Main

Its name suggests insignificance; one would think that in the age of super-sizing, something called microRNA, sporting a mere 22 nucleotides, would not attract a lot of attention. And yet, the more we learn about these miRNAs, the more impressive their role in genetic regulation becomes. Essential biological processes, such as embryonic development, cannot happen without miRNAs. The question of which genes are targeted by miRNAs has been of interest to many researchers, including the group of Stephen Cohen at the European Molecular Biology Laboratory.

For Cohen and his colleagues, the first brush with miRNA came when they found that a gene responsible for tissue growth in flies encodes an miRNA. Their next step was to find targets for this miRNA, but this proved to be harder than anticipated, because the miRNA, despite its short sequence, did not perfectly match any mRNA. After developing a rough target-prediction method, Cohen decided on a more systematic approach for the detailed analysis of criteria for binding between a target and an miRNA (Brennecke et al., 2005). The Cohen team developed a screen that allowed them to determine how well a particular miRNA recognized its target; by introducing changes at certain positions in the miRNA, they could identify positions critical for target binding.

The profile of a functional miRNA-target interaction yielded a few surprises, as Cohen pointed out: “I think the most striking finding is that sites which consist of no more than eight base pairs can function.” An eight-base-pair match at the 5′ end of the miRNA seems to be the best indicator for a good target site. But not all is lost for an miRNA with some mismatches in the first eight 5′ bases, as Brennecke et al. show that weak 5′ binding can be compensated for by strong binding at the 3′ end of the miRNA. These experiments underscore the complexity of miRNA target site prediction: out of the 22 possible, as few as 8 base pairs between miRNA and mRNA may be sufficient, and in certain cases paired bases at the 5′ and 3′ ends of the miRNA are needed. Moreover, bulges formed by mismatches are tolerated in certain positions but not in others. The Cohen team incorporated these findings into computational models for prediction of miRNA targets and concluded that a large number of genes in the fly are likely to be regulated by miRNAs.

In the end, any computationally predicted miRNA target has to be verified, and Cohen and coworkers are now testing some of their predicted targets experimentally. Their screen will tell them which of the predicted targets the miRNA actually binds to, but as Cohen points out, ”It will not necessarily mean that a particular site is regulated in the natural circumstances in a cell.” To prove that miRNA is working on its target in vivo, one will have to first mutate the endogenous miRNA and then demonstrate upregulation of the target gene.

Several groups are now working on such loss-of-function miRNA mutants to do proper genetic testing, whereas others are working to refine the computational prediction methods, and others still are focused on identifying more miRNA genes. It will take such multidisciplinary efforts to reveal just how influential the seemingly humble microRNA is.