Molecular cloning, DNA sequence analysis, and biochemical characterization of a novel 65-kDa FK506-binding protein (FKBP65).

We have identified a mouse gene encoding a 65-kDa protein (FKBP65) that shares homology with members of the FK506-binding protein (FKBP) class of immunophilins. Predicted amino acid sequence shows that this protein shares significant homology with FKBP12 (46%), FKBP13 (43%), FKBP25 (35%), and FKBP52 (26%). FKBP65 contains four predicted peptidylprolyl cistrans-isomerase (PPIase) signature domains, and, although similar in size, is distinct from FKBP52 (also identified as FKBP59, hsp56, or HBI), which contains three FKBP12-like PPIase domains. With N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide as the substrate, recombinant FKBP65 is shown to accelerate the isomerization of the prolyl peptide bond with a catalytic efficiency similar to other family members. This isomerization activity is inhibited by FK506 and rapamycin, but is not sensitive to Cyclosporin A. Based on Northern blot analysis, FKBP65 mRNA transcripts are present in lung, spleen, heart, brain, and testis. A polyclonal antibody, raised against a COOH-terminal peptide (amino acid residues 566-581), was used to immunoprecipitate FKBP65 from NIH3T3 cells and demonstrate that FKBP65 is a glycoprotein. In addition, [32P]orthophosphate labeling experiments show that FKBP65 is also a phosphoprotein. These results suggest that FKBP65 is a new FKBP family member.

We have identified a mouse gene encoding a 65-kDa protein (FKBP65) that shares homology with members of the FK506-binding protein (FKBP) class of immunophilins. Predicted amino acid sequence shows that this protein shares significant homology with FKBP12 (46%), FKBP13 (43%), FKBP25 (35%), and FKBP52 (26%). FKBP65 contains four predicted peptidylprolyl cistrans-isomerase (PPIase) signature domains, and, although similar in size, is distinct from FKBP52 (also identified as FKBP59, hsp56, or HBI), which contains three FKBP12-like PPIase domains. With N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide as the substrate, recombinant FKBP65 is shown to accelerate the isomerization of the prolyl peptide bond with a catalytic efficiency similar to other family members. This isomerization activity is inhibited by FK506 and rapamycin, but is not sensitive to Cyclosporin A. Based on Northern blot analysis, FKBP65 mRNA transcripts are present in lung, spleen, heart, brain, and testis. A polyclonal antibody, raised against a COOH-terminal peptide (amino acid residues 566 -581), was used to immunoprecipitate FKBP65 from NIH3T3 cells and demonstrate that FKBP65 is a glycoprotein. In addition, [ 32 P]orthophosphate labeling experiments show that FKBP65 is also a phosphoprotein. These results suggest that FKBP65 is a new FKBP family member.
FK506 and rapamycin are immunosuppressant macrolide antibiotics that mediate their activity in part by binding to members of a ubiquitous family of highly conserved intracellular receptors termed immunophilins (1,2). Although FK506 and rapamycin are structurally related and bind to a distinct class of immunophilins termed FK506-binding proteins (FK-BPs), 1 they exert their immunosuppressive effects by blocking different signal transduction pathways in normal T cells. FK506 selectively inhibits the Ca 2ϩ -dependent signal trans-duction events by specifically binding calcineurin, a Ca 2ϩ -and calmodulin-dependent serine/threonine phosphatase (3). Inhibition of calcineurin phosphatase activity interferes with an early event that is essential for T-lymphocyte growth and differentiation (4,5). In comparison, rapamycin appears to affect Ca 2ϩ -independent, IL-2-driven T-cell proliferation (6,7). A characteristic shared by immunophilins is peptidylprolyl cistrans-isomerase (PPIase) activity (8,9), which is inhibited upon drug binding (10). However, the inhibition of PPIase activity in itself does not appear to be responsible for the immunosuppressant effect of the drugs (7,11).
