The Distal ExsA-Binding Site in Pseudomonas aeruginosa Type III Secretion System Promoters Is the Primary Determinant for Promoter-Specific Properties

Abstract

Transcription of the Pseudomonas aeruginosa type III secretion system is controlled by ExsA, a member of the AraC/XylS family of regulators. Each ExsA-dependent promoter contains two adjacent binding sites for monomeric ExsA. The promoter-proximal site (binding site 1) consists of highly conserved GnC and TGnnA sequences that are individually recognized by the two helix-turn-helix (HTH) DNA-binding motifs of an ExsA monomer. While the GnC and TGnnA sequences are important for binding to site 1, the promoter-distal binding sites (site 2) lack obvious similarity among themselves or with binding site 1. In the present study, we demonstrate that site 2 in the P_exsC promoter region contains a GnC sequence that is functionally equivalent to the GnC in site 1 and recognized by the first HTH motif of an ExsA monomer. Likewise, the second HTH interacts with an adenine residue in binding site 2. Although several candidate GnC sequences are also present in site 2 of the P_exsD, P_exoT, and P_pcrG promoters, the GnC sequences were not required for ExsA-dependent transcription or ExsA binding. A comparison of hybrid promoters composed of binding site 2 from one promoter fused to binding site 1 derived from another promoter indicates that ExsA-binding affinity, promoter strength, and the degree of promoter bending are properties that are largely determined by binding site 2. Based on these data, we propose that the manner in which ExsA interacts with binding site 2 at the P_exsC promoter is distinct from the interactions occurring at other promoters.

INTRODUCTION

The Pseudomonas aeruginosa type III secretion system (T3SS) is an important virulence determinant used to translocate effector proteins into cells where they exert antihost properties by promoting phagocytic avoidance, tissue destruction, impaired wound healing, and dissemination (10). Expression of the T3SS is induced in response to calcium-limiting growth conditions or contact of P. aeruginosa with host cells through a signaling cascade that controls the activity of ExsA (6, 18). ExsA is the primary transcriptional activator of T3SS gene expression and a member of the AraC/XylS family of DNA-binding proteins (9). A total of 10 ExsA-dependent promoters are known to control expression of the ∼40 gene products that constitute the T3SS (8).

The DNA-binding properties of ExsA have been studied in vitro using purified amino-terminal histidine-tagged fusion protein (ExsA_His), which is monomeric in solution (3). Two distinct binding properties of ExsA_His are observed in electrophoretic mobility shift assays (EMSA) when using promoter probes derived from the P_exoT, P_exsD, P_exsC, and P_pcrG promoters. First, binding of ExsA_His to the promoter probes elicits DNA bending, the degree of which is promoter specific and ranges from 20° for the P_exoT promoter to 78° for the P_exsC promoter (3). The biological significance of promoter bending by ExsA is unclear. The second binding property observed by EMSA is that ExsA_His forms two distinct complexes (shift products 1 and 2) upon binding to each of the promoter probes (3; also see our accompanying paper, reference 12). Whereas the higher-mobility shift product 1 represents one bound ExsA_His monomer, the lower-mobility shift product 2 results from the binding of two ExsA_His monomers. Binding of ExsA_His to the P_exoT promoter occurs in an ordered fashion whereby occupation of binding site 1 is required for efficient occupation of site 2 (3). This cooperative binding mechanism is dependent upon monomer-monomer interactions mediated through the amino-terminal domain of ExsA (4). Although it is unclear whether ExsA binds to the other promoters in the same ordered fashion, strong cooperative binding of ExsA_His also occurs at the P_exsC promoter (3).

Mutagenesis and footprinting studies indicate that the two ExsA monomers bind to promoter regions in the same orientation (head-to-tail) at adjacent binding sites which are centered at the −44 (binding site 1) and −65 (binding site 2) positions relative to the transcriptional start site (3, 12). This binding arrangement is typical of class II activators, which bind in the forward orientation near the −35 region and typically recruit RNA polymerase holoenzyme through interactions with region 4.2 of the sigma subunit (2, 13, 14). ExsA activates transcription primarily by recruiting RNA polymerase to the promoter through interactions with region 4.2 of sigma, but a small role in facilitating isomerization to an open transcriptional complex has also been proposed (16, 17).

The defining feature of AraC/XylS proteins is the presence of two conserved helix-turn-helix (HTH) DNA-binding motifs (14). The second helix in each HTH motif functions as a recognition helix that inserts into adjacent major grooves of the DNA to make base-specific contacts with promoter elements separated by ∼10 bp (sometimes referred to as A- and B-box sequences). The recognition helices of an individual ExsA monomer, one from each HTH motif, are designated RH1 and RH2 (12). Base-specific contacts mediated by RH1 and RH2 have been defined for binding site 1 in the P_exoT promoter. The P_exoT promoter, as do all ExsA-dependent promoters, contains three highly conserved regions: a 5-bp adenine-rich region centered at −55 relative to the transcriptional start site a GnC sequence centered at −46, and a TGnnA sequence centered at the −36 position (3). Nucleotide substitutions in each of these conserved regions of P_exoT result in a significant decrease in ExsA_His binding in vitro and ExsA-dependent transcription in vivo. Although the role of the adenine-rich region has not been established, the GnC and TGnnA sequences constitute binding site 1 and are recognized by ExsA through interactions mediated by RH1 and RH2, respectively (12). Because the GnC and TGnnA sequences are conserved in binding site 1 of all known ExsA-dependent promoters, it has been proposed that the mechanistic basis for occupation of binding site 1 is similar at each promoter (12).

