Molecular Imprinting Inside Dendrimers

Article abstract of DOI:10.1021/ja0357240

Synthetic hosts capable of binding porphyrins have been produced by a mixed-covalent-noncovalent imprinting process wherein a single binding site is created within cross-linked dendrimers. Two synthetic hosts were prepared, using as templates 5,10,15,20-tetrakis(4-hydroxyphenyl)porphyrin and 5,10,15,20-tetrakis(3,5-dihydroxyphenyl)porphyrin. These two templates were esterified with, respectively, fourth- and third-generation Frechet-type dendrons containing homoallyl end-groups. The resulting tetra-and octadendron macromolecules underwent the ring-closing metathesis reaction using Grubbs' Type I catalyst, RuCl₂(P(C₆H₅) ₃)₂(CHCH₂C₆H₅), to give extensive interdendron cross-linking. Hydrolytic removal of the porphyrin cores afforded imprinted hosts whose ability to bind porphyrins with various peripheral substituents was investigated by UV-visible spectrophotometric titrations and size exclusion chromatography. The results indicate a high yield of imprinted sites that show high selectivity for binding of porphyrins capable of making at least four hydrogen bonds, but only a moderate degree of shape selectivity.

Full text of DOI:10.1021/ja0357240

Published on Web 10/10/2003
Molecular Imprinting Inside Dendrimers
Steven C. Zimmerman,* Ilya Zharov, Michael S. Wendland, Neal A. Rakow, and
Kenneth S. Suslick*
Contribution from the Department of Chemistry and Beckman Institute for AdVanced Science
and Technology, UniVersity of Illinois at Urbana-Champaign, Urbana, Illinois 61801
Received April 21, 2003; E-mail: sczimmer@uiuc.edu
Abstract: Synthetic hosts capable of binding porphyrins have been produced by a mixed-covalent-
noncovalent imprinting process wherein a single binding site is created within cross-linked dendrimers.
Two synthetic hosts were prepared, using as templates 5,10,15,20-tetrakis(4-hydroxyphenyl)porphyrin and
5,10,15,20-tetrakis(3,5-dihydroxyphenyl)porphyrin. These two templates were esterified with, respectively,
fourth- and third-generation Fre´chet-type dendrons containing homoallyl end-groups. The resulting tetra-
and octadendron macromolecules underwent the ring-closing metathesis reaction using Grubbs’ Type I
catalyst, RuCl2(P(C6H5)3)2(CHCH2C6H5), to give extensive interdendron cross-linking. Hydrolytic removal
of the porphyrin cores afforded imprinted hosts whose ability to bind porphyrins with various peripheral
substituents was investigated by UV-visible spectrophotometric titrations and size exclusion chromatog-
raphy. The results indicate a high yield of imprinted sites that show high selectivity for binding of porphyrins
capable of making at least four hydrogen bonds, but only a moderate degree of shape selectivity.
Introduction
have been among the most extensively studied host-guest
systems. MIPs have several drawbacks, however, including
Host-guest chemistry has emerged as a central paradigm
within organic chemistry.1 The design and synthesis of diverse
host molecules that selectively and tightly complex many
different classes of guest molecules have been notably success-
ful. As effective as this approach has been, especially for small
molecule hosts, the requirement to prepare hosts bond by bond
through multistep synthetic routes has limited their widespread
application. Furthermore, each new target guest typically
requires an entirely new host design and development program.
Two strategies that have the potential to significantly extend
the host-guest approach involve molding an organic receptor
around the guest “template”. The first, using molecularly
imprinted polymers (MIP), was initially described in Wulff’s
seminal 1972 report,2 in which a matrix was polymerized around
the template molecules, followed by removal of the template;
this leaves host cavities that, ideally, retain a shape and
functional group complementarity to the guest-template. This
early synthesis of a MIP is referred to as the covalent approach
because the template is reversibly linked to the matrix by
covalent bonds.3 A noncovalent approach, in which one or more
monomers complex the template, was pioneered by Mosbach
and co-workers and is now the most commonly used method
of MIP synthesis.4 Subsequently, mixed covalent-noncovalent
methods were developed5 as well as numerous related ap-
proaches.6,7 Indeed, molecularly imprinted polymers (MIPs)
incomplete template removal and slow mass transfer;6,7-9 their
practical application is also severely limited by the heterogeneity
of the binding sites for which a broad range of affinities are
observed.6,10
A second strategy for rapid host construction has emerged
more recently. It uses a dynamic combinatorial library (DCL)
of hosts, in which one or more members are bound to and
stabilized by the guest molecule.11-13 The molding process in
the DCL approach is different in two ways. First, the molding
uses reversible reactions so that ineffective hosts may be
sacrificed in favor of superior ones. The DCL approach is further
distinguished in that the molded receptors each contain a single
binding site so that individual receptors or classes of receptors
can be separated, characterized, and studied in solution.
We recently described a “monomolecular imprinting” ap-
proach, which contains elements of both the DCL and the
mixed-covalent-noncovalent imprinting approaches and pro-
(5) Whitcombe, M. J.; Rodriguez, M. E.; Villar, P.; Vulfson, E. N. J. Am.
Chem. Soc. 1995, 117, 7105-7111.
(6) For books and recent literature reviews of MIPs, see: Wu¨lff, G. Chem.
ReV. 2002, 102, 1-27. Ye, L.; Mosbach, K. J. Inclusion Phenom.
Macrocycl. Chem. 2001, 41, 107-113. Haupt, K.; Mosbach, K. Chem. ReV.
2000, 100, 2495-2504. Molecularly Imprinted Polymers.; Sellergren, B.,
Ed.; Elsevier: Amsterdam, 2001. Bartsch, R. A., Maeda, M., Eds. Molecular
and Ionic Recognition with Imprinted Polymers; ACS Symposium Series
703; Oxford Press: New York, 1998.
(7) Wulff, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 1812-1832.
(8) Katz, A.; Davis, M. E. Macromolecules 1999, 32, 4113-4121.
(9) Sellergren, B.; Shea, K. J. J. Chromatogr. 1993, 635, 31-49.
(10) Vlatakis, G.; Andersson, L. I.; Mu¨ller, R.; Mosbach, K. Nature 1993, 361,
645-647.
(1) ComprehensiVe Supramolecular Chemistry; Lehn, J.-M., Series Ed.; Elsevier
Science Ltd.: New York, 1996; Vols. 1-2. Lehn, J.-M. Angew. Chem.,
Int. Ed. Engl. 1988, 27, 89-112. Cram, D. J. Angew. Chem., Int. Ed. Engl.
1988, 27, 1009-1020.
(2) Wulff, G.; Sarhan, A. Angew. Chem., Int. Ed. Engl. 1972, 11, 341.
(3) Shea, K. J. Trends Polym. Sci. 1994, 2, 166-173.
(4) Andersson, L.; Sellergren, B.; Mosbach, K. Tetrahedron Lett. 1984, 5211-
5214.
(11) Cousins, G. R. L.; Poulsen, S.-A.; Sanders, J. K. M. Curr. Opin. Chem.
Biol. 2000, 270, 270-279.
(12) Lehn, J.-M.; Eliseev, A. V. Science 2001, 291, 2331-2332.
(13) Klekota, B.; Miller, B. L. Trends Biotechnol. 1999, 17, 205-209.
9
13504
J. AM. CHEM. SOC. 2003, 125, 13504-13518
10.1021/ja0357240 CCC: $25.00 © 2003 American Chemical Society

Molecular Imprinting Inside Dendrimers
A R T I C L E S
easily accessible.19,20 For example, the Grubbs’ ruthenium
benzylidene catalysts exhibit broad functional group tolerance
and are relatively insensitive to water, oxygen, or other
impurities.20 The commercially available Type I catalyst (2) is
particularly well studied and known to produce medium and
large ring systems,21 although the more recently reported Type
II catalyst has certain advantages in this regard.22 The Type I
catalyst has also been used to covalently capture supramolecular
assemblies.23 Most appealing from the perspective of dynamic
molding was the finding that the pendant alkene groups in 1,2-
polybutadiene underwent extensive ring-closing metathesis
(RCM) reactions and that undesirable ring closings were
reopened, ultimately allowing nearly all adjacent alkenes to
couple.24 To demonstrate this concept of monomolecular
imprinting, we decided to imprint a porphyrin within a Fre´chet-
type dendrimer25 for multiple reasons. First, the dendrimer shell
must be relatively nonpolar to permit hydrogen bond-mediated
recognition in the core.26 The Fre´chet-type phenyl-benzyl ether-
based dendrimers fulfill this requirement, and a range of other
recognition processes have been shown to occur readily within
the interior of these dendrimers.17 Fre´chet-type dendrimers are
also chemically robust, which allows considerable latitude in
the chemistry used to implement the imprinting process outlined
in Figure 1. Furthermore, they possess an intermediate degree
of flexibility in comparison to that of poly(propylene imine) or
phenylacetylene based systems. Most importantly, we reported
that, upon treatment with RCM catalyst 2, dendrimer 1
underwent extensive cross-linking to give 3, which subsequently
afforded 4 upon core removal (Scheme 1).27
Figure 1. Schematic representation of the monomolecular imprinting
process.
duces macromolecular hosts containing a single binding site.14,15
Herein, full details of those studies are described, including
binding kinetics and imprinting experiments using a different
but related template.
Design of a Monomolecular Imprinting System
To make a selective MIP, each template should produce an
effective binding site, and each binding site must originate from
template-mediated imprinting. The approach chosen to test the
basic strategy involves three distinct steps (Figure 1): (1)
covalent attachment of functional dendrons to the template with
cleavable bonds, (2) cross-linking of the dendron end-groups,
and (3) removal of the template. In principle, a single binding
site could be imprinted into virtually any macromolecule. Even
though they often require multistep synthesis, dendrimers were
attractive candidates for these early studies both because their
homogeneity assists in purification and characterization and
because the large number of end-groups should allow a
significant degree of cross-linking.16
To our knowledge, porphyrins have not been used as
templates for the synthesis of MIPs, although their incorporation
into MIPs for recognition and sensing was reported by Takeu-
chi,28 and porphyrins have been used extensively in molecular
recognition, shape selective oxidation, and self-assembly stud-
ies.29 In the current work, polyhydroxylated tetraphenylporphy-
rins were considered excellent candidates as templates because
the multiple hydroxyl groups provide sites for dendron attach-
ment. These, in turn, would produce an imprinted structure
capable of multipoint recognition. Furthermore, the visible
With respect to the cross-linking reaction, a reversible
chemical process was desirable as a way to effect a dynamic
molding that might result in a “best-fit” imprint for the template.
Recent advances in the development of novel catalysts for olefin
metathesis have made reversible, but robust, alkene cross-links
(17) Selected reviews on the supramolecular chemistry of dendrimers: Zeng,
F.; Zimmerman, S. C. Chem. ReV. 1997, 97, 1681-1712. Newkome, G.
R.; He, E.; Moorefield, C. N. Chem. ReV. 1999, 99, 1689-1746. Baars,
M. W. P. L.; Meijer, E. W. Top. Curr. Chem. 2000, 210, 131-182. Smith,
D. K.; Diederich, F. Top. Curr. Chem. 2000, 210, 183-227. Gorman, C.
B.; Smith, J. C. Acc. Chem. Res. 2001, 34, 60-71. Zimmerman, S. C.;
Lawless, L. J. Top. Curr. Chem. 2001, 212, 95-120.
(18) Selected reviews focusing on functionalized dendrimers and their applica-
tions including in catalysis: Kreiter, R.; Kleij, A. W.; Klein Gebbink, R.
J. M.; van Koten, G. Top. Curr. Chem. 2001, 212, 163-199. Bosman, A.
W.; Janssen, H. M.; Meijer, E. W. Chem. ReV. 1999, 99, 1665-1688. Smith,
D. K.; Diederich, F. Chem.-Eur. J. 1998, 4, 1353-1361.
(19) Schrock, R. R. Top. Organomet. Chem. 1998, 1, 1-36. Hoveyda, A. H.;
Schrock, R. R. Chem.-Eur. J. 2001, 7, 945-950. Randall, M. L.; Snapper,
M. L. J. Mol. Catal. A 1998, 133, 29-40. Fu¨rstner, A. Top. Organomet.
Chem. 1998, 1, 37-72.
(20) Grubbs, R. H.; Khosravi, E. Mater. Sci. Technol. 1999, 20, 65-104. Trnka,
T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18-29.
(14) Zimmerman, S. C.; Wendland, M. S.; Rakow, N. A.; Zharov, I.; Suslick,
K. S. Nature 2002, 418, 399-403.
(15) Shinkai and co-workers have pioneered the imprinting of a single binding
site on a single fullerene molecule: Ishi-i, T.; Shinkai, S. Chem. Commun.
1998, 9, 1047-1048. Ishi-I, T.; Iguchi, R.; Shinkai, S. Tetrahedron 1999,
55, 3883-3892.
(16) Selected reviews focusing on the synthesis and properties of dendrimers:
Fre´chet, J. M. J.; Tomalia, D. A. Dendrimers and Other Dendritic Polymers;
Wiley: New York, 2002. Newkome, G. R.; Moorefield, C. N.; Vo¨gtle, F.
Dendrimer and Dendrons: Concepts Syntheses, Applications; VCH:
Weinheim, 2001. Grayson, S. M.; Fre´chet, J. M. J. Chem. ReV. 2001, 101,
3819-3867. Crooks, R. M.; Lemon, B. I., III; Sun, L.; Yeung, L. K.; Zhao,
M. Top. Curr. Chem. 2001, 212, 81-135. van Manen, H.-J.; van Veggel,
F. C. J. M.; Reinhoudt, D. N. Top. Curr. Chem. 2001, 217, 121-162.
Wiesler, U.-M.; Weil, T.; Mu¨llen, K. Top. Curr. Chem. 2001, 212, 1-40.
Balzani, V.; Ceroni, P.; Juris, A.; Venturi, M.; Campagna, S.; Puntoriero,
F.; Serroni, S. Coord. Chem. ReV. 2001, 219-221, 545-572. Rockendorf,
N.; Lindhorst, T. K. Top. Curr. Chem. 2001, 217, 201-238. Frey, H.;
Schlenk, C. Top. Curr. Chem. 2000, 210, 69-129. Hawker, C. J.; Piotti,
M. ACS Symp. Ser. 2000, 755, 107-118. Majoral, J.-P.; Caminade, A.-M.
Chem. ReV. 1999, 99, 845-880. Janssen, H. M.; Meijer, E. W. Mater. Sci.
Technol. 1999, 20, 403-458. Chow, H.-F.; Mong, T. K.-K.; Wan, C.-W.;
Wang, Z.-Y. In AdV. Dendritic Macromol. 1971, 4, 107-133. Schluter,
A.-D. Top. Curr. Chem. 1998, 197, 165-191. Seebach, D.; Rheiner, P.
B.; Greiveldinger, G.; Butz, T.; Sellner, H. Top. Curr. Chem. 1998, 197,
125-164. Archut, A.; Vo¨gtle, F. Chem. Soc. ReV. 1998, 27, 233-240.
Moore, J. S. Acc. Chem. Res. 1997, 30, 402-413. Constable, E. C. Chem.
Commun. 1997, 12, 1073-1080. Hawker, C. J.; Wooley, K. L. AdV.
Dendritic Macromol. 1995, 2, 1-39.
(21) Delgado, M.; Martin, J. D. J. Org. Chem. 1999, 64, 4798-4816.
(22) Lee, C. W.; Grubbs, R. H. J. Org. Chem. 2001, 66, 7155-7158.
(23) Clark, T. F.; Ghadiri, M. R. J. Am. Chem. Soc. 1995, 117, 12364-12365.
Cardullo, F.; Calama, M. Crego; Snellink-Ruel, B. H. M.; Weidmann, J.-
L.; Bielejewska, A.; Timmerman, P.; Reinhoudt, D. N.; Fokkens, R.;
Nibbering, N. M. M. Chem. Commun. 2000, 5, 367-368.
(24) Coates, G. W.; Grubbs, R. H. J. Am. Chem. Soc. 1996, 118, 229-230.
(25) Hawker, C. J.; Fre´chet, J. M. J. J. Am. Chem. Soc. 1990, 112, 7638-7647.
(26) Zimmerman, S. C.; Wang, Y.; Bharathi, P.; Moore, J. S. J. Am. Chem.
Soc. 1998, 120, 2172-2173.
(27) Wendland, M. S.; Zimmerman, S. C. J. Am. Chem. Soc. 1999, 121, 1389-
1390.
(28) Matsui, J.; Higashi, M.; Takeuchi, T. J. Am. Chem. Soc. 2000, 122, 5218-
5219.
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13505

A R T I C L E S
Zimmerman et al.
Scheme 1
absorption bands of the porphyrin chromophore are intense and
sensitive to the local environment. Porphyrin-core dendrimers
have been studied as synthetic models of heme proteins, and
porphyrins have been shown to be excellent reporter groups in
studies of the photochemical and electrochemical properties of
dendrimers.30
Two synthetic hosts were prepared by imprinting 5,10,15,-
20-tetrakis(3,5-dihydroxyphenyl)porphyrin (6) and 5,10,15,20-
tetrakis(4-hydroxyphenyl)porphyrin (22) into octa- and tetra-
dendron macromolecules, respectively. Extensive complexation
studies with several potential porphyrin guests probed the
structural requirements for binding. The results suggest that this
strategy, wherein a single molecular template imprints a single
binding site within a single macromolecule, is a promising one
that merits further development.
interhost cross-linking. The 1H NMR spectrum of 9 showed the
expected broadening of all of its signals, as well as the nearly
complete loss of the alkene methine resonance at ca. 5.81 ppm
and the appearance of a new alkene peak at 5.60 ppm
corresponding to the disubstituted alkene group. Integration of
these overlapping signals suggested that the average number
of cross-links was about 29. Comparison of the MALDI-TOF
mass spectra of 7 and 9 (Figure 2A and 2B, respectively)
revealed a reduction in mass consistent with the formation of
28-32 cross-links, with the signal corresponding to 30 cross-
links being the most intense (Table 1).
Dendrimers 7 and 9 were studied by size-exclusion chroma-
tography (SEC). The apparent Mw of 7 as determined by SEC
in toluene is 8200 Da, 26% below its actual Mw as measured
by MALDI-TOF MS. This discrepancy is due to the globular
shape of 7 being more compact than the linear polystyrene
standards used for the Mw calibration. The SEC-determined
apparent Mw of the cross-linked dendrimer 9 is 5500 Da, nearly
54% below its actual molecular weight observed by MALDI-
TOF MS. This is consistent with the formation of an even more
compact structure upon cross-linking.
Results and Discussion
Imprinting 5,10,15,20-Tetrakis(3,5-dihydroxyphenyl)por-
phyrin (6) Inside of a Dendrimer. As shown in Scheme 2,
porphyrin-core dendrimer 7 was synthesized by DCC-mediated
esterification using dendron 8 and 5,10,15,20-tetrakis(3,5-
dihydroxyphenyl)porphyrin (6). Large-scale (ca. 40 g) prepara-
tion of dendron 8 was accomplished in six steps in 48% overall
yield from commercially available methyl 3,5-dihydroxyben-
zoate and homoallyl alcohol. Because of poor solubility, the
reaction was initially carried out in THF, but after partial
esterification the solvent was replaced with methylene chloride.
Typically, the esterification was partially incomplete with ca.
5% of the product containing only six or seven dendrons, as
discussed in greater detail below.
Cross-linked dendrimer 9 was hydrolyzed using a 2.5 M
aqueous KOH solution in THF to produce 10 in 43% yield.
The 1H NMR spectrum of the imprinted dendrimer 10 showed
loss of the porphyrin signals at 9.21, 8.14, 7.74, and -2.86 ppm.
Additionally, the aromatic protons ortho to the ester group and
appearing at 7.53 ppm in 9 were replaced by a signal at 7.20
ppm, consistent with loss of the porphyrin core and ester to
carboxylic acid conversion. Comparison of the MALDI-TOF
mass spectra of 9 and 10 (Figure 2B and 2D, respectively)
showed a difference of 605.4 Da corresponding to the loss of
the porphyrin core (C44H22N4: 606.7 Da).
Porphyrin-core dendrimer 7 was cross-linked in benzene (10-6
M) with 4 mol % Grubbs’ catalyst (2) per alkene to produce
the cross-linked dendrimer 9 in 88% yield, with less than 5%
The molecular weight of 10 could not be determined by SEC
because this rather polar compound did not elute from the
column. However, its octa-ethyl ester analogue 10-(CO2Et)8,
prepared by ethanolysis of 9, had an apparent molecular weight
of 4700 Da, 48% below its calculated molecular weight. In
comparison to 9, whose SEC-derived Mw was 5500, 10-
(CO2Et)8 appears to be even more compact. This result suggests
maintenance of a compact cross-linked dendrimer after the core
removal with partial or full contraction of the carboxylic acid
groups. It is not consistent with formation of a flexible, unfolded
structure. Elemental analysis of 10 confirmed that no nitrogen
was present, and, most significantly, the UV-visible spectrum
of 10 showed no detectable absorbance (i.e., less than 0.01
(29) Kosal, M. E.; Chou, J.-H.; Wilson, S. R.; Suslick, K. S. Nat. Mater. 2002,
1, 118-121. Sen, A.; Suslick, K. S. J. Am. Chem. Soc. 2000, 122, 11565-
11566. Rakow, N. A.; Suslick, K. S. Nature 2000, 406, 710-714. Suslick,
K. S. In The Porphyrin Handbook; Kadish, K., Smith, K., Guilard, R.,
Eds.; Academic Press: New York, 2000; Vol. 4, Chapter 28, pp 41-63.
Ogoshi, H.; Mizutani, T. Curr. Opin. Chem. Biol. 1999, 3, 736-739.
Sanders, J. K. M. Compr. Supramol. Chem. 1996, 9, 131-164. Suslick,
K. S.; Van Deusen-Jeffries, S. In ComprehensiVe Supramolecular Chem-
istry; Lehn, J. M., Ed.; Elsevier Publishers: Oxford, 1996; Vol. 5, pp 141-
170.
(30) Selected references: Bhyrappa, P.; Young, J. K.; Moore, J. S.; Suslick, K.
S. J. Am. Chem. Soc. 1996, 118, 5708-5711. Dandliker, P. J.; Diederich,
F.; Gisselbrecht, J.-P.; Louati, A.; Gross, M. Angew. Chem., Int. Ed. Engl.
1996, 34, 2725-2728. Sadamoto, R.; Tomioka, N.; Aida, T. J. Am. Chem.
Soc. 1996, 118, 3978-3979. Harth, E. M.; Hecht, S.; Helms, B.;
Malmstrom, E. E.; Fre´chet, J. M. J.; Hawker, C. J. J. Am. Chem. Soc. 2002,
124, 3926-3938.
9
13506 J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003

Molecular Imprinting Inside Dendrimers
A R T I C L E S
Scheme 2
absorbance, which is less than 0.1% of retained porphyrin) at
420 nm (Soret band). Thus, despite extensive cross-linking, the
porphyrin template can be quantitatively removed.
At issue are whether these isomers are kinetically or
thermodynamically favored and by what mechanism is 9 able
to become nearly fully cross-linked (i.e., 29-30 of 32 possible
cross-links)? Molecular modeling experiments to be published
elsewhere suggest that kinetic products are formed with several
alkenes left dangling sufficiently far from one another that they
cannot undergo the RCM reaction. Thus, a model is favored in
which the reversibility of the RCM reaction allows the cross-
linked dendrimer to dynamically mold itself around the por-
phyrin template.
Several aspects of the cross-linking reaction warrant additional
comment. First, cross-links may form within or between
dendritic wedges. As seen in Figure 3, there are two types of
intrawedge cross-links, a-b and a-c, and two types of
interwedge cross-links, a-d and a-e. Proximity likely promotes
intrawedge links, whereas interwedge connections are favored
statistically. Although none of the methods of characterization
used is capable of directly distinguishing the cross-link isomers,
if all were of type a-b and b-c, then the hydrolysis reaction
of 9 to 10 would lead to fragmentation with loss of one or more
cross-linked wedges. Neither the MALDI-TOF MS nor the SEC
traces of the crude reaction mixture or purified 10 showed
significant evidence of fragmentation. Thus, the cross-linking
reaction produces at least seven cross-links between the eight
dendritic wedges, seven interwedge cross-links being the
minimum needed to hold the cored dendrimer together.
In the RCM reaction of 1, it was estimated that nearly 1.7
million isomers are possible with just six cross-links. This
number, which was determined by enumeration, took into
account formation of both cis and trans alkenes, but neglected
topological isomers and the many isomers that would be
structurally untenable. Many more cross-link isomers are
possible for 9 in comparison to 3, but in both cases it is likely
that only a subset of isomers is formed.
Complexation Studies. In the noncovalent approach to MIPs,
the ligand that is the ultimate target for binding is also used as
the template in the imprinting reaction. In contrast, the target
ligand for covalently synthesized MIPs may not be an appropri-
ate template. For example, in the hydrolytic removal of template
6 from cross-linked dendrimer 9, eight additional hydroxyl
groups are added to the cored dendrimer, reducing the size of
the binding cavity. Thus, as seen by the very simple analysis
shown in Figure 4, a rigidly imprinted 10 would have a binding
cavity too small to complex 6, although the isomeric porphyrin
11 (Chart 1) could fit and potentially form multiple CO2H‚‚‚
OH hydrogen bonds. The same analysis indicates that the
covalent phenyl ester linkage is, to a good approximation,
isosteric with a pyridine-carboxylic acid hydrogen bond (Figure
4). As a result, a series of porphyrins (11-20, Chart 1) was
prepared and used to study the binding selectivity of imprinted
dendrimer 10.
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13507

