Fluorescent sensors for Zn2+ based on a fluorescein platform: Synthesis, properties and intracellular distribution

Article abstract of DOI:10.1021/ja010059l

Two new fluorescent sensors for Zn²⁺ that utilize fluorescein as a reporting group, Zinpyr-1 and Zinpyr-2, have been synthesized and characterized. Zinpyr-1 is prepared in one step via a Mannich reaction, and Zinpyr-2 is obtained in a multistep synthesis that utilizes 4′,5′-fluorescein dicarboxaldehyde as a key intermediate. Both Zinpyr sensors have excitation and emission wavelengths in the visible range (～500 nm), dissociation constants (K_d1) for Zn²⁺ of <1 nM, quantum yields approaching unity (Φ = ～0.9), and cell permeability, making them well-suited for intracellular applications. A 3- to 5-fold fluorescent enhancement is observed under simulated physiological conditions corresponding to the binding of the Zn²⁺ cation to the sensor, which inhibits a photoinduced electron transfer (PET) quenching pathway. The X-ray crystal structure of a 2:1 Zn²⁺:Zinpyr-1 complex has also been solved, and is the first structurally characterized example of a complex of fluorescein substituted with metal binding ligands.

Full text of DOI:10.1021/ja010059l

J. Am. Chem. Soc. 2001, 123, 7831-7841
7831
Fluorescent Sensors for Zn2+ Based on a Fluorescein Platform:
Synthesis, Properties and Intracellular Distribution
Shawn C. Burdette,§ Grant K. Walkup,† Bernhard Spingler,§ Roger Y. Tsien,†,‡ and
Stephen J. Lippard*,§
Contribution from the Department of Chemistry, Massachusetts Institute of Technology,
Cambridge, Massachusetts, 02139, and the Department of Pharmacology and Howard Hughes Medical
Institute, UniVersity of California at San Diego, La Jolla, California 92093-0647
ReceiVed January 5, 2001
Abstract: Two new fluorescent sensors for Zn2+ that utilize fluorescein as a reporting group, Zinpyr-1 and
Zinpyr-2, have been synthesized and characterized. Zinpyr-1 is prepared in one step via a Mannich reaction,
and Zinpyr-2 is obtained in a multistep synthesis that utilizes 4′,5′-fluorescein dicarboxaldehyde as a key
intermediate. Both Zinpyr sensors have excitation and emission wavelengths in the visible range (∼500 nm),
dissociation constants (Kd1) for Zn2+ of <1 nM, quantum yields approaching unity (Φ ) ∼0.9), and cell
permeability, making them well-suited for intracellular applications. A 3- to 5-fold fluorescent enhancement
is observed under simulated physiological conditions corresponding to the binding of the Zn2+ cation to the
sensor, which inhibits a photoinduced electron transfer (PET) quenching pathway. The X-ray crystal structure
of a 2:1 Zn2+:Zinpyr-1 complex has also been solved, and is the first structurally characterized example of a
complex of fluorescein substituted with metal binding ligands.
Introduction
neins (MTs),14-16 including MT-III and MT-IV which are
expressed mainly in the brain.17-20 ZnTs and MTs are probably
responsible for distributing the required amounts of Zn2+ to
proteins and enzymes, minimizing the amounts of free and
potentially toxic levels of Zn2+ present in cells. In addition to
protein regulators, Zn2+ can be released from synaptic vesicles6,21
and can enter cells through voltage-dependent Ca2+ channels,22
indicating that free Zn2+ is still available for neurological
functions. Despite much research, many aspects of ionic Zn2+
in neurobiology remain unclear because of the limited detection
methods currently available to the neuroscience community.
Fluorescent sensors are indispensable tools for visualizing
ions at the molecular level without the need for special
instrumentation.23 Practical fluorescent sensors must produce a
perceptible luminescent signal upon the selective binding of the
desired analyte. Careful engineering of the molecular properties
Zinc is a vital component in many cellular processes.1
Although traditionally the study of Zn2+ bioinorganic chemistry
has focused on its structural role and catalytic functions in
proteins,2 the neurobiology of Zn2+ has been a subject of
increasing attention.3,4 Whereas most Zn2+ in biological systems
is tightly bound in proteins, a pool of free Zn2+ has been imaged
in cells. Included are subnanomolar concentrations in undif-
ferentiated mammalian cells5 and higher concentrations, ap-
proaching 300 µM, in the mossy fiber terminals of the
hippocampus.6,7 The Zn2+ ion has the ability to modulate a
variety of ion channels,8 may play a role in neuronal death
during seizures,9 is pertinent to neurodegenerative disorders,10
and may be vital to neurotransmission.1,10 Because of these
diverse functions, Zn2+ continues to be an interesting subject
of research in neurobiology.
The levels of Zn2+ in the brain and other parts of the body
are regulated by at least three homologous Zn2+ transport
proteins (ZnT-1, ZnT-2, and ZnT-3)11-13 and by metallothio-
(11) Palmiter, R. D.; Findley, S. D. EMBO J. 1995, 14, 639-649.
(12) Palmiter, R. D.; Cole, T. B.; Findley, S. D. EMBO J. 1996, 15,
1784-1791.
(13) Palmiter, R. D.; Cole, T. B.; Quaife, C. J.; Findley, S. D. Proc.
Natl. Acad. Sci. U.S.A. 1996, 93, 14934-14939.
§ Massachusetts Institute of Technology.
(14) Ebadi, M. Methods Enzymol. 1991, 205, 363-387.
(15) Ebadi, M.; Iverson, P. L.; Hao, R.; Cerutis, D. R.; Rojas, P.; Happe,
H. K.; Murrin, L. C.; Pfeiffer, R. F. Neurochem. Int. 1995, 27, 1-22.
(16) Ebadi, M.; Perini, F.; Mountjoy, K.; Garvey, J. S. J. Neurochem.
1996, 66, 2121-2127.
† Department of Pharmacology, University of California at San Diego.
‡ Howard Hughes Medical Institute, University of California at San
Diego.
(1) Vallee, B. L.; Falchuk, K. H. Physiol. ReV. 1993, 73, 79-118.
(2) Lippard, S. J.; Berg, J. M. Principles of Bioinorganic Chemistry;
University Science Books: Mill Valley, 1994.
(3) Frederickson, C. J. Int. ReV. Neurobiol. 1989, 31, 145-238.
(4) Huang, E. P. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 13386-13387.
(5) Nasir, M. S.; Fahrni, C. J.; Suhy, D. A.; Kolodsick, K. J.; Singer, C.
P.; O’Halloran, T. V. J. Biol. Inorg. Chem. 1999, 4, 775-783.
(6) Frederickson, C. J.; Moncrieff, D. W. Biol. Signals 1994, 3, 127-
139.
(17) Palmiter, R. D.; Findley, S. D.; Whitmore, T. E.; Durnam, D. M.
Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 6333-6337.
(18) Pountney, D. L.; Fundel, S. M.; Faller, P.; Birchler, N. E.; Hunziker,
P.; Vasak, M. FEBS Lett. 1994, 345, 193-197.
(19) Tsuji, S.; Kobayashi, H.; Uchida, Y.; Ihara, Y.; Miyatake, T. EMBO
J. 1992, 11, 4843-4850.
(20) Uchida, Y.; Takio, K.; Titani, K.; Ihara, Y.; Tomonaga, M. Neuron
1991, 7, 337-347.
(7) Budde, T.; Minta, A.; White, J. A.; Kay, A. R. Neuroscience 1997,
79, 347-358.
(21) Slomianka, L. Neuroscience 1992, 48, 325-352.
(22) Atar, D.; Backx, P. H.; Appel, M. M.; Gao, W. D.; Marban, E. J.
Biol. Chem. 1995, 270, 2473-2477.
(8) Harrison, N. L.; Gibbons, S. J. Neuropharmacology 1994, 33, 935-
952.
(23) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley,
A. J.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. ReV. 1997, 97,
1515-1566.
(9) Choi, D. W.; Koh, J. Y. Annu. ReV. Neurosci. 1998, 21, 347-375.
(10) Cuajungco, M. P.; Lees, G. J. Neurobiol. Dis. 1997, 4, 137-169.
10.1021/ja010059l CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/24/2001

7832 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
Burdette et al.
of these probes can provide an effective methodology for
studying relevant intracellular metal ions, such as Ca2+ in
biological systems.24 In addition to selectivity for the analyte,
biosensors should have a Kd value near the median concentration
of the species under investigation to allow concentration changes
to be monitored. Desirable intracellular fluorescent sensors
should produce a signal with a high quantum yield (Φ) and
have excitation wavelengths exceeding 340 nm. The latter
property permits use with glass microscope objectives and
prevents UV-induced cell damage. A good sensor will also have
emission wavelengths approaching 500 nm to avoid autofluo-
rescence from species native to the cell and to facilitate use
with typical fluorescence microscopy optical filter sets.25 Finally,
sensors must have the ability to be passively and irreversibly
loaded into cells.
and preliminary characterization of Zinpyr-1,39 a new cell-
permeable fluorescent sensor for Zn2+ 40
. Our initial investiga-
tions indicated that Zinpyr-1 exhibits a 3.1-fold increase in
integrated fluorescence intensity upon the formation of a 1:1
ligand:Zn2+ complex. Zinpyr-1 can be passively loaded into cells
and used to detect Zn2+ introduced exogenously. Zinpyr-1 is
particularly amenable to intracellular work, because fluorescein
is the reporting group. This xanthenone-based chromophore has
a quantum yield approaching unity (Φ ) 0.95) and excitation
and emission wavelengths exceeding 490 nm.41
Previous approaches to preparing fluorescent sensors have
often suffered from a lack of generality. We have therefore
devised a general methodology for preparing fluorescent sensors
that is based on the fluorescein framework and demonstrate here
its utility through the synthesis of Zinpyr-2, a first-generation
fluorescent probe structurally similar to Zinpyr-1. In addition,
we describe a thorough investigation of the physical and
structural properties of these probes related to their use as Zn2+
sensors.
The majority of studies of intracellular Zn2+ with fluorescent
sensors have been performed with TSQ,26 Zinquin,27,28 and
TFLZn,7 all of which are aryl sulfonamide derivatives of
8-aminoquinoline. Recent investigations into the aqueous bind-
ing properties of these compounds have clarified many discrep-
ancies found in the literature on these probes, promoting the
accuracy of results for future application of these sensors.5,29
Although quinoline-based probes are useful, these sensors are
not ideal, because they require near-ultraviolet excitation and
Experimental Section
Materials and Methods. Chlorobenzene and 1,2-dichloroethane
(DCE) were distilled from calcium hydride under nitrogen. Dimethyl
sulfoxide (DMSO) was vacuum distilled from CaH2 and subsequently
dried over 3-Å molecular sieves. Deuterated chloroform was dried over
3-Å molecular sieves. Zinc chloride was fused prior to each use. Di-
(2-picolyl)amine (DPA) was prepared as previously described.42 All
other reagents were purchased and used as received. Flash column
chromatography was performed using silica gel-60 (230-400 mesh)
or Brockman I activated neutral aluminum oxide (150 mesh). Thin-
layer chromatography (TLC) analysis was performed using Merck F254
silica gel-60 plates or Merck F254 aluminum oxide-60 and viewed by
UV light, or developed with ceric ammonium molybdate, 2,4-
dinitrophenyl hydrazine, or iodine stain. Infrared spectra were recorded
on a BTS 135 FTIR instrument as KBr pellets. NMR spectra were
recorded on a Varian 500 MHz spectrometer at ambient probe
can form mixed complexes sensing partially coordinated Zn2+
.
Several approaches to fluorescent Zn2+ detection based on
polypeptide,30-32 protein,33 and macrocyclic34-36 sensors have
been reported, but none has been applied successfully to sensing
zinc in living cells. These alternative sensors are not suitable
for intracellular work because of either unsatisfactory binding
affinity, optical properties, and cell toxicity, or the need to
perform microinjections into cells. Several traditional Ca2+ and
Mg2+ probes have been utilized to study Zn2+ 22,37
from these applications can be ambiguous, because Zn2+
,
but results
-
induced signals are difficult to distinguish from those involving
alkaline earth metal ions. Recently, the synthesis and charac-
terization of ZnAF-1 and ZnAF-2, two aminofluorescein-based
Zn2+ sensors, were reported.38 The esterified forms of these new
sensors are cell-permeable, but details about their intracellular
behavior were not described.
1
temperature, 283 K, and were referenced to the internal H and 13C
solvent peaks. Fast atom bombardment (FAB) mass spectrometry was
performed in the MIT Department of Chemistry Instrumentation Facility
(DCIF) with the use of m-nitrobenzyl alcohol as the matrix. Melting
points were recorded on a Thomas-Hoover capillary melting point
apparatus. CAUTION: One of the isolated compounds below
contains perchlorate ion, which can detonate explosively and
without warning. Although we have encountered no incidents with
the reported compound, all due precautions should be taken.
4′,5′-Dimethylfluorescein Dibenzoate (1). Phthalic anhydride (16.7
g, 113 mmol) and 2-methylresorcinol (24.9 g, 201 mmol) were crushed
and melted into a brown liquid at 150 °C. Fused ZnCl2 (15 g, 110
mmol) was added slowly over 35 min, and the temperature was slowly
increased to 230 °C over 30 min until the material solidified. The brick
red solid was pulverized and boiled in 250 mL of 6 M HCl for 30 min.
The red solid was collected on a frit, washed thoroughly with distilled
water, and dried in vacuo at 50 °C for 2 h. The crude product was
combined with benzoic anhydride (115 g, 509 mmol) in 400 mL of
pyridine and refluxed at 140 °C for 2.5 h. The reaction mixture was
diluted with 700 mL of distilled water, and a dark brown solid formed
upon cooling. The solids were collected, washed thoroughly with water,
and dried. The dark brown solids were dissolved in 550 mL of boiling
toluene and decolorizing carbon was added. The hot mixture was filtered
through Celite, and the Celite/carbon mixture was washed with 250
mL of boiling toluene. The product was crystallized from toluene and
In an effort to facilitate the understanding of Zn2+ in cell
biology and neurology, we recently described the preparation
(24) Tsien, R. Y. Fluorescent and Photochemical Probes of Dynamic
Biochemical Signals Inside LiVing Cells; Czarnik, A. W., Ed.; American
Chemical Society: Washington D. C., 1993; Vol. 538, pp 130-146.
(25) Czarnik, A. W. Curr. Biol. 1995, 2, 423-428.
(26) Frederickson, C. J.; Kasarskis, E. J.; Ringo, D.; Frederickson, R.
E. J. Neurosci. Methods 1987, 20, 91-103.
(27) Zalewski, P. D.; Forbes, I. J.; Betts, W. H. Biochem. J. 1993, 296,
403-408.
(28) Mahadevan, I. B.; Kimber, M. C.; Lincoln, S. F.; Tiekink, E. R. T.;
Ward, A. D.; Betts, W. H.; Forbes, I. J.; Zalewski, P. D. Aust. J. Chem.
1996, 49, 561-568.
(29) Fahrni, C. J.; O’Halloran, T. V. J. Am. Chem. Soc. 1999, 121,
11448-11458.
(30) Godwin, H. A.; Berg, J. M. J. Am. Chem. Soc. 1996, 118, 6514-
6515.
(31) Walkup, G. K.; Imperiali, B. J. Am. Chem. Soc. 1997, 119, 3443-
3450.
(32) Walkup, G. K.; Imperiali, B. J. Org. Chem. 1998, 63, 6727-6731.
(33) Thompson, R. B.; Maliwal, B. P.; Fierke, C. A. Anal. Chem. 1998,
70, 1749-1754.
(34) Czarnik, A. W. Acc. Chem. Res. 1994, 27, 302-308.
(35) Koike, T.; Watanabe, T.; Aoki, S.; Kimura, E.; Shiro, M. J. Am.
Chem. Soc. 1996, 118, 12696-12703.
(39) The name Zinpyr indicates the structural composition of the ligand
(four pyridyl groups) as well as its ability to “peer” into the Zn2+
concentration of samples.
(40) Walkup, G. K.; Burdette, S. C.; Lippard, S. J.; Tsien, R. Y. J. Am.
Chem Soc. 2000, 122, 5644-5645.
(41) Lakowicz, J. R. Principles of Fluorescence Spectroscopy; 2nd ed.;
Kluwer Academic/Plenum: New York, 1999.
(36) Hirano, T.; Kikuchi, K.; Urano, Y.; Higuchi, T.; Nagano, T. Angew.
Chem., Int. Ed. 2000, 39, 1052-1054.
(37) Canzoniero, L. M. T.; Sensi, S. L.; Choi, D. W. Neurobiol. Dis.
1997, 4, 275-279.
(38) Hirano, T.; Kikuchi, K.; Urano, Y.; Higuchi, T.; Nagano, T. J. Am.
Chem. Soc. 2000, 122, 12399-12400.
