|Home | About | Journals | Submit | Contact Us | Français|
Yersinia pestis, the causative agent of plague, utilizes a plasmid-encoded type III secretion system (T3SS) to aid it with its resistance to host defenses. This system injects a set of effector proteins known as Yops (Yersinia outer proteins) into the cytosol of host cells that come into contact with the bacteria. T3SS is absolutely required for the virulence of Y. pestis, making it a potential target for new therapeutics. Using a novel and simple high-throughput screening method, we examined a diverse collection of chemical libraries for small molecules that inhibit type III secretion in Y. pestis. The primary screening of 70,966 compounds and mixtures yielded 421 presumptive inhibitors. We selected eight of these for further analysis in secondary assays. Four of the eight compounds effectively inhibited Yop secretion at micromolar concentrations. Interestingly, we observed differential inhibition among Yop species with some compounds. The compounds did not inhibit bacterial growth at the concentrations used in the inhibition assays. Three compounds protected HeLa cells from type III secretion-dependent cytotoxicity. Of the eight compounds examined in secondary assays, four show good promise as leads for structure-activity relationship studies. They are a diverse group, with each having a chemical scaffold not only distinct from each other but also distinct from previously described candidate type III secretion inhibitors.
Yersinia pestis, a gram-negative bacterium in the family Enterobacteriaceae, is the causative agent of plague (10). A hallmark of this devastating disease is the rapid development of systemic infection following the deposition of a small inoculum into the skin via the bite of an infected flea (9). This fulminant course is dependent on the ability of the bacteria to overwhelm host defenses, a feat which they accomplish by a combination of stealth and active suppression targeting innate immune mechanisms. The type III secretion (T3S) system (T3SS) of Y. pestis is crucial to the suppression arm of this strategy. This system injects a set of effector proteins directly into the cytoplasm host cells in contact with the bacteria. These act to derange key defensive functions, including phagocytosis and cytokine production (11, 12).
Although several other adaptations also play key roles in virulence, the loss of T3S is sufficient to render Y. pestis completely avirulent, even when the bacteria are introduced directly into the bloodstream (39). T3SS is shared by Y. pseudotuberculosis and Y. enterocolitica, which are also pathogenic for humans, and is encoded by a 70-kb plasmid, termed pCD1 in most Y. pestis strains (3, 11, 17). The major injected Yersinia outer proteins (Yops), usually designated YopH, YopE, YopM, YopJ, YopO, and YopT, target the cytoskeleton, proinflammatory signaling pathways, and probably other cell functions as well. T3SS involves at least 39 other proteins that play supporting roles in regulation, secretion, and translocation. Some notable members of this cast include the secreted translocon proteins YopB and YopD, which collaborate to form a pore in the host cell membrane; YscF, which forms the shaft of a needle-like channel through which the secreted proteins are thought to travel (16, 34); LcrV, which sits at the tip of this needle and which may serve as the interface with the YopB-YopD pore (26); and a set of chaperones (SycD, SycE, SycH, SycT, and SycN) that assist with the translocation of specific Yops (11, 36, 38, 40). The energy driving the secretion process is provided by the ATPase YscN (4).
T3SSs are also present in many other gram-negative pathogens. Examples include Salmonella spp., Shigella spp., Pseudomonas aeruginosa, enteropathogenic Escherichia coli (EPEC), enterohemorrhagic E. coli, and Chlamydia spp. Many details regarding their workings remain to be determined, but there is no doubt that they are both common and play important roles in bacterial virulence (11, 18). For these reasons, they constitute a potential target for novel therapeutics. Several attempts have been to identify small-molecule T3SS inhibitors, with some success (1, 19, 22, 23).
The regulation of the T3SS function in the yersinae is interesting and unusual, offering distinct advantages for in vitro manipulation and the development of screening techniques. At temperatures below about 34°C, the expression of many of the proteins required for assembly of the functional T3SS apparatus, known as the injectisome, is repressed. Above this temperature the apparatus is formed, but under physiological conditions it secretes little or no protein unless contact is made with cell membranes. Such contact both triggers secretion by an as yet undetermined mechanism and results in the increased expression of many Yop genes (5, 32, 33, 37).
