Facile and general synthesis of photoactivatable xanthene dyes

Article abstract of DOI:10.1002/anie.201104571

Despite the apparent simplicity of the xanthene fluorophores, the preparation of caged derivatives with free carboxy groups remains a synthetic challenge. A straightforward and flexible strategy for preparing rhodamine and fluorescein derivatives was developed using reduced, "leuco" intermediates. Copyright

Full text of DOI:10.1002/anie.201104571

Communications
DOI: 10.1002/anie.201104571
Caged Fluorophores
Facile and General Synthesis of Photoactivatable Xanthene Dyes**
Laura M. Wysocki, Jonathan B. Grimm, Ariana N. Tkachuk, Timothy A. Brown, Eric Betzig,
and Luke D. Lavis*
Photoactivatable “caged” fluorophores enable numerous
advanced biological imaging experiments,[1] including photo-
activated localization microscopy (PALM)[2] and related
super-resolution imaging techniques.[3] Of the extant fluoro-
phore scaffolds, caged rhodamines and fluoresceins display
properties that are exceptionally well suited for super-
resolution microscopy, exhibiting high contrast and photon
yields. The utility of these probes in PALM imaging has been
hampered, however, by inefficient syntheses. For example,
the existing route to caged Q-rhodamine, a promising PALM
probe,[2] requires harsh, strongly basic conditions and is
reported to “generate many products” with yields given as
“poor” and “variable”.[1a] Such difficult, inefficient syntheses
have rendered these important caged molecules unavailable
to the scientific community.
species[5] but essentially unexplored as synthetic intermedi-
ates for fluorogenic derivatives. We surmised reduction of the
xanthene core would increase the reactivity of the aniline
nitrogens in rhodamines and the phenolic oxygens in fluo-
resceins allowing installation of caging groups using mild
conditions. Here, we establish the use of leuco-dyes as an
effective method to prepare caged fluorophores. This efficient
and general route enables the preparation of the elusive caged
Q-rhodamine (RhQ) with exceptional ease, and can be
extended to rhodamine 110 (Rh110) and 2’,7’-difluorofluor-
escein[6] (Oregon Green) derivatives bearing free carboxyl
groups for bioconjugation. This synthetic method facilitated
the evaluation of these probes for super-resolution microsco-
py, culminating in the first PALM imaging of DNA in a
cellular context.
We set out to develop a general synthesis of photo-
activatable xanthene fluorophores and evaluate these dyes as
PALM labels. We recognized the synthetic difficulties of
caged xanthenes are inextricably linked with the fluorogenic
mechanism of the dyes. Rhodamines and fluoresceins exist in
equilibrium between a brightly fluorescent, “open” quinoid
structure and a colorless, “closed” lactone.[1b] This equilibrium
can be controlled in a light-dependent manner using several
strategies[1,4] with the most versatile involving attachment of
electron withdrawing photolabile groups to the aniline nitro-
gens of rhodamine or phenolic oxygens of fluorescein. While
the open–closed equilibrium is essential for the fluorogenic
properties of the caged dyes, this attribute also complicates
their syntheses. The quinoid form of the dye, which predom-
inates under the basic conditions required for functionaliza-
tion, exhibits poor solubility and low reactivity thereby
frustrating the installation of caging groups.
Our synthesis of a caged Q-rhodamine is shown in
Scheme 1. Condensation of trimellitic anhydride (1) and 7-
hydroxy-1,2,3,4-tetrahydroquinoline (2) gave a crude iso-
meric mixture of 5(6)-carboxy-RhQ.[1a] To install base-labile
protecting groups[7] on the nitrogen substitutents we treated
this material with trifluoroacetic anhydride (TFAA), afford-
ing 5(6)-carboxy-RhQ-bis(trifluoroacetamide). The trifluoro-
acetamides also facilitated isolation of 5-carboxy-RhQ-bis(tri-
fluoroacetamide) (3) by straightforward crystallization. Com-
pound 3 was reduced to the leuco-rhodamine by catalytic
Since the open–closed equilibrium is both the basis for
fluorogenicity and the cause of synthetic difficulties, we
envisioned eliminating this process in a reversible manner.
