HilD induces expression of a novel Salmonella Typhimurium invasion factor, YobH, through a regulatory cascade involving SprB

HilD is an AraC-like transcriptional regulator encoded in the Salmonella pathogenicity island 1 (SPI-1), which actives transcription of many genes within and outside SPI-1 that are mainly required for invasion of Salmonella into host cells. HilD controls expression of target genes directly or by acting through distinct regulators; three different regulatory cascades headed by HilD have been described to date. Here, by analyzing the effect of HilD on the yobH gene in Salmonella enterica serovar Typhimurium (S. Typhimurium), we further define an additional regulatory cascade mediated by HilD, which was revealed by previous genome-wide analyses. In this regulatory cascade, HilD acts through SprB, a LuxR-like regulator encoded in SPI-1, to induce expression of virulence genes. Our data show that HilD induces expression of sprB by directly counteracting H-NS-mediated repression on the promoter region upstream of this gene. Then, SprB directly activates expression of several genes including yobH, slrP and ugtL. Interestingly, we found that YobH, a protein of only 79 amino acids, is required for invasion of S. Typhimurium into HeLa cells and mouse macrophages. Thus, our results reveal a novel S. Typhimurium invasion factor and provide more evidence supporting the HilD-SprB regulatory cascade.


Results
HilD positively controls the expression of yobH (SL1344_1770). Previous RNA-sequencing (RNA-seq) analysis indicates that the HilD transcriptional regulator positively controls the expression of the S. Typhimurium SL1344_1770 gene 28 , which is located outside SPI-1 and codes for a hypothetical protein of 79 amino acids. Additional RNA-seq and co-expression analyses also support that HilD is involved in the expression of SL1344_1770 39,45 . Orthologs of SL1344_1770, which show high sequence identity and a conserved genomic context, are denominated yobH in Escherichia coli and several other bacteria; thus, we kept the name of yobH for SL1344_1770.
To confirm the regulation of yobH by HilD, a transcriptional fusion of the intergenic region upstream of yobH to the cat (chloramphenicol acetyl transferase) reporter gene was constructed in the pKK232-8 plasmid. Specific activity from this fusion was quantified in the wild type (WT) S. Typhimurium strain SL1344 and its derivative ∆hilD mutant, grown in nutrient-rich lysogeny broth (LB) at 37 °C, conditions that induce the expression of genes regulated by HilD 25,26,39 . The activity of the yobH-cat fusion showed a 3-fold reduction in the ∆hilD mutant, compared with its expression in the WT strain; in addition, the expression of HilD from the pK6-HilD plasmid increased around 5-fold the activity of this fusion in the ∆hilD mutant (Fig. 1A). To investigate if yobH indeed codes for a protein and to further confirm the positive regulation of yobH by HilD, we tested the expression of the YobH-FLAG putative protein (YobH tagged with a 3XFLAG epitope) in the WT S. Typhimurium strain and its derivative ∆hilD mutant. A specific signal for YobH-FLAG was detected in the WT strain, with the expected size for this protein (Fig. 1B). The amount of YobH-FLAG was almost abolished in the ∆hilD mutant; as expected, it was restored at WT levels by the expression of HilD from the pK6-HilD plasmid (Fig. 1B). Taken together, these results show that HilD induces the expression of yobH.
YobH is involved in the S. typhimurium invasion of host cells. HilD positively regulates expression of numerous genes mainly required for the invasion of Salmonella into host cells 2,28,39,45 . For instance, recently we found that HilD controls expression of the grhD1 invasion gene, which is located outside SPI-1 40 . To determine whether YobH is required for this Salmonella virulence phenotype, we evaluated the invasion ability of the WT S. Typhimurium strain and its isogenic ∆yobH mutant in HeLa cells and RAW264.7 mouse macrophages. Additionally, we constructed a complemented ∆yobH mutant (∆yobH + yobH-FLAG-kan), by inserting yobH into the chromosome of the ∆yobH mutant as described in Fig. S1, which was also assessed in the invasion assays. Furthermore, the ∆hilD and ∆ssrB mutants were also tested in these assays as positive and negative controls, respectively; SsrB is a transcriptional regulator that is required for Salmonella intracellular replication but not for invasion of host cells 2,40 . The ∆yobH mutant showed a ∼3-fold decrease in the invasion of both HeLa cells and RAW264.7 macrophages with respect to the WT and the complemented ∆yobH + yobH-FLAG-kan strains ( Fig. 2A,B). As expected, the ∆hilD mutant was unable to invade the HeLa cells and RAW264.7 macrophages, whereas the ∆ssrB mutant invaded these cells at similar levels to those showed by the WT and the complemented www.nature.com/scientificreports www.nature.com/scientificreports/ ∆yobH + yobH-FLAG-kan strains ( Fig. 2A,B). Important to note, the number of bacteria present in the starting inoculums used in the invasion assays showed a variation of only 18% between the different strains tested (Fig. S2). Together, these results indicate that YobH is a novel invasion factor of S. Typhimurium.
HilD controls expression of yobH through SprB. We sought to determine if HilD regulates the yobH virulence gene directly or indirectly. For this, we performed electrophoretic mobility shift assays (EMSAs) by using affinity-purified maltose-binding protein (MBP)-HilD and a DNA fragment spanning the intergenic region upstream of yobH. DNA fragments carrying the regulatory region of hilA or sopB were also assessed in these assays as positive and negative controls, respectively; HilD binds to hilA but not to sopB 25 . As show in Fig. S3A, MBP-HilD did not shift the yobH fragment or that of sopB, even at the highest protein concentration tested (1 µM). In contrast, MBP-HilD shifted the positive control, hilA, at concentrations from 0.1 to 1 µM (Fig. S3B). These assays indicate that HilD does not interact with the regulatory region of yobH; alternatively, HilD could require an additional factor to bind to yobH.
To investigate whether HilD requires another S. Typhimurium regulator to induce the expression of yobH, we monitored the activity of the yobH-cat fusion in the WT E. coli MC4100 strain, which lacks HilD and the other Expression of YobH-FLAG in WT S. Typhimurium SL1344 strain and its isogenic ∆hilD mutant containing the pMPM-K6Ω vector or the pK6-HilD plasmid, was analyzed by Western blotting by using an anti-FLAG monoclonal antibody. GroEL was detected as a loading control with an anti-GroEL polyclonal antibody. Blots were cropped from different parts of the same gel. CAT specific activity and YobH-FLAG expression were determined from samples of bacterial cultures grown for 9 h in LB at 37 °C.

Figure 2.
YobH is involved in the S. Typhimurium invasion of HeLa cells and macrophages. Epithelial HeLa cells (A) and murine RAW 264.7 macrophages (B) were infected with the WT S. Typhimurium SL1344 strain and its isogenic ∆ssrB, ∆hilD, ∆yobH and ∆yobH + yobH-FLAG-kan (∆yobH complemented) mutants. Invasion was quantified by enumerating the intracellular CFUs at 1 h post-infection, using a gentamicin protection assay. Means and standard deviations from three independent experiments performed in duplicate are shown. Statistically different values are indicated (***p < 0.001; ****p < 0.0001).
Salmonella-specific regulators, in the presence of the pK6-HilD plasmid expressing HilD or in the presence of the pMPM-K6Ω vector. As a positive control, activity of the hilA-cat fusion was also tested; hilA is directly regulated by HilD 47,48 . As expected, the yobH-cat and hilA-cat fusions showed low or undetectable expression levels in E. coli (Fig. S3C,D). The activity of hilA-cat, but not that of yobH-cat, was induced by HilD in E. coli (Fig. S3C,D), indicating that an additional factor, present in S. Typhimurium SL1344 but not in E. coli MC4100, is required for the HilD-mediated expression of yobH.
Several studies have shown that HilD induces expression of a high number of virulence genes through distinct regulatory cascades involving the HilA, InvF, HilC, RtsA, SsrA/SsrB and FlhDC transcriptional regulators 2,25,28,29,32,33 . To determine if any of these regulators are required for the HilD-mediated expression of yobH, activity of the yobH-cat fusion was measured in the WT S. Typhimurium strain and its isogenic ∆hilA, ∆invF, ∆hilC, ∆rtsA, ∆ssrB and ∆flhDC mutants. As positive controls, the ∆SPI-1 and ∆hilD mutants were also tested; SPI-1 encodes HilD, HilA, InvF and HilC 2 . Activity of the yobH-cat fusion was affected in the ∆SPI-1 and ∆hilD mutants, but not in the ∆hilA, ∆invF, ∆hilC, ∆rtsA, ∆ssrB and ∆flhDC mutants, with respect to the WT strain ( Fig. 3), suggesting that HilD induces expression of yobH through a regulatory cascade different to those well characterized before this study.
Previous RNA-seq analyses indicate that HilD induces expression of two additional Salmonella-specific transcriptional regulators, SprB and SinR 28,45 , encoded in SPI-1 and SPI-6, respectively. Moreover, we previously confirmed that HilD directly induces expression of sinR 39 . SprB has been involved in the expression of several S. Typhimurium virulence genes 28,46 , whereas SinR remains uncharacterized. To investigate whether SprB and/ or SinR are involved in the HilD-mediated expression of yobH, we monitored the activity of the yobH-cat fusion in the ∆sprB and ∆sinR mutants. Surprisingly, the activity of this fusion was reduced in the ∆sprB mutant, as in the ∆SPI-1 and ∆hilD mutants, whereas it was not affected in the ∆sinR mutant ( Fig. 3), suggesting that SprB is required for the expression of yobH. Expression of SprB from the pK6-SprB plasmid, under an arabinose inducible promoter, restored the activity of yobH-cat in both the ∆sprB and the ∆hilD mutants (Fig. 4A). In contrast, expression of HilD from the pK6-HilD plasmid induced the activity of yobH-cat in the ∆hilD mutant ( Fig. 1A), but not in the ∆sprB mutant (Fig. 4A). Similarly, SprB restored the expression of YobH-FLAG in the ∆hilD mutant, whereas HilD was unable to induce the expression of YobH-FLAG in the ∆sprB mutant (Fig. 4B). Altogether, these results support that SprB acts downstream of HilD for the expression of yobH.
Results from a chromatin immunoprecipitation sequencing (ChIP-seq) analysis revealed that SprB binds to the regulatory region of yobH in vivo 28 . We sought to confirm the SprB binding on yobH by EMSAs; however, we were unable to purify the 6XHis-tagged SprB protein, probably due to its high insolubility. Alternatively, we investigated if SprB requires any other Salmonella-specific regulator to induce the expression of yobH. For this, the activity of the yobH-cat fusion was tested in the WT S. Typhimurium strain and in the WT E. coli MC4100 strain carrying the pMPM-K6Ω vector or the pK6-SprB plasmid. Activity of a cat transcriptional fusion of sirA, a gene expected to be not controlled by SprB, was also tested as negative control; an ortholog of sirA (uvrY) is present in E. coli K-12 49,50 . The presence of SprB induced the activity of yobH-cat in E. coli to levels similar to those reached by this fusion in the WT S. Typhimurium strain (Fig. 5A); in contrast, SprB did not affect the activity of the sirA-cat fusion (Fig. 5B). These results are in line with the notion that SprB directly activates expression of yobH.
The results described above strongly suggest that HilD positively regulates the expression of sprB. To confirm this, we quantified the expression of sprB in the WT S. Typhimurium strain and its isogenic ∆hilD mutant. Expression of sprB seems to be controlled by both the regulatory region upstream of hilC, which generates a hilC-sprB transcript, and that located upstream of sprB 28 . HilD regulation on the promoter upstream of hilC has been extensively shown in previous studies 28,30,38,48,51 . Thus, we evaluated the effect of HilD on the regulatory region upstream of sprB by constructing and analyzing a sprB-cat transcriptional fusion carrying this region. As shown in Fig. 6A, activity of the sprB-cat fusion was 2-fold reduced in the ∆hilD mutant, compared with its activity in the WT strain; furthermore, expression of HilD from the pK6-HilD plasmid increased 3-fold the activity of this fusion in the ∆hilD mutant, indicating that HilD induces expression of sprB by also acting on the regulatory region upstream of this gene. www.nature.com/scientificreports www.nature.com/scientificreports/ To investigate whether HilD regulates expression of sprB (by acting on the regulatory region upstream of this gene) directly or through an additional Salmonella-specific factor, we determined the activity of the sprB-cat fusion in the WT E. coli MC4100 strain carrying the pK6-HilD plasmid or the pMPM-K6Ω vector. As expected, activity of the sprB-cat fusion was 3-fold lower in the E. coli strain than in the WT S. Typhimurium strain   6B). Expression of HilD from pK6-HilD induced 3-fold the activity of sprB-cat in the E. coli strain (Fig. 6B), supporting that HilD induces expression of sprB directly. In agreement to these results, EMSAs revealed that purified MBP-HilD binds to the regulatory region upstream of sprB, from a concentration of 0.5 µM, but it does not bind to the regulatory region upstream of ppK, used as a negative control in these assays (Fig. 6C). Furthermore, previous results from ChIP-seq analyses indicate that HilD binds to the intergenic region upstream of sprB in vivo 28,38 .
