https://orcid.org/0000-0002-8359-8399
Abstract
Mitochondria are believed to have originated from the endosymbiosis of bacteria and they still contain their own genome, which is called mitochondrial DNA (mtDNA). Under fluorescence microscopy of cultured mammalian cells, mtDNA is observed as numerous tiny dot-like structures called mitochondrial nucleoids. In live-imaging, the morphology and distribution of nucleoids are changed dynamically, but the molecular details remain poorly understood. In this study, we constructed a custom siRNA library targeting 1,164 human mitochondria-related genes, and from live-imaging-based screening of HeLa cells, we identified that mitochondrial calcium uniporter (MCU), a pore-forming subunit of the mitochondrial Ca2+ channel, is involved in nucleoid morphology. We found that suppression of MCU by RNAi induced the formation of highly enlarged nucleoids as well as respiratory dysfunction and that the re-introduction of MCU or treatment with Ca2+ ionophore recovered the enlarged nucleoid morphology. These results suggest that mitochondrial Ca2+ uptake via MCU is associated with nucleoid morphology. The constructed siRNA library might be widely applied to analyze the roles of mitochondrial proteins in various cellular events, making it useful to understand the multifaceted functions of mitochondria in human cells.
Mitochondria are double-membrane organelles that produce cellular ATP via oxidative phosphorylation in eukaryotic cells. Mitochondria are believed to have originated from the endosymbiosis of bacteria and they still contain their own genome, which is called mitochondrial DNA (mtDNA). In mammalian cells such as HeLa cells, there are more than several hundred copies of mtDNA per each cell (1, 2,). mtDNA encodes 13 proteins of essential respiratory subunits, 22 tRNAs, and 2 ribosomal RNAs (2, 3,). Mutation in mtDNA is reported to cause mitochondrial diseases due to defective respiratory complex formation and reduced respiratory activity (4).
In the nucleus, histone proteins bind to genomic DNA to form tightly packaged chromatin structures. In contrast, mtDNA binds with several DNA-binding proteins such as mitochondrial transcription factor A (TFAM) (5,), which bends and loosely folds mtDNA to form mitochondrial nucleoid structures (6–8,). Under fluorescence microscopy, nucleoids in mammalian cells are observed as numerous tiny dot-like structures distributed throughout the mitochondria (9–11 ). Although TFAM was originally identified as a mitochondrial transcription factor (12, 13,), suppression of TFAM by RNA interference (RNAi) leads to both decreased mtDNA copy number and enlarged nucleoid morphology, suggesting that TFAM also has an important role for mtDNA stabilization and nucleoid morphology (14, 15,). Other mtDNA binding proteins such as DNA helicase Twinkle and AAA-ATPase ATAD3A are also reported to act on nucleoid morphology and stability (16–18), although the molecular details of nucleoid morphology including dynamic features, how nucleoids move, and change their morphology and pathophysiological roles remain poorly understood.
Mitochondrial morphology is maintained by the balance between membrane fusion and fission (19, 20,), and dynamin-related protein 1 (Drp1) mediates mitochondrial fission with the Drp1 receptors such as Mff, MiD49, and MiD51 (21, 22,). We previously reported that suppression of Drp1 induced not only mitochondrial elongation by defective mitochondrial fission but also nucleoid clustering (23,); furthermore, co-suppression of the mitochondrial fusion factors Mfn1 and Mfn2 recovered both mitochondrial and nucleoid morphology (24,), suggesting that nucleoid morphology is related to mitochondrial membrane dynamics. We also found that respiratory activity was reduced in Drp1 knockout HeLa cells, and further repression of Mfn1/2 recovered mitochondrial function, suggesting that nucleoid morphology is also related to mitochondrial respiratory functions (24,). We recently found that AAA-ATPase protein ATAD3A in the inner membrane (IM) plays an important role for nucleoid trafficking along mitochondria, and that ATAD3A repression in Drp1 knockout cells recovered not only the nucleoid dispersion but also rescued respiratory complex formation (18), suggesting that the proper morphology and distribution of mtDNA is important to maintain mitochondrial function.
