The renal handling of hemoglobin. II. Catabolism.

The fate of small doses of isotopically labeled isologous hemoglobin was studied in the rat. When haptoglobin depleted animals were given 2.0 mg of (59)Fe hemoglobin intravenously, nearly half was trapped by the kidneys. Kidney (59)Fe activity disappeared slowly over several weeks. Whatever iron was lost from the kidneys was largely reutilized. In contrast, the porphyrin of hemoglobin absorbed by the kidneys appeared to be rapidly catabolized, since 5 hr after the injection of (14)C or (59)Fe heme-labeled hemoglobin only a small fraction was recovered as hematin. Likewise, after injection of globin-labeled hemoglobin, rapid disappearance of kidney protein activity indicated that the absorbed globin was readily catabolized in situ.

The renal handling of hemoglobin. I. Glomerular filtration.

The glomerular filtration of hemoglobin (alpha(2)beta(2)) was studied under conditions in which its dissociation into alphabeta dimers was experimentally altered. Rats receiving hemoglobin treated with the sulfhydryl reagent bis(N-maleimidomethyl) ether (BME) showed a much lower renal excretion and prolonged plasma survival as compared with animals injected with untreated hemoglobin. Plasma disappearance was also prolonged in dogs receiving BME hemoglobin. Gel filtration data indicated that under physiological conditions, BME hemoglobin had impaired subunit dissociation. In addition, BME hemoglobin showed a very high oxygen affinity and a decreased rate of auto-oxidation. Glomerular filtration was enhanced under conditions which favor the dissociation of hemoglobin into dimers. Cat hemoglobin, which forms subunits much more extensively than canine hemoglobin, was excreted more readily by the rat kidney. The renal uptake of (59)Fe hemoglobin injected intra-arterially into rabbits varied inversely with the concentration of the injected dose.

Differences in the interaction of 2,3-diphosphoglycerate with certain mammalian hemoglobins.

The hemoglobins of man, horse, dog, rabbit, guinea pig, and rat all have relatively high (nonphysiologic) oxygen affinity when stripped of organic phosphates, and a strong reactivity with 2,3-diphosphoglycerate (2,3-DPG). Appropriately, their red cells contain high levels of 2,3-DPG. In contrast, the sheep, goat, cow, and cat have low oxygen affinity hemoglobins which interact weakly with 2,3-DPG, and low concentrations of red cell 2,3-DPG. These hemoglobins have structural differences at the NH(2)-terminus of the beta chain, a site where 2,3-DPG is thought to bind.

Signal transduction. How do cells sense oxygen?

How do organisms sense the amount of oxygen in the environment and respond appropriately when the amount of oxygen decreases (a condition called hypoxia)? In their Perspective, Zhu and Bunn discuss new findings (Ivan et al., Jaakkola et al.) that reveal how the HIF transcription factor, which switches on a group of hypoxia-response proteins, is itself regulated by changes in oxygen tension. The authors are in the Hematology Division of the Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115, USA. E-mail: zhu@calvin.bwh.harvard.edu, bunn@calvin.bwh.harvard.edu

Regulation of erythrocyte cation and water content in sickle cell anemia.

The pathophysiological events in sickle cell disease are critically dependent on the intracellular concentration of hemoglobin S, which varies inversely with cell cation and water content. Erythrocytes of SS homozygotes exposed to oxygen or carbon monoxide decrease their potassium and water content through a pathway for potassium transport that is activated by both cell swelling and decrease in internal pH. This pathway is not inhibited by ouabain either with or without bumetanide. When SS erythrocytes were separated according to density, the pH- and volume-dependent potassium transport was greatest in the least dense fraction and was reduced in the densest cells. This pathway, which does not depend on polymerization of sickle hemoglobin, may be important in regulating the cation and water content of SS erythrocytes.

Reaction of monosaccharides with proteins: possible evolutionary significance.

