SteC represents a minimal kinase with a depleted C lobe

Despite SteC demonstrating kinase activity towards multiple substrates, a detailed understanding of its catalytic mechanism remains to be determined. The AlphaFold2 predicted structure of SteC shows an N-terminal regulatory domain and C-terminal kinase domain, connected by a linker (Figs 1A and S1A). Within the kinase domain, the N-terminal lobe (subdomains I-IV) mediates anchoring and orienting of the ATP molecule and comparison with the canonical eukaryotic kinase PKA (PDB: 1ATP [16]) demonstrates that this is largely conserved in SteC (Figs 1A and S1B). However, the C lobe, which normally consists of subdomains VIa-XI and mediates substrate binding and initiates phosphate transfer, is notably diminished in SteC. Instead, in SteC, this region spanning amino acids 320–457, is predicted to consist of αE, β6–9 and a further two α-helices (αF and αG) followed by a largely unstructured C-terminal tail (“C-tail”) (amino acids 429–457) (Figs 1A and S1B). Within the C lobe, it is unclear whether SteC contains a DFG motif, normally residing between β8 and β9, within which the aspartate mediates an interaction with the Mg²⁺ ion to aid positioning of the gamma-phosphate or an activation segment, which would canonically start at the DFG motif and extend to the APE motif within αEF. Furthermore, no APE-like motif, normally found at the C-terminal end of the activation segment, is evident in SteC (S1B Fig). Finally, the only crystalised fragment of SteC, covering amino acids 202–375, terminates at the end of αEF, leaving the role of the two remaining alpha helices (αF and αG) and unstructured C-tail unknown.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. SteC is monomeric in solution.

Top: Schematic of SteC, showing two domains connected by a linker. The kinase domain contains N lobe (green) and C lobe (light orange). The C-tail (amino acids 430–457) and the three kinase domain constructs used in this study are indicated. Bottom: cartoon representation of PKA (left) in complex with peptide substrate (black), ATP (red) and magnesium (blue) (PDB: 1ATP, left) and SteC kinase domain AF2 prediction (right) in the same orientation. The N lobes (green), C lobes (light orange) and loops (grey) are indicated, as are the catalytic clefts by a dotted line. adioactive kinase assays of 5 μM SteC_210-457 WT or K256H, SteC_210-429 WT or K256H, SteC_200-377 and SteC_210-375 expressed in E. coli with 100 μM FMNL1 peptide and 100 μM ATP with 20 kBq [γ-32P]ATP. The graph above represents the quantification of three repeats performed with ImageJ, the Coomassie and autoradiography images below show one representative experiment. HDX analysis of SteC_{210-457;K256H} in complex with FMNL1 peptide. Exchange is plotted as a relative difference in deuterium uptake of SteC_{210-457;K256H} alone or in the presence of the FMNL1 K190 peptide across the amino acid range. See S1E Fig for peptide coverage. SEC-MALLS analysis of the three SteC kinase domain fragments SteC_210-457, SteC_210-429, and SteC_210-375, expressed with K256H mutation. The dotted lines represent the corresponding estimated molar mass of the eluting species. Statistical analysis used one-way ANOVA (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001; ns, non-significant) with Dunnett’s multiple comparisons test. Assumption of normally distributed data was accepted using a Shapiro-Wilk and Kolmogoro-Smirnov test.

https://doi.org/10.1371/journal.ppat.1014424.g001

The depleted C lobe of SteC mediates substrate binding and phosphorylation

To investigate the role of the C lobe of SteC, and test whether catalytic residues reside within this region, we expressed and purified four truncated fragments of SteC: the entire predicted kinase domain (210–457), the kinase domain lacking the C-tail (210–429), which is predicted to be unstructured and of unknown function, and constructs 200–377 and 210–375 which lack αF and αG as well as the C-tail (Figs 1A and S1C) and which closely resemble the recently crystallised protein construct [8]. 1D ¹H-NMR analysis revealed that SteC_200-377, with the shortest C-terminal region, produced a spectrum with the sharpest peaks covering a wider range of chemical shifts (S1D Fig) indicative of a well folded construct. Although a subset of peaks in the spectra of the longer SteC constructs, SteC_210-457 and SteC_210-429, have the same chemical shifts as in the SteC_200-377 spectrum, they exhibited significant peak broadening. This effect could be attributed to the constructs’ higher molecular weights and the presence of chemical exchange from either inter- or intra-molecular interactions. Interestingly, given the similarity of SteC_210-457 and SteC_210-429 spectra, there is little evidence that the additional C-terminal residues are completely flexible. This suggests that contrary to the AlphaFold2 prediction, the C-tail of SteC might exist in equilibrium between free and bound states, where it interacts intramolecularly with the SteC core (S1D Fig).

Next, the ability of SteC_210-375, SteC_210-429 and SteC_210-457 to mediate substrate phosphorylation was analysed by radioactive kinase assays, using the reported substrate FMNL1 [7]. Incubation of FMNL1_1–458 with SteC and ATP resulted in the identification of four phosphorylation sites (S185, S199, S203 and S207) (S1 Table). As this protein is prone to aggregation, an FMNL1 peptide, spanning amino acids K190-K213 was generated. Whereas SteC_200-377 and SteC_210-375 were inactive, SteC_210-457 and SteC_210-429 phosphorylated the FMNL1 peptide, and this was dependent on the invariant lysine, K256 (Fig 1B). Absence of substrate phosphorylation by SteC_210-375 and SteC_200-377 corroborates previous findings [8] and is unsurprising as this truncated form of SteC lacks a significant portion of the C-lobe which forms part of the substrate binding module. To identify the regions of SteC involved in substrate binding, Hydrogen Deuterium Exchange (HDX) Mass Spectrometry was used to determine solvent accessibility of SteC_210-457 with and without the FMNL1 peptide. Detected SteC peptides covered the entire protein (S1E Fig). Amino acids 388–396 of SteC within αF of the kinase domain C lobe showed the strongest differential, indicating that this region becomes protected from solvent in the presence of FMNL1 peptide (Fig 1C). We conclude that the minimal kinase domain required for in vitro phosphorylation spans amino acids 210–429 of SteC with regions of the minimal C lobe essential for catalysis and substrate binding.

SteC kinase domain is monomeric in solution

Recently, it was suggested, based on a dimeric crystal structure of an SteC construct encompassing amino acids 202–375, that the kinase domain of SteC functions through an unusual dimeric arrangement in which one monomer supports the catalytic function of the other [8]. As our data reveal an essential role for additional regions of the C lobe, we used size exclusion chromatography coupled to multi-angle laser light scattering (SEC-MALLS) to determine the oligomeric state of both inactive and active SteC constructs in solution. Analysis of SteC_210-375, SteC_210-429 and SteC_210-457 at multiple concentrations clearly demonstrated that each fragment is monomeric in solution (Fig 1D). Therefore, even though dimerisation can activate or inactivate certain kinases [17], our in-solution analysis of the active kinase domain suggests that dimerisation does not represent a key regulatory feature of the activity of SteC.

