Red-emitting rhodamine dyes for fluorescence microscopy and nanoscopy

Article abstract of DOI:10.1002/chem.200902309

Fluorescent markers emitting in the red are extremely valuable in biological microscopy since they minimize cellular autofluorescence and increase flexibility in multicolor experiments. Novel rhodamine dyes excitable with 630 nm laser light and emitting at around 660 nm have been developed. The new rhodamines are very photostable and have high fluorescence quantum yields of up to 80 %, long excited state lifetimes of 3.4 ns, and comparatively low intersystem-crossing rates. They perform very well both in conventional and in subdiffraction-resolution microscopy such as STED (stimulated emission depletion) and GSDIM (ground-state depletion with individual molecular return), as well as in single-molecule-based experiments such as fluorescence correlation spectroscopy (FCS). Syntheses of lipophilic and hydrophilic derivatives starting from the same chromophore-containing scaffold are described. Introduction of two sulfo groups provides high solubility in water and a considerable rise in fluorescence quantum yield. The attachment of amino or thiol reactive groups allows the dyes to be used as fluorescent markers in biology. Dyes deuterated at certain positions have narrow and symmetrical molecular mass distribution patterns, and are proposed as new tags in MS or LC-MS for identification and quantification of various substance classes (e.g., amines and thiols) in complex mixtures. High-resolution GSDIM images and live-cell STED-FCS experiments on labeled microtubules and lipids prove the versatility of the novel probes for modern fluorescence microscopy and nanoscopy.

Full text of DOI:10.1002/chem.200902309

DOI: 10.1002/chem.200902309
Red-Emitting Rhodamine Dyes for Fluorescence Microscopy and Nanoscopy
Kirill Kolmakov, Vladimir N. Belov,* Jakob Bierwagen, Christian Ringemann,
Veronika Mꢀller, Christian Eggeling, and Stefan W. Hell*[a]
Abstract: Fluorescent markers emitting
in the red are extremely valuable in
biological microscopy since they mini-
mize cellular autofluorescence and in-
crease flexibility in multicolor experi-
ments. Novel rhodamine dyes excitable
with 630 nm laser light and emitting at
around 660 nm have been developed.
The new rhodamines are very photo-
stable and have high fluorescence
quantum yields of up to 80%, long ex-
cited state lifetimes of 3.4 ns, and com-
paratively low intersystem-crossing
rates. They perform very well both in
conventional and in subdiffraction-res-
olution microscopy such as STED
(stimulated emission depletion) and
GSDIM (ground-state depletion with
individual molecular return), as well as
in single-molecule-based experiments
such as fluorescence correlation spec-
troscopy (FCS). Syntheses of lipophilic
and hydrophilic derivatives starting
from the same chromophore-containing
scaffold are described. Introduction of
two sulfo groups provides high solubili-
ty in water and a considerable rise in
fluorescence quantum yield. The at-
tachment of amino or thiol reactive
groups allows the dyes to be used as
fluorescent markers in biology. Dyes
deuterated at certain positions have
narrow and symmetrical molecular
mass distribution patterns, and are pro-
posed as new tags in MS or LC-MS for
identification and quantification of var-
ious substance classes (e.g., amines and
thiols) in complex mixtures. High-reso-
lution GSDIM images and live-cell
STED-FCS experiments on labeled mi-
crotubules and lipids prove the versatil-
ity of the novel probes for modern
fluorescence microscopy and nanosco-
py.
Keywords: dyes/pigments · fluores-
cence · fluorescent probes · isotopic
labeling · microscopy
Introduction
field) optical microscopy. “Nanoscopic” techniques[2] such as
STED (stimulated emission depletion),[2a,d,h] PALM (photo-
activation localization microscopy),[2i,s] STORM (stochastic
reconstruction microscopy),[2j,t–v] and GSDIM (ground-state
depletion with individual molecular return)[3] allowed the
optical resolution to be improved from about 200–350 nm to
20–35 nm by switching between the dark and the bright
states of a fluorescent marker. However, the rhodamine
family lacks water-soluble compounds that combine high
photostabilities and fluorescence quantum yields with ab-
sorbance and fluorescence band maxima lying in the far-red
optical region.
Owing to their high photostabilities, high extinction, high
fluorescent quantum yields, and low degree of triplet forma-
tion, rhodamine dyes are widely used as laser dyes and fluo-
rescent markers for labeling proteins, nucleic acids, lipids,
carbohydrates, toxins, hormones, and other biomolecules.[1]
Besides their use in biological imaging and single-molecule-
based spectroscopy, rhodamines served well in the practical
implementation of some new physical concepts, for example,
those that helped to overcome the diffraction limit in (far-
Biological applications favor fluorescent dyes that absorb
in the far-red or even near-infrared (IR) optical region. In
contrast to visible and above all ultraviolet, excitation at
wavelengths longer than 600 nm is mostly noninvasive and
minimizes the unwanted background signal originating from
cellular autofluorescence. The most convenient near-IR
laser lines are between 630 and 650 nm, as supplied by He–
Ne, diode, or krypton ion lasers. A number of fluorescent
dyes have been prepared to match these excitation sour-
ces.[4] However, a survey of the patent literature revealed
[a] Dr. K. Kolmakov, Dr. V. N. Belov, Dipl.-Chem. J. Bierwagen,
Dr. C. Ringemann, Dipl.-Phys. V. Mꢀller, Dr. C. Eggeling,
Prof. Dr. S. W. Hell
Department of NanoBiophotonics
Max Planck Institute for Biophysical Chemistry
Am Fassberg 11, 37077 Gçttingen (Germany)
Fax : (+49)5512012505
E-mail: vbelov@gwdg.de
shell@gwdg.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200902309.
158
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 158 – 166

FULL PAPER
only one commercially available rhodamine dye, namely,
Alexa 633, which was initially reported as a “sulfonated
rhodamine derivative”.[5a] Its exact structure was disclosed in
2007.[5b] It can be deduced that the photostability of
Alexa 633 is similar to those of the cyanine dyes.[4]
cal assays or for proper cellular injection. In addition, being
lipophilic, Atto 647N produces a considerable background
in our immunostaining experiments largely due to its affinity
for mitochondria. Moreover, when coupled to an antibody
for immunostaining, Atto 647N sometimes displays a strong
increase in the intensity of an additional absorption peak at
around 605 nm that gives no emission at all. The complex
structure of Atto 647N and the other carbopyronine dyes
precludes any further chemical modifications aimed at in-
creased polarity and better solubility in water. On the other
hand, the cyanine dyes Alexa 647 and Cy 5 and the rhoda-
mine Alexa 633 are less photostable and have lower fluores-
cence quantum yields than Atto 647N. The following order
of the fluorescence quantum yields and photoresistance was
observed: Atto 633ꢀAtto 647N>Alexa 647>Alexa 633>
Cy 5.[4] Further, cyanine dyes have rather short fluorescence
lifetimes of about 1 ns, and we can rule them out as poten-
tial lead structures for further improvements. The short life-
times of the excited states, presumably low fluorescence
quantum yields, and moderate photostabilities of oxazine
dyes such as Evoblue 30 (for structures, see ref. [4]) make
them poor candidates for further optimization.
Here we present novel rhodamine dyes, excitable at
630 nm and emitting at around 660 nm. They are highly pho-
tostable, have fluorescence quantum yields of up to 80%,
relatively long fluorescence lifetime (3.4 ns), and low inter-
system-crossing rates. The same chromophore-containing
scaffold allows introduction of various functional groups and
preparation of fluorescent dyes ranging from lipophilic to
highly water soluble (hydrophilic), as well as reactive dyes
for specific biological labeling. High-resolution GSDIM
images of the labeled microtubular network in whole (fixed)
cells and live-cell, single-molecule-based STED-FCS (fluo-
rescence correlation spectroscopy) experiments with lipids
illustrate the potential of the novel markers in modern fluo-
rescence microscopy.
Results and Discussion
The commercially available dyes mentioned above are
widely used nowadays in numerous microscopic and nano-
scopic studies. Attempts to devise and improve photostable
red-emitting dyes of other classes are being made in various
research groups. Most recent publications on this topic de-
scribe water-soluble terrylenediimides,[8a,b] new hydrophilic
BODIPY derivatives,[8c] and dicyanomethylene dihydro-
furans.[8d] However, some important data on photophysical
properties of the bioconjugates and microscopic applications
of terrylene diimides and dicyanomethylene dihydrofur-
ans[8a,b,d] are lacking. Therefore, here we consider and refer
to commercially available dyes (mentioned above and in the
Supporting Information), whose properties and performance
are well described and are always possible to explore.
Background information on red fluorescent dyes: Fluores-
cence microscopy, nanoscopy, and FCS demand markers
with “red” excitation and emission bands, high fluorescence
quantum yields (Ffl), high oscillator strengths (absorption
coefficients), high photostability, low rates of formation of
the “dark” triplet state, relatively long lifetimes of the excit-
ed states (>3 ns), sufficient solubility in water, and a reac-
tive group with a linker for conjugation with biological ob-
jects or other structures of interest. Moreover, the availabili-
ty of lipophilic and hydrophilic derivatives of the same chro-
mophore is desirable, as long as it provides additional flexi-
bility in labeling substances of various polarities.
Hydrophobicity would be advantageous for labeling polar
substances such as lipid head groups, while lipophilic deriva-
tives are useful for labeling nonpolar domains (e.g., lipid
acyl chains).
Commercially available red-emitting dyes include the pre-
viously mentioned rhodamine Alexa 633, the cyanine dyes
Alexa 647 and Cy 5, oxazines such as Evoblue 30, and the
carbopyronine dyes Atto 647N and Atto 635, designed to
match the preferred characteristics. Carbopyronine dyes are
structurally similar to rhodamines and can serve as their
competitive substitutes in the far-red spectral region.[6] In
particular, Atto 647N has gained popularity as a dye for la-
beling in “nanoscopic” studies. Being very photostable,
Atto 647N has quite recently been applied for live-cell
video-rate STED imaging.[7] In the course of the study, the
movements of synaptic vesicles inside the axons of cultured
neurons were recorded with a spatial resolution of about
60 nm. However, the low polarity of Atto 647N is often dis-
advantageous.[4] The Atto 647N-labeled molecules may
strongly stick to glass, for example, to microscope cover
slides or to the walls of microcapillary injection tubes. As a
consequence, Atto 647N proves difficult for in vitro biologi-
The path to novel red-emitting rhodamines: Having ruled
out cyanines, oxazines, and carbopyronines as lead struc-
tures for the reasons mentioned above, we ultimately chose
rhodamines as development candidates. Although quite a
number of diverse rhodamines with a great variety of sub-
stituents have been explored, certain possibilities to shift the
emission of rhodamine dyes to the red region still remain.
A very large bathochromic shift was reported for Rhoda-
mine 700 with the skeleton of the well-known Rhoda-
mine 101, in which the benzoic acid residue is replaced by a
trifluoromethyl group. All rhodamines with a perfluoroalkyl
group at the 9-position absorb and emit above 600 nm.[8e]
Unfortunately, the presence of a small and very strongly
electron accepting group at C-9 of the xanthene fragment
(“opposite” to the oxygen atom) makes this position very
vulnerable to nucleophilic attack by water. As a result, such
rhodamines decolorize rapidly in aqueous solution and they
cannot be used as scaffolds.
On the other hand, Rhodamine 101 (Rh 101, for structure
see refs. [1a–c,10]) is one of the most stable and brightest
Chem. Eur. J. 2010, 16, 158 – 166
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
159

V. N. Belov, S. W. Hell et al.
fluorophores known (Ffl =
1[10]). Due to the high degree
of substitution and the planari-
ty of the molecule, its carboxyl
group is sterically hindered,
and therefore coupling reac-
tions with amines require dras-
tic conditions or powerful acti-
vating agents.[11] Importantly,
amidation of Rh 101 with sec-
ondary amines results in a red-
shift of the absorption and
emission bands from 560 and
589 nm to 583 and 604 nm, re-
spectively, with no significant
change in Ffl.[12]
There are further tools for
achieving the desired redshift.
First, the excitation and emis-
sion bands are considerably
shifted to the red if tetrafluor-
ophthalic anhydride is used for
rhodamine synthesis instead of
phthalic anhydride.[13] Second,
a C=C bond may be intro-
duced into the six-membered
ring, fused with the xanthene
fragment in such a way that it
conjugates with the adjacent
benzene ring. By these means,
the highest reported values of
the adsorption and emission
maxima (630 and 655 nm, re-
Scheme 1. Synthesis of lipophilic (5, 7) and hydrophilic (6) rhodamines. a) BrACTHNUGTRNE(UNG CH2)3Cl, Na2CO3, KI, CH3CN,
reflux, 25 h; b) 48% aq. HBr, glacial AcOH, 135–1408C, 22 h; c) tetrafluorophthalic anhydride, toluene,
reflux, 5 h; d) POCl3, ClCH2CH2Cl, 63–658C, 3.5 h; e) POCl3, ClCH2CH2Cl, reflux, 2.5 h; f) Et3N, CH3CN, RT;
g) 10% aq. KOH, THF, 08C; h) 2m HCl in 1,4-dioxane, RT; i) 96–98% H2SO4, RT, 40 h.
