The spatio-temporal dynamics of mitochondrial membrane potential during oocyte maturation

Abstract

Mitochondria are highly dynamic organelles and their distribution, structure and activity affect a wide range of cellular functions. Mitochondrial membrane potential (∆Ψm) is an indicator of mitochondrial activity and plays a major role in ATP production, redox balance, signaling and metabolism. Despite the absolute reliance of oocyte and early embryo development on mitochondrial function, there is little known about the spatial and temporal aspects of ΔΨm during oocyte maturation. The one exception is that previous findings using a ΔΨm indicator, JC-1, report that mitochondria in the cortex show a preferentially increased ΔΨm, relative to the rest of the cytoplasm. Using live-cell imaging and a new ratiometric approach for measuring ΔΨm in mouse oocytes, we find that ΔΨm increases through the time course of oocyte maturation and that mitochondria in the vicinity of the first meiotic spindle show an increase in ΔΨm, compared to other regions of the cytoplasm. We find no evidence for an elevated ΔΨm in the oocyte cortex. These findings suggest that mitochondrial activity is adaptive and responsive to the events of oocyte maturation at both a global and local level. In conclusion, we have provided a new approach to reliably measure ΔΨm that has shed new light onto the spatio-temporal regulation of mitochondrial function in oocytes and early embryos.

Introduction

Mitochondria regulate a number of cellular functions including ATP production, metabolism, Ca²⁺ homeostasis and cell death. Mitochondrial functions are primarily driven as a result of the bioenergetic properties of the electron transport chain, which pumps hydrogen ions across the inner mitochondrial membrane creating a highly negative mitochondrial membrane potential (∆Ψm). To generate ATP, hydrogen ions move back down this electrochemical gradient driving the ATP synthase in the process (Mitchell and Moyle, 1967). Mitochondrial ATP production is particularly critical in oocytes and early mammalian embryos because there is a block to glycolysis that is only relieved at the morula-blastocyst transition (reviewed in Grindler and Moley, 2013). The absolute reliance of oocytes and preimplantation embryos on mitochondrial ATP production has led to a significant focus on the role and function of mitochondria in the developmental potential of oocytes and early embryos (Eichenlaub-Ritter et al., 2004; Zhang et al., 2006; Wang et al., 2009; Ben-Meir et al., 2015; Boots et al., 2016).

Oocyte developmental competence and ability to support embryonic development are often linked with the mitochondria-dependent bioenergetic capacity (reviewed in Grindler and Moley, 2013). Interventions that inhibit mitochondrial function cause defects in meiotic and mitotic spindles and inhibit oocyte maturation and embryo development (Zhang et al., 2006; Dalton et al., 2014; Wu et al., 2015). Furthermore, environmental or pathophysiological challenges including obesity, metabolic disease, maternal ageing and oxidative stress are all associated with defects in oocyte mitochondria such as altered ∆Ψm, localization, altered cristae structure and reduced ATP production (Ottosen et al., 2007; Grindler and Moley, 2013; Tilly and Sinclair, 2013; Wu et al., 2015; Pasquariello et al., 2019). These functional effects are correlated with poor developmental outcomes in embryos, which place mitochondrial activity as a major determinant of oocyte quality or competence.

During in vitro oocyte maturation, there is a dramatic redistribution of mitochondria. In germinal vesicle (GV) stage mouse oocytes, they are relatively homogeneously distributed but after germinal vesicle breakdown (GVBD) and as the metaphase I spindle forms mitochondria aggregate with the endoplasmic reticulum (ER) around the developing first meiotic spindle (Van Blerkom and Runner, 1984; Dalton and Carroll, 2013). This is an active process dependent on dynein-mediated trafficking on microtubules. The cytoplasmic superstructure comprising the metaphase I (MI) spindle enveloped in a sheath of ER and mitochondria is then moved via actin-mediated forces to the cortex (Dalton and Carroll, 2013) where anaphase is initiated and the first polar body is extruded. Mitochondria then redistribute around the cytoplasm in preparation for fertilization at metaphase II (MII). These changes in mitochondrial distribution correlate with increased ATP levels during spindle formation and at polar body formation (Yu et al., 2010; Dalton and Carroll, 2013; Dalton et al., 2014), suggesting that mitochondrial activity may adapt to metabolic demand.