This report describes the molecular cloning, sequencing, and biochemical characterization of a novel 65-kDa protein, FKBP65, isolated from a NIH3T3 fibroblast cDNA library. We show that this protein, as well as containing four peptidylprolyl cis-trans-isomerase domains, possesses PPIase activity comparable to other FKBPs and appears to be a new and distinct FKBP family member. Cloning and Sequencing of Murine FKBP65 cDNA-Anti-Pep 2 antiserum was used to screen a JB6 murine epidermal cell (18) cDNA expression library (Stratagene). A 2468-bp clone was isolated and sequenced (19). A 1700-bp (EcoRI-SacI) segment of the 2.4-kb clone was then used to screen a Zap NIH3T3 fibroblast cDNA library (Stratagene). Hybridization was performed at 42°C for 48 h in 5 ϫ SSC, 5 ϫ Denhardt's solution, 40% formamide, 0.1% SDS, 100 g/ml sheared salmon sperm DNA. The most stringent wash was done at 0.2 ϫ SSC with 0.1% SDS at 65°C. Autoradiography was carried out at Ϫ70°C for 24 h with an intensifying screen. Eight positive clones were excised from the phagemid library and sequenced. Both strands of the largest insert (2.6 kb) were determined by the dideoxy chain termination method with Sequenase Version 2.0 (U. S. Biochemical Corp./Amersham Life Science) using [ 35 S]dATP and the Sequa Gel urea gel system (National Diagnostics, Atlanta, GA).
Northern Blot Hybridization-Total RNA (20 g) from various mouse tissues (Clontech, San Francisco, CA) was electrophoresed in a 3.9 M formaldehyde, 20 mM MOPS, pH 7.6, 1.0% agarose gel, transferred to MagnaNylon Membrane (Micron Separations Inc., Westborough, MA), * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) L07063.
Protein Expression, Purification, and Peptidylprolyl cis-trans-Isomerase Assay-A 1600-bp segment of the FKBP65 clone was subcloned into the pET21d ϩ vector, and protein was induced with isopropyl-1-thio-␤-D-galactopyranoside (Novagen, Madison, WI). Cell pellets were treated with lysozyme (Sigma), sonicated in 50 mM sodium phosphate, 0.1% Triton X-100, 0.2% ␤-mercaptoethanol, pelleted at 20,000 ϫ g, and solubilized in 6 M guanidine-HCl, 40 mM Tris, pH 8.0, 1% ␤-mercaptoethanol (20). Further purification was by RP-HPLC (Bondapak C18 PrepPak cartridge 10 m, 125 Å, 25 ϫ 100 mm, Waters, Bedford MA) and verified as FKBP65 by peptide sequencing. Recombinant FKBP65 (rFKBP65) was renatured prior to using in the PPIase assay. The isomerization activity of the renatured rFKBP65 was assayed as described (8,9) with minor modifications, using N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide (Sigma) as substrate at a final concentration of 108 M, preincubated with rFKBP65 (8 -80 nM final concentration) at ambient temperature. ␣-Chymotrypsin (Sigma) was used at a final concentration of 27 M. Absorbance readings were taken immediately upon mixing in a Perkin Elmer Lambda 6 spectrophotometer at 395 nm at 0.2-s intervals for 180 s. Drug inhibition studies were performed as described above except various amounts of FK506 or rapamycin (27-200 M final concentration) were added to the substrate/ FKBP65 mixture using 16 nM FKBP65. The k cat /K m was determined by plotting the observed rate constants (k obs ) versus rFKBP65 concentration (data not shown). The first order rate constant (k obs ) was calculated from the slope of the plot of the log of the difference between absorbance at steady state and absorbance at time t plotted against time (21,22).
In Vivo [ 35 S]Methionine Metabolic Labeling and Immunoprecipitations-NIH3T3 cells were washed with phosphate-buffered saline and incubated in methionine-free Dulbecco's modified Eagle's medium, 5% dialyzed fetal bovine serum for 1 h at 37°C, 5% CO 2 . The cells were labeled for 3 h with 100 Ci/ml [ 35 S]methionine (Amersham) and then lysed in TNT (20 mM Tris-HCl, pH 7.5, 200 mM NaCl, 1% Triton X-100) with 1% Aprotinin and 2 mM phenylmethylsulfonyl fluoride. Cell lysates were centrifuged at 14,000 rpm for 30 min at 4°C and precleared with protein A-Sepharose CL-4B (Pharmacia-LKB, Uppsala, Sweden) and preimmune serum followed by immune precipitation with protein A-Sepharose CL-4B and Pep 4 antiserum with or without peptide competitor. Samples were electrophoresed using the method of Laemmli (23), and the gels were enhanced with Fluoro-Hance (Research Products International, Mt. Pleasant, IL), dried, and exposed to film for 24 h at Ϫ70°C.
