Design and synthesis of C5 methylated L-arginine analogues as active site probes for nitric oxide synthase

Article abstract of DOI:10.1021/ja0746159

The role of nitric oxide (NO) as a biological signaling molecule is well established. NO is produced by the nitric oxide synthases (NOSs, EC 1.14.13.39), a class of heme proteins capable of converting L-arginine to NO and L-citrulline. Despite the large body of knowledge associated with the NOSs, mechanistic details relating to the unique oxidative chemistry performed by these enzymes remain to be fully elucidated. Furthermore, a number of disease states are associated with either the over- or underproduction of NO, making the NOS pathway an attractive target for the development of therapeutics. For these reasons, molecular tools capable of providing mechanistic insights into the production of NO and/or the inhibition of the NOSs remain of interest. We report here the stereospecific synthesis and testing of a number of new L-arginine analogues bearing a minimal substitution, methylation at position 5 of the amino acid side chain (such analogues have not been previously reported). The synthetic approach employed a modified photolysis procedure whereby irradiation of the appropriate diacylperoxide precursors at 254 nm gave access to the required unnatural amino acids in good yields. A heme domain construct of the inducible NOS isoform (iNOS_heme) was used to assess the binding of each compound to the enzyme active site. The compounds were also investigated as either inhibitors of, or alternate substrates for, the inducible NOS isoform. The results obtained provide new insight into the steric and stereochemical tolerance of the enzyme active site. These findings also further support the role of a conserved active site water molecule previously proposed to be necessary for NOS catalysis.

Full text of DOI:10.1021/ja0746159

Published on Web 09/25/2007
Design and Synthesis of C5 Methylated L-Arginine Analogues
as Active Site Probes for Nitric Oxide Synthase
Nathaniel I. Martin, Joshua J. Woodward, Michael B. Winter, William T. Beeson, and
Michael A. Marletta*
Contribution from the Departments of Chemistry, Molecular and Cellular Biology, and DiVision
of Physical Sciences, Lawrence Berkeley National Laboratory, UniVersity of California,
Berkeley, Berkeley, California 94720-1460
Received June 23, 2007; E-mail: marletta@berkeley.edu
Abstract: The role of nitric oxide (NO) as a biological signaling molecule is well established. NO is produced
by the nitric oxide synthases (NOSs, EC 1.14.13.39), a class of heme proteins capable of converting
L-arginine to NO and L-citrulline. Despite the large body of knowledge associated with the NOSs, mechanistic
details relating to the unique oxidative chemistry performed by these enzymes remain to be fully elucidated.
Furthermore, a number of disease states are associated with either the over- or underproduction of NO,
making the NOS pathway an attractive target for the development of therapeutics. For these reasons,
molecular tools capable of providing mechanistic insights into the production of NO and/or the inhibition of
the NOSs remain of interest. We report here the stereospecific synthesis and testing of a number of new
L-arginine analogues bearing a minimal substitution, methylation at position 5 of the amino acid side chain
(such analogues have not been previously reported). The synthetic approach employed a modified photolysis
procedure whereby irradiation of the appropriate diacylperoxide precursors at 254 nm gave access to the
required unnatural amino acids in good yields. A heme domain construct of the inducible NOS isoform
(iNOSheme) was used to assess the binding of each compound to the enzyme active site. The compounds
were also investigated as either inhibitors of, or alternate substrates for, the inducible NOS isoform. The
results obtained provide new insight into the steric and stereochemical tolerance of the enzyme active site.
These findings also further support the role of a conserved active site water molecule previously proposed
to be necessary for NOS catalysis.
Scheme 1. Reaction Catalyzed by NOS
The nitric oxide synthases (NOSs) are a family of enzymes
that catalyze the conversion of L-arginine to L-citrulline and
nitric oxide (NO) via the intermediate NG-hydroxy-L-arginine
(Scheme 1).1,2 Three distinct isoforms of NOS have been
identified, products of different genes, with different subcellular
localization, regulation, catalytic properties, and inhibitor sen-
sitivity. The human NOSs show 51-57% homology3 and
include two constitutive isoforms, neuronal nitric oxide synthase
(nNOS), which generates NO in the CNS4 and is involved in
NOS is active as a homodimer with each subunit containing
neurotransmission and long-term potentiation,5 and endothelial
a C-terminal reductase domain (with binding sites for NADPH,
nitric oxide synthase (eNOS), where the product NO is involved
FAD, and FMN) and a N-terminal oxygenase domain containing
in the regulation of smooth muscle relaxation and vascular
the heme prosthetic group. The substrate, L-arginine, and a redox
cofactor, (6R)-5,6,7,8 tetrahydro-L-biopterin (H4B), both bind
tone.6,7 A third, inducible isoform in macrophages (iNOS) is
important in immune system defense against pathogens and
near the heme center in the oxygenase domain.9 The overall
tumor cells.8
mechanism of the NOS reaction occurs via the two discreet
reactions shown in Scheme 1. In the first, L-arginine undergoes
two-electron oxidation (mediated by the redox active H4B
cofactor) with consumption of molecular oxygen to yield NG-
(1) Griffith, O. W.; Stuehr, D. J. Annu. ReV. Physiol. 1995, 57, 707-736.
(2) Kerwin, J. F., Jr.; Lancaster, J. R., Jr.; Feldman, P. L. J. Med. Chem. 1995,
38, 4343-4362.
(3) Alderton, W. K.; Cooper, C. E.; Knowles, R. G. Biochem. J. 2001, 357,
593-615.
(4) Schmidt, H. H.; Murad, F. Biochem. Biophys. Res. Commun. 1991, 181,
1372-1377.
(5) Schmidt, H. H.; Walter, U. Cell 1994, 78, 919-925.
(6) Forstermann, U.; Pollock, J. S.; Schmidt, H. H.; Heller, M.; Murad, F.
Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 1788-1792.
(7) Palmer, R. M.; Ferrige, A. G.; Moncada, S. Nature 1987, 327, 524-526.
(8) MacMicking, J.; Xie, Q. W.; Nathan, C. Annu. ReV. Immunol. 1997, 15,
323-350.
(9) Roman, L. J.; Martasek, P.; Masters, B. S. Chem. ReV. 2002, 102, 1179-
1190.
(10) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Chem. ReV. 1996,
96, 2841-2888.
(11) Shaik, S.; Kumar, D.; de Visser, S. P.; Altun, A.; Thiel, W. Chem. ReV.
2005, 105, 2279-2328.
(12) Denisov, I. G.; Makris, T. M.; Sligar, S. G.; Schlichting, I. Chem. ReV.
2005, 105, 2253-2277.
9
10.1021/ja0746159 CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 12563-12570
12563

A R T I C L E S
Martin et al.
Scheme 2. Oxygen Rebound Mechanism for the Hydroxylation of
L-Argininea
Figure 1. nNOSheme active site crystal structure (PDB ID 2GSK). A water
molecule is specifically oriented and within hydrogen-bonding distance of
the substrate, L-arginine, three active site residues (Val567 not shown for
clarity), and NO as the heme ligand (the three active site residues that
contribute to the orientation of the active site water molecule are completely
conserved for all three NOS isoforms).
a Indicated in bold are two mechanistically relevant protons which may
be provided via an active site water molecule.
Scheme 3. Proposed Mechanisms for the Oxidation of
NG-Hydroxy-L-arginine to L-Citrulline and NO (an Exogenous
Proton Source Is Not Required)
hydroxy-L-arginine. This is proposed to occur via an oxygen-
rebound mechanism analogous to that of the cytochrome P450-
type enzymes (Scheme 2).10-12 The second NOS reaction,
however, is unique in both its redox chemistry and substrate
specificity s there are no other enzymes known to perform one-
electron oxidations of NG-hydroxyguanidines to yield a urea and
NO as products. While the mechanistic details of the second
half-reactionhavebeenthesubjectofanumberofinvestigations,13-19
the precise mechanism of this process remains unclear. One
proposal for the second reaction involves a transient tetrahedral
intermediate formed by nucleophilic addition of the FeIII-peroxo
species (generated as in the first reaction) to NG-hydroxy-L-
arginine.20 The collapse of this addition complex is most often
proposed to proceed along one of two possible pathways
yielding L-citrulline and NO, with the concomitant one-electron
reduction of H4B, to return NOS to its resting state (Scheme 3).
While the two half-reactions by which NOS operates are
clearly distinct, there are common features. As seen in Schemes
2 and 3, both reactions involve the binding and reduction of
oxygen. While a proton source in the first NOS reaction step is
indicated, the need for externally supplied protons in the second
NOS reaction step is less clear. It has been recently speculated
that a specifically oriented water molecule in the NOS active
site may serve as such a proton source or act as a “shuttle” by
which protons from bulk solvent might be supplied in the first
NOS reaction.21,22 While a structure of the full-length protein
is not yet available, extensive crystallographic studies of the
NOSoxygenasedomainofeachisoformhavebeenreported.21,23-28
These structures consistently show a water molecule in a defined
location, held in place by a hydrogen-bonding network with
the substrate, active site residues, and the diatomic molecule
used as the heme ligand (Figure 1).
These considerations led us to speculate that modifications
to the substrate L-arginine at the δ-position (C5, adjacent to the
guanidine moiety) could provide useful molecular tools by which
to probe the NOS active site and catalytic mechanism. A
minimal modification was envisioned whereby either, or both,
of the enantiotopic protons at C5 are replaced with a methyl
(13) Crane, B. R.; Arvai, A. S.; Ghosh, S.; Getzoff, E. D.; Stuehr, D. J.; Tainer,
J. A. Biochemistry 2000, 39, 4608-4621.
(14) Stuehr, D. J.; Santolini, J.; Wang, Z. Q.; Wei, C. C.; Adak, S. J. Biol.
Chem. 2004, 279, 36167-36170.
(15) Huang, H.; Hah, J. M.; Silverman, R. B. J. Am. Chem. Soc. 2001, 123,
2674-2676.
(16) Tantillo, D. J.; Fukuto, J. M.; Hoffman, B. M.; Silverman, R. B.; Houk, K.
N. J. Am. Chem. Soc. 2000, 122, 536-537.
(17) Tierney, D. L.; Huang, H.; Martasek, P.; Masters, B. S.; Silverman, R. B.;
Hoffman, B. M. Biochemistry 1999, 38, 3704-3710.
(18) Wei, C. C.; Wang, Z. Q.; Hemann, C.; Hille, R.; Stuehr, D. J. J. Biol.
Chem. 2003, 278, 46668-46673.
(23) Pant, K.; Crane, B. R. Biochemistry 2006, 45, 2537-2544.
(24) Fedorov, R.; Vasan, R.; Ghosh, D. K.; Schlichting, I. Proc. Natl. Acad.
Sci. U.S.A. 2004, 101, 5892-5897.
(25) Crane, B. R.; Arvai, A. S.; Ghosh, D. K.; Wu, C.; Getzoff, E. D.; Stuehr,
D. J.; Tainer, J. A. Science 1998, 279, 2121-2126.
(19) Rosen, G. M.; Tsai, P.; Pou, S. Chem. ReV. 2002, 102, 1191-1200.
