Dynamic combinatorial libraries constructed on polymer scaffolds

Article abstract of DOI:10.1021/ol8011309

(Figure Presented) Functionalized polymers were prepared by grafting acylhydrazides onto a polyvinylbenzaldehyde scaffold through reversible hydrazone linkages. The dynamic nature of these linkages allows the functionalized polymers to exchange and reshuffle their appendages, and the resultant mixture of polymers can be considered as a dynamic combinatorial library constructed upon a polymer scaffold. The dynamic nature of these functionalized polymers was demonstrated.

Full text of DOI:10.1021/ol8011309

ORGANIC
LETTERS
2008
Vol. 10, No. 15
3291-3294
Dynamic Combinatorial Libraries
Constructed on Polymer Scaffolds
David A. Fulton*
School of Chemistry, Newcastle UniVersity, Newcastle Upon Tyne NE1 7RU, U.K.
d.a.fulton@ncl.ac.uk
Received May 20, 2008
ABSTRACT
Functionalized polymers were prepared by grafting acylhydrazides onto a polyvinylbenzaldehyde scaffold through reversible hydrazone linkages.
The dynamic nature of these linkages allows the functionalized polymers to exchange and reshuffle their appendages, and the resultant
mixture of polymers can be considered as a dynamic combinatorial library constructed upon a polymer scaffold. The dynamic nature of these
functionalized polymers was demonstrated.
The preparation of synthetic macromolecules which can
mimic natural proteins in terms of function such as catalysis
or molecular recognition presents a formidable challenge to
chemists. In the 1990s, Menger and co-workers utilized1 a
“combinatorial” approach to developing polymers possessing
catalytic activity, and more recently, Schrader and co-workers
have used2 a small set of functionalized monomers to prepare
libraries of copolymers and successfully screened these
libraries to find selective protein binders. Both of these
approaches to the design of functional synthetic macromol-
ecules are fundamentally different from that of Nature, where
through processes of selection and replication effective
functional macromolecules have evolved.
(DCLs), where all library members are in equilibrium and
can interconvert with one another through a reversible
chemical process. Addition of a template can lead to further
stabilization (e.g., through noncovalent interactions) of those
library members which can most effectively bind to the
template. Consequently, the entire DCL can re-equilibrate,
amplifying the concentration of effective binders of the
template and consuming library members which are inef-
fective binders.
The application of concepts from DCC to the discovery
of synthetic macromolecules which can mimic proteins is
intriguing. Natural proteins consist of amino acid residues
within very specific sequences, and it is the nature of the
amino acid side chains and their precise location upon the
peptide backbones which are largely responsible for protein
function.4 Importantly, peptides cannot exchange or reshuffle
their amino acid constituents without access to the complex
cellular machinery found in biology.5 We envisaged a wholly
synthetic polymer-based system where functionalized resi-
dues are grafted onto a preformed polymer scaffold through
Recently, dynamic combinatorial chemistry3 (DCC) has
emerged as a tool for the discovery of host-guest systems
with the potential for catalysis and molecular recognition.
DCC exploits collections of dynamic combinatorial libraries
(1) (a) Menger, F. M.; Eliseev, A. V.; Migulin, V. A. J. Org. Chem.
1995, 60, 6666–6667. (b) Menger, F. M.; West, C. A.; Ding, J. Chem.
Commun. 1997, 633–634. (c) Menger, F. M.; Ding, J.; Barragan, V. J. Org.
Chem. 1998, 63, 7578–7579.
(2) Koch, S. J.; Renner, C.; Xie, X.; Schrader, T. Angew. Chem., Int.
Ed. 2006, 45, 6352–6355.
(3) Corbett, P. T.; Leclaire, J.; Vial, L.; West, K. R.; Wietor, J.-L.;
Sanders, J. K. M.; Otto, S. Chem. ReV. 2006, 106, 3652–3711.
(4) Berg, J. M.; Tymoczko, J. L.; Stryer, L., M. Biochemistry, 5th ed.;
W. H. Freeman and Company: New York, 2002; Chapter 3, pp 41-73.
10.1021/ol8011309 CCC: $40.75
Published on Web 06/28/2008
2008 American Chemical Society

