Malachite Green Derivatives for Two-Photon RNA Detection

Article abstract of DOI:10.1002/cbic.201100747

The design, preparation and characterisation of a library of malachite green (MG) derivatives for two-photon RNA labelling is described. Some of these MG derivatives exhibit an increased affinity for an MG-aptamer, as well as improved two-photon sensitivi

Full text of DOI:10.1002/cbic.201100747

DOI: 10.1002/cbic.201100747
Malachite Green Derivatives for Two-Photon RNA
Detection
Jacques Lux,[a] Eduardo Josꢀ PeÇa,[b] Frꢀdꢀric Bolze,[a] Manfred Heinlein,[b] and
Jean-FranÅois Nicoud*[a]
The design, preparation and characterisation of a library of
malachite green (MG) derivatives for two-photon RNA labelling
is described. Some of these MG derivatives exhibit an in-
creased affinity for an MG-aptamer, as well as improved two-
photon sensitivity when compared to the classical malachite
green chloride. The underlying mechanisms and potential ben-
efits for in vivo RNA visualisation are discussed.
Introduction
Although the sequencing of the genomes of many organisms,
including humans, has been achieved, understanding how this
stored and inherited genetic information is used to organise
and maintain living cells and organisms remains a major chal-
lenge. The discovery of green fluorescent protein (GFP) in the
early 1960s[1] and its subsequent application for nonintrusive
imaging of cellular structures, organelles and proteins facilitat-
ed the functional analysis of genes and proteins in vivo. In-
creasing evidence indicates that cellular functions involve RNA
molecules targeted to specific locations within cells, and that
RNA processing pathways occur in association with specific
subcellular structures or complexes. For example, mRNAs exit
the nucleus only after they have been quality controlled,
spliced, poly-adenylated and capped.[2,3] The subsequent fate
of the mRNA molecules is often associated with proteins that
remain bound as the RNA enters the cytoplasm.[4] Here, the
RNA might be recruited by a signal-recognition particle for
translation at the endoplasmic reticulum surface.[5] Other
mRNAs might be transported along the cytoskeleton[6] or ER
membranes,[7] and during transport, translation might be tem-
porarily repressed to allow protein synthesis localised at final
destination.[8] Also, mRNA molecules might be targeted to spe-
cific storage organelles for later use, and those destined for
degradation are targeted to processing bodies (P-bodies).[9,10]
In addition to mRNAs, various other classes of RNAs function
in the regulation of gene expression and pathogen defence,
through mechanisms commonly known as “RNA silencing”.[11]
These RNA molecules might act in the nucleus or cytoplasm,
and might even have extracellular autonomous functions
(moving between cells and tissues).[11,12] In plants, intercellular
RNA transport is also common to certain RNA viruses that
move their genomes between cells and systemically through-
out plant tissues in a non-encapsidated form to cause systemic
infection.[13] Further insights into the function of RNA mole-
cules in relation to their associations with specific cellular
structures and localisations requires imaging tools with se-
quence-specific detection. Unlike proteins, which can be easily
imaged upon expression as translational fusions to GFP or
other fluorescent adjuncts, RNA molecules require in vitro la-
belling and microinjection for their direct visualisation,[14] or
they can be detected by indirect methods, either through hy-
bridisation with a fluorescently tagged probe (e.g. molecular
beacons),[15] or by sequence-specific binding of fluorescent pro-
teins.[16–19] An as yet not fully explored approach is the use of
aptamer-binding dyes. This RNA detection technique avoids
invasive methods or the binding of bulky proteins that might
interfere with RNA localisation, dynamics or function. Here, the
RNA of interest is tagged with an RNA sequence selected for
binding the cell-permeant dye or fluorophore through its sec-
ondary structure. The aptamer and corresponding fluorophore
are created either by identification of an RNA aptamer able to
bind a specific fluorophore, or by adapting a fluorophore to
bind a specific RNA aptamer. The first approach has been re-
cently applied by Paige et al. in the creation of GFP-like fluoro-
phores.[20] By following the second approach, we here report
the adaptation of malachite green (MG), a well known apta-
mer-binding dye.[21] To distinguish dye molecules bound to
aptamer from unbound dye in vivo, it is important that apta-
mer binding changes the dye properties such that the apta-
mer-bound dye can be detected in the microscope. MG ap-
pears to be particularly suited for this purpose as aptamer
binding induces a red shift in the dye absorbance and increas-
es its fluorescence.[22] A complication of MG aptamer (MGA) la-
belling is the fluorescence background that results from non-
specific stabilisation of a fluorescence-emitting conformation
of the dye.[23,24] In highly diffusive media and highly autofluor-
[a] Dr. J. Lux,+ Dr. F. Bolze, Prof. Dr. J.-F. Nicoud
Laboratoire de Biophotonique et Pharmacologie, UMR 7213 UdS/CNRS
74 route du Rhin, 67401 Illkirch (France)
E-mail: jfnicoud@unistra.fr
[b] Dr. E. J. PeÇa,+ Dr. M. Heinlein
Institut de Biologie Molꢀculaire des Plantes, CNRS UPR 2357
12, rue du Gꢀnꢀral Zimmer, 67084 Strasbourg (France)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201100747.
