Oncogene (2008) 27, 4733–4744 & 2008 Macmillan Publishers Limited All rights reserved 0950-9232/08 $30.00 www.nature.com/onc ORIGINAL ARTICLE CDK1 promotes cell proliferation and survival via phosphorylation and inhibition of FOXO1

P Liu1,2, TP Kao1,2 and H Huang1,2

1Cancer Center, University of Minnesota, Minneapolis, MN, USA and 2Department of Laboratory Medicine and Pathology, University of Minnesota, Minneapolis, MN, USA

The forkhead box O (FOXO) transcription factor (AFX) and FOXO6 in humans, belong to the super- FOXO1 functions as a tumor suppressor by regulating family of forkhead transcription factors (Greer and expression of involved in apoptosis, cell cycle arrest Brunet, 2005). Emerging evidence suggests that FOXO and oxidative detoxification. Here, we demonstrate that transcription factors possess tumor suppressor functions cyclin-dependent kinase 1 (CDK1) specifically phosphory- by regulating expression of genes involved in apoptosis, lates FOXO1 at serine 249 (S249) in vitro and in vivo. cell proliferation and genotoxic/oxidative stresses Coimmunoprecipitation assays demonstrate that both (Huang and Tindall, 2007). Activation of FOXO endogenous CDK1 and ectopically expressed CDK1 form upregulates proapoptotic genes, such as Fas a complex with FOXO1 in prostate cancer (PCa) ligand (FasL), Bim and tumor necrosis factor-related cells. In vitro protein binding assays reveal that CDK1 apoptosis-inducing ligand (TRAIL) (Brunet et al., 1999; interacts directly with FOXO1. Accordingly, overexpres- Dijkers et al., 2000; Modur et al., 2002; Gilley et al., sion of CDK1 inhibits the transcriptional activity of 2003). FOXO proteins also transcriptionally regulate FOXO1 in PCa cells through S249 phosphorylation on expression of an array of genes that encode proteins FOXO1. Consistent with the roles of FOXO3a and functioning at variouscell cycle checkpoints.FOXO FOXO4 (two other members of the FOXO family) in cell proteinsinduce cell cycle arrestat G1 by upregulating cycle regulation, forced expression of FOXO1 causes a cyclin-dependent kinase (CDK) inhibitors p27KIP1, delay in the transition from G2 to M phase. This effect is p21WAF1 and the -related protein blocked completely by overexpression of CDK1 and cyclin p130, and by downregulating cyclinsD1 and D2 B1. Ectopic expression of constitutively active CDK1 also (Medema et al., 2000; Nakamura et al., 2000; Kops inhibits FOXO1-induced apoptosis in PCa cells. More- et al., 2002; Ramaswamy et al., 2002; Schmidt et al., over, we demonstrate that the inhibitory effect of FOXO1 2002; Seoane et al., 2004). Expression of growth arrest on Ras oncogene-induced colony formation in fibroblasts and DNA-damaging-inducible protein 45a (Gadd45a), is diminished by overexpression of CDK1. Given that a protein with a critical function in the G2-to-M CDK1 and cyclin B1 are often overexpressed in human transition, is regulated by FOXO3a and FOXO4 cancers including PCa, our findings suggest that aberrant (Furukawa-Hibi et al., 2002; Tran et al., 2002), activation of CDK1 may contribute to tumorigenesis by suggesting a role of FOXO proteins in the regulation promoting cell proliferation and survival via phosphory- of mitosis. lation and inhibition of FOXO1. The importance of FOXO proteins in tumor suppres- Oncogene (2008) 27, 4733–4744; doi:10.1038/onc.2008.104; sion is further reflected by the fact that these proteins are published online 14 April 2008 often dysregulated either at the genomic level or by aberrant oncogenic signaling pathways. FOXO1, FOX- Keywords: FOXO1; CDK1; phosphorylation; cell cycle; O3a and FOXO4 genesare all affected by chromosomal apoptosis; prostate cancer translocations detected in solid tumors and leukemias (Galili et al., 1993; Borkhardt et al., 1997; Hillion et al., 1997). In addition, the activity of FOXO factorsis tightly regulated by post-translational modifications, including phosphorylation, acetylation and ubiquityla- Introduction tion. A number of inhibitory phosphorylation sites have been identified in FOXO proteins. For example, Akt O-class forkhead box (FOXO) proteins, which include inhibits the tumor suppressor function of FOXO FOXO1 (FKHR), FOXO3a (FLHRL1), FOXO4 proteinsby promoting phosphorylation of these proteinsin multiple sites(Biggs et al., 1999; Brunet et al., 1999; Kops et al., 1999; Rena et al., 1999; Tang Correspondence: Dr H Huang, University of Minnesota, MMC 806, et al., 1999). Akt phosphorylation not only causes 420 Delaware Street SE, Minneapolis, MN 55455, USA. E-mail: [email protected] nuclear export of FOXO proteinsbut alsoinduces Received 12 November 2007; revised 13 February 2008; accepted 1 polyubiquitylation and proteasome degradation of March 2008; published online 14 April 2008 FOXO1 and FOXO3a (Matsuzaki et al., 2003; Inhibition of FOXO1 by CDK1 PLiuet al 4734 Plasand Thompson, 2003; Aoki et al., 2004). Akt- greater than the phosphorylation of histone H1, a phosphorylated FOXO1 interacts with Skp2, the sub- known substrate of CDK1 (Figure 1b, upper panel). No strate-binding component of the Skp1/culin 1/F-box phosphorylation of the control GST protein was protein (SCFSkp2) E3 ligase complex and is targeted for detected under the same conditions. To rule out the ubiquitin-dependent proteasome degradation (Huang possibility that FOXO1 phosphorylation is mediated by et al., 2005). Phosphorylation of FOXO proteins by the other kinase(s) that may coimmunoprecipitate with other kinases, including serum- and glucocorticoid- CDK1, we performed in vitro kinase assays using the inducible kinases, protein kinase CK1, dual-specificity reconstituted CDK1/cyclin B1 kinase complex, and tyrosine-phosphorylated and regulated kinase 1A, and obtained similar results as seen with the endogenous IkB kinase, also inhibits the activity of FOXO proteins CDK1 immunoprecipitated from cells(Figure 1c). (Brunet et al., 2001; Woods et al., 2001; Rena et al., 2002; These data indicate that CDK1 can phosphorylate Hu et al., 2004). Our recent studies demonstrate that FOXO1 in vitro. Next, we sought to determine which CDK2, which isa kinasethought to play a critical role in residue in FOXO1 can be phosphorylated by CDK1. the G1-to-S transition, specifically phosphorylates CDKsoften recognize and phosphorylatethe serine/ FOXO1 at serine 249 (S249). Phosphorylation at this threonine-proline (S/T-P) motif. There isonly one such site leads to the nuclear exportation and subsequent motif (S249/P250) present in FO1-2, which is shared by transcriptional inhibition of FOXO1 (Huang et al., 2006). both FO1-2 and FO1-3. In addition, there are seven The cell division cycle is driven by the sequential more such motifs found in FO1-3 (Figure 1d). To activation of CDK complexes. CDK2 and other G1 determine which S/T-P motif in FOXO1 isphosphory- CDKs, including CDK4 and CDK6, along with their lated by CDK1, we generated three truncated associated cyclins, are known to be involved in the G1/S GST-FOXO1 recombinant proteinsfrom FO1-3 transition (Sherr and Roberts, 2004). The CDK1/cyclin (Figure 1d). Kinase assays demonstrated that CDK1 B kinase complex is the main regulator of the G2-to-M phosphorylates FO1-6, but not FO1-7 or FO1-8 transition. The cyclin B1/CDK1 kinase complex is (Figure 1e, left panel). Asshownin Figure 1d, there is believed to promote cell cycle transitions at late G2 only one S/T-P motif (S249/P250) present in FO1-6 and and mitosis by phosphorylating substrates and trigger- mutation of thisresidue to alanine (S249A) completely ing reorganization of the nuclear envelope, spindle abolished CDK1-mediated phosphorylation of FO1-3 apparatus and actin cytoskeleton. Interestingly, a recent (Figure 1e, right panel). S249A mutation also abolished mouse genetic study demonstrates that mammalian CDK1-mediated phosphorylation of FO1-2 (data not CDK1 is sufficient to drive the cell cycle in the absence shown). Collectively, these results suggest that CDK1 of G1 CDKs(Santamaria et al., 2007), highlighting an specifically phosphorylates FOXO1 at S249 in vitro. important role of CDK1 in regulation of the cell cycle. Indeed, cell cycle regulatory proteinsare often dysregu- CDK1 phosphorylates FOXO1 at S249 in vivo lated in human cancers. In this report, we demonstrate To determine whether CDK1 phosphorylates FOXO1 at that CDK1 bindsto and phosphorylatesFOXO1 at S249 in vivo, wild-type and S249A-mutated FLAG- S249 in vitro and in vivo, thereby abolishing the FOXO1 were transfected into LNCaP cells and inhibitory effect of FOXO1 on cell growth and survival. phosphorylation of FOXO1 was examined by immuno- Thus, our findings reveal a novel mechanism by which blotting with the phospho-specific antibody against CDK1 regulatescell proliferation and survival. phospho-S249. As shown in Figure 2a, wild-type (lane 4) but not S249A mutant FOXO1 (lane 2) was readily phosphorylated at S249. Consistently, forced expression of constitutively active CDK1 increases S249 Results phosphorylation of the wild-type FOXO1 but not the S249A mutant (Figure 2a, lane 5 versus lane 3). As CDK1 phosphorylates FOXO1 at S249 in vitro shown previously (Huang et al., 2006), ectopic expres- We demonstrated previously that CDK2 phosphorylates sion of constitutively active CDK2 increases FOXO1 FOXO1 at S249 in vitro and in vivo (Huang et al., 2006). phosphorylation at this residue (Figure 2a, lane 1). We also demonstrated that S249 phosphorylation is low Similar to CDK2, silencing of CDK1 using a pool of during the G1 phase and increased markedly in the -specific small interfering RNAs (siRNAs) resulted S phase. As FOXO1 S249 phosphorylation was also in a decrease in FOXO1 phosphorylation (Figure 2b). observed in cells in G2/M phases (Huang et al., 2006), FOXO1 phosphorylation was largely diminished in cells where CDK1 ishighly active, we sought to determine where both CDK1 and CDK2 geneswere silenced whether FOXO1 is phosphorylated by CDK1. First, we (Figure 2b). These observations indicate that in addition performed in vitro kinase assays using a series of to CDK2 (Huang et al., 2006), CDK1 can also glutathione S-transferase (GST)-FOXO1 recombinant phosphorylate FOXO1 at S249 in cells. proteins(Figure 1a and lower panel in Figure 1b) as substrates. As demonstrated in Figure 1b (upper panel), endogenousCDK1 immunoprecipitated from LNCaP CDK1 interacts with FOXO1 in vivo and in vitro prostate cancer (PCa) cells phosphorylates the FO1-2 Next, we sought to determine whether CDK1 interacts and FO1-3 GST-FOXO1 recombinant proteins. The with FOXO1 in cells. LNCaP cells were transfected with amount of CDK1 phosphorylation in FOXO1 was FLAG-tagged FOXO1 and V5-tagged CDK1. At 48 h

