Just FYI, the January 2007 issue of Scientific American has a fascinating article about riboswitches. A riboswitch is basically a mRNA transcript which can self-reguate it's own expression by controlling whether the gene it carries should be read and executed or destroyed without ever being read.
People used to think mRNA was passive - a 'photocopy' of a sequence of DNA - and required ribosomes to read it and a protein-based molecular machine called a protein supervisor to make the call as to whether the cell actually needs that mRNA expressed at that point in time time or not. But a mRNA sequence with a riboswitch leader self regulates by folding into a structure which acts as its own molecular machine, binding to whatever nutrients or other triggers in the cell are used to make the expression decision and therefor self-monitors the cell's need for the protein it encodes without any protein supervisor involved.
Personally, I find that totally cool. It's bascially a glimpse back into RNA world, a glimpse of what life may have been based on before DNA was invented. It shows that (unlike DNA, which is more durable and better suited for information storage, but cannot itself form molecular machinery) RNA can both carry information and form complex molecular machinery, machinery which at some point in the distant past, before it was replaced by today's elaborate mechanisms, may have allowed for a single naked RNA molecule to both carry genetic information and self replicate without the need for any other cellular machinery. The discovery of riboswitches and other RNA-based molecular machinery discussed in the article adds significantly to the evidence that an RNA based-replicator was perhaps a precursor to DNA-based life.
