ABSTRACT
Salmonella enterica serovar Typhimurium colonizes and invades host intestinal epithelial cells using the type three secretion system (T3SS) encoded on Salmonella pathogenicity island 1 (SPI1). The level of SPI1 T3SS gene expression is controlled by the transcriptional activator HilA, encoded on SPI1. Expression of hilA is positively regulated by three homologous transcriptional regulators, HilD, HilC, and RtsA, belonging to the AraC/XylS family. These regulators also activate the hilD, hilC, and rtsA genes by binding to the same DNA sequences upstream of these promoters, forming a complex feed-forward loop to control SPI1 expression. Despite the apparent redundancy in function, HilD has a unique role in SPI1 regulation because the majority of external regulatory inputs act exclusively through HilD. To better understand SPI1 regulation, the nature of interaction between HilD, HilC, and RtsA has been characterized using biochemical and genetic techniques. Our results showed that HilD, HilC, and RtsA can form heterodimers as well as homodimers in solution. Comparison with other AraC family members identified a putative α-helix in the N-terminal domain, which acts as the dimerization domain. Alanine substitution in this region results in reduced dimerization of HilD and HilC and also affects their ability to activate hilA expression. The dimer interactions of HilD, HilC, and RtsA add another layer of complexity to the SPI1 regulatory circuit, providing a more comprehensive understanding of SPI1 T3SS regulation and Salmonella pathogenesis.
IMPORTANCE The SPI1 type three secretion system is a key virulence factor required for Salmonella to both cause gastroenteritis and initiate serious systemic disease. The system responds to numerous environmental signals in the intestine, integrating this information via a complex regulatory network. Here, we show that the primary regulatory proteins in the network function as both homodimers and heterodimers, providing information regarding both regulation of virulence in this important pathogen and general signal integration to control gene expression.
INTRODUCTION
Salmonella enterica serovar Typhimurium is a foodborne pathogen that induces inflammatory diarrhea and invades the distal ileum by directly injecting effector proteins into the cytoplasm of host cells using a type three secretion system (T3SS) (1–4). The genes encoding the structural components of the T3SS, the primary effector proteins, and several regulators are located on Salmonella pathogenicity island 1 (SPI1), a distinct region in the chromosome that defines the genus Salmonella. (5, 6). HilA is the transcriptional regulator of SPI1 that directly activates expression of the needle complex structural genes and indirectly causes activation of genes encoding the translocon and effectors (7–16). Expression of hilA is regulated by the combined actions of three AraC-like transcriptional activators, HilD, HilC, and RtsA (17–19), each of which is capable of inducing transcription of the hilD, hilC, and rtsA genes. All three activators induce hilA and therefore the entire SPI1 T3SS (20). Thus, these regulators form a complex feed-forward regulatory loop, responding to the environment of the small intestine to fully induce production of the SPI1 T3SS (20, 21) (Fig. 1). Data obtained from intestinal invasion experiments in mice fully support this model for SPI1 regulation (20, 22).
SPI1 regulatory circuit. The blue lines indicate transcriptional regulation. The AraC-like proteins HilD, HilC, and RtsA activate transcription of hilD, hilC, rtsA, and hilA, the last encoding the transcriptional activator of the SPI1 T3SS structural genes (20, 26). Many regulatory factors are known to feed into the system at HilD. HilE is a primary negative regulator of HilD activity (30).
HilD, HilC, and RtsA are homologous to each other at the amino acid level. The N-terminal domains of HilD, HilC, and RtsA share approximately 10% identity, whereas the C-terminal DNA binding domains share 46% identity. Each of these proteins binds to the same sites in the hilD, hilC, rtsA, and hilA promoter regions to activate transcription (17–19, 23–25). Despite their similarities, HilD has the predominant role in hilA activation in that the majority of regulatory signals that affect SPI1 expression feed into the regulatory network exclusively through HilD (26–31). Thus, HilD acts as a switch to integrate environmental cues and initiate activation of hilC, rtsA, and hilA expression. Production of HilC and RtsA acts to amplify these signals (20, 21, 24, 26, 32).