Mitochondria act not only as the powerhouse of the cell but also as a major calcium ion (Ca2+) store, along with the endoplasmic reticulum. Ca2+ is incorporated into the mitochondrial matrix through several Ca2+ channels in the IM (25,). Evidence of mitochondrial Ca2+ uptake was found over 50 years ago (26, 27,), and MICU1 (28,) and mitochondrial calcium uniporter (MCU) (29, 30,) were recently identified as the regulatory subunit and a pore-forming subunit of the mitochondrial Ca2+ uniporter, respectively. It was also reported that the MCU complex, mitochondrial Ca2+ channel complex, is formed from several subunits, including MCU, EMRE, MICU1, MICU2, MICU3, MCUR1, and MCUb (25,). Although Ca2+ signaling is involved in the mitochondrial dynamics by regulating membrane fusion and fission (31–35,) as well as the regulation of the respiratory functions (36, 37), the relation between Ca2+ and mtDNA is not known.
In this study, we performed small interfering RNA (siRNA) screening to identify the regulatory factors of nucleoid morphology. First, we established a custom siRNA library of the 1,164 genes related to mitochondria, covering most of the mitochondrial proteins. Then, using this library, we identified that mitochondrial calcium uniporter MCU that is involved in nucleoid morphology. In MCU-knockdown cells, nucleoids were enlarged and respiratory activity was decreased, suggesting critical roles for MCU in mtDNA function. The Ca2+ ionophore rescued the dispersed nucleoids in the MCU-knockdown cells, suggesting that Ca2+ incorporated in the mitochondrial matrix is important for nucleoid morphology. These findings suggest a relationship among mitochondrial Ca2+ uptake, mtDNA distribution, and mitochondrial function in mammalian cells.
Materials and Methods
Cell culture and transfection
HeLa cells and HT1080 cells were grown in Dulbecco’s modified Eagle’s medium (Wako) supplemented with 10% fetal bovine serum (173012, Sigma-Aldrich) and 1% penicillin/streptomycin (100 U/ml penicillin and 100 μg/ml streptomycin; 09367–34, Nacalai Tesque). The cells were cultured in a CO2 incubator (5% CO2 and 95% air) at 37°C. The cells were transfected with plasmids using Lipofectamine 2000 (11668019, Thermo Fisher Scientific) or with siRNAs using Lipofectamine RNAiMax (13778150, Thermo Fisher Scientific). The siRNAs used in this study are shown in Supplementary Data S1 and Supplementary Data S2.
Antibodies
The following antibodies were used at the indicated dilutions. For immunoblotting: mouse monoclonal anti-β-actin antibody (1:1,000; A2228, Sigma-Aldrich), anti-Drp1 antibody (1:1,000; 611113, BD Biosciences), anti-α-Tubulin antibody (1:2,000; T9026, Sigma-Aldrich), mouse polyclonal anti-TFAM antibody (1:1,000; H00007019-B01P, Abnova), rabbit monoclonal anti-MCU antibody (1:1,000; 14997, Cell Signaling Technology), Total OXPHOS Rodent WB Antibody Cocktail (1:500; ab110413, Abcam), Goat anti-Mouse IgG (H + L) Cross-Adsorbed Secondary Antibody, HRP (1:2,000; G-21040, Thermo Fisher Scientific) and Goat anti-Rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody, HRP (1:2,000; G-21234, Thermo Fisher Scientific). For immunostaining: mouse IgM monoclonal anti-DNA antibody (1:100; 61014, PROGEN), mouse IgG1 monoclonal anti-cytochrome c antibody (1:500; 556432, BD Biosciences), anti-rabbit IgG polyclonal anti-FLAG antibody (1:500; F7425, Sigma-Aldrich), Alexa Fluor 488-anti-mouse IgM (1:1,000; A-21042, Thermo Fisher Scientific), Alexa Fluor 568-anti-mouse IgG1 (1:1,000; A-21124, Thermo Fisher Scientific), Alexa Fluor 568-anti-rabbit IgG (1:1,000; A-11036, Thermo Fisher Scientific), and Alexa Fluor 647-anti-mouse IgG1 (1:1,000; A-21240, Thermo Fisher Scientific).
