Abstract:
Spatial and temporal control of gene activity is a prerequisite of multicellular development, as the local accumulation of gene products is required to specify different cell fates. The regulation of gene expression involves a large number of different effectors, and also single-stranded RNAs of ~19-25 nucleotides in length. MicroRNAs (miRNAs) constitute a large group of endogenous small RNAs, and they negatively regulate their target genes by base-pairing to complementary nucleic acids. In animals, miRNAs typically trigger translational arrest of their targets, to which they pair with only limited complementarity. Plant miRNAs, on contrary, trigger cleavage of target transcripts, with which they share high sequence complementarity. Predictions of new plant miRNA targets have therefore focused on genes with only a small number of mismatches to miRNAs.
Studying the biological function of three plant miRNAs, I have over- and misexpressed their precursors in Arabidopsis thaliana. The different abnormalities, which could be observed in transgenic plants, strongly suggest a role of these miRNAs as regulators of plant development. Overexpression of miR156 extended the vegetative phase of Arabidopsis seedlings, and also increased the speed and number of organ initiation events. Conversely, miR172 was shown to decrease the time of vegetative development when overexpressed, and also to control floral organ identity. MiRNA164 was shown to control separation of above ground organs throughout the life cycle of Arabidopsis plants. The observed defects can be partially correlated with known loss-of-function mutants in predicted target genes, however they also postulate additional miRNA functions. These might be mediated by other predicted target genes, which had not been functionally characterized before.
I have studied direct effects of miRNA expression on target RNA accumulation and confirmed that many predicted target genes were strongly responsive to miRNA overaccumulation. This finding is consistent with phenotypes of miRNA overexpressers resembling mutants in target genes, and also with cleavage of mRNAs as the mode of plant miRNA function. In order to test for the presence of additional targets with similar or lower sequence complementarity, I monitored genome-wide expression changes caused by overexpression of five different miRNAs. In brief, these analyses suggest that plant miRNAs directly regulate only a very small number of target genes, to which they pair with high sequence complementarity.
Comparing authentic miRNA targets to other, non-responsive genes with similar mismatch numbers, I established positional determinants of plant miRNA target selection. These, unlike previous computational efforts, almost unambiguously discriminate targets from non-targets not only in the small set of miRNAs analyzed, but also when compared to other plant miRNAs.
Feedback regulation can add another level of complexity to miRNA mediated effects, as I have discovered for the case of miR172. It has been previously suggested that miR172 overexpression results in changes of target protein, rather than transcript abundance, and translational inhibition was proposed as a mechanistic basis. However, since cleavage of target transcripts was efficiently increased in miR172 overexpressing tissue, additional regulatory events have to be present. To uncouple miRNA mediated from other effects on target genes, I overexpressed a miRNA resistant version of the target gene AP2, and confirmed the presence of a negative feedback of AP2 on its own expression. This suggests that miR172 functions by simultaneously triggering cleavage and translational inhibition of its target genes.
Since the action spectrum of plant miRNAs is very narrow, it contrasts with the broad selectivity of animal miRNAs. This difference might reflect either only intrinsic properties of the plant miRNA machinery, or selection against miRNAs with broader specificity has reduced the number of plant miRNA targets. In order to distinguish between the two possibilities, I have generated artificial miRNAs (amiRNAs) targeting endogenous genes and found that their specificity was as high as that of natural plant miRNAs. This finding supports the idea that extensive basepairing with target genes is required for plant miRNA function. Since amiRNAs were efficiently produced and could be designed to specifically silence single, or groups of endogenous genes, they can easily be used as a tool for directed gene silencing in plants. In addition to conventional silencing of single genes, which is already possible by RNA interference, amiRNAs can specifically downregulate expression of multiple related genes. As they function with high specificity, they can potentially also function in strand or allele specific gene silencing, which is not possible by other means. Furthermore, introduction of amiRNA-insensitive variants of targets can be generates to compensate for defects in amiRNA expressing plants. A web-based tool has been established to automatically design amiRNAs for genes of interest and is available to the scientific community for further studies.