ChemBioChem 0000, 00, 1 – 9
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

J. F. Nicoud et al.
escent samples such as plant tissues, the signal-to-noise ratio
can be increased by localised excitation, for example by using
two-photon excitation (TPE).[25] By restricting excitation to the
focal point, the application of TPE also has the potential to
reduce the phototoxic effects caused by reactive oxygen spe-
cies produced by MG, other fluorophores and fluorescent pro-
teins.[26,27] To increase the applicability of MG to in vivo imag-
ing, we here report the design and biophysical characterisation
of MG-derived triphenylmethane dyes with improved aptamer
binding and two-photon (TP) brightness.
Scheme 1. Malachite Green (MG), with sites of modification highlighted.
para and/or meta positions of the phenyl ring. In order to
modulate both TPA and the pharmacokinetic properties of the
dye, we also tested the introduction of alkoxy and oligo(ethy-
lene glycol) chains of various lengths (12c–e). In addition, to
further increase the TPA, we extended the p system by re-
placement of the anisyl group of 8a, b with a methoxystyryl
group (14a, b). As the phenyl ring points out from the apta-
mer binding pocket, such long modifications were interesting
to study the size “tolerance” of the adaptive binding mecha-
nism. Modifications at the nitrogen atoms (N) included the ad-
dition of small lipophilic groups (4a) and of a hydroxyl group.
These modifications affect interactions with phosphates in the
RNA aptamer and were expected to influence aptamer binding
(4b).
N-modified MG derivatives were prepared by reaction be-
tween benzaldehyde 1 and the desired substituted anilines
2a, b in acidic conditions. The resulting leuco-malachite greens
3a, b were then oxidised by 2,3-dichloro-5,6-dicyano-1,4-ben-
zoquinone (DDQ) in the presence of an excess of hydrochloric
acid to yield the final MG derivatives 4a and 4b (Scheme 2).
MG derivatives mono-substituted at the phenyl ring 8a–d
were prepared by reaction of Michler’s ketone 5 with substitut-
ed iodo- and bromobenzenes 6a–d in the presence of nBuLi.
The resulting triphenylcarbinol derivatives 7a–d were dehy-
drated in acidic conditions to give the corresponding MGs 8a–
d (Scheme 3).
Results and Discussion
Synthesis
MG is commercially available with various counter anions such
as chloride, oxalate, or sulfate. Some of these, such as oxalate,
are particularly toxic and can interfere with cellular functions.
We prepared MG derivatives with chloride as the counter
anion, and their properties were compared with commercially
available MG chloride (MG-Cl) as reference. The structure of
the MGA and its interactions with MG or with its analogue tet-
ramethylrosamine have been studied by solid-state X-ray crys-
tallography,[28] and in solution by NMR spectroscopy.[29,30] Bind-
ing to the aptamer induces specific conformational changes in
both the RNA and the dye. The RNA folds into a pocket with
the dye ligand at its centre. Specific interactions reduce the
freedom of movement of individual phenyl rings of the dye,
and by aligning two rings in a planar conformation, the conju-
gated double bonds interact and cause fluorescence upon illu-
mination at the appropriate excitation wavelength. MG modifi-
cations have to maintain the electrostatic and base-stacking in-
teractions with the aptamer that are required for the formation
and stabilisation of the complex.[31,32] Nevertheless, as the bind-
ing process is adaptive and stabilised by an induced fit
(“ligand-induced folding” or “adaptative binding”),[33–35] small
modifications are possible, as they will be compensated for by
corresponding small structural changes in the aptamer confor-
mation. Several studies have investigated the structural adap-
tation of aptamer sequences to the corresponding dyes. As
described by Brackett and Dieckmann, the phenyl ring of MG
can be modified without compromising the interaction with
MGA.[36] In the present work, we modified the dye by adding
substituents at the meta and/or para positions of the phenyl
ring. In addition, as the MG dimethylamino groups are inserted
deep in the aptamer-binding pocket, we explored the effects
of changing the methyl substituents at the nitrogen atoms
(Scheme 1).
Modifications at the phenyl ring were mainly performed to
modulate the two-photon absorption (TPA) efficiency. By intro-
ducing electroactive groups to produce a mesomeric or induc-
tive effect (OMe, Me or I in 8a–d and 12a, b) we could induce
a donor-conjugated system-donor architecture that is known
to be efficient in TPA.[36] Furthermore, in the bound state, the
modified phenyl ring adopts a more coplanar alignment with
one of the other rings,[29] thus enhancing TPA by increasing p
electron conjugation. We introduced the substituents at the
Scheme 2. Preparation of N-modified MGs 4a, b. a) p-Toluenesulfonic acid in
benzene, reflux (808C) with a Dean–Stark device, 48 h; b) HCl 10m, DDQ, in
benzene and MeOH, 18 h; yield: 7%.
