BODIPY-containing aldehyde 7 was prepared according to a procedure reported in the literature (Xochitiotzi-Flores et al., 2016). After this, the next step consisted of the synthesis of the key α-isocyanoacetamide 9 to perform the Ugi-Zhu multicomponent reaction. This latter one was prepared also according to a method reported in the literature (Fayol et al., 2005) but with a slight modification. Once these key reagents were obtained, we continued with the synthesis of the target polyheterocyclic scaffold containing the BODIPY moiety.

Due to the novel characteristics of the aldehyde that was utilized in the preparation of the compounds, it was necessary to carry out a screening of the conditions for the MCR to obtain acceptable yields using a reaction model. To do this, the meso-(5-((4-formylphenyl)ethynyl)phenyl)-4,4-difluoro-4-bora-3a,4a-diaza-s-indacene (7), benzylamine (8a), 2-isocyano-1-morpholino-3-phenylpropan-1-one (9) and maleic anhydride (10) were combined sequentially to synthesize the corresponding BODIPY-pyrrolo[3,4-b]pyridin-5-one 11a, according to Scheme 1.

Table 1 shows the reaction conditions that were evaluated. First, the reaction conditions originally reported by Zhu (Janvier et al., 2002) were assayed (entry 1). However, even after 24 h of stirring, there was no change in the progress of the reaction. Microwave heating (entry 2) and the addition of a weak Brønsted acid (entry 3) were also unsuccessful to promote the reaction. Next, the non-polar solvent toluene (PhMe) was used (entry 4), in which the solubility of the aldehyde was improved, but the reaction still did not proceed at room temperature, even when microwave heating was applied (entry 5). A catalytic amount of scandium(III) triflate was utilized next (entry 6), as it has proved successful at promoting this reaction in previous reports of our research group (Zamudio-Medina et al., 2010; Vázquez-Vera et al., 2017; Zamudio-Medina et al., 2018; Segura-Olvera et al., 2019) and others (Rentería-Gómez et al., 2019; Pharande et al., 2020; Longo Jr et al., 2021). As expected, the addition of this Lewis’ acid promoted the reaction even at room temperature, albeit with a prolonged reaction time. The yield increased slightly when using microwave heating (entry 7). A two-fold increase of the catalyst loading improved the yield substantially (entry 8). A higher temperature (entry 9), however, did not improve the yield. Other Lewis’ acids were employed, namely, ytterbium(III) triflate (entry 10), which produced a lower yield than scandium(III) triflate, and indium(III) chloride (entry 11), which failed to promote the reaction at all. These Lewis’ acids were selected because they are considered as hard acids (small ionic radii, high oxidation state), they are stable in the presence of water and air, and, with the exception of indium(III) chloride, they have a weak coordinating counterion (OTf) which will not compete with the imine for coordination with the cation. Furthermore, the latter may be the reason why indium(III) chloride failed to promote the reaction.

Having established scandium(III) triflate as the best catalyst for the MCR, we revisited the use of MeOH as a solvent in the presence of this catalyst at RT and at 60°C. At RT (entry 12) the reaction did not proceed, and product formation was not observed. At 60°C (entry 13) the reaction did occur and 11a was obtained in 8% yield. This result was attributed to the reduced solubility of 7 in MeOH and possibly to the ring opening and esterification of maleic anhydride in the presence of MeOH.

Once the optimal reaction conditions were established (PhMe as the solvent, MW as heat source, at 60°C, and using a loading of 10 mol% of Sc(III) as the catalyst), the synthesis of the compound library 11a-g was carried out using the BODIPY-functionalized aldehyde 7, seven alkyl amines 8a-g, the α-isocyanoacetamide 9 and maleic anhydride (10) (Figure 2). Similar yields were obtained for all compounds which ranged from 20% to 46%. These yields are acceptable considering the structural complexity of the products, the number of reaction steps that occur and the remarkable atom economy (AE > 90% for all the cases) because only a couple of molecules of water and one molecule of carbon dioxide were released per synthesized compound. The reason for the moderate yields was due to incomplete consumption of the starting aldehyde and imine intermediate even under the optimized reaction conditions. Complete conversion of 5-aminooxazole to pyrrolo[3,4-b]pyridine-5-one was observed in all cases, according to the TLC monitoring. Formation of byproducts was not observed during the course of the reactions, even when using 3,4-dimethoxyphenethylamine (11f), which could potentially give the Pictet-Spengler product. This is in agreement with the original report by J. Zhu (Janvier et al., 2002) in which it is stated that the intermolecular reaction between imine and isocyanide is faster than the intramolecular trapping to form the corresponding tetrahydroisoquinoline. The overall atom economy of the process is remarkable considering that five new bonds were formed, even if the yields were low. However, it is important to note that atom economy is not an optimal parameter to determine whether a chemical reaction is efficient or not since it fails to take into account the incomplete conversion of reactants into products or formation of byproducts. The structure of all synthesized compounds was elucidated by NMR experiments (¹H, ¹³C, ¹¹B and ¹⁹F) and high-resolution mass spectrometry (HRMS). Assignments for all protons and carbons were performed by 2D NMR (COSY, HSQC and HMBC). See the Supplementary Material for further details.

The absorption and emission spectra of all compounds were acquired using dimethyl sulfoxide (DMSO) as the solvent, at a concentration of 10^–6 M and at room temperature. These spectra are detailed in Figure 3. Additionally, Table 2 summarizes the optical properties of each compound. All studied compounds exhibit absorption maxima at around 505 nm, and a second absorption band at around 390 nm, corresponding to the S₀→S₁ and S₀→S₂ transitions of the BODIPY core, respectively (Rojas-Montoya, et al., 2024). Additionally, there is an absorption band near 285 nm attributed to the pyrrolo[3,4-b]pyridin-5-one core (Flores-Reyes, et al., 2024). Compound 11e, which features a p-fluorobenzyl group in its structure, shows a bathochromic shift of approximately 5 nm for each of these bands. It is important to highlight that the substituent present in the pyrrolo[3,4-b]pyridin-5-one core influences its optical properties. For example, in the case of compound 11b, the aliphatic chain caused a decrease in its molar absorptivity coefficient (ε).

In the emission spectra, two bands are generally observed that correspond to each chromophore. The characteristic emission of the meso-phenyl BODIPY core is seen between 528 and 531 nm, while the emission of the pyrrolo[3,4-b]pyridin-5-one core appears between 438 and 450 nm, with a more pronounced hypsochromic shift in compound 11e. To support this statement, we acquired the absorption and emission spectra of an unsubstituted analog of 11a (12) in DMSO and compared them to the spectra of 7. The emission spectra of the dyads correspond to the overlapping of the emission spectra of both moieties. The absorbance spectra of dyads 11a-g show a more intense band between 250 and 300 nm than the one shown for compound 7, which corresponds to the pyrrolo[3,4-b]pyridin-5-one’s absorption. Emission of BODIPY 7 is shown in Supplementary Figure S40. The relative intensity of these bands varies among the different compounds. Generally, the emission from the BODIPY core is stronger. However, for compound 11f, the intensity of these two bands is similar. Notably, compound 11e shows a higher emission intensity of the pyrrolo[3,4-b]pyridin-5-one band. This dual emission centered into two different fluorophores allows organic compounds to cover a wide range of the visible spectrum and has awakened interest for designing white organic LEDs (Lee et al., 2020). Some examples of dual emitting organic materials containing the BODIPY core are the BODIPY-(1,8-naphtilamide) dyads separated by an oxoaryl bridge which showed the corresponding emissions for each fluorophore (Murkherjee and Thilagar, 2014), and phenyleneethynylene-BODIPY oligomers, which showed dual emission only when the oligomer chain was long enough to achieve an energy transfer quench (Reyes-Flores et al., 2021).

It is worth mentioning that the excitation spectra of compounds 11a-g showed a contribution of the pyrrolo[3,4-b]pyridin-5-one absorption towards the BODIPY emission (Supplementary Figure S41, measured at 540 nm). While BODIPY 7 exhibits absorption around 320 nm, dyads exhibit a band near 290 nm, this agrees with the computational studies shown below where the lowest energy transition has a charge transfer character contribution. Excitation spectra of the pyrrolo[3,4-b]pyridin-5-one is shown in Supplementary Figure S42.

Compared to our previous study where we synthesized meso-thienyl BODIPY-pyrrolo[3,4-b]pyridine-5-one conjugates (Flores-Reyes, et al., 2024), the absorption maxima of both chromophores are slightly blue shifted. The bathochromic shift in the meso-thienyl derivatives is attributed to a donor-acceptor structure arising from the inclusion of the thiophene moiety, a weak electron donor. The FQYs are very low and about the same in both series of compounds, which was expected due to the unrestricted rotation of the BODIPY moiety. The most striking difference between both types of compounds lies in the emission spectra. The meso-phenyl series exhibit a dual emission with a significantly blue shifted emission maxima for the BODIPY core (∼110 nm), while the meso-thienyl counterpart exhibited two distinct emission maxima for each chromophore with different excitation wavelengths. This also highlights the dramatic effect on emission of switching the thiophene for a phenyl ring.

It was investigated how the emission spectra of compounds 11a-g changed in response to varying viscosities in mixtures of dimethylsulfoxide (DMSO) and glycerol (Gly). The study began with a solution consisting entirely of DMSO, and glycerol was gradually added to increase the viscosity. The specific viscosities used are detailed in Table 3.