dynamic covalent linkages.6,7 The reversibility of these
linkages allows residues to exchange and reshuffle their
positions upon the polymer scaffold, and the resulting
mixture of constitutionally dynamic8 polymers can be
considered to possess the attributes of a DCL constructed
upon a polymer scaffold. We hypothesize that such DCLs
should be able to respond to the addition of templates in the
same manner as any other DCL. A further advantage of this
system is that polymer scaffolds of very precise lengths can
be prepared using controlled/“living” radical polymerization
methods9–11 offering a degree of flexibility over the size and
complexity of the resultant DCLs.
We report here the preparation of a “simple” polymer-
scaffolded DCL constructed from a preformed polymer
scaffold adorned through dynamic covalent hydrazone bonds
with simple acylhydrazide residues, and demonstrate the
dynamic nature of this system.
Recent reports describing the preparation of well-defined
polymers bearing aldehyde functionalities,14 coupled with
the apparent lack of reports of well-defined polymers bearing
acylhydrazide functionalities, led us to base our DCL upon
a polyaldehyde scaffold. Polyvinylbenzaldehyde (1) (DP )
50, PDI ) 1.11) was prepared by reversible addition
fragmentation chain-transfer polymerization9 of vinylben-
zaldehyde15 according to the method described by Wooley
and co-workers.14c Acylhydrazides (2-4) containing C5,
C10,or C17 alkyl chains, respectively, were chosen as simple
model residues of differing molecular weights to graft onto
the polyvinylbenzaldehyde scaffold. These acylhydrazides
were prepared in high yields by simple hydrazinolysis of
their corresponding methyl esters.
Acylhydrazone exchange was chosen as the dynamic
reaction to append residues to suitably functionalized poly-
mer scaffolds as it is a well-studied and successful reaction
in dynamic covalent chemistry.3,12 Acylhydrazone linkages
are formed from the acid-catalyzed condensation of acylhy-
drazides with aldehydes, and the resulting acylhydrazone
bonds are readily broken and formed under acidic condi-
tions,13 while neutralization yields kinetically stable products.
These reaction conditions are compatible with many molec-
ular recognition events that could potentially influence the
equilibrium distribution of the resulting DCL.
(5) For a report describing the generation of a DCL of peptides using
nonspecific proteases under conditions in which both hydrolysis and
synthesis occur, effectively allowing amino acid residues to exchange and
reshuffle their positions within a peptide framework without access to
complex celluar machinery, see: Swann, P. G.; Casanova, R. A.; Desai,
A.; Frauenhoff, M. M.; Urbancic, M.; Slomczynska, U.; Hopfinger, A. J.;
Scheme 1. Functionalization of Polyvinylbenzaldehyde with
Acylhydrazides through Hydrazone Bond Formation
Breton, G. C.; Venton, D. L. Biopolymers 1996, 40, 617–625
.
(6) To the best of our knowledge, the only reported example of the
functionalization of pre-formed polymer chains through dynamic covalent
bonds where dynamic nature has been demonstrated is: Higaki, Y.; Otsuka,
H.; Takahara, A. Macromolecules 2004, 37, 1696–1701
.
(7) (a) The functionalization of pre-formed polymers through dynamic
covalent bonds is conceptually different to dynamic covalent polymers
prepared by linking monomers together through dynamic covalent bonds.
For some examples of such dynamic covalent polymers, see: Otsuka, H.;
Aotani, K.; Higaki, Y.; Takahara, A. J. Am. Chem. Soc. 2003, 125, 4064–
4065. (b) Niu, W.; O’Sullivan, C.; Rambo, B. M.; Smith, M. D.; Lavigne,
J. J. Chem. Commun. 2005, 4342–4344. (c) Kamplain, J. W.; Bielawski,
C. W. Chem. Commun. 2006, 1727–1729. (d) Ono, T.; Fujii, S.; Nobori,
T.; Lehn, J.-M. Chem. Commun. 2007, 4360–4362. (e) Ruff, Y.; Lehn, J.-
M. Angew. Chem., Int. Ed. 2008, 47, 3556–3559
.
(8) (a) Lehn, J.-L. Prog. Polym. Sci. 2005, 30, 814–831. (b) Lehn, J.-
M. Chem. Soc. ReV. 2007, 36, 151–160.
(9) (a) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2005, 58,
379–410. (b) Perrier, S.; Takolpuckdee, P. J. Polym. Sci., Part A: Polym.
Chem. 2005, 43, 5347–5393. (c) Moad, G.; Rizzardo, E.; Thang, S. H. Aust.
J. Chem. 2005, 59, 669–692
.
(10) (a) Matyjaszewski, K.; Xia, J. Chem. ReV. 2001, 101, 2921–2990.
(b) Kamigaito, M.; Ando, T.; Sawamoto, M. Chem. ReV. 2001, 101, 3689–
The functionalization of 1 (Scheme 1) was performed
using an excess of acylhydrazides 2-4 in THF (concentration
3746
.
(11) Hawker, C. J.; Bosman, A. W.; Harth, E. Chem. ReV. 2001, 101,
3661–3688
(12) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.;
.
Stoddart, J. F. Angew. Chem., Int. Ed. 2002, 41, 898–952
.
(14) (a) Hwang, J. Y.; Li, R. C.; Maynard, H. D. J. Controlled Release
2007, 122, 279–286. (b) Li, R. C.; Hwang, J.; Maynard, H. D. Chem.
Commun. 2007, 3631–3633. (c) Sun, G.; Cheng, C.; Wooley, K. L.
Macromolecules 2007, 40, 793–795. (d) Yang, S.-K.; Weck, M. Macro-
molecules 2008, 41, 346–351.
(13) (a) Recent work as shown that the well known dynamic covalent
reactionsshydrazone and oxime formationscan be catalyzed in water using
aniline as a catalyst. These findings open up the possibility of developing
water-soluble polymer-scaffolded DCLs. See: Dirksen, A.; Hackeng, T. M.;
Dawson, P. E. J. Am. Chem. Soc. 2006, 128, 15602–15603. (b) Dirksen,
A.; Dirksen, S.; Hackeng, T. M.; Dawson, P. E. Angew. Chem., Int. Ed.
2006, 45, 7581–7584.
(15) (a) Rodgers, C. J.; Dickerson, T. J.; Wentworth, P.; Janda, K. D.
Tetrahedron 2005, 61, 12160–1216. (b) Manzotti, R.; Reger, T. S.; Janda,
K. D. Tetrahedron Lett. 2000, 41, 8417–8420.
3292
Org. Lett., Vol. 10, No. 15, 2008

