Calcineurin Regulation of the Mammalian G0/G1 Checkpoint Element, Cyclin Dependent Kinase 4

Calcineurin regulation of the mammalian G0/G1 checkpoint element, cyclin dependent kinase 4

Shairaz Baksh1, James A DeCaprio2 and Steven J Burako*,3,4

1Department of Pediatric Oncology, Dana Farber Cancer Institute, 44 Binney Street, Boston, Massachusetts, MA 02115, USA; 2Department of Adult Oncology, Dana Farber Cancer Institute, Boston, Massachusetts, MA 02115, USA; 3Department of Pediatrics, Harvard Medical School, Boston, Massachusetts, MA 02115, USA; 4Department of Medicine, Harvard Medical School, Boston, Massachusetts, MA. 02115, USA

Cyclin dependent kinase 4 (cdk4) activity is controlled by The activation of calcineurin and subsequent dephos- the binding of regulatory subunits and inhibitory factors, phorylation of its substrate, NFAT, is central to as well as tyrosine and serine/threonine phosphorylation. lymphokine activation (Northrop et al., 1994). Dephos- More recently the in¯uence of calcium levels have been phorylation of NFAT is required for its translocation to demonstrated. Using transient transfections in Jurkat the nucleus (in a complex with calcineurin) (Shibasaki et cells, we observed speci®c binding between cdk4 and the al., 1996), whereby NFAT forms heterodimers with AP-1 calcium and calmodulin activated serine/threonine phos- family members when binding to NFAT/AP-1 sites in phatase, calcineurin. Furthermore, we demonstrated that the interleukin (IL)-2 promoter (Chen et al., 1998). In the inhibition of the phosphatase activity of calcineurin addition to regulating the production of IL-2, NFAT with FK506 and cyclosporin A resulted in an overall family members also regulate numerous other lympho- increase in cdk4 kinase activity, suggesting that the kines including IL-4, IL-3, GM-CSF, and IFN-g phosphatase activity of calcineurin was inhibitory to the (reviewed in Rao et al., 1997). Inhibition of the kinase activity of cdk4. In contrast, we were not able to phosphatase activity of calcineurin is obtained by the observe a similar eect on the kinase activity of either binding of the fungal macrolides, FK506 and cyclosporin cdk6 or cdk2, indicating that the phosphatase activity of A (CsA). Speci®cally, FK506 and CsA bind to calcineurin was speci®c for cdk4. In addition, using an in endogenous proteins, the immunophilins ± the FK506 vitro phosphatase assay for calcineurin, we observed that binding proteins (FKBPs) and the cyclosporin A binding the exogenous addition of calcineurin resulted in the proteins (the cyclophilins ± CyPs) (Fruman et al., 1994). dephosphorylation of cdk4, an event that downregulated FK506, CsA, or the immunophilins by themselves do not the kinase activity of cdk4. Calcineurin could, therefore, inhibit the phosphatase activity of calcineurin. However, play an opposing role to the action of the cyclin as a complex of FK506/FKBP or CsA/CyP, they can activating kinase complex, an enzyme that upregulates inhibit the activity of calcineurin by sterically blocking the kinase activity of cdk4, an important G0/G1 the access of substrates to the phosphatase catalytic site checkpoint element in mammalian cells. Oncogene (Grith et al., 1995). (2000) 19, 2820 ± 2827. Cyclin dependent kinase 4 (cdk4) is a central element controlling cellular proliferation. Its activity is tightly Keywords: cell cycle; calcineurin; cdk4; cyclin activat- controlled by several mechanisms that include binding ing kinase; Jurkat cells of inhibitory proteins (such as the p16INK4 family members) (Sherr and Roberts, 1995), association with cyclin subunits, and serine/threonine and tyrosine Introduction phosphorylation (Morgan, 1995). All cyclin dependent kinase family members are regulated by an activating Lymphocyte activation involves the interaction of phosphorylation event (mediated by the cyclin activat- calcineurin with the nuclear factor of activated T cells, ing kinase or CAK) on threonine-172 for cdk4, NFAT (Klee et al., 1998). Calcineurin is a calcium threonine-176 for cdk6, and threonine-161 for cdc2 activated serine/threonine phosphatase that exists as a and cdk2 (Morgan, 1995). This active phosphorylation, heterodimer, formed by the association of a 60 kDa combined with binding of a cyclin counterpart (cyclin phosphatase catalytic subunit (calcineurin A) and a D for cdk4), allows the active form of cdk4 to 19 kDa high anity calcium binding subunit (calci- translocate to the nucleus where it can phosphorylate neurin B). Activation of calcineurin involves the target proteins and promote cell cycle entry. In the case binding of calcium bound-calmodulin (CaM) to the of cdk4, the main target protein is the p110 protein carboxy terminus of calcineurin A, resulting in the product of the retinoblastoma (Rb) tumor suppressor displacement of an autoinhibitory domain from the gene (Knudsen and Wang, 1996). In the quiescent phosphatase catalytic active site. These events then state, Rb associates with the transcription factor E2F allow calcineurin substrates to gain access to the and represses E2F-dependent gene transcription of catalytic site and undergo dephosphorylation. important S phase genes. Cyclin D/cdk4-dependent phosphorylation of Rb during late G0/G1 releases E2F, allowing E2F to activate E2F responsive elements on speci®c S phase genes, such as DNA polymerase II, *Correspondence: SJ Burako Received 21 January 2000; revised 16 March 2000; accepted thymidine kinase, and cyclin E (Weinberg, 1995). 22 March 2000 When cells are ready to return to the quiescent state, Calcineurin regulation of cdk4 S Baksh et al 2821 Rb becomes sequentially dephosphorylated by serine/ threonine phosphatase(s) and by the downregulation of cdk activity. The role played by calcium in the proliferation of eukaryotic cells has been mainly stimulated by eorts to understand how the high anity calcium binding protein, calmodulin, interacts with proteins involved in cell cycle control. We present data that suggest a role for the calcium and calmodulin regulated serine/ threonine phosphatase, calcineurin, in the control of the mammalian G0/G1 checkpoint kinase, cdk4. We observed a physical interaction with cdk4, a decrease in cdk4 kinase activity in the presence of active calcineurin, as well as an enhancement of the activity of cdk4 in the presence of calcineurin inhibiting drugs, FK506 and CsA. Thus, calcineurin may play an important role in the control of cellular proliferation in T lymphocytes by the regulation of cdk4 kinase activity.

