Advertisement

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Inheritance bias of deletion-harbouring mtDNA in yeast: The role of copy number and intracellular selection

  • Nataliia D. Kashko,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Faculty of Bioengineering and Bioinformatics, Lomonosov Moscow State University, Moscow, Russia

    ⨯
  • Georgii Muravyov,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliation Faculty of Bioengineering and Bioinformatics, Lomonosov Moscow State University, Moscow, Russia

    ⨯
  • Iuliia Karavaeva,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliation Faculty of Bioengineering and Bioinformatics, Lomonosov Moscow State University, Moscow, Russia

    ⨯
  • Elena S. Glagoleva,

    Roles Methodology, Resources, Writing – review & editing

    Affiliations Faculty of Biology, Lomonosov Moscow State University, Moscow, Russia, Vavilov Institute of General Genetics Russian Academy of Sciences, Moscow, Russia

    ⨯
  • Maria D. Logacheva,

    Roles Resources, Writing – review & editing

    Affiliations Vavilov Institute of General Genetics Russian Academy of Sciences, Moscow, Russia, Center for Molecular and Cellular Biology, Moscow, Russia

    ⨯
  • Sofya K. Garushyants,

    Roles Software, Validation, Writing – review & editing, Formal analysis

    Affiliation Faculty of Bioengineering and Bioinformatics, Lomonosov Moscow State University, Moscow, Russia

    ⨯
  • Dmitry A. Knorre

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    knorre@belozersky.msu.ru

    Affiliation A.N. Belozersky Institute of Physico-Chemical Biology, Lomonosov Moscow State University, Moscow, Russia

    ⨯

Abstract

During sexual reproduction, fungi usually inherit mtDNA from both parents, however, the distribution of the mtDNA in the progeny can be biased toward some mtDNA variants. For example, crossing Saccharomyces cerevisiae strain carrying wild type (rho+) mtDNA with the strain carrying mutant mtDNA variant with a large deletion (rho−) can produce up to 99–100% of rho− diploid progeny. Two factors could contribute to this phenomenon. First, rho− cells may accumulate more copies of mtDNA molecules per cell than wild-type cells, making rho− mtDNA the prevalent mtDNA molecule in zygotes. This consequently leads to a high portion of rho− diploid cells in the offspring. Second, rho− mtDNA may have a competitive advantage within heteroplasmic cells, and therefore could displace rho+ mtDNA in a series of generations, regardless of their initial ratio. To assess the contribution of these factors, we investigated the genotypes and phenotypes of twenty two rho− yeast strains. We found that indeed rho− cells have a higher mtDNA copy number per cell than rho+ strains. Using an in silico modelling of mtDNA selection and random drift in heteroplasmic yeast cells, we assessed the intracellular fitness of mutant mtDNA variants. Our model indicates that both higher copy numbers and intracellular fitness advantage of the rho− mtDNA contribute to the biased inheritance of rho− mtDNA.

Author summary

Eukaryotic cells contain multiple copies of mitochondrial DNA (mtDNA), which encode essential mitochondrial proteins required for cell functioning. However, some mtDNA molecules can acquire mutations – such as deletions – that eliminate protein-coding genes. Although these defective mtDNA molecules cannot sustain mitochondrial activity, certain mutant variants are thought to replicate more efficiently than functional ones. Over time, these mutants may outcompete functional mtDNA and dominate the cellular mtDNA pool. In this study, we used baker’s yeast as a model system to investigate how functional and mutated mtDNAs compete within cells. Yeast cells inherit mitochondria from both parental cells, making it possible to generate cells with mixed mtDNA types. We isolated and analysed 22 yeast strains, each carrying a different mutant mtDNA variant, and assessed how effectively these mutant mtDNAs can displace functional mtDNA. By measuring cell traits, such as growth rate and mtDNA copy number, and performing computer simulations, we discovered that some mutant mtDNAs gain a replicative advantage despite being detrimental to the cell. Moreover, we found that secondary deletions frequently occur following the initial deletion. These results shed light on the dynamics of mtDNA competition within cells and offer potential implications for understanding mitochondrial evolution.

Citation: Kashko ND, Muravyov G, Karavaeva I, Glagoleva ES, Logacheva MD, Garushyants SK, et al. (2025) Inheritance bias of deletion-harbouring mtDNA in yeast: The role of copy number and intracellular selection. PLoS Genet 21(6): e1011737. https://doi.org/10.1371/journal.pgen.1011737

Editor: Ashok Bhagwat, Wayne State University, UNITED STATES OF AMERICA

Received: October 15, 2024; Accepted: May 19, 2025; Published: June 24, 2025

Copyright: © 2025 Kashko et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All data are present within the manuscript and supplementary materials. Short-read sequencing results of rho− strains are available in the SRA archive: BioProject PRJNA1039586, SRA study SRP471445. Julia language script for simulation of mtDNA drift in yeast clonal populations is available under the link: https://github.com/dknorre/YeastHeteroplasmy

Funding: This work was supported by the Russian Science Foundation (project 22-14-00108, https://rscf.ru/en/). Grant was received by D.A.K. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Eukaryotic cells usually contain multiple not necessarily identical copies of mitochondrial DNA (mtDNA). This phenomenon is referred to as mitochondrial heteroplasmy. Heteroplasmy is linked to a variety of genetic diseases [1] and aging [2]. The presence of multiple mtDNA molecules within cells establishes a complex hierarchy of mtDNA populations. This hierarchy enables natural selection to operate at various levels of biological organisation, such as within individual cells and organisms, and between organisms [3–6]. As a result, “selfish” mtDNA variants may gain an advantage within individual cells, despite being neutral or exhibiting deleterious effects at the cellular or organismal levels [5,7–9].

Several experimental observations support the idea that mtDNA selection acts on different levels of organisation. First, pathogenic mitotypes containing mutations in the D-loop, a non-coding region of mtDNA controlling the initiation of its replication [10], gradually replace donor mtDNA during the development of a multicellular organism from an embryo [11]. Second, the common starling population in Australia was shown to carry two coexisting mtDNA variants (mitotypes). One of the mitotypes demonstrated a moderate advantage at the intracellular level while decreasing the fitness of the whole organism [12]. Next, natural isolates of nematodes Caenorhabditis briggsae harbour two distinct mtDNAs: a full-length variant and a deleterious Δnad5 variant lacking a portion of the NAD5 gene. The proportion of Δnad5 mtDNA increased across generations if the populations underwent frequent bottlenecks [13]. Finally, it has been shown that in the baker’s yeast Saccharomyces cerevisiae, the selection of mtDNA also depends on population size. In an experimental evolution study, a deleterious mtDNA variant displaced the parental wild-type variant in small populations, but not in the larger ones [14].

S. cerevisiae has been proven to be an invaluable model for exploring intracellular mtDNA selection. This yeast species can proliferate under fermentative conditions in the presence of mtDNA mutations that disrupt oxidative phosphorylation (referred to as the rho− genotype) or in the complete absence of mtDNA (rho0 genotype). Laboratory yeast strains often produce rho− mutants, i.e., having rho− genotype of their mtDNA. Such mutants typically contain large deletions in their mtDNA sequences but usually retain at least one of the eight replication origins [15–17]. Some of these replication origins, namely ORI2, ORI3 and ORI5, are reported to be more active than others [18]. Meanwhile, yeast cells with rho− and rho0 genotypes can be differentiated from the wild type (rho+) cells by their ‘petite’ phenotype. This phenotype is defined by slow growth and an inability to utilise non-fermentable carbon sources [19]. The growth rate of petite cells is also reduced because they have only limited capacity to synthesise some amino acids [20]. Additionally, S. cerevisiae has biparental mtDNA inheritance, where both gametes (mating haploid cells) transmit their mtDNAs to the diploid progeny [21], which allows easy production of heteroplasmic yeast cells.

Intriguingly, in the case of some rho− mitotypes, the rho− X rho+ cell crossings produce mainly petite diploid cells [22]. This phenomenon, known as mtDNA suppressivity, suggests that rho− mtDNAs might be positively selected, even though it would cause a detrimental (petite) phenotype. On one side, rho− mtDNAs could have a replication rate advantage in heteroplasmic cells. Indeed, some rho− mtDNAs incorporate radioactively labelled nucleotides faster than the rho+ full-length molecules [23]. On the other side, rho− haploid strains typically contain a significantly higher quantity of mtDNA molecules per cell [24]. The increased number of mtDNA molecules in rho− cells compared to rho+ cells may explain the preferential inheritance of rho− mitotypes in rho− x rho+ crossings, even in the absence of intracellular selection. However, the relative contribution of these factors to mtDNA suppressivity remains not fully characterised.

In this study, we generated a set of 22 rho− S. cerevisiae strains, mapped deletions in their mtDNA and analysed mtDNA copy number per cell and suppressivity for each mutant. To assess mtDNA copy number contribution to the degree of suppressivity, we simulated mtDNA selection and genetic drift during crossings using a stochastic model. Utilising this model, which incorporates estimates of mtDNA copy numbers per cell and growth rates of petite cells, we calculated possible suppressivity values under the assumptions of neutral drift and varying levels of replication advantage for rho− mtDNAs. Together, our experimental data and simulations support the assumption that intracellular selection favours rho− mtDNAs in heteroplasmic yeast cells.

Results

mtDNA copy number in rho− strains correlates with their suppressivity

To identify the genomic features that contribute to suppressivity, we generated 23 spontaneous rho− strains from a parental haploid strain [mat alpha rho+ URA+ leu−]. To measure their suppressivity, we crossed those strains with the rho+ strain of the opposing mating type [mat a rho+ ura− LEU+]. We defined the suppressivity of the rho− strains as the percentage of diploid progeny cells that are unable to grow on glycerol (Fig 1A). One of the low-suppressive strains was found to be rho0 and was therefore excluded from further analysis. The suppressivity of the remaining strains ranged from 19.4 to 90.2 percent (S1 Table).

thumbnail
Download:
Fig 1. Suppressivity of spontaneous rho− strains.

(A) Experimental design for estimating suppressivity; (B) S. cerevisiae mitochondrial DNA map featuring the locations of replication origins. The arcs on the map illustrate the mtDNA segments preserved in the spontaneous rho− mutants examined in this study. The colour of the arcs corresponds to the average suppressivity of the respective rho− strains. The arcs are sorted from the longest to the shortest rho− mitochondrial genomes, with the longest being the closest to the center. Replication origins with higher activity are marked with asterisks (*). (C) Correlation of rho− mtDNA copy number per nuclear genome estimated from the NGS data with suppressivity (Kendall’s rank correlation tau = 0.471, p-value = 0.001); (D) rho− mtDNA genome size does not show significant correlation with mtDNA copy number per nuclear genome (Kendall’s rank correlation tau = 0.105, p-value = 0.45); (E) The ability of rho− yeast cells to retain mtDNA upon growth with the DNA-intercalating agent ethidium bromide (BrEth) correlates with rho− mtDNA copy number (Kendall’s rank correlation tau = 0.72, p-value = 2✕10-5). Data for the rho+ strain is shown in all plots but is not included in the linear model and correlation tests.

https://doi.org/10.1371/journal.pgen.1011737.g001

Next, we isolated total DNA from the rho− strains and performed whole-genome sequencing to identify the position of deletions in their mtDNA. We show that the obtained rho− mtDNA mutants vary in the deletion length and position (Fig 1B). In 7 out of 22 strains, the position of the primary deletion was confirmed by Sanger sequencing (S1 Table). We found that the studied strains retained 76.8 to 14.8 percent of the unique parental mtDNA sequence. All strains retained at least one of the eight predicted replication origins (Fig 1B). We also sequenced a previously isolated hypersuppressive (HS) rho− strain which showed on average 99% suppressivity and retained only a short 2.2 kb fragment of mtDNA spanning a region flanked by ORI2 and ORI7 [25]. In our set of strains, we did not find a strong association between the retained replication origins and suppressivity, however, it may be noticed that two low-suppressive strains, namely IIa10 and 13, lack all three highly active replication origins: ORI2, ORI3 and ORI5 (Fig 1B).

To assess the relative mtDNA copy number for each rho− strain, we compared the average read coverage depth for mtDNA with the one for the nuclear genome (nDNA, S2 Table). However, because the yeast mitochondrial genome contains regions with very high AT content, these AT-rich sequences pose difficulties during sequencing and read mapping. As a result, mapping algorithms map fewer reads to AT-rich regions, which leads to the non-uniform read coverage along mtDNA (S1 Fig). To overcome this issue, other studies routinely assess mtDNA copy number only in the region of high GC-content spanning from position 14,000–20,000 [26], where read depth is indeed uniform (S1 Fig). Unfortunately, this region was absent in some of our rho− strains. Thus, we selected three short mtDNA regions (coordinates 8002–8153, 31222–31305, 48195–48296, based on the reference yeast mitochondrial genome) to ensure that each rho− strain retains at least one of these regions (S2 Table). Notably, these regions had the read depth 2.3-3.4 fold lower than the read depth of the GC-rich 14,000–20,000 region, suggesting that the actual mtDNA copy number estimate using these regions could be underestimated by several fold (S2 Table).

To estimate the mitochondrial DNA copy number in our rho− strains, we estimated the copy number for each of the available regions and selected the highest value among them. We validated results obtained from the procedure described above for a subset of rho− strains by quantitative PCR with three pairs of primers designed for the regions used for the NGS copy number assessment (S3 Table and S1 Fig). NGS and qPCR approaches for quantification of mtDNA/nDNA ratios showed concordant results, the values were positively correlated with Kendall’s rank correlation tau = 0.683, p-value = 8.2×10-5 (S2 Fig). However, the absolute values of the mtDNA/nDNA ratios obtained with NGS and qPCR methods substantially differed (see S2 and S4 Tables). For instance, according to NGS, the mtDNA copy number of the control wild-type strain ranged from 3.0 to 15.5 copies per haploid nuclear genome, depending on the mtDNA region selected as reference (S2 Table). In contrast, qPCR provided a higher estimate of 22.2-26.6 mtDNA copies per nuclear genome (S4 Table), which aligns better with the estimates reported in other studies [27–29].

We used the NGS-based evaluation of the mtDNA copy number and correlated it with the corresponding strain’s suppressivity. We showed that mtDNA/nDNA ratios of the rho− strains positively correlate with their average suppressivity (Figs 1C and S3, Kendall’s rank correlation tau = 0.47, p-value = 0.001). At the same time, we found no correlation between the length of the mtDNA remaining in the cells and mtDNA copy number (Fig 1D) or suppressivity (S4 Fig). Since both methods for determining the mtDNA/nDNA ratio (qPCR-based and NGS-based) relied on DNA amplification, we decided to complement those measurements with another approach. We hypothesised that rho− strains with increased mtDNA copy number would be more resilient to the loss of mtDNA in response to the inhibition of mtDNA replication. To test this, we treated rho− strains with the DNA-intercalating agent, ethidium bromide. Then, we stained the cells with DAPI to visualise nuclear and mitochondrial DNA in the cells. Our findings confirmed that rho− cells with high mtDNA/nDNA ratios were more resistant to mtDNA loss (Fig 1E). In the analysed set of rho− strains, the percentage of rho− cells retaining mtDNA after incubation with ethidium bromide correlated with the average mtDNA/nDNA ratios in the strain (Fig 1E, Kendall’s rank correlation tau = 0.72, p-value = 2✕10-5) and suppressivity (Kendall’s rank correlation tau = 0.56, p-value = 8✕10-4, S5 Fig). Taken together, three independent approaches to estimate mtDNA copy number show a positive correlation between the mtDNA copy number and suppressivity.