of 1 approximately 0.2 mmol) with TFA catalyst, and each
reaction was monitored by gel permeation chromatography
(GPC) until no further changes in retention times were
observed, indicating equilibrium was reached within 2 h in
all three cases.
Scheme 2. Hydrazone Exchange of Polymers 5 and 7
Analysis of each reaction at equilibrium by GPC (Tab-
le 1) revealed the formation of higher molecular weight
Table 1. Comparison of Mn (Measured by GPC), Mn
(Calculated), and PDIs for Polymers 1 and 5-7
polymer
Mn (GPC)
Mn (calc)a
PDIb
1
5
6
7
6660
7650
11600
14600
6964
12570
15980
20890
1.11
1.13
1.13
1.14
then added and the mixture allowed to equilibrate at room
temperature. GPC analysis showed the disappearance of the
peaks at 14.61 and 15.34 min and the appearance of a new
peak at 15.02 min. The mixture reached equilibrium within
2 h, and the GPC profiles of the reaction before the addition
of TFA and at equilibrium are shown in Figure 1. The same
a For polymer 1, calculated from 1H NMR. For polymers 5-7, the
calculated Mn was based on all aldehyde functions of polymer 1 reacting
with acylhydrazides to form hydrazones. b By GPC.
polymers 5-7, respectively, indicating the successful ligation
of the acylhydrazides onto the polymer scaffolds. Under these
conditions, the polydispersity index (PDI) of the function-
alized polymers was essentially identical to that of the bare
polymer 1, indicating that all aldehyde functions of the
polymer have reacted and that the functionalization does not
change basic polymer properties. The measured Mn of each
functionalized polymer, however, was lower than expected
if all aldehyde functions on the polymer had reacted.16 Each
reaction was neutralized by treatment with anion-exchange
resin (Amberlyst 21) and polymers 5 and 7 were purified
through Sephadex LH-20.17 The H NMR spectra of 5 and
1
7 were exceptionally broad,18 but IR spectroscopy of both
polymers displayed the absence of the carbonyl stretch at
1702 cm-1 present in 1 and the presence of an CdN stretch
at 1662 cm-1, suggesting that hydrazone formation is an
effective reaction for the functionalization of polymers.19
To demonstrate the dynamic nature of these functionalized
polymers, 5 and 7 were mixed (Scheme 2) in THF (concen-
tration of each approximately 0.2 mM). A catalytic quantity
of TFA (concentration of TFA approximately 5 mM) was
Figure 1. GPC profiles of (a) a THF solution of polymers 5 and 7
before the addtion of TFA catalyst and (b) the mixture at
equilibrium after the addition of TFA catalyst.
mixture can also be reached by reaction of 50 equiv of 2
and 50 equiv of 4 with 1 under similar conditions, affording
polymer with identical elution volume and PDI. These
findings suggest strongly that both the polymer functionalized
with C5 residues and that functionalized with C17 residues
have undergone hydrazone exchange to form a mixture of
polymers functionalized with both the C5 and C17 residues.
(16) Polymers 5-7 are structurally different from the polystyrene
standards used for GPC calibration, so differences between measured and
theoretical Mn are not unexpected.
(17) Comparison of the GPC profiles of the polymers before and after
purification indicated identical retention times and PDIs.
(18) The ligation of acylhydrazide residues onto 1 was investigated by
1H NMR titration experiments. These revealed that as hydrazone bonds
form, the aldehyde signal (δ ) 9.81 ppm) of the polymer deceases in
intensity as the CONH signal (δ ) 10.80 ppm) of the forming hydrazone
bonds increases in intensity. As the stoichiometry of the ligating residues
approaches that of the aldehydes of the polymer the 1H NMR spectra became
increasingly broad, preventing meaningful interpretation of the aldehyde
and hydrazone signals.
Such a polymer-based DCL possesses the potential to
display vast molecular diversity. The theoretical number of
unique members of such a DCL will depend upon the number
of aldehydes (or other functional groups) (x) upon the
polymer, the number of different types of residue available
(y) and the percentage of functional groups on the polymer
which have reacted. For the “simple” DCL described based
on a polymer with 50 aldehydes (x ) 50) and two different
residue types (y ) 2), assuming 100% ligation of the polymer
leads to a DCL theoretically possessing around 250 unique
(19) (a) The laboratory of Weck (ref 14d) have described recently the
functionalization of ketone-functionalized polymers with acylhydrazides
through hydrazone bonds, although the dynamic nature of these bonds has
not been investigated: Van Horn, B. A.; Iha, R. K.; Wooley, K. L.
Macromolecules 2008, 41, 1618–1626. (b) Van Horn, B. A.; Wooley, K. L.
Soft Matter 2007, 3, 1032–1040. (c) Taniguchi, I.; Mayes, A. M.; Chan,
E. W. L.; Griffith, L. G. Macromolecules 2005, 38, 216–219. See refs 14a
and 14b.
Org. Lett., Vol. 10, No. 15, 2008
3293

