Multiple Molecular Mechanisms Rescue mtDNA Disease in C. elegans.

Mitochondrial DNA (mtDNA) mutagenesis can be increased by inactivating the proofreading activity of the main (or only, but let’s not go there today) mitochondrial DNA-dependent DNA polymerase (POLG). During my PhD I have been working with these mtDNA mutators in flies (Bratic et al. 2015) and mice (Jiang et al. 2017). Therefore, it piqued my interest when I saw that a C. elegans mtDNA mutator was published (Haroon et al. 2018). Having a nematode, fly and a mouse carrying the analogous proofreading-deficient variant of POLG allows one to do some interesting comparisons.

To get it out of the system I just want to express first my annoyance. This nematode paper wanted specifically to establish a simpler mtDNA mutator organism which could be used to generate animals carrying mtDNA mutations and to perform suppressor screens. However, they failed to mention our mtDNA mutator fruit fly paper published 2 years ago (Bratic et al. 2015). On another note, they should have referenced the POLG knockout paper from 9 years ago (Bratic et al. 2009). Now to the actual paper.

The authors of this paper showed that the homozygous mutator nematodes have a 50% decrease in mtDNA copy number at stage L4 (Fig. 1B). Similar mtDNA depletion have been described in the mtDNA mutator fly and mouse models (Bratic et al. 2015,Jiang et al. 2017) but it should be also pointed out that the POLG knockout nematode has similar decrease (~41%) in mtDNA levels (Bratic et al. 2009). This makes one wonder whether the mutator POLG is replicating any mtDNA in nematodes. It is likely that most of this mtDNA depletion in contributed by the decrease in mtDNA levels in the gonads similar to the knockout (Bratic et al. 2009) and would also explain the fertility problems of the mtDNA mutator nematode as mentioned by Haroon et al. in the methods. One reason for this mtDNA depletion in mtDNA mutator nematodes could stem from our fly work where we showed that the fly POLG mutator variant should be considered dominant negative, because the hemizygous flies developed further in comparison with the homozygous (Bratic et al. 2015). Indeed, the exonuclease-deficient POLG is known to have increased strand-displacement activity (Farge et al. 2007) and even be insensitive to altered dNTP/NTP ratios (Forslund et al. 2018).

Interestingly, there aren’t any high heteroplasmy level mutations even after five generations of consecutive intercrosses of mtDNA mutator nematodes. It seems that like in our mtDNA mutator flies (Bratic et al. 2015), also in mtDNA mutator nematodes mtDNA mutations accumulate slowly across generations.

It has been now documented both in the mtDNA mutator mice (Trifunovic et al. 2004) and mutator flies (Bratic et al. 2015) that abolishing proofreading activity of POLG result in the formation of linear mtDNA with deletions. A model how these deletions are formed has also been put forward (Macao et al. 2015). Based on studies done using two-dimensional neutral agarose gel electrophoresis (2DNAGE), mtDNA replication in nematodes seems to be quite different to the ones in other metazoans (Lewis et al. 2015). I would have loved to see using Southern what happens to the mtDNA integrity of these mtDNA mutator nematodes but unfortunately the authors did not carry out this analysis.

They authors also tried to use this mtDNA mutator nematode to establish nematode lines carrying pathogenic mtDNA mutations. I feel they came a bit short on this one as they only established a nematode line, which mtDNA is littered with mutations. Some more careful breeding scheme, as we have used in the mouse (Kauppila et al. 2016), would have avoided this issue. The key is to carry the mtDNA mutator allele for only a one or two generations, so that each mtDNA molecule has only a few mutations. These lines should then be bred further to isolate lines with high heteroplasmy levels of certain mutations.

In the end the authors performed targeted RNAi screen against multiple cellular pathways to find suppressors of the neuromuscular defects of these mtDNA mutator nematodes. It has been shown in fruit flies that RNAi based screens have to be well controlled (Alic et al. 2012) and in this study the authors did not mention whether they included any negative controls, such as RNAi against GFP or LacZ. The reason I mention this is because nearly every pathway tested was able to suppress the neuromuscular defects, which feels a bit too good to be true (Cell Reports seem to have linked a wrong file to the Supplementary materials so I wasn’t able to see all the data). Also, they should have taken into account that the RNAi alone might be sufficient to affect the neuromuscular function of these nematodes. Some of these RNAi targets were verified further using hypomorphic and knockout mutant nematodes. In all cases the rescue of mtDNA copy number was modest so it is unclear to me how these mutants are rescuing the phenotype.

