Ca2+ Binding Effects on Protein Conformation and Protein Interactions of Canine Cardiac Calsequestrin”

Calsequestrin is a Ca2+-binding protein located intra- luminally in the junctional sarcoplasmic reticulum (SR) of striated muscle. In this study, Ca2+ binding to cardiac calsequestrin was assessed directly by equilibrium di-alysis and correlated with effects on protein confor- mation and calsequestrin’s ability to interact with other SR proteins. Cardiac calsequestrin bound 800-900 nmol of Ca2+/mg of protein (35-40 mol of Ca”+/ mol of calsequestrin). Associated with Ca2+ binding to cardiac calsequestrin was a loss in protein hydropho- bicity, as revealed with use of absorbance difference spectroscopy, fluorescence emission spectroscopy, and photoaffinity labeling with the hydrophobic probe 3-(trifluoromethyl)-3-(rn-[12S]iodophenyl)diazirine. Ca2+ binding to cardiac calsequestrin also caused a large change in its hydrodynamic character, almost doubling the sedimentation coefficient. We observed that car- diac calsequestrin was very resistant to several proteases after binding Ca2*, consistent with a global ef- fect of Ca2+ on protein conformation. Moreover, Ca2+ binding to cardiac calsequestrin completely prevented its interaction with several calsequestrin-binding proteins, which we identified in cardiac junctional SR vesicles for the first time. The principal calsequestrin-binding protein identified in junctional SR vesicles exhibited an apparent M, of 26,000 in sodium dodecyl


Ca2+ Binding Effects on Protein Conformation and Protein
Interactions of Canine Cardiac Calsequestrin" (Received for publication, May 27, 1987) Robert D. Mitchell Calsequestrin is a Ca2+-binding protein located intraluminally in the junctional sarcoplasmic reticulum (SR) of striated muscle. In this study, Ca2+ binding to cardiac calsequestrin was assessed directly by equilibrium dialysis and correlated with effects on protein conformation and calsequestrin's ability to interact with other SR proteins. Cardiac calsequestrin bound 800-900 nmol of Ca2+/mg of protein (35-40 mol of Ca"+/ mol of calsequestrin). Associated with Ca2+ binding to cardiac calsequestrin was a loss in protein hydrophobicity, as revealed with use of absorbance difference spectroscopy, fluorescence emission spectroscopy, and photoaffinity labeling with the hydrophobic probe 3-(trifluoromethyl)-3-(rn-[12S]iodophenyl)diazirine. Ca2+ binding to cardiac calsequestrin also caused a large change in its hydrodynamic character, almost doubling the sedimentation coefficient. We observed that cardiac calsequestrin was very resistant to several proteases after binding Ca2*, consistent with a global effect of Ca2+ on protein conformation. Moreover, Ca2+ binding to cardiac calsequestrin completely prevented its interaction with several calsequestrin-binding proteins, which we identified in cardiac junctional SR vesicles for the first time. The principal calsequestrinbinding protein identified in junctional SR vesicles exhibited an apparent M, of 26,000 in sodium dodecyl sulfate-polyacrylamide gels. This 26-kDa calsequestrin-binding protein was greatly reduced in free SR vesicles and absent from sarcolemmal vesicles and was different from phospholamban, an SR regulatory protein exhibiting a similar molecular weight. Our results suggest that the specific interaction of calsequestrin with this 26-kDa protein may be regulated by Ca2+ concentration in intact cardiac muscle, when the Ca2+ concentration inside the junctional SR falls to submillimolar levels during coupling of excitation to contraction.

~ ~ ~
Calsequestrin is a Ca2+-binding protein localized inside the junctional SR' of cardiac muscle and skeletal muscle which * This work was supported by National Institutes of Health Grants HL28556 and HL06308, a grant-in-aid from the American Heart Association, and the Herman C. Krannert Fund. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
-  binds Ca2+ with high capacity and moderate affinity (1-4) and has been shown to interact with the luminal face of the junctional SR membrane (5-8). Although the Ca2+-binding properties and associated structural changes of skeletal muscle calsequestrin have been extensively characterized (9), little is known about the Caz+-binding properties of cardiac calsequestrin. Indeed, the protein was not conclusively identified and purified from cardiac SR until relatively recently (3, 4).
The carbocyanine dye Stains All has proven very useful for detecting the various tissue forms of calsequestrin (10, 11) due to the dye's ability to stain calsequestrin an intense blue after sodium dodecyl sulfate-polyacrylamide gel electrophoresis (3, 4). Use of this protein stain along with antibodies specific for the cardiac form of calsequestrin has recently allowed detection of a similar protein in SR vesicles isolated from certain types of smooth muscle (12).
Skeletal muscle and cardiac calsequestrins have recently been sequenced by recombinant DNA methods; and although both proteins are found to be highly acidic, the sequence homology (approximately 60%) is not particularly striking (13, 14). However, both proteins are readily purified by Ca2+dependent elution from phenyl-Sepharose (4), a hydrophobic matrix, suggesting that in addition to the high density of negatively charged residues, hydrophobic regions on calsequestrin may be important for overall function. In view of the probable participation of cardiac calsequestrin in the Ca2+ release process at the junctional SR of cardiac muscle, this study was undertaken to quantitate Caz+ binding to the protein and to characterize the associated structural changes which account for calsequestrin's interaction with phenyl-Sepharose and the SR membrane. For the first time, a set of proteins localized to the junctional SR of cardiac muscle are identified which bind cardiac calsequestrin specifically with high affinity and in a Ca2+-sensitive fashion.

DISCUSSION
In this study, we have assessed Ca2+ binding to cardiac calsequestrin by five independent methods including equilibrium dialysis, spectrophotometric analysis, hydrophobic probe labeling, proteolysis susceptibility, and measurement of the ability of the protein to bind to receptor proteins in native membranes and on nitrocellulose blots. All methods are in general agreement, suggesting that cardiac calsequestrin binds Ca2+ with high capacity and moderate affinity and that there is an apparent specificity for Caz+ relative to M$+. The latter Portions of this paper (including "Experimental Procedures," "Results," Figs. 1-11, and Table I) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

Ca2+ Effects on
Cardiac Calsequestrin 1377 four methods and the hydrodynamic studies in particular suggest a rather large conformational change in cardiac calsequestrin upon binding Ca2+, which is associated with the burying of hydrophobic amino acid side chains and a resultant decrease in calsequestrin's ability to interact with other proteins. These findings are consistent with our earlier results (4) on purification of cardiac calsequestrin by phenyl-Sepharose chromatography. The Ca2+ concentration dependence for effects on structural parameters in all cases closely paralleled Ca2+ binding to the protein measured directly, suggesting that Ca2+ binding and resultant conformational changes occurred in a closely coordinated fashion. The number of Caz+-binding sites calculated for cardiac calsequestrin (35-40 mol/mol) appears to be very similar to that measured for the skeletal muscle protein (40-50 mol/ mol, Refs. 1, 2, and 40-42). The half-maximal Ca2+ concentrations required for binding were also approximately the same: 400-600 PM Ca2+ when measured in the presence of 150 mM KCl, and 100 JLM Ca2+ when measured in 20 mM KC1. The complete amino acid sequences of both rabbit skeletal muscle and canine cardiac calsequestrins have now been deduced by recombinant DNA methods. Skeletal muscle calsequestrin contains 103 acidic amino acid residues (13), with an excess of 73 negatively charged residues over basic residues; whereas cardiac calsequestrin contains 109 acidic amino acid residues, with an excess of 69 negatively charged residues over basic residues (14). The many acidic residues found for both cardiac and skeletal muscle calsequestrins are undoubtedly important in determining their large Caz+-binding capacities, and the similar numbers of these residues for both proteins are consistent with the two proteins binding similar amounts of Ca2+. Campbell et al.
(3) earlier reported that cardiac calsequestrin bound one-half to one-third as much Ca2+ as that reported here. However, in this earlier study, only one Ca2+ concentration was tested; and it is not clear if correction was made for protein loss during dialysis.
Earlier fluorescence studies of skeletal muscle calsequestrin (43) and our more recent observation of the interaction of the protein with phenyl-Sepharose (4) suggested that a hydrophobic patch was present on the protein's surface and that Ca2+ caused this hydrophobic site to internalize. In the work described here on cardiac calsequestrin, internalization of tryptophan residues was clearly indicated by the absorbance difference spectrum and the blue-shifted fluorescence spectrum obtained when Ca2+ was bound to the protein. Likewise, labeling of cardiac calsequestrin with the hydrophobic probe ['251]TID was substantially decreased when Ca2+ was bound to the protein. In a preliminary study to determine the site(s) labeled by ['251]TID, calsequestrin was cleaved with trypsin, and the products were fractionated by reverse-phase high performance liquid chromatography (data not shown). Only one peptide was observed to contain ['251]TID. These results obtained with cardiac calsequestrin are in agreement with the recent observations of Fliegel et al. (13), who identified a single hydrophobic site on skeletal muscle calsequestrin with the use of the hydrophobic probe [3H]trifluoperazine.
Sedimentation analysis demonstrated that cardiac calsequestrin underwent a dramatic shift in conformation upon binding Ca", as evidenced by a doubling of the sedimentation coefficient and a reduction in the partial specific volume. Our observations appeared consistent with those previously reported for skeletal muscle calsequestrin, which were attributed to an extended molecule becoming more compact upon binding ea2+ (9,34). With this in mind, we predicted that the Stokes radius (measured by gel filtration) of cardiac calsequestrin would decrease when Ca2+ was bound to the protein.
This has been noted previously for the skeletal muscle protein (34). We found, however, that the apparent Stokes radius of the molecule actually increased slightly in Ca2+-containing buffer, yielding data which were consistent with protein dimerization upon binding Ca". Since both aggregation and crystallization of skeletal muscle calsequestrin are well-known to occur in Ca2+-containing or low ionic strength buffers (9, 44), we were sensitive to this point. Sedimentation studies on cardiac calsequestrin were also performed in which the KC1 concentration was varied from 15 to 500 mM in either the absence or presence of Ca2+ (data not shown). The results clearly indicated that although a small degree of aggregation did occur at low KC1 concentrations, the predominant effect of the addition of CaZ+ was always a doubling of the sedimentation coefficient, again consistent with dimerization of the molecule. However, we detected no binding of 'z51-calsequestrin to nonradioactive calsequestrin in the protein blotting experiments. Therefore, if calsequestrin does dimerize under native conditions, it has probably lost this ability after sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transfer to nitrocellulose.
Of particular interest was the finding that cardiac calsequestrin in the Ca2+-bound state was highly protected from proteolytic digestion. This protective effect may have been due to one of several alterations in calsequestrin's physical properties upon binding Ca2+. As shown by our hydrophobic localization and hydrodynamic studies, in the presence of Ca2+, the protein evidently underwent a conformational change which 1) internalized hydrophobic residues, 2) resulted in a more compact structure, and 3) perhaps resulted in dimerization of the molecule. Since the proteases tested varied widely in their specificities, the loss in protease sensitivity associated with Ca2+ binding to calsequestrin suggested that the conformational change was global in nature and/or that charge alterations may have obscured protease specificity requirements. Thus, in the Ca2+-bound state, calsequestrin has apparently lost its ability to interact with proteins other than itself. This conclusion was substantiated by measuring 1251-calsequestrin binding to junctional SR vesicles and to junctional SR proteins transferred to nitrocellulose paper.
The specific association of calsequestrin with junctional SR proteins causing it to anchor to the membrane has long been postulated, both by ourselves (4) and by others (5-8, 35). In this study, we identified for the first time a major calsequestrin-binding protein of apparent M, = 26,000 which was localized to junctional SR vesicles isolated from both cardiac and skeletal muscle. Calsequestrin bound to this 26-kDa protein with high affinity (nanomolar concentrations of calsequestrin were used) and in a Ca2+-regulated fashion. Our in vitro results suggest that when Caz+ is stored inside the junctional SR of intact muscle, calsequestrin would not be able to bind to this protein. When Ca2+ is released from inside the junctional SR during muscle contraction, calsequestrin binding to the 26-kDa protein (as well as to the other minor proteins detected) could occur. Whether the 26-kDa protein is involved in the Ca2+ release process is presently unknown, but is an attractive hypothesis for future testing. Also, whether this protein is related to the 30-kDa protein previously localized to skeletal muscle junctional SR vesicles by Campbell et al. (7) is also presently unknown. Work in our laboratory is currently being directed toward purification and biochemical characterization of this 26-kDa calsequestrinbinding protein.