Typhoid toxin sorting and exocytic transport from Salmonella Typhi-infected cells

Typhoid toxin is an essential virulence factor for Salmonella Typhi, the cause of typhoid fever in humans. This toxin has an unusual biology in that it is produced by Salmonella Typhi only when located within host cells. Once synthesized, the toxin is secreted to the lumen of the Salmonella-containing vacuole from where it is transported to the extracellular space by vesicle carrier intermediates. Here, we report the identification of the typhoid toxin sorting receptor and components of the cellular machinery that packages the toxin into vesicle carriers, and exports it to the extracellular space. We found that the cation-independent mannose-6-phosphate receptor serves as typhoid toxin sorting receptor and that the coat protein COPII and the GTPase Sar1 mediate its packaging into vesicle carriers. Formation of the typhoid toxin carriers requires the specific environment of the Salmonella Typhi-containing vacuole, which is determined by the activities of specific effectors of its type III protein secretion systems. We also found that Rab11B and its interacting protein Rip11 control the intracellular transport of the typhoid toxin carriers, and the SNARE proteins VAMP7, SNAP23, and Syntaxin 4 their fusion to the plasma membrane. Typhoid toxin’s cooption of specific cellular machinery for its transport to the extracellular space illustrates the remarkable adaptation of an exotoxin to exert its function in the context of an intracellular pathogen.


INTRODUCTION 1
Identification of the cation-independent mannose-6-phosphate receptor as the typhoid 2 toxin cargo receptor for sorting into vesicle transport carriers 3 To identify the putative typhoid toxin sorting receptor, we sought to identify typhoid-toxin 4 interacting proteins in cultured human epithelial cells by affinity purification and liquid 5 chromatography-tandem mass spectrometry (LC-MS/MS) analysis (Fig. 1a). As a negative 6 control we used a mutant version of typhoid toxin that carries a single amino acid substitution 7 in its PltB subunit (PltB S35A ). This mutant cannot be packaged into vesicle carrier intermediates 8 because this residue is critical for its interaction with its putative sorting receptor (Chang et al., 9 2016). The cation-independent mannose-6-phosphate receptor (CI-M6PR) 10 2007; Stalder and Gershlick, 2020) was identified as a prominent interacting protein with wild 11 type typhoid toxin but not with its PltB S35A mutant control (Supplementary data set 1). To 12 validate this interaction, we infected cultured human epithelial cells with S. Typhi strains 13 expressing FLAG-epitope tagged versions of wild type typhoid toxin or its PltB S35A mutant form 14 and examined their interaction with CI-M6PR by immunoprecipitation and western blot 15 analysis. Similar to the in-vitro experiments, we detected the interaction of CI-M6PR with wild 16 type typhoid toxin but not with the PltB S35A mutant version (Fig. 1b). Consistent with these 17 observations, we found that CI-M6PR is recruited to the S. Typhi containing vacuole, 18 particularly later in infection ( Fig. 1c and 1d and Supplementary data set 2). This observation 19 is intriguing since it is well established that CI-M6PR is not recruited to the vacuoles that 20 contain S. Typhimurium(Garcia-del Portillo and Finlay, 1995;McGourty et al., 2012), which 21 does not encode typhoid toxin. Taken together, these results show that CI-M6PR is recruited 22 to the S. Typhi-containing vacuoles where it interacts with typhoid toxin through its PltB 23 subunit.

