Without plasma membrane proteins, every cell would be an island, unable to communicate with the outside world. So how does the cell ensure that newly synthesized plasma membrane proteins reach their destination? Proteins that don't fold properly or aren't correctly glycosylated are held back in the endoplasmic reticulum (ER), and an emerging mechanism that prevents the untimely escape of proteins from the ER is an ER-retention signal, RXR(R), which has to be masked by another protein to allow exit from the ER. But Ma and colleagues, reporting in the 12 January issue of Science, have found a new signal that, instead of holding proteins back, pushes them out.

The authors stumbled on this mechanism when they were trying to express different K+ channels in Xenopus oocytes. Two of the channels, Kir1.1 and Kir2.1, were efficiently expressed at the surface but others were not. So they swapped the carboxyl termini of the poorly expressed channels for those of the efficiently expressed channels, and got efficient expression. Conversely, removal of the C terminus of Kir2.1 beyond residue 374 reduced trafficking to the surface. This did not seem to be due to a fault in folding or assembly, because the small amount of surface-expressed truncated Kir2.1 had identical conductance properties to wild-type Kir2.1. Surface levels could also be restored by expression of full-length Kir2.1 together with truncated Kir2.1, indicating that the two forms could assemble together to create functional channels.

Scanning mutagenesis narrowed the signal down to FCYENE: even conservative mutations in the bold residues abolished surface expression of green-fluorescent-protein (GFP)-tagged Kir2.1. The position of the sequence seemed unimportant: it worked when inserted between GFP and Kir2.1, or at the extreme C terminus of Kir2.1. But does the export sequence work for other proteins? A truncated form of Kir3.1 normally remains stuck in the ER but adding FCYENE to its C terminus allowed it to escape. Likewise, FCYENE facilitated the export of a distantly related K+ channel, Kv1.2, which is normally helped to the surface by its partner, Kvβ2.

So which signal is the strongest if a protein contains both a FCYENE signal and an ER retention signal? On the cell's factory floor, safety and quality control are of paramount importance, so the ER retention signal overrides the export signal, presumably to prevent the plasma membrane from being flooded with misfolded proteins.

Is FCYENE the only signal for export? Apparently not; Kir1.1 has an unrelated sequence at its C terminus — VLSEVDETD — that also does the trick, and the vesicular stomatitis virus G-protein has a DXE motif that is thought to speed its transit to the cell surface. No doubt these hint at more to come.