spectively) for
rhodamine were achieved in
the rigidized xanthene derivative of structure
a
“regular”
4
(see
for aromatic methyl esters.[15] Conversion of the intermedi-
ate 2-H to rhodamine 4 was performed stepwise according
to the known method for the synthesis of asymmetric xan-
thene derivatives.[9a] Condensation of 3 and 2-H in the pres-
ence of POCl3 affords the corresponding acid chloride as an
intermediate. Aqueous workup leads to zwitterionic com-
pound 4, while quenching the reaction with methanol produ-
ces methyl ester 7, easily hydrolyzed to 4 under the proper
conditions. Other esters of rhodamine 4 can be easily ob-
tained in a similar way. In the long run, they may be inter-
esting intermediates, because one of the fluorine atoms in
the aromatic ring can be easily replaced with nucleophilic
(secondary) aliphatic amines.[2q] Such transformations may
open a new route to diverse fluorescent dyes. With com-
pound 4 as scaffold, attachment of the N-methyl-b-alanine
bridge to the rhodamine core (see Scheme 1 for the synthet-
ic sequence and structures) provided the crucial redshift (>
20 nm in a MeOH solution, see Table 1). Moreover, the ster-
ically unhindered carboxyalkyl or carboxyl groups in com-
pounds 5 are available for further smooth and high-yielding
transformations. Note that in compound 4 the positions of
the absorbance/emission maxima (616 and 641 nm in
Scheme 1).[9] However, 4 is unsuitable for bioconjugation. It
is poorly soluble in water and lacks a convenient binding
site. Moreover, its free carboxyl group may form colorless
and nonfluorescent cyclic esters (lactones). This carboxyl
group is sterically hindered, as in all rhodamines, and amida-
tion reactions with primary amines (e.g., in proteins) would
give primary amides that are known to form colorless and
nonfluorescent cyclic spiroamides (due to addition of the
NH group across the tetrasubstituted C9=C8a/8b bond in the
central xanthene ring).[2m,n,o]
Synthesis of tetrafluororhodamine 4 and its lipophilic and
hydrophilic derivatives: We chose compound 4 as a scaffold
for further derivatization. Synthesis of tetrafluororhodamine
4 (Scheme 1) started with 1, which was prepared from m-
anisidine and acetone in the presence of dry ytterbiumACHTUNTRGNEUNG(III)
triflate according to the known general method.[14] Alkyla-
tion of 1 with 1-bromo-3-chloropropane was carried out sim-
ilarly to the procedure described for its closest analogue, the
corresponding 10-pivaloyl ester.[9a] Demethylation of tricy-
clic ether 2-Me was carried out under conditions reported
160
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 158 – 166

Red-Emitting Rhodamine Dyes
FULL PAPER
Table 1. Spectroscopic properties of the fluorescent dyes prepared in this
study.[a]
water is high: up to 40–50 mg in 1 mL of water at room tem-
perature. The dark blue aqueous solutions produce intense
red fluorescence when irradiated with incandescent or halo-
gen lamps. Dissolution of 6 in methanol, aqueous methanol,
or any methanol-containing mixtures is accompanied by no-
ticeable esterification of the b-alanine fragment by metha-
nol. Probably, the close proximity of the sulfonic acid resi-
dues to the carboxyl group assists nucleophilic addition (a
case of an intramolecular acid catalysis). Unlike 6, lipophilic
rhodamine 5 is very poorly soluble in water. On the other
hand, the solubility of dye 5-H in DMF, THF, and even
CHCl3 is very good. Due to the presence of the asymmetri-
cally substituted secondary amide fragment, the molecules
of compounds 5 and 6 (and their derivatives, see Scheme 1)
Com- lmax lmax
Solvent
e
Ffl. [%] tfl
kISC/T
[10ꢁ6 sꢁ1
]
[c]
pound
G
A
[10ꢁ5 mꢁ1cmꢁ1
]
in
[ns]
N
H2O[b]
4
7
616
632
MeOH
655.5 MeOH
661 MeOH 0.92
661.5 MeOH 0.66
661.5 MeOH 0.73
63
5-tBu 638
5-Me 638
61
62
53
80
78
5-H
638
637
637
3.6
6[d]
660
660
H2O
H2O
0.94
0.92
3.4 2.5/0.3
3.4 2.5/0.3
6D
[a] lmax (abs.), lmax (fl.), e, Ffl, and tfl are absorption maxima, emission
maxima, extinction coefficient, fluorescence quantum yield, and lifetime,
respectively. [b] Atto 633 was used as reference (Ffl =64% in aqueous
buffer). [c] Intersystem crossing rate kISC and rate of T1!S0 transition kT
measured by FCS under single-molecule conditions in water at 228C.[17]
[d] The fluorescence quantum yields of bioconjugates were 48% (for dye
6 and sheep anti-mouse antibodies) and 40% (for dye 6 and goat anti-
rabbit antibodies).
ꢁ
have an N C chirality axis. The rotational barrier in asym-
metrically substituted tertiary amides is high enough (ca.
18 kcalmolꢁ1) that in the NMR spectrum two signals are ob-
served for the N-methyl group (in ca. 1:2–1:3 ratio). The
presence of the chirality axis makes the “opposite” methyl
groups and the CH atoms attached to the “left” and “right”
sides of the xanthene fragment nonequivalent (diastereotop-
ic), and consequently in the NMR spectra of compounds 5
and 6 two additional sets of equally strong signals can be ob-
served. For example, four signals of the methyl groups at
MeOH, respectively, see Table 1) are far from suitable. For
attaching the bifunctional b-alanine “bridge” to the carboxyl
group in 4 for providing the lipophilic fluorescent dyes 5-R,
we used our previously developed technique.[2p] We estab-
lished that utilization of N-methyl-b-alanine tert-butyl ester
instead of the methyl ester has certain advantages. In con-
trast to n-alkyl esters, tert-butyl esters are easily cleaved
under acidic conditions. Thus, in our synthesis of sulfonated
water-soluble hydrophilic dye 6, the subsequent alkali-assist-
ed saponification step becomes unnecessary. The absence of
this step not only shortens the reaction sequence, but also
prevents undesired fluorine substitution in the aromatic ring
of dyes 5-R. (Under basic conditions in methanol one of the
four fluorine atoms may be substituted by a strongly nucleo-
philic methoxide or hydroxide anion.) Another crucial step
of the synthesis is introduction of the two sulfo groups. This
is achieved by direct sulfonation, a straightforward proce-
dure which may be considered to be a case of so-called allyl-
ic sulfonation of olefins.[2r,13] Introduction of ester 5-tBu or
acid 5-H into a large excess of 96–98% H2SO4 at 08C fol-
lowed by prolonged exposure at room temperature afforded
the desired hydrophilic dye 6. Fortunately, these conditions
(concentrated H2SO4) left the secondary amido group intact.
Also importantly, the same starting material, tert-butyl ester
5-tBu, can be smoothly converted to the corresponding car-
boxylic acid 5-H by simple treatment with 2m HCl in 1,4-di-
oxane. In fact, the unsulfonated compound 5-H is the re-
quired lipophilic analogue of rhodamine 6. Later we man-
aged to cleanly saponify methyl ester 5-Me to carboxylic
acid 5-H using an excess of highly dilute aqueous KOH so-
lution at room temperature. Nonetheless, the absence of
methanol in the reaction mixture was important; with meth-
anol, the saponification reaction invariably gave mixtures
containing 5-H and a substance with very similar retention
parameters (both on silica gel and the reversed phase), pre-
sumably the derivative whose aromatic fluorine atom was
replaced by a methoxyl group. The solubility of 6-H in
1
the C=C bond are clearly seen in H NMR spectra of esters
5-Me and 5-tBu.[19]
Deuterated analogues: We also prepared the deuterated an-
alogue of fluorescent dye 6 (compound 6D in Scheme 1,
chemically identical to compound 6) for the following rea-
ꢁ
sons. High-frequency C H stretching vibrations (n
ꢂ3000 cmꢁ1) considerably contribute to the nonradiative
S1!S0 transition characterized by the Frank–Condon factor
f0, which is associated with the maximum possible rate of
nonemissive decay (f0 ꢂ1013 sꢁ1 in most chromophores with
[16]
ꢁ
C H bonds).