The AraC family of proteins, defined by their DNA binding carboxy-terminal domains that contain two helix-turn-helix motifs, can be further categorized according to the functions of their target genes (33). Members of the first group, which includes AraC, regulate genes involved in carbon metabolism (34–36). These regulators usually function as dimers to induce transcription of target genes (33, 37, 38), with the N-terminal domain responsible for dimerization and binding of a regulatory ligand (37, 39). A second group of the AraC family regulators includes MarA, SoxS, and Rob of Escherichia coli, which sense environmental stress (40, 41). They recognize and bind to similar sequences, known as Mar-Rob-Sox boxes, to induce or repress transcription of their target genes. Interestingly, MarA and SoxS consist of only the DNA binding domain and lack a distinct N-terminal domain, acting as monomers to regulate target gene transcription (33, 40, 42, 43). The third group of AraC family proteins includes those that regulate virulence gene expression in response to various environmental cues. Unlike the two previous groups of AraC regulators, there is little consensus on the oligomeric state of these proteins. For example, ToxT of Vibrio cholerae can function as either a monomer or a dimer to bind tox boxes and activate target gene transcription (44–46). ExsA in Pseudomonas aeruginosa binds DNA as a monomer and recruits another monomer to an adjacent binding site via the N-terminal domain to activate expression of the T3SS genes (47–53). On the other hand, VirF in Yersinia exists as a dimer in solution to stimulate transcription of the yop regulon and other target genes (54).
Unlike the well-conserved C-terminal DNA binding domain, the N-terminal domains of AraC proteins, usually encoding dimerization and/or ligand binding sites (33, 38), do not share much similarity at the amino acid level. However, careful analyses of predicted secondary structures of the N-terminal domains of AraC proteins and crystal structures of AraC (39), ToxT (44), ExsA (53), and BgaR from Clostridium perfringens (55) have revealed a conserved structure. Importantly, these proteins contain a hydrophobic α-helix at the end of the N-terminal domain that has been shown to be critical for dimer formation in AraC (39, 56), ToxT (44, 57), XylS of Pseudomonas putida (58), and ExsA (51).
To further understand the regulation of the SPI1 T3SS in Salmonella, we investigated the potential protein interactions between HilD, HilC, and RtsA both biochemically and genetically. We show that these proteins exist as dimers in solution and are capable of forming both homodimers and heterodimers without the presence of DNA. By examining the predicted secondary structures of HilD, HilC, and RtsA, we identified a conserved hydrophobic region predicted to form the last α-helix in each of the N-terminal domains. We show that changes in this domain affect the dimerization of HilD and HilC, disrupting their ability to activate hilA expression. Understanding the dimer interactions of these proteins provides important insight into the signal integration and activation of the SPI1 apparatus and Salmonella pathogenesis.
RESULTS
HilD, HilC, and RtsA exist as dimers in solution.To study the nature of the interactions between HilD, HilC, and RtsA, we first tested whether these proteins exist as monomers or dimers in solution. Functional Myc-tagged versions of the HilD (38.3 kDa), HilC (37.0 kDa), and RtsA (37.7 kDa) proteins encoded on plasmids were expressed in strains lacking a chromosomal copy of hilD, hilC, or rtsA. Whole-cell extracts were analyzed using size exclusion chromatography, and the presence of the proteins in elution fractions was analyzed via Western blotting using anti-Myc antibody. The results (Fig. 2) showed that HilD and HilC primarily eluted in fractions corresponding to the predicted size of dimers (∼75 kDa), confirming that they exist as dimers in solution. It is likely that a fraction of the protein exists as monomers in a dynamic equilibrium with dimers in vivo, but these results suggest that the dimer form is the preferred state. Unlike HilD and HilC, RtsA eluted in fractions corresponding to dimers or larger-order structures, suggesting that the protein was aggregating (not shown). Therefore, no conclusion could be drawn about the dimerization state of RtsA from these results.
Size exclusion chromatography of HilD and HilC. Whole-cell extracts of strains overexpressing Myc-tagged versions of HilD (A) or HilC (B) proteins were separated on a Superdex 75 gel filtration column. Fractions (0.25 ml) were concentrated using TCA precipitation and analyzed via Western blotting. The numbers correspond to the elution fractions collected, with the equivalent elution fractions of control proteins indicated. Quantification of band intensity was determined with ImageJ (68). The strain used was pLS118 or pLS119 in JS515.
To further examine the potential dimerization of these proteins, we purified both His-tagged HilC and SUMO-tagged HilD (30, 59) by nickel affinity and anion-exchange chromatography. The purified proteins were immediately analyzed via size exclusion chromatography. In each case, the fraction corresponding to the expected dimers was then rerun on the size exclusion column. The results are shown in Fig. 3. Only a fraction of the His-tagged HilC initially eluted as a dimer. However, the protein in these fractions remained in the dimeric state upon rerunning the gel filtration. The putative SUMO-HilD dimer consistently eluted earlier than expected; there was also a significant fraction representing degraded protein. Rerunning the dimeric fractions gave an almost identical elution profile. Separating the proteins in the various fractions via SDS-PAGE confirmed that the dimeric proteins were full length, whereas the proteins in the later-eluting peak contained significant degradation product (see Fig. S1 in the supplemental material). These results are very reproducible. Importantly, the minor peak corresponding to the putative monomeric SUMO-HilD (∼14.5 ml in Fig. 3B) also consistently eluted slightly earlier than expected from the column but in accordance with its being half the molecular weight of the major dimeric peak. Thus, both His-HilC and SUMO-HilD remained largely in the dimeric state through the gel filtration process. These proteins, like many AraC-like proteins, have proven challenging to purify and maintain in an active form, and given our experience, we did not include RtsA in this analysis, nor could we maintain SUMO-HilD, His-HilC, or other tagged versions of the proteins in a native state for further biochemical analyses.
Chromatograms of size exclusion chromatography performed using purified protein samples. His-HilC (A and B) or SUMO-HilD (C and D) was separated on a Superdex 200 column. The fractions corresponding to the dimeric size from the first run (A and C, arrows) were run again to demonstrate the stability of the multimer in a subsequent run (B and D). Relative elution peaks of standard proteins are indicated at the top of each panel. AU, absorbance units.
Dimerization of HilD, HilC, and RtsA was tested genetically using a LexA-based monohybrid system (60, 61). The system has been previously used to study protein interaction between SPI1 regulators HilD and HilE (28, 30) and AraC-like proteins ExsA (62) and ToxT (63). In this system, a plasmid-encoded LexA fusion protein is used as a transcriptional repressor of a sulA-lacZ transcriptional fusion. Fusion proteins contain an N-terminal DNA binding domain of LexA, while the C-terminal LexA dimerization domain is replaced with the protein of interest. For this study, we fused full-length HilD, HilC, or RtsA to the LexA DNA binding domain to study its protein-protein interaction. Consistent with the size exclusion chromatography data, HilD, HilC, and RtsA fusion proteins repressed sulA-lacZ reporter activity, suggesting the proteins form dimers in vivo (Fig. 4A). Taken together, our results suggest that HilD, HilC, and RtsA exist as dimers under physiological conditions.
Two-hybrid analysis of HilD, HilC, and RtsA. Reporter strains containing a LexA-repressible sulA-lacZ fusion (SU101 or SU202) were transformed with plasmids encoding LexA1–87-HilD, -HilC, or -RtsA (for monohybrid) (A) and LexAmut1–87-HilD, -HilC, or -RtsA (for two-hybrid) (B). D, HilD; C, HilC; A, RtsA. Positive-control (+) plasmids encoded LexA-Cat, and negative controls (−) contained empty vectors pSR658 (monohybrid) and pSR659 (two-hybrid). (B) The data represent expression when induced with 1 mM versus 0 mM IPTG for a given strain. β-Galactosidase activity units are defined as follows: (micromoles of ONP formed per minute) × 106/(OD600 × milliliters of cell suspension); they are reported as means ± standard deviation, where n = 4 to 8. **, P < 0.005; ***, P < 0.0005 (unpaired t test). For panel B, P values are relative to the negative control.
HilD, HilC, and RtsA can form homodimers and heterodimers.HilD, HilC, and RtsA are homologous to each other, sharing greater than 30% identity across the protein. Given their sequence similarity, we hypothesized that the proteins could interact with one another to form heterodimers as well as homodimers in solution. To test this possibility, we created strains with chromosomally encoded FLAG-tagged HilD, HilC, or RtsA and transformed with plasmids encoding Myc-tagged versions of the proteins, so that each strain had a different combination of FLAG- and Myc-tagged HilD, HilC, and RtsA. Coimmunoprecipitation (co-IP) was performed on whole-cell extracts incubated with anti-Myc–agarose beads. Western blot analysis was performed on co-IP samples to test whether the FLAG-tagged proteins coprecipitated. The results from the co-IP experiment showed that all the proteins formed homodimers (Fig. 5), confirming our findings from size exclusion chromatography and monohybrid experiments. Furthermore, all three proteins were detected in each of the HilD, HilC, and RtsA pulldown experiments, showing all possible combinations of heterodimer formation in vivo (Fig. 5). These data suggest that the proteins can interact with each other to form heterodimers, as well as homodimers. We observed that the band corresponding to the HilC-RtsA heterodimer was more predominant than other bands (Fig. 5B), suggesting that the HilC-RtsA heterodimer may be the preferred form among homo- and heterodimers when HilC is produced in excess. However, any conclusion drawn from this observation is complicated by the fact that HilD, HilC, and RtsA can autoactivate their own expression and activate each other’s expression to different levels (20, 32). The different levels of expression complicate this experiment. Moreover, the relative expression levels of the proteins under physiological conditions are unknown. Therefore, the relative ratio of homodimers and heterodimers in vivo cannot be determined or quantified from these results.
