Low-temperature turnover control of photosystem II using novel metal- containing redox-active herbicides

Article abstract of DOI:10.1021/ja994138x

A novel approach of using metal-containing redox-active herbicides to prepare and study the light-induced intermediates of the photosystem II (PSII) photocycle is described. The redox-active herbicides feature an iron(III) ethylenediaminetetracetate [Fe(I

Full text of DOI:10.1021/ja994138x

5180
J. Am. Chem. Soc. 2000, 122, 5180-5188
Low-Temperature Turnover Control of Photosystem II Using Novel
Metal-Containing Redox-Active Herbicides
Laba Karki, K. V. Lakshmi, Veronika A. Szalai,† and Gary W. Brudvig*
Contribution from the Department of Chemistry, Yale UniVersity, P.O. Box 208107,
New HaVen, Connecticut 06520-8107
ReceiVed NoVember 29, 1999
Abstract: A novel approach of using metal-containing redox-active herbicides to prepare and study the light-
induced intermediates of the photosystem II (PSII) photocycle is described. The redox-active herbicides feature
an iron(III) ethylenediaminetetracetate [FeIII-(EDTA)] electron-acceptor group linked to a QB-site binding
dimethylphenylurea moiety by a hydrocarbon spacer. Like the nitroxyl-based redox-active herbicides previously
described (Bocarsly, J. R.; Brudvig, G. W. J. Am. Chem. Soc. 1992, 114, 9762-9767), metal-containing
herbicides accept electrons from the donor side of PSII while bound to the QB site and restrict the S-state
cycling to two stable charge separations. The use of FeIII-(EDTA) as an electron acceptor allows turnover at
low temperatures. EPR studies of PSII upon continuous illumination at 225 K with 0.7 mM of redox-active
herbicide, FeIII-(EDTA) linked by an ethane spacer to a dimethylphenyl urea group (4), produced a stable
two-step S1 to S3 advance of the O2-evolving complex (OEC) and a stoichiometric reduction of the FeIII-
(EDTA) moiety of the herbicide, while a control sample with 0.02 mM DCMU [3-(3,4-dichlorophenyl)-1,1-
dimethyl-urea] and 0.7 mM of 4 exhibited only a one-step S1 to S2 advance of the OEC without significant
reduction of the FeIII-(EDTA) moiety of the herbicide. Similar EPR results were obtained for 7, FeIII-(EDTA)
linked to the dimethylphenylurea group by a pentane spacer. O2-evolution inhibition studies show that appending
the FeIII-(EDTA) moiety to the phenylurea herbicide causes a significant decrease in the binding affinity
compared to that of DCMU. On the basis of O2-evolution studies with various herbicide derivatives and different
PSII sample types, the observed decrease in binding affinities is attributed to the degree of accessibility of the
QB-binding pocket to the herbicides and to electrostatic and hydrophilicity factors. The present study describes
the use of novel metal-containing herbicides in studying long-range electron transfer in PSII and in trapping
photogenerated two-electron oxidized intermediate states of the O2-evolving complex.
Introduction
Scheme 1
Photosystem II (PSII), a multicomponent plant membrane
protein complex, harvests light energy to couple the oxidation
of water to dioxygen with the concomitant reduction of
plastoquinone.1 Absorption of photons by the primary chloro-
phyll electron donor, P680, initiates electron transfer from P680
through a pheophytin molecule (PheoA) to a primary quinone
electron acceptor (QA) and finally to a reversibly bound
plastoquinone (PQ) at the QB-binding pocket.2 After the second
photoinduced electron-transfer step, the plastoquinone becomes
fully reduced to PQH2 and leaves the binding pocket, creating
a two-electron gate to electron transfer. While the plastoquinone
accepts electrons from the excited P680* molecule, the O2-
evolving complex (OEC), consisting of a tetranuclear manganese
higher oxidation Si states (i ) 0-4) by successive photoinduced
electron-transfer steps and releases a molecule of dioxygen in
the S4 to S0 turnover (Scheme 1).1,4
Since the elucidation of the S-state cycle, spectroscopic
studies have been carried out to gain structural information on
each of the S states to understand their role in the water-
oxidation chemistry. However, complete characterization of the
higher oxidation S3 state has remained a challenge because of
its limited stability and the difficulty in isolating this state in
high yield. Although flash-induced methods5 have produced
∼50% yield of the S3 state in concentrated samples for
spectroscopic studies, characterization by parallel mode EPR6
and EXAFS7 has yielded only preliminary information. Besides
flash-induced methods, redox-active herbicide inhibitors such
+
cluster (Mn4), donates reducing equivalents of electrons to P680
via a tyrosine residue called TyrZ (Yz).3 According to the Kok
model,4 the tetranuclear manganese cluster (Mn4) accesses
* To whom correspondence should be addressed. Telephone: 203-432-
5202. Fax: 203-432-6144. E-mail: gary.brudvig@yale.edu.
† Current address: Department of Chemistry, University of North
Carolina at Chapel Hill, NC 27599-3290.
(1) Debus, R. J. Biochim. Biophys. Acta 1992, 1102, 269-352.
(2) Diner, B. A.; Babcock, G. T. In Oxygenic Photosynthesis: The light
reactions; Ort, D. R., Yocum, C. F., Eds.; Kluwer Academic Publishers:
Dordrecht, The Netherlands, 1996; Vol. 4, pp 213-247.
(3) (a) Debus, R. J.; Barry, B. A.; Sithole, I.; Babcock, G. T.; McIntosh,
L. Biochemistry 1988, 27, 9071-9074. (b) Metz, J. G.; Nixon, P. J.; Ro¨gner,
M.; Brudvig, G. W.; Diner, B. A. Biochemistry 1989, 28, 6960-6969.
(4) Kok, B.; Forbush, B.; McGloin, M. Photochem. Photobiol. 1970,
11, 457-475.
(5) Styring, S.; Rutherford, A. W. Biochemistry 1987, 26, 2401-2405.
(6) Matsukawa, T.; Mino, H.; Yoneda, D.; Kawamori, A. Biochemistry
1999, 38, 4072-4077.
10.1021/ja994138x CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/12/2000

TurnoVer Control of Photosystem II
J. Am. Chem. Soc., Vol. 122, No. 21, 2000 5181
dicyclohexylcarbodiimide (DCC, Acros Organics) and N-hydroxysuc-
cinimide (NHS, Pierce Chemical Co.) were also used as received.
n-dodecyl-â-D-maltoside was purchased from Anatrace and used as
received. Thin-layer chromatography (TLC) was performed on polymer
backed silica gel plates (60 Å, Aldrich). Column chromatography was
performed on silica gel 200-400 mesh (60 Å, Aldrich). High
performance liquid chromatography (HPLC) was performed with a
computer-controlled system (Rainin, model Dynamax SD-200) with a
diode array detector on a Vydac C-18 (1 cm × 25 cm) reverse phase
column. Visible spectra were measured on a Perkin-Elmer Lambda 3B
spectrophotometer. NMR spectra were recorded on an Oxford QE-Plus
300 MHz spectrometer.
Scheme 2
as dimethylureido phenyl phenyl nitroxide (PNPDU, Scheme
2), have also been used to generate the S3 state in 75% yield
(Scheme 3).8 The ability of the redox-active herbicides to bind
tightly to the QB-binding site and accept one electron from QA
enables the formation of a two-electron charge-separated state
in high yield. Although PNPDU accepts an electron in advancing
the Mn4 cluster from the S2 to S3 state, its applicability is limited
to temperatures of g250 K owing mainly to the proton-coupling
requirement for electron transfer. One approach to produce a
double turnover at temperatures lower than 250 K is by using
metal-containing redox-active herbicides that do not involve
proton-coupling in the electron-transfer step. Herbicides that can
accept electrons at lower temperatures would allow for new
possibilities to trap and study photogenerated intermediate states
such as the S2Yz• state that has previously been studied only in
inhibited samples.9
In light of the ongoing studies to unravel structural informa-
tion about the S3 and possibly the S2Yz• states, we demonstrate
the versatility of using novel metal-containing redox-active
herbicides with an FeIII-(EDTA) electron-acceptor group linked
by a hydrocarbon spacer to a phenyldimethyl urea herbicide
for studying photochemical turnover of PSII at temperatures
below 250 K. Metal-based acceptors provide additional struc-
tural flexibility in studying the mechanisms, energetics and rates
of long-range electron transfer to the QB site. The work reported
herein also demonstrates the use of a covalent tether between
an active-site binding moiety and a redox-active moiety for
controlled transfer of electrons out of a protein active site. In a
related study, Wilker et al.10 have recently demonstrated the
design and use of photosensitizers linked by hydrocarbon spacers
to binding groups that have high affinity for the cytochrome
P450cam heme pocket. By delivering holes or electrons directly
to the active site, these charge-transfer complexes allowed the
preparation and study of redox intermediates of cytochrome
P450. We demonstrate here the use of metal complexes linked
by hydrocarbon spacers to a phenyl urea herbicide with an
affinity for the QB site to prepare and study transient states in
the PSII photocycle.
