Application of computer-aided drug design for identification of P. falciparum inhibitors
- Authors: Diallo, Bakary N’tji
- Date: 2021-10-29
- Subjects: Plasmodium falciparum , Malaria -- Chemotherapy , Molecular dynamics , Antimalarials , Cheminformatics , Drug development , Ligand binding (Biochemistry) , Plasmodium falciparum1-deoxy-D-xylulose-5-phosphate reductoisomerase (PfDXR) , South African Natural Compounds Database
- Language: English
- Type: Doctoral theses , text
- Identifier: http://hdl.handle.net/10962/192798 , vital:45265 , 10.21504/10962/192798
- Description: Malaria is a millennia-old disease with the first recorded cases dating back to 2700 BC found in Chinese medical records, and later in other civilizations. It has claimed human lives to such an extent that there are a notable associated socio-economic consequences. Currently, according to the World Health Organization (WHO), Africa holds the highest disease burden with 94% of deaths and 82% of cases with P. falciparum having ~100% prevalence. Chemotherapy, such as artemisinin combination therapy, has been and continues to be the work horse in the fight against the disease, together with seasonal malaria chemoprevention and the use of insecticides. Natural products such as quinine and artemisinin are particularly important in terms of their antimalarial activity. The emphasis in current chemotherapy research is the need for time and cost-effective workflows focussed on new mechanisms of action (MoAs) covering the target candidate profiles (TCPs). Despite a decline in cases over the past decades with, countries increasingly becoming certified malaria free, a stalling trend has been observed in the past five years resulting in missing the 2020 Global Technical Strategy (GTS) milestones. With no effective vaccine, a reduction in funding, slower drug approval than resistance emergence from resistant and invasive vectors, and threats in diagnosis with the pfhrp2/3 gene deletion, malaria remains a major health concern. Motivated by these reasons, the primary aim of this work was a contribution to the antimalarial pipeline through in silico approaches focusing on P. falciparum. We first intended an exploration of malarial targets through a proteome scale screening on 36 targets using multiple metrics to account for the multi-objective nature of drug discovery. The continuous growth of structural data offers the ideal scenario for mining new MoAs covering antimalarials TCPs. This was combined with a repurposing strategy using a set of orally available FDA approved drugs. Further, use was made of time- and cost-effective strategies combining QVina-W efficiency metrics that integrate molecular properties, GRIM rescoring for molecular interactions and a hydrogen mass repartitioning (HMR) molecular dynamics (MD) scheme for accelerated development of antimalarials in the context of resistance. This pipeline further integrates a complex ranking for better drug-target selectivity, and normalization strategies to overcome docking scoring function bias. The different metrics, ranking, normalization strategies and their combinations were first assessed using their mean ranking error (MRE). A version combining all metrics was used to select 36 unique protein-ligand complexes, assessed in MD, with the final retention of 25. From the 16 in vitro tested hits of the 25, fingolimod, abiraterone, prazosin, and terazosin showed antiplasmodial activity with IC50 2.21, 3.37, 16.67 and 34.72 μM respectively and of these, only fingolimod was found to be not safe with respect to human cell viability. These compounds were predicted active on different molecular targets, abiraterone was predicted to interact with a putative liver-stage essential target, hence promising as a transmission-blocking agent. The pipeline had a promising 25% hit rate considering the proteome-scale and use of cost-effective approaches. Secondly, we focused on Plasmodium falciparum 1-deoxy-D-xylulose-5-phosphate reductoisomerase (PfDXR) using a more extensive screening pipeline to overcome some of the current in silico screening limitations. Starting from the ZINC lead-like library of ~3M, hierarchical ligand-based virtual screening (LBVS) and structure-based virtual screening (SBVS) approaches with molecular docking and re-scoring using eleven scoring functions (SFs) were used. Later ranking with an exponential consensus strategy was included. Selected hits were further assessed through Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA), advanced MD sampling in a ligand pulling simulations and (Weighted Histogram Analysis Method) WHAM analysis for umbrella sampling (US) to derive binding free energies. Four leads had better predicted affinities in US than LC5, a 280 nM potent PfDXR inhibitor with ZINC000050633276 showing a promising binding of -20.43 kcal/mol. As shown with fosmidomycin, DXR inhibition offers fast acting compounds fulfilling antimalarials TCP1. Yet, fosmidomycin has a high polarity causing its short half-life and hampering its clinical use. These leads scaffolds are different from fosmidomycin and hence may offer better pharmacokinetic and pharmacodynamic properties and may also be promising for lead optimization. A combined analysis of residues’ contributions to the free energy of binding in MM-PBSA and to steered molecular dynamics (SMD) Fmax indicated GLU233, CYS268, SER270, TRP296, and HIS341 as exploitable for compound optimization. Finally, we updated the SANCDB library with new NPs and their commercially available analogs as a solution to NP availability. The library is extended to 1005 compounds from its initial 600 compounds and the database is integrated to Mcule and Molport APIs for analogs automatic update. The new set may contribute to virtual screening and to antimalarials as the most effective ones have NP origin. , Thesis (PhD) -- Faculty of Science, Biochemistry and Microbiology, 2021
- Full Text:
- Date Issued: 2021-10-29
- Authors: Diallo, Bakary N’tji
- Date: 2021-10-29
- Subjects: Plasmodium falciparum , Malaria -- Chemotherapy , Molecular dynamics , Antimalarials , Cheminformatics , Drug development , Ligand binding (Biochemistry) , Plasmodium falciparum1-deoxy-D-xylulose-5-phosphate reductoisomerase (PfDXR) , South African Natural Compounds Database
- Language: English
- Type: Doctoral theses , text
- Identifier: http://hdl.handle.net/10962/192798 , vital:45265 , 10.21504/10962/192798
- Description: Malaria is a millennia-old disease with the first recorded cases dating back to 2700 BC found in Chinese medical records, and later in other civilizations. It has claimed human lives to such an extent that there are a notable associated socio-economic consequences. Currently, according to the World Health Organization (WHO), Africa holds the highest disease burden with 94% of deaths and 82% of cases with P. falciparum having ~100% prevalence. Chemotherapy, such as artemisinin combination therapy, has been and continues to be the work horse in the fight against the disease, together with seasonal malaria chemoprevention and the use of insecticides. Natural products such as quinine and artemisinin are particularly important in terms of their antimalarial activity. The emphasis in current chemotherapy research is the need for time and cost-effective workflows focussed on new mechanisms of action (MoAs) covering the target candidate profiles (TCPs). Despite a decline in cases over the past decades with, countries increasingly becoming certified malaria free, a stalling trend has been observed in the past five years resulting in missing the 2020 Global Technical Strategy (GTS) milestones. With no effective vaccine, a reduction in funding, slower drug approval than resistance emergence from resistant and invasive vectors, and threats in diagnosis with the pfhrp2/3 gene deletion, malaria remains a major health concern. Motivated by these reasons, the primary aim of this work was a contribution to the antimalarial pipeline through in silico approaches focusing on P. falciparum. We first intended an exploration of malarial targets through a proteome scale screening on 36 targets using multiple metrics to account for the multi-objective nature of drug discovery. The continuous growth of structural data offers the ideal scenario for mining new MoAs covering antimalarials TCPs. This was combined with a repurposing strategy using a set of orally available FDA approved drugs. Further, use was made of time- and cost-effective strategies combining QVina-W efficiency metrics that integrate molecular properties, GRIM rescoring for molecular interactions and a hydrogen mass repartitioning (HMR) molecular dynamics (MD) scheme for accelerated development of antimalarials in the context of resistance. This pipeline further integrates a complex ranking for better drug-target selectivity, and normalization strategies to overcome docking scoring function bias. The different metrics, ranking, normalization strategies and their combinations were first assessed using their mean ranking error (MRE). A version combining all metrics was used to select 36 unique protein-ligand complexes, assessed in MD, with the final retention of 25. From the 16 in vitro tested hits of the 25, fingolimod, abiraterone, prazosin, and terazosin showed antiplasmodial activity with IC50 2.21, 3.37, 16.67 and 34.72 μM respectively and of these, only fingolimod was found to be not safe with respect to human cell viability. These compounds were predicted active on different molecular targets, abiraterone was predicted to interact with a putative liver-stage essential target, hence promising as a transmission-blocking agent. The pipeline had a promising 25% hit rate considering the proteome-scale and use of cost-effective approaches. Secondly, we focused on Plasmodium falciparum 1-deoxy-D-xylulose-5-phosphate reductoisomerase (PfDXR) using a more extensive screening pipeline to overcome some of the current in silico screening limitations. Starting from the ZINC lead-like library of ~3M, hierarchical ligand-based virtual screening (LBVS) and structure-based virtual screening (SBVS) approaches with molecular docking and re-scoring using eleven scoring functions (SFs) were used. Later ranking with an exponential consensus strategy was included. Selected hits were further assessed through Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA), advanced MD sampling in a ligand pulling simulations and (Weighted Histogram Analysis Method) WHAM analysis for umbrella sampling (US) to derive binding free energies. Four leads had better predicted affinities in US than LC5, a 280 nM potent PfDXR inhibitor with ZINC000050633276 showing a promising binding of -20.43 kcal/mol. As shown with fosmidomycin, DXR inhibition offers fast acting compounds fulfilling antimalarials TCP1. Yet, fosmidomycin has a high polarity causing its short half-life and hampering its clinical use. These leads scaffolds are different from fosmidomycin and hence may offer better pharmacokinetic and pharmacodynamic properties and may also be promising for lead optimization. A combined analysis of residues’ contributions to the free energy of binding in MM-PBSA and to steered molecular dynamics (SMD) Fmax indicated GLU233, CYS268, SER270, TRP296, and HIS341 as exploitable for compound optimization. Finally, we updated the SANCDB library with new NPs and their commercially available analogs as a solution to NP availability. The library is extended to 1005 compounds from its initial 600 compounds and the database is integrated to Mcule and Molport APIs for analogs automatic update. The new set may contribute to virtual screening and to antimalarials as the most effective ones have NP origin. , Thesis (PhD) -- Faculty of Science, Biochemistry and Microbiology, 2021