Coimmunoprecipitation of HilD, HilC, and RtsA. The strains encoded 3×FLAG-tagged HilD (designated D), HilC (designated C), or RtsA (designated A) in the chromosome and overexpressed Myc-HilD (A), Myc-HilC (B), or Myc-RtsA (C) from plasmids The first three lanes of each gel are negative controls in which strains lacked the Myc-tagged proteins. All of the strains were grown for 2.5 h, and an equal number of bacteria were lysed and incubated with Myc-agarose beads overnight. The precipitated proteins were separated on an SDS-PAGE gel and subjected to Western blot analysis to detect Myc- or FLAG-tagged proteins. Immunoblots of whole-cell lysates are included. For DNase I treatment, co-IP samples were treated with DNase I (or buffer) for 1 h at 37°C prior to immunoprecipitation. The strains used were JS2082, JS2083, and JS2084 with pLS118, pLS119, or pCE81.
HilD, HilC, and RtsA are transcriptional regulators that bind to the same sequences within the hilD, hilC, rtsA, and hilA promoters (17, 18, 23, 24, 64). It is possible that the heterodimer interaction from the co-IP experiment is an artifact of two homodimers bound to DNA in close proximity. In order to test this, we treated the whole-cell lysate from the RtsA-Myc samples with DNase I prior to immunoprecipitation. To confirm DNA degradation, PCRs were conducted using primers spanning the hilA promoter region. While the expected fragment was amplified from untreated samples, those treated with DNase I did not yield any PCR product, suggesting degradation of HilD, HilC, and RtsA binding sites. Our results from an RtsA pulldown with DNase I treatment show the presence of heterodimers, as well as homodimers, suggesting that these proteins do not require DNA for their dimer interaction (Fig. 5C).
Heterodimer formation was also tested genetically using the LexA two-hybrid system (60, 61). Fusion proteins containing an N-terminal LexA DNA binding domain and a C-terminal full-length HilD, HilC, or RtsA protein were tested in all combinations for heterodimer formation; homodimers were included as internal controls (Fig. 4B). Our results confirm our findings that HilD, HilC, and RtsA can form both homodimers and heterodimers.
Dimerization domain.Although HilD, HilC, and RtsA share obvious homology to AraC in their C-terminal DNA binding domains, they do not share significant homology to AraC in their N-terminal domains. However, based on their predicted secondary structures and threading alignments to the ToxT structure (44, 65, 66), we have identified a conserved hydrophobic region in each of the proteins (HilD residues 180 to 192, HilC residues 161 to 178, and RtsA residues 150 to 167) that corresponds to α-helix 3 of ToxT (Fig. 6). The corresponding α-helices have been identified as the dimerization domain in other AraC-like proteins (39, 44, 51, 52, 55–58). In order to test whether this hydrophobic region is important for the dimer interaction, we performed alanine scanning mutagenesis and tested the abilities of various mutant proteins to form homodimers using the LexA monohybrid system. In the HilD protein, the L186A mutant was significantly less able to repress the sulA-lacZ fusion, suggesting a defect in dimerization. The L190A, L191A, and L192A mutants also showed slight defects at intermediate induction levels (Fig. 7A).
Sequence alignment of dimerization helices in various AraC-like proteins. The last 45 amino acids of the N-terminal domains of the indicated proteins were aligned using Clustal (69). The central 31 amino acids of the alignment are shown, with the amino acids from each protein numbered. The first and last amino acids in the dimerization helices as determined by crystal structure are boxed (39, 44, 53, 55). HilD, HilC, RtsA, CofS, and XylS were modeled onto the ToxT structure using I-Tasser (65, 66). The predicted first and last amino acids in the corresponding helices in these proteins are labeled with tildes. The amino acids for which experimental data suggest a direct role in dimerization are underlined (39, 44, 51, 56–58).
