Demonstration of Different Metal Ion-induced Calcineurin Conformations Using a Monoclonal Antibody*

It has been suggested that calcineurin, a calmodulin-stimulated phosphatase, may exist in different metal ion-dependent conformational states (Pallen, C.J., and Wang, J. H. (1984) J. Biol. Chem. 259, 6134-6141). Evidence in favor of this hypothesis comes from studies involving a monoclonal antibody, VA1, which is specific for the small (beta) subunit of calcineurin. This antibody inhibits Ni2+-stimulated but not Mn2+-stimulated phosphatase activity against p-nitrophenyl phosphate and phosphorylase kinase. Inhibition is not due to competition of the antibody with substrate or to interference with metal ion binding to the enzyme. Complex formation between the antibody and calcineurin can be demonstrated either in the presence of Mn2+ or Ni2+ or in the absence of metal ion activators. These results indicate that the active conformational states of calcineurin are metal ion dependent, that the monoclonal antibody VA1 affects the Ni2+-induced conformational change of the enzyme, and that the beta subunit of calcineurin plays a critical role in the expression of Ni2+-stimulated phosphatase activity.

interconverted. These results have led to the suggestion that calcineurin may exist in different metal ion-dependent conformations (10).
Recently several monoclonal antibodies to calcineurin have been developed and characterized in our laboratories.' One of these is specific toward the /3 subunit of the protein whereas others are a subunit specific. In this study we show that the /3 subunit-specific monoclonal antibody VA1 is capable of inhibiting Ni2'-activated calcineurin activity but not Mn2+activated enzyme activity. This inhibition by the antibody can be demonstrated when either protein or nonprotein substrates are used. The results suggest that multiple phosphatase activities are intrinsic properties of calcineurin and substantiate the suggestion that calcineurin can exist in different metal ion-dependent conformational states.

MATERIALS AND METHODS
Proteins-Calcineurin and calmodulin were purified from bovine brain as previously described (11). Phosphorylase kinase was purified from rabbit skeletal muscle according to a published procedure (12). The monoclonal antibody VA, was developed and purified from mouse ascites fluid using a DEAE'-Affi-Gel blue column (13).
Assay for Calcineurin-Phosphatase Activity-Calcineurin activity against pNPP was assayed as described previously (10). Briefly, calcineurin was preincubated with either 1 mM metal ion as indicated or 1 mM EDTA at 25 "C for an hour in the presence of BSA (0.5 mg/ ml). Calcineurin activity was assayed in 0.5 ml of a reaction mixture containing 50 mM Tris-HC1 (pH 7.2), 1 mM metal ion, 2.7 mM pNPP, 0.5 mg/ml BSA, and in the presence or absence of 1.5 W M calmodulin at 30 "C. The reaction was either terminated by the addition of 50 ~l of 13% KzHP04 and read at the wavelength 405 nm, or the time course of the reaction was followed on a Bausch and Lomb 2000 spectrophotometer.
NcrDodS04-Polyacrylnmide Gel Electrophoresis and Immunoblotting-Slab gel electrophoresis in the presence of 0.1% NaDodSO, was carried out according to the procedure of Laemmli (14) using 7.5-15% acrylamide gradients. Protein bands were visualized using the silver staining method as described by Wray et al. (15).
Immunoblotting was carried out essentially as described by Towbin et al. (16). The electrophoretic transfer of protein was carried out in the presence of 0.1% NaDodSO, to increase the transfer efficiency. The detection of the cross-reactive protein band was carried out using alkaline phosphatase-conjugated rabbit anti-mouse immunoglobulin G (Sigma). The blots were developed on an ultraviolet transilluminator using 4-methylumbelliferyl phosphate as a substrate (17).
High Performance Liquid Chromatography-The HPLC gel filtration column TSK-3000 SW (7.5 X 300 mm) from LKB was used. Elution buffer contained 20 mM Tris acetate and 0.1 M NalSOl (pH 7.2). The sample was prefiltered using HPLC nylon filters (3-mm H.-Y. P. Lam and J. H. Wang, manuscript in preparation. ' The abbreviations used are: DEAE, diethylaminoethyl; HPLC, high performance liquid chromatography; PAGE, polyacrylamide gel electrophoresis; NaDodSO,, sodium dodecyl sulfate; EDTA, ethylenediaminetetraacetic acid; pNPP, p-nitrophenyl phosphate; Tris, tris(hydroxymethy1)aminomethane; BSA, bovine serum albumin. diameter, 0.45-j~m pore size). Elution was done at room temperature and at a flow rate of 0.33 ml/min. The column eluate was monitored for UV absorbance at 280 nm and was fractionated every minute (0.33 ml/fraction). The fractionated samples were analyzed by Na-DodS0,-polyacrylamide gel electrophoresis and silver staining.

RESULTS
Further Characterization of the Monoclonal Antibody VAl-Five monoclonal antibodies have been purified and characterized in our laboratories. Only one of these antibodies, VA1, appears to be specific for the /3 subunit of calcineurin, as determined by immunoblotting. Immunoblotting analysis of VA1 reactivity toward purified bovine brain calcineurin and crude extract of rat brain is shown in Fig. 1. In both cases a polypeptide with a molecular weight identical to /3 subunit strongly reacts with the antibody, while a! subunit does not react at all. A protein band, present in both purified calcineurin and in crude rat brain extract, with a molecular weight lower than /3 subunit also reacts with VA1. In addition, several polypeptides with molecular weights higher than that of a! subunit showed consistent but weak reactivity with VA,.
These results indicate that VA, is specific for the /3 subunit of bovine and rat brain calcineurin but may possess weak cross-reactivity with several minor components of brain, some of which appear to be present in the purified calcineurin sample. Since the cell line producing the monoclonal antibody VA1 was subcloned three times, these reactions are unlikely Purified bovine brain calcineurin (1 pg) and rat brain crude extract (105,000 X g supernatant, 350 pg) were electrophoresed on NaDodS0,-PAGE. Proteins were electrophoretically transblotted onto nitrocellulose filters as described under "Materials and Methods." The nitrocellulose blots were incubated with 3% BSA in Tris saline for 1 h at 40 "C and then incubated with 27 pg/ml monoclonal antibody in Tris saline with 3% BSA for 2 h. The blot was then washed with Tris saline and incubated with alkaline phosphataseconjugated rabbit anti-mouse immunoglobulin G (Sigma). The blot was developed as described under "Materials and Methods." As a control the same samples were processed without incubation with the monoclonal antibody VAL solution. In this case no protein bands could be seen. Arrows show the position of each subunit (a and @) of calcineurin on the nitrocellulose. A , purified bovine brain calcineurin; B, rat brain crude extract.
to be due to contamination of VA1 with a different antibody clone. The nature of these minor components is unclear, but they may be higher molecular weight nondissociable forms of calcineurin (for example aP2) or, in the case of the smallest reactive polypeptide, a degradation product of the /3 subunit.
Effect of VA, on Calcineurin-Phosphatase Activity-The ability of the various monoclonal antibodies to inhibit pnitrophenyl phosphatase activity of calcineurin was tested. Only VA1 was found to have an inhibitory effect, and this effect is dependent upon the metal ion used to activate the enzyme. Previously we have shown that p-nitrophenyl phosphatase activity of calcineurin requires a metal ion activator, a requirement which may be satisfied by either Mn2+ or Ni2+. Fig. 2 shows that VA, markedly inhibits phosphatase activity of Ni2+-stimulated calcineurin but only marginally affects the activity of Mn2+-stimulated calcineurin.
The concentration dependence of VA1 inhibition of Ni2+stimulated calcineurin activity (Fig. 2) suggests that inhibition results from a stoichiometric interaction of 1 molecule of calcineurin with 1 molecule of antibody. At a molar ratio of antibody to enzyme of 1:1 about 75% inhibition of activity is observed (Fig. 2). Increasing antibody concentration up to a molar ratio of 41 resulted in only a small additional extent of inhibition. At the highest antibody concentration used, 20% of the original enzyme activity remained, indicating that the fully saturated immunoenzyme complex possesses residual catalytic activity.
Inhibition of Ni2+-activated p-nitrophenyl phosphatase activity by VA, is not competitive with respect to substrate (Fig.  3), indicating that inhibition does not arise from the binding of the antibody to the active site of the enzyme. Both V, , and K, of the dephosphorylation reaction are altered following immunoenzyme complex formation. The suggestion that the antigenic site is distinct from the active site on the enzyme is also compatible with the observation that the inhibition of enzyme activity is incomplete even at high antibody concentration.
The possibility that the antibody inhibits calcineurin-phosphatase by interfering with Ni2+ binding to the enzyme or by causing dissociation of Ni2+ from the enzyme has been con-  Calcineurin (0.2 mg/ml) was preincubated with 1 mM Ni2+. Preincubated calcineurin (10 pg) was then incubated with 15 pg of VA, antibody (-) or without antibody in the absence of calmodulin. The Ni2+ concentration was serially (---). Phosphatase activity was monitored spectrophotometrically increased. The assay was started with 0.1 mM Ni2+ and then increased to 0.3, 0.5, 1, and 2 mM Ni2+ at the times indicated by the arrows.
sidered. Similar extents of inhibition were observed if VA, was incubated with Ni2+-stimulated calcineurin or if the VA1calcineurin complex was incubated with Ni2+, indicating that antibody does not dissociate bound Ni2+ or prevent Ni2+ binding to calcineurin. Further support for eliminating the release of bound Ni2+ from calcineurin by antibody as the mechanism of inhibition comes from the observation that enzyme inhibition cannot be reversed by the addition of excess Ni2+ (Fig. 4). The evidence that VA1 does not interact directly with either the active site or the metal-binding site(s) of calcineurin led us to suggest that inhibition of calcineurinphosphatase by VA1 is the consequence of an antibody-induced conformational change of the enzyme to a form with low catalytic activity. The extent of inhibition of calcineurin by VA1 is dependent on calmodulin (Fig. 5A). In the presence of this protein activator, VAI effected a maximal inhibition of 82%, whereas in the absence of calmodulin a maximal inhibition of 45% was observed. While the free Ni2+-activated calcineurin may be activated 4.5-fold by calmodulin, the immunoenzyme complex is activated by calmodulin about 1.6-fold (Fig. 5B).
Effect of VA1 on Phosphophosphorylase Kinase Phosphatase Activity of Calcineurin-In addition to p-nitrophenyl phosphatase activity of calcineurin, phosphoenolpyruvate phosphatase activity is also inhibited by VA1 antibody in a metal ion-dependent manner (8). To test if phosphoprotein phosphatase activity of calcineurin was inhibited in a like way, the dephosphorylation of phosphorylase kinase was examined. Fig. 6 shows that both Ni2+ and Mn2+ support calcineurinphosphatase activity toward phosphophosphorylase kinase. In this case, the Mn2+-stimulated calcineurin-phosphatase activity is slightly higher than that of the Ni2+-stimulated enzyme, while Ni2+-stimulatedp-nitrophenyl phosphatase activity is higher than Mn2+-stimulated enzyme activity. Similar to reactions involving nonprotein substrates, VA1 only inhibits Ni2+-stimulated activity toward phosphorylase kinase.
Demonstration of Calcineurin and VAl Interaction by HPLC Gel Filtration-In view of the specific inhibition of Ni2+activated calcineurin activity by VA1, it is possible that the antibody binds only to the Ni2+-induced enzyme conformation and not to the Mn2+-induced form of calcineurin. Alternatively, the antibody, although complexing with all forms of calcineurin, may only affect the activity of the Ni2+-induced calcineurin conformation. To distinguish between these possibilities, direct interaction between various forms of calcineurin and VA1 monoclonal antibody was examined using   (0). Dephosphorylation experiments were started by the addition of phosphophosphorylase kinase (19.8 pg). At the times shown above, aliquots of the reaction mixture were removed, and the reaction was stopped with 5% trichloroacetic acid, and radioactivity in the supernatant counted.

Incubation Time (min)
HPLC gel filtration. Fig. 7 shows the gel filtration profiles of calcineurin and of VA1 alone and of calcineurin and VAl antibody preincubated in the presence of EDTA, Ni2+, or Mn2+. Purified calcineurin was resolved in a single protein peak (Fig 7A), while the VA, filtration profile showed one major protein peak and a minor shoulder peak, the latter of which represents transferrin contamination of VAl (Fig. 7B). On this gel filtration column calcineurin and VA, eluted at very similar positions despite their different molecular weights. Four poorly resolved peaks were seen when calcineurin and VA, mixtures were preincubated with EDTA or Ni2+ and chromatographed (Fig. 7, C and D). The elution times of the second, third, and fourth peaks correspond to those of VA,, calcineurin, and transferrin, respectively. The first peak, corresponding to a component with a higher molecular weight than any single protein applied to the column, contains both calcineurin and VA, as shown by NaDodS0,-PAGE and represents the enzyme-antibody complex. When calcineurin and VA1 were preincubated with Mn2+ and then chromatographed (Fig. 7E), the first peak seen in Fig. 7, C and D, was reduced to a shoulder of a larger second peak. Although the first peak was not as pronounced in this case, NaDodS04-PAGE revealed the presence of both calcineurin and VA1, indicating the occurrence of complex formation. A preincubated mixture of purified nonimmunized mouse immunoglobulin G and calcineurin was chromatographed as a control for the specificity of the calcineurin-VAl interaction (Fig. 7F). Two very poorly resolved peaks are apparent while no peak corresponding to an antibody-enzyme complex is seen.

DISCUSSION
Five hybridoma cell lines producing specific calcineurin antibodies have been produced in our laboratories. We have investigated the usefulness of these antibodies as probes to The volume of each sample was adjusted to 180 pl with elution buffer. Samples were filtered, and 160 pl of each sample were injected on the HPLC. The absolute protein amount injected on HPLC was 30 pg of calcineurin, 90 pg of VA, antibody, and 45 pg of mouse immunoglobulin G. The eluate (18-28 min) was fractionated every minute (0.33 ml/fraction) and analyzed by NaDodSO4-PAGE. Stained bands of transferrin, CY subunit of calcineurin, and the heavy chain of VAI are represented by T, a, and H, respectively. The solid arrow (+) and the broken arrow ( -4 indicate the elution times of VAI and calcineurin on HPLC, respectively. investigate the structural and functional properties of calcineurin. The studies described here exclusively involved the monoclonal antibody VA1, since it is the only antibody which significantly inhibits the phosphatase activity of calcineurin. While the other antibodies produced interact with the a subunit of calcineurin, this antibody binds to the @ subunit of the protein.
The inhibition of phosphatase activity by a calcineurinspecific monoclonal antibody demonstrates that this enzymatic activity is associated with calcineurin rather than due to a contaminant in the protein preparation. The observation that VA1 inhibits the activity of calcineurin toward certain protein and nonprotein substrates indicates that these multiple activities reside in the same protein species.
Marked inhibition of calcineurin-phosphatase activity was observed only when calcineurin was activated by Ni", while Mn2+-activated calcineurin activity was only marginally affected. The inhibition by the antibody of Ni2+-activated phosphatase activity is incomplete since significant activity is still observed at a saturating antibody level. Kinetic analysis indicates that inhibition of Ni2+-activated calcineurin activity by the monoclonal antibody does not appear to be due to the binding of the antibody to the active site of the enzyme; this is not surprising since the antibody is directed against the @ subunit of calcineurin while the catalytic site is reported to reside on the a subunit of the enzyme (18). The antibody also does not affect Ni2+ binding to calcineurin. Since both Ni2+ and Mn2+ appear to bind to the same site(s) on calcineurin (IO), the differential inhibition of these metal ion-stimulated enzyme activities by the antibody does not seem to be related to the interaction of antibody with the metal binding site(s) on the enzyme. The antibody VA1 complexes with Ni2+activated, Mn2+-activated, and metal-free calcineurin, demonstrating that the absence of enzyme inhibition observed with Mn2+-activated calcineurin is not due to a lack of enzyme-antibody interaction.
As well as inhibiting Ni2+-stimulated enzyme activity, VA1 antibody also reduces the calmodulin-stimulated activity of calcineurin. This may be due to one of the following possibilities: (i) it is possible that the antibody decreases the affinity of calcineurin for calmodulin; (ii) that calmodulin binding to calcineurin is blocked by antibody; or (iii) that the enzyme affinity for calmodulin binding is unaffected but that the calmodulin-induced conformational change of calcineurin is inhibited by the antibody. The observation that increasing amounts of calmodulin do not overcome the reduction in calmodulin-stimulated enzyme activity makes the first possibility seem unlikely. The second possibility is also unlikely, since some calmodulin stimulation, although reduced, is observed even in the presence of saturating levels of antibody and since the antibody binds to the @ subunit of calcineurin, while the calmodulin-binding site is located on the a subunit (1).
In a previous study (10) we have suggested that calcineurin may exist in various metal ion-and Ca2+/calmodulin-dependent forms, based on observations that the mechanism and rates of calcineurin activation by Ni2+ and Mn2+ are different and that Ni'+-and Mn2+-activated forms of calcineurin are not readily interconvertible except via an interaction with Ca2+/calmodulin. The differential inhibition of Ni2+-and Mnz+-activated calcineurin by the monoclonal antibody further supports such a suggestion. Our results indicate that the @ subunit is essential for Ni2+ stimulation but not Mn2+ stimulation of calcineurin activity. It is also possible that the / 3 subunit plays a role in both Ni2'-and Mn2+-stimulated enzyme activity and that the antibody is able to prevent the former but not the latter process. In either case, inhibition may only be due to interference with the ability of the @ subunit to modulate enzyme activity, suggesting that the conformational effects of antibody binding are confined to the @ subunit of calcineurin. However, the antibody-induced inhibition of Ni2+-stimulated calcineurin activity also results in a reduction of the calmodulin stimulation of enzyme activity. The location of the calmodulin-binding domain on the a subunit of calcineurin and our above discussion of the likelihood that the reduction in calmodulin-stimulated enzyme activity is due to an inhibition of the calmodulin-induced conformation change of calcineurin suggest that antibody binding to the @ subunit effects a conformational change in the @ subunit of the enzyme which is conferred to the a subunit. Since both the calmodulin-dependent and -independent Ni2+-stimulated but not Mn2+-stimulated enzyme activities are inhibited by the antibody, we would like to propose that Mn2+-and Ni2+-treated calcineurins possess different active conformations of which the Ni2+-induced conformation is prevented or reversed by VA1, while the Mn2+-induced conformational change occurs or is not reversed even after antibody complexation.
Since the inhibition of phosphatase activity is not due to the binding of VAI at the active site of calcineurin, our results do not indicate that the @ polypeptide is the catalytic subunit of the enzyme. It is interesting to note that all other monoclonal antibodies produced against the a subunit do not inhibit calcineurin activity (see Footnote 1 and Ref. 18). The results of this study suggest that @ subunit plays a critical role in the Ni2+ activation of calcineurin-phosphatase. The combined use of this monoclonal antibody and physical techniques may reveal the nature of the involvement of the @ subunit in the Ni2+-induced conformational changes of calcineurin.
The application of monoclonal antibody to the elucidation of structure-function relationships and regulatory properties of proteins has been stressed by many investigators (for review, see Ref. 19). The present study is limited to answering a few questiom which, in our opinion, are of fundamental importance to understanding the mode of action of calcineurin. The results demonstrate that both certain protein and nonprotein phosphatase activities are intrinsic to calcineurin, support the suggestion that calcineurin exists in various metal ion-dependent conformational states, and suggest that @ subunit plays a role in the Ni'+-induced conformational changes of the enzyme. Clearly, monoclonal antibodies may be used to investigate many other properties of calcineurin. For example, previous studies have suggested that calcineurin in cells or in crude tissue extracts may contain a tightly bound metal ion activator (10, 20). Monoclonal antibodies may be used to test such a possibility and to identify the physiological metal ion activator. In addition, monoclonal antibodies specific for the two different subunits of calcineurin may be used to test whether the two subunits have identical subcellular distributions under various physiological conditions.