Salmonella enterica serovar Typhi uses two type 3 secretion systems to replicate in human macrophages and colonize humanized mice

ABSTRACT Salmonella enterica serovar Typhi (S. Typhi) is a human-restricted pathogen that replicates in macrophages. In this study, we investigated the roles of the S. Typhi type 3 secretion systems (T3SSs) encoded on Salmonella pathogenicity islands (SPI)-1 (T3SS-1) and SPI-2 (T3SS-2) during human macrophage infection. We found that mutants of S. Typhi deficient for both T3SSs were defective for intramacrophage replication as measured by flow cytometry, viable bacterial counts, and live time-lapse microscopy. T3SS-secreted proteins PipB2 and SifA contributed to S. Typhi replication and were translocated into the cytosol of human macrophages through both T3SS-1 and T3SS-2, demonstrating functional redundancy for these secretion systems. Importantly, an S. Typhi mutant strain that is deficient for both T3SS-1 and T3SS-2 was severely attenuated in the ability to colonize systemic tissues in a humanized mouse model of typhoid fever. Overall, this study establishes a critical role for S. Typhi T3SSs during its replication within human macrophages and during systemic infection of humanized mice. IMPORTANCE Salmonella enterica serovar Typhi is a human-restricted pathogen that causes typhoid fever. Understanding the key virulence mechanisms that facilitate S. Typhi replication in human phagocytes will enable rational vaccine and antibiotic development to limit the spread of this pathogen. While S. Typhimurium replication in murine models has been studied extensively, there is limited information available about S. Typhi replication in human macrophages, some of which directly conflict with findings from S. Typhimurium murine models. This study establishes that both of S. Typhi’s two type 3 secretion systems (T3SS-1 and T3SS-2) contribute to intramacrophage replication and virulence.

T yphoid fever is a transmissible disease caused by the bacterium Salmonella enterica serovar Typhi (S. Typhi). Before widespread antibiotic use, over 50% of typhoid cases resulted in serious complications, with a mortality rate of over 10% (1). Outbreaks of multidrug-and extensively drug-resistant typhoid are increasing in magnitude and frequency, and climate change will likely exacerbate transmission, underscoring the need to identify new antimicrobial targets (2,3).
The molecular mechanisms underlying typhoid pathology are poorly understood due to the lack of experimental animal models, as S. Typhi is human restricted. By studying the related pathogen S. enterica serovar Typhimurium (S. Typhimurium), which causes a typhoid-like disease in mice, researchers have identified many molecular mechanisms underlying S. Typhimurium virulence. However, these two serovars cause distinct disease states during human infection. Although both pathogens can replicate inside phag ocytes, S. Typhimurium infection is usually restricted to the human gastrointestinal