24
To specifically examine the contribution of CI-M6PR to typhoid toxin sorting into vesicle 25 carriers we generated a CI-M6PR-defficient cell line by CRISPR/Cas9 genome editing (Fig. 26 2a). We found that inactivation CI-M6PR did not affect the ability of S. Typhi to gain access to 27 and replicate within these cells (Fig. 2b and Supplementary data set 3), neither it affected the 28 al., 2011). It is also well-established that the S. Typhimurium-containing vacuole does not 1 recruit CI-M6PR, and that the lack of recruitment of this lysosomal marker is also linked to the 2 activity of type III secreted effectors (McGourty et al., 2012) (Fig. 3a, Supplementary Fig. S2, 3 and Supplementary data set 4). In contrast, we found that the CI-M6PR is recruited to the S. 4 Typhi-containing vacuole where it serves as a sorting receptor for typhoid toxin export (Fig. 5 3a and 1d). These findings suggest that specific features of the S. Typhi-containing vacuole 6 determined by the type-III-secreted-effector repertoire are essential for typhoid toxin export. 7 To further explore this hypothesis, we cloned the entire S. Typhi genomic islet that encodes 8 typhoid toxin into S. Typhimurium and examined typhoid toxin export in cells infected with the 9 resulting strain. We reasoned that differences in the vacuolar environments of S. Typhimurium 10 may prevent the export of typhoid toxin when expressed in this strain. Cells infected with the 11 S. Typhimurium strain expressing typhoid toxin showed markedly reduced levels of vesicle 12 carrier intermediates when compared to S. Typhi infected cells, despite equivalent levels of 13 toxin expression (Fig. 3b, Supplementary Fig. S2, and Supplementary data set 4). This is 14 consistent with the observation that the typhoid toxin receptor CI-M6PR is not recruited to the 15 S. Typhimurium-containing vacuole (Fig. 3a). These observations are also consistent with the 16 notion that typhoid toxin packaging into vesicle carrier intermediates requires the specific 17 environment of the S. Typhi containing vacuole. Since the nature of the Salmonella-containing 18 vacuole is strictly dependent on the function of its type III secretion systems (Galán, 2001;19 Jennings et al., 2017), we reasoned that differences in the composite of effector proteins 20 expressed by S. Typhi and S. Typhimurium may be ultimately responsible for the differences 21 in the ability of typhoid toxin to be packaged into vesicle carrier intermediates. More 22 specifically, we reasoned that the action of some type III effector encoded by S. Typhimurium 23 but absent from S. Typhi may preclude the recruitment of CI-M6PR to the Salmonella-24 containing vacuole and thus prevent the packaging of typhoid toxin into transport carriers. To 25 test this hypothesis, we individually expressed in S. Typhi the S. Typhimuirum SPI-1 or SPI-2 26 T3SS effector proteins that are absent or pseudogenes in S. Typhi (Parkhill et al., 2001) (i. e. 27 SopD2, GtgE,SseJ,SteB,SlrP,SseK1,SseK2,SseK3,GtgA,SseI,SpvC,SpvD,GogB,and 28 SspH1). We infected cells with the resulting strains and examined the recruitment of CI-M6PR 1 to the S. Typhi-containing vacuoles. Of all the effectors tested, only the expression of SseJ 2 resulted in a significant reduction in the recruitment of CI-M6PR to the S. Typhi-containing 3 vacuole (Fig. 3c, 3d, Supplementary Fig. S2 and Supplementary data set 4). Consistent with 4 the reduced recruitment of its sorting receptor, the formation of the typhoid toxin carriers in 5 cells infected with the SseJ-expressing S. Typhi strain was significantly reduced (Fig. 3e,   6 Supplementary Fig. S2 and Supplementary data set 4). SseJ has been shown to modify the 7 lipid composition of the Salmonella-containing vacuole by esterifying cholesterol through its 8 glycerophospholipid:cholesterol acyltransferase activity (Kolodziejek and SI, 2015;Ohlson et 9 al., 2005). In addition, in a catalytic-independent manner, SseJ recruits the eukaryotic lipid 10 transporter oxysterol binding protein 1 (OSBP1) to the Salmonella-containing vacuole, thus 11 contributing to the integrity of this compartment (Kolodziejek et al., 2019). To investigate which 12 of these activities is responsible for the prevention of the recruitment of CI-M6PR to the S.

