Self-assembly, binding, and dynamic properties of heterodimeric porphyrin macrocycles

Article abstract of DOI:10.1021/jo0505369

A series of heterodimeric tetralactam macrocycles have been self-assembled using two kinetically labile zinc porphyrin-pyridine interactions. The stability constants have been determined by UV-vis titrations in CHCl₃. The stability constants depend on the degree of preorganization of the linker units connecting the interacting groups. The ability of the self-assembled macrocycles to bind a terephthalamide guest was also investigated. One of the macrocycles was used for the construction of a [2]rotaxane. The dynamic properties of this system provide insight into the exchange mechanisms that operate in complex noncovalent assemblies.

Full text of DOI:10.1021/jo0505369

Self-Assembly, Binding, and Dynamic Properties of Heterodimeric
Porphyrin Macrocycles
Pablo Ballester,*,† Antoni Costa,‡ Pere M. Deya`,‡ Antonio Frontera,‡ Rosa M. Gomila,‡
Ana I. Oliva,| Jeremy K. M. Sanders,§ and Christopher A. Hunter*,¶
ICREA and Institute of Chemical Research of Catalonia (ICIQ), 43007-Tarragona, Spain,
Departament de Quı´mica, Universitat de les Illes Balears, 07122-Palma de Mallorca, Spain,
University Chemical Laboratories, Lensfield Road, Cambridge CB2 1EW, United Kingdom, and
Centre for Chemical Biology, Krebs Institute for Biomolecular Science, Department of Chemistry,
University of Sheffield, Sheffield S3 7HF, United Kingdom
pballester@iciq.es; c.hunter@sheffield.ac.uk
Received March 16, 2005
A series of heterodimeric tetralactam macrocycles have been self-assembled using two kinetically
labile zinc porphyrin-pyridine interactions. The stability constants have been determined by UV-
vis titrations in CHCl3. The stability constants depend on the degree of preorganization of the
linker units connecting the interacting groups. The ability of the self-assembled macrocycles to
bind a terephthalamide guest was also investigated. One of the macrocycles was used for the
construction of a [2]rotaxane. The dynamic properties of this system provide insight into the
exchange mechanisms that operate in complex noncovalent assemblies.
Introduction
[2]pseudorotaxanes and [2]rotaxanes using a three-
molecule approach. In both cases, a tetralactam macro-
cycle was self-assembled using reversible coordinate
bonds and was employed as the wheel component of the
interlocked molecular system. Herein, we report our
continued efforts to devise rotaxane-like structures, that
because of the kinetic lability of the zinc-pyridine inter-
action can reversibly assemble from and dissociate into
their components under thermodynamic conditions. A
series of dinuclear macrocycles 1 (Figure 1) were designed
as the wheel component. The symmetry of the two
complementary components of 1 simplifies their synthesis
compared to the porphyrin dimer 2 (Figure 1) previously
reported.4 The fact that metallomacrocycle 1 is not
assembled through a self-coordination process implies
that its formation can be achieved in the presence of a
slight excess of bisporphyrin 4 with respect to bispyridyl
3 (Scheme 1) or vice versa. Under these conditions when
a suitable guest is added, it should be possible to study
Interlocked molecular complexes are appealing systems
for the construction of simple molecular-level machines.1
From this viewpoint, [2]pseudorotaxane and [2]rotaxane
systems comprising a macrocycle (wheel) threaded by a
linear molecule (axle) constitute a very attractive and
interesting proposition. In most cases, the construction
of these molecular architectures employs a two-molecule
approach (macrocycle and linear component), both linked
by covalent bonds.2 Recently, Jeong3 and one of us4
reported different examples of noncovalent assembly of
† ICREA and ICIQ. Phone: +34 977 920 206; fax: +34 977 920 221.
‡ Universitat de les Illes Balears.
§ University Chemical Laboratories.
| ICIQ.
¶ University of Sheffield.
(1) (a) Sauvage, J.-P. Molecular Machines and Motors; Springer-
Verlag: Berlin, 2001; Vol. 99. (b) Balzani, V.; Credi, A.; Raymo, F. M.;
Stoddart, J. F. Angew. Chem., Int. Ed. 2000, 39, 3348-3391.
(2) (a) Schalley, C. A.; Beizai, K.; Vo¨gtle, F. Acc. Chem. Res. 2001,
34, 465-476. (b) Bottari, G.; Leigh, D. A.; Perez, E. M. J. Am. Chem.
Soc. 2003, 125, 13360-13361 and references therein.
(3) Jeong, K.-S.; Choi, J. S.; Chang, S.-Y.; Chang, H.-Y. Angew.
Chem., Int. Ed. 2000, 39, 1692-1695.
(4) Hunter, C. A.; Low, C. M. R.; Packer, M. J.; Spey, S. E.; Vinter,
J. G.; Vysotsky, M. O.; Zonta, C. Angew. Chem., Int. Ed. 2001, 40,
2678-2682.
10.1021/jo0505369 CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/27/2005
6616
J. Org. Chem. 2005, 70, 6616-6622

Properties of Heterodimeric Porphyrin Macrocycles
SCHEME 1. Synthesis of the Bispyridyl Ligands 3
and Zinc Bisporphyrins 4a
a Reagents and conditions: (a) i-Pr2NEt, CH2Cl2, 0 °C to reflux.
(b) Et3N, CH2Cl2, 0 °C to reflux. (c) Zn(OAc)2/CH2Cl2:CH3OH (3:
1).
5a6 and 5b,7 4-amino-3,5-lutidine 6,8 and the aminopo-
rphyrin 79 were prepared according to literature proce-
dures. Reaction of 2.2 equivalents of 6 with the diacid
chlorides 5a-b in dry methylene chloride and in excess
Hu¨nig’s base afforded, after trituration with diethyl
ether, ligands 3a-b in 80 and 22% yield, respectively.
Coupling the diacid chlorides 5a-b with aminoporphyrin
7 was accomplished by dropwise addition of a methylene
chloride solution of the chloride into a solution of the
same solvent, previously cooled at 0 °C containing 2
equivalents of the aminoporphyrin and excess triethy-
lamine. Column chromatography of the crude products
allowed the isolation of the free base bisporphyrins in
46 and 42% yield, respectively. Metalation was ac-
complished almost quantitatively using a solution of zinc
acetate in CH2Cl2:CH3OH (3:1) affording the zinc bispor-
phyrins 4a-b.
The self-assembly of 1 was studied at 295 K in
chloroform using 1H NMR and UV-vis absorption spec-
troscopy. A chloroform solution of the bisporphyrin
component 4 maintained at constant concentration was
titrated by adding incremental amounts of the bispyridyl
component 3. The maximum in the UV-vis absorption
spectra for the Soret band of the free zinc bisporphyrins
4 is centered at 420 nm. On addition of bispyridyl ligands
3, a new Soret band appeared at 430 nm with an
isosbestic point in all cases. This 10-nm red shift is
characteristic of coordination of a pyridine ligand to a
zinc porphyrin. Analysis of the UV-vis binding curves
FIGURE 1. Structural formulas of the self-assembled mac-
rocycles. The arrows show intermolecular NOEs observed in
a ROESY experiment for 1aa. See text for details.
simultaneously the exchange dynamics of opening the
macrocycle and host-guest binding. The heteromeric
composition of 1 also allows us to assess the effects of
preorganization of the 2,6-dicarboxamidopyridine linker
versus the analogous 2,6-dicarboxamidobenzene for each
component in the self-assembly of the macrocycle.
The 2,6-dicarboxamidopyridine unit has internal hy-
drogen bonds and is expected to adopt a conformation
that orients its two arms approximately perpendicular
to one another.5 The self-assembly of rectangular mac-
rocycles using a self-complementary porphyrin-pyridine
unit as in 2 (self-coordination) or complementary bispy-
ridine and bisporphyrin units as in 1 is favored when
the corner angles are close to 90°. In our design, two of
the corners are provided by the conformation adopted by
the 2,6-dicarboxamidopyridine (or benzene) units and the
other two by the axial interaction of the pyridine ligand
with the zinc porphyrin.
(6) Markees, D. G.; Kidder, G. W. J. Am. Chem. Soc. 1956, 78, 4130-
4135.
Results and Discussion
(7) (a) Lee, S. B.; Hong, J.-I. Tetrahedron Lett. 1998, 39, 4317-4320.
(b) Fosdick, L. S.; Fancher, O. E. J. Am. Chem. Soc. 1941, 63, 1277-
1279. (c) Calandra, J. C.; Svarz, J. J. J. Am. Chem. Soc. 1950, 72,
1027-1028.
(8) Essery, J. M.; Schofield, K. J. Chem. Soc. 1960, 4953-4959.
(9) Gardner, M.; Guerin, A. J.; Hunter, C. A.; Michelsen, U.; Rotger,
C. New J. Chem. 1999, 23, 309-316.
The synthesis of the bispyridyl ligands 3 and zinc
bisporphyrins 4 is outlined in Scheme 1. The dichlorides
(5) Hunter, C. A.; Sarson, L. D. Angew. Chem., Int. Ed. Engl. 1994,
33, 2313-2316.
J. Org. Chem, Vol. 70, No. 17, 2005 6617

Ballester et al.
TABLE 1. Stability Constants (M-1) and Effective
Molarities (EM) Measured at 295 K for the
Self-Assembled Macrocycles 1 Determined by UV-Vis
Titrations
macrocycle
Kstability (M-1
2 × 108 a
)
EM (M)c
2
6
3a + 4a S 1aa
3b + 4a S 1ba
3a + 4b S 1ab
3b + 4b S 1bb
(1.3 ( 0.2) × 107 b
(8.7 ( 0.8) × 106 b
(1.1 ( 0.6) × 105 b
(2.9 ( 0.4) × 105 b
3
1
0.02
0.04
a Value taken from ref 5 and determined in CH2Cl2 solution.
b Determined in CHCl3 solution. c EM ) Kstability/Km2. Calculated
using the Km value determined in the corresponding solvent.
showed that a simple 1:1 complex was formed and from
fitting the data to the theoretical binding isotherm, the
stability constants were determined.10 Effective molari-
ties (EM) or chelation factors are useful for quantifying
cooperativity in intramolecular interactions and host-
guest complementarity in divalent binding. The EM for
these systems can be quantified using the following
relationship Kstability ) Km EM.11 A value of Km ) 5.6 ×
2
103 M-1 (intrinsic binding constant for the zinc-pyridine
interaction) was determined previously for a reference
complex of a monomeric zinc porphyrin with a similar
monopyridine ligand in methylene chloride.5 We deter-
mined that the association constants for the same refer-
FIGURE 2. Minimized structures (MMFF94 force field) of
1aa•8 and 2 showing the different geometries of the dimer.
