Structure of the modified heme in allylbenzene-inactivated chloroperoxidase determined by Q-band CW and pulsed ENDOR

Article abstract of DOI:10.1021/ja963684c

During the epoxidation of allylbenzene, chloroperoxidase (CPO) is converted to an inactive green species in which the prosthetic heme has been modified by addition of the alkene plus an oxygen atom (Dexter, A. F.; Hager, L. P. J. Am. Chem. Soc. 1995, 117,

Full text of DOI:10.1021/ja963684c

J. Am. Chem. Soc. 1997, 119, 4059-4069
4059
Structure of the Modified Heme in Allylbenzene-Inactivated
Chloroperoxidase Determined by Q-Band CW and Pulsed
ENDOR
Hong-In Lee,† Annette F. Dexter,‡ Yang-Cheng Fann,† Frederick J. Lakner,‡
Lowell P. Hager,‡ and Brian M. Hoffman*,†
Contribution from the Department of Chemistry, Northwestern UniVersity,
EVanston, Illinois 60208-3113, and Department of Biochemistry, UniVersity of Illinois,
Urbana, Illinois 61801
ReceiVed October 23, 1996X
Abstract: During the epoxidation of allylbenzene, chloroperoxidase (CPO) is converted to an inactive green species
in which the prosthetic heme has been modified by addition of the alkene plus an oxygen atom (Dexter, A. F.;
Hager, L. P. J. Am. Chem. Soc. 1995, 117, 817-818). We have used Q-band continuous wave and pulsed electron-
nuclear double resonance (ENDOR) spectroscopy to study the CPO heme in situ following inactivation with
allylbenzene, using samples prepared in natural isotopic abundance, with 15N-labeled enzyme, and with allylbenzene
2
labeled with H or 13C in specific vinylic positions. The electron paramagnetic resonance (EPR) spectrum of the
inactivated enzyme is dominated by a low-spin ferric signal (g1,2,3 ) 2.32, 2.16, 1.95). 14,15N ENDOR examination
of allylbenzene-inactivated CPO reveals that three nitrogens of the heme are similar, but the fourth nitrogen is markedly
different, suggesting that a single pyrrole ring has been covalently modified at the unique nitrogen. These studies
also reveal the orientation of the g tensor relative to the heme. 13C ENDOR of allylbenzene-inactivated CPO with
13C-labeled allylbenzene shows that the C-1 and C-2 carbons of allylbenzene are covalently connected to the heme
system. 1,2H ENDOR plus mass analysis of CPO heme after inactivation with deuterated allylbenzene show that all
three vinylic protons are retained in the heme adduct. No strongly-coupled exchangeable protons are observed,
indicating that the axially bound water of frozen native CPO has been displaced. The 1H at the C-2 position of the
alkene shows strong, mostly isotropic hyperfine coupling while the two hydrogens at the C-1 position show weak,
dipolar couplings. The hyperfine tensors of 1,2H of the C-1 position of allylbenzene have been determined, and give
the position of these atoms relative to the heme. These data, combined with molecular modeling calculations, have
been used to deduce that the allylbenzene-bound heme of inactivated CPO is an N-alkylhemin metallocycle with
C-1 of allylbenzene bonded to the pyrrole nitrogen and to obtain metrical details of its structure.
Chloroperoxidase (CPO)1 is a highly versatile heme enzyme
that catalyzes the oxidation of a broad range of substrates with
hydrogen peroxide as oxidant. During the catalytic cycle, native
ferric CPO reacts with hydrogen peroxide to form a two-electron
oxidized intermediate, compound I, which then reacts with an
organic or inorganic substrate to regenerate the ferric enzyme
and an oxidized species (Scheme 1). The first reaction reported
to be catalyzed by CPO was the halogenation of organic
substrates.2 Recent interest in CPO has focused on its capacity
for the epoxidation of alkenes.3 While the ability to epoxidize
alkenes in an oxo-transfer reaction is common to a number of
enzymatic and model heme systems, CPO is of interest in that
it catalyzes the oxo-transfer with high facioselectivity, leading
to enantiomerically enriched products.
Scheme 1
† Northwestern University.
‡ University of Illinois.
X Abstract published in AdVance ACS Abstracts, April 15, 1997.
(1) Abbreviations used: AB-CPO (allylbenzene-CPO), green enzyme
species produced by reaction of native CPO with allylbenzene and hydrogen
peroxide; CPO, chloroperoxidase; CW, continuous wave; DEAE, (diethy-
lamino)ethyl; EDTA, ethylenediaminetetraacetic acid; ENDOR, electron
nuclear double resonance spectroscopy; EPR, electron paramagnetic
resonance spectroscopy; hemin, iron(III) porphyrin; MS, mass spectroscopy;
NMR, nuclear magnetic resonance spectroscopy; P450, cytochrome P450;
RF, radio frequency; Rz, purity number A400/A280; THF, tetrahydrofuran.
(2) Shaw, P. D.; Hager, L. P. J. Biol. Chem. 1960, 236, 1626-1630.
(3) (a) Dexter, A. F.; Lakner, F. J.; Campbell, R. A.; Hager, L. P. J. Am.
Chem. Soc. 1995, 117, 6412-6413. (b) Colonna, S.; Gaggero, N.; Casella,
L.; Carrea, G.; Pasta, P. Teteahedron: Asymmetry 1993, 4, 1325-1330. (c)
Lakner, F. J.; Hager, L. P. J. Org. Chem. 1996, 61, 3923-3925. (d) Allain,
E. J.; Hager, L. P.; Deng, L.; Jacobsen, E. N. J. Am. Chem. Soc. 1993, 115,
4415-4416.
In a number of heme enzyme and heme model systems, the
catalytic epoxidation of monosubstituted alkenes is accompanied
by an autoinactivation reaction that produces modified hemins
containing the alkene skeleton. The formation of these “green
pigments” was first observed with hepatic microsomal cyto-
chrome P450 (P450) in ViVo,4 and has since been demonstrated
with several model hemins5,6 and more recently in our laboratory
with chloroperoxidase (CPO).3a,7 Surprisingly, the heme-
alkene adducts spontaneously revert to the native heme on
S0002-7863(96)03684-0 CCC: $14.00 © 1997 American Chemical Society

4060 J. Am. Chem. Soc., Vol. 119, No. 17, 1997
Lee et al.
standing, with both CPO7 and model systems.6 A similar
reversibility for the P450 inactivation reaction has been predicted,6c
but has not been reported.
In all alkene-modified heme systems studied to date, N-(2-
hydroxyalkyl)porphyrins (1) containing the alkene skeleton (or
their oxidized derivatives)5c,e can be isolated from the epoxi-
dation reaction following acid demetalation of the hemin. At
least five distinct heme adducts might give rise to this
chemistry: three metallocycles (2-4) with differing Fe,N bridge
lengths;6,8 a â-hydroxycarbene (5);9 and a σ-alkylhemin (6).10
enzymatic13 heme systems that organometallic heme species
including porphyrin iron carbenes and σ-alkyliron porphyrins
(and their σ-aryl or σ-vinyl analogs) (i) may be generated under
catalytic conditions and (ii) undergo a reversible, redox-induced
metal-to-nitrogen ligand migration to yield N-substituted por-
phyrin products. Ligand migration of this kind has been
proposed to be responsible for the isolation of N-substituted
porphyrins from heme enzyme systems following inactivation
with xenobiotics including alkyl (or aryl) hydrazines,13 syd-
nones,14 collidine derivatives,15 and griseofulvin.16
On the basis of results with model hemins, the recent
consensus is that the products of alkene-mediated inactivation
in both model and enzymatic systems are N-alkylhemin met-
allocycles of type 2.4-6 For the model systems, this structural
assignment appears well founded. NMR studies of an intact
alkene-modified tetraarylhemin6b have shown shifts in one set
of â-pyrrole proton resonances, consistent with covalent modi-
1
fication of a single pyrrole ring and with H shifts seen in
synthetic N-alkylhemins.17 While an observed shift in NMR
resonances for one set of pyrrole protons does not discriminate
between structures 2-4, reconstitution of N-(2-hydroxyalkyl)-
porphyrins with Fe(III) and appropriate ligands gives hemins
with electronic spectra similar to those of the primary products
of alkene-mediated inactivation,6a,b and the EPR and optical
spectra of these hemins are consistent with axial alkoxide
ligation, as opposed to σ-alkyl ligation.
To date, no comparable structural studies have been carried
out on the intact enzymatic heme adducts. For P450, in situ
studies of the modified hemin have been hampered by the
difficulty of obtaining sufficiently clean preparations of inac-
tivated enzyme, although detailed NMR studies have been
conducted on demetalated porphyrins isolated from P450. N-(2-
Hydroxyalkyl)protoporphyrin IX derivatives isolated from alk-
ene-inactivated P450 reveal a marked regioselectivity and
stereoselectivity of pyrrole N-alkylation, suggesting that the
pyrrole alkylation occurs within the enzyme active site.4a,b
However, enzyme reconstitution with an alkene-modified hemin
In each case, two orientations of addition are possible for a
monosubstituted alkene (R or R′ ) H), for a total of ten different
possible structures. The N-(2-hydroxyalkyl)porphyrins isolated
from alkene-inactivated enzymatic or model systems after acid
treatment are simple demetalation products of 2. However, the
same demetalated porphyrins might also arise by rearrangement
of the organometallic species 3-6 under the conditions of
workup.11 It is known from a variety of model8,10-12 and
(4) (a) Kunze, K. L.; Mangold, B. L. K.; Wheeler, C.; Beilan, H. S.;
Ortiz de Montellano, P. R. J. Biol. Chem. 1983, 258, 4202-4207. (b) Ortiz
de Montellano, P. R.; Mangold, B. L. K.; Wheeler, C.; Kunze, K. L.; Reich,
N. O. J. Biol. Chem. 1983, 258, 4208-4213. (c) Ortiz de Montellano, P.
R.; Correia, M. A. In Cytochrome P450: Structure, Mechanism, and
Biochemistry, 2nd ed.; Ortiz de Montellano, P. R., Ed.; Plenum Press: New
York, 1995; pp 305-364. (d) Shirane, N.; Sui, Z.; Peterson, J. A.; Ortiz
de Montellano, P. R. Biochemistry 1993, 32, 13732-13741.
