In Vitro and In Silico Evaluation of Indole-Bearing Squaraine Dyes as Potential Human Serum Albumin Fluorescent Probes

Abstract

The quantitative determination of proteins is an important parameter in biochemistry, biotechnology and immunodiagnostics, and the importance of serum albumin in clinical diagnosis should be highlighted, given that alterations in its concentration are generally associated with certain diseases. As possible probes for this purpose, squaraine dyes have been arousing the interest of many researchers due to their unique properties, such as absorption in the visible spectra, moderate relative fluorescence quantum yields and increased fluorescence intensity after non-covalent binding to specific ligands. In this work, five squaraine dyes, four of which have never been reported in the literature, were characterized and evaluated in vitro and in silico concerning their potential application as fluorescent probes for human serum albumin detection. After interaction with the protein, the fluorescence intensity increased from 12 to 41 times, depending on the dye under study. High sensitivity (1.0 × 10⁵–5.4 × 10⁵ nM), low detection limits (168–352 nM) and moderate quantitation limits (560–1172 nM) were obtained, proving the efficiency of the method. In addition, moderate-to-excellent selectivity was observed compared to γ-globulin proteins. Molecular docking suggests that the dyes interact more effectively with the Sudlow site I, and binding energies have been markedly higher than those of warfarin, a molecule known to bind to this site specifically.

1. Introduction

Human serum albumin (HSA) is the most abundant protein in the human circulatory system, is synthesized in the liver, and represents about 52% of the blood’s composition [1]. Due to their high bioavailability, low cost, stability, and ability to transport different ligands to specific sites, these proteins have been the most widely studied over the last few years [2,3]. Within its main functions, the transport of various macromolecules such as hormones, lipids, metal ions, amino acids, and some drugs in the body [4] is the most relevant. Thus, albumin plays an essential role in maintaining metabolic processes, such as determining oncotic plasma pressure, decreasing some toxins’ activity, and controlling the plasma’s antioxidant properties [5,6].

Quantifying proteins in biological fluids has shown great interest in several areas, such as biochemistry, biotechnology, and immunodiagnostics [7]. Since serum albumins are the main soluble protein constituents of the circulatory system, their quantification in biomedical diagnostics is a fundamental parameter in assessing a patient’s health status [8]. For example, changes in blood albumin levels can be used as a biomarker at an early stage of various disorders, such as of a cardiovascular origin, liver disease, nephrotic syndrome, diabetes, and severe dehydration [1,9]. Currently, there are several protein-detection methods, namely, colorimetric techniques, such as the Lowry and Bradford methods, immunoassays, cyclic voltammetry, and chemiluminescence. However, these more traditional methods present certain limitations, such as a long reaction time and the need for specific equipment and specialized personnel, increasing the cost of the processes and even creating complicated sample preparation protocols [10,11].

The development of new methods that can overcome these disadvantages and offer a faster, low-cost, and non-invasive evaluation has been an area of great interest. One of the alternatives is fluorescence spectroscopy, which has several advantages, such as high sensitivity, selectivity, simplicity, and compatibility with live cells and physiological assays, and allows for the real-time detection of proteins [12,13,14]. Several detection methods based on fluorescence have been developed in clinical and biological research [11], and several fluorescent dyes are commercially available, such as ethidium bromide, eosin B and eosin Y, albumin blue, and Coomassie Brilliant Blue [1].

Several fluorescent probes developed in recent years for the detection of serum albumins in vitro or in vivo have a low excitation wavelength (<500 nm), an emission wavelength < 600 nm, and a detection limit >30 mg/L, also showing photodegradation and interference caused by the autofluorescence of biomolecules, which limits their biological application in complex body fluids. Thus, it has been increasingly important to develop fluorescent probes that absorb in the red and near-infrared (NIR) region (>650 nm), which have high selectivity, sensitivity, stability, and a low detection limit for the recognition of serum albumins in biological samples [15,16]. In 2020, as an example, Aristova et al. [17] reported a study in which benzothiazole-derived pentamethine cyanines were evaluated for their potential application as fluorescent probes for serum albumins. These dyes exhibited maximum absorption near 600 nm and improved interaction with serum albumins compared to other globular proteins.

In addition to other biomedical applications, squaraine dyes have been used in the research on sensors to detect different types of analytes. Due to its strong absorption and emissions, it was possible to develop selective fluorescent probes for metal ions, explosives (such as picric acid [18]), and other important molecules. In the literature, there are a variety of dyes capable of selectively detecting some ions such as mercury [19,20], iron [21,22], copper [23,24], zinc [25,26], lead [27], sodium [28], silver [29], magnesium [30], and chloride [31,32]. The detection of metal ions has proved to be of great importance since they are not biodegradable and tend to accumulate in living organisms, and some heavy metals are known to be toxic and carcinogenic [33]. Some dyes are susceptible to the hydroxide ion and function as pH sensors [34,35].

In the recent literature, there are studies by Butnarasu et al. [36,37] where the authors evaluated the interaction of squaraine dyes with some proteins such as gastric mucins, transferrin, fibrinogen, trypsin, pepsin, and a gastric protease, to understand better the behavior of these compounds in the presence of different proteins and obtain helpful information in the development of new, more effective, and selective fluorescent dyes. In addition, intracellular biothiols, such as glutathione [38,39], cysteine, and homocysteine [39,40], are other components of biological systems which can be detected through the use of squaraines. Furthermore, since alterations in the levels of these components are strongly associated with diseases, such as Alzheimer’s, liver damage [41], cardiovascular diseases, and neural tube defects [42], their detection is also of great importance.

In this work, the potential application of squaraine dyes derived from 2,3,3-trimethylindolenine and 1,1,2-trimethylbenz[e]indole as fluorescent probes for HSA detection is evaluated. The four-membered central ring introduction of diethanolamino, picolyl-, and dipicolylamino groups was studied to understand if these functional groups improve their interaction with this protein compared to unsubstituted squaraines. This work is also comparative with analogous molecules whose albumin interaction results were previously reported in the scope of their potential as photosensitizers for photodynamic therapy (Figure 1) [43]. Thus, four new squaraine dyes similar to those previously reported were prepared and properly characterized to make this comparative study possible. The absorption of these dyes, emission, and relative fluorescence quantum yields were determined for the prepared dyes in several organic solvents. The interaction of the prepared squaraines with HSA was evaluated in vitro by drawing fluorescence spectra at increasing concentrations of this biomolecule. The relative fluorescence quantum yields in the buffer, with and without protein, method sensitivity, and limits of detection and quantification were calculated. Computational studies were performed to determine the sites with which the squaraines have the most significant affinity and the amino acid residues with which they may interact.

2. Materials and Methods

2.1. Instruments and Methods

All reagents and starting materials were of analytical grade and were used as received unless otherwise specified. The dichloromethane used in reactions was previously dried. Reactions under heating and/or magnetic stirring were executed using Velp Scientifica heating magnetic stirrers. The reaction monitoring and assessment of the purity of the synthesized compounds were performed by thin-layer chromatography using Merck 60 F₂₅₄ aluminum sheets coated with 0.25 mm silica gel. After elution in a mixture of 2–10% dichloromethane/methanol, the plates were controlled by the naked eye and, when required, visualized under 254 or 365 nm ultraviolet radiation.

Melting points (M.p.) were determined in a URA Technic hot plate binocular microscope apparatus and are uncorrected. After melting, the residue was eluted in proper solvent and compared chromatographically with the initial crystals to observe if melting led to decomposition. If so, the compounds’ characterization describes the abbreviation “dec.”. Ground-state visible (Vis) absorption spectra were recorded on a Perkin Elmer Lambda 25 spectrophotometer using an Hellma Analytics 10-mm-path-length two-sided clear quartz cuvette at room temperature with a slit width of 1 nm. Absorbance spectra of the dyes under evaluation were plotted in dimethyl sulfoxide (DMSO), dimethylformamide (DMF), chloroform (CFM), and phosphate buffer (PB). The maximum absorption wavenumbers are described in nm, and the molar absorptivity coefficients, slope of the line plotting the absorbance intensities of five dye’s concentration-known solutions versus the sample concentrations, are presented in M⁻¹⋅cm⁻¹. All fluorescence measurements were performed using a Varian Cary Eclipse fluorescence spectrophotometer and an Hellma Analytics 10 mm path length four-sided clear fluorescence/ultraviolet quartz cuvette. Fourier Transform Infrared spectra were measured on a Shimadzu IRAffinity spectrophotometer by means of Shimadzu LabSolutions IR software as KBr pellets. Sample spectra were recorded at room temperature in the range of 4000–500 cm⁻¹, with a spectral resolution of ±4 cm⁻¹. Bands’ wavenumber values (ʋ_max) are described in cm⁻¹, and assignments were made by indicating the strength of the vibration as being of weak (w), medium (m), or strong (s) intensity. The nuclear magnetic resonance of proton (¹H NMR) and of carbon-13 (¹³C NMR) spectra were recorded at 298.15 K in deuterochloroform (CDCl₃) or hexadeuterodimethyl sulfoxide (DMSO-d₆) on a Bruker NMR Avance III 400, observing ¹H at 400.13 MHz and ¹³C at 100.63 MHz, or on a Bruker NMR Avance III 600 spectrometer, observing ¹H at 600.13 MHz and ¹³C at 150.90 MHz. Chemical shifts (δ) are reported in parts per million (ppm) relative to tetramethylsilane or residual solvent signals. Coupling constants (J) are reported in Hertz. Proton splittings are defined as a singlet (s), broad singlet (br s), doublet (d), broad doublet (br d), triplet (t), broad triplet (br t), double triplet (dt), quintet (qt), or multiplet (m). Carbon-13 assignments were based on distortionless enhancement by polarization transfer 135° (DEPT 135) spectra. High-resolution electrospray ionization time-of-flight mass spectrometry (HRESI-TOFMS) was performed using a microTOF Bruker Daltonics spectrometer at CACTI (University of Vigo).

2.2. Squaraines’ Synthesis

2.2.1. 2-{2-[Bis(2-hydroxyethyl)amino]-3-(1-hexyl-3,3-dimethyl-3H-indolin-2-ylidenemethyl)-4-oxocyclobut-2-enylidenemethyl}-1-hexyl-3,3-dimethylindol-1-ium iodide (11a)

The title compound was obtained by reacting 0.25 g (0.34 mmol) of O-methylated squaraine 7a with 50 µL (0.51 mmol) of diethanolamine (8) in dry dichloromethane (25 mL) at room temperature under a nitrogen atmosphere, stirring for 1h. After removing the solvent to dryness on the rotary evaporator, 10 mL of methanol and 10 mL of 14% potassium iodide aqueous solution were added to the resulting residue. The mixture was allowed to stir at room temperature for 2 h. After decanting, the resulting residue was washed with distilled water, crushed with diethyl ether, and dried under vacuum to obtain bright green crystals. Yield: 41%. M.p.: 160–163 °C (dec.). Vis λ_max (DMF): 669 nm; Vis λ_max (DMSO): 673 nm; Vis λ_max (PB): 670 nm; Vis λ_max (CHCl₃): 667 nm, ε = 1.9 × 10⁵ M⁻¹·cm⁻¹. IR ʋ_max (KBr): 3333 (w, OH), 2951 (w, CH), 2924 (w, CH), 2859 (w, CH), 1742 (w, C=O), 1682 (w, C=C), 1539 (w), 1493 (m, ArC=C), 1479 (m), 1454 (m), 1435 (m), 1402 (w), 1360 (w), 1312 (m), 1287 (m), 1234 (w), 1225 (w), 1173 (m), 1105 (m), 1051 (w), 1020 (w), 989 (w), 924 (w), 824 (w), 802 (w), 752 (w), 671 (w) cm⁻¹. ¹H NMR (DMSO-d₆, 600 MHz) δ: 7.57 (2H, d, J = 7.2, ArH), 7.46 (2H, d, J = 7.8, ArH), 7.41 (2H, t, J = 7.5, ArH), 7.26 (2H, t, J = 7.5, ArH), 6.19 (2H, s, C=CH), 5.31 (2H, t, J = 4.8, N(CH₂)₂OH, exchanges with D₂O), 4.17 (4H, t, J = 7.2, NCH₂(CH₂)₄CH₃), 3.89 (4H, t, J = 5.1, NCH₂CH₂OH), 3.85–3.82 (4H, m, NCH₂CH₂OH, collapses into a triplet with D₂O), 1.71–1.66 (16H, s + qt, C(CH₃)₂ + NCH₂CH₂(CH₂)₃CH₃), 1.35 (4H, qt, J = 7.8, N(CH₂)₂CH₂(CH₂)₂CH₃), 1.28–1.24 (8H, m, N(CH₂)₃(CH₂)₂CH₃), 0.84 (6H, t, J = 6.9, N(CH₂)₅CH₃) ppm. ¹³C NMR (DMSO-d₆, 150.90 MHz) δ: 174.23, 172.13, 167.91, 158.60, 141.72, 141.62, 128.17 (ArCH), 124.78 (ArCH), 122.30 (ArCH), 111.41 (ArCH), 89.19 (C=CH), 58.58 (NCH₂CH₂OH), 53.29 (NCH₂CH₂OH), 49.39 (C(CH₃)₂), 44.27 (NCH₂ (CH₂)₄CH₃), 30.88 (CH₂), 26.71 (CH₂), 25.93 (C(CH₃)₂), 25.69 (CH₂), 21.91 (CH₂), 13.78 (CH₃) ppm. HRESI-TOFMS m/z: 652.4479 [M–I]⁺ (C₄₂H₅₈N₃O₃ calc. 652.4473).

