William T. Morgan

Human serum histidine-rich glycoprotein. I. Interactions with heme, metal ions and organic ligands.

The 3.8 S alpha2-histidine-rich glycoprotein of human serum is composed of two non-identical subunits, each of which contains carbohydrate. The far ultraviolet circular dichroism spectrum of alpha2-histidine glycoprotein indicates that the protein has little alpha-helix but apparently appreciable amounts of beta-sheet and non-regular structures. alpha2-Histidine-rich glycoprotein binds heme with concomitant changes in the electrophoretic mobility of the protein, in the fluorescence of tryptophan residues, and in the absorption and optical activity of the heme chromophore. By fluorescence quenching, the stoichiometry of binding is 1 heme per alpha2-histidine-rich glycoprotein molecule with an apparent Kd near 1.5 muM; however, by changes in absorbance, the interaction of 9 to 10 additional heme molecules with the alpha protein can be detected. The absorption spectra of heme . alpha2-histidine-rich glycoprotein complexes resemble those of low-spin hemoproteins. The ellipticity induced in the heme chromophore on binding by alpha2-histidine-rich glycoprotein increases linearly up to about 10 hemes bound per mol protein. No change in the conformation of alpha2-histidine-rich glycoprotein was indicated by circular dichroism when one or two heme molecules are bound by the protein. alpha2-Histidine-rich glycoprotein does not effectively compete with human serum albumin for heme, suggesting that alpha2-histidine-rich glycoprotein has no major function in serum heme transport. Nonetheless, the binding of heme by alpha2-histidine-rich glycoprotein provides a means of studying the structure of this protein using the heme chromophore as a probe. alpha2-Histidine-rich glycoprotein also binds other organic molecules including bilirubin, diaquocobinamide, Cibacron blue F3GA and rose bengal, and certain divalent metals. It is of interest that copper, zinc, nickel, cadmium and cobalt effectively inhibit the binding of heme by alpha2-histidine-rich glycoprotein, whereas other divalent metals tested, including calcium, magnesium and manganese do not appreciably affect the heme-alpha2-histidine-rich glycoprotein interaction.

The interaction of human serum albumin and hemopexin with porphyrins

Abstract The interactions of the porphyrin-binding proteins of human serum, albumin and hemopexin, with coproporphyrin I and III, uroporphyrin I, protoporphyrin IX and zinc-protoporphyrin IX were examined quantitatively by means of equilibrium dialysis and fluorescence. Albumin bound the porphyrins with a high affinity, dissociation constant ( K d ) in the μM range, with the exception of uroporphyrin I which was only weakly bound, K d of the order of 100 μM. Protoporphyrin IX and zinc-protoporphyrin IX were more strongly bound than coproporphyrin I and III suggesting common binding sites. Human hemopexin formed complexes with coproporphyrin I and III, K d near 0.5 μM, with protoporphyrin IX, K d near 2 μM, and with zinc-protoporphyrin IX, K d near 0.5 μM. All porphyrins were bound with an apparent stoichiometry of 1 : 1, except zinc-protoporphyrin, 2 mol of which appear to be bound per mol of hemopexin. As with albumin, only a weak interaction with uroporphyrin I is evident, K d near 50 μM. Formation of the heme-hemopexin complex almost completely abolished subsequent interaction with coproporphyrin, consistent with heme and porphyrins sharing the same primary binding site on hemopexin. Several types of competition experiment were conducted to test directly the relative affinities of the two proteins for these porphyrins. Under these conditions, albumin clearly has a higher affinity than hemopexin for both protoporphyrin and zinc-protoporphyrin. Hemopexin is able to compete effectively with albumin for coproporphyrin I or III only at low molar ratios of the two proteins and seems unlikely to bind a significant portion of these porphyrins at the molar ratios (about 1 : 40) of hemopexin and albumin found in human serum. Further evidence for species differences in the heme binding site of hemopexin was obtained from fluorescence measurements on the human and rabbit proteins. Firstly, human hemopexin displayed a higher apparent stoichiometry of zinc-protoporphyrin binding (1.9 : 1) than rabbit hemopexin (1.3 : 1), with zinc-protoporphyrin-human hemopexin exhibiting a slightly higher relative fluorescence at 595 nm. Secondly, protoporphyrin bound to human hemopexin had red-shifted excitation (410 vs. 407 nm) and emission (634 vs. 627 nm) maxima compared with rabbit hemopexin. Thirdly, coproporphyrin I produced little quenching of the human protein compared with that of the rabbit protein.

Transfer of heme from heme-albumin to hemopexin

Exchange of heme in vitro between two heme-binding serum proteins, albumin and hemopexin, was examined spectrophotometrically. Hemopexin, albumin and heme in molar ratios of 1 : 70 : 1 were incubated at 22 degrees C, pH 7.3. The heme was added as free heme, heme-hemopexin or methemalbumin. Due to the high affinity of hemopexin for heme, Kd near 10(-13) M, only negligible amounts of heme were transferred from hemopexin to albumin in 48 h. However, more than 80% of heme was transferred from methemalbumin to hemopexin within 24 h. Heme added to a 1 : 70 mixture of the apo-proteins is initially bound by albumin; but more than 90% is bound by hemopexin in 24 h. Addition of dithionite causes nearly all of the heme present, whether added as free heme or methemalbumin, to associate with hemopexin in 15 min. Albumin thus appears to have a much lower affinity for ferro- than for ferri-heme. Results obtained from similar experiments with human serum and human serum made hemopexin-free by immunoadsorption fully corroborate those obtained with mixtures of purified albumin and hemopexin. These observations suggest that the rate-limiting step in the heme transport function of hemopexin is the formation of the heme-hemopexin complex, rather than the uptake of the complex by the liver.

Hepatic subcellular metabolism of heme from heme-hemopexin: incorporation of iron into ferritin.

Abstract The subcellular distribution and metabolic fate of [59Fe]heme-[125I]-labeled hemopexin after receptor-mediated interaction with the liver was examined in the rat. After intravenous injection, [59Fe]heme from the complex and 59Fe from hepatic catabolism of this heme accumulate in the liver and undergo changes in their subcellular distribution over 2 hours. The amounts of [59Fe]heme and particularly of 59Fe increase in the cytosol while remaining constant or decreasing in membranous fractions. In contrast, [125I]-labeled hemopexin associated with the liver during heme transport is always a small fraction of the dose and is not measurably catabolized under these conditions. Gel filtration of the cytosol showed that 59Fe increased linearly with time in a high molecular weight fraction which was identified immunologically as ferritin. We conclude that heme transported by hemopexin is metabolized by the liver and the iron conserved.

Transport of heme by hemopexin to the liver: Evidence for receptor-mediated uptake

Abstract We used carefully defined heme-hemopexin complexes to investigate the role of hemopexin in the catabolism of heme in vivo . Uptake of rabbit [ 59 Fe]heme-[ 125 I]hemopexin by rat liver was rapid. The liver-associated 125 I reached a maximum 5 minutes after injection, nearly 7-fold higher than apo-hemopexin, whereas liver-associated 59 Fe increased with time. This together with an inverse relationship of [ 125 I]hemopexin in the liver and serum during the course of heme transport suggests that hemopexin was released from the liver back to the circulation. Saturation of uptake with heme-hemopexin, reaching about 170 pmol [ 125 I]hemopexin (gm liver) −1 5 minutes after injection of 11 nmol, indicates a receptor-mediated process. We conclude that hemopexin delivers heme to the liver via interaction with a finite number of receptors and returns to the circulation.

Human histidine-rich glycoprotein. II. Serum levels in adults, pregnant women and neonates.

Summary Concentrations of 3.8S histidine-rich glycoprotein were determined in the sera of healthy adults, pregnant women, neonates, and of persons afflicted with a variety of diseases. Quantitative differences were found between the HRG concentration in healthy adult serum (12.5 ± 3.2 mg/100 ml, mean ± SD) and in neonatal (2.3 ± 1.5 mg/100 ml) and cord serum (3.4 ± 1.1 mg/100 ml). In women the HRG concentration declines steadily during the last two trimesters of pregnancy reaching at parturition a value about 50% of that in adult serum (5.6 ± 2.3 mg/100 ml, P < 0.005), but returns to normal levels within 5-15 days postpartum. No difference was observed between the HRG levels of males and females in either adults or neonates. In general, mean HRG levels in the sera of patients with a variety of disease states were near normal but were more widely scattered than in healthy adults. However, the mean HRG concentrations in the sera of patients with a variety of heart ailments (16.6 ± 6.2 mg/100 ml, P < 0.01) and erythropoietic protoporphyria (14.5 ± 5.9 mg/100 ml, P < 0.05) were elevated compared to healthy adult levels. The levels of HRG in the sera of patients with neuromuscular diseases (8.0 ± 1.8 mg/100 ml), lead poisoning (10.8 ± 4.2 mg/100 ml), and porphyria cutanea tarda (11.5 ± 4.1 mg/100 ml) were lower than normal levels (P < 0.05).

The aromatic and heme chromophores of rabbit hemopexin: Difference absorption and fluorescence spectra

