(1, 2). Of the known cytosolic calcium-binding proteins, cal- modulin has received the most attention (for reviews, see. Refs. 3-7). Its broad phylogenetic distribution and regulatory ... bution of calmodulin (and other heat-stable calcium-binding.
THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1985 hy The American Society of Biological Chemists, Inc.
VOl. 260, No 3, Issue of February 10, pp. 1652-1660,1985 Printed in U.S. A.
Isolation and Characterizationof CAB-63, a Novel Calcium-binding Protein* (Received for publication, June 14, 1984)
David M. WaismanSj, Bharathi P.SalimathS, andM. John Andersonn With the technical assistance of Vilasini Sarangapani‘and Judit Muranyi From the #Departmentof Medical Biochemistry and the TDepartment of Pharmacology and Therapeutics, The University of Calgary, Calgary, Alberta,Canada TZN 4Nl
assay (10)) and calmodulin activity (assayed by phosphodiAnovelcalcium-bindingproteinnamedCAB-63 (formerly called calregulin) has been purifiedfrom esterase activation (11)).We concluded that calmodulin was bovine liver 100,000 X g supernatant. The purified responsible for only a small portion of the total calciumprotein has been characterized with respect physto its binding activity of the 100,000 x g supernatant. It was, ical, chemical, and calcium-binding properties. It has therefore, important to establish whether or not these calan apparent molecular weight of 63,000 by sodium cium-binding activity peaks were comprised of unknown dodecyl sulfate-polyacrylamide gel electrophoresis and (novel) or known proteins. Recently, we have purified and 55,000 by sedimentation equilibrium centrifugation characterized a protein responsible for one of the major calunder nondenaturing conditions. It is an asymmetric cium-binding activity peaks of bovine brain (12) and shown molecule with a friction@ coefficient of 1.69 and a this protein to be a novel calcium-binding protein. Stokes radium of 44.2 A. Amino acid analysis has Preliminary reports (13, 14) have detailed the chromatogrevealed 34.0% acidic residues, 14.0% basic residues, raphy of the 100,000 X g supernatant of bovine liver on DEAEand 4.0% tryptophan. The acidic nature of the molecule is further confirmed byits isoelectric pointof 4.65. In cellulose and the partial isolation of a single protein responthe presence of 3 mM MgC12 and 150 mM KCl, CAB-63 sible for the major calcium-binding activity peak. This protein binds 3.0 mol of calcium/mol of protein with an appar-was previously referred to as calregulin (13) and based on physical, chemical, and calcium-binding properties was sugent & = 0.1 HM. gested to be a novel protein (14). Immunoblotting and Ouchterlonydouble-diffusion The procedure for the purification of CAB-63 (14) has now procedures have identified CAB-63 in a variety of bovine tissues. Immunocytochemical staining of both been improved, and data arepresented on the physical, chemical, and calcium-binding properties of this novel protein. fibroblasts and cryotome-sectioned bovine liver furis re- Immunoblotting and Ouchterlony techniques have demonther indicates that CAB-63 immunoreactivity stricted to an elaborate system of perinuclear membra- stratedthe existence of CAB-63 immunoreactivity in the nous vacuolesand cisternae indistinguishable from im- 100,000 X g supernatants from a variety of bovine tissues. munocytochemical staining of the endoplasmic reticu- Immunofluorescent studies further indicate that CAB-63 is lum. It is concluded that CAB-63 represents a major localized primarily in a system of cytoplasmic organelles that calcium-binding protein whose subcellular organizaare virtually indistinguishable from immunocytochemically of the stained endoplasmic reticulum. tion suggests a possible role in the function endoplasmic reticulum. EXPERIMENTALPROCEDURES
The intracellular calcium-binding proteins play an important role in the transduction of hormonally induced changes in cytoplasmic free Ca2+concentration intocellular activation (1, 2). Of the known cytosolic calcium-binding proteins, calmodulin has received the most attention (for reviews, see Refs. 3-7). Its broad phylogenetic distribution and regulatory activities have led many investigators to the suggestion that calmodulin maybe the primary intracellular receptor for calcium (8). Previously (9) we have chromatographed the 100,000 X g supernatant of bovine brain and hearton DEAE-cellulose and analyzed the resulting fractions forboth calcium-binding activity (using the Chelex 100 competitive calcium-binding
* This work was supported by grants from the Medical Research Council of Canada and the Canadian Heart Foundation (D. M. W.) and the Alberta Heritage Foundation for Medical Research (M. J. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. § To whom all correspondence should be addressed.
Materials All chemicals were reagent grade, unless specified otherwise. Deionized water was used throughout. Chelex 100 was obtained from BioRad. “CaC12 was purchased from Amersham Corp. Proteolytic inhibitors were obtained from the following sources: leupeptin, pepstatin A, chymostatin, and antipain were obtained from Transformation Research (Framingham, MA), and diisopropyl fluorophosphate, benzamidine, and phenylmethylsulfonyl fluoride were obtained from Sigma; DEAE-cellulose (DE52) was obtained from Whatman. Hydroxyapatite was obtained from Calbiochem-Behring. Horseradish peroxidase-conjugated goat anti-rabbit IgG was obtained from BioRad. Methods’ Portions of this paper (including part of “Experimental Procedures” and Figs. 1 and 2) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of astandard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 84M-1807, cite the authors, and include a check or money order for$4.40 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.
1652
Calcium-binding Protein,Novel
CAR-63
1653
RESULTS
Heat Lability of the Calcium-binding Activity of the Bovine Liver 100,000X g Supernatant-Calmodulin has been viewed for some time as the majorcytosoliccalciumreceptor (8). Because its activity is retained after incubation at 98 "Cfor 2 min ( l l ) , thermal stability could be used to assess the contribution of calmodulin (and other heat-stable calcium-binding molecules) to the total calcium-binding activity of bovine liver 100,000 x g supernatant. Accordingly, the Chelex 100 competitive calcium-binding assay (10)was used to analyze both the 10,000 x g and 100,000 x g supernatants for calciumbinding activity before and after incubation at 98 "C for 2 min. The results of this experiment (Table I) indicate that the heat-stable calcium-binding components compriseonly about 30% of the total calcium-binding activity. As expected, 90-95% of the calmodulin activity, determined by the phosphodiesterase assay (1 I),was recovered after heat treatment. Since the majority of calcium-binding activity in the 100,000 X g supernatant of bovine liver could not be attributed to calmodulin, the existence of other calcium-binding proteins was suspected. The strategy of our subsequent experiments was thus to purify and characterize these substancesin order to determineif they were known or unknown (novel) calciumbinding proteins. Chromatography of Liver 100,000 x g Supernatant on DEAE-cellulose-The unheated 100,000 X g supernatant was fractionated on DEAE-cellulose. Four peaks of calcium-binding activity could be resolved (Fig. 1). Calmodulin activity, determined by the phosphodiesterase assay, resolved as two peaks. The minor peak did not show corresponding calciumbinding activity whereas the major peak did (Fig. 1, peak I V ) . The calcium-binding activities of peaks I, 11, and 111 were destroyed (80-9096)by heat treatment while the calciumbinding activity of peak IV remained largely unaffected by this procedure (data not shown). Therefore, as illustratedin Table I, the major activity of the 100,000 x g supernatant is heat labile and thus may be due to the presence of novel calcium-binding proteins. In fact, we have isolated and characterized a single protein which can account for the major peak of calcium-binding activity in this extract (Fig. 1, peak I I ) . This novel calcium-binding protein hasbeen named CAR63 (13). Isolation of CAR-6,3-As detailed under "Methods," purification of C A E " involved several chromatographic steps(see above) including ion exchange on DEAE-cellulose, adsorption on hydroxyapatite, and gel filtration on TSK 3000 (LKR). Approximately 80-100 mg of CAB-63 were purified from 1 kg of bovine liver. Radioimmunoassay of the original 100,000 X g supernatant has indicated that bovine liver contains about TARIXI Hrat stability of calcium-binding activityof bovine liver homogenate suprrnatant Homogenate supernatantswere incubated at 98"C for 2 min. ___ _ _ ~ ~ Binding capacity Supernatant I'reparat ion remaininf "
I I1 Ill IV V VI
X R
%
10,000 10,000 100.000 100,000 100.000 100.000
18 33 27 36 32 28
" A s measured hy the Chelex assay, % heat stahility = 100 X ((% totalcounts,calcium-bindingactivityafter incubation)/(% total counts, calcium-hinding activity hefore incubation)).
116.0 K 97.4 K 66.2 K
45.0 K
36.0 K 29.0 K 24.0 K 20.1 K 14.2 K -dye
f
e
d
b
C
a
FIG. 3. SDS-polyacrylamide electrophoretic analysis at different stagesofCAR-63 purification. Aliquotsof pnnled fractions ohtained during CAH-63 purification were subjected t o 5-15"; SIISPAGE. a . molecular weight standards (see Fig. 2 legend): h. I ~ H ) . ( H H ) X q , supernatant; c, DEAE-cellulosepeak 11: d. hydroxyapatite; r . Sephadex G-150;f, T S K 3000.
300 mg of CAR-S3/kg of liver.' Therefore, 30% of CAR-63 was recovered during purification. The data from the final step of purification. gel filtration on TSK 3000, is presented inFig. 2. A close correlation between protein concentration and calcium-binding activity is observed. SDS-PAGE:' of the purified protein reveals the presence of two bands (Fig. 2, inwt).the major of which had a M , of 63,000, and the other a M , of 61,000. In previous studies proteolytic fragmentation of each band was shown to result in peptide maps that were indistinguishable (141, suggesting that the band doublet was not due to the presence of a contaminant but was instead due to the presence of two closely related proteins. One obvious possibility is that these two forms simply result from proteolysis during the isolation procedure. However, analysis of the CAR-63-containingfractions by SDS-PAGE at thevarious steps of purification (Fig. 3) suggests that the minor lower-molecular-weight band mav not simply have been generated during purification. A close relationship between these two proteins is also suggested by the fact that analogous immunoreactive bands were alsoobserved in immunoblots of the original liver 100,000 X g supernatant (Fig. 41, indicating that both of the proteins present in purified calregulin were also present in the original tissue extract. Thus,in order for the minor band to have been an artifactof proteolysis, degradation must have occurred during or before the preparations of the 100.000 x g supernatant. ~~
~
~
* N. C. Khana and I). M.Waisman. manuscript
in preparation, ' T h e ahhreviationn used are: SDS.scdium dtdecvl sulfate: PAGE, polyacrylamide gel electrophoresis: [)'IT. dithiothreitol; MOPS. 3 (N-morpholino)propanesulfonicacid; THS, Trin-bufferedsaline; EGTA. ethylene glycol his(d-aminoethvl ether)-iV.N.,~".N'-tetraacetic acid.
Calcium-binding Protein,Novel
1654
CAR-63
Quantitation of the Two Forms of CAR-63"The doublet band of CAR-63 can be visualized on SDS-PAGE (Fig. 2), but the resolution of these bands is inadequate for quantitative determinations. On isoelectric polyacrylamide gels (16) a douI d v e' front ble-band pattern is also poorly resolved (data not shown). In contrast, t.he separation of the two forms of CAB-63 on nondenaturing polyacrylamide gels was notablyenhanced. Densit.ometric analysis (Fig. 5 R ) of Coomassie Blue-stained bands reveals that themajor component comprises about 83% of the total protein. Calcium-binding Actiuit-y of the Major and Minor Randsof CAB-63"Purified CAR-63 was analyzed by electrophoresis on nondenaturingpolyacrylamide gels after which the gel was sliced and assayed for calcium-binding activity and protein concentrated as out.lined under "Methods." Results of such 0.77 0.76 0.75 0.74 anexperimentarepresented in Fig. 6. Only one peak of protein could be distinguished,andthiscorrelated closely RELATIVEMOBILITY with a corresponding peak of calcium-binding activity. This Fl(;. 5 . Quantitationofthe t w o electrophoreticforms of suggesh that at least of one the electrophoretic forms of CARrln 63 must bind calcium. The gel-slicing technique is limited in CAB-63. Twenty p g o f purilied ('AH-tXI werc rI~~rtrc~~)hc,rc.st.d i.5'.; polvarrylamide gels in thr absrnre o f SI)S (11). The protrin that a proteinpeak corresponding to the minor band of CAR- concentration o f the stained gel was rstimatrd by drnsitcmwtrg ( I { ) . 63 was not resolved. Physical and Chemical Properties of CAR-63"Table I1 sum-
I '!
1
RELATIVE MOBILITY
FIG. 6 . Calcium-binding activity of the two forms of CAR6 3 . Sixty pg of purified ('AH-63 were ekctrophoresed on 7.r1'; polyacrylamide gels in the absence of SDS.2.0-mm slices o f gel were 0.2 M Nd'l. homogenized in 4 0 mM Tris-HCl (pH 7.51,0.2 mM .'T-l)1 and the calcium-hinding activity ). .".( and protein concentratinn (W) of each gel slice were determined.
TAR1.E
11
Physical proprrties -~ ~-
Sedimentation constant (sm,. ) Stokes radius ( A ) Frictional cnefficient C// , ) Partial volumespecific (V) Absorbance (A:?&) Isoelectric (IpH) Molecular weight From sedimentation and gel filtration From sedimentation equilibrium From SDS-gel electrophoresis
point 1
B.
FIG. 4. Identification of CAB-63 in bovine liver 100,000 X g supernatant. Samples were electrophoresed on 5-15"; polyarrvlamide gels in the presence of 0.1% SDS as described tly Laemmli (15). Gels were then stained (data not shown) with Coomassie Blue and destained or suhjected to immunohlotting as described under "Met hods." T h e immunohlot of gel B was treated with purified control serum ( 5 0 pg/ml) while that of A was treated with purified anti-CAB63 antillodv ( 5 0 pglml). Numbers refer to lanes which contain the following fractions: lanes 3 and 4, purified CAB-63 (5 p a ) ; lanes 2 and -5, liver 100.000 x p supernatant (I00 p g ) : and lanes I and 6. liver 100,000 X p supernatant (200 p g ) .
o/ ('AH-69
~~~~
:l.I): 44.2 1. 6 9
0,718 21 .o 4.65 5.5,WM) "
:):),~MM)
63.000
~
marizes some of the physical properties of purified CAR-63. The molecular weight of the protein determined by the method of Yphantis (18)under nondenaturing conditionswas 55,000. CAR-63 is, therefore, a monomer. The frictional ratio (f/f,)= 1.69) suggests the molecule has an asvmmetric shape. This is also consistent with the relativelv large Stokes radius
1655
Novel Calcium-binding Protein, CAB-63 TABLE 111 Amino acid composition of CAB-63 Values after hydrolysis for 24 h 48 h 72 h
Amino acid
A~~~~~~of extrapolated
residues/molb
92.2 88.2 90.4 Aspartic acid 20.1 20.4 19.9 Threonine Serine 19.8 19.6 16.9 98.4 95.0 96.5 Glutamic acid 37.9 34.6 Proline Glycine 37.3 37.1 37.0 26.0 24.7 24.2 Alanine 3.33.43.6 Half-cystine Valine 14.6 20.3 5.5 Methionine 5.7 Isoleucine 18.2 21.2 Leucine 23.1 23.5 Tyrosine 19.3 19.6 Phenylalanine 24.9 24.8 Lysine 51.6 52.3 Histidine 9.8 10.7 Arginine 12.5 12.7 Tryptophan 23.3 23.8 y-Carboxyglutamic acid -
33.0
21.8 4.45 22.7 23.7 19.2 24.1 52.5 10.9 12.8 23.8
90.3 20.0' 20.0' 96.9 35.2 37.1 25.0 3.4d 23.0' 5.22 24.0 24.0 19.4 24.6 52.1 10.5 12.7 24.0' >0.18
90 20 20 97 35 37 25 3 23 5 24 24 19 25 52 11 13 24.0 0
Average of values after 24-, 48-, and 72-h hydrolysis. chainassuming amonomericmolecular weight of 63,000 derived from SDS gels. Extrapolated to zero time hydrolysis. Determined as cysteic acid according to Ref. 20. From 72-h hydrolysis only. 'Determined after hydrolysis by methanesulfonic acid according to Ref. 21. Determined after alkaline hydrolysis according to Ref. 22.
* Residues/polypeptide
I
250
260
270
280
Calcium-binding Properties of CAB-63"Fig. 8 presents a saturation curve for calcium binding to CAB-63. Scatchard (28) analysis of the data(Fig. 8, inset and legend) reveals that in the presence of 3.0 mMM&12 and 150 mM KC1 CAB-63 binds 3.0 mol of Cazc/mol of protein with an apparent Kd = 0.1 pM. Calcium-binding studies were also performed in theabsence of MgC12 (data not shown), but reproducible differences between the binding of Caz+ to CAB-63 in the presence or absence of M&lZ could not be demonstrated. This result indicates that CAB-63 possesses both high affinity and selectivity for Ca2+.In contrast, calmodulin has been shown to bind 4.0 mol of Ca2+/mo1of protein with apparent K d = 6.7 ~ L Min the presence of millimolar M e and 4.0 p~ in the absence of Mg2+ (29). Therefore,CAB-63 has a higher affinity and specificity for Ca2+than calmodulin. Distribution of CAB-63 in Bovine Tissues-Immunochemical techniques were used to test for the presence of CAB-63 immunoreactivity in a variety of bovine tissues. Fig. 9, A and B, shows that precipitin bands wqre formed when several tissue extracts were allowed to react with CAB-63 antisera. In contrast the reaction of controlsera with thesetissue extracts consistently failed to produce precipitin bands (Fig. 9, C and D). The confluent pattern produced by the precipitin lines is consistent with the possibility that the CAB-63 present in the different tissue100,000 X g supernatants are closely related immunochemically. The tissue distribution of CAB-63 was also investigated by the immunoblot procedure (26). In Fig. 10, 200 pg of bovine tissue 100,000 x g supernatant and5 pg of purified liver CAB63 were subjected to SDS-PAGE and then immunoblotted using the anti-CAB-63 antibody. The results suggest that the CAB-63 immunoreactivity present in the 100,000 X g supernatants of these bovine tissues is of identical molecular weight to that of purified bovine liver CAB-63. Similar experiments
\I
290
WAVELENGTH (nm)
FIG. 7. Ultraviolet absorption spectrum of CAB-63. Purified CAB-63 (0.28 mg/ml) in 40 mM Tris-HC1 (pH7.5), 0.2 mM DTT, 0.1 M NaC1.0.1 mM EGTA was used. FREE Ca"(M)
and thesmall sedimentation constant. The amino acid composition of CAB-63 is shown in Table 111. The protein contains by mol % about 34% glutamic and aspartic acid and about 14% basic residues (lysine, histidine, and arginine) and, therefore, contains a high proportion of charged residues. The isoelectric point of 4.65 also indicates that the molecule is highly acidic. Interestingly, this protein has a relatively high content of tryptophan (4%) which presumably accounts for both its Ai%,,,,,of21.0 and characteristic UV absorption spectrum (Fig. 7).
FIG. 8. Calcium binding to CAB-63. The experiment was carried out by equilibrium dialysis as described under "Methods." Conditions: 150 mM KCI, 10 mM MOPS (pH 7.1), 3.0 mMMgC12, 1.0 mM DTT, and 0.1 mM EGTA. Inset, Scatchard plot (28) of data. V , mol of Ca2+ bound/mol of CAB-63; C, free [Ca"]. The deviation of the points at the low calcium concentration from the best fit regression line may be due to exaggeration of experimental deviation by this plot. These points also fall into the region where the free calcium concentration is most poorly regulated by the chelator method. Itis, therefore, impossible to rigorously determinewhetherthe curved Scatchard plot is due to experimental difficulties or due to DOSitiVe cooperativity multiple or classes of binding sites.
Novel Calcium-binding P r o t e i n , CAR-6?
1656
L.
U.
Fw. I). Ouchterlonyanalysisofdifferenttissueextracts. The renter t r d l s contained mhht rontrol serum ( ( ' and I ) ) or antiserum t o h v i n r liver CAll-ri:{ ( A a n d /{). 'To the o u t r r trrlls wrrr
added 800 g o f ' h v i n r tissrw 1 0 0 . 0 0 0 X g suprrnntants ns indirnted. I'lntr A : lung ( I ) , smooth musclr ( 2 ) ,pancreas (:{), hrnrt ( 4 ) . hrain (5). liver ( 6 ) . I'latc H : smooth muscle ( I ) , twain (2).livrr (.'I), lung ( 4 ) , kidney ( 5 ) , spleen ( 6 ) . I'lntc. I ) was identical to plntr A and plntr C' was idrnt ical t o plate /j rxrrpt that t h r ccntrr ~ r c l of l plotrs ( ' nnd I ) contninrd r n l h i t control srrum.
FIG. If). Distributionof C A R - 6 3 immunorrnrtivity i n hovine tissue 100.000 X supernatants. p~ r ~ t)rjvlnc* f tiswc. 100.000 X supcrnncnnts o r L' p g o f p t ~ r ~ f i etr~l o v i n c 1ivr.r (':\I!.(;:! wcrr rlectrophoresed o n 5 15'; ~ ) ~ l l v : t c . r ~ l : l m l~f :l fr ~: c d ~ c ~cf*l\ r ~ t 111 r h v prrsrncr nt 0.1'; .C;I)S. l
