Silencing of Poly(ADP-Ribose) Polymerase-2 Induces Mitochondrial Reactive Species Production and Mitochondrial Fragmentation

PARP2 is a DNA repair protein. The deletion of PARP2 induces mitochondrial biogenesis and mitochondrial activity by increasing NAD+ levels and inducing SIRT1 activity. We show that the silencing of PARP2 causes mitochondrial fragmentation in myoblasts. We assessed multiple pathways that can lead to mitochondrial fragmentation and ruled out the involvement of mitophagy, the fusion–fission machinery, SIRT1, and mitochondrial unfolded protein response. Nevertheless, mitochondrial fragmentation was reversed by treatment with strong reductants, such as reduced glutathione (GSH), N-acetyl-cysteine (NAC), and a mitochondria-specific antioxidant MitoTEMPO. The effect of MitoTEMPO on mitochondrial morphology indicates the production of reactive oxygen species of mitochondrial origin. Elimination of reactive oxygen species reversed mitochondrial fragmentation in PARP2-silenced cells.


Introduction
Optimal mitochondrial biogenesis requires the coordinated expression of nuclear and mitochondrial genes. Most mitochondrial genes are coded in the nucleus. There is a dominant flow of information and material from the nucleus and the cytosol to the mitochondria, termed anterograde signaling [1]. The appropriate flux of protein from the cytosol to the mitochondria requires the coordinated regulation of nuclear transcription by a large set of nuclear transcription factors [2]. This web of transcription factors responds to changes in nutrient levels (amino acids, oxygen, NAD + , etc.) and cellular energy charge. Therefore, the transcriptional programs can accommodate the actual nutrient and energy charge of cells [3][4][5][6]. The disruption of nutrient sensing and mitonuclear proteostasis plays a pathogenic role in metabolic and neoplastic diseases [7][8][9].
An increase in the mitochondrial content of cells is termed mitochondrial biogenesis [2,10]. Mitochondria undergo constant quality control [10][11][12][13][14][15], in which, among others, mitophagy [16][17][18] and the mitochondrial unfolded protein response (mtUPR) [19,20] µg/mL of puromycin for HepG2 cells. Cells that harbor the plasmid expressing the small hairpin RNA against PARP2 (i.e., PARP2 silenced) are referred to as shPARP2, while cells that harbor the plasmid expressing a small hairpin RNA that has no target in human or murine cells (controls) are referred to as scPARP2. Both cell lines were established earlier [54,55], and the effectiveness of silencing was verified before experimentation ( Figure  1A,B). or shPARP2 C2C12 cells were seeded into 6-well plates, and the expression of PARP2 was determined by RT-qPCR (n = 3) and Western blotting (n = 3). (B) A total of 250,000 scPARP2 or shPARP2 HepG2 cells were seeded into 6-well plates, and the expression of PARP2 was determined by RT-qPCR (n = 3) and Western blotting (n = 3). Representative Western blot images are presented. Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, two-tailed Student's t-test. *** Statistically significant differences between scPARP2 and shPARP2 C2C12 cells or scPARP2 and shPARP2 HepG2 cells at p < 0.001.

In Vitro Cell Proliferation Assay (SRB Assay)
Cellular proliferation was determined using Sulphorhodamine B (SRB) assay, as described in [68]. ,000 scPARP2 or shPARP2 C2C12 cells were seeded into 6-well plates, and the expression of PARP2 was determined by RT-qPCR (n = 3) and Western blotting (n = 3). (B) A total of 250,000 scPARP2 or shPARP2 HepG2 cells were seeded into 6-well plates, and the expression of PARP2 was determined by RT-qPCR (n = 3) and Western blotting (n = 3). Representative Western blot images are presented. Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, two-tailed Student's t-test. *** Statistically significant differences between scPARP2 and shPARP2 C2C12 cells or scPARP2 and shPARP2 HepG2 cells at p < 0.001.

In Vitro Cell Proliferation Assay (SRB Assay)
Cellular proliferation was determined using Sulphorhodamine B (SRB) assay, as described in [68].

Detection of Cell Death
To evaluate changes in apoptotic and necrotic cell death, an FITC Annexin V/Dead Cell Apoptosis Kit (cat. no. V13242, Invitrogen) was used according to the manufacturer's instructions. Cells were seeded into 6-well plates and treated with the different chemicals, as stated. Then, cells were collected and stained with 5 µL of FITC annexin V and 100 µg/mL of PI for 15 min at room temperature. Cells were analyzed by flow cytometry (FACS Calibur, Becton Dickinson Biosciences, San Jose, CA, USA), and data were analyzed using BD CellQuest Pro software v5.2 (Becton Dickinson Biosciences).

Determination of Cellular ATP Level
Cells were seeded into 6-well plates and treated, as indicated. ATP levels were determined using an ATP Assay Kit (cat. no. MAK190, Sigma-Aldrich, St. Louis, MO, USA). ATP concentration was measured in 96-well black plates using a fluorimeter (Spark 20M, Tecan Life Sciences, Männedorf, Switzerland). ATP levels were normalized to protein content, and normalized readings are presented.

MitoTracker Red Staining
Cells grown on glass coverslips were treated with the specified chemicals, as indicated, or transfected with the indicated siRNAs (see Figure legend). Mitochondria were stained with MitoTracker Red, as described in [69]. Confocal images were acquired with a Leica TCS SP8 confocal microscope (Leica, Wetzlar, Germany) and LAS X software v3. 5.5.19976 (Leica). Processed images were analyzed using ImageJ v1.44 software with Mito-Morphology Macro [70], yielding the mitochondrial number, content, circularity, and form factor. Form factor is derived from the area-to-perimeter ratio [71]; hence its decrease would signify fragmentation. Circularity increases if the shape of an object is closer to a circle; hence, increased circularity suggests that the mitochondria are not elongated, which is a feature of a disassembled mitochondrial network. Perimeter is a similar term; a decrease in perimeter indicates smaller mitochondria. For co-localization analysis and the assessment of the Pearson correlation coefficient, ImageJ software with EzColocalization plug-in was used [72].

Immunofluorescence
Immunofluorescence was described in [58]. Antibodies used in immunofluorescence are listed in Table 1.

Total RNA Preparation and Reverse Transcription-Quantitative PCR (RT-qPCR)
Total RNA preparation, reverse transcription, and qPCR were performed, as described in [58]. Expression was normalized to the geometric mean of murine 36B4 and cyclophilin values. Primers are listed in Table 2.

Determination of Lipid Peroxidation (TBARS Assay)
Lipid peroxidation was assessed by determining the production rate of thiobarbituric acid-reactive substances (TBARS) similar to [74].

Determination of the Oxygen Consumption Rate (OCR)
The oxygen consumption rate was determined using the XF96 Flux Analyzer (Agilent Technologies, Santa Clara, CA, USA). Cells were seeded into 96-well assay plates and treated with the indicated concentration of different chemicals for 48 h (see Figure legend). After recording the baseline, the OCR was recorded every 5 min, and 50 µM etomoxir (cat. no. E1905, Sigma-Aldrich) was used for determining the fatty acid oxidation. Data were normalized to protein content, and normalized readings are displayed.

Electron Microscopy
Data from a previous study [58] were re-analyzed. All relevant methodology can be found in [58].

Statistical Analysis
Statistical analyses were performed using GraphPad Prism v8 software (GraphPad Software, San Diego, CA, USA). The number of repetitions for each experiment is indicated in the figure legends. All groups were checked for normal distribution. For comparing two groups, two-sided Student's t-tests were applied. For multiple comparisons, oneway ANOVA was conducted, followed by Tukey's post hoc test (to compare all possible combinations) or Dunnett's post hoc test (to compare data to one selected group). The levels of significance are indicated in figure captions; n denotes the number of biological replicates.

Silencing of PARP2 Leads to Fragmented Mitochondria
We assessed the mitochondrial morphology in scPARP2 and shPARP2 C2C12 myoblasts. Silencing of PARP2 induced the mitochondrial content in cells (a readout called "Mito Content"), as visualized by Mitotracker Red (Figure 2A) and TOMM20 immunostaining ( Figure 2B), in good agreement with previous observations [27,55,57]. In addition, the silencing of PARP2 resulted in fragmentation of the mitochondrial network, marked by increased circularity and decreased individual mitochondria perimeters and form factors (Figure 2A,B). We re-analyzed electron microscopy sections from a previous study [58] and found that the number of mitochondrial cross sections increase in cells, which indicates the  Figure 2C). Acute silencing of PARP2 by siRNA in C2C12 cells led to the induction of mitochondrial content and fragmentation of the mitochondrial network, similar to the findings in the established cell line ( Figure 3A-C). Pharmacological inhibition of PARP1 by olaparib or nicotinamide was shown to induce mitochondrial fragmentation [20], similar to boosting in cellular NAD + levels [75][76][77][78][79][80] that aligns well with our observations. Numeric values are presented as the average ± SEM in (A) and (B) and as the average ± SD in (C). Statistical significance was determined using paired, two-tailed Student's t-test in (A-C). *, **, and *** Statistically significant differences between the indicated groups at p < 0.05, p < 0.01, and p < 0.001, respectively. Numeric values are presented as the average ± SEM in (A,B) and as the average ± SD in (C). Statistical significance was determined using paired, two-tailed Student's t-test in (A-C). *, **, and *** Statistically significant differences between the indicated groups at p < 0.05, p < 0.01, and p < 0.001, respectively.

