pH rate profiles of FnY356-R2s (n = 2, 3, 4) in Escherichia coli ribonucleotide reductase: Evidence that Y356 is a redox-active amino acid along the radical propagation pathway

Article abstract of DOI:10.1021/ja055927j

The Escherichia coli ribonucleotide reductase (RNR), composed of two subunits (R1 and R2), catalyzes the conversion of nucleotides to deoxynucleotides. Substrate reduction requires that a tyrosyl radical (Y ₁₂₂?) in R2 generate a transient cysteinyl radical (C ₄₃₉?) in R1 through a pathway thought to involve amino acid radical intermediates [Y₁₂₂? → W₄₈ → Y ₃₅₆ within R2 to Y₇₃₁ → Y₇₃₀ → C ₄₃₉ within R1]. To study this radical propagation process, we have synthesized R2 semisynthetically using intein technology and replaced Y ₃₅₆ with a variety of fluorinated tyrosine analogues (2,3-F ₂Y, 3,5-F₂Y, 2,3,5-F₃Y, 2,3,6-F₃Y, and F₄Y) that have been described and characterized in the accompanying paper. These fluorinated tyrosine derivatives have potentials that vary from -50 to +270 mV relative to tyrosine over the accessible pH range for RNR and pK_as that range from 5.6 to 7.8. The pH rate profiles of deoxynucleotide production by these F_nY₃₅₆-R2s are reported. The results suggest that the rate-determining step can be changed from a physical step to the radical propagation step by altering the reduction potential of Y₃₅₆? using these analogues. As the difference in potential of the F_nY? relative to Y? becomes >80 mV, the activity of RNR becomes inhibited, and by 200 mV, RNR activity is no longer detectable. These studies support the model that Y₃₅₆ is a redox-active amino acid on the radical-propagation pathway. On the basis of our previous studies with 3-NO₂Y₃₅₆-R2, we assume that 2,3,5-F₃Y₃₅₆,2,3,6-F₃Y₃₅₆, and F₄Y₃₅₆-R2s are all deprotonated at pH > 7.5. We show that they all efficiently initiate nucleotide reduction. If this assumption is correct, then a hydrogen-bonding pathway between W₄₈ and Y ₃₅₆ of R2 and Y₇₃₁ of R1 does not play a central role in triggering radical initiation nor is hydrogen-atom transfer between these residues obligatory for radical propagation.

Full text of DOI:10.1021/ja055927j

Published on Web 01/11/2006
pH Rate Profiles of FnY356-R2s (n ) 2, 3, 4) in Escherichia
coli Ribonucleotide Reductase: Evidence that Y356 Is a
Redox-Active Amino Acid along the Radical Propagation
Pathway
Mohammad R. Seyedsayamdost, Cyril S. Yee, Steven Y. Reece,
Daniel G. Nocera,* and JoAnne Stubbe*
Contribution from the Department of Chemistry, Massachusetts Institute of Technology,
77 Massachusetts AVenue, Cambridge, Massachusetts 02139-4307
Received August 28, 2005; E-mail: nocera@mit.edu; stubbe@mit.edu
Abstract: The Escherichia coli ribonucleotide reductase (RNR), composed of two subunits (R1 and R2),
catalyzes the conversion of nucleotides to deoxynucleotides. Substrate reduction requires that a tyrosyl
radical (Y122•) in R2 generate a transient cysteinyl radical (C439•) in R1 through a pathway thought to involve
amino acid radical intermediates [Y122• f W48 f Y356 within R2 to Y731 f Y730 f C439 within R1]. To study
this radical propagation process, we have synthesized R2 semisynthetically using intein technology and
replaced Y356 with a variety of fluorinated tyrosine analogues (2,3-F2Y, 3,5-F2Y, 2,3,5-F3Y, 2,3,6-F3Y, and
F4Y) that have been described and characterized in the accompanying paper. These fluorinated tyrosine
derivatives have potentials that vary from -50 to +270 mV relative to tyrosine over the accessible pH
range for RNR and pKas that range from 5.6 to 7.8. The pH rate profiles of deoxynucleotide production by
these FnY356-R2s are reported. The results suggest that the rate-determining step can be changed from
a physical step to the radical propagation step by altering the reduction potential of Y356• using these
analogues. As the difference in potential of the FnY• relative to Y• becomes >80 mV, the activity of RNR
becomes inhibited, and by 200 mV, RNR activity is no longer detectable. These studies support the model
that Y356 is a redox-active amino acid on the radical-propagation pathway. On the basis of our previous
studies with 3-NO2Y356-R2, we assume that 2,3,5-F3Y356, 2,3,6-F3Y356, and F4Y356-R2s are all deprotonated
at pH > 7.5. We show that they all efficiently initiate nucleotide reduction. If this assumption is correct,
then a hydrogen-bonding pathway between W48 and Y356 of R2 and Y731 of R1 does not play a central role
in triggering radical initiation nor is hydrogen-atom transfer between these residues obligatory for radical
propagation.
Introduction
two proteins has been generated using these structures on the
basis of their shape complementarity and on knowledge of
Class I Escherichia coli ribonucleotide reductase (RNR) plays
a crucial role in DNA replication and repair by catalyzing the
reduction of nucleoside diphosphates (NDPs) to deoxynucleoside
diphosphates (dNDPs).1,2 It is composed of two homodimeric
subunits, designated R1 and R2. A complex between R1 and
R2 is required for activity.3 R1 houses the NDP binding sites
and the binding sites for the effectors that control the specificity
and rate of nucleotide reduction.3-5 R2 harbors the diferric
tyrosyl radical (Y122•) cofactor proposed to initiate nucleotide
reduction by generating a transient thiyl radical (C439•) in the
active site of R1.6 The crystal structures of both R1 and R2
have been solved independently.7-9 A docking model of the
conserved residues.7 This model, a 1:1 complex of the R1 and
R2 homodimers, places the Y122• on R2 at a distance >35 Å
away from the terminal site of oxidation on R1, the C439 residue.
Electron tunneling between these residues according to Marcus
theory (kET ) 10-6 s-1 for â ) 1.2 Å-1 under activationless
conditions) is too slow to account for a kcat of ∼2-10 s-1 10
.
Thus, the radical generation process has been proposed to occur
via a hopping mechanism involving aromatic amino acid radical
intermediates, as shown in Figure 1.7,11 Mounting evidence is
calling into question the symmetry of the docking model.12,13
However, a strong case can be made for involvement of the
residues on this pathway, specifically Y356 on R2 and Y731 and
Y730 on R1. They are the only absolutely conserved residues
(1) Jordan, A.; Reichard, P. Annu. ReV. Biochem. 1998, 67, 71.
