Rational design of a new fluorescent 'ON/OFF' xanthene dye for phosphate detection in live cells

Article abstract of DOI:10.1039/c4ob00951g

A new fluorescein derivative with ON/OFF features, 9-[1-(4-tert-butyl-2- methoxyphenyl)]-6-hydroxy-3H-xanthen-3-one (Granada Green, GG), was designed and synthesised. The new dye has spectral characteristics similar to those of other xanthenic derivatives

Full text of DOI:10.1039/c4ob00951g

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Rational design of a new fluorescent ‘ON/OFF’
xanthene dye for phosphate detection in live cells†
Cite this: Org. Biomol. Chem., 2014,
12, 6432
A. Martínez-Peragón,a D. Miguel,a A. Orte,b A. J. Mota,c M. J. Ruedas-Rama,b
J. Justicia,a J. M. Alvarez-Pez,b J. M. Cuerva*a and L. Crovetto*b
A new fluorescein derivative with ON/OFF features, 9-[1-(4-tert-butyl-2-methoxyphenyl)]-6-hydroxy-
3H-xanthen-3-one (Granada Green, GG), was designed and synthesised. The new dye has spectral
characteristics similar to those of other xanthenic derivatives but shows a higher pKa value for the equili-
brium between its neutral and anionic forms. In addition, GG undergoes the same phosphate-mediated
excited state proton transfer (ESPT) reaction as other xanthenic derivatives, giving rise to fluorescence
decay traces that are dependent on both the phosphate concentration and pH of the medium. The phos-
phate-mediated ESPT reaction was employed to detect changes in the phosphate concentrations in live,
permeabilised MC3T3-E1 preosteoblasts at pH 7.35. Its high pKa value indicates that this new dye is more
sensitive as an intracellular phosphate sensor than other previously tested dyes, as experimentally demon-
strated by its ability to detect a wider range of phosphate concentrations in biomimetic media and by the
increased ratio of the phosphate concentration/decay time.
Received 8th May 2014,
Accepted 27th June 2014
DOI: 10.1039/c4ob00951g
www.rsc.org/obc
them had successfully estimated the real-time concentration of
phosphate inside live cells until the recently proposed xanthe-
Introduction
Fluorescence lifetime imaging microscopy (FLIM) is an attrac- nic-based method.11
tive fluorescence technique for quantitative real-time sensing
Fluorescein is a complex dye that, in aqueous solution at
of biologically important targets inside living cells.1–3 In physiological pH, can exist in mono- and dianion prototropic
addition to the high sensitivity and non-invasive character of forms. In two pioneering papers, we showed that at near-
fluorescence methods, FLIM is an excellent alternative to fluo- neutral pH and in the presence of 1 M phosphate buffer,
rescence intensity and fluorescence ratiometric measurements which acts as a suitable proton donor–acceptor, the excited-
because it is concentration independent and only a single exci- state reaction that interconverts the mono- and dianion forms
tation wavelength/emission interval is required.4 The detection occurs very efficiently, and the fluorescence decays of these
and quantification of phosphate inside live cells and in extra- excited mono- and dianions become coupled, i.e. the presence
cellular media are relevant in the study of bone mineralisation, of these excited-state proton-transfer (ESPT) reactions influ-
signal transduction, and energy storage in biological ences the decay traces from excited fluorescein and causes the
systems.5–7 Nevertheless, the development of fluorescent decay times to vary depending on both the pH and phosphate
sensors for measuring phosphate under aqueous physiological concentration. In contrast, at low phosphate concentrations
conditions is challenging due to strong hydration effects.8 (5 mM phosphate buffer or less), the excited monoanion and
Beyond invasive approaches employing radioactive phosphate, dianion are not coupled by the ESPT reaction and thus decay
which are associated with a number of drawbacks,9 there are independently of each other.12,13 The phosphate-mediated
few chemical sensors that are capable of detecting phosphate ESPT reaction is not exclusive to fluorescein, and the presence
anions in an aqueous physiological system,10 and none of of phosphate also induces the ESPT reaction at near-neutral
pH in other xanthenic derivatives, such as 2′,7′-bis-(2-carboxy-
ethyl)-5-(and-6)-carboxyfluorescein (BCECF).14 However, the
biexponential character of the fluorescence decays limits the
use of those dyes as possible lifetime-based sensors because
the close values of the two fluorescence decays prevent the use
of the fluorescence lifetime imaging microscopy (FLIM) tech-
nique.15 Moreover, although one of the coupled fluorescence
decay times depends on the phosphate concentration in the
medium, the other one is practically insensitive to phosphate.
aDepartment of Organic Chemistry, Faculty of Sciences, University of Granada,
C. U. Fuentenueva s/n, 18071 Granada, Spain. E-mail: jmcuerva@ugr.es
bDepartment of Physical Chemistry, Faculty of Pharmacy, University of Granada,
Cartuja Campus, 18071 Granada, Spain. E-mail: luiscrovetto@ugr.es
cDepartment of Inorganic Chemistry, University of Granada, C. U. Fuentenueva s/n,
18071 Granada, Spain
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4ob00951g
6432 | Org. Biomol. Chem., 2014, 12, 6432–6439
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
A few years ago, with the aim of developing new xanthenic shown that the dependence of the recovered long decay times
dyes with improved spectral properties for specific purposes, it from aqueous solutions of 2-Me-4-OMe TG on the phosphate
was proposed that the fluorescein structure could be divided concentration increases when the pH of the medium is near
into two parts: the benzoic acid moiety, as a photoinduced the pKa of the dye. Because the pKa values of 2-Me-4-OMe TG
electron transfer (PeT) donor, and the xanthene ring as the and 2-OMe-5-Me TG and related derivatives are far from the
fluorophore, in which PeT might determine the fluorescence physiological pH,21,22 it is desirable to obtain fluorescent dyes
quantum yield (ϕF). Thus, if the highest occupied molecular with pKa values that are higher and closer to the physiological
orbital (HOMO) energy level of the benzene moiety is higher pH than that of 2-Me-4-OMe TG.
than a certain threshold for electron transfer to the excited
In an attempt to increase the pKa, one possible solution is
xanthene fluorophore (a-PeT), the ϕF is small.16,17 In the oppo- to incorporate other aliphatic electron donating groups
site direction, electron transfer from the excited xanthene to without modifying the optical properties of the dye. In this
the benzene moiety (d-PeT) can occur if the lowest unoccupied context, some alkyl groups have been added to the xanthene
molecular orbital (LUMO) energy level of the benzene moiety core, thus causing an increase in the pKa values (fluorescein as
is low enough, also resulting in small ϕF.18 In contrast, fluore- a reference, pKa 6.43; BCECF, pKa 6.98; DHCF, pKa 6.60; DEF,
scein derivatives with high ϕF should have benzene moieties pKa 6.61).24–27 However, the synthetic procedures are not suit-
with both low HOMO and high LUMO energy levels. Thus, the able for a great variety of substituents because of the harsh
energy levels of the pendant aryl ring are directly related to the reaction conditions and the restriction thus imposed on the
ϕF. It has also been shown that the role of the carboxylic group starting materials. Recently, 9-alkyl xanthenones with different
is only to keep the benzene ring and the xanthene moiety aliphatic pendant groups have been easily prepared, and their
orthogonal to each other, and it can be replaced by any other photophysical behaviour has also been explored.28 Remark-
substituents. Although there is some theoretical controversy ably, some of them retained similar fluorescence properties to
about the direction of the PeT mechanism,19 based on these those of fluorescein, including the characteristic phosphate-
ideas, new 9-aryl substituted fluoresceins were developed mediated ESPT reaction and higher pKa values (6.20–6.67)
without the carboxylic acid in the pendant aromatic ring at compared with TGs (5.97–6.04).21,22 Nevertheless, this
C-9, which simplifies the number of prototropic species approach to increase the pKa without modifying the spectro-
in aqueous solution. Thus, the so-called Tokyo Green dyes, scopic properties of the dye is limited by the electron donating
9-[1-(2-methyl-4-methoxyphenyl)]-6-hydroxy-3H-xanthen-3-one nature of alkyl substituents. The best result has been achieved
(2-Me-4-OMe TG) and 9-[1-(2-methoxy-5-methylphenyl)]-6- using an isopropyl group as a substituent (pKa 6.67). Unfortu-
hydroxy-3H-xanthen-3-one (2-OMe-5-Me TG), do not have a nately, 9-alkyl xanthenones with better electron donating
carboxylic acid in the benzene ring, although this is kept groups, such as the tert-butyl group, could not be prepared.
orthogonal to the xanthenone core. Importantly, they have the Bearing all these precedents in mind, we thought that a tailor-
advantage of being highly fluorescent at basic pH values made TG-type fluorescein might exhibit lower acidity, with a
(anion form of the xanthene moiety), but their quantum yield pKa closer to the physiological pH. To this end, the substi-
is near zero at acidic pH values (protonated form of the tution of the phenyl at C-9 is critical, and two synergic effects
xanthene moiety).20
are required to destabilise the fluorescent anionic prototropic
In our aim to find fluorescein derivatives capable of under- species: (a) an electronegative oxygen atom must be placed
going the characteristic phosphate-mediated ESPT reaction near the xanthenone core to destabilise the anionic structure,
with a single lifetime in the near-neutral pH region, we study and (b) a good alkyl electron donating group must be used to
the photophysics of 2-Me-4-OMe TG and 2-OMe-5-Me TG in increase the electron density in this oxygen without altering
the presence of phosphate buffer in the pH range between 5 the desired molecular orbital ordering. After a theoretical
and 10.21,22 The results showed that both undergo the charac- screening, we prepared 9-[1-(4-tert-butyl-2-methoxyphenyl)]-
teristic ESPT reaction, in which the coupled fluorescence decay 6-hydroxy-3H-xanthen-3-one (Granada Green, GG), which is
exhibits a phosphate-sensitive component on the order of predicted to have a higher pKa value than previously described
nanoseconds and a second component on the order of sub- TGs.
nanoseconds, whose value becomes negligible at pH and phos-
phate concentrations greater than 6.0 and 0.02 M, respectively.
