Modular and tunable chemosensor scaffold for divalent zinc

Article abstract of DOI:10.1021/ja0355980

A modular peptide scaffold has been developed for fluorescent sensing of divalent zinc. The signaling component of the chemosensor is the chelation-sensitive fluorophore 8-hydroxy-5-(N,N-dimethylsulfonamido)-2-methylquinoline, which is prepared as the pro

Full text of DOI:10.1021/ja0355980

Published on Web 08/09/2003
Modular and Tunable Chemosensor Scaffold for Divalent Zinc
Melissa D. Shults, Dierdre A. Pearce, and Barbara Imperiali*
Contribution from the Department of Chemistry, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139
Received April 11, 2003; E-mail: imper@mit.edu
Abstract: A modular peptide scaffold has been developed for fluorescent sensing of divalent zinc. The
signaling component of the chemosensor is the chelation-sensitive fluorophore 8-hydroxy-5-(N,N-
dimethylsulfonamido)-2-methylquinoline, which is prepared as the protected amino acid derivative Fmoc-
Sox-OH and integrated into peptide sequences. Nineteen synthetic peptides incorporating the signaling
element exhibit a range of affinities for Zn2+ through variation of the type and number of Zn2+ ligands,
ligand arrangement and the â-turn sequence that acts as a preorganization element between the ligands.
The stoichiometry of the peptide-Zn2+ complexes is evaluated by several criteria. The fluorescence response
of these peptides to pH and various important metal ions is reported. Eleven of these sequences form only
1:1 complexes with Zn2+ and their affinities range from 10 nM to nearly 1 µM. When used in concert, these
sensors can provide Zn2+ concentration information in a valuable range.
introduction of ratiometric ZnAF-R11 and carbonic anhydrase-
based12 probes are also progress toward tools for determining
Introduction
Selective and tunable chemosensors for selected transition
metal ions have the potential to afford qualitative and quantita-
tive information about the presence, distribution and concentra-
tion of these metal ions in cells or tissues. Sensing divalent zinc
is particularly desirable because of its physiological importance.1
Zn2+ is involved in cellular signaling pathways,2 is implicated
in signaling at neural synapses,3 and is co-released with insulin
from glucose-stimulated pancreatic islet â-cells.4 A plethora of
examples exist in the literature of intensity-based Zn2+ fluoro-
phore sensors,5 from the earliest and widely used Zinquin and
TSQ6 to improved intensity probes such as ZPs,7 ZnAFs,8
FluoZin-3,9 and those based on carbonic anhydrase.10 The
Zn2+ concentrations. Optimized sensors are still needed to
quantify nanomolar to micromolar Zn2+ concentrations. Previ-
ously, this laboratory has developed peptide-based Zn2+
sensors13-15 and we now report a new generation of smaller,
more sensitive chemosensors with tuned affinity for divalent
zinc in this range originating from the modular peptide
architecture.
Quantifying free Zn2+ is only possible at Zn2+ concentrations
at or near the dissociation constant of the chemosensor. Many
of the currently available chemosensors are limited in that they
either form 1:2 Zn2+:chemosensor complexes5 and/or that they
have very tight affinities (sub nM).7-9,11 Progress toward tuning
selective Zn2+ chemosensors for monitoring Zn2+ at higher
concentrations has been achieved but not without shortcomings.
The dissociation constant of a dansylamide-polyamine mac-
rocycle Zn2+ sensor has been reduced to nearly 1 µM by altering
the chelator moiety.16 However, the Zn2+ complexes of chemosen-
sors containing aliphatic amine ligands exhibit significant
competition with protons at physiological pH. Gee et al. have
removed carboxylate ligands from Ca2+ chemosensors to obtain
(1) (a) Vallee, B. L.; Falchuk, K. H. Physiol. ReV. 1993, 73, 79-118. (b) Frau´sto
da Silva, J. J. R.; Williams, R. J. P. The Biological Chemistry of the
Elements: The Inorganic Chemistry of Life, 2nd ed.; Oxford University
Press: Oxford, 2001.
(2) Beyersmann, D.; Haase, H. BioMetals 2001, 14, 331-341.
