Downregulation of adherens junction proteins is a frequent event in carcinogenesis. How desmosomal proteins contribute to tumor formation by regulating the balance between adhesion and proliferation is not well understood. The desmosomal protein plakophilin 1 can increase intercellular adhesion by recruiting desmosomal proteins to the plasma membrane or stimulate proliferation by enhancing translation rates. Here, we show that these dual functions of plakophilin 1 are regulated by growth factor signaling. Insulin stimulation induced the phosphorylation of plakophilin 1, which correlated with reduced intercellular adhesion and an increased activity of plakophilin 1 in the stimulation of translation. Phosphorylation was mediated by Akt2 at four motifs within the plakophilin 1 N-terminal domain. A plakophilin 1 phospho-mimetic mutant revealed reduced intercellular adhesion and accumulated in the cytoplasm, where it increased translation and proliferation rates and conferred the capacity of anchorage-independent growth. The cytoplasmic accumulation was mediated by the stabilization of phosphorylated plakophilin 1, which displayed a considerably increased half-life, whereas non-phosphorylated plakophilin 1 was more rapidly degraded. Our data indicate that upon activation of growth factor signaling, plakophilin 1 switches from a desmosome-associated growth-inhibiting to a cytoplasmic proliferation-promoting function. This supports the view that the deregulation of plakophilin 1, as observed in several tumors, directly contributes to hyperproliferation and carcinogenesis in a context-dependent manner.
It is widely accepted that intercellular adhesion counteracts tumorigenesis by contact inhibition of cell growth. Loss of cell adhesion and alterations in the expression of cadherins are common features of malignant cells and markers for aggressive tumor growth. Whereas the role of adherens junctions and classical cadherins in tumorigenesis is well documented, studies analyzing the expression of desmosomal proteins in human tumors are contradictory with upregulation, downregulation or maintenance of desmosomal protein expression reported previously (Chidgey and Dawson, 2007; Dusek and Attardi, 2011). Attempts to clarify the role of desmosomes in the control of proliferation and apoptosis in vitro revealed that the enhanced expression of desmosomal components promoted proliferation, inhibited apoptosis and increased invasion in some cases (Hakimelahi et al., 2000; Kolligs et al., 2000; Furukawa et al., 2005). In contrast, a suppression of tumor-promoting characteristics such as anchorage-independent growth was proposed by other studies (Tselepis et al., 1998; Winn et al., 2002). These controversies can be explained by context-dependent regulation and function of desmosomal components. However, signaling pathways regulating the role of desmosomal proteins in adhesion or cell signaling are not well characterized. Inhibitors that block the epidermal growth factor receptor promoted desmosome assembly along with a change in morphology from a mesenchymal to an epithelial appearance, suggesting that growth factor signaling counteracts desmosomal adhesion (Lorch et al., 2004).
Desmosomes provide mechanical stability but in addition are thought to function in cell–cell communication and signal transmission. For example, desmosomal cadherins reveal distinct expression patterns in the epidermis and play a role in the morphoregulation of the skin. The plakophilins 1 to 3 (PKP1–PKP3) show tissue-specific expression patterns similar to the desmosomal cadherins, with PKP1 being predominantly expressed in the suprabasal layers of stratified epithelia (Hatzfeld, 2007; Bass-Zubek et al., 2009; Neuber et al., 2010). They function as scaffolds by interacting with various desmosomal proteins to promote desmosomal adhesion (Hatzfeld et al., 2000; Chen et al., 2002; Bonné et al., 2003; Hatzfeld, 2007; Bass-Zubek et al., 2009). Overexpression of PKP1 increased desmosome number and size in tissue cultured cells (Kowalczyk et al., 1999; Hatzfeld et al., 2000), and the lack of PKP1 in patients suffering from ‘ectodermal dysplasia/skin fragility syndrome’ (OMIM number 604536) correlated with reduced desmosome number and size, supporting its importance in regulating desmosomal adhesion in the skin (McGrath et al., 1997; McGrath and Mellerio, 2010). The localization of PKP1 in the cytoplasm and the nucleus indicates a dual-compartment role of the protein which likely involves functions in modulating cell signaling. In the cytoplasm, PKP1 promotes translation by interacting with components of the translation initiation complex (Wolf and Hatzfeld, 2010; Wolf et al., 2010). PKP1 stimulates the activity of the RNA helicase eIF4A leading to the upregulation of protein synthesis. This correlates with an increase in cell size and cell proliferation, events that are common in carcinogenesis (Bader and Vogt, 2004; Bjornsti and Houghton, 2004). Thus, PKP1 combines a function in strengthening desmosomal adhesion with a complementary role in promoting translation. As a consequence, it could either promote tumorigenesis by stimulating translation or act as a tumor suppressor by enhancing intercellular adhesion. Therefore, it is essential to elucidate how the equilibrium between these two distinct roles of PKP1 is regulated to finally unravel the context-dependent functions of PKP1.
Translation rates are primarily controlled at the initiation step. Growth factor signaling cascades modulate the activity of translation initiation factors by controlling their phosphorylation (Raught et al., 2004; Svitkin et al., 2005; Mamane et al., 2006; She et al., 2010). The phosphatidylinositol-3-kinase (PI3K) plays a key role in signaling downstream of many receptors such as those for insulin and insulin-like growth factors. Activated PI3K converts phosphatidylinositol-4,5-bisphosphate (PIP2) to phosphatidylinositol-3,4,5-trisphosphate (PIP3). The PIP3-dependent kinases PDK1 and Akt/protein kinase B (PKB) induce a kinase cascade that plays a central role in the regulation of translation (Osaki et al., 2004). Akt occurs in three isoforms, of which Akt1 (PKBα) is the dominant form whereas Akt2 (PKBβ) is predominantly expressed in insulin-responsive tissues and Akt3 (PKBγ) is highly expressed in testis and brain (Altomare et al., 1998; Nakatani et al., 1999). By phosphorylation of several downstream targets, Akts promote cell survival, proliferation and cell growth (Muise-Helmericks et al., 1998; Datta et al., 1999).
In this manuscript, we report that the equilibrium between the desmosomal and cytoplasmic pools of PKP1 is regulated by insulin/IGF1 signaling. Upon activation of the PI3K/Akt pathway, PKP1 was phosphorylated by Akt2, resulting in its delayed degradation and a cytoplasmic accumulation. This predominant cytoplasmic localization of PKP1 correlated with reduced intercellular adhesion but enhanced mRNA translation, proliferation, migration and anchorage-independent growth. These data suggest that PKP1 may contribute to tumorigenesis in a context-dependent way.
