Aphids nab pigment genes from fungus

Some of those critics who see neo-Darwinism as grossly insufficient assert that lots of adaptations result not from mutations that arise within a species, as evolutionary biology posits, but from “lateral gene transfer” (LGT) the capture of genes from one species by a different one.  This transfer can occur by either hybridization (mating with a different species) or assimilation (eating or absorbing foreign DNA).

This attack on neo-Darwinism is misguided on two counts.  First, neo-Darwinism doesn’t really require that genetic variation originate by mutation in the species in which that variation is selected.  The dynamics of natural selection will work on adaptive genetic variation no matter what its source. Second, while bacteria often acquire genes from different species (much of antibiotic resistance, for example, is carried on plasmids—small circular pieces of DNA—that one species acquires from another), there are only a handful of cases in which nonbacterial “eukaryotes” (animals, plants, and fungi) get genes from a different species (Wikipedia has a nice summary of these cases).

One reason we think that LGT is fairly rare in eukaryotes is because we’d detect it by making DNA based phylogenies.  Genes captured from a different species, especially one that is quite different, would stick out of these phylogenies like sore thumbs.

Indeed, that is the method used by two researchers in a really nice demonstration of LGT between fungi and aphids, reported by Nancy Moran and Tyler Jarvik in this week’s Science.

Moran and Jarvik were studying a color polymorphism in the pea aphid (Acyrthrosiphon pisum): some individuals are red, and others are green (Fig. 1).  This polymorphism behaves as if it were controlled by a single gene, with red color dominant to green.  Assays of differently colored aphids showed that they contain different kinds of the pigment carotene: green aphids have three types of carotenes, while red ones have those three and two additional ones.

Fig. 1.  Clones of red and green aphids (from Moran and Jarvik).

Further, the polymorphism is thought to be maintained by natural selection: ladybugs preferentially pick off the red aphids, while green aphids are more often destroyed by parasitoid wasps.  This may cause frequency-dependent selection, in which the color morphs are kept polymorphic because the rarer forms are eaten/parasitized less frequently.  This kind of selection can, theoretically, maintain both genes—and colors—in the population.

This puzzled Moran and Jarvik, because all carotenes in animals have been thought to come solely from diet, since no animal species is known to make the pigments with its own metabolic machinery. (In WEIT, I discuss how male house finches become more attractive mates if they have redder feathers; the red pigment derives from carotenoids in the seeds that the finches eat, and is a sign to a female of a healthy, well-fed male who would be a better father [see photos below].)  But it’s unlikely that aphids get carotenoids from plant sap (the aphids’ food), because those pigments are not soluble in sap, and the carotenoid polymorphism appears, as I said, to act as if it were produced by genes in the aphids themselves.  Indeed, DNA sequencing—the aphid genome was just sequenced completely—revealed that aphids do indeed carry carotenoid-synthesizing genes in their genome.  There were seven of them, coding for both carotenois desaturases and carotenoid synthases.

This led Moran and Jarvik to hypothesize that somehow the aphids acquired genes for making carotenoids from another species, presumably bacteria. They thus compared the aphid carotenoid genes to those from other species in the genome databank.  And they got a surprise.  Yes, the aphid genes did come from another species, but not a bacterial one.  They were closely related, instead, to genes from fungi.

Figure 2 gives a phylogenetic tree of the carotenoid desaturases, and shows clearly that the aphid genes nest within the group of desaturases from fungi.  This tree (and the tree for synthases as well) also show that all seven of the aphid genes were acquired from fungi in a single capture event between 80 and 30 million years ago.  We don’t know how this happened, but it’s possible that an ancestral aphid infected with a fungal disease captured some of the fungus DNA.

Fig. 2.  Phylogeny of carotenoid desaturase genes from various species.  Bacteria in black, plants in green, fungi in red, and aphids in blue.  Aphid genes cluster within fungus genes.

Moran and Jarvik also showed that the red-versus-green polymorphism is based on a mutation that presumably happened after the genes were captured:  green aphids derive from a “mutation” in one of the carotenoid desaturase genes, a mutation that deleted about 30,000 base pairs of the DNA.  Presumably the red color was ancestral, and the green resulted from an error in DNA replication.  (Moran and Jarvik also studied a mutation from red to green that spontaneously arose in the lab, and found that the new green form was based on a single amino-acid change, from glutamic acid to lysine, in the same carotenoid desaturase gene.)

This is a remarkable use of an acquired gene in an adaptive way, for the captured fungus DNA is the basis for the color polymorphism presumably maintained by natural selection.  It’s clear, then, that evolution in one species can be based on the acquisition of genetic information from a distantly related species.  Now that different species’ genomes can be sequenced quickly and reasonably cheaply, we’re bound to find more cases like this.  I don’t think they’ll be that common, simply because we don’t see evidence of LGT from existing gene trees in animals and plants. Nevertheless, Moran and Jarvik have shown that nature still has the capacity to surprise us.  And a good thing, too, because it makes our jobs as evolutionary biologists even more interesting.

Fig. 3.  Male house finches (Carpodacus mexicanus) showing color variation due to diet. Finch at bottom has had a lot more carotenoids. Photos from Project FeederWatch.


Moran, N. A. and T. Jarvik.  2010. Lateral transfer of genes from fungi underlies carotenoid production in aphids.  Science 328:624-627


  1. Thanny
    Posted May 1, 2010 at 9:04 am | Permalink

    I think it’s a mistake to say that female house finches in any sense see male redness as a sign of being well fed.

    They like the red, period.

    Even if the only way males can get the color is via food, it’s just confusing the issue to label the red preference as a food proxy, because there may be (and almost certainly are) a lot of male finches that are very well fed, but not with the kind of seeds that provide the red pigments.

    So it’s not a matter of females selecting males that are better able to find food, but females selecting males that are better able to make themselves red. And since the less red ones are more likely to avoid being eaten by a hawk, clearly the color red is not a proxy for survival in any sense. Again, it’s a proxy only for the male’s ability to make itself red.

    Conversely, if house finches find themselves in an area with very few seeds that are nutritious, but plenty that do little more than supply carotenoids, there will no connection between red feathers and the ability to find a sustaining diet. The female preference for red would then be even more at odds with general survivability, both of male and female offspring. A sustained environment like that could result in the population dying out completely, or losing the female preference for red feathers.

    • Tulse
      Posted May 1, 2010 at 9:48 am | Permalink

      The actual mechanism the females use is indeed probably just a preference for red, but it presumably evolved because in the original environment it served as an excellent signal of being well-fed. In other words, historically, red in males signals well-fed, hence females evolved a preference for red. Jerry is right that “the red pigment […] is a sign to a female of a healthy, well-fed male”, in those environments where, as was the case historically, there was a correlation between the red colour and being well-fed. The females are consciously looking for well-fed mates, but what they are looking for is a sign that historically was associated with being well-fed.

      • Tulse
        Posted May 1, 2010 at 9:50 am | Permalink

        Make that “aren’t consciously“, of course…

        • artikcat
          Posted May 1, 2010 at 11:20 am | Permalink

          “looking for”?

          • Tulse
            Posted May 1, 2010 at 1:59 pm | Permalink

            Is “have a mating bias for” less anthropomorphic?

            • artikcat
              Posted May 1, 2010 at 5:15 pm | Permalink

              I suggest one should say: “the birdies-the females-mate with red …..etc etc..”

            • Tulse
              Posted May 1, 2010 at 8:52 pm | Permalink

              But they don’t just “mate with red” males — they show a preference for red on males (i.e., are more likely to mate with a male the more red he is).

            • artikcat
              Posted May 2, 2010 at 8:36 am | Permalink

              they “dont show preference” they “mate with red..more often…” subtle but critical

            • Tulse
              Posted May 2, 2010 at 10:31 am | Permalink

              Perhaps you’re right to be so cautious in your language, artikat, but as I understand it, the formulation “mate with red more often” is almost certainly too narrow. While I admit that I don’t know a lot about finches in specific, it seems likely from Jerry’s description that we know more than just that females “mate with red [males] more often”. For example, it seems extremely probable to me that it is not the case that redder males are more able to force themselves on females despite females’ attempts to resist, even though such a scenario would produce the results that females “mate with red more often” (this would be in the same sense that gazelles “get eaten by lions more often than by hippos”). Also, given Jerry’s description, I’d be willing to bet a sizeable sum that one could devise lab experiments showing that females would expend more effort in an attempt to mate with redder males — for example, they might spend more time trying to get into a cage containing a red male. This would be true even if they didn’t end up actually mating, which means that describing the overall situation as “mate with red more often” is not only impoverished, but in a very real sense incomplete. Females show a preference for redder males, even if they cannot successfully act on that preference.

              I think it is a mistake to demand radical behaviourism as the mode of discourse for animal behaviour. My dog certainly has a preference for cheese over dog kibble, even though he does not “eat cheese more often” — if offered cheese and kibble at the same time, he will choose cheese, and if cheese and kibble are both shown to him then hidden in odour-proof containers, he will expend far more effort to open the cheese-containing one. My dog doesn’t “eat cheese more often”, he has a preference for cheese. I think a similar description is almost certainly true with female finches, and is certainly more complete than saying they simply “mate with red more often”.

      • Bimston
        Posted May 1, 2010 at 7:37 pm | Permalink

        I think you’re misinterpreting the motivations of the female finches. The female finch mates with the red-tinted finches not because it’s a decent signal of being well-fed, but because they like the red. That these red-tinted males were presumably well-fed and, therefore, had higher genetic fitness to pass onto their offspring is what maintains such a preference in the females: females that prefer high-fitness males produce high-fitness offspring and therefore pass more fit-male-preferring genes as well. The reason for that choice is completely immaterial (and may, in fact, derail from fitness into bizarre display adaptations as in many cases of sexual selection).

        • Tulse
          Posted May 1, 2010 at 9:12 pm | Permalink

          The female finch mates with the red-tinted finches not because it’s a decent signal of being well-fed, but because they like the red.

          No one is saying that the female finches are telling themselves “Hey, that males is really red, which is a signal to me that he much be very well fed.” The argument is that the female finches prefer redder mates, but that evolution has shaped that preference, and it was shaped in that way because it turns out that the red was historically a reliable phenotypic indicator correlated with male nutrition. Why else would the females “like the red”? (There may indeed be possible counter explanations, but presumably all such accounts must ultimately rely on the adaptive nature of such sexual selection.)

          There is no claim that female finches are doing sophisticated reasoning, or using colour as a cognitive marker. (Likewise, no one suggests that Tinbergen’s gull chicks told themselves “Hey, there’s a red dot, and I know that my mom has one of those on her bill, so I’ll peck at it to get food!”.)

          Slightly off topic, it seems to me that sexual selection is an excellent counter to F&P-P’s claim about the indeterminacy of the object of selection, since we can do very elegant experiments looking at what precisely is being selected. For example, we could presumably show that it is indeed redness, and not well-fedness in general, that is picked out by females finches through offering females scrawny males dyed red and robust males bleached out. But we could also presumably in principle show that, if we removed carotenoids from finch diets, and regularly dyed the wimpier males red, thus reversing the signalling value of red, after several generations the preference for red males would be weakened. It is hard to imagine how one could explain such presumed results without invoking natural selection.

          • Thanny
            Posted May 4, 2010 at 11:55 am | Permalink

            Your suggested experiment would produce interesting results. The latter part, about making less survivable males redder, that is. I’ve little doubt that we could demonstrate that it was redness being actually selected.

            It’s a complicated bag, sexual selection. I think the main mistake people make is giving a single answer, such as “the females choose red, because it indicates a well-fed male”. Even under normal circumstances, the correlation of color with diet quality is going to vary considerably. A male finch will be more likely to eat certain types of food to get the necessary pigments, even if they’re not the most nutritious types available.

            I think the strongest correlation is between redness and successfully mating, as a self-reinforcing phenomenon that doesn’t require any increase in survivability (females choosing red males are having sons that tend to become red, and daughters that tend to like red males).

            Now it may be that if the diet necessary to grow red feathers becomes sufficiently inferior to another available diet that doesn’t give you red feathers, then survivability will overrule mate choice, and a preference for red will begin to disappear. But if a finch can be well fed either way – choosing foods high in carotenoids, or foods with next to none – then I think it makes more sense to see red as indicative of nothing more than mating choice.

            What do the facts support? I don’t know, but I’d find the answer interesting.

            • Tulse
              Posted May 4, 2010 at 12:01 pm | Permalink

              if a finch can be well fed either way – choosing foods high in carotenoids, or foods with next to none – then I think it makes more sense to see red as indicative of nothing more than mating choice.

              But surely it is highly likely that there is a reason why such mating choice evolved in the first place, that such a choice has fitness implications for the female. And the likeliest explanation is that redness is a reasonably reliable proxy for mate health.

            • Thanny
              Posted May 5, 2010 at 12:29 pm | Permalink

              Well, that’s the big argument over sexual selection.

              Fisher showed that you don’t need a proper reason at all for the preference to start. All you need is a random preference for something like “more red” or “longer tail”, and the rest automatically follows.

              Again, it’s because when a female who prefers more X mates with a male who has more X, their offspring will include sons that tend to have more X, and daughters that tend to want more X.

              It’s a simple case of positive feedback, where mate selection forces the trait to grow out of control, while natural selection (increased predation, malnutrition, disease, etc.) puts on the brakes.

              Some biologists are scandalized by the notion that the trait need not be good for anything beyond attracting a mate. They think it must have some kind of positive impact on survival, if only you twist and turn just the right way before looking at it.

              I think the fact that there are so many different ways of being attractive is reason enough to conclude that the impetus was random preference. I also think there has been plenty of time for connections between survivability and preference to develop and become entangled, sometimes inextricably.

            • Tulse
              Posted May 5, 2010 at 1:05 pm | Permalink

              Thanny, you’re right that not all sexual selection has to offer a fitness advantage to the selected mate (and indeed, some have argued that in some cases the trait is selected because it disadvantages the mate, thus indicating that they are so fit they manage in spite of the handicap). With that said, however, Jerry claimed that the red “is a sign to a female of a healthy, well-fed male who would be a better father”. As I shamefully have not read his book, I don’t know if that is speculation, or if it has been demonstrated that redder males are healthier, but I was taking it as the latter.

  2. Sili
    Posted May 1, 2010 at 10:57 am | Permalink

    So they hypothesised that the genes came from bacteria, but they were wrong?


    Also: if aphids came from mushrooms, howcum there are still mushrooms?

    • DC
      Posted May 1, 2010 at 11:18 am | Permalink

      I really, really hope you’re being sarcastic. Or maybe just silly, as your handle suggests 😉

  3. Torbjörn Larsson, OM
    Posted May 1, 2010 at 11:03 am | Permalink

    I thought even bacteria didn’t acquire much genes by LGT, I believe I’ve seen numbers of merely ~ 1 % of genes estimated.

    • artikcat
      Posted May 1, 2010 at 11:27 am | Permalink

      Couple of decent reviews: Prokaryots: Gogarten JP, Doolittle WF, Lawrence JG (2002) Prokaryotic evolution in light of gene transfer. Molecular Biology and Evolution 19: 2226–2238

      Eukaryotes: Keeling PJ, Palmer JD (2008) Horizontal gene transfer in eukaryotic evolution. Nature Reviews Genetics 9: 605–618.

  4. Don
    Posted May 1, 2010 at 6:24 pm | Permalink

    Altogether as an aside, let me say thanks for this piece. Tyler Jarvik is the son of my college roommate, Rob Jarvik, inventor of the Jarvik 7 heart. Tyler visited me here with his parents about 30 years ago, when we all went fishing–lat time I saw him. It’s great to discover here this kind of connection.

    • Tyler Jarvik
      Posted October 8, 2010 at 7:14 pm | Permalink

      Did this fishing expedition take place near Jackson Hole?

  5. Andrew
    Posted May 1, 2010 at 7:35 pm | Permalink

    Out of place genes like this are a classic prediction of the Common Designer model. Expect a couple of triumphant creationist articles with a bit of fancy footwork to get around that whole “30-80 million years ago” thing.

    • igakusei
      Posted July 22, 2011 at 7:50 am | Permalink

      They’ll probably just ignore the dates altogether. Dating is one of those things that most young-Earth creationists just sweep under the rug with a few blanket dismissals about unprovable assumptions.

      I have met one young-Earth creationist who has spent the better part of the last decade working on a physics model that involves a 6,000 year old universe that is supposedly “consistent with all observable data.” He’s a riot.

  6. Hempenstein
    Posted May 1, 2010 at 8:33 pm | Permalink

    Way, way, way cool, and just as expected – there was one transfer event, and from there duplications with mutations at the separate loci. Proving once more that if something happens, and it results in an advantage, there is every reason for it to be maintained. Once the DNA has been incorporated in a non-disruptive manner (as, for instance, if it was incorporated in the middle of an exon for an essential gene product, it will not stand out as foreign. The only way the organism might discriminate against it is if the codon bias between donor and acceptor species is different enough to result in poor translation of the novel gene, which might or might not matter.

2 Trackbacks/Pingbacks

  1. […] very unrelated) species.  One of the most striking examples, which I wrote about last May, was an aphid’s capture of genes from a fungus, a capture that led to an adaptive polymorphism in the aphid’s body […]

  2. […] the source of a new gene that can spread through a species that didn’t originally have it.  (I’ve previously written about work showing that an adaptive color gene in aphids was acquired not by hybridization, but by […]

%d bloggers like this: