Oxidation of alpha -meso-Formylmesoheme by Heme Oxygenase
ELECTRONIC CONTROL OF THE REACTION REGIOSPECIFICITY*

(Received for publication, June 13, 1997, and in revised form, July 2, 1997)

Justin Torpey and Paul R. Ortiz de Montellano Dagger

From the Department of Pharmaceutical Chemistry, School of Pharmacy, and Liver Center, University of California, San Francisco, California 94143-0446

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES


ABSTRACT

The oxidation of heme to biliverdin IXalpha by heme oxygenase involves regiospecific alpha -meso-hydroxylation followed by extrusion of the alpha -meso-carbon as CO. In an earlier study, enzymatic oxidation of the four meso-methylmesoheme isomers suggested that the reaction regiospecificity is sensitive to the electronic properties of the meso-methyl group (Torpey, J. W., and Ortiz de Montellano, P. R. (1996) J. Biol. Chem. 271, 26067-26073), although we could not exclude the possibility that the altered reaction regiochemistry was due to perturbation of the porphyrin structure by the meso-substituent. To examine this possibility, we have synthesized the four meso-formylmesoporphyrin isomers and have examined their oxidation by heme oxygenase. The meso-formyl and meso-methyl substituents differ in that the former is electron withdrawing and the latter is electron donating. In contrast to alpha -meso-methylmesoheme, which is exclusively oxidized at the methyl-substituted position, alpha -meso-formylmesoheme is exclusively oxidized at a non-formyl-substituted meso-carbon. The finding that the methyl and formyl groups channel the reaction regiospecificity in opposite directions establishes that the regiochemistry of the heme oxygenase reaction is primarily under electronic rather than steric control. It also confirms that the oxidation involves electrophilic addition of the oxygen to the porphyrin ring.


INTRODUCTION

The oxidation of heme1 by heme oxygenase yields biliverdin IXalpha and CO (see Fig. 1), both of which have important physiological activities. Biliverdin is reduced to bilirubin, a powerful antioxidant that is highly lipophilic and, under conditions of impaired excretion, can reach concentrations at which it becomes neurotoxic (1). Carbon monoxide, the second product of the heme oxygenase reaction, is a putative physiological messenger akin to nitric oxide (2-4).


Fig. 1. Reaction intermediates in the heme oxygenase-catalyzed oxidation of heme to biliverdin. The substituents on the porphyrin are vinyl (V) and propionate (Pr). In the case of mesoheme, the vinyl substituents are replaced by ethyl groups. The alpha -, beta -, gamma -, and delta -meso positions are labeled.
[View Larger Version of this Image (26K GIF file)]

Human HO-1, a 32-kDa protein, is anchored to the endoplasmic reticulum via a C-terminal lipophilic domain (5). Removal of the 23 C-terminal amino acids by mutation of the cDNA coding for the protein, followed by heterologous expression in Escherichia coli, yields a truncated human HO-1 (hHO-1) that is both soluble and fully active (6). An active and soluble rat heme oxygenase without the 26 C-terminal amino acids and with mutations in the last two amino acids (S262R,S263L) has been independently reported (7).

The first step of the reaction catalyzed by heme oxygenase is the NADPH-, cytochrome P-450-reductase-, and O2-dependent oxidation of heme to alpha -meso-hydroxyheme (see Fig. 1) (8-12). In the second step, alpha -meso-hydroxyheme undergoes an O2-dependent but NADPH-independent (8) reaction that results in extrusion of the hydroxylated alpha -meso-carbon as CO with the concomitant formation of verdoheme (Fig. 1). Finally, in the third step, heme oxygenase catalyzes the formation of biliverdin from verdoheme in a reaction that also requires NADPH, cytochrome P-450 reductase, and O2 (10, 11). The heme oxygenase reaction is highly regiospecific and exclusively produces biliverdin IXalpha , the isomer resulting from oxidation of the alpha -meso-carbon (13).

The central role of alpha -meso-hydroxylation in the oxidation of heme by heme oxygenase led us earlier to synthesize the four possible meso-methylmesoheme isomers and to examine their enzyme-catalyzed oxidation (14, 15). Surprisingly, alpha -meso-methylmesoheme is still oxidized to mesobiliverdin IXalpha . The reaction does not result in the formation of CO, however, in accord with the fact that normal alpha -meso-hydroxylation is impossible due to the presence of the methyl substituent. Equally surprising was the finding that gamma -meso-methylmesoheme, despite the presence of an unsubstituted alpha -meso position, is oxidized exclusively at the gamma -meso position to give gamma -mesobiliverdin (15). delta -meso-Methylmesoheme was similarly oxidized at the delta -meso position to give delta -mesobiliverdin, but oxidation at the unsubstituted meso positions to give a mixture of methyl-substituted biliverdin isomers was also observed. The finding that the electron donating methyl group enhances reaction at the substituted carbon suggests that the oxidation reaction involves electrophilic addition of the activated oxygen to the meso carbon. An electrophilic oxidation mechanism is also suggested by the earlier finding that the oxidation of heme by heme oxygenase in the presence of ethylhydroperoxide rather than NADPH and P-450 reductase produces alpha -meso-ethoxyheme (16). Furthermore, the results argue that the regiochemistry of the heme oxygenase reaction is controlled not by steric orientation of the oxidizing species as previously suggested (17) but rather by changes in the frontier orbital electron density at the meso positions of the heme group. However, the possibility that the effect of the meso-methyl group was due to deformation of the heme structure rather than to an electronic effect could not be excluded.

