Many animals are venomous, but in most cases the exact proteins involved in causing pain or death are unknown, and even in those cases the genes producing them have not been identified, counted or mapped. If you’re interested in the evolution of venom, what its precursors are, and how venomous animals avoid poisoning themselves, you have to know this kind of stuff.
A new paper in Proc. Nat. Acad. Sciences (click screenshot below to read for free, or find the pdf here) answered several of these questions in the venomous caterpillar of the mottled cup moth (Doratifera vulnerans), shown below. It’s from Australia, and is described in Wikipedia this way:
It is known for its caterpillar having unique stinging spines or hairs that contain toxins, for which the scientific name is given that means “bearer of gifts of wounds”. Chemical and genetic analysis in 2021 show that its caterpillar contains 151 toxins, some of which have medicinal properties
That earlier paper, from 2021 and including some of the same authors as the one we discuss today, did indeed identify 151 proteins (peptide are bits of proteins or short chains of amino acids) that were in the toxins, but did not know which genes produced them, how the genes were arranged, what the closest relatives of the genes were, and how many of the 151 “toxins” were really toxic (the word “toxin” there and in the present paper do not mean that the substances were toxic, but that they were simply a component of the extracted toxins). However, the authors, some on the paper I’m highlighting today, did identify two genuine toxins that caused pain: the peptides Dv12 and Dv11.
Look at this thing! It’s clearly aposematic, meaning that it has bright warning coloration that predators can recognize and learn to avoid. And you can see those nasty spines. In the earlier paper they extracted toxins from related species and tested them by injecting them into mice tails, guinea pigs, and human volunteers. That earlier paper also adds this about the species name:
This species, whose binomial name etymologically means “bearer of painful gifts,” is a common culprit of caterpillar envenomations in Australia.
That means that many Aussies get stung by these things, probably inadvertently. Would you touch an animal that looks like this?:
On to the new paper, and I’ll try to be brief as it’s long and complicated.
1.) First, the authors sequenced the entire caterpillar genome (remember, it’s the same as the adult moth genome).
2.) Then, knowing the sequences of the proteins known from previous work on toxins, they could find the genes producing them by matching the protein sequence to the DNA sequence that could produce these proteins. Of the 151 proteins in caterpillar venom known from the prvious work, they mapped 149 of them to 115 sites in the genome
3.) Of the 115 sites, 35 were products of single genes, while 80 (70%) of the total, were members of gene families consisting of two or more similar genes (sometimes many genes) with similar sequences. Here’s a map of the “toxin gene” locations on the insect’s 13 chromosomes. The blue dots are the genes existing in single copies, orange dots are clusters of genes previously grouped together by protein-sequence similarity, and pink dots are genes that were newly identified, surely as part of gene families, in the present study. This conclusion comes from their sequence similarity and they physical grouping on two chromosomes.(The size of the dots indicates the number of genes that are part of a contiguous group. Click to enlarge:
So we know that genes found in venom are very often the product of gene duplications, either of single genes becoming two (this can happen via unequal crossing-over during meiosis or by other methods), producing two initially identical genes side by side or whole groups of them (“tandem duplications”). Once a gene has been duplicated, the original copy can then keep its original function, while the other copies, not being “needed,” are free to evolve other functions. Many genes we’re familiar with, like our own globins and immunoglobulins, evolved by gene duplication followed by divergence of the duplicated copies.
Where did the genes making venom proteins come from? This is the key evolutionary question answered here and, to some extent, in the previous paper. They evolved from ancestral genes in the moth’s immune system that evolved to attack microbes, the so-called “antimicrobial peptides” (AMPs), also known as cecropins. The ancestral AMP proteins, nearly identical to their original form and function, kill bacteria (prokaryotes) by disrupting the bacterial membranes. Insects still need to kill microbes!
Clearly, the proteins in venom have evolved by natural selection modifying ancestral genes used to kill bacteria. Now they are used to repel predators. Natural selection causing this divergence was implicated by looking at sequence differences, as there are ways of showing what sequence differences evolve faster than expected under either the slower processes of genetic drift or “purifying” selection that conserves structure. They found that most of the venom-adapted proteins that evolved from cecropins did evolve under natural selection, while the descendants of cecropins that retained their original anti-microbial proteins were under purifying selection to retain their sequence. It’s clear, then, that the insect still needs genes to attack bacteria. It’s just that some of them have been repurposed, often through gene duplication and divergence, to repel predators. (The authors have a way of assessing “pain” by measuring the increase in calcium concentration in cells grown in vitro and exposed to venom. This happens when the two investigated proteins are used.)
