DAP kinase mediates the phosphorylation of tropomyosin-1 downstream of the ERK pathway, which regulates the formation of stress fibers in response to oxidative stress

Tropomyosins constitute a family of nearly 40 closely related proteins generated from alternative splicing of four different genes (Perry, 2001). These different isoforms are selectively expressed in a cell-specific manner, and their number varies from one cell type to another. The balance between the isoforms present in a given cell determines the specificity of the tropomyosin functions (Bryce et al., 2003). In striated muscle, tropomyosin regulates the interaction between myosin and actin by means of a troponin complex that mediates muscle contraction in response to calcium (Cooper, 2002). In non-muscle cells, tropomyosins play a role in the formation and stabilization of stress fibers by facilitating actomyosin interactions and protecting existing fibers from the action of severing and deplolymerizing proteins (Cooper, 2002).

Tropomyosins are among the most abundant cytoskeletal proteins in endothelial cells (Patton et al., 1990). Accordingly, they play important roles in regulating cardiovascular homeostasis. For example, they modulate the functions of endothelial cells associated with cytoskeletal remodeling such as motility and permeability (Houle and Huot, 2006). In the heart, tropomyosin also plays essential functions, and tropomyosin mutations are associated with cardiac myopathies (Michele and Metzger, 2000). For example, mutations of cardiac tropomyosin-α, within the troponin-interaction region (Asp175Asn and Glu180Gly), impair contractility and relaxation in transgenic mouse heart (Muthuchamy et al., 1999; Prabhakar et al., 2001; Prabhakar et al., 2003). Mice homozygous for knockout of the gene encoding cardiac tropomyosin-α die between embryonic day 9.5 and 13.5, whereas heterozygous animals show compensatory expression and present no difference in tropomyosin-α expression in comparison with wild-type mice, suggesting the existence of tight auto-regulation of expression (Blanchard et al., 1997).

Intriguingly, loss of tropomyosin expression is found in tumor cells, which suggests that tropomyosin functions as a tumor suppressor (Bharadwaj and Prasad, 2002; Bharadwaj et al., 2005; Boyd et al., 1995; Varga et al., 2005). It is thought that the inhibition of tropomyosin expression is an essential step by which the oncogenic process pushes cells towards anchorage-independent cell growth. Accordingly, re-expression of tropomyosin in breast cancer cells reorganizes microfilaments and abolishes the anchorage-independent growth capacity of these cells (Bharadwaj et al., 2005).

Despite the knowledge that tropomyosins might act as tumor suppressors and that they play an essential role in cardiovascular homeostasis, little is known about how their functions are regulated. In particular, the posttranslational modifications of tropomyosins are still ill defined, especially in the context of their roles as modulators of tropomyosin functions. Nevertheless, it is known that the first methionine of tropomyosin-α is acetylated, which is essential for dimerization and interaction with actin (Hitchcock-DeGregori and Heald, 1987; Monteiro et al., 1994; Pittenger et al., 1995). Tropomyosin is also phosphorylated in skeletal, cardiac and smooth muscles, and this modification seems to modulate its interaction with other proteins, including caldesmon and HSP27 (Mak et al., 1978; Somara et al., 2005). More recently, it was reported that activation of the β-adrenergic receptor enhances the activity of phosphoinositide 3-kinase, leading to phosphorylation of tropomyosin and internalization of the receptor (Naga Prasad et al., 2005). By mass spectrometry, we identified tropomyosin-1 as an oxidative-stress-sensitive phosphoprotein in endothelial cells (Houle et al., 2003).

DAP kinase possesses a kinase domain that is closely related to that of members of the family of myosin light chain kinases (MLCKs) (Inbal et al., 2000; Bialik and Kimchi, 2006). It is a Ca²⁺/calmodulin-dependent Ser/Thr kinase that was identified initially in a functional knockout genetic screen in HeLa cells (Cohen et al., 1997; Deiss et al., 1995). Since then, four new members of the DAP kinase family have been cloned: zipper-interacting protein kinase (ZIP kinase), DAP-kinase-related protein kinase (DRP kinase) and DAP-kinase-related apoptosis-inducing protein kinases 1 and 2 (DRAK-1 and DRAK-2). All members phosphorylate the myosin light chain (MLC) and are implicated in the morphological changes associated with apoptosis (Cohen et al., 1997; Deiss et al., 1995; Inbal et al., 2000; Kawai et al., 1998; Murata-Hori et al., 2001; Sanjo et al., 1998). For example, DAP kinase activation by serum induces the formation of actin stress fibers before the onset of apoptosis in NIH 3T3 cells (Kuo et al., 2003). However, the signaling pathways involved remain to be determined.

Here, we show for the first time that H₂O₂ induces a time-dependent activation of DAP kinase, downstream of ERK. We found that DAP kinase promotes the in vitro and in vivo phosphorylation of tropomyosin-1 on Ser283 and that this phosphorylation is essential for induction by H₂O₂ of formation of actin stress fibers in endothelial cells. These findings strongly support a role for the phosphorylation of tropomyosin-1 in regulating the endothelial functions associated with cytoskeletal remodeling.

