Successful prediction of LC8 binding to intrinsically disordered proteins sheds light on AlphaFold’s black box DOI Creative Commons
Douglas R. Walker, Gretchen Fujimura, Juan M. Vanegas

et al.

Frontiers in Molecular Biosciences, Journal Year: 2025, Volume and Issue: 12

Published: April 23, 2025

Introduction LC8 is a hub protein involved in many processes from tumor suppression and cell cycle regulation to neurotransmission viral infection. Despite recent progress, prediction of binding sites for plagued by motif variability multitude weakly motifs, especially when depends on multivalency. Our site algorithm, LC8Pred has proven useful uncovering new binders, but insufficient finding all sites. Methods To address this, we probed the ability general structure predictor, AlphaFold, predict whether given sequence binds LC8. Certain combinations in-built AlphaFold scores were extracted distributions binders compared nonbinders. Results successfully places proteins at correct interface A set threshold values built-in enables differentiation between known nonbinders with minimal false positive (8%) acceptable negative rates (20%). This cutoff, along more inclusive was used elusive bind Discussion Correlations affinities provide insight into black box indicate that learned an inaccurate energy function nevertheless making inferences conclusions about physical systems. Binding predicted this method can be prioritized investigation comparing result LC8Pred, local structure, evolutionary conservation.

Language: Английский

Design of target specific peptide inhibitors using generative deep learning and molecular dynamics simulations DOI Creative Commons
Sijie Chen, Tong Lin,

Ruchira Basu

et al.

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: Feb. 21, 2024

We introduce a computational approach for the design of target-specific peptides. Our method integrates Gated Recurrent Unit-based Variational Autoencoder with Rosetta FlexPepDock peptide sequence generation and binding affinity assessment. Subsequently, molecular dynamics simulations are employed to narrow down selection peptides experimental assays. apply this strategy inhibitors that specifically target β-catenin NF-κB essential modulator. Among twelve inhibitors, six exhibit improved compared parent peptide. Notably, best C-terminal binds an IC

Language: Английский

Citations

82
From interaction networks to interfaces, scanning intrinsically disordered regions using AlphaFold2 DOI Creative Commons
Hélène Bret, Jinmei Gao, Diego Javier Zea

et al.

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: Jan. 18, 2024

The revolution brought about by AlphaFold2 opens promising perspectives to unravel the complexity of protein-protein interaction networks. analysis networks obtained from proteomics experiments does not systematically provide delimitations regions. This is particular concern in case interactions mediated intrinsically disordered regions, which site generally small. Using a dataset protein-peptide complexes involving regions that are non-redundant with structures used training, we show when using full sequences proteins, AlphaFold2-Multimer only achieves 40% success rate identifying correct and structure interface. By delineating region into fragments decreasing size combining different strategies for integrating evolutionary information, manage raise this up 90%. We obtain similar rates much larger protein taken ELM database. Beyond identification site, our study also explores specificity issues. advantages limitations confidence score discriminate between alternative binding partners, task can be particularly challenging small motifs.

Language: Английский

Citations

49
Systematic discovery of protein interaction interfaces using AlphaFold and experimental validation DOI Creative Commons
Chop Yan Lee, Dalmira Hubrich, Julia K. Varga

et al.

Molecular Systems Biology, Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 15, 2024

Abstract Structural resolution of protein interactions enables mechanistic and functional studies as well interpretation disease variants. However, structural data is still missing for most because we lack computational experimental tools at scale. This particularly true mediated by short linear motifs occurring in disordered regions proteins. We find that AlphaFold-Multimer predicts with high sensitivity but limited specificity structures domain-motif when using small fragments input. Sensitivity decreased substantially long or full length delineated a fragmentation strategy suited the prediction interfaces applied it to between human proteins associated neurodevelopmental disorders. enabled highly confident likely disease-related novel interfaces, which further experimentally corroborated FBXO23-STX1B, STX1B-VAMP2, ESRRG-PSMC5, PEX3-PEX19, PEX3-PEX16, SNRPB-GIGYF1 providing molecular insights diverse biological processes. Our work highlights exciting perspectives, also reveals clear limitations need future developments maximize power Alphafold-Multimer interface predictions.

Language: Английский

Citations

32
Primate-conserved carbonic anhydrase IV and murine-restricted LY6C1 enable blood-brain barrier crossing by engineered viral vectors DOI Creative Commons
Timothy F. Shay, Erin E. Sullivan, Xiaozhe Ding

et al.

Science Advances, Journal Year: 2023, Volume and Issue: 9(16)

Published: April 19, 2023

The blood-brain barrier (BBB) presents a major challenge for delivering large molecules to study and treat the central nervous system. This is due in part scarcity of targets known mediate BBB crossing. To identify novel targets, we leverage panel adeno-associated viruses (AAVs) previously identified through mechanism-agnostic directed evolution improved transcytosis. Screening potential cognate receptors enhanced crossing, two targets: murine-restricted LY6C1 widely conserved carbonic anhydrase IV (CA-IV). We apply AlphaFold-based silico methods generate capsid-receptor binding models predict affinity AAVs these receptors. Demonstrating how tools can unlock target-focused engineering strategies, create an LY6C1-binding vector, AAV-PHP.eC, that, unlike our prior PHP.eB, also works Ly6a-deficient mouse strains such as BALB/cJ. Combined with structural insights from computational modeling, identification primate-conserved CA-IV enables design more specific potent human brain-penetrant chemicals biologicals, including gene delivery vectors.

Language: Английский

Citations

29
Deep learning for advancing peptide drug development: Tools and methods in structure prediction and design DOI
Xinyi Wu, Huitian Lin, Renren Bai

et al.

European Journal of Medicinal Chemistry, Journal Year: 2024, Volume and Issue: 268, P. 116262 - 116262

Published: Feb. 19, 2024

Language: Английский

Citations

13
AlphaFold2 Predicts Whether Proteins Interact Amidst Confounding Structural Compatibility DOI
Juliette Martin

Journal of Chemical Information and Modeling, Journal Year: 2024, Volume and Issue: 64(5), P. 1473 - 1480

Published: Feb. 19, 2024

Predicting whether two proteins physically interact is one of the holy grails computational biology, galvanized by rapid advancements in deep learning. AlphaFold2, although not developed with this goal, promising respect. Here, I test prediction capability AlphaFold2 on a very challenging data set, where are structurally compatible, even when they do interact. achieves high discrimination between interacting and non-interacting proteins, cases misclassifications can either be rescued revisiting input sequences or suggest false positives negatives set. thus impaired compatibility protein structures has potential to applied large scale.

Language: Английский

Citations

8
Dietary Triterpenoids in Functional Food and Drug Ingredients: a review of structure-activity relationships, biosynthesis, applications, and AI-driven strategies DOI
Chao Fang, Haixia Yang, Daidi Fan

et al.

Trends in Food Science & Technology, Journal Year: 2025, Volume and Issue: unknown, P. 104961 - 104961

Published: March 1, 2025

Language: Английский

Citations

1
Structure Determination of Challenging Protein–Peptide Complexes Combining NMR Chemical Shift Data and Molecular Dynamics Simulations DOI
Arup Mondal, G.V.T. Swapna,

María M. López

et al.

Journal of Chemical Information and Modeling, Journal Year: 2023, Volume and Issue: 63(7), P. 2058 - 2072

Published: March 29, 2023

Intrinsically disordered regions of proteins often mediate important protein–protein interactions. However, the folding-upon-binding nature many polypeptide–protein interactions limits ability modeling tools to predict three-dimensional structures such complexes. To address this problem, we have taken a tandem approach combining NMR chemical shift data and molecular simulations determine peptide–protein Here, use MELD (Modeling Employing Limited Data) technique applied polypeptide complexes formed with extraterminal domain (ET) bromo (BET) proteins, which exhibit high degree binding plasticity. This system is particularly challenging as process includes allosteric changes across ET receptor upon binding, partners can adopt different conformations (e.g., helices hairpins) in complex. In blind study, new successfully modeled bound-state poses, using only protein backbone data, excellent agreement experimentally determined for moderately tight (Kd ∼100 nM) binders. The hybrid + required additional peptide ligand weaker ∼250 μM) partners. AlphaFold also predicts some these whereas provide qualitative rankings, directly estimate relative affinities. offers powerful tool structural analysis protein–polypeptide involving disorder-to-order transitions complex formation, are not most other prediction methods, providing both 3D their

Language: Английский

Citations

18
DeepRank-GNN-esm: a graph neural network for scoring protein–protein models using protein language model DOI Creative Commons
Xiaotong Xu, Alexandre M. J. J. Bonvin

Bioinformatics Advances, Journal Year: 2024, Volume and Issue: 4(1)

Published: Jan. 1, 2024

Protein-Protein interactions (PPIs) play critical roles in numerous cellular processes. By modelling the 3D structures of correspond protein complexes valuable insights can be obtained, providing, e.g. starting points for drug and design. One challenge process is however identification near-native models from large pool generated models. To this end we have previously developed DeepRank-GNN, a graph neural network that integrates structural sequence information to enable effective pattern learning at PPI interfaces. Its main features are related Position Specific Scoring Matrices (PSSMs), which computationally expensive generate, significantly limits algorithm's usability.

Language: Английский

Citations

8
AlphaFold2 enables accurate deorphanization of ligands to single-pass receptors DOI
Niels Banhos Danneskiold‐Samsøe,

Deniz Kavi,

Kevin M. Jude

et al.

Cell Systems, Journal Year: 2024, Volume and Issue: 15(11), P. 1046 - 1060.e3

Published: Nov. 1, 2024

Language: Английский

Citations

8