Synthesis of BOC-Protected Ethylenediamine, BOC-EN11 (1). To
a warm (45 °C) solution of ethylenediamine (34 mL, 0.51 mol) in 60
mL of THF, BOC-ON (25.0 g, 0.102 mol) in 80 mL of THF was added
dropwise continuously. The solution was stirred for 20 min, and a
yellow product was recovered after removing excess ethylenediamine
1
and THF in vacuo. Yield: 70%. H NMR in CDCl3 (ppm): δ 1.4 (9
H, s, BOC group); δ 3.0 (2 H, m, CH2); δ 3.2 (2 H, m, CH2); δ 4.0 (2
H, broad, NH2); δ 5.0 (1 H, t, N-H). Impurities due to the phenyl
groups of the BOC-ON were observed at δ 7.4 (2 H, m, aromatic); δ
7.8 (2 H, m, aromatic).
Synthesis of BOC-EN-SO2R (2).11,12 To a solution of 1,1-dimethyl-
3-[4-(chlorosulfonyl)phenyl]urea [RSO2Cl]8 (4.0 g, (15 mmol) in 150
mL of THF), 2.43 g (15 mmol) of 1 in 100 mL of THF was added
dropwise over 1 h with continuous stirring. After adding 20 mL of
water to this mixture, the solution was refluxed at 60-70 °C. After 2
h, the solution was allowed to cool to room temperature. The organic
layer was separated and washed with 2 × 5 mL portions of saturated
sodium bicarbonate (NaHCO3) followed by 2 × 5 mL portions of
saturated NaCl solution. The organic layer was dried over anhydrous
MgSO4, filtered, and concentrated under reduced pressure affording a
yellowish, sticky product. Performance of TLC with a solvent mixture
of (5:5:1) ethyl acetate:hexane:methanol resulted in three bands. The
slowest moving band was deduced to be the coupled product from a
comparison with the Rf values of the starting materials. Purification
was accomplished with silica gel column chromatography using a (4
× 30 cm) column, and eluting with (5:5:1) ethyl acetate:hexane:
methanol solvent mixture. The BOC-protected product eluted as the
final band. Yield: ∼60%. TLC: Rf ) 0.23 on SiO2 with hexane:ethyl
1
acetate:methanol (5:5:1). H NMR in (CD3)2O (ppm): δ 1.4 (9 H, s,
BOC group); δ 2.9 (2 H, m, ethyl); δ 3.2 (2 H, m, ethyl); δ 3.0 (6 H,
s, N(CH3)2); δ 6.1 (1 H, s, N-H); δ 6.4 (1 H, t, N-H); δ 7.8 (4 H, q,
aromatic); δ 8.2 (1 H, s, N-H). Anal. Calcd. for C16H26N4O5S1: C,
49.7; H, 6.78; N, 14.5. Found: C, 48.5; H, 6.42; N, 14.1.
Deprotection of the BOC Group, NH2-EN-SO2R (3). 0.5 g of 2
was dissolved in 5 mL TFA/30 mL CH2Cl2 and stirred under nitrogen
for several hrs. The solvent was removed under reduced pressure
affording a yellow sticky residue that was recrystallized from methanol/
ether solution yielding a white microcrystalline product. Yield: ∼90%.
1H NMR in (CD3)2O (ppm): δ 2.5 (2 H, d, NH2); δ 3.0 (6 H, s,
N(CH3)2); δ 3.25 (2 H, m, ethyl); δ 3.9 (2 H, m, ethyl); δ 7.7 (4 H, s,
aromatic); δ 7.9 (1 H, t, N-H); δ 8.2 (1 H, s, N-H). Anal. Calcd. for
C11H18N4O3S1: C, 46.13; H, 6.33; N,19.57. Found: C, 45.26; H, 6.44;
N, 19.04.
Synthesis of FeIII-(EDTA)-EN-SO2R (4)12 (Figure 1). 400 mg
(105.6 mmol) of FeIII-(EDTA)‚2H2O13 was dissolved in 10 mL of
anhydrous DMF and stirred at 0 °C under a dry nitrogen atmosphere.
After the solution turned clear, 0.216 g (105.6 mmol) of DCC and 0.060
g (105.6 mmol) of NHS were added and stirred. After 10 min, 0.15 g
(52.8 mmol) of 3 was added followed by a few drops of triethylamine
(base) while the reaction mixture was allowed to warm to room
Experimental Section
Materials and Instruments. All reagent grade chemicals and
solvents were purchased from commercial vendors. Iron(III) nitrate [Fe-
(NO3)3‚9H2O], BOC-ON [2-(tert-butoxycarbonyloxyimino)-2-phenyl-
aceto-nitrile] and ethylenediamine were purchased from Aldrich and
used as received. Other chemicals including mono-tert-butoxycarbonyl
1,5-diaminopentane toluenesulfonic acid salt (Nova BioChem), N,N′-
(7) Guiles, R. D.; Zimmerman, J.-L.; McDermott, A.; Yachandra, V. K.;
Cole, J. L.; Dexheimer, S. L.; Britt, R. D.; Wieghardt, K.; Bossek, U.; Sauer,
K.; Klein, M. P. Biochemistry 1990, 29, 471-475.
(8) Bocarsly, J. R.; Brudvig, G. W. J. Am. Chem. Soc. 1992, 114, 9762-
9767.
(9) (a) Szalai, V. A.; Ku¨hne, H.; Lakshmi, K. V.; Brudvig, G. W.
Biochemistry 1998, 37, 13594-13603. (b) Tang, X.-S.; Randall, D. W.;
Force, D. A.; Diner, B. A.; Britt, R. D. J. Am. Chem. Soc. 1996, 118, 7638-
7639. (c) Szalai, V. A.; Brudvig, G. A. Biochemistry 1996, 35, 1946-
1953. (d) Boussac, A.; Rutherford, A. W. Biochemistry 1988, 27, 3476-
3483. (e) MacLachlan, D. J.; Nugent, J. H. A. Biochemistry 1993, 32, 9772-
9780. (f) Lakshmi, K. V.; Eaton, S. S.; Eaton, G. R.; Frank, H. A.; Brudvig,
G. A. J. Phys. Chem. B 1998, 102, 8327-8335.
(11) Reifler, M. J. Ph.D. Thesis, Yale University, New Haven, 1998.
(12) Ermacora, R. M.; Delfino, J. M.; Cuenoud, B.; Schepartz, A.; Fox,
R. O. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 6383-6387. (b) Hertzberg,
R. P.; Dervan, P. J. Am. Chem. Soc. 1982, 104, 313-315. (c) Sluka, J. P.;
Griffin, J. H.; Mack, D. P.; Dervan, P. B. J. Am. Chem. Soc. 1990, 112,
6369-6374. (d) Ebright, Y. W.; Chen, Y.; Pendergrast, S. P.; Ebright, R.
H. Biochemistry 1992, 31, 10664-10670.
(13) Basallote, M. G.; Alcala, J. M. L.; Vizcaino, M. C. P.; Vilchez, F.
G. Inorg. Synth. 1986, 24, 207-208. (b) Hoard, J. L.; Kennard, C. H. L.;
Smith, G. S. Inorg. Chem. 1963, 2, 1316-1317.
(10) Wilker, J. J.; Dmochowski, I. J.; Dawson, J. H.; Winkler, J. R.;
Gray, H. B. Angew. Chem., Int. Ed. 1999, 38, 90-92.

5182 J. Am. Chem. Soc., Vol. 122, No. 21, 2000
Karki et al.
Scheme 3
by 2 × 5 mL of saturated NaCl solution. The organic layer was dried
over anhydrous MgSO4, filtered, and concentrated under reduced
pressure, yielding a pale yellow product. Performance of TLC in (5:
5:1) hexane:ethyl acetate:methanol showed two bands. From a com-
parison with the starting materials, the slower moving band was deduced
to be the BOC-coupled product. Purification of the coupled product
was accomplished with silica gel column chromatography using a (4
× 30 cm) column, and eluting with (5:5:1) ethyl acetate:hexane:
methanol solvent mixture. Yield ∼60%. TLC: Rf ) 0.27 on SiO2 (5:
5:1) ethyl acetate:hexane:methanol. 1H NMR in (CD3)2O (ppm): δ 1.4
(9 H, s, BOC); δ 1.5 (6 H, m, pentyl); δ 2.8 (4 H, m, pentyl); δ 3.0 (6
H, m, N(CH3)2); δ 5.9 (1 H, s, N-H); δ 6.23 (1 H, t, N-H); δ 7.7 (4
H, q, aromatic); δ 8.1 (1 H, s, N-H). Anal. Calcd. for C19H32N4O5S1:
C, 53.2; H, 7.50; N, 13.1. Found: C, 51.7; H, 7.03; N, 12.8.
Deprotection of the BOC Group, NH2-PN-SO2R (6). 0.5 g of 5
was dissolved in 5 mL TFA/30 mL CH2Cl2 and stirred under nitrogen
for several hours. The solvent was removed under reduced pressure
leaving behind a clear sticky residue. The product was recrystallized
from methanol/ether affording a white crystalline product. Yield ∼90%.
1H NMR in (CD3)2O (ppm): δ 1.4 (4 H, m, pentyl); δ 1.5 (2 H, m,
pentyl); δ 2.5 (2 H, d, NH2); δ 2.85 (2 H, q, pentyl); δ 3.1 (6 H, s,
N(CH3)2); δ 3.65 (2 H, t, pentyl); δ 6.43 (1 H, t, N-H); δ 7.65 (4 H,
q, aromatic); δ 8.2 (1 H, s, N-H). Anal. Calcd. for C14H24N4O3S1: C,
51.2; H, 7.36; N, 17.1. Found: C, 53.43; H, 7.58; N, 17.94.
Synthesis of FeIII-(EDTA)-PN-SO2R (7). 0.459 g (1.2 mmol) of
[FeIII-(EDTA)‚2H2O] was dissolved in 7 mL of anhydrous DMF and
stirred at 0 °C under dry nitrogen. To this solution, 0.250 g (1.2 mmol)
of DCC and 0.0698 g (0.6 mmol) of NHS were added and stirred.
After 10 min, 0.20 g (0.6 mmol) of 5 and a few drops of triethylamine
were added. After 30 min, another equivalent of DCC was added and
the solution was stirred for 36 h at room temperature. The reaction
was quenched by adding 50 mL of water. A white precipitate,
dicyclohexylurea (DCU), was isolated by filtration. The clear yellow
filtrate was concentrated under reduced pressure leaving the product
in neat DMF (2-3 mL). Neat toluene was added to the DMF solution
resulting in a yellow precipitate. This precipitate was isolated by
filtration, washed with toluene, acetone, and diethyl ether, and dried
in a desiccator. The FeIII-(EDTA)-coupled herbicide was purified by
running a HPLC column similar to that described for 4. The coupled
product 7 eluted at 25 min followed by 6 at 28 min. Yield: ∼15%.
Anal. Calcd. for C24H35N6O11S1Fe1: C, 42.9; H, 5.25; N, 12.5. Found:
C, 42.6; H, 5.71; N, 12.3.
Crystal Structure of 4. Single crystals of 4 were obtained by slow
vapor diffusion of acetone into a saturated solution of 4 in water.
Crystallographic data, intensity collection information and structural
refinement parameters are provided in Table 1. All measurements were
made on a Nonius Kappa CCD diffractometer with graphite mono-
chromated Mo KR radiation. The data were collected at -90 °C to a
maximum 2θ value of 61.0°. One φ scan consisting of 183 data frames
and omega scans consisting of 43 data frames were collected with a
scan width of 1° and a detector to crystal distance, Dx, of 33 mm.
Each frame was exposed twice (for the purpose of de-zingering) for
180 s. The data frames were processed and scaled using the DENZO
software package.14 Of the 9547 reflections collected, 7523 were unique
(Rint ) 0.043). No decay corrections were applied, but intensity data
were corrected for SORTAV absorption15 (absorption coefficient µ )
6.4 cm-1 for Mo KR radiation) and for Lorentz and polarization effects.
Figure 1. Synthesis of a redox-active herbicide, 4, by coupling FeIII-
(EDTA) to a primary amine group of a hydrocarbon spacer using N,N′-
dicyclohexylcarbodiimide (DCC).
temperature. After 30 min, 0.108 g (52.8 mmol) of DCC was added to
the reaction mixture to compensate for any loss of DCC due to
hydrolysis. After 36 h, the reaction was quenched by adding excess
(∼50 mL) of water. The resulting white precipitate, dicyclohexyl urea
(DCU), was isolated by filtration. The yellow filtrate was concentrated
under reduced pressure leaving the product in neat DMF (2-3 mL).
Neat toluene was then added dropwise, resulting in a yellow precipitate.
Finally, this yellow solid was isolated by filtration, washed with toluene,
acetone, and diethyl ether, and stored in a desiccator. Purification of
the coupled product was accomplished by using HPLC and eluting the
product with two different solvent mixtures: solvent A: 98% water,
2% acetonitrile, 0.06% TFA; solvent B: 80% acetonitrile, 20% water,
0.05% TFA. The HPLC solvent gradient went from 100% A, 0% B to
80% A, 20% B in 40 min. Uncoupled FeIII-(EDTA) eluted at 4 min
followed by the coupled product 4 at 21 min and 3 at 24 min. Yield:
∼15%. The UV-vis spectrum of 4 showed two bands: 250 nm (ꢀ ≈
26000 cm-1 M-1) (π-π*); 350 nm shoulder (ꢀ ≈ 5000 cm-1 M-1
)
(LMCT). The compound was crystallized from water/acetone solution
(see below). Anal. Calcd. for C21H29N6O11S1Fe1: C, 40.0; H, 4.61; N,
13.4. Found: C 39.3; H, 4.94; N, 12.9.
Synthesis of BOC-PN-SO2R (5). To 0.5 g (1.90 × 10-3 mol) of
RSO2Cl in 30 mL of anhydrous THF, mono-tert-butoxycarbonyl 1,5-
diaminopentane toluenesulfonic acid salt (BOC-PN) 0.855 g (1.2
equiv) in 60 mL of anhydrous THF was added dropwise over 1 h.
After the mixture stirred for 15 min, 10 mL of water was added into
the reaction mixture, and it was gently refluxed at 60 °C. After 2 more
hrs, the solution was cooled to room temperature, and the organic layer
was separated, washed with 2 × 10 mL of saturated NaHCO3 followed
The crystal structure was solved by the heavy atom Patterson
method16 and expanded using Fourier techniques.17 The non-hydrogen
(14) Otwinowski, Z.; Minor, W. Methods in Enzymology 1997, 276, 307-
326.