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4735

Figure 1 CDK1 phosphorylates FOXO1 at S249 in vitro.(a) A diagram showing GST-FOXO1 recombinant proteins. FKH, forkhead domain; NLS, nuclear localization signal; NES, nuclear export signal; TAD, transactivation domain. (b) Top: in vitro kinase assays with immunoprecipitated (IP) CDK1. Bottom: protein substrates indicated by Coomassie blue staining. Mr(K), relative molecular weight (kilodaltons). Histone H1, GST and GST-FOXO1 recombinant proteins at correct molecular mass are indicated by arrowheads. (c) Reconstituted CDK1/cyclin B1 kinase assays using the same set of proteins as substrates as indicated in panel b. (d) Schematic diagram of three GST-FOXO1 recombinant proteinsin the region of FO1-3. The locationsof eight S/T-P sitesare indicated. (e) Reconstituted CDK1/cyclin B1 kinase assays with various GST-FOXO1 recombinant proteins as substrates. Histone H1 was used as a positive control. All the experiments were performed in triplicate. CDK, cyclin-dependent kinase; FOXO, O-class forkhead box; S249, serine 249; GST, glutathione S-transferase.

after transfection, cells were collected and cell lysates 35S-labeled CDK1 proteinsproduced by in vitro were subjected to immunoprecipitation with either an transcription and translation (Figure 3d, middle panel) anti-FLAG antibody or an anti-V5 antibody. As and bacterially expressed GST-FOXO1 recombinant demonstrated in Figure, V5-tagged CDK1 was detected proteins(Figure 3d, lower panel). Asdemonstrated in in the protein complex immunoprecipitated by the anti- Figure 3d (upper panel), a strong interaction is detected FLAG antibody. Reciprocally, FLAG-tagged FOXO1 between CDK1 and FO1-3 (amino acids211–419 of wasdetected in the protein complex immunoprecipi- FOXO1). There also appears to be a weak interaction tated by the anti-V5 antibody (Figure 3b). We also between CDK1 and FO1-2 (amino acids149–267 of examined the interaction between endogenousFOXO1 FOXO1), albeit the detected band migratesslightly and CDK1 in LNCaP cells. As demonstrated in faster than expected (Figure 3d, upper panel) and the Figure 3c, endogenousCDK1 wasreadily immunopre- result was reproducible. These data suggest that CDK1 cipitated by an antibody against CDK1. Notably, interactsdirectly with FOXO1. FOXO1 proteinswere detected in the protein complex immunoprecipitated by the anti-CDK1 antibody (Figure 3c). These data suggest that CDK1 interacts CDK1 inhibits the transcriptional activityof FOXO1 with FOXO1 in vivo. To determine whether CDK1 AsFOXO1 functionsprimarily asa transcriptionfactor interactsdirectly with FOXO1 and to map the region(s) (Huang and Tindall, 2007), we sought to determine in FOXO1 that are required for thisinteraction, whether the transcriptional activity of FOXO1 is we performed in vitro protein binding assays using affected by CDK1. Asdemonstratedin Figures4a and

