In 2006, Hannah Seidel, a graduate student in Leonid Kruglyak's lab at Princeton University, performed an experiment that hundreds of C. elegans biologists had done before: She crossed two common worm strains, and looked at the progeny. Only this time, unlike previous experimenters, Seidel inspected the Petri plates a bit more carefully. And in the second generation, she noticed scores of dead eggs.
Seidel had decided to perform the experiment after evolutionary geneticist Matt Rockman, then a postdoc working with Kruglyak, had crossed the same parental strains to construct more than 200 recombinant inbred worms for mapping genes. He was inspecting the resulting worms' genotype data when he noticed that on the left arm of chromosome I, one of the parental strains had contributed far more than its fair share of DNA: The SNP ratio was nearly 50:1, not 50:50, as he expected. Rockman "basically gave me some worms and said 'See if you can figure out what's going on'," Seidel says.
Next, the two of them were staring at a collection of dead worms. How had Rockman—not to mention the hundreds of C. elegans biologists before him who had also crossed the same two strains for basic gene mapping and slews of other experiments—missed such an obviously deadly incompatibility? "It's because the worms live so fast," Rockman, now at New York University, professes. "Unless you count embryos as soon as they're laid, the plate becomes covered with worms."
Seidel crunched the numbers and realized that almost exactly one-quarter of the F2 offspring died. According to Mendelian theory, however, basic incompatibility genes should only produce a maximum lethality of three-sixteenths. She did some back-of-the-envelope calculations and realized that a maternal gene that affects developing embryos through secreted proteins in the egg—a so-called maternal-effect gene—could explain the data. She set up a couple more crosses, and pinpointed the offender: It was, indeed, a gene with cross-generational parental effects. But it was a paternal-, not a maternal-effect gene, that operated through the fathers—an extremely rare type of gene and only the second one discovered in C. elegans.
The gene—named peel-1—bestowed fathers with baby-killing, toxic sperm. Additional fine mapping also revealed a second gene, zeel-1, which was tightly linked to peel-1 and acted as an antidote, making offspring immune to their father's lethal sperm (Science 319:589-94, 2008).
In their cross, one of the parental strains had the toxin and the antidote alleles, while the other strain lacked both. Put them together, and toxic F1 fathers were killing off susceptible F2 offspring, but in a genotypically-biased fashion, with the offspring of one strain much more likely to die. The legacy of this one-sided infanticide was what Rockman had stumbled upon in the genomes of his recombinant inbred worms.
What's more, the deadly one-two genetic punch was not restricted to the worms used in the initial cross. Rockman and Seidel tested more than 50 different strains from around the world (including strains collected by this author for his PhD), and discovered both haplotypes in equal abundance.
Two big outstanding questions remain, says Rockman: Why have both allelic versions been preserved, and what's the modus operandi of the toxin and antidote? "It continues to be mysterious," he says.
"We can see these signatures of what must be selection and the population genetic process," says Patrick Phillips, an evolutionary geneticist at the University of Oregon. "But without understanding the ecology it will be difficult to get at the causation."
Seidel is making headway with the problem, though. She showed that the toxin encodes a small protein that is localized to the sperm cell membrane, while the antidote is activated at mid-embryogenesis, around six hours after the toxic sperm enters the egg. Without the antidote, however, embryos die around two hours later.
Importantly, this implies that the sperm's toxin protein stays intact for at least eight hours. So the big puzzle now, Seidel says, is how the protein remains latent but not broken down for several hours. She suspects that the toxin might be highly resistant to degradation; enough so that it manages to disrupt muscle development many hours later. "It's not impossible," she says, "but no one has ever seen a sperm protein stick around for that long."