High-Throughput Image-Based Quantification of Mitochondrial DNA Synthesis and Distribution

For most eukaryotic cells mitochondria are essential organelles, as they play a crucial role in numerous cellular processes. First and foremost, mitochondria are the key energy suppliers of cells¹. Mitochondria are also involved in regulating cellular homeostasis (for instance, intracellular redox² and the calcium balance³), cell signaling^4,5, apoptosis⁶, the synthesis of different biochemical compounds^7,8, and the innate immune response⁹. Mitochondrial dysfunction is associated with various pathological states and human diseases¹⁰.

The functioning of mitochondria depends on the genetic information located in two separate genomes: the nuclear and mitochondrial genomes. The mitochondrial genome encodes a small number of genes compared to the nuclear genome, but all the mtDNA-encoded genes are essential for human life. The mitochondrial protein machinery necessary to maintain the mtDNA is encoded by nDNA. The basic components of the mitochondrial replisome, as well as some mitochondrial biogenesis factors, have already been identified (reviewed in previous research^11,12). However, mitochondrial DNA replication and maintenance mechanisms are still far from being understood. In contrast to nDNA, the mitochondrial genome exists in multiple copies, which provides an additional layer for regulating mitochondrial gene expression. Much less is currently known about the distribution and segregation of mtDNA within organelles, to what extent these processes are regulated, and if they are, which proteins are involved¹³. The segregation pattern is crucial when cells contain a mixed population of wild-type and mutated mtDNA. Their unequal distribution may lead to the generation of cells with a detrimental amount of mutated mtDNA.

So far, the protein factors necessary for mtDNA maintenance have been identified mainly by biochemical methods, bioinformatic analyses, or through disease-associated studies. In this work, in order to ensure a high chance of identifying factors that have previously escaped identification, a different strategy is described. The method is based on the labeling of mtDNA during replication or repair with 5-bromo-2'-deoxyuridine (BrdU), a nucleoside analog of thymidine. BrdU is readily incorporated into nascent DNA strands during DNA synthesis and, in general, is used for monitoring the replication of nuclear DNA¹⁴. However, the procedure developed here has been optimized for detecting BrdU incorporated into mtDNA using the immunofluorescence of anti-BrdU antibodies.

The approach allows for the high-throughput quantification of mtDNA synthesis and distribution in human cells cultured in vitro. A high-throughput strategy is necessary to conduct tests under different experimental conditions in a relatively short time; therefore, it is proposed in the protocol to utilize a multi-well format for cell culturing and automated fluorescence microscopy for imaging. The protocol includes the transfection of human HeLa cells with an siRNA library and the subsequent monitoring of mtDNA replication or repair using the metabolic labeling of newly synthesized DNA with BrdU. This approach is combined with immunostaining of the DNA with the help of anti-DNA antibodies. Both parameters are analyzed using quantitative fluorescence microscopy. Additionally, mitochondria are visualized with a specific dye. To demonstrate the specificity of the protocol, BrdU staining was tested on cells devoid of mtDNA (rho0 cells), on HeLa cells upon the silencing of well-known mtDNA maintenance factors, and on HeLa cells after treatment with an mtDNA replication inhibitor. The mtDNA levels were also measured by an independent method, namely qPCR.

1. Preparation of the siRNA mixture

1. One day before the start of the experiment, seed cells (e.g., HeLa) on a 100 mm dish so that they reach 70%-90% confluence the next day.
   NOTE: All operations must be carried out under sterile conditions in a laminar flow chamber.
2. Prepare the appropriate amount of siRNA diluted to a concentration of 140 nM in Opti-MEM medium (see Table of Materials). A 96-well plate can be used as a reservoir.
3. Add 5 µL of the siRNA solution (or Opti-MEM medium for the control samples) to each well of a black 384-well cell culture microplate.
   NOTE: Depending on the number of siRNA samples tested, an electronic or multi-channel pipette can be used.
4. Prepare the appropriate amount of RNAiMAX transfection reagent solution (see Table of Materials) in Opti-MEM medium. Add 1 µL of RNAiMAX for every 100 µL of medium.
5. Add 10 µL of the transfection reagent solution to each well of the 384-well plate. The most convenient and fastest way to do this is with a reagent dispenser.
6. Incubate the siRNA with the transfection reagent for 30 min at room temperature.

2. Preparation of cells for transfection

1. During the incubation (step 1.6), aspirate the medium using a suction device (see Table of Materials), and wash the cells with 3 mL of PBS.
2. Add 1.5 mL of trypsin diluted 1:2 in PBS, and incubate the cells at 37 °C for 10 min.
3. Check if the cells have detached; if so, add 3 mL of DMEM medium with 10% FBS. Suspend the cells thoroughly, and transfer them to a 15 mL tube. Take at least 150 µL of the suspension, and count the cells in a cell-counting chamber (see Table of Materials).
4. Prepare a cell suspension at the appropriate concentration (35,000/mL for the HeLa line) in DMEM medium with 10% FBS, penicillin, and streptomycin.

3. Cell transfection

1. Add 20 µL of the cell suspension to each well of the 384-well plate using the reagent dispenser. This will result in 700 cells seeded per well.
2. Incubate the cells for 1 h at room temperature, and then place them in the incubator for 72 h (37 °C and 5% CO₂).

4. BrdU incorporation

1. Prepare a 90 µM solution of BrdU (see Table of Materials) in DMEM medium with 10% FBS.
2. At 56 h after the siRNA transfection (16 h before cell fixation), add 10 µL of 90 µM BrdU solution to each well of the 384-well plate (the final BrdU concentration is 20 µM).
   NOTE: The action should be performed as fast as possible; therefore, it is best to use a reagent dispenser, electronic pipette, or multi-dispenser pipette. Remember to prepare control wells without the addition of BrdU.
3. Incubate the cells for 16 h (37 °C and 5% CO₂).

5. Labeling of mitochondria

1. Prepare a 20 µM solution of BrdU in DMEM with 10% FBS, and add to it the mitochondria tracking dye solution (see Table of Materials) to a concentration of 1.1 µM.
2. At 15 h after the start of BrdU incorporation (1 h before cell fixation), add 10 µL of the mitochondria tracking dye solution (see Table of Materials) to each well of the 384-well plate (the final dye concentration is 200 nM).
   NOTE: Use a reagent dispenser, electronic pipette, or multi-dispenser pipette. Remember to retain control wells without the addition of BrdU but with the addition of the mitochondria tracking dye.
3. Incubate the cells for 1 h (37 °C and 5% CO₂).

6. Cell fixation

NOTE: All washing is most conveniently carried out using a microplate washer, while the addition of reagents is most quickly carried out using a reagent dispenser.

1. Using the microplate washer (see Table of Materials), rinse each well twice with 100 µL of PBS. After the second wash, leave 25 µL of PBS in the well.
   NOTE: Leaving PBS in the well reduces the likelihood of cell detachment during the addition of liquid in the next step. All the washes are completed with the PBS left in; therefore, the solution added in the next step should always be prepared at a 2x concentration as it will be added in an equal volume to the remaining PBS.
2. Fix the cells by adding 25 µL of an 8% formaldehyde solution in PBS with 0.4% Triton X-100 and Hoechst 33342 at a concentration of 4 µg/mL (the final concentrations of individual reagents are 4%, 0.2%, and 2 µg/mL, respectively) (see Table of Materials).
3. Incubate the plate in the dark at room temperature for 30 min.

7. Blocking

1. Rinse each well four times with 100 µL of PBS. After the last wash, leave 25 µL of PBS in the well.
2. Add 25 µL of 6% BSA in PBS (the final BSA concentration is 3%) to each well.
3. Incubate the plate in the dark at room temperature for 30 min.

8. Addition of the primary antibodies

1. Aspirate the BSA using a microplate washer, and leave 10 µL of solution in the well.
2. Add 10 µL of the primary antibody solution prepared in 3% BSA in PBS. Use anti-BrdU at 0.8 µg/mL and anti-DNA at 0.4 µg/mL (the final concentrations of antibodies are 0.4 µg/mL and 0.2 µg/mL, respectively) (see Table of Materials).
3. Incubate the plate in the dark at 4 °C overnight.

9. Addition of the secondary antibodies

1. Rinse each well four times with 100 µL of PBS. After the last wash, leave 10 µL of PBS in the well.
2. Add 10 µL of the 4 µg/mL secondary antibody solution prepared in 6% BSA in PBS (the final concentration is 2 µg/mL). Use isotype-specific antibodies (anti-mouse IgG1 and anti-mouse IgM) conjugated to fluorochromes such as Alexa Fluor 488 and Alexa Fluor 555 (see Table of Materials).
3. Incubate the plate in the dark at room temperature for 1 h.
4. Rinse each well four times with 100 µL of PBS. After the last wash, leave 50 µL of PBS in the well.
5. Seal the plate with an adhesive sealing film (see Table of Materials), and store it in the dark at 4 °C. Imaging must be performed within 2 weeks.

10. Imaging

NOTE: Imaging must be performed with an automated wide-field microscope; the microscope must be equipped with a motor stage supplied with controls to image individual areas of the plate automatically.

