Regulation of OmpA Translation and Shigella dysenteriae Virulence by an RNA Thermometer

RNA thermometers are cis-acting riboregulators that mediate the posttranscriptional regulation of gene expression in response to environmental temperature. Such regulation is conferred by temperature-responsive structural changes within the RNA thermometer that directly result in differential ribosomal binding to the regulated transcript. The significance of RNA thermometers in controlling bacterial physiology and pathogenesis is becoming increasingly clear.

ABSTRACT

RNA thermometers are cis-acting riboregulators that mediate the posttranscriptional regulation of gene expression in response to environmental temperature. Such regulation is conferred by temperature-responsive structural changes within the RNA thermometer that directly result in differential ribosomal binding to the regulated transcript. The significance of RNA thermometers in controlling bacterial physiology and pathogenesis is becoming increasingly clear. This study combines in silico, molecular genetics, and biochemical analyses to characterize both the structure and function of a newly identified RNA thermometer within the ompA transcript of Shigella dysenteriae. First identified by in silico structural predictions, genetic analyses have demonstrated that the ompA RNA thermometer is a functional riboregulator sufficient to confer posttranscriptional temperature-dependent regulation, with optimal expression observed at the host-associated temperature of 37°C. Structural studies and ribosomal binding analyses have revealed both increased exposure of the ribosomal binding site and increased ribosomal binding to the ompA transcript at permissive temperatures. The introduction of site-specific mutations predicted to alter the temperature responsiveness of the ompA RNA thermometer has predictable consequences for both the structure and function of the regulatory element. Finally, in vitro tissue culture-based analyses implicate the ompA RNA thermometer as a bona fide S. dysenteriae virulence factor in this bacterial pathogen. Given that ompA is highly conserved among Gram-negative pathogens, these studies not only provide insight into the significance of riboregulation in controlling Shigella virulence, but they also have the potential to facilitate further understanding of the physiology and/or pathogenesis of a wide range of bacterial species.

INTRODUCTION

Pathogenic bacteria of the genus Shigella (S. flexneri, S. boydii, S. sonnei, and S. dysenteriae) are the causative agents of shigellosis, a severe diarrheal disease of humans marked by intense cramping and frequent mucoid bloody diarrhea (1). Shigellosis remains a worldwide health concern, with a conservative estimate of 165 million illnesses and 1.1 million deaths worldwide each year (1, 2). The Centers for Disease Control and Prevention reports a consistent rate of approximately 500,000 cases of shigellosis in the United States annually (3, 4). Significantly, the incidence of antibiotic resistance among circulating strains of Shigella is increasing globally, a trend that has resulted in Shigella spp. receiving a threat level of “serious” in a recent CDC report (1, 5).

During infection within and spread between the host, Shigella spp. must adapt to a wide variety of environmental conditions. The ability of a bacterial pathogen to sense and adapt to changes in environmental conditions facilitates both survival and virulence, as the production of bacterial virulence factors is often regulated in response to specific environmental cues (6). It is well established that Shigella species utilize an arsenal of virulence factors to survive and cause disease within the infected host, and that the production of such factors is regulated in response to host-associated environmental cues including, but not limited to, temperature, pH, osmolarity, and nutrient availability (7, 8). Such environment-responsive regulation likely increases Shigella sp. fitness by ensuring optimal production of virulence-associated factors when the pathogen is within the human host, the precise environment in which such factors will provide the greatest advantage to the invading pathogen.

There is a growing list of chromosomally encoded factors now recognized as being critical to virulence-associated processes in Shigella spp., such as OmpA (9). First identified as a major component of the Gram-negative bacterial cell wall (10), OmpA is a highly conserved protein (11) that continues to be the focus of investigation in several bacterial species (12). Coined the “molecular Swiss Army knife” (13), OmpA has been implicated in a variety of cellular functions (13, 14). Importantly, OmpA is a known virulence factor in several bacterial pathogens shown to influence virulence-associated processes such as attachment, biofilm formation, and immune modulation, as well as invasion into, replication within, and spread between eukaryotic cells (9, 12, 13). In the case of S. flexneri, OmpA has been shown to be required for the virulence-associated process of efficient cell-to-cell spread by the pathogen (9).

Given its relative abundance and location within the bacterial outer membrane, as well as the number and variety of functions assigned to the protein, it is not surprising that OmpA production is extensively regulated (13). RNA-based regulation plays a central role in modulating OmpA production with both cis-acting and trans-acting riboregulation shown to modulate transcript stability (15–20). The studies presented within describe yet another mechanism of RNA-mediated regulation of ompA expression, that of temperature-dependent posttranscriptional regulation conferred by an RNA-thermometer harbored within the 5′ untranslated region (UTR) of the gene.

RNA thermometers are cis-encoded regulatory elements that function to modulate translation efficiency in response to changes in environmental temperature (21–23). Bacterial RNA thermometers share a key feature, the presence of a temperature-responsive element(s) that undergoes structural changes which modulate accessibility of the ribosomal binding site, and thus translation from the regulated transcript (21–25). RNA thermometers have been shown to control essential bacterial processes such as heat shock, virulence gene regulation, and, as in Shigella dysenteriae, nutrient acquisition (26–34).

The studies presented within identify and characterize a functional RNA thermometer controlling the temperature-dependent production of S. dysenteriae OmpA and directly implicate this riboregulator in controlling virulence-associated processes in this important bacterial pathogen.

RESULTS

OmpA is an S. dysenteriae virulence factor.

The role of OmpA in S. dysenteriae pathogenesis was first investigated using in vitro tissue culture-based plaque assays, well-established assays routinely utilized to assess Shigella virulence (9, 35–42). Specifically, the ability of S. dysenteriae lacking ompA (ΔompA mutant) to form plaques in a monolayer of human epithelial cells (Henle cells) was compared to that of the parental strain (wild type [WT]). The S. dysenteriae ΔompA mutant forms significantly fewer plaques than does the wild-type strain (Fig. 1A and B). The plaque formation defect of the S. dysenteriae ΔompA mutant can be complemented with a plasmid carrying a wild-type copy of the gene (pompA) (Fig. 1A and B). The reduced number of plaques formed by the S. dysenteriae ΔompA mutant is not a result of a decrease in the ability of the strain to invade the eukaryotic cells within the monolayer, as evaluated using an in vitro tissue culture-based invasion assay (Fig. 1C) or a generalized growth defect, as determined by growth analysis (Fig. 1D). These data clearly implicate OmpA as a virulence factor in S. dysenteriae.

FIG 1.