SteC is phosphorylated at S379

Recombinant SteC becomes phosphorylated in vitro upon incubation with ATP, suggesting it undergoes autophosphorylation, and this appears to be dependent on the C-tail (Fig 1B and Poh et al [9]). To identify putative sites of phosphorylation that might mediate activation of the kinase, WT and K256H (catalytically inactive) SteC_210-457 and SteC_210-429 were expressed and purified from E. coli and analysed by mass spectrometry (92% peptide coverage of the kinase domain). We identified one phosphorylation site in wild-type SteC present in a peptide with the sequence SVSLATR. The phosphorylated peptide identified has a unique signature by mass spectrometry which reveals the phosphorylation site to be S379 (Fig 2A). Next, we analysed the phosphorylation state of different constructs of SteC, in the absence of further incubation with ATP. In SteC_210-457, S379 was phosphorylated to approximately 40%, whereas in construct SteC_210-429 around 25% of peptides contained phosphorylated S379 (Fig 2B). Incubation with ATP did not change the level of phosphorylation in E. coli expressed SteC_210-457 or SteC_210-429, and no phosphorylated S379 peptides were identified when kinase inactive SteC_{210-457;K256H} was analysed, with or without the addition of ATP (Fig 2B, for quantification see S2 Table). We hypothesised that autophosphorylation occurs during the expression of SteC in E. coli, possibly with the protein in a semi-unfolded ‘prone-to-autophosphorylation’ state, as reported for other kinases [18,19]. Given the lack of additional phosphorylation at S379 upon the addition of ATP, it is likely that once the kinase domain is fully folded, it becomes unable to mediate autophosphorylation at S379. This seemingly contradicts the low autophosphorylation signal observed by autoradiography for SteC_210-457, which occurs in vitro upon the addition of radio-labelled ATP (Fig 1B). We speculate that this may be due to a low level of phosphorylation at a site not covered in the peptide map.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. SteC is phosphorylated on S379 mediating activity and nucleotide binding.

SteC expressed in E. coli and Sf9 cells was analysed by mass spectrometry. The peptide SVSLATR (S377-R383) was detected in S379-phosphorylated and unphosphorylated forms following fragmentation obtained for precursor ion at m/z 407.2. C-terminal (y) ions and N-terminal (b) ions are presented here. y ions at 627 and 726 and the b3 ion at 354 indicate presence of phosphorylation. 529 represents unphosphorylated y5 minus H₂O. S379 phosphorylated / non-phosphorylated peptides ratios are presented as determined by mass spectrometric analysis of recombinantly expressed protein for the indicated conditions. See also S2 Table. Radioactive kinase assays performed with 5 μM SteC_210-457 expressed in E. coli and Sf9, 100 μM FMNL1 peptide and 100 μM ATP with 20 kBq [γ-32P]ATP. Above, quantification of three repeats performed with ImageJ above, and below, a representative gel and autoradiography image. AlphaFold2 prediction of the SteC kinase domain. N lobe secondary structural elements are highlighted in green and C lobe in light orange. The putative activation segment is highlighted in teal from D364 of the DGD motif to αF, including β9 and αEF. S379, at the C-terminal end of the activation segment, is shown as a purple stick. Radioactive kinase assays were conducted with SteC_210-457 WT or mutant variants expressed in E. coli and analysed in vitro with 5 μM SteC protein, 100 μM FMNL1 peptide and 100 μM ATP with 20 kBq [γ-32P]ATP. Above, quantification of three repeats using ImageJ above, and below, representative Coomassie and autoradiography images below. Radioactive kinase assays with SteC_1-457 WT and mutant variants together with MYL12A. All proteins were expressed from E. coli. Kinase assays were conducted with 100 nM of SteC protein, 5 μM of MYL12A and 100 μM ATP with 20 kBq [γ-32P]ATP. Above, quantification of 3 repeats using ImageJ, below representative Coomassie -stained SDS gel and respective autoradiography images. Fluorescence titrations of 500 nM mant-AMPPNP with SteC_210-429 WT and mutants expressed in E. coli. The increase in fluorescence at increasing protein concentrations was measured. Data are representative of 2 repeats for each construct. Fitting curves are shown as purple lines. See also S2C Fig. Radioactive kinase assays of SteC_210-429 WT and mutants expressed in E. coli at 5 μM (left) or with 1 μL murine cell lysate (right) using 100 μM ATP with 20 kBq [γ-32P]ATP. Quantification of three repeats using ImageJ above, representative Coomassie and autoradiography images below. Statistical analysis used one-way ANOVA (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001; ns, non-significant) with Dunnett’s multiple comparisons test (C, E & F). Assumption of normally distributed data was accepted using a Shapiro-Wilk and Kolmogoro-Smirnov test.

https://doi.org/10.1371/journal.ppat.1014424.g002

We considered that the N-terminal domain of SteC may play a role in kinase autophosphorylation, yet full-length SteC_1-457 expressed in E. coli was aggregation-prone so couldn’t be purified to the degree necessary for this experiment. Therefore, we purified the full-length protein from insect cells (S2A Fig). Mass spectrometric analysis of this protein showed a very high degree of S379 phosphorylation (Fig 2B). Additionally, this protein became phosphorylated at S75 during an in vitro kinase assay (S3 Table). Interestingly, the kinase-deficient mutant, SteC_1-457;K256H, was also phosphorylated at S379 upon expression in Sf9 cells, with a similar proportion of pS379/S379 identified as in the wild type protein (Fig 2B). This observation suggests that phosphorylation at S379 of SteC is mediated by a eukaryotic kinase during insect cell expression and is not an autophosphorylation event. In summary, even though E. coli expressed SteC showed some degree of autophosphorylation at S379, with little pS379 peptide signal detected for the kinase-deficient mutant, when SteC was expressed in insect cells it was predominately in an S379 phosphorylated form, even in a mutant that is kinase deficient and hence unable to autophosphorylate [18].

S379 phosphorylation is required for the activity of SteC

Next, we compared the catalytic activity of SteC_210-457 after expression in E. coli or insect cells against the FMNL1 peptide substrate. This comparison showed that protein expressed from insect cells was significantly more active than from bacteria (Fig 2C). S379 is located between αEF and αF of the C lobe, which encompasses regions of the kinase often referred to as the activation segment (Figs 2D and S1B). As phosphorylation of residues within the activation segment often control kinase function [20,21], we hypothesised that phosphorylation of S379 might represent such a regulatory mechanism within SteC. To test whether phosphorylation of S379 activated kinase activity, mutation S379A, which cannot be phosphorylated and the phospho-mimetic variant, S379D, were introduced into SteC_210-457 and proteins were expressed and purified from bacteria. As expected, WT SteC_210-457 mediated phosphorylation of the FMNL1 peptide whereas SteC_{210-457;K256H} did not (Fig 2E). Phosphorylation of the peptide substrate was diminished upon incubation with SteC_{210-457;S379A}, with phosphorylation of the FMNL1 peptide restored and increased when S379 was mutated to aspartic acid (Fig 2E). Interestingly, SteC autophosphorylation was also evident (Fig 2E), revealing that additional SteC residue(s) beyond S379 become phosphorylated under in vitro conditions not identified by our mass spectrometry analysis. To investigate the role of S379 phosphorylation in the full-length protein and against another substrate MYL12A, the following SteC_1-457 variants K256H, S379A and S379D plus WT were expressed in E. coli. As expected, SteC_1-457 mediated phosphorylation of MYL12A⁸, whereas the kinase deficient mutant did not. As before, mutation of S379 to an alanine reduced the activity of SteC whereas SteC_1-457;S379D retained activity (Fig 2F). Therefore, we define S379 as a new phosphorylation site required for efficient activity of SteC towards known substrates.

S379 phosphorylation mediates nucleotide binding

Next, we interrogated why S379 is important for the catalytic activity of SteC. Phosphorylation of the activation loop as a mechanism to activate protein kinases [20] might occur through increased substrate binding [22]. To test whether S379 phosphorylation alters substrate binding we monitored the ability of the minimal active kinase domain, SteC_210-429, to interact with an N-terminally biotinylated FMNL1 peptide by biolayer interferometry. We determined the following binding constants; WT SteC_210-429 yielded a K_d = 7.9 ± 1.4 μM, the inactive K256H variant had a K_d = 22 ± 3 μM, S379A gave K_d = 21 ± 3 μM and S379D K_d = 14 ± 2 μM (S2B Fig). Given these similar affinities, we next tested the hypothesis that phosphorylation of S379 mediates allosteric conformational changes that alter nucleotide binding. This was monitored by analysing binding of a fluorescent non-hydrolysable ATP analog (mant-AMPPNP) to SteC_210-429 by fluorescence spectroscopy. The data showed that WT SteC_210-429, which was partially phosphorylated (Fig 2B) and the S379 phosphomimetic form, SteC_{210-429;S379D}, bound mant-AMPPNP with affinities of 5.4 ± 0.4 μM and 11.8 ± 1.5 μM respectively (Figs 2G and S2C). In contrast, when S379 was mutated to alanine (SteC_{210-429;S379A}), blocking S379 phosphorylation, or when SteC_{210-429;K256H} was analysed, which also lacks S379 phosphorylation (Fig 2B), mant-AMPPNP binding was significantly reduced and no binding affinity could be calculated (Figs 2G and S2C). From this, we propose that phosphorylation of S379 activates SteC by increasing its affinity for ATP.

S379 is phosphorylated by a mammalian kinase independent of SteC kinase activity

Kinases that require activation loop phosphorylation usually have an arginine residue adjacent to the catalytic aspartate and this forms the so called “HRD” motif within the C lobe of the kinase domain [14]. The aspartate is the most conserved residue of this motif and interacts directly with the substrate to orientate the hydroxyl acceptor group. Dai et al [8] proposed D364 performs this function, yet this conclusion is based on the structure of a minimal inactive construct of SteC. Based on sequence and structural predications (Figs 1A and S1B), we hypothesised that D344 represents part of an HRD-like motif, which in the case of SteC, lacks the arginine, but retains the histidine and aspartate conserved in all eukaryotic protein kinases [14]. To explore the hypothesis that a mammalian kinase phosphorylates SteC, the phosphorylation of bacterially expressed SteC_210-429 was analysed in vitro following incubation of recombinant protein with mammalian cell lysate. In this experiment the additional putative catalytic mutant, D344A, was analysed as an alternative negative control. Phosphorylation of SteC_210-429, SteC_{210-429;K256H} and SteC_{210-429;D344A} was observed but when S379 was mutated to alanine (SteC_{210-429;S379A}) no phosphorylation was detected (Fig 2H). This demonstrates that a mammalian kinase can phosphorylate SteC at residue S379. Furthermore, it suggests that without S379 phosphorylation, any subsequent phosphorylation of SteC is not observed.

SteC contains non-canonical catalytic motifs that facilitate C lobe function

So far, our data reveal a new phosphorylation site of SteC, S379, that mediates kinase activity. As noted above, SteC contains an HD at residues 343–344 (S1B Fig) and we therefore hypothesised that this motif, positioned just after β6, might be required for function (Fig 3A). Cells expressing GFP-tagged SteC showed evident actin foci formation, and this was absent in cells that expressed SteC_K256H or SteC_S379A (Figs 3B and S3A). Expression of either SteC_H343A or SteC_D344A was unable to induce actin foci formation, suggesting that these amino acids are required for the catalytic potential of SteC. Indeed, recombinant SteC_210-429 with an alanine substitution at D344 showed minimal phosphorylation of the FMNL1 peptide, demonstrating the importance of this residue for SteC activity and function (Fig 3C).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. S379 is essential for SteC induced actin polymerisation during infection.

AlphaFold2 prediction of SteC kinase domain with the putative activation segment highlighted in teal from D364 including the DGD motif, β9, αEF and S379 prior to αF, N-lobe in green and C-lobe in sand. H343 and D344 are shown in purple sticks, indicating their position in relation to the catalytic cleft and activation segment. HEK 293ET cells were transfected with vectors expressing SteC WT and mutants. 100 transfected cells were quantified by a blind scorer per coverslip and conducted in biological triplicate (see also S3A Fig). Each dot represents a biological repeat, and bars represent mean and standard deviation. Radioactive kinase assay of SteC_210-429 WT and mutants expressed in E. coli at 5 µM with 100 µM FMNL1 peptide. Above, quantification of three repeats using ImageJ, below representative Coomassie and autoradiography images. Representative super-resolution images of Swiss 3T3 mouse fibroblasts infected with ΔsteC Salmonella strains expressing the indicated SteC variants fixed at 8 hours post infection. Scale bars represent 5 μm. In merge panel, phalloidin is in magenta, CSA (Salmonella) in yellow and DAPI in cyan. In infected cells, white arrows denote dense actin poylmerisation and pink arrows denote speckled actin polymerisation. For each Salmonella strain indicated 3–5 coverslips from two independent infections were prepared at 8 hours post infection and 100 infected 3T3 cells were scored (blind) for the formation of dense actin polymerisation associated with the Salmonella microcolony. Each data point represents a coverslip, and bars represent the mean and standard deviation. For each Salmonella strain indicated 3–5 coverslips from two independent infections were prepared at 8 hours post infection and 100 infected 3T3 cells were scored (blind) for the formation of dense actin polymerisation associated with the Salmonella microcolony (data reported in Fig 3E). Infected cells lacking dense actin polymerisation were scored for speckled actin polymerisation and reported in this graph. Each data point represents a coverslip, and bars represent the mean and standard deviation. Significance calculation represents a paired comparison between WT and each other condition. Statistical analysis used one-way ANOVA (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001; ns, non-significant) with Dunnett’s multiple comparisons test. Assumption of normally distributed data was accepted using a Shapiro-Wilk and Kolmogoro-Smirnov test.

https://doi.org/10.1371/journal.ppat.1014424.g003

Interestingly, in line with the observations by Dai et al. [8], mutation of D362 or D364 to alanine also ablated SteC induced actin polymerisation whereas mutation of D279 did not (Figs 3B and S3A). This D₃₆₂GD motif is positioned just after β8, where the canonical DFG motif, which mediates interaction with the Mg²⁺ ion to aid positioning of the gamma-phosphate, is found (S3B and S1B Figs). This raises the intriguing possibility that D₃₆₂GD might represent a non-canonical DFG motif.

