Direct nitric oxide detection in aqueous solution by copper(II) fluorescein complexes

Article abstract of DOI:10.1021/ja064955e

A series of FL_n (n = 1-5) ligands, where FL_n is a fluorescein modified with a functionalized 8-aminoquinoline group as a copper-binding moiety, were synthesized, and the chemical and photophysical properties of the free ligands and their copper complexes were investigated. UV-visible spectroscopy revealed a 1:1 binding stoichiometry for the Cu(II) complexes of FL₁, FL₃, and FL₅ in pH 7.0 buffered aqueous solutions. The reactions of FL₂ or FL₄ with CuCl₂, however, appear to produce a mixture of 1:1 and 1:2 complexes, as suggested by Job's plots. These binding modes were modeled by the synthesis and X-ray crystal structure determination of Cu(II) complexes of 2-[(quinolin-8-ylamino)methyl]phenol (modL), employed as a surrogate of the FL_n ligand family. Two kinds of crystals, [Cu(modL) ₂](BF₄)₂ and [Cu₂(modL′) ₂-(CH₃OH)](BF₄)₂ (modL′ = 2-[(quinolin-8-ylamino)methyl]phenolate), were obtained. The structures suggest that one oxygen and two nitrogen atoms of the FL_n ligands most likely bind to Cu(II). Introduction of nitric oxide (NO) to pH 7.0 buffered aqueous solutions of Cu(FL_n) (1 μM CuCl₂ and 1 μM FL _n) at 37 °C induces an increase in fluorescence. The fluorescence response of Cu(FL_n) to NO is direct and specific, which is a significant improvement over commercially available small molecule-based probes that are capable of detecting NO only indirectly. The NO-triggered fluorescence increase of Cu(FL₅) occurs by reduction of Cu(II) to Cu(I) with concomitant dissociation of the N-nitrosated fluorophore ligand from copper. Spectroscopic and product analyses of the reaction of the FL₅ copper complex with NO indicated that the N-nitrosated fluorescein ligand (FL ₅-NO) is the species responsible for fluorescence turn-on. Density functional theory (DFT) calculations of FL₅ versus FL₅-NO reveal how N-nitrosation of the fluorophore ligand brings about the fluorescence increase. The copper-based probes described in the present work form the basis for real-time detection of nitric oxide production in living cells.

Full text of DOI:10.1021/ja064955e

Published on Web 10/12/2006
Direct Nitric Oxide Detection in Aqueous Solution by
Copper(II) Fluorescein Complexes
Mi Hee Lim,† Brian A. Wong,† William H. Pitcock, Jr.,‡ Deepa Mokshagundam,†
Mu-Hyun Baik,*,‡ and Stephen J. Lippard*,†
Contribution from the Department of Chemistry, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139, and the Department of Chemistry and School of Informatics,
Indiana UniVersity, Bloomington, Indiana 47405
Received July 13, 2006; E-mail: mbaik@indiana.edu; lippard@mit.edu
Abstract: A series of FLn (n ) 1-5) ligands, where FLn is a fluorescein modified with a functionalized
8-aminoquinoline group as a copper-binding moiety, were synthesized, and the chemical and photophysical
properties of the free ligands and their copper complexes were investigated. UV-visible spectroscopy
revealed a 1:1 binding stoichiometry for the Cu(II) complexes of FL1, FL3, and FL5 in pH 7.0 buffered aqueous
solutions. The reactions of FL2 or FL4 with CuCl2, however, appear to produce a mixture of 1:1 and 1:2
complexes, as suggested by Job’s plots. These binding modes were modeled by the synthesis and X-ray
crystal structure determination of Cu(II) complexes of 2-[(quinolin-8-ylamino)methyl]phenol (modL), employed
as a surrogate of the FLn ligand family. Two kinds of crystals, [Cu(modL)2](BF4)2 and [Cu2(modL′)2-
(CH3OH)](BF4)2 (modL′ ) 2-[(quinolin-8-ylamino)methyl]phenolate), were obtained. The structures suggest
that one oxygen and two nitrogen atoms of the FLn ligands most likely bind to Cu(II). Introduction of nitric
oxide (NO) to pH 7.0 buffered aqueous solutions of Cu(FLn) (1 µM CuCl2 and 1 µM FLn) at 37 °C induces
an increase in fluorescence. The fluorescence response of Cu(FLn) to NO is direct and specific, which is
a significant improvement over commercially available small molecule-based probes that are capable of
detecting NO only indirectly. The NO-triggered fluorescence increase of Cu(FL5) occurs by reduction of
Cu(II) to Cu(I) with concomitant dissociation of the N-nitrosated fluorophore ligand from copper. Spectroscopic
and product analyses of the reaction of the FL5 copper complex with NO indicated that the N-nitrosated
fluorescein ligand (FL5-NO) is the species responsible for fluorescence turn-on. Density functional theory
(DFT) calculations of FL5 versus FL5-NO reveal how N-nitrosation of the fluorophore ligand brings about
the fluorescence increase. The copper-based probes described in the present work form the basis for
real-time detection of nitric oxide production in living cells.
Introduction
rescent sensors for NO have the potential to provide a practical
method for visualizing its presence and movement in vitro and
in vivo.8,9
Nitric oxide (NO) is produced in a biological context by nitric
oxide synthases.1-6 The properties of NO allow for its varied
involvement in physiological and pathophysiological path-
ways,3,4,6,7 but the details of how it performs its biological roles
are not fully understood. The development of a method capable
of detecting NO in biology has been an intriguing challenge
for chemists, biologists, and engineers. Small-molecule fluo-
A few basic requirements are necessary for the design of
biologically useful fluorescent NO probes,8,9 including water-
solubility, cell-membrane permeability, and visible or near-IR
excitation and emission wavelengths. Most importantly, these
sensors must have the capability for direct, specific, and rapid
detection of NO. In addition, sensors that give fluorescence
enhancement by reaction with NO are preferred for imaging
NO in biological systems over those that respond to NO with
a fluorescence decrease.
† Massachusetts Institute of Technology.
‡ Indiana University.
(1) Murad, F. Angew. Chem., Int. Ed. Engl. 1999, 38, 1856-1868.
(2) Leone, A. M.; Palmer, R. M. J.; Knowles, R. G.; Francis, P. L.; Ashton,
D. S.; Moncada, S. J. Biol. Chem. 1991, 266, 23790-23795.
(3) Moncada, S.; Palmer, R. M. J.; Higgs, E. A. Pharmacol. ReV. 1991, 43,
109-142.
The commonly used, current generation of NO probes is
based on organic molecules.8,10 Although these probes satisfy
the basic requirements for bioimaging NO, they have the critical
limitation that their fluorescence response is not driven by NO,
but by an oxidized NO species. This requirement means that
(4) Conner, E. M.; Grisham, M. B. Methods Enzymol. 1995, 7, 3-13.
(5) Marletta, M. A.; Hurshman, A. R.; Rusche, K. M. Curr. Opin. Chem. Biol.
1998, 2, 656-663.
(6) Ricciardolo, F. L. M.; Sterk, P. J.; Gaston, B.; Folkerts, G. Physiol. ReV.
2004, 84, 731-765.
(9) Hilderbrand, S. A.; Lim, M. H.; Lippard, S. J. In Topics in Fluorescence
(7) Packer, L. Methods in Enzymology, Nitric Oxide. Part B, Physiological
and Pathological Processes; Academic Press: San Diego, CA, 1996; Vol.
269.
Spectroscopy; Geddes, C. D., Lakowicz, J. R., Eds.; Springer: 2005; pp
163-188.
(10) Gomes, A.; Fernandes, E.; Lima, J. L. F. C. J. Fluoresc. 2006, 16, 119-
(8) Nagano, T.; Yoshimura, T. Chem. ReV. 2002, 102, 1235-1269
139.
9
14364
J. AM. CHEM. SOC. 2006, 128, 14364-14373
10.1021/ja064955e CCC: $33.50 © 2006 American Chemical Society

Nitric Oxide Detection in Aqueous Solution
A R T I C L E S
Scheme 1
Figure 1. Chemical structures of FLn (n ) 1-5).
most available sensors cannot provide direct, real-time imaging
of NO. Since NO is relatively stable under physiological
conditions,11 it can be directly monitored. To achieve direct
detection of nitric oxide, new sensors are desired that allow for
better exploration of NO-induced signaling in biology.
