Oncogene (2010) 29, 1486–1497 & 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 $32.00 www.nature.com/onc ORIGINAL ARTICLE Depletion of WRN causes RACK1 to activate several protein C isoforms

L Massip1, C Garand1, A Labbe´1,E` Perreault1, RVN Turaga1, VA Bohr2 and M Lebel1

1Centre de Recherche en Cance´rologie de l’Universite´ Laval, Department of Molecular Biology, Medical Biochemistry and Pathology, Hoˆpital Hoˆtel-Dieu de Que´bec, Que´bec City, Que´bec, Canada and 2Laboratory of Molecular Gerontology, National Institute on Aging, National Institutes of Health, 5600 Nathan Shock Drive, Baltimore, MD, USA

Werner’s syndrome (WS) is a rare autosomal disease of life onwards, patients with WS develop pathologies characterized by the premature onset of several age- that prematurely resemble many traits of normal aging, associated pathologies. The protein defective in patients such as cardiovascular diseases, osteoporosis or type II with WS (WRN) is a helicase/exonuclease involved in diabetes mellitus (reviewed in Epstein et al., 1966). DNA repair, replication, transcription and Death generally occurs in the fourth decade of life from maintenance. In this study, we show that a knock down heart demise or . Fibroblasts isolated from WS of the WRN protein in normal human fibroblasts induces patients characteristically senesce prematurely in culture phosphorylation and activation of several C (Faragher et al., 1993) and display increased chromo- (PKC) . Using a tandem affinity purification somal aberrations (Melcher et al., 2000). The protein strategy, we found that WRN physically and functionally defective in WS (WRN) is a RecQ family 30–50 DNA interacts with receptor for activated C-kinase 1 (RACK1), helicase that also possesses a 30–50 exonuclease activity a highly conserved anchoring protein involved in various and is involved in DNA recombination, transcription, biological processes, such as cell growth and proliferation. repair and telomere maintenance (reviewed in Cheng RACK1 binds strongly to the RQC domain of WRN and et al., 2007). A depletion of the WRN protein from cells weakly to its acidic repeat region. Purified RACK1 has no renders them more sensitive to oxidative damage impact on the helicase activity of WRN, but selectively (Szekely et al., 2005). This is an important finding as inhibits WRN exonuclease activity in vitro. Interestingly, the plasma from several WS patients and tissues from knocking down RACK1 increased the cellular frequency mice lacking a functional Wrn helicase domain exhibit a of DNA breaks. Depletion of the WRN protein in return pro-oxidant state (Pagano et al., 2005; Massip et al., caused a fraction of nuclear RACK1 to translocate out of 2006). Oxidative damage that can lead to double- the nucleus to bind and activate PKCd and PKCbII in the stranded DNA lesions throughout the genome, includ- membrane fraction of cells. In contrast, different DNA- ing , has been proposed to have a causative damaging treatments known to activate PKCs did not role in organismal aging (Beckman and Ames, 1998; induce RACK1/PKCs association in cells. Overall, our Karanjawala et al., 2002; Parrinello et al., 2003). results indicate that a depletion of the WRN protein in Global expression analyses of mouse embryonic cells normal fibroblasts causes the activation of several PKCs lacking a functional Wrn helicase domain has shown through translocation and association of RACK1 with major changes in the expression of several and such kinases. phosphatases, at least at the mRNA level (Descheˆnes Oncogene (2010) 29, 1486–1497; doi:10.1038/onc.2009.443; et al., 2005). In agreement with this observation, western published online 7 December 2009 blot analyses indicated increased levels of intracellular protein phosphorylation in such mutant cells. Further- Keywords: Werner syndrome; mass spectrometry; more, protein extracts from the liver and cardiac tissues RACK1; PKC activation from Wrn mutant mice indicated an increase in serine and tyrosine phosphorylation of several unknown compared with wild-type animals. Several kinases exhibiting altered expression in this mouse Introduction model are known to be influenced by oxidative stress or are known to modify levels of cellular redox status Werner’s syndrome (WS) is an autosomal recessive (Descheˆnes et al., 2005). Human WS cells are also disorder that displays many clinical symptoms of known to show increased stress signaling through the normal aging at an early age. From their second decade p38 mitogen-activated protein kinase (Davis et al., 2007; Davis and Kipling, 2008). Correspondence: Dr M Lebel, Centre de Recherche en Cance´rologie, Several types of kinases can be activated upon cellular Hoˆpital Hoˆtel-Dieu de Que´bec, 9 McMahon St, Que´bec City, Que´bec, stress. These include ATM, protein kinase C (PKC), and Canada G1R 2J6. E-mail: [email protected] p38 mitogen-activated protein kinase (reviewed in Received 22 April 2009; revised 19 October 2009; accepted 3 November Kaneto et al., 2005; Nitti et al., 2008; Sedding, 2008). 2009; published online 7 December 2009 The exact mechanism by which these kinases are Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1487 activated upon cellular stress is still elusive, but may of DNA replication. Transfection efficiency determined require adaptor proteins. One such protein is the with an Alexa-488-labeled control siRNA was more receptor for activated C-kinase 1 (RACK1). RACK1 than 95% (Turaga et al., 2007). Forty-eight hours after is a conserved factor involved in a broad spectrum of transfection, proteins were extracted for antibody mechanisms, such as cell proliferation, motility, apop- microarray analyses. Two independent transfection tosis, or protein translation/degra- experiments (control siRNA and siWRN) were con- dation. It functions as a docking platform for various ducted in duplicate to extract proteins. To assess the molecules including protein kinases and phosphatases, efficiency of the siWRN-mediated knockdown, we thereby coordinating key signalization pathways examined WRN protein levels in transfected cells. As (McCahill et al., 2002). Interestingly, RACK1 function indicated in Figure 1a, the most efficient knock down of is intimately connected to its subcellular localization and WRN was achieved with HSS111385 siRNA (Invitrogen is related to typical age-related pathological processes, Inc.) and corresponded to B80% decrease compared such as cancer, cardiac dysfunction or neurodegenera- with control siRNA transfection. Fluorescence-acti- tion (McCahill et al., 2002). RACK1 expression vated cell sorting analyses were performed to determine decreases with age in the rat brain and human leucocytes the percentage of cycling cells 48 h after transfection (Pascale et al., 1996; Corsini et al., 2005), whereas it with control siRNA and HSS111385 siWRN (at the increases in different kinds of carcinomas as compared time of protein extraction). As indicated in Figure 1b, with normal tissues (Berns et al., 2000; Egidy et al., siWRN-transfected cells showed an increase in the 2008; Wang et al., 2008). number of cells in the G1 phase, from 71 to 77%, and In this study, we used mass spectrometry analysis to a decrease in the number of cells in the S phase, from 28 identify proteins that copurify with a WRN protein to 18%, compared with siRNA control-transfected containing a TAP tag (for tandem affinity purification). fibroblasts. Examples of fluorescence-activated cell We identified RACK1 as one of the proteins interacting sorting analyses are provided in Supplementary Figure S1. with this construct. RACK1 has no impact on the Proteins from transfected cells were subjected to a helicase activity of the WRN protein, but inhibits its Kinex antibody microarray screen at the Kinexus exonuclease activity. Interestingly, depletion of the WRN protein causes relocation of a sub-population of RACK1 protein in the cytoplasm where it activates several PKC isoforms.

Results

Short-term depletion of the WRN protein activates several PKC isoforms in normal human diploid fibroblasts We first determined which kinases were activated in WRN-depleted cells. Cells derived from WS patients exhibit aneuploidy and chromosomal rearrangements (Melcher et al., 2000). Such alterations might affect the expression of kinases in a manner only indirectly related to the principal WS or age-related defects. In addition, WRN mutant cells may have undergone an adaptation process in vitro. To avoid these problems, we used short-term small interfering RNA (siRNA)-based inhibition of WRN to test the direct consequences of Figure 1 Impact of WRN protein depletion on GM08402 diploid WRN protein loss on kinase activities in normal fibroblasts cell cycle and phospho-proteins after transfection with fibroblasts. Normal human diploid fibroblasts scrambled siRNA and WRN-specific siRNA. (a) WRN protein (GM08402 at passage 8) were transfected with a siRNA detection with an anti-WRN antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) in GM08402 cells 48 h after siRNA specific for WRN mRNA (referred to as siWRN molecule transfection. The b-actin protein present in the lysate was hereafter). Scrambled siRNA was used for control used as control. (b) Percentage of GM08402 cells (at passage 8) in transfections. Transfections were performed with each phase of the cell cycle 48 h after transfection with control Lipofectamine 2000 (Invitrogen Inc., Burlington, ON, siRNA and siWRN molecules determined by FACS analysis. Experiments were performed in duplicate. (c) Detergent-solubilized Canada) on 50% confluent cell culture accordingly to lysates from control siRNA- and siWRN-transfected fibroblasts the manufacturer’s instructions. Under these conditions, were subjected to Kinetworks custom multi-sample screen (KCSS- transfected cells only went through one population 1.0) analyses (http://www.Kinexus.ca). Antibodies used in this doubling. Although siWRN-transfected cells exhibited study (from Kinexus Bioinformatics Corp.) were against phospho- DNA damage during the first 48 h of WRN protein threonine 232 of FOS, phospho-serines 21 of GSK3a/b, IKKb, phospho-serine 729 of PKCe, PKCl/i and phospho-serine 910 of knock down (Turaga et al., 2007), this strategy PKD. WRN, protein defective in patients with Werner’s syndrome; permitted us to avoid the accumulation of fixed siRNA, small interfering RNA; FACS, fluorescence-activated cell that would occur only after several rounds sorting; PKC, protein kinase C.

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1488 Bioinformatics Corp. (Vancouver, BC, Canada). The results indicate that WRN-depleted cells behaved screen used antibodies to track expression levels and similarly to WS cells in these assays. phosphorylation states of 608 cell-signaling proteins in duplicate (using 258 phospho-site-specific and 350 pan- Identification of RACK1 as a cellular partner of the WRN specific antibodies). An image of the scans of the protein antibody microarray is shown in Supplementary Figure To identify new WRN protein partners that may affect S2. Raw data obtained from the scans are shown in the activity of PKCs, WRN cDNA was cloned in frame Supplementary Table 1. A list of proteins (some were with a TAP tag containing a calmodulin- and a phosphorylated) showing a two-fold difference between streptavidin-binding peptide for TAP. As the expression cells transfected with a control-scrambled siRNA and of the TAP-WRN construct was toxic to normal human siWRN molecules is shown in Table 1. Every scan fibroblasts (GM08402), we transfected the fibrosarcoma measurement that was flagged by the Kinexus service or cell line HT1080. Several stable viable clones were showed an error range >35% was excluded. As obtained. Copurification of proteins was achieved on indicated in Table 1, the expression level of one tyrosine exponentially growing HT1080 cells expressing either phosphatase (PTP1D), two tyrosine kinases (FGFR2 the TAP alone or the TAP-WRN construct. Unbound and phospho-PDGFRa) and four PKC isoforms proteins at each step of the chromatography procedure (phospho-PKCbII, phospho-PKCd, phospho-PKCe were removed by extensive washing, thus obtaining and PKCl/i) were changed in fibroblasts transfected proteins stringently bound to the TAP-WRN construct with siWRN compared with control siRNA molecules. in cell lysates. Bound proteins were identified by liquid In addition, siWRN transfection induced phosphoryla- chromatography tandem mass spectrometry. Proteins tion of FOS at threonine 232 and hypo-phosphorylation identified from both HT1080 TAP and TAP-WRN of the pRb tumor suppressor. The hypo-phosphoryla- expressing cells were considered artifacts, and were tion of pRb is consistent with the observed accumula- removed from the final list of potential WRN interacting tion of cells in the G1 phase of the cell cycle 48 h after proteins. Table 2 lists 12 potential proteins interacting siWRN transfection. Follow-up studies conducted using with the TAP-WRN construct in HT1080 cells. The first the Kinetworks immunoblot services (Vancouver, BC, three proteins identified with 100% confidence included Canada) confirmed the changes observed with the the heterodimer (XRCC5/XRCC6 or Ku80/70) and antibody screen for phospho-FOS, phospho-PKCe and the T-complex protein 1 subunit-g (a member of the PKCl/i (Figure 1c). The phosphorylation of GSK3a/b, chaperonine family). Thus, the liquid chromatography IKKb and PKD was also examined and considered as tandem mass spectrometry analysis confirmed the the negative control (on the basis of the antibody array). WRN–Ku complex interaction in cells (Cooper et al., These results indicate that the activity of several PKC 2000; Li and Comai, 2000). The interaction with the isoforms is affected by short-term WRN depletion in chaperonine may be due to misfolding of a number of normal fibroblasts. The same proteins were compared TAP-WRN molecules in cells. The fourth protein using normal fibroblasts (GM08402) and fibroblasts identified with 99% confidence was the Guanine derived from a WS patient (AG03141D). Supplementary nucleotide-binding protein (), b-polypeptide Figure S3 shows an increase in phospho-FOS, phospho- 2- like 1 variant also known as RACK1. The other PKCe and PKCl/i in WS cells compared with normal proteins listed in Table 2 were identified with 64% fibroblasts. There was a small change in the phosphor- o confidence and were not pursued further in this study. ylation of GSK3a/b and IKKb. However, there was a marked decrease in PKD expression in WS cells. These RACK1 physically associates with WRN To confirm the binding of RACK1 to the TAP-WRN Table 1 List of phospho-proteins exhibiting a greater than a two-fold construct and to exclude the possibility that this change in normal human GM08402 fibroblasts transfected with small association could result from independent binding of interference RNA against the WRN mRNA these proteins to nucleic acid, TAP purification was Protein Phosphorylation site % CFCa % Error repeated on cell extracts pretreated with benzonase and analyzed by western blotting. Figure 2a shows that FGFR2 Pan-specificb 99.7 29.27 FOS Threonine 232 155.7 16.96 RACK1 was still detected in the TAP-WRN precipitate, Paxillin 1 Tyrosine 118 184.0 16.89 but not in precipitates from cells expressing TAP alone, PDGFRa Tyrosine 742 97.0 9.33 suggesting that nucleic acids are not required for the PKCbII Threonine 641 162.4 34.51 WRN/RACK1 association. Coimmunoprecipitation PKCd Threonine 507 138.9 33.94 experiments were then conducted on endogenous PKCe Serine 729 309.9 21.86 PKCl/i Pan-specific 222.9 3.36 WRN proteins in HT1080 cells to detect endogenous PTP1D Pan-specific À99.1 17.50 RACK1. As indicated in Figure 2b, an antibody against Rb Serine 612 À96.4 20.99 the N-terminus portion of the WRN protein coimmuno- Rb Serines 807 þ 871 À169.2 26.09 precipitated RACK1, but an antibody against the C-terminus region of the WRN protein did not. Neither Abbreviation: WRN, protein defective in patients with Werner’s syndrome. of the anti-RACK1 antibodies tested immunoprecipi- a% (percentage) change from control siRNA-transfected cells. tated RACK1, preventing us from performing the bThe antibody also recognizes the unphosphorylated protein. reverse coimmunoprecipitation experiments.

