Abstract
Proteins destined for the nucleus contain nuclear localization sequences, short stretches of amino acids responsible for targeting them to the nucleus. We show that the first 29 amino acids of GAL4, a yeast DNA-binding protein, function efficiently as a nuclear localization sequence when fused to normally cytoplasmic invertase, but not when fused to Escherichia coli beta-galactosidase. Moreover, the nuclear localization sequence from simian virus 40 T antigen functions better when fused to invertase than when fused to beta-galactosidase. A single amino acid change in the T-antigen nuclear localization sequence inhibits the nuclear localization of simian virus 40-invertase and simian virus 40-beta-galactosidase in Saccharomyces cerevisiae. From these results, we conclude that the relative ability of a nuclear localization sequence to act depends on the protein to which it is linked.