Surprisingly, E272, previously predicted to stabilise the invariant lysine in subdomain II and ATP [9], was not required for the function of SteC (Figs 3B and S3B), nor its catalytic activity towards the FMNL1 peptide (Fig 3C). One possible explanation for this is that E272 is found outside alpha helix of subdomain III (αC), where the αC glutamate would normally be positioned (S1B Fig and PDB: 8JBI [8]). Regardless, it remains unclear how SteC fulfils this important function.

Finally, we tested the hypothesis that R280, which has a positive charge and resides on αC helix (S1B Fig), might be required to stabilise the negative charge upon S379 phosphorylation due to its position and orientation in the AlphaFold2 predicted structure (S3C Fig). GFP-tagged SteC_R280A was unable to induce actin polymerisation supporting this hypothesis (Figs 3B and S3A). We therefore propose that even though SteC lacks a classical HRD motif, the phosphate of pS379 might instead interact with R280 as a means of allosterically altering the active site to facilitate catalysis. Ultimately, why R280 and the putative ‘HD motif’ with the catalytically important D344 are required for function requires further investigation and the structure of the entire kinase module bound to a substrate. However, to date, protein stability has represented a limiting step towards achieving this. Furthermore, whether the N-terminal portion of SteC (amino acids 1–193) contributes to kinase activity or substrate selectivity remains enigmatic.

S379 and the C-tail of SteC is required for dense actin polymerisation at the microcolony of infected cells

SteC mediates the polymerisation of actin into dense foci associated with the Salmonella micro-colony [9]. To test whether S379 was required for dense actin foci formation during infection, 3T3 fibroblasts were infected with steC mutant bacteria expressing either HA-tagged WT or mutant SteC. Immunoblotting the pellet and the post-nuclear supernatant fractions of infected cells for HA and the Salmonella protein DnaK, demonstrated that each HA-tagged construct was expressed (pellet fraction containing bacteria) and translocated (supernatant fraction lacking bacteria) into the host cell cytosol (S3D Fig). As expected, cells infected with WT Salmonella showed association of dense actin, stained with phalloidin, at the microcolony (Fig 3D), and this occurred in approximately 40% of infected cells (Fig 3E). Actin polymerisation was significantly reduced when cells were infected with steC mutant Salmonella and restored upon expression of WT SteC from the bacteria (Fig 3D,3E). Cells infected with steC mutant Salmonella strains expressing either SteC_K256H or SteC_S379A were unable to induce dense actin polymerisation, though a less obvious ‘speckled’ actin polymerisation was associated with microcolonies of these strains (Fig 3D). Expression of SteC_S379D from steC mutant bacteria restored the ability of SteC to induce dense actin foci around the micro-colony to levels that were comparable to cells infected with bacteria expressing WT SteC (Fig 3D,3E). The Salmonella strain expressing SteC_1-429 was not associated with any actin polymerisation within the infected cell (Fig 3D,3E). Quantification of speckled actin polymerisation in each condition was performed by the same method as for dense actin polymerisation. 36% of cells infected with WT Salmonella showed speckled actin polymerisation. This was significantly lower for steC mutant bacteria, as well as for mutant bacteria complemented with pWSK29 empty plasmid, psteC and psteC_1-429; significantly higher for mutant strains complemented with psteC_K256H, psteC_S379A; and similar for the mutant strain complemented with psteC_S379D (Fig 3F).

To investigate this further, 3T3 cells infected with mutant strains of Salmonella were fixed and stained at 8 hours post infection and the intensity of polymerised actin associated with the Salmonella microcolony was analysed by automated 3D segmentation (S3E Fig). This showed that the mean intensity of actin at the microcolony was strongest in cells infected with steC mutant bacteria carrying WT SteC and significantly reduced in cells infected with bacteria expressing SteC_K256H, SteC_S379A and SteC_1-429 but not SteC_S379D. Together we conclude that both S379 and the C-tail of SteC control actin polymerisation. Both the C-tail and phosphorylation of S379 are required for SteC-induced dense actin polymerisation. The newly described speckled accumulation of actin at the micro-colony, whilst kinase-independent is dependent on the C-tail of SteC, overall highlighting an essential role of the C-tail.

S379 is conserved across Salmonella species and bacterial SteC homologs

SteC is highly conserved in most Salmonella serovars (Fig 4A). Analysis of 879 clinical isolates of Salmonella revealed that K256, H343, D344, D362, D364 and S379 residues are all conserved (S4A Fig). Furthermore, alignment of putative SteC homologues from Yokenella regensburgei (98% coverage, 41% amino acid identity), Cedecea neteri (92%, 42%), Sodalis praecaptivus (86%, 34%) and Erwinia mallotivora (42%, 49%) showed that these six residues are conserved across all the homologues (S4B Fig). These findings suggest a common catalytic and phosphorylation-regulated mechanism of SteC across diverse bacterial species.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. S379 and key catalytic motifs are conserved in SteC and its bacterial homologs.

