Ail Proteins of Yersinia pestis and Y. pseudotuberculosis Have Different Cell Binding and Invasion Activities

The Yersinia pestis adhesin Ail mediates host cell binding and facilitates delivery of cytotoxic Yop proteins. Ail from Y. pestis and Y. pseudotuberculosis is identical except for one or two amino acids at positions 43 and 126 depending on the Y. pseudotuberculosis strain. Ail from Y. pseudotuberculosis strain YPIII has been reported to lack host cell binding ability, thus we sought to determine which amino acid difference(s) are responsible for the difference in cell adhesion. Y. pseudotuberculosis YPIII Ail expressed in Escherichia coli bound host cells, albeit at ∼50% the capacity of Y. pestis Ail. Y. pestis Ail single mutants, Ail-E43D and Ail-F126V, both have decreased adhesion and invasion in E. coli when compared to wild-type Y. pestis Ail. Y. pseudotuberculosis YPIII Ail also had decreased binding to the Ail substrate fibronectin, relative to Y. pestis Ail in E. coli. When expressed in Y. pestis, there was a 30–50% decrease in adhesion and invasion depending on the substitution. Ail-mediated Yop delivery by both Y. pestis Ail and Y. pseudotuberculosis Ail were similar when expressed in Y. pestis, with only Ail-F126V giving a statistically significant reduction in Yop delivery of 25%. In contrast to results in E. coli and Y. pestis, expression of Ail in Y. pseudotuberculosis led to no measurable adhesion or invasion, suggesting the longer LPS of Y. pseudotuberculosis interferes with Ail cell-binding activity. Thus, host context affects the binding activities of Ail and both Y. pestis and Y. pseudotuberculosis Ail can mediate cell binding, cell invasion and facilitate Yop delivery.


Introduction
There are three Yersinia species pathogenic for humans.While the enteric pathogens Y. enterocolitica and Y. pseudotuberculosis cause primarily self-limiting gastroenteritis, and in some cases mesenteric lymphadenitis, Y. pestis causes the rapidly fatal disease plague [1,2].While Y. enterocolitica and Y. pseudotuberculosis both cause enteric infections via an oral route of infection, Y. pseudotuberculosis and Y. pestis are more closely related genetically, estimated to have evolved from one another between 1500 and 20,000 years ago [3].
Ail from Y. enterocolitica has been well characterized and many functions have been elucidated, including serum resistance and adhesion to and invasion into host cells [23,24,25,26,27,28]. Y. enterocolitica Ail binds cultured cells in a species-specific manner with adhesion to CHO and HEp-2 cells, but not MDCK cells [23].Further studies identified Ail point mutants with intermediate and severe serum sensitivity that also affected invasion [29].In particular, an aspartic acid (D67) and valine (V68) at the Cterminal end of loop 2 were required for both Ail functions.
Ail from Y. pseudotuberculosis is reported to lack adhesion and invasion activities when expressed in E. coli, although E. coli expressing Ail was still able to confer serum resistance [30] due to C4bp and factor H binding, like Y. enterocolitica, [31,32].Thus, Y. pseudotuberculosis Ail has long been believed to have no adhesion capacity.
Ail from Y. pestis has been shown to mediate serum resistance [22,33], auto-aggregation [22], and cell adhesion [6,22].Additionally, Ail is a key Y. pestis adhesin for Yop delivery and virulence [6,34].The extracellular matrix (ECM) proteins fibronectin and laminin are substrates for Y. pestis Ail and these Ail-ECM interactions are important for adhesion to host cells and Yop delivery [12,13,35].The crystal structure of Ail from Y. pestis has been determined [13] and it belongs to the OmpX family of proteins described as having a flattened b-barrel with four extracellular loops extending above the surface of the bacteria [36,37].The four extracellular loops of Y. pestis Ail contain 10-21 amino acids each.
Since Ail is a critical molecule for Y. pestis adhesion, Yop delivery, and virulence, we wanted to identify residues of Ail required for adhesion and Yop delivery.Although the previous studies of OmpX and Y. enterocolitica Ail are useful, the functions of these proteins cannot be translated to Y. pestis Ail, as even Y.enterocolitica Ail is only 26-80% identical to Y. pestis Ail within the four extracellular loops.Y. pseudotuberculosis Ail is 98.9-100% identical to Y. pestis Ail, depending on the Y. pseudotuberculosis strain, with only two amino acid changes, E43D and F126V in the most divergent Y. pseudotuberculosis derivatives.As Y. pseudotuberculosis YPIII Ail was previously reported to lack cell binding activity [30], we hypothesized these two amino acids should be important for the adhesive activity of Ail.In this study, we revisited the adhesion capacity of Y. pseudotuberculosis Ail and observed reduced but significant binding to cultured host cells, relative to Y. pestis Ail.Single mutations based on Y. pseudotuberculosis Ail introduced into the Y. pestis Ail molecule were analyzed for their cell adhesion, cell invasion and Yop delivery functions.Y. pseudotuberculosis YPIII Ail and single mutations were defective for host cell binding and invasion by up to 75%, relative to Y. pestis Ail, when expressed in E. coli and Y. pestis.However, the strongest defect observed for Yop delivery was 25%, and most forms had no decrease in Yop delivery relative to Y. pestis Ail.