We previously reported that tropomyosin-1 is a protein phosphorylated downstream of the ERK pathway in endothelial cells activated by oxidative stress (Houle and Huot, 2006; Houle et al., 2003). In this study, we investigated further the signaling mechanisms that lead to phosphorylation of tropomysoin-1 and regulate its functions. Given that tropomyosin has several isoforms that share sequence homology with tropomyosin-1 (Bruneel et al., 2005; Gunning et al., 2005), we first designed a set of experiments to demonstrate even more comprehensively that the tropomyosin-1 isoform is phosphorylated downstream of ERK in response to H₂O₂. We expressed exogenous human tropomyosin-1 in HEK293 cells as these cells have a very low basal level of tropomyosin-1 expression in comparison with human umbilical vein endothelial cells (HUVECs) and other cell types, including COS cells (Fig. 1A,B, lower panel). Thereafter, cells were incubated with H₃[³²P]O₄ and were treated or not for 30 minutes with 250 μM H₂O₂ in the presence or absence of PD098059 or UO126 to inhibit ERK activation. Protein extracts were prepared and tropomyosin-1 was immunoprecipitated and separated by SDS-PAGE. The results showed that H₂O₂ induced a strong increase in phosphorylation of tropomyosin-1 that was sensitive to ERK inhibition by both inhibitors (Fig. 1B). This indicates that the tropomyosin-1 isoform is phosphorylated downstream of ERK in response to oxidative stress. This conclusion was still further strengthened by the results of experiments performed on HUVECs electroporated with plasmids expressing haemagglutinin- (HA-) or FLAG-tagged tropomyosin-1 and then treated for 30 minutes with 250 μM H₂O₂ in the presence or absence of PD098059 or UO126 before being extracted, submitted to 2D gel analysis and immunoblotted for tagged tropomyosin-1. Using this approach, we found that treatment with H₂O₂ triggered a shift of the spot corresponding to non-phosphorylated tropomyosin-1 towards the acidic form, revealing its phosphorylation. The shift was inhibited by PD098059 or UO126, confirming that tropomyosin-1 is phosphorylated downstream of the ERK pathway in endothelial cells (Fig. 1C). Interestingly, we also found that ML-7, an inhibitor of cell contractility, blocks the phosphorylation shift of tropomyosin-1 induced by H₂O₂ (Fig. 1D). However, inhibition of actin polymerization by cytochalasin D did not impair the phosphorylation of tropomyosin 1, indicating that the phosphorylation of tropomyosin is not dependent on intact actin filaments (data not shown). Overall, these findings clearly indicate that the tropomyosin-1 isoform is phosphorylated downstream of the ERK pathway in response to oxidative stress and suggest the involvement of a kinase of the MLCK family, given that ML-7 inhibits the phosphorylation of tropomyosin-1.

Human tropomyosin-1 contains 14 serine and seven threonine residues that could be phosphorylated. However, none of them is within the minimal (S/TP) consensus motif for phosphorylation by ERK. Moreover, tropomyosin-1 does not contain the FxFP sequence that mediates the interaction of ERK with its substrate (Jacobs et al., 1999). This suggests that ERK is not the kinase that directly phosphorylates tropomyosin-1. To confirm this point, we set up an in vitro immunocomplex assay of ERK to test its potential tropomyosin-1 kinase activity. HUVECs were pre-treated with UO126, treated or not with H₂O₂, and then ERK was immunoprecipitated and incubated with recombinant human tropomyosin-1 (rh-tropomyosin-1) or myelin basic protein (MBP). As shown in Fig. 2A, ERK is activated by H₂O₂, as reflected by the phosphorylation of MBP, and its activation is sensitive to inhibition by UO126 (Fig. 2A). However, as expected, ERK did not phosphorylate rh-tropomyosin-1 (Fig. 2B). Hence, we next attempted to identify the tropomyosin-1 kinase (`TMK') that must act downstream of ERK.

As shown in Fig. 1D, H₂O₂-induced phosphorylation of tropomyosin-1 was inhibited by ML-7 even if this agent did not affect ERK activation (Houle et al., 2003). Given that ML-7 is a known inhibitor of MLCK, we inferred that the TMK was a member of the MLCK-related kinases. We then used purified activated MLCK to verify its ability to phosphorylate purified recombinant human tropomyosin-1. We found that activated MLCK could not induce the phosphorylation of rh-tropomyosin-1 (Fig. 3A). By contrast, MLCK phosphorylated myosin light chain (MLC) in a dose-dependent manner, indicating its constitutive activity (Fig. 3A). These results suggest that MLCK does not function as a TMK. Kinase assays were then set up using conditions that are specific for each of the other major members of the family, namely DRP, ZIP and DAP kinase, and using total extracts of HUVECs treated or not with H₂O₂ and rh-tromomyosin-1 as a substrate. We found that tropomyosin-1 was phosphorylated after H₂O₂ treatment, but only in assay conditions that were specific for DAP kinase (data not shown). We thus verified whether H₂O₂ activates DAP kinase. HEK293 cells were transfected with a wild-type form of FLAG-tagged DAP kinase and, after 24 hours, they were treated or not with H₂O₂ for increasing periods of time. Thereafter, FLAG-DAP kinase was immunoprecipitated with an antibody against the FLAG epitope and the immunoprecipitates were used to assay the activity of DAP kinase by evaluating its ability to phosphorylate MLC on Ser19, a typical substrate for this kinase. The results showed a time-dependent activation of DAP kinase that reached a peak 3.5-fold elevation at 10 minutes of treatment with H₂O₂ (Fig. 3B). Given that the phosphorylation of tropomyosin-1 was sensitive to ML-7, we investigated next whether DAP kinase activation by H₂O₂ was inhibited by this agent. FLAG-tagged DAP kinase was immunoprecipitated from HEK293 cells transfected with FLAG-DAP kinase and treated with H₂O₂. Thereafter, activation of immunoprecipitated DAP kinase was assayed in vitro in the presence of increasing concentrations of ML-7, using MLC as a substrate. As expected, the results presented in Fig. 3C showed that ML-7 inhibited in a dose-dependent manner the activation of DAP kinase in response to H₂O₂. However, it did not affect the autokinase activity of DAP kinase, as previously reported (Kuo et al., 2003).

We next confirmed that DAP kinase was indeed a kinase involved in phosphorylating tropomyosin-1. HEK293 cells were transfected with plasmids expressing FLAG-tagged versions of wild-type DAP kinase (wt-DAP kinase) or of a dominant-negative form of DAP kinase (DAP kinase K42A) (Cohen et al., 1997). Thereafter, the cells were treated or not with 250 μM H₂O₂ for 10 minutes, and the different forms of transfected FLAG-tagged DAP kinase were immunoprecipitated with an antibody against FLAG. The immunoprecipitates were used to assay the ability of DAP kinase to phosphorylate rh-tropomomyosin-1 used as substrate. As expected, we found that H₂O₂ induced the phosphorylation of rh-tropomyosin-1 with immunoprecipitated wt-DAP kinase but not with the immunoprecipitated dominant-negative DAP kinase K42A (Fig. 4A). The results indicated that DAP kinase can phosphorylate tropomyosin-1 in response to oxidative stress. We next verified whether DAP kinase could directly phosphorylate tropomyosin-1. We incubated purified activated DAP kinase with purified rh-tropomyosin-1. As expected, we found that DAP kinase phosphorylated rh-tropomyosin-1 in a dose-dependent manner, indicative of its TMK activity (Fig. 4B). From these in vitro kinase assays, we evaluated that 1 picomole of purified activated DAP kinase induces the incorporation of 0.9 μmoles PO₄/mmole rh-tropomyosin-1.

Thereafter, we ascertained whether DAP kinase is a tropomyosin-1 kinase that functions downstream of ERK. We first verified whether DAP kinase is phosphorylated in an ERK-dependent manner in response to H₂O₂. HEK293 cells were transfected with FLAG-tagged DAP kinase and incubated in medium containing H₃[³²P]O₄ in the presence or absence of H₂O₂ or/and PD098059. We found that phosphorylation of DAP kinase was induced by H₂O₂ and that it occurred downstream of ERK, being inhibited by PD098059 (Fig. 5A). Additionally, the phosphorylation of DAP kinase bearing the point mutation Ser735Ala remained at the basal level in response to H₂O₂, which is in line with the previous observation that this site is required for the activation of DAP kinase by ERK (Fig. 5B) (Chen et al., 2005). Interestingly, we further found that DAP kinase co-precipitated with ERK and that the association was increased by H₂O₂-induced activation of ERK and was reduced in the presence of the MEK inhibitor UO126 that inhibits ERK (Fig. 5C). Note that, in these assays, we used the DAP kinase-inactive mutant K42A to eliminate the possible involvement of the auto-kinase activity of DAP kinase. These findings are consistent with the hypothesis that DAP kinase is downstream of ERK. This conclusion is further highlighted by the fact that DAP kinase was activated downstream of ERK in endothelial cells. Indeed, we found that the H₂O₂-induced activation of endogenous DAP kinase immunoprecipitated from HUVECs pretreated with the ERK inhibitor UO126 was impaired following inhibition of ERK (Fig. 5D). Finally, the involvement of DAP kinase in vivo as a TMK was further supported by the finding that phosphorylation of tropomyosin-1 in an in vivo phosphorylation assay was impaired in HUVECs in which the expression of DAP kinase was knocked-down by siRNA and in HEK293 cells expressing a dominant-negative form of DAP kinase (Fig. 6A,B). Together, these findings constitute the first evidence indicating that tropomyosin-1 is phosphorylated by DAP kinase downstream of ERK in response to oxidative stress.