Therefore, isotopic substitution which cre-
ꢁ
ates C D bonds with considerably lower vibrational energy
(nꢂ2200 cmꢁ1) may reduce the rate of nonradiative deacti-
vation, improve the fluorescence quantum yield, and pro-
long the excited-states lifetime. These effects are likely to be
observed in rigid molecular systems (e. g., rhodamines) in
which the S1 and S0 states are known to have very similar
geometries, so that their potential-energy surfaces do not in-
tersect. Consequently, the overall deactivation rate may be
approximated by the internal conversion, which is deter-
mined by the Frank–Condon factor. Furthermore, deuterat-
ed and non-deuterated fluorophores with the same structure
are expected to have the same (or nearly the same) intersys-
tem-crossing rates (S1!T1). In view of all these factors, we
hoped that the deuteration would increase the fluorescent
quantum yields and excited-state lifetimes. Compound 6D
was synthesized similarly to compound 6.
Only a very large excess of [D6]acetone in the reaction
with m-anisidine catalyzed by anhydrous ytterbium
flate affords compound 1-Me with sufficiently high deuteri-
um content. The structural fragment C(CD3)=CD (the allyl
protons) in compound 1-Me is especially prone to the D/H
ACHTUNGTRENNUNG(III) tri-
AHCTUNGTRENNUNG
Chem. Eur. J. 2010, 16, 158 – 166
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
161

V. N. Belov, S. W. Hell et al.
exchange. For this reason, the
subsequent alkylation with Br-
ACHTUNGTRENNUNG(CH2)3Cl was performed in
[D3]acetonitrile in the pres-
ence of Na2CO3 (instead of
NaHCO3). Cleavage of deuter-
ated methyl ester 2-Me was
performed with 48% DBr in
D2O mixed with AcOD. Other
steps leading to compound 5-
tBu did not involve reagents or
solvents with easily exchangea-
ble protons and were per-
formed as for the undeuterat-
ed precursor. The reaction
with 97% D2SO4 in D2O (deu-
terated concentrated sulfuric
acid), due to the fast exchange
of the allyl protons, substan-
tially improved the degree of
deuteration in the fragment C-
ACHTUNGTRENNUNG(CD3)=CD. As a result, we
made up for the undesired ex-
change to hydrogen, and target
compound 6D had satisfactory
isotopic purity and
a very
narrow molecular mass distri-
bution in its mass spectrum
with two major peaks corre-
sponding to the presence of 17
and 18 D atoms (of 18 possi-
ble).
Reactive derivatives: To dem-
onstrate the potential and per-
formance of the new dyes as
fluorescent markers for bio-
conjugation and microscopy it
was necessary first to prepare
corresponding reactive deriva-
tives (starting from compounds 6/6D and their lipophilic
counterpart 5-H for comparison). Usually, amino- or thiol-
reactive derivatives of fluorescent dyes are used for labeling
of proteins. For conjugation with amines, the free carboxyl
group must be activated by formation of N-hydroxy-
Scheme 2. Synthesis of reactive derivatives 8, 9, and 11 of lipophilic (5-H) and hydrophilic (6/6D) rhodamine
dyes. a) 2,3,5,6-Tetrafluorophenol, N-ethyl-N’-(3-dimethylaminopropyl)carbodiimide (EDC), CH3CN, RT;
b) N,N,N’,N’-tetramethyl-O-(N-succinimidyl)uronium tetrafluoroborate (TSTU), Et3N, DMF, RT; c) O-(7-aza-
benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexafluorophosphate (HATU), DMF, Et3N, RT;
d) 10·CF3COOH, DMF, Et3N, RT, e) Et3N, DMF, MeOH, RT, overnight.
atively charged OHꢁ ion is invariably formed in the first
step of any esterification of the carboxyl group in 5-H. If it
could be exchanged quickly and under mild conditions with
another, less nucleophilic anion (e.g., perchlorate) before
workup of the reaction mixture, then the chances to isolate
the final ester 8 (or the corresponding NHS ester) in a pure
state would have been better. The NHS esters of hydrophilic
rhodamines 6 and 6D, compounds 9 and 9D, turned out to
be comparatively stable. They were isolated by reversed-
phase preparative HPLC with gradient elution with aqueous
acetonitrile containing 0.1% (v/v) of TFA, lyophilized, and
stored at ꢁ208C under argon. Under these conditions, the
content of the corresponding acids 6/6D (the hydrolysis
products) in these dark blue solids increased from 3–5% up
to 7–10% over two months. The higher stability of the NHS
esters 9 and 11 may be explained by the absence of any nu-
cleophilic counterion (the compounds themselves exist in
ACHTUNGTRENNUNG(sulfo)succinimidyl (NH(S)S) esters, phenyl esters with elec-
tron-acceptor groups in the benzene ring, or similar com-
pounds with good leaving groups. The corresponding NHS
ester of lipophilic rhodamine 5-H proved to be disappoint-
ingly unstable, and tetrafluorophenyl ester 8 was prepared
instead (Scheme 2). Ester 8 partly decomposed in the course
of chromatographic isolation on silica gel. Therefore, the re-
action mixture containing the active ester 8 was used for la-
beling d-erythro-sphingosine phosphocholine (lyso-SM,
Scheme 2). Low chemical stability of ester 8 and the corre-
sponding NHS derivative may be explained by the presence
of the nucleophilic hydroxide group as counterion. The neg-
162
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 158 – 166

Red-Emitting Rhodamine Dyes
FULL PAPER
the form of zwitterions, with sulfonate as anion). In the re-
action with lyso-SM, active ester 9 afforded 12-SO3H
(Scheme 2). Maleimide derivative 11 was prepared directly
from 9 and 10 by combining them in DMF with Et3N as
base. It was also isolated by HPLC. The maleimide ring in
11 proved to be sensitive towards methanol even at room
temperature (e.g., on dissolving the sample in methanol
during preparation of the mass-spectroscopic sample).
quantum yields after conjugation with antibodies is normal
and well-described.[2p]
Mass spectra of the deuterated analogues: Compound 6D,
the deuterated analogue of dye 6 (C42H41F4N3O10S2, M=
887), has up to 18 hydrogen atoms replaced with deuterium
at certain aliphatic groups and at C=C bonds (Schemes 1
and 2). As a result, in the negative-mode ESI mass spectrum
the highest peak with m/z 904 [MꢁH] corresponded to the
molecular formula C42H23D18F4N3O10S2 (M=905). The pres-
ence of 18 deuterium atoms indicates the maximal possible
degree of substitution with deuterated acetone as precursor.
