Skip to main content
Download PDF
Download ePub

Molecular exploration of natural and synthetic compounds databases for promising hypoxia inducible factor (HIF) Prolyl-4- hydroxylase domain (PHD) inhibitors using molecular simulation and free energy calculations

BMC Chemistry volume 18, Article number: 236 (2024) Cite this article

Abstract

Hypoxia-inducible factors (HIFs) are transcription factors that regulate erythropoietin (EPO) synthesis and red blood cell (RBC) production. Prolyl-4-hydroxylase domain (PHD) enzymes are key regulators of HIF’s stability and activity. Inhibiting PHD enzymes can enhance HIF-mediated responses and have therapeutic potential for diseases such as anemia, cancer, stroke, ischemia, neurodegeneration, and inflammation. In this study, we searched for novel PHD inhibitors from four databases of natural products and synthetic compounds: AfroDb Natural Products, AnalytiCon Discovery Natural Product (NP), HIM-Herbal Ingredients In-Vivo Metabolism, and Herbal Ingredients’ Targets, with a total number of 13,597 compounds. We screened the candidate compounds by molecular docking and validated them by molecular dynamics simulations and free energy calculations. We identified four target hits (ZINC36378940, ZINC2005305, ZINC31164438, and ZINC67910437) that showed stronger binding affinity to PHD2 compared to the positive control, Vadadustat (AKB-6548), with docking scores of − 13.34 kcal/mol, − 12.76 kcal/mol, − 11.96 kcal/mol, − 11.41 kcal/mol, and − 9.04 kcal/mol, respectively. The target ligands chelated the active site iron and interacted with key residues (Arg 383, Tyr329, Tyr303) of PHD2, in a similar manner as Vadadustat. Moreover, the dynamic stability-based assessment revealed that they also exhibited stable dynamics and compact trajectories. Then the total binding free energy was calculated for each complex which revealed that the control has a TBE of − 31.26 ± 0.30 kcal/mol, ZINC36378940 reported a TBE of − 38.65 ± 0.51 kcal/mol, for the ZINC31164438 the TBE was − 26.16 ± 0.30 kcal/mol while the ZINC2005305 complex reported electrostatic energy of − 32.75 ± 0.58 kcal/mol. This shows that ZINC36378940 is the best hit than the other and therefore further investigation should be performed for the clinical usage. Our results suggest that these target hits are promising candidates that reserve further in vitro and in vivo validations as potential PHD inhibitors for the treatment of renal anemia, cancer, stroke, ischemia, neurodegeneration, and inflammation.

Peer Review reports

Introduction

Hypoxia is an abnormal physiological state where the body demands oxygen in excess amount while cellular oxygen level decreases [1]. This process with remain continuous or intermittently, and can be divided in to acute or chronic condition [2]. Hypoxia results from various underlying factors and hence plays a significant role in the onset and progression of numerous oxygen-related diseases [2,3,4,5]. PHD enzymes play a crucial role in hydroxylating specific prolyl residues on hypoxia-inducible factor-alpha (HIF-α), thereby regulating the ubiquitination and degradation of HIF. This intricate process enables the precise control of human oxygen homeostasis by influencing the expression of numerous downstream genes [6, 7]. Consequently, the acquisition of PHD inhibitors/agonists provides an effective mean to modulate the oxygen regulation pathway in vivo, holding significant potential for treating diseases associated with this pathway [8]. Among the PHD enzymes, PHD2 is the most widely distributed isoform and plays a dominant role in regulating the steady-state levels of HIF-α [9, 10]. Currently, the HIF-PHD pathway has emerged as a critical target for new drug research across various diseases, attracting a multitude of researchers to engage in its exploration. In 2018, Roxadustat developed by FibroGen, in partnership with AstraZeneca, obtained approval in China for the treatment of anemia caused by chronic kidney disease (CKD) (dialysis & non-dialysis patients) [11]. Subsequently, in 2020, Mitsubishi Tanabe Pharma in Japan received approval for its newly developed drug, Vadadustat (AKB-6548), for treating renal anemia [12] these inhibitors such as Vadadustat and Roxadustat (FG-4592) exhibit heteroaromatic structure [12] that occupies the entrance of the 2OG binding site and consequently blocks the HIF-α substrate from reaching the catalytic sites [13]. Common features shared by these inhibitors include a bidentate coordinating fragment containing nitrogen or oxygen atoms, which chelate with ferrous iron, and a carboxylate group or esters that interact with Arg383 and Tyr329. The docking results of AKB-6548 revealed that its isoquinoline nitrogen atom and the amide carbonyl oxygen of glycinamide formed a bidentate chelation with the active site iron. The carboxylic acid in the glycinamide group occupied the 2OG binding pocket and established a hydrogen bonding interaction with Arg383 and Tyr329. The 3-hydroxyl group in its isoquinoline moiety formed a hydrogen bond with the adjacent Tyr303 residue in the PHD2 active site. Additionally, isoquinoline rings were involved in pi-pi interactions with Trp389 and His313 residues, further enhancing the binding affinity (Fig. 1) [14]. Based on these findings, our study aimed to identify compounds capable of effectively chelating iron and binding to the PHD2 active site by interacting with these critical residues.

Fig. 1
figure 1

A Structure of Vadadustat B Structure of Roxadustat C The catalytic domain of PHD2 complexed with Vadadustat (AKB-6548) and showing Bidentate coordination of iron (orange sphere) and hydrogen bonding with Arg383, Tyr329 and Tyr303. D The catalytic domain of PHD2 complexed with Roxadustat (FG4529) and showing Bidentate coordination of iron (orange sphere) and hydrogen bonding with Arg383, Tyr329, and Tyr303

Full size image

In drug development, it is important to maintain diversity in the chemical space to address upcoming challenges such as safety concerns related to on-/off-target effects and metabolism of certain chemical functional groups [15]. For instance, several PHD drug candidates have been associated with side effects such as hypertension, thrombosis, erythrocytosis, liver failure, and hypertension [16, 17]. Natural products, with their diverse and complex structures coupled with limited or no side effects, position them as promising sources for drugs and medicines [18]. Natural products have been used in traditional remedies for thousands of years and have contributed to the discovery of many modern drugs [17]. Similarly, synthetic compounds created via chemical synthesis are integral in drug development. Synthetic molecules offer advantages such as higher modifiability, affordability, purity, stability and can also be designed to target specific receptors or enzymes, to improve their safety, selectivity, and potency. Notable synthetic drugs include aspirin, sulfonamides, paracetamol, chloroquine, penicillin, and morphine [19, 20].

Molecular docking and simulation are powerful tools for identifying new potential drug candidates for the treatment of various diseases. First, an active site of the target protein should be identified then a drug candidate compound library is prepared either by virtual screening (high-throughput screening (HTS) & structure-based drug design (SBDD). Subsequently, compounds identified with high docking scores are finally isolated, synthesized, or purchased for further testing [21]. Researchers at Tokai University, Japan, employed docking and simulation techniques, and have discovered TM6008 and TM6089 as novel PHD inhibitors [22]. These structures were inspired by the previously reported inhibitor FG-0041 [23]. On the other hand, utilizing SBDD methods, Janssen company has successfully developed a series of benzimidazole-2-glycinecarboxamide PHD inhibitors [24, 25].

This study screened four databases of natural and synthetic compounds such as AfroDb Natural Products, AnalytiCon Discovery NP, Herbal Ingredients In-Vivo Metabolism, and Herbal Ingredients Targets, to search for novel, effective and safe potential PHD inhibitors. The clinical drug Vadadustat was used as a positive control and re-docked with the PHD2 protein (PDB ID: 2G19). The re-docking results were validated by comparing them with the X-ray co-crystal structure of Vadadustat-PHD2 isoform (PDB ID: 7UMP). Molecular simulation and binding free energy studies were performed on the top three hits out of four from the screening. These candidates showed promising computational activity against PHD2 and required further validation through in vitro and in vivo evaluation to assess their drug potential as PHD inhibitors.

