FK506 binding protein associated with the calcium release channel (ryanodine receptor).

The calcium release channel (CRC)/ryanodine receptor (RyRec) has been identified as the foot structure of the sarcoplasmic reticulum (SR) and provides the pathway for calcium efflux required for excitation-contraction coupling in skeletal muscle. The CRC has previously been reported to consist of four identical 565-kDa protomers. We now report the identification of a 12-kDa protein which is tightly associated with highly purified RyRec from rabbit skeletal muscle SR. N-terminal amino acid sequencing and cDNA cloning demonstrates that the 12-kDa protein from fast twitch skeletal muscle is the binding protein for the immunosuppressant drug FK506. In humans, FK506 binds to the 12-kDa FK506-binding protein (FKBP12) and blocks calcium-dependent T cell activation. We find that FKBP12 and the RyRec are tightly associated in skeletal muscle SR on the basis of: 1) co-purification through sequential heparin-agarose, hydroxylapatite, and size exclusion chromatography columns; 2) coimmunoprecipitation of the RyRec and FKBP12 with anti-FKBP12 antibodies; and 3) subcellular localization of both proteins to the terminal cisternae of the SR, and not in the longitudinal tubules of SR, in fast twitch skeletal muscle. The molar ratio of FKBP12 to RyRec in highly purified RyRec preparations is approximately 1:4, indicating that one FKBP12 molecule is associated with each calcium release channel/foot structure.

In skeletal muscle, the triad junction mediates signal transfer from the plasmalemma, resulting in release of Ca2+ from the terminal cisternae of SR,' a phenomenon referred to as excitation-contraction coupling (1). The RyRec of the SR has been purified from skeletal muscle and identified as the calcium release channel involved in excitation-contraction coupling (1)(2)(3)(4)(5)(6). The RyRec has been identified morphologically as the foot structure of the triad junction, involved in the association of the transverse tubule with the terminal cisternae of SR (5)(6)(7). Molecular cloning of the cDNA encoding the RyRec has provided its primary sequence and the predicted size of its protomer (565,000 daltons) (8)(9)(10). The foot structure has been termed the junctional channel complex to indicate that it not only contains the calcium release channel but is directly involved in junctional association and senses the depolarization in the transverse tubule (8,11). The junctional channel complex has a molecular mass of 2.3 million (12) and has 4-fold symmetry. It is, therefore, a homotetramer and the largest channel structure known (6,7,11).
The 12,000-Da binding protein for the immunosuppressant FK506 (FKBP12) is a cis-trans peptidyl-prolyl isomerase which has been isolated and cloned from human T cells (13,14). The FK506-FKBP12 complex blocks IL-2 gene transcription required for T cell activation in a calcium-dependent manner (15). Little is known about the cellular role of FKBPl2 protein in the absence of FK506. In the present study we examined the structure, expression, and subcellular localization of FKBPl2 in skeletal muscle where we show it to be tightly associated with the calcium release channel involved in excitation-contraction coupling.

EXPERIMENTAL PROCEDURES
Co-purification of FKBPl2 with the Ryanodine Receptor-Skeletal muscle RyRec was purified as described by Inui and Fleischer (16) on sequential columns: heparin-agarose, hydroxylapatite, and size exclusion chromatography using either a Pharmacia LKB Biotechnology Inc. Sepharose 6B or a Toso Haas TSK-gel G 4000 SW high pressure liquid chromatography column. Four separate preparations of purified RyRec were electrophoresed in reducing buffer through 12% gels in Tris-Tricine SDS-polyacrylamide (17) or through 5-15% gradient gels in Tris-glycine buffers (18). Each purified RyRec contained a low molecular weight polypeptide of relative molecular mass (M,) 12 kDa. This polypeptide was observed by Coomassie Blue staining of SDS-polyacrylamide gel electrophoresis gels in four out of four different RyRec preparations (data not shown). The amount of FKBPl2 in the purified RyRec was determined by gel densitometry using BSA (0.1-1 pg) as the protein standard with an automated gel analyzer and image processing system (Technology Resources, Nashville, TN). FKBP12 protein content for each RyRec preparation was expressed as molar ratios compared with the RyRec. Molar ratio calculations assume that the extinction coefficient of Coomassie Blue staining is the same for the FKBPl2 and the RyRec protomer.
Obtaining Amino Acid Sequence and Isolation of Rabbit Skeletal Muscle FKBPl2 cDNA-Amino acid sequence was obtained from the 12-kDa protein, which co-purified with the RyRec, for the purpose of designing synthetic oligonucleotides to be used as probes for cDNA cloning. Approximately 30 pg of rabbit skeletal muscle RyRec purified as described previously (16) were electrophoresed on a 15% polyacrylamide gel in 50 p1 of Laemmli sample buffer (18) after heating at 95 "C for 2 min. Cathode and anode buffers were as described (17).
The gel was electroblotted onto polyvinyldifluoride, and two bands were visualized on the blot after staining with Coomassie Brilliant Blue (data not shown). The RyRec had barely entered the 15% separating gel and was clearly visible at the uppermost part of the gel; a faint 12-kDa band, estimated by comparison with standards to be less than 100 ng of protein, was observed and used for direct sequencing. This low molecular weight band (12 kDa) was cut from the blot subjected to automated sequencing in an AB1 model 477A instrument. Standard sequencing procedures (19) were optimized for subpicomole analysis as previously described (20). The sequence obtained was as follows (yields, in picomoles, are between parentheses, single letter amino acid code is used, and X refers to residues which were not positively identified): X , V K(0.13). No additional unambiguous signals were observed past residue 17; secondary signals were detected in cycles 2 (1.2 pmol of Leu) and 11 (0.13 pmol of Phe). A data base search of GenBank revealed a perfect match with the human FK506-binding protein (residues 3-18).
A random primed, rabbit fast twitch skeletal muscle cDNA library was constructed in XZAPII as previously described (8). A synthetic oligonucleotide based on the sequence of amino acids 3-10 of the rabbit FKBPl2 was synthesized, labeled at the 5' end with y3*P, and used to screen 4 X lo6 recombinants. Five cDNA clones encoding portions of the rabbit skeletal muscle FKBPl2 were isolated and sequenced on both strands using dideoxy terminators.
Raising Anti-FKBPl2 Antibodies and Immunoprecipitation of the Ryanodine Receptor-Ten mg of a synthetic peptide (CVQVET-ISPGDGRTF, corresponding to amino acids 3-16 of the FKBP12, with a cysteine residue added at the carboxyl terminus) were coupled to keyhole limpet hemocyanin in the presence of 0.3% glutaraldehyde following standard protocols (21). A New Zealand White rabbit was immunized with 1 mg of peptide-keyhole limpet hemocyanin conjugate in complete Freund's adjuvant, followed by 0.5-mg booster injections subcutaneously at 3, 5, and 7 weeks in incomplete Freund's adjuvant. The rabbit was bled and serum obtained 12 days following the final injection. The specificity of the resulting anti-FKBPl2 antibody was determined by immunoblots of FKBPl2 purified from human T cells (Jurkats, data not shown). The anti-FKBP12 antibody was coupled to CNBr-activated Sepharose as per the manufacturer's instruction (Pharmacia LKB Biotechnology Inc.). The following adjustments were made: 30 mg/ml anti-FKBP12 antibody was dialyzed against 0.1 M NaHC03, 0.5 M NaCl (dialyzing buffer) at 4 "C. The antibody was centrifuged for 1 h at 14,000 rpm to remove aggregates. Protein concentration was determined (22) and adjusted to 5 mg/ml with the coupling buffer (0.1 M Tris-HC1, pH 8.0, 0.5 M NaCl). One gram of CNBr-Sepharose 4B (Pharmacia) was washed with 200 ml of 1 mM HC1, followed by one wash with coupling buffer. An equal volume of CNBr-Sepharose 4B beads was added to the dialyzed antibody solution and rotated end over end for 12 h at 4 "C. Excess ligand was removed by washing with coupling buffer, and binding sites were blocked with Tris-HC1 buffer (0.1 M, pH 8.0) for 2 h at room temperature followed by washing with three cycles of acetate buffer (0.1 M, pH 4.0) containing NaCl (0.5 M) followed by a wash with Tris-HC1 buffer (0.1 M, pH 8.0) at 4 "C. Highly purified RyRec was precleared by incubation with activated Sepharose (50 pl) for 1 h at 4 'C followed by centrifugation at 14,000 rpm for 30 min at room temperature, then mixed with either 50 pl of anti-FKBP12 antibody-Sepharose/TSA solution (0.01 M Tris-HC1 pH 8.0, 0.14 M NaC1, 0.025% azide) per ml of lysate or preimmune serum for 1 h at 4 "C. Beads were then washed as follows: two washes with TSA solution containing 0.1% Triton X-100, one wash with TSA solution. Sample buffer (0.05 M Tris-HC1, pH 6.8,0.1% SDS) was then added, and the sample was heated for 5 min at 100 "C. The supernatant was then loaded onto a 6% SDS-polyacrylamide gel. Following electrophoresis, gels were stained with Coomassie Brilliant Blue. Terminal cisternae and longitudinal tubules were purified as previously described (23,24), protein samples were electrophoresed through 15% SDS-polyacrylamide gels, transferred to nitrocellulose, and immunoblot blots performed with anti-FKBPl2 antibody (1/7,500 dilution) as described previously (25).