The requirements for occupation of binding site 2 are less clear, since this region shares no obvious sequence similarity between the different ExsA-dependent promoters or to binding site 1 (3). Footprinting data obtained with Fe-BABE (a protein-labeling reagent that can be conjugated to cysteine residues), however, indicate that RH1 of the ExsA monomer bound to site 2 interacts with the −70 to −75 region of the P_exoT, P_exsD, P_exsC, and P_pcrG promoters and that RH2 interacts with the −60 region of the P_exsC and P_pcrG promoters (12). An interaction between RH2 and the P_exoT and P_exsD promoters at binding site 2 was not detected. In the present study, we examine the requirements for recognition of binding site 2 by ExsA. As in binding site 1, occupation of binding site 2 in the P_exsC promoter also involves interactions between RH1 and a GnC sequence. Each of the candidate GnC sequences in binding site 2 of the P_exoT, P_exsD, and P_pcrG promoters, however, was dispensable for ExsA binding and reporter activity in vivo. This finding suggested that ExsA interacts with binding site 2 in at least two distinct ways. To determine whether these differences in binding were biologically significant, we constructed hybrid promoters and report that binding site 2 is the primary determinant for promoter activity, ExsA affinity, and the degree of promoter bending observed upon ExsA binding.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The bacterial strains, transcriptional reporters, expression plasmids, and sequences of the oligonucleotide primers used in this study are summarized in Tables S1 and S2 in the supplemental material. Escherichia coli DH5α was used for all cloning and maintained on Luria-Bertani agar plates with gentamicin (15 μg/ml) or ampicillin (100 μg/ml) as necessary. P. aeruginosa strains were cultured on Vogel-Bonner minimal medium (19) agar plates with 100 μg/ml gentamicin as required. For transcriptional reporter assays, P. aeruginosa strains were cultured in Trypticase soy broth supplemented with 100 mM monosodium glutamate, 1% glycerol, and 2 mM EGTA to an absorbance (A₆₀₀) of 1.0. β-Galactosidase activity was determined as previously described (5). The reported values for the β-galactosidase assays (Miller units) throughout this study represent the average of at least three independent experiments, and error bars represent the standard error of the mean (SEM). Two-tailed unpaired t tests were performed using Prism 5 GraphPad (GraphPad Software, Inc., La Jolla, CA).

Mutagenesis.

Unless stated otherwise, the P_exoT, P_exsC, P_exsD, and P_pcrG promoter mutants were constructed using a two-step PCR method with the oligonucleotide primer pairs indicated in Table S1 in the supplemental material to generate megaprimers (3). The megaprimers were then used in a second PCR with one of the following promoter-specific primers: 27978335 for P_exsC, 31216338 for P_exoT, 28433494 for P_pcrG, or 25843026 for P_exsD (Table S2). The resulting PCR products were cloned as HindIII/EcoRI restriction fragments into plasmid mini-CTX-lacZ and integrated onto the PA103 chromosome as previously described (1, 11) or used as PCR templates to generate DNA probes for EMSA experiments using the primer pairs indicated for each wild-type (wt) promoter. The P_exsC promoter truncations were amplified in a single-step PCR using the primer pairs indicated in Table S1 and cloned into mini-CTX-lacZ as described above. The promoter hybrids were constructed by PCR amplification of the site 1 or site 2 donors using the primer pairs indicated in Table S1. The resulting PCR products contained complementary ends that were then annealed and used in a second PCR with promoter-specific primers located at the 5′ and 3′ boundaries of each promoter region as defined in Table S1.

Binding assays.

EMSA experiments were performed using nonspecific (a 160-bp region of pscF) and specific DNA probes generated by PCR and end labeled with ³²P as previously described (3). EMSA reaction mixtures (19 μl) containing specific and nonspecific probes (0.05 to 0.25 nM each, as indicated), 25 ng/μl poly(dI·dC) (Sigma-Aldrich, St. Louis, MO), and ExsA DNA binding buffer (10 mM Tris [pH 7.9], 50 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 5% glycerol, and 100 μg/ml bovine serum albumin) were incubated for 5 min at 25°C. ExsA_His (concentrations as indicated in the figure legends) was added for a final reaction volume of 20 μl and incubated for 15 min at 25°C. Samples were subjected to electrophoresis on a 5% polyacrylamide glycine gel (10 mM Tris [pH 7.5], 380 mM glycine, 1 mM EDTA) at 4°C. Imaging was performed using an FLA-7000 phosphorimager (Fujifilm) and Multigage v3.0 software (Fujifilm) for data analyses.