Salmonella Typhi uses both T3SS-1 and T3SS-2 to replicate in human macrophages
To assess the relative contributions of T3SS-1 and T3SS-2 to S. Typhi replication within macrophages, we infected phorbol 12-myristate 13-acetate (PMA)-differentiated THP-1 macrophages, a human-derived promonocytic cell line commonly used to model Salmonella infection (12,(17)(18)(19). To measure S. Typhi replication in THP-1 macrophages, we used a previously published method of quantifying fluorescence dilution (pFCcGi), a dual-fluorescence tool which permits direct assessment of intramacrophage bacterial replication (11,20). By flow cytometry, the wild-type (WT) S. Typhi Ty2 strain replicated 8-to 11-fold over 24 h (Fig. 1A). In contrast, an isogenic S. Typhi ΔphoP strain, which is deficient for a virulence transcriptional regulatory protein and is unable to survive within macrophages, did not replicate in THP-1 macrophages (Fig. 1A), which is consistent with previous studies (12,13). To test the roles of the T3SS-1 and T3SS-2 in S. Typhi replication within human macrophages, we deleted InvA (ΔinvA) or SsaV (ΔssaV), the export gates in T3SS-1 and T3SS-2, respectively. The S. Typhi ΔinvA strain replicated in THP-1 macrophages to the same extent as WT S. Typhi (8-fold), whereas the S. Typhi ΔssaV strain replicated slightly less than the WT strain (7-fold) (Fig. 1A). To test the possibility that both T3SSs contribute to intramacrophage replication, a double knockout S. Typhi ΔinvA ΔssaV (T3SS-null) strain was constructed, and intramacrophage replication was measured. In contrast to the single T3SS-deficient strains, the strain that is deficient for both T3SSs had a severe defect in THP-1 macrophages (three-to fivefold replication), nearly phenocopying the ΔphoP strain (Fig. 1A). To validate the fluorescence dilution results, viable bacteria were enumerated at 2-and 24-h postinoculation (h.p.i.). by colony-forming unit (CFU) plating (Fig. S1A). In agreement with the fluorescence dilution assay results, WT S. Typhi replicated to significantly higher levels (~10-fold) compared with the ΔphoP strain (Fig. S1B). In contrast, the strains that are deficient for one of the T3SSs (the ΔinvA and ΔssaV strains) replicated to the same extent as the WT S. Typhi strain. Strikingly, the S. Typhi strain lacking both T3SSs (T3SS-null strain) had a significant replication defect compared with the WT S. Typhi strain (Fig. S1B). Importantly, the WT and T3SS-deficient bacterial strains were not defective for uptake by THP-1 macrophages (Fig. S1A), or for growth in a defined minimal medium (Fig. S1C), further validating an intramacrophage-specific replication defect. Finally, the levels of cell death induced during 24 h infections with WT or T3SS-deficient bacterial strains were low and similar for all strains (Fig. S1D), ruling out host cell death as a potential confounding factor.
To compare the kinetics of S. Typhi intramacrophage replication, we then performed live-cell imaging and quantified fluorescence dilution (pFCcGi) over 24 h for each strain relative to WT (Videos S1-2) (11,20). For each experiment, we obtained the ratio of mCherry to green fluorescent protein (GFP) signal within macrophages (Fig. S1E), subtracted background signal measured in uninfected wells, and then normalized to that of the WT strain at 24 h.p.i. to obtain a "replication index. " Throughout the infection, WT and ΔinvA S. Typhi strains replicated to similar levels in THP-1 macrophages (Fig. 1B). Although the S. Typhi ΔssaV strain replicated less than the WT strain at later timepoints, the T3SS-null strain had a significant replication defect throughout the time course, nearly phenocopying the ΔphoP control (Fig. 1B). To confirm that T3SS-dependent replication also occurs in primary human macrophages, we infected human blood monocyte-derived macrophages (hMDMs) and performed time-lapse microscopy to quantify replication at several timepoints during infection (Videos S3 and 4; Fig. S1F). Throughout infection, the T3SS-null mutant had a severe replication defect compared with the WT strain, phenocopying the ΔphoP strain, whereas the single ΔinvA and ΔssaV knockout S. Typhi strains had intermediate replication defects (Fig. 1C). Finally, we counted the number of intracellular bacteria per THP-1 macrophage or hMDM by confocal microscopy at 16 h.p.i. Although THP-1 macrophages contained a higher overall abundance of S. Typhi compared with hMDMs, the relative contributions of T3SS-1 and T3SS-2 to the abundance of bacteria inside hMDMs were similar to what was observed in THP-1 macrophages (Fig. 1D).
To confirm that gentamicin-protected S. Typhi was intracellular, hMDMs infected with either WT or T3SS-null bacteria were stained with an antibody to the endosomal membrane marker LAMP-1 and phalloidin to stain polymerized actin and analyzed by confocal microscopy. Both WT and T3SS-null S. Typhi colocalized with LAMP-1 and actin, suggesting that the gentamicin-protected bacteria reside within macrophages ( Fig. 1E and F).

S. Typhi T3SS-dependent effectors contribute to intramacrophage replication
We next interrogated which T3SS effectors contribute to S. Typhi replication in human macrophages. To this end, we constructed 25 mutant strains that are deficient for previously identified T3SS effectors and putative pseudogenes in S. Typhi (8). We then tested these mutants for replication defects in human macrophages using our time-lapse florescence dilution assay in 96-well dishes to facilitate higher throughput. Each dish contained wells infected with the WT S. Typhi strain and the ΔphoP and ΔinvA ΔssaV S. Typhi mutant strains, which do not replicate in THP-1 macrophages. As an additional control, we included an S. Typhi ΔsptP mutant, which lacks an effector gene that has been reported to be nonfunctional in S. Typhi (21). Consistent with these previous findings, we did not see a contribution of SptP to S. Typhi replication in macrophages ( Fig. 2A; Fig. S2A).
The results of this screen demonstrated that PipB2 and SifA contributed significantly to S. Typhi replication in human macrophages ( Fig. 2A; Fig. S2A). We confirmed an intramacrophage replication defect for the S. Typhi ΔpipB2 and ΔsifA mutant strains by plating viable bacteria (Fig. 2B). Importantly, the defects of the S. Typhi ΔpipB2 and ΔsifA mutant strains were rescued by providing WT copies of pipB2 and sifA, respectively (Fig.  2B). In addition, we noticed that S. Typhi strains lacking SteA and SopB have slight defects in the fluorescence dilution assay ( Fig. 2A), suggesting that they may contribute to intramacrophage replication, which is in agreement with previously published findings using S. Typhimurium models (22)(23)(24)(25)(26). However, we recovered the same CFUs for ΔsteA, and ΔsopB mutant S. Typhi strains as the WT S. Typhi strain (Fig. S2B). Collec tively, our results indicate that replication of S. Typhi in human macrophages is depend ent on both PipB2 and SifA. The S. Typhi ΔsifA mutant was more severely attenutated than the T3SS-2 mutant (ΔssaV) at 8 h.p.i. (Fig. 2C) with kinetics of intracellular replication similar to the ΔpipB2 mutant (Fig. 2C). Previous studies with S. Typhimurium infection of murine macrophages have shown that PipB2 is translocated through T3SS-1 at early timepoints, then by T3SS-1 and T3SS-2 at later timepoint (27). Therefore, we hypothe sized that the early translocation of SifA by S. Typhi is dependent on T3SS-1.

