Synchronous photoinitiation of endothelial NO synthase activity by a nanotrigger targeted at its NADPH site

Article abstract of DOI:10.1021/ja067543e

We designed a new nanotrigger to synchronize and monitor an enzymatic activity interacting specifically with the conserved NADPH binding site. The nanotrigger (NT) combines a docking moiety targeting the NADPH site and a chromophore moiety responsive to l

Full text of DOI:10.1021/ja067543e

Published on Web 01/31/2007
Synchronous Photoinitiation of Endothelial NO Synthase
Activity by a Nanotrigger Targeted at Its NADPH Site
Edward Beaumont,†,‡ Jean-Christophe Lambry,†,‡ Cle´ment Gautier,†,‡
Anne-Claire Robin,§ Said Gmouh,§ Vladimir Berka,| Ah-Lim Tsai,|
Mireille Blanchard-Desce,§ and Anny Slama-Schwok*,†,‡
Contribution from the Unite´ 696, INSERM, Palaiseau France, Laboratory for Optics and
Biosciences, UMR CNRS 7645, Ecole Polytechnique, Palaiseau, France, Photonique
Mole´culaire, UMR CNRS 6510, UniVersite´ de Rennes I, France, and DiVision of Hematology,
Internal Medicine, UniVersity of Texas Houston Medical School, Houston, Texas
Received October 25, 2006; E-mail: Anny.Schwok@polytechnique.edu
Abstract: We designed a new nanotrigger to synchronize and monitor an enzymatic activity interacting
specifically with the conserved NADPH binding site. The nanotrigger (NT) combines a docking moiety
targeting the NADPH site and a chromophore moiety responsive to light excitation for efficient electron
transfer to the protein. Specific binding of the nanotrigger to the reductase domain of the endothelial nitric
oxide synthase (eNOSred) was demonstrated by competition between NADPH and the nanotrigger on the
reduction of eNOSred flavin. A micromolar Ki was estimated. We had monitored initiation of eNOSred
activity by ultrafast transient spectroscopy. The transient absorption spectrum recorded at 250 ps fits the
expected sum of the reduced and oxidized species, independently obtained by other chemical methods, in
agreement with a photoinduced electron transfer from the excited nanotrigger to the flavin moiety of
eNOSred. The rate of electron transfer from the excited state of the nanotrigger (NT*) to the protein is
estimated to be kET ) (7 ( 2) × 109 s-1 using the decay of oxidized eNOSred-bound nanotrigger compared
against prereduced eNOSred or glucose 6-P dehydrogenase as controls. This fast electron transfer bypasses
the slow hydride transfer to initiate NOS catalysis as shown by ultrafast kinetics using the eNOSred mutated
in the regulatory F1160 residue. The selective targeting of the nanotrigger to NADPH sites should allow
controlled initiation of the enzymatic activity of numerous proteins containing an NADPH site.
of the inducible NO synthase (iNOS).5 The molecular wires
developed in the group of H. Gray were designed to drive
Introduction
A better understanding of protein functions and interactions
often requires the ability to trigger and initiate a specific
biological activity. In the past years, “caged nucleotides” have
been developed to trigger an enzymatic activity with light. These
compounds are derivatives of a natural protein cofactor, such
as ATP or NADPH, in which light activation leads to cleavage
of the “caging” moiety.1-2 The released cofactor is then free to
diffuse toward its binding site within the protein. Actually, the
intrinsic limitation of these compounds to millisecond timescales
originates from the rate of diffusion of the active moiety to the
catalytic site in the protein, especially with proteins with high
catalytic turnover.3 Recently, such caged compounds have been
used with multiphotonic excitation4 and coupled to an inhibitor
enzymatic reactions with light, and the approach was applied
to the cytochrome P450 and more recently to iNOS.6-7 The Re
or Ru organometallic complex is tethered to a heme binding
moiety such as imidazole or an enzyme substrate.6-7 Although
the electron transfer occurred within 300 ps upon 355 nm
excitation, it was not related to a physiological activation of
iNOS, and in some cases, the wire inactivated the protein by
sterically hindering its active dimer formation.
Thus, the innovative device that triggers and synchronizes
biological molecules in their native state “instantaneously” both
in time and in space would set the “zero” time of the catalysis
and thus allow time-resolved measurements of the enzymatic
activity in an ensemble of molecules. In this work, we develop
a novel approach to trigger enzymatic activity by light in a
synchronous manner. We designed a nanotrigger possessing two
main characteristics. The first one is the specific targeting of
the nanotrigger to the NADPH site within proteins. The second
† Unite´ 696, INSERM.
‡ Ecole Polytechnique.
§ Universite´ de Rennes I.
| University of Texas Houston Medical School.
(1) Stoddard, B. L.; Cohen, B. E.; Brubaker, M.; Mesecar, A. D.; Koshland,
D. E. Nat. Struct. Biol. 1998, 5, 891-897.
(2) Salerno, C. P.; Magde, D.; Patron, A. P. J. Org. Chem. 2000, 65, 3971-
3981.
(5) Perdicakis, B.; Montgomery, H. J.; Abbott, G. L.; Fishlock, D.; Lajoie,
G. A.; Guillemette, J. G.; Jervis, E. Bioorg. Med. Chem. 2005, 13 (1), 47-
57.
(3) Ursby, T.; Weik, M.; Fioravanti, E.; Delarue, M.; Goeldner, M.; Bourgeois,
D. Acta Crystallogr., Sect. D 2002, 58, 607-14.
(4) Allin, C.; Ahmadian, M. R.; Wittinhofer, A.; Gewert, K. Proc Natl. Acad.
Sci. U.S.A. 2001, 98, 7754-9.
(6) Dunn, A.; Dmochowski, I. J.; Winkler, J. R.; Gray, H. B. J. Am. Chem.
Soc. 2003, 125, 12450-12456.
(7) Bellinston-Bittner, W.; Dunn, A. R.; Le Nguyen, Y. H.; Stuehr, D. J.;
Winkler, J. R.; Gray, H. B. J. Am. Chem. Soc. 2005, 127, 15907-15915.