(42) Gruenwedel, D. W. Inorg. Chem. 1968, 7, 495-501.

Fluorescent Sensors for Zn2+
J. Am. Chem. Soc., Vol. 123, No. 32, 2001 7833
recrystallized (4:1 toluene/EtOH) to yield a white crystalline solid (32.3
g, 56.7%): TLC Rf ) 0.41 (3:1 hexanes/EtOAc); mp ) 240-42 °C.
1H NMR (CDCl3, 500 MHz) δ 2.42 (6H, s), 6.76 (2H, d, J ) 9.0 Hz),
6.93 (2H, d, J ) 8.5 Hz), 7.28 (1H, d, J ) 7.5 Hz), 7.55 (4H, t, J )
7.5 Hz), 7.65-7.74 (4H, m), 8.06 (1H, d, J ) 8.0 Hz), 8.25 (4H, d, J
) 7.5 Hz). 13C NMR (CDCl3, 125 MHz) δ 9.85, 82.97, 116.60, 118.09,
119.63, 124.47, 125.39, 125.79, 126.60, 128.91, 129.14, 130.16, 130.48,
134.11, 135.35, 150.44, 150.80, 153.03, 164.74, 169.42. FTIR (KBr,
cm-1) 1771, 1763, 1738, 1599, 1592, 1452, 1421, 1267, 1222, 1099.
HRMS (+FAB) calcd for MH+, 569.1600; found, 569.1588.
4′,5′-Bis(bromomethyl)fluorescein Dibenzoate (2). 4′,5′-Dimethyl-
fluorescein dibenzoate (5.00 g, 8.79 mmol) and 1,3-dibromo-5,5-
dimethylhydantoin (3.87 g, 13.6 mmol) were combined in 550 mL of
C6H5Cl, and acetic acid (133 µL, 2.32 µmol) and 1,1′-azobis-
(cyclohexanecarbonitrile) (0.181 g, 0.740 mmol) were added to the
solution with stirring. The solution was stirred at 40 °C for 72 h and
then washed four times with hot water (100 mL, 80 °C). Recrystalli-
zation (9:1 toluene:EtOH) and washing with n-pentane yielded the
product as a white crystalline solid (6.27 g, 98.2%). TLC Rf ) 0.34
(7:3 hexanes:EtOAc).; mp ) 300 °C dec. 1H NMR (CDCl3, 500 MHz)
δ 4.87 (2H, d, J ) 10.0 Hz), 4.88 (2H, d, J ) 10.5 Hz), 6.92 (2H, d,
J ) 8.5 Hz), 7.08 (2H, d, J ) 9.0 Hz), 7.32 (1H, d, J ) 8.0 Hz), 7.58
(4H, t, J ) 7.5 Hz), 7.68-7.78 (4H, m), 8.08 (1H, d, J ) 8.5 Hz),
8.28 (4H, d, J ) 7.0 Hz). 13C NMR (CDCl3, 125 MHz) δ 20.69, 81.60,
117.09, 119.21, 119.38, 124.47, 125.63, 126.36, 128.68, 128.99, 129.10,
130.58, 130.65, 134.50, 135.71, 149.63, 150.79, 152.48, 164.39, 169.04.
FTIR (KBr, cm-1) 1774, 1743, 1601, 1589, 1427, 1234, 1081, 1066.
HRMS (+FAB) calcd for MH+, 724.9811; found, 724.9824.
152.86, 158.19, 160.21, 169.64. FTIR (KBr, cm-1) 3448, 1755, 1591,
1435, 760. HRMS (+FAB) calcd for MH+, 755.2982; found, 755.2959.
9-(o-Carboxyphenyl)-2,7-dichloro-4,5-bis[bis(2-pyridylmethyl)-
aminomethyl]-6-hydroxy-3-xanthanone (Zinpyr-1, 5). DPA (1.59 g,
7.99 mmol) and paraformaldehyde (0.224 g, 7.47 mmol) were combined
in 20 mL of CH3CN and refluxed for 30 min. 2′,7′-Dichlorofluorescein
(1.00 g, 2.49 mmol) in 30 mL of CH3CN/H2O (1:1) was added to the
solution, and the reaction mixture was refluxed for 24 h. The CH3CN
was removed, and the product and residual water were triturated with
30 mL of boiling ethanol. The product was precipitated at -25 °C,
filtered on a frit, washed thoroughly with ice-cold water, and dried.
Flash chromatography on activated neutral alumina (45:1 CHCl3:
MeOH) yielded the product as a salmon-pink solid (960 mg, 46.8%).
TLC Rf ) 0.10 (alumina, 19:1 CHCl3:CH3OH); mp ) 185-187 °C
1
dec. H NMR (CDCl3, 500 MHz) δ 3.98 (4H, d, J ) 15.0 Hz), 4.01
(4H, d, J ) 15.0 Hz), 4.20 (4H, s), 6.64 (2H, s), 7.19 (5H, t, J ) 7.5
Hz), 7.36 (4H, d, J ) 8.0 Hz), 7.64-7.73 (6H, m), 8.04 (1H, d, J )
7.0 Hz), 8.60 (4H, dq, J ) 1.0, 5.0 Hz). 13C NMR (CDCl3, 125 MHz)
δ 49.38, 59.31, 89.44, 110.26, 111.93, 117.72, 122.67, 123.43, 124.27,
125.54, 127.28, 127.86, 130.31, 135.41, 137.32, 148.99, 151.83, 156.09,
157.81, 169.13. FTIR (KBr, cm-1) 3447, 1761, 1750, 1592, 1475, 1435,
761. HRMS (+FAB) calcd for MH+, 823.2202; found, 823.2229.
[Zn2(Zinpyr-1)(H2O)2](ClO4)2 (6). Zinpyr-1 (50.6 mg, 61.5 µmol)
was combined with zinc triflate (2.00 mL, 0.121 M, 243 µmol) and
sodium hydroxide (2.00 mL, 0.121 M, 243 µmol) in CH3CN (5 mL)
and sonicated for 10 min to dissolve all of the solids. Sodium perchlorate
(0.500 g, 4.70 mmol) was dissolved in water (5 mL) and added to the
CH3CN solution, then the combined solutions were filtered through
Celite. The product partially crystallized at room temperature after 12
h. The remainder of the product was crystallized at -25 °C over 12 h.
The orange crystalline material was filtered, powdered, and dried under
vacuum at 50 °C to yield 37.0 mg of dried product in 48.4% yield.
Analogous small-scale reactions at 25 °C yielded orange square prisms
4′,5′-Fluoresceindicarboxaldehyde (3). 4′,5′-Bis(bromomethyl)-
fluorescein dibenzoate (2.00 g, 2.75 mmol) and NaHCO3 (2.00 g, 23.8
mmol) were combined in 200 mL of DMSO and heated to 150 °C for
4 h. The dark red solution was cooled and then diluted into 700 mL of
2 M HCl and stirred for 2 h. The aqueous material was extracted
thoroughly with CH2Cl2 (8 × 100 mL), and the solvents were removed
to isolate an orange liquid. The orange solid that precipitated with the
addition of 300 mL of water was collected on a frit and washed
thoroughly with water. The orange solid was redissolved in CH2Cl2
and dried over Na2SO4. An orange solid was isolated after filtration
and solvent removal. Flash chromatography on silica gel (33:1 CHCl3:
MeOH) yielded the product as a yellow powder (391 mg, 36.7%). TLC
1
suitable for X-ray crystallography. H NMR (DMSO-d6, 500 MHz) δ
4.13-4.41 (12H, m), 6.14 (2H, s), 7.14 (1H, d, J ) 8.0 Hz), 7.54 (4H,
q, J ) 8.0 Hz), 7.604 (4H, t, J ) 8.0 Hz), 7.70 (1H, t, J ) 7.0 Hz),
7.84 (1H, t, J ) 7.5 Hz), 7.92 (1H, t, J ) 7.5 Hz), 8.07 (4H, quin.d,
J ) 1.5, 7.5 Hz), 8.68 (2H, d, J ) 5.0 Hz), 8.74 (2H, d, J ) 5.0 Hz).
FTIR (KBr, cm-1) 3441, 1748, 1610, 1571, 1457, 1363, 1121. Anal.
Calcd for Zn2C46H40N6O16 (6‚H2O): C, 45.84; H, 3.34; N, 6.97.
Found: C, 45.89; H, 3.13; N, 6.71.
