Biochimica et Biophysica Acta, 493 (1977) 78-92

© Elsevier/North-Holland Biomedical Press BBA 37702 T H E C R O S S - L I N K I N G OF RABBIT S K E L E T A L MUSCLE SARCOPLASMIC RETICULUM PROTEIN

CHARLES F. LOUIS, MICHAEL J. SAUNDERS and J. ANDREW HOLROYD Department of Biochemistry, University of Leeds, 9 Hyde Terrace, Leeds LS2 9LS (U.K.)

(Received December 24th, 1976)

SUMMARY Sarcoplasmic reticulum proteins have been cross-linked in situ with two reagents, the disulphide-bridged bifunctional imido ester, dimethyl-3,3'-dithiobispropionimidate dihydrochloride and the mild oxidant cupric phenanthroline. Analysis of proteins so cross-linked by electrophoresis on agarose/acrylamide gels reveals that a series of new polypeptides, up to a molecular weight of 900 000, are formed. These have molecular weights which are multiples of 100 000. Further analysis of samples by electrophoresis in a second dimension containing a reducing agent revealed the monomeric polypeptides from which the cross-linked polypeptides were formed. With dimethyl 3,3'-dithiobispropionimidate dihydrochloride homopolymers of the Ca2÷-stimulated ATPase, calsequestrin and/or calcium binding protein were formed. With cupric phenanthroline only the Ca2+-stimulated ATPase was involved in polymer formation. It has been confirmed on another gel system that these two proteins which are involved in Ca 2+ binding are not cross-linked intermolecularly with this latter reagent. We conclude that the 100 000 dalton CaZ+-stimulated ATPase polypeptides are within 2 A_ of each other in the membrane while calsequestrin and/or calcium binding protein are within 11 A of each other. Although there appears to be no limit to the extent of cross-linking of any of these polypeptides there is no indication of heteropolymer associations between them.

INTRODUCTION The arrangement of the protein components in biological membranes has been investigated with a variety of probes, both chemical and enzymatic [1-4]. The distribution of the proteins in the plane of the membrane has been examined with the electron microscope by the freeze fraction method [5] and several labelling techniques such as ferritin conjugates [6]. While these methods give a general indication of macromolecular protein locations they do not reveal sufficient information so that the actual arrangement of the individual polypeptide chains in the membrane may be Abbreviations: SDS, sodium dodecyl sulphate; imido ester, dimethyl-3,3'-dithiobisproprionimidate dihydrochloride; Cu/phenanthroline, CuSO4/1,10-phenanthroline.

79 interpreted. Thus, more sensitive techniques, capable of resolving to the level of polypeptide chains must be used. One such method which has been recently used is that of chemical cross-linking. Although the technique was originally developed by Davies and Stark [7] to interpret the stoichiometry of oligomeric proteins, it has now been used to investigate the interactions occurring in a variety of other biological systems such as ribosomes [8], histones [9], muscle [10] and biological membranes [11 ]. Of the many types of membranes that have now been investigated, that most studied is the erythrocyte. It has been shown that many individual polypeptides in this membrane are readily cross-linked and that in some cases cross-links between different polypeptides can occur [12, 13]. Thus specific associations between the polypeptides of this membrane have been suggested [12, 13]. Because of the complexity of the protein composition of these cells [14, 15] it was of interest to investigate a membrane with a relatively simple protein composition, the sarcoplasmic reticulum isolated from skeletal muscle [16]. This membrane has a single function namely the regulation of the myofibrillar Ca 2+ concentration [17]. The major protein component of this membrane, the Ca2+-stimulated ATPase, which has been purified by a variety of techniques has a molecular weight close to 100 000 [18-20]. Additional proteins which have now been identified in this membrane are a high affinity calcium binding protein [21], with a molecular weight of 55 000, calsequestrin [22, 23] with a molecular weight of 44 000-55 000, a 30 000 dalton protein [24] and a proteolipid with a molecular weight of 12 000 [24]. We have also previously indicated the presence of an additional protein with a molecular weight of 90 000 [25]. Since the functional groups to be cross-linked must be both accessible to the reagent being used and be in the correct geometry for cross-linking, the absence of cross-linking does not necessarily reflect the absence of contact between peptides in the membrane. Thus to obtain information with this technique it is important to use different cross-linking methods. In this paper we report the use of two reagents, dimethyl-3,3'-dithiobispropionimidate and cupric phenanthroline. The former should be able to cross-link membrane polypeptides with 11 A of each other [12], while the latter requires polypeptide-SH groups to be in molecular contact for cross-linking to occur, i.e. within 2 A. EXPERIMENTAL

Materials Both acrylamide and methylene bis-acrylamide were recrystallised from the minimum volume of boiling acetone. N,N,N',N'-Tetramethylethylenediamine was redistilled under reduced pressure. SDS was specially pure from B.D.H. Chemicals Ltd., Poole, Dorset. H232po4 (as orthophosphate in dilute HC1 solution, 10 mCi-ml) was obtained from the Radiochemical Centre, Amersham, Bucks. 3-Chloropropionitrile was obtained from Aldrich Chemical Co. Ltd., Gillingham, Dorset. Agarose (Seakem Brand) was obtained from Marine Colloids Inc., Springfield, N.J., U.S.A. All other chemicals were analytical grade. Water used in all solutions was double distilled.

