One of the more intriguing enzymes that handles RNA is polynucleotide phosphorylase (PNPase). This enzyme is a phosphorolytic 3’->5’ exonuclease; that means that it acts on the 3’ end of an RNA chain and moves towards the 5’ end, and that it adds phosphate (as opposed to water) to the broken phosphodiester bond. This means that the products of the nucleolytic reaction are a shortened RNA chain and a nucleotide 5′-diphosphate. The nucleolytic activity is appropriate, as the enzyme is a principal exonuclease component of the RNA-degrading machine known as the degradosome.
But RNA breakdown is not the only enzymatic activity possessed by PNPase. As I noted in an earlier essay, PNPase was a first (perhaps THE first) nucleotidyltransferase, or RNA polymerase. Indeed, it was an early candidate for the RNA polymerase (you know, the DNA-dependent RNA polymerases that are responsible for transcription). This activity reflects the fact that the nucleolytic activity, when reversed, is actually a nucleotidyl transferase activity, in which RNA chains can be extended (in a template-independent fashion) using nucleotide diphosphates as substrates. The clearest in vivo manifestation of this activity is evident in the many reports that show that PNPase can act as a poly(A) polymerase in vivo [see the review by Slomovic et al. for more on this]; this is true in bacteria and in organelles such as the chloroplast or mitochondria.
Posted by Arthur Hunt 
The RNA Underworld 