|Home | About | Journals | Submit | Contact Us | Français|
Yersinia type III secretion machines transport substrate proteins into the extracellular medium or into the cytoplasm of host cells. Translational hybrids, involving genes that encode substrates as well as reporter proteins that otherwise cannot travel the type III pathway, identified signals that promote transport of effector Yops into host cells. Signals for the secretion of substrates into high calcium media were hitherto unknown. By exploiting attributes of translational hybrids between yopR, whose product is secreted, and genes that encode impassable proteins that jam the secretion machine, we isolated yopR mutations that abolish substrate recognition. Similar to effector Yops, an N-terminal or 5’ signal in codons 1–11 is required to initiate YopR into the type III pathway. YopR secretion cannot be completed and translational hybrids cannot impose a block without a second signal, positioned at codons 131–149. Silent mutations in the second signal abrogate function and the phenotype of other mutations can be suppressed by secondary mutations predicted to restore base complementary in a 3’ stem-loop structure of the yopR mRNA.
During host infection, bacterial pathogens employ specialized secretion pathways that direct proteins to discrete locations (Lee and Schneewind, 2001). For example, pathogenic Yersinia species (Y. pestis, Y. pseudotuberculosis and Y. enterocolitica) use virulence plasmid-encoded type III machines to direct proteins into body fluids or into the cytosol of immune cells, thereby preventing phagocytic killing (Boland et al., 1996; Lee and Schneewind, 1999; Marketon et al., 2005; Rosqvist et al., 1994). Many different Gram-negative pathogens are endowed with type III secretion genes (Galán and Wolf-Watz, 2006). Although secretion substrates or host cells that are targeted for type III secretion vary between microbial species, evolutionary conservation of eleven machinery genes suggests that the mechanisms of substrate recognition and transport may be shared (Hueck, 1998; Rosqvist et al., 1995). This conjecture may apply also to flagella, a secretion system that shares nine genes with type III machines (Blocker et al., 2003).
The type III pathway of Yersinia spp. selects only few substrates (YscF, YscI, YscM1, YscM2, YscO, YscP, YscX, LcrV, YopB, YopD, YopE, YopH, YopM, YopN, YopO, YopP, YopQ, YopR & YopT) out of a cellular complement of more than four thousand (Michiels et al., 1990). Surprisingly, type III substrates do not share amino acid sequence homology or structural similarity that could explain the selective attributes of this pathway (Sorg et al., 2005b). A hierarchy in the synthesis of Yersinia substrates has not been observed. Nevertheless, substrate selection in the type III pathway appears ordered, beginning with factors that first extend the secretion machinery on the bacterial surface to form a hollow needle structure (YscF and YscP) (Journet et al., 2003; Sorg et al., 2007). Other proteins are deposited at the needle tip (LcrV) (Mueller et al., 2005) or into the plasma membrane of host cells (YopB and YopD) to complete the type III conduit into host cells (Håkansson et al., 1993; Håkansson et al., 1996).
Type III machines are activated by a drop in calcium concentration, which presumably is encountered as needles leave calcium-rich body fluids and penetrate into the low-calcium environment of the host cell cytoplasm (Lee et al., 2001; Pollack et al., 1986). The low calcium environment may impact needle structure and thereby trigger transport of effectors Yops into immune cells (Torruellas et al., 2005). During infection of tissue cultures, Yersinia assemble type III needles to specifically transport Yop effectors into HeLa cells (Lee et al., 1998). Nevertheless, Yersinia type III secretion can also be activated in laboratory media via the chelation of calcium ions, causing massive secretion of all substrates into the extracellular milieu and imposing a restriction of bacterial growth (Goguen et al., 1984; Michiels et al., 1990).
