Models for biological trinuclear copper clusters. Characterization and enantioselective catalytic oxidation of catechols by the copper(II) complexes of a chiral ligand derived from (S)-(-)-1,1'-binaphthyl-2,2'-diamine.

Article abstract of DOI:10.1039/b402539c

The dinuclear and trinuclear Cu(II) complexes of an octadentate ligand derived from (S)-1,1'-binaphthyl-2,2'-diamine have been prepared and characterized by UV/Vis, CD, EPR and NMR spectroscopy. The ligand contains two tridentate aminobis(benzimidazole) d

Full text of DOI:10.1039/b402539c

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Models for biological trinuclear copper clusters. Characterization
and enantioselective catalytic oxidation of catechols by the
copper(II) complexes of a chiral ligand derived from (S)-(−)-1,1′-
binaphthyl-2,2′-diamine†
Maria Chiara Mimmi,a Michele Gullotti,a Laura Santagostini,a Giuseppe Battaini,b
Enrico Monzani,b Roberto Pagliarin,c Giorgio Zoppellarod and Luigi Casella*b
a
Dipartimento di Chimica Inorganica, Metallorganica e Analitica, Università di Milano,
Istituto ISTM-CNR, Via Venezian 21, 20133 Milano, Italy
Dipartimento di Chimica Generale, Università di Pavia, Via Taramelli 12, 27100 Pavia, Italy.
b
E-mail: bioinorg@unipv.it; Fax: +39 0382 528544; Tel: +39 0382 507331
Dipartimento di Chimica Organica e Industriale, Università di Milano, Via Venezian 21,
c
20133 Milano, Italy
d
Max-Planck-Institut für Polymerforschung, Ackermannweg 10, 55128 Mainz, Germany
Received 19th February 2004, Accepted 19th May 2004
First published as an Advance Article on the web 11th June 2004
The dinuclear and trinuclear Cu(II) complexes of an octadentate ligand derived from (S)-1,1′-binaphthyl-2,2′-diamine
have been prepared and characterized by UV/Vis, CD, EPR and NMR spectroscopy. The ligand contains two tridentate
aminobis(benzimidazole) donor arms connected to a central bidentate diaminobinaphthyl linker, which hosts the chiral
unit. In the dinuclear Cu complex the ligation occurs essentially within the tridentate arms of the ligand. The two Cu
centers are EPR nonequivalent and noninteracting. The EPR data suggests that one of the Cu ions additionally interacts
with one of the tertiary aminonaphthyl donors. In the trinuclear complex the two aminonaphthyl donors bind the third Cu
ion. The EPR spectrum of this complex shows the signal for a mononuclear Cu(II) center bound to a tridentate arm, while
the remaining two Cu(II) centers are coupled through hydroxo groups. The CD spectrum shows that in the free ligand a
severe reduction of the dihedral angle between the naphthyl groups from the strain free range occurs. This conformation is
stabilized by ring stacking interactions with the benzimidazole groups. On complex formation this interaction is removed
because the benzimidazole groups are involved in metal binding. In the dinuclear Cu complex the conformation of the
binaphthyl chromophore probably approaches the strain free range, while in the trinuclear Cu complex a marked flattening
of the dihedral angle between the two naphthyl rings occurs. Both complexes are active catalysts in the oxidation of L-/D-
Dopa derivatives to quinones. High enantioselectivity is observed in the oxidation of L-/D-Dopa methyl ester catalyzed by
the dinuclear Cu complex, which exhibits strong preference for the D enantiomer. The enantioselectivity is largely lost for
the trinuclear Cu complex.
we described the ligand (R)–L carrying a chiral (R)-(+)-1,1′-binaph-
Introduction
thyl-2,2′-diamine central core acting as a spacer between two chelat-
Dinuclear and trinuclear copper complexes derived from polyden-
ing arms containing tridentate aminobis(benzimidazole) donors.10,11
tate ligands have received increasing attention in recent years1 for
The catalytic activity of the corresponding dinuclear and trinuclear
the challenge of mimicking various aspects of the structure and re-
copper(II) complexes in the oxidation of catecholic L,D-Dopa deriva-
activity that are generally associated with the type 3 or type 2–type
tives was briefly reported.10 Though, limited stereoselectivity effects
3 copper centers found in several proteins, such as hemocyanin,2 ty-
were observed in these reactions. In this paper we report the com-
rosinase3 and the multicopper oxidases laccase,4 ascorbate oxidase5
plete spectroscopic, magnetic and conformational characterization
and ceruloplasmin.6 In the trinuclear cluster of these enzymes, a
of the dinuclear and trinuclear copper(II) complexes derived from
total of eight histidine imidazole groups provide the protein ligands
the enantiomeric octadentate ligand (S)–L, derived from (S)-(−)-
for the metal centers, with the CuN3 donor set for the pair of type
1,1′-binaphthyl-2,2′-diamine (Scheme 1). The trinuclear complex
3 centers and the CuN2 donor set for the type 2 center. Mimicking
contains two metal coordination sites, labeled as type A sites, at
the cluster of the latter enzymes is therefore particularly challeng-
the tridentate aminobis(benzimidazole) units and a third bidentate
ing from the synthetic point of view, since it requires the design of
coordination site, labeled as B site, at the chiral 1,1′-binaphthyl
an octadentate ligand providing an appropriate distribution of the
residue, which simulate the protein donor environment of the Cu(II)
nitrogen donor groups among the metal centers and the necessity to
centers occurring in the biological clusters (Scheme 2).10 The cata-
keep these centers in close proximity (between 3.5 and 4.0 Å) and
lytic activity of the complexes toward the L,D-Dopa substrates was
electronically coupled.
also investigated, in order to ascertain whether the limited stereose-
We previously reported some biomimetic octadentate ligands with
lectivity observed with the catalysts of the (R)–L series was due to a
nitrogen donor atoms meeting the required trinucleating ability, and
mismatch between the catalysts and substrate chirality.
the corresponding metal complexes.7–10 In our most recent approach,
Results and discussion
Optical and CD spectroscopy
† Electronic supplementary information (ESI) available: Synthetic details
for the preparation of the ligand (S)-(+)-N,N′-dimethyl-N,N′-bis{3-[bis-(1-
The synthesis of (S)–L mirrors the multistep procedure followed
methyl-2-benzimidazolylmethyl)amino]-propyl}-1,1′-binaphthyl-2,2′-di-
for (R)–L;11 details on the characterization of the intermedi-
amine, (S)–L, with all the spectral data for the intermediates involved, and
ates involved are collected in the ESI.† As already reported for
(R)–L,10 the electronic spectrum of (S)–L in the near-UV region
contains contributions from the 1,1-binaphthyl and benzimidazole
representative paramagnetic 1H NMR spectra of the complex [Cu3(S)–L]6+
showing the pH and temperature dependence of the spectra (Figs. 1S
and 2S). See http://www.rsc.org/suppdata/dt/b4/b402539c/
2 1 9 2
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

View Article Online
Scheme 1 Structure of the ligand (S)–L.
Fig. 1 Electronic and circular dichroism spectra in acetonitrile solution of
(a) (−·−) (S)–L, (b) (−−−) [Cu2(S)–L]4+, (c) (—) [Cu3(S)–L]6+.
site B by the tertiary aminonaphthyl groups may require severe
distortion of the 1,1′-binaphthyl chromophore toward coplanarity
of the two rings.
Some interesting differences occur in the near-UV optical and
CD spectra of [Cu2(S)–L]4+ and [Cu3(S)–L]6+. The envelope of
the intense intraligand electronic bands between 250 and 300 nm
is slightly blue shifted in the spectra of the complexes and both
exhibit much better resolved vibrational structure for the benzimid-
azole absorptions around 280 nm, as a result of the strongly reduced
conformational freedom of these residues upon coordination to the
metal ions. In addition, the absorption spectrum of [Cu3(S)–L]6+
displays a prominent band centered at 362 nm, and extending from
320 and 400 nm, which is resolved in two components at 345 and
390 nm in the CD spectrum (both of negative sign) (Fig. 1). In the
same region, the spectrum of [Cu2(S)–L]4+ shows a much weaker
optical band which is nearly coincident with the (positive) CD peak
at 352 nm. The origin of these spectral features is clearly different
in the two complexes. For [Cu2(S)–L]4+ the near-UV band is es-
sentially an intraligand band, as testified by the similar optical and
CD features exhibited by (S)–L, which are only red shifted with
respect to those of the complex. In the case of [Cu3(S)–L]6+, the two
additional electronic bands of moderate intensity and opposite CD
activity in the 320–400 nm range can be attributed to LMCT tran-
sitions. The position of these transitions, at relatively low energy,
and their significant intensity, exclude that they can originate from
N(amino) → Cu(II) or N(benzimidazole) → Cu(II) LMCT.13,14 They
actually look very much like the OH− → Cu(II) LMCT transitions
originating from dissociated water molecules bound as bridging
ligands to pairs of Cu(II) centers.13–15 Support to this interpretation
comes from the EPR experiments, the pH dependence of the spectra
and the azide binding behavior (see below).
Scheme 2 Proposed structure for the trinuclear copper(II) complex
[Cu3(S)–L]6+, containing sites A and B. The dinuclear copper(II) complex
[Cu2(S)–L]4+ contains only sites of type A.
chromophores. Thus, we can assign to the 1,1-binaphthyl-2,2′-di-
substituted chromophore five main absorption bands at 210, 220,
257, 310, 350 nm,12 and to the benzimidazole chromophores the
bands at 270, 278 and 286 nm,13 while the latter chromophore also
contributes to the bands at 257 and 350 nm. As expected, the CD
spectrum of (S)–L displays essentially mirror image behavior to that
of (R)–L.11 The spectrum consists of a weak positive Cotton effect at
367 nm followed by two strong dichroic bands at 296 nm (negative)
and 264 nm (positive), by a weak band at 246 nm (negative), and
by an extremely intense exciton doublet featuring a positive peak at
225 nm and a negative peak at 215 nm (Fig. 1). This pattern indi-
cates that (S)–L assumes a conformation with P helicity between the
two naphthyl rings of the 1,1-binaphthalene core. The intensity of
the CD couplet, and the characteristic CD feature at 246 nm, attrib-
utable to an inter-naphthalenic CT transition,11 indicates that also for
(S)–L a significant distortion in the arrangement of the two naphthyl
rings from the strain free conformation of about 90° occurs.12 As
shown by the computational studies performed on (R)–L,11 this
conformational arrangement is due to strong ring stacking interac-
tion between the naphthyl groups and the benzimidazole residues
present in the lateral arms of the molecule.
The exciton couplet has somewhat reduced intensity in the CD
spectrum of [Cu2(S)–L]4+ (Fig. 1), suggesting that the dihedral
angle between the naphthyl rings undergoes some changes, as ob-
served for the parent dinuclear Zn(II) and Cu(II) complexes in the
(R)–L series,10,11 and probably relaxes approaching the strain free
range, as a result of the participation of the benzimidazole rings in
metal binding. In contrast, the CD couplet is so strongly perturbed
in the spectrum of [Cu3(S)–L]6+ (as it is for [Cu3(R)–L]6+)10 that
basically only the negative limb, at 210 nm, can be recognized,
while the positive limb, at 223 nm, virtually disappears. We can
explain these data considering that chelation of the Cu(II) ion in
The optical spectra of the Cu(II) complexes additionally possess
low energy absorption bands between 600 and 800 nm that encom-
pass the metal LF transitions. For [Cu2(S)–L]4+ the LF band occurs
at 698 nm, while for [Cu3(S)–L]6+ the LF bands appear as shoulders
at 646 and 730 nm on the low energy tail of the more intense LMCT
absorptions. The visible CD spectra exhibit weak positive Cotton
effects, at 619 nm for the dinuclear complex, and at 490 nm for the
trinuclear complex, with an additional extremely weak, broad CD
band of negative sign centered around 610 nm.
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1
2 1 9 3

View Article Online
It is worth noting that the chiroptical properties of the di-
nuclear complex [Cu2(R)–L]4+ in the near-UV and visible range
display some notable difference from the expected mirror image
behavior with respect to [Cu2(S)–L]4+. In fact, the CD spectrum of
[Cu2(R)–L]4+ exhibits a positive (instead of negative) CD band at
352 nm, and two bands at 524 nm (positive) and 635 nm (negative)
instead of the single band at 619 nm observed for [Cu2(S)–L]4+.10
On the contrary, the chiroptical behavior of the trinuclear complex
[Cu3(R)–L]6+ is as expected, with two positive CD bands in the
near-UV region, at 360 and 395 nm, and extremely weak CD
activity in the 500–700 nm range. The origin of the non-mirror
image behavior of the dinuclear complexes [Cu2(S)–L]4+ and
[Cu2(R)–L]4+ is not known, but is likely due to some diastere-
oselectivity occurring on complex formation, which affects the
solubility or rate of formation of the precipitate in the preparation
of the complexes. Among the possible sources of diastereoselec-
tivity are: (i) the folding of the chelating arms carrying the two
metal centers at sites A above and below the two naphthyl rings,
which can be induced by aromatic ring stacking interactions and
may generate a sort of (chiral) helical arrangement of the whole
molecule; (ii) the chirality at one or both Cu(II) centers, provided
that the two benzimidazole donors are not equivalent in the coor-
dination sphere (i.e. one axial and one equatorial). This point could
only be assessed by structural investigations, though we note that
a difference in coordination state of the Cu(II) centers between
[Cu2(S)–L]4+ and [Cu2(R)–L]4+ will also emerge from the EPR
spectra of the complexes described below.
EPR studies
The EPR spectra of diluted (0.3 mM) glassy solutions for the
dinuclear (line a) and trinuclear (line b) copper(II) complexes are
shown in Fig. 2. Both spectra are characteristic for axial systems
and despite the different number of copper centers they exhibit
similar EPR envelope. The signal of [Cu2(S)–L]4+ reveals that the
two coppers have a slightly different degree of distortion toward a
square pyramid (g|| > g⊥); one copper(II) ion, labeled Cua, undergoes
stronger axial interaction and shows the following spin-Hamiltonian
parameters, ga|| = 2.288, ga⊥ = 2.063, Aa|| = 151 × 10−4 cm−1, while
for the second copper(II) ion, labeled Cub, gb|| = 2.255, gb⊥ = 2.066,
Ab|| = 159 × 10−4 cm−1. In addition, the signal of Cua begins to satu-
rate at 127 mW while the signal of Cub remains unaffected (Fig. 2,
upper inset). The double integration of the signal intensity against
CuEDTAstandard accounts for 2.0 ± 0.25 spins and hence confirms
the dinuclear nature of the complex and the absence of significant
interaction between the Cu(II) centers. As it is known from struc-
turally characterized mononuclear16 and dinuclear17 Cu complexes
with analogous N3 donor environment, Cu binding at A sites favors
the adoption of five-coordinated structures. Therefore, we can con-
clude that in [Cu2(S)–L]4+ the arrangement of the ligand allows one
of the tertiary amino groups of the binaphthalenediamine residue to
interact with a Cu center (Cua), while solvent molecules are bound
to Cub (Scheme 4, structure Ia). On increasing the concentration of
the sample tenfold (3 mM) a broader and unresolved EPR pattern is
observed (Fig. 3, line a), with different EPR parameters (g|| = 2.240,
g⊥ = 2.065, A|| = 175 × 10−4 cm−1), accompanied by the presence of
the half-field transition at g ~ 4.28 (inset in Fig. 3). This phenom-
enon arises by dipolar interactions between copper(II) centers in
neighboring molecules. The appearance of a broad peak at high-
field (marked with an asterisk in Fig. 3, line a, at ~345 mT) further
supports the coexistence in solution of an appreciable amount of
dimeric species. The double integration of the signal intensity ac-
counts for 2.20 ± 0.15 spins. The intermolecular association likely
involves the naphthyl residues and the conformational rearrange-
ment produced induces the loss of the tertiary amine–Cua bond and
its replacement with a weaker solvent donor molecule, making the
two Cu(II) ions equivalent (Scheme 4, structure Ib). Interestingly,
the EPR spectrum of [Cu2(R)–L]4+ differs from that of [Cu2(S)–L]4+,
as it shows a single signal for two equivalent Cu(II) centers, with
EPR parameters (g|| = 2.275, g⊥ = 2.064 A|| = 155 × 10−4 cm−1) in-
termediate between those of the Cua and Cub centers of [Cu2(S)–L]4+
and no half-field transition at g ~ 4.3. We believe that the tertiary
amine interaction with one of the coppers (as in structure Ib in
Scheme 4) in this case is missing.
Ligand binding experiments
The addition of azide to acetonitrile solutions of [Cu2(S)–L]4+ pro-
duces the growth of a moderately intense absorption band in the
range between 350 and 450 nm, with a maximum at 404 nm. From
spectral titrations it is possible to estimate the strength of the bind-
ing of azide to the copper(II) centers. In general, the spectra, normal-
ized for dilution, show isosbestic points and it has been possible to
differentiate the binding of successive ligand molecules. The Hill
equation was used for determination of the equilibrium constants
(K) and stoichiometry (n) of the adducts. By selecting appropriate
−
ranges of [N3 ]:[Cu2], four equilibrium constants could be cal-
culated from the plots: K1 = 4800 M−1 (n = 1.01), K2 = 3400 M−1
(n = 1.03), K3 = 2500 M−1, (n = 1.00), and K4 = 1700 M−1,
(n = 1.02). The symmetric shape of the N3− → Cu(II) LMCT band
that develops shows that the ligand in all cases binds in the terminal
mode (see Scheme 3).13
The EPR spectrum of the complex [Cu3(S)–L]6+ in diluted glassy
solution (0.4 mM) is more intriguing and it is shown as line b in Fig.
2. It displays the following spin-Hamiltonian parameters, g|| = 2.277,
g⊥ = 2.065, A|| = 163 × 10−4 cm−1. Despite the fact that the complex
contains three copper(II) ions, the double integration of the signal
intensity against CuEDTA standard accounts for only 1.1 ± 0.15
spins. Neither extra signals are detected by increasing or decreasing
the microwave power nor half-field transition is observed, confirm-
ing the presence of a single EPR active copper(II) center. Thus, the
two remaining copper(II) ions are EPR non-detectable and need to
strongly antiferromagnetically interact. Since the EPR signal is
again typical for Cu(II) in the A site, the coupled dimer is formed by
the Cu(II) center at B site and the remaining Cu(II) at the other A site.
As suggested by the LMCT pattern observed in the UV-Vis and CD
spectra of the complex (Fig. 1), the antiferromagnetic coupling is
mediated by a double hydroxide bridge (Scheme 4, structure II). By
increasing the concentration of the sample tenfold (4 mM) a broader
EPR pattern is observed (Fig. 3, line b) with spin-Hamiltonian pa-
rameters (g|| = 2.240, g⊥ = 2.065, A|| = 177 × 10−4 cm−1) very similar
to those found for [Cu2(S)–L]4+ at high concentration. Though, un-
like the solution of [Cu2(S)–L]4+ at high concentration, no half-field
transition is observed here in the EPR spectrum, excluding the pres-
ence of dimeric forms. The double integration of the signal intensity
accounts for only 0.9 ± 0.2 spins. It appears, therefore, that the more
“folded” geometry of the trinuclear copper complex, with respect
Scheme 3 Schematic representation of the binding of azide to the
[Cu2(S)–L]4+ complex.
In the titration of [Cu3(S)–L]6+ with azide no significant spectral
change occurs until two equivalents of the ligand are added, but
−
upon further addition of azide, up to a ratio of [N3 ]:[Cu3] ~ 10:1,
a symmetric band develops between 350 and 450 nm, with a
maximum at 385 nm, but without isosbestic points. Thus, binding
of azide is initially inhibited by the existence of hydroxo ligands
bound to the Cu(II) centers. The plot of absorbance change against
azide concentration was sigmoidal, indicating multiple binding of
ligand molecules to the copper centers of the complex. A similar
behavior was noted for [Cu3(R)–L]6+,10 but in that case a complex
with a single bridging azide could be more clearly identified at low
azide concentration.
2 1 9 4
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1

View Article Online
Fig. 3 EPR spectra of the dinuclear (a) and trinuclear (b) Cu(II) com-
plexes of (S)–L recorded in frozen CH3CN/MeOH solutions (1/9 v/v). The
complex concentrations were 3 mM and 4 mM for (a) and (b), respectively.
Experimental parameters are the same as those reported in Fig. 2. The
upper inset shows the half-field transition (ms =2) for (a) recorded with
the following experimental parameters: 9.4040 GHz, 100 kHz modulation
frequency, 0.5 mT modulation amplitude, 40.0 mW power, 82 ms time
constant, 42 s sweep time, temperature 105 K. 30 scans were accumulated
and averaged and both the cell and cavity background (recorded in the same
conditions) were subtracted.
Fig. 2 EPR spectra of the dinuclear (a) and trinuclear (b) Cu(II)
complexes of (S)–L recorded in frozen CH3CN/MeOH solutions (1/9
v/v). The complex concentrations were 0.3 mM and 0.4 mM for (a) and
(b), respectively. Experimental parameters: 9.4040 GHz, 100 kHz modula-
tion frequency, 1.0 mT modulation amplitude, 1.27 mW power, 82 ms time
constant, 84 s sweep time, temperature 110 K. 6 scans were accumulated
and averaged. The upper inset shows the low-field envelope for the solution
(a) of the dinuclear complex recorded at different microwave powers. The
EPR spectrum of the dinuclear complex [Cu2(R)–L]4+ is shown as trace (c); it
was recorded at 9.4050 GHz with the same parameters as in (a) and (b). The
complex concentration was 0.4 mM in CH3CN/MeOH solution (1/9 v/v).
to the dinuclear complex, prevents the metal core to interact with
neighboring molecules. The EPR spectral properties of the parent
[Cu3(R)–L]6+ complex are quite similar.
A pH titration experiment of the complex [Cu3(S)–L]6+ followed
by EPR in diluted glassy solution (0.15 mM, MeOH/H2O/CH3CN)
showed a strong pH dependence of the signal. In the spectrum re-
corded at pH 10 (Fig. 4, line a) the presence of only one type of
copper, with g|| = 2.258 and A|| = 179 × 10−4 cm−1, can be detected.
The double integration of the signal intensity accounts for 1.0 ± 0.15
spins. The EPR parameters are comparable with those previously ob-
served in the solvent without water as outlined above. By decreasing
the pH down to neutral (pH 7, line b) a new dominating copper fea-
ture appears, with g|| = 2.28 and A|| = 154 × 10−4 cm−1, with additional
weaker signals (marked with arrows in the figure), with parameters
reminiscent of the species present at pH 10. The double integration
of the signal intensity accounts for 1.35 ± 0.25 spins. At pH 5.0 (Fig.
4, line c) clearly two distinct types of copper can be recognized in
the spectrum; one with identical parameters to those of the dominant
component observed at pH 7 (g1|| = 2.28 and A1|| = 154 × 10−4 cm−1)
and another with stronger tetragonal field (g2|| = 2.250 and
A2|| = 182 × 10−4 cm−1). Further, an unusual signal (marked with an
asterisk) appears in the high-field region of the spectrum (g ~ 1.95)
that cannot arise from a double-quantum transition since the micro-
Scheme 4 Schematic representation of the coordination state of the Cu(II)
centers in [Cu2(S)–L]4+ (Ia) and [Cu3(S)–L]6+ (II). Structure Ib represents
the coordination change undergone by the dinuclear Cu(II) complex upon
association, in a concentrate solution.
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1
2 1 9 5

