Picosecond photoreduction of inducible nitric oxide synthase by rhenium(I)-diimine wires

Article abstract of DOI:10.1021/ja0543088

In a continuing effort to unravel mechanistic questions associated with metalloenzymes, we are developing methods for rapid delivery of electrons to deeply buried active sites. Herein, we report picosecond reduction of the heme active site of inducible ni

Full text of DOI:10.1021/ja0543088

Published on Web 10/19/2005
Picosecond Photoreduction of Inducible Nitric Oxide
Synthase by Rhenium(I)-Diimine Wires
Wendy Belliston-Bittner,† Alexander R. Dunn,†,‡ Yen Hoang Le Nguyen,†
Dennis J. Stuehr,§ Jay R. Winkler,† and Harry B. Gray*,†
Contribution from the Beckman Institute, California Institute of Technology,
Pasadena, California 91125, Department of Biochemistry, Stanford UniVersity,
Stanford, California 94305, and Department of Immunology, Lerner Research Institute,
The CleVeland Clinic, CleVeland, Ohio 44195
Received June 29, 2005; E-mail: hbgray@caltech.edu
Abstract: In a continuing effort to unravel mechanistic questions associated with metalloenzymes, we are
developing methods for rapid delivery of electrons to deeply buried active sites. Herein, we report picosecond
reduction of the heme active site of inducible nitric oxide synthase bound to a series of rhenium-diimine
electron-tunneling wires, [Re(CO)3LL′]+, where L is 4,7-dimethylphenanthroline and L′ is a perfluorinated
biphenyl bridge connecting a rhenium-ligated imidazole or aminopropylimidazole to a distal imidazole (F8bp-
im (1) and C3-F8bp-im (2)) or F (F9bp (3) and C3-F9bp (4)). All four wires bind tightly (Kd in the micromolar
to nanomolar range) to the tetrahydrobiopterin-free oxidase domain of inducible nitric oxide synthase
(iNOSoxy). The two fluorine-terminated wires displace water from the active site, and the two imidazole-
terminated wires ligate the heme iron. Upon 355-nm excitation of iNOSoxy conjugates with 1 and 2, the
active site Fe(III) is reduced to Fe(II) within 300 ps, almost 10 orders of magnitude faster than the naturally
occurring reduction.
Introduction
module, an intervening calmodulin-binding domain, and an
N-terminal oxidase domain.8,9 The reductase module of one
Nitric oxide synthase (NOS) is a P450-type iron-heme
enzyme that catalyzes the synthesis of nitric oxide (•NO) and
L-citrulline from L-arginine and O2.1 There are at least three
types of NOS expressed by mammals: neuronal NOS (nNOS),
endothelial NOS (eNOS), and inducible NOS (iNOS). iNOS is
expressed transiently in a variety of cells as a response to
immunostimulation.2 Owing mainly to the role •NO plays as a
signaling molecule in an astonishing number of bodily processes,
NOS malfunction is implicated in a wide array of diseases.3-7
Specifically, •NO overproduction by iNOS is involved directly
in the pathology of inflammation in diseases such as rheumatoid
arthritis, septic shock, atherosclerosis, and diabetes.2,7 These
findings have brought NOS and its catalytic mechanism into
the spotlight as targets of great pharmacological interest.
All three mammalian NOS enzymes exist in vivo as homo-
dimers, with each monomer consisting of a C-terminal reductase
subunit provides reducing equivalents to the catalytically active,
heme-containing oxidase domain of the other subunit.10 NOS
•
turns over twice en route to NO production. The mechanism
of the first turnover (from L-arginine to N-hydroxy-L-arginine)
is expected to be similar to that of P450, while the second
turnover is believed to proceed differently (Scheme 1).8,9,11-13
The first steps of each oxygenation are well-characterized and
include reduction of the heme followed by O2 binding and a
second reduction.12,14,15 Because of the sluggishness of the
second electron transfer (∼12 s-1),12,14,15 and the speed of
subsequent reaction(s), very little is known about the active
intermediates.13,16,17
One goal of our research program is to use photoactive wires
to trigger the second electron transfer to the NOS heme in an
(8) Groves, J. T.; Wang, C. C. Y. Curr. Opin. Chem. Biol. 2000, 4, 687-695.
(9) Alderton, W. K.; Cooper, C. E.; Knowles, R. G. Biochem. J. 2001, 357,
593-615.
† California Institute of Technology.
‡ Stanford University.
(10) Crane, B. R.; Rosenfeld, R. J.; Arvai, A. S.; Ghosh, D. K.; Ghosh, S.;
Tainer, J. A.; Stuehr, D. J.; Getzoff, E. D. EMBO J. 1999, 18, 6271-
6281.
§ The Cleveland Clinic.
(1) Stuehr, D. J. Biochim. Biophys. Acta-Bioenergetics 1999, 1411, 217-
230.
(11) Wei, C. C.; Crane, B. R.; Stuehr, D. J. Chem. ReV. 2003, 103, 2365-
(2) Aktan, F. Life Sci. 2004, 75, 639-653.
2383.
(3) Luth, H. J.; Holzer, M.; Gertz, H. J.; Arendt, T. Brain Res. 2000, 852,
45-55.
(12) Stuehr, D. J.; Santolini, J.; Wang, Z. Q.; Wei, C. C.; Adak, S. J. Biol.
Chem. 2004, 279, 36167-36170.
(4) Luth, H. J.; Holzer, M.; Gartner, U.; Staufenbiel, M.; Arendt, T. Brain
Res. 2001, 913, 57-67.
(13) Davydov, R.; Ledbetter-Rogers, A.; Martasek, P.; Larukhin, M.; Sono, M.;
Dawson, J. H.; Masters, B. S. S.; Hoffman, B. M. Biochemistry 2002, 41,
10375-10381.
(5) Kendrick, K. M.; Guevara-Guzman, R.; Zorrilla, J.; Hinton, M. R.; Broad,
K. D.; Mimmack, M.; Ohkura, S. Nature 1997, 388, 670-674.
(6) Huang, P. L.; Huang, Z. H.; Mashimo, H.; Bloch, K. D.; Moskowitz, M.
A.; Bevan, J. A.; Fishman, M. C. Nature 1995, 377, 239-242.
(7) Hingorani, A. D.; Cross, J.; Kharbanda, R. K.; Mullen, M. J.; Bhagat, K.;
Taylor, M.; Donald, A. E.; Palacios, M.; Griffin, G. E.; Deanfield, J. E.;
MacAllister, R. J.; Vallance, P. Circulation 2000, 102, 994-999.
(14) Wei, C. C.; Wang, Z. Q.; Hemann, C.; Hille, R.; Stuehr, D. J. J. Biol.
