Autosomal dominant polycystic kidney disease (ADPKD) is characterized by the formation of renal, hepatic and pancreatic cysts and by non-cystic manifestations such as abnormal vasculature and embryo left–right asymmetry development. Polycystin-2 (PC2), in which mutations account for 10–15% of ADPKD, was previously shown to down-regulate cell proliferation, but the underlying mechanism was not elucidated. Here, we demonstrate that PC2, but not pathogenic mutants E837X and R872X, represses cell proliferation through promoting the phosphorylation of eukaryotic translation initiation factor eIF2α by pancreatic ER-resident eIF2α kinase (PERK). ER stress is known to enhance eIF2α phosphorylation through up-regulating PERK kinase activity (assessed by phosphorylated PERK). During ER stress, PC2 knockdown also repressed eIF2α phosphorylation but did not alter PERK phosphorylation, indicating that PC2 facilitates the eIF2α phosphorylation by PERK. PC2 was found to be in the same complex as PERK and eIF2α. Together, we demonstrate that PC2 negatively controls cell growth by promoting PERK-mediated eIF2α phosphorylation, presumably through physical interaction, which may underlie a pathogenesis mechanism of ADPKD and indicates that PC2 is an important regulator of the translation machinery.
Autosomal dominant polycystic kidney disease (ADPKD) is the most common form of PKD and occurs in 0.1–0.2% of the adults (1–3). Renal pathogenic polycysts in ADPKD are fluid-filled epithelial-lined cavities arising from glomeruli, nephron tubules and collecting ducts (4,5). ADPKD can also result in cerebral and intracranial aneurysms, cardiovascular abnormalities (e.g. cardiac valve disorders), aberrant embryonic left–right asymmetry development and hypertension (6,7). The majority (∼95%) of the ADPKD cases are caused by mutations in the PKD1 or PKD2 gene that encodes polycystin-1 (PC1) or polycystin-2 (PC2) (8,9). PC1 is an integral membrane glycoprotein and acts as a G-protein-coupled receptor. PC2 is a non-selective cation channel and is mainly localized on the ER membrane as a Ca release channel (10,11).
The mechanisms of how ADPKD results in over-proliferation of renal epithelial cells are not clearly understood, and different pathways have been reported to be implicated (12–15). PC1 regulates cell growth of renal tubular epithelial cells through p53 induction and JNK activation (15). PC2 was also reported to suppress cell growth and branching morphogenesis in kidney epithelial cells, but the underlying mechanism remains unknown (13). Bhunia et al. (12) showed that PC1 and PC2 together regulate cell cycle through inducing p21 and activating the JAK-STAT signaling pathway. On the other hand, Li et al. (14) demonstrated that PC2 undergoes PC1-dependent phosphorylation, which enhances its interaction with the helix–loop–helix inhibitor ld2, to regulate cell growth and differentiation.
ER is critical to the synthesis, modification, folding and quality control of both secretory and membrane protein (16). Conditions disrupting the ER homeostasis can cause unfolded protein accumulation that constitutes a fundamental threat to all living cells and triggers unfolded protein response (UPR), which is currently termed ER stress (17). During the ER stress response or UPR, ER-resident molecular chaperones and foldases are induced to enhance the folding capacity of the ER, and translation is attenuated to reduce the biosynthetic load of ER (17,18). Current studies of the UPR mechanism in mammalian cells have identified three branches of the signaling pathway, represented by three types of ER-transmembrane proteins: pancreatic ER eIF2α kinase (PERK) leading to phosphorylation of eIF2α that causes translational repression, activating transcription factor 6 (ATF6) which induces the expression of chaperones such as BiP, and inositol requiring 1 (IRE1) which initiates spliceosome-independent splicing of XBP1 mRNA, leading to the activation of ER chaperone genes.
Type I ER membrane protein PERK was recently reported to be involved in controlling cell growth (19,20). PERK contains a luminal ER stress-sensing N-terminal domain and a large C-terminal cytoplasmic kinase domain (18,21). During ER stress, PERK is activated by autophosphorylation via homodimerization. The phosphorylated (or activated) PERK (P-PERK) subsequently binds and phosphorylates its subunit eIF2α (21,22), which inhibits global protein synthesis and cell-cycle factors such as cyclin D1, and results in the repression of cell proliferation (19,20,23). We demonstrate here that PC2 negatively regulates cell proliferation through enhancing PERK-dependent eIF2α phosphorylation. Our study suggests that PC2-associated ADPKD pathogenesis is due at least in part to dysregulation of PERK-dependent cell growth.