Here is a complicated family tree of cecropin genes in black used to kill microbes. The genes found in venom are in the red box (“venom adapted”). You can see that they are related to cecropin genes but branched off fairly recently (probably four or five million years ago). The venom genes are in the red box that I’ve added, and their relationship as being derived from ancestral AMP genes is very clear. (The “canonical” genes in green are antimicrobial proteins closely related in sequence to the venom genes.
So, now we know where the genes in venom come from. What we do not know is how many of those genes are essential in venom, either causing pain or doing other stuff that venom needs to do. At least two of them cause pain, but there are probably more, for they haven’t all been tested. And some of the other genes are probably involved in dismantling cell walls in potential predators. The authors tested several of the venom proteins and also found that, as in their AMP ancestors, they disrupt cell covering, in this case eukaryotic cell membranes.
Finally, the big question: If the caterpillar makes venom, why doesn’t it poison itself? Here’s how the authors answer that question (I’ve put the answer for this species is in bold).
Animals that produce toxins, either for innate immunity or as venom toxins, must employ strategies to protect themselves from toxicity. Such protective mechanisms include production of toxin inhibitors, storage in inactive form, mutations in their own ion channels that confer resistance, alteration of lipid bilayer compositions, and compartmentalization of toxins separate from body tissues. In the case of limacodid venom peptides, the venom is compartmentalized into the cuticle-lined venom reservoir inside venom spines, preventing the toxin from coming into contact with cells other than the secretory cells that produce them. Thus, compared to canonical cecropins, venom-adapted cecropins may also be released from pressure to avoid activity against animal cells.
There are other findings in the paper that will be of interest primarily to those studying genomic evolution. For example, many of the venom proteins still retain some weak antimicrobial activity, so the idea that genes completely lose their ancestral function when they gain a new one doesn’t hold in this case.
Below you can see the adult moth because, remember, they studied caterpillar venoms, and many of those genes are probably turned off in the adult. But adult and caterpillar carry the exact same genes, of course; their different bodies, physiology, and behavior rest on the differential turning on and off of these genes at different life stages. And that remains a big mystery: how do such different life stages evolve, with each step of the evolution being adaptive?
From The Australian Museum, photo credits at bottom (click to enlarge), image by Lyn Craggs.
Totally cool, as the kids say.
Have people come up with theories to explain why Australia has so many of the world’s most venomous critters? They have the world’s most deadly snakes, nasty spiders, you name it. Even in the water they have the world’s deadliest jellyfish. A religious person might think that God hates Aussies. But is there an evolutionary theory?
Many years ago I watched an hour-long BBC programme about world-wide venomous animals. It spent 40 minutes on Australia.
Fascinating stuff! I’m not surprised that venom evolved from immune-system components. They’re already used in defense, and venom is another sort of defense. It also makes sense that duplicated immune-system genes were co-opted to create venoms. The immunity genes didn’t have that flexibility (since their job was immunity), but the duplicated genes—which already have a “killing” function—were free to go wherever selection took them. Killing in a new way. So cool.
Why don’t venoms kill their hosts? The sequestration of venom into venom spines makes sense, but I wonder if it always works or—as with the human immune system—things sometimes go awry.
Jerry, you mention “purifying” selection. Is “stabilizing selection” no longer used?
Finally, no. I would not touch that animal!
Very interesting study. Thank you for your excellent summary!
“Why don’t venoms kill their hosts? ”
🎯
Comment by Greg Mayer
Norm–
Stabilizing selection is usually used with reference to quantitative phenotypic traits that have an intermediate optimum. For example, human babies survive best if they are medium sized (8 lbs) rather than too small (4 lbs) or too big (10 lbs). Such traits are typically affected by many loci.
Purifying selection is usually used in reference to molecular traits (sequences) in which a currently common nucleotide or amino acid is favored by selection, so that any mutation changing the sequence will be selected against. For example, if a mutation causes an amino acid position on the surface of a protein to change it from a hydrophilic to a hydrophobic amino acid, this might change the solubility of the protein, and thus likely be selected against.
So the terms are related but, at least modally, applied to different situations.
GCM
Very interesting. Thanks!
Yes. I’d never thought of asking these questions, but when asked, they and their answers are really interesting.
Such an exciting combination of scientific fields interrogating Nature right where it conflicts – these intriguing little peptides – I want to follow this back further …
The structures of many types of peptides can be quite surprising, and their utility amazing e.g. charybdotoxin as a means to study ion-gated potassium channels….
🔥🧠
Thank you for sharing Jerry. Good ole’ solid science, and not a whiff of wokeness infecting this particular research topic. Perhaps the researchers have their own venom to ward this off…
I love the evolution posts
I like the bit about protective mechanisms. I suppose there are animals that prey upon venomous animals. I wonder what kind of protections they have. Or is ingesting the venom different?
