Nongenetic selection and evolution: flies use bacteria to adapt to parasitic worms

July 10, 2010 • 10:37 am

Evolution by natural selection requires three things:

1.  A trait shows variation

2. The variation in that trait must be capable of being passed on from parents to offspring (i.e., the variation is “heritable”)

3.  The variation in that trait must make a difference in its likelihood of being passed on from parents to offspring. (Usually, but not always, this requires that the trait affect the survival or reproduction of its carrier.)

If all this is true, those forms of a trait that are better at proliferating will gradually increase in a population.

Although these statements are the basis of Richard Dawkins’s book The Selfish Gene, do note that the above characterization doesn’t use the word “gene.”  Even if a trait has a nongenetic basis, it will evolve, via a form of natural selection, if it satisfies these conditions. Much of cultural evolution works in this way, although there are profound difference between cultural evolution based on cultural natural selection (or selection among “memes”) and biological evolution based on genes.

A new paper by John Jaenike and his colleagues in Science, however, shows a form of biological evolution by natural selection that isn’t based on changes in genes. It’s based on changes in the presence of symbiotic bacteria that protect a species from parasites.

The species in which the nongenetic evolution has occurred is the mushroom-feeding fruit fly Drosophila neotestacea in North America:


Fig. 1. Drosophila neotestacea

In the eastern U.S., this fly is afflicted by a parasitic nematode, Howardula aoronymphium (we’ll call it “the worm”), that renders females sterile, reduces the mating success of males, and reduces the survival of both sexes. (Fig. 2 gives a pretty disgusting picture of a worm-infected fly.) The nematode enters the fly larva when both are the mushroom, and persists in the adult fly.  Females, attempting to lay eggs then pass the nematode into the mushroom, where it mates and enters other fly larvae, completing the lifecycle.

Fig. 2.  D. neotestacea female dissected to show reproductive tract and its huge load of parasitic worms. Photo by J. Adam Fenster, University of Rochester

Based on genetic evidence, the nematode appears to have colonized North America from Europe fairly recently  (American worms have much less genetic variation than European ones, but represent a genetic subsample of them). But the worm is present in fly populations throughout North America, and in every population about 20% of the flies are infected with worms.

Some flies also carry another organism: the bacterial symbiont Spiroplasma, which is found in many insects.  In D. neotestacea, however, the presence of Spiroplasma protects the fly from the sterilizing effects of nematodes.  While flies with worms and no Spiroplasma are virtually sterile, the presence of the bacteria confers almost normal fertility on worm-ridden flies.  It’s not yet clear how this works, but worms in flies with Spiroplasma are much smaller than those without the bacteria. Presumably the bacteria does something to the worms (or to the flies) that makes the worms grow much more slowly.

So this is a good setup for natural selection. First, there is variation in a trait—some flies have Spiroplasma, others do not. That trait is heritable, for Spiroplasma are transmitted directly from mother to offspring in the egg (there’s no “horizontal” adult-to-adult transmission).  And there’s an environmental factor—the parasitic, sterilizing worms—that cause differential reprodution of flies depending on whether or not they carry Spiroplasma.  Those flies who carry Spiroplasma can still produce offspring, and hence pass on the Spiroplasma to the next generation; those flies who don’t carry the bacteria don’t get protection from nematodes, and leave no (or very few) offspring.  In the presence of worms, then, there’s a huge selective advantage in flies to carrying bacteria.

You’d expect, then, that in the presence of worms, the proportion of flies that carry bacteria would increase over time. This is a form of natural selection, but it’s selection on flies that carry bacteria.  Flies with bacteria are the ones who reproduce; they pass those bacteria on to their offspring, and so the proportion of bacteria-carrying flies goes up.  There’s no difference in the DNA of flies who reproduce and those who don’t, so the flies’ genomes themselves are not evolving.

Nor do the bacteria seem to be evolving, although there could theoretically have been mutations in the Spiroplasma that make them kill or inhibit nematodes (after all, any bacteria who could do this would leave more offspring).  But population-genetic studies of the bacteria suggest that this hasn’t happened.  The bacteria seem to confer resistance to nematodes simply through some innate feature of their biology.

What Jaenike et al. found was that the predicted selection occurred, and caused evolutonary change: the proportion of Spiroplasma-infected flies indeed seems to be increasing  within populations in the eastern U.S..  Moreover, the bacteria has started to spread from the eastern to the western U.S. in only three decades.  Here’s the evidence:

  • Museum specimens of flies collected in the eastern U.S. in the early 1980s show no Spiroplasma (you can screen museum specimens for bacterial DNA).  But now, in those same populations, the frequency of flies that carry bacteria ranges from 50% to 80%.
  • Worm-ridden flies collected in New York in the 1980s were largely sterile, having the same fertility profile as modern flies that are worm-ridden and don’t have the bacteria.  But in modern populations, most flies do carry the bacteria and so are fertile even if they have worms.
  • There’s a “cline” (a geographical gradient) in the presence of the bacteria from eastern to western North America.  As I mentioned above,  the proportion of flies with bacteria is high (50-80%) in the east, gets lower (10-25%) across the Great Plains, and is at or near zero in coastal British Columbia.  Given the presence of worms in all of these locations, there would be a strong advantage for the western populations of flies to acquire the bacteria too.
  • Finally, genetic evidence suggests that flies in western North American haven’t yet reached their equilibrium degree of bacterial infection (it should be about 0.8—it’s not complete at equilibrium because transmission of bacteria from mother fly to her offspring isn’t perfect).

All this suggests that the bacteria are in the process of sweeping from east to west in the flies through a natural selection-like process.  All populations of flies have worms, and should thus have bacteria, but the bacteria are just beginning to go west.  This is analogous to a new adaptive mutation working its way through a species.  And you can make a prediction: in another two or three decades, all western populations of flies should have a high frequency of Spiroplasma.

At the end of their paper, Jaenike et al. note that this example may suggest strategies for wiping out nematode-caused diseases like river blindness and filariasis (which produces the grotesque swelling of limbs called elephantiasis).  Those diseases, too, are transmitted by worm-ridden flies, and perhaps deliberate infection of those flies with bacteria like Spiroplasma could reduce the transmission of nematodes and help wipe out these diseases.

Here, then, we see how a species (the fly) has adapted to an environmental challenge not by changing its genes, but by acquiring a whole genome—the Spiroplasma genome.  It’s as if the whole bacterium was an adaptive mutation.  And we also have a case of selection in action, one that makes testable predictions. Creationists may dismiss it since it’s not “evolution” in the traditional sense, but it shows the principles of natural selection in any meaningful sense.


Jeanike, J., R. Unkless, S. N. Cockburn, L. M. Boelio, S. J. Perlman.  2010.  Adaptation via symbiosis: recent spread of a Drosophila defensive symbiont. Science 329:212-215

21 thoughts on “Nongenetic selection and evolution: flies use bacteria to adapt to parasitic worms

  1. A new paper by John Jaenike and his colleagues in Science, however, shows a form of biological evolution by natural selection that isn’t based on changes in genes.

    But it is a change of genes. It happens to be in a different body type that is acquired.

    Hmmm, no mention of red state/blue state effects on flies, bacteria or worms. 😉

  2. It doesn’t seem that surprising that a bacterium that gets into the next generation through a fly’s gametes would work hard to undermine a parasite that causes fly sterility. It’s super cool though! Isn’t this a perfect example of the extended phenotype?

  3. I have tough time seeing this as an “alternative form” of evolution or as “non-genetic selection”.

    At some point in the past, the flies and/or the bacteria developed some combination of genetic mutations which conferred this resistance to nematodes. It is probable that prior to the introduction of the nematodes, these mutations were selectively neutral. When the nematodes were introduced, these particular genes (whether in fly, or bacteria, or both) suddenly had a huge advantage that did not previously exist.

    In other words, what we have here is a sudden change in the environment which radically altered the selection pressures on existing genetic variations. This is why there is “no change in the genomes of fly or bacteria”. Natural selection *always* acts on pre-existing variations. New mutations cannot be selected for (or against) until after they come into existence. And the survival value of “old” traits does not remain static: i.e. thick fur loses it’s value and possibly even becomes a liability when the ice age ends and the climate warms.

    How is this example any different from the evolution of any symbiotic relationship, other than perhaps the sudden and serendipitous nature of the positive selection for whatever genes in the fly and bacteria provide the protection? Suppose we discover that there is a subpopulation of humans who possess mitochondria which are somehow 20% more efficient in producing energy and we can can trace the rapid spread of this subpopulation over the last few centuries. Is this suddenly an alternative mechanism of inheritance and selection?

    If there is no “horizontal” transmission of bacteria, this seems to imply that a strong, reproductively bound, symbiotic relationship already existed. Maybe not as tight as with mitochondria, but qualitatively similar. The process of mutations/natural selection leading towards or away from symbiotic dependence was already ongoing. This sudden change in the environment has simply altered that game so that some particular set of genetic variations in fly and bacteria are now favored.

    So what do we have here?

    The flies pass on their traits to the next generation via their genes.

    The bacteria pass on their traits to the next generation via genes.

    Circumstances in the environment favor some traits over others, resulting in variations in reproductive success.

    The traits which are most successfully reproduce in successive generations are becoming more prevalent in the populations.

    This sounds like plain vanilla evolution to me.

    We may envision the evolution of symbiotic relationships to ordinarily be “slow” with gradual changes in both species that make them more and more interdependent. But such is what we normally expect with all evolutionary adaptations. But when the environment changes suddenly, so too the selective benefit of certain traits may suddenly increase. And that is what has happened here: a rapid change in environment that has drastically increased the interdependence of these two species (actually, apparently the bacteria were already wholly dependent on the flies — if the only way they could be passed is from mother to offspring).

    1. “The flies pass on their traits to the next generation via their genes.”

      No, the trait in question (relative immunity to the parasite) is passed on by inheritance of the bacterium.

      The selection here is non-genetic in the sense that the selected trait (immunity) is conferred by the presence of the bacterium, not by genes.

      The evolution in question is not the evolution that gave the fly and bacterium the ability to have this symbiotic relationship in the first place. It’s the spread of the immunity trait (via the bacterium) through the population of flies by inheritance. This is analogous to the spread of a successful gene through a population. Moreover, the ability to have this symbiotic relationship appears not to be an adaptation. The flies can’t have adapted to it, since most of them are currently encountering the bacterium for the first time. And Jerry writes that population-genetic studies appear to show that the bacterium hasn’t adapted to this relationship either. It was apparently a co-option of “some innate feature of their biology”.

      Hope that helps.

      1. “No, the trait in question (relative immunity to the parasite) is passed on by inheritance of the bacterium.”

        No. The trait in question is carried in the genes of the bacterium (and there may be genes in the fly that enable this effect in some way). Just because the bacteria are “only passed from mother to offspring” does not negate the fact that the selection is on the gene(s) responsible for the immunity.

        Consider. What if a mutant bacterium arises which fails to confer the immunity? Their host will die off and therefore those bacteria will die off. In other words we have selection presure favoring whatever genes in the bacteria provide this benefit to its host. What if a mutation in a fly caused its immune system to kill off the helpful bacteria? The fly and its descendants might be lucky for a few generations, but will ultimately die out. Again we have selection pressure for genes in the fly which favor this particular partnership. Or what if a mutation in the fly left the bacteria alone but produced some variant of a hormone or enzyme that somehow neutralized the protective effect of the bacteria? Again those mutated genes are selected against. What if a mutation in the fly or bacteria increased the probability of a mother passing the bacteria to her young? That would be a selectively favored mutation. At all times natural selection favors those genes in both the fly and bacterium which inhibit nematode growth, and natural selection will favor genes that enhance this effect. Likewise any genetic variants of fly or bacteria which diminish the protective effect will be selected against.

        The only quirk here is that apparently the bacteria are only passed from mother to offspring. So in a completely trivial sense the bacteria are “carriers” of the trait. It is still ultimately genes which are responsible for the effect, and perpetuation of the ‘trait’ depends upon inheritance of the appropriate genes in both the fly and the bacteria. The exact same thing is true of mitochondria, which are no longer independent organisms, but they once were.

        Furthermore, given the untold billions or trillions of these flies, it is doubtful that the bacteria “never” pass “horizontally” from fly to fly. At best they can place an experimental upper bound on the frequency at which this happens. Given that the population of lab flies is a minute fraction of the wild population and only observed for a limited duration, there could easily be rare “horizontal” events in the wild.

        1. ..we have selection presure favoring whatever genes in the bacteria provide this benefit to its host.

          Agreed. And, as you say, we have selection pressure on the fly genes which sustain their infection by the bacteria.

          So natural selection acts to stabilize the bacterial and fly genomes.

          Yet evolution by natural selection is happening, because the trait of being infected by bacteria is spreading in the fly population, by virtue of superior reproductive success.

          How can the observed evolution of the fly population be happening by natural selection on genes, when natural selection is acting to stabilize the genome?

          1. Natural selection is not acting to “stabilize the genome”. What does that even mean? Perhaps you mean that it is stabilizing the gene pool(s) of fly and bacteria? That is also not the case, but if it was the case, then that would just be an example of natural selection preserving the locally optimal or evolutionarily stable distribution of genes within the population(s). That is natural selection acting on genes. In other words: ordinary selfish-gene evolution.

            Now what is really happening to the gene pool(s)?

            First the fly gene pool. Flies not infected with spiroplasma are dying off. It would be ridiculous to suppose that there is no variation in the degree of ‘friendliness’ of the flies immune systems and other factors to the bacteria. Ergo, right now, natural selection is favoring genes which make the flies bodies friendlier to the bacteria this could be anything from slight effects on body temperature to what the fly likes to eat to various enzymes or immune factors that do or do not attack the bacteria. The effects of these genetic variation may be very small, but given billions of flies with their rapid life cycle, even such small effects are almost certainly having an impact.

            Now the bacteria gene pool. Prior to the nematode invasion, the particular strain(s) of spiroplasma that confer the immunity were among a wide variety of other strains and species of bacteria that lived in or regularly infected the flies. The population of this particular strain of spiroplasma is obviously increasing rapidly and out-competing most if not all of it’s bacterial competitors (not within a single fly body, but in the larger population and environment — the spiroplasma are winning because they keep their hosts alive where other bacteria fail to do so). This means that the spiroplasma genes, in particular those genes which confer the nematode resistance, are rapidly increasing in frequency within the larger gene pool of “all bacteria that inhabit the flies”. It is perfectly appropriate to look at this larger gene pool since bacteria are well know for “horizontal” inheritance. It is entirely possible (and becoming more probable as the spiroplasma population becomes dominant) that the genes for nematode resistance could be horizontally transferred to other strains or species of bacteria, some of which may be communicable. Then suddenly inheriting the spiroplasma from the mother will become much less important for fly survival. But having the right “nematode resistance” genes remains very important for the bacteria.

            The point here is that far from stabilizing the ‘genome’ or gene pools, the nematode invasion has caused a sudden shift in the selection pressures on those gene pools. It is the genes that confer the resistance which ultimately matter.

            Yes, it is trivially true that inheriting the bacteria from the mother is, in a sense being selected for. But beneficial bacterial symbionts inherited exclusively (or nearly so) through the maternal line is nothing new. For that matter inheriting non-genetic (i.e. environmental) factors from parents is nothing new, though the exclusivity of passing the environmental factor from parent to offspring may not be as strong. But consider a case where some population of small birds are blown by a hurricane to some remote islands, each island too far from the others for the birds to fly. Some islands have better food. Others have worse food or no food. The offspring of each parent inherits its island from the parents. The birds on the food-poor islands die out after a few generations. Are we going to call island residence a carrier of inheritance for the “better food to eat” trait?

            With the exception of extremely mobile species, inheritance of local geography from exclusively from parents is the rule. And the parents living in the most favorable regions will have more descendants. We are free to call this non-genetic inheritance, if we want to. But it just muddies the waters. “Infection by spiroplasma” looks more like a non-genetic form of inheritance than “geographic location”, but it is still misleading to do so, as I have already explained it is still the genes which ultimately matter, especially in the long term.

            And as I have pointed out, in terms of evolutionary theory, nothing new is happening here. Exclusive matrilineal inheritance of beneficial microbial symbionts? Not new. Why weren’t we calling this an “alternative form of inheritance” previously?

            The result in this paper is very interesting and will be very significant for medical research… but the claim of a “new/alternative form of evolution/inheritance” just seems to me like hype to get more attention (however it is also entirely likely that the authors are completely sincere and do not realize that in terms of evolutionary theory, nothing new is happening.

      2. “‘The flies pass on their traits to the next generation via their genes.’

        No, the trait in question (relative immunity to the parasite) is passed on by inheritance of the bacterium.”

        I want to come back to this. “Trait” is a useful concept, but it is not fundamental. Is nematode resistance a ‘trait’ of the fly? Or a ‘trait’ of the bacterium? Is it an emergent property resulting from their interaction? Regardless of what you decide to call a trait, and which organism to which the trait is attributed, the underlying fact is that the genes of the fly determine it’s properties, including how it interacts with it’s environment (that I have scar on my hand is not a ‘trait’ programmed by my genes, but the proptery of “scar formation in response to wounds” is a trait determined by my genes). Likewise the genes of bacteria and nematode determine how they interact with their environments. The suppression of the nematodes within flies hosting spiroplasma is the result of some interaction among the three organisms. This interaction is determined by the genes of those organisms.

        The fly has the trait of “immune system does not attack and kill spiroplasma”. This trait is determined by and passed on through its genes. It is a very important trait to avoid death by nematode.

        It is genes that matter here and it is genes that are being favored, or not, by natural selection. The fact that the relevant genes were all in place before the nematodes arrived, and the fact that these genes serendipitously have the effect that some interaction among fly, bacteria, and nematode suppresses nematode developement does not make the genes involved any less significant or fundamental.

        Another way to look at this. The bacteria are going to produce occasional mutations and variants of varying degrees of success. But if the genes that give them the nematode suppressing ability mutate in such a way as to disable this effect, then those strains of bacteria will die out with their hosts. But other strains of bacteria that preserve those genes but change other genes will (all else equal) survive just as well as the current strain. So, again, we see that natural selection is favoring the particular genes involved.

        One last possibility to ponder. It is conceivable (however rare or unlikely) that a virus could come along that happens to snip the relevant genes out of a bacterium and splice it into a germ line cells of the fly, thus giving the fly’s descendants immunity from the nematodes without needing the bacteria, immunity that can be passed through males. These independently nematode-resistant flies will do just as well, maybe better than the spiroplasma-dependent flies. In any case, we once again see that it is the genes that matter.

    2. P.S. I can see how it feels a bit strange to call this evolution. We’re merely talking here about the spread of one trait through the population, and saying nothing about how the trait arose. We tend to think of evolution as a continuing process of variation, selection, variation, selection, etc. Here we’re just looking at one round of selection.

      1. “P.S. I can see how it feels a bit strange to call this evolution.”

        I have no problem calling what is going on evolution. I have a problem calling is a new or alternative form of evolution, or a new form of inheritance. There is nothing ‘new’ here in terms of evolution.

        What do we have?

        A bacterial symbiont providing a benefit to its host. Not new.

        Microbial symbiont passed exclusively (or almost exclusively) from mother to offspring. Not new.

        A sudden change in the environment which drastically changes the selective preference for a particular trait (which means favoring the gene(s) responsible for that trait, wherever they are located). Not new.

        Genes serendipitously providing a benefit in a completely new situation. Not new.

        Selection pressure against any variants/mutations in genes of fly or bacteria which reduce survival of either or both in the new environment. Not new.

        Selection pressure in favor of any genes of either symbiont which further enhance survival in the new environment. Not new.

        What, in terms of evolution, in light of the “selfish gene”, is new here?

  4. This is really cool to know. I wonder how many bacteria developed in us this way, or something like this. I think worms are disgusting and never would knowingly eat anything with worms. Is disgust a genetic quality that keeps us from liking other peoples body fluids and of course worms?

    1. Disgust is probably a combination of evolved and cultural features. Since many things that disgust are not innate, but learnt, and vary with culture. However some things are near universally disgusting, and the reaction has a biological basis (i.e. you need certain genes to learn to become disgusted – I wonder if insects that are coprophagous share those genes with us).

  5. An assumption from the article is that the bacteria, Spiroplasma, does not live in the nematodes, there is no mention that any testing was done on the worms. Did I miss something in the article?

    The bacteria as catalyst for the continued survival of both Fly and Worm

  6. From the article, the claim was made:

    “… it shows the principles of natural selection in any meaningful sense …”

    While I agree with this statement, it seems like a non-starter. Is the challenge to evolution on the grounds of natural selection alone or that random mutation *plus* natural selection leads to descent (new species)? Many would grant you natural selection … but that’s not the issue. There seems to be lacking evidence for “whole enchilada” (descent by modification through natural selection and mutation) … there is no repeatable, observable scientific evidence for the phenomenon.

  7. Where in ‘The Selfish Gene’ does Dawkins make the point at the opening of this post? I skimmed through the book but didn’t see it.

Leave a Reply