LcrQ Blocks the Role of LcrF in Regulating the Ysc-Yop Type III Secretion Genes in Yersinia pseudotuberculosis

Pathogenic Yersinia species employ the Ysc-Yop type III secretion system (T3SS) encoded by a highly conserved pYV virulence plasmid to export the virulence effectors into host cells. The Ysc-Yop T3SS is tightly regulated by multiple contributing proteins that function at different levels. However, systematic transcriptional regulation analysis of Ysc-Yop T3SS is lacking and the detailed mechanism under this regulation process is still elusive. Aimed at systematically characterizing transcriptional regulations of all T3SS genes in Y. pseudotuberculosis, we amplified 97 non-coding fragments from the pYV plasmid and analyzed transcriptional responses of the T3SS genes under different growth conditions. Transcriptions of T3SS genes were induced at 37°C and genes encoding T3SS effectors were highly induced by further depletion of Ca2+. The temperature induced gene transcription process is mediated by modules encoded on the chromosome, while the Ca2+ depletion-induced process is controlled by the positive regulatory protein LcrF as well as the negative regulatory protein LcrQ. In this process, LcrQ shares the same targets with LcrF and the effect of LcrQ is dependent on the presence of LcrF. Furthermore, over-expression of LcrF showed the same phenotype as that of the lcrQ mutant strain and intracellular amount balance of LcrQ and LcrF is important in T3SS regulation. When the expression level of LcrF exceeds LcrQ, expression of the Ysc-Yop T3SS genes is activated and vice versa. Together, these data support a model in which LcrQ blocks the activation role of LcrF in regulating the transcription of T3SS genes in Yersinia.


Introduction
The genus Yersinia includes three human pathogenic bacterial species, Y. pestis, Y. enterocolitica and Y. pseudotuberculosis. These pathogens harbor a conserved virulence plasmid called pYV (or pCD) which encodes the Ysc-Yop type III secretion system (T3SS) [1]. When contacting with host cells, pathogenic Yersinia can efficiently inject effectors termed Yops (Yersinia outer proteins) into eukaryotic host cells with this secretion system. These translocated Yops enable bacteria to evade host immune responses by modulating host cell signaling pathways and facilitate the following infection [1][2][3].
The Ysc-Yop T3SS in Yersinia is both positively and negatively regulated and is induced by low Ca 2+ conditions at 37uC in vitro or by direct contact with host cells in vivo [4,5]. Several regulators are involved in the negative regulatory process. The LcrQ protein (or the YscM1 and YscM2 homologues in Y. enterocolitica) is an antiactivator which causes the feedback inhibition for Ysc-Yop T3SS gene expression. This protein is secreted outside of cells by the secretory machinery under low Ca 2+ conditions at 37uC with the help from a chaperone protein named SycH [6][7][8]. This process relieves the repressive effect of LcrQ and results in the activation of the Ysc-Yop T3SS [7]. We have also recently characterized the role of an export apparatus component named YscV in LcrQ secretion and subsequent Ysc-Yops T3SS regulation [9]. However, the targets and roles of LcrQ during this regulatory process have not yet been fully illustrated. In addition to LcrQ, other negative regulatory elements, including SycN, YscB, YopN, TyeA and the lcrGVH-yopBD operon were also demonstrated to play important roles in regulating T3SS gene expression [10][11][12][13][14].
The AraC family activator LcrF (known as VirF in Y. enterocolitica) is the only protein from the pYV plasmid that positively regulates the transcription of T3SS genes [15]. LcrF directly binds to the promoter regions of several yop genes and also binds to RNA polymerase as supposed by the AraC family proteins [16]. With such interactions, LcrF enhances the binding of RNA polymerase to specific promoters to facilitate the transcription initiation. Transcription of a number of genes from the pYV plasmid, e.g., ylpA, yadA, yopE, yopH, virC and lcrGVHyopBD operons, are dependent on LcrF [17][18][19][20]. The direct bindings of LcrF to yopE, yopH, virC and lcrGVHyopBD have been confirmed and the ,40 bp AT-rich region in each of these promoters is characterized to be recognized by LcrF [20].
The Ysc-Yop T3SS in Yersinia is regulated at both transcriptional and post-transcriptional levels. LcrF is a classical activator that controls T3SS gene expression at the transcriptional level [20]. YopD is another regulator which binds to the mRNA of several yop genes to accelerate mRNA degradation or inhibit the ribosome binding in gene translation [21,22], which acts at the post-transcriptional level. The expression of lcrF has also been shown to be regulated at the post-transcriptional level. The mRNA structure around the ribosomal binding region of lcrF transcript is opened when the temperature is shifted from 26uC to 37uC [23], which would facilitate the gene translation. Several other regulators, such as the YopN-TyeA complex, also act at the post-translational level to inhibit T3SS secretion process [13,24]. Although LcrQ negatively regulates the expression of T3SS genes, it lacks obvious DNA or RNA binding domains, and does not act by obstructing the secretion channel [25], thus the regulatory mechanism of LcrQ is not fully understood.
In this study, we systematically analyzed the responses of each gene on the pYV plasmid at the transcriptional level under different growth conditions in Y. pseudotuberculosis YPIII strain. By comparing the activities of each promoter, the transcriptional responses of the Ysc-Yop T3SS genes were characterized. The role of LcrQ in regulating the Ysc-Yop T3SS genes at the transcriptional level was subsequently analyzed, and its interaction with LcrF in this transcriptional regulatory process was further investigated and discussed.

Materials and Methods
Bacterial strains, plasmids, growth media and oligos Bacterial strains and plasmids used in this study are summarized in Table 1. Yersinia pseudotuberculosis YPIII strains were grown in YLB medium (1% tryptone, 0.5% yeast extract, 0.5% NaCl) and E. coli strains were grown in Luria-Bertani (LB) medium. When appropriate, antibiotics were used at the following concentrations: nalidixic acid (Nal) 15 mg ml 21 , chloramphenicol 30 mg ml 21 , ampicillin 100 mg ml 21 , kanamycin 100 mg ml 21 . All oligos used in this study are listed in Table S1.

Promoter cloning and b-galactosidase assay
The sequence of the pYV virulence plasmid in Y. pseudotuberculosis YPIII strain is not available. We collected and analyzed available sequences of the virulence plasmids of different Y. pseudotuberculosis strains from the NCBI database, e.g., pYV from IP 32953 (NC_006153.2) and IP 31758 (NC_009704.1), pYPTS01 from PB1+ strain (NC_010635.1). There are 99 annotated genes on the pYV plasmid from IP32953, 66 genes from IP31758, and 87 genes on pYPTS01 from PB1+ strain. To cover all possible genes, we chose the sequence of the pYV plasmid from IP32953 as the template to design primers ( Table S1) for cloning of the noncoding fragments from the pYV plasmid in YPIII. All these amplified non-coding fragments were cloned into the promoter cloning vector pZT100 [9]. These clones were respectively transformed into YPIII or its derivative strains and the bgalactosidase assay was performed at least in triplicate as previously described [26] at 26uC, or 37uC or 37uC with Ca 2+ depletion.

RNA isolation and real-time PCR
Total RNA of Yersinia strains cultured at 26uC, or 37uC or 37uC with Ca 2+ depletion was extracted with TRIzol reagent (Invitrogen) according to the manufacturer's instruction. After treatment with RNase-free DNase I (Promerga), cDNA was obtained by reverse transcription (Random 9 mers, TaKaRa) using 2 mg of each RNA sample. Real-time PCR was performed in reactions containing 16iTaq TM Universal SYBR Green Supermix Reagent (Bio-Rad), cDNA templates and the gene specific primers using the CFX Connect Bio-Rad system. All these assays were tested in duplicates. The mRNA level in each sample was normalized to the level of 16S rRNA gene transcript and amplification specificity was assessed using melting curve analysis. The degrees of expression change were calculated using the 22DDCt method.

Gel-retardation assay
The gel-retardation was performed as described previously [26]. The LcrF protein was expressed in E. coli BL21(DE3) by pET28a plasmid carrying the lcrF gene from YPIII (named pET28a-LcrF), and purified by Ni-NTA resin. About 400 ng of each DNA fragment, which was amplified from the YPIII genomic DNA, was incubated with various amounts of purified LcrF protein in 20 ml of binding buffer (20 mM Tris-HCl pH 7.4, 4 mM MgCl 2 , 100 mM NaCl, 1 mM dithiothreitol, 10% glycerol, and 100 ng bovine serum albumin). A ,200 bp fragment amplified from the non-coding region upstream of the pYV0053 gene was used as a negative control. Probe binding were performed at 37uC for 1 h, and were then loaded onto a 6.5% native polyacrylamide gel. Electrophoresis was performed in 0.56TBE buffer on ice. The gel was stained with ethidium bromide for 20 min and scanned using a Syngene GeneGenius gel documentation system.

Mutant construction and complementation
For DlcrF construction, two DNA fragments flanking the lcrF gene were amplified by PCR from the genomic DNA of YPIII using two pairs of primers: named LcrF-in-upF/R, and LcrF-in-downF/R (Table S1). These two fragments were mixed and then used as the template DNA in a second round of PCR with primers LcrF-in-upF and LcrF-in-downR. The PCR product was digested with SalI and BglII and inserted into the suicide plasmid pDM4 [27]. This recombinant plasmid was transformed into E. coli S17-1 and transconjugation was performed as described [27] to obtain the DlcrF strain. To complement DlcrF, a plasmid named pOVR200 was constructed, which contains the pMB1 replicon, the bla gene and the lacI gene from pET21a (Novagen) and a lac promoter upstream of the multiple cloning site. The lcrF gene was PCR amplified from YPIII and was cloned into pOVR200 to obtain pOVR-LcrF, which was transformed into DlcrF. The pOVR200 plasmid was also transformed into DlcrF as a control.

Over-expression and co-expression of LcrF and LcrQ
The pOVR-LcrF plasmid obtained above was transformed into YPIII to over-express LcrF protein. To over-express LcrQ protein, lcrQ gene was amplified from YPIII and cloned into pOVR200 to obtain pOVR-LcrQ, which was transformed into YPIII. For coexpression of LcrF and LcrQ (or GST-LcrQ), the low copy plasmid named pKT100 [26], which contains the pOVR200 compatible P15A replicon and a promoter region of the chloramphenicol resistance gene, was used to express the target proteins together with the pOVR200 plasmid.

Yops analysis
Overnight cultures of Yersinia strains grown in Ca 2+ rich medium (YLB with 2.5 mM Ca 2+ ) were diluted (1:20) into fresh Ca 2+ -depleted medium (YLB containing 5 mM EGTA and 20 mM MgCl 2 ) [28] and incubated at 26uC for 1 h, then shifted to 37uC to induce Yops expression and secretion. When required, isopropyl b-D-thiogalactoside (IPTG) at a final concentration of 1 mM was added into the cultures before temperature changes. After 4 h of incubation at 37uC, 20 ml of each culture was harvested and centrifuged. The wet weight of the bacterial cell pellet was measured and the supernatant was collected and filtered by 0.22 mm filter (Millipore). Proteins from the supernatant were precipitated overnight at 4uC with 10% trichloroacetic acid as described [9]. The proteins from both supernatants and pellets were dissolved in the SDS loading buffer according to the wet weight of the bacteria and resolved by SDS-PAGE. Western-blot was carried out to detect the proteins following SDS-PAGE. The horseradish peroxidase-labeled anti-rabbit antibodies and a chemiluminescence detection kit (Beyotime Institute of Biotechnology) were used to develop the reaction following the protocol provided by the manufacturer.

pYV plasmid curing
The pYV plasmid in YPIII was cured as described by others [29]. Briefly, YPIII strain was grown on YLB plate supplemented with 20 mM of sodium oxalate, 20 mM MgCl 2 , and 0.05% Congo Red Dye at 37uC. White colonies were picked up and cultured overnight in YLB medium and then spread on Congo Red plate. After overnight culture at 37uC, a single white colony (loss of the functional pYV plasmid results in white colonies) was chosen for PCR to further confirm that the pYV plasmid was cured using five pairs of primers, which paired to different regions on the pYV plasmid.

Bacterial two-hybrid assay
The adenylate cyclase based bacterial two-hybrid system [30] was applied in analyzing the interaction between LcrF and LcrQ proteins. Briefly, pKT25 plasmid carrying T25 fragment was used to clone lcrF and sycH genes to obtain T25-LcrF and T25-SycH. The pUT18 carrying T18 fragment was applied to clone lcrF and lcrQ genes to obtain T18-LcrF and T18-LcrQ. These two pairs of plasmids were co-transformed into E. coli BTH101 and were cultured in LB medium with 1 mM IPTG at 30uC for 2 days. The b-galactosidase activities were determined as described [26].

Statistical analysis
All data for the b-galactosidase activity assays were shown as mean 6 standard deviation (SD) of the results of multiple independent experiments. Student's T test was used to compare the data between two relevant groups. DyscV 618-644 YPIII with deletion of a 81 bp fragment encoding amino acid 618-644 in yscV gene, Nal r [9] E. coli

Transcriptional responses of T3SS genes under inducible conditions
To systematically characterize the transcriptional responses of Ysc-Yop T3SS genes in YPIII under T3SS inducible conditions, we compared the promoter activities of all genes from the pYV plasmid. Using primers designed from Y. pseudotuberculosis IP 32953 strain, we successfully amplified 96 non-coding regions from the pYV plasmid in YPIII. Only three regions upstream of pYV0042, pYV0043 and pYV0044 could not be obtained. We next sequenced the region from pYV0041 to pYV0045, a deletion at the C-terminal of the coding region of pYV0041 and a full deletion of the pYV0042 and pYV0043 genes were found in YPIII. Using this new sequence, we designed another pair of primers to amplify the non-coding region of pYV0044. In total, 97 non-coding fragments were amplified and cloned into a lacZ reporter vector to obtain non-coding fragment-lacZ fusions. The promoter activities of these non-coding fragments under T3SS non-inducible (26uC and 37uC) and inducible (37uC with Ca 2+ depletion) conditions were subsequently analyzed. As shown in Fig. 1A, 68 out of the 97 non-coding fragments showed promoter activities compared with the pZT100 plasmid (202612, p,0.01) at least under one condition. We also compared the promoter activities of these 68 non-coding fragments at 26uC and 37uC, which is one of the key factors that induces T3SS in vitro. As shown in Fig. 1B, 34 out of 68 promoters showed higher (.2 folds) activities at 37uC, suggesting some factors activate the expression of T3SS genes when bacterial cells were shifted from 26uC to 37uC.
To test if these factors are encoded by genes on the pYV plasmid, we next compared this activation in YPIII parent strain with the pYV plasmid cured strain (DpYV). Our results showed that induced folds at 37uC for all these promoters are similar between YPIII parent and DpYV strains (Fig. 1B), which suggests that the activation at 37uC is largely mediated by factors encoded on the chromosome. To fully characterize the transcriptional response during the T3SS inducing process, we further compared the promoter activities of these 68 fragments at 37uC and at 37uC with Ca 2+ depletion. As shown in Fig. 1C, the activities of 9 promoters were induced more than 2-fold when Ca 2+ was depleted. On the contrary to temperature changes, this activation is dependent on factors encoded on the pYV plasmid, since all these activations were strongly decreased in the DpYV strain (Fig.  1C). To further confirm the lacZ fusions results, we applied qRT-PCR to test the mRNA levels of lcrG and lcrQ genes in YPIII under different conditions. Being consistent with the b-galactosidase activity assay, the mRNA level of lcrG gene was highly induced at 37uC with Ca 2+ depletion, and lcrQ gene is induced at 37uC (Fig.  S1).
Comparing Fig. 1B and Fig. 1C, we conclude that most of the promoters activated between 37uC and the Ca 2+ depleted conditions do not overlap, suggesting the Ysc-Yop T3SS activation process induced by 37uC and low Ca 2+ is different. As summarized in Table S2, the genes induced at 37uC could be divided into four categories according to their functions: I) pYV plasmid replication; II) secretion machinery components; III) effector proteins; and IV) hypothetical proteins. On the contrary, the genes induced by Ca 2+ -depletion mainly encode effectors for Ysc-Yop T3SS (Table  S3). Based on these analyses, we conclude that the Ysc-Yop T3SS inducing process first involves genes encoded from the YPIII chromosome, which activate the transcription of most of the secretion machinery components when the culture temperature is shifted from 26uC to 37uC, and full activation requires other key factors encoded on the pYV plasmid to maximally express the effector proteins.

Ca 2+ -depletion induced genes are regulated by LcrF
Since LcrF protein is the only characterized transcriptional regulator from the pYV plasmid, we next aimed to explore the role of LcrF in Ca 2+ -depletion induced gene expression. To systematically analyze the regulatory targets of LcrF, We compared the promoter activities from the pYV plasmid under T3SS inducible conditions in YPIII parent and the lcrF mutant (DlcrF) strains. Interestingly, activities of those nine Ca 2+ -depletion induced promoters, including promoters upstream of sycO (pYV0002), yadA (pYV0013), yopE (pYV0025), yopK (pYV0040), lcrG (pYV0058), yopN (pYV0065), yscN (pYV0067) and yopH (pYV0094), were all dramatically decreased in DlcrF (Fig 2A), which suggests all these genes are possible targets of LcrF. To further confirm these promoters are indeed regulated by LcrF, we also compared the promoter activities in the DlcrF complementary strain. As shown in Fig. 2B, expression of LcrF in DlcrF successfully restored the activities of these promoters.

Transcriptional regulation of LcrF to T3SS is not influenced by Ca 2+
Since transcription of all the targets of LcrF is induced by Ca 2+depletion, we next compared the regulatory effects of LcrF under Ca 2+ rich and depleted conditions. As shown in Fig. 2C, overexpression of LcrF in DlcrF induced the expression of Yops genes in the presence of Ca 2+ , which suggests the regulatory activity of LcrF is not influenced by Ca 2+ . To confirm this, we next tested the regulatory role of LcrF to its targets in the DpYV strain. Our results showed that expression of LcrF also increased the activities of these promoters both in Ca 2+ rich and depleted conditions (Fig.  S2). These data are inconsistent with the results from Y. enterocolitica in which the role of LcrF is not affected by Ca 2+ [31]. However, our data demonstrated that transcription of all the promoters targeted by LcrF are strongly induced by depletion of Ca 2+ at 37uC, which suggests other co-factor(s) may sense the change of Ca 2+ concentration and coordinate with LcrF in regulating the Ysc-Yop T3SS.

Role of LcrQ in the regulatory process of LcrF
The LcrQ protein globally regulates the Ysc-Yop T3SS in respond to Ca 2+ concentration changes, but the mechanism under this regulatory role has not been illustrated. To test whether LcrQ is involved in controlling the regulatory effect of LcrF, we next screened the targets of LcrQ by comparing the promoter activities in YPIII parent and LcrQ over-expressed strains using our promoter library. Interestingly, all and only the targets of LcrF showed decreased promoter activities in the LcrQ over-expressed strain compared with those in YPIII parent strain (Fig. 3A), which suggests LcrQ shares the same targets with LcrF in regulating the Ysc-Yop T3SS.
To further analyze the relationship between LcrF and LcrQ in regulating the Ysc-Yop T3SS, we next compared the phenotypes of DlcrF with LcrQ over-expressed strain and LcrF over-expressed strain with DlcrQ. The expression of Yops in the LcrQ overexpressed strain was the same as that in DlcrF, which were both decreased under T3SS inducible conditions (Fig. 3B, lanes 2&4), and Yops secretion was also decreased as confirmed by Western blot analysis (Fig. 3B). Over-expression of LcrF and deletion of the lcrQ gene both activated Yops expression and secretion under T3SS inducible conditions (Fig. 3C). Moreover, deletion of the lcrQ gene successfully secreted LcrV and YopD proteins in the presence of Ca 2+ at 37uC as reported by others [32], and overexpression of LcrF also promote the secretion of these two proteins (Fig. 3C). Together, these data suggest LcrQ and LcrF are closely related during the regulatory process.
We next compared the regulatory effects of LcrQ to its targets in YPIII parent and DpYV strains. As shown in Fig. 3D, promoter activities of the LcrQ targets were at basal levels in DpYV strain and over-expression of LcrQ could not repress these promoters in DpYV strain; however, these promoter activities were obviously inhibited by over-expressing LcrQ in YPIII parent strain (Fig.  S3), suggesting that LcrQ represses the promoter activities activated by pYV-encoded protein. To test whether this factor is LcrF, we tested the repressive effect of LcrQ in DlcrF strain and found no repressive role to these promoters (Fig. 3E). It is worthy to note that the activities of these promoters in LcrQ overexpressed YPIII strain were at similar levels as that in DlcrF strain, indicating that LcrQ only represses the LcrF-activated but not the basal level promoter activities. Together, these data suggest that the repressive effect of LcrQ requires LcrF.

Intracellular LcrQ inhibits the regulatory role of LcrF
Judging from the data we have obtained above, we hypothesized that intracellular LcrQ blocks the regulatory role of LcrF. If so, the expression levels of these two proteins are important for the regulatory process. To confirm this hypothesis, we over-expressed LcrF using the pOVR200 plasmid in the YPIII strain carrying pKT-LcrQ, in which LcrQ protein was also over-expressed by the pKT100 plasmid. As shown in Fig. 4A, over-expression of LcrF released the negative regulatory effect of LcrQ in this strain (compared lane 4 with lane 3). In agreement with this observation, the amount of intracellular LcrQ was obviously decreased in LcrF over-expressed strain (Fig. 4A). Therefore, these results can be explained by intracellular amount balance of LcrF and LcrQ proteins. When the amount of LcrF exceeds LcrQ, LcrF dominates the regulatory effect. To prove our hypothesis, we next exchanged the plasmid to over-express these two proteins. pOVR200, which contains a strong promoter, was used to overexpress LcrQ and pKT100 was used to express LcrF. In this case, the amount of intracellular LcrQ protein was slightly increased (Fig. 4A, compared lane 7 with lane 4), which in turn greatly inhibited the positive effect of LcrF (Fig. 4A, compared lane 7 with lane 5). All together, these results suggest that the intracellular amount of LcrQ is important for blocking the regulatory role of LcrF.
To further confirm our hypothesis, we used the pOVR200 plasmid to express GST-LcrQ (which is known to abolish transport of LcrQ [6,33]) in YPIII strain carrying the pKT-LcrF plasmid. As anticipated, GST-LcrQ is not secreted (Fig. 4B). And over-expression of LcrF by the pKT100 plasmid in this background did not show obvious effect on the expression or secretion of Yops (Fig. 4B, compared lane 7 with lane 5). In our previous work, we showed that mutation of the yscV gene (DyscV 618-644 ), which encodes a component of secretion machinery, accumulates the intracellular LcrQ protein and therefore represses the expression of Yops [9]. We next over-expressed LcrF protein in this background and found that both Yops expression and the promoter activities of its targets were all activated comparing with DyscV 618-644 carrying the control plasmid pOVR200 (Fig. 4C&D). Taken together, our data suggest that the amount of intracellular LcrQ and LcrF is vital for the regulation of Ysc-Yop T3SS. We hypothesize that intracellular LcrQ abolishes the role of LcrF at 37uC under Ca 2+ rich conditions, and this inhibition is released when LcrQ is secreted outside of cells under Ca 2+ depleted conditions.

LcrQ does not directly interact with LcrF
The regulatory mechanism for the LcrQ protein has not been fully understood. No DNA binding domain can be found on this protein [25], suggesting LcrQ does not directly interact with its target DNA. To explore how LcrQ functions in the regulatory process of LcrF, we applied the Bacterial Adenylate Cyclase Two-Hybrid (BACTH) system to detect possible interaction between these two proteins. As shown in Fig. 5, LcrF can form dimer as reported for the AraC family proteins [16], and LcrQ can interact with its chaperone protein SycH [6,7], which suggest the expression of these two proteins in the BACTH system is stable. However, we could not detect direct interaction between LcrF and LcrQ (Fig. 5), suggesting that other factor(s) is also required for LcrQ to interact with LcrF to block its regulatory role.

Discussion
The Ysc-Yop T3SS is a strictly regulated system and several modules have been demonstrated to play important roles in this regulatory process [34]. However, this regulation at the transcriptional level has not been systematically analyzed. In this study, we have cloned 97 non-coding fragments from the pYV plasmid into the lacZ fusion plasmid and compared the promoter activities of all these fragments under T3SS inducible and non-inducible conditions. The Ysc-Yop T3SS is induced by low Ca 2+ signal at 37uC in vitro [4]. In agreement with this phenotype, transcription of genes on the pYV plasmid is also regulated by low Ca 2+ signal at 37uC. Based on the promoter activity analyses, we proposed a model for transcriptional regulation of Ysc-Yop T3SS in Yersinia. As shown in Fig. 6, factors encoded on the chromosome first sense the temperature changes. At 37uC, these factors will activate the transcription of most genes on the pYV plasmid, including the secretion machinery components and two regulators LcrF and LcrQ. Under this condition, LcrQ blocks the role of LcrF. Therefore, the expression of T3SS genes could not be massively activated by LcrF at 37uC when LcrQ and LcrF are both intracellularly located. However, the low Ca 2+ signal will trigger the secretion of LcrQ, which subsequently release the repressive effect of LcrQ, the positive regulatory role of LcrF will thus be exerted.
LcrF and LcrQ are two characterized regulators for Ysc-Yop T3SS in Yersinia. LcrF is a classical transcriptional regulator that can directly binds to several promoters on the pYV plasmid [20]. In this study, we systematically characterized the targets of LcrF and identified 9 targets in YPIII. Being consistent with other reports in Y. enterocolitica [20], yopE, yopH, yopQ/yopK, yadA, lcrGVHyopBD operon and virC are all activated by LcrF. VirF is not essential for transcription of both virA and virB operons in Y. enterocolitica [15], but these two operons were both regulated by LcrF in Y. pseudotuberculosis. Binding of LcrF to these two promoter fragments were confirmed by in vitro gel-retardation assay (Fig.  S4), which suggests these two operons are also the direct targets of LcrF. The regulation of LcrF to its targets is not affected by Ca 2+ , but the transcription of all these genes in YPIII are induced by low Ca 2+ signal. These data suggest that some other factors are also involved in activating the role of LcrF under low Ca 2+ conditions. The role of LcrQ in T3SS regulation has not been documented. LcrQ has been proposed to form a complex with YopD-LcrH proteins [35]. A recent study has shown that YscM1 (a homologue of LcrQ in Y. enterocolitica) binds to ribosomes together with YopD-LcrH and subsequently inhibits translation of the yop genes [36]. Our current study suggests another mechanism for the regulatory role of LcrQ to block the role of LcrF. This blocking process is not in a direct way since we did not observe any direct interaction between these two proteins by a customized bacterial two-hybrid assay, which is in agreement with a previous study showing that LcrQ does not act as an anti-LcrF protein [25]. We suppose that other uncharacterized proteins may act as a linker between these two proteins in this process (Fig. 6). Functional homologues of LcrF and LcrQ are found in other pathogens. For example, the T3SS in Pseudomonas aeruginosa is triggered by low Ca 2+ at 37uC and is also globally regulated by two proteins with opposing effects: the AraC family positive regulator ExsA and the secreted negative regulator ExsE [37]. Similar to LcrQ and LcrF, ExsE blocks the role of ExsA by indirect interactions with two intermediate proteins ExsC and ExsD [37]. Further studies are needed to characterize functional homologues of ExsC and ExsD in Yersinia and find the binding proteins of LcrF and LcrQ, which will provide information to detail the mechanism of transcriptional regulation to the Ysc-Yop T3SS. In addition to regulators encoded by the pYV plasmid, several chromosome encoded proteins are also involved in the regulation of the Ysc-Yop T3SS. For example, YmoA, a chromosomeencoded histone-like protein, directly binds to the promoter region of the lcrF gene to inhibit its transcription. The ATP-dependent ClpXP and Lon proteases degrade YmoA when the environment temperature achieved at 37uC, which in turn releases the transcription of lcrF gene and subsequently activates the T3SS in Yersinia [38]. And the LysR family regulatory protein YtxR competes with LcrF in binding with the promoter regions of yopE and yopH genes in regulating T3SS [39,40]. All these factors regulate the Ysc-Yop T3SS indirectly through LcrF protein. In our study, we found that transcription of most of the pYV genes are activated by the chromosome-encoded factors at 37uC and these pYV-encoded genes are not all belongs to the targets of LcrF. These data suggest that chromosome-encoded factor is not one of the above proteins. Previous studies have detected that supercoiling of the pYV plasmid changes at 30uC and 37uC and supposed that this DNA topology change may be the temperature-sensing mechanism for T3SS gene expression [41,42]. Whether the uncharacterized factors are proteins participated in DNA architecture changes need further studies.
As stated in the introduction, the Ysc-Yop T3SS is regulated at both transcriptional and post-transcriptional levels. In this study, our proposed model is only for the transcriptional regulation of the Ysc-Yop T3SS, which could not explain all the phenotypes we have observed in our study. For example, as suggested in our model, deletion of the lcrQ gene would facilitate the role of LcrF to activate the Yops at 37uC in the presence of Ca 2+ , as demonstrated by the fact that blockage to the role of LcrF by LcrQ is diminished in the DlcrQ strain. However, as shown in Fig. 3C, the secretion pattern of the DlcrQ strain under Ca 2+ rich conditions is different from the normal secretion pattern, which suggests some other factors may also participate in controlling the secretion pathway of Ysc-Yop T3SS. In addition, the promoter activities of yopE and yopH remains high in DlcrF background (Fig. 3E), but Western blot results showed the expression of these proteins are undetectable in this strain (Fig. 3B), which indicate the regulation also occurs at the post-transcriptional level. Moreover, we found that the thermal induction of yopEp-lacZ is not dependent on the pYV plasmid, which is different from the results using a yopE-lacZ fusion (the yopE covers a promoter and a ,700 bp coding region in this fusion) [19]. These suggest some pYV-encoded elements may inhibit the expression of yopE at the post-transcriptional level, e.g., YopD protein [21]. The expression of LcrQ protein was increased in LcrF over-expressed strain (Fig. 4 & Fig. S5) but its promoter activity was not activated by LcrF protein. Transcription is only the first step for gene expression. Regulation at other levels should be considered together to fully clarify the complicated regulatory process of Ysc-Yop T3SS.
Among the 97 cloned non-coding fragments, 68 have promoter activities (summarized in Fig. S6), and several of these non-coding fragments located inside an operon, e.g., the non-coding fragment upstream of yopD gene (pYV0054), which is located in the lcrGVH-yopBD operon, were detected with promoter activity, suggesting the transcriptional regulation of pYV-encoded T3SS is complicated. On the other hand, we did not observe promoter activities of some genes located as the first gene in an operon. This may due to the fact that some fragments we chose do not cover the full  promoter regions, since open reading frame annotation is only based on bioinformatics analysis from the NCBI, which may introduce mis-annotation. Nevertheless, the promoter bank we have constructed in this study will facilitate further transcriptional regulation analysis in Yersinia and results obtained in this study would provide information for further transcriptional regulation analyses related to the Ysc-Yop T3SS. Figure S1 The mRNA levels of lcrG and lcrQ genes in YPIII at 266C, 376C or 376C with Ca 2+ depletion.  Figure S5 Relative level of LcrQ protein in LcrF overexpressed strains (quantified from Figure 4A). (TIF) Figure S6 Location of promoters on the pYV plasmid. The non-coding fragments with promoter activities detected in this study were labeled with black ''P'', and those located upstream of the first gene in an operon but showed no promoter activity in our test were labelled with grey ''P''. (TIF)