Structure and specificity of the Type VI secretion system ClpV-TssC interaction in enteroaggregative Escherichia coli

The Type VI secretion system (T6SS) is a versatile machine that delivers toxins into either eukaryotic or bacterial cells. It thus represents a key player in bacterial pathogenesis and inter-bacterial competition. Schematically, the T6SS can be viewed as a contractile tail structure anchored to the cell envelope. The contraction of the tail sheath propels the inner tube loaded with effectors towards the target cell. The components of the contracted tail sheath are then recycled by the ClpV AAA+ ATPase for a new cycle of tail elongation. The T6SS is widespread in Gram-negative bacteria and most of their genomes carry several copies of T6SS gene clusters, which might be activated in different conditions. Here, we show that the ClpV ATPases encoded within the two T6SS gene clusters of enteroaggregative Escherichia coli are not interchangeable and specifically participate to the activity of their cognate T6SS. Here we show that this specificity is dictated by interaction between the ClpV N-terminal domains and the N-terminal helices of their cognate TssC1 proteins. We also present the crystal structure of the ClpV1 N-terminal domain, alone or in complex with the TssC1 N-terminal peptide, highlighting the commonalities and diversities in the recruitment of ClpV to contracted sheaths.

Scientific RepoRts | 6:34405 | DOI: 10.1038/srep34405 protein is also part of the assembly platform composed of the TssEFGK proteins 30,31 . This platform -or baseplate -controls the polymerization of the tube/sheath structure and probably initiates sheath contraction 31,32 . This platform is tightly attached to the cell envelope through multiple contacts with components of the second sub-complex, called membrane complex 31,33 . The membrane complex is composed of three proteins: the inner membrane TssL and TssM proteins and the TssJ outer membrane lipoprotein [34][35][36][37][38] . These three proteins are present in 10 copies each and assemble a 1.7-MDa complex that crosses the cell envelope and delimits a channel for the passage of the inner tube during sheath contraction 39 . Once in contact with a target cell, or when the cell senses an attack by a competitor, the T6SS assembles the tubular structure and the sheath contracts, hence propelling the inner tube/spike towards the target cell and delivering toxins 20,40 . While the membrane complex is stable and can be reused for multiple injections 39 , the contracted sheath is disassembled by a dedicated AAA + ATPase of the Clp family, ClpV, and recycled for a new assembly and injection cycles 22,23,41 . In most T6SS, ClpV is recruited to the contracted sheath via interactions with the N-terminal helix of the TssC subunits that is thought to be accessible only in the contracted conformation 22,23,[41][42][43] . The N-terminal helix of TssC accommodates into a cleft constituted of charged residues located at the interface of two helices, called H1 and H2, at the N-terminus of ClpV 41,43 . However, in a subset of T6SS, the cleft is composed of uncharged residues, and the ClpV-TssC interaction is mediated or stabilized by the TagJ accessory protein 43 .
Interestingly, most bacteria encode several copies of T6SS gene clusters in their genomes [1][2][3]44 . These clusters are usually responsive to different environmental cues and hence subjected to different regulatory mechanisms 45 . It is thought that this diversity is responsible for the activation of distinct T6SS machineries in different conditions. However, little is known on how the subunits select their cognate machineries when multiple paralogous T6SS are present in the same cell. Enteroaggregative Escherichia coli strain 17-2 encodes two T6SS gene clusters of the T6SS-1 and T6SS-3 sub-families 44 , and it has been shown that the inner tube component Hcp encoded by the T6SS-1 cluster (sci-1) specifically interacts with the sheath component TssB1, but not with TssB2, which is encoded by the T6SS-3 cluster (sci-2), demonstrating specificity during assembly of the T6SS tails 21 . Here, we report that specificity also applies to disassembly of the contracted sheath. We demonstrate that over-production of ClpV2 cannot compensate for the absence of ClpV1 and cannot restore functionality of the T6SS-1 (and vice versa). We further provide evidence that ClpV1 specifically interacts with the N-terminal helix of TssC1 and not with that of TssC2. We then report the crystal structure of the ClpV1 N-terminal domain alone and in complex with the TssC1 N-terminal helix. Noteworthy, the crystal structure of the complex differs significantly from what was reported for ClpV-VipB complex from V. cholerae 41 . Our results suggest an alternative mode of binding between TssC1 and ClpV1, as the cleft in ClpV1 is essentially composed of uncharged residues but does not require TagJ to efficiently bind to the TssC1 N-terminal helix.

Experimental Procedures
Bacterial strains and media. The Escherichia coli K-12 DH5α , BTH101, W3110 and BL21(DE3) pLysS strains were used for cloning procedures, bacterial two-hybrid analyses, co-immunoprecipitations and protein production, respectively. Strain W3110 pUA66-rrnB (Kan R , GFP + ) 46 was used as prey in anti-bacterial competition experiments. Cells were grown in Lysogeny broth (LB), Sci-1-inducing medium (SIM) or Dulbelcco modified Eagle medium (DMEM), as specified. Plasmids were maintained by the addition of ampicillin (100 μ g/mL), chloramphenicol (40 μ g/mL) or kanamycin (50 μ g/mL). Table   S1. Polymerase Chain Reactions (PCR) were performed using a Biometra thermocycler using the Q5 high fidelity DNA polymerase (New England BioLabs). Custom oligonucleotides, listed in Supplemental Table S1, were synthesized by Sigma Aldrich. Enteroaggregative E. coli 17-2 chromosomal DNA was used as a template for all PRCs. The amplified DNA fragments correspond to the full-length ClpV1 (EC042_4530, GI: 284924251), ClpV2 (EC042_4577, GI: 284924293), TssC1 (EC042_4525, GI: 284924246) and TssC2 (EC042_4562, GI: 284924279) proteins, as well as the N-terminal domains of ClpV1 (residues 1-163) and ClpV2 (residues 1-147). Plasmids were engineered by restriction-free cloning 47 as previously described 35 . Briefly, genes of interest were amplified with oligonucleotides introducing extensions annealing to the target vector. The double-stranded product of the first PCR was then been used as oligonucleotides for a second PCR using the target vector as template. Deletion of TssC1 and TssC2 N-terminal helices as well as point mutations have been obtained by site-directed mutagenesis. All constructs have been verified by restriction analysis and DNA sequencing (Eurofins, MWG).

Plasmid construction for in vivo studies. Plasmids used in this study are listed in Supplemental
Bacterial two-hybrid assay. The adenylate cyclase-based bacterial two-hybrid technique 48 was used as previously published 49 . Briefly, compatible vectors producing proteins fused to the isolated T18 and T25 catalytic domains of the Bordetella adenylate cyclase were transformed into the reporter BTH101 strain and the plates were incubated at 30 °C for 24 hours. Three independent colonies for each transformation were inoculated into 600 μ L of LB medium supplemented with ampicillin, kanamycin and IPTG (0.5 mM). After overnight growth at 30 °C, 10 μ L of each culture were spotted onto LB plates supplemented with ampicillin (100 μ g/mL), kanamycin (50 μ g/mL), IPTG (0.5 mM) and Bromo-Chloro-Indolyl-Galactopyrannoside (X-Gal, 40 μ g/mL) and incubated for 16 hours at 30 °C. The experiments were done at least in triplicate and a representative result is shown.
Co-immunoprecipitations. 100 mL of W3110 cells producing the proteins of interest were grown to an absorbance at λ = 600 nm (A 600 ) ~ 0.4 and the expression of the cloned genes were induced with AHT (0.1 μ g/mL) and L-arabinose (0.2%) for 45 min. The cells were harvested, and the pellets were resuspended in 20 mM Tris-HCl pH 8.0, 100 mM NaCl, 30% sucrose, 1 mM EDTA, lysozyme 100 μ g/mL, DNase 100 μ g/mL, RNase 100 μ g/mL supplemented with protease inhibitors (Complete, Roche) to an A 600 of 80 and incubated on ice for 20 min. Cells were lysed by three passages at the French Press (800 psi) and lysates were clarified by centrifugation at 20,000 × g for 20 min. Supernatants were used for co-immunoprecipitation using anti-FLAG M2 affinity gel (Sigma-Aldrich). After 3 hours of incubation, the beads were washed three times with 1 mL of 20 mM Tris-HCl pH 8.0, 100 mM NaCl, 15% sucrose, resuspended in 25 μ L of Laemmli loading buffer, boiled for 10 min and subjected to SDS-PAGE and immunodetection analyses.
