12094
Biochemistry 2001, 40, 12094-12102
There Is Communication between All Four Ca2+-Bindings Sites of Calcineurin B† Stephen C. Gallagher, Zhong-Hua Gao,‡ Shipeng Li,‡ R. Brian Dyer, Jill Trewhella,* and Claude B. Klee‡ Bioscience DiVision, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, and Laboratory of Biochemistry, National Cancer Center Institute, National Institutes of Health, Bethesda, Maryland 20892-0520 ReceiVed October 30, 2000; ReVised Manuscript ReceiVed June 1, 2001
ABSTRACT: We have used site-directed mutagenesis, flow dialysis, and Fourier transform infrared (FTIR) spectroscopy to study Ca2+-binding to the regulatory component of calcineurin. Single Glu-Gln(E f Q) mutations were used to inactivate each of the four Ca2+-binding sites of CnB in turn, generating mutants Q1, Q2, Q3, and Q4, with the number indicating which Ca2+ site is inactivated. The binding data derived from flow dialysis reveal two pairs of sites in the wild-type protein, one pair with very high affinity and the other with lower affinity Ca2+-binding sites. Also, only three sites are titratable in the wild-type protein because one site cannot be decalcified. Mutation of site 2 leaves the protein with only two titratable sites, while mutation of sites 1, 3, or 4 leave three titratable sites that are mostly filled with 3 Ca2+ equiv added. The binding data further show that each of the single-site mutations Q2, Q3, and Q4 affects the affinities of at least one of the remaining sites. Mutation in either of sites 3 or 4 results in a protein with no high-affinity sites, indicating communication between the two high-affinity sites, most likely sites 3 and 4. Mutation in site 2 decreases the affinity of all three remaining sites, though still leaving two relatively high-affinity sites. The FTIR data support the conclusions from the binding data with respect to the number of titratable sites as well as the impact of each mutation on the affinities of the remaining sites. We conclude therefore that there is communication between all four Ca2+-binding sites. In addition, the Ca2+ induced changes in the FTIR spectra for the wild-type and Q4 mutant are most similar, suggesting that the same three Ca2+-binding sites are being titrated, i.e., site 4 is the very high-affinity site under the conditions of the FTIR experiments.
Calcineurin is a calcium- and calmodulin-dependent Ser/ Thr protein phosphatase that is found in all eukaryotic cells (reviewed in ref 1). It functions in T-cell activation and is a target for the immunosuppressive drugs cyclosporin and FK506 after they form complexes with the cytoplasmic binding proteins cyclophilin and FKBP12, respectively (2-5). The enzyme is made up of two tightly bound subunits: a 59 kDa component A (CnA)1 containing the active site, and a 19.3 kDa component B (CnB) that provides regulatory control via Ca2+ binding. Additional regulatory control is exerted by the binding of Ca2+/calmodulin to calcineurin. The regulatory CnB is closely related to calmodulin, although the proteins are not interchangeable and they bind to distinct sequence segments in CnA. Like calmodulin, CnB contains four helix-loop-helix [EF-hand] motifs (6-8). The Ca2+-binding sites are of the all-oxygen coordinating type with pentagonal bipyramidal geometry (7), and they share high sequence homology with the Ca2+-binding sites in calmodulin. The crystal structures of calcineurin complexed with FKBP12-FK506 (7, 8) show that, as in calmodulin, the Ca2+-binding sites of CnB occur in pairs † This work was performed under the auspices of the Department of Energy under contract to the University of California and was supported by National Institutes of Health project (GM40528, J.T.). * To whom correspondence should be addressed. Phone: (505) 6672690. Fax: (505) 667-2891. E-mail:
[email protected]. ‡ Laboratory of Biochemistry. 1 Abbreviations: CnA, active site containing subunit of calcineurin; CnB, regulatory subunit of calcineurin; FTIR, Fourier Transform InfraRed spectroscopy.
