Introduction

Mitochondria are critical to intracellular bioenergetic balance [1] and their genome, the mtDNA, encodes genes required to maintain the function of the electron transport chain (ETC). MtDNA maintenance defects are a hallmark of bioenergetic dysfunction in several diseases and aged tissue [2, 3]. In humans and mice, mtDNA is packaged into 70–110 nm ellipsoid DNA-protein complexes named nucleoids [4, 5] through interaction with its main protein component, the Mitochondrial Transcription Factor A (TFAM) [6, 7].

The maintenance and distribution of mtDNA are critical for cellular physiology [8,9,10,11,12,13] and are impaired during disease [14,15,16,17,18,19]. Mitochondria undergo constant fusion/fission cycles, which allow the exchange of its components and supports mtDNA maintenance [1, 8, 20]. In mammalian cells, nucleoids are semi-regularly distributed inside the mitochondrial population [5, 21, 22] and, in yeast, it was proven that regular distribution is necessary to accurately control nucleoid segregation to daughter cells [23]. Furthermore, nucleoid positioning on the sites of mitochondrial fission can ensure mtDNA inheritance to daughter mitochondria [24]. Regardless of the relevance of mtDNA abundance and maintenance, little is known about the physio-pathological relevance of nucleoid distribution in the mammalian mitochondrial population, partly because of the challenge of studying nanoscopic structures [25].

Nucleoids physically interact with the inner mitochondrial membrane (IMM), and are interspersed between the IMM cristae [4, 26, 27]. Cristae structure is tightly regulated by multiple proteins, such as the ATP synthase [28], the Mitochondrial Contact Site and Cristae Organization System (MICOS) [29, 30], and OPA1 [31,32,33]. OPA1 is a GTPase also involved in various processes, including mitochondrial fusion and fission [34,35,36]. Furthermore, the lack of Mgm1, the OPA1 yeast ortholog, leads to mtDNA depletion [37, 38] and Opa1−/− cells show a decrease in mtDNA and nucleoid number [39,40,41], and display mitochondria without nucleoids [39]. Finally, OPA1exon4b isoforms 3, 5, 6, and 8 physically interact with mtDNA to regulate its maintenance [39, 42]. Yet, the role of OPA1 in intramitochondrial nucleoid distribution in mammalian cells under pathophysiological conditions is underexplored.

OPA1 mutations lead to Autosomal Dominant Optic Atrophy (ADOA, MIM 165500), a disease that causes progressive loss of retinal ganglion cells resulting in optic nerve degeneration and blindness [43]. About 20% of ADOA patients manifest a more severe syndromic form of the disease, classified as ADOA plus (ADOA + , MIM: 125250) [44, 45]. Patients bearing mutations at the GTPase domain of OPA1 are more likely to develop ADOA+ [45]. Skeletal muscle samples from ADOA patients display defects in mtDNA integrity and distribution with bioenergetic dysfunction markers, as well as IMM ultrastructure alterations [18, 46, 47]. Accordingly, our previous study showed that OPA1 ADOA-causing mutants in the GTPase or GTPase effector domain (GED) display distinct effects on IMM ultrastructure and mitochondrial fusion and fission dynamics [48]. Nevertheless, it is unclear if these changes have an impact on mtDNA distribution and maintenance.

In this work we studied the effect of OPA1 and different ADOA-causing mutants on nucleoid distribution. We used Opa1−/− MEF and overexpression of OPA1 mutants on WT MEF along with patient-derived cells to study nucleoid distribution and nucleoid-cristae proximity. Our data showed that the lack of OPA1 and the overexpression of OPA1 or ADOA-causing mutants perturbed nucleoid cluster distribution within the mitochondrial population and disrupted cristae morphology. Furthermore, nucleoid-cristae proximity is perturbed in Opa1−/− cells which might contribute to local OXPHOS dysfunction. This work introduces new clues on the influence of OPA1 over mitochondrial nucleoid distribution and provides novel insights on the etiology of the mtDNA alterations as well as the local bioenergetic dysfunction found in ADOA patients.

