Thr kinase activity of the Yersinia protein kinase A (YpkA) is necessary for full virulence in the mouse, mollifying phagocytes, and disrupting the eukaryotic cytoskeleton

Microbial Pathogenesis 40 (2006) 234–243 www.elsevier.com/locate/micpath

The Ser/Thr kinase activity of the Yersinia protein kinase A (YpkA) is necessary for full virulence in the mouse, mollifying phagocytes, and disrupting the eukaryotic cytoskeleton David J. Wiley a, Roland Nordfeldth b, Jason Rosenzweig a, Christopher J. DaFonseca a, Richard Gustin a, Hans Wolf-Watz b, Kurt Schesser a,c,* a

Department of Microbiology and Immunology, University of Miami School of Medicine, Miami, FL, USA b Department of Molecular Biology, Umea˚ University, Umea˚, Sweden c Immunology Section, Department of Cell and Molecular Biology, Lund University, Lund, Sweden Received 22 July 2005; received in revised form 20 December 2005; accepted 10 February 2006

Abstract The Yersinia protein kinase A (YpkA) is injected into host cells by the yersinial type three secretion system (TTSS). YpkA is widely believed to function within the host cell based on the fact that its kinase domain is clearly homologous to eukaryotic Ser/Thr kinases and that its enzymatic activity, when assayed in vitro, is dependent on eukaryotic-derived host factors. Whether this activity is required for virulence has not been addressed. Here, we report that a Yersinia pseudotuberculosis strain expressing a kinase-inactive YpkAD270A variant is greatly attenuated in the mouse model of infection compared to the isogenic wild-type strain. The ypkAD270A mutant strain was likewise attenuated in a cell culture infection assay indicating that the kinase activity of YpkA enhances the viability of host cell-associated bacteria. To begin to understand what cellular activities are targeted, we expressed YpkA and its variants in two different yeast model systems. In agreement with previous studies, we found that when rapidly induced and expressed at high levels in Saccharomyces cerevisiae, YpkA-mediated toxicity occurred extremely swiftly. Under these conditions toxicity was dependent on the structurally distinct GTPase-binding domain of YpkA and was entirely independent of its kinase activity. Therefore, to probe for kinase-dependent effects we expressed YpkA and its kinase-inactive variant at comparatively moderate levels in the fission yeast Schizosaccharomyces pombe. S. pombe is particularly well suited for actin cytoskeletal studies due to its easily quantifiable, well defined pattern of actin localization. S. pombe transformed with a wild-type YpkA-encoding plasmid displayed a pronounced actin mislocalization phenotype, the severity of which was directly proportional to the level of YpkA expressed in the cell. In cells expressing the kinase-inactive YpkA variant, the degree of actin mislocalization was reduced, but not entirely abrogated, suggesting that YpkA affects the eukaryotic cytoskeleton through kinase-dependent and kinase-independent mechanisms. Collectively, our yeast-derived results show how critical expression levels and exposure periods are for assaying virulence factor activities in heterologous model systems. More generally, our finding that the ‘eukaryotic-like’ kinase domain of YpkA is important for virulence illustrates how a bacterium can utilize a host-like factor or activity in order to enhance its survival following host cell contact. q 2006 Elsevier Ltd. All rights reserved. Keywords: Pathogenesis; Yersinia; YpkA; Kinase; Fission yeast

1. Introduction Type three secretion systems (TTSSs) are found in several species of plant- and animal-interacting Gram* Corresponding author. Address: Department of Microbiology and Immunology, University of Miami School of Medicine, P.O. Box 016960 (R-138), Miami, FL 33101, USA. Tel.: C1 305 243 4760; fax: C1 305 243 4623. E-mail address: [email protected] (K. Schesser).

0882-4010/$ - see front matter q 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.micpath.2006.02.001

negative bacteria and have been shown to play important roles in the way these various bacteria interact with their hosts [1]. The pathogenic Yersiniae (Yersinia pseudotuberculosis, Yersinia pestis, and Yersinia enterocolitica) utilize TTSSs to inject 5 or 6 effector proteins, designated as Yersinia outer membrane proteins or Yops, directly into host cells [2]. Since, the Yops are delivered directly into host cells by surface-attached bacteria, their respective contributions to virulence are believed to be due to directly modulating the physiology of the infected cell. One relatively well-studied aspect of Yop biology is how they affect the cytoskeleton of the host cell. This activity is

D.J. Wiley et al. / Microbial Pathogenesis 40 (2006) 234–243

mediated by the GTPase activating protein YopE, the protein tyrosine phosphatase YopH, and the protease YopT (yopT is absent from the Y. pseudotuberculosis genome) [2]. An additional Yop that has been shown to affect the host cell cytoskeleton, the Yersinia protein kinase A (YpkA; YopO in Y. enterocolitica) possess two easily recognizable domains. Residues 136–408 exhibit extensive sequence similarity to eukaryotic Ser/Thr kinases of the PKA family, while the carboxyl-terminal region consists of sequences resembling GTPase-binding modules [3–6]. These latter sequences do indeed mediate the binding of YpkA to the small GTPases RhoA and Rac, and in an infection model in which YpkA is overexpressed in the absence of the other Yops, this binding activity is necessary for YpkA-dependent cytoskeletal disruption of the host cell [4]. A Y. pseudotuberculosis mutant strain encoding a YpkA variant possessing a large deletion within its kinase domain (YpkAD207–388) is greatly attenuated in the mouse model of infection [3]. Although YpkAD207–388 is stably expressed and secreted by this latter strain, it is unclear whether it is in fact translocated into host cells, thus complicating the interpretation of the role of the kinase domain. Here, we directly test whether the kinase activity of YpkA is required for virulence and bacterial survival following host cell contact. We also examine the relationship between YpkA biochemical activities and its cellular activity in two different yeast model systems. 2. Results 2.1. YpkA enzymatic activity and Yersinia–host interactions A D270A replacement within the catalytic core of YpkA completely abolishes its in vitro kinase activity without affecting either its expression or competency to be translocated into host cells by Yersinia’s type three secretion system [4]. We constructed a Y. pseudotuberculosis ypkAD270A mutant strain by allelic replacement and tested this strain, designated as YPIII/pIB47, in both animal and cell culture models of infection. Intraperitoneal infections of male C57 bl/6J mice were performed using various doses of either the wild-type Y. pseudotuberculosis YPIII/pIB102 strain, the previously described ypkAD207–388 mutant strain YPIII/pIB44 (Section 1), or the ypkAD270A mutant strain YPIII/pIB47. Using these infection conditions, the differences in the 50% lethal doses (LD50) between the wild-type and the ypkA mutant strains were 2–3 orders of magnitude (Table 1). These data demonstrate that

235

the Ser/Thr kinase activity of YpkA plays a significant role in enabling Y. pseudotuberculosis to establish a systemic infection in the mouse. Using a cell culture infection assay, we previously showed that the YopE and YopH virulence determinants (so defined based on their disease-promoting properties in the mouse) enhance the ability of Y. pseudotuberculosis to withstand the antimicrobial killing activity of host cells [7]. To determine whether YpkA similarly enhanced the ability of Yersinia to survive and proliferate vis-a`-vis host cells, we infected cultured mouse macrophages with either the wild-type, DyopB, ypkAD207–388, or the ypkAD270A strains and measured the number of viable cell-associated bacteria at various times following the removal of unattached bacteria (tZ0 h). Similar to what has been shown previously [7], there was a 28-fold increase in the number of viable wild-type bacteria recovered after 6 h compared to the number of viable bacteria recovered at the beginning of the infection (Fig. 1). In contrast, there was only a 3-fold increase of the DyopB strain during the 6 h infection period. The DyopB mutant strain expresses and exports the Yop effectors (YopE, YopH, YopM, YopT, YopJ, and YpkA) from the bacterial cell but fails to inject these effectors into the host cell [8]. During the 6 h infection period, the number of recovered ypkAD207–388 and ypkAD270A bacteria increased 9and 11-fold, respectively. Considerably, fewer viable DyopB, ypkAD207–388, and ypkAD270A bacteria were recovered early in the experiment compared to the wild-type strain despite all strains initially adhering similarly to the host cells (compare 0 and 2 h recoveries). The wild-type, DyopB, ypkAD207–388, and ypkAD270A strains displayed similar proliferation rates in tissue culture media alone (Fig. 1, insert). These data show that a catalytically active YpkA enhances the ability of the bacterium to resist the killing activity of macrophages. Although their modes of transmission differ [9], Y. pseudotuberculosis and Y. pestis are thought to have diverged recently, on the order of just a few thousand years ago [10]. Within the ypkA ORF, there are two silent and two missense differences between the species. To determine whether the Y. pestis TTSS can be assayed in a cell culture infection assay, we tested the relative proliferation rates of the Y. pestis wild-type (KIM5-3001) and isogenic DyopB and DypkA mutant strains using conditions similar to what was used for Y. pseudotuberculosis described above. The number of wild-type Y. pestis bacteria recovered following an 8 h infection period was 14-fold higher than the number of wild-type bacteria recovered at the beginning of the infection period (not shown). In contrast, there was only a 1.4-fold increase in the number of DyopB bacteria

Table 1 Virulence of wild-type and ypkA mutant strains in mice* Strains YPIII/pIB102 (wild type) YPIII/pIB44 (ypkAD207–388) YPIII/pIB47 (ypkAD270A)

Range of doses (number of groups) 1

5

7.0!10 –7.0!10 (5) 1.0!106–1.0!107 (2) 5.0!103–5.0!106 (4)

LD50 1.0!103 3.1!106 2.2!105

Groups of 5 male C57bl/6J mice were injected intraperitoneally with the indicated strain of Y. pseudotuberculosis and the 50% lethal dose (LD50) values were calculated by the method of Reed and Muench [30].

236

D.J. Wiley et al. / Microbial Pathogenesis 40 (2006) 234–243

Fig. 1. Evaluating the proliferation of various Y. pseudotuberculosis strains in a cell culture infection assay. The indicated strains were added to wells containing macrophage-like RAW 267 cells at a multiplicity of infection (MOI) of 1, and 30 min later unattached bacteria were removed. At various times thereafter, the number of viable cell-associated bacteria per well was determined by plating. Shown graphically is the actual number of colonyforming units (CFU) recovered from two independent wells per time point. Growth rates of the various Y. pseudotuberculosis strains in tissue culture media (in the absence of macrophages) were assessed by viable plating and the relative increases were plotted (insert).

recovered over the same time span. The number of DypkA bacteria recovered after an 8 h infection period increased 3.9fold. An additional independently derived Y. pestis mutant strain, in which the entire ypkA–yopJ operon was deleted, behaved similarly to the DypkA mutant strain in the cell culture infection assay (not shown). Disrupting either the yopB or ypkA loci in Y. pestis did not detectably affect the number of bacteria recovered following the 30 min attachment period, indicating that the corresponding gene products are not involved in bacterial adherence to host cells. These data indicate that, at least in the context of a cell culture infection assay, YpkA is likely performing similar duties for Y. pestis and Y. pseudotuberculosis. 2.2. YpkA activity in Saccharomyces cerevisiae Lesser and Miller [11] have previously shown that YpkA expression in S. cerevisiae results in a growth suppression phenotype on semi-solid media. We initially utilized the S. cerevisiae system developed by Lesser and Miller that results in high-level and relatively rapid YpkA expression. To monitored the induction of the GAL1 promoter (which controls test gene expression in our S. cerevisiae expression plasmid), we placed the gene encoding the green fluorescent protein (GFP) under the control of the GAL1 promoter and followed GFP expression by flow cytometry. The GAL1–yEGFP transformant was propagated in glucose-free raffinose-containing media and GFP expression was induced by simply adding galactose to the culture medium. Under these conditions there is a rapid and nearly complete (w90%) test gene induction in terms of the fraction of cells displaying increased fluorescence (relative mean fluorescence intensities: 0 h, 6; 1 h, 137; 2 h, 257; 3 h, 330).

