Evidence suggesting that the two forms of heme oxygenase are products of different genes

Recently, we have reported on the presence of two forms of heme oxygenase in rat liver and testis microsomes, referred to as HO-1 and HO-2 (M. D. Maines, G. M. Trakshel, and R. K. Kutty (1986) J. Biol. Chem. 261, 411-419; G. M. Trakshel, R. K. Kutty, and M. D. Maines (1986) J. Biol. Chem. 261, 11131-11137). Although the two forms differed in several biochemical properties, we could not ascertain whether they represented two isozymes or whether they were isoforms of heme oxygenase. In the present study, we provide evidence suggesting that the two forms are isozymes and represent different gene products. We also provide data suggesting that HO-1 is the commonly known heme oxygenase form. The molecular weight and immunochemical properties of HO-1 and HO-2 did not vary depending on the tissue source examined, i.e. liver and testis. Major differences, however, were noted in the amino acid composition of the two forms including the presence of 3 cysteine/cystine residues in HO-2 only. Using antibody to HO-2, four testis clones and two liver clones were isolated, and one liver and one testis clone were sequenced. Both clones revealed a 274-base-pair insert, and the sequence of both inserts was the same. The validity of assignment was confirmed by matching a 14-amino-acid peptide obtained from purified HO-2 with the sequence. Approximately 43% amino acid homology was detected between the HO-2 insert and the published amino acid sequence of heme oxygenase. However, amino acid homology search revealed the presence of two regions of homology: one 22-mer sequence with only one unmatched amino acid, and one 10-mer sequence with one unmatched amino acid. Heme oxygenase appeared to be the HO-1 form, an assignment based on its amino acid sequence matching the sequence of 2 peptides obtained from purified HO-1 and the immunochemical properties of the cobalt-, hematin-, and bromobenzene-induced rat liver enzyme. The secondary structure prediction analysis revealed an area of 100% structural homology with only 72% sequence homology. We predict this region may represent the catalytic site of the enzyme.     

Human heme oxygenase cDNA and induction of its mRNA by hemin

Hemin treatment increased both activity and mRNA level of heme oxygenase in human macrophages. Using poly(A)-rich RNA prepared from human macrophages treated with hemin, we have constructed a cDNA library in the Okayama-Berg vector. The human heme oxygenase cDNA was isolated by screening this library with a rat cDNA and was subjected to nucleotide sequence analysis. The deduced human heme oxygenase is composed of 288 amino acids with a molecular mass of 32,800 Da. The homology in amino acid sequences between rat and human heme oxygenase is 80%. Like rat heme oxygenase, human enzyme has a putative membrane segment at its carboxyl terminus, which is probably essential for the insertion of heme oxygenase into endoplasmic reticulum. Both rat and human heme oxygenase have no cysteine residues. Recently we have shown that rat heme oxygenase is a heat-shock protein [J. Biol. Chem. 262, 12889-12892 (1987)], and therefore we examined the effects of heat treatment on the induction of heme oxygenase in human macrophages and glioma cells. In contrast to hemin treatment, heat treatment had no apparent effects in either human cell line on the activity of heme oxygenase and its mRNA levels. These results suggest that human heme oxygenase may not be a heat-shock protein.     

Crystal structure of heme oxygenase from the gram-negative pathogen Neisseria meningitidis and a comparison with mammalian heme oxygenase-1

We report the crystal structure of heme oxygenase from the pathogenic bacterium Neisseria meningitidis at 1.5 A and compare and contrast it with known structures of heme oxygenase-1 from mammalian sources. Both the bacterial and mammalian enzymes share the same overall fold, with a histidine contributing a ligand to the proximal side of the heme iron and a kinked alpha-helix defining the distal pocket. The distal helix differs noticeably in both sequence and conformation, and the distal pocket of the Neisseria enzyme is substantially smaller than in the mammalian enzyme. Key glycine residues provide the flexibility for the helical kink, allow close contact of the helix backbone with the heme, and may interact directly with heme ligands.     

Degradation of heme in gram-negative bacteria: the product of the hemO gene of Neisseriae is a heme oxygenase

A full-length heme oxygenase gene from the gram-negative pathogen Neisseria meningitidis was cloned and expressed in Escherichia coli. Expression of the enzyme yielded soluble catalytically active protein and caused accumulation of biliverdin within the E. coli cells. The purified HemO forms a 1:1 complex with heme and has a heme protein spectrum similar to that previously reported for the purified heme oxygenase (HmuO) from the gram-positive pathogen Corynebacterium diphtheriae and for eukaryotic heme oxygenases. The overall sequence identity between HemO and these heme oxygenases is, however, low. In the presence of ascorbate or the human NADPH cytochrome P450 reductase system, the heme-HemO complex is converted to ferric-biliverdin IXalpha and carbon monoxide as the final products. Homologs of the hemO gene were identified and characterized in six commensal Neisseria isolates, Neisseria lactamica, Neisseria subflava, Neisseria flava, Neisseria polysacchareae, Neisseria kochii, and Neisseria cinerea. All HemO orthologs shared between 95 and 98% identity in amino acid sequences with functionally important residues being completely conserved. This is the first heme oxygenase identified in a gram-negative pathogen. The identification of HemO as a heme oxygenase provides further evidence that oxidative cleavage of the heme is the mechanism by which some bacteria acquire iron for further use.     

Importance of histidine residue 25 of rat heme oxygenase for its catalytic activity.

A truncated, soluble, and enzymatically active rat heme oxygenase lacking its membrane-associative, C-terminal segment was expressed in E. coli strain JM109. The roles of its four histidine residues were examined by determining the enzymatic activities of mutant enzymes in which each of these residues in turn was replaced by alanine. Mutation of histidine residue 25 to alanine resulted in marked decrease in activity for heme breakdown, indicating that this histidine residue has an important role in the heme oxygenase reaction.     

Reaction of the microsomal heme oxygenase with cobaltic protoporphyrin IX, and extremely poor substrate.

A reconstituted heme oxygenase system which was composed of a purified heme oxygenase from pig spleen microsomes and a partially purified NADPH-cytochrome c reductase from pig liver microsomes could not catalyze the conversion of cobaltic protoporphyrin IX (Co-heme) to biliverdin, although Co-heme could bind with the heme oxygenase protein to form a complex. The heme oxygenase system in the microsomes from pig spleen, rat spleen, and rat kidney also failed to oxidize Co-heme to biliverdin. Properties of the complex of Co-heme and heme oxygenase closely resembled those of cobalt myoglobin and cobalt hemoglobin; the Co-heme bound to the heme oxygenase protein did not react with cyanide and azide, the Co-heme moiety was reduced but only slowly with sodium dithionite, and the reduced form of the Co-heme did not appear to bind carbon monoxide. The co-heme bound to heme oxygenase was not reduced with the NADPH-cytochrome c reductase system in air. These findings further support the views that heme oxygenase may have a heme-binding crevice similar to those of myoglobin and hemoglobin and that reduction of heme is the prerequisite for the oxidative degradation of heme in the heme oxygenase reaction.     

Intracellular site of synthesis of microsomal heme oxygenase in pig spleen

In the pig spleen the specific activity of heme oxygenase was two to three times higher in smooth microsomes than in rough microsomes, whereas the total heme oxygenase activities recovered in the two microsomal fractions were similar. Free and bound polysomes were isolated from pig spleen and nascent peptides on these polysomes were analyzed by employing [3H]puromycin and a heme oxygenase-specific rabbit antibody (IgG). It was shown that free polysomes are the major site of heme oxygenase synthesis. In addition, cell-free synthesis of heme oxygenase was performed in a reticulocyte lysate system with free and bound polysomes isolated from pig spleen, and the results obtained again indicated that heme oxygenase is synthesized predominantly on free polysomes. The heme oxygenase newly synthesized on free polysomes may be incorporated first into the rough portion of endoplasmic reticulum either before or after its release from polysomes, although the specific activity of this enzyme at the steady state is considerably higher in the smooth region.     

Bacillus subtilis HmoB is a heme oxygenase with a novel structure

Iron availability is limited in the environment and most bacteria have developed a system to acquire iron from host hemoproteins. Heme oxygenase plays an important role by degrading heme group and releasing the essential nutrient iron. The structure of Bacillus subtilis HmoB was determined to 2.0 A resolution. B. subtilis HmoB contains a typical antibiotic biosynthesis monooxygenase (ABM) domain that spans from 71 to 146 residues and belongs to the IsdG family heme oxygenases. Comparison of HmoB and IsdG family proteins showed that the C-terminal region of HmoB has similar sequence and structure to IsdG family proteins and contains conserved critical residues for heme degradation. However, HmoB is distinct from other IsdG family proteins in that HmoB is about 60 amino acids longer in the N-terminus and does not form a dimer whereas previously studied IsdG family heme oxygenases form functional homodimers. Interestingly, the structure of monomeric HmoB resembles the dimeric structure of IsdG family proteins. Hence, B. subtilis HmoB is a heme oxygenase with a novel structural feature.     

Mechanism of action of heme oxygenase. A study of heme degradation to bile pigment by 18O labeling

The formation of bile pigment from heme by a reconstituted heme oxygenase system containing purified bovine spleen heme oxygenase, NADPH-cytochrome P-450 reductase, and biliverdin reductase was studied under an atmosphere containing 18,18O2. The product, bilirubin, was isolated and subjected to mass spectrometry, which revealed incorporation of 18O consistent with a two-molecule mechanism, whereby the product bile pigment contains oxygen atoms derived from two different oxygen molecules.     

Purification and characterization of heme oxygenase from chick liver. Comparison of the avian and mammalian enzymes

A major inducible form of heme oxygenase (EC 1.14.99.3) was purified from liver microsomes of chicks pretreated with cadmium chloride. The purification involved solubilization of microsomes with Emulgen 913 and sodium cholate, followed by DEAE-Sephacel, carboxymethyl-cellulose (CM-52) and hydroxyapatite chromatography, and FPLC through Superose 6 and 12 columns operating in series. The final product gave a single band on silver-stained SDS/polyacrylamide gels (Mr = 33,000). Optimal conditions for measurement of activity of solubilized heme oxygenase were studied. In a reconstituted system containing purified heme oxygenase, NADPH-cytochrome reductase, biliverdin reductase and NADPH, the Km for free heme was 3.8 +/- 0.5 microM; for heme in the presence of bovine serum albumin (5 mol heme/3 mol albumin) the Km was 5.0 +/- 0.8 microM; and the Km for NADPH was 6.1 +/- 0.4 microM (all values mean +/- SD, n = 3). Oxygen concentration as low as 15 microM, with saturating concentrations of heme and NADPH, did not affect the reaction rate, indicating that the supply of oxygen is not involved in the physiological regulation of activity of the enzyme. The pH optimum of the reaction was 7.4; at 37 degrees C, the apparent Vmax was 580 +/- 44 nmol biliverdin.(mg protein)-1.min-1 and the molecular activity was 19.2 min-1. Biliverdin IXa was the sole biliverdin isomer formed. In the presence of purified biliverdin reductase, biliverdin was converted quantitatively to bilirubin. Addition of catalase to the reconstituted system decreased the breakdown of heme to non-biliverdin products and led to nearly stoichiometric conversion of heme to biliverdin. Activity of the enzyme in the reconstituted system was inhibited by metalloporphyrins in the following order of decreasing potency: tin mesoporphyrin greater than tin protoporphyrin greater than zinc protoporphyrin greater than manganese protoporphyrin greater than cobalt protoporphyrin. Protoporphyrin (3.3 or 6.6 microM) (and several other porphyrins) and metallic ions (100 microM) alone had little if any inhibitory effect, except for Hg2+ which inhibited by 67% at 10 microM and totally at 15 microM. Following partial cleavage, fragments of the purified enzyme were sequenced. Comparison of sequences to those derived from cDNA sequences for the major inducible rat and human heme oxygenase showed 69% and 76% similarities, respectively. The histidine residue at position 132 of rat heme oxygenase-1 and the residues (Lys128-Arg136) flanking His132 were conserved in all three enzymes, as well as in the corresponding portion of a fourth less highly similar rat enzyme, heme oxygenase-2.(ABSTRACT TRUNCATED AT 400 WORDS)     