Rhodamines and fluoresceins can be reduced to “leuco”
derivatives, which are widely used sensors for reactive oxygen
[*] Dr. L. M. Wysocki, J. B. Grimm, A. N. Tkachuk, Dr. T. A. Brown,
Dr. E. Betzig, Dr. L. D. Lavis
Janelia Farm Research Campus, Howard Hughes Medical Institute
19700 Helix Drive, Ashburn, VA 20147 (USA)
E-mail: lavisl@janelia.hhmi.org
[**] This work was supported by The Howard Hughes Medical Institute.
We thank C. G. Galbraith, J. A. Galbraith, S. M. Sternson, and P. H.
Lee for contributive discussions.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201104571.
Scheme 1. Synthesis of NVOC2-5-carboxy-RhQ 8. a) TsOH, EtCO2H,
reflux. b) TFAA, py, CH2Cl2. c) H2(g), Pd/C, THF. d) 4, DCC, DMAP,
CH2Cl2. e) NH4CO3H, THF, H2O, CH3OH. f) 6, DIEA, CH2Cl2. g) DDQ,
CH2Cl2 (wet), reflux.
Re-use of this article is permitted in accordance with the Terms and
Conditions set out at http://angewandte.org/open
11206
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 11206 –11209

hydrogenation at ambient temperature and pressure.[8]
Esterification with 2,4-dimethoxybenzyl (DMB) alcohol (4)
using N,N’-dicyclohexylcarbodiimide (DCC) and catalytic 4-
dimethylaminopyridine (DMAP) furnished leuco-rhodamine
diester 5 in 74% yield over the two-step sequence. Hydrolysis
of the trifluoroacetamides, followed by acylation with chlor-
oformate 6 to install the ortho-nitroveratryloxycarbonyl
(NVOC) cages, yielded reduced rhodamine 7. The aniline
nitrogens in the leuco-Q-rhodamine are more reactive than in
Q-rhodamine, allowing high-yielding functionalization under
mild conditions. This step represents a significant improve-
ment over the reported reaction to install caging groups onto
RhQ.[1a] Treatment of the protected, reduced rhodamine
adduct 7 with the mild oxidant 2,3-dichloro-5,6-dicyanoben-
zoquinone (DDQ) allowed removal of the DMB esters[9] with
concomitant oxidation of the reduced dye core. This reaction
yields fully deprotected and oxidized NVOC2-5-carboxy-RhQ
(8).
aqueous acid and subsequent amidation with TFAA afforded
bis(trifluoroacetamide) 12 in excellent yield over two steps.
Reduction of 12 under H2(g) and Pd/C produced leuco-
rhodamine 110. The resulting free carboxyl groups were
esterified using benzyl alcohol 4, DMAP, and 3-(3-dimethy-
laminopropyl)carbodiimide (EDC) to produce diester 13.
Deprotection of the aniline groups with NH2OH,[7] followed
by acylation with 6 to install the NVOC cages, yielded
intermediate 14. Paralleling the RhQ example, oxidation of
leuco-rhodamine adduct 14 using DDQ gave NVOC2-5-
carboxy-Rh110 15.
We also used this approach with a fluorescein dye, the
photostable 2’,7’-difluorofluorescein,[6] as shown in Scheme 3.
Fluoresceins are easier to derivatize than rhodamines,[1a] but
To test the generality of this leuco-dye approach, we
applied this strategy to rhodamine 110. Installation of caging
groups onto Rh110 requires highly reactive electrophiles[10] and
the synthesis of caged 5-carboxy-Rh110 derivatives has not
been reported. Based on our success with RhQ, we first
attempted the preparation of 5-carboxy-Rh110-bis(trifluoroa-
cetamide) through the reaction of 3-aminophenol and trimel-
litic anhydride,[11] followed by treatment with TFAA. This
protocol delivered a complex mixture of rhodamine and
rhodol products that precluded purification by chromato-
graphy or crystallization. We therefore developed a novel
approach to the preparation of 5-carboxy-Rh110 derivatives as
shown in Scheme 2. 5-Carboxy-3’,6’-dibromofluoran 9[12] was
protected as the benzyl ester. Pd-catalyzed cross-coupling[12,13]
of the aryl bromide substituents in 10 with benzophenone
imine gave rhodamine 11. Hydrolysis of this diimine using
Scheme 3. Synthesis of NV2-5-carboxy-2’,7’-difluorofluorescein 21.