In line with our previous report (Houle et al., 2003), H₂O₂-mediated phosphorylation of tropomyosin triggers its association with actin stress fibers. In order to understand better the molecular mechanisms of tropomyosin-1 function in modulating cytoskeletal dynamics, we next attempted to identify the site/s phosphorylated within tropomyosin-1 after H₂O₂ treatment. As shown in Fig. 3B, DAP kinase phosphorylates MLC II on Ser19 in response to H₂O₂. Thus, we compared the sequence of MLC II with that of tropomyosin-1 and found a high level of similarity between the sequence spanning amino acids 13 to 20 within MLC II and that encompassing amino acids 276 to 284 within tropomyosin-1. In particular, Ser19 within MLC II is preceded by a threonine residue, just as is the penultimate Ser283 within tropomyosin-1 (Fig. 7A). Incidentally, mass spectrometry analysis of rh-tropomyosin-1 phosphorylated by purified activated DAP kinase confirmed that Ser283 within tropomyosin-1 is phosphorylated by DAP kinase (Fig. 7B). Accordingly, we mutated Ser283 to Ala (Ser283Ala) and asked whether this tropomyosin-1 mutant was still phosphorylated in response to H₂O₂. HEK293 cells were transfected with the wild-type or mutant forms of tropomyosin-1 or with plasmids expressing GFP and then were incubated with H₃[³²P]O₄ and were treated or not for 30 minutes with 250 μM H₂O₂. Thereafter, tropomyosin-1 was immunoprecipitated, and the immunoprecipitated proteins were separated by SDS-PAGE. We found that H₂O₂ induced strong phosphorylation of wild-type tropomyosin-1, but that no phosphorylation of tropomyosin-1 over the basal level was found in the Ser283Ala mutant protein (Fig. 7C, upper panel). The basal level probably arose from the small amount of endogenous tropomyosin-1 that was immunoprecipitated (Fig. 7B, lower panel). These results suggest that Ser283 is the sole site that is phosphorylated by DAP kinase in response to H₂O₂.

H₂O₂ induces an increase in actin polymerization and triggers actin remodeling into stress fibers in endothelial cells (Huot et al., 1997; Huot et al., 1998). Intriguingly, the inhibition of the ERK pathway is associated with a disappearance of the H₂O₂-mediated formation of stress fibers that is coincident with a decrease in tropomyosin phosphorylation and with the induction of intense membrane blebbing (Houle et al., 2003). This suggested that phosphorylation of tropomyosin-1 is an important trigger of stress fiber formation. Consistent with this concept, we found that HUVECs expressing the nonphosphorylatable mutant of tropomyosin-1 (Ser283Ala) are devoid of actin stress fibers, whereas the cells that express the pseudophosphorylated mutant of tropomyosin-1 (Ser283Glu) show levels of actin stress fibers higher than those of the cells overexpressing wild-type tropomyosin-1 (Fig. 8). We thus decided next to investigate the involvement of the DAP-kinase-mediated phosphorylation of Ser283 on H₂O₂-induced formation of actin stress fibers. HUVECs were electroporated with siRNA targeting mRNA encoding human DAP kinase to knockdown the expression of DAP kinase (Fig. 9C). Next, plasmids expressing GFP and the nonphosphorylatable tropomyosin-1 mutant Ser283Ala or the pseudophosphorylated mutant Ser283Glu were co-electroporated with siRNA against DAP kinase to knockdown DAP kinase and avoid the influence of the phosphorylation of the endogenous tropomyosin-1. Thereafter, the cells were treated for 30 minutes with 250 μM H₂O₂ and processed for examination of actin by fluorescence microscopy. Consistent with our previous findings (Houle et al., 2003), we found that the addition of H₂O₂ to HUVECs expressing only GFP was associated with the reorganization of cortical actin into transcytoplasmic stress fibers in 50% of the cells (Fig. 9Aa-d and 9B). By contrast, the H₂O₂ treatment was associated with membrane blebbing and did not result in the formation of stress fibers in cells depleted of DAP kinase as well as in cells depleted of DAP kinase and expressing the Ser283Ala mutant protein (Fig. 9Ae-l and 9B). However, the expression of the phosphorylation-mimicking mutant Ser283Glu in the DAP-kinase-depleted cells was associated with the formation of stress fibers in more than 50% of the cells, even in the absence of H₂O₂ treatment. In response to H₂O₂, cells expressing the Ser283Glu mutant exhibited a burst of stress fibers in 60% of the cells (Fig. 9Am-p and 9B), possibly as a result of an increase in actin polymerization following the concomitant activation of the p38-HSP27 pathway (Huot et al., 1998). Overall, these results indicate that phosphorylation of tropomyosin-1 on Ser283 is required for the formation of actin stress fibers in response to H₂O₂, and they confirm that DAP kinase is required to phosphorylate tropomyosin-1 in HUVECs.

Several results drawn from the present study converge towards the conclusion that tropomyosin-1 is a target of ERK signaling in response to oxidative stress. First, in response to H₂O₂, tropomyosin-1 is phosphorylated following its transfection in HEK293 cells that express very low level of tropomyosin-1. Second, FLAG-tagged tropomyosin-1 is phosphorylated in HUVECs treated with H₂O₂. Moreover, in both cases, the phosphorylation is mediated by ERK, being blocked by inhibiting MEK1/2, and thereby ERK activation, with PD098059 or UO126. Nevertheless, given that HUVECs express diverse isoforms of tropomyosins, namely tropomyosin-3 and tropomyosin-4 that are closely related to tropomyosin-1 (Bruneel et al., 2005), one cannot exclude the possibility that, in addition to tropomyosin-1, these forms of tropomyosin could also be phosphorylated after H₂O₂ treatment. In this context, the well-characterized TM311 antibody against tropomyosin that we used to immunoprecipitate phosphorylated tropomyosin-1 recognizes up to three phosphorylated bands that presumably correspond to different tropomyosin isoforms.