Results

Opa1 determines nucleoid abundance and distribution

We studied the role of OPA1 in nucleoid abundance using Opa1−/− MEF and rescue through expression of human OPA1 isoform 1. First, we quantified the mtDNA copy number and found a ~ 70% decrease in mtDNA levels in Opa1−/− cells, which was not explained by changes in mitochondrial biogenesis transcriptional regulator PCG1-α [1] (Fig. S1, Supplementary material), or mitochondrial protein mass decrease, as we previously reported [48]. The mtDNA levels were partially rescued upon acute expression of OPA1, and completely rescued in our Opa1−/− + OPA1 stable cell line ( + p-lenti OPA1 WT) (Fig. 1A). To accurately evaluate nucleoid distribution, we propose a novel model where we measure: (1) the nucleoids array in the mitochondrial population, which describes the density and spread of nucleoids within the mitochondrial network and includes the mitochondria with and without nucleoids and, (2) intramitochondrial longitudinal distribution, which describes nucleoid localization in relation to the length of single organelles (Fig. 1B). To apply our model, we developed a manual approach and a novel unbiased semi-automated method called MiNuD (Mitochondrial Nucleoids Distribution), as described in Methods (Fig. S2).

Fig. 1: Opa1 determines nucleoid abundance and array in the mitochondrial population and intramitochondrial longitudinal distribution.
figure 1

A Mitochondrial DNA abundance in WT MEF, Opa1−/− and Opa1−/− cells stably expressing a lentiviral plasmid carrying OPA1 cDNA, quantified by qPCR and expressed as mt-Nd4 to Gapdh ratio. Data are mean ± SEM of ≥ 3 independent experiments. B Schematic representation of the components of nucleoid distribution. C Representative images of WT MEF, Opa1−/− and Opa1−/− cells with WT OPA1 acute expression, co-transfected with mCherry-Omp25 (OMM) and Tfam-GFP (nucleoids) cDNA. Bottom panel, insets of Tfam-GFP (top) and merge (bottom). D Percentage of mitochondria with different number of nucleoid clusters. E Mean nucleoid cluster number per mitochondrion. F Mean nucleoid cluster number per mitochondrion area, manually quantified. G Number of nucleoid clusters per total mitochondrial area, semi-automatically quantified by the MiNuD algorithm. For Opa1−/− cells, DNA was labeled using PicoGreen, and mitochondria were labeled with mCherry-Omp25 (OMM) or MitoTracker Deep Red. H Mean nucleoid cluster number per mitochondrion length. I Simple regression of the relationship between nucleoid cluster number and mitochondrion length in individual organelles. Data are mean ± SEM of ≥ 281 objects from ≥ 15 cells of ≥ 3 independent experiments, and for G data are mean ± SEM of ≥ 10 cells of ≥ 2 independent experiments (****p < 0.0001; ***p < 0.0005; *p < 0.05).

Full size image

We applied our semi-automatic approach to quantify nucleoid distribution in WT MEF using two strategies to label nucleoids: (i) direct dsDNA labeling with PicoGreen or (ii) Tfam-GFP [49]. We performed live-cell imaging at a high-resolution confocal microscope, with sufficient resolution to identify nucleoid clusters, rather than individual objects [50]. Both techniques showed “point-like” signals with apparent uniform distribution throughout the mitochondrial network (Fig. S3A), and comparable number of nucleoid clusters per total mitochondrial area (Fig. S3B). Considering that PicoGreen displayed a nuclear stain (Fig. S3A), we continued our study with Tfam-GFP, excluding cells that show signs of major overexpression such as nuclear Tfam-GFP localization [49].

We evaluated mitochondrial nucleoid abundance and distribution using the manual quantification approach described in Methods (Fig. S2). We studied the number of nucleoids per mitochondria in maximum projections ( ~ 1 µm thickness) and found that in WT MEF, 13% of mitochondria lack apparent Tfam-GFP foci (Fig. 1C, D), thus we classified them as “empty”. We corroborated this result with anti-dsDNA and Picogreen staining (Fig. S3C).