Suppressivity drifts after genetic bottleneck

To explore the effects of genetic drift and selection on rho− mtDNA variants in yeast cell populations, we investigated whether suppressivity remains constant in rho− strains or is subject to drift. To test this, we took several strains with varying suppressivity from our set and derived subclones by selecting colonies that presumably originated from a single cell. We evaluated the suppressivity of these subclones as described earlier. Fig 2A shows that the suppressivity of a parental strain that did not undergo a bottleneck by a subclone selection remained stable in independent experiments. At the same time, some subclones obtained from single cells showed similar suppressivity, which, however, strongly differed from the average suppressivity of the parental rho− strain. This suggests the existence of secondary mtDNA mutations in rho− strains that affect their suppressivity. To test this possibility, we sequenced ten subclones of two rho− strains, specifically IIc11 and Ia14. We observed that all subclones of IIc11 had lost an additional various portions of mtDNA, whereas in the case of Ia14, one subclone exhibited a large secondary deletion. The other clones showed similar coverage maps, with only a small portion of mtDNA lost in subclones 1 and 3 (Fig 2B).

thumbnail
Download:
Fig 2. Genetic bottlenecks change the suppressivity and enable the fixation of clones with secondary deletions.

(A) Suppressivity was measured in parental rho− yeast strains and their subclones derived from each strain after a narrow genetic bottleneck (single colonies). (B) The read depth of the mitochondrial DNA of the parental strains and their subclones reveals secondary deletions in some of the subclones. The read depth was normalised to the average coverage of nuclear DNA (nDNA) in the same sample and log-transformed to highlight low-coverage regions.

https://doi.org/10.1371/journal.pgen.1011737.g002

Our experiments demonstrated that rho− strains can alter their suppressivity following a genetic bottleneck, prompting us to investigate whether there is a consistent trend towards increased or decreased suppressivity. To examine this, we performed a post hoc analysis of the suppressivity of 76 rho− strains previously obtained in our laboratory. These strains include the rho− strains sequenced in this study, as well as additional spontaneous rho− strains for which multiple suppressivity assessments were available. They include other spontaneous rho− strains originated from the same W303 URA+ mat alpha parental rho+ strain which were not selected for sequencing, and rho− strains obtained from different parental strains with the W303 background. The initial assessment was conducted immediately after the narrow bottleneck, following the isolation of each rho− strain (S6 Fig). Subsequent analyses were carried out after an additional (non-controlled) number of generations. We found that the initial suppressivity assessment was on average lower (5.9% ± 10.9%) than the subsequent ones (S6 Fig), although there were some isolates whose initial assessments were higher than the subsequent ones. Wilcoxon signed-rank test with continuity correction rejected the hypothesis that the medians of first and subsequent assessments are equal (p-value = 9.9✕10-6). At the same time, it might be noticed that suppressivity increased to a higher extent in the strains with low initial suppressivity (S6 Fig).

rho− mtDNA suppressivity does not affect the strain growth rate

In our experimental setup for the suppressivity assessment, variation in rho− strain suppressivity could be influenced by the difference in growth rates. For example, if a highly suppressive rho− strain has a higher proliferation rate than a low suppressivity rho− strain, diploid strains with homoplasmy of the highly suppressive rho− variant can undergo more divisions than the corresponding low suppressivity rho− strain. As a result, it would produce more petite cells in the suppressivity assay. To test this possibility, we measured the maximal growth rate of our rho− and rho+ strains, as well as rho0 cells as a control. We found no pronounced differences in growth rate between rho− and rho0 strains, while the rho+ strain expectedly displayed higher maximal growth rate (Fig 3A, 3B and S5 Table). As the conditions during crossings deviate from those in an exponentially growing yeast suspension, we additionally tested the change in the number of URA+ cells (colony-forming units) before and after the crossing experiments (Fig 3C). Surprisingly, there was no difference in growth rates of rho+ and rho−/rho0 strains under these conditions, as the number of URA+ cells increased approximately tenfold regardless of the mtDNA type for all tested strains/crossings (Fig 3D).

thumbnail
Download:
Fig 3. Phenotypes of rho+, rho− and rho0 strains during the suppressivity assay experiments.

Representative growth rate (A) and maximum specific growth rates (B) of rho+, rho− and rho0 strains (shown as mean ± standard deviation based on the data from four independent experiments); (C) Experimental setup to measure growth rates of strains during genetic crosses; (D) Relative growth rates of yeast strains specified on the plot (URA+ leu−) crossed with rho+ (ura− LEU+) strain; (E) URA+ LEU+ diploid cell count produced by crossing indicated strain with (ura− LEU+) rho+ strain, normalised to initial cell count; *P < 0.01, **P < 0.001 (Wilcoxon rank sum exact test) comparing rho− X rho+ to the rho0 X rho+ crosses.

https://doi.org/10.1371/journal.pgen.1011737.g003

We reasoned that the similar growth rates observed between rho+ and rho−/rho0 strains under crossing conditions stem from the inhibition of rho+ cell growth by the mating pheromone (alpha factor), which is secreted by cells of the opposite mating type. In this case, the presence of functional mitochondria in rho+ cells might be a non-limiting factor for the growth rate. Hence, we assessed the number of diploid cells during the crossing experiment. We calculated the proportion of diploid cells divided by the number of haploid cells at the beginning of the experiment (Fig 3E). We found that the growth rate of diploids obtained with rho+ X rho0 crossings exceeded that from the rho+ X rho− crossings. This result shows that the presence of rho− mtDNA in the diploid cells inhibits their growth.

Stochastic model suggests an intracellular fitness advantage of rho− mtDNAs with deletions

Our experiments yielded estimates for three parameters of rho− cells: suppressivity, their mtDNA copy number per cell, and growth rates in crossing experiments. We reasoned that a strain’s suppressivity could be primarily derived from its growth rate and mtDNA copy number. However, it can also be influenced by the intracellular fitness of the rho− mtDNA variant if this fitness value deviates significantly from one. In order to explore the association between these parameters, we developed a stochastic model that simulates the genetic drift and selection of mtDNAs in the heteroplasmic cells. This model takes as an input the following parameters: (1) relative growth rates of rho+ and rho− strains; (2) the initial proportion of two mtDNA variants (heteroplasmy level), with the range of possible values taken from mtDNA copy number estimations; (3) intracellular fitness (imputed from the value in a range from 0 to 4). In our model, intracellular fitness was defined as the relative increase in frequency of the rho− mtDNA variant compared to the rho+ mtDNA variant within a single heteroplasmic cell during one cell division; The model simulates random drift and selection on intracellular and intercellular levels and returns the distribution of mtDNAs in the population of cells following several cellular divisions (see S7 Fig for the in detail model description). At the end of the simulation, the model yields suppressivity as the ratio of cells devoid of wild type mtDNA to the total number of cells. We ran these simulations to estimate the suppressivity with varying starting parameters of the initial heteroplasmy level and intracellular fitness of rho− mtDNA (Fig 4A). Intracellular fitness values within the range of 2–4 produced similar results for most initial heteroplasmy levels. Therefore, for further analysis, we only used the data obtained from simulations with intracellular fitness values ranging from 0 to 2. Given that intracellular fitness was an unknown variable in our simulations, we replotted the data using alternative coordinates (Fig 4B). In this representation, suppressivity is plotted against the heteroplasmy level, with the model-imputed intracellular fitness shown by a colour gradient. We subsequently overlaid our experimental data points for all 22 rho− strains onto this plot.

thumbnail
Download:
Fig 4. Stochastic modelling suggests that the starting proportion of rho− and rho+ mtDNA molecules in the zygotes cannot explain the observed suppressivity values.

(A) in silico simulation provides a suppressivity estimate for varying intracellular fitness and starting heteroplasmy level parameters; (B) Simulation data and experimental data plotted in the coordinates of starting heteroplasmy level ~ suppressivity; Points represent simulated data, crosses represent experimental data points (suppressivity and relative copy numbers of wild-type rho+ and rho− mtDNA) for the 22 rho− strains. A line connects the points obtained with the simulation with no replication advantage of rho− mtDNA (Intracellular rho− mtDNA fitness equal to 1.0); (C) Relative intracellular fitness of rho− mtDNA variants calculated from the k-nearest neighbours in (B). P = 0.00006 for the hypothesis that the population distribution mean is equal to 1.0 according to Wilcoxon signed rank test with continuity correction.

https://doi.org/10.1371/journal.pgen.1011737.g004

Fig 4B shows that most of our experimental data points fall above the green line representing a neutral assumption that the intracellular fitness values of the rho− and rho+ mtDNAs are equal, indicating that most rho− mtDNAs have an intracellular fitness advantage compared to parental rho+ mtDNA. Furthermore, we estimated the intracellular fitness levels for our set of 22 rho− strains from the simulation data. To do this, we determined k-nearest neighbours for each strain in the coordinates (heteroplasmy level ~ suppressivity) (k = 6) and took their average intracellular rho− mtDNA fitness as the estimate for our strains. Fig 4C shows the distribution of the calculated fitness values and suggests that on average, the value is significantly higher than one (fitness equal to one is the neutral expectation, assuming that there is no intracellular fitness advantage of rho− mtDNA variants).