(2) Stubbe, J.; van der Donk, W. A. Chem. ReV. 1998, 98, 705.
(3) Thelander, L. J. Biol. Chem. 1973, 248, 4591.
(4) Kashlan, O. B.; Scott, C. P.; Lear, J, D.; Cooperman, B. S. Biochemistry
2002, 41, 462.
(9) Ho¨gbom, M.; Galander, M.; Andersson, M.; Kolberg, M.; Hofbauer, W.;
Lassmann, G.; Nordlund, P.; Lendzian, F. Proc. Natl. Acad. Sci. U.S.A.
2003, 100, 3209.
(10) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265.
(11) Stubbe, J.; Nocera, D. G.; Yee, C. S.; Chang, M. C. Y. Chem. ReV. 2003,
103, 2167.
(12) Erickson, H. K. Biochemistry 2001, 40, 9631.
(13) Ge, J.; Yu, G.; Ator, M. A.; Stubbe, J. Biochemistry 2003, 42, 10071.
(5) Kashlan, O. B.; Cooperman, B. S. Biochemistry 2003, 42, 1696.
(6) Stubbe, J.; Riggs-Gelasco, P. Trends Biochem. Sci. 1998, 23, 438.
(7) Uhlin, U.; Eklund, H. Nature 1994, 370, 533.
(8) Nordlund, P.; Sjo¨berg, B.-M.; Eklund, H. Nature 1990, 345, 593.
9
1562
J. AM. CHEM. SOC. 2006, 128, 1562-1568
10.1021/ja055927j CCC: $33.50 © 2006 American Chemical Society

pH Rate Profiles of FnY356−R2s in E. coli RNR
A R T I C L E S
Table 1. Fluorotyrosine Derivatives Used To Substitute Y356 of R2
and Their Physical Properties
fluorotyrosine
pKa
Ep(Y•
/Y-)/ mV
Ac-Y-NH2
9.9
7.2
7.8
6.4
7.0
5.6
642
755
810
853
911
968
Ac-3,5-F2Y-NH2
Ac-2,3-F2Y-NH2
Ac-2,3,5-F3Y-NH2
Ac-2,3,6-F3Y-NH2
Ac-F4Y-NH2
Figure 1. Putative PCET pathway of E. coli RNR based on the docking
model in ref 7. Y356 is not visible in the crystal structure of R2; it is proposed
to propagate the radical between W48 on R2 and Y731 on R1. Distances on
the R2 side are taken from the crystal structure of oxidized R2 at 1.4 Å
resolution, given in ref 9.
acid, Y356PheNH2-R2 construct,23 suggested that the rate of
radical transfer through position 356 is affected mainly by the
reduction potential of the residue at that position and not
controlled by hydrogen bonding or proton transfer. To assess
the interplay between the electron- and proton-transfer events
requires the systematic variation of the radical-reduction
potential and phenolic pKa at Y356, the proposed gate keeper
for radical transport between the R1 and R2 subunits. In the
accompanying paper,24 we reported the synthesis and charac-
terization of a number of unnatural N-acetyl and C-amide
protected fluorotyrosine analogues, Ac-FnY-NH2, (Table 1).
These derivatives have radical peak-reduction potentials that
vary from -50 to +270 mV relative to Ac-Y•-NH2 in the
pH region in which RNR activity can be measured and pKas
that range from 5.6 to 7.8. We suggested that these analogues
would be useful in probing the mechanism(s) of proton coupled
electron transfer (PCET) in systems that involve redox-active
tyrosines.
We now report the semisynthesis of R2, in which Y356 has
been replaced with fluorinated tyrosine derivatives of Table 1.
These R2 analogues have allowed us to study the role of Y356
in the radical-propagation process of E. coli RNR. The results
suggest that an increase in the peak-reduction potential of the
FnY analogue relative to Y by 80 mV results in a change in the
rate-limiting step from a conformational change to the radical
propagation process. These studies support the proposal that
the protonation state of the phenol of this residue is not important
in conformational gating, that the proton can be lost from this
pathway without affecting the overall enzymatic activity, and
that Y356 is a redox-active amino acid on the radical propagation
pathway. Additionally, the rigorous assessment of the enzymatic
activity afforded by the FnY356-R2 series establishes that the
energetics of radical hopping through Y356 is finely tuned in
wild-type (wt) class I RNR and can only operate within a ∼150
mV window for the Y•/Y redox couple.
that are not involved in assembly of the di-iron cofactor or in
the nucleotide-reduction process.13,14
Efforts to examine this pathway model experimentally and
identify the mechanism(s) for radical propagation have here-
tofore relied on site-directed mutagenesis of each amino acid
within the pathway on both the E. coli and mouse RNRs.15-18
More recently, in vivo complementation studies in E. coli using
pathway mutants to investigate radical migration have been
reported.19 These studies demonstrate the requirement for these
residues (Figure 1) in nucleotide reduction. However, they are
not amenable to addressing the mechanism of the radical
propagation process or the intermediacy of amino acid radicals
as the proteins are “inactive”. Investigation of the radical
propagation process is further complicated by results from
steady-state and pre-steady-state kinetic analyses of E. coli RNR,
which suggest that radical propagation and nucleotide reduction
are preceded by a slow physical step and that binding of
substrate and effector trigger this physical step.13 This confor-
mational gating makes the electron transfer (ET) step kinetically
invisible and, thus, impossible to study. One way to gain insight
into the propagation reaction, therefore, is to change the rate-
limiting step from this physical step to ET by altering the
reduction potential of a residue within the pathway.
The limited repertoire of natural amino acids does not allow
for an informative perturbation to be made and thus we have
recently turned to site-specific insertions of unnatural amino
acids into R2 at Y356 using protein-ligation methods.20
A
3-NO2Y356-R2 construct has allowed us to measure the pKa of
a single amino acid residue in a putative 260 kDa R1/R2
complex.21 A 2,3-F2Y356-R2 provided the first clues that the
phenolic proton at 356 in R2 is not required for conformational
gating of turnover.22 Results from studying the aniline amino
(14) Eklund, H.; Uhlin, U.; Farnegardh, M.; Logan, D. T.; Nordlund, P. Prog.
Biophys. Mol. Biol. 2001, 77, 177.
(15) Ekberg, M.; Sahlin, M.; Eriksson, M.; Sjo¨berg, B. M. J. Biol. Chem. 1996,
271, 20655.
(16) Ekberg, M.; Potsch, S.; Sandin, E.; Thunnissen, M.; Nordlund, P.; Sahlin,
M.; Sjo¨berg, B. M. J. Biol. Chem. 1998, 273, 21003.
(17) Rova, U.; Goodtzova, K.; Ingemarson, R.; Behravan, G.; Gra¨slund, A.;
Thelander, L. Biochemistry 1995, 34, 4267.
