WT EPEC O127:H6 strain E2348/69 (streptomycin-resistant strain) and EPEC-null mutants (ΔescN and ΔescV; Gauthier et al., 2003) were used to evaluate the T3SS and translocation activities (Table 1). E. coli BL21 strain (λDE3) and E. coli DH10B were used for protein expression and for plasmid handling, respectively (Table 1). E. coli FHK12, which encodes β-galactosidase under the control of the ctx promoter, and E. coli PD28 (a malE-deficient E. coli strain) were used for assessment of TMD oligomerization (Table 1). The strains were grown at 37°C in a Luria–Bertani (LB) broth or Dulbecco’s Modified Eagle Medium (DMEM) with or without IPTG in the presence of the appropriate antibiotics. Antibiotics were used at the following concentrations: carbenicillin (100 μg/ml), streptomycin (50 μg/ml), and chloramphenicol (30 μg/ml). Bacterial growth was recorded at 600 nm every 30 min on the Infinite 200 PRO multimode plate reader (Tecan Group Ltd., Switzerland).

FASTA sequence for the EscV protein was downloaded from GenBank on the NCBI and UniProt database. To predict the TMD sequences, we used the TMHMM software¹. ClustalW multisequence alignment algorithm was used to identify highly conserved protein domains by comparing sequences of EscV orthologs² (no changes were made in the default parameters of the server; Larkin et al., 2007).

EscV-expressing plasmids were constructed using the primers presented in Table 2. The pSA10 plasmid was amplified using the primer pair pSA10_F/pSA10_R (Table 2). The escV gene was amplified from EPEC genomic DNA using the primer pairs EscV_SA10_F/EscV_His_R1 and then EscV_SA10_F/EscV_His_R2, which fused a His tag to the coding region of EscV (Table 2). The polymerase chain reaction (PCR) products were subjected to digestion with DpnI, purified, and assembled by the Gibson assembly method (Gibson et al., 2008, 2009). An HA-labeled version of EscV was similarly cloned into pSA10; here, the double-HA tag was fused to the coding region of escV by amplification from EPEC genomic DNA using the primer pairs EscV_SA10_F/EscV_HA_R1 and then EscV_SA10_F/EscV_HA_R2 (Table 2). To clone EscV labeled with the V5 tag, we amplified the pEscD-V5 (pSA10) vector (Tseytin et al., 2018a) using the primer pair pSA10_V5_F/pSA10_R and the EPEC genomic DNA with EscV_SA10_F/EscV_V5_R (Table 2). The PCR products were subjected to digestion with DpnI, purified, and assembled by the Gibson assembly. To clone EscV-V5 into a low-copy-number plasmid (pACYC184), we amplified the pEscP-V5 (pACYC184) vector (Shaulov et al., 2017) using the primer pair pSA10_V5_F/pACYC_R and the EPEC genomic DNA with EscV-V5_184_F/EscV-V5_184_R (Table 2). The PCR products were subjected to digestion with DpnI, purified, and cloned using the Gibson assembly. The resulting constructs, pEscV_wt-V5 (in pSA10 and pACYC184), pEscV_wt-HA, and pEscV_wt-His expressed a full-length EscV protein with V5, HA, or His tag fused to its C-terminal, respectively.