Construction of siRNA library and siRNA screening via reverse transfection
A total of 1,164 genes were selected with reference to MitoCarta2.0 (38,) and related articles (22, 39–46,) (Supplementary Data S1). Three independent sequences of Silencer Select siRNAs targeting each gene were obtained from Thermo Fisher Scientific. Because only two sequences targeting C16orf91 and TSTD3 were available for a custom siRNA library, two siRNAs were used for these genes. The three siRNAs were mixed and plated on 96-well plates. The control siRNA sets are also shown in Supplementary Data S2. For siRNA screening, Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum was used as a culture medium. A total of 1,250 HeLa cells were added to the Lipofectamine RNAiMax–siRNA mixture on 96-well plates and incubated for 24 h at 37°C in a CO2 incubator. After that, the culture medium was changed to a fresh culture medium, and cells were further incubated for another 48 h. Cells were stained with MitoTracker Red CMXRos (M7512, Thermo Fisher Scientific) for 15 min followed by SYBR Green I (S7562, Thermo Fisher Scientific) for 5 min. After washing staining dyes, images were obtained using a BZ-X710 microscope (KEYENCE) with a ×20 objective lens. The screening results were quantified as follows using ImageJ (47,). Obtained images were processed using the Difference of Gaussians method (48,). First, the single-cell region was determined using cellpose (49). Next, nucleoid images were converted to 16-bit images, and the background was subtracted using the ‘Subtract Background’ command (hereafter, ‘subtracted images’). Next, the subtracted images were duplicated and one was processed with the ‘Gaussian blur’ command (hereafter, ‘Gaussian images’). To obtain the nucleoid signal, the Gaussian images were subtracted from the subtracted images (hereafter, ‘calculated images’). The calculated images were binarized using the ‘Threshold’ command followed by processing with the ‘Watershed’ command to distinguish two or more close nucleoids. Finally, the nucleoid number and nucleoid size in the single-cell region were analyzed using the ‘Analyze Particles’ command. The mean values of the number and maximum size of each cell in a well were normalized by the mean value of them from control cells, which were treated with both siRNA-Negative control and siRNA-Luciferase, in the same plate. The normalized values of the nucleoid number and maximum nucleoid size were plotted in Fig. 1C. Furthermore, nucleoid enlargement was judged visually from the obtained images.
Establishing a custom siRNA library to target mitochondria, and screening of a novel regulator for nucleoid morphology. (A) Summary of the siRNA library. Submitochondrial localizations and MitoPathways of each gene were summarized with reference to MitoCarta3.0 (51). Note that some genes have multiple MitoPathways. (B) Schematic image of the siRNA screening. (C) Summary of the first screening results. (D) Cells were treated with a mixture of three siRNAs targeting the indicated genes for 72 h, and the mitochondria and nucleoid morphologies in living HeLa cells were observed using a BZ-X710 fluorescence microscope.
Cell staining and observation
Prior to microscopy using living cells, mitochondria were stained with 100 nM MitoTracker Red CMXRos for 10 min and then at a final concentration of ×1/100,000 SYBR Green I for 5 min at 37°C in a CO2 incubator, after which the staining solution was washed out. All sample media were changed to culture media supplemented with 20 mM HEPES (15630080, Life Technologies) and the samples were incubated at 37°C for 30 min in a CO2 incubator before observation. Stained cells were observed using a DMi8 fluorescent microscope (Leica) with a ×100 objective lens, or an LSM710 confocal microscope (Carl Zeiss AG) with a ×63 objective lens. Images were processed using LAS X (Leica) or ZEN2010 (Carl Zeiss AG) software, respectively. Images from the DMi8 microscope were processed using THUNDER Computational Clearing, followed by deconvolution using LAS X.
Plasmid construction
ORF of human MCU (NM_138357.2) was amplified from HeLa cell cDNA. The sequences of these primers are listed in Supplementary Data S3. The PCR fragment was cloned into the BamHI and EcoRI sites of pcDNA3.1 containing FLAG-tag in the XhoI and XbaI sites.
Immunoblotting
The cells were lysed using SDS sample buffer (1% glycerol, 1% sodium dodecyl sulfate, 0.2 M Tris–HCl pH 6.8) and samples were prepared by the addition of a final concentration of 0.046% bromophenol blue (021–02911, Wako) and 7.58% β-mercaptoethanol (133–06864, Wako). Samples were separated by SDS-PAGE using 7.5 or 14% polyacrylamide gel and transferred onto a PVDF membrane (Immobilon-P Transfer Membrane; IPVH00010, Merck). After transfer, the membranes were incubated with a primary antibody followed by a horseradish peroxidase-conjugated secondary antibody. Chemiluminescent signal obtained using ECL Western Blotting Detection Reagent (RPN2106, Cytiva) or Immobilon Western Chemiluminescent HRP Substrate (WBKLS0500, Merck) was detected by FUSION SOLO 7S EDGE (Vilber) or LAS 4000 mini (GE). The expression levels of each protein were quantified using ImageJ as follows. The regions of each band were determined using the ‘Rectangle’ command, and the background signal was subtracted using the ‘Subtract background’ command. Each protein expression level was measured as the total signal intensity in each band region. The value of each band was normalized by the value of β-actin.
Immunostaining
The cells, grown on coverslips, were fixed with 4% paraformaldehyde (163–20145, Wako), then permeabilized with 0.2% Triton X-100 (169–21105, Wako) in phosphate-buffered saline (PBS). After washing with PBS, the cells were blocked with 5% skim milk in PBS, and incubated with primary antibodies for 1 h at room temperature. As a secondary antibody reaction, the cells were incubated with Alexa Fluor 488/568/647-conjugated secondary antibodies for 1 h at room temperature. Finally, the nuclei were stained with Hoechst 33258 (H3569, Thermo Fisher Scientific) and the coverslips were mounted using SlowFade Diamond antifade reagent (S36972, Thermo Fisher Scientific). The samples were observed under an LSM710 confocal microscope with a ×63 objective lens.
Nucleoid quantification
Obtained images from a DMi8 fluorescent microscope or an LSM710 confocal microscope were processed using the Difference of Gaussian method (48,) using ImageJ (47,) as follows. First, we determined the mitochondrial region of interest (ROI) in a single cell. The single-cell region was determined using cellpose (49) and modified manually using ImageJ. Prior to determining the mitochondrial ROI in the single cells, mitochondrial images were converted to 16-bit images, and the background signal was subtracted by the ‘Subtract’ command. After that, the single-cell region was set to processed mitochondrial images using the ‘ROI manager’, and the area outside of the single-cell region were processed with the ‘Clear Outside’ command. The mitochondrial ROI in the cell was obtained by processing with the ‘Threshold’ command followed by the ‘Analyze Particles’ command. Each mitochondrial region was combined using the ‘Combine’ command (hereafter ‘mitochondrial ROI’). Next, nucleoid morphology in the mitochondrial ROI was quantified. Nucleoid images were converted to 16-bit images, and the background was subtracted using the ‘Subtract Background’ command and the ‘Subtract’ command (hereafter, ‘subtracted images’). After that, the subtracted images were duplicated, and one was processed with the ‘Gaussian blur’ command (hereafter, ‘Gaussian images’). To obtain the nucleoid signal, the Gaussian images were subtracted from the calculated images (hereafter, ‘calculated images’). The calculated images were binarized using the ‘Threshold’ command and processed with the ‘Fill Holes’ command to fill the nucleoid area. To remove noise, processed calculated images were processed with the ‘Erode’ and ‘Dilate’ commands, followed by processing with the ‘Watershed’ command to distinguish two or more close nucleoids. Finally, the nucleoid size and number in the mitochondrial ROI were analyzed using the ‘Analyze Particles’ command. The nucleoid number and maximal nucleoid size in each cell were analyzed.
Measurement of respiratory activity
The cells, plated on a cell culture microplate for Seahorse XFe24 (Agilent), were washed three times and incubated in Seahorse XF Dulbecco’s modified Eagle’s medium pH 7.4 (Agilent) supplemented with 10 mM glucose, 1 mM pyruvate, and 2 mM glutamine at 37°C for 1 h. Respiratory activity was measured with a Mito Stress Test (103015–100, Agilent) using a Seahorse XFe24 (Agilent). After the measurement, the cells were stripped with trypsin and the cell number was counted using Cell Drop BF (DeNovix) to normalize the obtained data.
