6. Discussion
6.1 The role of MTERF in mitochondrial transcription termination
6.1.1 Effects of modulating MTERF levels on sense transcripts
One aim of the project was to study the effect of modulating the MTERF level on the relative activities of the human mitochondrial heavy-strand transcription units by measuring the relative levels of different mitochondrial transcripts under different conditions. I did this by means of Northern blot analysis and Q-RT-PCR using cells either over-expressing MTERF or cells knocked down for MTERF expression.
I studied the effect of over-expressing MTERF on the steady-state tRNAPhe/tRNALeu(UUR) ratio and established that there is no significant difference in it when studying different cell clones over-expressing the natural variant of MTERF compared to the control cells (Figure 5.5A). I also studied the effect of over-expressing MTERF on the 16S and ND1 transcript levels in Flp-In™ T-Rex™ -293 cells stably transfected with the MTERF-MycHis construct induced to over-express MTERF compared to non-induced cells and observed no effect on the relative quantities of the transcripts in question (Figure 5.5B). Neither did I observe any effect on mitochondrial RNA levels during the recovery period after EtBr depletion of mitochondrial RNA and DNA, when cells over-expressing MTERF were studied (Figure 5.5C).
As an opposite approach to MTERF over-expression I studied the effects of silencing MTERF by means of RNAi on different mitochondrial transcript levels. When the effect of silencing MTERF was studied 7 days after the initial RNAi treatment there was a
modest drop detected in the tRNAPhe/tRNALeu(UUR) ratio in HEK293T cells, with a slight overall increase in mitochondrial transcript levels. Once again, no changes were observed in similarly treated MTERF-MycHis over-expressing cells (Figure 5.6A). Knocking down MTERF did not bring about any detectable effects on mitochondrial RNA levels during the recovery period after EtBr depletion of mitochondrial RNA and DNA (Figure 5.6B).
However, my research produced two pieces of evidence supporting MTERF having a role in the transcription of the mitochondrial heavy strand. Firstly, after MTERF knock down in normal cells there was a small increase in the amount of the two mitochondrial tRNAs representing each of the heavy strand transcription units to be detected, relative to cytosolic 5S rRNA (Figure 5.6A). Secondly, the levels of sense-strand 12S rRNA relative to 16S or 18S rRNA gene transcripts, measured by means of Q-RT-PCR, was markedly increased when MTERF was knocked down (Figure 5.7F). As sense-strand 16S and ND1 transcript levels relative to each other or to 18S were not significantly affected, I suggest that there exists a compensatory mechanism (Figures 5.7F and S3C in II). According this suggestion, knocking down MTERF sends out a signal for globally upregulated mitochondrial transcription or decreased turnover, in order to overcome the potential impairment in 16S rRNA biogenesis. The degree of MTERF knockdown that was reached in these experiments was only 50%, for which reason the effects of the knockdown may be underestimated. It is likely that in these experiments up to half of the cells still express MTERF normally and in the remaining cells expression is significantly decreased.
I conclude here that the amount of MTERF in cells does not regulate the relative steady-state levels of transcripts representing the two heavy-strand transcription units in a simple manner, but it seems to be modified by compensatory mechanisms. Previous studies carried out to study the role of MTERF in transcription termination have yielded diverse and seemingly contradictory results. Martinet al. (2005) and Asin-Cayuelaet al.
(2004) reported that MTERF stimulates transcription in vitro from PH1 in a rather crude system. When purified recombinant proteins (Asin-Cayuela et al. 2005) or crude extracts
with DNA-affinity purified MTERF were used (Asin-Cayuela et al. 2004), MTERF was not found to stimulate transcription. Chomyn et al. (1992) studied the effect of the A3243G mutation on mitochondrial transcript levels and the possible role of MTERF in the pathogenesis of MELAS syndrome and reported impaired mitochondrial protein synthesis, defects in respiration and decreased binding affinity of purified MTERF to its target sequence due to the A3243G mutation; however, no major changes in mitochondrial rRNA or mRNA levels were observed. Fernandez-Silva et al. (1997) established in vitro that MTERF alone cannot bring about transcription termination.
Shang and Clayton (1994) studied whether the A3243G mutation impairs transcription termination in vivo but found no evidence for this even though the mutation caused a drop in the rate of transcription terminationin vitro, supposedly by reducing the MTERF binding affinity to its target site.