However, peaks with lower masses are present, too. Molecu-
lar mass distribution (negative mode) of the molecular ion
peaks in compound 6D was the following: m/z (%) 900 (15),
901 (36), 902 (67), 903 (94), 904 (100), 905 (46), 906 (18).
After subtraction of the natural deuterium content, this mo-
lecular mass distribution corresponds to the following num-
bers of new deuterium atoms introduced in the course of
the labeling procedures: 901 (D14): 8%; 902 (D15): 13%; 903
(D16): 24%; 904 (D17): 43%; 905 (D18): 5%.[19] Molecular
mass distribution patterns in compounds 6D and 9D were
found to be identical, and no stable isotope “leakage” was
observed in the course of the chemical transformations. Sim-
ilar molecular mass distribution patterns can be easily found
in the positive-mode ESI mass spectra of the sulfonated
rhodamines 6D and 9D, too. However, in this mode these
spectra are complicated by clustering with sodium and other
positively charged ions. Compounds 6D and 9D have the
same narrow molecular mass distribution patterns. There-
fore, all their derivatives can be easily recognized and iden-
tified by mass spectra. Their characteristic feature is the
symmetry of the molecular ion peaks due to the presence of
the ions with lower molecular masses.[19] Moreover, the se-
lective derivatization of amines (with compound 9D) or
thiols (with the deuterated analogue of the maleimide 11)
may enable group identification in biologically important
substance classes simply by comparison of the ESI mass
spectra in the native sample before and after the reaction
with the pairs of reagents 9 and 9D or 11 and its deuterated
analogue. Importantly, the deuterated reagents (e.g., 9D)
produce not only a new molecular mass distribution pattern
in the derivatives, but also develop a considerable isotopic
shift of 17–18 Da. Therefore, the new peaks do not overlap
with the peaks of the derivatives obtained from compounds
9 or 11. It may even be advantageous to use 9 and 9D (or 11
and its deuterated analogue) as mixtures for derivatization
of proteins, tryptic peptides, and other complex biological
samples. For very complex mixtures, HPLC separation prior
to mass spectroscopic analysis may be used. We established
that under reversed-phase HPLC conditions compounds 6
and 6D, as well as 9 and 9D, are inseparable and have the
same retention times. Thus, their derivatives are also expect-
ed to give the same peaks under HPLC conditions. An im-
portant feature is the lower peak intensity of the molecular
ions of the deuterated substances (6D and 9D) compared
with the undeuterated analogues (6 and 9) at the same con-
Photophysical properties of the new rhodamine derivatives:
Spectroscopic properties of fluorescent dyes 4–7 are given in
Table 1. Lipophilic compounds 4, 5, and 7 were studied in
methanolic solutions, and hydrophilic compounds 6/6D in
water. All the dyes form solutions of vibrant blue color with
intense red fluorescence, best visible on illumination with an
incandescent lamp. As can be seen in the table, the required
absorption and emission parameters of the lipophilic (5-H)
and hydrophilic (6 and 6D) fluorescent dyes have been ach-
ieved. The absorption and emission maxima were found to
be at 637 and 660 nm, respectively, similar to Alexa 633 and
Atto 647N. The Stokes shift (23 nm) is considerably larger
that that of Alexa 633 (15 nm) and similar to that of
Atto 647N (25 nm). All fluorescence quantum yields were
obtained for aqueous solutions. These values are considera-
bly higher for the hydrophilic (sulfonated) dyes 6/6D (78–
80%) than for the lipophilic substances: Atto 647N (64%)
or rhodamines 4, 5, and 7 (53–63%). Furthermore, in water
dyes 6/6D (or 5-H in MeOH) have relatively long fluores-
cence lifetimes of 3.4 ns, similar to that of Atto 647N
(3.4 ns) and much larger than those of cyanine dyes (ca.
1 ns). The rates of intersystem crossing kISC =2.5ꢂ106 sꢁ1
and triplet depopulation kT =3ꢂ105 sꢁ1 of both dyes 6/6D
are common to rhodamines,[17] and under STED conditions
they are as photostable as Atto 647N.[18] Isotopic substitution
did not affect the spectroscopic properties of the dye. Com-
parison of the photophysical characteristics of 6 with those
of the deuterated analog 6D revealed no changes in fluores-
cence quantum yield, lifetime, or crossing rates to the triplet
state.
Cellular labeling: The sulfo groups in rhodamines 6/6D,
apart from providing excellent solubility in water, prevent
aggregation of the dye molecules in water or aqueous buf-
fers. This is the most probable explanation for the drastic
rise in the fluorescence quantum yield that we observed
(from 53% for the lipophilic compound 5-H [MeOH] up to
80% for the hydrophilic derivative 6 [H2O]). As a result,
these dyes are superior in many respects to Atto 647N. The
polar hydrophilic dyes 6/6D are not integrated into biologi-
cal membranes and can be used at high concentrations in la-
beling experiments.[18] The degree of labeling achieved with
antibodies can be controlled by dosage of active NHS esters
9/9D. Typically anti-rabbit and anti-mouse antibodies were
labeled to a degree of 2.5 to 6. The fluorescence quantum
yields of the conjugates were found to be 48% (for dye 6
and sheep anti-mouse antibodies) and 40% (for the dye 6
and goat anti-rabbit antibodies). Decrease in fluorescence
Chem. Eur. J. 2010, 16, 158 – 166
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
163

V. N. Belov, S. W. Hell et al.
centrations. Thus, the relative content of a deuterated com-
pound in a reagent mixture with an unlabeled counterpart
must be 3–4 times higher, so that the intensities of the two
peaks in the mass spectrum of the derivatized analyte will
be comparable and easy to find. New molecular mass distri-
bution tags introduced here are colored and highly fluores-
cent. These properties are indispensable for analytical pur-
poses, because fluorescence and mass spectroscopy detection
limits are similar (picograms of an analyte can easily be de-
tected).
following excitation with less than 10 kWcmꢁ2 [19]
. Conse-
quently, GSDIM imaging of biological samples stained with
compound 6 showed significantly enhanced resolution over
conventional wide-field imaging. An illustrative example of
a GSDIM image of a cellular b-tubulin network in the
whole PtK2 cells immunolabeled with compound 6 is shown
in Figure 1. Comparison with the conventional wide field
Far-field fluorescence nanoscopy: Compound 6 performs ex-
cellently in far-field fluorescence imaging and especially as a
label in STED[18b] and GSDIM nanoscopy.[3] The spatial res-
olution of standard far-field optical microscopy is limited by
the diffraction of light to about 200 nm for visible light.[2o]
Breaking of the diffraction barrier has been achieved by
STED microscopy by transiently switching off the ability of
a dye to emit everywhere but at isolated subdiffraction-sized
spots.[2a,d,h, 18] In the STED method, switching off is realized
by stimulated emission induced by an additional laser light.
