Most missense errors have little effect on protein function, since they onl
y exchange one amino acid for another. However, processivity errors, frames
hifting or premature termination result in a synthesis of an incomplete pep
tide. There may be a connection between missense and processivity errors, s
ince processivity errors now appear to result from a second error occurring
after recruitment of an errant aminoacyl-tRNA, either spontaneous dissocia
tion causing premature termination or translational frameshifting, This is
dearest in programmed translational frameshifting where the mRNA programs e
rrant reading by a near-cognate tRNA; this error promotes a second frameshi
fting error (a dual-error model of frameshifting). The same mechanism can e
xplain frameshifting by suppressor tRNAs, even those with expanded anticodo
n loops. The previous model that suppressor tRNAs induce quadruplet translo
cation now appears incorrect for most, and perhaps for all of them. We sugg
est that the 'spontaneous' tRNA-induced frameshifting and 'programmed' mRNA
-induced frameshifting use the same mechanism, although the frequency of fr
ameshifting is very different. This new model of frameshifting suggests tha
t the tRNA is not acting as the yardstick to measure out the length of the
translocation step. Rather, the translocation of 3 nucleotides may be an in
herent feature of the ribosome.