PC2 facilitates the eIF2α phosphorylation by PERK
We recently reported that Herp, a regulator of ER-associated degradation (ERAD) (24–26, 28), interacts with PC2 and regulates its expression (27). Here, we started to test whether PC2 is involved in the regulation of UPR activated during ER stress (17,18). Immunoblotting experiments revealed that PC2 over-expression increases the phosphorylated eIF2α (P-eIF2α) level by 227 ± 7% (P < 0.001, N = 3) in human embryonic kidney (HEK) 293T cells (Fig. 1A), with quantitative analysis performed using the program Image J (v1.41c, NIH, http://rsb.info.nih.gov/ij/). Similar results were also shown in human melanoma A7 and Madin–Darby canine kidney (MDCK) cells (Fig. 1B and Supplementary Material, Fig. S1). In the presence of ER stress induced by thapsigargin (Tg) through ER Ca depletion, P-eIF2α substantially augmented, as expected. The combination of PC2 over-expression and Tg treatment, compared with Tg treatment alone, did not significantly further increase the P-eIF2α level (103% ± 5%, P = 0.29, N = 3), possibly because of a saturated P-eIF2α level induced by Tg. Of important note, PC2 over-expression had no effect on immunoglobulin heavy-chain binding protein (BiP, also called GRP78, an ER chaperone protein) (Fig. 1A and B) and splicing of XBP-1 mRNA (Supplementary Material, Fig. S2), both of which are activated during UPR (17,21), indicating that high levels of PC2 do not cause ER stress leading to UPR. This suggests that the effect of PC2 over-expression is not through triggering ER stress. It is known that increased P-eIF2α attenuates protein synthesis and selectively up-regulates the translation of the activating transcription factor 4 (ATF4) (29). We next tested the effects of PC2 on these downstream signals of P-eIF2α. Compared with control, GFP-expressing cells, GFP-PC2-expressing cells exhibited profound attenuation of global translation rates, as demonstrated using [35S]methionine/cystine incorporation (Fig. 1C). Consistently, the protein synthesis rate negatively correlated with the level of P-eIF2α (Fig. 1C). In addition, Western blotting showed that ATF4 expression is increased in cells expressing GFP-PC2, compared with control cells (Fig. 1D). Thus, these effects of PC2 are consistent with the reported downstream signaling of eIF2α phosphorylation. Also, we found that over-expressed human anion exchanger AE1 (Gift of Dr Joe Casey, University of Alberta) in HEK293T cells has no effect on the P-eIF2 level (Fig. 1E), further supporting the specificity of the PC2-mediated up-regulation of P-eIF2.
We further tested the effect of PC2 on eIF2α phosphorylation by small interference RNA (siRNA) knockdown (27,30). PC2 siRNA caused robust reduction in the PC2 and P-eIF2α levels in HEK293T cells and mouse inner medullary collecting duct (IMCD) cells (Fig. 2A and B). Thus, PC2 is a regulator of eIF2α phosphorylation. eIF2α is known to be phosphorylated by four Ser/Thr protein kinases (activated by different stress inducers): GCN2 (activated by nutrition limitation and UV irradiation), PERK (by ER stress), PKR (by double-stranded RNA and viral infection) and HRI (oxidative stress) (31). To identify the kinase(s) that is responsible for the PC2-regulated eIF2α phosphorylation, we examined effects of different stress inducers on eIF2α phosphorylation in HEK293T cells with or without PC2 knockdown. Our data showed that the effect of Tg (Fig. 2A), but not that of oxidative inducer arsenite (Fig. 2A), amino acid (aa) depletion or osmotic shock (data not shown), on eIF2α phosphorylation substantially diminished with the PC2 knockdown. A similar effect of PC2 knockdown was observed in IMCD cells (Fig. 2B). These results indicate that PC2 is crucial for ER stress-induced eIF2α phosphorylation. As PERK is known to be the major kinase phosphorylating eIF2α during ER stress (18,21), we tested whether PERK is involved. Indeed, PERK knockdown significantly repressed PC2-induced eIF2α phosphorylation in HEK293T cells (Fig. 2C). Together, our data showed that PC2 promotes eIF2α phosphorylation by the ER-resident eIF2α kinase PERK. In contrast, we found that PERK phosphorylation is not affected by either over-expression or knockdown of PC2 (Fig. 1A and B, and Fig. 2A), which indicates that PC2 is a crucial factor that facilitates eIF2α phosphorylation by phosphorylated (i.e. activated) PERK.
Positive correlation between the PC2 and P-eIF2α levels revealed by immunofluorescence
We performed co-immunofluorescence experiments to further document the regulation of P-eIF2α by PC2. To enlarge the range of PC2 expression levels in the same images taken by the microscope, 24 h prior to experiments, we mixed HEK293T cells transiently expressing PC2 with those with PC2 siRNA knockdown at the 20:80 ratio. We found that the PC2 expression level significantly correlates with that of P-eIF2α, with a correlation coefficient of 0.82 based on 290 randomly selected cells (Fig. 3A and B) and that PC2 and P-eIF2α partially colocalizes in perinuclear regions (Fig. 3A). Mixing native HEK293T cells with those with PC2 knockdown by siRNA at the 20:80 ratio produced similar results (Fig. 3C), with a correlation coefficient of 0.59 based on 107 randomly selected cells. These results support the concept that PC2 promotes the phosphorylation of P-eIF2α by the ER membrane kinase PERK, leading to the accumulation of P-eIF2α in perinuclear regions.
PERK is required for PC2-regulated cell proliferation
It was previously shown that PERK represses cell growth by inducing eIF2α phosphorylation to attenuate protein synthesis and by down-regulating cyclin D1 required for cell-cycle exit from the G1 phase (19,20,23). We thus reasoned that the reported down-regulation of cell growth by PC2 may be through the PERK-eIF2α signaling pathway. We measured cell growth rate using the CellTiter-Glo Luminescent Cell Viability Assay (Promega) as a function of alterations in the PC2 and PERK levels. We found that an increased (or decreased) PC2 expression significantly down-regulates (or up-regulates) proliferation of MDCK and/or HEK293T cells (Fig. 4), consistent with the reported anti-proliferative effect of PC2 (12–14). The PC2-mediated suppression of cell growth was significantly inhibited in HEK293T cells by PERK knockdown (Fig. 4D), which is in parallel to the inhibition of the PC2-promoted eIF2α phosphorylation by PERK knockdown (Fig. 2C). In contrast, PERK knockdown slightly, but not significantly, increased the rate of cell growth in our assays. Overall, these results indicate that PERK and eIF2α play key roles in the negative control of cell proliferation by PC2.
PC2 physically interacts with PERK and eIF2α
We next explored the possible association of PC2 with PERK and eIF2α. For this end, we immunoprecipitated extracts of IMCD cells using an anti-PC2 antibody (27,30). We found that PERK and eIF2α, but not the ER membrane protein ATF6α (negative control) (32), co-precipitate with PC2 (Fig. 5A). Reciprocally, PC2 was immunoprecipitated by PERK in HEK293T cells expressing Myc-tagged PERK (Fig. 5B). Also, PC2 was associated with PERK and eIF2α in HEK293T cells over-expressing GFP-PC2 (Fig. 5C). These data demonstrate that PC2 is in the same complex as PERK and eIF2α. Because the eIF2α phosphorylation by the activated PERK is facilitated by PC2, we wanted to examine whether PC2 also interacts with the activated PERK. For this, ER stress was induced by Tg in HEK293T cells to stimulate the P-PERK level. Our data showed that PC2 indeed interacts with P-PERK (Fig. 5D). Finally, we examined the role of PERK in the PC2-eIF2α interaction, utilizing PERK+/+ and PERK−/− murine embryonic fibroblast (MEF) cells. We observed that the amount of PC2-bound eIF2α is sharply decreased in PERK−/− MEF cells (Fig. 5E), suggesting that PERK mediates the PC2-eIF2α interaction, possibly by forming a PC2-PERK-eIF2α complex. Together, our data indicate that PC2 promotes the eIF2α phosphorylation by P-PERK, presumably through physically facilitating the action of P-PERK on eIF2α. Interestingly, we found that PC2 expression is significantly reduced in PERK-deficient MEF cells (Fig. 5E, input). How PERK reversely regulates PC2 is unclear and may be a subject of future studies.
Pathogenic mutations in PC2 abrogate the regulation of eIF2α phosphorylation and cell proliferation by PC2
We investigated how PC2 mutations affect its effect on cell proliferation and eIF2α phosphorylation, using HEK293T cells transiently expressing one of the following GFP-tagged PC2 mutants: PC2 (positive control), R872X, E837X, PC2ΔC (or S689X, aa 1–688, lacking the C-terminus), PC2ΔN (aa 209–968, lacking the N-terminus), PC2ΔNC (aa 209–688) and GFP (negative control) (27). We found that, unlike Tg treatment and WT PC2 (positive controls), these truncation mutants were unable to stimulate eIF2α phosphorylation (Fig. 6A). Using the cell viability assay described earlier, we found that pathogenic mutants R872X and E837X has no effect on the rate of cell growth (Fig. 6B). These data indicate that PC2 exerts its effect on eIF2α phosphorylation through physical association with PERK. Thus, the inability of PC2 pathogenic mutants in negatively controlling cell growth through stimulating eIF2α phosphorylation may lead to cyst formation. Further, PC2ΔC, PC2ΔN, PC2ΔNC and E837X lost the association with P-PERK (Fig. 6C), suggesting that the PC2-PERK physical interaction is critical for PC2-mediated cell proliferation and eIF2α phosphorylation.
ER stress increases the kinase activity of PERK to promote eIF2α phosphorylation, which results in translational repression and activation of downstream signals that down-regulate cell growth (18–23). Renal cystogenesis, caused by loss-of-function mutations in PC1 or PC2, is characterized by over-proliferation, de-differentiation and so on, but how PC2 negatively controls cell growth remains unclear. In this study, we demonstrate that PC2 regulates cell proliferation through the PERK-eIF2α phosphorylation signaling pathway.
Our data show that induction of eIF2α phosphorylation and inhibition of cell growth by PC2 are blocked in cells with the treatment of PERK siRNA, indicating that PERK plays a critical role in this event. Interestingly, among the four eIF2α kinases, only PERK is localized to the ER membrane where the ER Ca release channel PC2 is localized. PC2, PERK and eIF2α are in the same complex, suggesting that PC2 may play a role in recruiting eIF2α to form PERK–substrate complexes for eIF2α phosphorylation. This was confirmed by the observation that repression of PC2 expression dramatically inhibits ER stress-induced eIF2α phosphorylation, which is promoted mainly by PERK (18,21). Furthermore, PC2 pathogenic mutations lose the capability of inhibiting cell proliferation and of stimulating eIF2α phosphorylation. On the basis of these data, we propose that the PERK-eIF2α pathway be part of a novel molecular mechanism underlying the PC2-associated ADPKD.
Controlling cell over-proliferation in PKD is a key principle of therapeutic treatments such as the use of antagonists of vasopressin receptors and mammalian target of rapamycing (mTOR). mTOR antagonist, rapamycin, effectively alleviates cyst development in several animal models of PKD and in human ADPKD patients (33). In fact, PC1 interacts with mTOR to suppress the up-regulation of cell proliferation by mTOR, in which translation initiation factor eIF4E is involved (33). The current study shows that protein synthesis inhibitor P-eIF2α mediates negative control of cell growth by PC2. Therefore, the negative control of cell growth by PC1 and PC2, either individually or as a complex (12,14) may be through the same thread, i.e. protein synthesis initiation machinery. However, it cannot be excluded that PERK, by interacting with PC2, affects PC1-regulated cell proliferation, in which PC2 phosphorylation is required (12,14). In summary, this study shows that PC2 is an important regulator of the cellular translation machinery and may underlie a pathogenesis mechanism of ADPKD.
MATERIALS AND METHODS
Cell culture, DNA constructs and gene transfection
IMCD, MDCK and HEK293T cells were cultured in Dulbecco's modified Eagle's medium (high glucose; Invitrogen) containing 10% (v/v) fetal bovine serum, 1% penicillin and streptomycin at 37°C and 5% CO2. PERK+/+ and PERK−/− MEF cells were maintained as described previously (34). Plasmids pEGFP-PC2, pEGFP-R872X, pEGFP-E837X, pEGFP-PC2ΔC, pEGFP-PC2ΔN and pEGFP-PC2ΔNC were described previously (27,35). HEK293T cells were grown to ∼70% confluency prior to transfection using Lipofectamine 2000 (Invitrogen). MDCK cells stably expressing GFP-PC2 or GFP were selected as previously described (36) and maintained using G418 (300 µg/ml).
Immunoprecipitation and immunoblotting
Protein extraction, immunoblotting and immunoprecipitation (IP) were performed, as described earlier (27). Typically, 20 and 200 µg of total cellular protein were used for immunoblotting and IP, respectively. HEK293T cells were transiently transfected with pEGFP, pEGFP-PC2, pcDNA3.1 (Invitrogen) or pcDNA3.1-PERK (19,21) for IP. At 40 h post-transfection, cells were used for protein extraction and precipitation. To examine the effect of PC2 mutants on eIF2α phosphorylation, we transfected HEK293T cells with pEGFP-PC2, pEGFP-R872X, pEGFP-E837X, pEGFP-PC2ΔC, pEGFP-PC2ΔN, pEGFP-PC2ΔNC or vector pEGFP. At 40 h post-transfection, cell extraction was prepared for immunoblotting.
