Importance of Phenylalanine Residues of Yeast Calmodulin for Target Binding and Activation*

Recent genetic studies of yeast calmodulin (yCaM) have shown that alterations of different sets of Phe residues result in distinct functional defects (Ohya, Y., and Botstein, D. (1994)Science 263, 963–966). To examine the importance of Phe residues for target binding and activation, we purified mutant yCaMs containing single or double Phe to Ala substitutions and determined their ability to bind and activate two target proteins, calcineurin and CaM-dependent protein kinase (CaMK). Binding assays using the gel overlay technique and quantitative analyses using surface plasmon resonance measurements indicated that the binding of yCaM to calcineurin is impaired by either double mutations of F16A/F19A or a single mutation of F140A, while binding to CaMK is impaired by F89A, F92A, or F140A. These same mutant yCaMs fail to activate calcineurin and CaMK, respectively, in vitro. In addition, F19A exhibited a severe defect in activation of both enzymes. F12A activated calcineurin to only 50% of the level achieved by wild-type calmodulin but fully activated CaMK. These results suggest that each target protein requires a specific and distinct subset of Phe residues in yCaM for target binding and activation.

␣-helix (1)(2)(3). Each domain, comprising a pair of EF hand Ca 2ϩ -binding domains, has a concave hydrophobic surface in the center. The solution and crystal structures of CaM with CaM-binding peptides derived from three different target enzymes have demonstrated that the target peptide, adopting an ␣-helical conformation, is clamped by the two domains on their hydrophobic surfaces (4 -6). Many hydrophobic interactions as well as some electrostatic interactions appear to contribute to the formation of the CaM-peptide complexes. Thus, the molecular basis of CaM's versatility is explained by the flexible central linker and the clusters of hydrophobic residues on each lobe. However, while the function of the central linker has been biochemically examined (7)(8)(9)(10)(11)(12), the role of each hydrophobic residue of CaM has been analyzed by only a few biochemical studies (13)(14)(15). A possible concern in studying these residues is that the replacement of a hydrophobic residue by a smaller, less hydrophobic one may result in gross structural perturbations of CaM.
Among the hydrophobic residues of CaM, Phe residues are thought to be functionally important for the following reasons: (i) all of the eight Phe residues in CaM are highly conserved, and (ii) structural studies of CaM-peptide complexes have established that many of the Phe residues of CaM directly contact the hydrophobic residues of the target peptide (4 -6). Genetic studies of CaM from budding yeast (yCaM) have indicated that Phe to Ala replacements do not always perturb the gross structure of yCaM (16). 33 different yCaMs were generated by systematic substitution of one or more Phe residues with Ala. Of these 33 mutant yCaMs, 22 could support the vegetative growth of yeast, suggesting that many of these mutant proteins retained a relatively normal protein structure. Nevertheless, the Phe residues of yCaM were shown to be important for yCaM function. Of the 33 Phe to Ala yCaM mutants, 14 mutant strains were temperature-sensitive. These mutations were classified into four intragenic complementation groups, and the mutants in each group showed different characteristic functional defects (17). These genetic results suggest that each complementation group of the yCaM mutants may represent a single functional defect due to loss of regulation of one essential target and that the mutations that confer temperature sensitivity do not abolish the ability of yCaM to regulate other essential target proteins even at restriction temperature (17). Thus, to understand the diversity of yCaM functions, it is important to establish the role of Phe residues of yCaM. This can be done using the Phe to Ala mutant yCaMs that retain normal protein structure.
In order to examine the effect of Phe to Ala substitutions of yCaM on their ability to regulate target proteins, we used calcineurin and CaMK, because the activation of these enzymes by mutant yCaMs can be measured in vitro. In Saccharomyces cerevisiae, the catalytic and the regulatory subunits of calcineurin are encoded by CNA1/CNA2 and CNB1 (18 -20), respectively, while CaMK is encoded by CMK1/CMK2 (21,22). This study presents several lines of biochemical evidence indicating that Phe to Ala substitutions result in altered target binding and activation and that the role of each Phe residue is different for calcineurin and CaMK.

