Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 Apr 30;9(1):1727.
doi: 10.1038/s41467-018-04131-w.

Linear Mitochondrial DNA Is Rapidly Degraded by Components of the Replication Machinery

Affiliations
Free PMC article

Linear Mitochondrial DNA Is Rapidly Degraded by Components of the Replication Machinery

Viktoriya Peeva et al. Nat Commun. .
Free PMC article

Abstract

Emerging gene therapy approaches that aim to eliminate pathogenic mutations of mitochondrial DNA (mtDNA) rely on efficient degradation of linearized mtDNA, but the enzymatic machinery performing this task is presently unknown. Here, we show that, in cellular models of restriction endonuclease-induced mtDNA double-strand breaks, linear mtDNA is eliminated within hours by exonucleolytic activities. Inactivation of the mitochondrial 5'-3'exonuclease MGME1, elimination of the 3'-5'exonuclease activity of the mitochondrial DNA polymerase POLG by introducing the p.D274A mutation, or knockdown of the mitochondrial DNA helicase TWNK leads to severe impediment of mtDNA degradation. We do not observe similar effects when inactivating other known mitochondrial nucleases (EXOG, APEX2, ENDOG, FEN1, DNA2, MRE11, or RBBP8). Our data suggest that rapid degradation of linearized mtDNA is performed by the same machinery that is responsible for mtDNA replication, thus proposing novel roles for the participating enzymes POLG, TWNK, and MGME1.

