Reversing the discovery paradigm: A new approach to the combinatorial discovery of fluorescent chemosensors

Article abstract of DOI:10.1021/ja043682p

We report here a new approach to the discovery of fluorescent chemosensors in which a new signaling mechanism allows a core fluorophore to be used in a combinatorial search for new binding events, thus reversing the reigning discovery paradigm. Copyright

Full text of DOI:10.1021/ja043682p

Published on Web 06/30/2005
Reversing the Discovery Paradigm: A New Approach to the Combinatorial
Discovery of Fluorescent Chemosensors
Jesse V. Mello and Nathaniel S. Finney*,†
Department of Chemistry and Biochemistry, UniVersity of California, San Diego 9500 Gilman DriVe,
La Jolla, California 92093-0358
Received October 18, 2004; E-mail: finney@oci.unizh.ch
Scheme 1. Synthesis of the Chemosensor Librarya
Fluorescent chemosensors (small molecules whose fluorescence
emission changes in response to a binding event) have become
widely appreciated as tools for monitoring biological, biomedical,
and environmental processes.1 The reigning paradigm for chemosen-
sor discovery is to identify a binding event of interest and derivatize
the receptor component with an appropriate fluorophore. This makes
the development of new chemosensors extremely challenging as
chemists’ collective ability to design new binding events remains
limited. While the power of combinatorial methods in the discovery
a Reagents and conditions: a.(EtO)2P(O)CH2CO2H, DEPBt, DIEA, THF;
of new molecular recognition phenomena is now established,2 very
b. LiBr, NEt3, 1. CH3CN; c. TFA/CH2Cl; d. Fmoc-AA (Chart 1), DEPBT,
little progress has been made in the combinatorial discovery of fluor-
iPrNEt, CH2Cl2; e. piperidine/DMF; f. [t]: Ac2O, DMAP, CH2Cl2; [u-x]:
escent chemosensors despite the acknowledged appeal of this
approach.3,4 This reflects the limitations of the common mechanisms
for turning a binding event into a fluorescent response, which either
restrict the potential diversity of the binding domains or require
molecular architecture too complex for ready combinatorial varia-
tion.5
ArCO2H, as d; [y-z,aa]: ArCOCl/SO2Cl, NEt3, DMAP, CH2Cl2; [bb-cc]:
i
PhNCS/PhNCO, CH2Cl2 (Chart 1); g. TFA, Pr3SiH, CH2Cl2.
Chart 1. L-Amino Acid and Acyl End Cap Library Components
We have found that reversal of the discovery paradigm, using
fluorescence to search for new binding events, and the construction
of combinatorial chemosensor libraries can be simultaneously
effected by exploiting a new signaling mechanism, binding-induced
restriction of fluorophore biaryl torsion.6,7 This mechanism allows
the ready construction of solid-phase libraries of potential fluores-
cent chemosensors, as it does not restrict binding domain diversity
and relies on comparatively simple molecular architecture. Here
we describe the synthesis of a proof-of-principle library based on
a core biarylpyridine fluorophore, and the discovery of novel visibly
emissive Hg2+-responsive chemosensors from this library.
We previously found polyarylpyridines to be versatile scaffolds
for chemosensor development and selected a 2,6-biaryl-4-vinylpy-
ridine as the core fluorophore for our library.6b,f Following prepara-
tion of an appropriately functionalized, 2,6-biarylpyridine-4-car-
boxaldehyde (1, Scheme 1), this fluorophore precursor was attached
to PEG-derivatized aminomethyl polystyrene via coupling with an
immobilized phosphonate.8,9 Subsequent removal of the Boc groups
provided resin derivatized with the free diaminofluorophore.
