Amplifier-mediated activation of cell-penetrating peptides with steroids: Multifunctional anion transporters for fluorogenic cholesterol sensing in eggs and blood

Article abstract of DOI:10.1002/anie.200804422

(Figure Presented). Off-the-shelf, pret-a-porter multianalyte sensing systems are based on the ability of cell-penetrating peptides (CPPs; see picture, blue) to mediate the export of hydrophilic anions (yellow → green) from lipid bilayer vesicles (gray). Covalent capture of hydrophobic analytes (black) with hydrophilic anions (red) produces amphiphilic anions that can activate CPPs.

Full text of DOI:10.1002/anie.200804422

Angewandte
Chemie
DOI: 10.1002/anie.200804422
Biosensors
Amplifier-Mediated Activation of Cell-Penetrating Peptides with
Steroids: Multifunctional Anion Transporters for Fluorogenic
Cholesterol Sensing in Eggs and Blood**
Sara M. Butterfield, Tomohiro Miyatake, and Stefan Matile*
Cell-penetrating peptides (CPPs), such as HIV-1 Tat, poly-
arginine (pR), and a rich collection of synthetic mimics, are
attracting increasing attention owing to their usefulness in
[
1]
drug delivery, and, more recently, as optical transducers of
[
2]
reactions. We previously demonstrated that certain amphi-
philic anions can activate CPPs to move across bulk, lipid-
[
2,3]
bilayer as well as live-cell membranes. The concept of CPP
activators builds on the fact that arginine-rich polycations
permanently coexist with readily exchangeable counteranions
to minimize intramolecular charge repulsion between the
[
3]
weakly acidic guanidinium cations. Exchange of the hydro-
philic counterions of CPPs in water by amphiphilic anion
activators affords hydrophobic CPP–activator complexes that
[3]
can move across the membrane. A two-step counterion
exchange from hydrophilic to amphiphilic and back to
hydrophilic anions is sufficient for CPPs to overcome the
[
3]
membrane barrier and enter cells. Activated CPP–counter-
ion complexes can bind small hydrophilic anions such as 5(6)-
carboxyfluorescein (CF) by partial counterion exchange and
transport them across bulk and lipid bilayer membranes
(
Figure 1), although other mechanisms might also contribute
[
2,3]
to anion export from lipid bilayer vesicles.
Owing to their commercial availability, CPP–activator
complexes are attractive for sensing applications. However, to
[
4–6]
date they have proven to be inferior to synthetic pores
in
addressing critical issues such as the discrimination of
[2]
Figure 1. Fluorogenic steroid sensing systems with CPP transporters
adenosine tri- and diphosphate (ATP and ADP) or of
1
[5]
as signal transducers and signal amplifiers (e.g., A ) that react with
phytate and IP (diphosphoinositol pentakisphosphate). For
7
1
hydrophobic analytes (e.g., S ) to generate amphiphilic CPP activators.
sensing with pores, enzymes have been introduced to react
selectively with the analyte of interest, thus generating an
For cholesterol, signal generation with cholesterol oxidase (a) is
1
1 1
followed by signal amplification with A (b). The obtained A S
[4]
1
1
analyte-specific signal. Multianalyte sensing in complex
matrices became possible with the introduction of signal
conjugate and CPP (e.g., pR) form pR–A S transporters (c) that
mediate the export of intravesicular CF (d–f). Amphiphilic anions
[6]
1 1
(
e.g., A S ) are CPP activators below (d–f) and CPP inactivators above
amplifiers. Signal amplifiers are bifunctional molecules that
can react in situ with the product of enzymatic signal
generation and activate or deactivate signal transducers
their cmc (g).
[
6]
[6]
such as synthetic pores. Herein, we report that with the
cation methods, one of the most persistent sensing problems
with synthetic pores, the detection of hydrophobic analytes,
can be solved with CPPs. Sensing of hydrophobic analytes is a
challenge with pore sensors, because they tend to avoid the
hydrophilic interior of the pore and prefer partitioning into
the membrane. We demonstrate that covalent capture with
hydrophilic anions produces amphiphilic anions that can
activate CPP transporters. Coupled with enzymatic signal
generation, this hybridization of activators and amplifiers
affords cheap, fluorogenic, and interference-free CPP-based
sensors for hydrophobic analytes in complex matrices.
