A fluorogenic probe for the copper(I)-catalyzed azide-alkyne ligation reaction: Modulation of the fluorescence emission via 3(n, π*)-1(π,π*) inversion

Article abstract of DOI:10.1021/ja049684r

Chemoselective ligation reactions represent a powerful approach for labeling of proteins or small molecules in a biological environment. We report here a fluorogenic probe that is activated by click chemistry, a highly versatile bio-orthogonal and chemose

Full text of DOI:10.1021/ja049684r

Published on Web 06/30/2004
A Fluorogenic Probe for the Copper(I)-Catalyzed Azide-Alkyne Ligation
3
1
Reaction: Modulation of the Fluorescence Emission via (n,π*)- (π,π*)
Inversion
Zhen Zhou and Christoph J. Fahrni*
School of Chemistry and Biochemistry, Georgia Institute of Technology, 770 State Street, Atlanta, Georgia 30332
Received January 17, 2004; E-mail: fahrni@chemistry.gatech.edu
Chemoselective ligation reactions represent a powerful approach
for labeling of proteins or small molecules without the need of
laborious protection and deprotection steps. Recently, two highly
selective bioorthogonal ligation strategies have been introduced,
both of which utilize the azide moiety as the functional group. The
first strategy is based on an intramolecular variant of the Staudinger
reaction and has been applied to the labeling of metabolically
Scheme 1
1
engineered cell surfaces. The second approach is based on the
Cu(I)-catalyzed, Huisgen-like annealation of azides with terminal
alkynes,2 a very robust transformation that has been successfully
,3
4
used for the fluorescent labeling of virus capsid proteins, the cell-
5
surface labeling of Escherichia coli, and the activity-based protein
profiling in live mice.6
Here we report a fluorogenic probe (1), that takes advantage of
the electronic structure changes associated with the triazole ring
formed in the Cu(I)-catalyzed azide-alkyne ligation reaction
7
(
Scheme 1). We hypothesized that the electron-donating properties
Table 1. Photophysical Properties of Compounds 1 and 3
of the triazole ring might effectively modulate the emission response
of an electronically coupled coumarin. The fluorescence quantum
yield of coumarin is strongly influenced by substituents in the 7-
position, which control the relative energy ordering of the two
lowest excited states: a nonemissive (n,π*) state, which originates
from the carbonyl-oxygen lone pair, and an emissive locally excited
c
d
e
absorption
λ
emission
τ
f
k
r
k
nr
max [nm]a
λmax [nm]
Φ
b
[ns]
[108s-1
]
[109 s-1
>9.8
0.81
]
f
1
3
325 (7100)
328 (10 900)
388
415
0.014
0.25
<0.1
0.92
>1.4
2.7
a
Lowest-energy band, extinction coefficient ꢀ ([M-1 cm-1]) in paren-
8
state with (π,π*) character. Because substitution with an electron
theses. b Fluorescence quantum yield (quinine sulfate as standard). Fluo-
c
d
donor increases the charge-transfer character only of the (π,π*)
state but not the (n,π*) state, the former will be more effectively
stabilized in a polar solvent such as water. If the emissive (π,π*)
state lies then sufficiently below the energy level of the nonemissive
rescence lifetime (instrument resolution 100 ps). Radiative rate constant
(
kr ) Φf/τf). e Nonradiative rate constant (knr ) (1 - Φf)/τf).
(n,π*) state, the molecule should fluoresce with high quantum yield.
On the basis of this rational we designed fluorophore 1, in which
the reactive alkyne moiety is attached at the 7-position of the
coumarin framework. To increase the water solubility of the probe,
we further functionalized the 4-position with an electronically
decoupled carboxylate group (Scheme 1).
The photophysical data for 1 and its ligation product 3 are
compiled in Table 1. Under physiological conditions (pH 7.2, 0.1
M KCl) compound 1 exhibits a low fluorescence quantum yield of
1
.4% (Figure 1, right). Upon reaction with azide 2 the quantum
Figure 1. Absorption and fluorescence emission spectra (λex 320 nm) of
compound 1 (s) and 3 (- - -) (30 µM, 10 mM PIPES pH 7.20,
25 °C).
yield increases to 25%, and consistent with increased charge-transfer
character, the peak emission undergoes a bathochromic shift by 27
nm. The UV-vis spectrum of 1 shows two intense absorption
bands, which are reminiscent of symmetry-allowed π f π* tran-
sitions (Figure 1, left). Time-resolved emission spectroscopy of 3
revealed a monoexponential decay kinetics with a lifetime of 0.92
ns and a large radiative rate constant, both of which are in agreement
lifetimes of 1 (<7 ns) and 3 (3.7 ns) are similarly short, suggesting
1
that emission occurs for both compounds from a (π,π*) and not
1
(n,π*) state.8
While the kinetics for the [3 + 2] cycloaddition reaction of
1
with emission from the lowest (π,π*) state. Furthermore, the time-
alkynes with azides is slow, the reaction rate increases strongly in
2
resolved measurements indicate that the low quantum yield of 1 is
primarily due to a large increase of the nonradiative rate constant,
while the radiative deactivation rate remains essentially identical
within the experimental error (Table 1). Consequently, the radiative
the presence of Cu(I). Interestingly, Cu(I) complexes of tris-triazole
4
ligands such as 4 are highly effective catalysts for this reaction. If
5 µM of 1 is converted in the presence of 100 µM azide 2 and 60
µM of the Cu(I) catalyst, the reaction is complete in less than 5
8862
9
J. AM. CHEM. SOC. 2004, 126, 8862-8863
10.1021/ja049684r CCC: $27.50 © 2004 American Chemical Society

