In a handful of cases scientists have begun to pinpoint which of the parasite’s genes control their host’s behavior. Baculoviruses, for example, infect the caterpillars of gypsy moths and a number of other species of moths and butterflies. The parasite invades its host’s cells, hijacking them to make new baculoviruses. On the outside the caterpillar appears normal, continuing to munch on leaves as before. But the food it eats is not becoming more caterpillar tissue. Instead it’s becoming more baculoviruses.
When the virus is ready to leave its host, the caterpillars undergo a radical change. They become agitated, feeding without rest. And then they begin to climb. Instead of stopping in safe spots out of the way of predators, the infected caterpillars creep higher into the trees, remaining on top of leaves or on tree bark in daylight hours, when they are easily seen by predators.
The baculoviruses carry genes for several enzymes. When they’re ready to leave their host, certain genes become active in caterpillar cells, producing a torrent of enzymes that dissolve the animal into goo. As the caterpillars dissolve, clumps of viruses shower down onto the leaves below, to be ingested by new caterpillar hosts.
To Kelli Hoover and David Hughes of Penn State University and their colleagues, the climbing behavior of the caterpillars seemed like an exquisite example of an extended phenotype. By causing their hosts to move up in trees, the baculoviruses increased their chances of infecting a new host down below. To test Dawkins’s idea, they examined the genes in baculoviruses, to see if they could find one that controlled the climbing of caterpillars.
When the researchers shut down a single gene in the virus, called egt, it continued to infect caterpillar cells and replicate as before, even turning the caterpillars to goo as before. But baculoviruses without a working copy of egt could not cause the caterpillars to climb trees. It’s unlikely that many other parasites control their hosts with a single gene; an animal’s behavior is typically influenced by a number of its own genes, each contributing a small part to the sum. So it’s probable that many parasites control their hosts with a multitude of their own genes.