(3) (a) Frederickson, C. J.; Bush, A. I. BioMetals 2001, 14, 353-366. (b)
Frederickson, C. J.; Suh, S. W.; Silva, D.; Frederickson, C. J.; Thompson,
R. B. J. Nutr. 2000, 130, 1471S-1483S. (c) Cuajungco, M. P.; Lees, G. J.
Neurobiol. Dis. 1997, 4, 137-169.
(4) (a) Formby, B.; Schmid-Formby, F.; Grodsky, G. M. Diabetes 1984, 33,
229-234. (b) Grodsky, G. M.; Schmid-Formby, F. Endocrinology 1985,
117, 704-710.
(5) For reviews see: (a) Kimura, E.; Aoki, S. BioMetals 2001, 14, 191-204.
(b) Burdette, S. C.; Lippard, S. J. Coord. Chem. ReV. 2001, 216, 333-
361.
several chemosensors with micromolar affinity for Zn2+ 9b
. These
(6) See references within Canzoniero, L. M. T.; Sensi, S. L.; Choi, D. W.
Neurobiol. Dis. 1997, 4, 275-279 and within ref 5a.
(7) (a) Walkup, G. K.; Burdette, S. C.; Lippard, S. J.; Tsien, R. Y. J. Am.
Chem. Soc. 2000, 122, 5644-5645. (b) Burdette, S. C.; Walkup, G. K.;
Spingler, B.; Tsien, R. Y.; Lippard, S. J. J. Am. Chem. Soc. 2001, 123,
7831-7841. (c) Burdette, S. C.; Frederickson, C. J.; Bu, W. M.; Lippard,
S. J. J. Am. Chem. Soc. 2003, 125, 1778-1787.
(8) (a) Hirano, T.; Kikuchi, K.; Urano, Y.; Higuchi, T.; Nagano, T. J. Am.
Chem. Soc. 2000, 122, 12 399-12 400. (b) Hirano, T.; Kikuchi, K.; Urano,
Y.; Nagano, T. J. Am. Chem. Soc. 2002, 124, 6555-6562. (c) Ueno, S.;
Tsukamoto, M.; Hirano, T.; Kikuchi, K.; Yamada, M. K.; Nishiyama, N.;
Nagano, T.; Matsuki, N.; Ikegaya, Y. J. Cell Biol. 2002, 158, 215-220.
(9) (a) Gee, K. R.; Zhou, Z.-L.; Qian, W.-J.; Kennedy, R. J. Am. Chem. Soc.
2002, 124, 776-778. (b) Gee, K. R.; Zhou, Z.-L.; Ton-That, D.; Sensi, S.
L.; Weiss, J. H. Cell Calcium 2002, 31, 245-251. (c) Qian, W.-J.; Gee,
K. R.; Kennedy, R. T. Anal. Chem. 2003, 75, 3136-3143.
(10) (a) Fierke, C. A.; Thompson, R. B. BioMetals 2001, 14, 205-222. (b)
Thompson, R. B.; Peterson, D.; Mahoney, W.; Cramer, M.; Maliwal, B.
P.; Suh, S. W.; Frederickson, C.; Fierke, C.; Herman, P. J. Neurosci.
Methods 2002, 118, 63-75.
(11) Maruyama, S.; Kikuchi, K.; Hirano, T.; Urano, Y.; Nagano, T. J. Am. Chem.
Soc. 2002, 124, 10 650-10 651.
(12) Thompson, R. B.; Cramer, M. L.; Bozym, R.; Fierke, C. A. J. Biomed.
Optics 2002, 7, 555-560.
(13) (a) Walkup, G. K.; Imperiali, B. J. Am. Chem. Soc. 1996, 118, 3053-
3054. (b) Walkup, G. K.; Imperiali, B. J. Am. Chem. Soc. 1997, 119, 3443-
3450.
(14) Walkup, G. K.; Imperiali, B. J. Org. Chem. 1998, 63, 6727-6731.
(15) Jotterand, N.; Pearce, D. A.; Imperiali, B. J. Org. Chem. 2001, 66, 3224-
3228.
(16) Koike, T.; Abe, T.; Takahashi, M.; Ohtani, K.; Kimura, E.; Shiro, M. J.
Chem. Soc., Dalton Trans., 2002, 1764-1768.
9
10.1021/ja0355980 CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 10591-10597
10591

A R T I C L E S
Shults et al.