80

Methods Preparation of rabbit skeletal muscle sarcoplasmic reticulum microsomes. Rabbits were killed by cervical dislocation, the back and leg muscles were excised and muscle microsomes were then prepared by the method of Martonosi [26]. This included extraction with 0.6 M KC1, 5 mM histidine, pH 7.0, to remove any residual actomyosin contamination. The microsomes so isolated were stored in 3 5 ~ (w/v) sucrose at 0 °C. Preparation of the imido ester. 3-Isothioureidopropionitrile was prepared according to the procedure of Perham and Thomas [27]. It was then used, without further purification to prepare 3,3'-dithiobispropionitrile according to Ruoho et al. [28] and had a m.p. 47 °C, literature 48 °C [29], 49-51 °C [30]. This was then converted to its imido ester according to the procedure of Wang and Richards [31]. It was stored under vacuum at 4 °C. The imido ester melted, evolving gas at 125 °C, literature 125-128 °C [32] and had m.p. 174 °C, literature 175-176 °C [32]. It was used without further purification since any contact with moisture resulted in its hydrolysis. Preparation ofdimethyl suberimidate. This was prepared as described by Davies and Stark [7], m.p. 215 °C. Cross-linking of sacroplasmic reticulum proteins. Aliquots of microsome protein (2.5 mg) were suspended in 0.5 ml of 0.2 M triethanolamine. HCI buffer, pH 8.0. The solution of the imido ester was prepared immediately before use in the same buffer, readjusting the pH to 8.0 with 1 M NaOI-I. This was then added to the microsomes, in a volume of 25/zl, and after 15 min at room temperature the reaction was stopped by adding NH4C1 to a final concentration of 0.1 M. After 1 min 0.1 ml of 5~o SDS and iodoacetamide (30 mg) were added. This mixture was incubated for 30 min in the dark and then dialysed for 18 h at 4 °C against 0.1 ~o SDS. In some cases the reduction of the disulphide bonds present in cross-linked microsome proteins was required. This was performed by adding dithiothreitol (5 mg) and 0.1 ml of 5 ~ SDS to the microsomes after cross-linking had been stopped by the addition of NH4C1. After incubation for 2 h at 37 °C, the free sulphydryl groups were blocked with iodoacetamide (50 rag) and incubated for 30 min in the dark. To this was then added 50/A of/%mercaptoethanol and the mixture dialysed for 18 h at 4 °C. All samples were stored in ice at 0 °C and analysed within 24 h. Cross-linking of microsome samples using Cu/phenanthroline solutions was carried out in 50 mM sodium phosphate buffer, pH 8.0. The protein concentrations and solubilisation procedures were as described for imido ester cross-linking except for the substitution of 1 mM sodium ethylene diamine tetraacetate, pH 8.0, for 0.1 M NH4CI to stop the cross-linking reaction. SDS acrylamide gel electrophoresis. This was carried out in a conventional unit (Hoefer Scientific Instruments, San Francisco, Calif., U.S.A.). Analysis of proteins on 8 ~o acrylamide gels was carried out according to the method of Ugel et al. [33]. Gels containing 0.4 ~ agarose and 2.2 ~ acrylamide were prepared as described by Peacock and Dingman [34]. The gels containing 0.1 M sodium phosphate, pH 7.0, were inverted in their glass tubes after polymerisation in order to obtain a flat meniscus and held in the tubes during the electrophoresis by a fine nylon gauze. Dilute acrylamide solution containing agarose could not be successfully overlaid for preparing flat menisci by the conventional method. The buffer was 0.1 M sodium phosphate, pH 7.0, 0.1 ~o (w/v) SDS. Samples, which contained Pyronin Y as tracking dye, con-