13
Typhi-containing vacuole, we expressed a catalytic mutant of SseJ (SseJ S151A ) in S. Typhi and 14 examined the recruitment of CI-M6PR to the S. Typhi-containing vacuole. We found that in 15 contrast to wild type, expression of SseJ S151A had no effect in the recruitment of CI-M6PR to  To investigate potential mechanisms by which the typhoid toxin export carriers are formed, we 25 sought to identify specific coat proteins and cargo adaptors that promote the budding of toxin 26 carriers from the Salmonella-containing vacuole. Vesicle coats that initiate the budding subunits (Dell'Angelica and Bonifacino, 2019;McMahon and Mills, 2004). We specifically 1 examined the potential contribution to typhoid toxin packaging of four well-characterized coat 2 or adaptor proteins: clathrin (Briant et al., 2020 ), coat protein complex II (COPII)(McCaughey 3 and Stephens, 2018), and adaptor related protein complex 3 (AP3) and 4 (AP4) (Hirst et al.,4 2013; Odorizzi et al., 1998). We focused on these coat proteins because they carry out their 5 function at distinct compartments within the secretory pathway. Using CRISPR/Cas9 genome 6 editing, we generated HEK293T cells defective for CLTC (clathrin heavy chain), SEC23B (an 7 inner coat protein of COPII), AP3B1 (AP3 subunit Beta-1), and AP4M1 (AP4 subunit Mu-1), 8 whose deficiency abrogates the assembly of their respective vesicle coats (Supplementary 9 Fig. S3a). We found that typhoid toxin expression as well as the toxicity of exogenously  data set 5). If COPII plays a pivotal role in the formation of typhoid toxin export carriers we 7 reasoned that we should be able to observe a complex of typhoid toxin and components of 8 the COPII coat. To investigate this hypothesis, HEK293T cells transiently expressing GFP-9 tagged Sec23 were infected with an S. Typhi strain expressing FLAG epitope-tagged CdtB 10 and the formation of a typhoid toxin/Sec23 complex was investigated with immunoprecipitation 11 experiments. We found that typhoid toxin could be detected in complex with SEC23 ( Fig. 4i).

12
In contrast, this interaction was markedly reduced in cells infected with the S. Typhi 13 ∆spiA mutant strain (Fig. 4i). This mutant strain is defective in the SPI-2 T3SS and 14 consequently has an altered intracellular vacuole and is therefore unable to efficiently package 15 typhoid toxin into vesicle transport carriers (Chang et al., 2016). Taken together, our results 16 indicate that the packaging of typhoid toxin into vesicle carrier intermediates is dependent on 17 COPII.