(15) Blessing, R. H. Acta Cryst. 1995, A51, 33-38. (b) Blessing, R. H.
J. Appl. Cryst. 1997, 30, 421-426.

TurnoVer Control of Photosystem II
J. Am. Chem. Soc., Vol. 122, No. 21, 2000 5183
Figure 2. ORTEP view of FeIII-(EDTA)-EN-SO2R (4). Hydrogen atoms have been omitted for clarity. See Table 2 for the structure of R.
Table 1: Crystal Data for 4
contained 50 mM MES/KOH, pH 6.5, 10 mM NaCl, 20 mM CaCl2,
0.03% (w/v) n-dodecyl-â-D-maltoside, 1 mM K3FeIII(CN)6, 250 µM
DCBQ and 1 M sucrose. Generally, O2-evolution activity for a given
concentration of the herbicides was determined from an average of three
runs. Typically, O2-evolution activities of 375 and 3000 µmol O2 h-1
(mg chlorophyll)-1 were found for spinach and Synechocystis PSII
samples, respectively.
empirical formula
formula weight
crystal color, habit
crystal dimensions
crystal system
C21H29N6O13SFe
661.40
red, plate
0.07 mm × 0.10 mm × 0.14 mm
triclinic
lattice type
primitive
lattice parameters
Electrochemical Measurements. Electrochemical measurements
were performed on an EG&G Princeton Applied Research, (model 273),
potentiostat using aqueous solutions with 0.2 mM KCl supporting
electrolyte. Cyclic voltammograms were recorded with a platinum disk
as a working electrode, a platinum wire as a counter electrode referenced
to a standard calomel electrode (SCE). Cyclic voltammograms (CV)
showed reversible couples at -0.05 V vs SCE (0.2 V vs NHE) with a
peak-to-peak separation of 0.07 V for [FeIII-(EDTA‚2H2O] and both
FeIII-(EDTA)-coupled herbicides 4 and 7.
Electron Paramagnetic Resonance (EPR) Spectroscopy. EPR
measurements were performed on a Varian E-9 spectrometer equipped
with a TE102 cavity and a helium flow cryostat (Oxford Instruments),
and interfaced with a Macintosh IIci computer. Spectra of the FeIII-
(EDTA) moiety in the redox-active herbicides and the S2-state of PSII
were obtained under the following instrumental conditions: microwave
frequency, 9.28 GHz; magnetic field modulation frequency, 100 kHz;
magnetic field modulation amplitude, 20 G; scan width, 4000 G;
temperature, 10 K. Due to the different concentrations of the FeIII-
(EDTA)-containing herbicides and the Mn4 cluster in PSII, 0.1 mW
power was applied in acquiring spectra of the FeIII-(EDTA) species,
while 5 mW power was applied in acquiring the S2-multiline signals.
Metmyoglobin (Sigma) was added to the EPR samples of PSII as an
internal standard to quantitate the signal from the FeIII-(EDTA) moiety.
The EPR signal intensities of metmyoglobin (g ) 6) and FeIII-(EDTA)
(g ) 4.3) were measured from the peak-to-trough heights. The intensity
of the S2-multiline EPR signal (g ≈ 2) was estimated as the sum of the
heights of four hyperfine peak heights in the light-minus-dark difference
spectra. Each spectrum was an average of 2 to 4 scans. Power saturation
experiments for FeIII-(EDTA)-coupled herbicide and myoglobin show
no saturation effects up to 5 mW of applied microwave power at 10
K.
a ) 9.096(1) Å
b ) 11.389(1) Å
c ) 15.700(1) Å
R ) 70.752(3)°
â ) 79.524(4)°
γ ) 86.918(3)°
V ) 1509.8(2) Å3
P-1 (#2)
space group
Z value
Dcalc
2
1.455 g/cm3
686.00
F000
µ (Mo KR)
residuals: R; R
goodness of fitwindicator
6.38 cm-1
0.058; 0.083
2.12
maximum peak in final diff. map 1.24 e-/Å3
minimum peak in final diff. map -0.36 e-/Å3
R ) ∑||Fo| - |Fc||/∑|Fo|; R ) [(∑w(|Fo| - |Fc|)2/∑wFo )]1/2
.
2
w
atoms were refined anisotropically and hydrogen atoms were included
but not refined. In the case of methyl group hydrogen atoms, one
hydrogen was located in the difference map and included at an idealized
distance to set the orientation of the other two hydrogen atoms. The
hydrogen atoms were not observed and included for the four partial
occupancy water molecules in the lattice. The final cycle of full matrix
least squares refinement was based on 2906 observed reflections (I >
5.00 σ(I) and 397 variable parameters. The X-ray structure of 4 is shown
in Figure 2.
O2-Evolution Assays. O2-evolution assays for the determination of
the inhibition of PSII by the metal-containing redox-active herbicides
were performed using a YSI Clark-type electrode. Light from a 1200
W xenon lamp (Oriel), filtered through a 10 cm water filter, a heat
absorbing filter (Schott KG-5) and a long pass filter (Oriel LP 610),
was directed by a Pyrex light-pipe18 to a water-jacketed aluminum
chamber maintained at 25 °C by a Neslab RTE-9DD circulator bath.
Typically, 10 µg of chlorophyll spinach PSII was added in the dark
into a 2.5 mL buffer solution containing 20 mM MES/NaOH (pH 6.0),
10 mM NaCl, 1 mM K3FeIII(CN)6 and 250 µM 2,5-dichloro-p-
benzoquinone (DCBQ). The O2-evolution activity of Synechocystis PSII
core complexes was measured in the same cell except that the buffer
Photosystem II Sample Preparation. PSII membranes were isolated
from market spinach leaves by a modified version19 of the method of
Berthold et al.20 except that the thylakoid membranes were not frozen
at 77 K prior to isolating the PSII membranes. Histidine-tagged (His-
tag) PSII core complexes were isolated from Synechocystis PCC 6803
according to the method of Reifler et al.21 After extraction, the spinach
PSII membranes were suspended in buffer containing 20 mM MES/
NaOH, pH 6.0, 15 mM NaCl, 30% (v/v) ethylene glycol:water and
stored at 77 K. Typically, PSII samples with [chlorophyll] ([Chl]) )
8-9 mg of chlorophyll/ml (determined according to the method of
(16) Beurskens, P. T.; Admiraal, G.; Beurskens, G.; Bosman, W. P.;
Garcia-Granda, S.; Gould, R. O.; Smits, J. M. M.; Smykalla, C. PATTY;
the DIRDIF program system, Technical Report of the Crystallography
Laboratory, University of Nijmegen: The Netherlands, 1992.
(17) Beurskens, P. T.; Admiraal, G.; Beurskens, G.; Bosman, W. P.;
deGelder, R.; Israel, R.; Smits, J. M. M. DIRDIF94; the DIRDIF program
system, Technical Report of the Crystallography Laboratory, University of
Nijmegen: The Netherlands, 1994.
(19) Beck, W. F.; dePaula, J. C.; Brudvig, G. W. Biochemistry 1985,
24, 3035-3043.
(20) Berthold, D. A.; Babcock, G. T.; Yocum, C. F. FEBS Lett. 1981,
134, 231-234.
(21) Reifler, M. J.; Chisholm, D. A.; Wang, J.; Diner, B. A.; Brudvig,
G. W. In Photosynthesis: Mechanisms and Effects; Garab, G., Ed.; Kluwer
Academic Publishers: Dordrecht, The Netherlands, 1998; Vol. 2, pp 1189-
1192.
(18) Schweitzer, R.; Brudvig, G. W. Biochemistry 1997, 36, 11351-
11359.

5184 J. Am. Chem. Soc., Vol. 122, No. 21, 2000
Karki et al.
Arnon22), were treated with K3FeIII(CN)6 (1 mM final concentration)
and incubated in the dark at 0 °C to oxidize fully cytochrome b559 (Cyt
b559). Excess ferricyanide was removed by pelleting and resuspending
the PSII membranes in buffer with 20 mM MES/NaOH (pH 6.0) and
50% (v/v) ethylene glycol:water. For EPR studies, 200 µL of PSII
membranes were transferred to a 4 mm (OD) quartz EPR tube (Wilmad
glass) and were further treated with 50 µM ferricyanide. Herbicides 4
and 7 with the FeIII-(EDTA) acceptor were added to the sample in
the dark to a final herbicide concentration of 700 µM (∼12-fold excess
in comparison to the PSII concentration). The sample was then
incubated at 0 °C for 0.5 h to allow for equilibration and diffusion of
the herbicide into the QB-binding pocket. To this sample, 500 µM
myoglobin in a buffer containing 20 mM MES/NaOH (pH 6.0) and
50% (v/v) ethylene glycol:water was added as an internal spin standard
before rapidly freezing the EPR sample to 77 K. Myoglobin was chosen
as an internal spin standard as it does not chemically interfere with the
PSII samples and has a narrow EPR signal at g ) 6.0 that does not
overlap with the FeIII-(EDTA) signal (g ) 4.3) and the PSII signals
of interest in this study.
Illumination Protocol for Turnover Experiments. EPR tubes
containing the treated PSII samples were illuminated in a home-built
nitrogen-flow cryostat. Prior to each illumination, the sample was
incubated at the required temperature for 2-3 min to allow sufficient
time for thermal equilibration. Continuous illumination (500 W/m2)
was provided by a halogen lamp. The PSII samples were first
illuminated at 195 K for 3.5 min to induce the S2 state. The S2 to S3
turnover was monitored by recording EPR intensity changes at 10 K
following continuous illumination for varying time intervals at a given
temperature (195-245 K) in the presence of the different redox-active
herbicides.
Figure 3. O2-evolution activities as a function of concentration of 2
(b), 3 (9) and 4 (2) at pH 6.0. The BOC-protected herbicide derivative
shows one order of magnitude higher inhibition than either the free
amine herbicide derivative, 3, or the herbicide with an FeIII-(EDTA)
group, 4. The error bars reflect 8% uncertainty in the measurements.
The lines are least-squares fits of eq 1.
Table 2: Inhibition of O2-Evolution from Spinach PSII
Membranes by Various Herbicide Derivatives
herbicide derivative
DCMU
BOC-EN-SO2R (2)
BOC-PN-SO2R (5)
amino-EN-SO2R (3)
pI50
6.5
4.3
4.8
3.4
5.1
3.4
3.9
Results
Strategies for Designing Metal-Containing Redox-Active
Herbicides. There are basically four requirements for designing
efficient redox-active herbicides. First, the herbicide should have
a high affinity for the QB-binding pocket as evidenced by a
high pI50 value23 (typically pI50 > 3.0). Second, the herbicide
should accept an electron from QA- without protonation in order
to enable fast and/or low-temperature turnover. Third, the
acceptor’s reduction potential must be higher than that of QA
for it to accept an electron from QA-. And finally, the herbicide
should possess a spectroscopic handle that would enable
observation and quantification of the electron-transfer reaction.
Conceptually, a metal-containing herbicide featuring an FeIII-
(EDTA) electron-acceptor moiety linked via a hydrocarbon
spacer to a QB-site binding functionality (dimethylphenylurea)
meets all of these requirements (Figure 1). These herbicides
share the same dimethylphenylurea binding group as PNPDU
(Scheme 2), except that the electron-acceptor group is an FeIII-
(EDTA) complex. The reduction potential of 4, 0.2 V vs NHE,
is higher than that of QA (-0.10 V vs NHE), and electron-
transfer proceeds without protonation. Furthermore, MIII-
(EDTA) complexes (where M ) Co, Ni, Fe, Cu) are seven
coordinate24 with one protonated EDTA arm available for
covalent linkages via an amide bond to the dimethylphenylurea
moiety. The synthesis of 4 is shown in Figure 1 where one
carboxylic acid group of FeIII-(EDTA) is coupled to the primary
amine group of 3 using DCC as a coupling agent. It is evident
from Figure 1 that varying the hydrocarbon spacer would allow
for addressing the binding efficiency as well as the distance
dependence of electron transfer to the FeIII-(EDTA) acceptor.
In addition, the paramagnetic FeIII ion in these herbicides has
amino-PN-SO2R (6)
FeIII-(EDTA)-EN-SO2R (4)
FeIII-(EDTA)-PN-SO2R (7)
an EPR signal in the oxidized form that is eliminated upon
reduction. This EPR signature allows measurement of the extent
and rate of electron transfer.
O2-Evolution Inhibition Studies. Experiments to monitor
inhibition of O2-evolution activity were performed in the
presence of the electron acceptor DCBQ (250 µM) with 1 mM
ferricyanide to ensure complete oxidation of the DCBQ pool.
The inhibition rates were measured for a series of nonredox-
active herbicide derivatives as well as the FeIII-(EDTA)-coupled
herbicides. The observed O2-evolution rates in the presence of
the herbicides were modeled by eq 1:
Vmax
V )
(1)
(KI[I] + 1)
where ν is the observed rate of O2 evolution, νmax is the
normalized rate of O2-evolution in the absence of an inhibitor,
KI is the equilibrium association constant between the QB-
binding pocket and the herbicide and [I] is the inhibitor
concentration. Figure 3 shows the plot of inhibition of O2 rates
by 2, 3, and 4. The corresponding pI50 values are shown in Table
2. From Figure 3, it is evident that 2, for which the amino group
is protected, shows one order of magnitude higher inhibition
of O2-evolution rates compared to the herbicides 3 and 4. Unlike
2 where the BOC group protects the primary amine, the BOC-
deprotected primary amine, 3, is charged in the O2-evolution
assay buffer (pH 6.0). Thus, the charged amine group of 3 and
the hydrophilic FeIII-(EDTA) group of 4 interact unfavorably
with the polypeptide residues within the QB-binding pocket
(22) Arnon, D. I. Plant Physiol. 1949, 24, 1-15.
(23) The inhibitory potency of a herbicide is expressed by the equation
pI50 ) -log [I]50 where I50 is the concentration at which the O2-evolution
activity is half the initial value.
(24) Zubkowski, J. D.; Perry, D. L.; Valente, E. J.; Lott, S. Inorg. Chem.
1995, 34, 6409-6411.

TurnoVer Control of Photosystem II
J. Am. Chem. Soc., Vol. 122, No. 21, 2000 5185
Table 3: pI50 Values for Inhibition of O2-Evolution Activity from
Different Sample Types Using 4 at 25 °C
sample type
thylakoid membranes
spinach PSII membranes
pI50
∼ 2
3.4
Synechocystis PCC 6803 core complexes
3.7
Figure 4. O2-evolution activity as a function of concentration of 4 in
different sample types: spinach thylakoid membranes (b), spinach PSII
membranes (9), Synechocystis PSII core complexes (2). The error bars
reflect 8% uncertainty in the measurements. The lines are least-squares
fits of eq 1.
leading to a substantial loss in binding affinity.11,25 Moreover,
the different pI50 values for the series of herbicide derivatives
shown in Table 2 indicate that addition of functionalized groups
beyond the DCMU parent structure leads to less efficient binding
to the QB site.26 Nevertheless, adding hydrocarbon spacers via
a sulfonamide bond at the para-position of the phenylurea moiety
of the DCMU structure results in acceptable binding and good
inhibition. Furthermore, adding a longer hydrocarbon spacer
between the phenylurea moiety and the FeIII-(EDTA) group
leads to tighter binding as evidenced by the increase in the pI50
values for the pentane spacer herbicide derivatives compared
to those for the ethane spacer herbicide derivatives.27 One likely
explanation for the increased binding affinity with the pentane
spacer is the removal of the charged amine and hydrophilic
FeIII-(EDTA) groups farther away from the hydrophobic
residues in the QB-binding site, leading to less unfavorable
electrostatic interactions.