1
Rf ) 0.46 (19:1 CHCl3:CH3OH); mp ) 301-303 °C dec. H NMR
Collection and Reduction of X-ray Data. Crystals were covered
with Paratone-N oil and suitable specimens were mounted on the tips
of glass fibers at room temperature and transferred to a Bruker (formerly
Siemens) CCD X-ray diffraction system with a graphite-monochro-
matized Mo KR radiation (λ ) 0.71073 Å) controlled by a Pentium-
based PC running the SMART software package.43 Data were collected
at 188 K in a stream of cold N2 maintained with a Siemens LT-2A
nitrogen cryostat. Procedures for data collection and reduction have
been reported previously.44 The structure was solved by direct methods
and refined by full matrix least-squares and difference Fourier
techniques with SHELXTL.45 Empirical absorption corrections were
applied with the SADABS program,46 and the structure was checked
for higher symmetry by PLATON.47 The space group was determined
by examining systematic absences and confirmed by the successful
solution and refinement of the structure. All non-hydrogen atoms were
refined anisotropically. Hydrogen atoms were assigned idealized
locations and given isotropic thermal parameters 1.2× the thermal
parameter of the carbon atoms to which they were attached. In the
structure of 6‚6H2O, one perchlorate anion is disordered and several
water molecules are partially occupied; all were modeled and refined
accordingly. Relevant crystallographic information is summarized in
Table 1, and the 50% thermal ellipsoid plot is shown in Figure 1.
(CDCl3, 500 MHz) δ 6.74 (2H, d, J ) 8.5 Hz), 6.94 (2H, d, J ) 9.5
Hz), 7.18 (1H, d, J ) 7.5 Hz), 7.69 (1H, td, J ) 1.0, 7.5 Hz), 7.74
(1H, td, J ) 1.0, 7.5, Hz), 8.07 (1H, d, J ) 7.5 Hz), 10.67 (2H, s),
12.13 (2H, s). 13C NMR (CDCl3, 125 MHz) δ 80.80, 109.23, 109.71,
115.42, 123.90, 125.76, 126.73, 130.72, 135.86, 137.16, 151.87, 152.11,
164.87, 168.87, 192.00. FTIR (KBr, cm-1) 1769, 1656, 1600, 1228,
1172, 1092. HRMS (+FAB) calcd for MH+, 389.0661; found,
389.0674.
9-(o-Carboxyphenyl)-4,5-bis[bis(2-pyridylmethyl)aminomethyl]-
6-hydroxy-3-xanthanone (Zinpyr-2, 4). 4′,5′-Fluorescein-dicarbox-
aldehyde (200 mg, 0.515 mmol) and acetic acid (120 µL, 2.1 mmol)
were combined in 1,2-dichloroethane (DCE, 30 mL) and stirred. To
the resulting solution, DPA (215 mg, 1.08 mmol) in DCE (20 mL)
was added dropwise and stirred for 30 min. Sodium triacetoxyboro-
hydride (230 mg, 1.08 mmol) was added, and the reaction mixture was
stirred 12 h at room temperature. The reaction was chilled to 0 °C,
and water was added to the solution with stirring. The aqueous layer
was extracted with CH2Cl2, and the combined organic layers were
washed twice with saturated NaCl to give an orange solid after solvent
removal. The compound was dried by azeotroping with benzene. Flash
chromatography on activated neutral alumina (24:1 CHCl3:MeOH)
yielded the product as an orange solid (108 mg, 27.8%). TLC Rf )
0.10 (alumina, 19:1 CHCl3:CH3OH); mp ) 195-197 °C dec. 1H NMR
(CDCl3, 500 MHz) δ 3.98 (4H, d, J ) 15.0 Hz), 4.02 (4H, d, J ) 15.0
Hz), 4.18 (4H, s), 6.57 (2H, d, J ) 8.5 Hz), 6.64 (2H, d, J ) 8.5 Hz),
7.17-7.21 (5H, m), 7.37 (4H, d, J ) 7.5 Hz), 7.58-7.67 (6H, m),
8.00 (1H, d, J ) 7.5 Hz), 8.59 (4H, d, J ) 6.5 Hz). 13C NMR (CDCl3,
125 MHz) δ 49.24, 59.45, 109.84, 110.24, 113.38, 122.57, 123.52,
124.38, 125.10, 127.69, 128.30, 129.70, 134.92, 137.16, 149.02, 150.59,
(43) SMART, 5.05 ed.; Bruker AXS, Inc.: Madison, WI, 1998.
(44) Feig, A. L.; Bautista, M. T.; Lippard, S. J. Inorg. Chem. 1996, 25,
6892-6898.
(45) Sheldrick, G. M. SHELXL97-2: Program for the Refinement of
Crystal Structures; University of Gottingen: Germany, 1997.
(46) Sheldrick, G. M. SADABS: Area-Detector Absorption Correction;
University of Gottingen: Germany, 1996.
(47) Spek, A. L. PLATON, A Multipurpose Crystallographic Tool;
Utrecht University: Utrecht, The Netherlands, 1998.

7834 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
Burdette et al.
Table 1. Crystallographic Parameters for
[Zn2(Zinpyr-1)(H2O)2](ClO4)2‚6H2O (6‚6H2O)
of a Pentium II-based PC running the Windows NT ChemStation
software package or a Cary 1E scanning spectrophotometer under the
control of a Pentium PC running the manufacturer-supplied software
package. Spectra were routinely acquired at 25 °C, which was
maintained by a circulating water bath in 1-cm path length quartz
cuvettes having a volume of 1.0 or 3.5 mL.
Fluorescence Spectroscopy. Fluorescence spectra were recorded on
a Hitachi F-3010 spectrofluorimeter under the control of a Pentium-
based PC running the SpectraCalc software package. Excitation was
provided by a 150 W Xe lamp (Ushio Inc.) operating at a current of 5
A. All spectra were normalized for excitation intensity via a rhodamine
quantum counter, and emission spectra were normalized by the
manufacturer-supplied correction curves. Spectra were routinely ac-
quired at 25 °C, maintained by a circulating water bath in 1 cm × 1
cm quartz cuvettes using 3 nM slit widths and a 240 nm/min scan speed.
Fluorescence emission measurements were also acquired in a 1 cm ×
1 cm quartz cell using a Spex Fluorolog-2 instrument with 1-nm
bandwidth slits. All spectra were corrected for emission intensity by
using the manufacturer-supplied photomultiplier curves.
(a) pH-Dependent Fluorescence Studies. The apparent pKa was
measured by plotting the integrated area of the fluorescence emission
spectrum against pH recorded in the range from pH 12.5 to 2.5. A 1
µM solution of the Zinpyr dye (10 mL) containing ∼1 mM KOH and
100 mM KCl was adjusted to pH 12.5, and the UV-vis and
fluorescence spectra were recorded. The pH was lowered in steps of
∆pH ) 0.5 with the addition of appropriate amounts of 6, 2, 0.5, 1,
0.1, and 0.01 N HCl, recording the absorption and emission spectra at
each pH interval. The volume of the solution was controlled such that
the overall change in volume was <2%. Emission for Zinpyr-1 was
integrated from 450 to 700 nm. Emission for Zinpyr-2 was integrated
from 500 to 700 nm. The resulting integrated emission spectral areas
were normalized, plotted against pH, and fit to the nonlinear expression
in eq 1 to calculate the pKa value. In the equation, ∆F1max and ∆F2max
are the maximum fluorescence changes associated with the correspond-
ing pKa values.
formula
C46H34Cl4N6O21Zn2
formula wt
space group
a, Å
b, Å
c, Å
â, deg
V
1279.33
P21/n
14.1579(4)
18.2353(4)
21.7207(3)
90.705(2)
5607.3(2)
4
1.515
1.126
188(2)
28159
9552
Z
F
calcd(g cm-1
)
absorp. coeff. (mm -1
temp, K
)
total no. data
no. of unique data
obs dataa
5316
724
8.96
16.86
no. parameters
R, %b
wR2, %c
max/min peaks, e/Å3
0.867, -0.496
a Observation criterion: I > 2σ(I). b R ) Σ||Fo| - |Fc||/Σ|Fo|. c wR2
2
2
) {Σ[w(Fo - Fc2)2]/Σ[w(Fo )2]}1/2
.
∆F1max
(1 + 10(pH - pK
∆F2max
(1 + 10(pH - pK
∆F )
+
(1)
)
a1
)
a2
)
)
(b) Quantum Yield. The quantum yields for fluorescence were
obtained by comparison of the integrated area of the corrected emission
spectrum of the samples with that of a solution of fluorescein in 0.1 N
NaOH, which has a quantum efficiency of 0.95.41 The concentration
of the reference was adjusted to match the absorbance of the test sample.