GrapeTree [26] visualisation of SteC distribution in 879 Salmonella complete genomes. The MStree was constructed using the rMLST scheme, which could differentiate Salmonella at serovar level based on ribosome loci. Each bubble represents a distinct rST. The branch lengths correspond to the number of allele differences at the ribosome loci. Nodes with only a single allele difference were collapsed into bubbles, which exhibit high concordance with serovars. The size of each bubble is proportional to the number of genomes it represents. The colours of the bubbles represent presence, absence or pseudogenisation of SteC. The phylogenetic analysis is not rooted, meaning evolutionary relationships cannot be inferred. Representative super resolution microscopy analysing subcellular localisation of HA tagged SteC variants following the infection of 3T3 cells at 8 hours post infection. CSA labels Salmonella and Phalloidin stains actin. In the merge, CSA is shown in yellow, phalloidin in magenta and HA in cyan. Two examples are shown for mutant strains complemented with SteC WT and SteC 1–429. Scale bars represent 5 μm. The Pearson’s correlation coefficient of the mean intensity of CSA signal and HA signal within the cell mask was calculated for each strain. Each data point represents a field of view, each box represents the mean and interquartile range (IQR), while whiskers extend to 1.5xIQR. Independent two-sided T-test with Welch correction was used to determine p-values. * p < 0.05, ** p < 0.01, *** p < 0.001. Pearson’s correlation coefficient of the mean intensity of phalloidin and HA signals analysed as for Fig 4C.

https://doi.org/10.1371/journal.ppat.1014424.g004

Interestingly, however, the C-tail is not highly conserved in the putative homologues (S4B Fig). We noted that unlike WT SteC, translocation of SteC_1-429, which lacks the C-tail, was unable to polymerise actin (Fig 3D-3F). Yet, when assayed in vitro, SteC_1-429 phosphorylates the FMNL1 peptide (Fig 1B) and when ectopically expressed in HEK 293ET cells also retains activity (Figs 3B and S3A). To investigate this further, the subcellular localisation of HA-tagged SteC was analysed using super resolution microscopy, with steC mutant bacteria carrying an empty plasmid analysed as a control (Fig 4B, row 1). In 3T3 cells infected with Salmonella expressing WT SteC-HA, SteC formed structures that colocalised with actin foci (Fig 4B, row 2) and was occasionally found around individual bacteria (Fig 4B, row 3). In cells infected with Salmonella expressing SteC_1-429, the HA signal was readily apparent but was more diffuse and punctate (Fig 4B, rows 4 & 5). To try and quantify this, cells infected with steC mutant Salmonella carrying either an empty plasmid or expressing HA-tagged SteC_1-457 or SteC_1-429 were fixed, stained and imaged. Analysis by Pearson’s Correlation Coefficient of CSA(Salmonella):HA revealed a significant increase in correlation when comparing cells infected with bacteria carrying either HA-tagged SteC_1-457 or SteC_1-429 to cells infected with steC mutant Salmonella (Fig 4C). This corroborates our previous observation that the C-tail mutant of SteC is translocated and suggests that the C-tail does not significantly alter SteC localisation to the microcolony. Then, as expected, the correlation of HA (SteC) to actin was significantly increased when cells were infected with WT HA-tagged SteC and this was entirely dependent on the C-tail of SteC (Fig 4D). We conclude that during infection, the C-tail of SteC, despite being dispensable for SteC kinase activity and actin polymerisation upon ectopic expression, is an essential regulator of SteC function during infection. We hypothesise that such regulation of enzymatic activity, perhaps through substrate selection, provides an important layer of control that is only revealed under physiologically relevant conditions.

In summary, instead of becoming activated via dimerization, as suggested by a previous model, or through interaction with a binding partner, as observed for the minimal kinase Shigella effector OspG [15], the kinase activity of SteC is induced through phosphorylation at S379. Phosphorylation of SteC at S379, which is required for dense actin polymerisation, represents a catalytically activating event that induces a dramatic increase in kinase domain nucleotide binding affinity. Even though we are unable to fully tease apart the relative contributions of co-translational autophosphorylation versus phosphorylation by a host kinase, SteC expressed in E. coli and incubated with a mammalian cell lysate was readily modified at residue S379. Furthermore, both the WT and catalytically inactive SteC variants were phosphorylated at S379 upon expression in insect cells, pointing to a model involving an upstream kinase. Importantly, SteC purified from insect cells is three-fold more active than E. coli expressed protein, where only co-translational autophosphorylation occurs. Therefore, whilst it remains to be determined whether co-translational autophosphorylation can occur upon translocation of SteC from Salmonella into the host cell, our data suggest a model whereby a host kinase is able to phosphorylate S379 of SteC to yield a fully active kinase.

Several other Salmonella effectors are controlled by host-mediated PTMs. SifA, for example is prenylated and S-acetylated by host machinery which leads to both increased membrane binding and Salmonella survival in infected mice [23], and is also cleaved by caspase 3 to be activated [24], whereas S-palmitoylation of SseI regulates its plasma membrane localisation and function [25]. Altogether, the reliance of modification(s) by host enzymes might represent a mechanism to ensure that the activity of a given effector is only unleashed in the host and only in the correct subcellular localisation within the host.

Protein structure prediction

Experimentally derived protein structures were retrieved from the Protein DataBank (PDB) (https://www.rcsb.org/). AlphaFold2 (AF) predicted structures were downloaded from the AlphaFold Protein Structure Database ([href:https://alphafold.ebi.ac.uk/]https://alphafold.ebi.ac.uk/). The steC gene sequence with UniProt reference D0ZIB5 from Salmonella Typhimurium strain 14028s was used for all bioinformatic analyses and protein expression and purification. PDB and predicted structures were visualised using Pymol ([href:https://www.pymol.org/]https://www.pymol.org/).

Plasmid design and cloning

Expression plasmids were produced using Gibson assembly [27] (Tables 1 and 2). Vector backbones used were pET28 (either with C-terminal uncleavable His6 or N-terminal Twin-Strep tag with TEV protease cleavage site for E. coli expression), pACEBac (either N-terminal His6-3c cleavage site or N-terminal Twin-Strep TEV protease cleavage site for insect cell expression), pTCMV (N-terminal GFP tag, for ectopic expression in mammalian cells) and pWSK29 (C-terminal 2HA tag for transformation of Salmonella). For expression of SteC and associated variants (K256H, S379A, S379D, and SteC_1-429) from Salmonella, the steC open reading frame and C-terminal 2HA tag, was synthesised as a G-block from IDT technologies and inserted by Gibson Assembly into pWSK29 plasmid, which already contained the SteC promoter. Where required, mutagenesis was performed using the Quickchange protocol.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Recombinant DNA used for expression in E. coli and insect cells.

https://doi.org/10.1371/journal.ppat.1014424.t001

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Recombinant DNA used for expression in mammalian cells.

