Delineation of exoenzyme S residues that mediate the interaction with 14-3-3 and its biological activity Lubna Yasmin 1, *, Anna L. Jansson 1, *, Tooba Panahandeh 1 , Ruth H. Palmer 3 , Matthew S. Francis 2 and Bengt Hallberg 1 1 Department of Medical Biosciences ⁄ Pathology, Umea ˚ University, Sweden 2 Department of Molecular Biology, Umea ˚ University, Sweden 3 Umea ˚ Center for Molecular Pathogenesis, Umea ˚ University, Sweden 14-3-3 proteins are a group of highly conserved intra- cellular dimeric molecules, expressed in plants, inverte- brates and higher eukaryotes, with no intrinsic activity. 14-3-3 proteins play an important role in several signa- ling pathways and 14-3-3 interacts with proteins in a phospho-specific manner, using a defined consensus- binding motif [1–3]. Several of these interacting part- ners have recognized functions, which include enzymes in biosynthetic metabolism, ion channels and regula- tors of growth in plants [4–6]. It has been shown that many human proteins can also bind directly to 14-3-3 in a phosphorylation-dependent manner, placing 14-3-3 as a central regulatory molecule in several physiological processes such as biosynthetic metabo- lism, cell proliferation, and survival in human cells [3,7,8]. Crystal structure analyses of the 14-3-3 dimer alone or in complex with peptides or native binding partners has revealed the presence of a basic cluster in the amphipathic groove of each monomer which mediates the interaction of 14-3-3 with the phospho-amino acid residues in its interaction partners. Therefore it is likely that each dimer contains two binding pockets and can interact with a single target or with multiple binding partners. Further, it has been observed that interaction between 14-3-3 proteins and its target partner(s) can Keywords ADP-ribosylation; coenzyme binding site; cytotoxicity; NAD-dependent; cystic fibrosis; Pseudomonas aeruginosa Correspondence B. Hallberg, Department of Medical Biosciences ⁄ Pathology, Building 6 M, 2nd floor, Umea ˚ University, 901 87 Umea ˚ , Sweden Fax: + 46 90 785 2829 Tel: + 46 90 785 2523 E-mail: Bengt.Hallberg@medbio.umu.se *Both authors contributed equally to this work. (Received 5 October 2005, revised 7 December 2005, accepted 12 December 2005) doi:10.1111/j.1742-4658.2005.05100.x 14-3-3 proteins belong to a family of conserved molecules expressed in all eukaryotic cells, which play an important role in a multitude of signaling pathways. 14-3-3 proteins bind to phosphoserine ⁄ phosphothreonine motifs in a sequence-specific manner. More than 200 14-3-3 binding partners have been found that are involved in cell cycle regulation, apoptosis, stress responses, cell metabolism and malignant transformation. A phos- phorylation-independent interaction has been reported to occur between 14-3-3 and a C-terminal domain within exoenzyme S (ExoS), a bacterial ADP-ribosyltransferase toxin from Pseudomonas aeruginosa. In this study, we have investigated the effect of amino acid mutations in this C-terminal domain of ExoS on ADP-ribosyltransferase activity and the 14-3-3 interac- tion. Our results suggest that leucine-428 of ExoS is the most critical resi- due for ExoS enzymatic activity, as cytotoxicity analysis reveals that substitution of this leucine significantly weakens the ability of ExoS to mediate cell death. Leucine-428 is also required for the ability of ExoS to modify the eukaryotic endogenous target Ras. Finally, single amino acid substitutions of positions 426–428 reduce the interaction potential of 14-3-3 with ExoS in vitro. Abbreviations ADPRT, ADP-ribosyltransferase; BD, binding domain; ExoS, exoenzyme S; FAS, factor activating exoenzyme S; GAP, GTPase-activating protein; GEF, guanine exchange factor; GTPase, GTP binding protein; Ras, rat sarcoma. 638 FEBS Journal 273 (2006) 638–646 ª 2006 The Authors Journal compilation ª 2006 FEBS occur outside the amphipathic groove, which probably contributes to a stable three-dimensional configuration with an opportunity for conformational modulation of the target [2,9–13]. 14-3-3 also interacts in a phos- phorylation independent manner with some proteins and peptides, such as exoenzyme (Exo) S of Pseudo- monas aeruginosa, p190RhoGEF and the R18 peptide inhibitor [14–17]. P. aeruginosa is an opportunistic pathogen that cau- ses acute infections mainly in immunocompromised individuals, such as children and patients with cystic fibrosis, burn wounds or leukemia [18]. The virulence toxin ExoS from P. aeruginosa is first secreted and then translocated from the bacteria into the eukaryotic cell via a bacterial encoded type III secretion system [19]. ExoS is a bifunctional toxin with an N-terminal Rho GTPase activating protein (GAP) activity [20,21] and a highly promiscuous C-terminally encoded ADP- ribosylation activity towards small GTPases [21–23]. Its function is dependent on interactions with 14-3-3 and factor activating ExoS (FAS) protein cofactors [24–26]. As this interaction is necessary for