A R T I C L E S
Zimmerman et al.
Figure 3. Dendrimer 7 with wedges and subwedges color-coded to show
different types of cross-link isomers.
Figure 2. MALDI-TOF MS spectra of (A) porphyrin-core dendrimer 7;
(B) and (C) cross-linked dendrimer 9; and (D) cored dendrimer 10.
Table 1. A List of Observed Values, Peak Assignments, and
Calculated Values for the Mass Spectra of 7, 9, and 10
observed
value
calculated
value
compound
peak assignment
7
9
11 153
10 313
10 285
10 257
9680
M + H+
11 156.4
10 313.9
10 285.9
10 257.9
9679.2
M + H+ - 30C2H4
Figure 4. Schematic illustration of how covalent linkage between template
and dendrimer scaffold impacts the ultimate binding interaction between
imprinted dendrimer and ligand.
M + H+ - 31C2H4
M + H+ - 32C2H4
10
M + H+ - 31C2H4 - C44H22N4
Several control experiments were performed that are consis-
tent with reversible complexation of 11, 12, and 14-16 by
hydrogen bonding. First, no red shift was observed with
porphyrins lacking hydrogen bond donor-acceptor groups such
as 5,10,15,20-tetraphenylporphyrin, 13. The red shift observed
upon addition of 10 to a solution of 11 could be fully reversed
by increasing the EtOAc content from 5% to 15% (v/v) in
toluene. Likewise, addition of 10 to a solution of 11 in 50%
EtOAc did not produce a red shift. To demonstrate the
importance of the carboxylic acid groups of 10 in the binding,
the octa-ethyl ester dendrimer 10-(CO2Et)8 was prepared by
ethanolysis of 9 (K2CO3, toluene, ethanol, reflux). The structure
A typical binding experiment consisted of titrating a dilute
porphyrin solution with a concentrated imprinted dendrimer
solution in toluene (5% EtOAc-toluene for 6 and 11) by adding
a 10-2 M solution of 10 to a 10-5 M solution of porphyrin and
recording the absorbance of the Soret band region 10 min after
each addition of imprinted dendrimer. Complexation of the
porphyrin by the dendrimer host was signaled by a red shift of
the λmax of the Soret band. Factors that might produce such a
red shift include a general polarity change in the environment,
π-π stacking, hydrogen bonding between the pyrrole nitrogen
atoms and the carboxylic acid groups, and a complexation-
induced deplanarization of the porphyrin ring.31 Initial experi-
ments using 11 indicated that the extent of the red shift was
dependent both on the amount of 10 added and on the time
elapsed. Similar observations were made for 12 and 14-16,
and in each case it appeared that the red shift occurred in two
distinct phases, one fast (seconds/minutes) and one very slow
(hours/days). The apparent biphasic binding kinetics is discussed
in greater detail below, but the experiments described first
focused on the fast binding component.
1
of 10-(CO2Et)8 was verified by SEC, H NMR spectroscopy,
MALDI-TOF MS, and UV-visible spectroscopy. The addition
of approximately 60 equiv of 10-(CO2Et)8 to H2T(3-pyridyl)P
(15) did not produce a red shift even over the course of 95 h.
Likewise, addition of simple carboxylic acids did not cause a
red shift of H2T(3-pyridyl)P (15). Neither the addition of 640
equiv of dendron carboxylic acid 2 to a ca. 3 µM solution of
H2T(2,6-OHPh)P (11) in 5% EtOAc-toluene nor the addition
of 2000 equiv of 4-tert-butyl benzoic acid to a ca. 3 µM solution
of H2T(3-pyridyl)P (15) in toluene led to any red shift. Finally,
cored dendrimer 4 with three acid groups, which was imprinted
(31) Shelnutt, J. A.; Song, X.-Z.; Ma, J.-G.; Jia, S.-L.; Jentzen, W.; Medforth,
C. J. Chem. Soc. ReV. 1998, 27, 31-42.
9
13508 J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003

Molecular Imprinting Inside Dendrimers
A R T I C L E S
Chart 1
Table 2. Apparent Association Constants (KApp) for Complexes
Formed between Imprinted Dendrimer 10 and Porphyrins
a
porphyrin
∆λmax
K
app (×104 M)b
H2T(3,5-OHPh)P (6)
H2T(2,6-OHPh)P (11)
H2T(3,5-pyrimidyl)P (12)
H2TPP (13)
H2T(2-pyridyl)P (14)
H2T(3-pyridyl)P (15)
H2T(4-pyridyl)P (16)
H2-tris(4-pyridyl)P (17)
H2-trans-bis(4-pyridyl)P (18)
H2-cis-bis(4-pyridyl)P (19)
H2-mono(4-pyridyl)P (20)
CuT(3-pyridyl)P (15a)
0
<0.1
10
5
<0.1
5
6.8
0.8
0
3.0
5.8
5.4
0
0
0
0
4.6
14
13
<0.1
<0.1
<0.1
<0.1
11
a A value of ∆λmax ) 0 indicates no observed change in the absorption
spectrum. b Upper limits on solubility prevented a more accurate upper limit
from being obtained.
that the Kassoc values can be determined without knowledge of
the concentration of the free or complexed porphyrin. To
examine the fast binding component, the titrations were
completed within 1 h. All gave linear plots, which, by analogy
to Scatchard-type plots, indicates formation of 1:1 complexes.
The association constants, which are listed in Table 2, are
denoted as Kapp because of the likely heterogeneity in 10 and
time dependency of the binding. As predicted by Figure 4, H2T-
(3,5-pyrimidyl)P (12) is bound by 10, and its Kapp ) 5 × 104
M-1 is of the same magnitude as the value determined for the
10‚15 complex. The pyridyl porphyrins 14-16 bind equally well
despite their capacity to make only half as many hydrogen bonds
as 12. This may indicate that all eight carboxylic acids are not
involved in binding H2T(3,5-pyrimidyl)P (12); however, the
difference must at least partially originate in the 104-fold higher
basicity of pyridine relative to pyrimidine (pKa 5.23 vs 1.2).33
The Kapp for the 10‚14 complex (2-pyridyl isomer) is ap-
proximately 2-fold lower than that for the 3- and 4-pyridyl
isomers 15 and 16. Thus, there is a small degree of shape
selectivity in the imprint. No binding was detected for pyridyl
porphyrins 17-20, showing that a minimum of four binding
contacts are necessary for complexation under the conditions
used in this study.
with 1,3,5-tris(hydroxymethyl)benzene, did not alter the position
of the Soret band of a ca. 3 µM solution of 11 in toluene even
upon addition of 49 equiv in 95 h.
The results of two additional experiments indicated that
porphyrins 11, 12, and 14-16 were complexed within 10 and
that its binding sites arose from an imprinting process. First,
complexation of H2T(2,6-OHPh)P (11) by imprinted dendrimer
10 was observed by SEC. A 5% EtOAc-toluene solution
containing a 5:1 molar ratio of 10 to 11 was analyzed by
analytical SEC using 5% EtOAc-toluene as the eluent. Whereas
porphyrin 11 absorbs to the SEC matrix and does not elute by
itself, it coelutes with imprinted dendrimer 10 (as detected at
λmax ) 417 nm Soret band). Thus, complexation within the
dendrimer protects the polar porphyrin from SEC matrix
absorption during the course of its elution. A second, key
experiment involved adding 10 to template 6 in 5% EtOAc-
toluene. No change in the Soret band was observed upon titration
of 6 with 10. This finding supports the model presented in Figure
4 and is inconsistent with 10 being a flexible octa-acid able to
complex any guest with sufficient hydrogen bond donor and
acceptor groups.
Although the Kapp values for several of the porphyrin guests
are large, one might have expected them to be even larger given
that a typical benzoic acid-pyridine complex in apolar organic
solvents has a Kassoc ≈ 102 M-1 34
.
Several factors may be
responsible for lowering the absolute porphyrin affinity exhibited
by 10. Conformational reorganization may be necessary if the
empty imprint relaxes into a lower energy structure; such a state
might include partial collapse of the core site through internal
carboxylic acid dimerization or the internal binding of multiple
water molecules. The IR spectra of 10 were inconclusive with
regard to the latter possibility, but the elemental analysis did
reveal the presence of multiple water molecules, presumably
solvating the polar carboxylic acid groups. In studies of adenine
complexation by molecular tweezers with a single “active site”
carboxylic acid group, we reported that added water in
chloroform led to a measurable depression in the binding
constant.35
To determine more quantitatively the binding profile of 10,
the titration data for porphyrins 11-20 (Chart 1) were analyzed
using a modified form of the Drago equation: [10] versus [10]/
∆A, where ∆A is the change in absorbance upon addition of
10.32 The reciprocal of the intercept of this plot gives the
association constant (Kassoc). One advantage of this approach is
(32) Collman, J. P.; Brauman, J. I.; Doxsee, K. M.; Halbert, T. R.; Hayes, S.
E.; Suslick, K. S. J. Am. Chem. Soc. 1978, 100, 2761-2766.
(33) Stewart R. The Proton: Applications to Organic Chemistry; Academic
Press: Orlando, FL, 1985.
(34) Dega-Szafran, Z.; Szafran, M. Heterocycles 1994, 37, 627-659.
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13509

A R T I C L E S
Zimmerman et al.
Dendrimer Fractionation. The observation of slow and fast
binding processes may be consistent with two limiting popula-
tions of imprinted dendrimers, one that is tightly cross-linked
with a somewhat inaccessible core and another that is more
flexible and open. The latter properties are exactly those which
would be expected for an impurity of 10 missing one or more
dendron segments, and, thus, attention was focused on the ca.
5% impurity in 10 containing just six and seven dendrons (vide
supra). To determine if this imperfect component was respon-
sible for the observed two-phase binding, two series of experi-
ments were conducted. In the first, dendrimer 10 was synthe-
sized using a large excess of dendron 8. The new sample of 10
obtained in this way contained significantly less hexa- and
heptadendritic material. The titration of 15 with this new sample
of 10, however, was essentially identical to that with the original
(Table 2).
In the second set of experiments, the hexa- and heptadendrons
associated with impurities in 10 were isolated and examined
directly. Thus, 6-Zn was found to be significantly less reactive
in the Mitsunobu esterification reaction. Thus, treatment of 6-Zn
with 8 equiv of 8 under the same conditions used with 6 afforded
7-Zn containing a significant amount of hexa- and heptadendron
product. Cross-linking of 7-Zn under standard conditions gave
9-Zn whose MALDI-TOF spectrum is shown in Figure 5A. The
material containing predominantly six and seven dendrons was
separated from the dendrimers containing mainly eight dendrons
by selective absorption on a diethylenetriamine-modified poly-
styrene resin in benzene. The material left in solution contained
mostly seven and eight dendrons (Figure 5B), while the material
“trapped” on the resin and removed with 5% pyridine contained
mostly six and seven dendrons (Figure 5C). The selectivity in
the absorption was explained by a stronger amino group ligation
to the more accessible Zn(II) center in the imperfect dendritic
macromolecules. The dendrimers with six and seven dendrons
were cored, resulting in “imperfect” imprinted material. When
added to a toluene solution of 15, this material, which was
noticeably less soluble than previous preparations of 10, did
not induce a red shift. Therefore, it was concluded that hexa-
and heptadendritic structures do not participate in porphyrin
binding to a measurable extent and that the origin of biphasic
binding kinetics lies elsewhere.
Figure 5. MALDI-TOF MS spectra of (A) a mixture of porphyrin-core
dendrimer 9-Zn; (B) dendrimers with mainly eight dendrons left in solution;
and (C) dendrimers with mainly six and seven dendrons absorbed on the
resin.
Binding Site Heterogeneity, Complexation Kinetics, and
Cu-T(3-pyridyl)P. The incomplete reaction between 6 and 8
led to one type of heterogeneity in imprinted dendrimer 10. A
second type originates in the RCM reaction of dendrimer 7,
which can produce 9 containing a very large number of cross-
link isomers (vide supra). It is possible that the slow and fast
binding kinetics indicate two populations of cross-link isomers.
For example, it is likely that isomers containing cross-links
primarily between dendrons will be more compact and rigid
with slower binding kinetics than those isomers where intra-
dendron cross-links predominate. To examine more systemati-
cally the kinetics of binding, six 3 µM solutions of H2T(3-
pyridyl)P (15) containing 0.0, 0.5, 1, 1.5, 2.0, and 5.0 equiv of
10 were monitored over a period of 408 h (17 days). Shown in
Figure 6A and B are overlaid spectra of the solutions taken at
1 and 408 h, respectively. The overlaid spectra at intermediate
times similarly show an isosbestic point, whose position moves
only slightly with time. These data indicate that (1) a single
equivalent of 10 is able to complex most of H2T(3-pyridyl)P,
and (2) there are two classes of bound species. The former
observation has important implications for the monomolecular
imprinting approach in that it demonstrates that the chemistry
shown schematically in Figure 1 and in detail in Scheme 2
produces 10 containing >95% effective imprinted binding sites.
To gain insight into the nature of the bound species, the 35
spectra described above (5 solutions at 0.17, 1, 3, 28, 116, 206,
408 h) were deconvoluted using model spectra for three species
(Figure 6C): (1) free porphyrin (λmax ) 421 nm), (2) the fast
forming complex (Complex I), and (3) the slow forming
complex (Complex II). The model spectrum for Complex I (λmax
) 426 nm) was obtained 1 h after mixing 50 equiv of 10 and
H2T(3-pyridyl)P in toluene, conditions that led to >95% bound
(35) Zimmerman, S. C.; Wu, W.; Zeng, Z. J. Am. Chem. Soc. 1991, 113, 196-
201. In a receptor using two carboxylic acid groups, Adrian and Wilcox
reported an ca. 15% decrease in Kassoc and further showed a much larger,
but compensating, effect on the complexation enthalpies and entropies of
complexation: Adrian, J. C., Jr.; Wilcox, C. S. J. Am. Chem. Soc. 1991,
113, 678-680. It is possible with multiple carboxylic acids that the effect
is cumulative.
9
13510 J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003