Results

Calcineurin physically interacts with cdk4 In order to investigate potential substrates for calcineurin, we prepared whole cell lysates from an asynchronous population of a human leukemic T cell line, Jurkat, and precipitated calcineurin (Cn) binding proteins using a CaM/Cn anity matrix. Immunoblot- ting with an anti cdk4 antibody, we demonstrated that cdk4 speci®cally associated with the matrix of CaM/ Figure 1 Calcineurin can physically associate with cdk4. (a) Cn-Sepharose (Figure 1a, lane 3) and not with a Asynchronously growing Jurkat cells were lysed in DIP buer Sepharose matrix or with CaM-Sepharose devoid of and allowed to bind to either Sepharose beads (lane 1), CaM exogenously added calcineurin (Figure 1a, lanes 1 and Sepharose beads (lane 2), CaM/Cn Sepharose beads (lane 3), or CaM/Cn Sepharose beads that were preloaded with FK506/ 2 respectively). Preincubation of CaM/Cn-Sepharose FKBP12 (lane 4). Bound proteins were immunoblotted with a with the complex of FK506/FKBP12 dramatically rabbit anti-cdk4 antibody. (b) Jurkat cells were transiently inhibited the amount of cdk4 bound to CaM/Cn transfected with: pEBG, vector control (lane 1); GST-tagged wild Sepharose (Figure 1a, lane 4), suggesting a speci®c type calcineurin A (lane 2); GST-tagged constitutively active interaction of calcineurin with cdk4. Furthermore, calcineurin A (lane 3); GST-tagged phosphatase dead calcineurin A containing the H151A substitution (lane 4). All samples were under similar experimental conditions, neither cdk6 co-transfected with calcineurin B. The cells were harvested 19 h nor cdk2 bound to the CaM/Cn-Sepharose (data not later, lysed in DIP buer with 10 mM calcium added, immuno- shown), suggesting that calcineurin may have speci®city precipitated (IP) overnight, separated using 10% SDS ± PAGE, for cdk4. and then immunoblotted (IB) with a rabbit anti-GST antibody. To further de®ne the interaction of calcineurin with (b) is an example of one of three separate experiments cdk4, immunoprecipitation studies were performed using Jurkat cells transiently transfected with either: (1) wild type calcineurin A (CnA), (2) a constitutively immunoprecipitate the point mutant of calcineurin A active form of calcineurin A (containing a deletion of (Figure 1b, lane 4) suggests that, although devoid of the carboxy terminal autoinhibitory and CaM-binding catalytic activity, this mutant can still associate with domains and thus in the open conformation containing cdk4 and may possibly behave as a trapping mutant calcium-independent catalytic activity) (CnAd), or (3) a for calcineurin substrates (Zhu and McKeon, 1999). phosphatase dead mutant of calcineurin A (CnAd- H151A) (Figure 1b). CnAdH151A is also missing the Calcineurin activity inhibits the kinase activity of cdk4 carboxy terminal autoinhibitory and CaM-binding domains, but contained a histidine to alanine substitu- In order to determine whether calcineurin binding to tion in a critical residue important for stabilizing the cdk4 in¯uenced cdk4 kinase activity, we assessed the phosphate group and thus lacked catalytic activity kinase activity of cdk4 in an in vitro kinase assay using, (Grith et al., 1995; Shibasaki et al., 1996). Immuno- as a substrate, the carboxy terminal residues of the precipitation using an anti-cdk4 antibody demonstrated p110 protein product of Rb ligated to glutathione S- that the interaction between calcineurin and cdk4 transferase (GST-RbC) (Matsushime et al., 1994). preferentially occurred with the form of calcineurin in Lysates were prepared from Jurkat cells transiently the open conformation (CnAd and CnAdH151A) transfected with the indicated vectors and cdk4 kinase (Figure 1b, lanes 3 and 4), even though all forms of activity was assessed. The expression of active calcineurin A were expressed at comparable levels calcineurin resulted in a reduction of cdk4 kinase (Figure 1b, bottom panel, lanes 2 ± 4). Interestingly, the activity by 50% (Figure 2a, lane 3; Figure 2b), that was observation that cdk4 antibodies were able to reversed by the in vivo addition of the calcineurin