Opa1−/− cells displayed an increase in “empty” organelles (52% vs 13% WT) and a diminished number of nucleoid clusters per mitochondria, consistent with the loss in mtDNA levels (Fig. 1C, D). Acute rescue of OPA1 partially recovered the number of “empty” mitochondria (24%) and the number of nucleoid clusters per mitochondria (Fig. 1C–E). In addition, we measured the number of nucleoid clusters per mitochondrial area using our manual approach and MiNuD, and found the same tendency (Fig. 1F–G). Notably, in Opa1−/− cells, Tfam-GFP showed a combination of bright points and diffuse signals (Fig. 1C). The acute rescue of OPA1 re-established the nucleoid cluster phenotype and labeling. Alternative nucleoid labeling with PicoGreen or anti-dsDNA confirmed the low number of nucleoid clusters observed in the Opa1−/− cells (Fig. S4A–C).

Considering the high number of mitochondria lacking apparent Tfam-GFP foci and the changes in morphology shown in Opa1−/− cells, we evaluated intramitochondrial longitudinal distribution in organelles that: (1) showed Tfam-GFP foci; (2) were shorter than 3 µm, the longest organelles found in Opa1−/− cells (Fig. S5). Interestingly, the average intramitochondrial distribution of one nucleoid cluster per micrometer remained unchanged in Opa1−/− cells (Fig. 1H). To further examine this, we applied a simple linear regression to the number of nucleoid clusters found in each mitochondrion and its length. WT MEF showed a correlation between mitochondria length and nucleoid cluster number that is reduced in Opa1−/− cells and rescued after acute OPA1 expression (Fig. 1I).

Thus, OPA1 acute rescue in Opa1−/− cells partially recovered mtDNA abundance and the nucleoid cluster array in the overall mitochondrial population. Notably, the absence of OPA1 perturbed the intramitochondrial longitudinal distribution of the nucleoid clusters.

OPA1 rescues mitochondrial cristae periodicity and nucleoid-cristae proximity in Opa1 −/− cells

As OPA1 is key for IMM fusion and cristae organization [32, 40] and the cristae morphology is associated with nucleoid distribution and abundance [10, 26, 41], we performed transmission electron microscopy (TEM) (Fig. 2A). Based on the diversity of cristae phenotypes, we classified each mitochondrion according to their cristae shape and periodicity (Fig. 2B–D). Mitochondria with no discernible cristae were classified as organelles with “no cristae”.

Fig. 2: The stable expression of WT OPA1 rescues mitochondrial cristae periodicity and nucleoid-cristae proximity in Opa1−/− cells.
figure 2

A Representative TEM images of WT MEF, Opa1−/− and Opa1−/− cells stably expressing a lentiviral plasmid carrying OPA1 cDNA. B Mitochondrial cristae shape classification and C their quantification. D Mitochondrial cristae periodicity classification and E their quantification. Data are from ≥ 98 objects from ≥ 2 independent experiments. F Representative immunofluorescence images of WT MEF, Opa1−/− and Opa1−/− cells stably expressing a lentiviral plasmid carrying OPA1 cDNA. Cells were labeled with MitoTracker Deep Red and antibodies against dsDNA and Mic60. The white boxes correspond to 5 x 5 μm insets. (Right) Intensity profile of dsDNA and Mic60 signals along the yellow line from left to right (2 µm). G, H Colocalization coefficients for dsDNA and Mic60 in the cells shown in F. Each point represents one cell. Data are mean ± SEM from ≥ 70 cells from 4 independent experiments (****p < 0.0001; *p < 0.05). In A, B and D the scale bar is 1 μm.

Full size image

In Opa1−/− cells, we observed an increase in mitochondria with short (29% vs. 16% WT) and aberrant cristae (37% vs. 12% WT) (Fig. 2A–C). The aberrant cristae in Opa1−/− cells consisted mainly of onion-like and arch phenotypes (Fig. S4D). Importantly, the stable expression of OPA1 rescued the abundance of long cristae (65% vs. 34% Opa1−/−) (Fig. 2A–C). Similarly, cristae periodicity was disturbed in Opa1−/− cells. Mitochondria with no cristae were increased (12% vs. 6% WT) as well as non-periodic sparse cristae (78% vs. 12% WT) (Fig. 2E). The stable expression of OPA1 partially rescued the abundance of periodic cristae (47% vs. 7% Opa1−/−) (Fig. 2E).