Mitochondrial membrane potential can be measured using a number of potentiometric cationic fluorescent indicators such as tetraethylbenzimidazolylcarbocyanine iodide (JC-1), tetramethylrhodamine methyl ester (TMRM) and Rhod-2. Through the use of JC-1 in oocytes and embryos, it has long been reported and has become widely accepted that ∆Ψm is spatially organized such that mitochondria in the cortex have a higher ∆Ψm relative to those in the peri-nuclear region (Van Blerkom et al., 2000; Van Blerkom et al., 2008; Tatone et al., 2011; Zhao et al., 2011; Boudoures et al., 2017; Pasquariello et al., 2019). This reported high cortical ∆Ψm has implications for understanding the role of mitochondria in oocytes and embryos. For example, the high cortical ∆Ψm has been purported to be important for Ca²⁺ handling at fertilization, while the low ∆Ψm deeper in the cytoplasm implies that low levels of mitochondrial activity are sufficient to drive the events of spindle formation and function (Van Blerkom and Davis, 2007).

It is of interest that the highly polarized mitochondria in the cortex are reported in all studies using JC-1 but not in those that utilize TMRM (or other single wavelength indicators). JC-1 is lipophilic and cationic, which allows its accumulation in mitochondria. It has complex spectral properties such that at low concentrations, it is monomeric and emits in the green, while at high concentrations, there is a spectral shift to red emission because of so called ‘J-aggregate’ formation. As such, the red:green emission ratio gives a relative measure of ∆Ψm in the cell. The highly polarized cortical mitochondria indicated by JC-1 are not seen in studies using single wavelength indicators such as TMRM (Dumollard et al., 2004, Zhang et al., 2006, Igosheva et al., 2010). TMRM is a small cationic fluorescent indicator that accumulates in mitochondria purely based on ∆Ψm. These contrasting observations arising from different laboratories using different indicators suggests the need to revisit the question as to whether there is a spatial heterogeneity in ∆Ψm in oocytes and embryos.

The increasing appreciation of mitochondrial function in oocytes and the potential for using mitochondrial activators, or mitochondrial supplementation/donation for improving oocyte quality or avoiding mitochondrial genetic disease, makes it an imperative that accurate and reliable methods are available for measuring mitochondrial function. In this study, we have validated a novel ratiometric approach to measure ∆Ψm in oocytes and embryos. We use the assay to measure ∆Ψm through the time course of oocyte maturation and reveal for the first time ∆Ψm heterogeneity in sub-cellular compartments, most notably in the mitochondria surrounding the MI spindle. Our measurements of ∆Ψm using these approaches show no evidence for polarized cortical mitochondria.

Materials and Methods

Oocyte collection

Oocytes at the GV stage were collected from 3- to 4-week-old C57 female mice. These mice had been stimulated by intraperitoneal injection of 7 international units (IU) pregnant mare’s serum gonadotropin (PMSG; Intervet, Buckinghamshire, UK) 44–48 hours prior to oocyte retrieval. After removing adipose tissue surrounding the ovaries, cumulus oocyte complex (COC) containing fully grown GV-stage oocytes encircled by cumulus cells was released by puncturing the ovaries in M2 medium supplemented with 200 μM 3-isobutyl-1-methylxanthine (IBMX) (Sigma-Aldrich, St. Louis, Missouri, USA) to maintain GV arrest at 37°C. Attached cumulus cells were removed by repetitive pipetting of COCs. PhAM^loxP/loxP mice (Jax mice stock no: 018385) (Pham et al., 2012) were crossed with transgenic mice that carried Gdf-9 promoter-mediated Cre recombinase, which had a C57BL/6J background (Lan et al., 2004). Homozygous mutant female mice expressed a mitochondrial-specific version of Dendra2 green/red photo switchable monomeric fluorescent protein exclusively in oocytes (PhAM ^loxP/loxP; Gdf9-Cre). All mice were housed under controlled environmental conditions with free access to water and food. The Animal Ethics Committee of Monash University approved all animal handling and experimental protocols.

IVF and embryo culture

For in vitro fertilization, female mice (3–4 week) were primed with 5 IU PMSG and 5 IU human chorionic gonadotropin (hCG). One night before fertilization, human tubal fluid (HTF, Merck Millipore, Billerica, Massachusetts, USA) drops (50 μl/drop) and culture medium (CM, Life Global group, Belgium) overlaid with mineral oil (Sigma) were equilibrated in an incubator at 37°C and 5% CO₂ in air. One hour before oocyte collection, sperm were collected from cauda epididymides of 12-week-old C57 males into 1-ml HTF and placed in CO₂ incubator for capacitation. The COCs were mixed with capacitated sperm HTF drops and incubated for 2–4 hours. COCs were then washed through CM drops three times and were cultured in CM medium and monitored for the appropriate time to obtain 2-cell and 4-cell embryos.

Use of probes and inhibitors

All probes were loaded in drops of M2 under oil at 37°C. The concentration and loading times are provided in the relevant figure legends. In short, three different concentrations of JC-1 (0.5, 1 and 5 μg ml⁻¹), 25 nM of TMRM, and 100 nM of MitoTracker Green (MTG) were used to visualize mitochondria. In some experiments requiring dynamic time-lapse measurement of mitochondrial membrane potential, a low concentration of TMRM (5 nM) was included in the imaging drops. FCCP (Carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone) (0.5, 1 and 5 μM) was added directly to the imaging drops between acquisitions.

Confocal microscopy

Live-cell imaging was performed on an inverted confocal microscope (SP8, Leica, Germany) fitted with an environmentally controlled chamber to maintain a temperature of 37°C. Oocytes loaded with mitochondrially targeted probes, JC-1 (Sigma), TMRM (Sigma) and MTG (Thermo Fisher, Waltham, Massachusetts, USA) were placed in drops of M2 in an imaging chamber with a coverslip as its base. Images were acquired using a 40× water immersion 1.2 N A objective with line averaging set to 2–4. It should be noted that on exposure to UV, Dendra can be photoconverted to red emission wavelengths. We confirmed that using the excitation conditions in our experiments, Dendra did not undergo any photoconversion and was therefore suitable as a reliable and stable reporter of mitochondrial mass.

Ratiometric image analysis

Ratiometric analysis was performed using ImageJ software (NIH). Briefly, Dendra (green) and TMRM (red) channels of the confocal images were separated and converted to 32-bit images, and the background of each channel was set to not a number using thresholding. The thresholded TMRM fluorescence images were then divided by the Dendra fluorescence image pixel by pixel using the Image Calculation function in ImageJ software.

Experimental design and statistical analysis

Experiments were repeated at least three times and number of oocytes/embryos was taken as n for statistical analyses and mean and SD (or SEM) were calculated. Oocyte/embryos from 1–2 mice were used for each repeat. Error bars indicate SD or SE of the mean as mentioned in the individual results. Statistical analysis was performed using either student t-test or one-way ANOVA with Microsoft Excel or GraphPad Prism Software. Level of significance is denoted by ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 and ^****P < 0.0001 as shown in the figures.

Results

Monitoring spatial distribution of ∆Ψm with JC-1 and TMRM

The two most common approaches for monitoring ∆Ψm are through the use of either JC-1 or TMRM. Loading of mouse GV stage oocytes with JC-1 (1.0 μg/ml) for 30 minutes prior to imaging using confocal microscopy confirmed published observations that there was a bright cortical ring of red J-aggregate fluorescence and very low levels deeper in the cytoplasm (Fig. 1A, i), while the monomer appeared relatively evenly distributed across the cytoplasm (Fig. 1A, ii). The J-aggregate (red)/monomer (green) ratio was therefore much higher in the cortex than in the deeper cytoplasm, indicating, at face value, that ∆Ψm is significantly elevated in the cortex.