&2
&
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 9
ÝÝ
These are not the final page numbers!

Lighting RNA
Scheme 3. Synthesis of MGs 8a–d. a) nBuLi 1.6m in THF, À788C, 2 h, then Michler’s ketone in THF, RT, 18 h; b) HCl 10m in CH2Cl2 and MeOH, 758C, 18 h;
yield: 75–100%.
Scheme 4. Synthesis of MG 12a–e. a) Montmorillonite K10, 1008C, 18 h; b) HCl (10m), DDQ or PbO2, in CH2Cl2 and MeOH, RT, 18 h; yield: 5–10%.
One-photon properties
MG derivatives 12a–e were prepared by reaction of N,N-di-
methylaniline 9 with the corresponding benzaldehydes 10a–
e in the presence of montmorillonite K10, followed by PbO2
oxidation of the intermediary leuco-derivatives 11 a–e with an
excess of HCl (Scheme 4).
UV–visible absorption of the MG derivatives was studied in
aqueous solution. Figure 1 shows typical absorption spectra of
MG derivatives 4a, 8a and MG-Cl.
abs
max
The l and e values for the new MG derivatives are listed
In order to increase the length of the conjugated systems to
improve the TP properties, palladium Heck cross-couplings
were performed on the para- and meta-iodo derivatives 8c, d
with para-methoxystyrene 13 to yield MG derivatives 14a, b
(Scheme 5).
in Table 1. Absorption values lie in the 607–632 nm range, thus
abs
indicating a weak effect of the various modifications on l
.
max
The most bathochromically and hypsochromically shifted ab-
sorption bands were found for the isopropyl derivative 4a and
the m-iodo derivative 8d, respectively. We did not find any
correlation between the substitutions on the phenyl ring and
a bathochromic shift due to the increased conjugation length
1
All new compounds were characterised by H and 13C NMR
spectroscopy and high-resolution mass spectrometry. Experi-
mental details and the results of these characterisations are
provided in the Supporting Information.
Scheme 5. Preparation of MGs 14a, b. a) 13, Pd(OAc)2, tri-o-tolylphosphine
in triethylamine and xylene (or DMF), 1428C, 48 h, then PbO2 and HCl (10m)
in CH2Cl2 and MeOH, 18 h; yield: 25%.
Figure 1. Selected UV/Vis spectra of MG derivatives in water. MG-Cl (black),
8a (red) and 4a (blue).
ChemBioChem 0000, 00, 1 – 9
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
3
&
ÞÞ
These are not the final page numbers!

J. F. Nicoud et al.
abs
max
Table 1. UV–visible l and emax of new MG derivatives in HEPES buffer (pH 7.4).
abs
max
abs
max
Compound
l
emax
[mÀ1 cmÀ1
Compound
l
emax
[mÀ1 cmÀ1
]
[nm]
]
[nm]
4a
632
19900
4b
622
31100
8a
607
74100
8b
619
50500
8c
622
64900
8d
607
21100
12a
612
39600
12b
615
34300
12c
619
17000
12d
628
21200
12e
620
24000
14a[a]
624
47900
14b[a]
623
22300
MG-Cl
617
92400
[a] R=4-methoxystyryl:
&4
&
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 9
ÝÝ
These are not the final page numbers!

Lighting RNA
nor the presence of electron donating groups, thus indicating
weak conjugation between this phenyl ring and the two dime-
thylaniline residues. In contrast, molar absorption coefficients
are highly affected by these modifications, as described in a
previous study on fluorinated derivatives of MG.[37] The un-
modified MG chloride showed the highest emax
(92400mÀ1 cmÀ1), with p-iodo and p-methoxy derivatives in the
same high range. A reduced absorption coefficient was ob-
served for the p-methylated, p-methoxyvinylated, and m-me-
thoxylated derivatives. The N-modified MGs showed further
reduced emax values. The lowest absorption coefficients were
shown by the alkoxylated MGs. Very weak fluorescence was
detected for all new derivatives (quantum yields, 10À4–10À5),
thus indicating efficient nonradiative decay in solution except
for 14b, which was slightly fluorescent (in this case, the steric
hindrance of the large substituent in the meta position can
slow the rotation of the phenyl ring, and diminish the nonra-
diative decay rate).
MG derivatives and MGA interaction
Consistent with previous reports, we noticed a significant red
abs
max
shift (15–20 nm) in the l
for MG derivatives 8a–d, 12a–
c and 14a upon addition of MGA, which indicates nearly com-
plete binding (data not shown).[21] In contrast, 4a, b, 12d,
abs
max
e and 14b did not show any change in the l upon addition
of aptamer, thus indicating that the specific modification in
these derivatives interferes with aptamer binding. The interac-
tion of the MG derivatives with MGA was further tested by
one-photon fluorescence titration experiments in which in-
creasing quantities of the MGA were added to a fixed 0.32 mm
solution of the respective MG derivative, with fluorescence
measured after each addition (Figure 2).
The dissociation constants were extracted from these titra-
tion curves by nonlinear curve fitting for a 1:1 binding, and the
obtained values are shown in Table 2 (full data are available in
the Supporting Information).
Under our conditions, the dissociation constant (Kd) of the
reference MG-Cl was 170 nm, which is very similar to that pre-
viously published (117 nm).[22] We found that several derivatives
retained an ability to bind MGA, and two of them showed in-
creased affinity (Kd 150 nm (8b) and 110 nm (12b)). Four other
MG derivatives also exhibited Kd values in the nanomolar
range (8d, 8c, 14a and 12a: 300, 320 540 and 550 nm respec-
tively). Compounds 8a and 12c showed dissociation constants
in the micromolar range (1.43 and 1.82 mm, respectively). For
14b, which was already slightly fluorescent in solution, the ad-
dition of MGA resulted in only a weak increase in fluorescence.
The remaining four derivatives (4a, 4b, 12d and 12e) lost
binding to the aptamer. The two N-modified compounds
belong to this category, thus indicating that the modification
at the two nitrogen atoms (located in the centre of the recog-
nition pocket of the aptamer) impair the binding because of
steric hindrance. Compounds 12d and 12e also lost their
aptamer-binding abilities, thus indicating that the aptamer is
unable to adapt its conformation to a thrice-modified phenyl
ring. Interestingly, two substitutions, at meta and para posi-
Figure 2. Apparent quantum yields for MG derivatives as function of MGA
concentration.
tions of the phenyl ring (12c), preserved a moderate binding
capability. Thus, one free meta position appears to be necessa-
ry for aptamer binding. The dissociation constants of the
monosubstituted MG derivatives indicate that substitutions by
a small group (I, Me, OMe) are tolerated, and aptamer binding
prefers substitutions at the meta- over the para-position. More-
over, among the selected groups, adding electron-donor sub-
ChemBioChem 0000, 00, 1 – 9
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
5
&
ÞÞ
These are not the final page numbers!

J. F. Nicoud et al.
(Table 2). This increase in TP brightness was stronger for all MG
derivatives modified at the phenyl ring. The best values were
obtained with 8c and 8d (21- and 19.5-fold increase in TP
brightness, respectively). For the MG derivatives with the high-
est affinity for MGA (8b and 12b), the increase in TP bright-
ness was 3–3.5-fold. Thus, these derivatives are superior to
MG-Cl in terms of both aptamer binding and TP brightness.
The relative TP brightness can be correlated with the electronic
effects of the specific substituents. Relative to MG-Cl, the intro-
duction of an electroactive group (donor or acceptor) im-
proved TP brightness. The increase in TP brightness was more
pronounced with the more efficient acceptor groups (in our
case iodide, with Hammett constants of +0.28 and +0.34 for
the para (8c) and meta (8d) substitutions, respectively). The
meta-substituted iodo compound 8d appeared to be more ef-
ficient than the para compound 8c, likely attributable to the
distance dependence of the inductive effect. The meta-OMe
substitution resulted in a TP brightness similar to that of MG,
mainly because of its mesomeric effect (low for the nonconju-
gated meta position; Hammett constant +0.1 for 8b). The
para-OMe substitution provided a strong electron-donating
effect in a conjugated position (Hammett constant À0.28 for
8a), and thus a doubled TP brightness compared to MG. The
Me substitution induces an interesting increase of the TP
brightness, which is more pronounced in the meta than in the
para position, as would be expected for a dominant inductive
effect (Hammett constants of À0.14 and À0.28 for para and
meta substitution 12a and 12b, respectively).
Compared with the one-photon properties of RNA mimics of
GFP described by Paige et al.,[20] our MG derivatives exhibit
reduced one-photon brightnesses compared to that of their
RNA–fluorophore complex (“Spinach”). However, in contrast to
Spinach, our dyes were optimised for TP illumination. For
example, the TP brightness of 8d–MGA was about ten times
higher than that of GFP (see the Supporting Information).
As MG-Cl acts as sensitizer in one-photon chromophore-as-
sisted light inactivation (CALI),[26] that is, the degradation of the
MG-bound molecule,[21] it was important to demonstrate that
TPE imaging conditions do not trigger CALI and degradation
of the bound aptamer. Therefore, a solution of MGA:MG-Cl
was irradiated under a TP microscope (Supporting Informa-
tion). The RNA was then analysed by acrylamide gel electro-
phoresis. No degradation was detected, thus indicating that no
CALI occurred under these illumination conditions (Figure S15
in the Supporting Information). The same results were ob-
tained when using one photon microscopy with a 633 nm
laser and a Zeiss LSM510 confocal microscope (data not
shown).
Table 2. Dissociation constants of MGs with MG aptamer obtained by
one-photon fluorescence titration; two-photon diffusion times and two-
photon brightness of MG derivatives in the presence of the RNA aptamer
and in the presence of a negative control RNA.
MGs
Dissociation
constant [mm]
Diffusion time [ms]
neg./pos.
TP brightness
neg./pos.[d]
[b]
4a
4b
8a
8b
8c
n.b.[a]
n.b.
À/À
À/À
À/À
À/À
<0.01/0.04
<0.01/0.08
0.02/0.05
<0.01/0.02
<0.01/0.02
<0.01/0.08
À/0.13
1.43
0.15
0.32
0.30
0.55
0.11
1.82
n.b.
0.3/1.2
0.2/0.7
0.1/2.1
0.2/3.9
0.3/1.7
0.5/1.6
À/0.3
À/0.4
À/À
8d
12a
12b
12c
12d
12e
14a
14b
MG-Cl
À/<0.01
À/À
n.b.
0.54
l.b.[c]
0.17
<0.01/0.023
–
0.6/2.7
–
0.3/0.6
0.01/0.06
[a] n.b.: no binding. [b] Very low fluorescence signal inducing low quality
data for FCS measurements. [c] l.b.: low binding. [d] TP brightness of rho-
damine B measured under the same conditions is 7.9.
stituents (OMe, Me) in the para position resulted in increased
MG affinity. Interestingly, introduction of a rigid conjugated
linker in a para position (double bond) on the free phenyl ring
of MG allowed significant binding with the aptamer, even
when this linker was associated with a bulky 4-vinylanisyl
group (14a, 540 nm). In contrast, introducing the same substi-
tution in meta position interfered with induced fluorescence in
the presence of MGA (14b), as observed for the previously
described small groups.
Two-photon photophysical properties
To test the TP sensitivity of the new MG derivatives we per-
formed two-photon excited fluorescence correlation spectros-
copy (TP-FCS) measurements in solutions that contained either
MGA (positive control) or a modified MGA that did not bind to
MG (negative control; for sequences see the Supporting Infor-
mation).[21] This technique provided measurement of the rela-
tive TP brightnesses and diffusion times of the molecules
(Table 2). In FCS the diffusion time is related to the size of the
fluorescent moiety: a short diffusion time indicates a relatively
small size of molecule (typically free MGs), whereas a longer
diffusion time indicates the formation of a larger complex (i.e.,
MGA:MG). The low fluorescence of MG and its derivatives in
the absence of aptamer was usually sufficient for extracting
values for TP brightness and diffusion time. In the presence of
modified MGA we obtained values close to or below 0.01 ms
for all MGs, thus indicating a small diffusing molecule with low
TP brightness, as expected. This clearly indicated that no bind-
ing occurred between the modified RNA aptamer and the
dyes. As expected, based on the one-photon experiments, we
observed an increase in TP brightness in the presence of the
functional MGA. The reference MG-Cl showed a twofold in-
crease in TP brightness upon recognition of the aptamer
Cell toxicity
Cytotoxic and genotoxic effects have been attributed to MG in
fungi and bacteria,[38–40] whereas MG tolerance has been associ-
ated with detoxification systems linked to respiratory processes
and maintenance of the redox balance that protect cells from
oxidative damage.[41,42] During this work, cell toxicity of MG
and the MG derivatives was tested by measuring their effects
&6
&
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 9
ÝÝ
These are not the final page numbers!

Lighting RNA
on the propagation of cultured tobacco Bright Yellow 2 (BY-2)
cells in suspension.[43] As was reported previously for yeast,
worms and flies,[21] MG freely diffused into BY-2 cells (data not
shown), and the cells could tolerate MG in the culture
medium; that is, they grew normally in the presence of 1.0 mm
MG-Cl (Figure S16). The MG derivatives synthesised during this
work showed a broad spectrum of toxicity. Specific derivatives
(14a, 8d, 8c and 12a) affected growth (even at 1.0 mm), which
reflects a toxicity similar to that of MG-oxalate. Two derivatives
(12a and 8b) showed toxicity levels similar to that of MG-Cl.
