Mechanistic implications of a linear free-energy correlation of rate constants for the reduction of active- and Met-R2 forms of E. coli ribonucleotide reductase with eight organic radicals

Article abstract of DOI:10.1021/ja993412k

Cross-reaction rate constants k₁₂ (22 °C) at pH 7.0 have been determined for the reduction of Fe(III)₂ and tyrosyl-radical-containing active-R2 from E. coli ribonucleotide reductase with eight organic radicals (OR), e.g., MV·⁺ from methyl viologen. The more reactive OR's were generated in situ using pulse radiolysis (PR) techniques, and other OR's were generated by prior reduction of the parent with dithionite, followed by stopped-flow (SF) studies. In both procedures it was necessary to include consideration of doubly-reduced parent forms. Values of k₁₂ are in the range 10⁹ to 10⁴ M^-1 s^-1 and reduction potentials E₁/(o) for the OR vary from -0.446 to +0.l94 V. Samples of E. coli active-R2 also have an Fe₂(III) met-R2 component (with no Tyr·), which in the present work was close to 40%. From separate experiments met-R2 gave similar k₁₂ rate constants (on average 66% bigger) to those for active-R2, suggesting that reduction of the Fe₂/(III) center is the common rate-limiting step. A single Marcus free-energy plot of log k₁₂ - 0.5 log f vs - E₁/(o)/0.059 describes all the data, and the slope of 0.54 is in satisfactory agreement with the theoretical value of 0.50. It is concluded that the ratelimiting step involves electron transfer. In addition, the intercept at -E₁/(o)/0.059 = 0 is 5.94, where values of the reduction potential and self-exchange rate constant for met-R2 contribute to this value. To maintain electroneutrality at the ~10 A buried active site H⁺ uptake is also required. For both e^- and H⁺ transfer the conserved pathway Trp-48, Asp-237, His-118 to Fe(A) is a possible candidate requiring further examination.

Full text of DOI:10.1021/ja993412k

2206
J. Am. Chem. Soc. 2000, 122, 2206-2212
Mechanistic Implications of a Linear Free-Energy Correlation of Rate
Constants for the Reduction of Active- and Met-R2 Forms of E. coli
Ribonucleotide Reductase with Eight Organic Radicals
A. Mark Dobbing,† Christopher D. Borman,† Mark B. Twitchett,† David N. Leese,†
G. Arthur Salmon,‡ and A. Geoffrey Sykes*,†
Contribution from the Department of Chemistry, The UniVersity of Newcastle,
Newcastle upon Tyne, NE1 7RU, UK, and The UniVersity of Leeds, Cookridge Radiation
Research Centre, Leeds, LS16 6PB, UK
ReceiVed September 20, 1999
Abstract: Cross-reaction rate constants k12 (22 °C) at pH 7.0 have been determined for the reduction of FeIII
2
and tyrosyl-radical-containing active-R2 from E. coli ribonucleotide reductase with eight organic radicals (OR),
e.g., MV•+ from methyl viologen. The more reactive OR’s were generated in situ using pulse radiolysis (PR)
techniques, and other OR’s were generated by prior reduction of the parent with dithionite, followed by stopped-
flow (SF) studies. In both procedures it was necessary to include consideration of doubly-reduced parent forms.
Values of k12 are in the range 109 to 104 M-1 s-1 and reduction potentials Eo for the OR vary from -0.446
1
to +0.194 V. Samples of E. coli active-R2 also have an FeIII2 met-R2 component (with no Tyr•), which in the
present work was close to 40%. From separate experiments met-R2 gave similar k12 rate constants (on average
66% bigger) to those for active-R2, suggesting that reduction of the FeIII center is the common rate-limiting
2
step. A single Marcus free-energy plot of log k12 - 0.5 log f vs -Eo /0.059 describes all the data, and the
1
slope of 0.54 is in satisfactory agreement with the theoretical value of 0.50. It is concluded that the rate-
limiting step involves electron transfer. In addition, the intercept at -Eo /0.059 ) 0 is 5.94, where values of
1
the reduction potential and self-exchange rate constant for met-R2 contribute to this value. To maintain
electroneutrality at the ∼10 Å buried active site H+ uptake is also required. For both e- and H+ transfer the
conserved pathway Trp-48, Asp-237, His-118 to FeA is a possible candidate requiring further examination.
Introduction
a µ-oxo bridged FeIII2, but the µ-oxo is no longer present in
red-R2, and proton uptake is again evident. Therefore in (1)
three electrons and as many as three protons are taken up in
Ribonucleotide reductase (RNR)1-6 catalyzes the de novo
synthesis of the four deoxyribonucleotide components of DNA.
Of the three classes of RNR,5 those containing an FeIII2 center
are the best characterized/understood, and that from E. coli is
the most extensively studied. It consists of two nonidentical
protein components R1 (Mr ) 2 × 85.5 kDa) and R2 (Mr ) 2
× 43.5 kDa), each having two identical subunits (hence R2â2).7
Only 16 of the R2 residues are conserved within 12 known
sequences. The tyrosyl radical (Tyr•) of active-R2 is a depro-
tonated phenolate form.4 X-ray crystal structures of R1 and R2
from E. coli have been reported.8,9 In the R2 case the structures
determined are for the met-R2 and reduced R2 forms,9,10
the conversion of FeIII2‚‚‚Tyr• to FeII ‚‚‚(TyrH) with maintenance
2
of electroneutrality. Features of the met-R2 structure are the
position of the FeIII2 and tyrosyl-radical-forming Tyr-122, which
are buried and ∼10 Å from the protein surface. The µ-oxo
coupled FeIII is close to Tyr-122, with FeA 5.3 Å from the
2
phenolate O-atom. The FeIII2 stabilizes the Tyr• radical, and the
reduced FeII state is required for regeneration of Tyr•.11 The
2
mechanism of redox interconversion of R2 forms in (1) is a
subject of current interest.12-16
The aim in these studies was to determine rate constants k12
for cross reactions involving the reduction of active-R2 from
E. coli with eight organic radicals (OR), Figure 1. The rate
-
8
-
2e
8
e
9
FeIII ‚‚‚Tyr•
FeIII2‚‚‚(TyrH)
FeII ‚‚‚(TyrH)
(1)
2
2
(7) Nilsson, O.; Lundqvist, T.; Hahne, S.; Sjo¨berg, B.-M. Biochem. Soc.
Trans. 1988, 16, 91.
active-R2
met-R2
red-R2
(8) Uhlin, U.; Eklund, H. Nature (London) 1994, 70, 533.
(9) Nordlund, P.; Eklund, H. J. Mol. Biol. 1993, 232, 123.
(10) Logan, D. T.; Su, X.-D.; Åberg, A.; Regnstro¨m, K.; Hajdu, J.;
Eklund, H.; Nordlund, P. Structure 1996, 4, 1053.
(11) Ravi, N.; Bollinger, J. M., Jr.; Huynh, B. H.; Edmondson, D. E.;
Stubbe, J. J. Am. Chem. Soc. 1994, 116, 8007.
(12) Han, J.-Y.; Gra¨slund, A.; Thelander, L.; Sykes, A. G. J. Biol. Inorg.
Chem. 1997, 2, 287.
(13) Dobbing, A. M.; Han, J.-Y.; Sykes, A. G. J. Biol. Inorg. Chem.
1998, 3, 480.
(14) Han, J.-Y.; Swarts, J. C.; Sykes, A. G. Inorg. Chem. 1996, 35, 4629.
(15) Swarts, J. C.; Aquino M. A. S.; Han, J.-Y.; Lam, K.-Y.; Sykes, A.
G. Biochim. Biophys. Acta. 1995, 1247, 215.
(16) Davydov, A.; Schmidt, P. P.; Gra¨slund, A. Biochem. Biophys. Res.
Commun. 1996, 219, 213-218.
where protonation of the phenolate O-atom of the nonradical
met and red-R2 forms is indicated. Met-R2 and active-R2 have
† The University of Newcastle.
‡ The University of Leeds.
(1) Sjo¨berg, B.-M. Struct. Bonding 1997, 88, 139-173.
(2) Gra¨slund, A.; Sahlin, M. Annu. ReV. Biophys. Biomol. Struct. 1996,
25, 259-286.
(3) Wallar, B. J.; Lipscomb, J. D. Chem. ReV. 1996, 96, 2625-2657.
(4) Andersson, K. K.; Gra¨slund, A. AdV. Inorg. Chem. 1995, 43, 359-
408.
(5) Sjo¨berg, B.-M. Structure 1994, 2, 793.
(6) Fontecave, M.; Nordlund, R.; Eklund, H.; Reichard, P. AdV. Enzymol.
1992, 65, 147-183.
10.1021/ja993412k CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/29/2000

Reduction of E. coli Ribonucleotide Reductase
J. Am. Chem. Soc., Vol. 122, No. 10, 2000 2207
Figure 1. Formulas (pH ∼7) of eight parent reagents (X) used to generate the organic radicals OR (X•- in this study).
constants are appraised in terms of the Marcus equations for
outer-sphere electron-transfer reactions
is no longer appropriate since no account was taken of the direct
reduction of R2 by excess dithionite,13 which at the time was
reported to be negligible.
k122 ) k11k22K12f
(2)
(3)
Experimental Section
(log K12)2
Protein Source. Wild-type E. coli R2 was prepared as the ââ protein
from an E. coli overproducing strain,24 kindly provided by Professor
B.-M. Sjo¨berg, University of Stockholm, Sweden. Concentrated solu-
tions of R2 (∼0.3 mM), consisting of a mix of active-R2 (60%) and
met-R2 (40%) forms, were stored in small batches at -80 °C. Protein
was made air-free by dialyzing against deaerated buffer. Met-R2 was
prepared by reacting the above R2 mix with hydroxyurea (50 mM) at
25 °C for ∼30 min, after which excess hydroxyurea was removed using
a Sephadex G-25 column or by dialysis. The dilute met-R2 solution
was concentrated by ultra-dialysis against deaerated 50 mM Tris/HCl
at pH 7.5 containing 20% glycerol.25 The UV-vis spectra of active-
and met-R2 allow the composition to be determined using known
log f )
4 log(k11k22/Z2)
where k11 and k22 are self-exchange rate constants for the OR
and relevant R2 couple (which we specify later), respectively,
and the collision frequency Z is assumed to be 1011M-1 s-1 17
.
Using the Nernst equation the equilibration constant K12 for the
cross reaction is given by log K12 ) (Eo - Eo )/0.059, where
2
1
Eo1 and Eo2 are reduction potentials in volts (vs NHE) applying
to the OR and R2 couples, respectively. Equation 2 can be
expressed in the alternative form
absorption coefficients (ꢀ).26 In the case of active R2 a sharp band at
26
410 nm (ꢀ ∼6600 M-1 cm-1
)
is largely due to Tyr•. The Tyr• also
(log k12 - 0.50 log f) ) 0.50(log k11 + log k22
+
has a weak absorbance at ∼600 nm, which contributes to the
characteristic green color of active R2. Fully reduced FeII2 protein has
little or no absorbance at >400 nm. Concentrations of protein were
determined spectrophotometrically using the difference in absorption
Eo /0.059) - 0.50Eo /0.059 (4)
2
1
To test for (4) the left-hand side is plotted against -Eo /0.059.
1
coefficients ꢀ280 - ꢀ310 ) 1.2 × 105 M-1 cm-1 27
.
Hence the free-energy dependence of ∆G† on ∆Go is
12
12
explored, or more precisely the Eo term of ∆Go12. Active-R2
Buffers. For kinetic experiments all solutions were made up to pH
7.0, I ) 0.100 M, using sodium hydrogen phosphates (Sigma). During
the preparation tris(hydroxymethyl)aminomethane (Tris; pH range 7.1-
8.9) was used.
1
has an FeIII2 met-R2 component that has to be taken into account.
No preparation to date has <25% of met-R2.18 Rate constants
(k12) for the reaction of met-R2 were also determined by
independent experiments, and are an important mechanistic
contribution to the studies described.