Strains and Culture Conditions
Bacterial strains and plasmids used in this study are listed in Table S1.Y. pestis strains were cultivated in heart infusion broth (HIB) overnight or on heart infusion agar (HIA) for 48 hours at 28uC.Y. pseudotuberculosis strains were cultivated in brain heart infusion broth (BHI) overnight or on brain heart infusion agar for 48 hours at 28uC.E. coli strains were cultured in lysogeny broth (LB) or LB agar at 28uC or 37uC.Antibiotics were used at the following concentrations for E. coli and Y. pestis: chloramphenicol (Cm), 25 mg/ml; and ampicillin (Amp), 100 mg/ml.For Y. pseudotuberculosis strains, antibiotic concentrations were: kanamycin (Kan), 6 mg/ml; chloramphenicol (Cm), 5 mg/ml, tetracycline (Tet), 2.5 mg/ml; and ampicillin (Amp), 100 mg/ml.Isopropyl-ß-D-thiogalactopyranoside (IPTG) was used at a 100 mM concentration for E. coli and Y. pestis and 500 mM for Y. pseudotuberculosis, unless otherwise noted.The Y. pseudotuberculosis strain YPIII was obtained from Dr. Ralph Isberg and YPIII and IP2666 from Dr. James Bliska.

Construction of Ail-expressing Plasmids
pSK-ail Bluescript encoding wild-type Y.pestis Ail was constructed by PCR amplification of the ail locus from Y. pestis strain KIM5 [6] using primers 59-GCGCGGATCCTTGGCTGGCCACTTTAGTCT-39 and 59-GCGCCTGCAGGGTTAGGAGGACGTTAGAAC-39.The ail PCR product was digested with BamHI and PstI and ligated into BamHI/PstI-cut pSK Bluescript (Stratagene).ail from the Y. pseudotuberculosis strains YPIII and IP2666 were similarly PCR amplified and ligated into pSK Bluescript.All clones were also moved into the IPTG-inducible vectors pMMB207 [38] and pMMB66EH [39] for cell binding studies.

Generation of Mutations
PCR-mutagenesis was performed using the enzyme Pfu (Stratagene) and primer pairs encoding the mutations Ail-E43D and Ail-F126V.The primers used were complementary to one another and are listed here: ail-E43D top strand 59 -caaagtcgtgt-caagGACgatgggtacaagttgg and bottom strand 59 -ccaacttgtacc-catcGTCcttgacacgactttg; ail-F126V top strand 59 -catggaaaggc- taaaGTTtcctcaatatttggtc and bottom strand 59 -gaccaaatattgag-gaAACtttagcctttccatg.Ail-T7I/E43D/F126V was generated using pSK-ail-E43D/F126V as a template and T7I primers, T7I TOP 59-tttttatgaataagaTattactggtctcttc-39 and T7I BOTTOM 59-gaa-gagaccagtaatAtcttattcataaaaa-39.Following PCR amplification using a pSK-ail Bluescript-derived plasmid as a template, the PCR reactions were digested with DpnI to cut the template DNA and transformed into DH5alpha [40].Potential mutant clones were sequenced to confirm that only the target site was mutated and a BamHI/PstI fragment containing the entire open reading frame and ribosome-binding site, was liberated, purified and ligated into the IPTG-inducible plasmid pMMB207 (Cm R ) or pMMB66EH (Amp R ).

Adhesion Assays and Invasion
Adhesion assays were performed as described previously [12], unless otherwise indicated.HEp-2 cells were cultured in 24-well tissue culture plates until reaching 100% confluence.Bacteria at the proper dilution was added to cultured HEp-2 cells at a multiplicity of infection (MOI) of 50:1.After 1 hour 45 minutes incubation at 37uC in 5% CO 2 cell-associated bacteria were liberated by the addition of sterile H 2 O containing 0.1% Triton X-100 for 10-20 min.Percent adhesion was calculated by dividing bound CFU by total bacteria in the well and then multiplying by 100.For E. coli and Y. pestis adhesion, strains were grown overnight in LB or HIB respectively, and then diluted 1:50 or 1:10 respectively, into fresh media with 100 mM IPTG for 3 hrs.For Y. pseudotuberculosis adhesion, strains were grown overnight in BHI in the presence of 500 mM IPTG, and then diluted 1:10 into fresh media with 500 mM IPTG for 3 hrs.
Invasion assays were performed similarly, except that at the end of 1 hour 45 minute of bacterial binding, cells were washed once with phosphate-buffered saline (PBS) to remove unbound bacteria and minimal essential medium containing 7.5 mg/ml gentamicin was added for 1 hour at 37uC in 5% CO 2 to kill extracellular bacteria.Cells were then washed twice with PBS and lysed and plated as described for the adhesion assay.For E. coli invasion, strains were grown overnight in LB, then diluted 1:50 into fresh media with 100 mM IPTG for 3 hrs.For Y. pestis and Y.  pseudotuberculosis invasion assays, strains were induced overnight with 100 mM or 500 mM IPTG, respectively.Then cells were diluted 1:10 and grown for an additional 3 hrs in the continued presence of 100 mM or 500 mM IPTG prior to addition to cells.

Bacterial Binding to Fibronectin
Bacterial binding assays were performed as described previously [12].Briefly, bacterial cells were diluted and added to immobilized fibronectin (Sigma, F2006) and allowed to bind at 37uC for 2 hours.Bacteria bound to the wells were stained with 0.01% crystal violet.After washing away excess crystal violet, the bacterialassociated crystal violet stain was solubilized with an 80%methanol/20%acetone solution.The absorbance was measured at ABS 595 .

Cytotoxicity Assay
Cytotoxicity assays were performed as described previously [12].Briefly, Y. pestis KIM5 derivatives were added to HEp-2 cells at an MOI of 10.After 4 hours of incubation at 37uC in 5% CO 2 , cells were fixed with 0.5 ml methanol and stained with 0.76 mg/ ml Geimsa stain.Rounding was observed and pictures were taken under a phase-contrast microscope.Cytotoxicity was enumerated by counting total cells and the number of round dark purple (shrunken cytoplasm) cells experiencing cytotoxicity in three microscopic fields (,125 cells/field).Percent cytotoxicity was calculated by dividing rounded cells by total cells.This experiment was performed three times in duplicate, with two fields of cells enumerated/well n = 12.Statistical significance was assessed using the student t test.

Western Blot Assays
Cultures of bacteria normalized for their OD 600 (E. coli and Y. pseudotuberculosis) or OD 620 (Y.pestis) were resuspended in Laemmli sample buffer.Bacterial cell extracts were boiled and run on a 15% SDS-polyacrylamide gel electrophoresis (PAGE) gel.Proteins were either visualized by Coomassie Brilliant Blue straining or transferred to nitrocellulose and probed with rabbit anti-Ail serum (a kind gift from Dr. Ralph Isberg, [30]) at a 1:500 dilution.