19
The transport of typhoid toxin vesicle carriers is dependent on Rab11B and its 20 interacting protein RIP11

21
The transport of cargo proteins destined for targeting to other intracellular compartments or 22 the extracellular space most often entails multiple steps, which requires a specific set of 23 proteins that ensure the accuracy of cargo delivery. Among these proteins are Rab GTPases, 24 which regulate vesicular transport to most cellular compartments, and are therefore excellent 25 candidates to participate in the regulation of the exocytic transport of typhoid toxin (Pfeffer,26 2017; Stenmark, 2009;Zhen and Stenmark, 2015). To identify Rab GTPases that could be 27 potentially involved in typhoid toxin transport, we examined typhoid toxin transport in cell lines 28 deficient in a subset of Rab-family GTPases (or its regulators) that have been previously 1 implicated in various exocytic pathways. More specifically, using CRISPR/Cas9 genome 2 editing we generated cell lines deficient in Rab27A, Rab27B, Rab11A, Rab11B, and HPS4, a 3 component of the BLOC3 complex which is an exchange factor essential for the function of 4 both, Rab32 and Rab38 (Gerondopoulos et al., 2012) (Supplementary Fig. S6). The different 5 cell lines were then infected with S. Typhi and the levels of typhoid toxin in the infection media, 6 a direct measure of its export, were determined. We found that inactivation Rab27A, Rab27B, Most often Rab GTPases modulate vesicular transport by engaging specific 20 downstream effector proteins, which carry out specific functions. In the case of Rab11B in 21 particular, it has been shown that it regulates some exocytic events by interacting with Rab11-22 family-interacting protein 5 (Rab11-FIP5) also known as Rip11 (Horgan and McCaffrey, 2009). 23 This effector has been shown to be involved in regulating vesicular transport from recycling 24 endosomes to the apical plasma membrane, and in particular, in the regulation of insulin 25 granule exocytosis (Schonteich et al., 2008;Sugawara et al., 2009). It exerts its function at 26 least in part by linking vesicle carriers to kinesin motors (Schonteich et al., 2008), which 27 presumably move the vesicles through microtubule tracks. To determine whether Rab11-FIP5 also participates in toxin export, we generated Rab11-FIP5-defficent cells by CRISPR/Cas9 1 gene editing and examined typhoid toxin export after S. Typhi infection (Supplementary Fig.   2 S6). We found that the absence of Rip11 resulted in a significant reduction in the export of 3 typhoid toxin to the extracellular medium relative to parental cell (  The export of typhoid toxin is expected to involve steps in which the vesicle carriers harboring 13 the toxin fuse with the plasma membrane and release their cargo to the extracellular space. 14 In most exocytic pathways this step involves soluble N-ethylmaleimide-sensitive factor-15 activating protein receptors (SNAREs), which mediate the fusion between the carrier vesicle 16 and target membranes (Goda, 1997;Südhof and Rothman, 2009). Depending on whether they 17 are located in the vesicle or target membranes, SNAREs are referred to as v-or t-SNAREs, 18 respectively. The identities of the trans-SNARE pairing ensure the specificity of the fusion 19 process. We therefore hypothesized that a pair of v-and t-SNAREs must be involved in 20 targeting typhoid toxin-containing vesicles to the plasma membrane. We reasoned that a good 21 candidate tSNARE to be potentially involved in typhoid toxin export would be SNAP23, which 22 is ubiquitously expressed, localizes to the plasma membrane, and has been implicated in 23 several exocytic processes (Kádková et al., 2019). We therefore generated SNAP23-defficient Supplementary data set 8). We hypothesized that this increase was related to the failure of 4 those carriers to fuse with the plasma membrane due to the absence of SNAP23. Taken 5 together, these results are consistent with the hypothesis that SNAP23 is involved in the fusion 6 of the typhoid toxin vesicle carrier intermediates with the plasma membrane. 7 SNAP23 has been shown to cooperate with other SNARE proteins such as syntaxin 4 8 (STX4) and syntaxin 11 (STX11) to mediate membrane fusion (Lin et al., 2017;Ye et al., 2012). 9 Therefore, using CRISPR/Cas9 gene editing we generated STX4-and STX11-defficient cells

19
The SNAP23/STX4 complex has been shown to interact with the vSNARE VAMP- intermediates were not significantly altered in these cells ( Fig. 6c and Supplementary Fig. S8). the VAMP7/SNAP23/STX4 SNARE complex, which targets the typhoid toxin vesicle carrier 1 intermediates for fusion to the plasma membrane and subsequent release of the toxin cargo 2 to the extracellular space. Typhoid toxin has a very unusual biology for an exotoxin in that it is produced by intracellularly 7 localized bacteria, is secreted into the lumen of the S. Typhi-containing vacuole, is packaged 8 into vesicle intermediates, and then transported to the extracellular milieu (Chang et al., 2019;9 Chang et al., 2016;Geiger et al., 2020;Geiger et al., 2018;Hodak and Galán, 2013;Spano 10 et al., 2008). Here, we have investigated the mechanisms by which typhoid toxin is transported 11 from the S. Typhi containing vacuole to the extracellular space. Through an affinity purification 12 approach, we identified CI-M6PR as the typhoid toxin packaging receptor, which we found to 13 be robustly recruited to the S. Typhi-containing vacuole. This finding was surprising since it is 14 well established that this receptor is excluded from the S. Typhimurium-containing 15 vacuole(Garcia-del Portillo and Finlay, 1995;McGourty et al., 2012). Consistent with these 16 observations, we found that typhoid toxin is not packaged into vesicle carrier intermediates 17 when expressed in S. Typhimurium, indicating that the specific features of the S. Typhi-18 containing vacuole are essential for the formation of the typhoid toxin transport carrier, and 19 underlying the marked differences between the intracellular compartments that harbor these 20 pathogens. The properties of the Salmonella-containing vacuole are determined by the activity 21 of bacterial effectors delivered by either of its type III secretion systems, which differ 22 significantly between S. Typhi and S. Typhimurium. In fact, we found that expression in S. 23 Typhi of SseJ, a S. Typhimurium effector of its SPI-2 T3SS that is absent from S. 24 Typhi (Kolodziejek and SI, 2015;Ohlson et al., 2005), prevented the recruitment of CI-M6PR compartments that harbor S. Typhi and S. Typhimurium but also indicate that typhoid toxin 7 has evolved to adjust its biology to the intracellular biology of S. Typhi.

