Marilyn J. Telen
Evolution of Our understanding of the Erythrocyte
Erythrocytes were first described in the 17th century. The Dutch microscopist, Leeuwenhoek, took note of them, as did Malpighi, who mistook them for fat globules “looking like a rosary of red coral.” For many years, erythrocytes were not thought to be of any importance. The presence of iron in blood was demonstrated by Lemery in the 17th century, but not until 1851 did Funke isolate hemoglobin in crystalline form. The functional significance of red corpuscles was only appreciated, however, when Hoppe-Seyler demonstrated that hemoglobin has the property of readily taking up and discharging oxygen. This, then, was considered the primary or even sole function of the red cell (along with CO2 exchange) until the late 20th century. Now, however, our understanding of erythrocyte function has broadened to include O2, CO2, and NO exchange as well as immune clearance and, possibly, clearance of other soluble blood components, such as cytokines.
Structural Features
The uniquely anuclear mature human erythrocyte is one of the most highly specialized of cells. Lacking such cytoplasmic organelles as nucleus, mitochondria, or ribosomes, the red cell is unable to synthesize new protein, carry out the oxidative reactions associated with mitochondria, or undergo mitosis. More than 95% of the cytoplasmic protein is hemoglobin. The remainder includes those enzymes required for energy production and for the maintenance of hemoglobin in a functional, reduced state. However, the erythrocyte expresses a perhaps surprising number of proteins that subserve functions we associate with other cells, including a variety of transport proteins, adhesion molecules, receptors, and signaling pathways. Thus the red cell is now recognized as performing a number of crucial and complex functions in the human body.
Shape and Dimensions
At rest, the normal human erythrocyte is shaped like a flattened, bilaterally indented sphere, a shape often referred to as a biconcave disc (Fig. 7.1). In fixed, stained blood smears, only the flattened surfaces are observed; hence, on fixed blood films the erythrocyte appears circular, with a diameter of about 7 to 8 μm and an area of central pallor corresponding to the indented regions.
The dimensions of the red cell in the living state have been estimated by measurements made on photomicrographs (Fig. 7.2) (1,2,3). For this purpose, the cells are suspended in isotonic solutions and photographed on edge. With ordinary light microscopy, the potential error associated with imprecision of focus is about 0.5 μm. This value can be reduced to 0.02 μm by means of holography with the interference microscope (2). Dimensions obtained by the latter method are somewhat smaller than those reported in earlier studies (1). With drying and staining, the cell shrinks so that diameters measured in fixed preparations are somewhat smaller than those of cells in isotonic solutions. The painstaking studies of Price-Jones yielded an average normal value for red cell diameter of 7.2 to 7.4 μm (4).
Indirect measurements of cell volume are consistent with values obtained using the microscopic method. Average values for the mean cellular volume in normal subjects range from 85 to 91 fl, depending on the combination of methods used. The variation in cell size can be documented by means of a frequency distribution curve of red cell volumes generated from the output of a Coulter counter (Fig. 7.3). Although some studies have suggested that red cell volumes are log-normally distributed (5), the size distribution is symmetric (6). Ninety-five percent of normal red cells are between about 60 and 120 fl in volume (5). However, some workers have challenged these values using transmission electron microscopy and stereology. They estimate that the true volume of mature red cells is only 44 fl, and that only 51% of the volume of the red cell column observed in a hematocrit tube is occupied by erythrocytes (7). Interestingly, these newer methods also calculate that erythrocytes have 44% more membrane surface area than necessary to accommodate cytoplasmic contents in a sphere (7), a figure very close to that estimated using earlier techniques (8).
Total hemoglobin content and red cell volume vary considerably more than does hemoglobin concentration (9). It has been proposed that mature red cell size and hemoglobin content are primarily dependent on erythroid precursor cell size at the last cell division during erythropoiesis (10). Reticulocytes are 24% to 35% larger than mature red cells, although they have similar total hemoglobin content (and thus a lower hemoglobin concentration) (11).
The disc shape is well suited to erythrocyte function. The ratio of surface to volume approaches the maximum possible value in such a shape (12), thereby facilitating both gas transfer and deformability as the red cell traverses the microcirculation (13). When red cell movements within small blood vessels were observed by cinemamicrography (14), the plane of the biconcave disk was found to be oriented in the direction of flow. The leading edge became pointed and the following edge blunted (Fig. 7.4); thus, the shape was similar to that of a parachute or torpedo viewed from the side. When deformed in this way, the erythrocyte can pass through a vessel of about 4 μm in maximum diameter. Erythrocyte shape may also vary between large and small vessels and under conditions of high or low shear stress. Circulating erythrocytes rarely appear to be biconcave discs in the murine abdominal aorta, whereas typical biconcave discoid shapes are common in the inferior vena cava (15).
Deformability
The erythrocyte is remarkable for its ability to maintain membrane integrity while exhibiting extreme deformability under normal physiologic circumstances (16). Without undergoing extensive remodeling, the erythrocyte membrane withstands high shear stresses, rapid elongation and folding in the microcirculation, and deformation as the erythrocyte passes through the small fenestrations of the spleen. Cell deformability depends on both the membrane and the cytoplasm; however, the cytoplasm of normal erythrocytes (as opposed to sickled erythrocytes, for example) acts as an ideal liquid and, at physiologic concentration, has very low viscosity (16). Thus, it is the elasticity and viscosity of the membrane that are crucial for deformability.
|
|
|
Figure 7.1. The normal mature erythrocyte as visualized by the scanning electron microscope (×9,800). (Courtesy of Dr. Wallace N. Jensen.) |
At least three sets of circumstances result in a spherical erythrocyte shape: osmotic (hypotonic) swelling, discocyte– echinocyte transformation, and discocyte–stomatocyte transformation. To a point, all three types of shape change are reversible. Osmotic swelling (1) occurs when erythrocytes are suspended in hypotonic solutions. Under such circumstances, the cell acquires water and swells, first becoming cup-shaped and then spherical. These changes are associated with an increase in volume while the cell surface area remains the same or increases only slightly (2).
Discocyte–echinocyte transformation (17) takes place when intracellular adenosine triphosphate (ATP) is depleted (18), when intracellular calcium content is increased (19), when the cell is exposed to stored plasma, high pH, anionic detergents, lysolecithin, or fatty acids, or when the cell is washed and placed between a glass (not plastic) slide and a coverslip (1). ATP depletion reversibly transforms erythrocytes to spiculated sphero-echinocytes; restoration of ATP leads to gradual resumption of normal shape (18). Transformation because of ATP depletion or other factors proceeds through several recognizable stages (Fig. 7.5). First, the disk becomes irregularly contoured (echinocyte I); second, crenations (regularly spaced spicules or projections) appear on the flat surface (echinocyte II); third, the cell becomes ovoid or spherical, with about 30 evenly spaced spicules projecting from the surface (echinocyte III); fourth, the cell becomes more distinctly spherocytic, with tiny spicules (spheroechinocyte I); and ultimately, these spicules become so small that they can be appreciated only with the scanning electron microscope (spheroechinocyte II) (20). The last stage is irreversible and is sometimes referred to as a “prelytic” sphere.
|
|
|
Figure 7.2. Dimensions of a cross section of the erythrocyte in isotonic solution. Values are means ±1 standard deviation. (From Evans E, Fung YC. Improved measurements of erythrocyte geometry. Microvasc Res 1972;4:335.) |
|
|
|
Figure 7.3. Frequency distribution curve of erythrocyte volume. The cells are normally distributed about a mean volume of 90 fl. (Modified with permission from Bessman JD, Johnson RK. Erythrocyte volume distribution in normal and abnormal subjects. Blood 1975;46:369.) |
Discocyte–stomatocyte transformation (21) occurs when red cells are exposed to low pH, cationic detergents such as Tween-80, or phenothiazines. As the change proceeds, the cell loses the indentation on one side, and the opposite dimple increases in depth, producing a bowtie-, then cup-shaped cell. Because such cells appear on fixed smears to have a mouthlike “stoma” instead of a round area of central pallor, they are known as stomatocytes. As the change progresses, the cells become spherostomatocytes and, finally, spherocytes, with only a small hilum remaining at the former site of the stoma. Such cells lack the small spicules that characterize the end-stage prelytic spherocyte resulting from discocyte–echinocyte transformation.
Although many circumstances can lead to transformation of the red cell from a disc to a sphere, many of them appear to act through effects on the erythrocyte plasma membrane anion transporter (AE1, described more fully later) (22). AE1 also provides a critical linkage to the membrane skeleton, and it is the interrelationship of these two functions that is proposed to account for these shape transformations. Spherocytic transformation is linked to the ratio of chloride ion influx and efflux (Cl-i/Cl-o), and many if not all conditions that lead to spherocytic transformation alter this ratio. It is thus theorized that the same conformational change that determines whether AE1 facilitates the inward or outward flow of Cl- also alters the interaction of AE1 with ankyrin (and thus with spectrin), major components of the erythocyte cytoskeleton.
|
|
|
Figure 7.4. Human erythrocytes flowing through a vessel about 7μm in diameter. Direction of flow is indicated by the arrow. (From Skalak R, Branemark P-I. Deformation of red blood cells in capillaries. Science 1969;164:717, with permission.) |
|
|
|
Figure 7.5. Shape transformations of the discoid erythrocyte. Erythrocytes may undergo transformation from a biconcave disc to an echinocyte or to a stomatocyte. Each transformation occurs via different intermediate shape changes. (From Bull BS, Brailsford D. Red blood cell shape. In: Agre P, Parker JC, eds. Red blood cell membranes—structure, function, clinical implications. New York: Marcel Dekker, 1989, with permission.) |
Thus far, the factors acting within the membrane to maintain its elastic properties, as well as those that act on it, are still only partially understood. Among the factors that affect membrane deformability and stability are membrane lipid content, cytoskeletal proteins, and transmembrane proteins.
A large body of evidence supports the role of lipid type and distribution in maintenance of membrane integrity and normal function, and abnormalities in distribution of lipids between the inner and outer leaflets are associated with changes in mechanical properties of membranes as well as with the strength of association between the membrane and the cytoskeleton (23). The cytoskeleton, formed by a latticelike network of proteins, undoubtedly contributes to the bending energy necessary for assumption of the biconcave shape, as well as to membrane stability (24). Abnormalities in cytoskeletal proteins cause a variety of pathologically shaped red cells, including spherocytes and elliptocytes (25). Abnormalities of both expression and function of transmembrane proteins, such as AE1, also affect membrane mechanical properties (26,27). In addition, proteins adsorbed to the outer surface of the red cell, especially albumin, may also play a role in both maintaining normal cell shape and effecting changes in that shape under some conditions. Red cells suspended in isotonic medium tend toward an echinocytic shape until albumin is added, and increasing amounts of albumin move cells toward the discoid shape (28).
The Erythrocyte Membrane and Cytoskeleton
Membrane Structure
The central feature of membrane structure is a matrix formed by a double layer of phospholipids. The lipid bilayer hypothesis, first proposed in 1925 (29) and refined by Danielli and Davson in 1935 (30), is now generally accepted (10,31). Lipid molecules in the bilayer are oriented with the nonpolar groups directed toward one another, forming hydrophobic interactions. The hydrophilic polar head groups are directed outward, where they interact with the aqueous environment on both the cytoplasmic and plasma surfaces.
The best accepted concept of how proteins fit into the lipid membrane structure is the so-called fluid mosaic model (32). The lipid bilayer may be thought of as a two-dimensional viscous solution. Within this “sea of lipids” float globular proteins, some that penetrate the membrane completely and others that penetrate the membrane only partially and may be exposed at only one surface. Some proteins appear to have considerable lateral mobility, but in the red cell, many proteins interact with other membrane components, giving them a degree of immobility. Some proteins traverse the lipid bilayer once, whereas others have multiple membrane-spanning domains. On the cytoplasmic side of the membrane lies a network of structural proteins that form a cytoskeleton. Certain membrane-spanning proteins appear to interact with various cytoskeletal proteins (33). Some transmembrane proteins also appear to become covalently linked to lipid (34), and the so-called glycosylphosphatidylinositol-anchored class of proteins has no membrane-spanning domain but instead has phospholipid “tails” by which they are attached to the membrane (35).
Ultrastructure
In thin sections of erythrocyte membrane, fixed with either osmium tetroxide or potassium permanganate, three distinct layers are observed. Two electron-dense (osmophilic) layers approximately 2.5 nm (25 Å) in thickness are separated by an electron-penetrable layer about 2.0 nm thick, for a total thickness of some 7.0 nm (36). This appearance has often been cited in support of the lipid bilayer structure, with the electron-dense areas representing either membrane protein layers or the polar ends of the phospholipids (37).
With air-dried, metal-shadowed red cell membranes, features of the surface are made apparent (38). In such preparations, plaques about 3.0 nm thick and 10 to 50 nm in diameter are randomly distributed over the surface. These observations have been used to suggest the existence of lipid-protein subunits.
Still another technique used in electron microscopic analysis of membranes is that of “freeze-cleaving” (39). Erythrocytes are frozen rapidly at -150°C and fractured with a razor blade. The cleavage plane follows pathways of least resistance, often exposing large areas of membrane. These surfaces are replicated with condensed carbon and platinum, and the replicas are examined with the electron microscope. Two types of membrane surfaces are observed with this technique. Both surfaces are characterized by the presence of particles approximately 10 nm in diameter. The two surfaces differ in that one has four to five times as many of these particles as the other (2,600 to 3,800/μm2, compared with 575 to 1,400/μm2). Membrane cleavage likely occurs in the nonpolar region between the two lipid layers, the particles representing proteins suspended in the lipid layer, as predicted by the fluid mosaic model.
The “quick freezing” and “deep etching” method of examining unfixed erythrocytes by electron microscopy has succeeded in providing a three-dimensional picture of the membrane cytoskeleton, whose proteins are organized into structured but deformable arrays (40).
Chemical Composition of the Membrane
Much that is known about red cell membranes is derived from studies of the insoluble portion of the cell remaining after hemolysis induced by osmotic rupture. This material has been called stroma and, if the membrane remains intact after hemolysis, red cell “ghosts.” It consists largely of components of the membrane, including the cytoskeleton. With careful attention to pH and osmolarity, one can prepare ghosts with only small amounts of residual hemoglobin (41,44). About 230 to 300 mg of such relatively hemoglobin-free material can be extracted from 0.1 L of erythrocytes. Such preparations contain about 52% protein, 40% lipid, and 8% carbohydrate by weight (41). Most of the carbohydrate is accounted for by the oligosaccharide portion of glycoproteins, but a small fraction (about 7%) is carried by glycolipids (41,42).
Lipid Composition
Virtually all of the lipids in the mature erythrocyte are found in the membrane (43). Qualitative and quantitative analyses have been performed, and the data have been the subject of several reviews (41,42,44,45,46). These results are summarized in Table 7.1.
The majority of erythrocyte membrane lipids are either phospholipids or unesterified cholesterol, which are present in approximately equimolar quantities. Four classes of compounds account for most of the phospholipid: phosphatidylcholine (lecithin), phosphatidylethanolamine, sphingomyelin, and phosphatidylserine (Table 7.1). Two fatty acid side chains are attached to all of these lipids except sphingomyelin, which has only one. In addition, trace amounts of other phospholipids containing only one fatty acid (“lysophospholipids,” e.g., lysolecithin) or having a vinyl ether (plasmalogens) in place of a fatty acid are found.
|
Table 7.1 Lipids of the Normal Human Erythrocyte Membrane |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Phospholipids are distributed asymmetrically between the two lipid layers of the membrane (47,48). Eighty percent or more of the aminophosphatides (phosphatidylethanolamine and phosphatidylserine) lie within the inner (cytoplasmic) monolayer, whereas the choline-containing lipids (phosphatidylcholine and sphingomyelin) are the major components of the outer monolayer (Fig. 7.6). Little or no phosphatidylserine is detectable in the outer lipid layer of normal, nonsenescent red cells. The functional importance of the asymmetric distribution of phospholipids in the two erythrocyte membrane lipid layers is not completely understood, although study of certain pathologic conditions and experimental models has provided clues to both the role and regulation of this asymmetry. For example, when erythrocytes containing hemoglobin S and little or no hemoglobin A are subjected to reduced oxygen tension, the outer lipid layer shows increased amounts of phosphatidylethanolamine and phosphatidylserine, whereas the distribution of sphingomyelin is unchanged (42). The change in lipid sidedness correlates with loss of the ability to undergo echinocyte transformation, implicating a role for lipid distribution in regulating shape and also implying that the regulation of distribution of some phospholipids may be independent of the regulation of distribution of others. Maintenance of normal asymmetry results in improved mechanical membrane stability under applied shear stress and further appears to supply additional means for cytoskeleton attachment to the lipid bilayer through spectrin– phosphatidylserine interaction (27).
The lateral mobility of lipids in the outer membrane layer exceeds that of lipids in the inner layer. Although cholesterol is known to restrict lipid lateral mobility (49), the outer lipid layer likely is relatively enriched in cholesterol (50). Lipids in the inner layer may be restricted in their mobility because of interactions of phospholipids with cytoskeletal proteins (27,51,52).
An additional effect on lipid mobility and membrane deformability may come from the fact that the fatty acids found in erythrocyte phospholipids also are not distributed evenly between the two bilayers (47,53,54,55). Overall, about one half of the fatty acids in the membrane are unsaturated (Table 7.2). Unsaturated fatty acids, however, and particularly the polyunsaturated acyl chains with four or more double bonds, are a disproportionately large part of the inner leaflet phospholipids, phosphatidylethanolamine and phosphatidylserine. In contrast, phosphatidylcholine, which is predominantly in the outer lipid layer, contains most of the shorter-chain saturated fatty acids. Sphingomyelin is especially enriched in fatty acids, with a chain length longer than 20. Membranes rich in sphingomyelin are less “fluid” than those with relatively larger amounts of lecithin (56). An increased ratio of sphingomyelin to lecithin is found in abetalipoproteinemia and probably accounts for the erythrocyte abnormalities associated with that disorder (53).
|
|
|
Figure 7.6. Distribution of erythrocyte phospholipids between the inner and outer layers of the membrane. PC, phosphatidylcholine; PE, phosphatidylethanolamine; PS, phosphatidylserine; SM, sphingomyelin; TPL, total phospholipids. (From Rothman JE, Lenard J. Membrane asymmetry. Science 1977;195:743, with permission. Copyright 1977 by the American Association for the Advancement of Science.) |
The neutral lipid of the erythrocyte consists almost entirely of free, nonesterified cholesterol (56). The distribution of cholesterol in the two membrane layers has been extremely difficult to study, at least in part because the translocation rate of cholesterol between the two layers is extremely rapid. Although findings in one report suggested that cholesterol is somewhat overrepresented in the outer lipid layer (51), further data need to be gathered. Cholesterol has a pronounced effect on membrane fluidity (57,58). It interacts with phospholipids to form what has been called an “intermediate gel state.” Thus, compared with pure phospholipid membranes, membranes containing cholesterol are less fluid, that is, more viscous. Relatively modest increases in membrane cholesterol content decrease membrane deformability (59). Abnormally high levels of cholesterol lead to distortions in red cell shape; bizarre spicules form (“spur cells”), deformability of the cells is reduced, and they are destroyed in the spleen.
Glycolipids (or glycosphingolipids) make up a small fraction of the total lipids of the erythrocyte membrane (60). These glyco-lipids resemble sphingomyelin in that the lipid base is a unit known as ceramide, consisting of sphingosine and a long-chain fatty acid. Attached to the base is a variable number of hexose molecules. The predominant erythrocyte glycolipid, globoside (GL-4), contains four hexose units:
Ceramide–glucose–galactose–galactose–N-acetylgalactosamine
Globoside appears to be a component only of plasma membranes and is not present in the membranes of intracellular organelles, such as mitochondria. In this respect, it differs from cholesterol and the phospholipids, which are distributed widely. Globoside is particularly characteristic of the red cell membrane, but it has also been isolated from plasma membranes of several other cell types.
Normal human red cells also contain ceramide glycolipids with one glucose (GL-1), one glucose and one galactose (GL-2), and one glucose and two galactose (GL-3) molecules attached. The relative proportions of the glycolipids are as follows: GL-4, 69%; GL-3, 12%; GL-2, 14%; and GL-1, 5% (60). In addition, fucose-containing ceramide glycolipids are present in trace amounts. These are of interest because they are surface structures that have antigenic activity corresponding to the A, B, H, and Lewis blood groups. They are not the exclusive bearers of these antigens, however, because glycoproteins with similar hexose arrangements have similar antigenic properties.
