Parkinson’s disease is a common neurodegenerative disorder marked by the accumulation of the protein alpha synuclein. Studies have indicated the role of prolyl oligopeptidase (POP), a serine protease, in alpha synuclein accumulation. Therefore, POP emerges as an attractive medicinal target. Traditionally, most of the early medicines have been plant-based owing to their ready availability and negligible side effects. Alkaloids owing to their neurotransmitter modulatory, anti-amyloid, anti-oxidant, and anti-inflammatory activities have shown potential in neurodegenerative disease. In this work, we computationally evaluated alkaloid class of phytochemicals for their therapeutic efficacy against POP. Alkaloids were retrieved from the publically available database, Chemical Entities of Biological Interest (ChEBI), and screened for their drug likeness (Lipinski’s rule of 5) and absorption, distribution, metabolism, and excretion, and toxicity (ADMET) in Discovery Studio by ensuring parameters suitable for a central nervous system disease such as blood-brain barrier (BBB) level set to ≤2, absorption level set to 0 and solubility level permitted set to 2, 3, or 4. Next, molecular docking was performed to learn about the affinity of the filtered alkaloids with the POP. Subsequently, molecular dynamic simulations were conducted to assess the reliability and stability of the alkaloid-protein complex. Our study identified metergoline, pipercallosine, celacinnine, lobeline, cystodytin G, lycoperine A, hookerianamide J, and martefragin A as putative lead compounds against POP. Among these, metergoline, pipercallosine, hookerianamide J, and lobeline showed the most promising results. These compounds demonstrated better or equivalent molecular docking scores in comparison to three POP inhibitors that had reached clinical trials, i.e., Z-321, S-17092, and JTP-4819. MD simulations indicated that these compounds remained intact at the active site while adhering to the binding mode and interaction patterns as that of the reported inhibitors. The research conducted here, therefore, provides evidence for conducting in vitro POP inhibitory studies of these newly identified plant-based POP inhibitors.
Parkinson’s disease (PD) is a chronic and progressive neurodegenerative central nervous system (CNS) disorder. Affected individuals experience difficulty in walking, tremors, stiffness in limbs, or impaired balance (
POP (also known as PREP or prolyl endopeptidase; EC 3.4.21.26) is a large 80 KDa intracellular enzyme belonging to serine protease family. It is capable of cleaving peptides shorter than 30 amino acids after a proline residue [
POP in entirety assumes a cylindrical shape and is comprised of two domains: peptidase or catalytic domain (residues 1–72 and 428–710) containing the alpha/beta hydrolase fold, and a 7-bladed
Although several research groups have focused on POP inhibition [
In brief, structures of the alkaloids and target protein (POP) were retrieved. Alkaloids were screened for their drug likeness and ADMET properties, following which molecular docking was performed in order to access their binding to POP. Lastly, molecular dynamics simulations were conducted to evaluate the stability of the prospective drug candidates when bound to POP.
Structure of POP bearing PDB id: 3DDU [
The list of alkaloid compounds was downloaded from ChEBI (Chemical Entities of Biological Interest) (
To predict the binding affinity of the filtered alkaloids with POP, molecular docking was performed using Genetic Optimisation for Ligand Docking (GOLD) version 5.2.2 [
To determine the reliability and consistency of the binding of the prospective drug candidates to the target protein, molecular dynamics (MD) simulations were performed using GROningen MAchine for Chemical Simulations (GROMACS) v5.0.6 package [
ChEBI search resulted in the retrieval of 565 alkaloid compounds (accessed on 14th August 2019). To ensure a good pharmacokinetic profile of the potential drug candidates, ADMET tests were carried out in DS. By setting 0 as cutoff value for ADMET_ABSORPTION_LEVEL, it was ensured that potential inhibitors had a good intestinal absorption. Filtering of compounds where ADMET_SOLUBILITY_LEVEL was either 2, 3, or 4 ensured that the potential inhibitors were neither too soluble nor insoluble in water. The most important factor in identifying drug for neurological diseases is their ability to penetrate blood brain barrier. Usage of 0, 1, or 2 as cutoff for ADMET_BBB_LEVEL ensured only the compounds that could pass blood brain barrier would be retained. 190 of the initial 565 compounds passed the ADMET test. Drug likeness of these 190 compounds was accessed using “Filter by Lipinski and Veber Rules” with default parameters. Lipinski’s rule states that a drug-like molecule should not have more than 5 hydrogen bond donors and 10 hydrogen bond acceptors, molecular weight should not exceed 500 daltons, and LogP should not be more than 5 [
In order to determine whether the candidate inhibitors bind to POP, molecular docking was performed by employing Z-321and S-17092 as reference 1 (Ref1) and reference 2 (Ref2), respectively.
RMSD measures the average distance between the atoms of superimposed proteins and is used to measure the quality of reproduction of a known binding pose. Therefore, the lower the RMSD, the lesser the deviation from the known binding pose. The docked pose of co-crystal, i.e., GSK552 with POP resulted in an RMSD of 0.094 nm. The low RMSD ensured the credibility of our docking protocol. Using the affirmed docking parameters of GSK552, 189 candidate inhibitors were docked at the active site of POP. From the largest cluster obeying the binding mode of the co-crystal, best pose was selected based on higher score than that of reference compounds and interaction with key residues. GoldScore fitness for Ref1 and Ref2 was 70.5265 and 68.4808, respectively. Metergoline with a GoldScore fitness of 72.6547 was the only compound to score better than both Ref1 and Ref2. Pipercallosine with a GoldScore fitness of 68.6628 attained a slightly higher docking score than that of Ref2 (68.4808). Additionally, since JTP-4819 was also considered in clinical trials as POP inhibitor, we also docked it against POP. The GoldScore fitness value for JTP-4819 against POP was 60.0431. Apart from metergoline and pipercallosine, six other alkaloids, celacinnine, lobeline, cystodytin G, lycoperine A, hookerianamide J, and martefragin A, fared better than JTP-4819. These collectively will be referred to as hit candidates from now on. The docking scores of all these compounds are tabulated in Supplementary Table
After gaining knowledge on binding affinity via docking scores, we explored the interaction pattern of the hit candidates and reference compounds with POP. Because of its high docking score, we were mainly interested in learning about metergoline’s interaction with POP active site residues. “Show 2d diagram” of DS was utilized to inspect the molecular interactions with POP. Ref1, Ref2, and metergoline all interacted with the active site residues. Ref1 formed hydrogen bond with Trp595 and Arg643 and Ref2 formed hydrogen bond with Trp595, Tyr599, and Arg643. Metergoline like Ref1 and Ref2 was involved in hydrogen bond formation with Arg643 and Trp595, thus maintaining the necessary hydrogen bond interactions. These hydrogen bond interactions are depicted in Figures
Molecular docking based intermolecular interactions. Upper panel (a–c) demonstrates hydrogen bond interactions of POP residues (blue) with (a) Ref1, (b) Ref2, and (c) metergoline. Lower panel (d–f) is the 2D representation of all the molecular interactions between POP and (d) Ref1, (e) Ref2, and (f) metergoline. Green dashed lines in upper and lower panel represent hydrogen bond. All the other dashed lines represent various types of
We also performed interaction analysis of JTP-4819 and other hit compounds (pipercallosine, celacinnine, lobeline, cystodytin G, lycoperine A, hookerianamide J, and martefragin A) with POP. JTP-4819 demonstrated four hydrogen bonds: one each with Arg128 and Cys255 and two with Arg643.
We performed a 50 ns MD simulation using GROMACS in order to access the binding stability of POP when bound to Ref1, Ref2, and metergoline. RMSD for backbone atoms and potential energy calculations were executed to determine the stability of protein-ligand complex. As illustrated in Figure
Stability analysis from MD insights. (a) RMSD profiles for POP with Ref1, Ref2, and metergoline; (b) potential energy profiles for POP with Ref1, Ref2, and metergoline.
As the RMSD and potential energy profiles were reflecting stability of protein-ligand complexes, reference structures for all the three systems from last 5 ns were extracted and superimposed to check if the binding mode was retained. This analysis, as depicted in Figure
Binding mode analysis of POP with Ref1, Ref2, and metergoline. (a) Superimposed image of representative structures; (b) enlarged view. Protein is shown in grey wire model and compounds are depicted in stick models.
Further, hydrogen bond interactions sustained for Ref1 and Ref2 with Trp595 and Arg643. In case of metergoline, hydrogen bond with Trp595 was persistent and a new
Post-MD intermolecular interactions. Upper panel (a–c) demonstrates hydrogen bond interactions of POP residues (blue) with (a) Ref1, (b) Ref2, and (c) metergoline. Lower panel (d–f) is the 2D representation of all the molecular interactions between POP and (d) Ref1, (e) Ref2, and (f) metergoline. Green dashed lines in upper and lower panel represent hydrogen bonds. All the other dashed lines represent various types of
For JTP-4819 and other hit compounds, we performed 20 ns MD simulation with POP. The average RMSD for pipercallosine, lobeline, martefragin A, lycoperine A, cystodytin G, and celacinnine was observed to be 0.11 nm and for JTP-4819 and hookerianamide J, the average RMSD value was 0.12 nm. The RMSD profiles of all these systems were below the accepted value of 0.15 nm thus indicating stability when bound to POP. Potential energy profiles of these compounds with POP also strengthened and supported the stability observation. The RMSD and potential energy profiles are depicted in Supplementary Figure
Due to POP’s ability to alter several aspects and function of CNS such as learning, memory, mood, and hypertension [
Metergoline is an ergoline alkaloid obtained from taxa such as fungi and some higher plants [
Upon performing molecular docking of POP with prospective compounds, the GoldScore fitness of Z-321, S-17092, and JTP-4819 was 70.5265, 68.4808, and 60.0431, respectively. Metergoline with a GoldScore fitness of 72.6547 surpassed all the above mentioned inhibitors, while pipercallosine showed a slightly higher GoldScore fitness value of 68.6628 as compared to S-17092. Celacinnine, lobeline, cystodytin G, lycoperine A, hookerianamide J, and martefragin A showed higher GoldScore fitness as compared to JTP-4819 (Supplementary Table
Phe476, Asn555, Val580, Trp595, Tyr599, and Val644 form the S1 Subsite of active site, Arg643 forms the S2 subsite and S3 subsite is formed by Phe173, Met235, Cys255, Ile591, and Ala594. The best docked poses of all our hits molecules demonstrated interactions with all these residues, except only for cystodytin G which apart from Ala594 interacted with all the residues. Interaction monitoring after MD analysis of these hit compounds with POP revealed that metergoline, pipercallosine, hookerianamide J, and lobeline demonstrated crucial hydrogen bond interaction with either Trp595 or Arg643 or with both. Pipercallosine and hookerianamide J both showed both Trp595 and Arg643 hydrogen bond interactions. Metergoline and lobeline demonstrated hydrogen bond interaction with Trp595 and Arg643, respectively. All these compounds were retained at the active site as was observed from the aligned and superimposed MD poses (Figure
Left image displays superimposed metergoline-POP complex at 0, 25, and 49 ns. Right panel displays the enlarged view of Arg643 and metergoline while highlighting the distance between HH11 of Arg643 with O25 of metergoline at (a) 0 ns, (b) 25 ns, and (c) 49 ns.
GSK552 docking with POP resulted in a GoldScore fitness value of 73.88. Additionally, interaction analysis of docked pose of GSK552 also demonstrated hydrogen bonds with Trp595 and Arg643. We also performed a 20 ns MD for GSK552 in complex with POP. The average RMSD was observed to be 0.09 nm (Supplementary Figure
It has been reported that, for a drug to efficiently cross BBB, its molecular weight should be in the range of 314–420 Da, it should be highly lipophilic, i.e., logP should be in the range of 0.66–6, PSA less than 60–70 Å2, H-bond donors should be <2, H-bond acceptors should be less than <6, and total number of nitrogen + oxygen atoms should be less than 6 [
Literature survey indicated that while Ref1, Ref2 [
Comparison of BBB permeability properties.
Properties | Ref1 | Ref2 | JTP-4819 | Metergoline | Pipercallosine | Hookerianamide J | Lobeline |
---|---|---|---|---|---|---|---|
Molecular weight (Da) | 344.5 | 384.5 | 359.185 | 403.517 | 329.433 | 442.677 | 337.455 |
logP | 2.378 | 3.203 | 0.527 | 4.249 | 4.758 | 4.935 | 3.933 |
Polar surface area (Å2) | 65.92 | 65.92 | 89.94 | 46.5 | 47.56 | 52.57 | 40.54 |
Hydrogen bond donor | 0 | 0 | 2 | 1 | 1 | 2 | 1 |
Hydrogen bond acceptor | 3 | 3 | 4 | 3 | 3 | 3 | 3 |
N + O atoms | 4 | 4 | 7 | 5 | 4 | 4 | 3 |
Although this research was started keeping Parkinson’s disease in mind, literature survey during the process highlighted the role of POP inhibitors not just in PD but in various other disorders such as schizophrenia [
Safety of potential drug candidates is of utmost importance in development of new drugs. Due to their abundance and relative low toxicity, research in identifying plant-based drugs has seen an uptrend. Additionally, crossing of BBB is a major roadblock but an absolute necessary requirement for development of CNS drugs. Using in silico approaches, this work unravels the likely ability of alkaloids in inhibiting prolyl oligopeptidase, which has been indicated as therapeutic target in many diseases. By comparing POP inhibitors that had reached clinical trials, our work displayed better affinity of metergoline, pipercallosine, hookerianamide J, and lobeline for POP. Finally, MD simulations confirmed their stability in complex with POP. Therefore, the research conducted here provides promising results for further investigation of metergoline, pipercallosine, hookerianamide J, and lobeline as effective POP inhibitors.
The datasets generated or analyzed in the current study are available from the corresponding author upon request.
The authors declare no conflicts of interest.
This research was supported by the Bio & Medical Technology Development Program of the National Research Foundation (NRF) and funded by the Korean Government (MSIT) (no. NRF-2018M3A9A7057263). The authors would like to thank Dr. Rabia Mukhtar Rana and Ms. Shraddha Parate for the insightful discussions during the course of this research.
Supplementary Table 1: docking scores of hit compounds and reference inhibitors. Supplementary Figure 1: molecular docking-based intermolecular interactions of JTP-4819, GSK552, and other hit compounds. Supplementary Figure 2: RMSD and potential energy profiles of JTP-4819, GSK552, and other hit compounds. Supplementary Figure 3: binding mode analysis of POP with JTP-4819, GSK552, and other hit compounds. Supplementary Figure 4: post-MD intermolecular interactions of JTP-4819, GSK552, and other hit compounds. Supplementary Table 2: BBB permeability properties of celacinnine, cystodytin G, lycoperine A, martefragin A, berberine, and GSK552. Supplementary Figure 5: chemical structures of the proposed inhibitors.