EXPERIMENTAL PROCEDURES
Media and Strains-Media for growth of S. cerevisiae and Escherichia coli are as described (23). Genetic manipulation and yeast transformation were carried out as described (24). S. cerevisiae strains YOC200 (MATa ade2 leu2 lys2 trp1 ura3 cmd1-⌬1::TRP1 ade3::CMD1:HIS3) (16) and DCNB1-A (MATa ade2 leu2 lys2 trp1 ura3 cnb1::HIS3) 2 were used as a positive and a negative control in halo assays, respectively. The other YOC strains have the same genetic background as YOC200 except that each strain harbors a cmd1 allele instead of CMD1 (16). Escherichia coli strain DH5␣ was used for propagation of plasmids and expression of recombinant proteins. E. coli strain BL21 carrying pLysS (25) was used for production of yCaMs, calcineurin, and CaMK.
Materials-Restriction enzymes and modifying enzymes were purchased from Takara (Kyoto, Japan) and New England Biolabs Inc. (Beverly, MA). Synthetic ␣-factor was purchased from Peptide Institute Inc. (Osaka, Japan). Cyclic AMP-dependent protein kinase catalytic subunit was purchased from New England Biolabs. [␥-32 P]ATP was purchased from ICN Biomedicals Inc. (Costa Mesa, CA). Phenyl-Sepharose CL-6B, Sephadex G-25 medium, and PD-10 column were obtained from Amersham Pharmacia Biotech (Uppsala, Sweden). DNase I and ␣-casein were purchased from Sigma. Centriprep-10 was obtained from Amicon (Beverly, MA). Each oligopeptide containing the yCaM-binding sequence of Cna1p (residues 453-476) or Cmk1p (residues 314 -340) with a cysteine residue at its N terminus added (designated as Cna1pep and Cmk1pep, respectively) was custom-synthesized by Toray Research Center. The segment corresponding to Cna1pep in human calcineurin A-2 and the segment corresponding to Cmk1pep in Cmk1p have been shown to interact with bovine CaM and yCaM, respectively (22,26).
Plasmid Construction-Each plasmid for expressing wild-type or mutant yCaM (pET-cmd) was constructed by inserting the BstBI--SphI cmd1 fragment (16) into a BstBI--BamHI gap of pET-3a after BamHI linker ligation. Plasmids for expressing fusion proteins of Cna1p, Cmk1p, and Cmk2p were made using the GST Gene Fusion System (Amersham Pharmacia Biotech), and a plasmid for expressing a fusion protein of Cna2p was made using the Protein Fusion and Purification System (New England Biolabs) (see Fig. 2A). A plasmid for expressing Cna1p (pET-CNA1) was made as follows. An NcoI site and an XhoI site were introduced at the start codon and the 3Ј-noncoding region of CNA1, respectively, by in vitro mutagenesis. This caused a change of the second amino acid residue from Ser to Ala. A 1.7-kilobase pair NcoI-XhoI fragment of CNA1 was inserted into an NcoI-XhoI gap of pET-15b (pET-15b-CNA1) followed by inserting a 1.7-kilobase pair XbaI-XhoI fragment from pET-15b-CNA1 into an XbaI-XhoI gap of pET-3a. A plasmid for expressing Cnb1p (pET-CNB1) was made as follows. An NcoI site and a BamHI site were introduced at the start codon and 3Ј-noncoding region of CNB1 with an intron removed by two-step PCR, and a 0.5-kilobase pair NcoI-BamHI fragment was inserted into an NcoI-BamHI gap of pET-15b.
Protein Determination-Protein concentration was determined by the BCA method using the BCA protein assay reagent (Pierce) with bovine serum albumin as a standard.
