Like the rough draft of a novel, a newly transcribed pre-mRNA molecule undergoes plenty of polishing before it's fit to be read. Custodio et al. now report evidence for a previously undiscovered editing step in the production of mRNA. Cells are fussy about mRNA. They detain a would-be strand in the nucleus until enzymes cleave the 3′ end, excise introns, stick a cap on the 5′ end, and affix a tail of multiple adenines. The carboxyl end of RNA polymerase II, the enzyme that transcribes RNA, orchestrates processing by latching onto editorial proteins. This end normally carries 52 copies of a 7-amino acid sequence. By deleting different combinations of these duplications, researchers previously determined that certain repeats attract proteins that perform specific mRNA alterations. Custodio et al. engineered mouse cells to make RNA polymerase molecules with untested combinations of deletions. The protein carrying five repeats was nonfunctional. But the one with 31 repeats could transcribe a human β-globin gene and complete the four processing steps. Nevertheless, the RNA strand remained stuck at the transcription site. Its retention suggests that the pre-mRNA must pass through an as-yet undefined editing round before the cell will release it into the cytoplasm. RNA polymerase presumably draws in proteins that perform this new alteration. The researchers hope to pin down these proteins by comparing the binding partners of RNA polymerases with truncated and full-length carboxyl ends. Once the team knows the proteins' identities, they can work out their functions. Reference: Custodio, N., et al. 2007. J. Cell Biol. 179:199–207. [PubMed]