2. Materials and methods
2.1. Structure Collection
The structures of compounds from Z. budrunga were collected from previously published research papers. The structures were drawn on Chemdraw and compared to that on the PubChem database for all those whose structures were available. However, for one compound (pseudophrynamine), the structure was not found in any of the research papers and thus was considered from PubChem database.
2.2. Check for drug like properties of the compounds
Lipinski’s scoring function was used to determine the druggability of the compounds. According to Lipinski’s rule of 5, a compound satisfies the drug like status if there are not more than 5 hydrogen-bond donors (sum of –OHs and –NHs), molecular weight should not exceed 500 units, maximum value of LogP should be 5 and there should not be more than 10 hydrogen-bond acceptors (sum of Os’ and Ns’). Based on the fact that almost 80-90% of the known drugs when examined fell within the limits, this rule can be considered for the identification of lead compounds . Apart from the Lipinski’s rule of 5, we also calculated other molecular properties such as the number of rotatable bonds, an important parameter measuring flexibility and descriptor of oral bioavailability of drugs. Polar surface area is reported as good descriptor for characterization of drug absorption, bioavailability and blood brain barrier penetration . The molecular properties of the compounds were calculated using Mol Soft and Mol Inspiration online tools.
2.3. ADME prediction
ADME (Absorption, Distribution, Metabolism, and Excretion) prediction is an important criterion for selection of lead compounds because of their failure in the clinical phases. So an early prediction would be highly beneficial in avoiding the future failures due to poor ADME characteristics. ADME prediction was done using the Pre-ADMET online tool. Prediction using in silico HIA (Human Intestinal Absorption) model is useful to identify potential candidates for oral delivery, also HIA is an important parameter as it tells whether the drug is absorbed by the intestine or not. Ideally, if the HIA falls between 70-100% then it is said that the compound is well absorbed by the intestine or else it is moderately (20-70%) or poorly absorbed (0-20%). BBB (blood brain barrier) penetration can give information of therapeutic drug in the central nervous system (CNS). It is another important parameter as it tells us whether the compounds will pass the blood brain barrier or not. If the value is more than 2.00, then it is said that the compounds pass the barrier and are CNS active. If it is between 0.10-2.00, it is moderately absorbed by the CNS and if it is less than 0.10 then it is considered as CNS-inactive. Plasma protein binding model tells about the disposition and efficacy of a drug. If it is 90% and above, then it is said to be strongly bound and therefore, it is not a good drug as it is not available for the transport or interaction with the target. If it is less than 90%, then it is weakly bound to the proteins and thus readily available for the transport and can interact with the target.
2.4. Prediction for mutagenecity
AMES test is an important test which is used to identify whether a compound is mutagenic or not. AMES test was performed in silico using TEST (Toxicity Estimation Software Tool). The software detects frame-shift mutations or base-pair substitutions by exposing histidine-dependent strains of Salmonella typhimurium to a test compound. TEST software also estimates toxicity using various Quantitative Structure Activity Relationship (QSAR) methodologies. These methods are (i) Hierarchical method- measures toxicity using weighted mean of predictions from various models which are generated prior to the run time. (ii) FDA method- measures toxicity using a new model which is almost similar to the existing compound. The new model is generated during the run time. (iii) Nearest neighbour method- measures toxicity by taking the mean of the three chemicals from its library which is most similar to the test compound. (iv) Consensus method- measures toxicity by taking mean of the predicted toxicities from the above methods.