Oncogene Calcineurin regulation of cdk4 S Baksh et al 2822

Figure 2 Calcineurin inhibition of the kinase activity of cdk4. (a) Jurkat cells were transiently transfected with: pEBG (lane 1); GST-tagged wild type calcineurin A (lane 2); GST-tagged constitutively active calcineurin A (lanes 3 ± 5), treated with either 100 nM FK506 (lane 4) or with 5 mM CsA (lane 5) for 16 h. The cells were then harvested 19 h post-transfection, lysed and normalized to total protein content, and immunoprecipitated with a rabbit anti-cdk4 antibody and an IVK assay was carried out. (a) An example of one of four separate experiments performed and (b) is the quantitation of four separate cdk4 IVK assays using GST-RbC as the substrate. (c) The activity of calcineurin under the conditions of FK506 and CsA treatment was assessed using a peptide from the regulatory subunit of protein kinase A as the substrate (Fruman et al., 1992). The data is a representation of four separate experiments. (d) Jurkat cells were transiently transfected with the GST-tagged constitutively active calcineurin A (lane 1) and GST- tagged phosphatase dead calcineurin A (lane 2). Cdk4 IVK analysis was carried out as outlined in (a). WCL, whole cell lysate blotted with a rabbit anti-cdk4 antibody

inhibitors, FK506 or CsA (Figure 2a, lanes 4 and 5; Interestingly, the inhibition of the kinase activity of Figure 2b). Cdk4 kinase activity was enhanced by 45% cdk4 by active calcineurin (Figure 2d, lane 1) was above vector in the presence FK506 or CsA, suggesting reversed by the presence of the phosphatase dead form that the phosphatase activity of calcineurin could of calcineurin (CnAdH151A) (Figure 2d, lane 2) regulate the kinase activity of cdk4, and that FK506 further supporting the role of the phosphatase activity and CsA were present in amounts sucient to inhibit of calcineurin in the regulation of the kinase activity of the phosphatase activity of both the transfected and cdk4. The dramatic enhancement of the kinase activity endogenous pools of calcineurin. Both FK506 and CsA of cdk4 in the presence of the point mutant of inhibited the phosphatase activity of calcineurin by 10- calcineurin may have resulted from the point mutant fold under similar experimental conditions (Figure 2c); associating with the pool of endogenous calcineurin B and the amount of wild type, active calcineurin and and hence inhibiting the function of endogenous cdk4 were equivalent. calcineurin. Alternatively, the point mutant of calci-

