CYRI/FAM49B negatively regulates RAC1-driven cytoskeletal remodelling and protects against bacterial infection

Abstract

Salmonella presents a global public health concern. Central to Salmonella pathogenicity is an ability to subvert host defences through strategically targeting host proteins implicated in restricting infection. Therefore, to gain insight into the host–pathogen interactions governing Salmonella infection, we performed an in vivo genome-wide mutagenesis screen to uncover key host defence proteins. This revealed an uncharacterized role of CYRI (FAM49B) in conferring host resistance to Salmonella infection. We show that CYRI binds to the small GTPase RAC1 through a conserved domain present in CYFIP proteins, which are known RAC1 effectors that stimulate actin polymerization. However, unlike CYFIP proteins, CYRI negatively regulates RAC1 signalling, thereby attenuating processes such as macropinocytosis, phagocytosis and cell migration. This enables CYRI to counteract Salmonella at various stages of infection, including bacterial entry into non-phagocytic and phagocytic cells as well as phagocyte-mediated bacterial dissemination. Intriguingly, to dampen its effects, the bacterial effector SopE, a RAC1 activator, selectively targets CYRI following infection. Together, this outlines an intricate host–pathogen signalling interplay that is crucial for determining bacterial fate. Notably, our study also outlines a role for CYRI in restricting infection mediated by Mycobacterium tuberculosis and Listeria monocytogenes. This provides evidence implicating CYRI cellular functions in host defence beyond Salmonella infection.

Main

Enteropathogenic bacteria are a major public health concern, accounting for over 300 million foodborne illnesses and 60% of related fatalities worldwide. Among these bacteria, Salmonella enterica subspecies enterica serotypes are associated with the largest disease burden1, with non-typhoidal Salmonella alone responsible for approximately 93.8 million illnesses and 155,000 deaths annually worldwide2. This, together with the widespread emergence of antibiotic-resistant Salmonella strains3, calls for a better understanding of Salmonella virulence mechanisms and the associated host cellular responses.

Underlying Salmonella pathogenicity is an ability to hijack host cellular processes through targeting key host signalling proteins4. This is, in part, mediated by bacterial effectors that are delivered into the host cytosol5,6. Among the targeted host proteins, cytoskeletal regulators are particularly important, as their modulation by Salmonella effectors is crucial for the mediation of infection at different stages7. For example, Salmonella uptake depends largely on the timely activation of the host small Rho GTPase RAC1. By mimicking host guanine nucleotide exchange factors (GEFs), activators of small GTPases, the bacterial effector SopE stimulates RAC1 activity8,9. This triggers the recruitment of active RAC1 to the WAVE regulatory complex (WRC), a heteropentameric protein assembly comprising CYFIP1/2, NAP1, ABI2, HSPC300 and WAVE10,11. Binding of active RAC1 to the WRC through CYFIP proteins, in turn, drives actin polymerization and membrane ruffling, which are necessary for Salmonella engulfment by macropinocytosis12,13,14,15,16,17,18. Intriguingly, the Salmonella-mediated modulation of host cytoskeletal proteins also triggers antimicrobial immune responses19,20. However, Salmonella has evolved various strategies to evade and even exploit host defences, including strategically manipulating host factors implicated in restricting infection21. As such, deciphering anti-Salmonella host-defence mechanisms promises to reveal key host–pathogen interactions governing Salmonella infection.

To address this, we performed a genome-wide mutagenesis screen in mice to pinpoint host proteins that counteract Salmonella infection. This identified FAM49B as a host protein that confers Salmonella resistance. Structural and biochemical analyses revealed that FAM49B is related to CYFIP proteins and can similarly interact with RAC1. Based on this, we renamed FAM49B and its coding gene as CYRI (CYFIP-related RAC1 interacting protein). We show that, unlike CYFIP proteins, CYRI negatively regulates RAC1 signalling and restricts SopE-mediated bacterial entry. In addition, CYRI plays a critical role in myeloid-derived phagocytes, which can provide a Salmonella replicative niche and aid bacterial dissemination22,23,24,25. In particular, we demonstrate that CYRI impedes phagocytosis and phagocyte cell migration, thereby attenuating Salmonella infection. As a counterdefence, we found that Salmonella negatively modulates CYRI protein levels in a SopE-dependent manner. Together, these findings provide mechanistic insights into a previously uncharacterized Salmonella host defence mechanism and highlight additional strategies employed by Salmonella to subvert host responses. Our findings also implicate CYRI in the restriction of infection mediated by Mycobacterium tuberculosis and Listeria monocytogenes. This indicates that the cellular functions of CYRI might constitute a general antibacterial protective mechanism that is relevant in the context of other intracellular bacterial pathogens.

Results

An ENU-induced CYRI mutation confers susceptibility to Salmonella infection

Salmonella enterica subspecies enterica serovar Typhimurium (S. Typhimurium) is among the most predominant serotypes associated with human disease26. We therefore performed an N-ethyl-N-nitrosourea (ENU) genome-wide mutagenesis screen in mice to uncover proteins implicated in host defence against S. Typhimurium. As part of the screen, we identified a variant pedigree, namely Immunity to Typhimurium locus 15 (Ity15), that exhibited a significant reduction in survival following S. Typhimurium systemic infection (Fig. 1a,b and Supplementary Note 1). Despite having a broadly normal immune system under naive conditions (Supplementary Note 2 and Supplementary Table 1), in vivo imaging with bioluminescent Salmonella revealed a progressive increase in the bacterial load of Ity15 mutant (Ity15m/m) mice compared with their littermate controls (Ity15+/+; Fig. 1c). Consistent with this, infected Ity15m/m mice displayed significantly higher spleen and liver bacterial loads (Fig. 1d) and increased levels of serum pro-inflammatory (TNFα and IL-6) and anti-inflammatory (IL-10) cytokines relative to Ity15+/+ mice (Supplementary Fig. 1a). Moreover, Salmonella-infected Ity15m/m mice presented multiple abscesses in the spleen and liver. In particular, lesions were more diffuse in the spleens of Ity15m/m mice with the white pulp showing marked lymphocytolysis. In addition, we detected more parenchymal necrosis together with frequent fibrin thrombi in the microvasculature in liver sections of Ity15m/m mice (Supplementary Fig. 1b). Overall, these pathological changes are indicative of an overwhelming septicaemia in Ity15m/m mice. Indeed, elevated levels of circulating bacteria were detected in the blood of Ity15m/m mice (Fig. 1e), which provides further support for this idea. A similar trend was also observed following oral infection, with higher spleen and liver bacterial loads detected in mutant mice (Supplementary Fig. 1c). Histopathological analyses of the caeca from orally infected mice also revealed increased oedema, enhanced polymorphonuclear neutrophil infiltration of the gut submucosa, higher goblet cell counts and altered epithelial integrity in Ity15m/m mice one day post infection (Supplementary Fig. 1d). Collectively, these findings demonstrate the increased susceptibility of Ity15m/m mice to S. Typhimurium infection.

Fig. 1: The Ity15 pedigree displays increased susceptibility to S. Typhimurium infection.
figure1

a, Breeding scheme for the production of ENU-induced mutant mice. The G2 female (Chloe) was backcrossed to its G1 father (Jody) to generate the Ity15 pedigree G3 offspring. b, Survival curves for the indicated number of G3 offspring from WT 129S1 mice and the Ity15 pedigree following S. Typhimurium infection. Data were pooled from six independent experiments. The depicted significance is relative to the survival curve of the 129S1 mice. c, Xen26-luminescent S. Typhimurium replication kinetics in Ity15+/+ and Ity15m/m mice over seven days. The graph represents the log10 total photon flux per second ± s.e.m. from two independent experiments with three mice for each genotype per experiment. The image depicts representative Salmonella bioluminescence in Ity15+/+ and Ity15m/m mice at day 6 post infection. d, Bacterial load in the spleens and livers of Ity15+/+ and Ity15m/m mice at days 3, 4 and 5 post S. Typhimurium infection. The graphs represent the c.f.u. counts per weight of organs derived from the indicated number of mice and the mean ± s.e.m. from one experiment. Data are representative of five independent experiments. e, Bacterial load in the blood of Ity15+/+ and Ity15m/m mice at day 5 post S. Typhimurium infection. The graph represents the c.f.u. count per ml of blood derived from the indicated number of mice and the mean ± s.e.m. from two independent experiments. *P ≤ 0.05, ***P < 0.001 and ****P < 0.0001. The statistical analyses are detailed in Supplementary File 2.

To pinpoint the causative mutation underlying Salmonella susceptibility in Ity15m/m mice, we performed genome-wide linkage analysis in an F2 intercross and mapped the mutation to an interval between 63.4 and 67.8 Mb on chromosome 15 (Supplementary Fig. 2a and Supplementary Note 3). F2 mice homozygous for the mutant allele at the peak marker on chromosome 15 succumbed to Salmonella infection by day 8, whereas heterozygous (Ity15+/m) and Ity15+/+ mice survived infection effectively (Supplementary Fig. 2b). Exome sequencing identified a T to A substitution within the splice donor site of exon 9 in Cyri (Fam49b), a protein-coding gene that maps to the critical interval on chromosome 15 (Fig. 2a and Supplementary Note 3). This results in exon 9 skipping, causing a frameshift and the introduction of a premature stop codon within exon 10 (p.N212GfsX9; Fig. 2a and Supplementary Fig. 2c), thereby abrogating the expression of CYRI protein in Ity15m/m mice (Fig. 2b). Together, these findings support a role for CYRI in determining susceptibility to Salmonella.

Fig. 2: A mutation in Cyri confers susceptibility to S. Typhimurium infection in the Ity15 pedigree.
figure2

a, Schematic representation of the ENU-induced Cyri mutation. b, Western blot analysis of CYRI protein expression in the spleens and livers of Ity15+/+ and Ity15m/m mice at day 0 and 5 post S. Typhimurium infection. β-Actin was used as a loading control. Data are representative of three independent experiments. c, Allelic complementation assay. Survival curves following S. Typhimurium infection of the indicated number of F1 mice derived from a cross between Ity15+/m and Cyri+/− mice. Data were pooled from four independent experiments. The depicted significance is relative to the survival curve of Ity15+/Cyri+ mice. ****P < 0.0001. Statistical analyses are detailed in Supplementary File 2.

To further validate Cyri as the causal gene driving the phenotype observed in mutant mice, we performed an allelic complementation assay in which Ity15+/m mice were crossed to mice heterozygous for a knockout (KO) allele at Cyri (Cyri+/−). We then assessed the survival of F1 animals following Salmonella infection to ascertain susceptibility (Supplementary Fig. 2d). This revealed a lack of complementation with Ity15m/Cyri mice displaying increased susceptibility similar to that observed in Ity15m/m mice (Fig. 2c), thus confirming that the observed Cyri mutation mediates Salmonella susceptibility in Ity15m/m mice. Together, this identifies Cyri as a Salmonella resistance gene with the loss of CYRI expression resulting in enhanced Salmonella dissemination and septic shock.

CYRI is a CYFIP-related RAC1 effector

Given the uncharacterized nature of CYRI at the time of this study, we next explored its potential cellular functions. CYRI shares 80% sequence identity with FAM49A—a closely related uncharacterized protein—within a single domain, domain of unknown function 1394 (DUF1394). Interestingly, this domain is also present in two other proteins in the human proteome: CYFIP1 and CYFIP2 (Supplementary Fig. 3a). CYFIP proteins are part of the WRC and are known to interact with active RAC1 to drive actin polymerization14,15. This raised the possibility that CYRI might also bind to RAC1. Indeed, in vitro biochemical assays demonstrated that CYRI associates with RAC1 (Fig. 3a and Supplementary Fig. 3b). Consistent with this, Duolink in situ proximity ligation assays against endogenous RAC1 and exogenous CYRI performed in HeLa cells expressing CYRI–haemagglutinin (HA) under a doxycycline (dox)-inducible promoter further confirmed this interaction. In particular, dox-treated CYRI–HA-expressing cells displayed a markedly higher Duolink count compared with untreated cells, indicating increased CYRI–RAC1 binding (Fig. 3b,c). Glutathione S-transferase (GST) pulldown experiments using purified GST–RAC1 loaded with either GDP or GTPγS to mimic inactive and active RAC1, respectively, also revealed that CYRI preferentially binds to active RAC1 (Supplementary Fig. 3c). In addition, to further validate CYRI as an active RAC1 interacting protein, we assessed the binding of purified GST–CYRI to endogenous RAC1 from GDP- or GTPγS-pretreated HEK293T lysates. As expected, increased levels of active RAC1 were detected in the lysates pretreated with GTPγS following PAK-CRIB pulldown to specifically enrich the active form of RAC1 (Fig. 3d). Importantly, enhanced binding was observed specifically between GST–CYRI and endogenous RAC1 from GTPγS-pretreated lysates (Fig. 3e,f). This confirms that, similar to CYFIP proteins, CYRI–RAC1 binding also depends on RAC1 activation.

