Calcium-promoted aggregation of erythrocyte membrane proteins

Biochimica et Biophysica Acta, 379 (1975) 571-581

© Elsevier Scientific Publishing Company, Amsterdam - - Printed in The Netherlands BBA 36938 C A L C I U M - P R O M O T E D A G G R E G A T I O N OF E R Y T H R O C Y T E M E M B R A N E PROTEINS *

KERMIT L. CARRAWAY, RICHARD B. TRIPLETT and DARRELL R. ANDERSON Department of Biochemistry, Oklahoma State University, Stillwater, Okla. 74074 (U.S.A.)

(Received July 18th, 1974)

SUMMARY Introduction of Ca 2+ ( > 1 m M ) into erythrocytes during hemolysis causes formation of an aggregate which is highly resistant to disruption by sodium dodecylsulfate and other denaturing agents. The process is temperature dependent, but it does not require incubation in isotonic medium. Aggregation can be prevented but not reversed with chelating agents such as ATP or EDTA. The aggregate can be isolated by chromatography in dodecylsulfate on Sepharose 4B. Its amino acid composition indicates that it contains spectrin as the primary, but not exclusive, polypeptide component. Aggregate formation does not require increased Ca 2÷ binding to the membranes, and no 45Ca2÷ could be detected in the aggregate which had been separated by acrylamide electrophoresis on sodium dodecylsulfate. This indicates that the Ca 2+ is important in the formation of the aggregate, but not in its stabilization or maintenance once it has been formed.

INTRODUCTION The effect of Ca 2+ on erythrocyte membranes is of considerable interest because of its possible involvement in erythrocyte deformability and destruction [1, 2]. Introduction of Ca 2÷ into resealed ghosts from fresh erythrocytes causes significant volume loss [3] and transforms them into rigid spheres [1]. This loss of deformability can be prevented by A T P or E D T A introduced into the ghost [3]. Palek et al. [4] have suggested that a CaZ+-activated ATPase might be involved in the contraction of the cell. Alternatively, a sol-to-gel transformation of proteins at the interior surface of the cell membrane has been postulated to be responsible for the stiffening of the cell membrane [1]. Recently we have demonstrated an apparent aggregation of membrane proteins promoted by the introduction of Ca 2÷ into the resealed ghosts [5]. The present work describes studies on the means by which Ca z+ may promote the aggregation.

* Journal article J-2895 of the Agricultural Experiment Station, Oklahoma State University, Stillwater, Okla. This research was conducted in cooperation with the USDA, Agricultural Research Service, Southern Region.

572 EXPERIMENTAL P R O C E D U R E

Hemolysis Washed, packed erythrocytes were hemolyzed at 4 °C for 10 rain in 10 vol. of 10 mM Tris buffer (pH 7.4) containing appropriate concentrations of Ca z+ or chelating agents [5]. In some experiments the ghosts were subsequently incubated for 30 min at 4 or 37 °C in 0.17 M salt, obtained by adding 3.0 M NaCl or a mixture of 1.42 M KC1 and 0.28 M NaC1 to the hemolysis mixture. This procedure induces ghost resealing [6]. The resealed ghosts were then hemolyzed in 10 mM Tris at 4 °C and washed until essentially free of hemoglobin. In some experiments Ca z + was added at the second hemolysis step instead of the first. Analytical procedures Polyacrylamide gel electrophoresis of membranes was performed on 5 ~ gels by previously reported procedures [7, 8]. Gels were stained for protein by Coomassie Blue and for carbohydrate with periodate-Schiff [9]. Gels containing radioactive samples were sliced and counted as reported previously [10]. Column chromatography of membrane samples was performed in dodecylsulfate on Sepharose 4B (ref. 10). Circular dichroism spectra of selected eluted fractions in 1 ~ dodecylsulfate were recorded on a Cary Model 61 recording spectropolarimeter using a 1.0-mm path length, a scan rate of 0.2 nm/s and a period of 10 s [11]. The instrument was calibrated by the method of Cassim and Yang [12]. Ca 2+ binding Red cells or ghosts equivalent to I vol. of packed cells were suspended in 10 mM Tris buffer (pH 7.4) containing 0-10 mM Ca 2+ and 1/tCi 45Ca2+. The samples were made 0.17 M in salt by addition of 3 M NaCI, incubated for 30 min at 37 °C and washed 3 times with 10 mM Tris. Aliquots of the ghost suspension were solubilized in NCS (Amersham/Searle) containing 100 mM CaCI2 before counting in tolueneethanol. RESULTS

