Development of enzyme-activated photosensitizer based on intramolecular electron transfer

Article abstract of DOI:10.1016/j.bmcl.2010.06.091

Photosensitizers produce cytotoxic reactive oxygen species (ROS) upon light illumination, but it is difficult to ablate cells of a specific type (e.g., tumor cells) in the presence of other cell populations, because of the limited precision with which light illumination can be directed to small areas. Here, we report a strategy to achieve cell type-specific ablation by using an enzyme-activated off/on switch for oxidative stress induction. In the unactivated photosensitizer, induction of oxidative stress is quenched by intramolecular electron transfer. However, the target cells express an enzyme that hydrolyzes a substrate moiety of the photosensitizer and the activated photosensitizer induces oxidative stress. As proof of concept, we designed and synthesized a xanthene-based photosensitizer, TGI-βGal, whose oxidative stress induction ability is switched on following hydrolysis reaction with β-galactosidase, a widely used gene marker. TGI-βGal could selectively ablate lacZ-positive cells, whereas it showed no toxicity to lacZ-negative cells, upon light illumination.

Full text of DOI:10.1016/j.bmcl.2010.06.091

Bioorganic & Medicinal Chemistry Letters 20 (2010) 4320–4323
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Development of enzyme-activated photosensitizer based on
intramolecular electron transfer
Takatoshi Yogo a,b, Yasuteru Urano a,c, Mako Kamiya a,b, Kiminari Sano a, Tetsuo Nagano a,b,
*
a Graduate School of Pharmaceutical Sciences, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
b CREST, Japan Science and Technology Agency, 4-1-8 Honcho Kawaguchi, Saitama 332-0012, Japan
c Presto, Japan Science and Technology Agency, 3-5 Sanbancho, Chiyoda, Tokyo 102-0075, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Photosensitizers produce cytotoxic reactive oxygen species (ROS) upon light illumination, but it is diffi-
cult to ablate cells of a specific type (e.g., tumor cells) in the presence of other cell populations, because of
the limited precision with which light illumination can be directed to small areas. Here, we report a
strategy to achieve cell type-specific ablation by using an enzyme-activated off/on switch for oxidative
stress induction. In the unactivated photosensitizer, induction of oxidative stress is quenched by intramo-
lecular electron transfer. However, the target cells express an enzyme that hydrolyzes a substrate moiety
of the photosensitizer and the activated photosensitizer induces oxidative stress. As proof of concept, we
designed and synthesized a xanthene-based photosensitizer, TGI-bGal, whose oxidative stress induction
ability is switched on following hydrolysis reaction with b-galactosidase, a widely used gene marker.
TGI-bGal could selectively ablate lacZ-positive cells, whereas it showed no toxicity to lacZ-negative cells,
upon light illumination.
Received 25 April 2010
Revised 14 June 2010
Accepted 15 June 2010
Available online 19 June 2010
Keywords:
Photosensitizer
b-Galactosidase
Electron transfer
Oxidative stress
Ó 2010 Elsevier Ltd. All rights reserved.
Photosensitizers are chemical tools that produce reactive oxy-
gen species (ROS) upon light illumination and are commonly used
to cause light-induced cell killing, for example, in the treatment of
cancer by means of photodynamic therapy (PDT).1 Usually, the
area of cell ablation is regulated by controlling the area of light illu-
mination. However, it is difficult to distinguish and ablate a specific
type of cell (e.g., tumor cells) in the presence of other cell popula-
tions without affecting the non-targeted cells, because the preci-
sion with which light illumination can be directed to small areas
is limited. The development of photosensitizers that recognize spe-
cific types of cells and generate ROS only in them, enabling cell
type-specific ablation, would be extremely useful.
Silencing a specific cell population is also an important and fun-
damental technique in biological research. For example, ablating
specific cell populations of neural circuit in vitro and in vivo would
allow us to understand how the activity in specific neuronal popu-
lations contributes to physiological processes and behavioral re-
sponses.2,3 Light-triggered cell ablation using photosensitizers
potentially has a great advantage, compared to other methods,
including surgical excision and pharmacological techniques, be-
cause the timing and site of cell ablation can be precisely con-
trolled by adjusting the timing and area of light illumination. A
green fluorescent protein (GFP)-based photosensitizer (KillerRed)4
was recently developed for this purpose. KillerRed can be intro-
duced into and expressed in cells of interest, and light illumination
can then be applied to specifically ablate those cells by means of
light-induced ROS generation. However, the GFP-based photosen-
sitizer has some drawbacks, including low efficiency of ROS gener-
ation compared with small molecule-based photosensitizers.
In this study, we report an approach to achieve cell type-specific
ablation using a functional small-molecule-based photosensitizer
with an off/on switching device. We focused on b-galactosidase,5
a widely used gene expression marker, which can be easily intro-
duced into a selected cell population, and utilized its hydrolytic
activity as an off/on switch of ROS generation; that is, b-galactosi-
dase serves to activate the photosensitizer, and thereby switches
on local induction of oxidative stress by the photosensitizer inside
the target cells. We have recently developed highly sensitive fluo-
rescent probes for b-galactosidase (TG-bGal)6 based on precise
control of the photochemical properties of newly developed fluo-
rescein derivatives (TokyoGreens, TGs) by photoinduced electron
transfer (PeT) (Fig. 1a). The fluorescence of TG-bGal is efficiently
quenched by electron transfer from the benzene moiety to the xan-
thene moiety and is restored by the change in the reduction poten-
tial of the xanthene moiety after hydrolysis by b-galactosidase. We
hypothesized that if the xanthene moiety of TG-bGal could be con-
verted to a photosensitizer moiety, it would behave similarly and
could work as a b-galactosidase-dependent activatable photosensi-
tizer, which would generate ROS only after activation via hydroly-
sis reaction by b-galactosidase. Such a strategy might have broad
applicability.
* Corresponding author. Tel.: +81 3 5841 4950; fax: +81 3 5841 4855.
E-mail address: tlong@mol.f.u-tokyo.ac.jp (T. Nagano).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.06.091

