Ligand self-recognition in the stereoselective assembly of [2 + 2] metallomacrocycles from racemic chiral bisbipyridyl molecular clefts and zinc(II)

Article abstract of DOI:10.1039/b404054f

The synthesis of racemic and optically pure ligand L, in which two 6,6′-disubstituted bipyridines are connected by methyleneoxy linkers to the molecular cleft dibenzobicyclo[b,f][3.3.1.]nona-5a,6a-diene-6,12-dione, is reported. In the presence of 2 equiva

Full text of DOI:10.1039/b404054f

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Ligand self-recognition in the stereoselective assembly of [2 + 2]
metallomacrocycles from racemic chiral bisbipyridyl molecular
clefts and zinc(II)
Pia I. Anderberg, Jennifer J. Turner, Krystal J. Evans, Lauren M. Hutchins and
Margaret M. Harding*
School of Chemistry, University of Sydney, F11, N.S.W., 2006 Australia.
E-mail: harding@chem.usyd.edu.au; Fax: +61 2 9351 6650; Tel: +61 2 9351 2745
Received 17th March 2004, Accepted 21st April 2004
First published as an Advance Article on the web 7th May 2004
The synthesis of racemic and optically pure ligand L, in which two 6,6′-disubstituted bipyridines are connected by
methyleneoxy linkers to the molecular cleft dibenzobicyclo[b,f][3.3.1]nona-5a,6a-diene-6,12-dione, is reported. In the
presence of 2 equivalents of zinc(II) trifluoromethansulfonate (±)-L undergoes slow reversible coordination over 24 h
to form a pair of enantiomeric [2 + 2] metallomacrocycles, [Zn2(+)L2](OTf)4 and [Zn2(−)L2](OTf)4 respectively, that
contain either two (+)-L ligands or two (−)-L ligands. This assignment was confirmed by independent studies with either
(+)-L or (−)-L which formed the same complexes but at a significantly faster rate (3 h), and circular dichroism spectra
of [Zn2(+)L2](OTf)4 and [Zn2(−)L2](OTf)4 which gave signals of the same intensity with the opposite sign. Treatment
of (±)-L or optically pure L with copper(I) showed rapid formation of a mixture of oligomers as well as the [2 + 2]
metallomacrocycle. The complex Zn2L2(OTf)4 exhibits slow exchange between two species on the NMR time scale at
room temperature. The results are consistent with the formation of a library of metal complexes in which the zinc(II) binds
initially to the most accessible bipyridyl binding sites in (±)-L. Equilibration over several hours results in self-recognition
of enantiomeric ligands to form a pair of enantiomeric metallomacrocycles, which have been tentatively assigned as
having the helical configuration. Slow exchange is attributed to the preference for both metal centres to adopt 6-coordinate
geometries involving the linker oxygens, but are limited to exchanging 5-coordinate complexes due to the shape of the
cleft and the short linker.
base but contains different dimensions and additional carbonyl (or
Introduction
alcohol) groups that may be utilised in molecular recognition stud-
Numerous studies of the assembly of helicates, metallomacrocycles,
ies.22 This cleft was incorporated into the ligand design as it allows
catenanes and related supramolecular architectures that incorporate
entry into optically active ligands, and as the dimensions and rec-
pyridyl and bipyridyl groups have identified important features in
ognition properties of the resultant metallomacrocycles offer new
ligand design that allow the controlled assembly of both chiral and
opportunities to target the molecular recognition of different classes
achiral metal complexes.1–8 Most studies have used achiral ligands
of substrates. In the presence of zinc(II), racemic ligands undergo
that result in the formation of mixtures of complexes. However, the
an unusual slow equilibration to form the thermodynamic [2 + 2]
use of optically pure ligands has allowed the stereoselective assem-
metallomacrocycles which contain a single ligand enantiomer.
bly of metallomacrocycles and helicates in a predictable manner.8–15
In contrast, the controlled assembly of a single enantiomeric pair of
metal complexes from racemic ligands and two chiral metal centres
is challenging due to the number of possible stereoisomers that may
be formed. An elegant example demonstrating ligand self-recogni-
tion has been reported in which reaction of a racemic bispyridyl
ligand generated only homochiral complexes containing a single
enantiomeric ligand.16
Results
Synthesis
Ligand L was prepared under standard Williamson ether formation
conditions from the bisphenol cleft 4 and 6-methyl-6′-bromo-
methyl-2,2′-bipyridine (Scheme 1). As in our previous studies,17–20
the methylene group was included in the linker as a H NMR re-
porter group that could be used to monitor metal complexation, and
an aryl ether linker was selected due to synthetic ease.