SIRT1 Activation, Mitophagy, mtUPR, and the Deregulation of the Mitochondrial Fusion-Fission System Are Not Involved in Mitochondrial Fragmentation in PARP2-Silenced Cells
Multiple parallel pathways can govern the fragmentation of the mitochondrial network. Therefore, we assessed these possibilities. The silencing of PARP2 blocks autophagy [58], which may affect mitophagy, leading to mitochondrial fragmentation. We assessed the co-localization of LC3, a marker of autophagic vesicles, and TOMM20, a mitochondrial marker; however, there was no overlap between the LC3+ (a marker for autophagosomes) and MitoTracker Red+ or TOMM20+ (both mitochondrial markers) features ( Figure 4A). For a positive control demonstrating overlap of staining, C2C12 cells were co-stained with Tomm20 and MitoTracker Red (Supplementary Figure S1). Thus, PARP2 silencing did not increase mitophagy. SIRT1 activation is associated with PARP2 silencing [27,55,57], and SIRT1 regulates mitochondrial morphology [75,81]. However, genetic silencing of SIRT1 in shPARP2 cells did not reverse the changes to mitochondrial morphology ( Figure 4B). We also checked the involvement of mtUPR, which was shown to play a role in mitochondrial biogenesis induced by pharmacological PARP inhibition [82]. However, no consistent induction in the expression of key heat shock proteins (Hsp25, Hsp40, Hsp60, Hsp70, Hsp90, Hsp110) were detected when scPARP2 and shPARP2 C2C12 cells were compared under control conditions or after heat shock (Supplementary Figure S2). These pathways are, therefore, unlikely to contribute to mitochondrial fragmentation. We observed a mild increase in the mRNA and protein expression of OPA1, Parkin, and Pink1 ( Figure 5A,B).  Statistical significance was determined using ANOVA followed by Dunnett's post hoc test. *, **, and *** Statistically significant differences between the negative control and transiently transfected samples at p < 0.05, p < 0.01, and p < 0.001, respectively. In the same cells, mitochondrial structure was investigated using MitoTracker Red staining (n = 3) (B). Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20 from each group). Representative Western blot and immunofluorescence images are presented. Numerical values are presented as the average ± SEM. Statistical significance was determined using ANOVA followed by Dunnett's post hoc test. *, **, and *** Statistically significant differences between the negative control and transiently transfected samples at p < 0.05, p < 0.01, and p < 0.001, respectively.  . Mitochondrial fragmentation in shPARP2 C2C12 is not due to mitophagy or SIRT1 activation. (A) A total of 70,000 scPARP2 or shPARP2 C2C12 cells were seeded into 24-well plates on glass coverslips, and LC3-MitoTracker Red-DAPI or LC3-TOMM20-DAPI co-immunofluorescence was performed (n = 3). Co-localization analysis was performed using ImageJ software with the EzColocalization plug-in (measured cells: 100/100). (B) A total of 30,000 scPARP2 or shPARP2 C2C12 cells were seeded into 24-well plates, and SIRT1 was transiently silenced. After transfection, the expression of SIRT1 was determined using RT-qPCR (n = 3) and Western blotting (n = 3). The mitochondrial structure was investigated using MitoTracker Red staining (n = 3). Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20 from each group). Representative immunofluorescence and Western blot images are presented in the figure. Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, two-tailed Student's t-test in (A), while in (B), ANOVA followed by Tukey's post hoc test was used. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05, p < 0.01, and p < 0.001, respectively. ## and ### Statistically significant differences between the negative control and transiently transfected samples at p < 0.01 and p < 0.001, respectively.
sion of SIRT1 was determined using RT-qPCR (n = 3) and Western blotting (n = 3). The mitochondrial structure was inves-tigated using MitoTracker Red staining (n = 3). Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20 from each group). Representative immunofluorescence and Western blot images are presented in the figure. Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, two-tailed Student's t-test in (A), while in (B), ANOVA followed by Tukey's post hoc test was used. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05, p < 0.01, and p < 0.001, respectively. ## and ### Statistically significant differences between the negative control and transiently transfected samples at p < 0.01 and p < 0.001, respectively. . Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, two-tailed Student's t-test. * Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05.