T. Yogo et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4320–4323
4321
O
O
a
O
benzene moiety
(electron donor)
-
-
e
e
OH
O
β−galactosidase
HO
HO
pH. 7.4
xanthene moiety
-
(electron acceptor)
O
O
O
O
O
OH
2-Me-4-OMe TokyoGreen
(strongly fluorescent, Φfl=0.840)
TG-βGal
(almost non-fluorescent, Φfl=0.002)
b
Light illumination
Light illumination
O
O
O
-
-
electron
donor
e
e
OH
(1) Cellular uptake
O
O
HO
O
(2) Hydrolysis by
electron
acceptor
O
β−galactosidase
-
HO
O
O
O
O
OH
I
I
β-galactopyranosyl
TGI
residue
Cellular
oxidative stress
Cellular
oxidative stress
TGI-βGal
Figure 1. Development of b-galactosidase-dependent photosensitizer. (a) Reaction scheme of fluorescence probe for b-galactosidase (TG-bGal). (b) Reaction scheme of
b-galactosidase-dependent photosensitizer (TGI-bGal).
OAc
O
O
O
AcO
AcO
O
Br
OAc
OAc
, Cs2CO3
O
O
I
O
I2, HIO3
EtOH
OAc
AcO
AcO
AcO
AcO
DMF
O
O
O
O
HO
O
O
O
O
O
O
OAc
OAc
1
2
3
O
O
O
OH
O
O
O
I
β−galactosidase
NaOMe
MeOH
HO
HO
NaPi buffer (pH 7.4)
O
O
O
HO
O
OH
I
TGI-βGal
TGI
Scheme 1.
To test this hypothesis, we designed and synthesized TGI-bGal
(Fig. 1b and Scheme 1) by iodinating the xanthene moiety of TG-
bGal in order to increase the singlet oxygen generation making
use of the so-called internal heavy-atom effect,7,8 and to increase
the intracellular retention of the molecules by increasing their
hydrophobicity. TGI-bGal was converted to TGI (Fig. 1b) by b-galac-
tosidase without marked by-product generation (Supplementary
Fig. 1), indicating that TGI-bGal is a good substrate for b-galactosi-
dase. We then compared the ability of TGI-bGal to induce oxidative
stress with that of TGI by following the appearance of the 351 nm
(Ufl = 0.037) than that of fluorescein (Ufl = 0.85), suggesting that
the intersystem crossing efficiency from the lowest singlet excited
state to the triplet state had been greatly enhanced by the internal
heavy atom effect. Cellular phototoxicity of a photosensitizer upon
light illumination depends on both the extinction coefficient (
and /oxi, so the cellular phototoxicity of TGI-bGal (
pected to change dramatically (about 20 times) in the presence
of b-galactosidase.
ꢀ)
ꢀ
Á/oxi) is ex-
We then examined whether TGI-bGal could selectively ablate
b-galactosidase-expressing cells upon light illumination. HEK293
cells transduced or not transduced with lacZ (lacZ-positive or
9,10
absorbance band during oxidation of iodine ion (IÀ) to I3
.
À
TGI-bGal was unable to oxidize IÀ, whereas IÀ-oxidizing ability
appeared in the presence of b-galactosidase (Fig. 2).
lacZ-negative cells) were loaded with 25 lM TGI-bGal for 2 h.
Bright fluorescence was observed only in lacZ(+) cells (Fig. 3a–d),
owing to the higher quantum efficiency of fluorescence of TGI than
that of TGI-bGal (Table 1 and Supplementary Fig. 2), thereby
demonstrating that TGI-bGal permeates through the plasma
membrane and is converted into TGI in lacZ(+) cells. After wash-
Table 1 summarizes the rate of I3À production at the initial stage
(i.e., the slope), which corresponds to the efficiency of oxidative
stress induction (/oxi). TGI had a 4.8 times higher value of /oxi
and
a
much lower quantum efficiency of fluorescence