The TMD5-exchanged and TMD6-exchanged escV in pSA10 were generated by using the template of pEscV_wt-His (pSA10). To replace the TMD5 of EscV by a TMD backbone sequence of 7-leucine-9-alanine (7L9A), the EscV 222–675 amino acid sequence was amplified by using the primer pair EscV_Fc5_F/EscV_His_R2 (Table 2) from the pEscV_wt-His vector. The TMD 7L9A backbone was generated by annealing the primer pair 7L9A-F/7L9A-R (Table 2) by heating the sample to 95°C for 5 min and then decreasing the temperature to 20°C at a rate of 5°C/min. Then the 7L9A backbone was amplified by the primer pair EscV_TMD5-7L9A_F/EscV_TMD5-7L9A_R (Table 2). The EscV_222–675 PCR fragment and the 7L9A backbone were then ligated by using overlapping sequences and amplified by using the primer pair EscV_TMD5-7L9A_F/EscV_His_R2 (Table 2). The Gibson assembly was conducted by amplifying the pEscV_wt-His pSA10 vector with the primer pair pSA10_F/EscV_7L9A_Gib5_R (Table 2), followed by treating the reaction with DpnI and subjecting the amplified vector and the 7L9A-EscV_222–675-fused PCR fragment to ligation. The resulting construct, pEscV-TMD5_ex-His (pSA10), expressed a TMD5-exchanged EscV with a His tag at its C-terminus. The TMD6-exchanged escV in pSA10 was generated by a similar manner. The TMD 7L9A backbone was amplified by the primer pair EscV_TMD6-7L9A_F/EscV_TMD6-7L9A_R and fused to the EscV_259–675 PCR fragment, which was amplified by using the primer pair EscV_Fc6_F/EscV_His_R2 (Table 2). The Gibson assembly was conducted by amplifying the pEscV_wt-His pSA10 vector with the primer pair pSA10_F/EscV_7L9A_Gib6_R (Table 2), followed by treating the reaction with DpnI and subjecting the amplified vector and the 7L9A-EscV_259–675-fused PCR fragment to ligation. The resulting construct, pEscV-TMD6_ex-His (pSA10), expressed a TMD6-exchanged EscV with a His tag at its C-terminus. To move the TMD5-exchanged and TMD6-exchanged EscV into pSA10 that fuses the proteins with the V5 tag, pEscV-TMD5_ex-His and pEscV-TMD6_ex-His were amplified using the primer pair EscV_SA10_F/EscV_V5_R, and the pEscV_wt-V5 vector was amplified by the primer pair pSA10_V5_F/pSA10_R (Table 2). The PCR products were treated with DpnI and subjected to Gibson assembly.

Site-directed mutagenesis of G213A, G217A, G213L, and G217L within the EscV-V5 (pSA10) construct was performed using the primer pairs G213A_F/G213A_R, G217A_F/G217A_R, G213L_F/G213L_R, and G217L_F/G217A_L (Table 2). All constructs were verified by DNA sequencing.

The ToxR transcription activator can be successfully used to detect TMD–TMD interactions (Brosig and Langosch, 1998). DNA cassettes, encoding single TMDs of EscV (TMD1–TMD7.2), glycophorin A (GpA) TMD, E. coli aspartate receptor N-terminal TMD (Tar-1), 16-alanine backbone (A16), and no TMD (ΔTMD), were grafted between the cytoplasmic domain of the ToxR transcription activator protein (an oligomerization-dependent transcriptional activator) and the periplasmic domain of the maltose-binding protein (MBP). This was performed by aligning the oligonucleotides pairs encoding a NheI–BamHI TMD-DNA cassette of 16 core residues of the EscV TMD1–TMD7 (Table 2). The double-stranded fragments were oligophosphorylated and ligated between the toxR transcription activator and the malE (encodes E. coli MBP) within a NheI–BamHI digested ToxR-MBP plasmid (Table 2).

The MBP domain directs the chimera protein to the periplasm (Salema and Fernandez, 2013) where the TMD becomes embedded within the inner membrane. In the assay, we transform ToxR-TMD-MBP plasmids containing different TMDs into E. coli FHK12 cells, which contain a reporter gene, coding for β-galactosidase, under the control of the ctx promoter (Kolmar et al., 1995). Oligomerization of the investigated TMD derives ToxR oligomerization, which then can bind (in its oligomeric form) the ctx promoter and transcribe the reporter gene lacZ (Ottemann et al., 1992; Brosig and Langosch, 1998). Determination of the oligomerization level is calculated from the activity of β-galactosidase that is measured by the levels of a yellow color (OD₄₀₅) produced from the cleavage of the o-nitrophenylgalactose (ONPG), β-galactosidase substrate (Langosch et al., 1996; Fink et al., 2012). We monitored the activity of β-galactosidase for 20 min, at intervals of 30 s and calculated the V_max of the reaction. These were normalized to the original cell content (measured at OD₆₀₀) and are presented in Miller units. We used the GpA TMD sequence as a positive control for strong homo-oligomerization (Russ and Engelman, 2000), the N-terminal TMD of the E. coli aspartate receptor (Tar-1) as a reference for moderate oligomerization (Sal-Man et al., 2004), and the A16 sequence as a control for non-oligomerizing sequences (Bronnimann et al., 2013).

