Horizontal gene transfer in insects: widespread, but what does it mean?

July 22, 2022 • 12:45 pm

“Horizontal gene transfer”, henceforth “HGT”, is the process whereby a gene is moved between species by methods other than direct reproduction (the latter is called “vertical gene transfer”). Today I’ll write about a new survey published in Cell, trying to find out how often this process moves genes into the DNA of insects.

It’s surprisingly common, and the authors found one case in which a gene transferred into Lepidoptera from bacteria seems to help moths and butterfly males get mates. But HGT is not common enough to screw up the evolutionary trees of insects. As I said, it’s commoner than I thought, but it’s still pretty damn rare.

(I note that you can have gene transfer between species via reproduction, too, if you allow for a little bit of interbreeding between “species”. Although I think that Neanderthals should be considered members of our own species H. sapiens, we did have some vertical gene transfer in the past when members of the two groups mated. The hybrids passed on their genes to modern humans, and the result is that is that many of us carry a few Neanderthal genes. That’s via reproduction, though, not infection, which is the way HGT occurs.)

Typically, in HGT a bacterium, virus, or other microorganism that lives within a eukaryote host transfers some of its own genes into the host genome, or, in some cases, acts as a vector carrying genes from other organisms, including fungi, plants, or other animals.  But HGT is not really a “non-Darwinian process” that results in a drastic revision of the Modern Evolutionary Synthesis. No, Darwin didn’t know about this—he didn’t know about genes, either—but we can consider the introduction of a novel gene from a very different species as a type of mutation (granted, a big one), producing variation in the recipient species that can be acted on (increased or eliminated) by natural selection, or which could be “neutral”, neither enhancing nor reducing fitness.

And although HGT in principle could mess up evolutionary trees if it were very common, that doesn’t seem to be the case. In fact, HGT doesn’t occur that often, and is actually detected by finding a gene in an organism that, via sequencing, looks as if it came from a different species. Seeing such a discrepancy depends, as it did in this study, in knowing evolutionary trees. We wouldn’t be able to detect HGT if trees were all messed up by the process.  So, no, this new paper doesn’t show that “Darwin was wrong”, much as the Kuhnians want that.  I’ll emphasize this at the end!

With that out of the way, what do we know about the process of HGT? We know that it can be effected by various vectors, usually microorganisms, occurs in a lot more species than we thought, is a fairly recent discovery, and on occasion can create “mutations” that are good for the recipient species. It’s especially widespread among bacterial “species,” in which lots of genes, including antibiotic resistance genes (unfortunately for us) can be transferred widely. That’s how bacteria can quickly evolve resistance to new antibiotics: genes for resistance, favored by natural selection, can be spread among several species via HGT.

What about organisms with true cells—the “eukaryotes”? We know about lots of cases of HGT (see the Wikipedia article on HGT for a good summary), but only three in which HGT has transferred a gene that rose in frequency in the new host by natural selection. Here are the examples cited by Li et al.:

In addition to pieces of symbiont genomes introduced into insects via HGT, some studies have reported the transfer of a single or few genes from fungi, bacteria, plants, and viruses (e.g., Boto, 2014Husnik and McCutcheon, 2018Irwin et al., 2022Perreau and Moran, 2022). The functions of these transferred genes appear ecologically important; for example, carotenoid biosynthesis genes transferred from fungi to aphids contribute to aphid body coloration (Moran and Jarvik, 2010), genes that neutralize phenolic glucosides acquired by whiteflies from plants contribute to whitefly detoxification capabilities (Xia et al., 2021), and a parasitoid killing factor gene transferred from a virus to lepidopterans contributes to lepidopteran defense (Gasmi et al., 2021).

As far as we know, then, the vast majority of adaptations within eukaryotic species arose not via HGT, but via simple mutations in genes already present on a species’ genome: the “Darwinian way.”

However, research on HGT has been spotty, often highlighting single case studies. Now, a group of 13 authors from China have published a systematic survey of insects, examining the possibility of HGT in 218 species taken from 11 of the 19 orders of insects.

Click on the screenshot below to read the article for free, get the pdf here, or click on the reference at the bottom. If you can’t access the article, a judicious inquiry might yield a pdf.

The authors looked for horizontally transferred genes by examining whole-genome sequences of species and looking for regions of the DNA where an insect’s genome was not related to other insects in its group, but to a very distant group: a discrepancy in the “family trees” of genes. They then determined the protein sequences of those putative products of HGT to ensure that the genes that were transferred actually did something—i.e., make a protein.

They found that among the 218 species, 192 had protein-coding genes horizontally transferred from other species. This transfer involved 1,410 genes in 741 transfer events (more than one gene can be transferred in a single event). Most of the genes inserted into insect DNA came from bacteria: here’s the rundown:

Sources of HGT genes in insects:

Bacteria: 79.0% of the transferred genes
Fungi: 13.8%
Plants: 3.0%
Viruses: 2.8%
Other groups: 1.8%

As one would expect—given that virtually every animal harbors bacteria, and because bacteria are good at transferring their genes to other genomes as well as acquiring genes from hosts—we see the bulk of HGT-transferred genes coming from bacteria. It’s likely that genes from fungi or plants were transferred by viruses or bacteria, but there can also be direct transfer in a variety of ways, including ingestion. See the section on “eukaryote to eukaryote transfer” in Wikipedia, which also notes that there’s a possibility that some of our own genes are there because of HGT:

One study identified approximately 100 of humans’ approximately 20,000 total genes which likely resulted from horizontal gene transfer, but this number has been challenged by several researchers arguing these candidate genes for HGT are more likely the result of gene loss combined with differences in the rate of evolution

Not all groups of insects are equally susceptible to HGT. Here’s a figure you should enlarge, showing the phylogeny of insects studied, with the height of the bars representing the number of genes in each species acquired by HGT.  You’ll have to click on this one to see it all. But what’s clear, even in a small figure, is that Lepidoptera (butterflies and moths) carry a lot more horizontally transmitted genes than insects in other orders (look from 3-6 o’clock below):

As the authors note:

. . . . . we found that the order Lepidoptera acquired by far the highest average number of HGT-acquired genes (16 genes per species), followed by the orders Hemiptera (13 genes per species), Coleoptera (6 genes per species), Hymenoptera (3 genes per species), and Diptera (2 genes per species).

Many of the “HGT genes” in different Lepidopteran species are the same ones, implying that they had already been put via HGT into the common ancestor of the group and then were transferred vertically through reproduction, distributed into the various butterfly and moth species that evolved later. One of the transferred genes seemed to have adaptive value in Lepidoptera; more on that shortly.

Here’s a figure of who the donor species are. The overall picture was given in my list above, but this shows which subgroups of bacteria are the major vectors:

This also shows—take my word for it—that (mainly bacterial) groups that are symbiotic with an insect species (i.e., living within it or associated with it) were the most likely sources of horizontal gene transfer. That of course is also expected. You can’t get genes from an unrelated species unless the vector is nearby.

Finally, there’s one possible case of an adaptive HGT. That is the gene LOC105383139, horizontally transferred to the last common ancestor of Lepidoptera from the bacterial genus Listeria.  We don’t know what it does in the bacterium, but it does have one effect on a species of moth. Because LOC105383139 is in the genomes in all lepidopterans, Li et al. used the CRISPR-Cas9 gene editing technology to simply remove it—snip it precisely out of the genome—from one species of moth. the diamondback moth (Plutella xylostella), which can be reared fairly easily in the lab.

Li et al. found no significant effect of gene removal on any morphological or developmental character in the “snipped” moths, but the “knockout species” did have a much lower number of offspring from each mating compared to the normal moth having the gene. Upon further study, which I won’t describe, Li et al. found that the knockout males don’t court any females as ardently as normal males who carry the transferred gene. Normal moths also copulate more often than the edited ones. For some reason, since we don’t know what the gene does in either bacteria or moth, it is a mysterious “Casanova” gene”.

The diamondback moth

The upshot: In insects, at least, horizontal gene transfer is surprisingly common, with 88% of the 218 species examined having at least one gene in their DNA that came from another species. It’s widespread, too—at least one species in each of the 11 orders of insects examined (8 orders weren’t examined) having at least one HGT-acquired gene.

Of the species examined, butterflies are by far the most common carriers of transferred genes, with an average of 16 such genes per species, with the Hemiptera—true bugs—coming in second at 13.  In butterflies—at least in the one species examined—one of the HGT loci appears to have enhance mating and courtship, and that’s how it may have spread when it infected the genome of the common ancestor of all living moths or butterflies.

Finally, the big question: does this show that because evolutionary trees don’t exist, “Darwin was wrong”? Not on your life!

As I wrote in a previous post, there’s a difference between gene trees (the evolutionary history of a particular form of a gene) and species trees: the relatedness of groups of organisms that formed species as populations. Unless HGT is very, very common, the chance that it will mess up population trees—the trees of life that most concern evolutionists—is almost nil. Butterflies have thousands of genes, but each species has on average only 16 acquired from a different species. That is NOT going to mess up evolutionary trees. It’s a very small bit of noise in an overwhelming phylogenetic signal.

