GOLD: The All In One Molecular Docking Package
October 5, 2023
This blog presents the main features of GOLD, a validated, configurable protein-ligand docking software for expert drug discovery, alongside four case studies that use this versatile tool from the CSD-Discovery suite.
What Is GOLD?
GOLD (Genetic Optimisation for Ligand Docking) is a genetic algorithm for docking flexible ligands into protein binding sites and is relied on by researchers in academia and industry worldwide.
Using GOLD, it is possible to perform pose prediction to validate ligand docking results. It has proven success in virtual screening, lead optimisation, and in identifying the correct binding mode of active molecules.
A key aspect of GOLD is that it is very versatile. GOLD has four different scoring functions, and hence there is a greater chance of finding a protocol that suits a particular system. It can deal with key water molecules, rotating, translating, and turning them on and off during the docking run. GOLD presents several features for dealing with protein flexibility (rotatable side-chains, soft-potentials and ensemble docking). Finally, it contains several different constraints, including H-bond, hydrophobic region, scaffold, and similarity constraints.
All the functionalities are available through the Python API, where it is possible to run dockings programmatically for parameter optimisation and workflow incorporation.
Case Study 1: Novel Noncovalent Nuclear Export Inhibitors
In this work reported in the Journal of Medicinal Chemistry, scientists virtually screened a library of marketed drugs and identified an inhibitor of a key protein–protein interaction (PPI) in the fight against cancer, using GOLD and CHEMPLP-based ranking.
The group form Sichuan University aimed to find a non-covalent inhibitor of chromosome region maintenance 1 (CRM1), an attractive cancer drug target that is overexpressed in cancers and can promote their development. Although almost all the CRM1 inhibitors reported in the literature are covalent inhibitors, non-covalent inhibitors may offer better safety and efficacy profiles, along with increased resilience to resistance.
The structures of the approved drugs were obtained from the DrugBank database and docked into the CRM1 active site. The docking studies were performed using the software GOLD, and the docked poses were scored using CHEMPLP and ranked by CHEMPLP scores.
Zafirlukast (ZAF) was identified as a new noncovalent CRM1 inhibitor hit, and after a preliminary virtual screening analysis, biochemical experiments further verified this observation. The crystal structure of ZAF in the complex with CRM1 (Figure 1) confirmed that the ZAF-CRM1 interaction is purely hydrophobic, with no hydrogen bond forming. Finally, ZAF displayed activity against a broad range of cancer types in vitro, and synergy was observed when combined with doxorubicin.
To find out more about this work, follow the link to the article or read the blog “GOLD in Action: Foiling a Protein–Protein Interaction to Help Fight Cancer – Drug Repurposing using GOLD”.
Case Study 2: Ovarian Cancer Drug by Virtual Screening
This work reported the use of GOLD in identifying a promising next generation ribonucleotide reductase (RNR) inhibiting candidate for the treatment of ovarian cancer.
Poly ADP-ribose polymerase (PARP) inhibitors have been clinically approved as a treatment for epithelial ovarian cancer (EOC), demonstrating clinical efficacy by targeting BRCA gene mutations or defects in homologous recombination (HR) repair in breast and ovarian cancers. However, resistance to platinum and PARP inhibitors caused by BRCA mutation reversal and restoration of HR repair function was identified in clinical and preclinical studies.
This limitation encouraged scientists to develop inhibitors targeting HR repair. The group previously showed that triapine, a small molecule inhibitor of RNR, impaired HR repair and sensitized HR repair-proficient EOC to PARP inhibitors, restoring their effectiveness towards the treatment of EOC. In this work, scientists performed the docking of triapine into the triapine-binding pocket on RNR to identify a next generation RNR-inhibiting candidate for the treatment of ovarian cancer.
In silico screening of over 200,000 compounds was performed using GOLD, and hit clustering was carried out on the most promising 200 top-ranking compounds. This, along with considerations for the commercial availability of the hits, narrowed the candidate field down to 25 molecules. Additional physicochemical analysis allowed them to identify DB4 (Figure 2) as the most active compound, capable of causing a nearly complete DNA inhibition (> 99%).
To find out more about this work, follow the link to the article or read the blog “CSD in Action: Next Generation Ovarian Cancer Drug Identified by GOLD Virtual Screening and Confirmed Experimentally”.
Case Study 3: Structure-based Drug Design to Develop Novel Herbicides
In this example, researchers used GOLD to rationalize structural modifications of a new class of p-hydroxyphenylpyruvate dioxygenase (HPPD) inhibitors with potential use as herbicides.
The first step of the workflow (Figure 3) was the design of a family of benzyl-containing triketone-aminopyridines. GOLD was then used to identify the binding modes of the compounds to the HPPD, leading to the generation of a total of 300 conformations for each compound. The best conformations were selected based on the complex of the HPPD with the commercially available inhibitor mesotrione, and were further evaluated with additional computational methods.
The most promising 25 compounds were then synthesised, and their performance was compared to that of the commercially available HPPD inhibitor.
The best active compound exhibited a 5.8-fold enhancement in inhibiting Arabidopsis thaliana HPPD activity over that of mesotrione. Three key compounds displayed excellent greenhouse herbicidal effects. Another compound showed superior weed-controlling efficacy to multiple test dosages against the malignant Setaria viridis, when compared to mesotrione.
To find out more about this work, follow the link to the article or read the blog “GOLD in Action: Leveraging Structure-based Drug Design to Develop Novel Herbicides”.
Case Study 4: Application in Organometallic Compound Chemistry
Lomzik et al. used the GOLD docking tool for investigating the interaction of organometallic complexes of ispinesib, a known kinesin inhibitor.
The authors prepared the organometallic conjugates from the R and S-enantiomers of ispinesib. Ruthenium (Ru), Osmium (Os), Rhodium (Rh), Iridium (Ir) were used for the preparation of these complexes.
GOLD was used for comparing the docking of metal conjugates at the ispinesib-binding site of human kinesin KSP to that of ispinesib (R)-1 and its (S)-enantiomer (S)-1 in terms of the docking scores and the configuration relative to the ligands.
Results showed that when docking (R)-1 and (S)-1 into the same binding site, the quinazoline components overlapped. However, while for (R)-1 the p-toluyl group pointed towards the solvent exposed area of the binding site, in (S)-1 this position was held by the amino group. As a consequence, the binding site was more packed and hence the docking scores were lower.
When docking the complexes, the N,N-bidentate ligands instead occupied an inner hydrophobic region in the binding pocket, with the transition metal moiety pointing at a peripheral hydrophilic region, leading to the formation of π-π interactions with the binding site residues Trp127 and Tyr211.
To find out more about this work, follow the link to the article.
Organometallic Complexes (6)