Expression and Purification of Mutant yCaMs-E. coli strain BL21 carrying pLysS was transformed with each pET-cmd plasmid and incubated for 2 h at 37°C in TB medium containing 30 g/ml ampicillin. yCaM production was induced by adding isopropyl 1-thio-␤-D-galactoside to a final concentration of 1 mM. After 4 h of incubation, cells were collected by centrifugation, resuspended in 10 mM Tris/HCl (pH 8.0) containing 1 mM EDTA and 1 mM phenylmethylsulfonyl fluoride, and then freeze-thawed. MgCl 2 and DNase I were added to final concentrations of 5 mM and 40 g/ml, respectively, and incubated for 10 min at room temperature. Cellular debris was removed by centrifugation at 100,000 ϫ g for 1 h. yCaM was purified essentially as described previously (27). The supernatant containing 1 M ammonium sulfate and 1 mM CaCl 2 was applied to a bed of phenyl-Sepharose 6B equilibrated with 10 mM Tris/HCl (pH 8.0) containing 1 M ammonium sulfate and 5 mM CaCl 2 . After washing with 10 mM Tris/HCl (pH 8.0) containing 1 mM CaCl 2 and an appropriate concentration of ammonium sulfate (0.1-0.5 M) for each wild-type and mutant yCaM, yCaM was eluted with 10 mM Tris/HCl (pH 8.0) containing 1 mM EGTA and the same concentration of ammonium sulfate as used for washing. The eluate was concentrated by ultrafiltraton using Centriprep-10, and stored at Ϫ80°C.
Biotin Labeling of yCaM-Purified wild-type and mutant yCaMs were biotinylated using ImmunoPure Sulfo-NHS-LC-Biotin Kit (Pierce) according to the manufacturer's instructions.
Gel Overlay Assay-Fusion proteins containing yCaM-binding sequences of calcineurin and CaMK were expressed in E. coli strain DH5␣. Extracts containing individual fusion proteins were resolved by SDS-PAGE and proteins were transferred electrophoretically to a nitrocellulose membrane. The membrane was incubated at room temperature for 1 h with 2% (w/v) nonfat dry milk in 50 mM Tris/HCl (pH 8.0) containing 150 mM NaCl and 0.05% Tween 20. The blocked membrane was probed at room temperature for 1.5 h with 400 nM or 4 M biotinylated wild-type or mutant yCaM in Ca 2ϩ buffer (10 mM Tris/HCl (pH 7.5), 0.2% bovine serum albumin, and 1 mM CaCl 2 ). The membrane was washed with Ca 2ϩ buffer three times, each for 10 min, and air-dried. Biotinylated horseradish peroxidase-avidin complex was coupled to biotinylated yCaM using Vectastain Elite standard kit (Vector laboratories). After washing with Ca 2ϩ buffer three times, each for 10 min, the membrane was stained using the POD immunostain set (Wako).
Surface Plasmon Resonance (SPR) Measurements-Interactions of mutant yCaMs with Cna1pep and Cmk1pep were analyzed by SPR measurements using the IAsys biosensor (Fisons Applied Sensor Technology). All manipulations were carried out at 25°C. Cna1pep or Cmk1pep was covalently coupled to the carboxymethyldextran-coated biosensor cuvette via the thiol group of the N-terminal cysteine residue according to the manufacturer's instructions. After establishing a base line with 20 mM Tris/HCl (pH 7.5) containing 1 mM CaCl 2 , 250 mM NaCl, and 0.005% Tween 20, the binding of mutant or wild-type yCaM was monitored as an increase in the evanescent field response. The cuvette was regenerated with 20 mM Tris/HCl (pH 7.5) containing 1 mM EGTA, 250 mM NaCl, and 0.005% Tween 20.
