Interaction Study between ESIPT Fluorescent Lipophile-Based Benzazoles and BSA

In this study, the interactions of ESIPT fluorescent lipophile-based benzazoles with bovine serum albumin (BSA) were studied and their binding affinity was evaluated. In phosphate-buffered saline (PBS) solution these compounds produce absorption maxima in the UV region and a main fluorescence emission with a large Stokes shift in the blue–green regions due to a proton transfer process in the excited state. The interactions of the benzazoles with BSA were studied using UV-Vis absorption and steady-state fluorescence spectroscopy. The observed spectral quenching of BSA indicates that these compounds could bind to BSA through a strong binding affinity afforded by a static quenching mechanism (Kq~1012 L·mol−1·s−1). The docking simulations indicate that compounds 13 and 16 bind closely to Trp134 in domain I, adopting similar binding poses and interactions. On the other hand, compounds 12, 14, 15, and 17 were bound between domains I and III and did not directly interact with Trp134.

In this study, the binding affinity for BSA of photoactive lipophilic amines based on the benzazolic fluorophore, comprising benzoxazole and benzothiazole compounds, was evaluated. First, their electronic properties were investigated in phosphate buffer solution (PBS) in both the ground and excited states by UV-Vis absorption and steady-state fluorescence emission, respectively. Their application as optical sensors to detect proteins in PBS was successfully explored using BSA as a model. These compounds were chosen due to their high photostability and absent inner-filter effect, which could be interesting for association studies. In addition, the presence of different alkyl chains could also affect their interaction with BSA by suppression studies, provided by hydrophobic interactions with the macromolecule. Finally, molecular modeling studies were performed to better visualize the binding mode with BSA.

Synthesis and Photophysical Characterization
The fluorescent secondary amines used in this study are similar to those reported previously [41]. Generally, the reaction of the formyl derivatives 1-2 [14] with an equimolar quantity of amines 3-5 in dry isopropanol/acetic acid (catalyst) under reflux yielded the respective imines 6-11 (Scheme 1). In this step, the solvent was evaporated to dryness to afford 6-11 in a~90% crude yield, and no further purification was required to proceed with the next step. The desired amines were obtained by reducing the imines with sodium borohydride in a 1:2 (imine:NaBH 4 ) ratio. The final products were obtained after purification via column chromatography in satisfactory yields (~70%) using ethyl acetate as the eluent.
In this study, the binding affinity for BSA of photoactive lipophilic amines based on the benzazolic fluorophore, comprising benzoxazole and benzothiazole compounds, was evaluated. First, their electronic properties were investigated in phosphate buffer solution (PBS) in both the ground and excited states by UV-Vis absorption and steady-state fluorescence emission, respectively. Their application as optical sensors to detect proteins in PBS was successfully explored using BSA as a model. These compounds were chosen due to their high photostability and absent inner-filter effect, which could be interesting for association studies. In addition, the presence of different alkyl chains could also affect their interaction with BSA by suppression studies, provided by hydrophobic interactions with the macromolecule. Finally, molecular modeling studies were performed to better visualize the binding mode with BSA.

Synthesis and Photophysical Characterization
The fluorescent secondary amines used in this study are similar to those reported previously [41]. Generally, the reaction of the formyl derivatives 1-2 [14] with an equimolar quantity of amines 3-5 in dry isopropanol/acetic acid (catalyst) under reflux yielded the respective imines 6-11 (Scheme 1). In this step, the solvent was evaporated to dryness to afford 6-11 in a ~90% crude yield, and no further purification was required to proceed with the next step. The desired amines were obtained by reducing the imines with sodium borohydride in a 1:2 (imine:NaBH4) ratio. The final products were obtained after purification via column chromatography in satisfactory yields (~70%) using ethyl acetate as the eluent. The photophysical investigation was carried out in PBS (pH 7.2) at a concentration of 10 −5 M. The characterization data from the electronic ground and excited state are summarized in Table 1. As shown in Figure 1, the benzazoles exhibit absorption maxima in the UV region, as already observed in organic media [41]. The benzothiazoles produced red-shifted absorption maxima compared to their benzoxazole analogs, which can be explained by the better electron delocalization transferred from the sulfur to the oxygen [12]. Surprisingly, both derivatives with short alkyl chains (12 and 15) exhibited absorption maxima at high energies (~330 nm) despite the compounds with longer lipophilic chains, indicating that the alkyl chain somehow plays a role in their electronic properties in the ground state. In addition, for compounds 13-14 and 16-17, broad ab-Scheme 1. Synthesis of the ESIPT fluorescent lipophile-based benzazoles 12-17, under the following conditions: (i) iPrOH, CH 3 COOH, reflux, 5-24 h, and (ii) iPrOH, NaBH 4 .
