Deep learning designs proteins against deadly snake venom

(Press-News.org) New proteins not found in nature have now been designed to counteract certain highly poisonous components of snake venom. The deep learning, computational methods for developing these toxin-neutralizing proteins offer hope for creating safer, more cost-effective and more readily available therapeutics than those currently in use.  

Each year more than 2 million people suffer snakebites. More than 100,000 of them die, according to the World Health Organization, and 300,000 suffer severe complications and lasting disability from limb deformity, amputation or other aftereffects. Sub-Saharan Africa, South Asia, Papua New Guinea, and Latin America are among the places where poisonous snakebites pose the greatest public health concern.  

The computational biology effort to discover better antivenom therapeutics, led by scientists at the UW Medicine Institute for Protein Design and the Technical University of Denmark, is reported today, Jan. 15,  in Nature.  

The lead author of the paper is Susana Vazquez Torres of the Department of Biochemistry at the UW School of Medicine and the UW Graduate Program in Biological Physics. Her hometown is Querétaro, Mexico, which is located near viper and rattlesnake habitats. Her professional goal is to invent new drugs for neglected diseases and injuries, including snakebites.  

Her research team, which also included international experts in snakebite research, drugs and diagnostics, and tropical medicine from the United Kingdom and Denmark, concentrated their attention on finding ways to neutralize venom gathered from certain elapids. Elapids are a large group of poisonous snakes, among them cobras and mambas, that live in the tropics and subtropics. 

Most elapid species have two small fangs shaped like shallow needles. During a tenacious bite, the fangs can inject venom from glands at the back of the snake’s jaw. Among the venom’s components are potentially lethal three-finger toxins. These chemicals damage bodily tissues by killing cells. More seriously, by interrupting signals between nerves and muscles, three-finger toxins can cause paralysis and death.  

At present, venomous snakebites from elapids are treated with antibodies taken from the plasma of animals that have been immunized against the snake toxin. Producing the antibodies is costly, and they have limited effectiveness against three-finger toxins. This treatment can also have serious side effects, including causing the patient to go into shock or respiratory distress.  

“Efforts to try to develop new drugs have been slow and laborious,” noted Vazquez Torres. 

The researchers used deep learning computational methods to try to speed the discovery of better treatments. They created new proteins that interfered with the neurotoxic and cell-destroying properties of the three-finger toxin chemicals by binding with them. 

Through experimental screening, the scientists obtained designs that generated proteins with thermal stability and high binding affinity. The actual synthesized proteins were almost a complete match at the atomic level with the deep-learning computer design.  

In lab dishes, the designed proteins effectively neutralized all three of the subfamilies of three-finger toxins tested. When given to mice, the designed proteins protected the animals from what could have been a lethal neurotoxin exposure.  

Designed proteins have key advantages. They could be manufactured with consistent quality through recombinant DNA technologies instead of by immunizing animals. (Recombinant DNA technologies in this case refer to the lab methods the scientists employed to take a computationally designed blueprint for a new protein and synthesize that protein.)   

Also, the new proteins designed against snake toxins are small, compared to antibodies. Their smaller size might allow for greater penetration into tissues to quickly counteract the toxins and reduce damage.  

In addition to opening new avenues to develop antivenoms, the researchers think computational design methods could be used to develop other antidotes. Such methods also might be used to discover medications for undertreated illnesses that affect countries with significantly limited scientific research resources. 

“Computational design methodology could substantially reduce the costs and resource requirements for development of therapies for neglected tropical diseases,” the researchers noted.  

The senior researchers on the project to design protein treatments for elapid snakebites were Timothy J. Perkins at the Technical University of Denmark and David Baker of the UW Medicine Institute for Protein Design and the Howard Hughes Medical Institute. Baker is a professor of biochemistry at the UW School of Medicine.

The University of Washington has submitted a provisional U.S. patent application for the design and composition of the proteins created in this study. 

 

 

END