(Received for publication, June 13, 1997, and in revised form, July 2, 1997)
From the Department of Pharmaceutical Chemistry, School of Pharmacy, and Liver Center, University of California, San Francisco, California 94143-0446
The oxidation of heme to biliverdin IX by heme
oxygenase involves regiospecific
-meso-hydroxylation
followed by extrusion of the
-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
-meso-methylmesoheme, which is exclusively oxidized at the methyl-substituted position,
-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.
The oxidation of heme1
by heme oxygenase yields biliverdin IX 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).
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
-meso-hydroxyheme (see Fig. 1) (8-12). In the second step,
-meso-hydroxyheme undergoes an
O2-dependent but NADPH-independent (8) reaction
that results in extrusion of the hydroxylated
-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 IX
, the
isomer resulting from oxidation of the
-meso-carbon
(13).
The central role of -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,
-meso-methylmesoheme is still oxidized to mesobiliverdin
IX
. The reaction does not result in the formation of CO, however, in
accord with the fact that normal
-meso-hydroxylation is
impossible due to the presence of the methyl substituent. Equally
surprising was the finding that
-meso-methylmesoheme,
despite the presence of an unsubstituted
-meso position,
is oxidized exclusively at the
-meso position to give
-mesobiliverdin (15).
-meso-Methylmesoheme was
similarly oxidized at the
-meso position to give
-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
-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.
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 EsterThe 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 ( = 16 mg,
25.6 µmol;
= 14 mg, 22.4 µmol;
= 5 mg, 8.0 µmol;
= 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
(6.0 mg, 9.6 µmol, 38%),
(9.3 mg, 15.0 µmol, 67%),
(4.0 mg, 6.4 µmol, 80%), and
(11 mg, 17.7 µmol, 26%) meso-formylmesoporphyrins. The spectroscopic
and analytical data for the isomers are as follows:
isomer,
max (CH3Cl) 404, 504, 538, 576 nm;
1H NMR (CDCl3)
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)
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;
isomer,
max (CH3Cl) 404, 506, 538, 574 nm; 1H NMR (CDCl3)
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)
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;
isomer,
max (CH3Cl) 406, 504, 538, 574 nm; 1H NMR (CDCl3)
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)
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;
isomer,
max 434, 574, 628 (CH3Cl) nm;
1H NMR (CDCl3)
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)
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.
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
(-CHO, 0.29 mg, 9.8 nmol;
-CHO, 0.37 mg, 12.2 nmol;
-CHO, 0.23 mg, 7.8 nmol; and
-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
max 616 nm and the loss of the Soret band of the
starting complex.
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)
(-CHO, 0.27 mg, 9.1 nmol;
-CHO, 0.46 mg, 15.2 nmol;
-CHO, 0.23 mg, 7.8 nmol;
-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.
P-450 reductase (900 µg, 13.4 nmol) and NADPH (83.3 mg, 120 µmol) were added to a solution of the
-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 -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
-meso-formylmesobiliverdin dimethyl ester was then
determined.
To a 1-ml solution of the
-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
-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 (-CHO, 0.27 mg, 9.1 nmol;
-CHO, 0.21 mg, 7.0 nmol;
-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 -meso-formylmesoheme-hHO-1 complex (
-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.
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).
Formation of the Fe(II)Mesoverdoheme-CO Complex Using NADPH-P-450 Reductase
Incubation of the -,
-,
-, and
-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
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
- (Fig. 3A),
- (Fig. 3B), and
-meso-formyl (Fig. 3D) isomers. However, for
the
-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
-isomer is unclear but may be due to an
interaction of the
-formyl group with the flanking propionic acid
side chains.
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).
Regiochemistry of the hHO-1-catalyzed Oxidation of
The regiospecificity of the oxidation of
-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
-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.
The absorption and mass spectra of the product obtained from
-meso-formylmesoheme confirm that it is a mesobiliverdin
that retains the
-meso-formyl group. The maxima in the
absorption spectrum (
max 364, 640 nm) are slightly
shifted with respect to those of authentic mesobiliverdin IX
(
max 366, 636) (15), the product expected if the
-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 IX
and biliverdin IX
but not biliverdin IX
and biliverdin IX
. When the HPLC-purified mesobiliverdin dimethyl
ester obtained from reaction of the
-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
-
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
-
axis indicates that the parent porphyrin is oxidized by heme oxygenase
either at the
- or
-meso carbon. Thus, in contrast to
the
-meso-methyl analogue (15),
-meso-formylmesoheme is oxidized at one or more
non-formyl-substituted meso positions rather than at the
normally favored
-meso carbon.
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
-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
-meso-formylmesoheme.
-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
-meso carbon. Earlier
studies demonstrated that
-meso-methylmesoheme, despite
the
-meso-substituent, is enzymatically oxidized at the
-meso-carbon to give mesobiliverdin IX
, but the
reaction proceeds without the concomitant formation of CO (15). Thus,
the
-meso-carbon and the attached methyl group are
eliminated by a mechanism that does not converge on the normal
-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 (
) or
preferentially (
) oxidized in enzymatic turnover of the
- or
-meso-methylmesoporphyrin isomers, giving rise to the
corresponding unsubstituted mesobiliverdins (15). As found for the
-meso-methyl isomer, no CO is detected in the enzymatic
oxidation of
-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
-meso-formyl- and
-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
-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
- or
-meso position, which implies
that the heme oxygenase reaction occurred at one of these two positions
rather than at the formyl-substituted
-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
-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
p =
0.17), whereas a formyl
substituent is electron withdrawing (Hammett
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 -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
-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 -meso-methylmesoheme
regioisomer was unusual in that it was particularly slow and provided
low to negligible yields of mesobiliverdin-like products (15). In
contrast,
-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
-meso-formyl substrate disfavors the
-meso-carbon, whereas the
-meso-carbon is
favored in the case of the
-meso-methyl substrate. The
-formyl result clearly shows that a substituent at the
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
-meso position, an interaction that does not come into
play in the case of the
-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 -meso-formyl substrate, whereas no reaction was observed earlier with the
-meso-methyl substrate (15). This difference is again
rationalized by the fact that an
-meso-formyl substituent
directs the oxidation to the
- or
-unsubstituted meso
positions, whereas an
-meso-methyl favors oxidation of
the
-meso carbon. Thus, a specific interaction that
blocks the H2O2-dependent oxidation
of the
-meso-substituted carbon would not interfere in
the case of the
-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
-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
-meso carbon.
We thank Weiping Jia for obtaining the mass spectra of the mesobiliverdin products and Bettie Sue Siler Masters for the cytochrome P-450 reductase.