RESULTS

The GnC and TGnnA sequences in binding site 1 of the P_exsC promoter are required for maximal promoter activity and binding by ExsA.

Previous studies of the P_exoT promoter found that the GnC and TGnnA sequences in binding site 1 are required for maximal promoter activity in vivo and ExsA binding in vitro (3). To determine whether the same is true of the P_exsC promoter, we generated reporter fusions carrying nucleotide substitutions at the GnC and TGnnA positions in binding site 1 (reporters EB215 to EB220 [herein referred to collectively as EB215-EB220, etc.], summarized in Fig. 1A). The EB217 reporter, which carries a substitution at the noncritical −44 position, was included as a negative control. It should be noted that the GnC and TGnnA sequences are displaced by two nucleotides in P_exsC relative to P_exoT due to differences in the transcriptional start sites (i.e., the GnC corresponds to −47 and −45 in P_exoT and to −49 and −47 in P_exsC). The reporters were integrated at the CTX phage attachment site, and the resulting strains were cultured under noninducing (−EGTA, high-Ca²⁺) and inducing (+EGTA, low-Ca²⁺) conditions for T3SS gene expression and assayed for β-galactosidase activity. Under noninducing conditions, the EB215-EB216 and EB218-EB220 reporters demonstrated significant reductions in activity (<21% activity relative to the wt P_exsC reporter) (Fig. 1B). The same general trend was observed under inducing conditions, although the expression defects were of lesser magnitude than those observed under noninducing conditions. The latter finding likely reflects the fact that the amount of ExsA available for binding is more limiting under noninducing conditions because ExsA is sequestered in a complex with the ExsD antiactivator. As expected, the EB217 substitution had only a modest effect (∼75% of wt) on promoter activity.

Fig 1.

The GnC sequence in binding site 1 of the P_exsC promoter is required for maximal ExsA-dependent transcription and ExsA binding. (A) Map of the P_exsC promoter showing the conserved GnC and TGnnA sequences in binding site 1 and the GnC sequence in binding site 2 (bold with larger typeface). The solid line below the DNA sequence represents the promoter region protected by ExsA from DNase I cleavage on the bottom strand (3). Upward-pointing arrows indicate the nucleotide positions subjected to mutagenesis in binding site 1 that were used to generate reporters EB215-EB220 or in binding site 2, resulting in reporters EB204-EB211 (analyzed in Fig. 2B). The nature of the nucleotide substitutions is indicated above the arrows. The dashed arrows indicate the location of the truncated reporter fusions (reporters EB201-EB203 and EB225-EB226) analyzed in Fig. 2A. (B) PA103 exsA::Ω strain carrying the indicated lacZ transcriptional reporters were cultured under noninducing (−EGTA, hatched bars) or inducing (+EGTA, open bars) conditions for T3SS gene expression and assayed for β-galactosidase activity (*, P < 0.01). (C) Nonspecific (Non-SP) (derived from the pscF coding sequence) and specific (SP) end-labeled probes (0.05 nM) derived from the native P_exsC-lacZ reporter or the EB215-EB220 reporters were incubated in the absence (−) or presence (+) of 5 nM ExsA_His for 15 min at 25°C. Samples were analyzed by native polyacrylamide gel electrophoresis and phosphorimaging. The positions of shift products (Shift) 1 and 2 are indicated.

To corroborate our in vivo findings, we also performed EMSA experiments using specific promoter probes (202 bp) derived from the wt P_exsC reporter or the EB215-EB220 reporters and a nonspecific probe derived from the coding portion of pscF (160 bp) which served as a specificity control. The specific and nonspecific promoter probes were incubated with purified histidine-tagged ExsA (ExsA_His) for 15 min at 25°C and immediately subjected to native polyacrylamide gel electrophoresis and phosphorimaging. Consistent with a previous study (3), incubation of ExsA_His with the wt P_exsC promoter probe generated two distinct complexes: a higher-mobility complex (shift product 1) that represents one bound ExsA_His monomer and a lower-mobility complex (shift product 2) that represents two ExsA_His monomers bound to sites 1 and 2 (Fig. 1A and C). Similarly, binding of ExsA_His to the EB215-EB220 promoter probes also generated shift products 1 and 2, although the yield of both shift products was reduced in each case except that of EB217, which was only modestly reduced. These data demonstrate that the GnC and TGnnA sequences in binding site 1 are required for maximal activity of the P_exsC promoter and binding by ExsA.

The GnC sequence in binding site 2 of the P_exsC promoter is functionally equivalent to the GnC sequence in binding site 1.

Previous Fe-BABE footprinting experiments showed that RH1 interacts with the upstream half of binding site 1 (−50 region) and binding site 2 (−70 region) in the P_exsC, P_pcrG, P_exoT, and P_exsD promoters (12). Since ExsA binds in a head-to-tail orientation, and RH1 interacts with the GnC sequence in binding site 1, we hypothesized that RH1 of the ExsA monomer bound to site 2 also recognizes a GnC sequence. We focused our initial experiments on the P_exsC promoter, since it contains a single GnC sequence within the region previously shown to encompass binding site 2 by DNase I footprinting (Fig. 1A) (3). To define the minimal P_exsC promoter region and determine the relative importance of binding site 2, we constructed a series of transcriptional reporters bearing incremental deletions between bp −95 and −61 relative to the transcriptional start site (Fig. 1A). Whereas promoter fusions containing deletions upstream of bp −75 (EB226, EB225, and EB202) had activity comparable to that of the wt P_exsC-lacZ promoter fusion (−189 to +13), both reporter fusions lacking the putative GnC sequence in site 2 (EB203 and EB201) had significantly less activity (Fig. 2A). Thus, nucleotides located between the −67 and −75 positions are required for maximal promoter activity.

Fig 2.

The GnC sequence in binding site 2 of P_exsC is required for maximal promoter activity and binding by ExsA. (A and B) PA103 carrying the indicated reporter fusions (previously summarized in Fig. 1A) was cultured under noninducing (−EGTA) and inducing (+EGTA) conditions for T3SS gene expression and assayed for β-galactosidase activity (*, P < 0.05; **, P < 0.01). (C) Nonspecific (Non-SP) and specific (SP) end-labeled probes (0.05 nM) derived from the native P_exsC reporter or the EB204-EB211 reporters were incubated in the absence (−) or presence (+) of 5 nM ExsA_His for 15 min at 25°C. Samples were analyzed by native polyacrylamide gel electrophoresis and phosphorimaging. The positions of shift products (Shift) 1 and 2 are indicated.

To determine which nucleotides in the −75 to −67 region are important for promoter activity, we generated P_exsC-lacZ reporters carrying nucleotide substitutions at each position (reporters EB204-EB211, summarized in Fig. 1A). Under noninducing conditions, several of the mutant reporters (EB204, EB206-EB207, and EB210) demonstrated a significant increase in activity for reasons that are not clear. Only the EB209 and EB211 reporters, however, demonstrated a signification reduction in activity, indicating that the primary determinants in the −75 to −67 region are nucleotides G−70 and C−68 (Fig. 2B). Under inducing conditions, however, the EB209 and EB211 reporters had activity comparable to that of the native P_exsC reporter; as noted above, this likely reflects the increased concentration of ExsA available for DNA binding under inducing conditions. To examine this further, we performed EMSA studies and found that binding of ExsA_His to the EB209 and EB211 promoter probes was significantly reduced (Fig. 2C). These combined data indicate that the GnC sequence in binding site 2 of the P_exsC promoter is required for maximal activity and binding by ExsA.

In a previous study, we isolated gain-of-function mutants in ExsA that allowed activation of P_exoT-lacZ reporters bearing single-nucleotide substitutions in the GnC or TGnnA sequences at binding site 1 (3). For example, an ExsA T199S mutant was significantly impaired for activation of the wt P_exoT-lacZ reporter (12-fold lower than wt ExsA) but was 3-fold better than wt ExsA for activation of a P_exoT-lacZ reporter carrying the AnC sequence substituted for GnC. The two ExsA gain-of-function mutants identified for the GnC sequence and the cognate P_exoT promoter sequences used in their isolation are summarized in Fig. 3A.

Fig 3.

The GnC sequences located in bindings sites 1 and 2 of P_exsC are functionally equivalent. (A) Sequence of the P_exsC-lacZ reporter substitutions in binding sites 1 and 2 which are synonymous to those used in the P_exoT promoter to isolate the T199S, K202R, and T252S gain-of-function mutants as previously described (12). (B and C) PA103 exsA::Ω strain carrying the native P_exsC-lacZ reporter or the indicated mutant reporters and expressing either ExsA (wt), the T199S mutant (B), or the K202R mutant (C) were cultured under inducing conditions for T3SS gene expression and assayed for β-galactosidase activity (*, P < 0.05; **, P < 0.01).

Since the GnC and TGnnA sequences in binding site 1 are conserved in all ExsA-dependent promoters, we hypothesized that the T199S and K202R gain-of-function mutants would possess similar activities toward P_exsC-lacZ reporters carrying the synonymous P_exoT promoter substitutions in binding site 1. Indeed, as previously found with the P_exoT promoter, the T199S mutant demonstrated reduced activation of the wt P_exsC reporter and increased activation of the cognate reporter carrying the G−49A substitution (Fig. 3B). The finding with the K202R mutant, however, was not as clear, as the K202R mutant demonstrated reduced activation of the native P_exsC promoter but similar activity against the C−47A reporter (Fig. 3C). Nevertheless, the T199S and K202R gain-of-function mutants displayed the same general trend with respect to both the P_exoT and P_exsC promoters and proved to be useful tools for subsequent analyses of the P_exsC promoter (described below).

The T199S and K202R mutants were used to test the hypothesis that the GnC sequences in binding sites 1 and 2 of the P_exsC promoter serve a similar role with regard to ExsA-dependent transcriptional activation. In the first experiment, G-to-A substitutions were introduced at positions −49 and/or −70 in the P_exsC-lacZ reporter, and the resulting strains were tested for activation by wt ExsA and T199A. Whereas wt ExsA demonstrated a 2- to 3-fold reduction in activation of the reporters with single-nucleotide substitutions at positions −49 and −70 (P_{exsC(G−49A)-lacZ} and P_{exsC(G−70A)-lacZ}), activation of the reporter carrying both substitutions (P_{exsC(G−49A,G−70A)-lacZ}) was completely abolished (Fig. 3B). In contrast, the T199A mutant demonstrated the completely opposite pattern, being completely defective for activation of the wt P_exsC reporter but capable of activating the mutant P_{exsC(G−49A)-lacZ}, P_{exsC(G−70A)-lacZ}, and P_{exsC(G−49A,G−70A)-lacZ} reporters. Expressed quantitatively, the T199S gain-of-function mutant was 12-fold better at activating the P_{exsC(G−49A,G−70A)-lacZ} reporter when compared to the wt P_exsC-lacZ and nearly 30-fold more active against the P_{exsC(G−49A,G−70A)-lacZ} reporter compared to wt ExsA. Taken together, these data indicate that the ExsA monomers bound to sites 1 and 2 both utilize residue T199 to make base-specific contacts with the −49 and −70 guanines, respectively.

Applying the same strategy to the K202R mutant, we introduced C-to-A substitutions into the P_exsC-lacZ reporter at positions −47 and/or −68. Similar to the findings presented above, wt ExsA was able to activate the P_{exsC(C−47A)-lacZ} and P_{exsC(C−68A)-lacZ} single mutant reporters but was completely defective for activation of the P_{exsC(C−47A,C−68A)-lacZ} reporter bearing both substitutions (Fig. 3C). In contrast, activation of the P_{exsC(C−47A,C−68A)-lacZ} reporter by the K202R mutant was 20-fold higher than with wt ExsA, indicating that RH1 interacts with the −47 and −68 cytosines in binding sites 1 and 2 via base-specific contacts mediated by residue K202.

RH2 interactions at binding site 2.

Since RH1 contacts the GnC sequence in binding site 2 of the P_exsC promoter, it seemed probable that RH2 of the same ExsA monomer would contact a TGnnA sequence. Although no perfect match to the TGnnA sequence exists at the expected position in the P_exsC promoter, the adenine at −57 is separated by 10 bp from the GnC sequence (Fig. 4A). This same 10-bp spacing arrangement is common to binding site 1 in all of the ExsA-dependent promoters (Fig. 4A). To determine whether the adenine at positions −57 and −36 in binding sites 2 and 1 are functionally equivalent, we took advantage of the T252S gain-of-function mutant in RH2 (Fig. 3A) by introducing A-to-C substitutions individually at either position and in combination. Consistent with our previous findings for T252S when using the P_exoT promoter (12), the T252S mutant demonstrated reduced activation of the native P_exsC reporter, increased activation of the mutant reporter carrying the A−36C substitution and, as predicted by the spacing from the GnC sequence, increased activation of the reporter carrying the A−57C substitution (Fig. 4B). Although the reporter carrying both the A−36C and A−57C substitutions showed greatly reduced promoter activity compared to either single substitution, the residual activity was significantly more responsive to the T252S mutant than to wt ExsA. Importantly, the gain of function by T252S is specific to position −57, since A-to-C substitutions at the −58, −56, and −55 positions alone demonstrated reduced activity with T252S compared to that seen for wt ExsA. Similarly, the reporters carrying the A−36C substitution were largely unresponsive to activation by the T252S mutant when A-to-C substitutions were introduced into binding site 2 at positions −58, −56, and −55 (Fig. 4B).

Fig 4.

The adenines in the TGnnA sequence at binding sites 1 and 2 of the P_exsC promoter are functionally equivalent. (A) Map of the P_exsC promoter region showing the GnC sequences in binding sites 1 and 2 (bold with larger typeface), and the 10-bp spacing between each GnC sequence and the adenines at positions −36 and −57. (B) PA103 exsA::Ω strain carrying the indicated reporter fusions and expressing either wt ExsA (open bars) or the T252S mutant (hatched bars) was cultured under inducing conditions for T3SS gene expression and assayed for β-galactosidase activity (*, P < 0.05). For these experiments, the cells were cultured in the presence of 0.1% arabinose to induce expression of wt ExsA and the T252S mutant from the arabinose-inducible expression vector pJN105.

The GnC sequences in binding site 2 of the P_exsD, P_exoT, and P_pcrG promoters are not required for activation or ExsA binding.

