Among the vast array of chemical compounds that plants generate, some may offer solutions for treating human health issues. However, replicating these intricate, naturally occurring molecules in laboratories typically involves a lengthy and laborious trial-and-error method. Recently, chemists have demonstrated that new computational tools can expedite this process and make it more efficient.
Among the vast array of chemical compounds that plants generate, some may offer solutions for treating human health issues. However, replicating these intricate, naturally occurring molecules in laboratories typically involves a lengthy and laborious trial-and-error method.
Researchers at Scripps Research have illustrated how cutting-edge computational approaches can aid in the quicker and more efficient creation of complex natural compounds. Their method, detailed in Nature on December 23, 2024, successfully synthesized 25 different picrotoxanes, which are plant-derived compounds that have the potential to influence brain pathways.
“Manipulating these intricate plant compounds for drug design has been extremely challenging,” explains Ryan Shenvi, PhD, the study’s senior author and a professor at Scripps Research. “The integration of virtual predictions with hands-on experimentation represents a significant change in our molecule design and construction process.”
Picrotoxanes, located in the seeds of specific Asian and Indian shrubs, are known to interact with the mammalian nervous system; they attach to the same brain receptors that anxiety and sleep medications like Valium target. Some cultures have historically utilized these compounds as pesticides or to catch fish. Due to their ability to be consumed and their potential effects on brain function, researchers like Shenvi are intrigued about their possible therapeutic uses. Unfortunately, scientists have only been able to synthesize a few picrotoxanes in laboratories, making it challenging to analyze and manipulate them.
“Like numerous other plant metabolites, picrotoxanes have a complex atomic arrangement that makes their behavior difficult to predict,” states Shenvi. “We couldn’t assume that a method that worked for synthesizing one picrotoxane would apply to another, even if they appeared nearly identical.”
Facing difficulties in synthesizing picrotoxanes, Shenvi and graduate student Chunyu Li turned to advanced computer modeling to discover innovative ways to generate picrotoxanes from basic chemical components. They first created a virtual library of potential intermediate compounds that could emerge during the synthesis of picrotoxanes. They then employed a model called Density Functional Theory (DFT) to evaluate how these intermediates behaved, identifying those that were likely to succeed quickly and lead to neuroactive compounds.
When they tested five suggested picrotoxane synthesis pathways from their modeling — three expected to succeed and two expected to fail — all five predictions were accurate.
“DFT is typically used afterward to explain experimental data and the mechanics of chemical reactions, so I was initially skeptical about its predictive capability,” remarks Shenvi. “I was quite surprised to see how effective it was.”
Nonetheless, DFT remains somewhat time-consuming for evaluating each potential intermediate. Shenvi and Li aimed to enhance their approach and accelerate the process of creating more picrotoxanes. They implemented a pattern-recognition technology akin to that used in many contemporary artificial intelligence (AI) applications to detect patterns in the DFT output. This enabled them to develop a new statistical model that could predict the success of reactions much quicker. With this model, they pinpointed synthesis techniques for 25 picrotoxanes, confirming their functionality in the lab.
“This approach didn’t just allow us to create picrotoxanes,” Li says. “It opens the door for chemists to tackle other complex synthesis issues.”
Shenvi indicates that the laboratory is already utilizing this approach on additional challenges. They also plan to continue exploring the 25 picrotoxanes they can now produce to investigate their effects on mammalian biology.
This research received funding from the National Institutes of Health (GM122606) and a Skaggs Graduate School of Chemical and Biological Sciences Dale Boger Endowed Graduate Fellowship.