Materials and Methods
Materials. ATP, cytidine-5’-diphosphate (CDP), reduced â-nicoti-
namide adenine dinucleotide phosphate (NADPH), N-hydroxyurea,
trifluoroacetic acid (TFA), 2-mercaptoethanesulfonic acid (MESNA),
(18) Rova, U.; Adrait, A.; Potsch, S.; Gra¨slund, A.; Thelander, L. J. Biol. Chem.
1999, 274, 23746.
(19) Ekberg, M.; Birgander, P.; Sjo¨berg, B.-M. J. Bacteriol. 2003, 185, 1167.
(20) Muir, T. W. Annu. ReV. Biochem. 2003, 72, 249.
(21) Yee, C. S.; Seyedsayamdost, M. R.; Chang, M. C.; Nocera, D. G.; Stubbe,
J. Biochemistry 2003, 42, 14541.
(22) Yee, C. S.; Chang, M. C.; Ge, J.; Nocera, D. G.; Stubbe, J. J. Am. Chem.
Soc. 2003, 125, 10506.
(23) Chang, M. C. Y.; Yee, C. S.; Nocera, D. G.; Stubbe, J. J. Am. Chem. Soc.
2004, 126, 16702.
(24) Seyedsayamdost, M. R.; Reece, S. Y.; Nocera, D. G.; Stubbe, J. A. J. Am.
Chem. Soc. See accompanying paper.
9
J. AM. CHEM. SOC. VOL. 128, NO. 5, 2006 1563

A R T I C L E S
Seyedsayamdost et al.
Table 2. RP-HPLC and MALDI-TOF MS Characterization of
Fmoc-succinimide, Sephadex G-25 resin, DNase I (10 units/µL), and
Triton X-100 were purchased from Sigma. BL21 (DE3) RIL Codon+
competent cells were obtained from Stratagene. Calf-intestine alkaline
phosphatase (20 units/µL) was purchased from Roche. All amino acid
derivatives were purchased from Novabiochem, Fmoc-L-Leu-PEG-PS
resin and all other chemicals required for peptide synthesis were
obtained from Applied Biosystems. [2-14C]-CDP was purchased from
Moravek Biochemicals. E. coli thioredoxin (TR, SA of 40 units/mg)
and E. coli thioredoxin reductase (TRR, SA of 1800 units/mg) were
isolated as previously described.13
FnY-22mersa
MS of (tbuthio)-
MS of
RP-HPLC
Rt (min)
protected FnY-22mer
deprotected FnY-22mer
m/z [M −
H]- calcd (obs)
FnY-22mer
m/z [M
−
H]- calcd (obs)
Y-22mer
19
19
18.5
19
21.5
18
2548.6 (2548.2)
2584.6 (2585.2)
2584.6 (2584.0)
2602.6 (2601.9)
2602.6 (2601.5)
2642.6 (2641.4)b
2460.6 (2460.1)
2496.6 (2495.6)
2496.6 (2495.8)
2552.6 (2551.6)c
2552.6 (2551.6)c
2570.6 (2570.3)c
2,3-F2Y-22mer
3,5-F2Y-22mer
2,3,6-F3Y-22mer
2,3,5-F3Y-22mer
F4Y-22mer
Physical Measurements. 1H NMR spectra were recorded on a
Varian 300 MHz NMR spectrometer at the MIT Department of
Chemistry Instrumentation Facility. NMR samples were internally
referenced to tetramethylsilane. pH measurements were performed with
an Orion microelectrode. Absorption spectra were recorded on an
Agilent 8453 Diode Array spectrophotometer. EPR spectra were
recorded at 77 K on a Bruker ESP-300 X-band (9.4 GHz) spectrometer
equipped with an Oxford liquid helium cryostat.
a
See ref 21 for detailed description of HPLC and MALDI-TOF methods.
[M - 2H + Na]-. [M - 2H + K]-.
b
c
Table 3. Physical and Biochemical Characterization of
FnY356-R2s
ESI-MS
radical
contentb specific activity
(Y
/dimer) (nmol/min mg)c
maximal
yield
m/z [M
+
H]+
K
m for R1
M)
FnY356−R2
(mg)a
calcd (obs)
•
(µ
Synthesis of Fmoc-FnYs. Fmoc-FnYs were synthesized by the
method of Lapatsanis et al. with minor modifications.25 The tyrosine
analogue (700 µmol) was dissolved in 2.4 mL of 10% Na2CO3 and
mixed with Fmoc-succinimide (970 µmol) in 2.5 mL of dioxane at 4
°C. The reaction was warmed to room temperature and reacted for 20
min to 2 h (FnYs with higher fluorine substitution required longer
reaction times). The reaction was monitored via TLC using 10:1 CHCl3/
MeOH as the mobile phase. After completion, the reaction was
quenched with 25 mL of water and extracted twice with 10 mL of
EtOAc to remove Fmoc-succinimide. The pH was then lowered to 2-3
with 2 N HCl, and the solution was extracted with EtOAc (10×, 10
mL). The organic layer was washed with saturated NaCl (3×, 3 mL)
and water (2×, 3 mL) and finally dried over MgSO4. The solvent was
removed in vacuo, and the product was purified by silica gel
chromatography (16 g, 1.5 × 28 cm) using isocratic elution. The solvent
system, Rf, and the yield for each analogue are summarized in Table
S1 (Supporting Information).
Fmoc-L-3,5-F2Y: 1H NMR (300 MHz, CD3OD) δ ) 2.85 (dd, 1H,
Câ-H1, 9.7 Hz, 14.3 Hz), 3.12 (dd, 1H, Câ-H2, 5.7 Hz, 14.3 Hz), 4.2
(m, 3H, fluorenyl C-H and C-H2), 4.37 (dd, 1H, CR-H, 5.7 Hz, 10
Hz), 6.82 (m, 2H, PhOH C-H), 7.35 (dt, 4H, aromatic fluorenyl C-H,
7.6 Hz, 26.7 Hz), 7.62 (d, 2H, aromatic fluorenyl C-H, 7.5 Hz), 7.8
(d, 2H, aromatic fluorenyl C-H, 7.1 Hz).
Fmoc-L-2,3-F2Y: 1H NMR (300 MHz, CD3OD) δ ) 2.9 (dd, 1H,
Câ-H1, 9.6 Hz, 14.6 Hz), 3.23 (dd, 1H, Câ-H2, 4.8 Hz, 14.2 Hz),
4.24 (m, 3H, fluorenyl C-H and C-H2), 4.4 (dd, 1H, CR-H, 5 Hz,
9.7 Hz), 6.61 (m, 1H, PhOH C-H5), 6.82 (m, 1H, PhOH C-H6), 7.28
(m, 2H, aromatic fluorenyl C-H), 7.38 (t, 2H, aromatic fluorenyl C-H,
7.4 Hz), 7.6 (d, 2H, aromatic fluorenyl C-H, 7.4 Hz), 7.77 (d, 2H,
aromatic fluorenyl C-H, 7.4 Hz).
