Mechanism of inactivation of inducible nitric oxide synthase by amidines. Irreversible enzyme inactivation without inactivator modification

Article abstract of DOI:10.1021/ja0445645

Nitric oxide synthases (NOS) are hemoproteins that catalyze the reaction of L-arginine to L-citrulline and nitric oxide. N-(3-(Aminomethyl)benzyl) acetamidine (1400W) was reported to be a slow, tight-binding, and highly selective inhibitor of iNOS in vitr

Full text of DOI:10.1021/ja0445645

Published on Web 12/30/2004
Mechanism of Inactivation of Inducible Nitric Oxide Synthase
by Amidines. Irreversible Enzyme Inactivation without
Inactivator Modification
Yaoqiu Zhu,† Dejan Nikolic,‡ Richard B. Van Breemen,‡ and Richard B. Silverman*,†
Contribution from the Department of Chemistry, Department of Biochemistry, Molecular
Biology, and Cell Biology, and Drug DiscoVery Program, Northwestern UniVersity, EVanston,
Illinois 60208-3113, and Department of Medicinal Chemistry and Pharmacognosy, UniVersity of
Illinois at Chicago, Chicago, Illinois 60612-7231
Received September 8, 2004; E-mail: Agman@chem.northwestern.edu
Abstract: Nitric oxide synthases (NOS) are hemoproteins that catalyze the reaction of L-arginine to L-citrulline
and nitric oxide. N-(3-(Aminomethyl)benzyl)acetamidine (1400W) was reported to be a slow, tight-binding,
and highly selective inhibitor of iNOS in vitro and in vivo. Previous mechanistic studies reported that 1400W
was recovered quantitatively after iNOS fully lost its activity and modification to iNOS was not detected.
Here, it is shown that 1400W is a time-, concentration-, and NADPH-dependent irreversible inactivator of
iNOS. HPLC-electrospray mass spectrometric analysis of the incubation mixture of iNOS with 1400W shows
both loss of heme cofactor and formation of biliverdin, as was previously observed for iNOS inactivation by
another amidine-containing compound, N5-(1-iminoethyl)-L-ornithine (L-NIO). The amount of biliverdin
produced corresponds to the amount of heme lost by 1400W inactivation of iNOS. A convenient MS/MS-
HPLC methodology was developed to identify the trace amount of biliverdin produced by inactivation of
iNOS with either 1400W or L-NIO to be biliverdin IXR out of the four possible regioisomers. Two mechanisms
were previously proposed for iNOS inactivation by L-NIO: (1) uncoupling of the heme peroxide intermediate,
leading to destruction of the heme to biliverdin; (2) abstraction of a hydrogen atom from the amidine methyl
group followed by attachment to the heme cofactor, which causes the enzyme to catalyze the heme
oxygenase reaction. The second mechanistic proposal was ruled out by inactivation of iNOS with d3-1400W,
which produced no d2-1400W. Detection of carbon monoxide as one of the heme-degradation products
further excludes the covalent heme adduct mechanism. On the basis of these results, a third mechanism
is proposed in which the amidine inactivators of iNOS bind as does substrate L-arginine, but because of
the amidine methyl group, the heme peroxy intermediate cannot be protonated, thereby preventing its
conversion to the heme oxo intermediate. This leads to a change in the enzyme mechanism to one that
resembles that of heme oxygenase, an enzyme known to convert heme to biliverdin IXR. This appears to
be the first example of a compound that causes irreversible inactivation of an enzyme without itself becoming
modified in any way.
Introduction
synthase (nNOS), endothelial nitric oxide synthase (eNOS), and
inducible nitric oxide synthase (iNOS), which have different
Nitric oxide, which is produced by the enzyme nitric oxide
synthase (NOS, EC 1.14.13.39), is an important cell-signaling
agent.1 An explosion of research in this area has demonstrated
that nitric oxide fulfills a large range of biological functions as
both a messenger and a cytotoxic factor.2 Because of the harmful
effects of excess NO, inhibition of NO biosynthesis has been
an important approach to the design of new potential drugs to
treat those diseases caused by NO overproduction. However,
NOS inhibition may also be detrimental to the essential functions
of nitric oxide. This dilemma can be alleviated by selective
inhibition of the three NOS isoforms, neuronal nitric oxide
functions.
NOS converts L-arginine (1) and O2 to L-citrulline (2) and
NO (3) with concomitant oxidation of NADPH (Scheme 1).
The three isoforms of nitric oxide synthase are homodimeric
hemoproteins incorporating several domains in a single polypep-
tide: (a) a heme domain that also binds a tetrahydrobiopterin
cofactor and is the site of oxidation of L-arginine to nitric oxide
and L-citrulline,3 (b) a P450 reductase-like domain that has
binding sites for FMN, FAD, and NADPH and is responsible
for providing electrons to the heme domain,4 and (c) a
connecting peptide between the heme and flavin domains that
binds calmodulin in a Ca2+-dependent manner for the endothelial
† Northwestern University.
‡ University of Illinois at Chicago.
(3) Lambert, L. E.; French, J. F.; Whitten, J. P.; Baron, B. M.; McDonald, I.
A. Eur. J. Pharmacol. 1992, 216, 131-134.
(4) Bredt, D. S.; Hwang, P. M.; Glatt, C. E.; Lowenstein, C.; Reed, R. R.;
Snyder, S. S. Nature 1991, 351, 714-718.
(1) Kerwin, J. F.; Heller, M. Med. Res. ReV. 1994, 14, 23-74.
(2) Kerwin, J. F.; Lancaster, J. R.; Feldman, P. L. J. Med. Chem. 1995, 38,
4343-4362.
9
858
J. AM. CHEM. SOC. 2005, 127, 858-868
10.1021/ja0445645 CCC: $30.25 © 2005 American Chemical Society

Inactivation of Inducible Nitric Oxide Synthase
A R T I C L E S
Scheme 1
(eNOS) and neuronal (nNOS) forms and an essentially Ca2+
independent manner in the inducible enzyme (iNOS) because
it binds Ca2+ so tightly it is insensitive to changes in cellular
Ca2+ levels.5
Garvey and co-workers reported 1400W (4) to be a slow,
tight-binding, and highly selective
-
8), which differs from 1400W by only one methylene unit but
is nNOS selective,12 and N,N′-(1,3-benzyl)bisamidine (9).6,12
The recovery of 1400W with slowly reversible or irreversible
iNOS activity is reminiscent of the mechanistic behavior of
L-NIO7 (5). L-NIO was both a substrate and inactivator for iNOS.
After incubation with iNOS, most of the L-NIO was recovered;
a minor part was converted to NG-hydroxy-L-NIO (10) and
L-acetylornithine (11), which were irrelevant to the L-NIO
inactivation mechanism.7 Complete inactivation of iNOS by
L-NIO occurred concomitant with loss of heme cofactor from
the homodimer and formation of biliverdin (12).
inhibitor of iNOS in vitro and in vivo.6 1400W was the most
selective inhibitor of purified human iNOS reported to date
(5000- and 2000-fold more potent against purified human iNOS
than eNOS and nNOS, respectively). Inhibition of human iNOS
by 1400W did not reverse upon dilution; consequently, 1400W
was initially concluded to be either a slowly reversible inhibitor
or an irreversible inactivator. However, additional mechanistic
studies also established that 1400W was quantitatively recov-
ered, and no modification of iNOS could be detected, supporting
reversible inhibition. 1400W is one of several reported amidine-
containing inhibitors of NOS, including N5-(1-iminoethyl)-L-
ornithine (L-NIO, 5),7 a natural product8 known to be a substrate
and an inactivator of
Biliverdin is the major product of the heme-degradation
reaction catalyzed by the enzyme heme oxygenase;13 however,
there was no heme oxygenase in the iNOS preparation. Two
mechanisms to rationalize the formation of biliverdin that were
proposed earlier7 were (1) L-NIO attaches to the meso position
of the heme (13), which changes the chemistry to that
resembling heme oxygenase, producing biliverdin (12) via
verdoheme (14) (Scheme 2) and (2) the lack of a proton source
in the active site prevents the conversion of peroxoheme to the
iron oxo species, releasing excess peroxide, leading to R-meso-
hydroxyheme (Scheme 3). The first mechanism was based on
a variety of experiments using [14C]- and [2H]-labeled L-NIO,
especially the small deuterium isotope effect observed on
inactivation by trideuteriomethyl-L-NIO (15), suggesting a
possible C-H bond
iNOS, N5-(1-iminoethyl)-L-lysine (L-NIL; homo- L-NIO, 6),
which inactivates iNOS by modifying its heme cofactor and
destabilizing the iNOS homodimer into inactive monomers,9,10
2-amino-6-(1-imino-2-fluoroethylamino)-4,4-dioxo-4-thiohex-
anoic acid (7),11 N-(3-(aminomethyl)phenyl)acetamidine (5320W,
cleavage during inactivation.7 The intriguing aspect of the
first mechanism is that it accounts for little loss of inactivator
because the inactivator is regenerated by spontaneous decar-
boxylation of the proposed product. The first mechanism also
is supported by the observation that R-meso-methylheme is a
substrate for heme oxygenase and is converted to biliverdin;14
(5) (a) Abu-Soud, H. M.; Yoho, L. L.; Stuehr, D. J. J. Biol. Chem. 1994, 269,
32047-32050. (b) Sheta, E. A.; McMillan, K.; Siler-Masters, B. S. J. Biol.
Chem. 1994, 269, 15147-15153.
(6) Garvey, E. P.; Oplinger, J. A.; Furfine, E. S.; Kiff, R. J.; Laszlo, F.; Whittle,
B. J. R.; Knowles, R. G. J. Biol. Chem. 1997, 272, 4959-4963.
(7) Fast, W.; Nikolic, D.; Van Breemen, R.; Silverman, R. B. J. Am. Chem.
Soc. 1999, 121, 903-916.
(8) Scannell, J. P.; Ax, H. A.; Pruess, D. L.; Williams, T.; Demny, T. C.;
Stempel, A. J. Antibiot. 1972, 25, 179-184.
(9) Moore, W. M.; Webber, R. K.; Jerome, G. M.; Tjoeng, F. S.; Misko, T.
P.; Currie, M. G. J. Med. Chem. 1994, 37, 3886-3888.
(12) Collins, J, L.; Shearer, B. G.; Oplinger, J. A.; Lee, S.; Garvey, E. P.; Salter,
M.; Duffy, C.; Burnette, T. C.; Furfine, E. S. J. Med. Chem. 1998, 41,
2858-2871.
(13) Ortiz de Montellano, P. R. Curr. Opin. Chem. Biol. 2000, 4, 221-227.
(14) Torpey, J.; Ortiz de Montellano, P. R. J. Biol. Chem. 1996, 271, 26067-
26073.
(10) (a) Grant, S. K.; Green, B. G.; Stiffey-Wilusz, J.; Durette, P. L.; Shah, S.
K.; Kozarich, J. W. Biochemistry 1998, 37, 4174-4180. (b) Bryk, R.; Wolff,
D. J. Biochemistry 1998, 37, 4844-4852.
(11) Macdonald, J. E.Annual Reports in Medicinal Chemistry; Academic Press:
San Diego, 1996; Vol. 31, pp 221-230.
9
J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005 859

A R T I C L E S
Scheme 2
Zhu et al.
Table 1. KI and kinact Values for 1400W and L-NIO with iNOS
1400W
L-NIO
KI (µM)
k
k
23.8 ( 1.04
2.40 ( 0.28
0.10 ( 0.01
13.7 ( 1.60
inact (min-1
inact/KI (µM-1 min-1
)
0.073 ( 0.003
)
0.0053 ( 0.0007
first proposed mechanism above is not reasonable but a heme
oxygenase-like mechanism is reasonable.
Results
Inactivation of iNOS by 1400W. As in L-NIO, 1400W was
a time-, concentration-, and NADPH-dependent inactivator of
iNOS; 1400W is a more efficient inactivator than L-NIO (Table
1).
Kinetic Isotope Effect on Inactivation of iNOS by N-(3-
(Aminomethyl)benzyl)-[2H3]-acetamidine (12). N-(3-(Amino-
methyl)benzyl)-[2H3]-acetamidine (16) was synthesized from
CD3CN via 2-[2H3]-acetimidic acid ethyl ester (Scheme 4).
Inactivation by 16 shows a kinetic isotope effect on Hkinact/Dkinact
of 1.33 ( 0.18 and on H(kinact/KI/)/D(kinact/KI) of 1.39 ( 0.20
(Table 2 and Supporting Information).
Heme Loss during Inactivation by 1400W. iNOS incubation
with 1400W and all necessary cofactors was subjected to high-
performance liquid chromatography (HPLC) analysis after three
hours (Figure 1). The heme cofactor appears to be involved in
the inactivation event because the peak area for the heme
cofactor decreased with concomitant loss of enzyme activity.
A control solution made by either omitting NADPH or 1400W
maintained most of its activity. Inclusion of catalase in the
inactivation mixtures decreased the amount of heme loss in the
control solutions but did not diminish significant loss of heme
(40%) in the solution containing both 1400W and NADPH.
Formation of a New Peak after Inactivation. The HPLC
trace after iNOS inactivation also reveals a new NADPH- and
1400W-dependent peak at 6.5 min, which is consistent with
the retention time for commercial biliverdin IXR (Figure 1).
The formation of the new peak also occurs in a time-dependent
manner (data not shown).
Scheme 3
Characterization of the New Peak from 1400W Inactiva-
tion. LC-electrospray mass spectrometry was used to compare
the control (no 1400W), inactivation (with 1400W), and
standards of commercial biliverdin IXR and hemin (ferripro-
toporphyrin IX chloride). The mass spectra of the peaks in the
LC-MS traces (see Supporting Information) at 14.4 and 16.4
min from the inactivation mixture are shown in Figure 2. The
m/z values for the 14.4 min peak and 16.4 min peak are 583
(Figure 2A) and 616 (Figure 2B), respectively. MS/MS of the
14.4 min peak in the inactivation trace gave a spectrum (Figure
3A), which is nearly identical with that of the commercial
biliverdin IXR (Figure 3B). Both MS/MS spectra show a base
peak with m/z of 297.3. The 297.3 peak can be rationalized by
a “half-half” fragmentation pattern that is not consistent with
that expected for the â- and δ-isomers of biliverdin (Figure 4).
Identification of the Biliverdin Regioisomer Produced
from Inactivation of iNOS using MS/MS Analysis.
presumably, 13 (Scheme 2) could function the same as R-meso-
methylheme.
In the second mechanism, the heme cofactor of NOS is
hydroxylated at the R-meso-position to form R-meso-hydroxy-
heme, which is then decomposed to biliverdin and CO. The
second mechanism initially seemed less appealing because it
does not appear to rationalize the deuterium isotope effect on
inactivation.
Here, we present experiments to demonstrate that 1400W and
L-NIO apparently inactivate iNOS by the same mechanism, that
the biliverdin that is formed during inactivation is one regioi-
somer, namely, biliverdin IXR, the same isomer that is produced
by heme oxygenase-catalyzed oxidation of heme, and that the
9
860 J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005

Inactivation of Inducible Nitric Oxide Synthase
A R T I C L E S
Scheme 4
Table 2. Kinetic Isotope Effect on INOS Inactivation by 1400W
and d3-1400W
have the same retention time as that of the R-isomer reference
compound (the tiny peak at 29.3 min is the heme cofactor).
1400W
d3-1400W
H/D
Quantification of the Formation of Biliverdin and Loss
of Heme. HPLC traces of biliverdin IXR and hemin standards
at different concentrations were recorded and peak areas were
calculated to make calibration curves (see Supporting Informa-
tion) to quantify the biliverdin formed from heme degradation
by iNOS inactivation with 1400W and L-NIO. The ratio of
biliverdin to heme loss was calculated to be 0.90 ( 0.04 for
iNOS inactivation by 1400W and 0.92 ( 0.06 by L-NIO.
kinact (min-1
KI (µM)
)
2.40 ( 0.28
23.8 ( 1.04
0.101 ( 0.012
1.80 ( 0.12
24.8 ( 0.62
1.33 ( 0.18
0.96 ( 0.05
k
inact/KI (µM-1 min-1
)
0.073 ( 0.005 1.39 ( 0.20
Determination of Deuterium Loss During Inactivation of
iNOS by d3-1400W. To determine if a C-H bond of the
amidine methyl group is broken during iNOS inactivation, as
proposed in the alkylated heme mechanism (Scheme 2), d3-
1400W was incubated with iNOS. After the incubation mixture
fully lost its activity, the inactivator (heme:inactivator ) 1:5)
was separated from the enzyme by ultrafiltration and then
derivatized with OPA-DTT reagent (see Supporting Information)
to an OPA-DTT-d3-1400W adduct (M+H+ ) 433.6).15 By
comparing the calculated peak areas of the adducts in the HPLC
traces of inactivation (containing enzyme) and control (no
enzyme), the total amount of inactivator was fully recovered.
LC-APCI mass spectrometry was used to analyze the fully
recovered inactivator from iNOS inactivation in its adduct form
(M+H+ ) 433.6). The mass spectrum showed no peak with a
mass value of 432.6, which would have corresponded to OPA-
DTT- d2-1400W plus a proton (see Supporting Information).
Detection of CO as One of the Heme Metabolites. Freshly
prepared ferrous deoxymyoglobin was used as the trapping
reagent to detect the possible formation of CO as an oxidation
product of heme during inactivation of iNOS by L-NIO and
1400W. iNOS inactivation by L-NIO and 1400W produced a
shift of the Soret maximum from 430 nm (ferrous deoxymyo-
globin) to 418 nm (Figure 7), which indicates complete
complexation of CO with the ferrous ion.14 The positive control
(replacing the incubation mixture with CO-saturated buffer) and
the negative control (without inactivators) confirmed this result.
Figure 1. HPLC of heme loss and formation of a new peak (6 min) after
iNOS inactivation by 1400W. (a) Control (without 1400W); (b) inactivation
(with 1400W); (c) biliverdin and heme standards.
The four biliverdin IX regioisomers were synthesized24,38 and
subjected to MS/MS analysis under conditions used in the
characterization of the new peak during iNOS inactivation. The
MS/MS analyses (Figure 5) show that the R-, â-, γ-, and
δ-isomers have very different fragmentation patterns, and the
â-, γ-, and δ-isomers do not give a fragmentation with m/z of
297.2. Therefore, the unknown biliverdin from iNOS inactiva-
tion by both 1400W and L-NIO is biliverdin IXR.
Identification of the Biliverdin Regioisomer Produced
from Inactivation of iNOS by HPLC Analysis. The four
biliverdin IX regioisomers were separated by reversed-phase
HPLC with 370 nm wavelength detection in the order of R,
â+δ, γ with retention times of 24.7, 27.4, and 35.6 min,
respectively (Figure 6). The biliverdin peaks in the HPLC traces
of the incubation of iNOS inactivation by 1400W and L-NIO
Discussion
Our earlier mechanistic studies demonstrated that L-NIO
inactivates iNOS by degrading its heme cofactor into biliverdin
but with no apparent modification to L-NIO itself.7 Likewise,
the highly selective, non-amino acid-based, amidine-containing
iNOS inhibitor, 1400W, was reported to be a slowly reversible
inhibitor or irreversible inactivator, but no chemical modification
of the enzyme or of 1400W was detected after iNOS inactiva-
(15) Jones, B. N.; Paabo, S.; Stein, S. J. Liq. Chromatogr. 1981, 4, 565-86.
9
J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005 861

A R T I C L E S
Zhu et al.
Figure 2. Mass spectra of the 14.4 min (6.5 min in Figure 1b) peak A and the 16.4 min (20.5 min in Figure 1b) peak B in the inactivation LC trace.
both amidine-containing inhibitors have a common mechanism
of action and to characterize the enzyme modification product
further, we investigated the inhibition of iNOS by 1400W.