Molecular Imprinting Inside Dendrimers
A R T I C L E S
Figure 6. Overlaid spectra of the titration solutions taken at (A) 1 h and (B) 408 h; spectra of the model species (C) and the time course of the binding of
the porphyrin 15 with 1 equiv of 10 (D); blue line, 15; green line, Complex I; red line, Complex II.
porphyrin. The model spectrum for Complex II (λmax ) 432
nm) was that generated after equilibrating 5 equiv of 10 and
H2T(3-pyridyl)P in toluene for 408 min. The model spectra of
Complexes I and II are broader than that of the free porphyrin,
suggesting that each is a class of complexes that are similar in
their stability, formation kinetics, and UV-visible spectra. Using
the three model spectra, we deconvoluted each titration spectrum
with an average deviation of 0.019 au. The apparent molar
absorption coefficients determined for Complexes I and II are
2.07 × 105 and 1.92 × 105 mol1 cm-1, respectively.
nonplanarity of the porphyrin nucleus required for the additional
stabilizing hydrogen bond would explain the additional 6 nm
red shift.31 Direct support for such a mechanism is not available,
but the use of CuT(3-pyridyl)P (15-Cu) which lacks a free
pyrrole nitrogen atom completely eliminates the slow complex-
ation process. The dramatic change in the binding course is
clearly seen in the side-by-side comparison of the spectropho-
tometric titrations of H2T(3-pyridyl)P and CuT(3-pyridyl)P with
10 (Figure 7). Whereas the Soret band of 15 continues to shift
to longer wavelength after the initial rapid red shift, CuT(3-
pyridyl)P gives ∆λmax ) 4.6 nm in 10 min and then shows no
additional shift. The Kapp values determined for 10‚H2T(3-
pyridyl)P and 10‚CuT(3-pyridyl)P are quite similar (Table 2).
Imprinted Tetradendron Molecule. To examine whether a
porphyrin template with only four attachment points could
imprint an effective binding site containing four carboxylic acid
groups, dendritic porphyrin 21 was synthesized from 5,10,15,-
20-tetrakis(4-hydroxyphenyl)porphyrin and 23 (Scheme 3).
Thus, 8 was treated with oxalyl chloride in THF to produce 23
in 64% yield, which was reacted directly with 5,10,15,20-
tetrakis(4-hydroxyphenyl)porphyrin and DMAP in THF, af-
fording porphyrin core dendrimer 21 in 90% yield. The cross-
linking of 21 was accomplished in 79% yield using Grubbs’
catalyst 2 and the same conditions described for 7. Analysis by
MALDI-TOF MS indicated an average of 15 out of 16 possible
cross-links. However, upon core removal (same conditions as
for 10), MALDI-TOF mass spectra showed clear evidence of
fragmentation as peaks corresponding to cross-linked dendrimers
with only two and three dendrons were observed. The results
suggested that the dendrons in 21 were too far apart to promote
extensive interdendron cross-linking.
The time course of the binding as revealed by the spectral
deconvolution is shown in Figure 6D. Taken together, all of
the results are consistent with the free porphyrin rapidly forming
Complex I with an approximately 5 nm red shift. The decon-
volution indicates a maximum of 30% of H2T(3-pyridyl)P (at
10-5 M concentration of the porphyrin) is bound by a single
equivalent of 10, consistent with the amount expected given
the Kapp ) 1.4 × 105 M-1. Over time, Complex I is replaced
by Complex II whose Kapp is about 10-fold higher, leading to
70% H2T(3-pyridyl)P being bound. Although the two complexes
may arise from heterogeneity in the imprint (i.e., separate
isomers of 10), the stronger binding, the additional 6 nm red
shift observed for Complex II, and the evolution of the red shift
seen in Figure 6D are particularly well explained by a slow
conformational change to a tighter fitting complex.
Although there are several explanations for the observed fast
and slow complex formation, a conformation change represents
a likely candidate, and it is possible to speculate on its nature.
Given that only four of the carboxylic acid groups of 10 are
needed to complex H2T(3-pyridyl)P, a free carboxylic acid group
might hydrogen bond to an inner pyrrole nitrogen atom of the
porphyrin guest. Such a significant conformational reorganiza-
tion might take considerable time. Furthermore, the increased
To favor interdendron cross-linking, the larger, fourth-
generation dendron 24 was employed (Scheme 4). Dendron 24
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13511

A R T I C L E S
Zimmerman et al.
Figure 7. Overlaid spectra of the titration solutions for (A) H2T(3-pyridyl)P and (B) Cu-T(3-pyridyl)P after 408 h; fast binding is shown in blue, and slow
binding is shown in red.
Scheme 3
was prepared in 83% yield by basic hydrolysis of the corre-
sponding methyl ester, which was prepared in 89% yield by
reaction of methyl 3,5-dihydroxybenzoate with [G-3]-OH under
Mitsunobu conditions. Porphyrin-core dendrimer 25 was syn-
thesized in 46% yield by coupling 22 to 24 in THF with DPTS
and DCC (same conditions as for 7). Dendrimer 25 was cross-
linked in benzene using 5 mol % Grubbs’ catalyst per alkene
to afford the cross-linked dendrimer 26 in 88% yield, which
was hydrolyzed to produce the imprinted dendrimer 27 in 48%
yield.
As in the 7 f 9 f 10 conversion, the MALDI-TOF mass
spectra provide the most conclusive evidence for the success
of the cross-linking and coring reactions. The observed and
calculated values for the MALDI-TOF mass spectra peaks, as
well as the corresponding peak assignments for 25, 26, and 27,
are listed in Table 3. The data indicate an average of 31 of 32
cross-links for 26. Evidence of the removal of the porphyrin
core was obtained by comparing the MALDI-TOF mass spectra
of 26 and 27. No peak corresponding to 26 was observed in
the spectrum of 27, but a new peak was present consistent with
a loss of ca. 607 Da. The loss in mass was within experimental
error of that expected for loss of the core (C44H28N4, 614.7).
The mass spectrum of 27 also showed that, in contrast to the
cross-linked and cored product from 21, little if any of 26
fragmented upon coring.
The pyridyl porphyrins 14-16 were titrated with imprinted
dendrimer 27 in toluene, and their binding constants were
determined (Table 4). The decrease in binding constants in the
series 14-16 may result from a combination of the shape-
selective binding and decreasing basicity of the pyridyl groups.
The plots used to calculate binding constants for the complexes
formed between porphyrins 14-16 and 27 were linear, indicat-
ing the stoichiometry of these complexes was 1:1. A control
study was performed by titrating tetraphenylporphyrin (13) with
imprinted dendrimer 27, and no change in the Soret band
position of 13 was observed even upon the addition of 66 equiv
of 27, confirming that hydrogen-bonding to the imprinted
dendrimer is the cause for the red shift observed for the pyridyl
porphyrins. The titration solutions of 14 and 15 with 27 were
injected onto an analytical SEC column and eluted with toluene.
Porphyrins 14 and 15 coeluted with the imprinted dendrimer
27. When toluene was removed from the solution of 15 and 27
and the mixture redissolved in THF was injected into the SEC
9
13512 J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003

Molecular Imprinting Inside Dendrimers
A R T I C L E S
Scheme 4
Table 3. A List of Observed Values, Peak Assignments, and
Calculated Values for the Mass Spectra of 25, 26, and 27
tially allows the dendritic framework to reach a lowest-energy
“mold” around the template. Perhaps more importantly this
strategy produces soluble and sizable macromolecular hosts (Mw
≈ 10 kDa) but with a single imprinted site per molecule.
Although there was no strong evidence of binding site hetero-
geneity in the present system, in cases where mixed imprints
do arise, the potential exists for fractionation. In this regard,
the ability to select hosts based on binding kinetics, selectivity,
or affinity would be quite powerful.
observed
value
calculated
value
compound
peak assignment
25
26
11 526
10 624
10 653
10 041
10 069
M + H+
11 525.0
10 627.2
10 655.3
10 034.5
10 062.6
M + H+ - 32C2H4
M + H+ - 31C2H4
27
M + Na+ - 32C2H4 - C44H28N4
M + Na+ - 31C2H4 - C44H28N4
Table 4. Apparent Association Constants (KApp) for Complexes
Formed between Imprinted Dendrimer 27 and Porphyrins
In conventional polymer imprinting, removing the template
remains a challenge. Retained template has been shown to be a
serious obstacle to trace analysis applications36 and to be
involved in apparent chiral recognition exhibited by imprinted
polymers.8 In the two cases described here, the template was
quantitatively removed. Another significant finding is that 1
equiv of H2T(3-pyridyl)P, 15, was nearly fully complexed by a
single equivalent of 10, indicating that almost all of the imprints
were effective. Improvements in affinity and shape selectivity
are clearly needed but will come as the structure of the dendron
is tuned to provide a greater degree of rigidity. In this respect,
there are limitless variations possible, and one can even envision
highly rigid imprints with entry and exit channels designed into
the system.
porphyrin
∆λmax
K
app (×104 M)
H2T(2-pyridyl)P (14)
H2T(3-pyridyl)P (15)
H2T(4-pyridyl)P (16)
1.0
2.8
1.6
3.3
1.6
0.9
column, the porphyrin 15 eluted significantly later than the
imprinted dendrimer 27.
Conclusions
Described herein is a new strategy to make synthetic hosts
that combines elements of the traditional polymer imprinting
and dynamic combinatorial library approaches. There are several
appealing features of this monomolecular imprinting strategy.
The first is the use of the RCM reaction, which forms robust
carbon-carbon double-bond cross-links but is nonetheless
reversible. The ability to equilibrate cross-link isomers poten-
(36) Venn, R. F.; Goody, R. J. In Drug DeVelopment Assay Approaches Including
Molecular Imprinting and Biomarkers; Reid, E., Hill, H. M., Wilson, I.
D., Eds.; Royal Society of Chemistry: Cambridge, U.K., 1998; pp 13-
20.
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13513