(5) (a) Mansuy, D.; Devocelle, L.; Artaud, I.; Battioni, J.-P. NouV. J.
Chem. 1985, 9, 711-716. (b) Collman, J. P.; Hampton, P. D.; Brauman,
J. I. J. Am. Chem. Soc. 1986, 108, 7861-7862. (c) Artaud, I.; Devocelle,
L.; Battioni, J.-P.; Girault, J.-P.; Mansuy, D. J. Am. Chem. Soc. 1987, 109,
3782-3783. (d) Artaud, I.; Gre´goire, N.; Mansuy, D. New J. Chem. 1989,
13, 581-586. (e) Nakano, T.; Traylor, T. G.; Dolphin, D. Can. J. Chem.
1990, 68, 1504-1506. (f) Collman, J. P.; Hampton, P. D.; Brauman, J. I.
J. Am. Chem. Soc. 1990, 112, 2977-2986. (g) Collman, J. P.; Hampton,
P. D.; Brauman, J. I. J. Am. Chem. Soc. 1990, 112, 2986-2998.
(6) (a) Mashiko, T.; Dolphin, D.; Nakano, T.; Traylor, T. G. J. Am. Chem.
Soc. 1985, 107, 3735-3736. (b) Traylor, T. G.; Nakano, T.; Miksztal, A.
R.; Dunlap, B. E. J. Am. Chem. Soc. 1987, 109, 3625-3632. (c) Tian,
Z.-Q.; Richards, J. L.; Traylor, T. G. J. Am. Chem. Soc. 1995, 117, 21-29.
(7) Dexter, A. F.; Hager, L. P. J. Am. Chem. Soc. 1995, 117, 817-818.
(8) (a) Chevrier, B.; Weiss, R.; Lange, M.; Chottard, J.-C.; Mansuy, D.
J. Am. Chem. Soc. 1981, 103, 2899-2901. (b) Latos-Grazynski, L.; Cheng,
R.-J.; La Mar, G. N.; Balch, A. L. J. Am. Chem. Soc. 1981, 103, 4270-
4272. (c) Mansuy, D.; Morgenstern-Badarau, I.; Lange, M.; Gans, P. Inorg.
Chem. 1982, 21, 1427-1430. (d) Olmstead, M. M.; Cheng, R.-J.; Balch,
A. L. Inorg. Chem. 1982, 21, 4143-4148. (e) Balch, A. L.; Cheng, R.-J.;
La Mar, G. N.; Latos-Grazynski, L. Inorg. Chem. 1985, 24, 2651-2656.
(9) (a) Groves, J. T.; Avaria-Neisser, G. D.; Fish, K. M.; Imachi, M.;
Kuczkowski, R. L. J. Am. Chem. Soc. 1986, 108, 3837-3838. (b) Dolphin,
D.; Matsumoto, A.; Shortman, C. J. Am. Chem. Soc. 1989, 111, 411-413.
(10) (a) Lexa, D.; Save´ant, J.-M.; Battioni, J.-P.; Lange, M.; Mansuy,
D. Angew. Chem., Int. Ed. Engl. 1981, 20, 578-579. (b) Battioni, P.; Mahy,
J. P.; Gillet, G.; Mansuy, D. J. Am. Chem. Soc. 1983, 105, 1399-1401. (c)
Tabard, A.; Cociolos, P.; Lagrange, G.; Gerardin, R.; Hubsch, J.; Lecomte,
C.; Zarembowitch, J.; Guilard, R. Inor´g. Chem. 1988, 27, 110-117. (d)
Guilard, R.; Kadish, K. M. Chem. ReV. 1988, 88, 1121-1146.
(11) (a) Lange, M.; Mansuy, D. Tetrahedron Lett. 1981, 22, 2561-2564.
(b) Ortiz de Montellano, P. R.; Kunze, K. L.; Augusto, O. J. Am. Chem.
Soc. 1982, 104, 3545-3546. (c) Lanc¸on, D.; Cociolos, P.; Guilard, R.;
Kadish, K. M. J. Am. Chem. Soc. 1984, 106, 4472-4478. (d) Balch, A.
L.; Renner, M. W. J. Am. Chem. Soc. 1986, 108, 2603-2608. (e) Mansuy,
D.; Battioni, P.; Battioni, J.-P. Eur. J. Biochem. 1989, 184, 267-285.
(12) (a) Callot, H. J.; Schaeffer, E. Tetrahedron Lett. 1980, 21, 1335-
1338. (b) Mansuy, D.; Battioni, J.-P.; Dupre´, D.; Sartori, E.; Chottard, G.
J. Am. Chem. Soc. 1982, 104, 6159-6161. (c) Mansuy, D. Pure Appl.
Chem. 1987, 59, 759-770. (d) Artaud, I.; Gregoire, N.; Battioni, J.-P.;
Dupre, D.; Mansuy, D. J. Am. Chem. Soc. 1988, 110, 8714-8716.
(13) (a) Jonen, H. G.; Werringloer, J.; Prough, R. S.; Estabrook, R. W.
J. Biol. Chem. 1982, 257, 4404-4421. (b) Battioni, P.; Mahy, J.-P.;
Delaforge, M.; Mansuy, D. Eur. J. Biochem. 1983, 134, 241-248. (c)
Ringe, D.; Petsko, G. A.; Kerr, D. E.; Ortiz de Montellano, P. R.
Biochemistry 1984, 23, 2-4. (d) Raag, R.; Swanson, B. A.; Poulos, T. L.;
Ortiz de Montellano, P. R. Biochemistry 1990, 29, 8119-8126. (e)
Samokyszyn, V. M.; Ortiz de Montellano, P. R. Biochemistry 1991, 30,
11646-11653. (f) Mahy, J. P.; Gaspard, S.; Mansuy, D. Biochemistry 1993,
32, 4014-4021. (g) Gerber, N. C.; Ortiz de Montellano, P. R. J. Biol.
Chem. 1995, 270, 17791-17796.
(14) (a) Grab, L. A.; Swanson, B. A.; Ortiz de Montellano, P. R.
Biochemistry 1988, 27, 4805-4814. (b) Ortiz de Montellano, P. R.; Grab,
L. A. J. Am. Chem. Soc. 1986, 108, 5584-5589.
(15) (a) Ortiz de Montellano, P. R.; Beilan, H. S.; Kunze, K. L. J. Biol.
Chem. 1981, 256, 6708-6713. (b) Ortiz de Montellano, P. R.; Beilan, H.
S.; Kunze, K. L. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 1490-1494.
(16) (a) De Matteis, F.; Gibbs, A. H.; Martini, S. R.; Milek, R. L. B.
Biochem. J. 1991, 280, 813-816. (b) Holley, A. E.; Frater, Y.; Gibbs, A.
H.; De Matteis, F.; Lamb, J. H.; Farmer, P. B.; Naylor, S. Biochem. J.
1991, 274, 843-848.
(17) (a) Balch, A. L.; Chan, Y.-W.; La Mar, G. N.; Latos-Grazynski,
L.; Renner, M. W. Inorg. Chem. 1985, 24, 1437-1443. (b) Balch, A. L.;
La Mar, G. N.; Latos-Grazynski, L.; Renner, M. W. Inorg. Chem. 1985,
24, 2432-2436. (c) Balch, A. L.; Cornman, C. R.; Latos-Grazynski, L.;
Olmstead, M. M. J. Am. Chem. Soc. 1990, 112, 7552-7558.

Modified Heme in Allylbenzene-InactiVated CPO
J. Am. Chem. Soc., Vol. 119, No. 17, 1997 4061
be determined by analyzing the full 2-D pattern generated by collecting
ENDOR spectra across the whole EPR envelope, as described
elsewhere.23
has not been realized for either P450 or CPO. At the same
time, very few spectroscopic studies have been carried out on
the intact alkene-inactivated enzymes, hindering comparisons
with model hemins.
MM2 Calculation. MM2 energy minimization was performed by
using the molecular modeling and analysis program “CS Chem3D”
(CambridgeSoft Co., Cambridge, MA). The program employs Al-
linger’s MM2 force field for energy minimizations.24 In modeling the
heme structure of AB-CPO, the local geometry of Fe(III) was initially
assumed to be square pyramidal. The heme nitrogens were assigned
to pyrrole or ammonium nitrogen (see Results).
During the epoxidation of allylbenzene (7), chloroperoxidase
Materials. Phenylacetic acid-1-13C (99 atom % 13C) and io-
domethane-13C (99 atom % 13C) were from Sigma. Lithium aluminum
deuteride (96 atom % D) was from Aldrich. Na15NO3 (98 atom %
15N) was from Cambridge Isotope Laboratories. Phenylacetaldehyde
and pyridine were distilled prior to use. Tetrahydrofuran was distilled
from sodium benzophenone ketyl. Dess-Martin periodinane was
prepared as described.25 Other reagents were of the highest purity
available from Aldrich and were used as received. CPO was produced
in cultures of C. fumago grown in fructose-salts medium. For
preparation of uniformly 15N-labeled enzyme, Na15NO3 was used as
the sole nitrogen source in the fungal growth medium. CPO was
isolated from the growth medium by two cycles of ethanol precipitation,
followed by dialysis against 10 mM K+ phosphate at pH 5.8. Ethanol-
precipitated enzyme was purified further on DEAE-cellulose in a
gradient of 10-300 mM K+ phosphate at pH 5.8 containing 2 mM
EDTA. Native CPO purified by this method had RZ greater than 1.45
and was substantially free of spin-active metal contaminants.
(CPO) is converted to an inactive green species in which the
prosthetic heme has been modified by addition of the alkene
plus an oxygen atom.7 In the present work we have used
Q-band continuous wave (CW) and pulsed electron-nuclear
double resonance (ENDOR) spectroscopy to study in situ the
heme of CPO following inactivation with allylbenzene, using
samples prepared in natural isotopic abundance, with 15N-labeled
enzyme, and with allylbenzene labeled with 2H or 13C in specific
vinylic positions. These studies, in conjunction with molecular
modeling (MM2) calculations, have determined the in situ
structure of the alkene-modified heme.
Materials and Methods
Instrumentation. 1H and 13C NMR spectra were recorded on a
General Electric QE300 with tetramethylsilane as internal standard.
Fourier transform infrared spectra were recorded on a Mattson Galaxy
5000 spectrometer. Positive ion electrospray mass spectra were
recorded on a Fisons VG Quattro quadrupole-hexapole-quadrupole
mass spectrometer employing a flow system of 1:1 acetonitrile/water.
For mass analysis of the modified hemin present in inactivated CPO
following reaction with unlabeled or 2H-labeled allylbenzene, samples
containing 10-20 pmol/µL of CPO were denaturated by addition of
50% (v/v) acetonitrile 1-2 h prior to analysis then acidified with 0.1%
(v/v) formic acid immediately before injection.
Preparation of ENDOR Samples. Allylbenzene-inactivated CPO
was prepared by reacting the native enzyme with hydrogen peroxide
and alkene in an aqueous emulsion. In a typical procedure, native CPO
(85 mg) was added to a rapidly stirred suspension of 50 mg of
allylbenzene in 20 mM K+ acetate, pH 5.2 (40 mL). Hydrogen peroxide
(1.0 mM/min) was added over 25 min until the reaction turned bright
green. At this point, only trace chlorinating activity could be detected
in the enzyme preparation. Further workup of the inactivated enzyme
was carried out at 4 °C to slow the spontaneous reactivation. The
sample was passed over a 0.25 µm syringe filter to remove particulates,
exchanged three times into fresh 20 mM K+ acetate, pH 5.2, then
concentrated and frozen in ENDOR tubes and stored in liquid nitrogen.
Samples to be measured in D2O solution were prepared in H2O buffer
then exchanged six times into 20 mM K+ acetate-d4 at pD 5.2 in D2O.
Concentrations of allylbenzene-CPO were estimated with use of ꢀ419
) 75 000 M-1 cm-1 and were 2-3 mM in ENDOR samples.
Isotopically enriched allylbenzene-CPO was prepared as above, except
that allylbenzene-2-d, allylbenzene-1,1-d2, allylbenzene-1-13C, or al-
lylbenzene-2-13C was substituted for the unlabeled alkene or 15N-labeled
CPO for unlabeled enzyme.
CW-EPR and ENDOR spectra were recorded at 2 K on a modified
Varian E109 EPR spectrometer equipped with an E110 35 GHz
microwave bridge, using 100 kHz field modulation as described
elsewhere.18 The ENDOR response was observed as a change of EPR
intensity detected in the dispersion mode under condition of “rapid
passage”.19 Pulsed ENDOR spectra were recorded on a locally
constructed X-band (9.5 GHz)20 and Q-band (35 GHz)21 spectrometers,
using a Mims pulsed ENDOR sequence (π/2-τ-π/2-T-π/2).22
ENDOR Frequencies. The first-order ENDOR transitions for a
1
nucleus N with spin I ) /2 are given by
Synthesis of Labeled Allylbenzenes. Labeled allylbenzenes, except
for allylbenzene-1-d2, were prepared by subjecting appropriate pheny-
lacetaldehydes to Wittig methylenation. Labeled 2-phenylethanols were
made by reduction of phenylacetic acid with lithium aluminum deuteride
or reduction of phenylacetic acid-1-13C with lithium aluminum hydride.
Alcohols were oxidized to phenylacetaldehydes with use of Dess-Martin
reagent, which gave clean product nearly quantitatively. An attempt
to prepare allylbenzene-1-d2 with phenylacetaldehyde and CD2P(Ph)3
resulted in scrambled label. 1H NMR revealed one proton attached to
each of the three carbons of the allyl moiety, suggesting rapid H
exchange preceding olefination. This facile enolization probably
contributes to the modest yields associated with the olefination step.
Synthesis of allylbenzene-1-d2 was finally accomplished by to a
ν( ) |νN ( AN/2|
(1)
where νN and AN are the nuclear Larmor frequency and the hyperfine
coupling constant, respectively. For a nucleus with spin I g 1, the ν(
transitions will be further split into 2I lines by the quadrupole
interaction,
ν((mI) ) |νN ( AN/2 + (3PN/2)(2mI - 1)|
(2)
where PN is the quadrupole coupling constant and I g mI > -I + 1.
For a frozen-solution sample with a rhombic EPR signal (g1, g2, g3),
the full hyperfine tensor of a hyperfine-coupled nucleus, including both
the principal values and the orientation in the g-tensor axis frame, can
(23) (a) Hoffman, B. M.; Gurbiel, R. J.; Werst, M. M.; Sivara, M. In
AdVanced EPR. Applications in Biology and Biochemistry; Hoff, A. J.,
Ed.; Elsevier: Amsterdam, 1989; pp 541-591. (b) Hoffman, B. M.; De
Rose, V. J.; Doan, P. E.; Gurbiel, R. J.; Houseman, A. L. P.; Telser, J. In
EMR of Paramagnetic Molecules; Biological Magnetic Resonance 13;
Berliner, L. J., Reuben, J., Eds.; Plenum Press: New York, 1993; pp 151-
218. (c) Hoffman, B. M. Acc. Chem. Res. 1991, 24, 164-170.
(24) (a) Burkert, U.; Allinger, N. L. Molecular Mechanics; American
Chemical Society: Washington, DC, 1982. (b) Clark, T. C. Computational
Chemistry; Wiley: New York, 1985.
(18) Werst, M. M.; Davoust, C. E.; Hoffman, B. M. J. Am. Chem. Soc.
1991, 113, 1533-1538.
(19) (a) Mailer, C.; Taylor, C. P. S. Biochim. Biophys. Acta 1973, 322,
195-203. (b) Feher, G. Phys. ReV. 1959, 114, 1219-1244.
(20) Fan, C.; Doan, P. E.; Davoust, C. E.; Hoffman, B. M. J. Magn.
Reson. 1992, 98, 62-72.
(21) Davoust, C. E.; Doan, P. E.; Hoffman, B. M. J. Magn. Reson. 1996,
A119, 38-44.
(22) (a) Mims, W. B. Proc. R. Soc. London 1965, A283, 452. (b)
Gemperle, C.; Schweiger, A. Chem. ReV. 1991, 91, 1481-1505.
(25) (a) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155-4156.
(b) Ireland, R. E.; Liu, L. J. Org. Chem. 1993, 58, 2899.

4062 J. Am. Chem. Soc., Vol. 119, No. 17, 1997
Lee et al.
modified procedure of Kwart.26 Lithium aluminum deuteride reduction
of phenylpropionic acid was followed by a three-step elimination
process involving tosylation, displacement of tosylate with phenylse-
lenide, and oxidation to a selenoxide which spontaneously eliminates.
2-Phenylethanol-1-d2. Phenylacetic acid (1.00 g, 7.3 mmol) in 7
mL of THF was added slowly at room temperature and with good
stirring to lithium aluminum deuteride (350 mg, 8.3 mmol) dissolved
in 20 mL of THF. After 3.5 h, excess lithium aluminum deuteride
was quenched by slow addition of 1 mL of methanol followed by 10
mL of 5% HCl. The mixture was diluted with ether and 10% HCl,
and the layers were separated. The aqueous phase was extracted with
ether, and the combined organic phases were washed with saturated
NaHCO3 and brine. After drying (MgSO4) and removal of solvent, a
colorless oil remained (0.89 g): 1H NMR δ 2.85 (s, 2H, CH2CD2OH),
7.2-7.4 (m, 5H, aryl); 13C NMR δ 39.0, 62.9 (pentet, J ) 22 Hz),
126.4, 128.5, 129.0, 138.5; IR (thin film) 3600-3100, 2207, 2105, 1603,
mmol) in 3 mL of THF was added dropwise. Stirring was continued
for 1 h, after which the excess phosphorane was quenched with H2O.
The mixture was diluted with 25 mL of ether, and the layers were
separated. The aqueous layer was extracted with fresh ether, and the
combined ether layers were washed with brine, dried (MgSO4), and
concentrated. The residue was diluted with pentane, and triphenylphos-
phonium oxide was removed by filtration. After the mixture was chilled
overnight at 4 °C to complete precipitation, the supernatant was
decanted, the solvent was evaporated, and the remaining oil was flash
chromatographed on 50 g of silica with pentane as eluant. After
removal of solvent, 380 mg of pure alkene (46% overall yield) was
obtained: 1H NMR δ 3.41 (dd, J ) 6.7, 7.6 Hz, 2H, PhCH2), 5.07-
5.13 (overlapping multiplets, 2H, CHdCH2), 5.98 (dm, J ) 154 Hz,
1H, CHdCH2), 7.2-7.4 (m, 5H, aryl); 13C NMR δ 137.4; IR (thin
film) 1612, 1602, 1495, 1453, 1432, 990, 913, 740, 698 cm-1
.
Allylbenzene-1-13C. Triphenylphosphine (1.8 g, 7.0 mmol) was
dissolved in 5 mL of ether, and iodomethane-13C (1.0 g, 7.0 mmol)
was added. After 24 h in a sealed flask at room temperature, 2.65 g
(93%) methyltriphenylphosphonium iodide was collected by filtration.