Lipid Turnover and Acquisition
The mature erythrocyte is unable to synthesize lipids de novo; therefore, any lipid loss must be compensated for by renewal from pathways of interchange with the plasma (Fig. 7.7) (55). Quantitatively, the most important of these pathways are the transfer of cholesterol and phosphatidylcholine (lecithin) from plasma lipoproteins to red cells (pathways 1 and 3). The rates of transfer are functions of the relative plasma and red cell levels of these lipids and are indirectly affected by the activity of the cholesterol-esterifying enzyme in plasma, lecithin-cholesterol acyltransferase (LCAT) (61). This enzyme catalyzes the reaction in which a fatty acid in the 2 position of lecithin is transferred to free cholesterol, forming cholesterol ester and lysolecithin (Fig. 7.7, reaction 1A). Neither of the LCAT reaction products can enter the membrane. In patients with congenital LCAT deficiency, membrane cholesterol and lecithin are increased and the red cells are target-shaped (62).
The exchange of cholesterol and lecithin between red cells and plasma is also affected by the plasma bile salt concentration (63). If erythrocytes are incubated in normal plasma to which bile salts have been added, the cells acquire cholesterol, and this change is accompanied by an increase in surface area and the formation of target cells. Although the mechanism of bile salt action is not fully understood, at least two properties appear to be important: Bile salts inhibit the LCAT reaction and, in addition, they bring about a shift in the distribution of free cholesterol between plasma and cell.
Phospholipids also may be added to the membrane by three other reactions. Albumin-bound lysophospholipid may be transferred to the membrane (Fig. 7.7, pathway 4) and acylated (reactions 5a and 6a) to form a complete phospholipid (55). Of lesser quantitative importance is the transfer and conjugation of two lysophospholipid molecules to yield a phospholipid (reaction 5b) and glycerylphosphorylcholine (GPC), which returns to the plasma (64). Finally, some phosphatidylethanolamine (PE) is produced by transacylation of a fatty acid from lecithin to lyso-PE (reaction 6b). A congenital defect in the last reaction results in the formation of cells that possess increased membrane lecithin and decreased membrane PE (65). These changes are associated with the clinical picture of nonspherocytic hemolytic anemia.
|
|
|
Figure 7.7. Pathways of lipid acquisition and turnover in the mature red cell membrane. Reactions and pathways: (1) exchange of cholesterol with plasma lipoprotein; (1A) the LCAT reaction; (2a) transfer of FFA from albumin to membrane; (2b) penetration of FFA to a metabolically active site; (3) exchange of PC with plasma lipoprotein; (4) transfer of LPC from albumin to membrane; (5A) LPC + FFA → PC; (5b) 2LPC → FFA + GPC; (5c) LPC → FFA + GPC; (6a) LPE + FFA → PE; (6b) PC + LPE → LPC + PE; (7) PE → LPE + FFA; (7b) PC → LPC + FFA. Alb, albumin; C, cholesterol; CE, cholesterol ester; FFA, free fatty acid; GPC, glycerylphosphoryl choline; LCAT, lecithin-cholesterol acyltransferase; LPC, lysophosphatidyl choline; LPE, lysophosphatidyl ethanolamine; PC, phosphatidyl choline (lecithin); PE, phosphatidyl ethanolamine. (From Shohet SB. Hemolysis and changes in erythrocyte membrane lipids. N Engl J Med 1972;286:577, with permission.) |
The fatty acid composition, but not the major relative proportions of the phospholipid classes, may be altered by diet (66,67). With low-fat diets, linoleic acid levels decrease. With diets high in linoleic acid, the amount of red cell linoleic acid increases. These changes occur relatively slowly, over about 4 to 6 weeks. The levels of myristic, arachidonic, and oleic acids in plasma membrane lipids are also affected by diet.
Membrane Proteins
Early in the history of membrane biochemistry, erythrocytes became the model system for the study of plasma membranes because they lacked organelle and nuclear membranes, making the membranes easy to isolate. Solubilization of membrane proteins can be accomplished by the addition of detergent. Detergent usually is added to red cell ghosts, made relatively hemoglobin-free after hypotonic lysis of red cells (43). However, different erythrocyte membrane proteins are variably soluble in detergents. Nonionic detergents, which do not denature most proteins, efficiently solubilize some membrane proteins but leave several others, especially those with strong attachments to the cytoskeleton, unsolubilized. Use of an ionic detergent, most commonly sodium dodecyl sulfate (SDS), does accomplish solubilization of essentially all membrane proteins, although at the cost of denaturing the proteins, in some cases irreversibly.
Proteins extracted from membranes by detergent solubilization can be separated and analyzed with relatively high resolution by means of electrophoresis in polyacrylamide gels (60,68). Such gels are stained by protein stains, usually Coomassie Brilliant Blue or silver stain, or by reagents that react with carbohydrate, such as periodic acid-Schiff (PAS). Alternatively, proteins are first labeled with a variety of radioactive reagents, and then the gels are exposed to x-ray film, a process known as autoradiography. One radiolabeling method is radioiodination of tyrosine residues performed using intact erythrocytes. This method radiolabels many of the membrane proteins that have extracellular domains but fails, naturally, to label cytoskeletal proteins. Failure to be radioiodine-labeled also occurs, however, among membrane proteins such as glycophorin C, because they lack available extracellular tyrosine residues. Each of these methods thus yields a different picture of membrane proteins, because different proteins are stained or radiolabeled by each of these methods, and no method identifies all proteins. Additional methods of identifying proteins in gels include coupling of surface structures to small molecules such as avidin (which can then be stained using a biotin conjugate) or 3H-containing sugars (a nontyrosine but carbohydrate-dependent method of radiolabeling). Finally, many important membrane proteins are not expressed in sufficient quantity to be distinguishable among the many proteins of the membrane, although immunologic methods of detecting them, such as Western blotting (69), can readily identify many of these.
|
|
|
Table 7.3 Major Membrane Polypeptides |
|
|
|
Figure 7.8. Polyacrylamide gel electrophoresis of erythrocyte membrane proteins. In this system, polypeptides migrate according to molecular size, with the smaller molecules moving the farthest. Gels were stained with Coomassie Brilliant Blue (left) and periodic acid-Schiff (PAS) reaction (right). (From Steck TL. The organization of proteins in the human red blood cell membrane. J Cell Biol 1974;62:1, with permission.) |
On Coomassie-stained gels, seven major bands are usually identified, whereas PAS stains four major bands and several minor ones (Fig. 7.8). Originally, the seven major protein bands were referred to by their numeric designations (Table 7.3) As further refinements in SDS-PAGE have produced greater resolution, other bands have been given decimal or alphanumeric designations, such as bands 2.1 and 4.2, sometimes with even further divisions such as in 4.1a and 4.1b. At present, many of these proteins are no longer identified by this numeric nomenclature because they have been given specific names as their chemical structures have been defined and their cDNAs and genes have been cloned.
|
Table 7.4 Nomenclatures of PAS-Staining Erythrocyte Glycoproteins |
|||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||
Likewise, the PAS-stainable proteins are now well characterized. The major PAS-stained bands contain three proteins, two of which form both hetero- and homomultimers. These three proteins have been termed glycophorins, or sialoglycoproteins, with letter designations that vary according to the investigator (see Table 7.4) (68,70,71,72,73). Glycophorins A and B are proteins that are derived from highly homologous genes and form both homo- and heterodimers. Glycophorins C and D are two proteins produced by a single gene and are structurally unrelated to glycophorins A and B.
Historically, membrane proteins were thus first characterized by whether they were stainable by protein-binding or carbohydrate-specific dyes. Now, however, they are classified on the basis of their relationship to the membrane or their functions. One common classification of membrane proteins comprises the categories of integral membrane proteins and peripheral membrane proteins (32). Integral membrane proteins are most often globular and amphipathic; in their folded, three-dimensional form, they have distinct hydrophobic and hydrophilic domains. Of the major Coomassie-stainable proteins, only bands 3, 4.5, and 7 are integral membrane proteins. These proteins have one or more membrane-spanning domains. Band 3 has several extracellular domains, of which some are highly glycosylated. Band 4.5 is the glucose transporter. Less is known about band 7, a protein that has a small extracellular domain and relatively larger cytoplasmic domain. All other Coomassie-stained bands are situated within the cell, either as part of the cytoskeleton or bound in a more or less loose fashion to the inner leaflet of the membrane. All the PAS-stainable proteins are integral membrane proteins. Integral membrane proteins require detergent for removal from the membrane, whereas peripheral membrane proteins can often be extracted from “ghosts” by manipulation of the pH and ion content of buffers and tend to be soluble in neutral aqueous buffers. Proteins attached by phosphatidylinositol anchors to the outer membrane layer also require detergents or other reagents capable of disrupting the lipid bilayer for solubilization (35).
Transmembrane Proteins
The two predominant erythrocyte transmembrane proteins are glycophorin A (GPA) and the anion channel (AE1, formerly known as Band 3). However, whereas AE1 appears to play a number of crucial roles in red cell biology, the importance of GPA is less clear.
Glycophorin A is the principal PAS-stainable glycoprotein of erythrocyte membranes. There are between 0.5 and 1 × 106 copies of this protein per erythrocyte (74), accounting for approximately 85% of PAS-positive membrane protein. When analyzed by SDS-PAGE, GPA forms homodimers, heterodimers with glycophorin B, and higher-order multimers of up to several hundred kilodaltons. The monomeric form migrates anomalously, with an apparent molecular weight of 36,000 to 39,000, presumably because of its highly glycosylated state; approximately 60% of its weight is attributable to carbohydrate. Most of the carbohydrate is in the form of 15 O-glycosidically linked tetrasaccharides. The two sialic acid residues of each of these many O-glycosidically linked oligosaccharides account for 60% of the surface negative charge of the red cell. The other two components of these tetrasaccharides are one N-acetyl-galactosamine residue and one galactose residue. In addition, GPA bears one complex N-glycosidically linked oligosaccharide. GPA also bears blood group antigens. The N terminus of GPA bears the M or N antigen, depending on whether serine and glycine or leucine and glutamic acid are in amino acid positions 1 and 5, respectively. Although the antigenic polymorphism of the MN system thus depends on amino acid sequence differences, many human antibodies to these antigens do not recognize these antigens if sialic acid has been removed from the three O-linked oligosaccharides that are normally attached to amino acids 2, 3, and 4. GPA has also been found to be a binding site for several pathogens, including Plasmodium falciparum.
The complete amino acid sequence of GPA was determined in 1978 (75), and several investigators have since cloned the cDNA and explored the genomic structure of the glycophorin genes. GPA comprises 131 amino acids, of which 70 reside extracellularly, 22 within a single membrane-spanning domain, and the rest within the cytoplasm. The gene for GPA resides on chromosome 4 (76) and is organized into six exons that encode the leader peptide and mature protein, along with a seventh exon that encodes 3′-untranslated mRNA (77,78). Numerous variants of GPA have been described and may cause production of alloantibodies after transfusion. Rarely, some persons lack this protein totally. Although absence of GPA causes no clinically significant hematologic problems, persons with this deficiency may make antibodies that render blood transfusion difficult, given the general unavailability of blood from other GPA-deficient donors. In addition, some evidence suggests that the cytoplasmic C-terminal portion of GPA attaches loosely to the cytoskeleton via the cytoskeletal protein band 4.1 (79,80) and also is expressed as part of a macromolecular complex that includes a dimer of AE1 (see later) and as many as six GPA dimers. In experimental systems, GPA also acts as a chaperone to facilitate AE1 expression (81), and in humans, absence of GPA is associated with altered glycosylation of AE1 (82).
Glycophorin B, the second most abundant PAS-staining protein, is present on the erythrocyte in one tenth to one third the copy number of GPA (74). Also encoded by a gene on chromosome 4, glycophorin B is highly homologous to GPA. The N terminus of glycophorin B carries the “N” amino acid sequence of GPA and continues to be largely homologous to GPA for the first 27 amino acids, so glycophorin B always expresses the N antigen. At amino acid 29, within a region of glycophorin B less homologous to GPA, expression of methionine or threonine accounts for the S and s blood group antigens, respectively. The intramembranous portion of glycophorin B is likewise highly homologous to that of GPA, but glycophorin B lacks a large cytoplasmic domain corresponding to that of GPA. Like GPA, glycophorin B is encoded by a single-copy gene on chromosome 4; the GPA and B genes thus appear to have arisen from a single ancestral gene. A third related gene, encoding a protein that is expressed weakly or not at all, has also been described belonging to this gene family.
AE1, which is the erythrocyte anion-channel or anion-exchange protein, is expressed on erythrocytes in about the same copy number as GPA. Human erythroid AE1 cDNA has been cloned (83,84). Deduced cDNA sequence, as well as confirmatory biochemical analyses, has demonstrated that AE1 most likely tra-verses the membrane 12 times. Although it has several extracellular domains, the fourth is the major bearer of carbohydrate. This domain is heavily glycosylated and bears carbohydrate blood group antigens, including I and i and the antigens of the ABO major blood group system. Variations in the degree of glycosylation of this domain among individual protein molecules apparently accounts for the broad range of molecular weight (95,000 to 105,000) deduced for this protein from SDS gels. Gross changes in glycosylation of AE1 are also seen in certain inherited red cell abnormalities: AE1 bears increased sialic acid when GPA is absent, whereas glycosylation of AE1 is reduced in HEMPAS (hereditary erythroblastic multinuclearity with positive acidified serum test). Within the membrane, AE1 exists predominantly as a dimer, possibly in a macromolecular complex that includes GPA and the Rh proteins (81,85,86). Its function appears to be Cl–HCO3 exchange. AE1 also interacts with the erythroid cytoskeleton by binding ankyrin and binds NO, possibly facilitating its transit across the erythrocyte plasma membrane (87). Complete absence of erythroid AE1 is rare in humans and has been associated with severe hemolytic anemia and renal tubular acidosis. However, many different mutations of the membranous or cytoplasmic domain of AE1 have been described as leading to hereditary spherocytosis. In all, mutations of the AE1 gene account for about 30% of all cases of hereditary spherocytosis in Caucasians (88,89).
The Rh proteins are also important integral membrane proteins. Although they are present in only 100,000 copies per cell (74), these proteins are clearly important both to erythrocyte biology as well as to transfusion medicine. The RhD protein is the most immunodominant determinant of red cells outside the ABO antigens and is thus the only non-ABO determinant routinely taken into account when blood is selected for nonalloimmunized recipients. Absence of all Rh proteins, as in the Rhnull syndrome, is associated with multiple erythrocyte defects and mild hemolytic anemia (90). The proteins that carry the D, C (or c), and E (or e) Rh antigens are highly homologous to one another (91) and traverse the membrane multiple times (91). Two Rh genes situated very near each other on chromosome 1 encode the proteins that bear Rh antigens; in the normal situation, one gene encodes the D antigen and the other encodes a protein that bears both the E/e and C/c antigens (92). However, in the membrane, the Rh proteins are part of a multimeric protein complex, including both structurally related and unrelated proteins (93). The function of this complex remains uncertain, although it has been shown that the Rhnull syndrome occurs most frequently as a result of lack of expression of Rh50, one of the proteins in this complex, rather than a defect at the Rh gene locus itself (94). Recombinant Rh50 expressed in yeast and in Xenopus oocytes has been shown to facilitate ammonium transport (95,96), although some evidence also suggests that Rh proteins play a role in CO2 transport (97).
|
|
|
Figure 7.9. Model of the relationship between integral and cytoskeletal membrane proteins. Cytoplasmic domains of several integral membrane proteins interact with several cytoskeletal proteins. (From Gardner K, Bennett GV. Recently identified erythrocyte membrane-skeletal proteins and interactions. In: Agre P, Parker JC, eds. Red blood cell membranes—structure, function, clinical implications. New York: Marcel Dekker, 1989, with permission.) |
Cytoskeletal Proteins
The most abundant of the peripheral proteins are those that make up the so-called spectrin-actin cytoskeletal complex. These proteins, which can be extracted in the presence of EDTA and other chelating agents, or by reducing ionic strength and raising pH, account for about 35% of the membrane protein. The complex includes large α and β spectrin polypeptide chains (bands 1 and 2 on gel electrophoresis, molecular weight about 240,000 and 225,000, respectively) and the smaller actin chain corresponding to band 5. The relationship between the integral and peripheral proteins of the membrane is illustrated in Figure 7.9. Study of the erythrocyte cytoskeleton has led to the realization that similar spectrin-based skeletal structures are important not only in preserving erythrocyte integrity in the face of the shear stresses of the circulatory system and spleen but also in the function of more complex nucleated cells (98,99).
|
|
|
Figure 7.10. The α and β subunits of human erythrocyte spectrin. Each protein has numerous repeated segment domains, and the proteins combine to form antiparallel heterodimers. Most repeat domains (represented as rectangles) are homologous and are exactly 106 amino acid residues in length; nonhomologous domains are represented by squares. Roman numerals represent domains identified historically as numbered from the spectrin self-association site (at left). Peptides produced by limited trypsin cleavage, often used to identify the many spectrin variants associated with abnormal erythrocyte shape, are indicated by “T,” followed by the molecular weight (in thousands) of the fragment. (From Speicher DW. The present status of erythrocyte spectrin structure: the 106-residue repetitive structure is a basic feature of an entire class of proteins. J Cell Biochem 1986;30:245.) |
Spectrin proteins are long, rod-shaped molecules that self-associate into a two-dimensional network with the help of many other cytoskeletal proteins (33). The two forms of spectrin are homologous to each other, and both are composed primarily of 106-amino acid repeats (Fig. 7.10); however, they are encoded by genes on two different chromosomes, chromosome 1 (α spectrin) and chromosome 14 (β spectrin) (100,101,102). Each spectrin molecule appears to be folded over on itself an average of three times. The α and β spectrin molecules form heterodimers by aligning in antiparallel pairs. These heterodimers then form tetramers by head-to-head association (33). Incorporation of these tetramers into the latticework of the cytoskeleton then occurs with the interaction of other peripheral membrane proteins (Fig. 7.9) (33).
Erythrocyte actin, or band 5 as identified in Coomassie-stained gels of erythrocyte membrane proteins (Fig. 7.8), is an abundant (about half a million copies) erythrocyte protein of about 45,000 daltons. Like actin in skeletal muscle cells, erythrocyte actin can polymerize into long filaments and can activate myosin ATPase activity. It seems more likely, however, that erythrocyte actin normally forms only short filaments (103). Actin filaments are known to associate with spectrin tetramers at the ends containing the carboxy terminus of the α chain and the amino terminus of the β chain. This association, however, is a low-affinity interaction in the absence of other accessory proteins.
In addition to spectrin and actin, other cytoskeletal proteins appear to play a crucial role in membrane stability and maintenance of cell shape. Protein 4.1, of which approximately 200,000 copies are expressed per cell, is a 78,000-dalton protein that contains a spectrin-binding domain and also appears to bind to the transmembrane protein glycophorin C (104). Protein 4.1 is also known to promote spectrin-actin interaction, although this protein requires the presence of actin in order to interact directly with spectrin. Band 3 may also bind protein 4.1 to the membrane, because the two proteins can interact in solution. Protein 2.1, now known as ankyrin, also serves as a mode of attachment of the cytoskeleton to the membrane (105). About 100,000 copies of ankyrin are expressed per cell; at least two forms of ankyrin, a 206,000-dalton form and a 190,000-dalton form, are present in membranes. The smaller form has also been designated band 2.2. Both forms contain an amino-terminal domain that binds the anion exchanger band 3 and a domain more toward the carboxy terminal that binds spectrin. A peptide within the carboxy-terminal domain of protein 2.1 is absent from protein 2.2 because of differential processing of mRNA; this relatively small difference increases the affinity with which ankyrin binds both spectrin and band 3. Other proteins, although present in smaller quantities, nevertheless also play vital roles in formation and stability of the cytoskeleton. These proteins include protein (band) 4.9, tropomyosin, tropomodulin, and adducin.
Deficiencies or abnormalities in some of these cytoskeletal proteins have been associated with abnormal erythrocyte shapes, abnormal membrane stability and rigidity, and hemolytic anemias. Thus, elliptocytosis, spherocytosis, and pyropoikilocytosis can result from a variety of defects of spectrin, ankyrin, protein 4.1, and glycophorin C (106).
Acquired abnormalities of the membrane cytoskeleton can also occur. In sickle cell anemia and (89) thalassemia, for example, oxidation of protein 4.1 appears to change the affinity of protein 4.1 for spectrin (107,108). Also, in blood stored for transfusion, spectrin oxidation appears to occur with time, leading to loss of membrane surface area by formation of lipid vesicles comprising membrane unattached to the cytoskeleton. Other cytoskeletal proteins may also be affected by oxidative processes that alter their ability to interact with members of the spectrin-actin meshwork underpinning the erythroid membrane.
Membrane Transport Proteins and Functions
In general, the membrane acts as a partial barrier to penetration of all solutes. Nonpolar substances diffuse through the membrane at a rate proportional to their solubility in organic solvents. Polar solutes appear to cross the membrane at specialized sites. The erythrocyte membrane has a number of specialized transport proteins, including the anion transporter (AE1, band 3), several cation transporters, a glucose transporter, a urea transporter, and a water channel.
Erythrocytes have an abundant and highly active water channel protein, aquaporin-1, which contributes as much as 85% of the osmotic water permeability pathway (109). Acquaporin-1, originally described as CHIP-28 (channel-forming integral protein-28 kDa), occurs as a homotetrameric protein that expresses on its extracellular domain both ABH and Colton blood group antigens and can be inhibited by a variety of mercurial compounds (110). Red cells that lack aquaporin-1 have only slightly reduced lifespan in vivo (109).
|
|
|
Figure 7.11. A: Intracellular electrolyte composition of the erythrocyte as compared with serum. (From Guest GM. Organic phosphates of the blood and mineral metabolism in diabetic acidosis. Am J Dis Child 1942;64:401.) B: Role of band 3 in anion and CO2 transport. The ability of band 3 to accelerate anion transport across the membrane allows rapid equilibration of bicarbonate with the extracellular plasma and concomitant influx of chloride ion. CA, carbonic anhydrase; Hb, hemoglobin. (From Kopito RR, Lodish HF. Structure of the murine anion exchange protein. J Cell Biochem 1985;29:1, with permission.) |
Anions appear to cross the membrane by one of two discrete pathways. The first represents an exchange reaction in which an internal anion is exchanged for an external anion. This rapid exchange is mediated by the band 3 anion-exchange protein and plays an important role in the chloride–bicarbonate exchanges that occur as the red cell moves between the lungs and tissues (Fig. 7.11B) (111). The second anion pathway represents considerably slower ionic diffusion, accounting for net loss or gain of anions. This process is about 100 times as rapid as anion diffusion but 10,000 times slower than the anion-exchange reaction.
Glucose and other monosaccharides constitute an important exception to the generalization regarding nonpolar solutes in that the monosaccharides easily cross the membrane barrier, whereas the more lipid-soluble disaccharides do not. The speed of the process depends on molecular structure. Only d isomers are transported; l isomers are not. Of the common d-hexoses, glucose is transported most rapidly, followed by mannose, galactose, xylose, and arabinose. The concentration of solute at the half-maximal transport rate (Km) is 6.2 mM for glucose and 18.5 mM for mannose. Fructose is not transported under physiologic conditions (Km > 200 mM).
Glucose enters the erythrocyte by facilitated diffusion, mediated by a transmembrane protein designated the glucose transporter, encoded by the gene GLUT1 I (112). This protein constitutes about 5% of erythrocyte membrane protein and accounts for a diffusely migrating band (band 4.5) in Coomassie-stained polyacrylamide gels of erythrocyte membrane proteins. (Although it is a single protein, the glucose transporter is heterogeneously glycosylated, causing its apparent molecular weight to range from 45,000 to 75,000.) The protein appears to occur in the membrane as a homotetramer. The glucose transporter has an asymmetric effect on glucose transit across the membrane: Glucose influx exhibits a higher Vmax and a lower Km than does glucose efflux. The transport of glucose into the erythrocyte provides the energy substrate for anaerobic glycolysis; however, the energy requirement of the erythrocyte appears to be relatively low, and the efficiency of glucose transport is relatively high. Therefore, glucose transport does not appear to be rate-limiting for glucose use (113,114). Interestingly, heterozygosity for a defective glucose transport protein (GluT1) has been reported to result in a heritable seizure disorder (115).
The erythrocyte membrane is only slightly permeable to the major monovalent cations, sodium and potassium, and their movement depends greatly on an energy-requiring transport mechanism. Within the human erythrocyte, potassium is the predominant cation and sodium is a relatively minor constituent, whereas the relationship is reversed in plasma (Fig. 7.11A). Cation concentrations within the erythrocyte are approximately 130 mM K+ and 8 mM Na+, whereas the plasma contains approximately 140 mM Na+ and 4 mM K+. The preservation of these gradients is the result of the cation transport process. The steady-state cation concentrations within the erythrocyte are the result of an equilibrium between passive diffusion (“leak”) and active transport (“pump”). With respect to sodium, the direction of leak is inward and the direction of pump is outward; in contrast, potassium leaks out and is pumped in. The major cation pump represents a process in which sodium inside the cell is exchanged for potassium on the outside and energy is supplied by ATP. For each molecule of ATP converted to ADP, three sodium ions are pumped out and two potassium ions enter (116). The overall reaction has been expressed as follows (117):
3Nai+ + 2Ko+ + ATPi → 3Nao+ + 2Ki+ + ADPi + Pi
in which subscript i indicates intracellular and subscript o indicates extracellular (outside). In this process, the erythrocyte experiences a net loss of one positive charge per cycle. Some of this loss may be compensated by outward movement of a cation (most likely Cl-), but this cation pump also does contribute to the low resting potential of human erythrocytes (118).
Active Na+ and K+ transport depends on the activity of the membrane protein Na-K ATPase. This protein may exist in the membrane as part of a multienzyme complex, including glyceraldehyde 3-phosphate dehydrogenase and phosphoglycerate kinase (119). The pump protein itself comprises two noncovalently linked polypeptides. The α subunit has a molecular weight of about 112,500 daltons and contains the catalytically active domain; the β subunit, which has a molecular weight of about 45,000 daltons, is likely to be important to regulation of the molecule’s function (120,121). A small proteolipid γ chain of about 10,000 daltons also copurifies with the protein but does not appear to be required for ATPase activity (122). The α subunit contains both ATP- and ouabain-binding, as well as ATPase activity. The complete amino acid structure for the α chain of human Na-K ATPase has been determined (123). It is compatible with a protein structure that includes eight membrane-spanning domains, making the α subunit an integral membrane protein, as would be expected of an active membrane transport protein. The primary structure of the β subunit has also been characterized (123). It is also an integral membrane protein, with at least one transmembrane domain. Most of the β-subunit protein appears to be extracellular, whereas the nonmembrane-spanning domains of the α subunit are primarily cytoplasmic. In the membrane, the Na-K ATPase may exist as oligomers of the α, β, and γ units in a ratio of 1:1:1.
The erythrocyte also has a urea transporter that transports urea rapidly across the membrane and helps preserve red cell osmotic stability and deformability (124). Although this transporter has similar characteristics to the renal urea transporter, it appears to arise from a related but distinct gene (125,126).
Membrane and Membrane-Associated Enzymes
At least 50 enzymes are either membrane proteins or are bound to the erythrocyte membrane in some fashion; certain enzymes are both free in the cytoplasm as well as associated with the membrane. Their functions range from facilitating transport of a variety of molecules necessary to the erythrocyte to playing important roles in producing and using energy from glucose metabolism.
Some erythrocyte enzymes are externally oriented and can therefore react with substrates in the red cell environment. Certain hydrolytic enzymes, including glycosidases and acid phosphatases, are among the externally oriented enzymes (127). A classic example of an externally oriented enzyme is acetylcholinesterase. Acetylcholinesterase was the first of several membrane proteins discovered to be missing from the affected erythrocytes of patients with paroxysmal nocturnal hemoglobinuria (see Chapter 35) (128,129). It belongs to a class of proteins that are attached to the membrane by a phosphatidylinositol-glycan anchor, so that the entire polypeptide portion of the molecule is extracellular (35). Although the role of acetylcholinesterase on the red cell remains obscure (130), some other proteins in this class are complement regulatory proteins; it is the absence of these proteins that causes the classic hemolysis of paroxysmal nocturnal hemoglobinuria (129).
Among the enzymes required for the production and use of ATP are three enzymes that are thought to form a membrane-bound enzyme complex: aldolase, glyceraldehyde 3-phosphate dehydrogenase (G3PD), and phosphoglycerate kinase. Together, these three enzymes convert fructose diphosphate to 3-phosphoglycerate with the production of ATP. G3PD is the enzyme present in greatest amount in membrane preparations and is seen as band 6 in polyacrylamide gels. G3PD is also found in the erythrocyte cytoplasm and can be demonstrated to bind to a cytoplasmic segment of band 3. Although the exact physiologic role of membrane binding for this enzyme remains unclear, its possible role in regulating metabolic responses to cell injury and stress is intriguing (131).
ATP is not only generated by membrane-bound enzymes, it also is used by membrane-bound molecules, among which are adenylyl cyclase, which catalyzes the conversion of ATP to cyclic AMP (cAMP), protein kinases, and adenosine triphosphatases (ATPases). The latter are involved in several transport functions, as was discussed previously.
Protein kinases are enzymes that phosphorylate other proteins in the presence of ATP by forming phosphoserine, or phosphothreonine bonds. Phosphorylation is a major step in the regulation of a variety of target molecules, including structural proteins and enzymes. Erythrocytes contain numerous protein kinases, including both cytosolic and membrane-bound cAMP-dependent kinases, cytosolic and membrane-bound cAMP-independent protein kinases, protein kinase C, and a calcium-regulated protein kinase. Both membrane-bound and cytosolic kinases may phosphorylate membrane proteins (132). In general, phosphorylated structural proteins demonstrate lower-affinity binding to their target proteins than do unphosphorylated proteins. For example, phosphorylation of protein 4.1 leads to a decreased affinity for spectrin and a decreased ability of 4.1 to promote spectrin–actin association (33). Phosphorylation of spectrin, however, leads to little if any change in spectrin self-association or association of spectrin with other molecules, such as ankyrin and actin (33). Exactly how such processes regulate cell shape and membrane integrity has not been well worked out. In the case of protein 4.1, dephosphorylation because of ATP depletion, an event that might be associated with stress, would be predicted to lead to a more rigid spectrin-actin network and reduced membrane deformability (133). Phosphorylation of enzymes can likewise lead to activation of a variety of metabolic pathways; often, phosphorylation of one molecule leads to activation of both an enzyme system and the molecules that down-regulate activity of that system.
Enzymes that use and degrade ATP are also found in the membrane, although they are not present in large enough quantities to account for bands seen in Coomassie-stained polyacrylamide gels of membrane proteins. Like protein kinases, ATPases phosphorylate membrane proteins, but instead of forming phosphoserine or phosphothreonine bonds, they form acyl bonds as transient intermediates in the catalytic cycle. Important and well-studied ATPases of the erythrocyte membrane include Na-K ATPase, Ca-Mg ATPase, and Mg ATPase. The Na-K ATPase is also known as the sodium pump, or sodium-potassium pump (124). This enzyme is responsible for maintaining the high internal potassium and low internal sodium concentration of both erythrocytes and other mammalian cells. In the process of regulating the concentration of these cations, Na-K ATPase also affects regulation of cellular volume and maintains the potassium-dependent activation of some intracellular enzymes (134). Na-K ATPase is also known as the ouabain-inhibitable ATPase, because the enzyme is specifically sensitive to inhibition by ouabain.
Finally, the mature erythrocyte is capable of responding to its environment through classical signaling mechanisms. It has a wide range of signaling molecules (135) and is affected by exogenous stimuli such as epinephrine.(136,137). An adrenergic signaling mechanism has been implicated in the process that produces abnormal adhesion of sickle red cells to endothelial components (138).
Hemoglobin and Erythrocyte Function
Hemoproteins have been utilized by nature for processes as diverse as electron transfer down electrochemical gradients in the respiratory chain to the harvesting of light energy in photosynthesis. Hemoglobins are one of the most widespread and specialized hemoproteins existing in nature and have been found in prokaryotes, fungi, plants, and animals. These proteins permit the reversible binding of O2 to heme while keeping the iron in the +2 state. They also facilitate the exchange of carbon dioxide between the lungs and the tissues. Recent studies have also demonstrated the importance of hemoglobin in control of vascular tone mediated by NO. In vertebrates, hemoglobin is the major constituent of the red cell cytoplasm, accounting for about 90% of the dry weight of the mature cell.
In most invertebrates, oxygen-carrying pigment is transported freely in the plasma rather than within cells. This is a relatively inefficient delivery system. Hemoglobin, as a protein free in the plasma, would exert an osmotic pressure about five times greater than that produced by the plasma proteins. By the inclusion of this pigment in corpuscles, the viscosity of the blood can be maintained at a low level, water is not drawn from the tissues by it, and the flow of blood containing such a large amount of protein is made possible. Furthermore, free hemoglobin is not maintained in the circulation and is subject to oxidative denaturation. Attempts to make hemoglobin substitutes have revealed that infusion of free hemoglobin or derivatives causes a significant increase in blood pressure, as a result of the scavenging of NO produced by vascular endothelium.
In a human at rest, about 250 ml of oxygen are consumed and 200 ml of carbon dioxide are produced per minute. During exercise, these quantities increase 10-fold. If the respiratory gases were carried in physical solution in the plasma, human activity would be restricted to only one fiftieth of that possible in the presence of hemoglobin-containing red cells. Hemoglobin permits the transportation of 100 times more oxygen than could be carried by the plasma alone.
Evolution and Structure of Hemoglobin
Vertebrate hemoglobin is a conjugated protein with a molecular weight near 64,500 daltons. The hemoglobin molecule is roughly spherical, with a maximum molecular diameter of about 6.4 nm. It is a tetramer, consisting of two pairs of similar polypeptide chains called globins, that exhibits a diad axis of symmetry. To each of the four chains is attached a prosthetic group, heme, which is a complex of iron and protoporphyrin (Fig. 7.12).
|
|
|
Figure 7.12. Chemical structure of heme and its manner of union with globin to form hemoglobin. The carbon atoms derived from the α carbon of glycine are represented by •▴, those supplied from the methyl carbon of acetate by ▴, those derived from the carboxyl group of acetate by ×. The unmarked carbons are those derived from either the methyl carbon atom of acetate or from the carboxyl atom. (Prepared by Dr. G. E. Cartwright.) |
Human hemoglobins share a common ancestry with a simpler, single-chain molecule that was similar to myoglobin. The divergence of the invertebrate and vertebrate globin genes occurred >670 millions years ago, and the divergence of the α and non-α globin genes probably descended from a common gene >450 million years ago (139). Such an evolution would explain the high degree of homology between the α and non-α (∊, γ, β, and δ) chains, as well as the extraordinary similarities among the non-α globins. The non-α globin gene family is sometimes designated the β-globin gene family. The α and non-α peptide chains most likely arose because of gene duplication, after which time the genes for these individual proteins evolved independently (140). Likewise, the β, δ, γ, and ∊ globins probably also arose as a result of gene duplications (141). The occurrence of subunit cooperativity brought with it the advantage of increased physiologic effectiveness of the hemoglobin molecule and added a new pressure on further evolution of the hemoglobin chains (142).
Ontogeny of Hemoglobins
Erythroid development is divided into three developmental periods: embryonic, fetal, and adult. In each developmental period, the oxygen delivery requirements are different, and erythroid development has evolved to meet these needs. The globins of vertebrates belong to two families that are approximately 50% identical with each other. The genes that encode the α-globin gene family are on chromosome 16, whereas the genes that encode the members of the non-α-globin gene family are on chromosome 11 (143,144). These genes are developmentally and coordinately regulated, and through the production of different pairs of globins, different hemoglobins are produced to permit their appropriate expression during different developmental periods. The α-globin gene cluster is capable of producing two types of globins, zeta (ζ) and α. The ζ is an embryonic globin chain produced during the first 8 weeks of fetal development, whereas α is produced during the remainder of the fetal and adult developmental periods (145). The β-globin gene family members include the embryonic globin ∊, the fetal globin γ, and two globins expressed primarily during the adult period, δ and β.
The transition from one developmental period to another is accompanied by the sequential and coordinate expression of the genes in each family. During early embryonic development, the expression of ζ- and ∊-globin genes leads to the production and assembly of these globins into the embryonic hemoglobin Gower-1 (ζ2∊2) (146). α and γ globins are also produced at low levels during this period, and this permits the production of two other embryonic hemoglobins, designated Gower-2 and Portland (147). Gower-2 is composed of α- and ∊-globin chains (α2∊2), and Portland is assembled from ζ and γ chains (ζ2γ2). These embryonic hemoglobins display subunit cooperativity and serve as physiologic oxygen carriers in erythroid cells that were derived from yolk-sac hematopoietic progenitors (148). Embryonic erythrocytes carrying these hemoglobins have an affinity for oxygen that is similar to fetal blood.
After the eighth week of development, erythropoiesis shifts from the yolk sac to the fetal liver, and the embryonic hemoglobins normally are not detectable in fetal blood (149). Embryonic hemoglobins are replaced with hemoglobin F (α2γ2), which remains as the predominant hemoglobin until after birth. The γ chains are encoded by pairs of genes located near the normal β-globin gene on chromosome 11. The two γ genes encode nearly identical proteins: Gγ has a glycine at codon position, whereas Aγ has an alanine (150,151). In addition, many Aγ genes also encode a threonine-for-isoleucine substitution at position 75 of the protein (152,153). During fetal life, Gγ constitutes about 75% of γ chains, whereas hemoglobin F in adults contains about 60% Aγ chains (154,155). This has no known physiologic significance.
Red cells containing hemoglobin F have higher oxygen affinity than adult red cells. This permits the fetus to compete effectively for oxygen in the maternal blood. However, hemoglobin lysates from adult and fetal cells have nearly identical oxygen affinity when they are dialyzed against saline or a neutral buffer (156). This property of fetal hemoglobin is due to amino acid differences in the amino terminus of the γ chains that impair binding of 2,3,-DPG, an allosteric modifier of oxygen binding (157). Hemoglobin F also has an enhanced alkaline Bohr effect (158), whereby oxygen affinity is increased as hemoglobin passes through the pulmonary vasculature.
At about 20 weeks of fetal development, the site of erythropoiesis begins to switch from the liver and spleen to the bone marrow, where progenitors show increased expression of adult globins, α and β. Hemoglobin A may constitute 5% of β-family globin expression during this time. Beginning at the 30th week and proceeding to the time of birth, a significant switch from fetal to adult erythropoiesis takes place, such that at the time of birth, fetal hemoglobin constitutes approximately 80% of the total hemoglobin. Over the next 25 to 30 weeks following birth, fetal hemoglobin concentration decreases by approximately 10% every 2 weeks until it reaches its normal adult level of <2% by 30 weeks of age (159). Neonates with hemoglobinopathies or erythropoietic stress can have a greatly prolonged production of hemoglobin F, sometimes extending into adulthood (160). The proportion of hemoglobins produced during the different developmental periods is summarized in Table 7.5.
Hemoglobin A, α2β2, is the predominant adult hemoglobin and normally constitutes approximately 96% of the total adult hemoglobin. A minor adult hemoglobin, A2, is produced beginning at 35 weeks of gestation but has little physiologic relevance. Hemoglobin A2 is composed of α globins and the minor adult globin δ. It normally constitutes <3.5% of total adult hemoglobin; however, it is typically increased in β-thalassemias and may be increased in other conditions, to be described later. Clinically, its major importance is its value in diagnosing β-thalassemias (161).
|
Table 7.5 Normal Human Hemoglobins |
|||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||
Modifications of Normal Hemoglobin
Analysis of human red cell hemolysates by cation-exchange chromatography reveals several negatively charged minor hemoglobins that are designated AIa, AIb, and AIc, corresponding to their order of elution. These hemoglobins are formed by the nonenzymatic interaction of glucose with the α-amino groups of valine residues at the N terminus of the β chaines of hemoglobin (162). They account for the so-called hemoglobin A3 fraction on starch block electrophoresis (Fig. 7.13) (163).
The best characterized of the acquired variants is hemoglobin AIc, which constitutes about 3.5% of the hemoglobin in normal subjects and may be increased two- to threefold in individuals with diabetes mellitus. Its level is directly proportional to the time-integrated mean blood glucose concentration over the life of the red cell, typically the preceding 2 to 3 months (164). In the nonenzymatic glycosylation of hemoglobin A, a molecule of glucose forms a Schiff base with the N terminal of the β chain, then undergoes an Amadori rearrangement to a stable ketamine, 1-amino,1-deoxy fructose (165,166). With special techniques, two different derivatives can be detected in the AIa fraction. In so-called hemoglobin AIa1, a fructose 1,6-diphosphate molecule is attached to the β chain, whereas in hemoglobin AIa2, glucose 6-phosphate occupies the same site (164). Hemoglobin AIb has not yet been fully characterized, but it appears to be a glycosylated, nonphosphorylated derivative. Levels of hemoglobins AIa and AIb, like hemoglobin AIc, are also increased in persons with hyperglycemia. Recently, Hb A1d3 has been identified and found to be due to the addition of glutathione to the β chain of hemoglobin (167).
|
|
|
Figure 7.13. Column chromatograph of hemoglobin A on IRC-50. The main component, AIf, corresponds to the Hb-A1 demonstrated by starch-block electrophoresis, AIIIcorresponds to Hb-A2, and AI corresponds to A3. (Courtesy of Dr. Robert L. Hill.) |
Because the glycosylated hemoglobins are synthesized throughout the lifespan of the red cell, older cells contain a higher proportion of these modified hemoglobins than younger ones (168). Preferential destruction of older cells explains the observation that the proportion of these hemoglobins is reduced in hemolytic anemia (169). Because the rate of synthesis of hemoglobin AIc depends on the blood glucose level, the concentration of hemoglobin AIc at any one time is proportional to the average blood sugar over the previous 2 to 3 months (170). For this reason, the level of glycosylated hemoglobins is used widely as a measure of glucose control in diabetic patients, as well as a tool for diagnosis of diabetic states (170,171,172). In contrast, the hemoglobin AIacomponents are not significantly increased in diabetes (Table 7.6), presumably reflecting the fact that their phosphorylated substituents are not increased in the red cells of patients with diabetes (164).