C O MMU N I C A T I O N S
Figure 2. Kinetics for the reaction of 1 with 2 at pH 7.2 (10 mM PIPES,
3
7 °C). Conditions: 60 µM 1, 20 µM 2, in situ-formed catalyst: 60 µM 4,
6
0 µM CuSO4, 300 µM ascorbic acid as reducing agent. Control without
azide 2. Inset: identical conditions, but 5 µM 1, 100 µM 2 (λex 350 nm,
λem 420 nm).
Table 2. State Energies and Dipole Momentsa
Figure 3. Excited-state energy levels for 1a and 3a based on INDO1/s
calculations. The surface plots represent densities of the MOs with the largest
contribution to the CI-eigenvectors (dark- and light-shaded surfaces refer
to occupied and unoccupied MO densities, respectively).
1a
3a
b
GS dipole moment [D]
π-π* exptl [eV]
9.3 (7.2)
3.82
3.86 (3.94)
12.4 (5.1)
3.92 (3.70)
4.6 (2.3)
16.5 (13.0)
3.78
3.70 (3.90)
21.6 (14.1))
4.08 (3.80)
12.40 (8.1)
c
b
π-π* energy (S1) [eV]
1
b
lies now sufficiently above the (π,π*) state, rendering nonradiative
dipole moment [D]
b
deactivation energetically unfavorable via intersystem crossing to
n-π* energy (S2) [eV]
b
3
dipole moment [D]
the (n,π*) state.
In conclusion, the electron-donating properties of the triazole
ring formed in the azide-alkyne ligation reaction can be effectively
utilized for the design of a chemoselective fluorogenic probe that
may find a range of applications in biology, analytical chemistry,
or material science.
a
Based on INDO1/s CIS calculations using ground-state geometries
optimized at the (B3LYP/6-31G*) level. b SCRF solvent model with a0 )
.5 and 4.9 Å cavity radii for 1 and 3, respectively. Gas-phase values in
4
c
parentheses. 0.1 M KCl, pH 7.20 (PIPES 10 mM), 25 °C.
min. To evaluate the ligation kinetics under conditions that would
be more typical for biological labeling applications, we reacted 20
µM of azide 2 in the presence of 60 µM dye 1. Within ap-
proximately 30 min more than 90% of the azide had been converted,
and the reaction is completed in less than 1 h (Figure 2).
Acknowledgment. Financial support from the Georgia Institute
of Technology, IBM (SUR Grant), and NIH (R01GM67169) is
gratefully acknowledged. We thank Mostafa El-Sayed and K. Barry
Sharpless for their helpful advice, and David Bostwick for mass
spectral data.
To gain further insights into the nature of the fluorescence
switching mechanism we performed semiempirical quantum me-
Supporting Information Available: Experimental details, synthe-
sis, and characterization of compounds 1 and 3 and all intermediates.
This material is available free of charge via the Internet at http://
pubs.acs.org.
9
chanical calculations at the INDO1/s level of theory. The predicted
vertical excitation energies listed in Table 2 are based on the
optimized ground-state geometries of the model compounds 1a and
3a containing methyl groups in place of the succinyl- and benzyl-
substituents. While the calculated gas-phase excitation energies are
expectedly higher than the experiment by 0.1-0.2 eV, the SCRF
solvent model10 implemented in the ArgusLab software package9
provided a good approximation for the state energies in water.
In agreement with the experimental data, the calculations predict
References
(
1) Saxon, E.; Luchansky, S. J.; Hang, H. C.; Yu, C.; Lee, S. C.; Bertozzi, C.
R. J. Am. Chem. Soc. 2002, 124, 14893; Saxon, E.; Bertozzi, C. R. Science
2000, 287, 2007.
(
2) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. Angew.
Chem., Int. Ed. 2002, 41, 2596.
1
(3) Kolb, H. C.; Sharpless, K. B. Drug DiscoVery Today 2003, 8, 1128.
that emission in 1 occurs from a (π,π*) state (Figure 3). However,
(4) Wang, Q.; Chan, T. R.; Hilgraf, R.; Fokin, V. V.; Sharpless, K. B.; Finn,
1
1
the (n,π*) state lies only slightly above the (π,π*) state, suggesting
M. G. J. Am. Chem. Soc. 2003, 125, 3192.
3
1
(5) Link, A. J.; Tirrell, D. A. J. Am. Chem. Soc. 2003, 125, 11164.
a lower (n,π*) energy compared to (π,π*). According to the El-
(6) Speers, A. E.; Adam, G. C.; Cravatt, B. F. J. Am. Chem. Soc. 2003, 125,
Sayed rule11 such an energy ordering results in fast intersystem
4686.
1
3
crossing from the (π,π*) state to the nonemissive (n,π*) state,
and typically yields a large nonradiative deactivation rate, which
has been indeed observed for 1 (Table 1). In contrast, the triazole-
substituent in 3 greatly increases the charge-transfer character of
(7) For a fluorogenic probe activated by the Staudinger reaction see: Lemieux,
G. A.; de Graffenried, C. L.; Bertozzi, C. R. J. Am. Chem. Soc. 2003,
125, 4708.
(8) Seixas de Melo, J. S.; Becker, R. S.; Ma c¸ anita, A. L. J. Phys. Chem.
1994, 98, 6054.
(9) Thompson, M. A. ArgusLab 3.1; Planaria Software LLC, Seattle, WA,
1
1
the (π,π*) excited state, but has the opposite effect on the (n,π*)
http://www.arguslab.com, 2003.
(
10) Karelson, M. M.; Zerner, M. C. J. Phys. Chem. 1992, 96, 6949.
11) Lower, S. K.; El-Sayed, M. A. Chem. ReV. 1966, 66, 199; El-Sayed, M.
A. Acc. Chem. Res. 1968, 1, 8.
state as indicated by the differences of ground- and excited-state
(
1
1
dipole moments (Table 2). As a consequence the (n,π*)- (π,π*)
energy gap increases and yields an inverted energy ordering of the
(12) Seixas de Melo, J.; Fernandes, P. F. J. Mol. Struct. 2001, 565-566, 69.
1
3
(
π,π*) and (n,π*) states. Due to the inherently small singlet-
12
3
triplet splitting of (n,π*) states (∆EST ≈ 0.1 eV), the (n,π*) level
JA049684R
J. AM. CHEM. SOC.
9
VOL. 126, NO. 29, 2004 8863