Synthetic multifunctional pores with external and internal active sites for ligand gating and noncompetitive blockage

Article abstract of DOI:10.1021/ja045987+

Design, synthesis, and multifunctionality of p-octiphenyl β-barrel pores with external LRL triads and internal HH dyads are described. Molecular recognition of anionic fullerenes > calixarenes > pyrenes by guanidinium arrays at the outer pore surface is s

Full text of DOI:10.1021/ja045987+

Published on Web 10/05/2004
Synthetic Multifunctional Pores with External and Internal Active Sites for
Ligand Gating and Noncompetitive Blockage
Virginie Gorteau,† Florent Perret,† Guillaume Bollot,† Jiri Mareda,† Adina N. Lazar,§
Anthony W. Coleman,§ Duy-Hien Tran,† Naomi Sakai,† and Stefan Matile*,†
Department of Organic Chemistry, UniVersity of GeneVa, GeneVa, Switzerland, and Institut de Biologie et Chimie
des Prote´ines, CNRS UMR 5086, 7 passage du Vercors, F69367, Lyon, France
Received July 6, 2004; E-mail: stefan.matile@chiorg.unige.ch
Current research on the use of synthetic,1-5 bioengineered,6,7 and
biological8 pores as sensors focuses exclusively on pore blockage.
In this report, we introduce the synthetic multifunctional pore 1
that opens rather than closes in response to chemical stimulation
(Figure 1).9
This ligand gating was created as follows.9 Directed by the
nonplanarity of the p-octiphenyl staves, cylindrical self-assembly
of rigid-rods 2 was designed to produce multifunctional â-barrel
pore 1 with HH dyads at the inner surface and LRL triads at the
outer surface. Blockage of rigid-rod â-barrel pores with internal
HH dyads by guests that match their functionalized internal space
has been demonstrated previously.1 External LRL triads were
introduced to exploit anion scavenging by oligoarginines10 for ligand
gating. Namely, the self-assembly of p-octiphenyl peptide conjugate
2 should afford water-soluble rigid-rod â-barrels with hydrophilic
anions scavenged by the external oligoarginine arrays. Multifunc-
tional pores 1 should then form in response to external ion exchange
with amphiphilic anions 3 with high affinity for oligoarginine.10,11
p-Octiphenyl 2 was synthesized in 19 steps overall, following
protocols reported previously for other sequences.1 When added
to EYPC-LUVs⊃CF, this new rigid-rod molecule 2 did not mediate
substantial efflux of self-quenched intravesicular CF (Figure 2Aa,
EYPC-LUVs⊃CF: large unilamellar vesicles composed of egg yolk
phosphatidylcholine and loaded with 5(6)-carboxyfluorescein).
Subsequent addition of ligands 3, however, caused rapid, concentra-
tion-dependent CF efflux (e.g., Figure 2Ab-g). Hill analysis of
the resulting dose response curve provided an effective concentra-
tion, EC50 ) 14.8 µM, for pyrenebutyrate 3c to stimulate 50% pore
activity. EC50’s decreased from pyrene carboxylate 3c to pyrene
sulfate 3b, calix[4]arene 3a (Figure 2C(•)), and fullerene 3d, with
an EC50 ) 10.2 ( 1.9 nM (Table 1). Under experimental conditions,
all ligands 3 were membrane inactive without pores (e.g., Figure
2Ah). Replacement of the external arginines by leucines in pores
formed by 2a1 annihilated ligand gating (Figure 2Bc versus 2Bd).
Pores with internal rather than external arginine arrays formed by
rods 2b1 were partially blocked rather than opened by argininophile
3c. Once bound to bilayer membranes, pores 1 were not able to
permealize the newly added EYPC-LUVs⊃CF (Figure 2Be). This
inability of intervesicular transfer contrasted sharply with conven-
tional pores with external LLL triads,12 suggesting that external
ligands strengthened barrel-membrane interactions. This interpreta-
tion was supported by increasing the fluorescence resonance energy
transfer (FRET) from the p-octiphenyl staves (i.e., the intrinsic
fluorescent probe of rigid-rod â-barrels) to BODIPY lipids,13 with
increasing concentration of calixarene 3a (Figure 2C).
Figure 1. Ligand-gated self-assembly of rod 2 into pore 1 and noncompeti-
tive blockage of pore 1 with poly-L-glutamate (pE, random coil; R-pE,
R-helix) to give complex 4. Rigid-rod â-barrels are depicted in side (1)
and axial (4) views, with â-sheets as arrows (N f C, 1) or solid (backbone)
and dotted lines (H bonds, 4; external amino acid residues, dark on white
and internal ones, white on dark, single-letter abbreviations). (A) Molecular
mechanics simulations of complex 4 with 3a as the ligand in side (left)
and axial (right) view. Arginines are in blue, histidines in cyan, leucines
and calixarenes in gold, â-sheets and R-helix as ribbons, calixarenes space-
filling (right), and p-octiphenyls, in ball-and-stick, in tan color. We caution
that all shown suprastructures are, in part, speculative simplifications that
are, however, consistent with molecular models (A) and experimental data.
â-barrel pores are ideal for R-helix recognition (Figure 1A).1 It was,
therefore, not surprising that poly-L-glutamate (pE) blocked pores
1 with IC50’s, which were independent of the EC50’s of the external
ligands (Table 1 and Figure 2D(•),Bf). The IC50’s for blockage by
heparin, in contrast, varied with the EC50’s of ligands 3a-c (Table
1). This difference supported the view that oligoargininophile10
heparin acted, at least in part, competitively on the outer barrel
Force-field geometry optimizations (MMFF94/MacroModel)14
confirmed that the dimensions of the internal space of rigid-rod
† University of Geneva.
§ Institut de Biologie et Chimie des Prote´ines.
9
13592
J. AM. CHEM. SOC. 2004, 126, 13592-13593
10.1021/ja045987+ CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
Figure 2. Changes in activity and membrane affinity of 2 and 2a (Bc,d) in response to ligands 3c (A and B) and 3a (C), blockers (pE, Bf and D), and
enzymes (PLE, E and F) in lipid bilayers. (A) Fractional change in CF emission I (λex 492 nm, λem 517 nm) as a function of time during addition of rod 2
(100 nM (a-g) and 0 nM (h)) and ligand 3c (0 (a), 5 (b), 10 (c), 20 (d), 22 (e), 30 (f), and 50 µM (g and h)) to EYPC-LUVs⊃CF (250 µM EYPC) in buffer
(10 mM HEPES, 107 mM NaCl, pH 7.0), calibrated by final lysis (excess Triton X-100). (B) Same for addition sequences (a) 2 (100 nM), (b) 2 (100 nM)
f 3c (50 µM), (c) 3c (50 µM) f 2a (1 µM), (d) 2a (1 µM), (e) 2 (100 nM) f 3c (50 µM) f EYPC-LUVs⊃CF (250 µM EYPC), and (f) 3b (6 µM) f
2 (500 nM) f pE (1 µM). (C) Fractional activity of pore 2 (500 nM) as a function of the concentration of ligand 3a (b, fit to Hill eq) and BODIPY emission
under identical conditions (×). (D) Fractional activity of pore 2 (500 nM) with ligand 3a (1 µM) as a function of the concentration of blocker pE (b, fit to
Hill eq) and BODIPY emission under identical conditions (×). (E) Fractional change in CF emission I as a function of time during addition of rod 2 (100
nM), proligand 5 (600 µM), and pig liver esterase (0 (a), 0.10 (b), 0.25 (c), 0.50 (d), 1.00 (e), and 10.0 units/ml (f)) to EYPC-LUVs⊃CF (250 µM EYPC).
(F) Initial velocity of formation of ligand 3c as a function of esterase concentration (summary E).
Table 1. Data on Ligand Gating and Blockage of Pore 1a
Acknowledgment. We thank A. Som and N. Sorde´ for
IC50 (nM)c
IC50 (nM)c
heparin
experimental assistance, D. Jeannerat, A. Pinto, and J.-P. Saulnier
for NMR measurements, P. Perrottet and the group of F. Gu¨lac¸ar
for ESI-MS, H. Eder for elemental analyses, one referee for helpful
suggestions, and the Swiss NSF (200020-101486 and National
Research Program “Supramolecular Functional Materials” 4047-
057496; S.M.), Delta Proteomics (A.N.L.), and the CNRS (A.W.C.)
for financial support.
liganda
EC50
(
µ
M)b
polyglutamate
3a
3b
3c
0.44 ( 0.04
3.30 ( 0.50
14.80 ( 2.40
42.0 ( 2.2
47.0 ( 4.0
41.0 ( 7.1
24.6 ( 3.3
19.0 ( 1.2
4.2 ( 0.5
a See Figure 1 for structures. b Concentration of ligand 3 required for
50% pore activation. c Concentration of blocker required for 50% pore
blockage.
Supporting Information Available: Experimental details and brief
discussion of FRET experiments. This material is available free of
charge via the Internet at http://pubs.acs.org.
surface, whereas pE was recognized by spatially separated internal
active sites to give complex 4 (Figure 1). The functional evidence
for noncompetitive pore blockage was corroborated on the structural
level by unchanged FRET from fluorescent pore 1 to BODIPY
lipids during blockage by pE (Figure 2D(×)).
References
(1) Sakai, N.; Matile, S. Chem. Commun. 2003, 2514-2523.
(2) Matile, S.; Som, A.; Sorde´, N. Tetrahedron 2004, 60, 6405-6435.
(3) Gokel, G. W.; Mukhopadhyay, A. Chem. Soc. ReV. 2001, 30, 274-286.
(4) Kirkovits, G. J.; Hall, C. D. AdV. Supramol. Chem. 2000, 7, 1-47.
(5) Scrimin, P.; Tecilla, P. Curr. Opin. Chem. Biol. 1999, 3, 730-735.
(6) Bayley, H.; Cremer, P. S. Nature 2001, 413, 226-230.
(7) Terrettaz, S.; Ulrich, W.-P.; Guerrini, R.; Verdini, A.; Vogel, H. Angew.
Chem., Int. Ed. 2001, 40, 1740-1743.
(8) Deamer, D. W.; Branton, D. Acc. Chem. Res. 2002, 35, 817-825.
(9) The terms “ligand gating” and “blockage” together with “open(ing)” and
“close/closing” are used to refer explicitly to function without any
structural implications (i.e., increase and decrease in “pore activity” in
response to ligands and blockers, respectively). We emphasize that the
structural basis of “ligand gating” in this study (i.e., ligand-mediated
changes in pore-membrane interactions) is designed to differ from many
cases of biological “ligand gating” (i.e., ligand-mediated changes in pore/
channel conformation).
We have previously shown that synthetic multifunctional pores
are of practical use as adaptable detectors of enzyme activity.15,16
The continuous detection of enzyme activity was, however,
problematic with pores that operate on blockage because substrates
bound within pores seem to be less accessible to enzymes, whereas
continuous detection of pore closing upon production of blockers
is incompatible with routine fluorescence detection. This challenge
was therefore ideal for probing the practical usefulness of ligand-
gated pore sensors. Pyrenebutyrate methylester 5 was considered
as a substrate of pig liver esterase (PLE) that would yield ligand
3c as a product (Figure 1). Addition of PLE to a mixture of substrate
5, rod 2, and CF vesicles caused the CF efflux indicative for
“enzyme gating” (Figure 2E). Linear dependence on enzyme
concentration suggested that the observed initial velocity of CF
efflux reflected the initial velocity of ligand formation (i.e., enzyme
kinetics (Figure 2F)). We summarize that the rational design of
synthetic multifunctional pores that can be opened9 and closed
noncompetitively by external ligands9 and internal blockers is
possible, and that such pores can be of practical use for, namely,
the continuous fluorometric detection of chemical processes.
(10) Sakai, N.; Matile, S. J. Am. Chem. Soc. 2003, 125, 14348-14356.
(11) Da Silva, E.; Lazar, A. N.; Coleman, A. W. J. Drug Del. Sci. Technol.
2004, 14, 3-20.
(12) Das, G.; Matile, S. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5183-5188.
(13) Weiss, L. A.; Sakai, N.; Ghebremariam, B.; Ni, C.; Matile, S. J. Am.
Chem. Soc. 1997, 119, 12142-12149.
(14) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton,
M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440-467. (b) Halgren, T. A. J. Comput. Chem. 1999,
20, 720-748.
(15) Das, G.; Talukdar, P.; Matile, S. Science 2002, 298, 1600-1602.
(16) Sorde´, N.; Das, G.; Matile, S. Proc. Natl. Acad. Sci. U.S.A. 2003, 100,
11964-11969.
JA045987+
9
J. AM. CHEM. SOC. VOL. 126, NO. 42, 2004 13593