Probing inducible nitric oxide synthase with a pterin-ruthenium(II) sensitizer wire

Article abstract of DOI:10.1002/anie.200703743

(Figure Presented) Binding by the light: A photoactive ruthenium (II)-pterin wire designed to target the tetrahydrobiopterin (H₄B) binding site of inducible nitric oxide synthase (iNOS_heme) binds to iNOS_heme in a competitive fashion with the natural cofactor and other small molecules; thus it can be used to screen potential inhibitors. Rapid charge injection to the heme occurs upon 450-nm excitation of an iNOS _heme/wire conjugate.

Full text of DOI:10.1002/anie.200703743

Communications
DOI: 10.1002/anie.200703743
Enzyme Probes
Probing Inducible Nitric Oxide Synthase with a Pterin–Ruthenium(II)
Sensitizer Wire**
Edith C. Glazer, Yen Hoang Le Nguyen, Harry B. Gray, and David B. Goodin*
Nitric oxide synthase (NOS) is the primary biological source
of the ubiquitous signaling molecule nitric oxide (NO). The
enzyme utilizes tetrahydrobiopterin (H4B) as an essential
cofactor, which plays a key role in the catalytic conversion of
l-arginine to citrulline and NO.[1] Pterin has been shown to
serve both structural and catalytic roles in the enzyme by
affecting the monomer–dimer transition, promoting protein
stability, and forming a radical cation during catalytic
turnover.[2]
We have developed redox-active sensitizer wires to probe
the active sites of heme enzymes.[3] Herein, we describe a new
class of redox-active wires that target the pterin binding site of
NOS. The pterin group lies adjacent to the heme, and directly
interacts with the catalytic center through hydrogen bonds to
the heme propionate group. It is close in space, but physically
distinct from the face of the heme containing the arginine
Scheme 1. Synthesis of pterin probes. Reagents and conditions:
a) DMF dimethyl acetal (10 equiv), DMF, 508C, 2 h, 92%; b) ethyl
6-bromohexanoate (5 equiv), K2CO3 (2 equiv), DMF, 4 h, 68%; c) 1m
NaOH/DMF, 1 h, 82%; d) NH3/MeOH, 18 h, 85%; e) (with 4), 3-(3-
binding site, allowing both cofactor and substrate to bind
simultanously.[4,5] Thus, introducing a molecular wire at the
pterin site may allow photochemical triggering of enzyme
turnover, thereby offering an opportunity to study catalytic
dimethylaminopropyl)-1-ethylcarbodiimide (EDCI; 1.5 equiv), 4-dimeth-
intermediates and to shed new light on the cofactorꢀs role in
catalysis.[6] Herein we report the design and synthesis of a
ruthenium(II)–pterin wire, and investigate its interaction with
the heme domain of murine inducible nitric oxide synthase
(iNOSheme).
The wire combines two essential moieties: an analogue of
the cofactor to direct binding, and a redox-active sensitizer
that can be used for light-induced charge injection to generate
specific oxidation states of the heme (the pterin is believed to
provide an electron to the heme during catalysis).[7] The 6-
phenylpterin analogue was synthesized in the catalytically
inactive, fully oxidized form (1, Scheme 1). This moiety has
been shown to bind in a competitive manner with the fully
ylaminopyridine (DMAP; 1.5 equiv), pyridine, 12 h, 68%.
reduced cofactor.[8] The tether was attached to the pterin ring
at the 4-O position, which allows it to extend through the
iNOS active-site channel that is normally occupied by solvent
or other small molecules;[5] the linker length was chosen based
on modeling studies. In addition, we incorporated a ruth-
enium(II)–diimine sensitizer that acts as a luminescent bind-
ing probe, and as an excited-state oxidizing or reducing
agent.[9]
The pterin-linked wire 6 was synthesized from the
oxidized 6-phenyl pterin (1; see Scheme 1), which was
prepared according to the procedure of Storm et al. by an
Isay condensation reaction.[10] The 2-amino functionality was
protected as the (dimethylamino)methylene group to
improve solubility and facilitate further synthetic modifica-
tion.[11] Alkylation of the 4-hydroxy group was performed
under standard Williamson ether synthesis conditions. This
chemistry led to a combination of species alkylated at 4-O and
3-N which could be resolved using silica gel flash chromatog-
raphy.[12] Cleavage of both the ester and (dimethylamino)-
methylene protecting groups was accomplished in a single
step with NaOH in DMF (1m). The addition of the long-chain
alkane improved solubility, allowing the final conjugation to
the ruthenium(II) complex by esterification under standard
conditions; alternatively, it was converted to the amide 5 for
use as a model compound.[13]
[*] Dr. E. C. Glazer, Prof. D. B. Goodin
Department of Molecular Biology
The Scripps Research Institute
La Jolla, CA 92037 (USA)
Fax: (+1)858-784-2857
E-mail: dbg@scripps.edu
Dr. E. C. Glazer, Dr. Y. H. L. Nguyen, Prof. H. B. Gray
Department of Chemistry and Chemical Engineering
California Institute of Technology
Pasadena, CA 91125 (USA)
[**] We thank Yitzhak Tor, Doug Magde and M.G. Finn for synthetic
resources and many helpful discussions. Supported by NIH grant
GM070868 (to D.B.G. and H.B.G.), and NRSA fellowship
GM074406 (to E.C.G.) and the Ellison Medical Foundation (Senior
Scholar Award in Aging to H.B.G.).
Several spectroscopic methods were utilized to quantify
wire binding to iNOSheme. Upon binding H4B, the iNOS heme
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
898
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 898 –901