81 tained 1% (w/v) SDS to facilitate sample concentration at the top of the gel. This resulted in sharper bands on stained gels compared with samples containing only 0.1% (w/v) SDS. For analysis of proteins in two dimensions the first dimension was that described for 0.5 % agarose/2.2 % acrylamide gels. The second dimension was constructed in a slab gel electrophoresis apparatus model A-C 4-10 (E.C. Apparatus Corp., Philadelphia, Pa., U.S.A.). The method was as follows: a layer of 7 % acrylamide gel containing 11 mM phosphate, pH 7.0, was supported on a 1% agarose plug formed in the base of the gel apparatus. On top of this layer was formed one containing 0.5 % agarose, 1% fl-mercaptoethanol, 0.1% SDS, 11 mM phosphate, pH 7.0, approximately one-tenth the length of the 7 % acrylamide gel. When this layer had polymerized the first dimension agarose/acrylamide gel was rested on top of the reducing layer and covered with a further portion of the agarose mixture containing fl-mercaptoethanol. 5.5 mM phosphate, buffer, pH 7.0, used in the slab gel apparatus contained 0.1% SDS. All gels were stained as described by Fairbanks et al. [15]. All gel samples reported in this paper were 61ectrophoresed a minimum of three times to ensure reproducibility. Gels were photographed with an orange filter and when band intensities were to be calculated the gels were scanned at 540 nm on a Unicam SP 1800 UV spectrophotometer adapted for scanning polyacrylamide gels. Molecular weight calibration curve. Molecular weights of polypeptide species were determined from the calibration curve as shown in Fig. 1. Markers were prepared by cross-linking bovine serum albumin. The migration distances of the chemically cross-linked monomers exhibit a linear relationship between their molecular weight and migration distance on SDS-polyacrylamide gels [35]. Labelling of microsome proteins with Ey-32p]ATP. Microsomes were labelled as described previously [36]. Labelled proteins were analysed on agarose/acrylamide

87654"~ 3-

~2-

2

'9

-//, 0.4

ols

o.'6

0.'7

o'.8

o'.~

~'.o

RF

Fig. 1. Molecular weights of sarcoplasmic reticulum proteins, determined by their migration in agarose/acrylamide gels. Plot of the logarithm of the molecular weights of cross-linked bovine serum albumin polymers versus their relative mobility in agarose/acrylamide gels. Bovine serum albumin was cross-linked with dimethyl suberimidate (10 raM) in 0.2 M triethanolamine, pH 8.0, for 15 min and then reduced and carboxyamidomethylated as described by Louis et al. [36]. A, polymers of bovine serum albumin; 0, polymers of A2 polypeptide as enumerated in Fig. 2.

82 gels a n d 1 m M slices c o u n t e d for C h e r e n k o v r a d i a t i o n in a B e c k m a n L S 230 l i q u i d scintillation counter. Protein determination. P r o t e i n c o n c e n t r a t i o n s were d e t e r m i n e d by the m e t h o d o f L o w r y et al. [37] u s i n g b o v i n e s e r u m a l b u m i n as s t a n d a r d . RESULTS

Analysis of sarcoplasmic retieulum mierosomes on 0.5% agarose/2.2% acrylamide gels Analysis of sarcoplasmic reticulum microsomes on these gels (Fig. 2a) reveals the presence of a major component (Band A2). This protein has previously been identified as the Ca 2+-stimulated ATPase since it forms a stable phosphoprotein intermediate on incubation with [~,-3ZP]ATP [43]. When membranes so labelled were analysed on the agarose/acrylamide gel system reported here (Fig. 3) the major peak

Fig. 2. Cross-linking of sarcoplasmic reticulum proteins with the imido ester. Microsomes were cross-linked with varying concentrations of imido ester and analysed on agarose/acrylamide gels in the presence of SDS as described in Methods. Approx. 50 #g protein loaded on to each gel. a, untreated microsomes; b, 1.7 mM imido ester; c, 3.3 mM imido ester; d, 5.0 mM irnido ester; e, 5.0 mM imido ester, and then reduced with dithiothreitol.

83 of radioactivity coincided with the major polypeptide component A2 (Fig. 2a), confirming that it was the Ca 2+-stimulated ATPase polypeptide. Additionally in Fig. 2a there are two polypeptides with molecular weights less than 100 000. However, the lower of these bands is not always clearly visible on this gel system. When a sample of calsequestrin, as purified by the method of Ikemoto et al. [23], was analysed on this gel system it migrated in the region of Band C. Although no additional discrete band was observed, calcium binding protein would also migrate in this region of the gel since its molecular weight is close to that of calsequestrin [21]. There are in addition four bands of higher molecular weight, Bands 2, 5, 7 and one between 3 and 4; this latter component is somewhat variable in concentration from preparation to preparation (Fig. 2a). cpm A2

4000.

3000.

2000-

( A 2) 2 1000-

Migration

d i s t a n c e (cm)

Fig. 3. Labelling of sarcoplasmic reticulum proteins with [y-32p]ATP. Duplicate samples were analysed on agarose/acrylamide gels, one was sliced into 1-mm sections and its Cherenkov radiation measured (©--©), the other was stained with Coomassie Blue and then scanned at 550 nm( ).

The 200 000 dalton component was always observed in our preparations. Further washings of sarcoplasmic reticulum microsomes with 0.6 IV[, KC1, 5 mM histidine, pH 7.0, did not result in any significant loss of this component, indicating that it was unlikely to be due to contaminating myosin. Analysis of Fig. 3 indicates that there is a small but reproducible amount of 32p radioactivity associated with this band (A2)2. Effect of the imido ester on sarcoplasmic reticulum microsomes The effect of increasing concentration of imido ester on sarcoplasmic reticulum membrane polypeptides is shown in Fig. 2. It is clear that even at the lowest concentration of imido ester used (1.7 mM), modification results in the production of a series of high molecular weight components with molecular weights which are multiples of 100 000. The highest multiple observable was 700 000 and from the intensity of Coomassie Blue stain, the concentration of the polymeric species appeared to decrease with increasing molecular weight. The imido ester contains a disulphide

84

bridge and thus reduction results in the reappearance of the pattern characteristic of materials that had not been cross-linked (Figs. 2e and 2a). It has previously been shown that the Ca 2÷-stimulated activity of sarcoplasmic reticulum membranes is maximal at pH 7.0 [38] and decreases at higher pH. It was therefore of importance when undertaking any modification of these membranes to maintain the pH of the cross-linking incubation as near to pH 7.0 as possible. For this reason we investigated the effect of p H on the cross-linking of sarcoplasmic reticulum membranes with imido ester (Fig. 4). No cross-linking of membrane polypeptides was observed at pH 6.0 or 7.0. However, at or above pH 8.0 cross-linking of the membrane proteins occurred as indicated by the appearance of higher molecular weight polypeptides in these gels. When the imido ester concentration was increased from 3.3 to 10 mM very similar results were obtained except that with the pH 9.0 sample very little protein now entered the gel, i.e. the extent of cross-linking increases with increasing pH. Thus cross-linking was carried out at pH 8.0 as this was the pH closest to that of the maximal Ca 2÷-stimulated ATPase activity at which cross-linking of the sarcoplasmic reticulum proteins could be achieved.

Fig. 4. The effect of pH on cross-linking of sarcoplasmic reticulum proteins with the imido ester. Proteins were cross-linked with 3.3 m M imido ester for 15 min at room temperature in 0.2 M triethanolamine buffers and the samples then prepared for electrophoresis on agarose/acrylamide gels. Approx. 50 ~tg of protein was loaded on to each gel. Uncross-linked (a). Cross-linked at pH 6.0 (b), pH 7.0 (c), pH 8.0 (d), pH 9.0 (e) and pH 10 (f).

85

Rate of cross-linking of sarcoplasmic reticulum proteins To determine whether any proteins were preferentially cross-linked by the imido ester the concentration of the individual membrane polypeptides remaining after cross-linking with varying concentrations of this reagent was investigated. The membranes were cross-linked, prepared for electrophoresis and then analysed on 8 ~o acrylamide gels as described previously [25]. The three components detectable, the 100 000 dalton Ca2+-stimulated ATPase, calcium binding protein and calsequestrin were identified [21] and their concentration in the gels calculated from the peak heights on recordings of their gel scans. The identification of calsequestrin was by purification according to the method of Ikemoto et al. [23]. The other component was assumed to be Ca z+ binding protein since it was present in significant quantity and had a molecular weight approx. 10 000, greater than that of calsequestrin [21]. The results shown in Table I indicate that the relative concentrations of the three proteins remained constant as the extent of cross-linking increased. Thus none of these components is preferentially cross-linked with this reagent. TABLE I THE RATIOS OF CALSEQUESTRIN AND CALCIUM BINDING PROTEIN RELATIVE TO THE Ca2+-STIMULATED ATPase A2 POLYPEPTIDE IN SARCOPLASMIC RETICULUM SAMPLES CROSS-LINKED WITH THE IMIDO ESTER AND CU/PHENANTHROLINE 50 ktg samples of cross-linked sarcoplasmic reticulum membranes were analysed on 8 % acrylamide gels. These were stained and scanned; peak heights were calculated and the ratio of these heights for calcium binding protein and calsequestrin relative to the Ca2+-stimulated ATPase A2 polypeptide calculated. Results obtained with the imido ester and Cu/phenanthroline were with different sarcoplasmic reticulum preparations. Concentration terms for Cu/phenanthroline cross-linking gave CuSO+ concentration/1,10-phenanthroline concentration. Calcium binding protein

Calsequestrin

Imido ester Untreated 0.9 mM 1.8 mM 2.7 mM 4.5 mM

0.50 0.40 0.37 0.49 0.48

0.37 0.31 0.37 0.41 0.39

Cu/phenanthroline Untreated 0.1 mM/0.25 mM 0.2 mM/0.25 mM 0.3 mM/0.25 mM

0.55 0.51 0.67 1.35

0.26 0.28 0.30 0.66

Analysis of imido ester cross-linked sarcoplasmie reticulum microsomes by two-dimensional electrophoresis The appearance on cross-linking of polymeric species with molecular weights which were multiples of 100 000, as seen in Fig. 2, indicated that these species could be derived from the 100 000 dalton Ca2+-stimulated ATPase. However, it was of interest to obtain more positive p r o o f that this was in fact the case. To do this we devised a two-dimensional gel electrophoresis system, similar to that described by

86 Wang and Richards [12], whereby the peptides contained within a polymeric band could be analysed. By incorporating fl-mercaptoethanol in the second dimension the cross-linkages between polypeptides due to imido ester cross-linkages are cleaved, revealing the polypeptides present in the first dimension polymeric species. When membranes were cross-linked with the imido ester (Fig. 5), analysis in the second dimension reveals the monomeric components originally present in the polymeric species. The identification of the components present in the second dimension is achieved by comparison with the gel shown on the left of this figure which is an uncross-linked sample of sarcoplasmic reticulum proteins. The gel along the top of the figure is a sample run only in the first dimension. The diagonal spots (Fig. 5,

Fig. 5. Two-dimensional gel of sarcoplasmic reticulum proteins cross-linked with the imido ester. Microsomes were cross-linked with 3.3 mM imido ester as described in Methods. Approx. 100/~g of protein was loaded on to the gel in the first dimension and samples electrophoresed as described in Methods. all spots 1) are similar to those observed when membranes were cross-linked with the non-cleavable cross-linking reagent dimethyl suberimidate, i.e. they represent nonreducible monomeric polypeptides. The major spots identified correspond to the Ca2÷-stimulated ATPase (Band A2), (A2)2 and Band C which contains calsequestrin and calcium binding protein. The off-diagonal spots (Fig. 5, all spots except 1) revealed in the second dimension must therefore be derived from the polymeric species present in the single dimension gel shown at the top of this figure. On the horizontal axis of Band C, three species are apparent, on the same axis of A2, five

87 b a n d s a n d on the axis o f (A2)2, four bands. All these spots (except for B a n d C, 3) are on an exact vertical line a n d represent families o f p o l y p e p t i d e s which were derived f r o m p o l y m e r s h a v i n g the same m o l e c u l a r weight.

Cross-linking of sarcoplasmic reticulum membranes with Cu/phenanthroline The cross-linking of these membranes with different concentrations of this reagent (Fig. 6) results in the appearance of polymeric species, similar to those obtained with the imido ester. The molecular weights of these species were multiples of 100 000 when assessed using a calibration curve similar to that shown in Fig. 1. Although only polymers of 200 000 and 300 000 dalton are produced at the lowest concentrations of Cu/phenanthroline used (Fig. 6b), production of species with molecular weights in excess of 400 000 were only obtained with CuSO4 concentrations of 0.05 mM and 1,10-phenanthroline concentrations of 0.1 mM (Fig. 6c). At higher concentrations of these reagents (Fig. 6f) a considerable proportion of the protein did not now enter the agarose/acrylamide gel; in this gel polymers up to 900 000 dalton could be observed. Because one uses two reagents for the disulphide formation reaction, copper sulphate and 1,10-phenanthroline it is possible that we may not have had the optimal conditions of these reagents for producing all the polymeric species possible. In Figs. 6g-6i, different ratios of copper sulphate to 1,10-phenanthroline were used.

Fig. 6. Cross-linking of sarcoplasmic reticulum proteins with Cu/phenanthroline. Microsomes ~ere cross-linked with varying concentrations of Cu/phenanthroline and analysed on agarose/acrylamide gels as described in Methods. Approx. 50 ttg protein loaded on to each gel. a, untreated microsomes; b, 0.02 mM CUSO4/0.05 mM 1,10 phenanthroline; c, 0.05 mM/0.01 mM, d, 0.1 raM/0.2 mM, e, 0.2 mM/0.4 mM; f, 0.3 mM/0.6 mM; g, 0.3 mM/0.25 mM; h, 0.2 mM/0.25 mM; i, 0.1 mM/0.25 mM, and j, 0.3 mM/0.25 mM reduced with dithiothreitol.

88 Again only polymers of 100 000 dalton polypeptides were visible (although the pentamer in Fig. 6g appeared to be present in greater amount than that in Fig. 6b, this was not usually observed). As with the imido ester, reduction of cross-linked samples results in the reappearance of a pattern of polypeptides very similar to that shown for control, uncr0ss-linked microsomes (Figs. 6j and 6a). In light of the recent report of Murphy [44] that tetramers of the Ca 2÷stimulated ATPase were formed preferentially when sarcoplasmic reticulum was treated with cupric phenanthrolate, we repeated the experiment shown in Figs. 6a-6f using 50 mM morpholino ethane sulphonic acid or 50 mM morpholino propane sulphonic acid, pH 7.0, which contained 1 mM CaCI2. The cross-linking pattern was very similar to that observed using phosphate buffer in the absence of CaC12, i.e. there was no specific formation of tetramers.

Rate of cross-linking of sarcoplasmic reticulum proteins using Cu/phenanthroline In a similar manner to that described for the imido ester, the relative concentrations of calsequestrin and calcium binding protein relative to the 100 000 dalton CaZ+-stimulated ATPase were investigated at different cross-linked concentrations. At the highest Cu/phenanthroline concentrations used, it is clear that the Ca 2+stimulated ATPase has been cross-linked preferentially by comparison with the two other components (Table I).

Analysis of Cu/phenanthroline cross-linked sarcoplasmic reticulum microsomes by twodimensional electrophoresis In Fig. 7 is shown the analysis of sarcoplasmic reticulum polypeptides which were cross-linked with Cu/phenanthroline and then analysed on the two-dimensional SDS electrophoresis system which contained fl-mercaptoethanol in the second dimension. As with the imido ester a series of off-diagonal spots are apparent. On both A2 and (A2)z horizontal axes multiples of 100 000 dalton are observed, up to a pentamer of A2 is visible. All are in a vertical line beneath each other, i.e. represent families of polypeptides having the same molecular weights. In the horizontal axis of Band C there is in no case any appearance of off-diagonal spots, i.e. there is no cross-linking of Band C polypeptides observed using Cu/phenanthroline. DISCUSSION The electrophoretic analysis of sarcoplasmic reticulum proteins shown in Fig. 2a indicates the presence of a major protein component in this membrane (Band A2) the Ca2+-stimulated ATPase. In the present work electrophoresis was carried out in sodium phosphate buffer, pH 7.0, rather than pH 6.0 buffer as used by Martonosi [42]. Although the phosphoprotein has been shown to be unstable above pH 6.0 [40], little decomposition of this intermediate occurred during the electrophoretic analysis at pH 7.0 since there is no radioactivity at the end of the gel, where free 32po43- would be found (Fig. 3). The agarose/acrylamide electrophoresis system reported here did not resolve calsequestrin from calcium binding protein and both migrated to the same position on these gels. The higher molecular weight components which are visible may either be genuine components in intact sarcoplasmic reticulum membranes or non-specific

89 First dimension

+

Fig. 7. Two-dimensional gel of Cu/phenanthroline cross-linked sarcoplasmic reticulum proteins. Microsomes were cross-linked with 0.1 mM CUSO4/0.25mM 1,10-phenanthroline and analysed as described in Methods. Approx. 100ttg of protein was loaded on to the first dimension gel. polymerization products. The phosphorylation of the 200 000 dalton component under the sa~ne conditions as those used for the phosphorylation of the Ca 2+stimulated ATPase (Fig. 3) indicates that it is a dimer of the Ca2+-stimulated ATPase [20]. The effect of pH on cross-linking of sarcoplasmic reticulum membrane proteins with the imido ester confirms the observations of Browne and Kent [45] who showed that the rate of amidine formation when primary amines are reacted with imido esters increased with increasing pH above pH 8.0. Although considerable hydrolysis of the ester by water occurs at this p H [45] the cross-linked products appear to be the same as those obtained at pH 10.0 (Fig. 4), i.e. while the rate of cross-linking is increased at higher pH, the nature of the cross-linked products produced is pH independent. Previously we have shown that the proteins present in sarcoplasmic reticulum membranes may be successfully cross-linked with the diimido ester dimethyl suberimidate [25]. In the present report it is clear that a very similar electrophoretic pattern is obtained with the reducible cross-linking reagents, imido ester and Cu/phenanthroline. In both cases (Figs. 2 and 6) polymers with molecular weights that are multiples of 100 000 are produced. Although the highest molecular weight species observed is 700 000 dalton in the imido ester gels, polymers with molecular weights of 900 000 have been observed when using Cu/phenanthroline. This does not necessarily mean that this is the subunit size of a macromolecular association of certain peptides in sarcoplasmic reticulum membranes; only that this is the limit for the visualization of

90 the polymeric species since their total concentration will decrease with increasing degree of polymerization [46]. Analysis of the two-dimensional gel shown in Fig. 5 indicates three series of cross-linked polypeptides are obtained using the imido ester, namely bands C, A2 and (A2)2. Band C, which contains calcium binding protein and calsequestrin, contains at least three components in its horizontal axis. The question is whether these are homo- or heteropolymers of these polypeptides with the other major cross-linked species A2. Spot 2 on the horizontal axis of Band C (Fig. 5) clearly can only arise from dimer formation of Band C. The concentration of the imido ester used for this cross-linking was 1.7 mM, when a similar concentration was used in an attempted preparation of polymeric bovine serum albumin, no cross-linking was observed. In fact at least a 5-fold increase in cross-linked concentration was required for production of polymeric bovine serum albumin species (discounting the dimer which is present as a contaminant in the albumin supplied). Davies and Stark [7] showed that in oligomeric proteins the protomers are cross-linked much more readily than are the polypeptides of monomeric proteins such as bovine serum albumin. Since Band C components can be cross-linked with imido ester concentrations as low as 1.7 mM, this would indicate that the components exist in the membrane as homopolymers of at least two subunits. The suggested involvement of calsequestrin and calcium binding protein in Ca 2÷ transport [21] could indicate that these two proteins are associated in the membrane with the Ca2÷-stimulated ATPase. Spot 3 on the horizontal axis of Band C (Fig. 5) is derived from a cross-linked component having a molecular weight of 150 000. It could be derived from either homopolymer formation between Band C polypeptides or heteropolymer formation between a Band C polypeptide and an A2 polypeptide. If this spot did arise from heteropolymer formation there should be in addition to spot 3 on the horizontal axis of Band C a spot vertically above on the A2 horizontal axis, i.e. mid way between spots 1 and 2 on the A2 axis. No such spot is visible and thus spot 3 on the horizontal axis of Band C (Fig. 5) must be derived from a homopolymer trimer of Band C polypeptides. Although spot 4 on Band C horizontal axis (Fig. 5) is present in very small quantity it could derive from heteropolymer formation between A2 and band C. However, this seems unlikely in the absence of heteropolymer formation between one Band C and one A2 polypeptide. We therefore conclude that band C components, calsequestrin and/or calcium binding protein, exist as oligomers in the membrane but are not associated with the CaZ÷-stimulated ATPase polypeptide A2. Since Band C components are cross-linked by the imido ester but not Cu/phenanthroline, these components must be between 2 and 11 A of each other. The other major cross-linked species observed in Figs. 5 and 7 is that having a molecular weight of 200 000, (A2)2. The spots on the horizontal axis of this polypeptide could be derived from the cross-linking of A2 and (A2)2. The fact that the multimeric species observed on the horizontal axis are multimers of 100 000 indicates that they are either homopolymers of A2 or heteropolymers of the 100 000 and 200 000 dalton species. Examination of Figs. 2e and 6j indicates that reduction of cross-linked membranes completely abolishes the polymers visualized in gels of cross-linked samples. Therefore the spots on the horizontal axis of (A2)2 (Figs. 5 and 7) most likely represent uncross-linked 200000 dalton polypeptides. The staining

91 intensity of the spots on the (A2)2 axis decreases with increasing molecular weight in a similar manner to Bands A2 and C, i.e. the concentration of monomer in each band decreases with increasing molecular weight. There is no greater staining intensity at 400 000 and 600 000 as compared with 300 000 and 500 000 as would be expected if considerable homopolymer formation of the 200 000 dalton species was occurring. Thus the spots on the horizontal axis of (A2)2 in Figs. 5 and 7 most likely arose from heteropolymer formation between A2 and (A2)2 components. This would be as expected if (A2)z is in fact a dimer of the Ca2+-stimulated ATPase, Band A2 (Fig. 3). In addition, the CaZ+-stimulated ATPase 100 000 dalton polypeptide exists as an oligomer in the sarcoplasmic reticulum membrane since there appears to be no limit to the extent of its homopolymer cross-linking. Because Cu/phenanthroline is able to cross-link it in a similar manner to the imido ester, the A2 polypeptides must be within 2/~ of each other. If the Ca2+-stimulated ATPase were to exist as an oligomer in the membrane this could possibly be detected on cross-linking by the appearance of a single crosslinked species derived from the A2 component. None was detected in this work although there have recently been reports that the enzyme can exist as an oligomer in sarcoplasmic reticulum membranes. Using electron microscopy Malan et al. [47] compared the densities of the outer projections seen by deep etching with the particles seen on the concave faces after freeze fracture. They calculated that each particle contains between 2.5 and 3.1 of the 100 000 dalton polypeptides, leMaire et al. [48] employing ultracentrifugation in the detergent Tween 80 report that the smallest fully active particles contain between three and four of the 10 0000 dalton polypeptides. In a recent study using the cross-linking reagent Cu/phenanthroline Murphy [44] observed tetramer formation only and no dimer, or trimer, formation. In addition he observed a component of molecular weight 700 000 that was produced on cross-linking sarcoplasmic reticulum membranes with this reagent. Although we investigated a variety of concentrations of copper sulphate and 1,10-phenanthroline the products observed on cross-linking were always similar to those obtained with the imido ester. In our uncross-linked preparations we do observe polypeptides with molecular weights of 500 000 and 700 000 but these disappear very rapidly on crosslinking, presumably because any polymers of these polypeptides would be unable to penetrate the agarose/acrylamide gels. A possible explanation for the difference between the results of Murphy [44] and those reported here lies in the recent work of Scales and Inesi [49]. They presented evidence that the Ca2+-stimulated ATPase can exist in different oligomeric forms at the outer and inner surface of the membrane. At the inner face it exists as an oligomer with approximately four subunits while at the outer it is monomeric. Thus depending on the conditions used it may be possible to produce cross-linked C a 2 + - s t i m u l a t e d ATPase species either derived from the oligomer on the inner face (as in Murphy's [44] report) or the monomer on the outer face (as in our results reported here). REFERENCES 1 Morrison, M. and Gates, R. E. (1972) in The Molecular Basis of Electron Transport (Schultz, J. and Cameron, B., eds.), pp. 327-340, Academic Press, London 2 Wallach, D. F. H. (1972) Biochim. Biophys. Acta 265, 61-83

92 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 ,~5 46 47 48 49

Juliano, R. L. (1973) Biochim. Biophys. Acta 300, 341-378 Carraway, K. L. (1975) Biochim. Biophys. Acta 415, 379-410 Branton, D. (1971) Phil. Trans. R. Soc. B261, 133-138 de Petris, S. and Raft, M. E. (1974) Eur. J. Immunol. 4, 130-137 Davies, G. E. and Stark, G. R. (1970) Proc. Natl. Acad. Sci. U.S. 66, 651-656 Sun, T.-T., Bollen, A., Kahan, L. and Traut, R. R. (1974) Biochemistry 13, 2334-2340 Chalkley, R. (1975) Biochim. Biophys. Res. Commun. 64, 587-594 Lehrer, S. S. (1972) Biochem. Biophys. Res. Commun. 48, 967-976 Steck, T. L. (1972) J. Mol. Biol. 66, 295-305 Wang, K. and Richards, F. M. (1974) J. Biol. Chem. 249, 8005-8018 Wang, K. and Richards, F. M. (1975) J. Biol. Chem. 250, 6622-6626 Lenard, J. (1970) Biochemistry 9, 5037-5040 Fairbanks, G., Steck, T. L. and Wallach, D. F. H. (1971) Biochemistry 10, 2606-2617 MacLennan, D. H. (1975) Can. J. Biochem. 53, 251-261 Ebashi, S. and Endo, M. (1968) Prog. Biophys. Mol. Biol. 18, 125-183 MacLennan, D. H. (1970) J. Biol. Chem. 245, 4508-4518 Hasselbach, W. and Migala, A. (1972) FEBS Lett. 26, 20-24 Warren, G. B., Toon, P. A., Birdsall, N. J. M., Lee, A. G. and Metcalfe, J. C. (1974) Proc. Natl. Acad. Sci. U.S. 71,622-626 MacLennan, D. H. (1974) in Methods in Enzymology (Fleischer, S. and Packer, L., eds.), Vol. 32, pp. 291-302, Academic Press, New York MacLennan, D. H. and Wong, P. T. S. (1971) Proc. Natl. Acad. Sci. U.S. 68, 1231-1235 Ikemoto, N., Bhatnagar, G. M., Bagy, B. and Gergeley, J. (1972) J. Biol. Chem. 247, 7835-7837 MacLennan, D. H., Yip, C. C., Iles, G. H. and Seeman, P. (1972) Cold Spring Harbor Symp. Quant. Biol. 31,469-477 Louis, C. and Shooter, E. M. (1972) Arch. Biochem. Biophys. 153, 641-655 Mar~onosi, A. (1968) J. Biol. Chem. 243, 71-81 Perham, R. N. and Thomas, J. O. (1971) J. Mol. Biol. 62, 415-418 Ruoho, A., Bartlett, P. A., Dutton, A. and Singer, S. J. (1975) Biochem. Biophys. Res. Commun. 63, 417-423 Stroh, R. (1953) German Pat. 894, 244 Johnson, T. P. and Gallagher, A. (1963) J. Org. Chem. 28, 1436-1437 Wang, K. and Richards, F. M. (1974) Isr J. Chem. 12, 375-389 Miller, G. A. and Hausman, M. (1971) J. Heterocyclic Chem. 8, 657-658 Ugel, A. G., Chrambach, A. and Rodbard, D. (1971) Anal. Biochem. 43, 410~26 Peacock, A. C. and Dingman, C. W. (1968) Biochemistry 7, 668-674 Griffith, I. P. (1972) Biochem. J. 126, 553-560 Louis, C. F., Buonaffina, R. and Binks, B. (1974) Arch. Biochem. Biophys. 161, 83-92 Lowry, O. H., Rosebrough, N. J., Farr, A. L. and Randall, R. J. (1951) J. Biol. Chem. 193, 265275 Yamamoto, T. and Tonomura, Y. (1967) J. Biochem. Tokyo 62, 558-575 Makinose, M. (1969) Eur. J. Biochem. 10, 74-82 Martonosi, A. (1969) J. Biol. Chem. (1969) 613-620 Inesi, G., Maring, E. Murphy, A. J. and McFarland, B. H. (1970) Arch. Biochim. Biophys. 138, 285-294 Martonosi, A. (1969) Biochem. Biophys. Res. Commun. 36, 1039-1044 Louis, C. F. and Irving, I. (1974) Biochim. Biophys. Acta 365, 193-202 Murphy, A. J. (1976) Biochem. Biophys. Res. Commun. 70, 160-166 Browne, D. T. and Kent, S. B. H. (1975) Biochem. Biophys. Res. Commun. 67, 126-132 Hucho, F., Mullner, H. and Sund, H. (1975) Eur. J. Biochem. 59, 79-87 Malan, N. T., Sabbadini, R., Scales, D. and Inesi, G. (1975) FEBS Lett. 60, 122-125 leMaire, M., Moller, J. V. and Tanford, C. (1976) Biochemistry 15, 2336-2342 Scales, D. and Inesi, G. (1976) Biophys. J. 16, 735-751