Ca z + promoted aggregation of membrane proteins The preparation of erythrocyte membranes for these experiments involves a four-step process: l) hypotonic hemolysis; 2) resealing by incubation in isotonic salt solution; 3) re-hemolysis; and 4) washing of the membranes to remove hemoglobin [5]. Addition of Ca 2÷ in sufficient concentrations during the initial hemolysis causes an apparent protein aggregation (Fig. l, Arrow 1), determined by dodecylsulfateacrylamide electrophoresis. A concomitant apparent proteolysis [5] is also shown by Arrows 2 and 3. Since the mode of action of Ca z+ was of considerable interest, the effects of the addition of Ca z+ at the subsequent membrane-preparation steps were also investigated. Addition of Ca z+ up to 5 mM with the resealing buffer (Step 2) causes no effects (Fig. 2). However, addition of Ca 2+ at the second hemolysis step causes essentially the same effects as at the first hemolysis step. Thus it appears that either the ghosts must be "open" for Ca z+ to be effective, or the salt in the resealing medium must compete with Ca 2+ for the sites of Ca 2+ action.

573

III IV

Fig. 1. CaZ+-promoted aggregation of erythrocyte membrane proteins. Erythrocytes were hemolyzed in Caz+ solutions in 10 mM Tris (pH 7.4), and membranes were prepared as described previously. Membrane samples were subjected to electrophoresis on 5 ~ polyacrylamide gels in dodecylsulfate. Bands are numbered as in previous publications [8-10]. Ca2+ concentrations are 0, 0.1, 0.5, 1.0, 2.0 and 5.0 mM for Gels A-F, respectively.

Since the aggregation phenomenon might be induced by resealing in isotonic medium, the effects of C a 2+ w e r e tested under conditions in which the hemolysate from Step 1 was not brought to isotonic conditions. Instead the mixture was incubated at 37 °C under hypotonic conditions. The results are essentially equivalent to those obtained with resealing (Fig. 1), indicating that resealing is not required for aggregation if sufficient Ca 2÷ is present at hemolysis. I f the incubation following hemolysis is performed at 4 °C, the results are slightly different. Significant aggregation does not occur until the Ca 2÷ concentration reaches about 5 m M (Fig. 3). The two proteolytic processes also show different responses at this temperature. Addition of chelating agents with Ca 2+ during the initial hemolysis prevents the aggregation and proteolysis (Fig. 4). However, if the C a 2 ÷ is present at the initial hemolysis step, the chelating agents are not effective if added in the subsequent steps. Neither aggregation nor proteolysis is prevented or reversed by these treatments. The membrane glycoprotein(s) do not appear to be present in the aggregate. Periodate-Schiff staining of electrophoretic gels of CaE÷-treated samples showed no

574

Fig. 2. Effect of Ca2÷ during resealing. Erythrocytes were hemolyzed in 10 mM Tris and resealed in the presence of 0.17 M NaC1 containing increasing Ca2+ concentrations. Membranes were prepared as before. Ca2+ concentrations are the same as in Fig. 1. evidence of carbohydrate in the aggregate band and no significant decrease in the intensity of the major glycoprotein band. Therefore, it does not appear that the major sialoglycoprotein of the membrane is a part of the aggregate after dissociation in detergent. It was not possible to determine if other, less highly glycosylated proteins may be present.

Isolation of the aggregate and dissociation attempts Because of its large size in dodecylsulfate the aggregate can be isolated by chromatography on Sepharose 4B in the detergent. Fig. 5 shows elution profiles of membranes prepared from Ca 2+-treated and untreated cells. The profiles differ primarily by the presence of an excluded peak (Fractions 4 ~ 4 5 ) in the treated samples and by changes which occur in the areas of the proteolyzed components. Electrophoresis of the excluded peak in dodecylsulfate indicates that it bands near the top of the electrophoresis gels, as does the aggregate. The fractions containing the aggregate were pooled, dialyzed to remove detergent and hydrolyzed for amino acid analysis. The amino acid composition is shown in Table I, together with the composition for spectrin, the high-molecular-weight erythrocyte membrane protein which is denoted as

575

III IV

Vii VIII

A

B

¢

D

E

F

Fig. 3. Temperature effects on aggregation. Erythrocytes were hemolyzed in 10 mM Tris with increasing concentrations of Ca ~+ and maintained at 4 rather than at 37 °C for incubating in the isotonic medium. Ca 2+ concentrations are the same as in Fig. 1.