Endoglycosidase F Treatment-NIH3T3 cells were labeled in vitro with [ 35 S]methionine and immune-precipitated as described above, after which samples were denatured by boiling, cooled to 37°C prior to adding 0.4 unit of Endo-F (endoglycosidase F, N-glycosidase F-free; Boehringer Mannheim), and incubated for 3 h at 37°C (24). Samples were electrophoresed as described above.
In Vivo [ 32 P]Orthophosphate Labeling-The in vivo [ 32 P]orthophosphate labeling has been described in detail previously (25). Cells were lysed in TNT, immune-precipitated and electrophoresed (described above), electroblotted onto ImmobilonP TM (Millipore), and exposed to film. Western blotting was performed on the membranes after the 32 P had decayed by incubating overnight with a 1:1000 dilution of Pep 4 antiserum. The filter was washed and then incubated with 5 ϫ 10 5 cpm/ml 125 I-protein A and exposed to film.

RESULTS
Cloning, Sequencing, and mRNA Tissue Distribution of FKBP65-A unique 2468-bp clone was isolated from a mouse JB6 epidermal cell cDNA Zap expression library and sequenced (19). We considered this clone to represent a partial gene sequence because the first methionine residue did not completely comply with a Kozak consensus initiation sequence (26), and an upstream stop codon was not encoded within the 5Ј region of the JB6 clone. To obtain additional 5Ј and 3Ј sequences, an internal 1.7-kb segment of the 2.4-kb clone was used to screen an NIH3T3 fibroblast Zap cDNA library. A 2636-bp clone was obtained and sequenced as described under "Experimental Procedures" (GenBank TM accession L07063). The 2.6-kb clone was identical with the 2.4-kb clone except that it contained an additional 5Ј sequence, which included a TAG stop codon 48 bp upstream from the first methionine, a polyadenylation signal (AATGAAA) located at bp 2431-2437, a 3Ј-untranslated region from nucleotides 1837-2597 (Fig. 1A), and a 39-nucleotide poly(A) tail (data not shown). This clone contained an ORF of 1743 nucleotides (nucleotides 94 -1836) encoding a protein of 581 amino acids with a predicted molecular mass of 64,683 daltons and a calculated pI of 5.43 and predicted molar extinction coefficient at 280 nm of 52,170.
The predicted amino acid sequence of FKBP65 revealed a highly hydrophobic region (residues 14 -23) suggesting an NH 2 -terminal signal sequence (Fig. 1A, bold sequence). We analyzed the amino acid pattern of FKBP65 using the method of Von Heijne (27) for predicting potential cleavage processing sites in a pre-protein. Based on this method, we predict that FKBP65 is cleaved after Arg 33 , resulting in a mature protein with a calculated molecular mass of 60,576 daltons. Additionally, there is a potential membrane-spanning segment (amino acid residues 244 -258) consisting of hydrophobic residues (Fig.  1A, italicized sequence), as well as an ER retention sequence (HEEL) at the COOH terminus (28).
Analysis of the 2.6-kb clone, using the University of Wisconsin GCG software analysis program (29), revealed a domain containing significant amino acid sequence similarity to members of the FK506/rapamycin-binding protein (FKBP) family (19). In accordance with the convention for naming FKBP family members corresponding to their calculated molecular weights (17), we have renamed the protein previously designated FKBPRP (19) as FKBP65. Overall, FKBP65 shares 46% and 43% amino acid sequence identity with human FKBP12 and FKBP13, respectively, and 35% and 26% sequence homology with human FKBP25 and FKBP52 (FKBP59, hsp56, or HBI).