1. Check the autofocus settings in the corner areas of the plate (wells A1, A24, P24, P1) and in the center.
   NOTE: For imaging, it is recommended to use a 20x short working distance objective (see Table of Materials) with the highest possible numerical aperture.
2. Based on the intensity histograms generated by the imaging software in live view mode, select a sufficiently long exposure time for the individual fluorescence channels so that the resulting image is not oversaturated.
3. Set the appropriate number of planes for imaging on the z-axis.
   NOTE: The field of view at 20x magnification is large enough that not all cells are in the same plane of focus, so z-sectioning must be performed to view all the cells in their correct plane of focus. Usually, five planes are enough. Depending on the availability of disk space, individual z-stacks or only maximum intensity projections can be saved. The image analysis shown in the representative results section is based on maximum intensity projections.
4. Select the appropriate number of fields of view to display per well.
   NOTE: Depending on the confluence, approximately 60-300 cells can be imaged in one field of view. Typically, for HeLa cells with a confluence of 60%-90%, imaging five fields of view will allow one to analyze from 500 cells to over 1,000 cells.
5. Select the wells to be imaged, and start imaging.

11. Quantitative image analysis

NOTE: The quantitative analysis of acquired images can be performed using an open-source software such as Cell Profiler¹⁵. For the present study, the analysis was performed using the ScanR 3.0.0 software (see Table of Materials).

1. Start the analysis by performing background correction for all the images from all the fluorescence channels. Depending on the size of the analyzed objects, select the appropriate filter size: 80 pixels for cell nuclei and 4 pixels for BrdU and mtDNA spots.
2. Start segmenting the image by creating a mask of the main object based on the ratio of the intensity of the fluorescence signal to the background for the channel corresponding to the cell nuclei.
3. Create masks for the sub-objects representing the BrdU and mtDNA spots, respectively, using the fluorescence channels appropriate for the given structures.
   NOTE: Each sub-object is assigned to the closest main object (cell nucleus).
4. Set the parameters to be measured during the analysis, and measure the intensity of the individual pixels within each mask for all the fluorescence channels.
   NOTE: It is also necessary to calculate the number of sub-objects assigned to a given cell nucleus and the area of each sub-object.
5. Define the derived parameters to be calculated based on the obtained data. To obtain the total fluorescence intensities for the BrdU or mtDNA channel, sum up the intensities of all the sub-objects assigned to a given cell nucleus. To obtain the mean fluorescence intensity, divide the total fluorescence intensity by the sum of the area of the sub-objects assigned to a given cell nucleus.
   NOTE: To ensure that the created sub-object (BrdU or mtDNA spot) is actually located in the mitochondria, it is necessary to gate them based on the fluorescence intensity from the mitochondria tracking dye.
6. Perform further analysis of the data obtained with the ScanR software using the R 4.2.2 statistical Software¹⁶ along with the following packages: dplyr¹⁷, data.table¹⁸, ggplot2¹⁹, and ggpubr²⁰. This step is optional.

A scheme of the procedure for the high-throughput study of the dynamics of mtDNA synthesis and distribution is shown in Figure 1. The use of a multi-well plate format enables the simultaneous analysis of many different experimental conditions, such as the silencing of different genes using a siRNA library. The conditions used for the labeling of newly synthesized DNA molecules with BrdU allow for the detection of BrdU-labeled DNA in the mitochondria of HeLa cells (Figure 2A) but can also be used as starting conditions for setting up the assay for other cell types. The incubation time with BrdU should be selected for individual cell lines on the basis of a time-course experiment (Supplementary Figure 1). In the present study, HeLa cells were used, since the screening of siRNAs requires cells that can be transfected with high efficiency. Importantly, the signal obtained with anti-BrdU antibodies is specific, as it can only be observed in cells treated with BrdU (Figure 2). The specificities of anti-BrdU and anti-DNA labeling were further confirmed using rho0 cells lacking mitochondrial DNA²¹ (Supplementary Figure 2). As expected, in these cells, regardless of BrdU treatment, no signal was detected in the mitochondria for both the anti-BrdU or for anti-DNA antibodies (Figure 2B). Quantifying the fluorescent signal from the anti-BrdU antibodies indicated that rho0 cells treated with BrdU showed the same low level of fluorescence as the BrdU-untreated cells, while the signal for the parental lines A549 and HeLa was, on average, 50-fold higher (Figure 2C).