Chem. 2003, 278, 46668-46673.
(15) Wei, C. C.; Wang, Z. Q.; Arvai, A. S.; Hemann, C.; Hille, R.; Getzoff, E.
D.; Stuehr, D. J. Biochemistry 2003, 42, 1969-1977.
(16) Bec, N.; Gorren, A. C. F.; Voelker, C.; Mayer, B.; Lange, R. J. Biol. Chem.
1998, 273, 13502-13508.
9
10.1021/ja0543088 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 15907-15915
15907

A R T I C L E S
Belliston-Bittner et al.
Scheme 1. Proposed Mechanisms for the First and Second
Chart 1. Re Wires That Bind iNOSoxya
Turnovers of NOS9,12,13,18,a
a Known and hypothetical differences between the first and second
turnovers are noted in red. Putative intermediates of particular interest are
in the upper left ellipse.
attempt to observe high-valent intermediates directly. In recent
work, we took the first step toward this goal by building
ruthenium and rhenium wires (Re wires 1 and 3, Chart 1) that
bind tightly to the monomeric oxidase domain of iNOS
(iNOSoxy).19 These wires and analogues 2 and 4 (Chart 1) are
based on complexes that have been shown previously to bind
to and rapidly reduce P450cam upon photoexcitation.20-22
Strikingly, two of the bound wires, 1 and 2, photoreduce
iNOSoxy in less than 300 ps, nearly 10 orders of magnitude
faster than the natural reduction.23
Results
Synthesis and Characterization of Wires. Nucleophilic
substitution on decafluorobiphenyl proceeds readily (Scheme
2). Addition of two equivalents of nucleophile to decafluoro-
biphenyl will produce the doubly substituted product (such as
7 in Scheme 2) in approximately 100% yield. In our hands,
addition of fewer than two equivalents of nucleophile always
produces a mixture of singly and doubly substituted products.
These mixtures were easily separated by flash chromatography
or sublimation.
a Excitation of bound 1 or 2 triggers reduction of the enzyme within
300 ps.
Ligation of imidazole-based complexes 6, 7, and 8 to rhenium
is a slow and inefficient reaction. Attempts to speed up the
reaction by heating above 40 °C generally caused a decrease in
(17) In 1998, Bec et al.16 reported a low-temperature optical spectrum of a
complex they hypothesized to be compound I (ferryl heme). While this
paper has since been heavily cited, the lack of reference to this particular
spectrum and the lack of subsequent reproduction of this species calls their
assignment into question. Additionally, in 2002, Davydov et al.13 reported
EPR and ENDOR characterization of a peroxy heme intermediate at
cryogenic temperatures. These were the first and, to our knowledge, only
reported spectroscopic characterizations of post-oxyferrous heme-based
intermediates in NOS.
-
yield. In some samples, the BF4 counterion was exchanged
for trifluoromethanesulfonate (triflate, OTf-) due to observations
that the triflate salts were easier to purify. There were no
measurable differences in the solubility, protein binding, or
-
photophysical properties of the triflate and BF4 salts.
(18) Werner, E. R.; Gorren, A. C. F.; Heller, R.; Werner-Felmayer, G.; Mayer,
B. Exp. Biol. Med. 2003, 228, 1291-1302.
Typical of Re-diimines, all four complexes have very similar
spectroscopic behavior, with metal-to-ligand charge-transfer
absorption bands centered at ∼320 and 360 nm, emission
maxima at 560 nm, and microsecond luminescence lifetimes
(Table 1). All four wires have 4-5% luminescence quantum
yields in water, and 1 and 3 have Fo¨rster distances (R0) of 32
( 8 Å in the presence of ∆65 iNOSoxy.24
(19) Dunn, A. R.; Belliston-Bittner, W.; Winkler, J. R.; Getzoff, E. D.; Stuehr,
D. J.; Gray, H. B. J. Am. Chem. Soc. 2005, 127, 5169-5173.
(20) Wilker, J. J.; Dmochowski, I. J.; Dawson, J. H.; Winkler, J. R.; Gray, H.
B. Angew. Chem., Int. Ed. 1999, 38, 89-92.
(21) Dunn, A. R.; Dmochowski, I. J.; Winkler, J. R.; Gray, H. B. J. Am. Chem.
Soc. 2003, 125, 12450-12456.
(22) Dmochowski, I. J.; Crane, B. R.; Wilker, J. J.; Winkler, J. R.; Gray, H. B.
Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 12987-12990.
(23) Presta, A.; Siddhanta, U.; Wu, C. Q.; Sennequier, N.; Huang, L. X.; Abu-
Soud, H. M.; Erzurum, S.; Stuehr, D. J. Biochemistry 1998, 37, 298-310.
9
15908 J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005

Photoreduction of Inducible Nitric Oxide Synthase
A R T I C L E S
Scheme 2. Synthesis of Re-Diimine Wires
Table 1. Re Wire Characterization and Dissociation Constants in
Conjugates with ∆65 iNOSoxya
by the lack of a long-lived species). A dissociation constant for
the complex could not, therefore, be calculated from the transient
luminescence data. Instead, a dissociation constant of 130 (
70 nM was calculated from steady-state luminescence measure-
ments of 1 alone and in the 1:∆65 complex (Figure 3).19
Owing to low solubility and a tendency to aggregate in water,
the concentrations of wires 2 and 4 in buffer could not be
ꢀ360
(mM-1 cm-
τL
R0
K
d with
1
wire
)
(
µ
s)
φ
(Å)
∆65 ( M)
µ
1
4.1(4)
5.0(4)
4.1(6)
4.3(4)
0.97(3)
0.78(1)
0.95(4)
0.79(1)
0.055(2)
0.04(1)c
0.054(1)
0.042(6)c
32(8)
32(8)
0.13(7)
<10
1.4(5)
<10
2b
3
4b
(24) The 25% error in these calculations occurs due to the assumptions that κ2
is equal to 2/3. The danger in setting κ2 ) 2/3 is that the rhenium and
heme species are not likely to be freely rotating with respect to one another
in this bound system. This restricted motion is less of a problem for
calculating similar values for Ru(bpy)32+-based wires due to the existence
a Except where noted, numbers in parentheses are standard deviations
based on at least three separate measurements. b Measured in 50% glycerol.
c Average of two measurements.
2+
of three perpendicular transition dipoles for the octahedral Ru(bpy)3
Wire Binding to ∆65 iNOS. All four complexes described
above bind in the active-site channel of ∆65 iNOSoxy.
