of 51%.

Trisaccharide binding to the phage 1358 RBP

As with other lactococcal RBPs, the RBP of phage 1358 assembles as a trimer, with each monomer having a modular structure such as an embedded N-terminal domain and a C-terminal receptor binding ‘head’ domain (Farenc et al., 2014). The crystal structure of phage 1358 RBP in complex with TriS was determined by X-ray diffrac- tion at 2.10 Å resolution. Two RBP monomers are present in the asymmetric unit, each generating a biologically relevant trimer by applying crystallographic threefold sym- metry. A clear electron density was identified in the head domain of both independent monomers in the asymmetric unit, which was large enough to fit the length of approxi- mately three sugar moieties. The molecular structure of TriS obtained from CCP4 Sketcher module (Winn et al., 2011) was readily fitted in the electron density map without ambiguity for the direction of the saccharide chain. Refine- ment and manual rebuilding (Table 1) yielded an excellent model of TriS, although the C5, O5, C6 and O6 atoms of Galf were poorly defined in the electron density map (Fig. S1) and had higher B-factors than the rest of TriS. TriS is located in the middle of the RBP head domain, in a deep crevice (Fig. 3A) in which monosaccharides in complex with the phage RBP were observed (Farenc et al., 2014). Worth noting, TriS occupies about half of this crevice. Furthermore, the cavity becomes wider to the left of the GlcNAc 3 moiety (on Fig. 3B). The role of the remaining free volume of the cavity is discussed later in the text.

The TriS binds to the phage RBP through numerous interactions and its buried surface area in the complex is 254 Å2, 55% of its total accessible surface. Although the Glc 1 and GlcNAc 3 moieties strongly interact with the amino acids of the binding site, the middle Galf is mostly exposed and does not display any direct con- tact with the RBP (Fig. 4, S2). As often observed in other polysaccharide–protein interactions (Bourne et al., 1994a,b), TriS is wrapped around two hydrophobic resi- dues, Phe 240 and 243 (Table 2). Several TriS–RBP interactions are also mediated by chains of water mol-

ecules, also a common feature of protein–sugar interac- tions (Bourne et al., 1990). Glc 1 is strongly anchored in the binding site by hydrogen bonds between O1 and Ser 291 OH, O2 and the guanidinium group of Arg 23, and O3 and the NH2 group of Gln 341. TriS O4 interacts with a water molecule, itself strongly anchored by Arg 312 and Asn 289. The N-acetyl group of GlcNAc 3 is hydrogen- bound to the NH2 group of Gln 345 and the O2 atom with the main-chain NH of Ser 202. The O3 atom interacts with a water molecule bound to the carbonyl group of Val 200. Two water molecules are bound to the O4 and O5 atoms of Galf that are further hydrogen-bound to other atoms of TriS, Glc O5 and GlcNAc O7, respectively, con- tributing to the stabilisation of the saccharide geometry.

When comparing the position of TriS residues 1 and 3 to the previously determined positions of the monosac- charides bound to RBP (Farenc et al., 2014), we noticed that the GlcNAc 3 occupies exactly the same position and has the same orientation as the second GlcNAc from the GlcNAc complex (PDB 4L92) (Fig. 5A). In contrast, the TriS Glc 1 residue is rotated clockwise by ∼ 60° around the sugar center, as compared with the three free saccharides found at this position, GlcNAc, Glc-1P and

Table 1. Data collection and refinement statistics of the phage 1358 RBP crystals.

Data collection ORF20-TriS

PDB 4RGA Source Soleil PX 1 Space group, cell a = b = c (Å) P213, 166.4 Resolution limitsa (Å) 48.0–2.10

(2.16–2.10) Rmeasa (%) 11.0 (75.0) CC(1/2) 99.8 (85) Total reflectionsa 1012656 (82404) Unique reflectionsa 89232 (7218) Mean((I)/sd(I))a 12 (2.1) Completenessa (%) 99.9 (98.5) Multiplicitya 11.4 (11.4)

Refinement

Resolutiona (Å) 48.0–2.10 (2.15–2.10)

Nr of reflectionsa 89232 (6540) Nr protein/water/ligand 6078/927/76 Nr test set reflections 4462 (327) Rwork/Rfreea (%) 18.1/19.2

(18.9/20.6) r.m.s.d.bonds (Å)/angles (°) 0.010/1.12 B-wilson Å2 39.7 B-atoms (A/B) 46.0/45.5 B-ligands (A/B) 61.0/58.3 B-Waters 58.0 Ramachandran: preferred/allowed/outliers b (%) 97.0/2.48/0.52

a. Numbers in brackets refer to the highest resolution bin. b. Amino-acids Tyr182 and Gly317 are slightly outside the allowed area of the Ramachandran plot but have a clear electron density map. In complex with the trisaccharide βGlcNAc-Galf-αGlcNAc-OMe.

878 O. McCabe et al. ■

GlcNAc-1P [PDB IDs 4L92, 4L97, 4RGG (Farenc et al., 2014)]. Therefore, each Ci atom of the TriS Glc 1 residue occupies the position of the Ci+1 atom in the monosaccha- ride complexes. This rotation is indeed induced by the geometry constraint arising from the Glc-Galf linkage. It is worth noting that, despite this rotation, both saccharides keep the same hydrogen-bonds with the protein, but not with the same OH groups. Furthermore, the in silico sub- stitution of a NAc moiety at C2 of TriS Glc 1, presenting a GlcNAc residue in position 1, did not provoke steric clashes with the protein, aptly positioning the NAc group in a small pocket (Fig. 5B). The O7 atom of NAc replaces a water molecule in the experimental structure and pro- vides an extra hydrogen-bond with Asn 289 NH2 group.

