Protein pocketome mapped

Researchers have identified protein binding sites in eleven species from different kingdoms of life

Proteins perform their many different functions via physical interactions with other molecules, in particular, small molecules present in cells, such as metabolites. These interactions occur on distinct binding pockets on protein surfaces. Leveraging on the recent breakthroughs in computational structure predictions, scientists from the Max Planck Institute of Molecular Plant Physiolgy in Potsdam (Germany) now performed a systematic survey of all binding pockets - the “pocketome” - across eleven species from different kingdoms of life to explore the diversity and scaling behavior of number of binding pockets with the number of proteins found in the respective species.

To the point

• Protein structures from the AlphaFold database: The researchers identified nearly 100,000 potential binding sites on 220,000 proteins. The study covered eleven species, including humans, mice, and important crops.
• Understanding molecular interactions: The findings could promote the development of safer agrochemicals.
• Evolutionary trends: The study points to possible limits or limited need for structural and functional innovations.

Using over 220,000 AI-predicted protein structures from the AlphaFold database and sophisticated computational pocket detection tools, bioinformaticians Hanne Zillmer and Dirk Walther identified nearly 100,000 potential binding sites across the eleven species. They compared and clustered these sites within and across species, constructing a global "map" of binding sites and their associated features. Among the proteomes analyzed were those of human, mouse, yeast, the gut bacterium E.coli, and a nematode species, alongside key crop species such as rice and corn. “The availability of AI-based and confidently predicted protein structures as generated by AlphaFold now allows for an unprecedented breadth and depth of structural studies”, says PhD student Hanne Zillmer.

The study revealed a surprising evolutionary trend: as organisms evolved more complex proteomes - proteome: the entirety of all proteins in a species -, the diversity of their binding sites increased sub-linearly. In simpler terms, while the number of proteins increased with evolution, the number of distinct molecular interaction sites grew less than proportionally, suggesting possible limits of or need for structural and functional innovation.

Less interaction sites than expected

“We expected more proteins would mean proportionally more diversity of interaction sites”, says Walther, Professor of Bioinformatics, “but instead, evolution seems to either be constrained in designing altogether new binding sites, or, there is a limited need for radically new interaction modes.” “With this comprehensive view of the pocketomes of different species now available, a systematic investigation of molecular interactions between proteins and small molecules, and the understanding of species- or kingdom-specific binding modes is now greatly facilitated,” says Hanne Zillmer. 

Moreover, the research contributes an inter-species perspective that has been largely missing in the field to date. While most previous studies focused on biomedical applications and the druggability of binding sites, or differences between drug-protein vs. metabolite-protein binding events, this study emphasizes the evolutionary trends and structural diversity across the tree of life. Inter-species comparisons will, for example, also help design safer agrochemicals that specifically target pathogens but are inactive in pollinators or humans.