Models of nitric oxide synthase: Iron(III) porphyrin-catalyzed oxidation of fluorenone oxime to nitric oxide and fluorenone

Article abstract of DOI:10.1021/ja992373+

Nitric oxide synthase (NOS) is a heme-containing monoxygenase that catalyzes the oxidation of L-arginine to L-citrulline and NO in two steps. In the second step of the NOS reaction, citrulline and NO are generated from the heme-catalyzed 3-electron oxidation of L-N-hydroxyarginine. To model this unusual reaction, iron porphyrin-catalyzed oxygenations of oximes with O₂ were investigated. The oxidation of fluorenone oxime and a stoichiometric amount of hydroxoiron(III) porphyrin (Fe(OH)P, P = TMP and TPFPP) with O₂ in benzene generated Fe(NO)P, fluorenone, and O-(9-nitro-9-fluorenyl)fluorenone oxime. The X-ray crystal structure of the oxime ether product suggests that it originated from the dimerization of the fluorenyl iminoxy radicals. Detailed analysis of this reaction showed that the oxime reacted first with Fe(OH)P to generate a 5-coordinate, high-spin oximatoiron(III) porphyrin species [Fe(oximate)P]. The X-ray crystal structure of oximatoiron(III) tetrakis(2,6-dichlorophenyl)porphyrin [Fe(oximate)TDCPP] showed that the oximate ligand was monodentate, O-bound to Fe(III)P. The aerobic oxidation of Fe(oximate)P followed the characteristic kinetics of a metalloporphyrin- catalyzed radical-type autoxidation. O₂ surrogates, the π-acids NO and CO, induced the homolysis of Fe(oximate)P to generate Fe(NO)P or Fe(CO)P and the iminoxy radical, implicating a similar reaction mode for O₂ with Fe(oximate)P. Fe(oximate)TMP reacted with ¹⁸O₂ to generate predominantly ¹⁸O-labeled fluorenone (75% yield), while the reaction conducted under ¹⁶O₂ and H₂¹⁸O generated only ¹⁶O-labeled fluorenone. This reaction is proposed to proceed via an Fe-O bond homolysis of Fe(oximate)TMP followed by O₂ insertion to generate 9-nitroso-9-fluorenylperoxyFe(III)TMP, which decomposes via an O-O bond homolysis to generate NO, fluorenone, and oxoFe(IV)P. The implications of this system for the NOS reaction mechanism are discussed.

Full text of DOI:10.1021/ja992373+

12094
J. Am. Chem. Soc. 1999, 121, 12094-12103
Models of Nitric Oxide Synthase: Iron(III) Porphyrin-Catalyzed
Oxidation of Fluorenone Oxime to Nitric Oxide and Fluorenone
Charles C.-Y. Wang, Douglas M. Ho, and John T. Groves*
Contribution from the Department of Chemistry, Princeton UniVersity, Princeton, New Jersey 08544
ReceiVed July 8, 1999
Abstract: Nitric oxide synthase (NOS) is a heme-containing monoxygenase that catalyzes the oxidation of
L-arginine to L-citrulline and NO in two steps. In the second step of the NOS reaction, citrulline and NO are
generated from the heme-catalyzed 3-electron oxidation of L-N-hydroxyarginine. To model this unusual reaction,
iron porphyrin-catalyzed oxygenations of oximes with O2 were investigated. The oxidation of fluorenone oxime
and a stoichiometric amount of hydroxoiron(III) porphyrin (Fe(OH)P, P ) TMP and TPFPP) with O2 in benzene
generated Fe(NO)P, fluorenone, and O-(9-nitro-9-fluorenyl)fluorenone oxime. The X-ray crystal structure of
the oxime ether product suggests that it originated from the dimerization of the fluorenyl iminoxy radicals.
Detailed analysis of this reaction showed that the oxime reacted first with Fe(OH)P to generate a 5-coordinate,
high-spin oximatoiron(III) porphyrin species [Fe(oximate)P]. The X-ray crystal structure of oximatoiron(III)
tetrakis(2,6-dichlorophenyl)porphyrin [Fe(oximate)TDCPP] showed that the oximate ligand was monodentate,
O-bound to Fe(III)P. The aerobic oxidation of Fe(oximate)P followed the characteristic kinetics of a
metalloporphyrin-catalyzed radical-type autoxidation. O2 surrogates, the π-acids NO and CO, induced the
homolysis of Fe(oximate)P to generate Fe(NO)P or Fe(CO)P and the iminoxy radical, implicating a similar
reaction mode for O2 with Fe(oximate)P. Fe(oximate)TMP reacted with 18O2 to generate predominantly 18O-
labeled fluorenone (75% yield), while the reaction conducted under 16O2 and H218O generated only 16O-labeled
fluorenone. This reaction is proposed to proceed via an Fe-O bond homolysis of Fe(oximate)TMP followed
by O2 insertion to generate 9-nitroso-9-fluorenylperoxyFe(III)TMP, which decomposes via an O-O bond
homolysis to generate NO, fluorenone, and oxoFe(IV)P. The implications of this system for the NOS reaction
mechanism are discussed.
Introduction
Scheme 1
Nitric oxide (NO) is a ubiquitous biomessenger1,2 with a
variety of functions including blood vessel dilation,3 neuronal
signal transmission,4 cytotoxicity against pathogens and tumors,5
and cellular respiration activity.6 The over- or underproduction
of NO has also been implicated in a number of pathological
symptoms such as endotoxic shock,7 diabetes,8 allograft rejec-
tion,9 and myocardial ischimia/reperfusion injury.10
NO is generated by the oxidation of L-arginine by O2
mediated by nitric oxide synthase (NOS; EC 1.14.13.39). Four
isoforms of NOS have been discovered: a constitutive, cal-
modulin-dependent neuronal NOS (nNOS), endothelial NOS
(eNOS), mitochondrial NOS (mtNOS),11,12 and the calmodulin-
independent, cytokine-inducible NOS (iNOS). All NOS isoforms
contain the cysteinyl thiolate-ligated Fe(III) heme feature13-16
and extensive homology of their reductase domains to cyto-
chrome P450 reductase.17,18
NOS catalyzes the 5-electron oxidation of arginine to
citrulline and NO in two steps19 (Scheme 1). In the first step,
two NADPH-derived reducing equivalents and O2 afford
N-hydroxyarginine (NHA). The redox stoichiometry is that of
a typical P450-like hydroxylation. The second step of the NOS
reaction involves a 3-electron oxidation of N-hydroxyarginine
* To whom correspondence should be addressed
(1) Dangl, J. Nature 1998, 394, 525.
(2) Muller, U. Prog. Neurobiol. 1997, 51, 363.
(3) Hansen, M. B.; Dresner, L. S.; Wait, R. B. Physicol. Res. 1998, 47,
307.
(4) Agnihotri, N.; Lopez-Garcia, J. C.; Hawkins, R. D.; Arancio, O. Histol
Histopathol 1998, 13, 1155.
(5) Albina, J. E.; Reichner, J. S. Cancer Metastasis ReV. 1998, 17, 39.
(6) Brown, G. C. Biochim. Biophys. Acta: Bioenerget. 1999, 1411, 351.
(7) Shah, N. S.; Billiar, T. R. EnViron. Health Perspect. 1998, 106, 1139.
(8) Cosentino, F.; Luscher, T. F. J. CardioVasc. Pharmacol. 1998, 32,
s54.
(9) Liu, Z. Q.; Wildhirt, S. M.; Weismuller, S.; Schulze, C.; Conrad,
N.; Reichart, B. Atherosclerosis 1998, 140, 1.
(10) Hammerman, C. K. M. Clin. Perinatol. 1998, 25, 757.
(11) Ghafourifar, P.; Richter, C. FEBS Lett. 1997, 418, 291.
(12) Tatoyan, A.; Giulivi, C. J. Biol. Chem. 1998, 273, 11044.
(13) Mcmillan, K.; Bredt, D. S.; Hirsch, D. J.; Snyder, S. H.; Clark, J.
E.; Master, B. S. S. Proc. Natl. Acad. Sci. U.S.A. 1992, 84, 6369.
(14) Klatt, P.; Schmidt, K.; Mayer, B. Biochem. J. 1992, 288, 15.
(15) Stuehr, D. J.; Ikeda-Saito, M. J. Biol. Chem. 1992, 267, 20547.
(16) White, K. A.; Marletta, M. A. Biochemistry 1992, 31, 6627.
(17) Bredt, D. S.; Hwang, P. M.; Glatt, C. E.; Lowenstein, C.; Reed, R.
R.; Snyder, S. H. Nature 1991, 351, 714.
(18) Xie, Q. W.; Cho, H. J.; Calycay, J.; Mumford, R. A.; Swiderek, K.
M. Science 1992, 256, 225.
(19) Hemmens, B.; Mayer, B. In Nitric Oxide Protocols; Titheradge,
M. A., Ed; Humana: Totowa, NJ , 1998; Vol. 100, p 1.
10.1021/ja992373+ CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/15/1999

Models of Nitric Oxide Synthase
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 12095
Table 1. Summary of X-ray Crystallographic Data of
O-(9-Nitro-9-fluorenyl)fluorenone Oxime (1) and
Fe(oximate)TDCPP (3)
to NO and citrulline by O2 with the consumption of only one
NADPH-derived reducing equivalent.20,21 18O-labeling experi-
ments have shown that the urea oxygen of citrulline is derived
from O2, while the NO product originates exclusively from the
N-hydroxy group of NHA.22,23 Nevertheless, the mechanistic
details of this unusual second step of the NOS reaction
mechanism are still unclear.
A key unanswered question is the order of the redox events
involved in the oxidation of NHA. Particularly, it is unclear
whether the NOS Fe(III) heme is reduced by NHA or the
NADPH-derived reducing equivalent to initiate the second step.