Band I in Fig. 1. Comparison of these amino acid analyses with similar analyses of the other column chromatography fractions indicate that spectrin comprises a major fraction of the aggregate. The key observation is the high glutamic acid content of both spectrin and the aggregate. None of the other membrane proteins which might contribute have such a high glutamic acid content. It is also obvious, however, that spectrin cannot account for the analytical data completely. There must be at least one additional protein present. The conformation of the aggregate isolated by chromatography in detergent was examined by circular dichroism to determine if the apparent high molecular weight might be due to conformational restrictions which prevent the protein from assuming the appropriate shape in the detergent solution [13]. The spectrum is typical of that of other globular proteins in this detergent [14]. This suggests that a true aggregation of polypeptides rather than a conformational anomaly is responsible for the high molecular weight. A number of attempts have been made to disrupt the aggregate in the membrane

576

Fig. 4. Effect of chelating agents on aggregation. Erythrocytes were hemolyzed in the presence of 10 mM Tris and 1 mM Ca 2÷, except for Gel A, which contained no Ca2+. For Gels B-D the hemolyzing solution contained 2 mM ATP, EDTA or EGTA, respectively. The hemolysis solution for Gel A contained no chelator. before the membranes were subjected to dodecylsulfate electrophoresis. The dissociating agents included 3 ~ dodecylsulfate (zk heat), 6 M guanidine hydrochloride, 8 M urea (zk 0.2 9/o EDTA), phenol-acetic acid-water (1:1:1, by vol.), 90 ~ 2-chloroethanol and chloroform-methanol (2:1, v/v). Samples were dialyzed against distilled water and solubilized in dodecylsulfate for electrophoresis. In no case was any dissociation of the aggregate detected. I f any disaggregation occurred during these treatments, it must have been reversed during the removal of the dissociating agent. Further attempts have been made at disaggregation of the material isolated by Sepharose 4B chromatography. By heating in 4 70 dodecylsulfate containing 1 m M EDTA, a limited amount of dissociation could be achieved. Electrophoresis indicated that Component III, the 100 000-mol. wt. polypeptide which extends through the membrane, was the major protein released by disaggregation. Unfortunately, the dissociation experiments have not been highly reproducible. Several preparations of aggregate have been completely resistant. Components III and IV can be separated from the other membrane proteins by chromatography in dodecylsulfate. I f this fraction is dialyzed to remove the bulk of the detergent and treated with Ca 2+ (1 mM),

577 0.4

UNTREATE

0.3

0.2

0.1

O.C 0

/

TREATED 0.4

02

00.I C .0"2 2O t 40 60 80 Fraction No.

Fig. 5. Chromatography of membranes from CaZ+-treated and untreated erythrocytes. Membranes were prepared from cells treated with 5 m M Ca z+ during hemolysis and from control cells. The membranes were dissolved in 1 ~ sodium dodecylsulfate in Tris (pH 7.4) and chromatographed on Sepharose 4B as described previously. TABLE I A M I N O ACID ANALYSES O F A G G R E G A T E A N D SPECTRIN Values for aggregate are average of three separate preparations. A single preparation was analyzed for the spectrin data but the data compare well to previous samples analyzed in this laboratory and the results of Marchesi et al. [23]. Tryptophan was not analyzed. No amino sugars were detected. Amino acid

Aggregate (~mole ~ )

Spectrin (~mole ~ )

Lysine Histidine Arginine Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine ½-Cystine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine

6.1 2.5 5.2 8.9 4.9 6.8 16.2 5.2 7.4 7.7 0.5 5.1 2.0 3.4 12.2 2.2 3.8

6.8 2.8 6.1 10.7 4.0 6.0 19.5 2.6 5.9 7.3 0.5 4.9 2.3 3.0 11.6 2.3 3.8

578 an aggregate which has similar electrophoretic and stability properties to that formed in the membrane appears with the disappearance of Component III. This experiment simply shows that Component III can be involved in stable high-molecular-weight aggregates, as has also been shown by organic-solvent treatment of ghosts [15]. It does not necessarily indicate a direct interaction between Ca 2+ and the protein, since the C a 2 + may simply tie up detergent and prevent it from interacting with the protein to keep it disaggregated, nor does it constitute p r o o f of the presence of Component llI in the CaZ+-induced aggregate. Ca z + association with the membrane and aggregate

If Ca z+ is involved in stabilizing the aggregate in the membrane, an enhanced binding of Ca z+ to the membrane might be expected in the membranes containing aggregates. However, if radioactive Ca 2+ is added during hemolysis, the amount which remains bound to the membrane is actually less than that which is bound to isolated

240 20C , , 16C Z 0 ef~

.:

12~

8C 40

I

4

8

I

12

CoCl 2 mM

Fig. 6. Ca2+ binding to erythrocyte membranes. A ( A - - ~ ) 45Ca2+ was added to membranes during hemolysis. B (Q---Q) 45Ca2+ was added to isolated, washed membranes. Details of binding experiments are given in Experimental Procedures. Ca2+ bound is given as/~moles/mg cholesterol. hemoglobin-free membranes, which do not form an aggregate [5] (Fig. 6). The experiment does indicate that Ca 2+ bound to the aggregate does not represent an appreciable fraction of the total Ca 2+ bound to the membrane. Additional support for this was obtained by electrophoresis of 4SCa2+-containing membranes which had been treated with radioactive Ca z+ during hemolysis. After solubilization of the membranes in dodecylsulfate the 4SCa2+ w a s found to migrate about two-thirds of the distance down the gel, even though the aggregate barely moves into the gel (Fig. 7). A sample of 45Ca2+ alone moved about the same distance down the gel, but there was some discrepancy in the positions, probably due to the different amount of detergent available to interact with Ca 2+ in the absence of membrane.

579

Fig. 7. Electrophoresis of 4SCa2+-containing membranes. Gel A ( --) shows distribution of 45Ca2+ in gels from membranes prepared by hemolysis in 5 mM Caz÷ (aggregate forming conditions). Gel B (-- -- --) shows distribution of 4SCa2÷ in dodecylsulfate added directly to the gel.

These experiments clearly indicate that a significant amount of C a 2 + is not involved in stabilizing the aggregate, since the aggregate is still present in dodecylsulfate, but it does not have significant quantities of Ca z+ associated with it. DISCUSSION Addition of Ca 2+ to erythrocytes undergoing hemolysis results in the formation of a protein aggregate which is resistant to disaggregation by sodium dodecylsulfate and a variety of other denaturing agents. The aggregation is sensitive to temperature and Ca z+ concentration and is prevented by chelating agents, if they are added prior to or simultaneously with the Ca 2÷. The aggregate can be isolated by chromatography of solubilized membranes on Sepharose 4B in dodecylsulfate. Amino acid analysis comparisons indicate that it contains spectrin as a major, but not exclusive, component. The aggregate is extremely resistant to disaggregation. Only one identifiable polypeptide (Component III, the membrane-spanning protein) [16] has been discovered in disaggregation attempts. The role of Ca z ÷ in the aggregation has been investigated by examining the binding of Ca 2÷ to membranes during hemolysis and after washing. There is no evidence for an increased binding of Ca 2÷ as a result of aggregation. In addition no significant amount of radioactive Ca 2÷ could be found in the aggregate after its separation from the other membrane proteins by dodecylsulfate acrylamide electrophoresis. Calculations indicated the amount of Ca 2÷ present to be considerably less than 1 atom per spectrin chain. Thus it appears that Ca 2+ may be important in the aggregate formation but not in its stabilization once it is formed. These facts suggest a model for the aggregate formation based on current concepts of erythrocyte membrane-protein organization [17]. The model would depict the Ca 2+ as acting on spectrin to cause an enhanced association of the spectrin molecules. This in turn causes the associated Component III molecules to interact and

580 form a more stable complex with the spectrin. The model assumes a prior interaction of spectrin and Component III which would restrict the movement of the latter in the membrane. This is consistent with the limited mobility of the intramembrane particles, which presumably contain Component III, in the intact erythrocyte [17]. In the ghost the membrane particles are more mobile [18], and Ca 2÷ does not cause protein aggregation. This is probably because spectrin is partially dissociated from the membrane during ghost preparation and washing and no longer possesses the geometrical requirements for the aggregation [16]. The evidence for the involvement of Component III in aggregate formation is still somewhat circumstantial. It is included in this model as a representative of the other protein or proteins which must be present in the aggregate in order to explain the events which might take place during aggregate formation. Another important consideration of the aggregation is the stabilizing force. There is no compelling evidence to support or reject covalent-bond formation. Alternatively, hydrophobic interactions of the polypeptide side-chain residues may play an important role. The ability of sodium dodecylsulfate to break strong protein hydrophobic interactions may be questionable [19, 20]. This problem will have to be studied in more detail. The question of the physiological implications of protein aggregation must also be raised. Since the red cell will become rigid at a Ca 2÷ concentration 10-fold less than required for aggregation, the aggregation process probably does not occur in vivo. However, the course of events that occur in red cells with increasing internal Ca 2÷ concentrations suggest that the process at higher concentrations may be an extension of events that occur during cell rigidification [I ]. The Ca 2+ studies are also of interest as a perturbation method for studying erythrocyte membrane structure-function relationships. Enzyme and protein retention [21] and enzyme activities [22] are all affected by Ca 2÷ concentrations in the range studied here. Clearly further research into other aspects of the influence of Ca 2÷ on the behavior of red-cell membranes is in order. ACKNOWLEDGEMENTS We wish to thank Mr James Wingate for technical assistance with a portion of this work. This research was supported in part by grants from the National Institutes of Health ( G M 16870 and H L 15867), the American Cancer Society (P-563, BC-77 and IN-99B) and by the Oklahoma Agricultural Experiment Station.

REFERENCES 1 La Celle, P. L. (1970) Seminars in Hematol. 7, 355-371 2 La Cello, P. L. (1969) Transfusion 9, 238-245 3 Palek, J., Curby, W. A. and Lionetti, F. J. (1971) Am. J. Physiol. 220, 19-26 4 Palek, J., Curby, W. A. and Lionetti, F. J. (1971) Am. J. Physiol. 220, 1028-1032 5 Triplett, R. B., Wingate, J. M. and Carraway, K. L. (1972) Biochem. Biophys. Res. Commun. 49, 1014--1020 6 Redman, C. M. (1971) J. Cell Biol. 49, 35-48 7 Shin, B. C. and Carraway, K. L. (1973) J. Biol. Chem. 248, 1436-1444 8 Triplett, R. B. and Carraway, K. L. (1972) Biochemistry 11, 2897-2903

581 9 Kobylka, D., Khettry, A., Shin, B. C. and Carraway, K. L. (1972) Arch. Biochem. Biophys. 148, 475-487 10 Carraway, K. L. and Shin, B. C. (1972) J. Biol. Chem. 247, 2102-2108 11 Decker, R. V. and Carraway, K. L., Biochim. Biophys. Acta, in the press 12 Cassim, J. V. and Yang, J. T. (1969) Biochemistry 8, 1947-1951 13 Mitchell, E. D,, Riquetti, P., Loring, R. H. and Carraway, K. L. (1973) Biochim. Biophys. Acta 295, 314-322 14 Visser, L. and Blout, G. R. (1971) Biochemistry 10, 743-752 15 Fairbanks, G., Steck, T. L. and Wallach, D. F. H. (1971) Biochemistry 10, 2606-2617 16 Shin, B. C. and Carraway, K. L. (1974) Biochim. Biophys. Acta 346, 141-153 17 Juliano, R. L. (1973) Biochim. Biophys. Acta 300, 341-378 18 Marchesi, V. T., Tillack, T. W. and Scott, R. E. (1971) in Glycoproteins of Blood Cells and Plasma (Jamieson, G. A. and Greenwalt, T. J., eds), pp. 94-105, J. B. Lippincott Co., Philadelphia 19 Nelson, C. A. (1971) J. Biol. Chem. 246, 3895-3901 20 Knuferman, H., Bhakdi, S., Schmidt-Ulrich, R. and Wallach, D. F. H. (1973) Biochim. Biophys. Acta 330, 356-361 21 Burger, S. D., Fujii, T. and Hanahan, D. J. (1968) Biochemistry 7, 3682-3700 22 Heller, M. and Hanahan, D. J. (1971) Biochim. Biophys. Acta 255, 251-272 23 Marchesi, S. L., Steers, E., Marchesi, V. T. and Tillack, T. W. (1970) Biochemistry 9, 50-57