Imidazole-terminated wires 1 and 2 ligate the heme iron, causing
a characteristic red shift in the Soret band. Fluorophenyl-
terminated complexes 3 and 4 displace water from the active
site, causing a partial (blue) shift toward high-spin Fe(III)
(Figure 1).
species. The rhenium species being studied here, however, has a unidirec-
tional transition dipole in the plane of the phenanthroline ligand. This
suggests that assuming κ2 ) 2/3 could cause a large error in the calculated
R0 value. The largest error would occur if the rhenium transition dipole
was oriented exactly perpendicular to that of the heme. In this case, κ2
would be 0, and no FET would occur. Because quenching has been
observed, this extreme case is unlikely. Otherwise, because the sixth root
is taken to determine R0, κ2 values ranging from 1 to 4 produce up to a
25% error in the calculation. Crystal structures of the wire-NOS complexes
and anisotropy measurements will help to sort out this problem in the future.
Given the current limited understanding of the orientation of the rhenium
with respect to the heme in this system, the maximal 25% error was
assumed. Some error could also come from assuming that n ) 1.34 (the
refractive index of water). While studies were conducted in buffer solution,
the environment of the donor and acceptor is the protein channel. Because
the error due to the κ2 factor is likely to be much greater, however, the
error in n was not accounted for in these calculations.
The 3:∆65 conjugate has a dissociation constant of 1.4 (
0.5 µM, as determined by measurements of luminescence decay
kinetics (Figure 2A). These data are similar to those previously
described for the 3:∆114 iNOSoxy complex, suggesting Fo¨rster
energy-transfer quenching of the Re excited state and a Re-
heme distance of ∼18 Å.19
(25) Several attempts were made to measure the luminescence lifetimes of the
bound wires using the faster instrument on which the psTA measurements
were taken. These attempts were unsuccessful, however, due to the fact
that the protein itself is luminescent with a lifetime of approximately 200
ps when excited at 355 nm. The origin of this luminescence remains
unknown, but it dominates the decay kinetics in wire:protein mixtures to
the point that decay constants for the bound wires cannot be extracted.
They are, therefore, reported here as less than or equal to 200 ps.
Luminescence decay measurements on a 1:1 mixture of ∆65
and imidazole-terminated wire 1 (Figure 2B) show that Re(I)*
lifetime in the bound wire is much less than 10 ns.25 Further-
more, there is no apparent free wire in solution (as evidenced
9
J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005 15909

A R T I C L E S
Belliston-Bittner et al.
dominant phases, kb1 ) 6(1) × 106 s-1, kb2 ) 2.5(4) × 105 s-1
,
and kb3 ) 6(1) × 104 s-1 (Figure 4). Using ∆ꢀ445 ) 59.6 mM-1
cm-1 (see the Experimental Section), the calculated average
quantum yield of ferrous heme for both wires is 8 ( 2%. The
findings that the Fe(II) absorbance maximum does not shift from
445 nm up to 250 µs confirms that the wire does not dissociate
from the heme upon excitation.
Picosecond transient absorption measurements demonstrated
rapid formation of Fe(II) in the presence of wires 1 and 2. By
pumping with 70-ps, 355-nm pulses and probing with 442-nm
radiation from a continuous-wave He:Cd laser, we were able
to obtain transient absorbance traces that show the formation
of the ferrous heme at very short times (Figure 6 and Supporting
Information). The traces were fit to one exponential (eq 1) to
give kf ) 7(3) × 109 s-1 for formation of Fe(II). The lack of a
signal in the protein:4 trace confirms that the observed ab-
sorbance is attributable to Fe(II) and not to Re(I)* (Supporting
Information).
Discussion
Synthesis. Wires 1 and 3 were synthesized concomitantly.
Their design was based on those of Ru-bpy-based wires that
bind tightly to P450cam.20-22 In prior work, substituted per-
fluorobiphenyls were shown not only to be synthetically versatile
but also to be very efficient bridges for electron transfer. Further
investigation also has suggested that their hydrophobicity and
potential ability to π-stack facilitate binding to target proteins,
even when a bulky sensitizer prevents them from binding in an
active-site channel.19,21
The purpose of switching to a rhenium-based species was
two-fold: to take advantage of the smaller fac-tricarbonyl-
rhenium(I) phenanthryl sensitizer in comparison with the bulkier
Ru(bpy)32+ complex,19 and to increase the luminescence lifetime
(1 µs vs 200 ns) of the sensitizer in order to have a better chance
to quench reductively.
Luminescence Quenching and Electron Transfer. In the
case of 3, quenching of the bound wire is consistent with Fo¨rster
energy transfer from electronically excited rhenium to the heme
roughly 18 Å away. These data support a model in which the
wire is wedged far enough down the active-site channel for its
terminal fluorine to displace water from the sixth coordination
site on the iron (see Figure 7). This conclusion is further
supported by UV-visible absorption data that show the heme
spectrum shifting toward that of a five-coordinate, high-spin
structure (Figure 1A).
In the case of the other three wires, luminescence quenching
is too rapid (e200 ps)25 to be explained by Fo¨rster energy
transfer. Because of the rigidity of the perfluorobiphenyl moiety,
no sensible binding model can be created in which the terminal
imidazole or fluorine can come near the heme while allowing
the rhenium sensitizer to approach closely enough to be
quenched by energy transfer in less than 200 ps. Some other
mechanism surely must be operative. The production of Fe(II)
confirms that electron transfer must be involved in the lumi-
nescence quenching of 1 and 2.
Figure 1. UV-visible absorption spectra of ∆65:wire complexes. (A) ∆65
alone (5 µM; black) and bound to 1 equiv each of 1 (Re-Im-F8bp-Im; red)
and 3 (Re-Im-F9bp; blue). (B) ∆65 alone (5 µM; black) and bound to
approximately 1 equiv each of 2 (Re-Im-C3-F8bp-Im; red) and 4 (Re-Im-
C3-F9bp; blue).
measured precisely, so the micromolar dissociation constants
for both wires conjugated with ∆65 were estimated from UV-
vis spectra (Figure 1B) and luminescence decay kinetics (Figure
2C,D). Addition of approximately one equivalent of each wire
to micromolar solutions of ∆65 produced Soret band shifts
similar to those observed with the shorter, more soluble wires
1 and 3. Luminescence decay kinetics also indicated an
appreciable proportion of bound, quenched wire in each case.
In the protein:wire mixtures, the wires show an extremely fast
phase (below the instrument response time), corresponding to
bound wire, and two slower phases arising from free and
aggregated wire. The signal attributable to unbound wire was
similar to that observed for the protein:3 mixture, suggesting
that these wires likely bind about as tightly as 3.