Several proposals have been advanced to account for these
unusual redox events. They include an oxoFe(IV) porphyrin
radical cation active species reminiscent of the cytochrome P450
monooxygenase-type mechanism,24,25 and versions of a nucleo-
philic peroxoFe(III)P species analogous to that proposed for the
aromatase reaction.20,26-29
An understanding of the detailed reaction mechanism of the
NOS reaction is not only of intrinsic interest but is also important
for the rational design of selective NOS inhibitors.30,31 Chemical
reactions that mimic the important features of the NHA
oxidation could be powerful tools for deciphering its mechanism.
We have discovered that the Fe(III)P-catalyzed aerobic
oxidation of oximes produces NO and ketones. This reaction
was initiated by coordination of the oxime to the Fe(III)
porphyrin. Further, reactions conducted under 18O2 generated
18O-labeled product distribution patterns similiar to those of the
NOS reaction.
1
3
formula
C26H16N2O3
C57H28Cl8FeN5O‚
C6H5Cl‚0.5C7H16
1300.94
dark purple/prism
triclinic
FW
404.41
color/habit
cryst system
space group
a, Å
b, Å
c, Å
R, deg
â, deg
ø, deg
V, Å3
pale yellow/irregular
monoclinic
P21/n
9.8776(6)
11.9206(6)
17.4262(12)
90.00
101.550(3)
90.00
2010.3(2)
4
P1h
12.5029(2)
13.0424(4)
18.9793(5)
81.701(1)
88.304(2)
84.263(2)
3046.80(13)
2
Z
T, °C
F
25(2)
1.336
0.089
4.16-44.96
26 402
2622
1948
416
0.057
0.141
25(2)
1.418
0.690
3.58-54.92
79 647
13 775
7272
649
0.057
0.14
calcd, g cm-3
µ(Mo KR), mm-1
2θ range, deg
total no. of data
no. of unique data
no. of observed dataa
no. of parameters
Rb
wRc
b
2
a With I > 2σ(I). R ) ∑|Fo - Fc|/∑Fo. c wR ) [∑w(Fo - Fc2)2/
∑w(Fo )2]1/2
.
2
Scheme 2
Results and Discussion
Fluorenone oxime reacted with a stoichiometric amount of
hydroxoiron(III) porphyrins (Fe(OH)P, P ) TMP and TPFPP)
under O2 to generate nitrosyliron(II) porphyrin (Fe(NO)P),
fluorenone, and a novel dimer derivative (1; see Table 1) of
fluorenyl iminoxy radical (Scheme 2). For example, a one-to-
one ratio of fluorenone oxime and Fe(OH)TMP in benzene-d6
(0.75 mM) was oxidized with greater than 95% conversion after
exposure to 500 psi O2 for 9 h. The 1H NMR spectrum of this
reaction mixture showed that the iron porphyrin was converted
to Fe(OH)TMP, Fe(NO)TMP, Fe(NO3)TMP,32 and a new
5-coordinate high-spin Fe(III)TMP species, which we tentatively
assigned as a nitritoFe(III) complex (Fe(ONO)TMP).33 The
product distribution of the Fe(III)TMP-mediated aerobic oxida-
tion of the oxime was found to be dependent on the O2 pressure.
As shown in Table 2, the higher the headspace O2 pressure, the
higher the yields of fluorenone, Fe(NO3)TMP, and the putative
Fe(ONO)TMP. Concomitantly, the yields of the fluorene-based
dimer and Fe(NO)TMP decreased as the headspace O2 pressure
was raised. Fluorenone was always the major (>80% yield)
organic product, and 1 was the only other organic product
(<20% yield) of the reaction.
Coordination of Fluorenone Oxime to Fe(III)P. To eluci-
date the mechanism of this process, we have dissected and
characterized the reaction in detail. The reaction of a stoichio-
metric amount of fluorenone oxime and Fe(OH)P in benzene
was found to generate a new 5-coordinate, high-spin Fe(III)P
species. The 1H NMR spectrum of this reaction mixture showed
that, immediately upon mixing, the resonance signals of Fe-
(OH)TMP (δ80, pyrrole-H; δ 11 and 12, m-phenyl H) changed
(20) Korth, H. G.; Sustmann, R.; Thater, C.; Butler, A. R.; Ingold, K.
U. J. Biol. Chem. 1994, 269, 17776.
(21) Hevel, J. M.; Marletta, M. A. In AdVances in experimental medicine
and biology; Ayling, J. E., Nair, M. G., Baugh, C. M., Eds; Plenum Press:
New York, 1993; Vol. 338, p 285.
(22) Leone, A. M.; Palmer, R. M. J.; Knowles, R. G.; Francis, P. L.;
Ashton, D. S.; Moncada, S. J. Biol. Chem. 1991, 266, 23790.
(23) Kwon, N. Y.; Nathan, C. F.; Gilker, C.; Griffith, O. W.; Matthews,
D. E.; Stuehr, D. J. J. Biol. Chem. 1990, 265, 13442.
(24) Bec, N.; Gorren, A. C. F.; Voelker, C.; Mayer, B.; Lange, R. J.
Biol. Chem. 1998, 273, 13502.
(25) Klatt, P.; Schmidt, K.; Mayer, B. J. Biol. Chem. 1993, 268, 14781.
(26) Akhtar, M.; LeeRobichaud, P.; Akhtar, M. E.; Wright, J. N. J.
Steroid Biochem. Mol. Biol. 1997, 61, 127.
(27) Zhang, H. Q.; Dixon, R. P.; Marletta, M. A.; Nikolic, D.;
VanBreemen, R.; Silverman, R. B. J. Am. Chem. Soc. 1997, 119, 10888.
(28) Marletta, M. A. J. Biol. Chem. 1993, 268, 12231.
(29) Kerwin, J. J. F.; Feldman, P. L.; Lancaster, J. R. J. Med. Chem.
1995, 38, 4344.
(30) Fukuto, J. M.; Chaughuri, G. Annu. ReV. Pharmacol. Toxicol. 1995,
35, 165.
(33) (a) A 5-coordinate nitritoFe(III) porphyrin species has never been
characterized. Munro and Sheidt have shown that Fe(Tpiv)(NO2) is a low-
spin, N-bound nitro complex on the basis of EPR evidence. (Munro, O.
Q.; Scheidt, W. R. Inorg. Chem. 1998, 37, 2308.) Similar to Fe(Tpiv)-
(NO2), the high-spin Fe(III)TMP species reported here was detected by 1H
NMR as the only intermediate in the aerobic oxidation of Fe(NO)TMP to
Fe(NO3)TMP. (b) The same high-spin Fe(III)TMP species could also be
prepared by extensively evacuating the 6-coordinate low-spin Fe(NO)(NO2)-
TMP complex. Fe(NO)(NO2)TMP was prepared from NO and Fe(OH)-
TMP in toluene-d8 by modifying a literature method, (Ellison, M. K.; Schulz,
1
C. E.; Scheidt, W. R. Inorg. Chem. 1999, 38, 100), and it gave H NMR
(C7D8, 270 MΗz) signals at δ 1.75 (s, 12H), 2.0 (s, 12H), 2.5 (s, 12H), and
8.8 (s, 8H, pyrrole-H). (c) An O-nitrito analogue Fe(III)(TTP)[ONC(CN)2]
has been reported to be high-spin. (Bohle, D. S.; Conklin, B. J.; Hung,
C.-H. Inorg. Chem. 1995, 34, 2569.) Accordingly, we tentatively assigned
the high-spin Fe(III)TMP species as the O-bound nitritoFe(III)TMP complex
(Fe(ONO)TMP).
(31) Marletta, M. A. J. Med. Chem. 1994, 37, 1899.
(32) Philillipi, M. A.; Baenziger, N.; Goff, H. M. Inorg. Chem. 1981,
20, 3934.

12096 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Wang et al.
Table 2. Product Distribution of O2-Effected Oxidation of Fe(oximate)TMPa at Different O2 Pressures
dimer/FldO
pO2b (reaction time)c
dimerd
FldOe
ratio
Fe(NO)P
Fe(ONO)P
Fe(NO3)P
Fe(OH)P
14.7 (3 weeks)
100 (72 h)
200 (30 h)
0.10f
0.09
0.08
0.07
0.80
0.82
0.84
0.86
0.13
0.11
0.10
0.08
0.24
0.05
0.04
0.04
0.20
0.28
0.32
0.27
0.36
0.49
0.48
0.55
0.20
0.18
0.16
0.14
500 (6 h)
a A 0.8 mM benzene-d6 solution of Fe(oximate)TMP was used in each experiment, and the products were analyzed with 1H NMR spectroscopy.
e
bO2 pressures are reported in psi. c The time length required for 50% Fe(oximate)TMP to be consumed. d Dimer, compound 1. FldO, fluorenone.
fYields are expressed in equivalents relative to the Fe(oximate)TMP consumed.
Figure 2. Structure of Fe(oximate)TDCPP (3) showing 35% probability
thermal ellipsoids for all non-hydrogen atoms.
4H)) changed to a group of widely spread, broad signals (δ
9.2, 10.5, 14.7, 16.6, -20, -25, and -30, with one H at each
position).
The structure of the oximate-containing Fe(III)P species was
determined by an X-ray crystallographic analysis of a single
crystal of Fe(oximate)TDCPP (3) as shown in Figure 2.
Figure 1. 1H NMR spectra of the reaction of a one-to-one ratio of
fluorenone oxime and Fe(OH)TMP in benzene-d6 at 25 °C. Labels: #,
m-phenyl protons of the Fe(III)TMP species; *, protons of fluorenone
oxime and the oximate ligand.
Complex 3 contains a single fluorenone oximate ligand oxygen-
bound to a 5-coordinate Fe(III) cation. This observed binding
mode is rather unusual. More commonly, oximes and oximates
bind to metal ions through nitrogen.35,36 To date, the only iron
porphyrin complexes containing oxygen-bound monodentate
oximate ligands are complex 3 and Fe[OCN(CN)2]TTP.37
There are two notable features in the structure of 3 (Table
3). The first is the nearly equal C-N (1.334 Å) and N-O (1.343
Å) bond distances in the oximate ligand. The C-N bond is
slightly (0.04-0.07 Å) longer than C-N double bonds (1.26-
1.29 Å) in oximes38 but is much shorter (0.15-0.19 Å) than
C-N single bonds (1.48-1.52 Å) in nonconjugated C-nitroso
compounds.39 The N-O bond is 0.04-0.08 Å shorter than N-O
single bonds (1.38-1.42 Å) in oximes38 but is 0.13-0.2 Å
longer than N-O double bonds (1.14-1.21 Å) in nonconjugated
C-nitroso compounds.39 These observations suggest that the π
to those of a new Fe(III)TMP species (2; δ 78, pyrrole-H; δ
11.7 and 12.5, m-phenyl-H) (Figure 1). The pyrrole-H signal
of 2 was characteristic of a 5-coordinate, high-spin Fe(III)P
species,34 which indicated that 2 had only one axial ligand (eq
1). Concurrently, the 1H NMR resonance signals of fluorenone
oxime (δ 8.5 (d, 1H); 7.9 (d, 1H); 7.3 (m, 2H); 7.0-7.1 (m,
(34) Goff, H. M. In Physical Bioinorganic Chemistry Series. Iron
Porphyrins, Part I; Levere, A. B. P., Gray, H. B., Eds; Addison-Wesley:
Reading, MA, 1983; p 240.
(35) Chakravorty, A. Coord. Chem. ReV. 1974, 13, 1.
(36) Kukushkin, V. Y.; Tudela, D.; Pombeiro, A. J. L. Coord. Chem.
ReV. 1996, 156, 333.
(37) Bohle, D. S.; Conklin, B. J.; Hung, C.-H. Inorg. Chem. 1995, 34,
2569.
(38) Ianelli, S.; Nardelli, M.; Belletti, D.; Jamart-Gre´goire, B.; Caube`re,
P. Acta Crystallogr. 1992, C48, 1734.
(39) Miao, F.-M.; Chantry, D.; Harper, T.; Hodgkin, D. C. Acta
Crystallogr. 1982, B38, 3152.

Models of Nitric Oxide Synthase
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 12097
Table 3. Selective Bond Lengths (Å) and Angles (deg) of
O-(9-Nitro-9-fluorenyl)fluorenone Oxime (1) and
Fe(oximate)TDCPP (3)a
bond lengths
bond angles
Compound 1
1.442(12) O1*-C18-N2
O1*-N1*
109.7(3)
125.5(4)
97.8(4)
123.2(3)
117.1(3)
119.7(3)
106.0(8)
109.4(9)
O1*-C18
O2-N2
1.480(9)
1.188(3)
1.195(3)
1.286(12)
1.569(4)
1.499(5)
1.507(4)
O1*-C18-C17A
O1*-C18-C18A
O2-N2-O3
O3-N2
N1*-C9*
N2-C18
C18-C17A
C18-C18A
O2-N2-C18
O3-N2-C18
N1*-O1*-C18
C9*-N1*-O1*
Compound 3
Figure 3. Structure of O-(9-nitro-9-fluorenyl)fluorenone oxime (I)
showing 50% probability thermal ellipsoids for all non-hydrogen atoms.
Fe1-O1
Fe1-N1
Fe1-N2
Fe1-N3
Fe1-N4
O1-N5
N5-C45
1.836(2)
2.060(2)
2.073(2)
2.056(2)
2.058(2)
1.343(3)
1.334(5)
Fe1-O1-N5
O1-Fe1-N1
O1-Fe1-N2
O1-Fe1-N3
O1-Fe1-N4
O1-N5-C45
126.9(2)
105.6(1)
99.7(1)
99.6(1)
105.2(1)
112.4(3)
has been crystallographically characterized in the product of
metal-mediated, 1-electron oxidation of oximes.36 For example,
cis-[PtCl4(Me2CdNOH)2] spontaneously converted to [Pt(II)-
Cl2(N(dO)CMe2ONCMe2)] via the Pt(IV)-mediated 1-electron
oxidation of the oxime to form the O-(R-nitrosoisopropyl)-
acetone oxime ligand.50 The fluorenyl iminoxy radical is known
to dimerize rapidly (k ) 2 × 104 M-1 s-1).51 Thus, we conclude
that, during the oxidation of Fe(oximate)P, fluorenyl iminoxy
radicals were generated from the Fe-O bond homolysis with
subsequent dimerization to form O-(9-nitroso-9-fluorenyl)-
fluorenone oxime. Further oxidation of the nitroso group of the
iminoxy radical dimer by O2 would generate observed O-(9-
nitro-9-fluorenyl)fluorenone oxime52 (eq 2).
a Numbers in parentheses are estimated standard deviations. Atoms
are labeled as indicated in Figures 2 and 3.
system extends over the C-N and N-O bonds but with a
greater π bond density located on the C-N bond. Similar
observations have been reported for the structure of Fe[OCN-
(CN)2]TTP; i.e., the corresponding C-N and N-O bond
distances are 1.300 and 1.330 Å, respectively.
The other notable feature in 3 is a relatively short Fe-O bond
(1.84 Å). Monodentate oxyanion-ligated 5-coordinate Fe(III)P
complexes typically exhibit longer Fe-O bonds, e.g., 1.90-
2.07 Å. The Fe-O bond (1.945 Å) in Fe[OCN(CN)2]TTP is
such an example. The only exceptions have been Fe-O bonds
in methoxyFe(III)P (Fe-O bond length, 1.82-1.85 Å)40-43 and
phenoxyFe(III)P complexes (Fe-O bond length, 1.85-1.87
Å).44,45 It is interesting to note that all of the methoxyFe(III)P
complexes contain a planar porphyrin ring with the Fe atom
displaced 0.44-0.49 Å from the porphyrin plane. The porphyrin
ring in 3 is also planar, and the Fe atom is 0.45 Å out of the
porphyrin plane. By contrast, the porphyrin ring in Fe[OCN-
(CN)2]TTP is saddle-distorted, and the Fe atom is only 0.36 Å
out of the plane defined by the four porphyrin nitrogens.
Structure of the Organic Dimer 1. The dimer species 1
was characterized by X-ray crystallography to be O-(9-nitro-
9-fluorenyl)fluorenone oxime as shown in Figure 3. Species 1
contains two fluorene moieties linked by an oxime ether bridge,
and the nitro group is attached to the fluorenyl C-9 atom bonded
to the oximyl-O end of the bridge. Only two other compounds
with an O-(R-nitroalkyl)oxime ether functional group have been
structurally characterized.46,47 Mechanistic studies indicated that
both compounds were generated via the dimerization of iminoxy
radical species.48,49 An O-(R-nitrosoalkyl)oxime dimer structure
Kinetic Effects of Oxygen Pressure. The O2-effected
oxidation of Fe(III)(oximate)P was found to show the charac-
teristics of a metalloporphyrin-catalyzed radical-type autoxida-
tion.53 The kinetic profile of the reaction was biphasic with an
induction period followed by a fast phase of oxidation. This
1
feature was revealed by observing changes in the H NMR
spectra of benzene-d6 solutions of 2 (0.8 mM) under 500 psi
O2. As shown in Figure 4, complex 2 remained intact for 4.5 h
until the fast oxidation took place. The fast oxidation completely
consumed 2 in a period of another 5 h. Both the induction time
and the fast phase oxidation of the reaction could be shortened
by increasing the O2 pressure. As the O2 pressure was increased
(40) Hoard, J. L.; Hamor, M. J.; Hamor, T. A.; Caughty, W. S. J. Am.
Chem. Soc. 1965, 87, 2312.
(41) Lecomte, C.; Chadwick, D. L.; Coppens, E. D.; Stevens, E. D. Inorg.
Chem. 1983, 22, 2982.
(42) Hatano, K.; Uno, T. Bull. Chem. Soc. Jpn. 1990, 63, 1825.
(43) Johnson, M. R.; Seok, W. K.; Ma, W.; Slebodnick, C.; Wilcoxen,
K. M.; Ibers, J. A. J. Org. Chem. 1996, 61, 3298.
(44) Goff, H. M.; Shimomura, E. T.; Lee, Y. J.; Scheidt, W. R. Inorg.
Chem. 1984, 23, 315.
(45) Helms, A. M.; Jones, W. D.; McLendon, G. L. J. Coord. Chem.
1991, 23, 351.
(49) Oreshko, G. O.; Fadeev, M. A.; Lagodzinskaya, G. V.; Kozyreva,
I. Y.; Eremenko, L. T. IzV. Akad. Nauk. SSSR. Ser. Khim. 1986, 2533.
(50) Kukushkin, V. Y.; Belsky, V. K.; Aleksandrova, E. A.; Konovalov,
V. E.; Kirakosyan, G. A. Inorg. Chem. 1992, 31, 3836.
(51) Brokenshire, J. L.; Robert, J. R.; Ingold, K. U. J. Am. Chem. Soc.
1972, 94, 7040.
(46) Kokkou, S. C.; Rentzeperis, P. J. Acta Crystallogr. 1975, B31, 2793.
(47) Eremenko, L. T.; Atovmyan, L. O.; Golovina, N. I.; Oreshko, G.
O.; Fadeev, M. A. IzV. Akad. Nauk. SSSR. Ser. Khim. 1988, 1870.
(48) Lianis, P. S.; Alexandrou, N. E. Chim. Chron., New Ser. 1984, 13,
193.
(52) Larson, H. O. In The chemistry of functional groups; Feuer, H.,
Ed; John Wiley & Sons: New York, 1969; Vol. 1, pp 304, 339 and
references therein.
(53) Mlodnicka, T. In Metalloporphyrins in Catalytic Oxidations; Shel-
don, R. A., Ed; Marcel Dekker Inc.: New York, 1994; p 261.

12098 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Wang et al.