2.2.2. 2-{2-[Bis(2-hydroxyethyl)amino]-3-(3-hexyl-1,1-dimethyl-2H-benzo[e]indol-2-ylidenemethyl)-4-oxocyclobut-2-enylidenemethyl}-3-hexyl-1,1-dimethyl-1H-benzo[e]indol-3-ium iodide (11b)

The title compound was obtained by reacting 0.25 g (0.30 mmol) of O-methylated squaraine 7b with 87 μL (0.90 mmol) of diethanolamine (8) in dry dichloromethane (25 mL) at room temperature under a nitrogen atmosphere, stirring for 1 h 30 min. After removing the solvent to dryness on the rotary evaporator, 10 mL of methanol and 10 mL of 14% potassium iodide aqueous solution were added to the resulting residue. The mixture was stirred at room temperature for 2 h. After decanting, the resulting residue was washed with distilled water, crushed with diethyl ether, and dried under a vacuum to obtain yellowish-green crystals. Yield: 70%. M.p.: 114–117 ˚C. Vis λ_max (DMF): 703 nm; Vis λ_max (DMSO): 707 nm; Vis λ_max (PB): 705 nm; Vis λ_max (CHCl₃): 702 nm, ε = 1.7 × 10⁵ M⁻¹·cm⁻¹. IR ʋ_max (KBr): 3335 (w, OH), 2951 (w, CH), 2928 (m, CH), 2857 (w, CH), 1738 (w, C=O), 1622 (m, C=C), 1549 (m, ArC=C), 1520 (w), 1485 (s), 1454 (m), 1433 (m), 1337 (m), 1296 (m), 1246 (m), 1207 (w), 1180 (m), 1117 (m), 1092 (w), 1053 (w), 1015 (w), 986 (w), 934 (m), 891 (w), 806 (m), 746 (w), 671 (m) cm⁻¹. ¹H NMR (DMSO-d₆, 600 MHz) δ: 8.28 (2H, d, J = 8.4, ArH), 8.08 (2H, d, J = 8.4, ArH), 8.06 (2H, d, J = 8.4, ArH), 7.80 (2H, d, J = 8.4, ArH), 7.65 (2H, t, J = 7.8 ArH), 7.51 (2H, t, J = 7.5, ArH), 6.26 (2H, s, C=CH), 5.35 (2H, t, J = 5.1, N(CH₂)₂OH, exchange with D₂O), 4.32 (4H, t, J = 7.2, NCH₂(CH₂)₄CH₃), 3.94 (4H, t, J = 5.1, NCH₂CH₂OH), 3.88–3.86 (4H, m, NCH₂CH₂OH), 1.93 (12H, s, C(CH₃)₂), 1.74 (4H, qt, J = 7.2, NCH₂CH₂(CH₂)₃CH₃), 1.38 (4H, qt, J = 7.7, N(CH₂)₂CH₂ (CH₂)₂CH₃), 1.31–1.22 (8H, m, N(CH₂)₃(CH₂)₂CH₃), 0.83 (6H, t, J = 6.9, N(CH₂)₅CH₃) ppm. ¹³C NMR (DMSO-d₆, 150.90 MHz) δ: 174.67, 173.45, 167.81, 157.72, 139.28, 134.10, 131.30, 129.99 (ArCH), 129.74 (ArCH), 127.66 (ArCH), 124.86 (ArCH), 122.53 (ArCH), 111.85 (ArCH), 88.88 (C=CH), 58.69 (NCH₂CH₂OH), 53.37 (NCH₂CH₂OH), 51.12 (C(CH₃)₂), 44.63 (NCH₂ (CH₂)₄CH₃), 30.93 (CH₂), 27.08 (CH₂), 25.81 (C(CH₃)₂), 25.68 (CH₂), 21.94 (CH₂), 13.81 (CH₃) ppm. HRESI-TOFMS m/z: 752.4783 [M–I]⁺ (C₅₀H₆₂N₃O₃ calc. 752.4786).

2.2.3. 1-Hexyl-2-[3-(1-hexyl-3,3-dimethyl-3H-indolin-2-ylidenemethyl)-4-oxo-2-(pyridin-2-ylmethylamino)cyclobut-2-enylidenemethyl]-3,3-dimethylindol-1-ium iodide (12a)

The title compound was obtained by reacting 1.80 g (2.40 mmol) of O-methylated squaraine 7a with 800 µL (7.3 mmol) of 2-picolylamine (9) in dry dichloromethane (50 mL) at room temperature under a nitrogen atmosphere, stirring for 4 h. After removing the solvent to dryness on the rotary evaporator, 25 mL of methanol and 25 mL of 14% potassium iodide aqueous solution were added to the resulting residue. The mixture was stirred at room temperature for 2 h. After decanting, the resulting residue, which showed chromatographically to be a complex mixture of several compounds, was washed with distilled water, crushed with diethyl ether, and dried under a vacuum to obtain bright-green crystals. Yield: 12%. M.p.: 203–204 °C (dec.). Vis λ_max (DMF): 656 nm; Vis λ_max (DMSO): 660 nm; Vis λ_max (PB): 657 nm, Vis λ_max (CHCl₃): 657 nm, ε = 3.0 × 10⁵ M⁻¹·cm⁻¹. IR ʋ_max (KBr): 3447 (w, NH), 3092 (w, ArCH), 2955 (m, CH), 2926 (m, CH), 2857 (m, CH), 1632 (s, C=O), 1609 (w, C=C), 1566 (m, ArC=C), 1495 (s), 1462 (s), 1402 (w), 1358 (m), 1294 (m), 1238 (w), 1169 (m), 1130 (w), 1113 (m), 1051 (w), 1020 (w), 924 (w), 833 (m), 797 (w), 748 (m), 673 (w) cm⁻¹. ¹H NMR (CDCl₃, 600 MHz) δ: 10.10 (1H, br s, NH, exchange with D₂O), 8.52–8.51 (1H, m, ArH), 8.01 (1H, d, J = 7.8, ArH), 7.78 (1H, dt, J = 7.5, 1.8, ArH), 7.37 (2H, br s, ArH), 7.32 (1H, br s, ArH), 7.24 (1H, br d, J = 6.0, ArH), 7.23 (1H, br d, J = 5.4, ArH), 7.16 (2H, br s, ArH), 7.00 (1H, br s, ArH), 6.67 (1H, s, C=CH), 6.22 (1H, s, C=CH), 5.07 (2H, d, J = 6.6, NHCH₂C₅H₄N), 4.59 (2H, br s, NCH₂(CH₂)₄CH₃), 4.10 (2H, br s, NCH₂(CH₂)₄CH₃), 1.86 (2H, br s, NCH₂CH₂(CH₂)₃CH₃), 1.76 (6H, br s, C(CH₃)₂), 1.70 (6H, br s, C(CH₃)₂), 1.55 (2H, br s, N(CH₂)₂CH₂(CH₂)₂CH₃), 1.37–1.25 (12H, m, N(CH₂)₂CH₂(CH₂)₂CH₃ + N(CH₂)₃(CH₂)₂CH₃), 0.85 (6H, t, J = 6.9, N(CH₂)₅CH₃) ppm. ¹H NMR (CDCl₃ + D₂O, 600 MHz) δ: 8.50 (1H, d, J = 4.8, ArH), 8.01 (1H, d, J = 7.8, ArH), 7.75 (1H, dt, J = 7.5, 1.8, ArH), 7.37 (2H, br t, J = 6.0, ArH), 7.32 (1H, d, J = 7.2, ArH), 7.31 (1H, t, J = 7.2, ArH), 7.24 (1H, t, J = 6.6, ArH), 7.23–7.21 (1H, m, ArH), 7.15 (1H, t, J = 7.2, ArH), 7.14 (1H, d, J = 7.8, ArH), 6.99 (1H, d, J = 7.8, ArH), 6.65 (1H, s, C=CH), 6.22 (1H, s, C=CH), 5.04 (2H, s, NHCH₂C₅H₄N), 4.59 (2H, t, J = 6.9, NCH₂(CH₂)₄CH₃), 4.08 (2H, t, J = 7.2, NCH₂(CH₂)₄CH₃), 1.88 (2H, qt, J = 7.4, NCH₂CH₂(CH₂)₃CH₃), 1.75 (6H, s, C(CH₃)₂), 1.72–1.69 (8H, s + m, C(CH₃)₂ + NCH₂CH₂(CH₂)₃CH₃), 1,55 (2H, qt, J = 8.0, N(CH₂)₂CH₂(CH₂)₂CH₃), 1.37–1.24 (10H, m, NCH₂(CH₂)₃CH₃ + N(CH₂)₃(CH₂)₂CH₃), 0.84 (6H, br t, J = 6.6, N(CH₂)₅CH₃) ppm. ¹³C NMR (CDCl₃, 150.90 MHz) δ: 174.27, 173.57, 171.04, 168.76, 162.16, 157.84, 148.16 (ArCH), 142.93, 142.08, 141.97, 141.79, 137.75 (ArCH), 128.10 (ArCH), 127.87 (ArCH), 125.35 (ArCH), 124.26 (ArCH), 124.05, 122.89 (ArCH), 122.09 (ArCH), 111.22 (ArCH), 109.80 (ArCH), 90.80 (C=CH), 88.92 (C=CH), 50.03 (C(CH₃)₂), 49.21 (C(CH₃)₂), 48.27 (NHCH₂C₅H₄N), 45.22 (NCH₂(CH₂)₄CH₃), 43.81 (NCH₂(CH₂)₄CH₃), 31.55 (CH₂), 31.48 (CH₂), 27.80 (CH₂), 27.15 (CH₂), 26.73 (C(CH₃)₂), 26.34 (CH₂), 26.06 (C(CH₃)₂), 22.35 (CH₂), 13.95 (CH₃), 13.88 (CH₃) ppm. HRESI-TOFMS m/z: 655.4374 [M–I]⁺ (C₄₄H₅₅N₄O calc. 655.4370).

