Abstract
N-Terminal signal peptides direct secretory and most membrane proteins into the exocytic pathway at the endoplasmic reticulum. Signal sequences can function across kingdoms. However, our attempts at translocating variant surface glycoprotein (VSG) 117, VSG MVAT7, VSG 221 and BiP from Trypanosoma brucei and gp63 from Leishmania chagasi into canine pancreas microsomes failed. On replacing the signal peptide of VSG 117 with that from yeast prepro-alpha-mating factor (ppalphaMF) the chimaeric protein was imported, indicating that the signal sequence of VSG 117 was incompatible with the protein-import machinery of mammalian microsomes. Replacement of the gp63-h-region with a hybrid composed of the N-terminal nine residues from the h-region of gp67 from Autographa californica nuclear polyhedrosis virus and the C-terminal 10 residues from the h-region of gp63 from L. major produced a functional signal peptide. Thus, the h-region of kinetoplastid signal peptides appears to be the subdomain that is non-functional at the mammalian translocon. The calculated biophysical properties and computed discriminant scores (predictive of importability of signal peptides into mammalian microsomes) of the kinetoplastid signal sequences nevertheless are similar to those of ppalphaMF and Escherichia coli beta-lactamase both of which were imported. These signal peptides are the first collection from one biological family that have been found to fail to function across a species barrier. They indicate that signal peptides are not as universally interchangeable as previously believed. Intriguingly, endoplasmic reticulum signal peptides from Leishmania and Crithidia fasciculata are reminiscent of signal peptides from Gram-positive bacteria.