Diheteroarylethenes as thermally stable photoswitchable acceptors in photochromic fluorescence resonance energy transfer (pcFRET)

Article abstract of DOI:10.1021/ja016969k

We have employed diheteroarylethenes as acceptors for photochromic FRET (pcFRET), a technique introduced for the quantitative determination of fluorescence resonance energy transfer (FRET). In pcFRET, the fluorescent emission of the donor is modulated by cyclical transformations of a photochromic acceptor. Light induces a reversible change in the structure and, concomitantly, in the absorption properties of the acceptor. Only the closed forms of the selected diheteroarylethenes 2a and 2b have an absorption band overlapping the emission band of the donor, 1. The corresponding variation in the overlap integral (and thus critical transfer distance R_o) between the two states provides the means for reversibly switching the process of FRET on and off, allowing direct and repeated evaluation of the relative changes in the donor fluorescence quantum yield. The diheteroarylethenes demonstrate excellent stability in aqueous media, an absence of thermal back reactions, and negligible fatigue. The equilibration of these systems after exposure to near-UV or visible light follows simple monoexponential kinetics. We developed a general conceptual scheme for such coupled photochromic-FRET reactions, allowing quantitative interpretations of the photostationary and kinetic data, from which the quantum yields for the cyclization and cycloreversion reactions of the photochromic acceptor were calculated.

Full text of DOI:10.1021/ja016969k

Published on Web 06/04/2002
Diheteroarylethenes as Thermally Stable Photoswitchable
Acceptors in Photochromic Fluorescence Resonance Energy
Transfer (pcFRET)
Luciana Giordano,† Thomas M. Jovin,‡ Masahiro Irie,§ and
Elizabeth A. Jares-Erijman*,†
Contribution from the Departamento de Qu´ımica Orga´nica, Facultad de Ciencias Exactas y
Naturales, UniVersidad de Buenos Aires, PROPLAME-CONICET, Ciudad
UniVersitaria-Pabello´n II, 1428 Buenos Aires, Argentina, Department of Molecular Biology,
Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, D-37077 Go¨ttingen, Germany,
and Department of Chemistry and Biochemistry, Graduate School of Engineering, Kyushu
UniVersity, Higashi-ku, 812, Japan
Received August 30, 2001
Abstract: We have employed diheteroarylethenes as acceptors for photochromic FRET (pcFRET), a
technique introduced for the quantitative determination of fluorescence resonance energy transfer (FRET).
In pcFRET, the fluorescent emission of the donor is modulated by cyclical transformations of a photochromic
acceptor. Light induces a reversible change in the structure and, concomitantly, in the absorption properties
of the acceptor. Only the closed forms of the selected diheteroarylethenes 2a and 2b have an absorption
band overlapping the emission band of the donor, 1. The corresponding variation in the overlap integral
(and thus critical transfer distance Ro) between the two states provides the means for reversibly switching
the process of FRET on and off, allowing direct and repeated evaluation of the relative changes in the
donor fluorescence quantum yield. The diheteroarylethenes demonstrate excellent stability in aqueous media,
an absence of thermal back reactions, and negligible fatigue. The equilibration of these systems after
exposure to near-UV or visible light follows simple monoexponential kinetics. We developed a general
conceptual scheme for such coupled photochromic-FRET reactions, allowing quantitative interpretations
of the photostationary and kinetic data, from which the quantum yields for the cyclization and cycloreversion
reactions of the photochromic acceptor were calculated.
Introduction
with a spatial resolution by far exceeding the optical diffraction
limit. The photophysical behavior of extrinsic or intrinsic (e.g.,
genetically expressed) fluorophores also reflects conformational
changes and perturbations of the microenvironment.
Fluorescence resonance energy transfer (FRET) is a physical
process by which energy is transferred nonradiatively from an
excited molecular fluorophore (donor) to another chromophore
(acceptor) via long-range dipole-dipole coupling.1a-l The FRET
efficiency varies with the 6th power of the donor-acceptor
separation over the range of ∼1.5-10 nm, a range of distances
corresponding to most biologically significant molecular interac-
tions within cells. Applied in the fluorescence microscope, FRET
provides a tool for resolving molecular interactions (proximities)
There are two fundamental problems associated with quan-
titative FRET determinations in the microscope: (i) The local
donor and acceptor concentrations (densities) of donor and
acceptor moieties located on separate molecules can vary greatly.
Thus, the formalism for estimating the FRET efficiency E must
be appropriate for arbitrary stoichiometries. (ii) For some
preparations, for example, living cells undergoing a signaling,
metabolic, or other time-dependent process, continuous methods
of observation are required. Condition (i) can be met by various
combinations of the donor and sensitized emission signals, but
at least three separate images are required.1f,2 Alternatively, E
can be evaluated via the quenching of the donor alone, for
example, from determination of the fluorescence lifetime
(FLIM1c,e,h,j,3a-h) or employing various strategies based on
systematic photobleaching of the donor and/or acceptor
(pbFRET).1h,2,4a-h In the simple acceptor pbFRET procedure,
* To whom correspondence should be addressed. E-mail:
eli@qo.fcen.uba.ar.
† Universidad de Buenos Aires.
‡ Max Planck Institute for Biophysical Chemistry.
§ Kyushu University.
(1) (a) Fo¨rster, T. Naturwissenschaften 1946, 6, 166-175. (b) Fo¨rster, T.
Discuss. Faraday Soc. 1959, 27, 7-17. For recent reviews of relevance to
cell biology and imaging, see: (c) Centonze, V.; Takahashi, A.; Casanova,
E.; Herman, B. J. Histotechnol. 2000, 23, 229-234. (d) Chamberlain, C.;
Hahn, K. Traffic 2000, 1, 755-762. (e) Goedhart, J.; Gadella, T. W. J., Jr.
In Root Hairs: Cell And Molecular Biology; Ridge, R., Emons, A., Eds.;
Springer-Verlag: Tokyo, 2000; pp 65-94. (f) Gordon, G.; Berry, G.; Liang,
X.; Levine, B.; Herman, B. Biophys. J. 1998, 74, 2702-2713. (g)
Kenworthy, A.; Edidin, M. Methods Mol. Biol. 1999, 116, 37-49. (h)
Lippincott-Schwartz, J.; Snapp, E.; Kenworthy, A. Nat. ReV. Mol. Cell Biol.
