ZP4, an improved neuronal Zn2+ sensor of the Zinpyr family

Article abstract of DOI:10.1021/ja0287377

A second-generation fluorescent sensor for Zn²⁺ from the Zinpyr family, ZP4, has been synthesized and characterized, ZP4 (Zinpyr-4, 9-(o-carboxyphenyl)-2-chloro-5-[2-{bis(2-pyridylmethyl)-aminomethyl}- N-methylaniline]-6-hydroxy-3-xanthanone) is prepared via a convergent synthetic strategy developed from previous studies with these compounds, ZP4, like its predecessors, has excitation and emission wavelengths in the visible range (～500 nm), a dissociation constant (K_d) for Zn²⁺ of less than 1 nM and a high quantum yields (Π = ～0.4), making it well suited for biological applications, A 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 metal-binding stereochemistry of ZP4 was evaluated through the synthesis and X-ray structural characterization of [M(BPAMP)(H₂O)_n]⁺ complexes, where BPAMP is [2-{bis(2-pyridylmethyl)aminomethyl}-N-methylaniline]-phenol and M = Mn²⁺, Zn²⁺ (n = 1) or Cu²⁺ (n = 0).

Full text of DOI:10.1021/ja0287377

Published on Web 01/22/2003
ZP4, an Improved Neuronal Zn2+ Sensor of the Zinpyr Family
Shawn C. Burdette,† Christopher J. Frederickson,‡ Weiming Bu,† and
Stephen J. Lippard*,†
Contribution from the Department of Chemistry, Massachusetts Institute of Technology,
Cambridge, Massachusetts, 02139, and NeuroBioTex, Incorporated, GalVeston, Texas 77550
Received September 28, 2002 ; E-mail: lippard@lippard.mit.edu
Abstract: A second-generation fluorescent sensor for Zn2+ from the Zinpyr family, ZP4, has been
synthesized and characterized. ZP4 (Zinpyr-4, 9-(o-carboxyphenyl)-2-chloro-5-[2-{bis(2-pyridylmethyl)-
aminomethyl}-N-methylaniline]-6-hydroxy-3-xanthanone) is prepared via a convergent synthetic strategy
developed from previous studies with these compounds. ZP4, like its predecessors, has excitation and
emission wavelengths in the visible range (∼500 nm), a dissociation constant (Kd) for Zn2+ of less than 1
nM and a high quantum yields (Φ ) ∼0.4), making it well suited for biological applications. A 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 metal-binding stereochemistry of ZP4 was evaluated through the synthesis and X-ray structural
characterization of [M(BPAMP)(H2O)n]+ complexes, where BPAMP is [2-{bis(2-pyridylmethyl)aminomethyl}-
N-methylaniline]-phenol and M ) Mn2+, Zn2+ (n ) 1) or Cu2+ (n ) 0).
In addition to acute toxicity, disruption of Zn2+ homeostasis
may play a role in the pathology of several neurodegenerative
Introduction
Zinc is one of many transition metal elements that sustain
life.1 Once postulated to be readily available from a cytosolic
pool of free metal ion,2 mounting evidence suggests that Zn2+
uptake and distribution are regulated by a complex intracellular
mechanism.3-5 Disruption of Zn2+ homeostasis after ischemia,6,7
seizures,8-15 and following head injury16 induces neuronal death.
Although methods for detecting zinc show no evidence of
perikaryal Zn2+ in healthy cells,17,18 following these various
traumas, damaged neuronal cells are stained intensely for this
metal ion.16,19-21
disorders,22,23 like the formation of amyloid plaques in Alz-
heimer’s disease.24-26 The highly regulated delivery of zinc to
intracellular targets and the deleterious effects observed during
overexposure to the free metal ion suggest that healthy cells
contain no unbound Zn2+; however, free Zn2+ may play a vital
role in communication between cells. Zinc released from
glutamatergic nerve terminals plays a role in neurotrans-
mission,27-30 but the precise function of this synaptic Zn2+
remains unclear.31-33 Extracellular Zn2+ has been implicated
in the modulation of cellular processes through a zinc-sensing
receptor.34
† Massachusetts Institute of Technology.
‡ NeuroBioTex, Inc.
(1) Vallee, B. L.; Falchuk, K. H. Physiol. ReV. 1993, 73, 79-118.
(2) Williams, R. J. P.; Frau´sto da Silva, J. J. R. Coord. Chem. ReV. 2000,
200-202, 247-348.
(19) Sensi, S. L.; Canzoniero, L. M. T.; Yu, S. P.; Ying, H. S.; Koh, J. Y.;
Kerchner, G. A.; Choi, D. W. J. Neuroscience 1997, 17, 9554-9564.
(20) Budde, T.; Minta, A.; White, J. A. Neuroscience 1997, 79, 347-358.
(21) Suh, S. W.; Danscher, G.; Jensen, M. S.; Thompson, R.; Motamedi, M.;
Frederickson, C. J. Brain Res. 2000, 879, 7-12.
(3) Outten, C. E.; O’Halloran, T. V. Science 2001, 292, 2488-2492.
(4) Hitomi, Y.; Outten, C. E.; O’Halloran, T. V. J. Am. Chem Soc. 2001, 123,
8614-8615.
(22) Cuajungco, M. P.; Lees, G. J. Neurobiol. Dis. 1997, 4, 137-169.
(23) Bush, A. I. Curr. Opin. Chem. Biol. 2000, 4, 184-191.
(24) Suh, S. W.; Jensen, K. B.; Jensen, M. S.; Silva, D. S.; Kesslak, P. J.;
Danscher, G.; Frederiskson, C. J. Brain Res. 2000, 852, 274-278.
(25) Curtain, C. C.; Ali, F.; Volitakis, I.; Cherny, R. A.; Norton, R. S.;
Beyreuther, K.; Barrow, C. J.; Masters, C. L.; Bush, A. I.; Barnham, K. J.
J. Biol. Chem. 2001, 276, 20 466-20 473.
(5) Outten, C. E.; Tobin, D. A.; Penner-Hahn, J. E.; O’Halloran, T. V.
Biochemistry 2001, 40, 10 417-10 423.
(6) Koh, J. Y.; Suh, S. W.; Gwag, B. J.; He, Y. Y.; Hsu, C. Y.; Choi, D. W.
Science 1996, 272, 1013-1016.
(7) Tonder, N.; Johansen, F. F.; Frederickson, C. J.; Zimmer, J.; Diemer, N.
H. Neurosci. Lett. 1990, 109, 247-252.
(8) Choi, D. W.; Koh, J. Y. Annu. ReV. Neurosci. 1998, 21, 347-375.
(9) Canzoniero, L. M. T.; Turetsky, D. M.; Choi, D. W. J. Neurosci. 1999,
19: RC31, 1-6.
(26) Suzuki, K.; Miura, T.; Takeuchi, H. Biochem. Biophys. Res. Comm. 2001,
285, 991-996.
(27) Takeda, A. Brain Res. ReV. 2000, 34, 137-148.
(10) Sheline, C. T.; Behrens, M. M.; Choi, D. W. J. Neurosci. 2000, 20, 3139-
3146.
(28) Thompson, R. B.; Whetsall, W. O., Jr.; Maliwal, B. P.; Fierke, C. A.;
Frederickson, C. J. J. Neurosci. Methods 2000, 96, 35-45.
(29) Li, Y.; Hough, C. J.; Suh, S. W.; Sarvey, J. M.; Frederickson, C. J. J.
Neurophysiol. 2001, 86, 2597-2604.
(11) Kim, A. H.; Sheline, C. T.; Tian, M.; Higashi, T.; McMahon, R. J.; Cousins,
R. J.; Choi, D. W. Brain Res. 2000, 886, 99-107.
(12) Kim, Y.-H.; Kim, E. Y.; Gwag, B. J.; Sohn, S.; Koh, J.-Y. Neuroscience
1999, 89, 175-182.
(30) Li, Y.; Hough, C. J.; Frederickson, C. J.; Sarvey, J. M. J. Neurosci. 2001,
21, 8015-8025.
(13) Noh, K.-M.; Koh, J.-Y. J. Neurosci. 2000, 20: RC111, 1-5.
(14) Park, J. A.; Lee, J.-Y.; Sato, T.-A.; Koh, J.-Y. J. Neuroscience 2000, 20,
9096-9103.
(31) Cole, T. B.; Martyanova, A.; Palmiter, R. D. Brain Res. 2001, 891,
253-265.