Collectively, these results indicate that HilD positively regulates the expression of yobH through SprB.

HilD counteracts H-nS-mediated repression on sprB. HilD induces expression of target genes mainly
by counteracting H-NS-mediated repression on the respective promoters 2,36 . To know whether HilD induces expression of sprB by a similar way, we analyzed if inactivation of H-NS leads to HilD-independent expression of this gene. Since a Salmonella ∆hns mutant exhibits a severe growth defect 52,53 , we inactivated H-NS activity by the overexpression of the H-NS G113D dominant negative mutant, which is affected in DNA binding activity but still forms heterodimers with the WT H-NS monomers 54 . For this purpose, activity of the sprB-cat fusion was quantified in the WT S. Typhimurium strain and its isogenic ∆hilD mutant containing the pT6-HNS-G113D or pT6-HNS-WT plasmids, which express H-NS G113D and WT H-NS, respectively, or containing the empty vector pMPM-T6Ω. Expression of H-NS G113D , but not WT H-NS, increased the activity of sprB-cat in the ∆hilD mutant, at similar levels to those reached by this fusion in the WT strain (Fig. 7A). Consistently, activity of sprB-cat was also induced in an E. coli ∆hns mutant, compared with the WT E. coli strain (Fig. S4). Furthermore, EMSAs revealed that purified H-NS-FLAG-His (H-NS-FH) protein binds to the regulatory region upstream of sprB, from a concentration of 0.45 µM, but it does not bind to the regulatory region upstream of ppK, used as a negative control in these assays, even at the highest protein concentration tested (0.7 µM) (Fig. 7B). Previous genome-wide binding studies indicate that H-NS interacts with the region upstream of sprB in vivo 52,53 . In contrast to the observed for sprB, the activity of the yobH-cat fusion was not increased in the E. coli ∆hns mutant and H-NS-FH www.nature.com/scientificreports www.nature.com/scientificreports/ did not bind to the regulatory region upstream of yobH (Fig. S5). These results show that H-NS directly represses expression of sprB, but not of yobH, and that when the activity of H-NS is inactivated, or when H-NS is absent, expression of sprB becomes independent of HilD, which supports that HilD acts on this gene as an anti-H-NS factor.
To determine whether HilD indeed displaces H-NS from sprB, we performed competitive EMSAs. A DNA fragment carrying the regulatory region of sprB was first incubated with a constant concentration of H-NS-FH (0.6 µM) and then increasing amounts of MBP-HilD (0.2, 0.4, 0.6, 0.8 and 1 µM) were added. Binding reactions containing only H-NS-FH or MBP-HilD were also tested. The DNA-protein complexes were detected by staining the DNA fragments with ethidium bromide; additionally, the presence of H-NS-FH on these complexes was detected by Western blot with anti-FLAG antibodies. As shown in Fig. 7C, the DNA-H-NS complex was shifted by the presence of MBP-HilD to a slower-migrating complex similar to that formed only by MBP-HilD (upper panel); furthermore, the immunoblots showed that the presence of MBP-HilD decreased the amount of H-NS-FH bound to the tested DNA fragment (lower panel), which indicates that HilD is able to remove H-NS from sprB.
Altogether, these results demonstrate that HilD induces expression of sprB by antagonizing H-NS-mediated repression on this gene. the HilD-SprB regulatory cascade controls expression of the slrP and ugtL virulence genes. Previous RNA-seq analyses indicate that HilD and SprB positively controls expression of several other genes in common, in addition to yobH, including slrP and ugtL 28 , which have been involved in Salmonella virulence [55][56][57][58][59] . To further define if HilD and SprB also act in a cascade fashion on slrP and ugtL, we constructed and analyzed cat transcriptional fusions carrying the regulatory region of the slrP or ugtL genes. Activity of the slrP-cat and ugtL-cat fusions was quantified in the WT S. Typhimurium strain and its derivative ∆hilD mutant containing the pK6-SprB plasmid or the pMPM-K6Ω vector. As a negative control, an invF-cat transcriptional www.nature.com/scientificreports www.nature.com/scientificreports/ fusion was also assessed; HilD induces expression of invF through HilA 27,60,61 . The three fusions tested showed a decreased activity in the ∆hilD mutant, with respect to their activity in the WT strain ( Fig. 8A-C). Expression of SprB from pK6-SprB induced activity of the slrP-cat and ugtL-cat fusions, but not that of the invF-cat fusion, in the ∆hilD mutant ( Fig. 8A-C), supporting that HilD controls the expression of slrP and ugtL through SprB, which is in agreement with data from ChIP-seq analyses showing SprB binding, but not HilD binding, on the regulatory regions of slrP and ugtL 28 .
Thus, our results, together with previous studies, indicate that the regulatory cascade formed by HilD and SprB controls expression of a subset of Salmonella virulence genes, including yobH, slrP and ugtL.

Discussion
Acquisition of SPI-1 was a pivotal event for the evolution of Salmonella pathogenicity, not only by the virulence factors encoded in this island, which provide ability to invade host cells, but also by the additional factors for invasion encoded outside SPI-1 that have been recruited through the control of their expression by the SPI-1 regulator HilD 2,28 .
In this study, we identify a novel invasion factor, YobH, whose expression is controlled by HilD. Our results demonstrate that YobH is required for the S. Typhimurium invasion of HeLa cells and mouse macrophages. In agreement with these results, a previous analysis by transposon-directed insertion-site sequencing (TraDIS) supports that YobH plays a role in the intestinal colonization of S. Typhimurium in chicks and cows, but not in the systemic infection in the mouse model 56 . The yobH gene is located outside SPI-1, in a chromosomal region conserved in many bacteria, including S. bongori and E. coli K-12; YobH shares 79% sequence identity with its ortholog from E. coli K-12. A previous study indicates that HilD directly regulates expression of the flhDC operon, encoding the master regulator of the flagellar genes, which is also conserved in E. coli K-12 and many other bacteria 32,33 . YobH and its orthologs from different bacteria are on average 80 amino acids long and have no an assigned function; they are annotated as putative membrane, exported or uncharacterized proteins. Our preliminary results support that YobH is secreted in S. Typhimurium (data not shown). How is YobH secreted and what is the specific function of YobH for invasion, are topics of our current investigation.
HilD induces expression of a high number of target genes by acting directly or through distinct regulators, in growth conditions that somehow resemble the intestinal environment (SPI-1-inducing conditions), such as those that we assessed in this study 2 . At present, three different regulatory cascades headed by HilD have been well characterized: the HilD-HilA-InvF, HilD-SsrA/SsrB and HilD-FlhDC cascades 2,25,27,32,33,36,37 . Additionally, HilD forms a feed-forward positive loop with HilC and RtsA, which amplifies the activation of the HilD-HilA-InvF cascade 30,31 , and probably also the activation of the other regulatory cascades and genes controlled directly by HilD. Our data, together with previous results obtained from genome-wide expression and binding analyses 28 , define an additional cascade formed by HilD to induce expression of virulence genes. In this regulatory cascade, HilD induces expression of the yobH, slrP and ugtL virulence genes through SprB, a Salmonella-specific LuxR-like regulator encoded in SPI-1. Previous studies revealed that HilD and RtsA induce expression of slrP by an undefined way 62,63 . SlrP (Salmonella leucine-rich repeat protein) is an effector protein with ubiquitin ligase activity that is translocated into mammalian cells through both T3SS-1 and T3SS-2 63,64 . Previous reports support that SlrP plays a role in the intestinal colonization of S. Typhimurium in chicks, pigs, cows and mice, but not in the systemic infection in the mouse model 55,56 . UgtL is an inner membrane protein that mediates resistance to antimicrobial peptides by modifying lipid A in the lipopolysaccharide 57,59 ; furthermore, it is involved in the activation of the PhoP/PhoQ two-component regulatory system in response to mildly acidic pH 58 . UgtL is required by S. Typhimurium for killing 58 and for the intestinal colonization of mice 59 ; moreover, TraDIS analysis supports that UgtL is important for the intestinal colonization of S. Typhimurium in pigs 56 . Importantly to note, expression of both slrP and ugtL is also controlled directly by the PhoP/PhoQ two-component system, in growth www.nature.com/scientificreports www.nature.com/scientificreports/ conditions that somehow mimic the intracellular environment of host cells (SPI-2-inducing conditions) 63,65 , where the HilD-mediated regulation on target genes is not evident 25,36,63 ; in contrast, expression of yobH seems to be not regulated by PhoP/PhoQ 45 . PhoP forms with the SlyA regulator a feed-forward loop that controls expression of ugtL in SPI-2-inducing conditions 65,66 . Thus, expression of both ugtL and slrP is controlled by at least two distinct regulatory mechanisms that act in response to different environmental conditions. HilD-SprB and PhoP-SlyA would induce expression of ugtL and slrP in different niches where the activity of these genes is required for the Salmonella infection of hosts. For instance, activity of UgtL is needed for the intestinal colonization and for the systemic infection of mice 58,59 . On another hand, it is tempting to speculate that HilD-SprB helps to reach the levels of UgtL required for the subsequent UgtL-mediated activation of the PhoP/PhoQ system in response to acidic pH 58 , a cue present in the intracellular environment. Following this idea, it has been shown that activated PhoP represses expression of hilD, hilA and rtsA, and thus the SP-1 invasion genes 67 ; therefore, the HilD-SprB-UgtL-PhoP/PhoQ pathway could work as an additional negative feedback control in the complex and dynamic regulatory network governing expression of Salmonella invasion genes.
Global expression and binding analyses indicate that SprB positively controls expression of yobH, slrP, ugtL and 20 genes more 28 , all these genes located outside SPI-1, including the sifB, yhgE, yibP, SL1344_3112, SL1344_0336 and SL1344_0337 genes that have been associated to virulence 56,68 . Therefore, the HilD-SprB cascade represents an additional branch that further expands the HilD virulence regulon, connecting the activity of several genes located outside SPI-1 with the capability for invasion of host cells encoded within SPI-1.
Our results demonstrate that HilD positively controls expression of sprB by acting on the regulatory region upstream of this gene. Previous studies indicate that HilD can also control expression of sprB by acting on the hilC gene, located upstream of sprB; a hilC-sprB transcript was detected in a previous study 28 and direct regulation of hilC by HilD is well documented 28,30,38,48,51 . We show that HilD induces expression of sprB by directly displacing the repressor H-NS from the regulatory region upstream of this gene; a mechanism that HilD follows to induce expression of other target genes 36,37,[41][42][43] . H-NS represses expression of hilC 43,53,69,70 , which suggest that HilD induces expression of the hilC-sprB transcript also by antagonizing H-NS mediated repression. In contrast to HilD, which is required for the expression of target genes only in the presence of H-NS, we found that SprB is required for the expression of yobH even in the absence of H-NS, which supports that it does not act as an anti-H-NS factor. There is growing evidence to suggest that other LuxR-like regulators mainly act as classical activators, which induce expression of target genes by favoring binding of the RNA polymerase on promoters [71][72][73][74] . Whether SprB antagonizes a repressor different to H-NS or whether it acts as a classical activator remains to be elucidated.
Our data reveal a novel Salmonella invasion factor and further define an additional regulatory cascade mediated by HilD for the expression of Salmonella virulence genes. A model that summarizes the results from this study is depicted in Fig. 9.

Methods
Bacterial strains and growth conditions. Bacterial strains used in this study are listed in Table S1.
Bacterial cultures for the determination of chloramphenicol acetyl transferase (CAT) activity and for Western blot assays were grown in LB as described previously 25,26,36 . When necessary, the medium was supplemented with the following antibiotics: ampicillin (200 μg/ml), streptomycin (100 μg/ml), kanamycin (20 μg/ml) or tetracycline, (12 μg/ml). construction of plasmids. Plasmids and primers used in this study are listed in Tables S1 and S2, respectively. To generate the yobH-cat, sprB-cat, slrP-cat and ugtL-cat transcriptional fusions, the regulatory regions of yobH, sprB, slrP and ugtL were amplified by PCR using the primer pairs SL1770-FW22/SL1770-RV11, sprB-catF/ sprB-catR, slrPB2-Fw22/slrPH3-Rv11 and ugtL-Fw/ugtL-Rv, respectively, and chromosomal DNA from the WT S. Typhimurium strain as template. The resulting PCR products were purified with the Zymoclean Gel DNA Recovery Kit (Zymo Research), digested with BamHI and HindIII enzymes and then cloned into the pKK232-8 vector 75 digested with the same restriction enzymes. To construct the p2795-YobH-FLAG plasmid, the yobH::3X-FLAG gene was amplified by PCR using the primer pair SL1770-FW22/1770-SalIRv and chromosomal DNA from the DTM128 strain as template. This PCR product was digested with SalI and BamHI enzymes and then cloned into the p2795 vector 76 digested with the same restriction enzymes. To construct the pK6-SprB plasmid, the sprB structural gene was amplified by PCR using the primer pair sprB-K6NcoI/sprB-K6PstI and chromosomal DNA from the WT S. Typhimurium strain as template. This PCR product was digested with NcoI and PstI enzymes and then cloned into the pMPM-K6Ω vector 77 digested with the same restriction enzymes. pK6-SprB expresses SprB from an arabinose-inducible promoter.
construction of deletion and 3XFLAG-tagged S. typhimurium mutant strains. Non-polar deletion of the sprB, sinR or yobH genes in the S. Typhimurium SL1344 strain was performed by the λRed recombinase system, as reported previously 78 , using the respective primers described in Table S2, thus generating the strains DTM121, DTM123 and DTM124, respectively. The chromosomal yobH gene was 3XFLAG-tagged in the S. Typhimurium SL1344 strain using a previously reported method based on the λRed recombinase system 79 , thus generating the DTM127 (yobH::3XFLAG-kan) strain. P22 transduction was used to transfer the yob-H::3XFLAG-kan allele from the strain DTM127 into the strains JPTM25 and DTM122, generating the strains DTM129 and DTM131, respectively. The kanamycin resistance cassette was excised from the strains DTM121, DTM124, DTM127, DTM129 and DTM131, by using the pCP20 plasmid expressing the FLP recombinase, as described previously 78 , generating the strains DTM122, DTM125, DTM128, DTM130 and DTM132, respectively. The complemented DTM126 strain was generated by inserting the yobH::3XFLAG-kan into the chromosome (2019) 9:12725 | https://doi.org/10.1038/s41598-019-49192-z www.nature.com/scientificreports www.nature.com/scientificreports/ of the DTM125 strain, using a previously reported method based on the λRed recombinase system 76 and the p2795-YobH-FLAG plasmid. All modified strains were verified by PCR amplification and sequencing.
Chloramphenicol acetyltransferase (CAT) assays. The CAT activity and protein quantification to calculate CAT specific activities were determined as previously described 80 . Statistical analysis. Data were analyzed with GraphPad Prism 5.0 software (GraphPad Inc., San Diego, CA) using One-Way analysis of variance (ANOVA) with the Tukey's multiple comparison test. A P-value of <0.05 was considered significant.
Expression and purification of proteins. Expression and purification of MBP-HilD and H-NS-FH were performed as described previously 25,40 . Western blotting. Western blot assays were performed as described previously 26,36 . Anti-FLAG M2 monoclonal antibodies (Sigma) were used at 1:2,000 or 1:3,000 dilutions, for detection of YobH-FLAG and H-NS-FH, respectively. Anti-GroEL polyclonal antibodies were used at a dilution of 1:100,000. Horseradish peroxidase-conjugated secondary antibodies (Pierce), anti-mouse or anti-rabbit, were used at a dilution of 1:10,000. Blots were developed by incubation with the Western Lightning Chemiluminescence Reagent Plus (Perkin-Elmer) and then exposition to KodaK X-Omat films.
invasion assays. Invasion of HeLa cells or RAW264.7 macrophages was determined by gentamicin protection assays as described previously 40,81 . Briefly, HeLa cells or RAW264.7 macrophages were grown in high-glucose Dulbecco's Modified Eagle Medium (GIBCO 12100-046) supplemented with 10 mM sodium pyruvate solution, Figure 9. Model for the expression of YobH, SlrP and UgtL mediated by the HilD-SprB regulatory cascade. H-NS represses expression of sprB by binding the two promoter regions transcribing this gene. HilD binds to and thus displaces the H-NS repressor complex from these promoter regions, which allows expression of SprB that finally activates transcription of the yobH, slrP and ugtL virulence genes. Transcription of sprB from the promoter upstream of hilC and the effect of HilD and H-NS on this promoter were reported previously 28,30,43,48,69 . The previously defined regulation of ugtL and slrP by SlyA and/or PhoP is not depicted in the model but it is described in text.