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4736

Figure 2 CDK1 phosphorylates FOXO1 at S249 in vivo.(a) LNCaP cells were transfected with plasmids as indicated. At 48 h after transfection, cells were lysed and cell lysates were subjected to immunoprecipitation with anti-FLAG (M2) antibodies. Immunoprecipitants were analysed by western blot using antibodies against S249-p, FOXO1, V5 tag and cyclin B1. (b) LNCaP cells transfected with FLAG-FOXO1 expression vector in combination with control siRNA, CDK1 siRNA and/or CDK2 siRNA. At 72 h after transfection, cells were harvested and cell lysates were subjected to immunoprecipitation and western blot analysis as described in panel a. Similar results were obtained from three independent experiments. CDK, cyclin-dependent kinase; FOXO, O-class forkhead box; S249, serine 249; siRNA, small interfering RNA.

b, ectopic expression of either the wild-type or phosphorylation of S249. In line with these findings, constitutively active CDK1 decreased transcriptional silencing of endogenous CDK1 and/or CDK2 increased activity of FOXO1 in LNCaP cells. Although transfec- the activity of the FOXO1 reporter gene in LNCaP cells tion of cyclin B1 alone also decreased the transcriptional (Figure 4e). activity of FOXO1 (Figures4a and b), the inhibition of FOXO1 transactivation was largely enhanced by co- transfection of cyclin B1 and CDK1 (Figure 4a). To CDK1 overcomes FOXO1-induced delayin the G2-to-M determine whether inhibition of FOXO1 by CDK1 is transition mediated through phosphorylation, a number of Previous studies demonstrated that ectopic expression of phosphorylation-resistant mutants of FOXO1 were active FOXO3a and FOXO4 inducesincreasedGad- generated, and their transcriptional activities d45a expression and a delay in G2-to-M transition were examined in LNCaP cellsin which Akt is (Furukawa-Hibi et al., 2002; Tran et al., 2002). We constitutively active due to loss of phosphatase and sought to determine whether CDK1 regulates G2-to-M tensin homologue deleted on ten (PTEN) transition via inhibition of FOXO1. To this end, we (Wu et al., 1998). As expected, without overexpression examined cell cycle transition after synchronization by of CDK1 and cyclin B1, triple mutationsof the three treating cellswith the S-phase-blockingagent aphidico- Akt phosphorylation sites (FOXO1-A3) resulted in a lin. NIH 3T3 cells were transfected with an expression high basal activity of FOXO1 in LNCaP cells vector for green fluorescence protein (GFP) and various (Figure 4c). However, the activity of the triple mutant expression constructs as indicated in Figure 5a. Cells was inhibited by the expression of CDK1 and cyclin B1 were treated with 1 mg/ml aphidicolin for 24 h followed (Figure 4c). Similarly, restored expression of PTEN by phosphate-buffered saline washes to release the cells increased the transcriptional activity of FOXO1 in from the S-phase blockage. Cells were collected at LNCaP cells and overexpression of CDK1 and cyclin different time points(3, 5, 7 and 9 h), fixed, stainedwith B1 mitigated PTEN-induced increase in FOXO1 trans- propidium iodide and subjected to fluorescence-acti- criptional activity (Figure 4d). These data suggest that vated cell sorting (FACS) analysis. As demonstrated in CDK1 inhibitsthe transcriptionalactivity of FOXO1 in Figure 5a, the majority of the cellstransfectedwith the an Akt-independent manner. Moreover, introduction of empty vector (pcDNA3.1) were in G1 and S phases 3 h the S249A mutation that makesFOXO1 imperviousto after release from the S-phase blockage. Approximately the CDK1-mediated phosphorylation (Figure 2a) re- 20% of cells were in G2/M phases. Similar results were sulted in a higher transcriptional activity than FOXO1- obtained in cellstransfectedwith FOXO1-A3 at this A3 (Figure 4c). Importantly, the inhibitory effect of time point. In contrast, approximately 45% of cells CDK1 on FOXO1-A3 wasabrogated by the S249A transfected with CDK1-AF were in G2/M phases at 3 h mutation (Figure 4c). Together, these data suggest after release (Figure 5a). Although the majority of cells that CDK1-induced inhibition of transcriptional transfected with either empty vector or CDK1-AF were activity of FOXO1 ismediated primarily by the in G1 phase 9 h after release, the majority (approximately

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4737

Figure 3 CDK1 interactswith FOXO1 in vitro and in vivo. LNCaP cellswere transfected with FLAG-tagged FOXO1 and V5-tagged CDK1. At 48 h after transfection, cells were collected and lysed in IP buffer. Cells lysates were subjected to IP with an anti-FLAG antibody (a) or an anti-V5 antibody (b) or nonspecific IgG. Immunoprecipitated proteins were analysed by western blot using antibodiesasindicated. ( c) Interaction of endogenousCDK1 and FOXO1. Exponentially growing LNCaP cellswere harvestedand lysed. Cell lysates were subjected to immunoprecipitation with an anti-CDK1 antibody. Immunoprecipitants were analysed by western blot using antibodies against FOXO1 and CDK1. (d) Binding of 35S-labeled CDK1 proteinsproduced by in vitro transcription and translation to GST-FOXO1 recombinant proteins. Top: autography of in vitro produced CDK1 proteinsbound by GST-FOXO1 recombinant proteins. Middle: input of CDK1 proteins produced by in vitro transcription and translation. Bottom: GST and GST- FOXO1 recombinant proteinsfor in vitro protein binding assays indicated by Coomassie blue staining. GST and GST-FOXO1 recombinant proteins at correct molecular mass are indicated by arrowheads. All the experiments were repeated at least twice. CDK, cyclin-dependent kinase; FOXO, O-class forkhead box; IP, immunoprecipitation; GST, glutathione S-transferase.

60%) of cellstransfectedwith FOXO1-A3 were in G2/ reduced by ectopic expression of FOXO1-A3 M phases (Figure 5a). This effect was largely inhibited (Figure 5b), suggesting an inhibitory role of FOXO1 by co-transfection with CDK1-AF (Figure 5a). Next, we in mitosis. This inhibitory effect was abrogated by examined the effect of FOXO1 on mitosis by staining co-transfection with CDK1-AF (Figure 5b). Expression cells with antibodies against histone H3 phosphoryla- of FOXO1-A3/S249A also reduced the percentage of tion at serine 10 (H3-p), a known mitotic marker (Tran cells in mitosis (Figure 5b). However, forced expression et al., 2002). Asshownin Figure 5b, approximately 40% of CDK1-AF did not override the FOXO1-A3/S249A- of cellstransfectedwith empty vector were H3-p positive mediated inhibition of mitosis. Results from a repre- at 7 h after release from aphidicolin treatment. At this sentative experiment are shown in Figure 5c. In line with time point, fewer cellstransfectedwith CDK1-AF were these findings, forced expression of FOXO1-A3 H3-p positive than cells transfected with empty vector increased the levels of Gadd45a protein expression in (Figure 5b), which is consistent with the FACS data that NIH 3T3 cells(Figure 5b, lower panel). Thiseffect was expression of CDK1-AF accelerated cell cycle progres- inhibited in cellsthat were co-transfected with CDK1- sion (Figure 5a, the first column versus the third AF and cyclin B1 (Figure 5b). However, FOXO1-A3/ column). Importantly, H3-p-positive cells were largely S249A-induced expression of Gadd45a wasnot affected

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4738

Figure 4 CDK1 inhibitsthe transcriptional activity of FOXO1. ( a) Effect of CDK1 on transcriptional activity of the wild-type FOXO1. Cells were transfected with a firefly luciferase reporter construct, which contains three copies of FOXO response element (so-called insulin-responsive sequence (3xIRS)) from the of the IGFBP-1 gene, a renilla luciferase reporter construct and plasmids as indicated. At 24 h after transfection, luciferase activities were measured and analysed as described in Materials and methods. *Po0.05 comparing the luciferase activity in cells transfected with FOXO1 plus CDK1 and/or cyclin B1 to cells transfected with FOXO1 alone. (b) An additional set of samples from panel a was analysed by western blot for protein expression of FOXO1, CDK1 and V5-tagged cyclin B1. Erk2 isshownasa loading control. ( c) Effect of CDK1 on the transcriptional activity of the FOXO1- A3 mutant, which is resistant to Akt-mediated phosphorylation of FOXO1. LNCaP cells were transfected and luciferase activities were measured as described in panel a. *,**Po0.05. (d) Effect of CDK1 on PTEN-induced activation of FOXO1. LNCaP cellswere transfected and luciferase activities were measured as described in panel a. *Po0.05. (e) Effect of silencing of endogenous CDK1 and/ or CDK2 on the activity of the FOXO reporter gene 3xIRS. LNCaP cellswere transfectedwith 3xIRS and renilla reporter genesin combination with nonspecific control siRNA or a pool of CDK1- and/or CDK2-specific siRNA. At 48 h after transfection, cells were harvested and luciferase activities were measured as described in panel a. *,**,***Po0.05. (a, c–e) Error barsindicate s.d.among three individual experiments. All the experiments were repeated at least three times. CDK, cyclin-dependent kinase; FOXO, O-class forkhead box; Erk2, extracellular signal-regulated kinase 2; siRNA, small interfering RNA.

Figure 5 CDK1 inhibitsFOXO1-induced delay in the G2-to-M transition.( a) FACS analysis of the effect of expression of CDK1 and/or FOXO1 on G2/M progression. NIH 3T3 cells were transfected with plasmids as indicated and plated in culture media containing 1 mg/ml aphidicolin. At 24 h after aphidicolin treatment, cellswere releasedby culturing in aphidicolin-free media. At different time points as indicated, cells were harvested, fixed and stained with propidium iodide (PI), and FACS analysis was performed. Quantification of a representative experiment is shown in bar graphs. (b) Upper panel: mitosis marker (histone H3 phosphorylation at Ser10) analysis of effect of expression of CDK1 and/or FOXO1 on G2-to-M transition. NIH 3T3 cells were transfected with plasmids as indicated and plated in media containing 1 mg/ml aphidicolin. At 24 h after treatment, aphidicolin was washed out to release the cell cycle blockage. At 7 h after release, immunofluorescence chemistry experiments were performed using the anti-FLAG antibody (green) or GFP to detect cells expressing FLAG-FOXO1 or control cells expressing GFP, respectively, and the anti-phospho-histone H3 Ser10 antibody (red) to detect mitotic cells. Thirty random fields were scored for cells that are positive for green and/or red. Error barsindicate s.d.among three individual experiments.* ,**,***Po0.05. Lower panel: western blot analysis of proteinsfor Gadd45 a, FOXO1 and V5-tagged CDK1 and V5-tagged cyclin B1 in cells transfected with plasmids as indicated. (c) Representative fields of NIH 3T3 cells exhibiting GFP or staining of FLAG-FOXO1 (green) colocalized with (arrowheads) or without (arrows) staining of histone H3 phosphorylation at Ser10 (red). CDK, cyclin-dependent kinase; FOXO, O-class forkhead box; FACS, fluorescence-activated cell sorting; GFP, green fluorescence protein.

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4739

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4740 by CDK1-AF transfection (Figure 5b, lower panel). that activation of CDK1 overcomesFOXO1-mediated Taken together, these data suggest that FOXO1 is a inhibition on mitosis via phosphorylation of FOXO1 at critical negative regulator of the G2-to-M transition and S249.

Figure 6 CDK1 inhibits FOXO1-induced apoptosis in LNCaP PCa cells. (a) Cells were transfected with plasmids as indicated. Forty hours after transfection, cells were harvested, fixed and stained with propidium iodide (PI), and FACS analysis was performed. The percentage of cells at sub-G1 (apoptotic cells) is indicated in each experiment. Quantification results from three independent experimentsare shown( b). *Po0.05. (c) Effect of CDK1 on FOXO1-induced expression of Bim. LNCaP cells were transfected with or without constitutively active CDK1 in combination with active FOXO1-A3 or the mutated FOXO1 (FOXO1-A3/S249A) that is resistant to CDK1-mediated phosphorylation. At 48 h after transfection, cells were collected and lysed. Cell lysates were subjected to western blot analysis for expression of Bim, PARP, FOXO1, V5-tagged CDK1 and cyclin B1 proteins. (d) ChIP analysis of the effect of CDK1 on the binding of FOXO1 to the Bim promoter. LNCaP cells were transfected with plasmids as indicated. At 48 h after transfection, cells were fixed, lysed and subjected to ChIP analysis as described in Materials and methods. Top: the PCR results amplified from input samples (Input). Bottom: the PCR amplifications obtained from ChIP samples (ChIP). The ChIP PCR result from lane 2 was normalized with the PCR result from the corresponding input sample and the normalized value was arbitrarily set as 1. The same normalization was performed for the rest of samples and further normalized with the normalized value from lane 2. These experiments were repeated at least twice. CDK, cyclin-dependent kinase; FOXO, O-class forkhead box; Pca, prostate cancer; FACS, fluorescence-activated cell sorting; S249, serine 249; ChIP, chromatin immunoprecipitation.