Fmoc-L-2,3,6-F3Y: 1H NMR (300 MHz, CD3OD) δ ) 3.02 (dd,
1H, Câ-H1, 9.3 Hz, 14 Hz), 3.2 (dd, 1H, Câ-H2, 5.7 Hz, 13.9 Hz),
4.2 (m, 3H, fluorenyl C-H and C-H2), 4.41 (dd, 1H, CR-H, 5.7 Hz,
9.4 Hz), 6.46 (m, 1H, PhOH C-H), 7.33 (dt, 4H, aromatic fluorenyl
C-H, 7.1 Hz, 26.7 Hz), 7.6 (d, 2H, aromatic fluorenyl C-H, 7 Hz),
7.78 (d, 2H, aromatic fluorenyl C-H, 9 Hz).
Fmoc-L-2,3,5-F3Y: 1H NMR (300 MHz, CD3OD) δ ) 2.9 (dd, 1H,
Câ-H1, 9.6 Hz, 13.5 Hz), 3.25 (dd, 1H, Câ-H2, 4.2 Hz, 13.8 Hz),
4.25 (m, 3H, fluorenyl C-H and C-H2), 4.43 (dd, 1H, CR-H, 4.7 Hz,
9.2 Hz), 6.82 (m, 1H, PhOH C-H), 7.32 (dt, 4H, aromatic fluorenyl
C-H, 7.2 Hz, 26.7 Hz), 7.58 (d, 2H, aromatic fluorenyl C-H, 7.2
Hz), 7.75 (d, 2H, aromatic fluorenyl C-H, 7.2 Hz).
Y-R2
2,3-F2Y356-R2
3,5-F2Y356-R2
14.3 43360 (43360)
15 43396 (43399)
20 43396 (43392)
0.34
450
370
450
365
95
0.55 ( 0.18
0.62 ( 0.11
ndd
0.33
0.31
0.42
0.41
0.42
1.1
2,3,5-F3Y356-R2 13 43414 (43410)
0.65 ( 0.15
2,3,6-F3Y356-R2 10.5 43414 (43413)
ndd
F4Y356-R2
10 43432 (43434)
30
1420
0.59 ( 0.2
0.55 ( 0.21
V
353G/ S354C-R2e
-
-
a
Amount recovered from incubation with 45 mg of MESNA-activated
b
R2. Measured by the dropline correction method and quantitative EPR
c
methods using recombinant wt R2 as standard. Activity of intein-generated
FnY356-R2s is normalized for radical content of Y-R2. nd ) not
determined. Made by recombinant methods.
d
e
9.3 Hz), 7.34 (dt, 4H, aromatic fluorenyl-H, 7.7 Hz, 27 Hz), 7.61 (d,
2H, aromatic fluorenyl-H, 7.2 Hz), 7.78 (d, 2H, aromatic fluorenyl-
H, 7.5 Hz).
Peptide Synthesis. The R2 C-terminal peptide CSFnYLVGQID-
SEVDTDDLSNFQL was synthesized by a combination of standard
solid-phase and solution-phase peptide synthesis methods as previously
described.21 The first 19 residues were added using a Pioneer Peptide
Synthesizer from Applied Biosystems. The remaining 3 amino acids
(positions 20-22 on the peptide) were coupled manually. The manual
coupling reaction of Fmoc-FnYs were carried out for 1 h, and a typical
reaction contained 4 equiv of Fmoc-FnY, 3.6 equiv of O-(7-azabenzo-
triazole-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HATU),
and 8 equiv of diisopropyl ethylamine (DIPEA) in DMF. The purified
peptides were characterized by RP-HPLC and MALDI-TOF MS. The
results are summarized in Table 2.
Semisynthesis of R2 and FnY356-R2s. Culture growth, ligation,
and protein purification were carried out as previously described21 with
minor modifications. After MESNA-mediated cleavage of R2(1-353)
from the chitin resin, excess MESNA was removed using a Sephadex
G-25 column (200 mL, 3 cm × 30 cm) equilibrated in cleavage buffer
(50 mM HEPES, pH 7.6, 500 mM NaCl). Furthermore, prior to
purification by MonoQ anion exchange chromatography, unbound
peptide was removed by concentration/dilution cycles using a YM-30
membrane. MonoQ purifications were performed under reducing
conditions with 1.5 mM DTT. Each FnY356-R2 was judged to be >95%
pure on the basis of SDS-PAGE analysis. Each R2 was further
characterized by ESI-MS, and the tyrosyl radical was quantitated by
UV-vis and EPR spectroscopic methods (Table 3).
Purification of R1 and Removal of Contaminating R2. R1 (SA
1900 nmol/min/mg) was purified as previously described.26 To reduce
the redox-active cysteines of R1 and the Y• of contaminating R2, which
copurifies with R1, R1 (∼40 µM) was incubated with 30 mM DTT
for 25 min at room temperature. Hydroxyurea, ATP, and CDP were
Fmoc-L-2,3,5,6-F4Y: 1H NMR (300 MHz, CD3OD) δ ) 3.06 (dd,
1H, Câ-H1, 9.2 Hz, 14.2 Hz), 3.26 (dd, 1H, Câ-H2, 5.4 Hz, 14.2 Hz),
4.23 (m, 3H, nonaromatic fluorenyl-H), 4.38 (dd, 1H, CR-H, 5.4 Hz,
(26) Salowe, S.; Bollinger, J. M., Jr.; Ator, M.; Stubbe, J.; McCraken, J.; Peisach,
(25) Lapatsanis, L.; Milias, G.; Froussios, K.; Kolovos, M. Synthesis 1983, 671.
J.; Samano, M. C.; Robins, M. J. Biochemistry 1993, 32, 12749.
9
1564 J. AM. CHEM. SOC. VOL. 128, NO. 5, 2006

pH Rate Profiles of FnY356−R2s in E. coli RNR
A R T I C L E S
then added to final concentrations of 30, 3, and 1 mM, respectively,
and the incubation was continued for an additional 20 min at room
temperature. R1 was then isolated using a Sephadex G-25 column (∼35
mL, 1.5 cm × 23 cm), which had been equilibrated in assay buffer (50
mM HEPES, 15 mM MgSO4, 1 mM EDTA, pH 7.6).
reactions described here, with the exception of the 2,3-F2Y-
peptide, on the same batch of R2-thioester. As a control for
all of our experiments, wt R2 made by the intein procedure
(containing V353G/S354C/Y356 and from here on referred to as
Y-R2) was generated from the same batch and is reported in
Table 3. Its specific activity is identical to the double mutant
made by site-directed mutagenesis after normalization for the
amount of Y•. The amount of Y• radical was quantified by
X-band EPR spectroscopy and by the dropline correction method
using the vis spectrum.30,31
Determination of pH Rate Profiles of FnY356-R2s. 2-[N-mor-
pholino]ethanesulfonic acid (MES), N-2-hydroxyethylpiperazine-N-
2-ethanesulfonic acid (HEPES), N-[Tris(hydroxymethyl)methyl]-3-amino-
propanesulfonic acid (TAPS), and 2-[cyclohexylamino]ethanesulfonic
acid (CHES) were used for the pH rate profiles. Each buffer was
adjusted to the desired pH with either KOH or HCl. Each reaction
contained in a volume of 230 µL: 50 mM HEPES (or MES, TAPS,
CHES), 15 mM MgSO4, 1 mM EDTA, 3 mM ATP, 1 mM [2-14C]-
CDP (6 × 106 - 1.5 × 107 cpm/µmol), 30 µM TR, 0.5 µM TRR, 1
mM NADPH, 3 µM R1, and 3 µM FnY356-R2. For each reaction,
nucleotides, TR, TRR, and R1 were incubated at 25 °C for ∼1.5-2
min. The reaction was then initiated by the addition of FnY356-R2 and
NADPH. At defined time points, 40 µL were withdrawn and quenched
with 25 µL of 2% perchloric acid. At the end of the time course, the
reactions were neutralized with 20 µL of 0.5 M KOH. Each sample
was incubated at -20 °C overnight to ensure complete precipitation
of potassium perchlorate. The samples were then spun down for 3 min
in a tabletop centrifuge. Each supernatant was transferred to a 1.5-mL
screw-top microfuge tube, to which was added 14 units of calf-intestine
alkaline phosphatase, 120 nmol of carrier deoxycytidine (dC), and Tris-
EDTA buffer (pH 8.5) to a final concentration of 75 mM and 0.15
mM, respectively. The amount of dC was then quantitated by the
method of Steeper and Steuart.27
Problems Encountered in the RNR Activity Assay. Be-
cause we wished to set lower limits of detection on the rates of
deoxynucleotide formation with the mutant RNRs, we made
two changes to the standard assays.21 First, R1 (specific activity
of 1.9 µmol/min/mg), as isolated, always copurifies with a small
amount of R2 (estimated to be <1% by SDS PAGE). We
therefore inactivated this contaminating R2 by reducing the
tyrosyl radical with hydroxyurea in the presence of DTT, ATP,
and CDP and showed, with control experiments in the absence
of the R2 mutant, that no dNDP was detectable within the time
frame of the assay. (Figure S1, Supporting Information). Second
the peptide used to generate semisynthetic R2 is a known
competitive inhibitor of the interaction of R2 with R1. An
additional purification step for R2 mutants was thus added to
the protocol to ensure that all of the peptide (covalently and
noncovalently bound to R2) was removed. In each R2, there is
now a cysteine at position 354 in place of the serine of the wt
R2. R2 mutants containing this cysteine are inactivated in the
presence of R1, CDP, and ATP in a slow, time-dependent
reaction.21 This inactivation presumably results from an uncou-
pling of the radical propagation pathway, resulting in the
reduction of Y122•. Control experiments showed that this
uncoupling did not occur to any significant extent under the
conditions of the assays (Figure S2, Supporting Information).
Finally, to ensure that the single amino acid substitution does
not effect the interaction between the R1 and R2, the Km values
for several of the mutants were determined and found to be
identical to Y-R2, as listed in Table 3.
Results
Synthesis and Characterization of FnY356/V353G/
S354C-R2s. R2 is a homodimer containing 375 amino acids
per monomer, which can be prepared semisynthetically by intein
methods.21 Residues 1-353 of R2 with its C-terminus fused to
the VmaI intein and a chitin-binding domain were made via
molecular biological methods. Residues 357-375 were syn-
thesized by automated solid-phase peptide synthesis with
residues 354, 355, and 356 being added in solution manually.21-23
A number of minor changes have been made from our initially
reported procedure to enhance recoveries of R2 and in efforts
to increase the amount of Y• per R2. In general, C354 of the
C-terminal tail of R2 can form a disulfide with the excess
peptide used in the ligation reaction to generate full-length R2.
Removal of the peptide is essential for accurate quantitation of
RNR activity as the peptide is a competitive inhibitor of the
nucleotide-reduction process.28 The final purification steps,
therefore, used concentration/dilution cycles to remove non-
covalently bound peptide as well as anion exchange chro-
matographay on a MonoQ column with DTT in all the buffers
to reduce the peptide-R2 disulfide bond and remove the
peptide. Although DTT has previously been shown to reduce
the Y• in R2,29 the minimal time required for this purification
step ensures that this side reaction does not occur. A summary
of the recovery of semisynthetic R2s, their characterization by
ESI-MS, the amount of Y• per R2, and the maximum specific
activity of each R2, are presented in Table 3.
pH Rate Profiles for wt and FnY356-R2s. Kinetic assays
for nucleotide reduction over the RNR-accessible pH range from
6.5 to 9 were carried out to determine if direct evidence for the
role of Y356 as a redox-active amino acid on a pathway between
R1 and R2 could be obtained. The results of our kinetic assays
for the various FnY356-R2s are shown in Figure 2 and are quite
dramatic. First, all of the FnY356-R2s are active between pH
6.5 to 8.5, and the activity is relatively high despite the
complexity of the semisynthetic procedure for R2 generation
(Table 3). Second, the activities of 2,3-F2Y356-R2, 3,5-F2Y356
-
R2, and 2,3,5-F3Y356-R2 are very close to that of Y-R2 in
the pH range from 6.5 to 7.6. On the alkaline side, the activity
drops off for all FnY derivatives with the dropoff beginning at
a lower pH for the 2,3,5-F3Y356-, 2,3,6-F3Y356-, and 2,3,5,6-
F4Y356-R2s. With 2,3,6-F3Y356-R2 and 2,3,5,6-F4Y356-R2, the
activity relative to Y-R2 is substantially reduced throughout
the entire pH rate profile. At pH values greater than 8.5, the
activity is below the lower limit of detection of our assay
method.
Over the past few years we have continually modified various
aspects of the semisynthesis of R2 to improve yields and the
amount of Y• recovered. We have performed all of the ligation
(27) Steeper, J. R.; Steuart, C. D. Anal. Biochemistry 1970, 34, 123.