8
The engagement of the sorting receptor by typhoid toxin presumably initiates a 9 packaging and budding event that leads to the formation of the vesicle carrier intermediates. 10 We found that this step requires the activities of the coat protein complex COPII and the Sar1  Our studies also identified other components of the membrane trafficking machinery 20 that are presumably involved in the transport of the vesicle carriers from the SCV to the plasma 21 membrane. In particular, we identified Rab11B as required for the efficient transport of typhoid 22 toxin to the extracellular space. Although cells deficient in Rab11B showed reduced amount 23 of toxin in the infection media, they showed normal levels of vesicle carrier intermediates.

24
These results indicate that Rab11B must be involved in steps downstream from the vesicle 25 transport budding from the SCV but upstream from the fusion of these vesicles to the plasma 26 membrane. Rab11B has been implicated in various vesicle transport pathways to the plasma 27 membrane including melanin exocytosis (Tarafder et al., 2014 ). This observation is intriguing since the S. Typhi-containing vacuole (but not the S. Typhimurium vacuole) recruits Rab32 1 and Rab38, which are involved in melanosome biogenesis (Spanò and Galán, 2012;Spanò et 2 al., 2016). However, we found that Rab32 and Rab38 are not required for typhoid toxin 3 transport since cells deficient in Hsp4, an essential component of their exchange 4 factor (Gerondopoulos et al., 2012), were unaffected in typhoid toxin transport. The 5 mechanisms by which Rab11B may regulate typhoid toxin transport are unclear. However, we 6 hypothesize that this GTPase may be involved in the movement of the typhoid toxin carriers 7 along microtubule tracks. This hypothesis is based on the observation that Rab11B has been 8 shown to modulate the exocytic transport of insulin granules at least in part by linking these 9 transport carriers to kinesin motors, a function that requires the Rab11B effector 10 Rip11 (Sugawara et al., 2009). Consistent with this hypothesis, cells lacking Rip11 exhibited 11 reduced transport of typhoid toxin to the extracellular space although they showed normal 12 levels of typhoid toxin transport carriers.

13
Finally, we identified the machinery that is required for the fusion of the typhoid toxin 14 transport carriers to the plasma membrane. More specifically we found that the plasma and we found here that cells defective in VAMP-7 showed reduced levels of typhoid toxin in 23 the extracellular media. Therefore, we propose that the fusion of the typhoid toxin vesicle 24 carrier intermediates with plasma membrane is controlled by the plasma membrane tSNARES 25 SNAP23 and STX4, and the typhoid toxin vesicle carrier vSNARE VAMP7.

26
In summary, our studies have unraveled the mechanisms by which typhoid toxin is 27 transported from the Salmonella-containing vacuole to the extracellular space. The export pathway, which has evolved to be specifically adapted to the biology of Salmonella Typhi, has 1 coopted cellular machinery involved in various secretory and exocytic pathways (Fig. 7).

2
These findings reveal how vesicle trafficking pathways that are seemingly unconnected can 3 be coopted by microbial pathogens to carry out a specific function.

Bacterial strains and plasmids
The wild-type Salmonella enterica serovars Typhi ISP2825 (Galán and Curtiss III, 1991) and Typhimurium SL1344 (Hoiseth and Stocker, 1981) have been described previously. All the Salmonella mutant derivatives were constructed by standard recombinant DNA and allelic exchange procedures as previously described (Kaniga et al., 1994) using the E. coli β-2163 Δnic35 strain (Demarre et al., 2005) as the conjugative donor and are listed in Table S1.
All the plasmids listed in Table S2 were constructed using the Gibson assembly cloning strategy (Gibson et al., 2009) and verified by nucleotide sequencing.