the ADP-ribosyla- tion activity of ExoS, and more intriguingly appears to be independent of phosphorylation [14,15,26], we wanted to define individual residues within the 14-3-3 binding domain of ExoS that are important for the 14-3-3 interaction, as well as the resultant activity in vivo. We have approached these questions using a strategy of single amino acid site-directed mutagenesis of the cofactor interaction domain within ExoS. Various single mutant ExoS proteins were tested for their capacity to interact with 14-3-3 and subsequently for their cytotoxicity and ADP-ribosylation potential using Ras as a substrate in vivo. We show that the leu- cine residue at position 428 is necessary for both the ADP-ribosylation activity and the cytotoxic action of ExoS in vivo. Result and discussion Acidic residues within the 14-3-3 binding domain of ExoS are not strictly needed for phosphorylation-independent binding The interaction between 14-3-3 and ExoS is important for the ADP ribosylation activity of ExoS and even more intriguingly, appears to be independent of serine- phosphorylation [15,26]. The amino acid sequence between 419 and 428 of ExoS is known to be import- ant for this interaction [14]. To address exactly which amino acid residues in the ExoS sequence S 419 QGLLDALDL 428 are critical for 14-3-3 binding, a set of single substitution mutants of ExoS were con- structed together with some additional variants (Table 1). These variant alleles were then fused to GST giving rise to the following fusion proteins: GST-ExoS(wt), GST-ExoS(SD), GST-ExoS(LDL426– 428AAA), GST-ExoS(DALDL424–428AAAAA), GST- ExoS(D424A; D427A), GST-ExoS(S419I), GST-Exo- Table 1. Summary of the various GST-fusion protein constructs of ExoS used in the present study. Substituted amino acid(s) are underlined. GST-ExoS(88–453) is the parental allele (‘wild-type’), such that all other alleles listed differ only by the amino acid substitution indicated in parentheses. The number in front of plasmid indicates lane numbering in Fig. 1. Plasmid Substituted amino acid(s) Reference or source 2. GST alone Amersham 3. GST-ExoS(88–453), wild type S 419 QGLLDALDL 428 [26] 4. GST-ExoS(88–453; SD)M 419 AAAA 428 [14] 5. GST-ExoS(LDL426–428AAA) S 419 QGLLDAAAA 428 This study 6. GST-ExoS(DALDL424–428AAAAA) S 419 QGLLAAAAA 428 This study 7. GST-ExoS(D424A; D427A) S 419 QGLLAALAL 428 This study 8. GST-ExoS(S419I) I 419 QGLLDALDL 428 This study 9. GST-ExoS(Q420A) S 419 AGLLDALDL 428 This study 10. GST-ExoS(G421A) S 419 QALLDALDL 428 This study 11. GST-ExoS(L422A) S 419 QGALDALDL 428 This study 12. GST-ExoS(L423A) S 419 QGLADALDL 428 This study 13. GST-ExoS(D424A) S 419 QGLLAALDL 428 This study 14. GST-ExoS(A425K) S 419 QGLLDKLDL 428 This study 15. GST-ExoS(L426A) S 419 QGLLDAADL 428 This study 16. GST-ExoS(D427A) S 419 QGLLDALAL 428 This study 17. GST-ExoS(L428A) S 419 QGLLDALDA 428 This study 18. GST-ExoS(LD426–427AA) S 419 QGLLDAAAL 428 This study L. Yasmin et al. Delineation of ExoS residues FEBS Journal 273 (2006) 638–646 ª 2006 The Authors Journal compilation ª 2006 FEBS 639 S(Q420A), GST-ExoS(G421A), GST-ExoS(L422A), GST-ExoS(L423A), GST-ExoS(D424A), GST-Exo- S(A425K), GST-ExoS(L426A), GST-ExoS(D427A), GST-ExoS(L428A) and GST-ExoS(LD426–427AA) (Table 1). All GST-ExoS derivatives were expressed and purified, and were then employed in protein pull- down experiments (Fig. 1). HeLa cells were harvested and the lysates precleared with GST beads prior to 1-h incubation with each of the indicated GST-ExoS- fusion proteins. Samples were subsequently washed and run on SDS ⁄ PAGE, followed by immunoblotting with 14-3-3 antibodies. It should be noted that we did not investigate binding of different 14-3-3 isoforms or the specificity of different 14-3-3 isoform binding in this study, as we used a pan-14-3-3 antibody. It is established that GST-ExoS(wt) interacts with 14-3-3, but not GST-beads alone or the fusion protein, GST- ExoS(SD), in which the ExoS residues at positions 419–423 are substituted with alanine and residues 424–428 have been deleted [14] (Fig. 1, compare lane 3 with lanes 2 and 4). We also observed that both GST-ExoS(DALDL424–428AAAAA) and GST-ExoS- (LDL426–428AAA) lack the ability to interact with 14-3-3 proteins from whole cell lysates of HeLa cells (Fig. 1, lane 5 and 6). At first glance, none of the single amino acid substitutions of GST-ExoS between amino acid 419–428 showed any obvious inability to interact with endogenous 14-3-3 proteins (Fig. 1, lanes 8–17). The same was true for a series of double substi- tution mutants: GST-ExoS(D424A; D427A) (Fig. 1, lane 7), GST-ExoS(LD426–427AA) (Fig. 2, lane 13), GST-ExoS(DL427-428AA) and GST-ExoS(LL426 : 428AA) (data not shown). The basic cluster of amino acids in the binding groove of 14-3-3, including amino acids Lys-49, Arg- 56, Lys-120 and Arg-127, in an otherwise acidic mole- cule, are important for the interaction with ExoS, while residues on the hydrophobic surface of the groove are dispensable [27]. Moreover, an artificial nonphosphorylated peptide ‘R18’ from a phage display library, binds within the same amphipathic groove of 14-3-3 [28]. In this case the negatively charged aspartic (Asp-12) and glutamic acid (Glu-14) residues in the R18 peptide were found to interact in the 14-3-3 pocket. Furthermore, a peptide sequence from ExoS including the motif D 424 ALDL 428 has the same poten- tial as R18 to inhibit ExoS ADP-ribosylating activity [14]. One suggestion was that the negatively charged amino acids, such as glutamic and aspartic acid residues, are able to mimic the phosphorylated serine Fig. 1. Interaction of GST-ExoS variants with endogenous 14-3-3 proteins. HeLa cells were harvested and lysates were subjected to ‘pull- down’ analysis with 5 lg of individual GST-fusion proteins. Samples were separated on a SDS ⁄ PAGE, followed by immunoblotting with 14-3-3antibodies. Upper panel: Lane 1, control HeLa cell lysate, 2 lg; lane 2, GST alone; lane 3, GST-ExoS(wt); lane 4, GST-ExoS(DS); lane 5, GST-ExoS(LDL426–428AAA); lane 6, GST-ExoS(DALDL424–428AAAAA); lane 7, GST-ExoS(D424A; D427A); lane 8, GST-ExoS(S419I); lane 9, GST-ExoS(Q420A); lane 10, GST-ExoS(G421A); lane 11, GST-ExoS(L422A); lane 12, GST-ExoS(L423A); lane 13, GST-ExoS(D424A); lane 14, GST-ExoS(A425K); lane 15, GST-ExoS(L426A); lane 16, GST-ExoS(D427A); lane 17, GST-ExoS(L428A). Lower panel: Coomassie blue stained SDS ⁄ PAGE showing the purified GST-fusion proteins used in this study. The order corresponds to lanes 2–17 above. Delineation of ExoS residues L. Yasmin et al. 640 FEBS Journal 273 (2006) 638–646 ª 2006 The Authors Journal compilation ª 2006 FEBS motif of Raf-1, which would perhaps explain the binding of 14-3-3 proteins to this motif [28]. To test the hypothesis put forward by Petosa et al. [28], we used single or double amino acid substitutions of the aspartic acid residues at positions 424 and 427 of the ExoS binding site for 14-3-3. These substitutions did not alter the ExoS)14-3-3 interaction under the condi- tions tested (Fig. 1, lanes 7, 13 and 16). Although from this analysis it is not obvious how the interaction between 14-3-3 and ExoS occurs, our pull-down analysis with GST-ExoS(LDL426– 428AAA) still strongly suggests that ExoS must utilize a strategy for its interaction with 14-3-3 that is similar to that seen with R18 and serotonin N-acetyltrans- ferase. This is because R18 is also nonphosphorylated and serotonin N-acetyltransferase selectively utilizes a subset of residues both in the conserved basic binding groove and residues outside the groove [13,28,29]. To understand the molecular basis for why the triple substitution mutant ExoS(LDL426–428AAA) bound cellular 14-3-3 proteins poorly, we tested whether decreasing amounts of single amino acid substitution mutant of GST-fusion proteins containing Exo- S(L426A), ExoS(D427A) or ExoS(L428A) altered the outcome of our pull-down assay. A dilution series (2.5, 1.25 or 0.75 lg) of GST-ExoS(wt) or of GST- ExoS(D427A) gave similar pull-down equivalent amounts of 14-3-3 proteins (Fig. 2, lanes 1–3 and 7–9). In contrast, diluted GST-ExoS(L426A) and GST-Exo- S(L428A) precipitated fewer 14-3-3 proteins (Fig. 2, lanes 4–6 and 10–12). Thus, the two leucine amino acids at positions 426 and 428 might still play a role in the interaction between ExoS and 14-3-3. Leucine 428 is an important determinant for induced cell death by the ADP-ribosylating domain of ExoS Having shed some light on the residues more import- ant for the interaction between ExoS and 14-3-3, we wanted to investigate how they affected the biological function of ExoS in vivo. We first employed a live ⁄ dead assay, capitalizing on the fact that before the ADP-ribosylation activity of translocated ExoS causes cell death, the infected cells undergo a morphology change whereby they round up due to disruption of actin microfilaments [21,30]. HeLa cells were infected for 2 h with the surrogate bacterium Yersinia pseudotu- berculosis [21], which was engineered to express and translocate, under the control of arabinose [31], ExoS wild type as well as several single, double and triple amino acid substitution variants into target cells. Translocation of all ExoS variants resulted in a cyto- toxic phenotype, e.g., cells rounded up and became semidetached from the Petri dish. Both loose and semidetached cytotoxic cells were washed free from bacteria and transferred to a new Petri dish and incu- bated overnight with medium containing gentamicin. Bacterial growth of each strain was assessed by viable counts, both during initial infection and also after extended infection, to ensure the same constant bacter- ial load (data not shown). At the same time, we con- firmed equivalent levels of ExoS expression and secretion by each strain (Fig. 3B,C, lanes 2–9). We then quantitated cell death by a trypan blue exclusion assay performed 24 h after infection. Infection with wild-type ExoS mediated a nonreversible cell morphol- Fig. 2. Effect of using