Deep Learning-Initiated Closed-Loop Engineering of Brazzein for Enhanced Sweet-Taste Receptor Activation

Abstract Sweet-protein engineering is constrained by low-throughput functional assays. We developed VenusSweet, a deep learning-initiated workflow combining ProtSSN zero-shot prioritization, supervised LinearSVC ranking, and sweet-taste receptor assays. Four design-build-test-learn rounds evaluated 170 brazzein variants and identified VenusBraz. Across three matched experiments, VenusBraz showed a mean 36.9-fold increase in receptor-activation potency over wild type. It also outperformed H31R/E36D/E41A in side-by-side assays. A separate double-blind evaluation with 15 participants supported greater perceived sweetness; the receptor-derived fold change is not a sensory sweetness ratio. VenusBraz retained receptor activity after thermal treatment and storage for 30 days under acidic or alkaline conditions. No significant protein × pH interaction was detected in two end point experiments. Molecular dynamics suggested altered receptor-associated conformational sampling. VenusBraz warrants further formulation and safety evaluation.

Authors

Institutions

Publication Details

Journal
Journal of Agricultural and Food Chemistry
Published
2026-11-13
DOI
https://doi.org/10.1021/acs.jafc.6c11441
Primary Topic
Biochemical Analysis and Sensing Techniques
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Deep Learning-Initiated Closed-Loop Engineering of Brazzein for Enhanced Sweet-Taste Receptor Activation

Liang Hong, Banghao Wu, Lirong Zheng, Yunjie Liu et al.
2 links
Journal of Agricultural and Food Chemistry
Biochemical Analysis and Sensing Techniques
article

Deep Learning-Initiated Closed-Loop Engineering of Brazzein for Enhanced Sweet-Taste Receptor Activation

Liang Hong, Banghao Wu, Lirong Zheng, Yunjie Liu, Yang Tan, Yutong Sun, Bozitao Zhong, Pan Tan
article en

Abstract

Abstract Sweet-protein engineering is constrained by low-throughput functional assays. We developed VenusSweet, a deep learning-initiated workflow combining ProtSSN zero-shot prioritization, supervised LinearSVC ranking, and sweet-taste receptor assays. Four design-build-test-learn rounds evaluated 170 brazzein variants and identified VenusBraz. Across three matched experiments, VenusBraz showed a mean 36.9-fold increase in receptor-activation potency over wild type. It also outperformed H31R/E36D/E41A in side-by-side assays. A separate double-blind evaluation with 15 participants supported greater perceived sweetness; the receptor-derived fold change is not a sensory sweetness ratio. VenusBraz retained receptor activity after thermal treatment and storage for 30 days under acidic or alkaline conditions. No significant protein × pH interaction was detected in two end point experiments. Molecular dynamics suggested altered receptor-associated conformational sampling. VenusBraz warrants further formulation and safety evaluation.

Journal of Agricultural and Food Chemistry
Shanghai Jiao Tong University (CN), Shanghai Artificial Intelligence Laboratory (CN), Shanghai Innovation Institute (CN)
Science and Technology Commission of Shanghai Municipality, Shanghai Jiao Tong University, Shanghai Municipal Commission of Economy and Informatization, National Key Research and Development Program of China, Shanghai Artificial Intelligence Laboratory
Zero hunger
Biochemical Analysis and Sensing Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon