Could redox-switched binding of a redox-active ligand to a copper(II) centre drive a conformational proton pump gate? A synthetic model study

Article abstract of DOI:10.1002/chem.200390007

A proposal for a redox-linked conformational gate to proton translocation - a proton pump gate - based upon a transition-metal redox-switchable hemilabile ligand (RHL) system is made. Consideration of the requirements for such a system reveals copper-(II)

Full text of DOI:10.1002/chem.200390007

FULL PAPER
Could Redox-Switched Binding of a Redox-Active Ligand to a Copper(ii)
Centre Drive a Conformational Proton Pump Gate?A Synthetic Model Study
Zhicong He, Stephen B. Colbran,* and Donald C. Craig[a]
‡
Abstract: A proposal for a redox-linked
conformational gate to proton translo-
cation–a proton pump gate–based
upon a transition-metal redox-switch-
able hemilabile ligand (RHL) system is
made. Consideration of the require-
ments for such a system reveals copper-
(ii) to be the ideal metal centre. To test
the proposal and, thereby, to provide an
artificial proton pump gate, the copper
coordination chemistry of three tris-
(pyridylmethyl)amine (tpa) ligands with
one ™leg∫ (PY*) substituted at the
6-position of the pyridine ring by a
dimethoxyphenyl (L1), a hydroquinone
(H2L2) or a quinone (L3) substituent
has been investigated. Crystal structures
of sp-[Cu(k4N-L1)Cl]Cl ¥ 3H2O (sp ˆ
square pyramidal), sp-[Cu(k3N-H2L2)-
Cl2] and tbp-[Cu(k4N,kO-HL2)][PF6]
[Cu(L)(MeCN)n] (L ˆ L1, H2L2) dis-
play physicochemical properties consis-
tent with a ™dangling∫ PY* leg; from the
NMR spectra, the barriers to inversion
of the ligand amine donor for the CuI
complexes are estimated to be within
the range of about 30 45 kJmolÀ1. In
the CuII complexes, coordination of the
PY* leg is finely balanced and critically
depends on the nature of the PY*
substituent and the availability of po-
tential co-ligand(s). For example, tbp-
9.7 Â 102 and is at least 104-fold greater
‡
than that of tbp-[Cu(k4N-L3)Cl] . The
complex sp-[Cu(k3N-H2L2)Cl2] has
a
™dangling∫ PY* leg, in which an intra-
molecular OH(hydroquinone) ¥¥¥ N(pyr-
idine) hydrogen bond ™ties-up∫ the
pyridyl nitrogen atom, and reacts with
Br˘nsted bases to give tbp-[Cu(k4N,kO-
‡
HL2)] . Two-electron oxidation of sp-
[Cu(k3N-H2L2)Cl2] is linked to loss of
two protons and a conformational change,
‡
and affords tbp-[Cu(k4N-L3)Cl] . The
[Cu(k4N-L1)Cl] reacts cleanly with ClÀ
ions to afford sp-[Cu(k3N-L1)Cl2]; Vis/
[Cu(k3N-H2L2)Cl2] [Cu(k4N-L3)Cl]
‡
‡
system provides a first demonstration of
the critical step in the proposed proton
pumping cycle, namely a redox-driven
NIR spectrophotometric titrations sug-
gest the affinity of tbp-[Cu(k4N-L1)Cl]
‡
for ClÀ ion in dichloromethane is
and
proton-linked
conformational
change. The possible biological rele-
vance of this work to proton pumping
in cytochrome c oxidase is mentioned.
Keywords: bioinorganic chemistry ¥
molecular devices
N ligands ¥ proton transport
copper
¥
¥
(tbp ˆ trigonal
bipyramidal)
have
been determined. The CuI complexes
across the inner mitochondrial or a bacterial membrane
according to Equation (1).[3, 4] This is the last step in aerobic
respiration and accounts for the vast majority of oxygen
consumed by living organisms.
Introduction
Proton pumps are remarkable molecular machines capable of
translocating protons across a membrane against a proton
gradient,[1, 2] and are vitally important for all life since the
transmembrane proton gradient created by the proton
pumping proteins of the respiratory or photosynthetic path-
ways drives the synthesis of ATP, the energy currency of all
cells.[3] For example, cytochrome c oxidase, one inspiration for
this study, is respiratory complex IV and uses the free energy
released as oxygen is reduced by electrons supplied from
cytochrome c to pump protons against the proton gradient
‡
‡
out
8H ‡ O2 ‡ 4eÀ ! 4H
‡ 2H2O
(1)
in
The essential components of a proton pump include the
‡
proton-conducting (H ) channels, which convey protons
through the membrane, and the proton pumping ™gate∫, the
vital element that drives the unidirectional translocation of
protons.[2, 3] In real systems, for example proton pumping
‡
membrane proteins such as cytochrome c oxidase, the H
channels consist of (transient) chains of protonatable amino
acid side chains and bound water molecules along which the
protons can passively hop.[2, 3, 5] The proton pumping gate is
the site where the energy released by a chemical process is
used to drive the proton pumping; that is the proton pump
gate is the element that couples the chemical process to
vectorial proton transport. In cytochrome c oxidase, for
[a] Dr. S. B. Colbran, Z. He, D. C. Craig
School of Chemical Sciences
University of New South Wales
Sydney, NSW 2052 (Australia)
Fax : (‡61)2-9385-6141
E-mail: s.colbran@unsw.edu.au
Supporting information for this article is available on the WWW under
http://www.chemeurj.org or from the author.
116
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0116 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 1

116 129
example, the binuclear heme a3 ¥¥¥ CuB catalytic centre for
reduction of oxygen is also believed to be the proton pumping
gate.[6, 7] Although the structure of this centre is known to
atomic resolution[5, 8] and there is much information about the
states and intermediates in the catalytic cycle,[4, 9, 10] how
reduction of oxygen is actually coupled to proton pumping
remains a contentious issue that has occasioned much recent,
robust intellectual debate.[5, 9, 10] Two facts about cytochrome c
oxidase are mentioned here because they influenced this
work: Firstly, one of the three histidine ligands–which one is
membrane-spanning, artificial photosynthetic triad in the
presence of a lipid-soluble naphthaquinone.[17, 18] These devi-
ces, however, are not proton pumps, but photo-driven
Mitchell loops[3] that rely on diffusion of the lipid-soluble
naphthahydroquinone to transport the protons across the
vesicle membrane. Herein we outline a proposal for a simple
artificial redox-linked proton pump gate and describe our first
attempts to engineer a synthetic chemical system that
demonstrates the concept; the possible biological relevance
of our results is also mentioned.
I
unknown–to the CuB ion in the fully reduced state of
cytochrome c oxidase is more weakly bound than the others
and can be displaced by ClÀ ion;[11] secondly, one of the
histidine ligands (His240 in the numbering scheme for the
bovine enzyme) to CuB is covalently linked to a tyrosine
(Tyr244) side chain[5, 8]–the resulting imidazole-phenol co-
factor has recently been shown to be redox-active and is the
donor of one of the four electrons required to cleave oxygen
when it binds the heme a3 ¥¥¥ CuB active site.[12] Also essential
to redox-linked proton pumps–those driven by a redox
reaction such as cytochrome c oxidase–are one or more
electron transfer pathways,[13, 14] networks of appropriately
aligned and coupled covalent and hydrogen bonds and
through-space jumps, for delivery of electrons to the catalytic
centre where the redox reaction takes place.
Artificial proton pumps offer exciting prospects for new
technologies for energy conversion and storage.[1, 2] Great
recent progress has been made in the design, synthesis, and
understanding of artificial ion and proton-conducting chan-
nels[15] and of artificial electron-transfer pathways.[14, 16, 17] In
stark contrast, the synthesis and cycling of an artificial proton
pump gate are yet to be demonstrated.[1, 2] Perhaps closest is
the demonstration of proton transport across the lipid bilayer
membranes of artificial vesicles driven by illumination of a
Concept: transition-metal RHL complexes as redox-linked
proton pump gates
Figure 1 diagrammatically depicts a proposal for a proton
pump that has a transition-metal centre bound by a redox-
switchable hemilabile ligand (RHL)[19 22] as the proton pump
gate. RHLs are ligands in which the binding to a metal centre
of one of the donor groups is modulated by a covalently linked
redox centre. The essential features of this proton pump gate
are: 1) the discrete conformations for each redox-state of the
RHL transition-metal centre; 2) the proton-dependent redox
couple as the RHL redox-switch; 3) the fixed orientation for
the RHL transition-metal assembly, for example within a
membrane protein, with respect to the membrane across
which protons are to be pumped. Inductive withdrawal of
electron density upon oxidation of the RHL redox centre
weakens the binding of the covalently linked donor group
(Lstr ! Lwk), and it may be displaced from the metal centre by
other ancillary ligand(s), provided these are available (con-
formation I ! conformation II in Figure 1). Reduction of the
ligand redox centre increases the nucleophilicity of the donor
group (Lwk ! Lstr) that rebinds to the metal centre (confor-
mation II ! conformation I in Figure 1). The concept of
Figure 1. Cartoon of the proposed conformational proton pump cycle highlighting the role of the CuII RHL centre as the redox-linked proton pumping
‡
gate. Starting with the metal centre in conformation I (left-side), RHL-centred oxidation releases protons through the exit H channel and creates a weaker
donor group (Lwk) that is displaced by ancillary ligand(s) and swings away (the dashed arrow marked ™oxidation stroke∫) so that the CuII RHL centre adopts
‡
conformation II (right-side). RHL-centred reduction in conformation II is tightly linked to uptake of protons through the input H channel and creates a
stronger donor group (Lstr) that moves to rebind the CuII ion (the dashed arrow marked ™reduction stroke∫) thus regenerating I. Protons are only carried by
‡
the RHL during the ™reduction stroke∫, thereby leading to vectorial transport of protons from the input to the exit H channels. For further details, see the
text.
Chem. Eur. J. 2003, 9, No. 1
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0117 $ 20.00+.50/0
117