Effective fluorescence emission is constrained to a subdif-
fraction spot by overlaying the excitation focus with an in-
tensity distribution for stimulated emission featuring a local
zero.[2a,d,h] A high-resolution image is usually formed by
scanning the overlapped foci over the sample and recording
the fluorescence on a point detector. The final spatial reso-
lution is determined by the intensity of the stimulated emis-
sion laser. The GSDIM method,[3] on the other hand, relates
to single-molecule techniques such as PALM (photo-activat-
able localization microscopy)[2 i,s] or STORM (stochastical
optical reconstruction microscopy),[2j,k,t–v] in which imaging
of the stochastically switched-on single emitters is per-
formed. Spatial position of the single isolated molecules is
determined with subdiffraction precision from the signal de-
tected with a camera. Plotting of the positions of repeatedly
switched-on and -off molecules creates the final high-resolu-
tion image. Here, the spatial resolution depends on the
number of photons detected from an isolated single mole-
cule. In contrast to PALM and STORM, which apply special
photoswitchable or photoactivatable labels, in the GSDIM
method photoswitching is performed with conventional fluo-
rophores by randomly shelving more than 90% of the mole-
cules at a time into a dark state such as the triplet state.[3]
Over time, single molecules blink on and off. To achieve the
desired switch-off level of greater than 90%, the lifetime of
the dark state must be long (milliseconds to seconds), which
in the GSDIM approach is usually achieved by embedding
the dyes in poly(vinyl alcohol), PVA, or using oxygen-con-
sumption systems and exciting with a laser intensity greater
than 10–100 kWcmꢁ2. Recently, other approaches based on
reducing and oxidizing agents were presented for oxazine
and rhodamine dyes.[21] Compound 6 combines the parame-
ters most important for GSDIM: high quantum yield to
detect a large number of photons per single-molecule event
and optical shelving of 97% of all molecules into their dark
(triplet) state (with an average lifetime of ca. 50 ms in PVA)
Figure 1. Conventional wide-field (left) and GSDIM (right) images of b-
tubulin strands in whole fixed PtK2 cells stained with sheep anti-mouse
secondary antibodies labeled with 6 (taken as NHS ester 9). The spatial
resolution is approximately 300 nm in the wide-field image and 30 nm in
the GSDIM image.
image indicates that the optical resolution is tenfold im-
proved down to about 30 nm. Features of the network
hidden in the wide-field image can now clearly be resolved.
Besides the protruding tubulin fibers, some randomly dis-
tributed fluorescent spots are detected in the GSDIM
image. This is a well-known feature of nanoscopically re-
solved immunostained samples and is most likely explained
by unspecific antibody binding.[2m,o,r,3]
Compound 6 performs much better in GSDIM than other
red-emitting dyes such as Atto 647N or Alexa 647 and
Cy 5.[19] Atto 647N is unfavorable for GSDIM since its dark-
state transition probability is too low (kISC <106),[17b] and
consequently only up to 60% of all molecules can be
shelved into the dark at a time.[3] In contrast to the 97% of
dark-state shelving for 6, up to 75 and 90% of Alexa 647
and Cy5 molecules, respectively, can be switched off at a
time in PVA; the average dark-state lifetime is about 50 ms
for 6 and only 20–25 ms for the mentioned cyanine dyes.[19]
Moreover, the laser intensity needed to attain this maximum
switch-off ratio is less than 10 kWcmꢁ2 for 6 but 40 kWcmꢁ2
for Cy 5. Also, photobleaching is more pronounced for the
cyanine dyes: After ten switch-on/-off cycles more than
20% of all Cy 5 or Alexa 647 molecules are photobleached,
while almost 90% of 6 is still fluorescent.[19] High photosta-
bility is favorable in the GSDIM protocol for reaching the
highest possible contrast or for repetitive studies of dynami-
cal events
STED-FCS: We performed STED-FCS experiments on
sphingomyelin (SM) lipids labeled with hydrophilic com-
pound 6 or the lipophilic compound 5-H (Scheme 2).
Recent FCS studies on the dynamics of Atto647N-labeled
164
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 158 – 166

Red-Emitting Rhodamine Dyes
FULL PAPER
SM in the plasma membrane of living PtK2 cells revealed
that SM is transiently (ca. 10 ms) trapped on small spatial
scales (<20–30 nm) due to cholesterol-assisted complexa-
tion to other membrane constituents such as proteins.[22]
Trapping could, however, only be observed in spots smaller
than 100 nm formed by STED; conventional confocal spots,
about 250 nm in size, were too large to distinguish normal
from heterogeneous SM diffusion. Figure 2 shows FCS data
that expected from the model based on free-diffusing spe-
cies. The diffusion constants D=d2/(8ln2tD) calculated
from the focal diameters d and the decay times of the corre-
lation curves are much smaller for the STED than for the
confocal recordings (Figure 2) and render the heterogeneous
diffusion of the SM lipids in the plasma membrane of living
cells due to the mentioned nanoscale trapping: the decay of
the STED-FCS data is dominated by trapping. Most strik-
ingly, the STED-FCS data recorded for the SM analogues
labeled with dyes of different polarities (5-H is slightly more
lipophilic than Atto 647N, and both are much more lipophil-
ic than 6), show no significant differences. The trapping-do-
minated diffusion constants D are the same for all three SM
lipids. This negligible influence of the dye on the trapping
parameters strongly supports the conclusion that the ob-
served complexes are not caused or even influenced by the
dye label. Trapping seems to be assisted by molecular
groups of the lipid such as the hydrogen bonds of the sphin-
gosine residue. Consequently, our observations are very
likely to reflect the true behavior of the endogenous SM.[21]
Figure 2. STED-FCS measurements on fluorescent SM lipid analogues in
living Ptk2 cells: Correlation data of ATTO 647N- (black), 6- (red), and
compound 5-H-labeled SM (blue) for confocal (solid lines “Conf.”) and
STED recording (dotted lines “STED”), that is, focal spots of about
240 nm and about 45 nm in diameter (left panel) and corresponding aver-
age diffusion constants D calculated from the decay times of the correla-
tion data from confocal (white bars) and STED (gray bars) recordings
(right panel). The lower D in the STED recordings renders heterogene-
ous diffusion of the lipids due to trapping. The extent of trapping (i.e.,
values of D in STED) is the same regardless of the different polarities of
the labels Atto 647N, 5-H (“Cpd. 5-H”) or compound 6 (“Cpd. 6”). The
error bars reveal the standard deviation from the analysis of more than
30 correlation curves recorded on different parts of the same cell and
from different cells.