The photophysical investigation was carried out in PBS (pH 7.2) at a concentration of 10 −5 M. The characterization data from the electronic ground and excited state are summarized in Table 1. As shown in Figure 1, the benzazoles exhibit absorption maxima in the UV region, as already observed in organic media [41]. The benzothiazoles produced red-shifted absorption maxima compared to their benzoxazole analogs, which can be explained by the better electron delocalization transferred from the sulfur to the oxygen [12]. Surprisingly, both derivatives with short alkyl chains (12 and 15) exhibited absorption maxima at high energies (~330 nm) despite the compounds with longer lipophilic chains, indicating that the alkyl chain somehow plays a role in their electronic properties in the ground state. In addition, for compounds 13-14 and 16-17, broad absorption was observed in the visible region of the spectra. This feature could be a result of scattering and may be an indication of aggregation of these compounds in PBS (pH 7.2), due to the presence of the long alkyl chains in these compounds, which are extremely hydrophobic. Significantly, Molecules 2021, 26, 6728 3 of 15 the similarity in shape and maxima wavelength compared to those of organic solvents, as well as the absence of additional red-shifted bands in PBS, indicates the absence of ionized species in the ground state for derivatives in this media [42,43]. Table 1. Photophysical data of ESIPT fluorescent lipophilic benzoxazoles 12-17 from UV-Vis absorption and fluorescence emission spectroscopies, where ε is the molar extinction coefficient, λ abs and λ em are the absorption and emission maxima, respectively, f e is the calculated oscillator strength, k e 0 is the calculated radiative rate constant, τ 0 is the calculated pure radiative lifetime, ∆λ ST is the Stokes shift, and Φ FL is the fluorescence quantum yield.  [42,43]. Table 1. Photophysical data of ESIPT fluorescent lipophilic benzoxazoles 12-17 from UV-Vis absorption and fluorescence emission spectroscopies, where ε is the molar extinction coefficient, λabs and λem are the absorption and emission maxima, respectively, fe is the calculated oscillator strength, ke 0 is the calculated radiative rate constant, τ0 is the calculated pure radiative lifetime, ΔλST is the Stokes shift, and ΦFL is the fluorescence quantum yield.  Finally, the UV-Vis absorption spectra were used to calculate the experimental extinction coefficients ε from the oscillator strengths fe using the Strickler-Berg relationship represented by Equation (1) [44]. In this equation, the integral is related to the area of the absorption maxima from a plot of ε (M −1 ·cm −1 ) vs. ̅ (wavenumber, cm −1 ), correlated to a single electron oscillator. Finally, the UV-Vis absorption spectra were used to calculate the experimental extinction coefficients ε from the oscillator strengths f e using the Strickler-Berg relationship represented by Equation (1) [44]. In this equation, the integral is related to the area of the absorption maxima from a plot of ε (M −1 ·cm −1 ) vs. v (wavenumber, cm −1 ), correlated to a single electron oscillator.
The calculated molar absorptivity values (ε,~10 4 M −1 ·cm −1 ), as well as the radiative rate constants (k e 0 ,~10 8 s −1 ) indicate spin and symmetry-allowed electronic transitions, which could be related to 1 π-π* transitions. Similar values were obtained for the same compounds in organic media, as discussed in the literature [41]. Moreover, an almost constant radiative lifetime τ 0 indicates that these benzazoles seem to populate the same excited state after radiation absorption. The steady-state fluorescence emission spectra of the studied benzazoles are presented in Figure 2. The curves were obtained by exciting the compounds at the absorption maxima (Table 1). In general, all compounds produced the main emission band located above 450 nm with a large Stokes shift. However, the nature of the emission bands seems to be quite different depending on the hydrophobic portion of the molecule, once again indicating that the alkyl chains influence their photophysics. For the benzoxazole derivatives, compound 12 demonstrated a blue-shifted emission (464 nm, ∆λ ST 8751 cm −1 ) when compared to analogs 13 (502 nm, ∆λ ST 8651 cm −1 ) and 14 (492 nm, ∆λ ST 8246 cm −1 ). Considering the Stokes shift values and the emission maxima location [9], we believe that the emission from compound 12 was produced due to its ionized species. Moreover, from the same parameters, the emission from the additional derivatives 13 and 14 could be related to the ESIPT process (tautomeric emission). Benzothiazole derivatives 15-17 produced similar results. Compound 15, which has a short alkyl chain, emitted fluorescence at 467 nm with a Stokes shift of 8707 cm −1 . For 16-17, both the emission maxima (542 nm) and Stokes shifts (~8800 cm −1 ) were observed at higher wavelengths. Thus, the emission produced by compound 15 is probably related to its ionized species [42]. Compounds 16-17 exhibit proton transfer in the excited state, and the emission is a result of this process. In addition, compounds 16 and 17 produced a weak blue-shifted emission (462 nm), which is also believed to be related to their ionized species. The enol or normal emission was disregarded because of the high Stokes shift values related to these emission bands.
The calculated molar absorptivity values (ε, ∼10 4 M −1 ⋅cm −1 ), as well as the radiative rate constants (ke 0 , ∼10 8 s −1 ) indicate spin and symmetry-allowed electronic transitions, which could be related to 1 π-π* transitions. Similar values were obtained for the same compounds in organic media, as discussed in the literature [41]. Moreover, an almost constant radiative lifetime τ0 indicates that these benzazoles seem to populate the same excited state after radiation absorption.
The steady-state fluorescence emission spectra of the studied benzazoles are presented in Figure 2. The curves were obtained by exciting the compounds at the absorption maxima (Table 1). In general, all compounds produced the main emission band located above 450 nm with a large Stokes shift. However, the nature of the emission bands seems to be quite different depending on the hydrophobic portion of the molecule, once again indicating that the alkyl chains influence their photophysics. For the benzoxazole derivatives, compound 12 demonstrated a blue-shifted emission (464 nm, ΔλST 8751 cm −1 ) when compared to analogs 13 (502 nm, ΔλST 8651 cm −1 ) and 14 (492 nm, ΔλST 8246 cm −1 ). Considering the Stokes shift values and the emission maxima location [9], we believe that the emission from compound 12 was produced due to its ionized species. Moreover, from the same parameters, the emission from the additional derivatives 13 and 14 could be related to the ESIPT process (tautomeric emission). Benzothiazole derivatives 15-17 produced similar results. Compound 15, which has a short alkyl chain, emitted fluorescence at 467 nm with a Stokes shift of 8707 cm −1 . For 16-17, both the emission maxima (542 nm) and Stokes shifts (~8800 cm −1 ) were observed at higher wavelengths. Thus, the emission produced by compound 15 is probably related to its ionized species [42]. Compounds 16-17 exhibit proton transfer in the excited state, and the emission is a result of this process. In addition, compounds 16 and 17 produced a weak blue-shifted emission (462 nm), which is also believed to be related to their ionized species. The enol or normal emission was disregarded because of the high Stokes shift values related to these emission bands.

BSA Binding Study
Interaction studies were carried out to investigate the affinity of benzazoles 12-17 with BSA through fluorescence quenching assays at room temperature (298 K). BSA is composed of a single polypeptide chain of 583 amino acid residues [46] and three domains named I, II, and III, each of which is divided into two subdomains (A and B) [47]. Studies indicate that the main association sites are located in the hydrophobic cavities of the IIA and IIIA subdomains [48]. The BSA structure has two Trp residues, Trp-134 in the first domain, located on the surface of the molecule, and Trp-212 in the IIA subdomain, located in the hydrophobic region of the protein [49,50]. BSA produced an absorption band in the UV-Vis region, with a maximum of around 280 nm, and an intrinsic fluorescence emission with a maximum located between 340-350 nm, depending on the excitation wavelength, owing to the Trp residues [6]. Considering these photophysical properties, the interactions of small molecules with BSA generally lead to the suppression of fluorescence emission, which is a powerful indicator of their interaction mechanisms [3,4,8,51], which are generally classified as dynamic, static, or combined quenching. Static suppression mechanisms are induced by complex formation between BSA and the dye in the ground state. Dynamic suppression is caused by collisions between BSA and the dye in the excited state. In both cases, energy is transferred from the protein to the compound, resulting in the suppression of fluorescence [52]. Thus, to investigate the fluorescence quenching of BSA in the presence of the benzazoles, spectrophotometric titration was performed using different quantities of dye (2-20 µM) and constant BSA concentration (11 µM, 1 mL) in PBS (pH 7.2). This study was performed at room temperature (25 • C), and the suppression percentage was calculated from the emission intensities obtained under excitation at 277 nm. Figure 3 shows that the UV-Vis absorption bands of the dyes increase in intensity upon addition of the benzoxazoles 12-14 to the solution containing BSA. In addition, the absorption band around 275 nm, related to BSA, also increased due to interference from another interaction that causes fluorescence at a similar wavelength. A similar response was observed for the benzothiazole analogs 15-17 (Figures S1-S3 in Supplementary Materials).  BSA emitted intense fluorescence with a maximum at 334 nm under excitation at 277 nm. The fluorescence emission spectra of BSA in the presence of benzazoles 12-13 shows an initial decrease of approximately 10% upon the first addition of 2 μM quencher (12: 9%, 13: 7%, and 14: 8%) and a significant decrease in the fluorescence intensity at 20 μM dye concentration (12: 42%, 13: 50%, and 14: 41%) with no significant red-shift, suggesting that the interaction site may be close to the Trp residue (Trp-134 or Trp-212). Moreover, these results suggest that there is no perturbation of the microenvironment around the Trp binding site [52]. Similar behavior was observed for the benzothiazole analogs at 2 μM (15: 8%, 16: 6%, and 17: 5%) and at 20 μM (15: 48%, 16: 53%, and 17: 41%; Figures  S4-S6). Higher values for BSA quenching were obtained for compounds 13 and 15, which have 12 carbon aliphatic chains, probably due to an optimized balance between hydro- shows an initial decrease of approximately 10% upon the first addition of 2 µM quencher (12: 9%, 13: 7%, and 14: 8%) and a significant decrease in the fluorescence intensity at 20 µM dye concentration (12: 42%, 13: 50%, and 14: 41%) with no significant red-shift, suggesting that the interaction site may be close to the Trp residue (Trp-134 or Trp-212). Moreover, these results suggest that there is no perturbation of the microenvironment around the Trp binding site [52]. Similar behavior was observed for the benzothiazole analogs at 2 µM (15: 8%, 16: 6%, and 17: 5%) and at 20 µM (15: 48%, 16: 53%, and 17: 41%; Figures S4-S6). Higher values for BSA quenching were obtained for compounds 13 and 15, which have 12 carbon aliphatic chains, probably due to an optimized balance between hydrophobicity and spatial hindrance.
To identify the main fluorescence quenching mechanism induced by the benzazoles, the Stern-Volmer relationship, described in Equation (3), was applied [53] where F 0 is the fluorescence intensity of pure BSA and F is the fluorescence intensity in the presence of benzazoles (quencher [Q]). K SV is the Stern-Volmer constant, and k q is the bimolecular quenching rate constant, which is related to the suppression efficiency. τ 0 is the fluorescence lifetime of BSA in the absence of the suppressor (6.06 ns) [54]. According to this equation, K SV can be obtained from the slope of the linear fit, and k q is equal to K SV /τ 0 . The obtained curves are presented in Figure 4, and the relevant data are summarized in Table 2. The equations for the respective linear fits are listed in Table S1. High K SV values were obtained (10 4 M −1 ), especially for compounds 13 and 16, indicating that the benzazoles exhibit moderate to strong interactions with the BSA binding sites [55]. Additionally, the k q values (~10 12 M −1 ·s −1 ) exceed the maximum value for the diffusional collision quenching constant according to the Smoluchowski-Stokes-Einstein theory (k diff ≈ 7.40 × 10 9 M −1 s −1 ) [56], which indicates that fluorescence quenching occurred by a static mechanism, in which the formation of a benzazole-BSA conjugate takes place in the ground state.