Figure 5. EPR signals from the FeIII-(EDTA) moiety of 4 at g ) 4.3
and from myoglobin at g ) 6.0 added to a PSII membrane sample as
an internal standard. a) Dark scan, (b) scaled difference between the
spectrum after 8 min illumination at 225 K and the above dark scan.
The negative amplitude of the difference signal compared to no change
in the myoglobin signal is indicative of reduction of the FeIII-(EDTA)
moiety upon continuous illumination at 225 K.
branes, followed by PSII membranes and finally the PSII core
complexes. These observations are in agreement with experi-
ments by Gray and co-workers with charged pentammine
ruthenium complexes.28 They have reported that hydrophilic
charged metal-complexes prevent penetration into the hydro-
phobic sites within the protein. On the other hand, DCMU and
herbicides with BOC-protected amines are uncharged and
possess hydrophobic groups that can interact favorably with the
hydrophobic polypeptide residues of the binding pocket. These
herbicides are observed to bind the QB site with little hindrance
from the lipid network. Thus, the different pI50 values observed
for the series of herbicide derivatives and sample types suggest
that a combination of electrostatic, hydrophilicity and acces-
sibility factors are at play in determining the strength of
inhibition.
Turnover Control in PSII at 225 K. Turnover control
experiments were performed at 225 K in samples of PSII
membranes containing 0.7 mM of 4. On the basis of the pI50
value of 3.4 (Table 2), approximately a 12-fold excess of 4 was
added to ensure that at least 50% of the QB-binding site was
occupied and that the S2 to S3 conversion would produce a
detectable change in the EPR signal of the FeIII-(EDTA)
acceptor. Although adding a higher concentration of 4 would
lead to a higher occupancy of the QB-binding site, the change
of the FeIII-(EDTA) EPR signal would be more difficult to
measure. Figure 5a shows the low-field EPR spectrum of a dark-
adapted sample with the myoglobin and the FeIII-(EDTA)
signals at g ) 6.0 and g ) 4.3, respectively. Figure 5b shows
the light-minus-dark difference EPR signals in the same sample
after 8 min of illumination at 225 K. It can be observed that
there is a loss of 9 ( 4% of the FeIII-(EDTA) acceptor signal
and no change in the myoglobin (internal standard) signal. The
PSII sample concentration was 61 µM (based on the [Chl] and
calculated using 200 Chl/PSII). Because the QB sites are
expected to be 52% occupied by the herbicide (based on Figure
To address the ability of the metal-containing herbicides to
access the QB-binding pocket in PSII, rates of O2-evolution
inhibition were compared for thylakoid and PSII membranes
from spinach and for detergent-solubilized PSII core complexes
from Synechocystis PCC 6803.11 In the case of the core
complexes, the QB site is the most exposed to the surrounding
medium because the core complexes are solubilized and not
embedded in native membrane sheets. In addition, core com-
plexes lack associated light-harvesting complexes that could
further shield the QB-binding site. The PSII membranes,
however, occur as membrane fragments with significantly less
exposure of the QB sites to the aqueous solution. And, the
thylakoid membranes contain both PSI and PSII units within
intact membranes that block access to the intermembrane
proteins by charged hydrophilic groups. The O2-evolution
inhibition curves shown in Figure 4 and the corresponding pI50
values in Table 3 support the expected trend in the QB site
accessibility by 4 because the pI50 values appear to depend on
the degree of accessibility to the QB pocket. The core complexes
show the largest pI50 value followed by the PSII membranes
and finally by the thylakoid membranes. Because both 4 and 7
have a hydrophilic FeIII-(EDTA) group, accessibility from the
lipid surface would be restricted the most in thylakoid mem-
(25) Draber, W.; Tietjen, J.; Kluth, J. F.; Trebst, A. Angew. Chem., Int.
Ed. Engl. 1990, 30, 1621-1633.
(26) Omokawa, H.; Takahashi, M. Pest. Biochem. Physiol. 1994, 50,
129-137. (b) Fedke, C. Biochemistry and Physiology of Herbicide Action;
Springer-Verlag: New York, 1982.
(27) Asami, T.; Koike, H.; Inoue, Y.; Takahashi, N.; Yoshida, S. Z.
Naturforsch. 1988, 43c, 857-861.
(28) (a) Gray, H. B. Chem. Soc. ReV. 1986, 15, 17. (b) Bowler, B. E.;
Raphael, A. L.; Gray, H. B. Prog. Inorg. Chem. 1990, 38, 259-322.

5186 J. Am. Chem. Soc., Vol. 122, No. 21, 2000
Karki et al.
Figure 6. EPR spectra of the PSII membranes in the presence of 0.7
mM of 4 showing the S2-multiline signal. (a) Spectrum after 3.5 min
continuous illumination at 195 K, (b) spectrum after 2.0 min continuous
illumination at 225 K, (c) spectrum after a total of 8 min illumination
at 225 K. The narrow TyrD signal at 3300 G has been deleted for
clarity.
Figure 8. EPR spectra of the PSII membranes in the presence of 0.02
mM DCMU and 0.7 mM 4 showing the S2-multiline EPR signal. (a)
Spectrum after 3.5 min continuous illumination at 195 K, (b) spectrum
after 2.0 min continuous illumination at 225 K, (c) spectrum after a
total of 8 min illumination at 225 K. The narrow TyrD• signal at 3300
G has been deleted for clarity.
•
of the illumination time. As expected, the FeIII-(EDTA)
acceptor and the S2-multiline EPR signal intensities decrease
due to turnover of the S2 to the S3 state with a concomitant
FeIII to FeII reduction of the acceptor. As shown in Figure 7,
approximately half of an equivalent of FeIII-(EDTA) was
reduced during the first few minutes of illumination and
prolonged illumination gave no further turnover. This is
consistent with 4 accepting one electron while bound to the QB
site, and with no exchange of 4 into and out of the QB site
during the period of illumination at 225 K. The 60% loss of
the S2-multiline signal mirrors the percent O2-evolution inhibi-
tion by 4 at 700 µM concentration, because the inhibition curve
of 4 (Figure 3) shows that only 52% of the centers are inhibited
at 0.7 mM concentration of 4.
Figure 7. Percent EPR signal intensity changes upon continuous
illumination of PSII membranes with 0.7 mM 4 at 225 K: S2-multiline
signal (b), FeIII-(EDTA) of 4 (9). The amplitudes of the S2-multiline
and FeIII-(EDTA) of 4 signals measured following the 195 K
illumination were normalized to a value of 100% prior to proceeding
to the 225 K illuminations. The lines represent exponential fits to the
data for first-order decay kinetics at 225 K. The error bars represent
4% experimental uncertainty in the measurement of the data points.
To verify that the S2 to S3 turnover occurs via electron transfer
to 4 bound to the QB site, control experiments were performed
with samples treated with 200 µM DCMU prior to treatment
with or without 700 µM of 4 in the same PSII sample. Because
DCMU has a greater affinity for the QB-binding site (pI50
)
6.5, Table 2), it would preferentially occupy the QB-binding
site in PSII, while 4 would be randomly dispersed on the
surrounding surface sites or remain in solution. Figures 8 and
9 show the EPR spectra and the corresponding intensities of
the FeIII(EDTA) and S2-multiline signals as a function of the
illumination time at 225 K. It is apparent from Figures 8 and 9
that there is no significant decay of the S2 multiline and FeIII-
(EDTA) EPR signals even after 8 min of continuous illumina-
tion. The observed residual decrement of 15 ( 4% in the S2-
multiline signal may be due to incomplete occupancy of the
QB sites, to transfer of electrons to dioxygen18 dissolved in the
sample or to inherent experimental errors contributing to the
EPR measurement. The fact that the S2-multiline and the FeIII-
(EDTA) signal intensities remain constant with 225 K illumina-
tion further indicates that there are no surface pathways of
electron transfer to the FeIII-(EDTA) acceptor of 4. This control
experiment supports the conclusion that 4 accepts an electron
while bound in the QB site and allows for S2 to S3 state turnover.
