Explanatory AI helps to decipher the substrate specificity of γ-secretase enzymes

Using artificial intelligence, researchers show how gamma-secretases recognize substrates and substrates, an important advance in translation research.

The γ-secretase enzyme is capable of cleaving more than 150 different membrane proteins. These include amyloid precursor proteins, where deposits typical of Alzheimer's disease form, and Notch1 proteins, which play an important role in cell communication and carcinogenesis. However, for a long time, it was unclear how γ-secretase recognizes target proteins. Many proteases identify substrates by their distinctive amino acid sequences, whereas γ-secretases do not.

The interdisciplinary teams at LMU's Biomedical Centre, the Institute of Technology Munich (TUM) and the Centre for Neurodegenerative Diseases in Germany (DZNE) have now been able to clarify the mechanism details. Researchers show that enzyme substrates have complex physicochemical profiles that determine their recognition and cleavage.

New techniques will show hidden features

The team has developed a new technique called Comparative Physics and Chemical Profiling (CPP). This allows you to compare the physicochemical properties of known substrates with reference proteins and identify distinctive patterns. In combination with explanatory artificial intelligence (XAI), the team was also able to see distinctive features of the gamma-secretase substrate.

“The gamma-secretase substrates have specific physicochemical profiles that span the entire transmembrane domain and the entire adjacent sequence region,” explains Professor Harald Steiner (LMU and DZNE), who led the study along with Dmitrij Frishman (TUM). Researchers, particularly close to the cleavage sites of the substrate, found that substrates could develop expanded conformations as alternatives to helical structures.

“We wanted to understand what actually defines a board, not just generating black box predictions,” adds Dr. Stephan Breimann, who played a major role in developing the CPP method. “The use of explainable AI gave us this transparency accurately.”

Research and application perspectives

Using the CPP method, researchers were also able to identify several previously unknown substrates of the enzyme. This includes proteins that play an important role in immunoregulation and cancer development.

The authors of this study believe that their findings are far beyond γ-secretase.

Here we have a new approach to deciphering other proteases, or, for example, receptor sequence, structure, and function interactions. ”

Professor Harald Steiner, LMU, DZNE

In the long term, results may also contribute to the development of therapeutically relevant compounds, such as small molecule drugs, peptides, or antibodies with enhanced specificity.