Analyses of HilD and HilC dimerization mutants. (A and C) The reporter strain containing a LexA-repressible sulA-lacZ fusion (SU101) was transformed with plasmids encoding LexA1–87 fused to HilD (A) or HilC (C) wild type or plasmids encoding the designated mutant fusion proteins. (B and D) Expression was induced with 0, 10, or 100 μM IPTG. The ability of Myc-tagged HilD (B) or HilC (D) proteins with the designated mutations to activate a hilA-lacZ fusion was measured in a ΔhilCD ΔrtsA strain (JS515). Protein expression was induced with 10 mM l-arabinose (final concentration). Samples were split to determine the β-galactosidase activity produced from the fusion, while the steady-state level of Myc-tagged proteins was determined by Western analysis. β-Galactosidase activity units are defined as follows: (micromoles of ONP formed per minute) × 106/(OD600 × milliliters of cell suspension); they are reported as means ± standard deviations; n = 4. The P values (unpaired t test) were calculated relative to the corresponding wild type. ***, P < 0.0005.
To test whether these mutations affect the ability of an otherwise wild-type HilD to activate hilA expression, plasmids encoding mutant Myc-tagged proteins were transformed into a ΔhilCD ΔrtsA strain and tested for the ability to activate a hilA-lacZ fusion. Our results showed the L182A, L185A, L186A, and L188A mutants had reduced ability to activate hilA expression (Fig. 7B). To determine whether this change in protein activity was due to protein instability, mutant proteins from the same cultures were analyzed via Western blotting with anti-Myc antibody. Our results showed that reduced hilA activation levels in the L182A, L185A, and L190A mutants are likely due to protein degradation. However, the L186A protein was stable. Taken together, these data suggest that L186 is required for HilD dimerization.
Alanine scanning mutagenesis was also performed in the corresponding region of HilC (residues 166 to 178), and the mutants were tested for the ability to form dimers using the LexA monohybrid system. Mutations in F166, L168, L171, I172, F175, and V176 showed essentially complete loss of dimerization using the monohybrid analysis (Fig. 7C). When the equivalent HilC mutants were tested for the ability to activate hilA expression, L171A and F175A mutants showed complete loss of hilA activation, while F166A and Y170A mutants showed ∼33% reduction in protein activity, apparently due to protein instability (Fig. 7D). However, L171A and F175A mutants were stable, even though they could not activate hilA expression (Fig. 7D). Taken together, our results suggest that L171 and F175 play a crucial role in HilC dimerization. Note that these side groups are predicted to be on the same face of the helix. Interestingly, L168A, I172A, and V176A mutants were apparently deficient at forming dimers in the LexA monohybrid system, but they did not show any significant defect in activating hilA expression (Fig. 7C and D).
The above-mentioned results suggest that the HilD L186 and HilC L171 and F175 side groups are critical for the formation of homodimers, and hence, mutations at these sites, certainly when present in both monomers, block transcriptional activation of hilA. To test if these mutations also affect heterodimer formation or mixed mutant and wild-type dimers, we asked if the mutant proteins could activate transcription in strains that produce the other AraC-like regulators. Comparing the data in Fig. 7 and 8 shows that, although the HilD L186A mutant retains partial function, its ability to activate transcription is unaffected by the presence of wild-type HilD, HilC, or RtsA, activating at 20 to 25% compared to the wild-type protein. This is indicated by the fact that the ratios of activation between the wild-type plasmid and the mutant plasmid are approximately the same in the hilD hilC rtsA strain (Fig. 7) as in the wild-type background or the hilD strain (Fig. 8). In contrast, the HilC L171A and HilC F175A mutant proteins apparently retain partial ability to form heterodimers; whereas they activate at <6% of the activity of the wild-type HilC when produced alone (Fig. 7), they activate at ∼50% of the wild type in strains that produce HilC and/or HilD and RtsA (Fig. 8).
Abilities of HilD and HilC dimerization mutants to activate in concert with wild-type HilD, HilC, and RtsA. The abilities of Myc-tagged HilD or HilC proteins with the indicated mutations to activate a hilA-lacZ fusion were measured in the indicated background strains. Expression was induced with 10 mM l-arabinose (final concentration). β-Galactosidase activity units are defined as follows: (micromoles of ONP formed per minute) × 106/(OD600 × milliliters of cell suspension); they are reported as means ± standard deviations; n = 4. The strain used was JS575-77 with pLS118, pLS119, or corresponding mutant plasmids. **, P < 0.005; ***, P < 0.0005 (unpaired t test).