To further test the hypothesis that the regiochemistry of the heme oxygenase reaction is governed by electronic rather than steric effects, we have synthesized the four possible isomers of meso-formylmesoheme and examined their oxidation by heme oxygenase. The formyl group is an electron-withdrawing substituent and should exert an effect opposite to that of a methyl group if the reaction regiochemistry is electronically controlled. On the other hand, the effects of the methyl and formyl substituents on the reaction regiochemistry should be parallel if they stem from a deformation of the heme structure or a steric interaction of the meso-substituent with the protein. The finding that the formyl and methyl substituents channel the reaction regiochemistry in opposite directions provides persuasive evidence that the regiochemistry of the heme oxygenase reaction is electronically rather than sterically controlled.


EXPERIMENTAL PROCEDURES

General Methods

Human heme oxygenase truncated of its 23 C-terminal amino acids (hHO-1) was expressed in E. coli and purified as reported (6). Cytochrome P-450 reductase was kindly provided by Prof. B. S. S. Masters (University of Texas Health Sciences Center, San Antonio, TX). Mesoheme was prepared from mesoporphyrin IX dimethyl ester (Porphyrin Products, Logan, UT), and the individual meso-formylmesohemes were prepared from the corresponding synthetic meso-formyl mesoporphyrin IX dimethyl esters as described previously (14). Thin layer chromatography was done on silica gel GF (250 micron) plates (Analtech, Newark, DE). HPLC was performed on a Varian 9010 solvent delivery system equipped with a Hewlett Packard 1040A detector and a reverse-phase Whatman analytical (4.6 × 250 mm) Partisil 10-mm ODS-3 column. The mobile phase was either 100% Solvent A (acetone/0.1% aqueous formic acid, 50:50) monitored at 374 nm and referenced at 474 nm or 100% Solvent B (acetonitrile/water/AcOH, 55:40:10) monitored at 404 nm and referenced at 550 nm. Absorption spectra were recorded on a Hewlett Packard 8452A diode array spectrophotometer. Mass spectra were obtained by (+)LSIMS on a Kratos Concept instrument using a 1:1 (1% trifluoroacetic acid) glycerol-thioglycerol matrix. 1H NMR spectra were measured in deuterated chloroform (porphyrin concentration 3-4 mg/ml) on a General Electric QE 300-MHz instrument. 13C NMR spectra were acquired on the same 300-MHz instrument and are completely decoupled. The meso-formylmesoheme isomers were individually reconstituted into hHO-1 in a 2:1 heme to enzyme ratio, and the resulting mixtures were purified over a Bio-Rad HTP column to give the 1:1 complexes in 100 mM potassium phosphate buffer (pH 7.4) (6).