The identification of a functional GnC sequence in binding site 2 of the P_exsC promoter prompted us to examine whether other ExsA-dependent promoters possess the same determinant. Given the high G+C content of the P. aeruginosa genome (67%) (15), it was not surprising that several candidate GnC sequences are present in the footprinted region encompassing binding site 2 of the P_exsD, P_exoT, and P_pcrG promoters (3, 4, and 2 potential GnC sequences, respectively) (Fig. 5A). Although the GnC sequences in binding sites 1 and 2 are centered 21 bp apart in the P_exsC promoter (Fig. 4A), none of the putative GnC sequences in the other promoters possess the same spacing arrangement. To determine whether the GnC sequences in binding site 2 are important for ExsA-dependent activation of the P_exsD-lacZ, P_exoT-lacZ, and P_pcrG-lacZ reporters, we generated AnA substitutions at each GnC sequence. None of the substitutions, however, negatively impacted transcriptional activity in vivo (Fig. 5B, D, and F) or ExsA_His binding to the promoter probes in EMSA experiments (Fig. 5C, E, and G). A possible exception is the EB401 substitution in the P_pcrG-lacZ reporter, which showed a modest decrease in binding. We also examined the possibility that the GnC sequences were functionally redundant by simultaneously mutating each GnC sequence (reporters EB309, EB153, and EB405 for P_exsD, P_exoT, and P_pcrG, respectively) and again observed no significant effect on transcriptional activation (Fig. 5B, D, and F) or DNA binding by ExsA_His (Fig. 5C, E, and G). We conclude that the manner in which ExsA interacts with binding site 2 at the P_exsD, P_exoT, and P_pcrG promoter differs from that at P_exsC, which is reliant on the GnC sequence in binding site 2 for maximal activation.

Fig 5.

The GnC sequences located within site 2 of the P_exsD, P_exoT, and P_pcrG promoters are not required for transcription or ExsA binding. (A) Diagram of the P_exsD, P_exoT, and P_pcrG promoters showing the location of each candidate GnC sequence within binding site 2 (bold with larger typeface) and the reporters resulting from the GnC-to-AnA substitutions. (B, D, and F) PA103 carrying the native P_exsD-lacZ (B), P_exoT-lacZ (D), or P_pcrG-lacZ (F) transcriptional reporters or the indicated mutant reporters was cultured under noninducing (−EGTA, open bars) or inducing (+EGTA, hatched bars) conditions for T3SS gene expression and assayed for β-galactosidase activity. The EB309, EB153, and EB405 reporters combine each of the GnC-to-AnA substitutions described for EB303-EB305, EB146-EB149, and EB401-EB402, respectively. (C, E, and G) Nonspecific (Non-Sp) and specific (Sp) end-labeled probes (0.05 nM) derived from the native P_exsD-lacZ (C), P_exoT-lacZ (E), and P_pcrG-lacZ (G) reporters, and the indicated mutant derivatives of each, were incubated in the absence (−) or presence (+) ExsA_His (20 nM for C and E, and 10 nM for G) for 15 min at 25°C. Samples were analyzed by native polyacrylamide gel electrophoresis and phosphorimaging. The positions of shift products (Shift) 1 and 2 are indicated.

Binding site 2 determines ExsA binding properties and promoter bending.

Since the GnC and TGnnA sequences in binding site 1 are conserved in all ExsA-dependent promoters, we hypothesized that ExsA binds to site 1 through a common mechanism and that binding site 1 from one promoter could be substituted with the corresponding site from another promoter. Conversely, differences in interactions at site 2 may be responsible for previously observed differences in promoter properties such as ExsA affinity, the nature of cooperative binding, the degree of promoter bending, and promoter activity (3). To test these predictions, we constructed a series of hybrid promoters in which binding site 2 from one promoter was fused to binding site 1 from another using the conserved A-rich region as the junction point (see Fig. S1 in the supplemental material). The first hybrid promoters examined were fusions between P_exoT and P_exsC. These two promoters differ in the following ways, each of which can be seen in vitro through EMSA experiments or in vivo using transcriptional reporter assays (3): (i) ExsA binds with higher affinity to P_exsC than to P_exoT (Fig. 6A); (ii) although evidence of cooperative binding exists for each promoter, cooperative binding is more evident at the P_exsC promoter where the abundance of shift product 2 far exceeds the amount of shift product 1 in EMSA experiments; (iii) ExsA binding elicits a significant bend (∼78°) in the P_exsC promoter probe, thereby resulting in increased retardation of shift product 2 when examined by EMSA; and (iv) a P_exsC-lacZ transcriptional reporter has higher activity under both noninducing and inducing conditions for T3SS gene expression (Fig. 6B). Examination of the P_exsC2-exoT1 and P_exoT2-exsC1 promoter hybrids revealed that the P_exsC2-exoT1 hybrid adopted the characteristics of P_exsC, whereas P_exoT2-exsC1 more closely resembled P_exoT (Fig. 6A and B). To examine this further, we expanded the analyses to the P_exsD and P_pcrG promoters, which share the same general properties described above for P_exoT and P_exsC, respectively, with the exception that bending of the P_pcrG promoter has not been examined. Once again, the resulting hybrid promoters adopted the properties of the site 2 donor (Fig. 6C to F). From these combined data, we conclude that ExsA affinity/cooperative binding, promoter bending, and transcriptional activity are largely determined by binding site 2 at each promoter and by inference that ExsA interacts with binding site 2 in at least two different ways.

