05 Apr 1999

Fred Van Leeuwen and his colleagues reported last year that some neurons in Alzheimer's patients contain mutant RNA coding for APP and ubiquitin, despite the fact that the DNA from which the RNA had been transcribed was not mutated. By what process could "clean" DNA produce "dirty" RNA?

One candidate is a phenomenon called transcriptional bypass, wherein damage to the template DNA goes uncorrected and becomes responsible for coding an erroneous base into the RNA. This can happen if, for example, the DNA base cytosine is deaminated, producing the aberrant base uracil.

Rather than correcting this error, the transcriptional machinery treats uracil as the proper, and structurally similar, base thymine. Adenine, rather than the appropriate base, guanine, is then transcribed into the RNA, introducing an error that is propagated into protein production.

Transcriptional bypass has been demonstrated in vitro, where it can efficiently generate mutant proteins, but Viwanathan and colleagues have just published the first demonstration of this process in vivo (Science 2 Apr 1999). Working in mutant E. coli cells that are unable to correct DNA mutations containing uracil, they were able to transfect uracil into the cell DNA and show that the cells would transcribe the gene (for the fluorescing protein luciferase) and produce an altered protein. In their discussion, the authors point specifically to the work of Van Leeuwen and colleagues, suggesting that transcriptional bypass could account for the mutant RNA coding for APP in Alzheimer's patients.—Hakon Heimer

(See an accompanying editorial by Bridges in the April 2 issue of Science for discussion of these findings. Read also the Online Journal Club discussion with Van Leeuwen and colleagues.)