This labeled phosphonium salt was treated with phenylacetaldehyde
(0.7 mL) and potassium tert-butoxide (0.66 g) in a manner identical
with the preparation of allylbenzene-2-13C to give 100 mg (14% overall
yield) of product. 1H NMR δ 3.41 (dd, J ) 6.2, 6.1 Hz, 2H, PhCH2),
5.07-5.13 (dm, J ) 156 Hz, 2H, CHdCH2), 6.00 (m, 1H, CHdCH2),
7.2-7.4 (m, 5H, aryl); 13C NMR δ 115.8; IR (thin film) 1617, 1602,
1496, 1453, 1123, 1098, 744, 699 cm-1
.
2-Phenylethanol-1-13C. The synthesis was carried out as for
2-phenylethanol-1-d2, with phenylacetic acid-1-13C (1.00 g) substituted
for the unlabeled acid and lithium aluminum hydride substituted for
the deuteride. The yield was 0.89 g: 1H NMR δ 2.87 (dt, J ) 5.6, 6.7
Hz, 2H, CH2CH2OH), 3.85 (dt, J ) 143, 6.7 Hz, 2H, CH2CH2OH),
7.2-7.4 (m, 5H, aryl); 13C NMR δ 63.6; IR (thin film) 3600-3100,
1497, 1453, 1029, 746, 699 cm-1
.
3-Phenylpropanol-1-d2. The same procedure was used as for
2-phenylethanol-1-d2, except that 3-phenylpropanoic acid (2.1 g, 14
mmol) was used instead of phenylacetic acid. Lithium aluminum
deuteride (6450 mg, 15.4 mmol) was used as reductant and the reaction
scaled up proportionately to give 1.87 g of product: 1H NMR δ 1.90
(t, J ) 7.5 Hz, 2H, PhCH2), 2.73 (t, J ) 7.5 Hz, 2H, CH2CD2OH),
7.2-7.4 (m, 5H, aryl); 13C NMR δ 32.1, 34.0, 61.4 (quintet, J ) 21.6
Hz), 125.9, 128.4, 128.5, 141.9; IR (thin film) 3600-3100, 2203, 2106,
1495, 1453, 1432, 1030, 994, 906, 740, 698 cm-1
.
Allylbenzene-1-d2. Crude 3-phenylpropanol-1-d2 prepared above
was dissolved in 10 mL of CH2Cl2. p-Toluenesulfonyl chloride (2.9
g, 15.4 mmol) and pyridine (1.22 g, 15.4 mmol) were added. After
being stirred at room temperature for 65 h, the reaction mixture was
diluted with 70 mL of ether and extracted sequentially with 10 mL of
H2O and 10 mL of 10% HCl. The organic solution was stirred
vigorously with 25 mL of concentrated NH4OH for 1 h until TLC
revealed that the excess tosyl chloride had reacted. The layers were
separated, the organic layer was extracted sequentially with 2 N NaOH,
H2O, and brine dried (MgSO4), and the solvent was evaporated to give
3.1 g of colorless oil. Diphenyl diselenide (2.26 g, 7.2 mmol) was
dissolved in 40 mL of ethanol, and sodium borohydride (0.53 g, 14.0
mmol) was spooned in portionwise. After the mixture was cooled to
0 °C, the tosylate-d2 in 30 mL of THF was added. The reaction mixture
was warmed to room temperature and stirred for 4 h before being poured
into 5% K2CO3. Extraction with ether, drying (MgSO4), and removal
of solvent gave 3.1 g of crude selenide as a yellow oil. The residue
was taken up into 30 mL of CH2Cl2, and cumene hydroperoxide (2.04
g, 1.0 equiv based on 80% tech grade) was added. After 24 h of stirring
at room temperature, the mixture was diluted with an equal volume of
pentane and filtered through silica gel. The solvent was stripped, and
flash chromatography (42 g of silica gel, pentane) was performed to
give 340 mg of colorless oil. 1H NMR revealed minor contamination
by cumene, but this did not interfere with the use of allylbenzene-1-d2
for enzyme reactions: 1H NMR δ 3.43 (d, J ) 6.7 Hz, 2H, PhCH2),
6.00 (broad m, 1H, CHdCH2), 7.2-7.4 (m, 5H, aryl); 13C NMR δ
40.2, 126.1, 128.4, 128.6, 137.3, 140.1.
1602, 1496, 1454, 1191, 1134, 1099, 965, 747, 699 cm-1
.
Phenylacetaldehyde-1-d. 2-Phenylethanol-1-d2 (0.66 g, 5.3 mmol)
in 7 mL of CH2Cl2 was added to Dess-Martin periodinane (2.6 g, 6.9
mmol) in 25 mL of CH2Cl2 and stirred for 40 min. The mixture was
diluted with 100 mL of ether and poured into a mixture of saturated
NaHCO3 and saturated Na2S2O3. Stirring was continued until no solid
remained. The layers were separated, and the organic phase was
washed with saturated NaHCO3, then H2O. After drying (MgSO4),
filtration through silica gel, and evaporation of solvent, a light yellow
oil remained (640 mg): 1H NMR δ 3.69 (s, 2H, PhCH2), 7.2-7.4 (m,
5H, aryl).
Phenylacetaldehyde-1-13C. A scaled-up procedure as for pheny-
lacetaldehyde-1-d was used. 2-Phenylethanol-1-13C (0.85 g) gave a
yellow oil quantitatively (0.83 g): 1H NMR δ 3.70 (dd, J ) 2.3, 7.4
Hz, 2H, PhCH2), 7.2-7.4 (m, 5H, aryl), 9.75 (dt, J ) 2.3, 176 Hz, 1H,
CH2CHO); 13C NMR δ 199.5; IR (thin film) 2819, 1683, 1498, 1454,
749, 700 cm-1
.
Allylbenzene-2-d. To methyltriphenylphosphonium bromide (2.34
g, 6.5 mmol) and potassium tert-butoxide (0.67 g, 6.0 mmol) under
nitrogen was added 20 mL of THF. After the mixture was stirred for
30 min at room temperature, phenylacetaldehyde-1-d (0.64 g, 5.3 mmol)
in 3 mL of THF was added dropwise. Stirring was continued for 1 h,
excess phosphorane was quenched with silica gel, and the mixture was
filtered. The filtrate was concentrated and diluted with pentane, and
the triphenylphosphonium oxide precipitate was removed by filtration.
The filtrate was allowed to stand overnight at 4 °C to complete
precipitation, the supernatant was decanted, solvent was evaporated,
and the residue was distilled in an Aldrich Quick-distill apparatus to
give the pure alkene (170 mg, 24% overall yield): 1H NMR δ 3.41 (s,
2H, PhCH2), 5.09 (s, 2H, CHdCH2), 7.2-7.4 (m, 5H, aryl); 13C NMR
δ 40.2, 115.7, 126.1, 128.5, 128.6, 137.2 (t, J ) 24.2 Hz), 140.1; IR
(thin film) 3082, 3064, 3028, 2903, 1622, 1602, 1495, 1453, 1432,
Results
Mass Analysis of Allylbenzene-Modified Heme. When
AB-CPO is denatured with acetonitrile and analyzed by elec-
trospray MS, the major low molecular weight (i.e. non-protein)
component is observed at m/z 750.4.7 From the half-integer
spacing of isotope peaks, this metalloporphyrin species is
deduced to be doubly charged and hence, dimeric. This species
most likely represents a µ-oxo dimer of the modified hemin.
Formation of µ-oxo dimers from model N-alkylhemins has been
reported elsewhere.27 The formation of this dimer appears to
be facile not only under electrospray MS conditions but also in
larger-scale extractions. As extracted from AB-CPO, the
1076, 1030, 915, 842, 740, 698 cm-1
.
Allylbenzene-2-13C. To methyltriphenylphosphonium bromide (2.83
g, 7.9 mmol) and potassium tert-butoxide (0.89 g, 7.9 mmol) under
nitrogen was added 20 mL of THF. After the mixture was stirred for
30 min at room temperature, phenylacetaldehyde-1-13C (0.80 g, 6.6
(27) (a) Wyslouch, A.; Latos-Grazynski, L.; Grzeszczuk, M.; Drabent,
K.; Bartczak, T. J. Chem. Soc., Chem. Commun. 1988, 1377-1378. (b)
Bartczak, T. J.; Latos-Grazynski, L.; Wyslouch, A. Inorg. Chim. Acta 1990,
171, 205-212.
(26) Kwart, H.; Brechbiel, M.; Miles, W.; Kuart, L. D. J. Org. Chem.
1982, 47, 4524-28.

Modified Heme in Allylbenzene-InactiVated CPO
J. Am. Chem. Soc., Vol. 119, No. 17, 1997 4063
dimeric hemin is EPR-silent and exhibits broad electronic
absorption bands at 408, 504, and 596 nm (Dexter, A. F.; Xia,
Y.-M.; Debrunner, P. G. Unpublished data). The mass change
of the modified hemin relative to heme b (m/z ) +134 per
monomer) corresponds to the addition of allylbenzene plus an
oxygen atom, consistent with previous studies on alkene-
mediated heme modification. When allylbenzene-2-d is used
instead of the unlabeled alkene, the mass spectrum of the
porphyrin dimer has maximum intensity at m/z 751.4 (m/z )
2
+1 per monomer), indicating retention of the H label (data
not shown). With allylbenzene-1,1-d2, the maximum peak
intensity shifts to m/z 752.4, indicating retention of both vinylic
deuterons. Hence, all three vinylic protons of allylbenzene are
retained in the modified hemin.
EPR. The CW-EPR spectrum of native CPO at pH 5.2 in
low-temperature frozen solutions exhibits a rhombic, low-spin
ferric signal with g1,2,3 ) 2.61, 2.26, and 1.84.28 Sono et al.
carried out extensive EPR studies showing that this signal
corresponds to a heme center with a thiolate ligand,29a a result
also suggested by other spectroscopic methods.29b,c The pres-
ence of a proximal cysteine thiolate in CPO was confirmed by
the recent crystal structure.30 In the crystal structure, determined
at room temperature, the iron is displaced by 0.14 Å out of the
heme plane toward the thiolate ligand, and a water molecule is
positioned 3.4 Å from the iron on the distal site of the heme,
beyond bonding distance. This is consistent with the observed
high-spin properties of CPO at room temperature. At lower
temperatures, CPO undergoes a transition to a low-spin six-
coordinate form. 1H ENDOR studies have shown that the low-
spin CPO has a solvent water as a second axial ligand, thus
demonstrating that the hemes of low-spin CPO and P450 share
the same coordination.31 Following inactivation of CPO with
allylbenzene, the EPR signal of native CPO disappears a new
low-spin rhombic EPR signal appears with substantially smaller
g spread (g1,2,3 ) 2.32, 2.16, 1.95).
Figure 1. Definition of the angle φ and of the directions of the
pyrrole-nitrogen hyperfine and quadrupole tensors with respect to an
idealized heme plane, which includes g2 and g3 axes.
As will be seen below, determination of the metrical
parameters that describe the structure of the heme moiety of
AB-CPO requires a knowledge of the orientation of the g-tensor
axes relative to the heme plane. In general, the g1 axis of a
low-spin ferric heme is normal to the heme plane and g2 and g3
axes are in the plane.32 As this has been experimentally verified
for the similar, cysteine-bound heme of P450,31 we may assume
the same for CPO and AB-CPO. This leaves to be determined
the rotation (angle φ) of the g2, g3 axes about the g1 axis, relative
to the heme frame, namely relative to the Fe-N (pyrrole) bonds
(Figure 1). The solution to this question was obtained from
ENDOR studies of [14,15N]AB-CPO.
ENDOR of [14,15N]Allylbenzene-CPO. Single crystal-like
ENDOR spectra of [14,15N]AB-CPO collected at the low-field
(g1) and high-field (g3) edges of the EPR envelope are compared
in Figure 2. In the figure, only ν+ branches are shown; the ν-
branches are expected to fall near 1 MHz and are very difficult
to detect. In the 15N spectra at g1 and g3, four distinct ν+ peaks,
assigned to the four 15N pyrrole nitrogens, are observed. One
nitrogen, denoted Na, has a hyperfine coupling of less than 5
Figure 2. Q-band 14,15N CW-ENDOR spectra taken at the low-field
(g1 ) 2.32) and the high-field (g3 ) 1.95) edges of the EPR envelopes
of the naturally abundant and 15N-enriched AB-CPO samples. Experi-
mental conditions: microwave frequency, 35.00 (15N), 34.97 (14N) GHz;
modulation amplitude, 1.3 G; RF sweep, 0.4 (15N), 0.5 (14N) MHz/s.
14N quadrupole splittings are indicated by a “goal post”, and corre-
sponding 15N’s are connected by solid lines.
MHz; each of the others, denoted Nb-d, has a coupling greater
than 7 MHz (Table 1). The hyperfine values for the second
group are similar to those observed in 15N ENDOR of the
pyrrole nitrogens in native CPO, where the four pyrroles gave
rise to three resolvable 15N ENDOR peaks with hyperfine
couplings of 7.0, 7.9, and 8.7 MHz at g1 (data not shown). The
ca. 40% reduction in the hyperfine coupling to Na upon
allylbenzene inactivation, contrasted with the invariance of the
couplings to Nb-d, does not support a structure in which the
alkene fragment binds only to Fe (5, 6), and instead strongly
suggests one in which the alkene fragment is directly bonded
to Na of the AB-CPO heme (2-4).
(28) Hollenberg, P. F.; Hager, L. P.; Blumberg, W. E.; Peisach, J. J.
Biol. Chem. 1980, 255, 4801-4807.
(29) (a) Sono, M.; Hager, L. P.; Dawson, J. H. Biochim. Biophys. Acta
1991, 1078, 351-359. (b) Dawson, J. H.; Sono, M. Chem. ReV. 1987, 87,
1255-1276. (c) Yarmola, E. G.; Sharonov, Y. A. FEBS Lett. 1994, 35,
279-281.
(30) Sundaramoorthy, M.; Terner, J.; Poulos, T. L. Structure 1995, 3,
1367-1377.
(31) Fann, Y. C.; Gerber, N. C.; Osmulski, P. A.; Hager, L. P.; Sligar,
S. G.; Hoffman, B. M. J. Am. Chem. Soc. 1994, 116, 5989-5990.
(32) Murray, R. I.; Fisher, M. T.; Debrunner, P. G.; Sligar, S. G. In
Metalloproteins. Part 1: Metal Proteins with Redox Roles; Harrison, P.
M., Ed.; Verlag Chemie: Florida, Basel, 1985; Chapter 5, pp 157-206.
1
Comparison of the 14N (I ) 1) and 15N (I ) /2) ENDOR
spectra of AB-CPO in Figure 2 permits an analysis of the

4064 J. Am. Chem. Soc., Vol. 119, No. 17, 1997
Lee et al.
Table 1. 15N Hyperfine Tensors and 14N Nuclear Quadrupole
Tensors of the Heme Nitrogens of Allylbenzene-CPOa
the normal to heme, that is the g1 axis (see Figure 1).34 Hence,
a determination of the orientation of the pyrrole quadrupole
14
15N hyperfine
tensors (MHz)
14N nuclear quadrupole
tensors (MHz)
tensors through an analysis of the field-dependent
N
b-d
ENDOR spectra yields the angle φ by which the g tensor is
rotated about g1.
nucleus A1 A2 A3 φb (deg) P1 P(g3)c φb (deg) P3
P2
d
e
The analysis for a pyrrole 14N begins by taking advantage of
the fact that the quadrupole tensor is traceless, P1 + P2 + P3 )
0, which reduces the number of unknowns from four (three Pi
plus φ) to three. The quadrupole splittings in the ENDOR
spectrum taken at g1 (Figure 2) give 3P1(Nb, Nc, Nd) ) (0.2,
0.5, 0.5 MHz), leaving two unknowns which we take as the
angle φ and the largest principal value P3. The nuclear
quadrupole splitting measured at g3, 3P(g3; Nb, Nc, Nd) ) (1.8,
2.4, and 2.4 Mhz), can then be incorporated into a relationship
between φ and P3 for each of Nb-Nd. It can be written for Nb
as
Na
Nb
Nc
Nd
4.6 4.1 3.7 0-20f
7.6 7.6 7.6 0g
0.43 -1.20
g
g
g
0.06 -0.6 10-20 -0.68 0.62
0.17 -0.8 10-20 -1.12 0.97
0.17 -0.8 10-20 -1.12 0.97
8.7 7.9 7.3 ∼0g
9.0 8.3 7.5 ∼0g
a The tensor principal values (Ai, Pi) are defined in the local tensor
frame (see Figure 1). Uncertainties for Ai, (0.15 MHz, and uncertain-
ties for Pi, (0.1 MHz, are based on uncertainties of (5° in φ. b The
angle is defined in Figure 1. c The measured value along g3. d Calculated
from P1, P(g3), and φ ) 15° through use of eqs 3 and 4 (see text).
e Calculated from P1, P3, and the constraint P1 + P2 + P3 ) 0. f The
fit also suggests that A3 is ∼25° out of the g2, g3 plane (heme plane).
g Not well defined.
quadrupole splittings for the former. The 14N hyperfine
couplings are simply obtained by the ratio of the nuclear g
factors, |A(15N)/A(14N)| ) |g(15N)/g(14N)| ) 1.403, while, as
indicated in the figure, each ν+(15N) feature becomes a
quadrupole-split 14N doublet. As with the hyperfine interaction,
the quadrupole splitting of Na is significantly different from
those of the pyrrole 14Nb-d at g1 (Table 1). This confirms that
the nitrogen, Na, is significantly altered by the addition of
allylbenzene and supports the conclusion that allylbenzene binds
to Na. Among the pyrrole nitrogens, Nb-d, Nc, and Nd have
the same quadrupole coupling constants while Nb has a different
value. Based on this observation, we provisionally assign Nb
as trans to the unique nitrogen, Na.
In a “normal” heme, the orientation of the g tensor can be
deduced from its orientation relative to the hyperfine tensors
of the pyrrole nitrogens, because the latter normally can be taken
to have principal axes that include the Fe-N bond and the out-
of-plane (g1) normal to it.33 To this end, 15N ENDOR spectra
were collected across the EPR envelope, and this 2-D set of
orientation-selective spectra was analyzed as described else-
where.23 For 15Nb,c,d the hyperfine anisotropy is very small and
the signals are strongly overlapped. It was possible to estimate
principal values for the Nb-d tensors (Table 1) by assuming the
coincidence of the hyperfine- and g-tensor axes, namely rotation
angle φ ) 0° of the Fe-N framework relative to the g-tensor
axes; a reliable limit on deviations from zero could not be
determined. The signal for Na is well resolved, however, and
simulation of the field dependence of the 15Na ENDOR spectra
(data not shown) indicates that the tensor axes are rotated by
an angle of φ e 20° about the g1 axis (Table 1, see Figure 1).
This orientation of the g tensor relative to the heme plane is
quite comparable to that directly determined for native P450
by single-crystal EPR measurements where φ ) 18°.32
cos2φ ) [P(g3, Nb) - P1(Nb) - P3(Nb)]/[P1(Nb) + 2P3(Nb)]
(3)
and for Nc and Nd as
sin2 φ ) [P(g3, Nc,d) - P1(Nc,d) - P3(Nc,d)]/
[P1(Nc,d) + 2P3(Nc,d)] (4)
14N ENDOR spectra for Nb-d were calculated for fields across
the EPR envelope of AB-CPO through use of the hyperfine
couplings as scaled from the 15N values, the measured P1, and
for values of φ within the physically significant range, 0° e φ
e 45°; the above relationships were then used to calculate P3.
Comparison of the simulations to experimental spectra (data
not shown) confirmed that the angle between Fe-Na (Na-Nb)
and g3 is less than 20°; in particular, when the angle is taken to
be greater, the total breadth of the ENDOR pattern simulated
for g ) g2 is greater than that seen in the experimental data.
Overall, simulations of the 14N data (not shown) indicate that
the g tensor is rotated about the g1 (heme normal) axis by an
angle of φ ) 15° ( 5° (Figure 1), in good agreement with the
value of φ e 20° estimated from the analysis of the 15Na
hyperfine tensor. This result lays the foundation for analysis
of the following 1,2H ENDOR analysis of the structure of the
heme-bound allylbenzene.
13C ENDOR of Allylbenzene-CPO with 13C-Labeled
Allylbenzene. Well-resolved 13C ENDOR signals not visible
for natural-abundance samples are observed in spectra of both
AB-1-13C-CPO and AB-2-13C-CPO, as seen in data taken at
the low-field edge (g1) of the Ab-CPO EPR spectrum, Figure
3. This confirms that the allylbenzene framework is preserved
in AB-CPO. Somewhat surprisingly, the hyperfine couplings
in these spectra are comparable for the two locations (A(13C-1,
13C-2) ∼ 3 MHz). However, field-dependent 13C ENDOR
studies (data not shown) reveal a slightly larger maximum
component for 13C-2 (∼5 MHz vs ∼ 4 MHz), and while both
tensors are dominated by the isotropic contribution, the dipolar
term is more significant for 13C-2.
The above result was further tested because bonding of
allylbenzene to Na might produce distortions of the heme that
cause the principal axis of the Na hyperfine tensor not to lie
along the metal-nitrogen bond in AB-CPO. The results was
confirmed, and the best estimate for φ was obtained, by
considering the 14N quadrupole interactions for the unperturbed,
b-d, pyrrole nitrogens. The small anisotropy of the 15N
1H ENDOR of Allylbenzene-CPO in H2O and D2O. The
1H ENDOR spectra of AB-CPO in H2O and D2O buffers
collected at the low-field edge (g1) of the EPR envelope are
compared in Figure 4, parts A and B. 1H ENDOR from the
sample in H2O buffer (Figure 4A) shows intensity with A e 5
MHz and one strongly coupled proton with the hyperfine
hyperfine tensors translates to an even smaller anisotropy for
14
14N. As a result, the field dependence of the
N
b-d
ENDOR
spectra is almost completely determined by the quadrupole
interaction. Earlier studies showed that the axis of the maximum
nuclear quadrupole coupling (P3) for a pyrrole 14N lies along
the metal-nitrogen bond and the smallest coupling (P1) is along
(34) (a) Hsieh, Y. N.; Rubenacker, G. V.; Cheng, C. P.; Brown, T. L. J.
Am. Chem. Soc. 1977, 99, 1384-1389. (b) Ashby, C. I. H.; Cheng, C. P.;
Brown, T. L. J. Am. Chem. Soc. 1978, 100, 6057-6063. (c) Ashby, C. I.
H.; Cheng, C. P.; Duelsler, E. N.; Brown, T. L. J. Chem. Soc. 1978, 100,
6063-6067.
(33) (a) Brown, T. G.; Hoffman, B. M. Mol. Phys. 1980, 39, 1073-
1109. (b) Scholes, C. P.; Lapidot, A.; Mascarenhas, R.; Inubushi, T.;
Isaacson, R. A.; Feher, G. J. Am. Chem. Soc. 1982, 104, 2724-2735.

Modified Heme in Allylbenzene-InactiVated CPO
J. Am. Chem. Soc., Vol. 119, No. 17, 1997 4065
Figure 3. Q-band 13C CW-ENDOR spectra taken at the low-field edges
(g1 ) 2.32) of the EPR envelopes of AB-CPO samples with 13C labeled
at the allylbenzene C-1 (A) and C-2 (B) positions and (C) from naturally
abundant allylbenzene. Experimental conditions: microwave fre-
quency, 35.06 (A), 35.06 (B), 35.17 (C) GHz; modulation amplitude,
1.3 G; RF sweep, 0.5 MHz/s. All spectra are centered at 13C Larmor
frequency marked by 1.
coupling of 24 MHz as indicated by the “goal-post” mark. It
should be noted that in Q-band ENDOR one often sees unequal
intensity of the ν+ and ν- branches of a doublet; here ν+ appears
with an opposite phase. Figure 4B shows that this doublet
persists in D2O buffer. The reduced intensity of the ν+ signal
intensity of the 24 MHz coupled doublet is assigned to changes
in nuclear spin relaxation, not to a proton exchange.
1
Figure 4. Q-band H [(A) AB-CPO in H2O; (B) AB-CPO in D2O;
(C) AB-2-d-CPO] and 2H [(D) AB-2-d-CPO] CW-ENDOR spectra
taken at the low-field edge (g1 ) 2.32) of each EPR envelope.
Experimental conditions: microwave frequency, 35.17 (A), 35.42 (B),
35.30 (C, D) GHz; modulation amplitude, 0.7 (A, B, C), 1.3 (D) G;
RF sweep, 2.0 (A, B, C), 0.5 (D) MHz/s. The dotted line in part D
represents 14N resonances, and all spectra are centered at each Larmor
frequency marked by 1. The frequency of the H spectrum (D) is
scaled (×6.51) to match those of H spectra.
1,2H ENDOR of Allylbenzene-1,1-d2-CPO. The presence
of structure 2 and the inferred orientation of the alkylbenzene
is confirmed, and indeed a detailed structure of the heme moiety
in AB-CPO is obtained, by examining the signals from the
hydrogens attached to the C-1 position of allylbenzene. The
2
The absence of the exchangeable protons found in native CPO
implies that the axial HxOx-2 (x ) 1 or 2) ligand of the native
CPO is replaced in AB-CPO.31 The strongly coupled, non-
exchangeable proton (A(1H) ) 24 MHz) might arise either from
â-protons of the cysteinyl thiolate proximal ligand or from a
site on the allylbenzene substrate. To resolve this issue, as well
as to probe structural details of the AB-modified hemin, we
synthesized allylbenzene (7) specifically labeled with 2H in the
two vinylic positions (allylbenzene-1,1-d2, allylbenzene-2-d) and
prepared ENDOR samples from CPO inactivated with each of
the labeled alkenes.
1
1H (C-1) signals are best seen in the difference H ENDOR
spectra obtained by subtracting the H ENDOR of AB-1,1-d2-
CPO from those of AB-CPO, Figure 5A. In the figure, the
dashed lines represent one proton (denoted H1a); as will be
explained, the solid lines represent the other proton (denoted
1
1
1
2
1,2H ENDOR of Allylbenzene-2-d-CPO. The H and H
1
ENDOR spectra of AB-2-d-CPO at g1 are displayed in Figure
4, parts C and D. The H signal with the hyperfine coupling
H1b). The pattern of the H1a resonances that emerge in the
1
2-D display of spectra taken across the EPR envelope is typical
for a proton whose hyperfine coupling is dominated by dipolar
coupling to the spin on the ferric ion. As illustrated in Figure
5B, the pattern is well simulated by an almost purely dipolar
tensor, with components A ) [-2.8, -2.8, 5.0] MHz ) -0.2
+ [-2.6, -2.6, 5.2] MHz; here, the signs are fixed by the fact
that the interaction is governed by the through-space dipolar
coupling, the isotropic coupling is a ) -0.2 MHz, and the
[FeIII-H1a] distance within the point-dipole approximation is d
) 3.1 Å. The Fe-H1a vector corresponds to the unique axis
of the 1H1a tensor, A3, and the orientation of this axis with respect
to g-tensor frame (Euler angle: φ(H1a) ) 7°; θ(H1a) ) 56°)
corresponds to the orientation of the proton in the g-tensor frame.
The results of the 14,15N ENDOR measurements then can be
used to orient H1a relative to the heme plane as shown in Figure
6. This figure has assumed a particular resolution of the one
ambiguity in describing the projection of the FeIII f H1a vector,
d(H1a), onto the heme plane, as is better illustrated in Figure
6B. The ambiguity is that the g tensor is rotated relative to the
heme frame by φ(g) ) 15° about g1, and the A(H1a) tensor is
of 24 MHz (Figure 4A,B) disappears for AB-2-d-CPO (Figure
4C). The corresponding 2H ENDOR signal (Figure 4D) appears
as a doublet split by the hyperfine coupling of A(2H) ) 3.8
MHz as expected from the Larmor ratio, |ν(1H)/ν(2H)| ) |A(1H)/
A(2H)| ) 6.51. This result confirms that the vinylic proton of
allylbenzene is present in the modified hemin in situ.
1
Analysis of the full field-dependent H ENDOR pattern of
AB-CPO across the EPR envelope reveals that the hyperfine
coupling of the proton on the C-2 position (denoted H2a) of
AB-CPO is mostly isotropic, with tensor values of A ) (24.6,
17.0, 16.2) MHz. A glance at the spectrum shows that any
signal from a proton on the C-1 position must fall within the
central feature, and cannot have a 1H hyperfine splitting greater
than ca. 5-6 MHz at any field. This sharp disparity in the
couplings to the two types of protons implies that the heme of
AB-CPO must adopt a structure in which C-2 retains its bond
to the hydrogen and is closer to Fe(III) than C-1. Of structures
2-6, only 2 satisfies this description, provided R ) CH2Ph and
R′ ) H.

4066 J. Am. Chem. Soc., Vol. 119, No. 17, 1997
Lee et al.
Figure 5. (A) 1H ENDOR spectra calculated by subtracting the Q-band 1H CW-ENDOR of AB-1,1-d2-CPO from those of AB-CPO. Experimental
conditions: microwave frequency, 35.04 GHz; modulation amplitude, 1.3 G; RF sweep, 0.5 MHz/s. The dotted lines indicate the proton denoted
as H1a while the solid lines indicate the proton denoted as H1b (see text). (B) 1H ENDOR simulation with A ) [-2.8, -2.8, 5.0] MHz, Euler angles
of θ ) 56° and φ ) 7°, and ENDOR line width of 0.2 MHz. Here, A is the hyperfine tensor. All the spectra are centered at 1H Larmor frequency.
rotated relative to the g frame by φ(H1a) ) 7° about g1, but the
relative signs of these angles is not fixed. Thus, the angle
between the projection of d(H1a) and the FeIII-Na vector can
be either 8° or 22°, depending on whether the two φ angles
have the opposite or same signs. For reasons that will become
clear, Figure 6 assumes they have opposite signs.
Figure 5 showed one proton (H1a) on the C-1 position of the
allylbenzene adduct. The mass spectrum indicates that both
C-1 protons are present, but signals from the second one (H1b)
1
are not evident in these H CW-ENDOR spectra. The second
hydrogen, H1b, instead was detected in 2H Q-band Mims
ENDOR spectra of allylbenzene-1,1-d2-CPO (Figure 7). The
2H measurement at Q-band has no background and is specially
useful for detecting weak hyperfine couplings. The experi-
mental data represented by thin lines are the 2H ENDOR signal
of the deuteron (D1a) corresponding to H1a. They show ν+ and
ν- branches at each field separated by ∼0.45 MHz, as expected
by the Larmor ratio |ν(1H)/ν(2H)| ) |A(1H)/A(2H)| ) 6.51; each
branch is further split by the nuclear quadrupole interaction.
2
The simulation for D1a is seen as solid lines where the H
1
hyperfine tensor is determined from H tensor and the nuclear
g factors. The features depicted by thicker solid lines in Figure
7 are not predicted for the deuteron, D1a, and are from the second
deuteron, D1b, in AB-1,1-d2-CPO. The simulations for D1b, with
the hyperfine coupling of A ) [-0.29, -0.29, 0.49] MHz and
the Euler angle ) [36°, 71°, 0°] with respect to the g-tensor
frame, are shown as dashed lines. The corresponding H1b
hyperfine tensor is A ) [-1.9, -1.9, 3.2] MHz. The isotropic
Figure 6. (A) ENDOR-determined orientation of Na+, H1a, and H1b
(see text). The ENDOR measurement determines |φ|. Of the two
possibilities, one orientation is chosen for each with the aid of the MM2
calculation. (B) ENDOR (solid arrowhead) and MM2 (open arrowhead)
determined orientations of g-tensor axes and C-1 protons as projected
on the heme plane containing Fe(III) and pyrrole nitrogens.
1
coupling constant of H1b is a ) -0.2 MHz, the same as for
1H1a, within the point-dipole approximation the [FeIII-H1b]
distance is d ) 3.6 Å. From the hyperfine coupling and Euler
angle determined by 2H ENDOR of D1b, H1b can be positioned
in the heme plane as well as the g-tensor frame as in Figure 6,

Modified Heme in Allylbenzene-InactiVated CPO
J. Am. Chem. Soc., Vol. 119, No. 17, 1997 4067
Figure 8. Structure of AB-CPO (structure 8) calculated by MM2
energy minimization.
indicating close proximity to Fe. Those for H1a and H1b (C-1
position) are weak and overwhelmingly dipolar, and place these
hydrogens above the nitrogen, Na (Figure 6). These observa-
tions select structure 2 with R ) CH2Ph and R′ ) H. The
absence of solvent exchangeable protons in the electrospray
mass spectra and the 1H ENDOR further excludes the structures
3 and 4.36 Thus, we conclude that the alkene adds to the heme
so as to place the 2-carbon closest to the iron, generating a
“primary” N-alkylhemin that we now show to adopt struc-
ture 8.
Figure 7. Q-band 2H Mims ENDOR spectra of AB-1,1-d2-CPO.
Experimental Conditions: τ, 1300 ns; repetition rate, 50 Hz; RF pulse
width, 60 µs; average, 580 transients at g ) 2.18, 480 transients at g
) 2.16. Experimental data represented by a thicker line indicate D1b
while a thinner line represents D1a. The simulation displayed by a solid
line is for D1a and that represented by a dotted line is for D1b.
Simulation parameters for D1a are A ) [-0.43, -0.43, 0.77] MHz, θA
) 56°, φA ) 7°, P ) [-0.0275, -0.0275, -0.0550] MHz, θP ) 76°,
φp ) 69°; for D1b A ) [-0.29, -0.29, 0.49] MHz, θA ) 71°, φA
)
36°, P ) [-0.0275, -0.0275, -0.0550] MHz, θP ) 77°, φp ) 38°.
Here, A is the hyperfine tensor and P is the nuclear quadrupole tensor.
again with a particular resolution of an ambiguity analogous to
the one for H1a.
Structure of the Allylbenzene-Modified Hemin. The
results of the ENDOR studies and mass analysis are sufficient
to distinguish between structures 2-6 for the hemin present in
AB-CPO. First, the observation of a ferric EPR signal is
sufficient to rule out the carbene 5. Although stable iron(II)
carbenes (PFeIVdCRR′ T PFeII r CRR′) are known, both in
enzyme and model hemin systems,35 one-electron oxidation of
these carbenes leads to the formation of an Fe,N bridged ferric
species similar to 4,8 not a stable iron(III) carbene. The heme
of AB-CPO thus cannot be formulated as a â-hydroxycarbene.
Second, 14,15N ENDOR data show that one of the four pyrrole
nitrogens of AB-CPO has distinctive hyperfine and nuclear
quadrupole couplings. These suggests modification of the heme
at one of the pyrrole rings, consistent with structures 2-4 but
not the carbene 5 or the σ-alkyl species 6.
The absence of an ENDOR signal from strongly-coupled
exchangeable proton(s) argues against the presence of an axial
HxO. The strong coupling to H2a could not arise solely from
spin polarization through the Na-C1 linkage, and thus argues
for the presence of the metallocycle 8, formed by the Fe-O
bond of the bound alkoxide. Nevertheless, the ENDOR
measurements have not directly detected the axial alkoxide
ligand, and thus the presence of the metallocycle, and in any
case one wishes for greater structural detail. To both test and
refine structure 8, we have compared the ENDOR data with a
MM2 energy minimization of this structure.37 The lowest
energy conformation of structure 8 is shown in Figure 8. In
the absence of further information, we modeled the allylbenzene
heme adduct by assuming that the oxoferryl group of compound
I (Scheme 1) adds to the reface of the alkene.
The N-alkylated structures 2-4 are distinguished by the 1,2
H
Figure 9 expands the region of Figure 8 that contains the
AB-heme metallocycle and includes nitrogen, Na; it displays
both the metrical parameters given by ENDOR and those taken
from the MM2 refined structure 8. In general, the parameters
from the experiment and the refinement are in exceptionally
good agreement. The higher quality agreement for H1a,
ENDOR of AB-CPO and deuterated AB-CPO. The experiments
reveal different types of proton hyperfine interaction for protons
on the C-1 position and C-2 position of allylbenzene. The
coupling for H2a (C-2 position) is strong and mostly isotropic,
(35) (a) Mansuy, D.; Lange, M.; Chottard, J.-C.; Guerin, P.; Morliere,
P.; Brault, D.; Rougee, M. J. Chem. Soc., Chem. Commun. 1977, 648-
649. (b) Mansuy, D.; Lange, M.; Chottard, J. C.; Bartoli, J. F.; Chevrier,
B.; Weiss, R. Angew. Chem., Int. Ed. Engl. 1978, 17, 781-782. (c) Mansuy,
D.; Battioni, J.-P.; Chottard, J.-C.; Ullrich, V. J. Am. Chem. Soc. 1979,
101, 3971-3973 and references therein. (d) Mansuy, D. Pure Appl. Chem.
1980, 52, 681-690. (e) Balch, A. L.; Chan, Y. W.; Olmstead, M. M.;
Renner, M. W. J. Org. Chem. 1986, 51, 4651-4656. (f) Mansuy, D.;
Battioni, J.-P.; Lavallee, D. K.; Fischer, J.; Weiss, R. Inorg. Chem. 1988,
27, 1052-1056.
(36) The formation of hemin dimers in the MS samples shows that the
heme becomes separated from the apoprotein and is not sequestered from,
and thereby prevented from exchanging with, the aqueous medium.
(37) An analogous approach of comparing results from ENDOR and
molecular modeling has been used to study the structure of Rieske [2Fe-
2S] cluster: Gurbiel, R. J.; Doan, P. E.; Gassner, G. T.; Macke, T. J.; Case,
D. A.; Ohnishi, T.; Fee, J. A.; Ballou, D. P.; Hoffman, B. M. Biochemistry
1996, 35, 7834-7845.

4068 J. Am. Chem. Soc., Vol. 119, No. 17, 1997
Lee et al.
reaction with terminal alkenes. Formation of secondary N-
alkylhemins has not been demonstrated in any enzyme system
to date, nor do disubstituted alkenes inactivate P450 or CPO.
