FK506 Binding Protein Mutational Analysis
1995; Elsevier BV; Volume: 270; Issue: 32 Linguagem: Inglês
10.1074/jbc.270.32.18935
ISSN1083-351X
AutoresOlga Futer, Maureen T. DeCenzo, Robert A. Aldape, David J. Livingston,
Tópico(s)Cardiac Structural Anomalies and Repair
ResumoThe 12- and 13-kDa FK506 binding proteins (FKBP12 and FKBP13) are cis-trans peptidyl-prolyl isomerases that bind the macrolides FK506 (Tacrolimus) and rapamycin (Sirolimus). The FKBP12•FK506 complex is immunosuppressive, acting as an inhibitor of the protein phosphatase calcineurin. We have examined the role of the key surface residues of FKBP12 and FKBP13 in calcineurin interactions by generating substitutions at these residues by site-directed mutagenesis. All mutants are active catalysts of the prolyl isomerase reaction, and bind FK506 or rapamycin with high affinity. Mutations at FKBP12 residues Asp-37, Arg-42, His-87, and Ile-90 decrease calcineurin affinity of the mutant FKBP12•FK506 complex by as much as 2600-fold in the case of I90K. Replacement of three FKBP13 surface residues (Gln-50, Ala-95, and Lys-98) with the corresponding homologous FKBP12 residues (Arg-42, His-87, and Ile-90) generates an FKBP13 variant that is equivalent to FKBP12 in its affinity for FK506, rapamycin, and calcineurin. These results confirm the role of two loop regions of FKBP12 (residues 40-44 and 84-91) as part of the effector face that interacts with calcineurin. The 12- and 13-kDa FK506 binding proteins (FKBP12 and FKBP13) are cis-trans peptidyl-prolyl isomerases that bind the macrolides FK506 (Tacrolimus) and rapamycin (Sirolimus). The FKBP12•FK506 complex is immunosuppressive, acting as an inhibitor of the protein phosphatase calcineurin. We have examined the role of the key surface residues of FKBP12 and FKBP13 in calcineurin interactions by generating substitutions at these residues by site-directed mutagenesis. All mutants are active catalysts of the prolyl isomerase reaction, and bind FK506 or rapamycin with high affinity. Mutations at FKBP12 residues Asp-37, Arg-42, His-87, and Ile-90 decrease calcineurin affinity of the mutant FKBP12•FK506 complex by as much as 2600-fold in the case of I90K. Replacement of three FKBP13 surface residues (Gln-50, Ala-95, and Lys-98) with the corresponding homologous FKBP12 residues (Arg-42, His-87, and Ile-90) generates an FKBP13 variant that is equivalent to FKBP12 in its affinity for FK506, rapamycin, and calcineurin. These results confirm the role of two loop regions of FKBP12 (residues 40-44 and 84-91) as part of the effector face that interacts with calcineurin. The immunophilins are the intracellular high affinity receptors for the potent immunosuppressive agents cyclosporin, FK506, and rapamycin. The structure and function of the immunophilins have been investigated intensely since these proteins were first identified in thymus and T-cell extracts(1Armistead D.M. Harding M.W. Annu. Rep. Med. Chem. 1993; 28: 207-215Google Scholar, 2Fruman D.A. Burakoff S.J. Bierer B.E. FASEB J. 1994; 8: 391-400Crossref PubMed Scopus (236) Google Scholar, 3Galat A. Eur. J. Biochem. 1993; 216: 689-707Crossref PubMed Scopus (316) Google Scholar). The immunophilins FKBP12 1The abbreviations used are:FKBP12the 12-kDa FK506 binding proteinFKBP13the 13-kDa FK506 binding proteinGSTglutathione S-transferase. 1The abbreviations used are:FKBP12the 12-kDa FK506 binding proteinFKBP13the 13-kDa FK506 binding proteinGSTglutathione S-transferase. and FKBP13 are members of the multigene FK506-binding protein family, encoded by distinct genes (4DiLella A.G. Craig R.J. Biochemistry. 1991; 30: 8512-8517Crossref PubMed Scopus (19) Google Scholar, 5Hendrickson B.A. Zhang W. Craig R.J. Jin Y.-J. Bierer B.E. Burakoff S. DiLella A.G. Gene (Amst.). 1993; 134: 271-275Crossref PubMed Scopus (13) Google Scholar, 6Peattie D.A. Hsaio K. Benasutti M. Lipke J.A. Gene (Amst.). 1994; 150: 251-257Crossref PubMed Scopus (9) Google Scholar) and performing distinct cellular functions. FKBP12 and FKBP13 are homologous proteins of molecular mass 12 and 13 kDa respectively, with 43% amino acid sequence identity and 51% nucleotide sequence identity (Fig. 1)(7Jin Y.-J. Albers M.W. Lane W.S. Bierer B.E. Schreiber S.L. Burakoff S.J. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 6677-6681Crossref PubMed Scopus (156) Google Scholar). Upon binding of FK506 to FKBP12 in the T-cell, cytosolic signaling events are interrupted, resulting in inhibition of interleukin-2 expression during T-cell activation(8Sigal N.H. Dumont F. Annu. Rev. Immunol. 1992; 10: 519-560Crossref PubMed Scopus (657) Google Scholar). Inhibition of this signal transduction pathway is mediated through binding of the FKBP12•FK506 complex to the protein calcineurin (Ki = 6 nM)(9Aldape R.A. Futer O. DeCenzo M.T. Jarrett B.P. Murcko M.A. Livingston D.J. J. Biol. Chem. 1992; 267: 16029-16032Abstract Full Text PDF PubMed Google Scholar), a Ca2+-activated, calmodulin-dependent serine/threonine phosphatase(10Liu J. Farmer Jr., J.D. Lane W.S. Friedman J. Weissman I. Schreiber S.L. Cell. 1991; 66: 807-815Abstract Full Text PDF PubMed Scopus (3614) Google Scholar). Human FKBP13 is a membrane-associated protein that contains a 21-residue N-terminal signal peptide(7Jin Y.