Gene expression is very complex. My paper, which was published in Cell today, just shows that it is more complicated than previously realized.
Traditionally, eukaryotic gene expression is divided into five steps:
- Transcription (mRNA synthesis): this step is subdivided into initiation (i.e. recruitment of specific & general transcription factors, which in turn recruit/activate RNA polymerase II (pol II) up to the synthesis of the first few nucleotides; then comes the elongation step and finally the termination.
- mRNA maturation: includes capping, splicing and polyadenylation (which occur co- or post- transcriptionally).
- mRNA export: export of the mRNA from the nucleus into the cytoplasm
- mRNA decay (degradation): this step is subdivided into deadenylation (removal of the poly-A tail), followed by decapping (removal of the 5’ cap) and ending in exonucleolytic digestion to single nucleotides.
Recent research showed that all these steps are coupled, meaning that each step depends on the proper execution of the previous step (usually by common factor(s) that participate in both steps). Our lab and others showed that transcription regulatory elements (pol II subunits, as well as promoter sequences and transcription factors) can regulate the cytoplasmic steps (translation & decay) [see review reference at the bottom].
This traditional view sees this as a linear process, i.e. it begins with transcription and ends with mRNA decay. My work (in collaboration with 4 other labs) shows that the mRNA decay machinery is imported into the nucleus, where it associates with chromatin at promoters (exactly at the assembly site of the pre-initiation complex) and along the genes. In the nucleus, these decay factors stimulate transcription initiation and transcription elongation.
Importantly, this role in transcription is directly linked, and is dependent, on their role in mRNA decay, fitting the definition of “coupling” the processes. Thus, the gene expression process is not linear, but circular.
The initial observation that drove this project was that the steady-state level of most mRNAs in mutant strains that are defective in mRNA decay, are similar to the steady-state levels in wild-type strains (or even less!). Since we expect an increase in mRNA levels if they are not degraded, this means that the production of mRNAs (transcription) is down-regulated.
A key finding in my work is that a point mutant of the exonuclease, Xrn1, is more defective in transcription than the deletion strain (which is mildly defective in transcription), although the decay process is similarly defective in both mutants. This suggested a direct, rather than indirect, effect of these factors on transcription.
We have used a variety of methods to show that, including Northern, ChIP, FISH, genomic run-on and epifluorescence imaging. Since this is an imaging blog, I will dwell a little bit on the imaging techniques that we used.
One important aspect of my theory was that if decay factors affect transcription directly, they must reside in the nucleus. However, in most cases, these factors are detected only in the cytoplasm. I tagged the proteins with either GFP (that is GFPS65T) or RFP and tried to look for conditions when these proteins reside in the nucleus. The microscope in my lab then was (and still is) a very basic epifluorescent microscope Nikon Eclipse E400 (from the late 1990’s!) with a digital camera (DXM 1200F).
Though I did find some conditions or strains that show nuclear localization of some factors, I was not satisfied (particularly since I could not detect a major player, Xrn1p, in the nuclei at any condition).
Then I found a double mutant that prevents export of both proteins & mRNA out of the nucleus. The reasoning was that if export is inhibited, then the cytoplasmic protein will accumulate in the nucleus and I will be able to see it. Indeed that was the case. Better yet, I found some mutant forms of Xrn1 that prevented its import (as well as other factors), thus linking import (and transcription) to mRNA binding and degradation (the key parameter for success of this experiment was not to let the cells to grow too dense, keep them at 3-5×106 cells/ml).
FISHing for transcription
Our lab, as you can suspect from the previous paragraph had little experience with advanced imaging. So we outsourced…
Our collaborators in France performed FISH analysis to look at transcription sites in three different strains (wild-type [WT], deletion mutant and point mutant of the exonuclease Xrn1). They used 5 probes against the open reading frame (ORF) and one probe (in a different color) against an intron of the TEF4 gene.
As a result, the ORF probes identify both the cytoplasmic and nuclear mRNAs, whereas the intron probe identifies only the nuclear pre-spliced mRNA, which should co-localize with the transcription site (assuming splicing occurs co-transcriptionally).
Indeed, the two colors co-localize in the nucleus, confirming the localization of the transcription site.
Several calculations were then performed:
1) The average number of cytoplasmic spots was calculated and shown to be similar in all three strains (which confirms my Northern analysis), and to agree with previous published numbers for WT strain.
2) The percentage of cells that contain a transcription site (TS) is calculated for each strain. Confirming mine and our other collaborators’ results (using other methods), the WT and deletion strain has similar % of cells containing TS, but a lower % is seen with the point mutant.
3) The fluorescent intensity of the cytoplasmic spots is represented as a Gaussian distribution of intensities. This distribution denotes the spot intensities of a single mRNA.
4) The spot intensities of the TS were measured, and a distribution of the intensities vs. % cells is plotted. This revealed that in WT strain, some cells contain TS with intensities that are brighter than the intensity of a single mRNA. This indicated that in those cells, more than one polymerase is elongating simultaneously. However, in both deletion and point mutants, all cells with TS showed intensities distributions that fit only a single mRNA. Hence, in both strains there could be only one elongating polymerase at a time, indicating a defect in transcription initiation or elongation.
5) Just for neatness, a 3D representation of a single cell is shown, with a single spot in the nucleus and the cytoplasmic mRNAs “floating” outside the nucleus.
Using a variety of methods, I think we did a very good job to prove that mRNA decay factors directly affect transcription. Though the mechanism is not clear, we have some ideas…
This paper shifts the paradigm that gene expression is a linear process, to show that it is actually circular. Too much or too little mRNA can be detrimental to the cell. The cyclic nature of the gene expression process enables the cells to maintain a robust and “safe” level of mRNA in the cytoplasm.
Haimovich, G., Medina, D., Causse, S., Garber, M., Millán-Zambrano, G., Barkai, O., Chávez, S., Pérez-Ortín, J., Darzacq, X., & Choder, M. (2013). Gene Expression Is Circular: Factors for mRNA Degradation Also Foster mRNA Synthesis Cell, 153 (5), 1000-1011 DOI: 10.1016/j.cell.2013.05.012
Haimovich G, Choder M, Singer RH, & Trcek T (2013). The fate of the messenger is pre-determined: A new model for regulation of gene expression. Biochimica et biophysica acta, 1829 (6-7), 643-53 PMID: 23337853
[This is also an opportunity to thank Motti, my PhD mentor for all of his support and continued effort to get this work published in Cell, and to all of our collaborators who invested a lot of time and effort in this work, and really did a great job to produce a marvelous set of data.]