Electron Transfer. Transient absorbance measurements show
that, upon 355 nm excitation, imidazole-terminated wires 1 and
2 reduce iNOS (Figure 4 and Supporting Information). Reduc-
tion is indicated by a bleach centered near 420 nm (the
disappearance of six-coordinate Fe(III)) and an optical density
increase centered around 445 nm (the appearance of six-
coordinate Fe(II)). Difference spectra were constructed from
single-wavelength transient absorbance traces 80 ns after
excitation of the 2:∆65 complex (Figure 5, blue) and 3 µs after
excitation of the 1:∆65 complex (Figure 5, red). These spectra
bear a striking resemblance to that previously reported for
photoreduction of N-phenylimidazole-ligated P450cam, another
heme-thiolate enzyme.21
In follow-up experiments, the rate of formation of Fe(II) after
excitation of the samples was measured. Interference by stray
fluorescence from the protein was eliminated by placing the
sample far from the detector and using a laser rather than a
flash lamp to probe the absorbance before and after excitation.
The resulting traces (Figure 6 and Supporting Information)
With both wire complexes, the photoproduced Fe(II) signal
shows multiexponential decay kinetics that consist of three
9
15910 J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005

Photoreduction of Inducible Nitric Oxide Synthase
A R T I C L E S
Figure 2. Transient luminescence decays of wires in solution and bound to ∆65 iNOSoxy. (A) 1 (blue) and a 1:1 mixture of 1 and ∆65 (11 µM; red). The
lack of an apparent slow decay indicates that almost all of the wire is bound to the enzyme. (B) 3 (blue) and a 1:1 mixture of 3 and ∆65 (7 µM; red) and
biexponential fit (kL ) 8(1) × 106 s-1, k0 ) 1.0(3) × 106 s-1, black). The fast component of the luminescence decay (kL) corresponds to 3 bound to ∆65.
(C) 2 (blue) and a 1:1 mixture of 2 and ∆65 (5 µM; red). (D) 4 (blue) and a 1:1 mixture of 4 and ∆65 (2 µM; red). In both C and D, the fastest component
of the luminescence decay, corresponding to 2 or 4 bound to ∆65, is unresolvable on this time scale. The slower, visible, biexponential decay matches the
biexponential decay of the wire alone in water solution (blue trace) and likely corresponds to a mixture of free and aggregated wire (fits are shown in
Supporting Information).
Figure 3. Steady-state fluorescence spectra of 1 (2 µM; blue) and a 1:1
Figure 4. Transient absorbance of a 1:1 mixture of 1 and ∆65 (11 µM)
mixture of 1 and ∆65 (2 µM; red) with λex ) 355 nm.
showing the prompt formation and initial decay of a six-coordinate, ferrous
heme. λex ) 355 nm, λobs ) 445 nm (red), λobs ) 420 nm (blue),
biexponential fit (kb2 ) 2.1(1) × 105 s-1, kb3 ) 6(1) × 104 s-1, black).
demonstrate ferrous heme formation with a time constant less
than 300 ps with wires 1 and 2 bound and additionally confirm
the absence of an Fe(II) signal with wire 4 bound. This
Inset shows the data at 445 nm taken on a shorter time scale (red) fit to
two exponentials (kb1 ) 6(1) × 106 s-1, kb2 ) 2.5(4) × 105 s-1, black).
Taken together, these traces show that the Fe(II) signal has a complicated
remarkably rapid photoreduction occurs almost 10 orders of
magnitude faster than the initial reduction of NOS by its
reductase module.12
Mechanistic Considerations. The observed picosecond
reduction of the NOS heme by 1 and 2 is well outside the time/
distance range established for single-step electron tunneling
through a protein medium.27 How can an electron make its way
from a photoexcited rhenium to a heme approximately 20 Å
distant with such incredible speed? Based on related work with
decay pathway that requires at least three exponentials to be fit adequately.
cytochrome P450cam,21 our original thought was that the rapid
luminescence quenching was due to direct electron transfer (ET)
from photoexcited rhenium to the heme. The initial nanosecond
transient absorbance experiments with fully conjugated wire 1,
showing a prompt Fe(II) signal (Figure 4), appeared to support
this interpretation.
(27) Gray, H. B.; Winkler, J. R. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 3534-
(26) Crane, B. R.; et al. Science 1998, 279, 2121-2126.
3539.
9
J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005 15911

A R T I C L E S
Belliston-Bittner et al.
Figure 5. Difference spectra of a 1:1 mixture of 2 and ∆65 (5 µM, 80 ns
after 355-nm excitation, blue squares) and a 1:1 mixture of 1 and ∆65 (11
µM, 3 µs after 355-nm excitation, red triangles) showing a bleach of a
six-coordinate Fe(III) Soret (420 nm) and the appearance of a six-coordinate
Fe(II) Soret (445 nm). Individual points were taken from single-wavelength
transient absorbance traces.
Figure 7. Model of wire 1 bound to ∆65 iNOSoxy. The wire is shown in
orange, the protein backbone in purple, and the heme cofactor in green.
The modeled Re-Fe distance is 18 Å. The model was created using Insight
II from a crystal structure of pterin-bound, fully dimeric NOS26 that likely
has a more closed channel than is present in pterin-free iNOSoxy. The
rigidity of the wire and our finding that the distal imidazole ligates the
heme suggest, however, that the rhenium-heme distance should be fairly
independent of the channel conformation. The other three wires are expected
to bind in similar fashion.
Table 2. Reduction Potentials
redox couple
potential (V vs NHE)
[Re(CO)3(phen)(im)]II/I
*
-0.7a
∼1.5a
-0.75c
-0.35d
∼ -0.3e
[Re(CO)3(phen)(im)]I*/0
TrpH•+/TrpH
1.15b
[Ru(tmbpy)3]III/II
*
low-spin iNOSoxy (FeIII/II
low-spin P450cam (FeIII/II
)
)
a Reference 28. b Reference 29. c Reference 30. d Reference 31. e Ref-
erences 32 and 33.
Figure 6. Transient absorbance at 442 nm of ∆65 iNOSoxy alone (8 µM,
blue) and in the presence of excess 1 (red) with λex ) 355 nm. The red
trace shows the rapid formation of ferrous heme fit to one exponential (kf
) 7(3) × 109 s-1, black) with the residual shown above.
as poly-xylyl is invoked, the observed ET reaction is still orders
of magnitude faster than expected.27
Electronically excited [Re(CO)3(phen)(im)]+ is a powerful
oxidant; indeed, Re(I)* can oxidize tryptophan (Table 2).34 In
the model of wire 1 bound to ∆65 iNOSoxy (Figure 7), two
tryptophan residues, Trp490 and Trp73, lie near the mouth of the
active-site channel and in close proximity to the modeled wire.