Table 4. Mass Spectrometric Data of Fluorenone Generated from
the O2-Effected Oxidation of Fe(oximate)TMPa
m/e
fluorenone from 16O2b (%)
fluorenone from 18O2c (%)
179
180
181
182
183
184
0.4d
100.0
13.8
0.8
0.0
0.0
0.0
30.5
5.9
100.0
15.2
1.0
a The reaction was conducted in a 2 mM benzene-d6 solution of
Fe(oximate)TMP under 100 psi O2(g) at 25 °C for 4 days. b Relative
intensities of the mass spectrum of fluorenone generated from the
c
reaction conducted under 100 psi regular O2(g). Relative intensities of
the mass spectrum of fluorenone generated from the reaction conducted
18
under 100 psi
O
2(g)
.
Figure 4. Time course plot of the aerobic oxidation of Fe(oximate)-
TMP in benzene-d6 as monitored by 1H NMR spectrosopy. The reaction
was conducted by stirring a 0.75 mM Fe(oximate)TMP solution under
500 psi O2(g) at 25 °C in the dark. Note that each point on the plot
represents an individual experiment.
and H218O generated
2(g)
16
the reaction conducted under
O
fluorenone containing only 16O. These results prove that O2 is
the predominant oxygen source of the fluorenone product, a
characteristic similar to what has been observed in the NOS
reaction.22
from 14.7 to 100, 200, and 500 psi, the induction time was
shortened from 2 weeks to 50, 20, and 4.5 h, respectively.
Further, the oxidation phase lasted for 2 weeks at 14.7 psi O2,
while at 500 psi O2 it was shortened to 5 h. These results suggest
that O2 directly participates in both the induction and the fast
phase oxidation.
Reaction of Fe(III)(oximate)P with CO. The reaction of 2
and CO, an O2 surrogate, was investigated. Complex 2 was
found to react with CO(g) in benzene to generate Fe(CO)TMP
and Fe(CO)2TMP within a few hours (Scheme 3). The 1H NMR
resonance signals of 2 gradually converted (t1/2 ≈ 2 days at 25
°C) to two sets of resonance signals of diamagnetic products
under a slight excess of CO(g). One set of signals was identical
to those of the authentic Fe(CO)2TMP/Fe(CO)TMP mixture
prepared according to the literature method.55 The other set of
signals first appeared as broad bands, which gradually converted
to the sharp signals of fluorenone oxime. Interestingly, the
conversion of the signals took place ∼2 days after the emergence
of the Fe(CO)2TMP/Fe(CO)TMP signals.
Significantly, the rates of the formation of Fe(CO)P/Fe(CO)2P
from Fe(oximate)P and CO for several porphyrins paralleled
the redox potentials of the corresponding Fe(III)P species. The
order of the reduction potentials of the corresponding Fe(III)P
species is as follows (e.g., with Cl- as the axial ligand): Fe-
(Cl)TPFPP (-0.08 V vs SCE in CH2Cl2)56 > Fe(Cl)TDCPP
(-0.221 V)57 > Fe(Cl)TMP (-0.49 V).58 Consistent with this
trend, the order of the reaction rates of Fe(oximate)P and 1 atm
The rate of the aerobic oxidation of Fe(III)(oximate)P varied
1
with the structure of the iron porphyrin. H NMR spectra of
reaction mixtures showed that the 0.8 mM benzene-d6 solutions
of Fe(oximate)TPFPP were oxidized under 500 psi O2 with an
induction time of 3.5 h, and the overall oxidations was complete
within 7 h. As in the aerobic oxidation of Fe(oximate)TMP,
the organic products of the aerobic oxidation of Fe(oximate)-
TFPP in benzene-d6 were fluorenone (>70%) and the fluorene-
based dimer (<30%). The porphyrin products were Fe(NO)-
TPFPP (20%), µ-oxoFe(III)TPFPP dimer (<5%),54 and Fe(NO3)-
TPFPP (75%). By contrast, Fe(oximate)TDCPP (3) showed no
indication of aerobic oxidation in a variety of solvents, appar-
ently due to precipitation.
Origin of the Fluorenone Oxygen. The oxygen atom of the
fluorenone generated from the O2-effected oxidation of Fe-
(oximate)P was verified to be derived from O2. Because of the
simplicity of the reaction, the potential oxygen sources can only
be O2, H2O, or the oximate ligand. Two sets of 18O-labeling
experiments were conducted to determine the oxygen source
of fluorenone. A 2 mM benzene-d6 solution of 2 was mixed
with 1 µL of H218O (98% enriched, ∼30 equiv with respect to
2), and the reaction was brought to 100 psi 16O2(g). The second
CO(g) was found to be as follows: Fe(oximate)TPFPP (t1/2
)
10 min at 25 °C) > Fe(oximate)TDCPP (3) (t1/2 ) 30 min) >
Fe(oximate)TMP (2) (t1/2 ) 1.5 h). This correlation suggests
that the rate-limiting step of the formation of Fe(CO)P involves
the reduction of Fe(III)P to Fe(II)P.
Scheme 3
reaction contained the same solution of 2 except without H218O,
18
and the reaction was conducted under 100 psi
O
2(g)
(98%
enriched). Both reaction mixtures were stirred in the dark for 4
days before the reactions were stopped by releasing the
headspace O2 pressure. GC-MS spectrometric analysis showed
that more than 75% of the fluorenone generated from the 18O2-
effected oxidation of 2 contained 18O. (Table 4). By contrast,
The kinetics of the reaction of Fe(oximate)P and CO(g) have
been investigated in detail to further reveal the reaction
mechanism. Thus, the formation rates of Fe(CO)2TMP from 2
and 1 atm CO(g) were measured at temperatures ranging from
(55) Wayland, B. B.; Mehne, L. F.; Swartz, J. J. Am. Chem. Soc. 1978,
100, 2379.
(56) Grinstaff, M. W.; Hill, M. G.; Birnbaum, E. R.; Schaefer, W. P.;
Labinger, J. A.; Gray, H. B. Inorg. Chem. 1995, 34, 4896.
(57) Chen, H. L.; Ellis, P. E., Jr.; Wijesekera, T.; Hagan, T. E.; Groh, S.
E.; Lyons, J. E.; Ridge, D. P. J. Am. Chem. Soc. 1994, 116, 1086.
(58) Swistak, C.; Mu, X. M.; Kadish, K. M. Inorg. Chem. 1987, 26,
4360.
(54) Jayaraj, K.; Gold, G. E.; Toney, G. E.; Helms, J. H.; Hatfield, W.
E. Inorg. Chem. 1986, 25, 3516.

Models of Nitric Oxide Synthase
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 12099
the fluorenyl iminoxy radical dimerization product, 1, was not
detected in the products under these conditions. Instead, free
fluorenone oxime was found as the final organic product 2 days
after the emergence of Fe(CO)TMP/Fe(CO)2TMP. This delayed
emergence of fluorenone oxime in the reaction product suggests
that it derived from the decomposition of reaction intermediates,
such as adducts of CO and fluorenyl iminoxy radicals. One
likely intermediate is bis(fluorenoneoximyl)carbonate resulted
from the free-radical carbonylation of fluorenyl iminoxy radical
and CO.66 The slow hydrolysis of this carbonate species by the
residual water in the sample would generate fluorenone oxime
and carbonic acid (eq 4). Indeed, this transformation consumed
Figure 5. Eyring plot for the generation of Fe(CO)TMP/Fe(CO)2TMP
from a benzene solution of Fe(oximate)TMP (7 µM) under 1 atm CO(g).
The k values are the pseudo-first-order rate constants of the increase
of Fe(CO)2TMP (monitored at 426 nm) at 30, 40, and 50 °C in the
dark. The correlation coefficient (r) of the data is 0.998.
30 to 50 °C, and an Eyring plot was constructed based on the
pseudo-first-order rate constants (Figure 5). From this Eyring
plot, the enthalpy (∆Hq) and the entropy of activation (∆Sq) of
the reaction were calculated to be 14.0((1.5) kcal/mol and
-31.1((5.0) eu, respectively. The large, negative entropy term
indicates that the rate-limiting step of the reaction involves an
association of the reactants.59 Consistent with this notion, the
formation rate of Fe(CO)P/Fe(CO)2P from Fe(oximate)P and
CO was found to be directly proportional to the partial pressure
of CO(g) above the reaction mixture.60 On the basis of these
results, the rate law for the formation of Fe(CO)TMP is shown
as eq 3,61 where k ) 1.04 (( 0.09) × 10-1 M-1 s-1 at 30 °C.
the residual water in the benzene-d6 solution as was evident by
the diminishing of the H NMR resonance signal of water in
the reaction mixtures.
1
In analogy to the reaction of CO and Fe(oximate)P, the π-acid
NO also induced the Fe-O bond homolysis of Fe(oximate)P.
1H NMR analysis showed that a 2 mM benzene-d6 solution of
2 reacted with a substoichiometric amount of NO(g) to generate
Fe(NO)TMP as the only iron porphyrin product. This reaction
was so fast that it took place even when NO(g) was added to a
solution of 2 under argon purging. Concomitant with the
formation of Fe(NO)TMP, the oximate ligand of 2 was
converted to a one-to-one ratio of fluorenone oxime and the
fluorene-based dimer 1 in addition to trace amounts of fluo-
renone.
