JOURNAL OF BACTERIOLOGY, July 2010, p. 3491–3502 Vol. 192, No. 130021-9193/10/$12.00 doi:10.1128/JB.01493-09Copyright © 2010, American Society for Microbiology. All Rights Reserved.

Vp1659 Is a Vibrio parahaemolyticus Type III Secretion System 1Protein That Contributes to Translocation of Effector Proteins

Needed To Induce Cytolysis, Autophagy, and Disruption ofActin Structure in HeLa Cells�

Xiaohui Zhou,1 Michael E. Konkel,2 and Douglas R. Call1*Department of Veterinary Microbiology and Pathology1 and School of Molecular Biosciences,2

Washington State University, Pullman, Washington

Received 13 November 2009/Accepted 12 April 2010

Vibrio parahaemolyticus harbors two type III secretion systems (T3SSs; T3SS1 and T3SS2), of which T3SS1is involved in host cell cytotoxicity. T3SS1 expression is positively regulated by ExsA, and it is negativelyregulated by ExsD. We compared the secretion profiles of a wild-type strain (NY-4) of V. parahaemolyticus withthose of an ExsD deletion mutant (�exsD) and with a strain of NY-4 that overexpresses T3SS1 (NY-4:pexsA).From this comparison, we detected a previously uncharacterized protein, Vp1659, which shares some sequencehomology with LcrV from Yersinia. We show that vp1659 expression is positively regulated by ExsA and isnegatively regulated by ExsD. Vp1659 is specifically secreted by T3SS1 of V. parahaemolyticus, and Vp1659 isnot required for the successful extracellular secretion of another T3SS1 protein, Vp1656. Mechanical frac-tionation showed that Vp1659 is translocated into HeLa cells in a T3SS1-dependent manner and that deletionof Vp1659 does not prevent VopS from being translocated into HeLa cells during infection. Deletion of vp1659significantly reduces cytotoxicity when HeLa cells are infected by V. parahaemolyticus, while complementationof the �vp1659 strain restores cytotoxicity. Differential staining showed that Vp1659 is required to inducemembrane permeability in HeLa cells. We also show evidence that Vp1659 is required for actin rearrangementand the induction of autophagy. On the basis of these data, we conclude that Vp1659 is a T3SS1-associatedprotein that is a component of the secretion apparatus and that it is necessary for the efficient translocationof effector proteins into epithelial cells.

As a marine pathogen, Vibrio parahaemolyticus is frequentlyisolated from seafood products such as oysters and shrimp (19,45). The main symptoms of V. parahaemolyticus infection inhumans include diarrhea, nausea, and vomiting. In addition tothe gastrointestinal infection, necrotizing fasciitis and septicshock are reportedly associated with V. parahaemolyticus in-fection (37). V. parahaemolyticus can also cause wound infec-tions after contact with contaminated water (6, 7, 16, 37).

V. parahaemolyticus is able to adhere to and invade epithelialcells (1, 38, 43). Pili are involved in the adherence to theintestinal epithelium (32), but it is not clear what factors arerequired for V. parahaemolyticus to invade epithelial cells. He-molysins are considered primary factors involved in the patho-genesis of V. parahaemolyticus. For example, a thermostabledirect hemolysin (tdh) mutant strain loses the ability to causefluid accumulation in the intestinal lumen (33), while deletionof a tdh-related gene (trh) results in the complete loss ofhemolysis and the partial loss of fluid accumulation in a rabbitintestinal ligation model (42). Recent studies show that thedisruption of epithelial tight junctions, which is a hallmark ofbacterial dissemination into the circulatory system and subse-quent septicemia, is independent of the thermostable direct

hemolysin, suggesting that additional factors are required forthe pathogenesis of V. parahaemolyticus (27).

A broad range of Gram-negative bacteria employ type III se-cretion systems (T3SSs) to export virulence-related proteins intothe extracellular milieu and/or to deliver these proteins directlyinto host cells (5, 12, 13). T3SSs are composed of three parts: asecretion apparatus, translocators, and effectors (17, 18). Thesecretion apparatus and translocators are encoded by ca. 25 genesthat are conserved and usually located in a genomic island. Genesthat encode effectors are less conserved and can be found distalfrom the T3SS islands. The secretion apparatus serves to secreteboth effectors and translocators from bacterial cells, and translo-cators help the effectors cross into the eukaryotic cells, where theycan disrupt normal host cell signal functions.

Two distinct T3SSs (T3SS1 and T3SS2) were identified inthe genome of V. parahaemolyticus (28). On the basis of thesequence similarity and gene organization, T3SS1 was classi-fied as a member of the Ysc family of secretion systems, whileT3SS2 was classified as a member of the Inv-Mxi-Spa family(40). Functional analysis shows that deletion of T3SS1 de-creases cytotoxicity against HeLa cells, while deletion of T3SS2diminishes intestinal fluid accumulation (35). Interestingly, insome strains, T3SS2 can be involved in the cytotoxic effectspecifically against Caco-2 and HCT-8 cells (23). One studyshowed that T3SS1 of V. parahaemolyticus induces autophagy,but blocking autophagy does not completely mitigate cytotox-icity, indicating that other T3SS1-induced mechanisms contrib-ute to cell death (3, 4). Recent work from our laboratory

* Corresponding author. Mailing address: Department of Veteri-nary Microbiology and Pathology, Washington State University, 402Bustad Hall, Pullman, WA 99164-7040. Phone: (509) 335-6313. Fax:(509) 335-8529. E-mail: drcall@wsu.edu.

� Published ahead of print on 23 April 2010.

3491

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

showed that V. parahaemolyticus induces cell rounding, poreformation, and membrane damage, consistent with the induc-tion of an oncosis pathway (46). Importantly, treatment ofinfected cells with an osmoprotectant (polyethylene glycol3350) significantly reduced cytotoxicity, indicating that oncosisis the primary mechanism by which T3SS1 of V. parahaemo-lyticus causes cell death for in vitro cultures (46). Nevertheless,it is unknown which effector protein(s) is involved in cell cy-totoxicity. By comparing the secretion protein profiles of wild-type and T3SS1 mutant strains, four T3SS1 proteins have beenidentified (34). Among these, Vp1680 is translocated into hostcells and is required for the induction of autophagy duringinfection of HeLa cells (3, 34). Recent studies showed thatVopS is able to prevent the interaction of Rho GTPase with itsdownstream factors by a new modification mechanism, calledAMPylation (44), and this prevents the assembly of actin fi-bers. Two proteins (VopT and VopL) have been identified asT3SS2 substrates (23, 26). VopT is a member of ADP-ribosyl-transferase and is partially responsible for the cytotoxic effectspecific to Caco-2 and HCT-8 cells (23). VopL induces theassembly of actin stress fibers (26) and is potentially responsi-ble for the internalization of V. parahaemolyticus into Caco-2cells (1). Many other potential effector proteins are encodedproximal to T3SS1 and T3SS2 apparatus genes, but these havenot been functionally characterized. The function of structuralgenes has not been extensively studied for either T3SS1 orT3SS2 in V. parahaemolyticus.

T3SSs are expressed after contact with host cells or whencells are grown under inducing conditions (17). Expression ofT3SS1 in V. parahaemolyticus is induced when bacteria aregrown in tissue culture medium (Dulbecco’s minimal essentialmedium [DMEM]), although the secretion of one substrate(Vp1656) was not detected under this condition, probably dueto the low detection sensitivity (47). T3SS1 genes are not ex-pressed when bacteria are grown in LB medium supplementedwith 2.5% NaCl (LB-S). Disruption of the exsD gene or over-expression of exsA results in the constitutive expression ofT3SS1 genes and the secretion of Vp1656 even when bacteriaare grown in LB-S (47). For the present study, we took advan-tage of these regulatory mechanisms and compared the pro-teins secreted by the NY-4 (wild type), �exsD, �exsD::pexsD(exsD complement), and NY-4:pexsA strains. We identified twoproteins (VopS and Vp1659) that are present in the superna-tants of the �exsD and NY-4:pexsA strains but that are absentin the supernatants of the NY-4 and �exsD::pexsD strains.Herein we demonstrate that Vp1659 is secreted into the ex-tracellular milieu and is translocated into HeLa cells by T3SS1.Functional analysis is consistent with the hypothesis thatVp1659 plays a role in actin rearrangement and induction ofcytotoxicity and autophagy.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions. All V. parahaemolyticusstrains used in this study were derived from wild-type strain NY-4 (Table 1) (47).

TABLE 1. Strains and plasmids used in this study

Strain, genotype, or plasmid Description Referenceor source

E. coliS17 �pir thi pro hsdR hsdM� recA RP4-2-Tc::Mu-Km::Tn7 �pir 29S17-pMMB207-vp1659 S17 carrying pMMB207-vp1659 This studyS17-pDM4-vp1659-1F�2R S17 carrying pDM4-vp1659-1F�2R This study

Vibrio parahaemolyticusNY-4 Clinical isolate O3:K6�vp1659 vp1659 deletion mutant This study�vp1659::pvp1659 �vp1659 complemented with vp1659 gene located in plasmid pMMB207 This studyNY-4:pexsA Wild-type strain containing exsA gene located in plasmid pMMB207 47�vcrD::pexsA T3SS1 mutant strain containing exsA gene located in plasmid pMMB207 46

�escV::pexsA T3SS2 mutant strain containing exsA gene located in plasmid pMMB207 46�vcrD escV::pexsA T3SS1 and T3SS2 double mutant strain containing exsA gene located in plasmid pMMB207 46�exsA exsA (vp1699) deletion mutant 47�exsA::pexsA �exsA complemented with exsA gene located in plasmid pMMB207 47�exsD exsD (vp1698) deletion mutant 47�exsD::pexsD �exsD complemented with exsD gene located in plasmid pMMB207 47NY-4:pexsD Wild-type strain containing exsD gene located in plasmid pMMB207 47�vp1659::pexsA vp1659 deletion mutant strain containing exsA gene located in plasmid pMMB207 This study�exsD vp1659 exsD and vp1659 double-deletion mutant This study�exsD vp1656 exsD deletion and vp1656 insertional mutant 47

Yersinia enterocolitica JB580v Serogroup O:8, Nalr �yenR(r� m�) 41

PlasmidspMMB207 RSF1010 derivative, IncQ lacIq Cmr Ptac oriT 30apDM4 A suicide vector with ori R6K sacB Cmr 30pMMB207-RBS-vp1659 vp1659-coding sequences and sequences for 6� His amino acids at the C terminus cloned

into pMMB207This study

pDM4-1659-1F�2R Flanking region sequences of vp1659 cloned into pDM4 This studypMMB207-pexsA exsA-coding sequences and sequences for 6� His amino acids at the C terminus cloned

into pMMB20747

3492 ZHOU ET AL. J. BACTERIOL.

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

V. parahaemolyticus was grown in LB broth or LB agar supplemented with 2.5%NaCl to stationary phase at 37°C while the culture was shaken (200 rpm). Genedeletion experiments were performed by using Escherichia coli S17 �pir culturedin LB medium. Plasmid pMMB207 (Cmr) was used in complementation andprotein expression experiments, and plasmid pDM4 (Cmr) was used for genedeletion experiments. When appropriate, the following antibiotics were added atthe indicated concentrations: ampicillin, 100 �g ml�1, and chloramphenicol, 25�g ml�1 for E. coli and 5 �g ml�1 for V. parahaemolyticus.

Cell line. HeLa (CCL-2) cells were purchased from the American Type Cul-ture Collection (Manassas, VA). Monolayers of HeLa cells were maintained inDMEM (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum(FBS) at 37°C. Passage of HeLa cells was carried out every 1 to 2 days, and thecells were washed once before inoculation.

Construction of deletion mutants. The complete open reading frame (ORF) ofvp1659 was deleted from the chromosome by allelic exchange using a suicidevector, pDM4, carrying flanking regions on both sides of vp1659 (30). Primerpairs (primers 1659-1F and 1659-1R and primers 1659-2F and 1659-2R) (Table2) were used to amplify two DNA fragments corresponding to the flankingregions of vp1659 with V. parahaemolyticus strain NY-4 chromosomal DNA asthe template. Fragment 1 (approximately 1,000 bp upstream of the start codon ofvp1659) was amplified with primers 1659-1F and 1659-1R and digested with XhoIand XbaI. Fragment 2 (approximately 1,000 bp downstream of the stop codon ofvp1659) was amplified with primers 1659-2F and 1659-2R and digested with BglIIand XbaI. After digestion, these two DNA fragments were simultaneously li-gated into the pDM4 suicide vector that was predigested with XhoI and BglII.The resultant plasmid was designated pDM4-1659-1F�2R. This plasmid wastransformed into E. coli S17 �pir by electroporation, resulting in strain S17-pDM4-1659-1F�2R. This strain was grown overnight in the presence of chlor-amphenicol, and an aliquot (1 ml) was mixed with an overnight culture of theNY-4 strain (1 ml) for �12 h of conjugation at 37°C. Transconjugants wereselected from the LB-S agar plates containing both ampicillin and chloramphen-icol. We used ampicillin to select against E. coli because the NY-4 strain isnaturally resistant to ampicillin. Chloramphenicol was used to select againstnontransconjugants. Once pDM4-1659-1F�2R was conjugated from E. coli to V.parahaemolyticus, it integrated into the chromosome of V. parahaemolyticus. Theintegrated suicide plasmid was excised from the chromosome by growth on LB-Scontaining 5% sucrose. Once the allelic exchange was completed, clones thatwere sensitive to chloramphenicol were picked and screened by PCR with prim-ers vp1659_Forward and vp1659_Reverse, which lie adjacent to the vp1659sequence. One clone with the vp1659 gene deletion was selected and was desig-nated �vp1659.

Complementation. The complete sequence of vp1659 with a 6� His tag addedat the 3� terminus of the gene was amplified with primers vp1659-up and vp1659-down (Table 2). The amplified PCR product was purified before digestion withEcoRI and XbaI. The digested PCR product was ligated into plasmid pMMB207(digested with the same enzymes), resulting in plasmid pMMB207-RBS-vp1659.This plasmid was transformed into E. coli S17 �pir by electroporation, resultingin strain S17-pMMB207-RBS-vp1659, and was then conjugated from E. coli S17�pir into the �vp1659 strain, resulting in the �vp1659::pvp1659 strain.

Production of Vp1659 and generation of polyclonal antibody. Strain S17-pMMB207-RBS-vp1659 was transformed into Yersinia enterocolitica (strainJB580v) by conjugation for protein expression. We produced Vp1659 in Y.enterocolitica because a previous study showed that Y. enterocolitica is more likelyto yield soluble recombinant proteins (47). Y. enterocolitica containingpMMB207-RBS-vp1659 was grown at 26°C, and an aliquot of the overnightculture (5 ml) was diluted into 500 ml LB medium. After incubation for 2 h withshaking at 26°C, 1 mM isopropyl -D-thiogalactopyranoside (IPTG) was added.Expression of Vp1659 was induced for 12 h at 26°C with vigorous shaking, beforethe cells were pelleted. The pellets were resuspended in 15 ml binding buffer (50mM NaH2PO4, pH 8.0, 0.5 M NaCl, 10 mM imidazole) and sonicated for 15 min.After sonication, the whole-cell lysate was centrifuged at �3,000 � g for 10 minto remove the insoluble fraction and the supernatant containing soluble proteinswas loaded onto a Ni2� resin column (Invitrogen). After binding of the proteinsfor 1 h, the column was washed 15 times with washing buffer (50 mM NaH2PO4,pH 8.0, 0.5 M NaCl, 50 mM imidazole). The proteins were eluted with elutionbuffer (50 mM NaH2PO4, pH 8.0, 0.5 M NaCl, 300 mM imidazole). Proteinsamples from the elution were electrophoresed by SDS-PAGE and stained withCoomassie blue to check the purity of the protein. Purified Vp1659 was submit-ted to the Monoclonal Antibody Center at Washington State University (Pull-man, WA) for production of polyclonal antibody. The production of the anti-DnaK (Vp0653) polyclonal antibody will be described elsewhere. For antibodyabsorption, the whole-cell lysate of the �vp1659 overnight culture was mixed withthe anti-Vp1659 polyclonal antibody (1:1) for 12 h. After centrifugation at15,000 � g for 30 min, the supernatant was used for Western blot analyses.Rabbit anti-VopS polyclonal antibody was generously provided by Kim Orth(Southwestern Medical Center, University of Texas, Dallas, TX).

Preparation of proteins and Western blot analysis. To collect proteins in thesupernatant, an overnight culture of V. parahaemolyticus was diluted into LB-S(1:100) and antibiotics were added as necessary. Diluted samples were grown for4 h at 37°C with shaking (200 � g). IPTG (1 mM) was added after 1 h of shakingto induce protein expression. The bacterial culture (15 ml) was centrifuged at�3,000 � g for 15 min, and the supernatant from each sample was passedthrough a 0.22-�m-pore-size syringe filter to exclude any remaining bacteria. Toconcentrate the proteins in the supernatant, a final concentration of 10% (vol/vol) trichloroacetic acid was added to the supernatant, and the mixture wasincubated on ice for 2 h. After centrifugation at �12,000 � g and 4°C for 20 min,the pellets were resuspended in ice-cold acetone and the suspension was centri-fuged at �12,000 � g and 4°C for 20 min. Finally, the protein pellets wereresuspended in 50 �l 1� loading buffer (40 mM Tris-HCl, pH 8.0, 5% mercap-toethanol, 1% SDS, 5% glycerin, 1% bromphenol blue). After separation of thesupernatant proteins by SDS-PAGE, silver staining was performed by using aSilverSNAP stain kit (Pierce, Rockford, IL) to visualize the protein secretionprofile.

To collect proteins from whole-cell lysates for the detection of Vp1659, anovernight culture was diluted into LB-S (1:100) or DMEM (1:10) and grown for4 h with addition of antibiotics and IPTG as needed. The bacterial cultures (15ml) were centrifuged at �3,000 � g for 20 min, and the pellets were resuspendedin 500 �l phosphate-buffered saline (PBS). The bacterial suspension was soni-cated until it became clear and then an equal volume of 2� loading buffer wasadded. To collect the bacterial proteins after infection, an overnight bacterialculture was rinsed once with DMEM and added to six-well polystyrene cultureplates inoculated with �106 HeLa cells to achieve a multiplicity of infection(MOI) of 100 CFU per cell. Synchronization of the bacterium-host cell contactwas facilitated by centrifuging the plates at 600 � g for 5 min. The infected cellswere scraped from each well after 4 h of infection and resuspended in 200 �lPBS. The resuspended bacterial HeLa cell mixture was sonicated, and an equalvolume of 2� loading buffer was added to each sample for SDS-PAGE. All thesamples were boiled for 5 min at 100°C and loaded on a 12% (wt/vol) poly-acrylamide gel. After electrophoresis, the proteins were transferred to a nitro-cellulose membrane for 12 h using an electroblot apparatus (Bio-Rad, Hercules,CA). Skim milk (5%) in PBS containing 0.1% Tween 20 was used to block themembrane. After 1 h of blocking, the membrane was probed (1:1,000) withpolyclonal anti-Vp1659 (this study), anti-Vp1656 (47), or anti-DnaK (describedelsewhere) for 1 h at room temperature. As a secondary antibody, an anti-mouseimmunoglobulin conjugated to horseradish peroxidase was diluted in PBS with5% milk at 1:5,000. The membrane was developed by using a Western blottingkit, according to the manufacturer’s instructions (Bio-Rad).

Infection and LDH assay. An overnight culture of V. parahaemolyticus wasobtained as described earlier and pelleted by centrifugation at 4,000 � g at roomtemperature. The bacterial pellets were resuspended in DMEM containing 1%(vol/vol) FBS. Bacterial suspensions (10 �l) were inoculated into each well of a12-well polystyrene plate containing �106 HeLa cells/well to achieve an MOI of

TABLE 2. Primers used in this study

Primer Sequence

Vp1659-1F ................................AGGATAAACTCGAGTTTCTGAGCCTCTTTCGCAG

Vp1659-1R................................AGTTAGTCTAGA GGA TAA CCAACC ATG ACT AAA AC

Vp1659-2F ................................AGTTAGTCTAGA TCACCATACCGCGC

Vp1659-2R................................AGTTAGAGATCT CAA TTT GTTAGC GAC GTA AAA C

vp1659_Forward.......................TCT TCC AGA AAA GAC AGC AGvp1659_Reverse........................AAG AAC TCA TCG AAG AAG

TTCVp1659-up ................................AGGATAGAATTCTAAGGAGGTA

GGATAATATTGGATATGACGGATATGACAAC

Vp1659-down............................AGTTAGTCTAGATTAATGGTGATGGTGATGGTGAATGGCTCGTAGGATTTCTTG

VOL. 192, 2010 T3SS1 PROTEIN Vp1659 3493

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

�100 CFU per cell. For the lactate dehydrogenase (LDH) release assay, themedium was free of antibiotics before inoculation. The supernatant of HeLa cellsafter infection was collected at 1, 2, 3, and 4 h; and LDH activity was measuredwith a cytotoxicity detection kit (Promega, Madison, WI), according to themanufacturer’s instructions. The lysis buffer provided in the kit was added toachieve maximum LDH release. The LDH activity in the supernatant of unin-fected cells was also measured to obtain a spontaneous LDH value. Percentcytotoxicity was calculated with the following formula: (test LDH release �spontaneous release)/(maximal release � spontaneous release).

Detection of membrane permeabilization. Pore formation was analyzed byfluorescence microscopy, as described previously (21, 46). Briefly, monolayers ofHeLa cells were grown on coverslips. Each strain of V. parahaemolyticus wasinoculated into the appropriate wells, and after 2 h of infection, the coverslipswere dipped in PBS containing 25 �g ml�1 ethidium bromide (EtBr) and 5 �gml�1 acridine orange (AO) for 5 min. The coverslips were rinsed once andobserved under a Zeiss confocal microscope (Franceschi Microscopy and Imag-ing Center, Pullman, WA) using fluorescein isothiocyanate (543 nm) and rho-damine (488 nm) filters. The percentage of EtBr-positive cell was obtained byaveraging the data from three fields of HeLa cells (100 cells were counted foreach field) infected with different strains.

Translocation of Vp1659 and VopS into HeLa cells after infection. Mechanicalfractionation of infected HeLa cells was performed as described previously (14,15, 22). Briefly, HeLa cells were seeded in a six-well plate and allowed to growto confluence. One plate of HeLa cells was infected with each V. parahaemolyti-cus strain at an MOI of 100 for 2 h. The medium was aspirated and centrifugedat 4,000 � g for 10 min. The bacterial pellets were resuspended in 100 �l PBS,and the suspension was sonicated for 1 min and was used as the bacterial pelletfraction in the Western blot analysis. The infected cells as well as the adherentbacteria were scraped off the six-well plate and resuspended in 400 �l PBS. Thecells were then passed through a 27-gauge needle six to eight times, which wasexpected to lyse the HeLa cells and leave the bacteria intact. The homogenatewas centrifuged at 15,000 � g for 15 min, and the supernatant was used as thehost-cell membrane and cytoplasm fraction in the Western blot analysis. Westernblot analysis was performed by using mouse anti-Vp1659, mouse anti-Vp1656,and rabbit anti-VopS polyclonal antibodies, followed by Immun-Star horseradishperoxidase (HRP) chemiluminescence detection. To separate the cytosol andmembrane fraction of HeLa cells infected with the NY-4 strain, the HeLa cellslysed with a needle as described above were centrifuged to remove the intactbacteria and HeLa cell debris. The supernatant was further centrifuged at30,000 � g for 1 h to obtain the cytosol (supernatant) and membrane (pellet).Both the cytosol and the membrane fractions were probed with anti-VopS,anti-Vp1659, antitubulin (Sigma, St. Louis, MO), and anticalnexin (Santa CruzBiotechnology, Santa Cruz, CA) antibodies.

Actin staining and detection of LC3 conversion. Uninfected or infected HeLacells were washed with PBS two times and fixed with 4% paraformaldehyde for10 min. The HeLa cells were then blocked in 3% bovine serum albumin for 20min before they were stained with rhodamine-conjugated phalloidin, accordingto the manufacturer’s instruction (Invitrogen). The HeLa cells, on coverslips,were observed with a Zeiss confocal microscope using rhodamine (488 nm)filters. To detect a mass change in microtubule-associated protein light chain 3(LC3), which is indicative of autophagy, HeLa cells were infected with the NY-4,�vp1659, or �vp1659::pvp1659 strain for 6 h and the whole-cell lysate sampleswere collected. Western blot analysis was performed using rabbit anti-LC3(Novus Biologicals, Littleton, CO) and antiactin (Seven Hills Bioreagents, Cin-cinnati, OH) antibodies, followed by Immun-Star HRP chemiluminescence de-tection.

Statistical analysis. Student’s t test was used to compare the cytotoxicity andpercentage of EtBr-positive cells induced by the NY-4 and �vp1659::pvp1659strains and those induced by the �vp1659 strain. The data were the averages ofthree independent experiments with two replicates per experiment. A P value of0.05 was considered a statistically significant difference.