Anti-bacterial competition assay. Antibacterial competition growth assays were performed as previously described, in Sci-2-inducing 40 or Sci-1-inducing 29 conditions. The wild-type E. coli strain W3110 bearing the kanamycin-resistant GFP + pUA66-rrnB plasmid 46 was used as prey.
Plasmid construction for in vitro studies. The DNA sequence encoding the ClpV1 N-terminal domain (ClpV1-Nt, Gly1 to Leu163) was cloned into the pETG-20A expression vector using standard Gateway procedures to yield pETG20A-ClpV1-Nt. The resulting construction allows the production of ClpV1-Nt fused to an N-terminal Thioredoxin followed by a 6 × His tag and a Tobacco Etch Virus (TEV) cleavage site. Primers used for genes amplification are shown in Supplemental Table S1.
Peptide synthesis, protein production and characterization. E. coli BL21 (DE3) pLysS cells carrying the pETG20A-ClpV1-Nt plasmid were grown at 37 °C in TB medium (1.2% peptone, 2.4% yeast extract, 72 mM K 2 HPO 4 , 17 mM KH 2 PO 4 , and 0.4% glycerol). Expression of clpV1-Nt was induced at A 600 = 0.6 with 0.5 mM IPTG for 18 hours at 25 °C. Cells were then resuspended in lysis buffer (50 mM Tris-HCl pH 8.0, 300 mM NaCl, 1 mM EDTA, 0.5 mg/mL lysozyme, 1mM phenylmethylsulfonyl fluoride), submitted to several freeze-thawing cycles and sonicated after the addition of 20 μ g/mL DNase and 20 mM MgCl 2 . Pellet and soluble fraction were separated by centrifugation for 30 min at 16,000 × g. The soluble fraction containing ClpV1-Nt was loaded onto a 5-mL Ni 2+ affinity column (HisTrap ™ FF) using an ÄKTA Explorer apparatus (GE healthcare) and the immobilized proteins were eluted in 50 mM Tris-HCl pH8.0, 300 mM NaCl supplemented with 250 mM imidazole. The protein solution was desalted on a HiPrep 26/10 column (Sephadex TM G-25, Amersham Biosciences), and ClpV1-Nt was obtained by cleavage using 2 mg of TEV protease for 18 hours at 4 °C and collected in the flow-through of a 5-mL Nickel column. The protein was concentrated using the centricon technology (Millipore, 10-kDa cut-off). After concentration, the soluble ClpV1-Nt protein was passed through a Sephadex 200 26/60 column pre-equilibrated with 25 mM Tris-HCl pH7.5, 100 mM NaCl, 5% Glycerol. A similar procedure was applied for the ClpV2-Nt domain but the protein was insoluble and could not be purified.

Microscale thermophoresis (MST) experiments.
Microscale thermophoresis experiments were performed using a Monolith NT.115 apparatus (NanoTemper). ClpV1-Nt was labeled with the blue-fluorescent dye NT-495-NHS (NanoTemper) and the buffer was exchanged for the assay buffer (25 mM Tris-HCl pH 7.5, 200 mM NaCl, glycerol 5%, 0.05% Tween-20) using a Nap5 column (GE Healthcare). Titrations were conducted with a constant 200 nM fluorophore-labeled ClpV1-Nt against up to 500 μ M TssC N-terminal peptides in hydrophilic-coated capillaries. Each data point was measured in triplicate. Single-site fitting was performed using the NanoTemper data analysis software.
