Human transferrin (hTf) and vanillin were bought from Sigma Aldrich (St. Louis, USA). For buffer preparations, all the chemicals were of the best quality purchased from HiMedia. All the buffers were filtered before use and for spectroscopic assays, corresponding blanks were taken.

Molecular docking studies were conducted on a DELL^® workstation to elucidate the interactions between Vanillin and hTf. Computational tools were used to optimize visualization, evaluation, and analysis of the docking results. We obtained three-dimensional configuration of hTf (PDB ID: 3V83) and PubChem database (PubChem CID: 1183) was used to obtain vanillin structure. We prepared vanillin structure using MGL AutoDock tools (Huey et al., 2012). Further, docking simulations were carried out using AutoDock Vina (Trott and Olson, 2010). We employed PLIP to explore the intricate ligand-receptor interactions between vanillin and hTf (Salentin et al., 2015). To gain an in-depth understanding of the binding conformation and the nature of interactions between vanillin and hTf, PyMOL (DeLano, 2002) and Discovery Studio Visualizer (Biovia, 2017) were used.

MD simulations were conducted to explore the dynamic behavior of hTf and hTf-vanillin complex. We performed a simulation at 300 K temperature, using charmm36-jul2022 force field (Huang and MacKerell, 2013), implemented within GROMACS version 2020 (Van Der Spoel et al., 2005). The topology for vanillin was generated using CHARMM CGenFF program, using a Python script. hTf and hTf-Vanillin were later solvated in cubic boxes with 1 nm dimensions in the SPC216 water model using GROMACS gmx solvate tool (Glättli et al., 2002). Both systems were neutralized by adding suitable quantities of counterions to achieve a concentration of 0.15 M. Equilibration processes in the NVT and NPT ensembles were conducted, where the systems were slowly heated to 300 K over 1000 ps. Subsequently, solvent molecules were allowed to relax in the NPT ensemble for an additional 1000 ps without any restraints, while maintaining a pressure of 1 bar using the Berendsen Barostat method. To constrain bond lengths, the LINCS algorithm was applied, while SETTLE restrained water molecules. The system’s energy was minimized with 1500 iterations of the steepest descent approach. We employed a range of tools provided by GROMACS to analyze trajectories generated during MD simulations. The results obtained from the MD simulations were presented graphically using the XMGrace software (Turner, 2005). Graphs and figures were created to illustrate key findings, aiding in a clear and concise explanation of the obtained results.

Principal component analysis (PCA) is an essential tool for understanding the primary motions and conformational sampling of proteins in MD simulations. The MD trajectories of hTf and hTf-vanillin underwent PCA to explore their predominant modes of motion. This involved diagonalizing the covariance matrix to derive the first two eigenvectors (EVs), which depict the primary motions of the protein or protein-ligand complex. Additionally, we generated free energy landscapes (FELs) to assess the stability and folding dynamics of hTf and its complex with vanillin. Conformational sampling techniques were employed to construct FELs, enabling the visualization of energy landscapes and the identification of stable conformations or metastable states.

We performed fluorescence spectroscopy on Shimadzu (RF 6000) spectrofluorometer. The concentration of hTf was fixed at 4 µM and vanillin was varied in a range of “0–5 µM”. The quenching data was analyzed using “Modified Stern–Volmer” equation (MSV) to compute binding parameters of hTf-vanillin complex.

Vanillin showed remarkable interactions towards hTf, resulting in a highly favorable binding affinity in triplicate dodcking experiments. Vanillin binds to hTf with an average binding affinity of −5.8 kcal/mol, implying vanillin’s potential as a potent hTf binding partner. Vanillin is predicted to primarily interact with the hTf binding pocket (Figure 1A). Notably, vanillin forms three hydrogen bonds with key residues, viz. Ser409, Arg475, and Leu649 (Figure 1B; Supplementary Table S1). Tyr536 is a vital active site residue that plays a key role for iron (Fe³⁺) binding in hTf, which plays a key role in forming these interactions (Noinaj et al., 2012). Tyr536 forms a hydrophobic bond with vanillin, further stabilizing the complex (Figure 1B). The 2D and surface plots further shed light on other interactions playing a key role in stabilizing the hTf-vanillin complex. Arg475 is another active site residuewhich is responsible to bind with hydrogencarbonate. The surface representation of the docked complex further revealed that vanillin fits into the internal cavity of hTf (Figures 1C, D). Earlier studies have also reported iron binding to hTf showing similar interactions, thereby validating our findings (Noinaj et al., 2012; Ghanbari et al., 2017). Overall, it is indicated that vanillin binds to the hTf binding pocket.

MD simulations are a widely adopted computational technique for unraveling the underlying dynamics and structural alterations of proteins and protein-ligand complexes (Shukla and Tripathi, 2020). MD simulation serves as a valuable tool for probing the biomolecular interactions within a defined time frame (Mohammad et al., 2020). In this specific investigation, we conducted a comprehensive MD simulation spanning 200 nanoseconds, encompassing hTf and the hTf-vanillin docked complex. Our key goal was to scrutinize changes in conformation, assess stability, and unravel the fundamental mechanisms governing the interaction between vanillin and hTf. First, to evaluate the structural changes, we employed the root mean square deviation (RMSD) on the simulated trajectory. Our findings reveal that vanillin binding leads to the stabilization of hTf without inducing any significant alterations in its inherent configuration, as clearly depicted in Figure 2A. However, during the initial phase, the RMSD showed intermittent fluctuations, primarily attributable to the accommodation of vanillin within hTf’s binding site.

To delve into local structural changes and residue flexibility, we calculated and visually represented the root-mean-square fluctuation (RMSF) of the simulated systems (Figure 2B). The RMSF plot highlights varying levels of residue fluctuations across different zones of the hTf structure. Intriguingly, the overall average RMSF of hTf and hTf-vanillin exhibited a consistent pattern. Importantly, the introduction of vanillin resulted in heightened residue fluctuations, indicative of increased system dynamics. These enhanced fluctuations may be attributed to residual vibrations during the simulation process.

The radius of gyration (Rg) is a commonly utilized measure to glean insights into the folding and structural conformation of the proteins. A higher Rg typically implies a less condensed conformation. In the case of the hTf-vanillin, our observations indicate a minor decrease in Rg compared to the unbound hTf, signifying that the structure maintains compactness throughout the simulation, as illustrated in Figure 2C. Despite slight variations in Rg, there are no significant alterations in the packing of hTf following the binding of vanillin. These findings underscore that vanillin induces minimal structural deviations in hTf.

The solvent-accessible surface area (SASA) represents the fraction of a protein’s surface directly exposed to the adjacent solvent and is closely associated with the Rg. We calculated and visually depicted the average SASA for hTf and hTf-vanillin from the simulated trajectory. In the instance of the hTf-vanillin complex, we observed a minor decrease in SASA compared to the free hTf, implying the maintenance of structural compactness during the simulation. Despite minimal variations in SASA, no significant alterations in the packing of hTf were discernible following the binding of vanillin (Figure 2D).

Moreover, by employing MD simulation trajectories, we have successfully depicted the conformational snapshots showcasing the binding interactions between vanillin and hTf. This analysis enabled us to monitor the temporal evolution of the binding prototype. Structural assessments were performed at five distinct time intervals: 1, 50, 100, 150, and 200 ns, as illustrated in Figure 3. Notably, our investigation unveiled that vanillin maintained its initial docking positions consistently throughout the simulation, indicating minimal conformational alterations.

Evaluating the dynamics of hydrogen bonding is crucial for assessing the stability of the proteins and protein-ligand complexes. In this investigation, we quantified and explored the dynamics of hydrogen bonds intramolecularly within hTf and intermolecularly between hTf and vanillin (Figure 4). Upon the vanillin binding to hTf, we observed a minor increase in the number of intramolecular hydrogen bonds within hTf due to the higher compactness of the protein (Figure 4A). We also calculated the intermolecular hydrogen bonds formed between hTf and vanillin. Our analysis revealed the presence of up to 4 hydrogen bonds within the protein-ligand complex. We also observed that vanillin effectively nestled within hTf’s iron-binding pocket, forming 3-4 hydrogen bonds with some fluctuation and 1-2 hydrogen bonds with higher stability (Figure 4B). This thorough analysis provides additional supportive evidence for the stable binding of hTf and vanillin.

Principal component analysis (PCA) offers valuable insights into protein stability and conformational landscape (Stein et al., 2006). Here, we employed PCA to investigate the conformational dynamics of hTf and hTf-vanillin from the simulated trajectory. Figure 5 demonstrates the conformational dynamics of these systems along eigenvector 1 and eigenvector 2. The PCA plot indicates that in the presence of vanillin, hTf explores a more confined range of phase spaces (Figure 5A). We did not observe any significant overarching transitions or major conformational changes in the motion of hTf following vanillin binding (Figure 5B). This implies that the vanillin binding does not cause substantial conformational changes in hTf.

To gain deeper insights into the folding landscape of the systems, we delved into the free energy landscapes (FELs) generated through the simulated trajectories. Figure 6 presents a visual representation of the FELs for hTf and hTf-vanillin. In hTf’s FEL, it is indicated that hTf explores two distinct energy minima confined within 2-3 basins (Figure 6A). Conversely, when vanillin is introduced, a noticeable feature is a prominent single global minimum, mainly concentrated within a single basin, as illustrated in Figure 6B. Importantly, no significant or drastic conformational rearrangements are observed due to this binding event. This observation implies that vanillin binding introduces novel conformational states and influences hTf’s energy landscape. Examining these FELs provides valuable details about the conformational stability of hTf and hTf-vanillin.

After confirming hTf-vanillin complex formation through computational approaches, we used fluorescence binding to validate the obtained results. It is a powerful technique to study protein-ligand interactions, ascertaining the actual affinity of a ligand with the protein. Figure 7A shows the fluorescence emission spectra of hTf without vanillin and in the presence of varying vanillin concentrations (0–5 µM). It is apparent that with increasing vanillin concentrations, a decrease in the fluorescence intensity of hTf was apparent. The obtained quenching data was fitted into MSV equation as per earlier reports (Shamsi et al., 2020) with intercept of the obtained MSV plot (Figure 7B) giving the binding constant (K) of hTf-vanillin complex. Vanillin binds to hTf with a significant affinity forming a stable hTf-vanillin complex; evident from the obtained K of 1.92 × 10⁵ M^-1, suggestive of strong binding between vanillin and hTf.