The quantum efficiency of the metal-free ligand was measured by using
a dilute sample of Zinpyr dye (∼1 × 10-6 M, Abs e 0.1) in 10 mM
PIPES, pH ) 7.0; 100 mM KCl; and 50 µM EDTA. The quantum
efficiency of metal-bound ligand was measured by using a dilute sample
of Zinpyr (∼8 × 10-7 M, Abs e 0.1) in 10 mM PIPES, pH ) 7.0; 100
mM KCl; and 100 µM ZnCl2. The concentration of the reference was
adjusted to match the absorbance of the test sample at the wavelength
of excitation. Emission for Zinpyr-1 was integrated from 480 to 650
nm with excitation at 475 nm, and emission for Zinpyr-2 was integrated
from 496 to 600 nm with excitation at 492 nm. The quantum yields
were calculated with the expression in eq 2.
Figure 1. ORTEP diagram of [Zn2(Zinpyr-1)(H2O)2(ClO4)2]‚6H2O
(6‚6H2O) showing 50% thermal ellipsoids and selected atom labels.
Water molecules, hydrogen atoms, and disordered perchlorates are
omitted for clarity.
General Spectroscopic Methods. Ultrol grade HEPES (2[4-(2-
hydroxyethyl)-1-piperazinyl]ethanesulfonic acid) and PIPES (pipera-
zine-N,N′-bis(2-ethanesulfonic acid)) from Calbiochem and KCl
(99.997%) were purchased and used as received. All solutions were
filtered through 0.2-µm cellulose filters before measurements. Except
for the fluorescence titration experiment, Zn solutions were prepared
by the addition of appropriate amounts of 1.0 M, 100 mM, 10 mM or
1 mM Zn2+ stocks that were checked by atomic absorption spectroscopy
for concentration accuracy, or by titration with terpyridine and
measurement of the absorption spectra. The titration was performed
by treating a 70 µM solution of 2,2′:6′,2′′-terpyridine in 10 mM PIPES,
pH 7.0, with aliquots of 10 mM (nominal) ZnCl2 and determining the
equivalence point by monitoring the absorbance of the resulting complex
at 335 nm (ꢀ ) 28.5 × 10-3 M-1 cm-1). The Zn2+ stocks were prepared
from 99.999% pure ZnCl2. The purity of the Zinpyr probes was verified
to be greater than 99% by HPLC. Zinpyr was introduced to aqueous
solutions by addition of a stock solution in DMSO (0.67 µM). Graphs
were manipulated and equations were calculated by using Kaleidagraph
3.0. The pH values of solutions were recorded using an Orion glass
electrode that was calibrated prior to each use.
emission
∫
sample
Φsample ) Φstandard
×
(2)
emission
∫
standard
(c) Titration of Zn2+ Binding by Fluorescence Spectroscopy (Kd1).
The first Kd associated with Zn2+ binding was measured by a
fluorescence titration as described previously.48 Zinpyr-1 spectra were
acquired by exciting at 507 nm and integrating the emission from 475
to 650 nm. Zinpyr-2 spectra were acquired by exciting at 490 nm and
collecting and integrating data from 500 to 575 nm. The measurements
were performed in triplicate to ensure accuracy of the derived Kd value.
UV-Visible Spectroscopy. Absorption spectra were recorded on a
Hewlett-Packard 8453A diode array spectrophotometer under the control
(48) Ref. 40, Supporting Information.

Fluorescent Sensors for Zn2+
J. Am. Chem. Soc., Vol. 123, No. 32, 2001 7835
Scheme 1
(d) Titration of Zn2+ Binding by Absorption Spectroscopy (Kd2).
A 3.0-mL solution containing 10 µM Zinpyr-1 in PIPES buffer (100
mM KCl, 50 mM PIPES, pH 7) was prepared, and an initial absorption
measurement was made. A portion of Zn2+ was added from standardized
stock solutions to give an equimolar solution of ligand and metal ion,
and the absorbance was measured. Additional Zn2+ aliquots were added
to give final metal ion concentrations of 250 and 500 µM and 1, 5, 10,
and 15 mM. Calculation of the ∆A by subtracting the 10 µM Zn2+
spectrum from the higher concentration spectra, followed by plotting
against wavelength, revealed maximum changes at 497 (increase) and
520 (decrease) nm.
For Zinpyr-2, a 3.0-mL solution containing 19 µM Zinpyr-2 in PIPES
buffer, pH 7 (100 mM KCl, 50 mM PIPES) was prepared, and an initial
absorbance measurement was made. Zn2+ aliquots were titrated into
the solutions to give final concentrations of 1-20 µM (1 µM
increments), 30, 40, 50, 150, 250, 500, and 750 µM, and the absorption
spectra were recorded. The ∆A values were calculated by subtracting
the 0 µM spectrum from the subsequent spectra and were plotted against
the wavelength. The maximum absorption changes occurred at 486 nm
(increases) and 508 nm (decreases). The change at 508 nm corresponds
to the first binding event, but the increase at 486 nm is not indicative
of a single process.
compound in ∼60% yield after trituration with boiling ethanol
and washing with cold water. The crude material, which is
>90% pure after trituration (1H NMR), can be further purified
by chromatography on neutral alumina or reverse-phase silica
to give analytically pure material (HPLC). This synthesis pro-
vides easy access to gram quantities of the desired Zn2+ probe.
Although fluorescein undergoes electrophilic substitution
preferentially at the 4′ and 5′ positions,50,51 a mixture of
structural isomers with substitution at the 4′, 5′, 2′, and 7′
positions is obtained when the compound is subjected to the
Mannich conditions that were used to prepare Zinpyr-1. These
isomers are inseparable by conventional flash chromatography.
We therefore envisioned that a fluorescein compound function-
alized at the 4′ and 5′ positions with a synthetically useful group
such as an aldehyde could be employed as a scaffold for
preparing both the Zinpyr-1 analogue and a variety of related
sensors. Scheme 2 outlines the 4-step synthetic route devised
to prepare Zinpyr-2 in which the dialdehyde 3 is a key
intermediate.
The synthesis of 1 was achieved by modification of the
published procedure that involves installation of benzoate
protecting groups on the phenolic oxygen of 4′,5′-dimethylfluo-
rescein.52 In the initial report, 1 was prepared using benzoyl
chloride and characterized only by its melting point. Unprotected
fluoresceins are highly polar compounds that are only slightly
soluble in most common organic solvents. The benzoate
protecting groups provide a convenient method for purifying
fluorescein compounds by chromatography or crystallization and
facilitating subsequent chemical manipulation by enhancing their
solubility in organic media. In addition, protecting the phenolic
oxygen atoms forces the fluorescein to adopt the lactoid isomer,
preventing isomerization between the quinoid and lactoid forms.
Compound 2 can be formed in multigram quantities by
bromination of 1 under free radical conditions. The reaction
proceeded under facile conditions to give 2 in ∼95% purity
before recrystallization. No dibromination was observed. Com-
pound 2 can be carried on without purification to the next step,
or recrystallized to give highly pure material (+99%, TLC/1H
NMR). Attempts to prepare sensors directly from 2 by direct
displacement of the bromides were unsuccessful. Side reactions
involving the benzoate protecting groups and lactone ring are
suspected to be the problems.
The measurements were repeated at pH 7.5 in HEPES buffer (100
mM KCl, 50 mM HEPES) for both Zinpyr-1 and Zinpyr-2 using the
methods described above. The absorbance change for Zinpyr-1 was fit
to a binding curve, and the spectral response of Zinpyr-2 was corrected
for dilution and analyzed using SPECFIT, a nonlinear least-squares
fitting program.49
Cell Preparation. COS-7 cells were plated onto ethanol-washed,
UV-sterilized glass coverslips and grown to 60-80% confluence at
37 °C in DMEM (Gibco catalog no. 31600-075) supplemented with
10% v/v fetal bovine serum. For labeling, the cells were washed once
with HBSS (Gibco catalog no. 11201-092) supplemented with 2 g/L
D-glucose and 20 mM HEPES pH 7.4, then incubated with 5 µM
fluorophore in HBSS at 37 °C for 0.5-1h. Prior to imaging, the labeling
solution was aspirated and the cells were washed three times with
HBSS.
Cell Imaging. Cells were imaged on a Zeiss Axiovert microscope
equipped with a cooled CCD camera (Princeton Instruments, Trenton,
NJ). Excitation light was provided by a 150 W Xe arc lamp transmitted
through a 465-495-nm excitation filter and a 505-nm long-pass dichroic
mirror, and fluorescence was measured after passage through a 513-
558-nm emission filter. Filters were obtained from Omega Optical and
Chroma Technologies (Brattleboro, VT).
Results and Discussion
Synthesis. The synthesis of Zinpyr-1 was accomplished by
a Mannich reaction between bis(2-pyridylmethyl)amine (dipi-
colylamine or DPA), prepared from 2-pyridinecarboxaldehyde
and 2-aminomethylpyridine according to published procedures,42
and 2′,7′-dichlorofluorescein (DCF, Scheme 1). Formation of
the imminum cation by condensation of paraformaldehyde and
DPA and subsequent reaction with DCF afforded the desired
Oxidation of 2 with DMSO in the presence of NaHCO3
yielded the dialdehyde 3. Rigorous drying of the DMSO by
distillation from CaH2 followed by storage over molecular sieves
is required to obtain the product in ∼40% yield. Reactions using
undistilled DMSO reduced the isolated yields by >50%. The
(50) Sandin, R. B.; Orvis, R. L. J. Org. Chem. 1958, 23, 1234-1235.
(51) Shipchandler, M. T.; Fino, J. R. Anal. Biochem. 1986, 154, 476-
477.
(49) Binstead, R.; Zuberbuhler, A. D. SPECFIT Global Analysis System,
2.10 ed.; Spectrum Software Associates: Chapel Hill, NC 27515, 1998.
(52) Burton, H.; Kurzer, F. J. Soc. Chem. Ind. London 1948, 67, 345.

7836 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
Burdette et al.
Scheme 2
synthesis of 3 resulted in cleavage of the benzoate protecting
groups, but H-bonding between the phenolic hydrogens and the
aldehyde carbonyl oxygens enforces the lactoid isomer, as
shown by the NMR chemical shift of the phenolic hydrogen
atoms (δ ) 12.2), compensating for the loss of the protecting
groups. Compound 3 can be condensed with primary amines
to give imines, or it can be aminated under reducing conditions
to give amines. The synthetic chemistry of 3 and the use of the
compound as a scaffold for sensors as well as for other purposes
will be the subjects of future reports. Zinpyr-2 was prepared
by the reaction of 3 with DPA using NaBH(OAc)3 as the
reducing agent in fairly good yield. The zinc complex of
Zinpyr-1 crystallizes within 2 h from an aqueous solution of
CH3CN in the presence of excess NaClO4. The orange crystal-
line compound is nonfluorescent under UV light.
and other first-row transition metal ions including Cu+, Cu2+
Ni2+, Co2+, Fe2+, and Mn2+ quench the fluorescence.54
,
The fluorescent increase and binding affinity for Zn2+ were
characterized by using a dual-metal single-ligand buffer system55
comprising 1 mM EDTA, 2 mM Ca2+ and 0-1 mM total Zn2+
.
The formation of the [Zn(EDTA)]2- complex, which has an
apparent Kd of 2.11 nM under these conditions, allowed the
controlled variation of [Zn2+]. Figure 2a,b shows the results of
representative titrations for Zinpyr-1 and Zinpyr-2, respectively.
Whereas the quantum yield of Zinpyr-1 increased by 2.25-fold,
changes in the absorption properties of the bound and unbound
probe afforded a 3.1-fold change in integrated emission.
Similarly, the quantum yield of Zinpyr-2 increased by ∼3.7-
fold, and the integrated emission increased 6.0-fold.
The Zn2+ affinity of the Zinpyr sensors was determined by
performing these measurements in triplicate using different
Ca2+/EDTA/Zn2+ buffers. The measurements indicated that the
[(Zn)Zinpyr-2] complex has an apparent Kd at pH 7.0 of 0.5 (
0.1 nM (mean ( esd) that is slightly lower than that of Zinpyr-
1, for which the Kd is 0.7 ( 0.1 nM. The [Zn(DPA)]2+ complex
has an apparent Kd of 70 nM at pH 7, approximately 2 orders
of magnitude weaker than the observed Kd values for the Zinpyr
compounds.56 The tighter binding of the latter suggests that there
are additional groups involved in Zn2+ chelation. The fluores-
cence response fit to a Hill coefficient of 1, consistent with the
formation of a 1:1 Zinpyr:Zn2+ complex responsible for the
enhancement. The initial speculation that Zinpyr-1 might bind
Zn2+ in a “TPEN-like” manner, where TPEN is the intracellular
heavy metal ion chelator N,N,N′,N′-tetra(2-picolyl)ethylenedi-
amine, was not consistent with the shifting of the excitation
wavelengths or the binding affinity of TPEN (∼1 fM at pH 7).
Fluorescence Properties of Zinpyr Sensors. Under simu-
lated physiological conditions (50 mM PIPES, 100 mM KCl)
at pH 7 in the presence of EDTA to scavenge adventitious metal
ions, Zinpyr-1 and Zinpyr-2 have quantum yields of 0.38 and
0.25, respectively. The quantum yields increase to 0.87 for
Zinpyr-1 and 0.92 for Zinpyr-2 in the presence of 25 µM Zn2+
.
The excitation maximum of Zinpyr-1 shifts from 515 nm (ꢀ )
79.5 × 103 M-1 cm-1) to 507 nm (ꢀ ) 84.0 × 103 M-1 cm-1
)
upon Zn2+ complexation. Zinpyr-2 exhibits a similar shift from
498 nm (ꢀ ) 36.8 × 103 M-1 cm-1) to 490 nm (ꢀ ) 44.9 ×
103 M-1 cm-1) after the addition of Zn2+. The slight hypso-
chromic shift in the absorption wavelength is indicative of
coordination of the donor group (phenol) to Zn2+ 53
Since the
.
phenol is incorporated into the π-system of the fluorophore,
coordination to a metal is expected to perturb the electronic
structure of the system and produce a shift in the excitation
wavelength. The fluorescence response is Zn2+-selective. Ca2+
and Mg2+ concentrations as high as 5 mM produce no change,
(54) Details provided in Supporting Information.
(55) Wolf, H. U. Experientia 1973, 29, 241-249.
(56) Anderegg, G.; Hubmann, E.; Podder, N. G.; Wenk, F. HelV. Chim.
Acta 1977, 60, 123-140.
(53) Balzani, V.; Scandola, F. Supramolecular Photochemistry; Ellis
Horwood: New York, 1991.

Fluorescent Sensors for Zn2+
J. Am. Chem. Soc., Vol. 123, No. 32, 2001 7837
Figure 3. Plot of the normalized integrated emission intensity versus
pH for Zinpyr-1 (a) and Zinpyr-2 (b). The increase in fluorescence
corresponds to protonation of both tertiary amines, and the decrease
corresponds to the isomerization of the fluorescein to its nonfluorescent
zwitterion isomer. The emission intensity varies only minimally from
the physiological range, which is an important property for measuring
intracellular [Zn2+] changes.
Figure 2. Fluorescence emission response of Zinpyr-1 (a) and Zinpyr-2
(b) to buffered Zn2+ solutions. Spectra were acquired in 100 mM KCl,
50 mM PIPES, pH 7.00, at 25 °C. Excitation was provided at 490 nm
with 0.25-mm slit widths. Emission data were corrected for the response
of the detector using the manufacturer-supplied curve, and the emission
data points at 490 nm, which were perturbed by scatter, have been
removed for clarity. The spectra shown are for free zinc buffered at 0,
0.172, 0.424, 0.787, 1.32, 2.11, 3.34, 5.60, 10.2, and 24.1 nM,
respectively. For the final spectrum (containing 1 mM EDTA and 1
5.5, with a pKa of 2.8 for Zinpyr-1 and 3.9 for Zinpyr-2, is
attributed to the formation of the nonfluorescent diphenolic
zwitterion isomer.57
mM Zn2+), additional ZnCl2 was added to provide ∼25 µM free Zn2+
.
Inset: fluorescence response obtained by integrating the emission
At high pH, because the unbound sensor is more efficiently
quenched, the magnitude of Zn2+-induced fluorescence change
is more dramatic, increasing over 200-fold at pH 12. The
quenching of the fluorophore at high pH and enhancement by
H+ and Zn2+ are consistent with a photoinduced electron transfer
(PET) mechanism from the benzylic amines. Inhibition of PET
by coordination of d10 transition metals or protons to amines is
commonly observed mechanism for fluorescent enhance-
ment.23,34 The observed fluorescence increase with Cd2+ also
supports the validity of this suggested mechanism. Although
the fluorescence of the Zinpyr sensors is enhanced by protons
as well as Zn2+, the quantum yield experiment demonstrates
that the emission intensity of the protonated form is significantly
lower than that of the Zn complex. More importantly, the
emission intensity over the physiologically relevant pH range
spectra between 500 and 575 nm, subtracting the baseline (0 Zn2+
)
spectrum and normalizing to the full scale response (25 µM free Zn2+).
The fluorescence of Zinpyr was also enhanced by protonation
of the tertiary amines. Zinpyr fluorescence is almost completely
quenched at pH > 12, and reaches a maximum for both sensors
near pH 5.5. The pH change fits to an apparent pKa of 8.4 for
Zinpyr-1 (Figure 3a) and 9.4 for Zinpyr-2 (Figure 3b). The
plateau of the fluorescence corresponds to a state of the molecule
in which both tertiary amines are protonated. The difference in
the pKa values between the two sensors can be attributed to the
chlorine atoms on the xanthenone ring, which decrease the pKa
of the phenols. The increased acidity of the phenol groups should
have an effect on the fluorescence pKa by modifying the local
pH in the vicinity of the amines. In addition to the observed
difference in pKa values for the two sensors, each of the tertiary
amines of the individual molecules is predicted to have a
different pKa value because of the difference in environments
between the amines on the phenolic and keto oxygen halves of
the molecule. This difference is not manifest in a stepwise
change in the fluorescence intensity; however, the amine
adjacent to the keto oxygen should become protonated before
the more acidic site adjacent to the deprotonated phenol. The
fluorescence quenching observed for both molecules below pH
(pH 5.5 to pH 8) does not vary significantly (Figure 3), so Zn2+
induced changes can be readily measured.
-
Stoichiometry, Kd2, and Solution Behavior. Small increases
in the absorption spectrum of Zinpyr-1 can be observed by
titrating more than 1 equivalent of Zn2+ per ligand into buffered
solutions of Zinpyr-1 at pH 7. At Zn2+ concentrations in the
micromolar-to-millimolar range, spectral changes occur with
maxima at 497 nm (increases) and 520 (decreases) (Figure 4).
(57) Makuszewski, R.; Diehl, H. Talanta 1980, 27, 937-46.

7838 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
Burdette et al.
Figure 4. Plot of the difference in absorption between the 1:1 Zinpyr-
1:Zn2+ spectrum, and the absorption spectra after the addition of Zn2+
to give concentrations of 250 and 500 µM and 1, 5, 10, and 15 mM.
Inset: plot of the change in absorbance and the concentration of [Zn2+].
Figure 6. Plot of the percentage of each species present in the titration
of Zinpyr-2 with increasing concentrations of Zn2+. The titration was
performed with 19 µM Zinpyr-2 in PIPES buffer (pH 7, 100 mM KCl).
The total [Zn2+] represents the total concentration of Zn2+ titrated into
the solution. Zinpyr-2 (O) decreases upon the formation of the [(Zn)-
Zinpyr-2] complex (4). At higher concentrations of Zn2+, [(Zn)2Zinpyr-
2]2+ (0) forms. The formation of [(Zn)2Zinpyr-2]2+ is not detected until
all of the higher-affinity 1/1 sites are occupied.
orders of magnitude weaker than the first and is not ac-
companied by a fluorescence change at pH 7.
Scheme 3 shows our interpretation of the behavior of the
Zinpyr ligands in solution, which is consistent with all of the
spectral measurements. At pH 7, both benzylic amines are
expected to be largely protonated, inhibiting PET quenching of
the unmetalated fluorophore. Binding of the first Zn2+ more
efficiently interrupts the PET process and increases the HOMO-
LUMO gap between the S0 and S1 energy levels. The Kd1 value
is lower than that of DPA because of additional coordination
by the phenolic oxygen. It is difficult to measure accurately
the affinity of the second binding event, because Zn2+ is added
from unbuffered solutions, and the Zn2+ ion has limited
solubility at pH 7-7.5; however, a second binding event is
consistent with all of our observations, including the X-ray
crystal structure determination. The measured Kd2 of both ligands
is significantly weaker than Kd1, presumably because of loss of
coordination by the phenol and electrostatic repulsion between
the positively charged metalated ligand and the incoming Zn2+
cation. The lack of a significant shift in the absorbance, and
the persistence of a fluorescent signal at high concentrations of
Zn2+, suggest that the lactone ring remains open in aqueous
solution.
Figure 5. Absorption spectra of Zinpyr-2 [19 µM] after the addition
of Zn2+ to give final concentrations of 8-20 by increments of 1, 30,
40, 50, 150, 250, 500, and 750 µM. Addition of Zn2+ induces a decrease
and absorption at 508 nm and an increase at 486 nm.
Examination of the absorption changes suggested the occurrence
of a second binding event with a weak Kd2 in the vicinity of
∼85 ( 10 µM. The affinity of this weaker binding event Kd2
increased to ∼35 µM at pH 7.5, consistent with the second Zn2+
binding associated with displacement of a single proton.
The binding of Zn2+ by Zinpyr-2 was accompanied by slightly
different changes in the absorption spectra (Figure 5). Titration
through the first binding event showed that the absorption
changes around 485 indicate the occurrence of more than one
process. Addition of greater than one equivalent of Zn2+
produced small changes in the absorption spectra of Zinpyr-2
similar to the behavior of Zinpyr-1. Analysis of the spectral
data with SPECFIT suggest a Kd2 in the range of 9.0 ( 1.0
µM, which decreases to ∼2 µM at pH 7.5. Figure 6 shows the
percentage of each species that was detected during the titration
of Zinpyr-2 (19 µM) with increasing concentrations of total
Zn2+. The second binding event for both ligands was not
associated with a measurable change in fluorescence at physi-
ologically relevant pH. Formation of insoluble zinc(II) hydroxide
species at high [Zn2+] and pH 7 makes it difficult to estimate
free metal ion concentrations, so the values reported apply only
to the stated conditions. These experiments suggest, however,
that the affinity of the second binding site for Zn2+ is several
Structural Studies. The nature of the Zn2+ complex of
Zinpyr-1 in the solid state was investigated by X-ray crystal-
lography. Attempts to crystallize 6 with Cl-, I-, NO3-, PF6
,
-
-
2-
BF4-, BPh4-, CF3SO3 and SO4 counterions in a variety of
-
solvents failed to yield X-ray quality crystals. Only the ClO4
complex yielded single crystals of suitable quality for structural
analysis. The structure of 6 is displayed in Figure 1 as an
ORTEP diagram. Single-crystal X-ray diffraction results are
shown in Table 1, and selected bond distances and angles are
-
furnished in Table 2. Two ClO4 and six disordered H2O
molecules are also present in the asymmetric unit. The structure
of 6 is, to our knowledge, the first X-ray structure of a metal
bound to fluorescein molecule having appended donor moieties.
A survey of the CSD (Cambridge Structural Database)58
revealed only nine examples of crystal structures of any
(58) Allen, F. H.; Kennard, O. Chemical Design Automation News, 1993,
8, 31-37.

Fluorescent Sensors for Zn2+
J. Am. Chem. Soc., Vol. 123, No. 32, 2001 7839
Scheme 3
Table 2. Selected Interatomic Distances (Å) and Angles (deg) for
[Zn2(Zinpyr-1)(H2O)2](ClO4)2‚6H2O (6‚6H2O)a
Attempts to crystallize a 1:1 ligand:metal complex were
unsuccessful. Reactions of Zinpyr-1 in the presence of <1 equiv
of Zn2+ afforded only the 2:1 complex. Complexation of Zn2+
in the first binding site leaves a DPA arm uncoordinated, and
the resultant multiple degrees of freedom might disfavor crys-
tallization. The 2:1 complex is significantly more symmetric
and rigid, leading to more favorable crystal packing opportuni-
ties.
Bond Lengths
Bond Angles
Zn(1)-O(3)
1.942(6)
2.043(6)
2.178(7)
2.086(8)
2.072(8)
N(1)-Zn(1)-O(6)
170.0(3)
79.9(3)
97.4(2)
115.4(3)
119.7(3)
121.4(3)
171.5(3)
79.5(3)
Zn(1)-O(6)
Zn(1)-N(1)
Zn(1)-N(2)
Zn(1)-N(3)
N(1)-Zn(1)-N(2)
O(3)-Zn(1)-O(6)
N(2)-Zn(1)-N(3)
N(3)-Zn(1)-O(3)
O(3)-Zn(1)-N(2)
N(4)-Zn(2)-O(7)
N(4)-Zn(2)-N(5)
O(5)-Zn(2)-O(7)
N(5)-Zn(2)-N(6)
N(6)-Zn(2)-O(5)
O(5)-Zn(2)-N(5)
Formation of mixed complexes with partially coordinated
Zn2+ is one of the perceived shortcomings of the quinoline-
based sensors. Coordination of Zn2+ by the ligand arm of the
Zinpyr sensor leaves one open coordination site that is occupied
by a water molecule in the crystal structure. Although this open
coordination site offers the possibility to form mixed complexes,
binding at this open site should be relatively weak, because the
coordination sphere of the Zn2+ metal ion is nearly saturated.
The open site is also fairly crowded, which should hinder
coordination of macromolecular ligands.
Zn(2)-O(5)
Zn(2)-O(7)
Zn(2)-N(4)
Zn(2)-N(5)
Zn(2)-N(6)
1.954(6)
2.046(6)
2.178(7)
2.087(8)
2.086(8)
94.6(3)
124.4(3)
110.6(3)
121.5(3)
a Numbers in parentheses are estimated standard deviations in the
last digit(s). Atom labels are provided in Figure 1.
fluorescein-based molecules and only one metal complex, that
of Pb2+. The relatively few examples of fluorescein crystal
structures have been attributed to properties of the molecule
that disfavor crystallization.59
Intracellular Staining with Zinpyr-1 and Zinpyr-2. Incu-
bation of Cos-7 cells with 5 µM Zinpyr reagent produced bright
punctate staining patterns similar to observations made with
quinoline-based sensors.5,12 These observations were made with
both Zinpyr-1 and Zinpyr-2. The bright fluorescence colocalized
with the acidic compartment probe LysoTracker (Molecular
Probes, Eugene OR). The initial Zinpyr fluorescence was not
diminished by addition of TPEN, which suggests that the
concentration of free Zn2+ in these cells is lower than the
detection limit of the sensor (∼0.1 nM). The low free Zn2+ may
reflect buffering and chelation by exogenous Zinpyr. To draw
conclusions about endogenous Zn2+, further experiments will
be necessary in which the Zinpyr concentration is varied and
the observed free [Zn2+] extrapolated to zero dye. The fluores-
cence initially observed is most likely background induced by
protonation within vesicular membranes (Figure 7a). The
apparent propensity of these molecules to become trapped in
acidic compartments represents the principal challenge in the
The Zn2+ centers in [Zn2(Zinpyr-1)(H2O)2](ClO4)2‚6H2O are
trigonal bipyramidal, a relatively common coordination geom-
etry for this metal ion. Each Zn2+ ion is coordinated by the
three nitrogen atoms of one DPA arm, a phenolic oxygen atom,
and a water molecule. The tertiary amine atom and water
molecule occupy the axial positions, with the pyridine and
phenol oxygen atoms in the equatorial sites. The bond lengths
and angles are similar to those in X-ray structures of zinc
complexes of similar tripodal N3O ligands.60 The fluorescein
fragment adopts the lactoid form, which is consistent with the
lack of fluorescence of the orange crystals under UV light. Since
the complex fluoresces in solution, we conclude that crystal-
lization induces formation of the lactone.
(59) Tremayne, M.; Kariuki, B. M.; Harris, K. D. M. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 770-772.
(60) Trosch, A.; Vahrenkamp, H. Eur. J. Inorg. Chem. 1998, 827-832.

7840 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
Burdette et al.
Figure 7. (a) Fluorescence microscopy images of COS-7 cells labeled with 5 µM Zinpyr-1 for 0.5 h at 37 °C. The fluorescence is not diminished
by the addition of the high-affinity membrane-permeable heavy-metal chelator, TPEN, indicating the fluorescence is not Zn2+-induced. (b) The
bright perinuclear punctate staining increases upon the addition of the zinc ionophore Zn2+/pyrithione and can be reversed by treatment with TPEN.
monitoring the [Zn2+] change is straightforward and represents
a significant application for these sensors. The ability of these
probes to image exogenously applied Zn2+ demonstrates that
Zinpyr sensors can produce a discernible change in the
fluorescence signal even though the unmetalated/protonated
sensor has a relatively bright fluorescence. Examination of cell
types that undergo fluctuations in endogenous [Zn2+] as a result
of biochemical events would be of interest.
Conclusions
We have developed two new fluorescent Zn2+ sensors,
Zinpyr-1 (5) and Zinpyr-2 (4), utilizing fluorescein as the
reporting group. These new sensors are amenable to intracellular
studies as a result of their favorable optical properties, binding
affinity, water solubility, and their ability to be loaded passively
into cells. The primary shortcoming of the Zinpyr sensors is
Figure 8. Fluorescence intensity analysis of COS-7 cells shown in
their sensitivity to protons and the relatively modest fluorescence
Figure 7. The fluorescence intensity observed for two of the bright
enhancement upon binding of the analyte. Access to the
perinuclear punctate regions (Golgi), and one nucleus are shown.
fluorescein dialdehyde (3), the key intermediate in the synthesis
Exogenous Zn2+ was added using the zinc ionophore pyrithione 5 min
of Zinpyr-2, is significant because it can be utilized as a starting
after the initial treatment with Zinpyr-1, causing a dramatic increase
point for the design of future sensors. In addition to sensor
design, the aldehyde functionality of 3 provides a convenient
synthetic means for introducing a fluorescein molecule passively
into a variety of systems. The ability to modify synthetically
commercially available substituted fluorescein compounds such
as fluorescein amine is often limited because the compounds
are relatively unreactive, and require harsh conditions for
introduction into the desired system.
The Zinpyr dyes are the first generation of molecules in an
ongoing effort to develop Zn2+ sensors for the neurosciences.
Our current efforts are focused on preparation of sensors that
are completely quenched in the absence of Zn2+ and on
developing a series of probes with a range of binding affinities
to study free Zn2+ at the different concentrations proposed to
be present in living systems. The use and elaboration of the
synthetic methodologies developed here to prepare the dialde-
hyde 3 are key steps in accomplishing these goals. In addition
to new sensors, we are also continuing to apply the Zinpyr
sensors to investigating the roles of free Zn2+ in neurophysi-
ology.
in fluorescence in the puncta. The enhancement was reversed by
treatment with TPEN that was added at the 30 min mark. Cytosolic
and nuclear regions selected for analysis did not respond to exogenously
added Zn2+/pyrithione or TPEN.
measurement of Zn2+ within cells using Zinpyr sensors. The
nature of the fluorescence staining suggested that the sensor
localized in a specific cell organelle. Additional double-labeling
experiments with a galactosyl transferase-enhanced cyan fluo-
rescent protein fusion (GT-ECFP) that colocalizes in the medial/
trans-Golgi revealed that Zinpyr stains the Golgi or a Golgi-
associated vesicle.40
A significant enhancement of the fluorescence in the puncta
could be observed when the cells were treated with exogenous
Zn2+ (50 µM) using the zinc ionophore pyrithione (2-mercapto-
pyridine N-oxide, 20 µM, Figure 7b). Addition of TPEN (100
µM) reversed the fluorescence in the organelles to levels
observed before the treatment with Zn2+ (Figure 8). Although
the background fluorescence from the unmetalated sensor
presents a significant difficulty in measuring the free [Zn2+],

Fluorescent Sensors for Zn2+
J. Am. Chem. Soc., Vol. 123, No. 32, 2001 7841
Acknowledgment. This work was supported at MIT by seed
funds to launch new projects in the neurosciences and more
recently by a grant from the McKnight Foundation for the
Neurosciences. The NMR spectrometer at the MIT DCIF was
purchased with support from the National Science Foundation
under grant no. CHE9808061. S.C.B. and S.J.L. thank J. Du
Bois, K. J. Franz, D. A. Kopp, G. T. Gassner, and W. M. Davis
for helpful suggestions. S.C.B. thanks R. A. Binstead of
Spectrum Software Associates for assistance with SPECFIT.
B.S. was supported by the Swiss National Science Foundation
as a postdoctoral fellow. G.K.W. and R.Y.T. thank Dr. J. Llopis
for initial assistance with the GT-ECFP labeling experiments.
The award of a NIH training grant to support G.K.W. (GM19804-
02) is gratefully acknowledged.
Supporting Information Available: Experimental methods
for metal ion selectivity experiments; Figures S1 and S2 showing
the fluorescence response of Zinpyr-1 and Zinpyr-2 to various
divalent metal ions; Figures S3-S7 showing the 1H NMR spectra
of compounds 1-5; Figure S8 showing the fully labeled ORTEP
diagram; and Tables S1-S4 showing pertinent crystallographic
information for [Zn2(Zinpyr-1)(H2O)2](ClO4)2‚6H2O.
JA010059L