As the viscosity of the medium increases with the addition of glycerol, all compounds exhibit an increase in the emission intensity corresponding to the BODIPY core. Similarly, the pyrrolo[3,4-b]pyridin-5-one core shows a hypsochromic shift, with peaks initially around 444 nm in DMSO shifting to approximately 410 nm in glycerol-containing solutions. This shift may be attributed to the stabilization of a conformer favoured by the medium’s viscosity, which affects the rotation of the triple bond between the two fluorophores. Emission spectra across various viscosities for all compounds are shown in the Supplementary Figures S43–S49. This phenomenon has been previously observed in BODIPY dimers linked by butadiyne units (Zhang et al., 2017).

The dependence of fluorescence intensity on the viscosity of the solutions was analyzed according to the Förster-Hoffmann equation (Equation 1), where I _f represents the fluorescence intensity, η is the viscosity of the medium, c is a constant dependent on concentration and temperature, and χ is a constant related to the sensitivity of the fluorophores towards viscosity (Förster and Hoffmann, 1971; Su et al., 2016). Ln   I f = c + χ ⁡ ln   η (1)

Calibration curves were obtained by performing an excitation at 400 nm for each sensor. Among them, compound 11g exhibited the best linear trend (R ² = 0.9852) with a coefficient χ = 0.511 (Figure 4). This one suggests that the presence of smaller groups attached to the nitrogen atom of the pyrrolo[3,4-b]pyridin-5-one reduces potential energy losses, which might occur in other analogues into this compound family due to a higher number of degrees of freedom in the substituents. To see the coefficients obtained by the Förster-Hoffman analysis of all the studied compounds see Supplementary Table S2.

The curve obtained for this analogue was compared with the one obtained when performing an excitation at 290 nm, where the emission bands of the fluorophores in the 100% DMSO solution were studied (530 nm for BODIPY and 446 nm for pyrrolo[3,4-b]pyridin-5-one), as well as the new band observed at 406 nm with increased viscosity (Supplementary Figure S50). The bands corresponding to the pyrrolo[3,4-b]pyridin-5-one core did not show a linear behavior with respect to the natural logarithm of viscosity, whereas the BODIPY band demonstrated higher sensitivity (χ = 0.5551) and slightly lower correlation at this excitation wavelength.

Although the emission bands of the pyrrolo[3,4-b]pyridin-5-one core did not exhibit a linear trend, possibly due to non-radiative decay processes and conformer formation, it is noteworthy that the behavior of all studied compounds upon excitation at 490 nm follows a linear fitting. The correlation coefficients (R ²) are at least 0.9, and the χ values range from 0.5027 to 0.6082, indicating a good sensitivity to changes in viscosity. This correlation suggests a systematic response among the meso-phenyl BODIPY-containing molecules, demonstrating their high sensitivity to variations in viscosity.

For further elucidation of the electronic transitions within compounds 11a-g, the ground state geometries were optimized by DFT calculations using the program Gaussian 09, (Frisch et al., 2010), with the B3LYP functional and the 6-311G(d,p) basis set. The solvation model based on density (SMD) was applied using DMSO as solvent to enable a comparison with experimental results. To ensure that the optimized structures corresponded to energy minima, a frequency analysis was conducted to confirm the absence of imaginary frequencies. The optimized geometry of the compounds showed a torsion angle ranging from 53.8° to 54.2° between the BODIPY core and the phenyl substituent at the meso position, indicating a deviation from coplanarity (Table 4). These values are consistent with literature reports for meso-phenyl BODIPY compounds (Li et al., 1998).

To investigate how variations in the dihedral angle influence the energy of the compounds, a scan of the dihedral angle between the BODIPY core and the meso-phenyl ring was performed (Figure 5). This scan was conducted in the ground state by twisting the dihedral angle in 10° increments, covering a full range of 180°, using compound 11a as a model. The results revealed that the rotamers with the highest energies corresponded to dihedral angles of 2° and 172°, which represent nearly coplanar conformations of the two fragments. In contrast, two energy minima were observed for rotamers at dihedral angles of 52° and 122°, suggesting that a more planar orientation is favored rather than a completely orthogonal one (Dong et al., 2021). The lowest energy rotamer, with a dihedral angle of 52°, exhibited energy differences of 1.72 eV and 1.74 eV compared to the higher energy rotamers at 2° and 172°, respectively.