Intriguingly, none of the serine or threonine residues of tropomyosin-1 is within the minimal (S/TP) consensus motif for phosphorylation by ERK, and tropomyosin-1 does not encompass the FxFP sequence that mediates the interaction of ERK with its substrate (Jacobs et al., 1999). This suggests that ERK is not the kinase that directly phosphorylates tropomyosin-1. Accordingly, we found that immunoprecipitated ERK from H₂O₂-activated HUVECs does not phosphorylate rh-tropomyosin-1.

One of the major findings of our study is the evidence that DAP kinase functions as a TMK downstream of ERK. This is supported by several observations. First, DAP kinase is activated by H₂O₂ in an in vitro kinase assay, and the activation is sensitive to ML-7. Second, using the same kinase assay, we found that H₂O₂-induced phosphorylation of tropomyosin-1 is increased by the wild-type form of DAP kinase, whereas it is inhibited by a dominant-negative form of the kinase (DAPK K42A). This ability of immunoprecipitated DAP kinase to phosphorylate tropomyosin-1 suggests that the phosphorylation is direct. Third, purified activated DAP kinase phosphorylates tropomyosin-1 in a direct in vitro kinase assay, eliminating the possible involvement of a co-precipitated kinase in the immunocomplex assay. Fourth, the phosphorylation of tropomyosin-1 in HUVECs is impaired in cells in which DAP kinase expression is knocked-down by siRNA and by expressing a dominant-negative form of DAP kinase in HEK293 cells. Finally, the conclusion that DAP kinase activation is downstream of ERK is supported by the finding that it is sensitive to UO126. Moreover, both DAP kinase and ERK co-immunoprecipitate and their association is dependent on ERK activation being increased in the presence of H₂O₂ and is impaired by UO126. Consistent with these findings, it has been reported that DAP kinase is activated downstream of ERK by phosphorylation at Ser735 (Chen et al., 2005; Cohen et al., 1997), and we found that the phosphorylation of DAP kinase is impaired when it harbors a mutation to Ala735 at this site. Interestingly, the death domain of DAP kinase is also required for its activation by ERK (Chen et al., 2005).

MLC is a major substrate for DAP kinase, and the phosphorylation occurs on Ser19 (Kuo et al., 2003). By sequence alignment, we found similarities between the amino acid environment of Ser19 within MLC and Ser283 within tropomyosin-1 (Fig. 7A). Moreover, tropomyosin-α is phosphorylated only on Ser283 in frog and chicken muscle injected with H₃[³²P]O₄ (Mak et al., 1978). Together, these observations suggested that DAP kinase could phosphorylate tropomyosin-1 on Ser283. Accordingly, we found by mass spectrometry analysis that Ser283 within tropomyosin was phosphorylated by active DAP kinase. Accordingly, we generated nonphosphorylatable Ser283Ala and pseudophosphorylated Ser283Glu mutated forms of tropomyosin-1. Using the Ser283Ala mutant, we obtained functional confirmation indicating that Ser283 is the major site within tropomyosin-1 that is phosphorylated in response to H₂O₂. This is supported by the findings that the in vivo phosphorylation of the mutated Ser283Ala form of tropomyosin-1 remains at background levels in response to oxidative stress, whereas the phosphorylation of the wild-type form of tropomyosin-1 is strongly increased. Intriguingly, activation of the β-adrenergic receptor triggers the activation of phosphoinositide 3-kinase, which leads to the phosphorylation of mouse tropomyosin-2 on Ser61, which in turn is required for the internalization of the receptor (Naga Prasad et al., 2005). However, our mass spectrometry analysis does not indicate that Ser61 within tropomyosin-1 is phosphorylated by DAP kinase. Nevertheless, given that tropomyosins forms homodimers, one cannot exclude the possibility that the remnant basal level of phosphorylation emanates from the endogenous tropomyosin that co-precipitates with the exogenous mutated tropomyosin-1 in our experiments.

Another major accomplishment of the present study is to show that DAP kinase, by phosphorylating tropomyosin-1 on Ser283, is responsible for the actin-regulatory functions of tropomyosin-1. Indeed, we show for the first time that the phosphorylation-mimicking mutant of tropomyosin-1 (Ser283Glu) induces by itself the formation of actin stress fibers, whereas the nonphosphorylatable mutant (Ser283Ala) by contrast impairs the formation of actin stress fibers. This is further strengthened by our finding that knockdown of DAP kinase is associated with membrane blebbing in response to H₂O₂ and that the effect is impaired in the presence of the Ser283Glu mutant of tropomyosin-1. In fact, the expression of the exogenous pseudophosphorylated Ser283Glu tropomyosin-1 in cells in which expression of DAP kinase is knocked-down is associated with an increased level of actin stress fibers even in the absence of oxidative stress. By contrast, the expression of exogenous nonphosphorylatable Ser283Ala tropomyosin-1 does not induce actin stress fiber formation and is instead associated with membrane blebbing in the presence of H₂O₂. Incidentally, the finding that DAP-kinase-mediated phosphorylation of tropomyosin-1 on Ser283 underlies its actin-regulatory function constitutes further functional evidence in support of DAP kinase being a TMK and that Ser283 is the site that must be phosphorylated in order to regulate the assembly of stress fibers. Interestingly, the key role played by phosphorylation of tropomyosin-1 in the formation of stress fibers is not restricted to endothelial cells as similar results were obtained in MDA MB231 cells that stably express the tropomyosin mutant proteins (supplementary material Fig. S1). Further studies should be performed to ascertain the exact mechanism by which phosphorylation of tropomyosin-1 on Ser283 contributes to bundling of stress fibers. Theoretically, phosphorylation of tropomyosin-1 could trigger the formation of stress fibers or it might stabilize pre-existing fibers (Cooper, 2002). On the one hand, phosphorylation of tropomyosin-1 might increase the formation of stress fibers by contributing to an increase in the ATPase activity of myosin, thereby increasing the association of actin with myosin. On the other hand, tropomyosin-1 is a coiled-coil of two α-helical polypeptides that binds laterally to seven contiguous actin subunits and head-to-tail to neighbouring tropomyosin-1. By this structural arrangement, it acts as a shield that stabilizes stress fibers, protecting them by preventing breakdown by severing proteins such as gelsolin or by inhibiting actin depolymerization by cofilin. Notably, studies using viscometry have shown that phosphorylation of skeletal tropomyosin leads to stronger head-to-tail interactions between adjacent tropomyosin molecules and, more importantly, stronger interactions with actin (Naga Prasad et al., 2005), which strengthens evidence for the role of the phosphorylation of tropomyosin in the regulation of actin dynamics.