View Article Online
wave power applied in the measurement was kept very low. The
double integration of the signal intensity now accounts for 1.9 ± 0.3
spins. In all cases, no half-field transition was detected, even when
the concentration of the sample was increased tenfold, thus exclud-
ing the presence of triplet-state species. Note that the EPR changes in
the pH range examined are fully reversible, suggesting the presence
of a complex intramolecular equilibrium depending on pH, among
the three coordinated coppers, that could be mediated by one or more
protonable bridging groups. Parallel changes with pH are observed
in the optical spectrum of the trinuclear complex (Fig. 4, inset). On
increasing the pH (up to pH 10), a broad shoulder on the near UV
LMCT band, around 340 nm, progressively develops and also this
change appears to be fully reversible. Although it is difficult to fully
rationalize the pH dependent changes undergone by the trinuclear
complex, it seems clear the main species at neutral pH has a struc-
ture similar to that shown in Scheme 4 (structure II), with one or
two water molecules bound to the paramagnetic Cu A site. At basic
pH, one of these water molecules undergoes proton dissociation,
as shown by the intensity increase in the near UV region; the EPR
parameters testify the stronger tetragonal field in the resulting Cu(II)
center. In the acidic pH range several species are apparently pres-
ent, since also in these conditions the bridge between the dinuclear
Cu(A)–Cu(B) site can be cleaved. The EPR signal at g ~ 1.95 can be
interpreted as the high field component of a rhombic signal. Signals
with such small g values have little precedent in the Cu literature,
but a relevant example has been observed for the so called “native
intermediate” formed in the reduction of O2 at the trinuclear Cu clus-
ter of Rhus vernicifera laccase, in which all three Cu(II) atoms of the
cluster are coupled to give a ground state Stot = 1/2 spin system.18
EPR experiments were also carried out with [Cu2(S)–L]4+ and
[Cu3(S)–L]6+ in the presence of variable amounts of azide, in order
to probe the interactions of the two complexes with this exogenous
ligand. Upon addition of 1 equiv. azide to the solution of the bi-
nuclear complex (0.3 mM concentration), two different types of
copper signals can be observed in the EPR spectrum (Fig. 5, line
a). While the spin-Hamiltonian parameters for Cua remain almost
unchanged (ga|| = 2.280, ga⊥ = 2.059, Aa|| = 152 × 10−4 cm−1), which
seems to exclude a direct interaction of azide with this metal cen-
ter, those related with the Cub site shift as follows, gb|| = 2.245,
g
b⊥ = 2.049 and Ab|| = 175 × 10−4 cm−1. It is thus clear that the first
azide molecule binds to this Cub site and that the metal center is
now subject to a stronger tetragonal field. Surprisingly, the EPR
spectrum additionally shows the presence of the forbidden half-field
transition (ms = 2) (upper inset in Fig. 5, g ~ 4.29) accounting for
the presence of molecules in a triplet state. The spin quantitation
accounts for 2.35 ± 0.20 spins. In the absence of structural infor-
mation we suggest that an asymmetric end-on azido bridge (-1,1)
should be formed, with strong binding to Cub and weak binding to
Cua.19–22 This hypothesis satisfactorily explains the observed sym-
−
metric shape of the N3 → Cu(II) LMCT band, since this is essen-
tially determined by the interaction of the ligand with the Cub site.
Fig. 5 EPR spectra of the [Cu2(S)–L]4+ complex upon incubation with one
(line a), two (line b), and ten equiv. (line c) azide in CH3CN/MeOH solutions
(1/9 v/v). The final complex concentration was 0.3 mM for (a), (b), and
(c). Experimental parameters: 9.4040 GHz, 100 kHz modulation frequency,
0.5 mT modulation amplitude, 2.4 mW power, 82 ms time constant, 84 s
sweep time, temperature 120 K. 6 scans were accumulated and averaged.
The upper inset shows the half-field transition (ms =2) for (a) recorded
with the following experimental parameters: 9.4070 GHz, 100 kHz modula-
tion frequency, 0.5 mT modulation amplitude, 49.0 mW power, 82 ms time
constant, 42 s sweep time, temperature 120 K. 30 scans were accumulated
and averaged and both the cell and cavity background (recorded in the same
conditions) were subtracted.
Fig. 4 The pH titration of the [Cu3(S)–L]6+ complex followed by EPR: pH
10 (line a), pH 7 (line b) and pH 5 (line c). The final complex concentration
was 150 M for (a), (b) and (c); the solutions were in CH3CN/MeOH 1/9
(v/v) containing 4% (v/v) aqueous buffer (phosphate buffer at pH 5 and 7, and
carbonate buffer at pH 10). In all cases the copper complex was initially dis-
solved in CH3CN/MeOH, then aliquots of buffer solutions were added. Before
running the spectra, the samples were aged for 60 min at room temperature.
Experimental conditions: (a) 9.4054 GHz, (b) 9.3989 GHz, (c) 9.3962 GHz,
100 kHz modulation frequency, 1 mT modulation amplitude, 84 s sweep time,
82 ms time constant, 1.6 mW power, temperature 120 K, 10 scans were accu-
mulated and averaged. The inset shows for (a), (b) and (c) the corresponding
near-UV absorption spectra (pH 10, pH 7, pH 5) by using the same quartz
EPR cell tube as employed in the EPR experiments. The asterisk in spectrum
(c) marks an unusual peak with g < 2.0 (see text).
Incubation of a solution of [Cu2(S)–L]4+ with two equiv. azide
yields a single type of EPR signal (Fig. 5, line b), with the fol-
lowing spin-Hamiltonian parameters, g|| = 2.243, g⊥ = 2.049 and
A|| = 176 × 10−4 cm−1. The spin quantitation accounts for 1.90 ± 0.20
spins and no half-field transition is observed. Hence, the weak
asymmetric mono azido-bridge is displaced and geometrical rear-
rangements in the ligand coordination sphere at the Cua site occur,
leading to structurally identical Cu(II) centers. The two azide mol-
ecules bind each copper in the terminal mode, in agreement with
the ligand binding experiments. Upon further addition of up to ten
2 1 9 6
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1

View Article Online
equiv. azide, the EPR envelope is broadened (Fig. 5, line c), and
the following parameters can be estimated, g|| = 2.239, g⊥ = 2.090,
A|| = 180 × 10−4 cm−1. The spin quantitation accounts for 2.0 ± 0.15
spins and no half-field transition is observed; thus the azide ligands
bind in the terminal mode, with a stoichiometry of two azide mol-
ecules for each Cu(II) center (Scheme 3).
1H NMR spectra
1
As for the EPR spectra, also the paramagnetic H NMR spectra
of [Cu2(S)–L]4+ and [Cu3(S)–L]6+ appear qualitatively similar. As
shown in Fig. 7, the spectra of [Cu2(S)–L]4+ and [Cu3(S)–L]6+ in
CD3CN display, besides the broad paramagnetic peaks around 30,
20 and 0 ppm, an envelope of sharp and resolved resonances in the
aromatic region, between 7 and 9 ppm. This cluster of peaks prob-
ably belongs to the protons of the benzimidazole chelating arms
and the 1,1′-binaphthyl spacer of the ligand. It is interesting to note
that in the trinuclear complex the intensity of this cluster of peaks is
significantly reduced and at the same time, a new broad resonance is
detectable as a shoulder to the 0 ppm signal. This reflects the coordi-
nation of the 1,1′-binaphthyl moiety to the Cu(II) center in site B.
The EPR spectrum obtained for the trinuclear copper complex
in the presence of azide is very intriguing. Upon addition of four
equiv. azide to [Cu3(S)–L]6+ solution (0.4 mM) (Fig. 6, line a),
only the presence of a mononuclear Cu(II) species is observed,
with a spectral pattern which is reminiscent of Cu(II) in a strongly
distorted rhombic field but with an unusual component in the high
field region (upper inset in Fig. 6), with g = 1.96. This signal is un-
likely to come from “normal” copper since the spin–orbit coupling
constant is strongly negative and hence one would not expect a g
value less than 2.00. The presence of molecules in a triplet-state is
ruled out by the absence of a half-field transition. A similar EPR
signal (with g = 1.86) has been observed for Rhus vernicifera lac-
case upon azide binding; this feature has been attributed to the
establishment of an azide bridge between the type 2 and one of the
antiferromagnetically coupled type 3 coppers,23 but is observable
only at cryogenic temperature. We tentatively suggest that a similar
spin system is generated in the present case, where the spin density
seems distributed all over three bridged Cu(II) centers making an
unusual Stot = 1/2 system. Therefore, azide acts as a bridge between
the two previously separated copper systems, i.e. the Cu center
behaving as an isolated spin and the two Cu(II) centers bridged
and antiferromagnetically coupled by an hydroxo group. The fol-
lowing spin-Hamiltonian parameters are then associated with the
pseudo-rhombic spin system, g1 = 2.26, g2 = 2.045, g3 = 1.96 and
A1 = 175 × 10−4 cm−1. Upon further incubation of the [Cu3(S)–L]6+
complex with up to ten equiv. azide (Fig. 6, line b), no significant
changes in the EPR envelope can be observed except for small sig-
nals around 260 mT, that likely originate from a small fraction of
copper being released.
Fig. 7 1H NMR spectra of the compounds [Cu2(S)–L]4+ (a) and [Cu3(S)–
L]6+ (b) recorded at 298 K in CD3CN, with sample concentrations of about
1 mM for both the complexes. The insets, showing the spectral region
around 20 ppm, were recorded using a super WEFT sequence with a recycle
delay of 70 ms.
The pH and temperature dependence of the 1H NMR spectrum of
[Cu3(S)–L]6+ were studied in a mixed solvent of CD3CN-deuterated
buffer in the pH range between 5 and 10 and in the temperature
range between −35 and +40 °C. The main features of the spectrum
remain the same as those observed in pure CD3CN (Fig. 7b), with
the cluster of peaks in the aromatic region, the broad signals in the
downfield region and around 0 ppm, throughout the pH range inves-
tigated. Two representative series of 1H NMR spectra showing the
pH and temperature dependence are included in the ESI.† Changing
the temperature alters only slightly the broad downfield peaks (es-
sentially causing further broadening) and the peaks in the aromatic
region. The mostly affected signal is that around 0 ppm, which
undergoes an appreciable upfield shift of 1.5–2 ppm on decreasing
the temperature from +40 to −20 °C. These changes occur at all
pH values and also in pure CD3CN. The variations of the chemical
shifts with temperature are almost linear, preventing the determina-
tion of the diamagnetic contribution to the chemical shift; but it can
be anticipated that this value is far downfield from 0 ppm. Prob-
ably, the signal at about 0 ppm is connected to the protons of the
benzimidazole rings directly coordinated to the Cu(II) ions, which
undergo an upfield shift of dipolar origin.
Fig. 6 EPR spectra of the [Cu3(S)–L]6+ complex upon incubation with
four (line a) and ten equiv. (line b) of azide in CH3CN/MeOH solutions (1/9
v/v). The complex concentration was 0.4 mM. Experimental parameters
are the same as those reported in Fig. 5. The upper inset shows the singly
integrated EPR spectrum for (a), and the lower inset the magnified low-field
region for trace (b).
1
Fig. 8 shows the H-1H DQF-COSY (Double-Quantum Fil-
1
tered H-1H COSY) spectrum registered on a 8 mM solution of
[Cu2(S)–L]4+ in CD3CN. The cross-peaks in the aromatic region
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1
2 1 9 7

View Article Online
Table 1 Magnetic susceptibility data for the copper(II) complexes determined in CD3CN solution
[Cu2(S)–L]4+ [Cu3(S)–L]6+
T/K
/Hz
Mparaa /m3 mol−1 × 108

eff (B)b
T/K
/Hz
Mparaa /m3 mol−1 × 108

eff (B)c
1,4-Dioxane
298
308
318
328
28.61
2.26
2.24
2.30
2.21
2.07
2.09
2.16
2.15
298
308
318
328
27.29
25.87
24.66
23.68
3.09
2.95
2.82
2.73
2.42
2.40
2.39
2.39
28.28
29.27
27.95
tert-Butyl alcohol
298
308
318
328
29.49
29.05
30.04
28.61
2.30
2.27
2.33
2.24
2.09
2.11
2.17
2.16
298
308
318
328
28.06
26.64
25.32
24.34
3.15
3.00
2.87
2.77
2.44
2.43
2.41
2.41
a The Mpara are corrected with the diamagnetic contribution measured for the corresponding zinc derivatives. b The theoretical value of eff (B) for two non-
interacting Cu(II) centers in a dinuclear compound is 2.45. c The theoretical value of eff (B) for three noninteracting Cu(II) centers in a trinuclear compound
is 3.0.
clearly show the scalar connectivity between four and two adjacent
protons, most probably belonging to the 1,1′-binaphthyl residue. In
the aliphatic region of the COSY spectrum a cluster of cross-peaks
which may be associated to the propyl chain which connect the
1,1′-binaphthyl with the amino-bis(benzimidazole) residue can be
observed. It is worth noting that the COSY spectrum has been re-
corded on a concentrated complex solution and it is consistent with
structure Ib of Scheme 4, that according to the EPR measurement is
the major species above 4 mM concentration.
centration of sample needed for the Evans measurements. The tem-
perature variation does not markedly affect the magnetic moment
values, thus excluding the presence of thermally accessible excited
states. The magnetic moments calculated for [Cu3(S)–L]6+ are more
intriguing. The reduced eff values indicate that a more sizable cou-
pling occurs in this case; this coupling is quite certainly mediated
by some small bridging ligands like water or hydroxide. In contrast,
the temperature variation does not show appreciable change in the
magnetic moment. This behavior is not easily explainable because
various effects could contribute. The coupling likely involves only
two copper centers while the third metal center is either isolated or
gives dipolar interaction with the coupled dinuclear site.
Catalytic oxidation of L-/D-Dopa derivatives
The catalytic oxidations of L- and D-Dopa, and their methyl esters,
L-/D-DopaOMe, by the [Cu2(S)–L]4+ and [Cu3(S)–L]6+ complexes
were studied in the same conditions as for the corresponding com-
plexes in the (R)–L series.10 The unstable o-quinone products of the
catechol oxidation were trapped by formation of adducts with 3-
methyl-2-benzothiazolinone hydrazone (MBTH), that is the method
used in enzymatic studies with tyrosinase.24 Assuming that for the
present biomimetic catalytic reactions, a two-step mechanism of
catechol oxidation holds as in the case of our previous studies with
trinuclear7 and dinuclear14 Cu complexes, the following simplified
catalytic scheme can be hypothesized, where two molecules of cat-
echol (CatH2) per cycle are oxidized to quinone (Q):
CuII2 + CatH2  [CuII /CatH2] → CuI2 + Q + 2 H+
(1)
2
Fig. 8 1H–1H DQF-COSY spectrum of the [Cu2(S)–L]4+ complex (8 mM)
recorded at 293 K in CD3CN.
CuI2 + CatH2 + O2  [Cu2/O2/CatH2] → CuII2 + Q + 2 H2O (2)
A similar DQF-COSY spectrum recorded on the trinuclear
complex is much less resolved, with cross-peaks with intensities
comparable to the noise. This observation is consistent with the
coordination of Cu(II) at sites A and B as shown in structure II
of Scheme 4. The direct coordination of the amino group of the
1,1′-binaphthyl residue to the paramagnetic Cu(II) center strongly
increases the relaxation rates of the protons of this residue and
the attached propyl chain. As a consequence, the relaxation of the
protons in the NMR experiment occurs in the first instants of the
evolution time of the two-dimensional spectrum. Another possible
explanation of the lack of cross-peaks is that the proton relaxation
occurs during the delay after the gradient pulses.
Though, the kinetic experiments showed monophasic behavior and
it was impossible to separate the two steps. Thus, either the two
steps have a similar rate or the first one is slower. The dependence
of the rates of the catalytic reactions as a function of the substrate
concentration exhibited a hyperbolic behavior in all cases. This en-
abled the determination of the kinetic constants characterizing the
catalytic process, which are collected in Table 2. As for the Cu(II)
complexes in the (R)–L series, in the present case the dinuclear
complex is also more active and enantioselective than the trinuclear
complex. But while the chiral discrimination ability of [Cu2(R)–L]4+
toward the substrates was modest,10 that exhibited by [Cu2(S)–L]4+
toward L-/D-DopaOMe is remarkable, and obviously the faster
reacting enantiomer (D) is opposite to that preferred by the parent
(R)–L complex. Assuming the absolute value of the ratio R = [(kcat/
KM)L − (kcat/KM)D]/[(kcat/KM)L + (kcat/KM)D] as a reliable index of
enantioselectivity in catalytic reactions, the R index observed for
the L-/D-DopaOMe oxidation amounts to the notable value of 74%
for [Cu2(S)–L]4+, which is the highest reported so far for catalytic
oxidations of Dopa derivatives by biomimetic Cu complexes. For
instance, the R index recently reported for L-/D-Dopa oxidation by
another chiral trinuclear Cu(II) complex amounts to 61%.9 The en-
Solution magnetic susceptibility measurements
Also the magnetic susceptibility measurements performed in solu-
tion are consistent with the EPR results. As shown in Table 1, the
magnetic moments of [Cu2(S)–L]4+ suggest the presence of very
weak coupling between the Cu(II) centers, which is most probably
mediated by dipolar interaction. The possibility suggested in the
EPR analysis that this interaction originates from intermolecular
coupling is even more pertinent in this case, due to the high con-
2 1 9 8
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1