Oncogene Calcineurin regulation of cdk4 S Baksh et al 2823 neurin may compete with endogenous calcineurin for CaM-bound calcineurin, we observed a 70% reduction cdk4 binding and thus prevent the dephosphorylation in the kinase activity of cdk4 towards GST-RbC of cdk4, resulting in a greater proportion of the active (Figure 4a, lane 3). The loss of cdk4 kinase activity kinase. towards GST-RbC, as a result of the in vitro addition of calcineurin (Figure 4a, lane 3), would suggest a role for calcineurin in the regulation of threonine-172 Calcineurin can directly regulate the phosphorylation level phosphorylation ± a site regulated by CAK. We of cdk4 further tested this hypothesis in an in vitro kinase assay Cdk4 is known to undergo two distinct phosphoryla- for CAK phosphorylation of cdk4. CAK obtained tion events ± one that leads to maximal kinase activity from an asynchronous population of NIH3T3 cells (phosphorylation of threonine-172 by CAK) (Kato et phosphorylated GST-cdk4 (Figure 4b, lane 1). How- al., 1994); and one that leads to the inhibition of its ever, the addition of exogenous CaM-bound calcineur- kinase activity (phosphorylation of tyrosine-17 and in to the CAK phosphorylated GST-cdk4 resulted in a threonine-19) (Jinno et al., 1999). An asynchronous 30% reduction of the intensity of phosphorylated GST- population of COS-7 cells was transfected with either cdk4 (Figure 4b, lane 2) and this reduction was vector control (pEBG), wild type, constitutively active reversed by the addition of EDTA/EGTA, a known and phosphatase dead forms of calcineurin and cdk4 inhibitor of calmodulin activation of calcineurin was immunoprecipitated with a phosphospeci®c anti- (Figure 4b, lane 3). Therefore, the loss of calcineurin body for phosphorylated threonine-172 (PT101). Over- activity in the presence of EDTA/EGTA resulted in the expression of the constitutively active form of lack of dephosphorylation of GST-cdk4, and hence a calcineurin resulted in a loss of the ability of PT101 sustained phosphorylation of GST-cdk4 (Figure 4b, to immunoprecipitate cdk4 (Figure 3, lane 3); however, lane 3). Thus, calcineurin may act in direct opposition in the presence of the phosphatase dead form of to the action of CAK and regulate residue(s) targeted calcineurin, cdk4 remained phosphorylated and PT101 by CAK. regained its ability to immunoprecipitate cdk4 (Figure 3, lane 4). PT101 was speci®c for phosphorylated cdk4 The phosphatase activity of calcineurin is specific for cdk4 and we were not able to immunoprecipitate phosphorylated cdk6 under simliar transfection conditions To assess the selectivity of the phosphatase activity of (data not shown). calcineurin for cdk4, we transiently transfected Jurkat In order to further assess the direct regulation of cells with calcineurin, immunoprecipitated cdk6 and cdk4, an asynchronous population of Jurkat cells was cdk2 and carried out in vitro kinase assays using GST- treated with FK506 for 16 h, lysed and cdk4 complexes RbC as the substrate for cdk6 (Figure 4c) and histone were immunoprecipitated and associated kinase activity H1 as the substrate for cdk2 (data not shown). was assessed (Figure 4a). As observed for transiently Calcineurin did not change the kinase activity of transfected Jurkat cells, inhibition of the phosphatase cdk6 or cdk2, and the inhibition of calcineurin by activity of calcineurin by FK506 resulted in the FK506 did not aect the kinase activity of either cdk6 enhancement of the kinase activity of cdk4 (Figure towards GST-RbC (compare Figure 4c, lanes 3 and 4 4a; compare lane 2 to lane 1), supporting a role for with Figure 2a, lanes 3 and 4) or cdk2 towards histone calcineurin in the downregulation of the kinase activity H1 (data not shown). In addition, we looked at the of cdk4. Furthermore, when the immunoprecipitated ability of calcineurin to directly dephosphorylate active cdk4 complex from an asynchronous population immunoprecipitated cdk6 and cdk2 from an asynchro- of cells was in vitro dephosphorylated using puri®ed nous population of cells using an in vitro dephosphorylation assay with puri®ed CaM-bound calcineurin (Figure 4d). We did not observe a reduction in the kinase activity of either cdk2 or cdk6 as was observed for cdk4 (compare Figure 4d, lanes 2 and 3 with Figure 4a, lanes 2 and 3).

Calcineurin preferentially downregulates the kinase activity of cdk4 during mitosis To determine if calcineurin could modulate cdk4 kinase activity in a cell cycle dependent manner, Jurkat cells were synchronized at four distinct points of the cell cycle using 0.5% fetal calf serum (G0/G1 phase), 5 mg/ml aphidicolin (late G1 phase), 750 mM hydroxyurea (S phase), or 500 ng/ml nocodozole (M Figure 3 Active calcineurin can downmodulate the phosphor- phase). Cell cycle arrest experiments carried out in the ylation level of cdk4. COS-7 cells were transiently transfected presence and absence of FK506 revealed that the with: pEBG (lane 1); GST-tagged wild type calcineurin A (lane 2); GST-tagged constitutively active calcineurin A (lane 3); GST- kinase activity of cdk4 towards GST-RbC was tagged phosphatase dead calcineurin A (lane 4) and immunopre- increased primarily during the nocodazole synchro- cipitated with PT101 (a phosphospeci®c antibody for phospho- nized phase or M phase of the cell cycle having 40% threonine-proline sequences ± TpP). All lanes were also more activity when the phosphatase activity of cotransfected with pCMV-HA-cdk4 containing HA-tagged hu- calcineurin was inhibited by FK506 (Figure 5a). The man cdk4. Top panel: immunoprecipitated sample (IP); bottom panel whole cell lysate (WCL); IB: immunoblot. The data is a hydroxyurea synchronized phase or S phase kinase representation of one of three separate experiments activity of cdk4 was also elevated by FK506 treatment