2.5. Activity Prediction
The biological activity of a compound represents different pharmacotherapeutic effects, physiological and biochemical mechanisms of action and specific toxicity of the compound. Prediction of biological activity spectra was carried out by the use of PASS (Prediction of Activity Spectra for Substances) which predicts the activity based on the structure of the compound with the mean prediction accuracy of 95% . PASS predicts over 3500 different kinds of biological activity of the compound by comparing the structure of the compound with nearly 250,000 biological compounds which includes drugs, lead compounds, drug candidates and toxic substances compiled from different sources such as chemical databases, patents, publications and “gray” literatures. PASS prediction evaluates the biological activity spectra in terms of probabilities to be active (Pa) or to be inactive (Pi). As the biological activity is measured as the probability of whether the compound is active or inactive, therefore the values of Pa and Pi lies between 0.00-1.00. Interpretation of PASS prediction result wholly depends on two probabilities; only the activities whose Pa>Pi, are to be considered. (i) If Pa > 0.7, the probability of experimental activity is high. (ii) 0.5 < Pa < 0.7, the probability of experimental activity is less but also states that the compound is probably not same as the other known pharmaceutical compounds in the PASS library. (iii) Pa < 0.5, the probability to prove the experimental activity is less, but its activity can be increased by modifying the compound structurally. Unfortunately, PASS has some limitations; it is not able to predict the activity of a new compound which has not been reported in its library. Therefore, in this study we have included all the activities with Pa > 0.5, in order to avoid missing of any important biological activity.
3.1 Structure Collection
A total of 26 compounds were found to be extracted naturally from Z. budrunga (Table 2). This paper does not include the synthesized or slightly modified compounds from the plant.
3.2 Check for drug like properties of the compounds
The results of prediction of molecular properties for 26 compounds are given in Table 3. Except for 3,5-dimethoxy-4-geranyloxycinnamyl alcohol and Lup-20 (29)-en-3-one, all the compounds had their specific molecular properties values within the limits as specified by Lipinski’s rule of 5. The two important parameters that influence the permeability, bioavailability, absorption and distribution characteristics of any drug are LogP and LogS. The predicted LogP value for 3,5-dimethoxy-4-geranyloxycinnamyl alcohol (5.18) and Lup-20 (29)-en-3-one (8.14) exceeded the limits for LogP i.e 5. Lup-20 (29)-en-3-one has been predicted to be insoluble in water (Log S 0.00). The compounds which have less than 10 numbers of rotatable bonds and a polar surface area not greater than 140A2 are predicted to have good oral bioavailability. All the compounds have a polar surface area of not greater than 140A2. However, except for 3,5-dimethoxy-4-geranyloxycinnamyl alcohol (no of rotatable bonds – 10) all the compounds have less than 10 number of rotatable bonds. The drug likeness model score was highest and positive for arborine (1.21) followed by hydroxyevodiamine (0.90), (+)-evodiamine (0.86), 2-(2’,4’,6’-trimethyl-heptenyl)-4-quinozolone (0.85), 5-methoxy-7-hydroxy-flavonone (0.61), N-methylflndersine (0.53), dictamine (0.19) and rutaecarpine (0.10). Over all except for 3,5-dimethoxy-4-geranyloxycinnamyl alcohol and Lup-20 (29)-en-3-one, all the other compounds are good candidates for developing drugs as these two are found to be violating the Lipinski’s rule of 5 and therefore, they were not considered for further predictions.
3.3 ADME prediction
ADME prediction (Table 4) helps to identify the compounds with good ADME characteristics. The results showed that all the test compounds exhibited above 90% HIA. These compounds can thus be used effectively for oral delivery. Some compounds such as dictamine, N-methylflndersine and pseudophrynamine showed 100% HIA and other compounds such as Arborine (99.48), Canthine-6-one (99.00), Xanthyletin (98.96), 8-methoxy-N-methylflndersine (98.23), Zanthobungeanine (98.23), Dihydrochelerythrine (98.03) showed 98% and above HIA. BBB penetration prediction helps in identification of compounds which are CNS-active. The results showed that less than half of the compounds (41.6%) showed high BBB penetration property, thus these drugs which has high BBB penetration power can get delivered to CNS. Compounds such as arnottianamide, 6-acetyldihydrochelerythrine, N-norchelerythrine, dihydrochelerythrine, (-)-syringaresinol, (-)-seasamine, 5-methoxy-7-hydroxy-flavonone, Lunacridine, xanthoxylin, 8-methoxy-N-methylflndersine, N-methylflndersine, Xanthyletin, Zanthobungeanine and 11β,13-dihydro-1-epireynosin possessed very low capacity to cross the BBB ie. less than 2. Strong binding of the compounds to the plasma protein reduces the disposition and efficacy of the compounds. A strongly bound compound cannot diffuse or traverse through the cell membrane efficiently. PreADMET prediction showed that N-methylflndersine (61.9%) was least bound to the plama proteins followed by lunacridine (67.51%), Zanthobungeanine (70.00%), 8-methoxy-N-methylflndersine (70.01%), pseudophrynamine (71.35%) and (-)-Syringaresinol (74.53%), Xanthoxylin (75.05), 11β,13-dihydro-1-epireynosin (76.53), all other compounds have shown plasma protein binding above 80%. Overall, the ADME prediction showed that only pseudophrynamine could be considered as good candidates for developing drugs. However, Xanthoxylin, N-methylflndersine, 8-methoxy-N-methylflndersine and Zanthobungeanine can be considered a candidate for CNS-inactive drug development as it shows HIA (93.00%), plama protein binding (75.05%) but BBB penetration (0.62).
