Synthesis and characterisation of Thio-H, a new excitation and emission ratioable fluorescent Ca2+/Mg2+ indicator with high brightness

Article abstract of DOI:10.1039/b107042h

A newfluorescent ratiometric indicator for Ca²⁺ and Mg²⁺, tricaesium salt of {2-[bis(carboxymethyl)amino]-5-(5-phenylthiophen-2-yl)phenoxy}acetic acid, Thio-H, has been synthesised and evaluated for its cation binding propertiesvia fluorimetric titrations. Thio-H has the interesting property of being both excitation and emission ratioable, expanding the range of possible applications. The in vitro dissociation constant, K_d, measured at 20°C in 100 mM KCl, pH 7.20 for the Ca²⁺-Thio-H complex is 45 ± 13 μM; for Mg²⁺-Thio-H a K_d value of 5.6 ± 0.6 mM is found. The free form of Thio-H has a highfluorescence quantum yield (0.74), which decreases to 0.50 and 0.65 upon binding to Ca²⁺ and Mg²⁺, respectively. Measurements under physiological conditions show that increasing the temperature from 20 to 37°C decreases the affinity of Thio-H for Ca²⁺. Changing the pH from 7.05 to 7.40 does not affect the K_d value of the Mg²⁺ complex but it increases somewhat the affinity of the probe for Ca²⁺. In the presence of 1 mM Mg²⁺, the Ca²⁺ affinity of Thio-H decreases (K_d = 126 ± 46 μM at 20°C, pH 7.05). Time-resolved fluorescence measurements confirm that the inflection point in thefluorimetric titration curve can be correctly assigned to K_d. Since the dissociation constant for the Mg²⁺-indicator complex falls within the intracellular [Mg²⁺] domain, while the K_d value for Ca²⁺ is well above the basal Ca²⁺ levels, Thio-H has potential applications as a Mg²⁺ indicator. The low affinity for Ca²⁺ can be exploited for detecting intracellular Ca²⁺ levels in the micromolar range, on condition that [Mg²⁺]_i remains practically constant during the measurements. Thio-H provides an excellent addition to the commercial ratiometric low-affinity Ca²⁺ indicators Mag-fura-2, Mag-fura-5, Mag-indo-1, Fura-FF, and BTC.

Full text of DOI:10.1039/b107042h

View Online / Journal Homepage / Table of Contents for this issue
Synthesis and characterisation of Thio-H, a new excitation and
emission ratioable fluorescent Ca2؉/Mg2؉ indicator with high
brightness
2
Els Cielen,a Agnieszka Stobiecka,a,b Abdellah Tahri,a Georges J. Hoornaert,a
Frans C. De Schryver,a Jacques Gallay,c Michel Vincentc and Noël Boens*a
a Department of Chemistry, Katholieke Universiteit Leuven, Celestijnenlaan 200F,
3001 Heverlee, Belgium. E-mail: Noel.Boens@chem.kuleuven.ac.be; Fax: ϩ 32-(0)16-327 990
b Institute of General Food Chemistry, Technical University Lodz, ul Stefanowskiego 4/10,
90-924 Lodz, Poland
c Laboratoire pour l’Utilisation du Rayonnement Electromagnétique,
Centre Universitaire Paris-Sud, 91405 Orsay, France
Received (in Cambridge, UK) 2nd August 2001, Accepted 5th April 2002
First published as an Advance Article on the web 8th May 2002
A new fluorescent ratiometric indicator for Ca2ϩ and Mg2ϩ, tricaesium salt of {2-[bis(carboxymethyl)amino]-5-
(5-phenylthiophen-2-yl)phenoxy}acetic acid, Thio-H, has been synthesised and evaluated for its cation binding
properties via fluorimetric titrations. Thio-H has the interesting property of being both excitation and emission
ratioable, expanding the range of possible applications. The in vitro dissociation constant, Kd, measured at 20 ЊC in
100 mM KCl, pH 7.20 for the Ca2ϩ–Thio-H complex is 45 13 µM; for Mg2ϩ–Thio-H a Kd value of 5.6 0.6 mM
is found. The free form of Thio-H has a high fluorescence quantum yield (0.74), which decreases to 0.50 and 0.65
upon binding to Ca2ϩ and Mg2ϩ, respectively. Measurements under physiological conditions show that increasing
the temperature from 20 to 37 ЊC decreases the affinity of Thio-H for Ca2ϩ. Changing the pH from 7.05 to 7.40 does
not affect the Kd value of the Mg2ϩ complex but it increases somewhat the affinity of the probe for Ca2ϩ. In the
presence of 1 mM Mg2ϩ, the Ca2ϩ affinity of Thio-H decreases (Kd = 126 46 µM at 20 ЊC, pH 7.05). Time-resolved
fluorescence measurements confirm that the inflection point in the fluorimetric titration curve can be correctly
assigned to Kd. Since the dissociation constant for the Mg2ϩ–indicator complex falls within the intracellular [Mg2ϩ
]
domain, while the Kd value for Ca2ϩ is well above the basal Ca2ϩ levels, Thio-H has potential applications as a Mg2ϩ
indicator. The low affinity for Ca2ϩ can be exploited for detecting intracellular Ca2ϩ levels in the micromolar range,
on condition that [Mg2ϩ]i remains practically constant during the measurements. Thio-H provides an excellent
addition to the commercial ratiometric low-affinity Ca2ϩ indicators Mag-fura-2, Mag-fura-5, Mag-indo-1, Fura-FF,
and BTC.
because of the relatively low affinity of the APTRA-based indi-
Introduction
cators for Ca2ϩ. Some of the commercial fluorescent indicators
The calcium ion is one of the most important and widespread
for detecting changes in intracellular [Ca2ϩ] in the range
intracellular ions. Cytosolic Ca2ϩ regulates many cellular pro-
of > 1 µM (BTC, Fura-FF, Mag-fura-2, Mag-fura-5, Mag-
cesses, ranging from muscle contraction, neuronal signalling,
indo-1)7 exhibit spectral shifts upon ion binding, allowing
secretion, fertilisation, cell division, to metabolism and cell
dual-wavelength ratiometric measurements. Such ratiometric
death.1 Cells at rest have a free intracellular Ca2ϩ concentration
of about 100 nM; in active cells, this level can rise to well over
indicators allow a more reliable measurement of intracellular
[Ca2ϩ] than via single wavelength indicators, because the use
1 µM. The development of fluorescent indicators by Tsien and
of a ratio minimises the effect of artefacts (uneven indicator
colleagues entailed a real breakthrough in the elucidation of
loading, uneven cell thickness, photobleaching, and indicator
intracellular calcium dynamics.2 Although the vast majority of
leakage from the cell) that are unrelated to changes in intra-
experiments involves the measurement of cytoplasmic Ca2ϩ
cellular [Ca2ϩ]. It is obvious that new fluorescent ratiometric
concentrations, low-affinity Ca2ϩ indicators are very suitable for
Ca2ϩ/Mg2ϩ indicators can provide an interesting addition to the
the determination of Ca2ϩ in intracellular compartments3 or
list of existing Ca2ϩ/Mg2ϩ indicators.
for monitoring steep increases in intracellular [Ca2ϩ].4 Since
Therefore, we have developed Thio-H 1, tricaesium salt of
{2-[bis(carboxymethyl)amino]-5-(5-phenylthiophen-2-yl)phen-
oxy}acetic acid (Fig. 1), a new ratiometric fluorescent indicator
Ca2ϩ concentrations around 1 µM produce almost complete
binding saturation of the most frequently used Ca2ϩ indicator
Fura-2,5 but only very low fractional saturation of the low-
affinity analog Mag-fura-2,6 a series of indicators based on the
tricarboxylate chelator APTRA (o-aminophenol-N,N,O-
triacetic acid) with low affinity for Ca2ϩ has been developed.7
APTRA-based indicators are suitable for detecting elevated
Ca2ϩ levels and are sensitive also to Mg2ϩ concentrations from
0.1 to 6 mM, the commonly found range of intracellular free
Mg2ϩ levels. Typical physiological Ca2ϩ concentrations (10 nM–
1 µM) do not usually interfere with Mg2ϩ measurements
Fig. 1 Structure of Thio-H 1.
DOI: 10.1039/b107042h
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206
This journal is © The Royal Society of Chemistry 2002
1197

View Online
for Ca2ϩ/Mg2ϩ
,
consisting of the tricarboxylate chelator
fluorescence signal F(λex,λem,[X]) at cation concentration [X],
APTRA, linked via a Stille reaction to an arylthiophene
fluorophore.
due to excitation at λex and observed at emission wavelength λem
can be expressed by eqn. (2)10
In addition to the description of the synthesis of Thio-H,
we also report its spectroscopic and Ca2ϩ/Mg2ϩ binding
properties. The influence of temperature and pH on the spec-
tral parameters and the complex formation is investigated.