[2,3]
combination of counterion activation
and signal amplifi-
[
*] Dr. S. M. Butterfield, Prof. T. Miyatake, Prof. S. Matile
Department of Organic Chemistry, University of Geneva
Genꢀve (Switzerland)
Fax: (+41)22-379-3215
E-mail: stefan.matile@unige.ch
Homepage: http://www.unige.ch/sciences/chiorg/matile/
[**] We thank the University of Geneva and the Swiss NSF for financial
support.
Steroids provide attractive examples to illustrate this
approach because they are notoriously insoluble and difficult
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200804422.
Angew. Chem. Int. Ed. 2009, 48, 325 –328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
325

Communications
[
7]
1
to measure in aqueous media, and because their detection is
Table 1: Steroid detection with pR and amplifier A in the presence
(
[
a]
[8]
upper rows) and absence (lower rows) of nonmicellar triton X-100.
essential in domains such as medicinal diagnostics and
environmental monitoring. In the field of chemo- and
previous approaches to address these needs
included enzyme immobilization on electrochemical surfa-
[
9]
[
b]
[c]
Steroid
EC50 [mm]
Ymax [%]
[
10–12]
biosensors,
1
2
1
2
S
S
0.6ꢁ0.03
4.0ꢁ1.0
18.0ꢁ9.0
60
36
1
[
11]
ces
and chemical conjugation of steroid receptors to
[
12]
24
fluorescent proteins.
The concept of amplifier-mediated
[d]
[d]
–
–
fluorogenic steroid sensing with cell-penetrating peptides
builds on the ability of pR–counteranion complexes to export
CF from egg yolk phosphatidylcholine large unilamellar
vesicles (EYPC LUVs, Figure 1). In this assay, CF is loaded
within the vesicle at concentrations high enough for self-
quenching, and the local CF dilution during CF export by pR–
counteranion complexes is monitored as an increase in CF
fluorescence. The CF intensity is then normalized to an
activity factor, the fractional activity Y.
compatibility of this method with reactive signal amplification
strategies, we used the hydrazone approach
the steroid analyte S with anionic hydrazides and hydrazines
3
4
3
4
S
S
11.0ꢁ0.4
57
12
1
7.0ꢁ2.0
7.5ꢁ0.1
[
2]
39
37
9
.3ꢁ0.2
5
6
7
5
6
7
S
S
S
1.3ꢁ0.07
3.1ꢁ0.5
74
65
7.1ꢁ0.3
3.0ꢁ1.0
74
63
[
14]
To explore the
1
[
6]
[6,13]
10.0ꢁ0.2
8.0ꢁ0.4
50
30
and treated
1
1
1
5[14]
A –A
in DMSO at 608C for one hour. Without any further
sample workup or purification, the efficiency with which the
[
a] Determined from dose response curves for fluorogenic CF export
from EYPC-LUVsꢀCF in the presence of pR and after hydrazone
formation with one equivalent of CB amplifier A (Figures 1 and 2).
1
1
5 1
1
[14]
resulting hydrazone amphiphiles A S –A S activated pR was
measured in EYPC-LUVsꢀCF. Compared to the singly
1
n
[b] Concentration of activator A S needed to observe 50% of the
2
1
5
1
1 1
maximal pR activity Ymax, data ꢁstandard deviation. EC values depend
charged amphiphiles A S –A S , the A S amphiphile
50
[
6]
on parameters such as vesicle concentration (decreasing with decreas-
ing LUV concentration), lipid composition, surface potential, membrane
potential, temperature, and ionic strength. [c] Maximal fluorescence
obtained from the triply charged cascade blue (CB)
1
hydrazide A activated pR more efficiently (EC = (14.0 ꢁ
5
0
1
.0) mm), as determined from the resulting increase in
1 n
emission intensity found at maximal nonmicellar activator A S concen-
tration relative to emission after membrane lysis with triton X-100, ꢁ5%
error. [d] Not detectable.
fluorescence emission during the export of CF, and was
clearly the best in the series (Table 1). Hydrazide A alone did
1
not induce pR activation up to concentrations of 500 mm. For
comparison, CF efflux through synthetic pores was hindered
1
[6]
1 1
by both A alone (IC = 22 mm) and hydrazones A S
5
0
(
IC = 44 mm) with similar efficiencies. Originating presum-
50
1
1
ably from partial nonspecific binding of amphiphile A S to
the hydrophobic membrane rather than the hydrophilic pore
interior, this nearly negligible discrimination factor D = 2 was
insufficient for sensing. The interference-free (D = 1), sensi-
1
tive, and fluorogenic detectability of steroid S thus repre-
sented the first example in which CPP-based sensors are
[
2,5]
clearly superior to synthetic pores.