We propose that one of these tryptophan residues reductively
quenches (4,7-dmp)Re(I)* to generate TrpH•+ and “Re(0)”.
Because Trp73 is missing in ∆114 iNOSoxy, in which the same
rapid quenching of the wire 1 had earlier been observed,19 the
most likely candidate for reduction of Re(I)* is Trp490 (high-
lighted in blue in Figure 8).
Wires 2 and 4 were, therefore, designed to test this model.
The addition of a three-carbon linker between the sensitizer and
the heme should have slowed photoinduced ET, both by
increasing the Re-Fe distance and by disrupting the conjugation
in the bridging unit. Additionally, the extended Re-Fe distance
should have increased the luminescence lifetime of the Re
sensitizer in 4, the longer analogue of wire 3. Both new wires,
however, showed the same, dramatically fast luminescence
quenching that had been seen with wire 1. With wire 4, this
was true even though there was no Fe(II) signal observed.
Additionally, measurement of the rate of formation of Fe(II)
showed that heme reduction by wire 2 was as rapid as that
observed with wire 1 (Figure 6 and Supporting Information).
Finally, Fe(II) formation itself is simply too fast to be explained
by this model. The ReII/I* couple has approximately the same
reduction potential as the RuIII/II* couple in Ru(tmbpy)3, the
six-coordinate heme FeIII/II potentials for iNOSoxy and P450cam
are not very different (Table 2), and the bridging units between
sensitizer and heme for 1 and the Ru wires we reported21 are
virtually identical, leaving few possibilities to account for the
>100-fold faster heme reduction by the Re wires. Even if
activationless tunneling through a bridge as strongly coupled
On the basis of the structural model shown in Figure 8, we
propose the quenching and electron-transfer mechanisms out-
(28) Connick, W. B.; Di Bilio, A. J.; Hill, M. G.; Winkler, J. R.; Gray, H. B.
Inorg. Chim. Acta 1995, 240, 169-173.
(29) Harriman, A. J. Phys. Chem. 1987, 91, 6102-6104.
(30) Elliott, C. M.; Freitag, R. A.; Blaney, D. D. J. Am. Chem. Soc. 1985, 107,
4647-4655.
(31) Presta, A.; Weber-Main, A. M.; Stankovich, M. T.; Stuehr, D. J. J. Am.
Chem. Soc. 1998, 120, 9460-9465.
(32) Sligar, S. G.; Gunsalus, I. C. Proc. Natl. Acad. Sci. U.S.A. 1976, 73, 1078-
1082.
(33) Gunsalus, I. C.; Meeks, J. R.; Lipscomb, J. D. Ann. N.Y. Acad. Sci. 1973,
212, 107-121.
(34) Miller, J. E.; Grodinaru, C.; Di Bilio, A. J.; Wehbi, W. A.; Crane, B. R.;
Winkler, J. R.; Gray, H. B. J. Am. Chem. Soc. 2003, 125, 14220-14221.
9
15912 J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005

Photoreduction of Inducible Nitric Oxide Synthase
A R T I C L E S
that the ultrafast reduction involves electron hopping through
the perfluorobiphenyl bridge or possibly through a tyrosine near
the heme (see Figure 8). Simulations based on the Scheme 3
kinetics model that include reasonable estimates of driving forces
and reorganization energies indicate that two-step hopping could
account for the observed rate (Supporting Information).
Concluding Remarks
We have designed and constructed four Re(I)-diimine wires
that bind to the oxidase domain of inducible NOS. Upon
photoexcitation, two of these wires are capable of reducing the
enzyme almost 10 orders of magnitude faster than naturally
occurring electron transfer from the reductase module. The
demonstration of photoinduced, subnanosecond production of
Fe(II) represents an important step toward the unraveling of the
catalytic mechanism of NOS through wire-based technology.
Being able to trigger rapid heme reduction in the absence of
exogenous quenchers could lead to X-ray structural character-
ization of high-valent, catalytic intermediates in crystals of
enzyme:wire conjugates.
Figure 8. Cutaway view of the model in Figure 7, highlighting the putative
tryptophan quencher and three tyrosine residues that could be involved at
some stage of the redox chemistry.
Experimental Section
Syntheses. Dimethyl sulfoxide (DMSO) was purchased from EM
Science and dried over calcium hydride. Anhydrous, inhibitor-free
tetrahydrofuran (THF) was purchased from Aldrich and used without
further purification. All other solvents were either “OmniSolv” grade
reagents from EM Science or Burdick and Jackson High Purity solvents
and were used as received. Rhenium(I) pentacarbonyl chloride, silver
tetrafluoroborate, and silver trifluoromethane sulfonate were purchased
from Strem Chemicals. All other reagents were purchased from either
Aldrich or EM Sciences. All reagents were used as received. Preparative
chromatography on the wire complexes was performed using silica gel
(40 µm, 60 Å pore diameter) from JT Baker.
Electrospray ionization mass spectrometry (ESI/MS) was performed
on a Finnigan LCQ ion trap mass spectrometer. 1H and 19F NMR spectra
were recorded in CD2Cl2 on a Varian Mercury 300-MHz spectrometer.
The chemical shifts are reported relative to TMS for 1H NMR. 19F NMR
spectra were taken without a standard and were used solely for
verification of number of fluorine-containing species. Their quoted
chemical shifts are, therefore, qualitative.
(4,7-dmp)Re(CO)3Cl. Re(CO)5Cl (1.0 g, 2.8 mmol) and 4,7-
dimethyl-1,10-phenanthroline (4,7-dmp) (0.63 g, 3.0 mmol) were
suspended in 50 mL of toluene and stirred at 60 °C overnight. The
suspension was removed from heat, filtered, and washed with toluene
to obtain a bright yellow solid (1.436 g, 2.8 mmol, 100% yield) which
was used without further purification. 1H NMR, CDCl3 (ppm): 2.9 (s,
6H); 7.7 (d, 2H); 8.2 (s, 2H); 9.2 (d, 2H).
[(4,7-dmp)Re(CO)3(OH2)][BF4] (5). (4,7-dmp)Re(CO)3Cl (0.500 g,
0.98 mmol) was weighed into a sealed, 50-mL, three-neck round-bottom
flask and purged with argon for 30 min. Thirty milliliters of anhydrous
THF was added, the yellow suspension were stirred for 15 min, and
0.9 equiv of silver tetrafluoroborate (171 mg, 0.88 mmol) was
lined in Scheme 3. In the case of each wire bound to the protein,
excitation at 355 nm produces Re(I)*. Owing to its short length
or other binding constraints, excited wire 3 is quenched by
Fo¨rster energy transfer in approximately 30 ns. The other three
bound, photoexcited wires are quenched in less than 200 ps by
electron transfer that generates “Re(0)” and a tryptophan radical
cation. Recombination in this charge-separated state is likely
to occur rapidly. In the case of wire 4, in which no through-
bond pathway to the heme exists, it can be assumed that thermal
recombination is the dominant pathway back to the resting state.