Considering the major changes in IMM ultrastructure in Opa1−/− cells, we next looked at the nucleoid-cristae proximity, by immunofluorescence labeling of dsDNA and MICOS component Mic60. In WT MEF, the nucleoids are abundant and interspersed between the cristae as expected, displaying close mutual proximity (Fig. 2F). However, in Opa1−/− cells, there is an apparent disorganization of the nucleoid-cristae proximity. These data were further evaluated by analyzing the colocalization of mtDNA to Mic60 and vice versa. We found that mtDNA colocalization to Mic60 is decreased in Opa1−/− cells and is mildly rescued after OPA1 stable expression, hence, nucleoid-cristae proximity is perturbed in mitochondria that contain nucleoids (Fig. 2F–G). In contrast, Mic60 colocalization to mtDNA is rescued in Opa1−/− cells after OPA1 expression, which relates to the mtDNA content and cristae periodicity recovery (Fig. 2F, H).

Thus, in Opa1−/− cells, mitochondria predominantly display sparse cristae with decreased nucleoid-cristae proximity, that are partly rescued by OPA1 stable expression.

Acute expression of WT and pathogenic OPA1 variants perturb mitochondrial nucleoid abundance and distribution

We next inspected the effect of OPA1 ADOA-causing mutants on nucleoid distribution and abundance to further understand what causes the mtDNA defects found in ADOA patients [18, 19, 46, 47]. We used our previously characterized model of overexpression of OPA1 in WT MEF and expression of two OPA1 ADOA-causing mutants [48, 51]: OPA1 c.870+5 G > A (p. Lys262_Arg290Del), located in the GTPase domain; and c.2713 C > T (p.Arg905X), located in the GTPase-effector domain (GED) (Fig. 3A).

Fig. 3: Acute expression of WT and pathogenic OPA1 variants perturb mitochondrial nucleoid cluster abundance and distribution.
figure 3

A Schematic representation of OPA1 isoform 1 domains and ADOA-causing mutants: OPA1 c.870+5 G > A (GTPase domain) and OPA1 c.2713 C > T (GTPase effector domain, GED). MTS: mitochondrial targeting sequence. TM: transmembrane domain. S1: Processing site 1. CC1: coiled-coil domain 1. B Mitochondrial DNA abundance of WT MEF stably expressing a lentiviral plasmid carrying control cDNA (CTL), OPA1 WT, or pathogenic OPA1 variants; quantified by qPCR and expressed as mt-Nd4 to Gadph ratio. Data are mean ± SEM from ≥ 5 independent experiments. C Representative images of WT MEF, WT MEF overexpressing OPA1 WT, or pathogenic OPA1 variants, co-transfected with mCherry-Omp25 (OMM) and Tfam-GFP (nucleoids) cDNA. Bottom panel, insets of Tfam-GFP (top), and merge (bottom). D Percentage of mitochondria with a different number of nucleoid clusters. E Mean nucleoid cluster number per mitochondrion. F Mean nucleoid cluster number per mitochondrion area, manually quantified. G Mean nucleoid cluster number per total mitochondrial area, quantified semi-automatically by the MiNuD algorithm. H Mean nucleoid cluster number per mitochondrion length. I A simple regression of the relation between mitochondrial length and the number of nucleoids in single organelles. Data are mean ± SEM of ≥ 228 objects from ≥ 18 cells of ≥ 4 independent experiments. For G, data are mean ± SEM of ≥ 11 cells of ≥ 4 independent experiments (****p < 0.0001; **p < 0.005; *p < 0.05).