Methodology

Description of methodology

To conduct docking studies, the in-complex crystal structure of the HIF-PHD2 with 4HG ligand (PDB ID: 2G19) was obtained from the Protein Data Bank (PDB) maintained by the Research Collaboratory for Structural Bioinformatics (RCSB) [26]. The in-complex structure was processed in PyMOL to remove the ligand 4HG & water molecules to get the only clean PHD2 protein, The addition of hydrogen atoms and protein structure minimization were performed using Chimera software [27].

Data set preparation and docking

The 3D-SDF format of a wide range of natural products and synthesized compounds were obtained from AfroDb Natural Products (885) AnalytiCon Discovery NP (11247), Herbal Ingredients In-Vivo Metabolism (663), and Herbal Ingredients’ Targets (802) [28,29,30]. After preparing and filtering the databases, the compounds were subjected to screening against the active pocket of PHD2. To validate the effectiveness of the docking procedures, a positive control drug, Vadadustat (AKB-6548) was re-docked into the previously prepared cavity of the PHD2 protein with resolution 1.70 Å (PDB:2G19). To conduct ligand docking, we employed EasyDock Vina 2.0. Before the virtual screening, we focused on specific active site residues (Tyr310, His313, Asp315, Tyr329, His374, Tyr303, and Arg383) of the protein. A grid with dimensions X = 4.75, Y = 46.86, and Z = 24.5 was generated to accommodate these selected residues.

Virtual drug screening

The selected compounds were then subjected to a second screening at an exhaustiveness of 64 to refine the ranking and reduce false-positive outcomes. Induced fit docking (IFD), which allows flexibility in both the receptor and ligand, was applied to the top 10% of drugs from each database. IFD shares similarities with EasyDock Vina 2.0 but offers improved precision and reduced computational time. Finally, the best-hit compounds were subjected to further analysis using molecular dynamics simulation and free energy calculation [18].

Virtual screening of multiple databases was performed using EasyDock Vina 2.0. The screening process utilized the AutoDock4 algorithm to rank the most promising drug-like compounds. Before screening, all molecules were converted to PDBQT format. An initial quick screening was conducted with an exhaustiveness value of 16 to identify high-scoring compounds [31]. The selected compounds were then subjected to a second screening at an exhaustiveness of 64 to refine the ranking and reduce false-positive outcomes. Induced fit docking (IFD), which allows flexibility in both the receptor and ligand, was applied to the top 10% of drugs from each database [5, 32]. IFD shares similarities with EasyDock Vina 2.0 but offers improved precision and reduced computational time. Finally, the best-hit compounds were subjected to further analysis using molecular dynamics simulation and free energy calculation [33]. To observe the protein–ligand interactions Schrödinger Maestro free Academic version 2018-1 (for visualization only) and PyMOL software were used [34, 35].

Molecular and ADMET analysis of lead compounds

An examination of the Molecular, pharmacokinetic properties and ADMET profiles was conducted for the four hits leading molecular entities, along with a control substance, using the pkCSM online tool (https://biosig.lab.uq.edu.au/pkcsm/) [36]. The critical assessment utilized a comprehensive set of molecular and ADMET variables to form the narrative of each compound’s pharmacokinetic behavior. The essence of this analysis lies in its ability to provide a broad perspective on the physiological interactions of the molecular entities. The pkCSM platform offers predictions on various Molecule properties and pharmacokinetic aspects such as absorption rates, distribution within the system, metabolic pathways, excretion methods, and potential toxicity [37]. These prognostications are crucial during the early drug development phase, enabling the prioritization of compounds that exhibit the most favorable interaction with biological systems [38].

Molecular dynamics simulation

Atomistic investigation of the binding of the top hits to the PHD2 active pocket was achieved by using the AMBER21 simulation tool [39]. The antechamber, an integrated module in AMBER, was used to generate the drug topologies while the Amber general force field (GAFF) and ff19SB force fields were recruited for the solvated complexes to complete the simulation. An optimal point charge (OPC) solvation model and sodium (Na+) ions were added for neutralization. Multi-step energy minimization followed by heating and equilibration was completed. Mesh Ewald (PME) algorithm [40], long-range electrostatic interactions Van der Waals interactions, and Columbic interactions of short-range were used. Langevin thermostat and Berendsen barostat were for pressure and temperature control. A 100ns simulation for each complex with a time step of 2fs was performed. The dynamics, stability, SASA and other features of the ligand–protein complexes were evaluated by using CPPTRAJ and PTRAJ [18, 41].

Binding free energy calculation using MM/GBSA

The MM/GBSA technique is frequently employed in the field of drug development. This approach can be used to determine the most promising candidates by predicting the essential interactions and improving the quality of a lead molecule and specificity by computing the binding free energy of a ligand and a protein. This approach combines molecular mechanics simulations, which describe the interactions between atoms, with implicit solvent models, which describe the interactions between the protein and solvent [31, 42,43,44,45]. Hence, we also applied this approach here to accurately compute the total binding free energy of the top complexes. Mathematically the binding free energy can be estimated as:

$$"\Delta G(bind) = \Delta G(complex) - [\Delta G(receptor) + \Delta G(ligand)]"$$
(1)

Different contributing components of total binding energy were calculated by the following equation:

$$"G=Gbond + Gele +GvdW +Gpol +Gnpol"$$
(2)

It has a wide range of applications i.e., used to estimate the binding energy for proteins in different studies including SARS-CoV-2 and neurological disorders [46,47,48,49,50,51].

Results and discussion

PHD enzymes play an important role in oxygen sensing and HIF regulation and have been associated with various essential human diseases such as anemia, inflammation, cancer, rheumatoid arthritis, strokes, spinal cord injury, and von Hippel–Lindau disease-related renal cell carcinoma [52,53,54]. In this study, we explore four natural products’ chemical spaces for the discovery of novel PHD2 inhibitors. A total of 13,597 compounds have been screened from AfroDb Natural Products, AnalytiCon Discovery NP, Herbal Ingredients In-Vivo Metabolism, and Herbal Ingredients Targets against the PHD2 active site. The clean and prepared PHD2 protein (2G19) was first docked with a positive control drug, Vadadustat (AKB-6548) (Fig. 1A–D). Potential drug candidates were compared with the positive control, Vadadustat (AKB-6548), with a docking score of − 9.04 kcal/mol. The interaction patterns and binding energies of these targets are closely aligned with positive control and other PHD2 crystal structure complexes available in the Protein Data Bank (PDB), such as 7UMP, 2HBT, and 3OUH [18]. Figure 2A and B show 2OG and  NOG form five-membered chelates, while Daprodustat (GSK1278863) forms a six-membered chelate with the catalytic iron residue.

Fig. 2
figure 2

A) showing 2-oxoglutarate (2OG) and N-oxalyglycine (NOG) form five-membered chelates, while B) Daprodustat (GSK127863) forms a six-membered chelate with the catalytic iron residue. All inhibitors form a salt bridge with the binding pocket residue Arg383

Full size image

Compounds chemical space exploration

Initial screening of 13,597 compounds revealed that 63 compounds have better docking scores than the control drug AKB-6548, respectively. The top 10% of compounds were re-evaluated for interactions with the PHD2 active site residues. Induced fit docking (IFD) subsequently revealed docking scores ranging between − 12.76 kcal/mol and − 11.96 kcal/mol, respectively. The structures of the top hit compounds, their docking scores, interacting residues, interaction nature, and interaction distances (Å) are summarized in Table 1.

