Visualizing a Complete Siphoviridae Member by Single-Particle Electron Microscopy: the Structure of Lactococcal Phage TP901-1

Abstract

Tailed phages are genome delivery machines exhibiting unequaled efficiency acquired over more than 3 billion years of evolution. Siphophages from the P335 and 936 families infect the Gram-positive bacterium Lactococcus lactis using receptor-binding proteins anchored to the host adsorption apparatus (baseplate). Crystallographic and electron microscopy (EM) studies have shed light on the distinct adsorption strategies used by phages of these two families, suggesting that they might also rely on different infection mechanisms. Here, we report electron microscopy reconstructions of the whole phage TP901-1 (P335 species) and propose a composite EM model of this gigantic molecular machine. Our results suggest conservation of structural proteins among tailed phages and add to the growing body of evidence pointing to a common evolutionary origin for these virions. Finally, we propose that host adsorption apparatus architectures have evolved in correlation with the nature of the receptors used during infection.

INTRODUCTION

Bacteriophages of the order Caudovirales are exquisitely evolved nanomachines possessing a tail appendage used to recognize the host and ensure genome delivery with high specificity. They are the most abundant biological entity on earth, with an estimated 10³¹ tailed phages in the biosphere (1). Tail morphology serves as a basis to classify Caudovirales phages into three distinct families: Myoviridae, having a complex contractile tail (e.g., T4 [2]) Podoviridae, bearing a short noncontractile tail (e.g., P22 [3, 4]); and Siphoviridae, characterized by their long noncontractile tails (e.g., SPP1 [5]).

The first steps of phage infection require interactions between the phage receptor-binding proteins (RBPs) and the receptors at the host cell surface. Despite the diverse infection mechanisms displayed by Siphoviridae, using surface proteins and/or cell wall saccharides as receptors, their tail architecture is rather conserved. It is characterized by a long noncontractile tube, assembled by stacking several dozen homohexameric major tail protein (MTP) rings, and a central core formed by a few copies of the tape measure protein (TMP) extending between both tail extremities and determining its length. The proximal tail end harbors the homohexameric terminator that stops tube elongation during assembly, whereas the distal tail end is characterized by the presence of the tail adsorption apparatus. In phages of Gram-positive bacteria, this structure is composed of the distal tail protein (Dit), as well as the tail fiber, and is termed the baseplate or tip, depending on the presence or absence of peripheral proteins, respectively.

During the last few years, we have characterized the mechanisms underlying the initial steps of infection by bacteriophages targeting Gram-positive bacteria. Structural studies of the host adsorption apparatus of the Lactococcus lactis phages p2 and TP901-1 revealed distinct baseplate architectures and diverse strategies used by the two virions to initiate infection (6–11). The phage p2 baseplate undergoes large conformational changes in the presence of Ca²⁺ ions to appropriately orient its RBPs and establish multiple interactions with host saccharides at the onset of infection (8). In contrast, the TP901-1 baseplate harbors RBPs already pointing in the direction of the host, suggesting that the organelle is in a conformation ready for host adhesion (11). In vivo infection experiments confirmed and extended these observations by demonstrating that Ca²⁺ ions are required for host adhesion among p2-like phages (936 species) but have no influence on TP901-1-like phages (P335 species). Upon host recognition, a firing signal is generated and propagated along the tail up to the connector to inject the double-stranded DNA (dsDNA) genome into the host cell, which leads to the production of progeny virions (5, 12).

The highly flexible nature of Siphoviridae tails makes structural characterization of such phage particles difficult and explains the paucity of data reported for the organelle (5). We report here the electron microscopy (EM) reconstructions of the entire TP901-1 virion using single-particle protocols and a methodology specially implemented to characterize its tail. Mature TP901-1 virions have thin angular capsid shells filled with dsDNA and long tails when imaged by transmission electron microscopy. Based on our EM reconstructions and bioinformatics analyses, we propose pseudoatomic models for most parts of this Siphoviridae virion. The conservation of canonical phage structural protein modules supports the evolutionary connection proposed between all tailed phages and provides insights into the putative TP901-1 assembly and maturation pathway. We also put forward the idea that a striking correlation exists between host adsorption device architectures and the strategies employed to recognize and adsorb onto the host.

MATERIALS AND METHODS

Native phage production and purification.