1
Our contributions in this area have been in the assembly of
chiral and achiral [2 + 2] metallomacrocycles encoded with rec-
ognition features to allow binding of aromatic substrates.17–20 The
design features required to allow exclusive assembly of the chiral
[2 + 2] metallomacrocycle were identified by a systematic study of
bisbipyridyl ligands containing an aromatic spacer separated by a
short linker. The low stereochemical preference of zinc(II) allowed
a range of coordination geometries, and hence metallomacrocycles
of different shapes and dimensions, to be assembled. In addition,
zinc(II) coordination provided a mechanism whereby the dimen-
sions of the cavity were altered for optimal binding of aromatic
substrates.19 The rigidity and length of the linker allowed exclusive
assembly of the helical metallomacrocycle with a naphthalenedi-
imide spacer.17,19 In contrast, exclusive assembly of the non-helical,
achiral metallomacrocycle occurred with a longer, more flexible
linker and a pyrene aromatic spacer.18
Two routes to prepare the bisphenol 4 were investigated. In the
first route, the cleft was prepared using a similar procedure to that
reported for the parent cleft and derivatives from the methyl ether
3.21 Thus, p-methoxybenzylnitrile was converted to the mixture of
stereoisomeric bisnitriles 1 which were not purified but hydrolyzed
directly to the mixture of bisacids 2. In agreement with the literature
reports on a related compound,23 attempted intramolecular Friedel-
Crafts acylation of 2 in concentrated sulfuric acid, the standard
conditions that gave good yields of related clefts including the
bromocleft 5,21,24 was unsuccessful. The methoxy group activates
the aromatic rings to sulfonation, and a mixture of water soluble
products was obtained. Hence polyphosphoric acid was used to ef-
fect cyclization of the bisacids 2 to the molecular cleft 3. While the
bismethyl ether 3 was formed, the reaction was not reproducible
and highly variable yields of product were obtained (0–35%). The
best isolated yield (35%) was obtained only on small scale (<0.5 g)
reactions. Demethylation of 3 by slow addition of the BBr3 at a low
In this paper we report the stereoselective assembly of metal-
lomacrocycles from the optically active bisbipyridyl ligand L that
incorporates the chiral molecular cleft, dibenzobicyclo[b,f][3.3.1]n
ona-5a,6a-diene-6,12-dione.21 This cleft is reminiscent of Tröger’s
1 7 0 8
D a l t o n T r a n s . , 2 0 0 4 , 1 7 0 8
–
1 7 1 4
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 Reagents and conditions: (i) KOH, ; (ii) polyphos-
phoric acid (PPA), ; (iii) BBr3 −5 °C–rt; (iv) bis(pinacolato)diboron,
Pd(dppf)Cl2·CH2Cl2, KOAc followed by H2O2; (v) KOtBu, 6-bromomethyl-
6′-methyl-2,2′-bipyridine.
temperature (−5° to 0° C) gave the desired bisphenol 4. Treatment
of the bisphenol 4 with base followed by 6-bromomethyl-6′-methyl-
2,2-bipyridine25,26 afforded ligand L in 60% yield. Use of the bulky,
non-nucleophilic base, potassium tert-butoxide, allowed the reac-
tion to be performed without protection of the carbonyl groups on
the cleft.
1
Fig. 1 400 MHz H NMR spectra at 300 K of (a) (±)-L in CDCl3 and
(±)-L and Zn(OTf)2 (2 equivalents) in CD3CN after (b) 15 min, (c) 1 h,
(d) 6 h, (e) 24 h.
We have previously reported the successful resolution of the
bisbromocleft 5 and the corresponding bismethyl cleft using chiral
HPLC.24,27 However, attempts to resolve the enantiomers of either
the bimethoxycleft 3, the bisphenol 4, or ligand L by HPLC using
different chiral supports and a range of solvent systems were unsuc-
cessful. The alternate route to optically pure L involved the use of
the readily available optically pure clefts, (+)-5 and (−)-5.27 These
clefts were converted to the corresponding bisphenols, (+)-4 and
(−)-4 via the corresponding pinacol boronate esters under standard
Miyaura conditions,28 followed by conversion to (+)-L [(R)(R)–L]
and (−)-L [(S)(S)–L].
Metal complexation studies with zinc(II)
Racemic ligand. Addition of 2 equivalents of zinc trifluoro-
methanesulfonate (triflate, OTf) to a CD3CN solution of (±)-L
gave a pale yellow solution. Electrospray mass spectrometry was
consistent with the formation of a [2 + 2] complex Zn2L2(OTf)4
6 with peaks at m/z 1868 [Zn2L2(OTf)3]+, 1502.9 [ZnL2OTf]+ and
859.1 [Zn2L2(OTf)2]2+. The NMR spectrum of this solution (Fig.
1b) econtained multiple, broad signals consistent with the forma-
tion of several complexes and/or oligomers. However, over several
hours, the appearance of the spectrum changed significantly and
after 24 h a spectrum containing a single set of sharp signals was
obtained (Fig. 1e).Anotable feature of the spectrum was the change
in appearance of the methylene protons in the ligand from a singlet
( 5.23 ppm, Fig. 1a) to an AB quartet ( 5.35 ppm, Ha,b Fig. 1e).
2D COSY and NOESY spectroscopy showed that this AB quartet
was in exchange with a second AX system (Hc,d) in which one of
the two methylene protons was dramatically shifted upfield ( 2.34
ppm, Hc, Fig. 2).
Variable temperature measurements and NOESY spectra of
Zn2L2(OTf)4 6 were used to identify the dynamic process(es) oc-
curring on the NMR time scale. While NOESY spectra at 300 K
(Fig. 2, Fig. 3a) confirmed the presence of two species in exchange,
these spectra were complicated by transferred NOE peaks arising
from exchange. Hence tr-ROESY experiments, recorded at 250 K
where there was no exchange on the NMR time scale (Fig. 3b), were
used to identify the NOEs, and allowed full assignment of all signals
Fig. 2 400MHzNOESYspectrumofZn2L2(OTf)4 inCD3CNat300Kshow-
ing methylene protons. Exchange peaks in black, NOEs in grey; * = water.
in the spectrum. As shown in Fig. 3, two sets of bipyridyl reso-
nances were observed (labelled H3,4,5/3′4′5′ and H6,7,8/6′,7′,8′),
in exchange with one another. Similarly, unique signals for the six
cleft aromatic protons (He–j) were detected with He,f,g in exchange
with Hh,i,j.
Opticallypureligands.Identicalmetalcomplexationstudieswere
performed by addition of 2 equivalents of zinc(II) triflate to a CD3CN
solution containing either (+)-L or (−)-L to give [Zn2(+)L2](OTf)4
and [Zn2(−)L2](OTf)4 respectively. In each experiment, the same
1H spectrum as that shown in Fig. 1e was obtained. However, the
rate of sharpening of the signals was significantly faster than that
observed with (±)-L. Thus, while equilibration of (±)-L required
approximately 24 h until only signals arising from a single com-
plex 6 were observed, under the same conditions, (+)-L or (−)-L
required only 3 h to give 6. The CD spectra of [Zn2(+)L2](OTf)4 and
[Zn2(−)L2](OTf)4 (Fig. 4) confirm that the complexes are enantio-
mers, with each complex giving rise to signals of the same intensity
at the same concentration, but opposite in sign.