DISCUSSION
The three AraC-like regulators HilD, HilC, and RtsA form a complex feed-forward loop to control expression of the SPI1 T3SS (Fig. 1) (20). The proteins can each activate expression of the hilD, hilC, rtsA, and hilA promoters and do so by binding to the same specific sites in each of the promoter regions (17–19, 23–25). Certain proteins in the AraC family act as dimers, whereas others act as monomers (33). In some cases, binding of monomers to adjacent sites on the DNA facilitates interaction (47–53). Using both biochemical and genetic analyses, we have shown that the HilD, HilC, and RtsA proteins exist as dimers in solution. Moreover, they can form homodimers and all possible heterodimers with each other (Fig. 2 and 5). These interactions do not require DNA. Although the implications have not been completely investigated, these findings add an additional layer of complexity to the feed-forward loop model, providing a more comprehensive understanding of SPI1 T3SS regulation and Salmonella pathogenesis.
Although both homodimers and heterodimers are evident, we are unable to determine the relative ratios or amounts of the various species. Because HilD, HilC, and RtsA can each activate expression of the hilD, hilC, and rtsA genes, it is difficult to determine the relative amounts of the various proteins at any given moment. We did observe that the HilC-RtsA heterodimer was the dominant form when HilC was overexpressed on a plasmid (Fig. 5B). However, this heterodimer did not predominate when RtsA was overproduced. Due to these limitations, we cannot determine which species of homodimer or heterodimer is more stable or active under physiological conditions.
HilD, HilC, and RtsA bind to the same sites in the various promoters, although the absolute sequence requirements differ slightly between the proteins (18). Each of the sites is presumably recognized with subtly different affinity by each of the homo- and heterodimers. In the case of heterodimers, the affinity is also expected to be dependent on the orientation of binding. Thus, a given heterodimer or homodimer could have a more significant effect at a particular promoter based on subtle differences in the binding sites. These factors all add to the complexity of the system. Indeed, a previously published computational model of the SPI1 regulatory system explained numerous aspects of SPI1 regulation (21). However, the model was dependent on the assumption that the AraC-like proteins act as stable homodimers and that the promoters each have two binding sites, both of which must be occupied for activation. One of the binding sites was modeled to have high affinity for HilD and weak affinity for HilC or RtsA. Our new results suggest that these simplifications, while facilitating the mathematical modeling, are not physiologically accurate.
HilD has a unique role in SPI1 regulation, since it is the integration point for the majority of external signals controlling Salmonella invasion, many acting at the level of the HilD protein (20, 22, 26–28). Since neither the HilC nor the RtsA protein responds directly to regulatory inputs, it is almost certain that the HilC-RtsA heterodimer also does not respond to these inputs (20, 26, 27). However, the role of HilD heterodimers remains unclear. We hypothesize that HilD heterodimers, as well as homodimers, can respond to many of the regulatory signals that control SPI1 expression. For example, HilE is a negative regulator of SPI1 that exclusively interacts with the HilD protein (28, 30). Our data suggest that HilE binds HilD without disrupting dimerization (30). It is difficult to see how HilE could shut the system off if HilD heterodimers were not also subject to regulation. We predict that HilD heterodimers could also interact with other regulatory factors that control the SPI1 apparatus. We do not know, however, whether HilD heterodimers interact in a manner similar to HilD homodimers.
Despite relatively low amino acid identity, it seems clear that many AraC family proteins have essentially the same structure in their N-terminal domains, in addition to the more defined and defining C-terminal DNA binding domains (39, 44, 53, 55). This conservation apparently includes a hydrophobic α-helix in the N-terminal domain that is responsible for dimerization in these proteins (39, 44, 51, 55–58). Figure 6 shows an alignment of this region in several AraC-like proteins. Despite this apparent similarity and clear evidence that the helix is important for dimerization in several AraC-like proteins, there does seem to be significant structural variability in this dimerization helix. For example, the helix in AraC is ∼18 amino acids in length, as seen in the crystal structure (39). The prolines in HilD almost certainly limit the dimerization helix to 13 amino acids, while the HilC helix is predicted to be 18 amino acids. Moreover, the relative positions of the amino acids experimentally defined as being important for dimerization suggest that the helices are slightly out of phase in HilD and HilC, yet the proteins form both homo- and heterodimers. Perhaps the relative interactions are adjustable in the homodimers versus heterodimers. Consistent with this hypothesis, the HilC L171A and F175A mutations had an apparently greater effect on homodimers than on heterodimers (Fig. 7 and 8). We presume the corresponding helix in RtsA is also required for dimerization, but given the fact that the RtsA protein tended to be more insoluble in some of our experiments, we did not explicitly test this hypothesis. A more complete understanding of the interactions between HilD, HilC, and RtsA will require structural data. However, this is challenging given the relative instability of the proteins and AraC-like proteins in general.
MATERIALS AND METHODS
Bacterial strains and growth conditions.All the Salmonella strains used in this study are isogenic derivatives of Salmonella enterica serovar Typhimurium strain 14028 and are listed in Table S1 in the supplemental material. The strains were grown at 37°C in LB medium (1% tryptone, 0.5% yeast extract, 1% NaCl). When required, ampicillin (50 μg/ml) or tetracycline (15 μg/ml) was added to the LB medium for plasmid maintenance. If needed, 10, 100, or 1,000 μM isopropyl-β-d-thiogalactopyranoside (IPTG) or 10 mM l-arabinose (final concentrations) was added to induce gene expression.
Plasmid construction.Full-length hilD, hilC, or rtsA genes were cloned into plasmids containing the DNA binding domain of LexA (LexA1–87) so that the resulting fusion protein contained LexA1–87 in the N-terminal region. DNA fragments were amplified and cloned into pSR658 and pSR659 for mono- and two-hybrid assays, respectively (61). The plasmids were transformed into reporter strain SU101 or SU202 for monohybrid or two-hybrid experiments, respectively (60, 61). Alanine substitution mutations of HilD and HilC were introduced into both the LexA fusion plasmids (pKN8 and pKN12) and Myc-tagged constructs in plasmids pLS118 and pLS119 using QuikChange site-directed mutagenesis kits (Stratagene) according to the manufacturer’s recommendations. All the primers (IDT) used are listed in Table S2 in the supplemental material.
β-Galactosidase assays.Cultures were inoculated and grown overnight in LB medium, subcultured 1:100 in 1.5 ml LB medium in 13- by 100-mm test tubes, and allowed to grow on a roller drum for 4 h. For the LexA-based monohybrid and two-hybrid assays, the indicated amount of IPTG was added to the subculture for induction of fusion proteins. β-Galactosidase assays were performed using a microtiter plate assay as described previously (67). Activity units are defined as follows: (micromoles of ortho-nitrophenol [ONP] formed per minute × 106)/(optical density at 600 nm [OD600] × milliliters of cell suspension) and are reported as means ± standard deviations, where n was ≥4.
Size exclusion chromatography in crude extracts.Fast protein liquid chromatography (FPLC) was used to determine whether HilD, HilC, and RtsA existed as monomeric or dimeric proteins in solution. For whole-cell extract experiments, saturated overnight cultures were subcultured 1:100 into 100 ml of LB medium containing 10 mM l-arabinose (final concentration) to induce protein production and grown with shaking (200 rpm) for 3 h to an OD600 of ∼0.6. Each bacterial culture was centrifuged at 17,000 relative centrifugal force (rcf), and the cell pellet was washed and suspended in 3 ml of NPI buffer (50 mM NaH2PO4, 300 mM NaCl, pH 7.5) with protease inhibitor cocktail (Roche). Culture lysates were prepared by adding 750 μl of 0.1-mm sterile glass beads to the culture, which were vigorously oscillated three times for 3 min each time using a Disruptor Genie 2 (Scientific Industries Inc.); samples were cooled on ice between oscillations. Cellular debris was removed by a 30-min centrifugation (17,000 rcf) at 4°C. The supernatant was further centrifuged for 1 h using ultracentrifugation (160,000 rcf) at 4°C. Samples were loaded onto a Superdex 75 gel filtration column (Amersham). The column was eluted (0.4 ml/min) using degassed NPI buffer, and 250-μl fractions were collected. Each fraction was mixed 1:1 with 20% trichloroacetic acid (TCA) and stored on ice overnight. The samples were centrifuged at 17,000 rcf for 30 min at 4°C, and the supernatants were discarded. Cold acetone (750 μl) was added, and the samples were centrifuged at 17,000 rcf for 10 min at 4°C. The supernatant was discarded, and the pellets were allowed to air dry for 10 min and then heated for 10 min at 95°C. The pellets were boiled in loading buffer. Elution fractions were analyzed by Western blotting. A gel filtration standard (Bio-Rad) and bovine serum albumin (Thermo) were used to calibrate the gel filtration column.