https://doi.org/10.1371/journal.ppat.1014424.t002

Protein expression and purification

Competent BL21 (DE3) E. coli cells were transformed with relevant plasmids, grown to mid-log phase at 37°C in LB, measured by OD₆₀₀ of 0.7-1, and expression was induced with 0.5 mM Isopropyl-β-D-thiogalactoside (IPTG) at 20 °C. After 16–24 hours of expression, cells were pelleted and either frozen for future use or lysed directly. To lyse, the fresh or frozen cells were resuspended in a buffer of 50 mM HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) pH 7.5, 300 mM NaCl, 10% glycerol (v/v), 10 mM MgCl₂, 20 mM imidazole (for His-tagged proteins only), 0.5 mM tris(2-carboxyethyl)phosphine hydrochloride (TCEP), 1 x protease inhibitor cocktail (Roche) and 2 μg/ml DNAse (Merck). Lysis was performed by sonication. The lysate was cleared by ultracentrifugation at 48,000 x g for 1 hour at 4 °C. Lysate supernatant was passed over affinity resin, either using nickel nitrilotriacetic acid (NiNTA) or Streptactin resin affinity chromatography, in a gravity column at 4 °C. Resin washing was performed with Buffer A (50 mM HEPES pH 7.5, 300 mM NaCl, 10% glycerol plus 20 mM imidazole for His-tagged proteins) and high salt (50 mM HEPES pH 7.5, 1 M NaCl) prior to elution with buffer B (50 mM HEPES pH 7.5, 300 mM NaCl, 10 mM MgCl₂, 10% glycerol, 0.5 mM TCEP and either 300 mM imidazole or 2.5 mM desthiobiotin). All proteins were purified at 4°C and 10% glycerol added to prevent aggregation. SteC_1-457 aggregated on SEC column, making this protein hard to purify to large quantities. SteC_210-457 eluted from SEC but was aggregation-prone, so for some assays where the C-tail was not deemed to be mechanistically important, SteC_210-429 was used.

For insect cell expression, EmBacY cells were transformed with the pACEBac plasmid by electroporation pulsed at 2 kV with a Gene Pulser II (BioRad) and incubated at 37 °C for 6 hr to stimulate transposition. 100 μL of these cells were incubated for 48 hr on plates with antibiotics selecting for the bacmid, the transposed plasmid and a helper plasmid. DNA was extracted from positive colonies and transfected into Sf9 cells. Cells expressing GFP denoted the production of viable virus. Virus was amplified to reach the potency required for adequate protein expression. The P2 or P3 virus was deemed potent when > 80% of cells were infected and the cell viability was 85–92%. Sf9 at 2 x 10⁶ cells/ml were infected with potent virus at a ratio of 500:1 and incubated for 72 hours at 28 °C with shaking. Protein derived from insect cells was purified as described for E. coli expressed protein with the following small adjustments: Lysis buffer included 0.1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF) serine protease inhibitor. Cell lysis was performed with 0.1% Triton added to the lysis buffer with stirring at 4 °C for 45 min. For resin affinity purification, batch washing by resin centrifugation was performed prior to gravity column washing due to the viscosity of the lysate. Finally, elution of His-tagged proteins expressed in Sf9 required serial elution with buffers containing 300 mM and 500 mM imidazole.

For SteC kinase domains and MYL12A, size exclusion chromatography (SEC) was the final purification step. SEC buffer comprised 50 mM HEPES pH 7.5, 150 mM NaCl, 10 mM MgCl₂, 0.5 mM TCEP and 10% glycerol (v/v), unless otherwise stated. For large scale protein purifications, HiLoad 16/600 Superdex (200 or 75) columns were used on an AKTA prime system at 4 °C with a flow rate of 1 ml/min with 2 ml fractionation. For smaller scale purifications, Superdex 200 or 75 Increase 10/300 GL SEC columns were used, with a flow rate of 0.5 ml/min with 500 μL fractionation. In the cases of SteC full length and FMNL1_1–458, SEC was not possible due to protein aggregation and proteins were dialysed into SEC buffer overnight at 4 °C. Protein was concentrated using Amicon Ultra centrifugal filters, of 50, 30 or 10 kDa filter sizes. Protein concentration was measured using absorbance at 280 nm on a Nanodrop instrument. Aliquots of protein were flash frozen in liquid nitrogen and stored at -80 °C. For biophysical techniques potential aggregation was cleared by ultracentrifugation at 17,000 x g at 4 °C for 10 min, protein dialysed where necessary, and the concentration re-estimated using a Jasco V-760 Spectrophotometer.

1D NMR

One-dimensional ¹H nuclear magnetic resonance (1D NMR) protein spectra were recorded at 25 °C on Bruker AVANCE spectrometers operating at 800 MHz in NMR buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 10 mM MgCl2, 0.5 mM TCEP and 5% D₂O). Data were acquired and processed with Topspin (Bruker).

Hydrogen-deuterium exchange mass spectrometry

Hydrogen-deuterium exchange mass spectrometry (HDX-MS) was performed with 5 µL of 5 µM proteins (individually or in combination) incubated with 40 µL of D₂O buffer at room temperature for 3, 30, 300 and 3000 seconds in triplicate. The labelling reaction was quenched by adding chilled 2.4% v/v formic acid in 2 M guanidinium hydrochloride and immediately frozen in liquid nitrogen. Samples were stored at -80 °C prior to analysis. The quenched protein samples were rapidly thawed and subjected to proteolytic cleavage by pepsin followed by reversed phase high-performance liquid chromatography (HPLC) separation. Briefly, the protein was passed through an Enzymate ethylene-bridged hybrid (BEH) immobilised pepsin column (Waters, UK) at 200 µL/min for 2 min and the peptides trapped and desalted on a 2.1 x 5 mm C18 trap column (Acquity BEH C18 Van-guard pre-column, 1.7 µm, Waters, UK) then eluted. Peptides were separated on a reverse phase column (Acquity UPLC BEH C18 column 1.7 µm, 100 mm x 1 mm (Waters, UK). Peptides were detected on a Cyclic mass spectrometer (Waters, UK). Peptide identification was performed by MS^e [28]. The resulting MS^e data were analysed using Protein Lynx Global Server software (Waters, UK). Mass analysis of the peptide centroids was performed using DynamX software (Waters, UK). All time points in this study were prepared at the same time and individual time points were acquired on the mass spectrometer on the same day.

Size exclusion chromatography coupled to multi-angle laser light scattering

Size exclusion chromatography coupled to multi-angle laser light scattering (SEC-MALLS) was performed with 100 μL protein samples at concentrations of 0.25, 0.5, 1, 2, 4 and 8 mg/ml were first applied to a Superdex 200 10/300 INCREASE GL column equilibrated in 50 mM HEPES pH 7.5, 150 mM NaCl, 10 mM MgCl₂, 0.5 mM TCEP and 3 mM NaN₃ to separate species by hydrodynamic volume. Scattered light intensity was measured using a DAWN HELEOS II laser photometer, and dRI was measured using an OPTILAV-TrEX differential refractometer. The weight-averaged molecular mass of proteins was determined using the ASTRA software version 7.0 (Wyatt Technology Corp., Santa Barbara, CA) assuming Dn/dc to be 0.186 mL/g.