Fig 6.

Binding site 2 of ExsA-dependent promoters is the primary determinant for promoter activity, ExsA binding affinity, and promoter bending. (A, C, E) End-labeled probes (0.05 nM) derived from the indicated promoters or promoter hybrids were incubated in the absence (−) or presence of increasing concentrations of ExsA_His (5, 15, 45, 135 nM) for 15 min at 25°C. Samples were analyzed by native polyacrylamide gel electrophoresis and phosphorimaging. The positions of shift products 1 and 2 are indicated. (B, D, and F) PA103 carrying the indicated transcriptional reporters was cultured under noninducing (−EGTA, open bars) or inducing (+EGTA, hatched bars) conditions for T3SS gene expression and assayed for β-galactosidase activity.

DISCUSSION

Members of the AraC/XylS family regulate a wide array of cellular processes ranging from basic metabolism to virulence (7, 14). We have been using P. aeruginosa ExsA as a model system to understand how AraC/XylS family members control T3SS gene expression, with specific interest in regulatory systems that govern ExsA activity, the mechanism of transcriptional activation, and the DNA-binding properties of ExsA. Previous investigations found that the highly conserved GnC and TGnnA sequences located in binding site 1 of the P_exoT promoter region are required for maximal binding and activation by ExsA (3). Since there is near-perfect conservation of the GnC and TGnnA sequences in each of the 10 ExsA-dependent promoters, we previously proposed that the mechanisms of ExsA binding to site 1 are similar for each of the promoters (12). In this study, we confirm that the GnC and TGnnA sequences in binding site 1 of the P_exsC promoter are also required for maximal ExsA binding and transcriptional activation (Fig. 1). Experiments employing the T199S, K202R, and T252S gain-of-function mutants indicate that residues T199 and K202 in RH1 make base-specific contacts with the GnC sequence and that RH2 interacts with the TGnnA sequence in binding site 1 at both the P_exoT and P_exsC promoters (Fig. 3 and 4). These findings are consistent with previous Fe-BABE footprinting data showing that the cleavage patterns at binding site 1 are nearly identical on both the top and bottom strands of the P_exoT, P_exsC, P_exsD, and P_pcrG promoter regions (12). From these data, we conclude that ExsA binds in a similar manner to site 1 in the P_exoT and P_exsC promoters, a finding that is likely applicable to each of the known ExsA-dependent promoters. In support of this conclusion, the promoter hybrid experiments presented in Fig. 6 demonstrate that binding site 1 from the P_exoT, P_exsC, P_exsD, and P_pcrG promoters are functionally interchangeable.

Although important details on the interaction of ExsA with binding site 1 are emerging (3, 4, 12), there is a paucity of information regarding the binding of ExsA to site 2. Our truncation analyses of the P_exsC promoter found that at least 11 bp upstream of the adenine-rich region are required for ExsA-dependent activation (Fig. 2). This mapped region contains a single GnC sequence and coincides with binding site 2 as previously defined by DNase I and Fe-BABE footprinting (3, 12). Mutational analyses confirmed that the GnC sequence in binding site 2 is required for maximal binding and transcriptional activation of the P_exsC promoter by ExsA (Fig. 2). To further examine binding to site 2, we used the T199S and K202R gain-of-function mutants, originally isolated using mutant P_exoT promoters. We had previously speculated that gain-of-function (or new-specificity) mutants would be difficult to isolate because the screening strategy required that the mutant proteins be capable of binding to both the native sequence at binding site 2 and the mutant sequence at site 1 (12). Gain-of-function mutants selected for activation of the P_exoT promoter, therefore, may not be optimal for the P_exsC promoter. Despite this concern, the gain-of-function phenotype for the T199S mutant at the P_exsC promoter was particularly compelling because the defects in activation of the P_{exsC(G−49A,G−70A)-lacZ} reporter by wt ExsA and of the wt P_exsC promoter by T199S are both so severe. It is important to note that maximal transcriptional activation by ExsA depends upon the occupation of both binding sites 1 and 2 (3). For this reason, we conclude that the reduced activation of the P_{exsC(G−49A)-lacZ} and P_{exsC(G−70A)-lacZ} reporters by both ExsA and the T199S mutant results from a modest defect in occupation of either site 1 or site 2. For each promoter, therefore, there is one cognate site and one defective site: for the P_{exsC(G−49A)-lacZ} reporter, the interaction defect for wt ExsA would be at site 1 and for the T199S mutant at site 2; the opposite would be true for the P_{exsC(G−70A)-lacZ} reporter. While disruption of a single interaction is tolerated, simultaneous disruption of both interactions, as occurs for the T199S mutant at the wt P_exsC-lacZ reporter and for wt ExsA at the P_{exsC(G−49A,G−70A)-lacZ} reporter, results in a complete loss of activity. The exquisite specificities of the T199S and K202R mutants for their respective cognate promoters (P_{exsC(G−49A,G−70A)-lacZ} and P_{exsC(C−47A,C−68A)-lacZ}), by comparison to wt ExsA, therefore, provide strong genetic evidence that the recognition of binding site 2 in the P_exsC promoter involves base-specific contacts between the GnC sequence and residues T199S and K202R.