Additional sequence analysis of the 2.6-kb FKBP65 clone revealed four peptidylprolyl cis-trans-isomerase (PPIase) signature domains as defined by the Motifs program (GCG, University of Wisconsin, Madison, WI; Ref. 29). These four domains are located at amino acid residues 61-149, 173-261, 285-373, and 398 -485 and represent 67% of the protein (Fig.  1B). The regions flanking the four PPIase domains share 30% similarity; however, they appear to be unique as they show no homology with other FKBP family members or any sequences reported in available data bases. The FKBP65 PPIase Domain II conforms to the PPIase consensus motif, whereas the other three domains each contain one mismatch (Fig. 1B).
Crystallographic studies of FK506 bound to human FKBP12 (hFKBP12) have identified two regions within the PPIase domain that appear to be important for FK506-binding interactions. There are five amino acid residues (Tyr 26 , Phe 99 , Val 55 , Ile 56 , and Trp 59 ) proposed to form the hydrophobic drug binding cavity and three amino acid residues (Ile 56 , Asp 37 , and Tyr 82 ) of hFKBP12 that appear to form hydrogen bonds with FK506 (30,31). These seven amino acids are conserved in all FKBPs that have been identified to date. Sequence alignment of hFKBP12 with the four PPIase domains of FKBP65 reveals that Domains I-III of FKBP65 strictly conserve six of these seven amino acid residues and Domain IV conserves five of the seven amino acids (Fig. 1B).
Unlike the ubiquitously expressed FKBPs identified to date, FKBP65 mRNA shows restricted expression. Northern blot analysis reveals a unique 2.6-kb RNA band in mouse lung, spleen, heart, and brain that hybridizes to a radiolabeled 1.7-kb (ORF) fragment of the FKBP65 clone (Fig. 2). The testis mRNA appears to contain two transcripts, the 2.6-kb band and an additional 3.5-kb band. Although RNA was visualized in all lanes, by either ethidium bromide staining of the gel prior to blotting or hybridization with an actin probe (data not shown), no FKBP65-specific hybridization in liver RNA was seen.
Biochemical Characterization of FKBP65 in the Murine NIH3T3 Cells-To determine if FKBP65 possessed PPIase activity, recombinant FKBP65 (rFKBP65) was expressed using the pET21d ϩ expression vector (20) and purified by RP-HPLC as described under "Experimental Procedures" (Fig. 3A, lane   2). The purified rFKBP65 was renatured and utilized in the coupled peptidylprolyl cis-trans-isomerase assay developed by Fischer et al. (32). The PPIase assay measures the cis to trans isomerization of the prolyl peptide bond in the peptide, Nsuccinyl-Ala-Ala-Pro-Phe-p-nitroanilide. Chymotrypsin readily cleaves the trans form of the peptide, releasing p-nitroanilide, which is quantitated spectrophotometrically (Fig. 3B). The initial burst observed in the progress curves is a result of the large percentage of the peptide that exists in the trans conformation at ambient temperature. When rFKBP65 is preincubated with the peptide substrate (Fig. 3B, curve a), the rate of cis to trans isomerization is significantly higher than the chymotrypsin control (Fig. 3B, curve e). Recombinant FKBP65 appears to be fully active, and we have calculated the k cat /K m of rFKBP65 (described under "Experimental Procedures") for the cis-trans isomerization of N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide of 0.64 ϫ 10 6 M Ϫ1 ⅐s Ϫ1 , which is comparable to the PPIase activity of the other FKBPs (21,22). Additionally, the PPIase activity of rFKBP65 is specifically inhibited by the addition of FK506 in a dose-dependent manner (Fig. 3B, curves b-d). Since the assay is not under pure Michaelis-Menten conditions, only an IC 50 rather than a K i can be obtained. An IC 50 value of 45 nM was

FIG. 2. Northern blot analysis of FKBP65 in mouse tissues.
Total RNA (20 g) from various mouse tissues were electrophoresed in a 1% agarose gel, blotted to nylon membranes, and hybridized with the 1.7-kb ORF cDNA probe.

FKBP65, A Novel Immunophilin
found for FK506 (data not shown). The PPIase activity of rFKBP65 is also potently inhibited in a dose-dependent manner by rapamycin (27-200 nM), but is unaffected by the addition of 5 M Cyclosporin A (data not shown).
