Decreased Ca Pump ATPase Activity Associated ...

Blood

Cells

(1988)

14:139-148

DlOOCl UGIiS § Springer-Verlag New York Inc. 1988

Decreased Ca Pump ATPase Activity Associated with Increased Density in Human Red Blood Cells Frank F. VINCENZI and Thomas R. HINDS ABSTRACT. Red blood cells (rbcs) from five different normal humans were separated according to density using a simple procedure. The procedure in volved centrifugation for 30 minutes in small glass tubes in the absence of any density gradient medium. This produced a column of rbcs arranged accord ing to their density. Samples of the top 8% of the columns and bottom 8% of the columns were removed from the tubes with a micropipet. From each donor, samples of the least and most dense cells, respectively, were pooled from multiple tubes for each donor and designated "top" and "bottom" cells. These top and bottom cells were compared with unselected (total) cells from the same subjects, respectively. Top cells were larger and bottom cells were smaller than total cells. ATPase activities were operationally defined and mea sured in saponin lysates of these rbcs. The Ca pump ATPase (both in the calmodulin-activated and calmodu!in-indep< ndent states [achieved by addition of compound 48/80]) of the top cells exhibited greater activity, and the Ca pump ATPase of bottom cells exhibited lower activity han total cells. It was suggested that loss of Ca pump ATPase activity is associated with rbc aging and may be a determinant of rbc life span. A mechanism for the loss of Ca pump ATPase activity was suggested. This speculative mechanism is based upon selective proteolysis of the Ca pump ATPase by the Ca-activated protease, calpain. KEY WORDS: Ca pump ATPase - Density gradient - Calmodulin - Calpain INTRODUCTION The life-span of the human red blood cell (rbc) is well known, but the mechanism(s) that determine the life-span are not understood. We have been impressed that essentially, all rbcs have an active plasma membrane Ca pump, which is expressed as a (Ca2 + -f Mg2 + )-ATPase in isolated mem branes or broken cells [1]. The Ca pump ATPase is a membrane-bound protein activated by ATP, Ca. and Mg at the inner surface of the membrane [2]. Ca-dependent binding of calmodulin (CaM) at the inner surface further increases both the ATPase [3] and transport [4] activity of the Ca pump. Reprint requests to: Frank F. Vincenzi. Department of Pharmacology. SJ-30, University of Washington. Seattle. WA 98195. USA

140

F.F. Vincenzi and T.R. Hinds

The Ca pump exists in rbcs. although mechanisms for intracellular sig nalling by Ca such as contraction or secretion appear to be absent. We have suggested, as did Kretsinger [5], that maintenance of low intracellular Ca (and thus a Ca pump in .urnan rbcs lacking anv other mechanism for Ca extrusion) is a requisite for cell viability even in the absence of a signalling function. Another aspect of the relationship between intracellular Ca and cell viability has developed from the experiments of Farber and his coworkers [6, 7], as well as others. The general picture emerging is that the accumula tion of intracellular Ca precedes cell injury and death. It has been suggested that the accumulation of intracellular Ca may be a final common pathwav in cell death [7]. Thus, it seems reasonable to put forth and test the hypothe sis that loss of the activity of the Ca pump associated with aging of the rbc might lead to accumulation of Ca, activation of a similar final common pathway, and death of the rbc. This hypothesis thus suggests that the activitv of the Ca pump is a determinant of rbc life-span. The present work provides some preliminary data appearing to support the notion that Ca pump ATPase activity is decreased in association with increasing density (and presumably, increasing age) in human rbcs. For purposes of speculation and stimulating discussion, we also discuss but do not test here, one speculative mechanism by which such aging-associated loss of Ca pump ATPase activity may be mediated. Other mechanisms lor loss ot Ca pump activity are possible and would not alter the underlving hypothesis linking Ca accumulation and rbc death. BIOLOGIC MODEL Human subjects were apparently healthy normal volunteers. Following in formed consent, blood was drawn into heparinized tubes and stored on wet ice for up to 4 hours, blood specimens were centrifuged for 5 minutes at 5000 rpm in 0.9% NaCl in a Sorvall RC-5 centrifuge fitted with an SS34 rotor and maintained at 4°'C. For each of three washings, the supernatant fluid and buffy coat were removed by aspiration, with care to remove none ot the rbcs. The resultant packed washed cells were then diluted with *> vol 0.9 /o NaCl and layered over Ficoll-Paque (Pharmacia. Piscataway NJ) The rbcs were spun through the Ficoll-Paque for 5 minutes at 5000 rpm in the same rotor. After removal of the supernatant and most of the FicoliPaque by aspiration, rbcs were washed twice to remove the Ficoll-Paque again with care to remove none of the rbcs. The resultant packed rbcs were then used for further separation based upon density. Centrifugation without any density gradient or densitv reporting medium was used to obtain relatively light and dense rbcs, respectively. This applica tion is a modification and miniaturization of the method by Murphy [81 who separated cells according to their density without oils or other gradient matenals, using an angled rotor. For density separation in our experiments