DAP kinase is a pro-apoptotic kinase and a tumour suppressor that is implicated as a transducer of the morphological changes associated with apoptosis through the phosphorylation of MLC on Ser19 (Bialik et al., 2004). We found that the inhibition of ERK and the knockdown of DAP kinase are associated with apoptosis and early membrane blebbing in HUVECs (Huot et al., 1998) (Fig. 9). This suggests that oxidative-stress-induced activation of DAP kinase by ERK, by contrast, impairs apoptosis. In this context, we found that expression of the tropomyosin mutant Ser283Glu but not the Ser283A mutant protects HeLa cells against caspase-3-mediated apoptosis that is induced by the expression of active DAP kinase (supplementary material Fig. S2). These results highlight the dual opposing roles of DAP kinase with regard to apoptosis and are in line with recent findings showing that ERK-mediated phosphorylation of DAP kinase is associated with the suppression of its apoptotic activity in healthy cells and RAS/Raf-transformed cells (Anjum et al., 2005). They further emphasize the point that the role of DAP kinase in apoptosis is far from being clear. For example, in 293T cells, DAP kinase induces morphological changes similar to those seen in apoptosis without inducing any other markers of apoptosis (Wang et al., 2002). Furthermore, in NIH 3T3 cells, serum-induced activation of DAP kinase triggers the formation of transcytoplasmic actin stress fibers before the onset of apoptosis, and the effect is independent of its apoptosis-inducing death domain (Kuo et al., 2003).

In summary, we show, for the first time, that H₂O₂ induces in endothelial cells a time-dependent activation of DAP kinase downstream of ERK MAP kinase. Moreover, we report that DAP kinase, but not MLCK, promotes the in vitro and in vivo phosphorylation of tropomyosin-1 on Ser283 and that this phosphorylation is essential for the H₂O₂-induced organization of the assembly of actin stress fibers in endothelial cells. These findings strongly support a role for the phosphorylation of tropomyosin-1 in regulating the endothelial functions associated with cytoskeletal remodeling in the cardiovascular system.

H₃[³²P]O₄ was purchased from GE Health Care (Montréal, Qc, Canada). H₂O₂, FITC-phalloidin, ML-7, purified porcine MLC and endothelial cell growth supplement (ECGS) were from Sigma-Aldrich (Oakville, On, Canada). PD098059 and UO126 were purchased from Calbiochem (San Diego, CA), and Promega (Madison, WI), respectively, and were diluted in DMSO to make stock solutions of 20 mM. Chemicals for electrophoresis were obtained from Bio-Rad (Mississauga, On, Canada) and Fisher Scientific (Montréal, Qc, Canada). Purified activated MLCK (amino acids 1425-1771) and purified activated DAP kinase (amino acids 1-296) were purchased from Chemicon (Mississauga, On, Canada).

Anti-FLAG M2, anti-DAP kinase (clone DAPK-55), anti-tropomyosin (clone TM311) monoclonal mouse antibodies and rabbit polyclonal anti-FLAG and anti-actin (C-terminus) were purchased from Sigma-Aldrich (Oakville, On, Canada). Anti-living color (GFP) rabbit polyclonal antibody was purchased from BD Biosciences (Mississauga, On, Canada). Anti-HA tag (12Ca5) mouse antibody was purchased from Roche (Laval, Qc, Canada). Anti-MLC-phosphoSer19 rabbit polyclonal antibody was purchased from Cell Signaling (Pickering, On, Canada). Anti-ERK2 is a rabbit polyclonal antibody raised against a synthetic peptide that corresponds to the 14 C-terminal amino acids of rat ERK2 (Huot et al., 1997).