In our model, we incorporated several assumptions. First, we posited that mutant rho− mtDNAs have a threshold effect on yeast growth rate. Specifically, we proposed that yeast cells switch from slow to high growth when the proportion of rho+ mtDNA exceeds 0.5. This assumption aligns with the generally accepted notion that deleterious effects of mtDNAs, including rho−, are manifested when the proportion of mutants is above a certain threshold [30]. Second, we imputed into the model the copy number per cell equal to 20 which was concordant with our estimates (Fig 1C) and the published data [27]. To test the robustness of our estimates, we varied pathogenicity thresholds for rho− mtDNAs (S8 Fig) and mtDNA copy number per cell at the moment of cell division (S9 Fig) in the simulations. Using the same approach mentioned above, we calculated the intracellular fitness advantage of rho− mtDNA variants using simulations with different mtDNA copy numbers. The mean value of rho− mtDNA intracellular fitness remained above one (S10 Fig), suggesting that there is an advantage within the cell for mutant mtDNAs compared to parental wild-type mtDNA, and this result is robust. Overall, our experiments and the model show that on average spontaneous deletion in yeast mtDNA provides it with a selection advantage within the cell.

Discussion

Uniparental inheritance of mtDNA in most metazoan species prevents mtDNA recombination and imposes a problem of deleterious mutation accumulation in the mitogenomes [31]. Severe mtDNA copy number bottlenecks in the germline slow down this process [32]. In contrast to metazoa, fungi usually inherit mtDNA from both parents [33,34]. In addition, fungal mtDNAs experience frequent recombination in experimental conditions [35] and natural yeast populations [36]. Biparental inheritance enables recombination, which is likely to diminish the problem of mutation accumulation [32]. However, this makes the fungi prone to selfish mitochondrial genetic elements, which show preferential transmission to the progeny despite their deleterious effects on the whole organism. The simplest premise for the transmission bias of mtDNA variants is an increase in its copy number in one of the haploid cells — gametes. This would resemble the anisogamy typical for metazoa, where female gametes contain more mitochondria and mtDNA compared to male gametes [37].

To unravel how the mtDNA copy number in wild-type and rho− S. cerevisiae cells affects their inheritance bias, we studied several rho− strains harbouring large deletions in their mitochondrial genomes. mtDNA copy number per cell in all studied rho− haploid strains exceeded that of the parental rho+ strain (Fig 1C). This observation is concordant with a previous study showing that rho− yeast cells accumulated mtDNA to a copy number ~80 times higher than that in a wild-type strain [24]. Furthermore, we have recently found that the total amount of mtDNA is higher in rho− cells, whereas in some strains, the average mtDNA amount exceeded that of rho+ cells several fold [38]. We speculate that a negative feedback loop mechanism induces mtDNA biogenesis to compensate for a deficiency in mitochondrial function that eventually leads to excessive accumulation of mtDNA. A similar mechanism was described in nematodes, where the mitochondrial unfolded protein response transcription factor ATFS-1 helped to maintain high levels of mtDNA variant harbouring a deletion by enhancing the binding of mtDNA polymerase to mutated mtDNA molecules [39]. To the best of our knowledge, such a mechanism is still unknown in yeast. Meanwhile, deletion of the COX5 gene, encoding an essential component of cytochrome oxidase, the terminal part of the respiratory chain, increases mtDNA copy number per cell [40]. Together, these observations suggest that the ability to encode essential components of OxPhos is necessary for maintaining constant mtDNA levels across generations. If this is true, rho− cells lacking such regulation would experience drift in mtDNA copy number, resulting in highly heterogeneous mtDNA copy numbers both per cell and per clone. This hypothesis is consistent with the previous observation that rho− cell suspensions are more heterogeneous in terms of mtDNA levels per cell compared to their parental rho+ strains [38]. Moreover, this hypothesis is supported by the observed variability in mtDNA copy number among rho− strains in this study (Fig 1C–E). However, further studies are necessary to elucidate the properties and mechanisms underlying the regulatory processes linking OxPhos and mtDNA maintenance.

We did not find a positive correlation between genome size and either mtDNA copy number or suppressivity in the rho− strains (Figs 1D and S4). On the one hand, the absence of such correlation could be explained by the absence of contribution of the sequence length to the rho− mtDNA variant suppressivity. On the other hand, it may also result from the presence of mtDNAs with secondary deletions in the rho− strains. Such mtDNAs if accumulated in high copy numbers can strongly affect suppressivity.

The simplest explanation of the suppressivity phenomenon could be the observed high copy number of mtDNA molecules in the rho− strains followed by random segregation of the molecules (Fig 1C). However, by imputing mtDNA/nDNA ratios and relative growth rates of rho+ and rho− strains we were able to explain the rho− strains’ variance in suppressivity only partially (Fig 4B). It should be mentioned that in yeast, mtDNA molecules are usually concatenated [41,42]. Meanwhile, both methods (qPCR and NGS) used in our study to assess mtDNA copy numbers cannot distinguish molecules containing one repeat unit from concatenated molecules. To estimate the intracellular fitness of rho− mtDNA variants, we did not take into account that rho− mtDNAs consist of a smaller number of segregating units. However, if this were the case, we would overestimate the ratio of rho− to rho+ mtDNA in zygotes (the heteroplasmy parameter of the model) and underestimate intracellular fitness in our prediction. In other words, even if each rho− mitotype was represented by a separate molecule segregating independently during cell division, we could not explain the observed values of suppressivity by the initial ratio of mtDNA molecules in rho− and rho+ strains crossing. Therefore, we conclude that in yeast, some mtDNA variants with large deletions can have pronouncedly higher intracellular fitness than the full-size wild type mtDNAs.

In line with our results indicating that some rho− mtDNAs can have higher intracellular fitness than others, we have shown that the suppressivity of yeast rho− clones can be changed after the genetic bottleneck (Fig 2) and, on average, increase with generations (S6 Fig). We showed that passing a rho− strain through a single colony fixes secondary mutations in its mtDNA that can contribute to the increase in its suppressivity (Fig 2B). The gradual increase in suppressivity could be attributed to two factors. On the one hand, cells with secondary mutations in mtDNA leading to a higher suppressivity proliferate faster than parental cells with the original mutation. However, Fig 3 shows that there is no pronounced difference between the growth rate of rho− strains with different deletions. On the other hand, spontaneous secondary deletions in rho− mtDNA could lead to the appearance of mtDNA variants with intracellular fitness exceeding the intracellular fitness of the original rho− mitotype. Several experiments in our study support the second possibility. First, different rho− mtDNA molecules can have various intracellular fitness (Fig 4C). Therefore, rho− cells will eventually produce new rho− variants with increased intracellular fitness, as evidenced by the post hoc analysis in S6 Fig. Due to an increase in intracellular fitness, this variant is expected to displace parental rho− variants within the same cell and all descendants of this cell. In the presence of rho+ cells, such variants are expected to be removed from the population because of the selection on the cell level (see a hypothetical scheme in Fig 5A). Meanwhile, in petite cell populations, newly emerging mtDNA variants with low intracellular fitness are expected to be outcompeted and removed from the population, while the mtDNA with high intracellular fitness will repeatedly emerge and displace the parental variants within cell lineages (Fig 5B). We propose that as a result, a ratchet-like mechanism leads to a gradual increase in the proportion of cells carrying suppressive rho− mtDNA within small yeast populations. However, this process should be limited by the maximal possible intracellular fitness of the rho− mtDNA.

thumbnail
Download:
Fig 5. MtDNA dynamics in yeast suspension, a hypothetical scheme.

