How Scientists Gave Ordinary Soil Bacteria the Ability to Fix Nitrogen from the Air

In a greenhouse at Washington State University Vancouver, two strains of soil bacteria are being pressed together on a nutrient plate, swapping DNA at the rate that evolution normally manages only over millions of years. One bacterium carries a package of genes roughly 500 kilobases long, a genetic island dedicated entirely to a single metabolic trick: pulling nitrogen out of thin air and delivering it, usable, into a plant root. The other strain has never done anything like this. Within forty hours, some of the second strain’s descendants can. It’s a compressed replay of one of agriculture’s founding evolutionary events, running in a petri dish in southwest Washington state.

The experiment, published in Current Biology, is the culmination of roughly ten years of work in the lab of Stephanie Porter, and it takes on a particular urgency right now, when fertilizer prices have been erratic and farmers in many parts of the world are paying more for synthetic nitrogen than the crop can justify.

Nitrogen is the limiting nutrient in most agricultural soils. Plants need it desperately to build proteins and chlorophyll, but the atmosphere, which is about 78 percent nitrogen gas, might as well be empty for most of them: the molecule is chemically inert, triple-bonded, practically impenetrable. The industrial answer to this problem, the Haber-Bosch process, uses extreme heat and pressure to crack those bonds and synthesise ammonia, and it has fed billions of people since the early twentieth century, but it consumes around 1 to 2 percent of global energy in the process. Legumes, by contrast, solved the problem biologically. They host specialised bacteria called rhizobia inside root nodules, where the bacteria convert atmospheric nitrogen into ammonia for the plant. The plant, in return, feeds the bacteria with sugars. A mutualism that’s been ticking along for perhaps 60 million years.

What Porter’s team has now demonstrated is that the genetic machinery for this trick is, in principle, portable. It can be lifted out of its native bacterial chromosome and dropped into a stranger’s genome, with results that surprised even the researchers themselves.

The mechanism relies on a mobile genetic element called a symbiosis island (or SI), a cluster of genes encoding everything a bacterium needs to infect a plant root and start fixing nitrogen. These islands are known to jump between bacterial strains via a process called conjugation, essentially a form of bacterial mating. What wasn’t clear was how reliably they’d function in an unfamiliar genomic background, or whether the recipient bacteria would become mutualists, parasites, or something more ambiguous. Porter and her postdoctoral scholar Angeliqua Montoya, who led the study, set up 180 pairwise matings between 30 donor strains and 6 recipient strains, all drawn from wild Mesorhizobium populations across the western United States. They then grew the resulting converted bacteria alongside host plants and watched.

“We developed a new way to successfully move a big cluster of genes that allow the bacteria to harvest nitrogen and colonize host plants into new bacteria that could not do these things at all,” Porter said. “We can convert these regular bacteria into ones that are able to harvest nitrogen to fertilize plants in one single step.”

Transfer succeeded in 80 percent of the pairings, which is a remarkably high rate for this kind of horizontal gene transfer experiment; nobody had previously attempted anything close to this many pairwise matings in wild rhizobia populations, and the sheer breadth of the dataset let the team ask questions that smaller studies simply couldn’t address. They found, for instance, that the donor strain’s genotype explained nearly half (48 percent) of the variation in how often transfer succeeded, while the recipient’s genotype and the interaction between the two each accounted for roughly a quarter. In other words: who hands over the island matters more than who receives it. Once the symbiosis island had moved into its new host, it could, occasionally, displace other genomic elements already sitting at its preferred docking site, a feature consistent with the idea that these islands have evolved, over millions of years, to spread promiscuously through bacterial populations.

The functional results were, by the researchers’ own account, not what theory predicted. Endosymbiosis is conventionally thought to begin badly, with the incoming symbiont exploiting its host before subsequent evolution nudges the relationship toward mutualism. That wasn’t what happened here. Two-thirds of the novel transconjugants established what’s called commensal relationships, neither harming nor benefiting the host plant; this is consistent with the island incurring low costs even in unfamiliar genomic territory. More striking, roughly a third became de novo nitrogen-fixing mutualists straightaway, conferring measurable increases in leaf nitrogen to host plants. The expectation that new symbionts start out as exploiters looks, at minimum, incomplete.

There were constraints. Phylogenetic distance mattered a lot for function: bacteria that were more closely related to the donor tended to fix more nitrogen after acquiring its island, probably because the SI had co-evolved with certain genomic features its closest kin share. And the soil ecotype of donor and recipient predicted functional outcomes too, hinting at epistasis between the symbiosis island and other mobile elements already resident in the bacterial chromosome.

“These challenges emphasize the importance of finding more natural pathways for getting nitrogen to crops,” Porter said, speaking of the fertilizer cost pressures that motivate the work. The eventual target is something far more ambitious than engineering model bacteria: the team wants to identify which specific gene variants make transfers most successful, then apply that knowledge to bacteria that actually colonise maize, wheat, or soybean. Porter has said that if you know which microbes typically live in corn or soybean roots, this approach should let you transfer the nitrogen-fixing capability into exactly those strains.

There’s a long way between a proof-of-concept greenhouse study and field-ready nitrogen-fixing cereals. The converted bacteria still fix less nitrogen than native rhizobia, and there are competition dynamics to sort out: an engineered inoculant has to actually win the colonisation race against resident soil microbes. But the symbiosis island has been jumping between bacteria for millions of years without human help. Knowing how to steer that jump, and to whom, is a start worth having.

https://doi.org/10.1016/j.cub.2026.04.071

Frequently Asked Questions

Why can’t wheat and corn just fix nitrogen the way beans do?

The ability to form root nodules with nitrogen-fixing rhizobia evolved in legumes over tens of millions of years, involving a complex molecular dialogue between plant and bacterium. Cereal crops like wheat and corn lack the plant-side machinery to initiate that conversation. Engineering the bacteria is one approach to the problem; getting the plant to respond correctly is a separate, harder challenge that researchers are also working on.

What is a symbiosis island and how does it move between bacteria?

A symbiosis island is a large cluster of genes, roughly 500 kilobases in some species, that encodes everything a bacterium needs to infect plant root cells and fix nitrogen. It sits in the bacterial chromosome but retains the ability to excise itself and transfer into a neighbouring bacterium through a process called conjugation, a kind of bacterial mating involving direct cell-to-cell contact. The island essentially carries its own transfer machinery, which may explain why it spreads so readily even across distantly related strains.

Does this mean we could have nitrogen-fixing wheat or corn soon?

Not soon, no. The Washington State team’s result is a proof of concept: they’ve shown the genetic machinery can move into unfamiliar bacteria and function, which is a genuine step forward. But making that work reliably in the bacteria that colonise crop roots, at field scale, under real soil conditions, while competing with resident microbes, involves layers of additional engineering and testing. Most researchers would put a practical crop application at least a decade away, possibly much longer.

Why would a bacterium share genes that give it a competitive advantage?

The symbiosis island is a somewhat selfish element: it benefits from spreading into as many bacterial lineages as possible, even at a metabolic cost to the individual bacterium carrying it. This tension between the island’s evolutionary interests and those of its host chromosome is part of what makes the system so dynamic. In environments where host legumes are abundant, the island’s benefits probably outweigh its costs; where compatible plants are rare, the cost of maintaining all that extra DNA may select for bacteria that resist acquiring it in the first place.