S. Typhi translocates PipB2 and SifA into macrophages through both T3SS-1 and T3SS-2
To assess translocation of S. Typhi T3SS effectors into the macrophage cytosol, we constructed translational fusions between S. Typhi SifA and PipB2 with the TEM-1 βlactamase reporter. GST fused to TEM-1 β-lactamase was used as a negative control.

Research Article mBio
Translocation was detected in THP-1 macrophages using the fluorescent β-lactamase substrate CCF4-AM as described previously (28,29). A fraction of macrophages infected with WT S. Typhi containing PipB2-BlaM or SifA-BlaM emitted blue fluorescence at 450 nm, whereas a GST-BlaM control did not, suggesting that fusions between these effectors and TEM-1 are translocated into host macrophages ( Fig. 3A through C). To determine whether translocation of SifA and PipB2 is dependent on T3SS-1 or T3SS-2, we measured translocation of SifA-BlaM or PipB2-BlaM into THP-1 macrophages infected with T3SS-1, T3SS-2, and T3SS-null S. Typhi strains at 8 and 16 h.p.i. (27). PipB2-BlaM translocation was reduced when either T3SS-1 or T3SS-2 was disabled at both timepoints ( Fig. 3B and C). Surprisingly, SifA-BlaM translocation was consistently reduced in the T3SS-1 mutant at both 8 and 16 h.p.i. In contrast, the T3SS-2 mutant did not have a translocation defect at 8 h.p.i. but did have a translocation defect at 16 h.p.i. (Fig. 3B and Research Article mBio C). We further confirmed T3SS-1-dependent translocation of SifA-BlaM in primary human macrophages by confocal microscopy of hMDMs infected with WT, ΔinvA, ΔssaV, and ΔinvA ΔssaV S. Typhi strains (Fig. S3). Taken together, these results indicate that S. Typhi replication in human macrophages is dependent on the effectors PipB2 and SifA and that translocation of PipB2 and SifA into macrophages is dependent on the presence of either T3SS-1 or T3SS-2.

S. Typhi T3SS-1 and T3SS-2 contribute to virulence in a humanized mouse model of typhoid fever
Although S. Typhi is human restricted, researchers have described systemic infection in mice with "humanized" immune systems (30). A recent study aimed at identifying S. Typhi virulence factors required for acute infection in NOD.Cg-Prkdc scid IL2rg tm1Wjl /SzJ (NSG) mice engrafted with human CD34 + hematopoietic stem cells derived from umbilical cord blood (hu-SRC-SCID mice) identified Vi capsule, lipopolysaccharide, aromatic amino acid biosynthesis, and the siderophore salmochelin as essential for virulence (13). However, they found that the T3SS-2 did not provide a competitive advantage during the first 48 h of infection of hu-SRC-SCID mice. Based on our in vitro findings in human macrophages, we hypothesized that both S. Typhi T3SSs contribute to virulence in the hu-SRC-SCID model. To test this idea, hu-SRC-SCID mice were infected as described by Karlinsey et al. Briefly, hu-SRC-SCID mice were infected intraperitoneally (IP) with an equal mixture of S. Typhi WT and isogenic ΔinvA ΔssaV mutant strain (10 5 CFU of each). The competitive index (CI) was calculated for the spleen and the liver at 2 days p.i. The mutant lacking both T3SSs was not outcompeted by the WT strain (Fig. 4A). How ever, we reasoned that a later timepoint may reveal a role for the T3SSs because the bacteria would have more time to replicate in human-derived macrophages in the humanized mice. To test this idea, we infected the hu-SRC-SCID mice IP with a 10-fold lower dose that contained an equal mixture of two strains (10 4 CFU of each). Mice were infected with S. Typhi WT and isogenic mutants: either sptP::kan R as a control for kanamycin resistance inserted into a nonfunctional gene (21), a ΔssaV::kan R strain, or a ΔinvA ΔssaV::kan R strain. Strikingly, at 5 days p.i., the T3SS-null S. Typhi strain was significantly outcompeted by the WT strain in both the spleens and the livers (Fig. 4B). In contrast, the ΔsptP mutant S. Typhi strain was not outcompeted by the WT strain in the spleen or the liver (Fig. 4B). Although deletion of T3SS-2 alone had a defect relative to the control strain in the spleen, it was not significantly outcompeted by WT in the liver, and it was not outcompeted as significantly as the T3SS-null mutant in either organ. Previous publications using this humanized mouse model of typhoid infection have highlighted the heterogeneity in the levels of S. Typhi recovered from individual mice (13,30). However, we recovered greater than 10 3 WT S. Typhi CFUs per gram of tissue in 11 of the 12 spleens (Fig. S4A), indicating that the spleens contained an amount of WT bacteria well above the limit of detection. Liver burdens were more heterogenous than spleen burdens (Fig. S4B); however, the level of WT bacteria was 10-fold above the limit of detection, at 10 2 CFU per gram of tissue. Together, these data demonstrate that S. Typhi uses two T3SSs to colonize systemic sites in a humanized mouse model of typhoid fever.

DISCUSSION
Many studies have compared the genomes of the host generalist S. Typhimurium with S. Typhi and have speculated that the evolution of S. Typhi likely involved both gain of function and loss of function as it evolved to selectively infect humans (8). Although systemic murine infection with S. Typhimurium models human disease, fundamental genetic differences between typhoidal and nontyphoidal Salmonella serovars limit the ability of this model to identify human-specific virulence mechanisms. In this study, we used human-derived macrophages and a humanized mouse model to study S. Typhi virulence factors that contribute to intramacrophage replication (13,30). We show that both T3SS-1 and T3SS-2 promote S. Typhi replication in human macrophages and further that both T3SSs contribute to S. Typhi colonization of the spleen and the liver in a humanized mouse model of typhoid fever. There is clinical interest in determining if T3SS-null S. Typhi may be useful as an attenuated vaccine strain (31). This study provides evidence that S. Typhi mutants lacking both T3SS-1 and T3SS-2 are attenuated for intramacrophage replication. A previously published clinical study of an S. Typhi ΔaroA ΔssaV mutant, which lacks a functional T3SS-2, was administered to human volunteers, with infrequent adverse events, indicating that S. Typhi ΔaroA ΔssaV is likely attenuated within the specific patient population studied (31). The data presented here show T3SS-1 and T3SS-2 deletion attenuates the laboratory S. Typhi strain Ty2, which is closely related to the only currently licensed live, attenuated typhoid vaccine strain Ty21a (32). However, more rigorous clinical studies appropriately accounting for the diversity in genetics and lifestyle of all effected populations are necessary (3). Previous research has demonstrated that while T3SS-2 is required for S. Typhimurium replication in murine macrophages, T3SS-2 is not required for S. Typhi replication in human macrophages (12,33). Here, we used a fluorescence dilution assay that directly assesses bacterial replication in human macrophages ( Fig. 1B and C). We uncovered that T3SS-2 partially contributes to S. Typhi replication in human macrophages. We further showed that disabling both T3SS-1 and T3SS-2 renders S. Typhi unable to replicate in human macrophages.
To gain further insights into the potential roles of S. Typhi T3SS effectors, we screened 25 strains lacking individual T3SS effectors for replication defects in which mutant strains were simultaneously analyzed in a time-lapse fluorescence dilution assay. Many of the single mutants did not have significant replication defects, which would be expected for the reported pseudogenes (sptP, cigR, sopA, srfA, sopE2, slrP, sopD2, and sseJ). We also found that the S. Typhi ΔsteC mutant had the highest replication index at 24 h.p.i., although it didn't quite reach statistical significance ( Fig. 2A). This finding agrees with previously published data indicating that SteC may function to restrain intracellular growth during S. Typhimurium infection (34). Future studies are warranted to examine typhoidal SteC function in human macrophages specifically. Our screen also identified two T3SS effectors, PipB2 and SifA, that contribute significantly to replication in human macrophages ( Fig. 3A through C). We confirmed that the effector proteins PipB2 and SifA are important for intramacrophage growth of S. Typhi by plating for viable bacteria (Fig. 2B). We also show that both T3SS-1 and T3SS-2 contribute to translocation of these important effectors into the cytosol of human macrophages and that SifA translocation is T3SS-1 dependent at 8 h and that T3SS-2 contributes to translocation at 16 h.p.i. (Fig. 3B and C). Previous studies in S. Typhimurium have shown that PipB2 is translocated through both T3SS-1 and T3SS-2 (27). Similarly, our results indicate that PipB2 can be translocated by S. Typhi's T3SS-1 and T3SS-2 ( Fig.  3B and C). In contrast, others have shown that SifA is translocated through T3SS-2 in S. Typhimurium (35). Interestingly, our results show that in S. Typhi, SifA can be translocated into the macrophage cytosol by T3SS-1 at 8 h.p.i. and both T3SS-1 and T3SS-2 at 16 h.p.i. (Fig. 3B and C). Consistent with this finding, there are fewer replicating S. Typhi ΔsifA mutant bacteria at 8 h.p.i. compared with the S. Typhi T3SS-2 ΔssaV strain (Fig. 2C). Finally, the kinetics of intramacrophage replication of the S. Typhi ΔsifA and ΔpipB2 mutants were similar, indicating that these effectors play important roles in establishing and maintaining the SCV during S. Typhi infections of human macrophages (Fig. 2C). In S. Typhimurium, PipB2 and SifA are involved in SCV membrane dynamics (36). For example, both PipB2 and SifA recruit kinesin-1 to the SCV through a direct interaction or via Salmonella-kinesin interacting protein (SKIP), respectively (37)(38)(39). SifA also recruits Rab9 and is thought to suppress lysosome functions (40,41). Future studies focused on effector-host protein-protein interactions during S. Typhi infections of human macro phages will be important for increased knowledge of host-pathogen interactions during typhoid fever. Finally, Figueira et al. demonstrated in S. Typhimurium that poly-effector mutant strains are more severely attenuated in replication compared with single mutant strains, suggesting some redundancy in effector functions (11). For example, Research Article mBio the effectors SseF and SseG directly interact and cooperate during S. Typhimurium infection (25,42,43). This may partially explain why S. Typhi ΔsseF and ΔsseG mutant strains did not have significant intramacrophage replication defects. However, these two effectors each individually significantly contribute to S. Typhimurium intramacrophage replication (11). To probe specific differences in effector functions between serovars, a direct comparison of serovar replication within the same host cell context, including single-and double-effector knockouts, and heterologous expression of homologous effectors in different serovars, will be necessary. For example, a thorough study of the genetic and molecular reasons for SptP functional differences between S. Typhi and S. Typhimurium homologs revealed that typhoidal SptP loss of function is caused by a mutation in the chaperone-binding domain (21). It is possible that other typhoidal Salmonella effectors also have mutations that cause either divergent functionality or loss of function compared with their S. Typhimurium homolog. On the other hand, S. Typhi has a significantly reduced T3SS effector repertoire. For example, SopD2, which is pseudogenized in S. Typhi, modulates the trafficking of the SCV and has recently been shown to cooperate with SteA and PipB2 (11,25,44,45). Therefore, a possible explanation for why some individual effectors, such as PipB2, play a more significant role in intramacrophage replication of S. Typhi compared with S. Typhimurium is that the functionally redundant effectors are missing in S. Typhi (11). Future studies includ ing analysis of poly-effector mutants of S. Typhi will be important to increase our understanding of T3SS effector functional redundancies and possibly explain why PipB2 significantly contributes to S. Typhi intramacrophage replication.
Here, we show that S. Typhi intramacrophage replication is dependent on T3SSs activity. However, previous studies have shown that T3SSs also translocate flagellin and structural components of the T3SS into the cytosol which are pathogen-associated molecular patterns that are recognized by pattern recognition receptors (PRRs) (18,19,29,46). Indeed, human macrophages produce a cytosolic PRR known as hNAIP the NLR (nucleotide-binding domain, leucine-rich repeat-containing) family, apoptosis inhibitory protein that recognizes flagellin (19,46) and the T3SS-1 inner rod protein and the T3SS-1 and T3SS-2 needle proteins (18), triggering a signaling cascade that results in a pyroptotic cell death. However, we do not see high levels of cell death in THP-1 macrophages when infected with WT or mutant S. Typhi strains in this study (Fig. S1D). Future studies are, therefore, warranted to examine host cell detection of intramacrophage S. Typhi replication and whether S. Typhi has mechanisms to effectively evade detection by the macrophage.
Importantly, we demonstrated the pathophysiological relevance of our in vitro studies by showing that the T3SSs are critical for S. Typhi colonization in systemic tissues of humanized mice. Functional redundancy in S. Typhi's T3SSs likely explains why previous screens for S. Typhi genes required for intramacrophage survival and systemic coloniza tion of hu-SRC-SCID mice failed to identify T3SS-1 and T3SS-2 as crucial virulence factors (13,47). Overall, the data presented here demonstrate that both T3SS-1 and T3SS-2 are critical virulence factors enabling S. Typhi replication in systemic tissues and specifically within human macrophages.