Mitochondrial Fragmentation Is Not Specific for C2C12 Myoblasts
We assessed another cell line in which PARP2 was silenced (scPARP2 and shPARP2 cell lines). Similar to shPARP2 C2C12 cells, in shPARP2 HepG2 cells also, we observed mitochondrial fragmentation marked by an increased number of mitochondrial cross sections on EM sections, decreased perimeter of individual mitochondria, decreased form factor, and increased circularity of mitochondria in MitoTracker-stained and TOMM20stained cells ( Figure 6A-C). Nevertheless, similar to C2C12 cells, no overlap between LC3and TOMM20-positive features in shHepG2cells was detected ( Figure 7A), suggesting that mitochondrial fragmentation is not due to increased mitophagy. We also assessed the mRNA expression of a set of genes involved in regulating mitochondrial morphology. The silencing of PARP2 decreased the expression of Mitofusin1 (Mfn1) and Pink1 ( Figure  7B). The expression pattern of the genes involved in regulating mitochondrial morphology was not similar in shPARP2 C2C12 cells and shPARP2 HepG2 cells, making it unlikely that the deregulation of these genes could be responsible for mitochondrial fragmentation. . Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, two-tailed Student's t-test. * Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05.

Mitochondrial Fragmentation Is Not Specific for C2C12 Myoblasts
We assessed another cell line in which PARP2 was silenced (scPARP2 and shPARP2 cell lines). Similar to shPARP2 C2C12 cells, in shPARP2 HepG2 cells also, we observed mitochondrial fragmentation marked by an increased number of mitochondrial cross sections on EM sections, decreased perimeter of individual mitochondria, decreased form factor, and increased circularity of mitochondria in MitoTracker-stained and TOMM20stained cells ( Figure 6A-C). Nevertheless, similar to C2C12 cells, no overlap between LC3and TOMM20-positive features in shHepG2cells was detected ( Figure 7A), suggesting that mitochondrial fragmentation is not due to increased mitophagy. We also assessed the mRNA expression of a set of genes involved in regulating mitochondrial morphology. The silencing of PARP2 decreased the expression of Mitofusin1 (Mfn1) and Pink1 ( Figure 7B). The expression pattern of the genes involved in regulating mitochondrial morphology was not similar in shPARP2 C2C12 cells and shPARP2 HepG2 cells, making it unlikely that the deregulation of these genes could be responsible for mitochondrial fragmentation.

