Scoop up a handful of garden soil and you are almost certainly holding billions of Streptomyces bacteria. They are among the most abundant organisms on the planet, responsible for that petrichor smell after rain, and they have been quietly manufacturing antibiotics, anticancer drugs, and immunosuppressants for millions of years before we learned to harvest them. Scientists thought they had a reasonable handle on what these microbes get up to. They were, mostly, beneficial. Helpful. Mutualistic. That picture just got considerably more complicated.
A study published today in Nature Microbiology describes the discovery of a potent insecticidal toxin lurking within a specific lineage of Streptomyces: one distantly related to the toxin that causes diphtheria in humans, yet exquisitely tuned to kill insects and insects alone.
A Hundred-Million-Year-Old Killing Machine
The protein, named SAIP (Streptomyces antiquus insecticidal protein), shares the same fundamental mechanism as diphtheria toxin. It hijacks a cellular machine called eukaryotic elongation factor 2, or eEF2, and disables it through a chemical modification called ADP ribosylation, shutting down protein synthesis until the cell dies. Diphtheria toxin does this in humans. SAIP does it in insects. The difference comes down to a single cell-surface protein, a receptor called Flower (Fwe), which is found in insects but not in its human equivalent. Without Flower to grab onto, SAIP cannot get into a cell, which is why it is harmless to us at doses that obliterate insect tissue. The researchers confirmed this using CRISPR screens in fruit fly cells: knock out the gene encoding Flower, and cells become essentially impervious to SAIP, even at concentrations a thousand times higher than those lethal to normal insect cells.
The toxin is, in a word, ferocious. At concentrations below 10 picomolar (roughly a tenth of a billionth of a gram per litre), SAIP kills insect cells in culture. Injecting just four femtomoles into fruit flies killed about 80 percent of them within sixty hours. Flies fed SAIP-laced food gradually developed paralysis over several days, as the toxin accumulated in taste neurons and immune cells.
What makes the find so striking is how old it is, and how long it stayed hidden. Evolutionary analysis suggests the toxigenic lineage of Streptomyces arose roughly 125 million years ago, putting its emergence somewhere around the middle of the Cretaceous period, when dinosaurs still had fifty million years left in front of them. The toxin gene has been inherited vertically, passed faithfully from parent to daughter cell, generation after generation, and the pattern of genetic variation in the gene strongly suggests it has been under stabilising selection the entire time. Whatever SAIP does for these bacteria, it has been worth keeping.
“Streptomyces have primarily been known to have mutualistic relationships with insects, but we have discovered a clade of strains that are likely insect pathogens,” says Min Dong of Boston Children’s Hospital and Harvard Medical School, who co-led the research. The shift in framing matters. Most Streptomyces live on, in, and around insects in a broadly cooperative arrangement, protecting wasp larvae from fungal infection, helping beetles digest cellulose, hitching dispersal rides on soil arthropods attracted to their spores. This lineage, it seems, took a different evolutionary turn.
Kill, Consume, Defend
Cameron Currie at McMaster University, the other co-lead on the study, puts it bluntly: “They don’t just kill insects, they are also remarkably efficient at consuming them, using their dead hosts as a source of critical nutrients.” The team tested this by inoculating dead grasshoppers with a SAIP-producing strain. Within a week, the bacteria had colonised the exoskeleton, spread across the carcass, and digested it almost entirely. As they fed, they also produced conspicuous red-pigmented compounds: two antimicrobial molecules called undecylprodigiosin and streptorubin B, thought to ward off competing microbes attracted by the decomposing insect. It is a fairly complete strategy, really. Kill, consume, and defend the meal.
The evolutionary connection to diphtheria toxin is, frankly, one of the stranger aspects of the story. The bacterium that causes diphtheria, Corynebacterium diphtheriae, acquired its toxin through a bacterial virus at some point in evolutionary history, the gene shows clear signs of horizontal transfer, and where that ancestral toxin originally came from has remained an open question. Currie suggests the Streptomyces toxin family might represent something like an evolutionary precursor. “We know that the bacteria that causes diphtheria acquired its toxin from another species of bacteria long ago,” he says, “so it’s possible that these Streptomyces toxins were the crucible for the eventual emergence of the diphtheria toxin.” He is careful to call this speculative. The structural and mechanistic resemblances are real; the lineage question stays open.
From Soil Chemistry to Mosquito Control
The practical implications are more immediately tractable. Insect-specific toxins are in high demand, both in agriculture and in public health, and crucially, the SAIP receptor Flower is present in mosquitoes including Aedes aegypti and Anopheles species, the main vectors of malaria, dengue, and West Nile virus. The researchers confirmed that SAIP kills mosquito cells at concentrations similar to those effective against fruit fly cells. Whether this translates into a practical vector-control tool involves many more steps, but the team has already filed a patent. “A toxin like this could potentially help control vectors of human disease, like mosquitos, which can transmit malaria and West Nile virus, or perhaps be used to protect crops from insect pests,” Currie says. “It’s possible it could be used in a number of different ways.”
Importantly, SAIP’s mode of action is completely different from Bacillus thuringiensis Cry proteins, the insecticidal toxins currently engineered into transgenic crops. Resistance to Bt toxins is spreading in some pest populations, and an entirely new class of insecticidal protein, with a distinct receptor and a distinct mechanism, could eventually prove useful in managing that resistance by giving farmers something new to rotate through. The Currie lab has previous form here: in past research they have identified promising antibiotics from other Streptomyces strains, which perhaps explains why Currie sounds genuinely optimistic rather than merely cautious about what these particular bacteria might eventually yield.
“That we have found something so novel in one of the world’s most abundant and well-studied groups of bacteria underscores how little we actually know about them,” Currie says. The Streptomyces genus has been subject to perhaps unparalleled microbiological sampling; pharmaceutical companies screened millions of soil isolates through the mid-twentieth century in the search for new antibiotics. And yet here, buried in a lineage that barely registered in public sequence databases, was a hundred-million-year-old killing machine. “This toxin stands as a powerful reminder that bacteria are incredibly diverse organisms, with capabilities that continue to surprise us.” Given that Streptomyces have coexisted with insects for over four hundred million years, the full accounting of their interactions, cooperative, antagonistic, and everything in between, may have quite a lot further to run.
Frequently Asked Questions
Is this toxin dangerous to humans or pets?
No, and the reason is what makes the discovery interesting. SAIP can only enter cells that carry a specific surface protein called Flower, and while insects have a version of Flower that serves as the toxin’s entry point, the human and mouse equivalents don’t work that way. In lab tests, human and mouse cells showed no meaningful toxicity until SAIP concentrations reached levels roughly a thousand times higher than those lethal to insect cells. The specificity appears robust, though broader ecological safety research is still ongoing.
Could this actually replace chemical pesticides?
It’s too early to say, but the biology is genuinely promising. SAIP works through a mechanism entirely different from existing bio-insecticides like Bt toxins, which means it could potentially be rotated with current tools to manage resistance, a growing problem in agriculture. The researchers have filed a patent and are beginning to explore commercial pathways, with mosquito-control applications also on the table. The gap between a lab discovery and a registered product is wide, but the underlying specificity and potency give it a more credible starting point than most early-stage candidates.
How is this related to diphtheria, and should that worry us?
The connection is evolutionary, not clinical. SAIP and diphtheria toxin share the same underlying molecular architecture and the same mechanism for disabling cells, but the critical difference is their receptors: diphtheria toxin binds a human cell protein, SAIP binds an insect-only protein. The researchers speculate the ancient Streptomyces toxin family may have served as an evolutionary precursor to the diphtheria toxin that eventually ended up in a bacterial pathogen, though that remains speculative. There is no suggestion that SAIP itself poses any disease risk to humans.
Why hadn’t anyone found this toxin before, given how well-studied Streptomyces are?
Partly because the toxigenic lineage is a small and undersampled slice of an enormous genus, and partly because the field has historically focused on the small-molecule antibiotics these bacteria produce rather than their protein-based weapons. The researchers identified SAIP by searching genomic databases for structural relatives of diphtheria toxin, a different kind of search than the compound-screening approaches that dominated twentieth-century antibiotic discovery. It raises the question of what else might be hiding in one of the planet’s most abundant groups of microbes.
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.