Angewandte
Chemie
undergoes a conversion from low spin to high spin, with a
complete shift obtained after incubation with both pterin and
l-arginine. Addition of wire 6 to iNOSheme affects a change in
the UV/Vis spectrum that is consistent with partial conversion
from the DTT-ligated bis(thiolate) complex into a high-spin
state (Figure 1). This change is more subtle than observed
Figure 2. The steady-state ruthenium(II) luminescence spectrum of 6
(10 mm, red) is quenched in the presence of equimolar iNOSheme
(black); it recovers upon addition of 10 mm (blue) and 100 mm (green)
4AH4B.
that the interaction of 6 with the enzyme competes with pterin
cofactors.
Time-resolved emission experiments also confirmed ruth-
enium(II)–pterin binding to the iNOS heme (see Supporting
Information, Figure S4).[3] The emission decay of the wire is
monoexponential in the absence of protein (t = (368 Æ 2) ns),
but becomes biexponential (t1 = 370 ns, t2 = (40 Æ 4) ns) in
the presence of iNOSheme. Fits of the decays provided
estimates of the ratio of enzyme-bound/free ruthenium(II),
giving a dissociation constant of approximately 4 mm (see
Figure 3). As in the steady-state experiment, addition of the
cofactor analogue 4AH4Btriggers dissociation of the wire
from the protein, and recovery of the longer-lived,
unquenched species (see Supporting Information, Figure S7).
The affinity of 6 for iNOSheme appears to be determined by
the pterin and its linker rather than the ruthenium(II)
sensitizer. Both the pterin core 1 and intermediate wire 5,
lacking the ruthenium(II) complex, are inherently fluores-
Figure 1. Reconstitution of substrate- and H4B-free iNOSheme with wire
6. iNOSheme (ca. 10 mm) was incubated at 258C with the pterin
analogue in the presence of 1 mm DTT, and the absorbance spectrum
was taken after 2 h (10 mm, green, 30 mm, blue, 50 mm, red). The
spectrum of the DTT-equilibrated sample (black, without wire) shows
the formation of a bis(thiolate) heme. The inset shows the binding
curve obtained from titration data, plotting the fractional difference in
absorption n (DDA/DDAtotal, DDA=DA400ÀDA459) as a function of
concentration.
with the natural cofactor, but has the same features in the
difference spectrum, with a maximum at 400 nm and a
minimum at 459 nm (see Supporting Information, Figure S1).
These results are consistent with other synthetic pterin
analogues that have been found to produce partial low-spin
to high-spin shifts in optical spectra.[14]
The affinity (Kd ꢀ 8 mm) of 6 for the protein was obtained
from a plot of n = DDA/DDAtotal versus concentration (see
Figure 1, inset). Binding of the natural cofactor causes several
structural and electronic modifications in the enzyme, includ-
ing adjustments in the dimerization interface as well as the
substrate and pterin binding pockets, and the binding is
known to be quite slow.[15] The time dependence of the
absorption changes in the difference spectra for wire 6 was
analyzed, and the spectroscopic changes were complete
within about 60 min, in good agreement with the rate of
binding of the natural cofactor (see Supporting Information,
Figure S2). Thus, the structural and electronic modifications
within the enzyme upon binding wire 6 are fully consistent
with those observed upon binding H4B.
The binding of wire 6 to iNOSheme was confirmed by
analysis of changes in ruthenium(II) emission. As shown in
Figure 2, the emission centered at 610 nm is dramatically
quenched upon addition of iNOSheme, owing to Förster energy
transfer. Furthermore, the ruthenium(II) emission is restored
upon addition of the natural cofactor H4B, or the higher-
affinity pterin analogue 4-amino-H4B(4AH 4B), indicating
Figure 3. The relative fraction of the fast phase of luminescence decay
for compound 6 (taken as fraction bound) as a function of
[iNOSheme]free. An estimate for the dissociation constant of 6 to
iNOSheme was obtained by fitting the curve to a single-site binding
model (Kd =(3.8Æ1.4) mm), assuming that the saturation value of the
fractional fast phase represents a 1:1 complex.
Angew. Chem. Int. Ed. 2008, 47, 898 –901
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
899