Crystal structures determination. ClpV1-Nt was purified to homogeneity and the protein crystallized spontaneously during gel filtration. ClpV1-Nt crystals belong to space group P2 1 2 1 2 1 with cell dimensions a = 40.9Å, b = 58.7 Å, c = 65.6 Å (Supplemental Table S2). A complete data set was collected at the Soleil synchrotron (Saint-Aubin, France) at 2 Å. Data were integrated with XDS 50 and reduced with XSXALE 50 (Supplemental Table S2). The structure was solved by soaking crystals in a 0.5 M NaI and 0.5 M of CsI solution followed by data collection at 1.77-Å wavelength, allowing remote SAD phasing. The cesium/iodine positions were defined with SHELXC and SHELXD 51 and subsequently used for SAD phasing using PHASER 52 . These phases were improved by solvent flattening and histogram matching using the program PARROT 53 , resulting in interpretable maps from which the BUCCANEER program 54 was able to build an initial model. The model was completed and corrected manually with COOT 55 and refined with autoBUSTER 56 (Supplemental Table S2).
The ClpV1-Nt/TssC1 peptide complex was obtained by mixing the purified ClpV1 N-terminal domain with the peptide at a 1:4 molar ratio in 20 mM Tris-HCl pH8.0, 100 mM NaCl, 5% Glycerol and concentrated using the centricon technology (Millipore, kDa cut-off of 10) to 2 mg/mL. Crystals were obtained in one condition (0.2 M imidazole, malate pH 6.0, 8% w/v PEG 4000) using the STURA screen (Molecular dimensions) and diffracted at 2.5-Å resolution. The structure was solved by molecular replacement using the structure of ClpV1-Nt as starting model. The model was completed and corrected manually with COOT 55 and refined with autoBUSTER 56 (Supplemental Table S2).
Data deposition. The EAEC ClpV1-Nt and ClpV1-Nt/TssC1 peptide X-ray structures have been deposited in the Protein Data Bank (PDB) under accession numbers 4HH5 and 4HH6, respectively.

Results
ClpV1 and ClpV2 are not interchangeable. The enteroaggregative E. coli Sci-1 and Sci-2 T6SS are both involved in bacterial competition 29,40 . However, these two T6SS are active in different laboratory conditions (minimal medium for the Sci-1 T6SS and modified Eagle medium for the Sci-2 T6SS 57,58 ). To test whether the two ClpV proteins, ClpV1 (EC042_4530, GI: 284924251) and ClpV2 (EC042_4577, GI: 284924293), respectively encoded by the sci-1 and sci-2 gene clusters are interchangeable, we engineered ∆ clpV1 and ∆ clpV2 strains and complementation vectors (pIBA-ClpV1 and pIBA-ClpV2). The different combinations were tested for their ability to confer a growth advantage in Sci-1-or Sci-2-active conditions against E. coli K-12. Figure 1 shows that while clpV1 and clpV2 trans-expression restores the growth advantage of the ∆ clpV1 and ∆ clpV2 mutant strains respectively, over-production of ClpV2 does not restore ∆ clpV1 defects, and vice versa. From this experiment, we concluded that the ClpV1 and ClpV2 ATPase are not interchangeable and we hypothesized that this specificity could be imputable to specific contacts with partners.

ClpV specificity is dictated by ClpV-TssC complex formation.