HUVECs were isolated by collagenase digestion of umbilical veins from undamaged sections of fresh cords (Huot et al., 1997). Briefly, the umbilical vein was cannulated, washed with Earle's balanced salt solution (EBSS) and perfused for 10 minutes with collagenase (1 mg/ml) in EBSS at 37°C. After perfusion, the detached cells were collected, the vein was washed with 199 medium and the wash-off pooled with the perfusate. The cells were washed by centrifugation and plated on gelatin-coated 75 cm² culture dishes in 199 medium containing 20% heat-inactivated fetal bovine serum (FBS), ECGS (60 μg/ml), glutamine, heparin and antibiotics. Replicated cultures were obtained by trypsinization and were used at passages <5. The identity of HUVECs as endothelial cells was confirmed by their polygonal morphology and by detecting their immunoreactivity for factor-VIII-related antigens. HEK293 and HeLa cells were cultivated in DMEM containing 10% foetal calf serum. Cultures were incubated at 37°C in a humidified atmosphere containing 5% CO₂.

Tropomyosin-1 cDNA was cloned by PCR amplification from IMAGE clone 562592 (ATCC) into pIRES-hrGFP2a (Stratagene, La Jolla, CA) and pGEX-6P-3 (GE Health Care, Montréal, Canada) vectors using the following primers: 5′-TAGAATTCTATGGACGCCATCAAGAAGAAGATGCAGATGC-3′ and 5′-CCTGCTCGAGTATATGGAAGTCATATCGTTGAGAGC-3′. We proceeded similarly to generate HA- and FLAG-tagged version of tropomyosin-1 into pIRES-hrGFP2a (Stratagene, La Jolla, CA) and pCMV-4A, respectively, using the primers 5′-TAGAATTCTATGGACGCCATCAAGAAGAAGATGCAGATGC-3′ and 5′-TACTCTCGAGTATGGAAGTCATATCGTTGAGAGCGTG-3′. The tropomyosin-1 S283A was generated by PCR site-directed mutagenesis on pIRES-hrGFP2a-tropomyosin-1 construct using the primers 5′-ATGACTGCTATATAACTCGAGTACCCATATGACG-3′ and 5′-TTATATAGCAGTCATATCGTTGAGAGCGTGG-3′. The tropomyosin-1 S283E mutant was generated similarly using the primers 5′-CGATATGACTGAAATAATACTCGAGTACCCATATG-3′ and 5′-CGAGTTATATTTCAGTCATATCGTTGAGAGCG-3′ (NP_000357). DAP kinase constructs were obtained from Ruey-Hwa Chen (National Taiwan University, Taiwan). We purchased from Qiagen (Mississauga, Canada) the siRNA Hs-DAPK1-5 to perform the DAP kinase knockdown experiments.

Gene and siRNA transfer in HUVECs was by performed by electroporation (Le Boeuf et al., 2006). Co-electroporation of EGFP-encoding plasmids allowed evaluating transfer efficiency as being 30% following determination of the percentage of cells expressing EGFP. HEK293 cells were transfected using high molecular weight polyethylenimine (PEI) from Sigma. Plasmids (20 μg DNA/2×10⁶ cells) were mixed with 500 μl NaCl 150 mM and 10 μl PEI (0.43%) for 15 minutes at room temperature and then were put on cells. Cells were treated and extracted 24 hours post-transfection. Co-expression of GFP by using pIRES-rhGFP-2a (Stratagene) or a separate pEGFP construct enabled evaluating the transfection efficiency as being 40%.

After treatments, cells were lysed in 250 μl of buffer B (150 mM NaCl, 50 mM Tris-HCl pH 7.5, 0.5% Triton X-100, 0.1% sodium deoxycholate, 2 mM EDTA, 2 mM EGTA, 1 mM Na₃VO₄, 1 mM leupeptin, 50 μg/ml pepstatin and 1 mM PMSF). Samples were pre-cleared with 10 μl of a 50% v/v protein-G sepharose (GE Health Care, Montréal, Canada) suspension for 30 minutes. Supernatants were incubated overnight with 12 μl of mouse monoclonal antibody against tropomyosin. Then 10 μl of 50% v/v protein-G sepharose were added and incubation was extended for 30 minutes on ice with shaking. Antibody-antigen complexes were washed four times with buffer B and SDS-PAGE loading buffer was added. Proteins were separated through SDS-PAGE and the gel was dried or transferred onto nitrocellulose for western blotting using the antibody against tropomyosin.

Cells were extracted in 10 mM phosphate buffer pH 7.6 containing 100 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 5 mM EDTA, 1 μM leupeptin, 1 mM Na₃VO₄, 1 mM benzamidine, 1 mM PMSF and 10 mM NaF. DAP kinase activity was assayed as previously described (Cohen et al.,1997) in immune complexes after immunoprecipitation (for 2 hours only) of transfected cell extracts using the mouse monoclonal anti-FLAG M2 anti-DAP kinase or anti-ERK2. The DAP kinase assays were carried out in 36 μl of kinase buffer K: 50 mM ATP, 5 μCi of [γ³²P]-ATP (used only with tropomyosin as substrate), 50 mM Hepes pH 7.5, 8 mM MgCl₂, 2 mM MnCl₂, 0.1 mg/ml Bovine Serum Albumin (BSA) 0.5 mM CaCl₂, 1 μM Calmodulin. Purified porcine Myosin Light Chain (MLC) or purified recombinant human recombinant tropomyosin-1 (rh-tropomyosin-1) were used as substrates. The ERK assay was carried out as previously described (Houle et al., 2003) but MBP or rh-tropomyosin-1 was used as substrate. The kinase activity was assayed for 30 minutes at 30°C and the reaction was stopped by the addition of SDS-PAGE loading buffer. The reaction mixture was separated by SDS-PAGE and transferred to nitrocellulose membrane. The dry membrane was analyzed using PhosphorImager (Molecular Dynamics) in which [γ-³²P]ATP was used. Immunodetection of phospho-Ser19 within MLC was performed to monitor phosphorylation of MLC.

