Deep learning-based characterization and redesign of major potato tuber storage protein

被引：5

作者：

Luo, Xuming ^{[1
]}

Cao, Lijuan ^{[2
]}

Yu, Langhua ^{[2
]}

Gao, Meng ^{[2
]}

Ai, Ju ^{[2
]}

Gao, Dongli ^{[2
]}

Zhang, Xiaopeng ^{[1
,3
]}

Lucas, William John ^{[4
]}

Huang, Sanwen ^{[1
,5
]}

Xu, Jianfei ^{[6
]}

Shang, Yi ^{[2
]}

机构：

[1] Chinese Acad Agr Sci, Agr Genom Inst Shenzhen, Shenzhen Branch, State Key Lab Trop Crop Breeding,Guangdong Lab Lin, Shenzhen 518120, Guangdong, Peoples R China

[2] Yunnan Normal Univ, CAAS YNNU YINMORE Joint Acad Potato Sci, Yunnan Key Lab Potato Biol, Kunming 650500, Yunnan, Peoples R China

[3] Chinese Acad Agr Sci, Inst Vegetables & Flowers, Key Lab Biol & Genet Improvement Hort Crops, Sino Dutch Joint Lab Hort Genom,Minist Agr, Beijing 100081, Peoples R China

[4] Univ Calif Davis, Coll Biol Sci, Dept Plant Biol, Davis, CA 95616 USA

[5] Chinese Acad Trop Agr Sci, State Key Lab Trop Crop Breeding, Haikou 571101, Peoples R China

[6] Chinese Acad Agr Sci, Inst Vegetables & Flowers, State Key Lab Vegetable Biobreeding, Beijing 100081, Peoples R China

来源：

FOOD CHEMISTRY | 2024年 / 443卷

基金：

中国国家自然科学基金;

关键词：

Potato; Patatin; Deep learning; Protein design; Disulfide bond; PREDICTION; PATATIN;

D O I：

10.1016/j.foodchem.2024.138556

中图分类号：

O69 [应用化学];

学科分类号：

081704 ;

摘要：

Potato is one of the most important crops worldwide, to feed a fast-growing population. In addition to providing energy, fiber, vitamins, and minerals, potato storage proteins are considered as one of the most valuable sources of non -animal proteins due to their high essential amino acid (EAA) index. However, low tuber protein content and limited knowledge about potato storage proteins restrict their widespread utilization in the food industry. Here, we report a proof -of -concept study, using deep learning -based protein design tools, to characterize the biological and chemical characteristics of patatins, the major potato storage proteins. This knowledge was then employed to design multiple cysteines on the patatin surface to build polymers linked by disulfide bonds, which significantly improved viscidity and nutrient of potato flour dough. Our study shows that deep learning -based protein design strategies are efficient to characterize and to create novel proteins for future food sources.

引用

页数：11

共 38 条

[1] Food: use artificial intelligence to create new proteins [J].