To identify ClpV1 partners, we performed a systematic bacterial two-hybrid analysis. ClpV1 was fused to the N-or C-terminus of the Bordetella pertussis Cya T25 domain and these protein fusions were tested as baits against preys comprising fusion of the T18 domain to the individual T6SS core components. Growth on reporter plates demonstrated that ClpV1 interacts with itself ( Fig. 2A), in agreement with the ability of AAA + ATPases to assemble hexameric structures 22,59 . However, ClpV1 does not interact with ClpV2 suggesting that these two proteins cannot assemble hetero-hexamers (Fig. 2B). In addition, ClpV1 interacts with TssC1 (EC042_4525, GI: 284924246), one of the sheath components (Fig. 2C). The ClpV1-TssC1 interaction was further confirmed by co-immunoprecipitation (Fig. 2D). The ClpV-TssC interaction has already been reported in V. cholerae and P. aeruginosa and involves contacts between the N-terminal domain of ClpV and an N-terminal helix of TssC 22,41,43 . Similarly, co-immunoprecipitation assays showed that the N-terminal domain of ClpV1 (ClpV1-Nt, residues 1-163) is sufficient to mediate the interaction with TssC1, while a deletion of the TssC1 N-terminal helix (deletion of residues 23-25, TssC1∆ h) abolishes ClpV1-TssC1 complex formation (Fig. 2D, left panel). Identical results were obtained for ClpV2 and the Sci-2-encoded TssC2 protein (Fig. 2D, right panel). We then tested cross-interaction between the ClpV N-terminal domains and TssC proteins. Fig. 2D (right lanes in left and right panels) shows that each ClpV N-terminal domain interacts only with the TssC encoded within the same gene cluster. The lack of cross-interaction between the ClpV and TssC proteins from the two T6SSs was further confirmed by bacterial two-hybrid (Fig. 2E). From these results, we concluded that specificity determinants control ClpV-TssC complex formation and we hypothesized that these specificity determinants should be located within the TssC N-terminal helices. In support of this hypothesis, a sequence alignment of TssC1 and TssC2 showed that the two proteins share high level of homologies except in the N-terminal region that is, in average, less conserved (18% identity within the 55 N-terminal residues compared to 38% identity on full-length sequences, Supplemental Figure S1).
To gain further insights onto the binding and specificity of ClpV1 towards the TssC N-terminal helices, the ClpV1 N-terminal domain was purified. In addition, peptides corresponding to TssC1 (residues 23-35) and TssC2 (residues 15-29) N-terminal helices (Fig. 3A) were synthesized. Binding of the TssC peptides to ClpV1-Nt was assayed by microscale thermophoresis that allows the quantification of molecular interactions in solution. Figure 3B shows that the TssC1 peptide binds to ClpV1-Nt with a K D of ~26 μ M, a value that is comparable to the affinity measured between V. cholerae ClpV and the VipB (TssC) N-terminal peptide (K D ~ 39 μ M) 41 . By contrast, no binding of TssC2 peptide to ClpV1-Nt was observed (Fig. 3C).

Structures of the ClpV1 N-terminal domain alone and in complex with the TssC1 N-terminal peptide reveal a new mode of binding.
To obtain molecular details on the ClpV1-TssC1 interaction, we crystallized the ClpV1-Nt domain alone and in complex with the TssC1 N-terminal helix peptide. ClpV1-Nt crystallized in the P2 1 2 1 2 1 space group and diffracted to 2.0-Å resolution (Supplemental Table S2). ClpV1-Nt structure was solved using a CsI derivative and phasing was achieved with remote SAD methods, using a 1.77-Å wavelength for data collection. The amino-acid chain could be traced between residues 1 and 163. The structure comprises 10 α -helices (H0 to H9) (Fig. 4), with overall dimensions of a flat disc of 50 × 46 × 28 Å. Interestingly, helix H0 is oriented perpendicular to helix H1, delimiting a groove.
Crystals of ClpV1-Nt in complex with the TssC1 peptide were obtained. The crystals belong to space group P2 1 2 1 2 1 and diffracted to 2.5-Å resolution (Supplemental Table S2). The structure of the complex was determined by molecular replacement using the uncomplexed structure (Fig. 5A,B). ClpV1-Nt is defined between residues 6 and 162 in the electron density map and its structure shows that it is not subjected to significant variation compared to the unbound ClpV1-Nt domain (root mean square deviation of 0.8 Å). The TssC1 N-terminal peptide is well defined in the electron density map between the residue Asp23 and Leu35. The peptide forms a ~2-turn α -helix between Ser24 and Phe30 (SVYASLF), while the rest of the peptide (EKINL) is elongated (Fig. 5A,B). PISA analysis shows that the interaction between ClpV1-Nt and the TssC1 peptide covers 622 Å 2 on the peptide (38% of the total surface) and 520 Å 2 on the ClpV1 N-terminal domain (6.6% of the total surface) (Fig. 5C, Supplemental Table S3). The TssC1 peptide accommodates in the ClpV1-Nt groove delimited by helices H0 and H1, and the H4-H5 loop (Figs 4A and 5A-E, Supplemental Table S3). Most residues of the peptide are involved in the interaction, with the exception of Ser24 and Ala27. The most significant interactions involve Tyr26, Leu29, Phe30, Lys32, Ile33 and Asn34, the three last residues belonging to the elongated stretch (Fig. 5D,E, Supplemental Table S3). The co-structure of ClpV1-Nt with the N-terminal peptide of TssC1 therefore reveals a mode of binding distinct from previous co-structures 43

(see Discussion).