In the cases of imidazole-terminated wires 1 and 2, however, a
direct bond to the heme allows for rapid electron transfer in
competition with recombination, thereby producing Fe(II) within
a few hundred picoseconds after excitation. The reduced heme
then reacts with oxidized amino acids by multiple pathways
over a period of a few hundred microseconds.
In the kinetics model shown in Scheme 3, it is expected that
kr > kET, and that kr therefore dominates the rate of formation
of Fe(II). Thus, an estimate of kET can be extracted from the
data using eq 8 (see the Experimental Section). Measurement
of the rate of formation (kf ) 7(3) × 109 s-1) and quantum
yield (8(2)%) of Fe(II) in the reaction allows us to estimate kET
≈ 6 × 108 s-1 for the 1:iNOSoxy conjugate. Even for
activationless ET in a conjugated donor-bridge-acceptor
system, this rate is too high to be accounted for by single-step
tunneling over a distance g18 Å.27 Our tentative proposal is
Scheme 3. Proposed Quenching and ET Processes in Photoexcited Re-Diimine:iNOSoxy Conjugatesa
a Re(I)* either oxidizes a nearby Trp residue (kQ) or decays back to the ground state through a combination of Fo¨rster transfer (ken) and intrinsic decay
(k0). The transient “Re(0)” species either reduces the heme (kET) or reacts with the Trp radical (kr). The charge-separated species produced by this route then
decays back to the ground state through multiple pathways (kb1, kb2, kb3).
9
J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005 15913

A R T I C L E S
Belliston-Bittner et al.
transferred into the flask under argon flow. The reaction was sealed
and left to stir for 24 h. The suspension was then filtered over Celite
and rinsed with dry THF to remove silver chloride and obtain a bright
yellow solution. The solution was rotary evaporated to dryness to obtain
a yellow solid (theoretical yield: 0.88 mmol) that was used without
further purification. 1H NMR, CD3CN (ppm): 3.0 (s, 6H); 7.9 (d, 2H);
8.4 (s, 2H); 9.2 (d, 2H). 19F NMR, CD3CN (ppm): -152.2 (4F).
Im-C3-F9bp (8). Decafluorobiphenyl (3 g, 9.22 mmol) and K2CO3
(2.3 g, 16.76 mmol) were stirred in anhydrous DMF (15 mL) under
argon. Aminopropylimidazole (1 mL, 8.38 mmol) was added via
syringe, and the mixture was stirred overnight at room temperature.
Precipitated KF was removed by vacuum filtration through Celite. Water
was added to the filtrate, and the product was extracted with dichloro-
methane (3 × 10 mL). Organic layers were combined, dried over
magnesium sulfate, and concentrated. The product was purified by
column chromatography on silica gel (CH2Cl2:MeOH ) 50:1) to obtain
Im-F9bp (6) and Im-F8bp-Im (7). Potassium carbonate (1.24 g, 9
mmol) was added to a vacuum-dried 100-mL Schlenk flask under argon
flow and dried under heat and vacuum for 1 h. The heat was removed,
and imidazole (0.61 g, 9 mmol) was added to the flask under argon
flow. The flask was then returned to vacuum for 30 min, and then
∼50 mL of dry DMSO were vacuum-distilled into the flask. Deca-
fluorobiphenyl (2.0 g, 6 mmol) was added to the suspension under argon
flow. The reaction was then sealed under argon and stirred for 12 h at
30 °C. The reaction was removed from heat and stirring, and 100 mL
of water were added. The resulting mixture was extracted with
dichloromethane (3 × 100 mL). The combined organic layer was
washed with a further 100 mL of water to remove traces of DMSO.
The organic layer was then dried over magnesium sulfate, filtered, and
rotary evaporated to a light yellowish oil. The oil was redissolved in
minimal dichloromethane and flash chromatographed over silica gel
(2:1 ethyl acetate:hexanes). Three organic fractions were collected and
rotary evaporated to dryness. The first fraction contained unreacted
decafluorobiphenyl. The second contained product 6 as a white solid
1
a light brown oil (2.2 g, 5.01 mmol, 59.7% yield). H NMR, CDCl3
(ppm): 2.2 (m, 2H); 3.5 (t, 2H); 4.3 (t, 2H); 4.7 (s, 1H); 7.0 (s, 1H);
7.2 (s, 1H); 8.65 (s, 1H). 19F NMR, CDCl3 (ppm): -161.8 (2F); -160.2
(2F); -152.4 (1F); -141.0 (2F); -138.2 (2F).
[(4,7-dmp)Re(CO)3(Im-C3-F9bp)][BF4] (4). 5 (1 g, 1.71 mmol) and
8 (0.75 g, 1.71 mmol) were stirred in a 2:1 mixture of CH2Cl2:THF (5
mL) at 40 °C for 5 days. The reaction mixture was cooled to room
temperature, filtered through Celite, and washed with CH2Cl2. The
filtrate was concentrated and flash chromatographed over silica gel
(CH2Cl2:MeOH ) 50:3). The product fraction was rotary evaporated
to dryness to obtain a bright yellow solid (900 mg, 0.89 mmol, 52.2%
yield). ESI/MS (m/z)+: 917.9 (calcd 918.1). 1H NMR, CD2Cl2 (ppm):
1.95 (m, 2H); 3.0 (s, 6H); 3.3 (t, 2H); 3.9 (t, 2H); 4.6 (s, 1H); 6.5 (s,
1H); 6.8 (s, 1H); 7.25 (s, 1H); 7.9 (d, 2H); 8.3 (s, 2H); 9.3 (d, 2H). 19
F
NMR, DMSO (ppm): -161.8 (2F); -160.8 (2F); -152.2 (1F); -148.6
(4F) (BF4); -142.4 (2F); -139.4 (2F).
1
(0.461 g, 1.2 mmol, 20% yield). H NMR, CD2Cl2 (ppm): 7.30 (s,
[(4,7-dmp)Re(CO)3(Im-C3-F8bp-Im)][BF4] (2). Imidazole (74.3 mg,
1.09 mmol) and K2CO3 (150 mg, 1.09 mmol) were stirred in anhydrous
DMF (5 mL). 4 (1 g, 0.99 mmol) was added, and the reaction was
stirred under argon overnight at room temperature. The reaction mixture
was then filtered through Celite and washed with CH2Cl2. The filtrate
was concentrated and flash chromatographed over silica gel (CH2Cl2:
MeOH ) 50:3). The product fraction was rotary evaporated to dryness
to obtain a bright yellow solid (700 mg, 0.66 mmol, 67% yield). ESI/
1H); 7.35 (s, 1H); 7.84 (s, 1H). 19F NMR, CD2Cl2 (ppm): -161.2 (2F);
-150.5 (1F); -148.2 (2F); -138.0 (2F); -137.3 (2F). The third fraction
contained product 7 as a white solid (1.243 g, 2.9 mmol, 48% yield).