(20) Hurshman, A. R.; Marletta, M. A. Biochemistry 2002, 41, 3439-3456.
(21) Li, H.; Igarashi, J.; Jamal, J.; Yang, W.; Poulos, T. L. J. Biol. Inorg. Chem.
2006, 11, 753-768.
(26) Raman, C. S.; Li, H.; Martasek, P.; Kral, V.; Masters, B. S.; Poulos, T. L.
Cell 1998, 95, 939-950.
(27) Fedorov, R.; Hartmann, E.; Ghosh, D. K.; Schlichting, I. J. Biol. Chem.
2003, 278, 45818-45825.
(22) Cho, K. B.; Derat, E.; Shaik, S. J. Am. Chem. Soc. 2007, 129, 3182-
3188.
(28) Li, H.; Poulos, T. L. J. Inorg. Biochem. 2005, 99, 293-305.
9
12564 J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007

C5 Methylated
L
-Arg Analogues as Active Site Probes
A R T I C L E S
Table 1. 5-Substituted L-Arginines and NG-Hydroxy-L-arginines
Scheme 4. Synthesis of Orthogonally Protected C5-Methylated
L-Ornithine
Prepared
Scheme 5. Conversion of Orthogonally Protected C5-Methylated
L-Ornithine to (5R/S)-Methyl-L-arginine and NG-Hydroxy-(5R/S)-
methyl-L-arginine
group. Analogues wherein the pro-S hydrogen is replaced with
a methyl group may experience steric crowding with the active
site water molecule. Alternatively, compounds with substitution
of the pro-R hydrogen may interfere with or prevent binding
of heme ligands.
While no syntheses of C5 substituted L-arginine analogues
have been described, the preparation of analogues with methyl
substitution at positions C2(R), C3(â), and C4(γ) have been
reported, albeit as diastereomeric mixtures.29 The effects of these
compounds on purified NOS were not thoroughly investigated,
and the synthetic approach utilized is not applicable to the
preparation of C5 substituted L-arginine analogues. We here
report the stereospecific synthesis of all possible C5 methylated
L-arginine analogues and the effects observed when testing these
compounds with iNOS.
Results
The synthesis of C5 substituted L-arginines requires access
to the appropriately substituted and orthogonally protected
L-ornithine precursors. From here, existing methods can be used
to install either the guanidine or NG-hydroxy guanidine moi-
eties.30 The required orthogonally protected L-ornithines were
envisioned to be accessible via the diacylperoxide photolysis
method recently described by Vederas and co-workers.31 Initial
Figure 2. Spectral Kd determination for compound 6d. (A) Low- to
high-spin state transition effected by addition of 6d to iNOSheme
.
(B) Titration data and fit used in determination of the spectral Kd
for 6d.
investigation of this synthetic route made use of racemic
3-aminobutyric acid as starting material (Scheme 4). Carbamate
protection of the amine gave compound 1a which was coupled
with 2-methoxyprop-2-yl hydroperoxide32 to provide the perester
2a. Upon treatment with TFA the perester was converted to
(29) The Biology of Nitric Oxide (4) Enzymology, Biochemistry and Immunology;
Narayanan, K., Frey, C., Griffith, O. W., Eds.; Portland Press: London,
1993.
(30) Martin, N. I.; Woodward, J. J.; Marletta, M. A. Org. Lett. 2006, 8, 4035-
4038.
(31) Spantulescu, M. D.; Jain, R. P.; Derksen, D. J.; Vederas, J. C. Org. Lett.
2003, 5, 2963-2965.
(32) Dussault, P.; Sahli, A. J. Org. Chem. 1992, 57, 1009-1012.
9
J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007 12565

A R T I C L E S
Martin et al.
Table 2. Spectral Kd Values for 6b-d and 7b-d
a
compound
spectral Kd (µM)
L-arginineb
7.0 ( 0.7 (5.8 ( 1.6)
5.1 ( 0.4 (1.5 ( 0.3)
1.6 ( 0.3
NG-hydroxy-L-arginineb
(5R)-methyl-L-arginine (6b)
(5S)-methyl-L-arginine (6c)
5,5-dimethyl-L-arginine (6d)
NG-hydroxy-(5R)-methyl-L-arginine (7b)
NG-hydroxy-(5S)-methyl-L-arginine (7c)
NG-hydroxy-5,5-dimethyl-L-arginine (7d)
165 ( 35
1.2 ( 0.1
1.8 ( 0.5
270 ( 57
2.1 ( 0.5
a Spectral Kd values reported for 6b-d and 7b-d are averages obtained
from triplicate analysis. bPreviously reported values given in parentheses
(ref 20).
Figure 5. NOSheme active site crystal structure with 5R-methyl-L-arginine
modeled in place of L-arginine and oxygen in place of NO. The distance
between the carbon atom of the methyl substituent in the substrate analogue
and the oxygen molecule modeled at the heme center is calculated to be
2.23 Å (within van der Waals contact distance).
Table 3. Ki Values Determined for 6b-d, 7b, 7d and Km Value
Determined for 7c
compound
Kia
(µM)
NG-amino-L-arginineb
(5R)-methyl-L-arginine (6b)
(5S)-methyl-L-arginine (6c)
2.2 ( 0.5 (1.7 ( 0.6)
6.3 ( 0.1
570 ( 130
5,5-dimethyl-L-arginine (6d)
1.2 ( 0.4
NG-hydroxy-(5R)-methyl-L-arginine (7b)
NG-hydroxy-(5S)-methyl-L-arginine (7c)
NG-hydroxy-5,5-dimethyl-L-arginine (7d)
3.4 ( 0.6
327 ( 35c
12.0 ( 1.5
a Values reported are averages obtained from triplicate analysis. b Ki value
determined for NG-amino-L-arginine as a control; previously reported value
given in parentheses.33 c Km reported.
Figure 3. Km determination for compound 7c.
ornithine 4a was critical to the construction of the amino acid
backbone. The reaction conditions described in the literature31
for this type of conversion, however, provided mixed results
(low yields of 4a requiring extended reaction times of up to 5
days). Optimization of the process led us to a modified protocol
whereby the desired photolysis products were reliably obtained
in yields of 73-83% and in a much shorter period of time. In
the modified protocol, dissolution/evaporation cycles are per-
formed every 6 h allowing for unreacted diacylperoxide to be
evenly redistributed. Also of note is the inclusion of a short,
solution-phase photolysis step to end the reaction. We found
that this did not complicate formation of the desired products
(via the presumed radical cage-recombination process) as very
little of the undesired homocoupling products were detected.
Furthermore, as described in detail in the Experimental Section,
the equipment and setup used in this modified photolysis
procedure requires no custom glassware or instrumentation; all
necessary components are inexpensive or common in the
laboratory.
With access to orthogonally protected L-ornithine compounds
like 4a, their conversion to the desired guanidine or NG-
hydroxyguanidines was then performed following literature
methods (Scheme 5).30
Figure 4. Ki determination for compound 6d. Apparent Km for L-arginine
plotted as a function of inhibitor concentration.
the peracid followed by coupling with the appropriately
protected L-aspartic acid to give diacylperoxide 3a in good yield.
Photolysis of 3a provided access to the required L-ornithine
analogue 4a with methyl substitution at position C5.
As shown in Scheme 4, the photolytic conversion of dia-
cylperoxide 3a to the orthogonally protected (5R/S)-methyl-L-
The OTHP protected NG-hydroxyguanidine 5a was prepared
in a one-pot process by converting the free amine (obtained
after hydrogenation of 4a) to the intermediate thiourea by
treatment with CbzNCS. The thiourea was then directly
converted to 5a by treatment with EDCI and H2N-OTHP.
Owing to the reductive instability of OTHP protected NG-
9
12566 J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007

C5 Methylated
L-Arg Analogues as Active Site Probes
A R T I C L E S
Figure 6. Proposed interactions of (A) 6b, 7b and (B) 6d, 7d with the iNOS active site. The presence of a methyl substituent in place of the pro-R hydrogen
may interfere with oxygen binding and prevent catalysis.
aqua f FeIII-unligated) upon binding. This transition involves
a concomitant decrease in absorbance at ∼420 nm and an
increase at ∼396 nm in the UV-visible spectrum and is depicted
for compound 6d in Figure 2A. The magnitude of the spectral
change, ∆∆Abs (where ∆∆Abs ) ∆Abs396 - ∆Abs420), is
dependent on the concentration of the added analogue. Spectral
Kd values for each compound were then determined by fitting
the titration data to the saturation binding equation: ∆∆Abs )
(∆∆Absmax × [compound])/(Kd + [compound]) (Figure 2B
depicts this analysis for compound 6d). Table 2 summarizes
the results of the spectral binding assay for each of the
5-methylated and 5,5-dimethylated L-arginine and NG-hydroxy-
L-arginine analogues.
Working from the spectral Kd values determined for 6b-d
Figure 7. NOSheme active site crystal structure with 5S-methyl-L-arginine
and 7b-d, the ability of each compound to behave as either an
modeled in place of L-arginine and oxygen in place of NO. The distance
inhibitor or substrate for full-length iNOS was next investigated.
between the carbon atom of the methyl substituent in the substrate analogue
NOS activity was measured using a spectral assay making use
of the NO-induced oxidation of oxymyoglobin to metmyoglobin
observable at 405 nm. Initial investigation revealed that one of
the six compounds tested, 7c, led to detectable NO production
when added at saturating concentrations (>10× the spectral Kd).
The kinetics of NO production with 7c was further investigated
and showed the compound to be a substrate for iNOS. The Km
value for 7c was determined to be 327 ( 35 µM with a kcat
value approximately 35% of that observed for L-arginine (Fig-
ure 3).
and the active site water molecule is calculated to be 2.88 Å (within van
der Waals contact distance) and may lead to its extrusion.
hydroxyguanidines,30 compound 5a could be used as a precursor
to both the guanidine 6a and NG-hydroxyguanidine 7a final
products.
After establishing the validity of the approach, each diaste-
reomer of 5-methyl-L-arginine and NG-hydroxy-5-methyl-L-
arginine was prepared in pure form from the appropriate
enantiomer of 3-aminobutryric acid. The 5,5-dimethyl substi-
tuted compounds were prepared using 3-amino-3-methylbutyric
acid as starting material (Table 1).
Binding affinities for the active site of iNOS were determined
using a previously described spectral binding assay.20 For this
analysis, a truncated construct of iNOS (iNOSheme, comprised
of amino acids 1-490) was used. This construct has been
previously shown to bind heme, H4B, and substrate(s) in
identical fashion to that for the full-length enzyme and is
catalytically competent when provided with an exogenous source
of electrons such as sodium dithionite.20 Addition of 6b-d and
7b-d to H4B-bound iNOSheme induces a low- to high-spin state
conversion at the heme center resulting from loss of water (FeIII-
The five compounds that did not lead to NO production were
investigated as competitive inhibitors of iNOS. Figure 4
illustrates the Ki determination as performed for compound 6d,
and Table 3 summarizes the NOS inhibition/NO production
assay results for all of compounds 6b-d and 7b-d.
The five compounds displaying NOS inhibition (6b-c, 7b,
7d) were also investigated as time-dependent inactivators of the
enzyme. In all cases NOS activity was fully restored upon
addition of L-arginine to mixtures of each compound preincu-
bated with iNOS under turnover conditions, indicating no time
dependent or irreversible inactivation (data not shown).
9
J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007 12567

A R T I C L E S
Martin et al.
Figure 8. Proposed interactions of compounds 6c and 7c. (A) The methyl group in place of the pro-S hydrogen leads to extrusion of an active site water
molecule required for the first step of the NOS reaction. (B) While bearing a methyl group in place of the pro-S hydrogen, the NG-hydroxylated compound
7c is converted to NO, suggesting that an active site water molecule is not required for the second step of the NOS reaction.
Discussion
weak inhibitor, 7c is a substrate for iNOS. As compounds 6c
and 7c do not possess a methyl substituent in the pro-R position,
oxygen binding might not be affected in the manner proposed
for the other four analogues (Figure 6). While the affinity of 6c
and 7c for iNOSheme is relatively weak, we speculate that, upon
binding, both compounds lead to extrusion of the hydrogen-
bonded water molecule from the active site (Figure 7).21,23,28
Furthermore, the observation that the NG-hydroxylated com-
pound 7c leads to NO formation while the non-hydroxylated
compound 6c does not suggests that the proposed hydrogen-
bonded active site water molecule is required only for the first
step of the NOS reaction (conversion of L-arginine to NG-
hydroxy-L-arginine) (Figure 8). As described earlier, an active
site water molecule is proposed to be involved in the generation
of the high valent FeIV-oxo species required for the hydroxy-
lation of L-arginine to NG-hydroxy-L-arginine (Scheme 2). The
same process need not be invoked in a mechanistic explanation
for the conversion of NG-hydroxy-L-arginine to NO and L-
citrulline (Scheme 3).
The results obtained with the various C5-methylated ana-
logues of L-arginine and NG-hydroxy-L-arginine point to a
stereochemical preference in the ability of these compounds to
bind to the NOS active site with high affinity. In addition, as
discussed below, these results also provide further support for
the recently proposed role of an active site water molecule in
NOS catalysis.21,22
The Ki and Km values obtained for the six analogues tested
coincide very well with the spectral Kd values obtained.
Compounds 6b, 6d, 7b, and 7d all bear methyl substitution in
the pro-R position of C5 in the side chain of L-arginine and
NG-hydroxy-L-arginine and bind with high affinity. Despite the
tight binding of 6b, 6d, 7b, and 7d for iNOS, enzymatic
formation of NO was not detected. Substitution of the pro-R
hydrogen (as in 6b, 6d, 7b, and 7d) introduces a methyl group
that may be positioned above the heme upon binding of the
compound to NOS (Figure 5). In order for normal catalysis to
proceed in both steps of the NOS reaction (Schemes 2 and 3),
oxygen must coordinate to Fe in the heme. The presence of a
methyl substituent in this space may block or prevent oxygen
from binding in its required orientation and inhibit catalysis
(Figure 6).
These results further support the idea that an active site water
molecule is necessary for only the first reaction in the NOS
catalytic cycle. This interpretation is in agreement with previ-
ously suggested mechanisms involving the role of water in NOS
catalysis.21,23,28
Methyl substitution in the pro-R position (as for 6b, 6d, 7b,
and 7d) appears to have a stabilizing effect on the enzyme-
substrate analogue complex as reflected in the low Kd and Ki
values. The nature of this stabilization is not clear at this time.
We speculate that methyl substitution of the pro-R proton
facilitates the displacement of the heme iron coordinated water
molecule present in the resting state of the enzyme. The overall
effect of this on the detailed interactions at the active site awaits
structural studies that are underway. However, further support
for the stabilizing role of a methyl substituent at the pro-R
position is suggested by the observation that 6c and 7c, each
bearing methyl substitution solely in the pro-S position of C5,
bind much less tightly than the other four analogues tested. Of
special interest, however, is the observation that while 6c is a
Conclusion
In summary, analogues of L-arginine and NG-hydroxy-L-
arginine, bearing methyl substitution at the C5 position, have
shed new light on the steric and stereochemical tolerances of
the NOS active site. The construction of these analogues
required the use of an optimized diacylperoxide photolysis
procedure that should be of general synthetic use and applicabil-
ity. Methyl substitutions adjacent to the guanidinium group in
both L-arginine and NG-hydroxy-L-arginine have here been
shown to elicit varying effects on the enzyme. Of note is the
implication that disruption of an organized water molecule in
the NOS active site interferes with only the first reaction in the
9
12568 J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007

C5 Methylated
L-Arg Analogues as Active Site Probes
A R T I C L E S
room temperature and stirred for 3 h at which time TLC indicated
complete consumption of the peracid (Rf ) 0.13, 7:3 hexane/ethyl
acetate). The precipitated urea was removed by filtration through Celite
and washed with CH2Cl2, and the combined filtrates were evaporated
under vacuum. The product (Rf ) 0.50) was purified as a clear oil
(1.39 g, 88% over two steps) by flash chromatography (7:3 hexane/
ethyl acetate). Diacylperoxide 3a was found to be stable when stored
two-step NOS catalytic cycle. These findings may also serve
as a blueprint toward the design of new compounds capable of
acting as inhibitors or alternate substrates for NOS. Continuing
work is aimed at further characterizing the oxygen binding
capabilities of the enzyme when bound to the C5 methylated
analogues. Also, crystallographic studies with these methylated
analogues bound to the NOS active site are under way to help
establish the proposed modes of action for these compounds.