OmpA is a virulence factor in S. dysenteriae. (A) Representative images from in vitro plaque assays showing a monolayer of Henle cells with plaques formed by wild-type (WT) S. dysenteriae or the strain lacking ompA (ΔompA) carrying an empty vector or the ompA complementation plasmid (pompA). (B and C) Quantification of plaque formation (B) and percent invasion (C) by WT S. dysenteriae or the strain lacking ompA (ΔompA) carrying an empty vector or the ompA complementation plasmid (pompA). (D) Growth analysis of the WT and ΔompA mutant strains of S. dysenteriae over time. Data presented in panels B to D are the averages of the results from biological triplicate analyses. Error bars represent one standard deviation. *** and **** denote a statistically significant difference with P values of <0.005 and <0.001, respectively.

It is notable that in addition to fewer plaques, S. dysenteriae strains lacking ompA also form visibly smaller plaques than those formed by the parental strain (Fig. 1A). Data demonstrating a reduction in both the number and size of plaques formed by the S. dysenteriae ΔompA mutant, together with an invasion rate similar to that of the wild-type strain, are consistent with a role for OmpA in cell-to-cell spread by the pathogen, as was observed previously for S. flexneri (9).

S. dysenteriae ompA is subject to temperature-dependent posttranscriptional regulation.

It is well established that the production of Shigella virulence factors is extensively regulated, frequently in response to one or more environmental signals encountered throughout the course of a natural infection (7, 8). Such environment-specific signals include, but are not limited to, osmolarity, pH, nutrient availability, and temperature (7, 8). Previous studies in Escherichia coli have demonstrated that the production of OmpA is subject to multiple levels of regulation (13), including RNase E-mediated temperature-dependent regulation of transcript stability (15–20, 43), as well as temperature-dependent regulation of OmpA translation mediated via an as-yet-unknown molecular mechanism (43). Parallel measurement of relative ompA transcript levels and OmpA protein levels was performed to determine if S. dysenteriae OmpA production is subjected to temperature-dependent posttranscriptional regulation. Following the growth of wild-type S. dysenteriae to the mid-logarithmic phase at 25°C or 37°C, the total RNA and total protein were isolated, and the relative abundances of ompA transcripts and OmpA protein were measured using quantitative real-time PCR and Western blot analyses, respectively. Despite a significant reduction in the relative amount of ompA transcript following growth of S. dysenteriae at 37°C, compared to that those following growth of the strain at 25°C (Fig. 2A), no significant difference in the relative abundance of OmpA protein is observed (Fig. 2B). These data are consistent with observations made in E. coli (43) and indicate that translation from the less-abundant ompA transcript is greater at 37°C, demonstrating that S. dysenteriae ompA expression is subject to temperature-dependent posttranscriptional regulation.

FIG 2.

S. dysenteriae ompA is subject to temperature-dependent posttranscriptional regulation. (A) Quantitative real-time PCR of the relative levels of ompA transcript present in wild-type S. dysenteriae following growth of the strain at the indicated temperature. The amount of ompA transcript measured in each sample is normalized to that of rrsA and expressed relative to that measured in a single 37°C sample. (B) Western blot analysis of the relative abundance of OmpA present in wild-type S. dysenteriae following growth of the strain at the indicated temperature. The histogram represents the intensity of detected OmpA bands, expressed relative to a single 37°C sample set to a value of 1. The inserted images are representative of the detected bands quantified in the corresponding data set. All data presented in this figure are the averages of analyses performed in biological triplicate. Error bars represent one standard deviation. * denotes a statistically significant difference with a P value of <0.05.

In silico analyses suggest the presence of a putative 4U RNA thermometer within the 5′ UTR of S. dysenteriae ompA.

Based on analyses of E. coli ompA (44) and the similarity between the two genes (45), it is predicted that S. dysenteriae ompA contains a 133-nucleotide-long 5′ UTR. As an initial approach to characterization of the molecular mechanism underlying the observed temperature-dependent posttranscriptional regulation of S. dysenteriae ompA, the predicted ompA 5′ UTR was subjected to structural analysis using Mfold (46) (Fig. 3A). This in silico analysis predicts that RNA within the ompA 5′ UTR assumes a complex structure composed of three hairpins, the third of which displays characteristics conserved among members of the “4U” family of RNA thermometers (31). The specific features most consistent with the prediction of an RNA thermometer include the presence of a hairpin structure in which four consecutive uracil residues participate in canonical and noncanonical base pairing with sequences composing the ribosomal binding site (Fig. 3B). The prediction of such conserved features is consistent with the presence of a functional RNA thermometer within the 5′ UTR of S. dysenteriae ompA.

FIG 3.

A putative RNA thermometer is located within the 5ʹ UTR of S. dysenteriae ompA. (A) The predicted secondary structure formed by the nucleic acid sequencing of the S. dysenteriae ompA 5ʹ UTR. The translational start site is underlined. (B) The nucleic acid sequence and predicted structure of the putative RNA thermometer identified at the 3ʹ terminus of the S. dysenteriae ompA 5ʹ UTR. The boxed sequence represents the predicted ribosomal binding site (RBS) and consecutive uracil residues characteristic of a 4U RNA thermometer. The translational start site is underlined.

The nucleic acid sequence composing the putative ompA RNA thermometer is sufficient to confer temperature-dependent posttranscriptional regulation.

The prediction of an RNA thermometer within the 5′ UTR of ompA is consistent with the observed posttranscriptional temperature-dependent regulation of S. dysenteriae OmpA production. If the putative ompA RNA thermometer is a functional riboregulator, it, by definition, would be predicted to be sufficient to confer temperature-dependent posttranscriptional regulation. In order to experimentally test this prediction, the putative ompA RNA thermometer was isolated from the remainder of the ompA 5′ UTR and cloned between the constitutive promoter and reporter gfp gene on plasmid pXG-10 (47) to generate the translational reporter plasmid pWT-gfp (Fig. 4A). Such cloning results in a translational reporter in which gfp expression is driven by a constitutive plasmid promoter and subject to any regulation mediated by the cloned S. dysenteriae sequences, in this case, the putative ompA RNA thermometer.

FIG 4.

Nucleic acid sequence composing the ompA RNA thermometer is sufficient to confer temperature-dependent posttranscriptional regulation. (A) Schematic of the pompA translational reporter carrying the putative RNA thermometer cloned between the constitutive plasmid promoter (P_LtetO) and a gfp reporter gene. (B) Quantitative real-time PCR of the relative levels of gfp transcript present in wild-type S. dysenteriae carrying the pompA translational reporter following growth of the strain at the indicated temperature. The amount of gfp transcript measured in each sample is normalized to that of rrsA and expressed relative to that measured in a single 37°C sample. (C) Western blot analysis of the relative levels of GFP present in wild-type S. dysenteriae carrying the pompA translational reporter following growth of the strain at the indicated temperature. The histogram represents the intensity of detected GFP bands, expressed relative to a single 37°C sample set to a value of 1. The inserted images are representative of the detected bands quantified in the corresponding data set. All data presented in this figure are the averages of analyses performed in biological quadruplicate. Error bars represent one standard deviation. * and *** denote a statistically significant difference with P values of <0.05 and <0.005, respectively.