TP901-1 phage were purified as previously described (13). Briefly, phage were induced with 3 μg/ml mitomycin C from lysogenic L. lactis 901-1 grown at 30°C in GM17 broth. Following cell lysis, the phage particles were precipitated and purified by isopycnic centrifugation using a CsCl gradient.

Specimen preparation. (i) Negative staining.

Approximately 3 μl of sample was applied onto glow-discharged carbon-coated grids and incubated for 1 min. The grids were blotted, and 10 μl of a 2% uranyl-acetate solution was added and incubated for 30 s. The excess stain was then blotted, and the grids were transferred to the microscope or stored.

(ii) Capsid cryo-EM.

The sample (3 μl) was applied and incubated on glow-discharged Quantifoil grids for 1 min and subsequently blotted for 3 s before being plunged into liquid ethane for vitrification using an FEI Vitrobot.

Data collection. (i) Negative-stain data.

Approximately 1,000 charge-coupled device (CCD) images were collected using a Phillips CM200 microscope with a field emission gun operated at 200 kV (Imperial College London) and a 4,000 by 4,000 TVIPS CCD camera. We used a magnification of ×38,000, resulting in a pixel size of 2.32 Å (4.64 Å coarsened by 2) over a range of nominal defocus values between 0.5 and 1.5 μm.

(ii) Capsid cryo-EM.

We collected 200 CCD images using the same setup as for stain data but with a magnification of ×50,000 (resulting in a pixel size of 1.765 Å) and nominal defocus values ranging from −1.5 μm to −3 μm.

Image processing.

All processing was carried out using IMAGIC software (14). Defocus estimation and correction for the microscope contrast transfer function (CTF) was carried out using the IMAGIC CTF2D_FIND and CTF2D_FLIP programs. Particles were selected using the program PICK_M_ALL and filtered, normalized, and masked before further processing. The number of particles used in each reconstruction is presented in Table 1.

Table 1.

Summary of the data-processing strategies employed for the various TP901-1 reconstructions

Phage component	Method	Symmetry	Resolution (Å)	No. of particles
Capsid	Cryo-EM	Icosahedral	15	1,500
Connector	Negative-staining EM	c12	21	1,000
Tail	Negative-staining EM	Helical	20	4,000
Baseplate	Negative-staining EM	c6	25	10,000

(i) Full phage.

In order to evaluate the overall dimensions of the tail and the number of MTP rings, 1,000 particles were manually selected from the images where the phage were observed in isolation and with a relatively straight tail. The full-phage particles were extracted into boxes of 1,200 by 1,200 pixels, coarsened by 2 to speed up processing, and submitted to single-particle analysis imposing 6-fold (c6) symmetry. Particles were then pretreated as described above, submitted to five rounds of alignment by classification (15), and subsequently multivariate statistical analysis (MSA) classified with 10 images per class (16). An initial model was generated by back-projecting a selected class average with c6 symmetry. The initial model was reprojected, and the reprojections were used for the initial angular assignment of the aligned particles by projection matching (16, 17). As previously described (8), particles were positioned in a side view orientation with the symmetry axis perpendicular to the projection direction. Therefore, maps were reprojected along the equator (IMAGIC Euler angle β equal to 90°) with a difference of 2°. Subsequent cycles of refinement over the entire data set, including alignment, projection matching, and model calculations, were iterated until stabilized.

(ii) Fragment processing.

The EM map generated for the full phage was cut into 7 continuous segments of 72 by 72 by 72 pixels corresponding to the connector (segment 1), the tail (segments 2 to 6), and the baseplate (segment 7). The aligned particles resulting from the full-phage refinement described above were cut in the same way and remasked to generate 7 subsets. The 1st (connector) and 7th (baseplate) subsets were further refined by projection matching using the corresponding segment of the map obtained by cropping the full-phage initial map. Refinement was carried out for several rounds over 2°, imposing c12 symmetry for the connector and c6 symmetry for the baseplate. For further analysis and interpretation, however, we used the baseplate that we previously obtained (6). Fragments 2 to 6 (corresponding to the tail) were combined (∼4,000 particles) and submitted to helical processing.

(iii) Tail helical processing.

The 4,000 particles combined as described above were submitted to helical processing (Table 1). The helical map was produced using the package IHRSR++ (18). The rotational symmetry used was c6, and as the particles were already aligned, the maximum allowed in-plane rotational angle was set to 10°. The initial helical parameters were determined using the Brandeis Helical Package (19) to calculate the Bessel orders of the basic layer lines (6 and −6) (Fig. 1A and B) and the Ruby-Helix package to estimate a repeat distance of 110 Å (20). These were later refined by IHRSR to a helical rise of 38 Å and a rotation between subunits of 22.4°.

