Engineering more resilient crops for a warming climate
Google DeepMind
Scientists used AlphaFold to figure out why a key photosynthesis enzyme falls apart in heat, then rebuilt it to survive up to 65°C. That fix could mean crops that don't collapse during heatwaves and droughts.
Based on reporting by Google DeepMind — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Photosynthesis has a weak link, and it's been mostly invisible to scientists until now. Berkley Walker, a plant biologist at Michigan State University, has spent years staring at an enzyme called glycerate kinase, or GLYK, which helps plants recycle carbon while turning sunlight into glucose. Nobody had ever managed to pin down its exact 3D structure through lab experiments. That gap in knowledge mattered more than it sounds, because without seeing the enzyme's shape, nobody could figure out exactly why it breaks down when temperatures climb.
Walker's team turned to AlphaFold to fill that gap, predicting GLYK's structure not just in ordinary plants but also in a heat-tolerant algae species that lives in volcanic hot springs. Running those predicted structures through molecular simulations let researchers essentially watch the enzymes wiggle and warp as temperatures rose. And what they found was oddly specific: three flexible loops in the plant version of GLYK lost their shape under heat stress, while the algae's version stayed put.
That single observation opened a path forward. Walker's lab started swapping the wobbly loops from the plant enzyme with the sturdier ones borrowed from the heat-loving algae, building a series of hybrid enzymes. One combination worked far better than expected, holding its structure at temperatures up to 65 degrees Celsius — well beyond what a heatwave-stressed field would ever throw at a crop.
The bigger point here isn't really about GLYK specifically. It's about how researchers are starting to treat enzyme structure prediction as a design tool rather than just a research curiosity. Walker himself credits AlphaFold with revealing structural details that lab experiments alone couldn't touch, which let his team target exactly the right sections for modification instead of guessing.
Staple crops like wheat, rice, and soy already show yield losses tied to rising temperatures, and that trend isn't reversing anytime soon. Engineering heat-stable versions of enzymes like GLYK won't fix climate change, but it could buy plants — and the people who depend on them — some breathing room as growing seasons get hotter and less predictable.
My take — AI-written commentary, not fact-checked reporting
This is exactly the kind of AI application that deserves more attention than another chatbot benchmark: quiet, unglamorous protein engineering that could actually stabilize food supplies as the planet warms. I'd rather see DeepMind's compute budget go toward a hundred more projects like this than another flashy demo, and I say that as someone generally annoyed by AI hype cycles. Open tools like AlphaFold doing real agricultural science is the strongest argument for why this technology matters at all.
Read more about this at: Google DeepMind