Oncogene Calcineurin regulation of cdk4 S Baksh et al 2824

Figure 4 Calcineurin phosphatase activity is speci®c for cdk4. (a) Asynchronously growing Jurkat cells were left untreated (lane 1) or treated with FK506 for 16 h (lane 2) and then subjected to a cdk4 IVK assay. For lane 3, the Protein A-bound cdk4 kinase complex from lane 1 was incubated with 2.5 mg of exogenously added CaM-bound calcineurin for 20 min at 308C; calcineurin was then removed by washing the beads with three washes of 16 kinase buer, followed by a cdk4 IVK assay with GST-RbC as the substrate. (b) A CAK IVK assay ± 2 mg GST-cdk4/Histidine tagged (His)-cyclin D2 was added to 100 ml of a sonicated NIH3T3 lysate (56105 cells) as the source of CAK (Kato et al., 1994). The reaction was allowed to proceed for 1 h at 258C in CAK lysis buer with 10 mCi g-32P-ATP added to the reaction mixture. Lane 1 ± CAK phosphorylated GST-cdk4 (65 pmols) immobilized to Glutathione Sepharose beads. For lanes 2 and 3, 2.5 mg (43 pmols) of puri®ed CaM-bound calcineurin was added to Glutathione Sepharose beads containing immobilized GST-cdk4/His-cyclin D2, either without (lane 2) or with (lane 3) 2.5 mM EGTA/1 mM EDTA added in order to inactive the phosphatase activity of calcineurin. The resultant radiolabeled GST-cdk4 was separated by SDS ± PAGE, transferred to PVDF, and autoradiographed for 4 days to visualize radiolabeled proteins. (c) Jurkat cells were transiently transfected with: pEBG (lane 1); GST-tagged wild type calcineurin A (lane 2); GST-tagged constitutively active calcineurin A (lanes 3 ± 4), untreated (lane 3) or treated with 100 nM FK506 (lane 4) for 16 h and cdk6 kinase activity was assessed. (d) Same as (a), but a cdk6 IVK was carried out. All ®gures are the representation of three independently performed experiments

and would suggest that cdk4 kinase activity may start Much less is known about the phosphatases that are to become downregulated by calcineurin during the responsible for the dephosphorylation and subsequent DNA synthesis phase of the cell cycle and continue downregulation of the cyclin dependent kinases. In this into the mitotic phase resulting in further down- study, we demonstrated that calcineurin can physically modulation of the kinase activity of cdk4. Inhibition associate with cdk4, possibly dephosphorylate CAK of the activity of this phosphatase would result in the phosphorylated residues and thus downmodulate cdk4 continued activation of cdk4 kinase activity (Figure kinase activity. The regulation of cdk4 kinase activity is 5a). Cdk4 protein levels were comparable in the dependent upon the phosphatase activity of calcineur- presence and absence of FK506, as was the level of in, since the phosphatase dead form of calcineurin calcineurin (data not shown) and we were able to could not downmodulate the kinase activity of cdk4. In observe similar results in NIH3T3 and COS-7 cells addition, calcineurin demonstrated selectivity towards (data not shown). the G0/G1 kinase, cdk4, as we were not able to observe a similar modulation of cdk2 kinase activity nor of cdk6 kinase activity, a cyclin dependent kinase that is Discussion also abundant and active during the exit from G0/G1. The lack of modulation of cdk6 kinase activity by Cyclin dependent kinases (cdk) are regulated spatio- calcineurin and the lack of cdk6 association with a temperally as cells cycle through G0/G1 to the mitotic calcineurin anity matrix would support a direct phase. The activation state of these enzymes are interaction between calcineurin and cdk4 rather than dependent on phosphorylation events that possibly via the D cyclin family. Furthermore, FK506 inhibition control their cellular localization and substrate speci- of the activity of calcineurin resulted in the elevation of ®city. Cyclin activating kinase (CAK) has been shown the kinase activity of cdk4, suggesting that the to activate cdk4, cdk6 and cdk2 (Morgan, 1995). Cdk4 phosphatase activity of calcineurin was required for and cdk6 are important G0/G1 cyclin dependent the downregulation of the kinase activity of cdk4. kinases that use the family of D cyclins to assemble Recently, it was determined that members of the into an active kinase complex, while cdk2 is an major class of serine/threonine protein phosphatases important G2/M phase transition cyclin dependent could play essential roles in the regulation of cyclin kinase that associates with cyclin E. dependent kinases that control cellular proliferation.