Reagents for Organic Radicals. The following parent forms, Figure
1, with trivial, systematic, and abbreviated names as indicated, were
used to generate the one-electron reduced organic radicals (OR): methyl
viologen, C12H14N2Cl2 (1,1′-dimethyl-4,4′-bipyridinium dichloride),
MV2+; benzyl viologen, C24H22N2Cl2 (1,1′-dibenzyl-4,4′-bipyridinium
dichloride), BV2+; phenosafranin, C18H15N4Cl (3,7-diamino-5-phen-
ylphenazinium chloride), Pf+; riboflavin C17H20N4O6 (vitamin B2), Rb;
resorufin, C12H6NO3Na (7-hydroxy-3H-phenoxazin-3-one sodium salt),
Rf-; methylene blue C16H18N3SCl (3,7-bis(dimethylamino)phenazothion-
ium chloride), MB+; toluidine blue, (C15H6N3SCl)2‚ZnCl2 (approximate
formula only), TB+; and indo-phenol, C12H6Cl2NO2Na (phenolindo-
2,6-dichlorophenol sodium salt), IP-. These were obtained from Sigma
Chemicals, except MB+ and IP- (BDH), and TB+ (Fluka). Toluidine
Blue (1 g) was recrystallized by being dissolved in H2O (18 mL) at 80
°C. Ethanol (13.5 mL) was added with stirring over ∼10 min at 0 °C.
The solid obtained was filtered off, washed with cold ether, and dried
in a desiccator containing silica gel (yield 28%). Redox titration with
The procedures used have been tested previously for the
reduction of cytochrome c(III) with the same eight OR’s,19
where a plot of (4) has a slope of 0.49 (theoretical value 0.50)
and intercept at -Eo /0.059 ) 0 of 6.57. The latter was
1
calculated as 6.55 from known parameters and assuming a
common k11 of 1.0 × 106 M-1 s-1 for the self-exchange of the
different OR’s with the parent form.20-22 It is possible therefore
to proceed with this same value of k11 in the present studies,
and comment on the magnitude of Eo2 and k22 respectively for
R2. The procedure initially devised to study these reactions,23
(17) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265-
322.
(18) Lynch, J. B.; Juarez-Garcia, C.; Mu¨nck, E.; Que, L., Jr. J. Biol.
Chem. 1989, 264, 8091.
(19) Borman, C. D.; Dobbing, A. M.; Salmon, G. A.; Sykes, A. G. J.
Phys. Chem. B 1999, 103, 6605.
(20) de Oliveira, L. A. A.; Haim, A. J. Am. Chem. Soc. 1982, 104, 3363.
(21) Curtis, J. C.; Sullivan, B. P.; Meyer, T. J. Inorg. Chem. 1980, 19,
3833.
(22) Tsukahara K.; Wilkins, R. G. J. Am. Chem. Soc. 1985, 107, 2632.
(23) Lam, K.-Y.; Fortier, D. G.; Sykes, A. G. J. Chem. Soc., Chem.
Commun. 1990, 1019.
(24) Sjo¨berg, B.-M.; Hahne, S.; Karlsson, M.; Jo¨rnvall, H.; Goransson,
M.; Uhlin, B. E. J. Biol. Chem. 1986, 261, 5658.
(25) Barlow, T.; Eliasson, R.; Platz, A.; Reichard, P.; Sjo¨berg, B.-M.
Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 1492.
(26) Petersson, L.; Gra¨slund, A.; Ehrenberg, A.; Sjo¨berg, B.-M.; Reichard,
P. J. Biol. Chem. 1980, 255, 6706.
(27) Climent, I.; Sjo¨berg, B.-M.; Huang, C. Y. Biochemistry 1992, 31,
4801.

2208 J. Am. Chem. Soc., Vol. 122, No. 10, 2000
Dobbing et al.
dithionite gave 81% reactivity in accordance with that expected from
the formula (C15H6N3SCl)2‚ZnCl2, and was used in this work, see also
ref 19. Some difficulties were encountered in dissolving resorufin, and
solutions were prepared by sonicating appropriate mixes for 1 h and
then filtering through a 2 µm Acrodisc filter (Gellman).
Reactions with R2 were monitored at the peak positions λ/nm for
the OR’s as follows: MV•+ (606), BV•+ (555), Pf• (670), Rb•- (550
and 600), and Rf•2- (700). At these wavelengths absorption by the
radical is large and absorption by the protein and parent forms is
relatively small. Rate constants reported are average values using 3-4
fresh amounts of the same reactant solutions.
The parent forms MV2+ and BV2+ are colorless and the OR’s have
peaks λ/nm (ꢀ/M-1 cm-1) at 606 (12400) and 555 (11250), respec-
tively.28,29 All the other parent forms are colored with peak positions
Pf+ 521 (3.3 × 104), Rb 445 (1.1 × 104), Rf- 570 (4.5 × 104), MB+
660 (7.8 × 104), TB+ 632 (5.8 × 104), and IP- 603 (1.4 × 104).
Methylene Blue has been fairly extensively studied and the radical MB•
shown to disproportionate 2MB• a MB- + MB+ with K ∼ 2 × 105.
The rate constant for the forward reaction is in the range (1.5-3.0) ×
Stoichiometries for Parent Forms with Dithionite. In the stopped-
flow studies prior reduction of parent MB+, TB+, and IP- forms with
dithionite is required.19 The stoichiometries were determined by addition
of ∼2 mM sodium dithionite in air-free conditions (glovebox, O2 < 3
ppm) until only a trace of color remained.19 Dithionite is a two-
equivalent reducing agent S2O4 - 2e- f 2SO2 (i.e. SO32-). The
2-
number of moles of dithionite required for bleaching of the parent was
0.98 for MB+ and 0.96 for TB+, and 0.47 mol in the case of the quinone
type molecule IP-. In the first two instances therefore double reduction
and in the latter single reduction of the parent is observed.
109M-1s-1 30 Double reduction of parent forms (with S2O42-) and/or
.
formation of the double reduced form by disproportionation of OR has
to be considered in this work.
Electrochemistry. Reduction potentials (Eo1 vs NHE) for the parent/
OR couples were checked by cyclic voltammetry (CV) using a Princeton
Applied Research model 173 potentiostat, as previously described.19
In all but two cases these were within 10 mV of literature values,31 see
listing in Table 5. One exception is with Toluidine Blue, for which a
value of 0.015 V, as compared to a literature value of 0.034 V,32 has
been obtained. The other is with Indo-Phenol for which a 22 mV smaller
value is obtained.
Procedures for Pulse Radiolysis Studies. Experiments were carried
out on a Van de Graaff accelerator at the Cookridge Radiation Research
Centre, University of Leeds, using a triple-pass cell (6.9 cm light path
length) and a 2.5 MeV (∼4 × 10-16 J) beam of electrons.33 Pulse lengths
were 0.6 µs, and under the conditions adopted the production of radicals
by each pulse was as in
Procedure for Stopped-Flow Studies. An Applied Photophysics
UV-vis stopped-flow spectrophotometer was used to monitor reactions.
Rigorous air-free conditions were achieved by replacing connecting
leads on the stopped-flow by polyetheretherketone (PEEK) tubing which
has low O2 permeability, and by prior bubbling of N2 through all but
R2 solutions which were stored long-term under N2. The flow system
was washed with dithionite immediately prior to use. Reduction of the
three parent forms MB+, TB+, and IP- was by dropwise addition of
concentrated (∼10 mM) sodium dithionite using a Gilson pipetman in
a glovebox (O2 < 3 ppm) until there was little color remaining. An
excess of dithionite was avoided because of possible contributions from
direct reduction of active- and met-R2.13 Reactant concentrations used
were in the range 4.1-9.9 µM for parent and 30-47 µM for R2.
Conditions were 22.0 ( 0.1 °C, pH 7.0 (45 mM phosphate), I ) 0.100
( 0.002 M. Reactions were monitored at peak positions for reformation
of the parent forms λ/nm (ꢀ/M-1cm-1): MB+ 660 (7.4 × 104), TB+
632 (5.8 × 104), and IP- 603 (1.4 × 104). At wavelengths >600 nm
absorption coefficients for active- and met-R2 are <600 M-1cm-1. The
reactivities with TB+ and IP- were also monitored at the 410 nm R2
peak. At 410 nm, MB• absorbs strongly.30 Stopped-flow-loaded reactant
solutions were repeat triggered 4 times and averaging procedures carried
out.
4.23H2O f e-aq (2.8), OH• (2.8), H• (0.62), H+ (2.8), H2O2 (0.73)
(5)
where 107 G values in brackets correspond to the number of moles of
product per joule of energy absorbed.33 Solutions for kinetic studies
were at pH 7.0 (40.5 mM phosphate), contained 0.010 M sodium
formate, and were saturated with N2O, I ) 0.100 M. Subsequent
reactions are
Treatment of Data. In pulse radiolysis studies the program
FACSIMILE34 was used to fit UV-vis absorbance-time data. In the
stopped-flow studies the Applied Photophysics global analysis program
Glint (version 4.10) was used.
e-aq + N2O + H2O f N2 + OH- + OH•
OH•/H• + HCO2- f CO2•- + H2O/H2
CO2•- + X f CO2 + X•-
(6)
(7)
(8)
Results
•-
Initial Studies of R2 with CO2 and e-aq. A solution of
R2 (7.3µM) at pH 7.0 (40.5 mM phosphate), 0.010 M formate,
I ) 0.100 M, was subjected to a dose of 1 Gy of radiation,
where in the present case X•- is the OR and X the parent. A large
excess of X (∼1.0 × 10-4M) was used so that the formate radical CO2•-
is effectively scavenged, generating X•- as the only reducing species
present in solution. With such a choice of reactant concentrations double
•-
which gave 0.60 µM CO2 (reduction potential -1.9 V). No
reaction was observed at 410 nm over 10 µs to 10 s at 22 ( 1
•-
°C. At 350 nm rapid formation and decay of CO2 could be
reduction of the parent is precluded. The addition of 2X•- to X2 or
2-
monitored (ꢀ ) 250 M-1 cm-1) with rate constant 2k(CO2
+
•-
disproportionation to X and X2- also needs to be considered. Experi-
ments were at temperatures in the range 22 ( 1 °C.
CO2•-) ) 1.0 × 109 M-1 s-1 for the decay process. On repeating
with no phosphate or protein present a rate constant of 0.91 ×
109 M-1 s-1 was obtained as compared to a literature value of
Samples of R2 (∼100 µM) were dialyzed against two portions of
deaerated buffer (each g200 times the volume of R2 solution) and
stored under N2 before use. Prior to each experiment buffer solutions
containing 100 µM of the parent X (∼10 mL) were saturated with N2O
and argon by bubbling for at least 20 min. Calculated amounts of stock
deaerated protein were added to give the required concentration (∼10
µM). Gastight syringes were essential when loading the sample cell,
and a positive pressure of argon was maintained while loading and
draining the cell.
1.3 × 109 M-1 s-1 35
.
To generate e- (-2.9 V) a solution of R2 (5.1 µM) at pH
aq
7.0 (45 mM phosphate) containing tert-butyl alcohol (50 mM),
I ) 0.100 M, was subjected to a dose of 1 Gy of radiation,
which gave 0.30 µM of e-aq. At the 410 nm R2 peak no reaction
was observed over 10 µs to 10 s. Decay of the e-aq absorbance
at 650 nm (ꢀ ) 16.8 × 103 M-1 cm-1 33
)
was approximately
(28) Zare, R. N. Ber. Bunsen-Ges. Chem. 1974, 78, 153.
(29) Bird, C. L.; Kuhn, A. T. Chem. Soc. ReV. 1981, 10, 49-82.
(30) Keene, J. P.; Land, E. J.; Swallow, A. J., Ed. Baxendale J.H
Academic Press: London, 1965; p 225.
first order with a rate constant of 5.2 × 105s-1. This can be
largely accounted for by the reaction of e- with phosphate
aq
buffer. The rate constant for the reaction of e-aq with R2 would
(31) Clark, W. M. Oxidation and Reduction Potentials of Organic
Systems; Williams and Wilkins: Baltimore, Maryland, 1960.
(32) Steihler, R. D.; Chen, T.-T.; Clark, W. M. J. Am. Chem. Soc. 1933,
55, 891.
(34) Curtis, A. R.; Sweetenham, W. P. FACSIMILE/CHECKMAT Users
Manual, UKAEA, Atomic Energy Research Establishment, Harwell, 1987,
R12805.
(33) Salmon, G. A.; Sykes, A. G. Methods Enzymol. 1993, 227, 522.
(35) Sellers, R. M. Ph.D. Thesis, University of Leeds, 1972.

Reduction of E. coli Ribonucleotide Reductase
J. Am. Chem. Soc., Vol. 122, No. 10, 2000 2209
have to be >2 × 1010 M-1 s-1 to contribute >20% to the
changes observed.
•-
It is concluded that there is no measurable reaction of CO2
or e- with active- or met-R2.
aq
Rate Constants for the Disproportionation of OR’s. Rate
constants kD(22 °C) for the disproportionation of OR (here X•-)
2X•- f X + X2- (or X2
)
(9)
2-
where X is the parent form, were determined by monitoring
the absorbance decay at the OR peak positions Pf• (670 nm),
Rb•- (600 nm), and Rf•2- (700 nm). The PR dose gave OR
concentrations of ∼0.6 µM. Parent concentrations were 100 µM,
pH 7.0 (40.5 mM phosphate buffer), with 0.010 M sodium
formate present, I ) 0.100 M. Rate constants kD/M-1 s-1
obtained were 4.4(8) × 109 (Pf•), 1.9(8) × 109 (Rb•-), and 3.0(3)
× 108 (Rf•2-). Flash photolysis studies involving Rb•- dispro-
portionation have previously been reported.35 The radical MV•+
is long-lived and reaction 9 does not occur. With BV•+
dimerization rather than disproportionation occurs,36,37 but is
negligible for the concentrations used in these studies.
Reactions of OR’s with R2 by Pulse Radiolysis. Absorption
changes were monitored at wavelengths for decay of the OR’s
(see Experimental Section). At these wavelengths absorptions
of the parent and R2 forms are small. The rate constant k0 for
Figure 2. Fit of absorbance-time data from pulse radiolysis studies
to eqs 11-13 (as appropriate) for reactions of the radical Pf• (100 µM)
with E. coli R2 monitored at 670 nm: (A) the reaction of met-R2 (10
µM) and (B) the reaction of active-R2 (10 µM, 40% of which is met-
R2), using the FACSIMILE program. Conditions: 22 °C, pH 7.0 (40.5
mM phosphate), I ) 0.100 M.
(10) is >5 × 109 M-1 s-1
,
k0
9
CO2•- + X
8 CO2 + X•-
(10)
Table 1. Results from Pulse Radiolysis Experimentsa
OR
k1/M-1 s-1 b
k2/M-1 s-1 c
and is rapid compared to other steps (eqs 11-13).
MV•+
BV•+
Pf•
1.06(2) × 109
3.3(4) × 108
2.2(4) × 107
1.1(1) × 107
2.7(5) × 106
2.00(2) × 109
3.5(2) × 108
3.4(5) × 107
2.6(2) × 107
3.4(5) × 106
k1
X
•- + FeIII2‚‚‚Tyr• 98 X + FeIIFeIII‚‚‚Tyr•
(11)
(12)
(13)
Rb•-
Rf•2-
k2
X
•- + FeIII2 98 FeIIFeIII + X
a Rate constants (22 °C) for reactions of active-R2 (k1) and met-R2
(k2) with organic radicals (OR), pH 7.0 (40.5 mM phosphate), I ) 0.100
M. b Listed as k12 for active-R2 in Table 5. c Listed as k12 for met-R2
in Table 5.
kD
2X•- 98 X22- or X + X2-
Table 2. Absorption Coefficients (ꢀ) for the Parent, Organic
Radical (OR), and FeIIFeIII (R2) Intermediate (not previously
determined) from Pulse Radiolysis Experiments, Using the
FACSIMILE Program and Monitoring the Decay of the UV-Vis
Absorbance of the OR (wavelength indicated)a
Satisfactory fits to absorbance vs time data were obtained using
the FACSIMILE program.34 Solutions of met-R2 allowed k2 in
(12) to be determined by separate experiment. Solutions of
active-R2 have a met-R2 component (∼40%), and concurrent
reactions have to be considered. In the case of the MV•+ and
BV•+ (11) and (12) are favorable compared to (13),37,38 and k1
and k2 were obtained from first-order decay processes. For Pf•,
Rb•-, and Rf•2-, k1, k2, and kD are relevant, where kD and k2 are
as determined in separate experiments. Examples of fits of data
are shown for Pf• in Figure 2 and for Rb•- and Rf2- as
Supporting Information. No further steps corresponding to the
further reaction of X2- were required in the fitting process.
Values of k1 and k2 are listed in Table 1. The similarity of k1
and k2 values (k2 > k1 by an average 66%) implies an identical
λ/
nm
ꢀ(parent)/
ꢀ(OR)/
ꢀ(FeIIFeIII)/
OR
M-1 cm-1
M-1 cm-1
M-1 cm-1
Pf
670
550
600
700
950(80)
∼10
10440(300)
5390(240)
3730(90)
1310(150)
1500(100)
1090(30)
Rb•-
∼10
Rf•2-
400(60)
3550(260)
1010(100)
a pH 7.0 (40.5 mM phosphate), I ) 0.100 M.
Pf•, Rb•, and Rf•2-, and for FeIIFeIII obtained as part of the fitting
procedure, are listed. The ꢀ values for FeIIFeIII are small
compared to those for the OR’s. No major peak positions in
the 550-700 nm range were indicated, and further details were
not explored as a part of these studies. The reaction of MB•
was also studied by PR, but rates are too slow and the fit is
less precise than when the SF method is used.
redox step FeIII f FeIIFeIII. This is an unexpected result in
2
view of the reduction potential for the Tyr•/Tyr couple currently
believed to be ∼1.0 V.39 Details of the FeIIFeIII decay have not
been established, but the reaction is known to be rapid from
EPR freeze-quench experiments.40,41 In Table 2, ꢀ values for
(36) Ahmed, I.; Cusanovich, M. A.; Tollin, G. Proc. Natl. Acad. Sci.
U.S.A. 1981, 78, 6724.
(37) Van Leeuwen, J. W.; van Dijk, C.; Veeger, C. Eur. J. Biochem.
1983, 135, 601.
Reactions of OR’s with R2 by Stopped-Flow. The dithionite
procedure for the reduction of the parent X gives double-reduced
X2- forms of MB+ and TB+, and single reduced X•- in the
(38) Carey, J. G.; Cairns, J. F.; Colchester, J. E. J. Chem. Soc., Chem.
Commun. 1969, 1280.
(39) Silva, K. E.; Elgren, T. E.; Que, L., Jr.; Stankovich, M. T.
Biochemistry 1995, 34, 14093.
(40) Atta, M.; Anderson, K. K.; Ingemarson, R.; Thelander, L.; Gra¨slund,
A. J. Am. Chem. Soc. 1994, 116, 6429.
(41) Gerez, C.; Fontecave, M. Biochemistry 1992, 31, 780.

2210 J. Am. Chem. Soc., Vol. 122, No. 10, 2000
Dobbing et al.
Table 3. Stopped-Flow Studies (22 °C) on the Reduction of the
Active-R2 Component FeIII2‚‚‚Tyr• (46.7 µM) with Toluidine Blue
(4.4 µM)a
Table 4. Rate Constants from Stopped-Flow Studies on the
Reduction of the Active-R2 Component (FeIII2‚‚‚Tyr.) with Organic
Radicalsa
10-9kD(start)
10-5k1
10-5k3
10-9kD(final)
[parent]/
µM
[active-R2]/
10-5k1 /
10-5k3/
b
b
λ/nm
µM
M-1 s-1
M-1 s-1
0.01
0.30
1.5c
37(6)
7.1(3)
6.9(3)
3.2(7)
5.6(2)
5.7(2)
0.057(6)
1.2(3)
1.5d
reaction of methylene blue
660
660
660
4.5
7.4
9.9
30.8
30.8
46.7
7.5(2)
7.2(3)
7.3(3)
6.0(1)
5.8(1)
6.1(1)
a pH 7.0 (45 mM phosphate), I ) 0.100 M, monitored at 632 nm.
Solutions of the parent X were reduced with dithionite prior to stopped-
flow mixing. The program Glint was used to fit rate constants k1
(reaction of X•-) and k3 (reaction of X2-), using different starting values
for kD (examples given). A best fit value for kD is indicated. b Listed
as k12 in Table 5. c Best fit kD. d Does not vary during fit.
reaction of toluidine blue
46.7 6.9(3)
5.9(2)
6.32
410
632
4.4
7.0(2)
6.6
5.7(2)
30.8
30.8
6.8(2)
6.8(3)
5.8(1)
6.0(1)
632
9.1
case of IP-. Experiments on the reaction of dithionite reduced
MB+ and IP- (2.8µM) with met-R2 (17.5µM) were explored
(under N2) using a two compartment (split) optical cell. The
latter (total solution volume 2.6 mL) was loaded and initial
absorbance readings taken prior to mixing. No reaction was
observed over >2 h indicating a reduction potential for the met-
R2 FeIII2/FeIIIFeII couple of <0.011 V. This value is in keeping
with the two-equivalent reduction potential of -0.115 V for
the FeIII2/FeII2 couple.39 In the case of mouse R2 the reduction
potential for the FeIII2/FeIIFeIII couple has been estimated as
0.150 to 0.010 V,40 and if such a range holds also for E. coli
met-R2 then it is expected that some of the OR radicals will
have difficulty in bringing about reduction of met-R2. Reactions
with active-R2 could, however, be studied because the unfavor-
able reduction of FeIII2‚‚‚Tyr• to FeIIFeIII‚‚‚Tyr• is followed by
the favorable intramolecular redox change to give met-R2.
The reaction scheme to be considered is therefore
reaction of indo-phenolc
603
603
410
604
4.1
5.8
0.10(2)
8.1
30.8
30.8
0.11(1)
0.11(1)
46.7
0.13(1)
a Solutions of the parent X were reduced with dithionite prior to
stopped-flow studies. The program Glint was used to fit rate constants
for active-R2 with X• (k1) and doubly reduced X2- (k3), with kD fixed
.
at 1.5 × 109 M-1 s-1 b Listed as k12 in Table 5. c k3 (and kD) not
relevant.
Table 5. Summary of Rate Constants k12 (22 °C) at pH 7.0 for the
Reduction of Active- and Met-R2 from E. coli Ribonucleotide
Reductase with Organic Radicals (OR) Determined by Pulse
Radiolysis (PR) and Stopped-Flow (SF) Spectrophotometrya
OR
method (λ/nm)
Eo /V
log k12
-log f
1
active-R2 as oxidant
MV•+
BV•+
Pf•
PR(605)
-0.446
-0.355
-0.243
-0.200
-0.059
+0.006
+0.015
+0.194
9.04
8.51
7.26
7.04
6.43
5.86b
5.83c
4.04d
2.07
1.56
1.00
0.76
PR(555)
PR(670)
PR(550)
PR(700)
SF(660)
SF(632)
SF(603)
k3
Rb•-
Rf2-
MB•
TB•
X
2- + FeIII2‚‚‚Tyr• 98 X•- + FeIIFeIII‚‚‚Tyr• (14)
0.25
0.115
0.115
0.070
kD
2X•- 98 X2- + X
(15)
(16)
IP•2-
met-R2 as oxidant
-0.446
k1
•- + FeIII2‚‚‚Tyr• 98 FeIIFeIII‚‚‚Tyr•
MV•+
BV•+
Pf•
PR(606)
PR(555)
PR(670)
PR(550)
PR(700)
9.28
8.53
7.52
7.41
6.54
2.07
1.56
1.00
0.76
0.25
X
-0.355
-0.243
Concentrations of the parent were in the range 4.1-9.9 µM,
and of active-R2 in the range 30-47 µM. A fitting procedure
for toluidine blue was employed using absorbance-time data
at 632 nm, and the Glint program with different kD(start) values.
As can be seen from Table 3, k1 and k3 converge and give a
kD(final) value of 1.5 × 109 M-1 s-1 after a number of iterations.
The same kD was used for methylene blue, where values of kD
in the range (1.5-3.0) × 109 M-1 s-1 have been determined
previously.30 No double reduction is observed with indo-phenol
and a simpler treatment was possible. The TB+ reaction was
also monitored at 410 nm. The latter is an absorbance peak for
Rb•-
Rf•2-
-0.200
-0.059
a Reduction potentials Eo vs NHE are for OR’s. Values of log f
1
were obtained from eq 3. b Less precise k12 value obtained by PR, 12.9
× 105 M-1 s-1
.
c At 410 nm, k12 ) 7.0 × 105 M-1 s-1
.
d At 410 nm,
k12 ) 1.0 × 104 M-1 s-1
.
k12 - 0.5 log f against -Eo /0.059, Figure 4, gives a single line
1
for all the data of slope 0.54(2) in satisfactory agreement with
the theoretical value of 0.50, and an intercept of 5.94(9) equal
to 0.50 (log k11 + log k22 + Eo /0.059). Separate consideration
2
of the active-R2 points has negligible (∼2%) effect on slope
and intercept values. Assuming Eo2 for the E. coli met-R2 couple
FeIII2/FeIIFeIII to be close to zero (see above) and k11 for the
Tyr• with smaller absorbance contributions from FeIII 26 As can
.
2
be seen from Table 4 identical rate constants are obtained at
both wavelengths. The absorbance change at 410 nm was 0.016,
which is bigger than the absorbance change for FeIII f FeII
OR radicals to be 1.0 × 106 M-1 s-1 19
, the self-exchange rate
2
2
constant k22 for the same couple is 7.6 × 105 M-1 s-1
.
(0.009). On this evidence reduction of the Tyr• occurs rapidly
and within the first phase of reaction. The IP- reaction was
also monitored at 410 nm, Table 4. Since MB• absorbs strongly
at 410 nm, no similar studies were possible. Values of k1 are
listed in Table 5 as k12.
Discussion
With five OR’s as reductants it was possible to determine
rate constants for the reduction of active- and met-R2 forms by
PR, Table 5. The coincidence of k12 values for the two forms,
Figure 4, is significant since it indicates one-electron reduction
Free-Energy Plot. Rate constants (k12) at pH 7.0 for the eight
OR reactions with active-R2 and for five OR reactions with
met-R2 are listed in Table 5. In those cases in which both have
been studied the met-R2 rate constant is the larger, but only by
a small amount (average ∼66%). This is most likely due to
minor structural differences at the active site. A graph of log
of FeIII as rate limiting in both cases, with the Tyr• of active-
2
R2 playing a passive role. Subsequently, in the case of active-
R2, the Tyr• is reduced in an intramolecular step. Three other
less reactive OR’s (MB•, TB•, and IP•2-) were studied by the

Reduction of E. coli Ribonucleotide Reductase
J. Am. Chem. Soc., Vol. 122, No. 10, 2000 2211
Figure 5. Pathway proposed for transfer of an electron from the indole
ring of Trp-48 to His-118. A coordinated H2O (for mouse45) or
Asp-844 series as a contact to the phenolate O-atom of Tyr•.
FeIII2/FeIIFeIII couple is between -0.051 and +0.011 V. The
latter range is in keeping with an estimate of the mouse FeIII
/
2
FeIIFeIII couple of 0.010 to 0.150 V.40 Since there is no reduction
of met-R2 with MB•, TB•, and IP•2-, the mechanism of reduction
of active-R2 has to be carefully considered. Formation of FeII-
FeIII followed by rapid reduction of Tyr• is able to explain the
behavior observed. The latter does however need to be
significantly faster than the reverse step involving FeIIFeIII and
the parent form.
Figure 3. Fit of absorbance changes at (A) 632 and (B) 410 nm for
the stopped-flow reduction of the E. coli active-R2 component (46.7
µM) with toluidine blue (4.9 µM), prior reduced with dithionite.
Conditions: 22 °C, pH 7.0 (45 mM phosphate), I ) 0.100 M.
The single linear Marcus correlation of rate constants for
active- and met-R2, Figure 4, is in accordance with (4), and
indicates a common rate-limiting electron-transfer (ET) process
occurring between the OR (donor) and FeIII2 (acceptor) centers.
However, the Tyr• is a much stronger oxidant (∼1.0 V),39 and
the exclusiveness of the FeIII2 step, with reduction of active and
met-R2 occurring at similar rates, is an interesting outcome. A
pathway for ET from Trp-48 at the surface of R2 via Asp-237
and His-118 (which is a ligand to FeA) has been proposed.9
These amino acids are conserved,5,41,42 and are connected by
hydrogen bonds, Figure 5.42,43 Variant E. coli and mouse R2
forms in which amino acids in the ET pathway are replaced by
residues which are unable to H-bond severely impair redox
activity.42,43 Evidence has been presented that this pathway
mediates electron transfer during O2 activation.44 Reduction of
Tyr• may require a proton as well as an electron, in which case
proton transfer involving the H-bonded chain between the donor
and acceptor sites may be relevant by adaptation of Figure 5.
Alternatively, the electron and proton may have other means
of accessing the active site. For conversion through to red-R2
three electrons and three protons are required to retain electro-
neutrality at the active center.45,46 Theoretical calculations have
indicated unfavorable energetics for ET without the accompany-
ing proton transfer to maintain electroneutrality.46 Possible
pathways for ET can be examined using the criteria of Beratan
and Onuchic, and may involve atoms on the peptide backbone
as well as the H-bonded route.47,48 To initiate H+ transfer in
the present case a proton would have to be provided by solvent
Figure 4. Marcus free-energy plot for reduction of active-R2 (2) and
met-R2 (4) from E. coli RNR by eight organic radicals (OR) as listed
in Table 5. No reaction of met-R2 with MB•, TB•, and IP•2- is observed.
SF method, when only k12 values for reduction of active-R2
were obtained. With TB• and IP•2- the R2 absorbance decay at
410 nm gives identical rate constants to those obtained by
monitoring re-formation of the parent, Figure 3. Absorption
changes at 410 nm are bigger than those observed for the
(42) Persson, B. O.; Karlsson, M.; Climent, I.; Ling, J. S.; Loehr, J. S.;
Sahlin, M.; Sjo¨berg, B.-M. J. Biol. Inorg. Chem. 1996, 1, 247.
(43) Katterle, B.; Sahlin, M.; Schmidt, P. P.; Po¨tsch, S.; Logan, D. T.;
Gra¨slund, A.; Sjo¨berg, B.-M. J. Biol. Chem. 1997, 272, 10414.
(44) Parkin, S. E.; Chen, S.-X.; Ley, B. A.; Mangravite, L.; Edmondson,
D. E.; Huynh, B. H.; Bollinger, J. M. Biochemistry 1998, 37, 1124-1130.
(45) Schmidt, P. P.; Rova, U.; Katterle, B.; Thelander, L.; Gra¨slund, A.
J. Biol. Chem. 1998, 273, 21463.
(46) Siegbahn, P. E. M.; Blomberg, M. R. A.; Crabtree, R. H. Theor.
Chem. Acc. 1997, 97, 289.
(47) Beratan, D. N.; Onuchic, J. N. Photosynth. Res. 1989, 22, 173.
(48) Beratan, D. N.; Onuchic, J. N.; Betts, J. N.; Bowler, B. E.; Gray,
H. B. J. Am. Chem. Soc. 1990, 112, 7915.
reduction of FeIII alone, and the differences are assigned to
2
the decay of Tyr•. On this evidence it would appear that the
rate of the Tyr• oxidation of FeIIFeIII is comparable to the initial
reduction of FeIII
.
2
No reduction of met-R2 is observed for the radicals MB•,
TB•, or IP•2-, which have reduction potentials g0.011V, Table
5. Reduction of met-R2 with Rf•2- (-0.051 V) is observed,
however. It is concluded that the reduction potential for the

2212 J. Am. Chem. Soc., Vol. 122, No. 10, 2000
Dobbing et al.
H2O rather than the reductant. Electron-transfer from the
356, Tyr-731, and Tyr-730 to Cys-438 at the substrate binding
site of R1. The favorable self-exchange process k22 is relevant
as a part of the communication between R1 and R2.
reductant OR to the FeIII site may drive the protonation step
2
as in a similar instance involving redox coupled long-range
proton transfer to the [3Fe-4S] cluster of Azotobactor Vinelan-
dii.49 Interestingly, the Fe-S cluster is also buried and is
inaccessible to solvent water.50 A proton would thus travel to
the binuclear iron acceptor site in what has been described as
the “slipstream” of the electron.49 Compliance with Marcus
theory suggests that the rate controlling process is ET, and the
similarity of rate constants for active and met-R2 indicates that
ET terminates initially at FeA. In the case of E. coli R2, the
chelated carboxylate of Asp-84 has been suggested as a possible
mediator for proton transfer from the binuclear Fe to Tyr-122.14
To summarize rate constants (k12) for eight OR’s, reduction
potentials (Eo ) in the range -0.446 to +0.194 V, as reductants
1
for active- and met-R2, have been determined by PR and SF
techniques. Rate constants for any one OR with active- and met-
R2 forms are very similar (met on average 66% bigger), and
reduction of the FeIII (Eo close to zero) is rate controlling. A
2
2
single linear free-energy plot of log k12 vs Eo /0.059 conforms
1
to the Marcus equations, and ET is therefore the rate-controlling
step. Reduction of the more strongly oxidizing adjacent Tyr•
(reduction potential ∼1.0 V), in an overall process Tyr• + e-
+ H+ f TyrH, would serve to maintain electroneutrality of
the buried active site. A reasonable hypothesis, but one which
requires further examination, is that the electron and proton-
transfer steps may occur via the conserved pathway from the
surface Trp-48 to Asp-237, His-118, and FeA.
The intercept at -Eo /0.059 ) 0 in Figure 4 of 5.94 is equal
1
to 0.5 (log k11 + log k22 + Eo /0.059), (4). The OR’s are
2
assumed to have a common self-exchange rate constant k11 of
1.0 × 106 M-1 s-1, which gives an excellent correlation with
cytochrome c(III) as oxidant.19 As already discussed the
reduction potential Eo2 for the FeIII2/FeIIFeIII couple has a value
close to zero. Therefore the self-exchange rate constant k22 for
the FeIII2/FeIIFeIII exchange process is 7.6 × 105 M-1 s-1, which
is surprisingly favorable for a 10.9 Å (×2) separation of the
Acknowledgment. We are grateful for contributions to
earlier parts of this work particularly from Drs. J.-Y Han and
J.C. Swarts. Financial support was from the North of England
Cancer Campaign, and UK Engineering and Physical Sciences
Research Council. We also wish to acknowledge discussions
with Professors A. Gra¨slund and J. M. Bollinger.
binuclear Fe centers of two R2’s. Larger Eo values may be
1
considered with k22 decreasing by an order of magnitude for
each 0.059 V. The pathway for reduction of active- and met-
R2 is only a part of the total enzyme requirement. When R1
and R2 associate, the pathway extends from Trp-48 via Tyr-
Supporting Information Available: Figures for fits of
absorbance vs time data in PR experiments (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
(49) Hirst, J.; Duff, J. L. C.; Jameson, G. N. L.; Kemper, M. A.; Burgess,
B. K.; Armstrong, F. A. J. Am. Chem. Soc. 1998, 120, 7085.
(50) It has also been proposed49 that reoxidation of the [3Fe-4S]0H+ is
controlled by the rate of release of the proton.
JA993412K