Aminoacyl-tRNA synthetase urzymes optimized by deep learning behave as a quasispecies

Protein design plays a key role in our efforts to work out how genetic coding began. That effort entails urzymes. Urzymes are small, conserved excerpts from full-length aminoacyl-tRNA synthetases that remain active. Urzymes require design to connect disjoint pieces and repair naked nonpolar patches...

Full description

Saved in:
Bibliographic Details
Main Authors: Sourav Kumar Patra, Nicholas Randolph, Brian Kuhlman, Henry Dieckhaus, Laurie Betts, Jordan Douglas, Peter R. Wills, Charles W. Carter Jr.
Format: Article
Language:English
Published: AIP Publishing LLC and ACA 2025-03-01
Series:Structural Dynamics
Online Access:http://dx.doi.org/10.1063/4.0000294
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850281550663385088
author Sourav Kumar Patra
Nicholas Randolph
Brian Kuhlman
Henry Dieckhaus
Laurie Betts
Jordan Douglas
Peter R. Wills
Charles W. Carter Jr.
author_facet Sourav Kumar Patra
Nicholas Randolph
Brian Kuhlman
Henry Dieckhaus
Laurie Betts
Jordan Douglas
Peter R. Wills
Charles W. Carter Jr.
author_sort Sourav Kumar Patra
collection DOAJ
description Protein design plays a key role in our efforts to work out how genetic coding began. That effort entails urzymes. Urzymes are small, conserved excerpts from full-length aminoacyl-tRNA synthetases that remain active. Urzymes require design to connect disjoint pieces and repair naked nonpolar patches created by removing large domains. Rosetta allowed us to create the first urzymes, but those urzymes were only sparingly soluble. We could measure activity, but it was hard to concentrate those samples to levels required for structural biology. Here, we used the deep learning algorithms ProteinMPNN and AlphaFold2 to redesign a set of optimized LeuAC urzymes derived from leucyl-tRNA synthetase. We select a balanced, representative subset of eight variants for testing using principal component analysis. Most tested variants are much more soluble than the original LeuAC. They also span a range of catalytic proficiency and amino acid specificity. The data enable detailed statistical analyses of the sources of both solubility and specificity. In that way, we show how to begin to unwrap the elements of protein chemistry that were hidden within the neural networks. Deep learning networks have thus helped us surmount several vexing obstacles to further investigations into the nature of ancestral proteins. Finally, we discuss how the eight variants might resemble a sample drawn from a population similar to one subject to natural selection.
format Article
id doaj-art-d47d9b2150b448dd8e8240bdabe79d0d
institution OA Journals
issn 2329-7778
language English
publishDate 2025-03-01
publisher AIP Publishing LLC and ACA
record_format Article
series Structural Dynamics
spelling doaj-art-d47d9b2150b448dd8e8240bdabe79d0d2025-08-20T01:48:15ZengAIP Publishing LLC and ACAStructural Dynamics2329-77782025-03-01122024701024701-1510.1063/4.0000294Aminoacyl-tRNA synthetase urzymes optimized by deep learning behave as a quasispeciesSourav Kumar Patra0Nicholas Randolph1Brian Kuhlman2Henry Dieckhaus3Laurie Betts4Jordan Douglas5Peter R. Wills6Charles W. Carter Jr.7Department of Biochemistry and Biophysics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7260, USADepartment of Biochemistry and Biophysics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7260, USADepartment of Biochemistry and Biophysics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7260, USADivision of Chemical Biology and Medicinal Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7355, USADepartment of Biochemistry and Biophysics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7260, USADepartment of Physics, University of Auckland, Auckland, New ZealandDepartment of Physics, University of Auckland, Auckland, New ZealandDepartment of Biochemistry and Biophysics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7260, USAProtein design plays a key role in our efforts to work out how genetic coding began. That effort entails urzymes. Urzymes are small, conserved excerpts from full-length aminoacyl-tRNA synthetases that remain active. Urzymes require design to connect disjoint pieces and repair naked nonpolar patches created by removing large domains. Rosetta allowed us to create the first urzymes, but those urzymes were only sparingly soluble. We could measure activity, but it was hard to concentrate those samples to levels required for structural biology. Here, we used the deep learning algorithms ProteinMPNN and AlphaFold2 to redesign a set of optimized LeuAC urzymes derived from leucyl-tRNA synthetase. We select a balanced, representative subset of eight variants for testing using principal component analysis. Most tested variants are much more soluble than the original LeuAC. They also span a range of catalytic proficiency and amino acid specificity. The data enable detailed statistical analyses of the sources of both solubility and specificity. In that way, we show how to begin to unwrap the elements of protein chemistry that were hidden within the neural networks. Deep learning networks have thus helped us surmount several vexing obstacles to further investigations into the nature of ancestral proteins. Finally, we discuss how the eight variants might resemble a sample drawn from a population similar to one subject to natural selection.http://dx.doi.org/10.1063/4.0000294
spellingShingle Sourav Kumar Patra
Nicholas Randolph
Brian Kuhlman
Henry Dieckhaus
Laurie Betts
Jordan Douglas
Peter R. Wills
Charles W. Carter Jr.
Aminoacyl-tRNA synthetase urzymes optimized by deep learning behave as a quasispecies
Structural Dynamics
title Aminoacyl-tRNA synthetase urzymes optimized by deep learning behave as a quasispecies
title_full Aminoacyl-tRNA synthetase urzymes optimized by deep learning behave as a quasispecies
title_fullStr Aminoacyl-tRNA synthetase urzymes optimized by deep learning behave as a quasispecies
title_full_unstemmed Aminoacyl-tRNA synthetase urzymes optimized by deep learning behave as a quasispecies
title_short Aminoacyl-tRNA synthetase urzymes optimized by deep learning behave as a quasispecies
title_sort aminoacyl trna synthetase urzymes optimized by deep learning behave as a quasispecies
url http://dx.doi.org/10.1063/4.0000294
work_keys_str_mv AT souravkumarpatra aminoacyltrnasynthetaseurzymesoptimizedbydeeplearningbehaveasaquasispecies
AT nicholasrandolph aminoacyltrnasynthetaseurzymesoptimizedbydeeplearningbehaveasaquasispecies
AT briankuhlman aminoacyltrnasynthetaseurzymesoptimizedbydeeplearningbehaveasaquasispecies
AT henrydieckhaus aminoacyltrnasynthetaseurzymesoptimizedbydeeplearningbehaveasaquasispecies
AT lauriebetts aminoacyltrnasynthetaseurzymesoptimizedbydeeplearningbehaveasaquasispecies
AT jordandouglas aminoacyltrnasynthetaseurzymesoptimizedbydeeplearningbehaveasaquasispecies
AT peterrwills aminoacyltrnasynthetaseurzymesoptimizedbydeeplearningbehaveasaquasispecies
AT charleswcarterjr aminoacyltrnasynthetaseurzymesoptimizedbydeeplearningbehaveasaquasispecies