PHD2, also known as prolyl hydroxylase domain 2, plays a crucial role in both normal and disease physiology. In normal physiology, PHD2 is involved in the regulation of cellular responses to oxygen levels, primarily through its interaction with hypoxia-inducible factors (HIFs)45. By hydroxylating specific proline residues on HIFs, PHD2 targets them for degradation under normoxic conditions, preventing their accumulation and subsequent activation of hypoxia-responsive genes2. This process ensures proper oxygen sensing and adaptation in various tissues and is vital for maintaining homeostasis9.
Intermolecular interactions are essential in drug discovery as they govern the binding affinity and specificity of drug compounds to their target proteins. These interactions include hydrogen bonding, pi-pi cation, pi-pi stacking, and van der Waals forces of attraction. Hydrogen bonding stabilizes drug-protein complexes, while pi-pi cation and pi-pi stacking interactions contribute to altered pharmacological effects46. Van der Waals forces, though weak individually, collectively play a significant role in the overall binding process. Understanding these interactions aids in rational drug design, identifying new drug targets, and predicting pharmacokinetics and safety47.
Molecular Docking Studies
To identify suitable lead molecules, docking studies were performed with a natural product library, using Glide to validate the hypothesis. The HIF-1α -binding site of PHD2 contains key residues, including ARG322 and Fe2+, which play an active role in substrate binding through hydrogen bonding interactions. Specifically, the positively charged amino acid ARG322 interacts with PRO564 of the HIF-1α protein within pocket one of PHD2, making it a critical residue in the enzyme's active site and facilitating the interaction with the HIF-1α hydroxylation site, PRO56448.
Table 1 presents the binding scores and affinities of compounds Salmochelin SX, Mycobactin, Staphyloferrin A, and Enterobactin towards PHD2. Among these compounds, SALMOCHELIN SX exhibits the highest affinity with a binding score of -9.527 Kcal/mol (MM/GBSA score: -42.58 Kcal/mol). It engages in crucial interactions with the target protein, forming hydrogen bonds with ASP254, TYR310, and ASP315, while accepting hydrogen bonds from TYR303 and ARG322. Additionally, Salmochelin SX coordinates with Fe2+, contributing significantly to its interaction with the key residue ARG322 and the co-factor Fe2+. This coordination of Fe2+ is vital for stabilizing the ligand-protein complex and plays a pivotal role in determining Salmochelin SX's strong affinity towards PHD2.
Table 1
Prioritized Compounds List
Compound ID | Docking score (Kcal/mol) | MM/GBSA (Kcal/mol) | Interacting Amino acid residue and cofactor (Fe2+) | Interacting Amino Acids residue and co-factors during MD Simulation | Amino Acids bridged to Ferrous ion during simulation |
Salmochelin SX | -9.527 | -42.58 | ASP254, TYR303, TYR310, ASP315, ARG322, Fe2+ | ASP254, ARG383 | HIS313, ASP315, HIS374 |
Mycobactin | -9.166 | -34.25 | VAL314, ARG322, Fe2+ | ASP254, THR387 | HIS313, ASP315, HIS374 |
Staphyloferrin A | -7.819 | -25.74 | TYR310, ASP315, ARG322, Fe2+ | ASP254, ARG383, THR387 | HIS313, ASP315, HIS374 |
Enterobactin | -7.302 | -38.16 | TYR310, ARG322 | ASP254, TYR310, ASP320, TRP389 | HIS313, ASP315, HIS374 |
Salmochelin SX showed interaction with ASP315, ASP254, TYR303 and ARG322 residue. The carboxyl group of compound forms hydrogen bonds with a distance of 1.910 Å with TYR310 respectively. The oxan ring of 3,4,5-trihydroxy-6-(hydroxymethyl) oxan-2-yl] phenyl} formamido) accepts the hydrogen bond from TYR303 (2.178 Å). And the hydroxypropionic group donates hydrogen bond to ASP315 and accepts hydrogen bond from key residue ARG322 with bond distance (2.028 Å) (Table 2). The negatively charged ASP254 forms a two-hydrogen bond (1.985 Å and 2.392 Å) with the hydroxyl group of the phenyl ring and Fe2+ interaction with the OH group. In the same way, we tested the interactions of other lead molecules (Fig. 2(a)).