In tetraarylhemins where the aryl groups lack bulky ortho
substituents, secondary N-alkylhemins may be formed as major
products with terminal alkenes.5c Secondary N-alkylhemins may
also be formed in model systems as adducts with disubstituted
alkenes.5d,6b,c We have recently obtained evidence for steric
control in the partitioning between N-alkylhemin formation and
epoxidation during oxidation of terminal alkenes by CPO and
have shown that with appropriate alkene substrates, N-alkylation
is an order of magnitude more facile in CPO than previously
reported in any other heme system.3a The alkylation of a
specific pyrrole nitrogen indicated in this study, together with
demonstrations of regioselective heme alkylation in P450,4a
suggests that steric factors control the site of pyrrole alkylation
in enzyme systems, as well as the orientation of alkene addition
to the heme face. Based on the substrate selectivity and partition
numbers,3a heme N-alkylation in CPO appears to be under
greater steric control than in model systems,5g possibly as a result
of heme occlusion by the protein structure.
The results obtained in this study may comment to a limited
extent on the mechanism of CPO-catalyzed epoxidation. In
P450, the formation of organometallic porphyrin species during
alkene epoxidation was originally proposed to explain exchange
of the vinylic protons of propene with the aqueous medium
under oxidizing conditions,9a although an alternate explanation
for this result has since been offered.9b We do not observe
exchange of the vinylic protons of allylbenzene with water in
heme-alkylated CPO and can thus exclude species with ex-
changeable protons as long-lived precursors to 8. This rules
out Fe,N bridged species such as 3 and 4, as well as the carbene
5 and σ-vinylhemin 6, as precursor species partitioning between
epoxidation and heme N-alkylation. It does not, however,
exclude other proposed intermediates such as the four-membered
metallocycle 9,39 the cation 10,6b or the radical 11.40
Figure 9. Comparison of structures determined by ENDOR and MM2.
b and y are the projections of H1a and H1b onto the heme plane. The
(g1-FeIII-H1) MM2 angles are obtained by the assumption that the g1
axis is identical to the FeIII-O bond.
compared to H1b, reflects the fact that 2H ENDOR data for H1b
are available only at a few field positions, and thus the hyperfine
tensor is not as well constrained. Overall, the MM2 calculation
both confirms and extends the ENDOR analysis of the structure
of the heme of AB-CPO. The agreement of the measured and
calculated Fe-H1 distances confirms that inactivation occurs
by regioselective bonding of C-1 to the pyrrole nitrogen. The
agreement between the ENDOR and MM2 structures, both
distances and angles, coupled with the ENDOR evidence for a
large coupling to H2a and against an alternate proximal ligand
such as water or hydroxide in our view is conclusive confirma-
tion that the alkoxide oxygen ligates to the heme iron to give
the metallocycle of structure 8 as visualized in Figure 8.38
Discussion
In the present work, we have investigated for the first time
the in situ structure of an alkene-modified enzymatic heme.
While the results have confirmed earlier conclusions regarding
the occurrence of structure 2, they provide a detailed view of
the regiochemistry of alkene addition and indeed metrical
parameters, as well as serving as a foundation for further studies
of this and related systems. In particular, determination that 8
represents the heme structure of AB-CPO is an important
prerequisite for understanding the catalytic mechanism of CPO,
the mechanism of spontaneous heme dealkylation and restoration
of catalytic activity in this system,7 as well as the susceptibility
of the enzyme to inactivation with different olefinic substrates.3a
The ENDOR-based structural assignment may also serve as a
basis for studies of the modified enzyme using other spectro-
scopic techniques.
We find a radical species such as 11 an attractive proposal
for an intermediate partitioning between heme N-alkylation and
epoxidation, in part because of the specific susceptibility of iron
porphyrins to the “suicide” reaction. To date, inactivation and
pyrrole alkylation by alkenes has not been reported for Mn or
Cr porphyrins during epoxidation, although these metallopor-
phyrins are well-studied epoxidation catalysts. It is tempting
to propose that hemin N-alkylation is a specific consequence
of the structure of the two-electron oxidized iron porphyrin
species which reacts with alkenes. In iron porphyrins, this
species is understood to be a Fe(IV) porphyrin cation radical,41
while the corresponding two-electron oxidized species in Mn
The orientation of allylbenzene addition to the heme in CPO
is consistent with results obtained with terminal alkenes in P4504
and in model tetraarylhemins containing bulky o-aryl substit-
uents.5,6 In these sterically hindered hemins, “secondary”
N-alkylated hemins6c are minor and unstable products of the
(39) (a) Collman, J. P.; Brauman, J. I.; Meunier, B.; Hayashi, T.;
Kodadek, T.; Raybuck, S. A. J. Am. Chem. Soc. 1985, 107, 2000-2005.
(b) Collman, J. P.; Kodadek, T.; Raybuck, S. A.; Brauman, J. I.; Papazian,
L. M. J. Am. Chem. Soc. 1985, 107, 4343-4345. (c) Collman, J. P.;
Kodadek, T.; Brauman, J. I. J. Am. Chem. Soc. 1986, 108, 2588-2594.
(40) Luke, B. T.; Collins, J. R.; Loew, G. H.; McLean, A. D. J. Am.
Chem. Soc. 1990, 112, 8686-8691.
(38) This study provides no direct evidence about bonding between the
cysteine thiolate ligand and CPO heme in AB-CPO. To address this, we
plan to carry out 2H ENDOR studies of AB-CPO labeled with â-2H2-
cysteine.

Modified Heme in Allylbenzene-InactiVated CPO
J. Am. Chem. Soc., Vol. 119, No. 17, 1997 4069
Scheme 2
CPO to the native enzyme is 5.8 h at 25 °C.7 In model systems,
but not enzymatic systems, the N-alkylated metallocycle is still
an epoxidation catalyst, although it has reduced catalytic
activity.5f,6 Synthetic N-alkylhemins have been shown directly
to be capable of forming the two-electron oxidized species which
act as oxygen donors in epoxidation.45 The inability of alkene-
modified CPO to similarly form oxidized species and catalyze
substrate oxidation3a,7 might possibly be attributed to inacces-
sibility to oxidant of the second heme axial site in the intact
enzyme. In contrast, five-coordinate model N-alkylhemins can
function as oxygen donors on the remaining free heme face.
The 14N and 15N ENDOR of allylbenzene-inactivated CPO
indicates that the pyrrole alkylation is highly regioselective. A
single set of heme pyrrole signals (comprising three similar
nitrogens and one greatly shifted nitrogen) is observed in the
ENDOR samples, without detectable contributions from prod-
ucts in which a different pyrrole nitrogen has been alkylated.
Although we cannot exclude the possibility that small amounts
of alternate alkylation products are present in the intact enzyme,
the ENDOR results show that pyrrole alkylation occurs largely
at one site, consistent with results obtained in P450.4a Studies
are now underway to determine the specific site of alkylation
in the isolated N-alkylporphyrin, as well as the enantioselectivity
at C-2 of heme adduct formation.
and Cr systems are metal(V) porphyrins.42 Heme N-alkylation
might be considered as bond formation between an alkene-
derived free radical and the porphyrin ring radical in the oxidized
hemin (Scheme 2). Direct evidence has recently been obtained
for a radical species as an intermediate in epoxidation.43 While
the ENDOR and labeling studies described here would also be
compatible with concerted epoxidation and heme N-alkylation
mechanisms,44 it is difficult to see how this mechanism accounts
for the restriction of the pyrrole N-alkylation reaction to iron
porphyrins.
In summary, these first detailed spectroscopic and structural
studies of an alkene-inactivated heme enzyme show that the
heme of AB-CPO has structure 8 (Figure 8). It is hoped that
these results may stimulate similar studies of alkene inactivation
of P450 as the sister enzyme of CPO.
Model metallocycles of type 2 return to the native hemin on
standing, but do not revert at a sufficient rate to be an
intermediate in the epoxidation pathway.6b,c We have obtained
similar results in CPO, where the half-life for reversion of AB-
Acknowledgment. This work was supported by the National
Institutes of Health (HL 13531 (B.M.H.) and GM 07768
(L.P.H.)). A.F.D. is the recipient of a Predoctoral Fellowship
in Biological Sciences from the Howard Hughes Medical
Institute. Positive ion electrospray mass spectra were recorded
in the Mass Spectrometry Laboratory, School of Chemical
Sciences, University of Illinois, using a Fisons VG Quattro
quadrupole-hexapole-quadrupole mass spectrometer purchased
in part with a grant from the Division of Research Resources,
NIH (RR 07141).
(41) (a) Dolphin, D.; Forman, A.; Borg, D. C.; Fajer, J.; Felton, R. H.
Proc. Natl. Acad. Sci. U.S.A. 1971, 68, 614-618. (b) Schulz, C. E.;
Devaney, P. W.; Winkler, H.; Debrunner, P. G.; Doan, N.; Chiang, R.;
Rutter, R.; Hager, L. P. FEBS Lett. 1979, 103, 102-105. (c) Roberts, J.
E.; Hoffman, B. M.; Rutter, R.; Hager, L. P. J. Biol. Chem. 1981, 256,
2118-2121. (d) Roberts, J. E.; Hoffman, B. M.; Rutter, R.; Hager, L. P.
J. Am. Chem. Soc. 1981, 103, 7654-7656.
(42) (a) Lee, W. A.; Bruice, T. C. Inorg. Chem. 1986, 25, 131-135.
(b) Ostovic, D.; Bruice, T. C. Acc. Chem. Res. 1992, 25, 314-320.
(43) (a) Groves, J. T.; Watanabe, Y. J. Am. Chem. Soc. 1986, 108, 507-
508. (b) Gross, Z.; Nimri, S. J. Am. Chem. Soc. 1995, 117, 8021-8022.
(44) In principle, the degree of stereochemical retention in the heme
adduct could be determined by use of allylbenzene that has been stereospe-
cifically labeled with one deuterium at C-1, and would comment on a
stepwise Vs concerted mechanism of adduct formation.
JA963684C
(45) Balch, A. L.; Cornman, C. R.; Latos-Grazynski, L.; Renner, M. W.
J. Am. Chem. Soc. 1992, 114, 2230-2237.