Oncogene Calcineurin regulation of cdk4 S Baksh et al 2825 of p21CIP and p27KIP (but not p16) in Balb/c 3T3 cells (Yan and Mumby, 1999). In the budding yeast, Saccharomyces cerevisiae, both calcineurin (PP2B) (Mizunuma et al., 1998) and PP2C activity (Cheng et al., 1999) were demonstrated to downmodulate the activity of p34CDC28, an important G2/M checkpoint cdk in budding yeast. p34CDC28 becomes activated at the onset of mitosis by the coordinated action of kinases and phosphatases. In cooperation with Mpk1 (a component of the MAP kinase pathway), the activation of calcineurin was shown to downregulate p34CDC28 by increasing the activity of the inhibitory kinase, Swe1, on a trancriptional and translational level and thus prevent the entrance into mitosis. In contrast, PP2C-like activity was demonstrated to directly regulate the kinase activity of p34CDC28 in Sacchar- omyces cerevisiae and cdk2 and cdk6 kinase activity in HeLa cell extracts by directly dephosphorylating the activating phosphorylation event mediated by CAK. In this study, we were not able to observe changes in cdk6 and cdk2 kinase activity, nor changes in the levels of expression of p21CIP and p27KIP in either Jurkat or COS-7 cells when the phosphatase activity of calcineurin was inhibited (data not shown). We are proposing that calcineurin may play a role in speci®cally downregulating cdk4 kinase activity during the mitotic phase, possibly opposing the action of CAK. These results illustrate the complex nature of the regulation of cyclin dependent kinases and the important roles played by the various members of the serine/threonine phosphatase family in the cellular control of proliferation.

Materials and methods

Antibodies and immunoprecipitation The following polyclonal antibodies were used: rabbit anti- calcineurin B (Anity Bioreagents, Inc, Golden, CO, USA); rabbit anti-human cdk4 (C-22) and rabbit anti-human cdk6 (C-21) (Santa Cruz Biotechnology, Santa Cruz, CA, USA); PT101, a phosphothreonine-proline (TpP) speci®c antibody (New England BioLabs, Waltham, MA, USA). For cdk4 and cdk6 immunoprecipitations, cells (206106) were lysed in 880 ml of DIP buer containing 50 mM HEPES pH 7.2, Figure 5 Cell cycle dependent downregulation of the kinase 150 mM NaCl, 10% glycerol, 0.1% Tween-20, 1% Triton X- activity of cdk4 by calcineurin. (A) Asynchronously growing 100, 1 mM DTT, 1 mM sodium ¯uoride, 0.5 mM sodium Jurkat cells were allowed to undergo cell cycle synchronization vanadate, 10 mM b-glycerophosphate, 1 mg/ml aprotinin, using the indicated drugs for 22 h either with or without FK506 1 mg/ml leupeptin and 100 mg/ml PMSF. Lysed cells were (added at the same time as the addition of the synchronization spun down at 14 000 r.p.m. for 15 min, protein concentra- drugs). After 22 h the cells were harvested and lysed in DIP buer, normalized to total protein content, and then immuno- tions were normalized using the Bradford assay, and precipitated with a rabbit anti-cdk4 antibody and a cdk4 IVK immunoprecipitations were carried out by adding either assay (using GST-RbC as the substrate) was carried out as 10 mg of rabbit anti-calcineurin B; 0.5 mg rabbit anti-cdk4; outlined in Materials and methods. The IVK assay was repeated 1 mg of rabbit anti-cdk6 to the lysate; 0.5 mg of PT101. three times and the data normalized to the activity present in an Immunoprecipitations were allowed to proceed overnight at asynchronous population of Jurkat cells (100% activity). (b) Flow 48C, followed by the addition of 30 ml of a 50% slurry of cytometry (FACS) of synchronized cells using propidium iodide BSA blocked Protein A Sepharose beads. After 2 h the beads staining of the cells were washed extensively with DIP buer, and the proteins bound to the beads were resolved by SDS ± PAGE (Laemmli, Protein phosphatase (PP) 1 was found to directly 1970), transferred to polyvinylidene di¯uoride (PVDF), regulate the phosphorylation status of the p110 protein immunoblotted with the indicated antibody, and developed product of the retinoblastoma (Rb) tumor suppressor by the enhanced chemiluminescence (ECL) system (Amer- sham Corp.). gene in Balb/c 3T3 cells and thus directly regulate the transcriptional activity of E2F. PP2A phosphatase activity resulted in the decrease in cdk4, cdk6, and Cells lines and DNA transfections cdk2 kinase activity by decreasing cyclin D2 and cyclin Jurkat cells were grown in complete medium (RPMI 1640 E protein levels, while increasing the expression levels medium supplemented with 10% fetal bovine serum, 2 mML-

Oncogene Calcineurin regulation of cdk4 S Baksh et al 2826 glutamine, penicillin and streptomycin). Transfections were Cdk4 and cdk6 kinase assays carried out using electroporation with settings of 800 mF, low In vitro kinase assays (IVK) were performed for cdk4 and ohm and a voltage of 250 V. Cells (16107/500 ml) were used cdk6 as described elsewhere (Matsushime et al., 1994). with approximately 12 mg of CsCl prepared DNA. Inhibi- 2 Immunoprecipitations were carried out for 1 h at 48C. tion of calcineurin activity was carried out by the addition of Protein A Sepharose was added for 1 h, prewashed four 100 nM FK506 (Calbiochem, La Jolla, CA, USA) or 5 mM times with E-DIP lysis buer (DIP buer with 2.5 mM cyclosporin A (Sandoz Pharmaceuticals, NJ, USA) 3 h post EGTA/1 mM EDTA). After Protein A Sepharose precipita- transfection, followed by the incubation of the cells for 16 h tion, the beads were washed four times with E-DIP buer, at 378C. followed by two washes with 16 kinase buer (50 mM HEPES, pH 7.2, 10 mM MgCl2,5mM MnCl2,1mM DTT). Generation of calcineurin eukaryotic expression vectors The kinase reaction mixture (25 ml) contained 16 kinase buer, 10 mCi g-32P-ATP (New England Nuclear, Boston, Human calcineurin Aa (CnA) was cloned into the elongation MA, USA), 250 m ATP, and 0.65 mg GST-RbC fusion factor 1a-driven vector, pEBG; calcineurin B was cloned into M protein as substrate (GST-RbC: amino acids 772 ± 928 of the the vector, pEBB (generous gifts of Dr Bruce Mayer, The p110 protein product of Rb) (Boston Biologicals, Inc., Children's Hospital and Harvard Medical School, Boston, Waltham, MA, USA). The mixture was incubated at 378C MA, USA). pEBG encodes for glutathione S-transferase for 30 min and the reaction was stopped by adding SDS ± (GST), positioned 5' to the multiple cloning site. pEBB is the PAGE loading buer. The samples were boiled for 10 min parent vector for pEBG and does not contain GST. and resolved using a 10% SDS ± PAGE. The gel was then Therefore, all calcineurin A constructs were GST-tagged at dried, autoradiographed and quantitated using a BioRad the amino terminus. pET-CnA/B (encoding human calcineur- phosphoimager. in Aa) was used as the PCR template in the construction of the wild type (CnA) and constitutively active form of calcineurin A (CnAd); and pET-CnAdH151A/B was used in Cell cycle synchronization the construction of the phosphatase dead form of calcineurin Synchronized cells were prepared as previously described A (CnAdH151A) (both vectors were generous gifts of Dr (Krek and DeCaprio, 1995). Brie¯y, Jurkat cells (206106) Mondragon, Massachusetts Institute of Technology, Boston, were serum starved for 22 h at 378C in medium containing MA, USA) (Mondragon et al., 1997). For pEBG-CnA (wild 0.5% fetal bovine serum to achieve G /G arrest; with 5 mg/ type): 5' primer was GGTCGGATCCATGTCCGAGCC- 0 1 ml aphidicolin (Sigma Chemical CO., St. Louis, MO, USA) CAAGGC and 3' primer was AGCAGCAATATCCAGTG- to arrest cells in late G ; 750 mM hydroxyurea (Sigma AGGACCCTGC; pEBG-CnAd (amino acids 1 ± 368): 5' 1 Chemical CO., St. Louis, MO, USA) was used for S phase primer was the same one used for the wild type construct arrest; 500 ng/ml nocodazole (Sigma Chemical CO., St. and 3' primer was: AAGGAAGTCATCAGAAACG- Louis, MO, USA) was used for M phase arrest. Late G ,S GATCCCTGC. For the phosphatase dead form of calcineur- 1 and M phase cell cycle synchronizations were performed at in: same primers as for pEBG-CnAd. For calcineurin B, 378C in 15 ml of complete RPMI media. After 22 h the cells pGEX-3X-CnB was used as the PCR template using were lysed in 880 ml of DIP buer and then incubated with GGGGATCCCCATGGGAAGTGAGGCGAGT as the 5' the appropriate antibody for IVK assays. primer and GGGGATCCTCACACATCCACCACCAT as the 3' reverse primer.

