As for a possible mechanism of how the reading frame changes,
I can provide the following speculation. There is a special ribosomal
complex that recognizes the start codon (AUG) to initiate translation.
It is not inconceivable that this complex could be post-transcriptionally
altered so that it recognizes something different than AUG as the start
codon, thus changing the reading head. Or it recognizes AUG, but
due to the post-transcriptional modification its own bulk gets in the
way and the next nucleotide is covered up and not seen by the other
t-RNAs that come along bringing the amino-acids. There are a lot
of possibilities....
In general, there is a lot of post-transcriptional and post-
translational stuff going on in a cell. It isn't all just genome.
Kampis makes the analogy between reading frame in biochemistry
and computers. It is the hardware that controls the reading
frame in computers (the word length in bits if you like). In
biochemistry, there may be post-transcriptional and post-
translational events that change the reading frame. In computers,
there is only a single reading frame. The software can't change the
hardware in this way. If it could, then it would be something
quite a bit more interesting than the machine I'm typing on at
the moment.
Jeff Prideaux