1
over a desiccant at -80 °C in small (0.5 mmol) aliquots. H NMR
(CDCl3, 400 MHz) δ 7.38-7.29 (m, 5H), 5.47 (d, J ) 7.6 Hz, 1H),
5.19 (m, 1H), 5.09 (m, 2H), 4.52 (m, 1H), 4.19 (m, 1H), 3.02 (m, 2H),
2.68 (d, J ) 4.9 Hz, 2H), 1.45 (m, 18H), 1.28 (m, 3H); 13C NMR
(CDCl3, 100 MHz) δ 168.8, 166.9, 166.5, 155.4, 155.2, 136.3, 128.5,
128.1, 128.0, 83.1, 80.2, 66.7, 50.3, 43.9, 36.4, 33.0, 28.2, 27.7, 19.7.
HRMS (ESI, M + Na) Calcd for C25H36N2O10Na547.2268, found
547.2266.
Experimental Section
All reagents employed were of American Chemical Society (ACS)
grade or finer and were used without further purification unless
otherwise stated. 2-Methoxyprop-2-yl hydroperoxide was prepared as
described in the literature.32 1H NMR and 13C NMR spectra were
obtained on a Bruker AVQ-400 spectrometer. High-resolution FAB
and ESI mass spectra were obtained using VG ZAB2-EQ and Q-Tof
Premier (Waters) instruments, respectively. Detailed here is the
representative synthesis of compounds 1a-7a. Full experimental details
and data for all other compounds are given in the Supporting
Information.
(3R/S)-Benzyloxycarbonylamino-butyric Acid (1a). (3R/S)-Ami-
nobutyric acid (2.0 g, 19.4 mmol) was dissolved in 20 mL of 2 M
NaOH and cooled on ice. Benzyl chloroformate (1.2 equiv, 23.3 mmol,
3.4 mL) was slowly added, and 30 min after complete addition the
mixture was warmed to room temperature and stirred for an additional
2 h. The pH of the mixture was adjusted to 2 with concentrated HCl
and extracted with ethyl acetate (3 × 40 mL). The organic layer was
dried over Na2SO4 and concentrated under vacuum to a volume of 30
mL. Hexanes where then added and the product precipitated. Compound
1a was isolated as a white solid with spectral data identical to those
reported previously for the same compound.34
5(R/S)-Benzyloxycarbonylamino-(2S)-tert-butoxycarbonylamino-
hexanoic Acid tert-Butyl Ester (4a). Following a modified literature
protocol,31 diacylperoxide 3a (225 mg, 0.43 mmol) was dissolved in
CH2Cl2 (2 mL) and transferred to the photolysis vessel (see Supporting
Information for detailed description of photolysis vessel and procedure).
Solvent was removed by flowing dry N2 through the vessel until a thin
film of neat starting material remained. The vessel was placed in an
acetone/dry ice bath and continuously purged with dry N2 gas
throughout the course of the reaction. The neat diacylperoxide was
irradiated at 254 nm for 6 h after which the residue was redissolved,
re-evaporated, and reirradiated for an additional 6 h. This process was
repeated once more after which the mixture was dissolved in a minimum
volume of CH2Cl2 (1 mL) and the solution irradiated for 3 h in order
to ensure full consumption of starting material. Compound 4a (Rf )
0.55) was purified as a colorless oil (162 mg, 83%) by flash
chromatography on silica gel (3:1 hexane/ethyl acetate). 1H NMR
(CDCl3, 400 MHz) δ 7.35-7.28 (m, 5H), 5.15-5.04 (m, 3H), 4.74-
4.63 (m, 1H), 4.17 (m, 1H), 3.71 (m, 1H), 1.88-1.73 (m, 1H), 1.68-
1.55 (m, 1H), 1.55-1.34 (m, 20H), 1.14 (m, 3H). 13C NMR (CDCl3,
100 MHz) δ 171.7, 155.7, 155.4, 136.6, 128.5, 128.1, 81.9, 79.7, 66.5,
53.8, 47.1, 46.9, 32.7, 32.5, 29.7, 29.5, 28.3, 28.0, 21.2; HRMS (FAB,
M + H) Calcd for C23H37N2O6 437.2652, found 437.2657.
3(R/S)-Benzyloxycarbonylamino-butaneperoxoic Acid 1-Meth-
oxy-1-methyl-ethyl Ester (2a). Compound 2a was prepared by
coupling compound 1a to 2-methoxyprop-2-yl hydroperoxide. Com-
pound 1a (5.0 mmol, 1.2 g), DCC (6.0 mmol, 1.3 g), and DMAP (0.1
eq, 0.50 mmol, 61 mg) were added sequentially to a solution of
2-methoxyprop-2-yl hydroperoxide (1.5 equiv, 7.5 mmol, 0.80 g) in
CH2Cl2 (30 mL) at -20 °C. The reaction was slowly warmed to room
temperature and stirred for an additional 2 h. The precipitated urea
was removed by filtering through Celite and washed with several
volumes of CH2Cl2, and the filtrate was concentrated under vacuum.
The desired product (Rf ) 0.25) was purified as a colorless oil (1.14 g,
70%) by flash chromatography on silica gel (80:20 hexane/ethyl
NG-OTHP-NG-benzyloxycarbonyl-(5R/S)-methyl-N-tert-butoxy-
carbonyl-L-arginine-tert-butyl Ester (5a). Compound 5a was produced
using a one-pot procedure previously described in the literature.30
Compound 4a (150 mg, 0.34 mmol) was dissolved in ethyl acetate (6
mL) and treated with 10% Pd/C (100 mg). Hydrogen gas was
administered throughout the reaction using a balloon until all the starting
material had reacted (9 h). The deprotected amine was visualized with
TLC (3:1 hexane/ethyl acetate), appearing as a red spot along the
baseline using ninhydrin stain. Following concentration under vacuum,
the crude amine was dissolved in 6 mL of CH2Cl2. CbzNCS4 was added
dropwise (as a 0.5 M solution in CH2Cl2) until TLC indicated
consumption of the amine with concomitant production of the thiourea
(rapid conversion). The mixture was next directly treated with triethy-
lamine (1.1 equiv, 0.37 mmol, 53 µL), EDCI (1.1 equiv, 0.37 mmol,
71 mg), and OTHP-protected hydroxylamine (1.2 equiv, 0.41 mmol,
48 mg). The reaction mixture was stirred at room temperature, and
additional half equivalents of the reagents were added sequentially every
30-45 min until complete consumption of the starting material had
occurred, as indicated by TLC. The product, a colorless oil, was isolated
(156 mg, 80%) as the expected mixture of diastereomers (Rf ) 0.44)
1
acetate). H NMR (CDCl3, 400 MHz) δ 7.54-7.30 (m, 5H), 5.20 (br
m, 1H), 5.10 (s, 2H), 4.15 (m, 1H), 3.33 (s, 3H), 2.58 (m, 2H), 1.46 (s,
6H), 1.28 (d, J ) 6.8 Hz, 3H); 13C NMR (CDCl3, 100 MHz) δ 168.4,
155.4, 136.4, 128.5, 128.1, 128.0, 107.1, 66.7, 49.9, 44.0, 37.3 22.6,
22.4, 20.1; HRMS (ESI, M + Na) Calcd for C16H23NO6Na 348.1423,
found 348.1430.
3-((3R/S)-Benzyloxycarbonylamino-butyrylperoxycarbonyl)-(2S)-
tert-butoxycarbonylamino-propionic Acid tert-Butyl Ester (3a).
Compound 2a (3.0 mmol, 0.98 g) was dissolved in CHCl3 (40 mL)
and treated with 50% aqueous TFA (14 mL) at 0 °C. The reaction was
slowly warmed to room temperature and stirred for 30 min. The reaction
was quenched through the addition of saturated aqueous NaHCO3
(important to verify complete neutralization by checking pH). The
peracid was then extracted into ether (3 × 40 mL), dried with Na2SO4,
concentrated under reduced pressure, and used without further purifica-
tion. The crude peracid was dissolved in CH2Cl2 (50 mL) and cooled
on ice. Boc-Asp-OtBu (1 equiv, 3.0 mmol, 0.87 g) was added as a
solution in CH2Cl2 (10 mL) followed by DCC (1.1 equiv, 3.3 mmol,
0.68 g) as a solution in CH2Cl2 (10 mL). The reaction was warmed to
1
by flash chromatography on silica gel (3:1 hexane/ethyl acetate). H
NMR (CDCl3, 400 MHz) δ 7.89 (d, J ) 4.0 Hz, 1H), 7.42-7.35 (m,
5H), 6.19 (br q, J ) 6.0 Hz, 1H), 5.40-5.04 (m, 3H), 4.94 (m, 1H),
4.13 (m, 1H), 3.90 (m, 1H), 3.63 (m, 2H), 1.91-1.72 (m, 4H), 1.65-
1.50 (m, 6H), 1.49-1.41 (m, 18H), 1.15 (m, 3H); 13C NMR (CDCl3,
100 MHz) δ 171.84, 171.76, 155.5, 155.4, 152.9, 148.4, 148.3, 148.2,
135.1, 128.7, 128.4, 128.1, 101.3, 101.2, 81.7, 81.6, 81.5, 79.5, 79.4,
67.6, 64.0, 63.9, 54.2, 54.1, 53.8, 46.0, 36.6, 32.3, 32.3, 32.0, 31.9,
(33) Komori, Y.; Wallace, G. C.; Fukuto, J. M. Arch. Biochem. Biophys. 1994,
315, 213-218.
(34) Sutherland, A.; Willis, C. L. J. Org. Chem. 1998, 63, 7764-7769.
9
J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007 12569

A R T I C L E S
Martin et al.
29.7, 29.4, 29.14, 29.07, 28.3, 28.0, 25.1, 20.8, 20.6, 20.5; HRMS (FAB,
M + H) Calcd for C29H47N4O8 579.3394, found 579.3404.
the Kd was determined by fitting the data by a nonlinear least-squares
fit to the saturation binding equation: ∆∆Abs ) (∆∆Absmax
×
(5R/S)-Methyl-L-arginine Di-trifluoroacetate (6a). Compound 5a
(140 mg, 0.24 mmol) was dissolved in 5 mL of MeOH/H2O (1% acetic
acid), and the system was purged with N2(g) prior to addition of 10%
Pd/C (100 mg). Hydrogen gas was delivered with a balloon to the
stirring reaction mixture throughout the course of the reaction (3 h,
room temperature). Following reaction completion, the catalyst was
removed by filtering through Celite and washed with several volumes
of MeOH/H2O (1% acetic acid). The filtrate was concentrated under
vacuum followed by treatment with 10 mL of 25% thioanisole in TFA
for 5 h at room temperature. The mixture was then concentrated under
vacuum, and the residue was partitioned between ether (15 mL) and
water (15 mL) to remove nonpolar impurities. The ether layer was
further extracted with 5 mL of H2O, and the aqueous layers were
combined and concentrated under vacuum (final volume 5.0 mL). The
amino acid product was purified using a Varian C18 megabond elution
column (60 mL volume containing 10 g C18 silica). Prior to applying
the crude product, the column was pre-equilibrated with methanol (50
mL) followed by 0.1% TFA in H2O (50 mL). After loading the column,
the product was eluted using 0.1% TFA in H2O and collected in 10
mL fractions. Fractions 2-5 (positive ninhydrin stain) were pooled,
concentrated, and lyophilized to give the product as a clear, glassy foam
[compound])/(Kd + [compound]) (KaleidaGraph, Abelbeck Software).