members; it is likely that under normal experimental condi-
tions much of this diversity will be “virtual”.20
One drawback of utilizing polymer scaffolds such as
polystyrene which cannot be subdived is the inability to
establish sequence information, although the residual com-
position of a library member or group of members should
be starightforward to determine. Polymer scaffolds based on
peptides would present the possibility of sequencing with
Edman degradation methods after a library has been kineti-
cally fixed.
In summary, we have prepared mixtures of functionalized
polymers possessing the attributes of a DCL and demon-
strated their dynamic nature. We are currently expanding the
complexity of polymer-scaffolded DCLs by preparing resi-
dues incorporating functional groups found in amino acid
side chains and investigating the preparation and character-
ization of DCLs upon short oligomers. We then aim to
demonstrate how such polymer-scaffolded DCLs can respond
to transition-state analogue templates. We believe polymer-
based DCLs have great potential in (1) the discovery of
enzyme-like catalysts, (2) the discovery of polymer-based
receptors for small molecules, (3) the creation of polymer-
based protein-specific receptors, and (4) combinatorial
materials research.
Theoretical work21 on DCLs by the groups of Moore,
Severin, Sanders, and Otto make it possible to comment on
how such polymer-scaffolded DCLs will respond in the
presence of template. Amplification factors for library
members which are able to bind successfully a template are
predicted21b to be higher if the equilibrium distribution is
dominated by monomeric building blocks, and aggregates
are present in only small amounts. Matching this criteria with
polymer-scaffolded DCLs is straightforward, as the library
can simply be designed to contain a large excesses of residues
relative to polymer scaffold. Furthermore, the complexity
of polymer-scaffolded DCLs can by tailored simply be
controlling the length of the scaffold and the number of
residue types available. Work by Otto and Sanders
suggests21c that libraries containing 10-106 compounds still
allows useful amplification of the single best binder, and
oligomeric scaffolds would allow access to libraries pos-
sessing this order of diversity. For very large libraries,
theoretical work by Moore predicts21a that after reequilibra-
tion 5% of a library population can become up to 104 times
greater than the original library at binding a template. If this
5% of the library population can be isolated by, e.g., binding
to a resin-supported template followed by washing, theoreti-
cally one can evolve rapidly effective functional molecules,
although detailed information about structure will remain
elusive.
Acknowledgment. Research Councils UK (fellowship to
D.A.F.), the Royal Society (for provision of a University
Research Grant), and Newcastle University are gratefully
thanked for generous support.
Supporting Information Available: Experimental pro-
cedures, spectroscopic data for new compounds, and GPC
profiles of polymers. This material is available free of charge
via the Internet at http://pubs.acs.org.
(20) Lehn, J.-M. Chem. Eur. J. 1999, 5, 2455–2463.
(21) (a) Moore, J. S.; Zimmerman, N. W. Org. Lett. 2000, 2, 915–918.
(b) Grote, Z.; Scopelliti, R.; Severin, K. Angew. Chem., Int. Ed. 2003, 42,
3821–3825. (c) Corbett, P. T.; Otto, S.; Sanders, J. K. M. Org. Lett. 2004,
6, 1825–1827. (d) Corbett, P. T.; Otto, S.; Sanders, J. K. M. Chem. Eur. J.
2004, 10, 3139–3143.
OL8011309
3294
Org. Lett., Vol. 10, No. 15, 2008