Limits...
Enrichment of chemical libraries docked to protein conformational ensembles and application to aldehyde dehydrogenase 2.

Wang B, Buchman CD, Li L, Hurley TD, Meroueh SO - J Chem Inf Model (2014)

Bottom Line: Among individual MD snapshots, many exhibited enrichment that was significantly better than the crystal structure, but no correlation between enrichment and structural deviation from crystal structure was found.We found remarkable increase in enrichment power, particularly for p38, where the ROC-AUC increased by 0.30 to 0.85.We found that the use of randomly selected compounds docked to the target of interest using SVMSP led to notable enrichment for EGFR and Src MD snapshots.

View Article: PubMed Central - PubMed

Affiliation: Department of Biochemistry and Molecular Biology, ‡Melvin and Bren Simon Cancer Center, §Center for Computational Biology and Bioinformatics, and ⊥Stark Neurosciences Institute, Indiana University School of Medicine , 535 Barnhill Drive, Indianapolis, Indiana 46202, United States.

ABSTRACT
Molecular recognition is a complex process that involves a large ensemble of structures of the receptor and ligand. Yet, most structure-based virtual screening is carried out on a single structure typically from X-ray crystallography. Explicit-solvent molecular dynamics (MD) simulations offer an opportunity to sample multiple conformational states of a protein. Here we evaluate our recently developed scoring method SVMSP in its ability to enrich chemical libraries docked to MD structures of seven proteins from the Directory of Useful Decoys (DUD). SVMSP is a target-specific rescoring method that combines machine learning with statistical potentials. We find that enrichment power as measured by the area under the ROC curve (ROC-AUC) is not affected by increasing the number of MD structures. Among individual MD snapshots, many exhibited enrichment that was significantly better than the crystal structure, but no correlation between enrichment and structural deviation from crystal structure was found. We followed an innovative approach by training SVMSP scoring models using MD structures (SVMSPMD). The resulting models were applied to two difficult cases (p38 and CDK2) for which enrichment was not better than random. We found remarkable increase in enrichment power, particularly for p38, where the ROC-AUC increased by 0.30 to 0.85. Finally, we explored approaches for a priori identification of MD snapshots with high enrichment power from an MD simulation in the absence of active compounds. We found that the use of randomly selected compounds docked to the target of interest using SVMSP led to notable enrichment for EGFR and Src MD snapshots. SVMSP rescoring of protein-compound MD structures was applied for the search of small-molecule inhibitors of the mitochondrial enzyme aldehyde dehydrogenase 2 (ALDH2). Rank-ordering of a commercial library of 50 000 compounds docked to MD structures of ALDH2 led to five small-molecule inhibitors. Four compounds had IC50s below 5 μM. These compounds serve as leads for the design and synthesis of more potent and selective ALDH2 inhibitors.

Show MeSH

Related in: MedlinePlus

SVMSP rescoring of MD snapshots identifiesALDH2 inhibitors. Thedehydrogenase activity was screened by measuring the rate of increasein the fluorescence of NADH upon propionaldehyde oxidation. (A) Percentactivity of ALDH2 in the presence of 50 μM of each of the 111compounds that were tested; (B) concentration-dependent curves forfive compounds that inhibited in the initial screen; and (C) chemicalstructures for the five compounds (ALDH400, ALDH417, ALDH423, ALDH427,and ALDH440) along with their IC50s for inhibition of ALDH2 dehydrogenaseactivity.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4114474&req=5

fig6: SVMSP rescoring of MD snapshots identifiesALDH2 inhibitors. Thedehydrogenase activity was screened by measuring the rate of increasein the fluorescence of NADH upon propionaldehyde oxidation. (A) Percentactivity of ALDH2 in the presence of 50 μM of each of the 111compounds that were tested; (B) concentration-dependent curves forfive compounds that inhibited in the initial screen; and (C) chemicalstructures for the five compounds (ALDH400, ALDH417, ALDH423, ALDH427,and ALDH440) along with their IC50s for inhibition of ALDH2 dehydrogenaseactivity.

Mentions: We applied SVMSP rescoringof MD structures to the aldehyde dehydrogenase 2 (ALDH2) enzyme usingSVMSP as the scoring approach. The crystal structure of ALDH2 in itsapo form (PDB code: 1O04) was used to carry out explicit-solvent unbiased MD simulations.70−73 Five independent simulations with 7 ns in length (35 ns total) yielded25 000 snapshots. These were clustered by RMSD using ptraj(74) as described above. Aset of 50 representative snapshots were selected from the clusters.A focused set of the ChemDiv commercial library64 containing 50 000 compounds were docked to eachof the 50 snapshots by AutoDock Vina.61 Docked receptor–ligand complexes were rescored with SVMSP.For each of the 50 000 compounds, the 50 MD snapshots to whichthey were docked were ranked and the top score was selected. The scoreswere used to rank the 50 000 compounds. The top 1000 compoundswere clustered into 150 sets that led to the selection of a representativecompound from each set. Among the 150 compounds, 111 were commerciallyavailable and purchased for screening. A dehydrogenase assay thatwe have previously developed75 was usedto screen all 111 compounds at an initial concentration of 50 μM(Figure 6A). Compounds that inhibited ALDH2dehydrogenase activity by more than 50% were selected for a follow-upconcentration dependent study. Among them, five compounds inhibitedthe enzyme in a concentration-dependent manner (Figure 6B). The IC50s were 2.32, ∼23, 0.62, 1.58, and 3.51for ALDH400, ALDH417, ALDH423, ALDH427, and ALDH440, respectively(Figure 6C).


Enrichment of chemical libraries docked to protein conformational ensembles and application to aldehyde dehydrogenase 2.

Wang B, Buchman CD, Li L, Hurley TD, Meroueh SO - J Chem Inf Model (2014)

SVMSP rescoring of MD snapshots identifiesALDH2 inhibitors. Thedehydrogenase activity was screened by measuring the rate of increasein the fluorescence of NADH upon propionaldehyde oxidation. (A) Percentactivity of ALDH2 in the presence of 50 μM of each of the 111compounds that were tested; (B) concentration-dependent curves forfive compounds that inhibited in the initial screen; and (C) chemicalstructures for the five compounds (ALDH400, ALDH417, ALDH423, ALDH427,and ALDH440) along with their IC50s for inhibition of ALDH2 dehydrogenaseactivity.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4114474&req=5

fig6: SVMSP rescoring of MD snapshots identifiesALDH2 inhibitors. Thedehydrogenase activity was screened by measuring the rate of increasein the fluorescence of NADH upon propionaldehyde oxidation. (A) Percentactivity of ALDH2 in the presence of 50 μM of each of the 111compounds that were tested; (B) concentration-dependent curves forfive compounds that inhibited in the initial screen; and (C) chemicalstructures for the five compounds (ALDH400, ALDH417, ALDH423, ALDH427,and ALDH440) along with their IC50s for inhibition of ALDH2 dehydrogenaseactivity.
Mentions: We applied SVMSP rescoringof MD structures to the aldehyde dehydrogenase 2 (ALDH2) enzyme usingSVMSP as the scoring approach. The crystal structure of ALDH2 in itsapo form (PDB code: 1O04) was used to carry out explicit-solvent unbiased MD simulations.70−73 Five independent simulations with 7 ns in length (35 ns total) yielded25 000 snapshots. These were clustered by RMSD using ptraj(74) as described above. Aset of 50 representative snapshots were selected from the clusters.A focused set of the ChemDiv commercial library64 containing 50 000 compounds were docked to eachof the 50 snapshots by AutoDock Vina.61 Docked receptor–ligand complexes were rescored with SVMSP.For each of the 50 000 compounds, the 50 MD snapshots to whichthey were docked were ranked and the top score was selected. The scoreswere used to rank the 50 000 compounds. The top 1000 compoundswere clustered into 150 sets that led to the selection of a representativecompound from each set. Among the 150 compounds, 111 were commerciallyavailable and purchased for screening. A dehydrogenase assay thatwe have previously developed75 was usedto screen all 111 compounds at an initial concentration of 50 μM(Figure 6A). Compounds that inhibited ALDH2dehydrogenase activity by more than 50% were selected for a follow-upconcentration dependent study. Among them, five compounds inhibitedthe enzyme in a concentration-dependent manner (Figure 6B). The IC50s were 2.32, ∼23, 0.62, 1.58, and 3.51for ALDH400, ALDH417, ALDH423, ALDH427, and ALDH440, respectively(Figure 6C).

Bottom Line: Among individual MD snapshots, many exhibited enrichment that was significantly better than the crystal structure, but no correlation between enrichment and structural deviation from crystal structure was found.We found remarkable increase in enrichment power, particularly for p38, where the ROC-AUC increased by 0.30 to 0.85.We found that the use of randomly selected compounds docked to the target of interest using SVMSP led to notable enrichment for EGFR and Src MD snapshots.

View Article: PubMed Central - PubMed

Affiliation: Department of Biochemistry and Molecular Biology, ‡Melvin and Bren Simon Cancer Center, §Center for Computational Biology and Bioinformatics, and ⊥Stark Neurosciences Institute, Indiana University School of Medicine , 535 Barnhill Drive, Indianapolis, Indiana 46202, United States.

ABSTRACT
Molecular recognition is a complex process that involves a large ensemble of structures of the receptor and ligand. Yet, most structure-based virtual screening is carried out on a single structure typically from X-ray crystallography. Explicit-solvent molecular dynamics (MD) simulations offer an opportunity to sample multiple conformational states of a protein. Here we evaluate our recently developed scoring method SVMSP in its ability to enrich chemical libraries docked to MD structures of seven proteins from the Directory of Useful Decoys (DUD). SVMSP is a target-specific rescoring method that combines machine learning with statistical potentials. We find that enrichment power as measured by the area under the ROC curve (ROC-AUC) is not affected by increasing the number of MD structures. Among individual MD snapshots, many exhibited enrichment that was significantly better than the crystal structure, but no correlation between enrichment and structural deviation from crystal structure was found. We followed an innovative approach by training SVMSP scoring models using MD structures (SVMSPMD). The resulting models were applied to two difficult cases (p38 and CDK2) for which enrichment was not better than random. We found remarkable increase in enrichment power, particularly for p38, where the ROC-AUC increased by 0.30 to 0.85. Finally, we explored approaches for a priori identification of MD snapshots with high enrichment power from an MD simulation in the absence of active compounds. We found that the use of randomly selected compounds docked to the target of interest using SVMSP led to notable enrichment for EGFR and Src MD snapshots. SVMSP rescoring of protein-compound MD structures was applied for the search of small-molecule inhibitors of the mitochondrial enzyme aldehyde dehydrogenase 2 (ALDH2). Rank-ordering of a commercial library of 50 000 compounds docked to MD structures of ALDH2 led to five small-molecule inhibitors. Four compounds had IC50s below 5 μM. These compounds serve as leads for the design and synthesis of more potent and selective ALDH2 inhibitors.

Show MeSH
Related in: MedlinePlus