Membrane localization and correct orientation of the chimera proteins were examined as described previously (Langosch et al., 1996). Briefly, we transformed PD28 cells, E. coli strain deficient for malE (Duplay and Szmelcman, 1987), with the different ToxR-TMD-MBP constructs described above. The bacteria were grown overnight, washed, and then inoculated into M9 minimal medium supplemented with 0.4% maltose. Bacterial growth (OD₆₀₀) was measured over time. Since PD28 cells are unable to utilize maltose, only strains that translocate the chimera protein across the inner membrane were expected to grow in this medium. A construct without TMD (ΔTMD) served as a negative control. This chimera protein is expected to reside in the bacterial cytoplasm and, therefore, cannot grow in minimal medium.

Type 3 secretion assays were performed as previously described (Tseytin et al.,2018a,b, 2019). Briefly, overnight EPEC strains grown in LB were diluted 1:40 into preheated DMEM and grown for 6 h at 37°C in an atmosphere of 5% CO₂, statically, to an optical density of 0.7 (OD₆₀₀). Protein expression was induced by 0.1 mM IPTG. To separate bacterial cells from bacterial supernatants, we centrifuged the cultures at 20,000 × g for 5 min; the bacterial pellets were dissolved in an sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) sample buffer, and the supernatants, containing secreted proteins, were collected and filtered through a 0.22-μm filter (Millipore). The supernatants were normalized according to the bacterial OD₆₀₀ and precipitated with 10% (v/v) trichloroacetic acid (TCA) overnight at 4°C. To concentrate the secreted proteins, the supernatant samples were centrifuged at 20,000 × g for 30 min at 4°C and resuspended in the SDS-PAGE sample buffer. To neutralize the residual TCA, we added saturated Tris.

Samples were subjected to 12% SDS-PAGE and transferred to nitrocellulose (pore size, 0.45 μm; Bio-Rad) or polyvinylidene difluoride (PVDF; Mercury; Millipore) membranes. The blots were blocked with 5% (w/vol) skim milk–PBST (0.1% Tween in phosphate-buffered saline) for 1 h; then incubated with the primary antibody for 1 h at room temperature, or overnight at 4°C; washed with PBST three times; and then incubated with the secondary antibody (diluted in 5% skim milk–PBST, for 1 h, at room temperature). Chemiluminescence was detected with the EZ-ECL reagents (Biological Industries). A dilution of 1:1,000 was used for mouse anti-HA (Abcam), anti-V5 (Invitrogen), anti-His (Thermo Fisher Scientific), anti-HSV (Novagen), anti-actin (MPBio), anti-JNK (BD Pharmingen), anti-DnaK (Abcam), and rabbit anti-MBP (Thermo Fisher Scientific). Mouse anti-intimin (a gift from B. Brett Finlay, UBC) was diluted 1:2,000 (Gauthier et al., 2003), and horseradish peroxidase (HRP)-conjugated goat anti-mouse antibody (Abcam) and HRP-conjugated goat anti-rabbit antibody (Abcam) were diluted 1:10,000.