So, while Darwin was wrong about some things, we have no sign at all that he was wrong about the existence and importance of evolutionary trees that show how species are related. Nor was he wrong about variation that arose within species being the primary fuel for natural selection. HGT is fun and interesting, and a novel way of getting mutations and adaptations, but it hardly affects our general view of evolution.


Reference:  Li, Y. et al. 2022. HGT is widespread in insects and contributes to male courtship is lepidopterans.  Cell, DOI:https://doi.org/10.1016/j.cell.2022.06.014

28 thoughts on “Horizontal gene transfer in insects: widespread, but what does it mean?

  1. Is there any evidence that HGT being around since the beginning of life on earth or is this a more recent mechanism?

    1. HGT is suspected to have been around since the beginning of life, when life was just simple, non-nucleated cells in intimate co-existence in hydrothermal vents. Researchers speak of the “mangrove of life” (rather than the “tree of life”) since there were likely many interconnections thru HGT that built the shared metabolic pathways we see in all cells today. Mangrove roots interconnect with each other in a kind of web, so its a fair analogy.

  2. Looking to the far future, what will a thousand years of genetic engineering do to the Tree of Life?

    1. Probably almost nothing. The only species that would be worth the expense would be those where significant profit would be expected from the genetic engineering. So, Bt genes into a number of crop species, and possibly some anti-plasmodium characters into several species of mosquito, but not huge numbers of changes all through the “tree of life”.
      What reason would you have, for example, to dedicate a lab and several post-Docs for several years to the task of transferring canine smell biology into these “diamondback moth[s]”, of which I was blissfully unaware until a few minutes ago? TTBOMK, their interactions with humans are limited – perhaps only a modest amount of pollination service. So, who would spend the money, and why?
      (That’s one of the reasons I’m somewhat suspicious of the reported “explosion” in “biohacking” – the rationality of some of these people seems decidedly suspect, whereas the rationality of profit and loss is probably more predictable. The rattling, as Orwell put it, of the stick in the swill bucket.)
      On the other hand, I recall a number of years back (2000-odd) reports that people who were trying to resolve the very deepest roots of the “tree of life” in the bases of the Archaea and Bacteria, were struggling to find character sets that resolved into simple trees, but instead had a unresolvable mass. Which is exactly what you would expect from abundant HGT between taxa.

  3. The cause of successful horizontal transfer of genes like “Casanova” in moths & butterflies will be hard to figure out because the horizontal transfer event was so old. The new gene could have easily acquired other functions in some other lepidopterans, and could have been lost in many lepidopterans, so surveying the function of that gene across many species may show a diverse set of current functions. The causes and effects of differential gene loss after a new gene is acquired (by HGT or by gene or genome duplication) is an interesting and understudied problem.

  4. Thanks for this post; it was engrossing, interesting and fascinating. This study seems like it would be difficult work, and probably impossible without powerful computers. (Is that stating the obvious?)

    I wonder if Lepidoptera having the highest occurrence of HGT and Diptera the lowest is coincidental or is there a reason the Lepidoptera genome is/was more susceptible to HGT and Diptera not so. Metamorphosis? Just spitballing here.

  5. If Lynn Margulis were still with us, I’d imagine she’d claim vinidication for her Onychophoran caterpillar origin story!

    PS: Nooo!

        1. There’s a great Feynman quote… so… I’m gathering that …. [ deep breath] humans are not hybrids.

  6. Could it be said that all mitochondria-bearing life is the result of HGT since mitochondria evolved from invasive organelles? Plus, further gene migration occurred as genes from these ancestral organelles migrated into host genomes.

    1. Yes. The traditional view (a ‘la Lynn Margulis) is that mitochondria got their start as bacteria that lived as endosymbionts in early eukaryote cells. But there is consensus now that this view is fairly wrong. Instead, eukaryotes got their start as a fusion between bacteria and archaea (another kind of simple non-nucleated cell). The resulting interspecies hybrid cell became a eukaryote cell. Meanwhile, the remnant of the bacteria cell is now our mitochondria. So yes, the eukaryotes got their start from a humongous HGT event of sorts.

      1. The traditional view (a ‘la Lynn Margulis)

        It just doesn’t sound right. Even if it is true. What’s that rotating sound?

        So yes, the eukaryotes got their start from a humongous HGT event of sorts.

        And within the eukaryotes, a second symbiotic event leading to the incorporation of photosynthetic chlorophyll-containing into (mitochondria-containing) eukaryotes to form the Plantae.
        I remain unconvinced by Margulis’ argument for endosymbiotic origins for (IIRC) the endoplasmic reticulum, nucleus and Golgi bodies, but on general principle, if it’s happened twice, you’ve got to have a good reason for thinking it can’t happen a third, fourth, fifth time.

  7. Thanks for pointing this out, and, especially, for the analysis. From an information theory aspect (a big part of my primary field), this is pretty clear, though I would likely have not read the paper had you not pointed to it, since I generally don’t have the contextual background to follow the details.

    (I passed this on to the local AP biology teacher. I will be very unsurprised if she uses the paper and your post with the summer class, as evolution and genetics are a big focus, and she likes to use current materials with the students)

  8. Very cool! One can hope that this won’t provoke another “Darwin was wrong” type article. But it’s a forlorn hope.

    1. I sometimes wonder if it would not be better to speak more openly about where Darwin was wrong, such as with his pangenesis and gemmules. This might help to parry the accusation of Darwin worshipping that one occasionally sees. In other words, it is not the case that modern scientists treat Origins of Species like a bible and defend Darwin religiously, they evaluate his work in the face of old and new empirical evidence, keeping his successful hypotheses and discarding his flubs.
      Of course, the issue of VGT vis-à-vis HGT would be a case in which Darwin was “wrong” in as far as he had no idea about the mechanisms of inheritence, as we all know, but not really any “more wrong” than had already been obvious to all.
      Make any sense?

      1. People have done that ad infinitum, the most obvious error being about how genetics works and his view that changed “conditions of life” could create variation. I’ve written about how Darwin got speciation wrong, despite the title of his book, because he had no concept of what a species really was. He thought the Earth was a lot younger than it proved to be (but he had no way of knowing). What makes so many of us see Darwin as a role model is that he got so much RIGHT, and also constantly considered and rebutted objections to his theory.

        Look, lots of people, including creationists, New Scientist, and the Guardian, participate in “Darwin was wrong” exercises. I’d take a view opposite to yours: people should write about how much Darwin got RIGHT!

  9. HGT blew me away when I first read about the process, tricky little blighters I thought.
    It drove home, life doesn’t sit still and wait for you know who, it gets on with the game and some creature will luck out with an advantage.
    Don’t know about you but I tire of wokeness and this post is a good back-to-basic grounding.

  10. Thanks for an engaging post. The Wiki page is very well done also. I’m glad it agreed with you and pointed out why the Tree of Life is still well-rooted.

    1). The mechanism for the incorporation of donor DNA into the recipient cell is well-studied for prokaryotes, recombinational hot spots essentially. The mechanism appears to be less well established in eukaryotes. Is this because the DNA is less accessible except during mitosis, being tightly packed and coiled into histone-containing chromosomes?

    2). For the horizontally transferred DNA to be passed to progeny it has to infect the germ cells, sequestered away in gonads. Any thoughts on how this is achieved? Is this barrier why we don’t see it very often?

  11. Is there any clear relation between the frequency of gene transfers, and how distant the transfer event is from the root of the insect tree? The mass of lepidopteran finds seem to come from an event 5 branches from the root (at about 03:50 on the clock diagram), but the two viral HGT events at 09:40 (I can’t read the family names) are also about 5 levels from the root. A little further round at 10:20-ish is a spike of 170 HGT events in one family, about 7 branches from the root.
    I’m looking for a pattern there, but not seeing anything obvious. Does anyone know a better name for the sort of breakdown I’m trying to express. It’s not a clearly chronological succession – more related to the evolutionary radiations of the group than “absolute” dating.

    1. An intriguing idea. But also notice the generally low frequency of HGT in the Diptera (the order highest up in the tree). Thus, tree depth and HGT frequency cannot be strongly correlated. An explanation would have to account for that.

  12. Wowee zowee! Thanks for distilling and reporting on this paper, and for quickly dispelling the interpretations that so many, as you rightly put it, “Kuhnians” tend to jump the gun with. There are some intriguing patterns across taxa that are in those figures. What an exciting research avenue.

  13. Thanks for this! HGT varies between organisms, but the old rule of thumb on it’s average is “once per gene lifetime” making some gene trees a bit harder to tease out by necessitating methods to account for it (and its variation). But the point is that we have such methods now and we get the trees, as you note.

    I find it interesting that Archaea, which seems to be lumped with the 1-2 % “Others”, has little transfer into the many insects. They do transfer between themselves to various degrees, but they may be frugal in others. It could perhaps have something to do with their low (thus far non-existent) pathogenicity.

Leave a Reply