Assessing Structures and Conformational Ensembles of Apo and Holo Protein States Using Randomized Alanine Sequence Scanning Combined with Shallow Subsampling in AlphaFold2 : Insights and Lessons from Predictions of Functional Allosteric Conformations DOI Open Access
Nishank Raisinghani, Victoria N. Parikh, Brian Foley

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 6, 2024

Abstract Proteins often exist in multiple conformational states, influenced by the binding of ligands or substrates. The study these particularly apo (unbound) and holo (ligand-bound) forms, is crucial for understanding protein function, dynamics, interactions. In current study, we use AlphaFold2 that combines randomized alanine sequence masking with shallow alignment subsampling to expand diversity predicted structural ensembles capture changes between forms. Using several well-established datasets structurally diverse apo-holo pairs, proposed approach enables robust predictions structures ensembles, while also displaying notably similar dynamics distributions. These observations are consistent view intrinsic allosteric proteins defined topology fold favors conserved motions driven soft modes. Our findings support notion approaches can yield reasonable accuracy predicting minor adjustments especially moderate localized upon ligand binding. However, large, hinge-like domain movements, tends predict most stable orientation which typically form rather than full range functional conformations characteristic ensemble. results indicate modeling states may require more accurate characterization flexible regions detection high energy conformations. By incorporating a wider variety training including both model learn recognize occur

Language: Английский

AFsample2 predicts multiple conformations and ensembles with AlphaFold2 DOI Creative Commons
Yogesh Kalakoti, Björn Wallner

Communications Biology, Journal Year: 2025, Volume and Issue: 8(1)

Published: March 5, 2025

Understanding protein dynamics and conformational states is crucial for insights into biological processes disease mechanisms, which can aid drug development. Recently, several methods have been devised to broaden the predictions made by AlphaFold2 (AF2). We introduce AFsample2, a method using random MSA column masking reduce co-evolutionary signals, enhancing structural diversity in AF2-generated models. AFsample2 effectively predicts alternative various proteins, producing high-quality end diverse ensembles. In OC23 dataset, alternate state models improved (ΔTM>0.05) 9 out of 23 cases without affecting preferred generation. Similar results were seen 16 membrane transporters, with 11 targets showing improvement. TM-score improvements experimental substantial, sometimes exceeding 50%, improving from 0.58 0.98. Additionally, increased intermediate conformations 70% compared standard AF2, highly confident potentially representing states. For four targets, predicted structurally similar known homologs PDB, suggesting that they are true These findings indicate used provide proteins multiple states, as well potential paths between

Language: Английский

Citations

1
Predicting protein conformational motions using energetic frustration analysis and AlphaFold2 DOI
Xingyue Guan, Qian-Yuan Tang, Weitong Ren

et al.

Proceedings of the National Academy of Sciences, Journal Year: 2024, Volume and Issue: 121(35)

Published: Aug. 20, 2024

Proteins perform their biological functions through motion. Although high throughput prediction of the three-dimensional static structures proteins has proved feasible using deep-learning-based methods, predicting conformational motions remains a challenge. Purely data-driven machine learning methods encounter difficulty for addressing such because available laboratory data on are still limited. In this work, we develop method generating protein allosteric by integrating physical energy landscape information into methods. We show that local energetic frustration, which represents quantification features governing dynamics, can be utilized to empower AlphaFold2 (AF2) predict motions. Starting from ground state structures, integrative generates alternative as well pathways motions, progressive enhancement frustration in input multiple sequence alignment sequences. For model adenylate kinase, generated consistent with experimental and molecular dynamics simulation data. Applying another two KaiB ribose-binding protein, involve large-amplitude changes, also successfully generate conformations. how extract overall AF2 topography, been considered many black box. Incorporating knowledge structure algorithms provides useful strategy address challenges dynamic proteins.

Language: Английский

Citations

9
Bridging Prediction and Reality: Comprehensive Analysis of Experimental and AlphaFold 2 Full-Length Nuclear Receptor Structures DOI Creative Commons
Akerke Mazhibiyeva, Tri Thanh Pham, Karina Pats

et al.

Computational and Structural Biotechnology Journal, Journal Year: 2025, Volume and Issue: unknown

Published: May 1, 2025

Language: Английский

Citations

0
Predicting Mutation-Induced Allosteric Changes in Structures and Conformational Ensembles of the ABL Kinase Using AlphaFold2 Adaptations with Alanine Sequence Scanning DOI Open Access
Nishank Raisinghani, Mohammed Alshahrani,

Grace Gupta

et al.

International Journal of Molecular Sciences, Journal Year: 2024, Volume and Issue: 25(18), P. 10082 - 10082

Published: Sept. 19, 2024

Despite the success of AlphaFold2 approaches in predicting single protein structures, these methods showed intrinsic limitations multiple functional conformations allosteric proteins and have been challenged to accurately capture effects point mutations that induced significant structural changes. We examined several implementations predict conformational ensembles for state-switching mutants ABL kinase. The results revealed a combination randomized alanine sequence masking with shallow alignment subsampling can significantly expand diversity predicted shifts populations active inactive states. Consistent NMR experiments, M309L/L320I M309L/H415P perturb regulatory spine networks featured increased population fully closed state. proposed adaptation AlphaFold reproduce experimentally observed mutation-induced redistributions relative states on rearrangements kinase domain. ensemble-based network analysis complemented predictions by revealing hotspots correspond mutational sites which may explain global effect changes between This study suggested attention-based learning long-range dependencies positions homologous folds deciphering patterns interactions further augment predictive abilities modeling alternative sates, transformations.

Language: Английский

Citations

3
Probing Functional Allosteric States and Conformational Ensembles of the Allosteric Protein Kinase States and Mutants: Atomistic Modeling and Comparative Analysis of AlphaFold2, OmegaFold, and AlphaFlow Approaches and Adaptations DOI
Nishank Raisinghani, Mohammed Alshahrani,

Grace Gupta

et al.

The Journal of Physical Chemistry B, Journal Year: 2024, Volume and Issue: 128(45), P. 11088 - 11107

Published: Nov. 1, 2024

This study reports a comprehensive analysis and comparison of several AlphaFold2 adaptations OmegaFold AlphaFlow approaches in predicting distinct allosteric states, conformational ensembles, mutation-induced structural effects for panel state-switching ABL mutants. The results revealed that the proposed adaptation with randomized alanine sequence scanning can generate functionally relevant states ensembles kinase qualitatively capture unique pattern population shifts between active inactive Consistent NMR experiments, predicted G269E/M309L/T408Y mutant could induce changes sample significant fraction fully I2 form which is low-populated, high-energy state wild-type protein. We also demonstrated other mutants G269E/M309L/T334I M309L/L320I/T334I introduce single activating T334I mutation reverse equilibrium populate exclusively form. While precise quantitative predictions relative populations various hidden remain challenging, our provide evidence adequately detect spectrum redistributions structurally functional conformations. further validated robustness architecture BSK8 differences ligand-unbound apo ATP-bound forms BSK8. this comparative suggested AlpahFold2, OmegaFold, may be driven by memorization existing protein folds are strongly biased toward thermodynamically stable ground kinases, highlighting limitations challenges AI-based methodologies detecting alternative conformations, accurate characterization physically prediction changes.

Language: Английский

Citations

3
AlphaFold2-Based Characterization of Apo and Holo Protein Structures and Conformational Ensembles Using Randomized Alanine Sequence Scanning Adaptation: Capturing Shared Signature Dynamics and Ligand-Induced Conformational Changes DOI Open Access
Nishank Raisinghani, Victoria N. Parikh, Brian Foley

et al.

International Journal of Molecular Sciences, Journal Year: 2024, Volume and Issue: 25(23), P. 12968 - 12968

Published: Dec. 2, 2024

Proteins often exist in multiple conformational states, influenced by the binding of ligands or substrates. The study these particularly apo (unbound) and holo (ligand-bound) forms, is crucial for understanding protein function, dynamics, interactions. In current study, we use AlphaFold2, which combines randomized alanine sequence masking with shallow alignment subsampling to expand diversity predicted structural ensembles capture changes between forms. Using several well-established datasets structurally diverse apo-holo pairs, proposed approach enables robust predictions structures ensembles, while also displaying notably similar dynamics distributions. These observations are consistent view that intrinsic allosteric proteins defined topology fold favor conserved motions driven soft modes. Our findings provide evidence AlphaFold2 combined can yield accurate results predicting moderate adjustments especially localized upon ligand binding. For large hinge-like domain movements, predict functional conformations characteristic both ligand-bound absence information. relevant using this AlphaFold adaptation probing selection mechanisms according adopt conformations, including those competent indicate modeling states may require more characterization flexible regions detection high-energy conformations. By incorporating a wider variety training datasets, model learn recognize occur

Language: Английский

Citations

1
Assessing Structures and Conformational Ensembles of Apo and Holo Protein States Using Randomized Alanine Sequence Scanning Combined with Shallow Subsampling in AlphaFold2 : Insights and Lessons from Predictions of Functional Allosteric Conformations DOI Open Access
Nishank Raisinghani, Victoria N. Parikh, Brian Foley

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 6, 2024

Abstract Proteins often exist in multiple conformational states, influenced by the binding of ligands or substrates. The study these particularly apo (unbound) and holo (ligand-bound) forms, is crucial for understanding protein function, dynamics, interactions. In current study, we use AlphaFold2 that combines randomized alanine sequence masking with shallow alignment subsampling to expand diversity predicted structural ensembles capture changes between forms. Using several well-established datasets structurally diverse apo-holo pairs, proposed approach enables robust predictions structures ensembles, while also displaying notably similar dynamics distributions. These observations are consistent view intrinsic allosteric proteins defined topology fold favors conserved motions driven soft modes. Our findings support notion approaches can yield reasonable accuracy predicting minor adjustments especially moderate localized upon ligand binding. However, large, hinge-like domain movements, tends predict most stable orientation which typically form rather than full range functional conformations characteristic ensemble. results indicate modeling states may require more accurate characterization flexible regions detection high energy conformations. By incorporating a wider variety training including both model learn recognize occur

Language: Английский

Citations

0