Silencing of PARP2 Leads to Oxidative Stress, Originating Partly from the Mitochondria through Mitochondrial Protein Imbalance and Fragmentation
The redox state of cells changes in response to PARP inhibition (reviewed in [83,84]). The oxidative stress markers, dihydroethidine fluorescence, 4-hidroxynonenal (4HNE), and thiobarbituric acid-reactive substances (TBARS), increased in shPARP2 C2C12 cells ( Figure 8A-C) compared with controls. Nitrosative stress markers did not exhibit large changes ( Figure 8D-F).
To determine the effects of increased oxidative stress on mitochondrial fragmentation, we assessed how general reductants, such as reduced glutathione (GSH) or Nacetyl-cysteine (NAC), and a mitochondrial-targeted antioxidant MitoTEMPO affected mitochondrial morphology (similar to [69]). Although the antioxidants did not prevent mitochondrial biogenesis (marked by "Mito content" in Figure 9A), they affected the structure of the mitochondrial network ("Perimeter" and "Form factor" in Figure 9A). All three antioxidants increased the perimeters and form factors in mitochondria and slightly decreased the circularity of mitochondria in shPARP2 cells ( Figure 9A). These results suggest that mitochondrial fragmentation relies on oxidative stress of mitochondrial origin. HepG2 cells were analyzed by electron microscopy (n = 1; 45 cells were assessed from each group). Mitochondrial numbers were determined using ImageJ software, and mitochondrial intersections number/cell surface were plotted. (B,C) A total of 80,000 scPARP2 or shPARP2 HepG2 cells were seeded into 24-well plates on glass coverslips, and the cells were labeled with MitoTracker Red (n = 3) (B) or TOMM20 antibody (C) (n = 3). The nuclei were visualized using DAPI. Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20/20). Representative Western blot and electron microscopic and immunofluorescence images are presented in the figure. Numerical values are presented as the average ± SEM in (B) and (C) and as the average ± SD in (A). Statistical significance was determined using paired, two-tailed Student's t-test. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 HepG2 cells at p < 0.05, p < 0.01, and p < 0.001, respectively. HepG2 cells were analyzed by electron microscopy (n = 1; 45 cells were assessed from each group). Mitochondrial numbers were determined using ImageJ software, and mitochondrial intersections number/cell surface were plotted. (B,C) A total of 80,000 scPARP2 or shPARP2 HepG2 cells were seeded into 24-well plates on glass coverslips, and the cells were labeled with MitoTracker Red (n = 3) (B) or TOMM20 antibody (C) (n = 3). The nuclei were visualized using DAPI. Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20/20). Representative Western blot and electron microscopic and immunofluorescence images are presented in the figure. Numerical values are presented as the average ± SEM in (B,C) and as the average ± SD in (A). Statistical significance was determined using paired, two-tailed Student's t-test. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 HepG2 cells at p < 0.05, p < 0.01, and p < 0.001, respectively.
We assessed how the expression of mitochondrial genes responds to antioxidant treatment. The expression of Ndufa2, Ndufb3, Ndufb5, Cox17, and cyt C was induced in shPARP2 C2C12 cells compared with scPARP2 cells ( Figure 9B). The expression of all genes reduced upon treatment with GSH, NAC, and MitoTEMPO ( Figure 9B), similar to [69]. These responses were translated into physiological changes, as the antioxidants reduced mitochondrial oxygen consumption ( Figure 9C), fatty acid oxidation ( Figure 9C), and ATP content ( Figure 9D). There were no differences in cell proliferation and spontaneous cell death between scPARP2 and shPARP2 cells, even in the presence of GSH, NAC, and MitoTEMPO ( Figure 10A,B). Cells 2021, 10, x FOR PEER REVIEW 13 of 24 Figure 7. Mitophagy or deregulation of the mitochondrial fusion/fission system is not responsible for mitochondrial fragmentation in HepG2 cells. (A) A total of 80,000 scPARP2 or shPARP2 HepG2 cells were seeded into 24-well plates on glass coverslips, and LC3-MitoTracker Red-DAPI or LC3-TOMM20-DAPI co-immunofluorescence was performed (n = 3). Colocalization analysis was performed using ImageJ software with the EzColocalization plug-in (measured cells: 100/100). (B) A total of 250,000 scPARP2 or shPARP2 HepG2 cells were seeded into 6-well plates. The mRNA expression of mitochondrial fusion/fission genes was assessed by RT-qPCR (n = 3). Representative immunofluorescence images are presented in (A). Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, twotailed Student's t-test. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 HepG2 cells at p < 0.05, p < 0.01, and p < 0.001, respectively. . Mitophagy or deregulation of the mitochondrial fusion/fission system is not responsible for mitochondrial fragmentation in HepG2 cells. (A) A total of 80,000 scPARP2 or shPARP2 HepG2 cells were seeded into 24-well plates on glass coverslips, and LC3-MitoTracker Red-DAPI or LC3-TOMM20-DAPI co-immunofluorescence was performed (n = 3). Co-localization analysis was performed using ImageJ software with the EzColocalization plug-in (measured cells: 100/100). (B) A total of 250,000 scPARP2 or shPARP2 HepG2 cells were seeded into 6-well plates. The mRNA expression of mitochondrial fusion/fission genes was assessed by RT-qPCR (n = 3). Representative immunofluorescence images are presented in (A). Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, two-tailed Student's t-test. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 HepG2 cells at p < 0.05, p < 0.01, and p < 0.001, respectively.

Silencing of PARP2 Leads to Oxidative Stress, Originating Partly from the Mitochondria through Mitochondrial Protein Imbalance and Fragmentation
The redox state of cells changes in response to PARP inhibition (reviewed in [83,84]). The oxidative stress markers, dihydroethidine fluorescence, 4-hidroxynonenal (4HNE), and thiobarbituric acid-reactive substances (TBARS), increased in shPARP2 C2C12 cells ( Figure 8A-C) compared with controls. Nitrosative stress markers did not exhibit large changes ( Figure 8D-F). Representative Western blot images are presented in (C-E). Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, two-tailed Student's t-test. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05, p < 0.01, and p < 0.001, respectively.
To determine the effects of increased oxidative stress on mitochondrial fragmentation, we assessed how general reductants, such as reduced glutathione (GSH) or N-acetylcysteine (NAC), and a mitochondrial-targeted antioxidant MitoTEMPO affected mitochondrial morphology (similar to [69]). Although the antioxidants did not prevent mitochondrial biogenesis (marked by "Mito content" in Figure 9A), they affected the structure of the mitochondrial network ("Perimeter" and "Form factor" in Figure 9A). All three antioxidants increased the perimeters and form factors in mitochondria and slightly decreased the circularity of mitochondria in shPARP2 cells ( Figure 9A). These results suggest that mitochondrial fragmentation relies on oxidative stress of mitochondrial origin.
We assessed how the expression of mitochondrial genes responds to antioxidant treatment. The expression of Ndufa2, Ndufb3, Ndufb5, Cox17, and cyt C was induced in shPARP2 C2C12 cells compared with scPARP2 cells ( Figure 9B). The expression of all genes reduced upon treatment with GSH, NAC, and MitoTEMPO ( Figure 9B), similar to [69]. These responses were translated into physiological changes, as the antioxidants reduced mitochondrial oxygen consumption ( Figure 9C), fatty acid oxidation ( Figure 9C), . Numerical values are presented as the average ± SEM. Statistical significance was determined using paired, two-tailed Student's t-test. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05, p < 0.01, and p < 0.001, respectively.

