Abstract
It has been proposed that cauliflower mosaic virus 35S RNA with its 600 nt long leader uses an unusual translation process (the translational shunt). A wheat germ in vitro translation assay was used to improve the study of this mechanism. Deletions, the introduction of stable stem-loop structures, and the inhibitory effect of antisense oligonucleotides on gene expression were used to determine the roles of various parts of the leader. It was found that the 5'- and 3'-ends of the leader are absolutely required for translation whereas the middle part is apparently dispensable. These results confirm the data already reported from transient expression experiments with protoplasts. However, the in vitro data suggest in contrast to protoplast experiments that only two relatively short regions at both ends, approximately 100 nt each, are required. The in vitro system provides tools for further studying the shunt model at the molecular level and for examining the involvement of proteins in this mechanism. Shunting was also found to occur with the rice tungro bacilliform virus leader. As wheat is neither a host plant of cauliflower mosaic virus nor rice tungro bacilliform virus, the shunt seems to be host independent, a finding that deviates from earlier studies in protoplasts.