Mutagenesis studies confirm the ClpV1-TssC1 interface shown in the co-crystal. To test the physiological relevance of the ClpV1-Nt/TssC1 peptide co-structure, we engineered point mutations within the ClpV1 groove and the TssC1 N-terminal region, and tested their effects on ClpV1-TssC1 interaction and T6SS function. The Glu24-to-Lys (ClpV1 EK mutant) and Arg87-to-Glu (ClpV1 RE mutant) substitutions (as well as the double mutant variant, ClpV1 EKRE ) were introduced in ClpV1. These mutations did not interfere with ClpV1 oligomerization (Fig. 6A). The Glu31-to-Lys and Lys32-to-Glu substitutions were introduced in TssC1 (TssC1 EKKE mutant). Here again, these mutations did not prevent the ability of TssC1 to oligomerize or to interact with TssB1 (Fig. 6A), two complexes previously described in EAEC 60 . However, all these substitutions affected partly or totally ClpV1-TssC1 complex formation (Fig. 6A). While the ClpV1 RE mutation had only a mild effect on the interaction, the ClpV1 EK and EKRE and the TssC1 EKKE mutations prevented formation of the ClpV1-TssC1 complex. The ClpV1 and TssC1 variants were then tested for their ability to support T6SS function. The two double mutants, ClpV1 EKRE and TssC1 EKKE , did not restore the anti-bacterial activity of ∆ clpV1 (Fig. 6B) and ∆ tssC1 (Fig. 6C) EAEC cells, respectively. By contrast the ClpV1 RE and EK single substitutions caused a 9-and 310-fold decrease of the function of the EAEC Sci-1 T6SS, respectively (Fig. 6B).

Discussion
Bacterial genomes may encode several copies of Type VI secretion gene clusters. Although these gene clusters are usually subjected to different regulatory controls and expressed in different conditions, distinct Type VI secretion machineries may co-exist in the same bacterial cell 1,2 . There is therefore a need to control proper assembly of these machines and to allow recruitment of cognate subunits. Indeed, previous studies have demonstrated that specificity operates during machine assembly, more specifically for the interaction between the inner tube Hcp subunit and its cognate TssB sheath protein in enteroaggregative E. coli 21 . Here, we show that a second level of specificity exists, at the stage of sheath disassembly/recycling. The EAEC 17-2 strain genome encodes two T6SS gene clusters 44,58 , each containing its own ClpV ATPase. These two proteins share 38/65% sequence identity/similarity (Supplemental Figure S1). Here, we show that these two proteins are not interchangeable and that ClpV2 cannot rescue the anti-bacterial defects of a ∆ clpV1 strain, and vice versa. Using a combination of bacterial two-hybrid, co-immunoprecipitation and in vitro microscale thermophoresis assays, we provide evidence that specificity is dictated by binding of the TssC N-terminal helix to the ClpV N-terminal domain. The N-terminal regions of TssC1 and TssC2 (residues 1-163) are much more divergent (20/53% identity/similarity) than full-length proteins (38/65%) (Supplemental Figure S1). The crystal structure of the ClpV1 N-terminal domain in complex with the TssC1 N-terminal helix further showed that the TssC1 helix inserts into a cleft delimitated by the H0 and H1 helices and the H4-H5 loop (Figs 4A, 5 and 6, Supplemental Table S3). Interestingly, binding of the TssC1 N-terminal helix to ClpV1-Nt does not cause a significant conformational change within ClpV. This situation is similar to that reported for the V. cholerae ClpV Vc -VipB complex (see below) 41 . The weak K D value (26 μ M) is also very similar to that measured in V. cholerae (39 μ M) 41 . However, the K D values obtained in vitro might not represent the in vivo situation. In vivo, it is assumed that the ClpV ATPase is recruited to the contracted sheath 15,18,22,61 , and therefore hundreds of TssC N-terminal helices are available, probably resulting in avidity higher by several orders of magnitudes as compared to the affinity of isolated fragments.
