Interface engineering of synthetic pores: Towards hypersensitive biosensors

Article abstract of DOI:10.1002/chem.200701520

Hydrophilic anchoring is introduced as a promising strategy to constructively control the various interactions of synthetic pore sensors with the surrounding biphasic environment. Artificial rigid-rod β barrels are selected as classical synthetic multifun

Full text of DOI:10.1002/chem.200701520

FULL PAPER
DOI: 10.1002/chem.200701520
Interface Engineering of Synthetic Pores: Towards Hypersensitive Biosensors
Federico Mora,[a] Duy-Hien Tran,[a] Natalie Oudry,[b] Gerard Hopfgartner,[b]
Damien Jeannerat,[a] Naomi Sakai,[a] and Stefan Matile*[a]
Abstract: Hydrophilic anchoring is in-
troduced as a promising strategy to
constructively control the various inter-
actions of synthetic pore sensors with
the surrounding biphasic environment.
Artificial rigid-rod b barrels are select-
ed as classical synthetic multifunctional
pores and random-coil tetralysines are
attached as hydrophilic anchors. The
synthesis of this advanced pore is ac-
complished in 32 steps from commer-
cially available starting materials. With
regard to pore activity as such, the key
impact of hydrophilic anchoring is a
change from a Hill coefficient n<1 to
n=4. This change confirms successful
suppression of the competing self-as-
sembly with precipitation from the
aqueous phase as the origin of the ac-
complished increase in pore activity.
The hydrophilic anchors do not inter-
fere with the blockage of the synthetic
pore sensors by anionic analytes. In the
case of stoichiometric binding of block-
ers (KD =EC50 of the pore; EC50 =con-
centration needed to observe 50%
pore activity), however, the increase in
pore activity achieved by hydrophilic
anchoring results in improved pore
blockage under high dilution condi-
tions. Controls confirm that this in-
crease does not occur with analytes
that do not exhibit stoichiometric bind-
ing (KD >EC50). These results not only
reveal stoichiometric binding as the ex-
pected origin of the sensitivity limit of
synthetic pore sensors, they also pro-
vide promising solutions for this prob-
lem. The combination of hydrophilic
anchoring with targeted pore formation
emerges as a particularly promising
strategy to further reduce effective
pore concentrations. The scope and
limitations of this approach are exem-
plified with pertinent analyte pairs that
are essential for the sensing of sucrose,
lactose, acetate, and glutamate with
synthetic pores in samples from the su-
permarket.
Keywords: ion channels
· mem-
branes · peptides · self-assembly ·
sensors
Introduction
lytes could be easily monitored as blockage or deblockage
of the pore. This response could be used to reliably and rap-
idly quantify the analytes. Reactive amplifiers have been in-
troduced to covalently capture otherwise elusive analytes
after enzymatic signal generation and drag them into the
pore for transduction.[1] However, the detection limit of ana-
lytes by our pore sensors has so far been consistently in the
low micromolar range where the substrate selectivity was
also completely lost.[1,25] These phenomena suggested the
emergence of stoichiometric binding in this region in which
the pore concentration needed to detect activity becomes
similar to the dissociation constant KD of the pore–blocker
complex.[26,27] Lowering the effective pore concentration,
that is, increasing the activity of the pore as such would thus
be the key to increase both sensitivity and selectivity of syn-
thetic pore sensors. In other words, pore–membrane rather
than pore–blocker interactions would thus need improve-
ment to reach nanomolar sensitivity.
Recently, we obtained experimental evidence[1] for the abili-
ty of synthetic ion channels and pores[2–13] to serve as multi-
component sensors[12–24] in samples from the supermarket or
the hospital. These findings prompted us to construct re-
fined pore architectures for advanced applications. For mul-
ticomponent sensing in complex matrices,[1] synthetic pores
are used as general optical transducers of reactions together
with enzymes as selective signal generators.[12,13,25] Namely,
enzymatic formation or consumption of pore-blocking ana-
[a] F. Mora, D.-H. Tran, Dr. D. Jeannerat, Dr. N. Sakai, Prof. S. Matile
Department of Organic Chemistry, University of Geneva
Geneva (Switzerland)
Fax : (+41)22-379-3215
E-mail: stefan.matile@chiorg.unige.ch
[b] N. Oudry, Prof. G. Hopfgartner
School of Pharmaceutical Sciences, University of Geneva
Geneva (Switzerland)
To address this challenge, we noticed that, ironically, the
apparent activity of pores in the membrane often seems to
be determined by their solubility in water.[28–31] Namely, hy-
Supporting Information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2008, 14, 1947 – 1953
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1947

drophilic domains attached to one end of the pore may
assure delivery to the vesicle by preventing competing pre-
cipitation from the water and enforce vectorial partitioning
and transmembrane orientation followed by parallel self-as-
sembly. Exploited to perfection in biology, these various
benefits from hydrophilic anchoring are rarely considered in
synthetic functional systems.[28–31] Herein, we report hydro-
philic anchoring of rigid-rod b-barrel pores, such as 1, as a
promising approach toward multicomponent sensors that
can operate at low concentrations with high selectivity
(Figure 1).