A direct in vitro kinase assay of MLCK was performed by adding the constitutively active form of MLCK to 4 μg of purified porcine MLC or 4 μg of purified rh-tropomyosin-1 for 30 minutes at 30°C. Assays were performed in the following buffer: 40 mM Hepes pH 7.5, 10 mM MgCl₂, 0.5 mM CaCl₂, 100 μM ATP, 1 μM calmodulin and 5 μCi [γ-³²P]ATP. For DAP kinase, the assay was done similarly by adding the constitutively active form of DAP kinase to the same substrates for 30 minutes at 30°C. DAP kinase assays were performed in the following buffer: 50 mM Hepes pH 7.5, 8 mM MgCl₂, 2 mM MnCl₂, 0.1 mg/ml bovine serum albumin, 0.5 mM CaCl₂, 50 μM ATP, 1 μM calmodulin and 5 μCi [γ-³²P]ATP. The reactions were stopped by the addition of SDS loading buffer. Thereafter, the proteins were run through SDS-PAGE and transferred onto nitrocellulose membranes. Phosphorylation of MLC or rh-tropomyosin-1 was evaluated by monitoring the incorporation of ³²P into the corresponding band on the membrane using PhosphorImager (Molecular Dynamics).

HUVECs or HEK293 cells plated in Petri dishes were incubated in PO₄-free medium for 1 hour before labeling with H₃[³²P]O₄ (60-200 μCi/ml) for 90 minutes. Cells were treated and were then extracted and processed for immunoprecipitation experiments. In some experiments, NaF (1 mM for 1 hour) was added to the incubation medium to inhibit phosphatases and to allow the accumulation of phosphorylated tropomyosin. Endothelial cells remain 100% viable after 2 hours of treatment with a concentration of NaF of up to 20 mM (Wang et al., 2001).

Cells were extracted in IEF buffer (4% CHAPS, 20 mM dithiothreitol, 0.5% IPG buffer 4.5-5.5 (GE Health Care, Montréal, Canada), 7M urea). Next, extracts were spun for 10 minutes at 16,000 g. Supernatants were run for at least 16 hours on 18 cm Immobiline Drystrip pH 4.5-5.5 into IPGPhor electrophoresis apparatus (GE Health Care, Montréal, Canada). Gels were incubated for 10-15 minutes in loading buffer before being run by 8.5% SDS-PAGE. Thereafter, gels were transferred on nitocellulose membranes and were processed for immunodetection.

Purified active DAP kinase was used to phosphorylate in vitro rh-tropomyosin 1. Following fractionation by SDS-PAGE, tropomyosin-1 was excised from a Coomassie-stained gel. After digestion, samples were dried down and reconstituted in formic acid (0.2%) and ACN (5%) prior to injection in the LC-MS system. The LC system consists of an in-house 300-μm × 5-mm C₁₈ precolumn (Waters, Millford, MA) and a 150-μm × 10-cm in-house C₁₈ analytical column (Jupiter 5 μm, 300 Å, Phenomenex, Torrance, CA). Peptide elution was performed at 600 nl/min. Electrospray mass spectra were recorded on a quadrupole time-of-flight mass spectrometer Q-TOF Ultima (Waters, Millford, MA). Calibration of the instrument was made by infusion of a Glu-Fib B (Sigma, St Louis, MO) solution at 83 fmol/μl. For tandem mass spectra, collisional activation of selected precursors was obtained using argon as a target gas at collision energies of 30 eV (laboratory frame of reference). Fragment ions formed in the RF-only quadrupole were recorded by a time-of-flight mass analyzer.

Confocal microscopy was used for immunofluorescent visualization of F-actin and EGFP (Huot et al., 1997). The cells were plated on gelatin-coated LabTek dishes (Nalge Nunc, Naperville, IL). After treatment, they were fixed with 3.7% formaldehyde and permeabilized with 0.1% saponin in phosphate buffer, pH 7.5. F-actin was detected using Alexa-488-phalloidin (Invitrogen, Carlsbad, CA) diluted 1:400 in phosphate buffer. Living color rabbit antibody was used for EGFP staining. The antigen-antibody complexes were detected with an Alexa-568-rabbit-IgG secondary antibody. The cells were examined by confocal microscopy with a Bio-Rad MRC-1024 imaging system mounted on a Nikon TE2000 equipped with a 60× objective lens with a 1.4 numerical aperture (Le Boeuf et al., 2006).

We thank Ruey-Hwa Chen (National Taiwan University, Taiwan) for providing the DAP kinase plasmids. We thank Eric Bonneil at the Institute for Research in Immunology and Cancer (IRIC) at the University of Montreal, Canada, for providing technical assistance in mass spectrometry. We also thank Simon Rousseau (Dundee University, UK) for critical reading of the manuscript.