Search

Receptor mediated endocytosis appears to depend on the action of a transglutaminase (TGase). Endocytosis can be induced in intact human RBC by the action of several classes of drugs. We tested the hypothesis that drugs acted by stimulating TGase activity. Of the endocytosis inducing drugs tested, neither primaquine nor vinblastine nor chlorpromazine enhanced TGase activity. We next tested the hypothesis that TGase activity was required for drug endocytosis in RBC by adding known TGase inhibitors. Paradoxically, m-Dansyl cadaverine, the most potent TGase inhibitor, produces endocytosis in human RBC. Therefore despite apparent striking morphologic similarities, drug induced endocytosis in RBC appears to proceed via different mechanisms from those involved in receptor mediated endocytosis in other cells. In the receptor-mediated endocytosis of some hormones and growth factors, it appears that the receptor-ligand complex forms clusters over clathrin coated pits which are then internalized as endocytic vacuoles. Both the clustering and internalization of ligands are inhibited by a variety of agents shown to inhibit transglutaminase (TGase) and it is therefore proposed that TGase participates in receptor-mediated endocytosis (1–3). Human erythrocytes undergo endocytosis when exposed to drugs like primaquine, chlorpromazine, and vinblastine (4), all of which are amphipathic cations (4). However, the mechanism of drug action is not known nor is it clear that this is a form of receptor-mediated endocytosis (4). Furthermore, clustering of receptors can occur in neonatal but not adult human RBC (5). TGase has been measured in human red cells (6) although its physiologic role is unknown. Like all TGases, it is calcium dependent (6,7), and primaquine induced red cell endocytosis is enhanced by Ca ++ addition (8). Therefore, we tested the hypothesis that TGase participates in drug induced endocytosis in intact human red cells.

In order to study 45Ca distribution within erythrocytes, a method was devised that had minimal deleterious effects on the treated erythrocytes. It was observed that newly introduced 45Ca was predominantly recoverable from the cytosol and exchanged relatively slowly with membrane-associated Ca. Younger erythrocytes appeared to have relatively more 45Ca in membrane-associated sites. Erythrocytes from patients with sickle cell anemia had significantly more 45Ca in membrane-associated sites than did normal controls or patients with reticuloctosis due to a variety of disorders. There are theoretical reasons for considering the possibility that the distribution of 45Ca between cytosol and membrane-associated sites could modulate some of the properties of the erythrocyte membrane.

When amphipathic cationic drugs are added to intact human RBCs, the RBCs first undergo a stomatocytic shape change and then, if relatively large amounts of drug are added and if the metabolic state of the RBC is appropriate, endocytic vacuoles form. Vanadate has a structural similarity to the transition state of phosphate, which presumably accounts for its ability to inhibit phosphohydrolases, although other actions of vanadate have been described. Vanadate inhibited three forms of drug-induced endocytosis in intact RBCs despite the fact that the three drugs chosen (primaquine, chlorpromazine, and vinblastine) are known to have differing requirements for RBC ATP. Vanadate also inhibited the stomatocytic shape change produced by primaquine, chlorpromazine, and vinblastine, but not the stomatocytosis produced by low pH. Vanadate had no effect on RBC echinocytosis produced by lysophosphatidylcholine. In studying endocytosis in hypotonic, leaky, “white” ghosts, we discovered that vanadate inhibited only the endocytosis produced by Mg-ATP and not the endocytosis produced by manipulations that directly attack the cytoskeletal proteins. These findings suggest that ATP hydrolysis has a role in some forms of amphipathic cation-induced stomatocytosis and endocytosis in intact RBCs. In addition, studies in ghosts support the idea that Mg-ATP does indeed produce “energized” endocytosis dependent on utilization or hydrolysis of ATP.

Summary. There is impaired drug-induced vacuole formation in red cells from patients with hereditary spherocytosis (HS). The spherocytic shape per se accounts in part for the decreased vacuole formation seen, but there are constraints in HS red cell vacuole formation above that explicable by spheroidicity. Electron-microscopy indicates that the vacuoles produced in HS have one-third normal diameter. We propose that the decreased vacuole formation in HS is a result of a genetically determined abnormality of the membrane which limits its ability to invaginate and deform. The variability in vacuole formation from one HS family to another, but relative consistency within families, suggests that hereditary spherocytosis comprises a syndrome of related erythrocyte membrane disorders. Vacuole fromation in ghosts in normal in HS. This unanticipated finding suggests that the phenomenon of ghost endocytic vacuole formation may have requiements that by-pass the defect in the HS membrane.

Drug-induced endocytosis in intact human erythrocytes is a membrane-mediated event involving membrane invagination and fusion. The drugs capable of inducing endocytosis are amphipathic cations and it was thought that they intercalate themselves into the relatively electronegative cytosolic portion of the phospholipid bilayer. The current studies show that at concentrations that produce endocytosis vinblastine and chlorpromazine are not only concentrated in the red cell membrane compartment but reach high levels in the cytosol as well. The suspending medium can produce impressive effects on the extent of endocytosis. Sodium lactate enhances chlorpromazine and vinblastine endocytosis without altering the incorporation or distribution of these agents. Sucrose enhances chlorpromazine and inhibits vinblastine endocytosis, and while it produces no effects on chlorpromazine incorporation it inhibits vinblastine incorporation, thus in part accounting for its effects. It is proposed that the amphipathic dr...

The erythrocytes of the newborn infant have many properties that distinguish them from those of adults, and their membranes are also different from those of adult erythrocytes. We compared the ability of adult and neonatal RBCs to undergo endocytosis on exposure to drugs. Using a quantitative method, we showed that neonatal erythrocytes undergo a greater degree of endocytosis than do adult RBCs in response to primaquine, vinblastine, and chlorpromazine, and are sensitive to lower concentrations of the drugs. Some forms of drug-induced endocytosis are red cell age-dependent; when RBCs were separated by density gradient centrifugation, the membranes of the younger, less dense populations of both the neonatal and adult RBCs were capable of more extensive internalization than those of the denser, older RBCs. Neonatal RBCs of a given density undergo more endocytosis than do adult RBCs of the same density, suggesting that the membrane of the neonatal RBC is less stable and capable of more of the reorganization reflected in endocytosis than is the adult RBC membrane.

Human erythrocyte membranes were sheared into microvesicles with substantial loss of activity of the ATPases but not the nitrophenyl phosphatases. This loss of ATPase activity was not accompanied by alteration in lipid content, sialic acid content or orientation, or sulfhydryl group reactivity. Centrifugation in a linear sucrose gradient, resolved the microsvesicles into distinct classes which differed with respect to their lipid content, the reactivity of their sulfhydryl groups, and the activity of their ATPases and nitrophenyl phosphatases. Vesicles with low sulfhydryl reactivity had low activity of all enzymes tested. Vesicles containing relatively high lipid to protein and lipid to sialic acid ratios had low ATPase and K + -dependent nitrophenyl phosphatase activities. We have found that membranes could be sheared into readily separable membrane classes containing varying amounts of lipid relative to sialic acid and protein, and this observation supports a segmental concept of membrane structure. The parallelism of (Na + , K + )-ATPase and K + -dependent nitrophenylphosphatase activities in the separated microvesicle classes stands in contrast to their divergent behavior following shearing where the K + -dependent nitrophenyl phosphatase is not inactivated. This pattern suggested that K + -dependent nitrophenyl phosphatase probably accurately reflected the K + -dependent dephosphorylation phase of (Na + , K + )-ATPase; however, the prior step required a higher degree of structural integrity and was very susceptible to inactivation by chemical or mechanical perturbation.

The ability of human RBC and their ghosts to undergo reversible shape changes as well as endocytosis and exocytosis raised the likelihood that contractile proteins like actomyosin might be present. Actin, a component of actomyosin, is a constituent of the red cell membrane. Two generally accepted criteria for identifying a myosin include the ability of purified preparations to form arrowhead complexes with F actin fibers visible in the electron microscope, and the presence of an unusual enzyme activity where a low level of Mg++-ATPase is markedly enhanced by addition of F actin, a reaction reflecting the physiologic interaction of actin and myosin. Our studies on a RBC myosin were therefore based on the use of the actin-activated ATPase (AA-ATPase) as a reliable marker of a myosin-like protein. RBC contain AA-ATPase, all of which was recoverable in the ghost or membrane fraction prepared by either hypotonic or isotonic lysis. AA-ATPase was optimally stimulated by human erythrocyte F actin as compared to rabbit muscle F actin or human erythrocyte G actin. In common with the myosin AA-ATPase of macrophages there was no saturation effect seen upon F actin addition. Myosin AA-ATPase of platelets and macrophages is inhibited when the protein is dephosphorylated. Dephosphorylation of the RBC membrane resulted in almost complete inhibition of AA-ATPase. Another myosin enzyme marker is a K+, EDTA-ATPase, which is also detectable in the red cell membrane. The AA-ATPase does not appear to be related to the other RBC membrane ATPases. Attempts to solubilize, isolate, and purify the AA-ATPase have thus far been unsuccessful; however, the AA-ATPase is not a peripheral membrane protein and no AA-ATPase is found in purified spectrin preparations. The presence of AA-ATPase and K+, EDTA-ATPase in the red cell membrane as well as several distinguishing characteristics of activation and inactivation of the AA-ATPase presumptively establishes the likelihood that there is a myosin-like protein (which is not spectrin) in the red cell membrane.

We studied human erythrocyte membrane associated glyceraldehyde phosphate dehydrogenase in order to determine if a biological form of immobilization of the enzyme altered its kinetic properties. The results indicated that the partially purified solubilized glyceraldehyde phosphate dehydrogenase did not differ, with regard to Km for substrates or pH activity profile, from the same enzyme which was immobilized on the inner face of the erythrocyte plasma membrane.

In order to study 45Ca movements within erythrocytes, a method was devised that had minimal deleterious effect on the treated erythrocytes. Agents that induce endocytosis in intact erythrocytes (primaquine, vinblastine, and chlorpromazine) caused a prompt movement of 45Ca from cytosol to membrane-associated sites. This drug-induced movement of 45Ca to membrane sites was blocked by depleting erythrocytes of adenosine triphosphate (ATP) or by incubating them with known inhibitors of endocytosis, NaF of N-ethylmaleimide (NEM). It appears that endocytosis in intact human erythrocytes involves the movement and redistribution of 45Ca from cytosol to membrane-associated sites. Therefore, in the erythrocyte, as in perhaps other cells, movement of Ca from one site to another may modulate important cellular biologic functions.

A GROUP of hereditary non-spherocytic anaemias have been shown to be associated with unstable haemoglobins through the application of a simple screening test for thermolability involving incubation of haemolysate at 50° C in phosphate buffer (pH 6.5–7.4; 0.1 M)1–3. Unstable haemoglobins form a precipitate within a few hours but normal haemoglobin remains in solution. We have encountered two families with this type of anaemia whose haemoglobins demonstrated little or no precipitate in these conditions, but formed copious precipitate in the presence of non-phosphate buffers—observations which led us to wonder whether the recently described effects of phosphates on the oxygen-binding properties of haemoglobin4–7 might not also modify the heat stability of the molecule.