(A) rho− petite cells are continuously generated and selected out from the population of rho+ cells; (B) Intracellular selection favours the fixation of various hypersuppressive (HS) rho− variants within the rho− petite cell population.

https://doi.org/10.1371/journal.pgen.1011737.g005

What could be the mechanism ensuring rho− mtDNA intracellular fitness advantage in heteroplasmic cells? First, the advantage may be attributed to the reduced replication time of mtDNA molecules harbouring large deletions but retaining replication origins. Similar to PCR reactions, the amplification of shorter mtDNA molecules is more robust and proceeds more efficiently, especially under conditions of nucleotide deficiency. Indeed, the probability of rho+ mtDNA displacement appears to be negatively correlated with the size of the remaining mtDNA sequence [17] and positively correlated with the density of replication origins retained in a deletion variant of mtDNA [42]. Furthermore, the partial inhibition of dNTP synthesis by the overexpression of SML1, encoding the inhibitor of ribonucleotide reductase RNR1, strongly increased, and its deletion decreased the suppressivity [43]. Therefore, it might be that rho− mtDNAs simultaneously cause problems with the biogenesis of essential DNA replication components (e.g., nucleotides and ATP) and gain a replication advantage because of these deficiencies.

Second, mtDNA may gain an advantage via preferential diffusion of mutant mtDNAs to the daughter cells. Upon cell division, yeast daughter cells receive a limited number of mtDNAs amplified from a few randomly selected clones. This leads to a rapid loss of heteroplasmy in yeast cells because the progeny receives only one or another mitotype [44,45]. This mechanism is likely to rely on the limited diffusion of mtDNA molecules and mtDNA-encoded products along the mitochondrial network. The limited diffusion creates a linkage between mtDNA genotypes and phenotypes, which is considered to be essential for quality control of mtDNAs [46,47]. Cristae appear to be an essential factor preventing the diffusion of protein components of large protein complexes along mitochondria [48]. Accordingly, deletion of MIC (mitochondrial cristae organisation) genes prevents preferential inheritance of the wild type mtDNA to a bud [49]. We speculate that rho− mtDNAs showing intracellular advantage can avoid this sequestration, e.g., by the structural features of the spatial structure within the nucleoid. As a result, in contrast to the wild type mitotypes, all daughter cells can receive a portion of mutant mtDNAs. Whatever the mechanism, our study suggests that the inheritance bias of rho− mtDNAs is affected by variation in the intracellular fitness of the mutant mitotypes.

Biparental mtDNA inheritance repeatedly produces mitochondrial heteroplasmy, which creates the conditions for selfish mitochondrial genetic elements to spread in the fungal populations. Here, we provide rough estimates of the intracellular fitness ratio of mtDNA variants with deletions. Our data demonstrate that yeast mtDNA variants with large deletions can exhibit high intracellular fitness, challenging the fungal cells’ ability to maintain a functional mitochondrial genome containing all essential components for translation and oxidative phosphorylation.

Methods

Yeast strains, growth conditions and reagents

All strains used in this study are descendants from W303 (ade2–101 his3–11 trp1–1 ura3–52 can1–100 leu2–3, 112, GAL, psi+) mat a or mat alpha strains used in our previous study [25]. Rich and complete synthetic media compositions were prepared according to Sherman [50] and manufacturer’s instructions (Sigma-Aldrich). All incubations were performed at 30°C with shaking (200 rpm) if not indicated otherwise.

To obtain the spontaneous rho− strains, the parental rho+ strain W303 URA+ mat alpha pre-grown in liquid YPD to the exponential growth phase was plated onto YPDGly medium (containing 0.1% glucose and 2% glycerol). Colonies displaying a petite phenotype were amplified on solid YPD and used to assess suppressivity. Subsequent suppressivity measurements were performed from frozen culture stocks. For selected rho− strains, we also assessed suppressivity of individual subclones isolated from the stock culture by picking single colonies.

The rho0 strain was obtained by incubating exponentially growing rho+ cells in YPD with ethidium bromide added to a final concentration of 0.1 mg/mL. The cells were further incubated overnight (~16 hours) at 30°C with shaking, then diluted and plated on solid YPDGly medium (0.1% D-glucose and 2% glycerol) to obtain approximately 200 single colonies. Selected colonies with a petite phenotype were confirmed to be rho0 by their inability to grow on glycerol-containing media (YPGly) and the absence of DAPI-positive foci in the cytoplasm (see S11 Fig). For this, cells were fixed in 70% ethanol and incubated with DAPI at a final concentration of 2 µg/mL for 30 minutes, then resuspended in 1x PBS and analyzed under Olympus BX51 microscope with the following filter set: U-MNU2 filter for DAPI (excitation λ = 360 nm – 370 nm; dichroic mirror λ = 400 nm; emission wavelength > 420 nm).

Yeast DNA isolation and quantitative PCR

To measure mtDNA/nDNA ratio, we took exponentially growing yeast cells and isolated total DNA using a standard protocol which is based on the preparation of protoplasts and their lysis using SDS [51]. We carried out qPCR on the CFX96 Touch Real-Time PCR Detection System according to the manufacturer’s instructions. To quantify PCR products, we used EvaGreen-ROX (Syntol) as the dsDNA probe. To detect mtDNA copy numbers in rho− mutants, we designed three pairs of primers annealing to the different loci of the mtDNA genome (S3 Table). As a reference to the nuclear DNA we used primers to the ACT1 gene (S3 Table).

Library preparation and next generation sequencing of mitochondrial genomes

HS rho− strain was sequenced using Illumina MiSeq platform and MiSeq Reagent Kit Version 3 (150 cycle). Selected subclones of two rho− strains were sequenced with the Illumina NovaSeq6000 platform (100 cycle, paired-end sequencing). All other strains were sequenced using Illumina HiSeq platform, using HiSeq Flow Cell v3, TruSeq SBS Kit v3 (50 cycle) and TruSeq PE Cluster Kit v3 and on the Illumina Nextseq platform, using the Nextseq Mid Output Kit (150 cycle, paired-end sequencing). Libraries were prepared using the TruSeq RNA Library Prep Kit or NEBNext Ultra II DNA Library Prep Kit for Illumina and the NEBNext Multiplex Oligos for Illumina (Index Primers Set 1).

Obtained reads were trimmed with trimmomatic v. 0.32 with default parameters [52] and mapped with bowtie2 [53] to the reference S288C Saccharomyces cerevisiae genome (sacCer3, NCBI RefSeq assembly GCF_000146045.2).

Quantification of mtDNA to nDNA ratio using Next Generation Sequencing data

To quantify mtDNA copy numbers, we compared read coverage of the mitochondrial and nuclear genomes; this approach has been previously used in other studies [26,54]. First, we mapped the sequence reads to the reference S. cerevisiae genome using bowtie2 [53]. After this, we calculated the nDNA average read coverage using Rsamtools (https://bioconductor.org/packages/Rsamtool). To negate the effect of rRNA repeat and telomeric regions which have higher read counts, we discarded nuclear genomic regions with extreme quantiles from the analysis. Thereafter, we calculated average mitochondrial genome read coverage. Given poor coverage of AT-rich genomic regions, to quantify mitochondrial genome read coverage, we analysed the coverage only in the GC-rich regions of the mitochondrial genome. Next, we calculated the coverage in the positions used for the quantitative PCR (see S1 Fig). Given that rho− strains retained random regions of the parental genome, we calculated the average coverage in all three regions separately and then took the maximum of the obtained values.

Relative read depth in different regions of mtDNA is presented in S2 Table.

