« "Bigger splash in the gene pool" made a splash on the web | Main | Local food? »

Evolution-proof pest-resistant crops?

This week's paper is "Alarm pheromone habituation in Myzus persicae has fitness consequences and causes extensive gene expression changes", published in PNAS by Martin de Vos and others.

Aphids suck. This wouldn't be too big a problem for their host plants, except that they sometimes transmit viruses. Some plants repel these pests by giving off gases very similar to the chemical alarm signals aphids release when attacked by predators. Could crops be genetically engineered to do this? Probably, but would it work, or would the aphids evolve to ignore these signals and keep on sucking?

To answer this question, the authors studied aphids on plants genetically engineered to make aphid alarm signal.

The first generation of aphids on the false-alarm plants were often repelled by them, and also reproduced less. Maybe they were feeling stressed. But the second and third generations were about 50% and 25% as likely to be repelled.

So far, this looks like another example of pests quickly evolving resistance to our pest-control measures. But the alarm-ignoring aphids were twice as likely to get eaten by predators. Normally, they would detect alarm signals when other aphids nearby get eaten, and fly away or drop off the plant, escaping the predators. Ignoring the false-alarm from the plant "crying wolf" increased their chances of getting eaten. So maybe this approach would be relatively evolution-proof.

(I suggested as much in the draft of my book, "Darwinian agriculture: where does Nature's wisdom lie?" which I'll now have to revise to include this interesting paper. So far, Princeton University Press has sent me one detailed and positive review and one sketchy and negative review. They need another review, so the book will be delayed a bit.).

So what happened from generation 1 to generation 3? The authors don't think it was evolution (i.e., a change in the genetic composition of the population) because it happened to fast and because they could reverse the effect with three generations on plants that didn't make false alarm signals. I'm not so sure. With strong selection and some genetic variability in the population, changes at least as big as they found can certainly evolve.

They found changes in gene expression, but I don't see how that answers the question. An evolutionary change will usually lead to differences in gene expression, but you can also get differences in gene expression without evolution.

Two unusual features of aphids are (1) that they usually reproduce by parthenogenesis, with females giving birth, without sex, to daughters genetically identical to themselves (apart from any mutations) and (2) that adult females already contain tiny daughters and granddaughters. So here are two possible explanations for the changes observed:

1) Evolution. One member of the first generation had a mutation making her ignore alarm signals. Since responding to false alarm signals reproduces reproduction, she had more daughters and granddaughters than the others, increasing the frequency of alarm-ignoring genes in the population.

2) "Maternal effects." There are many examples of the environment to which a mother animal or plant is exposed affecting their offspring. Sometimes this involves some modification of their DNA, such as by methylation (see "epigenetics"). (The authors use the term "habituation", which I think it usually refers to a behavioral change in an individual, rather than a change over generations.)

Maternal effects aren't always beneficial, but they were in this case. Is this just chance, or did this tendency of daughters and granddaughters to ignore alarm signals if their mother or grandmother were exposed to false alarm signals evolve? Such responses rarely evolve "because they might be useful in the future" -- they must have been useful in the past. I would only expect such maternal effects to evolve if responding to alarm signals was sometimes beneficial and sometimes harmful over the evolutionary history of this species, and if conditions to which the mother was exposed were a consistently good predictor of the response that would be most beneficial to the daughter. This isn't impossible, but I wouldn't rule out the simpler alternative that the populations under study evolved over three generations. Maybe the authors have additional data that would disprove this, however.

It's also worth mentioning that parthenogenesis leads to clusters of genetically identical aphids. This genetic uniformity makes cooperation evolve more easily (e.g., alarm signals to warn others nearby when you're being eaten) than if these were aphids of different genotypes competing for resources.


// Alert: Enthusiastic Amateur that wants to rub shoulders with experts

** An intelligent comment! Welcome, though I know more about plants than about insects. See specific comments below -- Ford

Regarding the possibility of an evolved 'maternal affect' for the alarm signals: Could flight distance be worth considering?

Let's say that a given aphid finds itself in an environment where there's alarm signal around a lot of the time. That aphid will spend a lot of time fleeing. Time and energy spent fleeing is time and energy spent not feeding or reproducing.

Is it unthinkable that in times of heavy-alarm-signal, it might pay to reduce the flight distance, and in times of low-alarm-signal, it might pay to increase it?

** Seems plausible, although aphids don't necessarily have wings. Apparently they grow them when they're crowded or the plant they're on is less nutritious.
Wing dimorphism in aphids. Heredity 97:192–199
influence of natural enemies on wing induction in Aphis fabae... Bull Entomol Res 98:59–62.

Also: Is it unthinkable that a sub-population of aphid could evolve a tendency towards deceptively spraying an alarm signal to which they are partially immune, thus chasing off competitor aphids from food sources.

** This would be especially effective if they had a private "back-channel" signal they could use for honest alarms. But ignoring alarms from competitors could be risky, if those competitors are really getting eaten. "Quorum sensing" in bacteria raises similar issues: it may be useful to "lie" to competitors, but how much faith should you put in THEIR quorum-sensing signals?

Depending on the dynamics, could it be possible that there's an arms-race between a sub-population of cheaters and a greater population of non-cheaters - and in both cases, it would pay to have a maternal affect that could determine how 'suspicious' (resistant to alarm signal) the daughters would turn out to be.

** Not clear what kind of cheating you have in mind.

Just speculating in the vain hopes for intellectual approval from someone who actually knows what they hell they're talking about. Yes, I am that vain. Sue me. ^_^

** Speculation can be fun. You can think of it as "hypothesis generation." The next step would be to review the relevant literature looking for data that would disprove the hypothesis (not "your hypothesis", as you don't want to get emotionally attached to something you're going to subject to destructive testing) as well as any results that seem unlikely unless the hypotheses is true.

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Type the characters you see in the picture above.