- Full Text:
- Date Issued: 2021-10-29
In silico identification of natural inhibitory compounds against the Mycobacterium tuberculosis Enzyme Pyrazinamidase using high-throughput virtual screening techniques
- Authors: Kenyon, Thomas
- Date: 2021-10-29
- Subjects: Mycobacterium tuberculosis , Pyrazinamide , Molecular dynamics , High throughput screening (Drug development) , Mutagenesis , South African Natural Compounds database (SANCDB)
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10962/192074 , vital:45193
- Description: Tuberculosis (TB) is most commonly a pulmonary infection caused by the bacterium Mycobacterium tuberculosis. With the exception of the COVID-19 pandemic, TB was the most common cause of death due to an infectious disease for a number of years up until 2020. In 2019, 10 million people fell ill with TB worldwide and 1.4 million people died (WHO, 2020a). Additionally, multidrug-resistant TB (MDR-TB) remains a public health crisis and a health security threat. A global total of 206 030 people with multidrug- or rifampicin-resistant TB (MDR/RR-TB) were reported in 2019, a 10% increase from 186 883 in 2018. South Africa is ranked among the 48 high TB burden countries, with an estimated 360 000 people falling ill in 2019, resulting in 58 000 deaths, the majority of which being among people living with HIV. Unlike HIV, however, TB is a curable disease when managed correctly with long durations of antitubercular chemotherapy. Pyrazinamide (PZA) is an important first-line tuberculosis drug unique for its activity against latent TB. PZA is a prodrug, being converted into its active form, pyrazinoic acid (POA) by the Mtb gene pncA, coding for the pyrazinamidase enzyme (PZase). TB resistance to first-line drugs such as PZA is commonly associated with mutations in the pncA/PZase enzyme. This study aimed to identify potential novel inhibitors that bind to the active site of PZase. By making use of molecular docking studies and molecular dynamics (MD) simulations, high throughput virtual screening was performed on 623 compounds from the South African Natural Compounds database (SANCDB; https://sancdb.rubi.ru.ac.za). Ligands that selectively bound to the PZase active site were identified using docking studies, followed by MD simulations to assess ligand-PZase complex stability, Finally, hit compounds identified from the first round of MD simulations were screened again against PZase structures with high confidence point mutations known to infer PZA resistance in order to identify any novel compounds which had inhibitory potential against both WT and mutant forms of the PZase enzyme. , Thesis (MSc) -- Faculty of Science, Biochemistry and Microbiology, 2021
- Full Text:
- Date Issued: 2021-10-29
- Authors: Kenyon, Thomas
- Date: 2021-10-29
- Subjects: Mycobacterium tuberculosis , Pyrazinamide , Molecular dynamics , High throughput screening (Drug development) , Mutagenesis , South African Natural Compounds database (SANCDB)
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10962/192074 , vital:45193
- Description: Tuberculosis (TB) is most commonly a pulmonary infection caused by the bacterium Mycobacterium tuberculosis. With the exception of the COVID-19 pandemic, TB was the most common cause of death due to an infectious disease for a number of years up until 2020. In 2019, 10 million people fell ill with TB worldwide and 1.4 million people died (WHO, 2020a). Additionally, multidrug-resistant TB (MDR-TB) remains a public health crisis and a health security threat. A global total of 206 030 people with multidrug- or rifampicin-resistant TB (MDR/RR-TB) were reported in 2019, a 10% increase from 186 883 in 2018. South Africa is ranked among the 48 high TB burden countries, with an estimated 360 000 people falling ill in 2019, resulting in 58 000 deaths, the majority of which being among people living with HIV. Unlike HIV, however, TB is a curable disease when managed correctly with long durations of antitubercular chemotherapy. Pyrazinamide (PZA) is an important first-line tuberculosis drug unique for its activity against latent TB. PZA is a prodrug, being converted into its active form, pyrazinoic acid (POA) by the Mtb gene pncA, coding for the pyrazinamidase enzyme (PZase). TB resistance to first-line drugs such as PZA is commonly associated with mutations in the pncA/PZase enzyme. This study aimed to identify potential novel inhibitors that bind to the active site of PZase. By making use of molecular docking studies and molecular dynamics (MD) simulations, high throughput virtual screening was performed on 623 compounds from the South African Natural Compounds database (SANCDB; https://sancdb.rubi.ru.ac.za). Ligands that selectively bound to the PZase active site were identified using docking studies, followed by MD simulations to assess ligand-PZase complex stability, Finally, hit compounds identified from the first round of MD simulations were screened again against PZase structures with high confidence point mutations known to infer PZA resistance in order to identify any novel compounds which had inhibitory potential against both WT and mutant forms of the PZase enzyme. , Thesis (MSc) -- Faculty of Science, Biochemistry and Microbiology, 2021
- Full Text:
- Date Issued: 2021-10-29
- «
- ‹
- 1
- ›
- »