Isolation of RNA and Northern Blot Analysis-Total RNA was prepared using standard guanidinium-thiocyanate lysis buffer and centrifugation through cesium chloride cushion (26). Twenty pg of total RNA prepared from various rabbit tissues was size-fractionated on a 1% formaldehyde agarose gel, blotted to nitrocellulose, and probed with c~-~'P-labeled FKBPl2 and RyRec cDNA probes randomly labeled to a specific activity of 1 X lo9 cpm/pg. The FKBPl2 cDNA probe was a clone containing the entire coding region of the rabbit FKBP12; the RyRec cDNA probe was a 700-base pair rabbit skeletal muscle cDNA described previously (8). Hybridization was at 42 "C overnight and washing at 65 "C in 0.2 X SSC. Films were autoradiographed with a single intensifying screen at -70 "C for 24 h. To ensure that equivalent amounts of RNA were present in each lane, the same Northern blot was also hybridized with cDNA encoding the glyceraldehyde-3-phosphate dehydrogenase from chicken muscle (data not shown).

RESULTS
To date, the calcium release channel has been considered to consist solely of four 565-kDa protomers. We now report that an additional protein, FKBP12, has been found in four of four different samples of highly purified (>go% purity) RyRec prepared in our laboratory.
We have previously reported the extensive proteolytic mapping and microsequencing of highly purified RyRec from rabbit fast twitch skeletal muscle (8,27). Thirty out of thirtyone peptides mapped to the 5,037-amino acid sequence of the RyRec protomer (8,27). The only peptide which did not map to the deduced amino acid sequence of the RyRec was subsequently matched to the NH, terminus of the FKBPl2 (28). Thus, our earlier results provided the initial indication that FKBPl2 is tightly associated with the RyRec (8,27).
The RyRec was purified using three sequential columns (heparin-agarose, hydroxylapatite, and size exclusion chromatography with either Sepharose 6B or TSK-gel G-4000 SW) essentially as described previously (16). High molecular weight bands other than the intact RyRec, seen in heavily loaded protein gels (Fig. l), appear to be proteolytic breakdown products of the RyRec because: 1) addition of protease inhibitors greatly reduces their abundance (5); and 2) they react with polyclonal anti-RyRec antibodies.' In the low molecular weight range, the only prominent band seen by Coomassie staining was a 12-kDa protein (Fig. l). The NH2terminal amino acid sequence (amino acids 3-18) and the complete deduced amino acid sequence of the 12-kDa protein determined by cDNA cloning from skeletal muscle (Fig. 2) are identical to that reported for the human T cell FKBPl2 (14). The evolutionary conservation of sequence found in the FKBPl2 (100% identity over the 108 amino acid residues between the rabbit skeletal muscle and the human Jurkat T cell forms) suggests that it performs a fundamental function in diverse tissues.
Having identified the 12-kDa protein as FKBP12, we sought to demonstrate that it was associated directly with the RyRec. To do so we used an anti-FKBP12 antibody to immunoprecipitate the RyRec. This anti-FKBP12 antibody was incubated with highly purified RyRec, and the resulting immunoprecipitate was electrophoresed on a 6% SDS-polyacrylamide gel (Fig. 3a). A 565-kDa band which co-migrates with purified RyRec was seen after electrophoresis followed by staining with Coomassie Brilliant Blue (Fig. 3a). Immunoblotting showed that this 565-kDa band was recognized by a polyclonal anti-RyRec antibody but not by the FKBPl2 antibody (data not shown). The anti-FKBP12 antibody was not able to immunoprecipitate RyRec which had been electroeluted from a Coomassie-stained gel (and thus separated from FKBP12). Taken together these data indicate a tight association between the FKBPl2 and the RyRec.