Fig. 3: CYRI binds to the small GTPase RAC1 in a GTP-dependent manner.
figure3

a, GST pulldown of GST and GST–CYRI incubated with the lysates of HEK293T cells. Co-precipitated endogenous RAC1 was detected by western blotting. Both α-tubulin and Ponceau staining were used as loading controls. Data are from one experiment. b, Western blot analysis of the lysates from untreated (−dox) or dox-treated (+dox) HeLa Flp-In T-REx CYRI–HA cells. Endogenous and exogenous levels of CYRI protein were detected. α-Tubulin was used as a loading control. Data are representative of three independent experiments. c, CYRI–RAC1 binding Duolink count quantification using antibodies against endogenous RAC1 and exogenous CYRI in the indicated number of −dox and +dox HeLa Flp-In T-REx CYRI–HA cells. An HA intensity threshold of ≥50 was applied to +dox cells. The graph represents the values for individual cells and the mean ± s.e.m. from one experiment. Data are representative of three independent experiments. d, PAK-CRIB pulldown of GDP- and GTPγS-loaded HEK293T lysates. The levels of active and total RAC1 protein were detected by western blotting. α-Tubulin was used as a loading control. Data are representative of three independent experiments. e, GST pulldown of GST and GST–CYRI incubated with GDP- or GTPγS-loaded HEK293T lysates. Co-precipitated endogenous RAC1 was detected by western blotting. Both α-tubulin and Ponceau staining were used as loading controls. Data are representative of three independent experiments. f, Quantification of GDP- and GTPγS-loaded endogenous RAC1 binding to GST–CYRI normalized to the binding observed in GDP-loaded lysates. The graph represents the individual values and the mean ± s.e.m. from three independent GST pulldown experiments. g, GST pulldown of GTPγS-loaded GST and GST–RAC1 incubated with lysates from HEK293T CYRI/FAM49A KO cells expressing CYRI–HA WT, P150R, R161D, and P150R + R161D. Co-precipitated CYRI–HA was detected by western blotting. Both α-tubulin and Ponceau staining were used as loading controls. Data are representative of three independent experiments. h, Quantification of CYRI–HA binding to GST–RAC1 normalized to binding observed in the lysates of cells expressing CYRI–HA WT. The graph represents the individual values and the mean ± s.e.m. from three independent GST pulldown experiments. *P ≤ 0.05; **P < 0.01; ****P < 0.0001 and ns, non-significant. Statistical analyses are detailed in Supplementary File 2. WB, western blot.

Despite a low sequence homology of approximately 20%, analysis of CYRI using HHpred27 denoted a significant structural similarity between CYRI and the CYFIP1 DUF1394 domain (Supplementary Fig. 3d). A reported WRC structure identified cysteine 179 (C179) and arginine 190 (R190) within the CYFIP1 DUF1394 domain as part of a RAC1 binding site, with mutations of these residues to R and aspartic acid (D), respectively, abolishing RAC1 binding14. We therefore generated similar mutations in the corresponding CYRI residues, namely P150 and R161 (Supplementary Fig. 3e,f). We then individually expressed CYRI-HA wild type (WT), P150R, R161D and the P150R + R161D double mutant in a CYRI/FAM49A double KO HEK293T cell line and assessed their binding to GTPγS-loaded purified GST–RAC1. Consistent with the high conservation between CYRI R161 and CYFIP1 R190, the R161D mutant displayed significantly reduced RAC1 binding. In contrast, mutation of the CYRI P150 residue had no significant effect on the CYRI–RAC1 interaction (Fig. 3g,h). However, the P150R + R161D double mutant exhibited an even higher reduction in RAC1 binding compared with the R161D mutant alone (Fig. 3g,h), indicating that although individually non-essential, P150 contributes to an efficient CYRI–RAC1 interaction. Together, this demonstrates that CYRI is related to CYFIP proteins and can similarly interact with RAC1 through residues within the DUF1394 domain.

CYRI negatively regulates RAC1 signalling

Both the structural homology to CYFIP proteins and its ability to bind to RAC1 hinted at a potential role of CYRI in regulating RAC1-driven actin cytoskeletal modulation. To test this, we generated CYRI KO HeLa cells using clustered regularly interspaced short palindromic repeats (CRISPR)–CRISPR associated protein 9 (Cas9) technology (Supplementary Fig. 4a) and examined the changes in cell morphology. CYRI KO cells exhibited a round, unpolarized and flattened pancake-like morphology with evident membrane ruffling, increased cellular spread and circularity compared with parental and control KO cells (Fig. 4a–d and Supplementary Fig. 4b). To confirm that these effects were due to CYRI deletion, we next expressed single guide RNA (sgRNA)-resistant (sgRES) CYRI–HA WT, P150R and R161D in CYRI KO HeLa cells under a dox-inducible promoter. Following dox treatment, CYRI expression was detected in these cells, as evident by an increased HA signal compared with control KO and CYRI KO cells lacking the dox-inducible sgRES CYRI–HA system (Supplementary Fig. 4c). Despite selecting cells with similar expression levels (Supplementary Fig. 4d), reconstitution with CYRI–HA WT, but not the mutants, rescued the morphological changes associated with CYRI KO. Specifically, only CYRI–HA WT expression reduced the surface area and circularity to a level comparable to control KO cells (Fig. 4e,f). This demonstrates that the CYRI KO morphological phenotype is due to CYRI loss and that CYRI–RAC1 binding is important for CYRI-mediated effects.

Fig. 4: CYRI regulates cell morphology and actin cytoskeletal dynamics in a RAC1-dependent manner.
figure4

a, Fluorescence microscopy images of parental (left), control KO (middle) and CYRI KO (right) HeLa Flp-In T-REx cells depicting merged phalloidin and Hoechst channels to visualize the actin cytoskeleton and nuclei, respectively. Data are representative of three independent experiments. Scale bar, 100 μm. b, Percentage of cells with the outlined morphology categories in the indicated number of parental, control KO and CYRI KO HeLa Flp-In T-REx cells. The graph represents the mean ± s.e.m. from three independent experiments. c,d, Surface area (c) and circularity index (d) of the indicated number of parental, control KO and CYRI KO HeLa Flp-In T-REx cells. The graphs represent the average values per well and the mean ± s.e.m. from three independent experiments. e,f, Surface area (e) and circularity index (f) of the indicated number of +dox control KO, CYRI KO and CYRI KO sgRES CYRI–HA WT, P150R and R161D HeLa Flp-In T-REx cells. An HA intensity threshold of 10–35 was applied to the CYRI KO sgRES CYRI–HA HeLa Flp-In T-REx cells. The graphs represent individual values and the mean ± s.e.m. from one experiment. The dashed red lines highlight the parameter baseline based on the mean of the control KO cells. The depicted significance is relative to control KO cells. g, PAK-CRIB pulldown of the lysates from control KO and CYRI KO HeLa Flp-In T-REx cells. The levels of active, total RAC1 and endogenous CYRI protein were detected by western blotting. α-Tubulin was used as a loading control. Data are representative of three independent experiments. hj, Quantification of active RAC1 protein levels in the indicated cell lines normalized to the levels in the specified controls. The graphs represent individual values and the mean ± s.e.m. from three (h) and four (i,j) independent PAK-CRIB pulldown experiments. k, Quantification of G-actin protein levels in control KO and CYRI KO HeLa cells normalized to the levels in control KO lysates. The graph represents individual values and the mean ± s.e.m. from four independent G/F-actin biochemical assays. *P ≤ 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Statistical analyses are detailed in Supplementary File 2. ns, non-significant.

Interestingly, RAC1 hyperactivation has been shown to induce morphological changes comparable to CYRI KO28, which suggests a link between CYRI and RAC1 activity. To test whether RAC1 activation phenocopies CYRI KO, we generated a dox-inducible system for the expression of Strep-FLAG (SF)–RAC1 WT or SF–RAC1 P29S, a previously reported constitutively active RAC1 mutant29. PAK-CRIB pulldown confirmed that SF–RAC1 P29S is more active than SF–RAC1 WT (Supplementary Fig. 4e). Notably, HeLa cells expressing SF–RAC1 P29S displayed a round and flattened morphology similar to CYRI KO cells (Supplementary Fig. 4f,g), suggesting that RAC1 activity might be enhanced following CYRI deletion. Increased active RAC1 levels were indeed detected in CYRI KO cells compared with control KO cells (Fig. 4g,h). Importantly, reconstitution with sgRES CYRI–HA WT, but not P150R + R16D, significantly reduced the levels of active RAC1 detected in CYRI KO cells (Fig. 4i,j and Supplementary Fig. 4h,i). Together, these findings indicate that CYRI negatively regulates RAC1 signalling through its interaction with RAC1.

Based on these findings, we hypothesized that the morphological changes associated with CYRI KO are probably a consequence of increased RAC1 activation leading to intensified cytoskeletal modulation. To explore this, we examined the actin dynamics following CYRI deletion. Actin cytoskeleton remodelling is largely dictated by the balance between monomeric (G)-actin polymerization into filaments (F-actin) and depolymerization events30. The protein levels of G- and F-actin in control KO and CYRI KO HeLa cells were thus quantified using a biochemical approach. As experimental controls, we utilized lysates treated with phalloidin to drive actin polymerization in vitro as well as HeLa cells pretreated with latrunculin B, an inhibitor of F-actin assembly31, both of which demonstrated effective enrichment of F- and G-actin fractions, respectively (Supplementary Fig. 4j). Importantly, CYRI KO cells displayed reduced levels of G-actin protein compared with control KO cells (Fig. 4k), suggesting enhanced actin polymerization following CYRI KO. This indicates that, unlike CYFIP proteins, CYRI serves as a negative regulator of RAC1 signalling, thereby inhibiting RAC1-driven actin cytoskeletal remodelling.

CYRI negatively regulates SopE-mediated cellular effects

RAC1 activation by the Salmonella bacterial effector SopE and subsequent actin polymerization is crucial for bacterial entry into non-phagocytic enterocytes18. Given that CYRI negatively modulates RAC1 signalling, we postulated that CYRI might impede SopE-mediated cellular effects. To test this, we first examined the effect of green fluorescent protein (GFP)-tagged SopE expression in HeLa cells using a dox-inducible system. Consistent with the described role of SopE as a RAC1 activator8,9, GFP–SopE expression increased active RAC1 levels (Supplementary Fig. 5a). Interestingly, similar to SF–RAC1 P29S expression and CYRI deletion, cells expressing GFP–SopE exhibited a characteristic pancake-like morphology with increased surface area and circularity (Supplementary Fig. 5b–e). To demonstrate that this phenotype is SopE-dependent, we utilized a GFP intensity threshold to select low, medium and high GFP–SopE expressors (Supplementary Fig. 5f,g). This revealed a SopE-dose-dependent effect, with low-expressing cells exhibiting a significant increase compared with dox-untreated cells—a phenotype that was enhanced in cells with medium expression and further augmented in high GFP–SopE expressors (Fig. 5a,b). Next, we deleted CYRI in cells harbouring the GFP–SopE dox-inducible system to directly assess if CYRI influences SopE-mediated cellular effects. Consistent with our previous observations, CYRI KO in dox-untreated cells resulted in an increased cellular surface area and circularity compared with control KO cells. Similarly, control KO cells expressing GFP–SopE displayed a pancake-like morphology. Notably, GFP–SopE expression in CYRI KO cells further enhanced the cellular surface area and circularity when compared with changes induced following SopE expression in control KO cells (Supplementary Fig. 6a–c). This effect was more pronounced in cells expressing low levels of GFP–SopE (Fig. 5c,d and Supplementary Fig. 6d,e), indicating that CYRI deletion is advantageous in the context of low SopE cellular concentrations, which might be relevant during infection. To further explore the effect of CYRI on SopE-mediated cellular effects, we transiently expressed CYRI–HA WT and P150R + R161D in the dox-inducible GFP–SopE HeLa cell line. Using GFP and HA staining we selected cells with similar expression levels of GFP–SopE and CYRI–HA (Supplementary Fig. 6f–i) and assessed their cellular morphology compared with dox-untreated and treated mock-transfected cells. Strikingly, CYRI–HA WT expression completely reverted SopE-driven cellular spreading to a level comparable to dox-untreated mock-transfected cells, as evident by a significant reduction in surface area (Fig. 5e,f). Moreover, although circularity was not affected (Fig. 5g), GFP–SopE and CYRI–HA WT co-expressing cells did not display the characteristic flattened pancake-like morphology (Fig. 5e). In contrast, expression of the CYRI P150R + R161D mutant had no effect on either surface area or circularity following SopE expression (Fig. 5e–g). Collectively, this identifies a role for CYRI in attenuating SopE-mediated cellular events that is dependent on its ability to interact with RAC1.