peptidic anion channels.[29] Moreover, cationic anchors were
preferred over anionic or neutral anchors because repulsion
from the cationic interior of pore 1 was considered as essen-
tial to prevent backfolding of the anchor into the pore.
Rigid-rod molecule 4 was envisioned for the parallel self-
assembly into transmembrane b-barrel pore 1 with hydro-
philic K4 anchors. This multiple-substituted p-octiphenyl 4
was synthesized in 32 steps from commercially available
starting materials, including 14 steps of very straightforward
peptide synthesis.[37] The key challenge of this synthesis was
to attach two different peptide strands to one rigid-rod scaf-
fold. p-Octiphenyl 5 with carboxylic acids that carry orthog-
onal benzyl and tert-butyl protecting groups was conceived
to address this problem (Scheme 1). The synthesis of this
key intermediate 5 by Suzuki coupling of the two differently
substituted phenyl termini 6 and 7 to the previously report-
ed p-sexiphenyl scaffold 8[30] appeared not to be problemat-
ic.
The K4-terminal fragment 6 was prepared from resorcinol
9. Protected as a benzyl ester, one bromoacetate 10 was in-
troduced first by Williamson ether synthesis. Selective
ortho-iodination of the obtained phenol 11 afforded the aryl
iodide 12. This substrate was needed to introduce the
second bromoacetate 13 with the orthogonal tert-butyl pro-
tecting group. Pinacolboronate 6 was obtained by Pd-cata-
lyzed conversion[38] of aryl iodide 14.
The other rod terminus 7 was synthesized following re-
cently reported procedures.[39] In brief, ortho-iodination of
methylresorcinol 15 afforded regioisomer 16 chemoselec-
tively. This reaction was followed by Williamson ether syn-
thesis with tert-butyl bromoacetate 13. The obtained aryl
iodide 17 was transformed via boronate 18 to give 7, which
benefits from the increased stability, easier purification, and
higher reactivity of the solid potassium trifluoroborates.[40]
The p-sexiphenyl 8 was selected as an ideal building block
to attach the terminal rod fragments 6 and 7 because of its
rapid accessibility from the commercially available biphenyl
19 and tert-butyl bromoacetate 13.[30] In brief, diazide 19 was
converted into diiodide 20 and dipinacolboronate 21. These
two monomers were then polymerized under Suzuki cou-
pling conditions. Because of poor solubility, p-sexiphenyl 22
could be isolated directly and in good yield from the reac-
tion mixture. Rod 22 was treated with BBr3 to give oligo-
phenol 23. This rod 23 was treated with tert-butyl bromoace-
tate 13 to yield the target intermediate 8. Attachment of
fragments 7 to one end of the p-sexiphenyl 8 by Suzuki cou-
pling was accomplished following previously reported proce-
dures.[39] The obtained p-septiphenyl 24 was subjected to an-
other Suzuki coupling with the K4-terminal fragment 6. The
product was the desired p-octiphenyl scaffold with three dif-
ferent substituents, two of them being chemically differenti-
ated by orthogonal protecting groups.
Figure 1. Self-assembly from monomer 4 and the theoretical active struc-
ture of pore 1 with hydrophilic anchors together with anchor-free control
pores 2 and 3. b Sheets are shown as gray arrows in the theoretical pore
structure, and in the chemical structure, the external amino acid residues
are indicated within an empty circle, whereas the internal residues are in-
dicated within a filled circle (the residues are indicated with single-letter
abbreviations, see Scheme 2 for full structures).