Fig 1.

EM parameters of TP901-1 structure. (A and B) Determination of the helical parameters of the TP901 tail. An average of aligned tail tube segments (A) was used to generate the Fourier transform (B). The meridional line is marked by a dotted line. The layer lines are marked by arrows that also indicate their Bessel orders (6 and −6). This indexation showed the 6-fold rotational symmetry of the TP901 helical tail. (C) Fourier shell correlation (FSC) curves of the final three-dimensional (3D) reconstructions obtained by correlation of two different 3D reconstructions created by splitting the particle set into two subsets. The resolution was estimated by the 1/2-bit cutoff threshold criterion as 15 Å for the capsid, 21 Å for the connector, 21 Å for the helical tail, and 25 Å for the baseplate.

(iv) Reconstruction of the capsid.

A total of 1,500 particles were manually selected, extracted into boxes of 256 by 256 pixels, and submitted to MSA classification (Table 1). An initial model was created by back-projecting a single class average with icosahedral symmetry. The initial model was refined by projection matching over the entire data set with an angular sampling rate of 1° for several rounds, imposing icosahedral symmetry until stabilized.

(v) Resolution.

The resolution of the reconstructions was estimated by Fourier shell correlation and the 1/2-bit threshold correlation criterion (21) (Fig. 1C).

Fitting and analyses were carried out using the UCSF Chimera package (22) (Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, with support from the National Institutes of Health).

Sequence alignments were performed using the profile-profile alignment and fold recognition algorithm FFAS03 (Fold and Function Assignment System), as well as HHpred. Typically, predictions with FFAS03 scores lower than −9.5 contain <3% false positives.

Accession numbers.

The capsid, connector, and tail reconstructions have been deposited at the Electron Microscopy Data Bank (EMDB) with accession codes EMD-2133, EMD-2227, and EMD-2228, respectively.

RESULTS

The capsid.

Bacteriophage capsids are robust containers designed to carry and protect the viral genome that is packaged at liquid crystalline density within its interior (23). We computed a reconstruction of the TP901-1 capsid at 15-Å resolution using ∼1,500 particle images and applying icosahedral symmetry (Fig. 2A). The mature capsid is approximately 660 Å wide along its 5-fold axes and is made of 60 hexamers and 11 pentamers of the ORF36 major capsid protein (MCP), organized with a T = 7 symmetry, as well as a dodecamer of the portal protein occupying a unique vertex. Due to the 60-fold averaging applied during the reconstruction process, the portal density was averaged out, and its structure was independently investigated by reconstructing the connector region only. The capsid interior is filled with the dsDNA genome organized as concentric layers regularly spaced at ∼25 Å (Fig. 2B), as typically observed in other Caudovirales phages (23).

Fig 2.

The 15-Å-resolution cryo-EM reconstruction of the TP901-1 mature capsid. (A) Surface rendering of the icosahedral reconstruction low-pass filtered at 15 Å viewed along an icosahedral 2-fold axis. The capsid measures 660 Å along its 5-fold axis. (B) Cross section of the capsid reconstruction showing the layers of the dsDNA genome organized as concentric shells. (C) Pseudoatomic model of the TP901-1 mature capsid fitted in the reconstruction.

The plethora of MCP structures reported to date demonstrate the conservation of the so-called HK97 “Johnson fold” among tailed phages, herpesviruses, and some archaeal counterparts (23–25). Hence, we expect the TP901-1 MCP to exhibit a similar fold, and this is further supported by the detection of weak sequence similarity (but with high confidence) with the T7 and HK97 MCP sequences using the FFAS03 server and HHpred (Table 2) (26). A pseudoatomic model of the TP901-1 MCP shell was thus produced by rigid-body fitting of the icosahedral asymmetric unit of the mature HK97 capsid (Protein Data Bank [PDB] 1OHG) within the EM reconstruction (Fig. 2C). The seven subunits of the icosahedral asymmetric unit are well accommodated in the capsid density and form a 32-Å-thick shell surrounding the viral genome.

Table 2.