All analyses were performed in triplicate, and the average value was
reported.
NOS Inhibition/Activity Assays. A modified, high-throughput
oxymyoglobin assay35 using a 96-well plate reader was utilized to assess
the NO production or inhibition exhibited by compounds 6b-d and
7b-7d with full-length iNOS. Described here is the Ki determination
for 6d and the Km determination for 7c. Ki Determination for 6d:
Assay reactions were prepared containing L-arginine (0, 10, 20, 50,
100, 200, 400, and 1000 µM), 6d (2.8, 5.0, 10, and 15 µM), 12 µM
H4B, 5.0 µM DTT, 8.5 µM oxymyoglobin, and 15 nM iNOS in 100
mM Hepes (pH 7.4). The NOS reaction was initiated by addition of
NADPH (final concentration 100 µM) to each well using a multichannel
pipet. NO production was monitored by measuring the increase in
absorbance at 405 nm corresponding to the NO-mediated conversion
of oxymyoglobin to metmyoglobin. The data derived from the
Michaelis-Menten plots of iNOS inhibition by 6d were used to
generate an apparent Km versus6d replot from which the Ki for 6d was
determined to be 1.3 ( 0.4 µM. Each analysis was performed in
triplicate, and the average was reported. Km Determination for 7c:
Assay reactions were prepared containing 7c (350, 700, 1050, 1400,
2100, 2800, and 5600 µM), 12 µM H4B, 5 µM DTT, 8.5 µM
oxymyoglobin, and 15 nM iNOS in 100 mM Hepes (pH 7.4). The NOS
reaction was initiated by addition of NADPH (final concentration 100
µM) to each well using a multichannel pipet and NO production
monitored as described above. The data were fit to a standard
Michaelis-Menten plot (KaleidaGraph, Abelbeck Software) from which
the Km for 7c was determined to be 327 ( 35 µM. Each analysis was
performed in triplicate, and the average was reported.
1
in quantitative yield. H NMR (D2O, 400 MHz) δ 4.00 (app dt, J1 )
1.6 Hz, J2 ) 6.0 Hz, 1H), 3.51 (app sext, J ) 6.4 Hz, 1H), 2.03-1.81
(m, 2H), 1.73-1.46 (m, 2H) 1.13 (app dd, J1 ) 0.8 Hz, J2 ) 6.4 Hz,
3H); 13C NMR (D2O, 100 MHz) δ 171.7, 156.1, 52.50, 53.46, 47.4,
30.9, 30.8, 26.1, 26.0, 19.4, 19.3; HRMS (FAB, M + H) Calcd for
C7H17N4O2 189.1352, found 189.1357.
NG-Hydroxy-(5R/S)-methyl-L-arginine Di-trifluoroacetate (7a).
Compound 5a (150 mg, 0.25 mmol) was dissolved in a 10 mL solution
of TFA/25% thioanisole, and the mixture was stirred at room temper-
ature under N2 for 6 h during which time the solution’s color progressed
from bright green to dark red. The product was isolated as described
Acknowledgment. This work was supported by the Aldo
DeBenedictis Fund, the Natural Sciences and Engineering
Council of Canada, and the Alberta Heritage Foundation for
Medical Research (postdoctoral fellowships to N.I.M.). Nathaniel
Fernhoff is gratefully acknowledged for providing full-length
iNOS.
1
above for compound 6a and obtained in quantitative yield. H NMR
(D2O, 400 MHz) δ 4.00 (app dt, J1 ) 1.6 Hz, J2 ) 6.2 Hz, 1H), 3.51
(app sext, J ) 6.4 Hz, 1H), 2.03-1.81 (m, 2H), 1.76-1.49 (m, 2H),
1.13 (app d, J ) 6.0 Hz, 3H); 13C NMR (D2O, 100 MHz) δ 171.7,
158.1, 52.51, 52.48, 47.6, 30.8, 30.7, 26.2, 26.1, 19.4, 19.3; HRMS
(FAB, M + H) Calcd for C7H17N4O3 205.1301, found 205.1298.
Spectral Kd Titrations. iNOSheme was expressed and purified as
previously described.20 In a typical assay H4B-bound iNOSheme (500
µL of 0.5-1 µM) was titrated with the compound of choice working
at concentrations appropriate to detect the spectral transition from 420
nm (unbound FeIII-aqua) to 396 nm (saturated). Absorbance spectra
were corrected for the dilution due to the volume of titrant added. As
Supporting Information Available: Experimental procedures
and spectroscopic data for compounds 6b-d, and 7c-7d as
well as 1H and 13C NMR spectra for all compounds. Complete
results of the spectral binding and inhibition assays for
compounds 6b-d and 7c-7d. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA0746159
described in the text, absorbance changes (∆∆Abs ) ∆Abs396
-
∆Abs420) were plotted as a function of compound concentration, and
(35) Dawson, J.; Knowles, R. G. Mol. Biotechnol. 1999, 12, 275-279.
9
12570 J. AM. CHEM. SOC. VOL. 129, NO. 41, 2007