Table 1 2D structures of the top hits and control ligand with their names, database IDs, docking scores, interacting residues, and nature are given
Full size table

The first target hit, ZINC36378940, (2S)-2-[5,7-dihydroxy-3-oxo-6-[(2E,6E)-3,7,11-trimethyldodeca-2,6,10-trienyl]-1H-isoindol-2-yl] pentane dioic acid, gives a docking score of − 13.346. It targets essential residues such as Tyr329, and Arg383 which are required for ligand binding to PHD2. ZINC36378940 may block the 2OG binding pocket more potently than the control ligand AKB-6548, given its relatively high docking score. The 3D and 2D interaction patterns of ZINC36378940 in the PHD2 active site are shown in Fig. 3A and B. The carbonyl oxygens of its pyrrolidinone ring and N-acetyl-D-glutamic acid side chain chelate the Fe2+ ion in a bidentate manner. Salt bridge interactions were observed to form between the carboxylate end of the N-acetyl-D-glutamic acid with Arg383, occupying the 2OG pocket, further stabilizing the complex. In total, four hydrogen bonds were observed between residues Tyr303 (2.44Å), Tyr329 (1.67Å), and Arg383 (1.72Å, & 1.90Å). Moving on to ZINC2005305, also known as 5-methyltetrahydrofolate (5-MTHF), is the bio-active form of folate. 5-MTHF plays a vital role in serotonin synthesis and DNA production. Therapeutically, it prevents neural tube defects and improves vascular function. 5-MTHF has also been shown to prevent cervical dysplasia and neoplasia, and aid in vitiligo. Folate deficiency could contribute to neurological and psychiatric symptoms [55]. Supplementation of 5-MTHF is beneficial for folate repletion. Naturally occurring 5-MTHF has important advantages over synthetic folic acid, it is well absorbed even when gastrointestinal pH is altered and its bioavailability is not affected by metabolic defects. ZINC2005305 shows a high docking score of − 12.764. Figure 3C and D shows the 2D and 3D illustrations of ZINC2005305 in the PHD2 binding pocket. ZINC2005305 formed bidentate coordination bonds with the active site Fe, through its carbonyl oxygens of the N-acetyl-D-glutamic acid chain while its terminal carboxylic acid interacts via hydrogen bonding with Arg383 (1.58 Å & 1.76 Å) and Tyr329 (Å). Additionally, hydrogen bonding interactions were also observed between the N, NH, and NH2 of pyrimidinone with protein residues Thr236 (2.03Å), Ile256 (2.33Å), Trp258 (1.95Å), effectively locking the PHD2 entrance and hence contributing to the extra stability of the complex.

Fig. 3
figure 3

A Showing 2D interaction pattern of ZINC36378940 With PHD2 B shows the 3D interaction pattern ZINC36378940 With PHD2. C Showing 2D interaction pattern of ZINC2005305 With PHD2 D shows the 3D interaction pattern ZINC2005305 With PHD2

Full size image

Our next target hit is, ZINC31164438, (2S)-2-[(10R,11S,14R,15R,16S)-14-hydroxy-10,14-dimethyl-10-(4-methylpent-3-enyl)-5-oxo-9,18-dioxa-4-azapentacyclo[13.2.1.02,6.08,17.011,16] octadic-1,6,8 (17)-trien-4-yl]pentanedioic acid, with a docking score of -11.964. The binding mode of ZINC31164438 in the PHD2 active site is given in Fig. 4A and B. It shows bidentate coordination of Fe2+, via carbonyl oxygen of the d-pyrrolone ring and carboxyl group of the pentandioic acid chain. The key residues Tyr329 (1.91 Å), and Arg383 (1.71 Å & 2.01 Å) were also involved in the hydrogen bonding via the carboxyl region of the pentanedioic acid chain competing 2OG. The hydroxyl group of methylcyclohexane ring establishes a hydrogen bond with Asp254 (1.97 Å) further stabilizing the protein and ligand complex. Lastly, ZINC67910437, 2-(3-acetyloxy-4’-hydroxy-4,4,7,8a-tetramethyl-6’-oxospiro [2,3,4a, 5,6,7-hexahydro-1H-naphthalene-8,2’-3,8-dihydrofuro[2,3-e]isoindole]-7’-yl) pentanedioic acid, exhibited a docking score of − 11.41. The 2D and 3D binding mode of ZINC67910437 in the PHD2 active site is given in Fig. 4C and D. The analysis of interactions revealed bidentate coordination with Fe2+ through carbonyl oxygen of the pyrrole ring and carboxyl group from the pentanoic acid chain. Importantly, the compound also engaged in four hydrogen bonding interactions through the carboxyl region of the pentanedioic acid chain with key residues like Tyr329 (1.91 Å), Tyr303(2.17 Å) and Arg383 (1.81Å & 1.90 Å). Moreover, the hydrogen bond between the hydroxyl group of the benzene ring to Arg322 residue (1.97 Å) and the carbonyl oxygen of easter to Leu240 further locks the ligand at the PHD2 active site. Our explored compounds have tight-binding interactions with the catalytic iron via carbonyl oxygen and key restudies like Arg383, Tyr329, and Tyr303. The same binding mechanism can be seen in phase 3 clinical trials [2-(1,3-dicyclohexyl-6-hydroxy-2,4-dioxo-1,2,3,4- tetrahydro pyrimidine-5-carboxamides) acetic acid] (Daprodustat), which chelate the iron via two carbonyl oxygen and salt bridging the important residue Arg383, Figs. 3 and 4 [56]. Moreover, these shortlisted compounds possess better docking scores than our previously reported compounds that target PHD2 in different cancers [57].

Fig. 4
figure 4

A Showing 2D interaction pattern of ZINC31164438 with PHD2 B shows the 3D interaction pattern ZINC31164438 with PHD2. C Showing 2D interaction pattern of ZINC67910437 with PHD2 D shows the 3D interaction pattern ZINC67910437 With PHD2

Full size image

Molecular dynamics simulations analysis of drugs-PHD2 complexes

Examining the dynamic stability of molecular interactions within a binding cavity is crucial for understanding the binding strength of a small molecular ligand. Simulation trajectories are employed to uncover this stability, wherein various measures can be computed, including the root mean square deviation (RMSD). This calculation aids in illustrating the dynamic stability of molecules involved in interactions and offers valuable insights into the strength of binding. In this study, a 50 ns simulation was conducted to determine the RMSD and evaluate the stability of the leading protein-drug compounds within a dynamic environment. Based on the analysis of Fig. 1, it can be observed that throughout the 100 ns simulation, no substantial convergence was observed in the RMSD values of the four top hits. This indicates a stable behavior of the protein-drug complexes. Specifically, the tophit1-PHD2 system reached equilibrium at 10 ns and remained stable until the end of the simulation, with an average RMSD value of 2 Å (Fig. 5a). Both the top hit 2 and veda-PHD2 complexes followed a similar pattern, maintaining an average RMSD value of 2.5 Å throughout the simulation period. These systems equilibrated at 3 ns, and the RMSD values gradually increased, reaching 3 Å (Fig. 5b, c). However, in the case of top hit 2, the system equilibrated at 2 ns with an initial RMSD value of approximately 2 Å, which steadily increased and reached 3.5 Å at 100 ns. For top hit 1, the average RMSD was observed to be 2.5 Å, and little convergence was noticed between the 40 and 70 ns time period (Fig. 5d). These findings demonstrate that top hits exhibit consistent dynamics, indicating their stability and potential for interaction with PHD2. Dynamic stability is always associated with stronger pharmacological potential and therefore is an essential parameter in discovering novel therapeutics. For instance, our compounds possess a stably dynamic pattern which aligns with the previous literature reported to target PHD2 in various diseases [57,58,59].