One answer: for some reason that I don’t understand, mongooses are pretty much immune to cobra venom. I’m sure they know why, but there is differential resistance like that.
There’s an arms race between garter snakes and rough-skinned newts in our area. The garter snakes are fairly strongly resistant to the highly poisonous newts (skin glands). The newts are becoming more poisonous to get ahead of garter snake predation.* And garter snake resistance gets better in response, and so the circle goes. Super interesting, but I don’t know what the actual chemistry is.
*I’m using teleological shorthand here in saying “to get ahead of… .”
Years ago I did a comparative study of snake venoms from a wide range of species and it was clear that proteins in venom cocktails were rapidly evolving under positive selection.
There were dramatically high levels of divergence, not only between species, but also intra-specifically among samples from different geographic locations.
Literature that I found at the time showed that prey species are able to mitigate the effects toxins through various kinds genetic changes in toxin-neutralizing serum factors and in molecules targeted by venoms (e.g. ion channels). Experiments have shown high levels of resistance when snakes and prey are from the same area (e.g. Japan or Southeast Asia) and lower levels when the are from different areas. Thus snake and their prey are like Castor and Pollox, locked into an arms race.
I think similar situation exist in species of the deadly marine gastropod Conus.
Thanks to all above. According to Wikipedia, mulga snakes eat inland taipans. Blood curdling stuff. But the Wiki page on the mulga snake does not mention it, so I am not so sure.
But there is a toad that the mulga finds particularly disagreeable: https://australian.museum/learn/animals/reptiles/mulga-snake/
The caterpillar resembles a U.S. species called the Saddleback Caterpillar, which is notorious for also being painful to even touch. Indeed, they are in the same family.
I’ve only seen a few, but I have friends who had brushed up against one and they report being quite impressed with the effect.
Great example of evolving new functions for duplicated genes. That process is also common in the evolution of genes involved in sperm–egg recognition.
To answer your question, “Would you touch an animal that looks like this?” On a trip to Florida I noticed and picked up a spectacularly beautiful velvet ant, and, well, I found out. So maybe I would?
Also in response to our host’s statement about the methods,
“First, the authors sequenced the entire caterpillar genome (remember, it’s the same as the adult moth genome)”
there was a brief period where the same journal PNAS claimed that those two genomes were not the same.
https://www.pnas.org/doi/10.1073/pnas.0908357106
Jerry dubbed it “the worst paper of the year.”
https://whyevolutionistrue.com/2009/09/04/worst-paper-of-the-year/
“a spectacularly beautiful velvet ant”
Fortunately, the first time I came across one on our land, not knowing what it was, it evaded my grasp. So, yeah, curious and clueless!
The recluse and the copperhead give no warning; the yellow jackets have wonderful color–yet they attack before I ever see them.
Another thumbs up for the science posts.
Late to this site, I had not read that Jerry’s 2009 “worst paper post”, and I’m happy to see it now, with comments. I had run into some people, academics (cultural anthropologist) who bought into Margulis’ more recent outlandish notions on the basis of her prior reputation. My intemperate response cost me a friendship.
How fascinating! And I just read about how Yifat Merbl, a scientist at the Weizmann Inst, has discovered that human and mouse proteosomes (previously thought to be simply cellular garbage disposal units, grinding up old proteins) actually chop some proteins specifically into anti-microbial peptides capable of killing bacteria https://www.nature.com/articles/d41586-025-03846-3
Good to read this post — and the comments — after a day outside.
Can someone point me to a (relatively) brief and basic explanation of the evolution of the seemingly complex stages of development (i.e., going from larvae to vastly different adult stages) in various organisms, including insect, fish, amphibians, etc.?
TIA
Having watched several monarch caterpillars molt and grow this summer and then form chrysalids, it seems plausible to me that the caterpillar is an eating machine that builds up enough fat reserves to fuel the transformation of the embryonic discs (I think they’re called) into the adult butterfly during the time the chrysalis isn’t eating. It is still “gestating” and multiplying its own living body mass manyfold at the expense of the now-“dead” caterpillar.
Insects that undergo incomplete metamorphosis — and there are many quite complex ones like locusts — may have just learned other ways to have more adult -like structures emerge during their nymph phases without having to start from scratch. A helpless chrysalis is vulnerable. There must be some payoff for taking that risk. Maybe because locusts eat the same stuff from first hatching whereas caterpillars eat different things from butterflies — milkweed leaves vs several types of nectar in the case of monarchs. Nectar can be flown to and doesn’t require the flying adult to lift heavy biting and chewing mouthparts but nectar doesn’t have enough calories to fuel metamorphosis. Leaves can be crawled to.
A good place to start would be to see which appeared first: complete or incomplete metamorphosis, or did both appear here and there depending? Then let me know what you find. 🙂
There is a good summary for insects in Current Biology from 2019 – although it’s not brief and, depending on your background, might be considered too technical. As someone who knows nothing about insects I thought it was pretty accessible, but it’s not aimed at a pure lay audience
https://www.sciencedirect.com/science/article/pii/S0960982219313156
There is a shorter and less technical (and non-woke) Scientific American piece from 2012 that might be useful
https://www.scientificamerican.com/article/insect-metamorphosis-evolution/
For vertebrates try
https://www.sciencedirect.com/science/article/pii/S0960982211008311
Thanks!
The microbes strike back at the eukaryotic cells full of organelles which they see as a target-rich environment. Many bacteria produce potent exotoxins that disrupt protein synthesis on the eukaryotic ribosome and interfere with membrane-bound organelles. Still, sequestration of toxins is a good idea in case the eukaryotic ribosome is not so different from one’s own. But how to sequester in a bacterium which is just a bag of ribosomes with a big circular strand of DNA floating in it?
Many bacterial exotoxins are of the A-B structure, with different variations on the theme. The B(inding) subunit(s) recognizes the outside of the target cell and alters the membrane so the A(ctive) toxin itself can enter the cell and wreak havoc. Botulinum toxin is secreted as a single A+B polypeptide chain which is toxically inactive until a co-secreted protease cleaves it. The A-B subunits spontaneously (i.e., with a loss of free energy) reorganize, simultaneously “arming” the toxin and protecting the secreting bacterium from friendly fire, because Clostridium botulinum doesn’t have the ligands on its plasma membrane that the B subunit could recognize and inject the A toxin back in.
There are many cool things about botulinum toxin which, thanks to Botox (TM), there is Pharma money going to basic research on it. But I wanted to highlight the sequestration strategy used widely in the bacterial exotoxin world.
A challenge for natural selection theory is to propose how the secretion of botulinum toxin makes C. botulinum more fit. Most (but not all) clinical disease cases of botulism are from eating pre-formed toxin that built up in sausages or carrion (hey, there’s more in the world than just people, you know!) There don’t have to be any living bacteria in what you eat for you to get enough toxin to die. C. botulinum doesn’t know or care that the person it just killed even existed. It’s not a virulence factor that creates more florid disease that helps the bacterium spread faster, like spike protein in SARS-Cov2, or whacks the immune response, like exotoxins from staph and strep. People afflicted with botulism don’t even have diarrhea. (They’re usually constipated.) And even though cases of botulism are very rare, the organism itself remains widely present almost everywhere in the environment. (As spores. They germinate and produce toxin only in anaerobic conditions, like in roadkill that decay bacteria have been working on for a while.) Like many bacteria, C. botulinum does secrete bacteriocins that suppress the growth of surrounding bacteria (sort of like antibiotics) but botulinum toxin is known not to be a bacteriocin.
Almost all strains of C. botulinum produce toxin — it’s not just a fortuitous clade that survived — and there are at least six different serotypes of toxin. It’s doing something useful, but what? Is the tiny amount that each bacterium secretes a clue here? It’s the most potent toxin known to humanity, which I think everyone knows.
Most of the foregoing applies to the tetanus neurotoxin of C. tetani, too, even though a markedly different disease. So botulinum isn’t a one-off.
Classic WEIT post.
Fascinating subject matter, summarised but not dumbed down.
Clear referencing and attribution.
Cool figures and photos
Comments and clarifications from PhD-level bods
Agreed. The science posts and comments are fantastic!
151 toxins, because sometimes 150 just isn’t enough!
On a more serious note, I assume that the gradual accumulation of new toxins has proved beneficial over a long time, certainly enough to make each new toxin an advantage, and I have to wonder if there will come a variant that has 152 when an environmental change favours a mutation somewhere in that complicated chain.
Cone snails present a fascinating story where it comes to toxins, and that can’t even be summarized in a long reply here. Different families of cone snails (“fish-hunting”, “snail-hunting” and another that I’ve forgotten) use families of small peptides that are clearly the result of gene duplication and are heavily disulfide-bridged to exert immobilizing and killing effects on their prey. At least one painkilling peptide has been approved for clinical use (Prialt).
Much of the pioneering work was done by Baldomero “Toto” Olivera at the U Utah, and how that came about is a fantastic story in itself.
Some also hunt polychaetes.
This is one of my favourite Conus papers. The cone snails can switch between defensive toxins (to fend off larger predatory fish and snails) and predatory toxins (to kill smaller fish or snail prey), and the predatory toxins probably evolved from defensive molecules in a polychaete-eating ancestor.
Dutertre, S. et al. (2014). Evolution of separate predation- and
defence-evoked venoms in carnivorous cone snails. Nature Communications, 5, 3521. doi: 10.1038/ncomms4521