GST-ExoS fusion dilutions during pull-down analysis. Selected GST-ExoS variants were sequentially diluted prior to analysis of their interaction potential with endogenous 14-3-3 proteins from HeLa cell lysate. Lane 1, 2.5 lg of GST-ExoS(wt); lane 2, 1.25 lg of GST-ExoS(wt); lane 3, 0.75 lg of GST-ExoS(wt); lane 4, 2.5 lg of GST-ExoS(L426A); lane 5, 1.25 lg of GST-ExoS(L426A); lane 6, 0.75 lg of GST-ExoS(L426A); lane 7, 2.5 lg of GST-ExoS(D427A); lane 8, 1.25 lg of GST-ExoS(D427A); lane 9, 0.75 lg of GST-ExoS(D427A); lane 10, 2.5 lg of GST-ExoS(L428A); lane 11, 1.25 lg of GST-ExoS(L428A); lane 12, 0.75 lg of GST-ExoS(L428A); lane 13, 2.5 lg of GST- ExoS(LD426–427AA); lane 14, 1.25 lg of GST-ExoS(LD426–427AA); lane 15, 0.75 lg of GST-ExoS(LD426–427AA). Upper panel, 14-3-3 pro- teins were detected by immunoblotting with anti14-3-3 antibodies. Lower panel, Coomassie blue stained GST-fusion proteins used in the pull-down experiment. L. Yasmin et al. Delineation of ExoS residues FEBS Journal 273 (2006) 638–646 ª 2006 The Authors Journal compilation ª 2006 FEBS 641 ogy, concomitant with a disruption of actin microfila- ments, and ultimately cell death (compare Fig. 4B with 4F), corroborating with earlier studies [21]. In fact, only 9% of ExoS(wt) infected cells survived compared with noninfected cells (Fig. 3A, compare lane 2 with lane 1). We also observed that single substitutions of aspartic acid residues at position 424 or 427 of ExoS [ExoS(D424A) or ExoS(D427A)] and the double mutant ExoS(DD424 : 427AA) were as aggressive as wild-type ExoS in their ability to induce cell death, as infected cells were unable to recover from the initial infection (Fig. 3A, lanes 4, 6 and 8). Together with the results from the GST-pull-down assay (Figs 1 and 2) using ExoS mutants with the same amino acid substi- tution, it is noticeable that negatively charged amino acids at positions 424 and 427 do not mimic phosphor- ylated serine motifs important for the interaction between 14-3-3 and ExoS. Therefore, this interaction is more complex and must occur in another way, whereby amino acids 426 and 428 have a more prom- inent role. Significantly, the translocated triple mutant ExoS(LDL426–428AAA), which is unable to interact with 14-3-3 proteins in pull-down experiments, was sig- nificantly impaired in its ability to induce cell death, with the majority of cells (92%) surviving the infection with ExoS(LDL426–428AAA) toxin (Fig. 4, compare C with G and Fig. 3A, lane 3). Therefore, mutant ExoS(LDL426–428AAA) has a reduced ADP-ribosyla- tion activity, the main cause of cell death. This pheno- type is reminiscent of cells transiently infected with the ADP-ribosylation mutant ExoS(E381A), which recover their original cell structure and morphology overnight [21,30]. By analogy, ExoS(LDL426–428AAA) must still harbor wild-type GAP activity that enables actin reorganization through the ability to down regulate the activity of small GTP binding proteins, such as Rho and Cdc42 in HeLa cells [21]. However, a reduced ADP-ribosylation activity permits this phenotype to be reversed postinfection. This phenotype must be due to either leucine residues at positions 426 or 428, as a mutation of aspartic acid at position 427 aggressively induced cell deaths such as the wild type. Indeed, bac- teria translocating the ExoS(L428A) mutant poorly mediated cell death (90% survival) after a 2-h infec- tion, which is comparable to bacteria expressing the ExoS(LDL426–428AAA) mutant (Fig. 3A, lane 7, and Fig. 4, compare D with H). Curiously, this was despite an interaction between ExoS(L428A) and 14-3-3 in the pull-down experiment (Figs 1 and 2). In contrast, the ExoS(L426A) mutant killed all but 8% of infected cells similar to the wild-type protein (Fig. 3A, lane 5). To further support this important role for amino acid 428, a double mutant, ExoS(LD426 : 427AA), was con- structed. Bacteria translocating ExoS(LD426 : 427AA) still mediated significant cell death with only 20% sur- vival (Fig. 3, lane 9). This is similar to the lethal affects of the single substitution mutants ExoS(L426A) and ExoS(D427A). This is surprising, as this double mutant was rather impaired in 14-3-3 binding (Fig. 2, lanes 13–15). Why this weak interaction between ExoS(LD426 : 427AA) and 14-3-3 is still enough to mediate cytotoxicity is currently unclear. We can only A B C D E Fig. 3. Phenotypic analysis of ExoS during infection of HeLa cells in vivo. (A) Viability of HeLa cells are expressed as percentage survi- val rate. HeLa cells in the presence of 0.1% arabinose, were infec- ted for 2 h with Yersinia (YPIII) expressing different variants of ExoS, lane 1, noninfected cells; lane 2, YPIII(pMF384) expressing ExoS(wt); lane 3, YPIII(pMF516) expressing ExoS(LDL426–428AAA); lane 4, YPIII(pMF515) expressing ExoS(D424A); lane 5, YP- III(pMF582) expressing ExoS(L426A); lane 6, YPIII(pMF493) expres- sing ExoS(D427A); lane 7, YPIII(pMF583) expressing ExoS(L428A); lane 8, YPIII(pMF523) expressing ExoS(DD424 : 427AA); lane 9, YP- III(pMF518) expressing ExoS(LD426–427AA). Both loose and semi- detached cytotoxic cells were washed free from bacteria and transferred to a new Petri dish and incubated overnight with med- ium containing gentamicin. A trypan blue exclusion assay was per- formed 24 h after infection to quantitated the percentage of dead cells. Each bar represents the mean values of five independent experiments. (B) and (C) ExoS expression (B) and secretion (C) after each Y. pseudotuberculosis strain was induced in calcium-depleted medium in the presence of arabinose. Proteins were analyzed on SDS ⁄ PAGE followed by western blot using anti-ExoS antibodies. (D) and (E) Cells were lysed and samples were separated by SDS ⁄ PAGE. Western blot analysis was performed on immunoblot- ted filters with anti-Ras (D) and with anti-Erk 2 (E) antibodies. Delineation of ExoS residues L. Yasmin et al. 642 FEBS Journal 273 (2006) 638–646 ª 2006 The Authors Journal compilation ª 2006 FEBS speculate that the weak interaction is able to induce a conformational change of the ExoS protein that might be of importance for the activation of the ADP-ribosy- lation activity. Nevertheless, we define a second resi- due, leucine at position 428, which is an important determinant for induced cell death by the ADP-ribosy- lating domain of ExoS. Whether this serves a similar function to the critical glutamic acid residue at posi- tion 381 [32] remains a focus for our future research. ExoS-dependent in vivo ADP-ribosylation of Ras requires the Leu-428 residue Ras is modified by the ADP-ribosylating activity of ExoS expressed and delivered into the eukaryotic cells by genetically modified Y. pseudotuberculosis [14]. We used this assay to further assess the in vivo biological activity of our ExoS variants. HeLa cells were in- fected for 2 h with Y. pseudotuberculosis induced by arabinose to express and translocate ExoS(wt), ExoS(D424A), ExoS(L426A), ExoS(D427A), Exo- S(L428A), ExoS(D424A; D427A), ExoS(LDL426– 428AAA) and ExoS(LD426–427AA) into target cells. The cells were then harvested and the resultant lysate was separated on a SDS ⁄ PAGE followed by immuno- blotting with anti-Ras and anti-pan-Erk antibodies as a loading control (Fig. 3D and E respectively). Ras was modified in cells infected with bacteria expressing one of either wild-type ExoS, ExoS(D424A), Exo- S(L426A), ExoS(D427A), ExoS(D424A; D427A) or ExoS(LD426–427AA) (Fig. 3D, lanes 2, 4, 5, 6, 8 and 9). This paralleled our analysis of ExoS-induced HeLa cell death. Significantly, much less modification of Ras was observed in bacteria translocating either Exo- S(L428A) or ExoS(LDL426–428AAA) into infected cells (Fig. 3D, lanes 3 and 7), which again correlated to the extent of cell survival in these infections. While, for the most part, our results herein reflect the established principle that 14-3-3 proteins act as cofactors in activating ExoS located in the cytosol [14,26–28,33], a notable exception was revealed. The ExoS(LD426–427AA) double mutant and, to a lesser extent, the single mutant ExoS(L426A), showed weakened 14-3-3 binding potential. However, like ExoS(D427A), these toxin variants were still biologic- ally active. This suggests that the limited binding was still productive, in the sense that an initial contact of ExoS by 14-3-3 proteins or a fast on-off ratio is suffi- cient for ADP-ribosylation activation of ExoS. One goal of this study was to identify single ExoS amino acids residues, which are important for the phosphorylation independent interaction with 14-3-3. This is important considering that most interactions between 14-3-3 and cellular proteins require a phosphorylation-dependent event. At least for ExoS, however, earlier predictions that phosphorylation–inde- pendent interactions were mediated by acidic residues are not the whole truth. Interestingly, residues Leu-426 and Leu-428 were found to be most important for ini- tial binding in our assay. We interpret this to mean Fig. 4. Morphological analysis of HeLa cells infected with variants of ExoS. HeLa cells in the presence of 0.1% arabinose, were infected with: (B) and (F) YPIII (pMF384) expressing arabinose induced ExoS(wt); (C) and (G) YPIII(pMF516) expressing ExoS(LDL426–428AAA); (D) and (H) YPIII(pMF583) expressing ExoS(L428A). (A) and (E) Uninfected cells were used as a control. After infection for 2 h with bacteria translocating different variants of ExoS, all cells showed a cytotoxic phenotype in that they rounded up and became semidetached from the Petri dish (A–D). These infections were washed free from bacteria and transferred to new Petri dishes and incubated with medium contain- ing penicillin, streptomycin and gentamicin to ascertain the reversibility of this cytotoxic response (E–H). L. Yasmin et al. Delineation of ExoS residues FEBS Journal 273 (2006) 638–646 ª 2006 The Authors Journal compilation ª 2006 FEBS 643 that the phosphorylation-independent ExoS)14-3-3 interaction is complex, and is