Spectrin phosphorylation in senescent rat erythrocytes

Mechanisms of Ageing and Development, 22 (1983) 51-70

51

Elsevier Scientific P10qishel,s Ireland Ltd.

SPECTRIN PHOSPHORYLATION IN SENESCENT RAT ERYTHROCYTES*

MARY A. O'CONNELL** SloanoKettering Division, CorneU University, Graduate School of Medical Sciences, New York, N Y 10021 (U.S.A.)

NORBERT I. SWISLOCKI*** Department of Biochemistry, UMDNJ - New Jersey Medical School, Newark, NJ 07103 (U.S.A.)

(Received May 28th, 1982) (Revifion received November 19th, 1982)

SUMMARY The rates of phosphorylation and dephosphorylation of the erythrocyte cytoskeletal protein, spectrin, were analyzed in young and old rat erythrocytes. Endogenous membrane protein ldnase activity was measured in age-separated rat erythrocytes, and was found to decrease as a function of cell age. Membranes prepared from young and old erythrocytes contained comparable levels of protein phosphatase activity. Spectrin phosphatase activity was readily observed in erythrocyte membranes. Partially purified spectrin kinase and spectrin were prepared from membranes obtained from young and old erythrocytes, and the phosphorylation of the spectrin fractions was measured with the isolated kinases. The kinases prepared from young or old ceils phosphorylated spectrin from young cells to the same extent. When spectrin from old cells was used as the substrate, it was phosphorylated ten-fold less extensively by the spectrin kinase prepared from old cells than by the spectrin kinase from young cells. This finding indicated that the decreased phosphorylation of spectrin observed in membranes prepared from age-separated red cells was due to a structural alteration in the spectrin. A structural basis for the decreased phosphorylation of spectrin in older erythrocytes was sought. Treatment of erythrocyte membranes with malonyldialdehyde, a product of lipid peroxidation which accumulates in erythrocyte membranes during senescence, adversely affected spectrin phosphorylation. The results presented here indicate that intramolecular derivatization of spectrin was sufficient to impair its function as a substrate for protein kinase. *This work partially fulfilled the requirement (M.A.O'C.) for the Ph.D. from Sloan Kettering Division, Cornell University, Graduate School of Medical Sciences, New York, NY 10021, U.S.A. **Present address: Department of Biology, University of Virginia, Charlottesville, VA 22903, U.S.A. ***To whom reprint requests flaould be addressed. 0047 -6374/83/$03.00 Printed and Published in Ireland

© 1983 ElsevierScientific Publishers IreLandLtd.

52 Key words: Erythrocytes; Senescence; Phosphorylation; Cytoskeleton; Deformability

INTRODUCTION Senescent red cells are removed from the circulation, after a species-specific number of days, by macrophages in the erythroclastic organs, spleen and liver [1]. The nature of the signal on the surface of effete red cells which the macrophage recognizes is unknown, as is the mechanism for the generation of this signal. Allison [2] observed that the red cell undergoes a fixed number of oxygenation-deoxygenation cycles, approximately l0 s cycles/cell independent of species of origin, which suggested that peroxidative lesions in the red cell membrane might be suspected as causal in the generation of a senescent red cell signal. Jain and Hochstein [3] observed an increased concentration of malonyldialdehyde (MDA), a product of peroxide attack on fatty acids, in membranes obtained from senescent red cells. Pfeffer and Swislocki [4] found that exposure of red cells to peroxide resulted in changes in membrane structure and enzymatic activity similar to changes observed in senescent red cells [5-10]. One of the structural changes noted in senescent red cell membranes is an increase in the amount of a family of high molecular weight proteins with an apparent molecular weight greater than 500 000 [4-6]. These high molecular weight proteins can be formed in vitro by exposure of red cell membranes to MDA [3,4,10], and they are presumed to be the products of MDA cross-links of spectrin components, bands 1 and 2 (nomenclature of Steck [11]). The erythrocyte cytoskeleton is composed of spectrin which is in the form of a heterodimer of protein bands 1 and 2, actin and protein band 4.1 [12,13]. The cytoskeletal network is attached to the membrane through a specific association between spectrin and protein band 2.1 which in turn binds to the transmembrane protein band 3 [14-16]. The cytoskeleton maintains the discoid shape of the circulating erythrocyte [12]. Band 2 of the spectrin complex is multiply phosphorylated [17]. The level of spectrin phosporylation has been correlated with shape changes in red cell ghosts [18,19], but these findings were not corroborated when intact cells were used in similar studies [20]. Pfeffer and Swislocki [4,8] have shown that red cell membrane protein kinase activity in rat red cells declines during senescence. It was not determined in that study whether the decrease in protein kinase activity was due to a defect in the protein kinase or in the endogenous protein acceptors. The experiments detailed here were designed to distinguish between lesions that occurred during in vivo senescence in red cell membrane protein kinases and in their endogenous protein substrates. A further purpose of these studies was to determine whether exposure of red cell membranes to MDA could also alter a functional property of spectrin, to serve as a substrate for the endogenous protein kinase.

53 MATERIALSAND METHODS ATP, histone, casein, and protamine were purchased from Sigma. Acrylamide, bisacrylamide, sodium dodecyl sulfate (SDS), N,N,N',N'-tetramethylmethanediamine, and Triton X-100 were from Bio-Rad. Malonylaldehyde bis-(dimethyl acetal) was from Aldrich. [7-32P]ATP, 20-50 Ci/mmol, was from New England Nuclear. RP-X-Omat film was from Eastman Kodak. All other chemicals used were of reagent grade.