Mouse strains and husbandry
Experiments involving animals were performed in accordance with NIH guidelines, the Animal Welfare Act, and US federal law. All animal experiments were approved by the Stanford University Administrative Panel on Laboratory Animal Care and overseen by the Institutional Animal Care and Use Committee under Protocol ID 12826. Animals were housed in a centralized research animal facility accredited by the Association of Assessment and Accreditation of Laboratory Animal Care International. Female hu-SRC-SCID mice engrafted with CD34+ human umbilical cord blood stem cells were obtained from Jackson Laboratories. NOD-scid IL2Rg null mice engrafted with human CD34 + umbilical cord blood stem cells, 28-37 wk post-engraftment (aged) (Jackson, Strain #:005557), were housed under specific pathogen-free conditions in filter-top cages that were changed weekly. Sterile water and food were provided ad libitum. Mice were given at least 1 wk to acclimate prior to experimentation.

Cell culture and differentiation
THP-1 cells were obtained from ATCC and passaged a maximum of 10 times in all experiments. THP-1s were routinely cultured in Roswell Park Memorial Institute (RPMI) 1640 Medium with 10% heat-inactivated fetal bovine serum (FBS) and 2 mM Gluta MAX supplement. THP-1 cells were differentiated into macrophage-like adherent cells according to published protocols (17). Briefly, THP-1s were treated with 100 nM PMA for 48 h, then media were replaced, and THP-1s were infected 3 days after differentiation. To better adhere differentiated THP-1s, before seeding plates for differentiation, the surface of TC-treated plastic was coated with human plasma-derived fibronectin according to the manufacturer's instructions.
For primary human-derived macrophages, four individual TrimaAccel leukocyte reduction system (LRS) chambers recovered after plateletpheresis containing white blood cell concentrate were purchased from Stanford Blood Center. Each chamber contained samples from an individual platelet donor. Samples were entirely deidentified, products collected and sold by Stanford Blood Center for in vitro investigational use are not routinely tested for infectious disease markers and do not require institutional review board (IRB) approval. Peripheral blood mononuclear cells were isolated by gradient centrifugation using Ficoll-Paque PLUS, and adherent mononuclear cells were differentiated with 30 ng/mL hM-CSF in RPMI medium supplemented with GlutaMAX and 10% heat-inactivated FBS for 6-7 days prior to infection as previously described (48).

Bacterial strain construction
Genes in the Ty2 genome were targeted for deletion following an adjusted lambda red mutagenesis protocol (49). Briefly, primers provided in Table 1 were used to amplify the kanamycin resistance cassette of pKD4, and PCR products were purified and concen trated to 500 ng/µL using a MinElute PCR product purification kit (Qiagen). Strains containing pKD46 were grown at 30°C, 200 rpm until midlog phase, then L-arabinose was added to a final concentration of 50 mM, and incubated for an additional 1-2 h. Bacteria were pelleted and washed in ice-cold autoclaved DI water, concentrated 200×, and 1-5 µg of purified PCR product was added to each electroporation cuvette. Bacteria were electroporated and then recovered in SOC broth at 30°C for 3 h static. Cultures were plated on antibiotic selection plates and grown overnight at 42°C. Colonies were PCR verified for kanamycin resistance insertion and endogenous gene disruption. Then, pCP20 was electroporated into kanamycin-resistant strains, and removal of the kanamycin resistance cassette was PCR confirmed. Removal of both the pKD46 and pCP20 plasmids was confirmed by a loss of carbenicillin resistance.
pFCcGi was electroporated into strains and selected with 50 µg/mL of carbenicil lin. When indicated, 0.4% L-arabinose was added to pFCcGi+ cultures to induce GFP expression.
For chromosomal insertion of constitutive mCherry, the plasmid pMH21-mCherry was first constructed. The kanamycin resistance cassette from pKD4 was amplified using FRTkanFRT_fwd and FRTkanFRT_rev, and the backbone of pFPV-mCherry, designed for   constitutive mCherry expression in Salmonella, was amplified using mCherryBB-fwd and mCherryBB-rev. Linear PCR products were digested with DpnI, purified, and ligated with NEBuilder HiFi DNA Assembly Master Mix according to the manufacturer's instructions to create pMH21-mCherry. Then, pMH_sptP-ins_fwd and pMH_sptP-ins_rev were used to amplify the kanamycin resistance cassette, the PrpsM promoter, and mCherry gene from pMH21-mCherry, adding homology with the sptP locus. Mutagenesis proceeded as previously stated using the lambda red method, including using pCP20 to then remove the kanamycin resistance cassette. Each bacterium, thus, encodes one constitutive mCherry locus in place of the sptP gene.

β-Lactamase translocation constructs
The pFT-BlaM vector and pGST-BlaM were provided by the Baumler lab (28). To construct the sifA-FT and pipB2-FT plasmids, the pFLAG-TEM1 vector was digested with NdeI and XhoI, and fragments of Ty2 gDNA were amplified using primers listed in the Table 1. All purified plasmids were sequenced by Plasmidsaurus (Eugene, Oregon). Plasmids were electroporated into Ty2 sptP::mCherry.

THP-1 infections
Bacteria were grown in LB+Aromix broth with 0.4% arabinose at 25°C, 200 rpm (48) until reaching an OD of ~1.0. Bacterial pellets were spun down at 3,500 rpm and then washed and diluted with PBS to the appropriate colony-forming units per milliliter to reach a final multiplicity of infection of 1:20 and opsonized with 25% pooled human serum for 30 min at room temperature. Bacteria were then resuspended using a sterile 25G needle and into the infection medium (RPMI + 10% heat-inactivated FBS + 2 mM GlutaMAX).
To synchronize uptake, plates were spun at 250 × g for 5 min and then incubated at

hMDM infections
hMDM infections were performed similar to THP-1 infections, except infections were performed with an MOI of 10:1 and for 30 min, followed by 1.5 h of 100 μg/mL gent treatment.

Flow cytometry analysis of intramacrophage replication
At indicated timepoints, intramacrophage replication of S. Typhi was quantified by flow cytometry according to previously published protocols (11,20). Briefly, macrophages were lysed by incubation with 1% Triton-X at 37°C for 10 min. Samples were pelleted at 13,000 × g for 1 min and then fixed in 4% paraformaldehyde for 15 min at room temperature. Samples were then washed with PBS once and resuspended in fluorescentactivated cell sorting (FACS) buffer. A total of 100,000 events were collected per sample on a BD LSRII flow cytometer and analyzed using FloJo software. Bacteria were gated on size relative to buffer-only control and mCherry+ signal relative to nonfluorescent bacteria. The geometric mean of the GFP signal for mCherry+ events was then calculated for each sample. Technical replicates correspond to separate wells infected at the same time, and biological replicates correspond to infections performed on separate days.

Quantifying CFU/well from infected macrophages
At indicated timepoints, macrophages were lysed with 1% Triton-X at 37°C for 10 min. Samples were pelleted at 13,000 × g for 1 min, and media were aspirated until 100 µL remained. Samples were resuspended, serially diluted in PBS, and plated on LB + aromix plates. After overnight incubation at 30°C, CFUs were counted.

Time-lapse microscopy
An Incucyte S3 Live-Cell Analysis Instrument was used for time-lapse imaging. At indicated time intervals, Incucyte Base Software was used to collect 20× images at a set location, and three to four images per well were routinely collected. To quantify fluorescence, the Incucyte Cell-By-Cell Analysis Software Module segmented all images into extracellular and intracellular area. To indicate replication, the ratio of intracellular mCherry intensity was divided by the intracellular GFP integrated intensity and plotted over time. For each biological replicate, three wells, three to four images each well per timepoint, were averaged together to indicate one replicate. Biological replicates indicate experiments performed on separate days. To compare biological replicates, values from uninfected wells were subtracted, then values were normalized to make the greatest value the WT culture achieved in each experiment equal to one and the lowest value collected in each experiment equal to zero; this value is named "replication index" (11). Individual images from the Incucyte software were exported and compiled into videos using ImageJ.

