Computed structures of core eukaryotic protein complexes

By I. Humphreys, Jimin Pei, Minkyung Baek, Aditya Krishnakumar, Ivan V. Anishchenko, Sergey Ovchinnikov, Jing Zhang, Travis J Ness, Sudeep Banjade, S. Bagde and 20 more. 2021

442 citations recorded by the source corpus.

Description Deep learning for protein interactions The use of deep learning has revolutionized the field of protein modeling. Humphreys et al. combined this approach with proteome-wide, coevolution-guided protein interaction identification to conduct a large-scale screen of protein-protein interactions in yeast (see the Perspective by Pereira and Schwede). The authors generated predicted interactions and accurate structures for complexes spanning key biological processes in Saccharomyces cerevisiae. The complexes include larger protein assemblies such as trimers, tetramers, and pentamers and provide insights into biological function. —VV Proteome-wide coevolution and deep-learning methods identify and build accurate models of eukaryotic protein complexes. INTRODUCTION Protein-protein interactions play critical roles in biology, but the structures of many eukaryotic protein complexes are unknown, and there are likely many interactions not yet identified. High-throughput experimental methods such as yeast two-hybrid and affinity-purification mass spectrometry have been used to identify interactions in multiple organisms, but there are inconsistencies between different datasets, and the methods do not provide high-resolution structural information. Here, we use deep learning methods to systematically identify and build structures for the protein complexes that mediate key processes in eukaryotes. RATIONALE Interacting proteins often coevolve, and in prokaryotes, evolutionary information can be used to identify interactions on the proteome scale at an accuracy higher than that of experimental screens. Extending this method to eukaryotes is complicated because there are fewer genome sequences available, resulting in weaker coevolutionary signals. The deep learning methods RoseTTAFold and AlphaFold, have a rich understanding of protein sequence-structure relationships, and so could help overcome this limitation. RESULTS We developed a coevolution-guided protein interaction identification pipeline that incorporates a rapidly computable version of RoseTTAFold with the slower but more accurate AlphaFold to systematically evaluate interactions between 8.3 million pairs of yeast proteins. RoseTTAFold alone has comparable performance in identifying protein-protein interactions to that of large-scale experimental methods; combination with AlphaFold increases identification accuracy. In total, we constructed models for 106 previously unidentified assemblies and 806 that were structurally uncharacterized. These complexes provide rich insights into a range of biological processes from transcription, translation, and DNA repair to protein transport and modification. For example, Rad51 plays a pivotal role in DNA repair through homologous recombination, and mutations are associated with Fanconi anemia and cancer in humans. Rad55 and Rad57 are positive regulators of Rad51 assembly on single-stranded DNA. Our Rad55–Rad57–Rad51 complex model suggests that Rad55–Rad57 can bind at the 5′ end of the Rad51 single-stranded DNA filament and may stabilize the filament conformation of Rad51. Glycosylphosphatidylinositol transamidase (GPI-T) is a pentameric enzyme complex that catalyzes the attachment of GPI anchors to the C terminus of proteins. GPI-T is structurally uncharacterized, and mutations in subunits of the complex have been implicated in neurodevelopmental disorders and cancer in humans. Our model of the five-protein assembly shows that the previously identified catalytic dyad is positioned adjacent to a channel formed by three other subunits that could function in C-terminal GPI-T signal peptide recognition. CONCLUSION Our approach extends the range of large-scale deep learning–based structure modeling from monomeric proteins to protein assemblies. Following up on the many new interactions and complex structures should advance the understanding of a wide range of eukaryotic cellular processes and provide new targets for therapeutic intervention. Our results herald a new era of structural biology in which computation plays a fundamental role in both interaction discovery and structure determination. Examples of predicted complexes. Protein-protein interactions play critical roles in biology, but the structures of many eukaryotic protein complexes are unknown, and there are likely many interactions not yet identified. We take advantage of advances in proteome-wide amino acid coevolution analysis and deep-learning–based structure modeling to systematically identify and build accurate models of core eukaryotic protein complexes within the Saccharomyces cerevisiae proteome. We use a combination of RoseTTAFold and AlphaFold to screen through paired multiple sequence alignments for 8.3 million pairs of yeast proteins, identify 1505 likely to interact, and build structure models for 106 previously unidentified assemblies and 806 that have not been structurally characterized. These complexes, which have as many as five subunits, play roles in almost all key processes in eukaryotic cells and provide broad insights into biological function.

Subfields: Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

Citations in the corpus

Learning inverse folding from millions of predicted structures (2022, background): “Additionally, with the recent progress in structure prediction for multi-chain protein complexes (Evans et al., 2022; Humphreys et al., 2021), predicted complex structures could be another valuable source of data for learning protein-protein interactions.”
Scaffolding protein functional sites using deep learning (2022, background): “RosettaFold is able to predict the structures of protein complexes (40), and we hypothesized that it could generate additional binding interactions between hallucinated or inpainted binder and a target beyond the scaffolded motif.”
Towards a structurally resolved human protein interaction network (2021, methodology): “These approaches can correctly predict the structures of up to 60% of dimers 18 , and have been used to predict structures of 1,506 Saccharomyces cerevisiae protein interactions 22 .”
AlphaFold Models of Small Proteins Rival the Accuracy of Solution NMR Structures (2022, background): “…recognized that the structures of protein-protein complexes and multimeric assemblies can often be reliably modeled using modified approaches with these same AI platforms (Baek et al., 2021a; Baek et al., 2021b; Evans et al., 2021; Humphreys et al., 2021; Colman et al., 2022; Mondal et al., 2022).”
Benchmarking AlphaFold for protein complex modeling reveals accuracy determinants (2021, background): “58 Prospective developments that build upon and optimize the Alpha-Fold framework, or utilize other geometric deep learning methods, can bring the field closer to solving the long-standing challenge of predictive protein docking.”
Accurate prediction of nucleic acid and protein-nucleic acid complexes using RoseTTAFoldNA (2022, methodology): “…exceptions: a) only complexes fewer than 1000 residues plus nucleotides in length were considered; and b) for complexes containing more than one unique protein chains, paired MSAs were created by merging sequences from the same organism into a single combined sequence (following prior work [13]).”
Clustering predicted structures at the scale of the known protein universe (2023, background): “AlphaFold2 and its predicted structures have now been used for diverse applications, including studies of protein pockets 6 , prediction of structures of complexes 7,8 , studies of structural similarity 9 , novel fold predictions 10 and even improvement of genomic annotation 11 .”
Accurate prediction of protein–nucleic acid complexes using RoseTTAFoldNA (2023, methodology): “Paired MSAs were generated for complexes with multiple protein chains as described previously 14 .”
A chronotherapeutics-applicable multi-target therapeutics based on AI: Example of therapeutic hypothermia (2022, methodology): “RoseTTAFold is based on the Rosetta software, which is designed for macromolecular modeling, docking and design [20] RoseTTAFold also has good application [21] in the research of protein structure prediction.”
Benchmarking AlphaFold‐enabled molecular docking predictions for antibiotic discovery (2022, background): “The advent of AlphaFold2 and other deep neural networks for protein folding, such as RoseTTAFold (Baek et al , 2021; Humphreys et al , 2021), has been widely anticipated and celebrated in structural biology.”
RCSB Protein Data Bank (RCSB.org): delivery of experimentally-determined PDB structures alongside one million computed structure models of proteins from artificial intelligence/machine learning (2022, background): “• From AlphaFold DB: Generated by DeepMind using AlphaFold 2 • Model organism proteomes: 326 175 protein structures from 48 different model organisms • Global health proteomes: 238 274 protein structures from various disease-causing organisms • Swiss-Prot sequences (43): 542 380 protein structures, ”
AI-based structure prediction empowers integrative structural analysis of human nuclear pores (2022, background): “However, AlphaFold and RoseTTAfold have recently demonstrated unprecedented accuracy in predicting structures ofmonomeric proteins (26, 27, 60–65) and complexes (34, 36, 61, 66).”
AlphaFill: enriching AlphaFold models with ligands and cofactors (2022, methodology): “The multimer issue was addressed by the development of AlphaFoldMultimer 6 and RoseTTAFold 7 , that can predict complex protein assemblies.”
The STRING database in 2023: protein–protein association networks and functional enrichment analyses for any sequenced genome of interest (2022, background): “Recent developments have further improved our ability to unravel this connectivity, through techniques such as high-throughput genetic screens (3–5), systematic co-fractionation of proteomes (6,7), in-vivo proteome-wide cross-linking of proteins (8–10), and deep learning-based computational predicti”
GPCRdb in 2023: state-specific structure models using AlphaFold2 and new ligand resources (2022, methodology): “Lundbeck Foundation [R313-2019-526]; Novo Nordisk Foundation [NNF18OC0031226 to D.E.G.]; Hungarian Academy of Sciences [NAP3.”
AlphaFold2 models indicate that protein sequence determines both structure and dynamics (2022, methodology): “The simulations were conducted using a timestep of 2 fs and a non-bonded interaction cutoff switching of 9 to 11 Å .”