forming two globular domains (termed the N- and Cdomains), and there is a long C-terminal β-strand. The globular domains are connected via an R-helical segment that is strongly kinked at Gly85 such that they are brought into close contact and oriented to interact with the hydrophobic face of a helix in CnA. NMR data indicate that, when CnB is free in solution, residues 82-87 are flexible (9), similar to what was previously observed for the interconnecting helix region of calmodulin in solution (10, 11). In the wild-type calcineurin, one Ca2+ always remains bound to the calcineurin AB complex even with extensive dialysis against chelating agents (12). To identify this site and to gain further insights into the Ca2+-binding properties of CnB, we prepared four single-site mutants designed to deactivate each Ca2+-binding site in turn. Glu to Gln mutations were used, similar to those that have been shown to be effective in deactivating the Ca2+-binding sites in calmodulin (13). Flow dialysis and metal analysis data were used to evaluate how many Ca2+ bind to the wild-type and each mutant form of the protein, and to determine the association constants for each site. Fourier transform infrared (FTIR) difference spectroscopy was also used to monitor effects of Ca2+ binding. The FTIR data support the conclusions of the Ca2+-binding flow dialysis experiments with respect to the number of Ca2+ required to saturate each form of the protein. The FTIR data also suggest site 4 is the highaffinity Ca2+-binding site. In addition, they provide information on the protein’s secondary structure and changes that occur upon Ca2+ binding.
10.1021/bi0025060 CCC: $20.00 © 2001 American Chemical Society Published on Web 09/15/2001
Ca2+-Binding Sites Communicate in Calcineurin B
Biochemistry, Vol. 40, No. 40, 2001 12095
Table 1: Mutations in the Individual Ca2+-Binding Sites of Calcineurin B name
site mutated
mutagenic oligonucleotidesa
mutation
Q1 (Glu41Gln)
1
GAGfCAG
+BspHI
Q2 (Glu73Gln)
2
GAAfCAA
-AccI, -EcoRI
Q3 (Glu110Gln)
3
GAAfCAG
+PVuII
Q4 (Glu151Gln)
4
5′-CTGGTTCTTTGAGTGTGGAAcAG TTCATGagTCTGCCTGAGTTACAACAGA-3′ 5′-GATGGGAATGGAGAAGTAGAtTTTAAAcAA TTCATTGAGGGCGTCTCT-3′ 5′-ATGGCTATATTTCCAATGGGcAgCTgTTCCAGGTAT TGAAGATGAT-3′ 5′-GGAGATGGAAGAATATCCTTcGAAcAA TTCAAAGCTGTTGTAGGT-3′
GAAfCAA
+BstBI
a
restriction sitesb
Replaced bases are shown by lower case letters, Gln codons are shown in bold characters and underlined. b (+) Added sites; (-) deleted sites.
MATERIALS AND METHODS Mutagenesis. Restriction enzymes, ligations, DNA gel electrophoresis, and other recombinant DNA techniques were performed essentially as described by Sambrook et al. (14). The Altered Sites II-Ex1 in Vitro Mutagenesis System was from Promega and the oligonucleotides were purchased from Genosys (Texas). The plasmid pBAKE carrying the coding region of human CnB was used to create the four CnB mutants using the Altered Sites II-Ex1 in Vitro Mutagenesis System. The DNA fragment containing the entire coding sequence of CnB was excised by digestion with NcoI and BamH1 and subcloned into the pAlter-Ex1 vector converted to ampicillin resistance by inclusion of an ampicillin repair oligonucleotide in the mutation reaction according to the manufacturer’s instructions. Shown in Table 1 are the 5′phosphorylated mutagenic oligonucleotides with the silent mutations introduced to create or omit specific restriction sites to facilitate mutant selection. ES1301 muS cells were transformed with the heteroduplex DNA and the transformants selected on LB plates containing 12.5 mg of amplicillin/mL. The coding sequence, excised with NcoI and BamH1, were subcloned in the pET-11d expression vector and the entire coding sequence of the four mutants verified by DNA