Fig. 5: CYRI negatively regulates SopE-mediated cellular effects.
figure5

a,b, Surface area (a) and circularity index (b) of the indicated number of untreated or dox-treated HeLa Flp-In T-REx GFP–SopE cells. The specified GFP intensity thresholds (low, 18–20; medium, 20–25 and high, ≥ 25) were applied to the dox-treated cells. The graphs represent the values for individual cells and the mean ± s.e.m. from three independent experiments. c,d, Surface area (c) and circularity index (d) of the indicated number of −dox and +dox control KO and CYRI KO HeLa Flp-In T-REx GFP–SopE cells. A low GFP intensity threshold was applied to the +dox cells. The graphs represent the values for individual cells and the mean ± s.e.m. from three independent experiments. Dashed red lines highlight the parameter baseline based on the mean of the –dox control KO cells. e, Fluorescence microscopy images of −dox and +dox HeLa Flp-In T-REx GFP–SopE cells, transfected as indicated, depicting HA immunofluorescence (top) and merged HA, phalloidin and Hoechst channels (bottom) to visualize CYRI–HA, the actin cytoskeleton and nuclei, respectively. Data are representative of three independent experiments. Scale bar, 100 μm. f,g, Surface area (f) and circularity index (g) of the indicated number of −dox and mock-transfected or +dox and mock-, CYRI–HA WT- or P150R + R161D-transfected HeLa Flp-In T-REx GFP–SopE cells. Intensity thresholds for GFP (40–70) and HA (25–1,200) were applied to +dox mock (GFP)- and CYRI–HA (GFP and HA)-transfected cells. The graphs represent the values for individual cells and the mean ± s.e.m. from three independent experiments. The dashed red lines highlight the parameter baseline based on the mean of the −dox mock-transfected cells. h, Bacterial count in the indicated number of individual S. Typhimurium-infected parental, control KO and CYRI KO HeLa Flp-In T-REx cells. The box plots outline the data distribution (minimum, first quartile, median, third quartile and maximum) from one experiment. i, Bacterial load in the BMDMs from Ity15+/+ and Ity15m/m mice 1 h post infection with non-opsonized S. Typhimurium at the stationary (left) and late-logarithmic (right) growth phases. The graphs represent the c.f.u. counts in BMDMs derived from the indicated number of mice and the mean ± s.e.m. from one experiment. Data are representative of three independent experiments. *P ≤ 0.05; **P < 0.01; ****P < 0.0001. Statistical analyses are detailed in Supplementary File 2. ns, non-significant.

To ascertain the functional consequence of the identified CYRI–SopE interplay in the context of Salmonella infection, we next assessed bacterial entry in parental, control KO and CYRI KO HeLa cells as an in vitro model for bacterial internalization in non-phagocytic cells. Interestingly, CYRI deletion resulted in increased Salmonella uptake relative to parental and control KO cells (Fig. 5h and Supplementary Fig. 6j). Consistent with this, an increased bacterial load was detected in bone marrow-derived macrophages (BMDMs) isolated from Ity15m/m mice compared with Ity15+/+ mice following infection with non-opsonized S. Typhimurium to favour Salmonella-driven internalization over phagocytosis. Importantly, a similar increase was observed following infection of BMDMs with a SopE-deficient Salmonella strain (∆SopE) both at the stationary and logarithmic growth phase (Fig. 5i). These findings indicate that CYRI counteracts Salmonella infection through antagonizing SopE cellular functions, thus inhibiting the actin cytoskeletal rearrangements required for bacterial entry.

CYRI modulates cytoskeletal processes within the haematopoietic compartment

RAC1-driven actin cytoskeleton remodelling also plays a major role in phagocytosis32 and cell motility33. This is particularly relevant in myeloid-derived phagocytes, such as macrophages and neutrophils, which can foster bacterial replication and dissemination in vivo22,23,24,34,35. We therefore evaluated the ability of CYRI to confer Salmonella resistance in cells of haematopoietic origin by generating bone-marrow chimaeric mice using bone marrow cells from Ity15+/+ and Ity15m/m mice. Strikingly, Ity15+/+ mice receiving bone marrow cells from Ity15m/m mice displayed increased Salmonella susceptibility. Conversely, enhanced resistance was observed in Ity15m/m mice following the transfer of bone marrow cells from Ity15+/+ mice (Fig. 6a), thus demonstrating the importance of CYRI within the haematopoietic compartment for counteracting Salmonella infection in vivo.

Fig. 6: CYRI expression in haematopoietic cells confers resistance to S. Typhimurium infection.
figure6

a, Survival curves of the indicated number of bone-marrow chimaeras following S. Typhimurium infection. Bone marrow cells were transferred between mice as indicated by the arrows. Data are representative of two independent experiments. The depicted significance is relative to the survival curve of the Ity15+/+Ity15+/+ chimaeras. b, Survival curves of control (Cyri+/flox) and myeloid-specific conditional CYRI KO (Cyri−/flox) mice following S. Typhimurium infection. Data are representative of two independent experiments. The depicted significance is relative to the survival curve of Cyri+/flox mice. c, Bacterial load in the spleens and livers of Cyri+/flox and Cyri−/flox mice four days post S. Typhimurium infection. The graph represents the c.f.u. counts per weight of organs derived from the indicated number of mice and the mean ± s.e.m. from three independent experiments. d, Percentage of infected BMDMs in the indicated number of cells derived from Cyri+/flox and Cyri−/flox mice following infection with opsonized S. Typhimurium. The graph represents the number of infected cells per field and the mean ± s.e.m. from three independent experiments. e, Bacterial count per cell in the indicated number of BMDMs derived from Cyri+/flox and Cyri−/flox mice following infection with opsonized S. Typhimurium. The box plots outline the data distribution (minimum, first quartile, median, third quartile and maximum) from three independent experiments. f, Normalized tracks of the indicated number of neutrophils derived from Cyri+/flox (left) and Cyri−/flox (right) mice, migrating for 1 h. g, Average speed per time step of the indicated number of neutrophils derived from Cyri+/flox and Cyri−/flox mice. The graph represents the values for individual cells and the mean ± s.e.m. h, Average total distance travelled by the indicated number of neutrophils derived from Cyri+/flox and Cyri−/flox mice. The graph represents the values for the individual cells and the mean ± s.e.m. i, Three-dimensional mean squared displacement (3D MSD) plot of the indicated number of Cyri+/flox and Cyri−/flox-derived neutrophils. The data for fi were pooled from five (Cyri+/flox) or four (Cyri−/flox) independent experiments. **P < 0.01; ***P < 0.001; ****P < 0.0001. Statistical analyses are detailed in Supplementary File 2. ns, non-significant.

To further delineate the role of CYRI within the haematopoietic compartment, we next generated myeloid-specific conditional CYRI KO mice (Cyri−/flox; Supplementary Fig. 7a and Supplementary Note 4). The deletion of CYRI in myeloid cells significantly reduced survival following Salmonella systemic infection (Fig. 6b). Increased levels of several serum pro-inflammatory cytokines and chemokines produced by activated neutrophils and monocytes were also observed in Salmonella-infected Cyri−/flox mice (Supplementary Fig. 7b and Supplementary Table 2). In addition, Cyri−/flox mice displayed higher spleen and liver bacterial loads compared with control mice (Cyri+/flox; Fig. 6c). Similarly, oral infection resulted in decreased survival and increased bacterial loads in Cyri−/flox mice relative to Cyri+/flox mice (Supplementary Fig. 7c,d). Together, these findings indicate that CYRI plays an important host protective role in myeloid cells during Salmonella infection.

To interrogate whether the increased susceptibility to Salmonella observed in Cyri−/flox mice is a consequence of perturbed RAC1-driven cellular effects in myeloid cells, we analysed the phagocytic potential of BMDMs derived from Cyri+/flox and Cyri−/flox mice. Infection of BMDMs with opsonized bacteria revealed a significant increase in the percentage of Salmonella-infected cells and of BMDMs harbouring more than one bacterium per cell in BMDMs derived from Cyri−/flox mice compared with control BMDMs (Fig. 6d,e and Supplementary Fig. 7e). Similarly, higher Salmonella bacterial counts were detected in BMDMs derived from Ity15m/m mice compared with Ity15+/+ mice. Importantly, this increase was dependent on actin cytoskeleton remodelling, as pretreatment of BMDMs with cytochalasin D—an inhibitor of actin polymerization36—completely abrogated bacterial entry into BMDMs derived from both Ity15+/+ and mutant mice (Supplementary Fig. 7f). Intriguingly, CYRI deletion had no effect when BMDMs were exposed to zymosan particles (Supplementary Fig. 7g), which suggests that CYRI-mediated regulation of phagocytosis depends on the nature of the phagocytized particles. This underlines a role for CYRI in governing macrophage phagocytic potential following infection through its effect on the actin cytoskeleton.

In addition to utilizing phagocytes as a replicative niche, Salmonella also exploit phagocyte migration to aid dissemination to secondary tissues, such as the spleen and liver22,23,24. RAC1 is known to guide neutrophil trafficking between the bone marrow and blood as well as cell migration within tissues37. As such, we next investigated the cell motility of neutrophils derived from Cyri+/flox and Cyri−/flox mice in a three-dimensional collagen matrix. Loss of CYRI significantly affected the migratory behaviour of neutrophils (Fig. 6f), with cells displaying higher velocities (Fig. 6g) and travelling greater distances in all three dimensions compared with neutrophils derived from Cyri+/flox mice (Fig. 6h,i). CYRI deletion in neutrophils was also associated with significantly diminished neutrophil turning angles and directional persistence (Supplementary Fig. 7h), suggesting that neutrophils were switching to a more random mode of migration in the absence of CYRI, which is consistent with RAC1 hyperactivation38. These migratory pattern changes in neutrophils lacking CYRI could help explain the enhanced Salmonella dissemination observed in Cyri−/flox mice, implicating CYRI in restricting this process.

CYRI is targeted in a SopE-dependent manner following Salmonella infection

Considering the role of CYRI in restricting Salmonella infection, we postulated that Salmonella might employ a counterdefence mechanism to neutralize the cellular effects of CYRI. Indeed, the levels of CYRI protein were significantly reduced in Salmonella-infected BMDMs, which provides support for this idea. Strikingly, this effect was dependent on SopE, with BMDMs infected with the ∆SopE Salmonella strain displaying CYRI protein levels comparable to uninfected cells (Fig. 7a,b). Consistent with this, heterologous GFP–SopE expression in HEK293T cells also reduced the levels of CYRI protein. Conversely, SopE had no effect on CYFIP1/2 protein levels (Fig. 7c,d), thus identifying a role for SopE in selectively targeting CYRI.

