A Nigerian researcher, Muinat Zubair, has developed an innovative method for predicting protein structure and protein-protein interactions by combining experimental and computational approaches.

Zubair made this discovery during her time as a graduate teaching and research assistant at Tennessee Technological University, USA, where she earned her MSc in chemistry. Speaking about her work, she explained that her research contributes to solving the protein folding problem—a challenge that has perplexed scientists for over 50 years.

“I used artificial intelligence software to predict the full-length structures and interaction between two mammalian enzymes,” Zubair said. These proteins are part of the larger mitogen-activated protein kinase (MAPK) pathway, which is activated in response to cellular stress. According to her, malfunction in this pathway worsens symptoms of several neurodegenerative diseases, including Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis (ALS).

“Understanding the structures of these proteins and how they bind and activate each other will help us develop therapeutic targets to manage these diseases and improve outcomes for patients and their families,” she added.

Neurodegenerative diseases are increasingly becoming a global pandemic, affecting millions of people worldwide. Zubair highlighted that this number is expected to rise as the global population of older adults grows.

The steady-state kinetics of these proteins had never been studied prior to her research. Supervised by Dr. Xuanzhi Zhan, she determined their apparent Michaelis-Menten parameters through kinase assays and Western blot quantification.

In addition, Zubair predicted the full-length structures and protein-protein interactions using AlphaFold2 and AlphaFold-Multimer—two cutting-edge artificial intelligence tools designed to elucidate the structure and behavior of proteins and other macromolecules.

“Previously, X-ray diffraction studies had only solved the structures of the kinase domains of these proteins,” she explained. “However, prior research showed that other parts of the enzymes also contribute to binding and activation, emphasizing the need to predict the entire structure.”

To validate her findings, Zubair used additional computational methods. “I applied molecular dynamics simulations to test and confirm protein stability and performed protein frustratometer analysis, contact mapping, and other assessments on the interface between the docked proteins. The protein structures remained stable under simulated physiological conditions,” she noted.

Beyond pioneering kinetic data on the MAPK pathway, Zubair emphasized that the computational methods she employed could be applied broadly to study other proteins. Nevertheless, she acknowledges that computational results require validation through wet-laboratory experiments.

“That is the next stage of the project,” she said. “It involves site-directed mutagenesis and studying the other two isoforms of the enzyme to further validate our results.”

Originally from Iganna in Oyo State, Nigeria, Zubair grew up in Ibadan and earned her first degree in biochemistry from Ladoke Akintola University of Technology (LAUTECH), Nigeria. She is currently pursuing a doctoral degree in the highly selective Purdue University Interdisciplinary Life Science program (PULSe) and was awarded the prestigious Lynn Fellowship to support her PhD studies.

Looking ahead, Zubair expresses a strong commitment to advancing human health, particularly in the field of neurodegenerative diseases. “I find so much fulfillment knowing that I could potentially discover a druggable target or help improve our understanding of these diseases,” she said.
https://tribuneonlineng.com/researcher-develops-method-for-protein-protein-structure-prediction/