In order to experimentally determine if the putative ompA RNA thermometer is sufficient to confer posttranscriptional temperature-dependent regulation onto the reporter gfp gene, pWT-gfp was introduced into wild-type S. dysenteriae and the strain cultured to the mid-logarithmic phase of growth at either 25°C or 37°C. Total RNA and protein were isolated from the strain, and the relative amounts of gfp transcript and green fluorescent protein (GFP) were measured using quantitative real-time PCR and Western blot analyses, respectively. Despite a slight yet significant decrease in the relative abundance of gfp transcript levels (Fig. 4B), significantly more GFP is produced following growth of the reporter strain at 37°C compared to that at 25°C (Fig. 4C). The production of increased levels of GFP from lesser amounts of transcript following growth of the reporter strain at 37°C indicates that the cloned sequences are able to confer temperature-dependent posttranscriptional regulation, a defining feature of a functional RNA thermometer.

The stability of the inhibitory hairpin within the putative ompA RNA thermometer is predictably altered by changes in the primary nucleic acid sequence.

A defining feature of an RNA thermometer is the presence of a dynamic inhibitory hairpin for which structure, and thus function, is directly impacted by environmental temperature (21–23). Fluorescence-monitored thermodenaturation analyses (Fig. 5A) were performed to assess the effect of increasing temperature on the overall structure of the stem-loop portion of the putative ompA RNA thermometer (the ompA inhibitory hairpin) and to identify the thermodenaturation point (melting temperature [T_m]) of the element. This differential fluorimetry approach has been used previously for moderate-throughput assays investigating RNA stability and structure-function relationships (48, 49). For this analysis, the ompA inhibitory hairpin model RNA was labeled at the 5′ terminus with a HEX fluorophore and at the 3′ terminus with a DABCYL quencher (Fig. 5A), and fluorescence was measured across a range of temperatures to obtain a thermodenaturation (melting) curve. The observed fluorescence of the ompA inhibitory hairpin model RNA was sigmoidal as a function of temperature, indicating a transition from a folded state (fluorescence quenched) to a highly denatured state (highly fluorescent) (Fig. 5B). The T_m and shape of the melting curves were independent of the RNA strand concentration (our unpublished data), indicating that the transition represented by the melting curve is the monomer denaturation of an RNA hairpin (50). The sigmoidal shape of the melting curve is consistent with a two-state helix to random coil transition upon denaturation of the RNA, with a calculated T_m of 35°C (Fig. 5B).

FIG 5.

The stability of the inhibitory hairpin within the ompA RNA thermometer is temperature sensitive and predictably influenced by changes in the primary sequence. (A) Schematic of fluorescence-monitored RNA thermal denaturation assay and sequence of wild-type ompA model RNA and stabilized variant (noted by arrows). (B) Thermal denaturation profile and first derivative for wild-type ompA model RNA (solid line) and stabilized variant (dotted line).

Alterations in the nucleic acid sequence composing the inhibitory hairpin of a functional RNA thermometer would be expected to change the temperature responsiveness of the element in predictable ways. Sequence alterations that stabilize the inhibitory hairpin within a functional RNA thermometer would be expected to increase the T_m of the element. Site-specific nucleic acid changes were introduced into the ompA inhibitory hairpin model RNA in order to test the impact of primary sequence on the T_m of the predicted inhibitory hairpin. Two base changes were introduced into the ompA inhibitory hairpin model RNA that are predicted to stabilize the putative inhibitory hairpin (Fig. 5A). The measured stability of the ompA model RNAs is dependent on the strength of the base pairs in the folded region of the hairpin. Compared to a T_m of 35°C for the wild-type construct, the T_m of the stabilized model RNA was 78°C (Fig. 5B). These data indicate that alterations in sequence within the putative inhibitory hairpin within the ompA RNA thermometer have a predictable impact on the temperature responsiveness of the element, an essential feature of a functional RNA thermometer.

Sequences surrounding the ribosomal binding site within the putative ompA RNA thermometer become increasingly single stranded at 37°C.

The ability of an RNA thermometer to mediate temperature-dependent posttranscriptional regulation of gene expression is mediated directly by structural changes within the inhibitory hairpin that result in differential exposure of the associated ribosomal binding site. Structure-probing analyses were performed to measure the effect of temperature on the relative exposure of the ribosomal binding site within the larger context of the putative ompA RNA thermometer. An RNA molecule corresponding to the ompA RNA thermometer was generated, purified, radiolabeled, and subject to RNase-mediated digestion at 25°C and 37°C. The RNases utilized were RNase T1, RNase T2, and nuclease S1 that preferentially digest upstream of a single-stranded guanine residue, a single-stranded adenine residue, and any single-stranded residue, respectively. Analyses demonstrate that specific areas within the putative ompA RNA are increasingly susceptible to digestion at 37°C compared to at 25°C, while other areas are equally digested at the two temperatures. The area that displays the greatest level of temperature-responsive differential digestion by each RNase is that harboring the ribosomal binding site (Fig. 6). These data are consistent with increased exposure of the ribosomal binding site within the ompA RNA thermometer at 37°C, the temperature at which optimal expression of ompA as well as the plasmid-carried reporter gene is observed (Fig. 2 and 4, respectively).

FIG 6.

Exposure of the ribosomal binding site within the ompA RNA thermometer is temperature responsive and predictably influenced by the primary sequence. RNase-based structure probing using in vitro-synthesized RNA molecules representing the WT or stabilized version of the ompA RNA thermometer. Each RNA molecule was subjected to partial digestion at the indicated temperature (25°C or 37°C) with RNase T1, RNase T2, and nuclease S1 that preferentially digest 3ʹ to a single-stranded guanine residue, a single-stranded adenine residue, and any single-stranded residue, respectively. The control lanes are as follows: L_OH, alkaline ladder generated from the WT RNA molecule; L_T1 and L_T2, complete digestion of the WT RNA molecule with RNase T1 and T2, respectively; C, H₂O was added instead of RNase. TSS, translational start site; RBS, ribosomal binding sequence. The data presented in this figure are representative of analyses performed in biological duplicate.