Sequence analyses of TP901-1 structural proteins

TP901-1 ORF	Protein	FFAS03 scorea	FFAS03 % identity	HHpred E value	Identified similar protein
ORF36	Major capsid protein	−62.0	17	3.7 × 10−27	T7 MCP-10A
		−38.6	16	5.1 × 10−28	HK97 gp5
ORF32	Portal	−89.5	19*	3.4 × 10−60	SPP1 gp6
ORF38	Head completion protein	−33.2	14	1.8 × 10−19	SPP1 gp15
ORF39	Head completion protein, stopper	−12.8	10	0.0034	SPP1 gp16
ORF42	Major tail protein	−22.3	11	8.5 × 10−10	Lambda gpV
ORF47	Tail-associated lysin	−53.3	31	4.6 × 10−26	S. aureus glycyl-glycine endopeptidase LytM

^aFFAS03 scores lower than −9.5 are considered significant.

The head-to-tail connecting region.

The connector ensures the cohesion of the phage capsid with its tail and is often made of three different components organized as successive rings: the portal protein and two head completion proteins. It is located at a unique capsid vertex, where it replaces a penton motif (Fig. 3A to C). We achieved a reconstruction of the connector using ∼1,000 particles and applying 12-fold symmetry along the connector channel axis.

Fig 3.

The 20-Å-resolution reconstruction of the TP901-1 connector. (A to C) The connector occupies a unique capsid vertex. (A) Side view. (B) View from the distal extremity along the tail tube. (C) Cross section of the capsid showing the portal protruding into it. (D) The SPP1 portal and the first head completion protein dodecamer (SPP1 gp15) are fitted into the connector reconstruction. Note the additional density surrounding the gp15 ring that likely corresponds to the collar/whiskers. The stopper (equivalent to SPP1 gp16) and the tail terminator are postulated to account for the remaining density. (E) View along the tail axis showing the fitting of the SPP1 portal dodecamer into the corresponding TP901-1 EM density (the region corresponding to the capsid density has been computationally removed for clarity). (F) View along the tail axis showing the fitting of the SPP1 first head completion protein into the corresponding TP901-1 EM density.

The portal is a dodecameric protein, disclosing a conserved fold in tailed phages and herpesviruses (23, 24), that is involved in DNA packaging during assembly and allowing release at the onset of infection. The TP901-1 portal (ORF32) has an overall length comparable to that of the equivalent protein in phage SPP1 (452 versus 503 residues), and the two sequences share 26% identity and 45% similarity. We used an SPP1 dodecameric portal model (27) to fit into the proximal region of the connector reconstruction, revealing good agreement between the atomic model and the EM map at this resolution (Fig. 3D and E) and confirming the structural similarity between TP901-1 and SPP1 portal proteins suggested by FFAS03 and HHpred (Table 2). Sequence analysis of the TP901-1 portal indicates that no P22-like coiled-coil structure is present in its C-terminal region, suggesting that the phage relies only on the turbine region of the dodecamer to trigger packaging termination when the genome reaches the headful density (4, 28, 29).

The remaining region of the connector reconstruction reported here was assumed to account for the two rings of head completion proteins and the tail terminator. Sequence analyses using the FFAS03 and HHpred servers allowed us to link TP901-1 ORF38 and ORF39 to the SPP1 head completion proteins gp15 and gp16, respectively (Table 2). These two proteins form dodecameric rings with known structures in vivo (12). We docked the SPP1 gp15 (PDB 2KBZ) and gp16 (PDB 2KCA) rings directly underneath the portal dodecamer in the connector EM map. The SPP1 gp15 head completion protein ring matches the dimensions of the corresponding TP901-1 connector moiety reasonably well (Fig. 3D and F), while the SPP1 gp16 dodecamer is only partially accounted for by the density (data not shown). Due to the symmetry mismatch between the connector and the tail, we have not attempted to fit any tail terminator model in the connector reconstruction, and this region is analyzed below.

The tail.

Bacteriophage tails ensure genome delivery to the target cell with an efficacy unequaled in the viral world. To investigate the tail structure, we first produced a low-resolution, 6-fold-averaged reconstruction of the whole TP901-1 phage from selected virions exhibiting a straight (unbent) tail. We used this reconstruction to assess that the tail tube is made of 34 stacks comprising a tail terminator hexamer, located at the interface with the connector, and 33 MTP hexamers forming the rest of the tube. We then boxed small tail segments (each including seven complete MTP stacks) from the tails and combined them in one data set subsequently processed with the appropriate helical symmetry.