likely to involve coordi- nation of multiple discrete ExoS interaction motifs, some of which may be acidic in nature, but others not. It is easy to imagine that these molecular contacts could generate ExoS conformational changes necessary for the controlled induction of enzymatic activity or could even activate a cytosolic targeting mechanism. Understanding these molecular events will no doubt require detailed structural analysis, which is not cur- rently available. Numerous reports have described the importance of 14-3-3 proteins as a factor involved in the activation of ExoS [14,26–28,33]. We were therefore very sur- prised when the single substitution mutant Exo- S(L428A) lacked ADP-ribosylating activity in vivo, even though this mutant should still engage 14-3-3 proteins from HeLa cell lysates. This raises the notion that 14-3-3 binding is not the sole requirement for ExoS activity. Perhaps Leu-428 is even required for enzymatic activity per se, such as in directly engaging the molecular targets of ADP-ribosylation. This evokes the function of glutamic acid at position 381, which is a prerequisite for ADP-ribosylating activity. It has been proposed that E-381 functions in both catalysis and in contributing to the structural integrity of the active site [32]. Could it be that Leu-428 exhib- its similar properties? Another possibility is the Exo- S(L428A))14-3-3 interaction is not productive. While 14-3-3 can still bind to this mutant, perhaps it is unable to induce a putative conformational change that may be necessary for ExoS activation. If this were true, it would not appear to be due to a differ- ent fold in ExoS(L428A) compared to any other ExoS variant used in this study, because we did not detect any difference in protein production or stability (Fig. 3, and data not shown). In summary, we propose that ExoS of P. aeruginosa has evolved to recruit 14-3-3 to regulate its enzymatic activity, which is similar to many other signal-induced interactions between 14-3-3 and its targets ([3,7,8] and refs therein). It is noteworthy that 14-3-3 proteins are only expressed in eukaryotic cells, including plants, yeast and protozoa. No clear prokaryotic ancestor has been identified. Thus, it would be interesting to deter- mine if bacteria expressing a 14-3-3 isoform in the presence of ExoS can survive, as it may be the absence of 14-3-3 homologues in prokaryotes that safeguard them against the deleterious effects of their own toxins. This suggests that prokaryotic evolution has created a new way to take advantage of an evolutionary ‘novel’ eukaryotic 14-3-3 protein family, using them as a necessary cofactor to activate lethal bacterial toxins, but only after they have been safely transported from the bacteria into the eukaryotic cell. It is apparent that more secrets concerning this intriguingly complex interaction need to be uncovered. Many of these may be revealed only through compre- hensive structural analysis. No structural data exists for the phosphorylation-independent 14-3-3–ExoS complex, either using native ExoS domains or a syn- thetic peptide sequence encompassing the 14-3-3 bind- ing domain (this study) [14,15,26]. An enticing prospect for future research is to determine how amino acid Leu-428 of ExoS influences the interaction dynamics with 14-3-3. Experimental procedures Cell cultures, cell lysis HeLa cells were grown in RPMI 1640 supplemented with 10% (v ⁄ v) fetal bovine serum and 100 units ⁄ mL penicillin. Following bacterial infection cells were washed in ice-cold NaCl ⁄ P i and lysed on ice in lysis buffer [1%(v ⁄ v) Triton x-100, 100 mm NaCl, 50 mm Tris ⁄ HCl (pH 7.5), 1 mm EDTA supplemented with protease inhibitors (Complete, #1697498, Roche Diagnostics, Basel, Switzerland)]. Lysates were subsequently cleared by centrifugation at 15 000 g for 10 min at 4 °C. Lysates were precleared with glutathione S-transferase (GST) for 5 min, before incubation with var- ious GST-fusion proteins for 1 h prior to the addition of Glutathione Sepharose (GE Healthcare, Uppsala, Sweden) for 30 min. After three washes in lysis buffer, samples were boiled in SDS ⁄ PAGE sample buffer. Western analysis, peptides and antibodies Anti-14–3-3b (SC-629) was purchased from Santa Cruz (New York, NY, USA); monoclonal Ras (cat 610002) was obtained from BD Biosciences (Stockholm, Sweden). Anti- ExoS was from Agrisera AB, Sweden. Immunoblotting was performed according to the manufacturer’s instructions using secondary antibodies conjugated to horseradish peroxidase sheep antimouse or rabbit antibodies (Pierce, Rockford, IL, USA, and ECL Plus, Amersham-Biosciences). Plasmids pGEX-ExoS(SD) is derivative of pGEX-ExoS(88–453], both of which have been described previously [14]. The substi- tution mutants [Table 1], pGEX-ExoS(S419I), pGEX- ExoS(Q420A), pGEX-ExoS(G421A), pGEX-ExoS(L422A), pGEX-ExoS(L423A), pGEX-ExoS(D424A), pGEX-Exo- S(A425K), pGEX-ExoS(L426A), pGEX-ExoS(D427A), pGEX-ExoS(L428A), pGEX-ExoS(D424A:D427A), pGEX- ExoS(LD426–427AA), pGEX-ExoS(LDL426–428AAA), Delineation