Open Access Article
Gergő Riszterabc,
László Forgách
def,
Zoltán Mucsigh,
Laura Molnárb,
Dóra Bogdána,
Fatemeh Heydari
de,
Domokos Máthé
dei,
István Mándityacj,
Ágnes Gömöryk and
Zoltán Kaleta
*acj
aInstitute of Organic Chemistry, Semmelweis University, Hőgyes Endre Street 7, H-1092 Budapest, Hungary. E-mail: kaleta.zoltan@semmelweis.hu
bPROGRESSIO Engineering Bureau Ltd, Muhar u. 54, H-1028 Budapest, Hungary
cArtificial Transporters Research Group, Institute of Materials and Environmental Chemistry, Research Centre for Natural Sciences, HUN-REN Hungarian Research Network, Magyar Tudósok Körútja 2, Budapest, H-1117, Hungary
dDepartment of Biophysics, Semmelweis University, Tűzoltó utca 37-47, Budapest, 1094, Hungary
eIn Vivo Imaging Advanced Core Facility, Hungarian Center of Excellence for Molecular Medicine (HCEMM), Tűzoltó Utca 37-47, Budapest, 1094, Hungary
fDepartment of Nuclear Medicine, Medical Imaging Centre, Semmelweis University, Üllői út 78b, Budapest 1083, Hungary
gBrainVisionCenter, 1094 43-45. Liliom Str., Budapest, Hungary
hInstitute of Chemistry, University of Miskolc 3515 Miskolc-Egyetemváros, A/2 Building, Hungary
iCROmed Translational Research Centers, 37-47 Tűzoltó Street, Budapest, 1094, Hungary
jCentre for Pharmacology and Drug Research & Development, Semmelweis University, Budapest, Hungary
kMS Proteomics Research Group, Institute of Organic Chemistry, Research Centre for Natural Sciences, HUN-REN Hungarian Research Network, Magyar Tudósok Körútja 2, Budapest, H-1117, Hungary
First published on 1st June 2026
An effective, near-infrared (NIR) fluorescent Ca2+ sensor, Solomon-red-BAPTA, featuring a dioxothioxanthene moiety was developed for microscopy. A rapid and operationally simple synthetic route is described. The key step involves a room-temperature DDQ-mediated cross-dehydrogenative coupling (CDC) of the fluorescent precursor and BAPTA followed by a one-pot oxidation. The sensor exhibited excellent photophysical characteristics including a 160-fold fluorescence enhancement with excitation and emission maxima at 712 nm and 736 nm, respectively, suggesting high tissue-penetrating capacity. Based on the experimental data, the structure–property relationship was elucidated and further supported by DFT calculations, particularly with respect to the photoinduced electron transfer (PET)-based fluorescence turn-on mechanism.
As a rule of thumb, tissue penetration by the emitted light increases with wavelength.9 To visualize a plethora of biological processes deep inside the body, fluorescent two- and three-dimensional microscopic and mesoscopic in vivo imaging have gained prominence in other research areas as well, such as cancer and immune biology and stem cell studies. Nowadays, these fluorescent imaging applications are entering clinical practice, with a number of fluorescent dyes in large-scale clinical surgery trials around the world.10 Consequently, there is a high demand for the development and application of fluorescent dyes emitting in the far-red or near-infrared (NIR-I spectral window: 650–950 nm) regions.11,12 Numerous such fluorescent dyes have been synthesized and reported in the literature.13–18 However, the photostability of these dyes remains a frequently cited limitation that requires further improvement.19–22
Rhodamine fluorophores hold an outstanding position thanks to their excellent properties, such as good photostability, high molar extinction coefficients and quantum yields (QYs), resulting in high fluorescence brightness.22–24 By extending the conjugation of the π-aromatic framework or replacing the central oxygen atom, the initial excitation/emission maxima have been shifted from below 600 nm up to around 700 nm.25 Liu and co-workers recently demonstrated (summarized in Fig. 1) that dioxothioxanthene derivatives are very promising as NIR fluorescent dyes for bioimaging.26,27
The exchange of the central heteroatoms in the xanthene framework affects not only the spectroscopic characteristics, but also the aromaticity of the individual condensed rings A and B (Fig. 1).
The more electron-deficient the heteroatom, the more antiaromatic the central ring B becomes, as measured by nucleus-independent chemical shift (NICS) calculations.28–30
Despite its high brightness and photostability, the application of a dioxothioxanthene derivative dye is cumbersome and scarcely reported, mainly due to its low-yielding synthesis.31 This synthetic bottleneck has now been overcome, opening the way for broader exploration of these promising NIR fluorophores in practical imaging applications.
As shown in path “A” in Fig. 2, carbon 9 of the xanthene (anthracene) moiety originates from the coupling partner. If the coupling partner in the condensation is an acid or ester, no further oxidation is required;32,33 if the carbon is a carbonyl group, further oxidation is necessary.7
In path “B”, the xanthene part and a phenyl moiety are the two used synthons. According to this approach, the coupling is done with substituted phenyl-lithium and the appropriate oxo-compound.34,35 This synthesis requires using low temperatures and butyl-lithium and sometimes results in low yields. The use of Suzuki-coupling also follows path “B”. However, the Suzuki-coupling route, while enabling the synthesis of otherwise inaccessible compounds, and often requiring hardly accessible arylboroxins, is contingent upon extremely dry conditions. This strict requirement stems from the high reactivity of the triflated xanthone intermediates, which rapidly react with water or alcohols to regenerate the parent chalcogenoxanthones.36 The triflated xanthones are also the starting material for pyronin derivative synthesis.37 This derivatisation shows that position “9” of the xanthene ring system is easily attacked by nucleophiles. Both main synthetic approaches suffer from moderate or even low yields and require special reaction conditions.
Our goal was to conduct extensive research on the nucleophilic addition to the xanthene ring system and prepare a novel NIR calcium sensor for microscopic imaging, the Solomon-red-BAPTA (1).38 The origin of the synthesis was described as early as 1913.39 The oxidative coupling of xanthene derivatives has been extensively studied during the past decade using various coupling40 partners such as aldehydes,41–43 anisoles,44 heteroatoms,45,46 ketones,47–49 nitroalkanes.50 There are two possible ways to reach the active intermediate pyronin: either through the mild oxidation of the xanthene CH2 group or through the reduction of the 9-oxo derivative. Since the latter proceeds very sluggishly,27 we tried to find an appropriately mild and straightforward condition for oxidizing and coupling.
We synthesized the sulfone-pyronin-core (3) based on an already described convenient and scalable way from 2, as shown in Scheme 1.26,51
In the next step, as a model reaction, we studied the coupling between sulfone-pyronin (3) and N,N-dimethylaniline (5a), leading to compound 7a in one pot, as shown in Scheme 2.
![]() | ||
| Scheme 2 Coupling between sulfone-pyronin (3) and N,N-dimethylaniline (5a). Calculated enthalpy values are indicated in kJ mol−1 [B3LYP/6-31G(d,p)//PCM(tol)]. | ||
In order to find an optimal condition, various solvents and oxidizing agents were screened and the results are summarized in Table 1.
| Entry | Solvent | Oxidising agent | Yielda |
|---|---|---|---|
| a Yields of 7a are determined by HPLC using toluene as an internal standard.b Isolated yields of 7a. | |||
| 1 | AcOH | DDQ | 67%b |
| 2 | NMP | DDQ | 39% |
| 3 | DMF | DDQ | 51% |
| 4 | DMSO | DDQ | 50% |
| 5 | DCM | DDQ | 52% |
| 6 | MeOH | DDQ | 29% |
| 7 | EtOH | DDQ | 51% |
| 8 | Toluene | DDQ | 25% |
| 9 | HCOOH | DDQ | 15% |
| 10 | TFA | DDQ | 1% |
| 11 | ACN | DDQ | 19% |
| 12 | Pyridine | DDQ | 16% |
| 13 | AcOH/H2O 1/1 | DDQ | 40% |
| 14 | AcOH | (NH4)2S2O8 | 0% |
| 15 | AcOH | K2S2O8 | 0% |
| 16 | AcOH | H2O2 | 0% |
| 17 | AcOH | Na2CO3·1.5H2O2 | 0% |
| 18 | AcOH | Perurea | 0% |
The best result was obtained using glacial acetic acid and DDQ (Entry 1). Surprisingly DDQ and TFA or formic acid (Entry 10 and 9, respectively) gave very low yields, on the other hand aqueous 50% acetic acid gave a fair yield (Entry 13). Both a rapid colour-change and HPLC measurements showed that the oxidation of sulfone-pyronin (3) is completed in less than 10 min to yield the fluorescent intermediate (4).27 To avoid the side reactions between the oxidizing agent and the coupling anilinic partner, 5a was added after an initial period of time (ca. 5–10 min). The theoretical study [B3LYP/6-31G(d,p)//PCM(tol), Fig. 4] confirms that the coupling reaction (4 → 6a) exhibits a very smooth transition state [TS(4 → 6a), 51.2 kJ mol−1] and is followed by an exothermic deprotonation to 7a, using 4 as a reference point.
To evaluate the substrate scope and the general feasibility of the transformation, various nucleophilic coupling partners were investigated under the optimized conditions (Table 2).
| Entry | Coupling partner | Yielda |
|---|---|---|
| a isolated yields. | ||
| 1 | Aniline (5b) | 67% |
| 2 | N,N-Dimethylaniline (5a) | 39% |
| 3 | N,N-Diethylaniline (5c) | 51% |
| 4 | N,N-Dipropylaniline (5d) | 50% |
| 5 | N,N-Dibutylaniline (5e) | 52% |
| 6 | Pyridine (5f) | — |
| 7 | Benzimidazole (5g) | — |
| 8 | Indole (5h) | 80% |
| 9 | Benzidine (5i) | — |
| 10 | Anthranilic acid (5j) | 50% |
| 11 | Salicylic acid (5k) | — |
| 12 | 3-Hydroxy-2-naphthoic acid (5l) | — |
| 13 | 3-Amino-2-naphthoic acid (5m) | — |
| 14 | Pyrogallol (5n) | 75% |
| 15 | o-Phenylenediamine (5o) | 70% |
| 16 | 2-Aminophenol (5p) | 30% |
| 17 | 4-Aminophenol (5q) | — |
| 18 | p-Phenylenediamine (5r) | — |
| 19 | 3-Aminophenol (5s) | 43% |
In contrast to most amines, benzidine (5i) failed to react under these conditions. Regarding carboxylic acid derivatives, anthranilic acid (5j) gave a fair yield, whereas salicylic acid (5k), 3-hydroxy-2-naphthoic acid (5l), and 3-amino-2-naphtholic acid (5m) resulted in no product formation. The coupling resulted in good yields in case of aromatic compounds with multiple electron-donating groups: pyrogallol (5n), o-phenylenediamine (5o) and 2-aminophenol (5p) Notably, while 4-aminophenol (5q) and p-phenylenediamine (5r) did not yield ortho-adducts, 3-aminophenol (5s) exhibited unique reactivity. In this case, the reaction occurred at both the 4- and 6-positions, providing a diadduct in reasonable yield, demonstrating that ortho-substitution is possible under specific electronic configurations.
Aniline derivatives (5a–e) afforded the corresponding products in good to excellent yields, even with relatively short reaction times. While pyridine (5f) and benzimidazole (5g) proved to be unreactive, indole (5h) underwent coupling in high yield (80%).
Consequently, two primary conclusions can be drawn: first, the presence of an electron-donating group is essential for the reaction to proceed, and second, ortho-adducts are generally suppressed, unless favored by the substitution pattern of the nucleophile, as seen with 5s.
After the successful optimization and demonstration of the feasibility of DDQ-mediated cross-dehydrogenative coupling (CDC), we turned our attention to our main goal, namely, coupling with BAPTA esters (5t, 5u). Since this chelating compound is also an aniline derivative, we were expecting a successful reaction, however, an additional hydroxy substituent was required for a successful reaction as described by Minta and coworkers.52 The reaction time proved to be longer (2 h at RT) than that of unsubstituted anilines. The resulting non-fluorescent Et-BAPTA-sulfone-pyronin (7u) can be oxidized to fluorescent 8u either in a consecutive step after purification or in a one-pot procedure by adding an additional equivalent of DDQ to the reaction mixture as shown in Scheme 3.
Yields for the two-step and the one-pot methods are 37% and 25%, respectively. The synthesized ethyl ester was hydrolysed under conventional basic conditions (aq. LiOH) to afford the BAPTA-based sensor 1.
Following isolation, the photophysical properties were characterized, revealing a molar extinction coefficient t (ε = 6760 M−1 cm−1, apo-state, see Fig. SI[X]) and a respectable quantum yield (:φ ≈ 0.25) upon Ca2+ complexation. While the free sensor 1 is virtually non-fluorescent in aqueous media, it displays a dramatic “turn-on” response in the presence of Ca2+, Cd2+ and Hg2+ (Fig. 3). Due to this high signal-to-background ratio and striking visual brightness, sensor 1 was designated as Solomon-red-BAPTA.
![]() | ||
| Fig. 3 Solomon-red-BAPTA (1) under visible and NIR light: control, Co2+, Zn2+, Sr2+, Cd2+, Mn2+, Hg2+, Cu2+, Fe2+, Pb2+, Mg2+, Ni2+, Ca2+, and Ba2+. | ||
The selectivity of the probe was first screened against a series of biologically relevant metal ions. As shown in Fig. 3, cations such as Mg2+, Fe2+, or Sr2+ elicited negligible fluorescence responses, confirming the high specificity of the BAPTA scaffold within this chemical environment. These experimental findings are in excellent agreement with our computational models, which suggest that the hybrid system is optimally configured to favor the coordination of Ca2+ and its electronic analogues over smaller and harder cations.