The TP901-1 tail extends over 1,180 Å (Fig. 4A) between the connector and the baseplate, and its diameter varies between 110 Å (at the level of the MTP rings) and 90 Å (at the intersections between rings) (Fig. 4B). The MTP hexameric stacks are rotated by 22.4° between each other from the distal to the proximal tail extremities, and the interhexamer distance is 38 Å (Fig. 4B). The tail tube delineates a 42-Å-wide central channel that is continuous with the connector and baseplate channels to form the genome ejection pathway (5, 11, 12, 27, 30, 31) (Fig. 4C and D). We attributed the 28-Å-wide elongated density filling the tail interior to the TMP, the molecular ruler controlling the tail length during assembly (32) (Fig. 4C and D). The TMP is believed to be oligomeric and to form a long helical region extending through the tail tube and anchored by one globular domain at each extremity. Consistent with what has been observed in phage SPP1, no contacts were observed between the TMP and the surrounding MTP rings, either due to their nonexistence or because of a symmetry discrepancy between the two structures. Weak interactions between the TMP and the MTP hexamers probably facilitate ejection of the former before DNA ejection through the tail channel. Although of lower resolution, the overall dimensions of the TP901-1 tail components are in good agreement with their equivalents in the SPP1 tail, supporting the validity of our reconstruction.

Fig 4.

The 20-Å-resolution reconstruction of the TP901-1 tail. (A) Sixfold-averaged reconstruction of the TP901-1 phage tail from a few selected virions exhibiting an almost straight tail, making it possible to obtain the number of MTPs. (B) Detailed view of the reconstruction of a segment of the tail (7 MTP rings) using helical symmetry. The helical parameters of the tail are shown. (C) Cross section of the tail segment (blue) along its long axis revealing the internal TMP (violet). (D) Cross section of the tail segment orthogonal to its long axis (the color scheme is the same as in panel C).

The host adsorption device.

The baseplate is the control center for infectivity and is in charge of host recognition, attachment, and initiation of infection. Combining our recently reported TP901-1 baseplate crystal structure with the EM reconstruction (6) shows the detailed organization of this 280-Å-wide and 150-Å-high organelle exhibiting an overall 6-fold symmetry (Fig. 5A and B). From the proximal to the distal ends, it is formed by 18 copies of BppU (ORF48) arranged around a central Dit hexamer (ORF46) and holding 54 RBPs (ORF49) organized as 18 trimers (11) (Fig. 5A to C). The RBPs orient their 54 receptor-binding sites toward the distal extremity in a way suitable for establishing interactions with the pellicle layer of the host without requiring conformational changes (11, 33). The tail-associated lysine (Tal; ORF47) forms a 150-Å-long trimeric tail fiber appended to the Dit ring and extending beyond the baseplate core at its distal extremity. We modeled the tail fiber N-terminal domains using the closed p2 ORF16 trimer, which is expected to share a virtually identical fold and to undergo a similar conformational change to open the DNA ejection conduit during infection (8, 11, 24, 30, 31). While no structure is available to model the tail fiber central region, the last ∼150 residues of each monomer form a domain belonging to the peptidase M23 family that is probably involved in peptidoglycan digestion at the onset of infection to allow the virion to access the cytoplasmic membrane (Table 2) (13, 34).

Fig 5.

The 20-Å-resolution reconstruction of the TP901-1 distal tail region (the baseplate). The TP901-1 baseplate crystal structure was rigid-body fitted in the reconstruction. The color code is as follows: Dit (green), BppU (red), and BppL (light blue). The MTP hexamers were modeled using the Hcp type VI secretion system protein (dark blue). The N-terminal region of the tail fiber was modeled using the phage p2 ORF16 in closed conformation (pink). (A) Side view. (B) View along the tail axis from the distal extremity toward the capsid. (C) Cross section of the baseplate EM map. The assignment of the EM density to the different ORFs was performed using the X-ray structure.

DISCUSSION

Overall structure of the TP901-1 phage.

The structure determination of the TP901-1 phage capsid, tail, connector, and baseplate makes it possible to have an intermediate-resolution view of this large viral molecular machine (Fig. 6). The TP901-1 capsid is 660 Å wide along its 5-fold axes and is virtually identical to the phage HK97 capsid (25). Indeed, the TP901-1 MCP seems to harbor the so-called HK97 Johnson fold that is conserved among Caudovirales phages and in some viruses infecting eukaryotes and archaea (23–25). Analysis of the TP901-1 MCP sequence revealed that the protein does not harbor a scaffolding domain fused at its N terminus, in contrast to the HK97 situation. Instead, the virion genome exhibits an upstream open reading frame (ORF) encoding a protein product of ∼200 residues predicted to possess a high helical content and likely acting as a scaffolding protein. Based on these observations, we propose that the TP901-1 capsid assembly and maturation pathway are reminiscent of those of phage P22, which expels the intact scaffolding proteins upon initiation of dsDNA packaging rather than via proteolysis (35).