1
ESI/MS (m/z)+: 431.3 (calcd 431.3). H NMR, CD2Cl2 (ppm): 7.30
(s, 2H); 7.35 (s, 2H); 7.90 (s, 2H). 19F NMR, CD2Cl2 (ppm): -148.0
(4F); -136.9 (4F).
[(4,7-dmp)Re(CO)3(Im-F9bp)][BF4] (3). Product 5 was dissolved
in a mixture of 12 mL of dichloromethane and 5 mL of THF and added
to 6 (0.342 g, 0.88 mmol). The reaction was stirred at 30 °C. Aliquots
were tested by ESI/MS at various time points, and the reaction was
removed from heating when the product peak stopped growing (48-
120 h). The reaction was then rotary evaporated to an orange oil. The
oil was redissolved in minimal dichloromethane and flash chromato-
graphed over silica gel under the following conditions: 3% methanol
in dichloromethane until the first two bands (one fluorescent orange
band containing various rhenium species and one colorless band
containing free 6) were collected. The methanol was then gradually
increased to 40%. The fluorescent yellow-green product was collected
and rotary evaporated to dryness to yield a yellow solid (0.122 g, 0.128
mmol, 14.5% yield). ESI/MS (m/z)+: 860.8 (calcd 861). 1H NMR,
CD2Cl2 (ppm): 3.0 (s, 6H); 6.8 (s, 1H); 7.1 (s, 1H); 7.9 (s, d, 3H); 8.3
(s, 2H); 9.35 (d, 2H). 19F NMR, CD2Cl2 (ppm): -161.1 (2F); -152.2
(4F) (BF4); -150.2 (1F); -148.0 (2F); -137.6 (2F); -136.4 (2F).
1
MS (m/z)+: 965.1 (calcd 966.1). H NMR, CD2Cl2 (ppm): 1.95 (m,
2H); 3.0 (s, 6H); 3.3 (t, 2H); 3.9 (t, 2H); 4.6 (s, 1H); 6.5 (s, 1H); 6.8
(s, 1H); 7.25 (s, 1H); 7.3 (s, 1H); 7.35 (s, 1H); 7.8 (s, 1H); 7.9 (d, 2H);
8.3 (s, 2H); 9.3 (d, 2H). 19F NMR, CD2Cl2 (ppm): -161.2 (2F); -152.2
(4F) (BF4); -149.4 (2F); -142.2 (2F); -137.8 (2F).
Sample Preparation. ∆65 iNOSoxy samples were prepared as
described previously.35 Small aliquots of the protein solutions were
exchanged into phosphate buffer (50 mM potassium phosphate, 50 mM
potassium chloride, pH 7.4) (“buffer”) using a desalting column
immediately before use. The measurement of the heme Soret maximum
at 422 nm confirmed the presence of low-spin, water-bound heme.
Monomeric, heme-containing protein concentration was determined
using the extinction coefficient ꢀ422 ) 75 mM-1 cm-1 19
.
Due to their low solubilities in water (∼20 µM for 1 and 3 and e1
µM for 2 and 4), concentrated (0.5-1.5 mM) wire solutions were
prepared in absolute ethanol. Small aliquots were then added to buffer/
protein samples such that the overall ethanol concentration never
exceeded 2%.36 Repeated experiments indicated a large degree of error
in measured concentration by this method. As the measured concentra-
tion was consistently lower than expected, the error is likely due to
the affinity of the hydrophobic wires for either glass syringes or plastic
pipet tips. As such, protein:wire samples for rigorous quantitation were
prepared by adding wire solutions to buffer, measuring the concentration
by UV-visible absorption, and then adding a known quantity of protein.
Samples of 2 and 4 for rigorous quantitative characterization were
prepared in 50% glycerol in buffer. Attempts were made to conduct
protein:wire studies in this type of mixture, but both glycerol and
ethylene glycol blue-shifted the heme Soret, indicating that they were
binding to the enzyme. These mixtures also interfered with wire binding,
[(4,7-dmp)Re(CO)3(Im-F8bp-Im)] [BF4] (1). 7 (0.379 g, 0.88 mmol)
was added to a 100-mL, three-neck round-bottom flask fitted with a
stir bar and condenser. 5 (0.44 mmol by theoretical yield, 0.5 equiv to
prevent bis-Re-substituted wire formation) was dissolved in a mixture
of dichloromethane and THF added to the flask. The reaction was stirred
at 30 °C until ESI/MS showed that the product peak had stopped
growing (2-5 days). The reaction was removed from heat and stirring,
and the product was then extracted into dichloromethane and washed
with water. The organic solution was dried over magnesium sulfate
and rotary evaporated to dryness. The product was redissolved in
dichloromethane and flash chromatographed under the same conditions
as 3. The clean product was rotary evaporated to dryness to obtain a
yellow solid (0.0807 g, 0.076 mmol, 17.3% yield based on Re
1
concentration). ESI/MS (m/z)+: 908.7 (calcd 909). H NMR, CD2Cl2
(35) Ghosh, D. K.; Wu, C. Q.; Pitters, E.; Moloney, M.; Werner, E. R.; Mayer,
B.; Stuehr, D. J. Biochemistry 1997, 36, 10609-10619.
(36) Testing with ∆65 indicated that micromolar protein solutions could tolerate
5% ethanol with no spectroscopic evidence of heme loss or preciptation.
(ppm): 3.0 (s, 6H); 6.8 (s, 1H); 7.1 (s, 1H); 7.25 (s, 1H); 7.35 (s, 1H);
7.8 (s, 1H); 7.9 (d, 2H); 7.95 (s, 1H); 8.3 (s, 2H); 9.4 (d, 2H). 19F
NMR, CD2Cl2 (ppm): -152.2 (4F) (BF4); -147.9 (4F); -136.3 (4F).
9
15914 J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005

Photoreduction of Inducible Nitric Oxide Synthase
A R T I C L E S
supporting the hypothesis that hydrophobic burial of the perfluorobi-
phenyl unit provides a good deal of binding energy to the wire:protein
conjugates.
visibly deteriorate by precipitating from solution. A compromise
between sample concentration, probe laser power, and total laser shots
was therefore made in order to maximize signal-to-noise ratio.
Dissociation Constants. In the case of the ∆65:3 complex, the
equilibrium dissociation constant was calculated from the biexponential
fit of the luminescence decay (eq 1, n ) 2) as described previously.21
Due to the rapidity of the luminescence decay of bound wire 1, the
∆65:1 dissociation constant was calculated as described in earlier work
from steady-state luminescence spectra.19 Because of their low solubili-
ties and tendency to aggregate in water, dissociation constants for 2
and 4 with ∆65 could not be reproducibly determined. In approximately
1:1 wire:protein mixtures at 1-10 µM concentrations, however,
observation of significant ∆65 heme Soret shifts and visible lumines-
cence quenching of the wires allowed for the estimation of micromolar
dissociation constants for both wires.
UV-visible absorption spectra were taken on an Agilent 8453 UV
spectrometer. Steady-state emission measurements were made in
buffer using a Flurolog model FL3-11 fluorometer equipped with a
Hamamatsu R928 photomultiplier tube. Emission quantum yields (φ)
were determined in deoxygenated solutions relative to tris(2,2′-
bipyridine)ruthenium(II) (φ ) 0.042 in water).37 All steady-state
emission and transient spectroscopic measurements were made in sealed
quartz cuvettes with deoxygenated samples (at least 30 cycles of partial
evacuation followed by backfilling with argon) unless otherwise stated.
Transient Spectroscopy. Nanosecond luminescence decay and
transient absorption measurements were made using a frequency-tripled
Nd:YAG laser (λ ) 355 nm) as the excitation source in an apparatus
that has been described elsewhere.38,39 The instrument has a response
limit of approximately 10 ns. Luminescence decay curves for the wires
and ∆65:wire complexes were fit in Igor Pro using a nonlinear least-
squares algorithm according to the following expression (eq 1):
Fo1rster Energy Transfer. Fo¨rster energy transfer (FET) parameters
were obtained from eq 4:
ken ) k0(R0/R)6
(4)
cn e-k t
(1)
n
I(t) ) c0 +
where k0 is the intrinsic luminescence decay rate of the donor (the
rhenium sensitizer), R is the distance between donor and acceptor (the
iNOS heme group), and R0, the Fo¨rster distance, is the distance at which
half of the luminescence intensity of the donor is quenched by FET.
R0 was calculated (in cm) from eq 5:
∑
n
where n ) 1-3 for mono-, bi-, and triexponential decays, respectively.
Transient absorbance data were converted from intensity to optical
density using eq 2:
R06 ) 8.8 × 10-25(κ2n-4φDJ)
(5)
OD ) -log(I/I0)
(2)
where the overlap integral J ) ∫FD(λ) ꢀA(λ) λ4 dλ, κ2 is a measure of
the spatial orientation of the transition dipoles of the donor and acceptor
with respect to each other (κ2 ) 2/3 in randomly oriented pairs), n is
the refractive index of the solvent (n ) 1.34 for water), and φD is the
luminescence quantum yield of the donor.24
Estimation of ET Rate Constants. The kinetics for luminescence
quenching and electron transfer in this system were interpreted
according to the model in Scheme 3. The observed rate constant (kL)
for the decay of Re(I)* in the presence of iNOSoxy is
where I is the intensity of transmitted light after excitation and I0 was
taken as the average light intensity over the 200 ns prior to the laser
shot. Decays of the Fe(II) signals observed upon photoreduction of
the enzyme were fit using eq 1. The quantum yield of Fe(II) was
calculated at t ) 20 ns using the following expression (eq 3):
[Fe(II)] ) ∆OD445/∆ꢀ445
(3)
where ∆ꢀ445 (six-coordinate Fe(III) - six-coordinate Fe(II)) was
estimated by subtracting the absorptivity measured for wire-bound ferric
∆65 iNOSoxy from the absorptivity reported for six-coordinate ferrous
iNOSoxy (105.4 mM-1 cm-1).40 For these measurements, an excess of
protein was included in the sample in order to minimize contributions
to the transient absorbance signal from free Re(I)*, and sufficient laser
power was used to ensure that 100% of the wire in the sample was
excited.
kL ) k0 + ken + kQ
(6)
The rate of Fe(II) formation depends on the concentration of “Re(0)”,
so the observed rate constant (kf) is given by
kf ) kr + kET
(7)
For picosecond transient absorption measurements, the samples were
excited at 10 Hz with 70-ps, 355-nm pulses from a regeneratively
amplified, mode-locked Nd:YAG laser. To eliminate interference from
fluorescence either of free wire or protein, the sample was placed far
from the detector and probed with 442-nm light from a continuous-
wave HeCd laser. Stray light from the room was filtered with a 442-
nm notch filter, and light intensity with (I, 10 000 counts) and without
(I0, 10 000 counts) excitation of the sample was collected directly by
a fiber optic (Fiberguide Industries) and detected with a Hamamatsu
C5680 streak camera. The raw image files were converted to xy data
using an in-house macro for Matlab. The raw data were then converted
to optical density using eq 2, and the rate of formation of ferrous heme
was fit to one exponential using eq 1.
Because kr describes a nonproductive reaction, kET can be extracted
from kf through the following expression:
ΦET ) kET/(kET + kr) ) kET/kf
(8)
where ΦET is the quantum yield of Fe(II) as a percentage of Re(I)* in
the wire-bound sample assuming kQ . k0 + ken.
Acknowledgment. This research was supported by NIH
(DK19038; GM070868), the Parsons Foundation (W.B.-B.), the
Fannie and John Hertz Foundation (A.R.D.), the Ellison Medical
Foundation (Senior Scholar Award in Aging to H.B.G.), and
the Arnold and Mabel Beckman Foundation.
Because the resting (ferric) heme already has appreciable absorbance
at 442 nm, the initial probe light intensity, and therefore the overall
signal strength, was attenuated. Additionally, the protein would not
tolerate more than about 20 000 pump/probe shots before it began to
Supporting Information Available: Semilog plots with fits
of the luminescence decay data for the 2:∆65 and 4:∆65
complexes; single-wavelength transient absorbance data for the
1:∆65 complex; kinetics simulation for electron hopping. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(37) Van Houten, J.; Watts, R. J. J. Am. Chem. Soc. 1976, 98, 4853-4858.
(38) Low, D. W.; Winkler, J. R.; Gray, H. B. J. Am. Chem. Soc. 1996, 118,
117-120.
(39) Dmochowski, I. J.; Winkler, J. R.; Gray, H. B. J. Inorg. Biochem. 2000,
81, 221-228.
(40) Hurshman, A. R.; Marletta, M. A. Biochemistry 2002, 41, 3439-3456.
JA0543088
9
J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005 15915