Cytoplasmic tail length influences fatty acid selection for acylation of viral glycoproteins.

We report remarkable differences in the fatty acid content of thioester-type acylated glycoproteins of enveloped viruses from mammalian cells. The E2 glycoprotein of Semliki Forest virus contains mainly palmitic acid like most other palmitoylated proteins analysed so far. However, the other glycoprotein (E1) of the same virus, as well as the HEF (haemagglutinin esterase fusion) glycoprotein of influenza C virus, are unique in this respect because they are acylated primarily with stearic acid. Comparative radiolabelling of uninfected cells with different fatty acids suggests that stearate may also be the prevailing fatty acid in some cellular acylproteins. To look for further differences between palmitoylated and stearoylated glycoproteins we characterized stearoylation in more detail. We identified the acylation site of HEF as a cysteine residue located at the boundary between the transmembrane region and the cytoplasmic tail. The attachment of stearate to HEF and E1 occurs post-translationally in a pre-Golgi compartment. Thus, stearoylated and palmitoylated proteins cannot be discriminated on the basis of the fatty acid linkage site or the intracellular compartment, where acylation occurs. However, stearoylated acylproteins contain a very short, positively charged cytoplasmic tail, whereas in palmitoylated proteins this molecular region is longer. Replacing the short cytoplasmic tail of stearoylated HEF with the long influenza A virus haemagglutinin (HA) tail in an HEF-HA chimera, and subsequent vaccinia T7 expression in CV-1 cells, yielded proteins with largely palmitic acid bound. The reverse chimera, HA-HEF with a short cytoplasmic tail was not fatty acylated at all during expression, indicating that conformational or topological constraints control fatty acid transfer.