Earlier work mapped secretion signals with translational hybrids between genes that encode type III substrates or reporter molecules that otherwise reside in the bacterial cytoplasm (Michiels and Cornelis, 1991). When tested under conditions where secretion is induced via calcium chelation, all substrate genes examined thus far harbor a signal within the first seven to fifteen codons that promotes type III transport of fused reporter products into the extracellular medium (Anderson and Schneewind, 1997; Anderson and Schneewind, 1999; Schesser et al., 1996; Sory et al., 1995). Nevertheless, none of these N-terminal signals alone is sufficient to promote type III secretion of reporter proteins during Yersinia infection of tissue culture cells. The latter requires additional sequences in substrate genes, typically at least 50 codons (Boland et al., 1996; Lee et al., 1998; Sory et al., 1995). Mutational analysis of secretion signals and their associated phenotypes has led to diverse proposals for peptide or mRNA signals, for specific chaperones that interact with substrate residues following the N-terminal signal or even the structural fold of secretion substrate gene products that may be specifically recognized by the type III machinery (Lloyd et al., 2001; Sorg et al., 2005b). While all of these models contributed towards the exploration of type III secretion, the accumulated data can not yet be reconciled into a cohesive, unifying model that explains the mechanisms of substrate recognition (Cornelis, 2003; Ramamurthi and Schneewind, 2003).
Trapping intermediate steps in the housekeeping secretion process of Escherichia coli has permitted genetic and biochemical analysis of the canonical Sec pathway, which selects signal peptide bearing precursor proteins for secretion (Bassford Jr. et al., 1979; Oliver and Beckwith, 1981; Tian et al., 2000). To arrest secretion, export substrates were fused to the large, tightly-folding cytoplasmic protein LacZ (Kalnins et al., 1983). As signal peptides engage precursors into the Sec machinery in a specific and irrevocable manner, these impassable hybrids jam the entire secretion pathway for all substrates (Bieker and Silhavy, 1990). Attempts to similarly trap type III machines have revealed that most substrates cannot block the pathway (Lee and Schneewind, 2002; Sorg et al., 2005a). Only two secreted substrates, YscP and YopR, when fused as translational hybrids to impassable reporter proteins are able to block type III secretion (Riordan et al., 2008; Sorg et al., 2006).
What might be different about blocking substrates and non-blocking substrates, which can be rejected from the type III pathway? YopR and YscP do not share obvious physical features, though they do have a few functional attributes in common: neither YopR nor YscP is injected into eukaryotic cells and neither is thought to possess a chaperone (Lee et al., 1998; Payne and Straley, 1999). Both are phenotypically or genetically linked to the formation of the needle filament: yopR is immediately flanked by genes that encode the building blocks for the needle filament though the contribution of YopR to this process is unclear (Allaoui et al., 1995). YscP controls the length of the filament and is secreted during its polymerization (Journet et al., 2003). Both proteins contain N-terminal signals (Agrain et al., 2005b; Sorg et al., 2006) and YopR-GST or YscP-GST fusions that lack this signal no longer block the type III pathway (Riordan et al., 2008; Sorg et al., 2006). In the present work, we determined that the N-terminal signal of YopR, when fused to GST, was insufficient to create a blocking substrate. To find the additional secretion signals, we examined YopR-GST function following randomized mutagenesis and identified an mRNA motif that resides within codons 131–149 of yopR. This element acts in concert with the N-terminal signal to promote the export of YopR under physiological conditions.