Similar turnover experiments in PSII membranes were also
performed at 225 K using 500 µM of 7 which has an FeIII-
(EDTA) acceptor with a pentane spacer linked to the dimethyl
phenylurea moiety. It was observed that 65% of the S2-multiline
signal decayed away after 8 min of continuous illumination at
3), we expect that the FeIII-(EDTA) signal would decrease by
approximately 4.6% based on the ratio of the concentrations of
the QB site bound herbicide to the PSII sample. The slightly
greater measured loss in the FeIII-(EDTA) signal may be due
to experimental uncertainty.
Figure 6 shows the corresponding decrease in the S2-multiline
EPR signal upon illumination at 225 K. For untreated samples
with 195 K illumination, the dark adapted S1 state can progress
only to the S2 state in almost 100% yield without undergoing
further turnover.29 However, illumination at 225 K allows
advance to the S3 state, provided that there is an electron
acceptor species at the QB site. Figure 6a shows the S2-multiline
EPR signal acquired after a 3.5 min illumination at 195 K,
Figure 6b shows the S2-multiline signal upon continuous
illumination at 225 K for 2 min and Figure 6c shows the S2-
multiline signal after a total of 8 min (a sequence of four 2 min
illuminations). The intensity of the S2-multiline signal was
observed to decay during illumination at 225 K. Figure 7 shows
the normalized plots representing the percent change of the S2-
multiline and FeIII-(EDTA) acceptor EPR signals as a function
(29) dePaula J. C.; Innes, J. B.; Brudvig, G. W. Biochemistry 1985, 24,
28-34.

TurnoVer Control of Photosystem II
J. Am. Chem. Soc., Vol. 122, No. 21, 2000 5187
electron transfer to the FeIII-(EDTA) acceptor may not depend
on the hydrocarbon spacer length at least up to 5 carbon atom
spacer length.
Discussion
Previous EPR studies with the redox-active herbicide PNPDU
have shown that a controlled two-electron turnover from the S1
to S3 state can occur successfully by continuous illumination at
250 K, forming the S3 state with a 75% yield.8 However, two-
electron turnover of PSII at lower temperatures could not be
achieved with PNPDU because of the protonation requirement
for electron transfer to the nitroxyl group (Scheme 2). The
current results show that the two-step turnover from S1 to S3
turnover can be successfully achieved with 4 and 7 at much
lower temperatures than previously possible with PNPDU,
mainly because the FeIII-(EDTA) electron acceptor of the
herbicides does not require protonation for electron transfer.
Furthermore, because both 4 and PNPDU have only moderate
binding affinity for the QB binding site, the low-temperature
(<250 K) studies with 4 effectively minimize the problem of
multiple turnovers due to herbicide dissociation. The yield of
the S3 state was about 60-65% in the majority of the
experiments with spinach PSII. This yield correlates well with
the extent of O2-evolution inhibition in spinach PSII membranes
by the redox-active herbicides; because of the reduced affinity
for the QB site of 4 and 7 relative to PNPDU, only ∼50-60%
of the QB sites were occupied at the concentrations used in our
studies. Although the present study uses spinach PSII with
moderate (700 µM) concentrations of 4 and 7 to demonstrate
the S2 to S3 conversion of the OEC, a greater than 90% yield
of the S3 state could be achieved by using Synechocystis PCC
6803 core complexes and a higher concentration of 4. In
comparison to PNPDU, which exhibits a large EPR signal at g
≈ 2 that obscures the EPR signal from PSII species, the
nonoverlapping EPR signal of the FeIII-(EDTA) moiety at g
) 4.3 permits the observation of both the herbicide and the
OEC signals while monitoring the S2 to S3 conversion.
The O2-evolution studies with metal-containing redox-active
herbicides 4 and 7 suggest that the dimethylphenylurea group
enables the FeIII-(EDTA) acceptor to dock in the QB-binding
pocket because FeIII-(EDTA) by itself does not inhibit O2-
evolution activity or accept electrons from PSII at low temper-
atures. However, the QB site affinity of 4 and 7 are compromised
due to the presence of the hydrophilic FeIII-(EDTA) group.
The structural details of the QB-binding pocket in plant PSII
would benefit the design of tighter binding redox-active
herbicides. The currently available 8 Å resolution structure of
plant PSII, however, is still insufficient to provide a detailed
structure of the plant QB site.30 Nonetheless, the high-resolution
crystal structures of the analogous bacterial reaction center
(bRC) from Rb. sphaeroides and Rb. Viridis have provided
structural insight into the QB-binding pocket.31 The crystal
structure of the bRC QB site shows a hydrophobic pocket with
internal dimensions of ∼10 Å × ∼8 Å. Crystal structures with
the herbicides terbutryn and atrazine bound inside the QB-
binding pocket reveal specific hydrogen bonding interactions
of the triazine nitrogens with Ser L223, Ile L224, His L190,
Phe L216, Glu L212, and Tyr L222 in conferring the tight
binding affinity.32 Although crystal structures with DCMU
Figure 9. Percent changes of the S2-multiline (b) and FeIII-(EDTA)
of 4 (9) EPR signals in the presence of 0.020 mM DCMU and 0.7
mM of 4. The amplitudes of the S2-multiline and FeIII-(EDTA) of 4
signals measured following the 195 K illumination were normalized
to a value of 100% prior to proceeding to the 225 K illuminations.
The lines represent exponential fits of the data points and the error
bars reflect 4% uncertainty level in the measurements.
Figure 10. Arrhenius plot of the natural logarithm of the rate constant
(kobs) versus the inverse temperature. The line is the linear least-squares
fit to the data.
225 K. A slightly higher loss of the S2-multiline signal than
observed with 4 suggests that a slightly greater percentage of
the QB sites are occupied by 7, in agreement with the respective
pI50 values. As in the case of 4, a 9 ( 4% loss in the FeIII-
(EDTA) EPR signal was observed with a concomitant loss in
the S2-multiline signals in the 225 K illumination experiment
indicating that approximately one equivalent of the FeIII-
(EDTA) acceptor underwent reduction upon continuous il-
lumination.
Further S2 to S3 state turnover experiments were performed
with samples containing 4 at various temperatures ranging from
195 to 245 K to investigate the temperature dependence of the
rate of electron transfer (kobs) to the FeIII-(EDTA) acceptor of
4. As in the case of the 225 K illumination experiments, the
S2-multiline EPR signal intensity was observed to decrease by
60% upon continuous illuminations with a concomitant 9 (
4% loss in intensity of the FeIII-(EDTA) acceptor of 4 at 235
and 245 K illumination temperatures. In the case of 195 K
illumination, the S2-multiline signal decayed away by about 40%
of the initial value upon continuous stepwise illumination up
to 23 min. The observed S2-multiline EPR signal decay lifetimes,
(τ ) 1/e), varied with the temperature of illumination and the
observed decays were fit using an exponential function assuming
first-order decay kinetics as shown in Figure 7. An Arrhenius
plot of the observed rate constant, kobs, (sum of the various rates
including both donor and acceptor side effects) as a function
of the illumination temperature for 4 is shown in Figure 10,
(30) Rhee, K.-H.; Morris, E. P.; Barber, J.; Ku¨hlbrandt, W. Nature 1998,
396, 283-286.
(31) Deisenhofer, J.; Michel, H. Science 1989, 245, 1463-1473.
(32) Lancaster, C.; Roy, D.; Michel, H. Photosyn. Res. 1996, 48, 65-
74. (b) Lancaster, C.; Roy, D.; Michel, H. J. Mol. Biol. 1999, 286, 883-
898.
and gives an effective activation barrier, EA, of 35 ( 5 kJ mol-1
.