Additional properties of 1 were investigated to highlight and confirm its applicability. The absorption maximum was at 712 nm, while the emission maximum was at 736 nm, providing enough Stokes shift (24 nm) to significantly separate the absorption and emission spectra and minimize light scattering at the detection side. The growth of the emission maxima was as high as 160, 140 and 62 folds in the presence of Ca2+, Cd2+ and Zn2+, respectively. The fluorescent spectra did not change under various Ca2+ concentrations, as shown in Fig. 4, right. The calculated apparent binding constant
for Ca2+ is
(ca. 2 µM) as determined by a competitive fluorescent titration (Fig. 5). The dynamic range of sensor 1 is 0.3–10 µM, which is similar to other reported sensors. We have determined a relatively high water solubility of >10 mg l−1 for sensor 1, the estimated concentration exceeds the probe concentration needed in biology in general.
To further evaluate the analytical robustness of Solomon-red-BAPTA, competitive binding assays were performed (Fig. 6). At the investigated concentrations, the presence of Zn2+, Cd2+ or Hg2+ alone resulted in well-defined, discrete fluorescence responses (0.8–1.4 fold relative to the baseline). In the specific case of Hg2+, a slight attenuation of the background signal was observed. This phenomenon is likely attributable to a subtle heavy-atom effect, which is readily overcome by the strong chelation-enhanced fluorescence (CHEF) in the presence of Ca2+. In multi-ion systems containing Ca2+, a sustained and significant enhancement (up to 7.5-fold) was observed regardless of the presence of other divalent cations, demonstrating that the probe remains highly responsive even in complex ionic environments.
The computed photophysical properties, including absorption and emission wavelengths and oscillator strengths, of the free and complexed forms of 1 were calculated at the TD-B3LYP/6-31G(d,p)[PCM(water)]53,54 level of theory using Gaussian 16.55 The experimentally observed fluorescence quenching of the free sensor may be attributed to the photoinduced electron transfer (PET) mechanism. Here, the predicted electronic excitation of the most stable conformers of the uncomplexed form of 1 (with 2 Na+ ions) can be assigned to the HOMO − 1 (216) to LUMO (218) orbitals, as shown in Fig. 7, left. The two Na+ ions are included to ensure the overall charge neutrality during the complexation process with Ca2+. In this structure, the HOMO (217) is localized mainly on the electron-rich aromatic moiety of the BAPTA unit, spatially separated from the fluorophore, whereas the HOMO − 1 (216) corresponds predominantly to the chromophoric π-system. Following excitation, electron transfer from the energetically accessible BAPTA-centered HOMO (217) to the partially depleted chromophore-centered orbital can occur efficiently, providing a non-radiative relaxation pathway that competes with fluorescence emission. Here, the electron from HOMO (217) can occupy the vacant position of HOMO − 1 (216), preventing the radiative deexcitation of the electron from LUMO (218) and forcing non-radiative relaxations. This PET process therefore results in strong fluorescence quenching of the free sensor, which is similar to that of related BAPTA-based fluorescent probes.56–58
Upon complexation with Ca2+, substantial changes in the frontier molecular orbital energies and distributions were observed. The coordination of Ca2+ significantly stabilizes the electron-donating orbitals associated with the BAPTA unit, lowering their energies and thereby suppressing the thermodynamic driving force of the PET process. In the Ca2+-bound form of 1, the photochemical process represents a more complex situation, where the electronic excitation occurs from the mixed HOMO (216)/HOMO − 1 (215) to LUMO (217) (see Fig. 7, right). As a consequence, the PET-mediated non-radiative deactivation pathway becomes energetically unfavorable, allowing radiative decay to dominate and thereby restoring strong fluorescence emission. These computational findings are in good agreement with the experimentally observed fluorescence turn-on response upon Ca2+ binding.
Fig. 8 represents the biologically relevant free Ca2+ concentration ranges, spanning from resting cytosolic nM levels to extracellular and vesicular mM Ca2+ pools, together with several well-known fluorescent sensor molecules. The experimentally determined Ca2+
of Solomon-red-BAPTA is indicated at approximately 1.7–2.0 µM, placing it in the biologically very relevant medium-affinity region and showing that it is better suited for detecting pronounced, e.g., epileptic-type intracellular Ca2+ elevations rather than small near-resting Ca2+ fluctuations.
| This journal is © The Royal Society of Chemistry 2026 |