RESULTS

Identification of Vp1659 and VopS as secreted proteins. Aprevious study showed that the T3SS1 genes in V. parahaemo-lyticus NY-4 are not expressed when the bacteria are grown inLB-S (47). The T3SS1 genes are expressed, however, wheneither the exsD gene is disrupted or the exsA gene is expressedin trans in the NY-4 strain while it is cultured in LB-S. Wecompared the proteins in the supernatant of the NY-4 (wild

type), �exsD, �exsD::pexsD, and NY-4:pexsA strains of V. para-haemolyticus growing in LB-S. The supernatant proteins wereseparated by SDS-gel electrophoresis and silver stained. Twoproteins were present in the supernatant of the �exsD andNY-4:pexsA strains but were not visible in the supernatant ofthe NY-4 and �exsD::pexsD strains (Fig. 1A). These two pro-tein bands were excised from the gel and submitted for proteinidentification by matrix-assisted laser desorption ionizationtandem mass spectrometry (MALDI MS/MS; Laboratory forBiotechnology and Bioanalysis, Washington State University,Pullman, WA). Approximately 150 amino acid residues weredetected for the smaller protein band, and this sequence cor-responded to 39% of the deduced composition of VopS(Vp1686). Approximately 163 amino acid residues were de-tected for the larger protein, and this sequence correspondedto 27% of the deduced composition of Vp1659. To confirm ifthe larger protein band was Vp1659, we generated recombi-nant Vp1659 protein and polyclonal antibody. Western blotresults showed the presence of a clear band (67 kDa) for thesecreted proteins originating from the �exsD and NY-4:pexsAstrains, but no band was evident for the NY-4 or �exsD::pexsDstrain (Fig. 1B). The results of this experiment confirmed pre-

FIG. 1. (A) Protein samples from the supernatants of wild-typestrain NY-4 and the �exsD, �exsD::pexsD, and NY-4:pexsA strainswere separated by SDS-PAGE. Two protein bands that were present inthe �exsD and NY-4:pexsA strains but absent in NY-4 and the�exsD::pexsD strain were identified as Vp1659 and VopS (Vp1686) byMALDI MS/MS analysis. (B) Protein samples from the supernatantsof the NY-4, �exsD, �exsD::pexsD, and NY-4:pexsA strains wereprobed with anti-Vp1659 polyclonal antibody (Western blot assay) toverify the presence of Vp1659 in the supernatants of the �exsD andNY-4:pexsA strains.

3494 ZHOU ET AL. J. BACTERIOL.

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

vious reports that VopS is a secreted protein (2, 34) and iden-tified a novel secreted protein, Vp1659.

Expression of Vp1659 in V. parahaemolyticus is initiated bygrowth in DMEM and in contact with HeLa cells. To deter-mine if the expression of vp1659 is influenced by culture con-ditions, strain NY-4 was grown in LB-S or DMEM or in con-tact with HeLa cells, and Vp1659 was detected in whole-celllysates using a Western blot assay. The expression of vp1659was undetectable when the NY-4 strain was grown in LB-S(Fig. 2A, upper panel, lane 1). When NY-4 was grown inDMEM, a protein band (Fig. 2A, upper panel, lane 2) with a

similar size to that of the purified Vp1659 (Fig. 2A, upperpanel, lane 4) was detected by polyclonal anti-Vp1659 anti-body. For protein samples from NY-4 after contact with HeLacells for 4 h, a similarly sized protein band (Fig. 2A, upperpanel, lane 3) was detected by polyclonal anti-Vp1659 anti-body. More importantly, the band from NY-4 detected aftercontact with HeLa cells was more intense than that detectedfrom NY-4 grown in DMEM alone (compare lane 2 and lane3, Fig. 2A, upper panel). Because contact with HeLa cells wasconfounded by the presence of DMEM, we speculated thatcontact with HeLa cells had an additional role in facilitatingthe expression of vp1659, thereby producing a more intenseband, and this is consistent with the findings of a previouslyreported promoter fusion assay (47). The V. parahaemolyticusDnaK chaperone protein was detected by anti-DnaK poly-clonal antibody and served as a loading control (Fig. 2A, lowerpanel). To confirm that the growth conditions employed in ourstudy did not affect the expression of DnaK, we loaded theprotein samples extracted from similar numbers of CFU ofbacteria grown under different conditions and probed withanti-DnaK polyclonal antibody. The results showed that a pro-tein band with a similar intensity was detected in the wild-typestrain grown in LB-S, in DMEM, or in contact with HeLa cells(Fig. 2B). In addition, the �exsA and �exsD strains yieldedprotein bands of similar intensity (Fig. 2B).

Vp1659 is positively regulated by ExsA. ExsA is an AraC-like transcriptional factor that controls the transcription ofT3SS1 genes in V. parahaemolyticus (47). Vp1659 was not de-tected in whole-cell lysates when the �exsA strain was grown in

FIG. 3. (A) Western blot of whole-cell lysates from the NY-4,�exsA, �exsA::pexsA, and NY-4:pexsA strains grown in LB-S (lanes2, 4, and 6) or in DMEM (lanes 1, 3, 5, and 7); (B) Western blot ofwhole-cell lysates from the �exsD, �exsD::pexsD, and NY-4:pexsDstrains grown in LB-S (lanes 1, 3, and 5) or in DMEM (lanes 2, 4,and 6).

FIG. 2. (A) Western blot of (Vp1659 antibody) whole-cell lysatesfor the NY-4 strain after growth in LB-S (lane 1), after growth inDMEM (lane 2), and after contact with HeLa cells (lane 3) (all con-ditions were normalized to CFU counts of 5 � 107). Purified Vp1659(lane 4) was included as a positive control (upper panel). This Westernblot shows the specificity of the absorbed antisera, and DnaK served asa loading control in these experiments (lower panel). (B) Western blotshowing DnaK from the lysate isolated from NY-4 following growth inLB-S (lane 1), in DMEM (lane 2), or in contact with HeLa cells (lane3) for 4 h. The �exsA and �exsD strains (lanes 4 and 5) were grown inLB-S for 4 h. The number of bacterial CFU was determined to ensurethat each lane was loaded with protein samples extracted from similarnumbers of CFU.

VOL. 192, 2010 T3SS1 PROTEIN Vp1659 3495

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

LB-S or DMEM (Fig. 3A, lanes 2 and 3). In trans expression ofexsA in the �exsA strain restored the synthesis of Vp1659 whenV. parahaemolyticus was grown in LB-S or DMEM (Fig. 3A,lanes 4 and 5). Furthermore, the in trans expression of exsA inNY-4 activated the synthesis of Vp1659 when V. parahaemo-lyticus was grown in LB-S or DMEM (Fig. 3A, lanes 6 and 7).Qualitatively, it was also clear that overexpression of exsAproduced more protein than the amount produced by the wild-type strain under inducing conditions (Fig. 3B, lane 1). Theseresults indicate that ExsA is necessary for the expression ofvp1659.

Synthesis of Vp1659 is negatively regulated by ExsD. ExsDis a negative regulator for T3SS1 (47), and therefore, we de-termined if expression of Vp1659 is also affected by ExsD.Vp1659 was synthesized in the �exsD strain when V. parahae-molyticus was grown under noninducing conditions (in LB-S)or inducing conditions (in DMEM) (Fig. 3B, lanes 1 and 2),indicating that ExsD negatively affects the expression ofVp1659. Complementation of the �exsD strain with a wild-typeexsD gene in trans prevented vp1659 expression when the bac-teria were grown in LB-S or DMEM (Fig. 3B, lanes 3 and 4).When NY-4 was provided with an exsD gene in trans, vp1659expression was blocked when the bacteria were grown in LB-Sor DMEM (Fig. 3B, lanes 5 and 6). Thus, overexpression ofexsD represses the expression of vp1659 and inducing growthconditions do not override the inhibition from the in transexpression of exsD.

Vp1659 is secreted by T3SS1 but not by T3SS2. To furtherverify that Vp1659 is secreted by T3SSs in V. parahaemolyticus,we examined the supernatant of the NY-4, �vcrD (T3SS1 neg-ative [T3SS1�]), �escV (T3SS2 negative [T3SS2�]), and �vcrDescV (T3SS1� T3SS2�) strains. To induce the expression ofT3SS1 genes, NY-4 and each mutant strain were transformedwith an exsA plasmid for the in trans expression of this gene.Western blot analysis showed that Vp1659 is synthesized in allthe strains (Fig. 4, upper panel), while Vp1659 is present onlyin the supernatant of the NY-4:pexsA and �escV::pexsA strains

(Fig. 4, lower panel). These results demonstrated that Vp1659is secreted specifically through T3SS1.

