Active site models for galactose oxidase containing two different phenol groups

Article abstract of DOI:10.1016/S0020-1693(99)00579-4

Model complexes of the active site of galactose oxidase (GAO) have been developed using a new ligand carrying two different phenol groups, N-[(2-hydroxy-3-methylthio-5-tert-butylphenyl)methyl].N-[(2-hydroxy-3,5-di-te rt-butylphenyl)methyl].2-(2-pyridyl)ethylamine (L1H₂). Deprotonated ligand L1^2- forms a dimeric Cu(II) complex, [Cu(II)₂(L1^2-)₂], in the solid state, the structure of which has been determined by X-ray crystallographic analysis. The dimeric Cu(II)-diphenolate complex can be converted into the monomeric complex, [Cu(II)(L1^2-)(X)] (X = py, AcO, and PhCH₂OH), in solution by adding exogenous ligands such as pyridine (py), acetate (AcO^-), or benzyl alcohol (PhCH₂OH). The structure and physicochemical properties (UV-Vis, ESR, redox potential) of [Cu(II)(L1^2-)(X)] have been explored as a model for the resting state of the enzyme. One-electron oxidation of [Cu(II)(L1^2-)(py)] and [Zn(II)(L1^2-)(py)] by (NH₄)₂[Ce(IV)CNO₃)₆] (CAN) yielded the corresponding phenoxyl radical/phenolate complexes, Cu(II)(LI(·)^-) and Zn(II)(L1(·)^-), respectively, which have also been characterized by UV-Vis, resonance Raman, and ESR. The structure, physicochemical properties and reactivities of the diphenolate and phenoxyl radical/phenolate complexes of L1H₂ are compared to those of the corresponding monophenolate and monophenoxyl radical complexes in order to obtain further insight into the role of Tyr 495 in the native enzyme. (C) 2000 Elsevier Science S.A.

Full text of DOI:10.1016/S0020-1693(99)00579-4

www.elsevier.nl/locate/ica
Inorganica Chimica Acta 300–302 (2000) 622–632
Active site models for galactose oxidase containing two different
phenol groups
Masayasu Taki a, Hideyuki Kumei a, Shigenori Nagatomo c, Teizo Kitagawa c,*1,
Shinobu Itoh b,*2, Shunichi Fukuzumi a,*3
a Department of Materials and Life Science, Graduate School of Engineering, Osaka Uni6ersity, 2-1 Yamada-oka,
Suita, Osaka 565-0871, Japan
b Department of Chemistry, Graduate School of Science, Osaka City Uni6ersity, 3-3-138 Sugimoto, Sumiyoshi-ku, Osaka 558-8585, Japan
c Institute for Molecular Science, Myodaiji, Okazaki 444-8585, Japan
Received 22 September 1999; accepted 2 December 1999
Abstract
Model complexes of the active site of galactose oxidase (GAO) have been developed using a new ligand carrying two different
phenol groups, N-[(2-hydroxy-3-methylthio-5-tert-butylphenyl)methyl]-N-[(2-hydroxy-3,5-di-tert-butylphenyl)methyl]-2-(2-
pyridyl)ethylamine (L1H2). Deprotonated ligand L12− forms a dimeric Cu(II) complex, [Cu(II)2(L12−)2], in the solid state, the
structure of which has been determined by X-ray crystallographic analysis. The dimeric Cu(II)–diphenolate complex can be
converted into the monomeric complex, [Cu(II)(L12−)(X)] (X=py, AcO, and PhCH2OH), in solution by adding exogenous
ligands such as pyridine (py), acetate (AcO−), or benzyl alcohol (PhCH2OH). The structure and physicochemical properties
(UV–Vis, ESR, redox potential) of [Cu(II)(L12−)(X)] have been explored as a model for the resting state of the enzyme.
One-electron oxidation of [Cu(II)(L12−)(py)] and [Zn(II)(L12−)(py)] by (NH4)2[CeIV(NO3)6] (CAN) yielded the corresponding
−
phenoxyl radical/phenolate complexes, Cu(II)(L1 ) and Zn(II)(L1 −), respectively, which have also been characterized by
UV–Vis, resonance Raman, and ESR. The structure, physicochemical properties and reactivities of the diphenolate and phenoxyl
radical/phenolate complexes of L1H2 are compared to those of the corresponding monophenolate and monophenoxyl radical
complexes in order to obtain further insight into the role of Tyr 495 in the native enzyme. © 2000 Elsevier Science S.A. All rights
reserved.
’
’
Keywords: Galactose oxidase; Copper complexes; Zinc complexes; Phenoxyl radical; Cofactor
1. Introduction
Mechanistic studies on the enzymatic reaction have
shown that the Cu(II) complex of the phenoxyl radical
derived from Tyr 272 is the active species that oxidizes
primary alcohols to the corresponding aldehydes by the
following mechanism: (i) deprotonation of the alcohol
substrate by the axial tyrosinate ligand (Tyr 495); (ii)
inner-sphere electron transfer from the deprotonated
substrate to Cu(II); and (iii) hydrogen atom abstraction
of the resulting ketyl radical by the phenoxyl radical of
Tyr 272 (the ordering of electron transfer and hydrogen
atom abstraction steps could be reversed) [3–5].
Galactose oxidase (GAO) is a mononuclear copper
enzyme that catalyzes two-electron oxidation of pri-
mary alcohols to the corresponding aldehydes coupled
with the reduction of O2 to H2O2 (Eq. (1)) [1–4].
Structural details of the enzyme active site of GAO
have been disclosed recently by X-ray crystallographic
analysis [2], showing that the equatorial tyrosine (Tyr
272) is modified to make a covalent bond with the
adjacent cysteine (Cys 228) as illustrated in Scheme 1.
RCH2OH+O2“RCHO+H2O2
(1)
1*Corresponding author.
Model studies of the active site of GAO have recently
been performed to provide insight into the structure
and physicochemical properties of the inactive form
2*Corresponding author.
3*Corresponding author. Tel.: +81-6-6879 7368; fax: +81-6-6879
7370
0020-1693/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved.
PII: S0020-1693(99)00579-4

M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
623
Scheme 1.