To detect FKBP65 protein, two polyclonal antisera (Pep 2 and Pep 4) were raised in rabbits against synthetic peptides derived from the predicted sequence of FKBP65, as described under "Experimental Procedures." Although both peptide antisera gave similar results, we used Pep 4 antiserum, raised against the carboxyl terminus of FKBP65 (Fig. 1A, underlined sequence) because it reacted more strongly in immune precipitations and Western blots. Pep 4 antiserum was used to immunoprecipitate total cell lysates of [ 35 S]methionine-labeled NIH3T3 cells. Despite the calculated molecular mass of 64,683 or 60,576 for the NH 2 -terminal cleaved protein, two prominent bands migrating at approximately 72 and 68 kDa were observed (Fig. 4, lane 2) that were specifically competed when peptide was added during immune precipitation (Fig. 4, lane 3). The discrepancy between apparent and calculated molecular masses is observed with other FKBP family members (9,10,12,14,16,33).
The increase in apparent molecular weight may be due to the presence of post-translational modifications such as glycosylation and/or phosphorylation. The predicted amino acid sequence contains seven potential N-glycosylation sites and 14 potential casein kinase II phosphorylation sites. Additionally, FKBP65 contains one cAMP-dependent protein kinase phosphorylation site, one protein kinase C phosphorylation site, and two tyrosine kinase phosphorylation sites. Furthermore, FKBP65 contains nine proline-proline sequences that may alter the protein structure resulting in an altered migration on SDS-PAGE and increasing its apparent size.
To determine if FKBP65 was glycosylated, NIH3T3 cells were metabolically labeled with [ 35 S]methionine, precipitated with Pep 4 antiserum, and incubated for 3 h with N-endoglycosidase F (Endo-F; Fig. 4, lanes 4 and 5) or in Endo-F buffer alone (Fig. 4, lanes 1-3). The Endo-F-treated FKBP65 migrates at an apparent molecular weight of 58,000 (Fig. 4, lane 4). To confirm that the 58-kDa band was the unglycosylated form of FKBP65, Western blot analysis of Endo-F-treated lysates was performed (data not shown), and an immunoreactive 58-kDa band was visualized. Additionally, when NIH3T3 cells are pretreated with either 2-deoxyglucose or tunicamycin, agents that interfere with protein glycosylation, the unglycosylated 58-kDa species also was observed. These results demonstrate that FKBP65 is a glycoprotein.
The observance of two precipitated bands may reflect the difference between phosphorylated and unphosphorylated forms of FKBP65 or alternatively may represent the product from a different translational start site. There is a second methionine residue 60 amino acids downstream from the first ATG sequence. This second ATG contains a nearly perfect Kozak consensus sequence (26) and may encode a protein with a molecular weight of approximately 57,000; however, typically only a single 2.6-kb transcript is seen (Fig. 2). To determine if FKBP65 also is phosphorylated, NIH3T3 cells were labeled with [ 32 P]orthophosphate followed by immune precipitation with Pep 4 antiserum as described under "Experimental Procedures." The immune precipitates were visualized by autoradiography after SDS-PAGE and electroblotting onto Immo-bilonP TM . A specific band at a relative molecular mass of 72,000 (Fig. 5A, lane 2) was observed. The 32 P-labeled blot was allowed to decay prior to Western blot analysis (Fig. 5B), and alignment of the two autoradiographs confirms that the 32 Plabeled protein is FKBP65 and represents the top 72-kDa band observed during immune precipitation. DISCUSSION We have identified a mouse gene encoding a 65-kDa protein (FKBP65) that shares homology with members of the FK506binding protein class of immunophilins. In this study, we report the cloning, sequencing, and biochemical characterization of a unique 2636-bp clone isolated from a NIH3T3 fibroblast Zap cDNA library. The 2.6-kb FKBP65 clone contains an ORF encoding a protein with a predicted molecular mass of 64,683 daltons; however, the presence of a hydrophobic leader se- quence and a predicted cleavage site would result in a mature protein with a molecular mass of 60,576. Amino acid sequence comparison of FKBP65 to other FKBP family members revealed from 26% (human FKBP52) to 46% (hFKBP12) overall sequence homology. Additionally, FKBP65 contains four FK506-binding protein PPIase signature domains. The regions flanking the four PPIase domains retain 30% similarity and appear to be unique as they are not homologous to other FKBPs or any sequences reported in available data bases. The first three domains of FKBP65 conserve six of the seven amino acid residues reported to be involved with FK506-binding interactions.