Regulation of heme metabolism in rat hepatocytes and hepatocyte cell lines: delta-aminolevulinic acid synthase and heme oxygenase are regulated by different heme-dependent mechanisms

Regulation of delta-aminolevulinic acid (ALA) synthase and heme oxygenase was analyzed in primary rat hepatocytes and in two immortalized cell lines, CWSV16 and CWSV17 cells. ALA synthase was induced by 4,6-dioxohepatnoic acid (4,6-DHA), a specific inhibitor of ALA dehydratase, in all three systems; however, the induction in CWSV17 cells was greater than in either of the other two systems. Therefore, CWSV17 cells were used to explore the regulation of both enzymes by heme and 4,6-DHA. Data obtained from detailed concentration curves demonstrated that 4,6-DHA induced the activity of ALA synthase once ALA dehydratase activity became rate-limiting for heme biosynthesis. Heme induced heme oxygenase activity with increases occurring at concentrations of 10 microM or greater. Heme blocked the 4,6-DHA-dependent induction of ALA synthase with an EC50 of 1.25 microM. Heme-dependent decreases of ALA synthase mRNA levels occurred more quickly and at lower concentrations than heme-dependent increases of heme oxygenase mRNA levels. ALA synthase mRNA remained at reduced levels for extended periods of time, while the increases in heme oxygenase mRNA were much more transient. The drastic differences in concentrations and times at which heme-dependent effects were observed strongly suggest that two-different heme-dependent mechanisms control the ALA synthase and heme oxygenase mRNAs. In CWSV17 cells, heme decreased the stability of ALA synthase mRNA from 2.5 to 1.3 h, while 4,6-DHA increased the stability of the mRNA to 5.2 h. These studies demonstrate that regulation of ALA synthase mRNA levels by heme in a mammalian system is mediated by a change in ALA synthase mRNA stability. The results reported here demonstrate the function of the regulatory heme pool on both ALA synthase and heme oxygenase in a mammalian hepatocyte system.     

The oxidation of hemins by microsomal heme oxygenase. Structural requirements for the retention of substrate activity

The substrate specificity of microsomal heme oxygenase from rat liver was studied by introducing systematic structural changes in the array of substituents of the protohemin IX rings. Replacement of the vinyls by methyl groups resulted in hemins which were excellent substrates of the heme oxygenase. Replacement of the 4-vinyl group by a propionic acid chain (harderohemin), decreased substrate activity to 40%. The replacement of the vinyls by formyl residues strongly decreased substrate activity but the hemins were still substrates of heme oxygenase. The oxidation rates of Spirographis hemin and of 2,4-diformyldeuterohemin IX showed a time lag which was absent when isoSpirographis hemin was used as a substrate. This lag could be attributed to the formation of a transient hemiacetal between the 2-formyl group and the alpha-mesohydroxy residue. The isomeric protohemins I, XI, and XIV (Fischer's notation) were examined as possible substrates of microsomal heme oxygenase. In these protohemins the array of substituents of rings A and B was the same as in protohemin IX, but the methyl and propionic acid residues of rings C and D were at different positions from those of protohemin IX. None of them had substrate activity, indicating that the presence of two vicinal propionic acid side-chains at C6 and C7 was necessary for substrate activity. A hemin with only one propionic acid residue at C5 was not a substrate of the enzyme, either. When the propionic acid residues of protohemin IX were replaced by butyric acid residues, substrate activity decreased to 50% (as compared to protohemin IX), while when they were replaced by acetic acid residues, the substrate activity was entirely suppressed. The addition of dimethyl sulfoxide (25 mM) to the incubation mixture enhanced the oxidation of hemins with non-polar substituents in rings A and B by about 35%, while it was without effect on hemins with polar substituents in the same rings.     