It was also observed that both 4 and 7 show comparable S2-
multiline decay lifetimes at 225 K suggesting that the rate of

5188 J. Am. Chem. Soc., Vol. 122, No. 21, 2000
Karki et al.
bound in the QB pocket are not presently available, we can infer
that the urea group forms hydrogen bonds to the polypeptide
residues similar to triazines and quinones. The dimethylphenyl-
urea group of 4 and 7 can be envisioned to be docked at the QB
site and the FeIII-(EDTA) acceptor to project out of the QB-
binding pocket. Furthermore, by analogy to the bRC structure,
where the quinone tail points toward the PheoA molecule, it is
possible that the FeIII-(EDTA) moiety is oriented toward the
PheoA. This would account for the evidence that the FeIII-
QB. Because the bRC and PSII are structurally homologous with
similar electron-acceptor sides, the electron-transfer mechanism
from QA to QB in PSII could likely involve similar herbicide
and protein structural rearrangements in the QB-binding pocket
during the electron-transfer event.
-
The current results show that there are several advantages of
using metal-containing redox-active herbicides in studying
intermediate states of the PSII photocycle. First and most
importantly, these redox-active herbicides enable the high-yield
preparation of the S3 state for future spectroscopic studies of
the S3 state. Although low-temperature turnover studies of
spinach PSII membranes show that the S3 state can be formed
in 60-65% yield, inhibition studies suggest that yields as high
as ∼90% could be obtained when using high concentrations of
4 with PSII core complexes, because 4 binds much better to
the PSII core complexes owing to the greater accessibility of
the QB site to the aqueous phase. Second, these metal-containing
herbicides allow for the possibility of tuning the energetics to
investigate the long-range coupling and electron-transfer mech-
anisms in PSII by allowing variation in the coordinating ligands.
For example, [MIII-(EDTA)L] acceptors (where M ) Ru) has
an open sixth coordination site that could be coordinated by
various ligands (L) to tune the driving force (∆G°) of electron
transfer.35 The dependence of the rate (kobs) on the driving force
(∆G°) would provide further insight into the mechanism of
electron transfer in PSII.
-
-
(EDTA) moiety is reduced by PheoA rather than QA (see
below).
To address the pathway of electron transfer to the FeIII-
(EDTA) moiety, a dark-incubation experiment of PSII mem-
branes with 700 µM of 4 at 225 K was performed following
the 3.5 min illumination at 195 K. Dark-incubation of the
illuminated sample for 15 min at 225 K did not produce a
significant loss in the QA-, the S2-multiline or the FeIII-(EDTA)
acceptor EPR signals of the herbicide from that of the 195 K
illumination experiment. Electron transfer to the FeIII-(EDTA)
acceptor via QA- (Scheme 3) would result in a decrease of the
QA- EPR signal during dark incubation. However, the QA- EPR
signal remains constant indicating that QA- does not donate an
electron to the FeIII-(EDTA) acceptor group in the dark at 225
K. Continuous 2 min illumination at 225 K following the 15
min dark-incubation, however, produced a 50% loss in the S2-
multiline EPR signal and a 10 ( 4% loss in the FeIII-(EDTA)
signal indicating reduction of approximately one equivalent of
4. Although the assumed conventional electron-transfer pathway
Besides enabling electron-transfer studies and the high-yield
preparation of the S3 state, these metal-containing herbicides
could afford the possibility for studying previously inaccessible
intermediate states in photosystem II. For example, if the redox-
active herbicide accepts electrons at 195 K faster than the S2 to
-
to the QB site via QA cannot be ruled out, the current results
suggest that pheophytin (PheoA-) may participate in the electron-
transfer process as it is located in close proximity, ∼15 Å, to
the QB-binding site and has a lower reduction potential (-0.5
V vs NHE) than QA (-0.1 V vs NHE). Regardless of whether
electron transfer occurs via PheoA- or via QA-, electron transfer
•
Yz electron-transfer rate, then continuous illumination could
lead to the generation of the S2Yz• state in an uninhibited PSII
sample.9 This intermediate state is typically prepared when
inhibited PSII samples are continuously illuminated at 273 K
and rapidly frozen to 77 K. However, one drawback is that the
inhibitory treatments (Ca2+ depletion, Cl- depletion or addition
of acetate) could lead to structural perturbation of the OEC.
Thus, it would be of great interest to generate and study this
state without perturbing the OEC. Furthermore, using a redox-
active herbicide that can accept multiple electrons, for example
by coupling biferrocence36 (two iron atoms) to an atrazine or
DCMU binding functionality, could lead to trapping and study
of the elusive S4 state.
-
would still involve a long-range interaction because QA is
separated from QB by ∼ 15 Å based on homology with the
bRC.31 In addition, the FeIII-(EDTA) moiety of 4 is further
separated by ∼13 Å from the phenylurea moiety which is
expected to bind in the QB pocket. Thus, electron transfer to
the FeIII-(EDTA) acceptor would involve long-range through
space and sigma-bond coupling pathways. The hydrocarbon
chain linkers in these redox-active herbicides may provide an
efficient coupling pathway as in the cytochrome P450 electron-
transfer studies by Wilker et al.10
It was observed that the rate of two-electron turnover
depended upon the temperature of illumination. An Arrhenius
analysis yielded an activation energy of 35 ( 5 kJ/mol. In
previous studies with the bRC, a large activation barrier of 56
Acknowledgment. This work was supported by the National
Institutes of Health Grant GM 32715. We thank the research
group of Professor Schepartz at Yale University for use of their
HPLC instrument. We thank Susan deGala for obtaining the
X-ray crystal structure and Michael Reifler for helpful discus-
sions about synthesis and for preparing the His-tagged PSII core
complexes. We also thank Olaf Kievit for his assistance with
electrochemical measurements.
-
kJ/mol was found for QA to QB electron transfer.33 The large
activation barrier was thought to be due to a large reorganization
energy or a conformational gating for electron transfer.34 The
conformational gating process is associated with protein dynam-
ics that allow the quinone to move into the proximal position
from the distal position enabling electron transfer from QA- to
Supporting Information Available: Tables of crystal-
lographic data including atomic coordinates, bond angles and
bond lengths, and anisotropic displacement parameters (PDF).
An X-ray crystallographic file, in CIF format. This material is
available free of charge via the Internet at http://pubs.acs.org.
(33) (a) Macino, L. J.; Dean, D. P.; Blankenship, R. E. Biochim. Biophys.
Acta 1984, 764, 46-54. (b) Kleinfeld, D.; Okamura, M. Y.; Feher, G.
Biochim. Biophys. Acta 1984, 766, 126-140. (c) Li, J. L.; Gilroy, D.; Tiede,
D. M.; Gunner, M. R. Biochemistry 1998, 37, 2818-2829.
(34) According to studies performed on the bacterial reaction center
(bRC), conformational gating is thought to limit the rate of electron transfer
-
from QA to QB. For example, see: (a) Graige, M. S.; Okamura, M. Y.
JA994138X
Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 11679-11684. In addition, the
-
-
second electron-transfer step from QA to QB is limited by the rate of
proton transfer to QB-. See: (b) Graige, M. S.; Paddock, M. L.; Bruce, J.
M.; Feher, G.; Okamura, M. Y. J. Am. Chem. Soc. 1996, 118, 9005-9016.
(c) Graige, M. S.; Paddock, M. L.; Bruce, J. M.; Feher, G.; Okamura, M.
Y. Biochemistry 1999, 38, 11465-11473.
(35) (a) Masubara, T.; Creutz, C. Inorg. Chem. 1979, 18, 1956-1966.
(b) Diamantis, A. A.; Dubrawski, J. V. Inorg. Chem. 1983, 22, 1934-
1946.
(36) (a) Morrison, W. H.; Hendrickson, D. N. Inorg. Chem. 1975, 14,
2331-2346. (b) Creutz, C. Prog. Inorg. Chem. 1983, 30, 1-73.