PKP1 interacts with Akt2
A yeast two-hybrid (Y2H) screen revealed a putative interaction of PKP1 with Akt2, suggesting that PI3K-dependent signaling is involved in modulating functions of PKP1 in desmosomal adhesion versus translation. The initial Y2H results were validated by probing the association of all three PKP head and armadillo domains with Akt1 and Akt2. Akt3 was not included since its expression in normal tissue does not overlap with that of PKP1. For Akt1 no interaction with any of the PKPs was determined by Y2H analyses. In contrast, Akt2 associated specifically with the N-terminal head domain of PKP1, but not with the other PKPs (Fig. 1A). GST pull-downs with recombinant proteins confirmed a direct interaction between PKP1 and Akt2, and showed a considerably weaker association of PKP1 and Akt1 (Fig. 1B). In transiently transfected HEK293 cells, PKP1 co-immunopurified with Akt2 upon insulin stimulation, whereas hardly any association was observed in unstimulated cells. In contrast, Akt1 was not co-immunoprecipitated with PKP1 (Fig. 1C). This indicates that PKP1 associates preferentially with activated Akt2. In agreement, bimolecular fluorescence complementation (BiFC) performed in MCF-7 cells, which have numerous desmosomes but essentially no endogenous PKP1, confirmed that PKP1 only associated with wild type (wt) Akt2 and a constitutively active mutant (Akt2ca T309D/S474D). In contrast, association with kinase-inactive Akt2 (Akt2dn K181M/T309A/S474A) was reduced to background levels as observed for Akt1 (Fig. 1D). Notably, PKP1 association was observed throughout the cell and along the plasma membrane reflecting the localization of the exogenous proteins. In conclusion, these data indicate that PKP1 associates with Akt2 in an activation-dependent manner.
Akt2 phosphorylates PKP1 in vitro
The association of PKP1 with active Akt2 suggested PKP1 as a substrate of this kinase. To test this, in vitro phosphorylation assays were performed using recombinant PKP1 and active Akt2 as well as GSK3-β kinases. The latter was analyzed since this kinase was shown to regulate the PKP1-related cell contact proteins β-catenin (Ikeda et al., 1998) as well as p120ctn (Hong et al., 2010a). Akt2 phosphorylated PKP1 in vitro as demonstrated by 32P-incorporation, whereas GSK3-β did not (Fig. 2A). Auto-phosphorylation of GSK3-β confirmed that the kinase was active. To map the phosphorylation sites we used a PKP1 peptide microarray together with recombinant active Akt2 (Fig. 2B). Additionally, recombinant PKP1 was phosphorylated in vitro and subjected to mass spectrometry to identify the phosphorylation sites. These analyses led to the identification of four putative phosphorylation motifs in the PKP1 N-terminal domain, with 9, 2, 4 and 9 serine (S) or threonine (T) residues in the individual motifs (Fig. 2B and supplementary material Table S1).
Growth factor signaling stimulates PKP1 phosphorylation
To corroborate these findings, we next analyzed PKP1 phosphorylation upon Akt2 upregulation in transfected HEK293 cells. PKP1 was overexpressed together with activated Akt2 [myristoyl (myr)-Akt2] and immunoprecipitated. The precipitate was analyzed using an anti-phospho-PKA substrate antibody that was expected to recognize S118 located in motif 3 of PKP1 (Fig. 2C). Phosphorylation was observed for PKP1wt, but not for a mutant protein in which the S118 was converted to alanine, indicating that the antibody recognizes motif 3 (Fig. 2D). Notably, a band shift was observed for wt as well as mutant PKP1 upon co-transfection with Akt2, suggesting that the kinase phosphorylated additional sites in PKP1. In cells not transfected with active Akt2, PKP1 immunoprecipitates were not detected by the phospho-PKA substrate antibody (Fig. 2D), indicating that Akt2 modified PKP1 at S118. In agreement with the interaction studies, we observed that only myr-Akt2 but not myr-Akt1 efficiently phosphorylated PKP1 at this site (Fig. 2E). Phosphatase treatment of cell extracts confirmed that the band shift of PKP1 observed after co-transfection with active Akt2 depended on PKP1 phosphorylation (Fig. 2F). Phosphorylation of PKP1 at S118 was validated after insulin stimulation in PKP1-overexpressing HEK293 cells (Fig. 2G) as well as for the endogenous PKP1 after insulin and IGF1 but not after EGF stimulation of A431 cells (Fig. 2H). In summary, these studies indicate that PKP1 becomes phosphorylated by Akt2 in response to activation of the PI3K/Akt pathway.
PKP1 intracellular localization is regulated by insulin signaling
To characterize how phosphorylation of PKP1 modulates its cellular functions, we investigated the subcellular localization of the protein. In MCF-7 cells overexpressing activated myr-Akt2 together with PKP1, PKP1 localized preferentially in the cytoplasm. In contrast, coexpression of dominant-negative (kinase-inactive) Akt2 resulted in a punctate staining pattern of PKP1 at the plasma membrane, which is characteristic for its desmosome association. Moreover, PKP1 was essentially depleted from the cytoplasm in these cells (Fig. 3A). This indicates that Akt2-dependent signaling modulates the subcellular localization of PKP1 and suggests that Akt2-mediated phosphorylation enhances its cytoplasmic accumulation. In order to analyze if the shift in the PKP1 intracellular localization observed upon the overexpression of active Akt2 is also induced by growth factor signaling, we stimulated keratinocytes with either EGF, insulin or IGF1. An increased cytoplasmic pool of the endogenous PKP1 was observed after IGF1 and insulin but not after EGF treatment (Fig. 3B) in agreement with the finding that EGF treatment did not result in PKP1 phosphorylation at S118. To investigate if the cytoplasmic translocation elicited by insulin was dependent on Akt2 we combined insulin stimulation with Akt2 knockdown. In cells depleted of Akt2 no change in the PKP1 localization was observed upon insulin stimulation. In contrast, a considerable accumulation of PKP1 in the cytoplasm was detected in those cells that were not Akt2-siRNA transfected and still expressed Akt2 (Fig. 3C). This clearly indicates that insulin signaling induces an Akt2-dependent translocation of PKP1 from the desmosomes to the cytosol.
PKP1 intracellular localization is regulated by phosphorylation
Since these analyses did not allow distinguishing between direct or secondary effects of Akt2 or insulin we used site-directed mutagenesis to construct non-phosphorylatable variants of PKP1 (A-mutant; by exchanging S/T to alanine, A) on the one hand and phospho-mimetic mutants (E-mutant; S/T to glutamic acid, E) on the other hand to analyze the effects of PKP1 phosphorylation more directly. In order to determine the influence of each motif, we prepared single-motif mutants as well as mutants in which all four motifs were mutated (for nomenclature and sequence of the mutants see supplementary material Fig. S1). Their localization was analyzed in MCF-7 cells that contain numerous desmosomes but express virtually no PKP1 excluding bias by interference of endogenous PKP1. In these cells, all PKP1 mutants co-localized with desmosomes in the same manner as PKP1wt, indicating that no major folding defects were introduced by the mutations (Fig. 4A and supplementary material Fig. S2A). However, while A-mutants were almost exclusively at the desmosomes, E-mutants of the 1st and 4th motif (PKP1M1E and PKP1M4E) as well as PKP1M1-4E revealed considerably elevated cytoplasmic pools (Fig. 4A). The 2nd motif had no and the 3rd motif only a weak effect on the subcellular localization of PKP1 (supplementary material Fig. S2A). To exclude bias caused by varying expression levels of the individual mutants we compared expression levels by western blot analysis (supplementary material Fig. S2B). This analysis revealed increased protein amounts for PKP1M1E, PKP1M4E and PKP1M1-4E mutants suggesting that in MCF-7 cells PKP1M1E and PKP1M4E become stabilized whereas PKP1M2 and PKP1M3 are not (see below). All other mutants showed very similar expression levels which were similar to the amount of endogenous PKP1 observed in HaCaT cells (Fig. 4B).
The effect on the intracellular localization was confirmed in HaCaT keratinocytes which express endogenous PKP1. In these cells, elevated cytoplasmic localization was observed for M1E and M4E as well as M1-4E PKP1 mutants, whereas the non-phosphorylatable A-mutants associated exclusively with desmosomes (supplementary material Fig. S2C). Notably, cells expressing the PKP1M1-4E mutant tended to form extensive lamellipodia and long protrusions (Fig. 4A, arrowheads). Consistently, many of these cells detached from the monolayer and formed protrusions on top of their neighboring cells. In regions of lamellipodia formation desmosomes were reduced. This resembled the phenotype induced by overexpression of the PKP1 arm repeat domain (Hatzfeld et al., 2000) and may suggest that the cytosolic PKP1M1-4E mutant as well as the arm repeat domain modulate actin organization, possibly via Rho signaling (Hatzfeld, 2007). In summary, these findings support the view that non-phosphorylatable PKP1 associated preferentially with desmosomes, whereas phosphorylation as mimicked by the negative charge of the E-mutants promoted the accumulation of PKP1 in the cytoplasm.
Since we expected a higher solubility of the PKP1 cytoplasmic pool compared to the desmosomal one, we used mild detergent extraction of transfected MCF-7 cells to evaluate the solubility of PKP1 mutants. As expected, PKP1wt and the M1-4A mutant were essentially insoluble, whereas abundance of the M1-4E mutant was considerably elevated in the soluble fraction (Fig. 4C). The solubility of keratins was not affected. In conclusion, these studies indicate that Akt2-dependent phosphorylation of PKP1 enhances its cytoplasmic localization.
PKP1 also localizes in the nucleus where it was suggested to interact with ssDNA (Sobolik-Delmaire et al., 2010). In order to analyze if the phospho-site mutations affected its nuclear localization cells displaying nuclear PKP1 were counted. Both mutants revealed reduced abundance in the nucleus (supplementary material Fig. S2D). These results indicate that the increased solubility of the M1-4E mutant depended on its cytoplasmic and not on its nuclear pool.
To exclude abnormal sorting due to a shortage of interaction partners in cells that normally do not express PKP1 (MCF-7) or express PKP1 mutants in addition to the endogenous protein we used 3′-UTR-based knockdown of the endogenous PKP1 in HaCaT cells (Fig. 4D) followed by the overexpression of PKP1M1-4A and PKP1M1-4E. No differences in the localization were observed between control knockdown and PKP1 knockdown cells (Fig. 4E).
Phosphorylated PKP1 is more resistant to degradation
Our analysis suggested that the coexpression of Akt2 with PKP1 not only modified the latter but also enhanced PKP1 protein abundance (see Fig. 1C, Fig. 2D–F). In agreement, we observed that the phospho-mimetic mutant PKP1M1-4E accumulated at the protein level although mRNA levels were similar for all mutants (Fig. 5A and supplementary material Fig. S2B, Fig. S3A,B). This suggested that the modification of PKP1 by Akt2 enhanced the stability of the protein. To test this hypothesis, we determined PKP1 protein degradation after inhibiting protein synthesis using cycloheximide. The protein level of the M1-4A mutant was reduced to 50% after ∼5 hours, whereas the M1-4E mutant was very slowly degraded with >70% remaining at 9 hours (Fig. 5B). PKP1wt showed an intermediate degradation rate. Moreover, PKP1wt protein degradation was reduced by the coexpression of myr-Akt2, whereas the turnover of Myc protein remained essentially unaffected (Fig. 5C). Likewise, the turnover of PKP1 was reduced upon insulin stimulation of serum starved HEK293 cells (Fig. 5D). Hence, phosphorylation of PKP1 by Akt2 retarded its degradation and promoted its cytoplasmic accumulation.
PKP1 phosphorylation reduces intercellular adhesion and increases cell migration
The cytoplasmic accumulation of PKP1M1-4E and PKP1wt observed upon coexpression of Akt2 raised the question if phosphorylation affects intercellular adhesion. Dispase was used to detach confluent monolayers of MCF-7 cells stably expressing either PKP1wt, PKP1M1-4A or PKP1M1-4E before probing intercellular adhesion by subjecting the cells to mechanical stress. Whereas the monolayer of PKP1M1-4A-expressing cells remained essentially intact under the experimental conditions, the PKP1M1-4E-expressing cell layers dissociated into numerous fragments, indicating that the intercellular adhesion was reduced. PKP1wt-expressing cells revealed an intermediate phenotype, as indicated by the dissociation into fewer fragments when compared to cells expressing the phospho-mimetic mutant (Fig. 6A). PKP1wt-expressing cells were also subjected to insulin stimulation. This treatment strongly reduced intercellular adhesion. Inhibition of PI3K upstream of Akt reverted the effect (Fig. 6B).
Next we performed FRAP (fluorescence recovery after photobleaching) studies to determine the dynamic exchange of PKP1 mutants between distinct subcellular pools (Fig. 6C). In the desmosomal pool, PKP1M1-4A revealed a half-life of ∼12.5 seconds and a large immobile fraction (If = 0.63) compared to PKP1wt with a half-life of 6.6 seconds and a considerably smaller immobile fraction (If = 0.4). This supported a stable association of non-phosphorylatable PKP1 with desmosomes as suggested by the enrichment of PKP1M1-4A at these sites (see Fig. 4A) as well as the reduced solubility of the mutant protein (see Fig. 4C). In contrast, PKP1M1-4E showed a significantly reduced half-life of only ∼2.4 seconds with a small immobile fraction (If = 0.26) in desmosomes, suggesting an elevated exchange rate and a considerably reduced stability of its desmosomal association (Fig. 6C). In the cytoplasm and the nucleoplasm, diffusion rates were similar for wt and mutant PKP1s (supplementary material Fig. S4). These studies suggested that un-phosphorylated PKP1 is stabilized in the desmosome and promotes intercellular adhesion.
Since intercellular adhesion and cell migration are interdependent, we next analyzed the capacity of MCF-7 cells stably expressing PKP1 mutants to migrate into a cell-free gap (‘wound’) (Fig. 6D). Whereas PKP1wt- and M1-4E-expressing cells were able to close the ∼500 µm gap within 18 hours, a gap of ∼150–200 µm remained in the M1-4A-expressing cells suggesting that migration into the cell-free area was reduced (Fig. 6D, upper panel). Quantitative assessment of wound closure over time excluded bias by altered cell proliferation and confirmed that PKP1wt and M1-4E showed significantly elevated migratory capacity. In agreement, single-cell tracking revealed a reduction in the traveled distances for the PKP1M1-4A mutant compared to the M1-4E mutant (Fig. 6E). In conclusion, these analyses provide strong evidence that Akt2-directed phosphorylation of PKP1 interferes with intercellular adhesion and enhances the migratory capacity, presumably by inducing the cytoplasmic relocalization of PKP1.
PKP1 phosphorylation promotes translation and proliferation and confers the capacity for anchorage-independent growth
The cytoplasmic pool of PKP1 is supposed to stimulate translation (Wolf et al., 2010). This was tested by probing how PKP1 mutants affected the translation of cap-dependent luciferase reporters. To exclude bias by transfection variability, luciferase activity was normalized to reporter mRNA abundance. In agreement with the localization studies, the previously reported stimulation of reporter mRNA translation was significantly enhanced by PKP1M1-4E when compared to PKP1wt or PKP1M1-4A (Fig. 7A). We next asked if coexpression of active Akt2 would influence the effect of PKP1 on translation. In serum-starved cells, the effect of PKP1wt overexpression on mRNA translation was weaker than in serum stimulated cells (∼150% compared to ∼200%, Fig. 7A,B), suggesting that growth factor signaling affects the function of PKP1 in protein synthesis. Overexpression of myr-Akt2 alone enhanced translation only to ∼125%, whereas coexpression of myr-Akt2 and PKP1 upregulated translation of the reporter to ∼250%, indicating that the function of PKP1 in enhancing cap-dependent mRNA translation is stimulated by active Akt2 (Fig. 7B). Insulin stimulation of serum starved cells for 1 hour led to a minor upregulation of translation, whereas in PKP1wt-expressing cells the stimulation increased translation rates to ∼240% (Fig. 7C).
In vitro translation assays were performed to elucidate the role of PKP1 phosphorylation more directly. The encephalomyocarditis virus (EMCV) internal ribosomal entry site (IRES) depends on the same translation initiation factors as cap-dependent translation, and previous experiments indicated a stimulation of EMCV IRES-dependent translation by PKP1wt (Wolf et al., 2010). Although recombinant PKP1M1-4E was slightly more efficient in stimulating EMCV IRES-dependent translation than PKP1M1-4A the effect was much smaller than observed in a cellular context and was not significant (Fig. 7D). To rule out that this discrepancy was due to a folding defect of the recombinant M1-4E mutant we used in-vitro-phosphorylated PKP1wt compared to non-phosphorylated PKP1wt in the translation assay. In agreement, phosphorylated PKP1wt was not significantly more efficient in stimulating translation than un-phosphorylated PKP1. Taken together, these experiments demonstrate that both phosphorylated and un-phosphorylated PKP1 are able to stimulate translation in vitro to a similar extent, whereas in cells the phospho-mimetic mutant was much more efficient in upregulating protein synthesis. This suggested that the relocalization of phosphorylated PKP1 to the cytoplasm is essential for this function.
Translation and proliferation are tightly coupled processes and previous data indicated that loss of PKP1 expression reduced translation as well as cell proliferation (Wolf and Hatzfeld, 2010; Wolf et al., 2010). Therefore, we analyzed if the stimulation of translation by the phospho-mimetic mutants directly correlates with their capacity to stimulate proliferation. BrdU incorporation assays revealed slightly increased proliferation rates in PKP1wt- and PKP1M1-4A-expressing MCF-7 cells compared to control transfected cells (1.6- and 1.4-fold, respectively, Fig. 7E). In contrast, BrdU incorporation was severely (2.4-fold) upregulated in cells expressing the PKP1M1-4E mutant. To further validate these data we used a colorimetric growth assay based on crystal violet uptake to quantify cell numbers (Gillies et al., 1986). This confirmed that proliferation was slightly reduced in PKP1M1-4A-expressing cells whereas it was considerably upregulated by the PKP1M1-4E mutant (Fig. 7F). These findings correlated well with the predominant desmosomal association of PKP1wt and PKP1 A-mutants and an enhanced cytoplasmic pool of PKP1 E-mutants.
Anchorage-independent growth in soft agar is one of the hallmark characteristics of cellular transformation. Since the PKP1M1-4E mutant revealed increased proliferation rates and migratory capacity, which are both important features acquired by tumor cells, we tested if MCF-7 cells stably overexpressing PKP1M1-4E also acquired the ability to grow in soft agar. The expression of PKP1M1-4E resulted in a significant increase in the number and size of colonies formed whereas PKP1M1-4A-expressing cells formed very little colonies in soft agar (Fig. 7G).
PKP1 mutations modulate its interactions with desmosomes but not with eIF4A
We next asked if phosphorylation would modulate PKP1's protein interactions. Desmoplakin (DP) and desmoglein1 (Dsg1) are the preferential binding partner of PKPs in the desmosome (Kowalczyk et al., 1999; Hatzfeld et al., 2000; Hofmann et al., 2000). GST pull-down experiments were performed with PKP1wt, PKP1M1-4A and PKP1M1-4E and the DP-N-terminal peptide (DP-NTP) as well as with the Dsg1 cytoplasmic domain (Dsg1-CD). Both mutants were able to associate with DP and Dsg1 in vitro (supplementary material Fig. S5A), although the interaction of M1-4E appeared much weaker. Probing the PKP1–eIF4A interactions revealed no differences between the mutants and the wt protein (supplementary material Fig. S5A) indicating that the association did not depend on the PKP1 modification. BiFC studies confirmed that both mutants interacted with DP as well as with eIF4A (supplementary material Fig. S5B), supporting the notion that these mutants were functional and not misfolded. However, no interaction was detected between the Dsg1 cytoplasmic tail and PKP1M1-4E, suggesting that this interaction critically depended on the un-phosphorylated PKP1. Taken together, these experiments indicate that PKP1 phosphorylation is neither required nor does it prevent the association with DP or eIF4A but regulates its interaction with Dsg1. The finding that the PKP1–eIF4A interaction was not affected is in agreement with the in vitro translation assays that revealed only minor differences between PKP1wt, A- and E-mutants concerning their capacity to stimulate translation and further supports our conclusion that PKP1 functions are regulated predominantly via its Akt2-dependent subcellular relocalization.