Porphyrogenic effects and induction of heme oxygenase in vivo by δ-aminolevulinic acid

The effects of single large doses of the porphyrin-heme precursor ?d-aminolevulinic acid on tissue porphyrins and on δ-aminolevulinate synthase and heme oxygenase, the rate-living enzymes of liver heme synthesis and degradation respectively, were studied in the chick embryo in ovo, in the mouse and in the rat. δ-Aminolevulinic acid treatment produced a distinctive pattern characterized by extensive tissue porphyrin accumulation and alterations in these rate-limiting enzymes in the liver. Repression of basal or allylisopropylacetamide-induced liver δ-aminolevulinate synthase was observed and, in the mouse and the rat, induction of liver heme oxygenase after δ-aminolevulinic acid treatment, in a manner similar to the known effects of hemin on these enzymes. In the chick embryo liver in ovo heme oxygenase was substantially higher than in rat and mouse liver, and was not significantly induced by δ-aminolevulinic acid or other compounds, including hemin, CS₂ and CoCl₂. Levulinic acid, an analogue of δ-aminolevulinic acid, did not induce heme oxygenase in mouse liver. δ-Aminolevunilic acid treatment did not impair ferrochelatase activity but was associated with slight and variable decreases in liver cytochrome P-450. Treatment of chick embryos with a small ‘priming’ dose of 1,4-dihydro-3,5-dicarbethoxycollidine, which impairs liver ferrochelatase activity, accentuated porphyrin accumulation after δ-aminolevulinic acid in the liver. These observations indicate that exogenous δ-aminolevulinic acid is metabolized to porphyrins in a number of tissues and, at least in the liver, to a physiologically significant amount of heme, thereby producing an increase in the size of one or more of the heme pools that regulate both heme systhesis and degradation. It is also possible than when δ-aminolevulinic acid is markedly overproduced in vivo it may be transported to many tissues and re-enter the heme pathway and alter porphyrin-heme metabolism in cells and tissues other than those in which its overproduction primarily occurs.     

Heme oxygenase, steering dioxygen activation toward heme hydroxylation

The activation of dioxygen by heme oxygenase proceeds via formation of an obligatory ferric hydroperoxide intermediate (FeIII-OOH), as is the case in the activation of dioxygen by monooxygenase enzymes. This review summarizes current understanding of the structural and dynamic properties in heme oxygenase that channel the reactivity of the FeIII-OOH intermediate toward heme hydroxylation rather than oxoferryl formation. In addition, structural and electronic factors dictating the regiospecificity of heme oxygenation are analyzed in the context of recent X-ray and NMR spectroscopic studies. Differences in mechanism between heme hydroxylation, as carried out by heme oxygenase, and the coupled oxidation process, are also addressed.     

The complete amino acid sequence of bovine skeletal muscle acylphosphatase

The primary structure of bovine skeletal muscle acylphosphatase was determined by performing the sequence analyses of the complete series of tryptic peptides. The amino acid composition of the entire series of peptic peptides was used to reconstruct the sequence by the overlapping method. The proposed structure is further confirmed by analogy with known amino acid sequences of acylphosphatase from skeletal muscle of other vertebrate species. The length of the polypeptide chain is 98 residues, identical to the length of the enzymes from other known mammalian species, but different from that found in turkey. The enzyme is NH2-acetylated and a comparison with the analogous molecular forms from other vertebrate species indicates that there are several long polypeptide stretches strictly conserved (93-97% identical position among mammals, and about 80% between calf and turkey enzymes).     

Cyclic oscillations in rat hepatic heme oxygenase and delta-aminolevulinic acid synthetase following intravenous heme administration.

Heme administration in vivo results in the suppression of synthesis of rat hepatic δ-aminolevulinic acid (ALA) synthetase and induction of rat hepatic heme oxygenase. Intravenous heme administration in vivo results in the appearance of cyclic progressively damped oscillations of both hepatic ALA synthetase activity and hepatic heme oxygenase activity. Heme oxygenase induction precedes in time the induction of ALA synthetase. ALA synthetase oscillations are observed in hepatic cell cytosol and mitochondrial fractions as well as in the total homogenate. Cycloheximide pretreatment abolishes both the ALA synthetase and heme oxygenase oscillations, while actinomycin D pretreatment has only a minimal effect on the induction of heme oxygenase. These results suggest that hepatic heme metabolism is closely regulated by rapid changes in the capacity to synthesize and catabolize heme, and the cyclic oscillations following intravenous heme may be a manifestation of the feedback regulation processes involved. This regulatory capacity is dependent on protein synthesis, and the primary site of regulation may be at the translational level on the endoplasmic reticulum.