Fig 6.

Assembled complete structure of the TP901-1 phage. The complete phage was assembled by fitting the individually refined reconstructions into the map obtained for the full phage.

The TP901-1 connector structure is globally similar to that of SPP1: the SPP1 portal dodecamer, as well as the most proximal dodecamer of the head completion protein (SPP1gp15), is reasonably well accommodated in the corresponding regions of the TP901-1 reconstruction. The most distal ring of head completion protein should logically correspond to the SPP1 gp16 dodecamer (the “stopper”) according to sequence comparisons and to the observed density occluding the DNA exit channel in the reconstruction. However, the EM density appears to only partially account for it, probably due to the limited quality and resolution of the map in this region. Interestingly, an additional ring-like structure surrounds the connector at the level of the gp15 dodecamer, and we propose that this additional region of density might result from binding of additional proteins forming the collar/whiskers observed in micrographs of TP901-1, which have been averaged out during the reconstruction (13, 36). This additional ring might also be due to large conformational changes occurring upon tail attachment.

Considerable efforts have been made to understand how the dsDNA genome is driven from the phage capsid up to the host cytoplasm during infection. In the case of phage SPP1, it has been demonstrated that the high pressure with which the genome is packaged into the head is not enough to power its complete entry into the target cell (37). Other proteinaceous factors have been shown to participate in this phenomenon in bacteriophages T5 and T7 by pulling DNA into the host cell (38, 39). All the proteins building the central channel that allows DNA transit from the capsid to the host cell form an ∼40-Å-wide central channel with conserved negative electrostatic properties. No structure of the hexameric (biologically relevant) form of the MTP has been reported so far. However, considering the low pIs observed for most MTPs (e.g., ∼4.8 in TP901-1 or ∼4.7 in SPP1), it is likely that this protein contributes to the overall negative potential of the central ejection tunnel. The fact that the genome transit pathway is negatively charged has an obvious functional implication: as the dsDNA backbone is also negatively charged due to the presence of phosphate groups, a strong repulsion occurs, with the corresponding phage regions to which it is exposed favoring genome transit.

Besides ejection, this property might have an impact on phage assembly. A survey of the pIs of various TMP proteins revealed that these tail components are strongly basic (e.g., pI ∼8.8 in TP901-1 or pI ∼10 in SPP1). Therefore, association of the central TMP oligomer with the surrounding tail tube (formed of stacked MTP hexamers) is likely to rely, at least partially, on strong complementary electrostatic interactions.

The distal tail architecture is governed by the host adsorption strategy.

All bacteriophages belonging to the family Siphoviridae share a canonical tail organization but differ mainly in their distal tail part. Dairy phages belonging to the P335 and 936 species harbor complex baseplates in comparison to other phages of the same family, and this seems to correlate with the different host adsorption strategies. Indeed, they interact only with host cell wall saccharides, putatively the phosphosaccharides present in the pellicle of Gram-positive bacteria, for specific recognition and attachment (8, 10, 33, 40–42). Furthermore, many other siphophages, such as SPP1 and the lactococcal phages belonging to the c2 species, bind reversibly to saccharidic receptors in a first step before interacting irreversibly with a membrane protein that initiates infection (5, 43–49). As the affinity between phage antireceptors and their saccharidic partners is generally moderate (in the low micromolar range), several RBPs are involved in binding to ensure strong interactions based on avidity (50, 51). In contrast, the interaction between antireceptors and their proteinaceous receptors is strong, as illustrated by the tight binding reported between the T5 pb5 protein and the outer membrane transporter FhuA (52) or λ gpJ and LamB (53). Siphoviridae members can therefore be dichotomized in two categories based on the observation of their distal tail architecture. On one hand, some phages harbor a large baseplate accommodating up to several dozen RBPs to interact only with the saccharidic part of the host cell wall. On the other hand, bacteriophages devoid of any baseplate and possessing a simplified tail tip rely on irreversible binding with a transmembrane protein present in the target cell to ensure their commitment.