Vp1659 is not required for secretion of Vp1656 into theculture supernatant. Because the gene encoding Vp1659 islocated in the T3SS1 genomic island, it is possible that Vp1659is a structural protein required for the assembly of an intactsecretion apparatus. To test this possibility, vp1659 was dis-rupted and plasmid pexsA was transformed into both the NY-4and the �vp1659 strains to induce the expression of T3SS1genes. Vp1656 is known to be secreted specifically by T3SS1,and it may play a role in hemolytic activity (34). The resultsshowed that Vp1656 was synthesized and secreted in both

FIG. 4. Secretion of Vp1659 is T3SS1 dependent. Protein samplesfrom the pellet (upper panel) and the supernatant (lower panel) of theNY-4:pexsA (lane 4), �vcrD::pexsA (lane 3), �escV::pexsA (lane 2), and�vcrD escV::pexsA (lane 1) strains were collected, and Vp1659 wasdetected by Western blot assay. Purified Vp1659 was used as a control(lane 5).

FIG. 5. Deletion of vp1659 does not affect the secretion of Vp1656.(A) The NY-4:pexsA and �vp1659::pexsA strains were grown in LB-Sfor 4 h, and pellet (upper panel) and supernatant (lower panel) pro-teins were collected for detection of Vp1656 by Western blot assay.(B) The �exsD, �exsD vp1659, and �exsD vp1656 strains were grown inLB-S for 4 h; and pellet proteins were collected for detection ofVp1656 (upper panel) and DnaK (lower panel) by Western blot assay.(C) The �exsD, �exsD vp1659, and �exsD vp1656 strains were grown inLB-S for 4 h; and supernatant proteins were collected for detection ofVp1656 (upper panel) and DnaK (lower panel) by Western blot assay.

3496 ZHOU ET AL. J. BACTERIOL.

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

strains (Fig. 5A), indicating that Vp1659 is not required for thesecretion of Vp1656. As a confirmation, we deleted vp1659 inthe �exsD strain to produce a �exsD vp1659 strain. Vp1656 wasboth synthesized (Fig. 5B) and secreted (Fig. 5C) in both the�exsD and �exsD vp1659 strains. We mutated vp1656 in the�exsD strain to produce a �exsD vp1656 strain. Polyclonalanti-Vp1656 antibody did not detect a protein band consistentwith Vp1656, indicating that the protein detected in theseexperiments was Vp1656 (Fig. 5C). DnaK served as a loadingcontrol for these experiments (Fig. 5B and C). These resultsshowed that Vp1659 is not a structural protein required for theassembly of a functional secretion apparatus, although theresults of this experiment do not eliminate the possibility thatVp1659 is part of a terminal tip or translocator apparatus.

Vp1659 is translocated into HeLa cells in a T3SS1-depen-dent manner. We employed a mechanical fractionation ap-proach to determine if Vp1659 is associated with the host cellmembrane or cytosol (host membrane/cytosol fraction) afterinfection. Vp1659 was detected in the bacterial pellet of theNY-4, �vcrD, �escV, and �vcrD escV strains (Fig. 6A, 1st row),indicating that all these V. parahaemolyticus strains synthesizedVp1659. Vp1659 was detected in the membrane/cytosol frac-tion of HeLa cells infected with the NY-4 or �escV strain (Fig.6A, 2nd row, lanes 1 and 3). In contrast, Vp1659 was notdetected in this fraction of HeLa cells infected with the �vcrDor �vcrD escV strain (Fig. 6A, 2nd row, lanes 2 and 4). Theseresults indicated that Vp1659 was translocated into the hostcell membrane or cytosol through the T3SS1 of V. parahaemo-lyticus. To determine if the Vp1659 detected here was frombacterial lysis during the mechanical fractionation, we deter-mined if Vp1656 was present in this fraction. Vp1656 is ho-mologous to a translocator protein, YopD, in Yersinia, and wewere not able to detect Vp1656 in the membrane/cytosol frac-tion of HeLa cells infected with the NY-4, �vcrD, �escV, and�vcrD escV strains (Fig. 6A, 3rd and 4th rows). While Vp1656might be expected in this fraction, given that it is homologousto YopD, our results indicate that it is not secreted into thehost cell cytosol. Thus, assuming that this finding does notrepresent an assay sensitivity issue, Vp1656 is a useful negativecontrol for bacterial lysis (Fig. 6A). We further determined ifVp1659 is localized in the cytosol or membrane fraction byultracentrifugation. The membrane/cytosol fraction of HeLacells infected with NY-4 was centrifuged at 30,000 � g for 1 hto obtain the cytosol fraction (supernatant) and the membranefraction (pellet). Vp1659 was detected in the cytosol fractionbut not in the membrane fraction (Fig. 6B). Tubulin, a cytosolmarker, was exclusively present in the cytosol, while calnexin, amembrane marker, was exclusively present in the membranefraction (Fig. 6B), indicating that these two fractions wereseparated successfully by ultracentrifugation and were notcross contaminated.

Vp1659 is not required for translocation of the effector pro-tein VopS. To further determine if Vp1659 plays a role in thetranslocation of other known effectors, we infected HeLa cellswith the NY-4 and �vp1659 strains and determined the trans-location of VopS. VopS was detected in the bacterial pellet ofboth the NY-4 and �vp1659 strains (Fig. 7A, 1st row), indicat-ing that both wild-type and Vp1659 mutant V. parahaemolyti-cus strains synthesized VopS at similar levels. VopS was de-tected in the membrane/cytosol fraction of HeLa cells infected

with both the NY-4 and �vp1659 strains (Fig. 7A, 2nd row).The presence of VopS in the membrane/cytosol fraction ofHeLa cells was not caused by bacterial lysis during the me-chanical fractionation because Vp1656 was present in the bac-terial pellets (Fig. 7A, 3rd row) but was absent in the mem-brane/cytosol fraction of HeLa cells infected with either NY-4or �vp1659 strains (Fig. 7A, 4th row). Tubulin served as acontrol to ensure that similar concentrations of proteins wereloaded (Fig. 7A, 5th row). When we separated the cytosol andmembrane fractions by ultracentrifugation, we found that

FIG. 6. (A) Vp1659 is translocated into HeLa cells, and the trans-location is dependent on an intact T3SS1. HeLa cells were infectedwith the NY-4 (lane 1), �vcrD (lane 2), �escV (lane 3), and �vcrD escV(lane 4) strains. The bacterial pellet (1st and 3rd rows) of each strainwas probed with anti-Vp1659 and anti-Vp1656 antibodies. InfectedHeLa cell cultures were lysed by passage though a 27-gauge needle.After centrifugation (�15,000 � g), the supernatants in the 2nd and4th rows were assayed for the presence of Vp1659 and Vp1656, re-spectively. (B) Translocated Vp1659 is present in the cytosol fractionof HeLa cells infected with the NY-4 strain. The cytosol/membranefraction (see Materials and Methods) was centrifuged at 30,000 � g for1 h to obtain the cytosol (supernatant) and membrane (pellet) frac-tions. The cytosol and membrane fractions were probed with anti-Vp1659, antitubulin, and anticalnexin antibodies.

VOL. 192, 2010 T3SS1 PROTEIN Vp1659 3497

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

VopS was localized in the cytosol of HeLa cells infected withstrain NY-4 (Fig. 7B). These results indicate that the loss ofVp1659 does not prevent the translocation of VopS into thecytosol.

Vp1659 is required for cytotoxicity against HeLa cells. Be-cause T3SS1 is required for the cytotoxicity against HeLa cells(46), we determined if Vp1659 plays any role in this process.HeLa cells were infected with the NY-4, �vp1659, and�vp1659::pvp1659 strains; and cytotoxicity was quantified bythe LDH assay after 1, 2, 3, 4, and 8 h of infection. NY-4caused cytotoxicity against HeLa cells in a time-dependent

manner, with the maximum LDH release occurring at between2 and 3 h and nearly complete cell death being observed after4 h of infection. The cell lysis phenotype was clearly delayedfor the �vp1659 strain, with only 30% of the HeLa cells ap-parently being dead after 4 h of infection (Fig. 8A). The�vp1659::pvp1659 complement strain recovered the full cyto-toxicity phenotype, with approximately 90% of the HeLa cellsbeing lysed after infection for 4 h (Fig. 8A). The cytotoxicityinduced by NY-4 or the �vp1659::pvp1659 strain was signifi-cantly higher than that induced by the �vp1659 strain after 4 hof infection (P 0.05). As a control, analysis of whole-celllysates confirmed that VopS was synthesized at equivalent lev-els by the NY-4, �vp165, and �vp1659::pvp1659 strains in theseexperiments (data not shown). Microscope analysis showedthat HeLa cells infected with NY-4 and the �vp1659::pvp1659strain exhibited rounding and detachment from the bottom ofthe well after 2 h, while HeLa cells infected with the �vp1659strain maintained a uniform monolayer and a normal appear-ance compared with the appearance of the uninfected cells(data not shown).

Vp1659 is required to increase permeability of HeLa cellmembranes. Because the T3SS1 in strain NY-4 is involved inpore formation in the cell membrane of HeLa cells (46), wedetermined if Vp1659 is required for this process. Infectedcells were stained with a membrane-impermeant dye (EtBr)and a membrane-permeant dye (AO). When AO penetratesthe cell, it exhibits green fluorescence, whereas the combina-

FIG. 7. (A) Vp1659 is not required for the translocation of VopS.HeLa cells were infected with the NY-4 (lane 1) and �vp1659 (lane 2)strains. The bacterial pellet (1st and 3rd rows) of each strain wasprobed with anti-VopS antibody (1st row) and anti-Vp1656 antibody(3rd row) to verify the expression of VopS and Vp1656 by each strainduring 2 h of HeLa cell infection. The infected HeLa cells were lysedby passage though a 27-gauge needle. After centrifugation (�15,000 �g), the supernatant was assayed for the presence of VopS (2nd row),Vp1656 (4th row), and tubulin (5th row). (B) The cytosol/membranefraction was centrifuged at 30,000 � g for 1 h to obtain the cytosol(supernatant) and membrane (pellet) fractions. The cytosol and mem-brane fractions were probed with anti-VopS, antitubulin, and antical-nexin antibodies.

FIG. 8. Deletion of vp1659 reduced the ability of V. parahaemolyti-cus to kill HeLa cells. (A) LDH release of HeLa cells infected with theNY-4, �vp1659, and �vp1659::pvp1659 strains over an 8-h time course(expressed as percent cytotoxicity). (B) Percentage of EtBr-positivecells from HeLa cell cultures infected with the NY-4, �vp1659, and�vp1659::pvp1659 strains. The cells were stained with acridine orangeand EtBr; EtBr-positive cells have permeable outer membranes. As-terisk, significant difference compared to the results for the �vp1659strain (P 0.05).

3498 ZHOU ET AL. J. BACTERIOL.

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

tion of both AO and EtBr results in yellow or red fluorescence,thereby confirming membrane permeability. After 2 h of in-fection with NY-4 and the �vp1659::pvp1659 strain, a signifi-cant number of HeLa cells displayed red under a fluorescencemicroscope, indicating that both EtBr and AO entered thecells (data not shown). Quantitative analysis showed that EtBr-positive cells accounted for approximately 35% of the cells forcultures infected with the NY-4 or �vp1659::pvp1659 strain(Fig. 8B), while less than 1% of cells for cultures infected withthe �vp1659 strain were EtBr positive. These results showedthat Vp1659 is required directly or indirectly for pore forma-tion in the cell membrane, leading to the uptake of the mem-brane-impermeant dye (EtBr) (46).

Vp1659 is required for the disruption of actin structure inHeLa cells. V. parahaemolyticus is known to alter the host cellactin structure, but HeLa cells infected with the �vp1659 strain

do not exhibit obvious cell rounding (data not shown). Conse-quently, we confirmed that filamentous actin was disrupted inHeLa cells infected by the wild-type strain of V. parahaemo-lyticus (Fig. 9B) for 2 h. V. parahaemolyticus with the deletionof vp1659 did not affect the filamentous actin structure (Fig.9C), while complementation of the �vp1659 strain with in transvp1659 expression restored the ability of V. parahaemolyticus todisrupt the actin structure (Fig. 9D). Filamentous actin wasundisturbed in our negative control HeLa cells (Fig. 9A).These results indicate that Vp1659 is required for the disrup-tion of the filamentous actin structure in HeLa cells. It is alsonotable that there were no obvious changes for the actin struc-ture in HeLa cells even after 8 h of infection with the �vp1659strain (Fig. 9E).

Vp1659 is required to induce autophagy. When V. parahae-molyticus infects HeLa cells, it induces the conversion of mi-

FIG. 9. (A to E) HeLa cells stained with rhodamine-conjugated phalloidin to show actin structure after 2 h of infection. (A) Uninfected HeLacell control; (B) HeLa cell infected with the NY-4 strain (wild type); (C) HeLa cell infected with the �vp1659 strain; (D) HeLa cells infected withthe �vp1659::pvp1659 strain; (E) HeLa cell infected with the �vp1659 strain for 8 h.

VOL. 192, 2010 T3SS1 PROTEIN Vp1659 3499

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

crotubule-associated protein light chain 3 I (LC3-I) intoLC3-II (3, 4), which is a key marker of autophagy. The LC3-Iin HeLa cells infected with NY-4 was completely convertedinto LC3-II (Fig. 10, lane 1). LC3-I from HeLa cells infectedwith the �vp1659 strain was partially converted into LC3-II(Fig. 10, lane 2), but this result appeared to be no differentfrom that for the uninfected control (Fig. 10, lane 3). The rateof conversion of LC3-I into LC3-II in HeLa cells infected withthe �vp1659::pvp1659 strain was increased compared to thatfor the uninfected control (Fig. 10, lane 4). These results in-dicate that Vp1659 is required for the induction of autophagyby V. parahaemolyticus.

DISCUSSION

In this study, we overexpressed the T3SS1 of V. parahaemo-lyticus and detected two proteins, VopS (Vp1686) and Vp1659.The gene encoding Vp1659 is found in the T3SS1 genomicisland, and it is annotated as a hypothetical protein (28). Whenwe generated a deletion mutant for vp1659, we immediatelydiscovered that host cell death was attenuated (see below), andconsequently, we initiated a systematic series of experiments tocharacterize how Vp1659 contributes to this phenotype. Weinitially demonstrated that synthesis of Vp1659 is regulated ina manner consistent with that for a T3SS1 protein, where ExsAand ExsD serve as positive and negative regulators, respec-tively. Deletion of vp1659 had no apparent affect on the ex-pression of Vp1656 or the secretion of Vp1656 into the extra-cellular milieu, indicating that Vp1659 is not required to moveproteins through the T3SS1 secretion apparatus. Interestingly,mechanical fractionation and ultracentrifugation experimentsfollowed by Western blotting showed that Vp1659 is translo-cated into the cytosol of HeLa cells; VopS is also translocatedeven when vp1659 is deleted. At this point in our investigation,it appeared to be possible that Vp1659 was a potential effectorprotein that contributed directly to the oncosis phenotype thatis produced by V. parahaemolyticus infection (46).

Subsequent phenotypic studies showed that deletion ofvp1659 not only attenuated cell cytolysis and membrane per-meabilization but also blocked actin rearrangement in the hostcell and apparently interfered with the autophagy pathway.Actin rearrangement is linked to VopS (vp1686) (44), whichbelongs to the Fic (filamentation induced by cyclic AMP) pro-tein family. VopS is both necessary and sufficient to induceactin rearrangement by alteration of GTPase via a novel mech-anism called AMPylation (44). There is no evidence directlylinking actin rearrangement to host cell lysis, because deletionof VopS does not reduce cytotoxicity (34). Autophagy is initi-ated by VopQ (vp1680) via a PI3-kinase-independent pathway(3). While autophagy clearly contributes to host cell lysis, itdoes not explain this phenotype entirely (3). Importantly, whileVopS is sufficient to cause actin rearrangement and cell round-ing and VopQ is required to induce autophagy, there is noevidence that these pathways are linked in any manner or withoncosis. Consequently, the only way that deletion of vp1659could block multiple and independent pathways is if Vp1659 ispart of the tip of the injectisome or translocator apparatus,whereupon its absence either significantly reduces or elimi-nates translocation. Judging by the fact that VopS is translo-cated even when vp1659 is deleted and that VopS is sufficientto induce actin rearrangement, we surmise that the loss of thisphenotype with the vp1659 deletion mutant is due to an insuf-ficient quantity of VopS being translocated into the HeLa cellsto affect the actin structure.

The gene encoding Vp1659 is located within the T3SS1island of V. parahaemolyticus, and it is annotated as a 67-kDahypothetical protein. Analysis of protein homology (BLASTpqueries) indicates similarity between the C terminus of Vp1659and LcrV from Yersinia (37 kDa) and PcrV from Pseudomonas(32 kDa) (33% and 35%, respectively). Bioinformatics analysis(http://myhits.isb-sib.ch/cgi-bin/motif_scan) shows that Vp1659(GenBank accession number NP_798038.1) includes somefunctional domains that are not present in LcrV (GenBankaccession number YP_001004069.1), although the matcheswere very limited (e value, �10�3). Amino acids 329 to 607share close sequence homology with LcrV (e value, 1.3 �10�12).

LcrV from Yersinia has been linked to several phenotypes.One study showed that LcrV can be translocated into host cellsand that this translocation process is not dependent on the Yscapparatus (11). Translocated LcrV is colocalized with endoso-mal proteins during the early infection, and during later infec-tion, LcrV is colocalized with lysosomal, mitochondrial, andGolgi proteins (9). LcrV has also been shown to localize on thesurface of the bacterial cells, and treatment with anti-LcrVantibody can prevent the translocation of Yop effectors intothe host cell cytosol (36). Another study showed that treatmentwith anti-LcrV antibody does not block the delivery of Yopsinto the host cells, although this treatment can protect againstexperimental plague (10). Deletion of lcrV blocked the move-ment of Yersina effector proteins into the cytotosol of HeLacells (8, 24), suggesting that LcrV may compose part of thetranslocator apparatus. A recent study showed that LcrV ofYersinia forms a distinct structure at the tip of the secretionneedle to assist the pore formation and translocation of effec-tor proteins into host cells (31). Similarly, deletion of pcrV inPseudomonas does not affect secretion of other T3SS sub-

FIG. 10. Western blot analysis of uninfected HeLa cells or HeLacells that were infected with the NY-4 (lane 1), �vp1659 (lane 2), or�vp1659::pvp1659 (lane 4) strain for 6 h. The Western blot showedevidence for the conversion of LC3-I to LC3-II, which is consistentwith the induction of autophagy. Beta-actin served as a loading controlfor the HeLa cell lysates.

3500 ZHOU ET AL. J. BACTERIOL.

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

strates, but this does abolish translocation of effectors intofibroblasts (15). These results are consistent with our findings,and we hypothesize that Vp1659 in V. parahaemolyticus com-poses part of the secretion apparatus and that the efficiency ofeffector translocation is compromised in the absence ofVp1659. Considering that Vp1659 has a mass approximately2-fold larger than the masses of the LcrV and PcrV homo-logues and that Vp1659 is translocated into the host cell cy-tosol, it may have another role as an effector protein, but thereis no direct evidence of additional functions at this time.

It is important to note that in an earlier study we provision-ally rejected the conclusion that V. parahaemolyticus strainNY-4 induces autophagy in host cells (46). This conclusion wasbased on staining with monodansylcadaverine to reveal auto-phagic bodies, and as might be expected, stain-based assays canlack analytical sensitivity. Burdette et al. (3) used a much moresensitive Western blot assay to show that V. parahaemolyticusinduces the conversion of LC3-I to LC3-II, and this conversionis considered a hallmark of autophagy (20, 25). In the presentstudy, we also employed the LC3 Western blot assay and con-firmed that strain NY-4 also initiates autophagy in HeLa cells.

In conclusion, by overexpressing T3SS1, we identified aT3SS1 protein, Vp1659, which can be detected in the cytosolfraction of HeLa cells after infection. vp1659 expression ispositively regulated by ExsA and negatively regulated by ExsD.Deletion of vp1659 does not affect the secretion of Vp1656 orcompletely block the translocation of VopS, but deletion ofVp1659 limits the ability of V. parahaemolyticus to cause cellmembrane permeabilization, actin depolymerization, and au-tophagy. The mechanism by which Vp1659 induces these phe-notypes is probably through the loss of efficient translocationof relevant effector proteins into host cells. Consequently,Vp1659 is probably a component of the secretion apparatus tipor translocator used by V. parahaemolyticus during infection.

ACKNOWLEDGMENTS

We acknowledge the excellent technical assistance from Lisa Orfe,Stacey LaFrentz, Dan Erwin, Seth Nydam, and Pat Friel and servicesprovided through the Washington State University Monoclonal Anti-body Center.

This project was supported in part by the National Institutes ofHealth, U.S. Department of Health and Human Services, under con-tract number No1-AI-30055 and by the Agricultural Animal HealthProgram, College of Veterinary Medicine, Washington State Univer-sity.

REFERENCES

1. Akeda, Y., K. Nagayama, K. Yamamoto, and T. Honda. 1997. Invasive phe-notype of Vibrio parahaemolyticus. J. Infect. Dis. 176:822–824.

2. Bhattacharjee, R. N., K. S. Park, Y. Kumagai, K. Okada, M. Yamamoto, S.Uematsu, K. Matsui, H. Kumar, T. Kawai, T. Iida, T. Honda, O. Takeuchi,and S. Akira. 2006. VP1686, a Vibrio type III secretion protein, inducesToll-like receptor-independent apoptosis in macrophage through NF-kappaBinhibition. J. Biol. Chem. 281:36897–36904.

3. Burdette, D. L., J. Seemann, and K. Orth. 2009. Vibrio VopQ inducesPI3-kinase-independent autophagy and antagonizes phagocytosis. Mol. Mi-crobiol. 73:639–649.

4. Burdette, D. L., M. L. Yarbrough, A. Orvedahl, C. J. Gilpin, and K. Orth.2008. Vibrio parahaemolyticus orchestrates a multifaceted host cell infectionby induction of autophagy, cell rounding, and then cell lysis. Proc. Natl.Acad. Sci. U. S. A. 105:12497–12502.

5. Coburn, B., I. Sekirov, and B. B. Finlay. 2007. Type III secretion systems anddisease. Clin. Microbiol. Rev. 20:535–549.

6. Daniels, N. A., L. MacKinnon, R. Bishop, S. Altekruse, B. Ray, R. M.Hammond, S. Thompson, S. Wilson, N. H. Bean, P. M. Griffin, and L.Slutsker. 2000. Vibrio parahaemolyticus infections in the United States, 1973–1998. J. Infect. Dis. 181:1661–1666.

7. Daniels, N. A., B. Ray, A. Easton, N. Marano, E. Kahn, A. L. McShan II, L.Del Rosario, T. Baldwin, M. A. Kingsley, N. D. Puhr, J. G. Wells, and F. J.Angulo. 2000. Emergence of a new Vibrio parahaemolyticus serotype in rawoysters: a prevention quandary. JAMA 284:1541–1545.

8. DeBord, K. L., V. T. Lee, and O. Schneewind. 2001. Roles of LcrG and LcrVduring type III targeting of effector Yops by Yersinia enterocolitica. J. Bac-teriol. 183:4588–4598.

9. DiMezzo, T. L., G. Ruthel, E. E. Brueggemann, H. B. Hines, W. J. Ribot,C. E. Chapman, B. S. Powell, and S. L. Welkos. 2009. In vitro intracellulartrafficking of virulence antigen during infection by Yersinia pestis. PLoS One4:e6281.

10. Fields, K. A., M. L. Nilles, C. Cowan, and S. C. Straley. 1999. Virulence roleof V antigen of Yersinia pestis at the bacterial surface. Infect. Immun. 67:5395–5408.

11. Fields, K. A., and S. C. Straley. 1999. LcrV of Yersinia pestis enters infectedeukaryotic cells by a virulence plasmid-independent mechanism. Infect. Im-mun. 67:4801–4813.

12. Galan, J. E., and A. Collmer. 1999. Type III secretion machines: bacterialdevices for protein delivery into host cells. Science 284:1322–1328.

13. Galan, J. E., and H. Wolf-Watz. 2006. Protein delivery into eukaryotic cellsby type III secretion machines. Nature 444:567–573.

14. Gauthier, A., M. de Grado, and B. B. Finlay. 2000. Mechanical fractionationreveals structural requirements for enteropathogenic Escherichia coli Tirinsertion into host membranes. Infect. Immun. 68:4344–4348.

15. Goure, J., A. Pastor, E. Faudry, J. Chabert, A. Dessen, and I. Attree. 2004.The V antigen of Pseudomonas aeruginosa is required for assembly of thefunctional PopB/PopD translocation pore in host cell membranes. Infect.Immun. 72:4741–4750.

16. Hlady, W. G., and K. C. Klontz. 1996. The epidemiology of Vibrio infectionsin Florida, 1981–1993. J. Infect. Dis. 173:1176–1183.

17. Hueck, C. J. 1998. Type III protein secretion systems in bacterial pathogensof animals and plants. Microbiol. Mol. Biol. Rev. 62:379–433.

18. Johnson, S., J. E. Deane, and S. M. Lea. 2005. The type III needle and thedamage done. Curr. Opin. Struct. Biol. 15:700–707.

19. Joseph, S. W., R. R. Colwell, and J. B. Kaper. 1982. Vibrio parahaemolyticusand related halophilic vibrios. Crit. Rev. Microbiol. 10:77–124.

20. Kabeya, Y., N. Mizushima, T. Ueno, A. Yamamoto, T. Kirisako, T. Noda, E.Kominami, Y. Ohsumi, and T. Yoshimori. 2000. LC3, a mammalian homo-logue of yeast Apg8p, is localized in autophagosome membranes after pro-cessing. EMBO J. 19:5720–5728.

21. Kirby, J. E., J. P. Vogel, H. L. Andrews, and R. R. Isberg. 1998. Evidence forpore-forming ability by Legionella pneumophila. Mol. Microbiol. 27:323–336.

22. Kodama, T., H. Hiyoshi, K. Gotoh, Y. Akeda, S. Matsuda, K. S. Park, V. V.Cantarelli, T. Iida, and T. Honda. 2008. Identification of two transloconproteins of Vibrio parahaemolyticus type III secretion system 2. Infect. Im-mun. 76:4282–4289.

23. Kodama, T., M. Rokuda, K. S. Park, V. V. Cantarelli, S. Matsuda, T. Iida,and T. Honda. 2007. Identification and characterization of VopT, a novelADP-ribosyltransferase effector protein secreted via the Vibrio parahaemo-lyticus type III secretion system 2. Cell. Microbiol. 9:2598–2609.

24. Lee, V. T., C. Tam, and O. Schneewind. 2000. LcrV, a substrate for Yersiniaenterocolitica type III secretion, is required for toxin targeting into the cy-tosol of HeLa cells. J. Biol. Chem. 275:36869–36875.

25. Levine, B., and J. Yuan. 2005. Autophagy in cell death: an innocent convict?J. Clin. Invest. 115:2679–2688.

26. Liverman, A. D., H. C. Cheng, J. E. Trosky, D. W. Leung, M. L. Yarbrough,D. L. Burdette, M. K. Rosen, and K. Orth. 2007. Arp2/3-independent as-sembly of actin by Vibrio type III effector VopL. Proc. Natl. Acad. Sci.U. S. A. 104:17117–17122.

27. Lynch, T., S. Livingstone, E. Buenaventura, E. Lutter, J. Fedwick, A. G.Buret, D. Graham, and R. DeVinney. 2005. Vibrio parahaemolyticus disrup-tion of epithelial cell tight junctions occurs independently of toxin produc-tion. Infect. Immun. 73:1275–1283.

28. Makino, K., K. Oshima, K. Kurokawa, K. Yokoyama, T. Uda, K. Tagomori,Y. Iijima, M. Najima, M. Nakano, A. Yamashita, Y. Kubota, S. Kimura, T.Yasunaga, T. Honda, H. Shinagawa, M. Hattori, and T. Iida. 2003. Genomesequence of Vibrio parahaemolyticus: a pathogenic mechanism distinct fromthat of V. cholerae. Lancet 361:743–749.

29. Milton, D. L., A. Norqvist, and H. Wolf-Watz. 1992. Cloning of a metallo-protease gene involved in the virulence mechanism of Vibrio anguillarum. J.Bacteriol. 174:7235–7244.

30. Milton, D. L., R. O’Toole, P. Horstedt, and H. Wolf-Watz. 1996. Flagellin Ais essential for the virulence of Vibrio anguillarum. J. Bacteriol. 178:1310–1319.

30a.Morales, V. M., A. Backman, and M. Bagdasarian. 1991. A series of wide-host-range low-copy-number vectors that allow direct screening for recom-binants. Gene 97:39–47.

31. Mueller, C. A., P. Broz, S. A. Muller, P. Ringler, F. Erne-Brand, I. Sorg, M.Kuhn, A. Engel, and G. R. Cornelis. 2005. The V-antigen of Yersinia formsa distinct structure at the tip of injectisome needles. Science 310:674–676.

32. Nakasone, N., and M. Iwanaga. 1990. Pili of a Vibrio parahaemolyticus strainas a possible colonization factor. Infect. Immun. 58:61–69.

VOL. 192, 2010 T3SS1 PROTEIN Vp1659 3501

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

33. Nishibuchi, M., A. Fasano, R. G. Russell, and J. B. Kaper. 1992. Entero-toxigenicity of Vibrio parahaemolyticus with and without genes encodingthermostable direct hemolysin. Infect. Immun. 60:3539–3545.

34. Ono, T., K. S. Park, M. Ueta, T. Iida, and T. Honda. 2006. Identification ofproteins secreted via Vibrio parahaemolyticus type III secretion system 1.Infect. Immun. 74:1032–1042.

35. Park, K. S., T. Ono, M. Rokuda, M. H. Jang, K. Okada, T. Iida, and T.Honda. 2004. Functional characterization of two type III secretion systems ofVibrio parahaemolyticus. Infect. Immun. 72:6659–6665.

36. Pettersson, J., A. Holmstrom, J. Hill, S. Leary, E. Frithz-Lindsten, A. vonEuler-Matell, E. Carlsson, R. Titball, A. Forsberg, and H. Wolf-Watz. 1999.The V-antigen of Yersinia is surface exposed before target cell contact andinvolved in virulence protein translocation. Mol. Microbiol. 32:961–976.

37. Ralph, A., and B. J. Currie. 2007. Vibrio vulnificus and V. parahaemolyticusnecrotising fasciitis in fishermen visiting an estuarine tropical northern Aus-tralian location. J. Infect. 54:e111–e114.

38. Reyes, A. L., R. G. Crawford, P. L. Spaulding, J. T. Peeler, and R. M. Twedt.1983. Hemagglutination and adhesiveness of epidemiologically distinctstrains of Vibrio parahaemolyticus. Infect. Immun. 39:721–725.

39. Reference deleted.40. Troisfontaines, P., and G. R. Cornelis. 2005. Type III secretion: more sys-

tems than you think. Physiology (Bethesda) 20:326–339.

41. Warren, S. M., and G. M. Young. 2005. An amino-terminal secretion signalis required for YplA export by the Ysa, Ysc, and flagellar type III secretionsystems of Yersinia enterocolitica biovar 1B. J. Bacteriol. 187:6075–6083.

42. Xu, M., K. Yamamoto, T. Honda, and X. Ming. 1994. Construction andcharacterization of an isogenic mutant of Vibrio parahaemolyticus having adeletion in the thermostable direct hemolysin-related hemolysin gene (trh).J. Bacteriol. 176:4757–4760.

43. Yamamoto, T., and T. Yokota. 1989. Adherence targets of Vibrio parahae-molyticus in human small intestines. Infect. Immun. 57:2410–2419.

44. Yarbrough, M. L., Y. Li, L. N. Kinch, N. V. Grishin, H. L. Ball, and K. Orth.2009. AMPylation of Rho GTPases by Vibrio VopS disrupts effector bindingand downstream signaling. Science 323:269–272.

45. Yeung, P. S., and K. J. Boor. 2004. Epidemiology, pathogenesis, and pre-vention of foodborne Vibrio parahaemolyticus infections. Foodborne Pathog.Dis. 1:74–88.

46. Zhou, X., M. E. Konkel, and D. R. Call. 2009. Type III secretion system 1 ofVibrio parahaemolyticus induces oncosis in both epithelial and monocytic celllines. Microbiology 155:837–851.

47. Zhou, X., D. H. Shah, M. E. Konkel, and D. R. Call. 2008. Type III secretionsystem 1 genes in Vibrio parahaemolyticus are positively regulated by ExsAand negatively regulated by ExsD. Mol. Microbiol. 69:747–764.

3502 ZHOU ET AL. J. BACTERIOL.

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from