1400W was a time-, concentration-, and NADPH-dependent
inhibitor of iNOS. In fact, the kinact/KI value for 1400W is about
20 times larger than that for L-NIO (Table 1). Also, as was found
earlier for L-NIO, N-(3-(aminomethyl)benzyl)-2-[2H3]-acetami-
dine (d3-1400W) inactivated iNOS with a kinetic deuterium
isotope effect on kinact/KI of 1.39 ( 0.2 (Table 2). Initially, it
was not clear if this was a small primary kinetic isotope effect
or a large secondary kinetic isotope effect (see below). L-NIO
caused inactivation of iNOS with concomitant conversion of
its heme to biliverdin (12). HPLC analysis was used to detect
possible metabolites of 1400W-inactivated iNOS. Unlike that
reported previously,6 the heme cofactor of iNOS was diminished
with a concomitant decrease in enzyme activity, suggesting that
the heme cofactor was involved in the inactivation event (Figure
1). A new time-dependent, NADPH-dependent, and 1400W-
dependent small peak was found by HPLC with the same
retention time as that of a biliverdin standard (Figure 1), and
LC-MS analysis confirmed the new peak to be biliverdin
(Figures 2 and 3). On the basis of HPLC peak areas, it was
calculated that 0.90 ( 0.04 equivalent of biliverdin was formed
per 1.00 equivalent of heme loss during iNOS inactivation by
1400W, and 0.92 ( 0.06 equivalent of biliverdin was formed
per 1.00 equivalent of heme loss during iNOS inactivation by
L-NIO (calibration curves were made using commercial hemin
and biliverdin IXR after the identification of the unknown
biliverdin). These results support 1400W as an iNOS inactivator,
which acts very similarly to that of L-NIO. Previously, we found
that NG-allyl-L-arginine inactivates NOS by allylation and
reduction of the heme,16 indicating that nonamidine-containing
Figure 3. (A) MS/MS of the m/z 583 peak in Figure 2A. (B) MS/MS of
biliverdin IXR standard.
inactivators do not modify the heme cofactor by conversion to
biliverdin. This demonstrates that amidines inactivate NOS by
tion.6 A comparison of the structures of 1400W and L-NIO
suggests that they could bind to and inhibit iNOS similarly.
a unique mechanism.
However, our results with L-NIO and those of Garvey and co-
workers6 with 1400W appear to indicate that the mechanisms
of action of these two inhibitors are different. To determine if
(16) Zhang, H. Q.; Dixon, R. P.; Marletta, M. A.; Nikolic, D.; Van Breemen,
R.; Silverman, R. B. J. Am. Chem. Soc. 1997, 119, 10888-10902.
9
862 J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005

Inactivation of Inducible Nitric Oxide Synthase
A R T I C L E S
Figure 4. Expected m/z values for splitting of biliverdin IX R, â, γ, and δ.
Figure 5. MS/MS analysis of the four biliverdin IX regioisomers.
The commercially available biliverdin that was used for
product identification with LC-MS was biliverdin IXR. How-
ever, the reaction could occur at any or all meso positions of
the heme, leading to biliverdin IXR, â, γ, or δ. R-Meso oxidation
specificity is a key feature in the heme oxygenase-dependent
conversion of heme to biliverdin IXR,17 but the other three
isomers are found as heme metabolites in lower animals,18 in
body fluids of mammals,19 in the bile of mutant rats that cannot
esterify bilirubin IXR,20 with mutated heme oxygenase,21 and
when meso-substituted hemes are used as heme oxygenase
substrates.14,22 The MS/MS spectrum of the unknown biliverdins
from heme degradation had a base peak of 297 (Figure 3A),
which appears to represent half of the biliverdin product (Figure
4). To keep the mass values of the two “halves” of biliverdin
odd, the four nitrogen atoms have to be divided into two and
two. A comparison of the structures of the four possible
(17) Noguchi, M.; Yoshida, T.; Kikuchi, G. J. Biochem. 1982, 91, 1479-1483.
(18) (a) McDonagh, A. F. The Porphyrins; Academic Press: New York, 1979;
Vol. 6, Part A, pp 293-491. (b) Rudiger, W. Prog. Chem. Org. Nat. Prod.
1971, 29, 60-139. (c) Kayser, H.; Dettner, K. Comp. Biochem. Physiol.,
B: Comp. Biochem. 1984, 77, 639-643.
(19) (a) O’Carra, P.; Colleran, E. J. Chromatogr. 1970, 50, 458-468. (b)
Blumenthal, S. G.; Taggart, D. B.; Rasmussen, R. D.; Ikeda, R. M.; Ruebner,
B. H.; Bergstrom, D. E.; Hansen, F. W. Biochem. J. 1979, 179, 537-547.
(20) Blanckaert, N.; Fevery, J.; Heirweg, K. P. M. Compernolle, F. Biochem.
J. 1977, 164, 237-249.
(21) Caignan, G. A.; Deshmukh, R.; Wilks, A.; Zeng, Y.; Huang, H.; Moe¨nne-
Loccoz, P.; Bunce, R. A.; Eastman, M. A.; Rivera, M. J. Am. Chem. Soc.
2002, 124, 14879-14892.
(22) Torpey, J.; Ortiz de Montellano, P. R. J. Biol. Chem. 1997, 272, 22088-
22014.
9
J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005 863

A R T I C L E S
Zhu et al.
Figure 6. HPLC analysis (370 nm) of the biliverdins from iNOS
inactivation. (a) Four synthetic biliverdin IX regioisomers; (b) biliverdin
from iNOS inactivation with 1400W (the tiny peak at 29.3 min is heme);
(c) biliverdin from iNOS inactivation with L-NIO.
Figure 7. Formation of CO during iNOS inactivation by L-NIO or 1400W
using deoxymyoglobin as the trapping reagent.
The R-regiospecificity of the heme conversion to biliverdin
by iNOS incubation with amidines is the same as that observed
in the heme oxygenase-catalyzed conversion of substrate heme
and suggests that a heme oxygenase-like mechanism for iNOS
inactivation by 1400W and L-NIO may be relevant. Because
both heme and R-meso-methylheme are substrates for heme
oxygenase and both produce biliverdin IXR,14 we previously
proposed two possible inactivation mechanisms for L-NIO
inactivation of iNOS.7 In one of the inactivation mechanisms,
the amidine group of the inactivator is first oxidized by NOS,
which then covalently binds to the R-meso-position of the heme
to form an R-meso-alkylated heme (Scheme 2). This change in
the NOS oxygenase domain converts it into an “NOS heme-
oxygenase domain” which is capable of catalyzing the degrada-
tion of this R-meso-alkylated heme to biliverdin similarly to
the reaction of R-meso-methylheme with heme oxygenase. This
reaction would convert the R-meso-1400W-heme or R-meso-
L-NIO-heme into verdoheme (14) and carboxylated 1400W or
carboxylated L-NIO. These â-imino carboxylic acids should
undergo rapid and spontaneous decarboxylation to regenerate
1400W or L-NIO, which would account for why we7 and others6
have found that these inactivators do not appear to be modified
during inactivation.
In the second of the proposed inactivation mechanisms, the
binding of the amidines to NOS results in uncoupling of the
hydrogen peroxide produced, which may somehow lead to the
formation of R-meso-hydroxyheme, decomposing to biliverdin
and CO, as in the case of heme oxygenase.
Whereas the observed deuterium isotope effects with CD3-
1400W and CD3-NIO are easily rationalized by the first
mechanism (Scheme 2), which involves initial methyl C-H
bond cleavage, it would be difficult to rationalize the deuterium
isotope effect by a mechanism that involved uncoupling of
hydrogen peroxide. Furthermore, previous studies argue against
the hydroperoxide overproduction mechanism: (1) Although
hydrogen peroxide degrades free heme and cytochrome P450
bound heme, the products of peroxidative heme cleavage
consisted of various propentdyopents, maleimides, formic acid
(the meso-carbons), and hematinic acid; biliverdin, however,
biliverdin IX isomers indicates that to accommodate this base
peak, the â- and δ-isomers must be excluded.