Previously Selwood et al. (2000) reported that TAP (thiamphenicol) treatment does not affect 12S and 16S rRNA levels in HepG2 cells but instead increases the steady-state levels of both mRNAs and tRNAs transcribed from the IH2 transcription initiation site of the heavy strand. Thiamphenicol is an antibiotic which blocks mitochondrial protein synthesis. Their findings indicate that modulation of MTERF complex could be the limiting factor determining the mitochondrial gene expression at the level of transcription termination. The underlying mechanism was proposed to be TAP enhancing transcriptional initiation from the second HSP (IH2) transcription initiation site, which was first suggested by Montoya et al. (1983). This would explain the unchanged rRNA levels and also the enhanced transcription downstream of the MTERF binding site.
Findings from the study of thyroid hormone action (Enriquez et al. 1999) and of the effects of variation in ATP supply (Micol et al. 1997) are both in agreement with our results. The molecular mechanism behind suggested is that they have an effect on the relative rates of transcription of the two differentially transcribed transcription units of the mitochondrial heavy-strand but no effect on that at the high affinity MTERF binding site.
Some of the MTERF homologues in other organisms have been shown to affect mitochondrial transcription which argues against the idea that the effects on nucleic acid metabolism caused by altering MTERF levels are incidental while the real biological function of MTERF is still lurking somewhere else inside mitochondria. The Arabidopsis thaliana geneSOLDAT10 (singlet oxygen-linked death activator) encodes a protein that is related to the human MTERF protein. The soldat10 mutation has been reported specifically to decrease plastid rRNAs which indicates that this mutation does not universally impair chloroplast RNA accumulation (Meskauskiene et al. 2009). The soldat10 mutation has been also suggested to have more indirect effect on the ROS production or the redox state of the plastid. This mutation leads to disturbance of plastid homeostasis which suppresses ROS-mediated cell death (Meskauskieneet al. 2009). The MTERF protein homologue in Chlamydomonas reinhardtii, MOC1, on the other hand, has a role in restoring mitochondrial RNA levels after exposure to light (Schönfeld et al.
2004). MOC1 levels are upregulated after light exposure and the loss of MOC1 leads to light-sensitive phenotypes as well as impairing the chloroplast transcription and replication (Schönfeldet al. 2004).
6.1.2 MTERF regulates the levels of anti-sense transcript levels in human mtDNA
In this project I addressed the question whether MTERF could have a limiting role in transcription termination in human mitochondrial DNA in vivo. By modulating MTERF levels and studying the effect on sense and antisense 16S and ND1 transcripts, I found out that over-expressing MTERF decreases the anti-16S/anti-ND1 transcript ratio and knocking down MTERF increases it. I therefore established that MTERF regulates antisense transcript levels consistent with it facilitating of antisense transcription termination initiated at PL by binding to its canonical binding site. The effect of modifying MTERF levels on RNA19 (16S + ND1) remains to be studied. It should be interesting to do so, as it could shed more light on the antisense results.
Why would MTERF have a role in regulating mitochondrial antisense transcripts that have no documented function, and which are assumed to be destined for turnover? Asin-Cayuela et al. (2005) showed, using recombinant human MTERF in a highly purified reconstituted in vitro transcription system, that MTERF transcription termination exhibits clear polarity. When bound in the ‘forward’ orientation to its target sequence in the heavy strand promoter, MTERF arrested transcription completely, but when bound in the opposite orientation, it arrested transcription only partially. Asin-Cayuela et al. (2005) also showed that MTERF alone can terminate transcription in vitro and does not need to be post-translationally modified by phosphorylation. It might be that MTERF actually is capable of terminating transcription in a bidirectional manner alsoin vivo and the effects on the sense transcript levels are masked or complicated due to the post-transcriptional processing of rRNAs, the stabilization of rRNA into ribosomal subunits or due to compensatory effects on transcriptional initiation or RNA stability.
The biological significance of the antisense transcripts is still unclear but one possibility is that they could have a role in providing a stalled replication fork with a primer that is needed for the re-initiation of the replication machinery. Non-coding antisense RNA transcripts have been found in normal proliferating cells arising from the 16S rRNA gene, whereas in tumour cell lines the levels of these antisense transcripts are downregulated (Villegas et al. 2007, Burzio et al. 2009). It therefore seems that the antisense transcript levels may relate to the tumorigenicity or proliferation status of the cell. The exact physiological origin of these antisense transcripts is still unknown but they might be created post-transcriptionally or during transcription by template strand-switching.
My results suggest that MTERF has a role in mitochondrial transcriptionin vivo, but it does not seem to set the levels of mature mitochondrial transcripts encoded by the PH1 and PH2 heavy-strand transcription units in a simple manner. Rather, this is influenced by compensatory mechanisms.