2001, 2, 444-456. (i) Miyawaki, A.; Tsien, R. Methods Enzymol. 2000,
327, 472-500. (j) Selvin, P. Nat. Struct. Biol. 2000, 7, 730-734. (k)
Szo¨llo¨si, J.; Damjanovich, S.; Matyus, L. Cytometry 1998, 34, 159-179.
(l) Wouters, F.; Verveer, P.; Bastiaens, P. I. H. Trends Cell Biol. 2001, 11,
203-211.
the local (pixel-by-pixel) donor signal is measured prior to (IDA
)
(2) Jovin, T. M.; Arndt-Jovin, D. J. In Cell Structure and Function by
Microspectrofluometry; Kohen, E., Hirschberg, J. G., Ploem, J. S., Eds.;
Academic Press: London, 1989; pp 99-117.
9
10.1021/ja016969k CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 7481-7489
7481

A R T I C L E S
Giordano et al.
and following (ID) photodestruction of the acceptor.4a-f For the
case of total conversion (IDA f ID), the FRET efficiency E is
related to the fractional increase (â ) ID/IDA - 1; 0 e â e ∞)
or, if the final ID state serves as the reference, to the fractional
decrease (γ ) 1 - IDA/ID; 0 e γ e 1) in the donor signal.
The absorption spectra of both forms of 2a and 2b (Figure
1), determined according to the formalism and procedures
elaborated in this study, are given in Figure 2b and c,
respectively. The spectral bands of the closed forms of 2a and
2b in the visible are at 546 nm (ꢀ 12 000 M-1 cm-1) and 568
nm (ꢀ 11 000 M-1 cm-1), respectively. The closed and open
forms of 2a have isosbestic points at 248, 274, and 312 nm
and the corresponding forms of 2b at 233 and 290 nm. The
spectral constants are given in Figure 2.
Model compounds 3a and 3b were synthesized to evaluate
the performance of diheteroarylethenes as acceptors in FRET
(Figure 1). We selected Lucifer Yellow 1 (Figures 1 and 2a) as
the donor for the diheteroarylethenes because of the location
of its absorption maximum in a spectral region in which the
absorbances of both the closed and the open forms of the
diheteroarylethenes are minimal. In addition, the donor emission
spectrum of LY (Figure 2a) overlaps well the absorption of the
closed but not the open forms of 2a and 2b (Figure 2). 3a and
3b were designed with an aliphatic linker between the fluores-
cent donor and the photochromic acceptor moiety so as to
minimize interactions between the two. The photoconversion
difference spectra for 2a and 3a and for 2b and 3b, generated
by irradiation at 366 nm, confirmed that the donor absorption
properties did not change during the transition of the acceptor-
(s) between the open and closed forms. Conversely, the acceptor
difference spectra were not perturbed by the presence of the
donor.
The computed critical transfer distance Ro for the closed forms
of 3a and 3b was 38 and 35 Å, respectively. The corresponding
values for the open forms were 9 and 12 Å. According to an
energy minimization by Sybyl (AMBER 4.1 package, 3a:
Figure 1 bottom), the distance rDA between the central atom of
the LY and that of the acceptor was ∼20 Å in 3a and 22 Å in
3b, respectively. Thus, one would predict a FRET efficiency
(E+, see below for definition) of 0.98 for 3a and 0.94 for 3b
(closed forms), assuming an invariant rDA (eq 7).6 In the case
of the open forms, the corresponding computed FRET efficien-
cies (E-) were very low, 0.008 and 0.02, respectively.
Quantitative Formalism for the Photostationary State and
Photokinetics. The equilibrium state and kinetic processes
characterizing such a system are represented in Figure 3 and
by the following set of equations:
E ) (1 + â-1)-1 ) γ
(1)
However, the use of irreversible photobleaching of the
acceptor precludes reexamination of the same region at a later
time. Photochromic FRET overcomes this limitation by provid-
ing the required reference state (donor alone) in a local (single
pixel) and reversible manner. This is achieved by continued,
repeated, and light-driven on-off switching of the FRET
process. Our strategy is based on the use of photochromic
compounds as “programmable” acceptors for FRET.
A photochromic compound is characterized by the ability to
undergo a reversible transformation - in response to illumina-
tion at appropriate wavelengths - between two different
structural forms having different absorption (and, in some cases,
fluorescence) spectra. The two FRET probes (donor and
acceptor) are selected according to certain spectroscopic criteria,
including the key requirement that only one of the photochromic
forms of the acceptor has absorption properties rendering it
competent for the transfer of excitation energy from the donor.
As a consequence, the fluorescence emission (degree of quench-
ing) of the donor can be modulated reversibly by systematic
photochemical manipulation of the photochromic acceptor. That
is, the donor fluorescence is measured in the “presence” and
“absence” of the acceptor, the latter case corresponding to the
reference state described above.
We selected thermally stable thiophene-based dihet-
eroarylethenes as acceptors for pcFRET. We employed a
perfluorocyclopentene as the central bridging substructure
because of its demonstrated resistance to fatigue.5 The electron-
withdrawing property of the perfluoro substituents also promotes
the cyclization reaction.5
Results
Molecular Design, Syntheses, and Spectral Properties. The
diheteroarylethenes selected as photoswitchable acceptors for
FRET have closed forms with absorption maxima in the range
540-570 nm. The open and closed forms are both thermally
stable.
k-+
(DA-) y z (DA+)
(2)
k+-
(3) (a) Bastiaens, P. I.; Squire, A. Trends Cell Biol. 1999, 9, 48-52. (b) Gadella,
T. W. J., Jr.; Jovin, T. M. J. Cell Biol. 1995, 129, 1543-1558. (c) Harpur,
A. G.; Wouters, F. S.; Bastiaens, P. I. H. Nat. Biotechnol. 2001, 19, 167-
169. (d) Herman, B.; Wang, X. F.; Wodnicki, P.; Gordon, G. W.;
Seongwook, K.; Diliberto, P. A.; Periasamy, A. In CLEO ′95. Summaries
of Papers Presented at the Conference on Lasers and Electrooptics (IEEE
No. 95CH35800), Opt. Soc. Am. 1995; pp 274-275. (e) Murata, S.; Herman,
P.; Lakowicz, J. R. Cytometry 2001, 43, 94-100. (f) Schneider, P.; Clegg,
R. ReV. Sci. Instrum. 1997, 68, 4107-4119. (g) Szmacinski, H.; Lakowicz,
J. R. Sens. Actuators, B 1995, 29, 16-24. (h) Verveer, P.; Wouters, F.;
Reynolds, A.; Bastiaens, P. I. H. Science 2000, 290, 1567-1570.