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1489 Table 2 List of TAP-WRN-binding proteins in the HT1080 Depletion of RACK1 in normal cells induces DNA fibrosarcoma cell line damage MS/MS identified Accession MW % Confidence As RACK1 affects the exonuclease activity of WRN proteins no. (kDA) (no. of peptides in vitro, we next examined its impact on overall DNA found) damage in normal diploid fibroblasts (GM08402). Normal fibroblasts were transfected with a scrambled Werner’s syndrome protein Q14191 163 100 (49) XRCC5 protein (Ku80) Q4VBQ5 83 100 (4) control siRNA, siRNA against RACK1 (siRACK1), T-complex protein 1 P49368 61 100 (2) siWRN, or both siRACK1 and siWRN molecules. subunit-g Forty-eight hours later, cells were processed for Ku 70 protein P12956 70 100 (2) immunofluorescence analysis with an antibody against RACK1 protein Q53HU2 35 99 (2) Hypothetical protein DDX18 Q4ZG72 62 64 (1) g-H2AX, which marks double-stranded DNA breaks Hypothetical protein SAP130 Q53T46 78 64 (1) (Rogakou et al., 1998). The number of nuclear foci Leu-rich repeat-containing Q96CN5 76 62 (1) stained by anti-g-H2AX was calculated on the basis of protein 45 150 transfected cells. Supplementary Figure S4 shows an Protein FLJ25373 Q96LM3 46 58 (1) example of transfected cells stained with anti-g-H2AX WD40 protein Q6UXN9 35 55 (1) ASCC3 protein Q4G1A0 84 53 (1) and DAPI. As indicated in Figure 3c, the number of ATPase family, AAA domain Q5SV24 64 53 (1) fibroblasts with >20 nuclear foci (double-stranded containing 3A breaks) was 50% higher in cells transfected with Protein DKFZp686I1974 Q6AW91 50 51 (1) siRACK1 or siWRN molecules. The number of cells with >20 nuclear foci was increased twofold in cells Abbreviations: MS, mass spectrometry; MW, molecular weight; RACK1, receptor for activated C-kinase 1; TAP, tandem affinity transfected with both siRACK1 and siWRN molecules purification; WRN, protein defective in patients with Werner’s compared with control siRNA-transfected cells. syndrome. We also quantified the extent of double- and single- stranded breaks by alkaline comet assay in these transfected cells. As indicated in Figure 3d, the mean To determine whether the WRN/RACK1 association tail length of damaged DNA in siRACK1-transfected was direct, purified His-WRN protein was incubated cells was comparable with that in siWRN-transfected with purified GST-RACK1 immobilized on glutathione- fibroblasts. When both siRACK1 and siWRN molecules Sepharose beads (GE-Healthcare Bio-Sciences Corp., were transfected, the mean tail length increased above Piscataway, NJ, USA). GST containing beads were used those levels observed in the siRACK1-transfected cells as control. As shown in Figure 2c, WRN bound GST- (Figure 3d). Overall, these results indicate that a RACK1 but not GST, suggesting that this complex is depletion of RACK1 protein in fibroblasts induced formed through a protein–protein interaction. We next DNA damage in normal diploid fibroblasts. However, mapped the region of the WRN protein required for there was no synergistic effect when both RACK1 and RACK1 binding. Different GST-WRN fragment con- WRN proteins were depleted in fibroblasts. structs were immobilized on glutathione-Sepharose resins and incubated with HT1080 cell lysates. Bound proteins were then probed with an anti-RACK1 anti- WRN protein colocalizes with nuclear body. As shown in Figures 2d and e, RACK1 bound RACK1 in normal and tumor cell lines strongly to amino-acid residues 949–1092 of the WRN We next examined the localization of RACK1 and protein. RACK1 also bound weakly to residues 239–499 WRN proteins by immunofluorescence in HT1080 cells. and very weakly to residues 499–946 of the WRN RACK1 is localized in both the cytoplasm and the protein. These results indicate that RACK1 interacts nucleus of HT1080 cells (Figure 4, top row). Only a mainly with the RQC domain of the WRN protein and partial colocalization was observed between RACK1 weakly to the acidic repeats in the N-terminus region of and WRN in the cell nucleus. We also examined the WRN. It does not interact with residues 1072–1236 at localization of RACK1 and WRN proteins in normal the C-terminus region of WRN. diploid fibroblasts (GM08402) and several cancer cell lines. Figure 4 shows that the WRN protein is mainly RACK1 inhibits the exonuclease activity of WRN protein localized to the nucleoli of normal diploid fibroblasts as We next investigated whether RACK1 could affect the described earlier (Marciniak et al., 1998), with some in exonuclease and DNA helicase enzymatic activities of the nucleoplasm. Interestingly, a fraction of nucleoplas- WRN in vitro. WRN and RACK1 proteins were thus mic WRN colocalized with nuclear RACK1. There was purified in vitro, mixed, and the exonuclease/helicase no or very little RACK1 in the nucleoli (Figure 4, activity of this complex was assayed on a forked DNA bottom row). duplex. No DNA binding or helicase/exonuclease We then examined the impact of WRN depletion on activity was observed for purified RACK1 alone. RACK1 proteins in HT1080 cells. As shown in RACK1 had no impact on WRN helicase activity. In Figure 5a, the total cellular RACK1 levels did not contrast, a strong inhibition of the WRN exonuclease change upon WRN depletion in HT1080 cells activity was noted at all tested RACK1 concentrations, (Figure 5a). However, immunofluorescence analyses suggesting a specific inhibitory function for this inter- with anti-RACK1 antibodies indicated a partial loss of action (Figure 3a). RACK1 proteins in the nuclei of siWRN-transfected