Translocation assays were performed as previously described (Baruch et al., 2011). Briefly, EPEC strains were pre-induced for 3 h for T3SS activity (preheated DMEM, statically, in a CO₂ tissue culture incubator) and then were used to infect HeLa cells (8 × 10⁵cells per well) at a multiplicity of infection (MOI) of 1:300 for 3 h. Cells were then washed and lysed with RIPA buffer [10 mM Tris-Cl pH 8.0, 1 mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 140 mM NaCl; before use, add 1 mM phenylmethylsulfonyl fluoride (PMSF)]. Lysates were centrifuged at maximum speed for 5 min to remove non-lysed cells and supernatants containing cellular proteins, were collected, mixed with the SDS-PAGE sample buffer, and analyzed by western blot analysis with anti-JNK and anti-actin antibodies (loading control). Uninfected HeLa cell samples, and HeLa cells infected by EPEC ΔescN and EPEC ΔescV mutant strains were used as negative controls.

Bacterial cell fractionation was performed as previously described (Gauthier et al., 2003). Briefly, overnight EPEC cultures were subcultured at 1:50 in DMEM and grown for 6 h, at 37°C, in a CO₂ tissue culture incubator. Bacteria were collected, washed, and resuspended in buffer A [50 mM Tris (pH 7.5), 20% (w/v) sucrose, 5 mM EDTA, protease inhibitor cocktail (Roche Applied Science), and lysozyme (100 μg/mL)] to generate spheroplasts. The samples were incubated for 15 min, at room temperature, and while rotating, MgCl₂ (20 mM) was then added. The samples were spun for 10 min at 5,000 × g, and the supernatants, containing the periplasmic fractions, were collected. The pellets were resuspended in lysis buffer (20 mM Tris/HCl pH 7.5, 150 mM NaCl, 3 mM MgCl₂, 1 mM CaCl₂, and 2 mM β-mercaptoethanol with protease inhibitors) and kept at 4°C from this step on. The samples were sonicated (Thermo Fisher Scientific, 3 × 10 s) after addition of RNase A and DNase I (10 μg/ml) and centrifuged (2,300 × g for 15 min) to remove intact bacteria, and the supernatants, containing cytoplasmic and membrane proteins, were collected. To separate the cytoplasmic and membrane fractions, the samples were centrifuged (in a Beckman Optima XE-90 Ultracentrifuge with a SW60 Ti rotor) for 30 min at 100,000 × g. The supernatants, containing the cytoplasmic fraction, were collected, and the pellets, containing the membrane fractions, were washed with lysis buffer, and the final pellets were resuspended in lysis buffer with 0.1% SDS. We evaluated the protein content by Coomassie Plus protein assay. We used various proteins as membrane (Intimin), periplasm (MBP), and cytoplasm (DnaK) markers.

We isolated the bacterial membrane fraction using the bacterial fractionation protocol described above. To extract membrane proteins from the crude membranes, samples were resuspended in lysis buffer containing 1% n-dodecyl-β-D-maltoside (DDM) and incubated for 60 min on a rotary wheel, at 4°C. The samples were then centrifuged (20,000 × g, for 15 min, at 4°C) to remove non-solubilized material, and the supernatants, containing membrane proteins, were collected and analyzed by blue-native (BN)-PAGE.

We incubated the extracted membrane proteins for 15 min in a BN sample buffer (30% glycerol with 0.05% Coomassie Brilliant Blue G250), and then loaded the samples onto a Criterion XT Tris-Acetate 3–8% precast gradient native gel (Bio-Rad). Electrophoresis was carried out on ice using a cathode buffer (15 mM Bis-Tris and 50 mM bicine, pH 7) and anode buffer (50 mM Bis-Tris, pH 7) until full separation (5–6 h). The gel was then incubated in a transfer buffer and subjected to western immunoblotting.