(4) (a) Kenworthy, A. K.; Edidin, M. J. Cell Biol. 1998, 142, 69-84. (b) Varma,
R.; Mayor, S. Nature 1998, 394, 798-801. (c) Wouters, F.; Bastiaens, P.
I. H.; Wirtz, K. W.; Jovin, T. M. EMBO J. 1998, 17, 7179-7189. (d)
Bastiaens, P. I. H.; Jovin, T. M. In Cell Biology: A Laboratory Handbook,
2nd ed.; Celis, J. E., Ed.; Academic Press: New York, 1997; Vol. 3, pp
136-146. (e) Bastiaens, P. I. H.; Wouters, F. S.; Jovin, T. M. Proc. of the
2nd Hamamatsu Int. Symp. on Biomol. Mechanisms and Photonics;
Research Foundation for Opto-Sci. and Technology: Hamamatsu, Japan,
Feb. 20-22, 1997; pp 77-82. (f) Bastiaens, P. I. H.; Majoul, I.; Verveer,
P.; So¨ling, H. D.; Jovin, T. M. EMBO J. 1996, 15, 4246-4253. (g) Young,
R.; Arnette, J.; Roess, D.; Barisas, B. Biophys. J. 1994, 67, 881-888. (h)
Szabo, G., Jr.; Pine, P.; Weaver, J.; Kasari, M.; Aszalos, A. Biophys. J.
1992, 61, 661-670.
DA-
ex
DA+
ex
k-+ ) (k
+ kDexE-)Q-+
k+- ) (k
+ kDexE+)Q+- (3)
*
*
Q-+ ) [1 + kDd A /kopfcl
]
Q+- ) [1 + kDd A /kclfop
]
(4)
(5)
-1
-1
-
+
*
*
E- ) [1 + kDd /kFRET
]
E+ ) [1 + kDd A /kFRET
]
DA- -1
DA+ -1
-
+
in which DA- and DA+ denote the donor paired with the open
(-) or closed (+) forms of the acceptor, respectively. The rate
constants determining the kinetics of the system are those for
excitation, that is, absorption of light (kDA-, keDxA+, keDx; see
ex
Figure 3 for definitions of other terms in eqs 4 and 5). In this
(6) Realistically, one would envision a dynamic distribution of distances during
the excited-state lifetime of the donor because of conformational flexibility
of the spacer linking the donor and acceptor; this effect would serve to
increase E even further.
(5) Irie, M. Chem. ReV. 2000, 100, 1685-1716.
9
7482 J. AM. CHEM. SOC. VOL. 124, NO. 25, 2002

Diheteroarylethenes as Photoswitchable Acceptors
A R T I C L E S
Figure 1. Chemical structures of model compounds for pcFRET. Photochromic acceptors: 2a and 2b. The donor, Lucifer Yellow (LY) cadaverine (LYC,
1), is covalently bound to the diheteroarylethene photochromic molecules in 3a and 3b. Initially, the diheteroarylethenes are in the open form (middle left).
Upon near-UV irradiation, the photochromic compound is converted to the closed form (right), which now acts as an energy acceptor for the donor. Both
states are thermally stable. Lower panel: energy minimized structures of the open and closed forms of 3a (see text for details).
(diheteroarylethene) system, the thermal back reaction (-) f
(+) is negligible. As a consequence, both the off and the on
states are stable in the absence of light.
also monitored continuously by absorption at a single wave-
length. Monoexponential fitting yielded equilibration times
(k-eq1) of 5-30 s.
Photochromic Conversion Analyzed by Absorption Kinet-
Compound 2a was also converted by irradiation at 313 nm.
Except for scaling factors, the difference spectra for the
photostationary states of 2a and 2b achieved at 313 and 366
nm were indistinguishable (Figures S1c and S2a,b). However,
there was a significant increase in the conversion efficiency (Rps)
from 0.33 to 0.65 upon passing from 366 to 313 nm irradiation.
The degree of conversion in these reactions was established
by 1H NMR spectroscopy before and after irradiation with UV
light in the photostationary state. Integration of the signals of
the aliphatic methyl groups and selected aromatic resonances
of the spectra of 3a and 3b were evaluated before and after
1
ics and H NMR. The photochromic conversion efficiencies
of 2a, 3a, 2b, and 3b were determined in methanol. The kinetics
of photoconversion of 2a upon irradiation at 366 nm are shown
in Figure 4a. Two new bands emerged at 412 and 548 nm and
attained maximum absorbance at 40 min (photostationary state).
Compound 3a displayed identical behavior (data not shown),
confirming that the coupling of 2a to 1 did not alter the
photoconversion process. The photogenerated closed ring form
was stable at room temperature in the dark for at least 120 h.
The time-dependent absorption spectra during a cyclorever-
sion of 2a are shown in Figure 4b. Once in the photostationary
state induced by near-UV light (panel a), 2a was exposed to
green light, and the spectra were recorded every 10 s. After 50
s, the depletion of the 548 nm band was complete, and the 266
nm band recovered its original intensity. The identity of the
final and original spectra of the open form confirmed the
completion of the back conversion and the absence of photo-
degradation.
photoconversion. On the basis of these measurements, Rps
0.85 was determined for 3b.