The first eleven codons or amino-acids of YopR contain a secretion signal (Fig. 1A) and YopR-GST constructs lacking this signal do not block type III secretion (Sorg et al., 2006). Many Yop effectors associate with Syc chaperones that promote their type III transport and typically bind the gene product downstream from N-terminal secretion signals (Wattiau and Cornelis, 1993; Wattiau et al., 1994). To examine yopR sequence requirements for the blockade, full length yopR (1–165) or 3’ gene truncations were fused to dhfr, generating plasmids with the translational hybrids yopR1–165-dhfr, yopR1–150-dhfr, yopR1–100-dhfr and yopR1–50-dhfr (Fig. 1A). Type III secretion by Y. enterocolitica wild-type strain W22703 was activated by calcium chelation and expression of translational hybrids via the tac promoter was induced with IPTG. Following centrifugation of Yersinia cultures, proteins secreted into the supernatant (S) were separated from the bacterial pellet (P) and analyzed by Coomassie-stained SDS-PAGE. In contrast to YopR1–165-DHFR and YopR1–150-DHFR, the truncated hybrids YopR1–100-DHFR and YopR1–50-DHFR failed to block type III secretion, as Coomassie-stained Yop proteins could be detected in supernatant samples with similar abundance as in cultures where expression of YopR1–165-DHFR was not induced (- IPTG, Fig. 1B). This observation was corroborated by immunoblotting for the secretion substrates YopE and YopD as well as neomycin phosphotransferase (Npt), a control for a cytoplasmic protein (Fig. 1C). To test whether the defect in the type III blockade occurred also during Yersinia infection of HeLa tissue cultures, F-actin of infected cells was stained rhodamine-conjugated phalloidin to reveal cell rounding and actin rearrangement as a measure for type III effector injection. IPTG-induced expression of YopR1–165-DHFR and YopR1–150-DHFR abolished type III effector injection, apparent by the lack of cell rounding and actin cable rearrangements. In contrast, YopR1–100-DHFR and YopR1–50-DHFR failed to block type III effector injection (Fig. 1D). Thus, the N-terminal signal (codons 1–11) is not sufficient to impose the type III secretion blockade of YopR-DHFR, which requires an additional element located somewhere in yopR codons 12–150.
To characterize the second secretion signal of YopR, plasmids encoding yopR-lacZ or yopR-gst were first subjected to hydroxylamine mutagenesis in vitro and then transformed into Y. pestis KIM8. The low calcium response (LCR) of Y. pestis type III machines prevents bacterial growth on agar media with chelated calcium ions (Goguen et al., 1984; Kupferberg and Higuchi, 1958), however IPTG induced expression of blocking hybrids, yopR-lacZ or yopR-gst, restores Y. pestis growth (Sorg et al., 2006) (Fig. 2A). This feature is unique to Y. pestis, as blockades of type III secretion in Y. enterocolitica precipitate different rates but not a restriction of growth (data not shown). Four thousand transformants were scored for LCR and 122 plasmids reproducibly failed to enable Y. pestis growth on agar with chelated calcium. DNA sequencing of 122 plasmids identified 71 episomes with lesions in yopR (Fig. 2BC). All of these mutations recurred in independent experiments, suggesting that hydroxylamine mutagenesis was performed to saturation.
Treatment of DNA with hydroxylamine causes G to A and C to T transitions that often generate nonsense codons (TGA, TAA, TAG) from Trp (TGG, once in yopR), Gln (CAG and CAA, twelve times in yopR) or Arg (CGA, not in yopR) codons (Miller, 1991). Thus, scoring for the loss of yopR-lacZ or yopR-gst function, this screen should be answered by some missense and many nonsense mutations. Surprisingly, the screen isolated only six nonsense mutations in yopR but many more missense mutations, all of which mapped to a short segment near the 3’ end of the yopR gene: codons 131 to 149. Similar results were observed when screening yopR-gst and yopR-lacZ mutants for the LCR phenotype (Fig. 2BC). It is indeed unexpected that hydroxylamine mutagenesis of yopR, containing fourteen targets for nonsense mutations, generated such disproportionate ratio of nonsense to missense mutations (Fig. 2BC). As a control, we used hydroxylamine mutagenesis of yopR-lacZ and screened for LacZ− mutants (white instead of blue colonies on X-Gal containing agar), which generated the expected proportion of nonsense to missense mutations (Fig. 2D). Even more curiously, a silent mutation in codon 143 (Ile codon ATC to ATT) was observed fourteen times and thrice in isolation. These observations suggested to us that the loss of recognition of YopR-LacZ by the secretion apparatus entailed an mRNA element contained in codons 131 to 149 (Fig. 2E).