Clinical laboratories currently use a variety of assays to detect and quantify glycated hemoglobins, including cation-exchange chromatography, high-pressure liquid chromatography, immunoassays, and boronate-affinity methods (173). Commonly encountered hemoglobin variants, such as Hb S, Hb C, Hb E, and Hb F, can interfere with these assays, and may help explain discordant assay results.
Laboratory Analysis of Hemoglobins
Normal and variant hemoglobins can be detected and quantified by standard clinical laboratory techniques. Standard cellulose acetate electrophoresis performed at alkaline pH or isoelectric focusing (IEF) can detect most of the common variants (174). Usually, confirmation can be achieved by citrate agar electrophoresis, IEF, or high-performance liquid chromatography (HPLC). The analysis of hemoglobin A2 and fetal hemoglobin levels deserves special attention because the levels of these components are indicative of common conditions affecting hemoglobin synthesis. Hemoglobin A2normally constitutes <3.5% of the total hemoglobin in adults. In most cases of β-thalassemia and some α-thal-assemia, hemoglobin A2 levels may be increased, ranging from 3.6 to 8.0% (175,176). The subtle increase in hemoglobin A2 that is characteristic of these conditions cannot be quantified accurately by electrophoretic methods. Instead, most laboratories utilize ion-exchange-resin microchromatography (177). It is a common error to use hemoglobin electrophoresis as the sole method to rule out β-thalassemia, and most large laboratories add the chromatographic quantification of hemoglobin A2 as part of their routine hemoglobin analysis. In the most recent proficiency-testing survey for hemoglobinopathies, sponsored by the American College of Pathologists, approximately 40% of laboratories indicated that HPLC is utilized for hemoglobin identification (178). With this procedure there is a clear illusion and separation of hemoglobins A, F, and A2.
|
Table 7.6 Glycosylated Hemoglobins in Normal and Diabetic Individuals |
||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||
Hereditary persistence of fetal hemoglobin (HPFH) and δ-β–thalassemia are types of β-thalassemia caused by deletion of δ- and β-globin genes, but distinguished from typical β-thalassemia by more balanced α- and non–α-globin chain synthesis resulting from an increase in γ-globin production. In both of these thal-assemias, Hb A2 levels are reduced as a result of deletion of δ-globin genes.
Elevated levels of hemoglobin F can be caused by thalassemias, disorders of hematopoiesis, or hereditary disorders of globin synthesis such as HPFH and δ-β–thalassemia (179). Diagnosis requires the precise measurement of hemoglobin F levels, which can be achieved in clinical laboratories by taking advantage of the resis-tance of hemoglobin F to alkali denaturation (180). This resistance to alkali denaturation of fetal hemoglobin is due to the greater stability under these conditions of α-γ–globin dimers compared to α-β–globin dimers (181). In lysates exposed to alkali under controlled conditions, only fetal hemoglobin remains undenatured, and its concentration can be quantified after the other denatured hemoglobins are removed from solution (182). Routine electrophoretic procedures do not completely separate hemoglobin F from A, but a more adequate separation can be accomplished at pH 6.0 in agar gel, or now more commonly by HPLC (178,183).
In adults, fetal hemoglobin is unevenly distributed in erythrocytes, being restricted to between 0.1 and 7% of total cells (184). Cells containing fetal hemoglobin are designated F or A/F cells, wherein the hemoglobin F concentration is normally between 14 and 25% of the total hemoglobin (185). In certain thalassemias and HPFH, the number of F cells is increased. This condition can be detected by acid treatment of erythrocytes on a glass slide followed by elution of other unstable hemoglobins (Kleihaure-Betke technique). Counterstaining can identify hemoglobin F-containing cells (186).
Examination of normal blood by the Kleihaure-Betke method demonstrates both colorless cells and fetal hemoglobin containing (light pink) cells that vary in intensity. By comparison, analysis of cord blood mixed with adult blood demonstrates that true fetal cells in cord blood stain intensely, reflecting the high level of fetal hemoglobin in fetal cells. This assay can be performed on maternal blood to detect feto-maternal hemorrhage or other contamination of the maternal circulation with fetal blood. This method is easily performed in small laboratories, but some larger laboratories are now using flow cytometry with a phycoerythrin-conjugated antiglycophorin antibody to detect feto-maternal hemorrhage (187).
Structure of Globin
Proteins have at least four levels of structural organization: (a) primary structure, or the linear sequence of amino acids; (b) secondary structure, which describes how the amino acids within segments of the protein are spatially organized, e.g., by folding into an α helix or β-pleated sheet; (c) tertiary structure, which refers to the steric relationships of sequence domains separate from each other when analyzed as part of the linear sequence of the protein; and (d) quaternary structure, or the way in which several polypeptide chains join to form a single molecule.
The exact primary structure of all normal globin chains has been determined based on the DNA sequence of the individual globin genes (Table 7.7) (188), and the polypeptide chains in hemoglobin differ from one another in amino acid sequence. The α chain contains 141 amino acids and the non-α chains, 146. The members of the non–α-chain family are more similar than any member of the non-α chain and the α-chain family. The δ chain differs from the β chain in only 10 of the 146 amino acid residues, whereas the γ and β chains differ by 39 amino acids.
In spite of the differences in the primary structure of non–α-globin chains, their secondary structures are remarkably similar. Each has eight helical segments designated by the letters A through H (Table 7.7) (189). The helixes of all the non–α-chain members are of identical length; however, a significant difference exists between the α- and non–α-globin chains in the region of the D helix, which contains seven amino acids in the ∊, γ, δ, and β chains, but only two amino acids in the α chain. Because of the size of the D helix in the α-globin chains, many do not assign it a helix designation. The helixes make up about 75% of the molecule. Interspersed between them are seven nonhelical segments: NA, AB, CD, EF, FG, GH, and HC. This arrangement is important structurally, because the helixes are relatively rigid and linear, whereas the nonhelical segments allow bending.
A given amino acid in a polypeptide chain may be denoted either by its sequential number or by a helical number. In the sequential system, the N-terminal amino acid is assigned the number 1, and each succeeding amino acid receives the next higher number until the C terminal is reached. With this system, amino acids are numbered from 1 to 141 in the α chain and from 1 to 146 in the β, γ, and δ chains. In the helical system, each amino acid is designated by a letter and a number that indicate the helix and the position in the helix, respectively. The helical system is gradually gaining favor, because it illustrates the homology between chains and has more structural significance. For example, the histidine to which heme attaches is amino acid #87 in the α chain and #92 in the β, γ, and δ chains; the helical designation for this histidine is the same in all the normal chains, F8.
The tertiary and quaternary structures of hemoglobin have been studied by x-ray diffraction techniques, especially by Perutz and his coworkers (189,190). In aqueous solutions and in crystals, the polypeptide chains assume a structure in which the polar amino acids face the molecular surface, where they interact with water, rendering the molecule soluble. The groups directed toward the inner core of the molecule are all nonpolar, and the hydrophobic (van der Waals) bonding that occurs between them makes the structure stable. The resulting, roughly spherical, tertiary structure is similar for all the normal hemoglobin polypeptides (Fig. 7.14) as well as for certain other heme proteins, such as myoglobin.
|
Table 7.7 Primary and Secondary Structure of Hemoglobin Polypeptide Chains |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Figure 7.14. The tertiary structure of a single globin polypeptide chain. The helical segments, labeled A through H, are relatively linear; bending of the chains occurs between helices. Heme is suspended in a crevice between the E and F helices. (Courtesy of C. A. Finch.) |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The heme pocket is the site of many dynamic interactions involving oxygen binding to hemoglobin. Heme is suspended in a nonpolar crevice between the E and F helixes (Fig. 7.14), and helixes B, G, and H constitute the floor of the pocket. Heme iron forms a covalent bond with the imidazole nitrogen of the “proximal” histidine at F8. In addition, heme forms van der Waals bonds with many other parts of the molecule and in this way makes an important contribution to tertiary structure. If heme is extracted, the central helical regions, C, D, E, and F, unfold with a consequent decrease in solubility (191). Not surprisingly, some unstable hemoglobins (see Chapter 39) result from amino acid substitutions in the residues that line the heme pocket (192).
The binding of oxygen to the iron molecule causes the hemoglobin molecule to undergo conformational changes that affect the binding of oxygen to other heme sites. The mechanism for this property can be explained in part by the interactions in the heme pocket. The two histidines of globin (E7, F8) are located immediately above and below iron, which is in the plane of the pyrrole ring in oxyhemoglobin (193). In deoxyhemoglobin, the bond between the imidazole nitrogen of the proximal histidine and iron undergoes considerable strain, displacing iron from the plane of the ring. This strain is conveyed to other parts of the molecule and is in part responsible for the T or tense state of deoxyhemoglobin (194). The addition of two molecules of oxygen, which is bound to the iron atom in the heme ring by end-on geometry, results in the formation of a hydrogen bond between the oxygen atom that is not bound directly to the iron and the imidazole nitrogen of the histidine at E7 (the “distal” histidine) (194). The binding of oxygen to iron changes the electron spin state of iron and relaxes the covalent bond with the proximal histidine, permitting the iron to move into the plane of the ring and relaxing the molecule, contributing to the R or relaxed state (195). The overall conformational changes to hemoglobin appear to be the greatest after three molecules of O2 have been added. In general, proteins that undergo an allosteric change from the tense (T) to a relaxed (R) state are better able to interact with substrate in the relaxed state.
When four polypeptide chains combine to form the hemoglobin molecule, each chain lies approximately at the vertices of a regular tetrahedron. With high-resolution x-ray diffraction, the nature of the contacts between chains has been explored in detail for horse hemoglobin (189). Contacts between like chains, i.e., α1α2 and β1β2, are limited and of little importance. The two major contacts between unlike chains have been named α1β1 and α1β2, respectively (Fig. 7.15). (The α2β2 contact is the same as α1β1.) The α1β1 contact point is extensive and moves relatively little (<0.1 nm) when hemoglobin is oxygenated. The α1β2 contact is smaller and smoother, and movement on oxygenation is relatively great (as much as 0.7 nm). As a result, there are two quaternary structures for hemoglobin: one for the deoxygenated form and one for the liganded or oxygenated form. The main difference between the two is the nature of the α1β2 contact (Fig. 7.16).
|
|
|
Figure 7.15. Diagram of hemoglobin tetramer illustrates two types of αβ-contact points: a relatively extensive one (α1β1 or α2β2) and one that is smaller (α1β2 or α2β1). The actual molecular contact points are detailed in Figure 7.16. |
Assembly of Hemoglobin
Much of our overall knowledge of the synthesis of proteins was learned from the study of globins, and many of the specific details of the synthesis of globins are described in Chapter 6. Assembly of the hemoglobin tetramer from the monomers, however, has received relatively little attention, although an understanding of the process can provide some insight into the laboratory evaluation of hemoglobins.
In general, there is little posttranscriptional regulation of the synthesis of globins, although factors such as the availability of heme can affect translation of globin mRNA (196). Aside from several variants, such as hemoglobin Lepore, which are synthesized at a slower rate, the synthesis rate of most normal or mutant globins is the same (197). Nevertheless, individuals with β-chain variants often express less of the variant hemoglobin than hemoglobin A. This observation has been attributed to an increased rate of catabolism of newly synthesized globin chains resulting from decreased solubility, defective heme binding, or abnormal subunit assembly. The best data indicate that the variations seen in most stable hemoglobin variants result from differences in subunit assembly (181,198).
Following translation of globin mRNA and globin-chain synthesis, heme associates with globins and α-globin chains pair with members of the β-globin family. In large part this binding is a consequence of the different charges on α globins, which are positively charged (pI = 8.1), and β globins, which are negatively charged (pI = 6.6) (199). The greater the charge difference, the greater is the electrostatic attraction. Positively charged variants, such as βC and the uncharged βS, bind to α globin and assemble into α-β dimers at approximately half the rate of βA during in vitro mixing experiments (198). Conversely, more negatively charged variants such as βN-Baltimore bind with a greater association rate. This phenomenon has been suggested as an explanation for the ratio of 60:40 seen for hemoglobin A and hemoglobin S in heterozygotes for βS. Likewise the percentage of N-Baltimore is increased over that of hemoglobin A.
In conditions of α-globin deficiency, this competition is more pronounced and the percentage of the more positively charged variant is further reduced. Patients who are heterozygous for βS and α-thalassemia carry percentages of hemoglobin S of approximately 35, 30, and 25%, corresponding to one-, two-, or three-gene α-thalassemia (200).
|
|
|
Figure 7.16. Quaternary structure of hemoglobin. At left is the most extensive contact, α1β1, in which 16 amino acids in the α chain form bonds with 18 amino acids in the β chain. The α1β1contact does not change significantly on oxygenation. The smaller α1β2 contact, center and right, has two forms, depending on whether the hemoglobin is in the oxygenated (oxy) or deoxygenated (deoxy) form. At α1β2, 10 or 11 amino acids in the α chain form bonds with 9 amino acids in the β chain. Broken lines, hydrogen bonds; plain lines, van der Waals bonds. Numbers on lines give the number of atoms in contact. (From Perutz MF, Miurhead H, Cox JM, et al. Stereochemistry of cooperative effects in haemoglobin. Nature 1968;219:29.) |
Hemoglobin A2 is reduced in certain α-thalassemias and in iron deficiency, which causes an acquired reduction in α-globin synthesis as a result of decreased heme synthesis. Under these conditions, the more positively charged δ globin would be expected to compete less well with the normal β globin. On the contrary, during β-globin deficiency associated with β-thalassemia, δ globin would be expected to compete more effectively for α-globin chains, and the predicted increase in hemoglobin A2 is observed (201).
Oxygen Transport
In order to function as the primary medium of exchange of oxygen and carbon dioxide, hemoglobin must fulfill the four requirements first delineated by Barcroft (202): It must be capable of transporting a large quantity of oxygen, it must be highly soluble, it must take up and release oxygen at “appropriate pressures,” and it must also be a good buffer. Normal hemoglobin fulfills these requirements well, although many abnormal variants fail to meet one or more of these conditions.
Each gram of hemoglobin, when fully saturated, binds 1.39 ml of oxygen. The degree of saturation is related to the oxygen tension (Po2), which normally ranges from 100 mm Hg in arterial blood to about 35 mm Hg in veins. The relation between oxygen tension and hemoglobin oxygen saturation is described by the oxygen dissociation curve of hemoglobin (Fig. 7.17). The characteristics of this curve are related in part to properties of hemoglobin itself and in part to the environment within the erythrocyte, with pH, temperature, and concentration of 2,3-diphosphoglycerate (2,3-DPG) being the most important factors affecting oxygen affinity.
Oxygen affinity of a particular hemoglobin is generally expressed in terms of the oxygen tension at which 50% saturation occurs, the so-called P50. When measured in whole erythrocytes, this value averages about 26 mm Hg in normal, nonsmoking males and slightly higher in normal, nonsmoking females (203). When oxygen affinity increases, the dissociation curve shifts leftward, and the value for P50 is reduced. Conversely, with decreased oxygen affinity, the curve shifts to the right and P50 is increased.
The oxygen dissociation curve of single-subunit heme polypeptides (e.g., myoglobin) is hyperbolic, and oxygen affinity is considerably greater than that of hemoglobin (Fig. 7.17). Such a compound would function poorly in oxygen transport because little oxygen would be released until the tissue oxygen tension was very low. For example, at the usual tissue pO2 of 40 mm Hg, myoglobin would remain >90% saturated. In contrast, the oxygen-curve of hemoglobin is distinctly sigmoidal; the steepest part of its slope occurs at levels of oxygen tension corresponding to those found in tissues. This difference between the hemoglobin and myoglobin curves is the result of interaction between the four heme-polypeptide units of hemoglobin. Although it was called heme-heme interaction in the past, subunit cooperativity better describes the process whereby the binding of oxygen by one subunit increases the oxygen affinity of other subunits; no direct interaction among heme moieties is involved. This allosteric property of hemoglobin permits rapid changes in oxygen affinity during the time the red blood cell passes through the capillary bed.
|
|
|
Figure 7.17. Oxygen dissociation curve of hemoglobin, at three values for pH, compared with that of myoglobin. pCO2, partial pressure of carbon dioxide; pO2, partial pressure of oxygen. |
The change in oxygen affinity with pH is known as the Bohr effect (204,205). Hemoglobin oxygen affinity is reduced as the acidity increases (Fig. 7.17). Because the tissues are relatively rich in carbon dioxide, and because red cell carbonic anhydrase readily converts carbon dioxide to carbonic acid, the pH is lower there than in arterial blood; therefore, the Bohr effect facilitates transfer of oxygen to tissues. In the lungs, as oxygen is taken up and carbon dioxide is released, the pH rises and the oxygen-affinity curve shifts to the left. This event, termed the alkaline Bohr effect, increases the oxygen affinity of hemoglobin, helping to maximize oxygen uptake. Thus, the Bohr effect links and enhances the transport of both oxygen and carbon dioxide.
Another important factor affecting the oxygen affinity of hemoglobin is the concentration of 2,3-DPG (206,207,208,209). The molecule can insert into the pocket between β-globin subunits in tetrameric hemoglobin and reduces oxygen affinity. 2,3-DPG is synthesized from glycolytic intermediates by means of a pathway known as the Rapoport-Luebering shunt (Fig. 7.18). In the erythrocyte, 2,3-DPG is the predominant phosphorylated compound, accounting for about two thirds of the red cell phosphorus; in contrast, it is present in only trace amounts in other tissues.
The production of 2,3-DPG depends on the rate of formation of its precursor, 1,3-DPG, and the relative amounts of 1,3-DPG going into the Rapoport-Luebering shunt and into the ATP-forming glycolytic pathway. Actual 2,3-DPG concentration also depends on the rate of hydrolysis of 2,3-DPG. Although several steps in the glycolytic pathway may be sensitive to changing conditions, such as pH, ADP-ATP, and nicotinamide-adenine dinucleotide (NAD-NADH) ratios, and the concentration of inorganic phosphorus, it is the relative concentrations of ATP and ADP that appear to be most directly linked to the rate of 2,3-DPG production. A relative increase in the amount of ADP is associated with an increase in the production of 3-phosphoglycerate and a decrease in 2,3-DPG. Oxidants such as methylene blue and pyruvate increase 2,3-DPG synthesis by affecting the NAD–NADH ratio.
|
|
|
Figure 7.18. The synthesis of 2,3-diphosphoglycerate (2,3-DPG) or 3-phosphoglycerate (3-PG) and ATP from 1,3-diphosphoglycerate (1,3-DPG) by the Rapoport-Luebering cycle. (From Bunn HF, Forget BG. Hemoglobin: molecular, genetic and clinical aspects. Philadelphia, WB Saunders, 1986, with permission.) |
Two multifunctional enzymes are important to the synthesis and degradation of 2,3-DPG. Both enzymes are capable of promoting three reactions: the synthesis of 2,3-DPG from 1,3-DPG; the breakdown of 2,3-DPG into 3-PG, water, and phosphorus; and the conversion of 3-PG to 2-PG. The structures of these enzymes and their relative activities have been described extensively (210,211,212,213,214).
The most important function of 2,3-DPG is its effect on the oxygen affinity of hemoglobin. In the deoxygenated state, hemoglobin A can bind 2,3-DPG in a molar ratio of 1:1, a reaction that leads to reduced oxygen affinity and improved oxygen delivery to tissues. The increased oxygen affinity of fetal hemoglobin appears to be related to its lessened ability to bind 2,3-PDG. The cellular concentration of 2,3-DPG also affects intracellular pH, because 2,3-DPG is a highly charged, impermeant anion. Hemoglobin, also an impermeant anion, and 2,3-DPG together lower the intracellular concentration of chloride relative to the extracellular concentration. This effect, in turn, lowers intracellular pH relative to extracellular pH, which by the Bohr effect also lowers oxygen affinity and thus increases oxygen delivery to tissues (215).
The increased oxygen affinity of stored blood is accounted for by reduced levels of 2,3-DPG (216). Transfusion of such blood results in an in vivo increase in oxygen affinity that returns toward normal in 7 to 12 hours as the function of the glycolytic pathway is restored. The reduction in 2,3-DPG levels in stored blood can be mitigated by adding inosine or phosphate to the storage solutions (217). The actual clinical impairment in oxygen delivery resulting from low 2,3-DPG levels remains disputed, but it would be expected to have its greatest impact when large transfusions are required in critically ill individuals. In such situations, some centers administer fresh blood (218).
Changes in 2,3-DPG levels play a significant role in adaptation to hypoxia. In some situations associated with hypoxemia, 2,3-DPG levels in red cells increase, oxygen affinity is reduced, and delivery of oxygen to tissues is facilitated. Such situations include abrupt exposure to high altitude, anoxia resulting from pulmonary or cardiac disease, blood loss, and anemia (219,220,221). Increased 2,3-DPG levels also play a role in adaptation to exercise (222). The compound is not essential to life, however; an individual who lacked the enzymes necessary for 2,3-DPG synthesis was perfectly well except for mild polycythemia. Heme-heme interaction, the Bohr effect, and the effect of 2,3-DPG have been explained on a molecular basis in a model proposed by Perutz (223). In the completely deoxygenated state, hemoglobin assumes a quaternary structure termed T (“taut” or “tense”). This structure is stabilized by salt bridges involving the carboxy terminals of the peptide chains. The deoxy form is also stabilized by the presence of 2,3-DPG, which joins the β chains as shown in Figure 7.19.
The tertiary structure of deoxygenated subunits also differs from that of the oxygenated form. In the deoxy form, heme iron is in a high-spin state and, in this form, is displaced slightly from the plane of the porphyrin ring. The penultimate tyrosine is wedged firmly between the F and H helixes. When oxygen is added, iron changes to a low-spin state and moves to a position in plane with the porphyrin ring, a distance of about 0.2 nm, and pulls the attached F helix with it. This movement narrows the space between the F and H helixes, expelling the penultimate tyrosine from its pocket. The C-terminal amino acid moves with the tyrosine, thereby breaking the salt bridges with adjacent chains.
|
|
|
Figure 7.19. The 2,3-DPG–binding site. Part of one β chain is at upper left, the other at lower right. Salt bridges are found between the phosphates of 2,3-DPG and positively charged groups at 1 Val (a-NH2+), 2 His, and 143 His. The 82 Lys of one chain is also involved. (From Arnone A. X-ray diffraction study of binding of 2,3-diphosphoglycerate to human deoxyhemoglobin. Nature 1972;237:146, with permission.) |
Fully oxygenated hemoglobin assumes the R or “relaxed” structure (Fig. 7.20). The exact series of events that bring about this change, along with the change in oxygen affinity that accompanies it, has been difficult to ascertain, in part because of the large number of variables that need to be controlled (e.g., temperature, pH, 2,3-DPG concentration, and hemoglobin concentration) and the extraordinarily large number of thermodynamic measurements that need to be made. Nevertheless, at least some of the molecular events that contribute to this shift in quaternary structure and oxygen affinity have been delineated (224,225). Hemoglobin appears to exist in a third molecular form, intermediate between the T and R conformations. Achievement of the R conformation occurs when at least one oxygen molecule is bound on each side of the α1β2 interaction; however, significant subunit cooperativity also exists within each α1β1 dimer of the T-state tetramer. Conversion to the R form is also accompanied by expulsion of the 2,3-DPG and disruption of the salt bridges and hydrophobic interactions at the α1β2 contact point (Fig. 7.20). Oxygen affinity then becomes much increased, and oxygen is added to the remaining β chain or chains. Because hemoglobin A has four subunits, it might be expected that the switch from the T state to the R state would occur in four distinct steps. Experimentally, it appears to switch abruptly. This is best explained by the relationship of the four subunits. The α and β chains in the dimers have many more contacts than the two dimers with each other, and the dimers can move onto each other. As oxygen is either loaded or unloaded onto the hemoglobin tetramer, tension is created, and the point where the two dimer subunits slide onto each other is determined by the concentration of the allosteric mediaters in the cell, including pH, Cl_ concentration, and 2,3-DPG levels.
|
|
|
Figure 7.20. Diagrammatic representation of the subunit interaction in hemoglobin as oxygen is added. Deoxyhemoglobin (upper left), with low oxygen affinity, is in the T (taut) conformation, constrained by salt bridges (interconnecting lines) and the 2,3-DPG molecule. As O2 is added, salt bridges are broken, and eventually the DPG molecule is expelled, resulting in the R (relaxed) configuration with higher oxygen affinity. |
The alkaline Bohr effect is explained by rupture of the salt bridges involving the β-chain C-terminal histidine and the α N-terminal valine. When these bridges are broken, the pK of the dissociation of hydrogen ion is reduced. It has also been suggested that about 25% of the alkaline Bohr effect may be accounted for by histidine at α122 (226).
Alterations in the oxygen dissociation curve occurring in various hemoglobinopathies are discussed later in this book (Chapter 39).
Carbon Dioxide Transport
Transport of carbon dioxide by red cells, unlike that of oxygen, does not occur by direct binding to heme (227). In aqueous solutions, carbon dioxide undergoes a pair of reactions:
1. CO2 + H2O → H2CO3
2. H2CO3 → H+ + HCO3-
Carbon dioxide diffuses freely and rapidly into the red cell, where the presence of the enzyme carbonic anhydrase facilitates reaction 1. The H+liberated in reaction 2 is accepted by deoxygenated hemoglobin, a process facilitated by the Bohr effect. The bicarbonate formed in this sequence of reactions diffuses freely across the red cell membrane and a portion is exchanged with plasma Cl-, a phenomenon called the chloride shift. The bicarbonate is carried in plasma to the lungs, where ventilation keeps the Pco2 low, resulting in reversal of these reactions and excretion of CO2 in the expired air. About 85% of tissue carbon dioxide is processed in this way, and 5% is carried in simple solution.
The remainder of the CO2 is bound to the N-terminal amino group of each polypeptide chain by a carbamino complex, the result of an attack by the electron-poor carbon atom of CO2 on the electron-rich terminal amino acids. This nonenzymatic process varies directly with pH. Approximately 10% of CO2 is bound to deoxygenated hemoglobin, forming carbaminohemoglobin (Hb–NH–COO-).
Earlier, the effect of CO2 on oxygen affinity was noted and attributed to the Bohr effect. An additional, more direct effect results from CO2binding to hemoglobin. At a given pH, carbaminohemoglobin has a lower affinity for oxygen than has hemoglobin in the absence of CO2. This is felt to be a result of the stabilization of the T state through additional bonds, especially involving arginine 141 (228).
The carbon dioxide dissociation curve is analogous to the oxygen dissociation curve, in that it depicts the relationship between CO2 tension (pCO2) and CO2 content (Fig. 7.21). It is somewhat more nearly linear than the oxygen curve, especially in the physiologic range (pCO2 of 40 to 60 mm Hg). Because of the Bohr effect, blood containing deoxyhemoglobin has greater affinity for CO2 than does oxygenated blood. The shift in the CO2 dissociation curve related to this phenomenon, known as the Haldane effect, facilitates CO2 binding in the tissues and release in the lungs (Fig. 7.22).
Carbonic anhydrase is the second most abundant cytoplasmic protein of the erythrocyte. It is a zinc-containing enzyme for which three separately encoded isoenzymes, CAI, CAII, and CAIII, are recognized. The predominant form in the human erythrocyte is CAI, although CAII is also found. Largely absent from fetal erythrocytes, erythrocyte CAI expression reaches adult levels during the first year of life, in a manner analogous to the expression of adult hemoglobin (hemoglobin A) (229,230). All three isoenzymes of carbonic anhydrase have similar structures, and all appear to be encoded by genes on chromosome 8. CAI, however, contains within its active site a unique histidine that appears to stabilize the enzyme–HCO3 complex, thus effecting the maximal rate of CO2 hydration (231,232,233,234).
|
|
|
Figure 7.21. The carbon dioxide dissociation curve of whole blood. The curve is relatively linear between pCO2 40 and 60 mm Hg. The difference between the curves of oxygenated and deoxygenated blood is known as the Haldane effect. |
Although carbonic anhydrase clearly facilitates carbon dioxide metabolism, its presence in erythrocytes does not appear to be essential. Hereditary absence of CAI has been reported to occur without hematologic sequelae, and acquired deficiency is often seen with hyperthyroidism (235,236). CAII deficiency is likewise without hematologic effect, although it is associated with the syndrome of renal tubular acidosis and cerebral calcifications (237); enzyme activity with characteristics similar to those of CAIII is found to be increased in CAII-deficient erythrocytes (238). Also, complete inhibition of carbonic anhydrase by acetazolamide has little effect on carbon dioxide transport, at least under basal conditions (239). Nevertheless, carbonic anhydrase does appear to equalize the rate of hydration of carbon dioxide with the rate of protonation of hemoglobin, thus bringing into balance the rates of oxygen and carbon dioxide exchange. This equalization may be especially advantageous when the system is stressed, such as during exercise (240).
|
|
|
Figure 7.22. Interrelations of oxygen and carbon dioxide transport in the erythrocyte. Arrows to the left indicate direction of reactions taking place in the tissues; those to the right, in the lungs. In the tissue, CO2 diffuses into the red cell, and its hydration is catalyzed by carbonic anhydrase (CA). Dissociation of the resulting carbonic acid produces bicarbonate and a proton (H+). The bicarbonate is exchanged for chloride in the plasma. The proton is accepted by oxyhemoglobin (HbO2), a reaction that, by means of the Bohr effect, facilitates the dissociation of oxygen. These reactions are reversed in the lungs because of the low pCO2 and high pO2. |
Nitric Oxide (NO): Another Allosteric Effector of Hemoglobin
The discovery of nitric oxide as an important regulator of vascular and smooth muscle tone has provided many insights into our understanding of vascular physiology. Nitric oxide is produced by the vascular endothelium and relaxes muscles surrounding vessels, thereby controlling blood pressure. Subsequently, it was determined that free hemoglobin could act as a scavenger of nitric oxide and inactivate it, explaining the observation that the infusion of free hemoglobin results in significant elevations of blood pressure. This reaction occurs because the oxygenated heme irons scavenge nitric oxide in a reaction that yields methemoglobin.
Early studies of the interaction of hemoglobin with NO predicted that a function of hemoglobin was to eliminate or limit the biologic activity of NO and did not predict any functional effect on hemoglobin other than the oxidation of heme iron. This model did not explain how NO could be maintained at observed levels based on its known low levels of production. This was clarified with the discovery of S-nitrosohemoglobin (SNO-Hb) (241). When free hemoglobin is incubated with NO or S-nitrosothiols, S-nitrosothiols (SNOs) rapidly form on the two 93β cysteines of hemoglobin rather than reacting with the oxygenated heme groups as might be expected. The infusion of the SNO-hemoglobin results in no increase in blood pressure (242).
In the pulmonary circulation, coincident with oxygenation of hemoglobin, nitric oxide is added to hemoglobin, and rather than oxidizing the heme iron, it binds to the iron or forms SNO-Hb through the reactive sulfhydryl groups of cysteine 93β of hemoglobin (243,244). This function for cysteine 93β may explain why this amino acid is invariant in mammals and birds. Reactivity of nitric oxide with the SH groups of cysteine 93β is controlled by the allosteric transition of hemoglobin and the spin state of heme iron. Thus, oxygen binding and conversion to the R state increase reactivity (241).
Likewise, loss of oxygen in the peripheral tissues results in transition of hemoglobin to the T state and release of NO. NO is bound to either small thiols or the anion exchanger AE1. The latter is facilitated because hemoglobin can bind to the RBC membrane through specific, high-affinity binding to the amino-terminal cytoplasmic domain of the chloride/bicarbonate anion- exchange protein AE1 (band-3 protein) (87). The amino terminus of AE1 inserts into the cleft usually occupied by 2,3-DPG (87). NO-medicated vasodilatory activity is released from AE1 by deoxygeneation and binds to receptors on the vascular endothelium to induce vasodilation. S-nitrosohemoglobin is therefore detected in arterial but not in venous blood. Several recent reviews have been published on the relationship of nitric oxide and S-nitrosohemoglobin and the regulation of blood flow (245,246,247).
Oxidative Denaturation of Hemoglobin: Its Reversibility and Prevention
Oxyhemoglobin in solution gradually undergoes auto-oxidation, becoming methemoglobin (HbFe3+). The rate of oxidation is enhanced by conditions such as increased temperature, decreased pH, the presence of organic phosphate and of metal ions, and partial oxygenation of hemoglobin. To bind oxygen reversibly, however, the iron in the heme moiety must be maintained in the reduced (ferrous, Fe2+) state, despite exposure to a variety of endogenous and exogenous oxidizing agents. The red cell maintains several metabolic pathways to prevent the action of these oxidizing agents and to reduce the hemoglobin iron if it becomes oxidized. Under certain circumstances, these mechanisms fail and hemoglobin becomes nonfunctional. At times, hemolytic anemia supervenes as well. These abnormalities are particularly likely to occur (a) if the red cell is exposed to certain oxidant drugs or toxins (see Chapters 32, 36, and 39), (b) if the intrinsic protective mechanisms of the cell are defective (see Chapters 32 and 39), or (c) if genetic abnormalities of the hemoglobin molecule affect globin stability or the heme crevice (see Chapter 39).
The oxidation of hemoglobin occurs in a stepwise fashion from fully reduced hemoglobin to fully oxidized hemoglobin. Intermediate forms are called valence hybrids. In deoxyhemoglobin, the heme iron is in the “high-spin” ferrous state, in which six electrons are in the outer shell, four of which are unpaired. When oxygen is added, one of these electrons is partially transferred to the bound oxygen. Usually, when oxygen is given up, oxyhemoglobin dissociates into partially deoxygenated hemoglobin and molecular oxygen:
Hb(O2)4 → Hb(O2)3 + O2
A superoxide anion rather than molecular oxygen may dissociate, however, thus oxidizing the Fe to the ferric state, producing methemoglobin:
HbFe2+ + O2 → HbFe3+O2- → HbFe3+ + O2-
This type of dissociation is particularly likely if water gains access to the heme crevice. Methemoglobin formation may also occur in vivo as the result of exposure to superoxide anions:
2HbFe2+ + O2 + 2O2- + 4H+ → 2HbFe3+ + 3O2 + 2H2O
The formation of methemoglobin may also result from a direct reaction of reduced hemoglobin with the reduction product of the superoxide ion, peroxide:
2HbFe2+ + 2H2O2 → 2HbFe3+·H2O + O2
As a result of these processes, methemoglobin is formed in normal cells at the rate of about 0.5 to 3% per day (248).
Methemoglobin is unable to bind oxygen. It has a distinctive, pH-dependent spectrum (Fig. 7.23; Table 7.8), and, in concentrations >10% of the total hemoglobin, imparts to blood a distinctive brownish hue that does not disappear on vigorous shaking in air (248). When methemoglobin is present in vivo in concentrations >1.5 to 2.0 g/dl, patients appear visibly cyanotic. Methemoglobin combines readily with cyanide to form cyanomethemoglobin, a pigment so stable that it is used in laboratory procedures for quantifying hemoglobin.
As oxidative denaturation continues, methemoglobin is converted to derivatives known as hemichromes (Fig. 7.24) (249). Hemichromes also may form directly from hemoglobin without methemoglobin as an intermediate. The hemichromes are low-spin, ferric compounds with a greenish hue and a characteristic spectrum (Table 7.8, Fig. 7.23). They are formed when the sixth coordination position of iron becomes covalently attached to a ligand within the globin molecule, a change that requires alteration of tertiary protein structure. Probably the most common internal ligand is the so-called distal histidine at E7 (Fig. 7.24). The compound so formed has been called a “reversible” hemichrome, because relatively mild treatment with reducing agents and dialysis under anaerobic conditions converts it to deoxyhemoglobin. It may not be reversible in vivo, however, because it cannot be reduced by methemoglobin reductase.
|
|
|
Figure 7.23. Change in hemoglobin spectrum as oxyhemoglobin changes to methemoglobin. The numbers 1 through 6 represent curves taken in sequence as oxidation proceeds. Note in particular the disappearance of the band at 575 μm and the appearance of the band at 631 μm. |
In contrast, the “irreversible” hemichromes cannot be converted back to normal hemoglobin again in vivo or in vitro, implying that more severe distortions of tertiary protein structure have occurred. In one of the irreversible hemichromes, the histidine imidazole groups are protonated, that is, they participate in hydrogen bonding. The other irreversible hemichrome is characterized by a mercaptide and nitrogenous linkage at the fifth and sixth positions (Fig. 7.24). Presumably, the mercaptide link is provided by a cysteine residue in the globin chain, perhaps at β93.
As these changes occur in the vicinity of the heme group, oxidative changes also occur in other parts of the hemoglobin molecule. Once the cell’s supply of glutathione (GSH) is exhausted, the titrateable sulfhydryl groups at β93(F9)Cys are oxidized, often forming a mixed disulfide with glutathione (250). This change is reversible; however, as further alterations in globin conformation occur, normally protected or “buried” sulfhydryl groups at β112(G14)Cys and α104(G11)Cys become exposed and are oxidized, changes that disrupt the α1β2 contacts. These changes facilitate dissociation of polypeptide chains, first into αβ dimers and finally into monomers (249). In some instances, heme may dissociate from globin, particularly in the case of certain unstable hemoglobins.
The end products of these changes are precipitated hemichromes and precipitated hemefree globin. In intact erythrocytes, these precipitates take the form of globular inclusions known as Heinz bodies, which are not visible with ordinary Wright stain but can be seen easily after supravital staining with crystal violet or brilliant cresyl blue. Heinz bodies may become attached to the cell membrane and shorten red cell survival.
Another nonfunctional hemoglobin derivative that is occasionally formed during the oxidative denaturation of hemoglobin is sulfhemoglobin. This is a relatively stable pigment and, once formed, cannot be converted to hemoglobin in vivo. Instead, it tends to remain within the cell throughout the cell’s life. Sulfhemoglobin is bright green and has a distinctive spectrum characterized by an absorption band at about 618 nm (Table 7.7). It is a ferrous compound with one sulfur atom attached to each heme group. The sulfur is probably attached to a β carbon in the porphyrin ring, forming a thiochlorin (Fig. 7.25) (250,251).
Although the exact mode of synthesis of sulfhemoglobin remains to be established, proposed models suggest that methemoglobin is first converted to ferrylhemoglobin in the presence of hydrogen peroxide (251,252):
|
Table 7.8 Spectral Characteristics of Hemoglobin (Hb) and Its Derivatives |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
HbFe3+ + H2O2 → HbFe4+O + H2O + e-
With the addition of a sulfur-containing compound, such as hydrogen sulfide, the iron in ferrylhemoglobin is reduced and sulfur is incorporated into the porphyrin ring:
HbFe4+O + HS- + 2e- → HbSFe2+ + OH-
Although the iron in sulfhemoglobin is reduced, it binds oxygen with an affinity one hundredth that of unmodified hemoglobin.
Known mechanisms for preventing or reversing oxidative denaturation of hemoglobin in the erythrocyte include (a) the methemoglobin reductases, (b) superoxide dismutase, (c) glutathione peroxidase, and (d) catalase.
|
|
|
Figure 7.24. Stages in the oxidative denaturation of hemoglobin. Names of the stages are given on the left, proceeding from the most reduced form at the top to the most oxidized at the bottom. Partial structures are illustrated on the right. The heme group is denoted by a planar cross. For a more complete description, refer to text. |
Methemoglobin Reduction
Most methemoglobin in erythrocytes is reduced through the action of an enzyme, cytochrome b5 methemoglobin reductase, which acts in the presence of two electron carriers, cytochrome b5 and NADH. Only a small amount of methemoglobin is reduced by all other pathways of methemoglobin reduction together (Table 7.9). These other pathways involve two compounds that cause the reduction of methemoglobin nonenzymatically, ascorbic acid and glutathione, as well as a second enzyme, NADPH-flavin reductase. Deficiency of cytochrome b5 reductase, but not of NADPH-flavin reductase, is associated with methemoglobinemia, confirming that cytochrome b5 reductase is the most important physiologic means of reducing methemoglobin (253). In vitro evidence also confirms that cytochrome b5 reductase is the rate-limiting factor in methemoglobin reduction (254).
Cytochrome b5 reductase has been referred to by several other names, including diaphorase I, DPNH dehydrogenase I, NADH dehydrogenase, NADH methemoglobin reductase, and NADH methemoglobin-ferrocyanide reductase. Work in the 1940s established a relationship between the reduction of methemoglobin and the metabolism of lactate to pyruvate, thus implying an important role for NADH (255). Eventually, two methemoglobin-reducing enzymes were isolated. The NADH-dependent enzyme, which was absent from several patients with methemoglobinemia, has been shown to be a flavoprotein, with one mole of flavin-adenine dinucleotide (FAD) per mole of apo-enzyme. Its molecular weight is approximately 34,000 daltons. The corresponding cDNA has been identified and the gene localized to chromosome 22 (256,257). Most likely, erythrocyte cytochrome b5 reductase and hepatic cytochrome b5 reductase are products of a single gene.
|
|
|
Figure 7.25. The probable structure of sulfhemoglobin. One of the four pyrole rings in heme is illustrated at the left. Sulfur added at the β carbon forms a thiochlorin, accounting for the spectral changes that are characteristic of sulfhemoglobin. |
|
Table 7.9 Proportion of Methemoglobin Reduction by Various Erythrocyte Systems |
||||||||||||
|
||||||||||||
The reduction of methemoglobin by highly purified cytochrome b5 reductase in the presence of NADH is extremely slow, implying that another factor is most likely required as an electron carrier. In vitro, this role can be filled by dyes or by ferrocyanide. In vivo, cytochrome b5 acts as the intermediate electron carrier (258). Erythrocyte cytochrome b5 greatly accelerates reduction of methemoglobin by cytochrome b5 reductase and can also serve as a substrate for hepatic microsomal cytochrome b5 reductase. Congenital methemoglobinemia resulting from a deficiency in cytochrome b5 has been described (259). cDNA for human liver cytochrome b5 has been cloned; it encodes a protein of 134 amino acids (260). Erythrocyte cytochrome b5 probably results from proteolytic cleavage of the membrane-attached protein present in erythroid precursor microsomes, an event that then yields a soluble cytochrome b5 protein.
The process by which cytochrome b5 reductase and cytochrome b5 reduce methemoglobin in the presence of NADH probably involves three steps. In the first, NADH binds to the FAD-reductase complex and, in the presence of hydrogen ion, the NAD is converted to NAD+, and the FAD becomes FADH2. In the second step, cytochrome b5-Fe3+ is reduced to cytochrome b5-Fe2+, and the FADH2 reverts to FAD. Finally, methemoglobin forms a bimolecular complex with reduced cytochrome b5 through electrostatic interactions between negatively charged groups around the cytochrome heme and positively charged groups around the heme moieties of methemoglobin. The reduction of methemoglobin then takes place and can be represented as follows:
HbFe3+ Cytb5Fe2+ → HbFe2+ + Cytb5Fe3+
Of lesser physiologic importance is the enzyme system that depends on NADPH for its activity. It probably accounts for only about 5% of the methemoglobin-reducing activity of normal red cells (Table 7.9), and its hereditary deficiency does not lead to methemoglobinemia (261). The lack of physiologic activity may result from the absence of an intermediate electron carrier analogous to cytochrome b5. If methylene blue is supplied as the carrier, however, the NADPH-dependent enzyme becomes highly effective in methemoglobin reduction. This property is used in the therapy of methemoglobinemia from various causes.
Enzymes That React with Products of the Reduction of Oxygen
As molecular oxygen undergoes successive univalent reductions, a variety of reactive species are generated. These species constitute the oxidizing agents most likely to be responsible for the oxidative denaturation of hemoglobin, and they may damage other cellular components as well, especially lipid-containing elements such as the cell membrane (262,263). A variety of mechanisms has evolved in respiring organisms to deal with these potential toxins, and some are found within the erythrocyte.
Superoxide anions are produced in biologic tissues from several sources, including oxyhemoglobin itself, as well as oxidative reactions catalyzed by flavin enzymes, such as xanthine oxidase (263,264). In addition, many drugs and toxins have oxidant activity and appear to generate superoxide (265). Once superoxide has been generated in aqueous solution, additional toxic products of oxygen may form spontaneously (Fig. 7.26). Thus, superoxide can undergo spontaneous dismutation, yielding peroxide and oxygen:
|
|
|
Figure 7.26. Steps in the univalent reduction of oxygen and enzymatic pathways affecting the intermediates. The enzymatic pathways, shown on the right, provide the means for processing these intermediates without formation of the highly reactive hydroxyl radical. This potent oxidant can be formed by the reaction shown on the left if superoxide and peroxide concentrations are sufficient and if catalytic quantities of transition metals are present. |
O2 + O2 + 2H+ → H2O2 + O2
In addition, in the presence of catalytic quantities of transition metals, superoxide and peroxide may react to form the highly reactive hydroxyl radical (OH·):
O2 + H2O2 → OH· + OH- + O2
Any of these oxygen derivatives may exert toxic effects on cellular components. As previously noted, superoxide appears to induce methemoglobin formation (266,267). It may also bring about cell lysis via its effect on membranes (268). Hydrogen peroxide is the most stable intermediate in the reduction of oxygen. Although hydrogen peroxide has often been shown to induce the oxidative denaturation of hemoglobin in vitro, whether it does so directly or by giving rise to other products, such as the hydroxyl radical, is not clear.
The hydroxyl radical is one of the most potent redox agents known (263). Because it is generated by the radiolysis of water, it is thought to account for many of the effects of radiation in biologic tissue. It also, however, may be generated from superoxide and peroxide, as described previously, and from peroxide in the presence of certain metals:
Fe2+ + H2O2 → Fe3+ + OH- + OH·
Thus, enzymes that scavenge superoxide and peroxide may be viewed as mechanisms for preventing the accumulation of these intermediates in sufficient quantities to allow the hydroxyl radical to form (264).
The superoxide dismutases are enzymes that catalyze the dismutation of superoxide to oxygen and peroxide. Although this reaction occurs spontaneously, the presence of the enzyme speeds the reaction to a rate as much as 109 times faster than the spontaneous rate (264). In the erythrocyte, superoxide dismutase is a soluble, cuprozinc enzyme with a molecular weight of about 32,000 daltons. The enzyme accounts for most of the copper content of the red cell, and before its enzymatic function was determined, it was called erythrocuprein or hemocuprein.
The primary structure of human copper-zinc (Cu-Zn) superoxide dismutase has been determined, and the gene has been mapped to chromosome 21 (269). Although superoxide dismutase prevents the formation of methemoglobin in vitro under conditions in which superoxide forms, the relative importance of this reaction in vivo remains to be established (267,268).
Once hydrogen peroxide is formed, two enzymes catalyze the decomposition of hydrogen peroxide in erythrocytes. These are glutathione peroxidase and catalase. Glutathione (GSH) peroxidase is a component of the following reaction (270,271):
The enzyme is effective at very low concentrations of peroxide (Km = 1 × 10-6 M) (263).
Glutathione peroxidase is the major human selenoprotein, which may account for the antioxidant properties of selenium as a micronutrient (264,272). Human cells grown in the absence of selenium express significantly reduced glutathione peroxidase activity, despite normal glutathione peroxidase mRNA and transcription levels. The gene for glutathione peroxidase is on chromosome 3, although two homologous genes also appear to be present in the human genome (273,274). It has been proposed that a genetic defect in glutathione peroxidase may lead to a drug-sensitive hemolytic anemia (275). However, there is some doubt that acquired or genetic defects in this enzyme are associated with hemolysis (see Chapter 32).
Catalase, a heme enzyme, decomposes hydrogen peroxide to water and molecular oxygen (276). It appears to be less important to the red cell than peroxidase, presumably because it is effective only when the peroxide concentration is relatively high (277). Individuals with hereditary acatalasemia do not develop methemoglobinemia or hemolytic disease; an increase in glutathione peroxidase levels may compensate in part for the lack of catalase (278). Some evidence suggests, however, that erythrocyte catalase may be important in preventing oxidant damage to somatic tissues (279). Also, the level of catalase increases with physical conditioning, suggesting a physiologically significant role for erythrocyte catalase (280).
Catalase consists of a tetramer composed of 60,000-dalton subunits, with four heme groups per tetramer. It is encoded by a gene on chromosome 11 (281). Catalase is a major component of erythrocyte band 4.5 seen on Coomassie-stained gels of erythroid membrane proteins, as the enzyme interacts with the membrane in a calcium- and pH-dependent manner (282). Catalase also comprises a major reservoir of erythrocyte protein-bound NADPH. Each tetrameric molecule of erythrocyte catalase contains four molecules of tightly bound NADPH. Although it is not essential for enzymatic conversion of peroxide to oxygen, the NADPH appears to protect catalase from inactivation by peroxide (283).
Glutathione Metabolism
Glutathione is the principal reducing agent in erythrocytes and the essential cofactor in the glutathione peroxidase reaction. Reduced glutathione (GSH) is a tripeptide (γ-glutamyl-cysteinyl-glycine). Two ATP-dependent enzymatic reactions are required for the de novo synthesis of glutathione:
1. glutamic acid + cysteine → γ-glutamyl-cysteine
2. γ-glutamyl-cysteine + glycine → GSH
Reaction 1 is catalyzed by glutamyl-cysteine synthetase, reaction 2 by glutathione synthetase. Both reactions can take place in normal erythrocytes (284). The capacity of normal red cells to synthesize glutathione exceeds the rate of turnover by 150-fold. Deficiencies of both of these glutathione synthetic enzymes have been associated with hemolytic anemia (see Chapter 32).
|
|
|
Figure 7.27. Glutathione metabolism in the erythrocyte. |
In the course of reactions that protect hemoglobin from oxidation, GSH is oxidized, forming oxidized glutathione (GSSG), which consists of two GSH molecules joined by a disulfide linkage, and mixed disulfides with hemoglobin. GSSG rapidly leaves the erythrocyte (285). Thus, maintaining a continuous supply of GSH requires a system to reduce the oxidized forms of glutathione. Such a system is provided by glutathione reductase, which catalyzes the reduction of GSSG by NADPH, a product of the pentose phosphate pathway (Fig. 7.27). Glutathione reductase also catalyzes the reduction of hemoglobin-glutathione disulfides, yielding GSH and hemoglobin (286).
The gene for glutathione reductase maps to chromosome 8. This enzyme is a flavoprotein consisting of two identical peptide chains of 478 amino acids. Because of its flavin component, the activity of glutathione reductase depends on the dietary intake of riboflavin. Erythrocyte glutathione reductase activity may be increased by administration of riboflavin, even in apparently normal subjects (287,288). The enzyme is inhibited by hexavalent chromium in concentrations as low as 5 to 25 μg/mL (289).
Abnormally high levels of GSH have been noted in patients with a variety of diseases and conditions (290). Neonatal erythrocytes have a higher level of GSH than do adult erythrocytes. Increased GSH has also been noted in inherited pyrimidine 5′-nucleotidase deficiency and in association with many dyserythropoietic anemias. The molecular bases of the increased levels of GSH in these conditions remain unclear.
Energy Metabolism
Although the mature red cell contains the enzymes required for glycogen metabolism, the balance between synthesis and utilization is such that no significant amount of glycogen accumulates within the cell under normal circumstances (291). Glycogen may accumulate, however, in glycogen storage diseases types III and VI.
Lacking a storage compound, the normal erythrocyte must have constant access to glucose if its energy metabolism is to be sustained. As previously discussed, glucose enters the cell by means of a facilitated, carrier-mediated transport mechanism. Insulin or other hormones are not required, and transport is not ordinarily the rate-limiting factor in glucose utilization. Without mitochondria, erythrocytes must depend on two less efficient pathways for production of high-energy compounds, the anaerobic glycolytic (Embden-Meyerhof) pathway and the aerobic pentose phosphate pathway, also known as the hexose monophosphate shunt or the phosphogluconate pathway (Fig. 7.28). Under normal circumstances, about 90% of glucose entering the red cell is metabolized by the anaerobic pathway and 10% by the aerobic pathway (292). Under conditions of oxidative stress, however, the oxidative pentose pathway may account for up to 90% of glucose consumption (293).
|
|
|
Figure 7.28. Energy metabolism in the erythrocyte. Main pathways are shown as boxes; major substrates and products of each are shown outside the boxes. More details of the pathways are given in Figures 7.27 and 7.29. ADP, adenosine diphosphate; ATP, adenosine triphosphate; 2,3-DPG, 2,3-diphosphoglycerate; F-6-P, fructose 6-phosphate; G-6-P, glucose 6-phosphate; Ga-3-P, glyceraldehydes 3-phosphate; Hb, hemoglobin; MHb, methemoglobin; NAD, NADH, nico-tinamide adenine dinucleotide; NADP, NADPH, nicotinamide adenine dinucleotide phosphate. |
|
|
|
Figure 7.29. Energy metabolism in the erythrocyte. Circled numbers represent reactions referred to in the text. Enzymes are designated by abbreviations and are shown in bold to the right or above arrows representing reactions. Cofactors are shown in bold to the left of the arrows or above the enzymes. ADP, adenosine diphosphate; Ald, aldolase; ATP, adenosine triphosphate; DPGM, diphosphoglyceratemutase; Ep, epimerase; G6PD, glucose 6-phosphate dehydrogenase; HK, hexokinase; LDH, lactate dehydrogenase; NAD-NADH, nicotinamide adenine dinucleotide; NADP-NADPH, nicotinamide adenine dinucleotide phosphate; PFK, phosphofructokinase; PGD, phosphoglyceraldehyde dehydrogenase; 6PGD, 6-phosphogluconate dehydrogenase; PGI, phosphoglucose isomerase; PGK, phosphoglycerate kinase; PGM, phosphoglyceromutase; PK, pyruvate kinase; PRI, phosphoribose isomerase; TPI, triosephosphate isomerase; TRA, transaldolase; TRK, transketolase. |
Three important products are formed by the anaerobic glycolytic pathway: NADH, a cofactor in the methemoglobin reductase reaction; ATP, the major high-energy phosphate nucleotide that powers the cation pump; and 2,3-DPG, a regulator of hemoglobin function (Fig. 7.28). For each molecule of glucose that enters the pathway, two molecules of NADH are generated (Fig. 7.29, reaction 6). The yields of ATP and 2,3-DPG vary depending on the activity of the Rapoport-Luebering shunt (Fig. 7.29, reactions 7b and 7c), a side pathway unique to the red cell. Two molecules of ATP are used in the early steps of glycolysis (Fig. 7.29, reactions 1 and 3), and a maximum of four molecules is produced late in the pathway (two in reaction 7a and two in reaction 10). Thus, at maximum efficiency, a net yield of two molecules of ATP may be expected for each molecule of glucose catabolized. This net yield may be decreased, however, to the extent that 2,3-DPG is formed (Fig. 7.29, reactions 7b and 7c). For this reason, the DPG-forming step is sometimes referred to as an energy clutch.
Of the 11 enzymes in the glycolytic pathway, three appear to be particularly important in regulation of glycolytic rate. These are hexokinase, phosphofructokinase, and pyruvate kinase (Fig. 7.29, reactions 1, 3, and 10). Hexokinase is the least active enzyme in the series and is therefore often rate-limiting (Table 7.10). It is inhibited by its product, glucose 6-phosphate, and is stimulated by one of its substrates, Mg-ATP. The activity of phosphofructokinase is greatly affected by intracellular pH. Because the pH optimum of this enzyme is 8.0, the activity of the enzyme and the overall rate of glycolysis tend to increase with increased pH (alkalosis) and decrease with decreased pH (acidosis). Phosphofructokinase may also be activated by the product of the further phosphorylation of fructose 6-phosphate (294). Pyruvate kinase is strongly inhibited by its product, ATP, and pyruvate kinase activity may therefore be related to the rate at which ATP is used in the cell’s metabolic processes.
|
Table 7.10 Activity of Glycolytic Enzymes in Erythrocytes of Normal Adults |
||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||
The importance of glycolysis to the red cell is illustrated by the manifestations of inherited disorders, in each of which the activities of one of the glycolytic enzymes is impaired. Under such circumstances, the viability of the red cell is reduced, and hemolytic anemia results (Chapter 32).
The most important product of the pentose phosphate pathway in erythrocytes is reduced NADPH. The red cell lacks the reactions to use NADPH for energy; instead, NADPH, by serving as a cofactor in the reduction of oxidized glutathione (GSSG), is a major reducing agent in the cell and the ultimate source of protection against oxidative attack. The utilization of NADPH is the main stimulus to the utilization of glucose 6-phosphate by the pathway. Redox agents such as methylene blue, cysteine, ascorbate, and others induce up to 20-fold increase in pentose metabolism, presumably by bringing about oxidation of glutathione (295,296). This metabolic flexibility allows the red cell to respond to unexpected oxidant challenge. The initial reaction of the pentose phosphate pathway is catalyzed by the enzyme glucose 6-phosphate dehydrogenase (G6PD). Hereditary deficiency of red cell G6PD is one of the most common genetic abnormalities in the world associated with hemolysis (Chapter 32).
A second function of the pentose pathway is the conversion of hexoses to pentoses. For the most part, the latter are recycled into the glycolytic pathway; however, d-ribose 5-phosphate may be used for nucleotide synthesis.
References
1. Ponder E. Hemolysis and related phenomena. New York: Grune & Stratton, 1948.
2. Evans E, Fung YC. Improved measurements of erythrocyte geometry. Microvasc Res 1972;4:335–347.
3. Canham PB, Burton AC. Distribution of size and shape in populations of normal human red cells. Circ Res 1968;22:405.
4. Price-Jones C. Red blood cell diameters. London: Oxford Medical Publications, 1933.
5. England JM, Down MC. Red cell volume distribution curves and the measurement of anisocytosis. Lancet 1974;l:701.
6. Bessman JD, Johnson RK. Erythrocyte volume distribution in normal and abnormal subjects. Blood 1975;46:369–379.
7. Mayhew TM, Mwamengele GL, Self TJ, Travers JP. Stereological studies on red corpuscle size produce values different from those obtained using haematocrit- and model-based methods. Br J Haematol 1994;86:355–360.
8. Bull BS, Brailsford D. Red blood cell shape. In: Agre P, Parker JC, eds. Red blood cell membranes: structure, function, clinical implications. New York: Marcel Dekker, 1989.
9. Lew VL, Raftos JE, Sorette M, et al. Generation of normal human red cell volume, hemoglobin content, and membrane area distributions by “birth” or regulation? Blood 1995;86:334–341.
10. Danielli JF. The bilayer hypothesis of membrane structure. Hosp Pract 1973;Jan.
11. d’Onofrio G, Chirillo R, Zini G, et al. Simultaneous measurement of reticulocyte and red blood cell indices in healthy subjects and patients with microcytic and macrocytic anemia. Blood 1995;85:818–823.
12. Lenard JG. A note on the shape of the red cell. Bull Math Biol 1974;36:55.
13. Schmid-Schonbein H. Erythrocyte rheology and the optimization of mass transport in the microcirculation. Blood Cells 1975;1:285.
14. Skalak R, Branemark P-I. Deformation of red blood cells in capillaries. Science 1969;164:717.
15. Xue M, Kato Y, Terada N, et al. Morphological study by an “in vivo cryotechnique” of the shape of erythrocytes circulating in large blood vessels. J Anat 1998;193(part 1):73–79.
16. Evans EA. Structure and deformation properties of red blood cells: concepts and quantitative methods. Methods Enzymol 1989;173:3.
17. Brecher G, Bessis M. Present status of spiculed red cells and their relationship to the discocyte-echinocyte transformation. A critical review. Blood 1972;40: 333.
18. Nakao M, et al. Effect of inosine and adenine on adenosine triphosphate regeneration and shape transformation in long stored erythrocytes. Biochim Biophys Acta 1959;32:564; Nature 1960;187:945,954; Nature 1961;191:284; J Biochem (Tokyo) 1961;49:487; 1960;48:672.
19. White JG. Scanning electron microscopy of erythrocyte deformation: the influence of a calcium ionophore, A2387. Semin Hematol 1976;13:121.
20. Bessis M. Red cell shapes. An illustrated classification and its rationale. Nouv Rev Fr Hematol 1972;12:721.
21. Weed RI, Bessis M. The discocyte-stomatocyte equilibrium of normal and pathological red cells. Blood 1973;41:471.
22. Wong P. A basis of echinocytosis and stomatocytosis in the disc-sphere transformations of the erythrocyte. J Theor Biol 1999;196:343–361.
23. Manno S, Takauwa Y, Mohandas N. Identification of a functional role for lipid asymmetry in biological membranes: phosphatidylserine-skeletal protein inter actions modulate membrane stability. Proc Natl Acad Sci U S A 2002;99: 1943–1948.
24. Chasis JA, Shohet SB. Red cell biochemical anatomy and membrane properties. Annu Rev Physiol 1987;39:237.
25. Palek J, Lambert S. Genetics of the red cell membrane skeleton. Semin Hematol 1990;27:290.
26. Hoefner DM, Blank ME, Davis BM, Diedrich DF. Band 3 antagonists, p-azi-dobenzylphlorizin and DIDS, mediate erythrocyte shape and flexibility changes as characterized by digital image morphometry and microfiltration. J Membr Biol l994;141:91–100.
27. Heidi M, Van Dort DW, Knowles JA, et al. Analysis of integral membrane protein contributions to the deformability and stability of the human erythrocyte membrane. J Biol Chem 2001;276:46968–46974.
28. Jay AWL. Geometry of the human erythrocyte. I. Effect of albumin on cell geometry. Biophys J 1975;15:205.
29. Gorter E, Grendel F. On bimolecular layers of lipoids on the chromocytes of the blood. J Exp Med 1925;41:439.
30. Danielli JF, Davson HA. A contribution to the theory of permeability of thin films. J Cell Physiol 1935;5:483,495; J Cell Physiol 1936;7:393.
31. Stockenius W, Engelman DM. Current models for the structure of biological membranes. J Cell Biol 1969;42:613.
32. Singer SJ, Nicholson GL. The fluid mosaic model of the structure of cell membranes. Science 1972;175:720; Annu Rev Biochem 1974;43:805.
33. Bennett V. Spectrin-based membrane skeleton: a multipotential adaptor between plasma membrane and cytoplasm. Physiol Rev 1990;70:1029.
34. deVetten MP, Agre P. The Rh polypeptide is a major fatty acid-acylated erythrocyte membrane protein. J Biol Chem 1988;263:18,193.
35. Telen MJ. Phosphatidylinositol-linked red blood cell membrane proteins and blood group antigens. Immunohematology 1991;7:65.
36. Baker RF. Ultrastructure of the red blood cell. Fed Proc 1967;26:1785.
37. Robertson JD. The ultrastructure of biological membranes. Arch Intern Med 1972;129:202.
38. Hillier J, Hoffman JF. On the ultrastructure of the plasma membrane as determined by the electron microscope. J Cell Physiol 1953;42:203.
39. Weinstein RS, McNutt NS. Ultrastructure of red cell membranes. Semin Hematol 1970;7:259.
40. Terada N, Fujii Y, Ohno S. Three-dimensional ultrastructure of in situ membrane skeletons in human erythrocytes by quick-freezing and deep-etching method. Histol Histopathol 1996;11:787–800.
41. Cooper RA. Lipids of human red cell membrane: normal composition and variability in disease. Semin Hematol 1970;7:296.
42. Schwartz RS, et al. Plasma membrane phospholipid organization in human erythrocytes. Curr Top Hematol 1985;5:63.
43. Dodge JT, et al. The preparation and chemical characteristics of hemoglobin-free ghosts of human erythrocytes. Arch Biochem Biophys 1963;100:119.
44. Sweeley CC, Dawson G. Lipids of the erythrocyte. In: Jamieson GA, Greenwait TJ, eds. Red cell membrane. Philadelphia: JB Lippincott, 1969.
45. Van Deenen LLM, De Gier J. Chemical composition of lipids in red cells of various animal species. In: Bishop C, Surgenor DM, eds. The red blood cell. New York: Academic Press, 1964.
46. Ways P, Hanahan DJ. Characterization and quantification of red cell lipids in normal man. J Lipid Res 1964;5:318.
47. Bretscher MS. Membrane structure: some general principles. Science 1978; 181:623.
48. Rothman JE, Lenard J. Membrane asymmetry. Science 1977;195:743.
49. Cogan U, Schachter D. Asymmetry of lipid dynamics in human erythrocyte membranes studied with impermeant fluorophores. Biochemistry 1981;20:6396.
50. Fisher KA. Analysis of membrane halves: cholesterol. Proc Natl Acad Sci U S A 1976;73:173.
51. Haest CWM, Deuticke B. Possible relationship between membrane proteins and phospholipid asymmetry in the human erythrocyte membrane. Biochim Biophys Acta 1976;436:353.
52. Haest CWM, et al. Spectrin as a stabilizer of the phospholipid asymmetry in the human erythrocyte membrane. Biochim Biophys Acta 1978;509:21.
53. Cooper RA. Abnormalities of cell-membrane fluidity in the pathogenesis of disease. N Engl J Med 1977;297:371.
54. Dodge JT, Phillips GB. Composition of phospholipids and of phospholipid fatty acids and aldehydes in human red cell. J Lipid Res 1967;8:667.
55. Shohet SB. Hemolysis and changes in erythrocyte membrane lipids. N Engl J Med 1972;286:577,638.
56. Nelson GJ. Composition of neutral lipids from erythrocytes of common mammals. J Lipid Res 1967;8:374.
57. Bull BS, Brailsford JD. The biconcavity of the red cell: an analysis of several hypotheses. Blood 1973;41:833.
58. Cooper RA. Influence of increased membrane cholesterol on membrane fluidity and cell function in human red blood cells. J Cell Biochem 1978;8:413.
59. Tishler RB, Carlson FD. A study of the dynamic properties of the human red blood cell membrane using quasi-elastic light-scattering spectroscopy. Biophys J 1993;65:2586–6000.
60. Steck TL. The organization of proteins in the human red blood cell membrane. J Cell Biol 1974;62:1.
61. Glomset JA. The plasma lecithin: cholesterol acyltransferase reaction. J Lipid Res 1968;9:155.
62. Norum KR, et al. Familial lecithin-cholesterol acyltransferase deficiency. Scand J Clin Lab Invest 1967;20:231; 1968;21:327.
63. Cooper RA, Jandl JH. Bile salts and cholesterol in the pathogenesis of target cells in obstructive jaundice. J Clin Invest 1968;47:809.
64. Tarlov AR, Mulder E. Phospholipid metabolism in rat erythrocytes: quantitative studies of lecithin biosynthesis. Blood 1967;30:853.
65. Shohet SB, et al. Hereditary hemolytic anemia associated with abnormal lipids. Blood 1971;38:445.
66. Farquhar JW, Ahrens J Jr. Effects of dietary fats on human erythrocyte fatty acid patterns. J Clin Invest 1963;42:675.
67. Sutherland WH, Shilton ME, Nye ER, et al. Urban/rural differences in red blood cell fatty acid composition, plasma lipids and diet in Melanesian Fijians. Eur J Clin Nutr 1995;49:233–241.
68. Fairbanks G, et al. Electrophoretic analysis of the major polypeptides of the human erythrocyte membrane. Biochemistry 1971;10:2606.
69. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci U S A 1979;76:4350–4354.
70. Anstee DJ, et al. Abnormal blood-group-Ss-active sialoglycoproteins in the membrane of Miltenberger class III, IV and V human erythrocytes. Biochem J 1979;183:193.
71. Dahr W. Immunochemistry of sialoglycoproteins in human red blood cell membranes. In: Vengelen-Tyler V, Judd WJ, eds. Recent advances in blood group biochemistry. Arlington, VA: American Association of Blood Banks, 1986.
72. Dahr W, et al. Immunochemical aspects of the MNSs blood group system. J Immunogenet 1975;2:87.
73. Furthmayr H. Glycophorins A, B, C: a family of sialoglycoproteins. Isolation and preliminary characterization of trypsin-derived peptides. J Cell Biochem 1978;9:79.
74. Anstee DJ. The nature and abundance of human red cell surface glycoproteins. J Immunogenet 1990;17:219–225.
75. Tomita M, et al. Primary structure of human erythrocyte glycophorin A. Isolation and characterization of peptides and complete amino acid sequence. Biochemistry 1978;17:4756–4770.
76. Rahuel C, et al. Characterization of cDNA clones for human glycophorin A. Use for gene localization and for analysis of normal and glycophorin A deficient (Finnish type) genomic DNA. Eur J Biochem 1988;172:147–153.
77. Kudo S, Fukuda M. Structural organization of glycophorin A and B genes: glycophorin B gene evolved by homologous recombination at Alu repeat sequences. Proc Natl Acad Sci U S A 1989;86:4619–4623.
78. Vignal A, et al. Molecular analysis of glycophorin A and B gene structure and expression in homozygous Miltenberger class V (Mi. V) human erythrocytes. Eur J Biochem 1989;184(2):337–344.
79. Anderson RA, Lovrien RE. Glycophorin is linked by band 4.1 protein to the human erythrocyte membrane skeleton. Nature 1984;307:655–658.
80. Chasis JA, et al. Signal transduction by glycophorin A: role of extracellular and cytoplasmic domains in a modulatable process. J Cell Biol 1988;107:1351– 1357.
81. Groves JD, Tanner MJA. The effects of glycophorin A on the expression of the human red cell anion transporter (band 3) in Xenopus oocytes. J Membr Biol 1994;140:81–88.
82. Bruce LJ, Groves JD, Okubo Y, et al. Altered band 3 structure and function in glycophorin A- and B-deficient (Mk) red blood cells. Blood 1994;84: 916–922.
83. Lux SE, et al. Cloning and characterization of band 3, the human erythrocyte anion-exchange protein (AE1). Proc Natl Acad Sci U S A 1989;86: 9089–9093.
84. Tanner MJ, et al. The complete amino acid sequence of the human erythrocyte membrane anion-transport protein deduced from the DNA sequence. Biochem J 1988;256:703–712.
85. Beckmann R, Smythe JS, Anstee DJ, Tanner MJ. Functional cell surface expression of band 3, the human red blood cell anion exchange protein (AE1), in K562 erythroleukemia cells: band 3 enhances the cell surface reactivity of Rh antigens. Blood 1998;92:4428–4438.
86. Beckmann R, Smythe JS, Anstee DJ, Tanner MJ. Coexpression of band 3 mutants and Rh polypeptides: differential effects of band 3 on the expression of the Rh complex containing D polypeptide and the Rh complex containing CcEe polypeptide. Blood 2001;97:2496–2505.
87. Pawloski JR, Hess DT, Stamler JS. Export by red blood cells of nitric oxide bioactivity. Nature 2001;409:622–626.
88. Palek J, Jarolim P. Clinical expression and laboratory detection of red cell membrane protein mutations. Semin Hematol 1993;30:249–283.
89. Gallagher, PG, Update on the clinical spectrum and genetics of red blood cell membrane disorders. Curr Hematol Rep 2004;3(2):85–91.
90. Issitt PD. Null red blood cell phenotypes: associated biological changes. Transfus Med Rev 1993;7:139–155.
91. Cherif-Zahar B, et al. Molecular cloning and protein structure of a human blood group Rh polypeptide. Proc Natl Acad Sci U S A 1990;87:6243– 6247.
92. Smythe JS, Avent ND, Judson PA, et al. Expression of RHD and RHCE gene products using retroviral transduction of K562 cells establishes the molecular basis of Rh blood group antigens. Blood 1996;87:2968–2973.
93. Cartron JP. Defining the Rh blood group antigens. Biochemistry and molecular genetics. Blood Rev 1994;8:199–212.
94. Cherifzahar B, Raynal V, Gane P, et al. Candidate gene acting as a suppressor of the Rh locus in most cases of Rh-deficiency. Nat Genet 1996;12: 168–173.
95. Marini AM, Matassi G, Raynal V, et al. The human Rhesus-associated RhAG protein and a kidney homologue promote ammonium transport in yeast [see comments]. Nat Genet 2000;26:341–344.
96. Westhoff CM, Ferreri-Jacobia M, Mak DO, Foskett JK. Identification of the erythrocyte Rh blood group glycoprotein as a mammalian ammonium transporter. J Biol Chem 2002;277:12499–12502.
97. Soupene E, King N, Feild E, et al. Rhesus expression in a green alga is regulated by CO(2). Proc Natl Acad Sci U S A 2002;99:7769–7773.
98. Hartwig JH. Actin-binding proteins 1—spectrin superfamily protein profile. Protein Profile 1995;2:703–800.
99. Beck KA, Nelson WJ. The spectrin-based membrane skeleton as a membrane protein-sorting machine. Am J Physiol Cell Physiol 1996;39:C1263– C1270.
100. Huebner K, et al. The alpha-spectrin gene is on chromosome 1 in mouse and man. Proc Natl Acad Sci U S A 1985;82:3790–3793.
101. Prchal JT, et al. Isolation and characterization of cDNA clones for human erythrocyte beta spectrin. Proc Natl Acad Sci U S A 1987;84:7468–7472.
102. Winkelmann JC, et al. Molecular cloning of the cDNA for human beta spectrin. Blood 1988;72:328–334.
103. Schafer DA, Cooper JA. Control of actin assembly at filament ends. Annu Rev Cell Dev Biol 1995;11:497–515.
104. Conboy JG. Structure, function, and molecular genetics of erythroid membrane skeletal protein 4.1 in normal and abnormal red blood cells. Semin Hematol 1993;30:58–73.
105. Lambert S, Bennett V. From anemia to cerebellar dysfunction. A review of the ankyrin gene family. Eur J Biochem 1993;211:1–6.
106. Delaunay J. The molecular basis of hereditary red cell membrane disorders. Blood Rev 2007;21:1–20.
107. Liu SC, Derick LH, Palek J. Dependence of the permanent deformation of red blood cell membranes on spectrin dimer-tetramer equilibrium: implication for permanent membrane deformation of irreversibly sickled cells. Blood 1993; 81:522–528.
108. Advani R, Sorenson S, Shinar E, et al. Characterization and comparison of the red blood cell membrane damage in severe human alpha- and beta-thalassemia. Blood 1992;79:1058–1063.
109. Mathai JC, Mori S, Smith BL, et al. Functional analysis of aquaporin-1 deficient red cells. The Colton-null phenotype. J Biol Chem 1996;271:1309–1313.
110. Agre P, Preston GM, Smith BL, et al. Aquaporin CHIP: the archetypal molecular water channel. Am J Physiol 1993;265:F463–F476.
111. Tanner MJ. Molecular and cellular biology of the erythrocyte anion exchanger (AE1). Semin Hematol 1993;30:34–57.
112. Mueckler M, Caruso C, Baldwin SA, et al. Sequence and structure of a human glucose transporter. Science 1985;229:941–945.
113. Baly DL, Horuk R. The biology and biochemistry of the glucose transporter. Biochim Biophys Acta 1988;947:571–590.
114. Elbrink J, Bihler I. Membrane transport: its relation to cellular metabolic rates. Science 1975;188:1177–1184.
115. Klepper J, Willemsen M, Verrips A, et al. Autosomal dominant transmission of GLUT1 deficiency. Hum Mol Genet 2001;10:63–68.
116. Orringer EP, Parker JC. Ion and water movements in red blood cells. Prog Hematol 1973;8:1–23.
117. Kaplan JH. Active transport of sodium and potassium. In: Agre P, Parker JC, eds. Red blood cell membranes: structure, function, clinical implications. New York: Marcel Dekker, 1989.
118. Hoffman JF, et al. The Na:K pump in red cells is electrogenic. Fed Proc 1979;38:2440–2441.
119. Proverbio F, Hoffman JF. Membrane compartmentalized ATP and its preferential use by the Na, K-ATPase of human red cell ghosts. J Gen Physiol 1977;69:605–632.
120. Jorgenson PL. Mechanism of the Na+, K+ pump. Protein structure and conformation of the pure (Na+, K+)-ATPase. Biochim Biophys Acta 1982;694: 27.
121. Mercer RW, et al. Rat-brain Na, K-ATPase beta-chain gene: primary structure, tissue-specific expression, and amplification in ouabain-resistant HeLa C+ cells. Mol Cell Biol 1986;6:3884–3890.
122. Hardwicke PMD, Freytag JW. A proteolipid associated with Na, K-ATPase is not essential for ATPase activity. Biochem Biophys Res Commun 1982;102: 250–257.
123. Kawakami K, et al. Primary structure of the alpha-subunit of human Na, K-ATPase deduced from cDNA sequence. J Biochem 1986;100:389–397.
124. Brahm J. Urea permeability of human red cells. J Gen Physiol 1983;82:1–23.
125. Olives B, Neau P, Bailly P, et al. Cloning and functional expression of a urea transporter from human bone marrow cells. J Biol Chem 1994;269:31649– 31652.
126. Olives B, Martial S, Mattei MG, et al. Molecular characterization of a new urea transporter in the human kidney. FEBS Lett 1996;386:156–160.
127. Schrier SL. Human erythrocyte membrane enzymes: current status and clinical correlates. Blood 1977;50:227–237.
128. Chow F-L, et al. The acetylcholinesterase defect in paroxysmal nocturnal hemoglobinuria: evidence that the enzyme is absent from the cell membrane. Blood 1985;66:940–945.
129. Rosse WF. Phosphatidylinositol-linked proteins and paroxysmal nocturnal hemoglobinuria. Blood 1990;75:1595–1601.
130. Telen MJ. Erythrocyte blood group antigens: polymorphisms of functionally important molecules. Semin Hematol 1996;33:302–314.
131. Mallozzi C, Di Stasi AM, Minetti M. Free radicals induce reversible membrane-cytoplasm translocation of glyceraldehyde-3-phosphate dehydrogenase in human erythrocytes. Arch Biochem Biophys 1995;321:345–352.
132. Plut D, et al. Evidence for the participation of cytosolic protein kinase in membrane protein phosphorylation in intact erythrocytes. Eur J Biochem 1978;82: 333–337.
133. Ralston GB. Spectrin-actin interactions. In: Harris JR, ed. Blood cell biochemistry 1. Erythroid cells. New York: Plenum Press, 1990.
134. Mercer RW, et al. Na, K-ATPase structure. In: Agre P, Parker JC, eds. Red blood cell membranes: structure, function, clinical implications. New York: Marcel Dekker, 1989.
135. Minetti G, Low P. Erythrocyte signal transduction pathways and their possible functions. Curr Opin Hematol 1997;4:116–121.
136. Sager G. Receptor binding sites for beta-adrenergic ligands on human erythrocytes. Biochem Pharmacol 1982;31:99–104.
137. Sager G, Jacobsen S. Effect of plasma on human erythrocyte beta-adrenergic receptors. Biochem Pharmacol 1985;34:3767–3771.
138. Hines PC, Zen Q, Burney SN, et al. Novel epinephrine and cyclic AMP-mediated activation of B-CAM/LU-dependent sickle (SS) RBC adhesion. Blood 2003;101:3281–3287.
139. Hardison RC. A brief history of hemoglobins: plant, animal, protist, and bacteria. Proc Natl Acad Sci U S A 1996;93:5675–5679.
140. Braunitzer G, et al. Der homologe Chemische aufbau der Peptidketten in Human hamoglobin. A Z Physiol Chem 1960;320:170.
141. Goodman M. Decoding the pattern of protein evolution. Prog Biophys Mol Biol 1981;37:105–164.
142. Hill RL, Buettner-Janusch J. Evolution of hemoglobin. Fed Proc 1964;23: 1236–1242.
143. Deisseroth A, Nienhuis A, Turner P, et al. Localization of the human a-globin structural gene to chromosome 16 in somatic cell hybrids by molecular hybridization assay. Cell 1977;12:205–218.
144. Deisseroth A, Nienhuis A, Lawrence J, et al. Chromosome localization of the human b globin gene on human chromosome 11 in somatic cell hybrids. Proc Natl Acad Sci U S A 1978;75:1456–1460.
145. Huehns ER, Beaven GH. Developmental changes in human hemoglobins. Clin Dev Med 1971;37:175.
146. Huehns ER, et al. Human embryonic hemoglobins. Nature 1964;201:1095– 1097.
147. Kamuzora H, Lehmann H. Human embryonic hemoglobins including a comparison by homology of the human ζ and α chains. Nature 1975;256:511–513.
148. Huehns ER, Farooqqi AM. Oxygen dissociation properties of human embryonic red cells. Nature 1975;254:335–337.
149. Weatherall DJ, Pembrey ME, Pritchard J. Fetal hemoglobin. Clin Hematol 1974;3:467.
150. Bunn HF, Jandl JH. Control of hemoglobin function within the red cell. N Engl J Med l970;282:1414–1421.
151. Schroeder WA, et al. Evidence for multiple structural genes for γ-chain of human fetal hemoglobin. Proc Natl Acad Sci U S A 1968;60:537–544.
152. Ricco G, Mazza V. Significance of a new type of human fetal hemoglobin carrying a replacement isoleucine-threonine at position 75 (E19) of the gamma chain. Hum Genet 1976;32:305–313.
153. Schroeder WA, et al. Further studies on the frequency and significance for the Tg-chain of human fetal hemoglobin. J Clin Invest 1979;63:268–275.
154. Alter BP. The Gg:Ag composition of fetal hemoglobin in fetuses and newborns. Blood 1979;54:1158–1163.
155. Jensen M, et al. The developmental change in the Gg and Ag globin proportions in hemoglobin F. Eur J Pediatr 1982;138:311–314.
156. Allen DW, Wyman J, Smith CA. The oxygen equilibrium of foetal and adult human hemoglobin. J Biol Chem 1953;203:81–87.
157. Tyuma I, Shimizu K. Different response to organic phosphates of human fetal and adult hemoglobins. Arch Biochem Biophys 1969;129:404–405.
158. Poyart C, Bursaux E, Guesnon P, Teisseire B. Chloride binding and the Bohr effect of human fetal erythrocytes and Hb FII solutions. Pflugers Arch 1978;37:169–175.
159. Bard H. The postnatal decline of hemoglobin F synthesis in normal full-term infants. J Clin Invest 1975;55:395–398.
160. Dover GJ, Boyer SH, Charache S, Heintzelman K. Individual variation in the production and survival of F-cells in sickle cell disease. N Engl J Med 1978;299:1428–1435.
161. Ley TJ, Nienhuis AW. Induction of hemoglobin F synthesis in patients with beta thalassemia. Annu Rev Med 1985;36:485–498.
162. De Rosa MC, Sanna MT, Messana I, et al. Glycated human hemoglobin (HbAlc): functional characteristics and molecular modeling studies. Biophys Chem 1998;72:323–335.
163. Schneck AG, Schroeder WA. The relation between the minor components of whole human adult hemoglobin as isolated by chromatography and starch block electrophoresis. J Am Chem Soc 1961;83:1472.
164. McDonald MJ, et al. Glycosylated minor components of human adult hemoglobin. J Biol Chem 1978;253:2327–2332.
165. Spicer KM, et al. Synthesis of hemoglobin Alc and related minor hemoglobins by erythrocytes. J Clin Invest 1979;64:40–48.
