Identification of a set of calcium-binding proteins in reticuloplasm, the luminal content of the endoplasmic reticulum

DARRYL R. J. MACER and GORDON L. E. KOCH

Laboratory of Molecular Hiology, Hills Road, Cambridge CB2 2QII, luigland

Summary

A procedure was developed for the isolation of A calcium overlay test revealed that, in addition reticuloplasm, the luminal material of the endo- to endoplasmin, reticuloplasm contained at least plasmic reticulum (ER). A reticuloplasm-rich ex- three other calcium-binding proteins: i.e. BiP, PDI 3 tract was prepared from a murine plasmacytoma and the 55xlO Mr protein, respectively, with cell line that contains large amounts of ER, by first endoplasmin and the 55x10^, protein (CRP55) extracting the cytoplasmic contents using hypotonic accounting for the major proportion of the calcium- lysis to yield ER-rich 'shells' followed by mechan- binding activity. ical lysis to release the ER contents. The extract Treatment of cells with calcium ionophore led to contains five major proteins with apparent molecu- the specific over-expression of the major calcium- 3 lar weights of 100, 75, 60, 58 and 55 (xlO )Mr by binding reticuloplasmins endoplasmin, BiP and SDS— polyacrylamide gel electrophoresis. The 100, CRP55. 3 75 and 58 (Xl0 )Mr species were identified as the These studies show that the lumen of the ER known ER proteins endoplasmin, BiP and PD1, contains a family of proteins with the capacity to respectively. The ER association of the 60 and 55 bind significant amounts of calcium in the millimo- 3 (X10 ) Mr proteins was confirmed by conf ocal fluor- lar range and thereby to confer upon the ER the escence microscopy with affinity-purified anti- ability to perform a calcium storage function anal- bodies. ogous to that of the in Equilibrium dialysis with isolated reticuloplasm muscle cells. gave a calcium-binding capacity of 300 nmoles cal- cium per mg protein with half-maximal binding at 2+ 3 mM-Ca . Purified endoplasmin bound 280 nmoles Key words: endoplasmic reticulum, reticuloplasm, calcium- calcium per mg protein at a calcium concentration binding proteins, endoplasmin, BiP, protein disulphide 2+ of 5mM-Ca . isomerase, CRP55.

Introduction for removing Ca2+ from muscle cytoplasm to levels that would induce dissociation from troponin (MacLennan & Calcium ions play a vital role in a variety of cellular Holland, 1975). processes. This is achieved through a stringent control of It is now widely accepted that in non-muscle systems calcium levels within cells (Carafoli, 1987; Somlyo, the mitochondria play a relatively minor role in calcium 1984). One of the major mechanisms for maintaining homeostasis and attention has been focussed on the role intracellular calcium in the sub-micromolar range is the of the endoplasmic reticulum (ER) in the regulation of Ca2+-ATPase system, which pumps calcium actively out cytosolic calcium levels (Henkart, 1975; Somlyo, 1984; of the cytoplasm through the plasma membrane (Camp- Baker, 1986; Streb et al. 1983; Berridge & Irvine, 1984). bell, 1983). However, it is also known that intracellular In this respect it has been suggested that the ER might be membrane systems can sequester significant amounts of able to store and release calcium in a manner analogous to calcium and release this for use in intracellular signalling that involved in the sarcoplasmic reticulum (Baker, (Hasselbach & Makinose, 1961; Somlyo, 1984; Baker, 1986). Thus, it has been shown directly that the calcium 1986; Streb et al. 1983). One of the best-characterized concentration within the ER is relatively high and that intracellular systems operating under normal physiologi- the calcium can be released by agents such as inositol cal conditions is the sarcoplasmic reticulum (SR) in triphosphate and GTP (Streb et al. 1984; Gill et al. muscle, which has been clearly established as the system 1986). Isolated microsomes can accumulate calcium in an Journal of Cell Science 91, 61-70 (1988) Printed in Great Britain © The Company of Biologists Limited 1988 61 ATP-dependent process and release it under the in- LYS RET fluence of the above-mentioned agents (Spat et al. 1987). PS SH The existence of a Ca2+-ATPase activity capable of pumping calcium into the ER has also been suggested (Moore & Klaus-Freedmann, 1983). However, one major deficiency in this area is a lack of knowledge about the nature of the stored calcium in the ER. By analogy with SR one would expect the ER in non-muscle cells to contain one or more proteins in the luminal space or reticuloplasm that could carry out a role analogous to that of calsequestrin (MacLennan & Holland, 1975). Indeed, it has recently been established that the reticuloplasm is a protein-rich medium that contains a family of abundant resident proteins called reticuloplasmins (Koch, 1987), one of which is a major glycoprotein called endoplasmin, which is able to bind significant amounts of calcium in the millimolar range (Koch et al. 1986a). In this study we have carried out a systematic study of reticuloplasm with a view to identifying the repertoire of calcium-binding proteins associated with the ER. To achieve this, we have developed a general procedure for isolating reticuloplasm, identified and characterized the major abundant proteins in the isolate and examined these for calcium-binding activity.