His-HilC and SUMO-HilD expression and purification.The HilC open reading frame (ORF) was cloned into the pET19b vector to create plasmid pKN75 encoding an N-terminal His-tagged HilC. Plasmid pKN75 or the previously constructed His6-SUMO-HilD fusion vector (30, 59) were transformed into BL21(DE3) cells for expression. Overnight cultures of the strain were used as a 1/100 inoculum for six 500-ml cultures in LB in baffled flasks. The cultures were incubated at room temperature (His-HilC) or 37°C (SUMO-HilD) with shaking (200 rpm), and protein expression was induced with 0.4 mM IPTG when the OD600 reached ∼0.6. After 12 h of induction, the cells were harvested and resuspended on ice in lysis buffer (1× phosphate-buffered saline [PBS], 0.1% Triton-X, 10% glycerol, 60 μg/ml DNase I, 20 μg/ml RNase A, 0.3 mg/ml lysozyme, 1 mg/ml benzamidine, and 0.3 mg/ml phenylmethylsulfonyl fluoride [PMSF], 1× complete protease inhibitor cocktail). The suspensions were sonicated on ice for six 3-min cycles (450 W power, 0.7 duty cycle, 3 min rest on ice per cycle) until the cells were fully lysed. The lysates were then centrifuged at 20,000 rcf for 45 min and loaded on HisTrap 5-ml immobilized-metal affinity chromatography (IMAC) columns. The columns were then washed with binding buffer (25 mM NaPi [pH 8.0], 10% glycerol, 20 mM imidazole) supplemented with 500 mM NaCl. The columns were reequilibrated with binding buffer until the A280 stabilized. The proteins were subsequently eluted using the same buffer supplemented with 250 mM imidazole. The elution fractions were pooled, diluted 5 times with binding buffer (50 mM NaPi [pH 7.5], 10% glycerol), and loaded on a HiTrap Q HP 1-ml column. The protein was eluted with binding buffer containing 250 mM NaCl. A Superdex 200 10/300 GL from GE was equilibrated with 50 mM NaPi (pH 8.0), 10% glycerol, 250 mM NaCl. The fractions with the highest protein concentrations were pooled, and 500 μl (His-HilC) or 250 μl (SUMO-HilD) was loaded on the column using a 0.1-ml/min (His-HilC) or 0.3-ml/min (SUMO-HilD) flow rate. The elution volumes of the resulting peaks were compared to a standard curve generated using the same buffer and a Bio-Rad gel filtration standard (no. 151-1901). The fractions containing the protein peak centered around the expected homodimer elution volume (14.3 to 15.1 ml for His-HilC; 12.75 to 13 ml for SUMO-HilD) were combined, and 500 μl of the sample was reloaded on the column. Another round of size exclusion chromatography was performed under identical conditions.
Coimmunoprecipitation assays.Co-IP was used to test whether proteins formed homo- or heterodimers. Saturated overnight cultures were subcultured 1:100 in 70 ml of LB medium containing 10 mM l-arabinose and grown in a flask with shaking (200 rpm) for 3 h to an OD600 of ∼0.6. Each bacterial culture was centrifuged at 17,000 rcf, and the cell pellet was washed and suspended in 1.5 ml of lysis buffer (PBS, 0.5% Triton X-100). Culture lysates were prepared by adding 0.5 ml of 0.1-mm sterile glass beads and vigorously oscillated three times (3 min per oscillation) using a Disruptor Genie 2; samples were cooled on ice between oscillation cycles. Cell lysates were collected after a 30-min centrifugation (17,000 rcf) at 4°C. Fifty microliters of anti-Myc–agarose beads (Sigma-Aldrich) was added to soluble fractions and incubated overnight at 4°C, rotating slowly. Anti-Myc–agarose beads were centrifuged at 17,000 rcf, washed 6 times with lysis buffer (PBS, 0.5% Triton X-100), resuspended in SDS loading buffer, and boiled for 10 min. Samples were analyzed via Western blotting (see below). For DNase treatment, the cell lysate was separated into equal volumes after lysis with glass beads and incubated with either DNase I (Ambion) or buffer alone for 1 h at 37°C prior to incubation with Myc-agarose beads.
Western blot analysis.Size exclusion chromatography elution fractions and co-IP samples were analyzed using Western blot analysis. Proteins were separated on 10% SDS-PAGE gels and transferred to Hybond-ECL membranes (Amersham). The membranes were incubated with biotinylated anti-Myc antibody (primary) and extravidin peroxidase (both from Sigma-Aldrich). Bound antibodies were detected with a chemiluminescence system (Amersham) according to the manufacturer’s specifications.
ACKNOWLEDGMENTS
We thank members of the Slauch laboratory for helpful discussions.
This study was supported by National Institutes of Health grant R01 GM120182 to J.M.S. and partially funded by NIH Training Grant AI078876 to K.-E.N.
FOOTNOTES
- Received 7 January 2020.
- Accepted 6 February 2020.
- Accepted manuscript posted online 10 February 2020.
Supplemental material is available online only.
- Copyright © 2020 American Society for Microbiology.