S. B. Colbran et al.
FULL PAPER
RHLs was first demonstrated by ourselves for phosphinoqui-
none/hydroquinone RHLs[19, 20] and by Mirkin and co-workers
for ferrocene/ferrocenium-substituted RHLs.[21] As the RHL
redox-switch in the proposed proton pumping gate is a
proton-dependent couple, oxidation and reduction are ac-
companied by proton release and uptake in two discrete
conformational states, one for output and one for input of
electrons and protons. The protons enter and leave the active
this would be dictated by the protein structure. For models,
we targeted copper complexes of new tris(pyridylmethyl)-
amine (tpa) ligands derivatised with proton-dependent
redox centres. There were two reasons for this choice.
Firstly, tpa and derivatives are flexible and usually form
copper complexes exhibiting k4N-coordination, but k3N-
coordination with one pyridyl leg unbound, ™dangling∫
away from the copper centre is also found, especially with
more sterically encumbered pyridyl groups.[26] Secondly,
with a view to later studies of the reactions of CuI
complexes with dioxygen, we noted that copper tpa
complexes are commonly used to model the copper
centres in dioxygen binding/utilising copper proteins,[27]
including cytochrome c oxidases,[28] and promote forma-
tion of end-on rather than side-on dioxygen adducts (the
CuB ion in the resting, oxidised state of cytochrome c
oxidase may be bound by end-on peroxide[5]).
‡
centre through two (or more) H channels. Redox-linked
movement of the hemilabile donor group connects the input
‡
and output H channels and, since only the reduced state of
the redox-switch carries protons, allows only unidirectional
movement of protons. As depicted the movement of the labile
ligand group is exaggerated, for in a real system proton
tunnelling at comparable energies to electron transfer is
unlikely over distances greater than 0.25 ä.[23] The labile
ligand group, therefore, need only move about 0.3 ä to gate
proton movement, with any remaining distance between input
and exit channels spanned by a (transitory) hydrogen-bonded
chain of protonatable groups (e.g. protein residues or water
molecules) along which protons passively hop. Electrons are
supplied to/from the redox centre by a separate electron
transfer pathway(s), and electron transfer may be coupled to
3) The RHL redox switch must exhibit a proton-dependent
redox couple that must perturb the binding of a ligand
donor group to the metal centre. Both partners of the (2eÀ,
‡
2H ) p-quinone p-hydroquinone redox couple are sta-
ble[19, 20, 29, 30] and, therefore, the (hydro)quinone group is
an ideal first choice for the RHL redox centre. Direct
attachment of the redox centre to a pyridyl ring of tpa
serves the dual functions of modulation of the binding
ability of the pyridyl ring and providing the steric bulk to
direct binding of this pyridyl ring to along the weak field
axis in a CuII complex. Chloride ion was chosen as the
coligand/counterion, in part, because of the evidence for
ClÀ ion displacing a histidine ligand to the CuB centre in
cytochrome c oxidases.[11]
‡
the movement of one or more protons along an H channel
(represented in Figure 1 by the crossover of the proton and
electron pathways). Cycling between conformers I and II
leads to vectorial translocation of protons linked to the redox
changes at the RHL transition-metal centre. An oxidising
substrate could oxidise the redox centre of the ligand, in which
case the electron-transfer pathway need only deliver elec-
trons, and the reduction product(s) of the substrate may be the
ancillary ligand(s) that displaces the oxidised form (Lwk) of the
ligand-donor group in conformer II. The principle remains the
same for any proton-dependent redox couple as the RHL
redox switch.
In sum these requirements led us to investigate the copper
coordination chemistry of H2L2, L3 and, for comparison, L1, as
reported herein.
Ligand
tpa
R
H
System requirements and choice of synthetic models
For proof of this concept, we sought to demonstrate for a
synthetic model system, a transition-metal complex of a
proton-dependent RHL, that the ligand redox switch could
control the binding of the hemilabile donor to the metal ion,
with oxidation and reduction of the ligand redox centre
accompanied by proton loss and uptake at different spatial
locations with respect to the metal centre. There are several
requirements for the transition metal and the proton-depend-
ent RHL that led to a target system.
MeO
HO
O
L1
N
OMe
OH
O
N
N
H2L2
N
R
L3
1) The metal ion should exhibit fast exchange reactions so
that the hemilabile donor group can rapidly cycle on/off.
The other RHL donor group(s) should remain strongly
bound to the metal. These conditions lead naturally to CuII
ion. Although CuII ion exhibits the largest formation
constants for the first d series divalent ions (the Irving
Williams series),[24] it also exhibits the fastest rates for
transition metal ions because Jahn Teller effects lead to
weak binding of ligand(s) along (at least) one axis.[25]
2) The RHL must be flexible and coordination of the ligand
group substituted by the redox-centre to the metal ion
should be finely balanced. In a proton pumping enzyme
Results
Syntheses: The ligands L1 and H2L2 were prepared as follows.
1,4-Dimethoxybenzene or 1,4-di(tetrahydropyranyloxy)ben-
zene were treated with 1.1 equivalents of n-butyllithium to
generate the monolithio species that were allowed to react
with tri(isopropyl)boron to give borate intermediates, which
were treated directly in the same pot with 6-bromopyridine-2-
carboxaldehyde under Suzuki coupling conditions to afford
the corresponding 6-aryl-pyridine-2-carboxaldehydes in
118
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0118 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 1

Redox-Switched Binding
116 129
about 70% yield. These were coupled with di(2-pyridylme-
thyl)amine using sodium triacetoxyborohydride.[31] L1 was
obtained in 70% yield and, following treatment with etha-
nol HCl (pH 3) to remove the tetrahydropyranyl protecting
groups, H2L2 was obtained in about 60% yield.
Crystal structures: The X-ray crystal structure of [Cu(k4N-
L1)Cl]Cl ¥ 3H2O (1 ¥ 3H2O; Figure 2) reveals two independent
but near identical CuII complex ions (A and B), each bound by
the three pyridyl and the amine donor groups of L1 and by a
The CuII complexes, deep blue [Cu(L1)Cl]Cl (1) and jade-
green [Cu(H2L2)Cl2] (3), crystallised directly from solutions of
the appropriate ligand and CuCl2 ¥ 2H2O in dry methanol (for
1) or dry ethanol (for 3) that were placed under atmospheres
of dry diethyl ether. All procedures and measurements
involving 3 were performed under strictly anaerobic con-
ditions, because solutions of 3 decomposed upon exposure to
oxygen or moisture, but were indefinitely stable in dry,
deoxygenated solvents; attempts to identify the brown
oxidation product(s) obtained upon bubbling dry oxygen
through solutions of 3 were unsuccessful.[32] Treatment of 1
with excess K[PF6] in methanol gave deep indigo-blue
[Cu(L1)Cl][PF6] (2) in near quantitative yield. The blue p-
quinonyl-substituted complex, [Cu(L3)Cl]Cl (4), was obtained
by oxidation of 3. Three methods were employed: a) treatment
of
3 in acetonitrile solution with two equivalents of
[NH4]2[Ce(NO3)6]; b) treatment of 3 in dichloromethane or
acetone solutions with one equivalent of 2,3-dicyano-5,6-
dichlorobenzoquinone (DDQ); c) exhaustive controlled po-
tential electrolysis of 3 in acetonitrile 0.2m [NBu4][PF6] with
the potential poised at ‡0.47 V, just above that of the anodic
peak for the oxidation of the hydroquinone group in the cyclic
voltammogram of 3 (see below). The characteristic green
colour of solutions of 3 turned to blue as the oxidations to 4
proceeded. The samples of 4 obtained by the three methods
displayed identical spectroscopic properties (see below).
Reaction of 3 in dry methanol with one equivalent of NaOH,
followed by metathesis with K[PF6] gave a dark brown
solution, which afforded lustrous golden-brown crystals of
[Cu(HL2)][PF6] (5) when left to stand under a diethyl ether
atmosphere.
The CuI complexes, [Cu(L1)(MeCN)x][PF6] (6) and
[Cu(H2L2)(MeCN)x][PF6] (7), were prepared by three meth-
ods: a) mixing equivalent amounts of the appropriate ligand
(L1 or H2L2) and [Cu(MeCN)4][PF6] together in acetonitrile
solution; b) reduction of the corresponding CuII complex, 1 or
3, with one equivalent of cobaltocene in acetonitrile;
Figure 2. View of cation A from the crystal structure of [Cu(L1)Cl]Cl ¥
3H2O (1) (10% thermal ellipsoids shown). Key bond lengths [ä] and
À
À
À
angles [8]: CuA ClA 2.232(3), CuA N1A 2.351(6), CuA N2A 1.991(6),
À
À
CuA N3A 1.996(6), CuA N4A 2.052(8); ClA-CuA-N1A 117.3(2), ClA-
CuA-N2A 96.7(2), ClA-CuA-N3A 94.5(2), ClA-CuA-N4A 162.6(3), N1A-
CuA-N2A 85.9(2), N1A-CuA-N3A 103.0(2), N1A-CuA-N4A 80.1(3),
N2A-CuA-N3A 160.5(3), N2A-CuA-N4A 83.1(3), N3A-CuA-N4A
81.4(3). Data for cation B: CuB-ClB 2.234(3), CuB-N1B 2.362(6), CuB-
N2B 1.994(5), CuB-N3B 2.000(5), CuB-N4B 2.045(7); ClB-CuB-N1B
119.8(2), ClB-CuB-N2B 95.9(2), ClB-CuB-N3B 94.1(2), ClB-CuB-N4B
160.5(3), N1B-CuB-N2B 88.3(2), N1B-CuB-N3B 101.4(2), N1B-CuB-N4B
79.7(3), N2B-CuB-N3B 160.3(3), N2B-CuB-N4B 82.9(3), N3B-CuB-N4B
82.1(3).
chloro coligand. The CuII ions adopt square-pyramidal (sp)
geometries (cation A: t[34] ˆ 0.03; cation B: t < 0.01) with the
bulky aryl-substituted pyridyl (PY*) group directed to the
À
axial position at an average Cu N distance of 2.353 ä,
À
substantially longer than the equatorial Cu N distances (av
À
2.030 ä). The Cu Cl distances are normal and average
2.294 ä. In the crystal structure, the ClÀ counterions and the
three water molecules are disordered over four sites, well
isolated from the cations. The only notable inter-cation
interaction in the crystal structure is offset p-stacking at
about 3.5 ä between the dimethoxyphenyl rings of the bulky
PY* legs of the pairs of molecules A and B. The crystal
structure of [Cu(k4N-L1)Cl][PF6] (2), which exhibits sp-
c) exhaustive controlled potential reductions of
1 in
acetonitrile 0.2m [NBu4][PF6], which consumed about
1.0 FaradaymolÀ1. A controlled potential electrolysis of 3 at
À0.9 V resulted in decomposition and deposition of copper
metal. Likewise, attempts to concentrate solutions containing
7 caused disproportionation/decomposition and a pure solid
sample of this complex was not obtained. Solutions of 6 and,
particularly, 7 were not stable, decomposing rapidly in
chlorinated solvents and more slowly in acetonitrile (within
several hours), and were extremely dioxygen sensitive. Details
of the reactions of 6 and 7, which are not the subject of the
present investigation, will be reported elsewhere.[33] It is
noteworthy, however, that the reaction of 7 in dry acetone
with dioxygen at À908Cproduced a dark-brown intermediate
species that cleanly decomposed upon warming to give 5 in
about 75% yield with hydrogen peroxide detected as a
coproduct.
[Cu(k4N-L1)Cl] ions (t ˆ 0.15) similar to those found for 1,
‡
is reported elsewhere along with those of L1 and seven other
first d-series metal complexes of this ligand.[33]
Immediately apparent from the X-ray crystal structure of 3
(Figure 3) is that the complex is [Cu(k3N-H2L2)Cl2] with the
PY* leg uncoordinated and orientated away from the CuII ion,
its place taken by a second chloro coligand such that the
complex is sp (t ˆ 0.03). An intramolecular hydrogen bond
between the hydroquinonyl 2-OH group and the pyridine N
atom (O1 ¥¥¥ N1 2.559 ä) ™ties up∫ the PY* leg. The three
À
À
equatorial Cu N (av 2.022 ä) and the equatorial Cu Cl
Chem. Eur. J. 2003, 9, No. 1
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0119 $ 20.00+.50/0
119

S. B. Colbran et al.
FULL PAPER
by the ligand design, the distances from the Cu center to the
À
donor atoms of the PY* arm (Cu1 O2 1.899(4) ä and
À
Cu1 N4 1.978(4) ä) are significantly shorter than the other
coordinate bonds. The structure of 5 resembles that of
[CuL4][ClO4] (L4 ˆ bis(2-aminoethyl)(2-(4'-nitro-2'-hydroxy-
phenyl)iminoethyl)amine),[35] and analogous k4N,kO-binding
of a phenolate-tpa ligand was recently found in [Fe2(6-(2-
phenolato)-tpa)2(m-O)]2‡, although to an octahedral FeIII ion
in this case.[36]
Physicochemical properties
Figure 3. View of crystal structure of [Cu(H2L2)Cl2] (3) (10% thermal
À
ellipsoids shown). Key bond lengths [ä] and angles [8]: Cu Cl1 2.225(1),
Copper(ii) complexes: Emphasis is placed in the following
description on establishing the structures of the complexes in
solution and, where not known from X-ray crystallography, in
the solid-state. All complexes gave correct partial elemental
analyses.
À
À
À
À
Cu Cl2 2.585(1), Cu N2 2.011(3), Cu N3 2.004(3), Cu N4 2.052(3); Cl1-
Cu-Cl2 106.3(1), Cl1-Cu-N2 97.5(1), Cl1-Cu-N3 97.8(1), Cl1-Cu-N4
160.4(1), Cl2-Cu-N2 92.9(1), Cl2-Cu-N3 91.5(1), Cl2-Cu-N4 93.3(1), N2-
Cu-N3 162.3(1).
(2.225(1) ä) bond lengths are normal and comparable to
Infrared spectra: Data for selected key bands from FTIR
spectra recorded for the ground (micro)crystalline complexes
in KBr disks are given in Table 1. Free pyridines show two ring
À
those found in 1; in contrast, the axial Cu Cl bond is longer
(2.582(1) ä). The overall structure of 3 is similar to that of
[Zn(k3N-L1)Cl2],[33] but there are significant differences in the
detail: for example, the ZnII complex exhibits appreciably
Table 1. Pyridine-ring deformation band data from the FTIR spectra of
the solid complexes, and the deduced number of pyridine groups bound to
the metal ion.
À
longer metal N bond lengths (Zn N av 2.210 ä), near equal
À
Zn Cl bond lengths (2.235(1) and 2.244(1) ä), and consid-
erable trigonal distortion about the ZnII centre (t ˆ 0.29).
The crystal structure of 5 (Figure 4) reveals that the 2,5-
hydroquinonyl-substituent of H2L2 binds to the CuII ion as a
2-hydroquinonate O-donor along with the four nitrogen
atoms of the ligand. The [Cu(k4N,kO-HL2)] ion adopts a
distorted trigonal-bipyramidal (tbp) geometry (t ˆ 0.71) with
the amine N atom and the hydroquinonate O atom in the axial
positions and the three pyridyl groups equatorial. Constrained
nPY‡PY* [cmÀ1
]
No of bound PY
Ä
[Cu(L1)Cl]Cl (1)
[Cu(L1)Cl][PF6] (2)
[Cu(H2L2)Cl2] (3)
[Cu(L3)Cl]Cl (4)
[Cu(HL2)][PF6] (5)
[Cu(L1)][PF6] (6)
[Zn(L1)Cl2][a]
1608, 1573
1609, 1581
1608, 1596, 1567
1613, 1564
1608, 1567
1608, 1597, 1577
1605, 1585, 1572
3
3
2
3
3
2
2
‡
[a] Data from reference [33].
deformation bands at about 1590 1580 and about 1575
1565 cmÀ1 with the higher energy band shifting to about
1615 1605 cmÀ1 upon coordination to a metal ion.[37] These
bands are easily distinguished in the FTIR spectra of the
complexes allowing the unambiguous assignment of the
coordination mode of the tpa ligand, Table 1 and, for example,
Figure 6. Also of note, FTIR spectra of 4 in solution and in the
solid show a strong quinonyl carbonyl peak at 1662 cmÀ1.
Conductivity: Table 2 presents molar electrical conductivity
data obtained for 1.0 mm solutions of the complexes in
anhydrous dichloromethane and in dimethylformamide. In
Table 2. Molar electrical conductivity values (Æ10%) for the complexes.
LM [Scm2 molÀ1
]
CH2Cl2
dmf
CH3OH
[a]
[a]
[Cu(L1)Cl]Cl (1)
[Cu(L1)Cl][PF6] (2)
[Cu(H2L2)Cl2] (3)
[Cu(L3)Cl]Cl (4)
[Cu(HL2)][PF6] (5)
[Zn(L1)Cl2]
4
24
6
17
22
4[b]
26
44
64
42
47
59
9[b]
73
192
Figure 4. View of the cation from the crystal structure of [Cu(HL2)][PF6]
(5) (10% thermal ellipsoids shown). Key bond lengths [ä] and angles [8]:
[a]
80
À
À
À
À
Cu O1 1.899(4), Cu N1 1.978(4), Cu N2 2.129(5), Cu N3 2.019(5),
[a]
À
Cu N4 2.041(4); O1-Cu-N1 94.9(2), O1-Cu-N2 96.1(2), O1-Cu-N3
[NBu4][PF6]
96
100.7(2), O1-Cu-N4 175.8(2), N1-Cu-N2 98.1(2), N1-Cu-N3 133.2(2), N1-
Cu-N4 84.8(2), N2-Cu-N3 123.4(2), N2-Cu-N4 79.8(2), N3-Cu-N4 82.5(2).
[a] Not measured. [b] Data from [33].
120
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0120 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 1

Redox-Switched Binding
116 129
dichloromethane, the values
are consistent with 1 and 3
behaving as non-electrolytes,
and 2, 4 and 5 as 1:1 electro-
lytes. This suggests that in di-
chloromethane 1 is [Cu(L1)Cl2],
‡
not [Cu(L1)Cl] ClÀ as found in
the crystal structure, and 4 is
‡
[Cu(L3)Cl] ClÀ rather than
[Cu(L3)Cl2]. In dimethylforma-
mide, the molar conductivities
are consistent with all copper
complexes behaving as 1:1 elec-
trolytes. Thus, in dimethyl-
formamide the dichloride com-
‡
plexes are all [Cu(L)Cl] ClÀ (1:
L ˆ L1; 3: L ˆ H2L2; 4: L ˆ L4).
The lower conductivity values,
in both solvents, when ClÀ is the
counterion point to (stronger)
ion pairing.
X-band EPR and UV/Vis/NIR
spectra: Salient solution-state
data are presented in Table 3.
EPR spectra of 1 and 3 in
dichloromethane or acetonitrile
at 80 K show typical axial fea-
Figure 5. X-band EPR spectra of [Cu(L1)Cl2] (1) (a) and [Cu(L1)Cl][PF6] (2) (b) in frozen dichloromethane
solution at 80 K. c, d). Titration of [Cu(L1)Cl][PF6] (2, 0.59m) with ClÀ ion in dichloromethane at 295 K to afford
[Cu(L1)Cl2] (1): Vis-NIR spectra (c) and the corresponding double-reciprocal plot (À1/DA versus 1/[ClÀ]added
DA ˆ absorbance change at 735 nm) (d).
;
tures with gk > g? > 2.0 and jA? jꢀ 170 180 G indicative for
tetragonally distorted (dx2Ày2 ground-state) elongated octahe-
dral or sp CuII species in solution,[27, 38] whereas EPR spectra of
show intense ligand-centred bands below 350 nm, and weak
d
d transitions in the Vis/NIR region with profiles[27, 38]
characteristic for sp CuII centres for 1 and 3 and tbp CuII
centres for 2 and 4 (Figure 5
and Figure 6). Spectra of 5 show
intense ligand-centred bands
below 350 nm, and a character-
istic intense, sharp band at
about 400 nm, which was not
observed for the other com-
plexes, along with three weak,
broad shoulders to this at about
Table 3. Electronic and X-band EPR spectroscopic data for the CuII complexes.
UV-Vis-NIR (295 K)[a]
lmax [nm] (e [mÀ1 cmÀ1])
EPR (80 K)[a]
[Cu(L1)Cl2] (1)
257 (17700), 320 (8000),
770 (200), 1050 (130)
gk ˆ 2.236, g? ˆ 2.027,
Ak ˆ 176 G
[Cu(L1)Cl][PF6] (2)
[Cu(H2L2)Cl2] (3)[b]
[Cu(L3)Cl]Cl (4)
[Cu(HL2)][PF6] (5)
259 (13800), 311 (5050),
730 (105), 930 (130)
253 (16400), 336 (3600),
720 (130), 855 (120)
246 (10500), 265 (9800),
730 (100), 900 (130)
256 (13400), 286 (12000), 410 (4600),
520 (550), 650 (270), 1000 (90)
g? ˆ 2.235, gk ꢀ2.0,
A? ˆ 112 G, Ak ꢀ86 G
gk ˆ 2.239, g? ˆ 2.025,
Ak ˆ 171 G
520, 660 and 930 nm (see Sup-
porting Information).
g? ˆ 2.228, gk ꢀ 2.0,
A? ˆ 113 G, Ak ꢀ 85 G
g? ˆ 2.200, gk ꢀ 2.0,
A? ˆ 115 G, Ak ˆ 84 G
Titrations with ClÀ ion: Fig-
ure 5c shows Vis/NIR spectra
from a typical titration of 2
(lmax ˆ 930 nm) in dichlorome-
thane with [(Ph3P)2N]Cl. An
[a] In dichloromethane solution (unless stated otherwise). [b] [Cu(H2L2)Cl2] is poorly soluble in dichloromethane
and the data for this complex are from spectra of anhydrous acetonitrile solutions.
2, 4 and 5 in dichloromethane or acetonitrile have a typical
™inverted axial∫ appearance with g? > gk ꢀ 2.0 and jA? jꢀ
115–110 G >j Ak jꢀ 85 G indicative for (dz2 ground-state)
tbp species in solution,[27, 38] for example see Figures 5, 6, and
the Supporting Information. The solid samples of 4 (and 5)
display ™inverted axial∫ powder EPR spectra (g? ꢀ 2.014 >
gk ꢀ 2.0) at 80 K, which reveal 4 (and 5) retains a tbp
coordination geometry in the solid-state. In contrast, pow-
dered samples of microcrystalline 1 3 exhibit rhombic EPR
spectra (g1 ꢀ 2.2, g2 ꢀ 2.1, g3 ꢀ 2.0) at 80 K. UV/Vis/NIR
spectra of 1 4 in dichloromethane or acetonitrile solutions
isosbestic point at 630 nm reveals a direct conversion from
tbp reactant to sp product, 1 (lmax ˆ 767 nm), without inter-
mediate species (Scheme 1). A plot of the absorbance change
at 770 nm (DA770) against the ratio of ClÀ:2 indicates a single
ClÀ ion binds to 2 (two ClÀ bind Cu overall), and the chloride
affinity constant [KCl: Eq. (2)] at 295 K is estimated from the
KCl
‡
[Cu(L)Cl] ‡ ClÀ „[Cu(L)Cl2] (L ˆ tpa, L1, L4)
(2)
double reciprocal plot of the inverse of the absorbance change
at 735 nm (DA735) against the inverse of the ClÀ ion concen-
Chem. Eur. J. 2003, 9, No. 1
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0121 $ 20.00+.50/0
121

S. B. Colbran et al.
FULL PAPER
Formation of 5 from 3: The
EPR and the UV/Vis/NIR spec-
tra described above for 3 were
only obtained with some diffi-
culty, by using carefully purified
and rigorously dry solvents that
have little Br˘nsted basicity
such
as
dichloromethane,
chloroform and acetonitrile, in
which the complex was poorly
soluble. The resulting blue-
green solutions were extremely
sensitive to trace amounts of
moisture or other basic impur-
ities, rapidly turning greenish-
yellow to brown and with the
characteristic EPR and UV/Vis/
NIR spectra of 5 (see Support-
ing Information) replacing those
of 3 if, for example, the solutions
were exposed to moisture. Com-
plex 3 was more soluble in
alcohols such as methanol, but
EPR and Vis/NIR spectra of
these solutions show only peaks
for 5. These results highlight the
non-innocence of the hydroqui-
nonyl substituent in 3–in the
sense that it can coordinate to
Cu–with solutions of 3 being
difficult to obtain due to forma-
tion of 5 [Eq. (3): possible bases
include a protonatable solvent
or impurities]. The conductivity
data for 3 and 5 in methanol,
Table 2, are also consistent with
this conclusion. Furthermore, a
preliminary potentiometric ti-
tration of H2L2 and Cu2‡ ion
[as Cu(ClO4)2] in 95% metha-
nol-water reveals one proton is
lost from the ligand upon com-
plex formation (to give 5) which
is complete above pH 2 with the
value for the stability constant
uncertain but high (log K > 15).
Figure 6. a c) Controlled-potential electrolysis of [Cu(H2L2)Cl2] (3) in acetonitrile 0.4m [NBu4][PF6] at
‡0.47 V (Pt gauze working electrode) to afford [Cu(L3)Cl]Cl (4): X-band EPR spectra of samples taken before
(a) and after (c) the electrolysis and frozen at 80 K–the asterisks in c) mark peaks for residual 3 (see the text);
b) UV-Vis-NIR spectra before (––) and after electrolysis (- - - -). d) Pyridyl-ring deformation bands from the
FTIR spectra of [Cu(H2L2)Cl2] (3) (––) and isolated [Cu(L3)Cl]Cl (4) (- - - -).
solid
N
N
OMe
N
CuII
X
N
Cl
sp
dichloromethane
solution
OMe
1 or 2
X
= PF6
OMe
X
= Cl
MeO
N
N
N
CuII
N
N
N
Cl
± Cl
KCl = 9.7 × 102
Cl
N
CuII
MeO
N
OMe
Cl
sp
tbp
1
2
Scheme 1. [Cu(L1)Cl2] (1)>[Cu(L1)Cl][PF6] (2) equilibria.
tration[39] (Figure 5d) to be 9.7 Â 102. The titration was also
monitored by EPR spectroscopy with the spectrum of 1
replacing that of 2, and a conductiometric titration of 2 with
[(Ph3P)2N]Cl gave an endpoint at 1.0 equivalents, correspond-
ing to formation of the non-electrolyte 1. For comparison,
titrations in dichloromethane of [Cu(tpa)Cl][PF6] and of 4
with [(Ph3P)2N]Cl (up to 20 equivalents) were investigated. In
both cases no reaction was observed–as the increments of
[(Ph3P)2N]Cl were added, the EPR and Vis/NIR spectra of
the solutions remained unchanged whereas the conductivity
of the solutions monotonically increased. From these experi-
ments the chloride binding constants of [Cu(tpa)Cl][PF6] and
4 in dichloromethane [Eq. (2)] are estimated to be both < 0.1.
‡
‡
[Cu(k3N-H2L2)Cl2] (3) ‡ base >[Cu(k3NkO-HL2)] (5) ‡ Hbase ‡ 2 C Àl (3)
Electrochemistry: In the cyclic voltammograms (CVs) of
complexes 1 and 2 in dichloromethane (see Supporting
Information), the quasi-reversible (DEp(couple) ꢀ120
‡
160 mV> DEp(Fc À Fc) ꢀ 65 mV with the scan rate ˆ
100 mVsÀ1), one-electron oxidations at ‡1.02 and ‡0.86 V,
respectively, are attributed to the dimethoxyphenyl radical
I
cation/dimethoxyphenyl couple.[40] The CuII C u couples
appear as a cathodic peak at À0.66 V for 1 and at À0.85 V
for 2 with both complexes showing a broad anodic peak at
about À0.5 V in the reverse sweep. The poor electrochemical
122
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0122 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 1

Redox-Switched Binding
116 129
I
reversibility of the CuII C u couples is indicative of the
accompanying structural changes being rate-limiting on the
CV time scale.[38, 41] A constant potential electrolysis of 1 at
À0.95 V was a one-electron process and produced 6 (see
above).
[DEp(VI/VII) ꢀ 105 mV at n ˆ 100 mVsÀ1 and increases with
v). A comparison of the peak heights in the CVs of 2 (1 mm)
and 5 (1 mm), both in acetonitrile, indicates both couples for 5
to be one-electron processes. Hence the reduction is attrib-
I
uted to a CuII C u couple and the oxidation to a ligand-
Cyclic voltammograms of H2L2 in acetonitrile are typical
for a hydroquinonyl-containing compound and exhibit an
anodic peak (I) at ‡0.20 V and a cathodic peak (II) at
À0.59 V in the reverse sweep (Figure 7a). Peak I is attributed
centred hydroquinonate semiquinonate couple.[29, 30, 42] The
quasi-reversibility of the couples may arise from concomitant
rate-limiting structural changes. Also noteworthy in the CVof
5 is the cathodic ™daughter∫ peak (V) at À0.57 V, which is
found only after scanning through the III/IV couple. The
potentials of peak V and peak II, which corresponds to
‡
reduction of L3/H in the CVs of H2L2, are very close and,
accordingly, peak V is attributed to the reduction of a CuII
species with a pendant quinone group. Such a species is
expected: protonated semiquinones rapidly disproportion-
ate,[19, 20, 29, 30] and disproportionation of the semiquinonate
CuII product of peak III would generate equal amounts of
hydroquinonyl CuII (3 or 5) and quinonyl CuII species. The
hydroquinonyl CuII complexes (3 or 5) would be sponta-
neously oxidised at the potential of their formation (i.e., at all
potentials higher than peak III),[19, 20, 29, 30] and therefore on
the longer time scale of a bulk electrolysis 5 should undergo
an overall two-electron oxidation to a quinonyl CuII species.
Consistent with this interpretation, controlled potential
oxidation of 5 at ‡0.40 V consumed 1.94 FaradaymolÀ1.
Cyclic voltammograms of 5 in dichloromethane show qual-
itatively the same redox chemistry, but with redox couples at
‡0.22 and À1.08 V.
Cyclic voltammograms of 3 in acetonitrile (Figure 7c/d)
reveal peaks III/IV and VI/VII for 5 as well as a quasi-
reversible reduction couple (peaks IX/X) at À0.49 V (DEp-
(IX/X) ꢀ 170 mV at n ˆ 100 mVsÀ1 and increases with v),
which becomes less chemically reversible when the VI/VII
couple is traversed, and an irreversible oxidation (peak VIII)
at about 0.10 V. Vis/NIR and EPR spectra of the acetonitrile
solutions were recorded prior to and after the CVexperiments
and show only peaks for 3 (Figure 6). The peaks for 5 in the
CVs of 3 therefore suggest that pre-equilibration of 3 and 5
[Eq. (3)] occurs within the time scale of the CV experiment.
I
Peaks IX/X are attributed to the CuII C u couple for 3 with
the accompanying structural changes being rate-limiting and
leading to the quasi-reversible electrochemical response.
Peak VIII is attributed to the two-electron oxidation antici-
pated for the ™dangling∫ hydroquinone substituent in 3, which
should give the corresponding quinone in a chemically
reversible, but electrochemically irreversible (proton-depend-
ent) process;[20, 29, 30] thus oxidations of 3 (peak VII) and 5
(peak III) should both give quinonyl CuII product. Support-
ing these conclusions, controlled potential electrolysis of 3 at
‡0.47 V consumed 1.97 FaradaymolÀ1. Figure 6 depicts EPR
and Vis/NIR spectra recorded prior to and after electrolysis of
3–the spectra reveal 4 to be the electrolysis product
(Scheme 2). The small peaks for a sp species–the axial sub-
spectrum marked by asterisks–in the post-electrolysis EPR
spectrum (Figure 6c) are attributed to residual 3 for two
reasons: 1) They occur at identical magnetic field to peaks in
the spectrum of 3; 2) additional ClÀ ion did not affect the post-
electrolysis UV/Vis/NIR or EPR spectra (Figure 6b/c), ruling
out that they arise from the addition of ClÀ ion to 4 to give a
Figure 7. Cyclic voltammograms for: a) H2L2; b) [Cu(HL2)][PF6] (5); c and
d) [Cu(H2L2)Cl2] (3). Conditions: acetonitrile 0.1m [Bu4N][PF6]; 295 K;
1.0 mm Pt disk working electrode; scan rate ˆ 100 mV sÀ1
.
to the anticipated hydroquinone-centred oxidation[29, 30] of
H2L2 to afford quinone-substituted L3 [Eq. (4)], and peak II to
‡
the reverse reduction process (L3/H ! H2L2). Although
these processes are chemically reversible, they are electro-
chemically irreversible because the associated discharge or
uptake of protons changes the local pH at the electrode.[29, 30]
‡
H2L2>L3 ‡ 2H ‡ 2eÀ
(4)
Figure 7b presents a representative CV for 5 in acetonitrile
and reveals a quasi-reversible oxidation (the III/IV couple) at
‡
‡0.15 V (DEp(III/IV) ꢀ 260 mV cf. DEp(Fc À Fc) ꢀ 65 mVat
a scan rate (n) ˆ 100 mVsÀ1; DEp(III/IV) increases with v) and
a quasi-reversible reduction (the VI/VII couple) at À1.25 V
Chem. Eur. J. 2003, 9, No. 1
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0123 $ 20.00+.50/0
123

S. B. Colbran et al.
FULL PAPER
are diastereotopic and should
appear as a pair of AB methyl-
ene doublets as is found in
1H NMR spectra of tpa com-
plexes of substitutionally inert
metal ions such as octahedral
rhodium(iii) complexes.[43] The
methylene singlets in the
1H NMR spectra of L1, H2L2, 6
and 7 are accounted for by
rapid inversion of configuration
about the amine nitrogen,
which is typically a low energy
OH
N
N
N
OH
CuII
Cl
N
N
± 2e , ± 2H
N
Cl
Cl
N
O
CuII
N
sp
tbp
O
Cl
4
3
Scheme 2. Oxidation of sp-[Cu(H2L2)Cl2] (3) ! tbp-[Cu(L3)Cl]Cl (4).
sp-quinonyl CuII species. Cyclic voltammograms of 4, pro-
duced by the electrolysis of 3, revealed a broad cathodic peak
at about À0.6 V in the forward cathodic sweep attributed to
process for ™free∫ amines.[44, 45] The methylene singlets remain
sharp in NMR spectra of 6 recorded down to 260 K; based on
this observation the estimated upper limit for the free energy
barrier to exchange the diastereotopic methylene protons is
about 45 kJmolÀ1. As a comparison, a variable temperature
overlapping of the expected quinone- (L3/H ! H2L2)[20, 29, 30]
‡
and copper-(CuII ! CuI) centred reduction processes, and in
the reverse anodic sweep peak X and peak VIII as a
™daughter∫ peak. The CV is typical for a basic quinone,
which exhibit a cathodic peak for two-electron reduction of
300 MHz H NMR study of the isoelectronic ZnII complex,
1
[Zn(L1)Cl2], revealed a pair of AB doublets for the methylene
protons in lower temperature spectra that coalesce
(Tcoalescence ˆ 290 K) to a singlet in higher temperature spectra;
the free energy barrier for intramolecular exchange of the
diastereotopic methylene protons in this ZnII complex is 56 Æ
0.5 kJmolÀ1.[33]
‡
quinone/H followed by a ™daughter∫ anodic peak for two-
electron oxidation of the corresponding hydroquinone in the
reverse sweep rather than consecutive one-electron quinone/
semiquinone anion and semiquinone anion/hydroquinone
dianion couples, all due to extreme sensitivity to sub-
equimolar amounts of Br˘nsted acids or hydrogen bond
donors.[30] Controlled potential bulk electrolyses of 4 at
À0.90 V caused decomposition and deposition of copper
metal, a result suggestive for an unstable CuI product, such as
7 (see above), and consistent with the above assignment of the
À0.6 V cathodic peak.
Third, in the H NMR spectra of H2L2 and 7 the hydroxy
1
peaks appear as a sharp singlet at dH (CD3CN) ˆ 13.77 (H2L2)
and 13.68 ppm (7) and a broad singlet at dH (CD3CN) ꢀ 6.6
(H2L2) and ꢀ9.9 ppm (7). The chemical shifts of the sharp
low-field singlets show little concentration dependence in-
dicative for an intramolecular hydrogen bond, which must be
between the ortho-hydroxy group and the pyridyl nitrogen
atom of the PY* leg. This strongly suggests that the PY* leg in
7 is unbound, akin to in the crystal structure of 3 (see above).
Finally, the NMR peaks for 6 and 7 broaden slightly and shift
when ClÀ ion was present or was deliberately added (data are
presented in Table S1 in the Supporting Information),
indicative for some association between the CuI centres and
ClÀ ion.
‡
The 2eÀ, 2H quinone hydroquinone couple may become
electrochemically reversible in a buffered solution.[29, 30] Un-
fortunately, adding a pH-buffer to 3, in aqueous or non-
aqueous solvent, lead to complete conversion to 5 or, at very
low pH (< 2), to decomplexation of H2L2; conditions could
not be found where the 4-3 (hydro)quinone-centred couple
was electrochemically reversible.
In sum, these physicochemical data provide compelling
evidence for a non-coordinated PY* leg in 6 and 7 in the solid
state and in solution. Moreover, rapid inversion of the ligand
configuration in 6 and 7 requires at a minimum dissociation of
one of the three pyridylmethyl legs and the amine nitrogen
atom (Scheme 3). The estimated upper limit of about
45 kJmolÀ1 for exchange of the diastereotopic methylene
protons in 6 compares with 28 Æ 3 kJÀ1 mol for the energy
Copper(i) complexes
The CuI complexes 6 and 7 were not characterised by
elemental analysis because of their extreme air-sensitivity.
Positive ion ESI-mass spectra show strong peaks correspond-
‡
ing to the [CuL] molecular ions, at m/z 490.0 (calcd: m/z
490.07) for 6 and m/z 462.0 (calcd: m/z 462.01) for 7. The FTIR
spectrum of solid 6 exhibits three prominent pyridine
deformation bands indicative for an unbound, dangling PY*
leg (Table 1).
The 1H NMR spectra of 6 and 7 show several notable
features. First, distinct coordination shifts (Dd) for the ligand
protons are observed in the 1H NMR spectra upon forming 6
PY*
PY*
Ha
N
PY*
Hb
Hb
N
N
CuI
L'x
Hb
Ha
N
Ha
N
N
N
CuI
N
CuI
1
N
and 7 (see Supporting Information). Second, the H NMR
L'x
spectra of the ligands and the complexes show two sharp
methylene singlets, for example at dH(CD3CN) ˆ 4.07 (2H)
and 4.04 ppm (4H) for 6 and at dH(CD3CN) ˆ 4.00 (2H) and
3.92 ppm (4H) for 7, revealing the two 2-pyridylmethyl legs to
be equivalent. However, the methylene protons of these legs
L'x
Scheme 3. Possible mechanism for exchange of the diastereotopic meth-
ylene protons in the CuI complexes, [Cu(L)(L')x]z‡ (6: L ˆ L1, 7: L ˆ H2L2;
L' ˆ ancillary (solvent or ClÀ) ligand), involving dissociation and inversion
of the amine group.
124
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0124 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 1