Fig. 7: The Salmonella bacterial effector SopE negatively regulates CYRI.
figure7

a, Western blot analysis of the lysates from uninfected BMDMs and BMDMs infected with WT Salmonella (SL1344) or with a SopE-deficient Salmonella strain (∆SopE). Endogenous CYRI protein levels were detected. β-Tubulin was used as a loading control. Lanes represent technical replicas of infection. Data are representative of six independent experiments. b, Quantification of the levels of endogenous CYRI protein in uninfected, SL1344-infected and ∆SopE-infected BMDMs normalized to levels in the cellular lysates of uninfected cells. The graph represents the individual values and the mean ± s.e.m. from six independent experiments. c, Western blot analysis of the lysates from mock- and GFP–SopE-transfected HEK293T cells. The levels endogenous CYRI, CYFIP1/2 and exogenous GFP–SopE protein were detected. α-Tubulin was used as a loading control. Lanes represent western blot technical replicas. Data are representative of three independent experiments. d, Quantification of endogenous CYRI and CYFIP1/2 protein in mock- and GFP–SopE-transfected HEK293T cells normalized to levels in mock-transfected cellular lysates. Graph represents individual values and the mean ± s.e.m. from three independent experiments. e, PAK-CRIB pulldown of the lysates from mock-, GFP–SopE WT-, G168A- and G168V-transfected HEK293T cells. The levels of active, total RAC1, endogenous CYRI and exogenous GFP–SopE protein were detected by western blotting. α-Tubulin was used as a loading control. Data are representative of three independent experiments. f, Quantification of endogenous CYRI protein in mock-, GFP–SopE WT-, G168A- and G168V-transfected HEK293T cells normalized to the levels in cellular lysates of mock-transfected cells. The graph represents the individual values and the mean ± s.e.m. from three independent experiments. g, Schematic representation of the workflow for SopE-dependent ubiquitinome analysis using SILAC-diGly proteomics. SILAC light- or heavy-labelled HeLa Flp-In T-REx GFP–SopE cells were either left untreated or treated with dox, respectively, and subjected to SILAC-diGly proteomics followed by liquid chromatography-tandem mass spectrometry (LC–MS/MS) analysis. h, Overlap of total and two-fold upregulated or downregulated diGly peptides from three independent SILAC-diGly proteomics experiments (Exp.). i, SILAC ratios for selected diGly peptides. Graph represents the log2 heavy-to-light (H/L) SILAC ratios across three independent SILAC-diGly proteomics experiments. The dashed red line represents the cut-off applied to identify diGly peptides that exhibit increased ubiquitylation following SopE expression. The upregulated CYRI diGly sites are indicated in red. *P ≤ 0.05; **P < 0.01. Statistical analyses are detailed in Supplementary File 2. ns, non-significant.

Given that SopE predominantly functions as a RAC1 GEF8,9, we next examined the dependency of SopE-mediated CYRI regulation on SopE GEF activity. In contrast to GFP–SopE WT, expression of two previously reported SopE GEF-dead mutants, G168A and G168V39, failed to reduce CYRI protein levels concomitant with their inability to activate RAC1 (Fig. 7e,f). This indicates that SopE GEF activity is required for CYRI regulation. Strikingly, the expression of SF–RAC1 P29S in confluent, sub-confluent and sparse cells had no effect on the levels of CYRI protein (Supplementary Fig. 8a–d). Similarly, activation of RAC1 in vitro, through incubation of HEK293T cell lysates with GTPγS, resulted in no significant changes in the levels of CYRI protein despite an increase in active RAC1 levels compared with unloaded and GDP-loaded lysates (Supplementary Fig. 8e,f). Collectively, this demonstrates that although SopE GEF activity is required, RAC1 activation alone is insufficient to modulate CYRI protein levels. Notably, inactivation of SopE-meditated RAC1 activation in vitro, through incubation of lysates from GFP–SopE-transfected HEK293T cells with GDP following nucleotide chelating, did not revert SopE-induced reduction in the protein levels of CYRI (Supplementary Fig. 8e,f), demonstrating that SopE regulation of CYRI occurs in cells before cell lysis. Overall, these findings identify CYRI as a Salmonella-targeted host protein and demonstrate the dependency of Salmonella on SopE to negatively regulate CYRI following infection.

We previously identified an increase in CYRI ubiquitylation on lysine (K) 74 and K78 following proteasomal inhibition in Salmonella-infected cells40. This raised the possibility that SopE might regulate CYRI protein levels by mediating CYRI ubiquitylation and subsequent degradation. To test this, we utilized stable isotope labelling by amino acids in cell culture (SILAC) coupled with diGly proteomics to examine ubiquitinome changes following dox-induced GFP–SopE expression in HeLa cells (Fig. 7g). Three replicate experiments uncovered 5,742 unique ubiquitylation sites in 2,594 proteins (Supplementary Fig. 8g and Supplementary File 1), with replicate experiments displaying Pearson’s correlation coefficients of up to 0.64, indicating high reproducibility (Supplementary Fig. 8h). Among these diGly peptides, we identified 159 sites (143 proteins) displaying a twofold increase and 59 sites (49 proteins) with a twofold decrease in cells expressing GFP–SopE compared with dox-untreated cells in at least one experiment (Supplementary Fig. 8g and Supplementary File 1). Importantly, GFP–SopE expression was associated with increased ubiquitylation on CYRI K74 across all three replicates (Fig. 7h,i and Supplementary Fig. 8i). Further inspection of the SopE-regulated ubiquitinome dataset also revealed additional increased CYRI diGly sites, including K289 and K317, which were upregulated in one and two replicate experiments, respectively (Fig. 7i and Supplementary Fig. 8j,k). This illustrates the ability of SopE to induce global ubiquitinome changes and to stimulate CYRI ubiquitylation, thereby highlighting a mechanism by which SopE can regulate CYRI protein levels following infection.

CYRI regulates host susceptibility to other intracellular bacterial pathogens

Having identified an important protective role of CYRI during Salmonella infection, we were interested in evaluating its effects in the context of other intracellular bacterial pathogens that also rely on RAC1 signalling. To that end, Ity15m/m mice and their littermate controls were infected with M. tuberculosis. Analysis of the lungs of infected mice revealed large parenchymal consolidation in Ity15m/m mice (Supplementary Fig. 9a) and a higher microbial burden compared with Ity15+/+ mice (Supplementary Fig. 9b,c). Consistent with this, Ity15m/m mice exhibited increased parenchymal infiltration by granulomatous reaction and lymphohistiocytic inflammatory cells (Supplementary Fig. 9d). A similar phenotype was also observed following infection of Cyri−/flox mice (Supplementary Fig. 9e–h). Together, these results implicate CYRI in counteracting M. tuberculosis and indicate that CYRI function within the myeloid compartment is central for attenuating M. tuberculosis infection.

In addition to M. tuberculosis, we investigated the role of CYRI in L. monocytogenes infection as an example of another intracellular pathogen that utilizes macrophages as a protective niche41,42. Based on our observations from Salmonella-infected BMDMs we postulated that CYRI might also impede L. monocytogenes uptake. Indeed, infection of BMDMs derived from Ity15+/+ and Ity15m/m mice with opsonized L. monocytogenes revealed an increased bacterial count in BMDMs derived from Ity15m/m mice compared with control BMDMs. In addition, similar to Salmonella, inhibition of actin cytoskeleton remodelling abrogated L. monocytogenes internalization (Supplementary Fig. 9i). These findings suggest that the CYRI-mediated regulation of RAC1-driven cytoskeletal dynamics is also implicated in restricting L. monocytogenes infection.

Discussion

RAC1 signalling plays a critical role in the pathophysiology of intracellular bacterial pathogens, including Salmonella, M. tuberculosis and L. monocytogenes9,18,43,44. Here we show that, unlike the structurally related CYFIP proteins, CYRI negatively regulates RAC1 cellular events. In line with these findings, two recent studies also report similar observations in different cellular contexts45,46. It is therefore possible that CYRI protects against these pathogens by negatively modulating RAC1 signalling. Focusing on Salmonella, we demonstrate that CYRI restricts infection at multiple stages, all of which are dependent on RAC1 signalling (Supplementary Fig. 10a). Of particular note is the function of CYRI within the myeloid compartment. All three pathogens examined have evolved mechanisms that allow the evasion of host defences in macrophages, thus creating a favourable environment for survival and replication34,42,43,44,47. Our findings indicate that CYRI inhibits the phagocytic uptake of bacteria by macrophages. A recent genome-wide CRISPR–Cas9 screen also demonstrated that, unlike RAC1 and CYFIP1, CYRI KO is associated with increased phagocytosis48. Thus, through inhibiting phagocytosis CYRI might hinder the formation of a bacterial protective niche, thereby restricting bacterial infection. In addition, phagocytes are implicated in mediating systemic infections of Salmonella, M. tuberculosis and L. monocytogenes22,47,49. We therefore propose that the ability of CYRI to negatively regulate the cell migration of myeloid-derived phagocytes constitutes another potentially important defence mechanism that limits bacterial dissemination.

This study also describes an important host–pathogen signalling interplay involving CYRI and the Salmonella bacterial effector SopE. Intriguingly, our findings implicate SopE in reducing CYRI protein levels following infection and inducing ubiquitylation on K74, a site which was previously shown to undergo enhanced ubiquitylation following proteasomal inhibition in Salmonella-infected cells40. This suggests that SopE might drive the ubiquitylation and subsequent degradation of CYRI. Indeed, Salmonella has been shown to modulate host signalling cascades through ubiquitylation of host proteins40. However, to date this has been attributed to bacterial effectors that mimic host E3-ubiquitin (Ub) ligases, enzymes that facilitate the conjugation of Ub to substrates50. Here, we show that this can also be achieved via bacterial effectors lacking E3-Ub ligase activity, such as SopE, probably through the exploitation of host E3-Ub ligases (Supplementary Fig. 10b). This highlights an additional strategy employed by Salmonella to subvert host defences. Further characterization of CYRI regulation following infection by other intracellular pathogens and under different physiological and pathological conditions is warranted to ascertain whether the modulation of CYRI protein levels constitutes a general CYRI inhibitory mechanism.

Methods

Antibodies

The antibodies utilized in this study are outlined in Supplementary Table 3.

Mice

All animal experiments were performed under the guidelines specified by the Canadian Council for Animal Care. The animal-use protocol was approved by the McGill University Animal Care Committee (protocol no. 5797). C57BL/6J, 129S1, 129X1, DBA/2, B6.129S4-Meox2tm1(cre)Sor/J (Meox-Cre; JAX stock no. 003755) and B6.129P2-Lyz2tm1(cre)Ifo/J (LysM-Cre; JAX stock no. 004781) mice were purchased from The Jackson Laboratory. C57BL/6N-Fam49btm1a(KOMP)Wtsi/Tcp (Fam49btm1a) mice were obtained from the NorCOMM Project. Flp mice were a gift from Y. Yamanaka (McGill Cancer Research Centre). For mice infections, the sample sizes were determined by the availability of mice according to a batch heterogenization approach. All healthy mice aged between 7 and 12 weeks were included, at random, within experimental groups. Importantly, the mice used were littermates (heterozygous breeders), and were age and sex matched for each experiment. The investigators were blinded to groups during the monitoring of the Salmonella-infected mice. Original Ity15 mice were on a mixed background (129S1 × 129X1) and then transferred to 129X1 by serial backcrossing. For mapping purpose, F2 mice were on a mixed 129S1/129X1/DBA2/J background. Mice with a myeloid-specific deletion of Cyri were on a B6.BcA17-Slc11a1 background.

ENU mutagenesis

Generation 0 (G0) 129S1 males were mutagenized with a single intraperitoneal injection of 150 mg ENU per kilogram of body weight. The G0 males were outcrossed to 129X1 females to create G1 progeny that were subsequently outcrossed to 129X1 females. The resulting G2 females were backcrossed to the parental G1 males to produce the G3 mice that were assessed for susceptibility to Salmonella infection. The G1 male from the Ity15 pedigree was outcrossed to a DBA/2J female to generate F1 offspring. F1 mice were intercrossed to generate F2 offspring, which were then used for mapping. Crossing Ity15+/m with Cyri+/− generated mice that were used for the allelic complementation assay. The Ity15m/m allele was then transferred to a 129X1 background by serial backcrossing (n = 9).