It should be also pointed out that in Fig. 2C-E and Fig. 4 the authors performed multiple compaisons using t-tests whereas ANOVA would have been more appropriate instead.

In the end the authors didn’t really use this model to its full capacity to establish nematode lines carrying pathogenic mtDNA mutations but perhaps that will happen in the future. There is always the question though whether POLG and/or mtDNA disease nematode has scientific value, because the POLG knockout nematode is viable.

References:

Alic N, Hoddinott MP, Foley A, Slack C, Piper MD, Partridge L. Detrimental effects of RNAi: a cautionary note on its use in Drosophila ageing studies. PLoS One. 2012

Bratic A, Kauppila TE, Macao B, Grönke S, Siibak T, Stewart JB, Baggio F, Dols J, Partridge L, Falkenberg M, Wredenberg A, Larsson NG. Complementation between polymerase- and exonuclease-deficient mitochondrial DNA polymerase mutants in genomically engineered flies. Nat Commun. 2015.

Bratic I, Hench J, Henriksson J, Antebi A, Bürglin TR, Trifunovic A. Mitochondrial DNA level, but not active replicase, is essential for Caenorhabditis elegans development. Nucleic Acids Res. 2009

Farge G, Pham XH, Holmlund T, Khorostov I, Falkenberg M. The accessory subunit B of DNA polymerase gamma is required for mitochondrial replisome function. Nucleic Acids Res. 2007. PMID: 17251196

Forslund JME, Pfeiffer A, Stojkovič G, Wanrooij PH, Wanrooij S. The presence of rNTPs decreases the speed of mitochondrial DNA replication. PLoS Genet. 2018 PMID: 29601571

Haroon S, Li A, Weinert JL, Fritsch C, Ericson NG, Alexander-Floyd J, Braeckman BP, Haynes CM, Bielas JH, Gidalevitz T, Vermulst M. Multiple Molecular Mechanisms Rescue mtDNA Disease in C. elegans. Cell Rep. 2018

Jiang M, Kauppila TES, Motori E, Li X, Atanassov I, Folz-Donahue K, Bonekamp NA, Albarran-Gutierrez S, Stewart JB, Larsson NG. Increased Total mtDNA Copy Number Cures Male Infertility Despite Unaltered mtDNA Mutation Load. Cell Metab. 2017

Kauppila JHK, Baines HL, Bratic A, Simard ML, Freyer C, Mourier A, Stamp C, Filograna R, Larsson NG, Greaves LC, Stewart JB. A Phenotype-Driven Approach to Generate Mouse Models with Pathogenic mtDNA Mutations Causing Mitochondrial Disease. Cell Rep. 2016

Lewis SC, Joers P, Willcox S, Griffith JD, Jacobs HT, Hyman BC. A rolling circle replication mechanism produces multimeric lariats of mitochondrial DNA in Caenorhabditis elegans. PLoS Genet. 2015

Macao B, Uhler JP, Siibak T, Zhu X, Shi Y, Sheng W, Olsson M, Stewart JB, Gustafsson CM, Falkenberg M. The exonuclease activity of DNA polymerase γ is required for ligation during mitochondrial DNA replication. Nat Commun. 2015

Trifunovic A, Wredenberg A, Falkenberg M, Spelbrink JN, Rovio AT, Bruder CE, Bohlooly-Y M, Gidlöf S, Oldfors A, Wibom R, Törnell J, Jacobs HT, Larsson NG. Premature ageing in mice expressing defective mitochondrial DNA polymerase. Nature. 2004

The presence of rNTPs decreases the speed of mitochondrial DNA replication

In this time and age, it is delightful to see a publication which is not doing poorly controlled correlative stuff or omics for the sake of omics, but instead strong hypothesis driven biochemistry. The paper I am referring to comes from the lab of Sjoerd Wanrooij in Umea (Forslund et al. 2018).

Now, it has been known since the 70's that mitochondrial DNA (mtDNA) has ribonucleotides, but it has remained unclear whether they possess a function and how they affect mtDNA maintenance and gene expression. Until very recently, there has been very little attention given to these ribonucleotides, but it seems this subject is becoming trendy again. A recent paper from the Clausen lab showed for the first time the amount and distribution of ribonucleotides in whole mtDNA and also how the amount of ribonucleotides might be different in some patients with changes in nucleotide pools (Berglund et al. 2017). In this paper, Forslund et al. try to understand whether the mitochondrial DNA polymerase (POLG) is inhibited by ribonocleotides in the template strand or by changes in the relative ration of NTP to dNTP.

In the beginning of the paper Forslund et al. make a good point, which might be missed by a lot of people not working with POLG. Back in the 90's and early 00's, people performed all POLG in vitro experiments using the exonuclease-deficient mutant because it makes the experiments easier. However, it was later recognized that the exonuclease-deficient POLG might be doing some funky things such as increased strand-displacement (Farge et al. 2007). For this reason, one should interpret the old papers with some grain of salt.

I personally like these kind of experiments, where the models used are first established using previously published systems. Not only is this helpful in reproducing already published findings but it really makes you trust the experiment. I highly recommend reading the whole thing.

The weirdest finding of the paper is that only WT POLG but not the exonuclease-deficient POLG is sensitive to low NTP/high rNTP conditions. The authors suggest in the discussion that this could be caused idling of the WT POLG between polymerase/exonuclease modes. It is not obvious to be why the exonuclease-deficient POLG would not idle also. Could it be that the exo- mutation somehow alters the protein structure to have different nucleotide selection dynamics? Or perhaps the low NTP/high rNTP condition alters the efficiency of the WT POLG so that it will have increased removal of correct bases leading to decreased full length synthesis? I hope they will follow this up.

In the discussion the authors stated that their in vitro estimation of ribonucleotide content in mtDNA is somewhat lower than the values obtained from in vivo samples. One explanation for this could be that because the nucleotide pool sizes are so challenging to measure, it could be that the in vitro measurements are over-estimating the dNTP pools or under-estimating the NTP pools. To me, this is the most parsimonious answer. They also continue hypothesizing that some other polymerases, like Polβ or PrimPol, could introduce ribonucleotides in vivo. I'm yet to be convinced that either of these polymerases would be in mitochondrial matrix or have a mitochondrial function. For instance, depending on the study, small amount of PrimPol localizes to mitochondria or doesn't. Even when a fraction of PrimPol has been shown to localize within mitochondria, in subcellular fractination it seems to be degraded together with OPA1 (Torregrosa-Muñumer et al. 2017), suggesting PrimPol is in the inner membrane space. I put my money on over/under-estimated nucleotide pools.

In overall, once again strong biochemistry from Sjoerd's group.

References:

Berglund AK, Navarrete C, Engqvist MK, Hoberg E, Szilagyi Z, Taylor RW, Gustafsson CM, Falkenberg M, Clausen AR. Nucleotide pools dictate the identity and frequency of ribonucleotide incorporation in mitochondrial DNA. PLoS Genet. 2017. PMID: 28207748

Farge G, Pham XH, Holmlund T, Khorostov I, Falkenberg M. The accessory subunit B of DNA polymerase gamma is required for mitochondrial replisome function. Nucleic Acids Res. 2007. PMID: 17251196

Forslund JME, Pfeiffer A, Stojkovič G, Wanrooij PH, Wanrooij S. The presence of rNTPs decreases the speed of mitochondrial DNA replication. PLoS Genet. 2018 PMID: 29601571

Torregrosa-Muñumer R, Forslund JME, Goffart S, Pfeiffer A, Stojkovič G, Carvalho G, Al-Furoukh N, Blanco L, Wanrooij S, Pohjoismäki JLO. PrimPol is required for replication reinitiation after mtDNA damage. Proc Natl Acad Sci U S A. 2017. PMID:29073063

Reverse your mtDNA control region and you’ll reverse your mutation bias.

The Inversion of the Control Region in Three Mitogenomes Provides Further Evidence for an Asymmetric Model of Vertebrate mtDNA Replication, PlosOne 2014

As many of you are probably aware of there is an ongoing debate regarding the mechanism of mitochondrial DNA (mtDNA) replication (see for instance Holt and Reyes, 2012). The models were summarized well in a recent paper from Fonseca et al. in PlosOne 2014.

 

Fonseca MM et al. PlosOne 2014

They analyzed the base composition of the mtDNA protein-coding genes (2500 mtDNAs) and found a clear bias, which correlated with the direction of the control region. Based on this they concluded that the strand-displacement model (SDM) or the ribonucleotides-incorporation throughout the lagging strand model (RITOLS) is likely to be the replication mechanism instead of the strand-coupled model (SCM).

Of course one can always hypothesize that the SCM is specific for cell cultures (as in the original paper describing this model) or is tissue specific. However, from the evolutionary point of view the strand-displacement model seems to be the more important one. Whether the single-strand DNA is covered by mtSSB or RNA remains to be seen.