Redox-Switched Binding
116 129
barrier to inversion of configuration for dibenzylmethyl-
amine,[45] which also provides an estimate for the lower limit
to the barriers for inversion of configuration in ™free∫ L1 or
™free∫ H2L2. Thus the barriers to amine inversion in 6 and 7
should fall in the range of about 30 45 kJmolÀ1, which
compares with 56 kJmolÀ1 for [Zn(L1)Cl2][33] consistent with
the ligand binding to ZnII more strongly than to CuI.
as dimethylformamide (er (208C) ˆ 38.3), stabilises sepa-
‡
ration of [Cu(L1)Cl] and ClÀ ions, whereas a less polar
medium, for example in a solvent with a lower dielectric
constant such as dichloromethane (er (258C) ˆ 8.9), pro-
motes coordination of a second chloride coligand to afford
[Cu(k3N-L1)Cl2].
3) Ligand field stabilisation energy (LSFE) which increases
upon binding of a p-acceptor pyridyl over a p-donor
chloride coligand to the d9 CuII ion. A comparison of 1 and
[Zn(L1)Cl2][33] reveals the importance of LFSE: whereas
Discussion
‡
the crystal structure of 1 shows discrete [Cu(k4N-L1)Cl]
and ClÀ ions, its ZnII congener is [Zn(k3N-L1)Cl2] because
there is no LFSE for the d10 ZnII ion and, in this case,
minimisation of Coulombic and steric interactions leads to
coordination of the second chloride coligand over the
pendant PY* leg regardless of the medium.
Towards the goal of preparing a copper RHL system that
provides ™proof-of-concept∫ for our proposal for a redox-
linked conformational proton pump gate operating as depict-
ed in Figure 1, we have prepared a series of copper complexes
of tpa ligands derivatised on one leg (PY*) by dimethox-
yphenyl or (hydro)quinonyl groups. The redox states of the
copper and (hydro)quinonyl centres and the environment
control the geometries of the complexes (see Schemes 1 and
2). The interplay of the following effects is important:
4) Perturbation of the donor properties of a ligand by an
attached redox centre: A comparison of tbp-[CuCl(k4N-
‡
‡
L1)] (2) and tbp-[CuCl(k4N-L3)] (4), which differ only in
the 6-substituent to PY*, reveals the importance of this
effect. The d d band maximum of 2 is about 370 cmÀ1
lower in energy than that for 4 because conjugation
between the rings of the PY* legs results in L3 being a
stronger field ligand than L1 (Figure 8A). The higher
LFSE for 4 compared to 2 results in tbp-[Cu(k4N-
1) Steric interactions introduced by the 6-aryl substitution of
the PY* leg. These are minimised when the complexes
adopt an sp geometry with the PY* leg in the weak-field,
‡
axial position (e.g. the sp-[Cu(k4N-L1)Cl] ion in the
crystal structures of 1 and 2) over a tbp geometry, and more
so when the PY* leg is unbound
(e.g. sp-[Cu(k3N-H2L2)Cl2] (3))
rather than bound. Intramolec-
ular steric strain leads to hemil-
ability–the fine balance be-
tween binding of the PY* leg
versus ClÀ ion–in the 1 2 and
3
4 systems. In contrast, tbp-
[Cu(k4N-tpa)Cl] is unaffected
by excess ClÀ ion. In these
complexes, the particular coor-
dination geometry is unlikely
‡
to
significantly
influence
whether or not a pyridyl leg
undergoes substitution, be-
cause the ligand exchange re-
actions of five- (and lower[38]
)
coordinate CuII typically pro-
ceed by associative (A) or as-
sociative
interchange
(Ia)
mechanisms and exhibit little
dependence on the (sp tbp)
coordination geometry.[25]
2) Electrostatic interactions, which
favour coordination of ClÀ over
the PY* leg to copper; this
effect is minimised in more
polar/higher dielectric environ-
ments that better solvate ions
and stabilise separation of
charges. For 1, a more polar
medium, for example in the
crystal or in a solvent with a
higher dielectric constant such
Figure 8. Redox-switched donor properties for hydroquinone/quinone-pyridine (A) and phenol/phenoxyl
radical-imidazole (B) ligands.
Chem. Eur. J. 2003, 9, No. 1
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0125 $ 20.00+.50/0
125

S. B. Colbran et al.
FULL PAPER
‡
H2L3)Cl] for 4 in the solid and in dichloromethane or
copper centres with a substitutionally labile ligand. One such
acetonitrile, whereas 2 is tbp-[Cu(k4N-L1)Cl] in dichloro-
system is the CuB centre in cytochrome c oxidase,[4 which
‡
13]
methane but sp-[Cu(k4N-L1)Cl] in the crystal. This in turn
has a substitutionally labile histidine ligand proposed to be
crucial to proton pumping by this enzyme.[6, 9] In this case, the
substitutional lability would result from strain imposed by
protein structure, perhaps with specific hydrogen bonding
interactions between protein residues and the labile histidine
stabilising the ™off∫-copper state, and with the displacement
of the labile histidine requiring ancillary ligands (e.g. ClÀ or
OHÀ ion or water would need to be available to bind CuB) and
more likely for the labile CuB(i) state. In the light of the
present study, the facts that one histidine ligand to CuB is
labile and that during the enzyme cycle the tyrosine histidine
ligand to CuB is oxidised with a hydroxide ancillary ligand
simultaneously produced[6, 9, 12] become most intriguing. Argu-
ment based upon the mixing of s/p electrons inherent within
an imidazole[46] and the inter-ring conjugation anticipated for
the tyrosine-histidine cofactor suggests its redox-state should
strongly modulate its metal binding properties (Figure 8B).
The most recent evidence is that there indeed is electronic
delocalisation between the rings of a histidine tyrosyl
radical,[47] and we have shown that such inter-ring delocalisa-
tion between a ligand donor and its radical pendant can
significantly alter the reactivity of a (organo-) metal centre.[48]
The binding of the cofactor imidazole to CuB will be
weakened by oxidation of the tyrosine, the converse to found
for the pyridyl-based 3 4 system (Figure 8). Could it be that
redox-triggered dissociation of the tyrosine-histidine cofactor
from CuB is part of the mechanism of the proton pump in
cytochrome c oxidase?[49] Further investigations are needed to
address this important question.
‡
leads to the >104-fold difference in the affinities of the tbp-
[Cu(k4N-L1)Cl] (from 2) and tbp-[Cu(k4N-L3)Cl] (from
4) ions for ClÀ ion in dichloromethane (see above), and
hence to different structures when ClÀ ions are present–2
binds ClÀ to give sp-[Cu(k4N-L1)Cl2] (1), whereas 4 does
not.
‡
‡
5) Intramolecular hydrogen bonding, which may ™tie-up∫ the
s-lone pair of a potential ligand group thereby stabilising
™off∫-Cu states (Figure 8A). That
3
is sp-[Cu(k3N-
H2L2)Cl2] in the crystal, whereas 1 crystallises as sp-
[Cu(k4N-L1)Cl]Cl, is attributed to the intramolecular
hydrogen bond within the PY* leg of 3.
6) Oxidation state of copper: C u, along with other d10 metal
I
ions, exhibits little preference for one type of donor or one
particular geometry over another, and replacement of one
donor of a tripodal ligand by an ancillary ligand is expected
when this will reduce the repulsive and Coulombic
forces.[41] Hence, the PY* leg is ™dangling∫ in 6 and 7–in
acetonitrile this structure is likely stabilised by solvent
binding to the CuI ion. A number of other CuI complexes
of tpa-like ligands with bulky substituents, akin to in 6 and
7, also exhibit an unbound ligand leg.[26]
7) Availability of additional ligands: Last, and implicit
throughout the previous discussion, potential ancillary
ligands must be available for the substitutionally labile
PY* leg to be displaced from CuII; for example, the
reaction of 2 with ClÀ ion to afford 1. The shifts in the
NMR spectra of 6 and 7 when ClÀ ion is introduced are
indicative for association of ClÀ with the CuI centres.
A redox-driven switch in the conformation of a transition
metal RHL system tightly linked to proton uptake or output
is critical to the successful operation of the proton pump gate
proposed in Figure 1. The 3 4 system provides a first
demonstration of this crucial step–the oxidation of 3 to 4 is
linked to output of two protons and a switch of conformation
from k3N- to k4N-bound ligand. Previous demonstrations of
the use of a ligand-centred redox couple to switch the binding
of the ligand to a transition metal centre are all for systems
Experimental Section
Physical measurements
Elemental analyses for C, H and N were determined by the Australian
National University Microanalytical Unit. Electrospray ionisation mass
spectra (ESI-MS) were acquired on a VG Quattro mass spectrometer with
a capillary voltage of 4 kV and a cone voltage of 30 V. The solvent system
was 50:50 acetonitrile/water and, depending on the sample, with or without
1
1% acetic acid. H and 13C{1H} NMR spectra were recorded on a Bruker
where the redox centre is directly bonded to the donor atom
of the substitutionally labile group.[19
AC300F (300 MHz) spectrometer. Electronic spectra of the complexes
were recorded between 220 and 2000 nm on a CARY 5 spectrometer in
the dual beam mode; solution spectra were recorded in sealed 1 cm quartz
cuvettes and solid state spectra were recorded as KBr disks. Solutions
22]
The 3 4 system
clearly demonstrates for the first time that the redox-state of a
redox-centre covalently linked to a heterocyclic N-donor can
control the binding of the N-donor to a metal ion. The results
also illustrate that redox-linked changes to hydrogen bonding
may play an important role in the stabilisation of the different
conformations for the labile, redox-active ligand group.
Although complexes 3 and 4 could not be reversibly
interconverted due to the formation of unstable CuI species
and metal-binding by the hydroquinonyl redox group, the
results provide a solid foundation for future studies toward
artificial proton pumps.
of 3, 4,
6 and 7 were prepared under nitrogen in a M. Braun
glovebox operating with dioxygen and water levels below 2 ppm.
X-band EPR spectra of both frozen solution (at 80 K; liquid nitrogen
dewar) and solid (dispersed in a KBr matrix) samples were recorded using
a Bruker EMX 10 EPR spectrometer. Electrical molar conductivity
measurements were made using an in-house built conductivity bridge
and a YSI model 3403 conductivity cell thermostated at 258C. The cell
constant (K) was determined with a standard aqueous solution of KCl
(0.001m). The molar conductivity, LM, of a sample solution was determined
from Lm ˆ 1000 K/ cm, where cm is the molar concentration of the complex
(ca. 1 mm).
Cyclic voltammetry measurements were made using a Pine Instrument Co.
As far as we are aware, a biological proton pump operating
in the simple mode proposed in Figure 1 is yet to be
discovered. Nevertheless, our results may have implications
for biological systems. For instance, the structure-determining
factors delineated in this study will also apply to biological
AFCBP1 bipotentiostat interfaced to and controlled by
a Pentium
computer as previously described in detail elsewhere.[20] Solutions of the
compounds were 1.0 mm in anhydrous dichloromethane, acetonitrile or
dimethylformamide (Aldrich, used as received) with 0.1m tetra-n-butyl-
ammonium hexafluorophosphate. Solutions were deoxygenated by bub-
126
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0126 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 1

Redox-Switched Binding
116 129
bling with high purity nitrogen (presaturated with solvent) and then
blanketed with a cover of nitrogen for the duration of the experiment or,
alternatively, the electrochemical experiments were carried out in the
above M. Braun nitrogen-filled glove-box. Data are reported from cyclic
voltammograms recorded with a 1 mm Pt disk working electrode at a scan
a rotary evaporator. The residue was extracted with dichloromethane
(100 mL) and the extracts washed with saturated sodium bicarbonate
(40 mL), then water (40 mL), and dried over sodium sulfate. Rotary
evaporation of the solvent yielded a tacky yellow solid, which was
recrystallised from chloroform/hexane to give a yellow powder (1.2 g,
63%), m.p. 198 1998C; 1H NMR (300 MHz, CDCl3, 258C) d ˆ 8.52 (d,
2H; Py), 7.64 (m, 6H; Py), 7.36 (m, 1H; Py), 7.20 (m, 4H; Py), 7.13 (m, 2H;
Ph), 6.90 (m, 1H; Ph), 3.90 (m, 4H; CH2), 3.88 ppm (s, 2H; CH2);
elemental analysis calcd (%) for C24H22N4O2 ¥ H2O: C69.23, H 5.76, N
13.46; found: C69.74, H 5.45; N 13.26.
rate of 100 mVsÀ1
. Except where stated otherwise, electrochemical
potentials herein are quoted relative to the ferrocenium ferrocene
(FeIII FeII) couple measured under the same experimental conditions
(same concentrations, solvent, support electrolyte, electrodes, temperature,
and scan rate).
Preparations
[H3L2][ClO4]: [CAUTION: Although no problems were encountered with
this preparation, perchlorate salts are potentially explosive materials and
appropriate precautions should be taken when handling them.] 40%
Aqueous HClO4 was added dropwise to a solution of H2L2 (700 mg,
1.75 mmol) in 5% water in ethanol solution (20 mL) until pH ˆ 3.0. The
yellow solid that precipitated was collected by filtration. Recrystallisation
from ethanol/water gave clear yellow crystals of the product (700 mg,
Preparation of bis(2-pyridylmethyl)[6-(2',5'-dihydroxyphenyl)-2-pyridyl-
methyl]amine (H2L2): The preparation of H2L2 is described. The synthesis
of L1 employed a similar Suzuki coupling reaction, but is more straightfor-
ward and is described elsewhere.[33] Ligand H2L2 could not be obtained by
reaction of L1 with BBr3.
1',4'-Bis(tetrahydropyranyloxy)benzene: A suspension of 1,4-dihydroxy-
benzene (5.5 g, 50 mmol) in dichloromethane (125 mL) was treated with
dihydropyran (12.6 g, 150 mmol) and pyridinium tosylate (250 mg,
1 mmol). After stirring for 3 h a clear solution was obtained, which was
diluted with diethyl ether (250 mL) and washed with aqueous potassium
hydroxide (20 mL, 5%), until the aqueous layer was colourless. The organic
layer was dried over anhydrous sodium sulfate and the solvent evaporated
to give a white solid (13.8 g, ꢁ100%), m.p. 98 998C; 1H NMR (300 MHz,
CDCl3, 258C): d ˆ 6.97 (s, 4H; Ph), 5.29 (m, 2H; THP), 3.93 (2H, m; THP),
3.58 (2H, m; THP), 1.99 1.96 (m, 2H; THP), 1.85 1.82 (m, 4H; THP),
1
80%), m.p. 202 2038C; H NMR (300 MHz, [D6]DMSO, 258C): d ˆ 8.62
(d, 2H; Py), 7.92 (m, 4H; Py), 7.61 (d, 2H; Py), 7.45 (m, 3H; Py), 7.28 (m,
1H; Ph), 6.76 (m, 2H; Ph), 4.27 (s, 4H; CH2), 4.21 ppm (s, 2H; CH2);
elemental analysis calcd (%) for C24H22N4O2 ¥ HClO4 ¥ H2O: C55.76, H
4.84, N 10.84; found: C55.98, H 4.85, N 10.83.
[Cu(L1)Cl]Cl (1): L1 (102 mg, 0.24 mmol) and CuCl2 ¥ 2H2O (35 mg,
0.2 mmol) were dissolved in methanol (3 mL) and the solution was stirred
for 10 min. Diethyl ether (15 mL) was then added, and the light blue
microcrystalline solid which formed was collected by filtration and then
recrystallized from methanol/diethyl ether to afford the blue crystalline
product (115 mg, 95%), m.p. 1948C(decomp); X-band EPR (CH 2Cl2,
80 K): gk ˆ 2.236, g? ˆ 2.027, Ak ˆ 176 G; IR (KBr disc): nĈ 3435s, 2942w,
1608s, 1573s, 1498s, 1481m, 1455s, 1432s, 1399m, 1286s, 1206m, 1175w,
1096w, 1041m, 1028s, 890w, 875w, 819m, 801m, 782m, 721w, 655w cmÀ1; UV/
Vis/NIR (methanol): lmax (e) ˆ 259 (14600), 311 (5070), 701 (100), 902 (75);
UV/Vis/NIR (dichloromethane): lmax (e) ˆ 257 (17700), 320 (8000), 770
‡
1.64 1.58 ppm (m, 6H; THP); MS (70 eV): m/z (%): 278 (30) [M ], 194
(10), 110 (100), 85 (65).
6-[2',5'-Bis(tetrahydropyranyloxy)phenyl]-2-pyridylcarboxaldehyde:
A
solution of 1',4'-bis(tetrahydropyranyloxy)benzene (2.0 g, 7.2 mmol) in
THF (30 mL) was treated with n-butyllithium (3.5 mL (2.5 m in hexane),
8.5 mmol). After 2 h the solution was transferred using a cannula to a
solution of triisopropylborate (3.5 mL, 15 mmol) in THF (5 mL) cooled to
À808C. After 30 min, the temperature was allowed to rise to ambient and
stirring was continued for 16 h. Evaporation of the solvent in vacuo gave a
colourless oily residue. This was dissolved in toluene (40 mL), and
6-bromo-2-pyridylcarboxyaldehyde (1.4 g, 7.5 mmol), [Pd(PPh3)4] (0.4 g,
0.35 mmol), methanol (10 mL), and 2m aqueous sodium carbonate (6 mL)
were added. The mixture was heated at reflux for 8 h under nitrogen. After
cooling, dichloromethane (50 mL), 2 m aqueous sodium carbonate solution
(15 mL), and concentrated aqueous ammonia (2 mL) were added. The
mixture was extracted with further dichloromethane, and the organic layer
separated, dried over magnesium sulfate and the solvent removed using a
rotary evaporator. Flash chromatography (silica support; dichloromethane/
petroleum ether 3:1 eluent) of the residue yielded the product aldehyde as
a pale yellow solid (1.6 g, 60%), m.p. 76 778C; 1H NMR (300 MHz,
CDCl3, 258C): d ˆ 10.15 (s, 1H; CHO), 8.12 (m, 1H; Py), 7.88 (m, 2H; Py),
7.59 (s, 1H; Ph), 7.23 (s, 1H; Ph), 7.11 (s, 1H; Ph), 5.44 (m, 1H; THP), 5.36
(m, 1H; THP), 3.95 (m, 1H; THP), 3.82 (m, 1H; THP), 3.59 (m, 2H; THP),
‡
(200), 1050 (130); ESI-MS: m/z (%): 548 (40) {[Cu(L1)(AcO)] }, 526 (100)
‡
{[Cu(L1)Cl] }, 489 (10), 427 (2), 245 (10); elemental analysis calcd (%) for
C26H26Cl2CuN4O2 ¥ H2O: C53.94, H 4.87, N 9.68; found: C54.19, H 4.64, N
9.91.
[Cu(L1)Cl][PF6] (2):
A solution of K[PF6] (9.5 mg, 0.052 mmol) in
methanol (2 mL) was added to [Cu(L1)Cl]Cl (28 mg, 0.05 mmol) in
methanol (5 mL). A pale blue precipitate formed which was recrystallised
from dichloromethane/methanol to yield the blue microcrystalline product
(28.4 mg, 85%), m.p. 2388C(decomp); X-band EPR (powder, 80 K): g1 ˆ
2.22 , g2 ˆ 2.07, g3 ꢁ 2.0 (br); X-band EPR (CH2Cl2, 80 K): g? ˆ 2.235, gk ꢁ
2.0, A? ˆ 112 G, Ak ꢁ 86 G; IR (KBr disc): nĈ 3445br, 2940br, 1609s,
1581m, 1503s, 1461s, 1445s, 1293m, 1286m, 1265m, 1216s, 1163w, 1044s,
1023m, 876s, 816s, 774m, 557s cmÀ1; UV/Vis/NIR (dichloromethane): lmax
(e) ˆ 259 (13500), 311 (6000), 730 (106), 930 (125); ESI-MS: m/z (%): 526
‡
(100) {[Cu(L1)Cl] }, 508 (40), 489 (20); elemental analysis calcd (%) for
C26H26N4ClCuF6O2P: C46.58, H 3.91, N 8.36; found: C46.19, H 3.86, N
8.30.
2.05 (m, 2H; THP), 1.77 (m, 4H; THP), 1.61 ppm (6H, m; THP); IR (KBr
disc): nĈ 1735 (C O) cmÀ1; MS (70 eV): m/z (%): 386 (2) [M ‡ 1], 299
[Cu(H2L2)Cl2] (3): Standing the dark green solution produced by mixing
H2L2 (40 mg, 0.1 mmol) and CuCl2 ¥ 2H2O (17 mg, 0.1 mmol) in ethanol
(5 mL) under an atmosphere of diethyl ether gave jade green crystals of the
product (40 mg, 75%), m.p. 2018C(decomp); X-band EPR (powder, 80 K):
g1 ˆ 2.17, g2 ˆ 2.13, g3 ˆ 2.04; X-band EPR (CH3CN, 80 K): gk ˆ 2.236, g? ˆ
2.027, Ak ˆ 176 G; IR (KBr disc): nĈ 3435s, 3214s, 1608s, 1596s, 1567s,
1492s, 1470s, 1433m 1316m, 1283s, 1211s, 1166w, 1104w, 1052w, 1027m,
993w, 972w, 951w, 874w, 814s, 783s, 770s, 728m, 707m cmÀ1; UV/Vis/NIR
(acetonitrile): lmax (e) ˆ 253 (16400), 336 (3600), 740 (130), 855 (120); ESI-
‡
ˆ
(2.5), 215 (100).
Bis(2-pyridylmethyl){6-[2',5'-bis(tetrahydropyranyloxy)phenyl]-2-pyri-
dylmethyl}amine: 6-[2',5'-Bis(tetrahydropyranyloxy)phenyl]-2-pyridylcar-
boxaldehyde (1.35 g, 5.1 mmol) and bis(2-pyridylmethyl)amine (1.0 g,
5 mmol) were dissolved in 1,2-dichloroethane (30 mL). Sodium triacetoxy-
borohydride (1.6 g, 7.55 mmol) was added. The solution was stirred
overnight under nitrogen. The reaction was quenched with saturated
aqueous sodium bicarbonate solution. The organic layer was separated,
dried over anhydrous sodium sulfate, and the solvent was removed using a
rotary evaporator to give a clear brown oil (2.35 g, 100%); 1H NMR
(300 MHz, CDCl3, 258C): d ˆ 8.55 (d, 2H; Py), 7.72 (m, 6H; Py), 7.64 (m,
1H; Py), 7.50 (m, 1H; Py), 7.25 7.13 (m, 4H; Ph & Py), 7.03 (m, 1H; Ph),
5.39 (m, 1H; THP), 5.30 (m, 1H; THP), 3.94 (m, 4H; CH2), 3.92 (s, 2H;
CH2), 2.05 (m, 2H; THP), 1.77 (m, 4H; THP), 1.61 ppm (m, 6H; THP);
‡
MS: m/z (%): 460 (100) {[Cu(HL2)] }; elemental analysis calcd (%) for
C24H22N4Cl2CuO2 ¥ 1.5H2O: C51.48, H 4.50, N 10.11; found: C51.59, H 4.67,
N 9.96.
[Cu(L3)Cl]Cl (4): Three methods were employed to prepare this complex.
The EPR, FTIR and UV/Vis/NIR spectral data for the product from each
method were identical.
Method 1. Reaction of [Cu(H2L2)Cl2] with DDQ: A solution of DDQ
(1.08 mL, 49.5 mm) in dichloromethane (2 mL) was slowly added to a
stirred suspension of [Cu(H2L2)Cl2] (30 mg, 0.056 mmol) in dichloro-
methane (5 mL) cooled to À108Cby an external coolant bath. The
[Cu(H2L2)Cl2] dissolved and an off-white precipitate formed (DDQH2),
‡
ESI-MS: m/z (%): 566 (10) [M ], 482 (20), 398 (100).
H2L2: Aqueous 10% hydrochloric acid solution was added dropwise to
bis(2-pyridylmethyl)[6-(2',5'-di-tetrahydropyranyloxyphenyl)-2-pyridylme-
thyl]amine (2.35 g, 4.15 mmol) in ethanol (40 mL) until pH 3. The solution
was stirred at 55 608Cfor 3 h. The solvent was removed to near dryness on
Chem. Eur. J. 2003, 9, No. 1
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0127 $ 20.00+.50/0
127

S. B. Colbran et al.
FULL PAPER
which was removed by filtration through a small pad of filter-aid to give a
clear blue solution. Addition of an equal volume of diethyl ether caused the
blue product to precipitate, which was collected by filtration and dried in
vacuo (22 mg, 73%), m.p. 120 1228C; X-band EPR (powder, 80 K): g? ꢁ
2.2 (br), gk ˆ 2.00; X-band EPR (CH2Cl2, 80 K): g? ˆ 2.228, gk ꢁ 2.0, A? ˆ
112 G, Ak ꢁ 86 G; IR (KBr disc): nĈ 2940s, 1659s, 1613s, 1564s, 1530s,
[Cu(H2L2)][PF6] (7): In a nitrogen-filled glovebox, H2L2 (40 mg, 0.1 mmol)
was added to [Cu(MeCN)4][PF6] (37 mg, 0.1 mmol) in CH3CN (10 mL) to
afford a clear yellow solution containing the product. 1H NMR (300 MHz,
CD3CN, 258C): d ˆ 13.68 (s, 1H; OH), ꢀ 9.9 (br s, 1H; OH), 8.50 (d, 2H;
Py), 7.75 7.82 (m, 4H; Py), 7.26 (m, 6H; Py & Ph), 6.82 (m, 2H; Ph), 4.07 (s,
2H; CH ), 3.96 ppm (s, 4H; CH2); ESI-MS: m/z (%): 460 (100)
2
‡
1455m, 1264s, 1173m, 1078w, 1054w, 1029w, 887m, 857w, 827w, 772m cmÀ1
;
{[Cu(H2L2)] }. Attempts to concentrate the solutions containing 7 resulted
UV/Vis/NIR (CH3CN): lmax (e) ˆ 248 (10750), 277 sh (10700), 730 sh (120),
900 (150); UV/Vis/NIR (CH2Cl2): lmax (e) ˆ 246 (10500), 265 sh (9800), 730
sh (100), 900 (130); elemental analysis calcd (%) for C24H20N4Cl2CuO2 ¥
H2O: C52.51, H 4.04, N 10.21; found: C52.78, H 4.04, N 9.93.
Method 2. Reaction of [Cu(H2L2)Cl2] with cerium(IV) ammonium nitrate:
A solution of (NH4)2Ce(NO3)6 (2.75 mL, 40.8 mm) in acetonitrile was
added dropwise to a suspension of [Cu(H2L2)Cl2] (30 mg, 0.056 mmol) in
acetonitrile (5 mL) cooled to À108C. The [Cu(H2L2)Cl2] dissolved afford-
ing a clear solution of [Cu(L3)Cl]Cl, which was used for physical measure-
ments. The spectroscopic data were identical to those for samples of
[Cu(L3)Cl]Cl obtained by method 1 (or 3).
in disproportionation evidenced by the precipitation of copper metal and
the solutions turning green due to formation of unidentified CuII species.
Crystallography
Relevant crystal, data collection and refinement data are summarised in
Table 4. CCDC-175018–CCDC-175020 contain the supplementary crys-
tallographic data for this paper. These data can be obtained free of charge
via www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ,
UK; fax: (‡44)1223-336033; or e-mail: deposit@ccdc.cam.ac.uk).
Method 3. Controlled potential oxidation of [Cu(H2L2)Cl2]: This was carried
out in a conventional three-compartment ™H∫-cell adapted so that it could
be loaded and sealed under an inert atmosphere (high purity nitrogen). An
acetonitrile-filled Ag/AgCl quasi-reference electrode (the same as used in
CVexperiments) was placed in the working compartment along with the Pt
gauze (5 Â 2 cm2) working electrode and a Pt disk mini-electrode for
running CV experiments. The counter electrode was a Pt gauze (4 Â 2 cm2)
and was separated from the working compartment by two fine-porosity
glass frits. During the electrolysis the solutions in the working compartment
was stirred magnetically with a Teflon-coated stirring bar. Anhydrous
acetonitrile (Aldrich) was used as the solvent, and the concentration of
[Cu(H2L2)Cl2] was 2.0 mm and the support electrolyte was 0.2 m
[NBu4][PF6]. The potential used was ‡0.47 V and was applied using the
Pine potentiostat. The electrolysis consumed 1.97 FaradaymolÀ1. Post-
electrolysis electronic and EPR spectra are reproduced in Figure 6 and
Acknowledgement
We are grateful to Professor L. F. Lindoy, the University of Sydney, for the
CuII stability constant measurement for H2L2. This research was funded
from an UNSW Vice-Chancellor×s Goldstar Award (to S.B.C.).
[1] V. Balzani, A. Credi, F. M. Raymo, J. F. Stoddart, Angew. Chem. 2000,
112, 3484; Angew. Chem. Int. Ed. 2000, 39, 3349.
[2] J. K. Lanyi, A. Pohorille, Biotech. Trends 2001, 19, 140.
[3] D. G. Nicholls, S. J. Ferguson, Bioenergetics 2, Academic Press,
London, 1992.
[4] G. T. Babcock, M. Wikstrˆm, Nature 1992, 356, 301; S. Ferguson-
Miller, G. T. Babcock, Chem. Rev. 1996, 96, 2889.
‡
reveal that [Cu(L3)Cl] is produced.
[Cu(HL2)][PF6] (5): [Cu(H2L2)Cl2] (50 mg, 0.093 mmol) was dissolved in
methanol (10 mL) containing NaOH (3.8 mg, 0.095 mmol). A brown
solution formed and K[PF6] (18 mg, 0.1 mmol) in methanol was added to
the aforementioned solution. After 15 min the solvent was removed
yielding a brown solid that was recrystallised from acetone/diethyl ether to
afford golden brown crystals of the product (42 mg, 75%), m.p. 2228C
(decomp); X-band EPR (powder, 80 K): g? ꢁ 2.14 (br), gk ˆ 2.01; X-band
EPR (CH2Cl2, 80 K) g? ˆ 2.20, gk ꢁ 2.0, A? ˆ 115 G, Ak ꢁ 84 G; IR (KBr
disc): nĈ 3428m, 3098m, 1608s, 1567m, 1463s, 1443s, 1408s, 1278s, 1219s,
1211s, 1180m, 1140w, 1105w, 1052w,
[5] S. Yoshikawa, K. Shinzawaitoh, R. Nakashima, R. Yaono, E.
Yamashita, N. Inoue, M. Yao, M. J. Fei, C. P. Libeu, T. Mizushima,
H. Yamaguchi, T. Tomizaki, T. Tsukihara, Science 1998, 280, 1723.
[6] M. Wikstrˆm, Biochim. Biophys. Acta 2000, 1458, 188.
[7] A. Aagaard, G. Gilderson, D. A. Mills, S. Ferguson-Miller, P.
Brzezinski, Biochemistry 2000, 39, 15847.
[8] C. Ostermeier, A. Harrenga, U. Ermler, H. Michel, Proc. Nat. Acad.
Sci. USA 1997, 94, 10547.
[9] M. Wikstrˆm, Biochemistry 2000, 39, 3515.
1020m, 841vs, 814m, 800m, 765s,
Table 4. Numerical crystal and refinement data for the X-ray crystal structures.
559s cmÀ1; UV/Vis/NIR (CH2Cl2): lmax
(e) ˆ 232 (14600), 256 (13400), 286
(12000), 410 (4600), 523 (550), 650
(270), 1000 (90); ESI-MS: m/z (%):
Complex
[Cu(L1)Cl]Cl ¥ 3H2O
[Cu(H2L2)Cl2]
[Cu(HL2)][PF6]
formula (sum)
Mw
C26H32Cl2CuN4O5
615.0
C24H22Cl2CuN4O2
532.9
C24H21CuF6N4O2P
606.0
monoclinic
P21/c
16.738(9)
10.256(3)
13.818(8)
90
96.41(3)
90
2357(2)
4
26.66
4614
0.056 (155)
2627
‡
460 (100) {[Cu(HL2)] }; elemental
crystal system
space group
a [ä]
b [ä]
c [ä]
triclinic
P1
monoclinic
P21/c
7.270(3)
25.539(5)
13.153(5)
90
106.14(2)
90
2346(1)
4
analysis
calcd
(%)
for
≈
C24H21N4CuO2F6P: C47.57, H 3.49, N
9.25; found: C48.91, H 3.54, N 9.34.
9.931(7)
14.798(12
19.716(15
88.42(5)
86.34(5)
75.87(6)
2804(4)
4
[Cu(L1)][PF6] (6): In a nitrogen-filled
glovebox, L1 (42 mg, 0.1 mmol) was
added to [Cu(MeCN)4][PF6] (37 mg,
0.1 mmol) in CH3CN (10 mL). A clear
yellow solution formed. Removal of
the solvent in vacuo afforded a pale-
yellow solid (60 mg, 95%); 1H NMR
(300 MHz, CD3CN, 258C): d ˆ 8.44 (d,
2H; Py), 7.72 (m, 3H; Py), 7.57 (m,
1H; Py), 7.30 (m, 5H; Py), 7.13 (m,
1H; Ph), 6.99 (m, 2H; Ph), 4.04 (s, 2H;
CH2), 3.98 (s, 4H; CH2), 3.74 (s, 3H;
OCH3), 3.66 ppm (s, 3H; OCH3); IR
(KBr disc): nÄ/cmÀ1 ˆ 3063s, 1608s,
1597s, 1577s, 1523w, 1493s, 1460m,
1351m, 1322m, 1268 m, 1213s, 1079vs,
1011w, 854w, 818s, 622s; ESI-MS: m/z
a [8]
b [8]
g [8]
V [ä3]
Z
m [cmÀ1] (C u )
10.13
11.90
Ka
reflections collected
Rmerge (no. of equiv reflections)
observed reflections [I/s(I) > 3]
no. of parameters
5637
0.020 (7)
3227
4459
0.017 (341)
2494
344
298
309
observed reflections/no.
parameters
9.4
8.4
8.5
final R, Rw [I/s(I) > 3]
goodness-of-fit
0.058, 0.078
1.83
0.033, 0.044
1.27
0.061, 0.087
1.80
max., min. peaks in
final difference map (eäÀ3
)
1.09, À0.66
0.44, À0.56
0.80, À0.71
‡
(%): 489 (25) {[Cu(L1)] }, 427 (100).
128
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0128 $ 20.00+.50/0
Chem. Eur. J. 2003, 9, No. 1