Conclusions
New fluorescent dyes and markers for imaging and labeling
in the red optical region were devised. They may be excited
with irradiation at 633/635 nm, emit light at around 660 nm
with high fluorescence quantum yields in solution in the free
state or when conjugated with biological substrates. The
dyes are very photostable and have low intersystem-crossing
rates and relatively long (3.4 ns) excited-state lifetimes. We
prepared hydrophilic (soluble in water and aqueous buffers,
6/6D) and lipophilic (hydrophobic, insoluble in water or
aqueous buffers, e.g., 5-H) forms of the dyes. This feature
would have been much more difficult to achieve with the
spectroscopically similar carbopyronines, because the prepa-
ration of hydrophilic carbopyronines is a tedious and multi-
step procedure. A carboxyl reactive site for conjugation
with various functional groups (e.g., amines or thiols) is in-
troduced by using tert-butyl 3-(N-methylamino)propionate
as a bifunctional intermediate. If sulfonation follows, the
saponification step is skipped. Compounds 6/6D perform ex-
cellently in far-field fluorescence nanoscopy techniques such
as STED or GSDIM and are equally ideal for STED-FCS
studies on diffusing lipids in the plasma membrane of living
cells. We expect the novel dyes to be not only versatile
labels for nanoscopy and single-molecule based studies but
also helpful tools for multicolor experiments that can bring
forth new insights into physical, chemical, or biological
problems.
of fluorescent SM analogues labeled with Atto 647N, 6, or
5-H incorporated into the plasma membrane of living PtK2
cells for conventional confocal microscopy and high-resolu-
tion STED nanoscopy recordings. Diffusion of the fluores-
cent lipids through the focal spots with a diameter of about
240 nm (confocal) and about 45 nm (STED), which were
placed on the plasma membrane, resulted in FCS correlation
data dropping off with a decay time characteristic of the
lipid focal transit time. For the confocal recordings the
decay time of the correlation data recorded for 6-labeled
SM is slightly shorter than for 5-H- and Atto647N-labeled
SM. Labeling was performed by replacing the long native
acyl chain at the second position of the SM molecule by a
short acyl chain carrying the dye label, and thus compounds
12-H and 12-SO3H (Scheme 2) were synthesized starting
from lyso-SM and active esters 8 and 9. In 12-H the whole
lipophilic dye residue mimics the “second” long alkyl chain
of sphingomyelin, while in 12-SO3H only one fatty acid
chain is present. The higher polarity of 6 results in slightly
less tight membrane anchoring of the corresponding SM an-
alogue and thus in slightly faster diffusion, as was previously
more distinctly revealed for the even more polar dye
Atto 532 attached to lyso-SM.[22] Due to the smaller focal-
spot diameter, the correlation data of the STED recordings
are shifted to shorter times. However, this shift is less than
Acknowledgements
The authors are grateful to Donald Ouw for measuring the optical spec-
tra, fluorescence quantum yields, and degrees of labeling. Excellent assis-
tance of Dr. Ellen Rothermel, who performed the conjugation reactions,
Chem. Eur. J. 2010, 16, 158 – 166
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
165

V. N. Belov, S. W. Hell et al.
is acknowledged. We highly appreciate the assistance of Reinhard Machi-
nek, Dr. Holm Frauendorf, and their co-workers, who recorded numerous
NMR and mass spectra at the Institut fꢀr Organische und Biomolekulare
Chemie (Georg-August-Universitꢃt Gçttingen).
the corresponding rhodamines): a) J. Arden-Jacob, J. Frantzeskos, A.
Zilles (K.-H. Drexhage), WO 64986, 2000; b) A. Zilles, J. Arden-
Jacob, K.-H. Drexhage, N. U. Kemnitzer, M. Hammers-Schneider
(Atto-tec GmbH), WO 003086, 2005.
[7] V. Westphal, S. O. Rizzoli, M. A. Lauterbach, D. Kamin, R. Jahn,
S. W. Hell, Science 2008, 320, 246–249.
[8] For recent examples, see: a) K. Peneva, G. Mihov, F. Nolde, S.
Rocha, J. Hotta, K. Braeckmans, J. Hofkens, H. Uji-i, A. Hermann,
K. Mꢀllen, Angew. Chem. 2008, 120, 3420–3423; Angew. Chem. Int.
Ed. 2008, 47, 3372–3375; b) C. Jung, B. K. Mꢀller, D. C. Lamb, F.
Nolde, K. Mꢀllen, C. Brꢃuchle, J. Am. Chem. Soc. 2006, 128, 5283–
5291; c) S. L. Niu, G. Ulrich, R. Ziessel, A. Kiss, P.-Y. Renard, A.
Romieu, Org. Lett. 2009, 11, 2049–2052; d) Z. Lu, N. Liu, S. J. Lord,
S. D. Bunge, W. E. Moerner, R. J. Twieg, Chem. Mater. 2009, 21,
797–810; e) M. Sauer, K.-T. Han, R. Mꢀller, S. Nord, A. Schulz, S.
Seeger, J. Wolfrum, J. Arden-Jacob, G. Deltau, N. J. Marx, C.
Zander, K.-H. Drexhage, J. Fluoresc. 1995, 5, 247–261.
[9] a) L. G. Lee, R. J. Graham, W. E. Werner, E. Swartzman, L. Lu,
(Apptera Corp., USA), US 6372907, 2002. b) Similar spectral values
(624 and 644 nm for absorption and emission, respectively) have
been recorded in ethanol for ethyl esters of the tetrachlororhodami-
nes.[6b]
[10] T. Karstens, K. Kobs, J. Phys. Chem. 1980, 84, 1871–1872.
[11] T. Nguyen, M. B. Francis, Org. Lett. 2003, 5, 3245–3248.
[12] M. Bossi, V. N. Belov, S. Polyakova, S. W. Hell, Angew. Chem. 2006,
118, 7623–7627; Angew. Chem. Int. Ed. 2006, 45, 7462–7465.
[13] For an example, see: F. Mao, W.-Y. Leung, R. P. Haugland (Molecu-
lar Probes Inc., USA), US 6130101, 2000.
[14] L. A. Robinson, M. E. Theoclitou, Tetrahedron Lett. 2002, 43, 3907–
3910 (see also: C. S. Yi, S. Y. Yun, J. Am. Chem. Soc. 2005, 127,
17000–17006).
[1] For a general description and spectral range of rhodamines as fluo-
rescent dyes, see: a) M. Beija, C. A.. Afonso, J. M. G. Martinho,
Chem. Soc. Rev. 2009, 38, 2410–2433; M. S. T. GonÅalves, Chem.
Rev. 2009, 109, 190–212; b) R. P. Haugland, A Guide to Fluorescent
Probes and Labelling Technologies, Invitrogen, Carlsbad, 2005,
pp. 11–37; c) for the use of rhodamines in bioconjugation tech-
niques, see: G. T. Hermanson, Bioconjugate Techniques, Academic
Press, Amsterdam, 2008, pp. 415–430; for laser dye properties of
rhodamines, see: d) F. P. Schꢃfer in Dye Lasers (Ed.: F. P. Schꢃfer),
Springer, Berlin, 1990, pp. 3–4; e) K.-H. Drexhage, Dye Lasers
(Ed.: F. P. Schꢃfer), Springer, Berlin, 1990, p. 158; f) K.-H. Drexh-
age, Dye Lasers (Ed.: F. P. Schꢃfer), Springer, Berlin, 1990, pp. 175–
180; g) K.-H. Drexhage, Top. Appl. Phys. 1973, 1, 155–200.
[2] For reviews, see: a) S. W. Hell, Science 2007, 317, 1749–1753; b) M.
