In order to reduce background signals in Q beta replicase-mediated. bi
oassays, a target-dependent probe amplification strategy has been prop
osed that utilizes recombinant RNA hybridization probes that contain a
n inserted molecular switch.(1) A molecular switch is an internal regi
on of the probe that undergoes a conformational change when the probe
hybridizes to its target. We investigated whether non-hybridized probe
s (which cause background signals) could be selectively destroyed by i
ncubating the probe-target hybrids with ribonuclease ill, which should
cleave the non-hybridized probes and leave the hybridized probes inta
ct. Two problems with this assay design were observed. First, ribonucl
ease III cleaved probe-target hybrids non-specifically when the target
was an RNA, thereby destroying all of the bound probes. And second, t
he expected conformational change in the molecular switch did not occu
r when the probes were bound to their targets, apparently because the
hairpin stem formed by the molecular switch was too long. Although the
se results demonstrated that the original assay design could not work,
they provided insights that have led to better designs for target-dep
endent amplification assays. In these assays, the probes will be DNA m
olecules containing short-stemmed molecular switches. Non-hybridized p
robes will be selectively destroyed by incubation with a restriction e
ndonuclease. (C) 1997 Academic Press Limited.