Table 2
Hydrogen Bond distance of Salmochelin SX
Interaction | H Bond distance (Å) |
PHD2 TYR303: HH - D1:O1 | 2.178 |
PHD2 ARG322:HH12 - D1:O9 | 2.028 |
PHD2 ARG322:HH22 - D1:O9 | 2.237 |
PHD2: ARG383:HH12 - D1:O5 | 1.190 |
D1:H44 - PHD2: ASP254:OD2 | 1.985 |
D1:H45 - PHD2: ASP254:OD2 | 2.392 |
D1:H49 - PHD2: TYR310: OH | 1.910 |
Next, Mycobactin shows an affinity towards PHD2 with a binding score of -9.166 Kcal/mol. Like Salmochelin SX, it forms multiple interactions with the target protein, donating a hydrogen bond to ARG322 (Fig. 3(a)) and VAL314 donating and accepting hydrogen bonds from the compound. Furthermore, Mycobactin coordinates with Fe2+, which is crucial for its favorable association with the key residue ARG322 and the co-factor Fe2+. The coordination of Fe2+ enhances Mycobactin 's binding to PHD2, contributing to its significant affinity.
Table 3
Hydrogen Bond distance of Aerobactin
Interaction | H Bond distance (Å) |
PHD2:VAL314:H - Mycobactin: O10 | 1.972 |
PHD2:ARG322: HH12 - Mycobactin: O6 | 1.860 |
Mycobactin:H75 - PHD2:VAL314:O | 1.794 |
The two carboxyl groups of Staphyloferrin A form hydrogen bonds with ASP315 with distances (2.242 Å and 1.480 Å) and one of them, accepts hydrogen bond from key residue ARG322 (2.089Å). Similarly, at 2.182 Å TYR310 (Fig. 4(a) forms an interaction with the hydroxyl group of Staphyloferrin A (Table 4). There is also metal coordination with the carboxyl group. Finally, Enterobactin exhibits an affinity towards PHD2 with a binding score of -7.302 Kcal/mol. It forms bonds with the target protein by donating a hydrogen bond (Table 5) to TYR310 and engaging in a pi-cation, noncovalent interaction with the key residue ARG322(Fig. 5(a)).
Table 4
H Bond distance of Staphyloferrin A
Interaction | H Bond distance (Å) |
PHD2: TYR310: HH - Staphyloferrin A: O3 | 2.182 |
PHD2: ARG322:HH12 - Staphyloferrin A: O9 | 2.089 |
Staphyloferrin A:H51 - PHD2: TYR310: OH | 1.878 |
Staphyloferrin A:H54 - PHD2: ASP315:O | 2.242 |
Staphyloferrin A:H56 - PHD2: ASP315:OD1 | 1.480 |
Table 5
Hydrogen Bond distance of Enterobactin
Interaction | H Bond distance (Å) |
PHD2: TYR310: HH - Enterobactin:O11 | 2.174 |
Enterobactin:H70 - PHD2: TYR310: OH | 1.699 |
Enterobactin:H73 - PHD2: TYR310: OH | 2.170 |
The interaction analysis revealed that Salmochelin SX forms a variety of bonds with the target protein, including donating hydrogen bonds to ASP254, TYR310, and ASP315, while accepting hydrogen bonds from TYR303 and ARG322. Moreover, Salmochelin SX coordinates with Fe2+. These intricate bonds are crucial as they play a vital role in the compound's association with the key residue ARG322 and the co-factor Fe2+ within PHD2's active site. This strong and specific binding can modulate PHD2's enzymatic activity, potentially leading to therapeutic effects in the context of diseases associated with oxygen sensing and HIF regulation. Comparing the results to the Salmochelin SX, Aerobactin, Staphyloferrin A, and Enterobactin also exhibited significant affinities towards PHD2, albeit with slightly lower binding scores compared to Salmochelin SX. Aerobactin, like Salmochelin SX, forms multiple hydrogen bonds (Tables 2 & 3) with the target protein, while Staphyloferrin A (Table 4) and Enterobactin (Table 5) donate hydrogen bonds (Table 3) and engage in pi-cation, non-covalent interactions, respectively, with key residue ARG322.
Molecular Dynamic Simulation
The MD simulation revealed stable conformational changes in the target proteins upon ligand binding (Fig. 2(b) and (c). During the MD simulation of Salmochelin SX and PHD2 complex, the RMSD of alpha carbon and Salmochelin SX initially fluctuated but stabilized after 30 ns, with a final RMSD change of less than 3 Å. The Root Mean Square Fluctuation (RMSF) of the protein-ligand complex showed fluctuations only in the loop regions, a common occurrence during MD simulations (Fig. 2(d)). Ligand contact had no significant impact on protein RMSF, even in the beta-pleated sheets.
The simulation interaction fraction graph represented the strength of bonded interactions, with 40 bonds observed. Only a few bonds, like ASP254, HIS313, ASP315, HIS374, and ARG383 (Fig. 2(e)), remained in contact for more than 30% of the simulation time. ARG254 had the highest interaction, forming hydrogen bonds and water bridges, while ARG383 had similar interactions. The PHD2 enzyme's co-factor, Fe2+, interacted with histidine and aspartate, specifically HIS313, ASP315, and HIS374. This interaction was not present during molecular docking but formed during MD simulations. This interaction is crucial for inhibiting PHD2, as the ferrous ion plays a pivotal role in oxygen-mediated oxidoreductase reactions.
Mycobactin displayed similar results to the Salmochelin SX, with changed interactions (Fig. 3(b) compared to its docking results (Fig. 3(a)). RMSD values for the alpha carbon and ligands staying below 3 Å (Fig. 3(c). During the simulation, RMSF for the protein primarily fluctuated in the loop regions, as is typically observed in loop regions (Fig. 3(d)). Ligand interactions with the protein did not induce significant fluctuations. The Ligand contact graph, showing interaction fractions, revealed numerous interactions (Fig. 3(e)). Some of these interactions persisted with the protein for over 30% of the simulation time. In the two-dimensional pose view, ASP254, TYR310, and ARG383 formed hydrogen bonds with the protein-ligand, which had the highest interaction fraction scores (Fig. 3(e)). The ferrous ion's connection with HIS313, ASP315, and HIS374 played a substantial role in the ligand interaction (Fig. 3(b). These ionic interactions were nearly as significant as hydrogen bonding in terms of interaction strength. This interaction is crucial, and inhibiting the ferrous ion may reduce PHD2 enzyme activity. Furthermore, a few water bridges formed during the simulation, involving THR236 and THR387. The few interactions found during the docking studies were missing and additional interactions were observed. This may be due to the structural changes undergoes during MD simulation.
Further, Staphyloferrin A and Enterobactin have similar results like Salmochelin SX and Mycobactin in the case of protein and ligand RMSD (Figs. 4 & 5(c) and RMSF (Fig. 4&5(d). The post MD simulation interaction pose (Fig. 4(b) & 5(b)) showed changes in interaction compared to docking pose (Fig. 4(a) & 5(a)). RMSF results both siderophores (Staphyloferrin A and Enterobactin) was found changes in only loop regions of the PHD2(Fig. 4(d) & 5(d)). In Staphyloferrin A and PHD2 complex, ARG254, ARG383, and THR387 were formed hydrogen bonds (Fig. 4(e)). Ferrous ion showed the same interaction with Staphyloferrin A (Fig. 4(b)) as Salmochelin SX and Mycobactin. In this case water bridge observed with ASP254 and ASP325(Fig. 4(b)&(e)). Enterobactin was found stable throughout the simulations. The interaction of residues was changed during simulation compared to docked pose (Fig. 5(b)). Further, RMSD (Fig. 5(c)), RMSF (Fig. 5(d) were found stable and there are no changes were observed. Enterobactin showed ASP254, TYR310, ASP320 hydrogen bonded (Fig. 5(b)(e)) to the protein for more than 30% of simulation time and TYR389 showed pi-pi cationic interaction.
In a molecular dynamic simulation, we studied four compounds interacting with the PHD2 protein. These simulations showed that the protein and compounds changed their shapes but eventually stabilized. The compounds interacted with the protein without causing major fluctuations. We found specific interactions, like hydrogen bonds, that were crucial for inhibiting PHD2. These results differed from initial predictions using molecular docking, highlighting the importance of dynamic simulations in understanding how drugs interact with proteins.