Measurement of mitochondrial Ca2+ uptake
MCU-knockdown cells were transfected with matrix-targeting Cameleon (4mtD3cpv; plasmid #36324, Addgene) (50) and observed using a BZ-X710 microscope with a ×60 objective lens and two filter sets, CFP filter (ET436/20x, ET480/40m, and T455lp; CHROMA) and FRET-YFP filter (ET436/20x, ET535/30m, and T455lp; CHROMA). The cells were observed every 10 s for 370 s. Between the sixth and seventh observations, 100 μM histamine (18112–61, Nacalai) was added. Obtained images were processed using ImageJ as follows. The backgrounds were subtracted using the ‘Subtract background’ and ‘Subtract’ commands. After image calculation using the ‘Image Calculator’ command (equation = images of FRET-YFP/images of CFP), FRET intensity in each cell region was measured and normalized by the mean FRET intensity before histamine stimulation. The maximum responses were calculated from the maximum value of the relative FRET intensity after histamine stimulation.
A23187 treatment
Cells were stained with MitoTracker Red and SYBR Green I, followed by treatment with or without 2 μM A23187 (019–20111, Wako) for 1 h. Then, the cells were observed using a DMi8 microscope with a ×100 objective lens.
Quantification of mtDNA copy number
DNA from HeLa cells was extracted using a Puregene Cell Lysis Solution (158906, QIAGEN) and Puregene Protein Precipitation Sol. (158123, QIAGEN) according to manufacturer’s protocols. THUNDERBIRD Next SYBR qPCR Mix (QPX-201, TOYOBO) was used for quantitative PCR using the StepOnePlus Real-Time PCR System (Applied Biosystems). To produce a standard curve, 0.5, 1.0, 2.0, 4.0, and 8.0 ng of DNA (for mtDNA amplification) or 1.0, 2.0, 4.0, 8.0, and 16.0 ng of DNA (for nuclear DNA amplification) from control HeLa cells were used as described previously (23). The primer sets for amplification of mtDNA and β2M coding nuclear DNA are shown in Supplementary Data S3.
Data analysis, statistical analysis‚ and drawing graphs
The obtained data were processed using Microsoft Excel (Microsoft). For statistical analysis, analysis of variance and Tukey–Kramer post hoc test were performed using Python 3.9 (Spyder 5) for the siRNA screening or Statistics Kingdom (https://www.statskingdom.com/) for other analysis. Graphs were drawn using Microsoft Excel. All graphs show plots of the raw data and the mean ± standard deviation (SD).
Results
Constructing a custom siRNA library to target mitochondria-related genes
To explore the regulatory factors of nucleoid morphology, we established a novel siRNA library to target human mitochondria-related proteins encoded in the nuclear genome. To construct our custom library, 1,164 genes were selected with reference to MitoCarta2.0 (38,) and various articles on mitochondria-related factors (22, 39–46,). The submitochondrial localization and MitoPathways (51) of these genes are summarized in Fig. 1A. We prepared three independent siRNAs to target each gene.
To perform RNAi knockdown in multi-well plates, first, reverse transfection was performed by adding trypsinized HeLa cells (1,250 cells per well) to wells containing a mixture of siRNAs and lipofection reagent, followed by culturing for 72 h to repress the target genes. Then, mitochondria and nucleoids were stained with MitoTracker Red and SYBR Green I, respectively, and were observed using a fluorescence microscope (Fig. 1B).
MCU is involved in nucleoid morphology
Our first siRNA screening was performed with mixed siRNAs, in which three independent siRNAs repressing one target gene were mixed to treat HeLa cells. From the first screening, we identified MCU as a gene that is involved in nucleoid morphology, given that knockdown induced highly enlarged nucleoids (Fig. 1C and D). Although many other candidate genes have been found, their reproducibility has not yet been fully confirmed, and thus in this report, we show the data of MCU for which we were able to confirm reproducibility, as shown below.
When we performed RNAi using three independent siRNAs targeting MCU, nucleoids were highly enlarged by all three siRNAs (Fig. 2A). When we examined the protein levels of MCU by immunoblotting, MCU was only rarely observed by RNAi using all three siRNAs (Fig. 2B), suggesting that MCU protein was efficiently repressed by these siRNAs. Quantitative analysis of nucleoid morphology showed that the size of the nucleoids was increased and that the number of nucleoids was reduced in the MCU-knockdown cells (Fig. 2C and D). These results further confirmed that MCU is involved in nucleoid morphology. Nucleoid morphology was further analyzed by immunostaining because our previous finding shows that small nucleoids clusters to form enlarged nucleoids in mitochondrial fission factor Drp1-repressed cells (Fig. 2F and Supplementary Fig. S1A) (23). In the present study, we found that nucleoids in MCU-repressed cells were also clustered (Fig. 2F and Supplementary Fig. S1A). Furthermore, the re-introduction of MCU recovered the nucleoid morphology in the MCU-deficient cells (Fig. 2G–I and Supplementary Fig. S1B). We also examined the effect of MCU knockdown in the other human cell line, HT1080 cells. In HT1080 cells, nucleoids were also enlarged by suppression of MCU (Fig. 3A–D). From this result, we could confirm that MCU knockdown caused nucleoid enlargement in multiple cell lines. These data suggested that MCU is involved in nucleoid morphology and distribution.
MCU is involved in nucleoid morphology and distribution. (A) Cells were treated with siRNAs for 96 h and then observed using an LSM710 confocal microscope. Representative images are shown. The white dashed square indicates the magnified area. (B) Protein levels in the MCU-knockdown cells were determined by immunoblotting using the indicated antibodies. (C, D) The maximum nucleoid size (C) and nucleoid number (D) in each cell of (A) were quantified using ImageJ. (E) The relative mtDNA copy number in MCU-knockdown cells was measured using quantitative PCR. (F) Cells were treated with indicated siRNA for 96 h and then fixed. After that nucleoids were stained with an anti-DNA antibody, indicating nucleoid, and observed under an LMS710 confocal microscope. The representative images are shown. The overlay images of nucleoids with mitochondria are shown in Supplementary Fig. S1A. (G–I) Cells were treated with the indicated siRNA for 96 h and then transfected with MCU-FLAG. Subsequently, cells were fixed and co-stained with an anti-DNA antibody, an anti-cytochrome c antibody, and an anti-FLAG antibody, indicating the nucleoid, mitochondria, and MCU-FLAG, respectively. Representative images obtained from an LMS710 confocal microscope are shown (G). The nucleoid images and overview images are shown in Supplementary Fig. S1B. The maximum nucleoid size (H) and nucleoid number (I) in each cell were quantified as in (C, D). All graphs show plots of the raw data and the mean ± SD. n ≥ 3 for each quantitative dataset. All statistical analyses were performed using Tukey–Kramer test. *P < 0.05, **P < 0.01.
MCU suppression caused nucleoid enlargement in HT1080 cells. (A) HT1080 cells were treated with indicated siRNA for 96 h and then observed using a DMi8 fluorescence microscope. Representative images were shown. The white dashed square indicates the magnified area. (B) Protein levels in the MCU-knockdown HT1080 cells were determined by immunoblotting using the indicated antibodies. (C, D) The maximum nucleoid size (C) and nucleoid number (D) in each cell of (A) were quantified using ImageJ. All graphs show plots of the raw data and the mean ± SD. n ≥ 3 for each quantitative dataset. All statistical analyses were performed using the Tukey–Kramer test. **P < 0.01.
Respiratory activity was decreased in MCU-knockdown HeLa cells
Next, we analyzed the amount of mitochondrial DNA and mitochondrial function. When we measured the copy number of mtDNA, we did not find a significant change caused by the suppression of MCU (Fig. 2E). We further confirmed that expression levels of Drp1 and TFAM were not altered in MCU-knockdown cells (Fig. 2B). Next, we used an extracellular flux analyzer to measure oxygen consumption rates and found that both basal and maximal respiratory activities were partially but significantly decreased in MCU-knockdown cells (Fig. 4A–C). The reduction of respiratory subunits in complex I (NDUFB8) and complex IV (MTCO1) were also observed by immunoblotting (Fig. 4D–J). These data suggested that respiratory activity is altered in MCU-knockdown cells as seen in liver-specific MCU KO hepatocytes (52).
Mitochondrial respiratory activity was diminished in MCU-knockdown cells. (A–C) Mitochondrial respiratory activities in MCU-knockdown cells were measured using a Seahorse XFe24 extracellular flux analyzer. The trace shows the time-dependent change in the oxygen consumption rate. Graph shows the mean ± SD (A). Bar graphs show basal (B) and maximal (C) respiratory activity. These bar graphs show plots of the raw data and the mean ± SD. (D–J) The protein levels of several respiratory subunits in the MCU-knockdown cells were determined by immunoblotting (D). The protein levels relative to β-actin were shown in (E–J). n ≥ 3 for Fig. 4A–C, and n = 2 for Fig. 4E–J. All statistical analyses were performed using the Tukey–Kramer test. *P < 0.05, **P < 0.01.