Ca Pump ATPase Activity and RBC Density

141

washed packed rbcs were centrifuged for 30 minutes in a microhematocrit centrifuge. Washed, packed rbcs were loaded into a small glass tube (internal diameter = approximately 1.5 mm). The tube is designed for use in a positive displacement micropipette ("D" tube. SMI micro pettor. Scientific Manu facturing Industries. Emeryville. CA). The tube was originally slightly longer (95 mm) than the usual microhematocrit tube (75 mm by 6.5 mm internal diameter), but was cut to a length of 75 mm. A 70 mm column of washed, packed rbcs (equivalent to 138 ul of cells) was placed in the glass tube that was then sealed at the bottom using clay. The glass tube was then placed in the head of a microhematocrit centrifuge (Clay-Adams. Autocrit) and centrifuged for 30 minutes. To prevent overheating of the cells, the centrifuge was located in a refrigerator. Following centrifugation. tubes were removed and placed in a vertical orientation before being processed. Processing consisted of identifying the hematocrit of the packed cells, typi cally 90%. The watery layer above the densely packed cells vas removed by scoring the glass tube with a diamond scribe and breaking the section away from the lower portion of the tube. The open end was then daubed carefully with a laboratory tissue to remove fluid, but not cells. In order to obtain rbcs designated as "top cells" a small diameter micropipet (SMI, type B) was then inserted into the open end of the tube and lOjii were carefully removed. This resulted in the sampling of approximately the least dense 8% of the cells (10/(138 x .9)). Finally, in order to obtain rbcs desig nated as " bottom cells," the clay-containing section of the centrifuged tube vas removed by diamond scoring and breaking, and the same SMI pipet was used to remove 10 |il of cells from the bottom of the column of packed rbcs. Twenty-four such tubes were prepared for each blood sample. The top and bottom cells from each of these tubes was then pooled for further work. The^e pooled samples presumably represent, on average, relative:' "young" and "old" rbcs, respectively. Total cells were defined by removing an aliquot of the packed, washed cells before microhematocrit centrifuga tion. Prior to saponin lysis, aliquots of the pooled, packed rbc samples were taken for determination of cell number and volume on a Coulter (Coulter Electronics. Hialeah, FL). ATPase data were expressed on the basis of cell number. MATERIAL AND METHODS Three different operationally defined ATPase activities were determined in duplicate for each blood sample. The method of Vincenzi and colleagues [9] was used with certain modifications, as noted. The incubation media contained the following components in a final volume of 0.5 ml: invariable components 50 ul of rbc saponin lysate: 18 mM histidine/18 mM imidazole buffer (pH 7.0. 37°C): ? mM MgCK: 80 mM NaCl. 15 mM KC1: 0.1 mM ethylene glycol bis (/?-aminoeth\i ethen-N.N.N'.N'-tetraacetic acid: 0.1 mM ouabain, and (variable components) 3 mM adenosine triphosphate (Na2ATP). pH 7.0: 0.2 mM CaCU and 300 ug/ml of compound 48'80. Saponin lysates of rbcs [10] were prepared by placing 45 ul of 0.25 mg/ml saponin \i a series of tubes. To these were added 5 ul of pooi. : packed rbcs (top, bottom or


142

total, respectively). This mixture was allowed to stand for 15 minutes at room temperature before addition of ATPase assay reagents (final assay volume = 0.5 ml). All ATPase assay tubes were incubated at 37° C for 60 minutes in a shaking water bath. The addition of ATP initiated the reaction. The reaction was terminated by addition of 0.25 ml of 5% sodium dodecyl sulfate (SDS) solution. In the case of blank tubes. SDS was added before the incubation. After terminating the ATPase reaction, samples were mixed, transferred to autoanalyzer sampling cups, and analyzed for inorganic phos phate (Pi). The amount of Pi appearing in the tubes with SDS added prior to incubation was taken as the blank. Mg-ATPase activity was defined as the difference between the blank tubes and the tubes containing no added CaCl2. Ca pump ATPase was defined as the extra splitting induced by the addition of CaCl2 in the absence (maximal = CaM activated) or presence (basal = CaM independent) of added compound 48/80. Determination of Pi was performed as previously described [9] using a Technicon (Tarrytown, NY) Autoanalyzer Pump II with a 16-channel manifold and a Technicon Sampler II. Output of the autoanalyzer was connected to a flow through cell in a Gilford spectrophotometer (Oberlin. OH). Absorbance at 820 nm (provided by a Beckman model DU monochromator. Fullerton, CA) was recorded on a Gilford Model 6050 recorder. Autoanalyzer reagents included l.Acid molybdate (130 ml concentrated H-.SO*. 25 g ammonium molvbdate. (NH4)6Mo7024-4H,0 to prepare 21 2. SDS. 1% 3. SDS. 1.67%

4. ascorbic acid, 9% All reagents were stored at room temperature. A phosphate standard curve was deter mine at the beginning of each day's determinations. Standard concentrations of Pi ranging from 0 to 400 nmol/ml were employed. As with the analate for reaction tubes and blanks, the standards contained a final concentration of 1.67% SDS. Pi peaks on strip chart records from the Gilford spectrophotometer were measured using a Graphics Tablet or Apple Mouse II. each driven by appropriate software operat ing on an Apple He microcomputer (Apple Computer. Cupertino, CA), to digitize the data. Statistical analysis was performed using Microsoft (Redmond. WA) Excel operating on a Macintosh microcomputer (Apple Computer. Cupertino, CA). The data are ex pressed as the average ±SD. Statistical analysis employed the unpaired / test.

RESULTS Results presented here were obtained on blood samples from five different normal human volunteers. Centrifugation in a glass tube resulted in separa tion of rbcs, according to their density. As expected, the average cell volume of top cells was greater than the average volume of total cells. Similarly Table I. Volumes of Unseparated and Density Sepa rated rbcsa Top

Total

Bottom

99.3 + 3.6*

91.7 + 3.4

$6.7±3.3"

Total rbcs were unseparated. Top and Bottom rbcs were density separated as described in Methods. Data are expressed in fl: average±SD is basec >n results from five independent experiments Significantly different from total (Pi: piro. A., et al.: Oxidation as a possible mech anism ot cciinlar aging: Vitamin E def;...ncy causes premature aging and IeG bindine to ervthroc s. Proc. Natl. Acad. Sci. USA 83, 2463-2467. 1986

Blood Cells (1988) 14:153-156

Blood Cells C Springer-Verlag New York Inc. 1985'

Decreased Ca Pump ATPase Activity Associated with Increased Density in Human Red Blood Cells F.F. VINCENZI and T.R. HINDS

Commentary Is Loss of Calcium Pump ATPase Activity Really Related to Aging of Red Blood Cells? Antonio DE FLORA Although controversies in . be aging are becoming more vigorous, a subc \troversy that seems to be :rowing still more rapidly concerns the role of Ca in the aging and removal of senescent rbcs [1]. Far from providing final conclusions to the debate, the recent findings of compartmentalization of intracellular Ca2 + [2] complicates it. Whatever the critical Ca2*-depen dent event in rbc actually is, it cannot be safely assumed to occur in condi tions of severely perturbed rbc Ca2+ homeostasis. A related fact is that use of the A23187 ionophore is no longer regarded as a fully reliable tool for the study of a number of Ca2 + -linked functions, much in the same way (so to speak among hematologists) as phenylhydrazine-induced reticu locytes in rabbits are no longer considered true representatives of normal circulating reticulocytes. Perhaps at this point theories and speculations on Ca2"-related rbc pro cesses have expanded beyond the available base of experimental facts. I agree with the idea of Vincenzi and Hinds that a progressive failure of Ca21" pump ATPase activity may be a determinant of rbc I:'e-span. How ever. I do not think that their results unequivocally support :s conclusion. The experimental system described by Vincenzi and Hini>. -'Presenting a modification of Murphy"s method [3], is a useful procedure tor reparation of rbcs according to density, although analysis of markers of rbc aging other than cell volumes would demonstrate age-dependent fractionation too. My primary reservation deals with their assumption that assay of Ca2 * ATPase as well as other enzyme assays performed under optimal conditions Reprint requests to: A. De Flora. Institute of Biochemistry. School of Medicine. Viale Benedetto XVI. Genoa. Italy

154

A. De Flora

of pH. ionic strength, substrate saturation, etc. may reflect true alterations in the efficiency of actual Ca2* extrusion from circulating rbcs. Evaluation of Km values for Ca2* and ATP would in my opinion provide more accurate information on the molecular parameters of this transmembrane protein, allowing some educated guesses to be made on its physiologic function as related to rbc density. The variety of modulators of Ca2*-ATPase activity including calmodulin (CaM), acidic phospholipids, and polyunsaturated fat ty acids [4] introduces further complexity and considerably complicates the interpretation of the data provided by Vincenzi and Hinds. A major point that is not yet completely documented is whether Ca2 + levels increase during rbc aging. There are some reports indicating otherwise [5, 6], whereas other measurements do imply increased Ca2 * in high-density rbcs [7]. Moreover. Ca2+ accumulation following treatment of rbcs with ionophore A 23187 and Ca2* has been reported to take place preferentially in heavy rbc fractions, leading to selective loss of deformability of these presumptively "aged" rbcs [7]. The recent finding of apparent ionophoric effects elicited by physiologic deformation of rbcs in the microcirculation [8, 9] is in agreement with the view that Ca2+ concentrations may increase transiently. If not properly dealt with by efficient Ca2*-ATPase activity, such transient increases might lead to progressive and irreversible impair ment of cellular processes [10] including active Ca2+ extrusion by the occur rence of a vicious circle. However, as already emphasized, compartmentalizution of Ca2+ in membrane-associated vesicles is a further complicating factoi Many assumptions regarding the triggering of Ca2 "-stimulated pro" cesses now require reexamination. An example would be analysis of cytoskeletal proteins for the presence of proteolytic modifications and transglutaminase-mediated crosslinking. Both changes are often postulated, but supporting data is minimal. I am in substantial agreement with the possibility that Ca2*-ATPase is a substrate of calpain (the proteolytically modified active form of procalpain) due to a CaM-binding region in the Ca2+ pump and to a CaM-related domain in calpain itself. Led by this idea and in collaboration with the laboratory of Professor E. Carafoli (Zurich». we have already shown that, indeed, Ca2*-ATPase purified from human r cs undergoes limited digestion by purified calpain. The structural basis of this proteolytic digestion is con version of the native Ca2 +-ATPase (138 KD) to a modified 124 KD species that keeps its maximal (i.e.. assayed in the presence of CaM) activity unaf fected, although it is desensitized to CaM stimulation [IV. !n other words, purified Ca2+-ATPase is stimulated in its "basal" (with u: added CaM) activity by limited degradation catalyzed by calpain, similarly to the already describe:; proteolytic activation by trypsin [12] and chymotrypsin [11] through dnetically and structurally distinct degradation pathways. Obtained with the two purified enzyme proteins, these results clearly indicate that Ca2+-ATPase too is a target of calpain activity, adding to those cytoskeletal submembranous proteins known to be degraded in condi tions of intracellular calpain stimulation (mostly ankyrin and band 4.1). A further step, still under active investigation in laboratories at b. th Zurich

Commentarv

and Genoa, concerns the in situ degradation of Ca2 "-ATPase within white rbc membranes by purified calpain. Once again, our preliminary data .re consistent with a limited digestion taking place in this system and imitating the mechanism observed with the two purified proteins. Specifically, the sophisticated technique of in situ detection of the active E-32P complex following membrane incubation with ATP[-/-32P] [13] reveals a limited attack on the native Ca2' -ATPase species to yield a smaller, but still active, enzyme form [14]. The functional counterpart of these experiments implies concomi tant and progressive stimulation of basal Ca2 + -ATPase activity, whereas "maximal" activity is unaffected. Of course, much work remains to be done before concluding that Ca2 + ATPase activity in intact rbcs is or is not really inactivated by calpain. For instance, it does not escape us that the presence of resident ;..id endopeptidases in the rbc membrane r^ight explain further degradation of the calpain-nicked Ca2 "-ATPase form in the intact rbcs. regardless of their apparent switching off in the membrane-calpain reconstructed system. A reasonable approach to investigating such possibility is in situ detection of active Ca2'-ATPase in membranes from Ca2 + -loaded rbcs. a treatment that triggers conversion if procalpain to calpain [15]. This approach, how ever, has shown unexpected and still unsolved technical problems. We are currently trying to overcome ihes^ problems in an attempt to rationalize in ter~s of protein structure he rrevit usly observed loss of Ca2 + -ATPase activir in rbcs from favic patie:: 'hat is. in conditions of remarkably enhanced intraerythrocytic Ca2+ levjjs [6] and of consequent transient stim ulation of intracellular calpain activity [16], I believe that structural comparison of Ca2 "-ATPase in rbc membranes from favic and normal individuals, still not chieved. would clarify whether an elevation of intracellular Ca2* leads to calpain-catalyzed inactivation of the Ca2" pump. The experimental difficulties are more daunting when lower increases of intraerythrocytic Ca2+ are evaluated in respect to their influence on both structure and activity of membrane Ca2'-ATPase. The unequivocal proof of Vincenzi and Hinds' assumption is still an unrealized though perhaps not a hopeless task. Acknowledgments. A part of the research on proteolysis of Ca; "-ATPase by calpain was supported by the Special Project of C.N.R. "Genetic Engineering and Molecular Bases of Inherited Diseases". Rome. Italy, bv the Swiss Nationalt'onds. and bv the National Institutes of Health.

REFERENCES 1. Cellular and Molecular Aspects of Aging. The Red Cell as a Model. J.W. Eaton. D.K.. Konzen. JG. White, eds. Progress in Clinical and Bioloaical Research. Vol. 195. Alan R. Liss. Inc.. New York. 1985. pp. 1-454 2. Lew. V.L.. Hockaday. A.. Sepulveda. M.I.. et al.: Compartimentalization of sickle cell calcium in endocytic inside-out vesicles. Suture 315. 586-589. 1985 3. Murphy. J.R.: Influence of temperature and method of centrifusation on the separa tion of erythrocytes. J. Lab. Clin. Med. 82. 334--41. 1973

156

A. De Flora

4. Carafoli, E.: Intracellular calcium homeostasis. Ann. Rev. Biochem. 56. 395-433. 1987 5. Wiley. J.S.. Shaller. C.C.: Selective loss of calcium permeability on maturation of reticulocytes. J. din. Invest. 59. 1113-1119. 1977 6. De Flora. A.. Benatti. U.. Guida. L., etal.: Disordered ervthrocvte calcium ho meostasis. Blood 66. 294-297. 1985 7. Shiga, T., Sekiya. M.. Maeda. N.. et al.: Cell age-dependent changes in deformability and calcium accumulation of human erythrocytes. Biochim. Biophys. Acta 814. 289-299, 1985 8. Swislocki, N.I.. Kramer, J.J., O'Connell, M.A.. Cunningham, E.B.: Covalent modification of membrane components in the regulation of ervthrocvte shape. Ann. :V. Y. Acad. Sci. 416. 622-668. 1983 9. Vincenzi. F.F.. Cambareri, J.J.: Apparent ionophoric effects on red blood cell defor mation. In: Eaton. J.W.. Konzen, D.K.. White. J.G. eds.. Cellular and Molecular Aspects o' 'ing: The Red C 7 as a Model. New York. Alan R. Liss 1985. pp. 213-222 10. SwiSLor.i. ..I.. Tierney. I.: Covalent modification of membrane components during t. .rocyte aging. In. -aton. J.W.. Konzen. D.K., White. J.G.. eds.. Cellular and Molecular Aspects of Aging: The Red Cell as a Model. New York. Alan R. Liss. 1985. pp. 195-207 11. Carafoli. E.. Fisher. R., James. P.. et al.: The calcium pump of the plasma mem brane. Recent studies on the purified enzyme and on its proteolytic fragments, with particular attention to the calmodulin binding domain. In: Norman A.W.. etal.. eds. Calcium-Binding Proteins in Health and Disease. New York. Academic Press, 1987. pp. 78-91 12. Niggli, V., Adunyah. E.S.. Carafoli, E.: Acidic phospholipids, unsaturated fatty acids, and limited proteolysis mimic the effect of calmodulin on the purified erythro cyte Ca2'-ATPase. J. Biol. Chem. 256. 8588-8592. 1981 13. Sarkadi. 3.. Enyedi. A., Foldes-Papp. Z., Gardos. G.: Molecular characterization of the "in situ" red cell membrane calcium pump by limited proteolysis. J. Biol. Chem. 261. 9552-95"7. 1986 14. Morelli. A., Castello. G.. Piu. M.. et al.: Attivazione "in situ" della Ca:'-ATPasi in membrane di entrociti umani mediante trattamento con calpaina. Abstr. E 114. 33rd Meeting of the Italian Biochemical Society, Brescia-Gardone Riviera. September 26-28, 1987 15. Grasso. M., Morelli. A.. De Flora. A.: Calcium-induced alterations in the levels and subcellular distnbuu* n of proteolytic enzymes in human red blood cells. Biochem. Biophys. Res. Commun. i 38. 87-94, 1987 16. Morelli. A.. Grasso. M., Meloni. T.. etal.: Favism: Impairment of proteolytic systems in red blood cells. Blood 69, 1753-1758. 1987

Blood Cells (1988) 14:157-159

Blood Cells £ Springer-Verlag New York Inc. 1988

Decreased Ca Pump ATPase Activity Associated with Increasing Density in Human Red Blood Cells F.F. VINCENZI and T.R. HINDS

Reply to Commentaries Drs. Lowenson and Clarke suggest that transglutaminase and calpain are activated "only at relati\ Ay high Ca ion concentrations." This is certainly true for transglutaminase [1], but as far as we are aware, some forms of calpain are activated at micromolar levels of Ca [2]. This is so because calpain contains a calmodulin-like Ca2 + binding sequence as a part of its primary structure [3]. Thus, we view calpain as being potentially activated in cells at or near physiologic levels. It is our position that, during circulation in the microvasculature, intracellular Ca2+ in rbcs increases (physiologically, albeit transiently, because of the high capacity of the Ca pump) to micromo lar levels by shear-stress-induced Ca influx [4, 5]. We view transglutaminase as an enzyme activated in cells only under more *% catastrophic" conditions. This is part of the reason we did suggest that calpain. but not transglutamin ase is involved in normal aging of the rbc Ca Dump ATPase. Of course, another reason is the predicted relative selectivity of calpain toward the Ca pump ATPase (with which Professor De Flora and colleagues, would appear to agree). With respect to the data presented by Bookchin and associates [6], we must point out that the data were sketchy and were presented in generalities. Bookchin indicated that Ca2* loaded (for 1 hour) rbcs survived approxi mately as well as nonloaded cells when reinfused in a rabbit. Ca2 +-loaded rbcs survived approximately 75% of controls. Thus, in fact, a modest de crease of survival in vivo was observed, but was deemphasized. As Dr. Swislocki pointed out in the discussion of Dr. Bookchin's articles the dura tion of 1 hour for Ca2+ loading might not be long enough to mimic the effect of chronic, albeit transient, Ca2"' stresses in the circulation. We believe that it takes many passes of the rbc through the microcirculation (i.e., about 120 days worth) to accumulate sufficient Ca2* induced damage to result in the terminal aging of the cell by loss of Ca2 * pump ATPase activity. A simple analogy may help. Accumulated effects of Ca2* transients in the rbc may be likened to the effects of drop* of water on a rock over a long period of time. Bookchin's experiment was somewhat like dumping a bucket of water on the rock, seeing little or no effect, and the concluding that water could never wear away a rock.

158


We agree with Lowenson and Clarke that further studies are needed to clarify both the changes in calcium ion concentration with cell aging and the physiologic effects of these changes. Given adequate funding for our studies, this is one of the areas we wish to pursue. The data of Dr. Margaret Clark presented in this volume and similar data presented previously at the Symposium on Cellular Aging, the Red Blood Cell as a Model [7] were interpreted to demonstrate no difference in the Ca pump activity of the most-dense as compared to the least-dense human rbcs. On the other hand, if one expresses Dr. Margaret Clark's data on a per cell basis, then her data look very much like ours. We have made this point in our commentary on her paper in this volume. There is no question that, when expressed on a per cell basis, the two sets of data are similar. The statement of Lowenson and Clarke that "... Vincenzi and Hinds cannot explain the reason they observed changes in the calcium pump activity with cell density and Margaret Clark did not," has thus been answered. We agree with Dr. Margaret Clark that the rate of Ca pump activity measured in her experiments was probably less than the maximal capacity of the pump. We also agree that the rates in the least-dense and the mostdense cells are essentially indistinguishable, when expressed on the basis of the units she used, namely, per liter of cells. Because we interpreted the issue of red blood cell aging as refering to the cell (not a liter of cells), ve chose to express our data in "per cell" terms. We recalculated Dr. Margaret Clark's data on a per cell basis, using her cell volume data, and we came up with an approximately 20% decrease in the rate of the mostdense as compared to the least-dense cells. Thus, whether the maximal ca pacity (as in our ATPase assay) or the rate under some limiting condition (is in Dr. Margaret Clark's assay of transport) is measured, we believe •hat the data are consistent with the interpretation that, on a per cell basis, the activity of the rbc Ca pump decreases with increasing cell density, and by inference, with increasing cell age. In part of Dr. Margaret Clark's commentary, she compares our work to that of Luthra and Kim [8] and then discusses their data, not ours. We are left without comment because the results -A Luthra and Kim are qualitatively similar but not quantitatively the same as ours. Furthermore, their methodology was different from ours. Simple membrane loss may indeed account for some or all the decrease in Ca pump ATPase activity we found. It may be that the specific activity >f the Ca pump (per milligram membrane protein) decreases in association with cell aging. Decreased specific activity could be due to a loss of pump sites or a decrease in the activity of pump sites. It would be useful to quantify the number of Ca pump sites in the membranes of young vs old rbcs. By the use of azido-calmodulin labeling [9, 10] we intend u carry out such quantification in the near future. Professor De Flora used the commentary on our paper as a forum to mention a number of issues. Although most of what he said is indeed important and true, much of it was somewhat tangential to our article, in our opinion. We agree with much of what Professor De Flora says as

Reply to Commentaries

159

it pertains to our work. In particular, we agree that our results are not unequivocal. They are preliminary, but encourage rather than discourage us to continue to pursue the hypothesis in question. Many of the ideas expressed by Professor De Flora are parallel to those we have discussed in our laboratory for a number of years. Ve find particularly pithy his comment that "unequivocal demonstration of Vincenzi and Hinds' assump tion is still a remote, although not a hopeless task." We trust that it is not too remote to prove or disprove. Admittedly, as Professor De Flora states, "theories and speculations have expanded beyond the available base of experimental facts." If this were not the case, the f'eld would remain stagnant, caught in a quagmire of complacency and conformity. We are pleased, in fact, to have at least helped to cause a small ripple in complacency, and although it was suggested that we speculated wildly and must be willing to modify our speculations based upon new data, the data need no modification. They are, to us, patently clear. Finally, we would like to say that we have found that the intellectual j.xercise of this "meeting without walls" to be interesting and exciting. Our thanks to Dr. Brian Bull and the Board and staff of Blood Cells for he!: ng to make this possible. REFERENCES 1. Lorand. L.. Conrad. S.M.: Transglutaminases. Molec. Cell Biochem. 58. 9-35, 1984 2. Pontremoli. S.. Melloni, E.: Extr.ilvsosomal protein degradation. Ann. Rev Biochem. 55, 455-481. 1986 3. Ohno. S.. Emori. Y., Imajoh. S.. et al.: Evolutionary origin of a calcium-dependent protease bv fusion of eenes for a thiol protease and a calcium-binding protein? Nature 312. 566-570, 1984 4. Larsen. F.L.. Katz. S., Roufogalis. B.D.. Brooks. D.E.: Physiological shear stresses enhance the Ca2~ permeability of human erythrocytes. Nature 294. 667-668. 1981 5. Vincenzi, F.F.. Cambareri, J.: Apparent ionophoric effects of red blood cell deforma tion. In: Eaton, J.W., Konzen. D.K.. White. J.G.. eds.. Cellular and Molecular Aspects of Aging: The Red Cell as a Model. New York. Alan R. Liss. 1985. pp. 2136. Bookchin, R.M.. Lew. V.L.. Roth. E.F.: Elevated red cell calcium: Innocent by stander or kiss of death? In: Eaton. J.W.. Konzen. D.K.. White. J.G.. eds. Cellular and Molecular Aspects of Aging: The Red Cell as a Model. New York. Alan R. Liss. 1985. pp. 369-375 7. Clark. M.: Selected ionic and metabolic characteristics of human red cell populations separated on stractan density gradients. In: Eaton. J.W.. Konzen. D.K.. White. J.G.. eds.. Cellular and Molecular Aspects of Aging: The Red Cell .v a Model. New York. Alan R. Liss, 1985, 381-386 8. Luthra. M.G.. Kim, H.D.: (Ca2* + Mg2")- ATPase of density-separated human red cells. Effects of calcium and a soluble cytoplasmic activator (calmodulin). Biochim. Biophys. Acta 600. 480-488. 1980 9. Vincenzi. F.F.. Andreasen. T.J., Hinds. T.R.: Azido < 125I)-calmodulin, a photoaffjnity probe for the study of calmodulin as a regulator of plasma membrane Ca2 ~ transport. In: Bronner, F.. Peterlik. M.. eds.. Calcium and Phosphate Transport Across Biomemhranes. New York. Academic Press. 1981. pp. 45-50 10. Hinds. T.R.. Andreason. T.J.: Photochemical Crosslinking of Azido-Calmodulin to the (Ca2* -}-Mg2 +)ATPase of the Ervthrocvte Membrane. J. Biol. Chem. 256, 7877-7882, 1981