9
2178
J. AM. CHEM. SOC. 2007, 129, 2178-2186
10.1021/ja067543e CCC: $37.00 © 2007 American Chemical Society

Selective Targeting of Nanotrigger to NADPH Sites
A R T I C L E S
mide and an aromatic residue have been suggested.13 These
motions are often involved in the regulation of the catalysis, as
for example at the level of nNOS, Phe1395 by blocking access
to FAD.12 Thus, any designed molecule should overcome this
physical blockage and rapidly direct the photoinduced electron
transfer to the natural electron acceptor, FAD, in the case of
nNOS.
Scheme 1. Design of the Nanotrigger
Our design conserved the nucleotide recognition motif of
NADPH. This should confer the proper binding of the nano-
trigger to the NADPH site. The nicotinamide part was replaced
by a chromophore moiety, which, upon photoactivation, should
transfer electron(s) to the flavin center of the protein and initiate
the catalysis by selective light excitation (Scheme 1B). Pre-
liminary modeling studies optimized both the nature and the
length of the conjugated chromophore to fit in the NADPH
cavity of the cytochrome P450 reductase.13-14 Moreover, an
extended conformation of the nanotrigger by means of conju-
gated double bonds and a peptidic linker between chromophore
and nucleotidic moieties could “lock” the enzyme in an active
conformation, favorable for the electron transfer. The extended
conformation of the nanotrigger should also maximize the
delocalization of the π electrons on the chromophore, with
subsequent favorable properties of one and two photon
absorption.14-16 In this work, the proof of concept was tested
on the reductase domain of the endothelial nitric oxide synthase,
17
eNOSred, highly homologous to nNOSred and P450 reduc-
tase. First, the specific binding of the nanotrigger to the NADPH
site was demonstrated by measuring the rate of flavin reduction
by NADPH in stopped-flow experiments. These data showed a
competitive binding of the nanotrigger and NADPH in the
cofactor binding site. Moreover, the photophysical and photo-
chemical characteristics of the bound nanotrigger to eNOSred
were measured by ultrafast transient absorption in pump-probe
experiments. The transient spectrum recorded 250 ps after the
pulse fits the expected sum of the reduced and oxidized species,
independently obtained by another chemical method (reduction
with SbCl5). The decay of the transient absorption of the
nanotrigger bound to eNOSred was enhanced relative to the
control, and the rate of the electron transfer was estimated to
be kET ) (7 ( 2) × 109 s-1. Moreover, steady-state irradiation
also showed that nanotrigger NT triggering allowed subsequent
heme reduction of WT nNOS that became (partly) nitrosylated
in the presence of Ca2+/calmodulin. Thus, our novel nanotrigger
initiated eNOS activity in a synchronous manner by light.
characteristic is the ability to trigger the redox reaction leading
to the enzymatic activity by a short laser pulse. This nanotrigger
is a photoactivatable analogue of NADPH and should interact
specifically with the NADPH site, present in numerous proteins,
without the limitation of diffusion as encountered by the caged
NADPH. Moreover, the catalysis would be rapidly initiated at
the native state of the enzyme, and time-resolved spectroscopy
can be applied to record the kinetic steps.
The design of the nanotrigger was based on the conserved
motifs of the ubiquitous NADPH binding sites. Crystallographic
data of NADPH-containing proteins8-9 showed that the main
recognition motifs of NADPH in its binding site are found in
the nucleotidic part, with the sugar, adenine rings, and the final
phosphate linked by numerous H-bonding interactions with the
proteins, in particular by means of a highly conserved arginine
residue.8-10 The nicotinamide subunit of NADPH is on the other
hand often flexible, closer to the solvent interface, and its
recognition within the NADPH site is quite variable among
various proteins (Scheme 1A). NADPH can adopt an extended
conformation with the adenine and nicotinamide rings assuming
a perpendicular orientation and an end-to-end distance of
roughly 15 Å. Alternatively, NADPH can assume a fold back
conformation by the nicotinamide moiety stacking on the purine
ring.11-12 In other instances, coupled motions of the nicotina-
Materials and Methods
Materials: NADPH (99.8% purity), Tris-HCl, potassium ferricya-
nide, glycerol, and EDTA were purchased from Sigma. High purity
argon (99.9995%) was obtained from Alpha Gas.
Synthesis: The nanotrigger (NT) was prepared as described in
Scheme 2.15 A coupling reaction between the chromophoric moiety A
and the adenosine moiety B was first performed. The acetal protection
was then removed with trifluoroacetic acid, and a phosphate group was
(8) Carugo, O.; Argos, P. Proteins 1997, 28, 10-28.
(13) Zhang, J.; Martasek, P.; Paschke, R.; Shea, T.; Silver Masters, B. S.; Kim,
J. J. J. Biol. Chem. 2001, 276, 37506-13.
(9) Duax, W. D.; Pletnev, V.; Addlagatta, A.; Bruenn, J.; Weeks, C. M.
Proteins: Struct., Funct., Genet. 2003, 53, 931-943.
(14) The molecular modeling of NT in P450 reductase will be detailed elsewhere
(Slama-Schwok, A. and Lebret, M.).
(10) Tiso, M.; Konas, D. W.; Panda, K.; Garcin, E. D.; Sharma, M.; Getzoff,
E. D.; Stuehr, D. J. J. Biol. Chem. 2005, 280, 39208-39219.
(11) Malito, E.; Alfieri, A.; Frasije, M. W.; Mattevi, A. Proc Natl. Acad. Sci.
U.S.A. 2004, 101, 13157-62.
(12) Garcin, E. D.; Bruns, C. M.; Lloyd, S. J.; Hosfield, D. J.; Tiso, M.; Gachhui,
R.; Stuehr, D. J.; Tainer, J. A.; and Getzoff, E. D. J. Biol. Chem. 2004,
279, 37918-37927.
(15) The detailed synthesis is described elsewhere (Robin, A. C.; Gmouh. S.;
Mongin, O.; Jouikov, V.; Werts, M. H. V.; Gautier, C.; Slama-Schwok,
A.; Blanchard-Desce, M. Submitted). A french patent was deposited No.