PET-chemosensors experience significant fluorescence in the
absence of Zn2+ due to incomplete PET quenching. Carbonic
anhydrase-based Zn2+ sensors have been developed to bind Zn2+
with affinities over several orders of magnitude up to 1 µM.10
Several variants of the carbonic anhydrase-based sensors have
been developed for quantifying Zn2+ via ratiometric, anisotropic
or lifetime measurements on a single sensor; however, the
sensors are large and suffer from the necessity of ternary
complex formation.10 Recently, a “reagentless” carbonic anhy-
drase-based sensor has been reported, which avoids ternary
complex formation.17 The goal of the research presented herein
was to develop an addressable family of low molecular weight
chemosensors with a range of Zn2+ affinities, that can be used
to determine Zn2+ concentration via analysis of the relative
intensities of the sensor readout.
A metal ion chemosensor built from a modular peptide
scaffold offers many advantages. For example, the modular
architecture enables implementation of any one of a number of
fluorescence-based sensing mechanisms: chelation-enhanced
fluorescence,14,15 environment-sensitive fluorescence,13 fluores-
cence quenching,18 fluorescence resonance energy transfer,19 and
in some cases, metal-based fluorescence.20 Additionally, both
natural and nonnatural metal-binding amino acids can be
integrated via combinatorial or design approaches, made facile
via solid-phase peptide synthesis (SPPS). Binding optimization
can be achieved by altering the type of ligands as well as their
orientation and location within the peptide scaffold. Incorpora-
tion of other functionalities is also feasible via amide bond
formation. For example, an additional fluorophore could be
incorporated for ratiometric sensing, a cellular internalization
sequence21 could be appended to transport the sensors into cells,
or the sensors could be attached to a solid support for diagnostic
devices.
Figure 1. Schematic representation of modular Zn2+-binding peptides. To
tune the KD for Zn2+, the â-turn sequence, additional Zn2+ ligand, and the
ligand arrangement are altered.
phore into the novel amino acid, Sox, and the design, synthesis,
and evaluation of a family of chemosensors for Zn2+ is reported.
The basic design of the chemosensor peptides includes a Sox
residue and a residue that affords an additional ligand for Zn2+
flanking a â-turn sequence, which pre-organizes the metal-ion
binding site.23 The necessity of a turn sequence for metal
binding23b as well as its ability to tune affinity for the metal
ion23a have been demonstrated previously. The basic features
of the design are represented in Figure 1. To tune the affinity
for Zn2+, the additional ligand(s) for Zn2+ and the â-turn
sequence in the peptide are varied. Each peptide contains only
six to eight amino acids. Two of these are serine residues,
appended to the carboxy-terminus, to increase peptide solubility
without interfering with metal binding. Further variation may
be achieved through incorporation of an N-terminal module as
a functionalized capping group instead of an amino acid
derivative. Additional modules, of the types suggested earlier,
may be incorporated at either terminus. Because the chemo-
sensors reported herein are sufficiently polar not to be cell-
permeable, they may be useful for detection of extracellular Zn2+
or as a diagnostic tool for medical or environmental samples.