Figure 1.

Distribution of JC-1 and TMRM. (A) J-aggregate (i) and monomer (ii) distributions after loading with JC-1 (1 μg/ml). (B) TMRM (25 nM) fluorescence distribution in GV oocyte. (C) Ratio of the fluorescence intensities in cortex and cytoplasm. (D) J-aggregate and (E) TMRM-labeled 2-cell embryos. (F) Ratios of cortex/cytoplasm fluorescence intensities of J-aggregates and TMRM. (G) J-aggregate and (H) TMRM labeled 4-cell embryos. (I) Ratios of cortex/cytoplasm fluorescence intensities of J-aggregates and TMRM. Results are presented from three replicate experiments for each stage. ^*P ≤ 0.05, compared to TMRM; ^****P ≤ 0.0001, compared to TMRM. Error bars show SE of the mean. n = Total number of oocyte/embryos analyzed.

We next measured the cortical/cytoplasmic distribution of the J-aggregate and monomer, finding an approximate 8-fold accumulation of J-aggregate in the cortex (Fig. 1C). In contrast, the monomer was more evenly distributed with a ratio closer to 2 (Fig. 1C). To explore the spatial distribution of J-aggregate further and investigate the relevance to ∆Ψm, oocytes were loaded with the potentiometric indicator TMRM (25 nM for 30 minutes). In contrast to JC-1, no obvious gradient of fluorescence intensity was evident and this was reflected in a cortex/cytoplasm ratio that was consistently very close to unity (Fig. 1B and C). To examine if this difference between JC-1 and TMRM behaviors is specific to GV stage oocytes, we performed similar measurements in two-cell (Fig. 1D–F) and four-cell embryos (Fig. 1G–I). At both stages of development, the J-aggregate distribution ratio was greater than 5 while that of TMRM was consistently at unity (Fig. 1F and I).

JC-1 distribution is highly concentration- and time dependent

By varying the loading dose of JC-1, we found that the J-aggregate cortex/cytoplasm ratio was highly concentration dependent; but even concentrations as low as 0.5 μg/ml maintained a J-aggregate ratio of approximately 7 (Fig. 2A and B and Fig. S1). Since lower concentrations were not compatible with a reliable fluorescence signal, we next increased loading times to determine if an increased equilibration time could lead to a normalized distribution of J-aggregates. There was a strong inverse correlation between the J-aggregate cortical/cytoplasmic ratio and time, with a ratio of 8.19 ± 0.71 for a 30-minute loading decreasing to 3.42 ± 0.66 over a 2-hour loading period (Fig. 2C and D and Fig. S1). Furthermore, this was similar if the loading period was confined to 30 minutes and then followed by a prolonged equilibration period in JC-1-free medium (Fig. S2A and B). It is unclear if J-aggregate distribution would eventually reach equilibration with even longer loading times, but the potential deleterious impact of prolonged exposure to JC-1 may then bring into question the reliability and practicality in an experimental setting. This conclusion is supported by the finding that 2-cell embryos loaded with JC-1 (30 minutes; 1.0 μg/ml) failed to progress to the 4-cell stage and started to degenerate (Fig. S3 D and E); while after loading with TMRM, over 80% progressed to the 4-cell stage (Fig. S3 A, B and E). Taken together, these data demonstrate that JC-1 distribution is influenced by concentration and equilibration time and that these need to be balanced against potential toxicity.

Figure 2.

