Plants figured out hormones long before we did. About half a billion years ago, as green things crawled out of the ocean and onto land, they started cobbling together a signaling system that would eventually determine how tall your wheat grows, how much rice you harvest, and whether crops can survive on less fertilizer.

Gibberellins - GAs if you're into the whole brevity thing - are plant hormones that tell stems to grow, seeds to germinate, and flowers to bloom. A new review in Advanced Science traces how this system evolved piece by piece, like someone building IKEA furniture without instructions, except the furniture took hundreds of millions of years to assemble and accidentally saved a billion people from starvation.

Evolution: The Original Tinkerer

Here's the wild part: the GA signaling pathway wasn't designed. It was duct-taped together across geological time.

First came DELLA proteins, molecular brakes that slow down plant growth. These showed up in early land plants, probably helping them survive the harsh, dry conditions of terrestrial life. Think of DELLAs as the cautious friend who says "maybe we shouldn't" at every opportunity.

Then, much later, plants evolved the ability to make gibberellins themselves. But having a growth hormone without a way to detect it is like having a fire alarm with no sensor. Enter GID1, a receptor protein that finally connected GA production to DELLA destruction. When gibberellins bind to GID1, it tags DELLAs for removal, releasing the growth brake.

This GA-GID1-DELLA module became the core circuitry for hormone-controlled growth. And it worked so well that flowering plants ran with it, duplicating and diversifying these components into elaborate regulatory networks.

The Green Revolution's Dirty Secret

In the 1960s and 70s, plant breeders discovered that messing with GA signaling produced shorter, sturdier wheat and rice plants. These semi-dwarf varieties didn't fall over under the weight of their own grain, dramatically increasing yields. Norman Borlaug won a Nobel Prize. A billion people didn't starve. Pretty solid outcome.

But there was a catch.

Those same GA-related mutations that make plants shorter also make them worse at using nitrogen. Farmers compensated by dumping more fertilizer on fields, creating a cascade of environmental problems - nitrate runoff, dead zones in waterways, greenhouse gas emissions from fertilizer production.

The review by Fan and colleagues lays out this trade-off with uncomfortable clarity: the genetic changes that saved billions of lives in the 20th century are now part of the problem in the 21st [1]. It's like discovering your miracle drug has side effects that only show up decades later.

Can AI Redesign Plant Hormones?

This is where things get interesting for the machine learning crowd. The authors propose using protein language models and AI-driven design to rationally rewire GA signaling - essentially debugging half a billion years of evolutionary code.

The idea isn't as crazy as it sounds. Recent work has mapped how GA signaling connects to nitrogen metabolism, stress responses, and other pathways [2]. Researchers have identified specific regulatory nodes that could be tweaked to decouple plant height from nitrogen use efficiency.

Protein language models trained on millions of sequences can now predict how mutations affect protein function [3]. Combine that with structural biology data from cryo-EM and X-ray crystallography, and you've got the ingredients for targeted protein engineering. Instead of screening thousands of random mutants, you could computationally design DELLA variants that respond differently to GA, or GID1 receptors with altered binding properties.

AlphaFold and its descendants have already transformed how we think about protein structure [4]. Applying similar approaches to signaling pathways - predicting not just structure but dynamic behavior and interaction networks - could let us redesign plant hormones the way engineers redesign circuits.

The Bigger Picture

What makes this review valuable isn't just the science; it's the framing. By treating GA signaling as an evolutionary assembly rather than a fixed system, the authors open the door to thinking about plant hormones as hackable firmware.

Climate change is already stressing crops worldwide. We need plants that grow efficiently with less water, less fertilizer, and more heat tolerance. The GA pathway touches all of these - it's connected to drought responses, nutrient sensing, and developmental timing.

The tools are converging: comparative genomics to identify natural variation, structural biology to understand molecular mechanisms, and AI to predict and design new protein functions. The question is whether we can move fast enough to deploy these approaches before climate impacts outpace our ability to adapt crops.

If evolution could assemble this system through random tinkering across geological epochs, maybe directed intelligence can improve it in decades. That's the bet this research is making.

And honestly, given what's at stake, it seems worth taking.

References

1. Fan W, Chen X, Li D, Gao X, Fu X. Evolutionary Assembly and Future Design of Gibberellin Signaling. Adv Sci. 2025. DOI: 10.1002/advs.75049. PMID: 41902468

2. Wu K, et al. Enhanced sustainable green revolution yield via nitrogen-responsive chromatin modulation in rice. Science. 2020;367(6478):eaaz2046. DOI: 10.1126/science.aaz2046

3. Lin Z, et al. Evolutionary-scale prediction of atomic-level protein structure with a language model. Science. 2023;379(6637):1123-1130. DOI: 10.1126/science.ade2574

4. Jumper J, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583-589. DOI: 10.1038/s41586-021-03819-2

Disclaimer: This blog post is a simplified summary of published research for educational purposes. The accompanying illustration is artistic and does not depict actual model architectures, data, or experimental results. Always refer to the original paper for technical details.