of ExoS residues L. Yasmin et al. 644 FEBS Journal 273 (2006) 638–646 ª 2006 The Authors Journal compilation ª 2006 FEBS and pGEX-ExoS (DALDL424–428AAAAA) were construc- ted by digesting pGEX-ExoS (SD) with NdeI ⁄ NheI and inserting oligomers (supplementary material, Table S1) cor- responding to the appropriate amino acid substitutions, as outlined in Table 1. All constructs were confirmed by sequencing with DYEnamic ET terminal cycle sequencing kit (Amersham-Biosciences). Construction of arabinose inducible ExoS derivatives and infection of cells To ensure protein stability of full-length ExoS derivatives, mutant alleles were coexpressed with orf1, encoding the cog- nate nonsecreted chaperone of ExoS [30,34]. In all cases, DNA was amplified by PCR using conditions described pre- viously [35]. Construction of pMF384 containing arabinose inducible wild-type exoS has been described in detail previ- ously [14]. Arabinose inducible exoS variants on the plasmids pMF493, pMF515, pMF516, pMF518, pMF523, pMF582 and pMF583 were obtained by replacing the C-terminal ClaI ⁄ KpnI exoS fragment from pMF384 with DNA ampli- fied and restriction enzyme cut with ClaI ⁄ KpnI from pGEX-ExoS(D427A), pGEX-ExoS(D424A), pGEX- ExoS(LDL426–428AAA), pGEX-ExoS(LD426–427AA), pGEX-ExoS(DD424 : 427AA), pGEX-ExoS(L426A), and pGEX-ExoS(L428A), respectively (see Supplementary mater- ial, Table S1), using the exoS-specific primers, pexoSseq3 (position 973991; forward): 5¢-AAGTGATGGCGCTTGG TCT-3¢ and pexoSd (reverse): 5¢-ATGCATGGTACCTCAG GCCAGATCAAGGCCGCG-3¢. All constructs were main- tained in Escherichia coli DH5 and were confirmed by sequence analysis using the DYEnamic ET terminator cycle sequencing kit (Amersham Biosciences). Stable induc- tion of protein expression in strains grown in the presence of 0.02% l(+)-arabinose was confirmed by western analy- sis, as described previously [36], using polyclonal rabbit anti-ExoS [30]. Bacterial infection of cells was performed in the presence of 0.1% l(+)-arabinose, as described previously [14]. Acknowledgements Financial support for this work was from the Swedish Cancer Society, Carl Tryggers Foundation, and Riksfo ¨ rbundet Cystisk Fibros Forskningsfond. References 1 Muslin AJ, Tanner JW, Allen PM & Shaw AS (1996) Interaction of 14-3-3 with signaling proteins is mediated by the recognition of phosphoserine. Cell 84, 889–897. 2 Yaffe MB, Rittinger K, Volinia S, Caron PR, Aitken A, Leffers H, Gamblin SJ, Smerdon SJ & Cantley LC (1997) The structural basis for 14-3-3: phosphopeptide binding specificity. Cell 91, 961–971. 3 Mackintosh C (2004) Dynamic interactions between 14 and 3–3 proteins and phosphoproteins regulate diverse cellular processes. Biochem J 381, 329–342. 4 Wurtele M, Jelich-Ottmann C, Wittinghofer A & Oeck- ing C (2003) Structural view of a fungal toxin acting on a 14-3-3regulatory complex. Embo J 22, 987–994. 5 Roberts MR (2003) 14-3-3 proteins find new partners in plant cell signalling. Trends Plant Sci 8, 218–223. 6 Moorhead G, Douglas P, Cotelle V et al. (1999) Phos- phorylation–dependent interactions between enzymes of plant metabolism and 14-3-3 proteins. Plant J 18, 1–12. 7 Dougherty MK & Morrison DK (2004) Unlocking the code of 14-3-3. J Cell Sci 117, 1875–1884. 8 Wilker E & Yaffe MB (2004) 14-3-3 Proteins – a focus on cancer and human disease. J Mol Cell Cardiol 37, 633–642. 9 Rittinger K, Budman J, Xu J, Volinia S, Cantley LC, Smerdon SJ, Gamblin SJ & Yaffe MB (1999) Structural analysis of 14-3-3 phosphopeptide complexes identifies a dual role for the nuclear export signal of 14-3-3 in ligand binding. Mol Cell 4, 153–166. 10 Dubois T, Howell S, Amess B et al. (1997) Structure and sites of phosphorylation of 14-3-3 protein: role in coordinating signal transduction pathways. J Protein Chem 16, 513–522. 11 Liu D, Bienkowska J, Petosa C, Collier RJ, Fu H & Liddington R (1995) Crystal structure of the zeta iso- form of the 14-3-3 protein. Nature 376, 191–194. 12 Xiao B, Smerdon SJ, Jones DH, Dodson GG, Soneji Y, Aitken A & Gamblin SJ (1995) Structure of a 14-3- 3protein and implications for coordination of multiple signalling pathways. Nature 376, 188–191. 13 Obsil T, Ghirlando R, Klein DC, Ganguly S & Dyda F (2001) Crystal structure of the 14-3-3zeta: serotonin N-acetyltransferase complex: a role for scaffolding in enzyme regulation. Cell 105, 257–267. 14 Henriksson ML, Francis MS, Peden A, Aili M, Stefans- son K, Palmer R, Aitken A & Hallberg B (2002) A non- phosphorylated 14-3-3 binding motif on exoenzyme S that is functional in vivo. Eur J Biochem 269, 4921– 4929. 15 Masters SC, Pederson KJ, Zhang L, Barbieri JT & Fu H (1999) Interaction of 14-3-3 with a nonphosphory- lated protein ligand, exoenzyme S of Pseudomonas aeru- ginosa. Biochemistry 38, 5216–5221. 16 Zhai J, Lin H, Shamim M, Schlaepfer WW & Canete- Soler R (2001) Identification of a novel interaction of 14- 3-3 with p190RhoGEF. J Biol Chem 276, 41318–41324. 17 Wang B, Yang H, Liu YC, Jelinek T, Zhang L, Ruoslahti E & Fu H (1999) Isolation of high-affinity peptide antagonists of 14-3-3 proteins by phage display. Biochemistry 38, 12499–12504. L. Yasmin et al. Delineation of ExoS residues FEBS Journal 273 (2006) 638–646 ª 2006 The Authors Journal compilation ª 2006 FEBS 645 18 Stryjewski M & Sexton D (2003) Pseudomonas aerugi- nosa infections in specific types of patients and clinical settings. In Severe Infections Caused by Pseudomonas Aeruginosa (Hauser, A & Rello, J, eds), pp. 1–15. Kluwer Academic Publishers, Boston, Masachusetts. 19 Cornelis GR & Van Gijsegem F (2000) Assembly and function of type III secretory systems. Annu Rev Micro- biol 54, 735–774. 20 Goehring UM, Schmidt G, Pederson KJ, Aktories K & Barbieri JT (1999) The N-terminal domain of Pseudomo- nas aeruginosa exoenzyme S is a GTPase-activating pro- tein for Rho GTPases. J Biol Chem 274, 36369–36372. 21 Henriksson ML, Sundin C, Jansson AL, Forsberg A, Palmer RH & Hallberg B (2002) Exosenzyme S show selective ADP-ribosylation and GAP activities towards small GTPases in vivo. Biochem J 367, 617–628. 22 Coburn J, Wyatt RT, Iglewski BH & Gill DM (1989) Several GTP-binding proteins, including p21c-H-ras, are preferred substrates of Pseudomonas aeruginosa exo- enzyme S. J Biol Chem 264, 9004–9008. 23 Coburn J & Gill DM (1991) ADP-ribosylation of p21ras and related proteins by Pseudomonas aeruginosa exoenzyme S. Infect Immun 59, 4259–4262. 24 Coburn J, Kane AV, Feig L & Gill DM (1991) Pseudo- monas aeruginosa exoenzyme S requires a eukaryotic protein for ADP-ribosyltransferase activity. J Biol Chem 266, 6438–6446. 25 Fu H, Coburn J & Collier RJ (1993) The eukaryotic host factor that activates exoenzyme S of Pseudomonas aeruginosa is a member of the 14-3-3 protein family. Proc Natl Acad Sci USA 90, 2320–2324. 26 Henriksson ML, Troller U & Hallberg B (2000) 14-3- 3proteins are required for the inhibition of Ras by exoenzyme S. Biochem J 349, 697–701. 27 Zhang L, Wang H, Masters SC, Wang B, Barbieri JT & Fu H (1999) Residues of 14-3-3 zeta required for activa- tion of exoenzyme S of Pseudomonas aeruginosa. Bio- chemistry 38, 12159–12164. 28 Petosa C, Masters SC, Bankston LA, Pohl J, Wang B, Fu H & Liddington RC (1998) 14-3-3 zeta binds a phosphorylated Raf peptide and an unphosphorylated peptide via its conserved amphipathic groove. J Biol Chem 273, 16305–16310. 29 Rittinger K, Budman J, Xu J, Volinia S, Cantley LC, Smerdon SJ, Gamblin SJ & Yaffe MB (1999) Structural analysis of 14-3-3phosphopeptide complexes identifies a dual role for the nuclear export signal of 14-3-3 in ligand binding. Mol Cell 4, 153–166. 30 Frithz-Lindsten EY, Rosqvist R & Forsberg A (1997) Intracellular targeting of exoenzyme S of Pseudomonas aeruginosa via type III-dependent translocation induces phagocytosis resistance, cytotoxicity and disruption of actin microfilaments. Mol Microbiol 25, 1125–1139. 31 Francis MS, Lloyd SA & Wolf-Watz H (2001) The type III secretion chaperone LcrH co-operates with YopD to establish a negative, regulatory loop for control of Yop synthesis in Yersinia pseudotuberculosis. Mol Microbiol 42, 1075–1093. 32 Liu S, Kulich SM & Barbieri JT (1996) Identification of glutamic acid 381 as a candidate active site residue of Pseudomonas aeruginosa exoenzyme S. Biochemistry 35, 2754–2758. 33 Zhang L, Wang H, Liu D, Liddington R & Fu H (1997) Raf-1 kinase and exoenzyme S interact with 14- 3-3 zeta through a common site involving lysine 49. J Biol Chem 272, 13717–13724. 34 Yahr TL, Goranson J & Frank DW (1996) Exoenzyme SofPseudomonas aeruginosa is secreted by a type III pathway. Mol Microbiol 22, 991–1003. 35 Francis MS & Wolf-Watz H (1998) YopD of Yersinia pseudotuberculosis is translocated into the cytosol of HeLa epithelial cells: evidence of a structural domain necessary for translocation. Mol Microbiol 29, 799–813. 36 Francis MS, Aili M, Wiklund ML & Wolf-Watz H (2000) A study of the YopD–lcrH interaction from Yer- sinia pseudotuberculosis reveals a role for hydrophobic residues within the amphipathic domain of YopD. Mol Microbiol 38, 85–102. Supplementary material The following supplementary material is available online: Table S1. Construction of plasmids used in this study. pGEX-2TK-ExoS(Sn) [14] was digested with NdeI and NheI, followed by insertion of the annealed oligo- mers listed, which contained the appropriate amino acid substitutions corresponding to the ExoS variants outlined in Table 1. This material is available as part of the online article from: http://www.blackwell-synergy.com Delineation of ExoS residues L. Yasmin et al. 646 FEBS Journal 273 (2006) 638–646 ª 2006 The Authors Journal compilation ª 2006 FEBS . we used single or double amino acid substitutions of the aspartic acid residues at positions 424 and 427 of the ExoS binding site for 14-3-3. These substitutions. for its interaction with 14-3-3 that is similar to that seen with R18 and serotonin N-acetyltrans- ferase. This is because R18 is also nonphosphorylated and