ceRNAs: miRNA Target Mimic Mimics

In a recent issue of Cell, four papers described regulatory interactions among messenger RNAs (mRNAs) that share target sequences for the same microRNAs (miRNAs) (Cesana et al., 2011; Karreth et al., 2011; Sumazin et al., 2011; Tay et al., 2011). The authors state that these mRNAs, dubbed ‘‘competing endogenous RNAs (ceRNAs),’’ define a new layer of regulation of miRNA activity that has only recently been discovered. It may be new for animals and humans, but the principle of this phenomenon had been identified a while ago in plants, where it is known as ‘‘target mimicry.’’ Plant biologists not only played a central role in defining the function of small RNAs in gene silencing, but they also made important contributions to our understanding of miRNA action. In both plants and animals, miRNAs negatively affect their targets through a variety of transcriptional and posttranscriptional mechanisms. In addition, the biogenesis and activity of miRNAs themselves can be regulated at several levels. One of these is through target mimicry. In 2007, we reported that the IPS1 (INDUCED BY PHOSPHATE STARVATION1) RNA altered the protein levels of PHO2 (PHOSPHATE2) by modulating the effects of miR399 on the stability and translation of PHO2 mRNA (Franco-Zorrilla et al., 2007). The RNAs of both IPS1 and PHO2 have highly conserved sequence motifs complementary to miR399, with IPS1 having additional bases that interrupt the position where miR399 would normally guide cleavage of its target. A series of experiments supported a model in which noncleavable IPS1 sequesters miR399 and thus prevents it from inhibiting PHO2 mRNA accumulation and translation. We coined the term ‘‘target mimicry’’ for this endogenous regulatory mechanism of miRNA activity. Although IPS1 encodes a noncoding RNA, these early findings already pointed to the possibility that noncleavable target sites in plants could perhaps not only act in trans, but also in cis, as found to be the case in animals. We subsequently confirmed that artificial target mimic sites in plants could indeed reduce translation efficiency in cis (Todesco et al., 2010). A few weeks after natural and artificial target mimics had been described for plants (Franco-Zorrilla et al., 2007), artificial target mimics were introduced as ‘‘decoy targets’’ for miRNAs in animals (Ebert et al., 2007). Subsequently, a theoretical paper suggested that many natural miRNA target sites could act as miRNA decoys or target mimics in animals. In such a scenario, changes in the expression of some miRNA targets (or mimics) would alter the ability of a miRNA to reduce the activity of other targets (Seitz, 2009). Last year, Pandolfi and colleagues reported an exciting discovery of a naturally occurring noncoding RNA produced by a pseudogene that acts as a natural target mimic in human tumors (Poliseno et al., 2010). A recent Essay in Cell (Salmena et al., 2011) summarized the conclusions from this and some of the prior work, including our initial discovery of ‘‘target mimics’’ (Franco-Zorrilla et al., 2007) and the theoretical prediction by Seitz (2009). The Essay formally enumerated the many ways in which noncoding and coding RNAs with similar miRNA target sites could affect each others’ activity. The authors renamed such RNAs with shared and competing target sites as ceRNAs. They also stated that their theory ‘‘challenged the notion that a protein-coding mRNA must be translated into a protein to exert function,’’ although there has been ample prior evidence for noncoding functions of protein-coding mRNAs in animals and plants. Such examples include mRNAs that produce small RNAs because they form natural antisense transcripts or are routed through the trans-acting siRNA pathway. In the September 30, 2011 issue Cell, three groups described new ceRNAs in animal and human cells (Cesana et al.,