Effect of JC-1 concentration and time on spatial distribution of J-aggregates in oocytes. (A) Quantification of spatial distribution of J-aggregates after incubation with various JC-1 concentrations for the same period of time (30 minutes). Oocytes were also cultured in TMRM (25 nM) for 30 minutes. ANOVA was used to compare between groups. A statistical comparison was shown as letters, a versus b represents P ≤ 0.01; c versus a or b represents P ≤ 0.0001. (B) Representative images of oocytes used for quantification in A. (C) Quantification of outside/inside ratios of J-aggregates after incubation for different periods of time in the same concentration of JC-1 (1 μg/ml). For comparison, oocytes were also cultured in TMRM (25 nM) for 30 minutes. A statistical comparison was shown as letters after using ANOVA, all differences between groups were P ≤ 0.0001 except that comparison between 90 and 120 minutes (c versus d) was P ≤ 0.001. (D) Representative images of oocytes used for quantification in C. Results were combined from 20 oocytes from three replicate experiments. The ratio on the Y-axis was obtained by dividing the fluorescence intensity of the region of interest (ROI) in cortex by that in the cytoplasm. Error bars show SE of the mean.

TMRM and JC-1 can reliably report dynamic changes in ∆Ψm in oocytes

The data presented to date suggest that the properties of J-aggregates to preferentially accumulate in the cortex may make JC-1 an unreliable indicator for the spatial distribution of ∆Ψm in oocytes and embryos. Nevertheless, it may remain valid for the measurement of ∆Ψm at the whole-cell level. Oocytes were loaded with JC-1 (30 minutes; 1.0 μg/ml) and fluorescence was recorded using confocal microscopy. On application of FCCP, the J-aggregate signal decreased while the monomer signal increased (Fig. S4A), confirming that on depolarization of ∆Ψm, J-aggregates revert to monomers and the ratio of both wavelengths can be plotted to obtain a time course of mitochondrial membrane potential collapse (Fig. S4B). As expected, TMRM, which localizes to mitochondria exclusively due to its positive charge, re-equilibrates in the cytosol as the mitochondrial membrane potential collapses upon FCCP treatment (Fig. S4C and D). JC-1 and TMRM report a similar time course of FCCP-induced depolarization suggesting that both are reliable reporters of changes in ∆Ψm at the whole-cell level.

A new combination of probes for ratiometric measurement of ∆Ψm in oocytes

Our data suggest TMRM has several advantages over JC-1, including its more uniform potentiometric distribution into mitochondria and decreased toxicity. The one advantage of JC-1 is that the ∆Ψm-sensitive J-aggregate fluorescence can be expressed as a ratio against the fluorescence of the monomer, which in theory provides a measure of mitochondrial activity that is independent of mitochondrial mass. Thus, to take advantage of the properties of TMRM, we have combined it with other mitochondria-targeted fluorescent indicators that are insensitive to ∆Ψm (MTG or mito-Dendra); an approach that has been successful in other cell types (Lemasters and Ramshesh, 2007).

Oocytes were loaded with TMRM and MTG and imaged using confocal microscopy. Addition of FCCP caused the predicted decrease in TMRM fluorescence, while MTG remained essentially unchanged (Fig. 3A and B). The TMRM/MTG fluorescence ratio can be calculated and provides a means of measuring ∆Ψm independent of mitochondrial mass (Fig. 3C).

Figure 3.

Combination of probes for ratiometric measurement of ∆Ψm in oocytes. (A) After loading with TMRM and MTG for 30 minutes, loss in fluorescence intensities were monitored. (B) Representative images for MTG and TMRM fluorescence. (C) TMRM/MTG fluorescence ratio plotted after carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) addition. (D) Comparison of TMRM/MTG fluorescence over time after addition of different concentrations of FCCP. The experiments were repeated three times each and the values before FCCP addition were normalized to 1. Error bars show SE of the mean.

We next sought to examine whether the TMRM/MTG co-labeling could discriminate between conditions that lead to different resting ∆Ψm. Oocytes were loaded with TMRM and MTG as above and a concentration response to FCCP (0.5, 1 and 5 μM) was examined (Fig. 3D). To directly compare the three conditions, the baseline TMRM/MTG ratio was normalized to 1.0 and followed over time (Fig. 3D). The data clearly demonstrate that different concentrations of FCCP lead to different ratios (Fig. 3D), reflecting that this ratiometric measure of mitochondrial activity is sensitive to discrete perturbations in ∆Ψm. Thus, this indicator combination can be used to discriminate ∆Ψm in different populations of oocytes.