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1490

Figure 2 Interaction of RACK1 with the WRN protein. (a) Coprecipitation of RACK1 with the TAP-WRN protein in HT1080 fibrosarcoma cells. Proteins from TAP and TAP-WRN expressing cells were eluted from the streptavidin beads and analyzed by SDS– PAGE with antibodies against WRN and RACK1 proteins. (b) Coimmunoprecipitation of human WRN protein with RACK1. Approximately 2 mg of proteins from human HT1080 cells were immunoprecipitated with antibodies against the N- or C-terminus region of the human WRN protein. Control antibodies were of the same IgG species. Immunoprecipitates were analyzed by western blotting with the anti-WRN antibody (WRN; top panel) and an antibody against RACK1. Proteins were revealed with an ECL kit. The anti-WRN (N-ter) antibody is from Novus Biologicals (Littleton, CO, USA). The anti-WRN (C-ter) and anti-RACK1 antibodies are from Santa Cruz Biotechnology. The ‘input’ lane corresponds to 20 mg of total cell lysate. (c) Binding of purified WRN protein to GST-RACK1 affinity Sepharose beads as revealed by immunoblots with an anti-WRN antibody. GST Sepharose beads were used as a negative control. The experiment is presented in duplicate. (d) Interaction of RACK1 with different domains of WRN in whole-cell extract. Immunoblot against RACK1 protein bound to different GST-WRN affinity Sepahrose beads. Human HT1080 whole-cell extracts were incubated with 50 mg of the GST-WRN fragments or GST-linked glutathione-Sepharose beads overnight. Proteins bound to the affinity beads were analyzed by SDS–PAGE with antibodies against RACK1. (e) Schematic representation of different WRN fragments that were used in the WRN affinity chromatography experiments. Each domain of the WRN protein is indicated on the full WRN protein figure. The amino-acid residues of the WRN fragments used in this study are indicated on the top of each construct. Binding of RACK1 is indicated on the right by the ‘ þ ’ sign. The ‘À’ sign indicates no binding detected. RACK1, receptor for activated C-kinase 1; TAP, tandem affinity purification; WRN, protein defective in patients with Werner’s syndrome.

HT1080. The same result was obtained with normal RACK1 in siWRN-transfected cells compared with GM08402 fibroblasts transfected with siWRN molecules control siRNA-transfected cells (Figure 5c, right pa- (Figure 5b). RACK1 is an adaptor protein which nels). No change was observed in levels of cytoplasmic relocalizes to membrane fractions upon activation of b-tubulin or nuclear HNRPK splicing factor in all different signal transduction pathways (Battaini et al., transfected cells. 1997; Grosso et al., 2008). We thus examined levels of RACK1 protein in the membrane fractions of trans- fected cells. Fractionation of transfected cells into RACK1 interacts with several PKC isoforms nuclear, cytosolic and membrane fractions were per- upon WRN depletion formed, and the amount of RACK1 was quantified by RACK1 interacts with several PKC isoforms upon western blotting. As indicated in Figure 5c (left panels), different physiological conditions. These interactions there was an increase in RACK1 association with the are considered important for PKC activation and membrane fraction in siWRN-transfected HT1080 phosphorylation of specific substrates in cells. We thus cells compared with control siRNA-transfected cells. examined the association of RACK1 with three different Concomitantly, we detected a decrease in nuclear PKCs in WRN-depleted HT1080 cells. To look at

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1491

Figure 3 The impact of purified RACK1 on DNA helicase and exonuclease activities of purified WRN protein, and the impact of RACK1 depletion on DNA damage in cells. (a) The indicated concentration of purified human RACK1 protein was incubated with 3nM of purified WRN protein, and the indicated radioactive DNA substrate under standard conditions for helicase activity for 30 min at 37 1C. Reactions were stopped in the appropriate dye buffer, and DNA products were run on a 12% native polyacrylamide gel. The double- and single-stranded DNA structures are depicted on the left of the autoradiogram. The 50-labeled strand of the duplex is represented by an asterisk (*). The triangle represents heat-denatured DNA. (b) The indicated concentrations of purified human RACK1 and WRN proteins were incubated with a radioactive DNA-forked structure under the same buffer conditions as for the helicase assay for 30 min at 37 1C. Reactions were stopped in the appropriate dye buffer, heat denatured, and the DNA substrates were analyzed on a 12% denaturing gel. (c) The percentage of cells displaying foci of DNA damage detected by the anti-g-H2AX antibody. The percentage of cells exhibiting 0–10 foci, 11–20 foci or >20 foci is depicted for each transfection experiment (150 cells per transfection were analyzed). The resulting contingency table is displayed as a histogram. (d) Extent of DNA breaks quantified by the alkaline comet assay. The tails of broken DNA from 100 cells were measured to obtain the mean tail length (in mm). All experiments were repeated twice (unpaired Student’s t-test: *P ¼ 0.049; **Po0.001. Bars represent s.e.m.). RACK1, receptor for activated C-kinase 1; WRN, protein defective in patients with Werner’s syndrome.

