Dyes which absorb in the visible range and proteins which absorb in the UV range can be studied using UV-Visible spectroscopy. The aromatic amino acids tryptophan, tyrosine and phenylalanine absorb in the 230–300 nm range and the peptide bond at ~230 nm. In the case of hemoglobin, the absorption spectra of the heme prosthetic group in hemoglobin is very sensitive to changes in polypeptide environment—structure, oxidation state, and ligand binding, rendering the technique highly suitable to follow Hb-ligand binding (Messori et al., 2006). Typically the hHb spectrum shows two major absorbance in the 199–700 nm range—a peak at ~200 nm due to the CO carbonyl groups in the amino acids and a Soret band seen at 405–450 nm due to the π–π transitions due to the heme moiety embedded in a hydrophobic pocket in the protein backbone (Messori et al., 2006; Chatterjee and Kumar, 2016). Our data (Figure 2) for bovine Hb showed a broad band in the 420–530 nm range, with its peak at 478 nm and a slight shoulder at ~450 nm. Bovine hemoglobin shows very few structural and functional differences from human Hb (Safo and Abraham, 2001). However, the high absorbance peak or the Soret band appears at 478 nm and the very weak α and β bands typically seen between 540 and 600 nm cannot be distinguished from the baseline (hematoporphyrin and fully oxygenated Hb: Soret band at 422–450 nm, β-band at 540 nm and α band at 576 nm) at the concentrations used (5–10 μM).

In this study, absorption for the dye (λ_max) was followed using UV-Visible spectroscopy. The spectra for Congo Red (CR) is shown here as representative spectra (Figure 2A inset). Absorption data for CR from 200 to 800 nm, showed broad peaks with maxima at 343 and 486 nm. The broad peak around 485–490 nm could contain the azo-bond characteristic peaks at 493 and 488 nm (Wang and Zhang, 2012).

UV-visible spectra were taken of dye alone and of hemoglobin-dye complexes at increasing concentrations of dye. Changes in extinction coefficient and/or peak position of the dye on addition of the protein, or vice versa, are indicative of changes in the electronic environment of the molecule and hence complexation (Benesi and Hildebrand, 1949). The bands in the UV-Vis spectrum for the dye do not change position with the addition of Hb but show decreased absorption (Figure 2A inset), indicating interaction between the dye and Hb. Absorption of the complex showed ~86% deceased absorption relative to dye alone at 2 μM CR at 343 nm. This decrease in the absorption of the protein–dye complex was observed at all wavelengths between 200 and 600 nm, relative to the absorption spectrum of only dye. However, no significant change in peak positions in the spectrum of the complex relative to the spectrum of only dye was observed. In addition, the broad band with a peak at ~478 nm, seen in the Hb spectrum, shows decreased absorption with the addition of dye (data not shown), indicating a perturbation of the heme environment in the presence of dye (Franzen and Boxer, 1997; Dayer et al., 2010).

At each dye concentration, the absorption of only dye was subtracted from the absorption of the Hemoglobin–dye complex. The difference in absorption (ΔA) was then plotted against 1/[dye] (Figure 3) according to the Benesi-Hildenbrand equation (Fennell et al., 2015).

where ε is the extinction coefficient of bound and free states and M is protein concentration, b is bound ligand and f is free ligand and T is the total ligand concentration. ΔA is the change in absorbance at a given wavelength and K_a is the binding constant for the complex formation. Here, the value of K_a is given by the ratio of intercept to slope of the plot of 1/ΔA against 1/[dye] (Fennell et al., 2015). The values obtained for K_a, for all four dyes were similar [Fluorescein = 0.44 (± 0.51) × 10⁵ M, Congo Red = 0.2 (± 1.17) × 10⁵ M, Methyl Red = 0.55 (± 0.22) × 10⁵ M, Methyl Orange = 0.53 (± 0.188) × 10⁵ M, Table 1].

The Benesi-Hildenbrand equation assumes that the decrease in absorption at dye (λ_max) is entirely due to the formation of a 1:1 complex and hence is directly proportional to the concentration of the complex formed (Benesi and Hildebrand, 1949). A departure from linearity would indicate that the complex formed does not have the expected 1:1 protein–ligand stoichiometry or that less than expected ligand is available for complex formation. The linear fit for the data for all four dye molecules studied improves if data points at the concentration extremes are removed, removing the apparent non-linearity in the data. A non-linear model improves the fit to the data significantly only in the case of Fluorescein. The fit improves marginally in the case of Methyl Red and Methyl Orange, and not at all in the case of Congo Red (data not shown). The data for Fluorescein shows the greatest scatter and departure from a straight line. Fluorescein molecules are known to aggregate at higher concentrations which can lead to a lower than expected concentration of free Fluorescein being available for complex formation.

Parameters used to characterize an interaction between a protein and its ligand are its binding strength and the number of ligand binding sites on the protein. The intrinsic fluorescence of proteins, due to the amino acid residues tryptophan, tyrosine, and phenylalanine, change in response to conformational changes resulting from ligand/small molecule binding. Hemoglobin contains six Tryptophan amino acid residues, of which the β-37 Trp are the dominant fluorophores, sensitive to quaternary structural change (Alpert et al., 1980).

Quenching experiments were carried out with Hemoglobin and all four dyes. Quenching of Hemoglobin fluorescence was observed with increasing concentration of all four dyes. The fluorescent spectrum of hemoglobin with Methyl Red is shown as representative (Figure 3A). The Hemoglobin fluorescence emission peak between 290 and 400 nm showed quenching for all dyes studied and analyzed using Stern-Volmer and modified Stern-Volmer equations. The Stern-Volmer equation is given as:

Where F is fluorescence intensity with quencher, F₀ is fluorescence intensity without quencher, K_SV is the rate of quencher coefficient, and [Q] is the concentration of quencher (Trott and Olson, 2010). The modified Stern-Volmer equation is given:

Where F is fluorescence intensity with quencher, F₀ is fluorescence intensity without quencher, K_a is equilibrium binding constant, n is the number of binding sites, and [Q] is concentration of quencher (Lakowicz, 1999). Fluorescence data was analyzed using both equations to derive the interaction parameters for binding between hemoglobin and each dye, i.e., the Stern-Volmer constant (K_sv), the bimolecular quenching constant (K_q), the binding constant (K_a), and the number of binding sites (n, Figure 3). The results are shown in Table 1.

The Stern-Volmer plots are linear for all dyes under consideration here, indicating a single (static or dynamic) mode of fluorescence quenching. For each dye, the bimolecular collision constant K_q can be calculated from the value of K_sv and τ₀ (average lifetime of Hb without quencher, 10⁻¹⁰s, Berezin and Achilefu, 2010). The maximum quenching constant K_q with biopolymers is 2 × 10¹⁰ M L⁻¹ s⁻¹. The calculated values of K_q for the four dyes are greater than the maximum value for collision quenching indicative of static quenching and complex formation (Ware, 1962). [Fluorescein = 0.006 (± 0.0029), Congo Red = 0.005 (± 0.0013), Methyl Red = 0.009 (± 0.001), Methyl Orange = 0.003 (± 0.0051), Table 1]. The scatter seen in the Stern-Volmer plot for Fluorescein (Figure 4D) may be in part due to the aggregation seen in Fluorescein, especially with increasing concentrations.

The association/binding constant for Hemoglobin-dye complexation (K_a) and the number of binding sites (n) was determined from the intercept and slope, respectively, of the modified Stern-Volmer equation (given above), thus specifying the number of dye molecules bound to each hemoglobin molecule. Plots for both equations for all four dyes are shown in Figure 4. All four dyes showed similar association constants, which are within 2–6-fold of the association constants determined from UV-Visible experiments [Log Ka: Fluorescein = −1.3 (± 0.04), Congo Red = −2.113 (± 0.23), Methyl Red = −2.31 (± 0.22), Methyl Orange = −2.02 (± 0.72), Table 1]. The number of binding sites for each dye can be determined from the slope of the modified Stern-Volmer equation. The values shown in Table 1, indicate the expected single binding site for Congo Red (0.74 ± 0.3) and Methyl Red (1.16 ± 0.24). However, values for Fluorescein (0.19 ± 0.06) and Methyl Orange (0.424 ± 0.29) are unexpected and would indicate less than a full binding site for both dyes.

Molecular docking forms an important tool in drug design as it is allows estimation of the most stable binding conformation of small molecules to their target, hence allowing prediction of binding affinity, and activity of the molecule (Meng et al., 1992). The binding of Hemoglobin with each dye was simulated using the molecular docking software, Autodock Vina, from Scripps Research Institute, U.S.A. (Trott and Olson, 2010). The results of the docking program allows prediction of possible binding conformations of the ligand, binding site and binding partners for each dye with Hb. The results of the simulation showed multiple binding conformations (upto six different conformations) for each dye with very small differences in energy. The three lowest energy conformations for Methyl Red binding Hemoglobin have been shown as representative in Figure 5.

Interactions could be present between Methyl Red molecule and the same α- and β-chains of the protein. It is clear that all the dyes bind within the central cavity of the molecule. However, the exact binding partners or residues cannot be determined since a single binding conformation could not be identified. H-bonds and electrostatic interaction could be formed between the dye molecules and Hemoglobin residues falling within four angstroms be formed each conformation, stabilizing the hemoglobin-dye complex formed for all dyes.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.