Based on the ground state optimized geometry of compounds 11a-g, the electronic transitions were determined by time dependent DFT (TD-DFT) using the B3LYP functional and the 6-311+G(d,p) basis set with DMSO as solvent. Table 5 shows the experimental and theoretical absorption maxima for the BODIPY and pyrrolo[3,4-b]pyridin-5-one fragments, oscillator strengths and the orbitals involved in the transitions with the corresponding contributions. Analysis of the frontier molecular orbitals revealed that for compounds 11a, 11b, 11c, 11e and 11g, the highest occupied molecular orbital (HOMO) distribution is located mainly in the pyrrolo[3,4-b]pyridin-5-one core, including the morpholine fragment. For compounds 11d and 11f, the HOMO is also present in the electron rich 3,4-dimethoxybenzyl and 3,4-dimethoxyphenethyl groups, respectively. None of the compounds showed electron density contribution from the benzyl substituent. Regarding the lowest unoccupied molecular orbital (LUMO), it is distributed in the meso-phenyl BODIPY system on all compounds and no overlap with the HOMO can be seen. The LUMO energies are very similar among the compounds differing by only ±0.002 eV, however, the HOMO energies show a higher variation, attributed to the different substituents of the pyrrolo[3,4-b]pyridin-5-one (Figure 6).

In all cases, the lowest energy absorption band, attributed to the BODIPY fragment, consists predominantly of a HOMO→LUMO transition (≥58% contribution for all compounds) with considerable oscillator strength, and was computed with a hypsochromic shift between 16 and 25 nm compared to the experimental value. The HOMO is located within the pyrrolo[3,4-b]pyridin-5-one, while the LUMO is in the meso-phenyl BODIPY system, making this transition of charge-transfer character. Other transitions contributing to this absorption band are of the type of HOMO-x→LUMO, with x = 1, 2, 3. These higher HOMOs are located within the BODIPY fragment or in the phenyl-alkyne-phenyl bridge. On the other hand, the highest energy absorption band, which is attributed to the pyrrolo[3,4-b]pyridin-5-one, was reproduced with high precision as it only differed by 2–7 nm from the experimental results. For all compounds in general, the most contributing electronic transition consisted of a HOMO→LUMO+3 transition (≥89% contribution for all compounds). Both molecular orbitals are located within the pyrrolo[3,4-b]pyridin-5-one fragment.

Additional TD-DFT calculations were performed in CHCl₃ and PhMe to assess the influence of different solvents in the frontier molecular orbitals (Supplementary Material). Computed absorption maxima in CHCl₃ and PhMe are slightly blue shifted in comparison with DMSO, which was expected since DMSO is the more polar solvent. Oscillator strengths are higher in PhMe compared to DMSO and CHCl₃, indicating that in this solvent the compounds may show a more intense absorption and emission. The LUMO energies in CHCl₃ and PhMe showed a considerable increase with respect to DMSO, resulting in a larger HOMO-LUMO gap. This may be attributed to the capability of DMSO to form hydrogen bonds, thus lowering the energy of the LUMO. Another noticeable difference is in the distribution of the HOMO since in CHCl₃ and PhMe it shows a greater localization in the phenyl-alkyne-phenyl system than in DMSO (Supplementary Tables S5, S6).

¹H and ¹³C Nuclear Magnetic Resonance (NMR) spectra were acquired on a Bruker AMX Advance III spectrometer (500 MHz, Fällande, Uster, Switzerland). The solvent used for NMR experiments was deuterated chloroform (CDCl₃). Chemical shifts are reported in parts per million (δ/ppm). Coupling constants are reported in Hertz (J/Hz). Internal reference for NMR spectra was tetramethylsilane (TMS) at 0.00 ppm. Multiplicities of the signals are reported using the standard abbreviations: singlet (s), doublet (d), triplet (t), quartet (q), and multiplet (m). NMR spectra were analysed using the MestReNova software (Ver. 12.0.0-20080). HRMS spectra were acquired by electrospray ionization (ESI) on a Micro-TOF II spectrometer Bruker Daltonics GmbH (Bremen, Germany). HRMS samples were injected directly (Apollo source) and analysed by the Time-of-Flight method (TOF). HRMS spectra were analysed using the Compass analysis software (Ver. 1.5, Flex Control and Flex Analysis by Bruker Daltonics, Inc.). Microwave-assisted reactions were performed in closed-vessel mode on a CEM Discover SP MW-reactor (Matthews, North Carolina, CA, United States). Reaction progress was monitored by thin-layer chromatography (TLC) and the spots were visualized under ultraviolet (UV) light (254 or 365 nm). Flash columns packed with silica gel in a 0.063–0.200 mm mesh particle size were used to purify the products. Mixtures of hexanes (Hex) and ethyl acetate (EtOAc) in 2:3 or 3:2 (v/v) proportions were used to run TLC, silica gel columns, and to measure the retention factor (R _f ) values (using the same mobile phase for all the experiments). R _f values were measured after an elution of 4 cm. All starting reagents and solvents were used as received (without further purification, distillation, or dehydration). Chemical structures were drawn using the ChemDraw Professional software (Ver. 15.0.0.106, Perkin Elmer Informatics, Cambridge, MA, United States).