The identification of the structure of the biliverdin formed
in heme oxygenase-catalyzed reactions normally can be done
in a relatively straightforward manner because the quantity of
biliverdin produced from heme as a substrate is sufficient for
identification by two-dimensional NMR21 or by conversion to
the corresponding dimethyl esters, which were identified by
comparison of HPLC retention times with the four synthetic
biliverdin IX dimethyl esters.14,23 The biliverdins produced in
our studies is the degradation products of the heme cofactor in
iNOS, and the small amount produced (about 0.1-2.0 µM) for
analysis leads to inconclusive results by NMR methods. Con-
sequently, our identification relied on mass spectral approaches.
The four biliverdin IX isomers were synthesized from a
coupled-oxidation reaction24 and were used as reference stan-
dards to identify the biliverdin produced by iNOS inactivation.
It is apparent from the mass spectra of the four biliverdin isomers
under the same analysis conditions (Figure 5) that the only
isomer whose mass spectrum matches the biliverdin produced
by 1400W and L-NIO inactivation of iNOS is biliverdin IXR
(although the γ-isomer also has the same fragmentation pattern
theoretically, it might require a different collision energy).
Furthermore, conditions were developed to separate the four
biliverdin isomers into three peaks by HPLC (Figure 6), and
confirmation was obtained that the biliverdin produced in the
enzyme inactivation is biliverdin IXR. The identification of the
regioisomer of the biliverdin produced from iNOS inactivation
by 1400W and L-NIO is important to the elucidation of the
proposed heme degradation mechanism and will be very helpful
for the elucidation of the interaction between the inactivators
and the active site of the enzyme.
(23) (a) Sakamoto, H.; Omata, Y.; Adachi, Y.; Palmer, G.; Noguchi, M. J. Inorg.
Biochem. 2000, 82, 113-121. (b) Yamaguchi, T.; Yamaguchi, N.;
Nakajima, H.; Komoda, Y.; Ishikawa, M. Proc. Jpn. Acad., Ser. B 1979,
2, 79-83. (c) Zhang, X.; Fujii, H.; Matera, M. K.; Migita, C. T.; Sato, M.;
Ikeda-Saito, M.; Yoshida, T. Biochemistry 2003, 42, 7418-7426.
(24) Bonnett, R.; McDonagh, A. F. J. Chem. Soc., Perkin Trans. 1 1973, 881-
888.
9
864 J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005

Inactivation of Inducible Nitric Oxide Synthase
A R T I C L E S
Scheme 5
Because the alkylated heme mechanism (Scheme 2) is not
supported by our results and the excess hydrogen peroxide
mechanism is not consistent with previous results that show
that hydrogen peroxide does not convert heme to biliverdin and
also does not rationalize the deuterium isotope effect on the
methyl C-H bond during inactivation, a third mechanistic
possibility is required. If the inactivation mechanism is related
to the mechanism of heme oxygenase-catalyzed conversion of
heme to biliverdin, then carbon monoxide should be the other
product.13 To test this, the inactivation of iNOS by either 1400W
or L-NIO was carried out in a capped vial in the presence of a
limiting amount of freshly prepared ferrous deoxymyoglobin
as the CO trapping reagent in a separate tube inside the vial.
As shown in Figure 7, CO is produced by iNOS inactivation
with both L-NIO and 1400W, which strongly supports a
mechanism related to the heme oxygenase mechanism. This also
confirms the conclusion that an alkylated heme is not involved
in the inactivation of iNOS by amidines, because heme
oxygenase converts R-methylheme to biliverdin, but without
production of CO.14
was not a product or intermediate in this process;25 (2)
tetrahydrobiopterin, one of the cofactors included in the
inactivation mixture, was an inhibitor of the superoxide forma-
tion process by a mechanism that does not cause inhibition of
NADPH consumption;26,27 (3) inclusion of catalase does not
prevent significant heme loss during iNOS incubation with
L-NIO, which suggests that heme loss is not related to high
concentrations of hydroperoxide.7
The alkylated heme mechanism (Scheme 2) can be tested.
As shown in Scheme 5, oxidation of CD3-1400W or CD3-NIO
removes one of the three methyl deuteriums prior to alkylation
of the heme. Following oxidation of the inactivator-bound heme
and formation of the carboxylated inactivator, decarboxylation
results in the formation of the starting inactivator (1400W or
NIO), but with only two deuteriums in the methyl group. Mass
spectrometry can determine if one deuterium was lost. The
amount of inactivator used in the experiment was only in slight
excess (heme:inactivater ) 1:5) to keep inactivation fast and
to amplify the possible H-D exchange for easy mass spectral
detection. Derivatization of the amidines after inactivation with
o-phthalaldehyde/dithiothreitol reagent made their detection
straightforward. LC-MS of the derivatized inactivators clearly
showed that H-D exchange does not accompany iNOS
inactivation; no CHD2-1400W or CHD2-NIO was detected. The
alkylated heme mechanism (Scheme 2) is not consistent with
this result and is disproved.
This finding is consistent with the crystal structure of eNOS-
L-NIO,28 which shows that the amidine methyl carbon of L-NIO
is 3.95 Å from the R-meso-carbon of heme. This is too far for
formation of the alkylated R-meso-heme required for the
mechanism in Scheme 2. For bond formation to occur, there
has to be a movement of the inactivator toward the heme R-meso
position. This movement would break the hydrogen bonding
between the amidine nitrogen atoms and the glutamic acid
residue (E363) of eNOS, which is a high-energy barrier to cross.
Furthermore, the hydrophobic area around the heme R-meso-
carbon would not be sufficiently spacious to accommodate the
inactivator-heme adduct.
It appears, then, that the inactivation of iNOS by amidines is
related to the mechanism of heme oxygenase oxygenation of
heme to biliverdin. The amidines have a strong structural
relationship to L-arginine, and crystal structures of L-NIO and
1400W bound to all isozymes of NOS28,29 show that they all
bind similarly. This suggests that possibly the abnormal behavior
of amidines is a result of both the absence of the guanidino
group of L-arginine as well as the presence of the amidino group
of L-NIO and 1400W.
Speculation about the role of the amidino group in NOS
inactivation begins with a comparison of various heme-
dependent enzyme mechanisms. Consider the mechanisms of
cytochrome P450, NOS, and heme oxygenase. All three of these
enzymes catalyze oxygenation of substrates, and all three are
believed to proceed via a heme peroxide intermediate (Figure
8). In cytochrome P450, an amino acid residue in the active
site protonates the distal peroxide oxygen atom, which leads to
the loss of water and the formation of the high-energy iron-
oxo species (known as compound I).30 This is the same
mechanism proposed for the NOS-catalyzed oxygenation of
L-arginine to Nω-hydroxy-L-arginine,31 except that the proton
donor was proposed to be the guanidino amino group of the
substrate L-arginine.28 This conclusion is based on the crystal
structure of the NO complex of eNOS with L-arginine bound,28
where it is apparent that the guanidino amino group is hydrogen
bonded to the distal oxygen atom of NO, and no other active
site proton donors are available. In the heme oxygenase, the
proton donor is an active site residue, but it is close to the
proximal peroxide oxygen rather than to the distal oxygen.32
This allows a reaction to occur between the heme porphyrin
ring and the distal peroxide oxygen, leading to R-meso-
hydroxyheme and then on to biliverdin via verdoheme.