In vitro, high levels of secretion and Yop gene expression can be induced by a lack of Ca2+ ions in the medium during incubation at 37°C. However, when secretion is induced in this way, the bacteria are unable to grow (25). The increased transcription of yopE under these conditions was the basis of the technique used by Kauppi et al. to screen a library of 9,400 compounds for their ability to inhibit T3SS in Y. pseudotuberculosis (22). We took advantage of the inhibition of growth that accompanies the secretion induced by low levels of calcium to devise an alternative screening technique. Here, we describe this method and report on the initial characterization of a novel set of inhibitors of Yersinia T3S, identified through the screening of 70,966 compounds and mixtures from 13 small-molecule and extract libraries.
The Y. pestis strains used in this study were derivatives of Y. pestis strain KIM. The strains and the plasmids used are listed in Table Table1.1. Plasmid pMM85 (YopE-Bla) (24) was the gift of Olaf Schneewind (University of Chicago).
Y. pestis and Y. pseudotuberculosis were cultured in TB broth (1% [wt/vol] Bacto tryptose [BD Bioscience], 0.5% [wt/vol] NaCl, 0.3% [wt/vol] beef extract paste [BD Bioscience]) and brain heart infusion (BHI; BD Bioscience) broth, as indicated. For high-throughput screening, Y. pestis was grown in TB broth for 20 h at 26°C and diluted to 1 × 104 bacteria/ml, the compounds to be screened were added, and the bacterial cultures were then shifted to 37°C and incubated for 18 h. To monitor the cultures for the secretion of Yops, Yersinia strains were grown in BHI broth at 30°C overnight, diluted to an optical density at 600 nm (OD600) of 0.04 in fresh BHI broth, and incubated at 30°C to an OD600 of 0.2. The culture was then shifted to 37°C, and incubation was continued for an additional 2 h. EPEC strains were cultured in LB medium at 37°C, unless indicated otherwise.
The compounds and extract libraries used in the screening were provided by the NIH-supported National Screening Laboratory at Harvard Medical School. Larger quantities of the test compounds were obtained from commercial venders. Compounds 1, 2, and 3 were purchased from TimTec Inc.; compound 3 diproprionate was purchased from Sigma-Aldrich; and compound 4 and INP0007 were purchased from ChemBridge Corporation. Mouse monoclonal anti-YopM antibody (1A1), rabbit polyclonal anti-YopD antibody (5331), and rabbit polyclonal anti-YopH antibody (G5788) were gifts from Susan Straley (University of Kentucky). A rabbit polyclonal anti-YopE antibody was the gift of Olaf Schneewind (University of Chicago). Mouse anti-hemagglutinin (anti-HA) monoclonal antibody HA.11 was purchased from Covance, and rabbit anti-OmpA antibody was kindly provided by Carol Kumamoto (University of Massachusetts Medical School).
Random mutagenesis of pCD1 with the transposon magellen3 (35), which confers kanamycin resistance, was used to create pCD1 (Y0093::magellen3). Mutagenesis was performed with Y. pestis by using the suicide delivery plasmid pFD1 (35) in donor strain β2155 (14). For brevity, plasmid pCD1 (Y0093::magellen3) is hereafter referred to as pCD1K. Plasmid pML001 carries the lux operon from Photorhabdus luminescens. Both pCD1K and pML001 were introduced into Y. pestis JG154 by electroporation to generate JG401. JG406 is a spontaneous segregant of JG401 lacking pCD1K and was isolated by selection for growth at 37°C (31). The luminescence of the bacterial cultures was measured with a Packard Picolite luminometer (United Technologies) and an EnVision plate reader (Perkin-Elmer).
High-throughput screening (HTS) was conducted with luminescent Y. pestis strain JG401. Thirty microliters of JG401 culture (1 × 104 bacteria/ml) was added into each well of 384-well cell culture plates by a liquid-handling robot (a μFill plate dispenser with a Bio-Stack apparatus, both of which were from Bio-Tek Instruments). One hundred nanoliters of screening compound (5 mg/ml in dimethyl sulfoxide [DMSO]) was applied to each well of the assay plates by an Epson compound transfer robot with Epson standard-volume pin arrays. The plates were incubated overnight at 37°C, and the relative luminescence was measured by a plate-reading luminometer equipped with a 40-plate automatic loader (EnVision plate reader; Perkin-Elmer). The dwell time per well was set at 0.1 s. Statistical analysis of the primary screening results was carried out as described previously (31).
Overnight cultures of strains JG153 and JG154 were grown as indicated above in TB broth at 30°C and were used to inoculate BHI broth to an OD600 of 0.04. The cultures were then grown at 30°C in glass tubes (18 by 150 mm) and aerated by rotation on a Rollordrum apparatus (New Brunswick). When the OD600 reached 0.2, the compounds under test were dissolved in DMSO and were then added at the indicated concentrations. The final concentration of DMSO in all cultures except those with compound 3 diproprionate, including the controls, was 0.3%; in experiments with compound 3 diproprionate, no DMSO was used. The cultures were then shifted to 37°C to induce T3S and were incubated for an additional 2 to 3 h. The culture supernatants were collected following centrifugation (3,200 × g, 10 min). Proteins were precipitated from 2 ml of supernatant with 10% (final concentration) trichloroacetic acid (TCA), washed twice with 1 ml of 1:1 (vol/vol) ethanol-ether, and dissolved in 30 μl water. Ten microliters of this solution, which was mixed with sample buffer, was used to load a single lane for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequent preparation of Western blots for YopM, YopH, and YopD. Development of the Western blots was stopped before the positive Yop controls reached saturation, and the intensity and area of each band were quantified by video densitometry.
Mid-log-phase cultures were diluted to an OD600 of 0.05 in TB broth containing 2.5 mM CaCl2. Two hundred microliters of the culture was added to each well of 96-well tissue culture plates (Falcon Microtest, catalog number 353072; Becton Dickinson). A total of 0.6 μl of 20 mM test compounds was added to the wells (final concentration, 60 μM). The 20 mM stocks of compounds 1 to 4 were prepared in DMSO, and the compound 3 diproprionate stocks were prepared in water. The cultures were incubated in a Spectra Max 250 plate reader (Molecular Devices) at 37°C with continuous agitation (the automix function was activated). The OD600 of the 200-μl culture volume was measured at 6-min intervals for 24 h, and the values were corrected to the standard OD600 (with a cuvette with a 1-cm path length) by using equation 1. This equation was derived by measuring the OD600 values of 10 samples of various densities in both a standard cuvette with a 1-cm path length and the 96-well plate. The growth rates were calculated from the exponential portion of the resulting growth curves.
where R2 is equal to 0.999 for 1.6 ≥ OD600 ≥ 0.06.
HeLa cells were maintained at 37°C with 5% CO2 in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 U ml−1 penicillin, and 100 μg ml−1 streptomycin. A total of 1 × 105 HeLa cells/well were seeded in a 24-well tissue culture plate in DMEM with 3% heat-inactivated FBS 1 day prior to the assay. The test compounds were applied, and the cells were incubated at 37°C for 3 h. The level of disruption of the cell monolayer was recorded as the percentage of cell detachment from the culture plate. The amount of lactate dehydrogenase (LDH) released from the HeLa cell cultures was measured with a cytotoxicity detection kit (Roche), according to the manufacturer's protocol.
HeLa cell monolayers grown in DMEM supplemented with 3% heat-inactivated FBS and 24 μM HEPES (pH 7.4) were infected with Y. pestis, and the test compounds were immediately added. The infected cultures were then incubated at 37°C with 5% CO2. HeLa cells were infected with pCD1-positive Y. pestis strains, as indicated above. The cytotoxic response was initially observed as the rounding up of the cells. The level of inhibition of this response was obtained by determining the percentage of the HeLa cells that showed morphological changes among all treated cells. Strain JG154 was used as a T3S-negative control.
HeLa cells were maintained as described above. One day prior to infection, 3.5 × 104 cells were seeded in each well of 24-well plates. Two hours before infection, the cells were washed with phosphate-buffered saline and 350 μl of DMEM with 24 μM HEPES was added to each well. HeLa cells were infected with mid-log-phase YopE::Bla-expressing Y. pestis strains [either JG153(pMM85) or JG154(pMM85), as indicated] at a multiplicity of infection of 10 for 3 h. The β-lactamase substrate CCF2-AM (Invitrogen) was added to live cells at the concentration indicated by the manufacturer. The cells were incubated at room temperature for 1 to 2 h and were then analyzed by fluorescence microscopy with a Nikon Eclipse E800 microscope with a β-lactamase filter set (Chroma).
Assays for determination of EPEC T3S were performed as described previously (15). Briefly, KC14(pKC17) (8) or KC21(pKC17) (8, 27) cultures were grown in LB medium at 37°C for approximately 8 h, diluted 1:500 into prewarmed DMEM-HEPES, and grown in a 5% CO2 atmosphere overnight. Overnight cultures were further diluted 1:50 into 5 ml of prewarmed DMEM-HEPES and grown in a 5% CO2 atmosphere at 37°C for 2 h. Test compound dissolved in DMSO or DMSO alone was added to the cultures, and incubation was continued for an additional 8 h. To separate the bacteria from the secreted proteins, the cultures were centrifuged at 3,300 × g for 10 min and the supernatant was passed through a 0.45-μm-pore size low-protein-binding filter. The bacterial pellet was resuspended in 50 μl of loading buffer, boiled for 10 min, and subjected to SDS-PAGE; and immunoblots were prepared. The secreted proteins were precipitated with 10% cold TCA overnight at 4°C, and immunoblots were prepared as described above.
As noted above, our screening technique is based on peculiarities of the physiology of Y. pestis that are not well understood. The combination of low concentrations of calcium ions (<1 mM) and incubation at temperatures above 34°C strongly induces the expression of T3SS and results in the secretion of the Yop effector proteins directly into the surrounding medium (12, 25, 42). Note that this is distinct from the targeted secretion accomplished by this system during cell contact, in which effector proteins are directed into the cells at physiological Ca2+ concentrations (~2.5 mM). The low-calcium-concentration-dependent induction of T3S is also accompanied by the inhibition of bacterial growth. In certain media with very low sodium ion concentrations, this growth inhibition is weak or absent (6, 7), suggesting that the leakage of Na+ into the cells through the active section apparatus may be the key factor in growth inhibition. Whatever its true cause, we reasoned that at least some inhibitors of T3S would prevent this growth inhibition and could be detected on this basis.
To implement the screening, we constructed a specialized Y. pestis strain. This strain was rendered avirulent both by a large spontaneous chromosomal deletion that removed the genes needed for iron acquisition during infection and by the elimination of native plasmid pPCP1, which encodes an outer membrane protease (Pla) also required for full virulence. This combination of lesions has been shown to result in avirulence in a primate model of pneumonic plague (41). This strain was also made luminescent through addition of the lux operon from Photobacterium luminescens, carried by plasmid pML001. T3SS in Y. pestis is encoded by a 70-kb plasmid designated pCD1. Either segregation of this plasmid or internal deletions which compromise the T3SS function, both of which occur at appreciable frequencies, would give rise to false-positive results in a T3S inhibition assay. To circumvent this problem, we modified pCD1 by insertion of a version of a transposon mariner carrying a kanamycin resistance determinant, magellen3 (35). In addition to providing a means for imposing selection against segregation, we reasoned that the insertion of magellen3 at some locations within pCD1 might reduce the rate of spontaneous deletions affecting T3SS expression, either through tight linkage with the sequences most subject to spontaneous deletion or by some other means. Accordingly, we screened random magellen3 insertion mutants of pCD1 for their frequency of colony formation on medium with a low calcium concentration supplemented with kanamycin at 37°C: only bacteria with defects in T3S can form colonies on this medium. One mutant that showed markedly reduced plating efficiencies (<10−8) had an insertion in the putative helicase gene, Y0093. Due to its high degree of stability, pCD1(Y0093::magellen3) (i.e., pCD1K) was used in the screening strain, and that strain was designated JG401.
The HTS assay, illustrated in Fig. Fig.1A,1A, is extremely simple to perform. Y. pestis strain JG401 suspended in TB broth containing 1 mM ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid to sequester low levels of intrinsic Ca2+ was distributed into 384-well plates, followed by pin transfer of the compounds to be screened. After overnight incubation at 37°C, the luminescence of each well was measured. Elevated luminescence implies growth and, hence, the presumptive inhibition of T3S.
We screened 10 commercially available small-molecule libraries and three plant extract libraries, listed in Table S1 in the supplemental material. They contained 69,093 individual compounds and 1,873 plant extracts. Measurements of luminescence were interpreted relative to the luminescence of negative controls, to which no compound was added, and positive controls supplemented with CaCl2 (final concentration, 2.5 mM) rather than a test compound.
We performed statistical analysis as described previously (31). We chose a twofold elevation of luminescence relative to that for the controls receiving no test compound as the threshold for the indication of presumptive inhibition. The effects of the compounds on luminescence are summarized as a distribution in Fig. Fig.1B.1B. A total of 421 compounds and extracts were identified as primary HTS hits. On the basis of the intensity of the inhibition, the novelty of the chemical structures, commercial availability, and the correlation of our results with the results of other screens for unrelated activities performed at the National Screening Laboratory, we selected eight strong and moderate hits from the primary hits for further detailed study. Below, we report on the results of follow-up studies with these eight compounds which gave moderate or strong signals in the preliminary assay.
In the primary screening, we observed that the selected compounds were able to promote Y. pestis growth and/or luminescence at 37°C in the absence of Ca2+. To determine the validity of our assumption that some compounds with this effect would be bona fide T3S inhibitors, direct assays of T3S were performed. Three T3S substrates, YopH, YopM, and YopD, were chosen as secretion indicators to evaluate the compounds. Both YopH and YopM are delivered to the cytoplasm of target cells and play important roles in virulence. YopD is delivered to the cytoplasmic membrane of host cells, where, in combination with YopB, it forms pore-like structures that are required for the translocation of other proteins across the membrane (11, 12). The effects of the eight selected compounds on YopM, YopH, and YopD secretion were determined in assays without target cells, in which secretion was induced by shifting bacterial cultures in medium with a low concentration of Ca2+ from 26°C to 37°C (for details, see Materials and Methods). To detect the secreted proteins, culture supernatants were subjected to SDS-PAGE, followed by Western blotting with antibodies specific for each of the Yops. Four of the compounds, designated compounds 1 to 4, showed a concentration-dependent inhibition of secretion. The other four compounds did not inhibit secretion (data not shown). The structures of compounds 1 to 4 are shown in Fig. Fig.2,2, and their effects in secretion assays are shown in Fig. Fig.3.3. Interestingly, differential inhibition of secretion among the Yop species was observed with compounds 1 and 3. Compound 1 inhibited YopH and YopM secretion with a 50% inhibitory concentration (IC50) of approximately 10 μM, but for YopD secretion, it was clear that inhibition was very weak (Fig. (Fig.3A).3A). Compound 3 inhibited YopH secretion with an IC50 of 15 μM; at about 15 μM, its level of inhibition of YopD approached 50%, but the level of inhibition decreased at higher concentrations. Its inhibition of YopM secretion was barely observable (Fig. (Fig.3C).3C). The IC50 of compound 2 was about 10 μM for all three Yops.
The primary screening assay is dependent on increased luminescence, which is used as a surrogate for increased growth. One advantage of this approach is the ability to reject compounds that act indirectly on the inhibition of secretion via their toxic effects that result in growth inhibition. In order to confirm that the selected compounds do not negatively affect the growth of Y. pestis, we monitored the OD600s of the bacterial cultures in the presence of the compounds (60 μM) under conditions in which T3S is not induced. None of the four compounds showed an inhibitory effect on the growth of Y. pestis at the concentrations tested (Fig. (Fig.4A).4A). To our surprise, the previously identified T3S inhibitor INP0007 (compound 1 in reference 30) completely inhibited growth at 60 μM (Fig. (Fig.4A4A).
The toxicities of compounds 1 to 4 for HeLa cells were assessed in two ways: observing the effects on cell morphology (rounding, detachment) and obtaining quantitative measurements of LDH release. Compound 2 caused cell rounding and detachment, while the other compounds did not when they were used within the range of concentrations used in the T3S inhibition assays (data not shown). Consistent with this observation, compound 2 also caused concentration-dependent LDH release, while compounds 1, 3, and 4 also had no effect in this assay (Fig. (Fig.4B4B).
One result of the delivery of effector proteins by the Y. pestis system to cells in culture is a cytotoxic reaction characterized by cytoskeletal dysfunction, resulting in cell rounding and detachment (11). To determine if the compounds could inhibit the delivery of effector proteins to HeLa cells, we examined their ability to block this effect. Because compound 2 directly caused similar effects (see above), only compounds 1, 3, and 4 were tested in this assay. All three compounds protected infected HeLa cells from T3S-induced morphological changes. The protection afforded by all three compounds was concentration dependent (Fig. (Fig.5B).5B). Although we did not measure the rates of growth of the Y. pestis organisms in the medium supplemented with the inhibitors used in these experiments, this medium supplemented with the inhibitors at 60 μM supported the substantial growth of Y. pestis in overnight cultures.
The T3S-dependent morphological change induced in mammalian cells is primarily caused by the effector protein YopE, a GTPase-activating protein affecting actin polymerization (11). Thus, inhibition of these morphological changes implies the inhibition of YopE translocation or activity. A more direct assay for T3S effector translocation based on the use of effector::β-lactamase fusion proteins and the cell-permeant florescent β-lactamase substrate CCF2-AM has been devised (24). Using this method, we tested the abilities of compounds 1, 3, and 4 to block the translocation of a YopE::Bla fusion from Y. pestis to HeLa cells. At 80 μM, compound 3 (diproprionate salt) completely inhibited translocation in this assay, and at 40 μM, it yielded an estimated 50% inhibition (Fig. (Fig.6).6). Compounds 1 and 4 showed no obvious effect in this assay. The latter result is somewhat surprising, as both of these compounds inhibited T3S-mediated cytotoxicity (Fig. (Fig.5B),5B), a phenomenon dependent on Yop translocation. One possible explanation for this result is that the cytotoxicty reaction requires higher levels of Yop translocation to yield a positive result than does the YopE::Bla-based fluorescent substrate assay, at least under the conditions used in our experiments. It is also possible that the YopE::Bla fusion protein is less sensitive to inhibition than native Yops.
To determine if the inhibitory effects of compounds 1 to 4 might extend to T3SSs other than the T3SS of the yersiniae, we tested their ability to block T3S in EPEC. Unlike the Y. pestis system, the T3SS of EPEC secretes effectors into the medium during growth at 37°C without special treatment. In strain KC14 (8), in the presence of compounds 1 and 3, the level of secretion of the EPEC effector Tir was significantly reduced (Fig. (Fig.7).7). Compounds 2 and 4 were tested at concentrations up to 60 μM (data not shown), but no inhibition was observed. KC21 (27) was used as a T3S-negative control strain.
The HTS assay based on relieving the growth inhibition associated with the in vitro activation of T3S secretion in Y. pestis described above has proved to be both effective and convenient. The rate of false-positive results was low, in the sense that the screening of 70,966 compounds yield only 431 positive hits (0.6%), among which strong positive hits were observed at a rate of 0.01%. Z′ values (43), a statistical measure of assay quality, between 0.5 and 1 are generally regarded as excellent for HTS methods. In our HTS assay, the Z′ value for random assay plates averaged 0.7 (31). The assay clearly yields bona fide inhibitors. Addition of the test compounds to the bacterial suspension in the appropriate medium is the only preincubation manipulation required, and luminometry is the only postincubation step. The use of a luminescent Y. pestis derivative allows a reliable means of detection of modest growth at low bacterial densities and avoids interference by common artifacts (e.g., bubbles) that afflict direct measurements of growth by determination of the OD. This method is strongly biased against compounds that inhibit bacterial growth, an effect that would yield false-positive results by some other approaches. It also selects for compounds that act on the bacteria rather than on target cells, as the latter are not present in the assay. The chief disadvantage of this approach is that it may fail to detect some classes of inhibitors because they do not relieve growth inhibition. For example, compounds that interfere at the level of formation of the connection of the secretion apparatus with target cells would probably not be detected.
Several other techniques have been used to screen for T3S inhibitors. These include enzyme-linked immunosorbent assay-based detection of proteins secreted from EPEC (19), inhibition of sheep erythrocyte lysis by EPEC (1), and inhibition of induction of a yopE::luxAB fusion in Y. pseudotuberculosis. The erythrocyte lysis assay probably depends on pore formation by T3S translocon proteins (EspB and EpsD in EPEC, which are paralogues of Yersinia YopB and YopD, respectively), while the yopE::luxAB fusion assay depends on the observation that the induction of T3S in the yersiniae is accompanied by substantial increases in the levels of transcription of effector genes, including yopE (22). Each of these approaches has yielded some potential inhibitors. The yopE::luxAB fusion method has been used most extensively, because, like our method, it is well suited to HTS. Initial experiments by the use of this method yielded 4 lead compounds from the screening of 9,400 compounds from the ChemBridge DiverSet F library. Initial structure-activity relationship studies (13, 21, 30) have been performed with two of these compounds, a salicylaldehyde compound (previously designated INP0007 ) and a 2-arylsulfonylamino-benzanilide, by using activity assays based on whole bacteria, because the targets of these compounds are not established. The activities of INP0007 and selected analogs against both Salmonella enterica serovar Typhimurium have been tested in vitro and in a tissue culture model (20, 29), and the activities of these compounds against Chlamydia trachomatis have been tested in tissue culture (2, 28). The results of these studies are consistent with the inhibition of T3S in these species.
The inhibitory compounds identified in our assay are a diverse group and in all cases are distinct from those discussed above. One interesting feature of the inhibition data is the differential action of the inhibitors on the secretion of different Yop species. Only compound 2 acted effectively against all three Yop species examined, while each of the other compounds inhibited YopH secretion more effectively than they inhibited the secretion of YopD and YopM. Moreover, the relative activities against the secretion of YopD and YopM also varied among the compounds. These differences suggest that the compounds differ in their mechanisms of action. The abilities of compounds 1 and 3, but not those of compounds 2 and 4, to inhibit the secretion of Tir by EPEC provides additional evidence that the mechanisms of action are diverse. Like INP0007 and related compounds, the novel scaffolds of compounds 1 and 3 clearly indicate that they have spectra of activity that extend beyond the yersiniae. This confirms the utility of the Yersinia system as a general platform for identifying T3S inhibitors.
Compound 2 was remarkable in comparisons with the other T3S inhibitors described here and elsewhere for its low IC50 and its ability to cause very complete inhibition. While the toxicity of this compound for mammalian cells is unfortunate, studies of its mechanism of action against T3SSs may prove useful in the development of improved inhibitors. It is also possible that modifications of this compound may sufficiently reduce its toxicity to provide a useful therapeutic index.
One puzzling observation is that INP0007 inhibited the growth of both Y. pestis and Y. pseudotuberculosis, with stronger inhibition against the former being observed. This conflicts with the results of Kauppi et al. (22). We cannot explain this inconsistency and can only suggest that it may be due to differences in the growth medium used or may be due to the presence of a minor but potent growth inhibitor in our commercial preparation. Minor contaminants were observed by mass spectroscopy (see Fig. S1 in the supplemental material), but the effect of this compound on bacterial growth is an issue that should be revisited.
We thank Susan Straley for the kind gift of anti-Yop antibodies; Olaf Schneewind for plasmid pMM85; Su Chiang and Gerald Beltz for facilitating our interaction with the national screening facility; and Chrono Lee, Karen Ginglas, and Christina Anderson for technical assistance.
This study was supported by a grant to J.D.G. from the New England Regional Center of Excellence for Biodefense and Emerging Infectious Diseases Research (grant NERCE BEID; U54 AI1057159 from the National Institutes of Health) and by NIH grant R01 AI46454 to J.M.L.
Published ahead of print on 17 November 2008.
†Supplemental material for this article may be found at http://aac.asm.org/.