RACK1–PKC interactions, each PKC isoform was its threonine 507. Such phosphorylation is required for coprecipitated with a TAP-RACK1 construct. TAP- the activation of PKCbII (Edwards et al., 1999) and RACK1 and TAP expressing HT1080 cells were PKCd (Parekh et al., 1999) in cells, respectively. We next transfected with control siRNA or siWRN, and the determined whether phosphorylated PKCbII and PKCd TAP construct was purified along with their complex. (activated PKCs) were still bound to TAP-RACK1 in The eluted TAP–RACK1 complex was then analyzed by HT1080 cells. As shown in Figure 6b, both phosphory- western blotting with antibodies against PKCbII, PKCd lated PKCbII (at its threonine 642) and PKCd (at and PKCe. As indicated in Figure 6a, PKCbII and its threonine 507) were bound to TAP-RACK1 in PKCd isoforms were associated with RACK1 only in WRN-depleted cells. WRN-depleted cells. In contrast, PKCe did not A knock down of the WRN protein is known to cause associate with TAP-RACK1 under the same conditions. oxidative stress and DNA damage (Das et al., 2007; These results indicate that loss of WRN causes a sub- Turaga et al., 2007). We thus observed the interaction of population of RACK1 proteins to bind to PKCbII and PKCs with TAP-RACK1 at different time points after 2 PKCd. Noticeably, Table 1 shows that WRN depletion camptothecin (5 mM), ultraviolet (40 J/m )orH2O2 in normal diploid fibroblasts also increases the phos- (0.5 mM) treatments. No increase in RACK1/WRN or phorylation of PKCbII at its threonine 641 and PKCd at RACK1/TAP-WRN association was detected under

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1492

Figure 4 Nuclear RACK1 colocalizes with nucleoplasmic WRN protein in normal and tumor cells. Immuno-localization of RACK1 and WRN proteins in HT1080 fibrosarcoma cells is shown in the top row. The bottom row shows immuno-localization of RACK1 and WRN proteins in normal GM08402 fibroblasts. Anti-mouse Alexa-594-labeled and anti-rabbit Alexa-488-conjugated secondary antibodies were used to visualize WRN and RACK1 by confocal microscopy at 568 and 488 nm, respectively. Images depict representative cells at  600. In the merged images, a yellow color appears where RACK1 (red) and WRN (green) fluorescence signals coincide. RACK1, receptor for activated C-kinase 1; WRN, protein defective in patients with Werner’s syndrome.

these conditions (data not shown), although increased examined the impact of knocking down RACK1 or phosphorylation of PKCbII (at its threonine 642) PKCd in such cells. As shown in Figures 8a and b, the and PKCd (at its threonine 507) was detected after efficiency of RACK1 and PKCd knock down ultraviolet or peroxide treatments (Supplementary was approximately 64–42% in GM08402 fibroblasts. Figure S5a). Camptothecin and peroxide treatments Depletion of WRN increased ROS formation by 34% increased p38 mitogen-activated protein kinase phos- (Figure 8c). ROS levels were decreased to near basal phorylation in HT1080 cells, indicating that such levels when both siWRN and siRACK1 molecules were treatments induced cellular stress (Supplementary transfected. The decrease was not significant (P-value of Figure S5b). However, a depletion of the WRN protein 0.065; Student’s unpaired t-test), but there was an did not increase p38 phosphorylation in HT1080 obvious trend toward normalization. In contrast, (Supplementary Figure S5c), confirming the antibody siPKCd molecules significantly decreased ROS produc- microarray screen (Supplementary Table 1). tion in WRN-depleted fibroblasts (P-value o0.05; Figure 8c). RACK1 modulates PKCd and PKCbII phosphorylation in WRN-depleted cells We determined the impact of knocking down RACK1 Discussion protein levels on PKC phosphorylation in WRN- depleted normal fibroblasts. GM08402 fibroblasts were In this study, we examined the impact of WRN transfected with either control siRNA, siWRN or both depletion on the cellular phosphorylation of specific siWRN and siRACK1 molecules. As indicated in kinases in normal and cancer cells. Previous observa- Figure 7, depletion of the WRN protein increased the tions indicated that cells derived from mice with a phosphorylation of PKCd,PKCbII and PKCe. Knock- deletion in the helicase domain of the murine Wrn gene ing down RACK1 in WRN-depleted cells significantly exhibited changes in the phosphorylation status of decreased the phosphorylation of PKCbII (at its several kinases, including PKCd (Massip et al., 2009). threonine 642) and PKCd (at its threonine 507). The major kinases that exhibited expression and activity However, it had no impact on the phosphorylation of changes in short-term WRN-depleted normal fibroblasts PKCe (at its serine 729) induced by depletion of the were PKCbII, PKCd,PKCe and PKCl/i. Such kinases WRN protein (Figure 7, bottom panels). These results can be activated by changes in either insulin or glucose suggest that depletion of WRN activates PKCe levels and transport (for example, PKCbII and PKCl/i), independently of RACK1. shear stress (PKCbII), intracellular redox status (PKCbII and PKCd), cellular morphology changes Knock down of RACK1 and PKCd diminishes induction (PKCe), growth/differentiation factors (PKCbII and of production in WRN-depleted PKCe) or DNA damage (PKCd) (Li and Xu, 2000; normal fibroblasts Ottaviani et al., 2001; Besson et al., 2002; Yoshida et al., As WRN depletion increases reactive oxygen species 2003; Stawowy et al., 2005; Sajan et al., 2006; Min et al., (ROS) production in normal GM08402 fibroblasts, we 2009). In contrast to several PKCs, a knock down of the

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1493

Figure 5 The impact of WRN depletion on RACK1 in HT1080 and GM08402 cells. (a) Expression levels of total WRN and RACK1 proteins in control siRNA- and siWRN-transfected HT1080 cells. The b-actin protein is used as loading control. (b) RACK1 protein localization detected by immunofluorescence in HT1080 and GM08402 cells transfected with control siRNA or siWRN molecules. (c) Expression levels of RACK1 proteins in the membrane fraction of control siRNA- and siWRN-transfected HT1080 cells (panels on the left). TRAP1 protein was used as loading control for the membrane fraction (including mitochondrial membranes). b-Tubulin was used as a control for cytosolic fractions. Expression levels of RACK1 proteins in the nuclear and cytoplasmic fractions of control siRNA- and siWRN-transfected HT1080 cells (panels on the right). b-Tubulin was used as a control for the cytoplasmic fraction. Heterogeneous nucleoriboprotein K (HNRPK) was used as a control for the nuclear fraction. RACK1, receptor for activated C-kinase 1; WRN, protein defective in patients with Werner’s syndrome; siRNA, small interfering RNA.

WRN protein (up to 72 h) was insufficient to induce DNA damage, which causes replication fork collapse phosphorylation and activation of p38 as described for (Cheng et al., 2008). fibroblasts derived from WS patients (Davis et al., Our results indicate the fact that RACK1 is required 2007). Finally, phosphorylation of the ATM protein for the activation of PKCd and PKCbII in WRN- (and thus its activation) was not detected, although a depleted cells. Most importantly, RACK1 was the 48 h knock down of the WRN protein was sufficient to fourth most abundant protein partner to the TAP- cause double-stranded breaks (Turaga et al., 2007). This WRN construct in HT1080 cells. Other proteins result is consistent with a recent report indicating that included the Ku heterodimer, a known partner (Cooper the WRN protein is required for ATM activation upon et al., 2000; Li and Comai, 2000), which established the

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1494

Figure 6 The impact of WRN protein depletion on RACK1/PKC Figure 8 The impact of RACK1 protein knock down on reactive associations. (a) Coprecipitation of three different protein kinase C oxygen species (ROS) production and markers in with the TAP-RACK1 protein in HT1080 fibrosarcoma cells. WRN-depleted normal human GM08402 fibroblasts. (a) RACK1 Proteins from TAP and TAP-RACK1 expressing cells were eluted protein levels in GM08402 fibroblasts 48 h after transfection with from the streptavidin beads and analyzed by SDS–PAGE with control-scrambled siRNA or siRACK1 molecules. b-actin is antibodies against PKCd, PKCe and PKCbII. (b) Coprecipitation used as control. (b) PKCd protein levels in GM08402 fibro- of two phosphorylated PKC with the TAP-RACK1 protein in blasts 48 h after transfection with control-scrambled siRNA or HT1080 fibrosarcoma cells. Proteins from TAP and TAP-RACK1 siPKCd molecules. RACK1 protein level is used as a control. expressing cells were eluted from the streptavidin beads and (c) Intracellular ROS levels in GM08402 cells transfected analyzed by SDS–PAGE with antibodies against phospho-threo- with control siRNA (siControl), siWRN, siWRN þ siRACK1 or nine 507 of PKCd and phospho-threonine 641 of PKCbII. TAP- siWRN þ siPKCd molecules in combination. Experiments were RACK1 is shown in the top panel with an anti-RACK1 antibody. performed in triplicates (unpaired Student’s t-test; *Po0.05). Bars RACK1, receptor for activated C-kinase 1; PKC, protein kinase C; represent s.e.m. WRN, protein defective in patients with Werner’s syndrome; TAP, tandem affinity purification.

interacting proteins were detected in our screen, as we extensively washed the complexes at each step of the procedure to capture the most stringent WRN protein interactants. RACK1 binds mainly to the RECQ domain and more weakly to the acidic domain of WRN (between the exonuclease and helicase domains of WRN). It inhibits the exonuclease activity of WRN without affecting its helicase activity in vitro. Interest- ingly, RACK1 has the opposite effect on WRN exonuclease activity than the Ku complex (Cooper et al., 2000). Thus, RACK1 and Ku complex could specifically coregulate the exonuclease activity of WRN depending on the levels of DNA damage or cellular conditions. Notably, depletion of RACK1 causes DNA damage in cells. It is possible that in the absence of RACK1, there is a dysregulation of WRN exonuclease activities leading to increased DNA damage (single- stranded gaps or double-stranded breaks). RACK1 may also associate and regulate other DNA repair proteins. Figure 7 The impact of RACK1 on the phosphorylation of Experiments are currently underway to identify other PKCd, PKCbII and PKCe in WRN-depleted GM08402 cells. DNA repair enzymes that interact with the TAP- Immunoblot analyses with antibodies against WRN, RACK1, RACK1 construct. Although the binding of RACK1 PKCd, phospho-threonine 507 of PKCd, total PKCd, phospho- to the WRN protein is outside the exonuclease domain threonine 641 of PKCbII, total PKCbII, phospho-serine 729 of (and helicase domain), the reason for its inhibitory effect PKCe, and total PKCe 48 h after GM08402 cells were transfected with control-scrambled siRNA, and siWRN with or without on the exonuclease of WRN remains unknown. It is siRACK1 molecules. RACK1, receptor for activated C-kinase 1; noteworthy that there are other examples of WRN PKC, protein kinase C; WRN, protein defective in patients with interacting partners that regulate its exonuclease activ- Werner’s syndrome; siRNA, small interfering RNA. ity, but bind to other regions of the WRN protein. For example, the main binding regions of the KU hetero- dimer lie outside the exonuclease domain but strongly suitability of our TAP strategy to identify new WRN stimulates the exonuclease of WRN (Karmakar et al., cellular partners. We also detected the ATPase AAA 2002). The full-length WRN protein has not yet been protein, another protein interaction reported previously crystallized, and its structure has therefore not been (Indig et al., 2004). It is noteworthy that few TAP-WRN determined. Such crystallography analysis will be an

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1495 important advance in the visualization of the three- detected in siWRN-transfected cells, but we did not dimensional structure of the WRN protein and in the detect an association of TAP-RACK1/PKCe. Further- understanding of the regulation of the exonuclease. more, siRACK1 did not inhibit PKCe phosphorylation Finally, we have no evidence that RACK1 affects the in WRN-depleted cells. This suggests that PKCe nucleolar localization of WRN (data not shown) as it phosphorylation is independent of RACK1 in WRN binds to a region of WRN containing the nucleolar protein-depleted cells unlike PKCbII and PKCd. localization sequence (von Kobbe and Bohr, 2002). It is It is noteworthy that PKCd is known to modulate noteworthy that several proteins other than RACK1 ROS production from the mitochondria (Kohda and bind to this particular region 949–109 (for example, Gemba, 2005). As WS patients are known to exhibit a TRF2 and FEN1; Brosh et al., 2001; Opresko et al., pro-oxidant state (Pagano et al., 2005), there is the 2002), and we have no indication that any of these might interesting possibility that depletion of WRN activates affect WRN nuclear localization, but it does remain an PKCd through RACK1 displacement from the nucleus interesting question for further pursuit. to the cytoplasm, which in turn modulates ROS Depletion of WRN also induces oxidative stress in production from the mitochondria. Accordingly, we addition to DNA damage (Szekely et al., 2005; Turaga decreased the production of ROS in WRN-depleted cells et al., 2007). The exact mechanism by which depletion of by transfecting siPKCd molecules. The siRACK mole- WRN impacts the redox status of cells is unclear. cules were less efficient in reducing ROS production in However, it is known that WRN affects (directly or WRN-depleted cells maybe because of a partial knock indirectly) the expression of several involved in down of RACK1 proteins in cells. Indeed, we detected a lipidogenesis or oxidative metabolism (Massip et al., certain phosphorylation level of PKCd in siRACK1- 2009; Turaga et al., 2009). Such changes in the cellular transfected cells. Thus, application of appropriate redox status will affect PKC activities (Konishi et al., kinase inhibitors may well form the basis of antiaging 1997; Ohmori et al., 1998) through phosphorylation by therapies for individuals with WS as recently proposed an unknown kinase. In this study, we detected (Davis et al., 2007). phosphorylation of PKCbII and PKCd upon ultraviolet or peroxide treatments. However, we did not observe an association of these PKCs with the TAP-RACK1 Materials and methods construct in the lysate of treated cells. This suggests that the oxidative stress and DNA damage observed in The ‘Materials and methods’ section can be found at the WRN-depleted cells may not be the inducers of the Journal’s web site as Supplementary Materials. RACK1/PKCs association. A depletion of the WRN protein in cells causes a displacement of a portion of RACK1 from the nucleus to the cytoplasm in both Conflict of interest HT1080 and GM08402 cells, where it interacts with several PKCs. It is unknown whether RACK1 molecules The authors declare no conflict of interest. that were interacting with the WRN protein are the exact molecules that translocated into the cytoplasm upon WRN depletion. However, we observed an Acknowledgements increase in RACK1 association with PKCd and PKCbII We are grateful to Mrs N Roberge for the FACS analyses (Centre upon WRN depletion in HT1080 cells. RACK1 is well de Recherche en Cance´rologie, Quebec City, Qc). This study was known to regulate the activity of PKCbII, PKCd and supported by the Canadian Institutes of Health Research to ML PKCe (Pass et al., 2001; Besson et al., 2002; Osmanagic- and in part by funds from the intramural Program of the National Myers and Wiche, 2004; Grosso et al., 2008; Slager Institute on Aging, NIH to VAB. ML is a senior scholar from the et al., 2008). An increase in PKCe phosphorylation was Fonds de la Recherche en Sante´du Que´bec.

References

Battaini F, Pascale A, Paoletti R, Govoni S. (1997). The role of human flap endonuclease 1 and stimulates its cleavage activity. anchoring protein RACK1 in PKC activation in the ageing rat EMBO J 20: 5791–5801. brain. Trends Neurosci 20: 410–415. Cheng WH, Muftuoglu M, Bohr VA. (2007). Werner syndrome Beckman KB, Ames BN. (1998). The free radical theory of aging protein: functions in the response to DNA damage and replication matures. Physiol Rev 78: 547–581. stress in S-phase. Exp Gerontol 42: 871–878. Berns H, Humar R, Hengerer B, Kiefer FN, Battegay EJ. (2000). Cheng WH, Muftic D, Muftuoglu M, Dawut L, Morris C, Helleday T RACK1 is up-regulated in angiogenesis and human carcinomas. et al. (2008). WRN is required for ATM activation and the S-phase FASEB J 14: 2549–5258. checkpoint in response to interstrand cross-link-induced DNA Besson A, Wilson TL, Yong VW. (2002). The anchoring protein double-strand breaks. Mol Biol Cell 19: 3923–3933. RACK1 links protein kinase Cepsilon to integrin beta chains. Cooper MP, Machwe A, Orren DK, Brosh RM, Ramsden D, Bohr Requirements for adhesion and motility. JBiolChem277: 22073– VA. (2000). Ku complex interacts with and stimulates the Werner 22084. protein. Genes Dev 14: 907–912. Brosh Jr RM, von Kobbe C, Sommers JA, Karmakar P, Opresko PL, Corsini ER, Racchi M, Sinforiani E, Lucchi L, Viviani B, Rovati GE Piotrowski J et al. (2001). Werner syndrome protein interacts with et al. (2005). Age-related decline in RACK-1 expression in human