2.2.4. 2-[2-Bis(pyridin-2-ylmethyl)amino-3-(1-hexyl-3,3-dimethyl-3H-indolin-2-ylidenemethyl)-4-oxocyclobut-2-enylidenemethyl]-1-hexyl-3,3-dimethylindol-1-ium iodide (13a)

The title compound was obtained by reacting 0.25 g (0.34 mmol) of O-methylated squaraine 7a with 245 µL (1.4 mmol) of di-(2-picolyl)amine (10) in dry dichloromethane (50 mL) at room temperature under a nitrogen atmosphere, stirring for 24 h. After removing the solvent to dryness on the rotary evaporator, 5 mL of methanol and 5 mL of 14% potassium iodide aqueous solution were added to the resulting residue. The mixture was stirred at room temperature for 1 h. After decanting, the resulting residue was washed several times with distilled water, crushed with diethyl ether and petroleum ether, and dried under vacuum to obtain dark violet crystals. Yield: 37%. M.p.: 82–84 °C. Vis λ_max (DMF): 665 nm; Vis λ_max (DMSO): 668 nm; Vis λ_max (PB): 665 nm; Vis λ_max (CHCl₃): 664 nm, ε = 1.5 × 10⁵ M⁻¹·cm⁻¹. IR ʋ_max (KBr): 3048 (w, ArCH), 2953 (m, CH), 2826 (m, CH), 2857 (w, CH), 1744 (w, C=O), 1626 (w, C=C), 1543 (m, ArC=C), 1491 (s, ArC=C), 1479 (s), 1454 (s), 1435 (s), 1361 (m), 1312 (s), 1287 (s), 1238 (w), 1223 (w), 1175 (m), 1128 (w), 1105 (s), 1049 (w), 1020 (w), 984 (w), 922 (w), 754 (w), 681 (w) cm⁻¹; ¹H NMR (DMSO-d₆, 400 MHz) δ: 8.66 (2H, br s, ArH), 7.87 (2H, br t, J = 6.6, ArH), 7.56 (4H, d, J = 6.4, ArH), 7.43–7.35 (6H, m, ArH), 7.25 (2H, br s, ArH), 5.99 (2H, s, C=CH), 5.16 (4H, s, NCH₂C₅H₄N), 3.85 (4H, br t, J = 6.2, NCH₂(CH₂)₄CH₃), 1.63 (12H, s, C(CH₃)₂), 1.32 (4H, qt, J = 7.3, NCH₂CH₂(CH₂)₃CH₃), 1,16–1,05 (8H, m, N(CH₂)₃(CH₂)₂CH₃), 0.95 (4H, qt, J = 7.2, N(CH₂)₂CH₂(CH₂)₂CH₃), 0.77 (6H, br t, J = 6.2, N(CH₂)₅CH₃) ppm. ¹³C NMR (DMSO-d₆, 150.90 MHz) δ: 173.88, 171.70, 168.05, 158.27, 154.48, 149.14 (ArCH), 141.07, 140.93, 136.77 (ArCH), 127.56 (ArCH), 124.31 (ArCH), 122.61 (ArCH), 121.95 (ArCH), 121.67 (ArCH), 110.82 (ArCH), 88.13 (C=CH), 56.03 (N(CH₂C₅H₄N)₂), 48.86 (C(CH₃)₂), 43.45 (NCH₂(CH₂)₄CH₃), 30.10 (CH₂), 25.87 (CH₂), 25.28 (C(CH₃)), 24.84 (CH₂), 21.19 (CH₂), 13.08 (CH₃) ppm. HRESI-TOFMS m/z: 746.4784 [M–I]⁺ (C₅₀H₆₀N₅O calc. 746.4792).

2.3. Fluorescence Measurements

Squaraine dyes’ stock solutions were prepared at 6.7 × 10⁻⁴ M, dissolving each dye in analytical-grade DMF. Human serum protein stock solution was prepared at 14.0 × 10⁻⁴ M in PB (0.05 M, pH 7.4). All fluorometric measurements were performed in triplicate.

2.3.1. Relative Fluorescence Quantum Yields Calculations

Relative fluorescence quantum yields (Φ_F) were determined by the comparison of the integrated spectra area of each squaraine dye with the zinc phthalocyanine standard reference area, with a quantum yield for DMF of 0.17 and DMSO of 0.20, using the following equation (Equation (1); [44]):

$${\Phi }_{\mathrm{F}}={\Phi }_{\mathrm{F}\left(\mathrm{Std}\right)}\left(\frac{{\mathrm{Grad}}_{\mathrm{SqD}}}{{\mathrm{Grad}}_{\mathrm{Std}}}\right) \left(\frac{{\mathsf{\eta }}_{\mathrm{SqD}}^2}{{\mathsf{\eta }}_{\mathrm{Std}}^2}\right)$$

(1)

where the subscripts “Std” and “SqD” refer to standard and squaraine dye, respectively, “Grad” to the gradient from the plot-integrated fluorescence versus absorbance, and “η” to the refractive index of the solvent. Fluorescence spectra were displayed, exciting the compounds in varying wavelengths according to the absorption spectra of the squaraine dye in the organic solvent under study, with an excitation and emission slit width of 5 nm. The emission spectra areas of the standard dye and the squaraine dye were obtained using fluorometric measurements at the same excitation wavelength and slit width.

2.3.2. Squaraine–Protein Interaction Evaluation

Due to the insoluble character of squaraine dyes in aqueous media, compounds’ solutions for the protein–dye interaction assays were prepared in DMF and then diluted in PB. To this solution, HSA protein stock solution was added, where the squaraine concentrations were maintained (2.0 µM), and the protein concentrations varied from 0 to 3.0 µM (0, 0.05, 0.1, 0.2, 0.4, 0.6, 0.8, 1.0, 1.5, 2.0, 3.0 µM). The fluorescence intensity of the final solutions was read at t = 0, 1, 2, and 3 h of incubation with the protein, using the excitation wavelength and excitation and emission slit widths depending on each dye (

Table 1. Excitation wavelength and excitation/emission slit width values used for squaraine–protein binding fluorescence intensity measurements.

Squaraine Dye	Excitation Wavelength (nm)	Excitation/Emission Slits (nm)
5a	590	5/10
11a	630	10/10
11b	670	10/10
12a	595	10/10
13a	620	10/10

). The excitation wavelength of each dye was adjusted concerning their maximum absorption since the Stokes shifts are not large enough for us to be able to excite them to the maximum absorption values.

In both the absence and presence of proteins, relative fluorescence quantum yields were calculated as indicated in Section 2.3.1. In the latter case, the emission spectra at the maximum HSA concentration were used.

To prove the potential sensors’ selectivity, squaraines’ interaction spectra with human gamma globulin (γ-globulin) proteins were displayed at the highest protein concentration tested (3.0 µM), using the same excitation wavelengths and the same slit widths and compared with those obtained at the same serum albumin concentration.

Detection and Quantification Limits Calculation

Detection and quantification limits (DL and QL, respectively), the lowest concentration of an analyte that can be consistently detected and the lowest level of an analyte that can be quantified with any degree of certainty [45], respectively, were calculated from a calibration line plotting the maximum fluorescence intensity versus the concentrations of protein. From the slope of this line, or method sensitivity, Equations (2) and (3) were applied:

$$\mathrm{DL}=\frac{3\mathsf{\sigma }}{\mathrm{S}}$$

(2)

$$\mathrm{QL}=\frac{10\mathsf{\sigma }}{\mathrm{S}}$$

(3)

where σ refers to the blank standard deviation, and S (sensitivity) to the linear regression slope between the fluorescence intensity and albumin concentration.

2.4. Computational Studies

AutoDock Tools (autodock4.exe) 1.5.6 software (The Scripps Research Institute) was used to dock squaraine dyes 5a, 11a,b, 12a, and 13a with HSA in Sudlow sites I and II and considering the whole protein. Under the codes 2BXD and 2BXG, the crystal structures of HSA were retrieved from the Research Collaboratory for Structural Bioinformatics Protein Data Bank [RCSB PDB, http://www.rscb.org (accessed on 26 April 2022)] with R-warfarin and ibuprofen as co-crystallized ligands in the Sudlow site I (2BXD) and Sudlow site II (2BXG) crystal structures of HSA, respectively. The University of California San Francisco (UCSF) Chimera 1.13.1 software was used to prepare and isolate proteins and their co-crystallized ligands, with the chain B and water of the complex ligand–protein removed. Squaraines were sketched in ChemDraw software, queued into Chem3D for energy minimization to the lowest energy conformation at a Minimum RMS Gradient of 0.01, and exported in a “.mol2” format. The PDBQT file of the dyes was created using OpenBabel 3.1.1 from the “.mol2” format.

All hydrogen atoms were added to the protein structure, Gasteiger charges were computed, non-polar hydrogens were merged, and AD4-type atoms were allocated to the protein structure in ADT.

The coordinates (x, y, z) 3.376, (−9.600), 5.600, and the size (x, y, z) of 60 × 60 × 60 points with 0.375 spacing, were used to define the Grid Box for 2BXD. The coordinates (x, y, z) in 2BXG were 5.048, (−4.464), (−15.042), with a size (x, y, z) of 60 × 90 × 50 points and a spacing of 0.375. The Grid Box center was defined as the center of the macromolecule with the coordinates (x, y, z) −0.16, −0.696, 0.101 and the size of 90 × 126 × 102 (x, y, z) points with 0.800 spacing in the docking, considering all proteins using 2BXD. Docking was performed using the Lamarckian genetic algorithm (GA) with 60 GA runs, a 150 GA population size, a maximum number of generations of 27,000, a maximum number of energies evals of 5,000,000, and a maximum number of top individuals surviving to the next generation of 1. The conformation of each squaraine dye and co-crystallized protein ligand with the lowest estimated free binding energy (BE in Kcal/mol) was chosen and selected for the analysis of the ligand–protein interactions by means of Discovery Studio 2016 software.

The water solubility (WS) of squaraines dyes 5a, 11a,b, 12a, and 13a was predicted using the pkCSM tool [46] and is presented in log (mol/L).

3. Results

3.1. Chemistry

The synthetic strategy used to obtain the desired squaraine dyes 5a, 11a,b, 12a, and 13a (Scheme 1) was similar to that recently reported [47,48,49]. Briefly, after the N-alkylation reaction of heterocycles 1a and 1b with iodohexane (2) and their condensation with squaric acid (4), corresponding zwitterionic dyes 5a and 5b were obtained. The latter were methylated with an excess of methyl trifluoromethanesulfonate (6; CF₃SO₃CH₃) at room temperature under a nitrogen atmosphere, obtaining the corresponding O-methyl ethers 7a and 7b. By reacting these O-methyl ether derivatives with several amines, such as diethanolamine (8), 2-picolylamine (9), and di-(2-picolyl)amine (10), the triflate-counterioned aminosquaraine dyes 11a,b, 12a, and 13a were obtained, and then subjected to the counterion exchange to iodide after treatment with 14% potassium iodide aqueous solution.

¹H and ¹³C NMR spectra are included as supplementary material (Figures S1–S11) and are in agreement with what is expected for this kind of compounds [50,51]. The authors emphasize the spectra of compound 12a that show the typical signals of an asymmetric squaraine, in particular, the chemical shifts of methine protons in the ¹H NMR spectrum at 6.67 and 6.22 ppm and the corresponding carbons in the ¹³C NMR spectrum at 90.80 and 88.92 ppm.

3.2. Photophysical Properties

Visible absorption is an important parameter in the spectroscopic characterization of the prepared squaraine dyes since, to be considered suitable fluorescent probes, these molecules must exhibit high absorption and emission wavelengths (650–900 nm) to minimize interference caused by the biological samples’ fluorescence [52,53].

Through the visible spectra of squaraine dyes 5a, 11a,b, 12a, and 13a, it is observable that they present maximum absorption peaks in DMF between 642 and 703 nm, in DMSO between 645 and 707 nm, and in CFM between 636 and 702 nm (

Table 2. Maximum absorption and emission wavelengths (λ_abs and λ_em, respectively; in nm), Stokes shift (ΔS; cm⁻¹), and relative fluorescence quantum yields [Φ_F; in percentage (%)] of squaraine dyes 5a, 11a,b, 12a, and 13a in dimethylformamide (DMF), dimethyl sulfoxide (DMSO), and chloroform (CFM).

		Squaraine Dyes
Solvent		5a	11a	11b	12a	13a
DMF	λabs	642	670	703	656	665
	λem	655	687	718	671	677
	ΔS	309	369	297	341	267
	ΦF	21.1 ± 0.3	3.6 ± 0.2	1.7 ± 0.3	5.5 ± 0.4	5.4 ± 0.4
DMSO	λabs	645	673	707	660	668
	λem	658	685	725	674	680
	ΔS	306	260	351	315	264
	ΦF	58.9 ± 2.7	4.1 ± 0.3	2.3 ± 0.4	8.3 ± 0.6	7.1 ± 0.1
CFM	λabs	636	667	702	657	664
	ε	3.98	1.94	1.67	3.02	1.48
	λem	650	681	715	671	675
	ΔS	339	308	259	317	245
	ΦF	116.1 ± 10.4	2.7 ± 0.1	2.3 ± 0.3	10.7 ± 0.9	1.9 ± 0.1

and Figure S12). The absorption behavior of the dyes did not vary significantly in the three organic solvents studied, with a near overlap of these same absorption spectra being observed. In CFM, all squaraines exhibited very high molar absorptivity coefficients (ε = 1.48–3.98 × 10⁵ M⁻¹·cm⁻¹), demonstrating their ability to absorb radiation for a subsequent release of energy presented as fluorescence emission. In addition, moderate Stokes shifts were observed (ΔS = 245–369 cm⁻¹). In organic solvents, squaraine dye 5a was the most fluorescent, with quantum yields from 21.1% in DMF to greater than 100% in CFM. This quantum yield of fluorescence slightly above 100% is due to the fact that it was calculated by a comparative method as well as the errors associated with the experimental approach. Therefore, this value should be analyzed carefully since it only indicates that for approximately every 100 photons absorbed, 100 were emitted. Marked differences in the quantum yields of fluorescence of the squaraine 5a in the evaluated solvents are mainly due to its nonpolar nature; therefore, it exhibits higher quantum yields in solvents of a more lipophilic nature (chloroform) than in more hydrophilic ones (dimethylformamide and dimethyl sulfoxide). Ethanolamine- and picolylamine-bearing squaraine dyes 11a,b, 12a, and 13a showed considerably lower relative quantum yields (Φ_F = 1.7–10.7%). Given their poor emission ability in these organic solvents, these latter dyes have had normalized emission spectra with poor resolutions.

3.3. Squaraine–Protein Interaction Studies

To evaluate the HSA detection ability of each prepared squaraine dye, solutions of dyes in PB were used, keeping their concentration constant (2.0 μM), and then increasing concentrations of protein (0.0 to 3.0 μM) were added. The fluorescence intensity of the prepared protein–dye solutions was fluorometrically read after an incubation time (0 min, 1, 2 and 3 h) to observe when a complete interaction with the protein, and consequently the maximum fluorescence intensity, is reached.

Figure 2 shows the emission spectra of the squaraines under evaluation in PB and after the successive addition of concentration-increasing HSA solutions, as well as the respective plots that represent the linearity between the concentration of HSA and the fluorescence intensity of the squaraine–protein solutions obtained from at the wavelength of maximum emission. From these spectra, it is possible to verify that all dyes interacted with the protein since there is an increase in fluorescence intensity with increasing HSA concentrations. The linearity between the fluorescence intensity and the HSA concentration is also noticeable through the lines obtained with correlation coefficients (r²) between 0.9 and 1.0. For squaraine dyes 12a and 13a, the maximum fluorescence intensity was observed immediately after adding the protein to the respective dye solution in the buffer. For squaraine dyes 5a, 11a, and 11b, the highest fluorescence intensity was observed after one incubation hour. This analysis is completed by observing the linear regression graphs shown in Figure 2, in which the linear regressions with steeper slopes correspond to the time at which the dye showed a more robust interaction with the protein.

In the literature, it is already well described that squaraines in buffer solution with an absence of protein exhibit a very low fluorescence intensity (close to zero) due to the formation of non-fluorescent aggregates [54,55,56]. Similarly, all evaluated squaraines showed a very low fluorescence intensity and quantum yields in protein-free buffer (Figure 2 and

Table 3. Differences observed in the absorption and fluorescence properties of squaraine dyes 5a, 11a,b, 12a, and 13a in phosphate buffer in the absence and presence of human serum albumin (HSA). The maximum absorption and emission wavelengths (λ_abs and λ_em, respectively), as well as the Stokes shift (ΔS), are given in cm⁻¹, the molar absorptivity coefficients (ε) are shown in × 10⁵ M⁻¹·cm⁻¹, and the relative fluorescence quantum yields (Φ_F) are given in percentages.

Squaraine Dye	HSA Absence					HSA Presence
	λabs	λem	ΔS	ε	ΦF	λabs	λem	ΔS	ΦF
5a	640	649	217	0.81	1.0 ± 0.1	640	648	193	27.3 ± 0.3
11a	670	680	219	0.48	4.7 ± 0.9	667	682	330	80.1 ± 7.7
11b	705	715	198	0.44	1.9 ± 0.5	707	714	108	28.8 ± 6.3
12a	657	665	182	0.71	3.5 ± 1.0	656	665	206	74.9 ± 3.0
13a	665	675	223	0.23	19.4 ± 3.3	663	677	312	93.2 ± 22.0

). Interestingly, picolylamine-bearing dyes 12a and 13a exhibited fluorescence intensities in the order of 29 and 38 arbitrary units in the absence of protein, values considerably higher than those usually observed, even compared to the other dyes herein evaluated (

Table 4. Determined squaraine–albumin interaction fluorescence of squaraine dyes 5a, 11a,b, 12a, and 13a [fluorescence intensity in the absence of protein (F₀), squaraine fluorescence intensity observed at the highest concentration of protein (F), ratio between these last two parameters (F/F₀), sensitivity (S), and quantification and detection limits (QL and DL, respectively)]. Predicted water solubility (WS) of squaraine dyes is given in log (mol/L).

Squaraine Dye	WS	F0 (a.u.)	Squaraine-Protein Complex Fluorescence Properties
			F (a.u.)	F/F0	S (nM)	QL (nM)	DL (nM)
5a	−5.91	9	363	41	1.0 × 105	1047	314
11a	−4.75	19	564	31	2.0 × 105	635	190
11b	−3.54	8	166	21	5.5 × 105	560	168
12a	−4.99	29	968	36	3.0 × 105	692	208
13a	−4.52	38	463	12	1.0 × 105	1172	352

). Of the evaluated squaraines, the zwitterionic 5a was the one that showed the greatest interaction with the protein in the fluorescence assays, with an F/F₀ ratio of 41 times. Dyes 11a,b and 12a also showed interaction with serum albumin, with the order of interaction being 12a > 11a > 11b. Squaraine 13a was also shown to increase its fluorescence in samples with albumin, albeit with an F/F₀ ratio of 12, markedly lower than the other compounds.

Compared to CFM, the molar absorptivity coefficients of the squaraines were much more reduced in PB (Table 3). The presence of the protein has not been shown to change the absorption and emission wavelengths of the squaraines. As expected, according to the results shown in Figure 2, very high relative fluorescence quantum yields (Φ_F = 27.3–93.2%) were observed when incubated with the HSA, again proving the dye–protein interaction. Quantum yields with higher standard deviations, such as for squaraine dye 13a in the presence of albumin, are justified by the high sensitivity of the protein–dye complex to external factors such as, for example, small fluctuations in the temperature at which the measurements were performed. Furthermore, the relative fluorescence quantum yields of squaraines 11a,b, 12a, and 13a in HSA-containing buffer proved even higher than those calculated in organic solvents.

From the fluorescence data obtained previously, the detection limit, quantification limit, and method sensitivity parameters were calculated for each squaraine dye. The squaraine dyes that exhibited the lowest detection limits were 11a and 11b (DL = 190 and 168 nM, respectively). Consequently, they were also the ones that also exhibited the lowest quantification limits (QL = 635 and 560 nM, respectively). Proportionally, although compounds 5a and 13a showed higher detection limits, the quantification limits are markedly higher than the other dyes. The binding affinity was too weak to observe saturation in the evaluated protein concentration range.

Regarding the sensitivity of squaraine dyes as potential fluorescent probes, the method is more sensitive the higher the fluorescence variation at the lowest protein concentration. Squaraines 11a,b and 12a were the ones that caused the most significant increase in fluorescence intensity at the lowest protein concentration, so these are the ones with the most considerable sensitivity (S = 2.0 × 10⁵, 5.5 × 10⁵, and 3.0 × 10⁵ nM, respectively). The application of the method using the remaining dyes showed lower sensitivity, but in the same order of magnitude (×10⁵ nM).

As the adequate solubility of a protein sensor in aqueous media is essential, this squaraines’ property was predicted, showing oscillations depending on the structural changes carried out (Table 4). The unsubstituted squaraine dye 5a showed the worst solubility in water, which was expected since the introduction of amines modulates the lipophilicity of this class of compounds, increasing its polarity. According to the solubility categories provided by pkCSM, the introduction of the amines made the squaraines moderately soluble in water. On the other hand, the diethanolamine-bearing benz[e]indole-based squaraine 11b was the one that proved to exhibit more excellent solubility in water, as it presented the WS value closer to 0. Strategies that could significantly improve the solubility of these compounds would be reducing the carbon number of the N-alkyl chains and introducing groups that increase their polarity, such as primary and secondary amines and hydroxyl groups or carboxylic acids. Functionalizing the side chains with sultones, phosphoryl groups, and other charged groups are alternatives to modulate the squaraines’ hydro- and lipophilicity and markedly improve their biocompatibility.

Gamma globulins are the most abundant class of serum proteins after albumin. Given their high presence in serum, the study of selective fluorescence intensity increases for HSA compared to the presence of γ-globulins is of significant relevance. For this purpose, the squaraine–globulin fluorescence spectra were displayed at the highest protein concentration tested (3.0 µM) and compared with those obtained at the same serum albumin concentration to demonstrate the selectivity of the potential sensors. All dyes exhibited higher fluorescence intensities in the presence of albumin than γ-globulins, indicating a better interaction with HSA (Figure 3 and Figure S13). Despite this, for example, squaraine dye 11b was the one that showed the lowest selectivity, showing fluorescence intensities about three times higher when incubated with HSA. Squaraine dye 12a showed the most remarkable difference between the intensity emitted in the presence of albumin and γ-globulins, showing a fluorescence intensity about fifteen times lower in the presence of globulins. Compared with the fluorescence emitted by the squaraines alone in buffer, the fluorescence observed when incubated with the globulins was not markedly superior, regardless of the dye under study.

3.4. In Silico Studies

Human serum albumin contains three homologous α-helical domains I-III, each separated into subdomains A and B. It is involved in the transport of both drugs (e.g., warfarin) and endogenous ligands (e.g., fatty acids, bilirubin, heme group) in the bloodstream [57]. There are several known binding sites for their interaction with various drugs, with Sudlow sites I and II being the most frequently studied [57,58,59]. As a result, molecular docking of the squaraine dyes 5a, 11a,b, 12a, and 13a with HSA was carried out in Sudlow sites I (PDB code: 2BXD) and II (PDB code: 2BXG) [58]. Additionally, the co-crystallized ligands (warfarin and ibuprofen from Sudlow site I and II, respectively) were docked, and the results were consistent with previous studies (Tables S1 and S2), validating docking parameters [59,60].

3.4.1. Interaction with Sudlow Site I

Sudlow site I is the binding site for warfarin, a thrombolytic agent. The pocket interior wall consists of hydrophobic side chains, while the entrance is positively charged due to LYS195, LYS199, ARG218, ARG222, HIS242, and ARG257. TYR150, PHE211, TRP214, ALA215, LEU238, and ILE264 are among the residues remaining in Sudlow site I [57,58,60]. According to

Table 5. Estimated free binding energy of warfarin, ibuprofen, and squaraine dyes 5a, 11a,b, 12a, and 13a in Sudlow sites I and II and full human serum albumin (HSA).

Ligand		Estimated Free Binding Energy (Kcal/mol)
	Warfarin	Ibuprofen	5a	11a	11b	12a	13a
Protein 2BXD Sudlow Site I	−8.80	–	−12.08	−11.07	−12.15	−13.12	−12.34
Protein 2BXG Sudlow Site II	–	−7.42	−8.36	−8.83	−10.21	−8.87	−9.41
Protein 2BXG Full Protein	–	–	−9.09	−6.84	−9.67	−8.66	−9.25

, warfarin interacts with most of these amino acid residues, with an estimated BE of −8.79 Kcal/mol. Warfarin forms hydrogen bonds (H-bonds) with TYR150, ARG222, HIS242, and ARG257, and electrostatic interactions with the residue LYS199. LYS195, TRP214, ALA215, ARG218, LEU219, and LEU238 exhibited hydrophobic interactions (Figure 4 and Table S1).

Regarding the squaraine dyes, all revealed BE values lower than warfarin, indicating that the prepared dyes could be suitable substrates of HSA. Picolylamine-bearing squaraine dye 13a was the best ligand with a BE of −13.12 Kcal/mol (Table 5). The dyes that showed more hydrophobic interactions had a greater affinity to this active site. Squaraine dyes 5a, 11b, and 13a demonstrated charge repulsion interactions.

3.4.2. Interaction with Sudlow Site II

Sudlow site II is mainly constituted of hydrophobic side chains and some disulfide bridges with ARG410 placed at the pocket entrance with TYR411, which has a hydroxyl group facing towards the inside. In addition, the amino acid residues ARG348, GLU383, LYS414, VAL433, GLU450, ARG485, and SER489 are also present in this pocket [57,58,60].

Figure 5 shows that ibuprofen interacts with most of these amino acid residues, mainly through hydrophobic interactions. H-bonds were observed with ARG410 and TYR 411. The BE of the best conformation was −7.42 Kcal/mol (Table 5). All evaluated squaraine dyes demonstrated more proficient binding to this receptor than the co-crystallized ligand in Sudlow site II. In comparison, the interaction with this site was lower than that with Sudlow site I. The majority of the interactions of dyes were hydrophobic (Figure 5 and Table S2). On this site, only a few electrostatic interactions were observed. Squaraine dye 11b, which demonstrated more H-bonds and electrostatic interactions, was the best ligand among the tested dyes (BE = −10.21 Kcal/mol).

3.4.3. Docking with the Full Protein

Due to all the substrates of HSA and the several ligands of this transporter protein, molecular docking was performed considering the whole protein [57,58,59,60]. This simulation demonstrated squaraine 11b as having the best binding affinity to the HSA with a BE of −9.67 Kcal/mol (Table 5). The interactions observed are present in Table S3 and Figure S14.

4. Discussion

Based on the noteworthy results previously reported using squaraine dyes derived from benz[e]indole (1–3; Figure 1) with a view to being used as photosensitizers for photodynamic therapy [43], five squaraine dyes analogous to these were prepared in order to better understand the role of structural modifications in this core of compounds in its interaction with the HSA protein. Furthermore, in addition to the picolyl- and dipicolylamine groups that had already been recently considered, the introduction of the diethanolamine group was studied to provide better hydrophilicity to the compounds due to the hydroxyl groups present in them, improving their suitability for biological experiments.

Despite the low relative fluorescence quantum yields observed in PB, the dyes showed the ability to fluoresce after incubation with albumin. The fluorescence emitted by all squaraines was directly proportional to the protein concentration. The ideal incubation time for displaying the highest possible intensity varied according to the dye under study: while those loaded with picolyl- and dipicolylamino groups exhibited maximum fluorescence at t = 0, the remaining compounds required 1 h of incubation. However, this rapid interaction with the protein is not necessarily related to the picolyl- and dipicolylamino groups since squaraine 2, also loaded with the group with a single pyridine unit, required 3 h of incubation to exhibit maximum fluorescence [43]. The fact that the structure of dyes affects the response time had already been reported by Barbero et al., who found that the heterocyclic unit of indolenine, due to its smaller dimension, compared to the bulkier benz[e]indole-based derivatives, aimed at a faster and more effective interaction [2]. This last theory should be the explanatory basis for the results achieved. Furthermore, evidence shows that, after some time, the dye may again dissociate from albumin since a straight fluorescence intensity versus a protein concentration with lower slopes is obtained after enhanced interaction (more evident for squaraine 13a, which exhibits fluorescence intensity versus protein concentrations plots with decreased slopes after 1 and 2 h of incubation compared to t = 0).

At the time at which the fluorescence emitted by the protein–dye adducts is more evident and at the highest protein concentration tested, high quantum yields of fluorescence were obtained (Φ_F = 27.3–93.2%), highlighting that for dye 13a. However, this fact does not directly indicate a better interaction since squaraine 13a presents the highest relative fluorescence quantum yield in a protein-free buffer among the compounds presented here. The quantum yields after protein addition were higher for the indolenine-based compounds shown here than for the benz[e]indole derivatives 2 and 3 (Φ_F = 25.0 and 36.3%, respectively [43]). Comparing the fluorescence intensities emitted after incubation of the dye with the highest concentration of protein tested, squaraine dye 5a exhibited a 41-fold-higher peak fluorescence intensity in the presence of albumin. All others showed lower ratios (F/F₀ = 12–31), indicating a weaker interaction. The introduction of the dipicolylamino group markedly weakened its ability to interact in vitro with HSA. With an intermediate F/F₀ ratio, the diethanolamino-bearing benz[e]indole dye 11b was shown to be the most sensitive (S = 5.5 × 10⁵ nM), as well as the one with the lowest detection limit (DL = 168 nM), standing out as a potential means of detecting albumin in samples with very low levels of this protein. Given its low detection limit, squaraine 11b’s limit of quantification was also low (QL < 1.2 µM), desirable features for its potential application in protein quantification. Squaraine dye 5a exhibited a moderate detection limit (DL = 314 nM) and the highest quantification limit among the prepared dyes (QL = 1047 nM).

Compared to other dyes of the same family reported in the literature [61,62,63], those reported here generally showed slightly higher detection limits, but limits of quantification of the same order of magnitude or significantly higher (such as in the case of squaraine 5a). Based on the results obtained for compounds 2 and 3 [43] and those reported here, it can be concluded that the functionalization with amines induces, in general, a very marked reduction in both determined limits. The introduction of the diethanolamino group vaguely improved the method’s sensitivity, showing values two to five times higher than the indolenine zwitterionic dye 5a. Comparisons with the benz[e]indole zwitterionic dye 1 were not possible, as it was not studied due to its high insolubility. At the same time, the functionalization with picolylamines seems to have reduced this parameter more evidently for the dyes with two pyridine moieties.

For the potential application of these dyes as fluorescent probes for the detection of albumins in human plasma samples, selectivity is an essential factor, and the interaction of the evaluated dyes with γ-globulins was compared. Squaraine dye 12a showed the most remarkable difference between the intensity emitted in the presence of albumin and γ-globulins, showing a fluorescence intensity about fifteen times lower when incubated with globulins. The difference in fluorescence intensity emitted by squaraine 11b in the presence of albumin and globulins was relatively small compared to the other compounds, showing poor selectivity. The applicability of the remaining molecules as albumin fluorescent probes is not questioned, showing fluorescence intensities five to ten times higher for HSA.

Molecular docking of the various dyes at Sudlow site I showed that the most predominant interactions are of a hydrophobic nature, similar to warfarin. Squaraines 11a,b also showed interactions by hydrogen bonds, mainly involving the hydroxyl groups of the diethanolamino group. Electrostatic interactions were also observed for dyes 5a, 12a, and 13a. Although the squaraines exhibited an affinity for this receptor higher than warfarin, dyes 5a, 11b, and 13a exhibited a negative charge repulsion due to their positive charge in the indole moieties. These unfavorable interactions are explained by the positively charged entrance of Sudlow site I, making this active site more appropriate for anionic ligands [58,60]. For Sudlow site II, all squaraine dyes showed considerably lower BE values than for site I, thus displaying less affinity for this site on the protein. Similar to site I, hydrophobic interactions were the most predominant, and classic hydrogen bond interactions were also observed for diethanolamino-loaded squaraines 11a,b, and non-classical interactions for the benz[e]indole derivative 11b. In general, for this site, squaraines 11b and 13a showed the best affinity in silico (BE = −10.21 and −9.41 Kcal/mol, respectively), while the remaining ones showed equivalent negative energies. Considering the protein as a whole, all evaluated squaraines displayed similar BE values, except for dye 11a, which showed a lower interaction with human serum protein. Thus, the in silico studies, despite not allowing for differentiation of the interaction as accentuated as those observed in the fluorescence studies, allowed us to understand that Sudlow site I is the preferred active site for the interaction of all dyes with the HSA protein, since that it was for this that distinctly higher binding energies were calculated.

In the near future, our research group’s objective is to carry out more in-depth selectivity studies in which other proteins, such as fibrinogen and hemoglobin, and other molecules such as amino acids, vitamins, hormones, and ions recurrently found in biological samples will be subjected to in silico and in vitro interaction studies. Furthermore, having verified a good selectivity index, we intend to compare the method using the potential fluorescent probe with traditionally used procedures to prove its reliability and validity.

5. Conclusions

We report an in vitro and computational study to verify the potential application of squaraine dyes derived from benz[e]indole and indolenine as fluorescent probes for detecting HSA proteins. All dyes showed a structure–relationship influence on the kinetic interaction with serum albumin, resulting in a significant increase in fluorescence intensity. After adding the squaraine solution to the HSA solution, the dye molecules interact with the protein primarily through hydrophobic interactions. Depending on the dye structure and limits of detection, the quantification and sensitivity diverge considerably, with indolenine-based zwitterionic squaraine dye showing properties closer to those desirable for a fluorescent probe. Functionalization with amines has reduced the detection and quantitation limits of squaraine dyes while maintaining worthy sensitivity. In silico studies show these compounds may bind to Sudlow site I and site II of HSA, with a preference for site I. When incubated with γ-globulins, the dyes emitted considerably lower fluorescence intensities than in the presence of HSA, suggesting good to excellent selectivity. Other selectivity studies are the next step to reiterate the interest in these compounds from the perspective of application in biological samples.