Recording ∆Ψm through the duration of oocyte maturation

We next examined whether oocyte maturation is accompanied by any fluctuations in ∆Ψm. Given that TMRM equilibrates across the inner mitochondrial membrane according to Nernstian principles, decreases in ∆Ψm lead to the loss of TMRM down the potentiometric gradient and increases in ∆Ψm result in the accumulation of TMRM. As such, in order to measure any increases in ∆Ψm, a low concentration of TMRM (5 nM) was included during imaging.

MTG fluorescence remains constant during the 14 hours of oocyte maturation, while TMRM shows a 50% increase in fluorescence over the period of 100–400 minutes, after which it remains steady (Fig. 4A). The TMRM/MTG ratio shows an increase in ∆Ψm during oocyte maturation (Fig. 4A and B). This increase, which starts around the expected time of GVBD, indicates that maturation-dependent changes in ∆Ψm may be an important feature of successful oocyte maturation. In support of this, the increase in ∆Ψm is dependent on meiotic progression because no excursions in fluorescence or the TMRM/MTG ratio were observed if oocytes were maintained in meiotic arrest for the duration of the recording (Fig. 4C and D).

Figure 4.

Increase in ∆Ψm during oocyte maturation. GV oocytes were incubated with TMRM and MTG for 30 minutes and then matured in vitro in medium containing low TMRM concentration (5 nM). (A) Increase in TMRM fluorescence between 1.5 and 7 hours but with constant levels of MTG fluorescence. (B) Increase in TMRM/MTG ratio during oocyte maturation. (C) Oocyte meiotic maturation was blocked by adding 200 μM 3-isobutyl-1-methylxanthine (IBMX) to culture medium. This prevented the increase in TMRM fluorescence and the increase in TMRM/MTG fluorescence ratio (D). Each experiment was repeated at least three times. Error bars show SE of the mean.

Measuring spatial distribution of ∆Ψm

Having shown that long-term recordings of dynamic changes in ∆Ψm are feasible in oocytes, we next sought to determine whether combinations of indicators could be reliably used to measure the spatial distribution of ∆Ψm within the oocyte. Mito-Dendra expressing oocytes from PhAM^loxP/loxP; Gdf9Cre mice were loaded with TMRM as above and confocal imaging was performed to obtain high-resolution images at both wavelengths. The images were arithmetically divided pixel × pixel to provide a ratiometric image of ∆Ψm independent of concentration of mitochondria in different locations across the oocyte. It is clear from examining the ratio images that, consistent with TMRM distribution, there is no evidence of a peri-cortical increase in ΔΨm in oocytes at different stages of maturation. In fact, at the GV stage, peri-nuclear mitochondria (within 5 μM of GV, see Fig. S5A) show a small but significant increase in ΔΨm compared to the rest of the mitochondria (Fig. 5A–D). Further analysis revealed that at metaphase I, there was a significant 20% difference in ∆Ψm across the oocyte; spindle-associated mitochondria had higher ∆Ψm than non-spindle associated mitochondria (Fig. 5E–H, and S5B). Sub-cellular heterogeneity was also observed in MII stage oocytes (Fig. 5I–L and S5C), although less obvious than that seen at MI. In MII stage oocytes, we compared ΔΨm in the upper portion of the spindle-containing hemisphere with that at the opposite pole of the oocyte and found that ΔΨm was approximately 7% higher in the spindle region (Fig. 5L). To determine whether mitochondrial cluster size might have caused differences in ∆Ψm, we correlated ∆Ψm with mitochondrial cluster sizes in MI stage oocytes. The results show that ∆Ψm did not correlate with mitochondrial cluster size (Fig. S6). Thus, at each stage of oocyte maturation, mitochondrial activity is highest in the peri-nuclear region.

Figure 5.

Spatial distribution of ∆Ψm during oocyte maturation. Mito-Dendra expressing oocytes from PhAM^loxP/loxP; Gdf9Cre mice at GV, metaphase I and II (MI and MII) were loaded with TMRM. High-resolution confocal images were obtained at both wavelengths to calculate TMRM/Dendra ratios across various regions of the oocytes. (A–D) Mitochondria around the GV have higher membrane potential than the mitochondria not surrounding the GV. n = 28. ^****, P < 0.01. (E–H) GV oocytes were cultured in vitro for 8 hours to obtain MI oocytes. MI oocytes show higher ∆Ψm in mitochondria associated with spindle than in those away from the spindle. n = 32. ^****, P < 0.0001. (I–L) Ovulated MII oocytes were collected 14 hours after hCG injection. Mitochondria around the MII spindle (spindle pole) have higher ∆Ψm than those at opposite pole. n = 16. ^**, P < 0.01. Data shown as mean ± SEM.

Discussion

Mitochondrial function is critical to oocyte quality. Conditions that cause sub-fertility such as maternal ageing, obesity and metabolic disease are widely regarded to be acting through compromising mitochondrial activity, causing reduced ATP and increased ROS (Grindler and Moley, 2013; Schatten et al., 2014; Wu et al., 2015). The growing appreciation of mitochondrial function in oocyte and embryo development compelled us to investigate ΔΨm in oocytes with a particular focus on the spatial distribution of ΔΨm in the oocyte cytoplasm.

To tackle this question, it was first necessary to investigate the properties of the fluorescent indicators used for measuring ∆Ψm. The discordance in the literature regarding sub-cellular heterogeneity in ∆Ψm was confirmed in our investigations in that JC-1 and TMRM report very different distributions in ∆Ψm in mouse oocytes and early embryos. JC-1 showed high ΔΨm in the oocyte cortex relative to the inner cytoplasm, similar to previously reported patterns (Van Blerkom et al., 2000, 2002, 2003, 2008; Van Blerkom and Davis, 2006; Tatone et al., 2011; Zhao et al., 2011; Boudoures et al., 2017; Pasquariello et al., 2019), while in contrast, TMRM distribution showed a homogeneous ΔΨm distribution across the cytoplasm (Dumollard et al., 2004; Zhang et al., 2006; Igosheva et al., 2010). Despite efforts to adjust concentration and equilibration times, the J-aggregate distribution remained higher in the cortex. The ability of longer incubation times to reduce the gradient of J-aggregate across the oocyte indicates that diffusion and membrane permeability properties, combined with a non-linear J-aggregate formation, may contribute to the high levels of J-aggregate fluorescence in the cortical mitochondria. Nevertheless, while challenging to use for subcellular distribution, the relative level of red/green fluorescence at the whole cell level, is sensitive to membrane potential collapse so can be used to report overall differences in ΔΨm between groups of oocytes.

TMRM is conceptually a more predictable indicator to work with, as it is small and highly permeable and is predicted to distribute across membranes according to Nernstian principles (Nicholls and Ward, 2000). As such, it is reasonable to conclude that the failure of TMRM to report an increase ∆Ψm in the cortex indicates ∆Ψm is more homogeneously distributed in the oocyte cytoplasm. In addition, its lack of toxicity makes it ideal for long-term imaging experiments and it can also be used ratiometrically by combining with a second spectrally separable mitochondrial reporter that is insensitive to ΔΨm. This approach provides a reliable method for use of TMRM to measure ∆Ψ independent of mitochondrial mass.

We have found MTG, or mitochondria-targeted GFP/Dendra, to be reliable reporters for combining with TMRM in oocytes. While both work well, we find that mitochondrially targeted fluorescent proteins provide a cleaner signal than MTG, which can also bind to non-mitochondrial sites in the cytoplasm. Similar ratiometric approaches have been used in a variety of other cell types (Pendergrass et al., 2004; Lemasters and Ramshesh, 2007; Mourier et al., 2015). One important caveat with this approach is that the behavior of the second reporter must solely report mitochondrial mass, and not be influenced by the physiological state of the oocyte or the location of mitochondria in the cell. Careful consideration of these issues is needed for every experimental condition.