a) H2(g), Pd/C, THF. b) 4, DIC, DMAP, CH2Cl2. c) NH4CO3H, THF,
H2O, CH3OH. d) NH4HSO4, K2CO3, H2O, CH2Cl2. e) DDQ, CH2Cl2
(wet), reflux.
treatment with alkylating agents gives undesired, fluorescent
ether–esters as the major products, due to competing
reactivity of the ortho-carboxylate. The desired nonfluores-
cent caged fluorescein is typically obtained in low yield and
requires extensive purification.[14] Our synthetic strategy
eliminates this unproductive route while improving solubility
in organic solvents. 5-Carboxy-2’,7’-difluorofluorescein di-
acetate (16)[6] was reduced to the leuco-fluorescein diacetate
by catalytic hydrogenation. Esterification of the resulting
diacid with alcohol 4 using N,N’-diisopropylcarbodiimide
(DIC) gave the tetraester 17. Selective hydrolysis of the
acetate esters afforded diphenol 18, which was efficiently
alkylated with bromide 19 using phase-transfer conditions to
install the ortho-nitroveratryl (NV) photolabile groups in
diether 20. Treatment with DDQ gave the desired caged NV2-
5-carboxy-2’,7’-difluorofluorescein 21 in good yield, showing
the final oxidation step is general for both rhodamine and
fluorescein dyes.
We then investigated the chemical and photophysical
properties of the caged dyes. Biotin conjugates of compounds
8, 15, and 21 were synthesized by straightforward amidation
of the free carboxyl groups revealed during the final step of
the synthesis (see Schemes S1–S3 in the Supporting Informa-
tion). The chemical stability of these molecules was assessed
at pH 5–9; we observed negligible (< 1%) spontaneous
Scheme 2. Synthesis of NVOC2-5-carboxy-Rh110 15. a) BnOH, EDC,
DMAP, CH2Cl2. b) Pd(OAc)2, binap, Cs2CO3, toluene, 1008C. c) 5%
HCl/THF. d) TFAA, py, CH2Cl2. e) H2(g), Pd/C, THF. f) 4, EDC, DMAP,
CH2Cl2/EtOAc. g) NH2OH, CH3OH. h) 6, DIEA, CH2Cl2. i) DDQ,
CH2Cl2 (wet), reflux.
Angew. Chem. Int. Ed. 2011, 50, 11206 –11209
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 11207

Communications
uncaging after 48 h (Figure S1). We also determined the
croscopy images are also given for comparison. As expected
from previous in vitro experiments,[2] the RhQ derivative 8-
azide was a viable probe for PALM imaging (Figure 1a). Of
the shorter-wavelength probes, the caged Rh110 15-azide
derivative (Figure 1b) gave better PALM images than
fluorescein 21-azide (Figure 1c) due to the higher photon
yield and lower nonspecific staining. While super-resolution
microscopy of purified DNA has been performed by direct
stochastic optical reconstruction microscopy (dSTORM)[17]
and stimulated emission depletion (STED) microscopy,[18] this
is the first example of super-resolution microscopy of DNA in
a cellular context. Moreover, the PALM-caged fluorophore
system circumvents the redox buffer conditions necessary for
dSTORM and the sculpted light requirement of STED.