Fig. 5
figure 5

Stability analysis of drugs-PHD2 complexes by RMSD calculation. a Showing the RMSD value of the ZINC36378940-PHD2 complex. b Showing the RMSD value of the ZINC31164438-PHD2 complex. c Showing the RMSD value of ZINC2005305-PHD2 complex. d Showing the RMSD value of the AKB-6548-PHD2 complex

Full size image

In order to analyze the behavior of each complex in a dynamic environment and determine the occurrences of binding and unbinding events, the structural compactness was assessed. This involved measuring the radius of gyration (Rg) over time, which provided insights into the level of structural compactness. Figure 6 illustrates the results, showing a similar trend in compactness compared to the Root Mean Square Deviation (RMSD). In the case of the top hit 1, its structure remained compact throughout the simulation with an Rg value of 17.6 Å until 30 ns. After that point, the Rg value gradually increased to 18 Å, but no significant convergence was observed within the simulation timeframe (Fig. 6a). Similarly, for top hit 2, an average Rg value of 17.4 Å was maintained until 20 ns followed by a gradual decrease to 17.2 Å and reached back 17.6 Å till 80 ns (Fig. 6b). Top hit 3 exhibited a similar Rg pattern as top hit 2 (Fig. 6c). On the other hand, in the veda-PHD2 system, the Rg value increased until 40 ns with the 18.0 Å and then remained stable until 70 ns (Fig. 6d). A more compact nature determines the pharmacological potential of the binding molecules and therefore can be used as an essential approach to deciphering the pharmacological behavior of the respective molecules. These findings strongly align with the previous report where a more compact protein dynamics produces better results with minimal unbinding events than the others [60].

Fig. 6
figure 6

Compactness analysis of drugs-PHD2 complexes by Rg calculation. a Showing the Rg value of the ZINC36378940-PHD2 complex. b Showing the Rg value of the ZINC31164438-PHD2 complex. c Showing the Rg value of ZINC2005305 − PHD2 complex. d Showing the Rg value of AKB-6548-PHD2complex

Full size image

The calculation of root mean square fluctuation (RMSF) was performed to gain a more comprehensive understanding of the variations in flexibility at the residue level for both the wild type and its variants. The RMSF analysis provides valuable insights into the local flexibility of a molecule, which in turn influences crucial aspects such as intermolecular binding strength, molecular recognition, and overall biological function. Regions with higher RMSF values exhibit greater flexibility, while those with lower RMSF values indicate more stability. The results depicted in Fig. 7 illustrate a consistent trend in residual flexibility across all the complexes with a mean RMSF of approximately 1 Å. However, a significant fluctuation was observed in 50–70 amino acid residues. The result of RMSF is consistent with the RMSD indicating that the tophit1 has a stable binding affinity with the PHD2 protein.

Fig. 7
figure 7

Residual fluctuation analysis of drugs-PHD2 complexes by calculating RMSF

Full size image

In order to assess the strength of molecular interactions, hydrogen bond analysis is employed. To examine alterations in the hydrogen bonding pattern throughout the simulation, the total count of hydrogen bonds in each trajectory was determined. The outcomes of this computation, illustrating the hydrogen bond occurrences for each complex as a time-dependent parameter, are visually represented in Fig. 8. The average number of hydrogen bonds in the top hits complexes was found to be approximately 120, while the Veda-PHD2 complex exhibited an average of 110 hydrogen bonds. This indicates that all the complexes demonstrate a strong and well-established hydrogen bonding network, indicating stable interactions between the identified lead drugs and PHD2. These findings align with the results obtained from molecular docking, RMSD, and RMSF analyses, further reinforcing the stability of the complexes and providing additional supporting evidence.

Fig. 8
figure 8

The average hydrogen bonding analysis of lead drugs-PHD2 complexes. a Showing the number of hydrogen bonds in the ZINC36378940-PHD2 complex. b Showing the number of hydrogen bonds in the ZINC31164438-PHD2 complex. c Showing the number of hydrogen bonds in the ZINC2005305-PHD2 complex. d Showing the number of hydrogen bonds in AKB-6548-PHD2complex

Full size image

To show the binding variations caused by the simulation and to determine the bonding distinctions we used the structures from each trajectory and compared with the control complex. The post simulation hydrogen bonding analysis revealed significant variations in the interaction paradigm. The first target hit, ZINC36378940 along with iron chelation bidentate manner also targets essential residues such as Asp254, Tyr303, Tyr329, and Arg383 which are required for ligand binding to PHD2. The 2D interaction patterns of ZINC36378940 in the PHD2 active site are shown in Fig. 9A and B. The carbonyl and hydroxyl groups of its pyrrolidinone ring and N-acetyl-D glutamic acid side chain chelate the active Fe2+ ion engage in bidentate ligation and one salt bridge (2.12 Å, 1.97 Å & 1.96 Å). The key residues Arg383 (1.72 Å, & 1.69 Å), Tyr303 (1.63 Å), and Tyr329 (1.60 Å) establish hydrogen bond interactions via the carboxylate end of the N-acetyl-D-glutamic acid, occupying the 2OG pocket. Additional hydrogen bond interactions were also noted with Asp254 (1.55 Å) from resorcinol hydroxyl group further stabilizing the complex. The resorcinol ring also interacts with Trp389 (5.30 Å) via pi-pi stacking. The apparency of additional interactions like Hydrogen bonding with key residue Asp254 and Trp389 further enhance the stability of complex. Figure 9C and D Shows the 2D illustrations of second ranked hit ZINC2005305 in the PHD2 binding pocket. The carboxylate moiety near amide region of the N-acetyl-d-glutamic acid chain showed mono ligation and salt bridging (1.92 Å and 1.92 Å) with Fe2+. While the terminal, carboxylic acid establishes four hydrogen bonding with key residues Arg383 (1.58 Å), Tyr329 (1.59 Å), and Tyr303 (1.63 Å). The N, NH and NH2 moieties of methyl trihydropterin interaction with Thr336 (1.86 Å), Trp258 (1.97 Å) and Ile256 (2.54 and 1.88 Å). further locking the PHD2 entrance and hence contributing to the extra stability of the complex. There is shift of hydrogen bond with Thr236 from pyrimidinone to pyrazine NH During post simulation the appearance on additional hydrogen bond with Trp256 (1.88 Å) and Tyr303 (1.63 Å) were observed, similarly loss of one metal bond with Fe2+ were also noted when compared to pre simulation complex.

Fig. 9
figure 9

Post simulation interaction pattern A Showing 2D interaction pattern of ZINC36378940 With PHD2 B Showing 3D interaction pattern of ZINC36378940 With PHD2 C Showing 2D interaction pattern of ZINC2005305 With PHD2 D Showing 2D interaction pattern of ZINC2005305 With PHD2

Full size image

Our next target hit is, ZINC31164438, the 2D binding mode of ZINC31164438 in the PHD2 active site is given in Fig. 10A and B. The carbonyl oxygen of pyrrolone ring and carboxylate of pentandioic acid showed monodentate ligation and salt bridging with active site Fe2+ (2.05 and 1.88 Å). The key residues Tyr303 (1.94 Å), Tyr329 (1.62 Å), and Arg383 (1.79 Å & 1.73 Å) were involved in the hydrogen bonding via the extreme carboxyl region of the pentanedioic acid chain. The hydroxyl group of methyl cyclohexanol ring also establishes additional hydrogen bonding interaction with another Asp254 (1.70 Å) further stabilizing the protein and ligand complex. The appearance of hydrogen bond with key residue Tyr303 and loss of one metal bond with active site of Fe2+ from carbonyl oxygen carboxylate of pentandioic acid were resulted in post simulation. The 2D interaction of positive control is illustrated in Fig. 10C and D. In case of AKP-6548, The Nitrogen of Pyridine and carbonyl oxygen of glycinamide chain chelate bidentally the active site Fe2+ (2.34 and 2.46 Å). The carboxylate side end establishes three hydrogen bonding with key residues with Arg383 (1.73 and 1.78 Å) and Tyr329 (1.86 Å). The pyridine ring also establishes pi-pi stacking with His313 (5.27 Å) which stabilize the complex more. The interactions like hydrogen bonding with Tyr303 from pyridine Hydroxyl group and pi-pi stacking with Trp389 were absent in post simulations.