To determine whether the skeletal muscle FKBPl2 colocalized with the RyRec in the terminal cisternae of the SR, sucrose gradient fractions of SR longitudinal tubules (23), SR terminal cisternae (24), and highly purified preparations of RyRec were immunoblotted with anti-FKBP12 antibody (  FIG. 2. Comparison of the nucleotide sequence for the hum a n (Hum, first line) and rabbit (Rub, second line) F K B P l 2 ( t o p two lines) a n d the deduced amino acid sequence (third line) of the rabbit fast twitch skeletal muscle FKBP12. The deduced amino acid sequence of the rabbit skeletal muscle FKBPl2 is identical to that reported for the human T cell FKRP12 (14). The underlined sequence indicates the region near the amino terminus for which amino acid sequence was obtained by the microsequencing of the 12-kDa protein which co-purified with the RyRec (as shown in Fig. 1). Vertical lines (!) indicate sequence identity. IAwer case letters in the top line represent base pair mismatches between the human (lower case) and rabbit (upper case) sequences. The final TGA is the stop codon.

36).
These immunoblots showed that FKBP12 was present in the purified RyRec and the terminal cisternae but not in longitudinal tubules (Fig. 36). These results indicate that FKBP12 is associated specifically with the terminal cisternae of SR.
Northern blot analysis, using a rabbit skeletal muscle FKBP12 cDNA probe, showed that the 1.5-kb FKBPl2 mRNA was found in all forms of muscle examined, including smooth, cardiac, and fast (Fig. 4) and slow twitch skeletal muscles (data not shown).

DISCUSSION
One function of the FK506-FKBP12 complex is to hlock T cell activation. The FK506-FKRP12 complex blocks T cell activation by inhibiting the antigen-induced transcription of the IL-2 gene. It has been shown that the FK506-FKRP12 complex inhibits the calcium-induced translocation of the transacting factor NF-AT (nuclear factor of activated cells) which binds to the transcriptional regulatory region of the IL-2 gene (15). Independently, it has been shown that the FK506-FKBP12 complex inhibits the calcium-activatedphosphatase calcineurin (29). However, the cellular function of FKBPl2 in the absence of FK506 remains to be elucidated. FKBPl2 had previously been regarded as a cytosolic protein; the present study demonstrates the tight association between FKBPl2 and the ryanodine receptor, and the localization of FKBP12 to the terminal cisternae of the sarcoplasmic reticulum.
FKBPl2 which we have purified and cloned from rabbit skeletal muscle is identical in primary structure to the major cytoplasmic form of FKBPl2 in human T cells (14) (Fig. 2). The FKBPl2 associated with the RyRec in skeletal muscle is likely to be associated with the cytoplasmic side of the SR, as opposed to the luminal surface.
FKBPl2 is a member of a family of proteins known as immunophilins. This family of immunophilins includes cyclophilin, the binding protein for cyclosporin A (30,31). FKBPl2 and cyclophilin are both cis-trans peptidyl-prolyl isomerases whose primary structures are unrelated. FK506 and cyclosporin A have been reported to inhibit T cell activation pathways under conditions in which activation is associated with increased cytoplasmic calcium (32,33). On the basis of this observation it has been proposed that FKBP12 may modulate the conformation of ion channels involved in the signal transduction pathway for T cell activation (34). Furthermore, a recent study reported that FK506, cyclosporin A, and rapamycin all significantly decreased the magnitude of intracellular Ca2+ release typically seen after activation of the T cell receptor in human peripheral blood T lymphocytes (35). The present study is the first to report a direct association between FKBPl2 and a calcium channel. It remains to be determined whether FKBP12 is associated with calcium channels in T cells.
The cis-trans peptidyl-prolyl isomerase activity of the FKBP12 raises the question of whether it might isomerize a peptidyl-proline bond in the RyRec. Proline residues are found in or near the transmembrane regions of transport proteins (36). The putative M2 transmembrane region near the carboxyl terminus of the skeletal muscle RyRec contains a proline residue (residue 4641). The finding that FKBPl2 is present in the terminal cisternae of the SR and is tightly associated with the RyRec suggests that one of its cellular functions may be to alter channel conformation.