Mitochondrial Genome Instability in W303-SK1 Yeast Cytoplasmic Hybrids

Simple Summary

Mitochondrial respiratory chain subunits are encoded by both mitochondrial and nuclear genomes, imposing evolutionary constraints due to potential incompatibilities between mutations in these genomes. Our study demonstrates that swapping mitochondrial DNA (mtDNA) between two laboratory yeast strains increases the frequency of mtDNA loss and reduces their respiratory capacity, suggesting coadaptation of mitochondrial and nuclear genomes at the intraspecific level. We also show that the ability of selfish mtDNA to displace other variants in heteroplasmic yeast cells depends primarily on their mtDNA sequences rather than the nuclear genome background. These findings underscore the complex interactions between mitochondrial and nuclear genomes.

Abstract

Unlike most animals, some fungi, including baker’s yeast, inherit mitochondrial DNA (mtDNA) from both parents. When haploid yeast cells fuse, they form a heteroplasmic zygote, whose offspring retain one or the other variant of mtDNA. Meanwhile, some mutant mtDNA (rho⁻), with large deletions in the nucleotide sequence, can displace wild-type (rho⁺) mtDNA. Consequently, offspring of zygotes with such rho⁻ mtDNA predominantly carry the mutant variant. This phenomenon is called suppressivity. In this study, we investigated how the suppressivity of rho⁻ mtDNA depends on the mitochondrial and nuclear genomes of the rho⁺ strain during crossing. Comparing two diverged laboratory strains, SK1 and W303, we measured suppressivity in crosses with four rho⁻ strains. One rho⁻ strain showed significantly higher suppressivity when crossed with SK1 than with W303. We then created cytoplasmic hybrids by swapping mtDNAs between these strains. Surprisingly, we found that the mtDNA of the rho⁺ strain, rather than its nuclear DNA, determines high suppressivity in crosses of SK1 rho⁺ with the rho⁻ strain. Additionally, mtDNA replacement reduced respiration rate and growth rate on non-fermentable substrates while increasing the likelihood of functional mtDNA loss. Our data demonstrate that a mutant mtDNA variant’s ability to displace another mitochondrial DNA variant in a heteroplasmic cell depends more on mtDNA sequences than on the biochemical and structural context created by the nuclear genome background.

1. Introduction

Mitochondria of most eukaryotic species harbor their own DNA, with very few exceptions [1]. Meanwhile, the life cycle of many fungi involves the fusion of cytoplasms, which predates the fusion of nuclei [2]. As a result, a single cell can harbor mitochondria with diverged mitochondrial DNAs (mtDNAs); this condition is usually referred to as mitochondrial heteroplasmy. As a result of genetic drift, the descendants of heteroplasmic cells eventually retain only one of the mtDNA variants. In baker’s yeast, such segregation takes only a few cell divisions, with the exact number depending on parameters such as the structure of the mitochondrial cristae and the nature of substitution in mtDNA [3,4].

Laboratory yeast strains have high frequencies of mutations in mtDNA, which results in a high proportion, up to several percent, of petite yeast cells in suspensions of yeast cells [5]. Petite cells are incapable of oxidative phosphorylation and cannot utilize non-fermentable carbon sources, such as glycerol or ethanol [6]. The mtDNA of such cells usually harbors large deletions and is devoid of mitochondrially encoded protein-coding genes, tRNAs, and rRNAs (referred to as rho⁻ mtDNA) [7]. Despite this, such mutant mtDNAs can have a fitness advantage at the intracellular level over wild-type rho⁺ mtDNAs [8,9]. Therefore, during mating, after fusion of the yeast gametes, such rho⁻ mtDNAs can displace wild-type rho⁺ mtDNAs. As a result, the crossing of some rho⁻ haploid strains with a rho⁺ strain of opposing mating type could produce up to 100% rho⁻ diploids [10,11]. This phenomenon is called suppressivity and is usually defined as the proportion of rho⁻ diploids among all diploids produced in rho⁻ × rho⁺ strain crossings.

The mechanisms of suppressivity are still unclear. On the one hand, rho⁻ mutant mtDNAs with large deletions still contain highly active origins of replication and, therefore, their replication takes less time than the replication of full-length rho⁺ mitochondrial DNA [12], although the increased rate of rho⁻ mtDNA replication is observed only in some rho⁻ strains [10]. Similarly, the deficiency of deoxynucleotides, which hampers mitochondrial DNA replication, can also provide a replication rate advantage to the rho⁻ mtDNAs [13].

On the other hand, it has been shown that several mitochondrial enzymes are required for the suppression of wild-type mtDNA by rho⁻ mtDNA. For example, mitochondrial RNA polymerase Rpo41p unexpectedly impaired the maintenance of rho⁺ mtDNAs in heteroplasmic zygote cells, while the reduction in its activity decreased the suppressiveness of rho⁻ strains [14]. The activity of Rpo41 is regulated by mitochondrial RNA endonuclease Pet127p; however, the regulation mechanism is not associated with the RNase activity of Pet127p but rather with the physical interaction between Rpo41p and Pet127p [14]. Moreover, genetic screening revealed that the deletion of the CCE1 gene (alias MGT1) abolished the suppressiveness of highly suppressive rho⁻ strains [15]. CCE1 encodes the cruciform cutting endonuclease which is required for mitochondrial recombination [16]. In contrast to mammalian mtDNA [17], yeast S. cerevisiae mtDNA experiences frequent recombinations [18]. Moreover, rho⁻ mtDNAs can recombine with rho⁺ mtDNAs. This observation formed the basis of another hypothesis about the mechanisms of suppressivity. According to this suggestion, mutant rho⁻ mtDNA can enter into destructive recombination with rho⁺ mtDNA [19]. Considering the high copy number of rho⁻ mtDNA in the cells, this destructive recombination process removes most rho⁺ mtDNA molecules from the heteroplasmic cells, resulting in a transmission bias favoring rho⁻ mtDNAs.

Although there are multiple possible mechanisms that confer suppressivity of rho⁻ mtDNAs, they can be classified into two distinct categories. First, some mechanisms should depend on mtDNA genotypes. For example, if the main contribution to suppressivity is made by the relative activity of the origins of replication of rho⁻ and rho⁺ molecules, or if the consequences of destructive recombination are determined by the homology of mtDNA sequences, then suppressivity will depend on the mtDNA genotype. Second, suppressivity may depend on the presence of certain alleles in nuclear DNA (nDNA), for example, variants of CCE1, RPO41, PET127, or other yet unidentified nuclear-encoded genes regulating suppressivity.

In the study presented herein, we tested the contribution of these mechanisms by measuring the suppressivity of several previously characterized rho⁻ strains by crossing them with two diverged standard laboratory strains, SK1 and W303-1A, available in our laboratory. We found that one of the rho⁻ strains exhibited significantly different suppressivity while crossing with SK1 rho⁺ and W303-1A rho⁺ strains. Thus, we produced cytoplasmic hybrids by swapping the mtDNA of the SK1 and W303 strains and found that the differences in their mtDNA, rather than their nDNA, explain this variation.

2. Materials and Methods

2.1. Yeast Strains and Growth Conditions

This study utilized Saccharomyces cerevisiae strains W303-1A, SK1, and NAB69, along with their derivative mutants (Table S1). Yeast strains were cultured in standard rich and synthetic media as described by Sherman [20]. The rich medium components included peptone (DiaM (Moscow, Russia), Cat. No. HYP-A.5000) and yeast extract (BioSpringer (Cedar Rapids, IA, USA), Cat. No. 0207/0-PW-L.5000). Agar (DiaM (Moscow, Russia), Cat. No. 1923.5000) was used for solid media preparation. Carbon sources included D-glucose (Helicon (Moscow, Russia), Cat. No. H-0401-0.5), raffinose (Chimmed (Moscow, Russia), Cat. No 2012), and glycerol (Helicon (Moscow, Russia), Cat. No. I-195204-1.0). Carbon sources were autoclaved separately from other components. The medium’s pH was approximately 5.5. Synthetic media were prepared using Yeast Nitrogen Base without amino acids (Sigma (San Diego, CA, USA), Cat. No. Y0626-1KG) and yeast synthetic drop-out medium supplements (Sigma (San Diego, CA, USA). When required, G418 sulfate (PhytoTechnology Laboratories (Lenexa, KS, USA), Cat. No. 108321-42-2) was added for selection.

2.2. Comparison of SK1 and W303 Genomes

To distinguish the nuclear and mitochondrial genomes of SK1 and W303 (see Table S3) strains by PCR and following agarose gel electrophoresis, we found loci providing products of different sizes with the same primers. For that purpose, we made local similarity graphs (Figure 1A and Figure S1) using the “blastn” program [21] available at the NCBI. This showed us rearranged loci, and we designed the primers to produce significantly different products (see Figure 2).

To count SNPs in mitochondrial genomes, we performed global alignment with PairwiseAligner (scoring = ‘blastn’) from biopython [https://biopython.org/, Current Release—1.84—28 June 2024], removed all columns with gaps (Figure S2), and calculated SNPs in 1000 b.p. windows without overlaps.

2.3. Yeast Cybrid Strain Construction and Confirmation

To construct the cybrids, we utilized two rounds of crossing with the rho⁰ strains (see the scheme in Figure 2A). The NAB69 Δkar1-1 rho⁰ strain was then crossed with W303 KanMX6 rho⁺ and SK1 KanMX6 rho⁺ strains. We preincubated the cells in liquid YPD medium at 30 °C for at least 4 h. Crosses were performed by mixing the cells to an OD₅₅₀ of 0.1 in 200 μL of YPD and incubating them for 20 h at room temperature. The resulting suspension was diluted 100-fold and plated on solid YNBGly -Leu dropout medium. The plates were incubated for 48–72 h at 30 °C. Colonies were subsequently transferred to both YNBGly-Leu and YNBGly-Leu G418 plates to select for haploid cybrids containing the NAB69 Δkar1-1 nucleus and mitochondrial DNA from either W303 or SK1. To validate the resulting cybrids, they were crossed with SK1 TRP rho⁰ and W303 TRP rho⁰ strains.

The cybrids were confirmed by polymerase chain reactions (PCRs) with the primers designed to anneal unique regions of either SK1 or W303 mitochondrial and nuclear DNA (Tables S2 and S3). The PCRs were carried out using Phusion Hot Start II high-fidelity DNA polymerase. The PCR mixtures contained Phusion HF buffer, 200 μM dNTPs, the corresponding primers (at a final concentration of 0.5 μM), and yeast cells as the DNA template.

2.4. Suppressivity Test (rho⁻ mtDNA)

The cells were incubated in liquid YPD medium at 30 °C for at least 4 h. MAT a rho⁺ strains were then crossed with either MAT alpha rho⁻ strains or control MAT alpha rho⁺ and rho⁰ strains. Diploid selection was achieved using prototrophic selective markers (URA3 and TRP1). For crossing, the cells were mixed to an OD₅₅₀ of 0.1 in 200 μL of YPD and incubated for 20 h at room temperature. The resulting suspensions were diluted 100–1000-fold in sterile water and plated on synthetic dropout DGly (0.1% d-glucose, 2% glycerol) medium without uracil and tryptophan (-Ura, -Trp). The plates were incubated at 30 °C for 48–72 h, after which colonies were counted. Suppressivity was calculated as (number of petite CFUs/total CFUs) × 100% on YPDGly.

2.5. Growth Kinetics

We incubated the strains in liquid YPGly medium at 30 °C for at least 4 h. Exponentially growing cells were then diluted to an OD₅₅₀ of 0.05 and inoculated into a 96-well plate (Eppendorf (Hamburg, Germany). The plates were incubated at 30 °C in a SpectrostarNANO (BMG Labtech (Ortenberg, Germany) spectrophotometer with orbital shaking at 500 rpm for 2 min prior to each measurement. Optical density was measured at 5 min intervals over a period of 20–27 h. The maximal growth rate (μ_max), which is proportional to the inverse of duplication time, was calculated using linear model fits of the natural log-transformed OD₅₅₀ values. This analysis was performed using an R script as described in [22]. Along with μ_max, the script provides R² values that reflect the quality of the linear fit. We noticed that SK1 strain cells could aggregate and, therefore, in some experiments, the data were noisy, thus providing lower R² values.

2.6. Respirometry

We assessed the respiration of cells with a Clark-type oxygen electrode (Strathkelvin Instruments 782, Scotland, UK) at 25 °C. The incubation medium was 50 mM KH₂PO₄ and 0.05% glucose, pH = 5.5.

2.7. Statistics, Data Visualization, and Analysis

All experiments were performed with at least five independent replicates. The figures were made using the tidyverse packages of the R language [23]. The figures show boxplots where the lower and upper borders correspond to the 25th and 75th percentiles (InterQuartile Range, IQR). The upper whisker extends from the box border to the largest value within 1.5 × IQR. The lower whisker reaches the smallest value within 1.5 × IQR. The results of individual data points are shown as circles; points beyond whiskers are outliers. All analyses were performed using the base R programming language library. The significance difference was determined using the Wilcoxon rank-sum exact test. We plotted only the significance of the comparison with p-values below 0.05.

3. Results

To assess whether the mitochondrial or nuclear genome of the rho⁺ strains determines the suppressivity of the rho⁻ strains, we selected two laboratory strains, SK1 and W303, which harbored the required set of genetic markers for testing suppressivity. Importantly, the mitochondrial genomes of these strains exhibited pronounced differences in architecture and accumulated numerous single nucleotide polymorphisms (SNPs) (Figure 1A,B). Figure 1A displays a local similarity plot highlighting deleted and duplicated mtDNA regions in the W303 mitogenome compared to that of SK1. Figure 1B shows the number of SNPs in a 1000 bp sliding window along the genome alignments. These results demonstrate that the mitochondrial genomes of SK1 and W303 have diverged sufficiently to be distinguishable, while potentially inhibiting recombination between them. Meanwhile, the nuclear genomes of SK1 and W303 strains also show some differences in architecture, including duplications and rearrangements (see, for example, the local similarity graph of chromosome I in Figure S1). The genes CCE1, RPO41, and PET127, which affect mitochondrial DNA suppressivity, differ by only 2 to 4 missense non-radical substitutions in the variable regions of their orthologs in other Ascomycota species (Supplementary Text S1).

We then produced two cytoplasmic hybrids (cybrids) using a kar1-1 strain as an intermediate host for the mtDNA. The kar1-1 mutation inhibits nuclear fusion, often resulting in dikaryotic cells that enable the transduction of rho⁺ mtDNA into a new nuclear background (Figure 2A). Using this approach, we swapped mtDNA between W303 and SK1 strains, generating a W303 cybrid strain with SK1 mtDNA (W303mtSK1) and an SK1 cybrid strain with W303 mtDNA (SK1mtW303). To confirm the presence of the correct nuclear and mitochondrial DNAs in these cybrids, we performed colony PCR using strain-specific primers (Table S2). The primers used to verify mitochondrial replacement in the cybrids yielded PCR products of varying sizes for the SK1 and W303 genomes, as expected (Figure 2B).