Measurements were made of the rate of condensation of various monosaccharides with amino groups of hemoglobin to form Schiff base linkages. The reactivity of each sugar was dependent on the extent to which it exists in the open (carbonyl) structure rather than in the ring (hemiacetal or hemiketal) structure. Among the 15 monosaccharides tested, aldoses showed higher reactivities than ketoses. Glucose was the least reactive of the aldohexoses. The emergence of glucose as the primary metabolic fuel may be due in part to the high stability of its ring structure which limits potentially deleterious nonenzymatic glycosylation of proteins.

Hemoglobin switching in sheep and goats: occurrence of hemoglobins A and C in the same red cell.

Sheep and goats switch from the synthesis of hemoglobin A (alpha(2)beta(2)(A)) to hemoglobin C (alpha(2)beta(2)(C)) when made anemic. We have demonstrated the existence of the asymmetrical hybrid hemoglobin, alpha(2)beta(A)beta(C), in the circulating red cells of anemic sheep. These erythroid cells, therefore, synthesized both A and C hemoglobin simultaneously. Thus, the switch appears to be mediated by selective gene expression rather than by a clonal or cellular selective mechanism.

Regulation of the erythropoietin gene: evidence that the oxygen sensor is a heme protein.

Erythropoietin (Epo), the hormone that stimulates red blood cell production, is synthesized in the kidney and liver in response to hypoxia. The human hepatoma cell line Hep3B regulates its production of Epo in a physiologic manner. Either hypoxia or cobalt chloride markedly increases expression of Epo mRNA as well as production of biologically active and immunologically distinct Epo protein. New protein synthesis is required before the induction of increased levels of hypoxia- or cobalt-induced Epo mRNA. Hypoxia, cobalt chloride, and nickel chloride appear to stimulate Epo production through a common pathway. The inhibition of Epo production at low partial pressures of oxygen by carbon monoxide provides evidence that a heme protein is integrally involved in the oxygen-sensing mechanism. This hypothesis is further supported by the finding that when heme synthesis is blocked, hypoxia-, cobalt-, and nickel-induced Epo production are all markedly inhibited. A model is proposed in which a ligand-dependent conformational change in a heme protein accounts for the mechanism by which hypoxia as well as cobalt and nickel stimulate the production of Epo.

The glycosylation of hemoglobin: relevance to diabetes mellitus.

Glucose reacts nonenzymatically with the NH2-terminal amino acid of the beta chain of human hemoglobin by way of a ketoamine linkage, resulting in the formation of hemoglobin AIc. Other minor components appear to be adducts of glucose 6-phosphate and fructose 1,6-diphosphate. These hemoglobins are formed slowly and continuously throughout the 120-day life-span of the red cell. There is a two- to threefold increase in hemoglobin AIc in the red cells of patients with diabetes mellitus. By providing an integrated measurement of blood glucose, hemoglobin AIc is useful in assessing the degree of diabetic control. Furthermore, this hemoglobin is a useful model of nonenzymatic glycosylation of other proteins that may be involved in the long-term complications of the disease.

Subunit dissociation of certain abnormal human hemoglobins.

The extent of dissociation of various hemoglobins into subunits was estimated from their elution volumes (V(e)) on G-100 Sephadex. Under the same controlled conditions carboxyhemoglobins A, A3 (A(1)), F, S, and C all had the same elution volumes. The carboxy and cyanmet derivatives of hemoglobin Kansas (a variant with very low oxygen affinity) had a relatively high V(e), indicating a decreased mean molecular weight and therefore an increased tendency to form dimers and even monomers. Conversely, the liganded derivatives of hemoglobin Chesapeake (a variant with high oxygen affinity) had a relatively low V(e), suggestive of an impaired degree of subunit dissociation. Deoxyhemoglobin Chesapeake had a V(e) identical with that of deoxyhemoglobin A. Cat hemoglobin, known to have an unusually low oxygen affinity, was found to have a higher V(e) than human, dog, rabbit, rat, or guinea pig hemoglobins. Haptoglobin is thought to bind alphabeta dimers in preference to the alpha(2)beta(2)-tetramer. The comparative haptoglobin affinities of the human hemoglobins were measured by competition between the test hemoglobin and radioactive reference hemoglobin for haptoglobin binding sites. Hemoglobins A, F, S, and C all seemed to bind equally readily, but hemoglobin Kansas and cat hemoglobin showed a higher affinity, and hemoglobin Chesapeake a lower affinity. These results are in accord with recently proposed models which predict that hemoglobins which have an increased degree of subunit dissociation will have a low oxygen affinity, and vice versa.

The interaction of 2,3-diphosphoglycerate with various human hemoglobins.

Oxygen equilibria were measured on a number of human hemoglobins, which had been "stripped" of organic phosphates and isolated by column chromatography. In the presence of 2 x 10(-4) M 2,3-diphosphoglycerate (2,3-DPG), the P(50) of hemoglobins A, A(2), S, and C increased about twofold, signifying a substantial and equal decrease in oxygen affinity. Furthermore, hemoglobins Chesapeake and M(Milwaukee-1) which have intrinsically high and low oxygen affinities, respectively, also showed a twofold increase in P(50) in the presence of 2 x 10(-4) M 2,3-DPG. In comparison to these, hemoglobins A(IC) and F were less reactive with 2,3-DPG while hemoglobin F(I) showed virtually no reactivity. The N-terminal amino of each beta-chain of hemoglobin A(IC) is linked to a hexose. In hemoglobin F(I) the N-terminal amino of each gamma-chain is acetylated. These results suggest that the N-terminal amino groups of the non-alpha-chains are involved in the binding of 2,3-DPG to hemoglobin.

Nonenzymatic glycosylation of erythrocyte membrane proteins. Relevance to diabetes.

Nonenzymatic glycosylation of proteins of the erythrocyte membrane was determined by incubating erythrocyte ghosts with [3H]borohydride. The incorporation of tritium into protein provides a reliable assay of ketoamine linkages. The membrane proteins from 18 patients with diabetes incorporated twice as much radioactivity as membrane proteins from normal erythrocytes. After acid hydrolysis, amino acid analysis showed that the majority of radioactivity was localized to glucosyllysine. Autoradiograms showed that all of the major proteins of the erythrocyte membrane, separated by electrophoresis on sodium dodecyl sulfate gels, contained ketoamine linkages. No protein bands in either normal or diabetic erythrocytes showed significant preferential labeling. Erythrocyte membranes from three patients with hemolytic anemia showed reduced incorporation of tritium from [3H]-borohydride, indicating decreased nonenzymatic glycosylation. Two patients with diabetes and hemolytic anemia had incorporation of radioactivity similar to that of normal individuals. In these groups of patients the incorporation of tritium into erythrocyte membrane proteins correlated with levels of hemoglobin AIc. Thus the modification of membrane proteins like that of hemoglobin depends on blood glucose levels as well as erythrocyte age. These studies show that the enhanced nonenzymatic glycosylation of proteins in diabetics extends beyond hemoglobin to the proteins of the erythrocyte membrane and probably affects other proteins that have slow turnover and are exposed to high concentrations of glucose.

Regulation of cation content and cell volume in hemoglobin erythrocytes from patients with homozygous hemoglobin C disease.

Erythrocytes from patients with homozygous hemoglobin C disease (CC cells) contain less K, Na, and water than do erythrocytes from normal subjects that contain only hemoglobin A (AA cells). In this paper, we provide evidence that the reduced K content and volume of CC cells are due to the activity in these but not in AA cells of a K transport system that is: (a) insensitive to ouabain and bumetanide, and (b) stimulated by increased cell volume, and dependent on internal pH (pHi). When the cation and water content of CC cells was increased (by making the membrane temporarily permeable to cations with nystatin) and the cells were then incubated in an isotonic medium containing 140 mM NaCl and 4 mM KCl, they lost K and shrunk back toward the original volume. This regulatory K and volume decrease was not inhibited by ouabain or bumetanide. When CC cells were incubated in a hypotonic medium, with ouabain and bumetanide, they also lost K and shrunk toward the original volume. This behavior was not observed in control AA cells. The ouabain- and bumetanide-resistant K efflux from CC cells was volume and pH dependent: K efflux from CC cells rose from 5-6 to 20-25 mmol/liter of cells X h, when cell volume was increased by increasing cell solute content (nystatin method) or by exposure to hypotonic media. In CC cells, the dependence of K efflux on pHo had a bell shape, with a maximal flux (20-25 mmol/liter of cells X h) at pHo 6.8-7.0. In contrast, the K efflux from control cells was minimal at pH 7.4 (1.2 mmol/liter of cells X h) and was slightly stimulated by both acid and alkaline pH. In order to study the effect of pHi and pHo on K efflux, CC cells were incubated with 4,4'-diisothiocyanostilbene-2,2'-disulfonic acid (150 microM) and acetazolamide (1 mM) at different pHi (6.7, 7.3, and 7.8), and resuspended in media with different pHo (6.75, 7.4, and 8): K efflux was stimulated by reducing pHi but was independent of pHo. The ouabain- and bumetanide-resistant K efflux from CC cells was not inhibited by some inhibitors of the Ca2+-activated K permeability. It seems likely that the genetically determined change in the primary structure of hemoglobin C directly or indirectly causes this modification in K transport. One possible mechanism could involve an electrostatic interaction between C hemoglobin and components of the erythrocyte membrane.

Characterization of glycosylated hemoglobins. Relevance to monitoring of diabetic control and analysis of other proteins.

Boronate affinity chromatography and ion exchange chromatography were used to measure the levels of glycosylated hemoglobins in normal and diabetic hemolysates, as well as the distribution of glucose adducts on alpha-NH2-valine and epsilon-NH2-lysine residues. When analyzed by ion exchange chromatography on BioRex 70 resin, the Hb Alc peak comprised 4.4 +/- 0.6% of 15 normal hemolysates and 9.1 +/- 2.1% of 15 diabetic hemolysates. The "Hb Alc" was rechromatographed on GlycoGel B boronate affinity resin that binds vicinal hydroxyl groups of covalently linked sugars. Only 70 +/- 5% of the hemoglobin adhered to the resin. Analysis by the thiobarbituric acid colorimetric test confirmed that the affinity resin effectively separated glycosylated from nonglycosylated hemoglobin. When corrected for nonglycosylated contaminants, the mean level of Hb Alc in normal hemolysates was 2.9 +/- 0.4%, a value considerably lower than those previously reported. In addition to Hb Alc, 5.2 +/- 0.5% of the remaining hemoglobin (Hb Ao) was glycosylated. In diabetics, glycosylated Ao was increased in parallel with Hb Alc. After reduction with [3H]borohydride and acid hydrolysis, glycosylated amino acids were first purified on Affi-Gel boronate affinity resin and then analyzed by ion exchange chromatography. The glucose adducts on Hb Ao were distributed as follows: alpha-chain N-terminal valine, 14%; alpha-chain lysines, 40%; beta-chain lysines, 46%. This study has revealed several pitfalls in the analysis of nonenzymatically glycosylated proteins. Peaks isolated by ion exchange chromatography or electrophoresis are likely to be contaminated by nonglycosylated proteins. Furthermore, both the thiobarbituric acid test and [3H]borohydride reduction show variable reactivity depending upon the site of the ketoamine-linked glucose.

The acetylation of hemoglobin by aspirin. In vitro and in vivo.