Cell culture and bacterial infection
Human intestinal epithelial Henle-407 cells (obtained from the Roy Curtiss III collection in 1987) and HEK293T cells (from the American Type Culture Collection) were cultured in DMEM supplemented with 10% fetal bovine serum. Overnight cultures of the S. Typhi strains were diluted 1/20 in LB broth medium containing 0.3 M NaCl and grown until they reached an OD600 of 0.9. Culture cells were infected as described before (Chang et al., 2016). Briefly, cells were infected with the different S. Typhi or S. Typhimurium strains at a multiplicity of infections (MOI) of 10 or 30 (as indicated) in Hank's balanced salt solution (HBSS), and then incubated with culture medium containing 100 µg/ml gentamicin for 1 hr to kill extracellular bacteria. Cells were washed and incubated for the indicated times with medium containing 10 µg/ml gentamicin to avoid cycles of reinfection. All cell lines were routinely tested for a mycoplasma by a standard PCR method.

Visualization of typhoid toxin export carrier intermediates
The visualization of typhoid toxin vesicle carrier intermediates was carried out as previously described (Chang et al., 2016). Briefly, 24 hours after infection with the indicated S. Typhi strains expressing FLAG-epitope tagged CdtB, cells were fixed in 4% paraformaldehyde and then blocked with 3% BSA, 0.3% Triton X-100 in DPBS. Fixed cells were incubated with primary mouse monoclonal anti-FLAG M2 (Sigma) and rabbit polyclonal anti-S. Typhi lipopolysaccharide (Sifin) antibodies followed by Alexa 488-conjugated anti-mouse and Alexa 594-conjugated anti-rabbit antibodies (Invitrogen). Samples were visualized under a Nikon TE2000 fluorescence microscope equipped with and Andor Zyla 5.5 CMOS camera driven by Micromanager software (https://www.micro-manager.org). Quantification of typhoid toxin export carrier intermediates has been previously described (Chang et al., 2016). Briefly, images were analyzed using the open-source software ImageJ (https://imagej.nih.gov/). The LPS stain was used to identify the area corresponding to the bacterial cell body and this area was used to obtain the bacterial-associated typhoid toxin fluorescence signal, which was substracted from the typhoid toxin-associated fluorescence. The remaining fluorescence was defined as the area of typhoid toxin carrier intermediates. The intensity of fluorescence associated with toxin carriers was normalized using the fluorescence-associated with typhoid toxin within bacterial cells in the same field.

Immunofluorescent staining of CI-M6PR in infected cells
Henle-407 cells infected with the different S. Typhi or S. Typhimurium strains expressing superfolder GFP were fixed with ice-cold methanol for 5 min at -20˚C. Cells were then washed with DPBS three times, permeabilized with 0.1% Triton-100/DPBS for 20 min at room temperature, and then stained with the antibodies against CI-M6PR and GFP overnight at 4°C, and Alexa 488 and 594-conjugated antibodies (Invitrogen) for 1 hr at room temperature. Cells were then observed under a Nikon TE2000 fluorescence microscope and images captured with an Andor Zyla 5.5 CMOS camera driven by Micromanager software (https://www.micro-manager.org).

Bacteria internalization and intracellular replication
The gentamicin protection assay was used to assess the invasion of S. Typhi within host cells (Galán and Curtiss III, 1989). Briefly, HEK293T cells were cultured in 24-well plates and infected with the indicated Salmonella strains. One hour after infection, cells were incubated with culture medium containing gentamicin and then harvested in DPBS containing 0.1% sodium deoxycholate at 1 and 24 hours after addition of the antibiotic. The bacteria released from cultured cells were plated onto LB agar plates to determine colony-forming units.