In addition, the larger decay time is highly sensitive to the
phosphate concentration, while the presence of other ions not
Experimental
Materials and solutions
involved in the proton-transfer reaction has a negligible effect
on the fluorescence decay time.23 Therefore, the changes in For the photophysical studies, stock solutions of sodium phos-
the fluorescence decay time were considered a possible direct phate were prepared using NaH2PO4 × H2O and Na2HPO4
×
means of investigating the environmental phosphate concen- 7H2O (both Fluka, puriss. p.a.) in appropriate amounts to
tration in a small volume at near neutral pH. Thus, in a recent obtain the required pH. For the FLIM experiments, Dulbecco’s
paper, we used the ability of 2-Me-4-OMe TG to undergo phos- modified Eagle’s medium (DMEM, D-6546), α-minimum
phate-mediated ESPT to investigate the environmental phos- essential medium (α-MEM), foetal bovine serum (FBS),
phate concentration at physiological pH.11 In this paper, it is penicillin/streptomycin, trypsin-EDTA, and α-hemolysin from
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 6432–6439 | 6433

View Article Online
Paper
Organic & Biomolecular Chemistry
Staphylococcus aureus were obtained from Sigma Chemical Cell permeabilisation
(St. Louis, MO). To obtain the required phosphate concen-
MC3T3-E1 cells were deposited onto coverslips in 12-well
plates and incubated with the described medium. On the day
of the experiment, the cells were washed twice with phosphate-
buffered saline (PBS) and were perforated by incubation for
15 min at 37 °C with 2 µg ml−1 α-toxin in permeabilisation
buffer.30 Afterwards, the cells were washed twice with PBS and
analysed by FLIM.
tration and pH in the measurement solutions, the necessary
amount of phosphate stock solution was added to DMEM. The
permeabilisation buffer contained 20 mM potassium MOPS
pH 7.0, 250 mM mannitol, 1 mM potassium ATP, 3 mM
MgCl2, and 5 mM potassium glutathione. Commercially avail-
able phosphate-buffered saline (PBS) (Sigma-Aldrich) was also
used to wash the cells. All solutions were prepared using Milli-
Q water as the solvent. All chemicals were used as received
without further purification. The solutions were kept cool in
the dark when not in use to avoid possible deterioration by
light and heat. Aliquots of α-toxin from Staphylococcus aureus
were prepared and stored in a freezer (−20 °C) until use.
Ensemble time-resolved fluorescence instrumentation
Fluorescence decay traces at the ensemble level were recorded
using the time-correlated single photon timing (TCSPT)31,32
method with a FluoTime 200 fluorometer (PicoQuant, Inc.,
Germany). The excitation source consisted of an LDH-485
pulsed laser with a minimum pulse width of 88 ps. The pulse
repetition rate was 20 MHz. The laser pulse was directed to the
solution sample in 10 × 10 mm cuvettes. The lens-focused
fluorescence emission passed through a detection polariser set
at the magic angle and a detection monochromator prior to
reaching the photomultiplier detector. The fluorescence decay
histograms were collected in 1320 channels with a time incre-
ment of 36 ps per channel. Histograms of the instrument
response functions (using a LUDOX scatterer) and sample
decays were recorded until they reached approximately 2 × 104
counts in the peak channel.
Instrumentation
Absorption spectra were recorded on a Perkin-Elmer Lambda
650 UV/vis spectrophotometer with a temperature-controlled
cell.
Steady-state fluorescence emission spectra were collected
on a JASCO FP-6500 spectrofluorometer equipped with a
450 W xenon lamp for excitation and with an ETC-273T temp-
erature controller at 20 °C. The pH of the solutions was
measured just before and after the fluorescence measurements
at the same temperature.
Fluorescence decay traces of the compounds in the absence
and in the presence of phosphate buffer were recorded by the
single-photon timing method using a FluoTime 200 fluoro-
meter (PicoQuant, Inc.). The excitation was performed using
an LDH-485 (PicoQuant, Inc.), and the observation was made
through a monochromator at different wavelengths. The pulse
FLIM instrumentation
repetition rate was 20 MHz. Fluorescence decay histograms Fluorescence lifetime images were recorded with a MicroTime
were collected in 1320 channels using 10 × 10 mm cuvettes. 200 fluorescence lifetime microscope system (PicoQuant
The time increment per channel was 36 ps. Histograms of the GmbH) using the time-tagged time-resolved (TTTR) method-
instrument response functions (using a LUDOX scatterer) and ology, which allows fluorescence decay histograms to be recon-
sample decays were recorded until they typically reached structed from molecules in the confocal volume for each pixel
2 × 104 counts in the peak channel. The total width at half of the image. The excitation source consisted of the same LDH
maximum of the instrument response function was ∼60 ps. 485 nm pulsed laser described above, controlled using a
Fluorescence decays were recorded at three λem (505, 515 and “Sepia” driver (PicoQuant) and working at a repetition rate of
525 nm) for all samples. The fluorescence decay traces 20 MHz. The excitation light beam crossed a quarter-wave
were individually analysed using an iterative deconvolution plate and was directed into the specimen after being reflected
method with exponential models using the FluoFit software in the excitation dichroic mirror (490dcxr, AHF/Chroma) to the
(PicoQuant).