We cloned various YpkA-encoding gene fragments into the expression plasmid described above and ypkA-transformed S. cerevisiae strains were analyzed in a viability assay. The transformants containing either the empty vector control, or the YpkA-, YpkAD270A-, and YpkAD543–640-encoding plasmids were similarly cultured as the GFP-expressing strain described above and analyzed by optical density and viability plating (YpkAD543–640 is deficient for GTPase binding [4]). The optical density of the four cultures did not appreciably differ until around 4 h following induction of the test genes, after which time the optical densities of the YpkA- and YpkAD270Aexpressing yeast cultures started to level off relative to those of the empty vector control and YpkAD543–640-expressing cultures (Fig. 2). An even more striking difference between these cultures was observed in terms of their viability. The titer of viable cells in the empty vector control and YpkAD543–640expressing cultures displayed similar increases following induction in stark contrast to the YpkA- and YpkAD270Aexpressing cultures in which the titer of viable cells actually started to decrease w3 h following induction and continued decreasing over the next few hours. These results are consistent with a recent report by Nejedlik and colleagues showing that the shorter-term toxicity effects of YpkA are independent of its kinase activity and instead are mediated by it carboxyl-terminal sequences [12]. 2.3. YpkA activity in Schizosaccharomyces pombe Our animal and cell culture infection data strongly indicates that kinase-active YpkA possesses an anti-host activity. The results shown above for S. cerevisiae, as well as our previous data using an infection assay in which YpkA is overexpressed, suggest that at elevated expression levels the kinase domaindependent effects of YpkA are masked by its GTPase-binding activities. To achieve more subtle YpkA expression levels we employed the fission yeast S. pombe using inducible gene promoters of different strengths. YpkA and its kinase-inactive variant were cloned as GFP hybrid proteins in plasmids under the control of either the thiamine-repressible wild-type nmt1 promoter or a partially disabled nmt1 variant that results in reduced test gene expression. We used flow cytometry to follow the inductive and steady-state GFP–YpkAp levels in individual S. pombe cells (proteins expressed in S. pombe are by convention designated with a ‘p’). Cultures were maintained in logarithmic phase and test gene expression was induced by removing thiamine from the media. In cultures of S. pombe transformed with the lower expressed GFP–YpkAp-encoding plasmid (GFP–YpkAp(low)), cells displaying increased fluorescence started to be detected after 16.5 h of induction and by 20.5 h of induction these fluorescence cells constituted a distinct subpopulation (Fig. 3A; second column). In cultures transformed with the higher expressed GFP–YpkAp-encoding plasmid (GFP–YpkAp(High)), cells with increased fluorescence started to be detected after 12.5 h of induction and formed a substantial subpopulation by 16.5 h of induction (Fig. 3A; fourth column). In this latter culture we noted that a subpopulation of cells with distinct light scattering properties

D.J. Wiley et al. / Microbial Pathogenesis 40 (2006) 234–243

237

Fig. 2. YpkA activity in Saccharomyces cerevisiae. Transformants possessing the indicated YpkA expression plasmid were propagated in raffinose-containing media and YpkA expression was induced by the addition of galactose to the growth medium. At various times following induction (0 h), the OD600 (line plots) and viable cell titer (bars) of the cultures were recorded. The data presented represents the average of 6 independent cultures and is normalized to the 0 h values, which were OD600z0.20 and 2!106 CFU/ml.

increased during the course of the experiment (Fig. 3A; third column). These cells displayed reduced forward light scattering and increased side light scattering (shown within the R2 region of the light scattering dot plot) and based on their permeability to propidium iodine (not shown) had likely succumbed to YpkA toxicity. Increased levels of side light scattering and propidium iodine staining are both associated with dead or dying cells. The fact that these latter cells displayed relatively low levels of fluorescence is likely due to cytoplasmic leakage commonly observed in microscopic studies of S. pombe cells expressing high levels of GFP–YpkAp (see below). GFP–YpkAp and GFP– YpkA(D270A)p were expressed at comparable levels in S. pombe as analyzed by immunoblotting and flow cytometry (not shown). These data show that by using the nmt1 promotercontaining expression system we can achieve variable YpkA protein levels and/or induction kinetics in S. pombe. In parallel with the flow cytometric analysis shown in Fig. 3A, we assessed the growth of the S. pombe cultures by following their optical density. The growth profiles of cultures expressing lower levels of GFP–YpkAp or GFP–YpkA(D270A)p did not substantially differ from that of control cultures (Fig. 3B). When considered with the expression profile shown in Fig. 3A (left column), these results show that within the 12–20 h post-induction detection window, GFP–YpkAp and its kinase-inactive variant, do not achieve levels that are sufficient to alter the bulk growth of the culture. In contrast, the growth of cultures expressing relatively higher levels of GFP– YpkAp or GFP–YpkA(D270A)p started to diverge from the

control cultures (uninduced and GFP only) after w14 h of induction. From the expression profile shown in Fig. 3A (right column), there was a substantial increase in the fraction of cells possessing either heighten fluorescence or light scattering (indicative of GFP–YpkAp levels and death, respectively) during the 12–16 h post-induction window. Taken together these analyses allow us to correlate YpkA expression levels with bulk growth kinetics. A number of studies have shown that YpkA affects the eukaryotic cytoskeleton [4,6,13,14]. S. pombe is well suited for morphological studies due to its cylindrical shape and well defined cytoskeletal system. We, therefore, microscopically examined S. pombe cells at a variety of time points following the induction of YpkA expression and found that w16 h was a sufficient amount of time to observe cytoskeletal effects in cells transformed with the lower expressed GFP–YpkAp-encoding plasmid. Based on the flow cytometric data shown in Fig. 3A (left columns), detectable amounts of GFP–YpkAp appear in these cells between 12.5 and 16.5 h following induction; we therefore felt that examining cells at 16.5 h was more or less comparable to the YpkA exposure period that occurs during a cell culture infection assay such as shown in Fig. 1. For illustrative purposes Fig. 4A shows surface features, nuclear content, GFP signal, and actin localization of S. pombe cells transformed with the high-expressing GFP–YpkAp-encoding plasmid (GFP–YpkAp(High)). In cells from uninduced cultures, cells have a three-dimensional appearance with actin preferentially localizing to the cell tips or, if the cell is in mitosis, along

238

D.J. Wiley et al. / Microbial Pathogenesis 40 (2006) 234–243

Fig. 3. YpkA expression in the fission yeast Schizosaccharomyces pombe. The S. pombe strain PN567 was transformed with plasmids encoding the indicated GFP– YpkAp hybrid proteins. Test gene expression was under the control of either the full-strength thiamine-repressible nmt1 promoter (‘High’) or a nmt1 promoter variant that displays reduced activity (‘Low’). Transformants were placed in inductive conditions (thiamine free media, ‘KT’) and at the indicated times either (A) the light scattering and fluorescence of individual cells were measured by flow cytometry, or (B) the bulk growth of the cultures were followed by measuring their optical densities. Light scattering graphs were generated by plotting the forward light scattering data on the x-axis (in linear units) and the side light scattering data on the y-axis (in log units). The fluorescence intensity of individual cells within the two distinct light scattering populations (based primarily on differential side light scatter) is plotted on the histogram shown immediately to the right of each light scattering plot. The thin trace representing the R1-gated cells (displaying low side light scatter) and the thick trace representing the R2-gated cells (displaying a relatively higher amount of side light scatter).

the septum (e.g. Fig. 4A, cell no. 1). In contrast, many of the cells from the induced cultures (16.5 h) were visibly damaged and/or display a flattened morphology devoid of nuclei, fluorescence (GFP), and in some cases, actin (cells nos. 7 and 8), while still other GFP–YpkAp-expressing cells from the same culture appear morphologically normal with respect to surface features and nuclear content; this heterogeneity is likely due, at least in part, to the normal variation in plasmid copy number observed in yeast populations. Among cells from the induced cultures we observed a wide range of fluorescence intensities, from undetectable to nearly saturable (e.g. cells nos. 12 and 10, respectively). Similar to what has been shown in infected vertebrate and S. cereversiae cells [11,14], GFP–YpkAp localizes to the plasma membrane in S. pombe cells. Likewise GFP–YpkA(D270A)p, but not GFP only, also preferentially localized to the plasma membrane (not shown). We antidotally noted that in fluorescent S. pombe GFP– YpkAp transformants (i.e. GFP positive cells) actin appeared to be more or less evenly distributed throughout the cell compared to cells from uninduced cultures in which actin preferentially localized to the cell tips during interphase (Fig. 4A; compare cells nos. 9–11 with nos. 1–5). To quantify this effect, we analyzed individual cells transformed with either the high- or low-expressing GFP–YpkAp-encoding plasmids. Only those

cells that met a number of criteria were included in this analysis, limiting our analysis to apparently ‘healthy’, mono-nuclear cells that stained positively for actin (Section 4). In cells transformed with the high-expressing GFP–YpkAp-encoding plasmid the ratio of the mean intensity of actin staining at the cell tips compared to the mean intensity of the whole cell actin was nearly 1.4 when the cells were cultured in non-inducing conditions (Fig. 4B; nZ32); this is similar to the ratio observed in untransformed S. pombe cells (not shown). In contrast, if these transformants were cultured for 16.5 h in conditions that are inductive for GFP–YpkAp expression (‘KT’), the tip/whole actin staining intensity ratio was reduced to below 1.1 (nZ204; P!0.001 between uninduced and induced cells). If the latter cells were subdivided based on the intensity of their GFP-based fluorescence signal (designated as GFPC and GFPCC for moderate and high fluorescence, respectively), the tip/whole actin staining intensity ratio in the GFPCC cells was less than 1 (Fig. 4B). These data show that YpkA disrupts the actin-based cytoskeletal system and that this effect likely plays a primary role for YpkA-mediated yeast suppressive activity. We also examined S. pombe cells expressing relatively low levels of either GFP–YpkAp or GFP–YpkA(D270A)p, such levels not associated with growth suppression (see Fig. 3B). In cells expressing low levels of GFP–YpkAp the tip/whole actin

D.J. Wiley et al. / Microbial Pathogenesis 40 (2006) 234–243

Fig. 4. Analysis of YpkA-expressing S. pombe cells. Transformants were propagated and GFP–YpkAp expression induced as in Fig. 3. (A) GFP– YpkA(High)-transformed cells from cultures either left uninduced (CT) or induced for 16.5 h (KT) were analyzed by differential contrast (DIC), nuclear content (DAPI), fluorescence (GFP), and actin content (ACTIN). The bar in the lower right panel is 5 mm. (B) Following 16.5 h of induction, actin localization was analyzed in individual cells meeting a number of criteria (Section 4) and recorded as the ratio between the mean intensity of actin staining localized to the cell tip compared to the mean intensity throughout the cell. (By the Student’s t-test, *ZP!0.001; see main text for number of cells analyzed per condition)

intensity staining ratio was w1.1 (nZ128), which was substantially less than the uninduced controls and was nearly as low as the tip/whole actin intensity staining ratio observed in cells expressing high levels of GFP–YpkAp (Fig. 4B). In contrast, the tip/whole actin intensity staining ratio in S. pombe cells expressing low levels of GFP–YpkA(D270A)p was greater than 1.2 (nZ198; P!0.001 between GFP–YpkAp(low)- and GFP–YpkA(D270A)p(low)-expressing cells) indicating that YpkAs kinase activity contributed to YpkAs detrimental effects on the S. pombe cytoskeleton. The fact that the tip/whole actin intensity staining ratio of GFP–YpkA(D270A)p-expressing cells was still less than the uninduced controls indicate that other activities of YpkA play a role in disrupting the host cell cytoskeleton and that these latter activities predominate when YpkA is expressed at high levels. 2.4. YpkA-mediated inhibition of host cell bacterial internalization Internalization assays, as used for the yersiniae, are sensitive to the events that occur immediately following

239

bacterium–host cell interaction. Therefore, we tested whether an intact kinase domain was necessary for YpkA to block host cell uptake of Y. pseudotuberculosis. Employing an antibiotic protection assay [15], we recovered almost 10-fold less wild-type Y. pseudotuberculosis from infected host cells in the presence of gentamicin compared to infections in the absence of gentamicin (Fig. 5). This result indicates that the majority of wild-type Y. pseudotuberculosis bacteria are exposed to the antibiotic and is consistent with the majority of the host cell-associated bacteria being located extracellularly. In contrast, the number of DyopB bacteria recovered from infected host cells was not affected by gentamicin indicating that in the absence of a functioning TTSS Yersinia is rapidly internalized by host cells. The relative differences in the degree of internalization between the wild type and TTSS mutant strains observed in this assay are consistent with previously described antibiotic-sensitivity and microscopy-based studies [8,16]. Similarly to what was observed with the DyopB mutant strain and consistent with the findings of Grosdent et al. [17], the number of ypkAD207-388 and DyopE mutant bacteria recovered from infected cells was not affected by the presence of gentamicin (Fig. 5 and data not shown) indicating that these mutant bacterial strains are rapidly internalized by host cells. Strikingly, in contrast to the ypkAD207-388 and DyopE mutant ‘knockout’ strains, the number of ypkAD270A bacteria recovered from infected host cells in the presence of gentamicin was substantially less than the number of ypkAD270A bacteria recovered from infections proceeding in the absence of gentamicin. This indicates that under these conditions ypkAD270A bacteria are, like wild-type bacteria, not rapidly internalized following their attachment to host cells.

Fig. 5. Evaluating internalization of Y. pseudotuberculosis by culture mouse macrophages. The indicated strains were added to wells containing RAW 267 cells at a MOI of 60. Following a 30 min attachment period, unattached bacteria were removed and the infected cells were incubated an additional hour in either tissue culture media alone or in tissue culture media containing gentamicin. ‘Fraction protected’ represents the average fold differences (of three independent wells per condition) between the number of colonies recovered in the presence of gentamicin compared to the number of colonies recovered in the absence of gentamicin. By the Student’s t-test, P!0.001 for the differences between the wild-type and either the DyopB strains or ypkAD207– 388 mutant strains and PZ0.155 for the differences between the wild-type and the ypkAD270A mutant strains.

240

D.J. Wiley et al. / Microbial Pathogenesis 40 (2006) 234–243

3. Discussion Five of the six yersinial TTSS effector proteins have recognizable enzymatic activities. The enzymatic activity of two of these effectors, the protein tyrosine phosphatase YopH and the GTPase-activating protein YopE, have been previously demonstrated to be required for full virulence in the mouse [18,19]. Here, we show that the enzymatic activity of the yersinial TTSS effector YpkA is likewise required for full virulence in the mouse (Table 1). In assaying ypkA mutant strains in a cell culture infection assay (Fig. 1), it appears that YpkA and its enzymatic activity enhances the viability of Y. pseudotuberculosis (and likely Y. pestis) vis-a` -vis the macrophage. These findings indicate that the kinase activity of YpkA serves to neutralize the antimicrobial killing functions of the host cell. Previous work showed that YpkA disrupts the host cell cytoskeleton [4,13,14]. One study found that YpkA-mediated disruption of actin filaments in ypkA-transfected vertebrate cells was partially dependent on the enzymatic activity of YpkA [13]. Our lab, on the other hand, found that in an infection model in which YpkA was overexpressed in a multiple yop mutant strain of Y. pseudotuberculosis, YpkA-mediated disruption of host cell morphology was independent of its enzymatic activity [4]. Although seemingly contradictory, the disparate outcome of these two studies likely results from differing expression levels of YpkA. This possibility is strengthened by our findings presented here of dose-dependent YpkA and YpkAD270A cellular effects. Initially we utilized a S. cerevisiae model that had been originally developed by Lesser and Miller [11]. Expressing ypkA from the relatively strong GAL promoter results in suppression of growth as well as greatly reduced viability (Fig. 2). In general agreement with the results reported by Nejedlik and colleagues [12], under these assay conditions the majority of the growth suppressive activity was dependent on domains in the carboxyl terminus of YpkA and independent of its kinase activity. Hence, this situation may be akin to the infection studies cited above in which overexpressed YpkA exerted a kinase-independent effect on host cell morphology [4]. Apparently when YpkA is present at high levels its kinasedependent cellular effects are masked by its structurally distinct GTPase-binding domain. In the hope of unmasking kinase-dependent YpkA cellular effects we turned to the fission yeast S. pombe, which, in addition to having a well-developed variable expression system, is also well-suited for cytological studies. In S. pombe, YpkA affected the trafficking and/or retention of F-actin to the growing tip(s) of each daughter cell prior to mitosis (Fig. 4). The fact that YpkA-expressing S. pombe cells generally maintained their shape suggests that YpkA derails specific processes instead of causing a global disruption of cellular physiology. The degree of actin mislocalization was reduced, but not entirely abrogated, in YpkAD270A-expressing S. pombe as compared to YpkA-expressing S. pombe. These results suggest that under these conditions YpkA employs two activities to disrupt the eukaryotic cytoskeleton and cause

growth suppression. When expressed at sufficiently high levels, YpkA does not require its protein kinase activity to disrupt S. pombe growth (Fig. 3). This situation is likely analogous to what we observed in our S. cerevisiae model that displays a dependency on an intact GTPase-binding domain. When expressed at relatively lower levels, the kinase domain of YpkA becomes unmasked and can be demonstrated to exert a negative effect on the distribution of actin in S. pombe cells (Fig. 4). At these expression levels the kinase and carboxyllocated domains of YpkA appear to have an additive effect of mislocalizing actin during the cell cycle. Whether these two domains target the same or distinct cellular pathways remains to be determined. Surprisingly, we found that the ypkAD270A strain was fully competent to block the host cell phagocytic program (Fig. 5). This result was unexpected since we had supposed that the reduced viability of the ypkAD270A strain in the longer-term 6 h infection assay shown in Fig. 1 was due to the ypkAD270A bacteria failing to block the host cell phagocytic program. Our results, however, do not exclude the possibility that the kinase activity of YpkA is required to sustain Yersinia’s antiphagocytic effect. Testing for relatively longer-term effects (O1–2 h) in terms of alterations of host morphology and/or levels of bacterial internalization is problematic when assaying strains that express a normal complement of Yop proteins. Primarily due to the activity of YopE, the morphology of Yersinia-infected cells becomes highly perturbed and hence the ability to discern between extracellular and internalized bacteria in such cases is exceedingly difficult. It remains to be determined why the Y. pseudotuberculosis ypkAD207–388 (Fig. 5) and Y. enterocolitica yopOD65–558 strains [3] are impaired in their antiphagocytic activities. One possibility is that YpkAD207-388, which is stably expressed and secreted [3], interferes to some degree with the translocation of other Yops. There have been several reported instances in which altering the Yop expression profile leads to spurious, or at the very least misleading, outcomes in cell culture infection assays [17,20,21]. In conclusion, we have demonstrated a clear relationship between the enzymatic activity of YpkA and virulence and offer evidence that this activity of YpkA enhances bacterial survival following host cell attachment. Furthermore, we show that at least one of the cellular consequences of this activity is directed towards disrupting the eukaryotic cytoskeleton. 4. Materials and methods 4.1. Bacterial and yeast strains Y. pseudotuberculosis strains used in this study include YPIII/pIB102 (wild type), YPIII/pIB604 (DyopB), and YPIII/ pIB44 (ypkAD207–388) [3,8,22]. The YPIII/pIB47 (ypkAD270A) strain was constructed by a suicide plasmid-based allelic exchange method [23] using YPIII/pIB102 as the host strain. The wild-type Y. pestis strain used was KIM5-3001 (pCD1, PlaC, pMT1) [24] and yopB mutant strain (provided by Greg Plano) was constructed in a KIM5-3001 as described by Day and co-workers [25]. The Y. pestis DypkA mutant strain was

D.J. Wiley et al. / Microbial Pathogenesis 40 (2006) 234–243

241

constructed in a KIM5-3001 background using the method of Datsenko and Wanner [26]. The Y. pestis wild-type, DyopB, and DypkA strains grew equally well in tissue culture media. The Saccharomyces cerevisiae haploid strains RDY84 (MATa leu2 ura3 can1 trp1 his3 gal2 scd1-v pdr1DKAN pdr3DHISC) was used as host strain for the experiment shown in Fig. 2. The S. pombe strain PN567 (h-ade6-704 ura4-D18 leu1-32) was obtained from F. Verde (Department of Molecular and Cellular Pharmacology, University of Miami School of Medicine) and was used as the parental strain for all fission yeast experiments.

antibiotic-free RPMI1640/10% FCS (except in Fig. 5 in which some of the wells received RPMI1640/10% FCSC 10 mg/ml gentamicin). At the indicated time following the removal of the unattached bacteria, the overlaying media again was carefully removed and 0.5 ml of distilled water was added to the wells to lyse the eukaryotic cells. The resulting lysate was appropriately diluted, spread on LB plates, and incubated at 26 8C for 2 days at which time the number of colony forming units (cfus) were determined manually.

4.2. Plasmids and DNA methods

4.4. Yeast methods

For YpkA expression in S. cerevisiae (Fig. 2), ypkA sequences were generated by PCR using oligonucleotide primers flanked by attB sites using either pEG2 (a.k.a. pYpkA [3]) or its derivatives pYpkAD270A and pYpkAD543–640 as templates [4]. attB-flanked PCR products were cloned via recombination into pDONR221 (Invitrogen, Carlsbad, California) thereby generating ‘entry clones’. Following verification of gene integrity by sequencing, ypkA sequences were recombined into pYES-DEST52 (Invitrogen). pYES-DEST52-yEGFP was generated by amplifying GFP-encoding sequences from pKT209 [27] using attB-flanked oligonucleotide primers followed by recombination-based cloning as described above. pRep3X and pRep41X, containing either a relatively strong (wild type) or low (partially disabled) strength nmt1 inducible promoters, respectively [28], were used to express GFP–YpkA hybrid proteins in S. pombe (Figs. 3 and 4). pRep3X yEGFP– YpkA and pRep41X yEGFP–YpkA (encoding GFP–YpkA(High) and GFP–YpkA(Low), respectively) were constructed by first cloning a PCR-generated ypkA fragment into pRep3X and pRep41X at the XhoI and SmaI sites, followed by the insertion of yEGFP-encoding fragments, amplified from pKT209 [27], into the XhoI site. pRep3X yEGFP–YpkAD270A and pRep41X yEGFP–YpkAD270A (encoding GFP–YpkA(D270A)p(High) and GFP–YpkA(D270A)p(Low), respectively) were derived from their wild-type counterparts using Quickchange II XL (Stratagene).

