Assuming that the microRNA-Dicer-RISC algorithm (usually collectively known as RNA interference) did initially evolve as a defence against viruses*, it might seem odd that this isn't its chief function today. Rather, RNA interference predominantly earns its keep by downregulating the expression of various genes.
Assuming that the dicer-RISC sequence is already operational, how could you use it to modify gene expression of your own, normal genes? Hint: RNA interference only acts after the relevant gene has been transcribed into RNA (as its name suggests!), so we'll have to concentrate our efforts there.
Any guesses? Well, what we could do is simply synthesise a short bit of RNA that is complementary to a part of the transcribed RNA sequence. That way, it would automatically bind to the target gene's RNA and form ... double-stranded DNA. And, as we've shown in the previous post, the presence of double-stranded DNA provokes an alarm response in the cell ("perhaps it's a virus!") that rapidly degrades it, and any copies of it (single stranded or double-stranded), by the Dicer-RISC mechanism of RNA interference.
As so happens, that is exactly what occurs. Hundreds of our own genes encode small fragments of RNA complementary to other genes; these fragments are called microRNA (miRNA). And if these fragments bind to their counterpart RNA, the RNA interference mechanism is activated. In all there are at least 500 miRNAs present in mammalian cells, collectively down-regulating 30% of our genes.
Appropriately enough for such a new discovery, that isn't the end of the RNA interference story. For instance, it has also been implicated in keeping chromatin condensed, and in preventing transcription (in addition to translation). But that's enough for now...
Main source, including image: The Nobel Prize in Physiology or Medicine 2006 - Advanced Information.
* This is quite a big assumption, actually - the jury's still out.