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1
Degradation of linearized mtDNA in control and mutant mitoEagI-expressing HEK 293 cells. a Southern blot showing the degradation of mtDNA within the first 18 h of induced expression of mitoEagI (E) in control, MGME1 p.I9Qfs*32 knockout (‘MGME1 ko’), and exonuclease-deficient POLG (‘POLG p.D274A’) cells. BamHI endonuclease-linearized DNA (B) was labeled with a mitochondrial probe represented by an asterisk as well as a probe specific for nuclear 18S ribosomal DNA (‘18S’). Note that persistent bands with one end in the vicinity of oriL in MGME1 ko and mutated POLG cells are already present before induction (time point ‘0’, lowest arrowhead). These linear mtDNA species are due to leaky mitoEagI expression and their presence is not related to the induced massive double-strand breaks. b Quantification of full-length mtDNA confirms the efficient cleavage of mtDNA by mitoEagI and (c) the persistence of mitoEagI-linearized mtDNA in MGME1 ko and POLG p.D274A cells. Band intensities were first normalized to 18S ribosomal DNA intensities then to intensities of the full-length mtDNA in each cell line before induction. Error bars represent standard errors of the mean (SEM) in three independent experiments (including both available MGME1 knockout clones). Significance was calculated by applying one-way ANOVA test. *P < 0.05, **P < 0.01, *** P < 0.001. d, e Coverage ratios throughout the mitochondrial genome as determined by ultra-deep sequencing of mtDNA from cells 6 h after induced mitoEagI expression and normalized to values in non-induced cells. In control (gray), coverage ratio is the lowest around the mitoEagI cutting site (represented by the two ends of the x-axis) and gradually increases in both directions before reaching full coverage. In MGME1 ko cells (d, red) and in cells with exonuclease-deficient POLG (e, green), coverage ratio drops only in the immediate vicinity of the mitoEagI cutting site. Note that library preparation techniques used for ultra-deep sequencing result in underrepresented positions in the close vicinity of free DNA ends
Fig. 2
Fig. 2
Degradation of the 2.1-kb mtDNA fragment in control and siRNA-treated mitoPstI-expressing HEK 293 cells. a Southern blot showing the appearance and degradation of the 2.1-kb mtDNA fragment within the first 4 h of induced mitoPstI (P) expression in control cells and in MGME1, POLG, or TWNK siRNA-treated HEK 293 cells. BamHI endonuclease-linearized DNA (B) was labeled with a mitochondrial probe represented by an asterisk as well as a probe specific for nuclear 18S ribosomal DNA (‘18S’). b Quantification of full-length mtDNA confirms the efficient cleavage of mtDNA by mitoPstI and c the persistence of the 2.1-kb mtDNA fragment in MGME1, POLG, and TWNK knockdown cells. Band intensities were first normalized to 18S ribosomal DNA intensities then to intensities of the full-length mtDNA in each cell line at the starting time point and, in panel (c), additionally to the highest 2.1-kb fragment value on each blot. Error bars represent standard errors of the mean (SEM) in three independent experiments. Significance was calculated by applying one-way ANOVA test. *P < 0.05. d, e Cumulative relative frequencies (CRF) of ends around the 6910/6915 cutting site (represented by position 0) 2 h after mitoPstI induction as determined by single-molecule amplification of linker-ligated free mtDNA ends and subsequent Sanger sequencing. Positions indicated are relative to non-degraded ends at 6910 (d) or 6915 (e). CRFs were calculated from the following number of detected ends: d control (gray), n = 36; MGME1 siRNA (red), n = 39; POLG siRNA (green), n = 46; TWNK siRNA (blue), n = 37; e control (gray), n = 51; MGME1 siRNA (red), n = 50; POLG siRNA (green), n = 37; TWNK siRNA (blue), n = 42
Fig. 3
Fig. 3
Free mtDNA ends in HEK 293 cells expressing mitoEagI or mitoPstI. a Cumulative relative frequencies (CRF) of mtDNA ends detected by ultra-deep sequencing of linker-ligated mtDNA 6 h after induction of mitoEagI expression. CRF values for both orientations of ends were combined into a single curve. MGME1 knockout (red) results in persistence of non-degraded ends. In POLG p.D274A knockin cells (green), over 80% of ends are detected within a distance of 600 base pairs from the cutting site. Functionally relevant sites associated with prominent clusters of ends are marked on the top (mTER, mitochondrial transcription termination site; oriL, replication origin for the light strand; [oriH], replication origin region for the heavy strand). b, c Relative frequencies of blunt mtDNA ends at the vicinity of cutting sites detected by ultra-deep mtDNA sequencing 6 h after induction of mitoEagI (b) and mitoPstI (c). Balk heights represent proportions of ends at specific nucleotide positions among all detected ends of the same orientation. Shadings indicate the retained part of mtDNA. Gray, control; red, MGME1 knockout in mitoEagI and MGME1 siRNA knockdown in mitoPstI. Blue shading indicates mitoEagI and mitoPstI recognition sites (schematically shown on the top). d, e Frequent mtDNA ends distal to the cutting sites as observed by ultra-deep mtDNA sequencing 6 h after induction of mitoEagI (d) or mitoPstI (e) in control cells. Shaded areas indicate the retained mtDNA fragment. Associated GC stretches are indicated by red shading (at least 3 consecutive Gs or Cs starting at less than 3 nucleotides difference from the end). Note that prominent ends are located at different sides of the same GC stretch depending on the main direction of degradation in mitoEagI-expressing and mitoPstI-expressing cells (indicated by arched arrows in the schemes on the side). The presence of non-degraded and selected partially degraded ends was confirmed by single-molecule PCR (Supplementary Fig. 5b)
Fig. 4
Fig. 4
GC stretches decelerate mtDNA degradation. Relative frequencies of nucleotides at the vicinity of detected blunt double-stranded mtDNA ends are shown. Nucleotides surrounding all ends detected by ultra-deep sequencing of linker-ligated mtDNA were counted and their relative frequencies were normalized to overall nucleotide frequencies in the mitochondrial genome. Ends at the immediate vicinity of the cutting site were excluded from the analysis (mitoEagI: positions 2552–2585; mitoPstI: positions 6810–7015 and 8920–9125). To avoid bias through the most prominent ends in mitoEagI cells, corresponding positions were also excluded (positions 5732–5742 and 3208–3215). ‘dox+’, samples taken 6 h of inducing mitoEagI or mitoPstI expression. ‘dox−’, no induction, only leaky mitoEagI or mitoPstI expression. Upper set of panels shows ends generated by degradation in the forward direction, lower set of panels in reverse direction (according to reference numbering). Faded balks represent removed nucleotides. a Note that the high frequency of guanine and cytosine residues at the first 6 positions of linear mtDNA species upon induced cleavage in control mitoEagI cells. b Similar pattern can be observed in mitoPstI cells, although, cutting sites localize to different parts of the mitochondrial genome. c MGME1 knockout cells and d POLG p.D274A knockin cells do not show differences between induced and non-induced conditions. This is in line with the observation that newly generated ends do not undergo rapid degradation in the absence of MGME1 exonuclease or in the presence of exonuclease-deficient POLG
Fig. 5
Fig. 5
Rearranged mtDNA molecules in HEK 293 cells expressing mitoEagI or mitoPstI. a Positions of mtDNA breakpoints as detected by ultra-deep sequencing of mtDNA in control cells 6 h after induction of mitoEagI. Continuous lines represent retained parts of the mitochondrial genome. Dotted lines indicate deleted regions in rearranged mtDNA species. Note that the observed breakpoints correspond to frequent ends (cf. Figure 3d). b PCR detection of breakpoints corresponding to non-degraded and partially degraded ends of the 2.1-kb mtDNA fragment in mitoPstI-expressing cells. Amplification primers MT8282F and MT7682R (Supplementary Table 4) are shown in the scheme as arrowheads. Note that the majority of detected breakpoints correspond to partially degraded mtDNA ends in control cells, while MGME1, POLG, or TWNK siRNA treatments result in breakpoints mainly corresponding to non-degraded mtDNA ends. The exact positions of representative breakpoints determined by single-molecule PCR and sequencing are shown in Supplementary Table 2. c Concatemers of the 2.1-kb mtDNA fragment in mitoPstI-expressing cells as detected by long-extension PCR using primers MT8194F and MT8387R (Supplementary Table 4). The shortest bands represent amplification products from unique copies of the 2.1-kb mtDNA fragment (lower arrowhead). Longer PCR products indicate the presence of multimers of the mtDNA fragment corresponding to non-degraded (upper arrowhead) or partially degraded (middle part) ends. Knockdown of the LIG3 gene decreases the abundance of concatemer species
Fig. 6
Fig. 6
Models of double-strand degradation of mtDNA and its role in generation of rearrangements. a POLG, TWNK, and MGME1 play a role in both replication (upper half of the panel) and degradation (lower half of the panel) of mitochondrial DNA. Upon replication, the net movement of the complex (indicated by the large shaded arrow) corresponds to the polymerase activity of POLG. Under these conditions, MGME1 can remove flap structures, thus creating ligatable ends. During degradation, net movement is reversed and corresponds to the exonuclease activity of POLG. b Proposed role of linear mtDNA degradation in generation of rearrangements. Since double-strand breaks (DBS) are efficiently removed in normal tissues (blue), most somatic mtDNA deletions are generated by replication slippage, and thus are typically associated with direct repeats (black boxes) around the breakpoints (‘I’, class I type of mtDNA deletions, Table 1). Repair by homologous recombination (HR) or microhomology-mediated end joining (MMEJ) is not an efficient pathway in animal mitochondria. In the case of replication machinery dysfunction (red), frequent replication stalling leads to increased generation of double-strand breaks (DSB). Additionally, the breakdown of linear mtDNA is inhibited. The persistence of linear mtDNA favors the formation of class II deletions (‘II’) by ligase-III-dependent non-homologous end joining (NHEJ). Supporting this hypothesis, an increased frequency of class II mtDNA deletions (not being associated with direct repeats) is observed in patients carrying pathogenic mutations in MGME1, TWNK, and POLG (Table 1)

Similar articles

See all similar articles

Cited by 20 articles

See all "Cited by" articles

References

    1. Shokolenko IN, Alexeyev MF. Mitochondrial DNA: a disposable genome? Biochim. Biophys. Acta. 2015;1852:1805–1809. doi: 10.1016/j.bbadis.2015.05.016. - DOI - PMC - PubMed
    1. Moretton A, et al. Selective mitochondrial DNA degradation following double-strand breaks. PLoS One. 2017;12:e0176795. doi: 10.1371/journal.pone.0176795. - DOI - PMC - PubMed
    1. Bacman SR, et al. Specific elimination of mutant mitochondrial genomes in patient-derived cells by mitoTALENs. Nat. Med. 2013;19:1111–1113. doi: 10.1038/nm.3261. - DOI - PMC - PubMed
    1. Gammage PA, et al. Mitochondrially targeted ZFNs for selective degradation of pathogenic mitochondrial genomes bearing large-scale deletions or point mutations. EMBO Mol. Med. 2014;6:458–466. doi: 10.1002/emmm.201303672. - DOI - PMC - PubMed
    1. Reddy P, et al. Selective elimination of mitochondrial mutations in the germline by genome editing. Cell. 2015;161:459–469. doi: 10.1016/j.cell.2015.03.051. - DOI - PMC - PubMed

Publication types

MeSH terms

Feedback