[Cu(II)–phenolate state] and the active form [Cu(II)–
phenoxyl radical state] of the enzyme [6–12]. Efficient
catalytic oxidation of alcohols by Cu(II)–phenoxyl rad-
ical complexes has also been achieved by mimicking the
enzymatic functions in model systems [9,13]. Further-
more, detailed kinetic analysis of the oxidation of ben-
zyl alcohols by Cu(II)–phenoxyl radical complexes has
shown that the redox cycle between Cu(I) and Cu(II) as
well as the interconversion between the phenol and
phenoxyl radical states are very important for the effi-
cient two-electron oxidation of alcohols at a mono-
nuclear copper active site of GAO [6c,d]. However,
little attention has been focused on the catalytic role of
the axial tyrosinate ligand (Tyr 495) in model systems,
although some Cu(II) complexes with diphenol ligands
have been reported so far [7b,8,9,13]. In order to ad-
Chart (1)
2. Experimental
2.1. General
Reagents and solvents used in this study, except for
the ligand and the complexes, were commercial prod-
ucts of the highest available purity and were further
purified by standard methods, if necessary [14]. IR
spectra were recorded on a Shimadzu FTIR-8200PC
and UV–Vis spectra on a Hewlett Packard 8452A or a
Hewlett Packard 8453 photo diode array spectrophoto-
4
1
dress this issue, we have developed a new ligand L1H2
meter. H NMR spectra were recorded on a Jeol FT
carrying two different phenol groups as shown in Chart
(1). Ligand L1H2 is a hybrid of ligands L2H and L3H,
which we have developed in previous studies [6c,d].
Thus, a comparison of the structure and physicochemi-
cal properties of the Cu(II) and Zn(II) complexes of
L1H2 with those of the Cu(II) and Zn(II) complexes of
L2H and L3H, in both the phenolate and the phenoxyl
radical states, will give us valuable insight into the
different roles of the tyrosine residues (Tyr 272 and Tyr
495) in enzymatic systems. In this study, we have
examined the structure and physicochemical properties
of the diphenolate complexes as well as the stability and
spectroscopic features of the Cu(II) and Zn(II) com-
plexes of the phenoxyl radical/phenolate species derived
from L1H2.
NMR GSX-400 or a Jeol FT NMR EX-270 spectrome-
ter. ESR spectra were recorded on a Jeol JES-ME-2X
spectrometer. The g values were determined using a
Mn(II) marker as a reference. Computer simulation of
the ESR spectra was carried out using ESRAII version
1.01 (Calleo Scientific) on a Macintosh personal com-
puter. Mass spectra were recorded on a Jeol JNX-
DX303 HF mass spectrometer or
a
Shimadzu
GCMS-QP2000 gas chromatograph–mass spectrome-
ter.
2.2. Electrochemical measurement
The cyclic voltammetry and second harmonic ac
voltammetry (SHACV) measurements were performed
on a BAS 100B electrochemical analyzer or an ALS
electrochemical analyzer (model 600) in deaerated
CH3CN containing 0.10 M NBu4PF6 as supporting
electrolyte. The Pt and Au working electrodes (BAS)
were polished with BAS polishing alumina suspension
and rinsed with acetone before use. The counter elec-
trode was a platinum wire. The measured potentials
were recorded with respect to an Ag/AgNO3 (0.01 M)
reference electrode. The E1/2 values (versus Ag/AgNO3)
4 In order to distinguish the different states of the cofactor moiety
−
’
of the ligands more clearly, the symbols L1H2, L12−, and L1 are
used to denote the diphenol, diphenolate, and phenoxyl radical/phe-
nolate forms, respectively.

624
M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
are converted to values versus SCE by adding 0.29 V [15].
All electrochemical measurements were carried out under
nitrogen at atmospheric pressure.
1H), 8.52 (ddd, J=0.8, 1.6, 4.8 Hz, 1H). HRMS m/z:
330.1758 [M+]. Calc. for C19H26N2OS: 330.1766.
2.5.2. N-[(2-Hydroxy-3-methylthio-5-tert-
2.3. Visible resonance Raman measurement
butylphenyl)methyl]-N-[(2-hydroxy-3,5-di-tert-butyl-
phenyl)methyl]-2-(2-pyridyl)ethylamine (L1H2)
The 413.1 nm line of a Kr+ laser (model 2060 Spectra
Physics) was used as the exciting source. Visible reso-
nance Raman scattering was detected with a CCD
detector (model CCD3200, Astromed) attached to a 1 m
single polychromator (model MC-100DG, Ritsu Oyo
Kogaku). The slit width and slit height were set at 200
mm and 10 mm, respectively. Wavenumber ranges per
channel were 0.9 (Kr+ laser) and 0.5 cm−1 (dye laser).
The laser power used was 2.5 mW at the sample point.
All measurements were carried out at −60°C with a
spinning cell (1000 rpm). Raman shifts were calibrated
with indene and the accuracy of the peak positions of the
To a solution of 1 (1.47 g, 4.45 mmol) in CH3OH (5
ml) was added a suspension of paraformaldehyde (0.30
g, 10 mmol) in CH3OH (5 ml). The mixture was refluxed
for 1 h under N2. A methanol solution of 2,4-di-tert-
butylphenol (2.06 g, 10 mmol) was then added to the
mixture, and the resulting solution was further refluxed
for 16 h. After evaporation of the solvent, ligand L1H2
was obtained as a white powder by SiO2 column chro-
matography (CHCl3) in 30% yield. IR (KBr disk): 3390
cm−1 (OH). 1H NMR (CDCl3, 270 MHz): l 1.26 (s, 9H),
1.27 (s, 9H), 1.33 (s, 9H), 2.36 (s, 3H), 2.88 (t, J=7.3
Hz, 2H), 3.13 (t, J=7.3 Hz, 2H), 3.80 (s, 2H), 3.82 (s,
2H), 6.85 (d, J=2.2 Hz, 1H), 7.06–7.17 (m, 4H), 7.29
(d, J=2.2 Hz, 1H), 7.58 (td, J=1.7, 7.8 Hz, 1H), 8.58
(d, J=0.7, 5.0 Hz, 1H). EI MS (pos). m/z 548 [M+].
Anal. Calc. for C34H48N2O2S1: C, 74.41; H, 8.82; N, 5.10.
Found: C, 74.30; H, 8.75; N, 4.94%.
Raman bands was 91 cm−1
.
2.4. X-ray structure determination
X-ray diffraction data were collected with a Rigaku
RAXIS-CS imaging plate two-dimensional area detector
using graphite-monochromated Mo Ka radiation (u=
2.5.3. [Cu(II)2(L12−)2]
To a solution of ligand L1H2 (252 mg, 0.46 mmol) and
triethylamine (132 ml, 0.92 mmol) in CH3OH (50 ml) was
added Cu(ClO4)2·6H2O (170 mg, 0.46 mmol), and the
mixture was refluxed for 3 h. The resulting solution was
poured into excess water to provide a dark-violet solid
that was collected by filtration, washed with ether, and
dried in vacuo to give [Cu(II)2(L12−)2] in 87% yield.
