Introduction
The 5' end of newly synthesized RNAs bears a triphosphate derived from the first transcribed nucleotide. In eukaryotes, the distal phosphate is replaced by an inverted, methylated GMP to form the m7GpppX cap; however, no such modification occurs in prokaryotes. Instead, the 5'-triphosphate of the first transcribed nucleotide is retained as the defining feature of the mRNA 5' end, and this was thought to remain unchanged throughout the life of the mRNA. The hydrolysis of the cap to a 5'-monophosphate is the first committed step in eukaryotic mRNA decay, and now a study by Celesnik et al.1 in Molecular Cell shows that a remarkably similar process occurs in E. coli.
In eukaryotes, the process of mRNA decay generally involves loss of the 3' poly(A) tail and removal of the cap by a decapping enzyme (Dcp2). The body of the mRNA is then degraded from the 5' end by the 5'-3' exonuclease Xrn1 and from the 3' end by the exosome2 (Fig. 1a). Prokaryotes such as E. coli use a distinctly different process in which mRNA is cleaved downstream of the 5' end by the endonuclease RNase E to generate an upstream product with a 3' hydroxyl and a downstream product with a 5'-monophosphate. The upstream product is rapidly cleared by 3'-5' exonucleases, and the 5'-monophosphate end of each downstream cleavage product activates subsequent rounds of endonuclease cleavage and exonuclease clearance (bottom of Fig. 1b). Thus, on the surface bacterial mRNA decay appears to be distinctly different from the major decay process in eukaryotes.
Figure 1: E. coli mRNA decay is activated by the conversion of the mRNA 5' end from triphosphate to monophosphate.
![Figure 1 : E. coli mRNA decay is activated by the conversion of the mRNA 5|[prime]| end from triphosphate to monophosphate.](/nchembio/journal/v3/n9/thumbs/nchembio0907-535-F1.jpg)
(a) Eukaryotic mRNAs begin with an m7GpppX cap structure on the 5' end and end with a 150–200-residue poly(A) tail. Decay generally begins with shortening of the poly(A) tail to <30 residues by one or more deadenylases. The m7GpppX cap is cleaved by Dcp2 to generate RNA with a 5'-monophosphate end, which is subsequently degraded with 5'-3' polarity by Xrn1 and with 3'-5' polarity by the exosome. (b) Bacterial mRNAs begin with a 5'-triphosphate and end with a stem-loop structure, and decay involves endonuclease cleavage by RNase E. The preferred substrate for RNase E is RNA with a 5'-monophosphate, a property that is determined by the presence of a monophosphate-binding pocket within the catalytic domain of the enzyme. Celesnik et al.1 describe a previously unknown step in mRNA decay in which pyrophosphate is removed from the 5' end by an unknown enzyme(s) (scissors), in a manner analogous to decapping in eukaryotes, to generate a 5'-monophosphate substrate for cleavage by RNase E. This is the rate-limiting step in decay, and subsequent cleavage by RNase Egenerates an upstream product with a 3'-hydroxyl that is degraded by 3'-5' exonucleases and a downstream product with a 5'-monophosphate. This cycle is then repeated to complete degradation of the mRNA.
Full size image (46 KB)The 5' end of substrate mRNA dramatically affects its ability to be cleaved by RNase E: the rate constant for 5'-monophosphate RNA is more than an order of magnitude greater than that for RNA with a 5'-triphosphate or 5'-hydroxyl3, 4. The basis for this 5'-end selectivity became clear from the crystal structure of the homotetrameric catalytic domain5, which showed that the 5'-monophosphate end is hydrogen bonded into a pocket at the base of an RNA-binding channel. This interaction is thought to cause an allosteric change in the protein that juxtaposes a magnesium-coordinated hydroxyl with the scissile phosphate at a downstream site in the RNA substrate.
Given that the 5' end of bacterial mRNA is a 5'-triphosphate, Celesnik et al.1 suspected that there might be a previously unidentified step in E. coli that triggers RNase E cleavage by converting the 5' terminus to a monophosphate. To address this, they developed a splinted ligation assay in which an antisense DNA oligonucleotide that extends past the RNA 5' end is used to position another oligonucleotide (oligo X) adjacent to 5'-most nucleotide. DNA ligase will covalently join oligo X to the 5' end of the RNA if it has a 5'-monophosphate, but not a 5'-hydroxyl or 5'-triphosphate. By carefully optimizing reaction conditions and comparing ligated with unligated RNA, this assay yields quantitative data of the amount of RNA with a 5'-monophosphate end. Using this approach, the authors showed that a significant portion of the full-length mRNA is 5' monophosphorylated. With their assay established, the authors were then able to investigate the specificity of RNAse E: if mRNA with a 5'-monophosphate is indeed the proximal substrate for RNase E, the upstream product generated by the first cleavage event should be entirely monophosphorylated. Several approaches both confirmed this and showed that pyrophosphate removal is the rate-determining step in mRNA decay (Fig. 1b). Based on this, decay should be impaired if the mRNA 5' end cannot be processed to a 5'-monophosphate. To test this the authors joined the mRNA to a hammerhead ribozyme, which after self-cleavage generated mRNA with a 5'-hydroxyl end. The half-life of this mRNA was six times greater than that of the same mRNA with a 5'-triphosphate terminus, thus demonstrating that a 5'-monophosphate is required to activate the decay process.
mRNA decay is an important and tightly regulated process, and the results of this study provide a new and unexpected link with eukaryotic mRNA decay, where decapping to generate mRNA with a 5'-monophosphate end is considered the first irreversibly committed step in the decay process. Decapping is tightly regulated and coordinated by changes in nutritional state, various signaling processes and translation. The functional similarity of pyrophosphate removal in bacteria raises the possibility that this, too, is a tightly regulated process.
The results of this study raise many questions. What enzyme(s) catalyze(s) pyrophosphate removal? Does a single enzyme release pyrophosphate from the mRNA, or do multiple enzymes act to remove the two phosphates sequentially? Given that RNase E forms a complex (the RNA degradosome) with the ATPase RhlB and other proteins6, might this ATPase moonlight as an RNA pyrophosphatase? Does the RNA pyrophosphatase have to be part of the degradosome at all? How universal is this decay mechanism, and what controls its rate? Do the small untranslated RNAs that regulate gene expression in prokaryotes affect pyrophosphate removal? Much as the discovery of decapping led to new discoveries in eukaryotic mRNA decay, the results of this study are likely to generate new insights into prokaryotic decay for some time to come.