Linkage of the observed phenotypes to mutations in yopR was established by moving plasmids encoding wild-type and mutant yopR-gst alleles into Y. enterocolitica W22703. We chose E140K and the double mutant E135K, E140K for further analysis, because of the very high frequency (33 mutants out of 71) with which these alleles had been isolated. The function of wild-type and mutant yopR-gst was examined by measuring IPTG-inducible inhibition of type III secretion in low calcium cultures with Coomassie stained SDS-PAGE of culture supernatant samples (Fig. 3A) or by immunoblotting of culture supernatant and bacterial pellet samples (Fig. 3B). At a concentration of 10 μM IPTG, wild-type yopR-gst completely blocked the Yersinia type III pathway, whereas the E135K, E140K and E140K mutants were unable to impose a blockade even at 1 mM IPTG. To examine the ability of mutants to impose a type III blockade during Yersinia infection of tissue cultures, the expression of yopR-gst was induced with 1 mM IPTG as Y. enterocolitica W22703 harboring various plasmids were inoculated into HeLa cells. In contrast to uninfected HeLa cells or cultures where the type III pathway had been blocked by the expression of wild-type yopR-gst, neither the E135K, E140K nor the E140K mutant was able to cause a type III blockade, as evidenced by the rounding of HeLa cells and the rearrangement of rhodamine-phalloidin stained actin cables (Fig. 3C).
We assumed that mutations in yopR-gst, which reside in the yopR portion of the fusion and abolish the blocking phenotype, may also impact the recognition of YopR by the secretion apparatus. To investigate this, the Y. enterocolitica ΔyopR mutant was complemented with wild-type and mutant yopR alleles and, for consistency, we again used the E140K and E135K, E140K alleles. As outlined above, Yersinia species secrete YopR into the extracellular medium, where the concentration of calcium ions is high, but not into HeLa tissue cultures or immune cells, where calcium concentrations are low. We therefore assayed the type III secretion of YopR in the presence and absence of calcium, analyzing the abundance of secreted product on Coomassie stained SDS-PAGE or measuring secretion by immunoblotting of supernatant and pellet samples (Fig. 4A). In laboratory media, Y. enterocolitica secretes wild-type YopR in the presence and in the absence of calcium ions, whereas products of the E140K and E135K, E140K yopR mutants were only secreted in the absence of calcium, but not in the presence of 5 mM CaCl2. As controls, YopB was predominantly secreted in the absence of calcium, whereas the type III machine component YscD was not secreted (Fig. 4B)
Y. enterocolitica ΔyopR harboring plasmids that express wild-type or mutant yopR were able to promote type III effector injection of HeLa tissue culture cells (Fig. 4D). The culture medium of infected HeLa cells was decanted and centrifuged to separate proteins secreted by Yersinia into extracellular fluids from the bacterial sediment. Similar to wild-type parent Y. enterocolitica W22703 (Lee and Schneewind, 1999), the ΔyopR variant harboring plasmid pyopR secreted its YopR product into the conditioned medium of HeLa cultures (Fig. 4C). In contrast, ΔyopR variants harboring plasmids with the yopR mutants E140K or E135K, E140K failed to synthesize or secrete a YopR product during the infection of HeLa cells (Fig. 4C). The behavior of YopR in these experiments is consistent with the idea that codons 131–149 represent a secretion signal that is required under specific environmental conditions, namely when yersiniae have already assembled functional type III machines in body fluids but before they are engaged in the injection of effector Yops into immune cells.