In order to show the usefulness of the method presented here in the study of mitochondrial DNA synthesis and distribution, an experiment was conducted in which HeLa cells were treated with 2′,3′-dideoxycytidine (ddC), an inhibitor of mtDNA synthesis²², and the expression of genes known to be essential for mtDNA replication was silenced with siRNA (Figure 3). The treatment of cells with ddC led to the complete inhibition of BrdU incorporation, while the siRNAs used had varied effects. The downregulation of the Twinkle helicase (TWNK)²³ led to the most potent inhibition of BrdU incorporation; a weaker effect was observed for the TFAM protein²⁴, while the silencing of the mitochondrial DNA polymerase gamma (POLG)²⁵ led to a moderate decrease in BrdU incorporation compared to cells treated with negative control siRNA (Figure 3A,B). The use of anti-DNA antibodies in the procedure allows for the monitoring of the mtDNA distribution in the cell under the given experimental conditions. The downregulation of TFAM led to drastic changes in the mtDNA distribution; fewer mtDNA spots were detected compared to control cells, but those that remained had a very high fluorescence intensity (Figure 3A). Quantifying the mean fluorescence signal from the anti-DNA antibody showed a several-fold increase upon TFAM silencing compared to the other conditions tested (Figure 3C).

The specificity of the imaging results was verified by measuring the mtDNA and gene expression levels with the help of quantitative real-time PCR (qPCR) (Figure 4). The method has been described in detail elsewhere²⁶. Treatment with ddC resulted in a very strong decrease in the levels of mtDNA; a weaker effect was observed in the case of TFAM and TWNK silencing, while the transfection of cells with POLG siRNA had a very modest effect on the mtDNA copy number (Figure 4A). Quantifying the TFAM and TWNK expression confirmed an efficient reduction in their expression to ~15%-20% of the control levels, in contrast to POLG, whose expression at the mRNA level was reduced to 30% (Figure 4B).

Figure 1: Schematic representation of the protocol steps. The scale bars for the microscopic images are 10 µm. Please click here to view a larger version of this figure.

Figure 2: Specific detection of BrdU incorporation into mtDNA by anti-BrdU antibodies. Sample images of (A) HeLa and (B) A549 and A549 rho0 cells subjected to immunofluorescence staining with anti-BrdU (green) and anti-DNA (red) antibodies. The cells were treated or not treated with the BrdU solution. Nuclear DNA (blue) was labeled with Hoechst, and the mitochondria (cyan) were visualized with a mitochondria tracking dye. A merged image of all four fluorescence channels is shown. The scale bar represents 10 µm. (C) The result of the quantification of the signal from the anti-BrdU antibodies. The analysis was carried out for four independent biological replicates; each time, 100 randomly selected cells were analyzed for each experimental condition (N = 400 cells). On average, 18 BrdU foci were detected per cell. The boxes represent the first and third quartiles of the interquartile range (IQR), the median is represented by a horizontal line, the whiskers define the minimum and maximum, and the outliers are defined as 1.5 x IQR. A Kruskal-Wallis statistical test was performed. Please click here to view a larger version of this figure.

Figure 3: Outcomes of the inhibition of mtDNA synthesis, including reducing the efficiency of BrdU incorporation and affecting the distribution of nucleoids. (A) Sample fluorescence images of HeLa cells treated for 72 h with ddC or siRNA, which downregulate the proteins involved in mtDNA replication. The cells were treated with BrdU for 16 h before fixation. The nuclear DNA (blue) was stained with Hoechst, anti-BrdU (green) and anti-DNA (red) antibodies were used, and the mitochondria were labeled with a mitochondria tracking dye. A merged image of all four fluorescence channels is shown. The scale bar represents 10 µm. (B,C) Quantification of the fluorescence signals from (B) anti-BrdU and (C) anti-DNA antibodies. The analysis was carried out for four independent biological replicates; each time, 650 randomly selected cells were analyzed for each experimental condition (N = 2,600 cells). On average, 18 BrdU foci and 46 mtDNA foci were detected per cell. The boxes represent the first and third quartiles of the interquartile range (IQR), the median is represented by a horizontal line, the whiskers define the minimum and maximum, and the outliers are defined as 1.5 x IQR. A Kruskal-Wallis statistical test was performed. Please click here to view a larger version of this figure.

Figure 4: Validation of the efficiency of the inhibition of mtDNA synthesis and of the effectiveness of the tested siRNAs. HeLa cells were treated for 72 h with ddC or the indicated siRNAs and then subjected to DNA and RNA isolation. (A) Results of the mtDNA level analysis using qPCR. (B) qPCR analysis of gene expression changes under the indicated conditions. The bars represent the mean values of four independent biological replicates; the error bars represent the SEM; an ANOVA statistical test was performed; a t-test was performed for pairwise comparison against Untr (untreated) samples. Please click here to view a larger version of this figure.