Optical density growth curves
Cultures were grown in LB + aromix overnight at 37°C, 200 rpm shaking and then diluted 1:50 and grown in LB + aromix for 3 h to reach log phase. Aliquots of the cultures were then pelleted and washed three times in PBS, then optical density (OD600) was measured, and strains were back-diluted to a starting OD of 0.02 in 96-well plates. Plates were incubated at 37°C with shaking between reads on a Synergy HTX (BioTek) plate reader, and optical density was measured every 15 min. Biological replicates indicate experiments performed on separate days.

Confocal microscopy
Cover glass was coated with human fibronectin according to the manufacturer's instructions. Macrophages were seeded, differentiated, and infected on fibronectincoated glass. At indicated timepoints, media were removed, and infections were fixed with periodate-lysine-paraformaldehyde (PLP) fixative (2% paraformaldehyde in 75 mM NaPO 4 buffer pH 7.4, 2.5 mM NaCl) for 15 min at room temperature. Slides were gently washed after fixing and between stains with warm PBS supplemented with 9 mM CaCl 2 and 5 mM MgCl 2 . After permeabilization with 1% saponin and 3% BSA, slides were first stained with 1:200 Mouse monoclonal LAMP-1 (DSHB) and 1:1,000 Chicken anti-Salmonella (Aves Labs) for 1 h at room temperature, followed by incubation in secon dary antibodies 1:500 goat antichicken Alexa594 (Invitrogen), 1:500 donkey antimouse Alexa488 (ThermoFischer), and 1:100 Alexa660 Phalloidin (ThermoFischer) for 1 h at room temperature. Cover glass was then mounted on slides with ProLong Glass Antifade with NucBlue stain (ThermoFischer). Slides were cured for 24 h in the dark at room temperature and stored at −20°C until imaging on a Zeiss LSM 700 confocal microscope with the ZEN 2010 software. Images were processed using ImageJ. For imaging of β-lactamase translocation into hMDMs, cells were seeded and infected on poly-L-lysinecoated glass wells. At 8 h.p.i., cells were loaded with CCF4-AM dye according to the manufacturer's instructions for 1.5 h in the dark. Cells were then fixed with 3.2% PFA for 20 min in the dark. Cells were then washed once with PBS, and PBS was used to keep the cells hydrated during imaging. Cells were imaged with a Zeiss LSM 700 confocal microscope with the ZEN 2010 software with a blue diode 405 nm laser for excitation and with detection filters set at 410-450 nm for coumarin and 493-550 nm for fluorescein.

Flow cytometry analysis of effector translocation
At 8 h.p.i., THP-1 macrophages were acclimated to room temperature and loaded with CCF4-AM dye according to the manufacturer's instructions for 1.5 h in the dark. Cells were then fixed with 3.2% PFA for 20 min in the dark. Cells were washed and resuspen ded in FACS buffer using a cell scraper and analyzed on an LSRII analyzer as previously described (29). Cells were gated on size, singlets, mCherry (infected)+, and green (dye loaded)+, then analyzed for percentage of population positive for blue signal.

Humanized mouse infections
NSG mice (NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ, 005557) engrafted with CD34+ hemato poietic stem cells derived from umbilical cord blood were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). Successful humanization of each mouse is quantified by the supplier from mouse peripheral blood via flow cytometry using anti-hu-CD45+ and antimurine CD45+, approximately 2-month post-engraftment. Mice were shipped at 31-week post-engraftment, allowing increased tissue engraftment of myeloid cells (unpublished data from the Jackson Laboratory). After 7 days of acclimation, mice were injected IP with a 1:1 ratio of S. Typhi strains; either 2 × 105 CFUs for 2 days or 2 × 104 CFUs of S. Typhi for 5 days. The infected mice were closely monitored for signs of illness, and moribund animals were euthanized. At 2 and 5 days, entire livers and spleens were harvested and homogenized in PBS and plated for CFU on LB + Aromix and LB + Aromix + Kanamycin plates to quantify the ratio of WT to mutant bacteria in each organ. The CI was calculated as a ratio of (mutant/WT)output/(mutant/WT)input.

Quantification and statistical analysis
The statistical significance of all flow cytometry data, CFU counts, time-lapse replication measurements, LB growth curves, and cell death assays were determined by two-way ANOVA followed by Tukey's honestly significant difference (Tukey's multiple compari sons) to calculate multiple pairwise comparisons in Prism v. 8