A Ca2+-dependent Tryptic Cleavage Site and a Protein Kinase A Phosphorylation Site Are Present in the Ca2+ Regulatory Domain of Scallop Muscle Na+-Ca2+ Exchanger*

Digestion of scallop muscle membrane fractions with trypsin led to release of soluble polypeptides derived from the large cytoplasmic domain of a Na+-Ca2+exchanger. In the presence of 1 mm Ca2+, the major product was a peptide of ∼37 kDa, with an N terminus corresponding to residue 401 of the NCX1 exchanger. In the presence of 10 mm EGTA, ∼16- and ∼19-kDa peptides were the major products. Polyclonal rabbit IgG raised against the 37-kDa peptide also bound to the 16- and 19-kDa soluble tryptic peptides and to a 105–110-kDa polypeptide in the undigested membrane preparation. The 16-kDa fragment corresponded to the N-terminal part of the 37-kDa peptide. The conformation of the precursor polypeptide chain in the region of the C terminus of the 16-kDa tryptic peptide was thus altered by the binding of Ca2+. Phosphorylation of the parent membranes with the catalytic subunit of protein kinase A and [γ-32P]ATP led to incorporation of 32P into the 16- and 37-kDa soluble fragments. A site may exist within the Ca2+ regulatory domain of a scallop muscle Na+-Ca2+ exchanger that mediates direct modulation of secondary Ca2+ regulation by cAMP.

The Na ϩ -Ca 2ϩ exchangers of the plasma membrane catalyze a secondary active transport process dependent on the Na ϩ electrochemical gradient generated by the Na ϩ ,K ϩ -ATPase and play a major role in cellular Ca 2ϩ homeostasis in many tissues (1). Three Na ϩ -Ca 2ϩ exchanger proteins (NCX1, NCX2, and NCX3) have been described in vertebrates (2)(3)(4)(5). The molecular biology of the Na ϩ -Ca 2ϩ exchanger (Calx) from Drosophila has been described (6,7), and an exchanger from squid has been reported (8). An electroneutral Na ϩ -Ca 2ϩ antiporter has also been found in mitochondria, where it may be involved in modulating matrix Ca 2ϩ in response to changes in cytoplasmic Ca 2ϩ concentration (9,10).
The protein moieties of the plasma membrane Na ϩ -Ca 2ϩ exchanger proteins are close in overall size (ϳ103-106 kDa) (4,11) to the SERCA 1 -type Ca 2ϩ -ATPase pumps, which have a molecular mass of ϳ110 kDa (12). In the case of NCX1, a signal sequence of 32 amino acid residues at the N terminus of the protein is removed post-translationally (13)(14)(15). From the N terminus, structure prediction algorithms suggest that five transmembrane segments lead to a large cytoplasmic, followed by four C-terminal transmembrane helices (14 -16). Two internal stretches of the f loop, Phe 407(375) -Asp 478(443) and Thr 538(506) -Tyr 613(581) , 2 show similarity to one another, and are termed the ␤-1 and ␤-2 repeats (6). A high affinity Ca 2ϩ binding region (K d , 0.1-3 M) is present in the large cytoplasmic domain (17)(18)(19)(20). This binds Ca 2ϩ cooperatively and provides regulation of the exchanger through the I 2 mechanism, in which increased levels of cytoplasmic Ca 2ϩ activate the enzyme (19). The regulatory Ca 2ϩ binding region involves the ␤-1 repeat and extends through the variable region between the two ␤ repeats to the beginning of ␤-2 (18). Two acidic triads, one at the C terminus of ␤-1 (Asp 478(446) -Asp-Asp) and one in the variable region, (Asp 530(498) -Asp-Asp), just N-terminal to ␤-2, are important components of this high affinity Ca 2ϩ binding site (18).
The isolated cardiac exchanger shows three bands on silverstained SDS gels: a glycosylated 120-kDa species corresponding to the native exchanger, a glycosylated 160-kDa polypeptide representing oxidized exchanger, and an unglycosylated polypeptide of 70 kDa, which arises by proteolysis of the 120-kDa protein at the Asp 289(257) -Gly or Asp 303(270) -Gly bonds in the large cytoplasmic domain (11,(21)(22)(23). Digestion of the exchanger with chymotrypsin leads to a loss of regulatory function, probably through proteolysis localized in the cytoplasmic f loop (4).
The study of the protein chemistry of the Na ϩ -Ca 2ϩ exchanger has been hampered by its low abundance; even in cardiac muscle, the NCX1 exchanger represents only 0.1-0.2% (w/w) of the sarcolemmal membrane protein (24). The mitochondrial exchanger is present at 0.4 g/mg total protein (10). Thus, silver staining has often been necessary to detect the protein on SDS gels (21).
In the past, a number of studies of the Ca-ATPase from the cross-striated adductor muscle of the deep sea scallop have used a deoxycholate-extracted membrane fraction enriched in fragmented SR (25)(26)(27)(28). In the course of examining the effect of trypsin on this preparation, it was observed that soluble polypeptides were released by the action of the protease in sufficient amounts for them to be detected by conventional staining of SDS gels with Coomassie Blue. Sequencing of these tryptic fragments showed that they were not derived from the Ca-ATPase but instead possessed N termini identifying them as overlapping stretches of polypeptide originating in the large cytoplasmic domain (f loop) of a Na ϩ -Ca 2ϩ exchanger. Because this region of the Na ϩ -Ca 2ϩ exchanger is crucial for its regulation through physiological mechanisms and because it may represent a possible target for pharmacological intervention, the polypeptides were further investigated.

EXPERIMENTAL PROCEDURES
Deep sea scallops (Placopecten magellanicus) were obtained from the Marine Biology Laboratory (Woods Hole, MA).
Preparation of Membranes Enriched in Sarcolemma-The B 1 fraction was prepared as described above (25,29,30).
Preparation of Deoxycholate-extracted Scallop Membrane Fractions-This was carried out when necessary with both the SL-rich (B 1 ) and SR-rich (B 2 ) fractions, as described previously (25).
Preparation of Soluble Fragments from Tryptic Digests-Membranes were typically suspended at 5-10 mg ml Ϫ1 in standard media of 20% (v/v) ethylene glycol (Pierce), 0.15 M KCl, 1 mM CaCl 2 or 10 mM EGTA-Na, 50 mM MOPS-Na, pH 7.0. Digestions were at room temperature for 10 -15 min. with N-tosyl-L-phenylalanyl chloromethyl ketone-treated trypsin (Sigma, dissolved in 1 mM HCl at 12,000 units/mg) added in a 1:30 (w/w) ratio to total membrane protein (giving 400 units of activity/mg total membrane protein). Digestions were terminated by addition of 4-(2-aminoethyl)benzenesulfonyl fluoride (Calbiochem) to a final concentration of 20 mM, followed by transfer of the sample to ice. The samples were centrifuged at 16,000 ϫ g for 1 ⁄2 h at 4°C in an Eppendorf microcentrifuge. The supernatant was collected and recentrifuged for 1 h at 10 5 ϫ g in a Beckman TL100 bench top ultracentrifuge at 4°C to remove traces of contaminating membranes.
Concentration of Soluble Peptides for Use in SDS-PAGE-Sodium deoxycholate was added to the high speed supernatant from the tryptic digest to a final concentration of 0.025% (w/v), followed by trichloroacetic acid to a final concentration of 6% (w/v). The trichloroacetic acid also served to completely inactivate the trypsin. After incubation on ice for 15 min, the preparation was centrifuged at 16,000 ϫ g for 1 ⁄2 h, the supernatant was carefully removed, and the pellet was washed with 200 l of diethyl ether. After drying, the pellet was dissolved in Tricine or Laemmli sample denaturation buffer.
Gel Electrophoresis-Discontinuous SDS-PAGE was carried out with glycine or Tricine as the trailing anion (31,32). Sodium thioglycholate (0.1 mM) was present in the sample denaturation medium and added to the cathode buffer to scavenge the gel ahead of the peptides for free radicals and oxidants.
N-terminal Sequencing-The proteolytic fragments were separated by SDS-PAGE, and the gels were electroblotted onto Immobilon-PSQ polyvinylidene difluoride membrane (Millipore) in a medium of 10% (v/v) MeOH, 10 mM CAPS-Na, pH 11 at 4°C. After lightly staining with Coomassie Blue, the bands of interest were cut out and sent to the Protein Chemistry Core of the University of Florida.
Phosphorylation of the Membranes with Protein Kinase A-The DOCextracted B 2 membrane fraction (1 mg) was incubated with 50 units of the catalytic subunit of PKA in a medium of 0.27 M sucrose, 80 mM KCl, 5 mM EGTA, 0.08 mM CaCl 2 , 29 mM NaF, 8.3 mM MgCl 2 , 20 mM MOPS-Na, pH 7.0 containing 0.1 mM [␥-32 P]ATP (500 dpm pmol Ϫ1 ) for 2 min at 30°C. The membranes were sedimented at 16,000 ϫ g and washed free of unbound radioactivity. The phosphorylated membranes were then divided into two parts, one of which was resuspended in 20% (v/v) ethylene glycol, 0.4 M KCl, 1 mM CaCl 2 , 50 mM MOPS-Na, pH 7.0, while the other was resuspended in 20% (v/v) ethylene glycol, 0.4 M KCl, 10 mM EGTA, 50 mM MOPS-Na, pH 7.0. Digestion with trypsin was then carried out as described above. The soluble products were separated from insoluble (membrane-bound) material and concentrated by the DOC-trichloroacetic acid procedure described above. The phosphorylated samples were electrophoresed in the Tricine SDS gel system, and autoradiography of the dried gel was carried out using Kodak X-Omat film.
Production of Rabbit Polyclonal Antibody against the ϳ37-kDa Soluble Tryptic Fragment-The soluble tryptic fraction formed in the presence of Ca 2ϩ was run on a Tricine SDS gel. The gel was then lightly stained with Coomassie Brilliant Blue, and the ϳ37-kDa band was cut out with a razor blade. The gel slice was homogenized in PBS (0.15 M NaCl, 10 mM sodium phosphate, pH 7.2) and Freund's complete adjuvant before injection into New Zealand White rabbits. Booster injections were made every third week using Freund's incomplete adjuvant. Partial purification of antibody from serum was by ammonium sulfate fractionation, followed by DEAE-Sephacel anion exchange chromatography to separate IgG and IgM (33).
Western Blotting-SDS gels were first blotted onto polyvinylidene difluoride (Immobilon PSQ), as in the procedure for N-terminal sequencing (see above). Nonspecific binding was blocked with 5% w/v nonfat milk in PBS-T (0.1 M NaCl, O.1% v/v Tween-20, 0.1 M NaP i , pH 7.4) for 1 h at room temperature. The blots were washed twice quickly in PBS-T, followed by one wash for 15 min and two washes for 5 min in PBS-T. The blots were incubated in rabbit anti-37-kDa antibody diluted 1:1000 with 5% (w/v) nonfat dried milk, 0.1% v/v Tween-20 in PBS-for 1 h at room temperature with gentle agitation. The blots were then washed as before with PBS-T. The blots were then incubated for 1 h in goat anti-rabbit IgG secondary antibody conjugated with horseradish peroxidase (Amersham Pharmacia Biotech) that had been diluted 1:1,500 in 5% (w/v) dried milk in PBS-T. After washing again with PBS-T, the blots were incubated with the ECL luminol system (Amersham Pharmacia Biotech).
Protein Concentration-This was by the bicinchoninic acid method (34).

RESULTS
The starting materials for the work reported here were membrane fractions prepared from the cross-striated part of the adductor muscle of the deep sea scallop. One of these (the B 2 fraction; see "Experimental Procedures") is enriched in fragmented SR but still contaminated by membranes derived from the SL (25,29). This fraction can be extracted with low (nonsolubilizing) concentrations of DOC to remove peripheral membrane proteins and other contaminants (25). On SDS gels of the DOC-extracted B 2 fraction, ϳ90% of the Coomassie Blue-staining material has a mobility corresponding to a molecular mass of 105-115 kDa ( Fig. 1 and Ref. 25), consistent with the size of the Ca-ATPase (25,35,36). The second scallop muscle fraction (B 1 ), which was used later in the studies, is enriched in SL (29).
Tryptic digests of the DOC-extracted scallop muscle B 2 membrane fraction were examined for soluble peptide fragments released by the action of the protease. It was found that a polypeptide of ϳ37 kDa was the major soluble species formed in the presence of 1 mM Ca 2ϩ (Fig. 2), although smaller amounts of soluble ϳ16and ϳ19-kDa peptides were sometimes also present. In the presence of 10 mM EGTA, the pattern of the products was reversed, so that the ϳ37-kDa fragment was either absent or present in much smaller amounts, and the two smaller polypeptides were the main soluble products. The relative proportions of ϳ16and ϳ19-kDa peptides varied; sometimes the two species were present in approximately the same amounts, whereas in other digests one form predominated. This may mean that the 16-and 19-kDa fragments are susceptible to further proteolysis; in contrast, the ϳ37-kDa peptide was relatively resistant to the further action of trypsin. All three soluble peptides could be clearly visualized on the gels with standard Coomassie Blue stain. Occasionally, larger soluble polypeptides of ϳ60 and 40 kDa were observed.
The precursor polypeptide for the soluble peptides had to be very close in size to the scallop SERCA (molecular mass, 110 kDa), because only material of 105-115 kDa was present in sufficient amounts in the starting preparation to account for the amounts of soluble peptides formed ( Fig. 1 and Ref. 25).
There are usually only small and variable traces of other integral protein membrane components on Coomassie Blue stained SDS gels of the B 2 fraction after DOC extraction, the most significant of these being a 28-kDa protein associated with the SL (25,29). The latter peptide is too small to be the precursor of the 37-kDa soluble tryptic fragment and in any case is present only at low levels.
The results of N-terminal sequencing of the soluble fragments are shown in Table I, where the data have been aligned to show the apparent relationships between the peptides and some sodium-calcium exchangers. The ϳ37-kDa peptide produced in the presence of Ca 2ϩ and the ϳ16-kDa polypeptide formed by digestion in the absence of Ca 2ϩ had the same N-terminal sequence. Comparison of this N-terminal sequence with known sequences using the EMBL data base (Blitz) indicated that its start corresponded to residue 401(369) of the NCX1 Na ϩ -Ca 2ϩ exchanger, at the N-terminal end of the ␤-1 repeat in the large cytoplasmic domain (loop f) (2,4). A lysyl or arginyl residue must precede this sequence for the 37-and 16-kDa fragments to be produced by the action of trypsin. There was no significant similarity to any part of the scallop Ca-ATPase polypeptide chain (36). Assuming an average residue mass of 110 Da, the ϳ16-kDa fragment was approximately 145 residues long and was likely to correspond closely to the segment of loop f in NCX1 that starts and ends with the two acidic triads, Asp 478(446) -Asp-Asp at the carboxyl end of ␤-1 and Asp 530(498) -Asp-Asp preceding the ␤-2 repeat. The Ca 2ϩ -insensitive cleavage site on the scallop precursor represented by the N terminus of the ϳ16and ϳ37-kDa soluble fragments will be designated the T 1 site. The C terminus of the ϳ16-kDa fragment must be located in a stretch of polypeptide chain inaccessible to trypsin when Ca 2ϩ is bound to the precursor protein but more susceptible to proteolysis when Ca 2ϩ is unbound. The Ca 2ϩ -sensitive proteolytic cleavage site represented by the C terminus of the ϳ16-kDa fragment will be designated the T 2 site. Satisfactory N-terminal sequencing of the ϳ19-kDa band peptide was not possible, because of microheterogeneity within that band on SDS gels, but the size of the fragment is consistent with it representing that part of the 37-kDa fragment which is C-terminal to the T 2 cleavage site. There is other evidence supporting this conclusion from the immunological and phosphorylation studies reported below. The tryptic cleavage site corresponding to the C terminus of the 37-kDa fragment, and possibly the C terminus of the 19-kDa fragment, will be designated the T 3 site.

Polyclonal Antibody against the 37-kDa Soluble Fragment Cross-reacts with the 16-and 19-kDa Soluble Tryptic Fragments and Also a 105-110-kDa Polypeptide in the Undigested
Membranes-The soluble fragments formed in the presence of Ca 2ϩ were separated on a Tricine SDS gel, and a Western blot was obtained using polyclonal rabbit IgG raised against the ϳ37-kDa peptide. All three peptides bound the antibody, indicating common epitopes between the 37-kDa peptide and the 16-kDa peptide and between the 37-kDa peptide and the 19-kDa peptide (Fig. 3).
Western blots of Tricine SDS gels of the undigested DOC extracted B 2 membrane fraction probed with rabbit anti-37 kDa peptide antibody showed heavy labeling of the 105-110 band (Fig. 4), which was of the approximate size anticipated for a Na ϩ -Ca 2ϩ exchanger. Although only material in the 105-110-kDa region was detected on the blots with Coomassie Blue, Western blots showed in addition traces of a ϳ60-kDa peptide that bound the antibody. This ϳ60-kDa membrane-bound fragment may have represented the segment of the scallop muscle Na ϩ -Ca 2ϩ exchanger C-terminal to cleavage site T 1 , formed by

NDPV-SKVFFEPGTYQCLEN 418
RatNCX1 (9) DDDGASRIFFEPSLYHCLEN 408 RatNCX2 (7) PEDFASKVFFDPCSYQCLEN 410 RatNCX3 (8) proteolytic damage during the preparation and DOC extraction of the membranes. This polypeptide would have contained the stretch of peptide chain containing the epitopes of the 37-kDa tryptic fragment. (The predicted size for such a fragment, assuming a mean residue mass of 110 Da and a length of ϳ1,000 residues for the scallop exchanger, would be 62-63 kDa.) Thus, sequence comparisons indicated that the soluble tryptic fragments were not derived from the Ca-ATPase but from a Na ϩ -Ca 2ϩ exchanger, and Western blots suggested that the precursor of the soluble peptides was in the expected size range for a Na ϩ -Ca 2ϩ exchanger. The sarcoplasmic reticulum is an internal cytoplasmic cellular membrane and therefore cannot be associated with any secondary active transport process dependent on a Na ϩ electrochemical gradient. Although the Ca-ATPase must account for most of the material in the 105-115-kDa band on SDS gels of the scallop B 2 membrane fraction (25)(26)(27)(28)(29)(30), a second, non-Ca-ATPase protein, a Na ϩ -Ca 2ϩ exchanger, is also present. This enzyme not only has a similar size to the Ca-ATPase but also in addition coincidentally happens to resist extraction by low concentrations of DOC, like the Ca-ATPase. In fact, there is evidence that extraction of the membrane with low concentrations of DOC does not remove all of the non-SR material from the preparation. Thus, traces of a ϳ28-kDa protein associated with the SL remain in the DOCextracted B 2 membrane fraction ( Fig. 1 and Refs. 25 and 29). Again, the near UV absorption spectrum (not shown) of the DOC-extracted B 2 preparation dissolved in SDS shows that the material contains a trienoic chromophore known to be located in the SL (29). Therefore, some residual structures derived from elements of the SL that contaminate the fragmented SR-enriched B 2 fraction (29) may survive the DOC extraction step and provide a plausible location for the Na ϩ -Ca 2ϩ exchanger found in this work.
The Precursor of the Soluble Tryptic Fragments Is Also Present in a Membrane Fraction Enriched in Sarcolemma-Because there was a possibility that the precursor of soluble tryptic fragments was located in the SL, the B 1 scallop muscle membrane fraction enriched in SL was prepared as described under "Experimental Procedures." Digestion of the native (non-DOC extracted) B 1 membranes with trypsin yielded only traces of the soluble peptides (Fig. 5, lanes 2 and 3). However, when the B 1 fraction was first extracted with low concentrations of DOC, according to the protocol used with the B 2 fraction, and then treated with trypsin, the soluble fragments were formed in substantially larger amounts (Fig. 5, lanes 4 and 5). As with digestion of the DOC-extracted B 2 fraction, the 37-kDa fragment was the major soluble product in the presence of Ca 2ϩ , whereas the 16-and 19-kDa peptides were the predominant species released by digestion in the presence of EGTA.
The increased yield of the soluble tryptic fragments after treatment of the B 1 membrane fraction with DOC was consistent with those peptides originating in the large cytoplasmic domain (f loop) of a Na ϩ -Ca 2ϩ exchanger, if most of the SL vesicles in the initial preparation had a right-side-out orienta- The deoxycholate-extracted B 2 scallop muscle membrane fraction was digested with trypsin, the soluble fraction was concentrated by the DOC-trichloroacetic acid method, and the sample was electrophoresed in the Tricine SDS system. Western blotting using rabbit anti-37-kDa fragment IgG as probe was carried out as described under "Experimental Procedures." Both the 16-and 19-kDa fragments shared epitopes with the 37-kDa peptide.

FIG. 4. Western blot of deoxycholate-extracted scallop muscle B 2 membrane fraction probed with rabbit anti-scallop 37-kDa fragment antibody.
Undigested deoxycholate-extracted B 2 membrane fraction was electrophoresed on a Tricine SDS gel. After electrophoretic transfer onto polyvinylidene difluoride, the preparation was probed with rabbit anti-37-kDa peptide antibody as described under "Experimental Procedures." Lane A, Western blot developed by the luminol method. A strongly labeled zone was present in the 105-110-kDa region. The antibody also detected traces of membrane-bound breakdown products, particularly one species of ϳ60 kDa, possibly representing a C-terminal fragment produced by proteolysis during preparation. Lane B, the blot stained with Coomassie Blue. The dye detected polypeptide only in the ϳ110-kDa region. tion. Only after partial disruption of the SL membrane by the low (nonsolubilizing) concentrations of detergent used in the extraction would the loop, located on the internal surface of such vesicles, become accessible to the protease.
The Precursor of the Soluble Tryptic Fragments Is in a Membrane Domain Insoluble in Nonionic Detergent-The scallop SL is rich in cholesterol, whereas the SR has a low cholesterol content (29). Because cholesterol-rich domains of membranes are resistant to cold nonionic detergents (37), the DOC-treated B 2 fraction was extracted with the nonionic detergent C 12 E 9 to determine whether the precursor of the soluble fragments remained with the insoluble residue. (This procedure is routinely used to prepare solubilized Ca-ATPase (25,38).) The material resistant to C 12 E 9 was washed free of detergent and solubilized protein and then digested with trypsin in the presence and absence of Ca 2ϩ in the usual way. This treatment yielded the soluble tryptic fragments, with larger amounts of the 37-kDa fragment being formed in the presence of 1 mM Ca 2ϩ , whereas more of the 16-and 19-kDa peptides were produced in 10 mM EGTA (Fig. 6). This result was consistent with the precursor of the soluble fragments being an integral membrane protein in a cholesterol-rich membrane, such as the SL, contaminating the SR-rich B 2 fraction. (As discussed below, it is difficult to completely separate the scallop muscle SR and SL membrane fractions.) Treatment of Muscle Membranes with cAMP-dependent Kinase Leads to Phosphorylation of the Soluble Tryptic Fragments-Because there have been suggestions that the large cytoplasmic loop of the Na ϩ -Ca 2ϩ exchanger might be a substrate for protein kinase A (cAMP-dependent protein kinase (2)(3)(4)), the DOC-extracted B 2 scallop muscle membrane fraction was treated with the catalytic subunit of PKA and [␥-32 P]ATP (Experimental Procedures). After electrophoresis of a sample of the undigested phosphorylated membranes in the Tricine SDS system, autoradiography of the gel was carried out. A heavily labeled band was present in the 105-110-kDa region of the gel, and phosphorylation of a polypeptide of ϳ60 kDa was also seen (Fig. 7). The Ca-ATPase of the scallop SR lacks a PKA consensus phosphorylation sequence (R(R/K)X(S/ T)) (36). As with the rabbit SERCA1a enzyme, self-phosphorylation of the Ca-ATPase using ATP has an absolute requirement for Ca 2ϩ (27,28) and so cannot occur under the conditions of the PKA treatment, which is carried out in the presence of excess EGTA. Thus, the phosphorylated peptides were likely to correspond to intact exchanger and the 60 -65-kDa membranebound C-terminal proteolytic fragment detected by Western blotting (see above).
The phosphorylated membranes were then digested with trypsin separately in 1 mM Ca 2ϩ and 10 mM EGTA. Autoradiography of SDS gels of the soluble peptides in the supernatants showed that the ϳ37and ϳ16-kDa fragments were heavily phosphorylated, but the ϳ19-kDa peptide was unlabeled (Fig. 8). DISCUSSION Precise regulation of intracellular Ca 2ϩ levels is critical for the appropriate activation/inactivation of some key cell functions, such as motility and exocytosis. Recently, intracellular Ca 2ϩ has been revealed as a triggering factor for apoptosis (39). Thus, regulation of the channels and transporters determining cell Ca 2ϩ levels is of physiological and pharmacological interest. Sodium-calcium exchange represents one of the most important and widespread mechanisms available to the cell for controlling intracellular Ca 2ϩ concentration. The large cytoplasmic loop of the exchanger is accessible to internal signals reflecting cell status and so has evolved as a regulatory domain mediating the effect of changes in cell Ca 2ϩ concentration on the direction of net Ca 2ϩ movement across the plasma membrane. It is also apparent that external signals may affect the activity of the Na ϩ -Ca 2ϩ exchanger (40). The work reported here suggests that both types of signal may be transduced through the same region of the f loop.
Only material of 105-110 kDa was present in sufficient amounts in the initial DOC-extracted scallop muscle mem-FIG. 6. Tricine SDS gel of the soluble tryptic cleavage products from the C 12 E 9 -insoluble residue of scallop muscle membranes. After treatment of the DOC-extracted B 2 membrane fraction with 4% (w/w) C 12 E 9 on ice, as described previously (25), the preparation was centrifuged at 10 5 ϫ g for 1 ⁄2 h at 4°C, the supernatant was removed, and the pellet was resuspended and washed twice in 20% (v/v) ethylene glycol, 0.4 M KCl, 1 mM CaCl 2 , 50 mM MOPS-Na, pH 7.0. The insoluble residue was resuspended in the same medium and digested with trypsin as described under "Experimental Procedures." After concentration of the soluble tryptic products by the DOC-trichloroacetic acid method, these were electrophoresed on a Tricine SDS gel. Lane 1, soluble products from digestion of C 12 E 9 -insoluble pellet (240 g of total protein) in the presence of 1 mM Ca 2ϩ . Lane 2, soluble products from digestion of C 12 E 9 -insoluble pellet (240 g of total protein) in the presence of 10 mM EGTA. Lane 3, markers. Autoradiogram of a Tricine SDS gel of undigested scallop DOC-extracted B 2 membrane fraction treated with PKA and [␥-32 P]ATP as described under "Experimental Procedures." The 105-115-kDa region was strongly labeled with 32 P. A peptide of ϳ60 kDa, probably corresponding to the membrane-bound breakdown product detected by Western blotting, was also labeled with 32 P. brane preparation to act as a precursor for the soluble tryptic fragments. The origin of the soluble fragments in a ϳ110-kDa species was confirmed by Western blots using rabbit anti-37-kDa peptide IgG. Coincidentally, the precursor is not only close in size to the scallop SR Ca-ATPase, but in addition, like the Ca-ATPase, it resists extraction by low concentrations of DOC; however, unlike the Ca-ATPase it is not solubilized by cold nonionic detergent. The N-terminal sequence of the 37-and 16-kDa soluble fragments shows substantial similarity to a segment of the putative f loop of the well described NCX1 exchanger that is associated with regulation of the enzyme activity by Ca 2ϩ . This region in the predicted structure for NCX1 has a location compatible with the presence of a site susceptible to proteolytic enzymes and the production of watersoluble fragments. Because the precursor is also an integral membrane protein with an appropriate molecular mass for a Na ϩ -Ca 2ϩ exchanger, it is likely to represent an enzyme of that class located in scallop muscle membranes. Such transporters are ubiquitous in animal cells (41), and both plasma membrane (sarcolemmal) and mitochondrial Na ϩ -Ca 2ϩ exchangers must be present in scallop muscle. The possible overall relationships between the tryptic fragments and the putative parent molluscan Na ϩ -Ca 2ϩ exchanger are shown in Fig. 9. Gel shift experiments with fusion proteins containing the regulatory Ca 2ϩ -binding site of NCX1 have indicated that a large conformational change occurs in the ␤-1-␤-2 region when Ca 2ϩ becomes bound (18,19). The difference in accessibility to trypsin of the region represented by the T 2 cleavage site on the scallop exchanger between the Ca 2ϩ -bound and Ca 2ϩ -free states may be a reflection of such a refolding of the large cytoplasmic domain of the exchanger on binding and release of the ligand.
The scallop muscle Na ϩ -Ca 2ϩ exchanger reported here is 105-110 kDa in size, close to that of the mitochondrial exchanger (10), but smaller than the ϳ120 kDa of the glycosylated NCX1 plasma membrane exchanger. This observation might suggest that the precursor of the soluble tryptic fragments reported here is the scallop mitochondrial exchanger; however, oligomycin-sensitive Ca-ATPase and succinic dehydrogenase activities present in the native scallop membrane preparations (42) are removed by DOC extraction, and there is no indication from SDS gels of significant contamination of the DOC-extracted scallop membrane preparation by the F 0 F 1 -ATPase ( Fig. 1 and Ref. 25). Even in preparations of purified beef heart mitochondria, the Na ϩ -Ca 2ϩ exchanger is present in only small amounts (0.4 g/mg total protein (24)). There was direct evidence consistent with a sarcolemmal location for the precursor of the soluble tryptic fragments. Digestion of the DOC-extracted B 1 membrane fraction (SLenriched) with trypsin yielded soluble peptides, with more 37-kDa fragment being released in the presence of Ca 2ϩ and more 19-and 16-kDa material in the presence of EGTA. The fact that treatment of the membranes with deoxycholate before digestion with trypsin greatly increased the amount of soluble fragments released supports an origin for the soluble tryptic fragments in the Ca 2ϩ -regulatory binding site region of the large cytoplasmic domain (f loop). Presumably, treatment with low concentrations of DOC renders the SL permeable to trypsin and allows the protease access to the cytoplasmic loop of the exchanger on the interior aspect of the membrane. If most of the native vesicles derived from the SL in the preparation were right-side-out, a large increase in the extent of proteolysis of the cytoplasmic domain of the enzyme would be expected after DOC treatment. In fact, ϳ70% of the ouabain-inhibitable Na ϩ ,K ϩ -ATPase activity of the scallop muscle B 1 membrane fraction has to be unmasked by treatment with 0.2% (w/v) saponin. 3 Because the large cytoplasmic domain of the Na ϩ ,K ϩ -ATPase, which binds the substrate ATP, is located on the inner aspect of the plasma membrane, this finding is consistent with the majority of the native membrane vesicles derived from the scallop SL being right-side-out. The fact that the insoluble residue left after extraction of the membranes with cold nonionic detergent is a good source of the tryptic fragments fits a location for this scallop Na ϩ -Ca 2ϩ exchanger in some perhaps specialized domain of a cholesterol-rich membrane, such as the SL.
The presence of the precursor of the soluble tryptic fragments in both the B 1 (SL-enriched) and B 2 (SR-enriched) membrane fractions may be due to the difficulty in completely resolving the scallop SL and SR membrane fractions (29). This problem arises, at least in part, because in the scallop crossstriated adductor muscle, which has no transverse tubules, SR cisternae analogous to the terminal cisternae of vertebrate skeletal muscle lie directly beneath the SL (43). The two types of membrane are connected morphologically by foot-type structures resembling those associated with the ryanodine receptor complex seen in the triads of vertebrate muscle. Ryanodinetype Ca 2ϩ release channels have been observed in preparations of scallop muscle membranes (44). Possibly, the patches of SL overlying the junctions fulfill some of the functions of the transverse tubules of vertebrate cross-striated muscle. Disruption of the scallop muscle on homogenization may leave pieces of these domains of the SL attached to fragments of the SR cisternae. The resulting structures may collect in the B 1 or B 2 fraction, depending on the relative proportions of lipid-rich SL and protein-rich SR that remain associated. Thus, the scallop Na ϩ -Ca 2ϩ exchanger may be localized in SL domains associated with underlying cisternae of the SR. The presence of the scallop muscle exchanger in membranes resistant to cold nonionic detergent and its possible relation to underlying elements of the SR are consistent with previous studies of amphibian smooth muscle (45). That work demonstrated functional and structural linkages among the Na ϩ -Ca 2ϩ exchanger, Na ϩ -K ϩ pump, and the Ca 2ϩ pump of the SR and showed that the exchanger was concentrated in caveolae of smooth muscle sarcolemma that were in close proximity to the SR. Caveolae are rich in cholesterol and insoluble in cold nonionic detergent (46).
ATP increases the affinity of the transport sites for Ca 2ϩ and Na ϩ on the cardiac NCX1 exchanger and the squid giant axon (47). There is evidence that phosphoryl transfer is important for the action of ATP in the squid axon and that a Ca 2ϩ -dependent protein kinase may be involved in the effect (48). However, in the case of cardiac muscle, protein kinases do not appear to be implicated (49), and a mechanism involving PIP 2 interaction with the XIP region may be present (50,51). Nevertheless, a potential phosphorylation site for cAMP-dependent PKA and calcium calmodulin-stimulated protein kinase is present in the f loop of NCX1, at Ser 389(357) in the R 385(352) KAVS sequence, just N-terminal to the Ca 2ϩ regulatory region (2). A possible substrate site for these kinases is present at Thr 113 in NCX3, on the putative intracellular loop between transmembrane helices 1 and 2 (4). Possible sites for phosphorylation by PKA and protein kinase C exist at Thr 262 and Thr 267 in the XIP regions of NCX2 and NCX3 (3,5). All three vertebrate exchangers possess a potential site for tyrosine kinase phosphorylation in the ␤-2 repeat, and there may be mixed control of the exchangers through Ca 2ϩ and phosphorylation (4). The site(s) on the scallop Na ϩ -Ca 2ϩ exchanger phosphorylated by the catalytic subunit of PKA and detected on the 16-and 37-kDa soluble tryptic fragments must, of course, be C-terminal to the residue on the scallop Na ϩ -Ca 2ϩ exchanger equivalent to residue 401 of NCX1. The 16-kDa peptide, which is derived from the ␤-1-␤-2 stretch of exchanger polypeptide, spans that part of the f loop containing the Ca 2ϩ regulatory domain. A PKA substrate site must therefore exist in the Ca 2ϩ regulatory domain of the scallop Na ϩ -Ca 2ϩ exchanger, and thus there is a possibility that the Ca 2ϩ regulatory domain in the large cytoplasmic loop can be directly affected by agonists whose effects are mediated by the level of cellular cAMP. Western blotting indicated common epitopes between the 19-and 37-kDa soluble tryptic peptides. Thus, the complete absence of any phosphorylation of the 19-kDa soluble peptide by PKA and its size suggest that it represents the segment of the Na ϩ -Ca 2ϩ exchanger lying between cleavage sites T 2 and T 3 , i.e. the Cterminal segment of the 37-kDa peptide.
In cardiac muscle, Na ϩ -Ca 2ϩ exchange makes an extremely important contribution to the relaxation phase of the contractile cycle (1,40,47). In addition to its role in relaxation, when removal of Ca 2ϩ from the sarcoplasm by the exchanger is combined with Ca 2ϩ uptake into the SR, it has been suggested that the Na ϩ -Ca 2ϩ exchanger may also be involved in the initiation of contraction (52). This would be by supplementing the Ca 2ϩ influx into cardiac muscle through L-type Ca 2ϩ channels that is associated with Ca 2ϩ -induced Ca 2ϩ release from the cardiac SR. The role of the scallop exchanger in the physiology of the scallop muscle cell remains to be determined. Because scallop muscle Ca-ATPase lacks both a phospholamban binding site and a PKA phosphorylation consensus sequence (36), regulation of cell Ca 2ϩ by cAMP through effects on the pumping rate of the SERCA enzyme appears to be excluded. However, cAMP may affect Ca 2ϩ levels in the scallop muscle cell through PKA phosphorylation of the sarcolemmal Na ϩ -Ca 2ϩ exchanger. The fact that the exchanger copurifies so closely with the SR Ca-ATPase suggests a relationship between the two membrane Ca 2ϩ transport systems. In future studies, structural changes in the protein associated with Ca 2ϩ binding and phosphorylation will be examined, and the nature of the possibly specialized cellular structure containing the exchanger will be investigated in detail.