Haematology and flow cytometry analyses

Blood samples were collected in EDTA tubes from mice aged 8–10 weeks for complete blood counts with white blood cell differential. Analyses were performed at the Comparative Medicine Animal Resources Centre at McGill University. The data from the males and females were pooled. For flow cytometric analyses, the thymus and spleen from mice aged 8–12 weeks were collected from naive mice in 3–5 ml PBS under sterile conditions. Tissues were mechanically dissociated using the plunger end of 3 ml syringes into 70-μm cell strainers (Fisherbrand) to generate single-cell suspensions. The cells were treated with ACK lysis buffer, washed with cold PBS and passed through a second 70-μm cell strainer. Cell counts were determined using a Coulter Z2 particle counter with a 3–10 μM size-inclusion setting (Beckman Coulter) or a haemacytometer. Viability was assessed using Zombi Aqua Fixable viability dye (BioLegend). All antibodies were from eBioscience unless otherwise stated. The following antibody cocktail was used for granulocytes, monocytes, dendritic cells and B cells: CD19 PerCP–Cy5.5 (6D5, BioLegend), anti-CD11c eFluor450 (N418), anti-CD11b APC (M1/70), anti-Ly6C PE (HK1.4), anti-Ly6G APC–Cy7 (1A8, BioLegend), F4/80 PE–Cy7 and MHCII FITC. The following antibodies were used for T and natural killer cells in splenocytes: anti-CD4 PE–Cy7, anti-CD8α APC–Cy7, anti-TCRβ PerCP–Cy5.5, anti-CD69 FITC and anti-CD49b (DX5) Pacific Blue. The panel for thymocyte staining included: anti-CD4 PE–Cy7, anti-CD8α APC–Cy7, anti-TCRβ PerCP–Cy5.5, anti-CD44 PE (IM7, BioLegend) and anti-CD25 FITC (3C7, BioLegend). The cells were acquired on an eight-colour FACSCanto II using the FACS Diva software (BD). The data were analysed using FlowJo version 10.0.8r1 software. Doublets were removed by SSC-H versus SSC-W gating. The gating strategy for flow cytometry is detailed in Supplementary Fig. 11.

In vivo Salmonella infection and bioluminescence live imaging

The mice were injected intravenously in the caudal vein with S. Typhimurium (strain Keller; 5,000 c.f.u.). The mice were euthanized on days 3, 4 or 5 post infection to harvest the blood, spleen and liver as previously described51. For whole-body imaging, the mice were injected intravenously with 8.8 × 104 c.f.u. S. Typhimurium strain XEN26 (PerkinElmer, cat. no. 119230), anaesthetized with isoflurane, anteriorly shaved and imaged daily in an IVIS Spectrum system (PerkinElmer, cat. no. 124262) until they reached clinical endpoints, at which time they were euthanized. Bioluminescent images were acquired with an open emission filter, binning factor of 16 and exposure time ranging from 2 to 5 min. The regions of interest were selected, normalized across time points and quantified using Living Image software v.4.3.1 (PerkinElmer).

In vivo per os Salmonella infection

For the per os infections, the mice were first starved for 4 h (no food and water) and then gavaged with 20 mg streptomycin sulphate diluted in 100 μl sterile water. After 24 h following antibiotic administration, the mice were starved for 4 h before oral infection with SL1344 (5 × 107 c.f.u. in 100 μl sterile saline). Spleens and livers were collected either 24 h or four days post infection for c.f.u. enumeration. Haematoxylin and eosin-stained caeca were analysed in duplicate and blinded for genotype. A histopathological scoring system was adapted from a previously described model52. The score evaluated the submucosal oedema (0, 0%; 1, <10%; 2, 10–40% and 3, >40%), polymorphonuclear neutrophil infiltration of the lamina propria (0, <5; 1, 5–20; 2, 21–60; 3, 61–100 and 4, >100 polymorphonuclear neutrophils), the goblet cell count (0, >10; 1, 6–10; 2, 1–5 and 3, 0 per high-power field) and the epithelial integrity (0, no change; 1, desquamation; 2, erosion of the epithelial surface and 3, ulceration) of caecum cross-sections.

Histology

Tissues were collected and fixed in 10% neutral buffered formalin for 24 h at 20 °C, then placed in 70% ethanol at 4 °C before processing and embedding (Goodman Cancer Research Center histology facility, McGill University). The embedded tissues were sectioned and stained with haematoxylin and eosin.

Cytokine analysis

Blood was collected from uninfected and infected Ity15+/+, Ity15m/m, Cyri+/flox and Cyri−/flox mice by cardiac puncture into serum-separating tubes (Sarstedt). The levels of serum cytokines were quantified either using ELISA kits (TNFα, IL-6 and IL-10) purchased from eBiosciences (Ready-SET-Go! kits) or using a mouse cytokine/chemokine 32-multiplex Luminex array (Eve Technologies).

Genetic mapping, genotyping and exome sequencing

DNA was extracted from a tail biopsy by proteinase K digestion and phenol–chloroform extraction. The genome scan was performed using the medium density SNP panel from Illumina (708 single nucleotide polymorphisms (SNPs) were informative between 129S1 and DBA/2 strains; The Centre for Applied Genomics). The mapping was conducted using 24 mice (9 Salmonella susceptible and 15 Salmonella resistant). Exome sequencing was done for two Salmonella-susceptible mice. Briefly, exon capture was performed with the SureSelect mouse all exon kit (Agilent Technologies, cat. no. 5190-4641) and sequencing of 100-bp paired-end reads on an Illumina HiSEquation 2000 generating over 8 Gb of sequence for each sample (Centre National de Génotypage). Read alignment, variant calling and annotation were performed as previously described53. The Ity15 mutation in Cyri was subsequently genotyped using a custom Taqman SNP genotyping assay (Applied Biosystems).

Constructs

The details of the plasmids used in this study are summarized in Supplementary Table 4.

Cell lines

Neutrophils and BMDMs were obtained from Ity15 and/or Cyri+/flox and Cyri−/flox mice and confirmed using flow cytometry (Ly6G+ CD11b+ and CD11b+ F4/80+ respectively). HEK293T cells were provided by A. Malliri (Cell Signalling Group, Cancer Research UK Manchester Institute; purchased from the ECACC). HeLa cells harbouring the Flp-In T-REx system were purchased from the ATCC. The cell lines were not tested for mycoplasma contamination. For the genetic modification of cell lines, transfection with the indicated constructs was performed using GeneJuice (Nalgene, cat. no. 16211-034) according to the manufacturer’s instructions. For overexpression, CYRI–HA and GFP–SopE were sub-cloned into pcDNA5/FRT/TO and the dox-inducible cell lines were generated using HeLa cells harbouring the Flp-In T-REx system. HeLa cells containing the dox-inducible expression system were subjected to antibiotic selection with blasticidin (15 mg ml−1; Capricorn Scientific, BLA-5x) and hygromycin B (250 μg ml−1; Invitrogen, cat. no. 10687010). Retroviral transduction was performed as previously described54 to generate the HeLa Tet-On empty vector, SF–RAC1 WT and SF–RAC1 P29S cell lines. Three sgRNAs, outlined in Supplementary Table 5, were selected from the Broad Institute Genetic Perturbation Platform (https://portals.broadinstitute.org/gpp/public/analysis-tools/sgrna-design) for CRISPR–Cas9-mediated KO. The annealed oligos were cloned into the lentiCRISPR V2 plasmid and used to generate virus. Briefly, 1 × 106 HEK293T cells were plated per well in a six-well plate. The following day, the cells were transfected by mixing 1.1 μg of each sgRNA with 2.7 μg psPAX2, 1 μg pMD2.G and 21 μl TurboFect (Thermo Fisher Scientific, R0531) in 200 µl Opti-MEM (Thermo Fisher Scientific, cat. no. 31985-062) and incubating for 30 min at room temperature (RT) before adding to the cells. After 24 h, the supernatant was collected and stored at −80 °C; a further 2 ml of fresh media was added to each well and incubated for another 24 h. In the meantime, the recipient cells were plated at 1 × 105 cells well−1 in a six-well plate. The next day, the supernatant was collected from the transfected cells, mixed with virus collected the day before, centrifuged for 5 min at 1,200 r.p.m. and filter sterilized. To infect recipient cells, 500 μl of virus was mixed with 1.5 ml DMEM medium and polybrene (8 µg ml−1; Sigma-Aldrich, H9268-5G). To generate the double CYRI/FAM49A KO cells, 250 µl of each virus pool was mixed and used for cell transduction. As a control, parental cells were infected with virus generated from an empty LentiCRISPR V2 plasmid (control KO). The cells were re-plated 48 h post infection in selection DMEM medium with 2 μg ml−1 puromycin (Roth, cat. no. 0240.3). Antibiotic selection was withdrawn during experiments.

Cell culture

All cell lines were cultured in DMEM medium (Thermo Fisher Scientific, cat. no. 11960-044) supplemented with 10% (v/v) fetal bovine serum (Thermo Fisher Scientific, cat. no. 10270106) and 10 µg ml−1 penicillin-streptomycin (Sigma-Aldrich, P0781-100 ml) at 37 °C and 5% CO2.

CYRI in silico analysis

Uniprot was used to obtain human protein sequences for FAM49B (Q9NUQ9), FAM49A (Q9H0Q0), CYFIP1 (Q7L576) and CYFIP2 (Q96F07). The schematic representation of the domains present in these proteins was based on the Pfam database (http://pfam.xfam.org). HHpred software27 was used for CYRI remote protein homology detection and structural prediction. This identified homology to the CYFIP1 crystal structure (PDB-3P8C), which was then used to generate a three-dimensional model for CYRI using the MODELLER software27. The models presented in the manuscript were generated in MacPymol based on HHpred model predictions. The presence of CYRI, FAM49A, CYFIP1 and CYFIP2 in other species was determined using a Uniprot BLAST search (http://www.uniprot.org/blast). Important model organisms containing the respective proteins were used for sequence alignment, focusing on a region surrounding known RAC1 binding residues within the CYFIP1 DUF1394 domain. Sequence alignment and evolutionary conservation across the indicated species was conducted using Jalview.

Western blot analysis

Samples were mixed with the appropriate volume of 2×Laemmli sample buffer (BioRad, cat. no. 1610737) and resolved on 4–20% (BioRad, cat. nos 4561094, 4561095 and 4561096) or 12% Mini-protein TGX precast protein gels (BioRad, cat. nos 4561043, 451045 and 4561046) or 12% self-cast gels. Prestained protein ladder (BioFroxx, cat. no. 1123YL500) was run alongside the samples for protein size reference. Proteins were transferred onto Immobolin PVDF membranes (Millipore, IPFL00010). The membranes were stained with Ponceau S (Sigma-Aldrich, P7170) to determine transfer efficiency. Ponceau staining was also used as a loading control. Western blotting was performed using the antibodies listed in Supplementary Table 3, as outlined therein, and visualized on a Fuji medical X-ray super RX (cat. no. 4741008389) using the western blotting luminol reagent (SantaCruz Biotechnology, sc-2048 and sc-2049). Cells and tissue homogenate derived from mice were extracted using T-PER tissue extraction reagent (Thermo Scientific, cat. no. 78510) with a protease inhibitor cocktail (Sigma-Aldrich, P8465) as per the manufacturer’s protocol. The protein concentration in lysates was quantified using a Bradford assay (BioRad, cat. no. 500-0006) and the appropriate volume of lysates with equal protein concentrations were resuspended in Laemmli buffer before western blotting. For the generation of figures, western blot films were scanned and processed using Adobe Photoshop CS5 to generate the presented crops. Uncropped western blots are provided in Supplementary Figs. 12 and 13. Western blot scans were also processed by ImageJ, in which the band intensities were quantified using the ImageJ Gel analysis tool and the presented quantifications were generated as outlined in the respective Methods sections.

PAK-CRIB pulldown

Active RAC1 levels were assessed using PAK-CRIB pulldown, an established biochemical assay that utilizes the CDC42- and RAC1-interactive binding (CRIB) domain of PAK, a CDC42 and RAC1 effector, to enrich the active form of RAC1 and CDC42 (ref. 55). Briefly, cells were lysed in GST lysis buffer (25 mM Tris–HCl pH 7.2, 150 mM NaCl, 5 mM MgCl2, 1% (v/v) Nonidet P40, 5% (v/v) glycerol, 1% protease inhibitor cocktail tablet (Roche, cat. no. 4693132001), 1% (v/v) phosphatase inhibitor cocktails 1 and 2 (Sigma-Aldrich, P5726-5 ml and P0044-5 ml) in dH2O) containing 8 μg ml−1 biotinylated PAK-CRIB purified peptide and incubated for 30 min at 4 °C. Cleared lysates were then incubated with 50 μl Strep-Tactin superflow resin (IBA GmbH, cat. no. 2-1206-10) for 15 min at 4 °C. To assess the RAC1 activity of SF-RAC1 WT and SF-RAC1 P29S, lysates were incubated with GST-tagged PAK-CRIB and Glutathione Sepharose 4B beads for 1 h at 4 °C. Beads were washed following incubation and prepared for western blotting to quantify the levels of active and total RAC1. The biotinylated PAK-CRIB was a gift from A. Malliri. For the quantification of active RAC1 levels, RAC1 PAK-CRIB pulldown bands as well as RAC1 and α-tubulin input bands were analysed using the ImageJ Gel analysis tool. The RAC1 input band intensities were normalized to the α-tubulin-band intensities in the respective sample lanes. The intensities of the RAC1 PAK-CRIB pulldown bands were then normalized to the respective adjusted RAC1 input-band intensities. These values were then normalized to the control samples as specified in the presented graphs and figure legends.

GST pulldown

Rosetta competent cells were used to grow and purify the respective GST-tagged proteins. GST or GST–RAC1 immobilized on Glutathione Sepharose 4B (GE Healthcare, cat. no. 17-0756-01) were washed using GST lysis buffer and used immediately (unloaded) or subjected to nucleotide chelating using 0.5 M EDTA (10 mM final concentration). GST and GST–RAC1 beads were then incubated with guanosine 5ʹ-diphosphate sodium salt (GDP; 100 μM final concentration; Sigma-Aldrich, G7127) or guanosine 5ʹ-O-(3-thiotriphosphate) tetralithium salt (GTPγS; 1 mM final concentration; Sigma-Aldrich, G8634) for 15 min at 30–37 °C. Termination of the reaction was achieved by adding MgCl2 (60 mM final concentration). Cells were lysed 24 h post dox treatment (1 μg ml−1) or transfection in GST lysis buffer and equal protein amounts were incubated with unloaded, GDP-loaded or GTPγS-loaded GST and GST–RAC1 beads for 2–72 h at 4 °C. Both GST and GST–RAC1 were a gift from A. Malliri. For the GST and GST–CYRI pulldown experiments, HEK293T cells were lysed and either left untreated, or mixed with GDP or GTPγS as detailed above before overnight incubation with GST and GST–CYRI beads at 4 °C. The beads were washed following incubation and prepared for western blotting. Co-precipitated protein bands visualized by western blotting and GST-tagged protein bands visualized by Ponceau staining were analysed using the ImageJ Gel analysis tool for the quantification of protein binding. The band intensities of the co-precipitated proteins were then normalized to the intensities of the GST-tagged proteins in the respective GST-pulldown lanes. These values were then normalized to the control samples as specified in the presented graphs and figure legends.

Fluorescence microscopy

Cells grown on 96-well plates (Greiner Bio, cat. no. 655090) were fixed in 4% paraformaldehyde solution in PBS (ChemCruz, SC281692) for 15–20 min at RT and permeabilized by incubating for 3 min with 0.5% (v/v) Triton-X in PBS (Thermo Fisher Scientific, cat. no. 14190-169) at RT. The cells were next blocked in 1% BSA (Roth, T844.3) in PBS for 30 min before incubation with the relevant primary antibodies (in 1% BSA) for 1 h at RT. The respective fluorescent secondary antibodies were then added for an additional hour at RT in the dark. Fluorescence staining was performed using the antibodies/reagents listed in Supplementary Table 3, as outlined therein. Images were captured using the Yokogawa CQ1 confocal quantitative image cytometer platform (×40 magnification). To determine the surface area and circularity index of cells, the images were analysed with the built-in CQ1 image analysis software using the phalloidin and Hoechst staining to mark the cell body of individual cells. All imaging fields were analysed and fields with poor/abnormal staining and/or incorrect analysis due to cell confluency, cells exclusively on image edge or abnormal cell shape/blebbing due to fixation were excluded from the presented graphs. For cells expressing GFP–SopE and/or the different CYRI–HA variants, GFP and HA staining were used to determine the expression levels, respectively. For separation of cells based on GFP–SopE expression levels, the GFP staining of dox-untreated cells was used to generate three intensity thresholds that were roughly delineated as follows: high-expressing cells represent cells that display intensities higher than the majority of dox-untreated cells; medium-expressing cells represent cells with intensities between the average intensity of dox-treated cells and the threshold used for selecting high expressors; low-expressing cells represent cells that display intensities between the average intensity detected in dox-untreated cells and the average intensity detected in dox-treated cells. To select cells expressing similar levels of GFP–SopE and/or exogenous CYRI–HA, the GFP and/or the HA staining of the respective control cells was used to set an intensity threshold that was then applied to the dox-treated cells to exclude cells that lacked expression of the respective proteins as well as cells that had abnormal expression levels, which would therefore result in expression levels that were significantly different between dox-treated samples. Following cell population determination based on set thresholds, statistical analysis was used to confirm that the samples displayed no significant differences in protein expression of GFP–SopE and/or exogenous CYRI–HA.

Duolink in situ PLA

The Duolink in situ proximity ligation assay (PLA) was conducted as previously described54,56. Briefly, HeLa Flp-In T-REx CYRI–HA cells were plated in 96-well plates in the absence or presence of dox 24 h before fixation and permeabilization as indicated above. For the analysis of exogenous CYRI–endogenous RAC1 binding, cells were incubated with mouse anti-RAC1 antibody (BD Biosciences, cat. no. 610650) and rabbit anti-HA (Biolegend, cat. no. 902301) for 1 h at RT. The respective PLA probes (Olink Bioscience; anti-mouse, cat. no. 92004-0100 and anti-rabbit, cat. no. 92002-0100) and the detection reagent kit (Olink Bioscience, cat. no. 92014-0100) were then applied according to the manufacturer’s instructions. Alexa Fluor 568 phalloidin (Thermo Fisher Scientific, A12380) and NucBlue Live ReadyProbes reagent (Thermo Fisher Scientific, R37605) were used as fluorescence markers for the actin cytoskeleton and nuclei, respectively. Rat anti-HA (Roche, cat. no. 118674) was also used in conjunction with the respective fluorescent secondary antibody to assess CYRI–HA expression. Images were taken using the Yokogawa CQ1 confocal quantitative image cytometer platform (×40 magnification). The images were analysed using the built-in CQ1 image analysis software, which was set up to count the number of Duolink dots per cell using phalloidin and Hoechst to define the cell body of individual cells, for the quantification of the Duolink count per cell. Duolink dots with intensities below 1,000 were excluded from both the untreated and dox-treated samples. To specifically determine the Duolink count in CYRI–HA expressing cells, HA staining was used to set an intensity threshold based on the levels detected in dox-untreated cells and cells with HA intensities higher than the majority of dox-untreated cells were considered. The Duolink count per cell from dox-untreated cells was plotted together with the Duolink count per cell identified for the dox-treated CYRI–HA expressing cells to account for background.

Analysis of cell morphology

Cells were seeded in six-well plates for 24 h at varying cellular densities and phase-contrast images were taken on a Zeiss Axiovert 3 (×32 magnification). Cells were classified as normal or pancake-like according to their morphology. By examining parental cells and determining the morphology displayed by the majority of HeLa cells under resting conditions, we classified normal morphology as cells with an elongated mesenchymal-like phenotype. In contrast, as previously described28, cells displaying a flattened morphology with unpolarized membrane ruffling were classified as pancake-like.

Biochemical analysis of actin polymerization

G-actin and F-actin fractions were extracted from cell lysates by ultracentrifugation at 100,000 r.p.m. using the G-actin/F-actin in vivo assay biochem kit (Cytoskeleton, Inc., BK037) according to the manufacturer’s instructions. The levels of G-actin and F-actin were quantified by western blotting using the anti-actin antibody provided in the kit. As an assay control, G- and F-actin protein levels were determined in lysates treated with an F-actin-enhancing solution containing 100×phalloidin according to the manufacturer’s instructions. As a negative control, HeLa cells were pretreated with 5 µM latrunculin B for 30 min before cell lysis. The levels of G-actin were determined to assess the effect of CYRI deletion on the actin cytoskeleton dynamics. For the quantification of G-actin protein levels, the G-actin bands as well as α-tubulin input bands were analysed using the ImageJ Gel analysis tool. The intensities of the G-actin bands were normalized to the α-tubulin band intensities in the respective samples. These values were then normalized to the control KO sample as specified in the presented graph and figure legend.

In vitro Salmonella infection

Parental, control KO and CYRI KO HeLa Flp-In T-REx cells were plated on coverslips in six-well plates for 24 h, after which the cells were infected with S. Typhimurium (strain SL1344) at a multiplicity of infection (m.o.i.) of 8. After the addition of bacteria, the cells were incubated for 30 min at 37 °C and 5% CO2 before the addition of 100 μg ml−1 gentamicin to kill the extracellular bacteria. The cells were washed twice with PBS and prepared for immunofluorescence as outlined above to assess bacterial entry. To visualize intracellular bacteria, the BacTrace anti-Salmonella CSA-1 antibody (Seracare, 01-91-99) was utilized. Images were captured using confocal microscopy and analysed in Cell profiler 2.2.0 (rev 9969f42) to calculate the number of bacteria per cell.

Isolation of bone marrow-derived cells

The mouse femurs and tibias were collected aseptically. After removing most of the muscle and fat, the epiphyses were cut and the bones were placed into modified PCR tubes individually hung by the hinge into a 1.5 ml Eppendorf containing 300 μl sterile PBS. The bone marrow was flushed by short centrifugation at 5,000 r.p.m. for 10 s or using a 25 G needle. The red blood cells were lysed with a 4 min incubation at RT with RBC lysis buffer (Sigma-Aldrich, cat. no. 11814389001). Cells were pelleted and resuspended in 1 ml sterile PBS or cRPMI (10% fetal bovine serum, 1% penicillin-streptomycin) media. To generate BMDMs, the cells were plated in 10-cm petri dishes supplemented with 30% M-CSF every two days for a total of six days at 37 °C and 5% CO2. Bone marrow-derived neutrophils and BMDMs were confirmed using flow cytometry (Ly6G+ CD11b+ and CD11b+ F4/80+, respectively).

Mixed-infection-based virulence assay

BMDMs from Ity15+/+ and Ity15m/m mice were co-infected with a mix inoculum containing an equal number of non-opsonized SL1344 and a SopE-deficient strain (ΔSopE) generated using the Lambda RED recombination technique57 (total m.o.i. of 10). The ΔSopE Salmonella strain was enumerated from the colony counts on streptomycin/kanamycin plates (c.f.u.), whereas the number of SL1344 bacteria were obtained by subtracting the c.f.u. of the ΔSopE Salmonella strain from the total number of bacterial colonies counted on streptomycin plates.

Bone-marrow chimaeras

Five- to seven-week-old mice of both sexes were irradiated with two consecutive doses of 450 rads using a RS2000 X-ray machine. After irradiation, the mice were reconstituted by intravenous injection of 1.5 × 106 bone marrow cells in sterile PBS. The mice were kept on sterile tap water containing 2 g l−1 neomycin sulphate (Bioshop) for three weeks. Six weeks following irradiation, the mice were tested for chimaerism by collecting blood from the saphenous vein and evaluating the expression of the Cyri WT or mutant allele through isolating DNA and genotyping with TaqMan.