The current study suggests that the monomer assembly model for ExsA binding may not apply to the P_exsC promoter. The primary data supporting the monomer assembly model for the P_exoT promoter are the findings that ExsA is capable of binding to site 1 when site 2 is disrupted by mutation or deletion and that disruption of site 1 inhibits binding to both sites (3). The fact that sites 1 and 2 in the P_exsC promoter closely resemble one another and are recognized by ExsA through a common mechanism, however, suggests that an ExsA monomer might be capable of binding either site with similar frequency, followed by recruitment of a second monomer to the remaining site. Indeed, GnA and AnC substitutions in either binding site 1 or 2 in the P_exsC-lacZ reporter resulted in a similar reduction in activation. Site-directed mutagenesis of the GnC sequences in binding sites 1 or 2, however, results in a decrease in the yield of shift product 2 rather than a preferential accumulation of shift product 1 (Fig. 1 and 2). Both of these findings seem contrary to the monomer assembly pathway, because site 1 mutations should be more detrimental than site 2 mutations. These combined findings suggest that, irrespective of the binding order, cooperativity plays a major role in the binding of ExsA to the P_exsC promoter.

The spacing between the GnC sequence and adenine −57 in site 2 of the P_exsC promoter is identical to the spacing for the equivalent sequences at site 1 (Fig. 4A). Accordingly, A-57 was found to be required for maximal activation of the P_exsC-lacZ reporter and is likely recognized by the ExsA monomer bound to site 2 through interactions mediated by RH2 (Fig. 4B). A-57 is located within the conserved adenine-rich region and is positioned 8 bp upstream of the GnC sequence in site 1; this position is completely conserved in each of the known ExsA-dependent promoters, raising the possibility that ExsA bound to site 2 interacts with A-57 at each of the promoters. Consistent with this idea is the previous finding that mutations in the adenine rich region of P_exoT significantly reduced the amount of ExsA bound to site 2, while having little effect on binding at site 1 (3). Fe-BABE footprinting data, however, suggest that the RH2 recognition helix does interact directly with the adenine-rich region at P_exoT (12). Unfortunately, it remains unclear whether the adenine-rich region fulfills a structural/topological role (e.g., DNA bending) and/or is specifically recognized by ExsA, and the possibility remains that the functional role is promoter specific.

Previous Fe-BABE footprinting data for ExsA bound to site 2 at the P_exsC, P_pcrG, P_exsD, and P_exoT promoters found that the position of ExsA binding is similar in each case (12). Although these data suggested that ExsA might recognize site 2 via a common binding determinant, none of the putative GnC sequences located within binding site 2 of the P_exsD, P_exoT, or P_pcrG promoters are required for ExsA binding or transcriptional activation (Fig. 5). Whereas the GnC sequences in binding sites 1 and 2 of the P_exsC promoter are centered 21 bp apart, or two full helical turns of the DNA, none of the putative GnC sequences in site 2 at the other ExsA-dependent promoters possess this same spacing arrangement. This spacing difference may explain why the GnC sequences at those promoters are irrelevant with regard to ExsA binding/activation and could be examined further through insertion/deletion analyses to alter the spacing of the GnC sequences in sites 1 and 2 relative to one another. Irrespective of the spacing, however, the putative GnC sequences are not essential determinants for site 2 occupation at the P_exsD, P_exoT, or P_pcrG promoters.

How then might ExsA recognize site 2 at these promoters? One possibility is that occupation of site 2 involves nonspecific binding of an ExsA monomer to the DNA, driven by self-association with the ExsA monomer bound specifically to site 1. This is consistent with the monomer assembly model whereby occupation of site 2 is largely dependent upon the prior occupation of site 1 (3). The finding that all four of the putative GnC sequences in site 2 of the P_exoT promoter can be disrupted without significantly effecting ExsA binding is also consistent with the possibility that occupation of site 2 requires few, if any, specific binding determinants (Fig. 5). Although this model implies that occupation of site 2 is driven entirely through self-association of ExsA monomers, a previous study found that the isolated DNA-binding domain of ExsA, which is unable to self-associate, was still capable of binding to site 2 (4). This latter finding is more consistent with the existence of specific site 2 determinants that are sufficient for occupation of site 2 in the absence of ExsA self-association.

Our findings from the promoter-swapping experiments provide even more compelling evidence that the interactions occurring at site 2 are specific and result in important functional outcomes including ExsA-dependent promoter activity, the level of cooperative binding, and promoter bending. The P_pcrG promoter is of special interest because the binding properties appear to more closely match those of the P_exsC promoter, while lacking a GnC sequence utilized by ExsA (Fig. 6). The precise nature of the determinants required for recognition of site 2 by ExsA in these promoters will be the subject of future studies.