Single crystals (dark-brown prism) for the X-ray struc-
ture determination were obtained by recrystallization
from toluene–CH3OH–H2O. FAB MS: m/z 1221.6
[M++1]. A set of prominent peaks showing an isotope
distribution pattern consistent with the dimeric Cu(II)–
diphenolate complex was obtained. Anal. Calc. for
[Cu(II)2(L12−)2]·C7H8·H2O, C75H102N4O5S2Cu2: C,
67.69; H, 7.73; N, 4.21. Found: C, 67.83; H, 7.63; N,
4.41%.
,
0.71070 A) to 2qmax of 60.0°. The intensity was measured
by sealing samples in a glass capillary tube containing the
mother liquid. All the crystallographic calculations were
performed using the TEXSAN software package form the
Molecular Structure Corporation [TEXSAN: crystal struc-
ture analysis package, Molecular Structure Corp. (1985
& 1999)]. The crystal structure was solved by direct
methods and refined by full-matrix least-squares. All
non-hydrogen atoms and hydrogen atoms were refined
anisotropically and isotropically.
2.5. Synthesis
2.5.1. N-[(2-Hydroxy-3-methylthio-5-tert-
butylphenyl)methyl]-2-(2-pyridyl)ethylamine (1)
To a solution of 2-(2-pyridyl)ethylamine (2.44 g, 20
mmol) in CH3OH (10 ml) was added a suspension of
paraformaldehyde (0.9 g, 30 mmol) in CH3OH (5 ml).
The mixture was refluxed for 1 h under N2. A solution
of 2-methylthio-4-tert-butylphenol [6d] (3.92 g, 20 mmol)
in CH3OH (5 ml) was then added to the mixture, and the
resulting solution was further refluxed for 24 h. After
evaporation of the solvent, the expected product 1 was
obtained as a yellow powder by SiO2 column chromatog-
raphy (CHCl3–ethyl acetate) in 39% yield. IR (KBr disk):
3400 cm−1 (OH and NH). 1H NMR (CDCl3, 400 MHz):
l 1.28 (s, 9H), 2.43 (s, 3H), 3.06 (t, J=7.6 Hz, 2H), 3.21
(t, J=7.6 Hz, 2H), 4.03 (s, 2H), 6.78 (d, J=2.4 Hz, 1H),
7.08 (d, J=2.4 Hz, 1H), 7.10 (ddd, J=0.8, 4.8, 7.6 Hz,
1H), 7.17 (d, J=7.6 Hz, 1H), 7.58 (td, J=1.6, 7.6 Hz,
2.5.4. [Zn(II)2(L12−)2]
This complex was prepared in a similar manner as used
for the preparation of the copper complex, with
Zn(ClO4)2·6H2O instead of Cu(ClO4)2·6H2O, in 95%
yield. FAB MS: m/z 1223.6 [M++1]. A set of prominent
peaks showing an isotope distribution pattern consistent
with the dimeric Zn(II)–diphenolate complex was ob-
tained. Anal. Calc. for [Zn(II)2(L12−)2], C68H92N4O4S2-
Zn2: C, 66.91; H, 7.60; N, 4.59. Found: C, 66.78, H, 7.54,
N, 4.63%.
2.5.5. Generation of phenoxyl radical species
The phenoxyl radical species of the Cu(II) and the
Zn(II) complexes were generated in situ by adding a

M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
625
Table 1
CH3CN solution of an equimolar amount of
Summary of X-ray crystallographic data
(NH4)2[CeIV(NO3)6] (CAN, 5.0×10−4 M) into a deaer-
ated 9:1 THF–CH3CN solution of [M(II)(L12−)(py)]
(5.0×10−4 M) (M=Cu or Zn) in a UV cell (1 cm
path length, sealed tightly with a silicon rubber cap) at
−60°C.
Compound
[Cu(II)2(12−)2]·C6H5CH3
Empirical formula
Formula weight
Crystal system
Space group
C75H100N4O4S2Cu2
1312.85
monoclinic
P21/a (no. 14)
18.169(4)
22.295(5)
19.594(10)
105.29(3)
7656(8)
,
a (A)
3. Results and discussion
,
b (A)
,
c (A)
i (°)
3.1. Ligand synthesis
3
,
V (A )
Z
4
The new ligand L1H2, which has two different phenol
groups, was prepared by the double Mannich reaction
on 2-(2-pyridyl)ethylamine with 2-methylthio-4-tert-
butylphenol followed with 2,4-di-tert-butylphenol as
indicated in Scheme 2. Thus, the reaction of 2-(2-
pyridyl)ethylamine and 2-methylthio-4-tert-butylphenol
in the presence of paraformaldehyde in refluxing
CH3OH for 24 h gave the monophenol derivative 1, to
which another phenol group was introduced by the
second Mannich reaction with 2,4-di-tert-butylphenol
and paraformaldehyde in a similar manner.
F(000)
Dcalc (g cm−3
T (°C)
Crystal size (mm)
v (Mo Ka) (cm−1
Diffractometer
Radiation
2800.00
1.139
–50
0.5×0.5×0.5
6.56
RAXIS-CS
Mo Ka (0.71070 A)
60.0
)
)
,
2qmax (°)
No. of reflections measured
No. of reflections observed
[I\3|(I)]
11 867
11 714
No. of variables
1149
0.079
0.107
2.61
a
R
b
Rw
3.2. Synthesis of Cu(II) and Zn(II) complexes
Goodness of fit c
a R=SꢀFoꢁ−ꢁFcꢀ/SꢁFoꢁ.
The addition of Cu(ClO4)2·6H2O to the deprotonated
ligand L12− in CH3OH led to the formation of a
copper(II) complex that crystallized immediately as a
dimeric form, as in the case of ligand L2H [6c,d]. A
single crystal (dark-violet prism) for X-ray structure
determination was obtained by recrystallization from a
mixed solvent system of toluene–CH3OH–H2O. Crys-
tallographic data and selected bond distances and an-
gles for the X-ray structure are summarized in Tables 1
b [Sw(ꢁFoꢁ−ꢁFcꢁ)2/SwFo2]1/2; w=1/|2(Fo).
c [Sw(ꢁFoꢁ−ꢁFcꢁ)2/(No−Nv)]1/2
.
and 2, with an ORTEP drawing shown in Fig. 1. The
dimeric Cu(II) complex can be converted into the corre-
sponding monomeric complex by adding an external
ligand such as pyridine in solution (vide infra). The
Zn(II) complex of L12− was also isolated as a dimeric
Table 2
Selected bond lengths (A) and angles (°) of [Cu(II)2(12−)2]·C6H5CH3
,
Cu(1)ꢀO(1)
Cu(1)ꢀO(4)
Cu(1)ꢀN(4)
Cu(2)ꢀO(2)
Cu(2)ꢀN(1)
S(1)ꢀC(28)
S(2)ꢀC(2)
1.970(4)
1.881(5)
2.123(6)
1.962(4)
2.102(6)
1.760(9)
1.761(8)
Cu(1)ꢀO(2)
Cu(1)ꢀN(3)
Cu(2)ꢀO(1)
Cu(2)ꢀO(3)
Cu(2)ꢀN(2)
S(1)ꢀC(75)
S(2)ꢀC(32)
2.035(5)
2.190(7)
2.048(5)
1.880(4)
2.197(7)
1.77(1)
1.771(9)
O(1)ꢀCu(1)ꢀO(2)
O(1)ꢀCu(1)ꢀN(3)
O(2)ꢀCu(1)ꢀO(4)
O(2)ꢀCu(1)ꢀN(4) 146.1(2)
O(4)ꢀCu(1)ꢀN(4)
O(1)ꢀCu(2)ꢀO(2)
O(1)ꢀCu(2)ꢀN(1) 146.6(2)
O(2)ꢀCu(2)ꢀO(3) 162.6(2)
O(2)ꢀCu(2)ꢀN(2)
O(3)ꢀCu(2)ꢀN(2) 102.3(2)
Cu(1)ꢀO(1)ꢀCu(2) 92.4(2)
C(28)ꢀS(1)ꢀC(75) 102.8(6)
74.8(2)
89.8(2)
89.3(2)
O(1)ꢀCu(1)ꢀO(4)
O(1)ꢀCu(1)ꢀN(4)
O(2)ꢀCu(1)ꢀN(3)
O(4)ꢀCu(1)ꢀN(3)
N(3)ꢀCu(1)ꢀN(4)
O(1)ꢀCu(2)ꢀO(3)
O(1)ꢀCu(2)ꢀN(2)
O(2)ꢀCu(2)ꢀN(1)
O(3)ꢀCu(2)ꢀN(1)
N(1)ꢀCu(2)ꢀN(2)
Cu(1)ꢀO(2)ꢀCu(2)
C(2)ꢀS(2)ꢀC(32)
161.4(2)
93.1(2)
117.5(2)
106.3(2)
93.6(2)
89.8(2)
118.5(2)
93.9(2)
94.9(2)
92.8(2)
93.0(2)
102.5(5)
94.9(2)
74.7(2)
92.2(2)
Scheme 2.

626
M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
Fig. 1. ORTEP drawing of [Cu(II)2(L12−)2]·C6H5CH3. The solvate molecule and hydrogen atoms are omitted for clarity.
form, [Zn(II)2(L12−)2], in a similar manner using
Zn(ClO4)2·6H2O instead of Cu(ClO4)2·6H2O.
The copper–copper separation is 2.90 A. There is no
,
coordinative interaction between Cu(II) and sulfur
atoms of the methylthio groups, as is the case for the
enzymatic system (the distances Cu(1)ꢀS(1), Cu(1)ꢀS(2),
Cu(2)ꢀS(1), and Cu(2)ꢀS(2) are 4.57, 4.04, 3.92, and
3.3. Crystal structure of the dimeric Cu(II)–diphenolate
complex
,
4.57 A, respectively). The dihedral angles of the
The crystal lattice of the Cu(II) complex consists of a
dinuclear copper(II) complex and one toluene molecule
from the recrystallization solvent. The crystal structure
of the dimeric copper(II) complex of L12− shown in
Fig. 1 resembles the structures of the dimeric Cu(II)
complex of duncamine (dnc) ligand {(2-pyridylme-
thyl)[(2 - hydroxy - 3,5 - dimethylphenyl)methyl][(2hy-
droxy-5-methyl-3-methylthiophenyl)methyl]amine} re-
ported by Whittaker et al. [7b] and of the dimeric
Cu(II) complex of the monophenol ligand L2H,
[Cu(II)2(L2−)2](PF6)2 [6d]. Both Cu(II) centers of
[Cu(II)2(L12−)2] have a distorted square pyramidal
structure with basal planes consisting of the tertiary
amine nitrogen {N(1) and N(3)} and three phenolate
oxygen atoms {O(1), O(3), and O(2) or O(4)}, two of
which {O(1) and O(3)} in the phenolate ring with the
2-methylthio group act as the bridges for both copper
ions. The pyridine nitrogens {N(2) or N(4)} are the
axial ligands for both copper ions; the ~ values at the
Cu(1) and Cu(2) centers are 0.27 and 0.26, respectively.
methylthio substituents and the phenol rings defined by
C(7)ꢀS(1)ꢀC(6)ꢀC(5) and C(41)ꢀS(2)ꢀC(40)ꢀC(39) are
10.5 and 9.3°, respectively.
3.4. Characterization of the diphenolate complex
(model complex for the resting state of GAO)
3.4.1. UV–Vis
The dimeric copper(II) complex, [Cu(II)2(L12−)2], ex-
hibits a strong absorption band at 453 nm (m=2000
M−1 cm−1 per Cu ion) together with a small band at
680 nm (m=200 M−1 cm−1 per Cu ion) in CH3CN.
Those absorption bands have been assigned to the
phenolate-to-metal charge transfer (LMCT band) and
the ligand field excitation (Cu(II) d–d band), respec-
tively [6–8]. It is interesting to note that umax of the
LMCT band of [Cu(II)2(L12−)2] (453 nm) is very close
to that of the inactive form of GAO (451 nm) [3e], but
is considerably smaller than those of the monopheno-
late complexes [Cu(II)2(L2−)2]2+ (522 nm) and

M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
627
[Cu(II)(L3−)]+ (559 nm) [6c,d]. The higher energy of
the LMCT feature in the diphenolate complexes com-
pared with the monophenolate compounds may be due
to a greater electron density at the Cu(II) ion induced
by the second phenolate donor in the diphenolate com-
plexes. A similar tendency has been reported in Tol-
man’s TACN (triazacyclononane) ligand system [8b].
The Zn(II) complex, on the other hand, does not show
any characteristic absorption band in the visible region
(see, Fig. 5(B)).
Ka value for the titration of [Cu(II)2(L2−)2]2+ with
pyridine has been determined as 0.13 M−1 in a similar
manner [6d]. The significantly larger dissociation con-
stant of [Cu(II)2(L12−)2] compared with [Cu(II)2-
(L2−)2]2+ can be attributed to the steric hindrance of
the tert-butyl group in L12−, destabilizing the dimeric
form of the Cu(II) complex. The Ka values for the
titration of [Cu(II)2(L12−)2] with acetate ions and
PhCH2OH have also been determined as 36.4 M−1 and
2.7×10−5
M
−1, respectively. It is noteworthy that
The dimer complex can be converted into the corre-
even the neutral weak ligand PhCH2OH can coordinate
to the metal center of the Cu(II) complex of ligand
L12−, since such coordinative interaction of PhCH2OH
has never been observed in the Cu(II) complexes of
ligands L2− and L3−. The coordination of PhCH2OH
to the copper center is reminiscent of the substrate
binding process in the enzymatic system.
sponding monomeric Cu(II) complex, [Cu(II)(L12−
)
(X)] (X=py, AcO, PhCH2OH), by adding an exoge-
nous ligand such as pyridine (py), acetate (AcO−), or
benzyl alcohol (PhCH2OH) into a CH3CN solution of
the dimer. As a representative example, the spectral
change observed upon addition of pyridine (0–0.02 M)
into a CH3CN solution of [Cu(II)2(L12−)2] (2.5×10−4
M) is shown in Fig. 2. The LMCT band of the dimer at
453 nm shifted to 466 nm (m=1360 M−1 cm−1) by the
dissociation into the monomer complex [Cu(II)(L12−)-
(py)], while the d–d band at 660 nm shifted to 710 nm
(m=100 M−1 cm−1). In the monomer complex as well,
the energy of the LMCT feature (466 nm) is signifi-
cantly higher than that of the monophenolate complex
[Cu(II)(L2−)(py)]+ (533 nm) [6d]. From the spectral
change, the dissociation constant (Ka=[monomer]2/
[dimer][py]2) was calculated as 27.4 M−1 using the
3.4.2. ESR
The dimeric Cu(II) complex of L12− in a frozen
CH3CN solution (77 K) is ESR silent (Fig. 3(A)) owing
to a strong antiferromagnetic coupling interaction be-
tween the two cupric ions in the m-bridged Cu2O2 core
as in the case of [Cu(II)2(L2−)2]2+ and [Cu(II)2(dnc)2]
[6d,7b]. The addition of pyridine into a CH3CN solu-
equation
(A−A0)2/(A−Aꢀ)=(1/4)Ka(md−mm)[py]2,
where md and mm are molar absorption coefficients of the
dimer and monomer, respectively. A plot of (A−A0)2/
(A−Aꢀ) versus [py]2 gave a straight line passing
through the origin, as shown in the inset of Fig. 2, from
which (1/4)Ka(md−mm) was obtained as the slope. The
Fig. 3. (A) ESR spectrum of [Cu(II)2(L12−)2] (2.5×10−4 M) in
CH3CN at 77 K; microwave frequency 9.180 GHz, modulation
frequency 100 kHz, modulation amplitude 10 G, microwave power 3
mW. (B) ESR spectrum of [Cu(II)(L12−)(py)] generated from the
dimer (2.5×10−4 M) by adding pyridine (2.0×10−2 M) in CH3CN
at 77 K; microwave frequency 9.193 GHz, modulation frequency 100
kHz, modulation amplitude 10 G, microwave power 3 mW. (C) The
computer simulation spectrum with the parameters g1=2.262, g2=
2.071, g3=2.028, A1=186, A2=25, and A3=18 G.
Fig. 2. Spectral change of the titration of [Cu(II)2(L12−)2] (2.5×
10−4 M) by pyridine (0−0.02 M) in CH3CN. Inset: plot of (A−
A0)2/(A−Aꢀ) vs. [pyridine]2 for the titration.

628
M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
turned out not to be very stable at ambient temper-
ature, which is in contrast with the fact that the Cu(II)
’
and Zn(II) complexes of monophenoxyl radicals L2
’
and L3 are much more stable at room temperature
[6c,d]. The redox potentials of [Cu(II)(L12−)(py)] are
fairly similar to those of [Cu(II)(dnc)(py)] (E11/2=0.12 V
and E12/2=0.46 V versus Ag/AgNO3 in CH3CN) [7b]
t
and of Tolman’s diphenolate complex [Cu(II)(L Bu )]
4
t
(L Bu =diphenolate of 1,4-bis(2-hydroxy-3,5-di-tert-
4
butylbenzyl) - 7 - isopropyl - 1,4,7 - triazacyclononane)
(E11/2=0.50 V and E12/2=0.78 V versus SCE in CH2Cl2
thus, E11/2=0.21 V and E12/2=0.49 V versus Ag/
AgNO3) [8b]. The E11/2 value of [Cu(II)(L12−)(py)] is
also close to that of [Cu(II)(L2−)(AcO)] (0.14 V versus
Ag/AgNO3) in CH3CN [6d]. The slight differences in
the redox potentials among these complexes may reflect
the differences in the geometry at the copper center, but
there seems to be no relation between the radical stabil-
ity and the redox potential.
Fig. 4. Cyclic voltammogram of [Cu(II)(L12−)(py)] generated from
[Cu(II)2(L12−)2] (1.0×10−3 M) by adding pyridine (8.0×10−2 M)
in CH3CN containing 0.1 M NBu4PF6 at −40°C; working electrode
Pt, counter electrode Pt wire, reference electrode Ag/0.01 M AgNO3,
scan rate 0.1 V s−1
.
3.5. Spectroscopic characterization of the phenoxyl
radical complexes (model complex for the acti6e
form of GAO)
tion of [Cu(II)2(L12−)2] resulted in a drastic increase in
the ESR signal intensity (Fig. 3(B)). This indicates the
formation of the monomer Cu(II) complex
[Cu(II)(L12−)(py)] as demonstrated by the UV–Vis
titration shown in Fig. 2. Computer simulation of the
spectrum (Fig. 3(C)) provided the ESR parameters
g1=2.262, g2=2.071, g3=2.028, A1=186, A2=25
and A3=18 G, which are consistent with a d
As suggested by the electrochemical studies of
[M(II)(L12−)(py)], phenoxyl radical complexes of the
new ligand are unexpectedly unstable at ambient tem-
perature. Thus, the oxidation of the diphenolate com-
plexes by (NH4)2[Ce(IV)(NO3)6] (CAN) was examined
at a lower temperature (−60°C). The addition of an
equimolar amount of a CH3CN solution of CAN into a
9:1 THF–CH3CN solution of [Cu(II)(L12−)(py)]
(5.0×10−4 M) at −60°C resulted in an immediate
increase of a strong absorption band at 418 nm (m=
2100 M−1 cm−1) and a small one at 720 nm (m=100
M−1 cm−1) together with a decrease in the absorption
band of the diphenolate complex at 466 nm (Fig. 5(A)).
A characteristic absorption band at 398 nm (m=1500
2
2
−y
ground state with a minor rhombic perturbation.x
3.4.3. Redox potential
Two quasi-reversible redox couples of [Cu(II)(L12−)-
(py)] were only obtained at a low temperature (− 40°C)
at E11/2=0.19 V versus Ag/AgNO3 (0.48 V versus SCE)
and E12/2=0.42 V versus Ag/AgNO3 (0.71 V versus
SCE) in CH3CN 5 (see Fig. 4). These redox couples may
correspond to the sequential one-electron oxidation of
the diphenolate (L12−) complex to the phenoxyl radi-
M−1 cm−1
) was obtained in the oxidation of
−
−
[Zn(II)(L12−)(py)] by CAN under the same experimen-
tal conditions (Fig. 5(B)). The resulting absorption
bands at around 400 nm are the characteristic feature
of phenoxyl radical species commonly observed both in
enzymatic systems and in model complexes [3,6–
10,12,13], indicating formation of the phenoxyl radical
complexes as in the case of the CAN oxidation of
[M(II)2(L2−)2]2+ and [M(II)(L3−)(CH3CN)]+ (M=Cu
or Zn) [6c,d]. There is, however, a significant difference
in the absorption band in the NIR (near IR) region
’
’
cal/phenolate (L1 ) complex and that of L1 to the
2
’
diphenoxyl radical (L1 ) complex, respectively.
[Zn(II)(L12−)(py)], however, gave only an irreversible
oxidation peak at around 0.50 V versus Ag/AgNO3
(0.79 V versus SCE) even at the low temperature
(−40°C) 6. Thus, the phenoxyl radical complexes
5 At a higher temperature, only irreversible oxidation peaks were
observed under otherwise identical experimental conditions.
6 The E11/2 of [Zn(II)(L12−)(py)] was determined by the second
harmonic ac voltammetric (SHACV) method as 0.23 V versus Ag/
AgNO3 (0.52 V versus SCE). It has been well demonstrated that the
SHACV method provides a better approach to direct evaluation of
one-electron redox potentials in the presence of a follow-up chemical
reaction [16]. Well-defined symmetrical SHACV traces were obtained
for the one-electron oxidation of [Zn(II)(L12−)(py)] in CH3CN as
shown in Supporting Information (S1), in which the intersection with
the dc potential axis corresponds to the one-electron oxidation poten-
tial.
(above 800 nm) among the phenoxyl radical complexes
+
’
so far reported. For example, [M(II)(L2 )(NO3)]
(M=Cu or Zn) [6c,d] and Halcolm’s model complexes
[12] exhibit strong and broad absorption bands above
800 nm, which is closely related to that of the active
form of GAO [3]. However, most of the other phenoxyl
radical complexes show a relatively weak absorption

M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
629
Fig. 6. Resonance Raman spectra of the Cu(II)(L1 −) (bottom) and
’
Zn(II)(L1 −) (top) generated by CAN oxidation of [M(II)(L12−)(py)]
’
(M=Cu or Zn) in 9:1 THF–CH3CN at −60°C; excitation wave-
length: 413.1 nm.
also shows Raman bands at nearly the same positions
(Fig. 6). These Raman bands disappeared completely
when the solution temperature was raised to room
temperature. This is due to decomposition of the phe-
noxyl radical complexes at the higher temperature, as
suggested by the irreversibility in the electrochemical
measurements. Although detailed assignment of these
peaks has yet to be accomplished, the ꢁ1545 cm−1
and ꢁ1640/ꢁ1660 cm−1 features in the resonance
Raman spectra may be due to the modes w7a% and w8a%,
which predominantly include the CꢀO stretching and
the CorthoꢀCmeta stretching, respectively [8b,10]. If so,
the frequencies of these bands are somewhat higher
than those for the other metal-coordinated phenoxyl
radical species [6d,8b,10,13a], suggesting a larger con-
tribution of the quinonoid canonical form in this par-
ticular case [10b].
Fig. 5. (A) UV–Vis spectra of [Cu(II)(L12−)(py)] (dotted line) (5.0×
10−4 M) and of the corresponding phenoxyl radical/phenolate com-
plex, Cu(II)(L1 −) (solid line) generated by the CAN oxidation
’
(5.0×10−4 M) in 9:1 THF–CH3CN at −60°C. (B) UV–Vis spectra
The oxidation of [Cu(II)(L12−)(py)] by CAN at the
low temperature caused a significant decrease in the
ESR signal of Cu(II) (more than 90% decrease in peak
area), consistent with the formation of the Cu(II)–
monophenoxyl radical/monophenolate species. A mag-
netic coupling between the S=1/2 Cu(II) ion and the
S=1/2 phenoxyl radical will make the compound ESR
silent, although the type of coupling (antiferromagnetic
or ferromagnetic) is unclear at present [8b,13b]. On the
other hand, the solution ESR spectrum of the Zn(II)
complex measured at −80°C in CH3OH (Fig. 7(A))
exhibits an isotropic signal at g=2.0046 7. The g value
of [Zn(II)(L12−)(py)] (dotted line) (5.0×10−4 M) and of the corre-
sponding phenoxyl radical/phenolate complex, Zn(II)(L1 −) (solid
’
line) generated by CAN oxidation (5.0×10−4 M) in 9:1 THF–
CH3CN at −60°C.
band in the shorter wavelength region (600–700 nm)
[8–10,13]. Thus, the present model complexes belong to
the latter category, although the reason for such a
difference in the NIR features among the phenoxyl
radical complexes has yet to be clarified.
The formation of the phenoxyl radical species is also
supported by the resonance Raman spectra of the
solutions resulting from the CAN oxidation of
[M(II)(L12−)(py)] at −60°C. The Cu(II)–phenoxyl
radical complexes display characteristic peaks at
ꢁ1543 cm−1 together with small peaks at ꢁ1638 and
ꢁ1662 cm−1 as shown in Fig. 6. The Zn(II) complex
7 Judging from the peak area of Fig. 7(A), the yield of radical
formation is relatively low (ꢁ1%). Since the phenoxyl radical com-
plex is highly unstable even at low temperature, a large portion of the
radical species decomposes during the sampling for ESR measure-
ment.

630
M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
of this signal is slightly smaller than those of the
cofactor radical (2.0055) of the apo-enzyme and of the
phenoxyl radicals derived from our cofactor model
compounds (2.0052–2.0060) [3c,6b]. Although resolu-
tion of the spectrum is not sufficient owing to the
instability of the compound, the shape of the ESR
in Scheme 3, where the hyperfine coupling constants
(hfc) determined by the computer simulation are also
indicated. It is obvious that the spin density at the
−
’
benzylic methylene group in L1
(aH=1.38 G)8 is
’
significantly smaller than that of L3 (aH=7.19 G) but
’
rather similar to that of L2 (aH=2.43 G) in the Zn(II)
spectrum in Fig. 7(A) is very similar to that of
complexes [6d]. It has been suggested that the smaller
spin density at the benzylic position is due to the spin
+
’
[Cu(II)(L2 )(NO3)] [6d] but rather different from that
+
’
’
of [Cu(II)(L3 )(NO3)] [6d]. This indicates that the odd
delocalization into the alkylthio group in L2 [6d]. This
can also be applied to the case of L1 −. The smaller hfc
’
electron of the radical species exists mainly on the
phenol ring having the methylthio substituent as shown
−
’
value of the methylthio group in L1 (1.86 G) and the
−
’
’
smaller g value of L1
(2.0046) compared with L2
(aSMe=3.72 G and g=2.0055) [6d], however, indicate
that the spin delocalization into the methylthio group
−
’
’
in L1 is not as large as in L2 . This may be a reason
−
’
in part for the instability of L1
.
3.6. Summary
In this study, we have developed model complexes of
the active site of galactose oxidase (GAO) using a new
ligand carrying two different phenol groups (L1H2) in
order to examine the effect of the second phenol group
on the physicochemical properties as well as the redox
reactivity of the Cu(II)–phenoxyl radical complexes.
The nature of the ligand to ligand charge transfer
’
between Tyr 272 (tyrosyl radical) and Tyr 495 (tyrosi-
nate) in the active form of GAO and the acid-base
catalysis by Tyr 495 in the catalytic cycle of GAO are
of particular interest, since these issues have yet to be
disclosed at a molecular level. Unfortunately, however,
the phenoxyl radical/phenolate species (L1 −) in the
’
Cu(II) and Zn(II) complexes were too unstable to ex-
amine their reactivity in the alcohol-oxidation reac-
tions. This instability of the phenoxyl radical complexes
is an unexpected result, since the phenoxyl radical
complexes of the corresponding monophenol ligands
(L2H and L3H) are fairly stable [6c,d]. The reason for
such a difference in stability remains to be understood.
The low temperature experiments below −60°C al-
lowed us to examine the spectroscopic features (UV–
Vis, resonance Raman, and ESR) of the Cu(II) and
Zn(II) complexes of phenoxyl radical/phenolate species
Fig. 7. Solution ESR spectrum (A) of Zn(II)(L1 −), generated by the
’
CAN oxidation of [Zn(II)(L12−)(CH3OH)] in CH3OH at −80°C;
microwave frequency 9.210 GHz, modulation frequency 100 kHz,
modulation amplitude 0.5 G, microwave power 5 mW, and its
computer simulation spectrum (B) using the parameters (g and hfc
values) presented in Scheme 3.
8 It has been reported that hfc of benzylic methylene protons of
phenoxyl radicals can be estimated by the angle-dependent Mc-
Connell-type relationship: aCꢀH=zC1B cos2 q, where aCꢀH is the hfc
of the methylene proton, zC1 is the spin density at the C1 position, B
is a constant equal to 162 Hz, and q is the dihedral angle defined in
figure 8 in reference [17]. Thus, the ratio of the hfc values of two
methylene protons (aH1/aH2) is proportional to cos2 q1/cos2 q2. If one
accepts the assumption that the dihedral angles of benzylic methylene
protons do not change upon one-electron oxidation, the aH1/aH2 ratio
can be calculated as 75 by using the dihedral angles obtained in the
X-ray structure of [CuI2I(L12−)2] (q1(a6)=30.5,q2(a6)=84.3). Such a
large value of aH1/aH2 is consistent with the large difference between
aH1 and aH2 (1.38 G versus ꢁ0 G) determined by computer simula-
tion of the ESR spectrum (Scheme 3).
Scheme 3.

M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
631
(L1 −). To our surprise, the NIR features observed in
’
References
’
’
[Cu(II)(L2 )(NO3)] and [Zn(II)(L2 )(NO3)] [6c,d] disap-
peared when the second phenol group was introduced
into the ligand. This result indicates that the electronic
structure of the excited state is greatly altered by the
second phenolate group. The NIR features of
[Cu(II)(L2 )(NO3)] and [Zn(II)(L2 )(NO3)] have been
assigned to the intramolecular charge transfer from the
benzene ring to the methylthio groups in the phenoxyl
radical group [6]. Thus, neither intramolecular charge
[1] (a) G. Avigad, D. Amaral, C. Asensio B.L. Horecker, J. Biol.
Chem. 237 (1962) 2736. (b) D.J. Kosman, in: R. Lontie (Ed.),
Copper Proteins and Copper Enzymes, CRC Press, Boca Raton,
FL, 1985, pp. 1–26.
[2] (a) N. Ito, S.E.V. Phillips, C. Stevens, Z.B. Ogel, M.J. Mc-
Pherson, J.N. Keen, K.D.S. Yadav, P.F. Knowles, Nature 350
(1991) 87. (b) N. Ito, S.E.V. Phillips, K.D.S. Yadav, P.F.
Knowles, J. Mol. Biol. 238 (1994) 794.
[3] (a) M.M. Whittaker, J.W. Whittaker, J. Biol. Chem. 263 (1988)
6074. (b) M.M. Whittaker, V.L. DeVito, S.A. Asher, J.W.
Whittaker, J. Biol. Chem. 264 (1989) 7104. (c) M.M. Whittaker,
J.W. Whittaker, J. Biol. Chem. 265 (1990) 9610. (d) M.M.
Whittaker, J.W. Whittaker, Biophys. J. 64 (1993) 762. (e) M.M.
Whittaker, P.J. Kersten, N. Nakamura, J. Sanders-Loehr, E.S.
Schweizer, J.W. Whittaker, J. Biol. Chem. 271 (1996) 681. (f)
M.M. Whittaker, D.P. Ballou, J.W. Whittaker, Biochemistry 37
(1998) 8426.
[4] (a) B.P. Branchaud, M.P. Montague-Smith, D.J. Kosman, F.R.
McLaren, J. Am. Chem. Soc. 115 (1993) 798. (b) R.M. Wachter,
B.P. Branchaud, Biochemistry 35 (1996) 14 425. (c) R.M.
Wachter, M.P. Montague-Smith, B.P. Branchaud, J. Am. Chem.
Soc. 119 (1997) 7742.
’
’
transfer nor ligand-to-ligand charge transfer occurs in
−
’
Cu(II)(L1 ) and Zn(II)(L1 −) complexes. This phe-
nomenon of the present model complexes is rather
similar to that observed in Stack’s diphenolate system,
although their model complexes of the phenoxyl radi-
cal/phenolate species are very stable [9]. It is interesting
to note that the odd electron of the phenoxyl radical
species exists mainly on the phenol ring having the
methylthio substituent as illustrated in Scheme 3. This
is consistent with our previous conclusion that the
phenoxyl radical is stabilized to a greater extent by the
electron-sharing conjugative effect by the alkylthio
group [6a,b]. Overall, it became clear that the electronic
structure as well as the stability and the reactivity of
phenoxyl radical complexes are greatly influenced by
the coordination geometry of the metal center.
’
[5] (a) K. Clark, J.E. Penner-Hahn, M.M. Whittaker, J.W. Whit-
taker, J. Am. Chem. Soc. 112 (1990) 6433. (b) K. Clark, J.E.
Penner-Hahn, M.M. Whittaker, J.W. Whittaker, Biochemistry
33 (1994) 12 553.
[6] (a) S. Itoh, K. Hirano, A. Furuta, M. Komatsu, Y. Ohshiro, A.
Ishida, S. Takamuku, T. Kohzuma, N. Nakamura, S. Suzuki,
Chem. Lett. (1993) 2099. (b) S. Itoh, S. Takayama, R. Arakawa,
A. Furuta, M. Komatsu, A. Ishida, S. Takamuku, S. Fukuzumi,
Inorg. Chem. 36 (1997) 1407. (c) S. Itoh, M. Taki, S. Takayama,
S. Nagatomo, T. Kitagawa, N. Sakurada, R. Arakawa, S.
Fukuzumi, Angew. Chem. 111 (1999) 2944; Angew. Chem., Int.
Ed. Engl. 38 (1999) 2774. (d) S. Itoh, M. Taki, S. Takayama, S.
Nagatomo, T. Kitagawa, N. Sakurada, R. Arakawa, S.
Fukuzumi, submitted for publication.
[7] (a) M.M. Whittaker, Y.-Y. Chuang, J.W. Whittaker, J. Am.
Chem. Soc. 115 (1993) 10 029. (b) M.M. Whittaker, W.R. Dun-
can, J.W. Whittaker, Inorg. Chem. 35 (1996) 382.
[8] (a) J.A. Halfen, V.G. Young, Jr., W.B. Tolman, Angew. Chem.
108 (1996) 1832. (b) J.A. Halfen, B.A. Jazdzewski, S. Mahapa-
tra, L.M. Berreau, E.C. Wilkinson, L. Que, Jr., W.B. Tolman, J.
Am. Chem. Soc. 119 (1997) 8217.
4. Supplementary material
A summary of the fundamental crystal data and
experimental parameters for structure determinations is
given in Table 1. The experimental details including
data collection, data reduction, and structure solution
and refinement as well as the atomic coordinates and
iso/Beq, anisotropic displacement parameters, and in-
tramolecular bond distances and angles have been de-
posited in the Supporting Information (S2).
B
[9] (a) Y. Wang, T.D.P. Stack, J. Am. Chem. Soc. 118 (1996)
13 097. (b) Y. Wang, J.L. DuBois, B. Hedman, K.O. Hodgson,
T.D.P. Stack, Science 279 (1998) 537.
[10] (a) A. Sokolowski, J. Mu¨ller, T. Weyhermu¨ller, R. Schnepf, P.
Hildebrandt, K. Hildengrand, E. Bothe, K. Wieghardt, J. Am.
Chem. Soc. 119 (1997) 8889. (b) R. Schnepf, A. Sokolowski, J.
Mu¨ller, V. Bachler, K. Wieghardt, P. Hildebrandt, J. Am. Chem.
Soc. 120 (1998) 2353.
Acknowledgements
The present study was financially supported in part
by a Grant-in-Aid for Scientific Research on Priority
Areas (Molecular Biometallics: 10129218 and 11116219;
Electrochemistry of Ordered Interface: 10131242 and
11118244; Creation of Delocalized Conjugated Elec-
tronic System: 10146232) from the Ministry of Educa-
tion, Science, Sports, and Culture of Japan and by
Research Fellowships of the Japan Society for the
Promotion of Science for Young Scientists (no. 02410).
We also thank Professor Yasushi Kai and his co-work-
ers, particularly, Ms Eiko Mochizuki of Osaka Univer-
sity, for their assistance in the X-ray measurements.
[11] M. Ruf, C.G. Pierpont, Angew. Chem., Int. Ed. Engl. 37 (1998)
1736.
[12] (a) M.A. Halcolm, L.M.L. Chia, X. Liu, E.J.L. McInnes, L.J.
Yellowlees, F.E. Mabbs, J.E. Davies, Chem. Commun. (1998)
2465. (b) M.A. Halcolm, L.M.L. Chia, X. Liu, E.J.L. McInnes,
L.J. Yellowlees, F.E. Mabbs, I.J. Scowen, M. McPartlin, J.E.
Davies, J. Chem. Soc., Dalton Trans. (1999) 1753.
[13] (a) P. Chaudhuri, M. Hess, U. Florke, K. Wieghardt, Angew.
Chem., Int. Ed. Engl. 37 (1998) 2217. (b) P. Chaudhuri, M.
Hess, T. Weyhermu¨ller, K. Wieghardt, Angew. Chem., Int. Ed.
Engl. 38 (1999) 1095. (c) P. Chaudhuri, M. Hess, J. Mu¨ller, K.
Hildenbrand, E. Bill, T. Weyhermu¨ller, K. Wieghardt, J. Am.
Chem. Soc. 121 (1999) 9599.

632
M. Taki et al. / Inorganica Chimica Acta 300–302 (2000) 622–632
[14] D.D. Perrin, W.L.F. Armarego, D.R. Perrin (Ed.), Purification
of Laboratory Chemicals, Pergamon, Elmsford, NY, 1966.
[15] K. Manna, K.K. Barnes, Electrochemical Reactions in Non-
aqueous Systems, Marcel Dekker, New York, 1990.
[16] (a) T.G. McCord, D.E. Smith, Anal. Chem. 41 (1969) 1423. (b)
A.M. Bond, D.E. Smith, Anal. Chem. 46 (1974) 1956. (c) M.R.
Wasielewski, R. Breslow, J. Am. Chem. Soc. 98 (1976) 4222. (d)
E.M. Arnett, K. Amarnath, N.G. Harvey, J.-P. Cheng, J. Am.
Chem. Soc. 112 (1990) 344. (e) S. Fukuzumi, M. Fujita, J.
Maruta, M. Chanon, J. Chem. Soc., Perkin Trans. 2 (1994) 1597.
[17] G.T. Babcock, M.K. El-Deeb, P.O. Sandusky, M.M. Whittaker,
J.W. Whittaker, J. Am. Chem. Soc. 114 (1992) 3727.
.
.