The chemical modification of hemoglobin by aspirin (ASA) has been studied, both in intact human red cells and in purified hemoglobin solutions. After incubation of red cells with 20 mM [acetyl-1minus14C]ASA, incorporation of radioactivity into hemoglobin was observed in agreement with the results of Klotz and Tam (1973. Proc. Natl. Acad. Sci. U. S. A. 70: 1313-1315). In contrast, no labeling of hemoglobin was seen when [carbosyl-14-C]ASA was used. These results indicate that ASA acetylates hemoglobin. The acetylated hemoglobin was readily separated from unmodified hemoglobin by both gel electrofocusing and by column chromatography. Quantitation of the extent of acetylation by densitometric scanning of gels agreed very well with estimates obtained from radioactivity measurements. Hemolysates prepared from red cells incubated with ASA showed normal oxygen affinity and heme-heme interaction. Purified acetylated hemoglobin had a slightly increased oxygen affinity and decreased heme-heme interaction. There was no difference in the rate of acetylation of oxy- and deoxyhemoglobin. ASA acetylated column-purified hemoglobin A more readily than hemoglobin in crude hemolysate, but less rapidly than purified human serum albumin. The rate of acetylation of hemoglobulin increased with pH up to approximately pH 8,5. Structural studies were done on hemoglobin incubated with 2.0 mM and 20 mM [acetyl-1-14-C]ASA. Alpha- and beta-chains were acetylated almost equally. Tryptic digests of purified acetylated subunits were fingerprinted on cellulose thin layer plates and autoradiographed. Both alpha- and beta-chains showed a number of radioactive spots that were either ninhydrin negative or weakly ninhydrin positive. These results indicate that hemoglobin is acetylated at a number of sites, probably at the epislon-amino group of lysine residues. To determine whether ASA acetylates hemoglobin in vivo, hemolysates of 14 patients on long-term high-dose ASA therapy were analyzed by gel electrofocusing and compared to specimens of individuals not receiving ASA. The ASA-treated group had a twofold increase in a minor hemoglobin component having an isoelectric point lower than that of hemoglobin A, and indistinguishable from the minoe component which appears when hemoglobin is incubated with ASA in vitro.

Hemoglobin Syracuse (alpha2beta2-143(H21)His leads to Pro), a new high-affinity variant detected by special electrophoretic methods. Observations on the auto-oxidation of normal and variant hemoglobins.

Family members from four generations were found to have polycythemia and increased whole blood O2 affinity (P50; 11 mm Hg; normal, 27 mm Hg). No abnormal hemoglobin bands were seen after electrophoresis on starch gel at pH 8.6 or agar gel at pH 6.0. Analysis of the oxygenated hemolysate by isoelectric focusing on polyacrylamide gel revealed two closely spaced bands. When deoxygenated hemolysate was analyzed in oxygen-free gels, the two components were more widely separated. About 40% of the patient's hemoglobin focused at a more acid pH than hemoglobin A, indicating a hemoglobin variant with impaired Bohr effect. Chromatography of globin in 8 M urea revealed two beta-chain peaks, the first of which was eluted at a lower buffer molarity than normal beta chain. Fingerprints of tryptic digests of the aminoethylated chains were done on silica gel thin-layer plates. Tp 14 from the abnormal beta chain had slower electrophoretic mobility and a greater Rf value. Amino acid analyses of this peptide gave values identical with those of betaTp 14, except that it contained one proline residue and no histidine. Since the one His in betaTp 14 is in position 143, hemoglobin Syracuse in alpha2beta2-143 His leads to Pro. Native Hb Syracuse could be separated from hemoglobin A on a carboxymethylcellulose column. The inclusion of 0.1 mM EDTA in the preparative buffers proved very useful in reducing the formation of methemoglobin. Oxygen equilibria of purified hemoglobin Syracuse showed high oxygen affinity (P50 value 12% that of hemoglobin A) and lack of cooperativity between subunits (Hill's n equals 1.1). The alkaline Bohr effect was about half that of hemoglobin A. The proline substitution at betaH21 disrupts the helical configuration and probably prevents the formation of salt bonds that are important in stabilizing the deoxy structure and contribute to the alkaline Bohr effect. Since beta143 His is a binding site for 2,3-diphosphoglycerate (2,3-DPG), it is not suprising that hemoglobin Syracuse had markedly impaired reactivity with 2,3-DPG. Hemoglobin Syracuse auto-oxidized more slowly than hemoglobin A, probably reflecting a slower rate of dissociation of oxygen from fully liganded hemoglobin.

Oxygen equilibrium of hemoglobin E.

Oxygen equilibrium was determined on hemoglobin of individuals both heterozygous and homozygous for hemoglobin E. The whole blood oxyhemoglobin dissociation curve of AE blood was identical to that of normal AA blood. E hemoglobin, isolated by diethylaminoethyl Sephadex and carboxymethyl cellulose column chromatography, had oxygen affinity, heme-heme interaction, and Bohr effect identical to those of hemoglobin A prepared from the same column. Furthermore, the two hemoglobins had equal reactivity with 2,3-diphosphoglycerate. Phosphate-free hemolysates of blood from E and A homozygotes also had identical oxygen saturation curves. These results do not confirm earlier reports that hemoglobin E has an abnormally low oxygen affinity.

Effects of cyanate and 2,3-diphosphoglycerate on sickling. Relationship to oxygenation.

Cyanate and 2,3-diphosphoglycerate (2,3-DPG) both influence the oxygen affinity of hemoglobin. The studies presented here concern the effects of these compounds on the sickling phenomenon. The inhibitory effect of cyanate on sickling is largely due to the fact that it increases the percentage of oxyhemoglobin S at a given oxygen tension. In addition, cyanate inhibits sickling by a mechanism that is independent of oxygenation. In this paper, we have demonstrated that the viscosity of carbamylated sickle blood was lower than that of non-carbamylated controls at the same oxygen saturation. Furthermore, carbamylation resulted in an increase in the minimum concentration of deoxy-sickle hemoglobin required for gelation. Like cyanate, 2,3-DPG affected sickling of intact erythrocytes by two mechanisms. Since 2,3-DPG decreases the percentage of oxyhemoglobin S at a given oxygen tension, sickling is enhanced. In addition, 2,3-DPG had a direct effect. When the intracellular 2,3-DPG concentration was increased in vitro, a greater percentage of cells were sickled at a given oxygen saturation. Conversely, sickling was inhibited in cells in which 2,3-DPG was artificially lowered. These data indicate that the enhancement of sickling by 2,3-DPG is in part independent of its influence on oxygen affinity.

Hemoglobin function in stored blood.

Serial oxygen dissociation curves were performed on blood units preserved in acid-citrate-dextrose (ACD), ACD-adenine, and ACD-adenine-inosine. Dividing blood from a single donor into two or more bags allowed direct comparison between preservatives. During the 1st wk of storage in ACD, a progressive increase in oxygen affinity was observed. Thereafter, little further change was noted. Oxygen affinity increased even more rapidly during initial storage in ACD-adenine. However, with the inclusion of inosine as a preservative, oxygen affinity remained unaltered during the first 2 wk. Increases in oxygen affinity correlated well with falling levels of red cell 2,3-diphosphoglycerate (2,3-DPG) during storage. No significant changes in glutathione, reduced form (GSH), or A3 (A(I)) hemoglobin levels were noted during the first 3 wk of storage. No significant accumulation of ferrihemoglobin was detected. When blood stored 20 days in ACD or ACD-adenine was incubated with inosine for 60 min at 37 degrees C, 2,3-DPG and adenosine triphosphate (ATP) were resynthesized, and oxygen affinity was decreased. The distribution of 2,3-DPG in fresh and stored red cells appeared to influence experimental values for Hill's n, a measure of heme-heme interaction.

Nonenzymatic glycation of human lens crystallin. Effect of aging and diabetes mellitus.

We have examined the nonenzymatic glycation of human lens crystallin, an extremely long-lived protein, from 16 normal human ocular lenses 0.2-99 yr of age, and from 11 diabetic lenses 52-82-yr-old. The glucitol-lysine (Glc-Lys) content of soluble and insoluble crystallin was determined after reduction with H-borohydride followed by acid hydrolysis, boronic acid affinity chromatography, and high pressure cation exchange chromatography. Normal lens crystallin, soluble and insoluble, had 0.028 +/- 0.011 nanomoles Glc-Lys per nanomole crystallin monomer. Soluble and insoluble crystallins had equivalent levels of glycation. The content of Glc-Lys in normal lens crystallin increased with age in a linear fashion. Thus, the nonenzymatic glycation of nondiabetic lens crystallin may be regarded as a biological clock. The diabetic lens crystallin samples (n = 11) had a higher content of Glc-Lys (0.070 +/- 0.034 nmol/nmol monomer). Over an age range comparable to that of the control samples, the diabetic crystallin samples contained about twice as much Glc-Lys. The Glc-Lys content of the diabetic lens crystallin samples did not increase with lens age.