Structure probing analyses of the ompA RNA thermometer containing site-specific mutations shown previously to increase the thermodenaturation point of the element (Fig. 5) demonstrate that such stabilization prevents the temperature-induced exposure of the ribosomal binding site within the element (Fig. 6).

Ribosomal binding to the ompA 5′ UTR is temperature responsive, with increased binding occurring at 37°C.

Increased exposure of the ribosomal binding sequence within a bacterial RNA thermometer is the molecular mechanism underlying differential ribosomal binding and, thus, temperature-responsive posttranscriptional regulation. The impact of temperature on ribosomal binding within the ompA RNA thermometer was characterized directly using toeprinting assays (primer extension inhibition) (31, 51). In this assay, an in vitro-transcribed RNA molecule containing the ompA RNA thermometer and a portion of the regulated gene is used as the template for a reverse transcription assay primed by a radiolabeled oligonucleotide in the presence or absence of the 30S ribosomal subunit. The presence of a ribosomal subunit bound to the RNA template acts as a roadblock during the reverse transcription reaction, resulting in the generation of a short terminated product (toeprint). In the absence of a bound ribosomal subunit, the reverse transcription reaction continues unimpeded, generating a long full-length product. In the presence of ribosome (+ lanes), dramatically more truncated product is generated when the analysis is completed at 37°C compared to that generated when the analysis is completed at 25°C (Fig. 7). These data support the conclusion that ribosomal binding within the ompA RNA thermometer is influenced by temperature, with increased binding occurring at 37°C.

FIG 7.

Ribosomal binding to the ompA 5ʹ UTR is temperature sensitive and predictably influenced by the primary sequence. Toeprint analyses measuring ribosomal binding to the WT or stabilized versions of the ompA RNA thermometer at 25°C and 37°C are shown. The presence and absence of a 30S ribosome are indicated by “+” and “−”, respectively. ATG indicates the translational start site, while the full-length and truncated products are indicated by the dark and light arrows, respectively. The data presented in this figure are representative of analyses performed in biological duplicate.

Toeprint analyses of the ompA RNA thermometer containing site-specific mutations shown previously to increase the thermodenaturation point of the element (Fig. 5) and prevent temperature-responsive exposure of the ribosomal binding site (Fig. 6) demonstrate that these changes greatly reduce ribosomal binding to the ompA RNA thermometer at each temperature tested. A lack of ribosomal binding is consistent with a model in which the mutated ompA RNA thermometer remains in a “closed” position, one that would keep the ribosomal binding site sequestered, even at the previously permissive temperature of 37°C.

Mutations that alter in vitro measures of temperature-responsive structural changes and ribosomal binding alter the functionality of the ompA RNA thermometer.

Given that the regulatory activity of an RNA thermometer is mediated by differential ribosomal binding due to temperature-responsive structural changes, it is predicted that the introduction of point mutations that alter each of these molecular steps would result in predictable changes in the regulatory activity of the element. Specifically, sequence alterations that stabilize the inhibitory hairpin within a functional RNA thermometer would be predicted to prevent the temperature-dependent alterations in structure that mediate differential exposure of the ribosomal binding site. In such a case, it would be expected that the ribosomal binding site would remain occluded at previously permissive temperatures, resulting in decreased expression of the regulated gene.

Site-specific mutations predicted to stabilize the ompA RNA thermometer while not altering the overall structure of the element were introduced into the pWT-gfp translational reporter, generating the mutant reporter pS-gfp. As a means to characterize the impact of the engineered mutations on the activity of the putative ompA RNA thermometer, the relative amounts of gfp transcript and GFP produced from the mutated reporter (pS-gfp) were compared to those produced from the wild-type reporter (pWT-gfp) following the growth of wild-type S. dysenteriae carrying each at 37°C. While there was a 2-fold reduction in the abundance of gfp transcript following growth of S. dysenteriae carrying the stabilized reporter compared to that measured in the strain carrying the wild-type reporter (likely due to increased degradation of double-stranded RNA within the bacterial cell), a significant amount of reporter transcript was measured in both strains (Fig. 8A). Despite the presence of a significant amount of transcript in each strain, GFP was only detected in the strain carrying the wild-type reporter plasmid; no protein was detected in the strain carrying the stabilized plasmid reporter (Fig. 8B).

FIG 8.

Stabilization of the inhibitory hairpin within the ompA RNA thermometer results in decreased expression at the permissive temperature. (A) Quantitative real-time PCR of the relative levels of gfp transcript present in wild-type S. dysenteriae carrying the WT (pWT-gfp) or stabilized (pS-gfp) translational reporter following growth at the permissive temperature of 37°C. The amount of gfp transcript measured in each sample is normalized to that of rrsA and is expressed relative to that measured in a single WT sample. (B) Western blot analysis of the levels of GFP present in wild-type S. dysenteriae carrying the WT (pWT-gfp) or stabilized (pS-gfp) translational reporter following growth at the permissive temperature of 37°C. The histogram represents the intensity of detected GFP bands, expressed relative to a single wild-type sample set to a value of 1. B.L.D. indicates that protein levels were below the level of reliable detection. The inserted images are representative of the detected bands quantified in the corresponding data set. All data presented in this figure are the averages of analyses performed in biological triplicate. Error bars represent one standard deviation. **** denotes a statistically significant difference with a P value of <0.001.

These data demonstrate that, as predicted for a functional RNA thermometer, base changes that stabilize the inhibitory hairpin within the putative ompA RNA thermometer result in decreased translational efficiency at the previously permissive temperature of 37°C. Additionally, the data support the conclusion that mutations which alter in vitro measures of temperature-responsive structural changes and ribosomal binding alter the functionality of the ompA RNA thermometer within the bacterial cell.

The ompA RNA thermometer impacts S. dysenteriae virulence.

Given the established role of ompA in S. dysenteriae virulence processes (Fig. 1), elements that control the expression of this gene are potentially virulence factors in and of themselves. The role of the ompA RNA thermometer in S. dysenteriae virulence was investigated using in vitro tissue culture-based analyses. Specifically, the ability of the WT ompA to complement the mutant plaque phenotype of the S. dysenteriae ΔompA mutant (Fig. 1) was compared to that of ompA carrying the stabilizing mutations demonstrated to disrupt the regulatory function of the element (Fig. 5 and 8). Unlike a plasmid carrying a wild-type copy of full-length ompA (pompA), a plasmid carrying a copy of ompA with the stabilized RNA thermometer (pSompA) was unable to complement the reduced plaque phenotype of the S. dysenteriae ΔompA mutant (Fig. 9A and B). These data demonstrate that a functional ompA RNA thermometer is required for efficient plaque formation by S. dysenteriae, thus implicating the riboregulator as a virulence factor in this important bacterial pathogen.

FIG 9.

The ompA RNA thermometer controls a virulence-associated process, marking the riboregulator a S. dysenteriae virulence factor. (A) Representative images from in vitro plaque assays showing a monolayer of Henle cells with plaques formed by wild-type S. dysenteriae (WT) or the strain lacking ompA (ΔompA mutant) carrying an empty vector, the wild-type ompA complementation plasmid (pompA), or the stabilized ompA-complementing plasmid (pSompA). (B) Quantification of plaque formation by WT S. dysenteriae or the ΔompA mutant carrying an empty vector, pompA, or pSompA. Data presented in this figure are the averages of analyses performed in biological quadruplicate. Error bars represent one standard deviation. * and **** denote a statistically significant difference with P values of <0.05 and <0.001, respectively.

DISCUSSION

The transition from the nonhost environment to a warm-blooded host is associated with a substantial change in temperature. As such, it is not surprising that temperature is one of the multiple parameters to which many pathogenic bacteria specifically respond, and one that influences the expression of specific virulence-associated genes (22, 34). RNA thermometers have been implicated in this process in various pathogens. While several RNA thermometers control the translation of the master regulator of virulence, others regulate the production of various virulence factors involved in processes ranging from adhesion, quorum sensing, and iron acquisition to immune invasion (22, 34). Here, we describe the first RNA thermometer that controls the translation of an outer membrane protein, namely, S. dysenteriae OmpA. Furthermore, our findings implicate the ompA RNA thermometer as a virulence factor in S. dysenteriae.

As the list of characterized bacterial RNA thermometers grows, this broad group of riboregulators is being grouped into families, based largely on the conservation of specific structural features. The presence of four consecutive uracil bases within the inhibitory hairpin of the ompA RNA thermometer (Fig. 3) places this newly characterized riboregulator within the FourU RNA thermometer family (31). First identified in Salmonella spp. (31), FourU RNA thermometers are most often found in bacterial pathogens, where they function to control the production of stress response proteins (31, 52) or, as is the case with the S. dysenteriae ompA RNA thermometer, the production of virulence-associated factors (26, 27, 53, 54).

The ompA RNA thermometer is the third RNA thermometer characterized in S. dysenteriae to date (26, 33). Those previously characterized function to regulate the production of ShuA and ShuT (26, 33), factors required to utilize heme as a source of nutritional iron. In the case of ShuA and ShuT, the fitness advantage of RNA thermometer-mediated regulation is clear; the production of each factor is both optimized and restricted to the environmental temperature experienced within the infected host, the only environment in which heme will be encountered by the pathogen. The advantage of RNA thermometer-mediated regulation of ompA expression, however, is not as obvious. While the experimental design utilized in this study allowed for characterization of the ompA RNA thermometer in isolation, to understand the physiological implication of this regulation, these findings must be considered within the broader regulatory network known to control ompA expression. Studies in E. coli have demonstrated that OmpA production is highly regulated in response to growth phase and temperature (13, 15–17, 20, 43, 55, 56). This study, along with previously published work, demonstrates that temperature-dependent posttranscriptional regulation of ompA expression is conserved between E. coli and S. dysenteriae, with similar levels of OmpA observed at 25°C and 37°C despite a significant reduction in transcript levels at 37°C (Fig. 2) (43). Increased translational at 37°C accounts for the relative conservation of OmpA levels, despite the significant reduction in transcript levels at this temperature. In the case of Shigella spp., these data are consistent with the hypothesis that dramatic fluctuations in OmpA levels are detrimental to the pathogen; overproduction of OmpA is lethal (our unpublished data), while underproduction leads to reduced virulence (Fig. 1 and 9). Multiple layers of regulation likely ensure that S. dysenteriae OmpA levels remain within a narrow optimal range within each unique environment encountered throughout transmission and infection.

Despite the fact that RNA thermometers are often housed within known virulence-associated genes, evaluating the direct effect of the RNA thermometer itself on bacterial virulence is difficult and often limited by the availability of relevant virulence models. In this study, in vitro tissue-cultured based analyses were utilized to directly implicate the ompA RNA thermometer as an S. dysenteriae virulence factor (Fig. 9). As such, the identification and characterization of the S. dysenteriae ompA RNA thermometer lend further support to the growing recognition that RNA thermometers have significant impact, not just on the bacterial stress response, but also on the ability of bacterial pathogens to cause disease. Such findings highlight the potential of this type of work to contribute to the foundational knowledge of molecular mechanisms underlying Shigella sp. virulence.

Characterizing the molecular mechanisms controlling the expression of ompA is an important contribution to the foundational understanding of Shigella pathogenesis. The implications of these findings, however, are likely not limited to understanding virulence gene expression in S. dysenteriae. Specifically, the nucleic acid sequence contained within the ompA 5′ UTR is 96% identical between members of the Shigella genus (Fig. 10). Extending this analysis to the related Gram-negative pathogens of enterohemorrhagic E. coli (EHEC), enteropathogenic E. coli (EPEC), uropathogenic E. coli (UPEC), and neonatal meningitis-causing E. coli (NMEC), and Salmonella and Citrobacter spp. demonstrates an overall nucleic acid identity of 93% throughout the length of each 5′ UTR. Furthermore, the nucleic acid sequence composing the functional RNA thermometer (Fig. 3) is 100% identical in each of these Gram-negative enteropathogens (Fig. 10). The conservation of sequence alone does not indicate conservation of regulatory activity. It is, however, suggestive that OmpA production may be regulated by a functional RNA thermometer in a variety of Gram-negative bacterial pathogens, a prediction that may inform future investigations.

FIG 10.

The nucleic acid sequence composing the ompA RNA thermometer is conserved among a variety of Gram-negative enteropathogens. Shown is the alignment of the nucleic acid sequences within the 5ʹ UTR of ompA from each Shigella species as well as that from enteropathogenic E. coli (EPEC), enterohemorrhagic E. coli (EHEC), neonatal meningitis-causing E. coli (NMEC), uropathogenic E. coli (UPEC), Salmonella enterica, and Citrobacter koseri. Gray shading highlights residues that are not conserved in all species analyzed. Black boxes indicate sequences corresponding to the consecutive uracil residues conserved among members of the 4U RNA thermometer family and that within the RBS to which they are predicted to bind. Dashes indicate the conserved nucleic acid sequences shown to be sufficient to confer posttranscription temperature-dependent regulation. The translational start site is in bold.

Despite significant advances in the field over the last decade, much remains to be learned about bacterial RNA thermometers. Generalizations regarding conserved structural features and regulatory mechanisms are supported by the bacterial RNA thermometers characterized to date (22). However, given the limited number of RNA thermometers that have been identified and experimentally characterized, it would be premature to assume that the breadth of their regulatory functions, structures, and impacts on bacterial physiology and virulence is fully known. Do RNA thermometers within a given family of regulators display similar temperature responsiveness? Are RNA thermometers within a given species more or less similar in both structure and function than those within more distantly related species? What is the minimum temperature differential to which a bacterial RNA thermometer can detect and respond? What is the role of RNA thermometers in controlling virulence-associated gene expression in Gram-positive pathogens? These are just a few of the fundamental questions that remain to be answered by future studies on bacterial RNA thermometers.

MATERIALS AND METHODS

Bacterial growth conditions.

All bacterial strains and plasmids used in this study are detailed in Table 1. Escherichia coli was cultured at 37°C in Luria-Bertani (LB) broth or on LB agar plates (1% tryptone, 0.5% yeast extract, and 1% NaCl). Shigella dysenteriae was cultured at the indicated temperature in LB broth or on tryptic soy broth agar plates (Becton, Dickinson & Company, Sparks, MD) containing 0.01% Congo red dye (ISC BioExpress, Kaysville, UT). For the growth of E. coli- or S. dysenteriae-carrying plasmids, the appropriate medium was supplemented with ampicillin or chloramphenicol at a final concentration of 150 μg ml⁻¹ or 30 μg ml⁻¹, respectively.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Descriptiona	Source or reference
Strains
S. dysenteriae
ND-100	Wild type	57
ΔompA mutant	ND-100 lacking ompA	This study
E. coli
DH5α		Invitrogen
Plasmids
pXG-10	Low-copy-number plasmid carrying a pLTet0-1 constitutive promoter driving expression of a gfp reporter, Chlr	47
pWT-gfp	Translational fusion; WT ompA RNA thermometer cloned between constitutive promoter and gfp of pXG-10, Chlr	This study
pS-gfp	Translational fusion; mutant ompA RNA thermometer cloned between constitutive promoter and gfp of pXG-10, Chlr	This study
pXG-10-amp	pXG-10 containing bla ampicillin resistance gene, Ampr	This study
pompA	Complementation plasmid; WT ompA ORF, 5′ UTR, and promoter cloned in to pXG-10-amp, Ampr	This study
pSompA	pompA carrying site-specific stabilizing mutations in the cloned ompA 5′ UTR, Ampr	This study

^aChl^r, chloramphenicol resistant; Amp^r, ampicillin resistant.

Construction of S. dysenteriae ΔompA mutant strain.

The deletion construct used to replace S. dysenteriae ompA with kan was generated in a stepwise process, as cloning of the intact ompA open reading frame (ORF) into a plasmid proved lethal. First, nucleic acid sequences upstream and downstream of the S. dysenteriae ompA ORF were amplified and cloned directly into the pCR-BluntII-TOPO cloning plasmid of the Zero Blunt TOPO PCR cloning kit (Thermo Fisher Scientific, Waltham, MA), generating pTopo-ompA-5′ and pTopo-ompA-3′, respectively. Using the XbaI and SalI endonuclease restriction sites incorporated into the amplification oligonucleotides, the insert from pTopo-ompA-3′ was cloned into the corresponding site of pBlueScript KS (Stratagene, San Diego, CA), generating pKS-ompA-3′. Next, the insert of pTopo-0mpA-5′ was excised using the XbaI and SacI recognition sites incorporated into the amplification oligonucleotides and cloned into the corresponding sites of pKS-ompA-3′, generating pKS-ΔompA. pKS-ΔompA-kan was then generated by cloning a kan gene into the XbaI site of pKS-ΔompA, resulting in a final insert in which the nucleic acid sequence originally flanking the S. dysenteriae ompA ORF now flanks kan. The final deletion plasmid was constructed by moving the insert of pKS-ΔompA-kan into the suicide vector pCVD-442N2 (57). Allelic exchange was used as described previously (57) to replace S. dysenteriae ompA with kan. All generated plasmids as well as the final incorporated mutation were sequence verified.

Growth analysis of the S. dysenteriae ΔompA mutant.

A single colony of wild-type S. dysenteriae or the ompA mutant strain was used to inoculate a 3-ml LB broth culture. Following incubation with shaking overnight at 37°C, 200 μl of each culture was used to inoculate a fresh 3-ml LB broth culture. All cultures were incubated with shaking at 37°C and the optical density at 600 nm (OD₆₀₀) measured at each indicated time point using an ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). This analysis was performed in biological quadruplicate for each strain.

Wild-type and mutant ompA complementation plasmids.

The ompA open reading frame, along with the native promoter and 5′ UTR, was amplified from wild-type S. dysenteriae by PCR using 5′ and 3′ oligonucleotides containing AatII and XbaI endonuclease sites, respectively. The resulting product from each was purified, digested with AatII and XbaI, and cloned directly into the corresponding sites of the low-copy-number plasmid pXG-10 to generate the complementation plasmid pompA. The QuikChange protocol (Agilent, Santa Clara, CA) was used to introduce site-specific mutations into the ompA RNA thermometer cloned within pompA to generate the mutant complementation plasmid pSompA. All plasmids were sequence verified.

Translational reporter plasmids.

The translational reporters pWT-gfp and pS-gfp were constructed as detailed previously (26, 33). Briefly, complementary oligonucleotides (Integrated DNA Technologies, Skokie, IL) were annealed to generate double-stranded DNA molecules containing the wild-type or mutated ompA RNA thermometer flanked by overhanging ends compatible for ligation with DNA digested with NsiI or NheI at the 5′ or 3′ end, respectively. Each set of complementary oligonucleotides was annealed by boiling in STE buffer (0.1 mol liter⁻¹ NaCl, 10 nmol liter⁻¹ Tris-HCl, 1 mmol liter⁻¹ EDTA [pH 8.0]) for 10 min, followed by slow cooling to room temperature. The annealed DNA was then ligated into plasmid pXG-10 (47) previously digested with NsiI and NheI (New England BioLabs, Inc., Ipswich, MA), placing the RNA thermometer under investigation between the constitutive PLetO-1 promoter and the reporter gfp. Each generated plasmid was sequence verified.

DNA sequencing.

All DNA sequencing was performed by the Ohio University Genomics Facility (http://www.dna.ohio.edu/) on an Applied Biosystems 3130xL genetic analyzer.

In silico RNA structure prediction.

The secondary structures within the ompA 5′ UTR and RNA thermometer were predicted using the publicly available Mfold program with default settings in place, including a temperature of 37°C (http://unafold.rna.albany.edu/?q=mfold) (46).

RNA isolation.