(28) Climent, I.; Sjo¨berg, B. M.; Huang, C. Y. Biochemistry 1991, 31, 5164.
(29) Fontecave, M.; Gerez, C.; Mansuy, D.; Reichard, P. J. Biol. Chem. 1990,
265, 10919.
(30) Bollinger, J. M., Jr. Ph.D. Thesis, Massachusetts Institute of Technology,
1993.
(31) Bollinger, J. M., Jr.; Tong, W. H.; Ravi, N.; Huynh, B. H.; Edmondson,
D. E.; Stubbe, J. Methods Enzymol. 1995, 258, 278.
9
J. AM. CHEM. SOC. VOL. 128, NO. 5, 2006 1565

A R T I C L E S
Seyedsayamdost et al.
Figure 3. pH profiles of Ac-FnY•-NH2 peak-reduction potentials. The
Ep vs pH for reduction of (black s) Ac-Y•-NH2, (blue s) Ac-3,5-
F2Y•-NH2, (red s) Ac-2,3-F2Y•-NH2, (purple s) Ac-2,3,5-F3Y•-NH2,
(green s) Ac-2,3,6-F3Y•-NH2, and (orange s) Ac-F4Y•-NH224 in the
pH range accessible to RNR are shown. Lines indicate peak-reduction
potential differences of (40 mV (solid), 90 mV (short dashes), and 180
mV (long dashes) relative to those for Ac-Y•-NH2.
Figure 2. pH rate profiles of FnY356-R2s. The top panel shows an over-
lay of the pH rate profiles of Y-R2 and FnY356-R2s used in this
study. (Bottom panel) Blowup of the profiles of 2,3,6-F3Y356-R2 and
F4Y356-R2. Each point represents an average of two independent measure-
ments. Error was within 15% for all points: (black b) Y356-R2, (blue b)
3,5-F2Y356-R2, (red b) 2,3-F2Y356-R2, (purple b) 2,3,5-F3Y356-R2, (green
b) 2,3,6-F3Y356-R2, and (orange b) F4Y356-R2.
Discussion
Figure 4. Redox-potential regimes of RNR activity. Relative activities
of FnY356-R2s vs Y-R2, obtained from Figure 2, plotted as a function
of the peak-reduction potential difference between the corresponding
Ac-FnY•-NH2 and Ac-Y•-NH2:24 (blue b,O) 3,5-F2Y356-R2, (red
b,O) 2,3-F2Y356-R2, (purple O) 2,3,5-F3Y356-R2, (green b,O) 2,3,6-
F3Y356-R2, and (orange O) F4Y356-R2. Filled circles represent data points
where pH < pKa of the corresponding FnY; open circles represent data
points where pH > pKa of the corresponding FnY. Three different regimes
of RNR activity are highlighted as either gated by a physical/conformational
change (regime 1), rate limited by radical transport (regime 2), or reduced
to background levels (regime 3), depending on the peak-reduction potential
difference between the corresponding Ac-FnY•-NH2 and Ac-Y•-NH2.
The FnYs in Table 1 were incorporated into R2 by the intein
procedure. They were chosen to perturb the protonation state
and reduction potential of residue 356 in an effort to examine
its role in the radical-propagation process. For each mutant R2,
the ability to make deoxynucleotides was measured over the
accessible pH range for RNR (6.5-9). The results, Figure 2,
reveal that the FnYs markedly perturb the activity, suggesting
that these mutants will be useful for examining radical propaga-
tion.
The observed drop in RNR activity of FnY-R2 mutants
relative to Y-R2 may be related to changes in their reduction
potentials. Unfortunately, one cannot measure the reduction
potentials of these FnYs within R2. We can, however, measure
the peak-reduction potentials of Ac-FnY-NH2 derivatives.
These values were determined in the pH range in which RNR
activity can be measured (6.5-9, Figure 3) using differential
pulse voltammetry (DPV).24 Previous work has shown that the
peak potential for the tyrosyl radical as measured by DPV is
very similar to its reduction potential measured by cyclic
voltammetry and pulse radiolysis.32 The data in Figure 3 show
that the fluorinated tyrosine derivatives have potentials that vary
from -50 to +270 mV relative to tyrosine.
To gain insight into the relationship between RNR activity
and the reduction potential of residue 356, Figure 4 was
generated. In this Figure, the activity data of each mutant relative
to that of Y-R2 is plotted against the peak-reduction potential
difference between Ac-FnY-NH2 and Ac-Y-NH2. Analysis
of the data in this fashion allows for the removal of the pH-
dependent activity inherent to Y-R2 and assumes that all of
the mutants have the same inherent pH dependence. The basis
of this pH dependence is not understood. The construction of
Figure 4 also assumes that the reduction potential of Y356• and
(32) Tommos, C.; Skalicky, J. J.; Pilloud, D. L.; Wand, A. J.; Dutton, P. L.
Biochemistry 1999, 38, 9495.
9
1566 J. AM. CHEM. SOC. VOL. 128, NO. 5, 2006

pH Rate Profiles of FnY356−R2s in E. coli RNR
A R T I C L E S
FnY356• varies 59 mV/pH unit, as with Ep(Ac-FnY•-NH2/
Ac-FnY-NH2) and requires proton loss to bulk solution
concomitant with oxidation. Furthermore, we assume that if the
protein environment perturbs the reduction potential for Y in
Y-R2, it has the same relative effect on each FnY in the
corresponding FnY356-R2. We note that the data in Figure 4
reveal some scatter, which we attribute in part to the difficulties
with making R2 semisynthetically. Moreover, fluorine substitu-
tion may affect the conformation of residue 356 and its inter-
actions with its environment. Four fluorines certainly have
altered steric properties relative to four hydrogens, stemming
from the slightly larger van der Waals radius of F vs H (1.35
vs 1.2 Å, respectively) and the longer C-F bond vs the C-H
bond (1.38 and 1.09 Å, respectively). Finally, using the pKa of
Ac-FnY-NH2 as a benchmark for FnY356, a difference in charge
of this residue may result owing to the pKa differences of each
of these fluorinated tyrosines. Given the incongruities that may
be engendered by fluorine substitution, the data in Figure 4 are
quite striking.
When the driving force for Y356 oxidation exceeds 80 mV
but is less than 200 mV relative to Y, the rate of nucleotide
reduction correlates with the difference in peak-reduction
potential, suggesting that there has been a change in the rate-
limiting step to one involving radical transport through FnY356
.
This point is demonstrated for 2,3-F2Y356-R2 and 2,3,5-
F3Y356-R2. The short dashed lines in Figure 3 indicate two
pH values (8.6 and 7.8) at which the peak-reduction potential
difference between Ac-2,3-F2Y•-NH2 and Ac-Y•-NH2 and
between Ac-2,3,5-F3Y•-NH2 and Ac-Y•-NH2 is ∼90 mV.