In vitro interaction experiments Experiments using CaM-Sepharose beads were carried out as Acknowledgments outlined by Wesselborg et al. (1996). Brie¯y, CaM-Sepharose We would like to thank Mohammed Ladha for his beads were charged by washing in Tris-buered saline invaluable help with cdk4, cdk6 and cdk2 in vitro kinase containing 1 mM CaCl2. Calcineurin (Cn) (Sigma Chemical assays; Charles Sherr for providing the construct for GST- CO., St. Louis, MO, USA) was added to a ®nal concentra- cdk4, Histidine tagged-Cyclin D2 and hemataglutinnin tion of 55 nM and lysates were then applied to the CaM/Cn (HA)-tagged cdk4; Jim Koh for providing HA-tagged anity matrix and incubated for 1 h at 48C. Inhibition of cdk6; and Peter Adams for providing the GST-RbC binding was carried out by preincubation with 4 mM FK506 bacterial expression plasmid for the initial preparation of and 4 mM FKBP12 for 1 h at 48C in 250 ml of lysis buer. the substrate for cdk4 and cdk6 in vitro kinase assays. We Subsequently, lysates were added and incubated for 1 h at would also like to thank the members of the Burako and 48C. Beads were washed four times with precipitation wash DeCaprio laboratories for their helpful discussions. S buer and bound proteins resolved by SDS ± PAGE, Baksh is a recipient of the Medical Research Foundation immunoblotted with an anti-cdk4 antibody, and developed of Canada post-doctoral fellowship and an Abrahams post- by an enhanced chemiluminescence (ECL) system. doctoral Fellow.

References

Chen L, Glover JN, Hogan PG, Rao A and Harrison SC. Jinno S, Hung SC and Okayama H. (1999). Oncogene, 18, (1998). Nature, 392, 42 ± 48. 565 ± 571. Cheng A, Ross KE, Kaldis P and Solomon MJ. (1999). Genes Kato JY, Matsuoka M, Strom DK and Sherr CJ. (1994). Dev., 13, 2946 ± 2957. Mol. Cell. Biol., 14, 2713 ± 2721. Fruman DA, Burako SJ and Bierer BE. (1994). FASEB J., Klee CB, Ren H and Wang X. (1998). J. Biol. Chem., 273, 8, 391 ± 400. 13367 ± 13370. Fruman DA, Mather PE, Burako SJ and Bierer BE. (1992). Knudsen ES and Wang JY. (1996). J. Biol. Chem., 271, Eur. J. Immunol., 22, 2513 ± 2517. 8313 ± 8320. Grith JP, Kim JL, Kim EE, Sintchak MD, Thomson JA, Krek W and DeCaprio JA. (1995). Methods Enzymol., 254, Fitzgibbon MJ, Fleming MA, Caron PR, Hsiao K and 114 ± 124. Navia MA. (1995). Cell, 82, 507 ± 522.

Oncogene Calcineurin regulation of cdk4 S Baksh et al 2827 Laemmli UK. (1970). Nature, 227, 680 ± 685. Rao A, Luo C and Hogan PG. (1997). Annu.Rev.Immunol., Matsushime H, Quelle DE, Shurtle SA, Shibuya M, Sherr 15, 707 ± 747. CJ and Kato JY. (1994). Mol. Cell. Biol., 14, 2066 ± 2076. Sherr CJ and Roberts JM. (1995). Genes Dev., 9, 1149 ± 1163. Mizunuma M, Hirata D, Miyahara K, Tsuchiya E and Shibasaki F, Price ER, Milan D and McKeon F. (1996). Miyakawa T. (1998). Nature, 392, 303 ± 306. Nature, 382, 370 ± 373. Mondragon A, Grith EC, Sun L, Xiong F, Armstrong C Weinberg RA. (1995). Cell, 81, 323 ± 330. and Liu JO. (1997). Biochemistry, 36, 4934 ± 4942. Wesselborg S, Fruman DA, Sagoo JK, Bierer BE and Morgan DO. (1995). Nature, 374, 131 ± 134. Burako SJ. (1996). J. Biol. Chem., 271, 1274 ± 1277. Northrop JP, Ho SN, Chen L, Thomas DJ, Timmerman LA, Yan Y and Mumby MC. (1999). J. Biol. Chem., 274, 31917 ± Nolan GP, Admon A and Crabtree GR. (1994). Nature, 31924. 369, 497 ± 502. Zhu J and McKeon F. (1999). Nature, 398, 256 ± 260.

Oncogene