Amino Acid Sequence of Ail from Various Y. pseudotuberculosis Strains
Prior to determining the relative binding capacities of Y. pestis and Y. pseudotuberculosis Ail, we compared Ail sequences from representative strains of both Yersinia species.For Y. pestis, sequences from strains KIM5 and CO92 were used and for Y. pseudotuberculosis, sequences strains YPIII, IP2666, PB1/+, IP31758 and IP32953 were used.The sequence alignment (Fig. 1) shows Y. pseudotuberculosis YPIII and IP2666 Ail has an aspartate at position 43, instead of a glutamate acid (E43D) and a valine at position 126 instead of a phenylalanine (F126V).The E43D substitution is predicted to lie in surface exposed loop 1 while F126V is predicted to lie in loop 3.At these positions, these two residues may directly contact cell components such as fibronectin [12,35].Unlike a previously published YPIII Ail sequence [30], our Y.pseudotuberculosis YPIII Ail sequence did not contain an additional amino acid change at position 7, T7I.Analysis from other Y. pseudotuberculosis sequencing projects demonstrated that Y. pseudotuberculosis strains IP31758 and IP32953 only contain the F126V change relative to Y. pestis Ail, and Y. pseudotuberculosis strain PB1/+, has an Ail sequence identical to Y. pestis (Fig. 1).

Ail from Y. pestis and Y. pseudotuberculosis Exhibit Adhesion to Host Cells when Expressed in E. coli
To determine the relative binding efficiencies of Ail proteins from different Yersinia strains to host cells, Ail from the Y. pestis strain KIM5 [6], Y. pseudotuberculosis YPIII and Y. pseudotuberculosis IP2666 were expressed from an IPTG-inducible construct in E. coli AAEC185 (a non-fimbriated strain of E. coli, [41]), and adhesion to cultured HEp-2 cells was measured.Adhesion of E. coli expressing KIM5 Ail to HEp-2 cells was set to 100% (actual adhesion was 4.0%).In contrast to a previous report [30], we observed treatment for 1 hr after a 2 hour infection of cells, compared to the bacterial inoculum over the same time period on HEp-2 cells, n = 9.Ail expression levels were assessed by Coomassie gel staining.Actual invasion by Y. pestis Ail was ,1.0%.C) HEp-2 cells were infected with KIM5 DailDpla expressing the various Ail derivatives at an MOI of 10, to determine Ail-mediated cytotoxicity.After 4 hours of infection, cells were fixed and stained with Geimsa to show shrunken, round, darker cells, indicative of Yop-mediated cytotoxicity.Cells were counted and the percent of cytotoxic cells were calculated, n = 12.Again, Ail expression was assessed by Coomassie gel staining.*P,0.05,**P,10 211 .doi:10.1371/journal.pone.0083621.g005significant adhesion activity from the two Y. pseudotuberculosis alleles (both encode identical ORFs, Fig. 1).Ail from the two Y. pseudotuberculosis strains exhibited ,60% adhesion to HEp-2 cells, relative to Y. pestis Ail, while E. coli harboring the pMMB207 empty vector had ,2% adhesion relative to Y. pestis Ail (Fig. 2A).Whole cell lysates demonstrated E. coli expressing the various Ail constructs made similar amounts of Ail protein (Fig. 2B).These data indicate that while Y. pseudotuberculosis Ail has reduced adhesion ability relative to Y. pestis Ail when expressed in E. coli, it does mediate adhesion to cultured cells in vitro.
The various Ail proteins were also tested for their ability to facilitate E. coli invasion of HEp-2 cells, an activity previously reported for Y. enterocolitica [23,42], and Y. pestis Ail [22].While KIM5 Ail protein mediates low levels of invasion in this system (0.01%), this level is 50-fold above background E. coli invasion.Relative to KIM5 Ail-mediated invasion (set at 100%), Ail from Y. pseudotuberculosis YPIII and IP2666 had 25% invasion activity.E. coli expressing vector alone invaded with only 2% of the efficiency of KIM5 Ail (Fig. 2C).These results indicate that like adhesion function, invasion activity of Y. pseudotuberculosis YPIII Ail was significantly lower than Y. pestis Ail when expressed in E. coli.

