RNA interference sprays: the pesticides that switch a pest's genes off

The first time you hear about a pesticide that works by "silencing a gene", it sounds like science fiction with a farm-supply catalogue attached. The reality is more modest and, in places, more interesting. A sprayed RNA product leaves the crop's DNA alone, adds a kind of molecule that has always been present in everything we eat, and kills its target by a mechanism taught in second-year biology. It also raises questions about specificity, resistance and regulation that pesticide rules were never written to answer.

Below are the questions I am asked most often about these products, with answers that try to stay on the right side of what the papers actually show.

What is RNA interference, in one paragraph?

Cells use messenger RNA as working copies of genes. The DNA stays in the nucleus; the mRNA travels to the ribosomes and is translated into protein. RNA interference is a defence and housekeeping system that destroys specific mRNAs before they can be read. It is set off by double-stranded RNA (dsRNA), which cells treat as suspicious because many viruses make it when they replicate. An enzyme called Dicer chops the dsRNA into short pieces of roughly 21 to 24 nucleotides. One strand of each piece is loaded into a protein complex built around an Argonaute protein, which then pairs with any mRNA carrying a matching sequence and cuts it. No mRNA, no protein. The gene is still there. It has simply gone quiet.

Andrew Fire and Craig Mello worked out the mechanism in nematode worms in 1998 and shared a Nobel Prize for it in 2006. Plant scientists had seen the effect years earlier without knowing what it was. The classic case is the petunias engineered around 1990 to have deeper purple flowers, which came out white or streaked instead: the extra gene copies had triggered silencing of both the new gene and the plant's own pigment gene.

How do you turn that into a pesticide?

You choose a gene the pest cannot live without, design a stretch of dsRNA that matches part of it, and get the pest to swallow it. If the insect's gut cells take up the dsRNA and its own RNAi machinery processes it, the essential protein disappears. The insect stops feeding and dies, usually over several days rather than the minutes or hours of a nerve poison.

There are two ways to deliver it. One is to have the crop produce the dsRNA itself by inserting a gene that encodes it. That is a genetically modified plant, regulated as one. The other is to manufacture dsRNA in a factory and spray it on the leaves, where the pest eats it. Researchers call this spray-induced gene silencing, and the plant's genome is untouched.

Has any of this actually been approved?

Yes, both kinds, in the United States.

The engineered route came first. In 2017 the US Environmental Protection Agency registered a maize trait producing a dsRNA aimed at a gene called Snf7 in the western corn rootworm, a beetle larva that eats maize roots. It is sold stacked with Bt insecticidal proteins rather than on its own, partly so that the rootworm has to beat two mechanisms at once.

The sprayed route followed at the end of 2023, when the EPA registered a product whose active ingredient, ledprona, was developed by the US company GreenLight Biosciences against the Colorado potato beetle. The dsRNA targets a gene for part of the proteasome, the machinery cells use to break down worn-out proteins. The beetle was a sensible first target. It has a long record of evolving resistance to almost every insecticide used against it, and beetles as a group are unusually sensitive to ingested dsRNA.

GreenLight has also worked on a dsRNA against the varroa mite, the parasite behind so many honeybee colony losses, fed to hives in sugar syrup. Beyond those, the company and academic pipeline includes sprays against fungal diseases such as grey mould, against plant viruses, and against weeds that have become resistant to herbicides. Most of that is at the field-trial stage or earlier. Any claim that RNA sprays are about to replace chemical pesticides wholesale is running well ahead of the registrations.

Why beetles and not caterpillars?

Because insects differ enormously in how well RNAi works when the dsRNA comes from outside, and nobody fully understands why. Beetles tend to absorb dsRNA from the gut efficiently and amplify the response. Moths and butterflies, whose caterpillars include many of the worst crop pests, are notoriously hard to silence: their gut fluids are rich in enzymes that degrade dsRNA, and what survives is poorly taken up into cells. Aphids and other sap-suckers sit somewhere in between, and are awkward to dose anyway, since they drink from inside the plant rather than chewing its surface.

That matters for the business case. The products that work well are concentrated on a fairly narrow band of pests, and the pests that cost farmers the most are not always in it.

Isn't naked RNA too fragile to spray on a field?

It is fragile. On a leaf, unprotected dsRNA is broken down by ultraviolet light, washed off by rain and eaten by microbes. In soil it typically lasts days rather than weeks. That short life is a selling point with regulators, since the active ingredient does not linger, and a problem for farmers, who want protection that outlasts the next shower.

A lot of current research is therefore about formulation. One widely cited approach, published by a University of Queensland group in Nature Plants in 2017, loaded dsRNA onto clay nanosheets that released it slowly and stretched protection against a plant virus from days to weeks. Others use lipids, polymers or peptides as carriers. Each carrier solves one problem and creates another, because the formulation now has its own environmental fate to assess. We have looked at the regulatory gaps around nanomaterials in food and packaging; a nano-clay coating on a field spray belongs in the same conversation.

Cost was the other barrier. Making dsRNA used to be a laboratory luxury. Over roughly fifteen years, cell-free synthesis and fermentation-based production have brought the price down by orders of magnitude, to the point where spraying it on hectares of potatoes makes economic sense. That shift, more than any single scientific breakthrough, is why sprayable RNA moved from journals to product labels.

How specific is it, really?

More specific than a broad-spectrum insecticide. Less specific than the marketing suggests.

The specificity comes from sequence matching. A 21-nucleotide guide will only direct the silencing machinery to an mRNA with a closely matching sequence, and designers pick regions of the target gene that look distinctive to the pest. But a dsRNA molecule a few hundred base pairs long is diced into many different short pieces, and each is a separate chance of a partial match somewhere else. The machinery also tolerates some mismatches. Close relatives of the target, such as other leaf beetles or beneficial ladybirds, may share enough sequence to be hit.

This can be tested, species by species, and should be. Screening sequence databases helps, but genomes are known for only a fraction of the insects living in any given field. For honeybees, toxicity studies were part of the US registrations and the published results for the products assessed are reassuring. Beekeepers have heard reassuring things before, as the story of GM crops and honey shows, and they are entitled to ask for the data rather than the summary.

Can pests become resistant?

Yes, and it has already happened in the laboratory. Selection experiments with western corn rootworm produced a population that no longer responded to the Snf7 dsRNA. The resistance had nothing to do with the target gene. The insects had become worse at taking dsRNA up from the gut, and they shrugged off other dsRNAs as well.

That is worrying in a specific way. A change in the target gene would defeat one product, and you could redesign the dsRNA to hit a different stretch. A change in uptake could defeat every dsRNA aimed at that pest at once. The standard advice applies: rotate modes of action, keep refuges, do not lean on a single tool. Whether farmers follow it depends on pricing and farm economics as much as on biology.

Is it safe to eat?

Every plant and animal we eat is full of RNA, double-stranded and otherwise. The human gut breaks it down efficiently, and mammals lack the uptake and amplification systems that make some insects so sensitive. The amount of dsRNA left on a sprayed crop is tiny next to the RNA in the food itself.

There was a scare in 2012, when a study reported that a microRNA from rice was circulating in human blood and regulating a gene in the liver. Other groups tried and largely failed to reproduce the result, and contamination during sequencing was put forward as a likely explanation. The paper still turns up in campaign material. It shouldn't. Of all the questions around RNA sprays, the one about your dinner is among the least troubling. The ecological questions are where the real uncertainty sits.

Is a field sprayed with RNA a GM field?

Not under most current definitions. The crop's genome is unchanged, so in the European Union a sprayed dsRNA is handled under the pesticide rules, assessed as an active substance like any other, rather than under GMO law. That is a separate track from the EU's rules on new genomic techniques, which deal with edited plants, not with what is sprayed on them.

Regulators have had to adapt data requirements written for synthetic chemicals. The OECD published a working document in 2020 on environmental risk assessment of sprayed dsRNA, sorting out which standard tests still make sense (breakdown in soil and water, effects on non-target organisms) and which questions are new (sequence similarity in other species, uptake efficiency, the behaviour of carriers). Not every country has settled how to classify these products, and some will be tempted to treat them as low-risk biochemicals with lighter data packages than conventional chemistry gets.

What should we be watching?

A short list, in no particular order:

• Non-target testing. Whether regulators ask for actual data on relatives of the target pest, or accept a database search in its place.
• Formulations. Whether the carriers that make sprays last longer are assessed as carefully as the RNA itself.
• Resistance management. Whether labels require rotation, and whether anyone checks.
• Who owns the sequences. Patent claims on dsRNA designed against particular target genes could give a few companies control of whole pest-and-gene combinations. Agricultural biotech has been here before.
• Herbicide rescue. Using dsRNA to switch off the genes that make weeds resistant to an old herbicide, so that the herbicide works again, is scientifically elegant. It would also extend the commercial life of that herbicide. Both things can be true.

RNA sprays are neither a miracle nor a menace. They are a genuinely different mode of action arriving at a time when many insecticides are failing against resistant pests and others are being withdrawn on safety grounds, which is reason enough to get them right. The molecular mechanism is well understood. What happens to these molecules in a real field, among thousands of species nobody has sequenced, is understood much less well, and that gap is exactly where the regulatory effort belongs.