Results and Discussion
Synthesis of Fmoc-Sox-OH. The synthesis of the protected
amino acid is outlined in Scheme 1. The synthesis of 1 from
8-hydroxyquinaldine has been previously reported,22 but the
yield of the sulfonamide was improved by diluting the reaction
mixture significantly. The phenolic hydroxyl group of 1 was
then protected as the tert-butyldiphenylsilyl ether. The next three
steps were performed without intermediate purification and
include oxidation to the aldehyde with selenium dioxide,
reduction to the alcohol with sodium borohydride, and bromi-
nation. Both the intermediate aldehyde and alcohol were unstable
on silica gel. Purification of bromide 3 was accomplished via
chromatography on florisil. This purified bromide contains 15%
of 1 (by NMR) brought through from the selenium dioxide
oxidation step. This impurity does not affect future reactions
and can be washed away during peptide synthesis. Synthesis
of the amino acid from 3 was performed by Corey’s adaptation24
of O’Donnell’s asymmetric alkylation method,25 utilizing (8S,
9R)-O-(9)-allyl-N-9-anthracenyl methylcinchonidium bromide
as the phase-transfer catalyst. Because the benzophenone imine
We have chosen to use chelation-enhanced fluorescence
(CHEF) for sensing Zn2+. In this mechanism, metal ion binding
is accompanied by the generation of a new fluorescence
emission signal. The fluorophore chosen for this purpose need
not form a completely selective 1:1 complex with Zn2+ because
additional specificity determinants can be incorporated into the
peptide sequence. Previously, we have reported that the quino-
line derivative, 8-hydroxy-5-(N,N-dimethylsulfonamido)-2-me-
thylquinoline, binds Zn2+ and the complex has a much improved
quantum yield over the parent fluorophore, 8-hydroxy-2-
methylquinoline.22 Integration of the enhanced CHEF fluoro-
(17) Thompson, R. B.; Maliwal, B. P.; Feliccia, V. L.; Fierke, C. A.; McCall,
K. Anal. Chem. 1998, 70, 4717-4723.
(18) (a) Torrado, A.; Walkup, G. K.; Imperiali, B. J. Am. Chem. Soc. 1998,
120, 609-610. (b) Torrado, A.; Imperiali, B. J. Org. Chem. 1996, 61,
8940-8948.
(19) (a) Pearce, D. A.; Walkup, G. K.; Imperiali, B. Bioorg. Med. Chem. Lett.
1998, 8, 1963-1968. (b) Godwin, H. A.; Berg, J. M. J. Am. Chem. Soc.
1996, 118, 6514-6515.
(23) (a) Cheng, R. P.; Fisher, S. L.; Imperiali, B. J. Am. Chem. Soc. 1996, 118,
11 349-11 356. (b) Imperiali, B.; Kapoor, T. M. Tetrahedron 1993, 49,
3501-3510.
(20) Franz, K.; Nitz, M.; Imperiali, B. ChemBioChem 2003, 4, 265-271.
(21) (a) Wadia, J. S.; Dowdy, S. F. Curr. Opin. Biotech. 2002, 13, 52-56. (b)
Lindgren, M.; Ha¨llbrink, M.; Prochiantz, A.; Langel, U¨ . Trends Pharmacol.
Sci. 2000, 21, 99-103.
(24) (a) Corey, E. J.; Xu, F.; Noe, M. C. J. Am. Chem. Soc. 1997, 119, 12 414-
12 415. (b) Corey, E. J.; Noe, M. C.; Xu, F. Tetrahedron Lett. 1998, 39,
5347-5350.
(22) Pearce, D. A.; Jotterand, N.; Carrico, I. S.; Imperiali, B. J. Am. Chem.
Soc. 2001, 123, 5160-5161.
(25) O’Donnell, M. J. Aldrichimica Acta 2001, 34, 3-13.
9
10592 J. AM. CHEM. SOC. VOL. 125, NO. 35, 2003

Chemosensor Scaffold for Divalent Zinc
A R T I C L E S
Scheme 1
Table 1. Dissociation Constants for Zn2+ Peptidyl Chemosensors
and tert-butyldiphenylsilyl ether are relatively acid labile, the
adduct from the alkylation reaction was hydrolyzed in refluxing
6 N hydrochloric acid immediately upon completion of the
alkylation. The product, Sox-OH, was isolated as the hydro-
chloride salt and used in the subsequent step without purification.
A pure sample of this amino acid was derivatized with Marfey’s
reagent26 to determine that the reaction gave 96% e.e. in favor
of the L-enantiomer.27 For subsequent peptide synthesis, the free
amine was protected with the 9-fluorenyl-methoxycarbonyl
(Fmoc) group by treatment with (9-fluorenylcarbonyl)succin-
imidyl carbonate (FmocOSu) to give Fmoc-Sox-OH (70% yield
over 3 steps by NMR).
Peptide Synthesis. Fmoc-Sox-OH is used without purification
in standard Fmoc SPPS. The small amount of peptide containing
D-Sox was separated from the desired peptide during HPLC
purification. The identity of each of the peptides was confirmed
by mass spectrometry and purity was verified by RP-HPLC.
Nineteen peptide sequences were designed and characterized
(Table 1). The peptides differ in number and type of Zn2+
-
binding ligands, ligand arrangement or â-turn sequence.