Single Y. pseudotuberculosis-like Amino Acid Changes within Y. pestis Ail also Confer Reduced Adhesion and Invasion Function
We demonstrated that Ail from Y. pseudotuberculosis strains YPIII and IP2666 have reduced adhesive activity relative to Y. pestis Ail (Fig. 2A).Thus, we wanted to determine which of the two amino acid differences contributed to the reduced adhesion ability.We generated each of the single amino acid mutant derivatives, Ail-E43D and Ail-F126V, and tested their ability to mediate E. coli binding to HEp-2 cells.Adhesion to HEp-2 cells was normalized to the wild-type Y.pestis (KIM5) Ail construct.The Ail-E43D mutant exhibited about 32% adhesion when compared to KIM5 Ail, while the Ail-F126V mutant gave about 71% adhesion (Fig. 3A).Like previous results (Fig. 2A), Y. pseudotuberculosis YPIII Ail mediated 50% adhesion relative to Y. pestis Ail.Again, the expression of various Ail derivatives was analyzed by Western blotting and the levels of protein were comparable across all strains (Fig. 3B).
Invasion function of the single amino acid mutants in the KIM5 Ail protein was also assessed.The invasion capacity of the two single mutants were similarly reduced as the Ail-E43D and Ail-F126V mutations gave 45% and 55% invasion, respectively.For both adhesion and invasion, the Ail-E43D mutant had significantly less activity than the Ail-F126V mutant.We presume the slightly increased level of Y. pseudotuberculosis YPIII Ail-mediated invasion (45%) as compared to previous experiments (Fig. 2C) is due to experimental variation.

KIM5 Ail and YPIII Ail Bind to Purified Fibronectin with Different Efficiencies
We have previously shown that fibronectin is a host cell substrate for Y. pestis Ail [12].Therefore, we determined whether reduced binding of Y. pseudotuberculosis YPIII/IP2666 Ail to purified fibronectin could account for the difference in host cell binding.E. coli expressing various Ail derivatives were added to increasing concentrations of purified plasma fibronectin coated on microtiter plates.E. coli expressing KIM5 Ail on its surface bound to fibronectin in a manner that neared saturation as concentrations approached 20 mg/ml (Fig. 4).E. coli expressing the Ail-E43D, Ail-F126V, and Y. pseudotuberculosis YPIII/IP2666 Ail bound fibronectin less efficiency than Y. pestis Ail at coating concentrations of 10 mg/ml and 20 mg/ml.These decreases were significant for Ail-F126V (p = 0.02) and Y. pseudotuberculosis YPIII Ail (p = 0.03).Ail-E43D showed a trend toward reduced binding but did not reach statistical significance (p = 0.08).There was also a trend towards reduced fibronectin binding for Ail-E43D, Ail-F126V and YPIII Ail at 5 mg/ml, but those differences did not reach statistical significance (for Y. pestis Ail vs. YPIII Ail, p = 0.11).Thus, the various Ail proteins have similar binding patterns to purified fibronectin, but KIM5 Ail appears to bind fibronectin slightly more efficiently than the other derivatives tested.

Defects in Adhesion and Invasion Activity of Various Ail
Derivatives are Maintained in Y. pestis Y. pestis has distinct surface characteristics compared to E. coli strains such as AAEC185 owing to differences in their LPS core structure [43].Thus, we determined whether our various Ail derivatives had a defect in adhesion when expressed in the more natural context of Y. pestis.
Each derivative was expressed in a Y. pestis KIM5 DailDpla strain.Like our findings in E. coli, there was a significant difference in adhesion activity for Y. pestis Ail, as compared to Ail-E43D and Ail-F126V, although the trend toward reduced adhesion for YPIII Ail did not reach statistical significance (Fig. 5A) and adhesion defects were not as dramatic as in E. coli (Fig. 3A).Background adhesion was low in this strain (,0.25% in the presence of empty vector pMMB207) as it is deleted for both major Y. pestis adhesions, Ail and Pla [34].Invasion activities by the various Ail derivatives were also significantly different in Y. pestis KIM5 DailDpla (Fig. 5B), although one clone, Ail-F126V, did not reach statistical significance (P = 0.067).pla encodes plasminogen activator, the major mediator of cell invasion in Y. pestis [6,44].As with adhesion, in Y. pestis the invasion defects for the various Ail derivatives were not as drastic as in E. coli (Fig. 3C).
Perhaps reflecting the modest defects of Y. pseudotuberculosis forms of Ail on adhesion and invasion in Y. pestis, the ability of Ail to facilitate Type III secretion of Yop effector proteins into host cells was no different in Y. pestis DailDpla expressing Y. pestis Ail, Ail-E43D or YPIII Ail, and only a modest 25% reduction in Yop delivery was observed with Ail-F126V (Fig. 5C).While the KIM5 DailDpla derivative led to only 5% Yop-mediated cytotoxicity, comparable to the DyopB derivative, expression of all four Ail derivatives tested resulted in 35-45% cytotoxicity in 4 hours (Fig. 5C).YopB is part of the YopB/D translocon complex of the T3SS, and is required for Yop delivery [7].