Results and Discussion
In pore 1, which is introduced in this report, the rigid-rod
b barrel of the classical[32–36] pores 2 and 3 is elongated with
four of Tomichꢁs hydrophilic tetralysine (K4) anchors
(Figure 1).[29] Rigid-rod b barrels in general consist of p-octi-
phenyl staves and b-sheet hoops.[32,33] The selected peptide
sequence LRLHL produces a hydrophobic outer pore sur-
face (LLL) for contacts with the surrounding bilayer mem-
brane. Functional arginine–histidine (RH) dyads are placed
at the inner pore surface to interact with analytes passing
by, through the pore, and across the membrane.[12,13,32,33]
Except for the NMR tags in position 16 and 86, the
12,22,33,42,53,62,73,82 substitution pattern of the p-octiphenyl
stave in pore 2 is identical with that in pore 1.[34] The charac-
The NMR spectra of asymmetric oligomers such as 5 can
be challenging to fully understand because the presence of
several quasi-identical repeats causes extensive signal clus-
tering.[41] However, the recording of high-resolution
2D HSQC and HMBC NMR spectra allowed unambiguous
teristics of pore
2
and pore
3 with the classical
13,23,32,43,52,63,72,83 motif are essentially identical.[34–36]
The cationic K4 anchors were selected because they were
best in an extensive optimization with readily accessible
1948
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1947 – 1953

Synthesis of Synthetic Pore Sensors
FULL PAPER
Scheme 1. a) Cs2CO3, acetone, 608C, 30 min, yield=75%; b) I2, AgOTf, CHCl3, 2.5 h, RT, yield=40%; c) Cs2CO3, acetone, 658C, 30 min, yield=92%;
d) pinacolborane, [PdCl2(dppf)], Et3N, CH3CN, 2 h, 858C, yield=60%; e–h) see reference [39]; e) AgOTf, I2, yield=72%; f) Cs2CO3, yield=96%; g) pi-
nacolborane, [PdCl2(dppf)], yield=88%; h) KHF2, quantitative; i–m) see reference [30]; i), KI, yield=70%; j) pinacolborane, [PdCl2(dppf)], yield=
69%; k) [Pd(PPh3)4], yield=14%; l) BBr3; m) Cs2CO3, 92% (from 22); n) [PdCl2(dppf)], conversion yield 42% (from reference [39]); o) [Pd(PPh3)4],
toluene/EtOH 10:1, Na2CO3, yield=80%. dppf=1,1’-bis(diphenylphosphanyl)ferrocene.
ACHTREUNG
A
ACHTREUNG
A
N
ACHTREUNG
assignment of all carbon and hydrogen atoms of the key in-
termediate 5 (see Figure S1 in the Supporting Information).
The Z-protected K4 anchor 25 was newly prepared in 8
homogeneity and identity of the final product (see Figur-
es S2 and S3 in the Supporting Information).[37]
Pore 1 was characterized in egg yolkphosphatidylcholine
large unilamellar vesicles (EYPC-LUVs) that were loaded
with the fluorophore 8-aminonaphtalene-1,3,6-trisulfonate
(ANTS) and the quencher p-xylenebis(pyridinum)bromide
(DPX; EYPC-LUVsꢀANTS/DPX).[35,42] In this assay, pore
activity is detected as fluorogenic efflux of ANTS, DPX, or
both. Under these conditions, a Hill coefficient n=4.0ꢁ0.2
steps by
a
routine solution-phase peptide synthesis
(Scheme 2). The synthesis of the protected LRLHL peptide
26 has been reported previously.[34–36] Chemoselective hydro-
genolysis of the terminal benzyl ester of rod 5 liberated the
carboxylic acid at one rod terminus without deprotection of
those along the rigid-rod scaffold. The deprotected carboxyl-
ic acid at one end of rod 27 was coupled with the N termi-
nus of the Z-protected K4 anchor 25. The K4 rod 28 was ob-
tained in good yield. The carboxylic acids along the rigid-
rod scaffold 28 were deprotected next with trifluoroacetic
acid (TFA). The orthogonal Z protection of the K4 anchor
remained intact in product 29. The Pmc/Trt-protected
(Pmc=2,2,5,7,8-pentamethylchroman-6-sulfonyl, Trt=trityl)
LRLHL pentapeptides 26 were coupled to the liberated car-
boxylic acids along the scaffold of rod 29. The reasonably
complex product 30 was obtained in good yield. Removal of
all protecting groups with HBr/AcOH gave the target rod 4.
Reversed-phase HPLC, MS, and NMR spectra confirmed
*
was found for pore 1 (Figure 2 A, ). Moreover, pore 1
*
could reach full fractional activity Ymax ꢂ1.0 (Figure 2B, ).
This characteristic (n>1, Y=1) behavior demonstrates en-
dergonic (n>1) self-assembly of tetrameric pores (nꢂ4)
without competing precipitation from the water at higher
concentration (Ymax ꢂ1).[42–44]
These values differed clearly from the previously reported
n=0.6 and Ymax ꢂ0.4 for the anchor-free pore 3 (Figure 2,
[35,44]
*
).
This equally characteristic,[42–44] but undesired and
more complex (n=1, Y<1) behavior demonstrates exergon-
ic (n=1) self-assembly of pores of unknown stoichiometry
(n=1) that occurs already in the water. Exergonic assembly
Chem. Eur. J. 2008, 14, 1947 – 1953
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1949

S. Matile et al.
The biphasic Hill plot of pore 1 confirmed that values
measured above the effective pore concentration EC50 value
(i.e., concentration needed to observe 50% pore activity)
reflect the saturation of the assay rather than thermodynam-
ics and cooperativity of pore formation. They should not be
used for the determination of Hill coefficients.
In an attempt to further lower the EC50 value of pore 1,
the dependence of activity on the concentration of EYPC-
LUVs loaded with 5(6)-carboxyfluorescein (EYPC-
LUVsꢀCF)[1,25,42] was examined. In this more sensitive
assay, pore activity is detected as fluorogenic efflux of CF
because local dilution reduces self-quenching. With vesicle
dilution, the EC50 value of pore 1 in EYPC vesicles de-
creased to saturation around a minimal EC50,min =15 nm (Fig-
*
ure 3 A, ). Note that pore concentrations were calculated
Scheme 2. a) H2, Pd/C, THF, 1.5 h, RT, yield=96%; b) HATU, TEA,
DMF, 3.5 h, RT, yield=63%; c) TFA, 2 h, RT, yield=94%; d) HATU,
TEA, DMF, 5.5 h, RT, yield=75%; e) HBr/AcOH, thioanisole, TFA,
pentamethylbenzene, 1.5 h, RT, yield=68%. Bn=benzyl, DMF=N,N-di-
methylformamide, HATU=O-(7-azabenzotriazole-1-yl)-1,1,3,3-tetrame-
thyluronium hexafluorophosphate, TEA=triethylamine.
Figure 3. A) The dependence of the EC50 value of pore 1 on the concen-
*
*
tration of EYPC ( ) and EYPC/EYPG 1:1 ( ) LUVsꢀCF (given as the
total lipid concentration cL), and B) The dependence of the IC50 value of
*
*
analytes 31 ( ) and 32 ( ) to blockpore 1 on the concentration of
EYPC vesicles (given as lipid concentration cL).
as 25% of the known concentration of monomer 4 (n=4).
Considering the endergonic self-assembly of the active tetra-
mer 1, it is understood that this approximation represents a
clear, systematic overestimation.
[35]
*
*
)
Figure 2. Dependence of the activity Y of pores 1 ( ) and 3 (
in
EYPC-LUVsꢀANTS/DPX on the concentration of the monomeric rods
4. Data points for 3 (but not for 1) above 200 nm are not very well repro-
ducible because of competing precipitation from the media.
a-Ketoglutarate hydrazone 31 and pyruvate hydrazone 32
were attractive test analytes to elaborate on the impact of
vesicle dilution on stoichiometric binding in sensing applica-
tions (Scheme 3). The discrimination of a-ketoglutarate 33
and pyruvate 34 is, for example, of interest for umami sens-
ing with synthetic pores as transducers of signals generated
with transaminases. However, all attempts to do so failed so
far because of the unavailability of synthetic (or biological)
pores that would close or open in response to the recogni-
tion of either analyte 33 or analyte 34 at reasonable concen-
trations. This lackin sensitivity has been overcome with the
introduction of amplifiers such as 35.[1] This hydrazide can
react in situ with aldehydes and ketones produced or con-
sumed during enzymatic signal generation and drag them
into the pore for detection. Applied to umami sensing, reac-
tive amplification of a-ketoglutarate and pyruvate was
found to increase the sensitivity of synthetic pore sensors
similar to 2 and 3 by more than four orders of magnitude
without interference from amplifier 35 (IC50 =23 mm; IC50 =
blocker concentration required for 50% blockage). The dis-
crimination between a-ketoglutarate hydrazone 31 and pyr-
of pore 3 accounts for both better assembly formation (i.e.,
higher activity) at low concentrations as well as continuing
self-assembly at high concentrations[45] (i.e., precipitation
from the water before reaching the membrane, Ymax <1);
the result is simple saturation behavior.[42–45]
As elaborated in the introduction, the failure to reach full
pore activity because of competing precipitation from the
water (i.e., Ymax <1, Figure 2B, ) is one of the key obsta-
*
cles in practical applications of ion channels and pores as
drugs, sensors, or catalysts. The clear shift from n=0.6 to
n=4.0 confirmed that aqueous anchoring successfully sup-
pressed this exergonic self-assembly in the aqueous phase.
Without interference from precipitation at high concentra-
tions, pore 1 was able reach maximal activity (i.e., Ymax ꢂ1,
*
Figure 2B, ). This access to significant pore activity is,
from a practical point of view, one of the most important
findings with respect to hydrophilic anchoring.
1950
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1947 – 1953

Synthesis of Synthetic Pore Sensors
FULL PAPER
fluorometric detection of many enzymes (e.g., kinases) and
the sensing of their substrates.[1,12,13,23,25,36] Under standard
conditions, blockage of pore 1 by ATP occurred with IC50 =
44.2 mm and was clearly better than blockage by ADP
(Table 1, entries 5 and 6). The high discrimination factor of
D=6.1 suggested, together with relatively poor IC50 values,
that nucleotide recognition is not affected by stoichiometric
binding. Comparison with results from anchor-free pores ob-
tained under similar conditions demonstrated that ATP/
ADP discrimination is not significantly affected by the hy-
drophilic K4 anchors. Slightly weaker discrimination factors
were reported for anchor-free pores with internal RH dyads
Scheme 3. Sensing scheme for umami sensing with transaminase as the
signal generator, hydrazide 35 as the reactive amplifier, and synthetic
pores such as 1 as optical transducers.
as in 1 and 3[36] (IC50
discrimination factors were reported for pores with internal
KH dyads (IC
50(ATP)=22 mm, D=10.0).[1,12,13] Although
A
AHCTREUNG
less-specific and maybe transient interactions with anionic
guests are of course likely to occur, this overall excellent
similarity with anchor-free pores implied that the K4 anchors
of pore 1 do not contribute constructively to guest recogni-
tion.
The IC50 value of both nucleotides did not decrease and
their discrimination factor D did not increase with a de-
creasing concentration of pore 1 at a reduced vesicle con-
centration (Table 1, entries 7 and 8). This control experi-
ment confirmed that the increasing sensitivity and selectivity
obtained for the amplifier conjugates 31 and 32 originates
from a reduced interference from stoichiometric binding
(Table 1, entries 1–4 versus entries 5–8).
The determined increase in sensor sensitivity in response
to a reduction of the concentration of pores with increased
activity was consistent with the theory of stoichiometric
binding. This phenomenon has been studied in detail in the
context of enzyme inhibition.[26,27] Assuming one to one
binding of the blocker to the pore, one could correlate con-
centrations of pore [P], blocker [B], and the dissociation
constant KD as is shown in Equation (1).
uvate hydrazone 32 was, however, negligible because of, we
speculated, stoichiometric binding.[1]
The detection of amplified a-ketoglutarate 31 and pyru-
vate 32 by pore 1 under standard conditions occurred with
near nanomolar sensitivity (Table 1, entries 1 and 2). As ex-
Table 1. Blockage data for pore 1.
Entry
Blocker
Lipid [mm][a]
Pore [nM][b]
IC50 [mm][c]
D[d]
1
2
3
4
5
6
7
8
31
32
31
32
ATP
ADP
ATP
ADP
124
124
6.4
50
50
15
15
50
50
15
15
1.7ꢁ0.1
1.3
–
3.0
–
6.1
–
4.7
–
2.2ꢁ0.3
0.4ꢁ0.1
6.4
1.2ꢁ0.2
124
124
6.4
44.2ꢁ3.0
267.7ꢁ12.6
139.5ꢁ12.4
652.1ꢁ19.7
6.4
[a] Approximated from phosphate analysis of final vesicles, assuming a
reproducible yield. [b] Approximated as c(pore 1)=c(monomer 4)/4 (n=
AHCTREUNG
4, Figure 2 A); this is a significant overestimate because pore assembly is
endergonic. [c] Data ꢁ standard error. [d] Discrimination factor=IC50-
A
N
½BŠ=KD ¼ i=ð100ÀiÞ þ i½PŠ=100 KD
ð1Þ;
pected for molecular recognition in the zone affected by in-
terference from stoichiometric binding, the discrimination
between the two analytes was nearly negligible with D=1.3
(where D is the discrimination factor). The IC50 values of
both hydrazones decreased in response to the reduction of
the EC50 of pore 1 (Figure 3B and Table 1, entries 3 and 4).
These changes in the IC50 value resulted in an increase in
sensitivity of up to 4.3-fold (Table 1, entries 1 and 3). Also
exactly as predicted for reduced interference from stoichio-
metric binding, this increase in sensitivity coincided with a
quite remarkable increase in selectivity. The discrimination
factor D=3.0 found for the detection of amplified a-keto-
glutarate with nanomolar sensitivity (IC50 =400 nm) is suffi-
cient for sensing applications.[1,36]
In Equation (1), i is the percent blockage. As the IC50 value
is the concentration of the blocker that gives rise to 50%
pore blockage, the above equation can be simplified as is il-
lustrated in Equation (2).
IC50=KD ¼ 1 þ 0:5½PŠ=KD
ð2Þ
Taking this simplification into account, Equation (3) can be
derived.
IC50 ¼ KD þ 0:5½PŠ
ð3Þ
Therefore, if the KD value of the pore–blocker complex is
similar or less than the concentration of the pore (KD =[P]),
the IC50 values should decrease by lowering the pore con-
centration. On the other hand, if the KD value is far greater
than the concentration of the pore, the second term in
Equation (3) (0.5[P]) becomes negligible and thus the IC50
To confirm the validity of the interpretation of this break-
through, control dilution experiments without stoichiometric
binding were made. ATP/ADP discrimination (ATP=ade-
nosine 5’-triphosphate, ADP=adenosine 5’-diphosphate)
provides an ideal example because it is important for the
Chem. Eur. J. 2008, 14, 1947 – 1953
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1951

S. Matile et al.
value would be equal to KD.
*
). The determined decreasing activity of pore 1 with in-
The observed decrease in the IC50 value of blockers 31
and 32 by decreasing the concentration of the pore is there-
fore reasonable if we assume that [P] is similar to or higher
than KD (Table 1, Equation (3)). Note that the concentration
of active pores is unknown but certainly less than 25% of
the concentration of monomer 4 because of endergonic self-
assembly and nonideal partitioning. The KD value of 32 is
thus expected to be well below 15 nm and that of 31 is ex-
pected to be even lower because the response to reduced
stoichiometric binding is more pronounced (KD(31)<
KD(32)<[P]).
In contrast, the increase in the IC50 value of ATP and
ADP by decreasing the concentration of the pore is not ex-
plained by Equation (3). More-complex inhibitory mecha-
nisms may account for these results, including changes in
partitioning etc.
creasing surface potential could be explained by preferential
ion pairing of lipid phosphates with oligoarginines[47–50] from
the barrel rather than oligolysines from the anchor and thus
the destruction of the artificial b barrel. Synthetic efforts
toward anchor screening in situ and more specifically target-
ed pore formation for hypersensitive multianalyte sensing
are ongoing.
Acknowledgements
We thankA. Pinto and S. Grass for NMR measurements, P. Perrottet
and the group of F. GülaÅar for MS measurements, one referee for help-
ful suggestions, and the Swiss NSF for financial support.
[1] S. Litvinchuk, H. Tanaka, T. Miyatake, D. Pasini, T. Tanaka, G.
Bollot, J. Mareda, S. Matile, Nat. Mater. 2007, 6, 576–580.
[2] T. M. Fyles, Chem. Soc. Rev. 2007, 36, 335–347.
[3] A. P. Davis, D. N. Sheppard, B. D. Smith, Chem. Soc. Rev. 2007, 36,
348–357.
[4] J. T. Davis, G. P. Spada, Chem. Soc. Rev. 2007, 36, 296–313.
[5] R. S. Hector, M. S. Gin, Supramol. Chem. 2005, 17, 129–134.
[6] U. Koert, L. Al-Momani, J. R. Pfeifer, Synthesis 2004, 8, 1129–1146.
[7] S. Matile, A. Som, N. Sordꢂ, Tetrahedron 2004, 60, 6405–6435.
[8] G. W. Gokel, A. Mukhopadhyay, Chem. Soc. Rev. 2001, 30, 274–
286.
[9] P. Scrimin, P. Tecilla, Curr. Opin. Chem. Biol. 1999, 3, 730–735.
[10] Y. Kobuke, K. Ueda, M. Sokabe, J. Am. Chem. Soc. 1992, 114,
7618–7622.
[11] J.-H. Fuhrhop, U. Liman, V. Koesling, J. Am. Chem. Soc. 1988, 110,
6840–6845.
[12] G. Das, S. Matile, Chem. Eur. J. 2006, 12, 2936–2944.
[13] S. Matile, H. Tanaka, S. Litvinchuk, Top. Curr. Chem. 2007, 277,
219–250.
[14] Creative Chemical Sensing Systems. Topics in Current Chemistry
(Ed.: T. Schrader), Springer, Heidelberg, 2007.
[15] C. Zhang, K. S. Suslick, J. Agric. Food Chem. 2007, 55, 237–242.
[16] S. Tamaru, S. Kiyonaka, I. Hamachi, Chem. Eur. J. 2005, 11, 7294–
7304.
[17] J. J. Lavigne, E. V. Anslyn, Angew. Chem. 2001, 113, 3212–3225;
Angew. Chem. Int. Ed. 2001, 40, 3118–3130.
[18] A. Buryak, K. Severin, J. Am. Chem. Soc. 2005, 127, 3700–3701.
[19] K. Toko, Biosens. Bioelectron. 1998, 13, 701–709.
[20] C. A. Marquette, A. Degiuli, L. J. Blum, Biosens. Bioelectron. 2004,
19, 433–439.
[21] L. Campanella, A. Bonanni, E. Finotti, M. Tomassetti, Biosens. Bio-
electron. 2004, 19, 641–651.
[22] A. Hennig, H. Bakirci, W. M. Nau, Nat. Methods 2007, 4, 629–632.
[23] L. Vial, P. Dumy, J. Am. Chem. Soc. 2007, 129, 4884–4885.
[24] J. P. Goddard, J.-L. Reymond, Curr. Opin. Biotechnol. 2004, 15,
314–322.
[25] G. Das, P. Talukdar, S. Matile, Science 2002, 298, 1600–1602.
[26] B. K. Soichet, J. Med. Chem. 2006, 49, 7274–7277.
[27] O. H. Straus, A. Goldstein, J. Gen. Physiol. 1943, 26, 559–585.
[28] P. L. Boudreault, N. Voyer, Org. Biomol. Chem. 2007, 5, 1459–1465.
[29] J. R. Broughman, L. P. Shank, W. Takeguchi, B. D. Schultz, T. Iwa-
moto, K. E. Mitchell, J. M. Tomich, Biochemistry 2002, 41, 7350–
7358.
Conclusion
In summary, this report introduces hydrophilic anchoring as
a promising strategy to increase the sensitivity and the selec-
tivity of synthetic pore sensors. With regard to the activity
of the pores as such, the change in the Hill coefficient from
n<1 to n=4 is the most important impact of hydrophilic
anchoring. It demonstrates successful suppression of the
competing precipitation from the media. The presumably
disorganized tetralysine anchors do not interfere with the
molecular recognition of anionic blockers by the synthetic
pore sensors. The increase in pore activity achieved by hy-
drophilic anchoring is important because it allowed us to de-
mystify one of the key problems with pore sensors, that is,
the origin of the consistent cut-off in sensitivity and selectiv-
ity near nanomolar concentrations. Namely, efficiency and
selectivity of pore blockers that operate by stoichiometric
binding (KD =EC50) are shown to increase in response to a
reduction of the effective pore concentration. Controls con-
firm that this increase does not occur with analytes that are
too weakto exhibit stoichiometric binding ( KD >EC50).
This breakthrough is significant because it demonstrates
that sensitivity and selectivity of pore sensors can be im-
proved by increasing the activity of the pore. The lesson
learnt is that, with sufficiently sensitive analyte recognition
and amplification accomplished,[1] rational approaches
toward “hypersensitive” pore sensors will have to focus on
pore–membrane rather than pore–blocker interactions. Ex-
amples from biology in support of this conclusion include
nisin. This antibiotic peptide may act by targeted pore for-
mation with lipid II at concentrations far below the intrinsic
cut-off with other magainin-like natural antibiotics.[46]
This conclusion implied that the introduction of specific
lipid–pore interactions would be the key to further increases
in sensitivity and selectivity of pore sensors. Incorporation
of anionic EYPG into the bilayer so that it could participate
in an ion pair with the K4 anchor of pore 1 was, however,
too simplistic an approach toward this objective (Figure 3 A,
[30] N. Sakai, D. Gerard, S. Matile, J. Am. Chem. Soc. 2001, 123, 2517–
2524.
[31] K. Yamashita, V. Janout, E. Bernard, D. Armstrong, S. L. Regen, J.
Am. Chem. Soc. 1995, 117, 6249–6253.
[32] N. Sakai, J. Mareda, S. Matile, Acc. Chem. Res. 2005, 38, 79–87.
1952
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1947 – 1953

Synthesis of Synthetic Pore Sensors
FULL PAPER
[33] R. Bhosale, S. Bhosale, G. Bollot, V. Gorteau, M. D. Julliard, S. Lit-
vinchuk, J. Mareda, S. Matile, T. Miyatake, F. Mora, A. Perez-Velas-
co, N. Sakai, A. L. Sisson, H. Tanaka, D.-H. Tran, Bull. Chem. Soc.
Jpn. 2007, 80, 1044–1057.
[43] S. Bhosale, S. Matile, Chirality 2006, 18, 849–856.
[44] S. Litvinchuk, G. Bollot, J. Mareda, A. Som, D. Ronan, M. R. Shah,
P. Perrottet, N. Sakai, S. Matile, J. Am. Chem. Soc. 2004, 126,
10067–10075.
[34] D. Ronan, D. Jeannerat, A. Pinto, N. Sakai, S. Matile, New J. Chem.
2006, 30, 168–176.
[35] P. Talukdar, N. Sakai, N. Sordꢂ, D. Gerard, V. M. F. Cardona, S.
Matile, Bioorg. Med. Chem. 2004, 12, 1325–1336.
[36] N. Sordꢂ, S. Matile, Biopolymers 2004, 76, 55–65.
[37] See the Supporting Information.
[38] M. Murata, S. Watanabe, Y. Masuda, J. Org. Chem. 1997, 62, 6458–
6459.
[39] N. Sakai, A. L. Sisson, S. Bhosale, N. Banerji, A. Fürstenberg, E.
Vauthey, S. Matile, Org. Biomol. Chem. 2007, 5, 2560–2563.
[40] G. A. Molander, B. Biolatto, J. Org. Chem. 2003, 68, 4302–4314.
[41] D. Jeannerat, D. Ronan, Y. Baudry, A. Pinto, J.-P. Saulnier, S.
Matile, Helv. Chim. Acta 2004, 87, 2190–2207.
[45] G. Das, L. Ouali, M. Adrian, B. Baumeister, K. J. Wilkinson, S.
Matile, Angew. Chem. 2001, 113, 4793–4797; Angew. Chem. Int. Ed.
2001, 40, 4657–4661.
[46] E. Breukink, B. de Kruijff, Nat. Rev. Drug Discovery 2006, 5, 321–
332.
[47] N. Sakai, S. Matile, J. Am. Chem. Soc. 2003, 125, 14348–14356.
[48] J. B. Rothbard, T. C. Jessop, R. S. Lewis, B. A. Murray, P. A.
Wender, J. Am. Chem. Soc. 2004, 126, 9506–9507.
[49] T. Takeuchi, M. Kosuge, A. Tadokoro, Y. Sugiura, M. Nishi, M.
Kawata, N. Sakai, S. Matile, S. Futaki ACS Chem. Biol. 2006, 1,
299–303.
[50] D. Schmidt, Q.-X. Jiang, R. MacKinnon, Nature 2006, 444, 775–779.
[42] S. Matile, N. Sakai, The Characterization of Synthetic Ion Channels
and Pores. In Analytical Methods in Supramolecular Chemistry,
(Ed.: C. A. Schalley), Wiley, Weinheim, 2007, 391–418.
Received: September 25, 2007
Revised: November 6, 2007
Published online: December 7, 2007
Chem. Eur. J. 2008, 14, 1947 – 1953
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1953