Fluorescence Properties. In the presence of saturating ZnCl2,
the Sox peptides exhibit a maximum emission at 500 nm with
a maximum excitation at 360 nm. In addition, these peptides
maintain the striking luminescence properties of the fluoro-
phore.22 The extinction coefficient and quantum yield values
for representative chemosensor-Zn2+ complexes were deter-
mined following quantitative amino acid analysis (QAA). These
values are within 10% of 6200 M-1 cm-1 and 0.16, respec-
tively.28 In the absence of Zn2+, the quantum yields are less
than 0.005, indicating a 30-fold increase in fluorescence upon
a Obtained via direct fluorescence titration. Values reported with standard
deviation are obtained from triplicate titrations with a peptide concentration
near the KD. Values without standard deviations are obtained via a single
titration at [peptide] ) 50 nM. KD values are calculated by a global nonlinear
least-squares fit with the program Specfit/32.29 b Pen ) â-dimethylcysteine.
c Ipa ) 3-(imidazol-4-yl)propionic acid. d pSer ) phosphoserine. e âPhe )
3-amino-3-phenylpropionic acid. P9 and P18 each contain one of the
enantiomers of DL-âPhe.
binding to Zn2+
.
Summary of Binding Studies. Peptide binding characteristics
were assessed via titrations with ZnCl2 and Job’s plot analysis
to determine the affinity and stoichiometry of complex forma-
tion. Titration data is fit with the program Specfit/32,29 which
analyzes multiwavelength data using an iterative method to
obtain the association constant in terms of the free (unbound)
Zn2+ concentration, [Zn2+]f. Direct titrations were performed
for all peptide sequences under similar conditions for compari-
son purposes. Buffered Zn2+ solutions have been used for other
tighter-binding Zn2+ chemosensors and require only certain
buffer and ionic strength conditions where the binding constant
of a Zn2+ chelator has been previously determined.7-9,11 Only
(26) Fujii, K.; Ikai, Y.; Mayumi, T.; Oka, H.; Suzuki, M.; Harada, K. Anal.
Chem. 1997, 69, 3346-3352.
(27) The preference for L-enantiomer formation with the catalyst O-(9)-allyl-
N-9-anthracenyl methylcinchonidium bromide is well-established over a
wide range of electrophiles. See refs 24 and 25.
(28) Some variation between peptide sequences in the extinction coefficients
was observed due to the error associated with the stock solution concentra-
tion determined by QAA. The error in this measurement is at least 10%.
The values are detailed in the Supporting Information.
(29) SPECFIT/32 for Windows (version 3.0), Spectrum Software Associates,
Marlborough, MA.
9
J. AM. CHEM. SOC. VOL. 125, NO. 35, 2003 10593

A R T I C L E S
Shults et al.
Figure 2. Job plots for (a) P13 (forms only 1:1 complexes) and (b) P8 (forms both 1:1 and 1:2 complexes). The total [peptide] + [ZnCl2] ) 0.98 µM.
Figure 3. Scatchard plots for (a) P13 and (b) P8. Titrations were performed with 500 nM peptide. Data for entire titration is included.
Figure 4. Fit obtained from titration data for (a) P13 using a 1:1 complexation model and (b) P8 using a 1:1 and 1:2 complexation model. With these
models specified, a global nonlinear least-squares fit of the data was obtained via Specfit/32.29
the 1:1 complex dissociation constants are listed in Table 1 and
discussed in the following section because these depend the most
heavily on the â-turn sequence and the additional Zn2+ ligand.
Binding Stoichiometry. For the peptides in Table 1, four
criteria were evaluated to determine binding stoichiometry: a
Job’s plot analysis, a Scatchard plot, the model used for KD
calculation and the maximum fluorescence emission wavelength.
To illustrate the conclusion from each method, data for peptides
P8 and P13 are included. These data are representative of all
peptide sequences synthesized. A Job’s plot30 that is symmetrical
and exhibits a maximum at 0.5 mole fraction of Zn2+ indicates
that only a 1:1 complex is formed (Figure 2a). However, a Job’s
plot with a maximum at any other value besides 0.5 indicates
more complex equilibria (Figure 2b).30 A Scatchard plot of
extracted titration data at a single wavelength is linear only for
peptides forming solely a 1:1 complex (Figure 3).30 Peptides
forming a 1:1 complex exclusively throughout the course of a
titration obtain a better fit with a 1:1 complexation model (Figure
4). Addition of a 1:2 Zn2+:peptide complex into the model fits
the data for all other peptides. Additionally, the wavelength of
maximum emission for a peptide forming only a 1:1 complex
remains constant at 500 nm during the course of a titration
(Figure 5a). On the other hand, peptides forming 1:1 and 1:2
complexes exhibit shifts in the maximum emission wavelength
from 475 to 500 nm as increasing amounts of Zn2+ are added
(Figure 5b). This shift is seen because the 1:2 complex is also
fluorescent.
For some sequences, an increase of peptide concentration
increases the likelihood for 1:2 complex formation because any
two given peptides are in closer proximity. The rigid imidazole
(30) Connors, K. A. Binding Constants: The Measurement of Molecular
Complex Stability; John Wiley and Sons: New York, 1987.
9
10594 J. AM. CHEM. SOC. VOL. 125, NO. 35, 2003

Chemosensor Scaffold for Divalent Zinc
A R T I C L E S
Figure 5. Spectra recorded for the titration of (a) P13 (emission maximum remains constant at 500 nm) and (b) P8 (emission maximum shifts during the
titration). Only a few of the 25 spectra collected are shown for clarity.
Table 2. Dissociation Constants for 1:1 Zn2+:Peptide Complexes
Zn2+ ligands
(N C)
Apparent
KDa (nM)
f
Turn
P1
P2
P3
P4
P5
P6
P7
P10
P11
P12
P13
Sox, Pen
Ipa, Cys, Sox
Sox, Cys
Cys, Sox
Sox, Cys
Ipa, His, Sox
Sox, pSer
Ipa, Sox
Pro-DSer
DPro-Gly
Pro-Gly
9.6 ( 2.6
11 ( 2
12 ( 1
Pro-DSer
Pro-DSer
DPro-Ser
Pro-DSer
DPro-Gly
DPro-Ser
Pro-DSer
Pro-DSer
15 ( 4
19 ( 5
35 ( 5
130 ( 10
360 ( 20
530 ( 30
730 ( 80
910 ( 50
Ipa, Sox
Sox, Glu
Sox, Asp
a Obtained via direct fluorescence titration in triplicate. KD values are
calculated by a global nonlinear least-squares fit with the program Specfit/
32.
Figure 6. pH dependence of fluorescence intensity for P10 (open circles)
and the P10-Zn2+ complex (closed circles).
ligand was more sensitive to modification of the turn sequence
and the ligand arrangement; not all peptides synthesized with
this ligand formed only 1:1 complexes. With the exception of
P8, P15, P18, and P19, all peptides form only the 1:1 complex
at peptide concentrations of 50 nM. Many peptide sequences
form more favorable 1:1 complexes, indicated by their prefer-
ence for 1:1 complex formation at higher peptide concentrations.
Of the nineteen peptides synthesized, eleven form 1:1
complexes in a useful range for biological experiments (low
micromolar). Only these peptides (Table 2) will be further
discussed, as they are the most useful. For peptides that form
mixed complexes, determination of Zn2+ concentrations is
complicated by the fact that relative concentrations of the 1:1
and 1:2 complexes change depending on peptide concentration
and Zn2+ concentration.
Binding Trends. Inspection of the data presented in Table 2
allows the conclusion that the nonfluorophore Zn2+ ligands
provide rough tuning of binding affinity. A peptide’s affinity
for Zn2+ decreases along the following series: thiolate > two
imidazoles > phosphate > one imidazole > carboxylate. This
trend is expected based on the binding preferences of the Zn2+
ion, where “softer” ligands are preferred.1,31 The affinity of these
peptides for Zn2+ spans 2 orders of magnitude through ligand
variation. Fine-tuning of binding affinity was possible through
modification of the turn sequence and ligand arrangement. In
general, a more flexible Gly in the turn sequence leads to a
slightly tighter-binding peptide. In addition, incorporation of
two imidazole ligands greatly increased the affinity over one
imidazole ligand, but the addition of an imidazole ligand to a
cysteine-containing peptide barely changed the affinity.
Implications for Further Tuning. The possibility of ad-
ditional Zn2+ chemosensors with altered affinities is feasible
through continued modification of this architecture considering
the following trends. In general, affinity can be increased with
softer ligands, a more flexible turn, and additional ligands.
Affinity can be weakened with harder ligands and increased
turn rigidity. Synthesis and characterization of the new chemo-
sensor must be the final and necessary step to confirm the utility
of each new design.
Competition Studies. To evaluate the potential utility of these
sensors, the influence of protons and metal ions on the ability
of the chemosensor to sense Zn2+ was investigated. Two
separate experiments were performed to assess possible interfer-
ence on the detection of Zn2+: one to assess false positive
responses where the competing entity forms a fluorescence
complex with the chemosensor and another to identify false
negative responses where the ability of the sensor to bind Zn2+
is compromised.
The fluorescence of each peptide in the absence of Zn2+
remains essentially unchanged between pH 4 and 10, whereas
the fluorescence of each peptide-Zn2+ complex is quenched
below pH 5. All of the peptide-Zn2+ complexes are maximally
fluorescent at and above pH 7. Figure 6 shows the fluorescence
intensity of P10 and the P10-Zn2+ complex as a function of
(31) The phosphate ligand appears out of order if solely a “soft” ligand
explanation is invoked to explain increased Zn2+ affinity. The increased
affinity of P7 for Zn2+ relative to the “softer” imidazole ligand may be
due to its increased negative charge and/or its ability to bind as a bidentate
ligand.
9
J. AM. CHEM. SOC. VOL. 125, NO. 35, 2003 10595

A R T I C L E S
Shults et al.
Figure 7. Metal ion competition plots for representative peptide sequences, based on donor ligands. (a) Fluorescence intensity of peptides in the presence
of various cations. (b) Fluorescence intensity of peptide plus 1 equivalent of Zn2+ in the presence of various cations. All metal ions were evaluated at one
equivalent except Ca2+, Mg2+ and K+ which were used at 1 mM. Peptide concentrations were either 1 µM (P2, P5, P6 and P7) or 10 µM (P11, P12).
pH. This plot is representative of all peptide sequences; the pH
profiles differ only very slightly due to the ligand variation.
In general, the metal ion response variation between peptides
can be explained as due to donor-atom preference for each metal
ion.
The fluorescence response of each sensor in the presence of
each metal cation was investigated (Figure 7a), since metal ions
besides Zn2+ are known to form fluorescent complexes with
8-hydroxyquinoline, including Mg2+ and Ca2+ 32
. Only peptides
and most likely, similar regulatory proteins for other metal ions
exist.33,34 For diagnostic purposes, a sensor can be chosen based
on its lack of response to a competing metal ion in addition to
its dissociation constant.
In addition, these chemosensors all form fluorescent com-
plexes with Cd2+. The concentration of Cd2+ in biological
systems is typically very low1b and thus is unlikely to interfere
with measurements in such systems. However, environmental
and diagnostic applications exist for a Cd2+ chemosensor due
to the toxicity of this metal ion. Cd2+ forms complexes that are
as fluorescent as the ones with Zn2+ for all peptides, but does
not bind as effectively to peptides P6 and P7 as Zn2+. Sensors
are rarely able to differentiate between Zn2+ and Cd2+. It is
likely that Cd2+ is too large to efficiently bind to all four ligands
of P6. In addition, Cd2+ is considered an even “softer” metal
ion than Zn2+ and has less affinity for phosphate ligands than
Zn2+ in P7. These trends imply that the peptide scaffold could
be tuned for the specific detection of Cd2+ as well.
Probes of Zn2+ Concentration. For a given application,
peptides spanning the range of affinities may be selected for an
experiment and the Zn2+ concentration can be determined within
a defined range where one sensor gives a fluorescent response
when another does not. To illustrate this, vials with different
probes and varying amounts of Zn2+ were prepared (Figure 8).
The fluorescence of each peptide at each of the three concentra-
tions corresponds well with the data obtained from titrations.
Zn2+ concentrations can be narrowed into a range based on the
responses of two or more peptides. This measurement does not
require that the maximal fluorescence intensity be known, but
only that one peptide does not give a response. This method
avoids many of the well-known difficulties associated with
intensity-based measurements because the chemosensors are
sufficiently similar to be compared under identical experimental
conditions. For example, a solution that fluoresces in the
presence of P5 but does not fluoresce in the presence of P7
would contain a Zn2+ concentration between roughly 50 and
200 nM. Only very simple illumination equipment is required.
containing “hard” carboxylate and phosphate donor ligands (P7,
P12, P13) give a response with 1 mM Mg2+ and Ca2+. The
response to Mg2+ is greater likely because the size of Mg2+ is
more comparable to Zn2+ than that of Ca2+ 1b
P7 gives a
.
substantial response to Mg2+ (50%) because Mg2+ has a
particularly high affinity for phosphate, as seen by its frequent
coordination to phosphate ligands in ATP and oligonucleotide
sequences.1b Further details concerning Mg2+ and Ca2+ binding
to P7, P12, and P13 can be found in the Supporting Information.
Because many other metal ions do not form fluorescent
complexes with 8-hydroxyquinoline, the fluorescence response
of the peptide-Zn2+ complex in the presence of 1 equivalent of
several additional metal ions was investigated (Figure 7b). K+
does not interfere with fluorescence for any peptide-Zn2+
complex up to 1 mM. For weaker Zn2+-binding peptides, Mn2+
and Co2+ compete with Zn2+ binding by no more than 10%
and 20%, respectively. Fe3+ interferes for peptides with “hard”
oxygen donor atoms (P7, P12, P13). Ni2+ causes significant
quenching for peptides with imidazole and carboxylate ligands
but has little affinity for peptides with thiolate and phosphate
ligands. Cu2+ interferes with Zn2+ binding for all peptides, and
it entirely competes with Zn2+ for peptides containing thiolate
ligands (P1, P2, P5).
With the exception of Ca2+, Mg2+, and K+, all metal ions
investigated are unlikely to be present as the free metal ion in
biological systems. Regulatory proteins which bind Cu2+ and
Zn2+ ions sufficiently tightly to preclude the possibility of free
amounts of these metals in the cytoplasm have been discovered
(33) Outten, C. E.; O’Halloran, T. V. Science 2001, 292, 2488-2492.
(34) Rae, T. D.; Schmidt, P. J.; Pufahl, R. A.; Culotta, V. C.; O’Halloran, T. V.
Science 1999, 284, 805-808.
(32) Seitz, W. R. CRC Critical ReViews in Analytical Chemistry; CRC Press:
Boca Raton, FL, 1980; pp 367-404.
9
10596 J. AM. CHEM. SOC. VOL. 125, NO. 35, 2003

Chemosensor Scaffold for Divalent Zinc
A R T I C L E S
from the binding constants reported have implications for further
tuning to generate new sensors. Competition analysis with
protons and important metal ions show these probes to be
sufficiently selective for Zn2+ under conditions for biological
and diagnostic experiments. Chemosensors from this family can
be chosen for applications based on their Zn2+-affinity as well
as their metal ion preferences. The sensors can be used easily
to determine Zn2+ concentration when one sensor gives a
positive response and another does not. For Zn2+ concentration
determination with a single sensor, investigation of ratiometric
Zn2+ sensors based on this scaffold are in progress.
Acknowledgment. Research grants from the NSF (CHE-
9996335) and an NSF graduate research fellowship for M.D.S.
are gratefully acknowledged. The NMR spectrometers at MIT
were provided by grants from the NSF (DBI-9729592 and
CHE-9808061) and NIH (1S10RR13886-01). M.D.S. would like
to thank Dr. N. Jotterand for early procedures toward the
synthesis of the amino acid, Dr. K. J. Franz for helpful
discussions and Dr. R. A. Binstead of Spectrum Software
Associates for assistance with the Specfit/32 program.
Figure 8. (a) Visual representation of the range of Zn2+ affinities of five
of the peptides. Vials were loaded with a 1 µM solution of the appropriate
peptide in 50 mM HEPES buffer at pH 7 containing 150 mM NaCl and the
indicated concentration of ZnCl2. Vials were irradiated at 365 nm on a UV
transilluminator. (b) Comparison of the fluorescence responses in (a) with
the titration curves shows the probes behave as expected.
Conclusion
In conclusion, a nonnatural amino acid Sox has been
synthesized and used as a modular building block in peptide
sequences that bind Zn2+ in the nanomolar to micromolar range,
where few useful Zn2+ chemosensors currently exist. The
affinity for Zn2+ was tuned by variation of the ligands, scaffold,
and turn sequence of the peptide. Eleven peptide sequences that
bind to Zn2+ in a 1:1 complex with affinities spanning 10 nM
to 1 µM are summarized in Table 2. The general trends gleaned
Supporting Information Available: Synthetic procedures and
characterization for Sox and peptides, 1H NMR spectra of 2, 3,
Sox, and Fmoc-Sox-OH, experimental details for all fluores-
cence experiments (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA0355980
9
J. AM. CHEM. SOC. VOL. 125, NO. 35, 2003 10597