Regioisomers of Meso-Formylmesoporphyrin IX Dimethyl Ester

The regioisomeric meso-hydroxymethyl mesoporphyrin IX dimethyl esters were synthesized as previously reported (14). The meso-hydroxymethyl substituent was oxidized to the meso-formyl by dissolving the meso-(hydroxymethyl)-mesoporphyrin IX isomers (alpha  = 16 mg, 25.6 µmol; beta  = 14 mg, 22.4 µmol; gamma  = 5 mg, 8.0 µmol; delta  = 42 mg, 67.3 µmol) in CH2Cl2 (20 ml) and pyridine (5 ml) containing three equivalents of pyridinium chlorochromate. After standing overnight at room temperature in the dark, the reaction was partitioned between 1.0 N HCl (300 ml) and CH2Cl2 (50 ml). The organic phase was then washed with water (400 ml) and concentrated under vacuum. Thin layer chromatography of the reaction mixture with diethyl ether yielded the meso-formylmesoporphyrin dimethyl ester product (Rf = 0.88). The meso-formyl porphyrins were purified on silica gel columns with hexanes-diethyl ether (30:70) as the eluting solvent. The four isomers were oxidized in the same manner, yielding the alpha  (6.0 mg, 9.6 µmol, 38%), beta  (9.3 mg, 15.0 µmol, 67%), gamma  (4.0 mg, 6.4 µmol, 80%), and delta  (11 mg, 17.7 µmol, 26%) meso-formylmesoporphyrins. The spectroscopic and analytical data for the isomers are as follows: alpha  isomer, lambda max (CH3Cl) 404, 504, 538, 576 nm; 1H NMR (CDCl3) delta  1.69 (t, 2H, J = 7.3 Hz), 1.76 (t, 2H, J = 7.5 Hz), 3.23 (t, 4H, J = 6.5 Hz), 3.34 (s, 3H), 3.52 (s, 3H), 3.56 (s, 3H), 3.58 (s, 3H), 3.63 (s, 3H), 3.64 (s, 3H), 3.79 (m, 2H) 3.98 (m, 2H), 4.31 (m, 4H), 9.94 (s, 1H), 10.01 (s, 2H), 12.72 ppm (s, 1H); 13C NMR (CDCl3) delta  11.6, 15.2, 16.0, 16.7, 17.5, 18.5, 19.8, 21.7, 29.7, 36.8, 51.8, 98.4, 113.6, 115.2, 137.9, 138.6, 142.2, 143.1, 143.4, 144.0, 144.1, 145.1, 145.5, 150.2, 173.5, 197.6 ppm; HRMS m/z 622.3163, calculated for C37H42N4O6 622.3155; beta  isomer, lambda max (CH3Cl) 404, 506, 538, 574 nm; 1H NMR (CDCl3) delta  1.79 (brd m, 2H),1.81 (t, 2H, J = 7.5 Hz), 3.21 (m, 4H), 3.37 (s, 3H), 3.51 (s, 3H), 3.54 (s, 3H), 3.55 (s, 3H), 3.62 (s, 3H), 3.66 (s, 3H), 3.82 (m, 2H), 4.00 (m, 2H), 4.33 (m, 4H), 9.93 (s, 1H), 10.01 (s, 1H), 10.07 (s, 1H), 12.73 ppm (s, 1H); 13C NMR (CDCl3) delta  11.6, 14.1, 17.6, 19.6, 21.7, 22.0, 22.7, 27.4, 31.9, 36.8, 51.7, 97.2, 948.0, 98.5, 113.6, 114.4, 128.7, 135.1, 135.7, 136.2, 138.5, 140.7, 142.0, 142.9, 143.4, 144.7, 146.0, 146.2, 173.5, 174.0, 197.8 ppm; HRMS m/z 622.3149, calculated for C37H42N4O6 622.3155; gamma  isomer, lambda max (CH3Cl) 406, 504, 538, 574 nm; 1H NMR (CDCl3) delta  1.81 (t, 6H, J = 9.0 Hz), 2.33 (t, 2H, J = 7.5 Hz), 3.11 (brd t, 2H), 3.53 (, 3H), 3.54 (s, 3H), 3.55 (s, 3H), 3.56 (s, 3H), 3.63 (m, 2H), 3.72 (s, 3H), 3.73 (s, 3H), 3.99 (t, 2H, J = 9.0 Hz), 4.12 (t, 2H, J = 9.0 Hz), 4.29 (m, 2H), 9.92 (s, 1H), 10.02 (s, 2H), 12.73 ppm (s, 1H); 13C NMR (CDCl3) delta  11.4, 11.9, 17.5, 19.7, 24.4, 29.7, 35.5, 51.8, 98.5, 98.6, 98.8, 136.9, 139.7, 142.9, 143.1, 143.2, 143.8, 145.2, 173.4, 173.5, 196.0 ppm; HRMS m/z 622.3142, calculated for C37H42N4O6 622.3155; delta  isomer, lambda max 434, 574, 628 (CH3Cl) nm; 1H NMR (CDCl3) delta  1.77 (t, 2H, J = 7.5 Hz), 1.81 (t, 2H, J = 7.5 Hz), 3.18 (t, 2H, J = 7.5 Hz), 3.24 (t, 2H, J = 7.5 Hz), 3.37 (s, 3H), 3.38 (s, 3H), 3.55 (s, 3H), 3.57 (s, 3H), 3.63 (s, 3H), 3.65 (s, 3H), 3.98 (q, 4H, J = 7.5 Hz), 4.34 (m, 4H), 9.94 (s, 1H), 10.01 (s, 1H), 10.07 (s, 1H), 12.73 ppm (s, 1H); 13C NMR (CDCl3) delta  11.4, 11.5, 16.5, 16.7, 17.5, 19.7, 21.7, 29.7, 36.8, 51.7, 98.1, 98.3, 98.6, 129.2, 133.9, 135.7, 136.6, 138.8, 141.3, 142.4, 142.8, 143.2, 144.1, 144.7, 144.9, 145.3, 173.4, 174.0, 198.1 ppm; HRMS m/z 622.3151, calculated for C37H42N4O6 622.3155.

Fe(II)Mesoverdoheme-CO Complex Formation Using NADPH-P-450 Reductase

P-450 reductase (12 µg, 0.15 nmol) and NADPH (100 nmol) were added to a 1-ml cuvette containing, in a final 1-ml volume, a solution of one of the meso-formylmesoheme-hHO-1 complexes (alpha -CHO, 0.29 mg, 9.8 nmol; beta -CHO, 0.37 mg, 12.2 nmol; gamma -CHO, 0.23 mg, 7.8 nmol; and delta -CHO, 0.41 mg, 13.8 nmol) in 100 mM potassium phosphate buffer (pH 7.4). The solution was presaturated with CO by bubbling with the gas before addition of the P-450 reductase and NADPH. The progress of the reaction was followed spectrophotometrically by monitoring both the appearance of the Fe(II)mesoverdoheme-CO complex at lambda max 616 nm and the loss of the Soret band of the starting complex.

