Metal-driven self assembly of C3 symmetry molecular cages

Article abstract of DOI:10.1039/b001628o

A series of tris(2-aminoethyl)amine (TREN) derivatives functionalized on the arms with pyridine or methylimidazole ligands bind very strongly to a trisporphyrin derivative via apical coordination, with formation of molecular cages; the strength of the bin

Full text of DOI:10.1039/b001628o

Metal-driven self assembly of C3 symmetry molecular cages
Fulvia Felluga,a Paolo Tecilla,*a Louise Hillier,b Christopher A. Hunter,*b Giulia Licinic and Paolo Scrimin*c
a University of Trieste, Department of Chemical Sciences, via Giorgieri, 1 34127-Trieste, Italy
b University of Sheffield, Department of Chemistry, Sheffield, UK S3 7HF
c
University of Padova, Department of Organic Chemistry and CNR center CMRO, via Marzolo, 1 35131-Padova,
Italy. E-mail: scrimin@mail.chor.unipd.it
Received (in Liverpool, UK) 24th February 2000, Accepted 10th May 2000
A series of tris(2-aminoethyl)amine (TREN) derivatives
functionalized on the arms with pyridine or methylimidazole
ligands bind very strongly to a trisporphyrin derivative via
apical coordination, with formation of molecular cages; the
strength of the binding (Kb up to 108.8 mol21 dm3) depends
on the structure of the TREN derivative as well as on its
coordination to Zn(II).
The self assembly of proteins to elicit new and unique functions
is a common feature in the biological world.1 A case in point is
HIV-1 protease, whose activity is dependent on a hydrogen
bond-driven dimerization of two protein units.2 The possibility
of realizing molecules of controlled geometry and shape in the
laboratory by assembling easily synthesized subunits is an
appealing goal which is attracting enormous interest, demon-
strated by the continuous flow of contributions appearing in the
chemical literature.3 Such ‘supermolecules’ may find useful
application in molecular recognition and selective substrate
transformation, sensing, signal transduction and, eventually, in
the realization of miniaturized nanostructures. Metal–ligand
coordination is one of the driving forces exploited for the
assembly of these structures.4 Examples are provided by the
work of the Stang,5 Fujita,6 Hamilton,7 Sauvage,8 and Lehn9
groups. Self-assembling porphyrin oligomers towards macro-
cycles or polymers have been reported by Hunter et al.,10 while
Tabushi and coworkers, as early as 1985,11 elegantly demon-
strated allosteric binding to a Zinc-gable porphyrin.
We report here on the formation of molecular cages driven by
apical coordination of the terminal bases of tripodal ligands 1a–
c to the three zinc porphyrin units of 2. The possibility of
modifying the conformation of the tris(2-aminoethyl)amine
(TREN) platform of derivatives 1a–c via metal coordination
gives the system a tunable geometry for controlling the binding
to 2.
Ligand 1a was readily obtained by reaction of 3-pyr-
idinaldehyde with TREN followed by reduction with NaBH4.
Ligands 1b,c were prepared via condensation of TREN with the
4-formylbenzenamide derivatives, obtained by reaction of
4-formylbenzoyl chloride with 4-aminopyridine (1b) or me-
thylhystamine (1c), followed by reduction with NaBH4.† The
trisporphyrin derivative 2 was prepared by coupling 1,3,5-ben-
zene tricarbonyl trichloride with the corresponding aminopor-
phyrin,12 followed by metallation with zinc acetate.† Porphyrin
2, as well as the free bases 1a–c, are soluble in chlorinated
organic solvents. The solubility of the 1+1 Zn(II) complexes of
the TREN derivatives in these solvents depends on the nature of
the counterions: more hydrophilic nitrates are insoluble, while
more lipophilic hexafluorophosphates are freely soluble in these
media.
Trisporphyrin 2 shows a typical UV–vis spectrum, with
maxima at 420 nm (e = 430,000 mol21 dm3 cm21) and 550 nm
(e = 37,000 mol21 dm3 cm21), indicating that there is no
intramolecular interaction between the porphyrins. Binding of
an apical ligand causes a shift of these peaks to 430 and 565 nm,
respectively. The very intense absorptions in the 420–430 nm
region can be used for determination of the affinity constant
with apical ligands and, in the present case, for the formation of
cage complexes between 2 and 1. Fig. 1 shows the changes in
the absorption spectrum of 2 upon addition of increasing
amounts of 1a in CH2Cl2. The presence of a well-defined
isosbestic point is suggestive of the formation of a single
complex, and the fact that all the zinc porphyrin sites are fully
complexed after the addition of one equivalent of ligand implies
that a 1+1 cage complex has been formed. The complex so
formed does not convert into species of different stoichiometry
in the presence of up to a twofold excess of the TREN
derivative, contrary to recent observations for a bisporphyrin
receptor in the presence of multiarmed amines.13 Clearly, a key
role is played here by the complementary structure of the two
subunits involved in the formation of the supramolecular cage
complex. Analysis of the absorbance vs. concentration data‡
leads to a well-behaved complexation curve (Fig. 1, inset)
which gives a very high binding constant (log Kb = 8.81) for the
1+1 complex. Similar curves were obtained for tripodal ligands
1b,c and 1a,b–Zn(II) complexes.
DOI: 10.1039/b001628o
Chem. Commun., 2000, 1087–1088
This journal is © The Royal Society of Chemistry 2000
1087

Fig. 2 Schematic diagram of the molecular cage formed by binding of 1a
to 2.
Fig. 1 Change in the absorbance of a 60 mM solution of 2 in CH2Cl2 upon
addition of 1a. Inset: binding isotherm of the data at 427 nm. The solid line
represents the computer-generated best fit for a 1+1 complex.
imposing a more stringent geometric constraint for binding,
resulting in a smaller binding constant.
In conclusion, we have shown that tripodal molecules, based
on complementary porphyrins and pyridine or imidazole
ligands, give highly symmetrical molecular cages with very
strong binding constants which are potentially suitable for
molecular recognition and selective transformation of sub-
strates trapped in the cavity: work aimed at this goal is ongoing
in our laboratories.
This work was supported in part by a grant within the frame
of the British–Italian Collaboration in Research and Higher
Education.
The titration data are summarized in Table 1, along with the
values determined for model ligands 3–5. The values of the
parameters EM and b are also listed in the table. EM§
corresponds to the ratio {(Kb)cage/[(Kb)model]3}0.5 and is the
effective molarity for the intramolecular cyclizations required
to form the cage. b14 is simply the ratio (Kb)cage/(Kb)model and
gives an indication of the absolute strength of binding
associated with the formation of the cage, with respect to the
single apical coordination of the model compound to one
porphyrin. Inspection of Table 1 reveals that tripodal ligand 1a
shows the highest binding constant to 2, which reduces 30-fold
upon Zn(II) binding to TREN. In the case of 1b, the binding
constant is lower and coordination of Zn(II) to TREN does not
induce any significant change. Allowing for the stronger
binding interaction of methylimidazole to the Zn-porphyrin
(compared with pyridine), the least effective ligand is 1c, with
EM and b values significantly lower than those of 1a and 1b.
The b values indicate that cooperative binding of the tripodal
ligands to the three porphyrins leads to a higher affinity than
single apical ligands (up to five orders of magnitude in the case
of 1a). Inspection of molecular models reveals that, in the case
of 1b, the length of the tripodal arms and their flexibility is such
that coordination of Zn(II) does not significantly affect their
binding geometry. However, this flexibility reduces the strength
of binding with respect to 1a, due to the freezing out of
conformational mobility on complexation.14 The disadvantage
of flexibility is further evidenced by the relatively low values of
EM and b for 1c, which has even longer arms. For 1a, the length
of the arms¶ is such that apical binding to the porphyrins
requires the latter to move out of the plane of the central
benzene by rotation around the amide bonds pointing inwards
(Fig. 2). Complexation of Zn(II) to the TREN platform further
reduces the distance between the three pyridine nitrogens,
Notes and references
† All new compounds gave the expected 1H- and 13C-NMR spectra and the
correct elemental analyses (C, H, N).
‡ The binding isotherm was fitted using the program HOSTEST II. See:
C. S. Wilcox, in Frontiers in Supramolecular Organic Chemistry and
Photochemistry, H.-J. Schneider and H. Dürr, VCH, Weinheim, 1991.
§ In the cage formation the first binding is intermolecular (K), the second
binding is intramolecular (EM1 3 K) and the third binding is also
intramolecular (EM2 3 K). Thus, the overall observed binding constant is
EM1 3 EM2 3 K3. Assuming that EM1 = EM2, the effective molarity is the
square root of the product EM1 3 EM2.
¶ Energy minimization with the HyperChem program (Hyperchem 2 for
Windows, © 1991, Hypercube Inc. and Autodesk Inc.) indicates that two
Zn(II) ions in the porphyrin derivative 2 are ca. 20 Å apart when the
molecule is flat. The distance between two pyridine nitrogens of 1a is, in the
fully stretched conformation, ca. 15 Å.
1 W. Stites, Chem. Rev., 1997, 97, 1233; S. Jones and J. M. Thornton,
Proc. Natl. Acad. Sci. U.S.A., 1996, 93, 13.
2 R. Zutshi, J. Franciskovich, M. Shultz, B. Schweitzer, P. Bishop, M.
Wilson and J. Chmielewski, J. Am. Chem. Soc., 1997, 119, 4841.
3 M. M. Conn and J. Rebek, Chem. Rev., 1997, 97, 1647.
4 B. Linton and A D. Hamilton, Chem. Rev., 1997, 97, 1669.
5 P. J. Stang, D. H. Cao, S. Saito and A. M. Arif, J. Am. Chem. Soc., 1995,
117, 6273.
6 M. Fujita, F. Ibukuro, H. Hagihara and K. Ogura, Nature, 1994, 367,
720.
7 M. S. Goodman, J. Weiss and A. D. Hamilton, J. Am. Chem. Soc., 1995,
117, 8447.
8 N. Armaroli, F. Diederich, C. O. Dietrich-Buchecker, L. Flamigni, G.
Marconi, J.-F. Nierengarten and J.-P. Sauvage, Chem. Eur. J., 1998, 4,
406.
9 B. Hasenknopf, J.-M. Lehn, B. O. Kneisel, G. Baum and D. Fenske,
Angew. Chem., Int. Ed. Engl., 1996, 35, 1838.
10 X. Chi, A. J. Guerin, R. A. Haycock, C. Hunter and L. D. Sarson,
J. Chem. Soc., Chem. Commun., 1995, 2567.
11 I. Tabushi, S. Kugimiya, M. G. Kinnaird and T. Sasaki, J. Am. Chem.
Soc., 1985, 107, 4192.
Table 1 Binding constants of trisporphyrin 2 to tripodal ligands 1a–c and
model ligands 3–5 in CH2Cl2a at 25 °C
b
Ligand
log Kb
EMc
bc
3d
1a
1a·Zn
4d
3.75 ± 0.02
8.81 ± 0.07
7.36 ± 0.05
3.98 ± 0.03
7.75 ± 0.09
7.64 ± 0.05
5.37 ± 0.07
7.50 ± 0.09
—
—
6.0 3 1022
1.1 3 1022
—
1.1 3 105
4.1 3 103
—
1b
8.0 3 1023
7.1 3 1023
—
5.0 3 1025
5.9 3 103
4.6 3 103
—
1.3 3 102
1b·Zn
5d
1ce
a In the presence of 1% CH3CN. b Binding constant are expressed in mol21
dm3. c See the text for the definition of EM and b. The units of EM are mol
dm23, and b is dimensionless. d Binding constants of the model ligands are
the microscopic values determined assuming independent binding to each
porphyrin of 2. e The binding of 1c·Zn(II) to 2 does not follow a well
behaved isotherm; for this reason it has been omitted.
12 M. Gardner, A. J. Guerin, C. A. Hunter, U. Michelsen and C. Rotger,
New J. Chem., 1999, 23, 309.
13 J. N. H. Reek, A. P. H. J. Schenning, A. W. Bosman, E. W. Meijer and
M. J. Crossley, Chem. Commun., 1998, 11.
14 M. Mammen, S.-K. Choi and G. M. Whitesides, Angew. Chem., Int. Ed.,
1998, 37, 2754.
1088
Chem. Commun., 2000, 1087–1088