Agri-inputs & biocontrol
RNAi biopesticides
The Dicer-2 and RISC pathway, why systemic RNAi works in beetles and poorly in moths, nuclease degradation, and sequence homology as the real substance of the risk assessment.
Double-stranded RNA that enters an insect cell is read as the signature of a viral infection. Dicer-2 cuts it into small interfering RNAs about 21 nucleotides long, one strand is loaded into the RISC complex with Argonaute-2, and the complex cleaves any mRNA complementary to that strand. If the chosen target is an essential gene, the insect dies. Specificity here is set by sequence rather than by the biochemistry of a binding site, and that is what distinguishes the class from every other insecticide.
Why it works in beetles and struggles in moths
The cleavage machinery is universal, but the RNA has to reach it. In Coleoptera systemic RNAi is efficient: ingested dsRNA survives the gut contents, is taken up by enterocytes, and the silencing signal spreads through the body. This is why the first products target beetles — western corn rootworm Diabrotica virgifera virgifera, through the DvSnf7 construct expressed in maize and registered by EPA in 2017, and Colorado potato beetle Leptinotarsa decemlineata, through the sprayable active ledprona, which silences the proteasome subunit PSMB5 and was registered by EPA in 2023 as the first sprayable dsRNA pesticide.
In Lepidoptera the same approach gives a weak and irreproducible response, and the reasons are identified. Nuclease activity against double-stranded RNA is high in the gut lumen and the haemolymph; the midgut is strongly alkaline; and what is taken up by endocytosis is largely retained in endosomes and never escapes into the cytoplasm where RISC operates. These are three separate barriers, and nanocarriers, lipid complexes and nuclease protection each address a different one.
Resistance arrives from an unexpected direction
The intuitive expectation is resistance through mutation of the target sequence. Laboratory selection on western corn rootworm produced something else: resistance arising from impaired dsRNA uptake. Such a mechanism is independent of which target was chosen, so it confers cross-resistance to all sequences at once — changing the target gene does not recover control. That reshapes resistance management: what must be rotated is modes of action, not sequences.
Persistence and the risk assessment
Outside an organism dsRNA does not last. On the leaf surface and in soil it is degraded by microbial nucleases and damaged by ultraviolet light, with half-lives measured in hours to days. Nothing accumulates in the environment — but there is also almost no residual protection, so applications have to be timed onto the susceptible stage of the pest.
Non-target assessment is done not by toxicological screening but by sequence analysis: shared stretches of 21 nucleotides or more are sought between the construct and the transcriptomes of non-target species, because 21 nucleotides is the working unit of an siRNA. The method is direct but not exhaustive: RISC tolerates partial mismatches, so the absence of exact matches lowers the risk without eliminating it. This is what the EFSA scientific opinions and the EPA data requirements concentrate on.