Biology Reference
In-Depth Information
could have dozens, or even “hundreds,” of target mRNAs. In addition, the
finding that many miRNAs cause not only translation repression but also a
decrease in the levels of their target mRNA has led to the use of microarray
analysis to identify potential miRNA targets. In these studies, as well as in
proteomic analyses, typically the level of numerous mRNAs and proteins
change, in general quite subtly, when a given miRNA is knocked out or
overexpressed, and the subsequent statistical analyses in general show
enrichment among these for potential targets of the miRNAs under study
(based on the same prediction algorithms). This has been taken as proof that
each miRNA indeed targets a large number of mRNAs. Unfortunately, it
seems to be disregarded that our prediction algorithms still suffer from very
high false-positive rates and that in these genome-wide analyses the majority
of the changes observed could be indirectly caused by miRNA misexpres-
sion, even those that have a putative miRNA binding site. In addition,
methods to validate miRNA targets usually rely on overexpression of the
miRNA and a sensor in a heterologous system, which also likely produces
false positives. Moreover, as mentioned above, a miRNA can have distinct
targets in different cell types or even in the same cell type at distinct time
points. It is therefore crucial to conduct appropriate experiments to inter-
pret the target-prediction data correctly. Therefore, while it is a valid
possibility that a miRNA causes its effects by targeting dozens of mRNAs,
it has not been satisfactorily proven yet, and it is in fact very challenging to
do so, as it requires systematic testing of all predicted targets in the proper
experimental setup with rigorous quantitative approaches.
In contrast, it has been easier to validate those cases where a miRNA has
one or a few major targets. In these cases, the strongest evidence has come
from studying genetic interactions. For example, reduction in the dose of a
major target can fully (or almost fully) suppress the loss-of-function pheno-
type of the miRNA. Or alternatively, protection of the predicted target
from miRNA repression fully (or almost fully) suppresses the miRNA
overexpression phenotype. We have discussed some of these examples
along the previous sections, but to name a couple, the function of
lsy-6
can be fully accounted for by its effect on
cog-1
; mir-9 has a handful of
targets, but specific targets seem to mediate distinct functions in different
cell types or at different times. Other examples include
mir-8
and its target
atrophin
in the
Drosophila
nervous system (
Karres
et al
., 2007
) and mir-150
and its target c-Myb in the mouse B-cell lineage (
Xiao
et al
., 2007
).
Regarding the mode of action of miRNAs, experimental evidence
suggests two main types of miRNAs: those that are coexpressed with their
targets and thus likely modulate the target concentration and those that are
mutually exclusive and are therefore considered to have a switch-like
behavior. Whether a miRNA will act as a modulator or as a switch depends
on a number of variables. An important one to consider is the cellular
concentration of the miRNA, given that the number of turnovers a given