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1496 leukocytes is correlated to plasma levels of dehydroepiandrosterone. and cardiac fibrosis in mice with a in the helicase domain J Leuk Biol 77: 247–256. of the Werner syndrome gene homologue. Exp Gerontol 41: Das A, Boldogh I, Lee JW, Harrigan JA, Hegde ML, Piotrowski J 157–168. et al. (2007). The human Werner syndrome protein stimulates repair Massip L, Garand C, Paquet E, Cogger VC, O’Reilly J, Tworek L of oxidative DNA base damage by the DNA glycosylase NEIL1. et al. (2009). Vitamin C restores healthy aging in a mouse model for J Biol Chem 282: 26591–26602. Werner syndrome. FASEB J 2009 (e-pub ahead of print 9 Davis T, Wyllie FS, Rokicki MJ, Bagley MC, Kipling D. (2007). The September 2009; doi:fj.09-137133). role of cellular senescence in Werner syndrome: toward therapeutic McCahill A, Warwicker J, Bolger GB, Houslay MD, Yarwood SJ. intervention in human premature aging. Ann N Y Acad Sci 1100: (2002). The RACK1 scaffold protein: a dynamic cog in cell response 455–469. mechanisms. Mol Pharmacol 62: 1261–1273. Davis T, Kipling D. (2008). Assessing the role of stress signaling Melcher R, von Golitschek R, Steinlein C, Schindler D, Neitzel H, via p38 MAP kinase in the premature senescence of Ataxia Kainer K et al. (2000). Spectral karyotyping of Werner syndrome Telangiectasia and Werner syndrome fibroblasts. Biogerontology fibroblast cultures. Cytogenet Cell Genet 91: 180–185. 10: 253–266. Min W, Bin ZW, Quan ZB, Hui ZJ, Sheng FG. (2009). The signal Descheˆnes F, Massip L, Garand C, Lebel M. (2005). in vivo transduction pathway of PKC/NF-kappaB/c-fos may be involved in misregulation of genes involved in apoptosis, development and the influence of high glucose on the cardiomyocytes of neonatal rats. oxidative stress in mice lacking both functional Werner syndrome Cardiovasc Diabetol 8:8. protein and poly(ADP-ribose) polymerase-1. Hum Mol Genet 14: Nitti M, Pronzato MA, Marinari UM, Domenicotti C. (2008). PKC 3293–3308. signaling in oxidative hepatic damage. Mol Aspects Med 29: Edwards AS, Faux MC, Scott JD, Newton AC. (1999). Carboxyl- 36–42. terminal phosphorylation regulates the function and sub- Ohmori S, Shirai Y, Sakai N, Fujii M, Konishi H, Kikkawa U et al. cellular localization of protein kinase C betaII. J Biol Chem 274: (1998). Three distinct mechanisms for translocation and activation of 6461–6468. the delta subspecies of protein kinase C. Mol Cell Biol 18: 5263–5271. Egidy G, Jule´S, Bosse´P, Bernex F, Geffrotin C, Vincent-Naulleau S Opresko PL, von Kobbe C, Laine JP, Harrigan J, Hickson ID, et al. (2008). Transcription analysis in the MeLiM swine model Bohr VA. (2002). Telomere-binding protein TRF2 binds to and identifies RACK1 as a potential marker of malignancy for human stimulaˆtes the Werner and Bloom syndrome helicases. J Biol Chem melanocytic proliferation. Mol Cancer 7: 34. 277: 41110–41119. Epstein CJ, Martin GM, Schultz AL, Motulsky AG. (1966). Werner’s Osmanagic-Myers S, Wiche G. (2004). Plectin-RACK1 (receptor for syndrome: a review of its symptomatology, natural history, activated C kinase 1) scaffolding: a novel mechanism to regulate pathologic features, genetics and relationship to the natural aging protein kinase C activity. J Biol Chem 279: 18701–18710. process. Medicine 45: 177–221. Ottaviani E, Franchini A, Kletsas D. (2001). Platelet-derived growth Faragher RG, Kill IR, Hunter JA, Pope FM, Tannock C, Shall S. factor and transforming growth factor-beta in invertebrate immune (1993). The gene responsible for Werner syndrome may be a and neuroendocrine interactions: another sign of conservation ‘counting’ gene. Proc Natl Acad Sci USA 90: 12030–12034. in evolution. Comp Biochem Physiol C Toxicol Pharmacol 129: Grosso S, Volta V, Sala LA, Vietri M, Marchisio PC, Ron D 295–306. et al. (2008). PKCbetaII modulates translation independently from Pagano G, Zatterale A, Degan P, d’Ischia M, Kelly FJ, Pallardo´FV mTOR and through RACK1. Biochem J 415: 77–85. et al. (2005). in vivo prooxidant state in Werner syndrome (WS): Indig FE, Partridge JJ, von Kobbe C, Aladjem MI, Latterich M, Bohr results from three WS patients and two WS heterozygotes. Free VA. (2004). Werner syndrome protein directly binds to the AAA Radic Res 39: 529–533. ATPase p97/VCP in an ATP-dependent fashion. J Struct Biol 146: Parekh D, Ziegler W, Yonezawa K, Hara K, Parker PJ. (1999). 251–259. Mammalian TOR controls one of two kinase pathways acting upon Kaneto H, Matsuoka TA, Nakatani Y, Kawamori D, Matsuhisa M, nPKCdelta and nPKCepsilon. J Biol Chem 274: 34758–34764. Yamasaki Y. (2005). Oxidative stress and the JNK pathway in Parrinello S, Samper E, Krtolica A, Goldstein J, Melov S, Campisi J. diabetes. Curr Diabetes Rev 1: 65–72. (2003). Oxygen sensitivity severely limits the replicative lifespan of Karanjawala ZE, Murphy N, Hinton DR, Hsieh CL, Lieber MR. murine fibroblasts. Nat Cell Biol 5: 741–747. (2002). Oxygen metabolism causes breaks and is Pascale A, Fortino I, Govoni S, Trabucchi M, Wetsel WC, Battaini F. associated with the neuronal apoptosis observed in DNA double- (1996). Functional impairment in protein kinase C by RACK1 strand break repair mutants. Curr Biol 12: 397–402. (receptor for activated Ckinase 1) deficiency in aged rat brain Karmakar P, Snowden CM, Ramsden DA, Bohr VA. (2002). Ku cortex. J Neurochem 67: 2471–2477. heterodimer binds to both ends of the Werner protein and Pass JM, Gao J, Jones WK, Wead WB, Wu X, Zhang J et al. (2001). functional interaction occurs at the Werner N-terminus. Nucleic Enhanced PKC beta II translocation and PKC beta II-RACK1 Acids Res 30: 3583–3591. interactions in PKC epsilon-induced heart failure: a role for Kohda Y, Gemba M. (2005). Cephaloridine induces translocation of RACK1. Am J Physiol Heart Circ Physiol 281: H2500–H2510. protein kinase C delta into mitochondria and enhances mitochon- Rogakou EP, Pilch DR, Orr AH, Ivanova VS, Bonner WM. (1998). drial generation of free radicals in the kidney cortex of rats causing DNA double-stranded breaks induce histone H2AX phosphoryla- renal dysfunction. J Pharmacol Sci 98: 49–57. tion on serine 139. J Biol Chem 273: 5858–5868. Konishi H, Tanaka M, Takemura Y, Matsuzaki H, Ono Y, Kikkawa Sajan MP, Rivas J, Li P, Standaert ML, Farese RV. (2006). Repletion U et al. (1997). Activation of protein kinase C by tyrosine of atypical protein kinase C following RNA interference-mediated phosphorylation in response to H2O2. Proc Natl Acad Sci USA depletion restores insulin-stimulated glucose transport. J Biol Chem 94: 11233–11237. 281: 17466–17473. Li B, Comai L. (2000). Functional interaction between Ku and the Sedding DG. (2008). FoxO transcription factors in oxidative stress Werner syndrome protein in DNA end processing. J Biol Chem 275: response and ageing––a new fork on the way to longevity? Biol 28349–28352. Chem 389: 279–283. Li C, Xu Q. (2000). Mechanical stress-initiated signal transductions in Slager RE, Devasure JM, Pavlik JA, Sisson JH, Wyatt TA. (2008). vascular smooth muscle cells. Cell Signal 12: 435–445. RACK1, a PKC targeting protein, is exclusively localized to basal Marciniak RA, Lombard DB, Johnson FB, Guarente L. (1998). airway epithelial cells. J Histochem Cytochem 56: 7–14. Nucleolar localization of the Werner syndrome protein in human Stawowy P, Margeta C, Blaschke F, Lindschau C, Spencer-Ha¨nsch C, cells. Proc Natl Acad Sci USA 95: 6887–6892. Leitges M et al. (2005). Protein kinase C epsilon mediates Massip L, Garand C, Turaga RV, Descheˆnes F, Thorin E, Lebel M. angiotensin II-induced activation of beta1-integrins in cardiac (2006). Increased insulin, triglycerides, reactive oxygen species, fibroblasts. Cardiovasc Res 67: 50–59.

Oncogene Depletion of WRN causes activation of PKCs through RACK1 L Massip et al 1497 Szekely AM, Bleichert F, Numann A, Van Komen S, Manasanch E, von Kobbe C, Bohr VA. (2002). A nucleolar targeting sequence in the Ben Nasr A et al. (2005). Werner protein protects nonproliferating Werner syndrome protein resides within residues 949–1092. J Cell cells from oxidative DNA damage. Mol Cell Biol 25: 10492–10506. Sci 115: 3901–3907. Turaga RV, Massip L, Chavez A, Johnson FB, Lebel M. (2007). Wang Z, Jiang L, Huang C, Li Z, Chen L, Gou L et al. (2008). Werner syndrome protein prevents DNA breaks upon chromatin Comparative proteomics approach to screening of potential structure alteration. Aging Cell 6: 471–481. diagnostic and therapeutic targets for oral squamous cell carcinoma. Turaga RV, Paquet ER, Sild M, Vignard J, Garand C, Johnson FB Mol Cell Proteomics 7: 1639–1650. et al. (2009). The Werner syndrome protein affects the expression of Yoshida K, Wang HG, Miki Y, Kufe D. (2003). Protein kinase Cdelta genes involved in adipogenesis inflammation in addition to cell cycle is responsible for constitutive and DNA damage-induced phosphor- DNA damage responses. Cell Cycle 8: 2080–2092. ylation of Rad9. EMBO J 22: 1431–1441.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene