Synthesis and Photophysical Properties of Fluorescent 6-Aryl-D-π-A Coumarin Derivatives.

A series of 6-aryl coumarin dyes were synthesized in satisfactory yields by Pd-catalyzed Suzuki cross-coupling reactions with a panel of boronic acids and coumarin bromides. Photophysical studies highlighted a large Stoke shift and interesting fluorescence quantum yield for these compounds. Optical properties were also investigated with the aid of quantum chemical calculations. The treatment of selected coumarin dyes with increasing amounts of trifluoroacetic acid showed that their fluorescence can be strongly influenced by pH (fluorescence quenching at high acid concentrations), while the addition of Fe3+ and Al3+ metal ions allowed to highlight dichotomous behavior with the corresponding reduction in fluorescence with the increase of [Fe3+] or [Al3+]. Finally, biological assays and fluorescence microscopy imaging investigations indicated that these compounds can be used as potential biomarkers in living and fixed cells.

Introduction

Coumarin derivatives represent an important class of heterocyclic compounds possessing many significant electro-optical properties as well as showing different biological activities.[1−5] These molecular scaffolds are generally characterized by high fluorescence, large Stokes shifts, high photoluminescence quantum yield (PLQY), sensitivity, dual emission, internal charge transfer (ICT), and twisted ICT properties.[6,7] For all of these reasons, coumarin adducts find application in various research fields such as laser dyes,[8] organic light-emitting diode fabrication,[9] and dye-sensitized solar cells[10] as cell-imaging biomarkers[11] or optical brighteners.[12] The visible absorption properties of such compounds are highly controlled by the insertion of various substituents directly bonded to the heteroaromatic structure, and in particular, coumarin functionalization by the insertion of electron-donating groups (EDGs) in the positions C5 or C7 and electron-withdrawing groups (EWGs) in the positions C3 or C4 has been previously investigated.[13] By analyzing several coumarin derivatives used in dye laser applications, Zeidler and co-workers observed that the benzene rings of these compounds showed a predominant para-quinoidal resonance state, determined by a “push–pull” effect, which leads to intramolecular charge transfer (ICT) with consequent shortening of the C5–C6 and C8–C9 bonds. The increase in the “push–pull” effect[14a−14c] reduces the coumarin band gaps leading to a spectral redshift in UV–vis absorption. The increase in the redshift and molar extinction coefficients could also be obtained by selectively inserting an EWG in the C3 position.[14c] Moreover, Imai and co-workers, highlighted how high PLQY values could be achieved by using C6 or C7 EDG functionalized coumarins.[15]

Finally, π-expanded coumarin derivatives,[16] such as benzocoumarins or vertically π-expanded derivatives, revealed redshifted absorption and emission and improved quantum fluorescence yields compared to the simple bicyclic compounds (Figure ).[17] These achievements and the continuous investigations in this field indicate that the rational synthesis of donor-π-acceptor (D-π-A) fluorescent molecules is still of fundamental importance because of their centrality in material science. Herein, we report a rational study concerning the synthesis and the photophysical characterization of new coumarin structures containing various aromatic spacers between the heteroaromatic coumarin scaffold and the EDG, aimed to increase the push–pull system conjugation and to select new candidate molecules for the construction of performing nontoxic biomarker dyes with efficient emissions in the blue-green spectral region.

Figure 1

Representative push–pull fluorescent coumarin systems.

Results and Discussion

Synthesis and Theoretical Calculations

6-Aryl coumarins were synthesized from 1a by a simple two-step procedure consisting of Knoevenagel condensation (with diethyl malonate or malononitrile)[21] yielding the bromo derivatives 3a–3c,[21c−21e] followed by Pd(0)-catalyzed Suzuki cross-coupling reaction with selected boronic acids, affording the corresponding adducts 4a–4f in satisfactory yields. Following the same strategy, 6,8-diaryl-coumarins 5a–5b and 5d were prepared from the aldehyde 1b as summarized in Scheme . Then, quantum chemical calculations were performed in order to predict the electronic properties of these compounds. From this investigation, we observed that 6-aryl coumarins do not assume a planar geometry as expected because the EWGs tend to rotate with respect to the coumarin plane.

Scheme 1

Synthetic Route to 6- and 6,8-Aryl Coumarin Derivatives

Also, the analysis of the atomic distances and angles indicated that in the excited state, all the studied derivatives have shown a more planar structure than in the ground state, as a consequence of the redistribution of the electric charge. As shown in Figure , we report the calculated HOMO and LUMO for compound 4a(21e) as an illustrative example (details of the other coumarins 4 and 5 can be found in the Supporting Information).

Figure 2

Ball and stick representation of the 4a derivative (H = white, O = red, and C = gray). HOMO and LUMO (left and right, respectively) represent the charge distribution (isocontour value = 0.02 au).

The dihedral angle between the coumarin plane and the EDG (donor group) changes by about 11% in the excited state as compared to the ground state, while the electronic charge changes mainly on the carbon atoms of the coumarin rings, with the largest decrease of negative charge recorded in the C3 position, which is the one bridging the EWG acceptor group (COOEt for 4a). Similar results were observed for all the synthesized coumarins, with a variation of the dihedral angle in the 5–11% range, where the lowest variation was recorded for 4c. In the case of disubstituted coumarins bearing two EDGs in the C6 and C8 positions (Scheme , compounds 5a, 5b, and 5d), both substituents undergo similar variation in the excited state as compared to the ground state, leading to a more planar structure. These data are confirmed by the molecular orbital (MO) calculations shown in Figure (and Supporting Information) for the HOMO and LUMO. Upon excitation, these molecules undergo structural and electronic reorganization, leading to an electron density pinned on the coumarin plane and partially to the acceptor group, with no contribution from the donating ones. However, a complete charge transfer to the acceptor group is observed only for compound 4d. The TDDFT calculations allowed also us to estimate the absorption and emission features of these compounds, which are schematically shown in Figure (vacuum, see Supporting Information, Table S1 for details). The HOMO–LUMO transitions have been calculated at 420 nm, and the related emission was obtained at 512 nm. When the model solvent is considered, a redshift is reported for both absorption and emission features, leading to 442 and 549 nm, respectively, with similar structural and electronic properties in terms of spatial orientation and MOs. The simulated optical properties indicate that the substitution of the benzene fused ring with EDGs with increasing donating character, from phenyl (4a) to para methoxy-phenyl (4b)[21e] and 6-methoxy-naphthyl (4c) groups, lead to an estimated absorption wavelength increase, and consequently, an increase of the emissions of about 75 nm, moving the HOMO–LUMO transition from the near UV (362 nm for 4a) to the blue spectral region (435 nm for 4c) and the emission from blue (444 nm for 4a) to green (532 nm for 4c). Moreover, the variation of the acceptor group, from ethyl ester (4b) to nitrile (4e) or sulfone (4f) groups, shows a redshift of about 20 nm for 4a, while no significant differences were observed for the compounds 4b and 4e. However, C6 and C8 diaryl-disubstituted coumarins, such as 5a and 5b, show an absorption redshift of about 20 nm and emission of 40 nm, in comparison to the C6 monosubstituted derivatives. Finally, as already stated before, for all the simulated coumarins, the presence of solvent causes an almost rigid redshift for the calculated optical properties.

Figure 3

(a) Absorbance spectra of investigated coumarins. (b) Absorbance spectrum and simulated oscillator strength of the 4e sample in chloroform.

Optical Characterization of Aryl Coumarins 4 and 5

Figure a shows a comparison of the experimental absorption spectra of coumarins 4a–4c, 4f, and 5a-5b. All the spectra show two main excitations; the first one in the far UV around 300 nm and the second one in the 300–400 nm region. As depicted, the largest redshift is observed for 4e, in good agreement with the simulated results (Figure b). A detailed comparison of the coumarin absorption spectra as a function of the substitution is reported in the Supporting Information.

Absorption analyses conducted with compounds 4a (6-phenyl), 4b [6-(4-methoxyphenyl)], and 4c [6-(6-methoxy-naphthyl)] were in good agreement with these simulations (Figure b). Keeping the 6-(4-methoxyphenyl) unit and varying the acceptor group at the C3 position from the ethyl ester moiety (4b) to nitrile (4e) gave superposable trends with the calculated spectra. On the other hand, the absorption spectra of compound 4d, bearing 6-(4-N,N-dimethylphenyl) as a donor group and a carboxyethyl function as acceptor, showed a redshift of about 20 nm to respect the predicted ones. On the contrary, predictions regarding the photophysical properties of disubstituted coumarin 5 were not reflected in the experimental results because the recorded spectra did not show significant differences between mono- and disubstituted compounds. This discrepancy might be related to the poorly estimated interaction between the two substituents with the B3LYP pseudopotential. A direct comparison between the experimental and calculated spectra is shown in Figure b, where the oscillator strength of the calculated first six excitation channels and the experimental absorbance spectrum of the 4d sample are superimposed.

Correlations between the coumarin substitution and the emission properties are also detectable in Figure a, in which the recorded spectra show a large emission shift of about 70 nm between 4a and 4c. On the other hand, experimental and calculated emission peak data for compounds 5b, 4c, 4e, and 4f are almost overlapping. While in the calculated transitions, a shift of more than 25 nm was obtained, leading to the larger predicted redshift for the derivative 4c. Again, disubstituted coumarins 5a and 5b show an emission redshift of about 15 and 20 nm, respectively (Figure a), which is about half the predicted shift (see Supporting Information).

Figure 4

(a) PL spectra of investigated coumarins in chloroform excited at 340 nm. (b) PL spectra of the 4e sample in different solvents excited at 300 nm.

Nevertheless, a series of solvatochromic studies on coumarins 4a–4c, 4e, and 4f were carried out using n-hexane, toluene, diethyl ether, and chloroform. These investigations highlighted large emission shifts (>60 nm) for various coumarin dyes. As an example, in Figure b, we show the solvatochromism of the dye 4e, while the maximum peak values for the adducts 4a–4c and 4e-4f are listed in Table (further solvatochromism investigations have been reported in the Supporting Information).

Table 1

Emission Values of Coumarins 4a–4c, 4e, and 4f Measured in a Selected Panel of Solventsa

coumarin	λem-n-Hex (nm)	λem-Tol (nm)	λem-Et2O (nm)	λem-CHCl3 (nm)
4a	441	449	450	457
4b	454	470	475	498
4c	455	477	485	515
4e	454	480	493	512
4f	452	476	484	510

λem = maximum emission wavelength.

Coumarins 4 and 5 were excited at 350 nm in order to record their decay time. All the compounds showed a single exponential decay with lifetimes ranging from subnanoseconds as in the cases of compounds 4a and 5a to 9 and 13 ns for 4f and 4e, respectively (Figure ). These investigations also highlighted that by varying the acceptor group, we could observe a lifetime increase (SO2Me > CN > COOEt).

Figure 5

Decay time plot of coumarins 4a–4c, 4e-4f, and 5a-5b under excitation at 350 nm.

Besides, the larger Stoke shifts and quantum yield have been observed for the compounds 4e and 4f. Table summarizes the spectral features for both coumarins 4a–4f and 5a-5b in chloroform solutions.

Table 2

Photophysical Properties of Coumarinsa

coumarin	λabs (nm)	λPL (nm)	Φb	τ (ns)
4a	258, 295, 357	456	0.05	0.7
5a	252, 309, 365	465	0.03	0.6
4b	274, 370	500	0.15	6.0
5b	280	520	0.12	5.7
4c	303, 372	515	0.13	6.0
4e	283, 380	512	0.45	9.3
4f	288, 368	510	0.69	13.0

Fluorescence characteristics of coumarins 4a–4c, 4e-4f, and 5a-5b (chloroform solution).

Φ values were calculated using rhodamine 6G as ref (22).

The most promising coumarins 4b, 4e, and 4f (10 μM in chloroform) were submitted to a series of additional investigations. Fluorescence emission experiments were carried out in the presence of protic acids in order to evaluate the acidochromism properties of these compounds (Figure ). In this investigation, trifluoroacetyl (TFA) titrations caused a general variation (decrease) in the fluorescence maximum and sensible redshift from 512 to 536 nm for the coumarin 4e (Figure a), from 510 to 533 nm for 4f (Figure b), and from 496 to 508 for 4b (Figure S5, Supporting Information). These emission redshifts should be mainly due to the protonation of the nitrile group of 4e and the ester functions for the adducts 4b and 4f. Furthermore, for all three dyes, a decrease in luminescence was observed, revealing acidochromic activity (fluorescence turn-off). Then, in separated experiments, compounds 4c, 4e, and 4f were submitted to a series of titrations with C- or N-protected amino acids or with solutions of selected metal salts (Co2+, Sn2+, Ca2+, Cu2+, Al3+, Gd3+, and Fe3+) aiming to identify any selective interaction with organic or inorganic entities that would indicate potential sensor properties for these compounds. However, experiments conducted with ammino acids were ineffective (see Supporting Information).

Figure 6

Changes in the emission spectra of (a) 4e and (b) 4f (10 μM) with the incremental addition of TFA in chloroform (λex = 350 nm).

Coumarins 4e and 4f showed sensible variation in their emission spectra when Al3+ and Fe3+ (albeit less markedly Cu2+) salts were added to their solutions as illustrated in Figure a,b (titration methods and further emission/absorption spectra related to these investigations are reported in the Supporting Information). Notably, in the presence of Fe3+, the fluorescence spectra of adducts 4e and 4f exhibited an appreciable fluorescence turn-off response (Figure a). On the other hand, the addition of Al3+ salts to 4e and 4f solutions caused fluorescence enhancement and a slight blue shift, as shown in Figure b.

Figure 7

(a) Fluorescence emission spectra of coumarins 4e and 4f (10 μM) titrated with Fe3+ salts (0–100 equiv) in methanol. (b) Fluorescence spectra of coumarins 4e and 4f (10 μM) in the presence of different metal ions (100 equiv) in methanol.

Finally, the coumarins 4c (λabs = 303, 372), 4e (λabs = 283, 380), and 4f (λabs = 288, 368) were valued as potential biomarkers for bioimaging applications. For this purpose, cytotoxicity in cell-based assay was evaluated by using Nthy-ori 3–1 cells [Simian Virus 40 (SV40)-immortalized normal human thyrocytes]. Cultured cells were treated with different concentrations (10–100 μM) of 4c, 4e, and 4f. Then, cell viability was determined after 24 h at 37 °C by the MTT method.[23] As shown in Figure , both coumarins have proved to be noncytotoxic against cell monolayers (CC50 = >100 μM). In parallel, the Nthy-ori 3–1 cells were incubated with 4c, 4e, and 4f (1.0 μM) for bioimaging studies and then analyzed by an epifluorescence microscope using blue (excitation 470–495 nm; emission 510–550 nm) and orange filters (excitation 530–550 nm; emission 575 nm). As shown in Figure , no fluorescence was detected in the external buffer or the nucleus, whereas collected images showed that all the coumarin dyes are emissive in the cytoplasmic compartments.

Figure 8

Viability of the Nthy-ori 3–1 cell line in the presence of 4c (black), 4e (white), and 4f (gray). The cells were treated at different concentrations, and cytotoxicity was determined using the MTT assay by monitoring formazan absorbance at 570 nm. Data represent mean values (±SD) for three independent determinations.

Figure 9

Bioimaging of Nthy-ori 3–1 cells treated with 4c, 4e, and 4f coumarins (1 μM). (a) 20× and (b) 100× fluorescent images of Nthy-ori 3–1 cells treated with 4c coumarin; (c) 20× and (d) 100× fluorescent images of Nthy-ori 3–1 cells treated with 4e coumarin; and (e) 20× and (f) 100× fluorescent images of Nthy-ori 3–1 cells treated with 4f coumarin. Blue: cell nuclei. Orange: coumarins.

Furthermore, the intracellular environment does not affect the fluorescence properties of these compounds. These pieces of evidence prompted us to deduce that coumarins tend to interact at the cytoplasmic lipophilic substructures, helped by their poor solubility in water. However, to confirm this hypothesis, high-resolution confocal microscopy studies are ongoing in our laboratories.

Conclusions

The synthesis of new push–pull coumarins was successfully performed with a two-step strategy in good to satisfactory yields. Photophysical characterization allowed us to identify a series of derivatives endowed by good quantum fluorescence efficiencies, also characterized by solvatochromic and acidochromic properties. Compounds 4e and 4f showed certain selectivity for Fe3+ salts, causing fluorescence turn-off, while the addition of Al3+ enhanced their fluorescence. Nevertheless, in vitro bioassays highlighted the low toxicity of coumarins 4c, 4e, and 4f that were thus investigated as potential candidates for bioimaging applications. To this end, further studies are currently ongoing in our laboratory, aimed at identifying any selectivity of these compounds toward specific cytoplasmic subunits.

Methods

All the reagents were purchased from Merk or TCI and used without further purification.