Materials and methods

Growth and preparation of cells MOPC-315 cells (Eisen et al. 1968) were grown in Dulbecco's modified Eagle's medium (DMEM) with 10% foetal calf serum, * 100000 units ml"' penicillin/streptomycin and 4 mM-pyruvate to a density of ~5X 105 cells ml"', harvested by centrifugation « at 300if and washed thoroughly with phosphate-buffered saline (PBS).

Preparation of reticuloplasm Washed MOPC-315 cells were suspended at IX 107 cells min"' Fig. 1. Isolation of enriched reticuloplasm from a murine in lOmM-Tris-HC1, pH75, on ice and allowed to swell and plasmacytoma cell line (MOPC-315). PS, protein standards lyse completely. The resultant 'shells' were centrifuged out at (from top): glycogen phosphorylase (95X 10"'il/r); bovine s serum albumin (67X103); ovalbumin (45X103); carbonic 1000£, suspended in PBS at 1 X 10 cells equivalent per ml and 3 3 disrupted by syringing through an 18 gauge needle. The cell anhydrase (30X10 ); Q--lactalbumin (17X10 ). LYS, lysate debris was pelletted by centrifugingat 100 000g for 30 min. The from MOPC-315 cells; SH, hypotonic shells from MOPC-315 supernatant was collected and protease inhibitors (Koch et al. cells; RET, enriched reticuloplasm extracted from MOPC- 1986a) added. The supernatant was stored at 4°C. For longer 315 shells. The major proteins enriched in the reticuloplasm storage at —20°C glycerol was added to 50 % to prevent preparation are arrowed. freezing, and dialysed out before use. saponin was added to a final concentration of 001 %. After Preparation of sarcoplasmic reticulum vesicles 10 min on ice the membranes were centrifuged out at 100 000 £• The method used was based on that of MacLennan (1970), and for 50 min. The supernatant contains calsequestrin at >95 % was used to obtain vesicles from rabbit and munne skeletal purity. muscle. Muscle tissue was homogenized in an equal volume of 120niM-NaCl, 5 mM-imidazole, pH7-4. It was spun at 1000g Preparation of antibodies to individual reticuloplasmins for 5 min, and the pellet was re-extracted with buffer. The Rabbits (Dutch White or New Zealand White) were immunized combined superantants were squeezed through four layers of with 1-2 mg of purified reticuloplasm. The first injection was muslin, and the pH adjusted to 74. The suspension was spun at with Freund's complete adjuvant and subsequent injections 10 000 revs min"1 in a SS-34 rotor for IS min. The supernatant were at monthly intervals without adjuvant. Serum was col- was squeezed through four layers of muslin, and centrifuged at lected by standard procedures and stored in 002% sodium 16 000 revs min"1 in a SS-34 rotor for 50 min. The pellet is the azide. preparation of sarcoplasmic reticulum vesicles. Monospecific antibodies to particular proteins in the purified reticuloplasm were prepared, as described (Koch et al. 1986a). Purification of calsequestrin A sample of reticuloplasm was fractionated on a preparative Sarcoplasmic reticulum vesicles were suspended in PBS and SDS-polyacrylamide gel and the proteins were transferred to a

62 D. R. J. Macer and G. L. E. Koch .EP

P6O fDIP55

EP BiP

P60 PDl P55 Fig. 2. Monospecific antibodies to the major reticuloplasmins from MOPC-315 cells. Antibodies were prepared as described in Materials and methods, and tested by immunoblotting analysis on two-dimensional gels of lysates from MOPC-315 cells (the samples had been partially depleted for endoplasmin (Koch el al. 1986o) with ConA-agarose). Panel L shows the protein pattern from the lysate; and panels EP, BiP, P60, PDl and P55 show immunoblots developed with antibodies to endoplasmin, BiP, P60, PDl and CRPS5, respectively (see text). nitrocellulose sheet by electroblotting (see below). The nitro- (Koch el al. 1987) using the confocal microscope described by cellulose was stained with O'Ol % Ponceau Red in 5% acetic White el al. (1987). MOPC-315 cells were fixed in 3-5% acid and the strips corresponding to each protein component formaldehyde in PBS (5min), permeabilized with 1 % Nonidet were carefully excised. The strips were then immersed in the P40 (5 min), labelled with the appropriate antibody (5 min) and undiluted immune serum and mixed by rotation at 4°C for at developed with fluorescein-conjugated goat anti-rabbit Ig. least 24 h. The strips were removed, washed thoroughly with PBS and the antibodies eluted with O2M-glycine-HCl, pH2 Purification of pmteins fivrn reticuloplasm (lOmin at 4°C). The eluate was immediately neutralized to Endoplasmin was selectively extracted from purified reticulo- pH75 with 1 M-Tris-base and tested for antibodies by lmmu- plasm with Con A-Sepharose as described (Koch el al. 1986«). noblotting on cell lysates. Strips can be used repeatedly for the The non-bound fraction was dialysed against lOmM-Tns- HC1, extraction of antibody from unfractionated serum. pH75, and made up to 50% saturation with solid ammonium sulphate. The precipitate was centrifuged out and the super- Gel eiectmphoresis and immunoblolting natant made up to 85 % saturation with solid ammonium SDS—polyacrylamide gel electrophoresis was carried out ac- sulphate. The precipitate was collected by centrifugation, cording to the method of Laemmli (1970) and two-dimensional dissolved in lOmM-Tris- HC1, pH 7-5, and dialysed against the gel electrophoresis by the method of O'Farrell (1975). Immuno- same buffer. The sample was applied to DEAE-Sephadex blotting was as described by Towbin el al. (1979). Rabbit equilibriated with lOmM-Tris-HCI, pH75, and eluted with a IZ5 antibodies were developed with I-labelled Protein A (Amer- linear gradient (10 mM to 70mM-Tris- HCI, pH 7 5). Fractions sham). were monitored by SDS-polyacrylamide gel electrophoresis and immunoblotting and those containing (in order of elution) Immunojiiiorescence microscopy P58, BiP, PDl and CRP55 were pooled and concentrated. Each Immunofluorescence microscopy was carried out as described sample was fractionated by gel filtration on Sephacryl S200 to

Calcium-binding reticuloplasmins 63 Fig. 3. Immunofluorescence localization of the major reticuloplasmins from MOPC-315 cells. Samples were prepared and stained, as described (Koch et al. 1987). A. Endoplasmin; B, P60; C, PD1; D, CRPSS. give the purified proteins. All of the purified preparations were conditions resulted in the extraction of all the cvtoplasmic >90% pure for the respective protein (Macer, 1988). proteins, but endoplasmin itself was retained in the shells produced by the lysis. Fig. 1 shows that, in addition to 3 Results endoplasmin, the major 100xl0 /V/r protein in such preparations, several other major proteins are selectively Purification of reticuloplasm retained in the shells. When the shells were disrupted The approach used for the isolation of the luminal mechanically by repeated syringing through a hypoder- contents of the ER, referred to as reticuloplasm, was mic needle and the particulate material was pclletted by based on previous studies on the major glycoprotein high-speed centrifugation, the supernatant material con- endoplasmin (Koch et al. 1986a, 1987). We have pre- tained five major proteins with apparent molecular weights (on SDS-polyacrylamide gels) of 100, 75, 60, 58 viously established (Koch et al. 1988) that endoplasmin is 3 a major component of reticuloplasm and can therefore be and 55(XlO ).'l/r, respectively. used a marker for the isolation of the contents of the ER. It was found that the exposure of murine plasmacytoma Characterization of the major proteins in isolated cells (in this study we used the MOPC-315 line but any reticuloplasm plasmacytoma cell line can be employed) to hypotonic The major proteins in the soluble extract from the shells

64 D. R. J. Macer and G. L. E. Koch 300 a b c d e

-o

1 100 6 ~5

4 6 10 [Ca] (ITIM)

Fig. 4. Analysis of calcium-binding capacity of isolated reticuloplasm by equilibrium dialysis. Dialysis was carried out in 10niM-Tris-HCl, pH7-5, 100mM-K.Cl with varying concentrations of 45CaCl2 (Ikemoto, 1974). The protein concentration of the reticuloplasm sample was 8mgml~'. obtained from the plasmacytoma cells were characterized by determining the amino-terminal sequence of each. For this, each protein was purified by electroblotting onto PVDF membranes and the amino-terminal sequence was determined by the procedure described by Matsudaira (1987). The sequences obtained (Macer, 1988) confirmed 3 that the 100xl0 iV/r protein is endoplasmin and show 3 that the 75xl0 iV/r protein is another previously ident- ified ER-associated protein, the so-called immunoglobu- lin heavy-chain binding protein (BiP) (Lee et al. 1984; 3 Munro & Pelham, 1987). The 60 and 55(XlO )Mr proteins have amino-terminal sequences that are not identifiable with any known sequence in the major 3 sequence libraries. The 58xl0 yV/r protein did not yield any amino-terminal sequence, suggesting that it might be blocked. However, it was identifiable as the previously identified microsomal protein, protein disulphide isomer- ase (PDI) by immunoblotting with a monospecific anti- body (kindly provided by Dr R. Freedman). The ER location of the five major putative reticuloplas- mins was confirmed by immunofluoreseence microscopy. Monospecific antibodies to each protein were prepared by the nitrocellulose affinity strip procedure described (Koch et al. 1986a). Briefly, rabbits were immunized with the whole reticuloplasm extract. Nitrocellulose Fig. S. Detection of calcium-binding proteins in strips containing each protein were prepared, by running 4S reticuloplasm by Ca overlay. Samples were prepared and a preparative SDS-polyacrylamide gel on the reticulo- analysed as described in Materials and methods. Each panel plasm extract, electroblotting onto nitrocellulose and shows the protein (Pr) and 45Ca autoradiograph (Ca) from cutting out the strip bearing each band, and used to the same sample. Lanes a, protein standards (see Fig. 1); b, purify antibodies to each protein. The specificity of each whole cell lysate from MOPC-315; c, sarcoplasmic reticulum antibody was confirmed by immunoblotting on whole cell vesicles; d, reticuloplasm from MOPC-315 cells; e, purified lysates from the plasmacytoma cell line (Fig. 2). When calsequestrin. used in immunofluoreseence tests on plasmacytoma cells with the confocal fluorescence microscope, all antibodies material for the study of the calcium-binding properties gave the same pattern characteristic of the endoplasmic of reticuloplasm. reticulum of plasmacytoma cells (Fig. 3). The result for BiP is not shown, since its ER localization has been Calcium-binding capacity of isolated reticuloplasm established (Munro & Pelham, 1987). Equilibrium dialysis experiments were used to examine These studies confirm that all the major proteins in the the calcium-binding capacity of isolated reticuloplasm. sample of isolated reticuloplasm are reticuloplasmins, i.e. These experiments were carried out in lOmM-Tris- HC1, luminal proteins of the ER, and justify the use of this pH7-5, with lOOmM-KCl, since this is the standard

Calcium-binding reticuloplasmins 65 Ca Pr

Fig. 6. Identification of calcium-binding proteins in enriched reticuloplasm. Enriched reticuloplasm was purified, fractionated on two-dimensional polyacrylamide gels (O'Farrell, 1975), transferred to nitrocellulose (Towbin el al. 1979) and developed with 45Ca (Koch et al. 1986). The name spots were identified with the corresponding monospecific antibodies (Fig. 2). The sample was partially depleted for endoplasmin with ConA-agarose (Koch et al. 1986a). Spots 1, endoplasmin; 2, BiP; 3, PD1; 4, CRP55. buffer used in the study of calcium-binding by sarcoplas- (Koch et al. 19866). The advantage of this technique, mic reticulum vesicles and the major calcium-storage apart from its intrinsic convenience, is that it can be used protein calsequestrin (Campbell et al. 1983). Fig. 4 on complex mixtures as well as pure samples of proteins shows that calcium-binding saturates in the millimolar (Fig. 5) and with a sample of purified sarcoplasmic range, half-maximal binding occurring at about 2-5 raM- reticulum vesicles (Fig. 5). In the synthesis mixture, calcium. At lOmM-calcium the isolated reticuloplasm significant calcium-binding is observed with only the binds ~300nmoles calcium per mg protein. known calcium-binding protein a'-lactalbumin. In the Previous studies have shown that endoplasmin is one of sample of SR vesicles, the major calcium-binding species the major calcium-binding proteins in plasmacytoma cells has an apparent molecular weight of ~55xlO3 and co- (Koch et al. 1986a). However, the binding studies were migrates with pure calsequestrin, which itself binds not performed under equilibrium conditions so they were substantial amounts of calcium under the conditions repeated here using equilibrium dialysis, as described used. We therefore conclude that the 4SCa overlay tech- above. At 5 mM-calcium native endoplasmin bound nique was suitable for the detection of calcium-binding ~280nmoles calcium per mg protein, supporting the proteins of the calsequestrin-type, that is, high-capacity previous suggestions that endoplasmin is a major cal- and low-affinity. cium-binding protein of the ER. Fig. 5 shows that there are several calcium-binding proteins in a whole cell lysate derived from the plasma- Calcium-binding proteins in reticuloplasm cytoma cell. The major binding species have apparent The repertoire of calcium-binding proteins in reticulop- molecular weights of 100, 90, 75, 58 and 55X103, lasm was examined by the SDS-polyacrylamide gel/elec- respectively. Fig. 5 shows that the 100, 75, 58 and 4S 3 troblotting/ Ca overlay technique used previously 55(XlO )Mr calcium-binding species co-purify with re-