In summary, we have presented the first example of synthetic
iron porphyrin-catalyzed oxidation of oximes by O2. This
reaction is analogous to the O2-effected oxidation of the oximate
ligand of Co(III)(salen) reported by Nishinaga.67 The proposed
mechanism for that reaction involved the insertion of O2 between
Co(II) and an iminoxy radical. The long induction time of the
O2-effected oxidation of Fe(oximate)P suggests that its rate-
limiting step is the homolysis of the Fe-O bond, which would
generate fluorenyl iminoxy radical and Fe(II)P. Among the
several mechanisms that could accommodate our data, the most
probable one is the reaction of Fe(II)P, fluorenyl iminoxy
radical, and O2. This reaction would furnish an unstable
peroxyFe(III)P intermediate (4) in analogy to the cobalt salen
case (Scheme 4). If compound 4 were to decompose via O-O
bond homolysis, analogously to the decomposition of other
alkylperoxyFe(III)P species,68,69 fluorenone, NO, and an oxoFe-
(IV) porphyrin would be generated in a single step. NO would
be captured by either Fe(II)P or Fe(oximate)P to afford Fe-
-d[Fe(oximate)TMP]/dt ) k[Fe(oximate)TMP][CO] (3)
In light of the ∆Sq value, the dependence of Fe(CO)P
formation rates on both of the reduction potentials of Fe(III)P
and headspace CO(g) partial pressures, we conclude that Fe-
(CO)P was generated via the CO-induced reduction of Fe-
(oximate)P. In the rate-limiting step, CO would approach
Fe(oximate)P from the direction trans to the oximate ligand,
which was induced to dissociate from 2. The generation of Fe-
(CO)P indicates that Fe(oximate)P was reduced to Fe(II)P either
during or after the dissociation of the oximate ligand. Since
CO can only reduce Fe(III)P in aqueous solutions,62 it is unlikely
that CO could reduce Fe(III)(oximate)P to form Fe(II)P. Thus,
Fe(CO)P/Fe(CO)2P must come from the reaction of CO and
Fe(II)P, which would result from the homolysis of the Fe-O
bond of Fe(oximate)P. Hence, CO induces the Fe-O bond
homolysis via a mechanism similar to the π-acid-assisted
reduction of Fe(III)P by phosphine63,64 or isocyanide.65
Such a homolysis of the Fe-O bond of Fe(oximate)TMP
would also generate the fluorenyl iminoxy radical. However,
(59) Wilkins, R. G. Kinetics and mechanism of reactions of transition
metal, 2nd ed.; VCH Publishers, Inc.: New York, 1991.
(60) When the reaction was conducted under 1 atm 20/80 CO/Ar gas
mixture, the formation rate of Fe(CO)TMP/Fe(CO)2TMP was reduced to
18((2)% of that of the reaction conducted at 1 atm CO(g)
.
(61) The second-order rate constant was calculated from the pseudo-
first-order rate constant of the formation of Fe(CO)TMP [(8 ( 0.4) × 10-5
s-1] at 30 °C and the concentration of CO in benzene (7.37-8.05 mM)
(Cargill, R. W. In Carbon Monoxide; Cargill, R. W., Ed.; Pergamon Press:
New York, 1981; Vol. 43, p 114).
(62) Bickar, D.; Bonaventura, C.; Bonaventura, J. J. Biol. Chem. 1984,
259, 10777.
(63) Inoue, H.; Akahori, H.; Ohno, Y.; Nakazawa, K.; Nonomura, Y.;
Yoshioka, N.; Heckmann, G.; Fluck, E. Z. Naturforsch. Sect. B 1995, 50,
1222.
(66) Ryu, I.; Sonoda, N. Angew. Chem., Int. Ed. Engl. 1996, 35, 1050.
(67) Nishinaga, A.; Yamazaki, S.; Miwa, T.; Matsuura, T. React. Kinet.
Catal. Lett. 1991, 43, 273.
(68) Balch, A. L.; Cornman, C. R.; Safari, C. R.; Latos-Grazynski, L.
Organometallics 1990, 9, 2420.
(64) Del Gaudio, J.; La Mar, G. N. J. Am. Chem. Soc. 1978, 100, 1112.
(65) Nakano, T.; Kitamura, Y. Chem. Lett. 1989, 1207.
(69) Balch, A. L.; Hart, R. L.; Latos-Grazynski, L.; Traylor, T. G. J.
Am. Chem. Soc. 1990, 112, 7382.

12100 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Wang et al.
Scheme 4
Fe(II) heme has a lower affinity to CO than other hemopro-
teins.76 The binding constant of iNOS Fe(II) heme, the isoform
used in the CO binding assay by Pufahl et al., toward CO (Kco)
is 5 × 103 M-1 77
,
16 times smaller than those of other P450
monoxygenases such as P450cam and P450nor (Kco) 8.5 ×
104 and 8.8 × 104 M-1, respectively).78 Therefore, unless the
majority of the iNOS sample is reduced to Fe(II) and the solution
contains high concentration of CO, the UV-visible signal of
thiolate-ligated NOS Fe(II)(CO) heme species would be difficult
to detect.
Another likely reason for the absence of a P-450-type
difference spectrum is that the NHA-ligated 6-coordinate heme-
Fe(III) species could be in an equilibrium with the tightly
associated iminoxy radical-Fe(II) heme so that the Fe(II) heme
cannot react with CO. The NHA-derived iminoxy radical ligand
may associate with the Fe(II) heme so strongly that the complex
only decomposes upon reacting with O2 at the carbon center of
that radical. Similar tightly associated Fe(III)-substrate adducts
have been shown to be involved in both of the intradiol-cleaving
catechol dioxygenase and R-keto acid-dependent oxygenase
reactions.79,80
(NO)P. Further oxidation of Fe(NO)P would lead to Fe(NO3)P
and the putative Fe(ONO)P as is observed.
We suggest that the second step of the NOS reaction to form
NO is initiated by an Fe-O bond homolysis of the NHA-Fe-
(III) heme adduct. This process would generate an Fe(II) heme
and the NHA-derived iminoxy radical. The reaction of these
two species with O2 would reasonably afford the key peroxyFe-
(III) heme species (an analogue of 4) followed by its decom-
position to citrulline, NO, and oxoFe(IV) heme. The single
NADPH-derived reducing equivalent consumed in the second
step NOS reaction would then be used to reduce oxoFe(IV)
heme back to Fe(III) heme. In light of the direct contact of the
potential 1-electron cofactor H4B81a and NOS heme,82 it is likely
in this scenario that H4B is responsible for mediating this
reduction process affording an intermediate H3B• radical81b
(Scheme 5). Such a scheme has the intuitively attractive features
of producing a radical species, NO, via the participation of
single-electron redox cofactors, H4B and an NADPH-derived
reducing equivalent, by an radical-type autoxidation process.
This and other aspects of this interesting reaction are under
current study.
Implications for the NOS Reaction Cycle. The implications
of this model to the NOS reaction mechanism are intriguing.
First, the structure of the Fe(III) oximate 3 suggests an
unprecedented mode of interaction between the NOS Fe(III)
heme and the N-hydroxyguanidine function of NHA. This notion
is unusual given the fact that there is no evidence of an Fe-O
bond in the NHA-bound NOS.70,71 Nonetheless, the crystal
structure of L-arginine-bound iNOS shows that the guanidine-N
of arginine is 3.8 Å away from the heme-iron. This distance is
only ∼1 Å longer than that observed for Fe and oximyl-N in 3.
Further, the NHA-bound iNOS has been found to have a similar
structure with the N-hydroxy oxygen positioned directly above
the heme iron.72 Movement of NHA within the active site of
NOS could allow the N-hydroxy group to approach and ligate
to Fe(III) heme. Second, the occurrence of the Fe-O bond
homolysis of Fe(oximate)P under CO suggests that NHA is able
to reduce the Fe(III) heme to initiate the second step of the
NOS reaction. This notion is even more convincing when one
considers the redox potentials of the NOS Fe(III) heme (E°)
-248 to ∼-263 mV)73 and the conjugate base of the NHA
N-hydroxy group (E°e -200 mV).74 Thus, the Fe-O homolysis
of the presumed NHA-Fe(III) heme adduct could be an
energetically accessible process.
There have been reports suggesting that N-hydroxyarginine
does not reduce Fe(III) heme to initiate the second step of NOS
reaction.20,75 It has been shown that the reaction of NOS and
NHA under CO(g) does not give the characteristic P450-type
visible spectrum (λmax ) 450 nm).75 One possible reason for
this negative result is that the equilibrium concentration of the
CO-ligated Fe(II) heme relative to Fe(III) heme is too low to
be detected by the difference spectrum technique, for it has been
shown that tetrahydrobiopterin (H4B) and L-arginine restrict CO
access to NOS heme iron. Due to this steric restriction, NOS
Experimental Section
General Procedures and Methods. NMR spectra were collected
on JEOL GSX-270 (270 MHz) and Varian INOVA 500 (500 MHz)
1
NMR spectrometers. H NMR chemical shift values are reported in
ppm relative to the residual solvent resonances (1H NMR δ 7.26 for
C6HD5). UV-visible experiments were conducted on an HP 8452A
diode array spectrophotometer and a Varian Cary 2390 spectropho-
tometer. GC-MS data were obtained on an HP 5890 series II Plus gas
chromatograph equipped with a HP-5MS capillary column (30 m) and
an HP 5989B mass spectrometer. All anaerobic and gas-control
experiments were conducted either on a high-vacuum (<10-6 mmHg)
Schlenk preparative line or in an inert-air glovebox controlled by
(76) Wang, J. L.; Stuehr, D. J.; Rousseau, D. L. Biochemistry 1997, 36,
4595.
(77) Abu-Soud, H. M.; Wu, C. Q.; Ghosh, D. K.; Stuehr, D. S.
Biochemistry 1998, 37, 3777.
(78) Shiro, Y.; Minoru, K.; Iizuka, T.; Nakahara, K.; Shoun, H.
Biochemistry 1994, 33, 8673.
(70) Salerno, J. C.; Martasek, P.; Williams, R. F.; Masters, B. S. S.
Biochemistry 1997, 36, 11821.
(71) Sennequier, N.; Stuehr, D. J. Biochemistry 1996, 35, 5883.
(72) Tainer, J. A., personal communication on the crystal structure of
iNOS oxygenase dimer with bound cofactors and NHA.
(73) Presta, A.; Weber-Main, A. W.; Stankovich, M. T.; Steuhr, D. J. J.
Am. Chem. Soc. 1998, 120, 9460.
(74) Bordwell, F. G.; Ji, G.-Z. J. Org. Chem. 1992, 57, 3019.