oil immersion objective (1.4 NA, 100×) of an inverted confocal
microscope (IX-71, Olympus). The collected fluorescence light
was filtered by a long-pass filter HP500LP (AHF/Chroma) and
Cell culture
MC3T3-E1 preosteoblast (ECACC 99072810) cell lines were pro- focused onto a 75 μm pinhole. After the aperture, the fluore-
vided by the Cell Culture Facility, University of Granada. scence light crossed through an FF01-520/35 bandpass filter
MC3T3-E1 cells were grown in alpha minimum essential (Semrock) and was refocused onto an SPCM-AQR 14 single
medium (αMEM) containing 10% foetal bovine serum and 1% photon avalanche diode (Perkin Elmer). The data acquisition
penicillin–streptomycin in a humidified 5% CO2 incubator, as was performed with a TimeHarp 200 TCSPT module (Pico-
described previously.29 For the FLIM microscopy experiments, Quant GmbH) in TTTR mode, which enabled the reconstruc-
the cells were seeded onto 20 mm coverslips in 12-well plates tion of the FLIM image. Raw images were recorded by raster
at a density of 11 250 cells cm−2. The coverslips were translated scanning an area of 80 × 80 μm with a resolution of 512 × 512
to the MicroTime 200 fluorescence lifetime microscope system pixels. The photons of each pixel were temporally sorted with
(see below) and washed with PBS before adding the working respect to the excitation pulse in the histograms with a time
solutions.
resolution of 116 ps per channel.
6434 | Org. Biomol. Chem., 2014, 12, 6432–6439
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fluorescence decay traces and FLIM image analysis
The fluorescence decay traces at the ensemble level were indi-
vidually analysed using an iterative deconvolution method
with exponential models using the FluoFit software (Pico-
Quant). The FLIM images were analysed using the Sympho-
Time software (PicoQuant). To obtain the FLIM images, an
instrument response function (IRF) was reconstructed from a
monoexponential fluorescence decay trace of GG in solution.
This was employed to analyse the fluorescence decay histo-
gram in each pixel over the whole image by iterative deconvolu-
tion applying the maximum likelihood estimator (MLE), which
yields correct parameter sets for low count rates.33 Binning of
5 × 5 pixels and prehistogramming of 2 temporal channels (for
a final resolution of 232 ps per channel) were used to achieve a
larger number of counts in each pixel. For the analyses of GG
in biomimetic media in the presence of different phosphate
concentrations, a monoexponential decay model was set to fit
the decay traces. For the experiments in the intracellular cyto-
plasm, a biexponential model was employed with a fixed decay
time of 1.5 ns to account for the short-lived autofluorescence
of the cell, whereas the second decay time was a freely adjusta-
ble parameter. To obtain the average value of the GG lifetime
in the intracellular cytoplasm, the pixels in the region of inter-
est (cellular cytoplasm) were selected, and a total fluorescence
decay histogram of the region of interest was reconstructed
and fit to a biexponential function using a fixed short decay
time of 1.5 ns (lifetime arising from the interaction between
the dye and the intracellular components and cell autofluores-
cence) and a freely adjustable long decay time (lifetime sensi-
tive to the phosphate concentration). The value of the long
decay time is reported in Fig. 4–6.
Fig. 2 Proton-transfer energies of compounds 1 (GG), 5, 6, 7 (2-Me-
4OMe-TG).
Fig. 3 Schematic orbital diagram and isosurface plots of LUMO+1,
LUMO, HOMO and HOMO−6.
Although the theoretical estimation of pKa values is not a
simple task,35 the proton-transfer energy can be used qualitat-
ively to compare structures.36 In this case, we calculated the
proton-transfer energy of selected structures 1 (GG), 5, 6, and 7
(2-Me-4-OMe TG). Using the proton-transfer reaction of 2-Me-
4-OMe TG as a reference, we observed the highest destabilising
effect in compound 1 (Fig. 2).37 Moreover, the results for com-
pounds 5 and 6 suggested that the value of ΔG was not only
the sum of the effects of the oxygen atom at C-2′ (5) and the
alkyl groups at C-4′ (6). In fact, a direct destabilising effect of
the alkyl-donating group R on the xanthenone core could be
ruled out because the benzene and xanthenone moieties are
practically perpendicular.38 Taking into account that an
increase of one pKa unit corresponds to 5.7 kJ mol−1 in free
energy terms, a substantial increase in the pKa value of GG is
therefore expected.
Additionally, it is critical that the new substitution does not
interfere with the photophysical properties of the dye. In
this sense, Nagano proposed in 2004 that to preserve the
fluorescence of the anion, the HOMO and LUMO levels of the
xanthenone core must be placed between the HOMO
and LUMO levels of the aryl substituent at C-9.39 Theoretical
calculations showed that in compound 1, this prerequisite was
attained (Fig. 3). Bearing these precedents in mind, GG was
selected as our synthetic objective.
Results and discussion
We initially carried out theoretical calculations to find a suit-
able candidate to increase the pKa values of TGs without alter-
ing their photophysical properties. Within this context, we
focused our attention on a family of xanthenones, 1–6 and
2-Me-4-OMe TG. Several of these, xanthenones 1–5, present an
oxygenated function in the C-2′ of the aromatic ring located at
C-9. We assumed that an increase in the electronic density in
this oxygen atom would destabilise the anion formation.
