Evidence that GHN phase bias does not constitute a framing code.

It has been suggested that a triplet repeated pattern found in coding sequences, the G-nonG-N or GHN phase bias, serves a framing function during protein synthesis. To test this idea, the framing characteristics of a highly GHN biased sequence are examined. No effects on reading frame maintenance are observed despite the use of sensitive frameshift assays. Specifically, first the GHN phase is not more accurate than the alternative overlapping phases (i.e., HNG and NGH). Second, ribosomes do not exhibit any significant tendency to slip from the alternative frames into the GHN pahse. In addition, examination of Escherichia coli programmed frameshift sites does not support roles for GHN phase bias in programmed frameshifting. Framing functions for GHN phase bias, if they occur at all, must be extremely limited.