PKPs are considered as multi-functional proteins that localize in several cellular compartments. A function of PKP1 in regulating desmosomal adhesion is well established and supported by in vitro studies (Kowalczyk et al., 1999; Hatzfeld et al., 2000; South et al., 2003; South, 2004) as well as by the phenotype of a genetic skin disease caused by PKP1 mutations (McGrath and Mellerio, 2010). An additional cytoplasmic function of PKP1 in the regulation of translation and proliferation has recently been reported (Wolf et al., 2010). Here, we have characterized how PI3K/Akt2 signaling regulates the equilibrium between the desmosomal and cytoplasmic functions of PKP1. This is the first report showing the impact of growth factor signaling on a desmosomal protein at the molecular level. Importantly, these modifications affect not only the subcellular localization of PKP1 but exert a major effect on intercellular adhesion, cell migration, proliferation rates and significantly affect anchorage-independent cell growth, characteristics that are highly relevant in the context of tumorigenesis.
The effect of phosphorylation via Akt2 on the intracellular localization of PKP1 was conserved in several cell types. Coexpression of PKP1 and Akt2 induced relocalization of PKP1 to the cytoplasm. Consistently, non-phosphorylatable PKP1 was enriched in desmosomes, whereas the phospho-mimetic mutants of PKP1 accumulated in the cytoplasm. This suggests that un-phosphorylated PKP1 promotes desmosomal adhesion, whereas the phosphorylated form, even though able to associate with desmosomes, functions predominantly in the cytoplasm.
The fact that mRNA translation was increased by PKP1 phospho-mimetic mutants, the coexpression of PKP1 with active Akt2 as well as by insulin stimulation indicates that the phosphorylation of PKP1 enhances protein synthesis. Since in vitro assays did not reproduce the strong effect of PKP1 phosphorylation on mRNA translation, we conclude that the cellular regulation of PKP1 essentially relies on the control of its subcellular localization. Obviously, phosphorylation interferes not only with its binding to Dsg1 but also with the degradation of PKP1 and thus promotes its accumulation in the cytoplasm. This increased cytoplasmic pool presumably mediates an increase in mRNA translation by stimulating eIF4A resulting finally in enhanced cell proliferation. Although phospho-mimetic PKP1 associates with desmosomes and binds DP, exchange rates of the non-phosphorylated versus the phospho-mimetic forms in the desmosome differed considerably. Only the non-phosphorylatable mutant displayed a large immobile fraction in desmosomes in agreement with the enhancement of intercellular adhesion observed for this mutant. Since Akt2-dependent phosphorylation interfered with the PKP1–Dsg1 association we assume that DP association alone is not sufficient for the stable integration of PKP1 into desmosomes. Modification of PKP1 could alter its turnover at the desmosome by modulating internalization and endocytic trafficking. Endocytosis is an important mechanism to counteract junction assembly to maintain junctional homeostasis and it appears that dynamics of junctional complexes allows the rapid remodeling of these adhesive structures to adapt to various stimuli (Yap et al., 2007; Hong et al., 2010b). Accordingly, endocytosis of adherens junctions and E-cadherin controls processes such as cell movement, junction maintenance and polarity. Of note, mis-regulated E-cadherin endocytosis contributes to cancer (Sigismund et al., 2012). A known inhibitor of cadherin endocytosis is the associated armadillo protein p120-ctn, a plakophilin relative (Hatzfeld, 2005; Miyashita and Ozawa, 2007). Although desmosome turnover is less well characterized, caveolin- or raft-dependent endocytotic routes for desmosomal cadherins have been reported (Delva et al., 2008; Resnik et al., 2011; Brennan et al., 2012). In support, increased desmosomal internalization contributes to the autoimmune disease pemphigus (Jennings et al., 2011). Alternatively, the differential turnover of the PKP1 mutant proteins in the desmosome could reflect a modulation of the desmosome–cytoskeleton association.
Our data suggest that PKP1 may be one critical effector of IGF/Akt signaling in the skin. Interestingly, Akt1/Akt2 double knockout mice display a skin phenotype with hair loss and impaired proliferation of basal keratinocytes as well as strongly reduced body weight (Peng et al., 2003). A strikingly similar phenotype has been reported in patients carrying PKP1 mutations. These patients suffer from defects in hair follicle development (Bergman and Sprecher, 2005), growth retardation and a general failure to thrive (McGrath et al., 1999; Ersoy-Evans et al., 2006; McGrath and Mellerio, 2010), supporting our hypothesis that PKP1 mediates IGF/Akt induced effects in the skin. Moreover, a similar phenotype was reported in PKP3 knockout mice (Sklyarova et al., 2008). Although the PKP1 phosphorylation sites characterized here are not conserved in PKP3, both proteins appear to share a common function in RNA metabolism (Hofmann et al., 2006).
Reports on a role of desmosomes in cancer are controversial. While some experiments supported a tumor-suppressive role of desmosomes others suggested an oncogenic function (Chidgey and Dawson, 2007; Dusek and Attardi, 2011). Our data predict that this reflects context-dependent differences: unregulated activation of the growth factor/Akt2 pathway as described in many tumors (Cheng et al., 2008; Qiao et al., 2008; Steelman et al., 2008; Tokunaga et al., 2008; Memmott and Dennis, 2009) could induce the cytoplasmic accumulation of PKP1 accompanied by a destabilization of desmosomal adhesion on the one hand and an increase in translation and proliferation on the other hand. In such a context, PKP1 could actively contribute to carcinogenesis. Indeed, an overexpression of PKP1 and 3 has been described in several tumors including squamous cell carcinoma of the head and neck, lung carcinoma as well as Ewing sarcoma (Villaret et al., 2000; Furukawa et al., 2005; Cheung et al., 2007; Kundu et al., 2008; Valladares-Ayerbes et al., 2010). Interestingly, in actinic keratosis, a pre-cancerous lesion of the skin, as well as in oral squamous cell carcinoma (SCC) an elevated cytoplasmic pool of PKP1 has been reported (Kurzen et al., 2003; Narayana et al., 2010), supporting our notion that cytoplasmic PKP1 promotes proliferation. Although melanocytes typically do not express PKP1, some melanoma cell lines acquire PKP1 expression (Schmitt et al., 2007; Rickelt et al., 2008). A phospho-proteome screen of skin melanoma revealed PKP1 phosphorylation at several sites that we have identified here as Akt2 targets, namely S63, S65, S118, S121, S185 and S191 emphasizing the relevance of our findings in the context of tumor development (Zanivan et al., 2008). Interestingly, Akt2 or Akt3 appear upregulated in many melanoma biopsies (Stahl et al., 2004; Robertson, 2005; Shin et al., 2010). In a different context, when the PI3K/Akt pathway is not activated, the loss of PKP1 expression could contribute to reduced intercellular adhesion and thereby promote carcinogenesis. This correlates with reports showing reduced expression of PKPs in some tumor samples (Moll et al., 1986).
Thus, a simplified model of PKP1's functions suggests that in the absence of growth factor signaling un-phosphorylated PKP1 interacts with DP and associates with desmosomes leading to an increase in number and size of desmosomes thereby stabilizing desmosomal adhesion. Excess PKP1 in the cytoplasm is degraded in this situation. This scenario would correlate with a tumor suppressive function of PKP1. After growth factor stimulation however, PKP1 becomes phosphorylated and accumulates in the cytoplasm, where it stimulates translation and proliferation. It still associates with desmosomes, but is less stably integrated facilitating the remodeling of desmosomes. In such a context, PKP1 would acquire a growth-promoting function (Fig. 8). This scenario strongly resembles β-catenin regulation. β-catenin functions in E-cadherin-mediated cell–cell adhesion, where it supports the tumor suppressive function of E-cadherin. Upon stimulation of Wnt signaling, β-catenin accumulates in the cytoplasm and is translocated into the nucleus, where it stimulates together with LEF/TCF the transcription of several target genes involved in growth control, thereby acquiring an oncogenic function (Behrens, 2000; Polakis, 2000; Brembeck et al., 2006; Clevers, 2006). Mutations in either β-catenin or its regulatory proteins are common events in the development of colon and other cancers. Our data uncover a very similar mechanism in the regulation of PKP1: upon growth factor signaling phosphorylated PKP1 accumulates in the cytoplasm which shifts the balance from a predominantly adhesive function in cell contacts towards a growth-promoting function. A dual function in the control of adhesion and proliferation and its regulation by signaling factors is therefore not unique to β-catenin but might be a more widespread molecular mechanism of members of the armadillo protein family to fine-tune cell–cell adhesion and growth control.
Materials and Methods
Plasmids and primers
Human PKP1 and eIF4A1 constructs were described previously (Wolf et al., 2010). PKP1wt was subcloned into pDsRed-N1, pEGFP-C2 (BD Biosciences), pGEX-5x1 (Amersham Biosciences), pRSET and pcDNA3 (Invitrogen) using standard procedures. A FLAG tag was introduced into pcDNA3 for protein purification using M2–agarose. HA-Akt2 (16000), myr-HA-Akt2 (9016) and myr-HA-Akt1 (9008) plasmids were purchased from Addgene (Ramaswamy et al., 1999). HA-Akt2-K181M-T309A-S474A (Akt2dn), HA-Akt2-T309D-S474D (Akt2ca) and PKP1 mutants were generated using the QuikChange™ site-directed mutagenesis kit from Stratagene. For mutagenesis primer sequences, see the supplementary material Table S2. PKP1 mutants were inserted into pDsRed-N1, pEGFP-C2, pRSET and pGEX-5x1. PKP1 head (aa 1–286), PKP1 repeats (aa 287–726), PKP2 head (aa 1–394), PKP2 repeats (aa 395–837), PKP3 head (aa 1–342) and PKP3 repeats (aa 343–797) were cloned into pGBKT7 (BD Biosciences). Akt1 and Akt2 were inserted into pGADT7 (BD Biosciences) and pRSET. The human Dsg1 intracellular domain construct was described previously (Hatzfeld et al., 2000). eIF4A1, the N-terminal domain of desmoplakin (DP-NTP) and the cytoplasmic domain of Dsg1 (Dsg1-CD) were cloned into pRSET. Vectors used for BiFC analyses were as previously described (Wolf et al., 2010). Plasmids used in this study were: pVen1-FLAG-PKP1wt, pVen1-FLAG-PKP1M1-4A, pVen1-FLAG-PKP1M1-4E, pVen2-HA-Akt1, pVen2-HA-Akt2, pVen2-HA-Akt2ca (T309D-S474D), pVen2-HA-Akt2dn (K181M-T309A-S474A), pVen2-HA-eIF4A1, pVen2-HA-DP-NTP and pVen2-HA-Dsg1-CD. The EMCV-CAT plasmid was described previously (Ostareck et al., 2001).
Antibodies and reagents
The primary antibodies used for immunostaining, immunoblotting and immunoprecipitation were: PKP1 rabbit serum (667, (Hatzfeld et al., 2000)), PKP1 guinea pig sera against two ARM domain peptides (PQIARLLQSGNSDVVR, QGVLRQQGFDRNM), anti-GST (BD Biosciences), anti-His (Qiagen), anti-HA (Rockland), anti-FLAG M2, anti-BrdU BU-55, anti-α-tubulin DM1A, anti-c-myc 9E10, anti-vinculin hVIN-1, anti-β-Actin AC-74 (Sigma Aldrich), anti-Akt2, anti-phospho-(S473)-Akt, anti-phospho-(S422)-eIF4B, anti-phospho-PKA substrate antibody (Cell Signaling), anti-DP 1/2 DP-2.15 + DP-2.17 + DP-2.20 and anti-Cytokeratin 18 Ks 18.04/18.27/18.5/9B1/18.174 (Progen). Actin filaments were labeled with Alexa488-phalloidin or TRITC-phalloidin (Invitrogen Molecular Probes). Secondary antibodies were DyLight488 and DyLight594 anti-mouse, DyLight649 anti-rabbit (Invitrogen Molecular Probes), HRP-conjugated anti-mouse, anti-guinea pig and anti-rabbit (The Jackson Laboratory). DNA was stained with DAPI (Invitrogen Molecular Probes). Purified active Akt2 was from Upstate. GSK3-β and CIP (calf intestinal phosphatase) were obtained from Cell Signaling. LY294002 was purchased from Calbiochem, IGF-1 and EGF from R&D Systems. Insulin, crystal violet, cycloheximide and anti-FLAG M2 affinity gel were obtained from Sigma Aldrich and iodnitrotetrazolium chloride from AppliChem. Protein-A-conjugated agarose beads and glutathione agarose were supplied by Pierce Biotechnology.