If activity is determined solely by the reduction potential, then
the relative activity of 2,3-F2Y356-R2 at pH 8.5 should
approximate that of 2,3,5-F3Y356-R2 at pH 7.8. The relative
activities observed are 85% and 81%, respectively.
At potential differences greater than 150 mV (regime 3), the
rate of radical transfer appears to become so slow that nucleotide
reduction is minimal, and for potential differences >200 mV,
activity is abolished. Note that the relative activities of 2,3,6-
F3Y356-R2 at pH 8.4 and of F4Y356-R2 at pH 7.4, where the
peak-reduction potential difference between these corresponding
Ac-FnY•-NH2s and Ac-Y•-NH2 is 180 mV (long dashed
lines, Figure 3), are 11% and 8%, respectively. The direct
correlation between activity and peak-reduction potential strongly
suggests that Y356 is a redox-active amino acid on the pathway.
We also examined whether the protonation state of the
FnYs at position 356 affects RNR activity. Our previous
studies with NO2Y356-R221 showed that the pKa of this resi-
due at the R1/R2 interface is unperturbed relative to that of
Ac-NO2Y-NH2 in solution. Similar measurements of the pKas
of the FnYs within R2 cannot be made because these pheno-
lates do not absorb in the visible region of the spectrum as
does 3-NO2Y-. We have thus assumed that the pKas of the Ac-
FnY-NH2 are indicative of those within the R1/R2 complex.
We have incorporated this information into Figure 4 in an effort
to determine if the protonation state of FnY has any correlation
with RNR activity: Filled circles in Figure 4 represent data
points where pH < pKa of the corresponding FnY; conversely,
open circles represent data points where pH > pKa of the
corresponding FnY. This analysis reveals that most of the FnYs
used in this study are deprotonated or exhibit changes in
protonation state, whereas Y, because of its higher pKa, remains
protonated throughout the range where the activity of RNR can
be measured. In contrast with the peak-potential data, no clear
correlation exists between the protonation state of the FnY356
and the activity of the corresponding mutant.
Our results do not provide any insight about the destination
of the proton from Y356 in Y-R2 in the forward PCET event
or its delivery in the back PCET event. A comparison of the
sequences of the disordered C-terminal tails of >160 R2s
suggests that, although there are no absolutely conserved
residues, E350 is present in 129 of 162 R2 sequences.33 Thus, it
is possible that this residue functions as a proton acceptor in
the forward PCET reaction and a proton donor in the reverse
reaction. The importance of this residue is supported by
mutagenesis studies carried out on many of the residues within
the C-terminus of R2. Individual substitution of Y356 and E350
with alanine was found to affect activity. In the former case,
R2 was inactive, and in the latter case, the rate was reduced
250-fold.34
Three distinct activity regimes emerge from Figure 4. When
peak-reduction potential differences are -50 to 80 mV (regime
1), nucleotide-reduction activity falls within the range of that
observed for Y-R2; both 2,3-F2Y356-R2 and 3,5-F2Y356-R2
exhibit 90-120% of the activity obtained for Y-R2. As the
difference in peak-reduction potential increases from 80 to 200
mV (regime 2), the RNR activity drops off dramatically. At
>200 mV difference (regime 3), no detectable activity of RNR
is observed. 2,3,5-F3Y356-R2 spans the potential range from
∼25 to 145 mV, whereas 2,3,6-F3Y356-R2 and F4Y356-R2 have
very low activities and high peak-reduction potential differences
at all pHs accessible to RNR.
Our interpretation of these data is summarized by the three
activity regimes marked with dashed lines in Figure 4. When
the reduction potential difference is within the range of -50 to
80 mV (regime 1), the rate of nucleotide reduction in the
FnY356-R2s is governed by the same factors as nucleotide
reduction by Y-R2. Presteady-state experiments have shown
that the rate-limiting step in E. coli RNR is a physical step or
steps prior to the radical propagation process that is gated by
the binding of a substrate and allosteric effector.13 In the steady
state, the rate-determining step is proposed to be the same
physical step or, alternatively, a step involved in rereduction of
R1 or a conformational change associated with this rereduction,
after dNDP formation and regeneration of the Y•.21 In either
case, the rate of nucleotide reduction should be independent of
the Y356•-reduction potential, leading us to assign regime 1 to
the physically/conformationally gated step. The peak-reduction
potential of Ac-3,5-F2Y•-NH2 falls effectively in the confor-
mationally gated regime over all pHs, whereas the potentials
of Ac-2,3-F2Y•-NH2 and Ac-2,3,5-F3Y•-NH2 falls within
this regime for pH <∼8.4 and pH e∼7.6, respectively. At a
pH of 6.81 (solid line, Figure 3), the peak-reduction potential
difference between Ac-3,5-F2Y•-NH2 and Ac-Y•-NH2 is
-40 mV. Similarly, at pHs of 6.6 and 6.93 (solid lines, Figure
3), the peak-reduction potentials of Ac-2,3-F2Y•-NH2 and
Ac-2,3,5-F3Y•-NH2, respectively, are ∼40 mV higher than
that of Ac-Y•-NH2. The relative activities are all identical
within error (108, 99, and 100% of Y-R2, respectively, for
3,5-F2Y356-R2, 2,3-F2Y356-R2, and 2,3,5-F3Y356-R2).
(33) Ge, J. Ph.D. Thesis, Massachusetts Institute of Technology, 2003.
9
J. AM. CHEM. SOC. VOL. 128, NO. 5, 2006 1567

A R T I C L E S
Seyedsayamdost et al.
The mechanistic details for radical transport involving a
tyrosyl radical at this position are important to consider. Radical
transport involving the deprotonated fluorotyrosines can proceed
by simple ET (eq 1). Conversely, a stepwise ET/PT reaction of
a protonated tyrosine confronts a large thermochemical bias
because ∆G° for the initial ET or PT event is high.35-38
Accordingly, these thermodynamic constraints implicate pro-
ton-coupled electron transfer (PCET) for radical propagation
(eq 2).39-42
of 10-15 Å by a tunneling mechanism and, consequently, do
not involve aromatic amino acid radical intermediates.44-56 The
participation of tyrosine in the radical-initiation pathway of RNR
steps beyond simple ET because the management of proton and
electron inventories is required for charge transport involving
the amino acid. In the accompanying paper, we have shown
that fluorotyrosine derivatives have radical peak-reduction
potentials from -50 to +270 mV relative to Y and pKas that
range from 5.6 to 7.8. By exploiting these properties for FnYs
substituted at Y356, our data strongly suggest that this residue
in R2 is a redox-active amino acid on the radical-propagation
pathway in the class I RNRs. This result is important as Y356 is
the only residue on the pathway that is not observable
structurally. Different activity regimes of RNR may be accessed
by tuning the reduction potential of FnY356• such that radical
transport becomes rate limiting in nucleotide reduction. Future
experiments using presteady-state kinetic analyses will focus
on establishing that the radical-propagation process does, in fact,
become rate limiting at peak-reduction potential differences >80
mV, as proposed on the basis of our reported data, as well as
on detection of radical amino acid intermediates during turnover.
Studies reported herein further strongly suggest that hole
FnY• + e- f FnYO-
(1)
(2)
Y• + e- + H+ f YOH
In Y-R2, therefore, the proton must be lost concomitant with
Y356 oxidation via PCET.
Assuming the radical-propagation pathway shown in Figure
1, our results with the FnY356-R2s suggest that hole migra-
tion between the W48-R2-Y356-R2 pair and the Y356-R2-
Y731-R1 does not occur by an obligatory hydrogen atom
transfer as phenolates of the FnY356-R2s are active. Further-
more, this activity for the phenolates is inconsistent with the
models based on site-directed mutagenesis studies15 and the
theory43 that W48, Y356, and Y731 are connected by way of a
hydrogen-bonding network that is required for conformational
gating of the radical-propagation process. The turnover number
of 2,3,5-F3Y356-R2 at pH 7.6, for example, is identical to that
of Y-R2, where it is deprotonated if one assumes that the pKa
is not perturbed from that of Ac-2,3,5-F3Y-NH2, as indicated
by our studies on NO2Y356-R2.21 W48 and Y731 must, therefore,
communicate by ET involving the 2,3,5-F3Y-/2,3,5-F3Y• couple.
The insertion of a negative charge at the R1/R2 interface
apparently does not affect the overall RNR activity. These
results, taken together with our preliminary studies,22,23 lead us
to conclude that the proton from Y356 in the wt enzyme is
transferred (off pathway) to solvent, either directly or via a base
or bases in R2 or R1.
migration between the W48-R2-Y356-R2 pair and the Y356
-
R2-Y731-R1 does not occur by a hydrogen-atom transfer, as
phenolates of the FnY356-R2s are active. The obligation of
protons to the pathway from Y122 to Y356 in R2 and from Y731
into C439 at the active site of R1 remains an open question and
needs to be investigated, most efficiently by direct kinetic
analysis of the radical intermediates. These studies are currently
underway.
Acknowledgment. We thank John H. Robblee for help with
the quantitation of radical per R2 dimer via EPR and for helpful
discussions and the National Institutes of Health for support of
this work GM47274 (D.G.N.) and GM29595 (J.S.).
Supporting Information Available: Solvent systems and Rf
values for purification of the Fmoc-FnYs (Table S1); background
activity assay on prereduced R1 (Figure S1); control experiments
for uncoupling (Figure S2). This material is available free of
charge via the Internet at http://pubs.acs.org.
Conclusions
A major unresolved issue in class I RNRs is the mechanism
of radical propagationshow does Y122• on R2, transiently and
reversibly, generate a C439• on R1 proposed to be >35 Å
away? This problem is different from most ET reactions in
biology studied to date, which typically occur over a distances
JA055927J
(44) Page, C. C.; Moser, C. C.; Chen, X.; Dutton, P. L. Nature 1999, 402, 47.
(45) Electron Transfer in Chemistry; Balzani, V., Ed.; Wiley-VCH: Weinheim,
Germany, 2001; Vol. 3, Part 1.
(46) Gray, H. B.; Winkler, J. R. Annu. ReV. Biochem. 1996, 65, 537.
(47) Gunner, M. R.; Robertson, D. E.; Dutton, P. L. J. Phys. Chem. 1986, 90,
3783.
(48) Farver, O.; Pecht, I. Biophys. Chem. 1994, 50, 203.
(49) Namslauer, A.; Braenden, M.; Brzezinski, P. Biochemistry 2002, 41, 10369.
(50) Tollin, G. In Electron Transfer in Chemistry; Balzani, V., Ed.; Wiley-
VCH: Weinheim, Germany, 2001; Vol. 4, Part 1, p 202.
(51) Davidson, V. L. Acc. Chem. Res. 2000, 33, 87.
(34) Climent, I.; Sjo¨berg, B. M.; Huang, C. Y. Biochemistry 1992, 31, 4801.
(35) Mayer, J. M.; Rhile, I. J. Biochim. Biophys. Acta 2004, 1655, 51.
(36) Sjo¨din, M.; Ghanem, R.; Polivka, T.; Pan, J.; Styring, S.; Sun, L.;
Sundstro¨m, V.; Hammarstro¨m, L. Phys. Chem. Chem. Phys. 2004, 6, 4851.
(37) Sjo¨din, M.; Styring, S.; Wolpher, H.; Xu, Y.; Sun, L.; Hammarstro¨m, L. J.
Am. Chem. Soc. 2005, 127, 3855.
(38) Reece, S. Y.; Nocera, D. G. J. Am. Chem. Soc. 2005, 127, 9448.
(39) Chang, C. J.; Chang, M. C. Y.; Damrauer, N. H.; Nocera, D. G. Biochim.
Biophys. Acta 2004, 1655, 13.
(52) McLendon, G.; Hake, R. Chem. ReV. 1992, 92, 481.
(53) Millett, F.; Durham, B. Biochemistry 2002, 41, 11315.
(54) Nocek, J. M.; Zhou, J. S.; Forest, S. D.; Priyadarshy, S.; Beratan, D. N.;
Onuchic, J. N.; Hoffman, B. M. Chem. ReV. 1996, 96, 2459.
(55) Lewis, F. D. Electron Transfer and Charge Transport in DNA. In Electron
Transfer in Chemistry; Balzani, V., Ed.; Wiley-VCH: Weinheim, Germany,
2001; Vol. 3, Part 5, p 105.
(40) Cukier, R. I.; Nocera, D. G. Annu. ReV. Phys. Chem. 1998, 49, 337.
(41) Cukier, R. I. Biochim. Biophys. Acta 2004, 1655, 37.
(42) Hatcher, E.; Soudackov, A. V.; Hammes-Schiffer, S. J. Am. Chem. Soc.
2004, 126, 5763.
(43) Himo, F.; Siegbahn, P. E. M. Chem. ReV. 2003, 103, 2421.
(56) Boon, E. M.; Barton, J. K. Curr. Opin. Struct. Biol. 2002, 12, 320.
9
1568 J. AM. CHEM. SOC. VOL. 128, NO. 5, 2006