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4741 CDK1 inhibits FOXO1-induced apoptosis in PCa cells Hu et al., 2004; Huang et al., 2005). In an effort of Similar to previousreports(Nakamura et al., 2000; understanding the mechanism by which FOXO1 activity Huang et al., 2005), ectopic expression of constitutively isregulated, we demonstratedpreviouslythat CDK2 active FOXO1-A3 induced robust apoptosis in LNCaP interactswith and inhibitsFOXO1 via phosphorylation cells(Figure 6a, middle panel on the left). Thisis of FOXO1 at S249 (Huang et al., 2006). In thisstudy, presumably because LNCaP cells are PTEN-null and we provide evidence that CDK1 also interacts with and very sensitive to cell death induced by Akt-antagonistic specifically phosphorylates FOXO1 at S249, and signaling pathways (Nakamura et al., 2000). A similar accordingly, inhibitsthe transcriptional activity of result was obtained when cells were transfected with a FOXO1. In addition, we demonstrate an inhibitory role double mutant (FOXO1-A3/S249A) in which both Akt of FOXO1 in mitosis, which is overcome by CDK1 and CDK1 phosphorylation sites are mutated phosphorylation of FOXO1 at S249. We further (Figure 6a, lower panel on the left). Importantly, demonstrate that CDK1 phosphorylation of FOXO1 FOXO1-A3-induced cell death wasdiminishedby not only abolishes FOXO1-induced cell death but also ectopic expression of constitutively active CDK1-AF overridesFOXO1-mediated inhibition of growth and (Figure 6a, middle panel on the right). In contrast, survival of oncogene-transformed cells. Thus, inhibition FOXO1-A3/S249A-induced cell death wasnot affected of FOXO1 by CDK1 represents a novel mechanism that by CDK1-AF (Figure 6a, lower panel on the right). may play an important role in tumorigenesis. There waslittle or no effect of expressionof CDK1-AF Mouse embryos lacking all interphase CDKs (CDK2, alone on the death of LNCaP cellsin comparisonto CDK3, CDK4 and CDK6) undergo organogenesis and cellstransfected with an empty vector (Figure 6a, upper develop to midgestation (Santamaria et al., 2007). In panels). Results from three independent experiments contrast, embryos fail to develop to the morula and were quantified in Figure 6b. A small percentage of blastocyst stages in the absence of CDK1 (Santamaria apoptotic cell death wasdetected by FACS analysisin et al., 2007), indicating that CDK1 ablation leadsto a cellstransfected with an empty vector (Figure 6a, upper lethal phenotype. Analogously, generation of stable panel on the left), which wasfurther confirmed by the CDK1-depleted PCa cells has been unsuccessful (Chen cleavage of a small amount of poly (ADP-ribose) et al., 2006). Thus, CDK1 appears to be not only polymerase (PARP) proteins (Figure 6c, lane 1). This important for cell proliferation but also critical for cell background apoptotic cell death waslikely causedby survival. We demonstrate that silencing of endogenous the electroporation used in transfection. In line with the CDK1 results in a decrease in the phosphorylation of apoptotic results, ectopic expression of FOXO1-A3 FOXO1 at S249 and an increase in the transcriptional induced expression of Bim, a proapoptotic gene known activity of FOXO1 in PCa cells. It is likely that depletion to be regulated by FOXO proteins(Figure 6c). In of CDK1 leadsto cell death by activating the contrast, this effect was inhibited by expression of proapoptotic function of FOXO1. Further studies on CDK1-AF (Figure 6c). However, no inhibitory effect of the role of CDK1 phosphorylation of FOXO1 in the CDK1-AF on the expression of Bim was observed in survival of CDK1-null embryo or CDK1-null cells are cellstransfected with FOXO1-A3/S249A (Figure 6c). warranted. Moreover, chromatin immunoprecipitation (ChIP) Studieswith clinical cancer specimensimplicate that assays demonstrated that the binding of FOXO1-A3 CDK1 may play an important role in human cancers. but not FOXO1-A3/S249A to the promoter of the Bim Aberrant expression of CDK1 and cyclin B1 has been gene waslargely inhibited in cellstransfected with described in a number of primary tumors, including PCa CDK1-AF and cyclin B1 (Figure 6d). Finally, the long- (Kallakury et al., 1997; Soria et al., 2000). Several term effect of CDK1 phosphorylation of FOXO1 on cell groups have shown previously that expression of CDK1 growth and survival was evident in colony formation is increased in human PCa relative to normal prostatic assays (Supplementary Information). Together, these tissues (Mashal et al., 1996; Kallakury et al., 1997, data suggest that overexpressed CDK1 blocks FOXO1- 1998). Cdc25C, an upstream activator of CDK1, is induced death of LNCaP PCa cellsvia phosphorylation found not only upregulated in PCa but is also present of FOXO1 at S249. exclusively in its active form (Ozen and Ittmann, 2005). Cdc25B, another activator of CDK1, is overexpressed in PCa. Moreover, loss-of-function mutations of the Chk2 (also known Chek2) gene, an upstream negative Discussion regulator of CDK1, have been detected in PCa patients (Dong et al., 2003; Wu et al., 2006). In addition, cyclins Genetic and biochemical studies demonstrate that B1 and B2 are found overexpressed in the progression of FOXO proteins possess a tumor suppressor function PCa (Maddison et al., 2004; Stanbrough et al., 2006). that appearsto be evolutionally conserved(Huang et al., We demonstrate in this study that overexpression of 2006; Pinkston et al., 2006; Matsuoka et al., 2007; Paik CDK1 and cyclin B1 markedly inhibitsthe transcrip- et al., 2007; Pinkston-Gosse and Kenyon, 2007). The tional activity of FOXO1. Accordingly, the tumor importance of FOXO proteinsin human cancer is suppressor functions of FOXO1, which include the further evident in the fact that the functionsof these inhibitory effect of FOXO1 on the G2-to-M transition, proteinsare often inhibited by oncogenic proteins apoptosis and growth of the Ras-transformed cells, were such as Akt, IkB kinase and Skp2 (Brunet et al., 1999; abrogated by overexpression of CDK1 and cyclin B1. In