Redox-Switched Binding
116 129
[10] H. Michel, Biochemistry 1999, 38, 15129; D. Zaslavsky, R. B. Gennis,
Biochim. Biophys. Acta 2000, 1458, 164.
[11] M. Ralle, M. L. Verkhovskaya, J. E. Morgan, M. I. Verkhovsky, M.
Wikstrˆm, N. J. Blackburn, Biochemistry 1999, 38, 7185; J. P. Osborne,
N. J. Cosper, C. M. V. Stalhandske, R. A. Scott, J. O. Alben, R. B.
Gennis, Biochemistry 1999, 38, 4526.
[12] D. A. Proshlyakov, M. A. Pressler, G. T. Babcock, Proc. Nat. Acad.
Sci. USA 1998, 95, 8020; D. A. Proshlyakov, M. A. Pressler, C.
DeMaso, J. F. Leykam, D. L. DeWitt, G. T. Babcock, Science 2000,
290, 1588.
[28] Y. Naruta, T. Sasaki, F. Tani, Y. Tachi, N. Kawato, N. Nakamura, J.
Inorg. Biochem. 2001, 83, 239; K. D. Karlin, A. D. Zuberbuhler in
Bioinorganic Catalysis, 2nd ed. (Eds.: J. Reedjik, E. Bouwman), M.
Dekker, 1999, Chapter 14; R. A. Ghiladi, K. R. Hatwell, K. D. Karlin,
»
H. W. Huang, P. Moenne-Loccoz, C. Krebs, B. Hanh Huynh, L. A.
Marzilli, R. J. Cotter, S. Kaderli, A. D. Zuberb¸hler, J. Am. Chem.
Soc. 2001, 123, 4515.
[29] J. Q. Chambers, in The Chemistry of the Quinonoid Compounds, Vol.
II, (Eds.: Z. Rappoport, Z. Patai), Wiley, New York, 1988, pp. 719
757.
[13] D. M. Medvedev, I. Daizadeh, A. A. Stuchebrukhov, J. Am. Chem.
Soc. 2000, 122, 6571.
[14] J. Jortner, M. Bixon, Electron Transfer: From Isolated Molecules to
Biomolecules, Wiley, New York, 1999.
[15] B. Eisenberg, Acc. Chem. Res. 1998, 31, 117; D. T. Bong, T. D. Clark,
J. R. Granja, M. R. Ghadiri, Angew. Chem. 2001, 113, 1016; Angew.
Chem. Int. Ed. 2001, 40, 988; G. W. Gokel, A. Mukhopadhyay, Chem.
Soc. Rev. 2001, 30, 274; V. Sidorov, F. W. Kotch, G. Abdrakhmanova,
R. Mizani, J. C. Fettinger, J. T. Davis, J. Am. Chem. Soc. 2002, 124,
2267.
[30] N. Gupta, H. Linschitz, J. Am. Chem. Soc. 1997, 119, 6384.
[31] A. F. Abdelmagid, K. G. Carson, B. D. Harris, C. A. Maryanoff, R. D.
Shah, J. Org. Chem. 1996, 61, 3849.
[32] The oxygen reactivity is perhaps due to (intra- or inter-molecular)
hydroquinone to copper(ii) electron transfer to produce a copper(i)
semiquinone species, see: A. J. Evans, S. E. Watkins, D. C. Craig, S. B.
Colbran, J. Chem. Soc. Dalton Trans. 2002, 983; J. Rall, M. Wanner, M.
Albrecht, E. M. Hornung, W. Kaim, Chem. Eur. J. 1999, 5, 2802.
[33] Z. He, S. B. Colbran, D. C. Craig, J. Chem. Soc. Dalton Trans. 2002,
4224.
[16] M. N. Paddon-Row, Acc. Chem. Res. 1994, 27, 18.
[34] A. W. Addison, A. N. Rao, J. Reedjik, J. Rijn, G. C. Verschoor, J.
Chem. Soc. Dalton Trans. 1984, 1349.
[35] P. Bhattacharyya, J. Parr, A. M. Z. Slawin, J. Chem. Soc. Dalton Trans.
1998, 3263.
[36] S. J. Lange, H. Miyake, L. Que, Jr., J. Am. Chem. Soc. 1999, 121, 6330.
[37] M. M. da Mota, J. Rodgers, S. M. Nelson, J. Chem. Soc. (A) 1969, 2036.
[38] B. J. Hathaway, in Comprehensive Coordination Chemistry, Vol. 5
(Ed.: G. Wilkinson), Pergamon, Oxford, 1987, pp. 533 774.
[39] K. A. Conners, Binding Constants: The Measurement of Molecular
Complex Stability, Wiley-Interscience, New York, 1987.
[40] A. Zweig, W. G. Hodgson, W. H. Jura, J. Am. Chem. Soc. 1964, 86,
4124; V. Le Berre, L. Angely, G. Simonet, G. Mousset, M. Bellec, J.
Electroanal. Chem. 1987, 218, 273.
[41] E. A. Ambundo, M. V. Deydier, A. J. Grall, N. Aguera-Vega, L. T.
Dressel, T. H. Cooper, M. J. Heeg, L. A. Ochrymowycz, D. B.
Rorabacher, Inorg. Chem. 1999, 38, 4233.
[42] C. G. Pierpont, Coord. Chem. Rev. 2001, 216, 99.
[43] B. de Bruin, M. J. Boerakker, J. A. W. Verhagen, R. de Gelder,
J. M. M. Smits, A. W. Gal, Chem. Eur. J. 2000, 6, 298; P. S. Aburel,
C. Romming, K. Undheim, J. Chem. Soc. Perkin Trans. 1. 2001, 1024.
[44] H. Kessler, Angew. Chem. 1970, 82, 237; Angew. Chem. Int. Ed. 1970,
9, 219.
[45] M. J. S. Dewar, W. B. Jennings, J. Am. Chem. Soc. 1971, 93, 401.
[46] H. M. Marques, O. Q. Munro, T. Munro, M. de Wet, P. R. Vashi, Inorg.
Chem. 1999, 38, 2312; A. R. Katritzky, K. Jug, D. C. Oniciu, Chem.
Rev. 2001, 101, 1421.
[17] D. Gust, T. A. Moore, A. L. Moore, Acc. Chem. Res. 2001, 34, 40.
[18] G. Steinberg-Yfrach, P. A. Liddell, S. C. Hung, A. L. Moore, D. Gust,
T. A. Moore, Nature 1997, 385, 239; G. Steinberg-Yfrach, J. L. Rigaud,
E. N. Durantini, A. L. Moore, D. Gust, T. A. Moore, Nature 1998, 392,
479.
[19] S. B. Sembiring, S. B. Colbran, D. C. Craig, Inorg. Chem. 1995, 34, 761.
[20] S. B. Sembiring, S. B. Colbran, D. C. Craig, J. Chem. Soc. Dalton Trans.
1999, 1543.
[21] A. M. Allgeier, C. S. Slone, C. A. Mirkin, L. M. Liable-Sands, G. P. A.
Yap, A. L. Rheingold, J. Am. Chem. Soc. 1997, 119, 550; C. S. Slone,
C. A. Mirkin, G. P. A. Yap, I. A. Guzei, A. L. Rheingold, J. Am. Chem.
Soc. 1997, 119, 10743; E. T. Singewald, C. A. Mirkin, C. L. Stern,
Angew. Chem. 1995, 107, 1725; Angew. Chem. Int. Ed. 1995, 34, 1624.
[22] T. B. Higgins, C. A. Mirkin, Chem. Mat. 1998, 10, 1589; A. M. Allgeier,
C. A. Mirkin, Angew. Chem. 1998, 110, 936; Angew. Chem. Int. Ed.
1998, 37, 894; D. A. Weinberger, T. B. Higgins, C. A. Mirkin, L. M.
Liable-Sands, A. L. Rheingold, Angew. Chem. 1999, 111, 2748; Angew.
Chem. Int. Ed. 1999, 38, 2565; D. A. Weinberger, T. B. Higgins, C. A.
Mirkin, C. L. Stern, L. M. Liable-Sands, A. L. Rheingold, J. Am.
Chem. Soc. 2001, 123, 2503.
[23] K. S. Chen, J. Hirst, R. Camba, C. A. Bonagura, C. D. Stout, B. K.
Burgess, F. A. Armstrong, Nature 2000, 405, 814.
[24] H. M. N. H. Irving, R. J. P. Williams, J. Chem. Soc. 1953, 3192.
[25] A. Neubrand, F. Thaler, M. Kˆrner, A. Zahl, C. D. Hubbard, R.
van Eldik, J. Chem. Soc. Dalton Trans. 2002, 957; A. Pasquarello, I.
Petri, P. S. Salmon, O. Parisel, R. Car, E. Toth, D. H. Powell, H. E.
Fischer, L. Heim, A. E. Merbach, Science 2001, 291, 856.
[26] H. Nagao, N. Komeda, M. Mukaida, M. Suzuki, K. Tanaka, Inorg.
Chem. 1996, 35, 6809; M. Harata, K. Hasegawa, K. Jitsukawa, H.
Masuda, H. Einaga, Bull. Chem. Soc. Jap. 1998, 71, 1031.
[27] K. D. Karlin, S. Kaderli, A. D. Zuberbuhler, Acc. Chem. Res. 1997, 30,
139; M. Schatz, M. Becker, F. Thaler, F. Hampel, S. Schindler, R. R.
Jacobson, Z. Tyeklar, N. N. Murthy, P. Ghosh, Q. Chen, J. Zubieta,
K. D. Karlin, Inorg. Chem. 2001, 40, 2312; H. Hayashi, S. Fujinami, S.
Nagatomo, S. Ogo, M. Suzuki, A. Uehara, Y. Watanabe, T. Kitagawa,
J. Am. Chem. Soc. 2000, 122, 2124; C. He, J. L. DuBois, B. Hedman,
K. O. Hodgson, S. J. Lippard, Angew. Chem. 2001, 113, 1532; Angew.
Chem. Int. Ed. 2001, 40, 1484; A. Wada, M. Harata, K. Hasegawa, K.
Jitsukawa, H. Masuda, M. Mukai, T. Kitagawa, H. Einaga, Angew.
Chem. 1998, 110, 874; Angew. Chem. Int. Ed. 1998, 37, 798.
[47] J. A. Cappuccio, I. Ayala, G. I. Elliott, I. Szundi, J. Lewis, J. P.
¬
Konopelski, B. A. Barry, O. Einarsdottir, J. Am. Chem. Soc. 2002, 124,
1750.
[48] W. M. Harrison, C. Saadeh, S. B. Colbran, Organometallics 1997, 16,
4254.
[49] In this scenario a role for the covalent link within the His 240-Tyr 244
cofactor would be to ensure the redox-triggered dissociation of His
240 from CuB upon oxygen binding drives helix distortion, which
would provide an attractive mechanism for storing the energy for
proton pumping in the later steps (during the reductive phase) of the
enzyme×s catalytic cycle: S. B. Colbran, M. N. Paddon-Row, unpub-
lished results.
Received: May 8, 2002 [F4077]
Chem. Eur. J. 2003, 9, No. 1
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0947-6539/03/0901-0129 $ 20.00+.50/0
129