Generation of Cyri conditional KO in myeloid cells

Cyri KO-first allele (Fam49btm1a) on a C57BL/6N background (The Centre for Phenogenomics) was converted to a conditional allele (Fam49btm1c) through Flp recombination58. The C57BL/6N background carries a susceptibility allele at Slc11a1, which is permissive for exponential growth of Salmonella in cells and tissues. We therefore introduced the Slc11a1 WT allele (Slc11a1G169) in the early steps of breeding. Mice carrying the Fam49btm1c allele were crossed to the RCS strain BcA17 (87% C57BL/6J background and 13% A/J59) to introduce the Slc11a1 WT allele (Slc11a1G169). The resulting mice were serially backcrossed to mice carrying the Fam49btm1c allele to eliminate the A/J background. In parallel, Fam49btm1c/tm1c mice were crossed with Meox-Cre mice, leading to the elimination of the floxed exon 6 and the generation of the Fam49btm1d allele. From the resultant offspring, we selected Fam49btm1c/tm1d mice to be crossed with LysM-Cre, yielding Fam49b+/tm1d mice with a Cre recombinase under the control of the LysM myeloid-specific promoter (Tg(LysM-Cre)Fam49b+/tm1d). Finally, the Tg(LysM-Cre)Fam49b+/tm1d mice were crossed with B6.Fam49btm1c/tm1c-Slc11a1G169/G169 mice, yielding the experimental myeloid-specific conditional CYRI KO mice (Cyri−/flox) and control mice (Cyri+/flox). All mice carry one copy of the WT allele at Slc11a1. Despite the transfer of a Slc11a1-resistant allele, the C57BL/6N background is more permissive to Salmonella growth compared with the 129X1 background of the Ity15 mice.

Salmonella and Listeria phagocytosis and replication assays

Ex vivo phagocytosis and replication in BMDMs were assessed using an adapted gentamicin protection assay protocol60. The BMDMs were isolated and cultured as described above. Before use, BMDMs were seeded on 12-mm-diameter glass coverslips in a 24-well plate at 2–2.5 × 105 cells well−1 for 16–24 h. The cells were infected with opsonized S. Typhimurium at an m.o.i. of 10–20 or L. monocytogenes at an m.o.i. of 5. The plates were then centrifuged at 1,400 r.p.m. for 1 min and incubated for 20 min at 37 °C and 5% CO2 before the addition of 100 μg ml−1 gentamicin to kill extracellular bacteria. After 30 min, the cells were washed twice with warm PBS and fixed with 2.5% paraformaldehyde for 10–30 min at 37 °C and 5% CO2 to asses the phagocytosis of bacterial pathogens. Fixed cells were rinsed three times with PBS and permeabilized with 0.2% saponin in 10% normal goat serum overnight at 4 °C. To detect internalized Salmonella in BMDMs derived from Cyri+/flox and Cyri−/flox mice, an unconjugated rabbit anti-Salmonella antibody (Meridian Life Science, Inc., B65701R) was used. Images were obtained on a Zeiss LSM 710 confocal microscope and analysed using Zen software and ImageJ. Detection of S. Typhimurium and L. monocytogenes in BMDMs derived from Ity15+/+ and Ity15m/m mice was achieved using c.f.u. counts.

Zymosan phagocytosis assay

To assess the phagocytic potential of BMDMs derived from Ity15+/+ and Ity15m/m mice, the pH-sensitive pHrodo Red Zymosan A BioParticles conjugate (Life Technologies) was used according to the manufacturer’s instructions. Briefly, BMDMs were plated in a 96-well tissue-culture-treated, black, clear flat-bottom plate (Corning, CLS3904), incubated in Opti-MEM culture medium and left to adhere for at least 2 h at 37 °C and 5% CO2. The BMDMs were then either left untreated or treated with cytochalasin D (Sigma-Aldrich, C2618) at 10 μM for 30 min at 37 °C and 5% CO2. The BMDMs were incubated with 100 μl of zymosan bioparticles resuspended at 0.5 mg ml−1 in Live Cell imaging solution (Thermo Fisher Scientific, A14291DJ) for 1 h at 37 °C and 5% CO2 and fluorescence emission was measured using a microplate reader (PerkinElmer).

Neutrophil enrichment, labelling and three-dimensional migration assay

Total bone marrow neutrophils (~80–85%; Ly6G+CD11b+) were purified by negative selection using the EasySep mouse neutrophil enrichment kit (StemCell Technologies, cat. no. 19762) and dye labelled with 10 μM CellTracker Orange CMTMR (Invitrogen, C2927) in cRPMI medium for 15 min at 37 °C. The cells were suspended in PureCol (Advanced Biomatrix, cat. no. 5005) and cast in custom-built migration chambers at a final collagen concentration of 1.6 mg ml−1 for 40 min, as described previously61. The final cell concentrations in the assay were 1–2 × 106 cells ml−1 gel. The cells were imaged using a Zeiss 880 upright confocal fluorescence microscope for 1 h at 37 °C. The cells were tracked using the spots tool in Imaris (Bitplane), and the speed and directionality parameters were calculated using a custom-coded MATLAB program (MathWorks).

Biochemical analysis of CYRI and CYFIP1/2 protein levels

For the western blot analysis of CYRI protein levels following infection, BMDMs were infected at an m.o.i. of 20 with S. Typhimurium SL1344 or the ΔSopE strain for 45 min at 37 °C. A control group of uninfected macrophages was treated the same way but with medium lacking bacteria. After infection, the BMDMs were washed and cultured for an hour in the presence of 100 μg ml−1 gentamycin to kill all extracellular bacteria present in the medium. The cells were then washed and lysed in 2×Laemmli sample buffer. To determine the effect of SopE on CYRI and CYFIP1/2 protein levels, HEK293T cells were mock transfected or transfected with GFP–SopE 24 h post plating. For the analysis of CYRI protein levels following expression of SopE GEF-dead mutants, cells were transfected with GFP–SopE WT (10 μg per 1 × 107 cells), GFP–SopE G168A or GFP–SopE G168V (20 μg per 1 × 107 cells) 24 h post plating. To evaluate the effect of RAC1 activation, independent of SopE, on CYRI protein levels HeLa Tet-On empty vector, SF–RAC1 WT or SF–RAC1 P29S were plated at varying cellular densities as specified in the figure legend in the presence or absence of dox (1 μg ml−1). The cells were lysed 24 h post transfection or dox treatment in GST lysis buffer and cleared lysates were prepared for western blotting. For in vitro RAC1 activation, mock- or GFP–SopE (10 μg per 1 × 107 cells)-transfected HEK293T cells were lysed and either left untreated, or mixed with GDP or GTPγS, as detailed earlier, before western blot analysis. For the quantification of protein levels, CYRI (and CYFIP1/2) bands as well as α-tubulin bands were analysed using the ImageJ Gel analysis tool. The intensisites of the CYRI (and CYFIP1/2) bands were normalized to the α-tubulin band intensities in the respective samples. These values were then normalized to the control samples as specified in the presented graphs and figure legends.

SILAC-diGly mass spectrometry analysis

SILAC-diGly proteomics experiments were conducted as previously described40 to identify ubiquitylation-site-containing peptides in a quantitative manner. Briefly, HeLa Flp-In T-REx GFP–SopE cells were differentially light- or heavy-SILAC labelled and were either left untreated or dox-treated, respectively, to induce GFP–SopE expression. Following treatment and denaturing cell lysis, equal amounts of light- and heavy-labelled proteins were combined. The proteins were subjected to in-solution digestion. Generated diGly-remnant-containing peptides were enriched by immunoprecipitation and further fractioned by strong cation exchange chromatography before peptide identification by tandem liquid chromatography coupled mass spectrometry analysis.

In vivo M. tuberculosis infection

All experiments involving M. tuberculosis infection were conducted in the Containment Level 3 Platform located at the Research Institute of the McGill University Health Centre. M. tuberculosis H37Rv was grown at 37 °C in Middlebrook 7H9 medium (Difco Laboratories) containing 0.05% Tween 20 and 10% albumin dextrase catalase supplement (Becton Dickson and Co.). Bacteria were delivered by aerosol using an inhalation exposure system (In-Tox Products) and the infectious dose was confirmed by enumeration of bacteria within the lungs of control mice 24 h post infection. The mice were euthanized six weeks post infection. The lungs were homogenized in PBS and the bacterial burden was determined by serial dilution on Middlebrook 7H10 agar (Difco Laboratories) plates supplemented with OADC enrichment (Becton Dickson and Co.) and BacTac Panta Plus (Becton Dickson and Co.). Slides stained with haematoxylin and eosin, and Ziehl–Neelsen were assessed by a pathologist and scored for the degree of inflammation for histology. The score evaluated the cellular infiltration (0, no infiltrate; 1, predominantly lymphocytic; 2, lymphohistiocytic consolidation; 3, early granulomatous reaction; 4, well-formed granulomas and 5, necrotizing granuloma), the percentage of area affected (0, no consolidation; 1, 1–25%; 2, 26–50%; 3, 51–75% and 4, >75%) and mycobacteria presence (0, absent; 1, 1–2; 2, 3–5; 3, 5–10 and 4, >10 per high-power field).

Statistical analysis

Graphs were generated on GraphPad Prism 5 and 8. Statistical analyses were performed on GraphPad Prism 8. The datasets were tested for normality using the Anderson–Darling, D’Agostino and Pearson, Shapiro–Wilk or Kolmogorov–Smirnov tests, depending on the sample size, and the appropriate parametric or non-parametric statistical tests were used accordingly. The statistical tests employed, respective test statistics and exact P values are specified in Supplementary File 2. For the graphs presented in the figures, significance was denoted as follows: ns, non-significant (P > 0.05); *P ≤ 0.05; **P < 0.01; ***P < 0.001 and ****P < 0.0001.

Reporting Summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The data that support the findings of this study are available from the corresponding authors on reasonable request.

Code availability

All commercial and custom codes used to generate data presented in this study are available from the corresponding authors on reasonable request.

References

  1. 1.

    Kirk, M. D. et al. World Health Organization estimates of the global and regional disease burden of 22 foodborne bacterial, protozoal, and viral diseases, 2010: a data synthesis. PLoS Med. 12, e1001921 (2015).

  2. 2.

    Majowicz, S. E. et al. The global burden of nontyphoidal Salmonella gastroenteritis. Clin. Infect. Dis. 50, 882–889 (2010).

  3. 3.

    Prestinaci, F., Pezzotti, P. & Pantosti, A. Antimicrobial resistance: a global multifaceted phenomenon. Pathog. Glob. Health 109, 309–318 (2015).

  4. 4.

    LaRock, D. L., Chaudhary, A. & Miller, S. I. Salmonellae interactions with host processes. Nat. Rev. Microbiol. 13, 191–205 (2015).

  5. 5.

    Schlumberger, M. C. & Hardt, W. D. Salmonella type III secretion effectors: pulling the host cell’s strings. Curr. Opin. Microbiol. 9, 46–54 (2006).

  6. 6.

    Byndloss, M. X., Rivera-Chavez, F., Tsolis, R. M. & Baumler, A. J. How bacterial pathogens use type III and type IV secretion systems to facilitate their transmission. Curr. Opin. Microbiol. 35, 1–7 (2017).

  7. 7.

    Guiney, D. G. & Lesnick, M. Targeting of the actin cytoskeleton during infection by Salmonella strains. Clin. Immunol. 114, 248–255 (2005).

  8. 8.

    Hardt, W. D., Chen, L. M., Schuebel, K. E., Bustelo, X. R. & Galan, J. E. S. typhimurium encodes an activator of Rho GTPases that induces membrane ruffling and nuclear responses in host cells. Cell 93, 815–826 (1998).

  9. 9.

    Orchard, R. C. & Alto, N. M. Mimicking GEFs: a common theme for bacterial pathogens. Cell. Microbiol. 14, 10–18 (2012).

  10. 10.

    Humphreys, D., Davidson, A., Hume, P. J. & Koronakis, V. Salmonella virulence effector SopE and host GEF ARNO cooperate to recruit and activate WAVE to trigger bacterial invasion. Cell Host Microbe 11, 129–139 (2012).

  11. 11.

    Gautreau, A. et al. Purification and architecture of the ubiquitous Wave complex. Proc. Natl Acad. Sci. USA 101, 4379–4383 (2004).

  12. 12.

    Eden, S., Rohatgi, R., Podtelejnikov, A. V., Mann, M. & Kirschner, M. W. Mechanism of regulation of WAVE1-induced actin nucleation by Rac1 and Nck. Nature 418, 790–793 (2002).

  13. 13.

    Ismail, A. M., Padrick, S. B., Chen, B., Umetani, J. & Rosen, M. K. The WAVE regulatory complex is inhibited. Nat. Struct. Mol. Biol. 16, 561–563 (2009).

  14. 14.

    Chen, Z. et al. Structure and control of the actin regulatory WAVE complex. Nature 468, 533–538 (2010).