S. cerevisiae strains were propagated in 0.67% nitrogen base (Sigma Y0626) supplemented with adenine, amino acids and raffinose, and plasmids were maintained by nutritional selection. Transformants were grown to saturation, diluted 10-fold with fresh raffinose-containing media, and propagated for 3 h. At this point, expression of test genes were then induced by adding galactose to a final concentration of 2%. GFP levels in single cells were determined using a FacScan 488 flow cytometry (BD Biosciences), and the resulting data was analyzed with the Cellquest Pro software (BD Biosciences). In the experiment shown in Fig. 2 the cultures were dispersed into 96-well tissue culture plates immediately following the addition of galactose and incubated at 30 8C without shaking. At the indicated times clumped yeast were resuspended with a multi-channel pipette and the OD600 recorded with an automated plate reader and viable cell titer determined by plating. S. pombe strains were cultured in either YES media (Qbiogene) or minimal EMM media (Qbiogene) plus required supplements. Cultures were grown for at least eight generations prior to the start of each experiment. Exponentially growing cells were then washed twice with sterile water to remove thiamine and then resuspended in minimal media plus supplements. The cells were diluted to allow exponential growth for 12 h to a final OD595 of 0.2. At 11 h minus thiamine (start of experiment), the cells were diluted to an OD595 of 0.08 and allowed to continue to grow. The first experimental measurement was taken at 12.5 h and then continued every hour to monitor cell growth. From start to finish the cells were grown in minimal media with the required supplements in a shaking water bath at 32 8C. For light scattering properties and fluorescence levels in single cells (Fig. 3A), samples were removed during the growth assay at the indicated time points and analyzed by flow cytometry using a FacScan 488 flow cytometry (BD Biosciences). Collected data was evaluated using WinMDI 2.8 (http://facs.scripps.edu/software.html). The experiment shown in Fig. 3B was repeated three times and the average was plotted. For microscopy (Fig. 4), cultures were propagated as described above except that the OD595 was never allowed to exceed 0.2. At 16.5 h following induction, cells were collected and resuspended in 10 ml of freshly prepared solution of 40% formaldehyde (16%)C60% PM (35 mM potassium phosphate, 0.5 mM MgSO4, pH 6.8) for 5–6 min at room temperature.

4.3. Infection assays Mice were infected as described in Table 1 using the method of Rosqvist and co-workers [29]. The method of Bartra and co-workers [7], with some minor modifications, was used for cell culture infection assays. Bacterial cultures were grown to saturation by vigorous shaking at 26 8C in either Luria broth (Y. pseudotuberculosis) or RPMI1640/ 10% fetal calf serum (FCS) (Y. pestis). Saturated cultures were diluted appropriately (depending on the MOI as indicated in the figure legends) in a final volume of 0.4 ml of RPMI1640/10% FCS and added to 2!105 RAW mouse macrophage-like cells in 24-well plates that had previously been washed free of antibiotics. Infected cells were incubated at 37 8C for 30 min at which time the overlaying media (containing the unattached bacteria) was carefully removed and replaced with 0.4 ml of fresh

242

D.J. Wiley et al. / Microbial Pathogenesis 40 (2006) 234–243

Cells were then washed in 10 ml of PM, resuspended in 1 ml of PM, transferred to a microfuge tube, and then washed in another ml of PM. The cells were then permeabilized by treating the cells in PMC1% Triton for 2 min. Cells were then resuspended in PMC1% BSA (1 ml) for 5 min. This step was repeated two more times with an additional 5 min incubation each consecutive time. After the final wash the cells were resuspended in as little volume of PMC1% BSA as possible. Next 5 ml of BODIPYw 558/568 phalloidin (Molecular Probes) was added to the cells, the tubes were wrapped in aluminum foil and placed on a rotator for 60 min at room temperature. Cells were then washed with PM and resuspended in 50 ml of PM and then subjected to microscopic analysis in the presence of DAPI (4 0 ,6-diamidino-2-phenylindole; SIGMA) and Slowfade Antifade (Molecular Probes). Images were captured using an Olympus BX61 microscope equipped with Slidebook 4.0 software (Intelligent Imaging Innovations). Single layer images were captured for DIC and DAPI, and multi-layers were captured and subsequently deconvoluted for GFP and actin. In order to be used for actin analysis the cells were first required to pass the following criteria: (i) the cells must appear ‘healthy’ in the DIC images, meaning that they have a 3D, cylindrical appearance (as opposed to flatten) with a uniform cell wall, (ii) cells must be mono-nuclear as observed in the DAPI straining, indicating that the cells were in G1/S/G2 versus M phase when actin localizes to the septum, and (iii) the cells must be positively stained for actin. Cells that meet the criteria were then measured for mean GFP levels using Image J software provided by the NIH (http://rsb.info.nih.gov/ij). In addition, the whole cell mean actin levels were measured in parallel to the mean level of actin that was localized to the cell tips (approximately 2 mm from the distal end). For all measurements, background was subtracted from the values. The ratio of the mean values of actin at the cell tips versus the whole cell was then calculated. From previous experiments using these conditions this ratio is known to be w1.4. Acknowledgements We thank Blandine Ge´ry, Brian Dizon, Greg Plano, Tomas Leanderson, Fulvia Verde, Ulrich von Pawel-Rammingen, Kurt Thorn, and Sara Schesser Bartra, for their support, advice, and most of all, friendliness. Supported by the Swedish Medical Research Council, the Swedish Foundation of Strategic Research, Active Biotech (Lund, Sweden), the Glaser Foundation (Miami, FL), the Department of Microbiology and Immunology, University of Miami School of Medicine, and Public Health Service grant AI53459 from the National Institute of Allergy and Infectious Diseases. References [1] Gala´n JE, Collmer A. Type III secretion machines: bacterial devices for protein delivery into host cells. Science 1999;284:1322–8. [2] Cornelis GR. Yersinia type III secretion: send in the effectors. J Cell Biol 2002;158:401–8.

˚ , Wolf-Watz H. A secreted protein [3] Galyov EE, Ha˚kansson S, Forsberg A kinase of Yersinia pseudotuberculosis is an indispensable virulence determinant. Nature 1993;361:730–1. ˚ kerstro˝m B, Wolf[4] Dukuzumuremyi J-M, Rosqvist R, Hallberg B, A Watz H, Schesser K. The Yersinia protein kinase A is a host factor inducible RhoA/Rac-binding virulence factor. J Biol Chem 2000;275: 35281–90. [5] Maesaki R, Ihara K, Shimizu T, Kuroda S, Kaibuchi K, Hakoshima T. The structural basis of Rho effector recognition revealed by the crystal structure of human RhoA complexed with the effector domain of PKN/PRK1. Mol Cell 1999;4:793–803. [6] Barz C, Abahji TN, Trulzsch K, Heesemann J. The Yersinia Ser/Thr protein kinase YpkA/YopO directly interacts with the small GTPases RhoA and Rac-1. FEBS Lett 2000;482:139–43. ˚ , Schesser K. The Yersinia YopE and [7] Bartra S, Cherepanov P, Forsberg A YopH type III effector proteins enhance bacterial proliferation following contact with eukaryotic cells. BMC Microbiol 2001;1:22–32. [8] Ha˚kansson S, Schesser K, Persson C, Galyov EE, Homble F, WolfWatz H. The YopB protein of Yersinia pseudotuberculosis is essential for the translocation of Yop effector proteins across the target cell plasma membrane and displays a contact-dependent membrane disrupting activity. EMBO J 1996;15:5812–23. [9] Hinnebusch BJ. Bubonic plague: a molecular genetic case history of the emergence of an infectious disease. J Mol Med 1997;75:645–52. [10] Achtman M, Zurth K, Morelli G, Torrea G, Guiyoule A, Carniel E. Yersinia pestis, the cause of plague, is a recently emerged clone of Yersinia pseudotuberculosis. Proc Natl Acad Sci USA 1999;96:14043–8. [11] Lesser CF, Miller SL. Expression of microbial virulence proteins in Saccharomyces cerevisiae models mammalian infection. EMBO J 2001; 17:1840–9. [12] Nejedlik L, Pierfelice T, Geiser JR. Actin distribution is disrupted upon expression of Yersinia YopO/YpkA in yeast. Yeast 2004;21:759–68. [13] Juris SJ, Rudolph AE, Huddler D, Orth K, Dixon JE. A distinctive role for the Yersinia protein kinase: actin binding, kinase activation, and cytoskeleton disruption. Proc Natl Acad Sci USA 2000;97:9431–6. [14] Ha˚kansson S, Galyov EE, Rosqvist R, Wolf-Watz H. The Yersinia YpkA Ser/Thr kinase is translocated and subsequently targeted to the inner surface of the HeLa cell plasma membrane. Mol Microbiol 1996;20: 593–603. [15] Leong JM, Morrissey PE, Marra A, Isberg RR. An aspartate residue of the Yersinia pseudotuberculosis invasin protein that is critical for integrin binding. EMBO J 1995;14:422–31. [16] Mecsas J, Bilis I, Falkow S. Identification of attenuated Yersinia pseudotuberculosis strains and characterization of an orogastric infection in BALB/c mice on day 5 postinfection by signature-tagged mutagenesis. Infect Immun 2001;69:2779–87. [17] Grosdent N, Maridonneau-Parini I, Sory MP, Cornelis GR. Role of Yops and adhesins in resistance of Yersinia enterocolitica to phagocytosis. Infect Immun 2002;70:4165–76. [18] Black DS, Bliska JB. The RhoGAP activity of the Yersinia pseudotuberculosis cytotoxin YopE is required for antiphagocytic function and virulence. Mol Microbiol 2000;37:515–27. [19] Bliska JB, Guan KL, Dixon JE, Falkow S. Tyrosine phosphate hydrolysis of host proteins by an essential Yersinia virulence determinant. Proc Natl Acad Sci USA 1991;88:1187–91. [20] Schesser K, Frithz-Lindsten E, Wolf-Watz H. Delineation and mutational analysis of the Yersinia pseudotuberculosis YopE domains which mediate translocation across bacterial and eukaryotic cellular membranes. J Bacteriol 1996;178:7227–33. [21] Wong K-W, Isberg RR. Yersinia pseudotuberculosis spatially controls activation and misregulation of host cell Rac1. PLoS Pathogens 2005;1: e16. [22] Bo¨lin I, Wolf-Watz H. Molecular cloning of the temperature-inducible outer membrane protein 1 of Yersinia pseudotuberculosis. Infect Immun 1984;43:72–8. [23] Milton DL, O’Toole R, Horstedt P, Wolf-Watz H. Flagellin A is essential for the virulence of Vibrio anguillarum. J Bacteriol 1996;178:1310–9. [24] Lindler LE, Klempner MS, Straley SC. Yersinia pestis pH 6 antigen:

D.J. Wiley et al. / Microbial Pathogenesis 40 (2006) 234–243 genetic, biochemical, and virulence characterization of a protein involved in the pathogenesis of bubonic plague. Infect Immun 1990; 58:2569–77. [25] Day JB, Ferracci F, Plano GV. Translocation of YopE and YopN into eukaryotic cells by Yersinia pestis yopN, tyeA, sycN, yscB and lcrG deletion mutants measured using a phosphorylatable peptide tag and phosphospecific antibodies. Mol Microbiol 2003;47:807–23. [26] Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA 2000; 97:6640–5.

243

[27] Sheff MA, Thorn KS. Optimized cassettes for fluorescent protein tagging in Saccharomyces cerevisiae. Yeast 2004;21:661–70. [28] Forsburg SL, Sherman DA. General purpose tagging vectors for fission yeast. Gene 1997;191:191–5. ˚ , Rimpiainen M, Bergman T, Wolf-Watz H. The [29] Rosqvist R, Forsberg A cytotoxic protein YopE of Yersinia obstructs the primary host defence. Mol Microbiol 1990;4:657–67. [30] Reed LJ, Muench H. A simple method of estimating fifty percent endpoints. Am J Hyg 1938;27:493–7.