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4742 contrast, no inhibitory effect of CDK1 and cyclin B1 luciferase reporter construct, 3xIRS, which contains three was observed on the FOXO1 mutant that is resistant to copies of the FOXO response element in the promoter of the S249 phosphorylation. Thus, our findings suggest that IGFBP1 gene, were kindly provided by KL Guan (Tang et al., overexpression of CDK1 and cyclin B1 in human 1999). Plasmids for amino-acid substitution mutant of tumorsmay lead to the lossoffunction of FOXO1 FOXO1-A3/S249A were generated by PCR-based mutagenesis (Stratagene, La Jolla, CA, USA). The V5-tagged expression because of S249 phosphorylation, thereby promoting vector of CDK2 was described previously (Huang et al., 2006). cancer development and progression. Mammalian expression vectors of CDK1 and cyclin B1 were Increased activity of CDK1 in human cancers makes kindly provided by H Piwnica-Worms(Parker and Piwnica- it a potential target for cancer therapy. Currently, there Worms, 1992), and were subcloned into pcDNA3.1 (Invitrogen, are a number of ongoing clinical trialsusinginhibitors Carlsbad, CA, USA). The V5-tagged expression vector of of CDK1 and other CDKs, including flavopiridol, CDK1 wasgenerated in the pcDNA3.1D/V5-HisusingPCR. UCN-01, CYC202 (racemic roscovitine) and MMS- An active mutant of CDK1, CDK1-AF, wasgenerated by 387032 (Knockaert et al., 2002). More potent and converting Thr14 and Tyr15 to alanine and phenylalanine, selective small-molecule CDK1 inhibitors, such as respectively, via PCR-based mutagenesis (Stratagene). Expres- purvalanol, have been developed (Gray et al., 1998; sion vector for H-RasG12 V was kindly provided by R Urrutia (Fernandez-Zapico et al., 2003). Constructs for GST-FOXO1 Goga et al., 2007). Purvalanol treatment of - recombinant proteins were described previously (Huang et al., overexpressed tumors induces rapid apoptosis by redu- 2006). Amino-acid substitution mutants of GST-FOXO1 were cing the levels of the inhibitor of apoptosis survivin generated by PCR-based mutagenesis (Stratagene). Purified (Goga et al., 2007). It isworth noting that purvalanol recombinant GST-CDK1, GST-cyclin B and histone H1 were induces apoptosis in approximately 50% of cells, purchased from Cell Signaling Technology (Danvers, MA, whereas silencing of survivin leads to apoptosis in USA). The pEGFP vector was purchased from BD Biosciences approximately 20% of cells. This finding suggests that (Mountain View, CA, USA) (Clontech). A pool of siRNAs survivin-independent mechanisms also contribute to the for the human CDK1 and CDK2 gene and nonspecific cell death caused by CDK1 inhibition. Indeed, expres- control siRNAs were purchased from Dharmacon, Lafayette, sion of Bim, a known target gene of FOXO proteins CO, USA. (Dijkers et al., 2000; Stahl et al., 2002; Gilley et al., 2003), has been implicated in MYC-sensitized cell death Antibodies and chemicals (Green and Evan, 2002). We demonstrated previously A polyclonal antibody against phosphorylated S249 of that treatment with the CDK inhibitor roscovitine FOXO1 (S249-p) wasraisedby immunizing rabbitswith the increases the expression of Bim in LNCaP PCa cells phosphorylated peptide NPEGGKSGKpSPRRRAAS, and purified over a peptide-affinity column asdescribed(Huang (Huang et al., 2006). We demonstrate in this study that et al., 2006). Other protein-specific antibodies were as follows: forced expression of active FOXO1 results in an increase anti-FOXO1, anti-CDK1, anti-cyclinB1, anti-PARP, anti-Ras in the level of Bim protein and that overexpression of (Cell Signaling Technology); anti-CDK2 polyclonal (M2), CDK1 overridesFOXO1-induced expressionof Bim. anti-Gadd45a (H-165) polyclonal, anti-Erk2 monoclonal Thus, it would be of great value to assess the role, if any, (Santa Cruz Biotechnology, Carlsbad, CA, USA); anti-V5 of FOXO1 S249 phosphorylation in the anticancer (Invitrogen); anti-FLAG (M2) (Sigma, St Louis, MO, USA) effect of CDK inhibitorsthat are currently being and anti-phospho-histone H3 at serine 10 (H3-p) (Millipore, employed in clinical trials. Chicago, IL, USA). In vitro kinase assay buffer was purchased In summary, we demonstrate that CDK1 specifically from Cell Signaling Technology; crystal violet, RNase A, phosphorylates FOXO1 at S249, and accordingly, leads aphidicolin and propidium iodide were obtained from Sigma-Aldrich, St Louis, MO, USA. to the loss of transcriptional activity of FOXO1. We also demonstrate that CDK1 phosphorylation of FOXO1 not only abolishes FOXO1-induced delay in Cell culture, transfection, luciferase reporter assay and ChIP the G2-to-M transition but also mitigates FOXO1- assay induced apoptosis in PCa cells and long-term growth The PCa cell line LNCaP (purchased from the American Type and survival of fibroblasts transformed by oncogene Culture Collection, Manassas, VA, USA) was cultured in RPMI 1640 medium containing 10% fetal bovine serum, Ras. Given that CDK1 and cyclin B1 are often 100 U/ml penicillin and 100 mg/ml streptomycin. NIH 3T3 cells overexpressed in human cancers including PCa, were maintained in Dulbecco’smodified Eagle’smedium, CDK1-mediated inhibition of FOXO1 may very likely supplemented with 10% fetal bovine serum, 100 U/ml peni- contribute to tumorigenesis and therapy resistance in cillin and 100 mg/ml streptomycin. Cells were maintained at human cancer. 37 1C and 5% CO2. Transfections were performed by electroporation using an Electro Square Porator ECM 830 (BTX). Approximately 75–90% transfection efficiencies were routinely achieved. Luciferase reporter assays were performed Materials and methods asdescribed(Huang et al., 2006). ChIP assays were performed asdescribed(Huang et al., 2004b). The forward primer 50-CG Plasmids and siRNAs GTTCGCTGCGTTCC-30 and reverse primer 50-CGTC Plasmids for FLAG-tagged wild-type FOXO1 and Akt TCCGCTCGCACG-30 were designed based on the genomic phosphorylation-resistant mutant of FOXO1 (FOXO1-WT sequence of the promoter of human Bim gene, which flanksthe and FOXO1-A3, originally named asFKHR-WT and FKHR- forkhead response element, GTAAACAC, characterized AAA, where the three Akt phosphorylation sites Thr24, previously (Gilley et al., 2003). PCR amplification was Ser256 and Ser319 were converted to alanine) and the performed with 34 cyclesof 94 1C for 30 s, 62 1C for 30 sand

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4743 72 1C for 45 s. The PCR products were analysed by 2% 400 mg/ml. Two weeksafter selection,colonieswere fixed and agarose gels and stained with ethidium bromide. stained with crystal violet. Colonies with one hundred or more cellswere counted. Each group wasassayedin triplicate and Immunoprecipitation, western blot, immunofluorescence experimentswere repeated three times. chemistry and confocal microscopy Protein immunoprecipitationswere carried out usingan In vitro kinase assay immunoprecipitation kit (Roche Applied Science, Indianapolis, Kinase assays were carried out in the presence of [g-32P]ATP by IN, USA) asdescribed (Huang et al., 2006), and western blot using an in vitro kinase buffer system obtained from Cell wasperformed asdescribed(Huang et al., 2001). Immuno- Signaling Technology asdescribed(Huang et al., 2006). fluorescence chemistry and confocal microscopy were EndogenousCDK1 immunoprecipitated from LNCaP cells performed asdescribed(Huang et al., 2004a). or reconstituted CDK1 and cyclin B1 complexes obtained from Cell Signaling Technology were incubated with various Cell cycle analysis by flow cytometry substrates at 30 1C. Thirty minutesafter reaction, proteins NIH 3T3 cells were transfected or co-transfected with were subjected to SDS–polyacrylamide gel electrophoresis and indicated plasmids and seeded in dishes with the medium autography. containing 1 mg/ml aphidicolin. Twenty-four hoursafter incubation, cellswere releasedfrom aphidicolin-mediated S-phase blockage by washing cells with 1 Â phosphate- Statistics buffered saline. At different time points, cells were collected Experimentswere carried out with three or four replicates. and washed with 1 Â phosphate-buffered saline. After Statistical analyses were performed by Student’s t-test. Values fixation with 70% ethanol, cellswere washedtwice with of Po0.05 are considered significant. 1 Â phosphate-buffered saline and stained with a solution containing 20 mg/ml propidium iodide and 50 mg/ml RNase A. Acknowledgements Cellswere incubated for 30 min at the room temperature and cell cycle profileswere determined by flow cytometry usinga We thank DJ Tindall, KL Guan, H Piwnica-Worms, FACScan (Becton-Dickinson, San Jose, CA, USA). R Urrutia and F Farassati for plasmids and reagents, members of the Huang Lab for their stimulating discussion, V Shridhar Colony formation assay and K Schwertfeger for their critical reading of the manuscript Colony forming assays were performed as described and Jacie Maguire for her assistance in the preparation of the (Fernandez-Zapico et al., 2003). Briefly, NIH 3T3 cellswere manuscript. This work was supported in part by funds from transfected with the indicated plasmids. Approximately 1000 the Department of Defense, USA (W81XWH-07-1-0137), and cells were seeded in 100-mm tissue culture dishes. Transfected the Brainstorm Award from the University of Minnesota cellswere selectedby adding G418 at a final concentration of Cancer Center.

References

Aoki M, Jiang H, Vogt PK. (2004). Proteasomal degradation of the transcriptional activity of KLF11 with its role in growth regulation. FoxO1 transcriptional regulator in cells transformed by the P3k and EMBO J 22: 4748–4758. Akt oncoproteins. Proc Natl Acad Sci USA 101: 13613–13617. Furukawa-Hibi Y, Yoshida-Araki K, Ohta T, Ikeda K, Motoyama N. BiggsIII WH, Meisenhelder J, Hunter T, Cavenee WK, Arden KC. (2002). FOXO forkhead transcription factors induce G(2)–M (1999). Protein kinase B/Akt-mediated phosphorylation promotes checkpoint in response to oxidative stress. J Biol Chem 277: nuclear exclusion of the winged helix transcription factor FKHR1. 26729–26732. Proc Natl Acad Sci USA 96: 7421–7426. Galili N, DavisRJ, FredericksWJ, Mukhopadhyay S, Rauscher III Borkhardt A, Repp R, HaasOA, LeisT, Harbott J, Kreuder J et al. FJ, Emanuel BS et al. (1993). Fusion of a fork head domain gene to (1997). Cloning and characterization of AFX, the gene that fuses PAX3 in the solid tumour alveolar rhabdomyosarcoma. Nat Genet to MLL in acute leukemiaswith a t(X;11)(q13;q23). Oncogene 14: 5: 230–235. 195–202. Gilley J, Coffer PJ, Ham J. (2003). FOXO transcription factors Brunet A, Bonni A, Zigmond MJ, Lin MZ, Juo P, Hu LS et al. (1999). directly activate bim and promote apoptosis in Akt promotes cell survival by phosphorylating and inhibiting a sympathetic neurons. J Cell Biol 162: 613–622. forkhead transcription factor. Cell 96: 857–868. Goga A, Yang D, Tward AD, Morgan DO, Bishop JM. (2007). Brunet A, Park J, Tran H, Hu LS, HemmingsBA, Greenberg ME. Inhibition of CDK1 asa potential therapy for tumorsover- (2001). Protein kinase SGK mediates survival signals by phosphory- expressing MYC. Nat Med 13: 820–827. lating the forkhead transcription factor FKHRL1 (FOXO3a). Mol Gray NS, Wodicka L, Thunnissen AM, Norman TC, Kwon S, Cell Biol 21: 952–965. Espinoza FH et al. (1998). Exploiting chemical libraries, structure, Chen S, Xu Y, Yuan X, Bubley GJ, Balk SP. (2006). Androgen and genomics in the search for kinase inhibitors. Science 281: 533–538. phosphorylation and stabilization in prostate cancer Green DR, Evan GI. (2002). A matter of life and death. Cancer Cell 1: by cyclin-dependent kinase 1. Proc Natl Acad Sci USA 103: 19–30. 15969–15974. Greer EL, Brunet A. (2005). FOXO transcription factors at the DijkersPF, Medema RH, LammersJW, Koenderman L, Coffer PJ. interface between longevity and tumor suppression. Oncogene 24: (2000). Expression of the pro-apoptotic Bcl-2 family member Bim is 7410–7425. regulated by the forkhead transcription factor FKHR-L1. Curr Biol Hillion J, Le Coniat M, Jonveaux P, Berger R, Bernard OA. (1997). 10: 1201–1204. AF6q21, a novel partner of the MLL gene in t(6;11)(q21;q23), Dong X, Wang L, Taniguchi K, Wang X, Cunningham JM, defines a forkhead transcriptional factor subfamily. Blood 90: McDonnell SK et al. (2003). Mutations in CHEK2 associated with 3714–3719. prostate cancer risk. Am J Hum Genet 72: 270–280. Hu MC, Lee DF, Xia W, Golfman LS, Ou-Yang F, Yang JY et al. Fernandez-Zapico ME, Mladek A, Ellenrieder V, Folch-Puy E, Miller (2004). IkappaB kinase promotes tumorigenesis through inhibition L, Urrutia R. (2003). An mSin3A interaction domain linksthe of forkhead FOXO3a. Cell 117: 225–237.

Oncogene Inhibition of FOXO1 by CDK1 PLiuet al 4744 Huang H, Cheville JC, Pan Y, Roche PC, Schmidt LJ, Tindall DJ. Ozen M, Ittmann M. (2005). Increased expression and activity of (2001). PTEN induces chemosensitivity in PTEN-mutated prostate CDC25C phosphatase and an alternatively spliced variant in cancer cells by suppression of Bcl-2 expression. J Biol Chem 276: prostate cancer. Clin Cancer Res 11: 4701–4706. 38830–38836. Paik JH, Kollipara R, Chu G, Ji H, Xiao Y, Ding Z et al. (2007). Huang H, Muddiman DC, Tindall DJ. (2004a). Androgensnegatively FoxOs are lineage-restricted redundant tumor suppressors and regulate forkhead transcription factor FKHR (FOXO1) through a regulate endothelial cell homeostasis. Cell 128: 309–323. proteolytic mechanism in prostate cancer cells. J Biol Chem 279: Parker LL, Piwnica-WormsH. (1992). Inactivation of the p34cdc2– 13866–13877. cyclin B complex by the human WEE1 tyrosine kinase. Science 257: Huang H, Regan KM, Lou Z, Chen J, Tindall DJ. (2006). CDK2- 1955–1957. dependent phosphorylation of FOXO1 as an apoptotic response to Pinkston-Gosse J, Kenyon C. (2007). DAF-16/FOXO targets genes DNA damage. Science 314: 294–297. that regulate tumor growth in Caenorhabditis elegans. Nat Genet 39: Huang H, Regan KM, Wang F, Wang D, Smith DI, van Deursen JM 1403–1409. et al. (2005). Skp2 inhibits FOXO1 in tumor suppression through Pinkston JM, Garigan D, Hansen M, Kenyon C. (2006). Mutations ubiquitin-mediated degradation. Proc Natl Acad Sci USA 102: that increase the life span of C.elegans inhibit tumor growth. Science 1649–1654. 313: 971–975. Huang H, Tindall DJ. (2007). Dynamic FoxO transcription factors. PlasDR, ThompsonCB. (2003). Akt activation promotesdegradation of J Cell Sci 120: 2479–2487. tuberin and FOXO3a via the proteasome. JBiolChem278: 12361–12366. Huang H, Zegarra-Moro OL, Benson D, Tindall DJ. (2004b). Ramaswamy S, Nakamura N, Sansal I, Bergeron L, Sellers WR. Androgens repress Bcl-2 expression via activation of the retino- (2002). A novel mechanism of gene regulation and tumor suppres- blastoma (RB) protein in prostate cancer cells. Oncogene 23: sion by the transcription factor FKHR. Cancer Cell 2: 81–91. 2161–2176. Rena G, Guo S, Cichy SC, Unterman TG, Cohen P. (1999). Kallakury BV, Sheehan CE, AmbrosRA, FisherHA, Kaufman Jr RP, Phosphorylation of the transcription factor forkhead family member Muraca PJ et al. (1998). Correlation of p34cdc2 cyclin-dependent FKHR by protein kinase B. J Biol Chem 274: 17179–17183. kinase overexpression, CD44s downregulation, and HER-2/neu Rena G, WoodsYL, PrescottAR, Peggie M, Unterman TG, Williams oncogene amplification with recurrence in prostatic adenocarcino- MR et al. (2002). Two novel phosphorylation sites on FKHR that mas. J Clin Oncol 16: 1302–1309. are critical for itsnuclear exclusion. EMBO J 21: 2263–2271. Kallakury BV, Sheehan CE, AmbrosRA, FisherHA, Kaufman Jr RP, Santamaria D, Barriere C, Cerqueira A, Hunt S, Tardy C, Newton K Ross JS. (1997). The prognostic significance of p34cdc2 and et al. (2007). Cdk1 issufficientto drive the mammalian cell cycle. cyclin D1 protein expression in prostate adenocarcinoma. Cancer Nature 448: 811–815. 80: 753–763. Schmidt M, Fernandez de MattosS, van der HorstA, Klompmaker R, Knockaert M, Greengard P, Meijer L. (2002). Pharmacological KopsGJ, Lam EW et al. (2002). Cell cycle inhibition by FoxO inhibitors of cyclin-dependent kinases. Trends Pharmacol Sci 23: forkhead transcription factors involves downregulation of cyclin D. 417–425. Mol Cell Biol 22: 7842–7852. KopsGJ, De Ruiter ND, De Vries-SmitsAM,Powell DR, BosJL, Seoane J, Le HV, Shen L, Anderson SA, Massague J. (2004). Integration Burgering BM. (1999). Direct control of the forkhead transcription of Smad and forkhead pathwaysin the control of neuroepithelial and factor AFX by protein kinase B. Nature 398: 630–634. glioblastoma cell proliferation. Cell 117: 211–223. Kops GJ, Medema RH, Glassford J, Essers MA, Dijkers PF, Coffer PJ Sherr CJ, RobertsJM. (2004). Living with or without cyclinsand et al. (2002). Control of cell cycle exit and entry by protein kinase cyclin-dependent kinases. Genes Dev 18: 2699–2711. B-regulated forkhead transcription factors. Mol Cell Biol 22: Soria JC, Jang SJ, Khuri FR, Hassan K, Liu D, Hong WK et al. 2025–2036. (2000). Overexpression of cyclin B1 in early-stage non-small cell lung Maddison LA, Huss WJ, Barrios RM, Greenberg NM. (2004). cancer and itsclinical implication. Cancer Res 60: 4000–4004. Differential expression of cell cycle regulatory molecules and Stahl M, DijkersPF, KopsGJ, LensSM, Coffer PJ, Burgering BM evidence for a ‘cyclin switch’ during progression of prostate cancer. et al. (2002). The forkhead transcription factor FoxO regulates Prostate 58: 335–344. transcription of p27Kip1 and Bim in response to IL-2. J Immunol Mashal RD, Lester S, Corless C, Richie JP, Chandra R, Propert KJ 168: 5024–5031. et al. (1996). Expression of cell cycle-regulated proteins in prostate Stanbrough M, Bubley GJ, Ross K, Golub TR, Rubin MA, Penning cancer. Cancer Res 56: 4159–4163. TM et al. (2006). Increased expression of genes converting adrenal Matsuoka S, Ballif BA, Smogorzewska A, McDonald III ER, androgens to testosterone in androgen-independent prostate cancer. Hurov KE, Luo J et al. (2007). ATM and ATR substrate analysis Cancer Res 66: 2815–2825. reveals extensive protein networks responsive to DNA damage. Tang ED, Nunez G, Barr FG, Guan KL. (1999). Negative regulation Science 316: 1160–1166. of the forkhead transcription factor FKHR by Akt. J Biol Chem Matsuzaki H, Daitoku H, Hatta M, Tanaka K, Fukamizu A. (2003). 274: 16741–16746. Insulin-induced phosphorylation of FKHR (Foxo1) targets Tran H, Brunet A, Grenier JM, Datta SR, Fornace Jr AJ, DiStefano to proteasomal degradation. Proc Natl Acad Sci USA 100: PS et al. (2002). DNA repair pathway stimulated by the forkhead 11285–11290. transcription factor FOXO3a through the Gadd45 protein. Science Medema RH, KopsGJ, BosJL, Burgering BM. (2000). AFX-like 296: 530–534. forkhead transcription factors mediate cell-cycle regulation by Ras WoodsYL, Rena G, Morrice N, Barthel A, Becker W, Guo S et al. and PKB through p27kip1. Nature 404: 782–787. (2001). The kinase DYRK1A phosphorylates the transcription Modur V, Nagarajan R, EversBM, Milbrandt J. (2002). FOXO factor FKHR at Ser329 in vitro, a novel in vivo phosphorylation site. proteins regulate tumor necrosis factor-related apoptosis inducing Biochem J 355: 597–607. ligand expression. Implications for PTEN mutation in prostate Wu X, Dong X, Liu W, Chen J. (2006). Characterization of CHEK2 cancer. J Biol Chem 277: 47928–47937. mutationsin prostatecancer. Hum Mutat 27: 742–747. Nakamura N, Ramaswamy S, Vazquez F, Signoretti S, Loda M, Wu X, Senechal K, Neshat MS, Whang YE, Sawyers CL. (1998). The SellersWR. (2000). Forkhead transcription factorsare critical PTEN/MMAC1 tumor suppressor phosphatase functions as a effectors of cell death and cell cycle arrest downstream of PTEN. negative regulator of the phosphoinositide 3-kinase/Akt pathway. Mol Cell Biol 20: 8969–8982. Proc Natl Acad Sci USA 95: 15587–15591.

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

Oncogene