Scientists Reveal Novel Protein Structure Using AI-Enhanced Techniques

In a ground-breaking study published in the journal Structure (doi: 10.1016/j.str.2024.08.020), researchers at the University of Western Ontario have uncovered a unique protein structure that plays a crucial role in regulating calcium levels within mitochondria, the powerhouses of cells. The research, led by Dr. Qi-Tong (Tom) Lin, a recent PhD graduate from Dr. Peter Stathopulos's laboratory in the Department of Physiology and Pharmacology, combines artificial intelligence with traditional biophysical experimental methods to reveal new insights into cellular calcium regulation.   

                                                                                

The team focused on a protein called LETM1, which is deleted in Wolf-Hirschhorn syndrome, a rare genetic disorder. Using an innovative approach that merged AlphaFold AI predictions with nuclear magnetic resonance (NMR) spectroscopy, they discovered an unprecedented protein structure called the F-EF-hand domain. This domain functions as a two-way regulator of mitochondrial calcium levels, acting like a molecular switch that can either increase or decrease calcium concentrations within mitochondria.

"What makes this discovery particularly exciting is that this is a unique EF-hand and that we have integrated artificial intelligence programs such as AlphaFold that can be harnessed to improve and enrich experimental structure determination techniques," explains Lin. The F-EF-hand domain identified in LETM1 has a unique arrangement as a complete helix-loop-helix motif paired with a partial helix loop motif that hasn't been seen before in other calcium-binding proteins. The researchers demonstrated that mutations affecting calcium binding in this domain can significantly impact mitochondrial calcium levels. Enhanced calcium binding leads to increased mitochondrial calcium levels, while weakened binding results in decreased levels. This finding has important implications for understanding both normal cellular function and diseases involving mitochondrial dysfunction.


 

This study also showcases how artificial intelligence tools like AlphaFold can be effectively combined with experimental techniques to advance our understanding of protein structures and their functions. “AI is incredibly powerful, and we need to embrace it, we need to understand how it fits into our current methodology,” says Lin. The methodology developed by the team provides a new framework for future structural biology research. The findings could have significant implications for understanding and potentially treating conditions involving mitochondrial calcium regulation, including Wolf-Hirschhorn syndrome and other disorders affecting cellular energy production.