Molecules 2021, 26, x FOR PEER REVIEW 7 of 16 bimolecular quenching rate constant, which is related to the suppression efficiency. τ0 is the fluorescence lifetime of BSA in the absence of the suppressor (6.06 ns) [54]. According to this equation, KSV can be obtained from the slope of the linear fit, and kq is equal to KSV/τ0. The obtained curves are presented in Figure 4, and the relevant data are summarized in  [56], which indicates that fluorescence quenching occurred by a static mechanism, in which the formation of a benzazole-BSA conjugate takes place in the ground state.  Thus, given that the reaction was known to proceed via a static mechanism, the binding constant (KA) and the number of binding sites (n) between BSA and the benzazoles were also obtained by applying Equation (4) [57]:  Table 2. BSA-binding parameters for benzazoles 12-17, where Q is the quenching percentage, K SV is the Stern-Volmer quenching constant, k q is the bimolecular quenching rate constant, K a is the binding constant, n is the number of binding sites, and ∆G 0 is the Gibbs free energy for the BSA-benzazole interaction.  Thus, given that the reaction was known to proceed via a static mechanism, the binding constant (K A ) and the number of binding sites (n) between BSA and the benzazoles were also obtained by applying Equation (4) [57]:

Benzazole Q [a] (%)
where F 0 and F represent the fluorescence intensities in the absence and presence of the benzazoles, respectively, and [Q] is the concentration of the quencher (benzazole). The respective curves are shown in Figure 5. Moreover, the standard Gibbs free energy (∆G 0 ) of the benzazole:BSA conjugates was calculated from the K A values using Equation (5), where R is the gas constant (1.987 cal· K −1 ·mol −1 ) and T is the temperature (298 K). The results from the double logarithmic plot of the relationship between the fluorescence intensities of BSA and the benzazoles are summarized in Table 2, where the KA values are in the order of 10 4 M −1 , suggesting a strong interaction with BSA. The presence of different alkyl chains in the benzazoles is thought to cause variation in the binding constant values, with compounds 13 and 16 exhibiting higher values. In addition, the number of binding sites (n) for all compounds was in the range of 0.88-1.31, indicating the presence of one binding site in the interaction with BSA. The calculated Gibbs free energy did not show any trend, with values in the same order of magnitude indicating spontaneous interaction with BSA.
A BSA association study was also performed with different amounts of protein (0-12 μM in PBS) adding these solutions to the respective benzazole dye solution (2 μM in PBS). This study can be observed in Figure 6, using benzazoles 12 and 15 as models. Figure 6a-c depicts their UV-Vis absorption spectra in the absence and presence of BSA. Two distinct regions can be observed, related to the protein (250-300 nm) and the dyes (300-380 nm), depending on the benzazole. Increasing BSA raised the absorption intensity between 250-300 nm, and the dye region remained almost constant. A similar response was observed for all studied compounds (Figures S7-S10). However-and we would like to highlight these findings-based on the emission spectra (Figure 6b-d), the association was observed only for compound 12, where the fluorescence intensity increased with the BSA amount in solution. All sets of emission spectra and respective plots of fluorescence intensity maxima as a function of BSA concentration are presented in Figures S11-S15. The results from the double logarithmic plot of the relationship between the fluorescence intensities of BSA and the benzazoles are summarized in Table 2, where the K A values are in the order of 10 4 M −1 , suggesting a strong interaction with BSA. The presence of different alkyl chains in the benzazoles is thought to cause variation in the binding constant values, with compounds 13 and 16 exhibiting higher values. In addition, the number of binding sites (n) for all compounds was in the range of 0.88-1.31, indicating the presence of one binding site in the interaction with BSA. The calculated Gibbs free energy did not show any trend, with values in the same order of magnitude indicating spontaneous interaction with BSA.
A BSA association study was also performed with different amounts of protein (0-12 µM in PBS) adding these solutions to the respective benzazole dye solution (2 µM in PBS). This study can be observed in Figure 6, using benzazoles 12 and 15 as models. Figure 6a-c depicts their UV-Vis absorption spectra in the absence and presence of BSA. Two distinct regions can be observed, related to the protein (250-300 nm) and the dyes (300-380 nm), depending on the benzazole. Increasing BSA raised the absorption intensity between 250-300 nm, and the dye region remained almost constant. A similar response was observed for all studied compounds (Figures S7-S10). However-and we would like to highlight these findings-based on the emission spectra (Figure 6b-d), the association was observed only for compound 12, where the fluorescence intensity increased with the BSA amount in solution. All sets of emission spectra and respective plots of fluorescence intensity maxima as a function of BSA concentration are presented in Figures S11-S15.  Molecules 2021, 26, x FOR PEER REVIEW 9 of 16 The first addition of BSA (2 μM) increased the fluorescence intensity by around two times. After the last addition of protein (12 μM) the fluorescence increased by 3.6, and red-shifted the emission maxima 25 nm (Figure 6b). The observed bathochromic shift suggests that the BSA microenvironments are less polar than the PBS due to the hydrophobic groups present in the surface and interiors of the BSA [58]. On the other hand, the parent compound 15 presents an almost constant intensity in the presence of different amounts of BSA (Figure 6d). The additional compounds showed quite different results, sometimes increasing the intensity until reaching a maximum and then decreasing the intensity (compounds 13 and 16) or randomly varying the intensity (compound 14), and sometimes even showing fluorescence suppression (compound 17) (Figures S11-S15).
These results indicate that based on the compound's structure, the interaction with BSA from the point of view of the excited state of the dyes can be a tricky subject. The association (K) of compound 12 with BSA was also obtained by the Benesi-Hildebrand equation for 1:1 complex [59].  The first addition of BSA (2 µM) increased the fluorescence intensity by around two times. After the last addition of protein (12 µM) the fluorescence increased by 3.6, and red-shifted the emission maxima 25 nm (Figure 6b). The observed bathochromic shift suggests that the BSA microenvironments are less polar than the PBS due to the hydrophobic groups present in the surface and interiors of the BSA [58]. On the other hand, the parent compound 15 presents an almost constant intensity in the presence of different amounts of BSA (Figure 6d). The additional compounds showed quite different results, sometimes increasing the intensity until reaching a maximum and then decreasing the intensity (compounds 13 and 16) or randomly varying the intensity (compound 14), and sometimes even showing fluorescence suppression (compound 17) (Figures S11-S15).
These results indicate that based on the compound's structure, the interaction with BSA from the point of view of the excited state of the dyes can be a tricky subject. The association (K) of compound 12 with BSA was also obtained by the Benesi-Hildebrand equation for 1:1 complex [59].

Molecular Docking
To better understand the interaction between the studied compounds and BSA, and to identify the amino acid residues involved in their interactions, molecular docking simulations were performed. Initially, blind docking was performed for all protein structures. The results indicate that a possible binding site was in a region close to Trp134. BSA contains two Trp residues (Trp134 and Trp213), both of which are responsible for its fluorescence. Trp134 is located in domain I, while Trp213 is in domain II ( Figure 8) [60,61]. Second, we performed semi-flexible docking in the region close to Trp134 (with the lateral chain of residues Arg185, Lys136, Lys131, Trp134, Tyr137, and Tyr160 flexible) to improve the interactions between BSA and the compounds. The docking simulations indicate that compounds 13 and 16 bind closely to Trp134 in domain I ( Figure 8) and adopt similar binding poses and interactions (Figure 9b,e). These compounds exhibited π-π stacking between the benzene ring and the indole ring of the Trp134 residue (3.8-4.0 Å),

Molecular Docking
To better understand the interaction between the studied compounds and BSA, and to identify the amino acid residues involved in their interactions, molecular docking simulations were performed. Initially, blind docking was performed for all protein structures. The results indicate that a possible binding site was in a region close to Trp134. BSA contains two Trp residues (Trp134 and Trp213), both of which are responsible for its fluorescence. Trp134 is located in domain I, while Trp213 is in domain II (Figure 8) [60,61].

Molecular Docking
To better understand the interaction between the studied compounds and BSA, and to identify the amino acid residues involved in their interactions, molecular docking simulations were performed. Initially, blind docking was performed for all protein structures. The results indicate that a possible binding site was in a region close to Trp134. BSA contains two Trp residues (Trp134 and Trp213), both of which are responsible for its fluorescence. Trp134 is located in domain I, while Trp213 is in domain II ( Figure 8) [60,61]. Second, we performed semi-flexible docking in the region close to Trp134 (with the lateral chain of residues Arg185, Lys136, Lys131, Trp134, Tyr137, and Tyr160 flexible) to improve the interactions between BSA and the compounds. The docking simulations indicate that compounds 13 and 16 bind closely to Trp134 in domain I ( Figure 8) and adopt similar binding poses and interactions (Figure 9b,e). These compounds exhibited π-π stacking between the benzene ring and the indole ring of the Trp134 residue (3.8-4.0 Å),  (Figure 9b,e). These compounds exhibited π-π stacking between the benzene ring and the indole ring of the Trp134 residue (3.8-4.0 Å), hydrophobic interactions between the methyl group and the imidazole ring from the His18 residue and between the carbon chain of the compounds and the carbon chain from the Lys132 residue. They have hydrogen bonds (H-bonds) between the amine moiety and the carboxylic group of the Glu17 residue and between the hydroxyl moiety and the carbonyl group of the Asn158 residue. There is a π-H bond between the benzoxazole (or benzothiazole) group and the amide moiety of the Asn161 residue and a cation-π interaction between the benzene ring of compound 16 and the ammonium group of the Lys131 residue. These direct interactions between the compounds and the indole ring could influence the Trp fluorescence quenching.
Molecules 2021, 26, x FOR PEER REVIEW 11 of 16 hydrophobic interactions between the methyl group and the imidazole ring from the His18 residue and between the carbon chain of the compounds and the carbon chain from the Lys132 residue. They have hydrogen bonds (H-bonds) between the amine moiety and the carboxylic group of the Glu17 residue and between the hydroxyl moiety and the carbonyl group of the Asn158 residue. There is a π-H bond between the benzoxazole (or benzothiazole) group and the amide moiety of the Asn161 residue and a cation-π interaction between the benzene ring of compound 16 and the ammonium group of the Lys131 residue. These direct interactions between the compounds and the indole ring could influence the Trp fluorescence quenching. However, compounds 12, 14, 15, and 17 were bound between domains I and III ( Figure 8) and did not directly interact with Trp134 ( Figure 9). In general, these molecules interacted with the Arg458 residue via cation-π interactions and hydrophobic interactions between the benzoxazole (or benzothiazole) group and the lateral chain of the Ala193 residue and between the methyl moiety and the His145 and Pro146 residues. In addition, we observed that the alkyl chains of compounds 14 and 17 were located in different positions, while the carbon chain of 14 interacted with the Phe133 and Lys136 residues, and the carbon chain of 17 exhibited intramolecular hydrophobic interactions (Figure 9c,f). Previous studies have described this region as a BSA-binding site [63].
These observations are consistent with other studies where π-π interactions and H-bonds are involved in BSA-ligand complex stabilization [64]. In addition, the predicted thermodynamic data (∆Gbind) obtained by molecular docking, differently than observed experimentally, demonstrated that the compounds with short carbon chains (12 and 15) exhibited the lowest ∆Gbind, indicating a more spontaneous interaction with BSA compared with those of other molecules (Table 3). ∆G bind (kcal·mol −1 ) -8.0 -6.8 -6.5 -8.0 -6.9 -6.4 However, compounds 12, 14, 15, and 17 were bound between domains I and III ( Figure 8) and did not directly interact with Trp134 ( Figure 9). In general, these molecules interacted with the Arg458 residue via cation-π interactions and hydrophobic interactions between the benzoxazole (or benzothiazole) group and the lateral chain of the Ala193 residue and between the methyl moiety and the His145 and Pro146 residues. In addition, we observed that the alkyl chains of compounds 14 and 17 were located in different positions, while the carbon chain of 14 interacted with the Phe133 and Lys136 residues, and the carbon chain of 17 exhibited intramolecular hydrophobic interactions (Figure 9c,f). Previous studies have described this region as a BSA-binding site [63].

Photophysical Characterization
These observations are consistent with other studies where π-π interactions and Hbonds are involved in BSA-ligand complex stabilization [64]. In addition, the predicted thermodynamic data (∆G bind ) obtained by molecular docking, differently than observed experimentally, demonstrated that the compounds with short carbon chains (12 and 15) exhibited the lowest ∆G bind , indicating a more spontaneous interaction with BSA compared with those of other molecules (Table 3).

Photophysical Characterization
Spectroscopic grade solvents were used for fluorescence and UV-Vis absorption spectroscopies. UV-Vis absorption spectroscopy of 10 −5 M solutions was performed on a Shimadzu UV-2450 spectrophotometer (Shimadzu Corporation, Tokyo, Japan). Steadystate fluorescence spectroscopy was performed using a Shimadzu spectrofluorometer (RF-5301PC). The fluorescence quantum yields (Φ FL ) were measured at the optical dilute regime with solutions with an absorbance intensity lower than 0.05 and using quinine sulfate (Riedel-de-Haën, Seelze, Germany) in H 2 SO 4 1N as the quantum yield standard (Φ F = 0.55) [65].
