A. Materials

BSA (lyophilized powder > 98%), RB (> 99%) and all other chemical regents were purchased from Sigma-Aldrich Corporation. Deionized water (18.2 MΩ cm) from a Milli-Q system (Millipore, Bedford, MA) was used for all experiments. BSA was dissolved in Milli-Q water for a stock solution with a concentration of 600 μm. RB was dissolved in Milli-Q water with a stock concentration of 600 μm.

B. Spectroscopic measurements

UV-vis absorption spectra were recorded on a U-2900 UV-vis spectrometer (Hitachi).

C. SEC assays

The SEC assays were performed on a sephacryl S-300 column equipped with a high performance liquid chromatography (HPLC) system (Hitachi L2000 or Agilent 1260) with a flow phase of SB9 (sodium borate 50 mM, pH = 9) buffer and the flow rate of 1 mL/min. The spectra were monitored at 280 nm for UV detector and excitation (Ex): 555 nm / emission (Em): 575 nm for the fluorescence detector. For the fluorometric titration experiment, 995 μL solutions containing different RB/BSA ratios of 64/1, 32/1, 16/1, 8/1 and 4/1 with a fixed final BSA concentration of 2.29 nM was prepared and incubated at 298 K for 2 h.

D. Molecular docking

The three-dimensional (3D) structure of the BSA (PDB ID: 4JK4) was downloaded from Protein Data Bank (http://www.rcsb.org/pdb/home/home.dohttp://www.rcsb.org/pdb/home/home.do). The Auto Dock Tools 1.5.6 package (http://mgltools.scripps.eduhttp://mgltools.scripps.edu) was employed to generate the docking input files. The search grid of BSA was identified as center x: 97.127, center y: 24.933, and center z: 20.919 with dimensions size x: 15, size y: 15, and size z: 15. For Vina docking, the default parameters were used if it was not mentioned. The best-scoring pose as judged by the Vina docking score is chosen and visually analyzed using PyMOL software (http://www.pymol.org/http://www.pymol.org/).

A. Binding constant and binding sites

A number of methods can be used to calculate/measure the binding constant and binding sites. Some authors like using fluorescence quenching [17-19] to study interactions between proteins and other molecules. This method is facile and easily accessible, but it is limited when the molecules are not quenchers to proteins’ fluorescence. Other popular methods include quartz crystal microbalance (QCM)-based [20], and surface plasmon resonance (SPR)-based [21] techniques, but they cost quite a lot by consuming specific chips. The raw data obtained with these methods normally need further processing in combination with some classical equations, like Stern-Volmer equation, Scatchard equation and Hill equation, to get the binding constants, sites and even cooperativity.

SEC is a powerful tool for both analysis and separation. Because proteins are several times larger than the labeling fluorophores which are normally less than 1 nm, so we can readily separate the bound fluorophores and free ones, which is exactly the data required by the Scatchard equation,

where C_f is concentration of the free ligand, which in our case is the unbound RB’s concentration; n is the number of binding sites per protein molecule; K_a is the association/binding constant/affinity of RB for BSA; and ν is the real bound ratio defined as,

where C_t(RB) is concentration of total RB which is known before the binding assay; C_f(RB) is the free or unbound RB’s concentration which can be obtained by SEC measurement; and C_t(BSA) is concentration of total BSA for the binding assay, which is known and constant in the current assay.

We took RB and BSA for an example to justify the use of SEC in combination with Scatchard equation to calculate the binding constant and sites between molecules and proteins. To be more precisely, we made a pure dye’s SEC to get the SEC-based calibration curve of concentration (Fig. 1) by plotting the integrated elution peaks against the concentrations. From Fig. 2, it can be seen that, with the molar ratios of RB/BSA in Section II C, the RB’s fluorescence intensity decreases linearly with increasing BSA/RB ratios at a fixed RB molar concentration. The binding constant and binding sites were calculated at 2.2×10⁵ M^-1 and 1.3, respectively. Cai et al. reported that the binding constant and binding sites were 4.8×10⁴ M^-1 and 1.2, respectively. However, their further analysis by time-resolved fluorescence indicated at least two binding sites existed [22]. The real binding sites can be highly resolved by crystallographic studies of the complexes of proteins and molecules, for example Sekula B. and colleagues found there are four binding sites of 3,5-diiodosalicylic acid to BSA [23]. The binding affinity obtained by the current method seems much higher than Ref. [22]. Thus, we believe that different buffers and sensitivities of the instruments might be the culprit, and that the fluorescence quenching method can produce huge deviations for the binding affinity evaluation [24]. So, it would be helpful to compare different methods in the same conditions.

Fig. 1.Full-screen View

(Color online) Concentration calibration of RB obtained by SEC. (a) Fluorescence signals obtained by SEC at RB concentrations of (from a to e) 146.4 nM, 73.24 nM, 36.62 nM, 18.31 nM and 9.15 nM. (b) The calibration curve of free RB’s concentration (C_f(RB)) based on the SEC by plotting the elution peak area against the concentration.

Fig. 2.Full-screen View

(Color online) The binding assay studied by SEC and Scatchard equation. (a) Fluorescence signals derived from the SEC analysis at RB/BSA molar ratios of 64/1, 32/1, 16/1, 8/1 and 4/1 (from a to e). The peaks reflect the free RB’s concentration (C_f(RB)), and the conjugate’s concentration was too low compared with the free RB’s eluted out at 10 min. (b) The Scatchard plot and linear fit.

It is not difficult to find the intermolecular interaction studies by searching the scientific publication bank/library, but it may be difficult to figure out how those scientists get necessary values for the Scatchard plot. By using SEC assisted with an HPLC system, one can clearly learn how to do this plot step-by-step. With the well-known reputation of the repeatability of a commercial HPLC system such as the Agilent used in this paper, one can directly calculate both binding constant and binding sites just by simply preparing the mixture solution. In addition, the concentration calibration curve of free dye alone obtained by the same configuration constants of HPLC system such as the column, flow rates and loading volumes, is necessary in that it minimizes the measurement error.

B. Molecular docking of RB and BSA

To study the binding mode of RB to BSA, molecular docking was performed by using the Auto dock. As shown in Fig. 3, the RB molecule is docked into the binding pocket of the BSA. One of diethylamino group of RB fits into bottom of the binding pocket of BSA, surrounded by the residues Trp²¹³, Arg¹⁹⁴ and Arg¹⁹⁸, while the other diethylamino group located at the entrance of the pocket, interacted with the hydrophobic residues Val³⁴², Ala³⁴¹ and Pro⁴⁴⁶. Importantly, the carboxyl group forms two key hydrogen bonds (2.0 and 3.3 Å) with the residue Arg¹⁹⁴, which is important for the affinity between RB and BSA.

Fig. 3.Full-screen View

(Color online) Molecular docking of RB and BSA. (a) The RB is docked into binding pocket of the BSA’s surfaced structure. (b) Detailed analysis of main binding affinity between BSA’s amino acid residues and RB’s chemical groups. The distances are in Å (0.1 nm).

In summary, the computational approach helps in better understanding of inhibitors binding to the protein active site, which provides valuable information for further study of interaction between RB and BSA. Limited by the docking software which does not support the multi-molecular docking, we didn’t try more docking sites.