0511914: Synchronous trigger of the catalysis of NADPH enzymes by a
photoactivable analog of NADPH: A. Slama-Schwok, M. Blanchard-Desce,
S. Gmouh, J.-L. Martin.
9
J. AM. CHEM. SOC. VOL. 129, NO. 7, 2007 2179

A R T I C L E S
Beaumont et al.
Scheme 2. Synthetic Route to Nanotrigger NTa
addition of NT to the reduced protein caused unavoidable partial
reoxidation of the flavin thus yielded maximally a 60% net reduction.
Time-Resolved Spectroscopy. Multicolor pump-probe spectros-
copy was performed with a 500-Hz setup based on a home-built TiSa
oscillator and a 1-kHz regenerative amplifier (Spitfire, Spectra Physics),
delivering 50-fs 0.5 mJ at 810 nm. A synchronized chopper allowed a
reduction of the repetition rate to 500 Hz. Pump pulses centered at
405 nm were generated with a 0.5 mm BBO crystal used in the SHG
configuration. White light probe pulses ranging from 430 to 700 nm
were generated by focusing part of the fundamental on a CaF2 window
and separated in test and reference beams. These beams were, after
passing through the sample, dispersed and detected shot-by-shot on a
CCD camera configured as a dual-array detector. The sample was
renewed between subsequent pulse pairs by rotation perpendicularly
to the beams. 10 scans were usually recorded in a single experiment in
two timescales: 1-10 ps and 10-600 ps. An average of three
experiments was used for Figure 5. Carbon monoxide bound to reduced
myoglobin was a reference for an undecaying signal at 200-600 ps.19
Analysis of the data was performed by singular value decomposition
(SVD) of the time-wavelength matrix and by exponential fit of the
single wavelength data.
Results
1. Competition between the Nanotrigger and NADPH for
the NADPH Site of eNOSred. The reductase domain of eNOS
contains 1 equiv of FAD and FMN and binds NADPH very
similarly to the case of cytochrome P450 reductase (CPR).
Detailedstopped-flowstudiesonCPRhavebeenreported.20-21Previous
studies on nNOS22-26 showed that FAD reduction by NADPH
occurs in three distinct kinetic steps. In the first step, NADPH
binds to the enzyme; the second step involves a hydride transfer
from NADPH to FAD, with the third one being an intramo-
lecular electron transfer between the two flavins. The rebinding
of the oxidized NADP+ is inhibitory to steps two and/or three,
whose rates depend upon NADP+ release. We hypothesized that
closely related eNOS would undergo qualitatively similar
binding and reduction steps as those for nNOS. Thus, our
nanotrigger should interfere with NADPH binding in the first
step and affect the second and third kinetic steps as well. We
used the isolated reductase domain of eNOS in this study.17
The flavin reduction was monitored at 455 nm after rapid mixing
with either NADPH alone or NADPH in the presence of
nanotrigger. Absorption changes at 340 nm due to NADPH
consumption were also recorded. We always used a large excess
of NADPH over eNOSred to ensure “clean” pseudo-first-order
reactions. Typical data of eNOSred reduction by NADPH are
shown in Figure 1. The absorption changes were detected in
several timescales and analyzed by three exponentials and a
constant term at each NADPH concentration.22 We first found
a very fast component k1 of 190 ( 40 s-1 at 455 nm and of the
order of 400 s-1 at 340 nm using 100 µM NADPH (Figure 1A,
insert). Given the resolution of our stopped-flow measurement
(≈ 2 ms), the error on this last figure is high and we could not
a Reagents and conditions: (a) HBTU/DMF/Et3N; (b) TFA/THF/H2O;
(c)(i) i(Pr)2N-P(OBn)2/tetrazole, (ii) O2/CH2Cl2, (iii) Me3SiBr/CH2Cl2.
inserted with dibenzyldiisopropylphosphoramidite. The resulting phos-
phoramidate was oxidized, and the benzyl protective groups were
removed using bromotrimethylsilane to afford NT as a mixture of two
isomers (phosphate group on the 2 or 3 hydroxyl of the sugar). A
lipophilic model chromophore (MC) has been used for electrochemical
characterization in solution performed in a similar way as described.16
Expression of the Enzymes. The eNOS reductase domain was
overexpressed in yeast and purified as reported previously.17 F1160A
mutant was prepared by site-directed mutagenesis in the lab of Dr.
Tsai, UTH MS. Purified nNOS was purchased from Calbiochem.
Purified glucose 6-phosphate dehydrogenase (Baker yeast) was pur-
chased from Sigma.
Preparation of Samples for Stopped-Flow Measurements: Argon-
saturated solutions of 20 µM eNOSred alone or in the presence of the
nanotrigger were prepared apart from oxygen-free solutions of NADPH
(200 µM to 10 mM) and used for stopped-flow experiments. Removing
oxygen provided a better stability of the reducing agent. The stopped-
flow experiments were performed on an Edinburgh apparatus with
single wavelength detection at 455 nm or 340 nm with a dead time of
about 2 ms. The absorption decays were usually monitored at 1, 5,
and 10 s time scales. The data have been fitted by three exponentials
and a constant term using the Origin software.
Preparation of Samples for Ultrafast Kinetics. A concentrated
solution of proteins (10 to 20 µM) in 50 mM Tris-HCl buffer containing
150 mM NaCl at pH ) 7.5 and 5% glycerol was mixed with the NT
in an optical cuvette with a 1 mm path length. The extent of prereduction
achieved by light irradiation of an air-free solution of eNOSred
containing 5 mM EDTA was assessed by calculating the ratio of the
OD at 455 nm.18 We succeeded to obtain fully reduced eNOSred
indicated by a maximal decrease of A455. However, the subsequent
(19) Petrich, J.; Poyart, C.; Martin, J. L. Biochemistry 1988, 27, 4049-4060.
(20) Gutierrez, A.; Lian, L. Y.; Wolf, C. R.; Scrutton, N. S.; Roberts, G. C. K.
Biochemistry 2001, 40, 1964-1975.
(21) Grunau, A.; Paine, M. J.; Ladbury, J. E.; Gutierrez, A. Biochemistry 2006,
45, 1421-34.
(22) Knight, K.; Scrutton, N. S. Biochem. J. 2002, 367, 19-30.
(23) Dunford, A. J.; Marshal, K. R. L.; Munro, A. W.; Scrutton, N. S. Eur. J.
Biochem. 2004, 271, 2548-2560.
(16) Porre`s, L.; Alain, V.; Hapiot, P.; Blanchard-Desce, M. Phys. Chem. Chem.
Phys. 2003, 5, 4576-4582.
(17) Du, M.; Yeh, H. C.; Berka, V.; Wang, L. H.; Tsai, A. L. J. Biol. Chem.
(24) Daff, S. Biochemistry 2004, 43, 3929-32.
2003, 278, 6002-11.
(25) Wolthers, K. R.; Scrutton, N. S. Biochemistry 2004, 43, 490-500.
(26) Yamamoto, K.; Kimura, S.; Shiro, Y.; Iyanagi, T. Arch. Biochem. Biophys.
2005, 440, 65-78.
(18) Traber, R.; Kramer, H. E.; Hemmerich, P. Biochemistry 1982, 21, 1687-
93.
9
2180 J. AM. CHEM. SOC. VOL. 129, NO. 7, 2007

Selective Targeting of Nanotrigger to NADPH Sites
A R T I C L E S
Figure 2. Selective excitation of the nanotrigger compared to the absorption
of NADPH and eNOSred. (Upper panel) Dotted line, NADPH; dashed line,
reductase domain of eNOS (eNOSred); full line, nanotrigger in buffer
solution; note the much higher extinction coefficient of NT as compared to
both NADPH and eNOSred. (Bottom panel) The (equilibrium) absorption
maxima were found at λmax) 382 nm and λmax ) 361 nm for NT bound to
eNOSred and to glucose 6-phosphate dehydrogenase (G6PD), respectively,
shifted from λmax ) 376 nm for free NT in buffer.
decreased with the increase in nanotrigger concentration. The
observed k2 rate, the step attributed to the hydride transfer
reaction between NADPH (bound to the protein in step 1) and
k3 rates were best fitted by a competitive inhibition model that
keeps the same limit values as those obtained in the absence of
NT: k2 ) 115 ( 20 s-1 , k3 ) 8.6 ( 1.6 s-1. The inhibition
constant used to obtain the best fits (Figure 1B) is Ki ) 7 ( 3
µM. Thus, the nanotrigger is an efficient competitor of eNOS
catalysis initiated by NADPH.
2. Photophysics of the Bound Nanotrigger to eNOSred.
a. Selective Excitation of the Nanotrigger in the Presence of
eNOSred. Figure 2 compares the absorption spectra of the free
nanotrigger in buffer and of NADPH. The nanotrigger possesses
a ca. 10 times higher extinction coefficient of ꢀ ) 6 × 104 M-1
cm-1 at 376 nm relative to ꢀ ) 6.22 × 103 M-1 cm-1 at 340
nm of NADPH.28 Furthermore, the absorption of the nanotrigger
is well separated from that of the flavins of eNOSred,22-24
allowing a selective excitation of the nanotrigger in the spectral
range from 370 to 420 nm, and especially for the KHz laser
(the 405 nm wavelength we used). The bound nanotrigger
exhibits a broader absorption than that of the free NT, red-shifted
to λmax) 382 nm upon binding to eNOSred (Figure 2). Likewise,
NT binding to eNOSred induces a red-shift of its fluorescence
emission peak from 523 nm (free NT) to 533 nm (bound NT)
(Figure 3A, insert). By monitoring these fluorescence changes,
a binding constant of K ) (8 ( 2) × 105 M-1 was obtained.
This high affinity of NT for eNOSred insured that at least 80%
of the nanotrigger was bound to the protein at a 1/1 stoichi-
ometry used in the laser experiment described below.
Figure 1. Competition of NT with NADPH on the NADPH site of
eNOSred. (A) Raw decay of eNOSred flavin absorption monitored at 455
nm after addition of 2.15 mM NADPH to a solution of 10 µM eNOSred in
the presence of 0, 10, 50, and 120 µM nanotrigger; Insert: raw decay of
NADPH consumption at 340 nm after addition of 100 µM NADPH to 8
µM eNOSred alone (black) and with 8 µM NT (orange); the fits of the
experimental data are shown as black lines. (B) Fit of the observed k3 rate
constant at 455 nm according to a competitive model (kobs ) kmax
*
[NADPH]/{[NADPH] + K(NADPH) * [1 + [NT]/Ki]}. The nanotrigger is
taken as a competitive inhibitor, using kmax ) 8.1, 8.1, and 9.2 s-1 at [NT]
) 0, 10, and 50 µM, respectively, and Ki ) 6.1 ( 1.5 and 7.0 ( 1.4 µM
at [NT] ) 10 and 50 µM.
determine this first component at higher NADPH concentrations.
The second and third decay components were slower, k2 )
14 ( 4 s-1 while k3 ) 1.5 ( 0.5 s-1 at 455 nm using 100 µM
NADPH. These values determined for eNOSred are consistent
with the four observed rates of 214, 27, 3.7, and 0.3 s-1 reported
for 5 µM nNOS (without calmodulin) mixed with 50 µM
NADPH.22 Other authors reported rates of the main reduction
phase in nNOS in the presence of 50 µM NADPH to be 25.8
s-1 at 455 nm.27 The observed rate constants k2, k3 were found
to follow a saturation behavior as a function of the NADPH
concentration (Figure 1B and data not shown) as previously
shown for other flavoproteins.25 Peak values of 110 ( 15 s-1
and 8.6 ( 1.5 s-1 were determined for k2 and k3 at 455 nm,
respectively.
b. Transient Absorption of Bound NT to eNOSred. Need
for a Proper Control. The excited-state absorption and lifetime
of the nanotrigger are expected to depend on the solvent as
reported for many molecular wires.16,29 Thus, NT has to be in
an environment resembling the NADPH binding site for an
accurate measurement of the rate of its natural decay kN. The
The decay of NADPH oxidation at 340 nm and of the flavin
absorption at 455 nm resulting from the addition of NADPH
became slower in the presence of NT (Figure 1A). Figure 1B
details the analysis obtained with k3.22-26 The decay rates
(28) Racker, E. Methods Enzymol. 1955, 2, 272.
(29) Laage, D.; Plaza, D.; Blanchard-Desce, M.; Martin, M. M. Photochem.
Photobiol. Sci. 2002, 1, 526-35.
(27) Guan, Z. W.; Kamatani, D.; Kimura, S.; Iyanagi, T. J. Biol. Chem. 2003,
278, 30859-68.
9
J. AM. CHEM. SOC. VOL. 129, NO. 7, 2007 2181

A R T I C L E S
Beaumont et al.
excited-state properties without electron transfer is thus (see
Discussion):
NT + hV f NT*
kET
NT* + F 8 NT+ + F-
eNOSred with oxidized FAD
kN
NT* + F-98 NT + F- + hV′
eNOSred with prereduced FAD: Control 1
kN
NT*
98 NT + hV′ G6PD: Control 2
Transient absorption of the bound NT* to oxidized eNOSred
compared to the controls. Figure 3 compares the transient
spectra monitored at 3 ps of the bound NT* to the oxidized
eNOSred and NT* in the controls. Figure 3A shows that the
NT* maximum was sensitive to the redox state of eNOSred,
peaking at 670 nm in the oxidized eNOSred and shifted to 621
nm when eNOSred was prereduced. The excited-state absorption
of NT in G6PD peaked at 633 nm while NT* in 30% glycerol,
a solvent mimicking the protein interior, presented a maximum
at 623 nm (Figure 3A). A minimum was observed at 490-510
nm in all samples. This minimum was attributed to the
stimulated emission by comparison with the steady-state fluo-
rescence spectrum (Figure 3A, insert).
For NT bound to oxidized eNOSred, the transient absorption
decayed in two timescales, ca. 10 and 200 ps (Figure 4A, insert).
In parallel with the fast 10-ps decay, a progressive blue-shift
of the maximum from 670 to 643 nm was observed (see
Supporting Information Figure S1). This broad 643-nm peak
then decayed with no further spectral shift (Figure 3B). The
transient absorption spectrum was also modified in the 450-
500 nm range as shown by the comparison of the spectra
recorded at 3 and 200 ps. Most of the absorption decayed at
500 nm but distinct minima at ca. 450 and 475 nm were
simultaneously formed. This spectral region corresponds to the
absorption of the (oxidized) flavin (FAD). The insert of Figure
4A shows the parallel decays of the transient absorption of NT*
monitored at 640 nm, 600 nm, and then 478 nm where FAD
absorbs. Table 1 summarizes the fit of the data with three
exponentials and a constant value. The main decay was
characterized by τ3 ) 110 ( 26 ps, reaching subsequently an
asymptotic value (Figure 4A). This main component accounted
for ∼50-60% of the signal at 600 nm, whereas shorter time
scale components τ1 ) 1.3 ( 0.5 ps and τ2 ) 7 ( 3 ps
represented 30% and ∼10% of the total absorption changes,
respectively (insert).
The transient spectrum of NT in prereduced eNOSred decayed
more slowly than in oxidized eNOSred (compare Figure 4A
and B). The absorption maximum decayed away in hundreds
of picoseconds, characterized by τ3 ) 400 ( 100 ps obtained
for eNOSred at 60% prereduction. Moreover, in the case of
prereduced eNOSred, the slow decay monitored at 620 and 600
nm differed drastically from the fast decay followed at 478 nm
(Figure 4B and Table 1), without correlated decays observed
for NT* and FAD in oxidized eNOSred.
Figure 3. Transient spectra probed at 3 and 200 ps after the pump pulse
for the bound NT to oxidized eNOSred compared to the controls. (A)
Excited-state absorption recorded at a delay of 3 ps; note the large shift of
the maxima of the (raw) spectra when NT is bound to oxidized eNOSred
(λmax ) 670 nm, +) compared to NT in prereduced eNOSred (λmax ) 621
nm, black dashed) and NT in G6PD (λmax ) 633 nm, blue dashed); note
the similar maximum at 623 nm observed when the nanotrigger is in 30%
glycerol (dashed dotted). (Insert) Fluorescence spectrum of free NT (solid)
and NT bound to eNOSred (dashed), consistent with the minimum observed
in the transient spectrum at 490-510 nm. (B) Similar spectra as those in
part A but probed 200 ps after the pump.
latter is required to extract the electron-transfer rate between
NT* and FAD (F) from the observed decay rate: kobs ) kN +
kET. To confer the proper (proteic) environment to NT and NT*
with minimal disturbance, we utilized the nanotrigger bound to
prereduced eNOSred which already contains fully reduced flavin
(F-) that cannot be further (photo) reduced. The prereduction
of the protein was achieved by light irradiation of the protein
in the presence of 5 mM EDTA.18 A maximal extent of 60%
reduction could be reached by this method. This constituted our
first control. The protein glucose 6-phosphate dehydrogenase
(G6PD) was used as an additional control. This enzyme bears
an NADP site but lacks a flavin as the electron acceptor.30 The
shift of the absorption maximum to λmax ) 361 nm from λmax
) 376 nm for free NT provided evidence of NT binding to
G6PD (Figure 2). A good model for the determination of NT*
We also used NT bound to G6PD to monitor the rate of decay
of NT* in the absence of electron transfer. The insert of Figure
4B compares the decays of NT* in prereduced eNOSred with
that in G6PD at 600 and 478 nm. Both controls presented a
similar slow decay in the hundreds of picoseconds, characterized
(30) Kotaka, M.; Gover, S.; Vandeputte-Rutten, L.; Au, S, W. N.; Lam,
V. M. S.; Adams, M. J. Acta Crystallogr., Sect. D 2005, 61, 495-504.
9
2182 J. AM. CHEM. SOC. VOL. 129, NO. 7, 2007

Selective Targeting of Nanotrigger to NADPH Sites
A R T I C L E S
Figure 5. Comparison of the experimental and calculated transient
spectra: left scale indicates absorbance changes of spectral raw data
monitored 250 ps after the pulse, crosses; right scale represents the sum of
(60% NT cation + 40% dication + (reduced - oxidized flavin) - (reduced
NT), solid line; the experimental data were corrected for the direct excitation
of the protein. The spectrum of the cation of the NT chromophore was
generated by chemical oxidation with SbCl5 in CH2Cl2 (dashed line, the
dication is omitted for clarity); the dotted line represents the absorption
difference of the (unprotonated) semireduced - oxidized flavin.
reductase domain of eNOS in which Phe1160 was replaced by
Ala or F1160A mutant eNOSred. NT* bound to F1160A yielded
τ2 ) 20 ( 4 ps and τ3 ) 190 ( 20 ps (Table 1). NT bound to
F1160A showed a similar τ2 relative to that found in NT bound
to prereduced eNOSred, consistent with lacking complex
formation in the mutated F1160A as in prereduced eNOSred.
Consequently, no significant spectral shifts were observed (data
not shown). These data supported the hypothesis of an excited-
state complex between NT* and Phe1160 attributable to a minor
component, τ2. Moreover, a swinging of Phe1160 away from
both NADPH and FAD is likely to make room for the bound
nanotrigger, accounting for the main component of NT*.
The attribution of the main decay step (third step) to ET
implies that the transient spectra monitored 250 to 600 ps after
NT excitation correspond to the sum of the oxidized and reduced
species of the donor and acceptor pair corrected for the bleaching
of the reduced NT and oxidized eNOSred. NT is able to give
electron(s) but unable to transfer a proton to the flavin. It is
thus assumed that FAD is reduced but not protonated in the
experiment.31 The spectra of the cation and dication species of
NT were determined by chemical methods. The nanotrigger was
designed to reduce proteins with one or two electrons, and two
oxidation states have been reported for the related chromophore
MC.16 The redox activity of NT is localized in the chromophore
moiety, justifying our attempts to oxidize it and the further
assumption that the oxidized species of NT chromophore are
identical to that of the whole NT molecule. The NT chro-
mophore was chemically oxidized with SbCl5 yielding the
dication and monocation, the latter absorbing at λmax) 640 nm.
Figure 5 shows the fit of the experimental spectrum monitored
at 250 ps (crosses) with the calculated sum of the absorption of
the oxidized nanotrigger species plus the reduced flavin. This
sum estimated correctly the experimental maxima and minima
of the transient difference spectrum; discrepancies around 575
nm introduced uncertainties in the extent of the dication, ranging
from 25% to 40% of the oxidized NT, varying the corresponding
amount of the monocation from 75 to 60% accordingly.
Figure 4. Faster decay of the transient absorption of the bound nanotrigger
to oxidized eNOSred (A) compared to the controls (B). (A) (Raw) decays
observed at 640 nm (magenta), 600 nm (green), and 478 nm (cyan) with
NT in oxidized eNOSred; the main panel shows data recorded at the time
scales of 50 and 600 ps, and the insert presents the same experiments with
10 and 200 ps timescales. (B) Decays monitored at 620 nm (red), 600 nm
(green), and 478 nm (cyan) of NT in the presence of prereduced eNOSred
at the extent of 39% reduction; the insert presents the same experiments at
10 and 200 ps timescales (with the same colors); superimposed in the insert,
the decay of NT* bound to G6PD (shown in black) is monitored at 600 nm
(positive signal) or at 478 nm (second control); the full lines represent the
fit of the data summarized in Table 1. [eNOSred] ) 12 µM ) [NT],
50 mM Tris buffer at pH ) 7.5 containing 150 mM NaCl, 5% glycerol;
[G6PD] ) 117 µM, [NT] ) 25 µM in buffer containing 1 mM glucose
6-phosphate and 0.3 mM MgCl2.
by τ3 ) 400 ( 60 ps but differed by the value of the second
lifetime τ2 ) 54 ( 15 ps, longer in G6PD than in prereduced
eNOSred τ2) 20 ( 7 ps (Table 1).
The shorter lifetime τ3 ) 110 ( 26 ps for NT* in oxidized
eNOSred than in the two controls (τ3 ) 400 ( 60 ps) supported
the attribution of the third kinetic phase to electron transfer.
The minor component of NT*, τ2 ) 7 ( 3 ps, shortening in
oxidized eNOSred as compared to prereduced eNOSred or NT*
in G6PD remained puzzling. The large spectral shift only
observed during that time scale for NT* in oxidized eNOSred
suggested a possible excited-state complex. One critical residue
inserted between the NADPH and FAD in eNOSred is Phe1160.
To confirm the above attribution of electron transfer to the main
(third) step and to support the hypothesis of an excited-state
complex between NT* and Phe1160, we used the oxidized
(31) Sakai, M.; Takahashi, H. J. Mol. Struct. 1996, 379, 9-18.
9
J. AM. CHEM. SOC. VOL. 129, NO. 7, 2007 2183

A R T I C L E S
Beaumont et al.
Table 1. Kinetic Parameters for the Decay of NT* Bound to eNOSred and to Controls at Various Wavelengthsa
NT in protein
% prereduction
λ
(nm)
∆ODmax
A1
τ
1 (ps)
A2
τ2 (ps)
A3
τ3 (ps)
cte
prereduced
eNOSred
39
600
0.0255
0.20
1.1
0.35
142
0.45
551
0
620
510
600
620
510
478
600
478
600
0.0281
-0.0255
0.0271
0.40
0.69
0.23
0.39
0.58
0.76
0.40
0.80
0.28
1.7
1.7
0.8
0.8
1.8
2.0
2.0
1.8
0.6
0.39
0.15
97
13
0.21
0.15
0.77
0.61
0.28
0.24
0.35
0.20
0.44
673
471
347
320
400
447
410
400
106
0
0
0
0
0
0
0
0
60
0.032
-0.0196
-0.0065
0.020
-0.0072
0.0254
0.14
0.25
0.10
27
54
6.2
G6PDb
-
eNOSred
0
0.18
(oxidized)
620
640
510
478
600
0.0294
0.0320
-0.0041
-0.0045
0.0163
0.33
0.40
0.85
0.80
0.26
1.3
1.8
2.0
1.1
2.1
0.07
0.03
8.0
7.0
0.50
0.46
0.14
0.06
0.56
118
101
110
107
168
0.10
0.10
0.01
0.13
0.03
F1160A
(oxidized)
0
0.14
0.16
27
20
620
0.0264
0.41
2.1
0.42
200
0.01
a The sum of the amplitude terms A1 + A2 + A3 + cte was normalized to 1. [eNOSred] ) (10-20) µM, [eNOSred]/ [NT] ) 1. b [G6PD] ) 117 µM,
[NT] ) 25 µM in buffer containing 1 mM Glucose 6-phosphate and 0.3 mM MgCl2.