66 D. R. y. Macer and G. L. E. Koch Effect of calcium ionophore on the protein composition of plasmacytoma cells Fig. 8 shows that calcium ionophore has a specific effect on the protein composition of MOPC-315 cells. Even at the level of one-dimensional SDS-polyacrylamide gel electrophoresis, at least three protein bands are increased significantly and these bands co-migrate with endoplas- 3 min, BiP and the 55xlO Jl7r reticuloplasmin. Immuno- blotting studies (Fig. 8) confirm that these three proteins I 0 I I I are indeed over-expressed in cells treated with calcium o 0 10 3 ionophore. In contrast, PDI and the 60xl0 /l/r reticu- loplasmin are not increased by ionophore (Macer, 1988). 16 Studies on the kinetics of this overexpression show that, as observed previously for endoplasmin (GRP95) and BiP (GRP75) (Lee et al. 1984; Welch et al. 1983), the 12 increased expression of these proteins is apparent only after 3-4 h in the ionophore and reaches a maximum in 12-16 h. These studies extend the previous observations that endoplasmin and BiP are over-expressed in calcium ionophore-treated cells, by showing that a third protein, 3 the 55xl0 yV/r reticuloplasmin (referred to as CRP55 to emphasize its calcium-binding and calcium-regulated 10 characteristics), it also over-expressed in such cells, and [Cal (M) also show that all three proteins specifically increased in Fig. 7. Concentration dependence of calcium-binding these cells are prospective calcium-binding proteins in proteins in the 45Ca overlay assay. The assay was carried out, the reticuloplasm. as described in Materials and methods. (•), Rabbit calsequestrin; (•), human ; (D), murine endoplasmin; (A), murine CRPS5; (O), human transferrin. Discussion

The purpose of the present study was to determine ticuloplasm. Two-dimensional gel analysis (Fig. 6) whether reticuloplasm, the luminal material of the endo- showed that the major calcium-binding proteins in reticu- plasmic reticulum (Krstic, 1979; Koch, 1987), contains 3 loplasm were endoplasmin, BiP, PDI and the 55xl0 Mr proteins that could serve the calcium-storage function reticuloplasmin, respectively. performed by calsequestrin in the sarcoplasmic reticu- Sufficient quantities of the proteins, apart from endo- lum. To this end, we developed a simple procedure for plasmin, were not available for examining their calcium- isolating reticuloplasm from a cultured cell line. The binding capacities by equilibrium dialysis. However, the choice of the plasmacytoma cell was determined by the nitrocellulose binding test was adapted to permit a fact that it contains large amounts of endoplasmic reticu- comparison with calsequestrin under non-equilibrium lum. Consequently, when the cells are perforated by conditions. Briefly, small amounts of purified non-de- hypotonic lysis and the cytoplasmic contents released, the natured protein were spotted out on nitrocellulose filters resultant shells are essentially ER and nuclei only. and incubated with varying concentrations of 4SCa, the Fortunately, when these hypotonic shells are disrupted filters briefly washed and each sample counted to measure mechanically, the contents of the ER are released in a calcium binding. Fig. 7 shows the binding profiles for soluble form and are easily obtained in a highly purified several proteins obtained in this assay. Proteins like state. This was confirmed directly by showing that all the transferrin do not bind significant amounts of calcium major proteins in the extract are located in the endoplas- even at millimolar concentrations. Gelsolin, which is a mic reticulum. It is worth noting here that shells of very cytoplasmic calcium-binding protein, shows saturable similar composition can be obtained by perforating the binding in the region of 10^M-calcium. However, it is same cells with low concentrations of saponin under clear that low-affinity/high capacity calcium-binding pro- isotonic conditions. However, mechanical disruptions of teins such as calsequestrin can be distinguished from the such shells does not release the contents of the ER but high-affinity cytoplasmic calcium-binding proteins, such rather generates closed vesicles that retain the reticulo- as gelsolin, by this assay. Fig. 7 also shows the binding plasmins. These can be released from such preparations 3 profiles for endoplasmin and the 55xl0 il7r protein and only with stronger detergents. Thus, in order to obtain confirms that they resemble calsequestrin with respect to reticuloplasm without resorting to the use of detergents, both affinity and capacity. Thus, although the assay has it is preferable to use the hypotonic lysis procedure. the limitation of a non-equilibrium analysis, it does Equilibrium dialysis studies on isolated reticuloplasm indicate that the binding characteristics of the calcium- clearly showed that it has a significant calcium-binding binding reticuloplasmins are similar to those of the capacity. Calsequestrin, under identical conditions, sarcoplasmic reticulum protein, calsequestrin. binds 250-800 nmoles of calcium per mg protein and

Calcium-binding reticuloplasmins 67 Fig. 8. Increased expression of calcium-binding reticuloplasmins in calcium inophore-treated cells. MOPC-315 cells were incubated with ( + ) and without ( —) 5;tM-A23187 for 16h and analysed on SDS-polyacrylamide gels and immunoblotting using monospecific antibodies. The antibodies, to endoplasmin (EP), BiP and CRP55 (P55), were used successively on the same sample. A. Protein stain; arrows show proteins with increased expression in lonophore, which co-migrate with endoplasmin (EP), BiP and CRP55 (P55), respectively. B. Immunoblots showing, from the top, endoplasmin, BiP and CRP55. most of this binding occurs in the millimolar range (Cala calcium-binding proteins. This is probably not surpris- & Jones, 1983; Campbell et al. 1983; Mawere/ al. 1985). ing, since the calcium-binding sites in such proteins In the case of reticuloplasm, half-maximal binding is at appear to consist of clusters of negative charges, rather 2-5 mM and at lOmM-calcium the binding capacity is than any specific fold, such as the E-F hand found in the 300nmoles per mg protein. Thus, the calcium-binding high-affinity calcium-binding proteins. It is also worth activity in reticuloplasm is comparable to that of calse- emphasizing that the test is quite specific, in that it clearly questrin. The implication is that reticuloplasm contains distinguishes between proteins that are known to bind proteins that have the ability to bind calcium with high significant amounts of calcium and those that do not. In capacity in the millimolar range. However, in prelimi- the case of endoplasmin it was directly confirmed by nary studies with an antibody to calsequestrin, no evi- equilibrium dialysis that the protein does bind significant dence was obtained for a calsequestrin-like protein associ- amounts of calcium under the conditions used in the ated with the ER itself. Therefore, it was necessary to calsequestrin binding assay. In view of this, we conclude determine which proteins actually performed the cal- that endoplasmin, BiP, PDI and CRP55 are all signifi- cium-binding function in reticuloplasm. cant calcium-binding proteins of the reticuloplasm. Of The procedure used to identify the calcium-binding these, CRP55 appears to have a similar calcium-binding proteins in reticuloplasm was to fractionate the reticulo- capacity and affinity to calsequestrin, whilst endoplasmin plasm by gel electrophoresis, transfer the proteins to is of moderate capacity. BiP and PDI bind smaller nitrocellulose and probe directly for calcium-binding amounts of calcium but their capacity appears significant with 45Ca. This approach has been used previously to enough to make a contribution to the total calcium- identify proteins that bind calcium in the micromolar binding capacity of reticuloplasm. range (Koch et al. 1986a,6; Maruyama et al. 1984). The main conclusion to emerge from these studies is However, using calsequestrin as a model, we have shown that the lumen of the ER does contain a number of that the method also works for low-affinity/high-capacity proteins that could participate in the calcium storage

68 D. R. jf. Macer and G. L. E. Koch function attributed to this organelle in non-muscle cells. identification of the 53 000 dalton glycoprotein. J. biol Chem. 258, Since the proteins concerned are known to exist in the ER 1197-1204. CARAFOLI, E. (1987). Intracellular calcium homeostasis. A. Rev. of most, and probably all, eukaryotic cells (Macer, 1988), Diochem. 56, 395-433. it follows that the ER can function as a general calcium EISEN, H. N., SIMMS, E. S. & POTTER, M. (1968). Mouse myeloma storage organelle. proteins with anti hapten antibody activity. The protein produced The obvious question that arises now concerns the by plasma cell tumour MOPC-315. Biochemistry 7, 4126-4133. GILL, D. L., MEDA, T., CHEUCH, S. J. & NOEL.M. \V. (1986). demonstration that the proteins actually perform the 2+ Ca release from the endoplasmic reticulum is mediated by a storage function in vivo. This is not easy to do and even guanine nucleotide regulatory mechanism. Nature, Loud. 320, in the case of calsequestrin, it is based only on circum- 461-464. stantial evidence. However, at least one line of evidence HASSELBACH, W. & MAKINOSE, M. (1961). Die calcium pumpe dcr does exist for this in the case of the ER proteins. It has enschaffungs grana. Biochem. Z. 333, 518-528. been known for some time that calcium ionophores have a HENKART, M. (1975). Endoplasmic reticulum sequesters calcium in the squid giant axon. Biophys. J. 15, 267a. very specific effect on the protein composition of cells. IKEMOTO, N. (1974). The calcium binding sites involved in the Previous studies showed that endoplasmin and BiP were regulation of the purified adenosine triphosphatase of the over-expressed in ionophore-treated cells (Lee, 1987). In sarcoplasmic reticulum. J. biol. Chem. 249, 649-651. this study we have confirmed that a third protein, the KRSTIC, R. V. (1979). infrastructure of the Mammalian Cell. Berlin: 55xl03iV/ reticuloplasmin identified in this study Spnnger-Verlag. r KOCH, G. L. E. (1987). Reticuloplasmins, a novel group of proteins (CRP5S), is also over-expressed in ionophore-treated in the endoplasmic reticulum. J. Cell Sci. 87, 491-492. cells. Thus, the specific effect of altered calcium metab- KOCH, G. L. E., MACER, D. R. J. & SMITH, M. J. (1987). olism in cells is to induce the increased production of Visualization of the intact endoplasmic reticulum by three ER proteins that have the capacity to bind calcium. immunofluorescence with antibodies to the major ER glycoprotein, It is reasonable to speculate that such a response suggests cndoplasm. J. Cell Sa 87,535-542. KOCH, G. L. E., MACER, D. R. J. & WOODING, F. B. P. (1988). a role for the calcium-binding, calcium-regulated pro- Endoplasmin is a reticuloplasmin. J. Cell Set. 90, 485-491. teins in calcium homeostasis. One possibility is that the KOCH, G., SMITH, M., MACER, D., WEBSTER, P. & MORTARA, R. response leads to an increase in the intrinsic calcium (1986a). Endoplasmic reticulum contains a common abundant storage capacity of the ER and thereby permits the cell to calcium-binding glycoprotein endoplasmin. J. Cell Set. 86, counteract the damaging effects of increased cytosolic 217-232. KOCH, G., SMITH, M., TWENTYMAN, P. & WRIGHT, K. (19866). calcium caused by the ionophore. Such an explanation is Identification of a novel calcium-binding protein (CP22) in multi- also consistent with the established phenomenon that a drug resistant murine and hamster cells. FEBS Lett. 195, 275—279. variety of stresses can also induce the synthesis of the LEE, A. S. (1987). Co-ordinated regulation of a set of by calcium-binding reticuloplasmins. It is believed that one glucose and calcium ionophores in mammalian cells. Trends Biochem. Sci. 12, 20-23. of the common events following the exposure of cells to LEE, A. S., BELL, J. & TING, J. (1984). Biochemical characterisation stress is an alteration in calcium levels in the cytosol of the 94- and 78-kilodalton glucose regulated proteins in hamster (Trump et at. 1981). If not corrected, this can lead to cell fibroblasts. J. biol Chem. 259, 4616-4621. death. The observed increased synthesis of the prospec- LAEMMLI, U. K. (1970). Cleavage of structural proteins during the tive calcium storage proteins in the ER could be one way assembly of the head of bacteriophage T4. Nature, Land. 227, 680-685. of preventing this calcium-induced catastrophe. MACER, D. R. J. (1988). Identification and analysis of proteins of the In this study, we have been concerned with the reticuloplasm. Ph.D. thesis, University of Cambridge. question of whether the ER lumen does contain proteins MACLENNAN, D. H. (1970). Purification and properties of an that could carry out a calcium-storage function and we adenosine triphosphatase from sarcoplasmic reticulum. J. biol. Chem. 245, 4508-4518. conclude that it does. However, it should be emphasized MACLENNAN, D. H. & HOLLAND, P. C. (1975). Calcium transport in that, unlike the sarcoplasmic reticulum, the ER performs sarcoplasmic reticulum. A. Rev Biophys Bioengng. 4, 377—404. several other functions and the function of the relevant MARUYAMA, K., MIKAWA, T. & EBASHI, S. (1984). Detection of 45 calcium-binding proteins may not be confined to calcium- calcium binding protein by Ca autoradiography on nitrocellulose membranes after SDS clectrophoresis. J. Biochem. 95, 511-519. storage. MATSUDAIRA, P. (1987). Sequence from picomole quantities of proteins electroblotted onto polyvinyhdcne fluoride membranes. .7. biol. Chem. 262, 10035-10038." MAWER, A., TANAKA, M., OZAWA, T. & FLEISCHER, S. (1985). References Purification and crystallisation of the calcium binding protein of the sarcoplasmic reticulum from . Proc. natn Acad. BAKER, P. F. (1986). GTP and calcium release. Nature, Loud. 320, Sci. U.SA. 82, 4036-4040. 395. MOORE, P. B. & KLAUS-FREEDMAN, N. (1983). Hepatic microsomal BERRIDGE, M. J. & IRVINE, R. F. (1984). Inositol triphosphate, a Ca2+ -dependent ATPase. Biochem J 214, 69-75. novel second messenger in cellular . Nature, MUNRO, S. & PELHAM, H. R. B. (1987). A C-terminal signal Lond 312, 315-321. prevents the secretion of luminal ER proteins. Cell 48, 899-907. CALA, S. E. & JONES, L. R. (1983). Rapid purification of O'FARRELL, P. H. (1975). High resolution two-dimensional calsequestrin from cardiac muscle sarcoplasmic reticulum by Ca2+ electrophoresis of proteins. J. biol. Chem. 250, 4007-4021. dependent clution from phcnyl-Sepharosc. J. biol. Chew 258, SOMLYO, A. P. (1984). Cellular site of calcium regulation. Nature, 11932-11936. Land. 309, 516-517. CAMPBELL, A. K. (1983). lntracelhdar Calcium, its Universal Role SPAT, A., LUKACS, G. L., EBERHARDT, I., KIESEL, L. & as Regulator. New York: Wiley. RUNNEBAUM, B. (1987). Binding of inositol phosphates and CAMPBELL, K. P., MACLENNAN, D. H., JORGENSEN, A. O. & induction of Ca2+ release from pituitary microsomal fractions. MlNTZEL, M. C. (1983). Purification and characterisation of Biochem. J. 244, 493-496. calsequestrin from canine cardiac sarcoplasmic reticulum and STREB, H., IRVINE, R. F., BERRIDGE, M. J. & SCHULZ, I. (1983).

Calcium-binding reticuloplasmins 69 Release of calcium from a non-mitochondrial store in pancreatic WELCH, W. J., GARRELLS, J. 1., THOMAS, G. P., LIN, J. J. C. & acinar cells by inositol-l,4,5,-triphosphate. Nature, Land. 306, FERAMISCO, J. R. (1983). Biochemical characterisation of the 67-69. mammalian stress proteins and identification of two stresa proteins TOWBIN, H., STAEHLIN, T. & GORDON, J. (1979). Electrophoretic as glucose- and calcium-ionophore regulated proteins, jf. biol. transfer of proteins from polyacrylamidc gels to nitrocellulose Chew. 256, 5309-5312. sheets; procedure and some applications. Proc. naln. Acad. Sci. WHITE, J. G., AMOS, \V. B. & FORDHAM, M. (1987). An evaluation L'.SA. 76, 4350-4354. of confocal versus conventional imaging of biological structures by TRUMP, B. F., BEREGESKY, 1. K. & OSORNIS-VARGAS, A. R. (1981). fluorescence light microscopy. J. Cell Biol. 105, 41-48. In Cell Death m Biolog\' and Pathology (ed. I. D. Bowen & R. A. Lockshin), pp. 209-242. London, New York: Chapman & Hall. (Received 14 April I9SS - Accepted 2 June I9SS)

Note added in proof

45Ca overlay experiments using nitrocellulose filters were carried out in 25mM-Hepes, pH7-2, 100mM-KCl, 10mM-MgCl2, using the procedures described by Koch etal. (1986a).

70 D. R. J. Macer and G. L. E. Koch