We then appended the fluorophore with two identical arms
consisting of an amino acid followed by an acylating end cap. This
provided both synthetic simplicity and a high likelihood that library
elements would exhibit metal binding,10 in turn allowing us to
quickly test the central hypothesis that our fluorophore could be
used to search for new binding events. The amino acid and acyl
end cap components were selected to allow for direct interaction
with metal ions as well as additional hydrophobic interactions
induced by metal binding (Chart 1, a-r, s-cc).
removal of the trityl (Tr), methoxytrityl (Mtt), and Boc protecting
groups from all library members except those capped with phen-
ylthiourea (cc) provided the library of 198 spatially arrayed potential
fluorescent chemosensors.12
Library members were transferred to 96-well microtiter plates
and evaluated in parallel for their response to the addition of solu-
tions of 11 separate metal ions, as evaluated with a microscope in
a dark room using a hand-held TLC lamp as the illumination source.
An attractive feature of this library is that it is self-assaying: upon
substrate binding, “hits” reveal themselves through an increase in
visible emission.13 The ions Li+, Na+, K+, Cs+, Mg2+, Ca2+, Cu2+
,
Zn2+, Cd2+, Hg2+, and Pb2+ were chosen on the basis of physiologi-
cal or environmental relevance,14 and were added as 1 mM solutions
of metal perchlorate in 5 mM aqueous MOPS buffer (pH 7.5).
The majority of metal/library element combinations led to little
or no change in fluorescence emission. However, all library ele-
ments bearing a phenylthiourea end cap (cc) gave a readily seen
increase in visible green fluorescence emission upon addition of
Hg2+ (Figures 1, 2).15-18 While secondary to the apparent require-
ment for an Hg-thiourea interaction, the identity of the amino acid
also influenced emission intensity; for instance, after the addition
of Hg2+, the emission from library element k-cc (Figures 1, 2A)
was slightly brighter than that from g-cc and much brighter than
Suitably protected Fmoc amino acids were coupled by manual
i
parallel synthesis to the fluorophore with DEPBT and Pr2NEt in
CH2Cl2.11 Acyl end caps were appended after removal of the Fmoc
protecting groups and division of the resin (Scheme 1). Subsequent
† Present address: Organisch-chemisches Institut, Universita¨t Zu¨rich, Winter-
thurerstrasse 190, 8057 Zu¨rich, Switzerland.
9
10124
J. AM. CHEM. SOC. 2005, 127, 10124-10125
10.1021/ja043682p CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
This material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) For representative reviews of fluorescent chemosensors, see: (a) Rurack,
K.; Resch-Genger, U. Chem. Soc. ReV. 2002, 31, 116-127. (b) Lavigne,
J. J.; Anslyn, E. V. Angew. Chem., Int. Ed. 2001, 40, 3118-3130. (c)
Valeur, B.; Leray, I. Coord. Chem. ReV. 2000, 205, 3-40.
(2) Overviews of peptide-containing combinatorial libraries for the discovery
of new molecular recognition phenomena: (a) Wennemers, H. Chimia
2003, 57, 237-240. (b) Tunnemann, R.; Jung, G. Nachr. Chem. 2000,
48, 453-457. (c) Still, W. C. Acc. Chem. Res. 1996, 29, 155.
(3) Czarnik, A. W. Chem. Biol. 1995, 2, 423.
Figure 1. Representative solid- and solution-phase chemosensors for Hg2+
.
(4) For the few examples of combinatorial approaches to fluorescent
chemosensor development, see: (a) Chen, C.-T.; Wagner, H.; Still, W.
C. Science 1998, 279, 851-853. (b) Rothman, J. H.; Still, W. C. Bioorg.
Med. Chem. Lett. 1999, 9, 509-512. (c) Schneider, S. E.; O’Neil, S. N.;
Anslyn, E. V. J. Am. Chem. Soc. 2000, 122, 542-543. (d) Singh, A.;
Yao, Q.; Tong, L.; Clark Still, W.; Sames, D. Tetrahedron Lett. 2000,
41, 9601-9605. (f) Iorio, E. J.; Shao, Y.; Chen, C.-T.; Wagner, H.; Still,
W. C. Bioorg Med. Chem. Lett. 2001, 11, 1635-1638.
(5) Photoinduced electron transfer (PET), by far the most common signaling
mechanism, requires only simple molecules but is restricted to binding
domains containing anilinic or benzylic nitrogen atoms. Energy transfer
signaling (FRET, e.g.) is permissive in terms of binding domain structure
but requires large changes in interchromophore distance, thereby requiring
complex architectures such as proteins or partial protein sequences.
Displacement-based assays, in which substrate binding displaces a
fluorophore fall between these two extremes. For examples, see ref 1.
For an expanded discussion, see ref 6f.
(6) (a) McFarland, S. A.; Finney, N. S. J. Am. Chem. Soc. 2001, 123, 1260-
1261. (b) Mello, J. V.; Finney, N. S. Angew. Chem., Int. Ed. 2001, 40,
1536-1438. (c) McFarland, S. A.; Finney, N. S. J. Am. Chem. Soc. 2002,
124, 1178-1179. (d) McFarland, S. A.; Finney, N. S. Chem. Commun.
2003, 388-389. (e) Fang, A. G.; Mello, J. V.; Finney, N. S. Org. Lett.
2003, 5, 967-970. (f) Fang, A. G.; Mello, J. V.; Finney, N. S. Tetrahedron
2004, 60, 11075-11087.
(7) For other work on biaryl conformational restriction as a signaling
mechanism, see: (a) Takeuchi, M.; Yoda, S.; Imada, T.; Shinkai, S.
Tetrahedron 1997, 53, 8335-8348. (b) Costero, A. M.; Andreu, R.;
Monrabal, E.; Martinez-Manez, R.; Sancenon, F.; Soto, J. J. Chem. Soc.,
Dalton Trans. 2002, 1769. (c) Lee, D. H.; Im, J. H.; Lee, J. H.; Hong, J.
I. Tetrahedron Lett. 2002, 43, 9637. (d) Cody, J.; Fahrni, C. J. Tetrahedron
2004, 60, 11099-11107.
Figure 2. Solid- and solution-phase emission from g-cc, k-cc, and 8.
that from h-cc. That this variation is seen with the side-chain
protecting groups still in place bodes well for larger future libraries
with more robust sulfur-containing capping groups.
(8) See Supporting Information for complete details.
(9) (a) Johnson, C. R.; Zhang, B. Tetrahedron Lett. 1995, 36, 9253-9256.
(b) Rathke, M. W.; Nowak, M. J. Org. Chem. 1985, 50, 2624-2626.
(10) For influential examples of seemingly simple combinatorial libraries of
metal chelators, see: (a) Burger, M. T.; Still, W. C. J. Org. Chem. 1995,
60, 7382. (b) Frances, M. B.; Finney, N. S.; Jacobsen, E. N. J. Am. Chem.
Soc. 1996, 118, 8983-8984.
The response of library elements such as g-cc to Hg2+ was
confirmed by the synthesis of a solution-phase analogue, glycol
amide 2.8 While 2 was not sufficiently soluble in aqueous MOPS
buffer to allow an exact solution-phase repetition of the solid-phase
titration, comparison of the normalized emission spectra for a 10-5
M CH3CN solution of 2 and a suspension of resin bound g-cc in
CH3CN provides confirmation of the solid-phase results and shows
that the solution- and solid-phase steady-state fluorescence proper-
ties are nearly identical (Figure 2B).17,18 Titration of 2 in CH3CN
revealed Ka ) 1.8 × 10-6 M-1 for the formation of 2‚Hg2+.8 This
affinity is an order of magnitude greater than that of 18-crown-6
for K+,19 and represents both a remarkable first hit and a promising
starting point for the development of more sensitive Hg2+ sensors.20
In conclusion, we have demonstrated a new and broadly applic-
able approach to the identification of new fluorescent chemosensors.
This approach reverses the reigning discovery paradigm, which re-
quires a preestablished binding event. The ease with which we found
a new class of Hg2+-responsive chemosensors demonstrates the
power of this approach and confirms the hypothesis that our fluoro-
phore and signaling mechanism are well suited to such explorations.
Future work includes further study of 2 and its Hg(II) complex,
the development of analogues with increased solubility, selectivity,
and affinity, and the search for chemosensors for other metal ions.
(11) Li, H.; Jiang, X.; Ye, Y.-h.; Fan, C.; Romoff, T.; Goodman, M. Org.
Lett. 1999, 1, 91-94.
(12) Under the OtBu cleavage conditions we observed cleavage of the
fluorophore from the resin; thus, the OtBu groups were left in place.
Phenylthiourea-terminated library elements were left protected to avoid
Edman degradation under acidic conditions.
(13) It is, of course, impossible to exclude the possibility that there are binding
events that do not lead to increased emission. However, in every system
we have studied,6 we have observed perfect correlation between fluores-
cence enhancement and metal ion binding (as determined by 1H NMR).
(14) Since MOPS buffer contains a Na+ counterion, response to Na+ was also
evaluated in triethanolamine buffer. For a discussion of metals for which
fluorescent detection would be valuable, see ref 1 and citations therein.
(15) The addition of excess ethanedithiol reverses the increase in emission,
demonstrating reversibility of Hg(II) binding. This result was confirmed
in solution with 2, and 2 could be recovered unchanged from Hg(II)
titrations by preparative TLC.
(16) Library elements bearing 2-pyridyl (v), 2-pyrazyl (w), or phenylthiourea
(cc) end caps exhibited partial quenching upon addition of Cu2+
.
(17) Inspection of the emission spectra from suspended resin-bound fluorophore
(Figure 2) indicates that the blue “background emission” seen in the
absence of Hg2+ is scattered blue light from the illumination source. The
quantum yield of 2 in the absence of Hg(II) is ∼0.01 with ꢀ ≈ 12,000.
(18) Other examples of “turn on” fluorescent chemosensors for aqueous Hg-
(II): (a) Ono, A.; Togashi, H. Angew. Chem., Int. Ed. 2004, 43, 4300-
4302. (b) Guo, X.; Qian, X.; Jia, L. J. Am. Chem. Soc. 2004, 126, 2272-
2273. (c) Nolan, E. M.; Lippard, S. J. J. Am. Chem. Soc. 2003, 125,
14270-14271. (d) Rurack, K.; Kollmannsberger, M.; Resch-Genger, U.;
Daub, J. J. Am. Chem. Soc. 2000, 122, 968-969. (e) Prodi, L.; Bargossi,
C.; Montalti, M.; Zaccheroni, N.; Su, N.; Bradshaw, J. S.; Izatt, R. M.;
Savage, P. B. J. Am. Chem. Soc. 2000, 122, 6769-6770.
(19) (a) Izatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L. Chem. ReV.
1991, 91, 1721-2085. (b) Izatt, R. M.; Pawlak, K.; Bradshaw, J. S.;
Bruening, R. L. Chem. ReV. 1995, 95, 2529-2586.
Acknowledgment. We thank the National Science Foundation
for support of this research (CHE-9876333) and the Departmental
NMR facilities (CHE-9709183). We dedicate this manuscript to the
fond memory of Prof. Murray Goodman (UCSD).
(20) Fluorescence quenching by Hg2+ results from contact-mediated spin-
orbit coupling (McClure, D. S. J. Chem. Phys. 1952, 20, 682-686),
underscoring a key advantage of the spatial segregation of fluorophore
and binding domain.
Supporting Information Available: Complete experimental details;
spectral data for 1, 2 and precursors thereof; Ka determination for 2.
JA043682P
9
J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005 10125