Machine learning–assisted directed evolution of plant Rubisco
11 affiliations
- 1Department of Chemistry, Massachusetts Institute of Technology, Cambridge, MA 02139, USA.
- 2Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA.
- 3Sarafan ChEM-H, Stanford University, Stanford, CA 94305, USA.
- 4Department of Biochemistry, Stanford University School of Medicine, Stanford, CA 94305, USA.
- 5Plant Science Division, Research School of Biology, The Australian National University, Canberra, ACT 2601, Australia.
- 6Whitehead Institute for Biomedical Research, Cambridge, MA 02142, USA.
- 7Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA.
- 8Howard Hughes Medical Institute, Whitehead Institute for Biomedical Research, Cambridge, MA 02142, USA.
- 9Ragon Institute of Massachusetts General Hospital, Harvard, and MIT, Cambridge, MA 02139, USA.
- 10Broad Institute of MIT and Harvard, Cambridge, MA 02142, USA.
- 11Koch Institute for Integrative Cancer Research at MIT, Cambridge, MA 02141, USA.
Abstract
Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) is foundational to life on Earth, catalyzing carbon dioxide (CO 2 ) fixation to generate biomass. However, Rubisco is a slow and inefficient enzyme that has proven challenging to engineer. We applied the structure-informed machine learning (ML) model ESM-IF1 to identify plausible amino acid sites in the large subunit of Nicotiana tabacum Rubisco to target for directed evolution. ML-assisted library design followed by selection in Rubisco-dependent Escherichia coli identified multiple enriched variants displaying improved catalytic efficiency. Several improved variants carried amino acid changes not found in the evolutionary lineage of plants, despite being assembly competent in plant chloroplasts, demonstrating that ML-assisted protein design can explore functional sequence space beyond what is observed from natural sequence diversity. Most prominently, the T391I substitution improved carboxylation rate by 29% and aerobic carboxylation efficiency by 43%. Our findings demonstrate the utility of ML-assisted evolution for engineering Rubisco with improved carboxylation efficiency and potential for enhancing crop productivity.
Related
- Advances in Machine Learning for Directed Evolution: A Five-Year Retrospective
- DNA Fragment Length Analysis Using Machine Learning Assisted Vibrational Spectroscopy.
- Machine-learning-assisted integration of single-cell and bulk transcriptomics characterizes glioblastoma ecosystem states and immune-associated programs
- Machine learning-assisted chemometric analysis of salivary ATR-FTIR spectra for the diagnosis of oral submucous fibrosis
- Directed Evolution of Threonine Aldolases for the Photobiocatalytic Synthesis of β-Branched-α-Tetrasubstituted and Other Noncanonical Amino Acids
BibTeX
@article{mcdonald2026machine,
title = {{Machine learning–assisted directed evolution of plant Rubisco}},
author = {Julie L. McDonald and Jiacheng Lin and Yunlong Zhao and Brian L. Hie and Rosemary Birch and Mary Gehring and Bryan D. Bryson and Spencer M. Whitney and Matthew D. Shoulders and Robert H. Wilson},
journal = {Science Advances},
year = {2026},
month = oct,
doi = {10.1126/sciadv.aeh3246},
url = {https://www.science.org/doi/10.1126/sciadv.aeh3246},
note = {CC BY},
}
RIS
TY - JOUR
TI - Machine learning–assisted directed evolution of plant Rubisco
AU - Julie L. McDonald
AU - Jiacheng Lin
AU - Yunlong Zhao
AU - Brian L. Hie
AU - Rosemary Birch
AU - Mary Gehring
AU - Bryan D. Bryson
AU - Spencer M. Whitney
AU - Matthew D. Shoulders
AU - Robert H. Wilson
PY - 2026
DA - 2026/10/09
JO - Science Advances
DO - 10.1126/sciadv.aeh3246
UR - https://www.science.org/doi/10.1126/sciadv.aeh3246
AB - Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) is foundational to life on Earth, catalyzing carbon dioxide (CO 2 ) fixation to generate biomass. However, Rubisco is a slow and inefficient enzyme that has proven challenging to engineer. We applied the structure-informed machine learning (ML) model ESM-IF1 to identify plausible amino acid sites in the large subunit of Nicotiana tabacum Rubisco to target for directed evolution. ML-assisted library design followed by selection in Rubisco-dependent Escherichia coli identified multiple enriched variants displaying improved catalytic efficiency. Several improved variants carried amino acid changes not found in the evolutionary lineage of plants, despite being assembly competent in plant chloroplasts, demonstrating that ML-assisted protein design can explore functional sequence space beyond what is observed from natural sequence diversity. Most prominently, the T391I substitution improved carboxylation rate by 29% and aerobic carboxylation efficiency by 43%. Our findings demonstrate the utility of ML-assisted evolution for engineering Rubisco with improved carboxylation efficiency and potential for enhancing crop productivity.
SN - 2375-2548
ER -