We have taken advantage of this ratiometric approach to undertake the first recordings of ΔΨm during the entire time course of oocyte maturation. Interestingly, the increased ΔΨm we observe during oocyte maturation shows good temporal correlation with the previously reported increase in ATP during oocyte maturation (Yu et al., 2010; Dalton et al., 2014). Oocytes that remain arrested at the GV stage throughout the entire recording period show no excursions in ΔΨm or ATP. This indicates that on progression into meiosis, the energy demands increase causing an adaptive response that drives an increase in ΔΨm to provide the ATP necessary to drive the events of maturation.

The ability to perform arithmetic ratiometric image analysis of the TMRM/mito-Dendra couplet has allowed us to further interrogate the spatial distribution of ΔΨm across the oocyte. Our results provide no evidence for a peri-cortical increase in ΔΨm, but do reveal other heterogeneities may exist. At the GV and MII stages, modest increases (<10%) in ΔΨm were observed in the region of the GV and MII spindle. These differences, while statistically significant, are small and as such, the physiological relevance needs further investigation. One potential function of polarized mitochondria in the peri-nuclear region is the provision of ATP to fuel nuclear transport, while at the MII stage, they may serve to maintain function of the arrested but highly dynamic MII spindle (Nakagawa and FitzHarris, 2017; Zhang et al., 2017).

The most dramatic compartmentalization of ΔΨm was apparent in the spindle-associated mitochondria at the MI stage. Mitochondria in this region showed a ratio increase of around 25% compared to peripheral mitochondria. Like the local ΔΨm increases seen in regions of GV and MII stage oocytes, it is tempting to speculate the local increase in ΔΨm at the MI stage reflects an adaptive response to the increased metabolic demands of spindle formation and function during this critical phase of oocyte maturation. While recent studies in neurons predict mitochondrial localization can influence spatial gradients of ATP that play a role in localized translation (Cioni et al., 2019), it is unknown at this stage whether highly active mitochondria close to the MI spindle provide any advantage to spindle formation or fidelity.

How ΔΨm can be locally regulated also needs further investigation. Response to a local high rate of ATP consumption may be one possibility, but it may also reflect the mitochondrial organization in relation to the ER. In mouse oocytes, the ER is known to surround the MI spindle (FitzHarris et al., 2007; Dalton and Carroll, 2013) and local Ca²⁺ cycling between the ER-cytosol-mitochondria has been reported to influence mitochondrial function (Mak et al., 1999). Alternatively, if dynein-mediated mitochondrial trafficking favored those mitochondria with a higher ΔΨm, more highly polarized mitochondria would accumulate in the spindle region.

In summary, our data show that there is a maturation-dependent increase in ΔΨm and that mitochondria in the vicinity of the first meiotic spindle show a local increase in ΔΨm. We find no evidence for the previously reported sub-cortical increase in ΔΨm and suggest that these observations are a reflection of the behavior of the indicator, JC-1, rather than mitochondria. Our findings suggest that mitochondrial activity shows both temporal and spatial regulation during the process of mouse oocyte maturation in vitro.

Supplementary data

Supplementary data are available at Molecular Human Reproduction online.

Author’s roles

All authors qualify for authorship by contributing substantially to this article. UAZ, DA and JC developed the original concept of this study collectively. Data collection was performed by UAZ, JL, OC and DA., statistical analysis by UAZ. RLR contributed to the discussion and interpretation of the results and writing the manuscript. All authors have contributed to critical discussion, reviewed the final version of the article and approved it for publication.

Funding

The authors would like to acknowledge the financial support from National Health and Medical Research Council and Australian Research Council, Australia. Additionally, U.A.Z. was supported by the Iraqi Higher Education and Scientific Research Ministry Ph.D. scholarship and O.C. by TUBITAK-1059B191601275.

Conflict of interest

There are no competing interests related to this study.