In wild-type Y. enterocolitica, the gene encoding YopR is located on the pYV227 virulence plasmid where it is transcribed as part of the polycistronic virC message (Allaoui et al., 1995). Proteins encoded by genes upstream and downstream of yopR are required for assembly of the secretion apparatus, also named the injectisome (Cornelis, 2006). The preceding experiments therefore do not address how yopR mutations will affect the secretion of YopR expressed off its native mRNA. Secondly, the issue of silent mutations and their impact on YopR secretion demanded further investigation. We therefore engineered three of the mutants that we observed in our random mutagenesis onto pYV227. Two of these (E140K and P146I) represent missense mutations and one (I143I) is a silent mutation (Fig. 5A). For each of the three yopR mutations on pYV227, the mutant YopR products were not secreted in the presence of calcium or when yersiniae infected HeLa tissue culture cells (Fig. 5BC). The silent mutant I143I had the most drastic effect on YopR secretion, abolishing it completely and reducing synthesis (YopR was observed in pellet material that was concentrated 10x; data not shown). Moreover, as yopR mutations did not affect type III secretion of YopB or YopD (Fig. 5BC) or type III effector injection into HeLa cells (Fig. 5D), it seems reasonable to assume that the aforementioned mutations in yopR do not cause a general perturbation of the type III secretion pathway or of the products encoded by genes in the immediate vicinity of yopR (this was confirmed with immunoblots to YscJ; yscJ is the second gene downstream from yopR; data not shown).
The yopR mutations we examined here map to a region of the yopR mRNA that can be folded in silico into a secondary stem-loop structure (Mathews et al., 1999; Zuker, 2003) (Fig. 5A). In agreement with the possibility that this structure may contribute to signaling YopR secretion, two of the isolated yopR mutations (E140K and I143I) are predicted to disrupt a G:C base pair in the secondary structure (Fig. 5A). As a test for the possibility that the predicted mRNA structure and not the sequence per se harbors secretion information, the G:C base pair was inverted into a C:G base pair. The yopR mutant required to bring this about (E140Q, I143M) contains two lesions that—as predicted—enabled type III secretion of its YopR product when Yersinia were either grown in the presence of calcium ions or during infection of HeLa tissue culture cells (Fig. 5BC). The in silico folding of the mRNA structure of E140K and I143I or other relevant mutants revealed fundamental changes that perturb the entire predicted stem loop (Supplemental Fig. 1). The E140Q, I143M mutations, on the other hand, were able to restore the predicted stem loop structure of codons 131–149 observed in wild-type yopR. From these observations, two conclusions are borne out. First, physiological type III secretion of YopR by Yersinia requires an mRNA element located in the 3’ end of yopR, and second, a specific mRNA structure may be required for this to function.
One simple explanation for how an mRNA structure may contribute to secretion may be its ability to impart an ‘imprint’ such as a mis-incorporated or non-canonical amino-acid onto the final product of translation. Indeed, such post-translational modifications have been observed in the evolutionarily related secretion system for flagellar components. Campylobacter species do not secrete the flagellin unless it has been glycosylated (Goon et al., 2003). To test the possibility that secreted YopR carries a post-translational modification, we engineered Yersinia to express a yopR whose products carries an internal hexahistidine tag enabling purification of secreted YopR under ‘physiological’ conditions, i.e. from media supplemented with calcium (Fig. 6A). The expected mass of the primary translational product is 19,234.96 Da. Instead, we observed a mass of 19,104.10 as well as three additional compounds, which are explained as products of increased oxidation of 19,104.10 (+16, +32, +48 Da) (Fig. 6D). The observed mass of secreted YopR corresponds within 0.5 Da to a product lacking the initiator methionine and harboring three methionyl residues that are variably oxidized during purification. To confirm the absence of the initiator methionine, we subjected purified YopR to Edman degradation. The first cycle released phenylthiohydantoin-threonyl, the second amino-acid encoded by yopR and all subsequent residues confirmed the predicted YopR sequence (Fig. 6B). Purification and Edman-sequencing of YopR-GST mutants revealed that these also lacked the initiating methionine (data not shown). Moreover, the variant YopR(T2L)-GST, with a replacement of leucine for threonine at position 2, retained the initiator methionine as expected (Frottin et al., 2006) and this hybrid also retained the attribute of blocking the Yersinia type III secretion pathway (data not shown). We therefore conclude that while YopR is post-translationally modified by removal of its initiator methionine, this modification is unrelated to the function of the mRNA secretion signal in promoting Yersinia secretion in the presence of calcium or imposing a type III blockade when fused to an impassable reporter. Further, the mRNA signal located at codons 131–149 does not appear to impart a modification of the translational product and perhaps may exert its function by coupling translation and secretion (Anderson and Schneewind, 1997; Silhavy, 1997).