(32) Lee, J.-Y.; Cole, T. B.; Palmiter, R. D.; Koh, J.-Y. J. Neurosci. 2000, 20:
RC79, 1-5.
(15) Weiss, J. H.; Sensi, S. L.; Koh, J. Y. TiPS 2000, 21, 395-401.
(16) Suh, S. W.; Chen, J. W.; Motamedi, M.; Bell, B.; Listiak, K.; Pons, N. F.;
Danscher, G.; Frederickson, C. J. Brain Res. 2000, 852, 268-273.
(17) Frederickson, C. J. Int. ReV. Neurobiol. 1989, 31, 145-238.
(18) Danscher, G. Histochemistry 1981, 71, 1-16.
(33) Lee, J.-Y.; Park, J.; Kim, Y.-H.; Kim, D. H.; Kim, C. G.; Koh, J.-Y. Exp.
Neurol. 2000, 161, 433-441.
(34) Herchfinkel, M.; Moran, A.; Grossman, N.; Sekler, I. Proc. Natl. Acad.
Sci. U.S.A. 2001, 98, 11 749-11 754.
9
1778
J. AM. CHEM. SOC. 2003, 125, 1778-1787
10.1021/ja0287377 CCC: $25.00 © 2003 American Chemical Society

ZP4, an Improved Neuronal Zn2+ Sensor
A R T I C L E S
the magnitude of the fluorescence change upon metal binding
is diminished and background fluorescence from the unmetalated
probe is high at physiological pH.39 In the second generation
ZP sensor reported here, we have addressed these shortcomings
with an improved synthetic strategy and a modified Zn2+
binding ligand.
-
Experimental Section
Materials and Methods. Chlorobenzene and dichloroethane were
distilled from CaH2 under nitrogen. Acetonitrile was distilled from CaH2
under nitrogen and dried over 3 Å molecular sieves, nitrobenzene and
DMF were dried over 3 Å molecular sieves. THF was dried on an
aluminum oxide column followed by a column of molecular sieves.
CDCl3 was dried over 3 Å molecular sieves. Di-(2-picolyl)amine (DPA)
was prepared as previously described.43 All other reagents were
purchased and used as received. Flash column chromatography was
performed with silica gel-60 (230-400 mesh), octadecyl-functionalized
silica gel (RP18), or Brockman I activated basic aluminum oxide (150
mesh). Thin layer chromatographic (TLC) analysis was performed with
Merck F254 silica gel-60, Merck RP-18 F254S, or Merck F254
aluminum oxide-60 plates and viewed by UV light, or developed with
ceric ammonium molybdate, ninhydrin or iodine stain. NMR spectra
were recorded on a Varian 500 MHz or 300 MHz spectrometer at
Figure 1. First members of the ZP family of fluorescein-based Zn2+ sensors
ZP1 and ZP2. Both sensors incorporate a DPA ligand and have fluorescence
properties amenable to biological studies.
Fluorescent sensors facilitate study of spectroscopically silent
metal ions such as Zn2+ in biological systems35-37 because
specific properties can be engineered into probes through
chemical synthesis. Moreover, fluorescence microscopy is both
a sensitive and an innocuous investigative technique. To advance
our studies of Zn2+ in neurobiology, we have designed and
reported on a new family of Zn2+-sensitive fluorescent probes
(Figure 1).38,39 ZP1 and ZP2 were the first in a series of
fluorescein-based sensors designed to induce a positive fluo-
rescence response upon complexation of Zn2+. Although these
two dyes offer several advantages over traditional Zn2+ sensors,
improvements in the both the synthetic methodology and the
physical properties of the sensor are desirable for an even more
efficient investigation of Zn2+ ion in neurochemistry.
1
ambient probe temperature, 283 K, and referenced to the internal H
and 13C solvent peaks. Infrared spectra were recorded on a BTS 135
or an Avatar 360 FTIR instrument as KBr pellets or thin films on NaCl
plates. Electrospray ionization (ESI) mass spectrometry was performed
in the MIT Department of Chemistry Instrumentation Facility (DCIF)
with the use of m-nitrobenzyl alcohol as the matrix. CAUTION:
Several of the isolated compounds below contain perchlorate ion,
which can detonate explosively and without warning. Although we
have encountered no incidents with the reported compounds, all
due precautions should be taken.
ZP1 can be prepared by a Mannich reaction in one step,40
but the conditions limit the variety of metal-binding ligands that
can be incorporated into the sensor. The preparation of ZP2 is
more versatile,38 but a low-yielding synthesis of the key
intermediate reduces the rate at which compounds can be
produced. In addition to these synthetic restrictions, both
compounds have two distinct Zn2+ coordination sites. Although
the binding affinities of the two sites for Zn2+ differ signifi-
cantly, and only the first binding event is accompanied by a
fluorescence change,38 these properties may not apply to future
sensors. Multiple metal ion binding events under physiological
conditions complicate the analysis and can lead to ambiguous
conclusions. One of the advantages of ZP sensors is the
brightness of the Zn2+-bound complex, yet the tertiary amines
responsible for the fluorescence increase are susceptible to
protonation at physiological pH. The inhibition of a photoin-
duced electron transfer (PET) pathway either by coordination
to a closed-shell metal ion or protonation of an amine is a
common strategy utilized in the design of intensity-based
2′-Carboxy-5-chloro-2,4-dihydroxybenzophenone (5). Phthalic
anhydride (3, 9.0 g, 61 mmol) and 4-chlororesorcinol (4, 8.54 g, 59.1
mmol) were combined in C6H5NO2 (150 mL) and chilled to 0 °C.
Aluminum chloride (AlCl3, 18.4 g, 138 mmol) was added slowly over
1 h, and the slurry was stirred overnight while warming to room
temperature. The reaction mixture was heated to 120 °C for 6 h, and
then diluted with 0.1 M HCl (700 mL) and hexanes (100 mL) to
precipitate a black-brown solid. The solids were collected on a frit and
dissolved in a boiling solution of methanol and water (1:1) and
decolorizing carbon was added. The solution was filtered through Celite
and the product crystallized at -10 °C. The solid was recrystallized
twice (1:1 CH3OH/water), washed with ice cold water and dried to
give the product as a brown crystalline solid (8.45 g, 48.9%). TLC Rf
1
) 0.26 (7:3 CHCl3/MeOH). H NMR (CD3OD, 500 MHz) δ 6.47 (1
H, s), 6.94 (1 H, s), 7.38 (1 H, dd, J ) 1.2, 7.2 Hz), 7.65 (1 H, td, J
) 1.5, 7.5 Hz), 7.73 (1 H, td, J ) 1.5, 7.2 Hz), 8.11 (1 H, dd, J ) 1.5,
7.5 Hz).13C NMR (CD3OD, 125 MHz) 104.91, 113.32, 115.65, 128.64,
130.70, 131.14, 131.75, 133.78, 134.76, 141.54, 161.80, 164.78, 168.69,
202.41. FTIR (KBr) 3390, 2828, 2663, 2547, 1691, 1615, 1571, 1419,
1293, 1219, 1141, 925, 772. HRMS (ESI): Calcd for MH+ 293.0218;
Found 293.0212.
sensors.41,42 Although the fluorescence intensity of the Zn2+
-
bound complex is greater than that of the protonated sensor,
2′-Chloro-5′-methylfluorescein Di-t-butyldimethylsilyl Ether (7).
2′-Carboxy-5-chloro-2,4-dihydroxybenzophenone (5, 5.00 g, 17.0 mmol)
and 2-methylresorcinol (2.12 g, 17.0 mmol) were crushed and melted
into a brown liquid at 150 °C. Fused ZnCl2 (2.33 g, 110 mmol) was
added slowly over 35 min, and the temperature was slowly increased
to 250 °C over 30 min until the material solidified. The brick red solid
was pulverized and boiled in 250 mL of 3 M HCl for 30 min. The red
solid was collected on a frit, washed thoroughly with hot water and
dried in vacuo. The crude product was filtered through a plug of silica
(35) Tsien, R. Y. Chem. Eng. News 1994, 72, 34-44.
(36) Kimura, E.; Koike, T. Chem. Soc. ReV. 1998, 27, 179-184.
(37) Burdette, S. C.; Lippard, S. J. Coord. Chem. ReV. 2001, 216-217, 333-
361.
(38) Burdette, S. C.; Walkup, G. K.; Spingler, B.; Tsien, R. Y.; Lippard, S. J.
J. Am. Chem. Soc. 2001, 123, 7831-7841.
(39) See Supporting Information in ref 38.
(40) Walkup, G. K.; Burdette, S. C.; Lippard, S. J.; Tsien, R. Y. J. Am. Chem
Soc. 2000, 122, 5644-5645.
(41) Czarnik, A. W. Acc. Chem. Res. 1994, 27, 302-308.
(42) de Silva, A. P.; Gunaratne, H. Q.; Gunnlaugsson, T.; Huxley, A. J. M.;
McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. ReV. 1997, 97, 1515-
1566.
(43) Gruenwedel, D. W. Inorg. Chem. 1968, 7, 495-501.
9
J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003 1779

A R T I C L E S
Burdette et al.
(7:3 CHCl3:MeOH) to give a red solid after solvent removal. The crude
2′-chloro-5′-methylfluorescein (6, mixture of two components with
∼10% 4′,5′-dimethylfluorescein) was combined with imidazole (3.40
g, 101 mmol) in DMF (300 mL) and stirred. To the resulting red slurry
was added tert-butyldimethylsilyl chloride (6.92 g, 45.9 mmol). The
reaction mixture was stirred for 12 h at room temperature. A portion
of the DMF was removed (∼250 mL), and the reaction mixture was
diluted with saturated brine (∼300 mL). The aqueous layer was
extracted with EtOAc, and the combined organic extracts were dried
over MgSO4 to give a brown oil after filtration and solvent removal.
The crude product was filtered through silica (7:2:1 hexanes/C6H5CH3/
EtOAc) and the solvents removed. Flash chromatography (9:1 hexanes/
EtOAc) yielded an impure brown solid (3.8 g) containing ∼10% of
the corresponding 4′,5′-dimethylfluorescein disilyl ether. A yield of 37%
was calculated based on the integration of NMR peaks. TLC Rf ) 0.41
under a hydrogen atmosphere (1 atm) for 24 h with the addition of
300 mg of Pd/C after 12 h. The reaction mixture was filtered through
Celite to give a dark yellow oil after solvent removal. Flash chroma-
tography on basic alumina with a solvent gradient (CHCl3 to 9:1 CHCl3/
MeOH) yielded a yellow oil. Additional flash chromatography on basic
alumina (99:1 CHCl3/MeOH) yielded the product as a yellow oil (100
mg, 5%). TLC Rf ) 0.41 (97:3 CHCl3/MeOH). 1H NMR (CDCl3, 500
MHz) δ 3.65 (2 H, s), 3.79 (4 H, s), 4.94 (2 H, bs), 6.59-6.64 (2 H,
m), 7.02-7.05 (2 H, m), 7.12 (2 H, td, J ) 1.0, 5.0 Hz), 7.37 (2 H, d,
J ) 7.5 Hz), 7.59 (2 H, td, J ) 2.0, 7.5 Hz), 8.53 (2H, dt, J ) 1.0, 5.0
Hz).13C NMR (CDCl3, 125 MHz) δ 57.90, 60.22, 115.45, 117.23,
122.09, 122.33, 123.49, 128.55, 131.28, 136.43, 147.12, 149.18, 159.29.
FTIR (thin film) 3396, 3320, 3209, 3009, 2921, 2805, 1615, 1590, 1494,
1433, 751. HRMS (ESI): Calcd for MH+ 305.1766; Found 305.1769.
9-(o-Carboxyphenyl)-2-chloro-5-[2-{bis(2-pyridylmethyl)amino-
methyl}-N-methylaniline]-6-hydroxy-3-xanthanone (ZP4, 14). 2′-
Chloro-5′-bromo-methylfluorescein di-tert-butyldimethylsilyl ether (8,
160 mg, 234 µmol) and pyridine (85 µL, 1.1 mmol) were combined in
7 mL of CH3CN. AgNO3 (60 mg, 349 µmol) was added to the stirring
solution causing the reaction to change from a yellow solution to a
clear solution with the formation of a white precipitate. After 15 min,
2-[bis(2-pyridylmethyl)aminomethyl]-aniline (11) in 20 mL of CH3CN
was added to the solution, and the reaction mixture was stirred for
12 h at room temperature. The crude reaction mixture was filtered
through Celite and diluted with brine (∼300 mL), EtOAc (200 mL)
and CH2Cl2 (100 mL). The combined organics were washed with brine
(35 100 mL), dried over Na2SO4, and the solvents were removed to
give TBS-protected ZP4. The crude product was combined in 15 mL
of THF with AcOH (32 µL, 538 µmol) and 1.0 M tetrabutylammonium
fluoride (TBAF in THF, 485 µL, 485 µmol) and stirred for 36 h at
room temperature. Upon addition of the TBAF, the solution immediately
changed from orange to deep red. The reaction was diluted with 100
mL of H2O, and the aqueous solution was washed with hexanes (2 ×
100 mL) then saturated with NaCl. The product was extracted into
EtOAc, and the combined organics were washed with H2O (100 mL)
and brine (3 × 100 mL), dried over Na2SO4 to give a red solid after
filtration and solvent removal. Medium-pressure flash chromatography
on RP18 silica (3:1 0.1 N HCl:CH3CN) followed by removal of the
CH3CN afforded an acidic solution of ZP4. The acidic solution was
loaded onto a second RP18 silica column packed with millipure H2O,
and the column was washed with millipure water until the elutant from
the column reached neutral pH (6-7, pH paper). The product was
washed off the column (4:1 CH3CN/H2O), and the solvents were
removed to give an orange powder (48 mg, 30%). 1H NMR (DMF-d7,
500 MHz) δ 3.78 (1 H, d, J ) 22 Hz), 3.88 (1 H, d, J ) 24.5 Hz),
3.92 (2 H, d, J ) 25.5 Hz), 4.00 (2 H, d, J ) 24 Hz), 4.59 (2 H, s),
6.59 (1 H, t, J ) 12 Hz), 6.68 (1 H, d, J ) 14.5 Hz), 6.80 (1 H, s),
7.01 (1 H, d, J ) 13.5 Hz), 7.03 (1 H, d, J ) 14.5 Hz), 7.16-7.20 (2
H, m), 7.25-7.29 (3 H, m), 7.43-7.51 (3 H, m), 7.69 (2 H, t, J )
12.5 Hz), 7.81 (1 H, t, J ) 11.5 Hz), 7.90 (1 H, t, J ) 10.5 Hz), 8.07
(1 H, d, J ) 12.5 Hz), 8.40 (2 H, d, J ) 7.0 Hz). 13C NMR (DMF-d7,
125 MHz) δ 58.57, 60.02, 105.39, 111.18, 111.74, 112.30, 113.76,
113.87, 117.03, 117.61, 123.84, 124.87, 125.40, 125.93, 128.01, 128.74,
129.33, 130.37, 131.45, 132.62, 136.75, 138.95, 148.96, 149.32, 151.91,
152.16, 153.45, 156.91, 160.04, 169.82. FTIR (KBr) 3372, 3062, 1792,
1606, 1450, 1284, 1252, 754. HRMS (ESI): Calcd for MH+ 683.2061;
Found 683.2024.
1
(4:1 hexanes/EtOAc). H NMR (CDCl3, 500 MHz) δ 0.21 (6 H, s),
0.31 (6 H, s), 1.02 (9 H, s), 1.05 (9 H, s), 2.34 (3 H, s), 6.46-6.53 (2
H, m), 6.74 (1 H, s), 6.83 (1 H, s), 7.20 (1 H, d, J ) 7.5 Hz), 7.64 (1
H, t, J ) 8.0 Hz), 7.70 (1 H, t, J ) 7.0 Hz), 8.03 (1 H, d, J ) 7.5 Hz).
13C NMR (CDCl3, 125 MHz) δ -4.17, -4.10, -4.03, -3.94, 9.58,
18.47, 18.54, 25.79, 25.85, 25.91, 83.43, 108.69, 111.62, 113.19, 116.70,
121.35, 124.21, 125.32, 125.43, 127.02, 128.95, 130.04, 135.28, 150.64,
151.07, 152.79, 153.37, 155.61, 169.42. FTIR (thin film) 2955, 2930,
2858, 1769, 1606, 1488, 1411, 1281, 1257, 1218, 1183, 1089. HRMS
(ESI): Calcd for MH+ 609.2259; Found 609.2254.
2′-Chloro-5′-bromomethylfluorescein Di-t-butyldimethylsilyl Ether
(8). 2′-Chloro-5′-methylfluorescein di-tert-butyldimethylsilyl ether (7,
3.80 g, 6.24 mmol, contains 10% 4′,5′-dimethylfluorescein disilyl ether),
1,3-dibromo-5,5-dimethylhydantoin (2.0 g, 7.0 mmol), and 1,1′-azobis-
(cyclohexanecarbonitrile) (VAZO 88, 85 mg, 0.347 mmol) were
combined in C6H5Cl (150 mL). Acetic acid (100 µL, 1.70 µmol) was
added to the stirring solution, and the reaction mixture was stirred at
40 °C for 60 h. The crude reaction mixture was washed twice with hot
water (100 mL, 80 °C), and the solvent was removed. Flash chroma-
tography on silica (7:1 hexanes/EtOAc) yielded the product as a brown
1
oil, (2.83 g, 73.3%). TLC Rf ) 0.28 (7:1 hexanes/EtOAc). H NMR
(CDCl3, 500 MHz) δ 0.28 (6 H, s), 0.30 (6 H, s), 1.04 (9 H, s), 1.05
(9 H, s), 4.83 (1 H, d, J ) 9.5 Hz), 4.84 (1 H, d, J ) 9.5 Hz), 6.56 (1
H, d, J ) 9.0 Hz), 6.63 (1 H, d, J ) 9.0 Hz), 6.92 (1 H, s), 7.22 (1 H,
d, J ) 7.5 Hz), 7.62 (1 H, td, J ) 1.0, 7.5 Hz), 7.68 (1 H, td, J ) 1.0,
7.5 Hz), 8.01 (1 H, dd, J ) 1.0, 7.5 Hz). 13C NMR (CDCl3, 125 MHz)
δ -4.30, -4.24, -4.10, -4.01, 18.30, 22.00, 22.33, 24.37, 25.63, 25.76,
33.86, 82.38, 108.68, 111.98, 113.11, 114.57, 118.29, 121.72, 124.03,
125.20, 126.71, 128.68, 128.77, 130.13, 135.30, 150.36, 150.53, 152.12,
153.35, 155.83, 169.87. FTIR (thin film) 3208, 2954, 2931, 2858, 1730,
1608, 1582, 1453, 1429, 1282, 1257, 839. HRMS (ESI): Calcd for
MH+ 687.1364; Found 687.1359.
2-[Bis(2-pyridylmethyl)aminomethyl]nitrobenzene (10). DPA (950
mg, 4.77 mmol), K2CO3 (6.50 g, 47.0 mmol), 2-nitrobenzylbromide
(9, 980 mg, 4.54 mmol), and powdered 3 Å molecular sieves (750
mg) were combined in 20 mL of CH3CN and stirred for 12 h under
Ar. The crude reaction mixture was filtered through Celite to give a
brown oil after solvent removal. Flash chromatography on basic alumina
(7:3 CH2Cl2:EtOAc) yielded the product as an orange oil (997 mg,
1
65.7%). TLC Rf ) 0.26 (7:3 CH2Cl2:EtOAc). H NMR (CDCl3, 500
MHz) δ 3.79 (4 H, s), 4.08 (2 H, s), 7.14 (2 H, td, J ) 1.0, 5.0 Hz),
7.34 (1 H, td, J ) 1.0, 7.5 Hz), 7.40 (2 H, d, J ) 7.5 Hz), 7.49 (1H,
dd, J ) 1.0, 7.5 Hz), 7.64 (2 H, td, J ) 2.0, 7.5 Hz), 7.71 (1 H, dd, J
) 0.5, 7.5 Hz), 7.77 (1 H, dd, J ) 1.0, 8.0 Hz), 8.51 (2 H, dq, J ) 1.0,
5.0 Hz). 13C NMR (CDCl3, 125 MHz) δ 56.06, 60.57, 122.33, 123.48,
124.52, 128.12, 131.57, 132.56, 134.55, 136.68, 149.10, 150.13, 158.77.
FTIR (thin film) 3064, 3009, 2925, 2830, 1589, 1526, 1433, 1362, 766,
731. HRMS (ESI): Calcd for MH+, 335.1503; Found 335.1493.
[2-{Bis(2-pyridylmethyl)aminomethyl}-N-methylaniline]phenol
(BPAMP, 15). 2-[Bis(2-pyridylmethyl)aminomethyl]-aniline (11, 2.15
g, 7.06 mmol) was dissolved in 50 mL of EtOAc and salicylaldehyde
(0.72 mL, 6.76 mmol) was added dropwise via a syringe. The reaction
mixture was stirred for 12 h under Ar, and the solvent was removed.
The resulting yellow oil was dissolved in 50 mL of dichloroethane,
combined with NaBH(OAc)3 (1.58 g, 7.45 mmol), and stirred vigorously
for 12 h at room temperature. The excess NaBH(OAc)3 was quenched
with saturated NaHCO3, and the product was extracted into CH2Cl2
2-[Bis(2-pyridylmethyl)aminomethyl]aniline (11). Pd/C (300 mg,
10% activated) and 2-[bis(2-pyridylmethyl)aminomethyl]-nitrobenzene
(10, 2.48 g, 7.42 mmol) were combined in 150 mL of MeOH and stirred
9
1780 J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003

ZP4, an Improved Neuronal Zn2+ Sensor
A R T I C L E S
(3 × 50 mL). The combined organics were washed once with water
(50 mL), dried over MgSO4, filtered, and the solvent was removed to
give a dark yellow oil. Flash chromatography on silica (CHCl3/MeOH/
i-PrNH2 195:4:1) yielded the product as a yellow solid (1.98 g, 71.2%).
TLC Rf ) 0.26 (CHCl3/MeOH/iPrNH2 195:4:1). 1H NMR (CDCl3, 300
MHz) δ 3.48 (2 H, s), 3.56 (4 H, s), 4.27 (2 H, d, J ) 7.0 Hz), 6.45-
6.55 (2 H, m), 6.69-6.73 (2 H, m), 6.84-6.93 (4 H, M), 7.00-7.04
(3 H, m), 7.09 (1 H, d, J ) 12.5 Hz), 7.28 (2 H, td, J ) 1.5, 2.5 Hz),
8.25 (2 H, dt, J ) 1.0, 8.0 Hz).13C NMR (CDCl3, 125 MHz) δ 46.46,
58.74, 60.52, 112.07, 116.91, 117.71, 119.74, 122.15, 123.40, 123.49,
125.38, 128.46, 128.71, 129.16, 130.82, 136.64, 147.69, 148.69, 156.55,
158.40. FTIR (thin film) 3294, 3042, 2818, 2717, 1592, 1455, 1237,
750. HRMS (ESI): Calcd for MH+ 411.2179; Found 411.2160.
[Zn(BPAMP)(H2O)](ClO4) (16). A portion of BPAMP (15, 25 mg,
61 µmol) was combined with NaOH (0.81 mL, 0.075 M, 61 µmol)
and zinc triflate (0.81 mL, 0.075 M, 61 µmol) in 3 mL of CH3CN.
Sodium percholrate (200 mg, 1.6 mmol) was dissolved in water (1 mL)
and added to the CH3CN solution, then the combined solutions were
filtered through Celite. Crystallization at room temperature with
concurrent evaporation of some solvent over 2 days yielded yellowish
clear blocks suitable for X-ray crystallography. The crystalline material
was filtered, powdered, and dried under vacuum at 50 °C to yield 14.2
mg of product in 39.4% yield. FTIR (KBr) 3490, 3270, 3062, 2860,
1607, 1478, 1446, 1274, 1099, 761, 622. Anal. Calcd for ZnC26H27-
Cl1N4O6: C, 52.72; H, 4.59; N, 9.46. Found: C, 52.47; H, 4.52; N,
9.28.
[Mn(BPAMP)(H2O)](ClO4) (17). BPAMP (15, 40 mg, 97 µmol)
was combined with NaOH (1.3 mL, 0.075 M, 97 µmol) and manganese
perchlorate (1.3 mL, 0.075 M, 97 µmol) in 3 mL of MeOH. The MeOH
solution was layered with water (3 mL) containing sodium percholrate
(300 mg, 2.4 mmol). Crystallization at room temperature with concur-
rent evaporation of some solvent over 2 d yielded clear yellow blocks
suitable for X-ray crystallography. The crystalline material was filtered,
powdered and dried under vacuum at 50 °C to yield 36.2 mg of dried
product in 64.1% yield. FTIR (KBr) 3469, 1605, 1477, 1444, 1270,
1109, 1094, 764, 622. Anal. Calcd for MnC26H27Cl1N4O6: C, 53.66;
H, 4.68; N, 9.63. Found: C, 53.71; H, 4.78; N, 9.75.
[Cu(BPAMP)](ClO4) (18). This compound was prepared in a
manner identical to that described above for 17 using copper perchlorate
(1.3 mL, 0.075 M, 97 µmol) and crystallized as brown needles and
green blocks. Only the green blocks were suitable for X-ray crystal-
lography. The crystalline material was filtered, powdered and dried
under vacuum at 50 °C to yield 36.2 mg of dried product in 71.3%
yield. FTIR (KBr) 3431, 3239, 3072, 2926, 1607, 1478, 1446, 1276,
1092, 758, 621. Anal. Calcd for CuC26H25Cl1N4O5: C, 54.55; H, 4.40;
N, 9.79. Found: C, 54.31; H, 4.46; N, 9.89.
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) APEX X-ray diffraction system with a graphite-monochro-
matized Mo KR radiation (λ ) 0.710 73 Å) controlled by a Pentium-
based PC running the SMART software package.44 Data were collected
at 173 K in a stream of cold N2 maintained with a KRYO-FLEX
nitrogen cryostat. Procedures for data collection and reduction have
been reported previously.45 The structures were solved by direct methods
and refined by full matrix least-squares and difference Fourier
techniques with SHELXTL.46 Empirical absorption corrections were
applied with the SADABS program,47 and structures were checked for
higher symmetry by PLATON.48 Space groups were determined by
examining systematic absences and confirmed by the successful solution
and refinement of the structures. All non-hydrogen atoms were refined
anisotropically. Hydrogen atoms were assigned idealized locations and
given isotropic thermal parameters 1.2 times the thermal parameter of
the carbon atoms to which they were attached. In the structures of 16
and 18, the perchlorate anion is disordered, and a water molecule in
both 16 and 17 is partially occupied; all were modeled and refined
accordingly. Relevant crystallographic information is contained in the
Supporting Information, and the 50% thermal ellipsoid plots are shown
in Figure 2.
General Spectroscopic Methods. Ultrol grade PIPES (piperazine-
N,N′-bis(2-ethanesulfonic acid)) from Calbiochem and KCl (99.997%)
was purchased and used as received. All solutions were passed through
0.2-µM cellulose filters before measurements. Except for the fluores-
cence 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
100 mM solution of 2,2′:6′,2′′-terpyridine in buffered solution (50 mM
PIPES, 100 mM KCl, pH 7) with aliquots of 10 mM (nominal) ZnCl2
and determining the equivalence point by monitoring the absorbance
of the resulting complex at 321 nm (ꢀ ) 35.9 × 103 M-1 cm-1). The
Zn2+ stocks were prepared from 99.999% pure ZnCl2. The purity of
the ZP probes was verified by HPLC on an analytical RP18 column
using a 0.1% aqueous TFA:CH3CN solvent gradient. The purity of the
ZP probes was verified by HPLC. ZP was introduced to aqueous
solutions by addition of a stock solution in DMSO (0.67 mM). Graphs
were manipulated and equations calculated by using Kaleidagraph 3.0.
The pH values of solutions were recorded with an Orion glass electrode
that was calibrated prior to each use. The experiments for measuring
the pH-dependent fluorescence, quantum yield, Kd, metal ion selectivity
were performed as previously described.38-40,49
Titration of ZP1, ZP4, and BPAMP with Metal Ions. Solutions
(50 mM PIPES, 100 mM KCl, pH 7) containing either 10 µM ZP1, 10
µM ZP4 or 100 µM BPAMP were titrated with 10 mM stock solutions
of the divalent metal ion under consideration (CuCl2, MnCl2 or ZnCl2).
The absorption spectrum for each trial was measured before and
following the addition of each aliquot of stock solution. Subtraction
of the initial spectrum from subsequent ones generated absorption
difference plots. The absorption changes were plotted as a function of
[M2+].
UV-Visible Spectroscopy. Absorption spectra were recorded on a
Hewlett-Packard 8453A diode array spectrophotometer under the control
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, maintained by a
circulating water bath in 1-cm path length quartz cuvettes with 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
acquired at 25 °C, maintained by a circulating water bath in 1 × 1 cm
quartz cuvettes using 3 nM slit widths and a 240 nm/min scan speed.
All spectra were corrected for emission intensity by using the
manufacturer-supplied photomultiplier curves.
Tissue Preparation and Staining Procedures. Tissue from normal
rats suffering prior seizures was used. Methods of administering
(44) SMART; 5.626 ed.; Bruker AXS, Inc.: Madison, WI, 2001.
(45) Feig, A. L.; Bautista, M. T.; Lippard, S. J. Inorg. Chem. 1996, 25, 6892-
6898.
(46) Sheldrick, G. M. SHELXL97-2: Program for the Refinement of Crystal
Structures; University of Gottingen: Germany, 1997.
(47) Sheldrick, G. M. SADABS: Area-Detector Absorption Correction; Uni-
versity of Gottingen: Germany, 1996.
(48) Spek, A. L. PLATON, A Multipurpose Crystallographic Tool; Utrecht
University: Utrecht, The Netherlands, 1998.
(49) Walkup, G. K.; Burdette, S. C.; Lippard, S. J.; Tsien, R. Y. J. Am. Chem
Soc. 2000, 122, S1-S7.
9
J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003 1781

A R T I C L E S
Burdette et al.
Figure 2. ORTEP diagrams of [Zn(BPAMP)(H2O)](ClO4)‚0.5(H2O) (16‚0.5H2O), [Mn(BPAMP)(H2O)](ClO4)‚(CH3OH)‚0.5(H2O) (17‚CH3OH‚(0.5)H2O)
and [Cu(BPAMP)(H2O)](ClO4) (18) showing 50% thermal elipsoids and selected atom labels. Solvent molecules, hydrogen atoms and perchlorates are
omitted for clarity. Selected bond lengths (Å) and angles (deg) for 16: Zn(1)-O(101) ) 1.9795(15), Zn(1)-O(102) ) 2.1134(17), Zn(1)-N(101) ) 2.3048(17),
N(101)-Zn(1)-O(102) ) 175.97(6); 17: Mn(1)-O(101) ) 2.0546(19), Mn(1)-O(102) ) 2.182(2), Mn(1)-N(101) ) 2.360(2), N(101)-Mn(1)-O(102)
) 172.86(7); 18: Cu(1)-O(101) ) 1.894(2), Cu(1)-N(101) ) 2.004(3), O(101)-Cu(1)-N(101) ) 94.65(11), N(101)-Cu(1)-N(102) ) 95.92(12).
Scheme 1
pilocarpine by intraperitoneal injection were followed,50 and the rats
were allowed to survive overnight after drug administration. After
seizure or control (no) treatment, rats were killed by an overdose of
anesthesia, decapitated, and the brains were removed rapidly and frozen
by burying on dry ice “snow” for 2-3 min. Frozen brains were then
mounted on the chuck of a closed-cabinet cryostat (Harris/Jung
Reichert), and cut (at -13 °C) at anywhere from 6 to 30 µm thickness.
Immediately after thawing onto clean glass slides, the sections were
allowed to dry at room temperature for 0.5 to 3 h, then stained and
photographed. Through all steps, care was taken to keep sections
covered in dust-free containers.
Zeiss 25× (glycerine or water immersion, 1.3). Illumination was via
200 W HBO high-pressure mercury illumination, using either a 360 or
480 nm excitation band-pass filter, 500 nm dichroic splitter and either
a long-pass or a 550 band-pass emission filter. Additional images were
made on a Nikon Diaphot with Odyssey confocal, illuminating with
488 nm Argon laser and viewing through a 500 nm long-pass filter. In
this case, sections were mounted directly on coverslips and viewed
(from below) through the coverslip with oil immersion, using Nikon
40 and 100× objectives.
Results
Staining was done by applying 60 µL of staining fluid to sections
lying flat in a plastic box. After 1 min, the excess reagent was washed
off by gentle agitation in a 0.8% saline solution. Slides were viewed
and photographed while still damp without coverslips, immediately after
rinsing. To improve contrast, selected sections were also cleared by
immersion in increasing concentrations of glycerol (in water), specif-
ically 30, 60, 90, and 100%. These were then viewed under glycerine
immersion (Zeiss 25×, variable aperture) with 1.25 or 2.0-post
magnification.
ZP4 Synthesis. Scheme 1 outlines the synthesis of the
fluorescein scaffold from commercially available starting ma-
terials. Multigram quantities of 2′-carboxy-5-chloro-2,4-dihy-
droxybenzophenone (5) can be prepared in 49% yield by
reacting 1 equiv of phthalic anhydride (3) and 4-chlororesorcinol
(4) using AlCl3 as a catalyst. Subsequent fusion of 5 with
2-methylresorcinol using ZnCl2 as the catalyst yields a mixture
of the desired 2′-chloro-5′-methylfluorescein (6), 2′,7′-dichloro-
fluorescein and 4′,5′-dimethylfluorescein. The crude reaction
mixture consists of a 4:1 ratio of the desired asymmetric product
and the symmetric byproducts. The dichlorofluorescein is
removed during workup, leaving a mixture containing ∼90%
Microscopy Methods. Images were made on a Zeiss Universal with
SPOT II cooled CCD camera, using Olympus Plan-apo 10× (0.87) or
(50) Suh, S. W.; Thompson, R. B.; Frederickson, C. J. Neuroreport 2001, 12,
1523-1525.
9
1782 J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003

ZP4, an Improved Neuronal Zn2+ Sensor
A R T I C L E S
Scheme 2
6. Silylation of the phenols on both fluorescein components with
tert-butyldimethyl silyl chloride (TBS-Cl) under standard
conditions provides 2′-chloro-5′-methylfluorescein di-tert-butyl-
dimethylsilyl ether (7) in 37% yield based on NMR integration.
Bromination of the mixure of silylated products under mild
conditions and removal of the remaining symmetric product by
flash chromatography gives 2′-chloro-5′-bromomethylfluorescein
di-tert-butyldimethylsilyl ether (8) in 73% yield (13% overall).
Scheme 2 outlines the synthesis of the ligand fragment from
the commercially available 2-nitrobenzyl bromide (9) using
conventional methods. The overall yield for 11 is only 3%
because of the low-yielding reduction of the nitro group.
The combination of fluorescein scaffold 8 and the ligand
fragment 11 in the presence of AgNO3 provides the silyl-
protected final product (Scheme 3). Upon removal of the silyl
groups with fluoride ion, pure ZP4 (14) is obtained after
chromatography on reverse phase silica. Analysis of the crude
product suggests a >50% conversion to ZP4 that is >80% pure.
The most significant loss of product occurs during chromatog-
raphy. The overall yield of ZP4 is 4% from 4-chlororesorcinol.
Fluorescence Properties of ZP4. Figure 3 depicts the pH
dependent fluorescence changes measured for ZP4. The plot of
the integrated emission intensity reveals three fluorescence-
related pKa values: a slight emission enhancement with a pKa
of 10.0, a larger increase with a pKa of 7.2, and fluorescence
quenching with a pKa of 4.0.
Figure 3. Plot of the normalized integrated emission intensity versus pH
for ZP4. The pKa at 10.0 correspond to the aliphatic nitrogen and the one
at 7.2 to the aniline nitrogen. At low pH fluorescein adopts a nonfluorescent
isomer corresponding to the decrease in emission intensity. The quantum
yield of the sensor at pH 7 is 0.06.
Upon the addition of 25 µM Zn2+, the quantum yield increases
to 0.34 and the emission maximum shifts to 515 nm. The
addition of Zn2+ is also accompanied by a shift in the excitation
maximum from 506 nm (ꢀ ) 61.0 × 103) to 495 nm (ꢀ ) 66.7
× 103). The brightness (ꢀ × Φ) of the Zn2+-bound sensor is
(22.7 × 103).
The fluorescence response of ZP4 is Zn2+- and Cd2+
-
selective. Figure 4 shows the fluorescence response of ZP4 in
the presence of various divalent metal ions. Even high concen-
trations (2 mM) of Ca2+ and Mg2+ produce no appreciable
change in the fluorescence emission. Fluorescence enhancement
of ZP4 by Zn2+ occurs in the presence of these two metal alkali
earth metal ions. Other first row transition metals including
Cu2+, Ni2+, Co2+, Fe2+, and Mn2+ produce no discernible
change in the emission intensity, but only the sample containing
Mn2+ affords a fluorescence response upon the subsequent
In buffered solution (50 mM PIPES, 100 mM KCl, pH 7)
using EDTA to scavenge adventitious metal ions, ZP4 has a
quantum yield of 0.06 and an emission maximum at 521 nm.
addition of Zn2+
.
Scheme 3
9
J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003 1783

A R T I C L E S
Burdette et al.
Figure 4. Fluorescence response of ZP4 to various metal ions. Bars
represent the final integrated fluorescence response (Ff) over the initial
integrated emission (Fi). Initial spectra were acquired in 100 mM KCl, 50
mM PIPES, 10 µM EDTA pH 7.00 at 25 °C. Excitation was provided at
500 nm, and the emission was integrated between 503 and 635 nm. Aliquots
of concentrated stock solutions (10 mM) of each metal ion were added to
the solution to provide 50 µM total metal ion. Zn2+ solution was added to
the solution containing the metal ion similarly.
Figure 5. Fluorescence emission response of ZP4 to buffered Zn2+
solutions. Spectra were acquired in 100 mM KCl, 50 mM PIPES, pH 7.00
at 25 °C. Excitation was provided at 500 nm with 3 nm slit widths. Emission
data were corrected for the response of the detector, using the manufacturer
supplied curve, and the emission data points at 500 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 mM Zn2+) additional ZnCl2 was added to provide ∼25 µM free Zn2+
.
Inset: fluorescence response obtained by integrating the emission spectra
between 503 and 635 nm, subtracting the baseline (0 Zn2+) spectrum and
normalizing to the full scale response (25 µM free Zn2+).
The binding affinity of ZP4 was characterized by using the
dual-metal single-ligand buffer system described previously.38,40
Varying the total Zn2+ concentration between 0 and 1 mM while
maintaining constant concentrations of Ca2+ (2 mM) and EDTA
(1 mM) provides buffered free Zn2+ between 0.17 and 25 nM.
The quantum yield of ZP4 increases 5.7-fold upon the addition
of Zn2+; similarly, the integrated emission intensity increases
∼5-fold during the binding titration. Figure 5 shows a repre-
sentative titration of ZP4 and the fitting of the integrated
fluorescence emission. The titration was performed four times
using different Ca2+/Zn2+/EDTA buffers. Analysis of the
response indicates that the [Zn(ZP4)] complex has an apparent
Kd of 0.65 ( 0.10 nM.
Scheme 4
Structural Features of ZP4. Attempts to crystallize the Zn2+
complex of ZP4 were unsuccessful, so a truncated version that
includes the metal ion binding fragment, BPAMP, was prepared
from 11 and salicylaldehyde (Scheme 4) in 71% yield. The Zn2+,
Mn2+ and Cu2+ complexes of the BPAMP ligand were crystal-
lized from aqueous MeOH or CH3CN in the presence of excess
NaClO4. All three complexes are monomers containing a single
ligand and metal ion (Figure 2). The Mn and Zn compounds
are isomorphous with the pentavalent ligand and a water
molecule forming a distorted octahedral complex. The Cu
complex lacks the water molecule, adopting distorted trigonal
bipyramidal geometry.
Titrations of ZP4 and BPAMP in buffered solution with Zn2+
show 1:1 binding by optical spectroscopy. Titration of ZP1
exhibits 2:1 ZP1:Zn2+ binding accompanied by changes in the
absorption spectrum similar to those observed for ZP4. The
titration of ZP1 and ZP4 with Cu2+ and Mn2+ also reveals 2:1
and 1:1 binding, respectively.
neurons injured by prior excitotoxic insult (seizures), with
negligible labeling of zinc-containing vesicles (Figure 6a).
Control slices show no detectable free zinc except in presynaptic
vesicles that are only stained by TSQ.
Discussion
Synthesis. The synthetic approaches to both ZP1 and ZP2
result in a symmetric ligand substitution pattern at the 4′ and 5′
positions of the xanthenone ring (Figure 1).38 Computer model-
ing suggested that binding a single Zn2+ ion between ligands
emanating from these two positions would be geometrically and
entropically disfavored. One possible strategy for making a
sensor with a single Zn2+-binding site is to incorporate the metal-
binding moiety on the bottom (phthalate) ring. Synthesis of
bottom ring modified fluorescein derivatives from resorcinol
and phthalic anhydride derivatives inevitably requires a difficult
separation of structural isomers, however. Fluorescein amine
is a commonly employed, commercially available bottom ring
derivative; however, it is not an attractive starting material
because of its high cost and the relative inertness of the amino
group to chemical transformation. In addition to expense and
reactivity, bottom ring synthetic strategies often require low
Histological Staining Results. In contrast to 6-methoxy-(8-
p-toluenesulfonamido)quinoline (TSQ), which stains all of the
zinc-containing terminal fields in hippocampal slices (Figure
6b), slices treated with ZP4 only show intense staining of the
9
1784 J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003

ZP4, an Improved Neuronal Zn2+ Sensor
A R T I C L E S
Figure 6. Hilus of dentate gyrus stained with (a) ZP4 and (b) TSQ. The brain tissue is taken from a rat following seizure activity. The zinc-positive neurons
are more distinct when labeled with ZP4 because the neuropil are not stained. The background from Zn2+-containing vesicles that are stained by TSQ makes
identification of damaged cells more difficult.
yielding, multistep processes that subject the fluorescein to
rigorous reaction conditions and difficult purification after each
manipulation.51 Alternatively, we envisioned adapting the
method used to prepare ZP2 to access unsymmetrical fluorescein
derivatives with a single zinc-binding moiety on the xanthenone
ring.
product could be removed by chromatography. Although the
overall yield of 2′-chloro-5′-bromomethyl-fluorescein di-tert-
butyldimethylsilyl ether (8) is moderately low, multigram
quantities of this key fluorescein starting material can be
obtained in several days.
A pre-assembled fluorescein starting material is an advantage
for the convergent synthesis of target molecules. The primary
significance of the synthesis outlined in Scheme 3 is the ability
to modify the Zn2+-binding ligand easily and logically to prepare
future sensor candidates for screening. In addition to 2-nitroben-
zyl bromide (9), a number of commercially available starting
An unsymmetrical fluorescein-based Ca2+ sensor was the
inspiration for our synthetic strategy.52 In the conventional
synthesis of fluoresceins, 2 equiv of a resorcinol are allowed
to react with phthalic anhydride under rigorous conditions to
give the product. Two molecules of resorcinol condense to form
the xanthenone skeleton. To obtain a fluorescein with a single
metal-binding arm, a strategy that employs two different
resorcinols is required. Dihydroxybenzophenone derivatives such
as 2′-carboxy-5-chloro-2,4-dihydroxybenzophenone (5) are in-
termediates in the formation of fluorescein molecules. When
prepared independently from 4-chlororesorcinol (4) and phthalic
anhydride (3), 5 reacts with 2-methyl resorcinol to give an
asymmetric fluorescein (Scheme 1). Analysis of the reaction
product reveals minor amounts of 2′,7′-dichlorofluorescein and
4′,5′-dimethylfluorescein. The presence of these symmetric
fluoresceins indicates that the formation of the benzophenone
is to some extent reversible. A variety of methods and catalysts
were screened in an attempt to eliminate the formation of these
symmetric products, but the highest conversion to the desired
monofunctional fluorescein was 80% under ZnCl2 fusion
conditions.53
Separation of the 2′-chloro-5′-methylfluorescein (6) from the
symmetric products proved to be both difficult and unnecessary.
Installation of tert-butyldimethyl silyl (TBS) ethers on the
phenols under standard conditions provides a convenient method
to purify and manipulate the desired product. Unprotected
fluoresceins are notoriously difficult to handle because of their
limited solubility in conventional organic solvents and their
propensity to adopt different isomeric forms. Silyl ethers were
selected as the protecting groups for their stability to nucleo-
philes by comparison to esters that were utilized previously.38
The similar polarity of silylated 4′,5′-dimethylfluorescein and
2′-chloro-5′-methylfluorescein di-tert-butyldimethylsilyl ether
(7) prohibits separation, but following bromination the undesired
materials could serve as useful building blocks for the Zn2+
-
binding fragment. Systematic variation of the ligand building
block as well as the metal-binding fragment will permit
correlation of structural features with fluorescence properties
and the binding affinity of the probes. Another significant
improvement in this synthetic method is the ability to access
sensor molecules from the bromomethyl fluorescein derivative.
Formation of a nitro-oxy compound in situ by combining 8 with
AgNO3 activates the benzylic position toward nucleophilic
substitution. Displacement of the bromide facilitated by AgNO3
activation avoids the low yielding oxidation chemistry required
to prepare the dialdehyde precursor of ZP2.38
Fluorescence Properties of ZP4. Under simulated physi-
ological conditions ZP1 and ZP2 exhibit relatively intense
fluorescence in the absence of Zn2+ (Φ ) ∼0.3). This
background emission decreases the sensitivity of these probes
and interferes with the measurement of Zn2+ concentration
gradients. The pKa values of the amine responsible for the
fluorescence change of ZP1 and ZP2 are 8.4 and 9.4, respec-
tively. When the amines are completely deprotonated at high
pH, both probes emit a dim fluorescent signal, and exhibit a
dramatic fluorescence enhancement upon the addition of Zn2+ 38
.
Because protonation diminishes the magnitude of the fluores-
cence change, adjusting the pKa of the nitrogen responsible for
the PET quenching attenuates the fluorescence properties.
Aromatic nitrogen atoms typically have lower pKa values (pKa
) 4-6) than their aliphatic counterparts. In addition to having
a low affinity for protons, aniline nitrogen atoms seldom form
strong bonding interactions with metal ions. To avoid proton
interference while retaining an appreciable affinity for Zn2+
,
(51) Hirano, T.; Kikuchi, K.; Urano, Y.; Higuchi, T.; Nagano, T. J. Am. Chem.
Soc. 2000, 122, 12399-12400.
ZP4 integrates an aniline nitrogen into a chelating ligand
containing a di(2-picolyl)amine (DPA) moiety. Maintaining the
delicate balance between Zn2+-binding affinity and proton
(52) Smith, G. A.; Metcalfe, J. C.; Clarke, S. D. J. Chem. Soc., Perkin Trans.
2 1993, 1195-1204.
(53) Hilderbrand, S. H.; Lippard, S. J. 2000, unpublished results.
9
J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003 1785

A R T I C L E S
Burdette et al.
Scheme 5
insensitivity represents the most significant challenge in design-
ing PET-based Zn2+ sensors.
the absorbance spectrum is characteristic of binding by the
phenolic oxygen. DPA has a high affinity for Zn2+ (Kd ) 70
nM),54 and the sub-nM Kd requires a contribution from this
ligand fragment. The increase in quantum yield indicated
coordination of the aniline nitrogen that is primarily responsible
for PET quenching at physiological pH. Scheme 5 shows our
interpretation of the solution chemistry of ZP4. In the absence
of Zn2+, the aliphatic amine is protonated, giving the sensor an
overall -1 charge. Upon binding the chelating moiety wraps
around the Zn2+ ion, with the final coordination site being
occupied by a water molecule.
In the pH dependent fluorescence changes measured for ZP4,
the pKa at 10.0 corresponds to the protonation of the tertiary
amine. Although the aniline nitrogen would be predicted to
dominate PET, the aliphatic amine contributes to the quenching
at basic pH. The high pKa value of 10.0, compared to the value
of ∼9 for the previous ZP sensors, indicates that the environment
of the aliphatic amine in ZP4 promotes binding of a proton.
The pKa of 7.2 measured for the aniline nitrogen provides
additional evidence that advantageous H-bonding by the chelat-
ing ligand facilitates protonation. The pKa associated with
fluorescence quenching at low pH is consistent with the
formation of a nonfluorescent fluorescein isomer that was
observed for ZP sensors previously.38 Lowering of the aniline
pKa will be addressed in future sensor design. Incorporation of
electron withdrawing groups on the aniline ring, particularly
the para position, should dramatically influence the pKa of the
aromatic nitrogen. The new convergent synthesis should facili-
tate easy access to related compounds.
One of the most interesting aspects of ZP4 is the difference
in the fluorescence behavior toward transition metal ions
compared to properties of ZP1 and ZP2. This behavior indicates
some difference in either the metal ion coordination or the
electronic structure of the two generations of ZP metal
complexes. Transition metal ions commonly quench the fluo-
rescence of sensors.42 Unlike the fluorescence of ZP1 and ZP2,
which is quenched significantly by open shell metal ions, the
emission of ZP4 is unchanged upon binding Mn2+, Fe2+, Co2+
,
Ni2+, and Cu2+. Because subsequent addition of Zn2+ can only
displace the weakly binding Mn2+ ion, it appears that the other
transition metal ions bind tightly to the receptor of all three
sensors and compete for the binding of Zn2+. Examination of
The lower quantum yield (0.06) of the unmetalated sensor at
pH 7, a value ∼5-fold less than that observed for the previous
generation ZP sensors, indicates that the aniline nitrogen more
efficiently quenches fluorescence. The relatively low background
emission from the free probe makes ZP4 a significant improve-
ment over the first ZP sensors. Correlation of the Zn2+-induced
fluorescence enhancement with the pKa measurements provides
evidence that the aniline nitrogen also coordinates to the metal
ion. Although the brightness of the Zn2+-complexed sensor is
sufficient for imaging applications, the quantum yield of the
complex is less than half the value measured for ZP1 and ZP2.
the optical spectra from the titration of ZP1 and ZP4 with Zn2+
,
Cu2+, and Mn2+ supports this conclusion, because the only
apparent difference in the behavior of these two sensors is the
stoichiometry of metal ion binding.
The crystal structure of the zinc complex of ZP1 contains
two 5-coordinate Zn2+ ions each coordinated by the three
nitrogen atoms of a DPA ligand, a phenolic oxygen and a water
molecule.38 The zinc coordination is identical to that of zinc55
and other transition metal complexes56,57 with truncated versions
of this N3O ligand. Since attempts to crystallize the Zn2+
complex of ZP4 were unsuccessful, BPAMP was used for
structural studies. The BPAMP ligand shows Zn2+-binding
behavior similar to ZP4 in buffered solution. The crystal
structures of the Zn2+, Mn2+, and Cu2+ complexes (Figure 2)
display coordination consistent with the spectroscopic measure-
ments of ZP4, with all 5 donors from the chelating ligand bound
to the metal ions. The only difference between the 3 structures
is that the Cu2+ complex is distorted trigonal bipyramidal rather
Possible explanations for the lower quantum yield of the Zn2+
-
complex of ZP4 include the existence of a second unidentified
quenching mechanism, or inefficient disruption of PET from
the aniline nitrogen by Zn2+-binding. Future investigation of
structurally related ZP derivatives may help to understand the
basis for this fluorescence behavior.
Structural Aspects of ZP4 and Behavior Toward Transi-
tion Metal Ions. The sub-nM Zn2+-binding affinity of ZP4 is
nearly identical to that of ZP1 and ZP2. In addition to Zn2+
-
binding affinity, the hypsochromatic shift in the absorption
wavelength of ZP4 upon metal ion binding suggests that the
phenolic oxygen coordinates to the Zn2+ ion in solution in the
same manner as the previous two ZP sensors.38 The similarity
of the binding strength, the comparable absorbance and emission
properties, and the use of an identical metal-binding ligand
fragment indicate that the coordination of Zn2+ is probably
similar for all three ZP complexes. The Zn2+-induced shift in
(54) Anderegg, G.; Hubmann, E.; Podder, N. G.; Wenk, F. HelV. Chim. Acta
1977, 60, 123-140.
(55) Trosch, A.; Vahrenkamp, H. Eur. J. Inorg. Chem. 1998, 827-832.
(56) Uma, R.; Viswanathan, R.; Palaniandavar M.; Lakshminarayanan, M. J.
Chem. Soc., Dalton Trans. 1994, 1219-1226.
(57) Viswanathan, R.; Palaniandavar, M.; Balasubramanian, T.; Muthiah, T. P.
Inorg. Chem. 1998, 37, 2943-2951.
9
1786 J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003

ZP4, an Improved Neuronal Zn2+ Sensor
A R T I C L E S
than distorted octahedral because it lacks the coordinated water
molecule found in the Zn2+ and Mn2+ structures. From the
structural and spectroscopic experiments performed thus far, it
remains unclear why the fluorescence response of ZP4 to
transition metals is different from that of the first generation of
ZP sensors. Investigation of the electronic structure of ZP4
complexes by DFT as well as experiments on the fluorescence
properties is planned to investigate this behavior further.
These metal-binding properties have important implications
in biological studies where other metal ions may interfere with
Zn2+ measurements; however, there may be no such pools of
available transition metals in healthy cells. The reducing
environment within cells precludes the presence of Cu2+, which
is readily reduced to Cu+ and disproportionates to Cu metal
and Cu2+ in aqueous solution. Current evidence argues against
the presence of free copper in cells and reveals that chaperone
proteins transport Cu to its cellular targets.58 Free Fe2+ readily
neurons and the cells damaged during seizures by the release
of excitoxic Zn2+ are cut open by the microtome blade during
tissue slicing, permitting ZP4 to enter the cells. Only the
damaged neurons display intense fluorescence that indicates the
presence of Zn2+ when stained. Because membrane permeable
probes such as ZP1 and TSQ image both zinc-containing
vesicles and damaged neurons, discrimination between the two
is quite difficult. To date, ZP4 provides the most well resolved,
detailed images of zinc-damaged neurons using fluorescence
microscopy.64
Conclusions
We have designed, synthesized and characterized ZP4, a
second-generation member of the ZP family of Zn2+-selective
sensors. ZP4 is prepared by a convergent synthesis where the
fluorescein scaffold and the Zn2+-binding fragment are prepared
independently then linked to yield the final sensor. This new
synthetic method, along with a high yielding synthesis of the
key fluorescein intermediate, allows the efficient preparation
of additional sensor candidates for screening. The key feature
of ZP4 is the incorporation of an aniline nitrogen atom into the
Zn2+-binding ligand. The aniline nitrogen responsible for PET
quenching of the unmetalated fluorophore has a lower pKa than
the aliphatic nitrogen of ZP1 and ZP2, making it is less sensitive
to protonation under physiological conditions. Because the
background fluorescence from free probe is diminished, ZP4 is
more sensitive than previous ZP sensors.
oxidizes in aqueous solution to insoluble Fe3+ 59
drastically
,
reducing the presence of free metal ion. Such Fe species would
serve as a source of oxidative stress to cells,60,61 so sequestration
of Fe by macromolecules within cells seems extremely probable.
Because most “free” transition metal ions seem to have a
deleterious effect on living organisms,62 numerous unidentified
metalloregulatory proteins may exist, even for trace metals such
as Co2+. Despite the limited probability that other metal ions
will interfere with biological assays, the ability to bind Zn2+
selectively in the presence of other transition metal ions remains
an important goal in sensor design.
Acknowledgment. This work was supported at MIT to
launch new projects in the neurosciences and by grants from
the National Institute of General Medical Sciences (GM65519)
and 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.
Imaging Damaged Neurons with ZP4. Because synaptic
vesicles remain intact throughout slice preparation,63 the inability
of ZP4 to stain zinc-containing presynaptic terminals indicates
that this probe cannot penetrate cellular membranes. In contrast
to ZP4, membrane permeable probes such as TSQ and ZP1 stain
these vesicles intensely. The impermeability of ZP4 may arise
from either its charge at neutral pH, a dipole moment, or an
inability to adopt a permeable lactone isomer.64 Both the healthy
Supporting Information Available: Figure S1 showing the
fully labeled ORTEP diagrams of 16-18 and Tables S1-S12
showing pertinent crystallographic information for the 3 BPAMP
complexes. This material is available free of charge via the
Internet at http://pubs.acs.org.
(58) Rae, T. D.; Schmidt, P. J.; Pufahl, R. A.; Culotta, V. C.; O’Halloran, T. V.
Science 1999, 284, 805-808.
(59) Crichton, R. R.; Pierre, J.-L. Biometals 2001, 14, 99-112.
(60) Joshi, J. G. Ferritin: Intracellular Regulator of Metal AVailability; Connor,
J. R., Ed.; Plenum Press: New York, 1997; pp 131-148.
(61) Glinka, Y.; Gassen, M.; Youdim, M. B. H. Iron and Neurotransmitter
Function in the Brain; Connor, J. R., Ed.; Plenum Press: New York, 1997;
pp 1-22.
JA0287377
(62) Connor, J. R. Metals and OxidatiVe Damage in Neurological Disorders;
Plenum Press: New York, 1997.
(64) Frederickson, C. J.; Burdette, S. C.; Frederickson, C. J.; Sensi, S. L.; Weiss,
J. H.; Balaji, R. V.; Bedell, E.; Prough, D. S.; Lippard, S. J., to be submitted
for publication.
(63) Mui, B. L.; Cullis, P. R.; Pritchard, P. H.; Madden, T. D. J. Biol. Chem.
1994, 269, 7364-7370.
9
J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003 1787