Deep Neural Networks for Predicting the Affinity Landscape of Protein-protein Interactions

Overview

Journal iScience

Publisher Cell Press

Date 2024 Sep 23

PMID 39310756

Authors

Reut Meiri

Shay-Lee Aharoni Lotati

Yaron Orenstein

Niv Papo

Affiliations

Soon will be listed here.

Abstract

Studies determining protein-protein interactions (PPIs) by deep mutational scanning have focused mainly on a narrow range of affinities within complexes and thus include only partial coverage of the mutation space of given proteins. By inserting an affinity-reducing N-terminal alanine in the N-terminal domain of the tissue inhibitor of metalloproteinases-2 (N-TIMP2), we overcame the limitation of its narrow affinity range for matrix metalloproteinase 9 (MMP9). We trained deep neural networks (DNNs) to quantitatively predict the binding affinity of unobserved wild-type variants and variants carrying an N-terminal alanine. Good correlation was obtained between predicted and observed log enrichment ratio (ER) values, which also correlated with the affinity of N-TIMP2 variants to MMP9. Our ability to predict affinities of unobserved N-TIMP2 variants was confirmed on an independent dataset of experimentally validated N-TIMP2 proteins. This ability is of significant importance in the field of PPI prediction and for developing therapies targeting these interactions.

References

Yano H, Nishimiya D, Kawaguchi Y, Tamura M, Hashimoto R . Discovery of potent and specific inhibitors targeting the active site of MMP-9 from the engineered SPINK2 library. PLoS One. 2020; 15(12):e0244656. PMC: 7771667. DOI: 10.1371/journal.pone.0244656. View

Shirian J, Sharabi O, Shifman J . Cold Spots in Protein Binding. Trends Biochem Sci. 2016; 41(9):739-745. DOI: 10.1016/j.tibs.2016.07.002. View

Livesey B, Marsh J . Using deep mutational scanning to benchmark variant effect predictors and identify disease mutations. Mol Syst Biol. 2020; 16(7):e9380. PMC: 7336272. DOI: 10.15252/msb.20199380. View

Hu J, Van den Steen P, Sang Q, Opdenakker G . Matrix metalloproteinase inhibitors as therapy for inflammatory and vascular diseases. Nat Rev Drug Discov. 2007; 6(6):480-98. DOI: 10.1038/nrd2308. View

Song H, Bremer B, Hinds E, Raskutti G, Romero P . Inferring Protein Sequence-Function Relationships with Large-Scale Positive-Unlabeled Learning. Cell Syst. 2020; 12(1):92-101.e8. PMC: 7856229. DOI: 10.1016/j.cels.2020.10.007. View

Magoc T, Salzberg S . FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics. 2011; 27(21):2957-63. PMC: 3198573. DOI: 10.1093/bioinformatics/btr507. View

Chao G, Lau W, Hackel B, Sazinsky S, Lippow S, Wittrup K . Isolating and engineering human antibodies using yeast surface display. Nat Protoc. 2007; 1(2):755-68. DOI: 10.1038/nprot.2006.94. View

Jubb H, Pandurangan A, Turner M, Ochoa-Montano B, Blundell T, Ascher D . Mutations at protein-protein interfaces: Small changes over big surfaces have large impacts on human health. Prog Biophys Mol Biol. 2016; 128:3-13. DOI: 10.1016/j.pbiomolbio.2016.10.002. View

Arkadash V, Yosef G, Shirian J, Cohen I, Horev Y, Grossman M . Development of High Affinity and High Specificity Inhibitors of Matrix Metalloproteinase 14 through Computational Design and Directed Evolution. J Biol Chem. 2017; 292(8):3481-3495. PMC: 5336179. DOI: 10.1074/jbc.M116.756718. View

10.

Morrison K, Weiss G . Combinatorial alanine-scanning. Curr Opin Chem Biol. 2001; 5(3):302-7. DOI: 10.1016/s1367-5931(00)00206-4. View

11.

Angelini A, Chen T, de Picciotto S, Yang N, Tzeng A, Santos M . Protein Engineering and Selection Using Yeast Surface Display. Methods Mol Biol. 2015; 1319:3-36. DOI: 10.1007/978-1-4939-2748-7_1. View

12.

Starr T, Thornton J . Epistasis in protein evolution. Protein Sci. 2016; 25(7):1204-18. PMC: 4918427. DOI: 10.1002/pro.2897. View

13.

Naftaly S, Cohen I, Shahar A, Hockla A, Radisky E, Papo N . Mapping protein selectivity landscapes using multi-target selective screening and next-generation sequencing of combinatorial libraries. Nat Commun. 2018; 9(1):3935. PMC: 6158287. DOI: 10.1038/s41467-018-06403-x. View

14.

Currin A, Kwok J, Sadler J, Bell E, Swainston N, Ababi M . GeneORator: An Effective Strategy for Navigating Protein Sequence Space More Efficiently through Boolean OR-Type DNA Libraries. ACS Synth Biol. 2019; 8(6):1371-1378. PMC: 7007284. DOI: 10.1021/acssynbio.9b00063. View

15.

Wingfield P, Sax J, Stahl S, Kaufman J, Palmer I, Chung V . Biophysical and functional characterization of full-length, recombinant human tissue inhibitor of metalloproteinases-2 (TIMP-2) produced in Escherichia coli. Comparison of wild type and amino-terminal alanine appended variant with implications for.... J Biol Chem. 1999; 274(30):21362-8. DOI: 10.1074/jbc.274.30.21362. View

16.

Jenson J, Xue V, Stretz L, Mandal T, Reich L, Keating A . Peptide design by optimization on a data-parameterized protein interaction landscape. Proc Natl Acad Sci U S A. 2018; 115(44):E10342-E10351. PMC: 6217393. DOI: 10.1073/pnas.1812939115. View

17.

Heyne M, Papo N, Shifman J . Generating quantitative binding landscapes through fractional binding selections combined with deep sequencing and data normalization. Nat Commun. 2020; 11(1):297. PMC: 6962383. DOI: 10.1038/s41467-019-13895-8. View

18.

Hsu C, Nisonoff H, Fannjiang C, Listgarten J . Learning protein fitness models from evolutionary and assay-labeled data. Nat Biotechnol. 2022; 40(7):1114-1122. DOI: 10.1038/s41587-021-01146-5. View

19.

Sharabi O, Shirian J, Grossman M, Lebendiker M, Sagi I, Shifman J . Affinity- and specificity-enhancing mutations are frequent in multispecific interactions between TIMP2 and MMPs. PLoS One. 2014; 9(4):e93712. PMC: 3977929. DOI: 10.1371/journal.pone.0093712. View

20.

Aharon L, Aharoni S, Radisky E, Papo N . Quantitative mapping of binding specificity landscapes for homologous targets by using a high-throughput method. Biochem J. 2020; 477(9):1701-1719. PMC: 7376575. DOI: 10.1042/BCJ20200188. View