sequencing using the Dye Terminator Cycle Sequencing kit and oligonucleotides corresponding to the T7 promoter, TAATACGACTATAGGGGAATTG, and the CnB cDNA sequence, GACAATTCTGGTTCTTTGAGTGTGGA, with an ABI PRISM 310 genetic analyzer. Protein Preparations. All protein samples were prepared using the expression system described in Anglister et al. (9). After the phenyl-sepharose chromatography, the protein samples, concentrated to 7-13 mg/mL with a centriprep-10 Amicon concentrator, were passed through two successive Pharmacia PD10 columns equilibrated with 20 mM HEPESKOH buffer, pH 7.2, 0.1 M KCl, to remove EGTA. Ca2+ was removed from the samples by treatment with up to 2 mL/mg of protein prewashed Chelex-100. Residual Ca2+ was determined by atomic absorption spectroscopy. The molar ratio of Ca2+ to protein was 0.6, 0.2, 0.2, 0.1, and 0.1 for the wild-type CnB, Q1, Q2, Q3, and Q4 mutants, respectively. Protein concentrations for the decalcified proteins in 20 mM HEPES, pH 7.2, and 100 mM KCl were determined using an extinction coefficient at 276 nm of 4400 ( 100 based on protein concentrations measured in 6 M guanidine according to Edelhoch (15). The recombinant protein used for all studies in this paper is nonmyristylated and has Cys11 and Cys153 substituted by Ala and Lys, respectively, to avoid possible aggregation via disulfide bond formation. This recombinant protein has four Ca2+-binding sites with the
FIGURE 1: Amino acid sequence for human calcineurin B showing the Ca2+-binding sites and the Ca2+ ligands. The Ca2+-binding sites are indicated by the horizontal lines while the residues that coordinate Ca2+ are shown in bold. The mutants Q1 through Q4 have Gln substitutions for the last coordinating Glu in each site, i.e., Glu41Gln, Glu73Gln, Glu110Gln, and Glu151Gln, respectively. Note that for the experiments described in this paper the “wildtype” protein, CnB, and each mutant had the double mutation Cys11Ala, Cys153Leu to eliminate possible problems from intermolecular disulfide bond formation.
same structure as the wild-type protein (9) and, in this report, therefore, will be referred to as “wild-type.” Additional mutations were done to sequentially inactivate each of the Ca2+-binding sites. The substitutions were Glu-Gln in the Ca2+-binding sites (Table 1, Figure 1). These substitutions have been shown to prevent Ca2+ binding in calmodulin (13). The amino acid sequence for CnB is shown in Figure 1 with its four Ca2+-binding sites labeled 1-4 in order of their appearance (N- to C-terminus). The residues that contribute oxygen ligands to the Ca2+-binding site are shown in bold. The four single-site mutants each have the last glutamate residue in the sequence of one of the binding sites replaced by a glutamine and are designated Q1 ()Glu41Gln), Q2 ()Glu73Gln), Q3 ()Glu110Gln), and Q4 ()Glu151Gln). Ca2+ Binding by Flow Dialysis. Ca2+-binding determinations were performed using the flow dialysis procedure (16) as described previously (17) except that the data were corrected for the loss of 45Ca2+ (10-20%) from the well during the experiments (12). To accurately quantititate Ca2+ binding to the high-affinity sites of the wild-type and the Q1 and Q2 mutants, 45Ca2+ was allowed to exchange with endogenous Ca2+ overnight at 0-4 °C prior to the flow dialysis experiment. Interval times between collecting samples after each addition of Ca2+ for determination of free Ca2+ were chosen to maximize exchange of free and bound Ca2+ (2-3.2 min). Final free Ca2+ concentrations were 0.9 mM. FTIR Experiments. FTIR experiments were done using protein samples in D2O with 25 mM HEPES, pH 7.9 (uncorrected meter reading), and 100 mM KCl. The decalcified proteins were dialyzed twice against 100 vol of D2O (