View Article Online
Table 2 Kinetic parameters for the catalytic oxidations of D- and L-Dopa, and D- and L-DopaOMe in methanol/aqueous phosphate buffer, pH 8.6, with MBTH
at 20 °C
Complex
Substrate
kcat /s−1
KM/mM
kcat/KM/M−1 s−1
[Cu2(S)–L]4+
D-Dopa
L-Dopa
D-DopaOMe
L-DopaOMe
D-Dopa
L-Dopa
D-DopaOMe
L-DopaOMe
(3.6 ± 0.1) × 10−3
(3.3 ± 0.1) × 10−3
(15.0 ± 0.3) × 10−3
(8.1 ± 0.2) × 10−3
(4.2 ± 0.2) × 10−3
(3.9 ± 0.1) × 10−3
(9.1 ± 0.3) × 10−3
(8.9 ± 0.3) × 10−3
(6.2 ± 0.8) × 10−2
(8.5 ± 0.9) × 10−2
(1.0 ± 0.1) × 10−2
(3.6 ± 0.1) × 10−2
(11.8 ± 1.6) × 10−2
(12.7 ± 1.3) × 10−2
(3.8 ± 0.5) × 10−2
(3.2 ± 0.6) × 10−2
58
39
1500
225
36
31
239
278
[Cu3(S)–L]6+
antio-differentiating ability of [Cu2(S)–L]4+ drops to R = 20% in the
case of the L-/D-Dopa couple and it is still lower for the reactions
catalyzed by [Cu3(S)–L]6+ (as well as for [Cu3(R)–L]6+).10
complexes derived from (S)–L are active catalysts for enantiose-
lective biomimetic oxidations. In this case, the dinuclear Cu(II)
complex exhibits the most interesting behavior, because it allows
stronger chiral recognition by the binaphthyl residue. The situation
here is just opposite to that recently reported by our group for a dif-
ferent dinuclear/trinuclear pair of chiral Cu(II) complexes,9 where
the chiral centers of the ligand are within the tridentate donor units
of the A sites rather than the bidentate residue of the B site. The
enantioselectivity observed in the catalytic reactions was therefore
“metal induced”, because the chiral recognition effect was due to
substrate coordination to the catalytically nonactive Cu(II) center.
In the systems described here the origin of enantioselectivity is
“ligand induced”.
Previous studies on the catalytic oxidations of catechol deriva-
tives demonstrated that the reaction needs the cooperation of two
close copper centers,14 to enable the binding of the catechol as a
bridging ligand and allow a fast two-electron transfer process. In the
case of [Cu2(S)–L]4+, the Dopa substrate can only form a productive
complex by binding the catechol residue to the two Cu(II) ions in
the A sites, and this forces the amino acid portion of the molecule
to approach the 1,1′-binaphthyl residue so that chiral recognition
is possible. For L- and D-Dopa a strong, but little discriminating
electrostatic interaction between the carboxylate group and the ter-
tiary amine group, which is peripheral to the chiral binaphthalene
residue, likely occurs. When the carboxylate group is esterified,
this electrostatic interaction is lost and hydrophobic interactions
between the ester group and the binaphthyl unit can take place,
which apparently increases the chiral discrimination. It is possible
to account for the different enantio-differentiating ability exhibited
by the two dinuclear complexes [Cu2(R)–L]4+ and [Cu2(S)–L]4+ in
terms of the different structure of the metal centers. In fact, in the
former complex the tertiary amine group of the binaphthyl residue
does not bind to Cu(II), as indicated by the EPR spectrum and also
by the difference in the absorption and CD spectra in the near-UV
and visible region, with respect to those of [Cu2(S)–L]4+. Thus, in
the case of [Cu2(S)–L]4+ a more strongly discriminating interaction
with the substrate can occur by the closer proximity of the chiral
binaphthyl chromophore to one of the catalytic Cu(II) centers.
Regarding the trinuclear complexes, two effects limit the at-
tainment of significant chiral discrimination in the catalytic reac-
tions. The first one is due to the arrangement of the two naphthyl
rings of the binaphthyl chromophore, that in [Cu3(S)–L]6+ (and
[Cu3(R)–L]6+)10 is considerably flattened. In addition, as shown
by the EPR data, one of the Cu(II) centers at A sites and the Cu(II)
ion at B sites are coupled by hydroxo bridges, and therefore here
the catecholic residue of Dopa substrates is forced to bind to this
pair of Cu(II) ions for productive catalysis. As a result, the cata-
lyst–substrate interaction involves little discriminating contribution
between the chiral amino acid portion of the Dopa derivatives and
the binaphthyl residue of the ligand.
Experimental
Materials and physical methods
All reagents and solvents were obtained from commercial sources
and used without further purification. Acetonitrile (spectral grade)
was distilled from potassium permanganate, sodium carbonate, it
was then stored over calcium hydride and distilled before use under
an inert atmosphere. Tetrahydrofuran was dried by refluxing and
distilling from metallic sodium. Optical rotations were obtained
from a Perkin–Elmer 241 polarimeter at 25 °C using a quartz cell of
10 cm path length. MS spectra were obtained from a VG 7070 EQ
spectrometer and 1H-, 13C-NMR spectra, and magnetic susceptibil-
ity measurements were recorded with a Bruker AVANCE 400 spec-
trometer operating at 9.37 T. Elemental analyses were performed
at the microanalytical laboratory of the Chemistry Department in
Milano. Optical spectra were obtained from HP 8452A and 8453
diode array spectrophotometers equipped with a thermostated cell
holder maintained at 20 ± 0.1 °C. Circular dichroism (CD) spectra
were recorded with a Jasco J-500 spectropolarimeter using quartz
cells of 0.01–2 cm path length.
Synthesis of the ligand (S)–L
The
ligand
(S)-(+)-N,N′-dimethyl-N,N′-bis{3-[bis-(1-
methyl-2-benzimidazolylmethyl)]-amino]-propyl}-
1,1′-binaphthyl-2,2′-diamine, (S)–L, was prepared from
(S)-(−)-1,1′-binaphthyl-2,2′-diamine following the synthetic path-
way described for the (R)-(+)-enantiomer;11 the synthetic details are
reported in the ESI.† Anal. calcd. for C64H66N12.CHCl3 (1122.69):
C 69.54, H 6.02, N 14.97; found: C 69.98, H 6.16, N 14.70%.
[]2D0 = + 101 (c = 1 × 10−3 M, CHCl3). MS (FAB): m/z (%): 1003
Conclusion
We have described here the dinuclear and trinuclear complexes of
a chiral ligand mimicking the donor environment present in the
trinuclear cluster of multicopper oxidases. As we have been unable
to obtain suitable crystals for a structural characterization of the
complexes, we performed an accurate spectroscopic investigation
that enabled the main structural features of the new complexes to
be deduced. The EPR spectral properties of [Cu3(S)–L]6+, in par-
ticular, bear some relevance to the unusual features exhibited by
the corresponding cluster present in laccase; these features have
never been reproduced in biomimetic active site models. Accord-
ing to spectroscopic studies on various derivatives of laccase, the
EPR-active copper in the cluster appears to be one of the type 3
Cu centers, while type 2 Cu is coupled with the remaining type 3
Cu, at least at low temperature.25 This situation would be basically
reproduced by structure II in Scheme 4. In addition, the chiral Cu(II)
1
(100) [M + 1]+. H NMR (400 MHz, CD3COCD3, 25 °C, SiMe4):
 = 7.79 (d, 2H; CH), 7,72 (d, 2H; CH), 7.58 (d, 4H; CH), 7.47 (d,
4H; CH), 7.2–7.3 (envelope, 10H; CH), 7.07 (m, 2H; CH), 6.93 (m,
2H; CH), 6.87 (d, 2H; CH), 3.73 (s, 4H; CH2-benzimidazole), 3.72
(4H; CH2-benzimidazole), 3.65 (s, 12H; CH3–N-benzimidazole),
2.29 (s, 6H; CH3–N-binaphthyl), 2.2–2.4 (m, 4H; CH2–N-binaph-
thyl), 1.5–1.8 (m, 4H; CH2–N-benzimidazole), 1.2–0.95 (m, 4H;
CH2–CH2–CH2–N). 13C NMR (75.5 MHz, CD3COCD3, 25 °C,
SiMe4):  = 151.9, 150.8, 143.1, 136.9, 134.4, 130.4, 129.0 128.7,
128.1, 126.0, 126.0, 124.0, 122.4, 121.7, 121.5, 119.6, 109.9, 55.2,
51.2, 51.0, 39.3, 29.0, 23.3. UV/Vis max/nm (CH3CN) 220 sh (/dm3
mol−1 cm−1 161000), 257 (66500), 270 sh (50800), 278 sh (40100),
286 (35500), 305 sh (15100) and 350 (5900). CD max/nm (CH3CN)
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1
2 1 9 9