To characterize the phenotypes of the cybrids, we measured the specific maximal growth rates of the cybrids and parental strains. We tested their growth in rich media with different carbon sources: fermentable (YPD), poorly fermentable (YPRaf), and non-fermentable (YPGly). While no difference was detected in YPD, the W303mtSK1 cybrid exhibited slower growth in YPRaf and YPGly (Figure 3). Given that mitochondrial DNA encodes respiratory chain components, we also measured the respiration rate of the cybrids and their parental strains. Figure 4 shows that the oxygen consumption rate of SK1mtW303 cybrid cells was lower than that of the parental SK1 strain. This effect was pronounced in both tested conditions: with and without the addition of the protonophore Carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP). FCCP uncouples respiration and phosphorylation, thus enabling maximal activity of the respiratory chain. The W303mtSK1 cybrid exhibited a slower decrease in respiration rate compared to the parental W303 strain, although this difference was not statistically significant.

The premise of this study was the hypothesis that the suppressivity of rho⁻ strains can differ depending on the rho⁺ strain with which it is crossed. To test this, we selected four previously characterized MAT alpha rho⁻ strains with varying suppressivity [9] and crossed them with two MAT a strains, SK1 and W303. Our results showed that for highly suppressive strains (HS rho⁻, rho⁻ 5, and rho⁻ 10), substitution of the rho⁺ strain did not significantly affect suppressivity (Figure 5A). However, the low-suppressive strain rho⁻ 2 produced significantly more petite colonies when crossed with the SK1 rho⁺ strain compared to the W303 rho⁺ strain (Figure 5A). Notably, control crossings of the SK1 strain with rho⁰ and rho⁺ strains produced no petite colonies. Interestingly, the rho⁻ 2 strain exhibited lower suppressivity when crossed with SK1mtW303 than when crossed with the W303mtSK1 cybrid (Figure 5B). This result indicates that the mitochondrial DNA, rather than the nuclear DNA background of the rho⁺ strain, explains the variation in suppressivity values.

We hypothesized that the decreased resistance of the W303mtSK1 strain to rho⁻ 2 mtDNA is due to incompatibility between the W303 nuclear background and SK1 mtDNA. To test this, we measured the percentage of spontaneous petite cells in the W303mtSK1 suspension grown in YPD and YPGly media. In YPD medium, emerging petite cells can continue to divide, albeit at a slower rate than the parental rho⁺ strain. In contrast, YPGly medium contains only a non-fermentable carbon source, allowing only rho⁺ cells to proliferate. As expected, we found that all tested strains contained more petite cells when grown on YPD compared to YPGly medium (Figure 6). Importantly, on both tested media, the W303mtSK1 strain produced significantly more petite cells than the parental W303 strain. This observation suggests that the mitochondrial DNA of the SK1 strain cannot be stably maintained in the W303 background. Meanwhile, the complementary SK1mtW303 strain did not show mitochondrial genome instability.

4. Discussion

mtDNA encodes core subunits of the respiratory chain complexes and ATPase, while other subunits of these complexes are encoded in nuclear DNA. This arrangement imposes restrictions on the evolution of genes encoding these complexes, as mutations in one genome may become incompatible with the other [24,25,26]. Consequently, replacing mtDNA is only possible between closely related species. For instance, replacing human mtDNA with mtDNA from chimpanzees and gorillas produces cells capable of oxidative phosphorylation (OXPHOS). However, cytoplasmic hybrids of humans and orangutans are not viable [27]. Interestingly, the speciation of some organisms, including yeast, is driven by mitochondria–nucleus incompatibility [28,29].

Yeast cells have proven to be a valuable model for studying mito-nuclear interactions due to their ability to proliferate without mitochondrial DNA, which expands the available methods for manipulating their mitochondrial genomes. In this study, we swapped mtDNA and nDNA between two standard laboratory haploid strains: SK1 and W303. The W303 strain is one of the most commonly used laboratory strains [30]. Notably, the rho⁻ strains used in this study were derived from the isogenic W303 MAT alpha strain [9,11], and are therefore expected to have almost identical nuclear genomes with only minor MAT locus rearrangements in the nDNA. The SK1 strain is frequently employed in meiosis studies due to its high sporulation efficiency [31]. The genomes of these strains are pronouncedly diverged compared to other standard laboratory strains [32,33].