Theoretical calculations using Density Functional Theory
(DFT) showed a good correlation between the donor capabili-
ties of the alkyl group R34 with the charge at the oxygen
(Fig. 1).
Thus, GG was prepared in excellent yield (96%) using a
nucleophilic addition of the corresponding Grignard derivative
of aryl bromide (9) to the TBDMS-protected 3,6-dihydroxy-
xanthenone (10)20 and subsequent dehydration with aqueous
hydrochloric acid (Scheme 1). In this case, the tert-butyl group
present in the final product fulfilled a dual role. This group
was required to increase the pKa value and was also required
Fig. 1 Evaluated structures and calculated charge.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 6432–6439 | 6435

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 1 Synthetic sequence for the preparation of compound 1.
for the regioselective placement of the halogen in the bromide
precursor 8.
Photophysics
The photophysical behaviour of GG in aqueous solutions at
near-physiological pH has been studied by absorption and by
steady-state and time-resolved emission spectroscopy.
Fig. 5 Steady-state emission spectra (λex = 490 nm) of 7 × 10−6 M GG
aqueous solutions at pH values between 6 and 9.
The visible absorption spectra of aqueous solutions of GG
in the pH range between 6 and 9 were recorded. The absorp-
tion spectral changes occurring in this pH region are dictated
by the ground-state pKa values. Since the isosbestic point was
consistently around 455 nm at any phosphate concentration,
we concluded that phosphate buffer does not significantly
perturb the absorption spectrum of the aqueous GG solutions,
and therefore this compound does not form ground-state com-
plexes with phosphate buffer. Fig. 4 shows two prototropic
forms and only one isosbestic point. Thus, a single ground-
state pKa is present in this pH range. The similarity of the
spectral profiles of the two prototropic forms to those of the
Tokyo Green family21 suggests that the two visible absorbing
forms correspond to the neutral and the anionic form.
brium constant by fluorimetric titration. Therefore, steady-
state fluorescence spectra were collected at an excitation wave-
length of 490 nm from solutions in a pH range between 6 and
9. Fig. 5 shows the variations in fluorescence intensity with
pH. These changes were in good agreement with the changes
observed in the absorption spectra. A pKa value of 7.26 0.019
was obtained from the fit of the fluorescence experimental
results (Fig. S2†).
Quantum yield values from steady-state fluorescence
measurements were calculated for the anion forms using
fluorescein in 0.1 M NaOH as a reference (ϕfluo = 0.95). The
quantum yield of the neutral form was obtained by fitting the
steady-state fluorescence spectra to the equilibrium equation
(see ESI† for details) once the values of ϕA were known. The
recovered quantum yields were 0.97 and 0.10 for the anion and
the neutral form, respectively.
The fluorescence decay traces in the absence of phosphate
buffer were recorded in the pH range between 6 and 9 at one
λex = 485 nm and three λem (505, 515, and 525 nm). At all pH
values, the decays followed a monoexponential decay function
with a lifetime value of 3.97 ns, which was independent of pH.
This result can be explained by the small contribution of the
neutral form to fluorescence due to the low extinction coeffi-
cient at the excitation wavelength (485 nm) and its low
fluorescence quantum yield.
To recover the pKaapp and the molar absorption coefficients
εi (λabs) of the two acid–base species, a global fit of the absor-
12
bance vs. pH and vs. λabs to the corresponding acid–base
equilibrium equations was carried out (Fig. S1†). We obtained
a pKaapp value of 7.39 0.07. The anionic form showed a peak
at 495 nm with a value of εAnion(495) = 60 480 880.
At a constant excitation intensity, low absorbance values,
and negligible rates of proton transfer in the excited state,40,41
it is possible to determine the ground state acid–base equili-
Once the major spectral characteristics of GG near the physio-
logical pH were recovered, we focused on the ESPT reaction
between the dye and phosphate. We studied the absorption
and emission spectra as a function of the phosphate buffer
concentration at pH 7.35. We found that an increase in the
phosphate buffer concentration in the range of 5 mM to
600 mM showed pronounced effects on the emission spectra
but only weak effects (ionic strength) on the absorption
spectra. The pronounced effects on the emission spectrum
must therefore be due to the presence of the well-known ESPT
reaction.13 Thus, we studied the effect of the phosphate con-
centration by means of time-resolved fluorescence. The fluore-
scence decay traces should be influenced by the presence of
ESPT reactions, causing the decay times to be dependent on
Fig. 4 Absorption spectra of GG (at 7 × 10−6 M) in phosphate buffer
(0.005 M) at pH values between 6 and 8. The arrow indicates decreasing
pH values. The isosbestic point at 456 nm is clearly observed. For clarity,
Fig. 4 only contains spectra from six solutions.
6436 | Org. Biomol. Chem., 2014, 12, 6432–6439
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 6 Recovered decay times at pH = 7.35 vs. [phosphate].
the pH and on the phosphate concentration.13,22 We recorded
fluorescence decay traces from aqueous solutions of GG at pH
7.35 and at different phosphate buffer concentrations, CB,
ranging between 5 mM and 600 mM, as a function of λex
(485 nm) and λem (505, 515, and 525 nm). Fig. 6 shows the
recovered decay times at pH = 7.35 vs. phosphate concen-
tration. The standard errors are obtained from the diagonal
elements of the covariance matrix available from the global
analysis fit of decay traces recorded at the three different emis-
sion wavelengths and are between 0.020 and 0.032 ns. The
decays were well analyzed by monoexponential functions,
Fig. 7 FLIM images (A), average decay time (B), and frequency lifetime
histograms (C) of 1 × 10−7 M GG in Dulbecco’s modified Eagle’s medium
in 50 mM TRIS buffer at pH 7.35 and with varying phosphate concen-
trations in the range of 10 to 400 mM.
Fig. 7A, exemplary FLIM images from the samples are shown
after applying an arbitrary colour scale to the recovered fluo-
rescence decay times. The recovered images illustrate that the
phosphate-response of the dye was maintained even in these
crowded solutions. Considering the complexity of the matrix,
the results are very acceptable, as the differences in the phos-
phate concentrations in the media were clearly observable. The
average values of the fluorophore decay times (Fig. 7B) were in
good agreement with those obtained by conventional time-
resolved fluorimetry in aqueous solution (shown in Fig. 6).
The standard error calculated for the decay time of GG on the
glass surface was consistently low, even though the decay time
distributions were calculated from several FLIM images, which
indicates the robustness of the technique. The overall fre-
quency histograms of the decay times obtained from individ-
ual pixels of several images (at least three) collected under the
same conditions clearly showed a shift in the decay time of the
dye, allowing the phosphate response of the dye to be deter-
mined (Fig. 7C).
Notably, not only was the response range of GG wider, its
sensitivity was also slightly higher compared with 2-Me-4-OMe
TG. Indeed, the response range of GG extended to 400 mM,
whereas 2-Me-4-OMe TG was only effective at phosphate con-
centrations of up to 240 mM. In terms of sensitivity, the
maximum change in the fluorescence decay time within the
response range was ∼0.4 ns versus the 0.25 ns change shown
by TG. Even in the concentration range studied with 2-Me-4-
OMe TG, the decrease in the GG decay time was 15% higher.
2
given good fitting parameters (χg < 1.06). From this plot, we
observed that the change in the decay times was 0.1 ns higher
than that previously obtained with 2-Me-4-OMe TG in the
same phosphate buffer concentration range.21
Biomimetic media
FLIM technology is particularly attractive for intracellular
sensing because of the advantages of time-resolved fluore-
scence techniques. Sensors based on the emission intensity
have several disadvantages for quantitative measurement
because they are prone to concentration dependence, inter-
ference from cell autofluorescence, and several sources of
optical excitation power drift. In contrast to intensity measure-
ments, decay times are independent of the local concentration
of the probe and are inherently robust in the presence
of absorption and scattering, although they are still highly
sensitive to changes in the environment.
If GG is intended to be employed as an intra-cellular
sensor, the highly crowded intracellular environment could
alter its response. Therefore, we tested the capability of the dye
to estimate the concentration of phosphate in solutions
mimicking the cytoplasm. Thus, GG was dissolved in solutions
of Dulbecco’s modified Eagle’s medium at pH 7.35 in the pres-
ence of different phosphate concentrations in the 10 to
400 mM range. The dye was at a concentration of 1 × 10−7 M.
Fluorescence imaging microscopy of GG in preosteoblast cells
The solutions were deposited on the surface of a microscope The results obtained on glass slides suggest that this dye may
glass slide, and FLIM confocal images were recorded following have advantages over 2-Me-4-OMe TG as an intracellular probe
the procedure described in the Experimental section. In for phosphate using FLIM technology. First, we tested whether
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 6432–6439 | 6437

View Article Online
Paper
Organic & Biomolecular Chemistry
GG satisfies the additional requirements of water solubility colour scale allows a reasonably accurate estimate of the
and membrane permeability by performing experiments on different phosphate concentrations inside the permeabilised
live cells and obtaining images of the dye in the cell. Based cells.
on the fluorescence intensity upon excitation at 485 nm of the
cells during dye loading, it was concluded that the dye
accumulates efficiently inside the cytosol but is excluded from
the cell nucleus. To gain more insight into the temporal behav-
Conclusions
iour of the GG fluorescence emission in the intracellular In summary, we designed and synthesised a new dye that has
medium and the effect of different concentrations of phos- chemical reactivity and spectral characteristics similar to those
phate ions, we used MC3T3-E1 preosteoblast cells treated with of 2-Me-4-OMe TG but shows a higher pKa for the neutral and
α-toxin and various concentrations of pH 7.35 phosphate anion equilibrium. The new dye, referred to as GG, undergoes
buffer. The MC3T3-E1 murine preosteoblast cell line is a well- the same phosphate-mediated reaction in the excited state as
established model for osteoblast differentiation. The ability of does fluorescein and its derivatives, giving rise to fluorescence
osteoblasts to transport extracellular phosphate into the cell decay traces that are dependent on the phosphate concen-
through a transporter has been established.5
tration in the medium. The advantages of GG with respect to
Treatment with α-toxin generated 1.5 nm membrane pores, 2-Me-4-OMe TG, a previously tested intracellular phosphate
which allowed the diffusion of low molecular weight com- sensor, are its ability to detect a wider range of phosphate con-
pounds, including both GG and phosphate anions, without centrations along with an additional slight increase in sensi-
the loss of cytosolic proteins and high molecular weight com- tivity, measured as the ratio of the phosphate concentration/
pounds.30 FLIM images were recorded and analysed using the decay time. To prove the usefulness of GG as a phosphate
procedure described in the Experimental section. Fig. 8 shows sensor in the cellular cytoplasm, MC3T3-E1 preosteoblast cells
the FLIM images revealing changes in the decay time of the permeabilised with α-toxin and submerged in PBS solutions at
dye inside the MC3T3-E1 osteoblast cells treated with α-toxin different phosphate concentrations at pH 7.35 were used. The
versus the spiked phosphate concentration ranging from 10 to FLIM images recovered show that GG adequately reflects the
240 mM after applying an arbitrary colour scale. Although the intracellular phosphate concentration.
range of decay times is narrow, the use of an appropriate
Acknowledgements
This work was supported by grants CTQ2010-20507/BQU and
CTQ-2011.22455 from the Ministerio Español de Ciencia e
Innovación and P09-FQM-4571 from Regional Government of
Andalucía. AO acknowledges the grant P10-FQM-6154 from
the Conserjeria de Economia, Innovacion, Ciencia y Empleo
(Junta de Andalucia). DM thanks Spanish MICINN (project
CTQ-2011.22455) for her contract.
Notes and references
1 M. K. Kuimova, G. Yahioglu, J. A. Levitt and K. Suhling,
J. Am. Chem. Soc., 2008, 130, 6672–6673.
2 B. Hötzer, R. Ivanov, T. Brumbarova, P. Bauer and G. Jung,
FEBS J., 2012, 279, 410–419.
3 K. Okabe, N. Inada, C. Gota, Y. Harada, T. Funatsu and
S. Uchiyama, Nat. Commun., 2012, 3, 1–9.
4 M. Y. Berezin and S. Achilefu, Chem. Rev., 2010, 110,
2641–2684.
5 G. R. Beck Jr., B. Zerler and E. Moran, Proc. Natl. Acad.
Sci. U. S. A., 2000, 97, 8352–8357.
6 T. Schenk, N. M. G. M. Appels, E. D. A. van, H. Irth,
U. R. Tjaden and d. G. J. van, Anal. Biochem., 2003, 316,
118–126.
Fig. 8 FLIM images of 1 × 10−7 GG dye in the cytoplasm of MC3T3-E1
osteoblast cells treated with α-toxin after incubation in extracellular
7 J. Hirose, H. Fujiwara, T. Magarifuchi, Y. Iguti, H. Iwamoto,
media with different phosphate concentrations. The image labelled as
S. Kominami and K. Hiromi, Biochim. Biophys. Acta, 1996,
1296, 103–111.
autofluorescence corresponds to the FLIM image of the cells collected
without the dye.
6438 | Org. Biomol. Chem., 2014, 12, 6432–6439
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
8 H. Imamura, K. P. Huynh Nhat, H. Togawa, K. Saito, 28 A. Martinez-Peragon, D. Miguel, R. Jurado, J. Justicia,
R. Iino, Y. Kato-Yamada, T. Nagai and H. Noji, Proc. Natl.
Acad. Sci. U. S. A., 2009, 106, 15651–15656.
J. M. Alvarez-Pez, J. M. Cuerva and L. Crovetto, Chem. – Eur.
J., 2014, 20, 447–455.
9 M. Ito, S. Haito, M. Furumoto, Y. Uehata, A. Sakurai, 29 H. Sudo, H. A. Kodama, Y. Amagai, S. Yamamoto and
H. Segawa, S. Tatsumi, M. Kuwahata and K.-i. Miyamoto,
Am. J. Physiol., 2007, 292, C526–C534.
10 M. R. Ganjali, M. Hosseini, Z. Memari, F. Faridbod,
S. Kasai, J. Cell Biol., 1983, 96, 191–198.
30 M. D. Giron, C. M. Havel and J. A. Watson, Proc. Natl. Acad.
Sci. U. S. A., 1994, 91, 6398–6402.
P. Norouzi, H. Goldooz and A. Badiei, Anal. Chim. Acta, 31 J. R. Lakowicz, Principles of Fluorescence Spectroscopy,
2011, 708, 107–110. Springer, 3rd edn, 2006.
11 J. M. Paredes, M. D. Giron, M. J. Ruedas-Rama, A. Orte, 32 S. E. Braslavsky, A. U. Acuna, W. Adam, F. Amat,
L. Crovetto, E. M. Talavera, R. Salto and J. M. Alvarez-Pez,
J. Phys. Chem. B, 2013, 117, 8143–8149.
12 J. Yguerabide, E. Talavera, J. M. Alvarez and B. Quintero,
Photochem. Photobiol., 1994, 60, 435–441.
13 J. M. Alvarez-Pez, L. Ballesteros, E. Talavera and
J. Yguerabide, J. Phys. Chem. A, 2001, 105, 6320–6332.
14 N. Boens, W. Qin, N. Basaric, A. Orte, E. M. Talavera and
J. M. Alvarez-Pez, J. Phys. Chem. A, 2006, 110, 9334–9343.
15 J. Widengren, V. Kudryavtsev, M. Antonik, S. Berger,
M. Gerken and C. A. M. Seidel, Anal. Chem., 2006, 78,
2039–2050.
D. Armesto, T. D. Z. Atvars, A. Bard, E. Bill, L. O. Bjoern,
C. Bohne, J. Bolton, R. Bonneau, H. Bouas-Laurent,
A. M. Braun, R. Dale, K. Dill, D. Doepp, H. Duerr,
M. A. Fox, T. Gandolfi, Z. R. Grabowski, A. Griesbeck,
A. Kutateladze, M. Litter, J. Lorimer, J. Mattay, J. Michl,
R. J. D. Miller, L. Moggi, S. Monti, S. Nonell, P. Ogilby,
G. Olbrich, E. Oliveros, M. Olivucci, G. Orellana,
V. Prokorenko, K. R. Naqvi, W. Rettig, A. Rizzi, R. A. Rossi,
E. San Roman, F. Scandola, S. Schneider, E. W. Thulstrup,
B. Valeur, J. Verhoeven, J. Warman, R. Weiss, J. Wirz and
K. Zachariasse, Pure Appl. Chem., 2007, 79, 293–465.
16 T. Miura, Y. Urano, K. Tanaka, T. Nagano, K. Ohkubo and 33 M. Maus, M. Cotlet, J. Hofkens, T. Gensch, S. F. C. De,
S. Fukuzumi, J. Am. Chem. Soc., 2003, 125, 8666–8671.
J. Schaffer and C. A. Seidel, Anal. Chem., 2001, 73, 2078–
17 T. Ueno, Y. Urano, K. Setsukinai, H. Takakusa, H. Kojima,
2086.
K. Kikuchi, K. Ohkubo, S. Fukuzumi and T. Nagano, J. Am. 34 C. Hansch, A. Leo and R. W. Taft, Chem. Rev., 1991, 91,
Chem. Soc., 2004, 126, 14079–14085. 165–195.
18 T. Mineno, T. Ueno, Y. Urano, H. Kojima and T. Nagano, 35 K. S. Alongi and G. C. Shields, in Annual Reports in Compu-
Org. Lett., 2006, 8, 5963–5966.
19 R. Zhou and T. Ha, Methods Mol. Biol., 2012, 922, 85–100.
tational Chemistry, ed. A. W. Ralph, Elsevier, 2010, vol. 6,
pp. 113–138.
20 Y. Urano, M. Kamiya, K. Kanda, T. Ueno, K. Hirose and 36 K. C. Gross and P. G. Seybold, Int. J. Quantum Chem., 2001,
T. Nagano, J. Am. Chem. Soc., 2005, 127, 4888–4894. 85, 569–579.
21 J. M. Paredes, L. Crovetto, R. Rios, A. Orte, J. M. Alvarez-Pez 37 Structures 3 and 4 showed the same trend (8.16 and 7.74 kJ
and E. M. Talavera, Phys. Chem. Chem. Phys., 2009, 11,
5400–5407.
22 L. Crovetto, J. M. Paredes, R. Rios, E. M. Talavera and
J. M. Alvarez-Pez, J. Phys. Chem. A, 2007, 111, 13311–
13320.
23 J. M. Paredes, A. Garzon, L. Crovetto, A. Orte, S. G. Lopez
and J. M. Alvarez-Pez, Phys. Chem. Chem. Phys., 2012, 14,
5795–5800.
24 C. R. Schroeder, B. M. Weidgans and I. Klimant, Analyst,
2005, 130, 907–916.
mol−1 respectively). Nevertheless, for compound 2 the
calculated free energies for structures around the minimum
of energy oscillated up to 3 kJ mol−1, avoiding a reliable
comparison with the other values.
38 In this case, the torsion angle in the minimum energy
structure is about 66°. When we artificially force the angle
to 90° the situation is clearer and the proton transfer
energy is only 1.04 kJ mol−1 with respect to 2-Me-4-OMe
TG.
39 T. Nagano, Proc. Jpn. Acad., B, 2010, 86, 837–847.
25 A. M. Paradiso, R. Y. Tsien and T. E. Machen, Proc. Natl. 40 E. Cielen, A. Tahri, K. Ver Heyen, G. J. Hoornaert, F. C. De
Acad. Sci. U. S. A., 1984, 81, 7436–7440.
Schryver and N. Boens, J. Chem. Soc., Perkin Trans. 2, 1998,
1573–1580.
41 E. Cielen, A. Stobiecka, A. Tahri, G. J. Hoornaert, F. C. De
Schryver, J. Gallay, M. Vincent and N. Boens, J. Chem. Soc.,
Perkin Trans. 2, 2002, 1197–1206.
26 M. L. Graber, D. C. DiLillo, B. L. Friedman and
E. Pastoriza-Munoz, Anal. Biochem., 1986, 156, 202–212.
27 L. D. Lavis, T. J. Rutkoski and R. T. Raines, Anal. Chem.,
2007, 79, 6775–6782.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 6432–6439 | 6439