D a l t o n T r a n s . , 2 0 0 4 , 1 7 0 8 – 1 7 1 4
1 7 0 9

View Article Online
Fig. 3 400 MHz spectra of Zn2L2(OTf)4 in CD3CN showing peak assignments in (a) NOESY at 300 K, (b) tr-ROESY at 250 K. Exchange peaks in black are
highlighted for bipyridyl protons H5′/H8′ and cleft protons He/j. NOEs are shown in grey.
1
Fig. 5 400 MHz H NMR spectra of (±)-L and Cu(OTf)2 (a) in CD3CN
after 15 min at 300 K, (b) in CD3CN at 235 K, (c) in CD2Cl2 at 300 K.
Fig. 4 Circular dichroism spectra (CD3CN) of [Zn2(−)L2](OTf)4 (3.2 M,
solid line) and [Zn2(+)L2](OTf)4 (3.7 M, dotted line).
in CD3CN were recorded (Fig. 5b), and the solvent was changed to
CD2Cl2 (Fig. 5c). While some slight sharpening of the signals was
observed at lower temperatures, there were no significant changes
compared with the NMR spectrum obtained at room temperature.
An identical result was obtained with either (+)-L or (−)-L.
Stability of complex 6. The stability of 6 in different solvents
hampered efforts to obtain crystals suitable for X-ray diffraction.
Diffusion of either ether or methanol into a dichloromethane or
acetonitrile solution of 6 gave an amorphous solid with an NMR
spectrum similar to Fig 1a i.e., reformation of a mixture of metal
complexes followed by equilibration over several hours to form
metallomacrocycle 6. Once formed the complex 6 is quite stable.
Provided the solvent is unchanged, heating an CD3CN solution to
65 °C showed the expected signal broadening but signal coales-
cence would require the use of a higher boiling solvent. The appear-
ance of the NMR spectrum of 6 (Fig. 1e) was independent of sample
concentration, and removal of CD3CN followed by redissolving the
sample gave an identical spectrum to Fig. 1e.
Discussion
Unequivocal assignment of the structure of the [2 + 2] metallomac-
rocycle 6 formed from racemic and optically pure ligands L is chal-
lenging due to the need to consider both the chirality of the ligands
and the metal centres, and a number of unusual features exhibited
by the complex. First, slow equilibration of a library of oligomeric
complexes to form a single stable complex was observed with either
(+)-L, (−)-L or (±)-L (Fig. 1). The fact that identical NMR spectra
were obtained with either (+)-L or (−)-L or (±)-L, provides clear
evidence for the self-recognition of the enantiomeric ligands in the
assembly of the [2 + 2] metallomacrocycle from racemic (±)-L. No
evidence for the formation of a diastereomeric complex containing
one (+)-L and one (−)-L ligand was detected. However, since the
rate of formation of 6 was significantly slower with (±)-L than with
the pure enantiomers, initial coordination of zinc(II) to both (+)-L
and (−)-L must occur at the kinetically most accessible binding sites
followed by equilibration to the thermodynamic product 6.
Metal complexation studies with copper(I)
Fig. 5 shows the 1H NMR spectrum obtained on addition of
copper(I) triflate to a solution of (±)-L in CD3CN. In contrast to the
corresponding spectra obtained with zinc(II), the NMR spectrum
did not change in appearance with time, and the methylene protons
showed only minor upfield shifts compared to the ligand. Analysis
of the sample by electrospray mass spectrometry showed peaks
consistent with the formation of the [2 + 2] metallomacrocycle
Cu2L2(OTf)4, along with other unidentified complexes.
Second, the changes in chemical shift of the methylene protons
are significantly different to those observed in related copper(I) and
zinc(II) [2 + 2] metallomacrocycles in which the metals adopt tet-
In order to rule out signal broadening due to exchange between the
labile copper(I) centres and acetonitrile, variable temperature spectra
1 7 1 0
D a l t o n T r a n s . , 2 0 0 4 , 1 7 0 8 – 1 7 1 4

View Article Online
rahedral or distorted tetrahedral geometries.17,19,20 The large upfield
shift of the methylene proton Hc  = 3 ppm (Fig. 2) is reminiscent
of changes in a related achiral [2 + 2] zinc(II) metallomacrocycle in
which octahedral coordination of two ether oxygens to zinc(II) re-
sulted in significant shielding of these protons.18 The chemical shifts
of Hc,d (Fig. 2) strongly suggests coordination of the oxygen at-
tached to the methylene protons (Hc,d) to one of the zinc(II) binding
sites in the complex 6. The presence of 5- or 6-coordinate zinc(II) in
6 in which the vacant sites are occupied by oxygen from the linker
and/or solvent is consistent with the fact that the complex re-equili-
brates to mixtures in different solvents and is supported by studies
with copper(I) (Fig. 5) where multiple complexes/oligomers were
formed; the more rigorous requirement for tetrahedral complexation
with 6,6-disubstituted bipyridyls does not allow formation of 6.
Third, while the NMR spectra showed formation of a single com-
plex 6, slow exchange between two species occurred on the NMR
time scale (Figs. 2, 3). Prior to assigning the origin of this exchange,
the stereochemistry at the metal centres is addressed. There are 16
possible stereoisomers (8 pairs of enantiomers) that may be formed
in a [2 + 2] metallomacrocycle as a result of the new chirality from
the two metal centres and the two chiral ligands in (±)-L. The fact
that a single set of resonances was observed in Fig. 1e shows that
ligand–ligand recognition has occurred in the assembly process.