As expected, non-native mtDNA decreased the performance of yeast strains under respiratory conditions. We found that W303mtSK1 cybrids exhibited a decreased growth rate on non-fermentable and poorly fermentable carbon sources (Figure 3). Since oxidative phosphorylation is required for efficient utilization of these carbon sources, we suggest that even mild mito-nuclear incompatibility between SK1 and W303 mtDNA led to a decrease in fitness. However, we did not observe a difference in growth rate between parental W303 and W303mtSK1 strains in the YPD medium (Figure 3). This can be explained by the fact that under glycolytic conditions, mitochondrial functions are suppressed [34] and enzyme complexes containing proteins encoded by mtDNA do not play a significant role.

Meanwhile, we found that the SK1mtW303 cybrid exhibited a lower respiration rate compared to the parental SK1 strain (Figure 4). The W303mtSK1 strain also showed a decrease in respiration rate compared to its parental strain, although this difference was not statistically significant. It remains unclear why the decrease in respiration rate of the SK1mtW303 strain did not manifest as a growth defect when using respiratory substrates. We hypothesize that the SK1 strain’s growth rate may be additionally limited by other factors, such as its ability to flocculate, which could mask growth defects in the YPGly medium

The variation in individual yeast strain phenotypes is primarily explained by nDNA sequence and mito-nuclear interactions, while mtDNA sequence itself poorly explains phenotypic variation [35,36]. Our study demonstrates that mtDNA sequence can, in some cases, have a major effect on phenotype under conditions of mitochondrial heteroplasmy. We showed that mitochondrial DNA, rather than nuclear DNA, can determine rho⁺ mtDNA susceptibility to suppression by rho⁻ mtDNA (Figure 5). Specifically, the low-suppressive mtDNA, rho⁻ 2, displaced non-related SK1 mtDNA in the suppressivity test more efficiently than parental W303 mtDNA (Figure 5). Interestingly, we did not observe any difference in suppressivity for the other tested rho⁻ strains depending on the rho⁺ strain with which they were crossed. This may be explained by different mechanisms of suppressivity in highly and low-suppressive rho⁻ mtDNAs. Alternatively, in the case of the highly suppressive rho⁻ strain, the sensitivity of the method might not be sufficient to quantify the difference. Indeed, in the HS rho⁻ crosses, less than one percent of Grande colonies that were able to retain wild-type mtDNA formed (Figure 5).

5. Conclusions

To summarize, our study demonstrates that swapping mtDNA between laboratory yeast strains increases the rate of mtDNA loss while decreasing respiration and growth rates. Although expected, this observation suggests that mutual adaptation of mitochondrial and nuclear genomes occurs at the intraspecific level in yeast. At the same time, our data show that the mutant mtDNA variant’s ability to outcompete another variant in a heteroplasmic cell can be primarily determined by mtDNA sequences rather than the nuclear genome background.

In the state of mitochondrial heteroplasmy, over the course of generations, only one variant of mitochondrial DNA usually remains in the cells. This outcome is determined by both random genetic drift and the relative fitness levels of the different mtDNA variants [37]. Some variants have been shown to reproducibly displace others. The relative fitness of different variants depends on the context (e.g., tissue type of the multicellular organism), which is created in the mitochondrial matrix by proteins encoded in the nuclear genome [38,39]. However, the results of suppressivity tests in yeast SK1-W303 cybrids demonstrate that, in some cases, the relative competition of mtDNA in yeast is determined by differences in the mtDNA itself. These results demonstrate that different mtDNAs from standard laboratory strains can be subject to varying degrees of displacement by selfish mitochondrial DNA elements. We propose that mtDNA sequences evolved under constant pressure from such selfish elements and may contain adaptations to counteract them.