A R T I C L E S
Zimmerman et al.
shifts were referenced to the residual protio solvent peak at 7.26 ppm
in chloroform-d (CDCl3), 5.32 ppm in methylene chloride-d2 (CD2-
Cl2), and 2.09 in toluene-d8. The 13C NMR chemical shifts were
referenced to the solvent peak at 77.0 ppm in CDCl3 and 53.8 ppm in
The cross-linking of 7 and 25 was performed under high
dilution conditions that favored intramolecular cross-linking but
made it difficult to scale-up the preparations and obtain
significant quantities of the imprinted hosts. It is possible to
overcome this limitation given the recent finding that alkene
groups within the branching units of a dendrimer can avoid
intermolecular cross-linking at significantly higher concentra-
tions.37 One of the advantages of using dendrimers for imprinting
is in their monodispersity, which facilitates characterization. For
example, in the current work, this allowed the number of cross-
links to be determined. Ultimately, more rapid dendrimer
syntheses,38 or even hyperbranched polymerizations39 or single-
step syntheses of related macromolecules, might be applied.
Along with efforts in these areas, our current attention is focused
on expanding the types of binding contacts used, developing
new dendrons, and integrating reporter groups into the binding
site.
1
CD2Cl2. Unless stated otherwise, the H and 13C NMR spectra were
acquired in CDCl3.
All mass spectra were obtained in the Mass Spectrometry Laboratory,
School of Chemical Sciences, University of Illinois, Urbana-Champaign.
Mass spectra were measured by the technique of field desorption (FD)
and field ionization (FI) on a Finnigan-MAT 731 spectrometer, chemical
ionization (CI) and electron impact (EI) on a VG 70-VSE spectrometer,
and fast atom bombardment (FAB) on a VG ZAB-SE spectrometer.
Mass spectra were measured by matrix-assisted laser desorption
ionization time-of-flight (MALDI-TOF) on a PerSeptive Biosystems
Voyager DE-STR spectrometer; 2-(4-hydroxyphenylazo)benzoic acid
was used as the matrix to obtain the MALDI-TOF mass spectrum.
Infrared spectra were recorded on a Mattson Galaxy Series FTIR
5000 spectrometer and are reported in cm-1. Ultraviolet/visible (UV/
vis) absorbance spectra were recorded on a Hitachi U-3300 spectro-
photometer using a tungsten lamp light source, and λmax are reported
in nanometers. Elemental analyses were performed at the Microanalysis
Laboratory, School of Chemical Sciences, University of Illinois,
Urbana-Champaign.
Analytical size-exclusion chromatography (SEC) was performed on
a Waters Styragel HR3 column (molecular weight range 500-30 000)
coupled with a Water 410 differential refractometer and PD2000 dual
angle laser light scattering detectors or with an Hitachi L-4000H UV
detector using an Hitachi L-6000 pump. Molecular weights (Mw)
calculated from SEC were based on using polystyrene standards.
Preparative SEC was carried out on Bio-Beads S-X1 Beads gel
permeation gel 200-400 mesh (Bio-Rad Laboratories), which has
exclusion limits from 400 to 14 000.
In addition to a compound name, dendrons and dendrimers are
designated using the notation [G-x]-core, where [G-x] refers to the
generation number (x ) 1, 2, 3, or 4), and core refers to the functional
group at the focal point of the dendron or the molecule used as the
core to make the dendrimer. Because of the complexity of structure,
the cross-linked and cored dendrimers are named with the word cross-
linked or cored preceding the shorthand name for that compound. This
shorthand notation is used throughout the text to refer to these
compounds.
The titration data were analyzed by plotting [ID] versus [ID]/∆A,
where [ID] is the imprinted dendrimer concentration and ∆A is the
change in absorbance upon addition of 10 or 27. This method allows
association constants to be determined by finding the line intercept
that gives 1/Kapp and does not require knowledge of the concentration
of the free or complexed porphyrin. The titrations were all completed
within 1 h and all gave linear plots. As with other Scatchard-type plots,
linearity indicates the formation of 1:1 complexes.
Experimental Section
General. All of the reactions described below were run under a dry
nitrogen atmosphere, and all temperatures reported as reaction or drying
conditions were the temperatures of the heating medium. All solvents
and reagents were of reagent quality, purchased commercially, and used
without further purification, except as noted below. The following
solvents were freshly distilled prior to use: diethyl ether (Et2O) and
tetrahydrofuran (THF) from sodium-benzophenone; ethyl acetate
(EtOAc), methylene chloride (CH2Cl2), and toluene from calcium
hydride; benzene from sodium; and methanol (MeOH) from calcium
sulfate. Acetone and N,N-dimethylformamide (DMF) were stored over
4 Å molecular sieves. 3,5-Dihydroxybenzoic acid methyl ester,40 5,-
10,15,20-tetrakis(3′,5′-dihydroxyphenyl)porphyrin (6),41 5,10,15,20-
tetrakis(2′,6′-dihydroxyphenyl) porphyrin (11),42 5,10,15,20-tetrakis(2′-
pyridyl)porphyrin (14),43 5,10,15,20-tetrakis(3′-pyridyl)porphyrin (15),43
pyrimidine-5-carboxaldehyde,44 and 5,10,15,20-tetrakis(phenyl)por-
phyrin (13)45 were synthesized according to published procedures.
Thin-layer chromatography was performed on 0.2 mm silica 60
coated plastic sheets (EM Science) with F254 indicator. Flash chroma-
tography was performed on Merck 40-63 µm silica gel. Solvent ratios
for the purification of compounds by flash chromatography are reported
as percent volume (v/v). Dimensions of the columns used for the flash
chromatography are reported as (width × height).
All nuclear magnetic resonance (NMR) spectra were acquired in
the Varian-Oxford Center for Excellence in NMR Spectroscopy
(VOICE) laboratory at the University of Illinois, Urbana-Champaign.
1
All H and 13C NMR spectra were recorded on a Varian Unity 500
spectrometer (1H, 500 MHz; 13C, 125.7 MHz). Heteronuclear multiple
quantum coherence (HMQC) and heteronuclear multiple bond correla-
tion (HMBC) NMR experiments were performed on a Varian Unity
Inova 500NB spectrometer (1H, 500 MHz; 13C, 125.7 MHz). 1H NMR
3,5-Bis(3-buten-1-oxy)benzoic Acid Methyl Ester ([G-1]-CO2Me).
To a mixture of 75.0 g (1.04 mol) of 3-buten-1-ol, 79.4 g (472 mmol)
of 3,5-dihydroxybenzoic acid methyl ester, and 310.0 g (1.18 mol) of
triphenylphosphine (PPh3) in 1.0 L of THF cooled to 0 °C with an
ice/water bath was added dropwise a solution of 262.2 g (1.51 mol) of
diethyl azodicarboxylate (DEAD) in 3.0 L of THF. The reaction was
allowed to warm to room temperature and stirred until the reaction
was complete as indicated by TLC. The reaction was stopped by adding
700 mL of water and removing the THF under reduced pressure. The
resulting aqueous layer was extracted with Et2O (3 × 2 L). The organic
layers were combined and washed with an equal volume of 2.5 M
aqueous potassium hydroxide (KOH) and an equal volume of water.
The organic layers were dried over sodium sulfate and filtered. The
filtrate was reduced to one-third of its volume and placed in the freezer
(-17 °C) overnight. A white precipitate (triphenylphosphine oxide)
was filtered off and washed with cold Et2O. The resulting filtrate was
concentrated under reduced pressure. Approximately 25-30 g of the
1
coupling constants are reported in hertz (Hz). The H NMR chemical
(37) Schultz, L. G.; Zhao, Y.; Zimmerman, S. C. Angew. Chem., Int. Ed. 2001,
40, 1962-1966.
(38) Cf.: Zeng, F.; Zimmerman, S. C. J. Am. Chem. Soc. 1996, 118, 5326-
5327.
(39) Kim, Y. J. Polym. Sci., Polym. Chem. 1998, 36, 1685-1698. Sunder, A.;
Heinemann, J.; Frey, H. Chem.-Eur. J. 2000, 6, 2499-2506. Jikei, M.;
Kakimoto, M. Prog. Polym. Sci. 2001, 26, 1233-1285.
(40) Chakraborty, T. K.; Reddy, G. V. J. Org. Chem. 1992, 57, 5462-5469.
(41) Jin, R.; Aida, T.; Inoue, S. J. Chem. Soc., Chem. Commun. 1993, 1260-
1262.
(42) Bhyrappa, P.; Vaijayanthimala, G.; Suslick, K. S. J. Am. Chem. Soc. 1999,
121, 262-263. Momenteau, M.; Mispetter, J.; Loock, B.; Bisagni, E. J.
Chem. Soc., Perkin Trans. 1 1983, 189-196.
(43) Sugata, S.; Yamanouchi, S.; Matsushima, Y. Chem. Pharm. Bull. 1977,
25, 884-889.
(44) Rho, T.; Abuh, Y. Synth. Commun. 1994, 24, 253-256.
(45) Adler, A. D.; Longo, F. R.; Finarelli, J. D.; Goldmacher, J.; Assour, J.;
Korsakoff, L. J. Org. Chem. 1967, 32, 476.
9
13514 J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003

Molecular Imprinting Inside Dendrimers
A R T I C L E S
crude product was purified at a time by column chromatography. The
crude product was loaded onto a silica gel plug (8 × 12 cm) and eluted
with 1.5 L of petroleum ether (PE), 2.0 L of 5% EtOAc/95% PE, and
1.0 L of 10% EtOAc/90% PE. The product was dried under vacuum
(0.1 mmHg, 45 °C) overnight to afford 109.9 g (84%) of [G-1]-CO2-
Me as a clear, colorless oil. 1H NMR: δ 7.16 (d, 2H, J ) 2.4), 6.64 (t,
1H, J ) 2.4), 5.88 (ddt, 2H, J ) 17.2, 10.3, 6.8), 5.16 (ddt, 2H, J )
17.2, J ) 1.8, 1.5), 5.11 (ddt, 2H, J ) 10.3, J ) 1.8, 1.3), 4.01 (t, 4H,
J ) 6.6), 3.88 (s, 3H), 2.52 (dtdd, 4H, J ) 6.8, 6.6, 1.5, 1.3). 13C
NMR: 166.8, 159.9, 134.3, 131.9, 117.1, 107.8, 106.7, 67.4, 52.2, 33.5.
IR (neat): 3070, 2980, 1724, 1642, 1597, 1169, 1057, 996, 918. MS
(FAB): m/z 277.1 (M + H+). Anal. Calcd for C16H20O4: C, 69.55; H,
7.30. Found: C, 69.41; H, 7.12.
THF cooled to 0 °C with an ice/water bath was added dropwise a
solution of 64.5 g (103 mmol) of [G-2]-CO2Me in 750 mL of THF.
The suspension was allowed to warm to room temperature and stirred
until no starting material remained by TLC. The reaction was stopped
by quenching the excess hydride with 150 mL of water and neutralizing
the pH with 130 mL of 1 M aqueous HCl. The THF was removed
under reduced pressure, and the resulting aqueous layer was extracted
with Et
2O (3 × 1.5 L). The combined organic layers were washed with
water, dried over sodium sulfate, filtered, and concentrated under
reduced pressure. Approximately 20 g of the crude product was purified
at a time by column chromatography. The crude product was loaded
onto a silica gel column (8 × 26 cm) and eluted with 20% EtOAc/
80% PE. The product was dried under high vacuum (0.2 mmHg, 75
°C) overnight to afford 54.8 g (89%) of [G-2]-OH as a clear, colorless
3,5-Bis(3-buten-1-oxy)benzyl Alcohol ([G-1]-OH). To a suspension
of 24.6 g (649 mmol) of lithium aluminum hydride (LAH) in 500 mL
of THF cooled to 0 °C with an ice/water bath was added dropwise a
solution of 109.8 g (397 mmol) of [G-1]-CO2Me in 1.5 L of THF. The
suspension was allowed to warm to room temperature and stirred until
no starting material remained by TLC. The reaction was stopped by
quenching the excess hydride with 200 mL of water and neutralizing
the pH with 200 mL of 3 M aqueous hydrochloric acid (HCl). The
THF was removed under reduced pressure, and the resulting aqueous
layer was extracted with Et2O (4 × 750 mL). The combined organic
layers were washed with water, dried over sodium sulfate, filtered, and
concentrated under reduced pressure. The product was purified by
vacuum distillation (140-150 °C, 0.2 mmHg) to afford 96.4 g (98%)
1
oil. H NMR: δ 6.61 (d, 2H, J ) 2.2), 6.57 (d, 4H, J ) 2.2), 6.53 (t,
1H, J ) 2.2), 6.41 (t, 2H, J ) 2.2), 5.90 (ddt, 4H, J ) 17.2, 10.3, 6.8),
5.17 (ddt, 4H, J ) 17.2, 1.8, 1.5), 5.11 (ddt, 4H, J ) 10.3, 1.8, 1.3),
4.96 (s, 4H), 4.63 (d, 2H, J ) 6.2), 4.00 (t, 8H, J ) 6.6), 2.53 (dtdd,
8H, J ) 6.8, 6.6, 1.5, 1.3), 1.57 (t, 1H, J ) 6.2). 13C NMR: 160.3,
160.1, 143.5, 139.1, 134.4, 117.1, 105.9, 105.7, 101.3, 101.0, 70.0,
67.3, 65.3, 33.6. IR (neat): 3425 (bs), 3077, 2979, 1641, 1599, 1166,
1054, 993, 918. MS (FD): m/z 600.3 (M+). Anal. Calcd for C37H44O7:
C, 73.98; H, 7.38. Found: C, 73.92; H, 7.18.
3,5-Bis{3,5-bis[3,5-bis(3-buten-1-oxy)benzyloxy]benzyloxy}-
benzoic Acid Methyl Ester ([G-3]-CO2Me). To a mixture of 49.3 g
(82.1 mmol) of [G-2]-OH, 6.28 g (37.3 mmol) of 3,5-dihydroxybenzoic
acid methyl ester, and 24.5 g (93.5 mmol) of PPh3 in 200 mL of THF
cooled to 0 °C with an ice/water bath was added dropwise a solution
of 20.2 g (116 mmol) of DEAD in 300 mL of THF. The reaction was
allowed to warm to room temperature and stirred until the reaction
was complete as indicated by TLC. The reaction was stopped by adding
150 mL of water and removing the THF under reduced pressure. The
resulting aqueous layer was extracted with Et2O (3 × 500 mL). The
combined organic layers were washed with an equal volume of 2.5 M
aqueous KOH and an equal volume of water. The organic layers were
dried over sodium sulfate and filtered. The filtrate was reduced to one-
third its volume and placed in the freezer (-17 °C) overnight. A white
precipitate (triphenylphosphine oxide) was filtered off and washed with
cold Et2O. The resulting filtrate was concentrated under reduced
pressure. Approximately 40 g of the crude product was purified at a
time by column chromatography. The crude product was loaded onto
a silica gel column (10 × 25 cm) and eluted with 7.0 L of 20% EtOAc/
80% PE and 3.0 L of 30% EtOAc/70% PE. The product was dried
under vacuum (0.1 mmHg, 75 °C) overnight to afford 44.7 g (90%) of
1
of [G-1]-OH as a clear, colorless oil. H NMR: δ 6.51 (d, 2H, J )
2.2), 6.38 (t, 1H, J ) 2.2), 5.89 (ddt, 2H, J ) 17.2, 10.3, 6.8), 5.17
(ddt, 2H, J ) 17.2, 1.8, 1.5), 5.10 (ddt, 2H, J ) 10.3, 1.8, 1.3), 4.60
(s, 2H), 3.99 (t, 4H, J ) 6.6), 2.53 (dtdd, 4H, J ) 6.8, 6.6, 1.5, 1.3),
1.87 (bs, 1H). 13C NMR: 160.3, 143.3, 134.4, 117.1, 105.3, 100.7,
67.2, 65.3, 33.6. IR (neat): 3370 (bs), 3077, 1642, 1597, 1167, 1057,
992, 917. MS (FAB): m/z 249.2 (M + H+). Anal. Calcd for C15H20O3:
C, 72.55; H, 8.12. Found: C, 72.27; H, 8.35.
3,5-Bis[3,5-bis(3-buten-1-oxy)benzyloxy]benzoic Acid Methyl Es-
ter ([G-2]-CO2Me). To a mixture of 96.0 g (387 mmol) of [G-1]-OH,
29.6 g (176 mmol) of 3,5-dihydroxybenzoic acid methyl ester, and 115.6
g (441 mmol) of PPh3 in 500 mL of THF cooled to 0 °C with an ice/
water bath was added dropwise a solution of 99.6 g (572 mmol) of
DEAD in 1.5 L of THF. The reaction was allowed to warm to room
temperature and stirred until the reaction was complete as indicated
by TLC. The reaction was stopped by adding 500 mL of water and
removing the THF under reduced pressure. The resulting aqueous layer
was extracted with Et2O (4 × 750 mL). The combined organic layers
were washed with an equal volume of 2.5 M aqueous KOH and an
equal volume of water. The organic layers were dried over sodium
sulfate and filtered. The filtrate was reduced to one-third its volume
and placed in the freezer (-17 °C) overnight. A white precipitate
(triphenylphosphine oxide) was filtered off and washed with cold Et2O.
The resulting filtrate was concentrated under reduced pressure. Ap-
proximately 25-30 g of the crude product was purified at a time by
column chromatography. The crude product was loaded onto a silica
gel plug (8 × 12 cm) and eluted with 1.0 L of PE, 3.0 L of 5% EtOAc/
95% PE, and 3.0 L of 10% EtOAc/90% PE. The product was dried
under vacuum (0.15 mmHg, 68 °C) overnight to afford 94.0 g (85%)
1
[G-3]-CO2Me as a clear, colorless oil. H NMR: δ 7.29 (d, 2H, J )
2.4), 6.80 (t, 1H, J ) 2.4), 6.68 (d, 4H, J ) 2.20), 6.58 (d, 8H, J )
2.4), 6.57 (t, 2H, J ) 2.2), 6.42 (t, 4H, J ) 2.4), 5.90 (ddt, 8H, J )
17.2, 10.3, 6.6), 5.16 (ddt, 8H, J ) 17.2, 1.8, 1.5), 5.10 (ddt, 8H, J )
10.3, 1.8, 1.3), 5.01 (s, 4H), 4.96 (s, 8H), 4.00 (t, 16H, J ) 6.8), 3.91
(s, 3H), 2.53 (tddd, 16H, J ) 6.8, 6.6, 1.5, 1.3). 13C NMR: 166.8,
160.3, 160.2, 159.7, 139.0, 138.8, 134.4, 132.1, 117.1, 108.4, 107.2,
106.4, 105.9, 101.7, 101.0, 70.2, 70.1, 67.3, 52.3, 33.6. IR (neat): 3078,
2979, 1721, 1641, 1597, 1166, 1054, 994, 917. MS (MALDI-TOF):
m/z 1355.0 (M + Na+). SEC (toluene) Calcd Mw ) 1338. Anal. Calcd
for C82H92O16: C, 73.85; H, 6.95. Found: C, 74.16; H, 7.08.
1
of [G-2]-CO2Me as a clear, colorless oil. H NMR: δ 7.29 (d, 2H, J
3,5-Bis-{3,5-bis[3,5-bis(3-buten-1-oxy)benzyloxy]benzyloxy}-
benzoic Acid ([G-3]-CO2H) (8). A 40% (w/w) aqueous KOH solution
was made by dissolving 2.48 g (38.0 mmol) of KOH in 4 mL of water.
This solution was added to a solution of 4.85 g (3.64 mmol) of [G-3]-
CO2Me dissolved in 25 mL of THF. Ethanol (20 mL) was added to
make the reaction mixture homogeneous. The reaction was stirred at
reflux until no starting material remained by TLC. The reaction was
stopped by adding concentrated aqueous HCl to make the reaction
mixture acidic (pH < 3). The solvents were removed under reduced
pressure. The remaining residue was partitioned between 100 mL of
water and 100 mL of CH2Cl2. The aqueous layer was extracted with
) 2.4), 6.79 (t, 1H, J ) 2.4), 6.58 (d, 4H, J ) 2.2), 6.43 (t, 2H, J )
2.2), 5.90 (ddt, 4H, J ) 17.2, 10.3, 6.8), 5.18 (ddt, 4H, J ) 17.2, 1.8,
1.5), 5.11 (ddt, 4H, J ) 10.3, 1.8, 1.3), 4.99 (s, 4H), 4.01 (t, 8H, J )
6.6), 3.91 (s, 3H), 2.54 (dtdd, 8H, J ) 6.8, 6.6, 1.5, 1.3). 13C NMR:
166.8, 160.3, 159.7, 138.7, 134.4, 132.0, 117.1, 108.4, 107.2, 105.9,
101.1, 70.2, 67.3, 52.3, 33.6. IR (neat): 3094, 2980, 1734, 1656, 1597,
1174, 1062, 997, 924. MS (FD): m/z 628.3 (M+). Anal. Calcd for
C38H44O8: C, 72.59; H, 7.05. Found: C, 72.64; H, 6.84.
3,5-Bis[3,5-bis(3-buten-1-oxy)benzyloxy]benzyl Alcohol ([G-2]-
OH). To a suspension of 6.29 g (166 mmol) of LAH in 250 mL of
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13515

