Combining Deep Learning and Structural Modeling to Identify Potential Acetylcholinesterase Inhibitors from Hericium...

Area of Science:

• Computational chemistry
• Neuroscience
• Pharmacology

Background:

• Alzheimer's disease (AD) affects over 50 million globally, with current treatments like acetylcholinesterase (AChE) inhibitors causing side effects.
• Natural compounds are increasingly explored for AD treatment due to potentially fewer adverse effects.

Purpose of the Study:

• To computationally discover novel compounds from medicinal mushrooms that can inhibit AChE activity for potential AD treatment.
• To identify specific mushroom-derived compounds with high binding affinity to AChE.

Main Methods:

• Development of a deep neural network for rapid screening of compounds against AChE inhibition.
• Application of deep learning and molecular docking to screen the BACMUSHBASE database.
• Analysis of binding energy profiles of identified compounds against AChE.

Main Results:

• Five promising AChE inhibitory compounds were identified from the BACMUSHBASE database.
• Erinacerin A and hericenone B from *Hericium erinaceus* (lion's mane mushroom) were selected based on binding energy.
• These compounds exhibited binding energy profiles comparable to established AD drugs (donepezil, galanthamine).

Conclusions:

• Erinacerin A and hericenone B are potential candidates for developing new AD medications with improved side effect profiles.
• This study highlights the utility of computational approaches in discovering natural product-based therapeutics.
• Further research and development are warranted for these lion's mane mushroom-derived compounds.