Noteworthy, superposition of the RBP structure in complex with TriS with those of the complexes with GlcNAc, Glc1P and GlcNAc-1P yielded very low root mean square (r.m.s.) deviations, between 0.09 to 0.15 Å.

A closer look at the saccharide binding sites confirm their lack of conformational changes, also including the side- chains of the binding site. The r.m.s. deviation observed with the native RBP (in another crystal form) is slightly larger (0.45 Å). However, the saccharide binding site resi- dues are superimposable, including their side-chains.

Hexasaccharide modeling and docking into 1358 RBP binding site

We mentioned above that the TriS seemed to occupy only a part of the binding crevice volume identified at the RBP surface (Figs 3 and 6A). Because the structure of the surface pellicle of L. lactis SMQ-388 (Fig. 1, top), the host of phage 1358, has been determined (Farenc et al., 2014), we constructed this hexasaccharide in silico by adding monosaccharides 4, 5 and 6 (Fig. 6B) to TriS with the correct glycosidic linkages aided by the CPP4 option ‘sketcher’ (Winn et al., 2011) and with Coot (Emsley et al., 2010). The orientation of the 6-OH of GlcNAc 3 was turned, as this group, turned inside the crevice, had to be substi- tuted with Glc 6 (Figs 3B and 6A). The SMQ-388 hexasac- charide was the docked into the RBP crevice, and REFMAC5 (Murshudov et al., 2011) was used to optimise the docking by removing close contacts. The resulting model showed that the overall position of TriS remained unchanged and that the three added saccharides comple- mented the binding cavity shape. At the reducing extremity, the OMe moiety, a PO3 group in the pellicle, is favorably exposed to solvent, allowing the next pellicle phos- phohexasaccharides to be positioned freely (Fig. 6B, red arrow right). At the non-reducing end, the O3 atom of GlcNAc 5 is also accessible to solvent, indicating that the pellicle chain can be continued (Fig. 6B, red arrow left).

Phages–host interactions

Three different surface pellicle polysaccharides have been identified and analysed in three lactococcal strains,

Fig. 3. X-ray structure of the phage 1358 RBP in complex with the TriS molecule. Representation of the TriS molecule (sticks) in the complete trimeric RBP surface (blue, green and pink). The head domain is situated above the horizontal blue line. Inset: close-up of the binding site. GlcNAc, Galf and Glc saccharide atoms are identified by their number. Figure made with Pymol (Pymol, 2014). This figure is available in colour online at wileyonlinelibrary.com.

Table 2. Interactions between bound saccharides and the phage 1358 RBP.

GlcNAc-Galf- Glc-OMe BSA Å2; % of ASA

GlcNAc hydrogen bonds length (Å)

Total 254; 55% Val 200 # 13.4 Gln 201 # 4.4 Ser 202 # 5.4 N-H-OH23;2.8 Arg 237 • 12.6; 24% NH1,2-HO3 3.1; 3.2 Phe 240 • # 37; 50% C=O-OH24; 2.9. C=O-OH28; 2.8 Asp 243 # Asn 244 • # 3.4; 75% C=O-HO3; 3.35 Asn 289 • 25; 41% Ser 291 • 22; 25% Oγ-OH1; 2.81 Ala 292 • 13; 52% Asn 341 • 8.7; 95% C=O-HO3; 2.6 Phe 343 • 16; 78% Gln 345 # 11; C=O-HO27; 2.8

ASA, accessible surface area; BSA, buried surface area.

Receptor binding to lactococcal phage 1358 879

http://wileyonlinelibrary.com

Fig. 4. Details of the molecular interaction of TriS with the active site side-chains. A. Ligplot+ drawing (Laskowski and Swindells, 2011) of the interactions of Tris with protein and water. Only the ligand and protein bound water molecule are displayed. They are conserved between the two independent sites. B. The TriS moiety is represented as sticks (C, yellow; O, red; N, blue). The RBP backbone is drawn in cartoon representation (green), the side chains of the RBP binding sites residues are represented as sticks (C, white; O, red; N, blue). Blue transparent lines illustrate the hydrogen bonds. The red spheres are water molecules. Their lengths are indicated in Å. Figure made with Pymol (Pymol, 2014). This figure is available in colour online at wileyonlinelibrary.com.

880 O. McCabe et al. ■

http://wileyonlinelibrary.com

MG1363, 3107 and SMQ-388. As indicated in Fig. 1, the pellicle of strains L. lactis 3107 and SMQ-388 share almost the same trisaccharide core (GlcNAc-Galf-GlcNAc1P), whereas strain MG1363 is slightly different (GlcNAc-Galf- Glc1P). Phages 1358 infect its host SMQ-388 as well as strain 3107 but not MG1363 (data not shown). On the other hand, phage p2 infects only the strain MG1363 and phage TP901-1 only the strain 3107(data not shown).

We then performed phage adsorption assays in which we measured the percentage of phage adsorbed to the cells after an incubation of 15 min (see Materials and methods). As expected, phage p2 adsorbed only to its host strain MG1363, whereas TP901 adsorbed mainly to its host 3107 (Table 3). Phage 1358 had a lower adsorption to its hosts SMQ-388 as compared with the other two phage– host pair. In agreement, the adsorption of phage 1358 to

L. lactis 3107 was also low (Table 3). Of note the adsorp- tion level increased when phage 1358 was incubated for a longer period of time with its host (data not shown).

Discussion

Phage–host interactions are influenced by a number of factors including the presence of various phage defense mechanisms in lactococcal strains, which may prevent the lysis of a phage-infected cell (Labrie et al., 2010; Samson and Moineau, 2013). Recognition and binding of the phage to its host cell is key to the initiation of these interactions.