Fe(III)Mesoverdoheme Formation Using H2O2

H2O2 (one equivalent) was added to a cuvette containing, in a final 1-ml volume, one of the meso-formylmesoheme-hHO-1 complexes (pH 7.4) (alpha -CHO, 0.27 mg, 9.1 nmol; beta -CHO, 0.46 mg, 15.2 nmol; gamma -CHO, 0.23 mg, 7.8 nmol; delta -CHO, 0.33 mg, 11.11 nmol) in 100 mM potassium phosphate buffer. The progress of the reaction was monitored spectrophotometrically by the appearance of the Fe(III)mesoverdoheme absorption at 600-700 nm and the decrease in the intensity of the Soret absorbance of the starting complex.

Regiochemical Analysis

P-450 reductase (900 µg, 13.4 nmol) and NADPH (83.3 mg, 120 µmol) were added to a solution of the alpha -meso-formylmesoheme-hHO-1 complex (81.0 mg, 2.7 µmol) in 100 mM potassium phosphate buffer (pH 7.4). The final incubation volume was 40 ml. The reaction was allowed to stand at 25 °C for 60 min and was then extracted and analyzed as described below. The products formed in the reactions of the other three meso-formyl isomers were similarly obtained in smaller scale to determine their absorption spectra and HPLC properties.

To the alpha -meso-formylmesoheme-hHO-1 reaction mixture was added concentrated HCl (1 drop) and acetic acid (1 ml) before the solution was extracted with CH2Cl2 (50 ml), and the organic phase was washed with brine (50 ml). The concentrate was dissolved in 100% Solvent A and analyzed by HPLC at a flow rate of 1.0 ml·min-1. The UV-visible absorption spectrum was acquired, and the mesobiliverdin was dissolved in 50 ml of 5% H2SO4 in MeOH (v/v) and left at room temperature for 8 h. The mesobiliverdin dimethyl ester was extracted into the organic phase after adding CHCl3 (50 ml) and water (50 ml), and the organic phase was washed with water (100 ml) and concentrated under vacuum. The electron impact mass spectrum of the alpha -meso-formylmesobiliverdin dimethyl ester was then determined.

CO Assay

To a 1-ml solution of the alpha -meso-formylmesoheme-hHO-1 complex (1.6 mg, 54.0 nmol) was added P-450 reductase (0.24 mg, 3.1 nmol) and NADPH (0.42 mg, 500 nmol). The tube was immediately sealed with a rubber septum. After 15 min at 25 °C, the solution had turned green. By injection with a syringe through a septum, 525 µl of ferrous deoxymyoglobin (0.45 mg, 27 nmol) was added to the alpha -meso-formylmesoheme-hHO-1 reaction mixture. The solution was shaken, the septum was removed, and the UV-visible spectrum was recorded. The ferrous deoxymyoglobin used to detect CO was prepared by adding sodium dithionite (5 mg) to 2 ml of a solution of horse skeletal muscle myoglobin (1.7 mg, 100 nmol) in 100 mM potassium phosphate buffer (pH 7.4).

CO assays were similarly performed by sealing the tubes after adding cytochrome P-450 reductase (6.0 µg, 78 pmol) and NADPH (167 µg, 200 nmol) to 1-ml solutions (final volume) of the other meso-formylmesoheme-hHO-1 complexes (beta -CHO, 0.27 mg, 9.1 nmol; gamma -CHO, 0.21 mg, 7.0 nmol; delta -CHO, 0.26 mg, 8.6 nmol). After 5 min at 25 °C, the reactions had turned green, and freshly prepared ferrous deoxymyoglobin (88 µg, 5 nmol) was injected through a septum. The solutions were mixed, the septa were removed, and the UV-visible spectra were recorded.

In separate experiments, quantitative CO assays were repeated with 0.5 ml of a solution of the alpha -meso-formylmesoheme-hHO-1 complex (alpha -CHO = 1.1 mg, 35 nmol) to which were added cytochrome P-450 reductase (0.24 mg, 3.1 nmol) and NADPH (0.42 mg, 500 nmol). The sealed reaction tubes were allowed to stand for 15 min at 25 °C. Freshly prepared ferrous deoxymyoglobin (0.73 mg, 32 nmol) was then injected via syringe, the solutions were mixed, the septa were removed, and the UV-visible spectra were recorded.


RESULTS

Synthesis of the Fe(III)meso-Formylmesoheme Regioisomers

The four meso-formylmesoporphyrin IX dimethyl ester regioisomers were synthesized by pyridinium chlorochromate oxidation of the corresponding regioisomeric meso-hydroxymethylmesoporphyrin IX dimethyl esters (Fig. 2). The meso-hydroxymethylmesoporphyrin IX dimethyl esters were obtained by Vilsmeier formylation of the dimethyl ester of copper mesoporphyrin followed by reduction of the formyl to the hydroxymethyl group, as previously reported (14). Reduction of the formyl to the hydroxymethyl group was required to separate the four regioisomeric products because the alcohols are readily separated by chromatography but the formyl precursors are not. The position of the substitution in each of the meso-hydroxymethyl isomers was established by NMR in the earlier studies (14). Reoxidation of the separated hydroxymethyl isomers provides the individual, regiochemically defined, meso-formyl isomers. The synthesis of the requisite meso-formylmesoheme regioisomers is completed by ester hydrolysis and insertion of the iron atom (14).