(29) Fedorov, R.; Hartmann, E.; Ghosh, D. K.; Schlichting, I. J. Biol. Chem.
2003, 278, 45818-45825.
(25) Schaefer, W. H.; Harris, T. M.; Guengerich, F. P. Biochemistry 1985, 24,
3254-3263.
(30) Makris, T. M.; Davydov, R.; Denisov, I. G.; Hoffman, B. M.; Sligar, S. G.
Drug Metab. ReV. 2002, 34, 691-708.
(26) Vasquez-Vivar, J.; Hogg, N.; Martasek, P.; Karoui, H.; Pritchard, K. A.,
Jr.; Kalyanaraman, B. J. Biol. Chem. 1999, 274, 26736-26742.
(27) Heinzel, B.; John, M.; Klatt, P.; Bohme, E.; Mayer, B. Biochem. J. 1992,
281, 627-630.
(31) Marletta, M. A.; Hurshman, A. R.; Rusche, K. M. Curr. Opin. Chem. Biol.
1998, 2, 656-663.
(32) (a) Davydov, R.; Kofman, V.; Fujii, H.; Yoshida, T.; Ikeda-Saito, M.;
Hoffman, B. M. J. Am. Chem. Soc. 2002, 124, 1798-808. (b) Davydov,
R.; Matsui, T.; Fujii, H.; Ikeda-Saito, M.; Hoffman, B. M. J. Am. Chem.
Soc. 2003, 125, 16208-16209.
(28) Li, H.; Raman, C. S.; Martasek, P. Masters, B. S. S.; Poulos, T. L.
Biochemistry 2001, 40, 5399-5406.
9
J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005 865

A R T I C L E S
Zhu et al.
Figure 8. Comparison of the mechanisms for cytochrome P450, nitric oxide
synthase, and heme oxygenase.
Figure 9. Comparison of the reaction of NOS with L-arginine bound and
the proposed mechanism when L-NIO or 1400W is bound.
The active site region of all three isozymes of NOS around
the arginine-binding site is hydrophobic.28,29 The crystal structure
with L-NIO bound28 shows that the amidino methyl group faces
inward toward the heme peroxide, making the active site even
more hydrophobic. A water molecule that is observed between
the guanidino group of bound L-arginine and the proline residue
in the active site of eNOS is not present in the complex between
eNOS and L-NIO, presumably because of the increased hydro-
phobicity with the amidine bound.28 This hydrophobicity of the
active site coupled with the fact that the amidino methyl group
of the inactivators replaces the guanidino amino group of
L-arginine that provides the proton for activation of the heme
peroxide (or peroxy anion) rationalizes why the amidines convert
NOS catalysis to a heme oxygenase reaction (Figure 9).
Because there is no longer a proton source for conversion of
the heme peroxide (or peroxy anion) into the iron-oxo species
(the amino group of the guanidine is replaced by the methyl
group of the amidine) and the amidino methyl group is not
sufficiently activated for hydroxylation, the enzyme resorts to
the alternative heme oxygenase reaction. This leads to the
formation of R-meso-hydroxyheme, which then is converted to
verdoheme and biliverdin by further oxidation.
This proposed mechanism accounts for all of the experimental
observations except for the deuterium isotope effect on inactiva-
tion by the trideuteriomethylated amidines. If the mechanism
of inactivation of NOS by 1400W and L-NIO does not require
cleavage of the methyl C-H bond, then why is there a deuterium
isotope effect of about 1.3 on inactivation? Two effects may
be responsible. A CD3 group is slightly smaller than a CH3
group,33 and deuterium is less hydrophobic than hydrogen.34
Because both closeness of packing and hydrophobicity appear
to be important to the proposed inactivation mechanism, both
of these properties could easily be responsible for the small
kinetic isotope effect on inactivation.
The mechanism proposed for amidine inactivation of NOS
is quite novel. Irreversible inactivators typically cause inactiva-
tion by becoming covalently attached to the enzyme or, at least,
by undergoing a catalytic reaction to produce a molecule that
causes irreversible inactivation. We have not been able to
identify another example of a compound that causes irreversible
inactivation of a wild-type enzyme that does not itself become
modified, as is the case with the amidine inactivators of NOS.
In conclusion L-NIO, 1400W, and possibly all amidine-
containing inactivators of NOS appear to inactivate that enzyme
by preventing protonation of the heme peroxide intermediate,
thereby preventing the formation of the heme iron-oxo species
and normal substrate oxygenation. The alternative pathway, one
that is related to the reaction catalyzed by heme oxygenase,
results in self-mutilation and irreversible inactivation of NOS
by conversion of its heme cofactor into biliverdin.
Experimental Section
General Methods and Reagents. Optical spectra and enzyme assays
were recorded on a Perkin-Elmer Lambda 10 UV/vis spectrophotometer.
HPLC analyses were carried out on a Beckman 125 Solvent Module
delivery system equipped with a Beckman 166 detector. All NMR
spectra were recorded on a Mercury 400 spectrometer. Chemical shifts
(δ) are reported downfield from tetramethylsilane (Me4Si) in parts per
million (ppm). Mass spectra were recorded on G70-250SE (ESI),
Finnigan MAT900XL (EI), or Micromass Quattro II (APCI) spectrom-
eters.
Flash chromatography was performed with Merck silica gel (230-
400 mesh). TLC plates (silica gel 60-F254) were purchased from VWR
Scientific. L-Arginine, NADPH, HEPES, DTT, human ferrous hemo-
globin, OPA reagent, and myoglobin (from horse skeletal muscle) were
purchased from Sigma Chemical Co. Hemin and biliverdin IXR were
purchased from ICN Pharmaceuticals, Inc. Tetrahydrobiopterin (BH4)
was purchased from Alexis Biochemicals. 1-(N-Boc-aminomethyl)-3-
(aminomethyl)benzene was purchased from Aldrich Chemical Co. CD3-
CN was purchased from Acros Organic Co. L-NIO was purchased from
Research Biochemicals International. 1400W was either purchased from
(33) Kerstin, I.; Kohn, G.; Boeckmann, K.-U.; Kraft, R.; Holba-Schulz, P.; Ernst,
L. Org. Lett. 2000, 2, 4111-4113.
(34) El Tayara, N.; van de Waterbeemda, H.; Gryllakia, M.; Testa, B.; Trager,
W. F. Int. J. Pharm. 1984, 19, 271-281.
9
866 J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005

Inactivation of Inducible Nitric Oxide Synthase
A R T I C L E S
Alexis Biochemicals or synthesized. Acids, bases, and conventional
organic solvents were purchased from Fisher.
in ethyl acetate/hexane (2:1, v/v). After concentration by rotary
evaporation, the residues (5-20 mg) were dissolved in 10 mL of
MeOH, flushed with argon, and mixed with 5 mL of 1 M NaOH
containing 1 mM Na2EDTA. The mixture was kept at 37 °C for 1 h
for the R-, â-, and δ-isomers and overnight for the γ-isomer. After the
reaction was complete, 5 mL of glycine/HCl (pH 1.8) was added to
each, followed by 5 mL of 1 M HCl. The acidified mixture was then
extracted with an equal volume of CH2Cl2 and was dried over anhydrous
Expression and Purification of Recombinant iNOS. A recombi-
nant murine iNOS/calmodulin overexpression system in Escherichia
coli was obtained, and the enzyme was purified.35 The enzyme was
quantified according to the methods of Stuehr and Ikeda-Saito36 using
ꢀ397 ) 71 mM-1 cm-1 and a MW of 150 kDa.
Hemoglobin Assay for iNOS Activity. The production of nitric
oxide by iNOS was measured by the rapid oxidation of oxyHb to metHb
by nitric oxide.37 The assay mixture contained iNOS stock or aliquots
from incubation, L-arginine (0.1 mM), NADPH (0.1 mM), tetrahydro-
biopterin (10 µM), dithiothreitol (100 µM), and oxyhemoglobin (7.5
µM), diluted to a total volume of 600 µL with Hepes buffer (100 mM,
pH 7.4). The relative rate of nitric oxide synthesis was determined by
monitoring the NO-mediated conversion of oxyhemoglobin to meth-
emoglobin at 401 nm on a Perkin-Elmer Lambda 10 UV/vis spectro-
photometer. All assays were performed at 30 °C.
HPLC of Products formed from 1400W Inactivation of iNOS.
Chromatograms of iNOS inactivated by 1400W (107 µM) were obtained
on a C18 reversed-phase column (Vydac, 218TP54, 5 µm, 4.6 × 250
mm) at 401 nm with 60% H2O (0.1% TFA) and 40% CH3CN (0.1%
TFA). The flow rate was 1.0 mL/min. Control was made by either
omitting NADPH or 1400W. Standards were made with commercial
biliverdin IXR and hemin.
LC-Electrospray Mass Spectra of Inactivation Mixtures. iNOS
incubation with 1400W (107 µM) and other cofactors was kept at 37
°C for 2 h. A control was made by omitting the 1400W. A standard
sample was made with commercial biliverdin IXR and hemin. Aliquots
of 20 µL of these three samples were injected directly onto a YMC
ODS-AQ C18 column (2.0 × 250 mm). A Waters 2690 solvent delivery
system was used to elute the following gradient with a rate of 200
µL/min: 35-55% B over 10 min followed by 55-75% over 25 min
with solvent A consisting of water and solvent B of CH3CN. All
solvents contained 0.1% formic acid. After 8 min, the column outlet
was attached to a Q-TOF-2 quadrupole/time-of-flight mass spectrometer
without splitting for monitoring column output using positive mode
ionization. The scan range was 300-750, and the collision energy for
the MS/MS experiment was 35 eV.
1
Na2SO4 before rotary evaporation. In the H NMR spectra (CD3OD)
of the four biliverdin IX isomers, the peaks for the methyl ester
hydrogens (s, 3.66-3.72) disappeared after saponification. MS data
showed the M + H+ peak of all four hydrolyzed products to be 583.3-
583.5 (583.7 calculated).
MS/MS Analysis of the Four Biliverdin IX Isomers. Deprotonated
molecules of m/z 581 were detected for the biliverdin IX isomers using
negative ion mode mass spectrometry (Figure 5). Electrospray was used
for the analysis of the R-, â-, and δ-isomers on a VG70-250SE high-
resolution mass spectrometer and APCI for the γ-isomer on a
Micromass Quattro II triple quadrupole mass spectrometer. Commercial
biliverdin IXR was used to tune the instruments, and then MS-MS
spectra were obtained for all four isomers. The collision energy for
collision-induced dissociation during MS/MS was 8 kV for the VG70-
250SE instrument and 35 eV for the Quattro II.
Biliverdin Identification by HPLC Analysis. HPLC analysis was
performed using the same conditions for the analysis of the iNOS
inactivation mixture except for a different elution program: 100% A
for 2 min followed by 0-100% B over 10 min with solvent A consisting
of H2O (0.1% TFA) and solvent B of 25% KH2PO4 solution (20 mM,
pH 4.0) and 75% MeOH. The elution was monitored at 370 nm with
a flow rate of 0.5 mL/min.
Calibration Curves of Biliverdin and Heme and the Calculation
of the Ratio of Biliverdin/Heme Loss. Commercial biliverdin IXR
and hemin were dissolved in methanol to make a concentration of 1.0
mM, which was then diluted with H2O. A mixture of biliverdin and
heme in H2O (1.0, 2.0, 3.0, 4.0, 5.0 µM, each) was analyzed using the
same conditions as the HPLC analysis of the iNOS inactivation mixture.
The area of each peak was calculated with the Beckman Gold
Chromatography System (Version 1.6). Calibration curves of biliverdin
and heme were determined by plotting HPLC peak area versus
concentration (see Supporting Information). iNOS inactivation by L-NIO
was determined by replacing 1400W with L-NIO. Heme loss was
calculated by comparing the heme concentration in the control and
inactivation mixtures.
Synthesis of the Dimethyl Esters of Biliverdin IX Isomers. The
dimethyl esters of biliverdin IX isomers were obtained by oxidative
cleavage of hemin with O2-ascorbic acid followed by hydrolysis with
methanolic KOH and esterification with BF3-MeOH, as described by
Bonnett and McDonagh.24
Separation of the Dimethyl Esters of Biliverdin IX Isomers. The
biliverdin dimethyl ester isomers were isolated by preparative TLC
according to the method described by Heirwegh et al.38 with slight
modification. The mixture of the four esters was eluted five times with
chloroform/acetone (39:1, v/v) on a 20 × 20 cm preparative TLC plate,
and the four esters were distinctly separated into four bands with Rf
values of 0.62 (δ), 0.68 (γ), 0.72 (R), and 0.75 (â). The identity of the
compounds in the middle two bands was further confirmed by HPLC
(see Supporting Information): the retention time of the Rf 0.72 band is
the same as biliverdin IXR dimethyl ester, which was prepared from
commercial biliverdin IXR and BF3-MeOH;24 the Rf 0.68 band has the
longest retention time, which is consistent with all of the previous
reversed-phase HPLC analysis results.23,39 MS data showed M + H+
peaks of all four bands to be 611.5 (611.7 calculated).