PARP1 and PARP3 Are Not Involved in Mitochondrial Fragmentation Elicited by the Silencing of PARP2
PARP1 and PARP3 were implicated in regulating mitochondrial morphology [20,85,86]; hence, we assessed the possible involvement of PARP1 and PARP3. Silencing PARP1 in scPARP2 and shPARP2 C2C12 cells ( Figure 11A) led to the loss of PARP1 auto-PARylation ( Figure 11A). Silencing of PARP1 induced mild mitochondrial fragmentation in scPARP2 cells; nevertheless, it did not influence the morphology in shPARP2 cells ( Figure 11B), suggesting little or no involvement in the mitochondrial phenotype in shPARP2 cells. When PARP3 was silenced ( Figure 12A), upon the application of the negative control, unspecific siRNA, and two specific siRNA, activation of PARP1 was observed ( Figure 12A). Similar to the case of PARP1, silencing of PARP3 in scPARP2 C2C12 cells induces mild mitochondrial fragmentation but does not change the phenotype of the shPARP2 C2C12 cells ( Figure 12B), suggesting little or no involvement in the mitochondrial phenotype in shPARP2 cells. and ATP content ( Figure 9D). There were no differences in cell proliferation and spontaneous cell death between scPARP2 and shPARP2 cells, even in the presence of GSH, NAC, and MitoTEMPO ( Figure 10A,B). Figure 9. General reductants and MitoTEMPO can prevent mitochondrial fragmentation and mitochondrial biogenesis. (A) A total of 20,000 scPARP2 or shPARP2 C2C12 cells were seeded into 24-well plates on glass coverslips and treated with 5 mM glutathione (GSH), 5 mM N-acetyl-L-cysteine (NAC), or 10 µ M MitoTEMPO for 48 h. After treatment, the mitochondrial structure was investigated using MitoTracker Red staining (n = 3). Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20 from each group). (B) A total of 100,000 scPARP2 or shPARP2 C2C12 cells were seeded into 6-well plates and treated with 5 mM glutathione (GSH), 5 mM Nacetyl-L-cysteine (NAC), or 10 µ M MitoTEMPO for 48 h. After treatment, mRNA expression of the indicated mitochondrial genes was assessed by RT-qPCR (n = 4). (C) A total of 2000 scPARP2 or shPARP2 C2C12 cells were seeded into 96well assay plates and treated with 5 mM glutathione (GSH), 5 mM N-acetyl-L-cysteine (NAC), or 10 µ M MitoTEMPO for 48 h. After treatment, Seahorse assays were performed in the presence or absence of etomoxir to assess total and fatty aciddependent oxidation (n = 3). (D) A total of 100,000 scPARP2 or shPARP2 C2C12 cells were seeded into 6-well plates and treated with 5 mM glutathione (GSH), 5 mM N-acetyl-L-cysteine (NAC), or 10 µ M MitoTEMPO for 48 h. ATP content was assessed using the ATP Assay Kit. Data were normalized to protein content, and normalized readings are displayed. Representative immunofluorescence images are presented in (A). Numerical values are presented as the average ± SEM. Statistical significance was determined using ANOVA followed by Tukey's post hoc test, except for (B); in the shPARP2/scPARP ratio, Dunnett's post hoc test was used. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05, p < 0.01, and p < 0.001, respectively. # , ## , and ### Statistically significant differences between control and treated samples at p < 0.05, p < 0.01, and p < 0.001, respectively. (A) A total of 20,000 scPARP2 or shPARP2 C2C12 cells were seeded into 24-well plates on glass coverslips and treated with 5 mM glutathione (GSH), 5 mM N-acetyl-L-cysteine (NAC), or 10 µM MitoTEMPO for 48 h. After treatment, the mitochondrial structure was investigated using MitoTracker Red staining (n = 3). Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20 from each group). (B) A total of 100,000 scPARP2 or shPARP2 C2C12 cells were seeded into 6-well plates and treated with 5 mM glutathione (GSH), 5 mM N-acetyl-L-cysteine (NAC), or 10 µM MitoTEMPO for 48 h. After treatment, mRNA expression of the indicated mitochondrial genes was assessed by RT-qPCR (n = 4). (C) A total of 2000 scPARP2 or shPARP2 C2C12 cells were seeded into 96-well assay plates and treated with 5 mM glutathione (GSH), 5 mM N-acetyl-L-cysteine (NAC), or 10 µM MitoTEMPO for 48 h. After treatment, Seahorse assays were performed in the presence or absence of etomoxir to assess total and fatty acid-dependent oxidation (n = 3). (D) A total of 100,000 scPARP2 or shPARP2 C2C12 cells were seeded into 6-well plates and treated with 5 mM glutathione (GSH), 5 mM N-acetyl-L-cysteine (NAC), or 10 µM MitoTEMPO for 48 h. ATP content was assessed using the ATP Assay Kit. Data were normalized to protein content, and normalized readings are displayed. Representative immunofluorescence images are presented in (A). Numerical values are presented as the average ± SEM. Statistical significance was determined using ANOVA followed by Tukey's post hoc test, except for (B); in the shPARP2/scPARP ratio, Dunnett's post hoc test was used. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05, p < 0.01, and p < 0.001, respectively. # , ## , and ### Statistically significant differences between control and treated samples at p < 0.05, p < 0.01, and p < 0.001, respectively.

