The mutant LRRK2 G2019S has two active sites. The active site − 1 is where the unique ligand GDP (Guanosine diphosphate) is located while Active site − 2 is the ATP bond location. (Myasnikov et al. 2021). In this study, levodopa (Hornykiewicz 2010) was adopted as the reference drug on which selection of the lead compounds were based on. However, our interest is to select compounds as potent as levodopa but without side effects as the reference drug. The reference drug gave a binding score of -7 for both actives sites but − 6.2 and − 6.4 were got as binding score for active site 1 (GDP location) and active site 2 (ATP location) in a validation docking using Mcule webserver. The docking score threshold was set based on these values. Even though the docking score values of 4-Hydroxybenzhydrazide do not correspond with the threshold values set for docking score of active sites 1 and 2, the score obtained for active site 1 is close to the threshold score while the validated score is same as the threshold cut off. Table 1 shows the docking scores of the first 10 phytochemical compounds of safflower against the active sites of the LRRK2 ligand. Out of these top 10 compounds, only the first nine molecules have a docking score, less or equal to the threshold docking score. The phytocompound with the lowest score, Serotobenine was reported as a novel phenolic amide from safflower seeds in 1985 which was isolated together with the already known compounds; N-ferloyltyptamine and N-(P-coumaroyl)tryptamine (Sato et al. 1985). Also, by construction of an optically active dihydrobenzofuran through rhodium mediated intramolecular C-H insertion, (-) – serotobenine has been synthetically generated (Koizumi et al. 2008). Other methods have been employed in an attempt to carry out a total synthesis of serotobenine (Hu et al. 2013). Though no significant activity has been recorded for this compound, it’s derivatives alongside Decursivine which is structurally related to it have been shown to have anti- plasmodial activity (Qin et al. 2011). Quercetin has been reported to have various biological activity which includes but not limited to anti-inflammatory, antibacterial, anticancer and antioxidant activity, however it has mutagenic effects which corresponds to our toxicity predictions in this study (Batiha et al. 2020). The only structural difference between Kaempferol and Quercetin is the presence of an additional hydroxyl group in its benzene ring (Liu et al., 2008). Like quercetin it has numerous reported biological activities and their similarity in activity can be attributed to their structural resemblance, thus a proof of therapeutic dependence on molecular structure (Ren et al. 2019). Nb-P-Coumaroyltryptamine isolated from Phragmites australis (cav.) Trin. Ex steud has shown to have cytotoxic activity and even though it has been reported to inhibit the production of proinflammatory cytokins, it exhibited an activity lower than N(P-Coumaroyl)serotonin (Chen et al. 2021). More so, the antioxidant activity of N-Coumaroyl Serotonin has been reported alongside N-Feruloylserotonin of which both were isolated from safflower seeds (Lee et al. 2008). Scutellarin is another constituent of safflower that has been described as possessing multiple pharmacological activities including but not limited to myocardial protection, anticoagulation, antiplatelet, vascular relaxation, anti-inflammation and anti-oxidant. It has also been described to be effective towards diabetic complications, stroke and myocardial infarction treatment. Reports of its recent medicinal modification, formulation with improved safety, efficacy and bioavailability supports the appreciable admet and pharmacokinetics properties reported by our study on this compound and may thus be repurposed alongside its bioactive derivatives towards PD treatment (Wang and Ma 2018). Anticancer and anti-inflammatory activity have been reported for acacetin (Sun et al. 2018) while anti-inflammatory and alpha – amylase activity has also been reported for Transchalcone (Staurengo-Ferrari et al. 2018). Since oxidative stress has been reported to cause cognitive impairment in PD, the selected molecules that have previously shown to possess oxidative potency may likely be efficient in PD treatment therapy.
Table 2 presents some physicochemical properties of the lead compounds. These properties determine the pharmacodynamic process and pharmacokinetics of drugs and thus are given significant consideration in design, discovery and development of drugs (Jia et al. 2020). Molecular weight determines the drug-likeness of any molecule under consideration especially for oral drugs. While Lipinski’s rule assumes a molecular weight of ≤ 500 as an optimal weight of good oral drugs, lead-likeness prediction assumes a molecular weight in the range 250 ≤ MW ≤ 350 (Protti et al. 2021). The number of hydrogen acceptors and donors present in a compound has been described to affect its drug-likeness. Hydrogen bonding is characterized by the donor and acceptor group which creates the needed hydrophobic interactions that is essential for chemotherapeutic effects. It has been established decades ago that the action of local anesthetics is characterized by a formation of a hydrogen – bonded complex between the acceptor group and the drug (Sax and Pletcher 1969). Also, inhibitors and non-inhibitors of organic anion transporters (OAT) have number of hydrogen -bond donors and acceptors, topological surface area, molecular weight and number of rotatable bonds attributed to their inhibitory or non-inhibitory effect. OAT family is reported to mediate between clinically important drugs and the body disposition (Duan et al. 2012). Partition coefficient is as important as other physicochemical properties, thus an indispensable property in drug design and development. It measures the lipophilicity of a molecule and describes the extent to which a drug would remain either in the lipid membrane or in an aqueous medium. Swiss adme uses five (5) predictive models, however the consensus estimation is chosen for accuracy (Mannhold et al. 2009).
Among other hydrophobic interactions, non-covalent interactions, including the occurrences of hydrogen, Vanderwaals, π–π, π-sigma bond, conventional hydrogen bond, π – π T – shaped, π – alkyl and π – stacked interactions were observed. However, for serotonin and N-Feruloylserotonin, Vanderwaals, conventional hydrogen bond, π – π T – shaped, π – alkyl and π – stacked interactions were observed after visualization with discovery studio (Fig. 1). The conventional hydrogen bond, having a relative strongest interaction is responsible for the high stability observed in (N-Coumaroyl-Serotonin) – 7li3 complex during molecular dynamics study.
Examining the pharmacokinetics profile of a potential drug is one of the major routes to drug discovery and development. Carrying out all required processes in the wet laboratory is not only time consuming but also yield an extreme cost. Therefore, pharmacokinetics prediction is a route to salvaging time, resources and cost in finding potential lead molecule as it filters possible lead compounds at a minimal cost (Bandyopadhyay et al. 2021). Table 3 shows the drug likeliness and pharmacokinetics of the hit compounds. Swiss ADME uses three models in solubility prediction. However, we have presented the results from ALI model only, since it demonstrates a stronger linear relation between predicted and experimental values (R2 = 0.81). Most predicted molecules are soluble in this model, though N-Coumaroyl Serotonin and N-Feruloylserotonin are moderately soluble. Even though solubility is a key property in design of drugs, insoluble potent drug candidates can be converted to the salt of the parent compound to enhance its solubility. Some drugs such as diclofenac are marketed in this form (Al Ragib et al. 2018). Gastrointestinal absorption and brain accessibility prediction is one of the preliminary tests to discover the fate of a drug candidate within the biological system of an organism. The model employs the computation of polarity of small molecules and their lipophilicity. This physicochemical property determines the extent of a drug’s bioavailability. The bioavailability of a drug entails a systemic distribution of its molecules which also determines its therapeutic efficiency, metabolism and excretion (Maderuelo et al. 2019). Poor permeability of potential drugs has been revealed to be the cause of the prevailing limited therapeutic treatment for Parkinson’s diseases, thereby necessitating our study to also consider potential drug candidates with positive test result for blood brain barrier (BBB) permeability (Shaltiel-Karyo et al. 2013). Though, only Nb-PCoumaroyltryptamine and Trans-Chalcone show positive result for this test, other lead compounds can still be subjected to in vivo testing for further verification. It has been reported in previous research that water-soluble molecules can be transformed into lipid soluble molecules so as to enhance blood brain barrier permeability owing to the observation that lipid solubility plays significant roles in passive diffusion into BBB. This transformation introduces lipid groups at the polar ends of the drug candidate (Bellettato and Scarpa 2018). An alternative method reported to solve the challenge posed by this, is the use of intranasal drug delivery method which entails a direct transport into the cerebrospinal fluid; a process that delivers molecules of interest direct to the central nervous system in minutes time interval with a bypass of the BBB (Mathison et al. 1998). Other reported alternative methods include use of transport/carrier systems, inhibition of efflux transports that impede drug delivery, trojan horse approach, chimeric peptides, monoclonal antibody fusion proteins, pro-drug bioconversion strategy, nanoparticles-based technologies, gene therapy and intracerebral gene therapy (Bellettato and Scarpa 2018). All the selected lead molecules in this study show high GI (Gastrointestinal) absorption prediction. The P-gp (permeability glycoprotein) otherwise known as multi-drug resistance protein is a constituent of the cell membrane that exudes foreign substances out of the cell. The ability of a drug to act as either P-gp-substate or non-P-gp substrate determines both the total clearance time and bioavailability. Pumping of drugs back to lumen by this protein reduces absorption. In summary, drugs that inhibit P-glycoprotein would increase absorption while drugs that induce it leads to reduction in absorption.
CYPP450 (Cytochrome P450) enzymes are very essential in metabolism of many drugs. Although there exist about 50 cytochrome enzymes that are involved in drug metabolism, ninety percent of drugs are metabolized by only six of these enzymes. In this research, we present the possible inhibitors of CYP1A2, CYP2C19, CYP2C9, CYP2D6, CYP3A4 of which CYP3A4 and CYP2D6 have been described as the most important. Drug candidates that inhibit all of these enzymes may not be subjected to metabolism and thus would increase the chance of the drug’s toxicity (Das et al. 2009). In our study, each of the selected drug candidate is a non-inhibitor to at least one of these enzymes, making them possible of being metabolized and excreted.
