PRODUCTION OF AN ANTIBODY IN RABBIT AND DEVELOPMENT OF A RADIOIMMUNOASSAY*

Calmodulin, a heat-stable Ca2+-binding protein (Mr = 16,700) found in all eukaryotes, is a multifunctional modulator, mediating many of the effects of Ca2+ in cellular functions. The protein was derivatized with 1-fluoro-2,4-dinitrobenzene (DNB) to give 3 mol of DNB/mol of calmodulin (DNB3-calmodulin). The dinitrophenylated protein was almost as active as native calmodulin in stimulating bovine brain Ca2+-dependent phosphodiesterase. Incorporation of the dinitrophenyl groups renders calmodulin highly antigenic in the rabbit; native calmodulin is a weak antigen. Rabbits immunized with DNB3-calmodulin produced specific antibody against both DNB3-calmodulin and calmodulin. Using the immunized serum, a radioimmunoassay was developed for calmodulin, the sensitivity for DNB3-calmodulin and calmodulin being approximately 0.2 and 2 pmol, respectively. Although the sensitivity of the radioimmunoassay for calmodulin is comparable to the enzyme assay of calmodulin with Ca2+-dependent phosphodiesterase, the radioimmunoassay affords the detection of calmodulin on the basis of antigenic determinants, and thus measures calmodulin in terms of polypeptide structure instead of its ability to stimulate an enzyme. Further, the accuracy of the radioimmunoassay is not affected by the presence of a heat-labile inhibitor protein, which affects the enzyme assay to give an apparent underestimation.

Calmodulin, a heat-stable Ca'+-binding protein (Mr = 16,700) found in all eukaryotes, is a multifunctional modulator, mediating many of the effects of Ca2+ in cellular functions.
Incorporation of the dinitrophenyl groups renders calmodulin highly antigenic in the rabbit; native calmodulin is a weak antigen. Rabbits immunized with DNB3-calmodulin produced specific antibody against both DNB3-calmodulin and calmodulin. Using the immunized serum, a radioimmunoassay was developed for calmodulin, the sensitivity for DNB1-calmodulin and calmodulin being approximately 0.2 and 2 pmol, respectively.
Although the sensitivity of the radioimmunoassay for calmodulin is comparable to the enzyme assay of calmodulin with Ca2+-dependent phosphodiesterase, the radioimmunoassay affords the detection of calmodulin on the basis of antigenic determinants, and thus measures calmodulin in terms of polypeptide structure instead of its ability to stimulate an enzyme. Further, the accuracy of the radioimmunoassay is not affected by the presence .I a heat-labile inhibitor protein, which affects the enzyme assay to give an apparent underestimation.
Calmodulin,' a ubiquitous Ca"-dependent modulatory protein in all eukaryotes examined, was first discovered some 10 years ago in our laboratory as an activator of cyclic 3':5'nucleotide phosphodiesterase ( Ca2+ in a variety of cellular reactions and processes. The mode of action of calmodulin with respect to phosphodiesterase has been studied extensively. In the presence of micromolar Ca'+, calmodulin undergoes a conformational change to a more helical structure (6-9), which is the active configuration (10,11); the latter interacts with the apoenzyme of phosphodiesterase to form the active holoenzyme (12,13). The sequence of events leading to the stimulation of the enzyme may be outlined below: where CaM stands for calmodulin, E and E* . CaM* -Ca2', the apo-and holoenzyme of phosphodiesterase, respectively; and the asterisk (*) indicates the active conformation.
Implicit in this scheme is the regulatory role of Ca'+, the cellular flux of which is believed to govern the activity of the enzyme (14-16). This mechanism also appears applicable to brain adenylate cyclase (17) as well as erythrocyte Ca'+-ATPase (18). Calmodulin also regulates skeletal muscle phosphorylase kinase (19), myosin light chain protein kinase (20)(21)(22)(23), plant NAD kinase (24), Ca2+ transport in erythrocytes (25-29) and sarcoplasmic reticulum (30), phosphorylation of membranes (31,32), and the disassembly of microtubules (33). The mode of action of calmodulin on these systems has not been established.
Cahnodulin is a single polypeptide consisting of 148 amino acids, with a molecular weight of 16,700 and a p1 of 4.3 (34). The amino acid sequence of calmodulin corn bovine brain (34) has been completed, and it appears virtually identical with the partial sequence of bovine uterus (35) and rat testis (36). This, together with the biological (37-39) as well as immunological (40) cross-reactivity of calmodulin from phylogenetically distant organisms, suggests that the amino acid sequence of the protein has been highly conserved.
Although there have been numerous studies on the isolated protein, very little has been done on the physiblogical functions of calmoduiin at the cellular level. An antibody against calmodulin can be a highly useful tool in physiological studies. As part of our long term interest in the biological roles of calmodulin, we started several years ago to prepare an anticalmodulin serum in the rabbit. Repeated efforts failed to elicit an antibody against calmodulin, even when injected absorbed to poly(L-lysine) (41), polymerized with ethyl chloroformate to produce an insoluble complex (42), or crosslinked to thyroglobulin (43). In addition, the thyroglobulinlinked calmodulin was used to immunize chickens and frogs. In all cases, little or no antibody was detected that would Radioimmunoassay for Calmodulin recognize the native calmodulin, although using thyroglobulin-calmodulin as the antigen, an antibody was obtained in the rabbit. The antibody produced was directed against the thyroglobulin moiety rather than calmodulin. The difficulty in raising an anti-calmodulin probably reflects the fact that 6565 cahnodulin. Immunization of Rabbits-Two male New Zealand white rabbits were immunized with derivatized calmodulin containing 3 mol of DNB/mol of calmodulin (DNBo-cahnodulin). The antigen was prepared by mixing 1 ml of DNBS-calmodulin (0.84 mg of protein/ml) with 1 ml of Freund's complete adjuvant .  The suspension  was emulsified by a brief sonication,  injected  subcutaneously  into the animal  at four to five sites around  the scapula on Days 1, 17, 33, and 64. On Day 72, each rabbit was bled from the marginal vein of the ear. The blood was allowed to clot and the whole serum was stored in small aliquots at -90°C. For radioimmunoassay, the two sera were mixed in a 1:l volume ratio.
calmodulin is small and acidic and that it lacks tissue and species specificity, attributes generally characteristic of a weak immunogen.
We now report the production in the rabbit of an antibody specific for calmodulin using dinitrophenylated calmodulin as the antigen. While the incorporation of several dinitrophenyl groups into calmodulin did not significantly affect its biological activity, these groups render the derivatized calmodulin highly antigenic, and effectively evoke the production of an antibody against calmodulin.
In addition, we have developed a radioimmunoassay for calmodulin. While this work was in progress, Dedman et al. (44) reported the production of an immunoglobulin against native cahnodulin in the goat. The availability of an antibody against and a radioimmunoassay for this important regulatory protein will aid future studies on its cellular and subcellular localization. as well as its role in various biological functions.  (Fig. 2). On the next bleeding (Day 72), the serum reacted positively with both proteins; moreover, the serum titer increased with subsequent boostings. While the titer for DNBs-calmodulin leveled off after Day 72, that for calmodulin continued to increase until Day 140, and decrease gradually thereafter.
The serum from the first bleeding probably contained antibody directed against the dinitrophenyl groups, and additional boosting elicited the production of an antibody directed against other groups of the antigen. The serum obtained on Day 72 was found suitable to develop a radioimmunoassay. The serum titer for DNBs-cahnodulin was markedly higher than that for calmodulin throughout the course of immunization.
The results in Fig. 2 depict the immunization schedule and time course production of anti-calmodulin in one of the two rabbits; that of the other rabbit (not shown) was qualitatively similar.
Production of Anti-Calmodulin Serum in Rabbits-Preliminary efforts to elicit an antibody directed against calmodulin in rabbits have been unsuccessful. Calmodulin appears to be a poor antigen in rabbits, probably because the protein in small, acidic, and its amino acid sequence highly conserved in widely divergent species. These early efforts involved the use of cahnodulin in various states: they include polymerization of calmodulin with ethyl chloroformate (42) to yield an insoluble complex, forming a complex with poly(L-lysine) (41), or chemically coupling to thyroglobulin by the carbodiimide procedure (43). In addition, the thyroglobulin a calmodulin complex was used to immunize chickens and frogs. As mentioned before, all of these attempts elicited little or no production of antibody against calmodulin.
These results strongly suggest that both the native and the modified calmodulin were not sufficiently antigenic to evoke an anti-calmodulin response; a corollary is that calmodulin must be made more antigenic by the incorporation of certain immunogenic groups. The dinitrophenyl group is known to be particularly effective in eliciting antibody production (48). We reasoned that the incorporation of an appropriate number of dinitrophenyl groups into calmodulin might render it highly antigenic, and that the animal would not only make antibodies against the dinitrophenyl groups but probably also against some endogenous groups of the protein. This approach proved to be successful.
Calmodulin was derivatized with 1-fluoro-2,4-dinitrobenzene; the nitrophenylated calmodulin appeared yellow. The extent of derivatization was proportional to the incubation time; for example, a reaction time of 15 and 60 min incorporated 3.2 and 5.9 mol of DNB/mol of calmodulin, respectively. Although both calmodulin derivatives retained biological activity (Fig. l), the efficacy diminished with the extent of derivatization.
Half-maximum stimulation of phosphodiesterase was obtained with 16 ng of native calmodulin, 20 ng of DNBs-calmodulin, and 240 ng of DNBs-calmodulin.
Since the biological activity of DNBs-calmodulin did not differ signiticantly from that of the native molecule, the derivative was chosen as the antigen for subsequent immunization of the rabbit.
Two rabbits were immunized with DNBs-calmodulin emulsified with Freund's complete adjuvant as described under "Experimental Procedures." The immunized rabbits were bled one week following each injection, and the sera were tested for the presence of antibody directed against calmodulin or DNB3-calmodulin.
One week after the fist injection, the serum reacted positively with DNBa-calmodulin but not with 1 FIG. 1. Effect of dinitrophenylation on the biological activity of calmodulin. The derivative was assayed for its ability to stimulate calmodulin-deficient bovine brain phosphodiesterase (51). The number of DNB/calmodulin is indicated by the subscripts.
The immune serum from the rabbit did not give a positive Ouchterlony test, indicating that the anti-calmodulin is not a precipitating antibody. The goat antibody against calmodulin is also nonprecipitating (44). Radioimmunoassay of Calmodulin-Cahodulin iodinated according to the chloramine-T procedure yielded a specific activity of 90 @i/pg, equivalent to 0.8 mol of '251/mol of calmodulin.
Calmodulin has two tyrosines, one is buried while the other is exposed (7); iodination probably takes place primarily on the exposed residue. However, we did not determine whether the buried tyrosine was also iodinated, or to what extent. Richman and Klee (52) have recently reported that iodination of calmodulin, up to 1 mol of iodide/m01 of calmodulin, did not significantly affect its ability to stimulate Ca2+-dependent phosphodiesterase. Fig. 3A shows the binding of 1251-calmodulin or iz51-DNB3calmodulin to various concentrations of anti-calmodulin serum. Binding of 1251-cahnodulin was detected with 0.1~1 of the antiserum, whereas much less antiserum was needed to give a detectable binding of '251-DNB3-calmodulin.
Indeed, 0.001 ~1 of antiserum bound 15% of the '251-DNB3-calmodulin, indicating that the antiserum has a higher titer for DNBJcalmodulin than for calmodulin, in agreement with the data shown in Fig. 2. Fig. 3B  The sensitivities of the assay were approximately 30 ng (or 2 pmol) of calmodulin or 3 ng (or 0.2 pmol) of DNBs-calmodulin, while the upper limits of the assay for both molecules were approximately 1000 ng. These data demonstrate that the radioimmunoassay for nitrophenylated calmodulin is 10 times more sensitive. Conceivably, the increased sensitivity by nitrophenylation may be exploited in the future to give a more sensitive radioimmunoassay for calmodulin. Note that the level of calmodulin, which is usually high in many cells and tissues, can be accurately determined with the sensitivity of the radioimmunoassay.
The extent of binding of 1251-calmodulin to the antibody was affected by the Ca2+ concentration in the assay mixture (Fig.  4). In the presence of EGTA, the amount of '251-calmodulin bound was twice that in the presence of Ca'+. Calmodulin is known to undergo a conformational change to a more helical structure in the presence of Ca2+ (6-9). The altered conformation in the presence of Ca2+ may account for the apparent decrease in avidity of the immunoglobulin for calmodulin. An alternative explanation is that the rabbit serum contains other proteins which bind calmodulin in a Ca2+-dependent manner. In the presence of Ca'+, these components compete with the antibody for binding of '251-calmodulin, resulting in an apparent decrease of '251-calmodulin available to the antibody. In the presence of EGTA, this competition could be eliminated. To test the validity of this notion, the radioimmunoassay of calmodulin was carried out in the presence of a heat-labile calmodulin-binding protein, referred to as an inhibitor protein of CAMP metabolic enzymes (1,46,51,(53)(54)(55) and erythrocyte Ca2'-ATPase (18). As shown in Fig. 5, Fig. 4 when the radioimmunoassay was done in the presence of EGTA, the amount of '251-calmodulin bound to the antibodies was considerably higher; moreover, the titration curve is shifted to the left, giving a greater sensitivity for the assay. The reason for this is not apparent.
Specificity of Anti-Calmodulin-To examine the specificity of the antibody, its cross-reactivity with troponin and its individual subunits troponin-C, troponin-I, and troponin-T, as well as the heat-labile inhibitor protein of CAMP metabolic enzymes from bovine brain (46), was examined. Troponin-C possesses greater than 50% direct sequence homology and Radioimmunoassay for Calmodulin 25 protein profile. In the control experiment using the nonimmune serum, essentially all the protein passed through the to anti-calmodulin serum. Preparation of tissue extract and radioim-with the heat-denatured proteins, '251-calmodulin was added munoassay were performed as described under "Experimental Pro-to a 100,000 x g supernatant fluid of bovine brain extract (as cedures." Znh., inhibitor protein; TnZ, troponin-I; Z'nZ', troponin-?', described under "Experimental Procedures"), and the solution Tn, troponin; TnC, troponin-C.
was heated in a boiling water bath for 5 min. The denatured with a calmodulin-agarose affinity column. Panel A, immune serum; Panel B, nonimmune serum. Chromatography was carried out as described under "Experimental Procedures." NaSCN absorbs at 280 nm which accounts for the elevated baseline at the point of elution by NaSCN. PBS (NaCl/PJ, phosphate-buffered saline. Fraction greater than 70% conservative sequence homology with calmodulin (34). The inhibitor protein contains an l&500-dalton subunit which migrates electrophoretically similar to calmodulin in a sodium dodecyl sulfate-acrylamide ge1.3 As shown in Fig. 6, none of these proteins displaced significantly the binding of '251-calmodulin at the concentrations tested, demonstrating that the antiserum is specific for calmodulin.
Another way of showing that the immune serum from the rabbit contains specific antibodies directed against calmodulin is the use of a cahnodulin-agarose affinity column. Fig. 7 depicts the isolation of immunoglobulin from the immune (Panel A) and nonimmune (Panel B) sera, respectively. With the immune serum, the bulk of the protein passed through the calmodulin-affinity column unimpeded. Subsequent elution with NaCl/Pi containing 1 mu EGTA or 1 M NaCl eluted no significant amount of protein from the column. However, changing the solution to 4 M NaSCN which is known to break the antigen.antibody complex (56) eluted a second, minor protein peak (Panel A). An aliquot of the different fractions collected from the column was assayed for its ability to bind '251-calmodulin. Only the fractions under the second protein peak manifested binding activity, which coincided with the ' R. W. Wallace and W. Y. Cheung, unpublished observation. proteins were removed by centrifugation, and the amount of radioactivity remaining in the supernatant fluid was determined. Essentially all the radioactivity (90 to 95%) remained in the supernatant fluid, indicating that the denatured proteins did not remove any significant amount of '251-calmodulin under this condition. The radioactivity remaining in the supernatant after heating could be quantitatively precipitated with 12.5% trichloroacetic acid, suggesting that no isotope was released by heat treatment and that the label in '251-calmodulin was presumably unaltered. Fig. 6 also shows the competition by a bovine brain extract for the binding of '251-calmodulin in the radioimmunoassay. The inhibition curve appears parallel to that of pure calmodulin, indicating the absence of interfering substances in the brain extract. According to the radioimmunoassay, approximately 10% of the protein in the heat-treated supernatant is calmodulin, giving the equivalent of 110 mg of soluble calmodulin/kg of tissue. The enzyme assay of calmodulin is based on its ability to stimulate a calmodulin-deficient phosphodiesterase, as shown in Fig. 8. According to the enzyme assay, the amount of calmodulin in the extract was 74 mg/kg of tissue. Repeat assay of calmodulin in the extract by the two procedures showed that the enzyme assay consistently gave a lower value. Since the enzyme assay was done in the presence of Ca"+, any other calmodulin-binding proteins present in the brain extract would compete with phosphodiesterase for calmodulin, resulting in its apparent underestimation.
Calmodulin-binding proteins such as the heat-labile inhibitor protein (46) would be inactivated by the heat treatment. However, bovine brain contains a heat-stable inhibitor protein which binds calmodulin in a Ca'+-dependent manner (57). The presence of this and perhaps other heat-stable, calmodulin-binding proteins could give an apparently lower level of calmodulin obtained by the biological assay.
To demonstrate the interference of the enzyme assay by calmodulin-binding proteins, calmodulin in the brain extract was measured in the presence of two concentrations of the heat-labile inhibitor protein. As shown in Fig. 8, the addition of the inhibitor protein shifted the activity titration curves to the right; the extent of the shift was in direct proportion to the concentration of exogenous inhibitor.
In the presence of 3.8 and 11.4 pg of the inhibitor protein, the apparent amount of calmodulin was markedly decreased to 5.8 and 2.2 mg/kg of tissue, reppectively. This experiment clearly shows that the presence of the inhibitor gave an apparently lower amount of calmodulin detectable in the brain extract and is in sharp contrast to the radioimmunoassay in which the presence of a comparable concentration of inhibitor (3.8 pg/tube) did not affect the accuracy of the assay (Fig. 5).
Although the sensitivity of the radioimmunoassay is comparable to that of the biological assay, the usable range in the radioimmunoassay is considerably wider (30 to 1000 ng uersus 10 to 100 ng). Moreover, the radioimmunoassay measures calmodulin on the basis of antigenic determinants, rather than biological activity. In addition, the radioimmunoassay determines calmodulin accurately, even in the presence of potential interfering components, such as the heat-labile inhibitor protein. DISCUSSION The diitrophenylation of cahnodulin can be used effectively to provoke the production of an antibody against calmodulin in rabbits. The difficulty in eliciting an antibody response to native calmodulin in the rabbit, frog, and chicken probably reflects the structural similarities between calmodulin from different species. The incorporation of 3 mol of dinitrophenyl residues/m01 of calmodulin apparently did not affect appreciably its biological activity. However, the presence of these groups exerts a profound effect on the immunogenicity of calmodulin, as shown by the production of an antibody against DNBB-cahnodulin as early as 1 week following immunization of the rabbit (Fig. 2). The greatly increased antigenicity of calmodulin by the incorporation of dinitrophenyl groups is reminiscent of that of cytochrome c (58), which is also a weak immunogen.
Upon coupling with one to two dinitrophenyl groups, cytochrome c becomes strongly antigenic.
The controlled incorporation of dinitrophenyl groups as a means to increase antigenicity may represent a general approach applicable to other proteins that are inherently poor antigens.
Although native calmodulin was antigenically weak in the rabbit, it was effective in the goat. Dedman et al. (44) recently demonstrated the production of an antibody against calmodulin in the goat using native calmodulin as the immunogen. The use of rabbit for producing antibody does have practical advantages: the animal is hardy, easily accommodated with a general animal facility, and is less expensive in its upkeep. The goat, on the other hand, provides a much larger volume of serum with each bleeding. This advantage, however, is easily compensated for by using a number of rabbits.
The curves obtained for the differential binding of lz51calmodulin to the antibody in the presence and absence of Ca2+ indicate the presence of conformation-dependent antigenie sites. The difference appears paradoxical because the antibody shows a greater affinity for the less helical conformation that is presumably present in the absence of Ca'+. During the systemic circulation in the host animal, the antigen is exposed to a millimolar concentration of Ca2+ in the serum and should assume the more helical conformation; therefore, the antibody should have a higher affinity for calmodulin in the radioimmunoassay carried out in the presence of Ca'+. The results presented in Fig. 4 appear contrary to this argument. One explanation for this paradoxical result, in addition to the ones enumerated earlier, may be that the coupling of dinitrophenyl groups to calmodulin exposes other antigenic sites that are accessible only in the absence of Ca'+.
Although the radioimmunoassay does not offer a greater sensitivity than the enzyme assay, it has the advantage of detecting a wider span of sample concentrations.
The presence of numerous Ca'+-dependent, calmodulin-binding proteins in the brain3 (59) and perhaps in other tissues (60,61), could cause an underestimation of the tissue level of calmodulin by the enzyme assay. The radioimmunoassay, which measures calmodulin in the presence of EGTA, does not appear susceptible to this interference, and thus would give a more accurate assay. In addition, the radioimmunoassay permits quantitation of calmodulin on the basis of antigenic determinants rather than biological activity. Thus, the method would allow the study of the turnover rate of calmodulin, using an appropriate system, for example, neuroblastoma cells. Calmodulin and troponin-C have more than 70% conservative sequence homology; moreover, they have many physical chemical properties in common. Yet, the rabbit sera directed against calmodulin does not recognize troponin-C. Although calmodulin effectively substitutes for troponin-C in skeletal muscle myosin ATPase (62), troponin-C has only about ?&a the effectiveness of calmodulin in stimulating Ca2+-sensitive phosphodiesterase (63). Thus, similarity in physical chemical properties of the two proteins does not necessarily portend their potential biological cross-reactivities.
Although the antibody raised from the immunized rabbits is nonprecipitating, the results presented here demonstrate collectively that the antibody raised in the rabbit is specific for calmodulin.
First, nonimmune serum did not contain any immunoglobulin that binds '251-calmodulin; in fact, even rabbits that had been injected with the native or one of several forms of modified calmodulin for a prolonged time did not produce any immunoglobulin with specific binding for lZ51calmodulin.
Secondly, a calmodulin-specific immunoglobulin was isolated by a calmodulin-agarose affinity column from the immune, but not the nonimmune serum. Thirdly, binding of the antibody for calmodulin is Ca"-dependent. Calmodulin is known to undergo a conformational change in the presence of Ca'+. Finally, troponin-C, which shares more than 70% of conservative amino acid sequence, also undergoes a Ca'+dependent conformational change, but is not recognized by the anti-calmodulin serum. Since the discovery of calmodulin as an activator of cyclic nucleotide phosphodiesterase some 10 years ago (2), extensive studies have been carried out to elucidate its functions at the biochemical level. Increasing evidence from many laboratories suggests that calmodulin is a multifunctional modulatory protein, mediating many of the Ca2+ effects in the eukaryotes. Although its ubiquitous distribution in the eukaryotes suggests it may be involved in many of the Ca*+-mediated processes such as cell motility and movement, excitation-contraction and stimulus-secretion coupling, cytoplasmic streaming, chromosome movement, axonal flow, and transmitter release, direct evidence in support of these roles is lacking. The availability of specific antibodies against calmodulin would permit studies that may shed light on these questions. Using immunofluorescent-tagged anti-calmodulin, Welsh et al. (64) found that the antibody decorates the mitotic spindle during cell division of fibroblast, indicating a possible role of calmodulin in chromosomal movement. Wood et al. (65) noted that in mouse basal ganglia, the antibody labeled the postsynaptic, densities and the microtubules associated with postsynaptic dendrites, suggesting that calmodulin may play a postsynaptic role. Cellular and subcellular localization of calmodulin may shed light on its other biological functions.
Of the Ca*+-binding proteins known to act as cellular regulators, calmodulin emerges as a major mediator of Ca2+ functions. Future studies of calmodulin should continue to be fruitful at both the biochemical as well as the physiological levels.