Three derivatives showed slightly reduced toxicity (12b, 4a
and 12c) and the last group (derivatives 4b, 12d, 12e and
14b), did not show any toxic effect on BY-2 cells at the assayed
concentrations. These derivatives did not bind to MGA or
show high Kd values (Figure 3 and Table 2), thus indicating that
preserved upon introduction of modifications at the phenyl
ring. Substitutions at the meta position of the phenyl ring (in-
troducing small electron-donor groups) improved the affinity
for MGA. The introduction of electroactive groups at the meta
or para position of the phenyl ring significantly improved the
two-photon brightness of such molecules upon MGA recogni-
tion. Properties affecting MGA binding are involved in MG
toxicity in plant cells. meta- and para-iodo derivatives (8c, d)
showed the best signal-to-noise ratio, with TP brightnesses up
to ten times higher than that of GFP for 8d. As these deriva-
tives also exhibited strong MGA binding, they might be recom-
mended for in vivo visualisation of MGA-tagged RNA mole-
cules by TPE microscopy. Moreover, the MG aptamer is signifi-
cantly shorter (38 nt) than the Spinach aptamer (80 nt), which
might be a critical factor in in vivo RNA labelling.
Experimental Section
Synthesis: All chemicals and reagents were purchased from
Sigma–Aldrich or Acros Organics and were used as received unless
specified. THF was distilled over sodium and benzophenone under
argon atmosphere; dichloromethane was distilled over calcium
hydride under argon atmosphere; triethylamine was distilled over
potassium hydroxide under argon atmosphere; DMSO was distilled
over calcium hydride under vacuum and conditioned under argon
prior to use. 1H and 13C NMR spectra were recorded with a
400 MHz Advance 400 instrument (Bruker) in CDCl3 (internal stan-
1
dard 7.24 ppm for H, and 77 ppm, middle of the three peaks, for
13C spectra) or [D6]DMSO (internal standard 2.26 ppm, 39.5 ppm for
13C spectra). Fast atomic bombardment (FAB) mass spectra were
recorded with a ZA-HF instrument with 4-nitrobenzyl alcohol as a
matrix, and ESI spectra were obtained on a Bruker HTC ultra (ESI-
IT). TLC analyses were run on precoated aluminium plates (Si 60
F254; Merck). Column chromatography was run on Silica Gel (60–
120 mesh; Merck). MGAs were purchased from IBA (Gçttingen, Ger-
many); for sequences see the Supporting Information.
Figure 3. Relative BY-2 cell density six days after addition of MG-Cl or MG
derivatives at the given concentrations.
One-photon photophysics: UV–visible spectra were recorded on
a Carry 4000 spectrophotometer (Agilent Technologies, Santa Clara,
CA) with paired 1 cm optical path quartz cuvettes and HEPES
buffer in the reference cuvette. Fluorescence spectra were record-
ed on a Fluorolog spectrofluorimeter (Jobin–Yvon/HORIBA Scientif-
ic, Edison, NJ), in 1ꢂ0.4 cm fluorescence quartz cuvettes. Quantum
yields were calculated with cresyl violet as the reference. Fluores-
cence titration curves were obtained by addition of increasing con-
centrations of MGA and computed as described by Babendure
et al.[8] with Origin software (see the Supporting Information).
properties affecting MGA association are involved in MG toxici-
ty in plant cells. Interestingly, 14a and 14b differ only with re-
spect to the position of the 4-vinylanisole moiety at the phenyl
ring, but they are derivatives with very different toxicities.
Thus, the presence of this group at the para position (14a) al-
lowed MG to bind MGA but also resulted in its strong toxicity.
In contrast, this group in the meta position (14b) interfered
with MGA binding and abolished the toxic effect of the mole-
cule. However, changes in MGA binding affinity do not neces-
sarily correlate with changes in toxicity. Thus, although the
mono-methylated derivatives 12a and 12b differed in MGA
binding affinity they showed similar levels of toxicity.
Two-photon fluorescence correlation spectroscopy: TP FCS was
performed on a home-built setup. TPE was provided by a Tsunami
Ti:sapphire laser pumped with a Millennia V solid-state laser (Spec-
tra-Physics, Mountain View, CA) with 100 fs pulses (80 MHz,
760 nm). Following passage through a beam expander, the infrared
light was focused into the sample by a water-immersion Olympus
objective (60ꢂ, NA=1.2) mounted on an Olympus IX70 inverted
microscope. The back aperture of the objective was slightly over-
filled to create a diffraction-limited focal spot. Samples were
placed in eight wells of a Lab-Tek chambered cover glass (Nalge
Nunc International, Rochester, NY) and positioned in the X and Y
axes by a motorised stage (Mꢃrzhꢃuser Wetzlar, Germany). The
fluorescence from the samples was collected through the same ob-
jective and directed by a COWL750 dichroic mirror (Coherent,
Conclusion
A library of MG derivatives for TP detection of MGA-tagged
RNAs was designed, synthesised and characterised. In particu-
lar, their TP photophysical properties have been investigated
for the first time. Modifications performed at the nitrogen
atoms induced a loss of MGA binding, whereas binding was
ChemBioChem 0000, 00, 1 – 9
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
7
&
ÞÞ
These are not the final page numbers!