We have expressed rFKBP65 in the pET21d ϩ expression vector and demonstrate that rFKBP65 has PPIase activity with kinetics similar to other FKBPs that is inhibitable by FK506 and rapamycin, but not Cyclosporin A. Although FKBP65 is approximately the same relative size of the larger, multidomain FKBP52 that associates with steroid receptors (16,17,(33)(34)(35), FKBP65 is distinct from FKBP52 that contains three FKBP12-like PPIase domains. Likewise, FKBP65 appears to be distinct from p50 and p54, two FKBP-related proteins recently shown to complex with the avian progesterone receptor (36,37). Collectively, these observations strongly suggest that FKBP65 is a new FKBP family member.
Another distinguishing factor of FKBP65 is the tissue distribution. A unique 2.6-kb mRNA band was found in lung, spleen, heart, and brain, but not in liver. This result suggests that FKBP65, unlike other FKBP family members, may not be ubiquitously expressed. The second, larger band detected in the testis may reflect a related protein or an alternate form of FKBP65. We have recently isolated a partial human clone 2 that, when used as a probe, detects a 3.5-kb band in human mRNA. Using a rabbit polyclonal peptide antibody, two specific bands (72 and 68 kDa) were detected by immunoprecipitation and Western blot analysis in NIH3T3 cell lysates. The appearance of two precipitable bands may be the product resulting from a second, downstream methionine or reflect a difference in phosphorylation states. When cells were labeled with [ 32 P]orthophosphate, only the upper, 72-kDa band was visualized; however, both bands are observed on Western blot analysis. Although FKBP65 appears to be a phosphoprotein, there is no identified ATP binding site within the sequence. We have shown that FKBP65 is a glycoprotein, an observation that can account for the difference observed between the calculated and apparent molecular weights; however, the unglycosylated form migrates at an apparent molecular weight of 58,000 that is smaller than the predicated molecular weight of 60,576. This additional difference may be due in part to the large number of proline residues (37 in the mature protein) including nine Pro-Pro sequences that may alter protein structure leading to an altered migration on SDS-PAGE.
The functional diversity of the FKBP family members may be due to their distinct subcellular localization and association with different protein complexes. The multidomain FKBP52 has been shown to associate with hsp90, hsp70, and the unactivated steroid receptors (16,17,33), whereas the unidomain FKBP25, which is localized in the nucleus, associates with casein kinase II and nucleolin (38). The smaller FKBPs, FKBP12 and FKBP13, have been localized to the cytoplasm (8) and the endoplasmic reticulum (39), respectively. FKBP12 has been shown to associate with the ryanodine receptor (40), and, recently, the inositol 1,4,5-trisphosphate receptor (41), both of which mediate calcium release processes. The predicted amino acid sequence of FKBP65 reveals a potential ER retention sequence and a possible membrane-spanning sequence, suggesting localization to the ER. Using various cellular fractionation protocols, we have localized FKBP65 to the membrane and cytosolic fractions; however, due to the presence of Nlinked oligosaccharides, the appearance of FKBP65 in the cytosolic fraction may be a result of inadequate fractionation and/or contamination from the ER fractions.
The presence of four PPIase domains of FKBP65 suggest that they may function to interact with more than one molecule within a complex of proteins or act stoichiometrically with the same molecule. We are currently investigating proteins that associate with FKBP65, using affinity-purified protein and NIH3T3 lysates with or without FK506 or rapamycin treatment. To identify the possible functions of FKBP65 and associated proteins, we are currently generating FKBP65 mutants to assess the functional roles of the four PPIase domains. Using the Pep 4 polyclonal antibody, immunoreactive FKBP65 also has been detected by immunoprecipitation and Western blotting techniques in human cell lines, 2 and we are currently in the process of isolating the human homologue of FKBP65. This will allow precise comparisons of the various human FK506binding proteins and possible detection of additional genes related to FKBP65.