Ail Lacks Adhesion or Invasion Activity when Expressed in Y. pseudotuberculosis
We next assessed the adhesion and invasion activity of each Ail derivative expressed in Y. pseudotuberculosis, which has an Oantigen-containing LPS [45].The strain used was a YPIII derivative lacking Ail, invasin, pH 6 antigen and the YadAencoding virulence plasmid, pYV (YP18 P-; ail::Cm R , inv::Tet R , psaABC::Kan R , pYV-).Expression of KIM5 Ail or the other Ail derivatives in strain YP18 P-gave no adhesion or invasion above background levels observed in the presence of the empty expression vector, pMMB66EH (Fig. 6), even though the proteins were expressed (Fig. 6).It is notable that the levels of YP18 Padhesion and invasion in the presence of various forms of Ail is very low, 0.15-0.3%and 0.0004-0.0012%,respectively.This compares with 6% adhesion and 1% invasion by KIM5 Ail when expressed in Y. pestis.Thus, the context of the outer membrane environment in which Ail is expressed has an impact on the ability of Ail to function as an adhesin and direct invasion of host cells.

A T7I Mutation in Y. pseudotuberculosis Ail Renders the Protein Unstable
To address differences between our results and a previous report on Y. pseudotuberculosis Ail expressed in E. coli, we also generated the T7I mutation in the Y. pseudotuberculosis YPIII background.This Ail derivative has three amino acid changes relative to Y. pestis Ail (T7I, E43D, and F126V), and is the reported YPIII sequence from the previous study [30].This version of Ail was expressed in E. coli and tested for adhesion to HEp-2 cells.Y. pseudotuberculosis YPIII Ail harboring the T7I mutation was unstable and conferred no adhesive function (Fig. 7).Therefore, the T7I mutation likely explains why a previous report found no adhesion activity for Y. pseudotuberculosis Ail [30].We did not find the T7I substitution in two Y. pseudotuberculosis strains from which we sequenced ail, one of which is the same strain (YPIII) reported previously to contain the T7I substitution.Nor was the T7I substitution present in three other Y. pseudotuberculosis sequenced strains (Fig. 1).