Furthermore, a test based on time-resolved fluorescence
measurements indicates that the inflection point in the fluori-
metric titration curve is correctly assigned to Kd.8
(2)
where Fmin represents the fluorescence signal of the free form
F* of the indicator, and Fmax denotes the fluorescence signal
of the bound form B*. Fmin and Fmax thus correspond to
fluorescence signals at minimum and maximum X concentra-
tion. Fitting eqn. (2) to the fluorescence data F as a function of
Theory
[X] yields values for Kd, n, Fmin, and Fmax
.
Determination of Kd from direct fluorimetric titration
Eqn. (2) can be linearised in the form of a Hill plot [eqn. (3)],
The expression of the fluorescence signal F as a function of the
ion concentration has been derived by Kowalczyk et al.9 for the
case of a 1 : 1 complex between a fluorescent indicator and a
cation. This expression can be modified by taking into account
the formation of an n : 1 ion–indicator complex (n denotes the
stoichiometry of the formed complex, Scheme 1). Consider a
(3)
from which values for n and Kd can be derived. Since fitting
eqn. (2) to the fluorescence data F is a better way for recovering
reliable estimates for Fmin, Fmax, Kd, and n, we have used this
fitting method for obtaining Kd and n values.
Determination of Kd from ratiometric fluorimetric titration
In biological experiments, it is often difficult to keep the
analytical indicator concentration constant during the record-
ing of F, Fmin, and Fmax because of bleaching and leakage of the
fluorescent indicator from cells. In such a case, the ratiometric
method, first described by Grynkiewicz et al.,5 is the method of
choice. For each concentration of X, one can record the ratio
of the fluorescence signals (i) either at two different excitation
wavelengths and a common emission wavelength, or (ii) at
two different emission wavelengths and a single excitation
wavelength.
For the system depicted in Scheme 1, the plot of the ratio
R(λe1x/λe2x, λem, [X]) = F(λe1x, λem)/F(λe2x, λem) obtained from dual
excitation wavelengths at a common observation wavelength
as a function of Ϫlog[X] may have multiple inflections, unless
the rate of excited-state association can be neglected.11 (This
condition can be verified by the constancy of the fluorescence
decay times within the range of used [X];8 see section on Time-
resolved fluorescence). When the excited-state association rate
is negligible, we have
Scheme 1 Kinetic scheme of an intermolecular two-state excited-state
system.
photophysical system consisting of the free form of the indi-
cator (F) which in the ground state can undergo a reversible
association with cation X to form an n : 1 ion–indicator com-
plex [i.e., the bound form (B) of the indicator]. The association–
dissociation is described by the ground-state dissociation
constant Kd [eqn. (1)], assuming an n : 1 stoichiometry between
X and F,
(4)
Kd = [F][X]n/[B]
(1)
in which [ ] denote molarities. It should be emphasised that
Scheme 1 depicts only the overall association–dissociation
between an indicator and a cation,10 so that Kd represents the
composite dissociation constant. In Scheme 1, it is further
assumed that only species F and B absorb light at the excitation
wavelength λex to create the corresponding excited-state species
F* and B*, which can decay by fluorescence (F) and non-
radiative (NR) processes. The composite rate constant for
deactivation of species F* is denoted by k01 (= kF1 ϩ kNR1) and
by k02 for species B*. The overall rate constant describing the
with
(5)
Similarly, the ratio R(λex, λe1m/λe2m, [X]) = F(λex, λe1m)/F(λex, λe2m)
at dual emission wavelengths due to a common excitation wave-
length is given by eqn. (6) when the excited-state association can
be neglected:
association F* ϩ n X
B* is represented by k21, while the
overall rate constant for the dissociation of B* into F* and n X
is denoted by k12.
(6)
If the system depicted in Scheme 1 is excited with light of
constant intensity, and if the absorbance of the sample is less
than 0.1, and moreover, if one assumes that the intermediate
complexes (between F and one or more cations) are present in
concentrations low enough to give only a negligible contribu-
tion to the total fluorescence intensity, and that the rate of
cation binding in the excited state can be neglected, the total
with
(7)
1198
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206

View Online
Rmin is the ratio of the fluorescence intensities at two distinct
excitation wavelengths (respectively emission wavelengths) and
one emission wavelength (respectively one excitation wave-
length) of the free form of the indicator (minimum X con-
centration), Rmax the ratio of the fluorescence intensities of
the bound form of the indicator (maximum X concentration),
and R denotes the ratio of the fluorescence intensities of solu-
tions with intermediate X concentration. Non-linear fitting
eqn. (4) to the excitation ratiometric fluorescence data R as a
Results
Design of Thio-H
Based on promising spectroscopic properties of the model
compound, methyl 4-(5-phenylthiophen-2-yl)benzoate13 (ε346 nm
= 9500 MϪ1 cmϪ1 13
fluorescence quantum yield ꢀf = 0.36 in
,
methanol), a thiophene containing unit was chosen as fluoro-
phore. The ionophore consists of a derivative of the chelator
APTRA. Although the ion cavity formed by this chelating
function of [X] yields values for Kdζ(λe2x, λem), n, Rmin, and Rmax
.
group is of appropriate size to accommodate Mg2ϩ 14
, it has a
Since ζ(λe2x, λem)—the ratio of the fluorescence signal of the
free form of the indicator over that of the bound form at the
indicated wavelengths—is experimentally accessible, a value for
Kd can be recovered from ratiometric excitation fluorescence
data. In an analogous way, fitting eqn. (6) to the ratiometric
emission fluorescence data R as a function of [X] yields
values for Kdζ(λex, λe2m), n, Rmin, and Rmax. Since ζ(λex, λe2m)
can be determined from the fluorescence signals of the free
and bound forms of the indicator, a value for Kd can be
obtained.
much higher affinity for Ca2ϩ than for Mg2ϩ. This property has
been exploited frequently in the design of low-affinity Ca2ϩ
indicators.6,7 It must be noted that APTRA is a good ligand for
many cations, e.g. Zn2ϩ 15
.
Synthesis
Synthesis of the ionophore 6. Retrosynthetically, two alterna-
tive pathways may be proposed for the synthesis of the iono-
phore 6 starting from o-aminophenol 2: (i) either the bromin-
ation is carried out before the introduction of the N,N,O-
triacetic acid branches or (ii) the substitution is carried out in
the reverse order. The first method has been described in detail
in the literature. 5-Bromo-2-aminophenol can be obtained by
nitration of m-bromophenol,16 followed by reduction to an
amine according to the method of Bellamy and Ou.17 Since in
the first step, the desired isomer is obtained only in 20% yield,
this method is not recommendable. Better yields were reported
by Gershon et al. for the reaction of acylated 2-aminophenol
Eqns. (4) and (6) can be linearised in the form of a Hill plot:
(8)
with ζ given by eqn. (5) for excitation ratiometric fluorescence
data and by eqn. (7) for emission ratiometric fluorescence data.
Plotting the left-hand side of eqn. (8) as a function of log [X]
yields values for n and Kdζ. Since ζ is experimentally accessible,
a value for Kd can be determined. As is the case for direct
fluorimetric titrations, non-linear fitting of eqn. (4) or (6) is the
recommended approach to recovering reliable estimates for Kd
and n. Therefore, we have used this method here.
with N-bromosuccinimide in
a
five step protection–
deprotection procedure.18 After bromination, the triacetic acid
branches can be introduced as described for the preparation of
the Mg2ϩ selective indicator Mag-fura-2. By introducing the
N,N,O-triacetic acid branches prior to bromination (ii), we
simplified the sequence to a two-step procedure (Scheme 2).
Time-resolved fluorescence
If the system depicted in Scheme 1 with n = 1 is excited by a δ-
pulse which does not significantly change the concentrations of
the ground-state species (i.e., in the low excitation limit), the
fluorescence decays are dual exponential [eqn. (9)],12
f(λex, λem, t) = α1exp(Ϫτ/t1) ϩ α2exp(Ϫτ/t2), t ≥ 0
(9)
where the amplitudes αi (i = 1, 2) depend on the rate constants
kij, Kd, [X], λex, and λem. The decay times τi (i = 1, 2), however,
depend exclusively on kij and [X]. If the excited-state association
reaction rate is negligible, the decay times become virtually
independent of [X] and are given by
τ1 = 1/k01
(10)
(11)
τ2 = 1/(k02 ϩ k12)
A relatively simple experimental test based on time-resolved
and steady-state fluorescence measurements can be used to
assess the possible misevaluation of Kd from fluorimetric titra-
tions due to the excited-state binding reaction.8 Collecting
fluorescence decay traces in the same [X] range as used in the
fluorimetric titration and analysing them simultaneously as
biexponentials with linked decay times allows one to check the
invariance of the decay times. If an inflection occurs in the [X]
range of the fluorimetric titration curve F where the decay times
are invariant, this inflection point can be correctly associated
with pKd. This means that in this concentration range,
the excited-state binding reaction is negligible and therefore
does not interfere with the correct determination of Kd from
fluorimetric titration. In contrast, the inflection point(s) in
the [X] range where the decay times vary cannot be attributed
to Kd.
Scheme 2 Synthesis of trimethyl 2-amino-5-bromophenol-N,N,O-
triacetate 6. i: a) 2 equivs. “Proton Sponge”, 3 equivs. BrCH2CO2CH3,
0.02 equivs. NaI, CH3CN, 24 h at 70 ЊC, b) 3 equivs. BrCH2CO2CH3, 48
h at 70 ЊC, ii: 1.1 equiv. NBS, CH3CO2H, 30 h at rt.
The N,N,O-triacetic acid branches were introduced onto 2-
aminophenol 2 by application of a procedure based on the
preparation of the fluorescent magnesium selective chelator
Mag-fura-2 (Scheme 2).6 Compound 2 was heated at 70 ЊC in
acetonitrile in the presence of sodium iodide, 2 equivs. “Proton
Sponge” [1,8-bis(dimethylamino)naphthalene] and 3 equivs.
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206
1199

View Online
methyl bromoacetate. After 24 hours, another 3 equivs. methyl
bromoacetate were added, whereafter the mixture was heated
for 48 hours at 70 ЊC to yield trimethyl 2-aminophenol-N,N,O-
triacetate 3. Significant amounts of a lactone, methyl 2-(2-oxo-
2,3-dihydro-4H-1,4-benzoxazin-4-yl)acetate 4, and a lactam,
methyl 2-(3-oxo-2,3-dihydro-4H-1,4-benzoxazin-4-yl)acetate 5,
were formed by intramolecular cyclisation. Bromination of 3
with N-bromosuccinimide gave the desired ionophore 6 in good
yields.
The solutions contained 100 mM KCl and 10 mM MOPS
(3-[N-morpholino]propanesulfonic acid), and were adjusted
with potassium hydroxide or hydrochloric acid to the desired
pH. The free calcium in the separate solutions was adjusted
with Ca2ϩ–nitrilotriacetic acid buffers, as described by Fabiato
and Fabiato.22 Free [Mg2ϩ] was likewise controlled by Mg2ϩ
–
EGTA [ethylene glycol bis(β-aminoethyl ether)-N,N,NЈ,NЈ-
tetraacetic acid] buffers.23 The indicator concentration was
in the range (3.0–3.3) × 10Ϫ6 M, yielding an absorbance
per cm path length of approximately 0.1 at the absorption
maximum.
Synthesis of the fluorophore 10. The fluorophore was syn-
thesised via palladium catalysed cross-coupling reactions
(Scheme 3) using the conditions described by Bumagin and
Excitation and emission spectra. The spectral properties of
Thio-H 1 at different pH are summarised in Table 1. As an
example, the changes in the excitation spectra that accompany
Ca2ϩ and Mg2ϩ binding are illustrated in Figs. 2A and 2B,
respectively.
Comparison of these figures shows that the binding of the
two divalent metal ions by Thio-H results in a hypsochromic
shift of the excitation maximum of about 25 nm relative to the
free form. The excitation spectra as a function of Ca2ϩ or Mg2ϩ
show a crossover point around 330 nm, which corresponds to
a pseudo-isosbestic point. The real isosbestic point (i.e. the
wavelength where ε for the free and the bound species are equal)
was obtained from the absorption spectra and is located around
346 nm.
The emission intensity of the free form is reduced to half
when saturated with Ca2ϩ (Fig. 2C), while the position of this
maximum remains located around 500 nm. Furthermore, a
shoulder shows up around 390–395 nm with increasing [Ca2ϩ],
corresponding to the bound form. Increasing magnesium con-
centrations do not only decrease the fluorescence intensity but
also shift the emission maximum to around 489 nm (Fig. 2D).
Upon excitation at 363 nm, the emission spectra display a
crossover point around 431 nm with Ca2ϩ (Fig. 2C) and one
around 451 nm with Mg2ϩ (Fig. 2D). The position of these
pseudo-isoemissive points depends on the excitation wave-
length. The real isoemissive point can be determined by exciting
the indicator at λex = isosbestic point9 and occurs at 446 nm for
Ca2ϩ and 472 nm for Mg2ϩ. Thio-H is a very bright fluorescent
indicator with a high ꢀf value for the free as well as the Ca2ϩ or
Mg2ϩ bound forms. Because of the large value of the product of
the molar extinction coefficient ε and fluorescence quantum
yield ꢀf, Thio-H has a high fluorescence output per indicator,
allowing low indicator concentrations to be used,24 which
decreases the cation buffering effects of Thio-H.
Scheme 3 Synthesis of Thio-H 1. i: 2 mol% (MeCN)2PdCl2, DMF, air,
rt, ii: a) 9, BuLi, THF, Ϫ78 ЊC, b) SnBu3Cl, Ϫ78 ЊC, rt, iii: 1 mol%
Pd(PPh3)4, toluene, Ar atmosphere, 15 h reflux, iv: excess CsOH,
MeOH, overnight reflux.
Bumagina.19 The fluorophore skeleton of Thio-H, 9, was
obtained by stirring overnight at room temperature a mixture
of 2-(tributylstannyl)thiophene 8 and iodobenzene 7 in DMF
under air, catalysed by bis(acetonitrile)dichloropalladium().
Subsequently, the asymmetrical heterobiaryl compound 9 was
converted into the corresponding stannyl derivative 10 by
quenching of an in situ generated lithium intermediate with
tributyltin chloride.20
Synthesis of the indicator Thio-H 1. The ionophore 6 was
attached to the fluorophore 10 by a palladium-catalysed cross-
coupling reaction using the conditions previously described by
us for (hetero)aryl–thiophene coupling reactions10 (Scheme 3).
Therefore, the fluorophore 10 was coupled to trimethyl 2-
amino-5-bromophenol-N,N,O-triacetate 6 in toluene under
argon in the presence of 1 mol% tetrakis(triphenylphosphine)-
palladium to yield the ester form of the desired indicator 11.
Cation binding properties. From the excitation and emission
spectra, the dissociation constants Kd and the stoichiometry n
of the complexes of Thio-H with Ca2ϩ and Mg2ϩ can be deter-
mined. The results of the non-linear fitting of the fluorimetric
titration data obtained for both cations are shown in Fig. 3.
The estimated Kd and n values for the Ca2ϩ–Thio-H and
Mg2ϩ–Thio-H complexes are summarised in Tables 2 and 3,
respectively.
The first method for estimating Kd and n is the direct fluori-
metric titration [see eqn. (2)]. For the Ca2ϩ–Thio-H complex at
pH 7.20, Kd and n can be determined by monitoring the de-
crease of the fluorescence excitation intensity at the maximum
(Fig. 3A) or by monitoring the decrease of the fluorescence
emission intensity at the maximum (figure not shown). Kd
values for Ca2ϩ–Thio-H at pH 7.20 are 49 15 µM (excitation)
and 56 19 µM (emission) (Table 2), which are in the range of
elevated intracellular [Ca2ϩ]. The Kd value for Mg2ϩ–Thio-H at
pH 7.20, obtained from the fluorescence excitation intensities
at 364 nm with the emission wavelength set at 500 nm is 5.7
0.4 mM (Fig. 3C) and 6.3 0.5 mM from fluorescence emission
intensities at the same wavelength settings (Table 3). All n values
indicate a 1 to 1 stoichiometry between indicator and cation
(Ca2ϩ and Mg2ϩ). The same non-linear fitting method was used
Conversion to the water-soluble form. The final reaction step
comprised the conversion of the ester 11 into water-soluble
form, because the measurements have to be performed in
aqueous buffer solutions. The trimethyl ester 11 was converted
into the corresponding tricaesium salt 1 by reflux in methanol
with a large excess of caesium hydroxide, according to the
method of Minta and Tsien21 and was used as such for the
fluorescence measurements. Comparison of the UV spectra of
the ester in methanol and the tricaesium salt in water indicates
that the fluorophore structure remains unchanged.
Steady-state fluorescence
The spectroscopic measurements were performed on solutions
made up to mimic the intracellular milieu of biological cells.