)
Six singlet resonances corresponding to methyl groups
suggested the presence of two conformations of 2a, probably
parallel and antiparallel, as reported previously for other
diheteroarylethenes.7 The integration of the methyl group
resonances indicated an almost equal population of both
conformations. Irradiation with UV light to the photostationary
state led to a decrease in the integrated signals of all the methyl
groups corresponding to the open form. A new group of four
singlets between 2.15 and 2.00 ppm appeared and was assigned
to methyl resonances corresponding to two configurations of
The light-induced conversion between the open and closed
forms of 2b by irradiation at 340 nm is shown in Figure 4c. A
photostationary state (Rps ) 0.99) was achieved after 30 min.
Cycloreversion to the initial open form (Figure 4d) with visible
light (>520 nm) occurred within 2 min. The back reactions of
2a, 2b, 3a, and 3b induced by irradiation at >520 nm were
(7) Matsuda, K.; Irie, M. J. Am. Chem. Soc. 2000, 122, 8309-8310.
9
J. AM. CHEM. SOC. VOL. 124, NO. 25, 2002 7483

A R T I C L E S
Giordano et al.
Figure 3. Photophysical-photochromic scheme incorporating donor-
acceptor FRET (pcFRET). The central region comprises the ring closure
and opening reactions of the photochromic acceptor. The external pathways
represent the intervention of the donor via excitation and resonance energy
transfer reactions. (-) refers to states devoid of significant FRET between
donor and acceptor and (+) to states with very high FRET. In the present
system, the (-) state is the open molecular form and the (+) the closed
form (Figure 1). Because of steady-state conditions, the overall intercon-
version between the (-) and (+) forms can be represented by an apparent
first-order reaction, designated in the scheme by the pair of vertically
oriented arrows. The function of near-UV and visible light as reactants is
denoted by hc/λi (i ) 1, near-UV; i ) 2, visible).
The Rps values derived from the NMR measurements permit-
ted computation of the pure open and closed absorption spectra
of 2a and 2b (Figure 2b,c) according to eq S6a. The results of
these calculations, using data for the photostationary states
induced by 313 and 366 nm, were the same within experimental
error. The photoconversion efficiencies are summarized in
Table 1.
FRET Efficiencies for Open and Photostationary State
of 3a and 3b. The changes in the photostationary state
absorption spectra recorded after irradiation at 366 and 313 nm
were reflected in the fluorescence signals. Fluorescence emission
spectra of 3a and 3b were acquired for the initial (open ring
form) and the photostationary states achieved after irradiation
at 366 nm. 3a was quenched by 33% after irradiation at 366
nm and 65% after irradiation at 313 nm. The quenching of 3b
after exposure to 320 nm was 84%. From the conversion of 3a,
a value of E ) 1 was computed for the diheteroarylethene closed
form. Similarly, R was 0.85 for 3b with an E ) 1 for the closed
form. Irradiation for 60 s with visible light led to a complete
recovery of the initial fluorescence signal (E ) 0).
Frequency and time domain fluorescence lifetime analyses
of the donor 1 were carried out on the initial state (open form
of the diheteroarylethene acceptor) and the photostationary state
generated by irradiation at 366 nm. In both cases, a single
component sufficed for the analysis of the data (ø2 < 1.2 and
symmetrical residuals). The fluorescence lifetimes of 3a and
3b in the initial (open form) were 10.2 ( 0.4 and 9.5 ( 0.6 ns,
respectively, with no significant changes upon conversion to
the closed form (at 366 nm for 3a and 340 nm for 3b) and after
regeneration of the original open form.
Figure 2. Spectral properties for the completely off (-) and on (+) states
(open and closed forms of 2a and 2b, respectively) in pcFRET. (a)
Absorption (-O-) and emission (-b-) properties of the donor 1;
fluorescence excitation at 430 nm. (b) Absorption spectra of 2a in the open
(-O-) and closed (-b-) forms; kernel of the overlap integral in the open
(-0-) and closed (-9-) forms of 2a. (c) Absorption spectrum of the
open (-O-) and closed (-b-) forms of 2b; kernel of the overlap integral
in the open (-0-) and closed (-9-) forms of 2b. Peaks of 2a spectra:
open form, 226, 266, 298 nm; closed form, 230, 294, 413, 546 nm. Peaks
of 2b spectra: open form, 264 nm; closed form, 314, 424, 568 nm. Spectra
calculated from eq S6 based on extinction coefficients for the closed (+)
forms: 2a, ꢀ546 ) 12 000 M-1 cm-1; and 2b, ꢀ156cm8 ) 11 000 M-1 cm-1. In
1cm
the initial (-) state, there is no substantial overlap between the donor
emission spectrum and the absorption spectrum of the open form of the
diheteroarylethene. Near-UV irradiation (320-380 nm) generates the closed
(+) form, whose absorption spectrum overlaps well with the donor emission,
thereby potentiating FRET. Subsequent exposure to green light (500-580
nm) leads to a reversion of the closed to the open form, turning off the
FRET process. Applying eq S6 to the spectra of panels a and b, the following
relationships for 2a and 2b were derived, for which the selected reference
isosbestic points were 312 nm (2a) and 290 nm (2b). 2a, Rps ) -0.0048
Cyclic Modulation of Acceptor Absorption and Donor
Fluorescence. Absorption and fluorescence spectra determined
for the initial open form after several cycles of irradiation with
UV light and after irradiation with green visible light are shown
in Figures 5 and 6. The cyclic change in acceptor absorption
(Figure 5) and concomitant change in donor fluorescence
intensity and thus in FRET efficiency (Figure 6) induced by
+ 1.219A546/A312, ꢀ-546 ) 39, ꢀ312 ) 14 700, ꢀ+546 ) 12 000; 2b, Rps
)
ps
-+
-0.029 +ps1.390Ap5s68/A290, ꢀ5-68 ) 310, ꢀ290 ) 14 900, ꢀ+568 ) 11 000 (all ꢀ
ps
-+
in units of M-1 cm-1).
the closed form. The integrated signals indicated a 33%
conversion to the closed ring form.
9
7484 J. AM. CHEM. SOC. VOL. 124, NO. 25, 2002

Diheteroarylethenes as Photoswitchable Acceptors
A R T I C L E S
Figure 4. Time-resolved spectroscopic monitoring of the forward and backward photochromic interconversions between open and closed forms of the
diheteroarylethenes. (a) Absorption spectra of 2a undergoing photoconversion from the initial (-) form upon exposure to 366 nm light (-O-; 0.3 mW, 2
mW cm-2); 5 min intervals between spectra. After 50 min, the maximal change was achieved at 548 nm; spectra of the photostationary state (-b-). Inset:
A548 as a function of irradiation time (-0-). (b) Back photoconversion from the photostationary state of 2a produced by irradiation at >520 nm (-b-; 4
mW, 27 mW cm-2); interval between curves, 10 s. After 50 s, the spectrum of the original open (-) form was recovered (-O-). Inset: A548 as a function
of irradiation time (-0-). (c) Absorption spectra of 2b undergoing photoconversion from the initial (-) form upon exposure to 340 nm light (-O-; 16
mW, 107 mW cm-2); 5 s intervals between spectra. After 40 s, the maximal change was achieved at 568 nm; spectra of the photostationary state (-b-).
Inset: A568 as a function of irradiation time (-0-). (d) Back photoconversion from the photostationary state of 2b produced by irradiation at >520 nm
(-b-; 40 mW, 80 mW cm-2); interval between curves, 10 s. After 220 s, the spectrum of the original open (-) form was recovered (-O-). Inset: A568
as a function of irradiation time (-0-). The solid curves in the insets represent monoexponential fits according to the photokinetic formalism. The analyses
of these and other reactions are given in Table 1.
Table 1. Photochromic Conversion Efficiencies (from Absorption and 1H NMR) and Photokinetic Parameters (from Absorption)a
irrad.
mW cm
λirr
nm
ꢀ-
M-1 cm-1
ꢀ+
ꢀLY
M-1 cm-1
kλex-
s-1
kλex+
s-1
keq
s-1
Rps
abs
Rps
NMR
-2
cmpd
M-1 cm-1
Q-+
Q
+-
2a
2
110
27
3.7
110
27
110
80
110
80
366
340
548
366
340
548
340
568
340
568
410
1800
40
3200
5000
12 100
3200
5000
12 100
10 500
11 000
10 500
11 000
0.002
0.49
0.004
0.004
0.52
0.01
1.17
0.11
1.1
0.018
1.36
1.34
0.033
1.45
1.35
3.1
0.002
0.16
0.069
0.005
0.20
0.069
0.20
0.03
0.21
0.036
0.38
0.92
0.31
0.31
0.065
0.025
0.05
0.34
0.33
3a
410
2600
2900
0.28
0.76
0.31
0.29
0.05
1800
90
4000
310
4000
350
0.029
0.05
0.002
0.008
0.006
0.009
2b
3b
0.97
0.89
0.17
0.17
3.8
2900
2.9
3.8
0.85
0.12
a See Figure 3 and text for definitions and formalism. Repeated measurements yielded estimates of Q-+ and Q+- varying by ∼20% and ∼50%, respectively.
The greatest uncertainty was related to the irradiance and spectral distribution of the photochromic source(s).
UV-vis irradiation were repeated 40 times for 3a (Figure 5a,b
and Figure 6a,b) and 25 times for 3b (Figure 5c,d and Figure
6c,d). The absorption and fluorescence signals corresponding
to the initial form and the photostationary state remained stable
during these sequential reactions.
Computation of Photokinetic Parameters. The photokinetic
parameters corresponding to ring-closing reactions induced by
UV irradiation and to ring-opening back reactions initiated with
visible light were derived from the evolution of absorption
signals as a function of irradiation time (Figures 3, 4, and other
data). The initial state of the forward reactions and the final
state of the back reactions were the open ring configuration of
the acceptor [Ro ) Roff ) 0]. Forward reactions led to
photostationary states with values of Rps invariably <1, because
of the finite absorption of the closed state (and of the donor in
3a and 3b) in the near-UV. The corresponding fluorescence and
9
J. AM. CHEM. SOC. VOL. 124, NO. 25, 2002 7485

A R T I C L E S
Giordano et al.
Figure 5. Absorption monitoring of cyclical on and off (initial UV-vis) photoconversions of 3a and 3b. Initial and photostationary state spectra of 3a (a)
and 3b (c). Absorbance at 548 nm (3a, b) and 568 nm (3b, d) as a function of time during multiple forward and back conversion cycles initiated from the
diheteroarylethene open form. Photoconversion achieved by irradiation at 320 or 520 nm.
absorbance changes during such reactions are given by eqs S4
and S5; thus, for E- ≈ 0 and E+ ≈ 1 (our system), the fractional
quenching during such a back reaction should have been equal
to Ron.
In accordance with the formalism presented in this work (eqs
S8-S11), monoexponential processes were observed in all cases
(although at some wavelengths, a very slight transient overshoot
was observed in forward reactions). The parameters derived
yielded equilibration times of 4.9 and 7.2 s, respectively (average
of eight determinations). Photoreversion by irradiation at 568
nm (irradiance 207 mW cm-2) led to equilibration times of 19.9
( 0.8 and 13.8 ( 0.6 s (2b and 3b, respectively, average of
eight determinations). Thus, the cyclization was slightly retarded,
and the cycloreversion was slightly accelerated in the donor-
acceptor conjugate.
Discussion
from such data using eqs 3-5 and S8-S11, k+-, k-+, Rps, Q+-
,
and Q-+ are given in Table 1. The two isolated dihet-
eroarylethene compounds (2a, 3a) differed in important re-
spects: (i) Q+- was greater for 2a (0.31 vs 0.17); both values
were <0.5, in accordance with data reported for most dihet-
eroarylethenes of similar structure.5 (ii) Q-+ was also greater
for 2a (0.01-0.06 vs <0.01). (iii) Under identical irradiation
conditions for the forward reaction (340 nm), 2b achieved a
higher Rps and at a somewhat faster rate than 2a; these
differences could be attributed to the higher absorption of 2b
in the near-UV (eq S10). In accordance with the formalism
(Figure 3 and eqs 3-5 and S8-S11), conjugation to the donor
did not lead to a reduction in Q-+, but Rps diminished because
of the FRET-mediated donor contribution to the cycloreversion
reaction (∼17 and ∼8% for 2a f 3a and 2b f 3b, respec-
tively). There was good agreement between the values of Rps
obtained from the absorption and NMR measurements.
FRET can be activated and deactivated reversibly by switch-
ing on and off the closed form of diheteroarylethenes, as
demonstrated by the synchronous change in closed form
absorbance and donor fluorescence intensity induced by the
UV-vis irradiation cycles. Upon photogeneration of the closed
form of the diheteroarylethene, a concurrent reduction in the
donor fluorescence intensity was observed with E switching
from 0 to 1 for the model compounds 3a and 3b, reflecting the
changes in the critical transfer distance Ro. The photogenerated
closed form was stable at room temperature in the dark, as
evidenced both by absorption and by fluorescence signals, for
>120 h. Appropriate measurement conditions and controls
excluded inner filter effects arising from absorbance at the
excitation and photoconversion wavelengths as well as absorp-
tion of donor fluorescence by the diheteroarylethene closed form.
Low sample concentrations were used, such that the absorbance
at critical wavelengths was <0.05. Control samples of 1 showed
invariant fluorescence emission upon exposure to the same
experimental conditions. The positions of the absorption (430
In another series of measurements not represented in Table
1, the forward reactions of 2b and 3b induced by identical
conditions of irradiation at 340 nm (irradiance 107 mW cm-2
)
9
7486 J. AM. CHEM. SOC. VOL. 124, NO. 25, 2002

Diheteroarylethenes as Photoswitchable Acceptors
A R T I C L E S
Figure 6. Fluorescence monitoring of cyclical on and off (initial UV-vis) photoconversions of 3a and 3b. Initial and photostationary state emission spectra
of 3a (a) and 3b (c); excitation, 430 nm. Emission signals at 520 nm in the initial open form and in the photostationary state recorded for a number of cycles
during back and forward conversion initiated from the diheteroarylethene open forms of 3a (b) and 3b (d). Photoconversion achieved by irradiation at 320
or 520 nm.
nm) and fluorescence (530 nm) maxima in 1 before and after
conjugation were invariant, excluding the presence of ground-
state complexes.
than did exposure to 366 nm. The increase in the photogenera-
tion of the closed form resulted in a concurrent increase of the
fluorescence quenching from 33 to 65%. Diheteroarylethenes
do not display a thermal back reaction at temperatures below
80 °C, thereby enabling stable FRET determinations over long
measurement times. Moreover, the photocyclization reaction
takes place without a change in net charge, solvent-driven dark
processes are absent, and fatigue is negligible.
We computed the absorption spectra of the closed form using
the degrees of conversion determined from NMR.8a-e Integration
of the resonances corresponding to the closed form were
compared with corresponding signals of the open form to
determine the contribution of each species. The conversion
values agreed well with the extents of fluorescence quenching,
indicative of a FRET efficiency close to 100% expected for a
system (3a) with an Ro of ∼38 Å and a maximum donor-
acceptor distance of 20 Å (eq S12). The same situation applied
to 3b, with an Ro of ∼35 Å and a maximum donor-acceptor
distance of 22 Å.
The photochromic difference spectra for 2a and 3a and for
2b and 3b were virtually indistinguishable, attesting to the lack
of significant ground-state interactions between the LY donor
and the diheteroarylethene acceptors.
Irradiation of 2a at 313 nm gave rise to a greater (albeit
slower) conversion to the closed form of the diheteroarylethene
We have provided in this work a general, systematic
formalism accounting for the coupling between the photochro-
mic interconversion mechanisms and the photophysical states
of both donor and acceptor moieties. The photokinetic scheme
exhibits in theory and practice a simple monoexponential
behavior, from which rate constants and quantum yields for the
forward and cycloreversion photochromic reactions can be
derived. The formalism allowed the dissection of the photo-
conversion reaction from the FRET process. The quantum yields
for the cyclization and cycloreversion reactions were equal
within experimental error for 2a and 3a as well as for 2b and
3b (Table 1). We conclude that the donor moiety in the model
compounds did not affect the inherent properties of the
photoconversion reactions and that the observed slight changes
in the equilibration rates in the donor conjugate resulted from
the differences in the combined absorption cross-sections.
However, the conversion to the closed forms was somewhat
(8) (a) Norsten, T. B.; Branda, N. R. J. Am. Chem. Soc. 2001, 123, 1784-
1785. (b) Osuka, A.; Fujikane, D.; Shinmori, H.; Kobatake, S.; Irie, M. J.
Org. Chem. 2001, 66, 3919-3923. (c) Gauglitz, G.; Scheerer, E. J. J.
Photochem. Photobiol., A 1993, 71, 205-212. (d) Rau, H.; Greiner, G.;
Gauglitz, G.; Meier, H. J. Phys. Chem. 1990, 94, 6523-6524. (e) Wilkinson,
F.; Hobley, J.; Naftaly, M. J. Chem. Soc., Faraday Trans. 1992, 88, 1511-
1517.
9
J. AM. CHEM. SOC. VOL. 124, NO. 25, 2002 7487

A R T I C L E S
Giordano et al.
more efficient for 2a and 2b than for 3a and 3b. FRET-
augmented cycloreversion accounted for a <20% decrease in
the yield of the closed form upon irradiation with near-UV light.
According to the scheme of Figure 3, the FRET process
contributes to the back reaction via a second parallel pathway
leading to the excited state of the closed form of the dihet-
eroarylethene. The values of Q-+ computed from data obtained
with irradiation at 366 and 340 nm were the same, although at
wavelengths <340 nm, a distinct trend to lower conversion
efficiencies was observed. Under all conditions, the computed
values of Q-+ were <0.5, implying the existence of the two
states postulated for the open form, only one of which can
undergo cyclization. The presumption is that the equilibration
between the open-form conformers is very rapid, leading to a
preequilibrium that is implicit in the scheme of Figure 3.
The quenching of a donor engaged in a FRET process is
generally reflected in a corresponding fractional change (reduc-
tion) of its fluorescence lifetime. We observed two donor species
in the photostationary state after irradiation in the near-UV, for
example, at 366 nm. One of these corresponded to the population
of donor 1 attached to the diheteroarylethene in the closed form,
the latter acting as an efficient energy acceptor and thus
quenching the donor virtually completely. The other fraction
represented 1 bound to the diheteroarylethene in the open form,
characterized by an E- ≈ 0 and thus incapable of generating a
significant FRET effect. This case can be visualized in the time
domain with an expression for the decay of two fluorescence
species excited by a pulse of light:
results of such experiments. Other applications of pcFRET
include the potential for enhancing signal-to-background rela-
tionships in analytical procedures using flow cytometric or
imaging technology, which are characterized by very low levels
of specific complexes.
Experimental Section
Materials. (3-(4-{4-[3,3,4,4,5,5-Hexafluoro-2-(2-methoxy-benzo[b]-
thiophen-3-yl)-cyclopent-1-enyl]-3,5-dimethyl-thiophen-2-yl}-benzoy-
lamino)-propionic acid (2a) was prepared by treatment of acyl chloride
of the diheteroarylethene derived from the formyl derivative10 with
â-alanine in ether-water under alkaline condition. 4-(4-{4-[3,3,4,4,5,5-
Hexafluoro-2-(2-methoxy-benzo[b]thiophen-3-yl)-cyclopent-1-enyl]-
3,5-dimethyl-thiophen-2-yl}-phenyl)-butyric acid (2b) was prepared
from the butanol derivative.11a,b,12-14 The details of the synthetic
procedure will be published elsewhere. Lucifer Yellow cadaverine 1,
N-(5-aminopentyl)-4-amino-3,6-disulfo-1,8-naphthalimide, was from
Molecular Probes (Eugene, Oregon). N,N′-dicyclohexyl-carbodiimide
and N-hydroxysuccinimide were purchased from Fluka Chemie AG,
Buchs, Switzerland. We also refer to substructure 1 as LYC and
substructure 2a or 2b as DAE.
(3-(4-{4-[3,3,4,4,5,5-Hexafluoro-2-(2-methoxy-benzo[b]thiophen-
3-yl)-cyclopent-1-enyl]-3,5-dimethyl-thiophen-2-yl}-benzoylamino)-
propionic acid-2,5-dioxo-pyrrolidin-1-yl Ester (4a). Twenty milli-
grams (0.032 mmol) of 2a reacted with 8 mg (0.039 mmol) of DCC
and 4 mg (0.035 mmol) of N-hydroxysuccinimide in 5 mL of dry
acetonitrile at room temperature for 12 h. The dicyclohexylurea was
filtered, and the N-succinimide active ester 4a (Figure 1) was evaporated
in vacuo.
Disodium-6-amino-2-{5-[3-(4-{4-[3,3,4,4,5,5-hexafluoro-2-(2-meth-
oxy-benzo[b]thiophen-3-yl)-cyclopent-1-enyl]-3,5-dimethyl-thiophen-
2-yl}-benzoylamino)-propionylamino]-pentyl}-1,3-dioxo-2,3-dihydro-
1H-benzo[de]isoquinoline-5,8-disulfonate (3a). Sixteen milligrams of
Lucifer Yellow cadaverine (1, 32 µmol) was dissolved in 30 µL of 0.5
M borate buffer (pH 9.2) and reacted with 32 µmol of 4a dissolved in
200 µL of dry acetonitrile, for 2 h at room temperature. The product
was purified by HPLC using a RP-C18 stationary phase and an elution
solvent CH3CN:triethylammonium acetate (TEAA, 1 M, pH 7.0) 35:
65.
-t/τD+
F(t) ) Fo[(1 - Rps)e-t/τ + Rpse
]
(6)
D-
where Fo is the initial donor fluorescence signal, and τD
)
+
τ
D-(1 - E+)/(1 - E-). For the limit E+ ≈ 1, computed for
both 3a and 3b with the diheteroarylethenes in the closed form,
would be too short to be detected in the frequency-domain
τ
D+
system, regardless of the value of Rps. That is, the fluorescence
decay would be dominated by τD . The quenching of steady-
-
1H NMR (500 MHz, CD3OD): δ 9.05 (1H, H-7 LYC), 8.94 (1H, d,
1.5 Hz, H-5 LYC), 8.91 (1H, s, H-2 LYC), 8.49 (1H, s, NH, LYC),
7.20-8.00 (8H, Ar in DAE), 4.05 (2H, t, J ) 8.5 Hz, NCCCCCH2-
NHCO LYC), 3.72 (2H, m, -CONHCH2CH2CO- DAE), 3.20 (2H,
m, NCH2CCCCNHCO LYC), 2.62 (2H, m, CONHCH2CH2CO DAE),
2.47 (1.5H, s, CH3 DAE), 2.45 (1.5H, s, CH3 DAE), 2.42 (1.5H, s,
CH3 DAE), 2.21 (1.5H, s, CH3 DAE), 2.19 (1.5H, s, CH3 DAE), 1.93
(1.5H, s, CH3- DAE), 1.6-1.8 (6H, CH2 LYC). EM, FABMS,
M-H: 1061.3. Anal. Calcd for C47H38F6N4Na2O10S4: C, 52.03; H, 3.73;
N, 3.79. Found: C, 52.36; H, 3.74; N, 3.81.
1H NMR at the photostationary state by irradiation at 366 nm (500
MHz, CD3OD): δ 9.05 (1H, H-7 LYC), 8.94 (1H, d, 1.5 Hz, H-5 LYC),
8.91 (1H, s, H-2 LYC), 8.49 (1H, s, NH, LYC), 7.17-8.00 (Ar in
DAE), 4.05 (2H, t, J ) 8.5 Hz, NCCCCCH2NHCO LYC), 3.72-3.74
(CONHCH2), 3.20 (NCH2CCCCNHCO LYC), 2.62-2.63 (CH2CH2-
CO DAE), 2.47 (s, CH3 DAE), 2.45 (s, CH3 DAE), 2.42 (s, CH3 DAE),
2.35 (s, CH3-DAECF), 2.21 (s, CH3 DAE), 2.19 (s, CH3 DAE), 2.11
state donor fluorescence FD after photogeneration of the closed
form would have been RpsE+ (eq S4). From NMR measure-
ments, we determined Rps (366 nm) to be 0.33 in 2a and 0.85
in 2b, values which were reflected accurately in the extents of
fluorescence quenching (33 and 85%, respectively), allowing
the determination of the FRET efficiency (100%) for the
diheteroarylethene closed form.
Experiments involving cyclical ring closures and openings
demonstrated that the FRET process can be switched on and
off reversibly. We performed 40 cycles for 3a and 25 cycles
for 3b with no apparent fatigue (Figures 5 and 6). The number
of cycles of FRET activation-deactivation that can be achieved
by cyclic irradiation with UV and visible light is of central
importance, in that this parameter determines the feasibility of
carrying out continuous determinations in dynamic systems, for
example, with living cells.
Recently, we succeeded in conjugating diheteroarylethene-
based photochromic moieties to biomolecules and creating
donor-acceptor pairs via biotin-streptavidin. The donor fluo-
rescence was readily modulated in aqueous solution by cyclic
irradiation.9 The FRET E was 0 for the diheteroarylethene in
the open form and rose to 0.5 with the compound in the closed
form. The formalism for pcFRET presented here provides the
means for developing experimental strategies and analyzing the
(9) Giordano, L.; Macareno, J.; Song, L.; Jovin, T. M.; Irie, M.; Jares-Erijman,
E. Molecules 2000, 5, 591-593.
(10) Takeshita, M.; Irie, M. Chem. Lett. 1998, 1123-1124.
(11) (a) Nakashima, N.; Irie, M. Polym. J. 1998, 30, 985-989. (b) Corey, E. J.;
Schmidt, G. Tetrahedron Lett. 1979, 399-403.
(12) Gadella, T. W. J., Jr.; Jovin, T. M.; Clegg, R. M. Biophys. Chem. 1993,
48, 221-239.
(13) Morgan, C.; Hua, Y.; Mitchell, A.; Murray, J. G.; Boardman, A. ReV. Sci.
Instrum. 1996, 67, 41-47.
(14) Baumann, J.; Calzaferri, G.; Forss, L.; Hugentobler, T. J. J. Photochem.
1985, 28, 457-473.
9
7488 J. AM. CHEM. SOC. VOL. 124, NO. 25, 2002

Diheteroarylethenes as Photoswitchable Acceptors
A R T I C L E S
(s, CH3-DAECF), 2.05 (s, CH3-DAECF), 2.04 (s, CH3-DAECF), 1.93
(s, CH3- DAE), 1.6-1.8 (6H).
8 Hz, CH2-Ph), 2.20 (2H, t, CH2CO), 2.15 (3H, s, CH3 DAE), 1.95
(3H, s, CH3- DAE), 1.5-1.8 (8H).
4-(4-{4-[3,3,4,4,5,5-Hexafluoro-2-(2-methoxy-benzo[b]thiophen-
3-yl)-cyclopent-1-enyl]-3,5-dimethyl-thiophen-2-yl}-phenyl)-butyr-
ic acid 2,5-dioxo-pyrrolidin-1-yl Ester (4b). Ten milligrams (0.0164
mmol) was reacted in a similar manner as described for 4a, yielding
4b (Figure 1).
Acknowledgment. This work was supported by the Volk-
swagen Stiftung (Grant I/74924) and the Max Planck Society.
E.A.J.-E. would like to thank Fundacio´n Antorchas, CONICET,
ANPCyT, and UBACyT for financial support.
Disodium-6-amino-2-{5-[5-(4-{4-[3,3,4,4,5,5-hexafluoro-2-(2-meth-
oxy-benzo[b]thiophen-3-yl)-cyclopent-1-enyl]-3,5-dimethyl-thiophen-
2-yl}-phenyl)-butyrylamino]-pentyl}-1,3-dioxo-2,3-dihydro-1H-benzo-
[de]isoquinoline-5,8-disulfonate (3b). As described for 3a, 0.0164
mmol of 4b was reacted with equimolar amounts of 1. The product 3b
was purified by HPLC.
1H NMR (500 MHz, CD3OD): δ 9.06 (1H, d, 1.5 Hz, H-7), 8.90
(1H, d, 1.5 Hz, H-5), 8.87 (1H, s, H-2), 8.52 (0.5H, broad s), 7.70-
7.15 ppm (8H, Ar in DAE), 4.10 (2H, t, NCH2CCCCNHCO), 3.85
(3H, s, OCH3), 3.43 (2H, m), 2.6 (2H, t, J ) 8 Hz, CH2-Ph), 2.20
(2H, t, CH2CO DAE), 2.15 (3H, s, CH3 DAE), 1.95 (3H, s, CH3-
DAE), 1.5-1.80 (8H). EM, FABMS, M-H: 1048.3. Anal. Calcd for
C47H39F6N3Na2O10S4: C, 51.60; H, 3.59; N, 3.84. Found: C, 52.05;
H, 3.61; N, 3.87.
Supporting Information Available: Quantitative formalism
for the photostationary state and photokinetics containing eqs
S1-S11. Steady-state spectroscopy methods: photochromic
conversion light source, photochromic conversion, and photo-
kinetics. Methods: fluorescence lifetime and FRET methods.
Table S1: Integrated values for 1H NMR signals of 2a and 2b.
Figure S1: Spectral properties for the completely off (-) and
on (+) states (open and closed). Figure S2: Effects of
photochromic conversion wavelength on the degree of photo-
conversion and donor fluorescence quenching. Figure S3: 1H
NMR for 2a and 3b in the open form and in the photostationary
state. Figure S4: Photophysical-photochromic scheme incor-
porating donor-acceptor FRET (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
1H NMR of photostationary state by irradiation at 340 nm (500 MHz,
CD3OD): δ 9.06 (1H, d, 1.5 Hz, H-7), 8.90 (1H, d, 1.5 Hz, H-5), 8.87
(1H, s, H-2), 8.00-7.15 (DAE), 4.10 (2H, t, NCH2CCCCNHCO), 3.85
(s, OCH3DAE), 3.43 (2H, m), 3.36 (s, OCH3 DAECF), 2.6 (2H, t, J )
JA016969K
9
J. AM. CHEM. SOC. VOL. 124, NO. 25, 2002 7489