Fig. 2. Synthesis and structures of the four meso-formylmesoheme regioisomers. PCC is pyridinium chlorochromate.
[View Larger Version of this Image (26K GIF file)]

Formation of the Fe(II)Mesoverdoheme-CO Complex Using NADPH-P-450 Reductase

Incubation of the alpha -, beta -, gamma -, and delta -meso-formylmesoheme-hHO-1 complexes with NADPH and P-450 reductase under an atmosphere of O2 and CO causes a decrease in the intensity of the Soret band and the accumulation of an intermediate with an absorption maximum at lambda max = 616 nm (Fig. 3). The absorption spectrum of the intermediate resembles that of the Fe(II)mesoverdoheme-CO complex formed under similar conditions in incubations of Fe(III)mesoheme with hHO-1 (18, 19). The CO arrests the reaction at the spectroscopically convenient mesoverdoheme stage by chelating to the iron and thus preventing conversion of the mesoverdoheme to the final mesobiliverdin product. The Soret band changes, involving a broadening and decrease in the absorbance maximum with a small shift to lower wavelengths (from 400 to 396 nm), are similar in the reactions of the alpha - (Fig. 3A), beta - (Fig. 3B), and delta -meso-formyl (Fig. 3D) isomers. However, for the gamma -formyl isomer, the decrease in the Soret band intensity is associated with a marked red shift from 396 to 424 nm (Fig. 3C). The reason for the different spectroscopic change observed with the gamma -isomer is unclear but may be due to an interaction of the gamma -formyl group with the flanking propionic acid side chains.


Fig. 3. Spectrophotometric monitoring of the reactions of the meso-formylmesoheme-hHO-1 complexes with NADPH-P-450 reductase under an atmosphere of CO. A, alpha -meso-formyl isomer. B, beta -meso-formyl isomer. C, gamma -meso-formyl isomer. D, delta -meso-formyl isomer. The reaction details are given under "Experimental Procedures." The directions of the spectroscopic changes are indicated by the arrows.
[View Larger Version of this Image (17K GIF file)]

Formation of Fe(III)Mesoverdoheme with H2O2

Earlier studies have shown that H2O2 can substitute for O2 and NADPH-P-450 reductase in the hHO-1-catalyzed oxidation of heme to verdoheme but not in the subsequent conversion of verdoheme to biliverdin (6). The H2O2-dependent reaction can therefore be monitored without adding CO because the reaction is automatically arrested at the Fe(III)mesoverdoheme stage. As shown in Fig. 4, one equivalent of H2O2 similarly supports conversion of the meso-formylmesohemes to mesoverdohemes. For all four meso-formyl regioisomers, the reaction with H2O2 brings about a decay in the intensity of the Soret maximum at 400 nm without changing its position. This change is associated with the appearance of a broad absorbance with a maximum at 642 nm. These changes are similar to those reported for the H2O2-dependent oxidation of heme by hHO-1 (6).


Fig. 4. Spectrophotometric monitoring of the reactions of the meso-formylmesoheme-hHO-1 complexes with H2O2. A, alpha -meso-formyl isomer. B, beta -meso-formyl isomer. C, gamma -meso-formyl isomer. D, delta -meso-formyl isomer. The incubations contained one equivalent of H2O2, and the spectra were recorded at intervals of 10 s, the first spectrum being recorded before the addition of H2O2. The directions of the spectroscopic changes are indicated by the arrows.
[View Larger Version of this Image (17K GIF file)]

Regiochemistry of the hHO-1-catalyzed Oxidation of alpha -meso-Formylmesoheme

The regiospecificity of the oxidation of alpha -meso-formylmesoheme by hHO-1 was examined with the NADPH-P-450 reductase system. For these studies, the incubations were carried out in the absence of exogenous CO, and the final biliverdin product was isolated and characterized by HPLC, absorption spectroscopy, and mass spectrometry. Oxidation of the alpha -meso-formylmesoheme by NADPH-P-450 reductase followed by HPLC of the product yields a single broad mesobiliverdin-like peak (Fig. 5). A single product peak with a similar retention time is also obtained when each of the other meso-formylmesoheme isomers is oxidized by heme oxygenase. However, the HPLC system readily separates the four biliverdin regioisomers derived from heme and the methyl-substituted regioisomers from the unsubstituted mesobiliverdins, but it only partially separates the methyl-substituted mesobiliverdin regioisomers from each other (15). Because its ability to resolve the meso-formylmesoheme regioisomers has not been independently established, it cannot be inferred from the observation of a single peak in the chromatograms that a single isomer is formed. The HPLC retention times do show, however, that the products are not unsubstituted biliverdins and therefore that the biliverdin products retain the formyl group.


Fig. 5. HPLC chromatogram of the meso-formylmesobiliverdin extracted from the reaction of the alpha -meso-formylmesoheme-hHO-1 complex with NADPH and P-450 reductase. The absorbance was monitored at 364 nm. The peaks below 10 min are due to pyridine nucleotides.
[View Larger Version of this Image (11K GIF file)]

The absorption and mass spectra of the product obtained from alpha -meso-formylmesoheme confirm that it is a mesobiliverdin that retains the alpha -meso-formyl group. The maxima in the absorption spectrum (lambda max 364, 640 nm) are slightly shifted with respect to those of authentic mesobiliverdin IXalpha (lambda max 366, 636) (15), the product expected if the alpha -meso-carbon and the attached formyl group were eliminated in the reaction. The mass spectrum does not give a molecular ion peak but yields fragments that (a) confirm that the product retains the formyl function and (b) tentatively identify the axis along which the molecule fragments. Previous mass spectrometric studies have demonstrated that biliverdins fragment at the meso-carbon directly opposite the site of the original oxidative cleavage when subjected to electron impact ionization (20). The two fragments that are formed allow one to distinguish between, for example, biliverdin IXalpha and biliverdin IXbeta but not biliverdin IXalpha and biliverdin IXgamma . When the HPLC-purified mesobiliverdin dimethyl ester obtained from reaction of the alpha -meso-formylmesoheme-hHO-1 complex with NADPH-P-450 reductase was subjected to electron impact ionization, prominent fragments appeared in the mass spectrum at m/z 273 and 370 (Fig. 6). The sum of the two fragments is 643, a value close to that of the mass of the parent mesoformylmesoheme (642), in agreement with the earlier finding that the sum of the fragments gives a value that differs by 1-3 mass units from that of the parent molecule (20). Furthermore, the ions are most consistent with fragmentation of the molecule along the beta -delta axis to give structures similar to those shown in Fig. 7, although the fragment ion at 370 requires a dehydrogenation of the structure shown in the figure. Dehydrogenation reactions that occur in the ion source have been observed in related molecules (24). Fragmentation along the beta -delta axis indicates that the parent porphyrin is oxidized by heme oxygenase either at the beta - or delta -meso carbon. Thus, in contrast to the alpha -meso-methyl analogue (15), alpha -meso-formylmesoheme is oxidized at one or more non-formyl-substituted meso positions rather than at the normally favored alpha -meso carbon.


Fig. 6. Electron impact mass spectrum of the mesobiliverdin isolated from oxidation of alpha -meso-formylmesoheme by hHO-1, P-450 reductase, and NADPH.
[View Larger Version of this Image (16K GIF file)]


Fig. 7. Proposed structures for the principal ions observed in the electron impact mass spectrum of the mesobiliverdin isolated from oxidation of alpha -meso-formylmesoheme by hHO-1, P-450 reductase, and NADPH.
[View Larger Version of this Image (16K GIF file)]

CO Formation

The affinity of ferrous deoxymyoglobin for CO has been used to assay the formation of CO in the NADPH-P-450 reductase-dependent oxidation of alpha -meso-formylmesoheme by hHO-1. The assay is based on the shift of the Soret band of ferrous deoxymyoglobin at 434 nm to that of the CO-complex at 422 nm. Addition of deoxymyoglobin to the incubation mixture after completion of the reaction in a sealed tube resulted in immediate formation of the ferrous myoglobin-CO complex. For technical reasons, a substoichiometric amount (0.90 equivalents) of deoxymyoglobin was added in these studies, and all of the deoxymyoglobin was converted to the ferrous-CO complex. The same results are obtained with unsubstituted mesoheme, the control substrate. In view of our earlier demonstration that oxidation of the substituted position in meso-methylmesohemes does not produce CO (15), the extent of CO formation observed must be due to the nearly quantitative oxidation of unsubstituted positions in the alpha -meso-formylmesoheme.


DISCUSSION

alpha -meso-Hydroxylation, the first step in the hHO-1-catalyzed oxidation of heme (Fig. 1), does not proceed normally if a substituent is attached to the alpha -meso carbon. Earlier studies demonstrated that alpha -meso-methylmesoheme, despite the alpha -meso-substituent, is enzymatically oxidized at the alpha -meso-carbon to give mesobiliverdin IXalpha , but the reaction proceeds without the concomitant formation of CO (15). Thus, the alpha -meso-carbon and the attached methyl group are eliminated by a mechanism that does not converge on the normal alpha -meso-hydroxymesoheme intermediate. The excised two-carbon fragment is not released as acetaldehyde or acetic acid (21), but the identity of the fragment remains unknown. The methyl-substituted position is also exclusively (gamma ) or preferentially (delta ) oxidized in enzymatic turnover of the gamma - or delta -meso-methylmesoporphyrin isomers, giving rise to the corresponding unsubstituted mesobiliverdins (15). As found for the alpha -meso-methyl isomer, no CO is detected in the enzymatic oxidation of gamma -meso-methylmesoheme (21). A methyl substituent thus favors oxidation of the methyl-substituted meso-carbon, albeit by a mechanism that diverges from that which leads to formation of the normal meso-hydroxy intermediate.

