First, there's the initiation stage when a ribosome binds to the initiation codon in messenger RNA (mRNA) and the translation initiation complex forms by recruiting additional components.
Second, there's the elongation stage when the elongation complex moves along the mRNA translating the coding region and producing a polypeptide chain. The elongation rate is relatively constant but from time to time the elongation complex pauses at particular sites that are difficult to translate.
Finally, the elongation complex encounters a termination site where it disassembles and the ribosome with its various factors is released from the mRNA.
At any given time, a typical mRNA molecule will contain ribosomes and translation factors at all three stages. If initiation, elongation, and termination are rapid then each mRNA will be translated many times before it is degraded and there will be several elongation complexes working simultaneously. This gives rise to a "polysome" as shown in the figure below where multiple mRNAs are being translated in E. coli.
For many years it has been possible to treat cells with drugs that block elongation, for example, then isolate polysomes and digest all exposed RNA with nucleases. You are left with elongation complexes and small bits of mRNA that have been protected from the nuclease because they are bound to the complex. These little bits of RNA can be copied into DNA, cloned, and matched to the sequence of a particular gene. The result is a ribosome profile showing the average position of a translation complex inside the cell. One can also block initiation and ask how many ribosome are sitting at an initiation site at any given time. Same for termination.
Modern sequencing technology ("deep sequencing") allows for this type of analysis at a global level. Millions of protected RNAs are sequenced and matched to the genome of the species. This way you get information on ribosome profiling for thousands of genes at the same time.
Ingolia et al. (2011) did this experiment using mouse cells. There's a brief summary in Science (Weiss and Atkins, 2011).
This rate is a lot slower than the rate in bacteria. In E. coli, for example, the rate of protein synthesis is about 18 amino acid residues per second, which corresponds to a rate of movement of 54 nucleotides per second. This is the same as the rate of transcription (55 nucleotides per second) and that's why protein synthesis can keep up with transcription (see figure).
About 25% of all genes have, on average, one strong pause site where translation complexes halt for several seconds. Most pause sites have a glutamate or aspartate codon at the A site and a proline or glycine codon immediately downstream (already translated). An additional proline codon usually lies two codons downstream.
We used to think that pause sites were caused by rare codons where the abundance of the corresponding aminoacyl-tRNA was much lower than normal but apparently that's not correct.
There's no evidence of pileup at a pause site nor is there evidence of a depleted region downstream of a pause site ("shadow"). Pileup and depletion would be expected if translating ribosome were densely packed on an mRNA molecule as they are in bacteria. The evidence suggests that the average ribosome density on mRNA is not high enough to permit a trailing ribosome to catch up with one that's paused for several seconds. This, in turn, indicates that translation initiation is much slower in the mouse cells than in bacteria.
The most starling result of this study is the number of unannotated translation initiation sites. Recall that most start sites have the methionine codon AUG and in eukaryotes the pre-initiation complex binds to the 5′ cap and then moves down the mRNA until it encounters the first AUG where it pauses while the complete initiation complex is assembled. Then translation begins at this site.
Ingolia et al. discovered that 65% of their mRNAs (genes) contain more than one site where ribosomes are bound in the presence of the drug harringtonine—a drug that prevents translation initiation complexes from transitioning to elongation complexes. They refer to these as initiation sites but I'm not sure that's valid unless they have additional evidence that translation actually begins at these sites. Many of the sites cover codons that differ from AUG by a single nucleotide.
Most of these sites are downstream of the annotated start site, suggesting that many proteins with truncated N-terminii are being produced. However, in 280 genes there is no detectable binding at the annotated start site indicating that a downstream site is preferred or that the annotation (and/or sequence) is incorrect.
That's a problem with these studies. When the results conflict with the annotation you never know for sure whether there's a genuine conflict or whether the annotation is wrong. Most of these genes have not been well studied and there could be several mistakes: (a) the sequence could be wrong, (b) the transcription start site could be wrong, or (c) the predicted spliced products could be wrong.
Some of these issues could be resolved by looking at each individual gene but that's a lot of work when dealing with results from 5000 genes. Hopefully, there will be more studies like this on other types of mouse cells and on other species.
[Image Credit: Nobelorize.org]
Ingolia, N.T., Lareau, L.F., and Weissman, J.S. (2011) Ribosome profiling of mouse embryonic stem cells reveals the complexity and dynamics of mammalian proteomes. Cell 147:789-802. Epub 2011 Nov 3. [doi: 10.1016/j.cell.2011.10.002]
Weiss, R.B. and Atkins, J.F. (2011) Translation goes global. Science 334:1509-1510.