Piericidin A1 Blocks Yersinia Ysc Type III Secretion System Needle Assembly

The bacterial type III secretion system (T3SS) is widely used by both human and animal pathogens to cause disease yet remains incompletely understood. Deciphering how some natural products, such as the microbial metabolite piericidin, inhibit type III secretion can provide important insight into how the T3SS functions or is regulated. Taking this approach, we investigated the ability of piericidin to block T3SS function in several human pathogens. Surprisingly, piericidin selectively inhibited the Ysc family T3SS in enteropathogenic Yersinia but did not affect the function of a different T3SS within the same species. Furthermore, piericidin specifically blocked the formation of T3SS needles on the bacterial surface without altering the localization of several other T3SS components or regulation of T3SS gene expression. These data show that piericidin targets a mechanism important for needle assembly that is unique to the Yersinia Ysc T3SS.

IMPORTANCE The bacterial type III secretion system (T3SS) is widely used by both human and animal pathogens to cause disease yet remains incompletely understood. Deciphering how some natural products, such as the microbial metabolite piericidin, inhibit type III secretion can provide important insight into how the T3SS functions or is regulated. Taking this approach, we investigated the ability of piericidin to block T3SS function in several human pathogens. Surprisingly, piericidin selectively inhibited the Ysc family T3SS in enteropathogenic Yersinia but did not affect the function of a different T3SS within the same species. Furthermore, piericidin specifically blocked the formation of T3SS needles on the bacterial surface without altering the localization of several other T3SS components or regulation of T3SS gene expression. These data show that piericidin targets a mechanism important for needle assembly that is unique to the Yersinia Ysc T3SS. KEYWORDS T3SS, Yersinia, piericidin, type III secretion system, virulence blocker P iericidins are a family of natural products produced by Actinobacteria that resemble coenzyme Q, also known as ubiquinone, a molecule heavily involved in electron transport in prokaryotes and eukaryotes alike (1). The piericidin family member pieri-cidin A1 was originally shown to be a potent inhibitor of NADH oxidase in purified beef heart mitochondria (2). Subsequently, analysis of piericidin A1-resistant mutants of the bacterium Rhodobacter capsulatus suggested the molecular target to be the 49-kDa subunit of NADH-ubiquinone oxidoreductase, referred to as complex I (3). Complex I is a large protein complex containing at least 40 subunits with a mass of~1 MDa and is a central component of the electron transport chain in eukaryotes and prokaryotes (4). This complex functions to translocate protons and is integral in generating a proton gradient for ATP synthesis and other processes such as powering protein secretion. Piericidin A1 acts by blocking the reduction of coenzyme Q by complex I (3,5). Many bacteria are insensitive to piericidin, likely because of differences in complex I subunit composition and/or substrate and cofactor binding interfaces (6) that possibly prevent piericidins from binding. Piericidin A1 has been explored as a potential anticancer agent, as treatment with nanomolar concentrations of the compound prevented upregulation of the glucose receptor GRP78, resulting in rapid cell death in glucosestarved HT-29 cells (7), a human cultured colon cancer cell line. More recently, Kang et al. described piericidin A1 as an inhibitor of quorum sensing in Erwinia carotovora, a potato plant pathogen, although the mechanism underlying this remains unclear (8).
Recently, we identified piericidin A1 and the piericidin derivative Mer-A-2026B as inhibitors of the bacterial type III secretion system (T3SS) (9). The T3SS is a complex nanomachine composed of more than 20 different structural proteins that assemble into a needle-like injection apparatus that punctures eukaryotic cells and acts as a conduit through which bacterial effector proteins enter the host cell cytosol (10). Protein secretion through this apparatus requires a proton motive force (11), as well as the ATPase YscN (12), and the apparatus is assembled in three stages. The basal body, assembled first, is composed of rings that are embedded in the inner and outer membrane of the bacterium and bridge the gap of the periplasmic space. Upon completion of the basal body, secretion of the "early" T3SS substrates begins. The SctF needle subunit, referred to as YscF in Yersinia, is secreted at this stage and polymerizes from the basal body to form a hollow needle protruding~60 nm into the extracellular space (13). Upon completion of needle assembly, a substrate specificity switch leads to secretion of the "middle" substrates, which comprise the pore or translocon complex (10). These three additional proteins, in Yersinia, called YopD, YopB, and LcrV, are secreted and form the translocon complex at the tip of the newly polymerized YscF needle (14). YopK (YopQ in Yersinia enterocolitica) has been shown to regulate the translocation of the pore-forming translocon proteins YopB and YopD, as well as effector Yops (15). In addition to its role as a translocon protein, YopD is thought to form a complex with LcrQ and LcrH to bind Yop mRNA and the 30S ribosomal subunit, inhibiting translation until the T3SS is actively secreting cargo, including LcrQ and YopD (16,17). The final substrate switch, to the secretion of "late" cargo, T3SS effector proteins called Yops in Yersinia and regulatory proteins, is triggered in vivo by host cell contact, which may be sensed by the translocon (18). In Yersinia, this response can be mimicked through the chelation of calcium (19,20). Therefore, secretion of late Yops is prevented in the presence of high calcium concentrations and permitted at low calcium concentrations. In Yersinia, the regulatory proteins YopN and TyeA prevent the secretion of Yops prior to host cell contact or a decrease in the calcium concentration (21,22).
We previously identified piericidin A1 as an inhibitor of the T3SS in Yersinia pseudotuberculosis. In this study, we characterized the activity of piericidin A1 on the T3SSs of the related gut pathogens Y. pseudotuberculosis and Y. enterocolitica, as well as the opportunistic pathogen Pseudomonas aeruginosa.

RESULTS
Complex I inhibition by piericidin does not underlie its ability to block type III secretion. Piericidin A1 inhibits complex I of the electron transport chain in at least some bacterial species (3). Type III secretion requires a proton motive force to translocate substrates; treatment with the proton uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP) abolishes secretion (see Fig. S1A in the supplemental material). Therefore, it is possible that piericidins block type III secretion by blocking proton motive force generation. To investigate this possibility, we tested the effects of pyridaben and rotenone, two structurally distinct compounds with complex I inhibitory activity in mitochondria and certain bacteria, on Yop secretion in Y. pseudotuberculosis. Pyridaben and rotenone did not significantly block secretion of the T3SS effector YopE, while piericidin inhibited YopE secretion by 50% (see Fig. S1B), suggesting that compounds with complex I inhibitory activity do not all inhibit type III secretion.
As both nonflagellar and flagellar T3SSs require proton motive force, we tested whether piericidin A1 affects Y. pseudotuberculosis motility in a soft agar migration assay. Piericidin A1 (71 M) did not inhibit Yersinia motility, while the proton gradient uncoupler CCCP did (Fig. 1A). In addition, piericidin A1 did not alter the speed or direction of Yersinia motility, as observed by video microscopy (data not shown). Finally, we evaluated any effect piericidin A1 may have on membrane potential directly by using the membrane potential indicator dye JC-1 and observed no significant change (Fig. 1B). These data indicate that piericidin A1 does not inhibit the flagellar T3SS and that piericidins do not prevent the generation of a proton motive force needed to support type III secretion in Yersinia. Furthermore, these results suggest that piericidin A1 does not target a structure common to the flagellar and nonflagellar Ysc T3SSs or a regulator common to the two secretion systems, such as Hfq (23).
Coenzyme Q shows some structural similarity to piericidins and associates with complex I during electron transport (24) (see Fig. S2). Coenzyme Q is also sold as a dietary supplement for human consumption (25). Therefore, we tested the ability of coenzyme Q to inhibit type III secretion in Yersinia. Unlike piericidin, coenzyme Q did not inhibit YopE secretion in Y. pseudotuberculosis (see Fig. S3).
Piericidin A1 inhibits the Y. pseudotuberculosis and Y. enterocolitica Ysc T3SS but not the Y. enterocolitica Ysa T3SS. As we did not observe inhibition of the flagellar T3SS by piericidin A1, we tested its ability to inhibit the activity of T3SSs more closely related to the Y. pseudotuberculosis Ysc T3SS. Piericidin A1 reduced YopE secretion in Y. enterocolitica by 80% through the Ysc family T3SS in that organism ( Fig. 2A and B), similar to its previously reported effect on YopE secretion in Y. pseu- dotuberculosis ( Fig. 2A) (9). However, piericidin A1 did not block ExoU or ExoT secretion in P. aeruginosa (Fig. 2C), suggesting that piericidin does not inhibit the Pseudomonas T3SS, which is closely related to the Ysc T3SS (26). P. aeruginosa is known to exclude many antibiotics (27). Therefore, if the piericidin target of action is intracellular, it is possible that lack of target accessibility explains why piericidin cannot inhibit the P. aeruginosa T3SS.
To test the ability of piericidin to block a T3SS outside the Ysc family in an organism where the target is known to be accessible, we took advantage of the fact that Y. enterocolitica encodes a second T3SS on its chromosome, called the Ysa system (28,29). The Ysa T3SS is induced under high-salt conditions at low temperatures (26°C), reflecting its possible role during early stages of infection (28). We used a Y. enterocolitica yscL mutant defective in the Ysc T3SS to focus on the Ysa T3SS. We found that piericidin did not affect levels of the secreted Ysa effector proteins called Ysps (Fig. 2D). As a control, the yscL mutant did not show Ysp secretion under low-salt conditions. Lastly, a ysaT mutant previously shown to be defective in Ysa secretion (30) did not secrete Ysps. These data show that while piericidin can inhibit the Ysc T3SS in Y. enterocolitica, this compound cannot block the Ysa T3SS in the same organism.
Piericidin A1 traps T3SS effector proteins inside bacterial cells independently of the YopN, YopD, and YopK T3SS regulatory proteins. To begin to understand how piericidin inhibits Yersinia type III secretion, we tested whether piericidin A1 blocks type III secretion by blocking the expression of T3SS structural components or effectors. Transcript levels of the T3SS ATPase YscN and the effector YopE were unchanged following piericidin A1 treatment (see Fig. S4), suggesting that piericidin A1 prevents T3SS activity through a posttranscriptional mechanism. Therefore, we tested whether Piericidin A1 decreases the number of T3SS needles associated with the bacterial surface but does not alter the localization of basal body components. To determine if piericidin A1 inhibits Yop secretion by blocking the assembly of the T3SS needle, we incubated Y. pseudotuberculosis in the presence or absence of piericidin A1 and measured the number of T3SS needle complexes by staining the bacteria with antibodies recognizing the needle subunit YscF. One antibody, raised against fulllength, purified YscF, was previously used to analyze YscF puncta in Y. pseudotuberculosis (31), while the other antibody was raised against a YscF peptide for this study (see Materials and Methods). Regardless of which antibody was used, piericidin A1 reduced the number of bacterium-associated YscF puncta by approximately half (Fig. 4 and data not shown), suggesting that piericidin inhibits Yop secretion by blocking the formation of T3SS complexes, blocking the secretion of needle subunits, and/or leading to the assembly of nonfunctional needles that are not recognized by anti-YscF antibodies. We compared this phenotype to that of an inhibitor whose molecular target has been suggested through suppressor mutation analysis to be the YscF homolog in P. aeruginosa, PscF (32). Piericidin A1 reduced YscF puncta by the same magnitude as the PscF phenoxyacetamide inhibitor MBX1641 (Fig. 4).
To determine whether the reduction in needle numbers was caused (or accompanied) by a corresponding reduction in basal body assembly, we analyzed the subcellular localization of three sorting platform proteins, YscK, YscL, and YscQ, and of the basal body ring protein YscD by monitoring enhanced green fluorescent protein (EGFP)linked alleles of these proteins in Y. enterocolitica (33). While we observed a slight increase in spot intensity upon piericidin A1 treatment compared to that of the dimethyl sulfoxide (DMSO) control, there was no statistically significant change in the number of spots per unit of cell length (see Fig. S6 and S7), suggesting that the main effect of piericidin is not on the level of localization of the cytosolic or basal body T3SS components. Instead, piericidin A1 may block the secretion of T3SS cargo, including YscF or YscF needle assembly. Piericidin A1 blocks YscF higher-order structure formation and decreases the secretion of middle and late T3SS substrates. To determine if piericidin affects YscF needle polymerization, we employed a BS 3 cross-linking assay (34). We detected a decrease in higher-order YscF polymers between the control and piericidin-treated bacteria (Fig. 5A), suggesting that piericidin A1 decreases YscF needle assembly. In further support of this conclusion, we observed an accumulation of monomeric YscF in the supernatant upon piericidin A1 treatment (Fig. 5B). Taken together, these data support a model in which piericidin A1 prevents the proper assembly of YscF polymers, resulting in an accumulation of monomeric YscF in the supernatant.
The T3SS basal body secretes its cargo in a hierarchical manner, with early (ex-YscF), middle (ex-LcrV), and late (ex-YopE) substrates (35). As piericidin A1 caused an increase in monomeric YscF, decreased polymeric YscF, and decreased YopE in the supernatant, we tested the effect of piericidin A1 on the secretion of LcrV and YopD. As seen for late substrates, secretion of LcrV and YopD was decreased upon piericidin A1 treatment ( Fig. 5B and data not shown), suggesting that piericidin blocks YscF needle assembly, thereby inhibiting the secretion of middle and late T3SS substrates.
Piericidin A1 inhibition of type III secretion is reversible, and this reversibility is independent of new protein synthesis. We next sought to determine whether piericidin inhibition of Yop secretion is reversible. We incubated Y. pseudotuberculosis in the presence of piericidin A1 for 2 h under T3SS-inducing conditions, washed the bacteria to remove the compound, and allowed them to secrete T3SS effectors in fresh low-calcium medium with or without piericidin A1 for an additional 45 min. As expected, incubation of Y. pseudotuberculosis with piericidin A1 for the duration of the experiment significantly reduced the secretion of YopE (Fig. 6A). It should be noted that cells that were exposed to piericidin A1 only during the last 45 min of the experiment still had a significant decrease in secretion, indicating that piericidin can inhibit type III secretion after T3SS expression has already been induced. However, if piericidin was removed for the final 45 min of incubation, the YopE secretion level approached that of control bacteria treated with DMSO for the entire experiment (Fig. 6A). These data indicate that the action of piericidins on the T3SS is reversible, either because the piericidin A1 block in YscF needle assembly is reversible or because new needle complexes are assembled during the piericidin-free 45 min of incubation. To address the latter possibility, we repeated the washout treatment in the presence or absence of the translation inhibitor chloramphenicol during the final 45 min of incubation after the removal of piericidin A1 and analyzed YscF puncta. While chloramphenicol significantly decreased the number of YscF puncta associated with bacteria not exposed to piericidin A1, chloramphenicol did not alter the number of YscF puncta on Yersinia bacteria pretreated with piericidin A1 (Fig. 6B). These data suggest that reversal of piericidin A1 inhibition by washing away of the compound does not depend on new protein synthesis; rather, it allows for previously constructed needle complexes to regain secretion activity once piericidin is removed or for new needles to be polymerized from preformed needle subunits. Indeed, we observed an accumulation of cytoplasmic YscF in piericidin-treated bacteria under T3SS-inducing conditions (Fig. 6D), indicating that while YscF can be secreted in the presence of piericidin A1, the block in polymerization may eventually lead to a backup of YscF in the cytosol. This cytoplasmic pool may then be secreted and polymerized into a functional T3SS needle once piericidin is removed.

DISCUSSION
In this study, we characterize the activity of piericidin A1 as an inhibitor of the bacterial T3SS in Y. pseudotuberculosis and Y. enterocolitica. Piericidin A1 was originally described as an inhibitor of complex I in the mitochondrial electron transport chain (2) and in some prokaryotes (3). However, the data presented here demonstrate that piericidin A1 inhibits the T3SS independently of complex I. In addition, piericidin A1 was shown to affect quorum-sensing pathways in E. carotovora (8). However, studies on quorum sensing in Yersinia pestis indicate that quorum sensing negatively regulates the expression of T3SS genes (36). As our study shows that piericidin A1 does not affect T3SS gene expression, piericidin activity on the Yersinia T3SS is predicted to be independent of any effect on quorum sensing. Instead, piericidin blocks the assembly of the Yersinia YscF needle and subsequently the secretion of middle and late T3SS substrates without altering the localization of T3SS basal body components or inhibiting the secretion of early substrates. Furthermore, the effect of piericidin is specific, as it does not influence the Yersinia Ysa or flagellar T3SS. Thus, we propose that piericidins have an alternative molecular target specific to the Ysc T3SS in pathogenic Yersinia involved in YscF needle assembly.
Our data suggest that the molecular target of piericidin is unique to the Ysc secretion system and common to the two enteropathogenic Yersinia species but is either absent from or not well enough conserved in P. aeruginosa, which expresses a Ysc family T3SS. Alternatively, if piericidin must cross the bacterial cell envelope to carry out its T3SS inhibitory activity, P. aeruginosa may exclude or efflux piericidins. P. aeruginosa is known to have a high degree of drug resistance because of its unusually restrictive outer membrane permeability (27). However, we can conclude that piericidins do not inhibit the more distantly related Y. enterocolitica Ysa T3SS, though piericidin can inhibit the Ysc T3SS in this species.
Piericidin treatment reduced the number of YscF needle puncta visible by fluorescence microscopy at the Yersinia surface. Each punctum may represent multiple T3SS needles located in proximity to one another (37), so our quantitation of approximately two puncta per bacterium likely underestimates the actual number of T3SS needles. Nevertheless, both piericidin and the needle subunit-targeting T3SS inhibitor MBX-1641 reduced the number of YscF puncta to fewer than one per bacterium. This suggests that both compounds either disrupt needle assembly or alter needle conformation such that needles cannot be readily recognized by our antibody. As we observed a similar piericidin-induced decrease in YscF puncta when a different antibody raised against full-length YscF was used (data not shown), the latter explanation is unlikely. Our data showing that piericidin decreases the abundance of higher-order YscF polymers visible upon chemical cross-linking provides evidence for the former possibility.
Interestingly, while piericidin A1 treatment attenuates the secretion of middle and late T3SS substrates, it does not decrease effector protein expression but instead causes an accumulation of effectors in the bacterial cytosol. These results are surprising given the known T3SS feedback loop; Yop mRNA levels are positively regulated by active secretion (38). YopD and LcrQ have been shown to prevent Yop translation prior to host cell contact, and their subsequent secretion out of the bacterial cytosol through the T3SS relieves this repression (39). YopN is a regulatory protein involved in the triggering of effector secretion following host cell contact. Thus, we hypothesized that piericidin may block a regulatory pathway controlled by host cell contact, effectively causing a host contact "blind" phenotype. However, piericidin blocked type III secretion in the absence of YopD and YopN just as in wild-type (WT) bacteria. In addition, piericidin also inhibited type III secretion in the absence of YopK, a regulatory protein that controls T3SS translocon-mediated pore formation, a process recently linked to sensing of host cell contact. Collectively, these data suggest that piericidin may block the ability of the YscF T3SS needle to mediate Yop secretion by altering YscF needle assembly in a reversible manner that does not impact the T3SS feedback loop.
Whether piericidins have potential as antimicrobial therapeutics remains unclear. While their specificity and efficacy on preformed T3SSs are attractive, they have two obvious limitations. First, the micromolar concentration required to observe T3SS inhibition is impractical for use in the clinic as a treatment or in prophylaxis. A second, and related, concern is the compound's toxicity to mammalian cells. Several groups have found piericidins to be poorly tolerated in animal models, likely because of their inhibition of murine mitochondrial complex I (40). A solution may lie in structureactivity relationship studies. Perhaps more promising is the use of these molecules as laboratory tools. Given the surprising result that piericidin blocks the secretion of middle and late substrates without impacting the T3SS feedback loop, the compounds might be used to better understand T3SS regulatory pathways, as well as YscF needle formation and function. Furthermore, the compounds could be used to explore hostpathogen interactions involving temporary T3SS inactivation, as type III secretion can be restored by washing away the compound. Lastly, understanding why piericidin specifically targets the Yersinia Ysc T3SS may reveal novel mechanisms by which Yersinia bacteria have evolved to control the activity of this major virulence factor.

MATERIALS AND METHODS
Bacterial strains and growth conditions. The bacterial strains used in this study are listed in Table 1. Y. pseudotuberculosis was grown in 2ϫYT (yeast extract-tryptone) at 26°C with shaking overnight. The cultures were back diluted into low-calcium medium (2ϫYT plus 20 mM sodium oxalate and 20 mM MgCl 2 ) to an optical density at 600 nm (OD 600 ) of 0.2 and grown for 1.5 h at 26°C with shaking, followed by 1.5 h at 37°C to induce Yop synthesis as previously described (41). P. aeruginosa was grown in Luria-Bertani broth at 37°C with shaking. Y. enterocolitica was grown in 1% tryptone-0.5% yeast extract at 26°C with shaking overnight.
The full-length yopD deletion mutation (42), leaving only the first three and last three amino acids of YopD, was constructed with the primers described in reference 43. The yopK deletion mutation was introduced into Y. pseudotuberculosis with a pSR47s-derived suicide plasmid obtained from Molly Bergman (42). Y. pseudotuberculosis ΔyopD and ΔyopK mutants were isolated as previously described (42).
Piericidin A1. Piericidin A1 was isolated from Actinomyces sp. strain RL09-253-HVS-A, which was isolated from a marine sediment sample collected by scuba at Point Estero, CA, at a depth of 50 feet. This Actinomyces strain was identified as a Streptomyces sp. based on the typical morphology of streptomycetes, which form dry powdery spores on agar plates, and by sequencing of the 16S rRNA gene (data not shown). From a large-scale culture (4 liters) of the strain in SYP medium, we obtained an organic extract that was subject to a stepwise eluotropic gradient on a 10-g C 18  . A 0.33-g sample of fraction D was obtained, and the active constituent was purified by C 18 reverse-phase high-performance liquid chromatography (gradient profile, 58% to 88% MeOH-0.02% formic acid-H 2 O, 2 ml/min [Synergi 10 Fusion-RP column; Phenomenex, USA]; retention time, 30.5 min) to give 2.7 mg of piericidin A1. Electrospray ionization-time of flight high-resolution mass spectrometry analysis confirmed the molecular formula C 25 H 37 NO 4 . This structure was confirmed by one-dimensional 1 H nuclear magnetic resonance spectroscopy on a Varian Unity Inova spectrometer at 600 MHz and comparison with values in the literature. Piericidin A1 is commercially available; however, mass spectroscopy of commercially available piericidin A1 found it to be impure (data not shown). We have observed that these impurities can lead to inconsistent and artifactual results (data not shown). The concentration of piericidin A1 used in our experiments, 71 M, is approximately twice the 50% inhibitory concentration for Yop translocation (9).
Anti-YscF antibody. The anti-YscF antibody used in this study was raised in rabbits against the Y. pseudotuberculosis YscF peptide KDKPDNPALLADLQH (Pacific Immunology).
Measurement of membrane potential. Bacteria were grown under the type III secretion-inducing conditions described above, and membrane potential was measured as previously described (44). Briefly, during the last 30 min of the 37°C incubation, JC-1, a fluorescent dye widely used to measure membrane potential, was added. As a control, the protonophore CCCP was used. Samples were imaged with an LSM 5 PASCAL laser scanning microscope (Zeiss) fitted with a Plan-Apochromat 63ϫ/1.4 oil differential inference contrast (DIC) objective and analyzed with the LSM 510 software (Zeiss). Quantification of image intensities was performed with ImageJ.
Type III secretion assay. Visualization of Y. pseudotuberculosis T3SS cargo secreted in broth culture was performed as previously described (41). Y. pseudotuberculosis low-calcium medium cultures were grown for 1.5 h at 26°C. We added compounds or DMSO and switched the cultures to 37°C for another 2 h. Coenzyme Q (Sigma) was resuspended in warm DMSO. For the type III secretion reversibility assay, following the 2-h 37°C step, cultures were washed and incubated with piericidin A1 (71 M) or the equivalent volume of DMSO for another 45 min before secreted proteins were isolated (45). To test dependence on protein synthesis, following the 2-h 37°C step, cultures were washed and incubated with  (46). For Y. enterocolitica Ysa secretion, overnight cultures were grown in 1% tryptone-0.5% yeast extract (no-salt L medium) at 26°C with shaking overnight. The cultures were back diluted into L medium with either no salt or 290 mM NaCl added and incubated for 6 h at 26°C with shaking as previously described (30). Cultures were centrifuged at 13,200 rpm for 10 min at room temperature. Supernatants were transferred to a new Eppendorf tube. For Y. enterocolitica supernatants, the upper two-thirds only was removed and centrifuged again. The upper two-thirds was removed once again and passed through a 0.22-m filter prior to trichloroacetic acid (TCA) precipitation. Ten percent (final concentration) TCA was added, and the mixture was vortexed vigorously. Samples were incubated on ice for 20 min and then centrifuged at 13,200 rpm for 15 min at 4°C. The pellet was resuspended in final sample buffer (FSB) plus 20% dithiothreitol (DTT). Samples were boiled for 15 min prior to being subjected to 12.5% SDS-PAGE. Sample loading was normalized for the bacterial density (OD 600 ) of each sample. Y. pseudotuberculosis and P. aeruginosa samples were visualized by Coomassie blue staining, while Y. enterocolitica Ysa samples were visualized with SYPRO Ruby (Molecular Probes). Densitometric quantification of the bands was done with Image Lab software (Bio-Rad) with the relevant DMSO-treated Y. pseudotuberculosis YopE band set to 1.00.
Secreted and cytosolic YopE Western blot analysis. Supernatant samples were obtained as described above for the type III secretion assay. In parallel, after the supernatant was collected, the pellet was resuspended in the same volume of FSB plus 20% DTT as the corresponding supernatant sample. The pellet samples were then boiled for 15 min. Cytosolic and secreted protein samples were subjected to 12.5% SDS-PAGE and transferred to a blotting membrane (Immobilon-P) with a trans-Blot semidry transfer cell (Bio-Rad). Blots were blocked overnight in Tris-buffered saline with Tween 20 and 5% skim milk and probed with a YopE primary antibody and a horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotech). Following visualization, densitometric quantification of the bands was performed with Image Lab software (Bio-Rad) with the first DMSO-treated WT Y. pseudotuberculosis YopE band set to 1.00.
Detection of secreted monomeric YscF. As we could not detect secreted monomeric YscF from WT bacteria, presumably because it becomes rapidly polymerized into the needle structure, we used a strain of Y. pseudotuberculosis that ectopically expresses YscF from an isopropyl-␤-D-thiogalactopyranoside (IPTG)-inducible pTCR99A plasmid (pYscF). Cultures were grown overnight in 2ϫYT with 100 g/ml carbenicillin. The following day, cultures were back diluted to an OD 600 of 0.2 in low-calcium medium plus carbenicillin and grown at 26°C for 1.5 h. At the 37°C shift, 10 M IPTG and either 71 M piericidin A1 or DMSO were added to the cultures, which were incubated for 2 h. After incubation, bacteria were normalized by OD 600 and TCA precipitation was preformed as described above.
Cross-linking assay. Yersinia bacteria carrying YscF on an IPTG-inducible plasmid (pYscF) were grown under the T3SS-inducing conditions described above. At the 37°C shift, 10 M IPTG was added to induce YscF expression. After a 2-h induction, the cells were collected and the crosslinker BS 3 was added as previously described (34). Samples were boiled in Laemmli buffer and pulled through a 25-gauge needle several times to shear genomic DNA. Samples were loaded onto a 16.5% Tris-Tricine gel (Bio-Rad) and run at 120 V. Proteins were transferred to a polyvinylidene difluoride membrane (Immobilon-P) with a Mini-Trans-Blot cell (Bio-Rad) and probed with a YscF antibody (1:10,000) overnight, followed by a horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotech).
Motility assay. Following overnight growth, bacterial cultures were diluted to an OD of 2.5. One microliter of WT Y. pseudotuberculosis or a nonmotile mutant expressing flhDC Y. pestis (41) was spotted onto motility medium (1% tryptone, 0.25% agar) in six-well plates. Each well contained either 0.3% DMSO or 71 M piericidin A1, rotenone, or pyridaben. The plates were stored at 26°C for 24 h before the diameter of swimming motility from the center was measured. To increase bacterial visibility, 24 mM 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide was added on top of the motility agar just before imaging.
Real-time PCR. Overnight bacterial cultures grown in M9 were back diluted and grown in M9 at 37°C to induce type III secretion as described in reference 44, in the presence of DMSO, piericidin A1, rotenone, or pyridaben. RNA was isolated with an RNeasy Plus Micro kit (Qiagen) according to the manufacturer's instructions. SYBR green PCR master mix (Applied Biosystems) was used for quantitative PCRs (qPCRs) according to the manufacturer's instructions at a 60°C annealing temperature. The primers used are listed in Table S1. The YopE primers were designed with primer 3 software (http://fokker.wi.mit.edu/primer3/ input.htm). Results were analyzed with the Bio-Rad CFX software.
Immunofluorescence staining of YscF. Visualization of YscF was carried out as previously described (31), with the following modifications. Coverslips were blocked overnight with phosphate-buffered saline with Tween 20 (PBST) plus 3% bovine serum albumin (PBST-BSA) at 4°C. The blocking solution was removed, an anti-YscF primary antibody was added at 1:10,000 in PBST-BSA, and the mixture was rocked at 4°C for 4 h. The coverslips were carefully rinsed in ice-cold PBST plus 0.1% Tween 20 several times, incubated with an Alexa Fluor 594-conjugated anti-rabbit secondary antibody (Invitrogen) at 1:10,000 in PBST-BSA, rocked at 4°C for 3 h, and rinsed again in ice-cold PBST plus 0.1% Tween 20 several times. The coverslips were then stained for nuclear material with Hoechst 33342 (Thermo Scientific) at 1:10,000 in PBST-BSA and left in the dark at room temperature for 30 min. The coverslips were washed in ice-cold PBST plus 0.1% Tween 20 several times, allowed to dry briefly, mounted onto glass coverslips with ProLong Gold (Thermo Scientific), and sealed with clear nail polish. Images were taken with a Zeiss Axio Imager Z2 widefield microscope under 63ϫ/1.4 oil immersion with Zen software and pseudocolored and merged in FIJI. The conditions were blinded, and the puncta associated with a single bacterium were counted.
YscK, YscL, YscD, and YscQ analysis. Y. enterocolitica bacteria expressing N-terminal fluorescent fusions of cytosolic T3SS components, EGFP-YscK, mCherry-YscL, EGFP-YscQ, or EGFP-YscD joined by a glycine-rich flexible 13-amino-acid linker and expressed from the native genetic environment on the virulence plasmid (47,48) were grown overnight in BHI broth containing nalidixic acid (35 g/ml) and diaminopimelic acid (80 g/ml). Main cultures were inoculated to an OD of 0.12 in the same medium additionally complemented with 5 mM EDTA, 10 mM MgCl 2 , and 0.4% glycerol. Cultures were grown for 1.5 h at 28°C. A 71 M concentration of piericidin A1 or the equivalent volume of DMSO was added, and the cultures were switched to 37°C for another 3 h. A 1.5-l volume of the culture was layered on a patch of 1.5% agarose in M22 medium supplemented with 80 g/ml diaminopimelic acid, 5 mM EDTA, 10 mM MgCl 2 , 0.4% glycerol, and either 71 M piericidin A1 or the equivalent volume of DMSO. Fluorescence and DIC images of z stacks (n ϭ 4, Δz ϭ 250 nm) were taken on a Deltavision optical sectioning microscope (Applied Precision) equipped with a UPlanSApo 100ϫ/1.35 oil objective (Olympus). DIC frames were imaged for 10 ms, and mCherry and EGFP frames were imaged for 1 s with appropriate filter sets. Detection and quantification of fluorescent foci were performed as described in reference 33.
Statistical analysis. Statistical analysis was performed with Kaleidagraph (Synergy Software) or GraphPad as specified in each figure legend.