View Article Online
1H NMR spectra
215 (/dm3 mol−1 cm−1 −282.3), 225 (+169.1), 246 (−7.3), 264
(+52.5), 296 (−38.2) and 367 (+3.2).
1
All the H-NMR spectra were recorded on 8 mM solutions of the
[Cu2(S)–L](ClO4)4 and [Cu3(S)–L](ClO4)6 complexes. The samples
were dissolved in CD3CN or in a mixture of CD3CN containing 4%
(v:v) of a 50 mM deuterated buffer solution at different pH. The
deuterated buffer solutions were prepared from repeated cycles
of rotary evaporation and dissolution in D2O of the buffers at the
desired pH (phosphate buffer for pH 5 and 7 and tetraborate buffer
for pH 10) in Milli Q water. The isotopic effect on the pH has been
neglected. The experiments at variable temperature were recorded
changing the temperature from +40 to −35 °C; lower temperatures
cause freezing of the solutions containing the buffer. Typically 160
Copper(II) complexes
The dinuclear [Cu2(S)–L](ClO4)4·4H2O and the trinuclear [Cu3(S)–
L](ClO4)6·H2O complexes were obtained as described previously
for the corresponding (R)–L complexes.10
[Cu2(S)–L](ClO4)4·4H2O. Anal. calcd. for C64H66N12Cl4-
Cu2O16·4H2O (1600.26): C 48.04, H 4.66, N 10.50, Cu 7.94; found:
C 48.25, H 4.66, N 10.64, Cu 8.00%. UV/Vis max/nm (CH3CN)
220 sh (/dm3 mol−1 cm−1 130000), 255 (60000), 265 (57000),
272 (56000), 280 (50000), 300 sh (14000), 345 sh (5400) and 698
(200). CD max/nm (CH3CN) 215 (/dm3 mol−1 cm−1 −179.8), 225
(+142.9), 256 (+33.0), 294 (−16.7), 324 sh (+1.0), 352 (+2.3), 440
(+0.01) and 619 (+0.07).
scans were collected for each sample. The H–1H DQF-COSY
experiments on the [Cu2(S)–L](ClO4)4 and [Cu3(S)–L](ClO4)6 com-
plexes were performed at 20 °C.
1
Solution magnetic susceptibility measurements
[Cu3(S)–L](ClO4)6·H2O. Anal. calcd. for C64H66N12Cl6-
Cu3O24·H2O (1808.67): C 42.50, H 3.79, N 9.30, Cu 10.54; found:
C 43.22, H 3.91, N 9.53, Cu 10.69%. UV/Vis, max/nm (CH3CN)
220 sh (/dm3 mol−1 cm−1 120000), 250 (56000), 262 sh (40000),
272 (39000), 280 (38000), 302 sh (9400), 356 (7000), 560 sh
(300) and 648 sh (250). CD max/nm (CH3CN) 209 (/dm3 mol−1
cm−1 −77.5), 223 (+11.2), 236 (−8.8), 257 (+60.4), 293 (−10.6), 345
(−3.9), 395 (−0.9), 490 (+0.06) and 610 (−0.03).
The magnetic susceptibility of the [Cu2(S)–L](ClO4)4 and
[Cu3(S)–L](ClO4)6 complexes was measured by the modified Evans
method.27,28 A coaxial NMR tube was used with tert-butyl alcohol
or 1,4-dioxane (50 mM) as internal reference. The solution of the
paramagnetic species in CD3CN (~8 mM) was introduced into the
inner narrow-bore tube while the solvent containing the reference
compound was placed into the outer tube. The paramagnetic solu-
tion also contained the reference compound in an identical amount.
The methyl proton signals of tert-butyl alcohol from the inner and
outer solutions were recorded and the separation of the two signals
() was monitored and considered as the paramagnetic shift. Mass
susceptibility (Mpara) is correlated to the above measured  as fol-
lows:29
Caution!
Perchlorate complexes with organic ligands are potentially explo-
sive and should be handled with great care. Only small amounts of
material should be prepared. We did not have problems working
with small amounts of the perchlorate complexes described in this
paper.
 = 1000 M Mpara/3
(3)
where M is the concentration of the complexes in mol l−1 and 
Ligand binding experiments
para
is the shift expressed in ppm; M is obtained in m3 mol−1 units.
Spectrophotometric titrations of azide binding to [Cu2(S)–L]4+ and
[Cu3(S)–L]6+ were carried out by adding concentrate methanolic
solutions of the ligand to solutions of the complexes dissolved in
acetonitrile. Titration of [Cu2(S)–L]4+ (2.02 × 10−4 M) was per-
formed by addition of successive and equal amounts of an azide
solution (2.42 × 10−3 M) from 0.1 to 2.0 [azide]:[Cu2] ratios, and
from 2.2 to 4.0 [azide]:[Cu2] ratios. It was possible to separate the
In addition, the diamagnetic contribution to Mpara due to the ligand
was corrected by measuring the magnetic susceptibility of a CD3CN
solution of the dinuclear zinc complex (~8 mM),11 in the same con-
ditions and with the same procedure described above. The magnetic
moment of the complexes (eff) was then obtained from Mpara data
from the equation:29
−
titration steps for [N3 ]:[Cu2] between 0.1 and 1.0 (max = 402 nm,
eff2 = Mpara 3kT/NAeff
(4)
isosbestic point at 323 nm) and between 1.1 and 2.0 (isosbestic
point at 333 nm). In a similar way the binding process could be
separated into two additional steps for [N3 ]:[Cu2] between 2.2
The exact concentration of the metals in the diamagnetic and the
paramagnetic samples used for the calculations was checked imme-
diately after the NMR experiments by atomic absorption. The reli-
ability of the method was tested using CuSO4·5H2O as paramagnetic
standard. The eff value found for this compound agreed perfectly
with the values available in the literature.27,28
−
and 3.0 (max = 402 nm, isosbestic point at 360 nm) and between
3.2 and 4.0 (isosbestic point at 362 nm). Titration of [Cu3(S)–L]6+
(1.98 × 10−4 M) was carried out by addition of successive and equal
amounts of azide (2.42 × 10−3 M) from 0.1 to 2.0 [azide]:[Cu2]
ratios, and from 2.5 to 10.0 [azide]:[Cu2] ratios. It was not pos-
sible to separate the binding process into steps. The spectral data
in the LMCT region were analyzed as described previously,26 to
deduce equilibrium constants and stoichiometry of formation of
the adducts.
Catalytic oxidation of L,D-Dopa and L,D-Dopa methyl esters
(L,D-DopaOMe)
The kinetic experiments were performed in a magnetically stirred
and thermostated 1-cm path length cell at 20 ± 0.1 °C. A mixture
of methanol/aqueous phosphate buffer (50 mM, pH 8.6) 1:15 (v/v)
was used as solvent. The catalyst concentration was kept at 5 M by
adding 10 l of a 10−3 M solution of the copper complex dissolved in
CH3CN/MeOH 1:15 (v:v) in the reaction cuvette, while the substrate
concentration was varied from 6 M to 0.7 mM (final volume 2 ml).
To prevent further reactions of the Dopa-o-quinones and DopaOMe-
o-quinones initially formed, an excess amount of 3-methyl-2-benzo-
thiazoline hydrazone (MBTH, 1 mM final concentration) was added.
The formation of the stable Dopa-o-quinone-MBTH and DopaOMe-
o-quinone-MBTH adducts was followed through the development
of the intense absorption band at max = 500 nm (500 = 13400 M−1
cm−1 for the adduct with Dopa-o-quinone, and 500 = 11600 M−1 cm−1
for the adduct with DopaOMe-o-quinone). The noise during the
measurements was reduced by reading the absorbance difference
EPR measurements
EPR spectra of the [Cu2(S)–L](ClO4)4 and [Cu3(S)–L](ClO4)6 com-
plexes were recorded in degassed solution by using a Bruker X-
band spectrometer ESP300 E, equipped with an NMR gaussmeter
(Bruker ER035), a frequency counter (Bruker ER 041 XK) and a
variable temperature control continuous flow N2 cryostat (Bruker
B-VT 2000). The software WINEPR (v. 2.11) and SimFonia (v.
1.25) were provided by Bruker. A 1 mM CuEDTA solution was
used as standard for spin quantitation, while DPPH free radical
(g = 2.0036) was used as reference for g value corrections. The
azide adducts were prepared by incubation of the copper complex
solutions with the calculated amount of azide overnight at 4 °C in a
closed reaction vessel under anaerobic conditions.
2 2 0 0
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1

View Article Online
11 M. C. Mimmi, M. Gullotti, L. Santagostini, R. Pagliarin, L. De Gioia,
E. Monzani and L. Casella, Eur. J. Inorg. Chem., 2003, 21, 3934–
3944.
12 (a) L. Di Bari, G. Pescitelli and P. Salvadori, J. Am. Chem. Soc., 1999,
121, 7998–8004; (b) C. Rosini, L. Franzini, P. Salvatori and G. P. Spada,
J. Org. Chem., 1992, 57, 6820–6824.
between 500 and 800 nm; in this way the effect of the baseline oscil-
lation that occurs on stirring of the solution is cancelled (note that at
800 nm none of the reagents or products absorb appreciably). The
initial rates of the oxidations were obtained by fitting the absorbance
vs. time curves in the first few seconds of the reactions.
13 L. Casella, O. Carugo, M. Gullotti, S. Garofani and P. Zanello, Inorg.
Chem., 1993, 32, 2056–2067.
14 E. Monzani,G. Battaini,A. Perotti,L. Casella,M. Gullotti,L. Santagostini,
G. Nardin, L. Randaccio, S. Geremia, P. Zanello and G. Opromolla,
Inorg. Chem., 1999, 38, 5359–5369.
Acknowledgements
This work was supported by the Italian MIUR, through a PRIN
project, and the Italian CNR.
15 G. Battaini, E. Monzani, A. Perotti, C. Para, L. Casella, L. Santagostini,
M. Gullotti, R. Dillinger, C. Näther and F. Tuczek, J. Am. Chem. Soc.,
2003, 125, 4185–4198.
References
1 See for example: (a) K. D. Karlin and A. D. Zuberbühler, in Bioinor-
ganic Catalysis, ed. J. Reedijk and E. Bouwman, Dekker, New York,
2nd edn, revised and expanded, 1999, pp. 469–534; (b) H.-C. Liang,
M. Dahan and K. D. Karlin, Curr. Opin. Chem. Biol., 1999, 3, 168–175.
2 K. A. Magnus, H. Ton-That and J. E. Carpenter, Chem. Rev., 1994, 94,
727–735.
3 A. Sánchez-Ferrer, J. N. Rodriguez-Lopez, F. Garcia-Cánovas and
F. Garcia-Carmona, Biochim. Biophys. Acta, 1995, 1247, 1–11.
4 T. Bertrand, C. Jolivalt, P. Briozzo, E. Caminade, N. Joly, C. Madzak
and C. Mougin, Biochemistry, 2002, 41, 7325–7333.
16 L. Casella, O. Carugo, M. Gullotti, S. Doldi and M. Frassoni, Inorg.
Chem., 1996, 35, 1101–1113.
17 G. Battaini, L. Casella, M. Gullotti, E. Monzani, G. Nardin, A. Perotti,
L. Randaccio, L. Santagostini, F. W. Heinemann and S. Schindler, Eur.
J. Inorg. Chem., 2003, 21, 1197–1205.
18 S.-K. Lee, S. DeBeer George, W. E. Antholine, B. Hedman, K. O.
Hodgson and E. I. Solomon, J. Am. Chem. Soc., 2002, 124, 6180–6193.
19 J. Cabrero, C. de Graaf, E. Bordas, R. Caballol and J.-P. Malrieu, Chem.
Eur. J., 2003, 9, 2307–2315.
20 O. Kahn, Molecular-Magnetism, Wiley–VCH, New York, 1993.
21 M. A. Aebersold, B. Gillon, O. Plantevin, L. Pardi, O. Kahn, P. Bergerat,
I. von Seggern, F. Tuczek, L. Öhrström, A. Grand and E. Lelièvre-
Berna, J. Am. Chem. Soc., 1998, 120, 5238–5245.
5 A. Messerschmidt,R. Ladenstein,R. Huber,M. Bolognesi,L. Avigliano,
R. Petruzzelli, A. Rossi and A. Finazzi-Agrò, J. Mol. Biol., 1992, 224,
179–205.
6 (a) V. N. Zaitsev, I. Zaitseva, M. Papiz and P. F. Lindley, J. Biol. Inorg.
Chem., 1999, 4, 579–587; (b) N. Hakulinen, L.-L. Kiiskinen, K. Kruus,
M. Saloheimo,A. Paananen,A. Koivula and J. Rouvinen, Nature Struct.
Biol., 2002, 9, 601–605.
22 E. Ruiz, J. Cano, S. Alvarez and P. Alemany, J. Am. Chem. Soc., 1998,
120, 11122–11129.
23 E. I. Solomon, U. M. Sundaram and T. E. Machonkin, Chem. Rev.,
1996, 96, 2563–2605.
7 E. Monzani, L. Casella, G. Zoppellaro, M. Gullotti, R. Pagliarin,
R. P. Bonomo, G. Tabbì, G. Nardin and L. Randaccio, Inorg. Chim.
Acta, 1998, 282, 180–192.
8 L. Santagostini, M. Gullotti, R. Pagliarin, E. Bianchi, L. Casella and
E. Monzani, Tetrahedron: Asymmetry, 1999, 10, 281–295.
9 L. Santagostini, M. Gullotti, R. Pagliarin, E. Monzani and L. Casella,
Chem. Commun., 2003, 2186–2187.
24 R. P. Ferrari, E. Laurenti, E. M. Ghibaudi and L. Casella, J. Inorg. Bio-
chem., 1997, 68, 61–69.
25 T. L. Fraterrigo, C. Miller, B. Reinhammar and D. R. McMillin, J. Biol.
Inorg. Chem., 1999, 4, 183–187.
26 L. Casella, M. Gullotti, G. Pallanza and M. Buga, Inorg. Chem., 1991,
30, 221–227.
27 D. F. Evans, J. Chem. Soc., 1959, 2003–2005.
10 M. C. Mimmi, M. Gullotti, L. Santagostini, A. Saladino, L. Casella,
E. Monzani and R. Pagliarin, J. Mol. Cat. A: Chem., 2003, 204/205,
381–389.
28 S. K. Sur, J. Magn. Reson., 1989, 82, 169–173.
29 I. Bertini and C. Luchinat, NMR of Paramagnetic Substances, Elsevier,
Amsterdam, 1996.
D a l t o n T r a n s . , 2 0 0 4 , 2 1 9 2 – 2 2 0 1
2 2 0 1