Ca2+ incorporation across the mitochondrial inner membrane is critical for the regulation of nucleoid morphology
We further examined whether Ca2+ incorporation into mitochondria directly affects the nucleoid morphology. Given that MCU is the pore-forming subunit of the MCU complex and critical for mitochondrial Ca2+ uptake, the suppression of MCU was reported to reduce the Ca2+ uptake capacity into mitochondria (29, 30,). When we measured mitochondrial Ca2+ uptake after histamine stimulation using a matrix-targeting Ca2+ sensor, matrix-targeting Cameleon (4mtD3cpv) (50), the increasing rate of FRET intensity after histamine stimulation showing Ca2+ uptake into the matrix was decreased in MCU-knockdown cells compared to control cells (Fig. 5A and B). From this result, we confirmed that MCU suppression decreased transient Ca2+ uptake into mitochondria, as reported previously.
Mitochondrial Ca 2+ uptake was suppressed in MCU-knockdown cells. HeLa cells were treated with an indicated siRNA for 96 h and then transfected with matrix-targeting Cameleon (4mtD3cpv). After that, mitochondrial Ca2+ uptake after histamine stimulation was observed under a BZ-X710 fluorescence microscope in each 10 s for 370 s. After the sixth observation, 100 μM histamine was added. The trace shows the time-dependent change of the FRET intensity relative to the mean FRET intensity before histamine stimulation (Fbasal), showing the mean ± SD (A). The dot plot shows the maximum value of (A), showing plots of the raw data and the mean ± SD (B). n ≥ 3 for each quantitative dataset. All statistical analyses were performed using the Tukey–Kramer test. *P < 0.05, **P < 0.01.
To analyze the effects of mitochondrial Ca2+ on nucleoid morphology, MCU-knockdown cells were treated with Ca2+ ionophore A23187 to mediate Ca2+ incorporation into mitochondria. After 1-h treatment with A23187, the enlarged nucleoids in MCU-knockdown cells dispersed to form small nucleoids, as seen in the control cells (Fig. 6A–C and Supplementary Fig. S1C). These results indicate that mitochondrial Ca2+ incorporated via MCU directly affects the nucleoid morphology.
Ca 2+ ionophore A23187 repressed nucleoid enlargement in MCU-knockdown cells. MCU-knockdown cells were treated with 2 μM A23187 for 1 h and then nucleoid morphology was observed using a DMi8 fluorescence microscope. Representative images are shown (A). The white dashed square indicates the magnified area. The overlay images of nucleoids with mitochondria are shown in Supplementary Fig. S1C. The maximum nucleoid size (B) and nucleoid number (C) in each cell were quantified using ImageJ. All graphs show plots of the raw data and the mean ± SD. n ≥ 3 for each quantitative dataset. All statistical analyses were performed using the Tukey–Kramer test. *P < 0.05, **P < 0.01.
Discussion
In this study, we constructed a custom siRNA library to target mitochondrial proteins in human cells and used it to perform siRNA screening aimed at exploring novel nucleoid regulators. We found that the suppression of MCU, which is an essential core component of mitochondrial Ca2+ uniporter, caused nucleoid enlargement and reduced respiratory activity. The morphological change in nucleoids resulting from MCU deficiency was repressed by further Ca2+ ionophore treatment. These data clearly showed that mitochondrial Ca2+ uptake by MCU is involved in nucleoid morphology and distribution (see model in Fig. 7).
Schematic illustration of mitochondrial Ca 2+ uptake by the MCU complex to modulate mitochondrial DNA nucleoid morphology and distribution.
We constructed a novel siRNA library that should be widely applicable to analyze the roles of mitochondria in various cellular events, making it highly useful for understanding the multifaceted functions of mitochondria. For siRNA screening, we used a mixture of three siRNAs for each gene; however, the reproducibility of the data was often problematic, and the large number of false positive and false negative results made subsequent detailed analysis difficult to perform. Therefore, the overall screening results using this library are not presented in this paper, and further detailed screening using the three individual siRNAs is ongoing. During this first screening, we identified MCU as a factor modulating nucleoid morphology, which provided us the opportunity to analyze a new function of MCU in detail.