Molecular maps of cellular DNA will enable precise identifi-
cation of DNA–protein interactions and determination of the
location of these complexes within the cell.[19]
In summary, we have established a facile and general
strategy to prepare caged rhodamine and fluorescein photo-
activatable labels for super-resolution microscopy. The use of
leuco-dye intermediates overcomes difficulties with reactivity
and solubility, allowing the preparation of these valuable
molecules with reaction conditions accessible to any organic
chemistry laboratory. Our efficient final synthetic step utilizes
a mild oxidant to concurrently remove protecting groups and
oxidize the dye. We note this caged xanthene system is highly
modular, allowing independent tuning of the dye and the
cage. While the caged RhQ and Rh110 dyes already constitute a
potential pair of dyes for multicolor PALM imaging, our
divergent synthetic route will enable the synthesis of xan-
thene and isologous[20] dyes of different wavelengths bearing
cages with tailored chemical and optical properties. These
efforts will create a palette of probes to bolster the diminutive
collection of synthetic photoactivatable fluorescent labels
suitable for super-resolution imaging experiments.[1c,4,21]
Finally, this facile leuco-dye strategy is appropriate for the
installation of other blocking groups and will allow the
construction of a wide range of useful fluorogenic dyes.
average photon yields for the biotin adducts of 8, 15, and 21 as
well as biotinylated mEos2, a photoswitchable protein that is
used widely in PALM imaging and exhibits a localization
precision of 11 nm.[15] Both of the caged rhodamines, 8-biotin
and 15-biotin, showed excellent properties with mean photon
yields exceeding or equivalent to mEos2-biotin (139% and
97%, respectively; Figure S2). The caged Oregon Green 21-
biotin showed lower photostability than the rhodamine dyes
with an average photon yield at 60% relative to mEos2. Thus,
photoactivatable small-molecule fluorophores, especially
caged rhodamines, exhibit photon yields that are comparable
to an established photoactivatable label capable of high
resolution PALM.
Based on the favorable photon yields exhibited by
compounds 8, 15, and 21 we explored the utility of these
dyes in a cellular super-resolution microscopy experiment.
Again taking advantage of the free carboxyl groups in these
molecules, we prepared azide-containing derivatives of rhod-
amines 8 and 15 and fluorescein 21 (Schemes S4–S6). This
allowed metabolic labeling of cellular DNA by first incubat-
ing cells with 5-ethynyl-2’-deoxyuridine (EdU) followed by
the CuI-catalyzed Huisgen 1,3-dipolar cycloaddition between
the alkynyl nucleobase and the caged dye–azide conjugate
(i.e., “click chemistry”).[16] This labeling allowed super-
resolution localization microscopy of labeled cellular DNA
using all three dyes as shown in Figure 1. Diffraction limited
summed total internal reflection fluorescence (TIRF) mi-
Received: July 1, 2011
Revised: August 26, 2011
Published online: September 26, 2011
Keywords: caged compounds · fluorescence · imaging agents ·
.
rhodamine · super-resolution microscopy
[1] a) T. J. Mitchison, K. E. Sawin, J. A. Theriot, K. Gee, A.
Mallavarapu, G. Marriott, Methods Enzymol. 1998, 291, 63 –
78; b) L. D. Lavis, R. T. Raines, ACS Chem. Biol. 2008, 3, 142 –
155; c) M. Fernꢀndez-Suꢀrez, A. Ting, Nat. Rev. Mol. Cell Biol.
2008, 9, 929 – 943; d) D. Puliti, D. Warther, C. Orange, A. Specht,
M. Goeldner, Bioorg. Med. Chem. 2011, 19, 1023 – 1029.
[2] E. Betzig, G. H. Patterson, R. Sougrat, O. W. Lindwasser, S.
Olenych, J. S. Bonifacino, M. W. Davidson, J. Lippincott-
Schwartz, H. F. Hess, Science 2006, 313, 1642 – 1645.
Figure 1. PALM images of nuclei from fixed, cryosectioned 3T3 cells
labeled with photoactivatable dyes. Left panel: molecule localization
image; scale bar: 2 mm. Right panel: zoomed localization probability
image/summed-TIRF image; scale bar: 500 nm. a) 8-azide; b) 15-
azide; c) 21-azide.