3.4 Prediction for mutagenecity
The results of mutagenicity prediction by TEST are given in Table 5. Results showed that lunacridine, 5-methoxy 7-hydroxyflavonone, N-methylflndersine, pseudophrynamine, (-)-Seasamine, (-)-Syringaresinol, 6-Acetonyldihydrochelerythrin and xanthoxylin are not mutagenic while the rest were mutagenic.
3.5 Activity Prediction
Results for PASS prediction of Arborine (Figure 1) showed that arborine is highly predicted to act as nicotinic alpha2beta2 receptor antagonist (Pa 0.756), Tetrahydroxynaphthalene reductase inhibitor (Pa 0.661) and in treatment of male reproduction dysfunction (Pa 0.553) and as antiviral agent (Picorna virus) (Pa 0.539). Cantine-6-one (Figures 2 (a), (b) and (c)) is predicted to act as inhibitors for nearly 148 enzymes and also showed biological activities like nicotinic alpha2beta2 receptor antagonist (Pa 0.867), nicotinic alpha6beta3beta4alpha5 receptor antagonist (Pa 0.688), nootrpoic (0.679) and kidney function stimulant (Pa 0.592). The highest probability to act as an inhibitor is shown for glutathione thiolesterase (Pa 0.87), ferredoxin-NAD+ reductase (0.849), glycosylphosphatidylinositol phospholipase D (0.837), creatinase (0.778) and kinase (0.719). PASS prediction results showed that dictamine (Figures 3 (a) and (b)) is predicted to mainly act as inhibitor for many different enzymes as well as possess properties like antiviral activity against picorna virus (Pa 0.506), MAP kinase stimulant (0.507). Dictamine showed the highest probability as taurine dehydrogenase inhibitor (Pa 0.873), membrane permeability inhibitor (Pa 0.708) and JAK2 expression inhibitor (Pa 0.684). (+)-evodiamine (Figure 4) is predicted to show high biological activities like antihypoxic (Pa 0.84), 5-hydroxytryptamine release stimulant (Pa 0.837) and Mcl-1 antagonist (Pa 0.791). It was also predicted to possess anti neurotic therapeutic effect (Pa 0.547). Lunacridine (Figure 5) is predicted to show less biological activities. Among the bioactivities includes ability to act as inhibitor for exzymes such as ubiquinol-cytochrome-c reductase (Pa 0.682), chymosin inhibitor (Pa 0.604) and also act as a platlet aggregation stimulant (Pa 0.568). PASS prediction predicted biological activities for hydroxyevodiamine are shown in Figure 6. Biological activities such as 5-hydroxytryptamine release stimulant (Pa 0.927), nicotinic alpha2beta2 receptor antagonist (Pa 0.819) are predicted with high Pa value. Hyrdoxyevodiamine is also predicted to enhance TP53 expression (Pa 0.529) and show therapeutic effects like antihypoxic (Pa 0.908), antineurotic (Pa 0.704), anticonvulsant (Pa 0.634) and ability to be used in acute neurologic disorders treatment (Pa 0.591) and phobic disorder treatment (Pa 0.564). Some of the activities predicted for gamma fagarine (Figure 7) by PASS are membrane permeability inhibitor (Pa 0.698), JAK2 ecpression inhibitor (Pa 0.654), MAPK stimulant (Pa 0.55) and phobic disorders treatment (Pa 0.607). 5-methoxy-7-hydroxy-flavonone (Figures 8 (a) and (b)) is predicted to show very important activities like membrane permeability inhibitor (Pa 0.81), apoptosis agonist (Pa 0.749), mucous membrane protector (Pa 0.793), caspase 3 stimulant (Pa 0.68), hepatoprotectant (0.682), anti-inflammatory (Pa 0.627) and AR expression inhibitor (0.64). The predicted activity that has high Pa values includes mucous integrity agonist (Pa 0.957), HIF1A expression inhibitor (Pa 0.863) and Chlordecone reductase inhibitor (0.887). It is also predicted to enhance TP53 expression (Pa 0.801). PASS prediction also showed that both 8-methoxy-N-methylflndersine (Figure 9) and N-methylflndersine (Figure 10) to exhibit same biological activities such as HIF1A inhibitor (Pa 0.835; 0.859), 4-Nitrophenol 2-monooxygenase inhibitor (Pa 0.619; 0.669), CYP2H substrate (Pa 0.609; 0.572), CYP2A1 substrate (Pa 0.531; 0.551) and membrane permeability inhibitor (Pa 0.591; 0.658). 8-methoxy-N-methylflndersine is also predicted to show antiangial (Pa 0.612), antihypertensive (Pa 0.524) and antineoplastic activity (Pa 0.518). N-methyl flndersine is also predicted to act as an anticonvulsant (Pa 0.611) and antiangial (Pa 0.6). This pair is a good example to show the difference made by an extra methyl group on the activities of the same compound. Both pseudophrynamine (Figure 11) and rutaecarpine (Figure 12) are predicted to act as 5-hyroxytryptamine release stimulant (Pa 0.762, 0.73). Pseudophrynamine is also predicted to be a show good analgesic activity (Pa 0.767), anticonvulsant (Pa 0.542), kidney function stimulant (Pa 0.561) and neurotransmitter antagonist (Pa 0.517). Rutaecarpine is also predicted to possess antihypoxic effect (Pa 0.896), antineurotic (Pa 0.659) and nicotinic alpha2beta2 receptor antagonist (Pa 0.627). Figure 13 shows the PASS prediction predicted biological activity spectra for (-)-Seasamine. The results showed that (-)-seasamine is predicated to show high probability to be active as membrane integrity agonist (Pa 0.931), neurotransmitter uptake inhibitor (Pa 0.837) and caspase 3 stimulant (Pa 0.8). Other predicted biological activities include antidyskinetic (Pa 0.742), antineurotic (Pa 0.738), antineoplastic (Pa 0.739), TP53 expression enhancer (Pa 0.689), caspase 8 stimulant (Pa 0.605) and MAP kinase stimulant (Pa 0.6). Skimiammine (Figure 14) is predicted to act as JAK2 expression inhibitor (Pa 0.713), membrane permeability inhibitor (Pa 0.682), antineoplastic (Pa 0.65) and calcium channel activator (Pa 0.607). Xanthyletin (Figure 15) can act as HIF1A expression inhibitor (Pa 0.945), substrate for different enzymes, antineoplastic (Pa 0.781), membrane permeability inhibitor (Pa 0.737), hepatic disorders treatment (Pa 0.678), TP53 expression enhancer (Pa 0.68) and apoptosis agonist (Pa 0.649). Similar to most compounds (-)-syringaresinol (Figures 16 (a) and (b)) is predicted to show biological activities like antineoplastic (Pa 0.829), JAK2 expression inhibitor (Pa 0.81), TP53 expression enhancer (0.78), caspase 3 stimulant (Pa 0.712), antioxidant (Pa 0.672), cytoprotectant (Pa 0.659), HIF1A expression inhibitor (Pa 0.647), MAP kinase stimulant (Pa 0.543), free radical scavenger (Pa 0.521) and acetycholine neuromuscular blocking agent (Pa 0.553). 6-acetyldihydrochelerythrine (Figure 17) and arnottianamide (Figure 18) are predicted to show similar activities like neurotransmitter uptake inhibitor (Pa 0.63, 0.719), MAPK stimulant (Pa 0.551, 0.644), caspase-3 stimulant (Pa 0.536, 0.573) and ovulation inhibitor (Pa 0.569, 0.543). Individually, 6-acetyldihydrochelerythrine is predicted to exhibit antiprotozoal (Pa 0.519) and antineurotic (Pa 0.579) activities whereas arnottianamide can function as membrane integrity agonist (Pa 0.674) and MMP9 expression inhibitor (Pa 0.511). PASS predicts both dihydrochelerythrine (Figure 19) and N-norchelerythrine (Figure 20) to possess activities like antineoplastic (Pa 0.572, 0.759), caspase-3 stimulant (Pa 0.645, 0.777), caspase-8 stimulant (Pa 0.613, 0.625), ovulation inhibitor (Pa 0.643, 0.591), membrane integrity agonist (Pa 0.731, 0.696), JAK 2 expression inhibitor (Pa 0.533, 0.64), MMP-9 expression inhibitor (Pa 0.548, 0.583) and antineurotic (Pa 0.846, 0.76). Figure 21 shows the PASS predicted biological activity for 2-(2’,4’,6’-trimethyl-heptenyl)-4-quinozolone which predicts that 2-(2’,4’,6’-trimethyl-heptenyl)-4-quinozolone is expected to possess antineoplastic activity (Pa 0.549) and as a treatment for phobic disorders (Pa 0.655). PASS predicted zanthobungeanine (Figure 22) to exhibit activities like antineoplastic (Pa 0.518), HIF1A expression inhibitor (Pa 0.835), membrane permeability inhibitor (Pa 0.591) and antianginal (Pa 0.612). 11β, 13-dihydro-1-epireynosin (Figures 23 (a) and (b)) showed a wide range of activities such as antineoplastic (Pa 0.932), antieczematic (Pa 0.871), analeptic (Pa 0.782), immunosuppressant (Pa 0.702), antimetastatic (Pa 0.631), antifungal (Pa 0.605), anti-inflammatory (Pa 0.611), TP53 enhancer (Pa 0.61), apoptosis agonist (Pa 0.587) and hepatoprotectant (Pa 0.535). Similarly, a wide range of biological activities were predicted for xanthoxylin (Figures 24 (a), (b) and (c)). High Pa values were shown for activities like membrane integrity agonist (Pa 0.926), mucomembranous protector (Pa 0.83) and membrane permeability inhibitor (Pa 0.767). Other activities that were predicted include biological activities like apoptosis agonist (Pa 0.657), MMP9 expression inhibitor (Pa 0.636), antiseptic (Pa 0.629), cytoprotectant (Pa 0.619), HIF1A expression inhibitor (Pa 0.618) and anti-inflammatory (Pa 0.605).
In conclusion, the computer aided analysis of druggability and activity of lead compounds in Z. budrunga suggests that except for pseudophrynamine, other compounds could not satisfy all the criteria. However, if the compounds are to be considered as a drug for other disease except as CNS drugs then the BBB penetration cut off value of 2 can be neglected and N-methylflndersine and xanthoxylin can be considered too. Our main aim was to bring in note the druggable leads from Z. budrunga, so that drugs can be developed against a specific disorder in order to minimize the side effects and toxicity level. Certain compounds need structural modification to become a better drug; it is possible by combining cheminformatics and bioinformatics. Computers has eased the process of development of drug by allowing primary screening without the waste of cost as would have involved if were to carried out as a wet experiment. The current study basically is to help in screening and motivating a diverse array of compounds from plants or chemically synthesized in laboratory for their druggability. Today, in silico methods are playing an important part in the area of drug discovery. It is just that now one needs to understand the importance of traditional and modern concepts. Only after combining both the concepts one can successfully discover a drug.
...(download the rest of the essay above)