Preparation of rat erythrocyte membranes Rat erythrocyte membranes were prepared essentially as described by Dodge et al. [21]. Male Sprague-Dawley rats were anesthesized with ether and bled through the abdominal aorta. All subsequent procedures were carried out at 4°C. Cells were pelleted by centrifugation for 10 min at 1500 g and the plasma was removed. Care was taken during this procedure not to remove the uppermost red cells while aspirating the white cells, as the upper part of the cell column contains the younger erythrocytes. Reticulocyte contamination of the youngest cell population was estimated to be 10% and less than 1% of the oldest cell population. The cells were washed three times in 30 volumes of 0.167 M Tris-HC1 pH 7.4 containing 02% glucose. The cells were lysed by resuspension in 30 volumes of ten times diluted Tris buffer containing 1 mM EDTA, and centrifuged for 15 min at 25 000 g. The supernate was aspirated, the membrane pellets were transferred to clean centrifuge tubes, and the above wash was repeated twice more. Any hardpacked cell debris or precipitated hemoglobin under the membrane pellet was removed by rotating the tube and aspirating the underlying debris.

Age separation of rat erythrocytes Separation of rat erythrocytes according to cell age was conducted by the procedure of Murphy [22], which is based on the finding that as red cells age they become denser.

Preparation of spectrin Spectrin was eluted from rat erythrocyte membranes by the method of Bennett and Branton [23]. Membranes (4 mg/ml) were washed with 30 volumes of 0.3 mM sodium phosphate pH 7.4. One volume of membranes was incubated at 37°C for 25 min with 1 w~lume of the buffer. After incubation, the membranes were centrifuged for 15 min at 4°C at 198 700 g. The supernate contained the eluted spectrin.

Preparation of spectrin kinase Spectrin kinase was extracted by the method of Hosey and Tao [24]. Rat erythrocyte membranes (4 mg/ml) were incubated with an equal volume of 0.5 M NaCI for 10 min at O°C. The membranes were centrifuged for 30 min at 17 000 g. The supernate was harvested and the pellet was re-extracted and centrifuged. The supemates were pooled and dialyzed overnight against 10 mM Tris-HC1 pH 7.0, 10 mM magnesium acetate, and 4 mM /3-mercaptoethanol. The dialyzed supernate contained the spectrin kinase.

54 The pellet was resuspended in the dialysis buffer and assayed for residual protein kinase activity. Solubilization o f rat erythrocyte membranes with Triton X-I O0 Detergent solubilization was by the method of Grant et al. [25]. Rat erythrocyte membranes (4 mg/ml) were mixed with an equal volume of 0.4% (w/v) Triton X-100 and 8 mM/3-mercaptoethanol at 4°C. The samples were centrifuged at 198 700 g for 30 min and the supernates and pellets harvested. MDA incubations MDA was generated and detected by the method of Mengel [26]. Malonylaldehyde bis-(dimethyl acetal) was mixed with an equal volume of 1 mM HC1 to yield a 3.05 M MDA solution. One milliliter of membranes at 4 mg/ml was incubated with 2 ml of 0.017 M Tris-HC1 pH 7A, 1 mM EDTA, and the indicated concentration of MDA. After incubation the membranes were washed with 0.017 M Tris-HC1, pH 7.4, 1 mM EDTA and MDA determined. Samples containing 0 - 2 0 nmol of MDA were brought to 1 ml with water; 1.2 ml of 0.67% (w/v) thiobarbituric acid was added and the tubes were capped with a marble and placed in a boiling water bath for 15 rain. After cooling, the absorbance at 535 nm was determined; 2 nmol of MDA absorbed 0.1 units. Protein determination Protein was determined by the method of Lowry et al. [27] as modified by Dulley and Grieve [28]. Bovine serum albumin was used as the standard. SDS-polyaerylamide gel electrophoresis SDS-gel electrophoresis was performed essentially as described by Laemmli [29]. Following electrophoresis, gels were rinsed with distilled water and fixed in 10% (w/v) trichloroacetic acid for 30 min. Protein staining was accomplished as described by Kahn and Rubin [30] by immersing the gel in 300 ml of 0.1% (w/v) Coomassie Brilliant Blue R-250 in 25% (w/v) trichloroacetic acid for 1 h at 37°C. Destaining of the gels was done with several changes of 7% (v/v) acetic acid until the background gel was clear. For autoradiography, after destaining, the slab gel was soaked in 10% (v/v) glycerol for 1 h before drying under vacuum on a Bio-Rad dryer. The dried gel was sandwiched with 1-3 sheets of RP X-Omat film between two Cronex intensifying screens in an X-ray cassette. The film was exposed for 1-10 days at --20°C and developed in a Kodak X-ray film devel-

oper. Protein kinase assay Protein kinase was assayed by a combination of the methods of Grant et al. [25 ] and Delange et al. [31]. Membrane protein (100/~g) was incubated at a final concentration of 50 mM Tris-HC1 pH 7.0, 4 mM $-mercaptoethanol, 10 mM magnesium acetate, at 37°C. The reaction was started by the addition of 10 nmol of diTris ATP, containing

55 0.1 /aCi of [7-32p]ATP. The final reaction volume was 200 pl. Samples were incubated for 5 min. The reaction was stopped by the sequential addition of 1 ml of cold 0.2% (w/v) bovine serum albumin and 1 ml of cold 10% (w/v) trichloroacetic acid. Samples were kept on ice for 5 - 1 0 min and were then centrifuged for 10 rain at 1300 rpm in a Beckman TJ6 refrigerated centrifuge. Supernates were aspirated and the pellets dissolved in 1 ml of 0.1 N NaOH. The samples were reprecipitated with 1 ml of 10% trichloroacetic acid, centrifuged and the pellets were washed with 5 ml of 5% trichloroacetic acid. After recentrifugation the pellets were dissolved with 2 ml of 0.1 N NaOH, transferred to minivials and the Cerenkov radiation measured in a Packard scintillation counter. Boiled membrane blanks routinely gave values of 2 pmol of phosphate incorporated per sample. Casein, histone or protamine if present were added so that each assay contained 300 p.g of the protein. Preliminary studies showed that the phosphorylation of membrane protein was linear with respect to time for up to 5 min and linear with respect to protein up to 100 /ag. When nonmembranous sources of enzyme were used the final concentration of salts and reactants was as described above and in a final volume of 200/J1.

Protein phosphatase assay A 150-/.d volume of membrane at 4 mg/ml was incubated at 37°C with 150 gl of 100 mM Tris-HCl pH 7.0, 20 mM magnesium acetate, 8 mM /~-mercaptoethanol and 5/JCi of [7-a2P] ATP. After labeling for 5 min, 60/~1 of 300/~M ATP were added; samples were mixed and 60/J1 were removed and combined with 60 gl of 1% SDS, 1% ~mercaptoethanol, 1% glycerol, 0.05 M Tris-HCl pH 6.8, in a microfuge tube. This was called the zero time point, and the incubation was continued with 60-pl samples being withdrawn at 5, 10, 20 and 30 min. The samples were boiled for 2 rain and cooled; then the entire contents of the microfuge tube were loaded into a well of a polyacrylamide slab gel. After electrophoresis the gel was fixed in 10% trichloroacetic acid for 30 min, rinsed with distilled water, and wrapped in Saran Wrap. The gel was loaded in an X-ray cassette with a sheet of X-ray film and exposed for 18 h. The film was developed and the regions of the gel corresponding to bands on the autoradiograph were cut out, and the Cerenkov radiation determined.

EXPERIMENTALRESULTS

Fractionation of membrane-associated protein kinase activities Prior to an analysis of the effect of senescence on membrane protein phosphorylation, the number and nature of the protein kinase activities present in the rat red cell membrane were examined. Protein kinase activity was measured in red cell membranes, in membrane proteins extracted by 0.5 M NaCl, and in membrane proteins rendered soluble by 0.1% Triton X-IO0. These fractionation procedures have been used with human red cell membranes to obtain various protein kinase activities [2425, 32-34]. The three

56 protein kinase sources prepared here were assayed for their ability to phosphorylate several exogenous protein substrates. The results of these measurements are shown in Table I. The extract obtained with 0.5 M NaC1 contained phosphorylating activity from the red cell membrane which utilized casein and spectrin more efficiently than either protamine or boiled red cell membranes. The salt-extractable material displayed no autophosphorylating activity. Triton X-100 treatment o f the red cell membrane generated a soluble phosphorylating activity which utilized protamine more efficiently than any of the other substrates. 30% of the endogenous activity expressed in the intact membrane was rendered soluble by the Triton X-100 treatment. 60% of the membrane protein kinase activity was recovered in the insoluble fraction (340 pmol Pi per mg enzyme per 5 min). 30% (167 pmol Pi per mg enzyme per 5 min) o f the endogenous protein kinase activity exhibited by the intact membrane was resistant to extraction with salt. In contrast to the salt-extractable or detergent-soluble activities, the intact membrane did not phosphorylate any o f the exogenous substrates well. In fact, casein or spectrin, when added to assays containing intact membranes as the source of the kinase, inhibited the measured protein kinase activity by 1 0 - 1 5 % . While any single treatment of the membrane did not increase the specific activity of endogenous protein kinase activity, it is apparent that the Triton X-100 soluble material has an increased specific activity towards protamine, 327 pmol vs. 100 pmol for

TABLE I PROTEIN ACCEPTOR PREFERENCES OF RED CELL MEMBRANE PROTEIN KINASE ACTIVITIES Protein kinase activity was assayed on 80 #g of membrane protein, 25 ~g of protein of the 0.5 M NaC1 supernate, and 50 #g of the Triton X-100 soluble protein, in the presence of the following protein acceptors: 300 ~tg of casein, 300 #g of protamine, 50 #g of spectrin, and 300 #g of boiled red cell membrane. Protein kinase assays were conducted as described in Methods. Enzyme activities were corrected for endogenous protein kinase activities and the results are expressed as pmol Pi transferred to the indicated protein acceptor per mg enzyme source per min. Activities are shown as the mean +-standard deviation of triplicate values. Protein

Noaddition Casein Protamine B~RBC-Gb Spectrin

Protein kinase source Intact membrane

0.5 M NaCI supernate

Triton X-I O0 supernate

570 +-24 N.A.a 71 ± 39 136 ± 30 N.A.a

N.A.a 347 ~ 18 238 -+ 9 180 ± 12 320 ± 13

180 108 327 82 9

aN.A. = no activity. bB-RBC43 = boiled red blood cell ghosts.

-+ 7 +- 6 +- 19 -+ 10 ± 21

57 casein or 49 pmol for spectrin. The salt-extractable material has an increased specific activity towards casein and spectrin, 347 and 320 pmol vs. 238 pmol for protamine. The salt-extractable activity will be referred to as the spectrin kinase. These results suggest that, once dissociated from the membrane, the spectrin kinase can utilize a variety of protein acceptors. A corollary to this observation is that, while membrane-associated, the kinases are probably bound to their respective protein substrates. The rat red cell membrane was observed to possess at least two protein kinases, which behaved in an analogous manner to the enzymes present in human red cells. The detergent-soluble protein kinase present in the human red cell membrane is stimulated by cyclic AMP (cAMP) [32,33]. The detergent-soluble protein kinase in the rat red cell membrane was not activated by cAMP (data not shown). This insensitivity of the rat red cell membrane enzyme to cAMP has been reported by Pfeffer and Swislocki [8]. Endogenous protein kinase substrates The nature of the products of the endogeous phosphorylation reaction of the rat red ceil membrane was examined. Membranes were phosphorylated in the presence of [7-3~P]ATP, electrophoresed, stained and an autoradiogram was developed as described in the Methods section. The resulting Coomassie Brilliant Blue stained gel and its autoradiogram are shown in Fig. 1. The proteins were numbered after the scheme of Steck [11 ]. Band 2, the smaller component of spectrin, was the major product of the endogenous phosphorylation reaction in the intact membrane. Bands 2.9, 3.1, and a protein with an apparent molecular weight of 40 000 also served as substrates for protein kinases in the intact membrane. Effect of cell age on membrane protein kinase activity Red cells were separated into five fractions of increasing density, corresponding to fractions of cells of increasing age. Membranes were prepared from these cells and protein kinase activity was measured in the presence and absence of an exogenous substrate, histone. Histone was chosen as the exogenous acceptor in these experiments since it was the only reported kinase substrate which was readily utilized by the rat red cell membrane kinase [25]. The histone.specific phosphorylation was calculated as the difference in activity between samples incubated with and without the exogenous substrate. The results are shown in Table II. Both the endogenous and the histone-specific phosphorylating activities declined as a function of cell age. The endogenous protein kinase activity expressed by the oldest cell fraction was 40% of the protein kinase activity in the youngest cell fraction. Furthermore, in the youngest cell fraction, phosphorylated histone represents 36% of the total phosphorylated protein formed while in the oldest cell fraction, phosphorylated histone represents 50% of the total phosphoprotein formed. These results indicate that two separate processes are occurring during red cell senescence which result in decreased membrane protein phosphorylation. Firstly, the specific activity of the membrane pro-