Affinity purification-mass spectrometry analysis
HEK293T and Henle-407 cells were seeded onto each five 10 cm dishes at 70% confluency.
After 16-18 hours, culture cells were harvested, lysed in 5 ml of lysis buffer (0.5% Triton-X-100,150 mM NaCl, 50 mM Tris-HCl) containing protease inhibitors for 30 min on ice, and then centrifuged for 15 min at 14,000 × rpm at 4 °C. The supernatants were incubated with 30 μg of the purified typhoid toxin overnight at 4 °C. Next day, the protein complex was incubated with 20 μl of anti-FLAG M2 agarose for 2 hours at 4 °C. Immunoprecipitation of protein complexes was collected by centrifugation at 500 × g for 1 min, followed by washing with lysis buffer to avoid non-specific binding. The protein complexes were eluted by adding 0.1 M glycine HCL, pH 3.5, and digested in solution with trypsin overnight. The extracted peptides were subjected to LC-MS/MS analysis as previously described (Sun et al., 2018).

Co-immunoprecipitation Assay
HEK293T cells were seeded at a density of 2x10 6 onto 10 cm dishes. Twenty-four hours later, cells were infected with S. Typhi strains expressing FLAG-epitope tagged CdtB or mock-

Typhoid toxin export assay
Cultured cells were grown onto 6-well plates and infected with wild-type S. Typhi. Twenty-four hours after infection, the supernatant from infected cells was collected and filtered through 0.2 μm syringe filters, diluted as indicated in each experiment, and applied to fresh wild-type HEK293T cells. The toxic effect of different dilutions of the culture supernatant was determined at 48 hr post-treatment as previously described (Chang et al., 2016). Briefly, treated cells were trypsinized, fixed for 1 hr in 70% ethanol/DPBS at -20℃, washed with DPBS, and then resuspend in DPBS containing 50 µg/ml propidium iodide, 0.1 mg/ml RNase A, and 0.05% Triton X-100. Cell cycle analysis was carried out by flow cytometry, and the proportion of cells in the G2M phase was determined using FlowJo.

Toxin expression and purification
Purification of typhoid toxin was conducted as described previously (Song et al., 2013). Briefly, the genes encoding different versions of typhoid toxins in S. Typhi (as listed in Table S2) were cloned into the pET28a (Novagen) expression vector. Escherichia coli strains carrying the plasmids encoding the different toxins were grown at 37˚C in LB media to an OD600 of ~0.6, toxin expression was induced by the addition of 0.5 mM IPTG, and cultures were further incubated at 25˚C overnight. Bacterial cell pellets were resuspended in a buffer containing 15 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.1 mg/ml DNase, 0.1 mg/ml lysozyme, and 0.1% PMSF and lysed by passaging through a cell disruptor (Constant Systems Ltd.). Toxins were then purified from bacterial cell lysates through affinity chromatography on a Nickel-resin (Qiagen), ion exchange, and gel filtration (Superdex 200) chromatography as previously described (Song et al., 2013). Purified toxins were examined for purity on SDS-PAGE gels stained with coomassie blue.

Statistical analysis
The p values were calculated using a two-tailed, unpaired Student's t-test for two-group comparisons in GraphPad Prism (GraphPad software). p values below 0.05 were considered statistically significant.

Data and software availability
The following software was used in this study: Graphpad Prism (plotting data), Micro-Manager, Slidebook 6, Adobe Illustrator & Adobe Photoshop (image preparation), FlowJo (analysis of flow cytometry data), and Image Studio Lite (Li-COR Biosciences) (quantification of the band intensity of western blot).

Competing interests
The authors declare no competing interests.     Values were normalized relative to wild-type cells, which was considered to be 100, and are the mean ± SEM. ****: p < 0.0001, unpaired two-sided t test. TT: typhoid toxin.  represent the degree of co-localization between CI-M6PR and Salmonella-containing vacuoles and are the mean ± SEM. ****: p < 0.0001, unpaired two-sided t test. The results shown are from two replicates of the experiment shown in Fig. 3a. (b) Quantification of the intensity of typhoid toxin-associated fluorescent puncta associated with typhoid toxin carrier intermediates in infected cells. Cells infected with either S. Typhi or the S. Typhimurium strain expressing FLAG-tagged