4322
T. Yogo et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4320–4323
tion of TGI-bGal, expression of lacZ, and light illumination resulted
a
b
0.1
0.08
0.06
0.04
0.02
0
in cellular toxicity (Fig. 3 and Supplementary Fig. 3). Significant
morphological changes were observed only in lacZ(+) cells
(Fig. 3e and f), and the occurrence of lethal damage was confirmed
by cell viability staining (Supplementary Fig. 3). These results dem-
onstrate that TGI-bGal could selectively ablate lacZ-positive cells
upon light illumination.
β
TGI- Gal
In conclusion, we have developed a b-galactosidase-activated
photosensitizer, which induces oxidative stress only after having
been hydrolyzed by b-galactosidase, based on quenching of oxida-
tive stress induction by the photosensitizer via intramolecular elec-
tron transfer, and activation by enzymatic removal of the quenching
substituent. The enzymatic reaction results in a 20-fold increase of
oxidative stress induction ability, and selective ablation of lacZ-po-
sitive HEK 293 cells was demonstrated in the presence of TGI-bGal
under light illumination. Activatable photosensitizers such as TGI-
bGal are expected to offer unique advantages in biological and
clinical applications. First, selective ablation of a cell population
of interest in a temporally well controlled manner would enable
us to investigate the roles of specific cell populations in neural
circuits or the fate of cells in developing and adult organisms,
although bystander effects11 may limit a spatial resolution of cell
ablation. Second, activatable photosensitizers would open up new
possibilities for PDT without causing prolonged light sensitivity,
which is a serious clinical side effect of photosensitizers that do
not specifically localize to tumor tissues by designing photosensi-
tizers which recognize environments of specific tumor. We now
plan biological studies to confirm the usefulness of TGI-bGal in
silencing specific populations of cells, especially neurons.
300 350 400 450 500 550 600
wavelength (nm)
0.1
TGI
0.08
0.06
0.04
0.02
0
300 350 400 450 500 550 600
wavelength (nm)
Figure 2. Time profile of absorption spectra during iodine ion oxidation. A 1 lM
solution of (a) TGI-bGal or (b) TGI in 100 mM sodium phosphate buffer (pH 7.4),
containing 100 mM KI was illuminated with Xe light source, which was filtered to
around 490 nm, and absorption spectra were measured every 2 min.
out of excess substrate, the cells were illuminated with blue light
(BP 470–490 nm, 7 mW/cm2 at 490 nm) for 1 min. The combina-
Table 1
Summary of photochemical and photosensitization properties of TGI-bGal and TGI
Slopeb
e490
/
e
Á/oxi (rel.)
a
c
d
e
Absorption
max (nm)a
Extinction coefficient
Emission
Ufl
oxi
a
(Â104 MÀ1 cmÀ1
)
max (nm)a
[Â10À4/min] =
eÁ/oxi
[Â104 MÀ1 cmÀ1
]
[Â10À8
]
TGI-bGal
TGI
Fluorescein
470
502
2.6
9.4
—
526
ꢀ0
0.037
2.6
52
14
2.2
6.4
7.9
1.2
8.2
1.7
1
20
a
All data were measured in 0.1 M sodium phosphate buffer, pH 7.4. Quantum efficiency of fluorescence (Ufl) was determined using that of fluorescein (0.85) in 0.1 M NaOH
aq as a standard.
b
À
Rate of I3 production monitored by measuring the absorbance 351 nm in 100 mM sodium phosphate buffer, pH 7.4, containing 100 mM KI.
c
Molar extinction coefficient at 490 nm.
Relative efficiency of oxidative stress induction (/oxi) determined by calibrating the slope in terms of e490
The relative value of
d
.
e
eÁ/oxi of TGI taking that of TGI-bGal as 1.
Figure 3. Selective killing of lacZ-positive cells using TGI-bGal. (a–d) Differential interference contrast (DIC) and fluorescence images of lacZ-positive HEK 293 cells (a and b)
or lacZ-negative cells (c and d), loaded with TGI-bGal. (e) DIC images of HEK 293 cells (lacZ(+)) after light illumination in the presence of TGI-bGal. (f) HEK 293 cells (lacZ(À))
were not affected by light illumination in the presence of TGI-bGal. Scale bar indicates 5 lm.