RNA isolation was carried out as detailed previously (33). Following growth to mid-logarithmic phase at the indicated temperature, bacterial cultures were mixed with RNA preserving buffer (95% ethanol and 5% phenol [pH 4.5]) at a ratio of 4:1, and the mixture was incubated overnight at 4°C. Next, 3 ml of the mixture was pelleted by centrifugation at 17,000 × g for 2 min, and the pelleted bacteria were lysed by resuspension in 357.3 μl diethyl pyrocarbonate (DEPC)-treated water, 40 μl of 10% (wt/vol) sodium dodecyl sulfate (SDS), and 2.67 μl of 3 mol liter⁻¹ sodium acetate (pH 5.2) and incubated for 7 min at 90°C. Following lysis, 1 ml of TRIzol reagent (Ambion, Foster City, CA) was added to each sample and the mixture transferred to a 2-ml Phase-Lock tube (5 Prime, Inc., Gaithersburg, MD) and incubated for 5 min at room temperature. Next, 250 μl chloroform was added to each sample and the samples subjected to vigorous shaking for 1 min prior to incubation for 2 min at room temperature and centrifugation for 2 min at 17,000 × g. The nucleic acid-containing supernatant was transferred to a clean tube, 1 ml of 100% ethanol was added, and samples were incubated overnight at –80°C. The nucleic acid in each sample was pelleted by centrifugation for 17,000 × g for 15 min at 4°C and the pellet washed with 1 ml cold 70% ethanol, as described above. Following removal of the supernatant, each pellet was dried and resuspended in 53 μl nuclease-free water. The DNA present in each sample was removed using the Ambion Turbo DNA-free kit, as per the directions, and the efficient elimination of DNA was confirmed by PCR analysis. The final concentration of RNA in each sample was measured by spectrophotometry using a ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE).

Quantitative real-time PCR analyses.

Quantitative real-time PCR analyses were completed as detailed previously (33). Purified RNA was used as the template for the generation of cDNA with the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA), as per the instructions. Each cDNA sample was diluted 1:10 in nuclease-free water. Five microliters of diluted cDNA was combined with 10 μl of iTaq Universal SYBR green Supermix (Bio-Rad) and 5 μl of each primer set at an optimum concentration to generate a 20-μl reaction mixture. Amplification and detection were performed using a CFX96 real-time system (Bio-Rad, Hercules, CA) under conditions optimized for each target. Each reaction plate contained a six-point standard curve to ensure acceptable amplification efficiency and that all experimental samples fall within the linear range portion of the standard curve. The relative amount of each target transcript was calculated using the ΔΔC_T calculation method with experimental threshold cycle (C_T) values normalized to that of rrsA present in each sample and expressed relative to a selected control sample. All primers used were designed using Beacon Designer 7.5; sequences are available upon request.

Western blot analyses.

All Western blot analyses were carried out using whole-cell extracts, as detailed previously (26, 33). Briefly, bacteria were cultured to the mid-logarithmic phase of growth at the indicated temperature and the OD₆₀₀ measured using an ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). A volume corresponding to 5 × 10⁸ bacteria was transferred to a clean tube and the bacteria pelleted by centrifugation at 17,000 × g for 2 min at room temperature. Each pellet was resuspended in 200 μl Laemmli protein dye (Bio-Rad, Hercules, CA) containing 5% β-mercaptoethanol prior to boiling for 10 min. All samples were stored at –20°C until use.

For analysis, the protein present in 15 μl of each sample was resolved on a 7.5% gel by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Following pretreatment of a gel-sized piece of polyvinylidene difluoride (PVDF) membrane by soaking in methanol for 10 min and rinsing with water, proteins were transferred from the gel to the membrane at 350 mA for 1 h. The protein-containing membrane was incubated overnight in a blocking solution of 10% milk in phosphate-buffered saline (PBS) with 0.1% (wt ml⁻¹) Tween 20 (PBST) at 4°C with rocking. The membrane was then incubated in a primary antibody mixture of mouse anti-GFP antibody (Roche, Basel, Switzerland) diluted 1:1,000 or mouse anti-OmpA (58) diluted 1:500 in PBST with 5% (wt ml⁻¹) milk for 1 h at 4°C. After being washed 3 times for 15 min each in PBST, the membrane was incubated in a secondary-antibody mixture of goat anti-mouse horseradish peroxidase (HRP)-conjugated IgG (Bio-Rad, Hercules, CA) diluted 1:20,000, for GFP detection, or 1:10,000, for OmpA detection, in PBST with 5% (wt ml⁻¹) milk for 1 h at 4°C. Following washing as detailed above, detection was achieved using chemiluminescent HRP substrate (Millipore Corp., Billerica, MA) as directed and the resulting protein bands visualized and quantified using the Molecular Imager ChemiDox XRS+ imaging system (Bio-Rad). The total protein on each membrane was visualized by staining with a solution of Ponceau S dye.

Thermodenaturation curves and T_m determination.

Fluorescently labeled synthetic nucleotides were purchased from TriLink Biotechnologies (San Diego, CA), dialyzed, and diluted to a final concentration of 100 nM in a buffer composed of 150 mM NaCl, 1.0 mM MgCl₂, 20 mM morpholinepropanesulfonic acid (MOPS), and 0.01 mM EDTA (pH 6.5). Immediately before the acquisition of data, oligonucleotide samples were heated to 90°C for 1.5 min and then cooled at room temperature for 20 min. Quartz cuvettes (5.0 mm), sealed with AluminaSeal adhesive aluminum foil, were used during data acquisition. Fluorescence was measured on a Jobin Yvon Horiba FluoroMax-3 spectrophotometer equipped with an F-3004 Peltier sample heater-cooler, using an excitation wavelength of 535 nm and emission wavelength of 551 nm. Data were acquired at 0.5°C intervals, from 2°C to 92°C, with a heating rate of 1.0°C per minute. The sample chamber was purged with argon gas to prevent condensation on the cuvette at subambient temperatures. After acquisition, data were analyzed using Prism 5 for Mac OSX. Melting curve data were normalized to the maximum emission intensity in order to compare the melting curves corresponding to different oligonucleotides. Thermodenaturation points (T_m) were determined by taking the first derivatives of the melting curves, where the T_m is defined as the temperature corresponding to the maximum first derivative.

RNase-based RNA structure-probing analyses.

Enzymatic RNA structure probing was completed as detailed previously (30, 31, 59). Briefly, the RNA thermometer under investigation was generated by runoff in vitro transcription from a linearized pUC18 plasmid in which a nucleic acid sequence encoding a T7 promoter followed by the WT or stabilized ompA RNA thermometer along with the ATG and one additional codon were cloned. The synthesized RNA was then purified by resolution within, and elution from, a 6% polyacrylamide urea gel prior to being dephosphorylated using calf intestinal phosphatase (Thermo Scientific, Waltham, MA). Fifteen picomoles purified RNA was radiolabeled at the 5′ end using γ-ATP and polynucleotide kinase (PNK), as detailed previously (60), followed by resolution within and purification from a 6% polyacrylamide urea gel. Purified radiolabeled RNA was then subjected to partial (experimental lanes) or complete (size control lanes L_T1 and L_T2) digestion with the indicated RNase at the indicated temperature. The RNA-only control lanes (C) contain no enzyme and thus indicate bands resulting directly from the instability of the starting material. RNase T1 (Thermo Fisher, Waltham, MA) and RNase T2 (MoBiTec, Göttingen, Germany) and nuclease S1 (Thermo Fisher) were used at final dilutions of 1:2,000, 1:2,000, and 1:6,000, respectively. RNase T1 and RNase T2 digestions were carried out using 5× TN buffer (100 mM Tris-acetate [pH 7.5], 500 mM NaCl), while S1 digestion was carried out in the supplied buffer. The alkaline ladder was generated by exposure of the WT RNA molecule to 0.05 M NaOH and 1 mM EDTA. All reactions were stopped by the addition of formamide stop buffer, and the sample was incubated on ice prior to resolution on an 8% polyacrylamide urea gel. Alkaline ladders were generated as described previously (60).

Toeprinting ribosomal binding assay.

Ribosomal binding was evaluated using the toeprinting assays detailed previously (30, 31, 51). Briefly, the RNA thermometer along with 90 nucleotides of the downstream ompA open reading frame were cloned into a plasmid. Using this plasmid as the template, in vitro transcription was carried out to generate a target RNA molecule. Following purification by resolution on, and extraction from, a 6% polyacrylamide urea gel, 1 pmol the target RNA molecule was incubated with 2 pmol a radiolabeled oligonucleotide complementary to the cloned portion of the ompA ORF. The target RNA along with the annealed radiolabeled primer were incubated for 10 min at the indicated temperature prior to the addition of 30S ribosome (experimental) or buffer alone (control) and 16 pmol uncharged tRNA fMet (Sigma-Aldrich, St. Louis, MO). Next, 2 μl of Moloney murine leukemia virus (MMLV) mix (high-performance reverse transcriptase; Epicentre, Madison, WI) was added to each reaction, and the mixture was incubated for 10 min at 37°C. The reaction was terminated by the addition of formamide stop buffer and the products visualized on an 8% polyacrylamide urea gel. The size ladder was generated using the radiolabeled oligonucleotide detailed above and the cloned RNA thermometer as the template in a Sequenase reaction (Thermo Fisher Scientific, Waltham, MA).

In vitro tissue culture-based virulence assays.

All tissue culture assays were performed using Henle cells (intestine 407; ATCC-VR-106) cultured at 37°C under 5% atmospheric CO₂ in Gibco minimal essential medium (Invitrogen Corp., Carlsbad, CA) with 10% fetal bovine serum (FBS), 2 mM glutamine, and 1× nonessential amino acids (Lonzo, Basel, Switzerland).

(i) Plaque assays. Plaque assays were performed as described previously (35), with minor modifications. Briefly, the S. dysenteriae strain being investigated was cultured to the mid-logarithmic phase of growth at 37°C with 0.01% deoxycholate (DOC) to prepare the bacteria for invasion and, if appropriate, 50 μg ml⁻¹ ampicillin to maintain the indicated plasmid. The OD₆₀₀ of the culture was measured using an ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE), and 5 × 10⁴ bacteria were added to a single well within a six-well tissue culture plate containing an 80% confluent monolayer of Henle cells, 2 ml Henle cell medium, and, when appropriate, 50 μg ml⁻¹ ampicillin. Following the addition of bacteria, tissue culture plates were centrifuged at room temperature for 10 min at 600 × g in a Beckman Coulter Allegra 25R centrifuge (Brea, CA) and incubated at 37°C under 5% CO₂ for 90 min. Following this incubation, the supernatant was removed, the monolayer was washed with Henle cell medium, and 2 ml of Henle cell medium containing 20 μg ml⁻¹ gentamicin, 0.3% glucose, and, when appropriate, 50 μg ml⁻¹ ampicillin were added to each well. Following incubation for 72 h at 37°C under 5% CO₂, the supernatant was removed, and the monolayer was stained with Giemsa-Wright stain (Camco, Ft. Lauderdale, FL) and washed twice with double-distilled water. Plaques present in each well were counted by eye. Diluted cultures used to inoculate the Henle cell monolayer for the in vitro plaque were plated to ensure equal inoculation of each tissue culture well within a given experiment.

(ii) Invasion assay. As detailed previously (42), the in vitro invasion assay is a modification of the plaque assay described above. Following growth as described above, 2 × 10⁸ bacteria were used to infect a 60% confluent monolayer of Henle cells. After centrifugation as detailed above, the plates were incubated for 30 min at 37°C under 5% CO₂ prior to the removal of the supernatant, washing of the monolayer, and the addition of Henle cell medium containing 20 μg ml⁻¹ gentamicin, 0.3% glucose, and, when appropriate, 50 μg ml⁻¹ ampicillin. Plates were then incubated for 90 min at 37°C under 5% CO₂ prior to being stained with Giemsa-Wright stain (Camco, Ft. Lauderdale, FL), washed twice with double-distilled water, and allowed to dry at room temperature. One hundred Henle cells not in contact with any neighboring cell were randomly selected for scoring. Invasion was scored if the selected Henle cell contained ≥3 bacteria.

Sequence alignment.

Sequence alignment was performed using the CloneManager Software package and the sequenced genomes of the following bacteria: S. dysenteriae (accession no. CP000034.1), S. flexneri 2a strain 301 (accession no. AE005674.2), S. boydii CDC3083-94 (accession no. CP001063.1), S. sonnei (accession no. CP000038.1), E. coli O127:H6 E2348/69 (EPEC; accession no. FM180568.1), E. coli O157:H7 strain EDL993 NC (EHEC; accession no. AE005174.2), E. coli S88 (NMEC; accession no. CU928161.2), E. coli CFT073 (UPEC; accession no. AE014075.1), Salmonella enterica (accession no. AE006468.2), and Citrobacter koseri ATCC BAA-895 (accession no. NC_009792).

Statistical analysis.

Unless otherwise noted, all presented analyses were performed in biological triplicate or quadruplicate. An F-test was used to determine variance between groups, followed by the appropriate Student t test to determine significance.