The regiochemistries of the hHO-1-catalyzed oxidations of alpha -meso-formyl- and alpha -meso-methylmesoheme are diametrically opposed in that the former results exclusively in oxidation of the substituted position and the latter in oxidation of an unsubstituted position. A single biliverdin-like product peak is observed when the alpha -meso-formylmesoheme reaction product is analyzed by HPLC (Fig. 5). The retention time and the absorption and mass spectra of the product clearly identify it as a formyl-substituted mesobiliverdin. Although a molecular ion is not observed in the mass spectrum (Fig. 6), the two principal peaks observed at m/z 273 and 370 are consistent with fragmentation of a formyl-substituted mesobiliverdin with a molecular mass of 642. Earlier studies showed that biliverdins fragment to give ions such as those observed here (Fig. 7) (20). The molecular masses of the fragments are consistent with fragmentation of the parent mesobiliverdin at either the beta - or delta -meso position, which implies that the heme oxygenase reaction occurred at one of these two positions rather than at the formyl-substituted alpha -meso position. Independent evidence for oxidation at the unsubstituted positions is provided by the finding that CO is formed in amounts (>90%) approaching those expected for quantitative oxidation of the unsubstituted rather than the formyl-substituted, meso positions. The results unambiguously establish that a meso-formyl substituent, in contrast to a methyl substituent, directs the oxidation away from the substituted carbon toward the unsubstituted positions.

The opposite effects of the formyl and methyl groups could be due to the differences in their steric or electronic properties. In steric terms, a formyl group is more compact and therefore presents a smaller steric profile than a methyl group. A value for the steric parameter Es is not available for a formyl group, but the value for the slightly more compact cyano group is -0.51. The corresponding value for methyl group is -1.24 (22). Steric effects therefore cannot be used to rationalize the opposite effects on the reaction regiospecificity of the meso-methyl and meso-formyl substituents, particularly when the substituent is at the normally favored alpha -meso carbon. However, the changes in regiospecificity are readily rationalized by the differences in the electronic effects of the two substituents. A methyl is electron donating (Hammett sigma p = -0.17), whereas a formyl substituent is electron withdrawing (Hammett sigma p = +0.42) (22).

The opposite effects of meso-formyl and meso-methyl substituents on the heme oxygenase reaction regiochemistry indicate that electron donating substituents promote oxidation of the substituted carbon, whereas electron-withdrawing substituents disfavor it. A corollary of this inference is that the regioselectivity of the heme oxygenase reaction is primarily determined by electronic rather than steric features of the active site. A second corollary is that the reaction involves electrophilic attack at the meso carbon, because electron donation facilitates the reaction, whereas electron withdrawal impedes it. Electrophilic attack is consistent with the finding that the ethylhydroperoxide-dependent oxidation of heme by hHO-1 produces alpha -meso-ethoxyheme (16). Furthermore, the 1H NMR spectrum of the hHO-1-heme complex suggests that the distribution of electron density in the highest occupied porphyrin molecular orbital is similar to that seen when the heme bears an electron donating or withdrawing meso-substituent (23). Although the normal heme substrate does not have such a substituent, the results suggest that the protein favors alpha -meso-hydroxylation by causing an analogous asymmetry in the heme electronic structure.

A comparison of the enzymatic oxidation of the meso-methyl and meso-formyl hemes reveals two additional significant differences. The oxidation of the beta -meso-methylmesoheme regioisomer was unusual in that it was particularly slow and provided low to negligible yields of mesobiliverdin-like products (15). In contrast, beta -meso-formylmesoheme is oxidized just as readily as the other meso-formyl isomers and gives similar amounts of mesoverdoheme-like products (Fig. 5). This difference is presumably due to the fact that oxidation of the beta -meso-formyl substrate disfavors the beta -meso-carbon, whereas the beta -meso-carbon is favored in the case of the beta -meso-methyl substrate. The beta -formyl result clearly shows that a substituent at the beta  position does not cause a general perturbation that abrogates the normal catalytic mechanism. This conclusion strengthens our earlier suggestion (15) that a specific steric interaction prevents oxidation of the beta -meso position, an interaction that does not come into play in the case of the beta -meso-formyl heme because oxidation occurs at other positions.

The second anomaly is provided by the finding that the H2O2-dependent heme oxygenase reaction proceeds normally with the alpha -meso-formyl substrate, whereas no reaction was observed earlier with the alpha -meso-methyl substrate (15). This difference is again rationalized by the fact that an alpha -meso-formyl substituent directs the oxidation to the beta - or delta -unsubstituted meso positions, whereas an alpha -meso-methyl favors oxidation of the alpha -meso carbon. Thus, a specific interaction that blocks the H2O2-dependent oxidation of the alpha -meso-substituted carbon would not interfere in the case of the alpha -meso-formyl substrate, which channels the oxidation to the other meso positions.

The key finding of this study is that meso-methyl and meso-formyl substituents have opposite effects on the regiospecificity of the heme oxygenase reaction, the former favoring oxidation of the substituted position and the latter oxidation of an unsubstituted position. This finding leads to two conclusions: (a) that the regiochemistry of the reaction is primarily under electronic rather than steric control and (b) that oxidation involves electrophilic addition to the heme. The normal alpha -meso regiospecificity of the heme oxygenase reaction thus appears to be enforced by interactions of active site residues with the heme that selectively enrich the frontier orbital electron density at the alpha -meso carbon.