166. Bunn HF, et al. Further identification of the nature and linkage of the carbohydrate in hemoglobin A1c. Biochem Biophys Res Commun 1975; 67(1):103–109.
167. Al-Abed Y, VanPatten S, Li H, et al. Characterization of a novel hemoglobin-glutathione adduct that is elevated in diabetic patients. Mol Med 2001;7:619–623.
168. Fitzgibbons JF, et al. Red cell age-related changes of hemoglobins Aia+b and Alc in normal and diabetic subjects. J Clin Invest 1976;58:820–824.
169. Abraham EC, Cameron B, Abraham A, Stallings M. Glycosylated hemoglobins in heterozygotes and homozygotes for hemoglobin C with or without diabetes. J Lab Clin Med 1984;104:602–609.
170. Bernstein RE. Nonenzymatically glycosylated proteins. Adv Clin Chem 1987; 26:1–78.
171. Lawrence AM, Abraira C. New modalities in diabetes treatment. Am J Med 1988;85:153–158.
172. Salemans TH, et al. The value of HbAl and fructosamine in predicting impaired glucose tolerance—an alternative to OGTT to detect diabetes mellitus or gestational diabetes. Ann Clin Biochem 1987;24:447–452.
173. Bry L, Chen PC, Sacks DB. Effects of hemoglobin variants and chemically modified derivatives on assays for glycohemoglobin. Clin Chem 2001;47:153–163.
174. International Hemoglobin Information Center. IHIC variants list. Hemoglobin 1989;13:223–297.
175. Schmidt RM, Rucknagel DL, Necheles TF. Comparison of methodologies for thalassemia screening by Hb A2 quantitation. J Lab Clin Med 1975;86: 873–882.
176. Steinbert MH, Adams JG 3rd. Hemoglobin A2: origin, evolution, and aftermath. Blood 1991;78:2165–2177.
177. Huisman THJ, Schroeder WA, Brodie AN, et al. Microchromatography of hemoglobins. III. A simplified method for the determination of hemoglobin A2. J Lab Clin Med 1975;86:700–702.
178. Proficiency Testing Survey for Hemoglobinopathies. College of American Pathology, 2007.
179. Bollekens JA, Forget BG. Delta beta thalassemia and hereditary persistence of fetal hemoglobin. Hematol Oncol Clin North Am 1991;5:399–422.
180. Pembrey ME, McWade P, Weatherall DJ. Reliable routine measurement of small amounts of foetal haemoglobin by alkalai denaturation. J Clin Pathol 1972;25:738–740.
181. Bunn HF. Subunit assembly of hemoglobin: an important determinant of hematologic phenotype. Blood 1987;69:1–6.
182. Chernoff AL. A method for the quantitative determination of HgbA2. Ann NY Acad Sci 1964;119:557–560.
183. Robinson, AR, et al. A new technique for differentiation of hemoglobin. J Lab Clin Med 1957;50(5):745–752.
184. Boyer SH, Belding TK, Margolet L, Nouyes AN. Fetal hemoglobin restriction to a few erythrocytes (F cells) in normal human adults. Science 1975;188: 361–363.
185. Zago MA, Wood WG, Clegg JB, et al. Genetic control of F cells in human adults. Blood 1979;53:977–986.
186. Kleihaure E, Braun H, Betke K. Demonstration von fetalem Hamoglobin in den Erythrozyten eines Blutausstrichs. Klin Wochenschr 1957;35:637–638.
187. Lloyd-Evans P, Guest AR, Austin EB, Scott ML. Use of a phycoerythrin-conjugated anti-glycophorin A monoclonal antibody as a double label to improve the accuracy of FMH quantification by flow cytometry. Transfus Med 1999;9:155–160.
188. Lehmann H, Carrell RW. Variations in the structure of human hemoglobin. Br Med Bull 1969;25:14–23.
189. Bolton W, Perutz MR. Three dimensional fourier synthesis of horse deoxyhemoglobin at 2.8 Angstrom units resolution. Nature 1970;228:551–552.
190. Shaanan B. The structure of human oxyhaemoglobin at 2.1 A resolution. J Mol Biol l983;171:31–59.
191. Hrakl Z, Vodrazka Z. A study of the conformation of human globin in solution by optical methods. Biochim Biophys Acta 1967;133:527–534.
192. deCastro CM, Lee M, Devlin B, et al. A novel β-globin mutation, β Durham-N.C. [β114 LeuSPro], detected by PCR-SSCP, produces a dominant thalassemia-like phenotype. Blood 1994;83:1109–1116.
193. Perutz MF, Kilmartin JV, Nagai K, et al. Influence of globin structures on the state of the heme. Ferrous low spin derivatives. Biochemistry 1976;15:378–387.
194. Baldwin J, Chothia C. Haemoglobin: the structural changes related to ligand binding and its allosteric mechanism. J Mol Biol 1979;129:175–220.
195. Chance MR, Parkhurst LJ, Powers LS, Chance B. Movement of Fe with respect to the heme plane in the R-T transition of carp hemoglobin. An extended x-ray absorption fine-structure study. J Biol Chem 1986;261: 5689–5692.
196. Kurlekar N, Mehta BC. Hemoglobin A2 levels in iron-deficiency anemia. Isr J Med Res 1981;73:77–81.
197. Forget BG, Cavallesco C, Benz EJ Jr, et al. Studies of globin chain synthesis and globin mRNA content in a patient homozygous for hemoglobin Lepore. Hemoglobin 1978;2:117–128.
198. Shaeffer JR, McDonald MJ, Bunn HF. Assembly of normal and abnormal human hemoglobins. Trends Biochem Sci 1981;6:158.
199. Shaeffer JR, Kingston RE, McDonald MJ, Bunn HF. Electrostatic interactions in the assembly of human hemoglobin. Nature 1983;306:498.
200. Embury SH, Dozy AM, Miller J, et al. Concurrent sickle-cell anemia and alpha-thalassemia. Effect on severity of anemia. N Engl J Med 1982;306:270–274.
201. Steinberg MH. Case report: effects of iron deficiency and the -88 C→T mutation on HbA2 levels in beta-thalassemia. Am J Med Sci 1993;305:312–313.
202. Barcroft J. The respiratory function of the blood. Part II. Haemoglobin. London: Cambridge University Press, 1928.
203. Humpeler E, Amor H. Sex differences in the oxygen affinity of hemoglobin. Pflugers Arch 1973;343:151–156.
204. Astrup P. Red-cell pH and oxygen affinity of hemoglobin. N Engl J Med 1970;283:202–204.
205. Riggs A. Functional properties of hemoglobins. Physiol Rev 1965;45:619–673.
206. Benesch R, Benesch RE. The effect of organic phosphates from the human erythrocyte on the allosteric properties of hemoglobin. Biochem Biophys Res Commun 1967;26:162–167.
207. Brewer GJ, Eaton JW. Erythrocyte metabolism: interaction with oxygen transport. Science 1971;171:1205–1211.
208. Marschner JP, Seidlitz T, Rietbrock N. Effect of 2,3-diphosphoglycerate on O2-dissociation kinetics of hemoglobin and glycosylated hemoglobin using the stopped flow technique and an improved in vitro method for hemoglobin glycosylation. Int J Clin Pharmacol Ther 1994;32:116–121.
209. Chanutin A, Curnish RR. Effect of organic and inorganic phosphates on the oxygen equilibrium of human erythrocytes. Arch Biochem Biophys 1967;121: 96–102.
210. Haggarty NW, et al. The complete amino acid sequence of human erythrocyte diphosphoglycerate mutase. EMBO J 1983;2:1213–1220.
211. Reynolds CH. Activation of human erythrocyte 2,3-bisphosphoglycerate phosphatase at physiologic concentrations of substrate. Arch Biochem Biophys 1986;250:106–111.
212. Rosa R, Calvin MC. Comparative effects of two sulfhydryl reagents on the activities of a multifunctional red cell enzyme: bisphosphoglycerate mutase. Biochem Biophys Res Commun 1988;153:945–951.
213. Rose ZB. The enzymology of 2,3-bisphosphoglycerate. Adv Enzymol 1980;51: 211–253.
214. Yu KT, et al. 2,3-Diphosphoglycerate phosphatase/synthase: a potential target for elevating the diphosphoglycerate level in human red cells. J Pharmacol Exp Ther 1990;252:192–200.
215. Salhany JM, et al. The rate of deoxygenation of red blood cells. Effect of intracellular 2,3-diphosphoglycerate and pH. FEBS Lett 1971;16:257–261.
216. Bunn HF, et al. Hemoglobin function in stored blood. J Clin Invest 1969;48: 311–321.
217. Oski FA, et al. The in vitro restoration of red cell 2,3-diphosphoglycerate levels in banked blood. Blood 1971;37:52–58.
218. Tinmouth A, Chin-Yee I. The clinical consequences of the red cell storage lesion. Transfus Med Rev 2001;15:91–107.
219. Lenfant C, et al. Effect of altitude on oxygen binding by hemoglobin and on organic phosphate levels. J Clin Invest 1968;47:2652–2656.
220. Edwards UJ, Canon B. Oxygen transport during erythropoietic response to moderate blood loss. N Engl J Med 1972;287:115–119.
221. Torrance J, et al. Intraerythrocytic adaption to anemia. N Engl J Med 1970;283: 165–169.
222. Oski FA, et al. Exercise with anemia. Ann Intern Med 1971;74:44.
223. Perutz MF. Stereochemistry of cooperative effects in haemoglobin. Nature 1970;288:726–739.
224. Ackers GK. The energetics of ligand-linked subunit assembly in hemoglobin required a third allosteric structure. Biophys Chem 1990;37:371–382.
225. Daugherty MA, et al. Identification of the intermediate allosteric species in human hemoglobin reveals a molecular code for cooperative switching. Proc Natl Acad Sci U S A 1991;88:1110–1114.
226. Rorth M. Hemoglobin interactions and red cell metabolism. Ser Haematol 1972;5:1–104.
227. Surgenor DM. Transport of oxygen and carbon dioxide. In: Bishop C, Surgenor DM, eds. The red blood cell. New York: Academic Press, 1964.
228. Arnone A. X-ray studies of the interaction of CO2 with human deoxyhemoglobin. Nature 1974;247:143–145.
229. Boyer SH, et al. Developmental changes in human erythrocyte carbonic anhydrase levels: coordinate expression with adult hemoglobin. Dev Biol 1983;97: 250–253.
230. Brady HJ, et al. Expression of the human carbonic anhydrase I gene is activated late in fetal erythroid development and regulated by stage-specific trans-acting factors. Br J Haematol 1990;76:135–142.
231. Barlow JH, et al. Human carbonic anhydrase I cDNA. Nucleic Acids Res 1987;15:2386.
232. Lloyd J, et al. Nucleotide sequence and derived amino acid sequence of a cDNA encoding human muscle carbonic anhydrase. Gene 1986;41:233–239.
233. Venta PJ, et al. Polymorphic gene for human carbonic anhydrase II: a molecular disease marker located on chromosome 8. Proc Natl Acad Sci U S A 1983;80:4437–4440.
234. Behravan G, et al. Fine tuning of the catalytic properties of carbonic anhydrase. Studies of a Thr300-His variant of human isoenzyme II. Eur J Biochem 1990;190:351–357.
235. Tashian RE, et al. Inherited variants of human red cell carbonic anhydrases. Hemoglobin 1980;4:635–651.
236. Kiso Y, Yoshida K, Kaise K, et al. Erythrocyte carbonic anhydrase-I concentrations in patients with Graves’ disease and subacute thyroiditis reflect integrated thyroid hormone levels over the previous few months. J Clin Endocrinol Metab 1991;72:515–518.
237. Sly WS, et al. Carbonic anhydrase II deficiency in 12 families with the autosomal recessive syndrome of osteopetrosis with renal tubular acidosis and cerebral calcification. N Engl J Med 1985;313:139–145.
238. Carter ND, et al. Red cells genetically deficient in carbonic anhydrase II have elevated levels of a carbonic anhydrase indistinguishable from muscle CA III. Ann NY Acad Sci 1984;429:284–286.
239. Lee TS. End-tidal partial pressure of carbon dioxide does not accurately reflect PaCO2 in rabbits treated with acetazolamide during anaesthesia. Br J Anaesth 1994;73:225–226.
240. Discussion. In: Rorth M, Astrup P, eds. Oxygen affinity of hemoglobin and red cell acid base status. New York: Academic Press, 1972.
241. Lancaster JR Jr. Simulation of the diffusion and reaction of endogenously produced nitric oxide. Proc Natl Acad Sci U S A 1994;91:8137–8141.
242. Jia L, Bonaventura J, Stamler JS. S-nitrosohaemoglobin: a dynamic activity of blood involved in vascular control. Nature 1996;380:221–226.
243. Gow AJ, Luchsinger BP, Pawloski JR, et al. The oxyhemoglobin reaction of nitric oxide. Proc Natl Acad Sci U S A, 1999;96:9027–9032.
244. Gow AJ, Stamler JS. Reactions between nitric oxide and haemoglobin under physiological conditions. Nature 1998;391:169–173.
245. Gladwin MT. Role of the red blood cell in nitric oxide homeostasis and hypoxic vasodilation. Adv Exp Med Biol 2006;588:189–205.
246. Sonveaux P, et al. Transport and peripheral bioactivities of nitrogen oxides carried by red blood cell hemoglobin: role in oxygen delivery. Physiology (Bethesda) 2007;22:97–112.
247. Singel DJ, Stamler JS. Chemical physiology of blood flow regulation by red blood cells: the role of nitric oxide and S-nitrosohemoglobin. Annu Rev Physiol 2005;67:99–145.
248. Hsieh H-S, Jaffe ER. The metabolism of methemoglobin in human erythrocytes. In: Surgenor DM, ed. The red blood cell, ed 2. New York: Academic Press, 1975.
249. Rachmilewitz EA. Denaturation of the normal and abnormal hemoglobin molecule. Semin Hematol 1974;11:441–462.
250. Berzofsky JA, et al. Sulfheme proteins: optical and magnetic properties of sulfmyoglobin and its derivatives. J Biol Chem 1971;246:3367–3377.
251. Morell DB, et al. The structure of the chromophore of sulphmyoglobin. Biochim Biophys Acta 1967;136:121–130.
252. Berzofsky JA, et al. Sulfheme proteins. IV. The stoichiometry of sulfur incorporation and the isolation of sulfhemin, the prosthetic group of sulfmyoglobin. J Biol Chem 1972;247:3783–3791.
253. Sass MD, et al. TPNH-methemoglobin reductase deficiency. A new red cell enzyme defect. J Lab Clin Med 1967;70:760–767.
254. Mansouri A. Methemoglobin reduction under near physiological conditions. Biochem Med 1989;42:43–51.
255. Gibson QH. The reduction of methaemoglobin in red blood cells and studies on the cause of idiopathic methaemoglobinemia. Biochem J 1948;42:13–23.
256. Bull PC, et al. Cloning and chromosomal mapping of human cytochrome b5 reductase (DIA1). Ann Hum Genet 1988;52(Pt 4):263–268.
257. Yubisui T, et al. Molecular cloning of cDNAs of human liver and placenta NADH-cytochrome b5 reductase. Proc Natl Acad Sci U S A 1987;84: 3609–3613.
258. Hultquist DE, et al. Catalysis of methaemoglobin reduction by erythrocyte cytochrome b5B5 and cytochrome b5 reductase. Nature 1971;229:252–254.
259. Hegesh E, et al. Congenital methemoglobinemia with a deficiency of cytochrome b5. N Engl J Med 1986;314:757–761.
260. Miyata M, et al. Isolation and characterization of human liver cytochrome b5 cDNA. Pharmacol Res 1989;21:513–520.
261. Bloom GE, Zarkowsky HS. Heterogeneity of the enzymatic defect in congenital methemoglobinemia. N Engl J Med 1969;281:919–922.
262. Carrell RW, et al. Activated oxygen and hemolysis. Br J Haematol 1975;30: 259–264.
263. Fridovich I. Oxygen radicals, hydrogen peroxide and oxygen toxicity. In: Pryor WA, ed. Free radicals in biology, vol. I. New York: Academic Press, 1976.
264. Fridovich I. Superoxide dismutases. Annu Rev Biochem 1975;44:147–159.
265. Misra HP, Fridovich I. The oxidation of phenylhydrazine: superoxide and mechanism. Biochemistry 1976;15:681–687.
266. Lynch RE, et al. Inhibition of superoxide dismutase of methemoglobin formation from oxyhemoglobin. J Biol Chem 1976;251:1015; Biochemistry 1977;16:4563–4567.
267. Sutton HC, Roberts PB, Winterbourn CC. The rate of reaction of superoxide radical ion with oxyhaemoglobin and methaemoglobin. Biochem J 1976;155: 503–510.
268. Lynch RE, Fridovich I. Lysis of red cell vesicles exposed to superoxide anion. Blood 1977;50(suppl l):81.
269. Jabusch JR, et al. Some sulfhydryl properties and primary structure of human erythrocyte superoxide dismutase. Biochemistry 1980;19:2310–2316.
270. Cohen G, Hochstein P. Glutathione peroxidase. Biochemistry 1963;2:1420–1428.
271. Hill AS Jr, et al. The role of nonhemoglobin proteins and reduced glutathione in the protection of hemoglobin from oxidation in vitro. J Clin Invest 1964;43: 17–26.
272. Ganther HE, et al. Studies on selenium in glutathione peroxidase. Fed Proc 1974;33:694(abstr).
273. Chada S, et al. Isolation and chromosomal localization of the human glutathione peroxidase gene. Genomics 1990;6:268–271.
274. McBride OW, et al. Gene for selenium-dependent glutathione peroxidase maps to human chromosomes 3, 21, and X. Biofactors 1988;1:285–292.
275. Necheles TF, et al. Erythrocyte glutathione-peroxidase deficiency. Br J Haematol 1970;19:605–612.
276. Tudhope GR. Red cell catalase in health and in disease, with reference to the enzyme activity in anaemia. Clin Sci 1967;33:165–182.
277. Jacobs HS, et al. Oxidative hemolysis and erythrocyte metabolism in acatalasia. J Clin Invest 1965;44:1187.
278. Aebi HE, Wyss SR. Acatalasemia. In: Stanbury JB, et al., eds. The metabolic basis of inherited disease. New York: McGraw-Hill, 1978.
279. Agar NS, et al. Erythrocyte catalase. A somatic oxidant defense. J Clin Invest 1986;77:319–321.
280. Ohno H, et al. Physical training and fasting erythrocyte activities of free radical scavenging enzyme systems in sedentary men. Eur J Appl Physiol 1988;57: 173.
281. Schroeder WT, Saunders GF. Localization of the human catalase and apolipoprotein A-l genes to chromosome 11. Cytogenet Cell Genet 1987;44:231–233.
282. Aviram I, Shaklai N. The association of human erythrocyte catalase with the cell membrane. Arch Biochem Biophys 1981;212:329–337.
283. Kirkman HN, Gaetani GF. Catalase: a tetrameric enzyme with four tightly bound molecules of NADPH. Proc Natl Acad Sci U S A 1984;81:4343–4347.
284. Minnich V, et al. Glutathione biosynthesis in human erythrocytes. J Clin Invest 1971;50:507–513.
285. Srivastava SK, Beutler E. The transport of oxidized glutathione from the erythrocytes of various species in the presence of chromate. Biochem J 1969;114: 833–837.
286. Srivastava SK, Beutler E. Glutathione metabolism of the erythrocyte. The enzymatic cleavage of glutathione-haemoglobin preparations by glutathione reductase. Biochem J 1970;119:353–357.
287. Beutler E. Glutathione reductase: stimulation in normal subjects by riboflavin supplementation. Science 1969;165:613–615.
288. Beutler E. Effect of flavin compounds on glutathione reductase activity. J Clin Invest 1969;48:1957–1966.
289. Koutras GA, et al. Effect of sodium chromate on erythrocyte glutathione reductase. J Clin Invest 1964;43:323–331.
290. Valentine WN, Paglia DE. Syndromes with increased red cell glutathione (GSH). Hemoglobin 1980;4:799–804.
291. Moses SW, et al. Glycogen metabolism in the normal red blood cell. Blood 1972;40:836; 1974;44:275–284.
292. Murphy JR. Erythrocyte metabolism. J Lab Clin Med 1960;55:281,286.
293. Schuster R, Holzhutter HG, Jacobasch G. Interrelations between glycolysis and the hexose monophosphate shunt in erythrocytes as studied on the basis of a mathematical model. Biosystems 1988;22:19–36.
294. Tanaka KR, Zerez CR. Red cell enzymopathies of the glycolytic pathway. Semin Hematol 1990;27:165–185.
295. Davidson WD, Tanaka KR. Factors affecting pentose phosphate pathway activity in human red cells. Br J Haematol 1972;23:371–385.
296. Szeinberg A, Marks PA. Substances stimulating glucose catabolism by the oxidative reactions of the pentose phosphate pathway in human erythrocytes. J Clin Invest 1961;40:914–924.