As TMDs are known to be involved in protein–protein interactions, we examined the ability of isolated EscV TMDs to support self-interaction. For that purpose, we employed the ToxR assembly system (Figure 2A), which monitors the strength of TMD–TMD interactions within the bacterial inner membrane (Langosch et al., 1996; Joce et al., 2011). We compared the oligomerization level of EscV TMDs with that of GpA’s TMD sequence, which contains a GxxxG motif and is used as a reference for strong homo-oligomerization (Lemmon et al., 1992; Adair and Engelman, 1994; Russ and Engelman, 2000). We also compared the EscV TMD oligomerization levels with the N-terminal TMD of the E. coli aspartate receptor (Tar-1), which has moderate oligomerization (Sal-Man et al., 2004), and polyalanine’s (A16) sequence as a non-oligomerizing sequence (Langosch et al., 1996; Sal-Man et al., 2005). Since the amino acid sequence of TMD7 was significantly longer than that of the other TMDs and the ToxR system responded differently to various TMD lengths (Langosch et al., 1996), we decided to test two different forms of this TMD, TMD7.1, and TMD7.2, which are of a similar length as the other examined TMDs. The sequences of the TMDs studied are presented in Figure 2A. We observed strong TMD self-oligomerization activity for EscV’s TMD5, TMD6, and TMD7.2 compared to the activities of the GpA and Tar-1 TMDs, whereas EscV’s TMD1, TMD2, TMD3, TMD4, and TMD7.1 showed reduced oligomerization activities compared to GpA (Figure 2B). As expected, the oligomerization of the A16 background control was low (Figure 2B). These findings suggested that TMD5, TMD6, and TMD7.2 of EscV might be involved in the oligomerization of the full-length protein EscV, through TMD–TMD interactions. To exclude the possibility that the high self-oligomerization activity of EscV’s TMD5, TMD6, and TMD7.2 resulted from higher expression levels of these chimera proteins, we subjected the bacterial samples to SDS-PAGE and western immunoblotting analysis with an anti-MBP antibody. All samples showed comparable expression levels (Figure 2B). To verify that the ToxR-TMD-MBP chimera proteins correctly integrated into the inner membrane, we employed the maltose complementation assay. For that purpose, we used an E. coli strain with a malE gene (PD28) deletion, which cannot produce endogenous MBP and therefore cannot support bacterial growth in minimal medium with maltose as the sole carbon source (Langosch et al., 1996). Only strains that express the chimera protein ToxR-TMD-MBP and orient it across the inner membrane, with MBP facing the periplasm, will support bacterial growth. We observed that all examined strains demonstrated bacterial growth, which indicated proper membrane integration, while the negative control that did not contain a TMD (ΔTMD) showed no growth, as expected (Figure 2C). Overall, these results suggest that TMD5, TMD6, and TMD7.2 of EscV are involved in EscV self-oligomerization through TMD–TMD interactions. However, due to the high conservation of TMD6 and the GxxxG motif within TMD5, on one hand, and the unclear boundaries of TMD7, on the other, we decided to focus on EscV’s TMD5 and TMD6.

To examine whether EscV’s TMD5 and TMD6 serve solely as membrane anchors or have a functional role within the full-length protein, we constructed EscV mutant proteins lacking TMD5 or TMD6 sequences. Since EscV lacking its TMD5 or TMD6 will likely adopt alternate protein folding compared to the native protein or have impaired localization, we constructed TMD5- and TMD6-exchanged EscV proteins, where the native core TMD5 and TMD6 sequences (16 amino acids in length) were replaced by a hydrophobic sequence. We chose a hydrophobic sequence of seven consecutive leucine residues followed by nine alanine residues (7L9A), which was previously shown to be sufficiently hydrophobic to support protein integration into the membrane yet cannot support TMD–TMD interactions (Sal-Man et al., 2005). To determine the biological effect of this replacement, we transformed the TMD5- and TMD6-exchanged EscV (EscV-TMD5_ex-His and EscV-TMD6_ex-His), as well as EscV_wt-His, into the escV-null strain (ΔescV) and examined their ability to restore T3SS activity. However, when EscV overexpression was induced by addition of IPTG to a concentration of 0.25 mM, growth rate was reduced in all strains (Figure 3). To determine the conditions that allow EscV expression without severe reduction of bacterial fitness, we grew WT EPEC, EPEC ΔescV, and EPEC ΔescV carrying either pEscV_wt-His, pEscV-TMD5_ex-His, or pEscV-TMD6_ex-His in LB or in DMEM (which is used for T3SS-inducing conditions), in the presence (0.1 or 0.25 mM) or the absence of IPTG. Optical density at 600 nm was measured over time (Figure 3). We observed that expressions of EscV WT and TMD-exchanged versions have reduced fitness when induced with IPTG at a concentration higher than 0.1 mM (Figure 3). These results suggest that overexpression of EscV is toxic to bacteria and therefore negatively affects bacterial growth. Based on these results, we used 0.1 mM IPTG for our experiments.