A R T I C L E S
Zimmerman et al.
CH2Cl2 (3 × 100 mL). The combined organic layers were washed with
water, dried over sodium sulfate, and filtered. The resulting filtrate was
concentrated under reduced pressure. The product was loaded onto a
silica gel column (10 × 12 cm) and eluted with 1% acetic acid/30%
EtOAc/69% PE. The product was dried under vacuum (0.15 mmHg,
100 °C) overnight to afford 4.08 g (85%) of 8 as a beige oil. 1H NMR:
δ 7.35 (d, 2H, J ) 2.4), 6.84 (t, 1H, J ) 2.4), 6.68 (d, 4H, J ) 2.2),
6.57 (d, 8H, J ) 2.2), 6.56 (t, 2H, J ) 2.2), 6.42 (t, 4H, J ) 2.2), 5.89
(ddt, 8H, J ) 17.2, 10.3, 6.6), 5.16 (ddt, 8H, J ) 17.2, 1.8, 1.5), 5.10
(ddt, 8H, J ) 10.3, 1.8, 1.3), 5.01 (s, 4H), 4.96 (s, 8H), 4.00 (t, 16H,
J ) 6.8), 2.52 (tddd, 16H, J ) 6.8, 6.6, 1.5, 1.3). 13C NMR: 171.2,
160.3, 160.2, 159.8, 139.0, 138.7, 134.4, 131.1, 117.1, 108.9, 108.0,
106.5, 105.9, 101.8, 101.0, 70.2, 70.1, 67.3, 33.6. IR (neat): 3196,
3078, 2977, 1692, 1642, 1594, 1163, 1055, 993, 917. MS (MALDI-
TOF): m/z 1342.7 (M + Na+). Anal. Calcd for C81H90O16: C, 73.73;
H, 6.87. Found: C, 73.86; H, 6.80.
the reaction was complete by TLC. The reaction was stopped by
removing the THF under reduced pressure. The resulting aqueous layer
was extracted with CHCl3 (3 × 25 mL). The combined organic layers
were washed with 1 M aqueous HCl, washed with water, and
concentrated under reduced pressure. The product was dried under
vacuum (0.2 mmHg, 70 °C) to afford 41.5 mg (43%) of 10 as a beige
powder. 1H NMR: δ 7.23 (bs, 16H), 6.51 (bs, 152H), 5.86 (bs, ∼4H),
5.60 (bs, ∼60H), 5.12 (bs, ∼8H), 4.86 (bs, 96H), 3.92 (bs, 128H),
2.45 (bs, 128H). MS (MALDI-TOF): m/z 9679.8 (M + H+ - 31C2H4
- C44H22N4). SEC (toluene) Calcd Mw ) 4656. Anal. Calcd for
C588H600O128: C, 72.70; H, 6.22; N, 0.00; Ru, 0.00. Found: C, 70.83;
H, 6.19; N, 0.00; Ru, 0.00.
5,10,15,20-Tetrakis(3′,5′-pyrimidyl)porphyrin (12). Pyrimidine-
5-carboxaldehyde (245 mg, 2.27 mmol) and 160 µL (2.31 mmol) of
freshly distilled pyrrole were added to 12 mL of refluxing propionic
acid. The solution quickly turned black and was maintained at reflux
for 1 h. The reaction mixture was cooled, and the propionic acid was
removed under reduced pressure to give a tarry residue. The resulting
residue was partitioned between water and CHCl3. The aqueous layer
was extracted with CHCl3. The organic layers were combined and
concentrated under reduced pressure. The resulting residue was loaded
onto a silica gel column and eluted with 50% CHCl3/50% acetone.
The second, more intense band was collected and concentrated under
reduced pressure. The product was dried under vacuum to afford 2.8
5,10,15,20-Tetrakis{3,5-bis[3,5-bis(3,5-bis(3,5-bis(3-buten-1-oxy)-
benzyloxy)benzyloxy)benzoyloxy]phenyl}porphyrin ([G-3]8-T(3,5-
OHPh)P) (7). [G-3]-CO2H (8) (283 mg, 215 µmol), 13.2 mg (17.8
µmol) of 5,10,15,20-tetrakis(3′,5′-dihydroxyphenyl)porphyrin (H2T(3,5-
OHPh)P) 6, 519 mg (2.5 mmol) of DCC, 78.6 mg (643 µmol) of
DMAP, and 118 mg (621 µmol) of p-toluenesulfonic acid were
dissolved in 15 mL of THF. The reaction was stirred overnight at room
temperature. The reaction mixture was vacuum filtered to remove the
DCU formed. The filtrate was concentrated under reduced pressure and
dissolved in 15 mL of CH2Cl2. To this solution, 510 mg (2.5 mmol) of
DCC, 75.8 mg (620 µmol) of DMAP, and 117 mg (615 µmol) of
p-toluenesulfonic acid were added. This solution was stirred at room
temperature and monitored by TLC and SEC. The reaction was
determined to be complete after 3 days. The reaction mixture was
vacuum filtered to remove the DCU, and the resulting filtrate was
concentrated under reduced pressure. The remaining residue was loaded
onto a silica gel column (4 × 15 cm) and eluted with 30% EtOAc/PE.
The product was further purified by loading it onto a SEC column and
eluting it with toluene. The product was dried overnight under vacuum
(1.0 mmHg, 98 °C) to afford 107 mg (54%) of 7 as a deep red oil. 1H
NMR (CD2Cl2): δ 9.22 (s, 8H), 8.15 (d, 8H, J ) 2.1), 7.75 (t, 4H, J
) 2.1), 7.54 (d, 16H, J ) 2.1), 6.81 (t, 8H, J ) 2.1), 6.64 (d, 32H, J
) 2.1), 6.48 (m, 80H), 6.31 (t, 32H, J ) 2.1), 5.81 (ddt, 64H, J )
17.2, 10.2, 6.7), 5.07 (ddt, 64H, J ) 17.2, 1.8, 1.5), 5.01 (ddt, 64H, J
) 10.2, 1.8, 1.2), 4.98 (s, 32H), 4.87 (s, 64H), 3.88 (t, 128H, J ) 6.7),
2.42 (tddd, 128H, J ) 6.7, 6.7, 1.5, 1.2), -2.86 (s, 2H). 13C NMR
(CD2Cl2): 165.1, 160.7, 160.5, 160.4, 150.3, 144.2, 139.6, 139.3, 135.0,
131.6, 126.3, 18.9, 17.0, 109.3, 108.1, 106.7, 106.2, 101.9, 101.0, 70.5,
70.3, 67.6, 33.9. MS (MALDI-TOF): m/z 1153.4 (M + H+). SEC
(toluene) Calcd Mw ) 8155. UV/vis (CH2Cl2) λmax ) 275.0, 420.5
(Soret). Anal. Calcd for C692H734N4O128: C, 74.51; H, 6.63; N, 0.50.
Found: C, 74.56; H, 6.41; N, 0.48.
1
mg (1%) of 12 as a purple solid. H NMR: δ 9.72 (s, 4H), 9.59 (s,
8H), 8.94 (s, 8H), -2.87 (s, 2H). MS (FAB): m/z 623.4 (M + H+).
UV/vis (chloroform) λmax ) 420.0 (Soret), 515.5, 549.5, 590.5, 646.0.
Mixed Porphyrin Condensation. Pyrrole (12.5 mL, 180 mmol) was
added dropwise to a refluxing solution of benzaldehyde (14.0 mL, 138
mmol) and pyridine-4-carboxaldehyde (6.0 mL, 63.0 mmol) in propionic
acid (500 mL). The solution quickly changed color from yellow to
black. The solution was stirred at reflux for 1.5 h. The reaction flask
was cooled in an ice/water bath. Diethylene glycol (350 mL) was added
to the cooled reaction solution. The solution was refrigerated overnight.
The purple precipitate that formed was isolated by vacuum filtration.
The solid was washed with MeOH. The crude product mixture (1.25
g) was loaded on a silica gel column and eluted with CHCl3. The
following compounds were isolated from the column in this order:
5,10,15,20-tetrakis(phenyl)porphyrin (13), 5-(4′-pyridyl)-10,15,20-tris-
(phenyl)porphyrin (20), cis-5,10-bis(4′-pyridyl)-15,20-bis(phenyl)por-
phyrin (19), trans-5,15-bis(4′-pyridyl)-10,20-bis(phenyl)porphyrin (18),
and a mixture of 5,10,15-tris(4′-pyridyl)-20-(phenyl)porphyrin (17) and
5,10,15,20-tetrakis(4′-pyridyl)porphyrin (16). Porphyrins 16 and 17 were
isolated by loading the mixture of the two from the first column onto
a second silica gel column. The column was eluted with 2% MeOH/
98% CHCl3 with 17 eluting from the column first.
5,10,15-Tris(4′-pyridyl)-20-(phenyl)porphyrin (17). Compound 17
was isolated in 2.3% yield from the mixed porphyrin condensation
reaction as a reddish-purple solid. 1H NMR (CDCl3): δ 9.08 (m, 6H),
8.94 (d, 2H, J ) 4.8), 8.88 (s, 4H), 8.84 (d, 2H, J ) 4.8), 8.23 (m,
2H), 8.18 (m, 6H), 7.81 (m, 3H), -2.86 (s, 2H). MS (high-resolution
FAB): m/z 618.24 (M + H+). UV/vis (CHCl3) λmax ) 417.5 (Soret),
513.5, 548.0, 589.0, 645.0.
Cross-linked [G-3]8-T(3,5-OHPh)P (9). To a solution of 497 mg
(44.6 µmol) of 7 in 4.4 L of benzene was added 99.6 mg (120 µmol)
of the Grubbs’ catalyst 2. The reaction was stirred at room temperature
for 22 h. The benzene was removed under reduced pressure. The crude
product was loaded onto a silica gel plug (4 × 6 cm) and eluted with
400 mL of 50% PE/50% CH2Cl2 and 400 mL of 5% EtOAc/CH2Cl2.
The EtOAc/CH2Cl2 solution was concentrated under reduced pressure.
The product was dried overnight under vacuum (2 mmHg, 85 °C) to
afford 409 mg (89%) of 9 as a dark red powder. 1H NMR (d8-
toluene): δ 9.17 (bs, 8H), 8.15 (bs, 8H), 7.70 (bs, 20H), 6.59 (bs, 152H),
5.78 (bs, ∼4H), 5.51 (bs, ∼60H), 5.04 (bs, ∼8H), 4.82 (bs, 96H), 3.70
(bs, 128H), 2.33 (bs, 128H). MS (MALDI-TOF): m/z 10 257.1 (M +
trans-5,15-Bis(4′-pyridyl)-10,20-bis(phenyl)porphyrin (18). Com-
pound 18 was isolated in 0.7% yield from the mixed porphyrin
1
condensation reaction as a reddish-purple solid. H NMR (CDCl3): δ
9.04 (m, 4H), 8.91 (d, 4H, J ) 4.6), 8.81 (d, 4H, J ) 4.6) 8.22 (m,
4H), 8.17 (m, 4H), 7.78 (m, 6H), -2.85 (s, 2H). MS (high-resolution
FAB): m/z 617.24 (M + H+). UV/vis (CHCl3) λmax ) 418.0 (Soret),
514.0, 548.5, 590.5, 645.0.
H+ - 32C2H4), 10 285.2 (M + H+ - 31C2H4), 10 312.3 (M + H+
30C2H4). SEC (toluene) Calcd Mw ) 5509. UV/vis (CH2Cl2) λmax
276.0, 421.0 (Soret).
-
)
cis-5,10-Bis(4′-pyridyl)-15,20-bis(phenyl)porphyrin (19). Com-
pound 19 was isolated in 1.2% yield from the mixed porphyrin
1
condensation reaction as a reddish-purple solid. H NMR (CDCl3): δ
Imprinted [G-3]8-T(3,5-OHPh)P (10). To a solution of 102 mg
(9.88 µmol) of 9 dissolved in 15 mL of THF was added 10 mL of 2.5
M aqueous KOH. The reaction was stirred vigorously at reflux until
9.05 (m, 4H), 8.91 (d, 2H, J ) 4.9), 8.87 (s, 2H), 8.84 (s, 2H), 8.81 (d,
2H, J ) 4.9), 8.22 (m, 4H), 8.17 (m, 4H), 7.77 (m, 6H), -2.85 (s,
9
13516 J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003