The structure of the T6SS contracted sheath 23,24,62 revealed that the TssC N-terminal helices are not visible in the electron density map, probably due to their high flexibility. However, the position of the first visible TssC residue (residue 61, pointed by the red arrow in Fig. 7A) strongly suggests that the N-terminal VipB/TssC helix protrudes far out of the sheath cylinder and hence might be accessible to ClpV. This observation supports the current model in which sheath contraction extricates the TssC N-terminal helix, leading to ClpV recruitment and sheath subunits recycling 18,22 . However, this model assumes that the TssC N-terminal helix is buried within the extended sheath. To date, the molecular structure of the T6SS sheath under the extended conformation is not available. In a recent study, a low resolution model of the extended V. cholerae VipAB sheath was modeled using the low resolution EM map of the extended T4 phage tail sheath 23 . By superimposing the VipAB EM map to the gp18 T4 phage sheath protein, gross features of the sheath structure were obtained 23 . To obtain high-resolution details, we applied this approach by fitting the recently released VipAB molecular model into the extended T4 phage tail sheath. In the extended model, VipB residue 61 is also located near the sheath surface, but packed against the cylinder surface (red arrows in Fig. 7B). We noticed an empty cavity nearby residue 61 that is large enough to accommodate the VipB N-terminal segment 1-60 (Fig. 7B). In this configuration, the TssC N-terminal helix will be buried and therefore not accessible to ClpV.
The structures of the V. cholerae (PDB 3ZRI) 41 and EAEC (this study) ClpV N-terminal domains share very close folds (Fig. 8). They display an elongated crevice at the same location, in which the TssC (VipB) helix inserts. While the overall scheme is the same, one can notice several differences. First, the position and orientation of the V. cholerae VipB and EAEC TssC1 N-terminal α -helices within the groove are slightly different. The TssC1 peptide starts with an alpha helical structure followed by an extended segment (Fig. 8). By contrast, the VipB peptide is entirely helical and locates within ClpV Vc at a position that is occupied by the TssC1 extended segment in ClpV1 (Fig. 8). This indicates, as already suggested 43 , that although ClpV exhibits a common fold in different T6SS, the nature and position of residues involved in TssC binding vary significantly enough to elicit variations in the mode of interaction. In addition, the ClpV1 groove that accommodates the TssC1 helix is constituted of a hydrophobic central area borded by charged residues, Glu24 of H1 in contact with the Lys29 residue of TssC1, and Arg87 of the H4-H5 loop in contact with a main-chain carboxylic group of TssC1 (Fig. 5D,E). It has been proposed that ClpV proteins could be categorized in two phylogenetic groups differing in the nature of residues within the groove (charged or uncharged residues) 43 . It has also been suggested that ClpV proteins with charged grooves cannot bind TssC N-terminal helices and require the assistance of an accessory protein, TagJ 43,63 . However, although the EAEC ClpV1 protein presents a central hydrophobic groove borded by charged residues, (i) it is able to bind directly to the TssC N-terminal helix, (ii) the ClpV1 charged residues are critical for the ClpV1-TssC1 interaction,    and (ii) no TagJ homologue is encoded within the EAEC genome. These observations suggest that interaction between ClpV1 and TssC1 in EAEC represents a third mode of binding dependent on a partially charged interface but independent on TagJ. This variability in the mechanism of ClpV binding to TssC supports the notion that specific determinants regulate recruitment of ClpV to contracted sheaths in cells with distinct T6SS machines.