14 April 2010

Mind controlling an ant... with a fungus

ResearchBlogging.orgRecently, I talked about a wasp that is able to use its venom to turn a cockroach from a freely moving individual into a soulless zombie. It’s a feat that stretches your imagination of how one animal could exert that much control over the nervous system of another.

As it happened, I cam across a paper that might just go one better. Another insect is the victim (an ant, this time), but the Svengali isn’t another animal. It’s a fungus.

The fungus is Ophiocordyceps unilateralis, here pictured growing out of a very, very dead ant. Andersen and colleagues wanted to test one of Richard Dawkins’ famous ideas: that the idea of the “extended phenotype.” After infection, how precise is the ant’s subsequent change of behaviour? And can it be shown that this actually benefits the infecting agent in a way much greater than chance?

This fungus infects an ant species (Camponotus leonardi) that lives high up in the forest canopy, many meters above ground. But the corpses of infected ants were in very precise locations that were very different from the location of the ants in the colony:

  • Jaws clamped into the vein of a leaf
  • Close to the ground (but not on it)
  • On the north-northwest side of a tree

It’s so precise, you almost can’t help but imagine the inside of the head of an ant, with a fungus sitting behind a steering wheel. The authors were able to show that this position was particularly advantageous to the fungus by moving the corpses of the ants to different locations, which quickly resulted in the bodies being lost to the wild.

The authors also did some detailed analysis of how the fungus spread through the ant’s body after the ant died. This is fascinating stuff, I’m sure... to a mycologist. But one key point is that the fungus can’t start reproducing until some days or weeks after the ant has died.

For me, this raises so many interesting neurobiological questions. Some seem fairly easy to explain, in principle; ending up on the north side of a plant is probably due to changing the animal’s response to light. The final “death bite” of the ant into a leaf seems particularly tricky to explain. What chemicals are the fungus releasing into the ant’s brain that is leading the ant to such specific locations, and perform such exacting behaviours?

And if all of that wasn’t enough specificity: there’s a few other ant species in the are (Polyrhachis spp.) that the fungus can infect, but rarely does. When those animals are infected, their corpses are found significantly higher in the trees than the preferred species. This hints that the fungus can’t control this ant species so well as the one it usually infects. Is this a difference in the brains of the ant, where the fungus infects, the cocktail the fungus produces, the immune response of the hosts, or something else?

Answering some of these neurobiological questions may be a bit tricky. Ants aren’t the easiest things to do neuro on – very small. But there will be some great questions for the masochistic scientist willing to do it.

Reference

Andersen, S., Gerritsma, S., Yusah, K., Mayntz, D., Hywel‐Jones, N., Billen, J., Boomsma, J., & Hughes, D. (2009). The life of a dead ant: The expression of an adaptive extended phenotype. The American Naturalist, 174 (3), 424-433 DOI: 10.1086/603640

Photo by myriorama on Flickr, and used under a Creative Commons license.

4 comments:

Anonymous said...

Fascinating, if unsettling. I'm steering clear of evil fungi.

Dan said...

Another great tale of the parasite. Keep them coming. I was still more disturbed about the wasp though. Don't know why?

Zen Faulkes said...

I think the fungus is less disturbing because it has the common decency to kill its host first. Leaving the roach alive is the freaky bit.

The Alabamycologist said...

You're darn tootin' it's interesting to a mycologist! Entomophthora muscae does a similar trick to houseflies. It's a Zygomycete, which is an entirely different phylum of Fungi, so the phenotype has evolved multiple times independently