-J. Albers M.W. Lane W.S. Bierer B.E. Schreiber S.L. Burakoff S.J. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 6677-6681Crossref PubMed Scopus (156) Google Scholar). Gene expression and immunolocalization studies of FKBP13 indicate that this protein may act as an endoplasmic reticulum chaperone and that its expression is induced by cellular stress such as heat shock(11Partaledis J.A. Berlin V. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 5450-5454Crossref PubMed Scopus (85) Google Scholar, 12Nigam S.K. Jin Y.-J. Jin M.-J. Bush K.T. Bierer B.E. Burakoff S.J. Biochem. J. 1993; 294: 511-515Crossref PubMed Scopus (54) Google Scholar). Its FK506 complex does not inhibit calcineurin, as demonstrated herein. the 12-kDa FK506 binding protein the 13-kDa FK506 binding protein glutathione S-transferase. the 12-kDa FK506 binding protein the 13-kDa FK506 binding protein glutathione S-transferase. The FKBPs share the common activity of catalyzing the cis-trans isomerization of proline amide bonds of polypeptide substrates (the peptidyl-prolyl isomerase reaction), but beyond this, their biochemical properties diverge. Different substrate specificities are observed for catalysis of the peptidyl-prolyl isomerase reaction as well as in binding affinity for FK506 and rapamycin(13Peattie D.A. Harding M.W. Fleming M.A. DeCenzo M.T. Lippke J.A. Livingston D.J. Benasutti M. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 10974-10978Crossref PubMed Scopus (230) Google Scholar). High resolution three-dimensional structures have been determined of FKBP12(14Moore J.M. Peattie D.A. Fitzgibbon M.J. Thomson J.A. Nature. 1991; 351: 248-250Crossref PubMed Scopus (136) Google Scholar, 15Wilson K.P. Yamashita M.M. Sintchak M.D. Rotstein S.H. Murcko M.A. Boger J. Thomson J.A. Fitzgibbon M.J. Black J.R. Navia M.N. Acta Crystallogr. 1995; (in press)Google Scholar), FKBP12 in complex with FK506 and rapamycin(15Wilson K.P. Yamashita M.M. Sintchak M.D. Rotstein S.H. Murcko M.A. Boger J. Thomson J.A. Fitzgibbon M.J. Black J.R. Navia M.N. Acta Crystallogr. 1995; (in press)Google Scholar, 16Van Duyne G.D. Standaert R.F. Karplus P.A. Schreiber S.L. Clardy J. J. Mol. Biol. 1993; 229: 105-124Crossref PubMed Scopus (1087) Google Scholar), and the FKBP13•FK506 complex(17Schultz L.W. Martin P.K. Liang J. Schreiber S.L. Clardy J. J. Am. Chem. Soc. 1994; 116: 3129-3130Crossref Scopus (32) Google Scholar). A network of hydrophobic interactions and hydrogen bonds from macrolide ligand to protein side chain and backbone heteroatoms are known to stabilize these complexes. A ternary structure at atomic resolution of the FKBP12• FK506•calcineurin complex is not yet available. However, by studying the effects of mutations on the affinity of FKBP12 for calcineurin, we reported the first direct evidence that the calcineurin-inhibitory binding surface of the FKBP12•FK506 complex must include the Arg-42, His-87(9Aldape R.A. Futer O. DeCenzo M.T. Jarrett B.P. Murcko M.A. Livingston D.J. J. Biol. Chem. 1992; 267: 16029-16032Abstract Full Text PDF PubMed Google Scholar), and Ile-90 2D. J. Livingston, R. A. Aldape, O. Futer, M. T. DeCenzo, B. P. Jarrett, and M. A. Murcko, poster presented at New York Academy of Sciences Conference on “Immunomodulating Drugs,” June 27, 1992. 2D. J. Livingston, R. A. Aldape, O. Futer, M. T. DeCenzo, B. P. Jarrett, and M. A. Murcko, poster presented at New York Academy of Sciences Conference on “Immunomodulating Drugs,” June 27, 1992. surface residues of the protein itself, a finding later confirmed by other groups(18Yang D. Rosen M.K. Schreiber S.L. J. Am. Chem. Soc. 1993; 115: 819-820Crossref Scopus (97) Google Scholar, 19Becker J.W. Rotonda J. McKeever B.M. Chan H.K. Marcy A.I. Wiederrecht G. Hermes J.D. Springer J.P. J. Biol. Chem. 1993; 268: 11335-11339Abstract Full Text PDF PubMed Google Scholar). Yang et al.(18Yang D. Rosen M.K. Schreiber S.L. J. Am. Chem. Soc. 1993; 115: 819-820Crossref Scopus (97) Google Scholar) raised questions about the role of Arg-42 in calcineurin binding versus its role in stabilization of local structure. In this investigation, we have further explored the effects of FKBP12 surface mutations and we have examined the role of the homologous surface residues of FKBP13 in calcineurin inhibition. Interpretation of these data was aided by the availability of the high resolution structures of several of these mutant FKBPs solved by high field NMR (20Lepre C.A. Pearlman D.A. Cheng J.-W. DeCenzo M.T. Livingston D.J. Moore J.M. Biochemistry. 1994; 33: 13571-13580Crossref PubMed Scopus (12) Google Scholar) and x-ray crystallography 3S. Itoh and M. A. Navia, submitted for publication. 3S. Itoh and M. A. Navia, submitted for publication.(21Itoh S. DeCenzo M.T. Livingston D.J. Pearlman D.A. Navia M.N. Bioorg. Med. Chem. Lett. 1995; 5 (in press)Crossref PubMed Scopus (29) Google Scholar). We report here the separable effects of these FKBP surface mutations on binding interactions with peptidyl-prolyl isomerase substrates, macrolides, and calcineurin. Restriction enzymes and polynucleotide kinase were from New England Biolabs. Sequenase™174; and other sequencing reagents were from U. S. Biochemical Corp. Phagemid, helper phage, the Muta-Gene kit, and protein assay reagents were from Bio-Rad. Alkaline phosphatase was from Boehringer Mannheim; Gene-Clean kits were from Bio-101; and Sep-Pak cartridges were from Millipore Corp. Peptide substrates for peptidyl-prolyl isomerase assay were from Bachem Biosciences (Philadelphia, PA), and the calcineurin peptide substrate was from Penninsula Labs. Glutathione-Sepharose 4B was purchased from Pharmacia Biotech Inc. FK506 and rapamycin were kindly provided by Chugai Pharmaceuticals. For site-directed mutagenesis, the dut−ung−Escherichia coli strain CJ236 was used for uracil enrichment of single strand DNA, and the Nova Blue strain (NovaGene) was used for selection of heteroduplex DNA after extension-ligation reactions. The CAG626 host strain, obtained from C. Gross (University of Wisconsin, Madison, WI) was used for expression of FKBP12 mutants. The XA90 strain was used as a host for expression of recombinant glutathione S-transferase (GST) FKBP13 fusion proteins. A pKEN2 phagemid (G. Verdine, Harvard University, Cambridge, MA) was used for site-directed mutagenesis. FKBP12 proteins were expressed in this vector as described previously(22Park S.T. Aldape R.A. Futer O. DeCenzo M.T. Livingston D.J. J. Biol. Chem. 1992; 267: 3316-3324Abstract Full Text PDF PubMed Google Scholar). The pGEX-2T phagemid (Pharmacia) was used for expression of fusion GST-FKBP13 proteins. This vector is a derivative of pGEX-1 vector, which includes an isopropyl-1-thio-β-D-galactopyranoside-inducible tac promoter, complete coding sequence of glutathione S-transferase followed by oligonucleotides encoding the cleavage-recognition sequence of thrombin, and a polylinker containing unique recognition sites for BamHI, SmaI, and EcoRI. The phagemid also contains an amprgene and a lacIq allele of the lac repressor(23Smith D.B. Johnson K.S. Gene (Amst.). 1988; 67: 31-40Crossref PubMed Scopus (5046) Google Scholar). A 362-base pair BamHI fragment derived from a cDNA encoding mature human FKBP13 and a 431-base pair EcoRI fragment encoding human FKBP12 cDNA were ligated into phagemid pKEN2. All oligonucleotide-directed mutagenesis was performed on the pKEN2•FKBP12 and pKEN2•FKBP13 constructs using uracil-enrichment of single strand DNA by the modification of Kunkel (24Kunkel T.A. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 488-492Crossref PubMed Scopus (4900) Google Scholar, 25Kunkel T.A. Roberts J. Zakour R. Methods Enzymol. 1987; 154: 367-382Crossref PubMed Scopus (4558) Google Scholar) of the method originally described for M13 mutagenesis(26Zoller M.J. Smith M. Nucleic Acids Res. 1982; 10: 6487-6500Crossref PubMed Scopus (467) Google Scholar, 27Zoller M.J. Smith M. Methods Enzymol. 1983; 100: 468-500Crossref PubMed Scopus (661) Google Scholar). Two-primer mutagenesis was performed to construct the double mutants. All mutants were sequenced by dideoxy method (28Sanger F. Nicklen S. Coulson A.R. Proc. Natl. Acad. Sci. U. S. A. 1977; 74: 5463-5467Crossref PubMed Scopus (52649) Google Scholar) from the promoter through the region surrounding the mutagenesis site, and most were sequenced in their entirety in the coding region. E. coli strain CAG626 (a lon− strain on a SC122 background) was used as host for expression of FKBP12 mutants in the pKEN2 expression vector(22Park S.T. Aldape R.A. Futer O. DeCenzo M.T. Livingston D.J. J. Biol. Chem. 1992; 267: 3316-3324Abstract Full Text PDF PubMed Google Scholar). For expression of FKBP13, a cDNA encoding the 13-kDa mature form was used for mutagenesis and expression. cDNA inserts carrying a confirmed mutation were cut out of the pKEN2 vector by BamHI restriction digest, purified from the agarose gel using Gene-Clean kit, and subcloned into BamHI site of the pGEX-2T vector. The 5′ ends of the BamHI-cut vector were dephosphorylated with calf intestine alkaline phosphatase. Orientation of the insert was determined by restriction digest with PstI. Positive clones were sequenced in their entirety. The GST-fusion expression system allowed us to obtain expression levels of 3-12 mg of the GST-FKBP13 chimera/g of cell paste. Mutant duplex DNA was transformed into XA90 cells for expression and plated on LB plates containing 100 μg/ml ampicillin. Five-ml overnight cultures were started from multiple colonies and grown at 37°C on LB-ampicillin medium. For expression, cultures were grown at 37°C in 2-liter shake flasks or in a 10-liter fermentor (Biostat ED, B. Braun). Culture medium (LB-ampicillin) was inoculated at a density of 1:500 and grown to A600 of 0.5-1.0. Protein expression was then induced by the addition of isopropyl-1-thio-β-D-galactopyranoside at 1 mM final concentration. Cells were then grown for 15-19 h, and harvested by centrifugation at 4230 × g for 15 min at 4°C. Cell pastes were washed in 0.1 M Tris, pH 7.4, and frozen at −70°C. Purification of Wild-type and Mutant Proteins-The purification scheme used for the wild-type and mutant FKBP12 has been described previously(22Park S.T. Aldape R.A. Futer O. DeCenzo M.T. Livingston D.J. J. Biol. Chem. 1992; 267: 3316-3324Abstract Full Text PDF PubMed Google Scholar). Wild-type and mutant FKBP13 were purified by a procedure modified from Smith and Johnson(23Smith D.B. Johnson K.S. Gene (Amst.). 1988; 67: 31-40Crossref PubMed Scopus (5046) Google Scholar). Cell paste (9-12 g) was resuspended in 10-15 ml of phosphate-buffered saline buffer containing 1% Triton X-100, incubated with lysozyme (0.4 mg/g of cell paste, 10 mg/ml stock solution) for 15-20 min at room temperature, sonicated 3 × 30 s at 20% with microtip sonicator, and spun for 15 min at 31,000 × g. Supernatant fractions were transferred to a clean centrifuge tube and spun again to remove residual particulates. The cell pellet was again resuspended in a smaller volume of buffer, sonicated once for 30 s, and centrifuged. The resulting supernatant fractions were pooled. The purification was monitored by peptidyl-prolyl isomerase enzyme activity, using FK506-inhibitable activity to distinguish FKBP13 activity from endogenous cyclophilin activity. All subsequent purification steps were performed at 4°C unless otherwise indicated. The clarified cell lysate was loaded on glutathione-Sepharose column (20 ml bed volume) preequilibrated with phosphate-buffered saline/Triton buffer. The column was washed with at least 10 bed volumes of phosphate-buffered saline and/or until A280 returned to base line. GST fusion protein was eluted with 5 mM glutathione in 50 mM Tris buffer, pH 8.0, concentrated, and dialyzed overnight versus 50 mM Tris, pH 8.0. The >50% pure protein was comprised of two major bands on SDS-polyacrylamide gel electrophoresis (Fig. 2) corresponding to fusion protein (43 kDa) and free GST (29 kDa). The protein was then subjected to cleavage by thrombin (1.2 NIH units/mg of fusion protein) by incubating at room temperature for 1 h in 50 mM Tris, pH 8.0, 150 mM NaCl, 2.5 mM CaCl2 as described in (23Smith D.B. Johnson K.S. Gene (Amst.). 1988; 67: 31-40Crossref PubMed Scopus (5046) Google Scholar). The reaction was stopped by addition of phenylmethylsulfonyl fluoride and EGTA to 0.025 and 2.0 mM final concentration, respectively. Thrombin-cleaved protein was loaded on a glutathione-Sepharose column (20 ml bed volume) equilibrated with 50 mM Tris, pH 8.0; the column was washed with the same buffer, and free FKBP was collected in the flow-through. The remaining GST was retained on the column, while FKBP13 was collected in the flow-through fractions. To remove remaining contaminants, the sample was purified further by hydrophobic interaction chromatography. Active fractions were pooled and concentrated to ~10 ml and brought to 2 M potassium phosphate, pH 7.0. The sample was then loaded onto a 10 × 100 mm hydrophobic interaction chromatography column (Hydropore-HIC, Rainin) mounted on a Waters 650E HPLC. The conditions used were as described previously(21Itoh S. DeCenzo M.T. Livingston D.J. Pearlman D.A. Navia M.N. Bioorg. Med. Chem. Lett. 1995; 5 (in press)Crossref PubMed Scopus (29) Google Scholar), except we used a 2-h gradient. All purified proteins were stored in 50 mM Tris, pH 8.0, at 4°C. 1% polyethylene glycol(8000) was added for longer storage of proteins at −70°C. Measurement of catalytic efficiency (kcat/Km) for the peptidyl-prolyl isomerase reaction was performed essentially according to Harrison and Stein (29Harrison R.K. Stein R.L. Biochemistry. 1990; 29: 1684-1689Crossref PubMed Scopus (166) Google Scholar) with the modifications described previously (22Park S.T. Aldape R.A. Futer O. DeCenzo M.T. Livingston D.J. J. Biol. Chem. 1992; 267: 3316-3324Abstract Full Text PDF PubMed Google Scholar). The substrate used for determinations of kcat/Km and inhibition constants (Ki) was Suc-Ala-Leu-Pro-Phe-pNA. kcat/Km and Ki were measured at enzyme concentrations such that kobs was at least 5-fold higher than knon-enz. Stock solutions of FK506 and rapamycin were prepared in Me2SO. Final Me2SO concentration in the peptidylprolyl isomerase assay was 0.5% (v/v). Reactions were initiated with chymotrypsin (Sigma) at a final concentration of 50 μg/ml. Absorbance at 400 nm was monitored at 1-s intervals on a Hewlett-Packard 8452A spectrophotometer with a thermostatted cuvette holder, interfaced to a model 300 Hewlett-Packard computer. The calcineurin assay was performed essentially as described by Klee and Cohen(30Klee C.B. Cohen P. Mol. Aspects Cell Regul. 