If YopR secretion necessitates the presence of two elements, positioned at codons 1–11 and 131–149, then what is the relative contribution of each? To address this question, we constructed a yopR-dhfr fusion where the first fifteen codons were exchanged for those of yopE. The N-terminal or 5’ secretion signal of yopE was selected for several reasons. The destiny of the YopE effector during infection is divergent, i.e. YopR is extracellular and YopE is intracellular. Further, on alignment, only one amino acid (serine at position 8) out of 15 residues is conserved between YopE and YopR (Fig. 7). Finally, earlier work suggested that features of yopE mRNA may contribute to secretion signaling (Anderson and Schneewind, 1997; Ramamurthi and Schneewind, 2005). Upon IPTG induction, YopE1–15YopR16–165-DHFR blocked the type III secretion of Yops just as effectively as YopR-DHFR (Fig. 7AB). It therefore appears that the N-terminal or 5’ signal secretion signal of YopR-DHFR can be exchanged with a heterologous signal and that the hybrid retains its blocking function. Whether YopR codons 131–149 are of themselves sufficient as a second secretion signal is a question we cannot answer at this time; protein abundance of relevant YopE-YopR hybrids was problematic (data not shown). In concert with other investigators, we speculate that N-terminal or 5’ signals are generic and perhaps universal to type III substrates and that secondary or 3’ signals direct protein products to their final destinations (Bröms et al., 2007; Rosqvist et al., 1995; Sorg et al., 2007).
It is not unusual for mRNA structures to regulate the transcripts they reside in; mRNA structures can affect transcript half-life, length or translational efficiency. For example, examination of the type III substrate FliC revealed mRNA structures around the initiating AUG which impact secretion indirectly; these stem loops impact translation (Aldridge et al., 2006). In the case of YopR, the initiation of translation is 400 bases upstream of the stem loop we have characterized and our data does not consistently or unequivocally suggest an expression component to its function.
The idea of an mRNA signal promoting protein secretion is not new. It was proposed after observing that the secretion signals in YopE1–15Npt and YopN1–15Npt can be scrambled by shifting the reading frame without affecting the export of the fusion protein (Anderson and Schneewind, 1997). The notion that 5’ codons contain an mRNA signal was buttressed by the observation that a silent mutation in YopE1–7Npt abolishes secretion of the fusion protein (Ramamurthi and Schneewind, 2005). All of these experiments assayed the secretion of fusion proteins into media containing calcium chelators, however the in vivo destination of YopE is within host cells and no fusion protein with less than YopE1–50 is observed in this compartment (Lee et al., 1998; Sory and Cornelis, 1994; Sory et al., 1995). The underlying rationale is that binding of a chaperone—SycE—to YopE15–100 is required for YopE injection into eukaryotic cells (Birtalan et al., 2002; Cheng et al., 1997; Cheng and Schneewind, 1999). YopE can therefore be thought of as containing two secretion signals, one embedded in N-terminal codons and another which encompasses the binding site for the SycE chaperone. Both signals must be present for YopE injection into HeLa cells to occur (Schesser et al., 1996).
The notion that secreted proteins require multiple, non-contiguous secretion signals makes sense given the complex developmental program that type III secretion represents with early, middle and late substrates (Cornelis, 2006). In all likelihood, each class of substrates fits like a ‘lock and key,’ with three-dimensional cues enabling secretion and changes in the lock to accommodate different classes of substrates. While many of the secreted substrates are known to possess chaperones which are required for their secretion (Wattiau et al., 1996), others do not. The obvious question has been, how do these substrates exit? The best-studied examples, YscP and now YopR, possess two secretion signals, one at the N-terminus and another further downstream (Agrain et al., 2005a; Riordan et al., 2008).