Suppressivity assessment

Suppressivity was estimated as the proportion of petite diploid colonies in the total number of diploid progeny as described previously [55] with several modifications. Prior to the experiment, strains of two different mating types were grown in rich medium YPD to exponential phase (OD550 = 0.4-0.8). Strains were mixed in a 2 mL eppendorf tube or 96-well plate to the final OD550 = 0.1 of each strain (2 × 106 cells/mL) in the total volume of 200 µL. Matings were performed at room temperature without mixing. After 20 h of mating cell suspensions were diluted with mQ water (10–1000 times depending on the mating efficiency for the given pair of strains) and plated on double-selective media (e.g., YNB -LEU -URA) containing 2% glycerol and 0.1% glucose as carbon sources. Such media allows selecting only diploid cells carrying both selective markers and furthermore distinguishing petite and grande diploid colonies. After 2–3 days we counted the number of petite and grande colonies and calculated the suppressivity.

Assessment of ethidium-bromide-induced mtDNA loss

To quantify the proportion of rho− cells retaining mtDNA after incubation with ethidium bromide, cells were pre-grown to OD550 = 0.1 in YPD, then incubated in 2 mL YPD with 25 µg/mL ethidium bromide for 16 hours. After that, cells were washed once with YPD and incubated in fresh YPD for 6 more hours. To check for mtDNA retention, cells were stained with DAPI as described above. 300–1000 cells were counted per strain for the presence of the DAPI signal in the cytoplasm.

Growth rate assessment

Growth rate measurement was performed in a spectrophotometer (SpectrostarNANO) with automated OD550 measurement at 5-minute intervals. Analysed strains were pre-grown in YPD to exponential phase, then diluted in YPD to an equal density of 106 cells/mL (OD550 = 0.05) and inoculated into a 96-well plate (Eppendorf) at 100 µL per well. Plates were incubated in the spectrophotometer at 30°C for up to 18 hours, with orbital shaking before measurements (500 rpm for 30 sec).

To determine the maximum growth rates, we ln-transformed the OD values and computed the slopes for each sliding window (50 points in size). The maximum value was selected; see the script and example fitting available on GitHub: https://github.com/Knorre-lab/GrowthCurves). It should be noted that the calculated maximal growth rate (µmax) values may be biased, as they were measured in batch culture experiments where OD may not linearly depend on cell density in dense cultures.

To assess growth rates of rho+ and rho− strains in the conditions of the suppressivity-measurement experiment, we quantified CFUs in the spot assays sampled from the crossing cultures. Crossings were set as described above, by mixing two strains to a final OD550 = 0.1 of each strain (2 × 106 cells/mL) in a total volume of 200 µL, and incubated at room temperature for 20 h without mixing. At the first (0 h) and last (20 h) timepoints, 10 µL of the mating cultures were used to prepare five serial dilutions in sterile mQ water (1:10–1:10-6 depending on the mating efficiency). Each dilution was plated on YPD, single-selective (YNB -LEU and YNB -URA) and double-selective (YNB -LEU -URA) media containing 2% glucose. CFUs were quantified after 1 (for YPD) or 2 days (for selective media) of incubation at 30°C. Only diploid progeny could form colonies on the double-selective medium, while the single selective medium allowed growth of only one of the parental strains, and YPD allowed growth of all cells. These numbers were used to calculate the growth rate of both diploid progeny and either of the parental strains.

Stochastic model

To simulate genetic drift and dual level of selection in yeast we wrote a stochastic model using Julia language. The script is available here: (https://github.com/dknorre/YeastHeteroplasmy and illustrated in S7 Fig. An individual cell was defined as an entity which can harbor a predefined number of wild type and mutated mtDNA molecules, has an individual grande or petite phenotype tag and the parameter indicating the time spent after the previous cell division. During each simulation, the total number of mtDNA molecules within a cell remains constant, while the proportion of mutated and wild-type mtDNA can change due to selection and stochastic drift.

In each simulation, (1) a set of 103 cells with specified levels of rho− and rho+ molecules is generated. The time elapsed since the last cell division is assigned randomly to each cell, which helps to avoid potential artifacts that might arise from the simultaneous division of cells, specifically the synchronisation that occurs at the start of the simulation. (2) For each cell, the algorithm calculated the phenotype which was set as a function of rho−/rho+ mtDNA molecules ratio. In this study, we assumed the threshold dependencies of the phenotype on the ratio of rho+/rho− mtDNAs, although to improve robustness, we tested different values of the threshold (S8 Fig). The petite and grande phenotypes were represented by duplication time, i.e., the number of minutes necessary to produce a new daughter cell. The values were taken from the experiments in Fig 3B and set as 140 minutes for grande phenotype and 210 minutes for petite phenotype. However, it should be noted that the actual growth rates of grande and petite cells in the diploid state during crossing experiments may differ from these values. (3) The simulation incremented time which relates to cell division. Each cell that reached its duplication time (according to the phenotype) divided and produced two cells with randomly distributed rho− and rho+ mtDNAs. To do this, we randomly sampled rho+ and rho− mtDNAs into the new cells. Importantly, before the sampling, we doubled the number of rho+ mtDNAs and multiplied the number of rho− mtDNAs by 2 ✕ Intracellular fitness value. (4) After this, we returned to step two and repeated the simulation for 1440 minutes.

Data analysis and visualisation

All data analysis and generation of figures were performed using R language version 4.4.0 and tidyverse libraries [56]. Where appropriate, we have displayed individual data points alongside boxplots using default parameters from the R ggplot2 library. These individual data points represent either separate strains or measurements taken on different dates, depending on the experimental setup. By default, the boxplot borders denote the first and third quartiles, while the whiskers extend from these hinges to the largest and smallest values within 1.5 times the interquartile range (IQR). The values outside of this range are plotted as outliers. Circular map with mtDNA deletion variants were generated with circlize R package [57].

Supporting information

S1 Table. Deletion coordinates of rho− strains.

The coordinates were deduced from short-read sequencing assemblies and, for some strains, verified by PCR and Sanger sequencing.

https://doi.org/10.1371/journal.pgen.1011737.s001

(DOCX)

S2 Table. Quantification of mtDNA copy number from NGS data: depth of reads mapped to mtDNA normalised to depth of reads mapped to the nuclear genome.

Trim mean and median values were calculated for the entire mitochondrial genomes. In the rest of the columns, the average read depth was calculated in specified positions.

https://doi.org/10.1371/journal.pgen.1011737.s002

(DOCX)

S3 Table. Primers used for quantitative PCR (qPCR) of mtDNA to nDNA ratio.

https://doi.org/10.1371/journal.pgen.1011737.s003

(DOCX)

S4 Table. Quantification of mtDNA copy number using qPCR.

The mtDNA copy number was determined as the ratio of mitochondrial to nuclear genome loci abundance, calculated as -logCt. The positions of the amplified regions 1 and 3 used for the mtDNA quantification are illustrated in S1 Fig.

https://doi.org/10.1371/journal.pgen.1011737.s004

(DOCX)

S5 Table. Raw data of growth characteristics for yeast haploid and diploid strains used in Fig 3.

https://doi.org/10.1371/journal.pgen.1011737.s005

(XLS)

S1 Fig. Relative position of the regions used for mtDNA quantitative PCR and the regions retained in rho− mitochondrial genomes.

https://doi.org/10.1371/journal.pgen.1011737.s006

(PDF)

S2 Fig. Concordance between mean qPCR estimates and NGS-based estimates of mtDNA copy numbers. Kendall’s rank correlation tau = 0.683, p-value = 8.266e-05.

https://doi.org/10.1371/journal.pgen.1011737.s007

(PDF)

S3 Fig. Correlation of rho− mtDNA copy number per nuclear genome estimated from NGS data with suppressivity (as in Fig 1C, but with the strain names added as point labels).

https://doi.org/10.1371/journal.pgen.1011737.s008

(PDF)

S4 Fig. No apparent correlation between the genome lengths of rho− mtDNA and their suppressivity.

https://doi.org/10.1371/journal.pgen.1011737.s009

(PDF)

S5 Fig. The ability of rho− yeast cells to retain mtDNA upon growth with the DNA-intercalating agent ethidium bromide correlates with the rho− strain’s suppressivity (Kendall’s rank correlation tau = 0.56, p-value = 8×10−4).

https://doi.org/10.1371/journal.pgen.1011737.s010

(PDF)

S6 Fig. Suppressivity increases within yeast subclones.

(A) Schematic showing the experimental setup for measuring the suppressivity drift over generations; (B) Change in suppressivity between the first (green circles) and subsequent (orange and blue circles) assessments (n = 76). The orange vs blue color of the circle illustrates the direction of change in subsequent experiments compared to the first one, with orange indicating a decrease and blue indicating an increase of suppressivity over generations. X labels the results of individual experiments.

https://doi.org/10.1371/journal.pgen.1011737.s011

(PDF)

S7 Fig. mtDNA selection in heteroplasmic yeast cells: a simulation guided by experimental data.

(A) Simulation algorithm, see details in the Methods section. (B) The simulation predicts the proportion of cells containing only rho+ mtDNA (represented by the blue area) and those with only rho− mtDNA (represented by the red area). Each plot represents an overlay of the results of ten simulations, to overlay blue and red areas that illustrate homoplasmic cells we set their opacity to 0.1. The not coloured areas correspond to the cell retaining heteroplasmy. Three plots represent simulation results with different intracellular fitness levels of rho− mtDNA (1, 1.3, and 1.6). The relative cell-level fitness of rho− cells was simulated using doubling times of 140 minutes and 190 minutes. The pathogenicity threshold was set at 0.5, indicating that a rho− cell-level phenotype only manifests when more than half of the mtDNA in a cell is rho−.

https://doi.org/10.1371/journal.pgen.1011737.s012

(PDF)

S8 Fig. Variation of pathogenicity threshold in the simulations of suppressivity.

Simulation data and experimental data plotted in the coordinates of starting heteroplasmy level ~ suppressivity (same as in the Fig 4B); Points represent simulated data, crosses represent experimental data points for the 22 rho− strains; a line connects the points obtained with the simulation with no replication advantage of rho− mtDNA (Intracellular rho− mtDNA fitness equal to 1.0).

https://doi.org/10.1371/journal.pgen.1011737.s013

(PDF)

S9 Fig. Variation of mtDNA copy number (nmtDNA) in the simulations of suppressivity.

Simulation data and experimental data plotted in the coordinates of starting heteroplasmy level ~ suppressivity (same as in the Fig 4B); Points represent simulated data, crosses represent experimental data points for the 22 rho− strains; a line connects the points obtained with the simulation with no replication advantage of rho− mtDNA (Intracellular rho− mtDNA fitness equal to 1.0).

https://doi.org/10.1371/journal.pgen.1011737.s014

(PDF)

S10 Fig. Relative intracellular fitness of rho− mtDNA variants calculated from the k-nearest neighbours for simulations with different mtDNA copy numbers (number of mtDNA segregating units).

Individual lines illustrate the predictions for individual rho− strains, and the red bold line shows the average value.

https://doi.org/10.1371/journal.pgen.1011737.s015

(PDF)

S11 Fig. DAPI staining of rho0 and rho− cells.

Arrows indicate putative mitochondrial nucleoid foci in rho− cells. Images are representative and were acquired and processed independently on different days.

https://doi.org/10.1371/journal.pgen.1011737.s016

(PDF)

Acknowledgments

We are grateful to Maria Kasyanova for the valuable help in confirming deletion sites in mtDNAs and Ekaterina Smirnova for conducting some of the crossing experiments. We acknowledge the Skoltech Genomics Core Facility for their assistance in sequencing the yeast subclones.

References

  1. 1. Taylor RW, Turnbull DM. Mitochondrial DNA mutations in human disease. Nat Rev Genet. 2005;6:389–402.
  2. 2. Kraytsberg Y, Kudryavtseva E, McKee AC, Geula C, Kowall NW, Khrapko K. Mitochondrial DNA deletions are abundant and cause functional impairment in aged human substantia nigra neurons. Nat Genet. 2006;38(5):518–20. pmid:16604072
  3. 3. Rand DM. The units of selection on mitochondrial DNA. Annu Rev Ecol Syst. 2001;32:415–48.
  4. 4. Ma H, O’Farrell PH. Selfish drive can trump function when animal mitochondrial genomes compete. Nat Genet. 2016;48(7):798–802. pmid:27270106
  5. 5. Klucnika A, Ma H. A battle for transmission: the cooperative and selfish animal mitochondrial genomes. Open Biol. 2019;9(3):180267. pmid:30890027
  6. 6. Gitschlag BL, Tate AT, Patel MR. Nutrient status shapes selfish mitochondrial genome dynamics across different levels of selection. Elife. 2020;9:e56686. pmid:32959778
  7. 7. Haig D. Intracellular evolution of mitochondrial DNA (mtDNA) and the tragedy of the cytoplasmic commons. Bioessays. 2016;38(6):549–55. pmid:27062292
  8. 8. Havird JC, Forsythe ES, Williams AM, Werren JH, Dowling DK, Sloan DB. Selfish mitonuclear conflict. Curr Biol. 2019;29:R496–511.
  9. 9. Sequeira AN, O’Keefe IP, Katju V, Bergthorsson U. Friend turned foe: selfish behavior of a spontaneously arising mitochondrial deletion in an experimentally evolved Caenorhabditis elegans population. G3 (Bethesda). 2024;14(4):jkae018. pmid:38261394
  10. 10. Shadel GS, Clayton DA. Mitochondrial DNA maintenance in vertebrates. Annu Rev Biochem. 1997;66:409–35. pmid:9242913
  11. 11. Kang E, Wu J, Gutierrez NM, Koski A, Tippner-Hedges R, Agaronyan K, et al. Mitochondrial replacement in human oocytes carrying pathogenic mitochondrial DNA mutations. Nature. 2016;540(7632):270–5. pmid:27919073
  12. 12. Rollins LA, Woolnough AP, Fanson BG, Cummins ML, Crowley TM, Wilton AN, et al. Selection on mitochondrial variants occurs between and within individuals in an expanding invasion. Mol Biol Evol. 2016;33: 995–1007.
  13. 13. Clark KA, Howe DK, Gafner K, Kusuma D, Ping S, Estes S, et al. Selfish little circles: transmission bias and evolution of large deletion-bearing mitochondrial DNA in Caenorhabditis briggsae nematodes. PLoS One. 2012;7(7):e41433. pmid:22859984
  14. 14. Taylor DR, Zeyl C, Cooke E. Conflicting levels of selection in the accumulation of mitochondrial defects in Saccharomyces cerevisiae. Proc Natl Acad Sci U S A. 2002;99(6):3690–4. pmid:11891344
  15. 15. Dimitrov LN, Brem RB, Kruglyak L, Gottschling DE. Polymorphisms in multiple genes contribute to the spontaneous mitochondrial genome instability of Saccharomyces cerevisiae S288C strains. Genetics. 2009;183(1):365–83. pmid:19581448
  16. 16. Stenberg S, Li J, Gjuvsland AB, Persson K, Demitz-Helin E, González Peña C, et al. Genetically controlled mtDNA deletions prevent ROS damage by arresting oxidative phosphorylation. Elife. 2022;11:e76095. pmid:35801695
  17. 17. de Zamaroczy M, Marotta R, Faugeron-Fonty G, Goursot R, Mangin M, Baldacci G, et al. The origins of replication of the yeast mitochondrial genome and the phenomenon of suppressivity. Nature. 1981;292(5818):75–8. pmid:7024821
  18. 18. Turk EM, Das V, Seibert RD, Andrulis ED. The mitochondrial RNA landscape of Saccharomyces cerevisiae. PLoS One. 2013;8(10):e78105. pmid:24143261
  19. 19. Sherman F. Respiration-deficient mutants of yeast. I. Genetics. Genetics. 1963;48(3):375–85. pmid:13977171
  20. 20. Vowinckel J, Hartl J, Marx H, Kerick M, Runggatscher K, Keller MA, et al. The metabolic growth limitations of petite cells lacking the mitochondrial genome. Nat Metab. 2021;3(11):1521–35. pmid:34799698
  21. 21. Linnane AW, Saunders GW, Gingold EB, Lukins HB. The biogenesis of mitochondria. V. Cytoplasmic inheritance of erythromycin resistance in Saccharomyces cerevisiae. Proc Natl Acad Sci U S A. 1968;59:903–10.
  22. 22. Ephrussi B, de Margerie-Hottinguer H, Roman H. Suppressiveness: a new factor in the genetic determinism of the synthesis of respiratory enzymes in yeast. Proc Natl Acad Sci U S A. 1955;41(12):1065–71. pmid:16589797
  23. 23. Chambers P, Gingold E. A direct study of the relative synthesis of petite and grande mitochondrial DNA in zygotes from crosses involving suppressive petite mutants of Saccharomyces cerevisiae. Curr Genet. 1986;10(8):565–71. pmid:3327610
  24. 24. Fangman WL, Henly JW, Churchill G, Brewer BJ. Stable maintenance of a 35-base-pair yeast mitochondrial genome. Mol Cell Biol. 1989;9:1917–21.
  25. 25. Karavaeva IE, Golyshev SA, Smirnova EA, Sokolov SS, Severin FF, Knorre DA. Mitochondrial depolarization in yeast zygotes inhibits clonal expansion of selfish mtDNA. J Cell Sci. 2017;130(7):1274–84. pmid:28193734
  26. 26. Puddu F, Herzog M, Selivanova A, Wang S, Zhu J, Klein-Lavi S, et al. Genome architecture and stability in the Saccharomyces cerevisiae knockout collection. Nature. 2019;573(7774):416–20. pmid:31511699
  27. 27. Göke A, Schrott S, Mizrak A, Belyy V, Osman C, Walter P. Mrx6 regulates mitochondrial DNA copy number in Saccharomyces cerevisiae by engaging the evolutionarily conserved Lon protease Pim1. Mol Biol Cell. 2020;31(7):527–45. pmid:31532710
  28. 28. Galeota-Sprung B, Fernandez A, Sniegowski P. Changes to the mtDNA copy number during yeast culture growth. R Soc Open Sci. 2022;9(7):211842. pmid:35814911
  29. 29. Potapenko EY, Kashko ND, Knorre DA. Flow-cytometry reveals mitochondrial DNA accumulation in Saccharomyces cerevisiae cells during cell cycle arrest. Front Cell Dev Biol. 2024;12:1497652. pmid:39737342
  30. 30. Rossignol R, Faustin B, Rocher C, Malgat M, Mazat J-P, Letellier T. Mitochondrial threshold effects. Biochem J. 2003;370(Pt 3):751–62. pmid:12467494
  31. 31. Popadin K, Polishchuk LV, Mamirova L, Knorre D, Gunbin K. Accumulation of slightly deleterious mutations in mitochondrial protein-coding genes of large versus small mammals. Proc Natl Acad Sci U S A. 2007;104: 13390–5.
  32. 32. Edwards DM, Røyrvik EC, Chustecki JM, Giannakis K, Glastad RC, Radzvilavicius AL, et al. Avoiding organelle mutational meltdown across eukaryotes with or without a germline bottleneck. PLoS Biol. 2021;19(4):e3001153. pmid:33891583
  33. 33. Birky CWJ. The inheritance of genes in mitochondria and chloroplasts: laws, mechanisms, and models. Annu Rev Genet. 2001;35:125–48.
  34. 34. Wilson AJ, Xu J. Mitochondrial inheritance: diverse patterns and mechanisms with an emphasis on fungi. Mycology. 2012;3(2):158–66.
  35. 35. Fritsch ES, Chabbert CD, Klaus B, Steinmetz LM. A genome-wide map of mitochondrial DNA recombination in yeast. Genetics. 2014;198(2):755–71. pmid:25081569
  36. 36. Leducq J-B, Henault M, Charron G, Nielly-Thibault L, Terrat Y, Fiumera HL, et al. Mitochondrial recombination and introgression during speciation by hybridization. Mol Biol Evol. 2017;34(8):1947–59. pmid:28444332
  37. 37. Sato K, Sato M. Multiple ways to prevent transmission of paternal mitochondrial DNA for maternal inheritance in animals. J Biochem. 2017;162(4):247–53. pmid:28981751
  38. 38. Potapenko EY, Kashko ND, Knorre DA. Spontaneous mutations in Saccharomyces cerevisiae mtDNA increase cell-to-cell variation in mtDNA amount. Int J Mol Sci. 2023;24.
  39. 39. Yang Q, Liu P, Anderson NS, Shpilka T, Du Y, Naresh NU, et al. LONP-1 and ATFS-1 sustain deleterious heteroplasmy by promoting mtDNA replication in dysfunctional mitochondria. Nat Cell Biol. 2022;24(2):181–93. pmid:35165413
  40. 40. Lebedeva MA, Shadel GS. Cell cycle- and ribonucleotide reductase-driven changes in mtDNA copy number influence mtDNA Inheritance without compromising mitochondrial gene expression. Cell Cycle. 2007;6(16):2048–57. pmid:17721079
  41. 41. Dujon B. Mitochondrial genetics revisited. Yeast. 2020;37(2):191–205. pmid:31691343
  42. 42. Nunn CJ, Goyal S. Contingency and selection in mitochondrial genome dynamics. Elife. 2022;11:e76557. pmid:35404229
  43. 43. Bradshaw E, Yoshida M, Ling F. Regulation of small mitochondrial DNA replicative advantage by ribonucleotide reductase in Saccharomyces cerevisiae. G3 (Bethesda). 2017;7(9):3083–90. pmid:28717049
  44. 44. Ling F, Shibata T. Mhr1p-dependent concatemeric mitochondrial DNA formation for generating yeast mitochondrial homoplasmic cells. Mol Biol Cell. 2004;15(1):310–22. pmid:14565971
  45. 45. Roussou R, Metzler D, Padovani F, Thoma F, Schwarz R, Shraiman B, et al. Real-time assessment of mitochondrial DNA heteroplasmy dynamics at the single-cell level. EMBO J. 2024;43(22):5340–59. pmid:39103491
  46. 46. Busch KB, Kowald A, Spelbrink JN. Quality matters: how does mitochondrial network dynamics and quality control impact on mtDNA integrity?. Philos Trans R Soc Lond B Biol Sci. 2014;369(1646):20130442. pmid:24864312
  47. 47. Knorre DA. Intracellular quality control of mitochondrial DNA: evidence and limitations. Philos Trans R Soc Lond B Biol Sci. 2020;375(1790):20190176. pmid:31787047
  48. 48. Wilkens V, Kohl W, Busch K. Restricted diffusion of OXPHOS complexes in dynamic mitochondria delays their exchange between cristae and engenders a transitory mosaic distribution. J Cell Sci. 2013;126(Pt 1):103–16. pmid:23038773
  49. 49. Jakubke C, Roussou R, Maiser A, Schug C, Thoma F, Bunk R, et al. Cristae-dependent quality control of the mitochondrial genome. Sci Adv. 2021;7(36):eabi8886. pmid:34516914
  50. 50. Sherman F. Getting started with yeast. In: Guthrie C, Fink GR, editors. Methods in Enzymology. Academic Press; 2002. pp. 3–41.
    • 51. Heintz N, Gong S. Small-Scale Preparations of Yeast DNA. Cold Spring Harb Protoc. 2020;2020(10). pmid:32763978
    • 52. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30(15):2114–20. pmid:24695404
    • 53. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat Methods. 2012;9(4):357–9. pmid:22388286
    • 54. De Chiara M, Friedrich A, Barré B, Breitenbach M, Schacherer J, Liti G. Discordant evolution of mitochondrial and nuclear yeast genomes at population level. BMC Biol. 2020;18(1):49. pmid:32393264
    • 55. Galkina KV, Zyrina AN, Golyshev SA, Kashko ND, Markova OV, Sokolov SS, et al. Mitochondrial dynamics in yeast with repressed adenine nucleotide translocator AAC2. Eur J Cell Biol. 2020;99(2–3):151071. pmid:32057484
    • 56. Wickham H, Averick M, Bryan J, Chang W, McGowan L, François R, et al. Welcome to the Tidyverse. J Open Source Softw. 2019;4(43):1686.
    • 57. Gu Z, Gu L, Eils R, Schlesner M, Brors B. circlize Implements and enhances circular visualization in R. Bioinformatics. 2014;30(19):2811–2. pmid:24930139