To examine whether EscV TMD5 and TMD6 sequences are critical for the activity of the full-length protein, we examined whether EscV-TMD5_ex-His and EscV-TMD6_ex-His can restore the T3SS activity of the EPEC ΔescV strain. Only functional EscV can restore the T3SS of the ΔescV strain, which is measured by the ability of EPEC strains to secrete three T3SS translocators (EspA, EspB, and EspD) into the culture supernatant, when grown under T3SS-inducing conditions.

First, we evaluated the ability of WT EscV to restore the T3SS activity of ΔescV. We observed that the expression of EscV_wt-His within the ΔescV strain restored secretion of translocators but also resulted in hypersecretion of effectors (Tir and NleA; Figure 4A). To evaluate whether this phenotype occurs due to the labeling of EscV or to its expression from a plasmid, we examined the T3SS activity of ΔescV-carrying plasmids with unlabeled EscV or EscV labeled with various tags and expressed from low- and high-copy-number plasmids. We observed that transformation of unlabeled EscV_wt resulted in a milder phenotype and that only a slight elevation in effector secretion was observed. In contrast, expression of labeled EscV, regardless of the tag type, resulted in hypersecretion of effectors (Supplementary Figure 2). Interestingly, expressions of both EscV-TMD5_ex-His and EscV-TMD6_ex-His failed to complement the T3SS activity of the ΔescV strain and demonstrated a secretion profile similar to that of ΔescV and ΔescN (Figure 4A). Comparable protein expression of the WT and the exchanged versions was observed by analyzing whole-cell lysates by western blot analysis using anti-His antibody (Figure 4A).

To analyze whether the unregulated secretion of ΔescV transformed with pEscV_wt-His affected the ability of the bacteria to infect host cells, we examined the ability of the strain to infect and translocate effectors into the HeLa cells. For this purpose, we infected HeLa cells with WT, ΔescN, ΔescV, and ΔescV transformed with pEscV_wt-His and examined the cleavage pattern of JNK, a cellular protein that is cleaved by NleD, a translocated EPEC effector (Baruch et al., 2011). WT EPEC induced extensive degradation of JNK, as expected, relative to the uninfected sample and to the samples infected with ΔescN or ΔescV mutant strains (Figure 4B). EPEC ΔescV transformed with the plasmid encoding EscV_wt-His showed a JNK degradation profile, indicating functional complementation by His-labeled EscV (Figure 4B). In addition, the ΔescV strain transformed with EscV TMD-exchanged versions (pEscV-TMD5_ex-His or pEscV-TMD6_ex-His) showed no degradation of JNK, as observed for the uninfected sample (Figure 4B). Overall, our results suggest that His-labeled EscV functionally complements the T3SS activity; however, replacing the native TMD5 or TMD6 sequences of EscV with an alternative hydrophobic sequence (7L9A) impairs the function of the T3SS (Figure 4B).

To exclude the possibility that EscV-TMD5_ex-His and EscV-TMD6_ex-His failed to complement the ΔescV T3SS activity due to impaired subcellular localization, we grew the strains under T3SS-inducing conditions and fractionated them into periplasmic, cytoplasmic, and membrane fractions. Our results showed that EscV-TMD5_ex-His and EscV-TMD6_ex-His localized mostly to the membrane fraction, as was seen for EscV_wt-His (Figure 5). Correct bacterial fractionation was confirmed by analyzing the samples with anti-MBP (periplasmic marker), anti-DnaK (cytoplasmic marker), and anti-intimin (membrane marker) antibodies. Overall, our results indicated that replacement of TMD5 and TMD6 did not disrupt EscV localization to the bacterial membrane.

To investigate whether the EscV TMD-exchanged variants fail to complement the T3SS activity of the ΔescV-null strain due to their inability to properly integrate into the T3SS complex, we prepared crude membrane samples of EPEC ΔescV and EPEC ΔescV-null strains transformed with EscV_wt-His, EscV-TMD5_ex-His, and EscV-TMD6_ex-His grown under T3SS-inducing conditions. The samples were then analyzed by BN-PAGE and immunoblotting. BN-PAGE analysis revealed that EscV_wt-His and EscV-TMD5_ex-His preserved the ability to integrate into the T3SS complex, as they migrated primarily as a large complex (>1 MDa) to the top of the gel. However, EscV-TMD6_ex-His integration into the complex appeared to be impaired (Figure 6). To verify that the modified running pattern of the EscV TMD6-exchanged version was not due to reduced protein expression, we analyzed the crude membrane extracts by SDS-PAGE and western blotting using anti-His antibody. Similar expression levels were observed for all EscV variants (Figure 6). These results suggest that TMD5 and TMD6 are not critical for the integration of EscV into the T3SS complex, as EscV-exchanged versions enabled the formation of high-molecular-weight complexes. EscV-TMD5_ex-His fully preserved the ability to integrate into the full or intermediate T3SS complexes, while integration of EscV-TMD6_ex-His was impaired.

To examine whether the GxxxG motif identified within TMD5 is critical for protein activity, we mutated the glycine residues at positions 213 and 217 to either alanine or leucine (G213A, G217A, G213L, and G217L). Due to expression challenges of the mutated proteins tagged with His-tag, we labeled EscV WT and single mutants with the V5 tag, which resulted in a similar secretion profile to EscV_wt-His (Supplementary Figure 2). The single mutants were transformed into ΔescV, and their T3SS activity was examined. We observed that mutations G213A and G217A had similar secretion profiles to the ΔescV strain transformed with EscV_wt-V5, while the single mutation G213L completely abolished T3SS activity (Figure 7A). The effect was much milder when the escV strain was transformed with the EscV G217L mutant (Figure 7A). To confirm proper expression of the EscV point mutation variants, whole-cell lysates were submitted to western blot analysis using anti-V5 antibody. Comparable protein expression was detected for the WT and the single mutants (Figure 7A). Our results suggest that replacement of the glycine residues of the GxxxG motif found in TMD5 by a large residue (leucine) disrupts the activity of the protein while replacement by a small residue (alanine) does not.

To investigate the effect of the single mutation G213L on the assembly of the T3SS complex, we examined the ability of mutant EscV proteins to properly integrate into the T3SS complex. For this purpose, we grew the EPEC WT and EPEC ΔescV strain transformed with EscV_wt-V5, EscV_G213A-V5, and EscV_G213L-V5 under T3SS-inducing conditions. We prepared crude membranes and analyzed them by BN-PAGE and immunoblotting. BN-PAGE analysis showed that the ΔescV mutant strain transformed with EscV_wt-V5 and EscV_G213A-V5 migrated mainly as a large complex at the top of the gel, while the EscV_G213L-V5 integration into the complex appeared to be impaired (Figure 7B). To confirm that the altered running pattern of the EscV_G213L-V5 mutant form was not due to reduced protein expression, the crude membrane extracts were analyzed by SDS-PAGE and immunoblotting using the anti-V5 antibody. We detected a lower expression level of EscV_G213L-V5 relative to EscV_wt-V5 and EscV_G213A-V5, but not to a level that explains the significant reduction in complex formation (Figure 7B). Overall, our results indicate that the GxxxG motif and more specifically the glycine at position 213 are critical for the proper EscV integration into the T3SS complex.