If substrates possess multiple secretion signals, what is their relative contribution? The genetic analysis of yopR-gst (Fig. 2) provides some answers for yopR. No missense mutations were observed in the secretion signal at codons 1–11 (Sorg et al., 2006). While this may be due to the limited repertoire of our mutagen, perhaps this is due to the mutability of these signals in general (Anderson and Schneewind, 1997; Bröms et al., 2007; Lloyd et al., 2001; Lloyd et al., 2002) and by the functionality of the yopE secretion signal in the context of lcrV (Sorg et al., 2007) or the yopR-dhfr fusion described here (Fig. 7). The second signal, located in the mRNA sequence of codons 131–149, is much more sensitive to mutagenesis, however it alone is not able to promote YopR secretion, which absolutely requires the N-terminal or 5’ signal. Thus, as in the case of YopE, the two signals cooperate and are mutually dependent on one another.
The ultimate goal of this field of study is to appreciate how the lock and key of secretion machine and substrates fit together over the course of a developmental program. Having unexpectedly discovered an mRNA element and the absence of a post-translational modification that would easily explain its contribution, we are once again left with the question: how does an mRNA signal promote the secretion of its encoded protein product? We can only invoke the ribosome and imagine that perhaps a translational pause (Nakatogawa and Ito, 2002) or localized protein synthesis (Aldridge et al., 2006) couple an mRNA structure to type III secretion (Silhavy, 1997).
Y. pestis KIM8 (Williams and Straley, 1998) and Y. enterocolitica W22703 (Cornelis and Colson, 1975) were used for all Yersinia work and grown in BHI or TSB, respectively. Plasmids and techniques for allelic replacement have been described previously (Cheng et al., 1997). In brief, a 2 kb fragment of pYV centered on yopR was cloned into the suicide plasmid pLC28 and modified appropriately before being mobilized into Yersinia. For the generation of ΔyopR, the 8 bp sequence TAAGATCT was inserted after the seventh codon of yopR, resulting in a stop codon at position 8. yopR alleles P146I, E140K, I143I, and E14Q & I143M were the result of modification of the yopR-containing suicide plasmid with primers 5’-GAATTAATGATCTTATTAttGTATAACTCGATTGTAGA, 5’-TCAGAGTTAATAAATAAAaAATTAATGATCTTATTAC, 5’-ATAAATAAAGAATTAATGATTTTATTACCGTATAACTCG and 5’-CAGAGTTAATAAATAAAcAATTAATGATgTTATTACCGTATAACTC respectively (and their reverse complements). Lower case letters in the primers indicate the site of the lesion(s) created. Creation of YopE-YopR-DHFR was accomplished by PCR priming into yopR with a primer containing the first fifteen codons of yopE (5’-AACATATGAAAATATCATCATTATTTCTACATCACTGCCCCTGCCGGCATCGTCTCA GGCAGTCTCTACG) with 5’-AAAGATCTTGTATCCATATCAATTTGATGGCTGTTATGAA and repeating the cloning which created the original YopR-DHFR fusion, pJS95 (Sorg et al., 2006). The same cloning was repeated for the construction of YopR1–50, 1–100, and 1–150-DHFR, YopR was amplified with 5’-AACATATGACGGTTACCCTTAATAGAGGTTCCAT and : 5’-AAAGATCTCAGAACTTCCCGTGTCTTTTCA, 5’-AAAGATCTCAATACACTACGCAATTCAGGTAAATCTA, or 5’-AAAGATCTAATCGAGTTAGACGGTAATAGGATCA respectively. YopR-LacZ was generated by amplification of lacZ from E. coli K12 with primers 5’-AAAGATCTATGACTATGATTACGGATTCT and 5’-AATCTAGATTATTTTTGACACCAGACCAA, which was digested with BglII and XbaI, as was the vector pJS116 which contains a BglII site just downstream of Ptac-yopR (Sorg et al., 2006). To assess the secretion of YopR in Yersinia, Ptac-yopR was amplified from pJS116 (Sorg et al., 2006) or the mutant derivatives from our mutagenesis with primers 5’-AAGGATCCCACTGCATAATTCGTGTCGCTC and 5’-AACTCGAGTTATGTCTCCATATCAATTTG and digested with BamHI and XhoI; the medium-copy vector pCDF (Invitrogen) was digested with BglII and XhoI and the fragments ligated. A hexahistidine tag was added after the fortieth residue of YopR by taking the plasmid encoding Ptac-yopR and amplifying the plasmid with 5’-GATCATAAGCTAAAAAGTcaccatcaccatcaccatGAATCCGCTGAAAAGAC and its reverse complement.