Our laboratory has designed NO sensors employing transition-
metal complexes. Interaction of these complexes with NO offers
the promise of investigating nitric oxide itself in biological
media. In the past few years, we have constructed several
metal-fluorophore scaffolds as potential NO sensors, including
those based on cobalt,12-16 copper,17-19 iron,15 rhodium,20 and
ruthenium21 chemistry. Our general strategy for NO sensing is
to coordinate a fluorophore ligand to a metal ion to quench the
fluorescence, which is restored by the interaction of NO with
the metal center, sometimes with release of the fluorophore and
concomitant emission turn-on (Scheme 1a).9 Although all of
the metal complexes reported by us can directly interact with
NO, most of the metal-fluorophore platforms are either not
soluble or are unstable in aqueous solutions. In some cases,
water can outcompete the fluorophore for coordination to the
metal center.
Recently, we have been exploring Cu(II)-based probes for
NO detection. The strategy behind these Cu(II)-based sensors
is that formation of a diamagnetic Cu(I) species via NO-triggered
reduction alleviates the fluorescence quenching associated with
coordination to a paramagnetic Cu(II) center (Scheme 1b).17,18
Utilizing this strategy, we constructed Cu(II) complexes con-
taining dansyl groups as the fluorophore, which are able to detect
NO in pH 9.0 buffered solutions.17 These Cu(II) dansyl
compounds were unable to sense NO at a physiologically more
relevant pH, however. Recently, a Cu(II) cyclam complex was
reported that demonstrated the potential of Cu(II) systems as
suitable NO indicators by a strategy different from that used
for the Cu(II) dansyl compounds.22 This complex exhibited
fluorescence emission turn-on via NO-induced nitrosation of
the fluorophore, which dissociated from the reduced Cu(I) in
the buffered methanolic solution (Scheme 1c). Although these
various Cu(II)-based systems are not satisfactory as biological
NO sensors, the studies formed the basis of a successful
approach for developing metal-based sensors to image NO
production in live cells.
To achieve NO sensing in a physiological context, five
fluorescein derivatives (Figure 1) were prepared for binding
Cu(II). These ligands are based on molecules previously
described for zinc-sensing bioapplications by our laboratory.23,24
The copper(II) fluorescein complexes of these ligands react
directly with NO, evoking a concomitant emission increase in
pH 7.0 buffered solutions. Moreover, the detection of NO by
these Cu(II) probes is specific over other reactive nitrogen or
oxygen species, including HNO, ONOO-, NO2-, NO3-, H2O2,
O2-, and ClO-. The synthesis, NO-sensing ability, and mech-
anism of fluorescence turn-on of the copper-fluorescein
complexes are described in the present work. Density functional
theory (DFT) calculations are also presented that address the
mechanism of fluorescence turn-on. The imaging of NO
production in macrophage and neuroblastoma cells by Cu(FL5)
has recently been described elsewhere.19
Experimental Section
Materials and Procedures. All chemical reagents and solvents were
obtained from commercial suppliers and used as received. Whatman
F254 silica gel-60 plates of 1 mm thickness were used for preparative
TLC. Nitric oxide (NO) (Matheson, 99%) was purified by a previously
reported method.20 NO was transferred to the reaction solutions by a
gastight syringe in an anaerobic chamber. Fluorescence emission spectra
were recorded on a Photon Technology International fluorescence
spectrophotometer at 25 or 37 °C. A Varian 300 or 500 NMR
(11) Wink, D. A.; Grisham, M. B.; Mitchell, J. B.; Ford, P. C. Methods Enzymol.
1996, 268, 12-31.
(12) Franz, K. J.; Singh, N.; Spingler, B.; Lippard, S. J. Inorg. Chem. 2000, 39,
4081-4092.
(13) Franz, K. J.; Singh, N.; Lippard, S. J. Angew. Chem., Int. Ed. 2000, 39,
2120-22.
1
spectrometer was used to record H and 13C NMR spectra. IR spectra
were measured on an Avatar 360 FTIR instrument. UV-vis spectra
were obtained on a Hewlett-Packard 8453 diode-array or a Cary 1E
(14) Hilderbrand, S. A.; Lippard, S. J. Inorg. Chem. 2004, 43, 4674-4682.
(15) Hilderbrand, S. A.; Lippard, S. J. Inorg. Chem. 2004, 43, 5294-5301.
(16) Lim, M. H.; Kuang, C.; Lippard, S. J. ChemBioChem 2006, web release
date June 21, 2006; DOI: 10.1002/cbic.200600042.
(17) Lim, M. H.; Lippard, S. J. J. Am. Chem. Soc. 2005, 127, 12170-12171.
(18) Smith, R. C.; Tennyson, A. G.; Lim, M. H.; Lippard, S. J. Org. Lett. 2005,
7, 3573-3575.
(22) Tsuge, K.; DeRosa, F.; Lim, M. D.; Ford, P. C. J. Am. Chem. Soc. 2004,
126, 6564-6565.
(23) Chang, C. J.; Lippard, S. J. In NeurodegeneratiVe Diseases and Metal
Ions: Metal Ions in Life Sciences; Sigel, A., Sigel, H., Sigel, R. K. O.,
Eds.; John Wiley & Sons: 2006; Vol. 1, pp 321-370.
(19) Lim, M. H.; Xu, D.; Lippard, S. J. Nat. Chem. Biol. 2006, 2, 375-380.
(20) Hilderbrand, S. A.; Lim, M. H.; Lippard, S. J. J. Am. Chem. Soc. 2004,
126, 4972-4978.
(24) Nolan, E. M.; Jaworski, J.; Okamoto, K.-I.; Hayashi, Y.; Sheng, M.; Lippard,
S. J. J. Am. Chem. Soc. 2005, 127, 16812-16823.
(21) Lim, M. H.; Lippard, S. J. Inorg. Chem. 2004, 43, 6366-6370.
9
J. AM. CHEM. SOC. VOL. 128, NO. 44, 2006 14365

A R T I C L E S
Lim et al.
spectrophotometer. High-resolution mass spectrometric measurements
were performed by staff at the MIT Department of Chemistry
Instrumentation Facility (DCIF).
anions in general, but found the corrections to be unimportant for the
aims of this study. We utilize the 6-31G** geometries for simplicity.
All orbital energies discussed in this work were computed with the
Amsterdam Density Functional (ADF) 2005.01 software package,38
using a triple-ú quality basis (TZP), the BLYP functional,34,39 and the
COSMO solvation model40,41 with ꢀ ) 78.4.
Syntheses. Precursors of FL1 and FL2 were prepared by modification
of a previously reported method.42 The syntheses of 2-[2-chloro-6-
hydroxy-5-(quinolin-8-ylaminomethyl)-3-oxo-3H-xanthen-9-yl]benzo-
ic acid (FL4 ) QZ1), 2-{2-chloro-6-hydroxy-5-[(2-methylquinolin-8-
ylamino)methyl]-3-oxo-3H-xanthen-9-yl}benzoic acid (FL5), and 2-{2-
chloro-6-hydroxy-5-[((2-methylquinolin-8-yl)(nitroso)amino)methyl]-
3-oxo-3H-xanthen-9-yl}benzoic acid (FL5-NO) were described
previously.19,24
X-ray Crystallographic Studies. Single crystals suitable for data
collection were mounted in Infineum V8512 on the tips of glass
capillaries and frozen in a -100 °C nitrogen cold stream. Data were
collected on a Bruker APEX CCD X-ray diffractometer with Mo KR
radiation (λ ) 0.71073 Å) controlled by the SMART software
package.25 The general procedures used for data collection are reported
elsewhere.26 Empirical absorption corrections were calculated with the
SADABS program.27 Structures were solved by direct methods and
refined with the SAINTPLUS and SHELXTL software packages.28,29
All non-hydrogen atoms were refined anisotropically. Hydrogen atoms
were assigned idealized positions and each was given a thermal
parameter equivalent to 1.2 times the thermal parameter of the atom to
which it was attached. All structure solutions were checked for higher
symmetry with PLATON.30 In the structure of [Cu(modL)2](BF4)2‚
2CH3OH, the hydrogen atoms of the lattice CH3OH molecule were
not assigned, since the oxygen atom of a CH3OH molecule was
disordered over two positions with occupancy factors of 0.15 and 0.85,
8-Nitro-2-quinolinecarboxylic Acid Methyl Ester (1). To a solution
(20 mL of CH3OH and 50 mL of Et2O) of 8-nitro-2-quinolinecarboxylic
acid (1.0 g, 4.3 mmol) was added Me3SiCHN2 (15 mL, 30 mmol, 2.0
M in Et2O) until the color of the solution became light yellow. The
reaction solution was stirred for 4 h at room temperature. Removal of
the solvent provided a light brown solid. The product (0.47 g, 2.0 mmol,
47%) was purified by column chromatography (SiO2, 1:1 hexanes/
-
respectively. Four fluorine atoms of one BF4 anion were also
1
EtOAc Rf ) 0.55 by TLC). H NMR (500 MHz, CD2Cl2): δ (ppm)
disordered over two positions with occupancy factors of 0.28 and 0.72,
respectively, in the structure of [Cu2(modL′)2(CH3OH)](BF4)2‚CH3OH.