1200
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206

View Online
Table 1 Spectral properties of Thio-H 1 in free and bound (Ca2ϩ/Mg2ϩ) forms, determined at 20 ЊC in solutions containing 10 mM MOPS and 100
mM KCl. The isosbestic points were determined from the absorption spectra, while the isoemissive points were determined from the emission spectra
upon excitation at the isosbestic points
Isosbestic
point/nm
Isoemissive
point/nm
c
Absorptiona λmax/nm
Emissionb λmax/nm
ꢀf
pH
Free
Ca2ϩ
Mg2ϩ
Ca2ϩ
Mg2ϩ
Free
Ca2ϩ
Mg2ϩ
Free
Ca2ϩ
Mg2ϩ
Ca2ϩ
Mg2ϩ
7.05
7.20
7.40
351 (31)
359
359
332 (31)
335
335
335 (30)
335
335
345
347
346
345
347
346
505
500
500
505 (390)
499 (395)
498 (395)
488
489
491
0.73
0.74
0.75
0.50
0.50
0.55
0.57
0.65
0.61
449
446
449
449
472
472
a Numbers in parentheses correspond to the molar extinction coefficients, εmax × 10Ϫ3 (MϪ1 cmϪ1). b Numbers in parentheses correspond to emission
wavelength at which a shoulder shows up at increasing [Ca2ϩ]. c Estimated error not higher than 10%.
Fig. 2 Fluorescence excitation spectra of Thio-H 1 in solutions containing zero to 1.07 mM Ca2ϩ (A) or zero to 99.7 mM Mg2ϩ (B). Corresponding
fluorescence emission spectra of Thio-H 1 as a function of [Ca2ϩ] (C) or [Mg2ϩ] (D). All spectra were recorded at 20 ЊC in solutions containing 100
mM KCl and 10 mM MOPS pH 7.20.
to estimate Kd and n at other pH values by monitoring changes
in fluorescence emission or excitation intensities.
45 13 µM; for the Mg2ϩ complex at the same pH an average
Kd value of 5.6 0.6 mM is found.
Secondly, for in vivo determinations of the intracellular Ca2ϩ
and Mg2ϩ concentrations, the ratiometric method [see eqns. (4)
and (6)] is preferred, because experimental artefacts can be
eliminated using this method. For example, for Ca2ϩ binding
at pH 7.20 (Table 2), the ratio of the fluorescence excitation
intensities at 338 and 363 nm with observation at 500 nm as a
function of [Ca2ϩ] was used to estimate Kd and n (Fig. 3B). The
Influence of pH, temperature and the presence of competitive ions
on the binding properties of Thio-H
As the promising in vitro properties of Thio-H point towards
possible biological applicability, the influence of some physio-
logical conditions on the complexation behaviour with Ca2ϩ
was investigated.
estimated Kd value of 67
32 µM agrees well with those
obtained via the direct method. Again the stoichiometry
between probe and Ca2ϩ was 1 to 1. Analogous results were
found for Mg2ϩ binding by Thio-H. For example, at pH 7.20,
the ratio of the fluorescence intensities at two excitation wave-
lengths (λ1 = 341 nm and λ2 = 364 nm) and at the common
According to the results of the fluorimetric titrations, the Kd
value estimated for the Ca2ϩ–Thio-H complex (Table 2) is pH
dependent: an increase in pH from 7.05 to 7.40 increases some-
what the affinity of the probe for Ca2ϩ (the average Kd drops
from 56 5 to 36 3 µM in this pH range). In contrast, in the
observation wavelength (λem = 500 nm) as a function of [Mg2ϩ
]
pH range between 7.05 and 7.40 the average Kd value of Mg2ϩ
–
(Fig. 3D) yields Kd = 6.0 0.4 mM and n = 1.05 0.03. When
the emission ratiometric mode was used (λ1 = 482 nm and λ2 =
500 nm) at the common excitation wavelength (λex = 346 nm),
the estimated Kd value at pH 7.20 was found to be 5.5 0.5 mM
Thio-H remains constant around 5.5 mM (Table 3).
The influence of temperature was investigated by carrying
out binding measurements for Ca2ϩ at 20 ЊC and 37 ЊC. Due to
the temperature dependence of the pH of the used buffer, the
pH of 7.20 measured at 20 ЊC decreases to 7.01 at 37 ЊC.
Increasing the temperature causes a decrease of the affinity
of Thio-H for Ca2ϩ, in contrast to the observations for the
complex formation between Ca2ϩ and the tetracarboxylate
chelators EGTA,25 BAPTA [1,2-bis(o-aminophenoxy)ethane-
N,N,NЈ,NЈ-tetraacetic acid],25 Fura-2,26,27 and Indo-1.27
with n = 0.92
0.05 (figure not shown). When n was kept
fixed at unity in the fitting, the estimated Kd value was 6.2
0.4 (Table 3). As shown in Tables 2 and 3, the Kd and n values
estimated via the emission and excitation ratiometric methods
are in good agreement with those obtained by the direct titra-
tion. The average Kd value for the Ca2ϩ complex at pH 7.20 is
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206
1201

View Online
Table 2 Binding properties of Thio-H 1 with Ca2ϩ determined from fluorimetric titrations. The experimental conditions were the same as in Table 1
except when indicated. All Kd and n values were estimated by non-linear fitting of eqns. (2), (4), or (6)
pH
Experimental excitation and emission wavelengths/nm
Kd/µM
n
7.01a
λex = 362 λem = 503
96 18b
90 17c
85 15
50 14c
1.05 0.05
1.04 0.05
1.03 0.04
1.02 0.03
1.19 0.06
1.03 0.02
—
1.01 0.09
0.99 0.08
1.0 0.1
0.87 0.04
1e
λe1x/λe2x = 337/362 λem = 500
λex = 360 λem = 500
7.05
7.20
λe1x/λe2x = 350/370 λem = 500
λex = 350 λe1m/λe2m = 400/502
λex = 360 λem = 500
60
58
4
6
126 46c, d
56 19b
49 15c
67 32
26 3b
40 3b
43 9b
48 3b
28 3b
45 3b
42 15b
38 14c
λex = 363 λem = 500
λe1x/λe2x = 338/363 λem = 500
λex = 335 λem = 395
λex = 347 λem = 500
λex = 347 λem = 395
λex = 364 λem = 500
0.96 0.06
1e
0.86 0.03
1e
1.1 0.1
1.0 0.1
0.87 0.06
0.97 0.09
1e
7.40
λe1x/λe2x = 339/364 λem = 500
λex = 335 λem = 395
34
7
31 10b
34 3b
35 10b
36 3b
λex = 347 λem = 395
1.00 0.07
1e
a Measurements performed at 37 ЊC (pH at 20 ЊC is 7.20, at 37 ЊC pH is 7.01). b Determined by monitoring the decrease in the fluorescence emission
intensity. c Determined by monitoring the decrease in the fluorescence excitation intensity. d Kd was determined in solutions containing 1 mM Mg2ϩ
e n fixed at 1.
.
Fig. 3 Estimation of Kd and n for Ca2ϩ–Thio-H 1 and Mg2ϩ–Thio-H 1 complexes using direct fluorimetric titration method [eqn. (2); A and C,
respectively] and ratiometric method [eqn. (4); B and D, respectively]. Data were obtained from experiments performed at 20 ЊC and pH 7.20.
Binding of Thio-H to Ca2ϩ is thus less efficient at higher
temperatures (average Kd = 90 6 µM at 37 ЊC, pH 7.01 vs. Kd =
56 5 µM at 20 ЊC, pH 7.05).
of Thio-H for Ca2ϩ (Kd = 126
46 µM at 20 ЊC, pH 7.05).
Similar observations were made for the binding of Ca2ϩ by
the tetracarboxylate chelator BAPTA28 and for the affinity
of Fura-25 and BTC29 for Ca2ϩ in the absence and presence
While the presence of physiological Ca2ϩ concentrations does
not affect the determination of [Mg2ϩ], on condition that
[Ca2ϩ]i remains relatively constant during the measurements,
Ca2ϩ binding is markedly perturbed by physiological levels
of Mg2ϩ. From the Kd determination of Ca2ϩ–Thio-H, it was
found that the presence of 1 mM Mg2ϩ halves the affinity
of Mg2ϩ
.
Because of the importance of pH, temperature and com-
petitive ions in intracellular environments, these parameters
should be more extensively investigated if Thio-H turns out to
be a useful in vivo fluorescent indicator of Ca2ϩ/Mg2ϩ
.
1202
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206

View Online
Table 3 Binding properties of Thio-H 1 with Mg2ϩ determined from fluorimetric titrations. The experimental conditions were the same as in
Table 1. All Kd and n values were estimated by non-linear fitting of eqns. (2), (4), or (6)
pH
Experimental excitation and emission wavelengths/nm
Kd/mM
2c
n
7.05
λex = 350 λem = 420
5
1.08 0.03
1.02 0.07
1.05 0.01
1.00 0.05
1.00 0.04
1.05 0.03
0.92 0.06
1a
λe1x/λe2x = 350/370 λem = 500
λex = 350 λe1m/λe1m = 420/500
λex = 364 λem = 500
6.1 0.8
5.3 0.9
6.3 0.5b
5.7 0.4c
6.0 0.4
4.8 0.4b
5.3 0.3b
4.7 0.3
5.0 0.3
5.5 0.5b
6.1 0.4b
5.5 0.5
6.2 0.4
5.2 0.3b
4.6 0.2c
6.0 0.2
7.20
λe1x/λe1x = 341/364 λem = 500
λex = 335 λem = 400
λex = 335 λe1m/λe2m = 479/500
λex = 346 λem = 400
0.93 0.05
1a
0.91 0.06
1a
λex = 346 λe1m/λe2m = 482/500
λex = 364 λem = 500
0.92 0.05
1a
7.40
0.91 0.03
0.89 0.03
1.05 0.02
0.93 0.01
1a
λe1x/λe2x = 341/364 λem = 500
λex = 335 λem = 500
6
6
1b
1b
λex = 335 λem = 400
5.2 0.3b
5.6 0.2b
4.9 0.5
5.2 0.4
5.3 0.5b
5.9 0.5b
5.7 0.2b
6.3 0.3b
5.8 0.2
6.5 0.3
0.94 0.03
1a
λex = 335 λe1m/λe2m = 479/500
λex = 346 λem = 500
0.94 0.07
1a
0.89 0.07
1a
λex = 346 λem = 400
0.91 0.02
1a
λex = 346 λe1m/λe2m = 479/500
0.91 0.02
1a
a n fixed at 1. b Determined by monitoring the decrease in the fluorescence emission intensity. c Determined by monitoring the decrease in the
fluorescence excitation intensity.
Table 4 Global biexponential analysis for each [Ca2ϩ] of the fluor-
Time-resolved fluorescence
escence decays of Thio-H 1 in the presence of Ca2ϩ. Columns 2 and 3
Before a new fluorescent indicator can be released for in vivo
measurements, the correct evaluation of Kd has to be checked.
In the previous sections, Kd values of the complexes of Thio-H
with Ca2ϩ and Mg2ϩ were determined from fluorimetric titra-
tion data. However, since fluorescence depends on both ground-
state and excited-state parameters, the excited-state association
of the cation X with the indicator may influence the value of Kd
derived from a fluorimetric titration.8 A test based on time-
resolved measurements has been developed to investigate the
possible interference of the excited-state association with the
fluorimetric determination of Kd.8 This test indicates that, if an
inflection point occurs in the plot of the fluorescence signal
vs. Ϫlog [X] in the [X] range where both fluorescence decay
times are invariant, this inflection point can be associated with
the correct Kd. In contrast, inflection points in the [X] range
where the decay times vary cannot be attributed to Kd.
summarise the globally fitted decay times τL (L for long) and τS (S for
short). # indicates the number of decay traces that were analysed for
each concentration. The last two columns describe numerical criteria
for the quality of the fits [see eqns. (12) and (13)]
[Ca2ϩ]/M
τL/ns
τS/ns
#
Zχ
χg2
2
g
1 no added Ca2ϩ
2 6.11 × 10Ϫ6
3 4.41 × 10Ϫ5
4 5.73 × 10Ϫ4
5 4.07 × 10Ϫ2
6 1.14 × 10Ϫ1
7 4.86 × 10Ϫ1
8 7.29 × 10Ϫ1
3.25 0.01
3.25 0.01
3.23 0.01
3.24 0.01
3.32 0.01
3.33 0.01
3.27 0.01
3.30 0.01
0.19
0.16
0.15
0.31 0.05
0.12 0.04
0.81 0.04
0.66 0.03
0.30 0.01
6
6
0.54
0.42
0.16
3.06
0.19
2.23
1.40
2.91
1.01
1.01
1.00
1.07
1.00
1.06
1.04
1.07
6
8a
6
6
6
8a
a At these concentrations, two extra decay traces were recorded with
excitation at 300 nm and emission at 400 nm.
To apply this test, time-resolved fluorescence measurements
were performed on the system Thio-H–Ca2ϩ. Decay traces of
Thio-H at different [Ca2ϩ] ranging from 6.11 µM to 0.73 M were
collected at three different sets of excitation/emission wave-
0.01 ns and τS = 0.35 0.01 ns. They are represented graphically
in Fig. 4. As the decay times remain constant in the range
of [Ca2ϩ] used in the fluorimetric titrations, the inflection
point in the fluorimetric titrations can be associated with the
correct Kd.
lengths: (i) λex = 330 nm, λem = 520 nm; (ii) λex = 370 nm, λem
=
520 nm; (iii) λex = 330 nm, λem = 480 nm. Each sample was
measured twice under the same circumstances. The decay traces
of solutions with identical Ca2ϩ concentrations were analysed
globally as biexponential functions with the decay times linked
over the emission and the excitation wavelengths at each
[Ca2ϩ] (Table 4). The short decay time τS could not be estimated
accurately because its contribution to the total fluorescence is
small.30 The decay traces of all samples within the Ca2ϩ concen-
tration range used for the fluorimetric titration (up to 5.73 ×
10Ϫ4 M) were then analysed globally by linking the fluorescence
lifetimes over the different [Ca2ϩ], λex, and λem. Excellent fits
were obtained for this extended global biexponential analysis
(χ2 = 1.04, Zχ2 = 3.12). The recovered lifetimes are τL = 3.25
Conclusions
A new ratiometric indicator for Mg2ϩ and Ca2ϩ, Thio-H, has
been synthesised via a convergent route in which the key steps
consisted of Pd catalysed cross-coupling reactions. By modify-
ing the order of reactions, the synthesis of the ionophore syn-
thon could be simplified from an elaborate five-step literature
procedure to a more efficient two-step route.
Thio-H provides an excellent addition to the commercially
available fluorescent Ca2ϩ/Mg2ϩ indicators, because of its
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206
1203

View Online
Table 5 Comparison of spectroscopic properties and Kd (dissociation constant) values of Thio-H 1 and commercial fluorescent low-affinity Ca2ϩ
and Mg2ϩ indicators available from Molecular Probes7
a
c
Kd
ꢀf
ε (at λex)d
Indicator
(T in ЊC)
Ratiometric
modeb
Ca2ϩ/µM
Mg2ϩ/mM
Free
Ca2ϩ
Mg2ϩ
0.306
Free
Ca2ϩ
Mg2ϩ
Mag-fura-2 (22)
Mag-fura-5 (22)
Mag-indo-1 (22)
BTC (22)
25
28
35
7
1.9
2.3
2.7
—
Ex 340/380
Ex 340/380
Em 405/485
Ex 400/480
0.246
22 (369)e
23 (369)e
38 (349)e
29 (464)e
25 (364)e
31 (351)h
26 (329)f
25 (330)f
35 (328)f
20 (401)f
28 (335)f
31 (332)h
24 (330)g
25 (332)g
33 (330)g
—
0.127
0.50h
Fura-FF (22)
5.5
—
Ex 340/380
—
Thio-H (20)h
45h
5.6h
0.74h
0.65h
30 (335)h
h
a Kd values for the Ca2ϩ–indicator complex have been measured in vitro in 100 mM KCl, 10 mM MOPS pH 7.20. Kd values of the Mg2ϩ–indicator
complex have been measured in vitro in 115 mM KCl, 20 mM NaCl, 10 mM Tris pH 7.05, unless otherwise noted. b The indicated wavelengths are
given in nm. Ex = excitation ratiometric mode. Em = emission ratiometric mode. c Fluorescence quantum yield of the free indicator (zero [Ca2ϩ] and
[Mg2ϩ]) and of the Ca2ϩ and Mg2ϩ complexes. d Molar extinction coefficient × 10Ϫ3 MϪ1 cmϪ1 at the indicated excitation wavelength (in nm). e Free
form measured in aqueous buffer containing 10 mM EGTA. f Ca2ϩ bound form measured in aqueous buffer containing a > 10-fold excess of free
.
Ca2ϩ relative to the Kd. g Mg2ϩ bound form measured in aqueous buffer containing 35 mM Mg2ϩ h For the experimental conditions of the
measurements on Thio-H 1, see text.
Experimental
Materials and methods
When required, solvents and reagents were dried prior to use.
Tetrahydrofuran (THF) and toluene were distilled over sodium.
1H NMR spectra were recorded on a Bruker Avance 300 or a
Bruker AMX 400 instrument. The spectra were measured using
deuteriochloroform (CDCl3) or deuterated DMSO as solvent;
1
the H and 13C chemical shifts are reported in ppm relative
to tetramethylsilane or the deuterated solvent as an internal
reference; the coupling constants J are given in Hz. Mass
spectra were recorded using a Kratos MS50TC instrument and
a DS90 data system. IR spectra were taken on a Perkin Elmer
1720 Fourier transform spectrometer. Melting points were
determined with a Reichert Thermovar apparatus and are
uncorrected.
The absorption measurements were performed on a Perkin
Fig. 4 Decay times of Thio-H 1 as a function of Ϫlog [Ca2ϩ]. The
symbols ᭿ and ꢀ represent the decay times determined by global
biexponential analysis. The symbols marked with * correspond to the
samples without added Ca2ϩ. The solid line represents the globally
analysed decay times linked over the first four concentrations of
Table 4.
Elmer Lambda
6 UV/VIS spectrophotometer. Corrected
steady-state excitation and emission spectra were recorded on
a SPEX Fluorolog 212. Fluorescence quantum yields of the
free and bound forms of the indicators were determined using
quinine sulfate in 0.1 N sulfuric acid as reference. The fluor-
escence quantum yield of this reference was taken to be 0.50
0.02.31 The samples were measured in Milli-Q water. No correc-
tion for the refractive index was necessary. The fluorescence
decay traces were collected by the single-photon timing tech-
nique32,33 using the synchrotron radiation facility SUPER-ACO
(Anneau de Collision d’Orsay at LURE, France) as described
elsewhere.34 The storage ring provides vertically polarised light
pulses with a full width at half maximum of ≈ 500 ps at a
frequency of 8.33 MHz in double bunch mode. A Hamamatsu
microchannel plate R1564U-06 was utilised to detect the
fluorescence photons under magic angle (54Њ 44Ј). The instru-
ment response function was determined by measuring the
light scattered by glycogen in aqueous solution at the emission
wavelength. All decay traces were collected in 1K channels of a
multichannel analyser and contained approximately 2 × 103
peak counts. The time increment per channel was 30 ps.
interesting property of being both excitation and emission
ratioable. An extra advantage is its high fluorescence quantum
yield: Thio-H is about three times brighter than the commercial
low-affinity Ca2ϩ indicator Mag-fura-2.6 In Table 5, the spectro-
scopic properties and Kd values of Thio-H are compared to
those of the commercial low-affinity Ca2ϩ indicators.
The dissociation constant for Ca2ϩ is 45 µM at ionic strengths
near 0.1 M and pH 7.20; titration with Mg2ϩ yields a dissoci-
ation constant of 5.6 mM under the same experimental con-
ditions. Measurements under physiological conditions show
that binding of Ca2ϩ to Thio-H is somewhat less efficient at
elevated temperatures, while an increase in the pH causes the
opposite effect. In the presence of 1 mM Mg2ϩ the Ca2ϩ affinity
of Thio-H halves. Time-resolved fluorescence measurements
confirm that the inflection point in the fluorimetric titration
curve can be correctly assigned to Kd.
Since the dissociation constant for the Mg2ϩ–Thio-H com-
plex falls within the intracellular [Mg2ϩ] range, while the
Kd value for Ca2ϩ–Thio-H is well above the basal Ca2ϩ levels,
this fluorescent indicator has potential applications as a Mg2ϩ
indicator. The low affinity for Ca2ϩ can be exploited for the
determination of Ca2ϩ concentrations in the micromolar range,
on condition that [Mg2ϩ]i remains practically constant during
the measurements. In the future, it will be investigated whether
these promising in vitro properties will be confirmed in intra-
cellular environments.
Data analysis
The existing general global analysis program35 based on Mar-
quardt’s algorithm36 was used to obtain estimates of the decay
times τi (i = 1, 2) [eqn. (9)].
The fitting parameters were determined by minimising the
global reduced chi-square χ2g:
where the index l sums over q experiments, and the index i
(12)
1204
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206

View Online
sums over the appropriate channel limits for each individual
experiment. yoli and ylci denote respectively the experimentally
measured (observed) and fitted (calculated) values corre-
sponding to the ith channel of the lth experiment. wli is the
corresponding statistical weight. ν represents the number of
degrees of freedom for the entire multidimensional fluorescence
ation of the solvent, the yellow oil was chromatographed on
silica (CH2Cl2–AcOEt, 95 : 5), yielding a crystalline product
(3 g, 60 %); mp 58–60 ЊC; νmax(KBr)/cmϪ1 2960, 2912 (CH3,
CH2, CH), 1741 (CO), 1598, 1504 (Ar), 1143, 1060 (C-O-C),
710 (C-Br); δH(CDCl3) 3.68 (6 H, s, CH3(N)), 3.76 (3 H, s,
CH3(O)), 4.14 (4 H, s, CH2(N)), 4.62 (2 H, s, CH2(O)), 6.7 (1 H,
d, J 8.5, 3-H), 6.88 (1 H, d, J 2.1, 6-H), 7.01 (1 H, dd, J 8.5 and
J 2.1, 4-H); δC(CDCl3) (Scheme 2) 51.7–52 (CH3), 53.4
(CH2(N)), 66 (CH2(O)), 114.2 (5-C), 117.88 (6-C), 121.08 (3-C),
125.4 (4-C), 138.7 (2-C), 150.19 (C-1), 168.7 (CO(O)), 171.3
decay surface. χ2g and its corresponding Z [eqn. (13)] provide
χg2
numerical goodness-of-fit criteria for the entire fluorescence
decay surface.
(CO(N)); m/z 404 (MHϩ, 88%), 343 (Mϩ Ϫ CO2Me, Ϫ NH,
ؒ
(13)
ϩ
18), 324 (Mϩ Ϫ Br, 8), 147 (M Ϫ 3 × CO2Me, Ϫ Br, 16);
ؒ
ؒ
Found: 403.0285, Calc. for C15H18BrNO7: 403.0267.
2
Using Zχ the goodness-of-fit of analyses with different ν can
g
2-Phenylthiophene 9. To a mixture of iodobenzene 7 (4.9
mmol, 1 g) and 2-(tributylstannyl)thiophene 8 (4.9 mmol, 1.8 g)
in anhydrous DMF (30 mL), (MeCN)2PdCl2 (2 mol%, 9.8 ×
10Ϫ5 M, 25 mg) was added. After stirring overnight at room
temperature, water (100 mL) was added to the reaction mixture,
which was then extracted with chloroform (4 × 40 mL). The
combined organic layers were washed with water (2 × 20 mL)
and then dried over MgSO4. After evaporation, the residue was
dissolved in hexane. A black precipitate was formed, which was
filtered off. The filtrate was concentrated and the crude product
was purified by vacuum distillation (bp 96 ЊC at 10Ϫ2 atm). A
colourless oil was obtained, which recrystallised upon cooling
(0.55 g, 71%); mp 30.8–32.3 ЊC; νmax(KBr)/cmϪ1 2921, 2871
(CH), 1602 (Ar); δH(CDCl3) 7.04 (1 H, dd, J 3.9 and J 5.1,
4Ј-H), 7.22 (1 H, dd, J 1.2 and J 5.1, 5Ј-H), 7.25 (1 H, m, 4-H),
7.27 (1 H, dd, J 1.2 and J 3.9, 3Ј-H), 7.33 (2 H, t, J 6.3, 3-, 5-H),
7.58 (2 H, dd, J 3.3 and J 6.3, 2-, 6-H); δC(CDCl3) (Scheme 3)
123.0 (3Ј-C or 4Ј-C), 124.7 (3Ј-C or 4Ј-C), 125.9 (2-C, 6-C),
be readily compared. The additional criteria that were used to
judge the quality of the fits are described elsewhere.33 Standard
error estimates were obtained from the parameter covariance
matrix available from the analysis. All quoted errors are one
standard deviation.
Synthesis
Trimethyl 2-aminophenol-N,N,O-triacetate 3. To a solution
of 2-aminophenol 2 (92 mmol, 10 g) in anhydrous acetonitrile
(100 mL) were added sequentially “Proton Sponge” [1,8-
bis(dimethylamino)naphthalene] (183 mmol, 2 equivs., 39.7 g),
methyl bromoacetate (283 mmol, 3 equivs., 43.3 g) and sodium
iodide (0.02 equivs., 0.27 g), after which the mixture was heated
under argon to 70 ЊC. After heating for 24 h, methyl bromo-
acetate (3 equivs., 43.3 g) was added, whereupon the mixture was
heated at 70 ЊC for another 48 h. The reaction mixture was then
filtered on Celite and the obtained organic layer was dried over
MgSO4. After evaporation of the solvent, the crude product
was chromatographed on silica (CH2Cl2–AcOEt, 95 : 5), yield-
ing pale yellow crystals (15%); mp 75 ЊC; νmax(KBr)/cmϪ1 3025
(CH3, CH2, CH), 1736 (CO), 1609, 1567, 1502 (Ar); δH(CDCl3)
3.70 (6 H, s, CH3(N)), 3.78 (3 H, s, CH3(O)), 4.24 (4 H, s,
CH2(N)), 4.70 (2 H, s, CH2(O)), 6.85 (4 H, m, Ar); δC(CDCl3)
(Scheme 2) 51.7 (CH3(N)), 52.1 (CH3(O)), 53.6 (CH2(N)), 66.2
(CH2(O)), 114.9 (6-C), 119.9 (3-C), 122.6 (5-C), 122.7 (4-C),
127.4 (4-C or 5Ј-C), 127.9 (4-C or 5Ј-C), 128.8 (3-C, 5-C), 134.4
ϩ
(1-C), 144.4 (2Ј-C); m/z 160 (Mϩ , 100%), 128 (M Ϫ NS, 19),
ؒ
ؒ
115 (Mϩ Ϫ SCH, 49), 102 (M Ϫ SC2H2, 15), 45 (SCH , 53);
ϩ
ϩ
ؒ
ؒ
Found: 160.0344, Calc. for C10H8S: 160.0346.
2-Phenyl-5-(tributylstannyl)thiophene 10. In a three-necked
flask equipped with a septum and a stopcock, 2-phenylthio-
phene 9 (8.4 mmol, 1.3 g) was dissolved in anhydrous THF
(50 mL). After cooling to Ϫ78 ЊC, a solution of butyllithium
(1.2 equivs., 10 mmol) in hexane was added via a syringe. The
mixture was stirred at Ϫ78 ЊC under argon for 30 min, after
which tributyltin chloride (1.2 equivs., 10 mmol, 3.25 g) was
added dropwise via a syringe. The reaction mixture was stirred
at room temperature and then poured into glacial brine
(75 mL), which was washed with ether (3 × 75 mL). After
evaporation, the unreacted starting material was removed by
vacuum distillation, after which the residue was purified by
column chromatography (Al2O3; hexane). A yellow oil was
obtained (80%); νmax (KBr)/cmϪ1 2958–2853 (CH3, CH2, CH),
1599 (Ar), 1572 (Ar); δH(CDCl3) 0.91 (9 H, m, CH3), 1.12 (6 H,
m, CH2), 1.35 (6 H, m, CH2), 1.58 (6 H, m, CH2), 7.19 (1 H, d,
J 3.37, 3Ј-H or 4Ј-H), 7.30 (1 H, m, 4-H), 7.42 (2 H, t, J 8.63,
3-, 5-H), 7.47 (1 H, d, J 3.37, 3Ј-H or 4Ј-H), 7.67 (2 H, dd,
J 3.21 and J 8.63, 2-, 6-H); δC(CDCl3) (Scheme 3) 8.7 (CH2Sn),
13.58 (CH3), 28.95 (CH2), 30.11 (CH2), 125.04 (3Ј-C), 126.45
(2-C, 6-C), 127.4 (4-C), 129.02 (3-, 5-C), 134.8 (1-C), 142.9
139.5 (2-C), 149.7 (1-C), 169.4 (CO(O)), 171.8 (CO(N)); m/z
ϩ
325 (Mϩ , 24%), 266 (M Ϫ CO2Me, 100).
ؒ
ؒ
Besides compound 3, two cyclic side products were formed.
A lactone, methyl 2-(2-oxo-2,3-dihydro-4H-1,4-benzoxazin-4-
yl)acetate 4, was isolated in 23% yield; νmax(KBr)/cmϪ1 3018,
2952 (CH3, CH2, CH), 1778, 1739 (CO), 1613, 1591, 1507 (Ar);
δH(CDCl3) (Scheme 2) 3.76 (3 H, s, CH3), 4.68 and 4.69 (2 H, s,
CH2 and 2 H, s, 3-H), 6.85 (4 H, m, Ar); δC(CDCl3) 50.2 (CH3),
50.8 (CH2), 51.4 (3-C), 115.7 (8-C), 118.9 (5-C), 122.3 (7-C),
122.5 (6-C), 136.4 (4a-C), 145.6 (8a-C), 167.2 (2-C), 170.2
ϩ
(CO(O)); m/z 221 (Mϩ , 64%), 162 (M Ϫ CO2Me, 60).
Furthermore, a lactam, methyl 2-(3-oxo-2,3-dihydro-4H-1,4-
benzoxazin-4-yl)acetate 5, was isolated in 35% yield. νmax(KBr)/
cmϪ1 3013, 2956 (CH3, CH2, CH), 1742, (CO), 1687 (amide),
1606, 1597, 1507 (Ar); δH(CDCl3) (Scheme 2) 3.76 (3 H, s,
CH3), 4.00 (2 H, s, CH2), 4.09 (2 H, s, 2-H), 6.85 (4 H, m,
Ar); δC(CDCl3) 50.1 (CH3), 52.3 (CH2), 66.7 (2-C), 114.3 (8-C),
119.2 (5-C), 123.2 (7-C), 123.4 (6-C), 137.8 (4a-C), 147.7
ؒ
ؒ
(8a-C), 165.1 (3-C), 169.6 (CO(O)); m/z 221 (Mϩ , 64%), 162
ؒ
(2Ј-C), 146.0 (5Ј-C), 150.7 (4Ј-C); m/z: 450 (Mϩ , 5%), 393
ؒ
(Mϩ Ϫ CO2Me, 60).
ؒ
ϩ
ϩ
(Mϩ Ϫ Butyl, 100), 336 (M Ϫ 2 × Butyl, 54), 279 (M Ϫ 3 ×
Butyl, 52), 235 (Sn(Bu)2ϩ, 15), 177 (SnBuϩ, 14); Found:
450.1398, Calc. for C22H34SSn: 450.1403.
ؒ
ؒ
ؒ
Trimethyl 2-amino-5-bromophenol-N,N,O-triacetate 6.
A
mixture of trimethyl 2-aminophenol-N,N,O-triacetate 3 (12.4
mmol, 4.02 g) and N-bromosuccinimide (13.6 mmol, 2.42 g)
(freshly recrystallised from hexane) in acetic acid (25 mL) was
stirred at room temperature for 30 h. After evaporation of
acetic acid, water (100 mL) and ether (80 mL) were added to the
black residue. After separation of the two layers, the aqueous
solution was further extracted with ether (80 mL). The ether
fractions were combined and dried over MgSO4. After evapor-
Trimethyl ester of {2-[bis(carboxymethyl)amino]-5-(5-phenyl-
thiophen-2-yl)phenoxy}acetic acid 11. To a solution of 2-phenyl-
5-(tributylstannyl)thiophene 10 (1.2 equivs., 1.2 mmol, 0.54 g)
in anhydrous toluene (20 mL), trimethyl 2-amino-5-bromo-
phenol-N,N,O-triacetate 6 (1 mmol, 0.34 g) and tetrakis-
(triphenylphosphine)palladium(0) (1 mol%, 11 mg) were
added. The reaction mixture was refluxed for 15 h under argon
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206
1205

View Online
5 G. Grynkiewicz, M. Poenie and R. Y. Tsien, J. Biol. Chem., 1985,
260, 3440.
6 B. Raju, E. Murphy, L. A. Levy, R. D. Hall and R. E. London, Am.
J. Physiol., 1989, 256, C540.
7 R. P. Haugland, Handbook of Fluorescent Probes and Research
Products, Molecular Probes, Eugene, OR, USA, 8th edn. (CD-
version), 2001.
8 A. Kowalczyk, N. Boens, K. Meuwis and M. Ameloot, Anal.
Biochem., 1997, 245, 28.
9 A. Kowalczyk, N. Boens, V. Van den Bergh and F. C. De Schryver,
J. Phys. Chem., 1994, 98, 8585.
10 E. Cielen, A. Tahri, K. Ver Heyen, G. J. Hoornaert,
F. C. De Schryver and N. Boens, J. Chem. Soc., Perkin Trans. 2,
1998, 1573.