Plasmid DNA was mixed with an equal volume of 1 M hydroxylamine and incubated at 68°C for 30–60 minutes. DNA was precipitated and washed to remove hydroxylamine and used for transformation of Y. pestis. Transformants were patched onto BHI and after a day replica-plated onto two plates containing BHI with 20 mM MgCl2, 20 mM sodium oxalate and 1 mM IPTG; one plate was incubated at 28°C while the other was at 37°C. Plates were scored for the low calcium response one day later (Sorg et al., 2005a).
One liter of Y. enterocolitica culture expressing histidine-tagged YopR was incubated at 37°C in the presence of 5 mM calcium chloride. After three hours, the bacteria were sedimented by centrifugation at 6,000 ×g for 10 minutes and supernatants were filter-sterilized. Proteins in the filtrate were precipitated with 600 g ammonium sulfate at 4°C overnight. Proteins were sedimented by centrifugation at 6,000 ×g for 60 minutes and subsequently dialysed against 150 mM NaCl, 50 mM Tris-HCl (pH 8.0), 10 mM imidazole. The sample was applied to Ni-NTA sepharose (Qiagen), washed and eluted with dialysis buffer containing 500 mM imidazole. The eluate was further purified by reversed phase high-pressure liquid chromatography on a PLRP-S column (Varian) and eluate was subjected to mass spectrometry at the Pasarow Mass Spectrometry Laboratory, UCLA. For Edman degradation, this same material was separated on SDS-PAGE, transferred to PVDF, and stained with Amido black. Edman degradation was performed at the University of Illinois, Urbana Champaign protein sciences facility.
Secretion assays, HeLa cell manipulations and infections and rhodamine-phalloidin staining for microscopy are standard assays which have been described in detail elsewhere (Cheng et al., 1997; Lee et al., 1998; Sorg et al., 2006). Briefly, secretion assays are conducted at 37°C in M9 medium (Cheng et al., 1997) supplemented with 1 mM IPTG, 5 mM CaCl2 or 5 mM EGTA wherever appropriate. To monitor secretion in the presence of HeLa cells, tissue culture flasks containing ~107 HeLa cells are infected at an MOI of 10. After 3 hours, the media bathing the HeLa cells (DMEM with 1 mM IPTG wherever necessary) is decanted, and centrifuged at 6,000×g. This pellet contains non-adherent bacteria. 7 ml of supernatant is precipitated with 7 ml methanol and 1.75 ml chloroform and the sediment washed with methanol. This pellet contains secreted proteins. For microscopy of HeLa cells, bacteria were mixed at an MOI of 10 and incubated for 3 hours. The samples were fixed with 3.7% formaldehyde for twenty minutes, permeabilized with 0.1% Triton X-100 for 30 minutes and labeled with 3 units rhodamine-phalloidin (Invitrogen) for 20 minutes. Samples were washed four times with PBS and viewed as DIC and fluorescence microscopy images.
We thank Vincent Lee and Mailin Chu for the creation of the yopR strain and Kym F. Faull for help with the mass spectrometry experiments. B.B. and O.S. acknowledge membership in and support from the RegionV “Great Lakes” Regional Center of Excellence in Biodefenseand Emerging Infectious Diseases Consortium (NIH award 1-U54-AI-057153).