Discussion
The Ail protein of pathogenic Yersinia species is an outer membrane protein comprised of eight transmembrane ß-strands and four extracellular loops of 10-21 amino acids available for substrate binding [13,29,46].Those substrates include the host cell ECM proteins, fibronectin and laminin [12,13,35], as well as the complement regulatory proteins C4bp and factor H [26,27,28,31,32].While the Ail extracellular loops of Y. enterocolitica have diverged significantly from Y. pestis and Y. pseudotuberculosis (55% identity over the four loops), the Y. pestis and Y. pseudotuberculosis Ail molecules differ in zero, one or two amino acid residues depending on the strain (Fig. 1).Based on a previous report that Y. pseudotuberculosis Ail from strain YPIII is unable to mediate E. coli binding to host cells [30], we sought to identify residues important for the interaction of Y. pestis Ail with host cells.Two mutants, Ail-E43D and Ail-F126V, derived from Y. pestis Ail, and Y. pseudotuberculosis YPIII Ail (Ail-E43D/F126V, relative to Y. pestis Ail) were analyzed for cell binding and cell invasion capacity in E. coli.We found the Ail-E43D mutation is more defective than Ail-F126V for host cell binding and invasion, although both mutations affected these activities (Fig. 3).When expressed in Y. pestis, Ail-E43D, Ail-F126V and YPIII Ail also had defects in cell binding and invasion activities, albeit slightly less defective than what was found in E. coli (Figs. 3 and 5).Since Y. pestis contains an unusually short LPS (no O-antigen [47]), we assessed the cellbinding activity of KIM5 Ail, Ail-E43D, Ail-F126V and YPIII Ail in the Y. pseudotuberculosis YPIII strain background, which has a full complement of LPS with O-antigen.In a YPIII strain background lacking other adhesins (YP18 P-; ail::Cm R , inv::Tet R , psaABC::Kan R , pYV-), none of the Ail derivatives, including KIM5 Ail were able to mediated cell adhesion or invasion (Fig. 6).Previous studies have demonstrated that the proteolytic and adhesive activities of another Y. pestis outer membrane protein, the protease/adhesin plasminogen activator (Pla), is inhibited by the longer LPS present in Y. pseudotuberculosis [48].Thus, Ail and Pla may share a preference for certain outer membrane environments for function.It is clear that some activities of Ail are maintained in Y. pseudotuberculosis as it has been demonstrated that Ail mediates serum resistance in Y. pseudotuberculosis [30,31].However, serum resistance protects against host complement components that must reach the bacterial outer membrane to mediate their effects.Thus, Ail would have access to such host proteins, whereas access to extracellular matrix proteins and host cell receptors by the relatively short Ail protein may be sterically occluded by the lengthy LPS in Y. pseudotuberculosis.It remains possible that under certain conditions at 37uC in vivo Y. pseudotuberculosis O-antigen may be down-regulated, revealing Ail for cell-binding and matrixbinding activity [45,49].Additionally, wild-type Y.pseudotuberculosis has a number of other adhesins to facilitate cell binding, including YadA, invasin and pH 6 antigen.
We previously reported that the extracellular matrix component, fibronectin, is a host cell substrate for Y. pestis Ail [12].We hypothesized the defect observed in E. coli host cell binding by the single mutants as well as Y. pseudotuberculosis YPIII Ail, was due a lower affinity for fibronectin.Upon binding to immobilized fibronectin Ail-E43D, Ail-F126V, and Y. pseudotuberculosis Ail had slightly reduced binding to fibronectin as compared to Y. pestis Ail at 10 mg/ml and 20 mg/ml fibronectin (Fig. 4; although the defect for the E43D mutant did not reach statistical significance; p = 0.13 at 10 mg/ml and p = 0.08 at 20 mg/ml.)It is unclear why defects in binding to purified fibronectin do not precisely reflect defects in host cell binding, although we have noted previously that Ail recognizes cell-deposited and assembled fibronectin matrices somewhat differently than immobilized purified fibronectin [12].
Y. pseudotuberculosis Ail was previously reported to lack adhesion and invasion ability to HEp-2 cells [30].The previous experiments were done similarly to our experiments in a heterologous E. coli expression system.The previously published YPIII Y. pseudotuberculosis Ail sequence from that study contained an additional T7I substitution that we did not observe in either of our Y.pseudotuberculosis strains, YPIII or IP2666, or in other Y. pseudotuberculosis sequenced strains (Fig. 1).Furthermore, upon generation of the Y. pseudotuberculosis Ail-T7I substitution, we found this particular protein (Ail-T7I/E43D/F126V) was unstable, potentially explaining the previous findings (Fig. 7).
Studies presented here indicate that various versions of Ail expressed by Y. pseudotuberculosis strains do have adhesion and invasion activity, but the adhesion and invasion capacity is less than Y. pestis Ail when expressed in E. coli (Fig. 3) or Y. pestis (Fig. 5).Future studies will be aimed at identifying residues that when substituted eliminate Ail adhesive activity in the natural context of Y. pestis.