Fernꢄndez-Suꢄrez, A. Y. Ting, Nat. Rev. Mol. Cell Biol. 2008, 9,
929–943; c) D. Evanko, Nat. Methods 2009, 6, 19–20; d) S. W. Hell,
Nat. Methods 2009, 6, 24–32; e) X. Zhuang, Nat. Photonics 2009, 3,
365–367; f) J. Lippincott-Schwartz, S. Manley, Nat. Methods 2009, 6,
21–23; g) B. Huang, M. Bates, X. Zhuang, Annu. Rev. Biochem.
2009, 78, 993–1016; for recent reports, see: h) S. W. Hell, J. Wich-
mann, Opt. Lett. 1994, 19, 780–782; i) E. Betzig, G. H. Patterson, R.
Sougrat, O. W. Lindwasser, S. Olenych, J. S. Bonifacino, M. W. Da-
vidson, J. Lippincott-Schwartz, J. F. Hess, Science 2006, 313, 1642–
1645; j) M. J. Rust, M. Bates, X. W. Zhuang, Nat. Methods 2006, 3,
793–795; k) M. Bates, B. Huang, G. T. Dempsey, X. Zhuang, Science
2007, 317, 1749–1753; l) S. T. Hess, T. P. K. Girirajan, M. D. Mason,
Biophys. J. 2006, 91, 4258–4272; m) J. Fçlling, V. Belov, R. Kunet-
sky, R. Medda, A. Schçnle, A. Egner, C. Eggeling, M. Bossi, S. W.
Hell, Angew. Chem. 2007, 119, 6382–6386; Angew. Chem. Int. Ed.
2007, 46, 6266–6270; n) J. Fçlling, V. Belov, D. Riedel, A. Schçnle,
A. Egner, C. Eggeling, M. Bossi, S. W. Hell, ChemPhysChem 2008,
9, 321–326; o) M. Bossi, J. Fçlling, V. N. Belov, V. P. Boyarskiy, R.
Medda, A. Egner, C. Eggeling, A. Schçnle, S. W. Hell, Nano Lett.
2008, 8, 2463–2468; p) V. P. Boyarskiy, V. N. Belov, R. Medda, B.
Hein, M. Bossi, S. W. Hell, Chem. Eur. J. 2008, 14, 1784–1792;
q) S. M. Polyakova, V. N. Belov, C. Eggeling, S. F. Yan, C. Ringe-
mann, G. Schwarzmann, A. de Meijere, S. W. Hell, Eur. J. Org.
Chem. 2009, 5162–5177; r) V. N. Belov, M. L. Bossi, J. Fçlling, V. P.
Boyarskiy, S. W. Hell, Chem. Eur. J. 2009, 15, 10762–10776; s) H.
Schroff, C. G. Galbraith, H. White, J. Gilette, S. Olenych, M. W. Da-
vidson, E. Betzig, Proc. Natl. Acad. Sci. USA 2007, 104, 20308–
20313; t) M. Bates, B. Huang, G. T. Depsey, X. Zhuang, Science
2007, 317, 1749–1753; u) B. Huang, W. Wang, M. Bates, X. Zhuang,
Science 2008, 319, 810–813; v) B. Huang, S. A. Jones, B. Branden-
burg, X. Zhuang, Nat. Methods 2008, 5, 1047–1052.
[15] R. Hermann, H.-P. Josel, K.-H. Drexhage, J. Arden-Jacob (Boeh-
ringer Mannheim GmbH), US 57550409, 1998.
[16] a) W. Siebrand, D. F. Williams, J. Chem. Phys. 1968, 49, 1860–1871;
b) C. Dosche, M. U. Kumke, H.-G. Lçmannsrçben, F. Ariese, A. N.
Bader, C. Gooijer, O. S. Milanic, M. Iwamoto, K. P. C. Vollhardt, R.
Puchta, N. J. R. van Eikema Hommes, Phys. Chem. Chem. Phys.
2004, 6, 5476–5483; c) M. J. Frampton, G. Accorsi, N. Armaroli,
J. E. Rogers, P. A. Fleitz, K. J. McEwan, H. L. Anderson, Org.
Biomol. Chem. 2007, 5, 1056–1061.
[17] a) J. Widengren, U. Mets, R. Rigler, J. Phys. Chem. 1995, 99, 13368–
13379; b) C. Eggeling, J. Widengren, L. Brand, J. Schaffer, S. Feleky-
an, C. Seidel, J. Phys. Chem. A 2006, 110, 2979–2995.
[18] a) S. W. Hell, V. N. Belov, K. Kolmakov, V. Westphal, M. A. Lauter-
bach, S. Jakobs. C. A. Wurm, C. Eggeling, C. Ringemann (Max-
Planck-Gesellschaft zur Fçrderung der Wissenschaften e.V.), EP-A
09005886.8, 2009; b) R. Schmidt, C. A. Wurm, A. Punge, A. Egner,
S. Jakobs, S. W. Hell, Nano Lett. 2009, 9, 2508–2510.
[19] See the Supporting Information for figures and experimental data.
[20] E. Abbe, Arch. f. Mikroskop. Anat. 1873, 9, 413–420
[21] a) S. Van de Linde, U. Endesfelder, A. Mukherjee, M. Schꢀttpelz, G.
Wiebusch, S. Wolter, M. Heilemann, M. Sauer, Photochem. Photo-
biol. Sci. 2009, 8, 465–469; b) J. Vogelsang, T. Cordes, C. Forth-
mann, C. Steinhauer, P. Tinnefeld, Proc. Natl. Acad. Sci. USA 2009,
106, 8107–8112; c) M. Heilmann, S. Van de Linde, A. Mukherjee,
M. Sauer, Angew. Chem. 2009, 121, 7036–7041; Angew. Chem. Int.
Ed. 2009, 48, 6903–6908.
[3] J. Fçlling, M. Bossi, H. Bock, R. Medda, C. A. Wurm, B. Hein, S.
Jakobs, C. Eggeling, S. W. Hell, Nat. Methods 2008, 5, 943–945.
[4] See the Supporting Information for detailed characteristics of com-
mercially available dyes whose fluorescence can be exited with the
He–Ne red laser (633 nm) or the 635 nm spectral line of red diode
lasers.
[22] C. Eggeling, C. Ringemann, R. Medda, G. Schwarzmann, K. Sand-
hoff, S. Polyakova, V. N. Belov, B. Hein, C. von Middendorff, A.
Schçnle, S. W. Hell, Nature 2009, 457, 1159–1163.
[5] a) J. E. Berlier, J. Histochem. Cytochem. 2003, 51, 1699–1712; b) B.
Agnew, K. R. Gee, T. G. Nyberg (Invitrogen), US 0249014, 2007.
[6] In carbopyronines, as compared to rhodamines, the oxygen atom at
position 10 of the xanthene fragment is “replaced” with a geminal
dimethyl group [C
G
Received: August 20, 2009
in the absorption and emission bands (ca. 50 nm in comparison with
Published online: November 30, 2009
166
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 158 – 166