To study the interaction modes of BSA with the benzazoles, a stock solution of BSA (2.0 mg·mL −1 ) with a concentration of 31 µM was initially prepared in PBS at pH 7.2. To analyze the fluorescence quenching of BSA, solutions were prepared with a final volume of 3 mL with a fixed BSA concentration of 11 µM in PBS, pH 7.2. Benzazoles were added to these solutions at different concentrations (0-20 µM) from a stock solution in dimethylformamide. UV-Vis absorption and fluorescence emission spectra were obtained at 25 • C. The emission curves were obtained under excitation at 277 nm, relative to the absorption wavelength of BSA, using Exc/Em slits of 5.0/5.0 nm, respectively. The BSA association study was performed keeping the dye concentration constant (2 µM in PBS). In this study, different amounts of previously prepared BSA solutions (0-10 µM in PBS, pH 7.2) were added. The final solution was kept to rest for 1 h. The fluorescence spectra were obtained at 25 • C and under the excitation of each absorption maxima.

Molecular Docking
The 3D structure of BSA was obtained from the Protein Data Bank (http://www. rcsb.org/pdb, ID: 4f5s, accessed 1 October 2019) [66]. The Chimera 1.8 software [67] was used to remove chain B, water, and other molecules, and add hydrogen atoms to the BSA protein. The ligands were built in the software Avogadro 1.1.1 [68], following semi-empirical PM6 [69] geometry optimization using the program MOPAC2012 [70]. The ligands and proteins in the pdbqt format were generated by AutoDockTools, where the ligands were considered flexible (with PM6 charges) and the enzyme rigid (with Gasteiger charges) [71]. The AutoDock Vina 1.1.1 program was used for blind docking [72], with a gridbox of 92 × 62 × 86 and the coordinates x = 9.457, y = 23.359, and z = 98.149 (exhaustiveness of 150). Semi-flexible docking was performed for the region surrounding the Trp134 residue, with a gridbox of 30 × 30 × 35 and the coordinates x = 20.324, y = 33.690, and z = 97.801 (exhaustiveness of 150). The lateral chain of the residues Arg185, Lys136, Lys131, Trp134, Tyr137, and Tyr160, was considered flexible during docking to improve the interactions between compounds and proteins. As a model for the binding pose, the compound conformer with the lowest binding energy (∆G) was selected from the semi-flexible docking experiment. The docking results were analyzed using the Accelrys Discovery Studio 3.5 software [73].

Conclusions
In summary, we report here the photophysical characterization of ESIPT fluorescent lipophilic benzazoles and their binding affinity towards BSA by BSA fluorescence quenching and molecular docking. In PBS solution (pH 7.2), these compounds produce absorption maxima in the UV region related to 1 π-π* electronic transitions. The derivatives with short alkyl chains exhibited absorption maxima at higher energies, indicating that the alkyl chain influences the electronic properties in the ground state. In addition, compounds with longer alkyl chains demonstrated broad absorption in the visible region, which could be related to the aggregation of these compounds. The benzazoles are fluorescent, with the main emission band located in the blue-green regions with a large Stokes shift. The nature of the emission bands seems to be different depending on the hydrophobic portion of the molecule, once again indicating that the alkyl chains influence their photophysics. The fluorescence quenching of BSA in the presence of the benzazoles demonstrated a significant decrease in the fluorescence intensity without a significant red-shift. The k q values of approximately 10 12 M −1 ·s −1 indicate that the interactions proceed according to a static mechanism. In addition, higher association constant values (K A~1 0 4 M −1 ) were obtained from the double logarithmic plot relating the fluorescence intensities from BSA and the benzazoles, suggesting a strong interaction with BSA. Association with BSA could be observed only for benzazole 12. The docking simulations indicate that compounds 13 and 16 bind closely to Trp134 in domain I and that the additional compounds were bound between domains I and III, and did not directly interact with Trp134. Finally, the obtained binding parameters indicate that all the studied benzazoles could be efficiently transported and biodistributed by BSA in the bloodstream.

Conflicts of Interest:
The authors declare no conflict of interest.