Immunofluorescence microscopy and quantitative analysis
Immunofluorescence microscopy was performed as described earlier (27). Briefly, HEK293T cells were transiently transfected with Myc-PC2 or siRNA, mixed at the 20:80 ratio after 24 h of transfection, and then grown on cover slips. At 48 h post-transfection, cells were subject to Tg treatment for 1 h before fixation. P-eIF2α and PC2 antibodies were used for double staining. Fluorescence images were captured on a motorized Olympus IX81 microscope with a CCD cooling RT SE6 monochrome camera (Diagnostic Instruments). Final composite images were made using Image-Pro Plus 5.0 (Media Cybernetics). The quantification of expression density was performed using Image J. The area of each cell was manually defined for the analysis of both green and red signal densities. A cluster of confluent cells was defined as one cell for density calculation. For each image that contains multiple cells, the average density for red or green color was normalized to 100 after background subtraction. The correlation coefficient was obtained from linear regression using Sigmaplot 10 (Systat Software Inc.).
Rabbit antibodies against eIF2α and P-eIF2α (S51), and mouse anti-Myc antibody were purchased from Cell Signaling Technology. Human anti-PERK and anti-P-PERK antibodies were from R&D Systems and Biolegend, respectively. β-actin antibody was from Sigma-Aldrich, Canada. Goat anti-BiP, rabbit anti-ATF6α/ATF4 and mouse anti-PERK were purchased from Santa Cruz. GFP antibodies were a gift of Dr Luc Berthiaume (University of Alberta; also available at www.eusera.com). Mouse anti-PC2 antibody was described as before (27,30). Secondary antibodies were from Amersham or Promega.
PERK and PC2 knockdown by siRNA
PERK Stealth siRNA (Invitrogen, EIF2AK3-HSS114058, -059 and -060) and control siRNA (Invitrogen, Cat# 46-2002) were used to transfect HEK293T cells using Lipofectamine 2000 reagent following the manufacturer's instructions. PC2 knockdown was described previously (30). The efficiency of the siRNA knockdown was assessed by immunoblotting.
Cell proliferation assays
HEK293T cells were transfected with plasmid pEGFP-PC2 or PC2 siRNA in 35 mm dishes (Fig. 4B and C). At 24 h post-transfection, cells were split and seeded into eight separate wells of a 96-well plate. After incubation for another 30 h, luminescence activity was measured using a cell viability assay kit (Promega, Cat# G7571) and an illuminometer (Fluoroskan Ascent FL, Thermo Labsystems). The remaining cells in the 35 mm dishes were further cultured for immunoblotting. For testing the effect of PERK knockdown on PC2-regulated cell proliferation (Fig. 4D), HEK293T cells were transfected with pEGFP-PC2 or vector pEGFP using 35 mm dishes. At 6 h post-transfection, cells were split into two equal fractions for subsequent transfection with PERK or control siRNA at 24 h post-transfection. Forty-eight hours after pEGFP-PC2 transfection, cells were cultured in a 96-well plate for the cell growth assay described above. MDCK stable cell lines were serum-starved overnight and then grown in the presence of serum for 48 h in a 96-well plate before measuring luminescence activity. The efficiency of siRNA knockdown and PC2 expression were assessed by immunoblotting and/or immunofluorescence microscopy.
To measure protein translational rates, HEK293T cells were transfected with plasmid pEGFP-PC2 or vector pEGFP. At 40 h post-transfection, cells were starved for 1 h in the pre-labeling medium (l-methionine and l-cystin-depleted DMEM with 10% fetal bovine serum and penicillin/streptomycin) (Invitrogen), followed by pulse-labeling with 50 µCi of [35S]methionine/cystine (35S-Protein Labeling Mix, Perkin Elmer) for 10 min. Cell extracts were applied for sodium dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis (PAGE) and autoradiography.
Student's t-test was used for statistical significance analyses. Data were expressed in the form of mean±SE (N), where SE is the standard error and N is the number of experiment repeats. Probability values (P) of less than 0.05 and 0.01 were considered significant and very significant, respectively.
This work was supported by the Canadian Institutes of Health Research and the Kidney Foundation of Canada (to X.-Z.C.). X.-Z.C. is a Senior Scholar of the Alberta Heritage Foundation for Medical Research.
Myc-tagged PERK plasmid, PERK+/+ and PERK−/− MEF cells were kindly provided by Dr David Ron (New York University).
Conflict of Interest statement. None declared.