J. F. Nicoud et al.
[16] E. Bertrand, P. Chartrand, M. Schaefer, S. M. Shenoy, R. H. Singer, R. M.
Long, Mol. Cell 1998, 2, 437–445.
[17] J. Chen, O. Nikolaitchik, J. Singh, A. Wright, C. E. Bencsics, J. M. Coffin, N.
Ni, S. Lockett, V. K. Pathak, W.-S. Hu, Proc. Natl. Acad. Sci. USA 2009, 106,
13535–13540.
[18] N. Daigle, J. Ellenberg, Nat. Methods 2007, 4, 633–636.
[19] T. Ozawa, Y. Natori, M. Sato, Y. Umezawa, Nat. Methods 2007, 4, 413–
419.
[20] J. S. Paige, K. Y. Wu, S. R. Jaffrey, Science 2011, 333, 642–646.
[21] D. Grate, C. Wilson, Proc. Natl. Acad. Sci. USA 1999, 96, 6131–6136.
[22] J. R. Babendure, S. R. Adams, R. Y. Tsien, J. Am. Chem. Soc. 2003, 125,
14716–14717.
Orsay, France) toward a 50 mm diameter optical fibre coupled to
a SPCM 200 FC avalanche photodiode (EG&G, Vaudreuil, Canada).
Residual infrared light was rejected by a BG39 filter (Coherent). For
FCS measurements, the normalised autocorrelation function (ACF),
G(t), of the fluorescence intensity fluctuations was calculated
online by an ALV5000E digital correlator card (ALV, Langen, Germa-
ny). Calibration of the system was performed with 50 nm tetrame-
thylrhodamine (TMR). By assuming a diffusion constant of 2.8ꢂ
10À10 m2 sÀ1 [44]
volume were 0.29 and 1.3 mm, respectively (effective volume
0.2 fL). By assuming three-dimensional Gaussian-distributed
,
the equatorial (r0) and axial (z0) radii of the focal
a
[23] D. F. Duxbury, Chem. Rev. 1993, 93, 381–433.
[24] R. Y. Tsien, FEBS Lett. 2005, 579, 927–932.
excitation intensity, the ACF curves were fitted as described by
Clamme et al.[44]
[25] M. Pawlicki, H. A. Collins, R. G. Denning, H. L. Anderson, Angew. Chem.
2009, 121, 3292–3316; Angew. Chem. Int. Ed. 2009, 48, 3244–3266.
[26] T. Surrey, M. B. Elowitz, P. E. Wolf, F. Yang, F. Nꢀdꢀlec, K. Shokat, S. Lei-
bler, Proc. Natl. Acad. Sci. USA 1998, 95, 4293–4298.
[27] S. J. Remington, Curr. Opin. Struct. Biol. 2006, 16, 714–721.
[28] C. Baugh, D. Grate, C. Wilson, J. Mol. Biol. 2000, 301, 117–128.
[29] D. H. Nguyen, T. Dieckmann, M. E. Colvin, W. H. Fink, J. Phys. Chem. B
2004, 108, 1279–1286.
Cell toxicity: BY-2 cells were grown under standard conditions[26]
in 24-well plates. Increasing concentrations of MG or MG deriva-
tives (0, 0.125, 0.25, 0.5, 1.0, 2.0, 4.0 mm) were added, and after six
days of incubation, pictures were taken, then pixel intensities were
measured by using ImageJ and used as a measure of cell accumu-
lation.
[30] D. H. Nguyen, S. C. DeFina, W. H. Fink, T. Dieckmann, J. Am. Chem. Soc.
2002, 124, 15081–15084.
[31] L. T. Nguyen, P. H. S. Kwakman, D. I. Chan, Z. Liu, L. de Boer, S. A. J. Zaat,
H. J. Vogel, Antimicrob. Agents Chemother. 2011, 55, 2074–2083.
[32] T. Dieckmann, E. Suzuki, G. K. Nakamura, J. Feigon, RNA 1996, 2, 628–
640.
[33] J. Flinders, S. C. DeFina, D. M. Brackett, C. Baugh, C. Wilson, T. Dieck-
mann, ChemBioChem 2004, 5, 62–72.
[34] D. J. Patel, A. K. Suri, F. Jiang, L. Jiang, P. Fan, R. A. Kumar, S. Nonin, J.
Mol. Biol. 1997, 272, 645–664.
Acknowledgements
The authors thank Dr. Ludovic Riechert for TP-FCS experiments
and Dr. Julien Godet for assistance in Kd determination. This
work was performed with support from the ANR (grant ANR-08-
BLAN-0244–02).
[35] D. M. Brackett, T. Dieckmann, ChemBioChem 2006, 7, 839–843.