(75) Pufahl, R. A.; Marletta, M. A. Biochem. Biophys. Res. Commun.
1993, 193, 452.
(79) Bugg, T. D.; Winfield, C. J. Nat. Prod. Rep. 1998, 15, 513.
(80) Que, L., Jr.; Ho, R. Y. N. Chem. ReV. 1996, 96, 2607.
(81) (a) Lunte, C. E.; Kissinger, P. T. Anal. Chem. 1983, 55, 1458. (b)
The H3B• radical has very recently been detected in a single-turn-over
experiment with the iNOS heme domain: Hurshman, A. R.; Krebs, C.;
Edmondson, D. E.; Huynh, B. H.; Marletta, M. A. Biochemistry 1999, 38,
15689-15696.
(82) Crane, B. R.; Arvai, A. S.; Ghosh, D. K.; Wu, C.; Getzoff, E. D.;
Stuehr, D. J.; Tainer, J. A. Science 1998, 279, 2121.

Models of Nitric Oxide Synthase
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 12101
H), 15.2 (s, 4H, m-phenyl H), 16.4 (s, 4H, m-phenyl H), 4.0 (s, 12H,
p-methyl H),2-6 (broad, 24H, o-methyl H); UV-visible (C6H6) λmax
(ꢀ × 10-3) 330 (67), 418 (100), 580 nm (7.7).
Scheme 5
Nitrato(5,10,15,20-tetrakis(pentafluorophenyl)porphyrinato)iron-
(III) [Fe(NO3)TPFPP] was prepared with the same procedure above
using Fe(OH)TPFPP instead of Fe(OH)TMP: 1H NMR (C6D6, 25 °C,
270 MHz) δ 81 (pyrrole H); UV-visible (C6H6) 408, 507 nm.
Nitrosyl(5,10,15,20-tetrakis(mesityl)porphyrinato)iron(II) [Fe-
(NO)TMP] was prepared by reductive nitrosylation of Fe(Cl)TMP.86,87
NO gas was introduced to the anaerobic benzene solution of a 1:1
mixture of methanol and Fe(Cl)TMP. The solution was reduced to
dryness after reacting for overnight. The solid product was isolated
under ambient air: 1H NMR (C6D6, 25 °C, 270 MHz) δ 2.60 (s, 8H,
p-methyl H), 2.1-2.8 (24H, o-methyl H), 5.6-6.4 (8H, pyrrole H),
7.8-8.6 (8H, m-phenyl H); UV-visible (CH2Cl2) λmax (ꢀ × 10-3) 408
(92), 539 (11), 611 nm (2.3); IR (KBr) νNO 1677 cm-1(vs).
Nitrosyl(5,10,15,20-tetrakis(mesityl)porphyrinato)iron(II) [Fe-
(NO)TPFPP] was prepared with the same procedure above using Fe-
(Cl)TPFPP instead of Fe(Cl)TMP: 1H NMR (C6D6, 25 °C, 270 MHz)
δ 5.4-6. 6 (pyrrole H); UV-visible (C6H6) 402, 544 nm.
Monocarbonyl(5,10,15,20-tetrakis(mesityl)porphyrinato)iron-
(II) and Biscarbonyl(5,10,15,20-tetrakis(mesityl)porphyrinato)iron-
Vacuum Atmosphere Co. Dri-Train model MO40-2 with O2 concentra-
tion maintained under 2 ppm.
(II) [Fe(CO)TMP and Fe(CO)2TMP] were prepared from Fe(II)TMP
55
and CO(g)
.
Fe(II)TMP was prepared by reducing Fe(Cl)TMP with
Materials. All reagents and chemicals were obtained from com-
mercial sources and used as received unless otherwise described.
Solvents were dried and distilled under N2 immediately before use
according to literature procedures.83 Fluorenone oxime was prepared
by refluxing a mixture of hydroxylamine hydrochloride and fluorenone
in ethanol.84 The 15N-labeled fluorenone oxime was prepared from 15N-
hydroxylamine hydrochloride (> 96% 15N enriched). Fe(III)tetrakis-
(mesityl)porphyrin chloride (Fe(Cl)TMP), Fe(III)tetra(pentafluorophenyl)-
porphyrin chloride (Fe(Cl)TPFPP), and Fe(III)tetra(2,6dichlorophenyl)-
porphyrin chloride (Fe(Cl)TDCPP) were purchased from Midcentury
Inc. and used directly without further purification. 18O2(g) (98% enriched)
and H218O (98% enriched) were both purchased from Cambridge Isotope
Laboratories. C. P. grade NO(g) and CO(g) were purchased from
Matheson Gas Products. NO(g) was passed through a column of KOH
pellets to remove higher nitrogen oxides immediately before use.
Hydroxo(5,10,15,20-tetrakis(mesityl)porphyrinato)iron(III) [Fe-
(OH)TMP].85 A total of 10 mg of Fe(Cl)TMP dissolved in minimum
amount of CH2Cl2 was eluted through a column packed with a mixture
of basic alumina and water (10 wt %) with purified CH2Cl2. The solvent
was immediately removed in vacuo from the green elute to give
quantitative a yield of Fe(OH)TMP: 1H NMR (C6D6, 25 °C, 270 MHz)
δ 78 (8H, pyrrole H), 11 (s, 4H, m-phenyl H), 12 (s, 4H, m-phenyl H),
3.3 (s, 12H, p-methyl H), 2-6 (broad, 24H, o-methyl H); UV-visible
(C6H6) λmax (ꢀ × 10-3) 346 (30), 416 (98), 510 nm (11).
sodium dithionite in benzene/H2O under N2. The mixture was stirred
until all Fe(III)TMP was consumed. The organic layer was separated
from the aqueous layer under CO to generate a mixture of Fe(CO)-
TMP/Fe(CO)2TMP. The UV-visible spectrum of the crude mixture
under CO showed that, upon either heating or irradiation, the 410 nm
peak increases while the 426 nm peak decreases. Due to the instability
of the second CO ligand, the species giving the 426 nm peak was
assigned as Fe(CO)2TMP while the species giving the 410 nm peak
was assigned as Fe(CO)TMP: 1H NMR (C6D6, 25 °C, 270 MHz) δ
1.95 (s, 24H, o-methyl H), 2.43 (s, 12H, p-methyl H), 7.09 (s, 8H,
m-phenyl H), 8.79 (s, 8H, pyrrole H); UV-visible (CH2Cl2) λmax
Fe(CO)2TMP 426, 510 nm; Fe(CO)TMP 410, 502 nm.
Fluorenoneoximato(5,10,15,20-tetrakis(mesityl)porphyrinato)-
iron(III) [Fe(oximate)TMP] (2). Ten milligrams of Fe(OH)TMP (11.7
µmol) and 2.4 mg of fluorenone oxime (12.2 µmol) were mixed in 20
mL of benzene. The mixture was stirred for 10 min, and then the solvent
was removed in vacuo. The solid was dissolved in a minimum of CHCl3,
and CH3CN was diffused in over 2 days to precipitate small crystals
of 2: 1H NMR (C6D6, 25 °C, 270 MHz) δ 78 (8H, pyrrole H). 12.5 (s,
4H, m-phenyl H), 11.7 (s, 4H, m-phenyl H), 16.6 (s, 1H), 14.7 (s, 1H),
10.5 (s, 1H), 9.2 (s, 1H), -20 (s, 1H), -25 (s, 1H), -30 (s, 1H); UV-
visible (CH2Cl2) λmax (ꢀ × 10-3) 332 (51), 422 (102), 580 (12), 636
nm (7.5).
Fluorenoneoximato(5,10,15,20-tetrakis(2,6-dichlorophenyl)por-
phyrinato)iron(III) [Fe(oximate)TDCPP] was prepared with the same
procedure above using Fe(OH)TDCPP instead of Fe(OH)TMP: 1H
NMR (C6D5Cl, 25 °C, 270 MHz) δ 76 (8H, pyrrole H), 11.5 (s, 4H,
m-phenyl H), 10.8 (s, 4H, m-phenyl H), 7.9 (s, 4H, p-phenyl H), 17.1
(s, 1H), 15.6 (s, 1H), 9.7 (s, 1H), -17 (s, 1H), -21 (s, 1H), -28 (s,
1H); UV-visible (CH2Cl2) λmax 338, 416, 578 nm.
Fluorenoneoximato(5,10,15,20-tetrakis(pentafluorophenyl)por-
phyrinato)iron(III) [Fe(oximate)TPFPP] was prepared with the same
procedure above using Fe(OH)TPFPP instead of Fe(OH)TMP: 1H
NMR (C6D6, 25 °C, 270 MHz) δ 79 (8H, pyrrole H), 17.5 (s, 1H),
16.7 (s, 1H), 11.4 (s, 1H), 10.3 (s, 1H), -16 (s, 1H), -22 (s, 1H) -29
(s, 1H); UV-visible (CH2Cl2) λmax (ꢀ × 10-3) 404 (76), 570 nm (10.8).
O-(9-Nitro-9-fluorenyl)fluorenone Oxime. Four milligrams of Fe-
(OH)TDCPP and 27 mg of fluorenone oxime were dissolved in 100
mL of thiophene-free benzene in a silanized Pyrex flask. The solution
was bubbled with O2 and irradiated with a 200 W tungsten lamp filtered
with a No. 3-73 Corning color glass filter (λ > 420 nm). After
fluorenone oxime completely reacted (∼3 days), the solvent was
removed in vacuo. The solid was triturated with CH3CN, and the pale
yellow insoluble residue was isolated. Yield is 50% based on fluorenone
Hydroxo(5,10,15,20-tetrakis(pentafluorophenyl)porphyrinato)-
iron(III) [Fe(OH)TPFPP] was prepared with the same procedure above
using Fe(Cl)TPFPP instead of Fe(Cl)TMP. The eluate containing Fe-
(OH)TPFPP was immediately reduced to dryness to avoid the formation
of the µ-oxo dimer:54 UV-visible (CH2Cl2) λmax (ꢀ × 10-3) 406 (76),
563 nm (11.9).