11 V. Van den Bergh, N. Boens, F. C. De Schryver, M. Ameloot,
P. Steels, J. Gallay, M. Vincent and A. Kowalczyk, Biophys. J., 1995,
68, 1110.
atmosphere. After evaporation of toluene, the product was
purified on alumina (EtOAc–hexane 30 : 70 with a gradient to
pure EtOAc) and recrystallised from chloroform by addition of
cold hexane. A yellow powder was obtained (66%); mp 81–83
ЊC; νmax(KBr)/cmϪ1 2955, 2924 (CH3, CH2), 2855 (CH), 1728
(CO), 1596, 1569, 1492 (Ar); δH(CDCl3) (Scheme 3) 3.73 (6 H,
s, CH3(N)), 3.80 (3 H, s, CH3(O)), 4.24 (4 H, s, CH2(N)), 4.72
(2 H, s, CH2(O)), 6.89 (1 H, d, J 8.32, 3-H), 7.06 (1 H, d, J 1.88,
6-H), 7.14 (1 H, d, J 3.7, 3Ј-H or 4Ј-H), 7.20 (1 H, dd, J 1.88 and
J 8.3, 4-H), 7.24 (1 H, d, J 3.7, 3Ј-H or 4Ј-H), 7.23–7.27 (1 H,
m, 9Ј-H), 7.37 (2 H, t, J 7.46, 8Ј-, 10Ј-H), 7.60 (2 H, d, J 7.46,
7Ј-, 11Ј-H); δC(CDCl3) 51.8 (CH3(N)), 52.1 (CH3(O)), 53.5
(CH2(N)), 66.2 (CH2(O)), 112.4 (6-C), 119.8 (3-C), 120.1 (4-C),
123.8 and 123.9 (3Ј-, 4Ј-C), 125.4 (7Ј-, 11Ј-C), 127.3 (9Ј-C),
128.8 (8Ј-, 10Ј-, 5-C), 134.2 (2-C), 139.0 (6Ј-C), 142.8 and 143.1
(2Ј-, 5Ј-C), 149.5 (1-C), 169.1 (CO(O)), 171.5 (CO(N)); m/z 483
12 J. B. Birks, Photophysics of Aromatic Molecules, Wiley, New York,
1970.
13 G. Kirsch, D. Prim, F. Leising and G. Mignani, J. Heterocycl. Chem.,
1994, 31, 1005.
ϩ
ϩ
(Mϩ , 94%), 424 (M Ϫ CO2Me, 100), 351 (M Ϫ 2 × CO2Me,
ؒ
ؒ
ؒ
Ϫ CH2, 24), 278 (Mϩ Ϫ 3 × CO2Me, Ϫ 2 × CH2, 18), 235 (M
ϩ
ؒ
ؒ
14 L. A. Levy, E. Murphy, B. Raju and R. E. London, Biochemistry,
1988, 27, 4041.
15 O. Reany, T. Gunnlaugsson and D. Parker, J. Chem. Soc., Perkin
Trans. 2, 2000, 1819; O. Reany, T. Gunnlaugsson and D. Parker,
Chem. Commun., 2000, 473.
16 H. H. Hodgson and F. H. Moore, J. Chem. Soc., 1926, 155;
C. Wright, M. Shulkind, K. Jones and M. Thompson, Tetrahedron
Lett., 1987, 28, 6389; R. P. Hanzlik, P. E. Weller, J. Desai,
J. Zheng, L. R. Hall and D. E. Slaughter, J. Org. Chem., 1990,
55, 2736.
17 F. D. Bellamy and K. Ou, Tetrahedron Lett., 1984, 25, 839.
18 H. Gershon, D. D. Clarke and M. Gershon, Monatsh. Chem., 1993,
124, 367.
Ϫ N(CH2CO2CH3)2, Ϫ OCH2CO2CH3, 11); Found: 483.1361,
Calc. for C25H25NO7S: 483.1352.
Tricaesium salt Thio-H 1. This was prepared from compound
11 according to the procedure described by Minta and Tsien.21
A solution of the ester (1.03 × 10Ϫ5 mol) and anhydrous
caesium hydroxide (1.03 × 10Ϫ4 mol, 10 equivs.) in methanol
(3 cm3) was refluxed overnight. After evaporation of the
methanol, the product was dissolved in water (100 cm3) and
used as such for the fluorescence measurements.
Acknowledgements
19 N. A. Bumagin and I. G. Bumagina, Dokl. Akad. Nauk SSSR, 1983,
274(4), 2547.
The authors are indebted to Professor Dr S. Toppet and R. De
Boer for the NMR and mass spectra, respectively. K. Van
Werde is thanked for assistance with the time-resolved
fluorescence measurements. E.C. is an Aspirant and N.B. is
an Onderzoeksdirecteur of the Fonds voor Wetenschappelijk
Onderzoek (FWO). A.T. is a postdoctoral fellow at the
K. U. Leuven. A.S. thanks the Flemish Ministry of Science and
Technology for a postdoctoral fellowship through the Bilateral
Scientific and Technological Cooperation Programme (BIL 97/
09). The continuing financial support of DWTC through
IUAP-4-11 is gratefully acknowledged.
20 S. Gronowitz, A.-B. Hörnfeldt and Y. Yang, Chem. Scr., 1988, 28,
281.
21 A. Minta and R. Y. Tsien, J. Biol. Chem., 1989, 264, 8171.
22 A. Fabiato and F. Fabiato, J. Physiol. (Paris), 1979, 75, 219.
23 R. Y. Tsien, Biochemistry, 1980, 19, 2396.
24 R. Y. Tsien, in Fluorescent Chemosensors for Ion and Molecule
Recognition, A. W. Czarnik, Ed., ACS Symposium Series,
Washington, 1993, p. 538.
25 S. M. Harrison and D. M. Bers, Biochim. Biophys. Acta, 1987, 925,
133.
26 P. W. Marks and F. R. Maxfield, Anal. Biochem., 1991, 193, 61.
27 T. J. Shuttleworth and J. L. Thompson, J. Biol. Chem., 1991, 266,
1411.
28 R. Pethig, M. Kuhn, R. Payne, E. Adler, T. H. Chen and L. F. Jaffe,
Cell Calcium, 1989, 10, 491.
29 H. Iatridou, E. Foukaraki, M. A. Kuhn, E. M. Marcus,
R. P. Haugland and H. E. Katerinopoulos, Cell Calcium, 1994, 15,
190.
References
1 M. J. Berridge, P. Lipp and M. D. Bootman, Nature Rev. Mol. Cell
Biol., 2000, 1, 11.
2 R. Y. Tsien, Annu. Rev. Neurosci., 1989, 12, 227; W. J. J. M.
Scheenen, A. M. Hofer and T. Pozzan, in Cell Biology: A Laboratory
Handbook, 2nd Edn., J. E. Celis, Ed., Vol. 3, Academic Press, 1998,
p. 363; A. Takahashi, P. Camacho, J. D. Lechleiter and B. Herman,
Physiol. Rev., 1999, 79, 1089.
3 A. M. Hofer and T. E. Machen, Proc. Natl. Acad. Sci. USA, 1993,
90, 2598; V. A. Golovina and M. P. Blaustein, Science, 1997, 275,
1643; G. Csordás, A. P. Thomas and G. Hajnóczky, EMBO J., 1999,
18, 96; J. Meldolesi and T. Pozzan, Trends Biochem. Sci., 1998, 23,
10.
4 A. L. Escobar, J. R. Monck, J. M. Fernandez and J. L. Vergara,
Nature, 1994, 367, 739; A. M. Hofer, B. Landolfi, L. Debellis,
T. Pozzan and S. Curci, EMBO J., 1998, 17, 1986; A. K. Stout and
I. J. Reynolds, Neuroscience, 1999, 89, 91.
30 L. D. Janssens, N. Boens, M. Ameloot and F. C. De Schryver,
J. Phys. Chem., 1990, 94, 3564.
31 W. R. Dawson and M. W. Windsor, J. Phys. Chem., 1968, 72,
3251.
32 D. V. O’Connor and D. Phillips, Time-Correlated Single Photon
Counting, Academic Press, New York, 1984.
33 N. Boens, Luminescence Techniques in Chemical and Biochemical
Analysis, Eds. W. R. G. Baeyens, D. De Keukeleire and K. Korkidis,
Marcel Dekker, New York, 1991, p. 21.
34 O. P. Kuipers, M. Vincent, J. C. Brochon, H. M. Verheij,
G. H. De Haas and J. Gallay, Biochemistry, 1991, 30, 8771.
35 N. Boens, L. D. Janssens and F. C. De Schryver, Biophys. Chem.,
1989, 33, 77.
36 D. W. Marquardt, J. Soc. Ind. Appl. Math., 1963, 11, 431.
1206
J. Chem. Soc., Perkin Trans. 2, 2002, 1197–1206