  15. 15.

    Chen, B. et al. Rac1 GTPase activates the WAVE regulatory complex through two distinct binding sites. eLife 6, e29795 (2017).

  16. 16.

    Francis, C. L., Ryan, T. A., Jones, B. D., Smith, S. J. & Falkow, S. Ruffles induced by Salmonella and other stimuli direct macropinocytosis of bacteria. Nature 364, 639–642 (1993).

  17. 17.

    Steele-Mortimer, O. The Salmonella-containing vacuole: moving with the times. Curr. Opin. Microbiol. 11, 38–45 (2008).

  18. 18.

    Ham, H., Sreelatha, A. & Orth, K. Manipulation of host membranes by bacterial effectors. Nat. Rev. Microbiol. 9, 635–646 (2011).

  19. 19.

    Schleker, S. et al. The current Salmonella-host interactome. Proteom. Clin. Appl. 6, 117–133 (2012).

  20. 20.

    Mostowy, S. & Shenoy, A. R. The cytoskeleton in cell-autonomous immunity: structural determinants of host defence. Nat. Rev. Immunol. 15, 559–573 (2015).

  21. 21.

    Behnsen, J., Perez-Lopez, A., Nuccio, S. P. & Raffatellu, M. Exploiting host immunity: the Salmonella paradigm. Trends Immunol. 36, 112–120 (2015).

  22. 22.

    Worley, M. J., Nieman, G. S., Geddes, K. & Heffron, F. Salmonella typhimurium disseminates within its host by manipulating the motility of infected cells. Proc. Natl Acad. Sci. USA 103, 17915–17920 (2006).

  23. 23.

    Ohl, M. E. & Miller, S. I. Salmonella: a model for bacterial pathogenesis. Annu. Rev. Med. 52, 259–274 (2001).

  24. 24.

    Fabrega, A. & Vila, J. Salmonella enterica serovar Typhimurium skills to succeed in the host: virulence and regulation. Clin. Microbiol. Rev. 26, 308–341 (2013).

  25. 25.

    Vazquez-Torres, A. et al. Extraintestinal dissemination of Salmonella by CD18-expressing phagocytes. Nature 401, 804–808 (1999).

  26. 26.

    Hendriksen, R. S. et al. Global monitoring of Salmonella serovar distribution from the world health organization global foodborne infections network country data bank: results of quality assured laboratories from 2001 to 2007. Foodborne Pathog. Dis. 8, 887–900 (2011).

  27. 27.

    Soding, J., Biegert, A. & Lupas, A. N. The HHpred interactive server for protein homology detection and structure prediction. Nucleic Acids Res. 33, W244–W248 (2005).

  28. 28.

    Michiels, F., Habets, G. G., Stam, J. C., van der Kammen, R. A. & Collard, J. G. A role for Rac in Tiam1-induced membrane ruffling and invasion. Nature 375, 338–340 (1995).

  29. 29.

    Krauthammer, M. et al. Exome sequencing identifies recurrent somatic RAC1 mutations in melanoma. Nat. Genet. 44, 1006–1014 (2012).

  30. 30.

    Pollard, T. D. & Cooper, J. A. Actin, a central player in cell shape and movement. Science 326, 1208–1212 (2009).

  31. 31.

    Spector, I., Shochet, N. R., Kashman, Y. & Groweiss, A. Latrunculins: novel marine toxins that disrupt microfilament organization in cultured cells. Science 219, 493–495 (1983).

  32. 32.

    May, R. C. & Machesky, L. M. Phagocytosis and the actin cytoskeleton. J. Cell Sci. 114, 1061–1077 (2001).

  33. 33.

    Krause, M. & Gautreau, A. Steering cell migration: lamellipodium dynamics and the regulation of directional persistence. Nat. Rev. Mol. Cell Biol. 15, 577–590 (2014).

  34. 34.

    Wijburg, O. L., Simmons, C. P., van Rooijen, N. & Strugnell, R. A. Dual role for macrophages in vivo in pathogenesis and control of murine Salmonella enterica var. Typhimurium infections. Eur. J. Immunol. 30, 944–953 (2000).

  35. 35.

    Richter-Dahlfors, A., Buchan, A. M. & Finlay, B. B. Murine salmonellosis studied by confocal microscopy: Salmonella typhimurium resides intracellularly inside macrophages and exerts a cytotoxic effect on phagocytes in vivo. J. Exp. Med. 186, 569–580 (1997).

  36. 36.

    Lin, D. C. & Lin, S. Actin polymerization induced by a motility-related high-affinity cytochalasin binding complex from human erythrocyte membrane. Proc. Natl Acad. Sci. USA 76, 2345–2349 (1979).

  37. 37.

    Campa, C. C. et al. Rac signal adaptation controls neutrophil mobilization from the bone marrow. Sci. Signal. 9, ra124 (2016).

  38. 38.

    Pankov, R. et al. A Rac switch regulates random versus directionally persistent cell migration. J. Cell Biol. 170, 793–802 (2005).

  39. 39.

    Schlumberger, M. C. et al. Amino acids of the bacterial toxin SopE involved in G nucleotide exchange on Cdc42. J. Biol. Chem. 278, 27149–27159 (2003).

  40. 40.

    Fiskin, E., Bionda, T., Dikic, I. & Behrends, C. Global analysis of host and bacterial ubiquitinome in response to Salmonella Typhimurium infection. Mol. Cell 62, 967–981 (2016).

  41. 41.

    de Chastellier, C. & Berche, P. Fate of Listeria monocytogenes in murine macrophages: evidence for simultaneous killing and survival of intracellular bacteria. Infect. Immun. 62, 543–553 (1994).

  42. 42.

    Cossart, P. & Lebreton, A. A trip in the “New Microbiology” with the bacterial pathogen Listeria monocytogenes. FEBS Lett. 588, 2437–2445 (2014).

  43. 43.

    Rajaram, M. V. S. et al. M. tuberculosis-initiated human mannose receptor signaling regulates macrophage recognition and vesicle trafficking by FcRγ-Chain, Grb2, and SHP-1. Cell Rep. 21, 126–140 (2017).

  44. 44.

    Ireton, K., Rigano, L. A. & Dowd, G. C. Role of host GTPases in infection by Listeria monocytogenes. Cell Microbiol. 16, 1311–1320 (2014).

  45. 45.

    Shang, W. et al. Genome-wide CRISPR screen identifies FAM49B as a key regulator of actin dynamics and T cell activation. Proc. Natl Acad. Sci. USA 115, E4051–E4060 (2018).

  46. 46.

    Fort, L. et al. Fam49/CYRI interacts with Rac1 and locally suppresses protrusions. Nat. Cell Biol. 20, 1159–1171 (2018).

  47. 47.

    Queval, C. J., Brosch, R. & Simeone, R. The macrophage: a disputed fortress in the battle against Mycobacterium tuberculosis. Front. Microbiol. 8, 2284 (2017).

  48. 48.

    Haney, M. S. et al. Identification of phagocytosis regulators using magnetic genome-wide CRISPR screens. Nat. Genet. 50, 1716–1727 (2018).

  49. 49.

    Drevets, D. A. Dissemination of Listeria monocytogenes by infected phagocytes. Infect. Immun. 67, 3512–3517 (1999).

  50. 50.

    Maculins, T., Fiskin, E., Bhogaraju, S. & Dikic, I. Bacteria-host relationship: ubiquitin ligases as weapons of invasion. Cell Res. 26, 499–510 (2016).

  51. 51.

    Roy, M. F. et al. Pyruvate kinase deficiency confers susceptibility to Salmonella typhimurium infection in mice. J. Exp. Med. 204, 2949–2961 (2007).

  52. 52.

    Barthel, M. et al. Pretreatment of mice with streptomycin provides a Salmonella enterica serovar Typhimurium colitis model that allows analysis of both pathogen and host. Infect. Immun. 71, 2839–2858 (2003).

  53. 53.

    Eva, M. M. et al. Altered IFN-γ-mediated immunity and transcriptional expression patterns in N-ethyl-N-nitrosourea-induced STAT4 mutants confer susceptibility to acute typhoid-like disease. J. Immunol. 192, 259–270 (2014).

  54. 54.

    Marei, H. et al. Differential Rac1 signalling by guanine nucleotide exchange factors implicates FLII in regulating Rac1-driven cell migration. Nat. Commun. 7, 10664 (2016).

  55. 55.

    Benard, V., Bohl, B. P. & Bokoch, G. M. Characterization of Rac and Cdc42 activation in chemoattractant-stimulated human neutrophils using a novel assay for active GTPases. J. Biol. Chem. 274, 13198–13204 (1999).

  56. 56.

    Marei, H., Carpy, A., Macek, B. & Malliri, A. Proteomic analysis of Rac1 signaling regulation by guanine nucleotide exchange factors. Cell Cycle 15, 1961–1974 (2016).

  57. 57.

    Karlinsey, J. E. λ-Red genetic engineering in Salmonella enterica serovar Typhimurium. Methods Enzym. 421, 199–209 (2007).

  58. 58.

    Skarnes, W. C. et al. A conditional knockout resource for the genome-wide study of mouse gene function. Nature 474, 337–342 (2011).

  59. 59.

    Fortin, A. et al. Recombinant congenic strains derived from A/J and C57BL/6J: a tool for genetic dissection of complex traits. Genomics 74, 21–35 (2001).

  60. 60.

    Cunnington, A. J., de Souza, J. B., Walther, M. & Riley, E. M. Malaria impairs resistance to Salmonella through heme- and heme oxygenase-dependent dysfunctional granulocyte mobilization. Nat. Med. 18, 120–127 (2011).

  61. 61.

    Sixt, M. & Lammermann, T. In vitro analysis of chemotactic leukocyte migration in 3D environments. Methods Mol. Biol. 769, 149–165 (2011).

Download references

Acknowledgements

We thank L. Larivière, E. Flamant, L. Rached-D’Astous, J. Kim, F. Pampaloni and M. Sader for technical assistance, N. Prud’homme and P. D’Arcy for mouse breeding and screening, S.-J. Pilon for histopathology scoring, and K. Koch and T. Maculins for reviewing the manuscript. We also acknowledge A. Malliri for reagents and G. White for the preparation of the PAK-CRIB peptide. This work was supported by the Canadian Institutes of Health Research (grant no. MOP133700 to D.M.), the DFG-funded Collaborative Research Centre on Selective Autophagy (grant no. SFB 1177), the European Research Council under the European Union’s Horizon 2020 research and innovation program (grant agreement no. 742720), the DFG-funded Cluster of Excellence ‘Macromolecular Complexes’ (grant no. EXC115), the DFG-funded SPP 1580 program ‘Intracellular Compartments as Places of Pathogen-Host-Interactions’, and by the LOEWE program Ubiquitin Networks (Ub-Net) and the LOEWE Centre for Gene and Cell Therapy Frankfurt, which are both funded by the State of Hessen, Germany (to I.D.). Funding was provided to M.M.E. by Le Fonds de Recherche Santé du Québec and to H.M. by the Alexander von Humboldt foundation as a Humboldt research fellowship for postdoctoral researchers.

Author information

K.E.Y., H.M., E.F., M.M.E., S.M.V., M.C., D.M. and I.D. designed the conceptual framework of the study and experiments. K.E.Y. and M.M.E. contributed to ENU mutation identification and performed all in vivo and ex vivo experiments involving Cyri mutant and conditional alleles and analysed the data. H.M. designed, performed and analysed all biochemical and cell biology assays. E.F. designed, performed and analysed all mass spectrometry and in vitro Salmonella-infection experiments and performed some biochemical assays. J.A.S. and J.M. performed the analyses of exome sequences. A.A.G. and J.N.M. designed, performed and analysed the experiments with neutrophil mobility assays. D.M. and I.D. supervised the project. All authors wrote the manuscript.

Correspondence to Danielle Malo or Ivan Dikic.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary Figs. 1–13, Supplementary Tables 1–5, Supplementary Notes and Supplementary References.

Reporting Summary

Supplementary File 1

Ubiquitylation sites quantified from three replicate SILAC-diGly proteomics experiments in HeLa Flp-In T-REx GFP–SopE cells, as described in Fig. 7.

Supplementary File 2

Statistics for main and supplementary figures.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark