Red Cell Aging II. ANOMALOUS ELECTROPHORETIC PROPERTIES OF NEURAMINIDASE TREATED HUMAN ERYTHROCYTES F. J. NORDT, R. J. KNOX AND G. V F. SEAMAN Department of Neurology, University of Oregon Health Sciences Center, Portland, Oregon 97201

ABSTRACT Desialylation of human red blood cells (RBC) by Vzbrio cholerue neuraminidase (VCN) was found to produce cells with electrophoretic properties which were inconsistent with the view of simple loss of N-acetylneuraminic acid (NANA) as the sole effect of VCN treatment. Modification of human RBC with 50-350 U VCNA0'' RBC for one hour a t 37°C releases 90100% of the NANA and produces a progressive decrease towards zero in their electrophoretic mobilities when measured in 0.15 M NaCl (pH 7.2) at 25°C. The appearance of positive groups on the desialylated cells was indicated by the VCN-treated cells displaying positive mobilities below - pH 5.5 and increased negative mobilities at - pH 9 as well as substantial increases in their mobility a t neutral pH following treatment with formaldehyde. Adsorption of about 95% of the VCN activity at 0°C to the RBC did not produce any significant change in their electrophoretic mobilities thus indicating that the observed changes in the electrophoretic properties of the RBC following VCN treatment could not be attributable to adsorption of VCN. These studies indicate that the cationic charge groups which appear a t the electrophoretic surface of the RBC after VCN treatment are probably of endogenous origin. I t is suggested that this alteration rather than simple NANA release may operate to shorten the in vivo survival time of desialylated red cells. Terminal sialic acid residues occur on the carbohydrate portions of several glycoproteins present a t the exoface of various mammalian red cells (Klenk and Uhlenbruck, '58; Seaman and Uhlenbruck, '63) and their ionized carboxyl groups account for a major portion of the negative surface charge of these cells (Cook e t al., '61; Eylar et al., '62; Seaman and Uhlenbruck, '63). In many instances treatment of these erythrocytes with neuraminidase (acylneuraminyl hydrolase EC 3.2.1.18) from Vibrio cholerae releases virtually all of the external membrane-bound sialic acid (Cook et al., '61; Eylar et al., '62; Seaman and Uhlenbruck, '63). For the human erythrocyte the particular sialic acid, Nacetylneuraminic acid (NANA), is removed by Vibrio cholerue neuraminidase (VCN) with a concomitant marked decrease in red cell charge (Cook e t al., '61; Eylar e t al., '62). The sialic acid content of circulating human 3. CELL. PHYSIOL. (1978)97: 209-220.

red blood cells is a function of cell age with the older red cells containing about 10%less sialic acid than the younger red cells (Greenwalt and Steane, '73; Cohen e t al., '76; Seaman et al., '77). However the constancy in electrophoretic mobility for red cells of different ages (Seaman et al., '77) indicates that the effective net number of negative charges per unit area of membrane a t the electrophoretic surface remains essentially constant. Nevertheless the diminished levels of sialic acid in senescent red cells has kindled speculation that loss of sialic acid during the in vivo life of the red cell constitutes a normal determinant of red cell life span (Bocci, '76). Many investigators have found that reinjection of autologous red cells treated in vitro with neuraminidase results in a markedly shortened red cell life span in a variety of animals inReceived Feb. 22, '78. Accepted Apr. 28, '78.

209

210

F. J. NORDT, R. J. KNOX AND G . V. F. SEAMAN

cluding dogs and goats (Aminoff et al., '76), mice (Landaw et al., '77), rabbits (Marikovsky e t al., '77) and rats (Aminoff, et al., '77) as well as in humans (Jancik et al., '75). Aside from studies on the biological roles of red cell sialic acid, VCN has been used extensively to elucidate the contribution of sialic acid to the physicochemical properties of the red cell surface. The mild conditions required for enzyme activity, the completeness of removal of sialic acid by VCN, and the commercial availability of a highly purified preparation of VCN, made it an attractive tool for studies of the biological roles of sialic acid on the surfaces of red cells as well as other cell types (Jeanloz and Codington, '76). Critical to the interpretation of such experiments is a knowledge of the purity of the enzyme preparation, the specificity of its action on the cell membrane and the extent to which adsorption of the enzyme or other components of the preparation influence the measurements in question. Seaman ('75) recently reviewed studies of the electrokinetic properties of red cells, the use of neuraminidase in studies of these properties and the consequences of alterations produced by VCN with regard to colloid stability, cell-cell interaction and interactions with macromolecules. This communication reports recent studies in which it is shown that the effect of neuraminidase treatment on human erythrocytes is not limited simply to the removal of NANA from the cell surface. The roles of enzyme adsorption and structural rearrangement of the cellular peripheral zone in the production of anomalous electrokinetic properties of red cells by VCN treatment are examined. Finally we examine critically some of the experimental evidence for a primary role of sialic acid in determining normal red cell life span. MATERIALS AND METHODS

Blood collection and processing Human blood was drawn by venipuncture from healthy male and female adult donors (aged 23-44 years) into disodium ethylenediamine tetraacetate (Na2EDTA.2 H20) to a final concentration of 1.5 mg/ml blood. The blood was centrifuged for 15 minutes a t 2,000 g at room temperature and the supernatant plasma and the buffy coat were removed. The red cell pack was washed three times in about 25 volumes of 0.15 M aqueous sodium chloride solution buffered to pH 7.3 -+ 0.2 with 0.15 M aqueous sodium bicarbonate solution (stan-

dard saline) with careful aspiration of residual buffy coat after each wash. Red cells were density fractionated according to the method of Murphy ('73) as described by Seaman et al. ('77). Neurarninidase treatment of erythrocytes Vibrio cholerae neuraminidase was purchased from Behring Diagnostics. The enzyme which is prepared as described by Schick and Zilg ('76) was supplied in 0.05 M sodium acetate buffer, pH 5.5, containing 0.15 M NaCl and 0.01 M CaClz (enzyme buffer) in 1-ml vials containing 500 Units of activity where one Unit is the amount of enzyme required to release 1pg of N-acetylneuraminic acid from human a,-acid glycoprotein in 15 minutes at 37°C. A VCN preparation (500 U/mU having approximately a 40% higher specific activity was obtained as a gift from Behring Diagnostics. The standard VCN treatment consisted of 20 one wash of the packed red cells in volumes of the media to be used during the enzyme treatment followed by their suspension in the medium to a volume concentration of 20% v/v. Volume concentrations were calculated from the microhematocrit readings obtained after five minutes of centrifugation a t 15,00Og, assuming a packing fraction of 0.99. Red cell concentrations were either measured by electronic particle counting with an Electrozone/Celloscope (Particle Data, Inc.) or were calculated from the hematocrit assuming 1.1 x 10'' cells/ml. Incubations of red cells with VCN were initiated by the addition of the enzyme to the cell suspension which had been equilibrated to the incubation temperature. Control cell suspensions were incubated in the same manner except that an equivalent volume of pH 5.5 enzyme buffer was added instead of enzyme. The final pH of all incubation mixtures was seven. Different enzyme concentrations (expressed as enzyme Units/lO1° RBC), incubation media, and incubation times were employed at either 0°C or 37°C as indicated under

-

-

RESULTS.

Supernatant fluids were recovered from the incubation mixtures after centrifugation for five minutes at 2,500g at either room temperature or 0-4OC. Aliquots were assayed for NANA content by the thiobarbituric acid method (Aminoff, '61). The results were computed as fg NANA released per cell as described previously (Seaman et al., '77) or as

211

ERYTHROCYTE MODIFICATION BY NEURAMINIDASE

percent of total release where the total releasable NANA was the observed quantity released at 37°C in three hours by 2 60 U VCN/lO1° RBC. Neuraminidase activity was measured by assaying t h e amount of NANA released from neuraminlactose (Calbiochem, A grade) by the thiobarbituric acid method (Aminoff, ’61). The basic assay procedure consisted of incubating a 0.25 ml aliquot of sample with 0.25 ml of enzyme buffer containing 200 pg neuraminlactose for 15 to 60 minutes a t 37°C. The reaction was stopped by the addition of the strongly acidic periodate reagent employed in t h e NANA assay method.

binding of protein to Amidoschwartz 10B (Shaffner and Weissman, ’73); (b) binding of protein to Coomassie Brilliant Blue G-250 (Bradford, ’76); and (c) measurement of nitrogen by the microkjeldahl procedure using Nessler’s reagent (Campbell et al., ’63). Standard crystallized bovine albumin (BSA, Pentex, Inc.) was used to calibrate the dye binding methods. The BSA concentration of a 1%w/v solution was measured a t 278 nm using c$F8 = 6.6 (Tanford and Roberts, ’52). The presence of proteolytic activity in the VCN preparation was checked with Azocoll (Seaman et al., ’67). RESULTS

The protein concentration of the purchased VCN solutions as measured by both dye binding methods was 10 pglml. The protein nitrogen concentration from t h e microkjeldahl assay was 1.94 pg/ml which corresponds to 12 pg proteinlml assuming a n average nitrogen content of 16%wlw. No proteolytic activity was detectable in the VCN preparation by the Azocoll method. The quantities of NANA released from red cells treated for one hour at 37°C with VCN at concentrations ranging from 1.5-350 U/lOIO RBC are shown in figure 1with the resulting Analytical particle electrophoresis electrophoretic mobilities of the cells in stanElectrophoretic mobilities were measured dard saline a t 25°C. Relatively low concentraw i t h a c y l i n d r i c a l c h a m b e r a p p a r a t u s tions of enzyme (-- 40 UllO’o RBC) released 90%of the total NANA in one hour at 37°C equipped with AgIAgC1 electrodes (Seaman, ’75). The chamber was immersed in a tempera- in both 0.15 M NaCl buffered with 0.01 M soture bath maintained at either 25.0 0.1”C or dium phosphate to pH 7.4 and 0.15 M NaCl0.005 M CaCl, (Ca-saline) and reduced the at 0.2 0.2”C. - 0.2 The electrophoretic suspending media were mobilities of the cells by about 80%to 0.15 M NaCl and 0.03 M NaC1-0.22 M sorbitol, pm sec-’ V-’ cm in standard saline. With the pH of which were adjusted to the indicated higher levels of enzyme the mobilities of the values with HCl, NaOH or NaHC03 made to cells approached zero so that accurate mothe corresponding ionic strengths and where bility measurements were difficult to collect. zero in standard saline at appropriate rendered isotonic with sorbitol. A mobility of During t h e collection of pH-mobility data the 25°C appeared to represent the “endpoint” of suspension pH was checked before and after the VCN treatment. This result was obtained electrophoresis and was found to vary by less for cells treated at 37°C with high concentrathan 0.1 pH unit. The limits of pH within tions of VCN (700 U/lOIO RBC) for one hour as which t h e cells displayed electrokinetic stabil- well as for cells treated for three hours with 60 ity were established as described by Heard and U/10lo RBC. Slight hemolysis (