Partial Purification of Recombinant Calcineurin-E. coli strain BL21 carrying pLysS was transformed individually with pET-CNA1 and pET-CNB1 and grown to an optical density at 600 nm of 0.5 at 23°C and of 0.25 at 37°C, respectively, in TB medium containing 100 g/ml ampicillin and 25 g/ml chloramphenicol. The production of Cna1p and Cnb1p was induced by adding isopropyl 1-thio-␤-D-galactoside to final concentrations of 0.1 and 1 mM, respectively. Cells were collected by centrifugation; resuspended in 50 mM Tris/HCl (pH 7.5) containing 10 mM EDTA, 10 mM DTT, and protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine HCl, and 5 g/ml each of leupeptin, pepstatin A, aprotinin, and chymostatin); and then freezed-thawed. After sonication, the lysate was clarified by centrifugation at 100,000 ϫ g for 1 h. Supernatants containing Cna1p and Cnb1p were mixed and then incubated at 4°C for 24 h. Reconstituted calcineurin was precipitated by adding 30% ammonium sulfate, and the precipitate was resuspended in buffer A (50 mM Tris/HCl (pH 7.5), 1 mM DTT, 1 mM CaCl 2 , and protease inhibitors) and was dialyzed against buffer A at 4°C. The dialysate was applied to yCaM-conjugated Sepharose 4B equilibrated with buffer A. After it was washed with 50 mM Tris/HCl (pH 7.5) containing 1 mM DTT, 1 mM CaCl 2 , 0.5 M NaCl, and protease inhibitors, calcineurin was eluted with 50 mM Tris/HCl (pH 7.5) containing 1 mM DTT, 5 mM EGTA, and protease inhibitors. The eluate was dialyzed against 50 mM Tris/HCl (pH 7.5) containing 5 mM DTT, 5 mM sodium ascorbate, 0.5 mM FeSO 4 , and 50% glycerol.
Phosphatase Assay-Phosphatase activity was assayed at 30°C for 20 min in a 50 l reaction mixture containing 50 mM MOPS (pH 7.0), 1 mg/ml bovine serum albumin, 1 mM CaCl 2 , 5 mM DTT, 5 mM sodium ascorbate, 32 P-labeled ␣-casein corresponding to 100 pmol of P i , yCaM, and calcineurin. The release of P i was measured by a modified version 2 I. Tanida, Y. Ohya, and Y. Anraku, unpublished result. of the organic extraction procedure (28,29). The reaction was terminated by adding 400 l of 5 mM silicotungstate in 5 mM H 2 SO 4 and 600 l of a 1:1 mixture of isobutyl alcohol and benzene. After vortexing briefly, 80 l of 5% (w/v) ammonium molybdate in 2 M H 2 SO 4 was added, and the solution was vigorously mixed for 10 -15 s. The organic and inorganic phases were separated by centrifugation in a tabletop centrifuge for 5 min, and the amount of radioactivity in 400 l of the organic phase was determined.
Purification of Recombinant CaMK-CaMK (Cmk1p) was purified as described previously with some modification (21). E. coli strain BL21 carrying pLysS was transformed with pET-CMK1 (21) and grown to an optical density at 600 nm of 0.1 at 37°C in TB medium containing 100 g/ml ampicillin and 25 g/ml chloramphenicol. Cmk1p production was induced by adding isopropyl 1-thio-␤-D-galactoside to a final concentration of 0.5 mM. Cells were collected by centrifugation, resuspended in buffer C (20 mM Tris/HCl (pH 7.5), 2 mM EDTA, 1 mM DTT, and protease inhibitors containing 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine HCl, and 5 g/ml each of leupeptin, pepstatin A, aprotinin, and chymostatin), and then freeze-thawed. After sonication, the lysate was clarified by centrifugation at 100,000 ϫ g for 1 h. Nucleic acids were removed by adding 5% (w/w) protamine sulfate followed by centrifugation at 100,000 ϫ g for 20 min. The supernatant was adjusted to 0.2 M NaCl and applied to phenyl-Sepharose 6B equilibrated with buffer C containing 0.2 M NaCl. After washing with two bed volumes of buffer C containing 0.2 M NaCl, proteins were eluted with a linear gradient from 0.2 to 0 M NaCl in buffer C. The peak fractions of CaMK activity were pooled. Proteins in the pooled fractions were concentrated by 80% ammonium sulfate. The precipitate was resuspended in buffer C and dialyzed against buffer D (20 mM Tris/HCl (pH 7.5), 1 mM DTT, 1 mM CaCl 2 , 10% glycerol, and protease inhibitors) at 4°C. The dialysate was applied to yCaM-conjugated Sepharose 4B equilibrated with buffer D. After washing with 20 mM Tris/HCl (pH 7.5) containing 1 mM DTT, 1 mM CaCl 2 , 0.5 M NaCl, and protease inhibitors, CaMK was eluted with 20 mM Tris/HCl (pH 7.5) containing 1 mM DTT, 5 mM EGTA, and protease inhibitors. The eluate was dialyzed against 20 mM Tris/ HCl (pH 7.5) containing 1 mM DTT and 50% glycerol.
Determination of Kinetic Parameters-Dissociation constants from binding assays were determined as follows. First, the response change at equilibrium was determined by nonlinear regression with an equation representing one-site binding and plotted for each yCaM concentration. Dissociation constants were determined by nonlinear regression of this plot with the Michaelis-Menten equation. Kinetic parameters for enzyme activation were determined by nonlinear regression with the Michaelis-Menten equation.
Halo Assay-MATa haploids were picked as colonies from fresh plates and grown at 23°C in YPD medium. The fully grown culture was suspended in soft (1.5%) melted YPD agar that had been warmed at 55°C to make 1-2 ϫ 10 6 cells/ml of suspension. Cells were then plated on prewarmed YPD-Suc agar plates containing 1% Bacto-yeast extract, 2% polypeptone, 2% glucose, 2% succinate and 2% agar. Glucose and succinate were autoclaved separately. After solidification of the top agar, sterile filter disks (diameter 6 mm) were placed aseptically on the surface of the nascent lawn, and 1.6 nmol of synthetic ␣-factor (5 l) was pipetted onto sterile filter disks. The plates were incubated at 23°C for 3, 5, and 7 days before photographs were taken. Because of the extended period of incubation, this assay has been useful to measure the capacity of cells to recover from ␣-factor arrest and resume growth. These assays were performed for each yCaM mutant at least three times.  (Fig. 1). In contrast, some of the mutants with two or more Phe to Ala substitutions showed significantly altered mobility in the presence of EGTA. For example, F65A/F68A exhibited a larger apparent molecular mass of 16.9 kDa in the presence of EGTA (Fig. 1). However, another double mutant, F16A/F19A, showed a mobility similar to that of wild-type yCaM (data not shown), indicating that double mutations do not always result in the irregular mobility. Thus, the single substitution Phe to Ala mutations and the F16A/F19A double mutation combination did not significantly perturb the global protein structure of yCaM. A triple mutant protein, F12A/F16A/F19A ran more slowly than wild-type yCaM, exhibiting an apparent molecular mass of 17. can be used (14). All of the single substitution mutant yCaMs exhibited nearly identical retention times, further suggesting that the single substitution mutations do not cause global structural perturbations of yCaM (Table I).

SDS-PAGE and Gel Filtration
Gel Overlay Assays of Calcineurin and CaMK with Mutant yCaMs-We performed gel overlay assays to investigate the allele-specific binding of mutant yCaMs to calcineurin and CaMK. Fragments of the target enzymes containing the yCaMbinding sequences were expressed in E. coli as fusions with GST for Cna1p, Cmk1p, and Cmk2p or as a fusion with MalE for Cna2p ( Fig. 2A). Fig. 2B clearly shows that the binding of mutant yCaMs to each target protein is allele-specific and that binding specificity differs for calcineurin and CaMK. yCaM binding to GST-Cna1 and MalE-Cna2 were severely impaired for mutant yCaMs possessing the substitutions F19A or F140A (lanes 4, 5, 8, and 9). On the other hand, yCaM binding to GST-Cmk1 or GST-Cmk2 was not detected for mutant yCaMs possessing a mutation in the C-terminal half, F89A, F92A, or F140A (lanes 6 -8). A triple mutant protein, F12A/F16A/F19A, also failed to bind to both GST-Cmk1 and GST-Cmk2. For each yCaM, equivalent binding was always observed to the two different isoforms of calcineurin, GST-Cna1 and MalE-Cna2.
However, slightly different results were observed for GST-Cmk1 and GST-Cmk2. F19A and F16A/F19A were shown to have slightly reduced affinities for GST-Cmk1 but not for GST-Cmk2, whereas the binding of F89A was more weakened for GST-Cmk2 than GST-Cmk1.
SPR Measurements of the Binding of Mutant yCaMs to Calcineurin and CaMK-In order to quantify the binding of wildtype or mutant yCaMs to calcineurin and CaMK, we determined the kinetic parameters of binding using an optical biosensor (Table II). Cna1pep and Cmk1pep, which contain the yCaM-binding sequences of Cna1p (positions 453-476) and Cmk1p (positions 314 -340), respectively, were individually immobilized on cuvettes via their N-terminal cysteine residues, and their interactions with wild-type and mutant yCaMs were monitored in real time.
For binding to Cna1pep, F16A/F19A and F140A, both of which showed no detectable binding activity to GST-Cna1 in the gel overlay assay, exhibited dissociation constants that were more than 10-fold larger than that of wild-type yCaM. In contrast, F19A, which showed severely decreased binding to GST-Cna1 in the gel overlay assay, exhibited only a slightly increased dissociation constant. F12A, F16A, and F65A/F68A exhibited dissociation constants that were similar to that of wild-type yCaM. These observations are consistent with the finding that these yCaMs showed no defect in yCaM binding in the gel overlay assay.
For binding to Cmk1pep, the responses of F89A, F92A, and F140A, all of which failed to bind to GST-Cmk1, were too small to determine the dissociation constants, indicating that the affinities of these mutant yCaMs to Cmk1pep are considerably decreased. The dissociation constants of F12A, F16A, F19A, and F65A/F68A are similar to (and that of F16A/F19A is slightly larger than) that of wild-type yCaM. These results are consistent with the findings from the gel overlay assay.   Table II). To assay calcineurin activity, the catalytic subunit (Cna1p) and the regulatory subunit (Cnb1p) were individually expressed in E. coli, and recombinant calcineurin was reconstituted. We found that F16A/F19A exhibited a more than 13-fold increase in the activation constant for calcineurin. The other mutant yCaMs exhibited small increases in the activation constants for calcineurin, with F140A exhibiting the largest (less than 4-fold) increase. Some mutant yCaMs showed a significantly decreased ability to activate calcineurin. The most striking observation was that F19A and F16A/F19A activated calcineurin to only 13 and 20%, respectively, of the level achieved by wildtype yCaM. F12A and F140A were also poor activators of calcineurin and induced only 50 -60% as much phosphatase activity as did wild-type yCaM. Compared with the activation of calcineurin, a different set of Phe residues in yCaM was involved in the activation of CaMK. The C-terminal mutant yCaMs, F89A, F92A, and F140A, showed negligible activation of CaMK. All of the Nterminal mutant yCaMs exhibited small increases in the activation constants for CaMK, with F16A exhibiting the largest (about a 5-fold) increase. F19A and F16A/F19A exhibited severe defects in CaMK activation, showing a 70 -100% decrease in maximal activity, while F16A induced only 77% as much CaMK activity as did wild-type yCaM.

Activation of Calcineurin and CaMK by Mutant yCaMs-The
Mating Pheromone Response of yCaM Mutants-It has been reported that calcineurin and CaMK are required to maintain viability during exposure to mating pheromone and that their effects are additive (30,31). This survival defect of calcineurindeficient cells can be detected using a halo assay with mating pheromone. Wild-type cells exhibit a turbid halo in response to mating pheromone; however, with calcineurin-deficient cells, a clear halo is observed (18,20). CaMK-deficient cells are indistinguishable from wild-type cells when examined by halo assay (data not shown). However, CaMK-deficient cells are slightly compromised for survival during response to pheromone, and cells lacking both calcineurin and CaMK activity display a clearer halo than cells lacking only calcineurin (30). Thus, we predict that yCaM mutants that are defective either for activation of calcineurin only or for activation of both calcineurin and CaMK in vivo will display a clear halo in response to mating pheromone. As expected from the biochemical results for calcineurin, yeast strains expressing the mutant yCaM harboring the F16A/F19A substitutions showed clear halo phenotypes like ⌬cnb1 (Fig. 5). A strain expressing F12A/F140A showed an intermediate phenotype (Fig. 5). All other mutants except strains expressing F12A/F92A showed turbid halo phenotypes (Table III). In vitro, we observed that F12A/F92A can activate calcineurin (data not shown); therefore, the mutations of F12A/F92A might impair another yCaM regulatory pathway other than calcineurin that is essential for adaptation to mating pheromone. The clear halo phenotypes we observed did not simply reflect the slower growth of the mutants, because prolonged incubation for more than a week did not change the phenotype.

DISCUSSION
In this study, we examined the ability of mutant yCaMs containing Phe to Ala substitutions to bind and activate two target proteins using a gel overlay assay, SPR measurements, enzymatic assays, and genetic analyses. Our results clearly TABLE II Kinetic parameters for binding and activation of enzymes by mutant yCaMs K d is the concentration of yCaM that produces half-maximal binding level. Maximal binding level and K d were determined by nonlinear regression as described under "Experimental Procedures." K CaM is the concentration of yCaM that produces half-maximal activation and was determined by nonlinear regression as described under "Experimental Procedures." V max is the maximal observed activity, expressed as the percentage of the maximal activity observed for wild-type yCaM. Maximal activity was determined by nonlinear regression as described under "Experimental Procedures." yCaM Calcineurin CaMK  3. Activation of calcineurin by mutant yCaMs. The activity of calcineurin was measured at different concentrations of yCaM as described under "Experimental Procedures." Maximal activity (100%) was set at the V max value of wild-type yCaM obtained from nonlinear regression of the Michaelis-Menten equation (Table II) indicate that a distinct set of Phe residues of yCaM is required to bind and activate each target enzyme.
The binding of mutant yCaMs to calcineurin and CaMK was examined by three independent binding assays: gel overlay assay, SPR measurements, and the assay of enzyme activation. The inability of F16A/F19A to bind to the GST-Cna1 and MalE-Cna2 fusions on a membrane and the significantly larger dissociation and activation constants of F16A/F19A for calcineurin indicate that Phe-16 and Phe-19 of yCaM are required for this interaction. Phe-19 seems slightly more important for the binding than Phe-16, since the single substitution mutation of F19A itself has a more profound effect as determined by gel overlay assay and SPR measurements. In addition, Phe-140 is also important for the interaction with calcineurin based on these binding assays. In contrast, binding to CaMK is markedly diminished by the C-terminal mutations of F89A, F92A, or F140A. The importance of the C-terminal Phe residues for CaMK binding suggests that the C-terminal domain of yCaM dominantly contributes to the whole binding energy of the yCaM⅐CaMK complex. Although the binding of F12A/F16A/ F19A to CaMK was also not detected by gel overlay assay, this finding may be due to structural perturbations caused by the multiple mutations (see discussion below).
The dissociation and activation constants of F19A are similar to those of F89A and F92A, both of which could bind to the fusions of calcineurin on a membrane. However, gel overlay assays showed that F19A failed to bind to fusion proteins of calcineurin on a membrane. This discrepancy might arise from the denatured forms of the fusion proteins used in gel overlay assays. We assume that F19A has the ability to bind to native calcineurin but fails to bind to denatured calcineurin. Alterna- FIG. 4. Activation of CaMK by mutant yCaMs. The activity of CaMK was measured at different concentrations of yCaM as described under "Experimental Procedures." Maximal activity (100%) was set at the V max value of wild-type yCaM obtained from nonlinear regression of the Michaelis-Menten equation (Table II)   The activation and dissociation constants of F140A were different for calcineurin. This discrepancy might result from the immobilization of the N terminus of Cna1pep in the binding assay. For optimal binding, F140A might position its C-terminal lobe slightly differently from that of wild-type relative to the yCaM-binding domain of calcineurin. If the C-terminal lobe of yCaM binds to the N terminus of the yCaM-binding domain of calcineurin as observed in the three structures of CaMpeptide complexes, such alteration in the positioning of the C-terminal lobe of F140A might be sterically hindered in binding to Cna1pep with its N terminus immobilized on carboxymethylated dextran.
The ability of each mutant yCaM to activate target enzymes was assessed by comparing the maximal activation achieved by each mutant yCaM relative to wild-type yCaM. Calcineurin is hardly activated by F19A and F16A/F19A. F12A and F140A also exhibit some defects in activation of calcineurin. Consistent with these results, yeast strains expressing only F16A/ F19A, F12A/F16A/F19A, or F12A/F140A showed the same phenotypes as that of a calcineurin-deficient strain, suggesting that these mutant yCaMs cannot stimulate calcineurin in vivo to the level required to display the wild-type phenotype. CaMK is activated only slightly by F19A and F16A/F19A as well as F89A, F92A, and F140A. The inability of F89A, F92A, and F140A to activate CaMK may be due to defective binding to CaMK at the concentrations of yCaM used in the assay. A specific role for F19A in CaMK activation is suggested, since F19A retains most of its binding ability.
We examined whether Phe to Ala mutations in yCaM affect its overall structure and Ca 2ϩ -dependent conformational change. The mutations of a single Phe residue and the double substitution F16A/F19A in yCaM do not have significant effects on the global protein structure as judged by their interaction with phenyl-Sepharose, their mobility during SDS-PAGE in the presence of Ca 2ϩ or EGTA, and their retention time in gel filtration chromatography in the presence of Ca 2ϩ . Although F65A/F68A exhibits a slower electrophoretic mobility than wild-type yCaM in the presence of EGTA (apparent molecular mass of 16.9 kDa), it shows the same mobility as wild type in the presence of Ca 2ϩ , suggesting that Ca 2ϩ binding restores the native conformation. This may explain why F65A/ F68A has the ability to bind to both target proteins well. In contrast, some yCaMs containing multiple Phe to Ala substitutions showed aberrant mobility in the presence of EGTA or Ca 2ϩ (16). A triple mutant protein, F12A/F16A/F19A, and a quintuple mutant protein, F12A/F16A/F19A/F65A/F68A, both show a much slower electrophoretic mobility than wild-type yCaM. Multiple replacements of Phe by Ala could cause a considerable gap in the hydrophobic core of yCaM, thereby destabilizing the native conformation of yCaM. That normal overall protein structure is preserved in single substitution mutant proteins is further suggested by the in vivo phenotypes of yCaM mutants. Yeast strains expressing each single-substitution mutant yCaM are all viable at 30°C (16). This indicates that these mutant yCaMs can interact with all of the targets that are essential for cell proliferation at 30°C. Taken together, all of the single mutant yCaMs as well as F16A/F19A and F65A/F68A have the same global conformation as wildtype yCaM, at least in the presence of Ca 2ϩ . Subsequent to our analyses of the functional importance of Phe residues in yCaM, a mutant of human CaM containing the F92A substitution was analyzed (32). In contrast to the F92A mutant yCaM, the F92A mutant human CaM displays a major defect in Ca 2ϩ -induced conformational transition. First, the F92A human CaM runs significantly more slowly than wildtype CaM during gel electrophoresis (32), while the F92A mutant yCaM is indistinguishable from wild-type yCaM in mobility (Fig. 1). Second, the Ca 2ϩ -dependent binding of hydrophobic probes to the F92A human CaM is lower than that of wild-type human CaM, while the Ca 2ϩ -dependent binding of the hydrophobic probes to the F92A yCaM and wild-type yCaM is very similar (data not shown). These results suggest that the F92A substitution causes much less of a change in the exposure of hydrophobic surfaces in yCaM than it does in human CaM. yCaM mutants harboring one or several Phe to Ala mutations were classified into four intragenic complementation groups (17). Each group showed a different characteristic functional defect in actin organization, yCaM localization, nuclear division, or bud emergence. Based on analyses of the complementing yCaM mutants, we have proposed that each complementation group is specifically defective for the activation of one essential target of yCaM (17). This study has revealed that distinct subsets of Phe residues in yCaM are in fact required for activation of calcineurin and CaMK, thus providing a biochemical basis for the observed complementation. Phe-92 of yCaM has recently been shown to be essential for the interaction with Myo2p (33), one of the essential calmodulin targets in S. cerevisiae (34). Thus, given that Myo2p is biochemically assigned to one of the intragenic complementation groups, Phe residue(s) other than Phe-92 must participate in the regulation of another known essential target, Nuf1p/Spc110p (35,36). Furthermore, yCaM binding to the other unknown essential targets must also require specific Phe residues.