[36] M. Albota, D. Beljonne, J.-L. Brꢀdas, J. E. Ehrlich, J.-Y. Fu, A. A. Heikal,
S. E. Hess, T. Kogej, M. D. Levin, S. R. Marder, D. McCord-Maughon, J. W.
Perry, H. Rçckel, M. Rumi, G. Subramaniam, W. W. Webb, X.-L. Wu, C. Xu,
Science 1998, 281, 1653–1656.
[37] G. A. Kraus, I. Jeon, M. Nilsen-Hamilton, A. M. Awad, J. Banerjee, B.
Parvin, Molecules 2008, 13, 986–994.
[38] S. J. Culp, P. W. Mellick, R. W. Trotter, K. J. Greenlees, R. L. Kodell, F. A.
Beland, Food Chem. Toxicol. 2006, 44, 1204–1212.
[39] R. A. Mittelstaedt, N. Mei, P. J. Webb, J. G. Shaddock, V. N. Dobrovolsky,
L. J. McGarrity, S. M. Morris, T. Chen, F. A. Beland, K. J. Greenlees, R. H.
Heflich, Mutat. Res. 2004, 561, 127–138.
[40] T. Zahn, T. Braunbeck, Toxicol. In Vitro 1995, 9, 729–741.
[41] M. V. Bianco, F. C. Blanco, B. Imperiale, M. A. Forrellad, R. V. Rocha, L. I.
Klepp, A. A. Cataldi, N. Morcillo, F. Bigi, BMC Infect. Dis. 2011, 11, 195–
204.
[42] Y.-H. Kim, C. Lee, H. Go, K. Konishi, K. Lee, P. C. K. Lau, M.-H. Yu, Arch.
Biochem. Biophys. 2010, 494, 159–165.
[43] T. Nagata, Y. Nemoto, S. Hasezawa, Intl. Rev. Cytol. 1992, 132, 1–30.
[44] J. P. Clamme, J. Azoulay, Y. Mꢀly, Biophys. J. 2003, 84, 1960–1968.
Keywords: aptamers · fluorescent probes · malachite green ·
RNA visualization · two-photon absorption
[1] O. Shimomura, F. H. Johnson, Y. Saiga, J. Cell. Comp. Physiol. 1962, 59,
223–239.
[2] M. Schmid, T. H. Jensen, Chromosoma 2008, 117, 419–429.
[3] P. Vinciguerra, F. Stutz, Curr. Opin. Cell Biol. 2004, 16, 285–292.
[4] S. L. Erickson, J. Lykke-Andersen, J. Cell Sci. 2011, 124, 293–297.
[5] K. Nagai, C. Oubridge, A. Kuglstatter, E. Menichelli, C. Isel, L. Jovine,
EMBO J. 2003, 22, 3479–3485.
[6] D. St. Johnston, Nat. Rev. Mol. Cell Biol. 2005, 6, 363–375.
[7] B. Pyhtila, T. Zheng, P. J. Lager, J. D. Keene, M. C. Reedy, C. V. Nicchitta,
RNA 2008, 14, 445–453.
[8] F. Besse, A. Ephrussi, Nat. Rev. Mol. Cell Biol. 2008, 9, 971–980.
[9] A. Eulalio, I. Behm-Ansmant, E. Izaurralde, Nat. Rev. Mol. Cell Biol. 2007,
8, 9–22.
[10] J. Xu, N.-H. Chua, Curr. Opin. Plant Biol. 2011, 14, 88–93.
[11] M. Ghildiyal, P. D. Zamore, Nat. Rev. Genet. 2009, 10, 94–108.
[12] D. H. Chitwood, M. C. Timmermans, Nature 2010, 467, 415–419.
[13] M. Heinlein, B. L. Epel, Int. Rev. Cytol. 2004, 235, 93–164.
[14] N. Christensen, J. Tilsner, K. Bell, P. Hammann, R. Parton, C. Lacomme, K.
Oparka, Traffic 2009, 10, 536–551.
Received: November 26, 2011
[15] S. Tyagi, F. R. Kramer, Nat. Biotechnol. 1996, 14, 303–308.
Published online on && &&, 0000
&8
&
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 9
ÝÝ
These are not the final page numbers!

FULL PAPERS
A library of malachite green deriva-
tives was designed in order to improve
two-photon detection of a malachite
green RNA aptamer. Some of the pre-
pared derivatives show remarkable af-
finity for the aptamer and improved
two-photon brightness as measured by
two-photon fluorescence correlation
spectroscopy. These new derivatives
show lower toxicity than the classical
malachite green oxalate.
J. Lux, E. J. PeÇa, F. Bolze, M. Heinlein,
J.-F. Nicoud*
&& – &&
Malachite Green Derivatives for Two-
Photon RNA Detection
ChemBioChem 0000, 00, 1 – 9
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
&
9
&
ÞÞ
These are not the final page numbers!