Communications
cent, and display marked quenching in the presence of protein
(see Supporting Information, Figure S3). The efficient
quenching of pterin fluorescence by the heme was used to
characterize wire binding to the protein. The 4 mm dissocia-
tion constant for 6 is in agreement with that for compound 5
((5.1 Æ 1) mm), which was obtained by steady-state fluores-
cence quenching (see Supporting Information, Figure S8); it
is significantly lower than the value (26 mm) reported for
pterin 1.[8] Thus, the addition of the tether rather than the
ruthenium(II) complex is likely to be responsible for the
increased binding affinity. Indeed, other pterin analogues
have been reported that show enhanced binding upon
alkylation at the 4-position,[8a,16] and wires designed for
cytochrome P450 from P. putida (P450cam) show higher
affinities than substrate alone.[3b] It is likely that the additional
hydrophobic contacts provided by the linker in the access
channel are responsible for each of these enhanced affinities.
A model for the interaction of wire 6 with iNOSheme is
shown in Figure 4. The pterin portion of the wire was overlaid
with the H4Bbound at the dimer interface, and the linker and
Figure 5. Photoreduction of iNOSheme using wire 6. Illumination of the
iNOSheme–wire complex under an atmosphere of CO causes an
increase in absorption at 420 nm, consistent with the formation of the
reduced CO-bound P420 species (black line, 0 min, blue line, 30 min).
1 mm TMPD was used for the experiment; no spectral change was
observed without TMPD, or without light. Chemical reduction with
Na2S2O4 in the dark results in production of a combination of P450
and P420 species (red line).
demonstrate the instability of the P450 form, and show
conversion into the P420 form with time or exposure to
450 nm light (see Supporting Information, Figure S9). It thus
appears that photolysis of the CO-bound heme in the
presence of the natural cofactor results in conversion into
the five-coordinate form; it follows that any P450 species
formed upon irradiation would be converted to the P420
form.
In summary, our pterin-based photoactive wire binds in
competition with native pterins to the heme domain of murine
iNOS. Charge injection from the metal complex should
facilitate mechanistic investigations of the enzyme, and may
help determine how the pterin acts as an essential redox-
active cofactor at specific points of the catalytic cycle.
Importantly, the wire also can be used as a fluorescent
probe for rapid screening of small molecule inhibitors that
target the pterin binding site.
Figure 4. Structural model of wire 6 docked into iNOSheme. Heme is
shown in red; for 6: gray C, red O, blue N, Ru.
ruthenium(II) conformations were adjusted to minimize
steric conflict. The distance between the heme edge and the
nearest diimine ring of the ruthenium(II) complex is 17 Š in
this model, which agrees well with the value of (15.7 Æ 1.4) Š
obtained by assuming that the quenched (40 ns) component
of the emission decay is attributable solely to Förster energy
transfer to the heme in the wire–enzyme conjugate.[17]
Upon steady-state illumination at 450 nm, we observe
reduction of the wire-bound enzyme in the presence of CO
and the reductive quencher TMPD (N,N,N’,N’-tetramethyl-
phenylenediamine). Changes in the UV/Vis spectra are
dependent both on the concentration of the wire and the
irradiation time. Photoreduction generates a species with an
absorption maximum at 420 nm, rather than one with a peak
at 450 nm associated with the cysteine-ligated FeII(CO) form
of the enzyme (see Figure 5). In a control experiment,
reduction of the wire-bound enzyme with sodium dithionite
produces a combination of P450 and P420 species, with the
P420 species predominating.[18] In addition, experiments
performed on the enzyme in the presence of 50 mm H4B
Received: August 15, 2007
Revised: September 7, 2007
Published online: December 17, 2007
Keywords: electron transfer · metalloenzymes ·
.
nitric oxide synthase · ruthenium · sensitizers
[1] W. K. Alderton, C. E. Cooper, R. G. Knowles, Biochem. J. 2001,
357, 593 – 615.
[2] C. C. Wei, B. R. Crane, D. J. Stuehr, Chem. Rev. 2003, 103, 2365 –
2383.
[3] a) J. J. Wilker, I. J. Dmochowski, J. H. Dawson, J. R. Winkler,
H. B. Gray, Angew. Chem. 1999, 111, 93 – 96; Angew. Chem. Int.
Ed. 1999, 38, 89 – 92; b) A. R. Dunn, I. J. Dmochowski, J. R.
Winkler, H. B. Gray, J. Am. Chem. Soc. 2003, 125, 12450 – 12456;
c) A. R. Dunn, A. M. A. Hays, D. B. Goodin, C. D. Stout, R.
Chiu, J. R. Winkler, H. B. Gray, J. Am. Chem. Soc. 2002, 124,
900
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 898 –901