1988; 5: 225-248Google Scholar). A commercial preparation of bovine brain calcineurin was used (Sigma, catalog no. C-1907, specific activity = 16 nmol/min/mg under the conditions of the assay). Radiolabeled phosphorylated peptide substrate, derived from the serine phosphorylation site sequence of the RII subunit of cAMP-dependent protein kinase, was prepared as described previously(9Aldape R.A. Futer O. DeCenzo M.T. Jarrett B.P. Murcko M.A. Livingston D.J. J. Biol. Chem. 1992; 267: 16029-16032Abstract Full Text PDF PubMed Google Scholar). The serine phosphatase assay was performed in 60 μl of buffer containing 20 mM Tris, pH 8.0, 0.1 M NaCl, 6 mM MgCl2, 0.1 mM CaCl2, 0.5 mM dithiothreitol, and 0.1 mg/ml bovine serum albumin(30Klee C.B. Cohen P. Mol. Aspects Cell Regul. 1988; 5: 225-248Google Scholar). The following ordered additions were made for the assays: 5 nM to 15 μM FKBP, 5 nM to 15 μM FK506, 160 nM bovine calmodulin (Sigma, catalog no. P-2277), and 40 nM bovine brain calcineurin. 32P-Labeled phosphorylated peptide was added to 1-2 μM final concentration, followed by a 15-min incubation at 30°C. Reactions were quenched with 540 μl of 0.1 M potassium phosphate, 5% trichloracetic acid (w/v). Cation-exchange columns (Dowex AG1-X8, 0.6 ml) were used for separation of free 32Pi(31Hubbard M.J. Klee C.B. Chad J. Wheal H. Molecular Neurobiology: A Practical Approach. Oxford University Press, Oxford, United Kingdom1991: 135-149Google Scholar). The quenched reaction mixtures (0.6 ml) were applied to the columns, followed by a 0.6-ml H20 wash, and the effluents were collected in scintillation vials and counted with 5 ml of scintillation mixture (Beckmann Liquiscint). All assays were performed in duplicate. Affinity of the mutant FKBP•FK506 complexes for calcineurin was determined by varying the concentrations of mutant FKBP and FK506 at 30°C, using a drug/FKBP ratio of 1.35:1. FK506/FKBP ratios were increased appropriately for the lower affinity mutants to ensure saturation of the mutant FKBP with drug. Control reactions at 30°C were run in the presence of the same concentration of Me2SO (0.1% (v/v)) as reactions containing FK506. For the peptidyl-prolyl isomerase reaction data, absorbance data points were fit to a first-order function using Hewlett-Packard data analysis software. Ki values were calculated using nonlinear fitting to a competitive tight-binding equation (32Morrison J.F. Biochim. Biophys. Acta. 1969; 185: 269-286Crossref PubMed Scopus (725) Google Scholar) using KineTic software version 3.0 (BioKin, Ltd.) running on a Macintosh IIcx. An example of a fit of Ki data is shown in Fig. 3A. In control experiments with the homogeneous major isoform (Aα) of bovine calcineurin (provided by M. Fitzgibbon and J. Thomson, Vertex), we developed an active site titration method for determining active calcineurin concentration. We observed identical inhibition constants (Kic = 5.5-6 nM) for the commercial bovine and homogenous Aα isoform preparations. The inhibition constant for calcineurin by the mutant FKBP•FK506 complexes (Kic) was calculated by computer-fitting the fractional inhibition data as a function of concentration of free FKBP mutant and FK506 to an the equilibrium equation derived by Liu et al.(33Liu J. Albers M.W. Wandless T.J. Luan S. Alberg D.G. Belshaw P.J. Cohen P. MacKintosh C. Klee C.B. Schreiber S.L. Biochemistry. 1992; 31: 3896-3901Crossref PubMed Scopus (501) Google Scholar). Quadratic equations were first used to calculate the free concentrations of these reaction components from the concentrations of calcineurin, FKBP mutant, and FK506 in the experiment, as well as the Ki of the FKBP mutant for FK506. A rearranged form of Equation 7 from (33Liu J. Albers M.W. Wandless T.J. Luan S. Alberg D.G. Belshaw P.J. Cohen P. MacKintosh C. Klee C.B. Schreiber S.L. Biochemistry. 1992; 31: 3896-3901Crossref PubMed Scopus (501) Google Scholar) was then used to calculate the calcineurin affinity of the FKBP mutant•FK506 complex: I/(1 − I) = [FK506]free[FKBP12]free/(KiKic), where I is the fractional inhibition of calcineurin, and (1 − I) is the fractional activity remaining. Kic and the associated standard deviation were calculated from linear regressions performed on MiniTab (Addison-Wesley). A replot of calcineurin inhibition data to the above equation is shown in Fig. 3B. Human FKBP13 is a membrane-associated protein that contains a 21-residue N-terminal signal peptide(7Jin Y.-J. Albers M.W. Lane W.S. Bierer B.E. Schreiber S.L. Burakoff S.J. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 6677-6681Crossref PubMed Scopus (156) Google Scholar). The cDNA fragment of FKBP13 that we used for mutagenesis and expression did not contain the sequence for the signal peptide and thus produced a mature form of the protein of 13 kDa molecular mass. A GST-fusion expression system allowed us to obtain expression levels of 3-12 mg of the GST-FKBP13 chimera/g of cell paste. Induction of the cell culture at higher optical density (A600 of 0.9-1.1) with isopropyl-1-thio-β-D-galactopyranoside helped to increase expression level. We used a purification scheme for FKBP13 proteins that was modified from one developed for isolation of foreign polypeptides as GST fusion proteins(23Smith D.B. Johnson K.S. Gene (Amst.). 1988; 67: 31-40Crossref PubMed Scopus (5046) Google Scholar). Affinity chromatography on a glutathione-Sepharose column allowed us to obtain fusion protein that was >50% homogeneous in a single purification step. Two major bands of mobilities corresponding to 43 and 29 kDa were visualized on the Coomassie Blue-stained SDS-polyacrylamide gel electrophoresis (Fig. 2), corresponding to the fusion protein and free GST, respectively. The same glutathione Sepharose column was used in a second purification step after thrombin cleavage of the fusion protein. Most of the GST was retained on the column, while GST-free FKBP 13 was collected in the flow-through fractions. However, the protein obtained at this step was still slightly contaminated with free GST and other minor protein species. To remove these impurities, we used hydrophobic interaction chromatography as a final purification step. FKBP13 mutants resulting from this purification scheme were ≥95% pure as determined by densitometry of SDS-polyacrylamide gel electrophoresis (Fig. 2). Each FKBP12 mutant generated in this investigation was also purified to homogeneity by methods described previously(9Aldape R.A. Futer O. DeCenzo M.T. Jarrett B.P. Murcko M.A. Livingston D.J. J. Biol. Chem. 1992; 267: 16029-16032Abstract Full Text PDF PubMed Google Scholar, 22Park S.T. Aldape R.A. Futer O. DeCenzo M.T. Livingston D.J. J. Biol. Chem. 1992; 267: 3316-3324Abstract Full Text PDF PubMed Google Scholar). Site-directed mutagenesis was used to make conservative mutations of FKBP12 binding pocket and surface residues (Fig. 1). All FKBP12 and FKBP13 mutants produced in this investigation are active catalysts of the peptidyl-prolyl isomerase reaction. Mutation of two hydrophilic residues in the solvent-exposed interface of the binding pocket cause more significant decreases in peptidyl-prolyl isomerase activity. The D37V mutant has 10% of the catalytic efficiency of the wild-type protein. The R42I mutant has approximately 60% of the specific activity of the wild-type protein, while the other Arg-42 mutants affect activity less. Of the His-87 mutations, only H87L significantly decreases the catalytic efficiency. Specific activities of FKBP13 variants range from 39 to 110% of the wild-type protein (Table1). Only a few of these mutations, however, cause greater than 2-fold decreases in the catalytic efficiency (kcat/Km) of this reaction.Table I Open table in a new tab The D37V surface residue mutant has an affinity for FK506 and rapamycin 580- and 180-fold lower, respectively, than the wild-type protein (the most dramatic change in drug affinity caused by a point mutation that we have observed). This effect is likely due to disruption of the electrostatic interaction formed by Asp-37 carboxylate oxygen and the Arg-42 guanidino NH in the FK506 and rapamycin complexes(15Wilson K.P. Yamashita M.M. Sintchak M.D. Rotstein S.H. Murcko M.A. Boger J. Thomson J.A. Fitzgibbon M.J. Black J.R. Navia M.N. Acta Crystallogr. 1995; (in press)Google Scholar, 16Van Duyne G.D. Standaert R.F. Karplus P.A. Schreiber S.L. Clardy J. J. Mol. Biol. 1993; 229: 105-124Crossref PubMed Scopus (1087) Google Scholar). The R42I and R42K mutants have unaltered affinities for FK506, while R42I has a 20-fold lower affinity for rapamycin than the wild-type protein. However, mutation of this residue to Gln, the residue found at the homologous position in FKBP13, causes a greater loss in affinity for these ligands (Table1). Mutation of the His-87 surface residue to either Phe, Val, or Ala has little effect on the specific activity or affinity of FKBP12 for ligands. However, mutation of this residue to Leu decreases FK506 affinity 40-fold and rapamycin affinity 100-fold. Mutations at the Ile-90 surface residue of FKBP12 have minimal effects on drug binding. We have reported that mutation of FKBP12 surface residue His-87 to Val decreases calcineurin affinity 4-fold(9Aldape R.A. Futer O. DeCenzo M.T. Jarrett B.P. Murcko M.A. Livingston D.J. J. Biol. Chem. 1992; 267: 16029-16032Abstract Full Text PDF PubMed Google Scholar). Subsequently we generated three more mutations at this residue and observed that each of them decreases calcineurin binding by no more than 2-fold (Table1). We do not observe inhibition of calcineurin by the wild-type FKBP13•FK506 complex at concentrations up to 10 μM. To determine whether mutation of the corresponding residue of FKBP13 to His confers calcineurin affinity to that protein, we expressed and purified the single FKBP13 A95H mutant. Its FK506 complex does not inhibit calcineurin. However, we determined that mutation of two other FKBP13 residues to their FKBP12 counterparts (Q50R and K98I) generate potent inhibitors of calcineurin when complexed with FK506 (Ki values of 100 and 23 nM, respectively), representing a gain of at least 3 orders of magnitude in calcineurin affinity relative to the wild-type protein. The reverse mutation of FKBP12, I90K, results in 2600-fold loss in calcineurin affinity without affecting affinity of this mutant to either FK506 or rapamycin. Combination of the two FKBP13 mutations in the double mutant Q50R/K98I results in even tighter binding of the FK506 complex to calcineurin (Ki = 16 nM). It is interesting to note that combination of the mutation A95H with either Q50R or with the double Q50R/K98I mutant, increases calcineurin affinity in each case. The FK506 complex of the triple mutant Q50R/A95H/K98I FKBP13 binds calcineurin as tightly (within experimental error) as does wild-type FKBP12. Structure-activity data with FK506 analogs in immunosuppression assays (19Becker J.W. Rotonda J. McKeever B.M. Chan H.K. Marcy A.I. Wiederrecht G. Hermes J.D. Springer J.P. J. Biol. Chem. 1993; 268: 11335-11339Abstract Full Text PDF PubMed Google Scholar, 34Bierer B.E. Somers P.K. Wandless T.J. Burakoff S.J. Schreiber S.L. Science. 1990; 250: 556-559Crossref PubMed Scopus (298) Google Scholar) led to the proposal that the macrolide contained effector elements that are in direct contact with a target protein and responsible for inhibition or antagonism of this protein. Such data, along with the three-dimensional structure of the FKBP12•FK506 complex(35Van Duyne G.D. Standaert R.F. Karplus P.A. Schreiber S.L. Clardy J. Science. 1991; 252: 839-842Crossref PubMed Scopus (579) Google Scholar), enabled identification of the probable effector region of FK506 as the cyclohexyl ring (C26-C34) and the C18-C23 solvent-exposed portion encompassing the C21 allyl group. According to the original proposal, FKBP12 acts as a “presenter” of the immunophilin ligand(36Schreiber S.L. Science. 1992; 251: 283-287Crossref Scopus (1345) Google Scholar). Based on our preliminary analysis of FKBP12 mutants as calcineurin inhibitors, we concluded that the actual effector face was a composite of the protein and ligand and must include several FKBP12 protein side chains(9Aldape R.A. Futer O. DeCenzo M.T. Jarrett B.P. Murcko M.A. Livingston D.J. J. Biol. Chem. 1992; 267: 16029-16032Abstract Full Text PDF PubMed Google Scholar). This modified hypothesis was later supported and extended (18Yang D. Rosen M.K. Schreiber S.L. J. Am. Chem. Soc. 1993; 115: 819-820Crossref Scopus (97) Google Scholar, 19Becker J.W. Rotonda J. McKeever B.M. Chan H.K. Marcy A.I. Wiederrecht G. Hermes J.D. Springer J.P. J. Biol. Chem. 1993; 268: 11335-11339Abstract Full Text PDF PubMed Google Scholar), and there is now consensus that the so-called “80s loop” (residues 84 − 91) of FKBP12 is an important binding determinant for calcineurin(1Armistead D.M. Harding M.W. Annu. Rep. Med. Chem. 1993; 28: 207-215Google Scholar, 37Clardy J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 56-61Crossref PubMed Scopus (75) Google Scholar). Mutation of the homologous Lys-98 residue of FKBP13 to Ile imparts very high calcineurin affinity to the FK506 complex of that protein. A similar result was reported by Rosen et al.(38Rosen M.K. Yang D. Martin P.K. Schreiber S.L. J. Am. Chem. Soc. 1993; 115: 821-822Crossref Scopus (50) Google Scholar), who argued for the requirement of a simultaneous Pro-97 to Gly mutation to impart high affinity, due to a predicted distortion in the 80s loop of FKBP13. We show here that mutation of Lys-98 alone generates a higher affinity calcineurin inhibitor (Ki = 23 nM) than the double P97G/K98I FKBP13 mutant reported in that work (Ki = 44 nM). A high resolution structure of the K98I mutation in the FKBP13 Q50R/A95H/K98I triple mutant indicates that no such loop distortion is present in the FK506 complex. 4J. P. Griffith, K. P. Wilson, O. Futer, D. J. Livingston, and M. A. Navia, manuscript in preparation. We have also determined that a single mutation of the corresponding Lys-121 residue of FKBP52 to Ile, generates a protein whose FK506 complex has high calcineurin affinity (Ki = 90 nM), whereas the co-complex of wild-type FKBP52 has no endogenous calcineurin affinity(13Peattie D.A. Harding M.W. Fleming M.A. DeCenzo M.T. Lippke J.A. Livingston D.J. Benasutti M. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 10974-10978Crossref PubMed Scopus (230) Google Scholar). The mutagenesis results reported herein demonstrate that the peptidyl-prolyl isomerase catalytic efficiency of FKBP12 is relatively insensitive to mutation at the active site. All mutants generated in this investigation are catalytically active. We reported previously a 4-fold decrease in calcineurin affinity of the FK506•FKBP12 H87V complex, with no significant decrease in FK506 and rapamycin affinity (9Aldape R.A. Futer O. DeCenzo M.T. Jarrett B.P. Murcko M.A. Livingston D.J. J. Biol. Chem. 1992; 267: 16029-16032Abstract Full Text PDF PubMed Google Scholar). It was later proposed by Rosen et al.(38Rosen M.K. Yang D. Martin P.K. Schreiber S.L. J. Am. Chem. Soc. 1993; 115: 821-822Crossref Scopus (50) Google Scholar), that the β-branching of the Val side chain should force this residue to occupy a conformation that would introduce strain into the 80s loop. We show here that the FK506 complex of the H87V FKBP12 mutant indeed suffers a greater decrease in calcineurin binding affinity than the FK506 complexes of H87L, H87F, or H87A. Structural analysis of the FK506•FKBP12 H87V complex by x-ray crystallography (21Itoh S. DeCenzo M.T. Livingston D.J. Pearlman D.A. Navia M.N. Bioorg. Med. Chem. Lett. 1995; 5 (in press)Crossref PubMed Scopus (29) Google Scholar) confirms that the branched Val side chain makes van der Waals' contact with Tyr-82, distorting its position relative to that observed in the wild-type protein. This in turn shifts the position of the 80s loop away from its wild-type position. In concert with the R42K mutation, the H87V mutation can also propagate a change in the FK506 backbone conformation near C16, as well as reorienting the 13- and 15-MeOH groups, as observed in solution by NMR(20Lepre C.A. Pearlman D.A. Cheng J.-W. DeCenzo M.T. Livingston D.J. Moore J.M. Biochemistry. 1994; 33: 13571-13580Crossref PubMed Scopus (12) Google Scholar). A second region important for calcineurin binding is composed of residues 40-44 of FKBP12(9Aldape R.A. Futer O. DeCenzo M.T. Jarrett B.P. Murcko M.A. Livingston D.J. J. Biol. Chem. 1992; 267: 16029-16032Abstract Full Text PDF PubMed Google Scholar). Becker et al.(19Becker J.W. Rotonda J. McKeever B.M. Chan H.K. Marcy A.I. Wiederrecht G. Hermes J.D. Springer J.P. J. Biol. Chem. 1993; 268: 11335-11339Abstract Full Text PDF PubMed Google Scholar) argued for the importance of this region in the effector face of FKBP12•FK506, based on sequence homologies among FKBP12 from different species and structural analysis of an inactive FK506 analog. We showed that mutation of Arg-42 to Lys, Ile, or Gln significantly decreases calcineurin affinity of the FK506 complex, with negligible effects on FK506 affinity. Furthermore, the structure of the FK506•FKBP12 R42I complex3 provides direct evidence that this substitution causes no perturbation in global protein conformation or in FK506 conformation, although higher conformational disorder of the polypeptide backbone is observed in the region of the mutation. However, Yang et al.(18Yang D. Rosen M.K. Schreiber S.L. J. Am. Chem. Soc. 1993; 115: 819-820Crossref Scopus (97) Google Scholar), who characterized FKBP12/13 chimeras, interpreted their data by claiming that residues within this region were unlikely to play a role in direct interactions with calcineurin, and were only important for conformational effects on other parts of the protein. It has been suggested that these two conflicting hypotheses needed to be resolved(37Clardy J. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 56-61Crossref PubMed Scopus (75) Google Scholar). The evidence provided by the mutation of FKBP13 residue Gln-50 to Arg is compelling. This mutation imparts substantial calcineurin affinity to the FKBP13•FK506 complex (Table1). The same mutation on an FKBP13 A95H or A95H/K98I double mutant background also increases calcineurin affinity of these FK506 complexes. In fact the resulting FK506•FKBP13 Q50R/A95H/K98I triple mutant complex has a calcineurin affinity equivalent to that of the FK506•FKBP12 complex within experimental error (Table1). Structural analysis of the Q50R/A95H/K98I triple mutant FKBP13•FK506 complex4 suggests multiple FK506 conformations but a striking overlap of the effector face residues of the triple mutant FKBP13 and wild-type FKBP12. These structural observations support the hypothesis that the 40s loop of FKBP12, and Arg-42 in particular, is part of the calcineurin effector surface of FKBP12. The definitive evidence would be provided, of course, by a high resolution structure of the FKBP12•FK506•calcineurin complex. Such structural and mutagenesis information will be useful in the design of novel FKBP12 peptidyl-prolyl isomerase and calcineurin inhibitors(39Armistead D.M. Badia M.C. Deininger D.D. Duffy J.P. Saunders J.O. Tung R.D. Thomson J.A. Futer O. DeCenzo M.T. Livingston D.J. Murcko M.A. Yamashita M.M. Navia M.A. Acta Crystallogr. 1995; (in press)Google Scholar). We thank Mark Fleming for oligonucleotide synthesis. We also thank Drs. D. Armistead, C. Lepre, M. Navia, J. Saunders, and K. Wilson for discussions and critical reading of the manuscript.
Referência(s)