Hydroxo(5,10,15,20-tetrakis(2,6-dichlorophenyl)porphyrinato)-
iron(III) [Fe(OH)TDCPP] was prepared with the same procedure
above using Fe(Cl)TDCPP instead of Fe(Cl)TMP: UV-visible (CH2-
Cl2) λmax 332, 414, 578 nm.
Nitrato(5,10,15,20-tetrakis(mesityl)porphyrinato)iron(III) [Fe-
(NO3)TMP] was prepared by modification of the literature method32
by mixing a one-to-one ratio of Fe(OH)TMP and HNO3 in benzene-
d6. The progress of the reaction was monitored by 1H NMR. The signals
for Fe(OH)TMP were completely converted to another Fe(III)TMP
species upon the mixing. On the basis of the literature precedent and
the stoichiometry of the reaction, the new Fe(III)P was assigned as
Fe(NO3)TMP: 1H NMR (C6D6, 25 °C, 270 MHz) δ 79 (8H, pyrrole
(83) Perrin, D. D. Purification of Laboratory Chemicals, 3rd ed.;
Pergamon Press: Oxford, U.K., 1988.
(84) Vogel’s Textbook of Practical Organic Chemistry, 4th ed.; Furniss,
B., Hannaford, A. J., Rogers, V., Smith, P. W. G., Tatchell, A. R., Eds;
Wiley: New York, 1987; p 1113.
(86) Wayland, B. B.; Olson, L. W. J. Am. Chem. Soc. 1974, 96, 6037.
(87) Mu, X. H.; Kadish, K. M. Inorg. Chem. 1988, 27, 4720.
(85) Groves, J. T.; Gross, Z.; Stern, M. K. Inorg. Chem. 1994, 33, 5065.

12102 J. Am. Chem. Soc., Vol. 121, No. 51, 1999
Wang et al.
oscillation range of 2°/frame and an exposure time of 60 s/deg.88
oxime: 1H NMR (C6D6, 2mM, 25 °C, 500 MHz) δ 6.80 (t, J ) 7.5
Hz, 1H, 2-fluorenoneoximyl H), 6.93 (t, J ) 7.5 Hz, 1H, 7-fluorenone-
oximyl H), 6.95 (t, J ) 7.5 Hz, 1H, 3-fluorenoneoximyl H), 6.97 (t, J
) 7.5 Hz, 2H, 2-fluorenyl H), 7.01 (t, J ) 7.5 Hz, 1H, 6-fluorenone-
oximyl H), 7.06 (t, J ) 7.5 Hz, 1H, 3-fluorenyl H), 7.11 (d, J ) 7.5
Hz, 1H, 4-fluorenoneoximyl H), 7.12 (d, J ) 7.5 Hz, 1H, 5-fluorenone-
oximyl H), 7.20 (d, J ) 7.5 Hz, 2H, 4-fluorenyl H), 7.64 (d, J ) 7.5
Hz, 1H, 8-fluorenoneoximyl H), 8.02 (d, J ) 7.5 Hz, 2H, 1-fluorenyl
H), 8.30 (d, J ) 7.5 Hz, 1H, 1-fluorenoneoximyl H); IR (CDCl3) νNO2
1561 (vs), 1349 (s); νNO 989 (vs); νCO 1036 cm-1 (s).
X-ray Crystallography. General Procedure. X-ray crystallographic
studies were carried out on a Nonius KappaCCD diffractometer
equipped with graphite-monochromatized Mo KR radiation (λ )
0.710 73 Å). Samples were either mounted on a glass fiber with epoxy
cement or sealed in a glass capillary, and the diffraction data were
collected at room temperature. Relevant crystallographic information
is listed in Table 1.
A
total of 79 647 reflections (θmax ) 27.46°) were indexed, integrated,
and corrected for Lorentz and polarization effects using DENZO-SMN
and SCALEPACK.89 Data reduction yielded 13 775 unique reflections
(Rint ) 0.0490) of which 7272 had I > 2σ(I). Postrefinement of the
unit cell parameters gave a ) 12.5029(2) Å, b ) 13.0424(4) Å, c )
18.9793(5) Å, R ) 81.701(1)°, â ) 88.304(2)°, γ ) 84.263(2)°, and
V ) 3046.80(13) Å3. Axial photographs and a lack of systematic
absences suggested that the compound had crystallized in the triclinic
space group P1 or P1h. The latter space group P1h (No. 2) was selected
on the basis of an observed mean value of 0.901 for |E* E - 1| (versus
the expectation values of 0.968 and 0.736 for centric and noncentric
data, respectively).
The structure was solved by direct methods and refined by full-
matrix least squares on F2 using SHELXTL90 to R(F) ) 0.1545 and
wR(F2) ) 0.3066 for 13 775 unique reflections. A disordered chlo-
robenzene solvent molecule was included in the refinement, and two
additional carbon atoms were placed at a second (presumably heptane)
solvent site. The chlorobenzene did not refine well, and the identity of
the second solvent molecule could not be assigned via a discrete-atom
approach. Therefore, the analysis was resumed with the following
solvent-free model.
X-ray Crystal Structure Determination of O-(9-Nitro-9-fluore-
nyl)fluorenone Oxime (1). Single crystals of 1 suitable for X-ray
crystallographic analysis were obtained by slow diffusion of CH3CN
into a benzene solution of 1 over a period of 1 week. A pale yellow
irregular chunk (0.15 mm × 0.20 mm × 0.25 mm in size) was used
for the diffraction experiment (Table 1). A total of 608 frames of data
were collected at 298(2) K with an oscillation range of 1°/frame and
The SQUEEZE/BYPASS procedure93 implemented in PLATON94
was used to account for the solvent electron density. Two solvent-
accessible voids were detected. The first of these two voids is situated
along the crystallographic c-axis (0, 0, z) and has a volume and electron
count of 482 Å3 and 92 e, respectively. From our discrete-atom approach
mentioned above, we know that this void is occupied by two
chlorobenzene molecules (339 Å3, 116 e). The second void is situated
at the center of the cell (0.5, 0.5, 0.5) and has a volume and electron
count of 245 Å3 and 55 e, respectively. These results suggest that the
occupant of this second void may be a molecule of the precipitating
solvent heptane (244 Å3, 58 e) used in recrystallizing the iron complex.
Since there are two molecules of the iron complex per cell, the chemical
formulation C57H28Cl8FeN5O‚6H5Cl‚.5C7H16 is proposed.
an exposure time of 120 s/deg.88 A total of 26 402 reflections (θmax
22.48°) were indexed, integrated, and corrected for Lorentz and
polarization effects using DENZO-SMN and SCALEPACK.89 The θmax
limit was reduced from the standard 27.50° due to the sample being a
)
poor scatterer. Data reduction yielded 2622 unique reflections (Rint
)
0.0530) of which 1948 had I > 2σ(I). Postrefinement of the unit cell
parameters gave a ) 9.8776(6) Å, b ) 11.9206(6) Å, c ) 17.4262-
(12) Å, â ) 101.550(3)°, and V ) 2010.3(2) Å3. Axial photographs
and systematic absences were consistent with the compound having
crystallized in the monoclinic space group P21/n (No. 14).
The structure was solved by direct methods and refined by full-
matrix least squares on F2 using SHELXTL.90 All of the non-hydrogen
atoms were refined with anisotropic displacement coefficients. The
hydrogen atoms were assigned isotropic displacement coefficients U(H)
) 1.2U(C), and their coordinates were allowed to ride on their
respective carbons. Nearly 50% of the structure was found to be
disordered. The disorder was treated with a two-site model as follows:
[O(1), N(1), C(1), C(2), C(3), C(4), C(4A), C(4B), C(5), C(6), C(7),
C(8), C(8A), C(9), C(9A)] and [O(1*), N(1*), C(1*), C(2*), C(3*),
C(4*), C(4A*), C(4B*), C(5*), C(6*), C(7*), C(8*), C(8A*), C(9*),
C(9A*)]. The partial atoms at these two sites were assigned occupancy
factors of a half and the carbons refined with mild distance restraints.
The SQUEEZE-processed data were used in all subsequent cycles
of least squares. All of the nonhydrogen atoms were refined with
anisotropic displacement coefficients. The hydrogen atoms were
assigned isotropic displacement coefficients U(H) ) 1.2U(C), and their
coordinates were allowed to ride on their respective carbons. The
weighting scheme employed was w ) 1/[σ2(Fo ) + (0.0838P)2], where
2
2
P ) (Fo + 2Fc2)/3. The refinement converged to R(F) ) 0.0571,
wR(F2) ) 0.1490, and S ) 1.32 for 7272 reflections with I > 2σ(I),
and R(F) ) 0.1155, wR(F2) ) 0.1681, and S ) 1.07 for 13 775 unique
reflections and 649 parameters. The maximum ∆/σ in the final cycle
of least squares was less than 0.001, and the residual peaks on the
final difference Fourier map ranged from -0.298 to 0.443 eÅ-3
.
The weighting scheme employed was w ) 1/[σ2(Fo ) + (0.0551P)2 +
2
0.7277P], where P ) (Fo2 + 2Fc2)/3. The refinement converged to R(F)
) 0.0569, wR(F2) ) 0.1407, and S ) 1.20 for 1948 reflections with I
> 2σ(I), and R(F) ) 0.0772, wR(F2) ) 0.1529, and S ) 1.09 for 2622
unique reflections, 416 parameters, and 72 restraints. The maximum
∆/σ in the final cycle of leas -squares was less than 0.001, and the
residual peaks on the final difference Fourier map ranged from -0.094
to 0.160 eÅ-3. Scattering factors were taken from the International
Tables for Crystallography, Volume C.91,92
Scattering factors were taken from the International Tables for
Crystallography, Volume C.91,92
Titration of Fe(OH)TMP with Fluorenone Oxime. To 0.8 mL of
a benzene-d6 solution of Fe(OH)TMP (2.4 mM) was added small
aliquots of solid fluorenone oxime. A 1H NMR spectrum of the sample
was taken after thorough mixing. The spectrum showed that the
resonance signals of Fe(OH)TMP changed to those of 2 within minutes,
and the amount of 2 generated was proportional to that of the oxime
added. The conversion was complete upon the addition of 1 equiv of
fluorenone oxime. Further addition of fluorenone oxime led to the
emergence of the signals of free fluorenone oxime without interfering
signals of any other species.