The toxicity predictions of the hit compounds (Table 4) show that, Quercetin and 4-Hydroxybenzhydrazide are not suitable as drugs as they are predicted to be ames mutagenic. A potential drug candidate with no tolerable toxicity profile would not be included in the next stage of a drug discovery process, although there should a consideration of concentration/dosage dependence of toxicity. The Oral rat acute toxicity (LD50) estimates the amount of a drug candidate that would lead to 50% death of the test animal group. The prediction reveals the relative toxicity profile of the selected compounds. Ames mutagenicity involves the use of bacteria to investigate mutagenic potential of a given molecule in the DNA (deoxyribonucleic acid) while the toxicity threshold dose in humans is determined by maximum tolerated dose in human. CarcinoPred EL was used to predict the carcinogenicity of the selected compounds. Quercetin in previous research has been demonstrated to be toxic and precisely has exhibited carcinogenicity in the kidney of male rats, leading to benign tumors of the renal tubular epithelium, however, further research has demonstrated that synergistic therapy can attenuate its possible toxicity by reducing the dosage in take (Zou et al. 2021), a proof why the phytocompound does not pose much risk to the consumption of fruits that contains it, since there is a synergistic action with other phytochemicals. Also, this compound has been evaluated previously for its ability to attenuate cognitive impairment with 6-hydroxydopamine induced PD in rats and the result proved that quercetin enhances spatial memory (Sriraksa et al. 2012) and has been considered as PD supplemental therapy (Tamtaji et al. 2020).
LRRK2 being a 286-kDa multidomain protein makes it more computationally expensive to carryout out a simulation using the entire protein. In order to minimize time and computational cost, Kinase (KIN), ROC, and COR domains which contain the mutation sites (Myasnikov et al. 2021) were used for a complete model simulation, though RMSD (Root Mean Square Deviation) was obtained using the full protein structure (7LI3) (Fig. 2). The RMSD values for the COR domain of 7LI3 complex and apo were within 2.5 to 5.6 Amstrong. The RMSD of the COR domain of 7LI3 is stable for the apoprotein after about 15ns, for C4B, stability is attained after about 80ns. C1B showed a stark increase in its RMSD after about 80ns and remains stable throughout the rest of the duration. This is suggested to be as a result of change in conformation (Bandyopadhyay et al. 2021); however, it maintains an RMSD similar to the apoprotein indicating that the C1B complex results in a similar conformation as the native conformation (Swargiary et al. 2022). The RMSD values for the ROC domain for 7LI3 complex and apo were within 1 to 5 Amstrong. It is stable for the apoprotein, C1B and C4B after about 5ns. C1B showed a stark increase in its RMSD after about 124 and 175ns. This is perhaps due to a conformational change; however, it maintained an RMSD similar to the apoprotein indicating that the C1B complex results in a similar conformation as the native conformation. C1B remains stable at around 3 Armstrong throughout the rest of the duration, maintaining the lowest deviation from its native conformation (Lower than the apo state). The RMSD values for the KIN domain for 7LI3 complex and apo were within 1 to 4 Amstrong. It is stable for the apoprotein, C1B and C4B after about 2ns. C1B remains stable at around 2.5 Armstrong throughout the rest of the duration, maintaining the lowest deviation from its native conformation (Lower than the apo state).
For the 200ns simulation, RMSF (Root mean square fluctuation) (Fig. 3) was carried out to detect the local changes along 7LI3 amino acid residues (Ishola et al. 2021). It was observed that the alpha helices and bet strands of the apo structure and docked systems oscillated within 5 to 30 Amstrong with the most fluctuation observed in C4B. The loop regions of all the simulations showed large fluctuations up to 40, 29 and 35 Armstrong for C4B C1B and the apo state respectively for the COR domain. For the ROC domain, loop regions of all the simulations showed large fluctuations up to 45, 27 and 40 Armstrong for C4B, C1B, and apo state respectively while large fluctuations were observed up to 45, 27 and 47 Armstrong for C4B, C1B, and apo state respectively in the KIN domain.
In MD simulation, the gyration radius is an index to monitor the structural formation process. A relative steady value of Rg (Radius of Gyration) (Fig. 4) indicates that the complex is stable (Morgan et al. 2017).
The results of our study reveal that C1B and C4B complex with 7LI3 were found to be stable in the binding sites with insignificant structural movements and lesser conformational changes to the overall structure hence, presenting them as potential inhibitors against LRRK2 (Bandyopadhyay et al. 2021).