TriS saccharide synthesis

Although the hexa/pentasaccharide motifs of the three pellicles could be purified directly from the L. lactis strains, the quantities available were not enough for X-ray crys- tallographic studies. Thus, we decided to embark on a synthetic approach, which comprise the ability to generate not only the pellicle motifs, but also fragments and deriva- tives thereof. The TriS synthesis, reported here, shows the usefulness of this approach in an efficient synthesis. Starting from appropriately protected monosaccharide building blocks, the target trisaccharide was produced on a ∼ 20 mg scale. These building blocks, with minor adjust- ments, are now being further used in the continued syn- theses of other pellicle oligosaccharide structures to assist in future structural and biophysical studies.

The RBP structure of phage 1358 and its saccharide-binding site

The RBP from phage 1358 displays large differences from those of phages p2 and TP901-1 (Fig. S3A–C). In particu- lar, the receptor binding site has been located between two monomers in phages p2 and TP901-1 (Fig. S3D and E) while it is in the middle of the monomers of the RBP of phage 1358 (Fig. S3F). The saccharide-binding site of phage 1358 is an elongated crevice, which can be filled in part by our synthetic TriS. This TriS defines a core trisac- charide present in the three wall polysaccharide (WPS) identified to date, the position 1 being Glc-1P or GlcNAc-

Fig. 5. Analysis of the TriS geometry. A. TriS geometry and position compared to those of GlcNAc monosaccharides bound to the RBP of 1358. Sticks representations and atom mode coloring (Tris: C, white; O, red; N, blue. GlcNAc: C, orange; O, red; N, blue ). B. TriS is compared with a model in which Glc 1 has been replaced by a GlcNAc saccharide. The RBP surface is green. Stick representations and atom mode coloring (TriS: C, white; O, red; N, blue. GlcNAc Model: C, yellow; O, red; N, blue). Figure made with Pymol (Pymol, 2014). This figure is available in colour online at wileyonlinelibrary.com.

Table 3. Adsorption assay of phages p2, TP901-1 and 1358 on the Lactococcus lactis strains MG1363, 3107 and SMQ-388.

Percentage of phage adsorption

Strain/phage p2 TP-901 1358 MG1363 82.7 ± 9.4 21.5 ± 5.0 18.7 ± 8.0 3107 0 85.2 ± 7.8 31.5 ± 11.2 SMQ-388 0 5.8 ± 5.0 60.4 ± 9.8

Bold numbers refer to the host’s specific phage.

Receptor binding to lactococcal phage 1358 881

http://wileyonlinelibrary.com

1P. TriS (or TriS-NAc) would therefore be a seeding module, attaching to all RBPs binding site, whereas selec- tivity might occur from the rest of the hexasaccharide pellicle motif. In this context, we notice that two main differences between the three WPS: position 4 exhibits a Rha (MG1363) or a Galf (SMQ-388, 3107), both 1→3 linked to the core trisaccharide with elongation through position 3. However, with the Galf residue being a 5-membered ring, the angle formed by the saccharides bound to Galf 4 is smaller than in the case of a pyranose saccharide (6-membered ring). The Galf at position 4 may, therefore, introduce steric clashes with the RBP binding site of a WPS possessing GlcNAcp at this position. The second large difference is the substitution of a Glc at 6-OH of GlcNAc 3. This substitution would introduce steric clashes in a site for which the specific WPS is not substi- tuted. In contrast, binding a non-substituted WPS to a site accepting Glc substituted WPS would be less dramatic, the only difference being probably a small loss in extra binding energy arising from interactions of the branched Glc in the RBP site.

The adsorption of phage 1358 to L. lactis strains

The adsorption of phage 1358 at the surface of the L. lactis strain SMQ-388 results from a compatibility between the

pellicle repeating phosphohexasaccharide and the binding crevice of the phage RBP. The hexasaccharide spans the width of the binding crevice, leaving the terminal ends exposed and allowing the rest of the polymer to be posi- tioned freely (Fig. 6B). Furthermore, all the nooks of the binding cavity are filled, leading to optimal interaction energy. Concerning the lower adsorption of phage 1358 at the surface of its other host L. lactis strain 3107, we noticed that the only differences between both receptors are the absence of a branched Glc 6 and a Glc 5 instead of GlcNAc 5 in strain 3107. We suggested above that the absence of branched Glc should have minor effects (less interaction energy). Furthermore, in our hexasaccharide model, a NAc group at position 2 of Glc 5 would be exposed to solvent. Phage 1358 does not infect L. lactis MG1363, and its adsorption level was the lowest (Table 1). We noticed a unique difference between both saccharide motifs (besides Glc 1), a Rha instead of Galf in strain MG1363. The angle between the saccharides attached to Rha should be larger than for Galf, hence we observe in our model that Rha 4 and GlcNAc 5 of the MG1363 pellicle are further from the RBP crevice and thus less prone to estab- lishing favourable interactions. However, this kind of sub- stitution (not the contrasting one) does not induce steric clashes but only a milder loss of favourable interactions. As we do not know the structure of a complex between TriS

Fig. 6. A model of the SMQ-388 hexasaccharide in interaction with the RBP. A. The original TriS structure in sphere representation at the RBP surface. Inset: close-up of the binding crevice; saccharides are numbered 1–3 and their sequence is given below. B. The modelled SMQ-388 hexasaccharide structure in sphere representation at the RBP surface. Inset: close-up of the binding crevice; saccharides are numbered 1–6 and the sequence of the modelled motif is given below. The two red arrows indicate the spread of the polysaccharide at the reducing and non-reducing end. Figure made with Pymol (Pymol, 2014).

882 O. McCabe et al. ■

and the p2 and TP901-1 RBPs, analysis of the adsorption data in terms of pellicle/RBP interaction is not possible at this time.

Conclusion

In summary, these results provide a better understanding of the adhesion properties of phage 1358 towards its L. lactis host strains, as well as phage–host interactions in general. Overall, our data suggest that the binding mechanism of a lactococcal phage to its host may involve two different interactions of the RBP with the hexasaccha- ride receptor: the TriS moiety, being similar in the three receptors sequenced, may provide a non-specific initial interaction, whereas the rest of the hexasaccharide may increase the interaction for strain-specific phages, or abolish it for non-specific phages.

These results also provide a stimulating illustration of what can be achieved by combining synthetic carbohy- drate chemistry, X-ray protein crystallography analysis and in vivo viral infections. Finally, the TriS synthesis paves the way for the synthesis of larger pellicle saccharides with different specificities, making it possible to experimentally verify our suggested recognition mechanism.

Experimental procedures

Chemical synthesis of a conserved core trisaccharide

Unless noted, chemical reagents and solvents were used without further purification from commercial sources. Reac- tions were magnetically stirred. Concentration in vacuo was generally performed using a Buchi rotary evaporator. The 1H/13C NMR spectra (δ in ppm, relative to TMS in CDCl3) were recorded with Varian spectrometers (Varian, Palo Alto, CA, USA) (400/101 MHz or 500/125 MHz) at 25°C. Assignments were aided by 1H-1H and 1H-13C correlation experiments. HRMS spectra were recorded on a micromass LCT instru- ment from Waters. Optical rotations were recorded on a Perkin-Elmer polarimeter (Model 343) at the sodium D-line (589 nm) at 20°C using a 1 dm cell and are not corrected. Silica gel chromatography was carried out using Davisil LC60A (Grace tech., Columbia, MD, USA) SiO2 (40–63 μm) silica gel. All reactions were monitored by thin-layer chroma- tography (TLC). TLC was performed on Merck DC-Alufolien plates precoated with silica gel 60 F254. They were visual- ised with UV-light (254 nm) fluorescence quenching, and/or by charring with an 8% H2SO4 dip (stock solution: 8 ml conc. H2SO4, 92 ml EtOH), and/or ninhydrin dip (stock solution: 0.3 g ninhydrin, 3 ml AcOH, 100 ml EtOH).

1,2:5,6-Di-O-isopropylidene-3-O-triflate-α-D-glucofuranose (2). Pyridine (1.4 ml, 17.7 mmol) was added to a solution of 1,2:5,6-di-O-isopropylidene-α-D-glucofuranose 1 (2.0 g, 7.7 mmol) in dry CH2Cl2 (60 ml). The mixture was cooled to −20°C, and a solution of triflic anhydride (1.6 ml, 9.2 mmol) in CH2Cl2 (20 ml) was added over 30 min. The reaction was

stirred at −20°C for an additional 30 min, before being diluted with CH2Cl2 (60 ml) and washed successively with H2O (60 ml), sat. aq. NaHCO3 (60 ml) and H2O (60 ml). The organic layer was dried over MgSO4, filtered and concen- trated under diminished pressure to give a yellow solid. The crude product was used for the next step without purification. Rf 0.79 (toluene/EtOAc, 1:1 +1% Et3N).

3-Deoxy-1,2:5,6-di-O-isopropylidene-α-D-erythro-hex-3-eno- furanose (3). Compound 2 (3.0 g, 7.7 mmol) was dissolved in dry DMSO (90 ml). DBU (1.4 ml, 9.2 mmol) was added, and the reaction was stirred at 85°C for 1.5 h. Upon com- pletion, the mixture was poured onto H2O (200 ml) and extracted with EtOAc (6 × 50 ml). The organic phases were combined, dried over MgSO4, filtered and evaporated in vacuo. The crude product was then purified by silica gel chromatography (toluene/EtOAc, 6:1 v/v + 1% Et3N), which gave compound 3 (1.69 g, 91% over two steps) as an off- white solid.

1,2:5,6-Di-O-isopropylidene-α-D-galactofuranose (4). Com- pound 3 (200 mg, 0.8 mmol) was dissolved in dry THF (20 ml) and cooled to 0°C. A solution of BH3·THF in THF (1M, 1.7 ml) was added, and the resulting solution was stirred at 40°C. After stirring for 3 h, a substantial amount of starting material was still detected by TLC (toluene/EtOAc, 4:1 v/v). The reaction was cooled to 0°C and additional BH3·THF in THF (1M, 2.0 ml) was added. The reaction was stirred at 40°C for 1 h. Once complete, the solution was cooled to 0°C and H2O (0.8 ml), 10% aq NaOH (2.4 ml) and 35% aq H2O2 (6.0 ml) were sequentially added. The mixture was stirred at room temperature for 30 min. H2O (20 ml) was added, and the aqueous solution was extracted with CH2Cl2 (3 × 20 ml). The organic phase was combined, dried over MgSO4, filtered and concentrated in vacuo. Purification by silica gel chroma- tography (toluene/EtOAc, 2:1 v/v) was carried out, yielding 4 (131 mg, 61%) as a white solid.

1,2-O-Isopropylidene-α-D-galactofuranose (5). Compound 4 (2.0 g, 7.7 mmol) was treated with 70% aq AcOH (30 ml). The solution was stirred at room temperature for 6 h. Once com- plete, the mixture was concentrated in vacuo followed by co-evaporation with H2O and toluene, successively. The crude product was then purified by silica gel chromatography (EtOAc/MeOH, 9:1 v/v) giving compound 5 (1.33 g, 79%) as a white solid.