Angewandte
Chemie
10254 – 10255; d) A. R. Dunn, W. Belliston-Bittner, J. R. Win-
kler, E. D. Getzoff, D. J. Stuehr, H. B. Gray, J. Am. Chem. Soc.
2005, 127, 5169 – 5173; e) W. Belliston-Bittner, A. R. Dunn,
Y. H. L. Nguyen, D. J. Stuehr, J. R. Winkler, H. B. Gray, J. Am.
Chem. Soc. 2005, 127, 15907 – 15915.
allows further chemistry to be carried out, but is still sufficiently
labile, in contrast to amide and thiol protecting groups.
[12] A more direct approach to the 4-O species was attempted using
Mitsunobu chemistry (see ref. [11]); however, it was found that
this route also led to a combination of O- and N-alkylated
species, but with an overall lower yield.
[13] The amide was used instead of the carboxylic acid to prevent any
undesired reactions with the protein.
[14] A. Presta, U. Siddhanta, C. Wu, N. Sennequier, L. Huang, H. M.
Abu-Soud, S. Erzurum, D. J. Stuehr, Biochemistry 1998, 37, 298 –
310.
[4] B. R. Crane, A. S. Arvai, D. K. Ghosh, C. Wu, E. D. Getzoff, D. J.
Stuehr, J. A. Tainer, Science 1998, 279, 2121 – 2126.
[5] C. S. Raman, H. Li, P. Martasek, V. Kral, B. S. S. Masters, T. L.
Poulos, Cell 1998, 95, 939 – 950.
[6] a) D. J. Stuehr, C. C. Wei, Z. Wang, R. Hille, Dalton Trans. 2005,
3427 – 3435; b) D. J. Stuehr, C. C. Wei, J. Santolini, Z. Wang, M.
Aoyagi, E. D. Getzoff, Biochem. Soc. Symp. 2004, 39 – 49.
[7] a) A. R. Hurshman, C. Krebs, D. E. Edmondson, B. H. Huynh,
M. A. Marletta, Biochemistry 1999, 38, 15689 – 15696; b) A. R.
Hurshman, M. A. Marletta, Biochemistry 2002, 41, 3439 – 3456.
[8] a) P. Kotsonis et al., J. Biol. Chem. 2001, 276, 49133 – 49141 (see
Supporting Information for full reference); b) H. Matter, P.
Kotsonis, O. Klingler, H. Strobel, L. G. Froehlich, A. Frey, W.
Pfleiderer, H. H. H. W. Schmidt, J. Med. Chem. 2002, 45, 2923 –
2941.
[9] J. Berglund, T. Pascher, J. R. Winkler, H. B. Gray, J. Am. Chem.
Soc. 1997, 119, 2464 – 2469.
[10] C. B. Storm, R. Shiman, S. Kaufman, J. Org. Chem. 1971, 36,
3925 – 3927.
[11] Q. Yao, W. Pfleiderer, Helv. Chim. Acta 2003, 86, 1 – 12. Pterins
are notoriously insoluble; we have found that the DMF dimethyl
acetal protecting group is far superior to many others in that it
[15] A. R. Hurshman, C. Krebs, D. E. Edmondson, M. A. Marletta,
Biochemistry 2003, 42, 13287 – 13303.
[16] L. G. Froehlich, P. Kotsonis, H. Traub, S. Taghavi-Moghadam, N.
Al-Masoudi, H. Hofmann, H. Strobel, H. Matter, W. Pfleiderer,
H. H. H. W. Schmidt, J. Med. Chem. 1999, 42, 4108 – 4121.
[17] Förster energy transfer calculations were performed as de-
scribed in ref. [3d]. See Supporting Information for further
details.
[18] The instability of the P450 form of NOS is well documented; it is
clear that the pterin plays a central role in determining the
degree of conversion to the P420 form. Pterin-free NOS is
converted over time completely to the P420 form, and variable
fractions of the two forms are observed with subsaturating levels
of H4Bor synthetic pterin analogues. See refs [7b, 15], and J.
Wang, D. J. Stuehr, D. L. Rousseau, Biochemistry 1995, 34,
7080 – 7087.
Angew. Chem. Int. Ed. 2008, 47, 898 –901
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
901