Nonpolar hydrogen atoms and pentyl substituents have been
omitted for clarity.
ence complex in chloroform is Km ) 2.2 × 103 M-1
.
The stability constants and EM values for each com-
plex are summarized in Table 1. To take solvent effects
into account, the EM values are calculated using the
binding constant of the reference complex (Km) measured
in the same solvent where the titration was performed.
Thus, the relative stabilities of the macrocycles are based
on the reported EM values. When these titrations were
studies12 show that the formation of macrocycle 1 re-
quires that the zinc bisporphyrin component has a syn
arrangement of the amides. This conformation is favored
in the 2,6-carboxamidopyridine linker, placing the two
porphyrin rings in an almost perpendicular orientation.
This linker has internal hydrogen bonds which reduces
the 120° angle expected for meta-substituted arenes to
approximately 96° (Figure 2).5,12
In contrast, the bispyridyl component can span the gap
between the two porphyrins equally well regardless of
the nature of the linker.13 The most likely explanation
for the difference between the EM values of 2 and 1aa is
due to a geometric mismatch in the heteromeric system.
Macrocycle 2 forms a rectangular structure with four
right-angled corners and two sets of parallel sides of
equal length. The X-ray crystal structure shows that 2
is essentially strain-free.4 In 1aa, the parallel sides are
not of equal length, and the energy-minimized structure
of 1aa reveals a quadrilateral geometry that requires
distortion of the 2,6-dicarboxamidopyridine linkers away
from the ideal 90° geometry. If right angles are main-
tained for the two zinc-pyridine interactions, the bispor-
phyrin unit needs to reduce the 2,6-dicarboxamidopyri-
dine angle and the bispyridyl component needs to open
1
repeated at a concentration of 10-3 M using H NMR
spectroscopy, the signals due to the R-pyridine proton (∆δ
) - 6.2 ppm) and the pyridine methyl group (∆δ ) -
1.3 ppm) were highly upfield shifted on complexation,
indicating that the pyridine sits over the face of the
porphyrin ring. Analogous upfield shifts were also ob-
served for these protons when the other three complexes
1ab, 1ba, and 1bb were formed. A ROESY experiment
on 1aa revealed intermolecular contacts between the
pyridine methyl protons and the aromatic protons of the
meso-phenyl substituents on the porphyrin ring (see
Figure 1). This result is consistent with the macrocyclic
structure proposed for 1aa.
The data in Table 1 indicate that all macrocycles 1 are
assembled quantitatively at mM concentrations. In all
cases, the complementarity between the bispyridine and
bisporphyrin units, quantified by the EM value, is
inferior to that observed between the two aminoporphyrin
units that self-coordinate in macrocycle 2. When the 2,6-
dicarboxamidopyridine linker is present in the bispor-
phyrin component (4a), the most stable macrocycle is
obtained. However, the nature of the linker in the
bispyridyl component has little or no effect on the
stability of the macrocycle. Simple molecular modeling
(12) The MMFF94 was used as implemented in Macromodel 6.0,
except that the charges on the Zn atoms and N atoms coordinated to
it were modified. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.;
Liskamp, R.; Lipton, M.; Caulfield, C.; Chang, G.; Hendrickson, T.;
Still, W. C. J. Comput. Chem. 1990, 11, 440.
(13) Differences in the conformation adopted by the bisporphyrins
4a and 4b may cause changes in the geometry of the zinc porphyrin-
pyridine interaction contributing also to the observed difference in
stability. Gomila, R. M.; Quin˜onero, D.; Rotger, C.; Garau, C.; Frontera,
A.; Ballester, P.; Costa, A.; Deya`, P. M. Org. Lett. 2002, 4, 399-401.
(10) SPECFIT, ver 3.0, from Spectra Software Associates. Gampp,
H.; Maeder, M.; Meyer, C. J.; Zuberbu¨hler, A. D. Talanta 1985, 32,
95-101.
(11) Anderson, H. L.; Anderson, S.; Sanders, J. K. M. J. Chem. Soc.,
Perkin Trans. 1 1995, 2231-2245.
6618 J. Org. Chem., Vol. 70, No. 17, 2005

Properties of Heterodimeric Porphyrin Macrocycles
FIGURE 4. Left: downfield regions of 1H NMR spectra in
CDCl3 at 253 K (a) [1aa] ) 1 mM; (b) [1aa] ) 1 mM and [9]
) 4 mM; (c) [1aa] ) 1 mM and [10] ) 4 mM. Right: Expansion
of the EXSY experiment at 253 K; [1aa] ) 1 mM, [4a]free
0.2 mM, and [10] ) 4 mM.
)
FIGURE 3. Structures of the amides used in the titrations
and the VT NMR experiments.
TABLE 2. Association Constants (Ka ( 20%) Determined
by 1H NMR Titrations in CDCl3 of Macrocycles 1 with
Diamide 8 at 298 K
ously for covalent tetralactam macrocycles.16 The sub-
stitution of the 2,6-carboxamidopyridine linker with a
closely related benzene linker has a relatively small effect
on the binding constants. The signal due to the aromatic
protons of 8 shows an upfield complexation induced
change in chemical shift in all of the complexes (esti-
mated ∆δmax ) -1.0 to -1.5 ppm) which suggests that
the mode of binding involves the formation of a threaded
H-bonded complex (Figure 2).17
macrocycle
Ka (M-1
1aa
160
1ab
550
1ba
260
1bb
320
)
this angle slightly to achieve a ditopic interaction (Figure
2).14
Next, we set out to investigate the binding properties
of macrocycles 1 using 1H NMR titration with N,N,N′,N′-
tetraethyl terephthalamide 8 (Figure 3). This compound
can form up to four hydrogen bonds with the amide
groups directed into the cavity of 1 and has been used
previously as a guest with similar tetralactam recep-
tors.3,4,15 For the titration of 1aa, the NH signals of the
bisporphyrin and bispyridine components (see Figure S4)
shifted downfield 0.16 and 0.44 ppm, respectively, upon
addition of 4.6 equiv of 8, indicative of the formation of
hydrogen bonds between the amide NHs of 1aa and the
carbonyl oxygens of 8. For 1ba, the same amount of 8
caused a downfield shift of the NHs of the bisporphyrin
and bispyridine components of 0.14 and 0.28 ppm,
respectively. In the case of 1ab, it was only possible to
follow and measure the downfield shift of the NHs of the
bispyridine component which was 0.27 ppm when 4.6
equiv of 8 are added. The NH signal of the bisporphyrin
unit of 1ab was too broad to be used for the titration
experiment. For 1bb, both types of NH signals were
either too broad or overlapped with other proton signals
in the aromatic region. In this case, we followed the
downfield shift experienced by the proton signal assigned
to the methyl ether of the bisporphyrin component upon
addition of 8 (0.19 ppm when 4.6 equiv of 8 are added).
The titration data obtained for the chemical shift change
of the above-mentioned signals were fitted to a 1:1
binding model, and the calculated association constants
are shown in Table 2. When more than one signal was
used to calculate the association constant, the value
reported is an average value.
Other evidence for the formation of pseudorotaxane
and rotaxane structures with these macrocycles came
from dynamic 1H NMR studies carried out with diamides
9 and 10. While macrocycle 1aa forms complexes with
similar structure and stability with 9 and 10 (Ka ) 184
and 160 M-1, respectively), their dynamic properties are
quite different.
On cooling a 1 mM chloroform solution of 1aa contain-
ing 4 equivalents of diamide 10 to 253 K, the signal due
to the amides of the bisporphyrin component (4a) split
into two signals. The rest of the signals of 1aa become
broad upon cooling and no other splitting of signals could
be clearly observed. One signal had a chemical shift very
close to that of pure 1aa,18 and the other was downfield
shifted, indicative of hydrogen bonding. We conclude that
the splitting is due to slow exchange between free and
bound 1aa. Under identical conditions with a mixture of
1aa and 9, no splitting was observed for the same signal.
In this case, only broadening of the signal was observed,
and the chemical shift was intermediate between that
of the free and bound signals of 1aa (Figure 4). This is
due to fast exchange between the free and bound species
in this case. Although the two complexes, 1aa•9 and
1aa•10, have similar structures and thermodynamic
stabilities, they clearly have quite different dynamic
properties. The most likely explanation is that complex
1aa‚9 has a pseudorotaxane structure, while 1aa‚10 is
a rotaxane. Thus, while the exit or entry of 9 into the
cavity of 1aa does not require disruption of the macro-
cycle, the rate-determining step for exchange of 10
involves opening the macrocycle by breaking zinc-pyri-
dine interactions, because the bulky stopper groups do
not fit into the cavity (Figure 5).
The binding affinities are all of the same order of
magnitude and are comparable to those reported previ-
(16) Chang, S.-Y.; Kim, H. S.; Chang, K.-J.; Jeong, K.-S. Org. Lett.
2004, 6, 181-184.
(17) Intermolecular NOEs observed in a ROESY experiment on 1aa‚
8 are also in agreement with a threading binding mode.
(18) The small chemical shift difference between the bisporphyrin
amide signals of pure 1aa compared with free 1aa in the presence of
10 is probably due to some aggregation at low temperatures.
(14) Bisporhyrin 4a is a flexible molecule, for example, about the
porphyrin meso bond and so a range of strained conformations are
possible for macrocycle 1aa. The structure depicted in Figure 2 is only
an energy minimum.
(15) Chang, S.-Y.; Um, M.-C.; Uh, H.; Jang, H.-Y.; Jeong, K.-S.
Chem. Commun. 2003, 2026-2027.
J. Org. Chem, Vol. 70, No. 17, 2005 6619

Ballester et al.
temperature changes, the second-order rate constant for
macrocycle formation can be estimated kformation ≈ 2.9 ×
107 M-1 s-1 (Kstability ) kformation/kmacrocycle). This implies that
the formation of the highly stable macrocycle 1aa is close
to diffusion controlled, as expected for a system where
there are no significant kinetic barriers. The binding sites
of 3a and 4a are sterically accessible, and desolvation
during complexation requires only a small activation
energy.
Figure 5b illustrates the processes that are important
in the case of the exchange of rotaxane 1aa‚10 and its
components. Here, the key intermediate is I where only
one of the zinc-pyridine interactions is broken. This
species can close again to regenerate the rotaxane
1aa•10, and this process must be very rapid setting up
an equilibrium between the rotaxane and I. We can
estimate the equilibrium constant Kopen from the EM for
cyclization of 1aa (EM ) 3 M, T ) 295 K, Table 1) and
the binding constant of the reference complex (Km ) 2.2
× 103 M-1). Thus, Kopen ) 2/EM Km ) 3 × 10-4. The factor
of 2 accounts for the probability of opening one of two
zinc-pyridine interactions. Intermediate I could subse-
quently lose either the guest 10 with rate constant k1 or
break the second zinc-pyridine interaction to lose the
bisporphyrin 4a with rate constant k2. The rate constants,
k1 and k2, for these processes are determined by the
strengths of the interactions to be broken. Thus, if we
assume a diffusion-controlled rate for the formation
process of intermediate I from its components II and 10
FIGURE 5. Processes involved in the exchange of macrocycle
1aa and its components (a) and rotaxane 1aa‚10, macrocycle
1aa, and free bisporphyrin 4a (b). The key intermediate I is
highlighted.
The rotaxane exchange mechanism can be probed by
comparing the rate of exchange of free 1aa with the
rotaxane 1aa‚10 and the rate of exchange of free 4a with
the macrocycle 1aa.19 Both exchange processes are slow
on the NMR time scale at 253 K, and the rate constants
were therefore determined simultaneously using two-
dimensional exchange spectroscopy (2D-EXSY).20 The
rate constants were determined using the diagonal and
cross-peak areas for the NH signals of 1aa (free and
bound) and the free bisporphyrin 4a.21 The rate constant
krotaxane for the interconversion of 1aa‚10 and free 1aa is
13 ((1) s-1, corresponding to an activation energy (∆Gq)
of 13.4 kcal mol-1. The rate constant kmacrocycle for ex-
change between the bisporphyrin in the macrocycle 1aa
and free bisporphyrin 4a is 2.2 ((0.3) s-1, giving an
activation energy (∆Gq) of 14.4 kcal mol-1. No cross-peaks
were observed between the amide of the rotaxane 1aa‚
10 and the free bisporphyrin 4a. Consequently, at this
temperature, interconversion between the rotaxane 1aa‚
10 and its components macrocycle 1aa and diamide 10
is faster than any of the possible exchanges involving the
free bisporphyrin 4a. Why is this type of dynamic
behavior observed in this system?
or III and 4a (from above, kformation ) 2.9 × 107 M-1 s-1
)
and estimate association constants for each process on
the basis of the interactions involved (Ka ) 160 M-1 for
the formation of I from II and 10, see Table 2, and 2Km
) 5 × 103 M-1 for the formation of I from III and 10), we
can estimate the first-order rate constants for the two
competitive processes as k1 ) kformation/Ka ) 18 × 104 s-1
,
and k2 ) kformation/2Km ) 5.8 × 103 s-1 (the statistical factor
of 2 relates to the asymmetry of I). The rate of loss of 4a
from I is an order of magnitude smaller than the rate
constant of loss of 10, because the strengths of the
interactions holding the different units in place are very
different. The partially open macrocycle (II) formed by
the loss of 10 can now either lose 4a (k4 in Figure 5b) or
close to regenerate 1aa (k3 in Figure 5b). No interactions
must be broken in the latter process, so k3 . k4. This
dynamic scheme explains why there is no direct exchange
of 4a with the rotaxane 1aa•10 in the EXSY spectrum
in Figure 4. The analysis above allows us to estimate the
expected rate constant for the exchange of the rotaxane
with the macrocycle, krotaxane ) Kopen k1 ) 54 s-1. This
The exchange of the macrocycle 1aa with its compo-
nent parts, 3a and 4a, can be considered as a straight-
forward dissociation process breaking two zinc-pyridine
interactions (Figure 5a). For this dissociation, the first-
order rate constant for fully breaking the macrocycle
(kmacrocycle) was determined using the EXSY experiment
as kmacrocycle ) 2.2 ((0.3) s-1 (T ) 253 K). Using the
macrocycle stability constant (1.3 × 107 M-1, T ) 295 K,
Table 1) and assuming that it is not sensitive to small
value is higher than the observed value of 13 s-1
,
indicating that there is an additional 0.7 kcal mol-1 strain
energy in the transition state for escape of 10.
Conclusion
In conclusion, we have demonstrated the quantitative
self-assembly of a series of heterodimeric tetralactam
macrocycles at mM concentrations in CDCl3 using two
kinetically labile zinc porphyrin-pyridine interactions. We
have studied the effect of preorganization of the linker
on the stability of the macrocycles. The macrocycles
feature a cavity with inwardly directed hydrogen bond
donors that make them molecular hosts for terephthala-
(19) It has been proposed for similar systems that the exchange
between the components of the rotaxane occurs by breaking just one
coordinative bond (see refs 3 and 4), but no determination of the rate
constant for the exchange of the subunits of the self-assembled
macrocycle has been carried out.
(20) Pons, M.; Millet, O. Prog. Nucl. Magn. Reson. Spectrosc. 2001,
38, 267-324. Perrin, C. L.; Dwyer, T. J. Chem. Rev. 1990, 90 (6), 935-
67.
(21) D2DNMR (PC version). Abel, E. W.; Coston, T. P. J.; Orrell, K.
G.; Sik, V.; Stephenson, D. J. Magn. Reson. 1986, 70 (1), 34-53.
6620 J. Org. Chem., Vol. 70, No. 17, 2005

Properties of Heterodimeric Porphyrin Macrocycles
164.19, 160.49, 149.55, 141.70, 135.62, 129.68, 117.98, 116.86,
50.04, 15.47; HRMS m/z (ESI) calcd for C23H25N4O3 [M+H+]
405.1927; found, 405.1939.
mide and adipamide guests. We have used one of the self-
assembled macrocycles for the construction of a [2]-
rotaxane. The dissociation rate for the interconversion
between the components of the [2]rotaxane (macrocycle
and linear bisamide) is faster than complete dissociation
of the macrocycle in both the free and bound states.
Free Base Bisporphyrin (4aH2). A solution of aminopo-
rphyrin 7 (100 mg, 0.12 mmol), freshly distilled dry triethy-
lamine (40 µL, 0.28 mmol), and a catalytic amount of 4-(dim-
ethylamino)pyridine (DMAP) in dry CH2Cl2 (10 mL) was cooled
at 0 °C in an ice-water bath, and diacid chloride 5a (13 mg,
0.05 mmol) dissolved in a minimum amount of dry CH2Cl2 was
added in one portion. After stirring for 3 h at room temperature
under argon atmosphere, the mixture was heated at reflux for
2 h and was cooled again at room temperature. The organic
layer was washed with 0.1 N HCl, saturated NaHCO3, and
brine; dried over Na2SO4; filtered; and concentrated in vacuo
to yield a solid residue. The product was separated from the
unreacted porphyrin by flash chromatography of the solid
residue on silica gel eluting with CH2Cl2:hexanes (2:1) to
recover first the unreacted aminoporphyrin 7 and affording
next the free base bisporphyrin 4aH2 as a purple solid (50 mg,
46%): 1H NMR (300 MHz, CDCl3) δ 10.04 (s, 2H), 8.89 (m,
16H), 8.33 (d, J ) 8.5 Hz, 4H), 8.25 (d, J ) 8.5 Hz, 4H), 8.19
(s, 2H), 8.10 (d, J ) 7.9 Hz, 12H), 7.54 (m, 12H), 4.43 (q, J )
6.7 Hz, 2H), 2.94 (m, 12H), 1.89 (m, 12H), 1.62 (t, J ) 6.7 Hz,
3H), 1.51 (m, 24H), 1.02 (m, 18H), -2.77 (s, 4H).
Free Base Bisporphyrin (4bH2). Aminoporphyrin 7 (220
mg, 0.26 mmol), freshly distilled dry triethylamine (60 µL, 0.4
mmol), and a catalytic amount of 4-(dimethylamino)pyridine
(DMAP) were dissolved in dry CH2Cl2 (30 mL), and the
resulting solution was cooled at 0 °C in an ice-water bath.
Diacid chloride 5b (31 mg, 0.13 mmol) was added in one
portion to the previously cooled reaction mixture. After stirring
for 4 h at room temperature under argon atmosphere, the
organic layer was washed with saturated NaHCO3 and brine,
dried over Na2SO4, filtered, and concentrated in vacuo to yield
a solid residue. The product was separated from the unreacted
porphyrin by flash chromatography of the residue on silica gel
eluting with CH2Cl2:THF (99:1) to recover first the aminopo-
rphyrin 7 (39 mg) and affording next the free base bisporphyrin
4bH2 as a purple solid (100 mg, 42%):1H NMR (300 MHz,
CDCl3): δ 8.89 (s, 8H), 8.87 (s, 8H) 8.35 (s, 2H), 8.29 (d, 4H,
J ) 8.2 Hz), 8.26 (s, 2H), 8.12 (d, 4H, J ) 8.2 Hz), 8.11 (d,
12H, J ) 7.8 Hz), 7.84 (s, 2H), 7.55 (d, 12H, J ) 7.8 Hz), 2.94
(t, 12H, J ) 7.5 Hz), 1.90 (m, 12H), 1.52 (m, 24H), 1.02 (t,
18H, J ) 7 Hz), -2.75 (s, 4H).
General Procedure for the Preparation of the Zinc
Bisporphyrins 4a-b. The free base bisporphyrin (100 mg,
0.05 mmol) was dissolved in CH2Cl2-CH3OH (3:1, 30 mL), and
zinc acetate (180 mg, 0.98 mmol) was added. The reaction
mixture was protected from light and was stirred at room
temperature for 1 h. After removal of the solvents under
reduced pressure, the product was purified by column chro-
matography on basic alumina eluting with CH2Cl2:CH3OH (99:
1). The product was recrystallized from CH2Cl2:CH3OH yield-
ing the zinc bisporphyrin 4.
Zn Bisporphyrin (4a). (87 mg, 82%); mp > 220 °C dec; IR
(KBr, cm-1) 3414, 2924,1618, 1525, 1384, 1338, 1205, 998, 796,
719, 621;1H NMR (300 MHz, CDCl3) δ 10.03 (s, 2H), 8.99 (m,
16H), 8.32 (d, 4H, J ) 8.8 Hz), 8.24 (d, 4H, J ) 8.8 Hz), 8.16
(s, 2H), 8.10 (d, 12H, J ) 8.0 Hz), 7.53 (dd, 12H, J ) 8.0, 2.5
Hz), 4.42 (q, 2H, J ) 6.9 Hz), 2.94 (m, 12H), 1.91 (m, 12H),
1.62 (t, 3H, J ) 6.9 Hz), 1.52 (m, 24H), 1.02 (m, 18H); HRMS
m/z (ESI) calcd for C127H123N11O3Zn2 1977.8393; found,
1977.8372.
Experimental Section
Titrations and Data Analysis. 1H NMR and UV-vis
titrations were performed by adding solutions of the guest (8,
9, 10 for the 1H NMR titrations or 3a-b for the UV-vis
titrations) to a solution of the host (macrocyle 1 for the 1H
NMR titrations or zinc bisporphyrin 4a-b for the UV-vis
titrations) in either a 5-mm NMR tube or a 1-cm path cuvette
using microliter syringes. In both types of titration experi-
ments, the host was present in the guest solution at the same
concentration as that in the NMR tube or cuvette to avoid
dilution effects. Deacidified chloroform and deacidified deutero-
chloroform were used as solvents for the UV-vis and 1H NMR
titrations, respectively. UV-vis spectrophotometric titrations
were analyzed by fitting the whole series of spectra at 1-nm
intervals using the software SPECFIT 3.0 from Spectrum
Software Associates, PMB 361, 197M Boston Post Road West,
Marlborough, MA 01752 (e-mail: SpecSoft@compuserve.com),
which uses a global system with expanded factor analysis and
Marquardt least-squares minimization to obtain globally
optimized parameters. The titration data obtained using the
1H NMR titration method were analyzed by fitting to a simple
1:1 binding isotherm using a nonlinear curve fitting program
developed by one of us (C.A.H.).
Synthesis. Compounds 5a,6 5b,7 6,8 and 79 were prepared
as described previously. Compound 8 was prepared by coupling
reaction of the corresponding commercially available acid
chloride with diethylamine.
N,N′-Bis(3,5-dimethyl-4-pyridinyl)-4-ethoxy-2,6-py-
ridinedicarboxamide (3a). To a CH2Cl2 (10 mL) solution of
4-ethoxy-2,6-pyridinedicarbonyl dichloride (5a) (660 mg, 2.3
mmol) cooled at 0 °C were added a solution of 4-amino-3,5-
lutidine (6) (607 mg, 5 mmol) in CH2Cl2 (10 mL) and N,N-
diisopropylethylamine (1 mL, 5.8 mmol). The solution was
stirred under argon atmosphere at room temperature for 2 h
and was subsequently heated at reflux for 2 h. The reaction
mixture was diluted with CH2Cl2 (10 mL), washed twice with
saturated NaHCO3 solution and once with brine, dried over
anhydrous MgSO4, filtered, and evaporated in vacuo. The solid
residue was washed with diethyl ether and was recrystallized
in CH2Cl2/EtOAc solvent to yield 3a as a white solid (190 mg,
56%): mp 238-240 °C; IR (KBr, cm-1) 3294, 1669, 1587, 1506,
1
1342, 1166, 1044, 685; H NMR (300 MHz, CDCl3) δ 9.60 (s,
NH, 2H), 8.30 (s, 4H), 7.97 (s, 2H), 4.31 (q, J ) 6.9 Hz, 2H),
2.25 (s, 12H), 1.50 (t, J ) 6.9 Hz, 3H); 13C NMR (75 MHz,
CDCl3) δ168.7, 161.5, 150.6, 149.7, 142.1, 130.3, 112.7, 64.5,
15.9, 14.7; HRMS m/z (ESI) calcd for C23H25N5O3 [M + H+]
420.2036; found, 420.2047; calcd for C23H25N5O3Na [M + Na+]
442.1853; found, 442.1866.
N,N′-Bis(3,5-dimethyl-4-pyridinyl)-5-methoxy-1,3-ben-
zenedicarboxamide (3b). To a CH2Cl2 (6 mL) solution of
4-methoxy-1,3-benzenedicarbonyl dichloride (5b) (187 mg, 0.8
mmol) cooled at 0 °C were added 4-amino-3,5-lutidine (6) (206
mg, 1.7 mmol) and N,N-diisopropylethylamine (365 µL, 2.1
mmol). The reaction mixture was stirred under argon atmo-
sphere for 2 h at room temperature followed by heating at
reflux for an additional 2 h. The reaction mixture was filtered,
and the filtrate was washed twice with saturated NaHCO3
solution and once with brine, dried over anhydrous Na2SO4,
filtered, and evaporated in vacuo. The solid residue was
triturated with diethyl ether affording 3b as a white solid (70
mg, 22%). mp 229-234 °C; IR (KBr, cm-1) 3415, 3232, 1653,
Zn Bisporphyrin (4b). (96 mg, 94%); mp 295-300 °C; IR
(KBr, cm-1) 3415, 2926, 1617, 1526, 1384, 1339, 1207, 1000,
1
797, 720; H NMR (300 MHz, CDCl3) δ 8.99 (s, 8H), 8.98 (d,
4H, J ) 5.3 Hz), 8.96 (d, 4H, J ) 5.3 Hz), 8.26 (d, 4H, J ) 8.5
Hz), 8.19 (s, 2H), 8.12 (d, 4H, J ) 8.2 Hz), 8.11 (d, 8H, J ) 8.2
Hz), 7.94 (d, 4H, J ) 8.5 Hz), 7.82 (s, 1H), 7.57 (s, 2H), 7.56
(d, 4H, J ) 8.2 Hz), 7.53 (d, 8H, J ) 8.2 Hz), 3.96 (s, 3H), 2.95
1
1592, 1514, 1383, 1344, 1162, 1055, 879, 700; H NMR (300
MHz, CDCl3) δ 8.36 (s, 4H), 8.09 (s, 1H), 7.91 (s, 2H), 7.67 (s,
2H), 3.95 (s, 3H), 2.25 (s, 12H); 13C NMR (75 MHz, CDCl3) δ
J. Org. Chem, Vol. 70, No. 17, 2005 6621

Ballester et al.
(m, 12H), 1.92 (m, 12H), 1.50 (m, 24H), 1,02 (m, 18H); HRMS
m/z (ESI) calcd for C127H122N10O3Zn2Na 1989.8187; found,
1989.8251.
calculated integrations were processed using the program
D2DNMR21 considering a three-site exchange to give the
chemical exchange rate constants kmacrocycle and krotaxane
.
N,N′-Dioctyl-hexanediamide (9). A suspension of adipic
acid (0.3 g, 2 mmol) in thionyl chloride (4 mL) containing a
catalytic amount of triphenyl phosphine was heated at reflux
until a clear solution was obtained (∼2 h). The excess of thionyl
chloride was removed and the residue was suspended in dry
CH2Cl2 (10 mL). To this solution, octylamine (680 µL, 4.1
mmol) and triethylamine (600 µL, 4.3 mmol) were added. The
reaction mixture was stirred overnight; washed with aqueous
HCl (5%), NaOH solution (10%), and brine; dried over Na2-
SO4; filtered; and concentrated. The residue was purified by
column chromatography (CH2Cl2/MeOH 5%) to give 9 as a
white solid (320 mg, 43%): IR (KBr, cm-1) 3313, 1632; 1H NMR
(300 MHz, CDCl3) δ: 5.62 (s, 2H), 3.24 (m, 4H), 2.19 (m, 4H),
1.65 (m, 4H), 1.49 (t, 4H, J ) 5.8 Hz), 1.27 (m, 20H), 0.88 (m,
6H). Anal. Calcd (%) for C22H44N2O2 (368.59): C 71.69, H 12.03,
N 7.60. Found: C 71.42, H 12.20, N 7.62.
Acknowledgment. We thank Direccio´n General de
Investigacio´n, Ministerio de Ciencia y Tecnolog´ıa
(BQU2002-04651) for grant support. P.B. and A.I.O. also
thank ICIQ Foundation for financial support. A. F.
thanks MCyT for a Ramo´n y Cajal contract. Thanks are
also due to Professor Kyu-Sung Jeong for providing us
with a sample of compound 10.
Supporting Information Available: UV-vis and 1H
NMR spectra recorded during titrations of 4a with 3a.
Aromatic region of the 1H NMR spectra recorded during
titration of 1aa with 8. Selected expansions of the ROESY
experiments performed on (1) the self-assembled macrocycle
1
1aa and on (2) the pseudorataxane complex 1aa‚8. H NMR
2D-EXSY Experiment. The 2D-EXSY experiment was
recorded at 253 K on a 300 MHz spectrometer using an inverse
probe. The mixing time was 200 ms. Exchange cross-peaks
were integrated using XWINNMR Bruker software. The
spectra of compounds 3a and 4a. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO0505369
6622 J. Org. Chem., Vol. 70, No. 17, 2005