T. Yogo et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4320–4323
4323
Acknowledgements
References and notes
1. Lissi, E. A.; Encinas, M. V.; Lemp, E.; Rubio, M. A. Chem. Rev. 1993, 93, 699.
We thank Dr. M. Yamada for the gift of HEK293 cells transduced
with lacZ. This study was supported by research grants (Grant Nos.
19021010, 19205021, and 20117003) from the Ministry of Educa-
tion, Culture, Sports, Science and Technology of the Japanese Gov-
2. Marek, K. W.; Davis, G. W. Curr. Opin. Neurobiol. 2003, 13, 607.
3. Miesenbock, G. Curr. Opin. Neurobiol. 2004, 14, 395.
4. Bulina, M. E.; Chudakov, D. M.; Britanova, O. V.; Yanushevich, Y. G.; Staroverov,
D. B.; Chepurnykh, T. V.; Merzlyak, E. M.; Shkrob, M. A.; Lukyanov, S.; Lukyanov,
K. A. Nat. Biotechnol. 2006, 24, 95.
5. Spergel, D. J.; Kruth, U.; Shimshek, D. R.; Sprengel, R.; Seeburg, P. H. Prog.
Neurobiol. 2001, 63, 673.
6. Urano, Y.; Kamiya, M.; Kanda, K.; Ueno, T.; Hirose, K.; Nagano, T. J. Am. Chem.
Soc. 2005, 127, 4888.
ernment and
a grant from the Kato Memorial Bioscience
Foundation to Y.U.
7. Yuster, P.; Weissman, S. I. J. Chem. Phys. 1949, 17, 1182.
8. McClure, D. S. J. Chem. Phys. 1949, 17, 905.
9. Frei, H. Chimia 1991, 45, 175.
Supplementary data
10. Mosinger, J.; Micka, Z. J. Photochem. Photobiol., A Chem. 1997, 17, 77.
11. Dahle, J.; Angell-Petersen, E.; Steen, H. B.; Moan, J. Photochem. Photobiol. 2001,
73, 378.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.06.091.