FOOTNOTES

*   This work was supported by Grant DK30297 from the National Institutes of Health. Mass spectra were obtained at the Biomedical, Bioorganic Mass Spectrometry Facility of the University of California, San Francisco, which is supported by National Institutes of Health Grants RR 01614 and 5 P30 DK26743.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    To whom correspondence should be addressed. Fax: 415-502-4728; E-mail: ortiz{at}cgl.ucsf.edu.
1   The abbreviations used are: heme, iron protoporphyrin IX regardless of the iron oxidation and ligation states; HO-1, heme oxygenase isozyme-1; hHO-1, truncated human HO-1; HPLC, high pressure liquid chromatography; P-450 reductase, cytochrome P-450 reductase; mesoheme, iron mesoporphyrin IX; CHO, Chinese hamster ovary.

ACKNOWLEDGEMENTS

We thank Weiping Jia for obtaining the mass spectra of the mesobiliverdin products and Bettie Sue Siler Masters for the cytochrome P-450 reductase.


REFERENCES

  1. Maines, M. D. (1992b) Heme Oxygenase: Clinical Applications and Functions, pp. 203-266, CRC Press, Boca Raton, FL
  2. Verma, A., Hirsch, D. J., Glatt, C. E., Ronnett, G. V., and Snyder, S. H. (1993) Science 259, 381-384 [Medline] [Order article via Infotrieve]
  3. Stevens, C. F., and Wang, Y. (1993) Nature 364, 147-149 [CrossRef][Medline] [Order article via Infotrieve]
  4. Marks, G. S. (1994) Cell. Mol. Biol. 40, 863-870
  5. Yoshida, T., Biro, P., Cohen, T., Müller, R. M., and Shibahara, S. (1988) Eur. J. Biochem. 171, 457-461 [Abstract]
  6. Wilks, A., and Ortiz de Montellano, P. R. (1993) J. Biol. Chem. 268, 22357-22362 [Abstract/Free Full Text]
  7. Ishikawa, K., Sato, M., Ito, M., and Yoshida, T. (1992) Biochem. Biophys. Res. Commun. 182, 981-986 [Medline] [Order article via Infotrieve]
  8. Liu, Y., Moënne-Loccoz, P., Loehr, T. M., and Ortiz de Montellano, P. R. (1997) J. Biol. Chem. 272, 6909-6917 [Abstract/Free Full Text]
  9. Matera, K. M., Takahashi, S., Fujii, H., Zhou, H., Ishikawa, K., Yoshimura, T., Rousseau, D. L., Yoshida, T., and Ikeda-Saito, M. (1996) J. Biol. Chem. 271, 6618-6624 [Abstract/Free Full Text]
  10. Saito, S., and Itano, H. A. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, 1393-1397 [Abstract]
  11. Yoshida, T., and Noguchi, M. (1984) J. Biochem. (Tokyo) 96, 563-570 [Abstract]
  12. Schacter, B. A., Nelson, E. B., Marver, H. S., and Masters, B. S. S. (1972) J. Biol. Chem. 247, 3601-3607 [Abstract/Free Full Text]
  13. Tenhunen, R., Marver, H. S., and Schmid, R. (1969) J. Biol. Chem. 244, 6388-6394 [Abstract/Free Full Text]
  14. Torpey, J., and Ortiz de Montellano, P. R. (1995) J. Org. Chem. 60, 2195-2199
  15. Torpey, J. W., and Ortiz de Montellano, P. R. (1996) J. Biol. Chem. 271, 26067-26073 [Abstract/Free Full Text]
  16. Wilks, A., Torpey, J., and Ortiz de Montellano, P. R. (1994) J. Biol. Chem. 269, 29553-29556 [Abstract/Free Full Text]
  17. Brown, S. B., Chabot, A. A., Enderby, E. A., and North, A. C. T. (1981) Nature 289, 93-95 [Medline] [Order article via Infotrieve]
  18. Yoshida, T., Noguchi, M., Kikuchi, G., and Sano, S. (1981) J. Biochem. (Tokyo) 90, 125-131 [Abstract]
  19. Lagarias, J. C. (1982) Biochim. Biophys. Acta. 717, 12-19 [Medline] [Order article via Infotrieve]
  20. Bonnett, R., and McDonagh, A. F. (1973) J. Chem. Soc. Perkin Trans. I 881-888
  21. Torpey, J. (1997) Mechanistic Studies of Home Oxygenase. Ph.D. Dissertation, University of California at San Francisco
  22. Hansch, C., and Leo, A. (1995) Exploring QSAR. Fundamentals and Applications in Chemistry and Biology, pp. 1-96, American Chemical Society, Washington, D.C.
  23. Hernández, G., Wilks, A., Paolesse, R., Smith, K. M., Ortiz de Montellano, P. R., and La Mar, G. N. (1994) Biochemistry 33, 6631-6641 [Medline] [Order article via Infotrieve]
  24. Lightner, D. A., Moscowitz, A., Petryka, Z. J., Jones, S., Weimer, M., Davis, E., Beach, N. A., and Watson, C. J. (1969) Arch. Biochem. Biophys. 131, 566-576 [Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.