Supplementary Figure 1: Dynamics of BrdU incorporation into the mtDNA of HeLa cells. The results of a time-course experiment in which the cells were incubated with 20 µM BrdU for a given time are shown. The means of four independent replicates for each time point are shown; 900 randomly selected cells were analyzed in each replicate. The bars represent the means of four replicates; an ANOVA statistical test was performed. Please click here to download this File.

Supplementary Figure 2: Validation of the A549 rho0 cells. (A) Electrophoretic analysis of the qPCR products from amplifying DNA fragments of the mtDNA-encoded ND1 and nuclear DNA-encoded B2M genes. Total DNA from A549 or A549 rho0 cells was used as templates in the reaction. (B) Results of the mtDNA level analysis using qPCR. The bars represent the mean values of four independent biological replicates; the error bars represent the SEM; a t-test was performed. Please click here to download this File.

Historically, DNA labeling by BrdU incorporation and antibody detection has been used in nuclear DNA replication and cell cycle research^14,27,28. So far, all the protocols for detecting BrdU-labeled DNA have included a DNA denaturation step (acidic or thermal) or enzyme digestion (DNase or proteinase) to enable epitope exposure and facilitate antibody penetration. These protocols were developed for tightly packed nuclear DNA. However, the different organization of mtDNA enabled the development of a procedure without the denaturation step; this has simplified the procedure and made it more suitable for high-throughput applications. This approach has brought excellent results because omitting the denaturation step results in the detection of BrdU-labeled DNA only outside the nucleus (Figure 2 and Figure 3). Consequently, the strong signal from nuclear DNA, which is always present when a denaturation step is included and which may cause a severe problem by masking the signal from BrdU-labeled mtDNA²⁹, can be avoided with this protocol. In addition, the lack of harsh denaturation ensures better preservation of the cellular structures and does not affect the efficiency of staining by fluorescent stains such as Hoechst or mitochondria tracking dye (Figure 2 and Figure 3).

The dynamics of BrdU incorporation into mtDNA is cell-line dependent. Doing a time-course experiment on BrdU incorporation is recommended whenever using a new cell line. This study has established 16 h as an optimal BrdU incubation time for HeLa cell lines. However, due to the biological variation in HeLa lines from different laboratories³⁰, it is suggested to perform a BrdU time-course experiment on the particular HeLa variant one plans to work with.

The combination of anti-BrdU and anti-DNA labeling in this protocol allows for the monitoring of mtDNA synthesis and the simultaneous study of the mtDNA distribution in the cell. This can be seen very well in TFAM-silenced cells (Figure 3). Decreasing the level of TFAM leads to a reduction in mtDNA synthesis (a decreased total anti-BrdU signal) and to the clustering of mitochondrial nucleoids, which is manifested by a substantial increase in the mean mtDNA fluorescence signal (Figure 3). It should be noted that, in this case, the increase in the mean mtDNA fluorescence intensity is caused by the changed distribution of the mitochondrial nucleoids and does not reflect the upregulation of mtDNA levels; in fact, the mtDNA levels are reduced, as revealed by quantitative PCR analysis (Figure 4A). Surprising results were obtained for the cells transfected with POLG siRNA (Figure 3). One could expect that the transfection of cells with siRNA-targeting POLG, which is the replicative DNA polymerase in mitochondria, would significantly reduce BrdU incorporation into the mtDNA. However, in this study, the effect on BrdU incorporation was negligible (Figure 3), as further confirmed by measuring the mtDNA levels using qPCR (Figure 4A). The poor downregulation of POLG expression can explain this seemingly contradictory result upon the transfection of cells with the applied siRNA (Figure 4B). This highlights a potential disadvantage of using siRNA for functional studies and suggests that there is a need for gene silencing validation.

Overall, an assay for studying the dynamics of mtDNA metabolism in cultured cells has been developed that uses a multi-well plate format and immunofluorescence imaging to detect and quantify mtDNA replication, repair, and distribution. The method allows for the simultaneous study of many different experimental conditions. It can be used to investigate the effects of various inhibitors, cellular stresses, and gene silencing on mtDNA metabolism. While the assay has a high potential for use in studying mtDNA metabolism in various physiological and pathological contexts, it should be noted that the assay does not allow replication and repair-linked mtDNA synthesis to be distinguished. This should be considered when interpreting the results and planning subsequent functional experiments. Notably, the assay has further potential for development. For example, replication-inactive (or repair-inactive) mitochondrial nucleoids are not detected with anti-BrdU antibodies. Therefore, the application of anti-DNA antibodies enables the quantitation of the mtDNA state levels, the number of replication-silent nucleoids, and the distribution of nucleoids in the cell.