Identification of YsaP, the Pilotin of the Yersinia enterocolitica Ysa Type III Secretion System

ABSTRACT

Secretins are multimeric outer membrane pore-forming proteins found in complex export systems in Gram-negative bacteria. All type III secretion systems (T3SSs) have a secretin, and one of these is the YsaC secretin of the chromosomally encoded Ysa T3SS of Yersinia enterocolitica. In some cases, pilotin proteins, which are outer membrane lipoproteins, are required for their cognate secretins to multimerize and/or localize to the outer membrane. However, if secretin multimers mislocalize to the inner membrane, this can trigger the protective phage shock protein (Psp) stress response. During a screen for mutations that suppress YsaC toxicity to a psp null strain, we isolated several independent mutations predicted to increase expression of the YE3559 gene within the Ysa pathogenicity island. YE3559, which we have named ysaP, is predicted to encode a small outer membrane lipoprotein, and this location was confirmed by membrane fractionation. Elevated ysaP expression increased the steady-state level of YsaC but made it less toxic to a psp null strain, and it also decreased YsaC-dependent induction of psp gene expression. Subsequent experiments showed that YsaP was not required for YsaC multimerization but was required for the multimers to localize to the outer membrane. Consistent with this, a ysaP null mutation compromised protein export by the Ysa T3SS. All these observations suggest that YsaP is the pilotin for the YsaC secretin. This is only the second pilotin to be characterized for Yersinia and one of only a small number of pilotins described for all bacteria.

IMPORTANCE Secretins are essential for the virulence of many bacterial pathogens and also play roles in surface attachment, motility, and competence. This has generated considerable interest in understanding how secretins function. However, their fundamental differences from typical outer membrane proteins have raised various questions about secretins, including how they are assembled into outer membrane multimers. Pilotin proteins facilitate the assembly of some secretins, but only a small number of pilotins have been identified, slowing efforts to understand common and distinct features of secretin assembly. This study provides an important advance by identifying a novel member of the pilotin family and also demonstrating a method of pilotin discovery that could be broadly applied.

INTRODUCTION

Yersinia enterocolitica is a bacterial pathogen that causes gastrointestinal syndromes in humans (1). In addition to being an important pathogen in its own right, it has also been a useful model organism for studying conserved aspects of bacterium-host interactions. For example, Y. enterocolitica provided one of the first reports of protein secretion from what was later found to be a type III secretion system (T3SS) (2). T3SSs are widely conserved multicomponent secretion machines in Gram-negative bacteria that export effector proteins from the bacterial to the host cell cytoplasm (for a recent review, see reference 3). The T3SS first discovered in Y. enterocolitica is the Ysc system, which is encoded by a plasmid that is also conserved in the other human-pathogenic Yersinia species, Y. pseudotuberculosis and Y. pestis (4). The Ysc T3SS is essential for the virulence of all three species, because it exports effectors (Yops) that disarm the innate immune response (5). A second T3SS, known as the Ysa system, is encoded on the chromosome of highly virulent Y. enterocolitica biovar 1B strains (6, 7). The Ysa system exports multiple effectors (Ysps) (8, 9), although the exact roles of the Ysa system and the Ysps are unclear. However, mutants with a nonfunctional Ysa system have propagation defects during the early stages of a mouse infection (7, 8, 10) and a major defect in intracellular replication within Drosophila melanogaster S2 cells (11). Furthermore, the Ysa system has also been shown to export some of the Yops (12).

An essential component of all T3SSs is a secretin, which is an outer membrane (OM) multimeric pore-forming protein that is also found in type II secretion systems and type IV pili (13). Filamentous phages also encode a secretin that is part of a system used to export new phage particles from bacterial cells (14). Some secretins have been implicated as being pivotal for the assembly of their entire export systems (reviewed in reference 13). For example, the YscC secretin has been proposed to be the first component of the Ysc T3SS assembled (15). For this and other reasons, there is interest in understanding how secretins assemble into multimeric OM pores. However, secretins are distinct from typical β-barrel OM proteins (Omps), which are inserted into the OM by the β-barrel assembly machine (Bam complex) (16). The Bam complex is not required for the OM insertion of some secretins, and even when it is required, this might be due to its role in inserting accessory proteins rather than the secretin itself (17–19). One accessory protein involved in secretin assembly is a pilotin, which is a small OM lipoprotein (20, 21). Pilotins work with one specific secretin and are required for its multimerization and/or OM insertion. Cotransport of a pilotin-secretin complex to the OM by the localization of lipoprotein (Lol) system has been proposed as a mechanism by which pilotins might facilitate secretin assembly (22, 23). However, only a small number of pilotins have been described, perhaps because their diverse sequences, structures, and genomic contexts make them difficult to identify (21).

Remarkably for an OM protein, secretin multimers can also insert into the inner membrane (24). This aberrant event triggers an extracytoplasmic stress response known as the phage shock protein (Psp) system (24, 25). In Y. enterocolitica, production of the YscC secretin of the Ysc T3SS induces psp gene expression and is toxic to psp null mutants (26). However, these effects are reduced when YscC is coproduced with its pilotin, YscW, which facilitates the insertion of YscC multimers into the OM (26, 27). Similarly, induction of the Escherichia coli Psp response by production of the Klebsiella oxytoca PulD secretin is reduced by coproduction with its pilotin, PulS (20).

The mechanism by which the Psp system prevents secretin toxicity is unknown. In ongoing experiments to investigate this phenomenon, we screened for transposon overexpression mutations that suppress the toxicity caused by production of the Ysa T3SS secretin YsaC in a psp null strain. Half of the suppressor mutations were predicted to overexpress a gene within the Ysa pathogenicity island. Here we show that this gene encodes the pilotin for YsaC, which we have now named YsaP. YsaP mitigates YsaC toxicity to a psp null strain by facilitating the insertion of YsaC multimers into the OM, and it is required for the function of the native Ysa T3SS. YsaP provides an important addition to the relatively short list of known pilotins and is only the second one to be described for Yersinia. This work also highlights the possibility that the specific relationship between secretins and the Psp system could be exploited to identify other pilotins and accessory proteins involved in secretin assembly.

MATERIALS AND METHODS

Bacterial strains, primers, and routine growth.

Bacterial strains and plasmids are listed in Table 1, and primer sequences are listed in Table 2. The DNA sequences of all PCR-generated fragments were checked. The standard growth medium was Luria-Bertani (LB) broth or LB agar. Antibiotics were used as described previously (28).

TABLE 1.

Strains and plasmids used for this study

Strain or plasmid	Genotype and/or features	Reference or source
Y. enterocolitica strains
AJD3a	ΔyenR (R− M+)	51
AJD977	ΔyenR (R− M+) ΔaraGFB::[Φ(pspAp-lacZY)]	52
AJD1171	ΔyenR (R− M+) Δ(pspF-ycjF) ΔpspG	31
AJD4991	ΔyenR (R− M+)::[araC-araBp] ΔysaC Δ(pspF-ycjF) ΔpspG	This study
AJD4992	ΔyenR (R− M+)::[araC-araBp-ysaC-his6] ΔysaC Δ(pspF-ycjF) ΔpspG	This study
AJD5201b	ΔyenR (R− M+) ΔysaC	This study
AJD5441c	ΔyenR (R− M+)::[araC-araBp-ysaC-his6] ΔysaC Δ(pspF-ycjF) ΔpspG YE3558::TnMod-RKm′-lacIq-tacp	This study
AJD5442c	ΔyenR (R− M+)::[araC-araBp-ysaC-his6] ΔysaC Δ(pspF-ycjF) ΔpspG YE3558::TnMod-RKm′-lacIq-tacp	This study
AJD5487	ΔyenR (R− M+) ΔysaP	This study
Plasmids
pBAD33	Cmr p15A ori; araBp expression vector	53
pSR47S	Kmr R6K ori mob+ (RP4) sacB+	54
pVLT35	Smr Spr RSF1010 ori; tacp expression vector	55
pAJD428	R6K ori; TnMod-RKm′-lacIq-tacp delivery plasmid	28
pAJD801	tacp-ysaC-his6 in pVLT35	28
pAJD935	araBp-ysaC-his6 in pBAD33	25
pAJD2372	araBp-ysaP in pBAD33	This study
pAJD2373	araBp-ysaP-FLAG in pBAD33	This study
pAJD2474	araBp-ysaP(T28D) in pBAD33	This study
pAJD2497	araBp-ysaP(T28D)-FLAG in pBAD33	This study

^aAJD3 is a virulence plasmid-cured derivative of strain JB580v (51). All other Y. enterocolitica strains listed are derivatives of AJD3, unless stated otherwise.

^bAJD5201 is a virulence plasmid-cured derivative of YVM1178 (9).

^cIn the genome sequence (56), the transposon insertion is between nucleotides 3873295 and 3873296 in AJD5441 and between nucleotides 3873432 and 3873433 in AJD5442.

TABLE 2.

Primers used in this study

Primer	Sequence (5′–3′)a
321	CTGGACTAGTCAAATGGACGAAGCAGGGATTCTGC
433	TTTATCAGACCGCTTCTGCG
1835	GGCTCTAGAACGCCTTGAACTTAGTTAATC
1836	CGCGGATCCACGAGATTTCATCGGATTATCC
1839	CGGGAGCTCTTTTACCGCAGGATAATCCG
1840	GGCTCTAGAAGCCTCGATCAAGTGAG
1841	GGCTCTAGACTATTTATCGTCGTCATCTTTGTAGTCACTGTCATTTACTGTCAGTAATAACTTCTC
2024	CTGGTTTTATCCGGCTGTGACCTGCCGACTGACAATAGC
2025	GCTATTGTCAGTCGGCAGGTCACAGCCGGATAAAACCAG
2087	CGCGGATCCACGCTGCCGACTGACAATAG
2088	CGGGAGCTCGGGTGTATTACTGACACA

^aRestriction sites are underlined.

Plasmid construction.

araBp-ysaP and araBp-ysaP-FLAG expression plasmids were made by amplifying ysaP from the Y. enterocolitica chromosome by using primer pairs 1839/1840 and 1839/1841, respectively. The fragments were digested with SacI-XbaI and ligated into pBAD33. The araBp-ysaP(T28D) expression plasmid was made by splicing overlap extension (SOE) PCR (29). Two fragments were amplified from the Y. enterocolitica chromosome by using primer pairs 1839/2025 and 1840/2024, joined by SOE PCR with primer pair 1839/1840, digested with SacI-XbaI, and ligated into pBAD33 to make pAJD2474. The araBp-ysaP(T28D)-FLAG expression plasmid was made by amplifying the insert of pAJD2474 by using primer pair 1839/1841, digesting the product with SacI-XbaI, and then ligating it into pBAD33.

Construction of chromosomal araBp expression strains.

For construction of AJD4991, an araC-araBp fragment was amplified from pBAD33 by using primers 321 and 433, digested with SpeI-XbaI, and cloned into the unique XbaI site of a previously described sacB⁺ allelic exchange plasmid that integrates into the yenR deletion site in all our Y. enterocolitica strains (30). The araC-araBp cassette was then introduced into the chromosomal yenR deletion site of ΔysaC strain YVM1178 (9) by plasmid integration, followed by selection for sucrose-resistant segregants and confirmation by colony PCR. The Δ(pspF-ycjF) and ΔpspG deletions were then introduced sequentially by using previously described suicide plasmids and procedures (31). Finally, the virulence plasmid was cured as described before (26). AJD4992 was made in exactly the same way, except that in the first step, an araC-araBp-ysaC-his₆ fragment was amplified from pAJD935 by using primers 321 and 433.

Construction of a ysaP in-frame deletion mutant.

Two ∼550-bp fragments corresponding to the regions immediately upstream and downstream of the deletion site were amplified with primer pairs 1835/1836 and 2087/2088, respectively. The fragments were digested with XbaI-BamHI (upstream fragment) and SacI-BamHI (downstream fragment) and ligated into plasmid pSR47S that had been digested with XbaI-SacI. This construct was used to introduce the ysaP deletion into strain AJD3 by plasmid integration, followed by selection for sucrose-resistant segregants and confirmation by colony PCR.

Transposon mutagenesis.

Y. enterocolitica strain AJD4992 was mutagenized with the Tn5-based transposon encoded by plasmid pAJD428, which was delivered from E. coli by conjugation, as described previously (28, 32). Mutants were isolated on LB agar containing nalidixic acid, kanamycin, 0.5% (wt/vol) arabinose, and 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside). Mutants with IPTG-dependent suppression of YsaC-induced toxicity were identified as described in Results. Southern hybridization using the kanamycin resistance gene of the transposon as a probe was used to confirm a single transposon insertion. Insertion sites were determined by arbitrarily primed PCR and DNA sequencing (33).

YsaC-induced stress tolerance assay.

Strains were grown to saturation at 26°C in LB broth containing appropriate antibiotics. Optical densities at 600 nm (OD₆₀₀) were determined and, if necessary, adjusted to be equivalent for all strains by bacterial cell concentration (centrifugation and resuspension in appropriate volumes of culture medium). Three-microliter aliquots of undiluted sample and serial 10-fold dilutions of each sample were placed onto the surface of LB agar containing appropriate antibiotics as well as arabinose and, in some cases, IPTG, as indicated in the figure legends. Plates were incubated at 37°C for approximately 24 h.

β-Galactosidase assays.

Saturated cultures were diluted in 5 ml of LB broth in 18-mm-diameter test tubes to an optical density (600 nm) of approximately 0.04. The growth medium contained antibiotics for plasmid selection and 0.2% (wt/vol) arabinose to induce araBp-ysaP expression. Cultures were grown on a roller drum at 37°C for 4 h, and β-galactosidase enzyme activity was determined at room temperature in permeabilized cells, as described previously (34). Activities are expressed in arbitrary Miller units (35). Individual cultures were assayed in duplicate, and values obtained for three independent cultures were averaged.

Polyclonal antisera and immunoblotting.

Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes by electroblotting. Enhanced chemiluminescence detection followed sequential incubation with polyclonal antiserum or monoclonal antibodies and then horseradish peroxidase-conjugated goat anti-rabbit IgG or goat anti-mouse IgG (Sigma). Polyclonal antisera were diluted 20,000-fold for anti-PspB (36) and 10,000-fold for anti-FtsH (30) and anti-OmpA (Antibody Research Corporation). Monoclonal antibodies were diluted 5,000-fold for anti-THE-His (GenScript) and 10,000-fold for anti-FLAG M2 (Sigma).

Membrane fractionation.

Three hundred milliliters of culture at an OD₆₀₀ of 0.6 to 0.8 was collected by centrifugation at 4°C, and the pellet was resuspended in 13.5 ml of 20% (wt/wt) sucrose in 25 mM Tris-HCl (pH 8). The suspension was supplemented with DNase I and RNase A, each to a final concentration of 10 μg/ml, as well as with complete EDTA-free protease inhibitors (Thermo Scientific). Cells were disrupted by two passages through a French pressure cell (1,100 lb/in²), and unbroken cells were removed by centrifugation at 8,000 × g for 10 min at 4°C. Membranes were collected from the supernatant by ultracentrifugation at 150,000 × g for 60 min at 4°C. The membrane pellet was homogenized in approximately 2 ml of 25% (wt/wt) sucrose in 25 mM Tris-HCl, 1 mM EDTA (pH 8), and then approximately 1.7 ml was placed on top of a discontinuous sucrose gradient that was made by layering 1.69 ml each of 55%, 50%, 45%, 40%, 35%, and 30% (wt/wt) sucrose in 25 mM Tris-HCl, 1 mM EDTA (pH 8), into a 14- by 89-mm Ultra-Clear tube (Beckman Coulter). The gradients were centrifuged using a Beckman Coulter SW41 Ti rotor at 230,000 × g for 40 to 42 h at 4°C. Twenty fractions (592 μl) were collected from the tops of the tubes and analyzed by SDS-PAGE followed by immunoblotting or silver staining (37).

Ysp secretion assay.

Saturated cultures were used to inoculate 5 ml of LB broth containing 290 mM NaCl in an 18-mm-diameter test tube to an OD₆₀₀ of 0.1. The growth medium contained chloramphenicol for plasmid selection and 0.2% (wt/vol) arabinose to induce araBp-ysaC or araBp-ysaP expression. The cultures were incubated at 26°C on a roller drum for 6 h. Bacterial cells were removed by centrifugation at 8,000 × g for 5 min. The upper 3.5 ml of supernatant was passed through a 0.2-μm-pore-size low-protein-binding filter (Millipore). Proteins were precipitated with 10% (vol/vol) trichloroacetic acid, washed with acetone, and resuspended in SDS-PAGE sample buffer. Samples from supernatant equivalent to that derived from 150 μl of culture at an OD₆₀₀ of 1 were separated by SDS-PAGE (10% polyacrylamide) and visualized by silver staining (37).

Modeling of the YsaP structure.

HHpred (38) at the Bioinformatics Toolkit website (http://toolkit.tuebingen.mpg.de) was used to search for YsaP homologs based on predicted secondary structure similarity. Default search settings were used, except that all standard hidden Markov model databases were selected and the alignment mode was set to global. Search results were used in conjunction with MODELLER (39; also available at http://toolkit.tuebingen.mpg.de) to generate a three-dimensional (3D) structural model of YsaP by using the “automatically select best templates” option, with the template automatically chosen being Pseudomonas aeruginosa ExsB (PDB accession no. 2YJL). The PDB data file of the predicted YsaP structural model was then used with Jmol software (http://jmol.org) to view the structure as a ribbon model.

RESULTS

Overexpression mutations in the Ysa pathogenicity island suppress YsaC toxicity to a psp null strain.

Our laboratory is interested in understanding how the Psp system prevents mislocalized secretins from killing bacterial cells (25). We have begun to explore the approach of isolating mutations that suppress secretin toxicity in a Y. enterocolitica psp null strain. An arabinose-inducible araBp-ysaC-his₆ expression cassette was integrated into the chromosome of a strain with all psp genes deleted (strain AJD4992) (Table 1). At 37°C, this strain grew poorly in medium containing arabinose due to YsaC secretin production, whereas it grew normally on medium without arabinose (data not shown). A psp⁺ strain with the same araBp-ysaC-his₆ cassette grew normally with or without arabinose (data not shown). Strain AJD4992 was mutagenized with a mini-Tn5 transposon carrying an outward-facing tac promoter (28). Large colonies selected after overnight incubation on agar containing arabinose and IPTG were screened for poor growth without IPTG, which indicated that transposon tacp-dependent expression of a downstream gene(s) suppressed the YsaC toxicity.

Forty-three IPTG-dependent suppressor mutants were isolated from a library of ∼40,000 random transposon mutants. Nine had insertions downstream of the araBp-ysaC-his₆ cassette with the tac promoter facing the end of the ysaC-his₆ gene (the ysaC gene tagged with the coding sequence for a His₆ tag). These mutants had reduced YsaC-His₆ protein production, probably because of antisense interference with araBp-ysaC-his₆ expression (data not shown). Of the remaining 34 mutants, 17 had an insertion in the Ysa pathogenicity island, and these are the focus of this report (Fig. 1A) (the other mutants will be described elsewhere). Two of the Ysa pathogenicity island transposon mutants were chosen for further analysis (Fig. 1A). On medium containing arabinose, the parental strain without a transposon insertion grew poorly with or without IPTG, due to toxicity caused by YsaC-His₆ production. In contrast, the poor growth of the suppressor mutants was alleviated by IPTG (Fig. 1B) (all three strains grew well on medium that did not contain arabinose to induce YsaC-His₆ production [data not shown]). We also monitored the steady-state level of YsaC-His₆ protein by immunoblotting of boiled cell lysates (boiling dissociates the YsaC-His₆ secretin multimers). This revealed that the suppressors increased the YsaC-His₆ steady-state level, even though they reduced the toxicity it caused (Fig. 1C).

FIG 1.

Ysa pathogenicity island Tn-tacp insertions. (A) Partial map of the Ysa pathogenicity island showing the locations of the 17 Tn-tacp insertions. Each flag shows the approximate location of each transposon insertion, with the direction of the arrow representing the orientation of the tac promoter. The mutants used to generate the data in panels B and C are labeled 1 and 2 (AJD5441 and AJD5442, respectively). (B) Suppression of YsaC-His₆ toxicity to a psp null strain. Serial 10-fold dilutions of saturated cultures were placed as spots on LB agar containing 0.5% (wt/vol) arabinose to induce YsaC-His₆ production, with or without 1 mM IPTG to induce the transposon tac promoter, and incubated at 37°C. No suppressor, parental strain AJD4992, which has no transposon insertion. (C) Suppressor mutations increase the steady-state level of YsaC-His₆. The blots show an anti-His₆ immunoblot of boiled whole-cell lysates (YsaC-His₆) and Ponceau S total protein staining (protein) of the same region of the nitrocellulose membrane. Strains were grown in LB broth with 0.5% arabinose to induce YsaC-His₆ production, with or without 1 mM IPTG to induce the transposon tac promoter. No YsaC, strain AJD4991, which does not have the araBp-ysaC-his₆ expression cassette or a transposon insertion; no supp, strain AJD4992, which has the araBp-ysaC-his₆ expression cassette but no suppressor (transposon) mutation.

YsaP reduces YsaC toxicity to a psp null strain and YsaC-dependent induction of the Psp response.

The transposon had inserted at a different position in each of the 17 Ysa pathogenicity island insertion mutants (Fig. 1A and data not shown). However, the insertions were clustered close to one end of the pathogenicity island, and the transposon had always inserted in the same orientation (Fig. 1A). Based on the transposons' locations, common orientation, and IPTG dependence, it was likely that increased expression of the YE3559 gene, which we have now named ysaP, was responsible for suppressing YsaC toxicity. All of the insertions were upstream of ysaP, whereas none were within ysaP or downstream of it. Therefore, ysaP was cloned into an arabinose-inducible expression plasmid and tested for the ability to suppress YsaC-His₆ toxicity. For this experiment, ysaC-his₆ was expressed from a multicopy tacp expression plasmid, which was more toxic to a Δpsp strain than the single-copy araBp-ysaC-his₆ cassette used in the screen (compare Fig. 1B and 2A). This analysis showed that overexpression of ysaP alone was sufficient to suppress YsaC-His₆ toxicity in a Δpsp strain (Fig. 2A). It remains possible that other genes affected by some of the transposon insertions play additional roles related to YsaC toxicity, but we did not investigate that possibility in this study.

FIG 2.

YsaP suppresses the toxicity of YsaC to a psp null strain and reduces YsaC-dependent induction of the Psp response. (A) Suppression of YsaC-His₆ toxicity to a psp null strain. psp⁺ (AJD3) and Δpsp (AJD1171) strains contained a tacp-ysaC-his₆ plasmid as well as an araBp-ysaP plasmid (↑YsaP) or the empty araBp vector (−). Serial 10-fold dilutions of saturated cultures were placed as spots onto LB agar containing 0.2% (wt/vol) arabinose to induce YsaP production and then incubated at 37°C (ysaC-his₆ expression without IPTG is sufficient for toxicity to a psp strain at 37°C). (B) Reduced YsaC-His₆-dependent induction of the Psp response. Derivatives of Φ(pspA-lacZ) operon fusion strain AJD977 carried the araBp-ysaP plasmid (+YsaP) or the empty araBp vector (−YsaP) and were grown and had their β-galactosidase activity determined as described in Materials and Methods. Gray bars, strains with the empty tacp vector control; black bars, strains with a tacp-ysaC-his₆ plasmid. β-Galactosidase activities were averaged for three independent cultures, and the error bars indicate the positive standard deviations from the means. The panels at the bottom show an anti-His₆ immunoblot of the boiled total cell lysates from one set of cultures and Ponceau S total protein staining (protein) of the same region of the nitrocellulose membrane.

When the Psp system is intact, secretin production induces pspA operon expression (e.g., see references 26 and 28). Therefore, we also tested the effect of increased YsaP production on YsaC-dependent induction of Φ(pspA-lacZ) operon fusion expression. YsaC-His₆ production induced Φ(pspA-lacZ) expression approximately 4-fold (Fig. 2B). However, this was reduced to approximately 2-fold when YsaC-His₆ was coproduced with YsaP, even though the YsaC-His₆ steady-state level increased (Fig. 2B).

YsaP is an outer membrane protein.

The N terminus of the predicted YsaP protein has the classic features of a lipoprotein: a predicted signal sequence cleaved by signal peptidase II followed immediately by a cysteine residue that is the target of lipidation (Fig. 3A) (40). Furthermore, the amino acid following cysteine (referred to as the +2 position) is threonine (Fig. 3A). In E. coli, the +2 amino acid determines whether a lipoprotein is retained in the inner membrane or transferred to the OM by the Lol system (41). Threonine does not cause inner membrane retention in E. coli (41, 42). Therefore, if the LolA avoidance signals of E. coli and Y. enterocolitica are similar, the results suggest that YsaP is an OM lipoprotein.

FIG 3.

YsaP is an outer membrane protein. (A) The first 33 amino acids encoded by the ysaP gene are shown, with the predicted signal peptidase cleavage site indicated by a hyphen. The lipobox characteristic of lipoprotein signal sequences is labeled and underlined, with the consensus sequence shown below, in parentheses. The threonine at position 28 of the predicted preprotein, corresponding to the +2 position of the mature protein, is also labeled. (B) Membrane fractionation analysis. ΔysaP strain AJD5487 contained plasmid pBAD33 derivatives encoding YsaP or YsaP-T28D, both of which also had a C-terminal FLAG tag for immunodetection. Strains were grown in LB broth containing 0.001% (wt/vol) arabinose, and membranes were prepared and fractionated in discontinuous sucrose gradients as described in Materials and Methods. Fractions 3 to 20 were boiled and analyzed by SDS-PAGE and immunoblotting with anti-FLAG antibodies (to detect YsaP), anti-OmpA antiserum (outer membrane control), and anti-FtsH antiserum (inner membrane control). The approximate locations of peak fractions for the inner (IM) and outer (OM) membranes, based on the fractionation of the control proteins, are indicated.

To investigate the location of YsaP, we constructed an araBp-ysaP expression plasmid encoding YsaP with a C-terminal FLAG epitope for immunodetection (the FLAG tag did not abolish YsaP function [see below]). We also constructed a plasmid encoding a derivative with the threonine at position +2 of the mature protein changed to aspartate (YsaP-T28D). In E. coli, aspartate at position +2 is a Lol avoidance signal that promotes inner membrane retention of lipoproteins (41, 42). Membranes were separated into fractions of various densities by centrifugation in discontinuous sucrose gradients, and the fractions were analyzed by SDS-PAGE and immunoblotting (Fig. 3B) (YsaP did not accumulate in the soluble fraction [data not shown]). YsaP fractionated similarly to the OM control protein OmpA, with the peak fractions for both proteins coinciding. However, YsaP-T28D showed an altered localization pattern that was more similar to that of the inner membrane control protein FtsH (the T28D mutation also reduced the abundance of YsaP, most likely because it affected its stability). These results confirm that YsaP is an OM protein and that the +2 position of the predicted mature protein is important for this localization.

YsaP is required for the localization of YsaC secretin multimers in the outer membrane.

All of the preceding data are consistent with YsaP being the pilotin for YsaC. For example, pilotins are small OM lipoproteins (21) (the predicted molecular mass of the unlipidated mature YsaP protein is 13.2 kDa). Also, in common with YsaP, the only characterized pilotin in Y. enterocolitica, YscW, reduces the toxicity of its secretin, YscC, to a psp null strain and reduces YscC-dependent induction of Φ(pspA-lacZ) expression (26). YscW does this because it promotes the correct localization of YscC multimers into the OM rather than into the inner membrane (24, 25, 27). Therefore, if YsaP is the YsaC pilotin, we predicted that it would affect the localization of YsaC multimers.

Membranes of ΔysaP strains containing the tacp-ysaC-his₆ expression plasmid and either an araBp-ysaP plasmid or the empty araBp plasmid control were fractionated by discontinuous sucrose gradient centrifugation. Unheated fractions were analyzed by SDS-PAGE and immunoblotting to detect SDS-resistant YsaC-His₆ multimers. However, protein staining of gels after electrotransfer revealed that some multimers remained in the gel regardless of the transfer method used (data not shown). Similar problems were reported with YscC multimers (27). Therefore, like the authors of that study, we detected YsaC multimers by direct silver staining of SDS-PAGE gels. The YsaC multimers fractionated with the OM when YsaC was coproduced with YsaP (Fig. 4). However, the YsaC multimers peaked in the inner membrane fractions when YsaP was absent. The multimers were also less abundant without YsaP, which is consistent with the impact of YsaP on the steady-state level of YsaC in the earlier experiments (Fig. 1C and 2B). Immunoblot detection of the YsaC monomer band in the same unheated fractions revealed a localization and relative abundance in each strain similar to those of the multimer (data not shown). However, we cannot safely draw any conclusions regarding YsaC monomer distribution in cells, because YsaC multimers are easily dissociated (e.g., by heating), and some multimers might break down into monomers during sample processing. Finally, immunodetection of the inner membrane protein PspB showed that it was far more abundant when YsaC was produced without YsaP (Fig. 4). This was expected and is consistent with increased mislocalization of YsaC multimers to the inner membrane inducing the Psp response (Fig. 2B). These data show that YsaP is not required for YsaC multimerization but is required for the efficient OM localization of the multimers. This supports the hypothesis that YsaP is the pilotin for YsaC.

FIG 4.

Membrane fractionation analysis of YsaC, with and without YsaP coproduction. ΔysaP strain AJD5487 contained a tacp-ysaC-his₆ plasmid as well as an araBp-ysaP plasmid (+ YsaP) or the empty araBp vector (− YsaP). Cultures were grown in medium containing 0.2% (wt/vol) arabinose to induce YsaP production (leaky IPTG-independent expression of ysaC-his₆ was used as described in the legend to Fig. 2). Membranes were prepared and fractionated on discontinuous sucrose gradients as described in Materials and Methods. Unboiled fractions 3 to 20 were analyzed by SDS-PAGE on a 6% polyacrylamide gel, and YsaC multimers were detected by staining the top of the gel with silver. The same fractions were boiled and analyzed by SDS-PAGE on a 13.5% polyacrylamide gel followed by immunoblotting with anti-OmpA antiserum (outer membrane control), anti-FtsH antiserum (inner membrane control), and anti-PspB antiserum (additional inner membrane control that also served to show induction of the Psp response). The approximate locations of peak fractions for the inner (IM) and outer (OM) membranes, based on the fractionation of the control proteins, are indicated. For the multimer control, unboiled total membranes from strain AJD5487, with the araBp-ysaP plasmid and either a tacp-ysaC-his₆ plasmid (+) or the empty tacp vector (−), were separated by SDS-PAGE on a 6% polyacrylamide gel, and the top of the gel with stained with silver. This confirms that the high-molecular-weight silver-stained band corresponds to YsaC multimers.

YsaP is required for the function of the Ysa T3SS.

Our assignment of YsaP as the YsaC pilotin means that it should be required for the function of the endogenous Ysa T3SS. To test this, we monitored the export of Ysp proteins into the culture supernatant after growth at 26°C in medium containing 290 mM NaCl, conditions known to induce the Ysa T3SS (43). Proteins in the supernatant of the wild-type strain were identified as Ysa T3SS substrates (Ysp proteins) based on their absence from the supernatant of a ΔysaC negative-control strain (Fig. 5). Note that some proteins originally identified as Ysps were found to be Yop proteins encoded by the virulence plasmid (12). However, we used strains without the virulence plasmid, so these protein bands were absent from the supernatants in our analysis.

FIG 5.

YsaP is required for normal Ysa T3SS-dependent secretion. The ΔysaC (AJD5201), wild-type (WT; AJD3), and ΔysaP (AJD5487) strains contained the empty vector plasmid pBAD33 (−) or derivatives encoding wild-type YsaC, YsaC-His₆, YsaP, YsaP-T28D, or YsaP-FLAG. Strains were grown and culture supernatant proteins prepared as described in Materials and Methods. For each strain, the amount of sample loaded was derived from the equivalent of 150 μl of supernatant from a culture at an optical density (600 nm) of 1.0. The samples were separated in a 10% SDS-PAGE gel that was stained with silver. Asterisks indicate the most clearly Ysa T3SS-dependent proteins, identified by their complete absence in the negative control (ΔysaC mutant with pBAD33) compared to the wild type. Numbers to the left indicate the sizes of protein standards (lane M) in kilodaltons.

An in-frame deletion within ysaP reduced the secretion of all Ysp proteins, although small amounts were still detectable (Fig. 5). This defect was complemented by a plasmid encoding wild-type YsaP. Therefore, YsaP is required for the normal function of the Ysa T3SS, but the residual Ysp secretion detected in the ΔysaP mutant suggests that some T3SS function remains. This is consistent with our observation that small amounts of YsaC multimers could still localize into the OM when YsaP was absent (Fig. 4). The ΔysaP secretion defect was not complemented by the YsaP-T28D mutant, which suggests that the OM localization of YsaP is essential for its function (Fig. 5). It is also possible that the apparent instability of YsaP-T28D (Fig. 3) explains its failure to complement the defect. However, to attempt to compensate for this, the araBp-ysaP expression plasmids were induced with a 200-fold higher concentration of arabinose than that used for Fig. 3. Finally, we extended the Ysp secretion analysis to test the function of the His₆-tagged YsaC protein that was used throughout this study, as well as the YsaP-FLAG protein, which was used only to monitor YsaP localization (Fig. 3). Both epitope-tagged proteins were able to complement their respective null mutations.

YsaP is predicted to be a close structural homolog of the YscW/ExsB pilotin family.

Database searches did not identify any known pilotin as a close primary sequence homolog of YsaP (data not shown). Therefore, we used HHpred to search for YsaP homologs based on predicted secondary structure similarity (38). Remarkably, this suggested that YsaP has a secondary structure that is highly similar to those of two pilotins: YscW (probability = 99.9%; E value = 1e−27; P value = 5.2e−33) and ExsB (probability = 97.9%; E value = 3.4e−05; P value = 1.8e−10). These two pilotins are from the related Ysc and Psc T3SSs of Yersinia and P. aeruginosa, respectively, and the primary sequences of the mature YscW and ExsB proteins are 61% similar and 27% identical (data not shown, but see references 44 and 45). In contrast, the primary sequence of YsaP is not obviously related to that of YscW or ExsB. For example, pairwise alignments using LALIGN revealed only 43% similarity and 17% identity to YscW (data not shown) and only 47% similarity and 16% identity to ExsB (Fig. 6A). A 3D structural model of the mature YsaP protein was generated using the solved ExsB structure (44) as a template, as described in Materials and Methods. A side-by-side comparison of the two structures highlights their predicted similarity (Fig. 6B). When the ExsB structure was first reported, it was noted to be distinct from those of all other pilotins, including other T3SS pilotins (although structural similarity to YscW was predicted based on their sequence homology [44]). Our analysis suggests that YsaP might be a new member of this unique structural family of pilotins, even though the Ysa T3SS is not closely related to the Ysc and Psc T3SSs (6). Indeed, the YsaP primary sequence distinguishes it from other YscW/ExsB-like proteins (Fig. 6A; see Discussion).

FIG 6.

Sequence and structural comparisons between YsaP and the P. aeruginosa T3SS pilotin ExsB. (A) Alignment of the predicted mature YsaP and ExsB protein sequences by LALIGN, with similar (.) and identical (:) amino acids labeled. A pair of cysteines that are conserved in other ExsB homologs but not in YsaP are shown in red (see Discussion). (B) Comparison of the ExsB structure and the predicted YsaP structure (without their signal sequences). The ribbon model of ExsB shows the structure as viewed from published data, using the 3D viewer at http://pdb.rcsb.org (two sulfate ions and one nickel ion ligand are also shown). HHpred and MODELLER from the Bioinformatics Toolkit website (http://toolkit.tuebingen.mpg.de) were used to search for predicted structural homologues of YsaP and then to produce a YsaP structural model as described in Materials and Methods.

DISCUSSION

Secretins are an intriguing group of proteins with unique but incompletely understood properties (13). Among these are the processes involved in their assembly into OM multimers. Pilotins are accessory proteins that facilitate this process for some secretins (21). However, pilotins are similarly mysterious because they have divergent primary sequences and very different structures. Only a small number of pilotins have been identified, so discovering more will undoubtedly facilitate our understanding of common and distinct features in the ways they function. Here we describe the discovery of YsaP as a new member of the pilotin family. Although the primary sequence of YsaP is not highly similar to those of other pilotins, it is predicted to be a close structural homolog of the YscW/ExsB group.

Secretin multimers are toxic to a Δpsp strain when they mislocalize into the inner membrane (24, 25). Therefore, any protein that helps a secretin to localize correctly into the OM might be found in a screen for suppression of secretin toxicity in Δpsp cells. A pilotin is just such a protein, even though pilotin discovery was not our original motivation. Indeed, the other suppressor mutations we found are not specific to YsaC and will be described elsewhere (R. Rau, D. Srivastava, and A. J. Darwin, unpublished data). Nevertheless, an interesting consequence of our findings is that, in principle, a similar approach could be used to identify the pilotins for other secretins. The overproduction of any multimeric secretin, from many different species, is toxic in a Δpsp strain (25, 46, 47). Therefore, candidate pilot proteins, or even random plasmid expression libraries, could be tested for the ability to alleviate this secretin toxicity. Another intriguing possibility is that in any psp⁺ species, a psp gene expression reporter could be used to identify mutations that increase or decrease secretin multimerization and/or OM localization. This nonbiased approach could identify any trans-acting factors that affect secretin assembly.

The diversity among pilotins is striking. In a 2012 review article, pilotins were separated into three classes (21). Class 1 pilotins are comprised entirely of α-helical tetratricopeptide repeats (TPRs) and are approximately twice as large as other pilotins, class 2 pilotins are comprised predominantly of β-strands, and class 3 pilotins are predominantly α-helical non-TPR proteins. However, we now know that these three classes do not tell the full story. Shigella flexneri MxiM and Y. enterocolitica YscW were both assigned to the β-strand-rich class 2 pilotins (21). However, when the structure of the YscW homolog ExsB was solved, it was distinct from that of MxiM (44). Although both have stacked antiparallel β-sheets, MxiM has two features not found in ExsB: an elongated α-helix and a linear hydrophobic cavity at its core that binds lipids (48). Indeed, among known structures, that of ExsB is unique compared to all other OM lipoproteins involved in the assembly of secretion systems or pili (44). Therefore, the YscW/ExsB family defines a unique structural class of pilotins, and our analysis suggests that YsaP is likely to be a new member of this class (Fig. 6).

Despite the predicted structural similarity between YsaP, YscW, and ExsB, YsaP is more divergent from YscW and ExsB than they are from each other. The Ysc and Psc T3SSs are closely related, and their YscW and ExsB pilotins share obvious sequence similarity (44, 45). In contrast, the Ysa T3SS is related to the Mxi T3SS of Shigella and the Inv SPI-1 T3SS of Salmonella enterica (6). However, the MxiM and InvH pilotins are not YscW/ExsB family members (21). In addition, the primary sequence of YsaP is much less similar to those of YscW and ExsB than they are to each other (Fig. 6A and data not shown). YsaP also lacks a pair of C-terminal cysteines that form a disulfide bond in the ExsB structure and are conserved in all ExsB-like proteins from organisms with Ysc T3SS-related systems (44) (Fig. 6A). The conservation of these cysteines hinted at a conserved but unknown role. However, their absence from YsaP argues against this. Also, when the two cysteines in YscW were mutated to serine, there was no Ysc T3SS-related phenotype (45).

A functional similarity between YscW, ExsB, and YsaP is supported by the fact that all three have exactly the same effect on their secretins. Without the pilotin, the cognate secretin still forms multimers, but the multimers mislocalize to the inner membrane (27, 49; our data presented here). In addition, YscW and ExsB are required for the stability of their secretins, and the lower steady-state level of YsaC in the absence of YsaP suggests that the same is true for YsaP. Therefore, all three of these pilotins probably function similarly. Studies of MxiM led to the piloting model to explain how pilotins might traffic secretins to the OM (23). In this model, after a pilotin is engaged by the Lol system, it forms a complex with its secretin so that Lol transports the pilotin-secretin complex to the OM. An alternative is the docking model, in which the pilotin is targeted to the OM alone and acts as a receptor for the secretin (50). As with other pilotins, the OM location of YsaP appears to be important for its function (Fig. 5), but this does not provide insight into either model. However, the failure to detect ExsB or YscW in complex with their secretins suggests a transient interaction, which has been proposed to fit better with the piloting model (27, 49). Interestingly, even in the absence of the ExsB, YscW, or YsaP pilotin, some of their secretin multimers can still make it into the OM fraction (27, 49) (Fig. 4). Furthermore, some T3SS function remains in both exsB and ysaP null strains (49) (Fig. 5). This might suggest the involvement of other proteins or an inherent ability of the secretins to localize correctly, albeit inefficiently, into the OM. Finally, two observations suggest that YsaP might not directly catalyze secretin multimerization. First, YsaC multimers can still form in the absence of YsaP (Fig. 4). Second, when the YsaC fractionation experiment (Fig. 4) was done in the presence of the YsaP-T28D mutant, the amounts and localization of YsaC multimers were identical to those observed when YsaP was absent (data not shown). Therefore, tethering of YsaP-T28D to the inner membrane did not promote increased YsaC multimer formation at that location.

In summary, we have identified the pilotin component of the Y. enterocolitica Ysa T3SS. This is only the second pilotin to be described for Yersinia, but remarkably, it might be a close structural homolog of the other, YscW. In fact, YsaP is predicted to be the first member of the YscW/ExsB pilotin family that is not part of a Ysc-related T3SS. Its sequence divergence from the other YscW/ExsB family members could make it a valuable tool in future studies aimed at understanding key functional features within this family. Of course, much more work will be needed to understand how this family of pilotins functions and whether they do so similarly to or differently from other pilotins that do not share their structure. Finally, we note that the secretin-Psp relationship could be exploited to help discover other pilotins, and perhaps other factors involved in the assembly of various secretin-containing systems.