De novo designed proteins neutralize lethal snake venom toxins

Area of Science:

• Biochemistry
• Toxicology
• Computational Biology

Background:

• Snakebite envenoming is a neglected tropical disease causing over 100,000 deaths annually.
• Three-finger toxins (3FTx) are key venom components responsible for severe pathologies like neurotoxicity and tissue damage.
• Current antivenom treatments are costly and have limited efficacy against 3FTx.

Purpose of the Study:

• To utilize deep learning for *de novo* protein design targeting 3FTx.
• To develop novel protein therapeutics for snakebite envenoming.
• To create a cost-effective and accessible alternative to traditional antivenoms.

Main Methods:

• Application of deep learning algorithms for *de novo* protein design.
• Computational modeling and *in vitro* experimental validation of designed proteins.
• Assessment of protein binding affinity, thermal stability, and neutralizing capacity against 3FTx.

Main Results:

• Designed proteins demonstrated high thermal stability and binding affinity.
• Near-atomic level agreement between computational models and experimental structures.
• Effective *in vitro* neutralization of all three 3FTx sub-families.
• Protection of mice against lethal neurotoxin challenge.

Conclusions:

• Deep learning-designed proteins are potent neutralizers of 3FTx.
• These proteins represent a foundation for next-generation, cost-effective antivenoms.
• The computational design methodology can accelerate therapeutic discovery for neglected diseases.