Fig. 10
figure 10

Post simulation interaction pattern A Showing 2D interaction pattern of ZINC31164438 With PHD2 B Showing 3D interaction pattern of ZINC31164438 With PHD2 C Showing 2D interaction pattern of AKB-6548 with PHD2 D Showing 2D interaction pattern of AKB-6548 with PHD2

Full size image

Solvent-accessible surface area (SASA) analysis

Solvent-accessible surface area (SASA) analysis is essential for understanding protein–ligand interactions and binding mechanisms. It measures the surface area of a protein and ligand that is accessible to solvent molecules, identifying the exposed and buried regions when a complex forms. SASA offers insights into the hydrophobic and hydrophilic properties of binding interfaces, conformational changes, and the energetics of molecular recognition, which are crucial for structure-based drug design and protein engineering. It can be seen that all the complexes demonstrated a SASA range in between 10,500 A2 to 12,500 A2. In the ZINC36378940-PHD2 complex the between 10,500 Å2 and 12,500 Å2 and maintain a stable state thus implies a steady state interaction between the receptor and ligand with no significant conformational changes or exposure of new surface areas to the solvent. The SASA for the ZINC36378940-PHD2 complex is given in Fig. 11a. On the other hand, the ZINC31164438-PHD complex reported as increasing trend over time. The increasing SASA suggests for the ZINC31164438-PHD2 complex that is a testimony of the undergoing conformational changes which consequently indicate the opening of binding pockets potentially impacting the stability and binding efficiency of the ligand. The SASA for the ZINC31164438-PHD2 complex is given in Fig. 11b. The SASA values, which range from around 11,000 Å2 to 12,500 Å2, for the ZINC2005305-PHD2 complex fluctuate but typically display a minor rising trend. Variations in solvent exposure are caused by conformational changes in the ZINC2005305-PHD2 complex, as indicated by the SASA, which is variable and slightly rising. These alterations might be the result of dynamic interactions between the protein and the ligand, which could have an impact on the effectiveness and stability of binding. The SASA for the ZINC2005305-PHD2 complex is given in Fig. 11c. Finally, The SASA values for the AKB-6548-PHD2 complex show an increasing trend, starting at approximately 11,000 Å2 and peaking at around 13,000 Å2. This growing SASA, similar to that observed in the ZINC311644438-PHD2 complex, indicates conformational changes that expose more surface area to the solvent. This suggests a less compact structure and potential alterations in the binding interface, which could affect the ligand’s stability and strength of interaction. The SASA for the AKB-6548-PHD2 complex is given in Fig. 11d. Overall this shows the stability and exposure of ligands into solvent and thereby reports that the shortlisted hits are stable and remained bound in the cavity.

Fig. 11
figure 11

The Solvent-accessible surface area (SASA) analysis over 100 ns MD of lead drugs-PHD2 complexes. a Showing the number of SASA of ZINC36378940-PHD2 complex. b Showing the SASA of the ZINC31164438-PHD2 complex. c Showing the SASA of the ZINC2005305-PHD2 complex. d Showing the SASA of AKB-6548-PHD2complex

Full size image

Evaluation of the binding free energy

For the accurate re-evaluation of the docking results, the binding free energy is the best, cheaper and alternative approach to validate the docking results. It has been more widely used than the other alchemical and traditional approaches available. We calculated the binding free energy (BFE) for each complex and compared it with the control drug which reported that ZINC36378940 is the best hit against the PHD2. The vdW results revealed that the control has a vdW of − 26.90 ± 0.38 kcal/mol while the identified best three hits reported a vdW of − 38.64 ± 0.38 for the ZINC36378940, − 31.79 ± 0.32 for the ZINC31164438 and ZINC2005305 reported a vdW of − 30.58 ± 0.44 kcal/mol. This shows that the vdW of our top hits are better than the control and therefore may produce stronger pharmacological inhibition than the others. Moreover, the electrostatic energy for each complex was calculated to be − 91.86 ± 0.81 kcal/mol for the control, for the ZINC36378940 the electrostatic energy was − 145.92 ± 2.10 kcal/mol, for the ZINC31164438 the electrostatic energy was − 125.15 ± 1.7 kcal/mol while the ZINC2005305 complex reported electrostatic energy of − 112.18 ± 1.89 kcal/mol. Then the total binding free energy was calculated for each complex which revealed that the control has a TBE of − 31.26 ± 0.30 kcal/mol, ZINC36378940 reported a TBE of − 38.65 ± 0.51 kcal/mol, for the ZINC31164438 the TBE was − 26.16 ± 0.30 kcal/mol while the ZINC2005305 complex reported electrostatic energy of − 32.75 ± 0.58 kcal/mol. This shows that ZINC36378940 is the best hit than the other and therefore further investigation should be performed for the clinical usage. The binding free energy results are given in Table 2.

Table 2 Binding free energy calculation using the MM-GBSA approach. All the results are given in kcal/mol
Full size table

Evaluation of ADMET properties of lead compounds

In the domain of medicinal chemistry, the molecular descriptors of ZINC36378940, ZINC2005305, ZINC31164438, ZINC67910437, and AKB-6548 are scrutinized and summarized in Table 3 to elucidate their physicochemical constitution and its implications on pharmacodynamics and pharmacokinetics. The top hit ZINC36378940, with a molecular mass of 499.60 g/mol and a partition coefficient (LogP) of 4.80, manifests a lipophilic propensity, facilitating membrane permeation, articulated through 13 rotatable bonds, 7 hydrogen bond acceptors, and 4 hydrogen bond donors, within a molecular surface area of 135.37 Å2. In juxtaposition, ZINC2005305 as second hit, has a molecular weight of 459.46 g/mol and a LogP of -0.25, exhibits an augmented hydrophilicity, which may enhance aqueous solubility and renal clearance, reflected in a more intricate hydrogen bonding network of 9 acceptors and 7 donors, across a surface area of 187.55 Å2. Advancing to third hit ZINC31164438, the entity weighs 513.58 g/mol with a LogP of 3.86, suggesting a balanced lipophilic-hydrophilic interplay, potentially influencing its oral bioavailability and tissue distribution, as denoted by 8 rotatable bonds, 6 hydrogen acceptors, and 3 hydrogen donors, extending over a surface area of 215.60 Å2. The last hit ZINC67910437, the most massive at 557.64 g/mol with a LogP of 4.14, intimates a moderate lipophilicity, which may affect its pharmacokinetic profile, including absorption, distribution, metabolism, and excretion (ADME), characterized by 6 rotatable bonds, 7 hydrogen acceptors, and 3 hydrogen donors, over a molecular surface area of 233.182 Å2. Finally, the reference molecule, AKB-6548, presents a molecular weight of 306.75 g/mol and a LogP of 1.92, indicating a lower lipophilicity compared to the other compounds, which may result in distinct ADME characteristics, delineated by a simpler structural flexibility with 4 rotatable bonds, 4 hydrogen acceptors, and 3 hydrogen donors, within a confined surface area of 124.56 Å2. Collectively, these molecular depictions provide a comprehensive understanding of the compounds’ physicochemical profiles, which are paramount in predicting their behavior in biological systems and their potential as therapeutic agents. The interplay of these molecular properties with biological membranes, enzymes, and receptors will ultimately dictate their efficacy and safety as drug candidates.

Table 3 ADMET properties analysis of the identified hits using mCSM-PK
Full size table

The ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) properties of the lead compounds are critical for evaluating their potential as therapeutic agents. Upon examining the ADMET properties of the listed compounds, it is evident as summarized in Table 3 that most exhibit significant or comparable characteristics to the positive control drugs, AKB-6548. The ADMET analysis of the Top hit compound identified as ZINC36378940 reveals a comprehensive profile that is critical in the assessment of its pharmacokinetic and toxicological properties. The compound exhibits a logP value of − 2.858, indicating its hydrophilicity which may affect its absorption and distribution. The water solubility is logged at 0.167, suggesting moderate solubility which is beneficial for bioavailability. The topological polar surface area (TPSA) is 38.828, which is within the range that allows for adequate oral absorption. The logS value of − 2.735 further supports its solubility characteristics. The compound is predicted to have no inhibitory effects on the CYP450 enzymes, which is favorable as it reduces the potential for drug-drug interactions. The bioavailability score is 0.808, indicating a high likelihood of good bioavailability. The BBB permeability is − 0.951, suggesting that the compound may not easily cross the blood–brain barrier, which could be advantageous or disadvantageous depending on the therapeutic target. The human intestinal absorption (HIA) is 0.258, which is relatively low and could impact the efficacy of oral administration. The P-glycoprotein substrate and inhibitor predictions are negative, which is beneficial for drug efflux and resistance concerns. The renal organic cation transporter (OCT) predictions are also negative, indicating no significant interaction. The AMES toxicity and hERG inhibition predictions are negative, suggesting a lower risk of genotoxicity and cardiotoxicity, respectively. The skin sensitization and eye irritation predictions are also negative, which is promising for safety profiles.

Moving to second hit ZINC2005305 the ADMET profile presents a set of properties that are crucial for determining its pharmacokinetic behavior and potential as a therapeutic agent. The logP value of -2.857 suggests a high degree of hydrophilicity, which often correlates with good solubility but may pose challenges for cellular membrane permeability. The blood–brain barrier (BBB) permeability score of − 0.931 indicates that the compound is unlikely to penetrate the BBB, which is important for drugs targeting the central nervous system. A topological polar surface area (TPSA) of 5.645 and a logS value of − 2.735 reflect on its solubility and absorption characteristics, which are essential for oral bioavailability. The compound is not an inhibitor of the CYP450 enzymes, suggesting a reduced risk of drug-drug interactions. The human intestinal absorption (HIA) score is 0.472, which is moderate and may require optimization for effective oral administration. The P-glycoprotein substrate and inhibitor predictions are negative, indicating no significant interaction with this efflux transporter. The renal organic cation transporter (OCT) predictions are also negative, which is favorable for renal clearance. The bioavailability score of 0.161 is low, suggesting that the compound may have limited bioavailability in its current form. The AMES toxicity and hERG inhibition predictions are negative, which is promising for its safety profile. The skin sensitization and eye irritation predictions are negative, indicating a lower risk of adverse dermatological or ocular effects.

The ADMET analysis of third hit, ZINC31164438 is characterized by several key parameters that inform its pharmacokinetic and toxicological potential. The logP value of − 2.809 indicates a preference for aqueous environments, which can influence absorption and distribution within the body. A water solubility log value of 0.261 suggests moderate solubility, which is generally favorable for oral administration. The topological polar surface area (TPSA) is 55.89, a value that is indicative of good oral absorption potential. The logS value of − 2.735 aligns with these findings, supporting the compound’s solubility profile.

The compound is not an inhibitor of the CYP450 enzymes, which is advantageous as it minimizes the risk of metabolic drug-drug interactions. The bioavailability score of 0.121 is relatively low, indicating that the compound may have limited bioavailability in its current form. The blood–brain barrier (BBB) permeability score of − 0.884 suggests that the compound is less likely to cross the BBB, which is relevant for drugs intended for central nervous system activity. The human intestinal absorption (HIA) value of 0.2 is on the lower end, which could impact the drug’s efficacy when administered orally.

The compound is not a substrate for P-glycoprotein, which is beneficial for drug resistance profiles. It is, however, a substrate for the renal organic cation transporter (OCT), with a score of 0.378, indicating potential interaction with this transporter. The predictions for AMES toxicity and hERG inhibition are negative, suggesting a lower risk of genotoxicity and cardiotoxicity, respectively. The compound is also predicted to be non-sensitizing to the skin and non-irritating to the eyes, which is promising for its safety profile.

Last hit ZINC67910437 ADMET analysis indicates specific pharmacokinetic properties. The logP value of − 2.813 suggests that the compound is hydrophilic, which may influence its absorption and distribution. A negative blood–brain barrier (BBB) permeability score of − 0.4 implies that the compound is unlikely to cross the BBB, which is significant for drugs targeting the central nervous system. The topological polar surface area (TPSA) of 49.58 and a logS value of − 2.735 indicate moderate solubility, which can affect bioavailability.

The compound is not an inhibitor of the CYP450 enzymes, which is favorable as it reduces the potential for drug-drug interactions. The human intestinal absorption (HIA) score is 0.255, suggesting moderate absorption potential. The compound is not a substrate for P-glycoprotein, which is beneficial for drug resistance profiles. However, it is a substrate for the renal organic cation transporter (OCT), with a score of 0.783, indicating potential interaction with this transporter.

The bioavailability score of − 0.045 is low, suggesting that the compound may have limited bioavailability in its current form. The predictions for AMES toxicity and hERG inhibition are negative, which is promising for its safety profile. The compound is also predicted to be non-sensitizing to the skin and non-irritating to the eyes, which is favorable for its overall safety.

Lastly The ADMET profile for the compound known as AKB-6548, which is a positive control, indicates a set of pharmacokinetic properties that are significant for its evaluation as a therapeutic agent. The logP value of − 2.612 suggests that AKB-6548 is hydrophilic, which can influence its absorption and distribution within the body. The water solubility log value of 0.646 indicates good solubility, which is beneficial for bioavailability. The topological polar surface area (TPSA) of 46.504 is within the range that allows for good oral absorption, while the logS value of − 2.735 supports its solubility profile.

The compound does not inhibit the CYP450 enzymes, which is favorable as it reduces the potential for drug-drug interactions. The blood–brain barrier (BBB) permeability score of -1.031 suggests that AKB-6548 may not easily cross the BBB, which could be relevant depending on the intended therapeutic target. The human intestinal absorption (HIA) score of 0.375 indicates moderate absorption potential when administered orally.

The bioavailability score of 0.022 is low, suggesting that AKB-6548 may have limited bioavailability in its current form. The compound is not a substrate for P-glycoprotein, which is beneficial for drug resistance profiles. However, it has a significant interaction with the renal organic cation transporter (OCT), with a score of 1.252, indicating potential for renal excretion.

The predictions for AMES toxicity and hERG inhibition are negative, which is promising for its safety profile. The compound is also predicted to be non-sensitizing to the skin and non-irritating to the eyes, which is favorable for its overall safety.

In conclusion, the ADMET analysis suggests that the four compounds have comparable or better profiles than the positive control AKB-6548 in terms of solubility, metabolism, and safety. However, their bioavailability and permeability vary, with some compounds showing potential for further optimization (Table 3). Overall, these compounds exhibit promising ADMET characteristics that could be advantageous over AKB-6548, warranting further investigation and optimization for therapeutic applications.

Conclusions

In the current study, chemical space with diverse compounds was explored which revealed ZINC36378940, ZINC31164438, ZINC2005305, and ZINC67910437 as promising hits by targeting PHD2 active site Tyr303, Tyr310, Tyr329, Arg383, Thr236, ThR256 and Trp258 residues. Moreover, molecular simulation and binding free energy studies revealed that these drugs possess relatively good activity in comparison with AKB-6548 and need further in vitro and in vivo validation for the possible usage as potential drugs against HIF-PHD-associated diseases. Despite the robustness of our three-step virtual screening and reliable MD simulations, these methods may miss off-target effects and cannot fully replicate in vivo complexities. Binding free energy calculations, though predictive, depend on computational models and databases with inherent limitations. Ultimately, experimental validation is necessary to confirm the efficacy and safety of drug candidates for clinical use. Our findings also contribute to overcoming safety issues, liver failure, and problems associated with on‐/off‐targets associated with HIF-PHD.

Data availability

All the data is available on RCSB PDB ID: 7UMP (https://www.rcsb.org/structure/7ump), UniProt (Q9GZT9) are attached as Supplementary File S1 as well as deposited online on the Zenodo (Springer Nature recommended public repository) (https://zenodo.org/records/13890923) and GitHub (https://github.com/Abbas24-AI/BMC_6e84fa7d-caa3-49d6-b6b5-ce1860412fa9), publicly and any simulation data would be provided on demand. The accession numbers to access this data are given in the manuscript.

References

  1. Zhang Y, et al. Screening of prolyl hydroxylase 2 inhibitors based on quantitative strategy of peptides. J Chromatogr A. 2022;1679: 463411.

    Article  CAS  PubMed  Google Scholar 

  2. Wu Y, et al. Inhibition of the oxygen-sensing asparaginyl hydroxylase factor inhibiting hypoxia-inducible factor: a potential hypoxia response modulating strategy. J Med Chem. 2021;64(11):7189–209.

    Article  CAS  PubMed  Google Scholar 

  3. Wang J, et al. Glucose transporter GLUT1 expression and clinical outcome in solid tumors: a systematic review and meta-analysis. Oncotarget. 2017;8(10):16875.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Pugh CW, Ratcliffe PJ. New horizons in hypoxia signaling pathways. Exp Cell Res. 2017;356(2):116–21.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Apte RS, Chen DS, Ferrara N. VEGF in signaling and disease: beyond discovery and development. Cell. 2019;176(6):1248–64.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Vion D, et al. Manipulating the quantum state of an electrical circuit. Science. 2002;296(5569):886–9.

    Article  CAS  PubMed  Google Scholar 

  7. Hon W-C, et al. Structural basis for the recognition of hydroxyproline in HIF-1α by pVHL. Nature. 2002;417(6892):975–8.

    Article  CAS  PubMed  Google Scholar 

  8. Haase VH. Therapeutic targeting of the HIF oxygen-sensing pathway: lessons learned from clinical studies. Exp Cell Res. 2017;356(2):160–5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Berra E, et al. HIF prolyl-hydroxylase 2 is the key oxygen sensor setting low steady-state levels of HIF-1α in normoxia. EMBO J. 2003;22(16):4082–90.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Li Z, You Q, Zhang X. Small-molecule modulators of the hypoxia-inducible factor pathway: development and therapeutic applications. J Med Chem. 2019;62(12):5725–49.

    Article  CAS  PubMed  Google Scholar 

  11. Chen N, et al. Roxadustat treatment for anemia in patients undergoing long-term dialysis. N Engl J Med. 2019;381(11):1011–22.

    Article  CAS  PubMed  Google Scholar 

  12. Markham A. Vadadustat: first approval. Drugs. 2020;80:1365–71.

    Article  PubMed  Google Scholar 

  13. Islam MS, et al. Structural analysis of the 2-oxoglutarate binding site of the circadian rhythm linked oxygenase JMJD5. Sci Rep. 2022;12(1):20680.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Zuk A, et al. Preclinical characterization of vadadustat (AKB-6548), an oral small molecule hypoxia-inducible factor prolyl-4-hydroxylase inhibitor, for the potential treatment of renal anemia. J Pharmacol Exp Ther. 2022;383(1):11–24.

    Article  CAS  PubMed  Google Scholar 

  15. Schmidt F, et al. Predictive in silico off-target profiling in drug discovery. Future Med Chem. 2014;6(3):295–317.

    Article  CAS  PubMed  Google Scholar 

  16. Wu Y, et al. Small-molecule inhibitors of HIF-PHD2: a valid strategy to renal anemia treatment in clinical therapy. Medchemcomm. 2016;7(7):1271–84.

    Article  CAS  Google Scholar 

  17. Maxwell PH, Eckardt K-U. HIF prolyl hydroxylase inhibitors for the treatment of renal anaemia and beyond. Nat Rev Nephrol. 2016;12(3):157–68.

    Article  CAS  PubMed  Google Scholar 

  18. Sayaf AM, et al. Exploring the natural products chemical space through molecular search to discover potential inhibitors to target Hypoxia Inducible Factor (HIF) Prolyl Hydroxylase Domain (PHD). Front Pharmacol. 2023;14:1202128.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Alkhzem AH, Woodman TJ, Blagbrough IS. Design and synthesis of hybrid compounds as novel drugs and medicines. RSC Adv. 2022;12(30):19470–84.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Alamgir A, Alamgir A. Drugs: their natural, synthetic, and biosynthetic sources. Therapeutic use of medicinal plants and their extracts: Volume 1: Pharmacognosy. 2017: p. 105–123.

  21. Gopinath P, Kathiravan M. Docking studies and molecular dynamics simulation of triazole benzene sulfonamide derivatives with human carbonic anhydrase IX inhibition activity. RSC Adv. 2021;11(60):38079–93.

    Article  Google Scholar 

  22. Nangaku M, et al. A novel class of prolyl hydroxylase inhibitors induces angiogenesis and exerts organ protection against ischemia. Arterioscler Thromb Vasc Biol. 2007;27(12):2548–54.

    Article  CAS  PubMed  Google Scholar 

  23. Ivan M, et al. Biochemical purification and pharmacological inhibition of a mammalian prolyl hydroxylase acting on hypoxia-inducible factor. Proc Natl Acad Sci. 2002;99(21):13459–64.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Kim SY, Yang EG. Recent advances in developing inhibitors for hypoxia-inducible factor prolyl hydroxylases and their therapeutic implications. Molecules. 2015;20(11):20551–68.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Barrett TD, et al. Pharmacological characterization of 1-(5-chloro-6-(trifluoromethoxy)-1H-benzoimidazol-2-yl)-1H-pyrazole-4-carboxylic acid (JNJ-42041935), a potent and selective hypoxia-inducible factor prolyl hydroxylase inhibitor. Mol Pharmacol. 2011;79(6):910–20.

    Article  CAS  PubMed  Google Scholar 

  26. Rose PW, et al. The RCSB protein data bank: integrative view of protein, gene and 3D structural information. Nucleic Acids Res. 2016: p. gkw1000.

  27. Haddad Y, Adam V, Heger Z. Ten quick tips for homology modeling of high-resolution protein 3D structures. PLoS Comput Biol. 2020;16(4): e1007449.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Ntie-Kang F, et al. AfroDb: a select highly potent and diverse natural product library from African medicinal plants. PLoS ONE. 2013;8(10): e78085.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Sorokina M, Steinbeck C. Review on natural products databases: where to find data in 2020. J Cheminform. 2020;12(1):20.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Kang H, et al. HIM-herbal ingredients in-vivo metabolism database. J Cheminform. 2013;5(1):1–6.

    Article  CAS  Google Scholar 

  31. Teli MK, et al. In silico identification of hydantoin derivatives: a novel natural prolyl hydroxylase inhibitor. J Biomol Struct Dyn. 2021;39(2):703–17.

    Article  CAS  PubMed  Google Scholar 

  32. Ravindranath PA, et al. AutoDockFR: advances in protein-ligand docking with explicitly specified binding site flexibility. PLoS Comput Biol. 2015;11(12):e1004586.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Ravindranath PA, et al. AutoDockFR: advances in protein-ligand docking with explicitly specified binding site flexibility. PLoS Comput Biol. 2015;11(12): e1004586.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Yuan S, Chan HS, Hu Z. Using PyMOL as a platform for computational drug design. Wiley Interdiscip Rev Comput Mol Sci. 2017;7(2): e1298.

    Article  Google Scholar 

  35. Maestro S. Maestro. Schrödinger, LLC, New York, NY, 2020. 2020.

  36. Azzam KA. SwissADME and pkCSM webservers predictors: an integrated online platform for accurate and comprehensive predictions for in silico ADME/T properties of artemisinin and its derivatives. Kompleksnoe Ispolzovanie Mineralnogo Syra Complex use of mineral resources. 2023;325(2):14–21.

    Article  Google Scholar 

  37. Pires DE, Kaminskas LM, Ascher DB. Prediction and optimization of pharmacokinetic and toxicity properties of the ligand. Computational drug discovery and design. 2018: p. 271–284.

  38. Wu F, et al. Computational approaches in preclinical studies on drug discovery and development. Front Chem. 2020;8:726.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Case DA, et al. The Amber biomolecular simulation programs. J Comput Chem. 2005;26(16):1668–88.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Price DJ, Brooks CL III. A modified TIP3P water potential for simulation with Ewald summation. J Chem Phys. 2004;121(20):10096–103.

    Article  CAS  PubMed  Google Scholar 

  41. Roe DR, Cheatham TE III. PTRAJ and CPPTRAJ: software for processing and analysis of molecular dynamics trajectory data. J Chem Theory Comput. 2013;9(7):3084–95.

    Article  CAS  PubMed  Google Scholar 

  42. Li J, et al. The VSGB 2.0 model: a next generation energy model for high resolution protein structure modeling. Proteins Structure Function Bioinformatics. 2011;79(10):2794–812.

    Article  CAS  Google Scholar 

  43. Khan A, et al. Dynamics insights into the gain of flexibility by helix-12 in ESR1 as a mechanism of resistance to drugs in breast cancer cell lines. Front Mol Biosci. 2020;6:159–159.

    Article  PubMed  PubMed Central  Google Scholar 

  44. Khan A, et al. Higher infectivity of the SARS-CoV-2 new variants is associated with K417N/T, E484K, and N501Y mutants: an insight from structural data. J Cell Physiol. 2021; n/a(n/a).

  45. Khan A, et al. The Omicron (B.1.1. 529) variant of SARS-CoV-2 binds to the hACE2 receptor more strongly and escapes the antibody response: insights from structural and simulation data. Int J Biol Macromol. 2022;200:438.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Arshia AH, et al. De novo design of novel protease inhibitor candidates in the treatment of SARS-CoV-2 using deep learning, docking, and molecular dynamic simulations. Comput Biol Med. 2021;139: 104967.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Fu T, et al. VARIDT 2.0: structural variability of drug transporter. Nucleic Acids Res. 2022;50(D1):D1417-d1431.

    Article  CAS  PubMed  Google Scholar 

  48. Fu T, et al. Exploring the binding mechanism of metabotropic glutamate receptor 5 negative allosteric modulators in clinical trials by molecular dynamics simulations. ACS Chem Neurosci. 2018;9(6):1492–502.

    Article  CAS  PubMed  Google Scholar 

  49. Jomhori M, Mosaddeghi H, Farzin H. Tracking the interaction between single-wall carbon nanotube and SARS-Cov-2 spike glycoprotein: a molecular dynamics simulations study. Comput Biol Med. 2021;136: 104692.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  50. Xue W, et al. What contributes to serotonin-norepinephrine reuptake inhibitors’ dual-targeting mechanism? The key role of transmembrane domain 6 in human serotonin and norepinephrine transporters revealed by molecular dynamics simulation. ACS Chem Neurosci. 2018;9(5):1128–40.

    Article  CAS  PubMed  Google Scholar 

  51. Xue W, et al. Molecular mechanism for the allosteric inhibition of the human serotonin transporter by antidepressant escitalopram. ACS Chem Neurosci. 2022;13(3):340–51.

    Article  CAS  PubMed  Google Scholar 

  52. Kiriakidis S, et al. Complement C1q is hydroxylated by collagen prolyl 4 hydroxylase and is sensitive to off-target inhibition by prolyl hydroxylase domain inhibitors that stabilize hypoxia-inducible factor. Kidney Int. 2017;92(4):900–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. Kaelin WG Jr. The von Hippel-Lindau tumor suppressor gene and kidney cancer. Clin Cancer Res. 2004;10(18):6290S-6295S.

    Article  CAS  PubMed  Google Scholar 

  54. Semenza GL. Oxygen sensing, hypoxia-inducible factors, and disease pathophysiology. Annu Rev Pathol. 2014;9:47–71.

    Article  CAS  PubMed  Google Scholar 

  55. Zaric BL, et al. Homocysteine and hyperhomocysteinaemia. Curr Med Chem. 2019;26(16):2948–61.

    Article  CAS  PubMed  Google Scholar 

  56. Ariazi JL, et al. Discovery and preclinical characterization of GSK1278863 (daprodustat), a small molecule hypoxia inducible factor–prolyl hydroxylase inhibitor for anemia. J Pharmacol Exp Ther. 2017;363(3):336–47.

    Article  CAS  PubMed  Google Scholar 

  57. Sayaf AM, et al. Exploring the natural products chemical space through a molecular search to discover potential inhibitors that target the hypoxia-inducible factor (HIF) prolyl hydroxylase domain (PHD). Front Pharmacol. 2023;14:1202128.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  58. Chandrasekaran J, Balasubramaniam J. A computational evaluation of FDA medicines’ ability to inhibit hypoxia-inducible factor prolyl hydroxylase-2 (PHD-2) for acute respiratory distress syndrome. Struct Chem. 2022;33(5):1391–407.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. Yu Z, et al. Discovery of prolyl hydroxylase 2 inhibitors with new chemical scaffolds as in vivo active erythropoietin inducers through a combined virtual screening strategy. Chem Biol Drug Des. 2020;95(2):270–8.

    Article  CAS  PubMed  Google Scholar 

  60. Camagni GF, Minervini G, Tosatto SCE. Structural characterization of hypoxia inducible factor α-prolyl hydroxylase domain 2 interaction through MD simulations. Int J Mol Sci. 2023;24(5):4710.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

This work was also supported by the Ministry of Higher Education Malaysia (MOHE), Fundamental Research Grant Scheme (FRGS/1/2021/STG04/USM/02/14). This work was also supported by Qatar University Grants No. QUT2RP-CPH-24/25-477 and No. QUPD-CPH-23/24-592. The statements made herein are solely the responsibility of the authors. The computational resources were provided by College of Pharmacy, Qatar University to conduct the Molecular dynamics simulation of the complexes.

Funding

This work was supported by the Ministry of Higher Education Malaysia (MOHE), Fundamental Research Grant Scheme (FRGS/1/2021/STG04/USM/02/14). This work was also supported by Qatar University grants No. QUT2RP-CPH-24/25-477 and No. QUPD-CPH-23/24-592. The statements made herein are solely the responsibility of the authors.

Author information

Authors and Affiliations

  1. School of Chemical Sciences, Universiti Sains Malaysia, 11800, Penang, Malaysia

    Abrar Mohammad Sayaf & Kar Kheng Yeoh

  2. Department of Healthcare Biotechnology, Atta Ur Rahman School of Applied Biosciences, National University of Science and Technology Islamabad, Islamabad, Pakistan

    Kafila Kousar

  3. Centre for Biotechnology and Microbiology, University of Swat, Swat, KPK, Pakistan

    Muhammad Suleman & AbdulRahman Alshammari

  4. Department of Pharmacology and Toxicology, College of Pharmacy, King Saud University, Post Box 2455, 11451, Riyadh, Saudi Arabia

    Norah A. Albekairi

  5. Department of Biochemistry and Molecular Biology, Dasman Diabetes Institute, Dasman, Kuwait

    Anwar Mohammad

  6. Department of Pharmacology, College of Pharmacy, Qatar University, Doha, Qatar

    Abbas Khan & Abdelali Agouni

  7. Department of Biological Sciences, School of Medical and Life Sciences (SMLS), Sunway University, 47500, Bandar Sunway, Selangor Darul Ehsan, Malaysia

    Abbas Khan

Authors
  1. Abrar Mohammad Sayaf
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Kafila Kousar
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Muhammad Suleman
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Norah A. Albekairi
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. AbdulRahman Alshammari
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Anwar Mohammad
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Abbas Khan
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Abdelali Agouni
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Kar Kheng Yeoh
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Abrar Mohammad Sayaf, Muhammad Suleman, Kafila Kousar, Norah A. Albekairi, AbdulRahman Alshammari = conceptualization, data curation, formal analysis, visualization, validation and writing original draft. KarKheng Yeoh, Anwar Mohammad, Abdelai Agouni, Abbas Khan = conceptualization, project administration, resources, supervision, writing the original draft and final draft.

Corresponding authors

Correspondence to Abbas Khan, Abdelali Agouni or Kar Kheng Yeoh.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Sayaf, A.M., Kousar, K., Suleman, M. et al. Molecular exploration of natural and synthetic compounds databases for promising hypoxia inducible factor (HIF) Prolyl-4- hydroxylase domain (PHD) inhibitors using molecular simulation and free energy calculations. BMC Chemistry 18, 236 (2024). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13065-024-01347-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13065-024-01347-4

Keywords

BMC Chemistry

ISSN: 2661-801X

Contact us