“There is a dogma that translation only occurs in the cytoplasm,” says Miles Wilkinson (M.D. Anderson Cancer Center, Houston, TX), “but there's really no hard evidence—there's just a dogma.” Cook did not set out to disprove this dogma—his original experiments involved a high-resolution method for localizing protein synthesis in the cytoplasm. But to his surprise, he saw evidence for nuclear translation. He went back to the literature, found 30-year-old papers supporting his findings, and set out to improve upon them.
Cook used two different labels to modify lysine tRNAs. In permeabilized cells, the labeling lit up discrete nuclear sites and was sharply reduced by the addition of cycloheximide. Similar labeling was obtained both in the presence of an inhibitor of nuclear transport that blocked the entry of a labeled dextran, and with isolated nuclei that lacked >95% of cytoplasmic ribosomes.
“Nuclear translation is certainly an exciting possibility,” says Lynne Maquat (University of Rochester, Rochester, New York). “What remains to be resolved is the nature of the template,” such as whether it is a completed and spliced transcript. Cook claims that transcription and nuclear translation are coupled, as in bacteria. He bases this conclusion on the reduction of nuclear (but not cytoplasmic) translation after the addition of an RNA polymerase II inhibitor, and suggests that the ribosome is following on the tail of the polymerase, scanning for errors.
Wilkinson agrees, at least in theory. “I don't think he has very good evidence that [transcription and translation] are coupled,” he says. “But I'm assuming the purpose of this is not to make protein, but to scan messages for problems.”
For now, Cook has enough controversy with the central idea of the paper: that nuclear translation occurs at all. “Either there is something we haven't thought about, in which case we have egg on our face, or those guys were right 30 years ago,” he says. ▪