1,2-O-Isopropylidene-3,5,6-tri-O-benzyl-α-D-galactofuranose (6). A solution of compound 5 (667 mg, 3.0 mmol) and BnBr (2.2 ml, 18.2 mmol) in dry DMF (20 ml) was added to a cooled suspension of 60% NaH (848 mg, 21.0 mmol) in dry DMF (10 ml) at 0°C. The reaction was stirred at room temperature overnight. MeOH was added at 0°C, to quench excess NaH, and the mixture was concen- trated in vacuo. The residue was diluted with H2O (40 ml) and extracted with CH2Cl2 (10 ml × 4). The organic phases were combined and washed with sat. aq. NaHCO3 (2 × 20 ml) and H2O (20 ml). The organic layer was dried over MgSO4, filtered and concentrated in vacuo. Purification by silica gel chromatography (toluene/EtOAc, 15:1 v/v) afforded 6 (1.38 g, 93%).

Receptor binding to lactococcal phage 1358 883

1,2-O-Benzoyl-3,5,6-tri-O-benzyl-D-galactofuranose (7). Compound 6 (1.83 g, 3.7 mmol) was dissolved in 85% aq AcOH (24 ml) and stirred at 70°C overnight. The solution was poured onto excess sat. aq NaHCO3 and extracted with CH2Cl2 (20 ml × 6). The organic extracts were dried over MgSO4, filtered and concentrated in vacuo. The crude product was used for the next step without purification. Rf 0.38 (toluene/EtOAc, 6:1). BzCl (1.7 ml, 15.0 mmol) was added to a cooled solution (0°C) of the crude product in dry pyridine (20 ml). The resulting mixture was stirred at room temperature overnight. The solvent was evaporated under diminished pressure and the residue was re-dissolved in EtOAc (40 ml). The solution was washed with H2O (20 ml), 1M HCl (2 × 20 ml), sat aq NaHCO3 (2 × 20 ml) and H2O (20 ml). The organic phase was then dried over MgSO4, filtered and concentrated in vacuo. Purification by silica gel chromatography (cyclohexane/EtOAc, 12:1–10:1 v/v) gave 7 (2.23 g, 91%) as a white solid.

Ethyl 2-O-benzoyl-3,5,6-tri-O-benzyl-1-thio-D-galactofurano- side (8). Compound 7 (2.0 g, 3.0 mmol) was dissolved in dry CH2Cl2 (30 ml) and cooled to 0°C. Ethanethiol (1.1 ml, 15.2 mmol) was added, and the mixture was stirred for 15 min. BF3·OEt2 (0.6 ml, 4.6 mmol) was added dropwise, and the reaction was stirred for 1 h at 0°C. Upon completion, the solution was neutralised with Et3N, and the solvent was concentrated in vacuo. Silica gel chromatography (cyclohexane/EtOAc, 20:1–15:1 v/v) of the crude product gave 8 (1.14 g, 63%) as an anomeric mixture (β:α = 5:1).

Methyl (2-O-benzoyl-3,5,6-tri-O-benzyl-β-D-galactofuranosyl)- (1→6)-2,3,4-tri-O-benzyl-α-D-glucopyranoside (10). A solu- tion of methyl 2,3,4-tri-O-benzyl-α-D-glucopyranoside (9,87 mg, 0.19 mmol), compound 8 (135 mg, 0.23 mmol) and 4Å MS (350 mg) in dry CH2Cl2 (5 ml) was cooled to 0°C. The mixture was stirred for 15 min before addition of N-iodosuccinimide (63 mg, 0.28 mmol) and AgOTf (12 mg, 0.05 mmol). The reaction was stirred at 0°C for 45 min and then neutralised with Et3N. The mixture was filtered through Celite (Imerys Mineral, CA, USA) and washed with 10% aq Na2S2O3 (5 ml) and H2O (5 ml). The organic phase was dried over MgSO4, filtered and concentrated in vacuo. The crude product was subjected to silica gel chromatography (cyclohexane/EtOAc, 10:1–5:1 v/v) to obtain disaccharide 10 (173 mg, 92%).

Methyl (3,5,6-tri-O-benzyl-β-D-galactofuranosyl)-(1→6)-2,3, 4-tri-O-benzyl-α-D-glucopyranoside (11). Disaccharide 10 (158 mg, 0.16 mmol) was dissolved in dry MeOH (1.6 ml) and cooled to 0°C. NaOMe was added until a pH of 12 was reached. The solution was stirred at room temperature over- night. Dowex-50WX8 resin was used to neutralise the reac- tion. The mixture was filtered and concentrated in vacuo. Purification by silica gel chromatography (toluene/EtOAc, 8:1–6:1 v/v) gave 11 (122 mg, 86%).

Methyl (3,4,6-tri-O-benzyl-2-deoxy-2-phthalimido-β-D-glu- copyranosyl)-(1→2)-(3,5,6-tri-O-benzyl-β-D-galactofurano- syl)-(1→6)-2,3,4-tri-O-benzyl-α-D-glucopyranoside (13). A solution of ethyl 3,4,6-tri-O-benzyl-2-deoxy-2-phthtalimido-1- thio-β-D-glucopyranosyl 12 (88 mg, 0.14 mmol), disaccha-

ride acceptor 11 (85 mg, 0.10 mmol) and 4Å MS (230 mg) in dry CH2Cl2 (4.5 ml) was cooled to −15°C. The mixture was stirred for 15 min before addition of NIS (32 mg, 0.14 mmol) and AgOTf (6 mg, 0.02 mmol). The reaction was stirred at −15°C for 40 min and then neutralised with Et3N. The mixture was filtered through Celite and washed with 10% aq Na2S2O3 (3 ml) and H2O (3 ml). The organic phase was dried over MgSO4, filtered and concentrated in vacuo. The crude product was subjected to silica gel chromatography (cyclohexane/EtOAc, 8:1–6:1–5:1 v/v) to obtain trisaccharide 13 (112 mg, 82%).

Methyl (2-amino-3,4,6-tri-O-benzyl-2-deoxy-β-D-glucopy- ranosyl)-(1→2)-(3,5,6-tri-O-benzyl-β-D-galactofuranosyl)- (1→6)-2,3,4-tri-O-benzyl-α-D-glucopyranoside (14). Ethyl- enediamine (0.14 ml, 2.1 mmol) was added to a solution of disaccharide 13 (76 mg, 0.05 mmol) in EtOH (4 ml). The reaction was stirred at 70°C overnight. The mixture was co-evaporated with acetonitrile under diminished pressure. Purification by silica gel chromatography (toluene/EtOAc, 6:1 v/v +1% Et3N) afforded 14 (57 mg, 83%).

Methyl (2-acetamido-3,4,6-tri-O-benzyl-2-deoxy-β-D-glu- copyranosyl)-(1→2)-(3,5,6-tri-O-benzyl-β-D-galactofurano- syl)-(1→6)-2,3,4-tri-O-benzyl-α-D-glucopyranoside (15). Acetic anhydride (8 μl, 0.08 mmol) was added to a solution of trisaccharide 14 (54 mg, 0.04 mmol) in pyridine (0.5 ml) at room temperature. After 3 h of stirring, the solvent was removed under reduced pressure, and the residue was re-dissolved in CH2Cl2 (2 ml). The solution was washed with 1M HCl (1 ml), sat aq NaHCO3 (1 ml) and H2O (1 ml). The organic phase was dried over MgSO4, filtered and concen- trated in vacuo. Silica gel chromatography (toluene/EtOAc, 6:1 v/v) of the crude product gave compound 15 (51 mg, 91%).

Methyl (2-acetamido-2-deoxy-β-D-glucopyranosyl)-(1→2)- (β-D-galactofuranosyl)-(1→6)-α-D-glucopyranoside (16). Pd/C 10% molar (33 mg, 0.31 mmol) was added to a solution of trisaccharide 15 (47 mg, 0.034 mmol) in EtOAc/EtOH/H2O (0.5 ml, 6:4:1). The reaction was stirred vigorously under an atmosphere of H2 gas (20 bar) at room temperature for 2 days. The mixture was filtered through a pad of Celite and concentrated in vacuo. Reverse phase chromatography gave trisaccharide 16 (19 mg, 95%) as a white solid.

RBP/Trisaccharide crystallisation. The orf-20 of phage 1358 was cloned into the Gateway™ (Invitrogen, Grand Island, NY, USA) destination vector pETG-20A for protein production in E. coli BL21, purified by Ni affinity, and gel filtration chroma- tography according to standard procedures (Vincentelli et al., 2003; 2005) as previously described (Farenc et al., 2014). A solution of RBP in 10 mM HEPES, pH 7.5, 150 mM NaCl was mixed with TriS (dissolved in water) to final concentrations of 13 mg ml−1 of RBP and 100 mM of TriS. Cocrystals were obtained by mixing 300 nl of the protein/TriS solution with a precipitant solution containing 10 mM Zinc Sulfate Heptahy- drate, 100 mM MES buffer pH 6.5, and 25 % v/v polye- thyleneglycol monomethyl ether 550. Crystals were cryopro- tected with the mother liquor supplemented with 15% poly- propylene glycol and immediately flash-frozen under a stream of nitrogen.

884 O. McCabe et al. ■

X-ray crystallography. Crystals were grown by hanging-drop vapor diffusion, and datasets were collected at the synchro- tron Soleil (PROXIMA-1). Data were treated by XDS and XSCALE (Kabsch, 2010). The structure was determined by molecular replacement with Molrep (Vagin and Teplyakov, 2010) using the apo RBP structure as a search model (Farenc et al., 2014). The crystals belong to the cubic space group P213 with unit cell dimension of a = b = c = 165.4 Å, and a Vm of 4.7 with 74% solvent for two monomers in the asymetric unit. They diffracted to a resolution of 2.1 Å (Table 1). Each of the two monomers in the asymmetric unit is part of a trimer that can be reconstituted by the threefold crystallographic axis. Refinement was performed with AutoBUSTER (Blanc et al., 2004) alternated with manual reconstruction with Coot (Emsley et al., 2010). TriS was built from monosaccharides and regularised with Coot (Emsley et al., 2010). The omit map was calculated after four cycles of Cartesian simulated annealing using Phenix (Adams et al., 2010). Figures were made with the molecular graphics programme Pymol (Pymol, 2014).

Molecular modelling. The hexasaccharide was modelled by extension of the trisaccharide. The coordinates of the extra residues were obtained from the trisaccharide itself or from high-resolution structures of protein–saccharide complexes in the Protein Data Bank (PDB). A unique Galf structure was found from the entry 2VK2. The TriS molecule was generated by the CCP4 building option ‘Sketcher’ according to standard bond lengths. The resulting structure was docked ‘manually’ (by rotation-translation of the whole mol- ecule and rotation of the dihedral angles) into the phage 1358 RBP receptor-binding groove using Coot (Emsley et al., 2010). The resulting complex was idealised using REFMAC5 (Murshudov et al., 2011) in refinement mode in order to obtain a reasonable and clashless model as well as favourable contacts. Figures were made with the molecular graphics programme Pymol (Pymol, 2014).

Phages adsorption to L. lactis strains

Phage adsorption assays were performed as described pre- viously (Sanders and Klaenhammer, 1980) with the following modifications. One hundred microlitres of phage (104 pfu ml−1) were mixed with 900 μl of bacteria (OD600 of 0.6 to 0.8). After incubation at 30°C for 10 min, the mixture was centrifuged at 16,000 × g for 1 min. The supernatant was then titrated. The percentage of adsorption was calculated with the formula: 100 × ((phage titer in adsorption assay without bacteria – phage titer in supernatant after adsorption assay) / Phage titer in adsorption assay without bacteria). All the assays were performed in triplicates.

Accession numbers

X-ray structure and structure factors were deposited in the Protein Data Bank with the ID code 4RGA.

Conflict of interest

None

Funding

This work was supported by a grant from the Agence Nation- ale de la Recherche (grants ANR-11-BSV8-004-01) and grants from Science Foundation Ireland (Grants 08/RSC/ B1393, 08/In.1/B2067, and 13/IA/1959). SM acknowledges funding from NSERC of Canada (Strategic program). SM holds a Tier 1 Canada Research Chair in Bacteriophages.

Acknowledgements

We thank the synchroron Soleil (Saint-Aubin, France) for beam time allocation, and the staff of Proxima 1 beamline for their assistance.

References

Adams, P.D., Afonine, P.V., Bunkoczi, G., Chen, V.B., Davis, I.W., Echols, N., et al. (2010) PHENIX: a comprehensive Python-based system for macromolecular structure solu- tion. Acta Crystallogr D Biol Crystallogr 66: 213–221.

Ainsworth, S., Sadovskaya, I., Vinogradov, E., Courtin, P., Guerardel, Y., Mahony, J., et al. (2014) Differences in lactococcal cell wall polysaccharide structure are major determining factors in bacteriophage sensitivity. MBio 5: e00880–14.

Alonso, J.C., Tavares, P., Lurz, R., and Trautner, T.A. (eds) (2006) Bacteriophage SPP1. New York: Oxford University Press.

Andre, G., Kulakauskas, S., Chapot-Chartier, M.P., Navet, B., Deghorain, M., Bernard, E., et al. (2010) Imaging the nanoscale organization of peptidoglycan in living Lactococ- cus lactis cells. Nat Commun 1: 1–8.

Bebeacua, C., Tremblay, D., Farenc, C., Chapot-Chartier, M.P., Sadovskaya, I., van Heel, M., et al. (2013) Structure, adsorption to host, and infection mechanism of virulent lactococcal phage p2. J Virol 87: 12302–12312.

Blanc, E., Roversi, P., Vonrhein, C., Flensburg, C., Lea, S.M., and Bricogne, G. (2004) Refinement of severely incom- plete structures with maximum likelihood in BUSTER-TNT. Acta Crystallogr D Biol Crystallogr 60: 2210–2221.

Bourne, Y., Rouge, P., and Cambillau, C. (1990) X-ray struc- ture of a (alpha-Man(1–3)beta-Man(1–4)GlcNAc)-lectin complex at 2.1-A resolution. The role of water in sugar- lectin interaction. J Biol Chem 265: 18161–18165.

Bourne, Y., Bolgiano, B., Liao, D.I., Strecker, G., Cantau, P., Herzberg, O., et al. (1994a) Crosslinking of mammalian lectin (galectin-1) by complex biantennary saccharides. Nat Struct Biol 1: 863–870.

Bourne, Y., Mazurier, J., Legrand, D., Rouge, P., Montreuil, J., Spik, G., and Cambillau, C. (1994b) Structures of a legume lectin complexed with the human lactotransferrin N2 frag- ment, and with an isolated biantennary glycopeptide: role of the fucose moiety. Structure 2: 209–219.

Breyton, C., Flayhan, A., Gabel, F., Lethier, M., Durand, G., Boulanger, P., et al. (2013) Assessing the conformational changes of pb5, the receptor-binding protein of phage T5, upon binding to its Escherichia coli receptor FhuA. J Biol Chem 288: 30763–30772.

Chapot-Chartier, M.P., Vinogradov, E., Sadovskaya, I.,

Receptor binding to lactococcal phage 1358 885

Andre, G., Mistou, M.Y., Trieu-Cuot, P., et al. (2010) Cell surface of Lactococcus lactis is covered by a protective polysaccharide pellicle. J Biol Chem 285: 10464–10471.

Cirla, A., McHale, A.R., and Mann, J. (2004) Synthesis of analogues of calicheamicin and neocarzinostatin chromo- phore. Tetrahedron Lett 60: 4019–4029.

Deveau, H., Labrie, S.J., Chopin, M.C., and Moineau, S. (2006) Biodiversity and classification of lactococcal phages. Appl Environ Microbiol 72: 4338–4346.

Emsley, P., Lohkamp, B., Scott, W.G., and Cowtan, K. (2010) Features and development of Coot. Acta Crystallogr D Biol Crystallogr 66: 486–501.

Farenc, C., Spinelli, S., Vinogradov, E., Tremblay, D., Blangy, S., Sadovskaya, I., et al. (2014) Molecular insights on the recognition of a Lactococcus lactis cell wall pellicle by the phage 1358 receptor binding protein. J Virol 88: 7005– 7015.

Garegg, P.J., Oscarson, S., and Szönyi, M. (1990) Synthesis of methyl 3-o-(alpha-d-glucopyranosyl)-7-o-(l-glycero- alpha-d-manno-heptopyranosyl)-l-glycero-alpha-d-manno- heptopyranoside. Carbohydr Res 205: 125–132.