Full size image

We first analyzed mtDNA maintenance. Interestingly, acute, and stable overexpression of OPA1 showed no differences in mtDNA copy number (Fig. S6A, Fig. 3B). Likewise, mtDNA abundance was unchanged after stable overexpression of OPA1 mutants (Fig. 3B) or TFAM overexpression (Fig. S6B). Moreover, the stable expression of OPA1 mutants did not cause mtDNA deletions (Fig. S6C–D) or changes in PCG1-α levels (Fig. S1).

The nucleoid array showed that acute OPA1 overexpression caused a slight reduction in empty mitochondria (10% vs. 13% WT MEF) and an increase in the number of nucleoid clusters per mitochondrion and per mitochondrion area (Fig. 3C–G). Conversely, cells with OPA1 c.870+5 G > A expression showed an increase in “empty” mitochondria (17% vs. 10% WT OPA1) and a lower abundance of nucleoids clusters per mitochondrion area (Fig. 3C–G). Furthermore, cells expressing the mutant OPA1 c.2713 C > T showed an increase in “empty” mitochondria (22% vs. 10% WT OPA1) and a reduction in the number of nucleoid clusters per mitochondrion and area (Fig. 3C–G). These changes were independent of the mitochondrial morphology since WT OPA1 and c.870+5 G > A mutant overexpression showed elongated mitochondria, and cells expressing OPA1 c.2713 C > T exhibited fragmented mitochondria (Fig. S7A–D), as we previously described in this model [48]. Intramitochondrial longitudinal distribution of nucleoids was quantified in mitochondria up to 6 µm, the longest organelles found in OPA1 c.2713 C > T (Fig. S7C). Interestingly, we found no changes in this parameter upon the expression of WT OPA1 or the studied mutants (Fig. 3H–I).

Thus, in our model neither OPA1 overexpression nor the expression of ADOA-causing mutants induced changes in mtDNA abundance and integrity. Yet, OPA1 overexpression caused mitochondrial elongation and an increased nucleoid cluster abundance in the mitochondrial population. Surprisingly, cells expressing OPA1 ADOA-causing mutants showed alterations in mitochondrial morphology and nucleoid array in the mitochondrial population without altering the intramitochondrial longitudinal distribution of the nucleoids. Hence, this heterozygous-like system constitutes a pathologically relevant model to study nucleoid distribution upon alterations of mitochondrial dynamics without mtDNA depletion.

Cristae ultrastructure and nucleoid-cristae proximity in cells expressing OPA1 ADOA-causing mutants

Previous data indicate that ADOA patients show IMM cristae defects [40, 48, 52]. Thus, we further evaluated the IMM ultrastructure. Our data showed that stable OPA1 overexpression leads to an increase in short cristae (17% vs. 8% CTL) and aberrant cristae (23% vs. 16% CTL) (Fig. 4A–B). OPA1 c.870+5 G > A mutant’s expression caused the same tendency, with an increase in short cristae (21% vs. 17% WT OPA1) and aberrant cristae (30% vs. 23% WT OPA1). Interestingly, cells expressing the mutant c.2713 C > T showed similar abundance of aberrant cristae (21% vs. 23% WT OPA1) but a decrease in short cristae (8% vs. 17% WT OPA1) (Fig. 4A–B).

Fig. 4: Cristae ultrastructure and nucleoid-cristae proximity in cells expressing OPA1 ADOA-causing mutants.
figure 4

A Representative TEM images of WT MEF stably expressing a lentiviral plasmid containing control cDNA (CTL), OPA1 WT, or pathogenic OPA1 variants. Scale bar is 1 μm. B Quantification of mitochondrial cristae shape descriptors. C Quantification of mitochondrial cristae periodicity. Data are from ≥ 112 objects from ≥ 3 independent experiments. D Representative immunofluorescence images of Skeletal muscle-derived fibroblasts from ADOA patients carrying OPA1 variants. Cells were labeled with antibodies against dsDNA and Mic60. The white boxes correspond to the 5 x 5 μm insets. (Right) Intensity profile of dsDNA and Mic60 signals along the yellow line from left to right (2 µm). E, F Colocalization coefficients for dsDNA and Mic60 in the cells shown in D. Each point represents one cell. Data are mean ± SEM from ≥ 70 cells from 4 independent experiments (****p < 0.0001).

Full size image

We then quantified cristae periodicity and found that cells with stable overexpression of OPA1 showed a decrease in periodic cristae (Fig. 4A–C). The expression of both mutants caused an increase in the interrupted cristae phenotype (31% c.870+5 G > A; 38% c.2713 C > T vs. 25% WT OPA1) (Fig. 4A–C).

We further explored if the cristae defects impact nucleoid-cristae proximity in Skeletal muscle-derived fibroblasts of ADOA patients. In all the conditions, nucleoids seemed in close proximity to the cristae, which was corroborated by the mtDNA to Mic60 colocalization (Fig. 4D–E). However, the Mic60 colocalization to mtDNA was decreased in the cells carrying the mutant c.870+5 G > A (Fig. 4F), which might be explained by the increased aberrant cristae present in these cells (Fig. 4B).

Therefore, OPA1 is relevant for cristae structuration and nucleoid array in the mitochondrial population. Nevertheless, we did not observe major domain-specific differences of OPA1 mutants on nucleoid distribution, IMM structure, and nucleoids-to-cristae proximity.

Discussion

Here, we demonstrated that the absence of OPA1 or the expression of OPA1 ADOA-causing mutants, lead to an altered nucleoid array, suggesting that the execution of mitochondrial fusion is needed to keep nucleoid distribution amongst the mitochondrial population. Also, single-organelle analysis showed that the absence of OPA1 alters the cristae periodicity and nucleoids-to-cristae proximity, leading to perturbed longitudinal nucleoid distribution within each mitochondrion. Yet, the expression of OPA1 disease-causing mutants showed some defects in cristae periodicity, but no changes in longitudinal nucleoid distribution (Fig. 5). Thus, these data allow us to dissect the role of OPA1, with fusion supporting nucleoids array amongst organelles, while cristae shaping determining intra-mitochondrial mtDNA distribution.

Fig. 5: Summary of main findings.
figure 5

The cartoon and table show mitochondrial morphology and cristae ultrastructure changes, as well as nucleoid (green circles) abundance and nucleoid distribution alterations in the studied models. Mitochondrial fusion data were obtained from Cartes-Saavedra, et al. [48].

Full size image

Different groups have studied the role of OPA1 in maintaining mtDNA and nucleoid abundance in mammalian cells [8, 33, 39, 41, 42, 53]. Yet, we emphasize the importance of describing nucleoid distribution considering the mitochondria available to contain them. Hence, we proposed a new paradigm based on our nucleoid distribution model (Fig. 1B), that incorporates mitochondrial characteristics and single-organelle analysis. Nucleoid distribution has gained attention as a functional parameter related to mtDNA distribution and inheritance [23, 54, 55]. Using this new paradigm, we showed that OPA1 plays a vital role in the distribution of the nucleoids in the mitochondrial population and at a single organelle level.

We found that in the absence of Opa1 and further rescue with OPA1 isoform 1, there is a correlation between mitochondrial morphology, cristae ultrastructure, mtDNA levels, and nucleoid number, consistent with previous reports [41]. Thus, the rescue of OPA1 isoform 1 restores mitochondrial fusion [48], morphology and cristae ultrastructure supporting the semi-regular distribution of the nucleoids and mtDNA levels, independently from the direct interaction proposed between OPA1exon 4b isoforms and mtDNA [39, 42]. Previous studies showed that rescue of the mtDNA levels upon the expression of different OPA1 isoforms in Opa1−/− cells are related to rescue of fusion and IMM ultrastructure and not to morphology changes [41], highlighting these two components as possible regulators of mtDNA abundance.

OPA1 overexpression increased mitochondrial fusion rate [48], but the relative abundance of mtDNA molecules was unaltered. However, nucleoids are more spread within the mitochondrial population than in WT cells. This supports the idea that fusion-induced mitochondrial elongation or material exchange play a role in the distribution of nucleoids amongst mitochondria, as has been showed in the Mfn 1/2 KO model [56].

We did not find mtDNA deletions in our stable cell lines with mutant OPA1, although multiple mtDNA deletions have been found in skeletal muscle biopsies from ADOA patients [46]. The accumulation of mtDNA damage is characteristic for post-mitotic cells and it emphasizes the relevance of studying nucleoid distribution in post-mitotic tissues and on subcellular level, as the local accumulation of damaged mtDNA causes bioenergetic dysfunction in these patients [2, 18].

The nucleoid array in the mitochondrial network describes how they are spread, which is relevant for the local bioenergetic function of individual mitochondria and the cells. It is proposed that mtDNA has a “sphere of influence” over the nearby cristae, locally controlling OXPHOS function [57, 58]. Since mitochondria without nucleoids (“empty”) are accumulated in Opa1−/− cells and WT cells expressing the ADOA-mutants, this phenotype might be related to the overall bioenergetic dysfunction found in these cells [51, 59], and most importantly, in ADOA patients [18]. In WT MEF overexpressing OPA1 and the c.2713 C > T ADOA-causing mutant, nucleoid array was altered on the organelle’s population, associated with mitochondrial shape and cristae disorganization. However, expression of the c.870+5 G > A mutant caused cristae alterations but a milder effect on the nucleoid array in the mitochondrial population, suggesting that nucleoid distribution amongst mitochondria is not exclusively directed by cristae ultrastructure.

Interestingly, intramitochondrial longitudinal distribution of nucleoids was a robust parameter only altered in Opa1−/− cells, which also showed altered cristae periodicity and nucleoids-to-cristae proximity. Both parameters were partially rescued after OPA1 isoform 1 expression, suggesting a key role of this isoform in the arrangement of nucleoids within each mitochondrion, and the periodicity of IMM folding. This finding aligns with the super-resolution data that showed the nucleoid interspersed between the cristae in live cells [60] and the nucleoid disorganization upon cristae disorganization due to Mic60 silencing [61].

In conclusion, our new paradigm allowed us to characterize nucleoid distribution in relation to the characteristics of the mitochondria that contain them. We showed that isoform 1 of OPA1 plays a vital role in the distribution of the nucleoids in the mitochondrial population and single organelle level by means of IMM fusion and cristae organization. Furthermore, OPA1 disease-causing mutants lead to altered nucleoid distribution irrespective of the protein domain where the mutation is located, associated with fusion inhibition and cristae morphology defects, providing new insights into the pathophysiology of ADOA.

Materials and methods

Extended Materials and Methods can be found in Supplementary Information.

Live-cell nucleoid cluster imaging

MEF were plated on 25 mm glass coverslips and grown in culture up to 80% confluence. Nucleoids were stained using 1 µL/mL PicoGreen (15 min, 37 °C; Thermo Fisher Scientific, Massachusetts, USA) or by transfection with Tfam-GFP. Mitochondria were stained with MitoTracker Deep Red 75 nM (15 min, 37 °C; Thermo Fisher Scientific) or by transfection of mCherry-Omp25. Live-cell imaging was performed using extracellular media at 37 °C, as previously described [51]. Image acquisition was performed with 5 Z-stacks (0.19 µm interval) in a confocal microscope Zeiss LSM 880 with an Airyscan detector (63X PlanApo 1.4 NA) on the Fast SR mode, at the Advanced Microscopy Facility UMA UC. After acquisition, images were deconvolved in 3D using the ZEN Black software.

Nucleoid cluster distribution quantification

We manually quantified nucleoid distribution and abundance inside single organelles (Fig. S1A). The manual analysis was performed in FIJI ImageJ. We detected a high Tfam-GFP intramitochondrial background noise Opa1−/− cells, previously described as “diffused TFAM” [39]. Thus, we adjusted the B/C settings to distinguish the brightest puncta (Fig. S3A–C). To strengthen our distribution data in the mitochondrial population, we developed a semi-automated method called Mitochondrial Nucleoid Distribution (MiNuD; Fig. S1B). The input for MiNuD are the .tif files of light-microscopy images of mitochondria, nucleoids, and their respective masks obtained from FIJI threshold or Trainable Weka.