UV–vis absorbance spectra were carried out with a Varian Cary 60 spectrophotometer (200–600 nm). Emission spectra were collected (90° geometry) with a HORIBA Jobin Yvon FluoroMax 3.0 spectrofluorometer (300–600 nm range). In both cases, a quartz cuvette (10 mm optical path) was used. Time-resolved measurements were carried out by exciting the samples in the UV range with 200 fs long pulsed laser light delivered by an optical parametric amplifier (Light Conversion TOPAS-C) pumped by a regenerative Ti/sapphire amplifier (Coherent Libra-HE, repetition frequency 1 kHz). The PL signals were recorded by a streak camera (Hamamatsu C10910) equipped with a grating spectrometer (Princeton Instruments Acton SpectraPro SP-2300). The emission signals were gathered in the front face mode in order to avoid inner filter effects. 1H NMR spectra were recorded on a 500 MHz Varian spectrometer at 25 °C using CDCl3 (ref. 7.26 ppm) as the solvent. 13C NMR spectra were recorded at 126 MHz (ref. CDCl3 77.16 ppm) using CDCl3 as the solvent. The chemical shifts (δ) are given in ppm. The coupling constants (J) are reported in Hz. Low mass spectra analysis was recorded using Agilent-HP GC–MS (E.I. 70 eV). High-resolution mass spectra (HRMS) of compounds 3, 4, and 5 were obtained using a high-resolution mass spectrometer in fast atom bombardment ionization mode acquired using Bruker micrOTOF-Q II or/and Agilent Q-TOF 6520. The melting points were determined with Büchi M-560 (°C). Analytical thin-layer chromatography was performed using 0.25 mm Aldrich silica gel 60-F plates. Flash chromatography was performed using Merck 70-200 mesh silica gel. Yields refer to chromatography and/or spectroscopically pure materials. Acetone, acetonitrile, and ethyl acetate were used as received (HPLC grade >99%) or distilled with the appropriate procedure. THF and toluene were distilled from sodium/benzophenone ketyl. All the simulations were performed with the Gaussian 16 package[18] to optimize the structures down to the self-consistent field energy and to calculate their optical properties. Ground and excited states were simulated within DFT and TDDFT frameworks, respectively, by exploiting the B3LYP hybrid functional with the 6-31G(d,p) basis set as already reported for similar compounds.[19] The simulations were accomplished under vacuum and with a model solvent, chloroform. In the latter cases, solvation effects arising from the interaction of coumarin derivatives with chloroform were treated with the self-consistent reaction field model by simulating the dielectric solvent through the polarizable continuum model calculation with-in the integral equation formalism.[20]

The Nthy-ori 3–1 cell line [Simian Virus 40 (SV40)-immortalized normal human thyrocytes] was purchased from the Health Protection Agency Culture Collections (Health Protection Agency Culture Collections; 2011 http://www.hpa.org.uk, last accessed on 20 May 2021).

The absence of mycoplasma contamination was checked periodically by the Hoechst staining method. The cells were grown as monolayers in DMEM/Ham’s F-12 (DMEM/F12) supplemented with 10% FBS (Life Technologies, Milan, Italy), 100 UI/mL penicillin, and 100 μg/mL streptomycin (Sigma-Aldrich, Milan, Italy), at 37 °C in a humidified 5% CO2 atmosphere.

To evaluate the cytotoxic effect of 4c, 4e, and 4f, MTT assay was performed. Briefly, the cells were seeded at a density of 7.5 × 103 cells in a 96-well plate and incubated for 24 h. Then, the compounds were added at different concentrations (10–100 μM), and the cells were further incubated for 24 h. After incubation, 50 μL of MTT reagent (1 mg/mL in DMEM/F12) was added, and then, the cells were incubated for an additional 4 h. The resulting formazan crystals were dissolved in 100 μL of DMSO. The absorbances were measured at 570 nm using a Tecan microplate reader (Infinite 200, Tecan, Salzburg, Austria). The extent of cell growth/viability at each coumarin concentration tested was expressed as a percentage of untreated controls. Concentrations resulting in 50% inhibition (CC50) were determined by linear regression analysis.

For cellular bioimaging assay, the cells were grown on coverslips. After 24 h with the designed compound solution (1 μM), the cells were washed three times in PBS 1× (pH 7.4) and then fixed in 3.7% formaldehyde for 30 min. The slides were then washed with PBS 1× and mounted in antifading solution with 0.15 μg/mL DAPI as a counterstain. The samples were analyzed using a digital image analysis system based on an epifluorescence Olympus BX41 microscope using blue (excitation 470–495 nm; emission 510–550 nm) and orange filters (excitation 530–550 nm; emission 575 nm) and a charge-coupled device camera (Cohu, San Diego, CA) interfaced with the CytoVysion System (Applied Imaging).