Kabsch, W. (2010) XDS. Acta Crystallogr D Biol Crystallogr 66: 125–132.

Konradsson, P., Udodong, U.E., and Fraser-Reid, B. (1990) Iodonium promoted reactions of disarmed thioglycosides. Tetrahedron Lett 31: 4313–4316.

Labrie, S.J., Samson, J.E., and Moineau, S. (2010) Bacterio- phage resistance mechanisms. Nat Rev Microbiol 8: 317– 327.

Laskowski, R.A., and Swindells, M.B. (2011) LigPlot+: multi- ple ligand-protein interaction diagrams for drug discovery. J Chem Inf Model 51: 2778–2786.

Mahony, J., and van Sinderen, D. (2012) Structural aspects of the interaction of dairy phages with their host bacteria. Viruses 4: 1410–1424.

Murshudov, G.N., Skubak, P., Lebedev, A.A., Pannu, N.S., Steiner, R.A., Nicholls, R.A., et al. (2011) REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallogr D Biol Crystallogr 67: 355–367.

Parent, K.N., Erb, M.L., Cardone, G., Nguyen, K., Gilcrease, E.B., Porcek, N.B., et al. (2014) OmpA and OmpC are critical host factors for bacteriophage Sf6 entry in Shigella. Mol Microbiol 92: 47–60.

Plancon, L., Janmot, C., le Maire, M., Desmadril, M., Bonhivers, M., Letellier, L., and Boulanger, P. (2002) Char- acterization of a high-affinity complex between the bacte- rial outer membrane protein FhuA and the phage T5 protein pb5. J Mol Biol 318: 557–569.

Pymol. (2014) The PyMOL Molecular Graphics System, Version 1.5.0.4 Schrödinger, LLC.

Ricagno, S., Campanacci, V., Blangy, S., Spinelli, S., Tremblay, D., Moineau, S., et al. (2006) Crystal structure of the receptor-binding protein head domain from Lactococ- cus lactis phage bIL170. J Virol 80: 9331–9335.

Samson, J.E., and Moineau, S. (2013) Bacteriophages in food fermentations: new frontiers in a continuous arms race. Annu Rev Food Sci Technol 4: 347–368.

Sanders, M.E., and Klaenhammer, T.R. (1980) Restriction and modification in group N streptococci: effect of heat on

development of modified lytic bacteriophage. Appl Environ Microbiol 40: 500–506.

Sao-Jose, C., Lhuillier, S., Lurz, R., Melki, R., Lepault, J., Santos, M.A., and Tavares, P. (2006) The ectodomain of the viral receptor YueB forms a fiber that triggers ejection of bacteriophage SPP1 DNA. J Biol Chem 281: 11464– 11470.

Sato, K.-I., Akai, S., Sakuma, M., Kojima, M., and Suzuki, K.-J. (2003) Practical synthesis of [1-C-13]- and [6-C-13]- D-galactose. Tetrahedron Lett 44: 4903–4907.

Spinelli, S., Campanacci, V., Blangy, S., Moineau, S., Tegoni, M., and Cambillau, C. (2006a) Modular structure of the receptor binding proteins of Lactococcus lactis phages. The RBP structure of the temperate phage TP901-1. J Biol Chem 281: 14256–14262.

Spinelli, S., Desmyter, A., Verrips, C.T., de Haard, H.J., Moineau, S., and Cambillau, C. (2006b) Lactococcal bac- teriophage p2 receptor-binding protein structure suggests a common ancestor gene with bacterial and mammalian viruses. Nat Struct Mol Biol 13: 85–89.

Spinelli, S., Veesler, D., Bebeacua, C., and Cambillau, C. (2014) Structures and host-adhesion mechanisms of lac- tococcal siphophages. Front Microbiol 5: 1–13.

Tremblay, D.M., Tegoni, M., Spinelli, S., Campanacci, V., Blangy, S., Huyghe, C., et al. (2006) Receptor-binding protein of Lactococcus lactis phages: identification and characterization of the saccharide receptor-binding site. J Bacteriol 188: 2400–2410.

Vagin, A., and Teplyakov, A. (2010) Molecular replacement with MOLREP. Acta Crystallogr D Biol Crystallogr 66: 22–25.

Vincentelli, R., Bignon, C., Gruez, A., Canaan, S., Sulzenbacher, G., Tegoni, M., et al. (2003) Medium-scale structural genomics: strategies for protein expression and crystallization. Acc Chem Res 36: 165–172.

Vincentelli, R., Canaan, S., Offant, J., Cambillau, C., and Bignon, C. (2005) Automated expression and solubility screening of His-tagged proteins in 96-well format. Anal Biochem 346: 77–84.

Winn, M.D., Ballard, C.C., Cowtan, K.D., Dodson, E.J., Emsley, P., Evans, P.R., et al. (2011) Overview of the CCP4 suite and current developments. Acta Crystallogr D Biol Crystallogr 67: 235–242.

Xia, G., Kohler, T., and Peschel, A. (2010) The wall teichoic acid and lipoteichoic acid polymers of Staphylococcus aureus. Int J Med Microbiol 300: 148–154.

Xia, G., Corrigan, R.M., Winstel, V., Goerke, C., Grundling, A., and Peschel, A. (2011) Wall teichoic Acid-dependent adsorption of staphylococcal siphovirus and myovirus. J Bacteriol 193: 4006–4009.

Zissis, E., and Fletcher, H.G., Jr (1970) Benzyl 2,3,4-tri-o- benzyl-beta-d-glucopyranosiduronic acid and some related compounds. Carbohydr Res 12: 361–368.

Supporting information

Additional supporting information may be found in the online version of this article at the publisher’s web-site.

886 O. McCabe et al. ■