The independent experiments with (+)-L (and (−)-L) have al-
lowed us to determine that the ligand enantiomers in (±)-L undergo
self-recognition to give rise to a pair of enantiomeric complexes,
[Zn2(+)L2](OTf)4 and [Zn2(−)L2](OTf)4 respectively, that give rise
to identical NMR spectra.
proposed exchange is also consistent with initial formation of oligo-
mers and polymers in solution; initial zinc(II) coordination to the
most accessible bipyridyl sites is followed by slower reorganisation
to form the metallomacrocycle. Steric effects prevent zinc(II) from
simultaneous formation of 6-coordinate geometries by binding to
both linker oxygens adjacent to the bipyridyls at both metal centres.
Slow exchange between one metal centre in an octahedral geometry
from coordination of both bipyridyls and the two linker oxygens,
and the second metal in a tetrahedral geometry was also considered.
However, model building suggested that the aromatic rings of the
cleft would prevent simultaneous binding of two oxygens to zinc(II),
and NOE data was not consistent with this structure.
While the assignment of the helical configuration is supported
by model building and a weak NOE between Hi and He (data not
shown), formation of the non-helical complex, either 6c or 6d was
also considered. This complex lacks a symmetry axis and CPK
models suggested that this complex is more strained than the cor-
Fig. 6 shows a schematic of the four diastereomeric helical (6a,
6b) and non-helical complexes (6c, 6d) that may be formed from two
(+)-L ligands. The presence of a single set of resonances when (+)-
L was treated with zinc(II) indicates that a single diastereomer was
formed. We have assigned the complexes Zn2L2(OTf)4 tentatively
as having the helical configuration (illustrated in Fig. 6 for 6a). This
structure is consistent with the asymmetry of the ligands in the com-
plex, and the exchange on the NMR time scale, which we propose is
due to slow exchange in geometry at each metal centre (Fig. 7). The
distinct chemical shifts of Ha,b and Hc,d suggest that each metal
centre is 5-coordinate (or 6-coordinate including solvent), which
results in shielding of Hc by the aromatic ring of the cleft. This
Fig. 6 Schematic of four possible stereoisomers that can be formed from
two (+)-L ligands and two metal ions in a tetrahedral or pseudo-tetrahedral
geometry.
Fig. 7 Proposed structures of enantiomeric helical complexes formed from either two (−)-L or two (+)-L ligands, showing slow exchange of ether oxygen
coordination to zinc(II) on the NMR time scale and peak assignments as shown in Fig. 3. Coordination geometry at each metal centre unknown but assumed
to be 5-coordinate involving one linker oxygen or 6-coordinate including solvent.
D a l t o n T r a n s . , 2 0 0 4 , 1 7 0 8 – 1 7 1 4
1 7 1 1

View Article Online
responding helical complexes 6a or 6b. In addition, this structure
is not consistent with the NOE detected between Hi and He. The
CD spectra of the complexes formed from either (+)-L or (−)-L are
equal in intensity and opposite in sign (Fig. 4), and thus provide
direct experimental evidence for the formation of enantiomeric
complexes consistent with the fact that identical NMR spectra were
obtained with (+)-L and (−)-L and zinc(II). However, due to the chi-
ral ligands in each complex, both the helical and non-helical com-
plexes are chiral and therefore would give rise to signals in the CD
spectrum and detailed analysis of the transitions that give rise to the
CD signals are required. Compared to the CD spectra of (+)-L and
(−)-L (data not shown), the CD spectra of the complexes gave an
increase in absorption at 258 nm which supports the assignment of
the complex as the chiral helicate rather than the achiral complex.
In principle, CD can be used to assign the absolute configuration
of M- and P-helicates.13 However, direct correlation of the sign of
the CD spectra and Cotton effects with absolute configuration is not
straightforward in complexes that are chiral as a result of optically
active ligands plus the chirality introduced by the metal centres.13
In studies where CD spectra of helicates have been used to deduce
chirality, these assignments have been supported independently by
X-ray crystallography, or the absolute configuration has been as-
sumed on comparison of the spectra to related published structures.
In the absence of further data, assignment of the M or P configura-
tion to the complex has not been attempted.
X-ray diffraction is required to confirm the full structure of the
complexes. However, obtaining crystals of pure 6 is almost impos-
sible due to the labile coordination geometry at zinc(II) in different
solvents. Similar difficulties have been noted in a recent study by
Prabaharan and Fletcher in which labile, 5-coordinate zinc(II) com-
plexes involving solvent prevent crystallisation for unambiguous
determination of stereochemistry.10 Equilibration over several hours
to the thermodynamically stable chiral metallomacrocycle (Fig. 1b)
is consistent with the proposed structure of complex 6 in which
both metal centres prefer to adopt 6 coordination geometries but are
limited to exchanging 5-coordinate complexes due to the shape of
the cleft and the short linker.
Experimental
General
Melting points were determined using a Reichert heating stage and
are uncorrected. 1D 1H NMR spectra were recorded using a Bruker
AC 200B, Avance DPX 300 or AMX 400 spectrometer and are
reported as parts per million (ppm) downfield shift from residual
solvent peak. Peak assignments in the experimental are labeled ac-
cording to the numbering scheme shown in Figs. 2 and 3. IR spectra
were recorded on Perkin-Elmer 1600 fourier transform spectropho-
tometer. UV VIS spectra were recorded on a Carey 5E UV-Vis-NIR
spectrophotometer at 20 °C. Optical rotations were obtained using
a POLAAR 2001 Polarimeter at 589.3 nm in the indicated solvent
and concentration (g/100 cm3) at 22 °C. LR electron impact (EI)
mass spectra were recorded on a Finnigan PolarisQ ion trap mass
spectrometer using electron impact ionisation mode at 40 or 70 eV.
LR electrospray ionisation (ES) spectra were recorded on a Finni-
gan LCQ Ion Trap spectrometer, using a capillary voltage of 10 V.
High resolution electron impact spectra were recorded on a Kratos
MS902 mass spectrometer. High resolution electrospray mass spec-
trometry were recorded on a Bruker ApexII Fourier Transform Ion
Cyclotron Resonance mass spectrometer, 7.0 T magnet at the Uni-
versity of New South Wales. Microanalyses were performed by the
Microanalytical Unit at the University of Otago, New Zealand. CD
measurements were recorded on a Jasco J-710 spectropolarimeter
using a 0.1 cm cell and the following parameters; range 200–500
nm, accumulation 10 scans, temperature 20 °C, step 0.5 nm, speed
50 nm min−1, response 2 s and bandwidth 1.0 nm.
NMR spectroscopy
2D Spectra were recorded on a BrukerAMX 400 spectrometer using
a 5 mm probe fitted with z-gradients and standard Bruker pulse pro-
grammes. Spectra were acquired in the phase sensitive mode using
time-proportional phase incrementation. Data sets resulting from
256 increments of t1 were acquired and zero filled to 1024 points,
with each free induction decay composed of 4096 data points. Typi-
cally 24 transients were recorded for each increment of t1 with a
recycle delay of 1.3–1.5 s. NOESY29,30 spectra were recorded with
gradient pulses in the mixing time of 600 ms. ROESY31,32 spectra
were recorded with a spinlock pulse of 600 ms.
While it has not been possible to assign the absolute configura-
tion of the complex 6, this study illustrates that chiral ligand design
in combination with the choice of transition metal can direct the
self-assembly of metallomacrocycles to result in self-recognition
of ligand enantiomers in a racemic mixture. The use of zinc(II) is
essential to this assembly process by providing reversible labile
coordination to the kinetically accessible binding sites that are
present in both ligand enantiomers, followed by formation of the
thermodynamic complex. Oxygen coordination to zinc(II) is also a
requirement for the stereoselectivity observed.
(±)-2,8-Dimethoxydibenzobicyclo[b,f][3.3.1]nona-5a,6a-diene-
6,12-dione 3
p-Methoxybenzylnitrile (3.7 cm3, 27 mmol) and powdered potas-
sium hydroxide (1.52 g, 27 mmol) were dissolved in diiodometh-
ane (1.3 cm3, 16 mmol) and refluxed at 155 °C for 45 min. The
reaction mixture was cooled to room temperature, diluted with
water, and extracted with dichloromethane (5 × 50 cm3). The com-
bined organic extracts were washed with water (100 cm3), dried
(Na2SO4) and concentrated in vacuo to give crude (±, meso)-2,4-
bis(p-methoxyphenyl)pentanedinitrile 1 as a red/brown oil (5.8 g).
The crude bisnitrile 1 was dissolved in ethanol (40 cm3), aqueous
potassium hydroxide (40%, 30 cm3) was added and the mixture
heated at 80 °C for 18 h. The ethanol was removed in vacuo and
the resultant thick oily residue was diluted with water (50 cm3) and
washed with dichloromethane (4 × 75 cm3) until the organic phase
was clear. The aqueous phase was acidified to pH < 1 with aqueous
HCl (3 M, 100 cm3) to give a cloudy cream coloured suspension,
which was extracted with ethyl acetate (4 × 75 cm3). The combined
organic extracts were dried (Na2SO4) and concentrated in vacuo to
give crude (±, meso)-2,4-bis(p-methoxyphenyl)pentanedioic acid 2
as a light yellow residue (4.615 g) which was used without further
purification.
Conclusions
The addition of zinc(II) to the racemic molecular cleft (±)-L
results in the formation of [2 + 2] metallomacrocycles in which
the (+) and (−) ligands undergo self-recognition to form a pair
of enantiomeric metallomacrocycles. To our knowledge, the
slow equilibration of zinc(II) complexes on the NMR time scale
observed with L is unprecedented with bisbipyridyl ligands. The
equilibration may be attributed to incorporation of the cleft in the
ligand design which orients the metal binding groups (including
the oxygen in the linker) in restricted locations. The assembly
involving the racemic ligand is one of the few examples where
enantiomeric ligand ligand self-recognition has been achieved in
the assembly of metallomacrocycles. The formation of oligomers
with copper(I) suggests that the methyleneoxy linker is slightly
too long and flexible to promote exclusive assembly of [2 + 2]
metallomacrocycles. This result, along with the coordination of
oxygen to zinc(II) in 6 suggest that ligands containing a methy-
lene linker directly connecting the bipyridyls to the cleft may
allow controlled assembly of a single, stable diastereomer with
both copper(I) and zinc(II). Further studies on the assembly and
molecular recognition properties of these, and related cleft-con-
taining metallomacrocycles, are underway in our laboratories.
Polyphosphoric acid (85%, 10 cm3) was heated to 100 °C, until it
began to reflux and was immediately cooled to 85 °C and the crude
acid 2 (0.5074 g) was added. The resulting reaction mixture was
stirred for 10 min at 85 °C resulting in formation of an intense scar-
let solution. The reaction was quenched, without cooling, by pour-
ing onto ice (100 g) with repeated heating to the reaction flask to
promote the transfer of the viscous fluid. The resultant green/brown
1 7 1 2
D a l t o n T r a n s . , 2 0 0 4 , 1 7 0 8 – 1 7 1 4

View Article Online
suspension was extracted with dichloromethane (8 × 75 cm3). The
combined organic extracts were washed with aqueous potassium
hydroxide (5% w/v, 500 cm3), dried (Na2SO4) and concentrated in
vacuo to give the crude product as a red/brown residue (0.170 g).
Purification by flash chromatography on silica (100% dicholo-
romethane) afforded a white solid, which was recrystallised from
dichloromethane to give bismethoxycleft 3 as white fluffy crystals
(160 mg, 35%, mp 194.0–195.0 °C). Anal. calc. for C19H16O4 [308]
C, 74.0; H, 5.23; O, 20.76. Found: C, 70.41; H, 5.14; O, 24.45%);
was triturated with hot methanol to produce a white solid that
was collected and washed with methanol (51.8 mg). Recrystalli-
sation from dichloromethane with infusion of methanol afforded
ligand (±)-L as a fine white solid (35 mg, 52%, mp 187.5–
189.0 °C). Anal. calc. Found: C 76.1, H 5.0, N 8.8 C41H32N4O4
[644] requires C 76.4, H 5.0, N 8.7%.; max/cm−1 (CHCl3) 1574,
1604, 1682; max/nm (CHCl3) 291 (/dm−3 mol−1cm−1 32 000),
302 (sh), 324 (sh); H (400 MHz, CDCl3) 2.63 (6H, s, CH3), 2.95
(2H, app.t, J 2.8 Hz, H-l), 3.94 (2H, app.t, J 2.8 Hz, H-k, H-m),
5.23 (4H, s, H-a,b /H-c,d), 7.16–7.20 (4H, m, H-f, H-i, H-5/8),
7.38 (2H, d, J 8.6 Hz, H-g, H-h), 7.44 (2H, d, J 7.6 Hz, H-5′/8′),
7.61 (2H, d, J 2.7 Hz, H-e, H-j), 7.70 (2H, t, J 7.8 Hz, H-4/7),
7.80 (2H, t, J 7.8 Hz, H-4′/7′), 8.19 (2H, d, J 7.8 Hz, H-3/6), 8.32
(2H, d, J 7.8 Hz, H-3′/6′); ES m/z 645 ([M + H]+, 100%); HRMS:
C41H33N4O4 requires 645.2502; found 645.2496.

max/cm−1 (CH2Cl2) 1704; max/nm (CH2Cl2) 253 (/dm−3 mol−1cm−1
33500), 323 (sh), 345 (sh), 355 (sh), 375 (sh); H (300 MHz; CDCl3)
2.96 (2H, t, J 3.0 Hz, H-l), 3.78 (6H, s, OCH3), 3.94 (2H, t, J 3.0 Hz,
H-k, H-m), 7.06 (2H, dd, J 8.5, 2.9 Hz, H-f, H-i), 7.36 (2H, d, J
8.5 Hz, H-g, H-h), 7.43 (2H, d, J 2.8 Hz, H-e, H-j). EI m/z 308 (M+˙,
100), 291 (25), 280 (40), 237 (26), 165 (20%).
Using the same method as above and diol (+)-4 (68 mg,
243 mol), ligand (+)-L was obtained as as a white solid (26 mg,
17%, mp 185.2–187.0 °C). []D22 +0.333 (c, 0.145, CH2Cl2); H
(400 MHz, CDCl3) 2.63 (6H, s, CH3), 2.95 (2H, app.t, J 2.9 Hz, H-
l), 3.94 (2H, app.t, J 2.9 Hz, H-k, H-m), 5.25 (4H, s, H-a,b /H-c,d),
7.15–7.20 (4H, m, H-f, H-i, H-5/8), 7.39 (2H, d, J 8.5 Hz, H-g, H-
h), 7.44 (2H, d, J 7.6 Hz, H-5′/8′), 7.61 (2H, d, J 2.8 Hz, H-e, H-j),
7.69 (2H, t, J 7.8 Hz, H-4/7), 7.80 (2H, t, J 7.8 Hz, H-4′/7′), 8.19
(2H, d, J 7.8 Hz, H-3/6), 8.32 (2H, d, J 7.8 Hz, H-3′/6′); ES m/z 645
([M + H]+, 100%), 667 ([M + Na]+, 25%); HRMS: C41H32N4O4Na
requires 667.2315; found 667.2301.
Using the same method as above and (−)-4 (26 mg, 93 mol),
ligand (−)-L was obtained as a white solid (16 mg, 27%, mp
190.5–192.5 °C). []D22 −0.333 (c, 0.165, CH2Cl2); H (400 MHz,
CDCl3) 2.63 (6H, s, CH3), 2.95 (2H, app.t, J 2.9 Hz, H-l), 3.94 (2H,
app.t, J 2.9 Hz, H-k/H-m), 5.25 (4H, s, H-a,b /H-c,d), 7.16–7.20
(4H, m, H-f, H-i, H-5/8), 7.39 (2H, d, J 8.5 Hz, H-g, H-h), 7.44
(2H, d, J 7.7 Hz, H-5′/8′), 7.61 (2H, d, J 2.8 Hz, H-e, H-j), 7.70
(2H, t, J 7.8 Hz, H-4/7), 7.80 (2H, t, J 7.8 Hz, H-4′/7′), 8.18 (2H,
d, J 7.8 Hz, H-3/6), 8.32 (2H, d, J 7.8 Hz, H-3′/6′); ES m/z 645
([M + H]+, 100%), 667 ([M + Na]+, 25%); HRMS C41H32N4O4 re-
quires 645.2502; found 645.2495.
(±)-2,8-Dihydroxydibenzobicyclo[b,f][3.3.1]nona-5a,6a-diene-
6,12-dione 4
Method 1. Dione 3 (195 mg, 0.63 mmol) was dissolved in dry,
distilled dichloromethane and cooled to −5 °C. Boron tribromide
was added in aliquots (2 × 640 mm3, 1.27 mmol) over 1 h, and the
reaction was stirred at room temperature for 17 h. The reaction was
quenched by addition of water (10 cm3) and was extracted with
ether (5 × 18 cm3). The combined ether extracts were washed with
a saturated aqueous solution of sodium hydrogen carbonate (2 ×
10 cm3) and brine (2 × 10 cm3), dried (Na2SO4) and the solvent re-
moved to give a brown/white solid (183 mg). Recrystallisation from
dicholoromethane afforded the product 4 as white crystals (102 mg,
58%, mp 206–207 °C). max/nm (MeOH) 224 (/dm−3 mol−1cm−1
25900), 252 (sh), 326 (sh), 350 (sh); H (200 MHz, MeOD) 2.90
(2H, t, J 3.6 Hz, H-l), 3.84 (2H, t, J 3.6 Hz, H-k, H-m), 6.98 (2H, dd,
J 2.8, 8.5 Hz, H-f, H-i), 7.26 (2H, d, J 8.4 Hz, H-g, H-h), 7.27 (2H,
d, J 2.8 Hz, H-e, H-j); EI m/z 280 (M+˙, 100), 252 (20), 223 (20%);
HRMS: C17H12O4 requires 280.0730; found 280.0736.
Method 2. A solution of dibromide 5 (50 mg, 0.12 mmol)
in DMSO (2 cm3) was added to
a stirred mixture of
bis(pinacolato)diboron (78.8 mg, 0.31 mmol), Pd(dppf)Cl2·CH2Cl2
(6.1 mg, 0.01 mmol) and anhydrous potassium acetate (97.4 mg,
0.99 mmol) under nitrogen. The mixture was stirred at 80 °C under
nitrogen for 22 h. Ethyl acetate (10 cm3) was added, and the mixture
filtered through sintered glass. The filtrate was added to aqueous
sodium chloride (20%, 10 cm3) and the aqueous phase extracted
with ethyl acetate (2 × 10 cm3). The combined organic layers were
washed with brine (3 × 10 cm3), dried (Na2SO4) and concentrated in
vacuo to give the crude product as a brown oil. Purification by flash
chromatography on silica (50% ethyl acetate/dichloromethane) af-
forded the boronate as white crystals (37 mg, 60%). H (200 MHz;
CDCl3) 1.29 (12H, s, 4 × CH3), 2.98 (2H, t, J 2.9 Hz, H-l), 4.03
(2H t, J 2.9 Hz, H-k, H-m), 7.44 (2H, d, J 7.6 Hz, H-g, H-h), 7.89
(2H, dd, J 1.4, 7.6 Hz, H-f, H-i), 8.38 (2H, d, J 1.2 Hz, H-e, H-j); EI
(m/z) 500 (100, M+˙), 414 (20), 229 (35), 227 (20%). The boronate
was immediately dissolved in methanol (4 cm3) and a solution of
hydrogen peroxide in water (30%, 1 cm3) added. The mixture was
stirred for 1.5 h, extracted with dichloromethane (5 × 10 cm3), dried
(Na2SO4) and concentrated in vacuo to give 4 as a yellow solid in
quantitative yield. The product has spectroscopic data identical to
that produced by method A.
Zinc(II) complexes
Ligand (±)-L (2.1 mg, 3.3 mmol) and zinc(II) triflate (2.3 mg,
6.3 mmol) were suspended in CD3CN (500 mm3). Vigorous stir-
ring resulted in dissolution to give a clear, colourless solution. After
30 h the sample was analyzed. H (400 MHz, CD3CN) 2.09 (3H, s,
CH3), 2.12 (3H, s, CH3), 2.34 (1H, d, J 15.0 Hz, H-c), 2.97 (2H, qt,
J 2.8 Hz, J 13.4 Hz, H-l), 3.87 (2H, br. d, J 10.5 Hz, H-k, H-m), 4.48
(1H, d, J 15.0 Hz, H-d), 4.93 (1H, dd, J 2.8 Hz, J 8.4 Hz, H-i), 5.20
and 5.35 (2H, ABq, J 16.4 Hz, H-b, H-a), 5.68 (1H, dd, J 2.9 Hz,
J 8.5 Hz, H-f), 6.35 (1H, d, J 2.7 Hz, H-j), 6.57 (1H, d, J 8.4 Hz,
H-h), 6.84 (1H, d, J 8.6 Hz, H-g), 7.16 (1H, d, J 8.0 Hz, H-8′), 7.18
(1H, d, J 2.9 Hz, H-e), 7.60 (1H, d, J 7.8 Hz, H-3/5), 7.72 (1H, d,
J 7.9 Hz, H-3/5), 7.74 (1H, d, J 7.8 Hz, H-8), 7.94 (1H, t, J 7.9 Hz,
H-4), 8.10 (1H, d, J 8.1 Hz, H-3′/5′), 8.13 (1H, d, J 7.9 Hz, H-3′/5′),
8.40 (1H, t, J 8.0 Hz, H-7), 8.46 (1H, t, J 8.0 Hz, H-4′), 8.51 (1H, t,
J 8.0 Hz, H-7′), 8.66 (1H, d, J 8.1 Hz, H-6), 8.80 (1H, d, J 8.1 Hz,
H-6′); ES m/z 1502 ([ZnL2OTf]+, 3%), 1352 ([ZnL2 − H]+, 1), 1008
([ZnLOTf2 + H]+, 2), 645 ([L + H]+, 74), 323 ([L + H]2+, 100).
Ligand (+)-L (2.4 mg, 3.73 mol) and zinc(II) triflate (2.7 mg,
7.45 mol) were suspended in CD3CN (500 mm3). Vigorous stir-
ring resulted in dissolution to give a clear, colourless solution. After
3 h the sample was analysed. H (400 MHz) gave identical spectra
to (±)-6 with Zn.
(±)-2,8-Bis[1′-(6-methyl-2,2-bipyridin-6-yl)methoxy]diben
zobicyclo[b,f][3.3.1]-nona-5a,6a-diene-6,12-dione L
Ligand (−)-L (2.1 mg, 3.26 mg) and zinc(II) triflate (2.5 mg,
6.52 mol) were suspended in CD3CN (500 mm3). Vigorous stir-
ring resulted in dissolution to give a clear, colourless solution. After
3 h the sample was analysed. H (400 MHz, CD3CN) gave identical
spectra to (±)-L with Zn.
Potassium tert-butoxide (28 mg, 0.25 mmol) was added to a so-
lution of bisphenol (±)-4 (29 mg, 0.103 mmol) in THF (3 cm3),
under nitrogen.After 1 h, a solution of 6-bromomethyl-6′-methyl-
2,2′-bipyridine (65 mg, 0.25 mmol) in THF (2 cm3) was added
dropwise to the yellow, opaque solution. The reaction mixture
was refluxed for 40 h, filtered to remove a fine brown precipitate,
and the solvent removed to give crude ligand (64 mg) as a pale
yellow oil. Purification by flash chromatography on silica (1%
methanol/dichloromethane) afforded a clear yellow oil, which
Copper(I) complexes
Ligand L (2.5 mg, 3.9 mmol) and copper(II) triflate (1.4 mg,
3.9 mmol) were suspended in CD3CN (500 mm3), resulting in
D a l t o n T r a n s . , 2 0 0 4 , 1 7 0 8 – 1 7 1 4
1 7 1 3

View Article Online
12 T. Bark, M. Duggeli, H. Stoeckli-Evans and A. Von Zelewsky, Angew.
Chem. Int. Ed., 2001, 40, 2848.
13 M. Ziegler and A. Von Zelewsky, Coord. Chem. Rev., 1998, 177, 257.
14 R. Prabaharan, N. C. Fletcher and M. Nieuwenhuyzen, J. Chem. Soc.,
Dalton Trans., 2002, 602.
15 A. Lutzen, M. Hapke, J. Griep-Raming, D. Haase and W. Saak, Angew.
Chem. Int. Ed., 2002, 41, 2086.
16 M. Massod, E. J. Enemark and T. D. P. Stack, Angew. Chem. Int. Ed.,
1998, 37, 928.
dissolution to give a yellow/orange solution. Hydrazine hydrate
(0.2 mm3, 4 mmol) was added and the reaction mixture darkened
to give a red/orange solution. 1H NMR (400 MHz, CD2Cl2) H 2.27
(s, CH3), 2.80 (s, CH3), 2.88 (m, H-l), 3.77 and 3.91 (m, H-k, H-m),
4.35 and 4.56 (ABq, J 9.4 Hz), 4.85 and 4.94 (ABq, J 15.6 Hz); ES
m/z 1738 ([Cu2L2OTf2 + Na]+, 1%), 1582 ([Cu2L2OTf + H2O]+, 2),
1565 ([Cu2L2OTf]+, 4), 707 ([Cu2L2]2+/[CuL]+, 100%)
17 A. Bilyk and M. M. Harding, Chem. Commun., 1995, 1697.
18 A. Bilyk, M. M. Harding, P. Turner and T. W. Hambley, J. Chem. Soc.,
Dalton Trans., 1994, 2783.
19 M.A. Houghton,A. Bilyk, M. M. Harding, P. Turner and T. W. Hambley,
J. Chem. Soc., Dalton Trans., 1997, 2727.
20 A. Bilyk and M. M. Harding, J. Chem. Soc., Dalton Trans., 1994, 77.
21 H. Tatemitsu, F. Ogura, Y. Nakagawa, M. Nakagawa, K. Naemura and
M. Nakagawa, Bull. Chem. Soc. Jpn., 1975, 48, 2473.
22 J. D. Field, P. Turner, M. M. Harding, T. Hatzikominos and L. Kim, New
J. Chem., 2002, 26, 720.
23 F. Ogura, A. Nakao and M. Nakagawa, Bull. Chem. Soc. Jpn., 1979, 52,
1165.
24 M. C. Kimber, A. C. Try, L. Painter, M. M. Harding and P. Turner, J.
Org. Chem., 2000, 65, 3042.
25 G. R. Newkome, V. K. Gupta and F. R. Fronzcek, Inorg. Chem., 1983,
122, 171.
26 R. Stiller, Eur. J. Inorg. Chem., 1998, 977.
27 A. C. Try, L. Painter and M. M. Harding, Tetrahedron Lett., 1998, 39,
9809.
28 N. Miyaura and A. Suzuki, Chem. Rev., 1995, 95, 2457.
29 J. Jeener, B. H. Meier, P. Bachmann and R. R. Ernst, J. Chem. Phys.,
1979, 71, 4546.
30 R. Wagner and S. Berger, J. Magn. Reson., 1996, 123A, 119.
31 A. Bax and D. G. Davis, J. Magn. Reson., 1985, 63, 207.
32 T.-L. Hwang and A. J. Shaka, J. Am. Chem. Soc., 1992, 114, 3157.
Acknowledgements
Financial support from theAustralian Research Council is gratefully
acknowledged. We thank L. Kim for resolution experiments and Dr
I. Luck for assistance with low temperature NMR experiments.
References
1 J.-M. Lehn, in Supramolecular Chemistry, Concepts and Perspectives,
VCH, Weinheim, 1995.
2 Comprehensive Supramolecular Chemistry, ed. J. L.Atwood, J. E. Davies,
D. D. MacNicol and F. Vögtle, Elsevier Science, London, 1996, vol. 6.
3 W.-Y. Sun, M. Yoshizawa, T. Kusukawa and M. Fujita, Curr. Opin.
Chem. Biol., 2002, 6, 757.
4 C. Kaes, A. Katz and M. W. Hosseini, Chem. Rev., 2000, 100, 3553.
5 C. J. Jones, Chem. Soc. Rev., 1998, 27, 289.
6 M. Albrecht, Chem. Rev., 2001, 101, 3457.
7 N. C. Fletcher, J. Chem. Soc., Perkin Trans. 1, 2002, 1831.
8 C. Piguet, G. Bernardinelli and G. Hopfgartner, Chem. Rev., 1997, 97,
2005.
9 G. Baum, E. C. Constable, D. Fenske, C. E. Houecroft and T. Kulke,
Chem. Eur. J., 1999, 1862.
10 R. Prabaharan and N. C. Fletcher, Dalton Trans., 2003, 2558.
11 Y. Cui, H. L. Ngo and W. Lin, Chem. Commun., 2003, 1388.
1 7 1 4
D a l t o n T r a n s . , 2 0 0 4 , 1 7 0 8 – 1 7 1 4