Radioactive kinase assays

Reactions were performed in standard SEC buffer. Each 15 μL reaction mixture contained 50 nM - 5 μM kinase (details stated in Figure legends), 100 μM ATP, 20 kBq [γ-32P]ATP (Hartmann Analytic, 9.25 MBq, 25 μL pot, 5000 Ci/mmol). Given the aggregation-prone nature of SteC_210-457, for concentrations below that detectable by Coomassie staining, protein aliquots were thawed, thoroughly mixed and concentration recalculated using Nanodrop measurement, prior to serial dilution to the concentration required. FMNL1 K190 24mer peptide (pep_K190) was used at 100 μM (given its small molecular weight, this concentration was required for visualisation by Coomassie staining). Reactions were incubated at 30 °C for 30 min prior to SDS-PAGE analysis, Coomassie staining, gel drying and exposure to radiography film in the dark for between 30 min and 18 hours before developing. 1 μl Swiss 3T3 mouse fibroblast lysate (see below details of its preparation) was used to spike relevant kinase assay samples. Quantification of kinase assays was performed using band quantification in ImageJ.

Phosphorylation mass spectrometry

SteC_1-457 WT and K256H and FMNL1_1–458 expressed in insect cells were reduced and alkylated in-gel (10 mM TCEP, 40 mM chloroacetamide for 20 min at 70°C) prior to trypsin digestion (modified sequencing grade, Promega) in 10 mM NH₄HCO₃ overnight at 37°C. Acidified supernatant was separated by high-performance liquid chromatography and loaded into a Lumos Tribrid Orbitrap mass spectrometer (all Thermo Scientific). Raw files were searched using Maxquant (maxquant.org) against FASTA sequences of relevant recombinant constructs, recent downloads of UniProt baculovirus related databases and a common contaminants database and modified residues were assigned. Visualisation was in Perseus (maxquant.net/perseus) and Skyline (skyline.ms).

Biolayer interferometry

Bio-Layer Interferometry (BLI) was performed on an Octet Red instrument (Fortebio/Sartorius) operating at 25 °C. The N-terminally biotinylated FMNL1 peptide was synthesised by the Chemical Biology STP (Francis Crick Institute). The peptide was solubilised in, and proteins were dialysed into 50 mM HEPES pH 7.5, 150 mM NaCl, 10 mM MgCl₂ and 0.5 mM TCEP. 0.05% Tween-20 was added to the samples prior to the experiments. Octet Streptavidin (SA) biosensors were loaded with the biotinylated peptide (1 μg/ml) and then exposed to SteC protein concentrations ranging from 7.5 nM to 200 µM. Association and dissociation curves were recorded for each concentration. Control experiments with no peptide loaded on the sensors were recorded to correct for non-specific interactions of SteC proteins with SA sensors. Data were analysed using Octet BLI Analysis software (Sartorius) and in-house software [29]. The equilibrium dissociation constant (K_d) was determined from the instrument response against SteC proteins concentration using least squares non-linear regression. The reported error is the mathematical error of the fit.

Nucleotide binding

SteC proteins were analysed in SEC buffer. Mant-Adenylyl imidodiphosphate (AMPPNP) was acquired from Jena Bioscience GmbH. Data were collected on a Jasco FP-8500 Spectrofluorometer using an excitation wavelength of 355 nm and recording emission at 390–550 nm, in a 0.3 cm path length quartz cuvette (Hellma Analytics). Protein titrations were performed by recording full spectra and adding small volumes of protein to mant-AMPPNP at 500 nM. The signal at 442 nm was baseline subtracted and corrected for protein samples contributions and dilution, prior to plotting against the protein concentrations. Data were fitted using non-linear least squares regression with in-house software [29]. Average values and standard deviations were calculated from two independent measurements.

Salmonella strains

Salmonella WT 14028s and the same strain in which genomic steC had been replaced with a kanamycin cassettes were from Poh et al., 2008 [9]. To generate Salmonella strains expressing SteC:2HA and associated variants from pWSK29 the ΔsteC strain was transformed by electroporation. Salmonella ΔsteC was grown in Luria Broth media (LB) to an OD₆₀₀ of 0.4. The culture was cooled on ice for 30 min prior to washing the bacteria three times with autoclaved MilliQ-purified H₂O and then autoclaved MilliQ H₂O with 10% glycerol (once). The bacterial cell pellet was resuspended in 200 µl 10% glycerol and electroporation was performed with 50 µl of cell suspension and 100 ng DNA in a 0.2 mm cuvette at 2.5 kV, 200 ohms and 25 µF with a Pulse Controller Plus (BioRad). The product was incubated for 1 hr at 37 °C in 500 µl SOC and plated on LB agar plates with kanamycin and carbenicillin. Single colonies were selected and stored in 33% glycerol at -80 °C.

Cell lines and cell lysate preparation

Human Embryonic Kidney 293 cells (HEK 293ET cells) and Swiss 3T3 mouse fibroblasts (3T3 cells) were grown in Dulbecco’s Modified Eagle’s Medium (DMEM) (Sigma, USA) supplemented with 10% foetal bovine serum (FCS) (Gibco Life Sciences, UK) at 37 °C in 5% CO₂. To produce murine cell lysate, two confluent 10 cm plates of 3T3 cells were washed in phosphate buffered saline (PBS) once then lysed in the plate with 1 ml Lumier ++ (Tris pH 7.4, 150 mM NaCl, 0.1% Triton, 1% EDTA, protease inhibitor cOmplete (Roche) and phosphatase inhibitor phosSTOP (Roche)) with 0.3% Triton for 15 min on ice, before pelleting at 17,700 x g for 10 min at 4 °C. 20 µl aliquots of the supernatant were stored at -20 °C.

Salmonella infection of mammalian cells

Overnight cultures of S. Typhimurium strains were diluted 1:33 in fresh LB and grown for 3.5 hrs at 37 °C with shaking (200 rpm) to obtain logarithmic phase bacteria. Cultures were added directly to 3T3 cells at a multiplicity of infection of 100:1. Bacterial invasion was allowed to proceed for 20 min at 37 °C in 5% CO₂ after which the cells were washed twice in PBS and fresh media containing 100 μg / ml gentamicin was added for one hour prior to exchange with media containing 20 μg / ml gentamicin for the remainder of the experiment. A list of all bacterial strains is presented in Table 3.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. List of bacterial strains.

https://doi.org/10.1371/journal.ppat.1014424.t003

In cellulo analysis of SteC translocation

3T3 cells seeded on 6-well plates were infected with Salmonella WT, or the indicated ΔsteC strains expressing HA-tagged variants of SteC. Eight hours after invasion, cells were washed in PBS and lysed with 50 μl Lumier ++ (Tris pH 7.4, 150 mM NaCl, 0.1% Triton, 1% EDTA, protease inhibitor and phosphatase inhibitor) for 10 min on ice and centrifuged at 17,700 x g at 4 °C. After centrifugation the soluble layer representing the post-nuclear supernatant (PNS) was denatured by addition of 20 μl of 5x SDS buffer (125mM Tris-Cl pH 6.8, 4% SDS, 10% glycerol, bromophenol blue and 5% β-mercaptoethanol). The pellet fraction, containing bacteria and mammalian cell nuclei, was denatured with addition of 100 μl of 2x SDS loading buffer. Samples were then boiled at 95°C for 7 min. Pellet samples were sonicated briefly to reduce viscosity prior to SDS-PAGE and immunoblot analysis with antibodies against HA, DnaK and GAPDH (Table 4).

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. List of antibody reagents used in the study.

https://doi.org/10.1371/journal.ppat.1014424.t004

Immunoblotting

Denatured protein samples were run on 10–14% polyacrylamide gels by electrophoresis. A constant voltage of 110 V was applied for 90 min. Proteins were transferred onto a 0.2 μm polyvinylidene fluoride or polyvinylidene difluoride (PVDF) transfer membrane (Millipore) using a Trans-Blot Turbo Transfer System (BioRad). Membranes were blocked in 5 ml 5% bovine serum albumin (BSA) in Tris-buffered saline with 100 mM Tris pH 7.4, 150 mM NaCl, 0.1% Tween 20 (TBS-T) for 30 min at room temperature on a roller. The membranes were incubated with primary antibodies in 5% BSA overnight at 4 °C followed by three washes in TBS-T before being exposed to the appropriate secondary antibody in 5% BSA for 1 hr at room temperature. Next, after three further washes in TBS-T, membranes were incubated with either ECL detection reagents or Pierce ECL Plus Western Blotting Substrate and imaged on an iBright FL1500 Imaging System (Invitrogen, USA). See antibody list for details (Table 4).

Transfection of mammalian cells

HEK 293ET cells were seeded onto sterile glass coverslips coated with Poly-L-lysine (Sigma, USA) 24 hours prior to transfection. The cells were transfected with 3 μL Lipofectamine 2000 (Invitrogen, USA) and 600 ng of the indicated pTCMV GFP-SteC mutant vectors (Table 2) in 250 μL Opti-MEM Reduced Serum Medium (Gibco, UK). The cells were incubated at 37 °C for 24 hours prior to analysis.

Immunofluorescence microscopy

3T3 or 293ET cells, seeded in 24 well plates on glass coverslips were infected or transfected as described. At the desired time point, cells were washed in PBS and fixed using 3% paraformaldehyde (PFA, Sigma-Aldrich) in PBS at room temperature for 30 min. After washing in PBS, the cells were quenched with 100 mM NH₄Cl in PBS and then permeabilised and blocked using 0.1% Triton with 10% horse serum (Sigma) or, in the case of experiments including HA staining, 0.1% Saponin with 10% horse serum in PBS at room temperature for 1 hour. Cells were then incubated with the indicated primary antibodies as detailed in the Table 4. Actin was stained using phalloidin and DNA with stained with 4’6-diamidino-2-phenylindole (DAPI). After several washes in PBS coverslips were then incubated with the associated secondary antibodies. Coverslips were mounted onto slides using Aqua-Poly/Mount (Polysciences, Inc.) and dried overnight at room temperature in the dark.

For transfection experiments, images were captured using a confocal laser scanning microscope (LSM710) (Zeiss GmbH) with a 63x objective. Counting was performed manually, by assessing the actin polymerisation of 100 transfected cells for each cover slip, with 3 coverslips per condition across 2 independent infections. To capture super-resolution images of infected cells, an Olympus IX-83 inverted confocal microscope with Yokogawa CSU-W1 SoRa spinning disc module was used, equipped with Olympus UPLAPO OHR 60x/1.50 oil immersion objective in 3.2x magnification mode, Photometrics Prime BSI sCMOS camera, and laser lines at 405, 488, 561, and 640 nm. Fields of view were chosen using the 488 channel, corresponding to CSA, to ensure imaging of infected cells without biasing on secondary phenotypes. Olympus cellSens Dimension 4 software was used for z-stack acquisition, with a pixel size of 33.854 nm/px, and a z-spacing of 0.21 μm. Images were post-processed using Olympus Super Resolution (OSR) with medium setting for the 488nm channel and low for the others, and subsequently underwent Constrained Iterative deconvolution with 5 iterations.

Image analysis proceeded by thresholding and 3D segmentation of Salmonella signal and actin to determine extents of microcolonies and cells respectively. Segmented microcolonies were modelled as convex hulls and Delaunay triangulation was used to determine points enclosed within the bounds of the microcolony. Segmented cells were modelled as stacks of 2D concave hulls to define areas of the images within and exterior to cells. The relation between Salmonella microcolony and actin formation was then quantified by the ratio of the mean intensity of actin within 5 pixels of microcolonies compared to the mean intensity within cell bodies for each field of view. All images were processed in Fiji (Image J) and are displayed in colour-blind friendly combinations. To quantify actin foci during in infection, an Axio Imager Upright Microscope (Zeiss) was used. The scorer was blind to which condition was being scored. 3–5 coverslips from 2 independent infections were analysed.

Protein sequence analysis

The NCBI Basic Local Alignment Search Tool (BLAST) was used to identify homologues of SteC (UniProt reference D0ZIB5) ([href:https://blast.ncbi.nlm.nih.gov/Blast.cgi]https://blast.ncbi.nlm.nih.gov/Blast.cgi). To compare the different SteC protein sequence types among Salmonella serovars, 879 complete Salmonella genomes were downloaded from Enterobase by searching “Complete Genome” in the “Status” field, which represents the highest assembly quality with circular chromosomes and plasmids ([href:https://enterobase.warwick.ac.uk/]https://enterobase.warwick.ac.uk/, accessed on 2023/06/30). The SISTR1 results from Enterobase were used to identify the subspecies and serovars of the genomes. The steC nucleotide sequence from Salmonella Typhimurium LT2 (RefSeq: GCF_000006945.2) was used as a reference. A BLAST database was constructed from the steC sequence. Each of the 879 Salmonella genomes was queried against the steC database using BLASTn v2.14.0 + ³⁰. The aligned DNA sequences were then extracted and translated into protein sequences using Seqkit v2.4.0 [31]. The unique SteC protein sequences were summarised and aligned using Clustalo v1.2.4 [32]. To visualise the SteC types in the Salmonella subspecies and serovars, an MStree of the 879 complete Salmonella genomes was generated on Enterobase using the rMLST scheme with the MSTree2 algorithm [33]. The tree was visualised with GrapeTree [26]. The sequence logo figure for the regions of interest was generated using the ggseqlogo package ([href:https://github.com/omarwagih/ggseqlogo]https://github.com/omarwagih/ggseqlogo) in R [34].

Statistical analysis

In most cases, statistical significances were calculated using an ordinary one-way analysis of variance (ANOVA) complemented with a post-hoc test for multiple comparison’s corrections, as described in figure legends. When comparing automatically generated image data, independent two-sided T-test with Welch correction was used to determine p-values. * – < 0.05, ** – < 0.01, *** – < 0.001. For the analysis of correlation between actin and HA signal, Pearson’s Correlation Coefficient was applied. All analyses were completed on GraphPad Prism (Version 10.1.1).