Oxidation of Fe(oximate)TMP (2) at High O2 Pressures. A high-
pressure reaction vessel was made from stainless steel by our
departmental machine workshop. The reactor contains a 30 mL Teflon
liner, and the high-pressure seal was achieved by pressing an O-ring
between the stainless steel screw-cap and the Teflon liner. The pressure
gauge on top of the reactor monitoring the presure in the liner and two
separate gas inlet and outlet values allow the control of the gas content
inside the Teflon liner.
X-ray Crystal Structure Determination of Fe(oximate)TDCPP
(3). Single crystals of 3 suitable for X-ray crystallographic analysis
were obtained by slow diffusion of n-heptane into a chlorobenzene
solution of 3 over 3 days. A dark purple prism cut to 0.12 mm × 0.28
mm × 0.35 mm in size was used for the diffraction experiment (Table
1). A total of 493 frames of data were collected at 298(2) K with an
(88) COLLECT Data Collection Software. Nonius B. V. Rontgenweg 1
P.O. Box 811, 2600 AV, Delft, The Netherlands, 1998.
(89) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307.
(90) Sheldrick, G. M. SHELXTL Version 5.04.; Siemens Analytical X-ray
Instruments: Madison, WI, 1996.
(91) Maslen, E. N.; Fox, A. G.; O’Keefe, M. A. In International Tables
for Crystallography: Mathematical, Physical and Chemical Tables; Wilson,
A. J. C., Ed: Kluwer: Dordrecht, The Netherlands, 1992; Vol. C, p 476.
(92) Creagh, D. C.; McAuley, W. J. In International Tables for
Crystallography: Mathematical, Physical and Chemical Tables; Wilson,
A. J. C., Ed: Kluwer: Dordrecht, The Netherlands, 1992; Vol. C, p 206.
A 2 mL silanized glass vial containing a 0.8 mL benzene-d6 solution
of 0.75 mM 2 was enclosed in the high-pressure reactor. The reaction
(93) Van der Sluis, P.; Spek, A. L. Acta Crystallogr., A 1990, 46, 194.
(94) Spek, A. L. Acta Crystallogr., A 1990, 46, C34.

Models of Nitric Oxide Synthase
J. Am. Chem. Soc., Vol. 121, No. 51, 1999 12103
mixture was stirred with a magnetic stirrer. The oxidation was started
by introducing various pressures of O2 (>99.8% purity, from BOC gas)
into the reactor. After a certain length of reaction time, the reactions
was stopped by venting the O2 from the reactor, and the content in the
vial was analyzed with 1H NMR spectroscopy. The reactions were
conducted at 100, 200, 330, and 500 psi O2.
then once each half hour for 12 h. The experiment was conducted at
30, 40, and 50 °C.
Aerobic Oxidation of Fe(NO)TMP. A benzene-d6 solution of 1
mM Fe(NO)TMP was sealed in an NMR tube under 14.7 psi O2. The
1
sample was kept in a 50 °C bath in the dark and analyzed with H
1
NMR spectroscopy every 12 h. The H NMR spectrum showed that
Oxidation of Fe(oximate)TMP (2) under 18O2. The 18O2 (98%
enriched, from Cambridge Isotopes Laboratories) experiment was
conducted on a high-vacuum Schlenk line. Into a 80 mL silanized flask
was added a 50 mL benzene solution of 20 µM 2, which was deaeriated
(via three freeze-pump-thaw cycles) to remove the dissolved gas.
Then 250 mL of 1 atm 18O2 was introduced into the flask and condensed
with liquid N2 cold bath. The flask was then sealed with a J-Young
valve, and the solution was warmed to 25 °C and stirred in the dark.
The final O2 pressure in the reaction flask was measured with a pressure
gauge to be 100 ( 5 psi. The reaction was stopped 4 days later by
releasing the O2 pressure, and the solvent was removed by evacuation.
Fe(NO)TMP converted gradually via a 5-coordinate high-spin Fe(III)-
TMP species to Fe(NO3)TMP (t1/2 ≈ 2 day). The new Fe(III)TMP
species gave resonance signals at δ 76 (8H, pyrrole H), 13.3 and 14.6
(8H, m-phenyl H), 3.8 (12H, p-methyl H), and 2-6 (24H, o-methyl
H).34 The same reaction conducted at 25 °C showed that Fe(NO)TMP
was oxidized much slower (t1/2 ≈ 1 week). When the reaction was
conducted under 500 psi O2 at 25 °C, the half-life of Fe(NO)TMP was
shortened to 10 h.
Preparation and Characterization of the Putative Fe(ONO)TMP
Complex. Two milliliters (100-fold excess) of NO(g) was introduced
to a 2.5 mM toluene-d8 solution of Fe(OH)TMP in an air-free NMR
tube. The solution immediately changed from dark green to bright red
upon contact with NO. After the NMR tube was sealed with a J-Young
valve, the sample was analyzed with 1H NMR spectroscopy. In addition
to Fe(NO)TMP, the 1H NMR spectrum showed that a diamagnetic Fe-
(II)TMP species was the only other product, with resonance signals at
δ 1.75 (s, 12H), 2.0 (s, 12H), 2.5 (s, 12H), and 8.8 (s, 8H, pyrrole-
H).34
1
The solid product was dissolved in benzene-d6 and analyzed with H
NMR spectroscopy, which showed that all 2 was consumed. The NMR
sample was reduced to dryness and then triturated with acetonitrile to
remove the iron porphyrins. The fluorenone-containing solid residue
was redissolved in CH2Cl2 and analyzed with GC-MS. The relative
ion abundances of the fluorenone peak in the mass spectrum are as
follows: m/z (relative intensity) 179 (0.0), 180 (30.5), 181 (5.9), 182
(100.0), 183 (15.2), 184 (1.0). For comparison, the mass spectrum of
fluorenone generated in the 16O2-effected oxidation of 2 showed the
following ratios: 179 (0.4), 180 (100.0), 181 (13.8), 182 (0.8), 183
(0).
This diamagnetic Fe(II)TMP species was unstable and tended to lose
one ligand upon extensive evacuation. After evacuating the sample to
<1 µmHg vacuum for 2 days, the 1H NMR spectrum of the sample in
anaerobic benzene-d6 revealed that the diamagnetic Fe(II)P species had
changed to a new 5-coordinate high-spin Fe(III)TMP species with
resonance signals at δ 76 (8H, pyrrole H), 13.3 and 14.6 (8H, m-phenyl
H), 3.8 (12H, p-methyl H), and 2-6 (24H, o-methyl H). By contrast,
Fe(NO)TMP remained unchanged upon extensive evacuation. Upon
exposing the sample to O2 for a few days, the 1H NMR spectrum showed
that the new Fe(III)TMP species and Fe(NO)TMP were both converted
to Fe(NO3)TMP. This new 5-coordinate high-spin Fe(III)TMP species
was tentatively assigned as Fe(ONO)TMP.
Aerobic Oxidation of Fluorenone Oxime. A 0.75 mM benzene-d6
solution of fluorenone oxime was purged with and kept under 1 atm
1
O2 at 35 °C in the dark. The reaction mixture was analyzed by H
1
NMR spectroscopy every 24 h for 2 weeks. The H NMR spectrum
showed that fluorenone oxime was oxidized in 1 week to 1, which
gradually decomposed to a one-to-one ratio of fluorenone and 9,9-
dinitrofluorene over another week.
Reaction of Fe(oximate)TMP (2) and CO(g). (a) UV-Visible
Experiment. A 3 mL benzene solution of 7 µM 2 in a cuvette was
purged with CO(g) for 5 min at 10 °C. The cuvette was equipped with
an enlarged headspace (∼80 mL). The cuvette was sealed with a
J-Young valve under 1 atm CO(g), and the sample was warmed to room
temperature and analyzed with UV-visible spectroscopy immediately.
The UV-visible spectrum of the mixture showed that the absorbance
due to 2 was gradually converted to λmax at 410, 426 (sh), 502, and
510 nm, which were the same as those of Fe(CO)TMP/Fe(CO)2TMP.
(b) Kinetic Experiment for the Eyring Plot. To avoid the
interference of irradiation on the product ratio, the same reaction mixture
described above was prepared in the dark. A modified Cary 2390 UV-
visible spectrophotometer was used to monitor the progress of the
reaction. Located between the light source and the sample cuvette in
the spectrophotometry was an electrical shutter controlled with a manual
switch. The shutter was open for 2 s at each measurement. The
absorbance at 426 nm was measured every 3 min for the first 2 h and
Aerobic Oxidation of Fe(NO)TMP. A benzene-d6 solution of 1
mM Fe(NO)TMP was sealed in an NMR tube under 14.7 psi O2. The
1
sample was kept in a 50 °C bath in the dark and analyzed with H
1
NMR spectroscopy every 12 h. The H NMR spectrum showed that
Fe(NO)TMP converted gradually via the putative Fe(ONO)TMP species
to Fe(NO3)TMP (t1/2 ≈ 2 day). The same reaction conducted at 25 °C
showed that Fe(NO)TMP was oxidized more slowly (t1/2 ≈ 1 week).
When the reaction was conducted under 500 psi O2 at 25 °C, the half-
life of Fe(NO)TMP was shortened to 10 h.
Acknowledgment. Financial support of this research from
the National Institutes of Health (Grant GM36928) is gratefully
acknowledged.
JA992373+