6
small quantities of the fractions, accurate determination of the
weights of the product was not possible.
of template 2 to the linear and cyclic dimers, 3 3 10 and 2 3
7
21
10
M
respectively, are consistent with the observed
8
In the absence of the template only traces of the capped dimer
could be isolated (0.009 au—absorbance unit; less than 2%
yield). However, in the templated experiments the isolated yield
of pure 3 increased 6-fold to approximately 10% (0.055 au). The
structure of the ring-closed dimer was confirmed by HMBC-
templating effect. However, the relatively small increase in the
binding constant upon cyclization highlights the flexibility of
the host and rationalises the low yield of 3, i.e. the equilibrium
is shifted only slightly from oligomeric species towards the
formation of 3.‡
It was possible to identify mono- and disubstituted linear
dimers as intermediates of the reaction by MALDI-TOF and H
1
3
NMR showing a cross-peak for the single enriched C signal
d = 172.4 ppm) with the allylic CH protons (d = 4.63 ppm).
1
(
2
This significant and reproducible increase in the yield confirms
the templating effect of the bidentate ligand and represents the
first thermodynamic capping of a linear porphyrin dimer using
reversible p-allyl palladium chemistry. To test the generality of
the reaction, and to illustrate the feasibility of utilising libraries
of capping groups at a later stage, cyclic dimers containing
glutaric and 1,3-phenylenediacetic acid as capping groups were
also prepared in comparable yields. In addition to NMR
characterisation, MALDI-TOF mass spectrometry was used to
distinguish between linear and cyclic species.
The control experiments without template and those in the
presence of pyridine (2 eq.) gave the same results. Pyridine is a
monodentate equivalent of the template 2, i.e. the pyridine units
are not covalently linked as for the bidentate template. This
result confirms that the dipyridyl compound 2 templates the
formation of the cyclic dimer by stabilising it and that the
pyridyl functionality of the template does not influence the
outcome of the transesterification. The large binding constants
NMR. However, the relative quantities of all intermediates
could not be estimated accurately due to the low solubility of
oligomers and disubstituted linear dimers with two free
carboxylic acid groups.§ It should be noted that the solubility of
the mixture was not problematic during the reaction itself as the
triethylamine salts of all intermediates are soluble in chloro-
form.
In summary, we have shown the synthesis under reversible
conditions of three cyclic porphyrin dimers using palladium-
catalysed transesterification. The reaction was templated by a
bidentate pyridyl ligand which improved the yield of the cyclic
product 6-fold. To our best knowledge, this is the first
supramolecular application of reversible p-allyl palladium
chemistry; it is also one of the most effective examples yet
reported of amplification of a specific product from an
exchanging library of components.
We thank the BBSRC, AstraZeneca and the Cambridge
European Trust for financial support and Dr Y.-F. Ng for
recording some of the NMR spectra.
Notes and references
†
The molecular weights of linear dimer and unlabelled succinic capped
dimer differ by only 2 a.m.u., precluding the use of MALDI-TOF mass
spectrometry.
‡
This point is also apparent in the uniformity of yield from the three
different capping acids of very different geometry.
The overall amount of isolated porphyrinic material was measured by
§
UV–Vis spectroscopy and marked differences between the templated and
non-templated reactions were observed. The overall recovery of soluble
porphyrinic material was considerably lower for the non-templated reaction
(0.1 au compared with 0.2 au for the templated reaction). This result implies
that the template suppresses the formation of insoluble oligomers in
addition to promoting the formation of cyclic product. The decrease in total
oligomer formation is greater than the increase in cyclic dimer formation,
suggesting that these are two separate effects, i.e. the dipyridyl compound
2
is a positive template for the cyclic dimer formation and a negative
template for the oligomer formation.
1
2
For reviews covering this area see: G. R. L. Cousins, S.-A. Poulsen and
J. K. M. Sanders, Curr. Opin. Chem. Biol., 2000, 4, 270; J.-M. Lehn,
Chem. Eur. J., 1999, 5, 2455; B. Klekota and B. L. Miller, Trends
Biotechnol., 1999, 17, 205.
For more recent examples see: R. L. E. Furlan, G. R. L. Cousins and
J. K. M. Sanders, Chem Commun., 2000, 1761, previous paper in this
issue (DOI: 10.1039/b006149m); M. G. Simpson, S. P. Watson, N.
Feeder, J. E. Davies and J. K. M. Sanders, Org. Lett., 2000, 2, 1435; O.
Ramström and J.-M. Lehn, CHEMBIOCHEM, 2000, 1, 41.
3
4
D. G. Hamilton, N. Feeder, S. J. Teat and J. K. M. Sanders, New J. Chem.,
1
998, 22, 1019; T. J. Kidd, D. A. Leigh and A. J. Wilson, J. Am. Chem.
Soc., 1999, 121, 1599.
M. Nakash, Z. Clyde-Watson, N. Feeder, J. E. Davies, S. J. Teat and
J. K. M. Sanders, J. Am. Chem. Soc., 2000, 122, 5286; M. Nakash and
J. K. M. Sanders, J. Org. Chem., in the press.
5
6
7
8
L. J. Twyman and J. K. M. Sanders, Tetrahedron Lett., 1999, 40,
6
681.
J. Tsuji, Palladium Reagents and Catalysts, John Wiley & Sons,
Chichester, 1995.
C. Amatore, A. Jutand, G. Meyer and L. Mottier, Chem. Eur. J., 1999, 5,
4
66.
The binding constants were measured in dry CHCl
and guest concentrations of 10 M and 10 M respectively. The data
were analysed using the commercial program Specfit.
3
at 25 °C with host
25
24
Scheme 2 Reagents and conditions: concentration of porphyrin dimer 5
mM, Pd(PPh ) (0.1 eq.), Et N, CHCl , 55 °C, 6 h.
3 4 3 3
1764
Chem. Commun., 2000, 1763–1764