[3] a) S. T. Hess, T. P. K. Girirajan, M. D. Mason, Biophys. J. 2006,
91, 4258 – 4272; b) M. J. Rust, M. Bates, X. Zhuang, Nat.
Methods 2006, 3, 793 – 796.
[4] a) J. Fçlling, V. Belov, D. Riedel, A. Schonle, A. Egner, C.
Eggeling, M. Bossi, S. W. Hell, ChemPhysChem 2008, 9, 321 –
11208 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 11206 –11209

326; b) V. N. Belov, C. A. Wurm, V. P. Boyarskiy, S. Jakobs, S. W.
Hell, Angew. Chem. 2010, 122, 3598 – 3602; Angew. Chem. Int.
Ed. 2010, 49, 3520 – 3523.
[14] a) G. A. Krafft, W. R. Sutton, R. T. Cummings, J. Am. Chem.
Soc. 1988, 110, 301 – 303; b) T. J. Mitchison, J. Cell Biol. 1989,
109, 637 – 652.
[15] S. A. McKinney, C. S. Murphy, K. L. Hazelwood, M. W. David-
son, L. L. Looger, Nat. Methods 2009, 6, 131 – 133.
[16] A. Salic, T. J. Mitchison, Proc. Natl. Acad. Sci. USA 2008, 105,
2415 – 2420.
[17] C. Flors, C. N. J. Ravarani, D. Dryden, ChemPhysChem 2009, 10,
2201 – 2204.
[18] F. Persson, P. Bingen, T. Staudt, J. Engelhardt, J. Tegenfeldt,
S. W. Hell, Angew. Chem. 2011, 123, 5696 – 5698; Angew. Chem.
Int. Ed. 2011, 50, 5581 – 5583.
[5] J. A. Royall, H. Ischiropoulos, Arch. Biochem. Biophys. 1993,
302, 348 – 355.
[6] W.-C. Sun, K. R. Gee, D. H. Klaubert, R. P. Haugland, J. Org.
Chem. 1997, 62, 6469 – 6475.
[7] R. P. Haugland, M. T. Z. Spence, I. D. Johnson, Handbook of
Fluorescent Probes and Research Chemicals, 6th ed., Molecular
Probes, Eugene, 1996.
[8] T. O. Zaikova, A. V. Rukavishnikov, G. B. Birrell, O. H. Griffith,
J. F. W. Keana, Bioconjugate Chem. 2001, 12, 307 – 313.
[9] C. Kim, P. Misco, Tetrahedron Lett. 1985, 26, 2027 – 2030.
[10] J. Ottl, D. Gabriel, G. Marriott, Bioconjugate Chem. 1998, 9,
143 – 151.
[11] I. S. Ioffe, V. F. Otten, Zh. Obshch. Khim. 1961, 31, 1511 – 1516.
[12] C. C. Woodroofe, M. H. Lim, W. M. Bu, S. J. Lippard, Tetrahe-
dron 2005, 61, 3097 – 3105.
[13] X. Jin, C. Uttamapinant, A. Y. Ting, ChemBioChem 2011, 12,
65 – 70.
[19] J. Yao, R. D. Fetter, P. Hu, E. Betzig, R. Tjian, Genes Dev. 2011,
25, 569 – 580.
[20] K. Kolmakov, V. N. Belov, J. Bierwagen, C. Ringemann, V.
Mꢁller, C. Eggeling, S. W. Hell, Chem. Eur. J. 2010, 16, 158 – 166.
[21] a) G. T. Dempsey, M. Bates, W. E. Kowtoniuk, D. R. Liu, R. Y.
Tsien, X. Zhuang, J. Am. Chem. Soc. 2009, 131, 18192 – 18193;
b) H. D. Lee, S. J. Lord, S. Iwanaga, K. Zhan, H. Xie, J. C.
Williams, H. Wang, G. R. Bowman, E. D. Goley, L. Shapiro, R. J.
Twieg, J. Rao, W. E. Moerner, J. Am. Chem. Soc. 2010, 132,
1642 – 1645.
Angew. Chem. Int. Ed. 2011, 50, 11206 –11209
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 11209