Spectrophotometric and fluorimetric techniques were employed to charcterize the environment of the heme chromophore of rabbit hemopexin and to monitor changes in the environment of aromatic amino acid residues induced by the interaction of hemopexin with porphyrins and metalloporphyrins. Difference spectra showed maxima at 292 and 285 nm when hemopexin binds heme or deuteroheme but not deuteroporphyrin. These maxima are attributed to alterations in the local environment of tryptophan and tyrosine residues. Spectro-photometric titrations of the tyrosine residues of hemopexin, heme-hemopexin and hemopexin in 8 M urea showed apparent pK values at 11.4, 11.7, and 10.9 respectively. Perturbation difference spectra produced by 20% v/v ethylene glycol are consistent with the exposure of 6-8 of the 14 tyrosine residues and 6-8 of the 15 tryptophan residues of rabbit hemopexin to this perturbant. Only small differences were found between the perturbation spectra of apo- and heme-hemopexin near 290 nm, suggesting that slight or compensating changes in the exposure to solvent of tryptophan chromophores occur. In the Soret spectral region, the exposure of heme in the heme-hemopexin complex to ethylene glycol was 0.7, relative to the fully exposed heme peptide of cytochrome c. The fluorescence quantum yields of rabbit apo- and heme-hemopexin were estimated to be 0.06 and 0.03, respectively, compared to a yield of 0.13 for L-tryptophan. Iodide quenched 50% of the fluorescence of the deuteroheme-hemopexin complex. Cesium was not an effective quencher. Modification of approximately, 4 tryptophan residues with N-bromosuccinimide also decreased the relative fluorescence of apo-hemopexin by 50% and concomitantly reduced the heme-binding ability of the protein by 70%. The existence of sterically unhindered tryptophan residues in either apo- heme-hemopexin is unlikely since no charge transfer compelxes between these proteins and N-methylnicotinamide were detected.

Chemical modification of histidine residues of rabbit hemopexin

Abstract Treatment of rabbit hemopexin with bromoacetic acid (BrAc) or with diethylpyrocarbonate (DEP) modified histidine residues and produced a concomitant decrease in the proteins ability to form a low-spin hemichrome complex with deuteroheme (ferrideuteroporphyrin IX). Deuteroheme bound to hemopexin before treatment decreased the extent of inactivation by either reagent. After exposure of deuteroheme-hemopexin to 0.16 m BrAc at pH 6.9 for 120 h, 10–11 of the 16 histidine residues of hemopexin were carboxymethylated, but 90–95% of the deuteroheme-hemopexin complex remained intact. Under the same conditions, 12 histidine residues of apo-hemopexin were carboxymethylated, and 95% of the proteins ability to form its normal hemichrome complex with heme (ferriprotoporphyrin IX) was abolished. The alkylated apo-protein, however, did retain a potential to interact with deuteroheme. The apparent dissociation constants for the complexes of metal-free deuteroporphyrin and deuteroheme with BrAc-treated apo-hemopexin were both about 10 −6 m and nearly equal to that of the native deuteroporphyrin-hemopexin complex, as assessed by quenching of tryptophan fluorescence. Approximately 10 histidyl residues of the deuteroheme-hemopexin complex, but only about 4 residues of the apo-protein, were modified by DEP before heme-binding was appreciably affected. The effects of DEP on hemopexin were reversed by hydroxylamine at neutral pH, indicating that ethoxyformylation of histidine residues caused the observed inactivation of hemopexin. This and the results of BrAc treatment suggest that hemopexin contains several easily accessible histidine residues which are not critical for its interaction with heme. The conformation-sensitive positive ellipticity at 231 nm of hemopexin was affected by carboxymethylation and ethoxyformylation. Treatment with BrAc had only a small effect on the intrinsic ellipticity of apo-hemopexin, but eliminated the increase in ellipticity produced by interaction of unmodified hemopexin with heme. Treatment with DEP, on the other hand, decreased both intrinsic and extrinsic ellipticity. These results provide further evidence that the heme-hemopexin complex involves histidyl-heme iron coordination. In addition, they show that formation of the histidyl-heme complex not only greatly enhances the strength of the heme-hemopexin interaction but also is important for triggering conformational changes in the protein.

Circular dichroism of C5a anaphylatoxin of porcine complement

Abstract The far-ultraviolet circular dichroism spectrum of C5a anaphylatoxin of porcine complement implies that it has a substantial content of helical structure. The circular dichroism spectra of C5a in the 200–250 nm region at pH 7.2 and 3.7 are nearly identical and resemble those of C3a anaphylatoxin. Treatment of C5a with 2-mercaptoethanol progressively diminishes the ellipticity at 222 nm and its anaphylatoxic activity to limiting values. Removal of the reducing agent by dialysis completely restored both the ellipticity at 222 nm and the activity. This finding indicates that the integrity of the secondary conformation of the C5a molecule is largely dependent on disulfide bonds and is essential for its full activity.

Interaction of rabbit hemopexin with copro- and uroporphyrins

Rabbit hemopexin forms equimolar complexes in vitro with the I and III isomers of both coproporphyrin and uroporphyrin. The apparent dissociation constants (Kd) of these complexes are estimated to be 4-10(-7) M for coproporphyrin-hemopexin and 10(-6) M for uroporphyrin-hemopexin by equilibrium dialysis and quenching of protein fluorescence. Results of competitive binding experiments suggest that all four porphyrins bind at the heme-binding site of hemopexin, and that the relative affinity of rabbit hemopexin for these porphyrins is: deuteroheme greater than coproporphyrin I or III greater than uroporphyrin I or III. These findings provide further evidence that hemopexin may function as a transport protein for circulating coproporphyrins as well as for heme.