Figure 1 .
Figure 1.Protein alignment of the various Ail proteins.Ail from Y. pestis strains KIM5 ands CO92 and five Y. pseudotuberculosis isolates, IP2666 and YPIII, are shown.The two variant amino acids; positions 43 and 126 are highlighted.The arrow indicates the predicted processed cleavage site.The putative extracellular loop regions are indicated with black lines.The alignment was performed with the MegAlign program from Lasergene.KIM5, YPIII and IP2666 ail were sequenced in our laboratory.PB1/+, IP31758 and IP32953 ail sequences were obtained from genome sequencing projects.YPIII (Yang et al) refers to the YPIII sequence reported previously[30].doi:10.1371/journal.pone.0083621.g001

Figure 2 .Figure 3 .
Figure 2. Y. pseudotuberculosis Ail exhibits reduced adhesion and invasion function in E. coli.(A) Cultured HEp-2 cells were infected with E. coli AAEC185 expressing various forms of Ail from the indicated Yersinia strains.The empty vector (pMMB207) serves as a negative control.Percent adhesion was calculated by dividing the number of cell-associated CFU by the total number of bacteria in the well and multiplying by 100.The adhesion of E. coli AAEC185 expressing KIM5 Ail was set to 100% (actual adhesion was 4.0%).(B) Ail expression levels were determined in whole cell extracts were separated by SDS-PAGE followed by anti-Ail Western blotting.(C) Invasion assays were performed similarly to adhesion assays except infected cells were treated with gentamicin to kill external bacteria.Percent invasion was normalized to 100% for E. coli+KIM5 Ail (actual invasion was ,0.05%).Data are from two independent experiments performed in triplicate (n = 6).*p,0.02;**p,0.00002.Significance was calculated using the Student t test.doi:10.1371/journal.pone.0083621.g002

Figure 4 .
Figure 4. Y. pseudotuberculosis YPIII Ail is attenuated for binding to purified fibronectin relative to Y. pestis Ail.Purified plasma fibronectin was immobilized on 96-well plates.E. coli AAEC185 derivatives expressing the indicated Ail derivatives were added to wells and allowed to bind at 37uC.Bound bacteria were stained with 0.01% crystal violet.Stained bacterial cells were solubilized and the plates were read at ABS 595 .Shown is representative data from two independent experiments done in triplicate (n = 6).*p,0.05 compared to E. coli expressing pMMB207-ail KIM5.doi:10.1371/journal.pone.0083621.g004

Figure 5 .
Figure 5. Y. pseudotuberculosis Ail and the Y. pestis Ail single mutants have defects in HEp-2 cell adhesion and invasion relative to Y. pestis Ail, when expressed in Y. pestis.A) When expressed in the KIM5 DailDpla strain, Ail-E43D and Ail-F126V have significantly lower levels of cell adhesion to HEp2-cells than KIM5 Ail; MOI = 50 bacteria/cell.Actual adhesion by Y. pestis Ail was ,6%, n = 6.B) Levels of HEp-2 cell invasion by KIM5 DailDpla expressing Ail-E43D, Ail-F126V and YPIII Ail, were also less than that by KIM5 Ail, although the Ail-F126V result had a P value of 0.067, rather than 0.05, MOI = 50.Invasion refers to the number of Y. pestis surviving gentamicin

Figure 6 .
Figure 6.Ail has no adhesive or invasive activity when express in Y. pseudotuberculosis.Various Ail derivatives were expressed in the Y. pseudotuberculosis strain YP18 P-that lacks expression of numerous Y. pseudotuberculosis adhesins including: invasin, pH 6 antigen, YadA and Ail.All four Ail derivatives, including KIM5 Ail, failed to conferred adhesive (A) or invasive (B) ability on YP18 P-.YP18 P-strains were induced overnight with 0.5 mM IPTG to induce Ail expression and then diluted 1:50 into fresh media with 0.5 mM IPTG for 3 hrs prior to infecting cells.Ail levels were assessed by Coomassie gel straining.Assays were performed at an MOI of 50 bacterial/cell, n = 6.doi:10.1371/journal.pone.0083621.g006

Figure 7 .
Figure 7. Y. pseudotuberculosis Ail containing the T7I mutation is unstable.HEp-2 cells were infected with E. coli AAEC185 expressing the Y. pestis KIM5 Ail, Y. pseudotuberculosis Ail, and Y. pseudotuberculosis Ail harboring a T7I substitution.Percent adhesion was calculated by dividing the number of cell-associated CFU by the total number of bacteria in the well and multiplying by 100.HEp-2 adhesion average = 4.7%.Ail expression levels were determined in whole cell extracts separated by SDS-PAGE followed by anti-Ail Western blotting.Data are from two independent experiments performed in triplicate (n = 6).doi:10.1371/journal.pone.0083621.g007