Figure 6. Light-induced reduction of the flavin and heme of WT eNOS by NT. An aerated solution of 3.4 µM nNOS containing 1 mM NOHA, 50 µM BH4,
Ca2+/calmodulin (Cd), and 5µM NT was irradiated at 400 nm with a 50 W halogen lamp equipped with a BG12 interference filter. The left panel shows the
absorption spectra recorded before (dashed dotted) and after (solid) irradiation. Note the decrease in the flavin bands and the increase of the broad visible
R band, consistent with reduction of the heme. The difference spectrum formed by irradiation, shown in the upper right panel (solid), presents similar
maxima in the Soret and R bands as that obtained by addition of NADPH to nNOS (2.5 µM) in the presence of NOHA, BH4, and Cd shown as a reference
(dashed). The bottom right panel shows the difference spectrum obtained after irradiation of nNOS in the absence of Ca2+/calmodulin in the presence of NT
(solid) or obtained with NADPH in the absence of NT and light (dashed). Note that in this case only the flavins were reduced by NT, the absence of
calmodulin blocking the (photoinduced) electron flow from the reductase to the oxygenase domain.
c. Light Irradiation of nNOS in the Presence of NT. The
above ultrafast kinetics demonstrated the electron transfer from
NT to FAD after excitation with a laser pulse. To ensure that
this excitation triggered the electron flow associated with NOS
catalysis, steady-state irradiation experiments were performed
in an aerated solution of full length nNOS in the presence of
N-hydroxyarginine, NOHA, and BH4. We used nNOS instead
of eNOS because the ferrous nitrosyl complex was shown to
be formed in the steady-state catalysis of nNOS.30 The difference
spectra monitored before and after addition of NADPH (in the
absence of NT and light) were compared to those obtained
before and after light irradiation in the presence of NT (without
NADPH). In the absence of calmodulin, thus the electron flow
from the reductase domain cannot activate the oxygenase
domain; consequently, only the flavins got reduced either by
NADPH or by light activation with NT, as shown by the
bleaching of the oxidized flavins at 457 and 487 nm (Figure 6,
bottom right panel). In the presence of calmodulin, the NADPH-
induced catalysis yielded the ferrous nitrosyl complex with the
typical absorption difference in the Soret band at 440 nm and
the broad visible band at ca. 575 nm. These maxima are similar
to those monitored after irradiation in the presence of NT,
peaking at 445 and 568 nm with minima at 390 and 486 nm.
The formation of a single broad visible band at 568 nm is typical
for a reduced heme (partly) nitrosylated, and the 486 nm
minimum accounts for the reduction of flavins (left panel). Both
the disappearance of the reduced NT and the oxidized heme
Soret band likely contributed to the minimum at 390 nm (upper
9
2184 J. AM. CHEM. SOC. VOL. 129, NO. 7, 2007

Selective Targeting of Nanotrigger to NADPH Sites
A R T I C L E S
Scheme 3. Decay Mechanism of the Protein-Bound NT*
NT* transient absorption decays to zero in three steps in the
absence of electron transfer. The first excited state NTCT* has
the shortest lifetime, τ1 ) 1.0-2.0 ps. Similar lifetimes of 2 to
3 ps have been reported for polyenes substituted with donor
and acceptor end groups and attributed to a charge transfer
state.29 In contrast, the p-coumaric acid chromophore of the
photoactive yellow protein readily isomerizes by twisting in 1
ps.32 In NT, τ1 is unaffected within experimental error by an
increase of the viscosity up to 60% glycerol, showing that this
first excited state is likely to be a (trans) charge transfer state.
a In the presence of eNOSred, A* decays in 7 ps to a complex NT-Phe*
or to B* in prereduced eNOSred and in glycerol. NT-Phe* deactivates to
the ground state via presently unidentified (radiative and nonradiative)
processes. B* returns to the ground state NT0 in prereduced eNOSred or
undergoes ET in oxidized eNOSred. Similar steps, except the conversion
of A* to B* were also found in F1160A.
The second lifetime A* is shortened to τ2 ) 7 ( 2 ps in
oxidized eNOSred compared to τ2 ) 20 ( 4 ps in prereduced
eNOSred and in F1160A (Table 1). Thus, this second step is
attributed to an excited-state complex (NT-Phe)* that contributed
to only a minor part 11% of the eNOSred decay and was absent
in F1160A. It probably reflected a minor Phe conformation
stacked on FAD, unproductive for ET and therefore not found
in prereduced eNOSred. The extended shape of NT required
for an effective charge delocalization probably locked the
NADPH site in an active conformation with most of the Phe1160
moved away from NT.
right panel). Thus, irradiation of NT bound to nNOS induced
heme reduction and subsequent formation of some heme-NO
complex, as attested by absorption spectroscopy.
Discussion
Innovative Design of the Nanotrigger: The nanotrigger was
designed according to a number of criteria listed below and
substantiated by the experimental data:
The third component is the most significantly shortened in
comparison with the control and represented the main decay
(45-60%) at 600 nm (Table 1). The ET rate was estimated
from the difference: (1/110 × 10-12) - (1/400 × 10-12) )
(7 ( 2) × 109 s-1 in eNOSred, based on τ3 ) 400 ( 100 ps for
NT in prereduced eNOSred and confirmed by the use of G6PD
(Table 1). Phe residues are usually found redox inert in contrast
to Trp or Tyr,33 consistent with electron transfer taking place
in F1160A but at a slower rate τ3 ) 190 ( 20 ps. This result
showed that F1160 was not involved in the electron transfer
and identified FAD itself as the electron acceptor. The fit of
the experimental transient spectrum involving the reduced FAD
confirmed this assertion (Figure 5 and see below).
(1) Selective Excitation. A selective excitation of the
nanotrigger is possible, due to a high extinction coefficient
combined with a maximal absorption shifted from that of
eNOSred (Figure 2).
(2) Specific Binding to the NADPH Site of the Protein.
The stopped-flow experiments support the specific binding of
the nanotrigger to the NADPH site: a direct competition
between the nanotrigger and NADPH on its natural binding site
was observed with an inhibition constant of Ki ) 7 ( 3 µM
(Figure 1). Our competition kinetic results are consistent with
a higher affinity of NT for the NADPH site than the natural
23
cofactor as judged by the comparison of Ki with KM as well
as the binding constant of K ) (8 ( 2) × 105 M-1 determined
The ET yield was estimated from the remaining absorption
at 600 ps using ∆ꢀ(600) ) 2.6 × 104 M-1 cm-1 relative to that
of MbCO as a reference (assuming a quantum yield (Φ) of CO
photodissociation of 1), Φ ) 0.19 ( 0.05 and Φ ) 0.09 (
0.03 in WT oxidized eNOSred and F1160A, respectively. This
difference in yield suggested that some charge recombination
likely took place within 600 ps in F1160A; this is consistent
with a structural role of Phe1160 in optimizing the conformation
of the NADPH site for ET.
for NT by fluorescence.
(3) Light-Induced Electron Transfer from the Nanotrigger
to the Protein Initiated the Enzymatic Activity. To calculate
the electron-transfer rate between NT* and FAD (F) from the
observed decay rate, the rate of natural decay kN of the
nanotrigger was determined in an environment resembling the
NADPH binding site. NT in prereduced eNOSred was used as
our first control. Experimental difficulties limited the extent of
flavin prereduction to more than 60%. This incomplete prer-
eduction could actually lead to an underestimation of kN.
Therefore, NT in G6PD was taken as a second control. This
protein carries an NADP site30 but lacks FAD as an electron
acceptor. This enzyme converts glucose 6P to 6-phosphoglu-
cono-lactone while reducing NADP+ to NADPH and definitely
is not a reductase. Moreover, inspection of its structure did not
reveal a nearby alternative aromatic residue that could function
as an electron trap. Evidence for NT binding to G6PD was given
by the blue shift of the absorption to 361 nm as compared to
376 nm for NT in buffer (Figure 2). Thus, G6PD-bound NT
served as an additional control for the determination of kN.
The second evidential support for electron transfer is the
proper fit of the experimental transient spectrum with the sum
of (semireduced - oxidized)flavin + (monocation + dication
radical of NT) (Figure 5); in this calculation, we assume that
the protonation of FAD does not occur in the time scale of our
experiment.29 The calculation showed that 25 to 40% of NT
yielded the dication, with concomitant delivery of two electrons.
The mechanism of this process will require further investigations
since the presence of the dication may result from alternative
pathways, either by two sequential (photoinduced) electron
transfers or by an intrinsic instability of the cation when
dissociated from the protein milieu to the solution.16 Thus, the
Evidence for the photoinduced electron transfer is supported
by three consistent data sets.
First, the main decay of the transient absorption is faster when
the nanotrigger is bound to oxidized eNOSred as compared to
that observed in the controls (Figures 3, 4 and Scheme 3).
(32) Lee, I. R.; Lee, W.; Zewail, A. H Proc Natl. Acad. Sci. U.S.A. 2006, 103,
258-62.
(33) Byrdin, M.; Eker, A. P. M.; Vos, M. H.; Brettel, K. Proc. Natl. Acad. Sci.
U.S.A. 2003, 100, 8676-81.
9
J. AM. CHEM. SOC. VOL. 129, NO. 7, 2007 2185

A R T I C L E S
Beaumont et al.
presence of the dication in addition to the monocation may arise
by its diffusion out from eNOSred and thus reflect the koff rate
of the NT.
slow activation and the slow rate-limiting binding step. Fur-
thermore, the rapid electron transfer mainly bypasses the slow
hydride transfer and thus is very useful to investigate the
subsequent electron-transfer events of NOSred or the full length
NOS. The photoinduced electron transfer is faster than that
reported for the rhenium-diimine wires, although the redox
partner is not the same in both cases: FAD in the present study,
compared to the heme in the latter case.5 The selective binding
to the NADPH site by NT should enable activation of other
proteins involved in bioreductive reactions as suggested by the
binding of NT to G6PD. Such a versatile tool is also expected
to be useful for time-resolved X-ray crystallography, presently
limited to the study of proteins with slow turnover.3 Work along
these lines is in progress in our laboratories.
Third, evidence of electron transfer is given by the shift of
NT* maximum to 670 nm when ET occurs as observed in
oxidized eNOSred or in F1160A in contrast to 620 nm as a
maximum observed in the controls lacking electron transfer
(Figures 3 and 5). Thus, the excited state in the controls
presented only a partial charge delocalization over the chro-
mophore chain. In contrast, the large shift of the NT* maximum
attested to a complete charge delocalization in the excited state
required for charge separation (ET) in oxidized eNOSred.
Concluding Remarks
In this work, the successful targeting of the designed
nanotrigger to the NADPH site of eNOSred and its subsequent
photoactivation by light are demonstrated. Such a compound
enables a time-resolved study of electron transfer within the
reductase domain of NOS, which contains a number of negative
regulating elements. Protein activation is selectively and rapidly
triggered at the natural site by light absorption of the nanotrigger;
this should allow detection of short-lived intermediate(s) that
would not be possible with NADPH or with diffusion-limited
compounds like caged nucleotides because of their (relatively)
Acknowledgment. The authors gratefully acknowledge Dr.
Marc Lebret for preliminary modeling studies, Dr. Yves
Meshulam for his help with the stopped-flow experiments, and
Dr. Viatcheslav Jouikov for electrochemical measurements.
Supporting Information Available: Experimental details:
Transient absorption spectra. This material is available free of
charge via the Internet at: http://pubs.acs.org.
JA067543E
9
2186 J. AM. CHEM. SOC. VOL. 129, NO. 7, 2007