58

1 2

2

2.9 3.1

3 4.5

5 Vlr 4 0 , 0 0 0 6

1

2

Fig. 1. Endogenous phosphorylation in rat red cell membranes. 100 ~g of rat red cell membrane were incubated for 5 min at 37°C with 1 ~Ci of [3,-32P]ATP in 50 mM Tris-HC1 pH 7.0, 10 mM magnesium acetate, and 4 mM t3-mercaptoethanol. After phosphorylation the sample was loaded on a 7.5% polyacrylamide slab gel and electrophoresed. The stained and dried gel was exposed to Kodak X-Omat film for 7 days. Lane 1 is a photograph of the Coomassie Brilliant Blue stained gel, lane 2 is the autoradiogram of lane 1.

tein kinases is decreased, and secondly, the ability of the remaining active protein kinases to recognize their endogenous protein acceptors is impaired. These results suggest that the concentration of functional endogenous protein acceptors is reduced in senescent erythrocytes.

59 TABLE II HISTONE PROTEIN KINASE ACTIVITY IN AGE-SEPARATED RAT ERYTHROCYTES Membranes were prepared from age-separated rat erythrocytes, Protein kinase activity was measured (see Methods) on 80 tag of membrane protein in the presence or absence of 250 ~tg of histone. Enzyme activity was expressed as pmol Pi transferred per mg membrane protein per 5 rain. Histone-specffic phosphorylation is presented as the difference in activity between samples incubated with and without the exogenous substrate. Activities are reported as the mean -+standard deviation of quadruplicate values.

Membranes

Young 1 2 3 4 Old 5

Enzyme activity Endogenous

Histone

1043 -+54 938 -+24 703 -+80 620 -+19 388 -+46

594 +-56 379 -+47 321 -+82 173 +-24 344 -2_43

Endogenous membrane protein phosphorylation in age-separated red cells The decrease in total membrane protein kinase activity measured as a function of red cell age might have been at the expense of only one of the phosphorylated protein products. To determine if one of the phosphorylation reactions was singularly affected by senescence, autoradiograms of SDS-polyacrylamide gels which contained membrane proteins phosphorylated with [7-3~P]ATP were prepared. Membranes were obtained from age-separated red cells and phosphorylated at two different ATP concentrations, 50/aM (Fig. 2) and 0.5 tAVl(Fig. 3). The decrease in endogenous protein kinase activity as a function of cell age, reported in Table II, is reflected in the intensity o f the bands seen in the autoradiogram in Fig. 2. The level of phosphorylation o f the phosphorylated proteins decreased as a function o f age. When phosphorylation reactions were conducted at a lower ATP concentration (0.5 taM), there was no apparent decrease in the extent of protein phosphorylation seen as a function of cell age (Fig. 3). The same spectrum of proteins was phosphorylated and in the same ratio as was observed at higher ATP concentrations. No particular endogenous phosphorylation reactions appear to be specifically impaired as a consequence o f cell aging. At very low ATP concentrations the reaction rates are undoubtedly limited by ATP, and all the available ATP is utilized. When the respective reaction rates are allowed to reach their maximum by increasing the ATP concentration, the effects o f aging on protein phosphorylation become apparent. Since the KIn,ATP of the spectrin kinase does not change with red cell age (see below), the reduced reaction rate seen in membranes from older cells likely reflects a decreased ability to form effective spectrin-spectrin kinase complexes.

60

2

2.9 3.1

40 K

1

2

3

4

Fig. 2. Endogenous protein phosphorylation at 50 t~M ATP, in membranes obtained from age-separated rat erythrocytes. 100 t~g of membrane protein from age-separated rat erythrocytes were phosphorylated for 5 rain at 37°C with 1 #Ci of [3,-32P]ATP (50 tzM ATP) in 50 mM Tris-HC1 pH. 7.0, 10 mM magnesium acetate and 4 mM fl-mercaptoethanol. Samples were electrophoresed on 7.5% polyacrylamide gels, dried and exposed to Kodak X-Omat film for 14 days. Lane 1 contained membranes from the youngest cells and lane 4 contained membranes from the oldest cells.

Initial velocity analysis of spectrin kinase from young and old cells Protein kinase assays were conducted with the spectrin kinase extracted from membranes of young and old red cells in the presence o f varying concentrations of ATP ( 1 - 2 0 /~M) and a saturating concentration of casein (1.5 mg/ml). The concentration o f

61

2 2.9 3.1

Mr 4 0 , 0 0 0

1

2

3

4

5

Fig. 3. Endogenous phosphorylation, at 0.5 ~M ATP, in membranes obtained from age-separated rat erythrocytes. 100 /zg of membrane protein from age-separated rat erythrocytes were phosphorylated for 5 min at 370C with 1 /zCi of ['y-32P]ATP(0.5 ~,M ATP) in 50 mM Tris-HC1 pH 7.0, 10 mM magnesium acetate, and 4 mM #-mercaptoethanol. Samples were electrophoresed on 7.5% polyacrylamide gels, dried and exposed to Kodak X-Omat film for 7 days. Lane 1 contained membranes from the youngest calls and lane 5 contained membrane from the oldest ceils.

ATP in the reaction mixture at the onset of incubation and after 5 min was monitored by chromatography o f nucleotides on polyethyleneimine thin-layer plates. After 5 min o f incubation, the concentration of ATP was observed to be at least 95% of the initial concentration (data not shown). Hots of the inverse o f the reaction velocity as a function of the reciprocal concentration o f ATP gave straight lines (Fig. 4). It can be seen graphically that the Km~TP does not change as a function of cell age. The calculated values for K m and Vmax for the spectrin kinase extracted from membranes prepared from young cells were 7.2 + 2.4 /zM and 260 --- 10 pmol Pi per rain per mg, and from old cells were 8.6 + 1.7/aM and 180 + 18 pmol Pi per min per mg. There is a 30% decrease in Vmax as the red cells age. This decrease in Vmax probably reflects the lowered concentration of active spectrin

62

.06 -

OLD IO) E

,_

- .04

E o;

NG

0

.02 O.

~"

.ol

-

S

i

i

J

I

,05

1.0 1/ATP

IJM"1

Fig. 4. Initial velocity of spectrin kinase v e r s u s ATP concen~ation. Spectrin kinase was extracted from membranes prepared from young and old erythrocytes as described in the Methods, and I0/~g of spectrin kinase were assayed for protein kinase activity in the presence of the listed ATP concentrations for 5 min at 37°C with 1.5 mg/ml casein as the substrate.

kinase in the extract of membranes from older ceils rather than a structural alteration in the active site of the enzyme (see below).

Effect of cell age on spectrin phosphorylation The effect of senescence on the major membrane protein kinase reaction of erythrocytes was examined in greater detail. Spectrin kinase and spectrin were extracted from membranes prepared from young and old erythrocytes. The spectrin fractions extracted from the young and old cells contained primarily bands 1 and 2 with small amounts of band 5. All the other proteins present constituted less than 1% of the total protein as judged by Coomassie Brilliant blue staining of polyacrylamide gels of the spectrin preparations. The same amount of spectrin was extracted from equivalent amounts of young and old red cells. The specific activity of the spectrin kinases prepared from membranes obtained from young and old ceils when assayed with casein as the substrate were 2500 and 1700 pmol Pi per mg enzyme per 5 min, respectively. The ability of both young and old preparations of spectrin kinase to phosphorylate spectrin fractions obtained from young and old cells was measured, and the results of these assays are shown in Table III. Spectrin prepared from young cells was phosphorylated to a similar extent by spectrin kinases obtained from either young or old cells. Both kinases could detect a difference in spectrin prepared from old cells, however, and phosphorylated this substrate less

63 TABLE III PHOSPHORYLATION OF SPECTRIN BY SPECTRIN KINASE E X T R A C T E D FROM MEMBRANES OBTAINED FROM Y O U N G A N D OLD R E D CELLS Spectrin kinase was extracted from m e m b r a n e s obtained from y o u n g and old red ceils as described in Methods. Spectrin was also prepared from similar m e m b r a n e preparations as described in Methods. An e n z y m e unit was arbitrarily defined as the a m o u n t o f spectrin kinase which transfers 50 pmol of phosphate to casein in 5 min at 37°C. 20 t~g of e n z y m e from old cells contain 0.66 units, 20 t~g from y o u n g ceils contain 1.02 units. Protein kinase assays contained 20 t~g of spectrin kinase and 35 ug of spectrin. The assay was conducted as described in Methods. E n z y m e activity was expressed as pmol Pi transferred to spectrin per e n z y m e unit. Results are shown as the mean ± standard deviation of triplicate values.

Spectrin

Young Old

Spectrin kinase Young

Old

4.0 -+ 1.0 2.3 -+ 0.4

4.4 -+ 0.5 0.4 -+ 1.3

efficiently than spectrin from young ceils. This discrimination was greater with the preparation o f kinase made from old cells which phosphorylated spectrin from young cells ten times more readily than that obtained from the older preparation. These results suggest that, in the spectrin-spectrin kinase reaction, the decreased phosphorylation observed in membranes prepared from older cells is due to a lesion in spectrin which effectively decreases the concentration of functional substrate.

Membrane protein phosphatase activity Membranes from young and old erythrocytes were prepared and assayed for protein phosphatase activity in a pulse-chase type experiment as described in the Methods section. The hydrolysis of phosphate from three membrane proteins as a function of time is shown in Fig. 5. Membranes prepared from young and old cells possess equivalent levels of protein phosphatase activities. The incorporation of phosphate into these proteins is at sites which turn over with half-lives of 5 - 1 0 min. Virtually all of the phosphate incorporated into spectrin can be hydrolyzed by resident membrane phosphatases. The turnover of phosphate on protein Mr 40 000 was found on subsequent experiments to be much more rapid, equilibrium of on-and-off rates had already been established before the start of the cold ATP chase. These results rule out the possibility that the decreased membrane protein kinase activity observed in senescent erythrocytes was due to a reduction in protein phosphatase activity. Effect o f MDA exposure on membrane protein structure Red cell membranes were exposed to MDA and the amount of MDA associated with washed membranes was determined. After 1 h incubation at 37°C, membranes incubated

64

I00~% 751-

SPECTRIN

251-

.......

G}N2

75k

5

e-.-._..__~

~

Mr40,000

~'%.

50t

25

l

5

I

I0

time

I

20

,

'

30

(minutes)

Fig. 5. Endogenous protein phosphatase activity in membranes prepared from young and old rat erythrocytes. Protein phosphatase activity was measured in membranes prepared from age-separated red cells. 150 ~1 of membrane proteins (4 mg/ml) were phosphorylated for 5 min with [7-32P]ATP at 37°C after addition to 150 /zl of 100 mM Tris-HC1 pH 7.0, 20 mM magnesium acetate, 8 mM p-mercaptoethanol. After labeling, samples were removed and combined with a stopping solution containing 1% SDS, 1% 13-mercaptoethanol, 1% glycerol in 50 td of 50 mM Tris-HC1 pH 6.8. Samples were taken at 5,10, 20 and 30 min, boiled for 2 min, cooled and applied onto gels for electtophoresis. After fixation and autoradiograpliy gel sections corresponding to bands were cut out and Cerenkov radiation determined. Solid line represents percentage phosphate remaining in proteins listed in membranes obtained from young cells, dashed line refers to old cells.

65 with 0, 10, 50 or 100 mM MDA had accumulated 1.4, 35, 59, and 60 nmol o f MDA per mg of membrane protein, respectively. The production of high molecular weight protein aggregates (>500 000) in membranes exposed to MDA was observed in polyacrylamide gels. The amount of high molecular weight aggregate formed was dependent on the time o f exposure, the concentration o f MDA, and the temperature at which the membranes were incubated with MDA. Maximal production of the high molecular weight protein complexes was evident at 100 mM MDA for incubations o f 30 min at 2 5 - 3 7 ° C . Shorter times, lower temperatures, or lower MDA concentrations resulted in decreased production o f crosslinked proteins. While the high molecular weight protein aggregates were easily and reproducibly observed by eye in Coomassie-stained gels, they were present in such low concentrations relative to the other membrane proteins that it was not possible to quantitate the amount o f aggregate accurately using spectrophotometric scans o f the gel. The high molecular weight proteins observed in these gels resembled the crosslinked products reported by Pfeffer and Swislocki [4], Kadlubowski [10], and Jain and Hochstein [3], all working in similar red blood cell systems.

Effect o f MDA exposure on membrane protein kinase activity Red cell membrane protein kinase activity was measured in membranes that had been incubated with 0, 50, or 100 mM MDA for 30 min at 37°C. Protein kinase activity was determined in the presence and absence of two exogenous substrates, protamine and histone (Table IV) The ability o f membrane protein kinase to phosphorylate endogenous membrane proteins declined in membranes that had been exposed to increasing

TABLE IV EFFECT OF MDA TREATMENT OF RED CELL MEMBRANES ON MEMBRANE PROTEIN KINASE ACTIVITY A 1-ml volume of red cell membranes was incubated with the indicated concentrations of MDA for 30 min at 37°C. Membranes were washed and protein kinase activity measured (see Methods). 80 #g of membrane protein were assayed in the presence and absence of 300 #g of protamine or 250 ~ag of histone. Endogenous protein kinase activity was that activity measured in the absence of added protein acceptor. Protamine or histone-specific protein kinase activity was calculated as the difference between assays with and without exogenous acceptor. Protein kinase activity was expressed as pmol Pi tramferred per mg membrane per 5 min. Results of two separate experiments are reported and the activity is shown as the mean + standard deviation of triplicate values.

MDA {raM)

0 50 100

Experiment I

Experiment H

Endogenous

Protamine

Endogenous

Histone

443 ± 33 415 ± 17 313 ± 20

0 141 ± 27 183 ± 28

525 ± 19 519 ± 35 439 ± 20

142 -+ 37 284 + 39 292 ± 22

66 concentrations of MDA, In contrast, the activity of the protein kinase towards the exogenous acceptors increased in membranes exposed to increasing concentrations of MDA. These results indicated that MDA affected red cell membrane phosphorylation by altering the suitability of the endogenous substrates to serve as phosphate acceptors in the membrane.

Effect of MDA exposure on spectrin phosphorylation The following experiment was designed to quantitate the decrease in phosphorylation of one of the membrane proteins, spectrin, which is presumed to be preferentially affect. ed by MDA [10]. Membranes which had been incubated with MDA were phosphorylated by endogenous kinases in the presence of [TP2P]ATP. The samples were then electrophoresed on gels, the regions of the gel containing spectrin were excised and the Cerenkov radiation determined. The a2p radioactivity associated with spectrin was expressed as a percentage of control membranes, and the results are shown in Table V. Treatment o f red cell membranes with MDA resulted in a 44% decrease in the phosphorylation of spectrin. This level of inhibition was observed in membranes after treatment with 100 mM MDA; 50 mM MDA resulted in a 12% decrease in spectrin phosphorylation. The level of inhibition of total phosphorylation by treatment with 100 mM MDA was only 1 5 - 3 0 % (Table IV). While the effects of MDA on the phosphorylation of other individual membrane proteins were not determined, it would appear that MDA treatment decreases total membrane protein phosphorylation primarily by affecting spectrin phosphorylation. More importantly, it should be noted that these effects were seen with spectrin molecules whose mobility had not been altered on SDS-polyacrylamide gels. The decrease in spectrin phosphorylation after exposure to MDA was measured in spectrin molecules which had not been intermolecularly crosslinked.

TABLE V EFFECT OF MDA TREATMENT OF RED CELL MEMBRANES ON SPECTRIN PHOSPHORYLATION A 1-ml volume of red cell membranes was incubated for 30 min at 37°C in the presence of the listed MDA concentrations. 100 mg of membrane protein were phosphorylated for 2 min at 37°C with 1 tzCi of [-r-32P]ATP,in 50 mM Tris-HCl pH 7.4, 10 mM magnesium acetate, and 4 mM ~-mercaptoethanol. The samples were boiled for 2 min in SDS-gel buffer, loaded on 7.5% polyacrylamide slab gels and electrophoresed. The region of the gel containing spectrin was cut out and the Cerenkov radiation counted. Results are expressed as percentage of control + standard deviation of triplicate values.

mM MDA

Percentage of 32P.cpm in spectrin

0

100 ±0

50 100

88 + 1

56 -+8

67 DISCUSSION AND CONCLUSIONS The cytoskeleton of the erythrocyte membrane maintains the biconcave discoid shape of the erythrocyte [12]. This shape allows the cell to deform readily so that it can pass through capillaries whose diameter is narrower than the cell's. Senescent erythrocytes are less deformable than young cells [35], and it is assumed that decreased deformability contributes to the sequestration of the senescent erythrocytes by the spleen. The data presented here clearly show that a functional property of the cytoskeleton, the extent of spectrin phosphorylation, is impaired during senescence. A mechanism for the generation in vivo of defective spectrin molecules was proposed and investigated. The extent of membrane protein phosphorylation was observed to decline in a progressive fashion as a function of red cell age (Table II). Two possible explanations for this decrease in phosphorylation were presented: (1) the concentration of protein substrate decreased in older cells, or (2) the concentration of active protein kinase decreased. The concentration of ATP has been reported to remain unchanged throughout the lifetime of the red cell [36]. To distinguish between these alternative explanations, it was necessary to focus further experiments on a single phosphorylation reaction. The effect of red cell aging on spectrin phosphorylation was examined, and the results indicated that the primary cause for the decrease in spectrin phosphorylation that was observed when membranes were phospho,rylated (Fig. 2), was a decreased concentration of functional spectrin (Table III)o There was a 65% decrease in the specific activity of spectrin kinase prepared from older erythrocytes, as compared with younger cells. One might argue that the decreasing amount of active spectrin kinase present in older cells was responsible for the decline in spectrin phosphorylation. Since erythrocytes are incapable of protein synthesis, they cannot replace damaged spectrin or spectrin kinase. However, unless spectrin kinase is operating very close to its maximal turnover rate, the active enzymes in the older erythrocytes could be able to maintain the required phosphorylation level of the membrane. Furthermore, the magnitude of the impairment on the part of the spectrin component, a ten-fold decrease in phosphorylation (Table III), indicated that the primary cause for the decline in spectrin phosphorylation in senescent red ceils was due to the accumulation of defective spectrin. The effect of MDA exposure on membrane phosphorylation reactions was investigated since this natural product of lipid peroxidation has been reported to accumulate during erythrocyte senescence [3]. MDA can act as a crosslinking agent reacting with free amino groups on lipids and proteins. Of the spectrum of products formed in reactions of this type, intermolecular crosslinks are the least prevalent, the majority of the added bifunctional MDA reacts at only one end to derivatize the available free amino groups. Since the parameter used to optimize the conditions of MDA exposure was intermolecular crosslinking, all of the available free amino groups in the membrane were undoubtedly reacted with MDA. Reaction of the e-amino group of lysine with MDA would reduce the positive charge on a protein. If charge plays any part in the binding of the

68 spectrin kinase to its substrate, MDA derivativization of the lysine residues in band 2 could prevent enzyme-substrate interaction. The effect of MDA on spectrin phosphorylation was determined by analyzing the incorporation of phosphate into band 2 which had been located on SDS-polyacrylamide gels, The analyzed spectrins could not have been crosslinked by virtue of their electrophoretic mobility in SDS-gels. The MDA reacted spectrin was shown to be an unaccept. able substrate for spectrin kinase. These alterations in spectrin structure which were not detected by gel-electrophoretic techniques were easily recognized by the spectrin kinase. The endogenous production of MDA upon lipid peroxidation during the course of an erythrocyte's lifetime is a plausible mechanism for the accumulation of damaged spectrin molecules. It has recently been reported that the levels of spectrin phosphorylation do not correlate with cell shape changes in a causal fashion [20], nor is phosphorylated spectrin required to polymerize actin [37]. These recent reports cast serious doubt on the previously prevailing assumption that there was a causal relationship between spectrin phosphate levels and cell shape changes [18,19]. The hypothesis that the dephosphorylation of spectrin is of regulatory interest has been implied but not clearly tested [38]. The full significance of the age-related decline in spectrin phosphorylation will not be appreciated until the effects of cytoskeletal phosphorylation reactions on red cell function are fully elucidated. The function of the cytoskeleton and the role of spectrin in modulating surface properties of erythrocytes is not fully understood. At this juncture it is tempting to speculate that rearrangements in cytoskeletal architecture alter the topographic distribution of transmembrane proteins, such as glycophorin. Age-related changes in availability of surface antigens may in turn provide the appropriate signal for erythrophagocytosis. A mechanism which implicates the appearance of an antigen in the removal of senescent cells from the circulation has been described by Kay [39]. We suggest that the expression of antigenic sites may be modulated by alterations in cytoskeletal components which become less reactive as their suitability for covalent modification is reduced during aging [40]. ACKNOWLEDGEMENTS We would like to thank Drs. Jack Kyte and Earlene Cunningham for helpful discussions, Dr. Jon Richards for the statistical analysis of the kinetic data, and Ms. Joan Tierney for technical advice. This research was supported by NIH grant AM23040. REFERENCES 1 J.W. Harris and R.W. Kellermeyer, The Red Cell, Harvard Univerity Press, Cambridge, MA, 1974. 2 A.C. Allison, Turnovers of erythrocytes and plasma proteins in rnammala. Nature, 188 (1960) 37-40. 3 S.L. Jain and P. Hoch~tein, Polymerization of membrane componenta in aging red blood cell~. Biochem. Biophys. Res Commun., 92 (1980) 247-254. 4 S.R. Pfeffer and N.I. Swislocki, Role of peroxidation in erythrocyte aging.Mech. Ageing Dev., 18 (1982) 355-367.

69 5 R.G. Chapman and L. Schaumberg, Glycolysis and glycolytic enzyme activity ofagingred cells in man. Changes in hexokinase, aldolase, gtyceraldehyde-3-phosphate dehydrogenase, pyruvate kinase and glutamic-oxalacetic transaminase. Br. Z Haematol., 13 (1967) 665 -678. 6 N.S. Cohen, J,E. Ekholm, M.G. Luthra and D.J. Hanahan, Biochemical characterization of densityseparated human erythrocytes. Biochim. Biophys. Acta, 419 (1976) 229-242. 7 M. Kadlubowski and P.S. Aggutter, Changes in the activities of some membrane-associated enzymes duringin vivo aging of normal human erythrocyte. Br. J. Haematol., 37 (1977) 111-125. 8 S,R. Pfeffer and N.I. Swislocki, Age-related decline in the activities of erythrocyte membrane adenylate cyclase and protein kinase. Arch. Biochem. Biophys., 177 (1976) 117-122. 9 M. Kadlubowski, The effect of in vivo aging of the human erythrocyte on the protein of the plasma membrane. A characterization. Int. J. Biochem., 9 (1978) 67-78. 10 M. Kadlubowski, The effect of in vivo aging of the human erythrocyte on the proteins of the plasma membrane. A comparison with metabolic depletion and blood bank storage, lnt. Z Biochem., 9 (1978) 79-88. 11 T.L. Steck, The organization of proteins in the human red blood cell membrane../. CellBiol., 62 (1974) 1-19. 12 G.B. Ralston, The structure of spectrin and the shape of the red blood cell. Trends Biol. Sci., 3 (1978) 195-198. 13 V.T. Marchesi, Functional proteins of the human red blood cell membrane. Semin. Hematol., 16 (1979) 3-20. 14 M.P. Sheetz, Integral membrane protein interaction with triton cytoskeletons of erythrocytes. Biochim. l~'ophys. Acta, 557 (1979) 122-134. 15 J. Yu and S.R. Goodman, Syndeins: The spectrin-binding protein(s) of the human erythrocyte membrane. Proc. Natl. A cad. ScL U.S.A., 76 (1979) 2340-2344. 16 V. Bennett and P.J. Stenbuck, The membrane attachment protein for spectrin is associated with band 3 in human erythrocyte membranes. Nature, 280 (1979) 468-473. 17 L.C. Wolfe and S.E. Lux, Membrane protein phosphorylation of intact normal and hereditary spherocytic erythrocytes. Z Biol. Chem., 253 (1978) 3336-3342. 18 M.P. Sheetz and S.J. Singer, On the mechanism of ATP-induced shape changes in human erythrocyte membranes. I. The role of the spectrin complex. J. Cell Biol., 73 (1977) 638-646. 19 W. Birchmeier and S.J. Singer, On the mechanism of ATP-induced shape changes in human erythrocyte membranes. II. The role of ATP. J. Cell Biol., 73 (1977) 647-659. 20 J.Mo Anderson and J.M. Tyler, State of spectrin phosphorylation does not affect membrane shape or spectrin binding to erythrocyte membranes. J. Biol. Chem., 255 (1980) 1259-1265. 21 J.T. Dodge, C. Mitchell and D.J, Hanahan, The preparation and chemical characteristics of hemoglobin-free ghosts of human erythrocyte$. Arch. Biochem. Biophys., 100 (1963) 119-130. 22 J.R. Murphy, Influence of temperature and method of centrifugation on the separation of erythrocytes. Z Lab. Clin. Med., 82 (1973) 334-341. 23 V. Bennett and D. Branton, Selective association of spectrin with the cytoplasmic surface of human erythrocyte plasma membranes. Quantitative determination with purified [3aP]spectrin. J. Biol. Chem., 252 (1977) 2753-2763. 24 M. Hosey and M. Tao, Protein kinases of rabbit and human erythrocyte membranes. Biochim. Biophys. Acta, 482 (1977) 348-357. 25 B.F. Grant, T.B. Breithaupt and E.B, Cunningham, An adenosine 3',5'-monophosphate-dependent protein kinase from the human erythrocyte membrane. Z Biol. Chem., 254 (1979) 5726-5733. 26 C.E. Mengel, The effects of hyperoxia on red cells as related to tocopherol deficiency. Ann. N. Y. Acad. ScL, 203 (1972) 163-171. 27 O.H. Lowry, N.J. Rosebrough, A.L. Farr and R J . Randall, Protein measurement with the Folin phenol reagent. Z Biol. Chem., 193 (1951 ) 265-275. 28 J.R. Dulley and P.A. Grieve, A simple technique for eliminating interference by detergents in the Lowry method of protein determination. Anal, Biochem., 64 (1975) 136-141. 29 U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227 (1970) 680-685.

70 30 R. Kahn and R.W. Rubin, Quantitation of submicrogam amounts of protein using coomassie brilliant blue R on sodium dodecyl sulfate-polyacrylamide slab-gels. Anal. Biochem., 67 (1975) 347-352. 31 R.J. I)eLange, R.G. Kemp, W.D. Riley, R.A. Cooper and E.G. Krebs, Activation of skeletal muscle phosphorylase kinase by adenosine triphosphate and adenosine 3',5t-monophosphate. J. Biol. Chem., 243 (1968) 2200-2208. 32 J. Avruch and G, Fairbanks, Phosphorylation of endogenous substrates by erythrocyte membrane protein kinases. I. A monovalent cation-stimulated reaction. Biochemistry, 13 (1974) 55075513~ 33 G. Fairbanks and J. Avruch, Phosphorylation of endogenous substrates by erythrocyte membrane protein-kinases. II. Cyclic adenosine monophosphate-stimulated reactions. Biochemistry, 13 (1974) 5514-5521. 34 C. Rubin, J. Erlichman and O.M. Rosen, Cyclic adenosine 3',5'-monophosphate-dependent protein kinase of human erythrocyte membranes. J. Biol. Chem., 247 (1972) 6135-6139. 35 P.L. LaCelle, F.H. Kirkpatrick, M.P. Udkow and B. Arkin, Membrane fragmentation and Ca 2. membrane interaction: Potential mechanisms of shape change in the senescent red cell. In M. Bessis, R.1. Weed and P.L. Leblond (eds.), Red Cell Shape, Springer-Verlag, New York, 1973, pp. 69-78. 36 F.H. Kirkpatrick, A.G. Muhs, P.K. Kostuk and C.W. Gabel, Dense (aged) circulating red cells contain normal concentrations of adenosine triphosphate (ATP). Blood, 54 (1979) 946-950. 37 S.L. Brenner and E.D. Korn, Spectrin-actin interaction. Phosphorylated and dephosphorylated spectrin tetramer cross-link F-actin. J. Biol. Chem., 254 (1979) 8620-8627. 38 H.W. Harri~ and S.E. Lux, Structural characterization of the phosphorylation sites of human erythrocyte spectrin. J. Biol. Chem., 255 (1980) 11512-11520. 39 M.M.B. Kay, Isolation of the phagocytosis-inducing IgG-binding antigen on senescent somatic ceils. Nature, 289 (1981) 491-494. 40 N.1. Swislocki, J.J. Kramer, M.A. O'Connell and E.B. Cunningham, Covalent modification of membrane components in the regulation of erythrocyte shape. Ann. N. Y. Acad. Sci., in press.