We previously reported that mitochondrial respiratory activity is reduced with nucleoid clustering (24,) and that nucleoid dispersion by suppression of ATAD3A recovers respiratory complex formation (18). In the present study, we found that the suppression of MCU induces enlarged nucleoids and reduced respiratory capacity, consistent with our previous conclusion that nucleoid morphology and distribution are associated with mitochondrial function.
Ca2+ incorporation into mitochondria was reported to be important for the regulation of membrane permeability transition (53, 54,), respiratory function (36, 37,), and fission of IM (33,). Furthermore, MICU1, a subunit of the MCU complex, was reported to be important for cristae formation (55,). We previously demonstrated that nucleoid clustering affects cristae structure (23), leading us to speculate the possibility that Ca2+ regulates the cristae structure of IM via MICU1, which may be involved in nucleoid morphology, although further detailed analysis is needed to elucidate how Ca2+ regulates the morphology and distribution of the mitochondrial membrane and genome.
In this article, we showed that the suppression of MCU reduced respiratory capacity (Fig. 4A–C), consistent with a previous finding showing a reduction of respiratory activity in MCU KO hepatocytes (52,). However, several reports showed that respiratory activity was not affected in MCU-knockdown HeLa cells by short hairpin RNA (30,) and MCU−/− MEF (56). These results suggest that the relationship between MCU and respiratory activity is controversial and remains to be analyzed.
In our previous study, we found that nucleoid clustering, caused by suppression of Drp1, decreased the expression of the respiratory subunits (24,). Furthermore, nucleoid dispersion by ATAD3A knockdown recovered the expression of the respiratory subunits (18,). These results suggest that nucleoid morphology should affect the formation of respiratory complexes. In this study, we showed that MCU suppression by RNAi caused nucleoid enlargement and decreased the subunit of complex I and complex IV, consistent with our previous studies (18, 24). However, the mechanisms of how nucleoid morphology affects respiratory complex formation remain to be understood.
The regulatory mechanisms of nucleoid morphology, distribution and dynamics observed in live-imaging remain largely unknown, although we previously found that nucleoid morphology and distribution are affected by mitochondrial fission factors. Nucleoids are also considered to be related to liquid–liquid phase separation, given that TFAM, a major component of nucleoids, was reported to have an ability to induce liquid droplet formation (57, 58,). In this study, we showed that treatment with Ca2+ ionophore A23187 rescued the enlarged nucleoids in MCU-deficient cells (Fig. 6). The concentration of Ca2+ ions is thought to affect the efficiency of droplet formation by liquid–liquid phase separation (59,), assuming that the Ca2+ ion status in the matrix modulated by MCU affects TFAM droplet formation and the nucleoid morphology. Although MCU was reported to interact with several proteins related to mtDNA maintenance and transcription, such as ATAD3A, TFAM‚ and POLG2 (60, 61), here, we showed that Ca2+ should have critical roles in the regulation of nucleoid morphology (Fig. 6). We need further analysis to understand the molecular details of the nucleoid morphogenesis.
Screening using the custom siRNA library established in this study is expected to help identify related factors and to elucidate the regulatory mechanism of mitochondrial nucleoids, which may lead to the development of new technologies for modulating mitochondrial activity as well as new methods for activating mitochondrial respiratory activity in the near future.
Supplementary Data
Supplementary Data are available at JB Online.
Acknowledgments
We thank Dr. Keisuke Takeda (Osaka University) for assistance with image quantification, and Michael Tiger Chang (UCLA) for technical support. We also thank the members of the Ishihara lab for insightful discussions.
Funding
This work was supported by Grant-in-Aid for Japan Society for the Promotion of Science Fellows grant number 22KJ2068 (H.K.); Japan Society for the Promotion of Science KAKENHI grant numbers 24H02276 and 24H00558; AMED-CREST grant number JP23gm1110006; the Daiichi Sankyo Foundation of Life Science; the Naito Science & Engineering Foundation; and the Uehara Memorial Foundation (N.I.); AMED BINDS grant number JP24ama121032 (K.T.)
Conflict of Interest
None declared.
References
Author notes
These authors contributed equally to this work.
