Cell, Vol. 83,463-472, November 3, 1995, Copyright 0 1995 by Cell Press Calcineurin Associated with the lnositol 1,4,5=Trisphosphate Receptor-FKBPI 2 Complex Modulates Ca*+ Flux Andrew M. Cameron,* Joseph P. Steiner,* such as the transcription factor NFAT (for nuclear factor A. Jane Roskams, l Siraj M. Ali, * Gabriele V. Ronnett,? of activated T cells) (Liu et al., 1991; O’Keefe et al., 1992; and Solomon H. Snyder**5 Clipstone and Crabtree, 1992; for review of calcineurin, *Department of Neuroscience see Klee et al., 1988). In its phosphorylated state, NFAT tDepartment of Neurology cannot translocate to the nucleus, where it regulates ex- $Department of Pharmacology and Molecular Sciences pression of genes critical for T cell activation such as in- SDepartment of Psychiatry and Behavioral Sciences terleukin-2 (IL-2). Rapamycin acts differently than cyclo- Johns Hopkins University School of Medicine sporin A and FK506, as it blocks the actions of IL-2 rather Baltimore, Maryland 21205 than its synthesis (Bierer et al., 1990). Rapamycin binds with high affinity to FKBP12 and inhibits its rotamase activ- ity, but the rapamycin-FKBP12 complex does not interact Summary with calcineurin. Rather, this immunosuppressant-immu- nophilin complex binds to a recently identified target pro- The immunosuppressant drug FK506 binds to the im- tein designated RAFT (for rapamycin and FKBP12 target) munophilin protein FKBPl2 and inhibits its prolyl iso- or FRAP (for FKBP-rapamycin-associated protein) (Sa- meraseactivity. Immunosuppresiveactions, however, batini et al., 1994; Brown et al., 1994). are mediated via an FK506-FKBP12 inhibition of the Marks and colleagues have observed an association Ca*+-activated phosphatase calcineurin. Physiologic of FKBP12 with the ryanodine receptor (RyR) of skeletal cellular roles for FKBPl2 have remained unclear. muscle in the absence of FK506 (Jayaraman et al., 1992; FKBPl2 is physically associated with the RyR and IP3R Timerman et al., 1993; Brillantes et al., 1994). RyR is a Ca*+ channels in the absence of FK506, with added tetrameric Ca*+ channel that mediates Ca’+-induced Ca2+ FK506 disrupting these complexes. Dissociation of release in muscle, brain, and other tissues (Takeshima FKBPl2 results in alteration of channel Ca% conduc- et al., 1989; McPherson and Campbell, 1993). We have tance in both cases. We now report that calcineurin is demonstrated that the inositol 1,4,5-trisphosphate (IP,) re- physiologically associated with the IP3R-FKBP12 and ceptor (IPBR), a structurally and functionally related tetra- RyR-FKBP12 receptor complexes and that this inter- merit Ca2+ channel that shares up to 40% homology to action can be disrupted by FK506 or rapamycin. Cal- RyR in some regions, is also a direct target of FKBP12 cineurin anchored to the IP3R via FKBP12 regulates (Cameron et al., 1995; for review see Furuichi and Miko- the phosphorylation status of the receptor, resulting shiba, 1995; Snyder and Sabatini, 1995). Whereas FK506 in a dynamic Ca*+ -sensitive regulation of IP3-mediated stimulates the binding of FKBP12 to calcineurin, FK506 Ca2+ flux. dissociates FKBP12 from RyR and from IP$R. When FKBP12 is “stripped” from the RyR-FKBPlP or IP3R- FKBP12 complex, the Ca2+ channels of these two proteins Introduction become “leaky.” Accordingly, net accumulation of Ca’+ into RyR-or IPsR-gated stores is diminished, and releasing The immunophilin proteins mediate the clinical effects of agents such as caffeine or IPa cause increased Ca*+ flux at the immunosuppressant drugs cyclosporin A, FK506, and lower concentrations. It was assumed that the isomerase rapamycin. Cyclophilins, thefirstclassof immunophilinsto activity of FKBP12 was the means by which this protein be identified, bind to cyclosporin A, acyclic undecapeptide modulates the RyR and IP3R. However, recent evidence (Handschumacher et al., 1984; for review see Walsh et obtained from mutant FKBP12 proteins lacking isomerase al., 1992). FK506 and rapamycin, structurally related mac- activity suggests that this is not the case (Timerman et rolides, instead bind to the FK506-binding protein (FKBP) al., 1995). Accordingly, we have attempted to investigate class of immunophilins. Although there are a number of alternative means by which FKBP12 modulates IP,R- and membersoftheimmunophilinfamily, immunosuppressant RyR-mediated Ca*+ flux. actions appear largely attributable to cyclophilin A and IP3R can be phosphorylated at three distinct sites by B and FKBP12 (Bram et al., 1993). The cyclophilins and protein kinase C (PKC), Caz+/calmodulin~dependent pro- FKBPs do not share amino acid sequence similarity, but tein kinase II (CaMKII), and cyclic AMP-dependent protein both display peptidylprolyl-cis-trans isomerase (rota- kinase (PKA), respectively (Ferris et al., 1991). Whether mase) activity thought to be associated with protein folding phosphorylation and dephosphorylation physiologically (Schmid, 1993). Furthermore, immunophilin rotamase ac- regulate IP,R function has not been established. The asso- tivity is inhibited by their respective immunosuppressant ciation of FKBP12 with IP3R suggested a potential func- ligands. lmmunosuppressant actions, however, do not ap- tional link of calcineurin and IP3R. We now demonstrate a pear to derive from inhibition of this isomerase activity. physical association of calcineurin with the IP,R-FKBP12 Rather, the drug-immunophilin complex of cyclosporin and RyR-FKBP12 complexes that is potently disrupted A-cyclophilin or FK506-FKBP12 binds to the protein by treatment with FK506 or rapamycin. We also show that phosphatase calcineurin to inhibit its activity, thereby aug- complexed calcineurin modulates the phosphorylation menting levels of phosphorylated calcineurin substrates status and Ca’+ flux properties of IP3R. Cell 464 Results point that reflects the molecular mass of the IP3R tetramer, which is greater than 1.2 x 1 O6 Da. Subjecting the identical Copurification and Coimmunoprecipitation FPLC fractions to Western blot analysis with an antibody Demonstrate an FKSOBSensitive Association specific for calcineurin reveals that calcineurin immunore- of Calcineurin with IP3R and RyR activity closely parallels the distribution of IP3R (Figure To look for an association of calcineurin with IP3R, we 1B). lmmunoreactivity for both IP3R and calcineurin is purified lPBR from rat cerebellum to apparent homogeneity blocked by preabsorption with antigen (data not shown). by sequential fractionation on heparin and concanavalin Thus, not only is calcineurin present in a highly purified A columns as previously described (Ferris et al., 1989). preparation of IPzR, but when such a preparation is frac- The purified receptor was then subjected to further frac- tionated by gel filtration calcineurin comigrates with the tionation on a fast protein liquid chromatography (FPLC) lPBR even though its molecular mass is only 10% that of Superose 6 column, which separates proteins according the IP3R tetramer. In contrast, when pure calcineurin is to their molecular mass. Purified IP& immunoreactivity fractionated by FPLC in an identical fashion, calcineurin is concentrated in fractions 9-12 (Figure lA), an elution immunoreactivity migrates in fractions 18-19, a later elu- tion point more consistent with a protein of molecular mass FPLC fraction # around 80 kDa (Figure 1C). Treating our preparation of IPzR with rapamycin, which disrupts the IP3R-FKBP12 in- teraction (Cameron et al., 1995), before FPLC fraction- ation causes calcineurin immunoreactivity to shift back to fractions 18 and 19 (Figure 1 D), while the IP3R continues A to run in fractions 9-12 (data not shown). We also demonstrate the specific association of cal- cineurin with IP3R and RyR by immunoprecipitation ex- periments. We employed antibodies to IP9R with crude preparations of rat cerebellar membranes. The immuno- precipitate displays immunoreactive calcineurin as well as IP3R (Figure 2). Inclusion of increasing concentrations of FK506 in the membrane preparation decreases the B amount of calcineurin in the immunoprecipitate. Quantifi- cation of band intensity reveals that half-maximal reduc- tion of the presence of calcineurin in the immunoprecipi- tate occurs between 10 and 100 nM FK506. This range is similar to the ECso obtained for FK506 dissociation of FKBP12 from RyR reported by Timerman and colleagues to be in the range of 120-500 nM (Timerman et al., 1993). Cyclosporin A fails to influence the calcineurin content of C the immunoprecipitate, while rapamycin, which like FK506 binds tightly to FKBP12, markedly reduces the calcineurin content of the immunoprecipitate. The association of cal- cineurin with the IP,R-FKBP12 complex appears to re- quire Ca’+, as EGTA depletes the calcineurin content of the immunoprecipitate (Figure 2). Control immunopre- cipitation experiments were carried out with antibodies against calbindin and microtubule-associated protein 2 D (MAP-P), which like IPsR are proteins of high abundance in cerebellar Purkinje cells. Calcineurin immunoreactivity was not seen in the immunoprecipitate in either case, nor in experiments using the anti-lP3R antibody preimmune serum (Figure 2). Figure 1. Copurification of Calcineurin with IP,R Is Disrupted by Treat- We have also demonstrated a similar association of cal- ment with Rapamycin cineurin with RyR (Figure 3A). lmmunoprecipitation of cer- In (A), (B) and (D), purified preparations of IP,R were subjected to ebellar homogenates with anti-RyR antibodies results in FPLC gel filtration and analyzed via Western blot