The highest electron density in the final difference Fourier maps for
[Cu2(modL′)2(CH3OH)] (BF4)2‚CH3OH was 1.368 e/Å3 in the region
4.05 (3H, s), 7.74 (1H, t, J ) 7.5), 8.11-8.15 (2H, m), 8.31 (1H, d, J
) 8.5), 8.43 (1H, d, J ) 8.0). HRMS (m/z): [M + Na]+ calcd for
C11H8N2NaO4 255.0382, found 255.0374.
-
of the disordered BF4 anion.
8-Amino-2-quinolinecarboxylic Acid Methyl Ester (2). Ethyl
acetate (20 mL) was added to a mixture of 1 (0.18 g, 0.76 mmol) and
10% Pd/C (63 mg), and the solution was deoxygenated by purging
with Ar for 20 min. One atmosphere of H2(g) was introduced to the
deoxygenated solution overnight. The solid residues were removed by
filtration over Celite. Concentration of the filtrate produced the desired
Computational Details. All geometries were optimized within the
density functional theory31 framework using Jaguar 6.032 at the B3LYP/
6-31G** level of theory.33,34 A standard geometry optimization of a
fluorescein skeleton in its deprotonated carboxylate form gives the
lactone form. This result arises because the fluorescein lactone is the
most stable form in the gas phase. Because the LUMO of the xanthene
ring is utilized in forming the lactone ring, the lactone is not the active,
fluorescent form of the molecule. To obtain the more relevant, ring-
opened fluorescein structure, we carried out a full geometry optimization
of its protonated carboxylic acid form. The proton was then removed
and the carboxylate portion of the molecule reoptimized, keeping the
xanthene fragment frozen. Whereas this procedure is an approximate
solution to the intrinsic problem of lactone formation, being unavoidable
in a gas phase geometry optimization, it is expected to give a
qualitatively correct model of the solution-phase geometry of fluores-
cein. A second problem that needed to be addressed is that both FL5
and FL5-NO are dianions at neutral pH, which gives rise to a number
of positive orbital energies in gas phase simulations and complicates
the interpretation of the computed orbital energies. In previous work,
we showed that the perturbative addition of the continuum solvation
potential in self-consistent-field calculations leads to physically more
meaningful Kohn-Sham wave functions for anionic species.35,36 We
also considered the use of a more elaborate basis set, such as cc-pVTZ-
(-f)++ including diffuse functions,37 which are more appropriate for
1
product (0.13 g, 0.66 mmol, 86%). H NMR (500 MHz, CDCl3): δ
(ppm) 4.05 (3H, s), 6.97 (1H, dd, J ) 7.5, J ) 1.5), 7.18 (1H, dd, J )
8.3, J ) 0.5), 7.44 (1H, t, J ) 8.0), 8.13 (1H, d, J ) 8.5), 8.19 (1H,
d, J ) 8.5). HRMS (m/z): [M + Na]+ calcd for C11H10N2NaO2
225.0640, found 225.0638.
(8-Amino-2-quinolinyl)methanol (3). To an ethanol (8 mL) solution
of NaBH4 (0.19 g, 4.9 mmol) at 0 °C was added dropwise a THF
solution (8 mL) of 2 (0.10 g, 0.50 mmol) over 5 min. The solution
was allowed to warm to room temperature with stirring overnight. A
saturated aqueous NaHCO3 solution (3 mL) and water (2 mL) were
added to the reaction, and the solution was extracted with CHCl3 three
times. The collected CHCl3 layer was washed with brine and water
and dried with MgSO4. Removal of the solvent provided the desired
1
product (63 mg, 0.36 mmol, 73%). H NMR (500 MHz, CDCl3): δ
(ppm) 4.95 (2H, s), 7.00 (1H, dd, J ) 7.5, J ) 1.5), 7.20 (1H, dd, J )
8.0, J ) 1.0), 7.29 (1H, d, J ) 8.5), 7.36 (1H, t, J ) 7.5), 8.12 (1H,
d, J ) 8.5). HRMS (m/z): [M + Na]+ calcd for C10H10N2NaO 197.0691,
found 197.0676.
2-{2-Chloro-6-hydroxy-5-[(2-(methylcarboxy)quinolin-8-ylami-
no)methyl]-3-oxo-3H-xanthen-9-yl}benzoic Acid (FL1). An ethyl
acetate (3 mL) solution of 2 (15 mg, 76 µmol) and 7′-chloro-4′-
fluoresceincarboxaldehyde43 (30 mg, 76 µmol) was stirred overnight
at room temperature. The resulting red residue was collected, dried in
vacuo, and dissolved in dichloroethane (2 mL). To this solution was
added NaB(OAc)3H (19 mg, 0.15 mmol) and the reaction solution was
stirred overnight at room temperature. A portion of the crude material
was purified by preparative TLC on reverse phase silica gel (Rf ) 0.32,
(25) SMART: Software for the CCD Detector System, version 5.626; Bruker
AXS: Madison, WI, 2000.
(26) Kuzelka, J.; Mukhopadhyay, S.; Spingler, B.; Lippard, S. J. Inorg. Chem.
2004, 43, 1751-1761.
(27) Sheldrick, G. M. SADABS: Area-Detector Absorption Correction, Uni-
versity of Go¨ttingen: Go¨ttingen, Germany, 1996.
(28) SAINTPLUS: Software for the CCD Detector System, version 5.01; Bruker
AXS: Madison, WI, 1998.
(29) SHELXTL: Program Library for Structure Solution and Molecular
Graphics, version 6.1; Bruker AXS: Madison, WI, 2001.
(30) Spek, A. L. PLATON, A Multipurpose Crystallographic Tool, Utrecht
University: Utrecht, The Netherlands, 2000.
(38) te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca Guerra, C.; van
Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. J. Comput. Chem. 2001,
22, 931-967.
(39) Becke, A. D. Phys. ReV. A 1988, 38, 3098-3100.
(40) Klamt, A.; Schu¨u¨rmann, G. J. Chem. Soc., Perkin Trans. 2 1993, 799-
805.
(41) Pye, C. C.; Ziegler, T. Theor. Chem. Acc. 1999, 101, 396-408.
(42) Roth, R.; Erlenmeyer, H. HelV. Chim. Acta 1954, 37, 1064-1068.
(43) Nolan, E. M.; Burdette, S. C.; Harvey, J. H.; Hilderbrand, S. A.; Lippard,
S. J. Inorg. Chem. 2004, 43, 2624-2635.
(31) Parr, R. G.; Yang, W. Density Functional Theory of Atoms and Molecules,
Oxford University Press: New York, 1989.
(32) Jaguar 6.0; Schro¨dinger, Inc.: Portland, Oregon, 2003.
(33) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
(34) Lee, C. T.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789.
(35) Baik, M.-H.; Friesner, R. A. J. Phys. Chem. A 2002, 106, 7407-7415.
(36) Baik, M.-H.; Schauer, C. K.; Ziegler, T. J. Am. Chem. Soc. 2002, 124,
11167-11181.
(37) Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007-1023.
9
14366 J. AM. CHEM. SOC. VOL. 128, NO. 44, 2006

Nitric Oxide Detection in Aqueous Solution
A R T I C L E S
Scheme 2 a
CH3OH/H2O ) 3:0.7) to yield a red solid (16 mg, 27 µmol, 36%).
Mp: 155-157 °C (dec). 1H NMR (500 MHz, CD3OD): δ (ppm) 3.97
(3H, s), 4.79 (2H, s), 6.53 (1H, d, J ) 9), 6.61 (1H, s), 6.63 (1H, d, J
) 8.5), 6.99 (1H, s), 7.04 (1H, d, J ) 8.0), 7.12 (1H, d, J ) 8.0), 7.19
(1H, d, J ) 7.5), 7.46 (1H, t, J ) 8.0), 7.68 (1H, t, J ) 7.5), 7.73 (1H,
t, J ) 7.5), 7.99 (2H, t, J ) 8.5), 8.15 (1H, d, J ) 9.5). FTIR (KBr,
cm-1): 3421 (br, m), 2962 (w), 2917 (vw), 2849 (vw), 1759 (m), 1722
(m), 1627 (m), 1610 (m), 1577 (w), 1565 (w), 1519 (m), 1489 (w),
1445 (w), 1427 (m), 1377 (w), 1283 (w), 1263 (s), 1217 (w), 1149
(w), 1100 (m), 1092 (m), 1063 (w), 1027 (m), 873 (w), 803 (s), 769
(w), 701 (w), 617 (vw), 575 (vw), 545 (vw), 469 (vw), 403 (vw). HRMS
(m/z): [M - H]- calcd for C32H20ClN2O7 579.0959, found 579.0963.
Anal. Calcd for C32H21ClN2O7: C, 66.16; H, 3.64; N, 4.82. Found: C,
66.18; H, 3.53; N, 4.98.
a Reagents: (a) Br2, AcONa, and AcOH (ref 42); (b) 20% H2SO4(aq)
(ref 42); (c) Me3SiCHN2; (d) Pd/C, H2; and (e) NaBH4.
2-{2-Chloro-6-hydroxy-5-[(2-carboxyquinolin-8-ylamino)methyl]-
3-oxo-3H-xanthen-9-yl}benzoic acid (FL2). Portions of 8-amino-
quinoline-2-carboxylic acid (14 mg, 76 µmol) and 7′-chloro-4′-
fluoresceincarboxaldehyde43 (30 mg, 76 µmol) were added to 2 mL of
EtOAc and the reaction solution was stirred overnight at room
temperature. After removing the solvent, the resulting red residue was
collected and dissolved in dichloroethane (2 mL) followed by the
addition of NaB(OAc)3H (19 mg, 0.15 mmol). The reaction solution
was stirred and the solvent was removed after 1 d. The crude product
was twice purified by preparative TLC on silica gel (first purification,
Rf ) 0.11, 5:1 CH3OH/CH2Cl2; second purification, Rf ) 0.61, 6:1
CH3OH/0.1 M HCl), affording a red solid (9.7 mg, 17 µmol, 23%).
Mp: 240-242 °C (dec). 1H NMR (300 MHz, CD3OD/CD2Cl2): δ
(ppm) 4.72 (2H, s), 6.57 (1H, d, J ) 8.1), 6.68 (1H, s), 6.93 (0.5H, d,
J ) 7.5), 7.00-7.13 (3H, m), 7.20 (1H, d, J ) 6.3), 7.32 (0.5H, t, J )
7.5), 7.48 (1H, t, J ) 8.1), 7.53-7.61 (2H, m), 8.03-8.14 (3H, m).
FTIR (KBr, cm-1): 3404 (br, m), 3048 (vw), 2962 (w), 2923 (w), 2845
(w), 1635 (m), 1609 (m), 1576 (s), 1512 (w), 1459 (m), 1340 (w),
1303 (w), 1261 (m), 1221 (w), 1148 (s), 1094 (s), 1015 (s), 938 (vw),
878 (vw), 859 (vw), 821 (m), 796 (s), 745 (vw), 713 (vw), 690 (vw),
661 (vw), 628 (w), 599 (w), 582 (vw), 550 (w), 527 (vw), 516 (vw),
492 (w), 472 (m), 460 (w), 442 (w), 435 (w), 422 (w), 413 (w), 405
(w). HRMS (m/z): [M + Na]+ calcd for NaC31H19ClN2O7 589.0779,
found 589.0786. Anal. Calcd for C31H19ClN2O7: C, 65.67; H, 3.38; N,
4.94. Found: C, 65.58; H, 3.32; N, 4.79.
Quantities of [Cu2(modL′)2(CH3OH)](BF4)2‚CH3OH sufficient for full
characterization were not obtained. FTIR (KBr, cm-1): 3485 (br, w),
3216 (m), 3071 (vw), 2952 (vw), 2931 (vw), 2849 (vw), 1617 (vw),
1598 (vw), 1570 (vw), 1513 (m), 1485 (m), 1454 (m), 1384 (m), 1322
(w), 1265 (s), 1199 (vw), 1130 (sh, w), 1081 (br, vs), 1042 (br, vs),
936 (sh, vw), 903 (vw), 881 (w), 859 (vw), 834 (w), 804 (w), 764 (s),
730 (w), 656 (vw), 633 (vw), 618 (vw), 598 (vw), 581 (vw), 551 (w),
519 (w), 504 (w), 463 (w), 415 (w). Characterization of [Cu(modL)2]-
(BF4)2‚2CH3OH. Mp: 223-225 °C (dec). FTIR (KBr, cm-1): 3379
(br, m), 3246 (m), 3119 (vw), 3070 (w), 3056 (w), 3026 (vw), 2940
(vw), 2864 (vw), 1616 (w), 1596 (m), 1516 (s), 1479 (w), 1462 (s),
1400 (w), 1382 (w), 1358 (w), 1332 (w), 1316 (vw), 1270 (w), 1248
(w), 1188 (w), 1175 (w), 1162 (sh, w), 1133 (sh, w), 1100 (br, vs),
1080 (br, vs), 1039 (sh, vw), 936 (sh, vw), 885 (w), 867 (w), 848 (w),
832 (m), 794 (w), 763 (s), 731 (w), 721 (w), 664 (vw), 559 (vw), 519
(w), 491 (vw). Anal. Calcd for CuC32H28B2F8N4O2‚2CH3OH: C, 50.93;
H, 4.53; N, 6.99. Found: C, 50.41; H, 4.18; N, 6.98.
Results and Discussion
Design Considerations for Cu(II)-Based NO Sensors. Of
the various metal-based fluorescent sensors for nitric oxide
prepared and evaluated in our laboratory, Cu(II) systems proved
to have the greatest potential for application in a physiologically
relevant setting.17-19 Fluorescein-based ligands for Cu(II) were
chosen, since fluorescein is water-soluble and highly emissive
(Φ ) 0.95) and has excitation and emission wavelengths in the
visible region,44,45 which minimizes harm to cellular components
during NO imaging. Incorporation of copper-binding units to
the fluorescein moiety was achieved in a manner analogous to
the synthesis in our laboratory of fluorescent sensors for
Zn(II).23 The target ligands are displayed in Figure 1, FL4 being
the same as QZ1 described previously.24
Synthesis of Fluorescein-Based Ligands (FLn). The general
assembly of fluorescein-based sensors with one metal ion
binding site containing two nitrogen and one or more oxygen
donor atoms (N2On, n ) 1 or 2) has been previously achieved
in our laboratory by using 7′-chloro-4′-fluoresceincarboxalde-
hyde.43 This general synthesis was adopted to prepare
copper(II)-based NO sensors in the present study.
2-{2-Chloro-6-hydroxy-5-[(2-hydroxymethylquinolin-8-ylamino)-
methyl]-3-oxo-3H-xanthen-9-yl}benzoic Acid (FL3). A solution
(EtOAc, 3 mL) of 3 (33 mg, 0.19 mmol) and 7′-chloro-4′-fluores-
ceincarboxaldehyde43 (75 mg, 0.19 mmol) was stirred overnight at room
temperature. The resulting red residue was collected, dried in vacuo,
and dissolved in dichloroethane (2 mL). To the dichloroethane solution
was added a portion of NaB(OAc)3H (48 mg, 0.23 mmol). The solution
was stirred overnight and purified by preparative silica TLC (Rf ) 0.22,
1:1:0.9 EtOAc/Hx/CH3OH), affording a red solid (23 mg, 42 µmol,
22%). Mp: 230-231 °C (dec). 1H NMR (500 MHz, CD3OD): δ (ppm)
4.74 (2H, s), 4.77 (2H, s), 6.53 (1H, d, J ) 9.5), 6.73 (1H, s), 6.97-
7.00 (2H, m), 7.09-7.17 (3H, m), 7.32 (1H, t, J ) 8.0), 7.45 (1H, d,
J ) 8.0), 7.49-7.56 (2H, m), 7.99 (1H, d, J ) 9.0), 8.05 (1H, d, J )
10.5). FTIR (KBr, cm-1): 3427 (br, m), 3059 (vw), 2996 (vw), 2963
(w), 2932 (vw), 2900 (vw), 1634 (vw), 1575 (s), 1519 (w), 1457 (m),
1419 (w), 1376 (m), 1342 (w), 1305 (vw), 1263 (m), 1223 (w), 1151
(w), 1092 (w), 1039 (vw), 925 (vw), 881 (vw), 822 (m), 806 (m), 712
(vw), 649 (vw), 619 (vw), 600 (vw), 550 (vw), 469 (vw). HRMS
(m/z): [M - H]- calcd for C31H20ClN2O6 551.1010, found 551.1003.
[Cu2(modL′)2(CH3OH)](BF4)2‚CH3OH and [Cu(modL)2](BF4)2‚
2CH3OH. Two kinds of crystals were grown by vapor diffusion of
Et2O into a methanol solution (3 mL) of 2-[(quinolin-8-ylamino)methyl]-
phenol (modL) (10 mg, 40 µmol) and copper(II) tetrafluoroborate (9.5
mg, 40 µmol) at room temperature overnight. Violet {major product,
[Cu(modL)2](BF4)2‚2CH3OH} and green {minor product, [Cu2(modL′)2-
(CH3OH)](BF4)2‚CH3OH, modL′ ) 2-[(quinolin-8-ylamino)methyl]-
phenolate} crystals were manually separated for characterization.
The synthesis of amine ligands 2 and 3 was accomplished
by modification of a previously reported method, as shown in
Scheme 2.42 8-Nitro-2-quinolinecarboxylic acid was generated
by bromination of 8-nitroquinaldine followed by hydrolysis
as described.42 To prepare 8-nitro-2-quinolinecarboxylic acid
methyl ester (1), trimethylsilyldiazomethane (Me3SiCHN2),
(44) Sjo¨back, R.; Nygren, J.; Kubista, M. Spectrochim. Acta Part A 1995, 51,
L7-L21.
(45) Brannon, J. H.; Madge, D. J. Phys. Chem. 1978, 82, 705-709.
9
J. AM. CHEM. SOC. VOL. 128, NO. 44, 2006 14367

A R T I C L E S
Scheme 3
Lim et al.
Table 1. Spectroscopic Resultsa
1
absorption
unbound
λmax (nm),
ꢀ
(×
104 M-1 cm-
)
emission
unbound (
λmax (nm), Φ
b or %
Cu(II)c
Φ
)
Cu(II)d (%)
ref
FL1
FL2
FL3
FL4
FL5
504, 4.3 ( 0.1
503, 3.8 ( 0.5
503, 3.9 ( 0.1
505, 6.9 ( 0.1
504, 4.2 ( 0.1
499, 4.0 ( 0.1
496, 3.8 ( 0.3
497, 3.9 ( 0.5
496, 5.7 ( 0.1
499, 4.0 ( 0.1
520, 0.083 ( 0.004
520, 0.084 ( 0.002
520, 0.31 ( 0.01
520, 0.024 ( 0.001
520, 0.077 ( 0.002
520, 32 ( 2
520, 26 ( 3
520, 19 ( 2
520, 30 ( 3
520, 18 ( 3
this work
this work
this work
24 and this work
19 and this work
a All spectroscopic measurements were performed at pH 7.0 in 50 mM PIPES, 100 mM KCl buffer. b Reported quantum yields are based on fluorescein,
Φ ) 0.95 in 0.1 N NaOH.45 c One equivalent of CuCl2 was added. d Percent fluorescence decrease upon addition of 1 equiv of Cu(II) to a pH 7.0 buffered
solution (50 mM PIPES, 100 mM KCl) of FLn (1 µM), relative to that of free FLn (50 mM PIPES, 100 mM KCl).
which is a safer alternative to the use of diazomethane, was
employed in the present work, affording 1 as a light-brown solid
in moderate yield (47%) after purification on silica gel (1:1
hexanes/EtOAc). Hydrogenation of 1 in EtOAc using 10% Pd/C
gave pure 2 in 86% yield without further manipulation. Upon
reduction of 2 with sodium borohydride in a mixture of THF
and EtOH, 3 was obtained in 73% yield without purification.
Scheme 3 illustrates the syntheses of the fluorescein-based
ligands FL1-FL5. Combination of 2 or 3 (Scheme 2) with 7′-
chloro-4′-fluoresceincarboxaldehyde in dry EtOAc resulted in
precipitates of the respective intermediate imines, as shown in
the scheme. The imines were reduced under mild conditions
by using NaB(OAc)3H in dichloroethane. A portion of the crude
materials was purified by preparative TLC, affording desired
products FL1 and FL3, respectively, in moderate yield. Com-
mercially available amine moieties were assembled with 7′-
chloro-4′-fluoresceincarboxaldehyde to obtain pure FL2, FL4,24
and FL519 by a similar procedure. The spectroscopic properties
of the FLn (n ) 1-5) compounds are summarized in Table 1.
Preparation and Fluorescence Studies of Copper Fluo-
rescein Compounds. The Cu(II) fluorescein complexes Cu(FLn)
(n ) 1-5, Figure 1) were generated in situ by combining FLn
with CuCl2 in a 1:1 ratio in 50 mM PIPES, pH 7.0, 100 mM
KCl buffer. Upon Cu(II) binding to the fluorescein ligands FLn,
a blue-shifted absorption band was observed, compared to that
of free FLn, as summarized in Table 1. UV-visible spectroscopy
was employed to verify the binding stoichiometry of the FLn:
Cu(II) complexes at pH 7.0. For FL1, FL3, and FL5,19 Job’s
plots of CuCl2:ligand mixtures revealed a break at 0.5, sug-
gesting the formation of 1:1 complexes (Figure 2). For the
reactions of FL2 or FL4 with CuCl2, however, breaks occurred
at values between 0.33 and 0.5, which may indicate the
generation of a mixture of 1:1 and 1:2 complexes in pH 7.0
buffered solution. This mixed binding mode of FL4 with Cu(II)
was also observed in the reaction of Cu(BF4)2 with 2-[(quinolin-
8-ylamino)methyl]phenol (modL), investigated as a model for
the reaction of FL4. Two kinds of crystals, [Cu(modL)2](BF4)2
and [Cu2(modL′)2(CH3OH)](BF4)2 (modL′ ) 2-[(quinolin-8-
ylamino)methyl]phenolate), were obtained by vapor diffusion
of Et2O into a methanol solution of Cu(BF4)2 and modL in a
ratio of 1:1 (Figure 3). The observation of 1:1 and 1:2 complexes
in the Cu(II) modL system suggests that a similar mixed binding
mode for FL2 and FL4 with Cu(II) might occur. Also, these
crystal structures reveal that one oxygen and two nitrogen atoms
of the ligands FLn might bind to Cu(II), as in the crystal structure
of [Cu2(modL′)2(CH3OH)](BF4)2. Crystallographic data for
[Cu(modL)2](BF4)2 and [Cu2(modL′)2(CH3OH)](BF4)2 are sum-
marized in Tables S2 and S7 (Supporting Information), and
selected bond lengths and angles are listed in Table 2.
As shown in Table 1, the fluorescence emission intensities
of 1 µM FLn (n ) 1-5) pH 7.0 buffered solutions at 37 °C are
slightly diminished upon addition of 1 equiv of CuCl2. Introduc-
tion of excess NO to buffered Cu(FLn) (1 µM FLn and 1 µM
CuCl2) solutions (50 mM PIPES, pH 7.0, 100 mM KCl) causes
an increase in fluorescence. A 2.5((0.1)-46 or 8.3((0.9)-fold
increase in fluorescence was observed in the reaction of Cu(FL1)
or Cu(FL2) with excess NO over 60 or 70 min (Figure 4a,b).
Copper complexes Cu(FL3) and Cu(FL4) displayed
3.4((0.1)-46 and 31((1)-fold fluorescence enhancements upon
treatment with excess NO over 15 and 20 min, respectively
(46) Preliminary measurements of the reactions of Cu(FLn) (n ) 1, 3) with
NO: Tio, J.; Lim, M. H.; Lippard, S. J. Nucleus 2006, 84, 8-12.
9
14368 J. AM. CHEM. SOC. VOL. 128, NO. 44, 2006

Nitric Oxide Detection in Aqueous Solution
A R T I C L E S
Figure 2. Job’s plot for the formation of the FLn:Cu(II) complex, where n ) 1 (a), 2 (b), 3 (c), 4 (d), and 5 (e),19 determined by using UV-vis spectroscopy
in a pH 7.0 buffered solution (50 mM PIPES, 100 mM KCl). The concentrations of the initial FLn (n ) 1-5) and Cu(II) solutions are 10, 10, 5, 5, and 5
µM, respectively.
Figure 3. ORTEP diagrams of [Cu(modL)2](BF4)2 (left) and [Cu2(modL′)2(CH3OH)](BF4)2 (right) showing 50% probability thermal ellipsoids. The Cu1-
-
O1 distance of 2.492 Å in the structure of [Cu(modL)2](BF4)2 indicates a weak interaction. The BF4 anions are omitted in the figure.
Table 2. Selected Bond Lengths (Å) and Angles (deg)a
(Figure 4c,d). Last, a 16((1)-fold fluorescence increase was
[Cu(moL)2](BF4)2·
2CH3OH
[Cu2(modL)2(CH3OH)]-
(BF4)2·CH3OHb
observed at 5 min when Cu(FL5) was allowed to react with
excess NO (immediate 11((2)-fold increase in fluorescence,
Figure 4e).19 The lower limit of Cu(FL5) for NO detection is 5
nM.19 The fluorescence response of the Cu(FL5) probe is specific
for NO over other reactive species present in biological systems,
including ClO- (1.9((0.1)-fold increase), H2O2 (1.2((0.1)-fold
increase), NO3 (1.2((0.2)-fold increase), NO2 (1.2((0.2)-
fold increase), ONOO- (2.6((0.3)-fold increase), and HNO
(1.4((0.2)-fold increase).19 A 2.0((0.2)-fold increase in fluo-
rescence was observed when Cu(FL5) was allowed to react with
superoxide ion (O2-) that was generated by xanthine and
xanthine oxidase (Figure 5).47,48 Although Cu(FL5) exhibited a
slight fluorescence increase in the presence of peroxynitrite or
superoxide ion, it is significantly less than that caused by NO.
Cu1-N1
1.997(2)
2.033(2)
1.956(4)
2.009(4)
1.928(3)
1.947(3)
85.39(17)
Cu1-N2
Cu1-O1
Cu1-O2
N1-Cu1-N2
N1-Cu1-N2A
O1-Cu1-O2
Cu1-O1-Cu2
Cu1-O2-Cu2
N3-Cu2-N4
O1-Cu2-O2
97.24(10)
82.76(10)
-
-
77.47(14)
100.85(15)
99.96(15)
85.32(18)
76.68(14)
a Numbers in parentheses are estimated standard deviations in the last
significant figures. Atoms are labeled as indicated in Figure 3. b Bond
lengths and angles of only one of the two crystallographically independent,
but chemically identical, molecules are listed.
(47) Sono, M. J. Biol. Chem. 1989, 264, 1616-1622.
(48) Yang, D.; Wang, H.-L.; Sun, Z.-N.; Chung, N.-W.; Shen, J.-G. J. Am. Chem.
Soc. 2006, 128, 6004-6005.
Fluorescence enhancement in the NO reactions of Cu(FL5)
is not significantly affected in different buffered solutions and
9
J. AM. CHEM. SOC. VOL. 128, NO. 44, 2006 14369

A R T I C L E S
Lim et al.
Figure 4. Fluorescence emission spectra of a solution of Cu(FLn) (1 µM FLn and 1 µM CuCl2) in deoxygenated buffered solution (50 mM PIPES, pH 7.0,
100 mM KCl) before (dashed line) and after (solid lines) addition of 1300 equiv of NO(g) at 37 °C.
Figure 6. Fluorescence emission spectra of a solution of Cu(FL5) upon
Figure 5. Specificity of Cu(FL5) for NO at pH 7.0, determined as the
fluorescence response of Cu(FL5) after addition of 100 equiv of ClO-
(sodium hypochlorite), O2- (0.1 U xanthine oxidase and 100 µM xanthine),
H2O2, NO3-, NO2-, ONOO- (sodium peroxynitrite), NO (1.9 mM NO-
saturated aqueous solution), and HNO (Angeli’s salt Na2N2O3) for 2 h.
The excitation wavelength was 503 nm. All data (F) are normalized with
respect to the emission of Cu(FL5) [F(Cu(FL5))]. The Cu(II) species Cu-
(FL5) was prepared in situ by reacting FL5 (1 µM) with CuCl2 (1 µM).
addition of excess NO. (a) Fluorescence response of Cu(FL5) (1 µM FL5
and 1 µM CuCl2) in a deoxygenated pH 7.0 buffered solution (20 mM
phosphate) before (dashed line) and after (solid lines) addition of 1300 equiv
of NO(g) at 37 °C. (b) Fluorescence emission spectra of an aerobic solution
of Cu(FL5) (1 µM FL5 and 1 µM Cu(NO3)2) upon admission of 1300 equiv
of NO at 37 °C (50 mM PIPES, pH 7.0, 100 mM KNO3).
turn-on under aerobic conditions, indicating that it is an indirect
NO sensor.19
is independent of Cl- ion. When 20 mM potassium phosphate
buffer was used to follow the NO reaction of Cu(FL5), a
10((2)-fold fluorescence increase occurred over 5 min (Figure
6a). When Cu(FL5) was prepared in situ by the reaction of
Cu(NO3)2 with FL5 in a 1:1 ratio, a 12((2)-fold increase was
observed upon addition of excess NO over 5 min (50 mM
PIPES, pH 7.0, 100 mM KNO3), as shown in Figure 6b.
These results demonstrate that Cu(FLn) can directly detect
NO with significant emission turn-on at a physiologically
relevant pH. A comparison of the NO reactions of Cu(FLn) with
those of a commercially available NO probe DAF-2 (o-
diaminofluorescein) clearly highlights the unique ability of
copper-based sensors for direct NO sensing. DAF-2 displayed
no increase in fluorescence upon addition of excess NO for 1
h in the absence of O2, but it did exhibit immediate fluorescence
Mechanism of Fluorescence Detection of NO by Cu(FLn).
When FL1 or FL5 was titrated with CuCl2 at 25 °C, the optical
absorption spectral changes could be fit to a one-step binding
equation, affording a dissociation constant (Kd) of 0.25((0.14)
or 1.5((0.3) µM19 for Cu(II) ion (Figure 7). These Kd values
indicate that the reaction solution of FL1 or FL5 (1 µM) with
CuCl2 (1 µM) contains Cu(II) ion, free FLn, and Cu(FLn), as
described in Scheme 4. The 6-fold increase in binding affinity
of FL1 compared to FL5 is consistent with the presence of an
additional oxygen atom for copper coordination. To identify
the species responsible for NO detection, the fluorescence of a
Cu(II)-free FL5 solution was monitored during treatment with
excess NO. There was none of the fluorescence enhancement
that was observed in the reaction of Cu(FL5) with NO.19
Moreover, no fluorescence increase was observed upon addition
9
14370 J. AM. CHEM. SOC. VOL. 128, NO. 44, 2006

Nitric Oxide Detection in Aqueous Solution
A R T I C L E S
Scheme 5
solution of FL5 containing excess NaNO2(aq) over ice with
HCl(aq) immediately induced a color change from red-orange
to bright yellow. A precipitate immediately appeared and the
supernatant was collected after 30 min by centrifugation.
Dialysis of the supernatant against water was performed to
remove residual NaNO2. An orange solid was obtained in
moderate yield upon lyophilization. The isolated solid analyzed
by LC-MS was identical to the material obtained in the reaction
of Cu(FL5) with NO.19 In addition, the preparation of 15N-labeled
nitrosated product FL5-15NO allowed us to elucidate the site
Figure 7. Measurement of the dissociation constant (Kd) of Cu(FL1) ([)
or Cu(FL5)19 (b). CuCl2 was titrated into a 5 µM FL1 or FL5 solution (50
mM PIPES, pH 7.0, 100 mM KCl). The formation of Cu(FL1) or Cu(FL5)
was followed by the absorbance change (∆A) at 512 nm. The titration trace
was fit to the equation described previously.19
1
of nitrosation by H and 15N NMR spectroscopy. Nitrosation
of FL5 occurs at the secondary amine metal-binding site to create
a N-nitrosamine, FL5-NO, as shown in Scheme 4, causing
dissociation of nitrosated ligand from the copper center. We
next addressed the question of whether FL5-NO is the species
responsible for enhanced emission. The quantum yield of
synthetic FL5-NO was measured to be Φ(FL5-NO) ) 0.58 (
0.02 compared to fluorescein (Φ ) 0.95).45 Thus, FL5-NO is
brighter than FL5 or Cu(FL5), for which Φ ) 0.077 ( 0.002
and 0.063 ( 0.002, respectively.19 These mechanistic studies
clearly demonstrate that Cu(FL5) is capable of fluorescence-
based NO sensing at pH 7.0 by NO-induced metal reduction
followed by release of the N-nitrosated fluorescein ligand from
the copper center, triggering concomitant fluorescence enhance-
ment (Scheme 4). The reason for the enhanced emission of the
N-nitrosated species FL5-NO, compared to FL5, was addressed
by density functional theory (DFT) calculations (vide infra).
Further elaboration of the reaction pathway shown in Scheme
4 is possible by considering the different observed fluorescence
responses of Cu(FLn) (n ) 1-5) to NO. Compound FL5-NO
might be formed by several plausible mechanisms. One pos-
sibility (Scheme 5a) would involve initial NO coordination to
Cu(II) followed by internal electron transfer and migration of
NO+ from the copper(I) center to the amine functionality with
loss of a proton to produce the N-nitrosamine. A subsequent
step would be dissociation of the N-nitrosamine from the Cu(I)
center. An alternative mechanism of N-nitrosamine formation
would be direct interaction of NO with the deprotonated amine
by an inner sphere electron transfer from the ligand to the copper
center to form the N-nitrosamine, as described for related
chemistry in a recent report (Scheme 5b).22
A comparison of the NO reaction rates of the Cu(II)
complexes of FL1, FL3, and FL5 reveals that the NO-induced
fluorescence increase is slower when FLn has one additional
donor atom coordinated to the Cu(II) center (Figure 1 and
Scheme 4). This observation might be explained by one or both
of the following possibilities. The Cu(II) complexes containing
tetradentate ligands FL1 or FL3 may be less able to accommodate
the structural rearrangement required for a planar-to-tetrahedral
conversion upon Cu(I) complex formation as compared to the
tridentate ligand FL5. As a result, the rates of NO reactions of
Cu(FLn) (n ) 1, 3), compared with that of Cu(FL5), are lower,
reducing the rate of formation of the FLn-NO species respon-
sible for fluorescence enhancement. The second possibility
concerns the dissociation of FLn-NO from copper. The presence
of an additional donor atom in the FLn chelate may decelerate
Scheme 4
of excess NO to a Cu(FL5) solution containing the Cu(II)
chelator N,N′-1,2-ethanediylbis(N-(carboxymethyl)glycine)-
(EDTA).19 These observations demonstrate that Cu(FL5), and
not FL5 or Cu(II) ion alone, is the nitric oxide indicator with
fluorescence turn-on.
The mechanism of turn-on emission of Cu(FL5) by NO
requires NO-induced reduction of Cu(II) to Cu(I), forming NO+.
Loss of Cu(II) was monitored by EPR spectroscopy.19 Formation
of a diamagnetic Cu(I) complex during NO reactions of
Cu(II) complexes can restore the quenched fluorescence of a
fluorophore coordinated to the paramagnetic metal.17,18,22
The fluorescence of FL5 is unchanged in the presence of
[Cu(CH3CN)4](BF4), however.19 This result indicates that reduc-
tion of the copper center from Cu(II) to Cu(I) by NO alone
does not cause fluorescence enhancement in the NO reaction
of the Cu(FL5) complex, which differs from prior observations
for Cu(II)-based systems.17,18
Examination of the reaction of Cu(FL5) with NO by UV-
vis spectroscopy and liquid chromatography-mass spectrometry
(LC-MS) revealed that the FL5 ligand becomes nitrosated to
generate FL5-NO, which dissociates from the copper center.19
The FL5-NO species is stable in pH 7.0 buffered solution for
several days, as monitored by LC-MS, indicating the irrevers-
ible nature of NO sensing by Cu(FL5). To characterize further
the nitrosated product from the reaction of Cu(FL5) with NO,
we independently prepared FL5-NO.49 Treatment of a CH3OH
(49) Lee, J.; Chen, L.; West, A. H.; Richter-Addo, G. B. Chem. ReV. 2002,
102, 1019-1065 and references therein.
9
J. AM. CHEM. SOC. VOL. 128, NO. 44, 2006 14371

A R T I C L E S
Lim et al.
Figure 8. DFT-optimized structures of (a) FL5 and (b) FL5-NO.
dissociation of FLn-NO from copper, which would also result
in a slower fluorescence increase. A comparison of the Kd values
for Cu(FL1) and Cu(FL5), 0.25 ( 0.14 and 1.5 ( 0.3 µM (Figure
7), respectively, supports this notion, assuming that the Kd values
for the N-nitrosated amine product may follow this trend. At
present, we are unable to determine whether one or both of these
factors play a significant role in influencing the NO reactivity
of Cu(FLn). More detailed kinetic studies of these and additional
complexes are required to establish how the structure and
composition of the copper-fluorophore complex affects its NO
reactivity.
Introduction of additional donor atoms to FLn should also
affect the electronic properties of the copper center. To address
this issue, electrochemical studies were attempted. Measure-
ments of the Cu(FLn) redox potentials in a mixture of 1:9 DMSO
and DMF, however, were unsuccessful, the redox potential of
free Cu(II) ion being mainly observed. This result might be a
consequence of the micromolar Kd values for the equilibrium
in Scheme 4, measured for Cu(FLn) (n ) 1, 5) in aqueous buffer
and assuming similar values in the DMSO/DMF mixed solvent,
and the corresponding presence of significant concentrations
of free Cu(II) ion in solution.
Although the mechanism of the NO reactions with Cu(FLn)
is not unequivocally established by these experiments, the
present work suggests that both electronic effects and the
equilibrium between bound and free copper will affect formation
of the N-nitrosamine. Irrespective of the mechanism, however,
the ability of the Cu(FLn) complexes to detect NO described
here provides a valuable NO sensing strategy, and improved
versions based on this platform should enable the elucidation
of NO function in vivo.
Theoretical Analysis of the Relative Fluorescence of FL5
versus FL5-NO. The “off-on” properties of fluorescence-
based sensors such as FLn may operate by the well-established
photoinduced electron transfer (PET) mechanism.50-52 In PET,
the presence of a covalently bound nitrogen atom with a lone
pair of electrons will quench the fluorescence of the sensor in
the absence of an analyte. Perturbation of this nitrogen atom
lone pair, such as by coordination to a diamagnetic metal ion,24
Figure 9. (a) Relative energies of the molecular orbitals for the ground
states of FL5 (left) and FL5-NO (right). The NNO-based orbital, HOMO-
14, of FL5-NO is expected to be too low in energy to serve as a donor for
fluorescence quenching. (b) Qualitative molecular orbital diagram for the
ground (left), excited (middle), and charge-transfer (right) states of FL5.
The lowering of the HOMO and LUMO associated with the fluorophore in
the excited state is due to electronic relaxation, which would not be expected
to alter significantly the energy of the nitrogen-based MO. Nonradiative
relaxation of the charge-transfer state returns the molecule to the ground
state. R1 ) 2-methylquinoline and R2 ) fluorophore.
will alleviate this quenching process, causing fluorescence “turn-
on.” Note that binding to a paramagnetic metal ion such as
Cu(II) is expected to preserve or enhance the quenching behavior
of fluorophores.53
For the FL5/FL5-NO system, both species have a nitrogen
atom-based lone pair of electrons that might be expected to
contribute to fluorescence quenching, raising the important
question of why fluorescence “turn-on” occurs at all in the
nitrosated species. The DFT-computed geometries (Figure 8)
illustrate an important structural difference between FL5 and
FL5-NO. Whereas the amino functionality in FL5 adopts a
trigonal pyramidal geometry, a trigonal planar structure occurs
at the lone-pair-carrying N atom in FL5-NO. This result
indicates that the amino functionality becomes rehybridized to
a sp2 center upon nitrosation. The MO diagrams for FL5 and
FL5-NO given in Figure 9a illustrate the electronic difference
between the two lone pair orbitals. The energy of the donor
lone pair orbital in FL5 (HOMO-1) is -4.665 eV, being
essentially isoenergetic with the HOMO. In FL5-NO, however,
the analogous donor orbital (HOMO-14) displays an orbital
energy of -6.400 eV, 1.82 eV lower in energy than the HOMO.
The HOMO and LUMO orbitals of FL5-NO are largely
identical in character and energy with those computed for FL5.
(50) Chanon, M.; Hawley, M. D.; Fox, M. A. In Photoinduced Electron Transfer;
Fox, M. A., Chanon, M., Eds.; Elsevier: Amsterdam, 1988; Part A, pp
1-60.
(51) Callan, J. F.; de Silva, A. P.; Magri, D. C. Tetrahedron 2005, 61, 8551-
8588.
(52) Tanaka, K.; Miura, T.; Umezawa, N.; Urano, Y.; Kikuchi, K.; Higuchi,
T.; Nagano, T. J. Am. Chem. Soc. 2001, 123, 2530-2536.
(53) Chang, J. H.; Choi, Y. M.; Shin, Y.-K. Bull. Kor. Chem. Soc. 2001, 22,
527-530.
9
14372 J. AM. CHEM. SOC. VOL. 128, NO. 44, 2006

Nitric Oxide Detection in Aqueous Solution
A R T I C L E S
based complexes that could serve as biosensors for nitric oxide.
The Cu(II) species Cu(FLn) generated in situ by reaction of FLn
with CuCl2 exhibit a significant increase in fluorescence upon
addition of NO in pH 7.0 buffered aqueous solution. Turn-on
emission of Cu(FL5) by NO occurs by reduction of Cu(II) to
Cu(I), forming NO+, which nitrosates FL5 to create FL5-NO,
as revealed by spectroscopic and product analyses of the
reaction. Dissociation from the copper center leads to emission
turn-on. DFT calculations of FL5-NO relative to FL5 also
indicate how N-nitrosation of FL5 triggers fluorescence en-
hancement. The Cu(FLn) probes facilitate direct NO detection,
using metal coordination chemistry with fluorescence enhance-
ment at pH 7.0, when excited at a visible wavelength. The
fluorescence response to NO of Cu(FLn) is specific over other
-
biologically relevant reactive species such as O2-, H2O2, NO2
,
NO3-, HNO, ONOO-, and ClO-. Therefore, the Cu(II) fluo-
rescent complexes can directly and specifically detect NO at a
physiologically relevant pH.
Figure 10. Isosurface plots (isodensity value ) 0.5 au) of the key molecular
orbitals in the FL5/FL5-NO system. The HOMO, LUMO, and HOMO-1
of FL5 and HOMO-14 of FL5-NO are illustrated.
Although these compounds represent a significant advance,
there are some important improvements that need to be
addressed in future work. The binding of NO is not reversible,
so strictly speaking the molecules are detectors of NO, not true
sensors. A reversible sensor for biological NO would be a
valuable contribution. Second, the fluorescein construct needs
to be modified to allow the molecules to become trapped in
cells and not to diffuse out, an important property for imaging
live brain slices.54-56 Finally, it would be useful to have a
ratiometric version, in which NO binding causes a change in
the fluorescence excitation or emission wavelength, as well as
derivatives that can be positioned in a site-specific manner in
living tissue.
These most important orbitals are depicted in Figure 10. For a
fluorophore to be quenched by an electron donor group, the
MO associated with that donor must lie energetically within
the region of the frontier orbitals associated with the fluorophore
(Figure 9b), in addition to having the correct orientation to
couple effectively with the acceptor orbital, to maximize the
electron-transfer efficiency. For FL5, the nitrogen atom lone pair
is conjugated into the quinoline group, where an antibonding
π-interaction helps to increase its energy, as can be seen from
the node between these two groups [FL5-(HOMO-1), Figure
10]. Excitation of the fluorophore causes a reordering of these
MOs due to electronic relaxation in the excited state, which
does not affect the energy of the donor orbital to a significant
extent. Electron transfer from the donor orbital into the now
lower-lying, half-occupied MO of the fluorophore brings about
the observed quenching (Figure 9b).
In the case of FL5-NO, the donor nitrogen atom has a nearly
planar geometry with a normal vector in the plane of the
quinoline group [FL5-NO-(HOMO-14), Figure 10], thus
preventing the same out-of-phase π-interaction conjugation
present in FL5. Instead, the nitrogen atom lone pair interacts
with the π-system of the formally cationic NO+ group,
producing a stabilized MO at a significantly lower energy. This
MO lies well below the frontier orbitals of the fluorophore and
cannot serve as an electron donor for the purposes of quenching.
Although it is impossible to determine quantitatively the orbital
energies of the excited state within the framework of DFT, it is
clear that lowering the donor orbital energy in the ground state
will translate analogously to the excited state, offering a
qualitative rationale for the experimentally observed fluores-
cence “turn-on” of the N-nitrosated species.
Acknowledgment. This work was supported by grants from
the National Science Foundation (S.J.L.) and the National
Institutes of Health (M.-H.B.). The NSF is acknowledged for
computational resources (0116050 at IU). M.H.L. thanks the
Martin Family Society at MIT for fellowship funding. Spec-
troscopic instrumentation at the MIT DCIF is maintained with
funding from NIH Grant 1S10RR13886-01 and NSF Grants
CHE-9808063, DBI9729592, and CHE-9808061. We thank Dr.
Rodney P. Feazell for assistance with the X-ray structural work
and Dr. Elizabeth M. Nolan for helping with ligand synthesis
and providing FL4 (QZ1).
Supporting Information Available: Cartesian coordinates of
computed structures, tables of complete bond distances and
angles, and other X-ray crystallographic files (CIF files). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA064955E
(54) Tsien, R. Y. Nature 1981, 290, 527-528.
(55) Tsien, R. Y.; Pozzan, T.; Rink, T. J. J. Cell Biol. 1982, 94, 325-334.
(56) Woodroofe, C. C.; Masalha, R.; Barnes, K. R.; Frederickson, C. J.; Lippard,
S. J. Chem. Biol. 2004, 11, 1659-1666.
Summary and Perspective. Five fluorescein-based ligands
(FLn, n ) 1-5) were synthesized as a framework for Cu(II)-
9
J. AM. CHEM. SOC. VOL. 128, NO. 44, 2006 14373