N-(3-(Aminomethyl)benzyl)acetamidine (1400W, 4). Acetimidic
acid ethyl ester hydrochloride (247 mg, 2.00 mmol), prepared according
to the method of Liu et al.,40 was dissolved in 10 mL of ethanol, and
then 567 mg (2.40 mmol) of 1-(N-Boc-aminomethyl)-3-(aminomethyl)-
benzene was added all at once. The reaction mixture was stirred at 0
°C for 3 h and then partitioned between 20 mL of diethyl ether and 20
mL of H2O. The aqueous layer was washed with 10 mL of diethyl
ether twice before concentration by high-vacuum rotary evaporation
to a dry white solid of N-(3-(N-Boc-aminomethyl)benzyl)-acetamidine
hydrochloride (577 mg, 1.84 mmol, 92%): 1H NMR (D2O) d 1.41 (s,
9H), 2.28 (s, 3H), 4.24 (s, 2H), 4.48 (s, 2H), 7.18-7.42 (m, 4H).
The above compound (160 mg, 0.51 mmol) was stirred in a mixture
of 30 mL of 1,4-dioxane and 4 mL of 4 N HCl at ambient temperature
for 3 h. The reaction mixture was dried by high-vacuum rotary
evaporation to yield N-(3-(aminomethyl)benzyl)acetamidine dihydro-
chloride (125 mg, 0.50 mmol, 98%) which was further purified by three
consecutive recrystallizations with ethanol and diethyl ether to give
the product as a white granular crystalline solid: 1H NMR (D2O) d
2.27 (s, 3H), 4.18 (s, 2H), 4.52 (s, 2H), 7.30-7.52 (m, 4H); HRMS
Preparation of the Free Acids of Biliverdin IX. The four bands
from preparative TLC were scraped off, and the esters were dissolved
(35) (a) Hevel, J. M.; White, K. A.; Marletta, M. A. J. Biol. Chem., 1991, 222,
22789-22791. (b) Calaycay, J. R.; Kelly, T. M.; MacNaul, K. L.;
McCauley, E. D.; Qi, H.; Grant, S. K.; Griffin, P. R.; Klatt, T.; Raju, S.
M.; et al. J. Biol. Chem. 1996, 271, 28212-28219.
(EI): calcd for C10H15N3, M+
) 177.1260, M+ ) 177.1261.
(36) Stuehr, D. J.; Ikeda-Saito, M. J. Biol. Chem. 1992, 267, 20547-20550.
(37) Hevel, J. M.; Marletta, M. A. Methods Enzymol. 1994, 233, 250.
(38) Heirwegh, K. P. M.; Blanckaert, N.; Van Hesss, G. Anal. Biochem. 1991,
195, 273-278.
calcd
obs
(40) Liu, M. T. H.; Chishiti, N. H.; Burkholder, C. D.; Jones, W. E.; Wasson,
(39) Noguchi, M.; Yoshida, T.; Kikuchi, G. J. Biochem. 1982, 91, 1479-1483.
J. S. J. Org. Chem. 1980, 45, 4515-4519.
9
J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005 867

A R T I C L E S
Zhu et al.
N-(3-(Aminomethyl)benzyl)-2-[2H3]-acetamidine (d3-1400W, 11).
d3-1400W (126 mg, 0.50 mmol, 90% overall) was synthesized by the
same route used to make 1400W except that CD3CN was substituted
for CH3CN to make 2-[2H3]-acetimidic acid ethyl ester hydrochloride:
1H NMR (D2O) d 4.18 (s, 2H), 4.52 (s, 2H), 7.30-7.52 (m, 4H); HRMS
pipet) containing 55.0 µL of freshly prepared ferrous deoxymyoglobin
(2.0 µM), prepared by adding sodium dithionite (5 mg) to an 80.0 mL
solution of horse skeletal muscle myoglobin (2.7 mg, 16.0 nmol) in
100 mM Hepes buffer (pH 7.4), was carefully put into the 10 mL vial,
which was sealed with a rubber septum immediately. The 1400W
solution (67.2 µL, 500 µM) was injected into the enzyme solution via
a syringe, and the mixture was incubated at 37 °C with mild stirring
for 30 min. After removal of the septum, 50.0 µL of the ferrous
deoxymyoglobin solution was diluted to 500.0 µL with Hepes buffer
(100 mM, pH 7.4), and the spectrum was recorded against Hepes buffer.
The same experiment was repeated on iNOS inactivated with L-NIO
(7.5 mM). A control was made by replacing the 1400W or L-NIO with
Hepes buffer.
(EI): calcd for C10H122H3N3, M+
) 180.1448, M+ ) 180.1449.
calcd
obs
Irreversible Inhibition Kinetics. iNOS (0.1-0.3 µM) was incubated
at 37 °C in Hepes buffer (100 mM, pH 7.4), containing catalase (500-
1000 units), NADPH (9 mM), tetrahydrobiopterin (1 mM), dithiothreitol
(10 mM), glycerol (10%, v/v), and inactivator in a total volume of
70-280 µL. The reactions were initiated by the addition of iNOS, and
10 µL aliquots were removed to assay for enzyme activity at various
times. Controls were made by omitting the inactivators or NADPH.
LC-MS of OPA-DTT-derivatized d3-1400W after iNOS Inactiva-
tion. OPA-DTT reagent was prepared by mixing o-phthalaldehyde
(OPA, 1 mg/mL, “incomplete”) reagent with DL-dithiothreitol (DTT)
to a concentration of 100 mM. Inactivation mixtures were prepared as
described before in a volume of 210 µL with 5 µM or 10 µΜ d3-
1400W. Controls were prepared by omitting the enzyme. After the
enzyme fully lost its activity, the mixture (200 µL) was diluted with
Hepes buffer (100 mM, pH 7.4) to 1.0 mL, which was then ultrafiltered
with a stirring cell (NMWL: 10 000). The filtrate was collected, and
50 µL of the solution was derivatized with 10 µL of OPA-DTT reagent
for 5 min. HPLC analysis confirmed that d3-1400W was fully recovered
by comparing the adduct peak with that of the control. LC-MS (APCI
detector) studies were performed on the OPA-DTT-derivatized inactiva-
tion mixture and control in parallel.
Acknowledgment. The authors are grateful to the National
Institutes of Health for financial support of this work to R.B.S.
(GM49725). The electrospray mass spectrometer is supported
by an instrument grant (S10 RR10485) and core facility grant
(R24 CA83124) from the NIH to R.B.v.B. Y.Z. carried out all
of the experiments except for the LC-MS analysis with modified
heme. D.N. and R.B.v.B carried out the LC-MS analysis with
the modified heme. Y.Z. thanks Dr. Walter Fast (University of
Texas, Austin) and Michelle M. Spiering (University of
California, Berkeley) for providing help on NOS purification.
Supporting Information Available: Six figures of data and
a scheme. This material is available free of charge via the
Internet at http://pubs.acs.org.
Carbon Monoxide Detection. Into a 10-mL vial was added 82.8
µL of glycerol, 192.0 µL of NADPH (39.0 mM), 94.2 µL of
tetrahydrobiopterin (8.0 mM), 403.8 µL of iNOS (2.0 µM), and a stirring
bar. A small thin tube (made by sealing the tip of a piece of Pasteur
JA0445645
9
868 J. AM. CHEM. SOC. VOL. 127, NO. 3, 2005