PARP1 and PARP3 Are Not Involved in Mitochondrial Fragmentation Elicited by the Silencing of PARP2
PARP1 and PARP3 were implicated in regulating mitochondrial morphology [20,85,86]; hence, we assessed the possible involvement of PARP1 and PARP3. Silencing PARP1 in scPARP2 and shPARP2 C2C12 cells ( Figure 11A) led to the loss of PARP1 auto-PARylation ( Figure 11A). Silencing of PARP1 induced mild mitochondrial fragmentation in scPARP2 cells; nevertheless, it did not influence the morphology in shPARP2 cells ( Figure 11B), suggesting little or no involvement in the mitochondrial phenotype in shPARP2 cells. When PARP3 was silenced ( Figure 12A), upon the application of the negative control, unspecific siRNA, and two specific siRNA, activation of PARP1 was observed ( Figure 12A). Similar to the case of PARP1, silencing of PARP3 in scPARP2 C2C12 cells induces mild mitochondrial fragmentation but does not change the phenotype of the shPARP2 C2C12 cells ( Figure 12B), suggesting little or no involvement in the mitochondrial phenotype in shPARP2 cells.   A,B) A total of 30,000 scPARP2 or shPARP2 C2C12 cells were seeded into 24-well plates, and PARP1 was transiently silenced using three types of siRNAs (n = 3). (A) Cells were assessed 48 h post-transfection, and then the expression of PARP1 was determined using RT-qPCR (n = 3). Expression of PARP1 and PAR was determined using Western blotting (n = 3). (B) In the same cells, the mitochondrial structure was investigated using MitoTracker Red staining (n = 3). Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20 from each group). Representative Western blot and immunofluorescence images are presented in the figure. Numerical values are presented as the average ± SEM. Statistical significance was determined using ANOVA followed by Tukey's post hoc test. *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05. # , ## , and ### Statistically significant differences between the negative control and transiently transfected samples at p < 0.05, p < 0.01, and p < 0.001, respectively.  A,B) A total of 30,000 scPARP2 or shPARP2 C2C12 cells were seeded into 24-well plates, and PARP1 was transiently silenced using three types of siRNAs (n = 3). (A) Cells were assessed 48 h post-transfection, and then the expression of PARP1 was determined using RT-qPCR (n = 3). Expression of PARP1 and PAR was determined using Western blotting (n = 3). (B) In the same cells, the mitochondrial structure was investigated using MitoTracker Red staining (n = 3). Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20 from each group). Representative Western blot and immunofluorescence images are presented in the figure. Numerical values are presented as the average ± SEM. Statistical significance was determined using ANOVA followed by Tukey's post hoc test. *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05. # , ## , and ### Statistically significant differences between the negative control and transiently transfected samples at p < 0.05, p < 0.01, and p < 0.001, respectively.  A,B) A total of 30,000 scPARP2 or shPARP2 C2C12 cells were seeded into 24-well plates, and PARP3 was transiently silenced using three types of siRNAs (n = 3). (A) Cells were assessed 48 h post-transfection, and then the expression of PARP3 was determined using RT-qPCR (n = 3). Expression of PARP3 and PAR was determined using Western blotting (n = 3). (B) In the same cells, the mitochondrial structure was investigated using MitoTracker Red staining (n = 3). Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20 from each group). Representative Western blot and immunofluorescence images are presented in the figure. Numerical values are presented as the average ± SEM. Statistical significance was determined using ANOVA followed by Tukey's post hoc test. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05, p < 0.01, and p < 0.001, respectively. # , ## , and ### Statistically significant differences between the negative control and transiently transfected samples at p < 0.05, p < 0.01, and p < 0.001, respectively.  A,B) A total of 30,000 scPARP2 or shPARP2 C2C12 cells were seeded into 24-well plates, and PARP3 was transiently silenced using three types of siRNAs (n = 3). (A) Cells were assessed 48 h post-transfection, and then the expression of PARP3 was determined using RT-qPCR (n = 3). Expression of PARP3 and PAR was determined using Western blotting (n = 3). (B) In the same cells, the mitochondrial structure was investigated using MitoTracker Red staining (n = 3). Mitochondrial morphology was analyzed using ImageJ software with Mito-Morphology Macro (measured cells: 20 from each group). Representative Western blot and immunofluorescence images are presented in the figure. Numerical values are presented as the average ± SEM. Statistical significance was determined using ANOVA followed by Tukey's post hoc test. *, **, and *** Statistically significant differences between the scPARP2 and shPARP2 C2C12 cells at p < 0.05, p < 0.01, and p < 0.001, respectively. # , ## , and ### Statistically significant differences between the negative control and transiently transfected samples at p < 0.05, p < 0.01, and p < 0.001, respectively.

Discussion
Previous studies have shown that the genetic deletion of PARP1 or PARP2 [55,[87][88][89][90][91][92][93] or pharmacological PARP inhibition [82,87,94] induces mitochondrial biogenesis in multiple organs and tissues, including skeletal muscle. In this study, we provide evidence for fragmentation of the mitochondrial network via induction of oxidative stress in cells. Reduced glutathione (GSH) and N-acetyl-cysteine (NAC), which are known antioxidants that minimize the effect of oxidative stress, reversed the changes to mitochondrial morphology that occur when PARP2 expression is decreased. The mitochondria-targeted antioxidant MitoTEMPO also affected the structure of the mitochondrial network, suggesting that the fragmentation of the mitochondrial network leans on the oxidative stress of mitochondria. GSH, NAC, and MitoTEMPO also reduced mitochondrial oxidative phosphorylation. In good agreement with these results, pharmacological PARP inhibition by olaparib or nicotinamide, as well as increased cellular NAD + levels [75][76][77][78][79][80], also induce mitochondrial fragmentation [20]. PARP1 or PARP3 does not compensate for the loss of PARP2 in C2C12 cells. Furthermore, the silencing of PARP1 or PARP3 can bring about signs of mitochondrial fragmentation.
Although SIRT1 activation in the absence of PARP2 was shown to induce mitochondrial biogenesis [27,55,57] and SIRT1 can regulate mitochondrial morphology [75,81], modulating SIRT1 activity does not influence mitochondrial fragmentation. Silencing of PARP2 induces autophagy [58], making it likely that mitochondrial fragmentation is due to enhanced mitophagy. However, we ruled out the involvement of mitophagy, despite the indication that PARP1 inhibition can induce mitophagy [18]. Furthermore, in the absence of PARP2, the consistent dysregulation of the mitochondrial fusion/fission machinery and the induction of mtUPR were both ruled out.
Treatment of cells with GSH, NAC, or MitoTEMPO reduced the expression of mitochondrial genes and reduced mitochondrial output but did not reduce mitochondrial content in shPARP2 cells. Taken together, mitochondrial biogenesis upon the absence of PARP2 is dependent on the combined action of increased expression and NAD + -dependent activation of SIRT1 and mitochondrial reactive species production. SIRT1 activation [20,82,86] and reactive species production [95][96][97] were linked with mitonuclear proteostasis in other models. Yet, the source of reactive species has not been identified. Nevertheless, higher mitochondrial activity in a cell leads to higher levels of oxygen-centered reactive species in the Szent-Györgyi-Krebs cycle and the mitochondrial electron transport chain [98].
PARP2 has physiological roles in skeletal muscle. Namely, the silencing of PARP2 supports isotype switching, inducing proportions of slow-twitch, type I fibers and supporting mitochondrial biogenesis through inducing SIRT1 [55,57]. Furthermore, the silencing of PARP2 induces the expression of myogenic transcription factors in differentiating C2C12 myoblasts [58], suggesting pro-differentiation effects, while at the same time, the silencing of PARP2 leads to an aberrant stress fiber-like actin structure [58]. The role of PARP2 in skeletal muscle biology warrants further examination. Nevertheless, there is a large set of human diseases involving PARP activation and mitochondrial damage [53,[99][100][101][102][103]; hence, our results, in terms of mitochondrial output and mitochondrial morphology, have implications for understanding the protective action of PARP inhibitors in these diseases.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/10 .3390/cells10061387/s1, Figure S1: Mitochondrial fragmentation in C2C12 is not due to mitophagy or SIRT1 activation, Figure S2: Lack of consistent difference in HSP expression and induction in scPARP2 andshPARP2 cells.