Disruption of type III secretion in Salmonella enterica serovar Typhim
http://www.100md.com
《核酸研究医学期刊》
1 Department of Molecular, Cellular and Developmental Biology, Yale University, New Haven, CT 06520, USA, 2 Department of Pediatrics, Division of Infectious Diseases, Yale University School of Medicine, New Haven, CT 06520, USA and 3 Section of Microbial Pathogenesis, Yale University School of Medicine, New Haven, CT 06536, USA
*To whom correspondence should be addressed. Tel: +1 314 286 2912; Fax: +1 314 286 2895; Email: McKinney_J@kids.wustl.edu
Present address:
Jeffrey S. McKinney, Departments of Pediatrics and Molecular Microbiology, Washington University, Saint Louis, MO 63110, USA
ABSTRACT
The type III secretion system involved in Salmonella enterica serovar Typhimurium invasion of host cells has been disrupted using inducibly expressed oligonucleotide external guide sequences (EGSs) complementary to invB or invC mRNA. These EGSs direct single site cleavage in these mRNAs by endogenous RNase P, and their expression in Salmonella results in invC mRNA and InvC protein depletion, decreased type III secretion and interference with host cell invasion. Comparison of these effects with those from studies of Salmonella invB and invC mutants suggests that invB EGSs have polar effects on invC mRNA.
INTRODUCTION
External guide sequence (EGS) oligonucleotides target complementary mRNA for specific cleavage catalyzed by RNase P (1). EGS oligonucleotides require an accessible single-stranded region on their target mRNA to base-pair with and create the stem structure recognized as a cleavage substrate by RNase P (1). Using EGSs complementary to essential genes, Escherichia coli viability can be decreased in a manner which is EGS oligonucleotide sequence specific, dose dependent and dependent on time elapsed after EGS expression (2). Here, EGS studies are extended to Salmonella, using EGSs complementary to two Salmonella pathogenicity island SPI-1 genes (3), invB and invC, neither of which are essential for bacterial viability (4). The invB and invC DNA sequences occur directly adjacent to each other in the multigene pathogenicity island SPI-1 of Salmonella, with the last nucleotide in the final codon of invB also serving as the first nucleotide in the first codon of invC (4). Prior studies of Salmonella invB and invC mutants have shown that invC is required for host cell invasion and that the gene encodes a protein with ATPase activity (4). The ATPase encoded by invC is postulated to provide energy to power the type III secretion system involved in host cell invasion (4) and pathogenesis (5) by Salmonella. In contrast, invB does not appear necessary for invasion. InvB is a type III secretion chaperone specific for SipA, a translocated Salmonella protein which facilitates actin rearrangements within infected eukaryotic cells (6). Mutations in invB do not alter the secretion of other type III secreted proteins (6) and do not disrupt invasion (4). Using a tightly regulated inducible EGS expression system in Salmonella (7), we show that EGSs complementary to either invB or invC mRNA can disrupt type III secretion and Salmonella invasion assayed in vitro.
MATERIALS AND METHODS
Plasmids and bacterial strains
The EGSs listed below were cloned as previously described into high copy number EGS expression plasmids, derived from pUC19 (2) or into low copy number plasmids derived from pWKS30 (8). These plasmids were transformed into the Salmonella enterica serovar Typhimurium strain SB300A#1 (7). SB300A#1 has a T7 RNA polymerase gene integrated with an adjacent araC-P(BAD) control element into the bacterial chromosome of parent strain SB300. SB300A#1 allows tightly controlled arabinose-inducible T7 promoter-driven transcription of our EGSs in Salmonella (7). The invA-deficient Salmonella strain SB136 (4), which is disrupted for type III secretion, was used as a control. An invC deletion mutant Salmonella (J. E. Galán and Y. Akeda) was used as a negative control strain for studies of InvC intracellular protein level and of type III secretion. Salmonella was grown in 0.3 M NaCl Luria–Bertani (LB) medium. Liquid culture incubation conditions and EGS induction with arabinose at 0.2% final concentration are as previously described (7). Following addition of arabinose for EGS induction, Salmonella liquid cultures were grown to late log phase prior to northern blot analysis, assay of Salmonella type III secretion or quantification of bacterial entry, as detailed below.
Design of external guide sequences
EGS oligonucleotides were designed to be complementary to single-stranded regions of invB and invC mRNA, followed by an additional 3'-ACCA EGS terminal sequence. This strategy allows formation of a duplex EGS–mRNA molecule recognized as a substrate by endogenous RNase P with resultant cleavage of target mRNA (9). The individual EGS oligonucleotide sequences were named according to their predicted site of target mRNA cleavage. For example, invB 108 EGS (5'-AAUGCAAAUAAAUCCacca-3') is complementary to invB mRNA nucleotides 108–122 (5'-GGAUUUAUUUGCAUU-3') and will result in RNase P cleavage of invB mRNA at nucleotide number 108. The other invB or invC EGS sequences were: invC 98 EGS (5'-GGCGUGAUUUCACAAacca-3'), invC 269 EGS (5'-ACCGCGCCUAAUACCacca-3') and invC 293 EGS (5'-ACGAUUUUCCCUGUCacca-3'). Two previously characterized EGSs which are not complementary to invB or invC were also used: synthC5 EGS 21 and synthC5 EGS 45 (2). The EGSs synthC5 EGS 21 and synthC5 EGS 45 are complementary to, and can guide the RNase P cleavage of, mRNA used for the recombinant synthesis of the C5 protein subunit of the RNase P holoenzyme of E.coli, but are not complementary (containing at least five unpaired nucleotides per EGS) to the mRNA encoding C5 in Salmonella. Herein, the EGSs synthC5 EGS 21 and synthC5 EGS 45 are referred to as synthC5 EGS 1 and 2, respectively.
Partial RNase T1 digest mapping of invB and invC mRNA
Single-stranded regions of invB and invC mRNA were identified using RNase T1 digestion (1). Two mRNAs were digested: (i) a joint in vitro transcript containing invC mRNA 3' to invB mRNA, transcribed from the plasmid pSB553 (4) DNA after digestion with BamHI; and (ii) an invC mRNA in vitro transcript alone, expressed from plasmid pIC001 (a pSB553 derivative, with invB coding sequence removed via KpnI and BspEI excision) DNA after digestion with EcoRI.
In vitro RNase P assays
Assays of mRNA cleavage in vitro by RNase P were performed as previously described (10), using the EGS sequences and the invB and invC mRNA targets detailed above. RNase P M1 RNA was folded in a buffer containing 10 mM magnesium, using a heat block to first heat the sample at 65°C for 5 min and then slowly cool the sample to room temperature. For conditions of substrate excess, reagent concentrations were: 11 fmol labeled substrate (1100 c.p.m.), 1, 5 and 10 pmol EGS, and 1 pmol of enzymatically active recombinant E.coli RNase P M1 RNA. For conditions of limited substrate, 10 fmol of labeled substrate RNA (1000 c.p.m.) and 50, 100 and 500 fmol of EGS were used. Samples were electrophoresed in 5% polyacrylamide–7 M urea gels.
Northern blots
Northern blots were performed on total RNA extracts of Salmonella, using previously published techniques (11). Each lane of a 2.5% agarose gel was loaded with 4 μg of total RNA. The UV transillumination pattern of rRNA bands after separation of each sample on an agarose gel revealed similarity in rRNA band patterns in terms of both gross quantity and quality. Northern blot probes were 5'-end-labeled DNA oligonucleotide probes. They included probes complementary to invC mRNA, 5S rRNA and each of the four invB and invC EGS oligonucleotides listed above. In each case, 8 pmol of oligonucleotide labeled with 4 pmol (30 μCi) of ATP was used per 40 ml of rapid hybridization buffer (Amersham). Signal was detected using a phosphoimager (Fuji) and quantitated using image analysis software (Fuji ImageGauge). Quantitative results are reported as the ratio, expressed as a percentage, of northern blot invC mRNA band signal intensity for the matched culture specimens of a given Salmonella strain with EGS induction versus without EGS induction.
Assay of Salmonella type III secretion
Salmonella culture supernatant proteins were prepared and analyzed as previously described (12). Western blots were probed with polyclonal antibodies against SipB and SipC and chemiluminescent signals produced using ECL Plus western blotting detection reagents from Amersham Biosciences. Signal was detected using autoradiograph film (Kodak) and quantitated using image analysis software (Fuji ImageGauge). Quantitative results are reported as the ratio, expressed as a percentage, of western blot band signal intensity for the matched culture specimens of a given Salmonella strain with EGS induction versus without EGS induction.
Matched culture specimens were employed as internal controls for western blots. Specifically, a given Salmonella transformant was grown to early log phase in a single liquid culture, then split into paired cultures which were grown simultaneously either with or without arabinose induction of EGS expression. Equal volumes of paired cultures were harvested for protein preparation at the same point of their late log phase growth, as assessed by optical density. Equal volumes of these protein preparations were loaded per well for western blot analysis.
Assay of intracellular InvC protein level
Salmonella cultures in liquid media were pelleted and resuspended in one-twentieth volume of phosphate-buffered saline (PBS)–Tris (77 mM Tris–HCl, pH 8.0). Samples were denatured by boiling with SDS–PAGE loading buffer and separated on a 9% polyacrylamide–SDS gel. Western blots were probed with a polyclonal antibody against InvC (J. E. Galán and Y. Akeda), and chemiluminescent signals produced using ECL Plus western blotting detection reagents from Amersham Biosciences. Signals were detected using autoradiograph film (Kodak). Quantitation of results, using matched culture specimens, was performed as described for type III secretion assays, above.
Quantification of bacterial entry
Entry of different Salmonella strains into Henle-407 cells in a gentamicin protection assay of bacterial entry into host tissue culture cell monolayers was performed and quantified as previously described (13).
RESULTS
Design of EGSs for invB and invC mRNA in vitro
To design the EGSs reported here, mRNA transcripts of invB and invC made in vitro were mapped using partial RNase T1 nuclease digestion to suggest EGS-accessible single-stranded mRNA regions. The first nucleotide of the start codon of each gene is labeled as nucleotide 1; single-stranded guanine residues of invB mRNA were identified via partial RNase T1 nuclease digestion at invB nucleotides 108 and 217. For invC mRNA, single-stranded guanines were identified at invC nucleotides 98, 237, 269 and 293. Given uncertainty about whether invC mRNA exists in cells independently from invB mRNA or as a joint transcript with invB, both possibilities were examined in RNase T1 digestion in vitro. RNase T1 digestions were performed on two in vitro transcripts: invC mRNA alone, as well as a tandem transcript of invC mRNA immediately 3' to invB mRNA. Single-stranded regions of invC identified in the joint invB–invC in vitro transcript were notably also found for the invC in vitro transcript alone (data not shown). EGS oligonucleotides were designed to be complementary to the RNase T1-accessible mRNA sequences invB 108–122, invC 98–112, invC 269–283 and invC 293–307, and were named for their predicted nucleotide cleavage sites by RNase P: invB 108, invC 98, invC 269 and invC 293, respectively.
RNase P-specific cleavage of mRNA in vitro
RNase P hydrolyzes the phosphodiester bond (in the target mRNAs) that precedes the first base pair in the 5' end of the target mRNA–EGS complex, akin to the site-specific cleavage reaction the enzyme catalyzes in the 5' processing of precursor tRNA (9). The reaction ingredients for RNase P assays in vitro were E.coli RNase P, internally radiolabeled invB and invC mRNA target transcribed in vitro, and EGSs complementary to portions of invB or invC mRNA. All four EGSs guide RNase P to cleave the mRNA at the predicted sites of EGS mRNA hybridization, yielding appropriately sized 5' and 3' cleavage products (Fig. 1). RNase P cleavage of mRNA increases with increasing EGS dose, with the EGSs invB 108 and invC 293 most efficient at guiding mRNA cleavage in vitro in conditions of limited substrate (data not shown).
Figure 1. RNase P–EGS cleavage of mRNA in vitro. Target substrates for cleavage include a joint transcript including both invB and invC mRNA (lane 1), and a transcript of invC mRNA alone (lane 6). The former was incubated with increasing amounts of the invB EGS 108 (lanes 2–4); the latter as a target for invC EGSs invC 98 (lanes 7–9), invC 269 (lanes 10–12) and invC 293 (lanes 13–15). Lanes labeled N (lanes 5 and 16) lack any EGS, but do have active RNase P with the invB and invC joint transcript (lane 5) or the invC transcript (lane 16) and show no non-specific target cleavage in the absence of EGS. Cleavage products were separated by size using electrophoresis in a 5% polyacrylamide–7 M urea gel. Cleavage product sizes are consistent with RNase P cleavage occurring at the predicted site at the 5' end of the mRNA region to which each EGS hybridizes. Predicted sizes of reaction products following RNase P enzymatic cleavage of in vitro transcripts are listed on the left and right of the image (e.g. invB and invC joint transcript mRNA cleavage products of 585 and 123 nucleotides for EGS invB 108; invC transcript mRNA cleavage products of 485 and 198 nucleotides for EGS invC 98, etc.).
Specific mRNA targeting in Salmonella
Induction of expression of invB or invC EGSs in Salmonella is followed by a decrease in invC mRNA compared with identical Salmonella transformants lacking EGS induction. Northern blots of equal microgram amounts of total RNA isolated from various Salmonella liquid cultures are shown in Figure 2, where northern blot signals for invC mRNA in matched Salmonella cultures decreased between 27 and 50% following relevant EGS induction. The invB or invC EGS expression effects appear specific, in that there is no similar change detected in the level of constitutive 5S rRNA after EGS induction (Fig. 2). Arabinose was used to induce Salmonella EGS expression. It did not have as pronounced an effect on invC mRNA in the absence of EGS expression plasmids, either in the case of Salmonella which was not transformed with an EGS expression plasmid, or in the case of Salmonella for which the EGS expression plasmid was presumably lost after prolonged culture (Fig. 2). Plasmid maintenance in transformants was assessed by parallel quantitative plating on LB or LB ampicillin plates as previously reported (2,11). Plasmid loss was considered to have occurred when colony counts on LB plates were greater than on LB ampicillin plates by at least an order of magnitude. As previously demonstrated for other EGS transcription in this Salmonella system (7), northern blot probes complementary to EGS oligonucleotides detected EGS expression only after arabinose induction (data not shown).
Figure 2. Northern blots in Salmonella with inducible expression of EGS molecules. RNA was isolated from Salmonella, electrophoresed in a 2.5% agarose gel, and probed for invC mRNA or for constitutively produced non-targeted 5S rRNA. The partial invC in vitro transcript expressed from plasmid pIC001 DNA after digestion with EcoRI (as described for T1 digest mapping) serves as a size marker (683 nt). The invA mutant is SB136 (also used for invasion assays). SB300 is the parent Salmonella strain, from which SB300A#1 was constructed. Other lanes are paired by SB300A#1 Salmonella transformant type, either with (+) or without (–) the addition of arabinose for the induced expression of the EGS molecules listed. Longer term induction (++) was suspected to be accompanied by the loss of EGS expression plasmid invC 293, and an accompanying lack of effect on target mRNA. Note the decrease in invC mRNA after the induction of EGS expression, whereas non-targeted constitutive 5S rRNA levels are independent of EGS expression.
The effect of expression of EGSs on type III secretion by Salmonella
A standard functional assay of InvC-dependent type III secretion was employed, in which proteins secreted by the type III secretion system were measured in cell culture supernatants (12). These secreted proteins, SipB and SipC, are also encoded in the Salmonella pathogenicity island I gene complex and are reviewed elsewhere (14,15).
A panel of Salmonella transformants, containing EGSs in various inducible expression plasmid vectors, was used for InvC-dependent type III secretion assays. The concurrent expression of two EGSs using any one of four high copy number plasmids derived from pUC19 (2) consistently decreased SipB secretion by 65% from that detected for the same Salmonella transformant grown in parallel under non-EGS-inducing conditions, based on the relative signal intensities of SipB western blot bands for a given Salmonella strain with versus without EGS induction (Fig. 3). Using the same method of quantitative analysis, induction of the same set of EGS pairs in Salmonella transformed with a low copy number plasmid derived from pWKS30 showed a 20–30% decrease in secretion (data not shown), as did any of the four EGSs when expressed alone from pUC19-derived plasmids (data not shown). These findings are consistent with prior studies of EGS dose–response features in E.coli comparing EGS expression plasmids encoding single versus multiple EGSs (2) and comparing EGS expression plasmids with strong versus weak promoters (10).
Figure 3. Western blots of cell culture supernatants from Salmonella with (+) or without (–) the induced expression of EGS molecules listed. Western blot detects the proteins SipB and SipC, secreted into cell culture supernatant by the type III secretion system and subsequently electrophoresed in a 9% polyacrylamide–SDS gel. The transformants shown contain a pUC19-derived high-copy number plasmid (2) from which two EGSs are concurrently expressed. InvBB denotes inducible expression of invB 108 and invB 108 EGSs in tandem, InvBC denotes invB 108 and invC 98 EGS expression, InvCB denotes invC 293 and invB 108 EGS expression, and invCC denotes invC 293 and invC 98 EGS expression. The control concurrently expresses two EGSs targeting the mRNA used for synthetic C5 protein over-expression in E.coli. Migration patterns of SipB and SipC bands relative to protein molecular weight 66 and 45 kDa markers are as shown in Figure 4.
To assess sequence specificity, the negative control EGSs, synthC5 EGS 2 and 1, against the synthetic C5 component of E.coli RNase P, described above and in McKinney et al. (2), was also used. Induction of this negative control EGS was accompanied by a <10% decrease in SipB secretion as compared with non-induced parallel cultures (Fig. 3). In addition, the loss of an invB or invC EGS expression plasmid from a bacterial strain was accompanied by a loss of previously observed inhibitory effects on secretion (data not shown).
The effect of EGSs on intracellular InvC protein in Salmonella
A polyclonal antibody (J. E. Galán and Y. Akeda) raised against recombinant InvC protein was used for western blot analysis of cellular InvC protein level as an assessment of EGS effects (Fig. 4, top). A band consistent with the 47 kDa InvC protein is found in the Salmonella strain SB300A#1 with no EGS. An SB300 strain (constructed by Y. Akeda) in which InvC expression has been disrupted via an invC deletion mutation serves as a negative control for InvC expression. Intracellular expression of InvC was also assessed in SB300A#1 transformed with the EGS expression plasmid for the concurrent expression of either the EGSs invB 108/invC 98 or the negative control EGSs synthC5 EGS 2 and 1 (also shown in Fig. 3). After the induction of EGS expression, the relevant protein band is not detected for SB300A#1 with the invB/invC EGSs. No effect on the detection of this protein band is seen after the induction of the negative control EGS. Figure 4 is representative of repeated assays in that the decrease in putative 47 kDa InvC signal cannot be explained by a general decrease in the intensity of other bands detected by the polyclonal antibody. Indeed, for the two cases of EGS induction shown here, invB/invC EGS induction is accompanied by decreased 47 kDa InvC signal in the context of strong signals for other bands, whereas negative control EGS induction is accompanied by strong 47 kDa InvC signal in the context of relatively weak signals for other bands.
Figure 4. Western blots for Salmonella intracellular InvC (top) and for SipB and SipC secreted into Salmonella cell culture supernatants (bottom, as for Fig. 3). SB300A#1 without an EGS expression plasmid (no EGS) and a SB300 invC deletion mutant which does not express InvC (invC) (produced by J. E. Galán and Y. Akeda) are shown to the left of protein molecular weight markers. At right: SB300A#1 transformed with the high copy number plasmids for arabinose-inducible expression of either invB 108 and invC 98 EGSs (InvB/C EGS) or the control EGSs (Control EGS) described for Figure 3. EGS induction status by arabinose addition is shown as (+) or (–). Arrowheads denote the predicted locations of InvC, SipB and SipC after electrophoresis in 9% polyacrylamide–SDS gels.
Cell supernatants from the same experiment were used for type III secretion assays (Fig. 4, bottom) as above. In each case, detection of the band representing intracellular InvC protein correlates with type III secretion. Type III secretion is decreased when intracellular InvC protein levels are decreased, either by the static invC deletion mutation, or by the dynamic disruption of InvC expression following the induction of the invB/invC EGSs. While the presumptive InvC protein band disappears in both situations, the decrease in type III secretion is more pronounced following the static deletion of the invC gene than following the induction of the invB/invC EGS. This functional assay of InvC cellular activity suggests that our invB/invC EGS system inhibits InvC-dependent type III secretion less completely than does the invC deletion-mediated ablation of InvC expression. Wild-type levels of secretion are observed for SB300A#1 with no EGS plasmid, without induction of invB/invC EGSs, or with negative control EGS expression.
EGS impact on Salmonella invasion into host cells
Salmonella strains containing EGSs in various inducible expression plasmid vectors were also tested for their ability to invade Henle-407 cells in tissue culture. In each case, a given Salmonella strain was grown in parallel liquid cultures, either with arabinose added to induce EGS induction or without the addition of arabinose. For invasion, these Salmonella strains were first incubated with Henle cells for 45 min in Hanks buffered salt solution. This was followed by a 2 h treatment with gentamicin to kill extracellular bacteria, and subsequent washes with tissue culture buffer. Invasion was quantified as the percentage of bacteria inoculated into the tissue culture wells which were recovered from lysed Henle cells.
As shown in Table 1, a control Salmonella strain, SB136, with a null mutation in invA and a previously documented functional defect in type III secretion and host cell invasion (4), achieves <1% invasion into Henle cells, independent of arabinose addition. Salmonella SB300A#1 transformants which express an EGS complementary to mRNA encoding synthetic C5 (also shown as a negative control for type III secretion assays, above) have a 20% rate of invasion, with a small decrease in invasion observed after arabinose addition. Salmonella SB300A#1 transformants which inducibly express EGSs against invB or invC from a high copy number pUC19-derived plasmid exhibit an 8–12% rate of invasion without arabinose addition, which decrease to 2.8–3.5% after arabinose induction of EGS expression. The same EGS constructs, expressed from low copy number pWKS30-derived plasmids, did not affect invasion rates after EGS induction.
Table 1. Invasion of Henle-407 cells by Salmonella strains
DISCUSSION
The pathogenicity island genes invC and invB of Salmonella provide intriguing targets for gene product disruption. In the case of invC, mutagenesis studies clearly show that the ATPase encoded by invC is required for type III secretion in assays in vitro and is important for pathogenicity in animal models (4,5). Following appropriate EGS expression, we observe a decrease in invC mRNA, InvC intracellular protein, InvC-powered type III secretion and type III secretion-dependent host cell invasion.
The inhibition of type III secretion and of Salmonella invasion using EGSs to disrupt invC mRNA is less complete than that resulting from invC deletion mutagenesis. This suggests that a certain critical level of mRNA disruption is able to partially inhibit type III secretion and host cell invasion. The level of mRNA disruption required for phenotypic changes probably varies for different target mRNAs, depending on factors such as the ratio of EGS to target mRNA (2,11), the relative efficiency of various EGSs and the functional reserve capacity a cell has for a given target mRNA and the protein that mRNA encodes. The EGSs reported here show greater phenotypic effects when expressed from high copy number, rather than low copy number, plasmids. This is consistent with prior EGS dose–response observations in bacteria. Phenotypic effects in E.coli are greater following EGS expression driven by a strong promoter as compared with a weak promoter (11), and the concomitant expression of different EGSs in E.coli results in phenotypic effects exhibiting additive synergy (2). The application of EGS technologies to regulate gene expression in bacteria in in vivo models of infection may benefit from EGS expression plasmids which can be stably maintained in bacteria within an animal. We produced the low copy number plasmids reported here in an initial effort toward this end. The fact that EGSs expressed from our low copy number plasmid system did partially inhibit type III secretion but showed no apparent effect on host cell invasion suggests a possible threshold effect, in which inhibition of type III secretion must reach a critical threshold to result in inhibition of cell invasion.
In contrast to the static effects on gene product disruption produced by mutagenesis techniques, our techniques of gene product disruption involve a more dynamic process of EGS-guided, RNase P-mediated, mRNA cleavage commencing after EGS induction. As observed in past studies of EGSs complementary in sequence to essential genes (2) or to antimicrobial resistance genes (11) in E.coli, a lag occurs between initiation of EGS expression and the observed molecular and phenotypic effects. This lag is consistent with the time required to accrue sufficient amounts of mRNA cleavage for observable phenotypic effects.
Reconciling the relative magnitude of effects we observe after EGS expression upon invC gene product expression and functions (Fig. 5) requires consideration of the kinetics of several steps of cellular physiology and our measurement methods. For example, the assays we employed showed a smaller relative effect of EGSs on type III secretion than on invasion. Our invasion assays detect host cell invasion over a short time period commencing at a time when differences between conditions of EGS or no EGS are relatively great. In contrast, our secretion assays detect the accumulated SipB or SipC secreted into supernatants throughout the period of EGS induction and include a lag time when EGS effects are minimal. Accordingly, the relative difference between conditions of EGS expression and no EGS expression which we detect for secretion might be predicted to be less than the relative difference we detect for invasion.
Figure 5. Summary of invB and invC EGSs and their effects. A schematic of the invB and invC mRNA transcripts is shown, with the predicted sites of RNase P-mediated cleavage identified for each invB or invC EGS (invB 108 EGS, invC 98 EGS, invC 269 EGS and invC 293 EGS). The last nucleotide in the final codon of invB also serves as the first nucleotide in the first codon of invC. Effects of these EGSs on Salmonella invC mRNA, type III secretion, InvC protein and host cell invasion are summarized, and relevant figures cited. Further details are provided in the text.
Effects following expression of invB EGSs were somewhat surprising given the results of prior invB studies. InvB is a type III secretion chaperone specific for SipA (6). Mutations of invB do not affect the type III secretion of SipB or SipC (6) or the invasion of host cells (4). Yet, an EGS complementary in nucleotide sequence to invB and which cleaves invB mRNA in vitro, does inhibit type III secretion of SipB and SipC and host cell invasion. Furthermore, this EGS against invB decreases invC mRNA levels in Salmonella.
The coding regions of the invB and invC genes overlap, such that the 3' nucleotide of the final codon of invB serves as the 5' nucleotide of the first codon of invC. We postulate that the effects of invB EGSs are consistent with the existence of a joint invB–invC transcript in Salmonella (Fig. 5), which becomes destabilized or functionally disrupted after cleavage of its invB portion. If this is true, the fact that an EGS targeting invB mRNA also has effects on invC suggests that, in polycistronic mRNA transcripts, EGS targeting the 5' portions of mRNAs may also have effects on mRNAs encoded by genes 3' to the EGS target site.
Ultimately, pathogenicity island gene product disruption may lead to novel anti-pathogenicity agents and strategies. Since these genes are not essential to bacterial viability, interfering with their functions could conceivably decrease morbidity without the relatively broad-spectrum effects and selective pressure for resistance seen with many current antimicrobial agents.
ACKNOWLEDGEMENTS
We thank Yukihiro Akeda for antibody against InvC and for the invC Salmonella mutant. This research was conducted while J.M. was a Pfizer postdoctoral fellow. The work was supported by an NIH research training program (NIH T32 AI07210-17), a Yale Children’s Health Center Scholar Award and a Washington University Digestive Diseases Research Core Center Grant (DDRCC P30 DK52574) to J.M., and NIH grants AI030492 to J.E.G. and GM19422 to S.A.
REFERENCES
Guerrier-Takada,C. and Altman,S. (2000) Inactivation of gene expression using ribonuclease P and external guide sequences. Methods Enzymol., 313, 442–456.
McKinney,J., Guerrier-Takada,C., Wesolowski,D. and Altman,S. (2001) Inhibition of Escherichia coli viability by external guide sequences complementary to two essential genes. Proc. Natl Acad. Sci. USA, 98, 6605–6610.
Groisman,E.A. and Ochman,H. (1996) Pathogenicity islands: bacterial evolution in quantum leaps. Cell, 87, 791–794.
Eichelberg,K., Ginocchio,C.C. and Galán,J.E. (1994) Molecular and functional characterization of the Salmonella typhimurium invasion genes invB and invC: homology of InvC to the F0F1 ATPase family of proteins. J. Bacteriol., 176, 4501–4510.
Porter,S.B. and Curtiss,R.R. (1997) Effect of inv mutations on Salmonella virulence and colonization in 1-day-old White Leghorn chicks. Avian Dis., 41, 45–57.
Bronstein,P.A., Miao,E.A. and Miller,S.I. (2000) InvB is a type III secretion chaperone specific for SspA. J. Bacteriol., 182, 6638–6644.
McKinney,J., Guerrier-Takada,C., Galán,J. and Altman,S. (2002) Tightly regulated gene expression system in Salmonella enterica serovar Typhimurium. J. Bacteriol., 184, 6056–6059.
Wang,R.F. and Kushner,S.R. (1991) Construction of versatile low-copy number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene, 100, 195–199.
Forster,A.C. and Altman,S. (1990) External guide sequences for an RNA enzyme. Science, 249, 783–786.
Guerrier-Takada,C., Li,Y. and Altman,S. (1995) Artificial regulation of gene expression in Escherichia coli by RNase P. Proc. Natl Acad. Sci. USA, 92, 11115–11119.
Guerrier-Takada,C., Salavati,R. and Altman,S. (1997) Phenotypic conversion of drug-resistant bacteria to drug sensitivity. Proc. Natl Acad. Sci. USA, 94, 8468–8472.
Russmann,H., Kubori,T., Sauer,J. and Galán,J.E. (2002) Molecular and functional analysis of the type III secretion signal of the Salmonella enterica InvJ protein. Mol. Microbiol., 46, 769–779.
Galán,J.E. and Curtiss,R.R. (1989) Cloning and molecular characterization of genes whose products allow Salmonella typhimurium to penetrate tissue culture cells. Proc. Natl Acad. Sci. USA, 86, 6383–6387.
Collazo,C.M. and Galán,J.E. (1997) The invasion-associated type III system of Salmonella typhimurium directs the translocation of Sip proteins into the host cell. Mol. Microbiol., 24, 747–756.
Komoriya,K., Shibano,N., Higano,T., Azuma,N., Yamaguchi,S. and Aizawa,S.I. (1999) Flagellar proteins and type III-exported virulence factors are the predominant proteins secreted into the culture media of Salmonella typhimurium. Mol. Microbiol., 34, 767–779.(Jeffrey S. McKinney*,1,2, Haifeng Zhang1)
*To whom correspondence should be addressed. Tel: +1 314 286 2912; Fax: +1 314 286 2895; Email: McKinney_J@kids.wustl.edu
Present address:
Jeffrey S. McKinney, Departments of Pediatrics and Molecular Microbiology, Washington University, Saint Louis, MO 63110, USA
ABSTRACT
The type III secretion system involved in Salmonella enterica serovar Typhimurium invasion of host cells has been disrupted using inducibly expressed oligonucleotide external guide sequences (EGSs) complementary to invB or invC mRNA. These EGSs direct single site cleavage in these mRNAs by endogenous RNase P, and their expression in Salmonella results in invC mRNA and InvC protein depletion, decreased type III secretion and interference with host cell invasion. Comparison of these effects with those from studies of Salmonella invB and invC mutants suggests that invB EGSs have polar effects on invC mRNA.
INTRODUCTION
External guide sequence (EGS) oligonucleotides target complementary mRNA for specific cleavage catalyzed by RNase P (1). EGS oligonucleotides require an accessible single-stranded region on their target mRNA to base-pair with and create the stem structure recognized as a cleavage substrate by RNase P (1). Using EGSs complementary to essential genes, Escherichia coli viability can be decreased in a manner which is EGS oligonucleotide sequence specific, dose dependent and dependent on time elapsed after EGS expression (2). Here, EGS studies are extended to Salmonella, using EGSs complementary to two Salmonella pathogenicity island SPI-1 genes (3), invB and invC, neither of which are essential for bacterial viability (4). The invB and invC DNA sequences occur directly adjacent to each other in the multigene pathogenicity island SPI-1 of Salmonella, with the last nucleotide in the final codon of invB also serving as the first nucleotide in the first codon of invC (4). Prior studies of Salmonella invB and invC mutants have shown that invC is required for host cell invasion and that the gene encodes a protein with ATPase activity (4). The ATPase encoded by invC is postulated to provide energy to power the type III secretion system involved in host cell invasion (4) and pathogenesis (5) by Salmonella. In contrast, invB does not appear necessary for invasion. InvB is a type III secretion chaperone specific for SipA, a translocated Salmonella protein which facilitates actin rearrangements within infected eukaryotic cells (6). Mutations in invB do not alter the secretion of other type III secreted proteins (6) and do not disrupt invasion (4). Using a tightly regulated inducible EGS expression system in Salmonella (7), we show that EGSs complementary to either invB or invC mRNA can disrupt type III secretion and Salmonella invasion assayed in vitro.
MATERIALS AND METHODS
Plasmids and bacterial strains
The EGSs listed below were cloned as previously described into high copy number EGS expression plasmids, derived from pUC19 (2) or into low copy number plasmids derived from pWKS30 (8). These plasmids were transformed into the Salmonella enterica serovar Typhimurium strain SB300A#1 (7). SB300A#1 has a T7 RNA polymerase gene integrated with an adjacent araC-P(BAD) control element into the bacterial chromosome of parent strain SB300. SB300A#1 allows tightly controlled arabinose-inducible T7 promoter-driven transcription of our EGSs in Salmonella (7). The invA-deficient Salmonella strain SB136 (4), which is disrupted for type III secretion, was used as a control. An invC deletion mutant Salmonella (J. E. Galán and Y. Akeda) was used as a negative control strain for studies of InvC intracellular protein level and of type III secretion. Salmonella was grown in 0.3 M NaCl Luria–Bertani (LB) medium. Liquid culture incubation conditions and EGS induction with arabinose at 0.2% final concentration are as previously described (7). Following addition of arabinose for EGS induction, Salmonella liquid cultures were grown to late log phase prior to northern blot analysis, assay of Salmonella type III secretion or quantification of bacterial entry, as detailed below.
Design of external guide sequences
EGS oligonucleotides were designed to be complementary to single-stranded regions of invB and invC mRNA, followed by an additional 3'-ACCA EGS terminal sequence. This strategy allows formation of a duplex EGS–mRNA molecule recognized as a substrate by endogenous RNase P with resultant cleavage of target mRNA (9). The individual EGS oligonucleotide sequences were named according to their predicted site of target mRNA cleavage. For example, invB 108 EGS (5'-AAUGCAAAUAAAUCCacca-3') is complementary to invB mRNA nucleotides 108–122 (5'-GGAUUUAUUUGCAUU-3') and will result in RNase P cleavage of invB mRNA at nucleotide number 108. The other invB or invC EGS sequences were: invC 98 EGS (5'-GGCGUGAUUUCACAAacca-3'), invC 269 EGS (5'-ACCGCGCCUAAUACCacca-3') and invC 293 EGS (5'-ACGAUUUUCCCUGUCacca-3'). Two previously characterized EGSs which are not complementary to invB or invC were also used: synthC5 EGS 21 and synthC5 EGS 45 (2). The EGSs synthC5 EGS 21 and synthC5 EGS 45 are complementary to, and can guide the RNase P cleavage of, mRNA used for the recombinant synthesis of the C5 protein subunit of the RNase P holoenzyme of E.coli, but are not complementary (containing at least five unpaired nucleotides per EGS) to the mRNA encoding C5 in Salmonella. Herein, the EGSs synthC5 EGS 21 and synthC5 EGS 45 are referred to as synthC5 EGS 1 and 2, respectively.
Partial RNase T1 digest mapping of invB and invC mRNA
Single-stranded regions of invB and invC mRNA were identified using RNase T1 digestion (1). Two mRNAs were digested: (i) a joint in vitro transcript containing invC mRNA 3' to invB mRNA, transcribed from the plasmid pSB553 (4) DNA after digestion with BamHI; and (ii) an invC mRNA in vitro transcript alone, expressed from plasmid pIC001 (a pSB553 derivative, with invB coding sequence removed via KpnI and BspEI excision) DNA after digestion with EcoRI.
In vitro RNase P assays
Assays of mRNA cleavage in vitro by RNase P were performed as previously described (10), using the EGS sequences and the invB and invC mRNA targets detailed above. RNase P M1 RNA was folded in a buffer containing 10 mM magnesium, using a heat block to first heat the sample at 65°C for 5 min and then slowly cool the sample to room temperature. For conditions of substrate excess, reagent concentrations were: 11 fmol labeled substrate (1100 c.p.m.), 1, 5 and 10 pmol EGS, and 1 pmol of enzymatically active recombinant E.coli RNase P M1 RNA. For conditions of limited substrate, 10 fmol of labeled substrate RNA (1000 c.p.m.) and 50, 100 and 500 fmol of EGS were used. Samples were electrophoresed in 5% polyacrylamide–7 M urea gels.
Northern blots
Northern blots were performed on total RNA extracts of Salmonella, using previously published techniques (11). Each lane of a 2.5% agarose gel was loaded with 4 μg of total RNA. The UV transillumination pattern of rRNA bands after separation of each sample on an agarose gel revealed similarity in rRNA band patterns in terms of both gross quantity and quality. Northern blot probes were 5'-end-labeled DNA oligonucleotide probes. They included probes complementary to invC mRNA, 5S rRNA and each of the four invB and invC EGS oligonucleotides listed above. In each case, 8 pmol of oligonucleotide labeled with 4 pmol (30 μCi) of ATP was used per 40 ml of rapid hybridization buffer (Amersham). Signal was detected using a phosphoimager (Fuji) and quantitated using image analysis software (Fuji ImageGauge). Quantitative results are reported as the ratio, expressed as a percentage, of northern blot invC mRNA band signal intensity for the matched culture specimens of a given Salmonella strain with EGS induction versus without EGS induction.
Assay of Salmonella type III secretion
Salmonella culture supernatant proteins were prepared and analyzed as previously described (12). Western blots were probed with polyclonal antibodies against SipB and SipC and chemiluminescent signals produced using ECL Plus western blotting detection reagents from Amersham Biosciences. Signal was detected using autoradiograph film (Kodak) and quantitated using image analysis software (Fuji ImageGauge). Quantitative results are reported as the ratio, expressed as a percentage, of western blot band signal intensity for the matched culture specimens of a given Salmonella strain with EGS induction versus without EGS induction.
Matched culture specimens were employed as internal controls for western blots. Specifically, a given Salmonella transformant was grown to early log phase in a single liquid culture, then split into paired cultures which were grown simultaneously either with or without arabinose induction of EGS expression. Equal volumes of paired cultures were harvested for protein preparation at the same point of their late log phase growth, as assessed by optical density. Equal volumes of these protein preparations were loaded per well for western blot analysis.
Assay of intracellular InvC protein level
Salmonella cultures in liquid media were pelleted and resuspended in one-twentieth volume of phosphate-buffered saline (PBS)–Tris (77 mM Tris–HCl, pH 8.0). Samples were denatured by boiling with SDS–PAGE loading buffer and separated on a 9% polyacrylamide–SDS gel. Western blots were probed with a polyclonal antibody against InvC (J. E. Galán and Y. Akeda), and chemiluminescent signals produced using ECL Plus western blotting detection reagents from Amersham Biosciences. Signals were detected using autoradiograph film (Kodak). Quantitation of results, using matched culture specimens, was performed as described for type III secretion assays, above.
Quantification of bacterial entry
Entry of different Salmonella strains into Henle-407 cells in a gentamicin protection assay of bacterial entry into host tissue culture cell monolayers was performed and quantified as previously described (13).
RESULTS
Design of EGSs for invB and invC mRNA in vitro
To design the EGSs reported here, mRNA transcripts of invB and invC made in vitro were mapped using partial RNase T1 nuclease digestion to suggest EGS-accessible single-stranded mRNA regions. The first nucleotide of the start codon of each gene is labeled as nucleotide 1; single-stranded guanine residues of invB mRNA were identified via partial RNase T1 nuclease digestion at invB nucleotides 108 and 217. For invC mRNA, single-stranded guanines were identified at invC nucleotides 98, 237, 269 and 293. Given uncertainty about whether invC mRNA exists in cells independently from invB mRNA or as a joint transcript with invB, both possibilities were examined in RNase T1 digestion in vitro. RNase T1 digestions were performed on two in vitro transcripts: invC mRNA alone, as well as a tandem transcript of invC mRNA immediately 3' to invB mRNA. Single-stranded regions of invC identified in the joint invB–invC in vitro transcript were notably also found for the invC in vitro transcript alone (data not shown). EGS oligonucleotides were designed to be complementary to the RNase T1-accessible mRNA sequences invB 108–122, invC 98–112, invC 269–283 and invC 293–307, and were named for their predicted nucleotide cleavage sites by RNase P: invB 108, invC 98, invC 269 and invC 293, respectively.
RNase P-specific cleavage of mRNA in vitro
RNase P hydrolyzes the phosphodiester bond (in the target mRNAs) that precedes the first base pair in the 5' end of the target mRNA–EGS complex, akin to the site-specific cleavage reaction the enzyme catalyzes in the 5' processing of precursor tRNA (9). The reaction ingredients for RNase P assays in vitro were E.coli RNase P, internally radiolabeled invB and invC mRNA target transcribed in vitro, and EGSs complementary to portions of invB or invC mRNA. All four EGSs guide RNase P to cleave the mRNA at the predicted sites of EGS mRNA hybridization, yielding appropriately sized 5' and 3' cleavage products (Fig. 1). RNase P cleavage of mRNA increases with increasing EGS dose, with the EGSs invB 108 and invC 293 most efficient at guiding mRNA cleavage in vitro in conditions of limited substrate (data not shown).
Figure 1. RNase P–EGS cleavage of mRNA in vitro. Target substrates for cleavage include a joint transcript including both invB and invC mRNA (lane 1), and a transcript of invC mRNA alone (lane 6). The former was incubated with increasing amounts of the invB EGS 108 (lanes 2–4); the latter as a target for invC EGSs invC 98 (lanes 7–9), invC 269 (lanes 10–12) and invC 293 (lanes 13–15). Lanes labeled N (lanes 5 and 16) lack any EGS, but do have active RNase P with the invB and invC joint transcript (lane 5) or the invC transcript (lane 16) and show no non-specific target cleavage in the absence of EGS. Cleavage products were separated by size using electrophoresis in a 5% polyacrylamide–7 M urea gel. Cleavage product sizes are consistent with RNase P cleavage occurring at the predicted site at the 5' end of the mRNA region to which each EGS hybridizes. Predicted sizes of reaction products following RNase P enzymatic cleavage of in vitro transcripts are listed on the left and right of the image (e.g. invB and invC joint transcript mRNA cleavage products of 585 and 123 nucleotides for EGS invB 108; invC transcript mRNA cleavage products of 485 and 198 nucleotides for EGS invC 98, etc.).
Specific mRNA targeting in Salmonella
Induction of expression of invB or invC EGSs in Salmonella is followed by a decrease in invC mRNA compared with identical Salmonella transformants lacking EGS induction. Northern blots of equal microgram amounts of total RNA isolated from various Salmonella liquid cultures are shown in Figure 2, where northern blot signals for invC mRNA in matched Salmonella cultures decreased between 27 and 50% following relevant EGS induction. The invB or invC EGS expression effects appear specific, in that there is no similar change detected in the level of constitutive 5S rRNA after EGS induction (Fig. 2). Arabinose was used to induce Salmonella EGS expression. It did not have as pronounced an effect on invC mRNA in the absence of EGS expression plasmids, either in the case of Salmonella which was not transformed with an EGS expression plasmid, or in the case of Salmonella for which the EGS expression plasmid was presumably lost after prolonged culture (Fig. 2). Plasmid maintenance in transformants was assessed by parallel quantitative plating on LB or LB ampicillin plates as previously reported (2,11). Plasmid loss was considered to have occurred when colony counts on LB plates were greater than on LB ampicillin plates by at least an order of magnitude. As previously demonstrated for other EGS transcription in this Salmonella system (7), northern blot probes complementary to EGS oligonucleotides detected EGS expression only after arabinose induction (data not shown).
Figure 2. Northern blots in Salmonella with inducible expression of EGS molecules. RNA was isolated from Salmonella, electrophoresed in a 2.5% agarose gel, and probed for invC mRNA or for constitutively produced non-targeted 5S rRNA. The partial invC in vitro transcript expressed from plasmid pIC001 DNA after digestion with EcoRI (as described for T1 digest mapping) serves as a size marker (683 nt). The invA mutant is SB136 (also used for invasion assays). SB300 is the parent Salmonella strain, from which SB300A#1 was constructed. Other lanes are paired by SB300A#1 Salmonella transformant type, either with (+) or without (–) the addition of arabinose for the induced expression of the EGS molecules listed. Longer term induction (++) was suspected to be accompanied by the loss of EGS expression plasmid invC 293, and an accompanying lack of effect on target mRNA. Note the decrease in invC mRNA after the induction of EGS expression, whereas non-targeted constitutive 5S rRNA levels are independent of EGS expression.
The effect of expression of EGSs on type III secretion by Salmonella
A standard functional assay of InvC-dependent type III secretion was employed, in which proteins secreted by the type III secretion system were measured in cell culture supernatants (12). These secreted proteins, SipB and SipC, are also encoded in the Salmonella pathogenicity island I gene complex and are reviewed elsewhere (14,15).
A panel of Salmonella transformants, containing EGSs in various inducible expression plasmid vectors, was used for InvC-dependent type III secretion assays. The concurrent expression of two EGSs using any one of four high copy number plasmids derived from pUC19 (2) consistently decreased SipB secretion by 65% from that detected for the same Salmonella transformant grown in parallel under non-EGS-inducing conditions, based on the relative signal intensities of SipB western blot bands for a given Salmonella strain with versus without EGS induction (Fig. 3). Using the same method of quantitative analysis, induction of the same set of EGS pairs in Salmonella transformed with a low copy number plasmid derived from pWKS30 showed a 20–30% decrease in secretion (data not shown), as did any of the four EGSs when expressed alone from pUC19-derived plasmids (data not shown). These findings are consistent with prior studies of EGS dose–response features in E.coli comparing EGS expression plasmids encoding single versus multiple EGSs (2) and comparing EGS expression plasmids with strong versus weak promoters (10).
Figure 3. Western blots of cell culture supernatants from Salmonella with (+) or without (–) the induced expression of EGS molecules listed. Western blot detects the proteins SipB and SipC, secreted into cell culture supernatant by the type III secretion system and subsequently electrophoresed in a 9% polyacrylamide–SDS gel. The transformants shown contain a pUC19-derived high-copy number plasmid (2) from which two EGSs are concurrently expressed. InvBB denotes inducible expression of invB 108 and invB 108 EGSs in tandem, InvBC denotes invB 108 and invC 98 EGS expression, InvCB denotes invC 293 and invB 108 EGS expression, and invCC denotes invC 293 and invC 98 EGS expression. The control concurrently expresses two EGSs targeting the mRNA used for synthetic C5 protein over-expression in E.coli. Migration patterns of SipB and SipC bands relative to protein molecular weight 66 and 45 kDa markers are as shown in Figure 4.
To assess sequence specificity, the negative control EGSs, synthC5 EGS 2 and 1, against the synthetic C5 component of E.coli RNase P, described above and in McKinney et al. (2), was also used. Induction of this negative control EGS was accompanied by a <10% decrease in SipB secretion as compared with non-induced parallel cultures (Fig. 3). In addition, the loss of an invB or invC EGS expression plasmid from a bacterial strain was accompanied by a loss of previously observed inhibitory effects on secretion (data not shown).
The effect of EGSs on intracellular InvC protein in Salmonella
A polyclonal antibody (J. E. Galán and Y. Akeda) raised against recombinant InvC protein was used for western blot analysis of cellular InvC protein level as an assessment of EGS effects (Fig. 4, top). A band consistent with the 47 kDa InvC protein is found in the Salmonella strain SB300A#1 with no EGS. An SB300 strain (constructed by Y. Akeda) in which InvC expression has been disrupted via an invC deletion mutation serves as a negative control for InvC expression. Intracellular expression of InvC was also assessed in SB300A#1 transformed with the EGS expression plasmid for the concurrent expression of either the EGSs invB 108/invC 98 or the negative control EGSs synthC5 EGS 2 and 1 (also shown in Fig. 3). After the induction of EGS expression, the relevant protein band is not detected for SB300A#1 with the invB/invC EGSs. No effect on the detection of this protein band is seen after the induction of the negative control EGS. Figure 4 is representative of repeated assays in that the decrease in putative 47 kDa InvC signal cannot be explained by a general decrease in the intensity of other bands detected by the polyclonal antibody. Indeed, for the two cases of EGS induction shown here, invB/invC EGS induction is accompanied by decreased 47 kDa InvC signal in the context of strong signals for other bands, whereas negative control EGS induction is accompanied by strong 47 kDa InvC signal in the context of relatively weak signals for other bands.
Figure 4. Western blots for Salmonella intracellular InvC (top) and for SipB and SipC secreted into Salmonella cell culture supernatants (bottom, as for Fig. 3). SB300A#1 without an EGS expression plasmid (no EGS) and a SB300 invC deletion mutant which does not express InvC (invC) (produced by J. E. Galán and Y. Akeda) are shown to the left of protein molecular weight markers. At right: SB300A#1 transformed with the high copy number plasmids for arabinose-inducible expression of either invB 108 and invC 98 EGSs (InvB/C EGS) or the control EGSs (Control EGS) described for Figure 3. EGS induction status by arabinose addition is shown as (+) or (–). Arrowheads denote the predicted locations of InvC, SipB and SipC after electrophoresis in 9% polyacrylamide–SDS gels.
Cell supernatants from the same experiment were used for type III secretion assays (Fig. 4, bottom) as above. In each case, detection of the band representing intracellular InvC protein correlates with type III secretion. Type III secretion is decreased when intracellular InvC protein levels are decreased, either by the static invC deletion mutation, or by the dynamic disruption of InvC expression following the induction of the invB/invC EGSs. While the presumptive InvC protein band disappears in both situations, the decrease in type III secretion is more pronounced following the static deletion of the invC gene than following the induction of the invB/invC EGS. This functional assay of InvC cellular activity suggests that our invB/invC EGS system inhibits InvC-dependent type III secretion less completely than does the invC deletion-mediated ablation of InvC expression. Wild-type levels of secretion are observed for SB300A#1 with no EGS plasmid, without induction of invB/invC EGSs, or with negative control EGS expression.
EGS impact on Salmonella invasion into host cells
Salmonella strains containing EGSs in various inducible expression plasmid vectors were also tested for their ability to invade Henle-407 cells in tissue culture. In each case, a given Salmonella strain was grown in parallel liquid cultures, either with arabinose added to induce EGS induction or without the addition of arabinose. For invasion, these Salmonella strains were first incubated with Henle cells for 45 min in Hanks buffered salt solution. This was followed by a 2 h treatment with gentamicin to kill extracellular bacteria, and subsequent washes with tissue culture buffer. Invasion was quantified as the percentage of bacteria inoculated into the tissue culture wells which were recovered from lysed Henle cells.
As shown in Table 1, a control Salmonella strain, SB136, with a null mutation in invA and a previously documented functional defect in type III secretion and host cell invasion (4), achieves <1% invasion into Henle cells, independent of arabinose addition. Salmonella SB300A#1 transformants which express an EGS complementary to mRNA encoding synthetic C5 (also shown as a negative control for type III secretion assays, above) have a 20% rate of invasion, with a small decrease in invasion observed after arabinose addition. Salmonella SB300A#1 transformants which inducibly express EGSs against invB or invC from a high copy number pUC19-derived plasmid exhibit an 8–12% rate of invasion without arabinose addition, which decrease to 2.8–3.5% after arabinose induction of EGS expression. The same EGS constructs, expressed from low copy number pWKS30-derived plasmids, did not affect invasion rates after EGS induction.
Table 1. Invasion of Henle-407 cells by Salmonella strains
DISCUSSION
The pathogenicity island genes invC and invB of Salmonella provide intriguing targets for gene product disruption. In the case of invC, mutagenesis studies clearly show that the ATPase encoded by invC is required for type III secretion in assays in vitro and is important for pathogenicity in animal models (4,5). Following appropriate EGS expression, we observe a decrease in invC mRNA, InvC intracellular protein, InvC-powered type III secretion and type III secretion-dependent host cell invasion.
The inhibition of type III secretion and of Salmonella invasion using EGSs to disrupt invC mRNA is less complete than that resulting from invC deletion mutagenesis. This suggests that a certain critical level of mRNA disruption is able to partially inhibit type III secretion and host cell invasion. The level of mRNA disruption required for phenotypic changes probably varies for different target mRNAs, depending on factors such as the ratio of EGS to target mRNA (2,11), the relative efficiency of various EGSs and the functional reserve capacity a cell has for a given target mRNA and the protein that mRNA encodes. The EGSs reported here show greater phenotypic effects when expressed from high copy number, rather than low copy number, plasmids. This is consistent with prior EGS dose–response observations in bacteria. Phenotypic effects in E.coli are greater following EGS expression driven by a strong promoter as compared with a weak promoter (11), and the concomitant expression of different EGSs in E.coli results in phenotypic effects exhibiting additive synergy (2). The application of EGS technologies to regulate gene expression in bacteria in in vivo models of infection may benefit from EGS expression plasmids which can be stably maintained in bacteria within an animal. We produced the low copy number plasmids reported here in an initial effort toward this end. The fact that EGSs expressed from our low copy number plasmid system did partially inhibit type III secretion but showed no apparent effect on host cell invasion suggests a possible threshold effect, in which inhibition of type III secretion must reach a critical threshold to result in inhibition of cell invasion.
In contrast to the static effects on gene product disruption produced by mutagenesis techniques, our techniques of gene product disruption involve a more dynamic process of EGS-guided, RNase P-mediated, mRNA cleavage commencing after EGS induction. As observed in past studies of EGSs complementary in sequence to essential genes (2) or to antimicrobial resistance genes (11) in E.coli, a lag occurs between initiation of EGS expression and the observed molecular and phenotypic effects. This lag is consistent with the time required to accrue sufficient amounts of mRNA cleavage for observable phenotypic effects.
Reconciling the relative magnitude of effects we observe after EGS expression upon invC gene product expression and functions (Fig. 5) requires consideration of the kinetics of several steps of cellular physiology and our measurement methods. For example, the assays we employed showed a smaller relative effect of EGSs on type III secretion than on invasion. Our invasion assays detect host cell invasion over a short time period commencing at a time when differences between conditions of EGS or no EGS are relatively great. In contrast, our secretion assays detect the accumulated SipB or SipC secreted into supernatants throughout the period of EGS induction and include a lag time when EGS effects are minimal. Accordingly, the relative difference between conditions of EGS expression and no EGS expression which we detect for secretion might be predicted to be less than the relative difference we detect for invasion.
Figure 5. Summary of invB and invC EGSs and their effects. A schematic of the invB and invC mRNA transcripts is shown, with the predicted sites of RNase P-mediated cleavage identified for each invB or invC EGS (invB 108 EGS, invC 98 EGS, invC 269 EGS and invC 293 EGS). The last nucleotide in the final codon of invB also serves as the first nucleotide in the first codon of invC. Effects of these EGSs on Salmonella invC mRNA, type III secretion, InvC protein and host cell invasion are summarized, and relevant figures cited. Further details are provided in the text.
Effects following expression of invB EGSs were somewhat surprising given the results of prior invB studies. InvB is a type III secretion chaperone specific for SipA (6). Mutations of invB do not affect the type III secretion of SipB or SipC (6) or the invasion of host cells (4). Yet, an EGS complementary in nucleotide sequence to invB and which cleaves invB mRNA in vitro, does inhibit type III secretion of SipB and SipC and host cell invasion. Furthermore, this EGS against invB decreases invC mRNA levels in Salmonella.
The coding regions of the invB and invC genes overlap, such that the 3' nucleotide of the final codon of invB serves as the 5' nucleotide of the first codon of invC. We postulate that the effects of invB EGSs are consistent with the existence of a joint invB–invC transcript in Salmonella (Fig. 5), which becomes destabilized or functionally disrupted after cleavage of its invB portion. If this is true, the fact that an EGS targeting invB mRNA also has effects on invC suggests that, in polycistronic mRNA transcripts, EGS targeting the 5' portions of mRNAs may also have effects on mRNAs encoded by genes 3' to the EGS target site.
Ultimately, pathogenicity island gene product disruption may lead to novel anti-pathogenicity agents and strategies. Since these genes are not essential to bacterial viability, interfering with their functions could conceivably decrease morbidity without the relatively broad-spectrum effects and selective pressure for resistance seen with many current antimicrobial agents.
ACKNOWLEDGEMENTS
We thank Yukihiro Akeda for antibody against InvC and for the invC Salmonella mutant. This research was conducted while J.M. was a Pfizer postdoctoral fellow. The work was supported by an NIH research training program (NIH T32 AI07210-17), a Yale Children’s Health Center Scholar Award and a Washington University Digestive Diseases Research Core Center Grant (DDRCC P30 DK52574) to J.M., and NIH grants AI030492 to J.E.G. and GM19422 to S.A.
REFERENCES
Guerrier-Takada,C. and Altman,S. (2000) Inactivation of gene expression using ribonuclease P and external guide sequences. Methods Enzymol., 313, 442–456.
McKinney,J., Guerrier-Takada,C., Wesolowski,D. and Altman,S. (2001) Inhibition of Escherichia coli viability by external guide sequences complementary to two essential genes. Proc. Natl Acad. Sci. USA, 98, 6605–6610.
Groisman,E.A. and Ochman,H. (1996) Pathogenicity islands: bacterial evolution in quantum leaps. Cell, 87, 791–794.
Eichelberg,K., Ginocchio,C.C. and Galán,J.E. (1994) Molecular and functional characterization of the Salmonella typhimurium invasion genes invB and invC: homology of InvC to the F0F1 ATPase family of proteins. J. Bacteriol., 176, 4501–4510.
Porter,S.B. and Curtiss,R.R. (1997) Effect of inv mutations on Salmonella virulence and colonization in 1-day-old White Leghorn chicks. Avian Dis., 41, 45–57.
Bronstein,P.A., Miao,E.A. and Miller,S.I. (2000) InvB is a type III secretion chaperone specific for SspA. J. Bacteriol., 182, 6638–6644.
McKinney,J., Guerrier-Takada,C., Galán,J. and Altman,S. (2002) Tightly regulated gene expression system in Salmonella enterica serovar Typhimurium. J. Bacteriol., 184, 6056–6059.
Wang,R.F. and Kushner,S.R. (1991) Construction of versatile low-copy number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene, 100, 195–199.
Forster,A.C. and Altman,S. (1990) External guide sequences for an RNA enzyme. Science, 249, 783–786.
Guerrier-Takada,C., Li,Y. and Altman,S. (1995) Artificial regulation of gene expression in Escherichia coli by RNase P. Proc. Natl Acad. Sci. USA, 92, 11115–11119.
Guerrier-Takada,C., Salavati,R. and Altman,S. (1997) Phenotypic conversion of drug-resistant bacteria to drug sensitivity. Proc. Natl Acad. Sci. USA, 94, 8468–8472.
Russmann,H., Kubori,T., Sauer,J. and Galán,J.E. (2002) Molecular and functional analysis of the type III secretion signal of the Salmonella enterica InvJ protein. Mol. Microbiol., 46, 769–779.
Galán,J.E. and Curtiss,R.R. (1989) Cloning and molecular characterization of genes whose products allow Salmonella typhimurium to penetrate tissue culture cells. Proc. Natl Acad. Sci. USA, 86, 6383–6387.
Collazo,C.M. and Galán,J.E. (1997) The invasion-associated type III system of Salmonella typhimurium directs the translocation of Sip proteins into the host cell. Mol. Microbiol., 24, 747–756.
Komoriya,K., Shibano,N., Higano,T., Azuma,N., Yamaguchi,S. and Aizawa,S.I. (1999) Flagellar proteins and type III-exported virulence factors are the predominant proteins secreted into the culture media of Salmonella typhimurium. Mol. Microbiol., 34, 767–779.(Jeffrey S. McKinney*,1,2, Haifeng Zhang1)