1
1
The dose response curve of A S activators at constant pR
concentration was paraboloidal (Figure 2a, *) rather than
Figure 2. a) Dose response curves for pR activation with increasing
1
1
concentrations c of A S in the presence of 0 (*), 50 (&), and 80 mm
*) triton X-100. b) Dose response curve for pR activation with
increasing volumes of egg yolk extract after signal generation with
1
1
sigmoidal as usual. This result suggested that A S activators
form micelles at higher concentrations; these micelles then
act as hydrophilic polyanions, binding CPPs and keeping them
away from the membrane (Figure 1g). The maximal activity
Ymax in the dose response curve should thus correspond to the
critical micelle concentration (cmc) of the activator (ca. 50 mm
(
ꢂ1
ꢂ1
cholesterol oxidase (*: 0.45 unitsmL , *: 0 unitsmL (no enzyme)
ꢂ1
plus peroxidase (37 unitsmL ) and subsequent signal amplification
1
with A .
1
1
for A S , Figure 2a, *). This behavior is contrary to that of
common surfactants such as triton X-100, which, in the
absence of CPPs, is membrane-inactive as a monomer and
membrane-disruptive as micelle (EC = cmc = 200 ꢁ 8 mm).
aqueous phase. Increasing the concentration of triton X-100
1
1
up to 80 mm not only lowered the EC50 for A S to activate the
subsequently added pR transporters but also raised the
5
0
1
1
1 1
To increase the relatively modest Y = 36% of pR–A S
maximal activity of the resulting pR–A S complexes up to
an excellent Ymax = 60% (Figure 2a, *, and Table 1).
max
complexes (Figure 2a, *), the use of solubilizing additives was
considered. In our hands, b-cyclodextrin, DMSO, dioxane,
DMF, acetonitrile, and tert-butyl alcohol were ineffective as
additives. However, submicellar concentrations of triton X-
n
Fluorometric detection of the steroid series S was
possible at low micromolar concentrations (Table 1). This
1
detection was achieved by covalent capture with A and
1
1
1
00 proved perfect to deliver activator A S to the membrane
subsequent activation of pR transporters in fluorogenic
1
n
with minimal losses from competing precipitation from the
vesicles with the obtained A S conjugate. The least respon-
3
26
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 325 –328

Angewandte
Chemie
1
2
1 [a]
sive activator was estrone hydrazone A S , for which no pR
Table 2: Cholesterol sensing with pR and amplifier A .
activation was observed without additives, and a Ymax of only
Entry Sample
Expected
Found
[mgg
UV control
ꢂ1 [b]
ꢂ1 [c]
ꢂ1 [d]
2
2% was reached in the presence of triton X-100 (Table 1,
[mgg
]
]
[mgg
]
entry 2). The pR activating efficiencies (i.e., 1/EC ) of the
5
0
1
2
egg yolk
black lumpfish 3.0–3.5
eggs
human blood
serum
9–20
10.2ꢁ1.0
3.5ꢁ0.5
n.d.
1
1
steroid amphiphiles followed the order: A S (choleste-
3.5ꢁ0.1
1.0ꢁ0.1
1
5
1
6
1
4
none) > A S (progesterone) > A S (corticosterone) ꢃ A S
1
3
1
7
(
pregnenalone) > A S
Table 1). This trend suggested that increasing hydrophilicity
(testosterone) ꢃ A S
(cortisone;
3
1.5–2.5
1.2ꢁ0.2
reduces activator efficiency. The outstanding EC values of
A S and progesterone amphiphile A S were independent of
5
0
[a] Determined from dose response as activation of pR in fluorogenic
1
1
1
5
[16–18]
vesicles (Figure 2). [b] From supplier/literature.
[c] Calibrated
1
1
number and location of the reactive ketone groups, but
against the dose response curve of A S . [d] Determined using a
[
14]
[
15]
colorimetric method for cholesterol detection.
correlated with the maximal logP values of the steroids and
the intact hydrophobicity of their dimethylated b faces
(
Table 1, entries 1 and 5). The detectability of many different
1
1
analytes is important to assure broad applicability of CPP
sensors. The selective detection of single components in a
collection of accessible analytes is achieved in this case by the
selectivity of the enzymatic signal generator (cholesterol
oxidase); selective detection of the other steroids in complex
matrices is conceivable by appropriate choice of enzyme.
Intrinsic differences in reactivity implied that di- and
triketones S –S would give monohydrazone amphiphiles
when treated with equimolar amounts of A . ESI-MS analysis
supported the in situ formation of monohydrazones A S –
A S . The declining activation efficiency of corticosterone S
the calibration curve for A S (Table 2, entry 1). This result
was in excellent agreement with the typical range of 9–
20 mgg of cholesterol in egg yolk reported in the literature.
Following the same procedure, cholesterol contents were
determined in black lumpfish eggs
The obtained values were in agreement with
expectations from the literature and independent measure-
ments with a conventional UV assay for cholesterol (Table 2,
ꢂ1
[
17]
and human blood
[
18]
serum.
5
7
1
[14]
entries 2 and 3).
1
5
In summary, the present study introduces cell-penetrating
peptides (CPPs) as fluorogenic multianalyte sensors in
complex matrices, particularly for otherwise intractable
hydrophobic analytes. The approach combines two recently
described key concepts, that is, signal amplification by
1
7
6
1
with increasing molar fraction of A during incubation was
consistent with the formation of the hexaanionic bolaampi-
philic species A S A with increased pR affinity but reduced
affinity for the membrane. The same trend was found for
1
6
1
[6]
covalent analyte capture after enzymatic signal generation
and the activation of CPP transporters by anionic amphi-
7
triketone cortisone S .
[
2,3]
The usefulness of amplifier-mediated CPP activation to
detect the activity of enzymes with hydrophobic substrates
was explored next (Figure 1). Cholesterol was incubated with
cholesterol oxidase, peroxidase, and triton X-100 as solubil-
izer for varying reaction times, treated with signal amplifier
philes,
and it is exemplified with cholesterol sensing in
blood serum and other matrices. The exceptionally broad and
attractive perspectives assured by the general nature of the
sensing system are currently under investigation.
1
A , and subsequently added with pR to EYPC-LUVsꢀCF.
Received: September 8, 2008
Published online: December 3, 2008
Increasing CF emission was found with increasing enzyme
reaction time and with increasing enzyme concentration. This
result was consistent with increasing concentration of pR–
Keywords: amphiphiles · biosensors · ion transport ·
.
1
1
membranes · steroids
A S resulting from the enzymatic oxidation of cholesterol to
S .
1
[16]
Because of its high cholesterol content, egg yolk was an
ideal candidate to explore the compatibility of amplifier-
mediated CPP activation with sensing applications in samples
from supermarkets and hospitals (Table 2, entry 1). A dried
organic extract was prepared from egg yolk following a
standard protocol. This extract was exposed to enzymatic
signal generation with cholesterol oxidase, peroxidase, and
triton X-100 solubilizer and subsequent signal amplification
[
1] a) I. Nakase, T. Takeuchi, G. Tanaka, S. Futaki, Adv. Drug
Delivery Rev. 2008, 60, 598 – 607; b) E. A. Goun, T. H. Pillow,
L. R. Jones, J. B. Rothbard, P. A. Wender, ChemBioChem 2006,
7, 1497 – 1515; c) T. Shimanouchi, P. Walde, J. Gardiner, Y. R.
Mahajan, D. Seebach, A. Thomae, S. D. Kraemer, M. Voser, R.
Kuboi, Biochim. Biophys. Acta 2007, 1768, 2726 – 2736; d) S.
Pujas, J. Fernandez-Carneado, C. Lopez-Iglesias, M. J. Kogan, E.
Giralt, Biochim. Biophys. Acta Biomembr. 2006, 1758, 264 – 279.
2] T. Miyatake, M. Nishihara, S. Matile, J. Am. Chem. Soc. 2006,
[14]
1
with A . Increased volume of the resulting egg yolk extract
caused increasing activation of pR transporters (Figure 2b,
[
128, 12420 – 12421.
*
). Controls demonstrated that the egg yolk extract did not
[3] a) 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; b) F. Perret, M. Nishihara, T. Takeuchi, S. Futaki,
A. N. Lazar, A. W. Coleman, N. Sakai, S. Matile, J. Am. Chem.
Soc. 2005, 127, 1114 – 1115; c) M. Nishihara, F. Perret, T.
Takeuchi, S. Futaki, A. N. Lazar, A. W. Coleman, N. Sakai, S.
Matile, Org. Biomol. Chem. 2005, 3, 1659 – 1669.
activate pR without incubation with cholesterol oxidase
1
(
Figure 2b, *). This finding confirmed that the product S of
enzymatic oxidation is the only source of pR activator in the
egg yolk matrix and that subtraction of background contri-
butions is thus redundant.
[
5,6]
A cholesterol content of (10.2 ꢁ
ꢂ1
1
.0) mgg in egg yolk was obtained from comparison of the
[
4] a) G. Das, P. Talukdar, S. Matile, Science 2002, 298, 1600 – 1602;
b) S. Litvinchuk, N. Sordꢀ, S. Matile, J. Am. Chem. Soc. 2005,
dose response curve of egg yolk extract (Figure 2b, *) with
Angew. Chem. Int. Ed. 2009, 48, 325 –328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
327

Communications
1
27, 9316 – 9317; c) S. Matile, H. Tanaka, S. Litvinchuk, Top.
Angew. Chem. 2001, 113, 3212 – 3225; Angew. Chem. Int. Ed.
2001, 40, 3118 – 3130; f) N. A. Rakow, K. S. Suslick, Nature 2000,
406, 710 – 713.
Curr. Chem. 2007, 277, 219 – 250.
[
5] S. M. Butterfield, D.-H. Tran, H. Zhang, G. D. Prestwich, S.
Matile, J. Am. Chem. Soc. 2008, 130, 3270 – 3271.
6] a) S. Litvinchuk, H. Tanaka, T. Miyatake, D. Pasini, T. Tanaka, G.
Bollot, J. Mareda, S. Matile, Nat. Mater. 2007, 6, 576 – 580; b) S.
Hagihara, H. Tanaka, S. Matile, J. Am. Chem. Soc. 2008, 130,
[11] Z. Matharu, G. Sumana, S. K. Arya, S. P. Singh, V. Gupta, B. D.
Malhotra, Langmuir 2007, 23, 13188 – 13192.
[
[12] S. S. Muddana, B. R. Peterson, ChemBioChem 2003, 4, 848 – 855.
[13] a) P. T. Corbett, J. Leclaire, L. Vial, K. R. West, J.-L. Wietor,
J. K. M. Sanders, S. Otto, Chem. Rev. 2006, 106, 3652 – 3711; b) B.
Levrand, Y. Ruff, J.-M. Lehn, A. Herrmann, Chem. Commun.
2006, 2965 – 2967; c) Y. Takaoka, H. Tsutsumi, N. Kasagi, E.
Nakata, I. Hamachi, J. Am. Chem. Soc. 2006, 128, 3273 – 3280;
d) I. C. Carrico, B. L. Carlson, C. R. Bertozzi, Nat. Chem. Biol.
2007, 3, 321 – 322.
5
656 – 5657; c) S. Hagihara, H. Tanaka, S. Matile, Org. Biomol.
Chem. 2008, 6, 2259 – 2262.
[
7] a) P. Wallimann, T. Marti, F. Diederich, Chem. Rev. 1997, 97,
1
1
567 – 1608; b) A. Fꢁrer, T. Marti, F. Diederich, Helv. Chim. Acta
999, 82, 1843 – 1859.
[
8] F. R. Maxfield, I. Tabas, Nature 2005, 438, 612 – 621.
[
9] P. Westerhoff, Y. Yoon, S. Snyder, E. Wert, Environ. Sci. Technol.
[14] See the Supporting Information.
2
005, 39, 6649 – 6663.
[15] F. A. A. Nfflæez, S. H. Yalkowsky, J. Pharm. Sci. 1997, 86, 1187 –
1189.
[16] R.-Z. Zhang, L. Li, S.-T. Liu, R.-M. Chien, P.-F. Rao, J. Food
Biochem. 1999, 23, 351 – 361.
[17] C. Achard, J. Levy, N. Georgiakakis, Arch. Mal. Appar. Dig. Mal.
Nutr. 1934, 24, 785 – 793.
[
10] a) Creative Chemical Sensing Systems. Topics in Current Chemis-
try (Ed.: T. Schrader), Springer, Heidelberg, 2007; b) A. Hennig,
H. Bakirci, W. M. Nau, Nat. Methods 2007, 4, 629 – 632; c) J. Liu,
Y. Lu, Angew. Chem. 2006, 118, 96 – 100; Angew. Chem. Int. Ed.
2006, 45, 90 – 94; d) J.-P. Goddard, J.-L. Reymond, Curr. Opin.
Biotechnol. 2004, 15, 314 – 322; e) J. J. Lavigne, E. V. Anslyn,
[18] S. Westin, I. Heath, Br. Med. J. 2005, 330, 1461 – 1462.
3
28
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 325 –328