Cell culture, transfection and treatments
MCF-7, HaCaT, HeLa, A431 and HEK293 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS). MCF-7 cells stably expressing DsRed-tagged PKP1wt, PKP1M1-4A and PKP1M1-4E were grown in medium containing 600 µg/ml G418 (PAA). Transfection of plasmid DNA was performed using a standard Ca-phosphate precipitation protocol, Truefect (United BioSystems) or Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocols. For knockdown analysis in HaCaT cells, siRNAs were transfected by Nucleofector II (Lonza). siRNAs used in this study were listed in supplementary Table S4. For growth factor stimulation cells were maintained in serum free medium for 24 hours and stimulated with 4 µg/ml insulin (Sigma Aldrich), 100 ng/ml IGF1 or 100 ng/ml EGF (R&D Systems), or insulin in the presence or absence of 20 µM LY294002 (Calbiochem) for 48 hours.
Cells were fixed in 3.7% formaldehyde in PBS for 15 minutes at room temperature or in methanol at −20°C for 10 minutes and permeabilized in 0.5% Triton X-100 in PBS for 15 minutes. Cells were mounted with Mowiol, and images were acquired using an Eclipse E600 microscope (Nikon) with a 60× objective (Apo TIRF; Nikon) connected to a CCD camera (CCD-1300QLN; VDS Vosskühler) running NIS Elements AR 2.30 software (Nikon) or a Zeiss Axio Imager microscope (Carl Zeiss) equipped with a camera (AxioCam MR3; Carl Zeiss), a Plan Apo 63× objective (NA 1.4), and the Axio Vision 4.7 software (Carl Zeiss). Photoshop CS5 (Adobe) was used for image processing (adjustment of brightness and contrast).
Fluorescence recovery after photobleaching
The dynamic localization of GFP-tagged PKP1wt, PKP1M1-4A and PKP1M1-4E in MCF-7 cells plated on glass-bottomed dishes was analyzed by FRAP 48 hours after transfection. Bleaching and imaging was performed on a confocal microscope (TCS SP5; Leica) using an oil-immersion 63× objective (NA 1.4) and an argon laser at 488 nm. Ten images were taken before the bleach pulse and 300 images after bleaching with an image acquisition frequency of 5 frames/second at 7% laser transmission. For each time point, the background intensity was subtracted and the data were normalized by dividing each data point by the average unbleached value. The average fluorescence recovery curve was fitted to a single exponential function, given by F(t) = A (1−e−tτ)+B, where F(t) is the intensity at time t, A the plateau level after recovery, B the level after bleaching and τ the slope of the exponential term (time constant). The recovery half-times were obtained by calculating the time when the cells reached half of the final intensity. The mobile (Mf) and immobile fractions (If) were determined by calculating Mf = A/(1−B) and If = 1−Mf.
Cell migration assays
For single cell migration analyses, cells were plated on collagen-coated culture plates for 4 hours before conducting time-lapse imaging using a Zeiss Axio Imager microscope (Carl Zeiss). Brightfield and fluorescence images were acquired every 15 minutes for a period of 14 hours using a 10× Ex Plan Neofluar objective (NA 0.3) and standardized microscope settings. The traveled path (total distance [µm]) of random migration was quantified using the manual tracking plugin (http://rsbweb.nih.gov/ij/plugins/track/track.html) for ImageJ (Wayne Rasband, NIH) based on the coordinates obtained from the translocation of the nuclei of the cells.
To analyze cell migration into a cell free gap IBIDI culture inserts (IBIDI) consisting of two reservoirs separated by a 500 µm thick wall were placed onto a culture plate. MCF-7 cells expressing DsRed, PKP1wt-, PKP1M1-4A- or PKP1M1-4E-DsRed (4×105 cells/ml) were seeded in dublicate into the two reservoirs and grown to confluency for 48 hours before the insert was removed creating a gap of ∼500 µm. Migration was observed by live cell imaging using a Zeiss Axio Imager microscope (Carl Zeiss). Brightfield images were aquired every 1 hour for a period of 24 hours using a 10× Ex Plan Neofluar objective (NA 0.3) and standardized microscope settings. The area of migrated cells into the gap was quantified using the MiToBo plugin for ImageJ (Glaß et al., 2012).
Soft agar assay for colony formation
MCF-7 cells stably expressing DsRed, PKP1wt-, PKP1M1-4A- or PKP1M1-4E-DsRed were suspended in a final top agarose concentration of 0.35% and added in sextuple in 48-well plates (1000 cells/well) on a base layer of 0.5% agar containing culture medium and incubated for 3 weeks. Colonies were stained using iodonitrotetrazolium chlorid (5 mg/ml, AppliChem) and counted.
Immunoprecipitation and dephosphorylation
Immunoprecipitation was performed as described previously (Keil et al., 2009). Phosphatase treatment was carried out by incubating protein lysates with 20 units of CIP for 30 minutes at 37°C.
Solubility and protein stability assays
Cells were suspended for 15 minutes in non-denaturing lysis buffer (20 mM Tris-HCl pH 7.6, 140 mM NaCl, 10% glycerol, 1.5% Triton X-100, 2 mM EDTA, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 1 mM PefaBloc, 1 mM NaF, 1 mM NaVO3). After centrifugation for 30 minutes at 13,000 g corresponding amounts of soluble and insoluble fractions were resolved by SDS-PAGE and immunoblotted. For protein stability analyses, cells were treated with 50 µg/ml cycloheximide (CHX).
Yeast two-hybrid and GST pull-down
Yeast two-hybrid analysis and GST pull-down experiments were performed as described previously (Wolf et al., 2010).
Bimolecular fluorescence complementation
MCF-7 and HeLa cells were co-transfected with the indicated pVen1 plus pVen2 constructs. Yellow fluorescence was detected at 24 hours post transfection. Transfected cells were identified by antibody staining against the FLAG and HA tags.
Luciferase reporter gene assay
Firefly luciferase FFL-pcDNA3 was co-transfected with EGFP-C2 (control), EGFP-PKP1wt or EGFP-PKP1 mutants or with EGFP-C2 + pcDNA3, EGFP-C2 + myr-Akt2, EGFP-PKP1wt + pcDNA3 or EGFP-PKP1wt + myr-Akt2 into HEK293 cells. In some cases cells were starved of serum for 24 hours (as indicated in the figure legends) and luciferase activities were measured (48 hours post transfection) by using a GloMaxTM 96 Luminometer (Promega). FFL-activities were normalized to FFL-mRNA levels. qRT-PCR and in vitro reporter assays were performed as described previously (Wolf et al., 2010). Primer sequences are given in supplementary material Table S3.
Transfected cells were serum starved for 24 hours (0.1% FBS), refreshed with 10% FBS for 16 hours and incubated with 10 µM BrdU (Roche) for 1 hour. After fixation in 3.7% formaldehyde in PBS for 20 minutes, the DNA was denatured with 4 M HCl for 10 minutes. Cells were washed in PBS containing 1% BSA and 0.1% Triton X-100 followed by BrdU antibody staining. More than 200 transfected cells were checked for BrdU incorporation in each independent assay.
Cell viability assay
MCF-7 cells stably expressing DsRed, PKP1wt-, PKP1M1-4A- or PKP1M1-4E-DsRed were seeded in triplicate into 98-well plates (5000 cells/well) and cultured for 24, 48, 72 and 96 hours. Cells were fixed in 3.7% formaldehyde/PBS and stained for 1 hour with 0.05% crystal violet in PBS. After washing the plates were left to dry at 37°C and the dye was washed out by shaking the cells in 1% SDS in PBS for 30 minutes. Absorption was measured at 550 nm. To avoid bias by variability in cell seeding, all data were normalized to values determined at 24 hours.
For analyses of cell-cell adhesive function, MCF-7 clones were seeded in triplicate onto 12-well plates and treated with control medium, insulin or insulin + LY294002 for 48 hours. The confluent cell cultures were washed twice with PBS and incubated with 1 ml of dispase (2.5 U/ml; Roche) for 30 minutes at 37°C/5% CO2. Free-floating monolayers were shaken on an orbital shaker (200 rpm) for 1 hour before imaging. The final images were acquired using a EOS 40D with a EFS 60 mm macro lens (both Canon).
In vitro kinase assay
Purification of His-tagged PKP1 was performed as described previously (Wolf et al., 2010). Recombinant purified PKP1 was incubated with recombinant Akt2 (250 ng) or GSK3-β (250 ng) and 2 µCi of [γ-32P]ATP (Perkin Elmer) in 25 µl of kinase buffer (5 mM MOPS pH 7.2, 2.5 mM β-glycerophosphate, 1 mM EGTA, 0.4 mM EDTA, 10 mM MgCl2, 0.05 mM DTT) for 30 minutes at 30°C. Reactions were terminated by addition of SDS-PAGE loading buffer. Proteins were resolved by 10% SDS-PAGE, transferred to nitrocellulose membranes, and 32P incorporation was detected by autoradiography.
Peptide array phosphorylation assay
Peptides (15 aa in length with five overlapping aa) were covalently immobilized to the surface of a glass slide and printed in triplicate (JPT Peptide Technologies). Full-length control proteins with suitable phosphorylation sites were included. 300 µl of kinase reaction solution containing 10 µg Akt2, 5 mM MOPS pH 7.2, 2.5 mM β-glycerophosphate, 1 mM EGTA, 0.4 mM EDTA, 10 mM MgCl2, 0.05 mM DTT 1% BSA and 150 µM ATP was added to the array and incubated for 4 hours at 30°C. Arrays were stained with Pro-Q Diamond (Invitrogen), and phosphorylation was quantified by fluorescence imaging (excitation/emission at 550/580 nm) using a GenePix 4000B Microarray Scanner (Molecular Devices).
Phosphorylated proteins were digested with trypsin and proteolytic peptide mixtures were separated on an Ultimate nano-HPLC system (Dionex Corporation) using reversed phase C18 columns (precolumn: Acclaim PepMap, 100 µm ×20 mm, 5 µm, 100 Å, separation column: Acclaim PepMap, 75 µm×250 mm, 3 µm, 100 Å, Dionex Corporation). After washing the peptides on the precolumn for 15 minutes with 0.1% TFA at a flow rate of 20 µl/minute, peptides were eluted and separated using 90-minute gradients from 0 to 40% solvent B (solvent A: 5% ACN containing 0.1% FA, solvent B: 80% ACN containing 0.08% FA) at a flow rate of 300 nl/minute. The nano-HPLC system was directly coupled to the nano-ESI source (Proxeon) of an LTQ-Orbitrap XL hybrid mass spectrometer (Thermo Fisher Scientific). MS data were acquired in data-dependent MS/MS mode: Each high-resolution full scan (m/z 300 to 2000, R = 60,000) in the orbitrap was followed by five product ion scans in the LTQ on the five most intense signals in the full-scan mass spectrum (isolation window 1.5 u). Additional neutral loss triggered MS3 experiments (neutral losses 98, 49 or 32.7 Thomson) were conducted to confirm the phosphorylation sites. Dynamic exclusion (exclusion duration 180 seconds, exclusion window −1 to 2 Th) was enabled to allow detection of less abundant ions. Data acquisition was controlled via XCalibur 2.0.7 (Thermo Fisher Scientific) in combination with DCMS link 2.0 (Dionex). Identification of phosphorylation sites was done by Mascot search against the PKP1 sequence using the Proteome Discoverer 1.2 (Thermo Fisher Scientific).
For all measurements, the error bars represent the standard deviation (s.d.). Student's two tailed t-test was used to determine the respective statistical significance. Probability (P) values are given in the figure legends where appropriate.
We thank K. Green for providing the DP-NTP construct, N. Stöhr and the Core Facility Imaging at the Faculty of Medicine of the Martin-Luther-University Halle-Wittenberg for assistance with time-lapse microscopy and T. Magin and F. Berthelmann for critically reading the manuscript.
A. Wolf performed and evaluated most of the experiments, K.R. carried out localization studies of PKP1 mutants and dispase assays, M.S. and A. Wolf carried out the kinase arrays, A.S. and C.I. performed mass spectrometry analyses, A. Wingenfeld performed BiFC studies, Akt transfection experiments, the solubility assay and cloning work, A.M. did recombinant protein expression, purification and GST pull-down experiments and in vitro mutagenesis. M.G. and S.H. contributed to live-cell imaging and FRAP analyses and S.H. contributed ideas and discussion. M.H. performed the Y2H screen, designed and analyzed most experiments, supervised the study and wrote the manuscript.
This work was supported by the Deutsche Forschungsgemeinschaft (DFG) [grant number Ha1791/8-1 to M.H.]; and the Federal Ministry of Education and Research (BMBF) (ProNET T3) to M.H.
Supplementary material available online at http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.118992/-/DC1
- Accepted January 20, 2013.
- © 2013. Published by The Company of Biologists Ltd