Molecular Imprinting Inside Dendrimers
A R T I C L E S
2H). MS (high-resolution FAB); m/z 617.24 (M + H+). UV/vis (CHCl3)
λmax ) 418.0 (Soret), 514.0, 549.0, 590.0, 645.0.
stirred at reflux until complete as indicated by TLC. The reaction was
stopped by removing the THF under reduced pressure. The resulting
residue was loaded onto a silica gel column (3 × 17 cm) and eluted
with 1% acetic acid/30% EtOAc/69% PE. The product was further
purified by loading it onto a series of preparatory SEC columns and
eluting it with toluene. The product was dried overnight under vacuum
(0.2 mmHg, 75 °C) to afford 450 mg (90%) of 21 as a dark red oil. 1H
NMR (CD2Cl2): δ 8.97 (s, 8H), 8.31 (m, 8H), 7.69 (m, 8H), 7.62 (d,
8H, J ) 2.3), 6.95 (t, 4H, J ) 2.3), 6.76 (d, 16H, J ) 2.3), 6.59 (t, 8H,
J ) 2.3), 6.58 (d, 32H, J ) 2.3), 6.39 (t, 16H, J ) 2.3), 5.88 (ddt,
32H, J ) 17.1, 10.4, 6.7), 5.13 (m, 48H), 5.07 (ddt, 32H, J ) 10.4,
1.8, 1.2), 5.01 (s, 32H), 3.99 (t, 64H, J ) 6.7), 2.50 (tddd, 64H, J )
6.7, 6.7, 1.8, 1.2), -2.80 (s, 2H). MS (MALDI-TOF): m/z 5884.0 (M
+ H+).
Cross-linking and Hydrolysis of [G-3]-T(4-OHPh)P. To a solution
of 50.0 mg (8.49 µmol) of 21 in 1.15 L of benzene was added 8.87 mg
(10.8 µmol) of the Grubbs’ catalyst. The reaction was stirred at room
temperature for 18 h. The benzene was removed under reduced pressure.
The crude product was loaded onto a silica gel plug (4 × 6 cm) and
eluted with 250 mL of 50% PE/50% CH2Cl2 and 300 mL of 5% EtOAc/
95% CH2Cl2. The EtOAc/CH2Cl2 solution was concentrated under
reduced pressure. The product was dried overnight under vacuum (0.2
mmHg, 80 °C) to afford 36.5 mg (79%) of a dark red powder. 1H NMR
(CD2Cl2): δ 8.98 (bs, 8H), 8.30 (bs, 8H), 7.60 (bs, 16H), 6.96 (bs,
4H), 6.45 (bs, 72H), 5.87 (bs, ∼2H), 5.65 (bs, ∼30H), 5.04 (bs, ∼52H),
3.99 (bs, 64H), 2.45 (bs, 64H), -2.74 (bs, 2H). MS (MALDI-TOF):
m/z 5440.9 (M + H+ - 16C2H4), 5468.9 (M + H+ - 15C2H4), 5498.2
(M + H+ - 14C2H4).
A 25% (w/w) aqueous KOH solution was made by dissolving 48.3
mg (740 µmol) of KOH in 145 µL of water. This solution was added
to a solution of 36.5 mg (6.68 µmol) of cross-linked [G-3]-T(4-OHPh)P
dissolved in 2.0 mL of THF. Ethanol (145 µL) was added to make the
reaction mixture homogeneous. The reaction was stirred at 85 °C until
no starting material remained by TLC. The reaction mixture was
allowed to settle such that two layers formed. The organic layer was
decanted off, and the remaining aqueous layer was washed two more
times with THF. The THF extracts were combined and concentrated
under reduced pressure. The remaining residue was partitioned between
5.0 mL of 2.5 M aqueous KOH and 5.0 mL of CH2Cl2. The aqueous
layer was extracted with CH2Cl2 (3 × 5 mL). The organic layers were
combined and washed with 2.5 M aqueous KOH, 1 M aqueous HCl
(twice), and water. The organic layer was dried over sodium sulfate
and filtered. The filtrate was concentrated under reduced pressure and
dried under vacuum (0.1 mmHg, 75 °C) overnight to afford 24.0 mg
of a beige powder. 1H NMR (CD2Cl2): δ 7.23 (bs, 8H), 6.74 (bs, 4H),
6.45 (bs, 72H), 5.85 (bs, ∼2H), 5.64 (bs, ∼30H), 5.10 (bs, ∼4H), 4.99
(bs, 48H), 3.99 (bs, 64H), 2.45 (bs, 64H). MS (MALDI-TOF): m/z
4855.9 (M + Na+ - 16C2H4 - C44H22N4), 4883.3 (M + Na+ - 15C2H4
- C44H22N4), 4911.3 (M + Na+ - 14C2H4 - C44H22N4), and peaks
corresponding to the loss of one and two dendrons.
5-(4′-Pyridyl)-10,15,20-tris(phenyl)porphyrin (20). Compound 20
was isolated in 2.9% yield from the mixed porphyrin condensation
reaction as a reddish-purple solid. 1H NMR (CDCl3): δ 9.04 (m, 2H),
8.90 (d, 2H, J ) 4.9), 8.87 (s, 4H), 8.80 (d, 2H, J ) 4.9), 8.22 (m,
6H), 8.17 (m, 2H), 7.77 (m, 9H), -2.80 (s, 2H). MS (FAB): m/z 616.3
(M + H+). UV/vis (CHCl3) λmax ) 418.0 (Soret), 515.0, 549.0, 589.5,
645.0. Anal. Calcd for C43H29N5: C, 83.88; H, 4.75; N, 11.37. Found:
C, 83.13; H, 4.57; N, 11.13.
Imprinted Octakis(ethyl ester) [G-3]8-T(3,5-OHPh)P (10-(CO2Et)8).
To a solution of 409 mg (39.7 µmol) of cross-linked [G-3]8-T(3,5-
OHPh)P dendrimer (9) in 10 mL of toluene were added 2.0 mL of
EtOH and 88.0 mg (637 µmol) of potassium carbonate. The suspension
was stirred at reflux until the reaction was complete by TLC. The
toluene and EtOH were removed under reduced pressure. The resulting
residue was partitioned between 100 mL of EtOAc and 30 mL of water.
The pH was brought to 7 by adding 1 M aqueous HCl. The aqueous
layer was further extracted with EtOAc (3 × 100 mL). The combined
organic layers were washed with an equal volume of water, dried over
sodium sulfate, and filtered. The filtrate was concentrated under reduced
pressure. The crude product was loaded onto a silica gel plug (4 × 6
cm) and eluted with 5% EtOAc/95% CH2Cl2. The EtOAc/CH2Cl2
solution was concentrated under reduced pressure. The product was
dried overnight under vacuum (0.2 mmHg, 75 °C) to afford 116 mg
1
(30%) of 10-(CO2Et)8 as a beige powder. H NMR (toluene-d8): δ
7.47 (bs, 16H), 6.62 (bs, 152H), 5.75 (bs, ∼4H), 5.48 (bs, ∼60H),
5.03 (bs, ∼8H), 4.78 (bs, 96H), 4.17 (bs, 16H), 3.67 (bs, 128H), 2.31
(bs, 128H), 1.12 (bs, 24H). MS (MALDI-TOF): m/z 9906.8 (M +
Na+ - 32C2H4 - C28H18N4), 9934.6 (M + Na+ - 31C2H4 - C28H18N4),
9963.2 (M + Na+ - 30C2H4 - C28H18N4), 9990.1 (M + Na+ - 29C2H4
- C28H18N4). SEC (toluene) Calcd Mw ) 4656.
3,5-Bis{3,5-bis[3,5-bis(3-buten-1-oxy)benzyloxy]benzyloxy}-
benzoyl Chloride ([G-3]-COCl) (23). [G-3]-CO2H (8) (351 mg, 266
µmol) was dissolved in 10 mL of THF. To this solution were added
10 µL (129 µmol) of DMF and 50 µL (537 µmol) of oxalyl chloride.
The reaction mixture was stirred at room temperature for 30 min. The
reaction mixture was heated to 50 °C and stirred at this elevated
temperature for 15 min. An aliquot of the reaction mixture was placed
in dry MeOH (to form the methyl ester), and TLC of this solution
showed that the starting material has disappeared. The reaction was
stopped by removing the solvent (and the HCl formed) under reduced
pressure. The crude material was loaded onto a silica gel column (2.5
× 20 cm) and eluted with 30% EtOAc/70% PE. This afforded 227 mg
(64%) of 23 as clear, colorless oil. It was discovered that the product
partially hydrolyzed on silica gel as the remainder of the material was
isolated as the starting material. Subsequent formation of acyl chloride
dendrons followed this procedure with the modification that once the
reaction was determined to be complete by TLC, the solvent was
removed under reduced pressure, and the resulting material was dried
gently with heat under vacuum until it was considered dry and the
3,5-Bis{3,5-bis[3,5-bis(3-buten-1-oxy)benzyloxy]benzyloxy}-
benzyl Alcohol ([G-3]-OH). To a suspension of 886 mg (23.4 mmol)
of LAH in 50 mL of THF cooled to 0 °C with an ice/water bath was
added dropwise a solution of 18.4 g (13.8 mmol) of [G-3]-CO2Me in
150 mL of THF. The suspension was allowed to warm to room
temperature and stirred until no starting material remained by TLC.
The reaction was stopped by quenching the excess hydride with 100
mL of water and neutralizing the pH with 20 mL of 1 M aqueous HCl.
The THF was removed under reduced pressure, and the resulting
aqueous layer was extracted with Et2O (3 × 400 mL). The combined
organic layers were washed with water, dried over sodium sulfate,
filtered, and concentrated under reduced pressure. Half of the crude
product was purified at a time by column chromatography. The crude
product was loaded onto a silica gel column (10 × 20 cm) and eluted
with 30% EtOAc/70% PE. The product was dried under vacuum (0.02
mmHg, 75 °C) overnight to afford 16.3 g (90%) of [G-3]-OH as a
1
excess HCl was removed. H NMR: δ 7.34 (d, 2H, J ) 2.2), 6.88 (t,
1H, J ) 2.2), 6.67 (d, 4H, J ) 2.2), 6.59 (t, 2H, J ) 2.2), 6.58 (d, 8H,
J ) 2.2), 6.43 (t, 4H, J ) 2.2), 5.90 (ddt, 8H, J ) 17.2, 10.3, 6.6),
5.17 (ddt, 8H, J ) 17.2, 1.8, 1.5), 5.11 (ddt, 8H, J ) 10.3, 1.8, 1.3),
5.01 (s, 4H), 4.97 (s, 8H), 4.00 (t, 16H, J ) 6.8), 2.54 (tddd, 16H, J )
6.8, 6.6, 1.5, 1.3). IR (neat): 3078, 2980, 1760, 1642, 1593, 1163,
1055, 992, 919.
5,10,15,20-Tetrakis{4-[3,5-bis(3,5-bis(3,5-bis(3-buten-1-oxy)-
benzyloxy)benzyloxy)benzoyloxy]phenyl}porphyrin ([G-3]-T(4-
OHPh)P) (21). [G-3]-COCl (23) (684 mg, 511 µmol) was dissolved
in 15 mL of THF and added to a flask containing 57.8 mg (85.2 µmol)
of H2T(4-OHPh)P (22) and 251 mg (2.05 mmol) of DMAP in 10 mL
of THF. The reaction was stirred at reflux for 17 h. TLC showed the
reaction was progressing very slowly. Additional DMAP (252 mg, 2.07
mmol) was added to the reaction mixture. The reaction mixture was
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13517

A R T I C L E S
Zimmerman et al.
1
clear, colorless oil. H NMR: δ 6.66 (d, 4H, J ) 2.20), 6.60 (d, 2H,
J ) 2.2), 6.56 (d, 8H, J ) 2.2), 6.54 (t, 2H, J ) 2.2), 6.52 (t, 1H, J )
2.2), 6.41 (t, 4H, J ) 2.2), 5.89 (ddt, 8H, J ) 17.2, 10.3, 6.6), 5.16
(ddt, 8H, J ) 17.2, 1.8, 1.5), 5.10 (ddt, 8H, J ) 10.3, 1.8, 1.3), 4.97
(s, 4H), 4.95 (s, 8H), 4.62 (d, 2H, J ) 6.0), 3.93 (t, 16H, J ) 6.8),
2.52 (tddd, 16H, J ) 6.8, 6.6, 1.5, 1.3), 1.72 (t, 1H, J ) 6.0). 13C
NMR: 160.3, 160.1, 160.0, 143.5, 139.3, 139.1, 134.4, 117.1, 106.3,
105.9, 105.7, 101.6, 101.3, 101.0, 70.1, 70.0, 67.3, 65.3, 33.6. IR
(neat): 3531 (bs), 3077, 2979, 1641, 1591, 1140, 1058, 994, 916. MS
(M + Na+). Anal. Calcd for C169H186O32: C, 74.37; H, 6.87. Found:
C, 74.45; H, 6.86.
5,10,15,20-Tetrakis{4-[3,5-bis(3,5-bis(3,5-bis(3,5-bis(3-buten-1-
oxy)benzyloxy)benzyloxy)benzyloxy)benzoyloxy]phenyl}porphy-
rin ([G-4]-T(4-OHPh)P) (25). [G-4]-CO2H (24) (296 mg, 109 µmol),
11.8 mg (17.4 µmol) of H2T(4-OHPh)P (22), 13.4 mg (110 µmol) of
DMAP, and 20.3 mg (107 µmol) of p-toluenesulfonic acid were
dissolved in 10 mL of THF. In a separate flask, 86.5 mg (419 µmol)
of DCC was dissolved in 5 mL of THF. This solution was added to
the reaction mixture. The reaction was stirred overnight at room
temperature. The reaction mixture was vacuum filtered to remove the
DCU formed. The filtrate was concentrated under reduced pressure and
redissolved in 10 mL of CH2Cl2. To this solution were added 286 mg
(105 µmol) of 24, 13.2 mg (108 µmol) of DMAP, and 20.2 mg (106
µmol) of p-toluenesulfonic acid. In a separate flask, 85.5 mg (414 µmol)
of DCC was dissolved in 5 mL of CH2Cl2. This solution was added to
the reaction mixture. The reaction mixture was stirred at room
temperature until complete as indicated by both TLC and SEC. The
reaction was stopped by vacuum filtering the mixture and condensing
the filtrate under reduced pressure. The resulting residue was loaded
onto a silica gel column (4 × 17 cm) and eluted with 5% EtOAc/95%
CH2Cl2. The crude product was further purified by loading it onto a
large preparatory (7 × 120 cm) SEC column and eluting it with toluene.
The product was dried under vacuum (0.2 mmHg, 75 °C) overnight to
(MALDI-TOF): m/z 1328.3 (M + Na+). SEC (toluene) Calcd Mw
1324. Anal. Calcd for C81H92O15: C, 74.52; H, 7.10. Found: C, 74.52;
)
H, 7.05.
3,5-Bis{3,5-bis[3,5-bis(3,5-bis(3-buten-1-oxy)benzyloxy)benzyloxy]-
benzyloxy} Benzoic Acid Methyl Ester ([G-4]-CO2Me). To a mixture
of 1.13 g (862 µmol) of [G-3]-OH, 61.6 mg (366 µmol) of methyl-
3,5-dihydroxybenzoate, and 241 mg (920 µmol) of PPh3 in 10 mL of
THF cooled to 0 °C with an ice/water bath was added dropwise a
solution of 197 mg (1.13 mmol) of DEAD in 10 mL of THF. The
reaction was allowed to warm to room temperature and stirred until
the reaction was complete as indicated by TLC. The reaction was
stopped by adding 20 mL of water and removing the THF under reduced
pressure. The resulting aqueous layer was extracted with Et2O (3 × 75
mL). The combined organic layers were washed water, dried over
sodium sulfate, filtered, and concentrated under reduced pressure. The
resulting residue was loaded onto a silica gel column (4 × 20 cm) and
eluted with 30% EtOAc/70% PE. The product was dried under vacuum
(0.05 mmHg, 80 °C) overnight to afford 889 mg (89%) of [G-4]-CO2-
Me as a clear, colorless oil. 1H NMR: δ 7.30 (d, 2H, J2,4 ) 2.0), 6.81
(t, 1H, J ) 2.0), 6.68 (m, 12H), 6.57 (d, 16H, J ) 2.0), 6.56 (m, 6H),
6.42 (t, 8H, J ) 2.0), 5.89 (ddt, 16H, J ) 17.2, 10.3, 6.6), 5.16 (ddt,
16H, J ) 17.2, 1.8, 1.5), 5.10 (ddt, 16H, J ) 10.3, 1.8, 1.3), 5.01 (s,
4H), 4.97 (s, 8H), 4.95 (s, 16H), 3.99 (t, 32H, J ) 6.8), 3.90 (s, 3H),
2.52 (tddd, 32H, J ) 6.8, 6.6, 1.5 1.3). 13C NMR: 166.7, 160.3, 160.1,
159.8, 139.2, 139.1, 138.9, 134.4, 132.1, 117.1, 108.5, 107.1, 106.4,
106.0, 101.7, 101.6, 101.0, 70.2, 70.1, 67.3, 52.3, 33.6. IR (neat): 3077,
2976, 1723, 1642, 1595, 1164, 1052, 992, 916. MS (MALDI-TOF):
m/z 2766.7 (M + Na+). SEC (toluene) Calcd Mw ) 2699. Anal. Calcd
for C170H188O32: C, 74.43; H, 6.91. Found: C, 74.46; H, 6.82.
3,5-Bis{3,5-bis[3,5-bis(3,5-bis(3-buten-1-oxy)benzyloxy)benzyloxy]-
benzyloxy} Benzoic Acid ([G-4]-CO2H) (24). A 40% (w/w) aqueous
KOH solution was made by dissolving 1.20 g (18.4 mmol) of KOH in
1.8 mL of water. This solution was added to a solution of 4.81 g (1.75
mmol) of [G-4]-CO2Me dissolved in 25 mL of THF. Ethanol (13 mL)
was added to make the reaction mixture homogeneous. The reaction
was stirred at reflux until no starting material remained by TLC. The
reaction flask was removed from the heat and cooled to 0 °C in an
ice/water bath. Water (5 mL) was added to the reaction mixture. The
reaction was stopped by adding 1 M aqueous HCl until the reaction
mixture was acidic (pH < 3). The solvents were removed under reduced
pressure. The remaining residue was partitioned between 60 mL of
water and 100 mL of CH2Cl2. The aqueous layer was extracted with
CH2Cl2 (3 × 100 mL). The combined organic layers were washed with
water, dried over sodium sulfate, and filtered. The resulting filtrate was
concentrated under reduced pressure. The remaining residue was loaded
onto a silica gel column (8 × 30 cm) and eluted with 3.0 L of 5%
EtOAc/95% CH2Cl2 and 2.0 L of 10% EtOAc/90% CH2Cl2. The product
was dried under vacuum (0.05 mmHg, 70 °C) overnight to afford 3.99
1
afford 92.1 mg (46%) of 25 as a deep red oil. H NMR (CD2Cl2): δ
9.01 (s, 8H), 8.34 (m, 8H), 7.73 (m, 8H), 7.68 (d, 8H, J ) 2.1), 7.02
(t, 4H, J ) 2.1), 6.80 (d, 16H, J ) 2.1), 6.73 (d, 32H, J ) 2.3), 6.64
(t, 8H, J ) 2.1), 6.58 (d, 64H, J ) 2.3), 6.57 (t, 16H, J ) 2.3), 6.41
(t, 32H, J ) 2.3), 5.90 (ddt, 64H, J ) 17.2, 10.4, 6.7), 5.16 (ddt, 64H,
J ) 17.2, 1.8, 1.5), 5.09 (ddt, 64H, J ) 10.4, 1.8, 1.2), 5.05 (s, 32H),
4.99 (s, 16H), 4.98 (s, 64H), 3.98 (t, 128H, J ) 6.7), 2.51 (tddd, 128H,
J ) 6.7, 6.7, 1.5, 1.2), -2.78 (s, 2H). MS (MALDI-TOF): m/z 11 546.3
(M + Na+). SEC (toluene) Calcd Mw ) 10 793.
Cross-linked [G-4]-T(4-OHPh)P (26). To a solution of 90.0 mg
(7.81 µmol) of 81 in 2.0 L of benzene was added 20.5 mg (24.9 µmol)
of the Grubbs’ catalyst. The reaction was stirred at room temperature
for 20 h. The benzene was removed under reduced pressure. The crude
product was loaded onto a silica gel plug (4 × 5 cm) and eluted with
250 mL of 50% PE/50% CH2Cl2 and 250 mL of 5% EtOAc/95% CH2-
Cl2. The EtOAc/CH2Cl2 solution was concentrated under reduced
pressure. The product was dried under vacuum (0.05 mmHg, 65 °C)
overnight to afford 73.2 mg (88%) of 26 as a dark red powder. MS
(MALDI-TOF): m/z 10 624.4 (M + H+ - 32C2H4), 10 652.7 (M +
H+ - 31C2H4).
Imprinted [G-4]-T(4-OHPh)P (27). To a solution of 73.2 mg (6.87
µmol) of cross-linked [G-4]-T(4-OHPh)P (26) dissolved in 100 mL of
THF was added 25 mL of 2.5 M aqueous KOH. The reaction was stirred
vigorously at reflux until the reaction was complete by TLC. The
reaction was stopped by removing the THF under reduced pressure.
The resulting aqueous layer was extracted with CH2Cl2 (3 × 25 mL).
The organic layers were combined and washed with an equal volume
of 2.5 M aqueous KOH and water. The organic layer was stirred
vigorously with an equal volume of 1 M aqueous HCl for 3 h at room
temperature. The organic layer was concentrated under reduced pressure.
The product was dried under vacuum (0.2 mmHg, 75 °C) overnight to
afford 33.2 mg (48%) of 27 as a beige powder. MS (MALDI-TOF):
m/z 10 040.9 (M + Na+ - 32C2H4 - C44H22N4), 10 069.1 (M + Na+
- 31C2H4 - C44H22N4).
1
g (83%) of 24 as a beige oil. H NMR: δ 7.38 (d, 2H, J ) 2.4), 6.87
(t, 1H, J ) 2.4), 6.71 (d, 4H, J′ ) 2.1), 6.70 (d, 8H, J ) 2.1), 6.60 (t,
2H, J ) 2.1), 6.59 (d, 16H, J ) 2.1), 6.57 (t, 4H, J ) 2.1), 6.43 (t, 8H,
J ) 2.1), 5.90 (ddt, 16H, J ) 17.2, 10.3, 6.7), 5.18 (ddt, 16H, J )
17.2, 1.7, 1.5), 5.12 (ddt, 16H, J ) 10.3, 1.7, 1.3), 5.03 (s, 4H), 4.99
(s, 8H), 4.96 (s, 16H), 4.00 (t, 32H, J ) 6.7), 2.54 (tddd, 32H, J )
6.7, 6.7, 1.5, 1.3). 13C NMR: 171.4, 160.5, 160.4, 160.1, 139.4, 139.3,
139.1, 134.7, 131.4, 117.3, 109.2, 108.2, 106.8, 106.7, 106.2, 102.0,
101.9, 101.2, 70.5, 70.3, 67.5, 33.8. MS (MALDI-TOF): m/z 2753.6
Acknowledgment. This work was funded by the NIH
(GM61067 and HL 25934) and U.S. Army Research Office
(DAAG55-97-0126). I.Z. is grateful to the Arnold and Mabel
Beckman Foundation for a Beckman Fellowship.
JA0357240
9
13518 J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003