MicroRNA‐382 silencing induces a mitonuclear protein imbalance and activates the mitochondrial unfolded protein response in muscle cells

Abstract Proper mitochondrial function plays a central role in cellular metabolism. Various diseases as well as aging are associated with diminished mitochondrial function. Previously, we identified 19 miRNAs putatively involved in the regulation of mitochondrial metabolism in skeletal muscle, a highly metabolically active tissue. In the current study, these 19 miRNAs were individually silenced in C2C12 myotubes using antisense oligonucleotides, followed by measurement of the expression of 27 genes known to play a major role in regulating mitochondrial metabolism. Based on the outcomes, we then focused on miR‐382‐5p and identified pathways affected by its silencing using microarrays, investigated protein expression, and studied cellular respiration. Silencing of miRNA‐382‐5p significantly increased the expression of several genes involved in mitochondrial dynamics and biogenesis. Conventional microarray analysis in C2C12 myotubes silenced for miRNA‐382‐5p revealed a collective downregulation of mitochondrial ribosomal proteins and respiratory chain proteins. This effect was accompanied by an imbalance between mitochondrial proteins encoded by the nuclear and mitochondrial DNA (1.35‐fold, p < 0.01) and an induction of HSP60 protein (1.31‐fold, p < 0.05), indicating activation of the mitochondrial unfolded protein response (mtUPR). Furthermore, silencing of miR‐382‐5p reduced basal oxygen consumption rate by 14% ( p < 0.05) without affecting mitochondrial content, pointing towards a more efficient mitochondrial function as a result of improved mitochondrial quality control. Taken together, silencing of miR‐382‐5p induces a mitonuclear protein imbalance and activates the mtUPR in skeletal muscle, a phenomenon that was previously associated with improved longevity.


| INTRODUCTION
Mitochondria play a central role in the regulation of cellular metabolism.
However, in recent years it has become increasingly clear that mitochondrial function is complex and depends on far more factors than merely mitochondrial abundance and OxPhos enzyme capacity (reviewed in Dahlmans, Houzelle, Schrauwen, & Hoeks, 2016;Quiros, Mottis, & Auwerx, 2016). Although mitochondria were classically seen as individual organelles, it is now evident that mitochondria form a dynamic network that is constantly being remodeled by the removal of damaged mitochondria (fission) and fusion of healthy mitochondria (Smirnova, Shurland, Ryazantsev, & van der Bliek, 1998), to maintain proper mitochondrial morphology and function.
Furthermore, additional mitochondrial quality control mechanisms have been identified, which are crucial for efficient mitochondrial function and also play a fundamental role in health and disease (Haelterman et al., 2014;Imatoh et al., 2009;Quiros et al., 2016). An important mitochondrial quality control mechanism is the mitochondrial unfolded protein response (Zhao et al., 2002;mtUPR), an evolutionary conserved protein quality control mechanism that has been shown to be deregulated in Parkinson's disease (reviewed in Haelterman et al., 2014) and has also been linked to T2DM . The main role of the mtUPR is to resolve proteotoxic stress, that is, the accumulation of harmful protein aggregates consisting of misfolded and damaged proteins, through the actions of chaperones and proteases (Haynes, Petrova, Benedetti, Yang, & Ron, 2007), which assist in the folding of misfolded or unfolded proteins and the degradation of misfolded proteins (reviewed in Haynes et al., 2007;Jovaisaite & Auwerx, 2015).
Interestingly, during the last decade miRNAs have been reported to not only be present in mitochondria, but also to regulate various aspects of mitochondrial function (Barrey et al., 2011;Bian et al., 2010;El Azzouzi et al., 2013;Li et al., 2014;Mohamed et al., 2014;Sripada, Tomar, & Singh, 2012). In this context, we recently performed an unbiased, hypothesis-free, high throughput miRNA silencing screen in C2C12 muscle cells, and identified 19 miRNAs, which upon silencing, had a positive impact on mitochondrial metabolism (Dahlmans et al., 2017).
Here we extend these findings and measured the expression of 27 genes involved in different aspects of mitochondrial function upon individual silencing of these 19 candidate miRNAs in fully differentiated C2C12 myotubes. Based on the results, we singled out miR-382-5p and show that silencing of miR-382-5p leads to a collective downregulation of the transcripts encoding for the mitochondrial ribosomal proteins, induces a mitonuclear protein imbalance and activates the mtUPR.

| Cell culture and transfection
Myoblasts were seeded in 12-well plates at 100,000 cells/cm 2 in growth media (Dulbecco's modified Eagle medium 4.5 g/l glucose, with 10% fetal bovine serum, 2% HEPES, and 1% nonessential amino acids) and reached full confluence in 24 hr. Once confluence was reached, differentiation was initiated by exchanging the growth media for differentiation media (Dulbecco's modified Eagle medium 4.5 g/l glucose, with 2% horse serum, 2% HEPES, and 1% nonessential amino acids). The media was changed regularly and mature myotubes were obtained after 5 days of differentiation.
Transfections were either conducted at Day 5 or 6 of differentiation. To this end, the cell media was changed to 400 µl fresh differentiation media and the cells were transfected with miRNA inhibitors using lipofectamine RNAimax, according to the manufacturers protocol, reaching a total volume of 500 µl per well.
The final concentration for transfection (50 nM) was previously established in optimization experiments using increasing concentrations of a miRNA inhibitor against the muscle-specific miRNA-206 (Dahlmans et al., 2017). Assays were always performed at Day 7 of differentiation, either 24 or 48 hr after the transfection.

| Individual miRNA inhibitors
Individual mouse miRCURY LNA™ microRNA inhibitors were purchased from (Exiqon A/S (Vedbaek, Denmark). MiRCURY LNA™ microRNA inhibitor control/negative control A (Exiqon A/S, Vedbaek, Denmark) was used as negative control. This control (sequence: TAACACGTCTATACGCCCA) does not target any known mature miRNA of the online miRbase database.

| Microarray analysis
Raw microarray data derived from Affymetrix MoGene 1.1 ST arrays (Santa Clara, CA) were collected and processed per data set into normalized gene expression values using RMAExpress (Bolstad, Irizarry, Astrand, & Speed, 2003) as previously described (Sanderson et al., 2009). Volcano plot was created by using GraphPad v6. Heatmaps were created using GENE-E software (www.broadinstitute.org/cancer/software/GENE-E), which transformed log2-scale gene expression values in each row into z-scores and subsequently converted these to heatmap colors. Raw microarray data were deposited in the GEO archive (GSE116786). Gene Set Enrichment Analysis (GSEA) was performed using the GSEA tool (Subramanian et al., 2005), which was obtained from Broad Institute (http://software.broadinstitute.org/ gsea/index.jsp), with mitochondrial gene sets defined previously (Andreux et al., 2014). Gene sets with a false discovery rate below 0.25 were considered significant.

| Protein extraction and immunoblot analysis
Cells were washed with 1 ml cold 1xphosphate-buffered saline (PBS) and scraped in 40 µl bio-plex cell lysis buffer (Biorad), and processed according to the manufacturers guidelines. OXPHOS proteins were detected using a cocktail of five monoclonal antibodies directed against structural components of the different OXPHOS complexes (MS601, MitoSciences, Eugene, OR), as previously described (Hoeks et al., 2010).
2.5 | RNA extraction, cDNA synthesis, and quantitative RT-PCR analysis Cells were washed with 1 ml ice-cold 1xPBS, harvested in 700 µl Trizol reagent (Invitrogen, Breda, The Netherlands) and stored at −80°C for later use. RNA was isolated using the RNeasy mini kit (Qiagen, Venlo, The Netherlands). RNA quantity and quality were assessed spectrophotometrically (ND-1000, NanoDrop Technologies, Wilmington) and with 6000 Nano chips (Bioanalyzer2100; Agilent, Amstelveen, The Netherlands). cDNA was synthesized using the High-Capacity RNA-to-cDNA Kit (Applied Biosystems). A SensiMix SYBR Hi-ROX kit (Bioline, London, United Kingdom) was used for PCR amplification according to manufacturer's instructions. The following protocol was used for every amplification: 50°C for 2 min, 95°C for 10 min and 40 cycles of 95°C for 15 s followed by 60°C for 1 min. Relative expression was calculated using the 2 −ΔΔCt method, comparing the expression of miRNA silenced C2C12 myotubes to C2C12 myotube transfected with negative control A, using the geometrical mean of 36B4 and B2M gene expression as a reference.
Heatmaps were generated from the average relative expression of three independent experiments, unless stated otherwise, using R (R Foundation for Statistical Computing, Vienna, Austria).
For the determination of mitochondrial DNA copy number, total DNA was isolated using a DNeasy Blood & Tissue Kit (Qiagen, Venlo, The Netherlands). Total DNA was quantified and the integrity was checked by spectrophotometry using the Nanodrop. Relative amounts of nuclear and mtDNA were quantified by quantitative polymerase chain reaction (qPCR), in which nuclear DNA was represented by the UCP2 gene and mitochondrial DNA by the COX2 gene (Table 1). qPCR was performed using HOT FIREPol DNA Polymerase (Solis Biodyne, Tartu, Estonia) in the ABI Prism 7900HT Real-Time PCR system (Applied Biosystems) with the following protocol: 50°C for 2 min, 95°C for 10 min and 40 cycles of 95°C for 15 s followed by 60°C for 1 min. MtDNA copy number was calculated with the following formula: 2ΔCt (ΔCt = CtCOX2 − CtUCP2).

| Cellular respiration in C2C12 myotubes
Oxygen consumption rate (OCR) was measured using the Seahorse XF96 equipment (Seahorse bioscience Inc., North Billerica, MA). On the day of the assay, the cell media was exchanged by 175 µl XF media, according the manufacturer's guidelines. Cells were placed at 37°C with ambient CO 2 concentrations for 45 min. Subsequently, the assay was started and OCRs were measured, three consecutive times with 3-min intervals. First at basal conditions to determine routine respiration, after an oligomycin injection to determine oligomycin-induced leak respiration, after an FCCP injection to measure maximal respiration and finally after an injection containing both antamycinA and rotenone, to determine nonmitochondrial respiration. After the assay, cells were washed twice with PBS and stored at −20°C for protein determination.
Total protein concentrations were determined using the Pierce™ Coomassie (Bradford) Protein Assay Kit (ThermoFisher, Bleiswijk, Netherlands) according to manufacturer's guidelines.

| Statistical analyses
Reported data are expressed as means ± standard error of the mean (SEM) where indicated. Differences in gene expression, protein abundance, respiration, and citrate synthase activity were assessed using Student's t tests. Array data was analyzed using GSEA software (Subramanian et al., 2005). Several gene sets were tested and the q value of the false discovery rate control inferior to 0.25 was used as a significance threshold.

| miRNA-382-5p silencing induces changes in genes involved in mitochondrial dynamics and biogenesis in C2C12 myotubes
To investigate the role of previously identified (Dahlmans et al., 2017) miRNAs in the regulation of skeletal muscle mitochondrial function, we measured the expression of 27 genes involved in different aspects of mitochondrial function, such as mitochondrial biogenesis, metabolism, dynamics, and quality control (

| 6603
The analysis revealed distinct expression patterns as visualized in a clustered heatmap (Figure 1a 3.2 | miRNA-382-5p silencing collectively downregulates mitochondrial ribosomal subunit gene expression To gain more insight in the pathways which are affected by silencing miR-382-5p silencing, we next conducted microarray analyses in C2C12 T A B L E 1 Primer overview. All primer oligonucleotide sequences are shown from 5′ to 3′. Genes with an asterisk represent primer pairs used for mitochondrial DNA copy number experiments

| miRNA-382-5p silencing induces a mitonuclear protein imbalance
To test if the collective downregulation of genes encoding mitochondrial ribosomal protein subunits upon miR-382-5p silencing was indeed accompanied by a mitonuclear protein imbalance, we

| DISCUSSION
We previously conducted an unbiased high throughput miRNA silencing screen in C2C12 myoblasts and myotubes to identify novel miRNAs involved in the regulation of skeletal muscle mitochondrial metabolism. To that end, we individually silenced > 700 miRNAs and investigated several functional parameters for mitochondrial function, and identified 19 miRNAs putatively involved in the regulation of muscle mitochondrial function (Dahlmans et al., 2017). In the current study, we extended these findings and determined the expression of 27 genes related to mitochondrial metabolism, after individual silencing of these 19 candidate miRNAs in C2C12 To this end, we measured the stoichiometric balance of mitochondrial OXPHOS components over the course of a 48-hr timespan. We found that the collectively reduced expression of the MRP gene set upon miR-382-5p silencing was associated with the gradual development of a mitonuclear protein imbalance over time, which started to change 18 hr posttransfection, but was most pronounced 48 hr posttransfection ( Figure 3) (Wang & Auwerx, 2017).
The mismatch between the nDNA-and mtDNA-encoded subunits of mitochondrial components is characteristic of a mitonuclear protein imbalance, and has been shown to cause unfolded protein stress, which subsequently activates mitochondrial protein quality control mechanisms (i.e., mtUPR; Houtkooper et al., 2013;Jovaisaite & Auwerx, 2015). The mtUPR assists in protein folding and degradation, and thereby protects against damaged proteins or protein aggregates. Moreover, in addition to a mitonuclear protein imbalance, the mtUPR is also activated upon mitochondrial biogenesis, mtDNA depletion (Yoneda et al., 2004), ETC subunit loss (Durieux, Wolff, & Dillin, 2011) and interference with mitochondrial architecture (Zhang et al., 2016), and protein import (Rainbolt, Atanassova, Genereux, & Wiseman, 2013) and translation (Houtkooper et al., 2013). Furthermore, previous reports demonstrated that activation of the mtUPR through cco-1 (a nuclear encoded subunit of the ETC) and spg-7 (a mitochondrial protein quality control protease) loss of function, activated the mtUPR and increased lifespan in C. elegans (Durieux et al., 2011;Wu et al., 2014;Yoneda et al., 2004). Therefore, we investigated the abundance of heat shock protein 60 (HSP60), a key protein chaperone of the mtUPR that resides in the mitochondrial matrix. Of note, 48 hr posttransfection when the mitonuclear protein imbalance was most pronounced, activation of the mtUPR was also observed ( Figure 4).
Furthermore, it was reported that the Mrps5 silenced worms displayed a reduced OCR and lowered citrate synthase activity (Houtkooper et al., 2013). Although the worms used less oxygen and had lower citrate synthase activity, their longevity was improved, whereas the worms were more active and moved twice as much compared with the control worms (Houtkooper et al., 2013). In our study, we also observed reduced OCRs in miR-382-5p silenced C2C12 myotubes 48 hr posttransfection, albeit to a lesser extent. In contrast to the findings in worms, however, we did not observe reduced citrate synthase activity or reduced mitochondrial DNA copy number, two commonly used markers of mitochondrial content (Wang, Hiatt, Barstow, & Brass, 1999), indicating that the reduction of OCR is not due to loss of mitochondria.
Besides longevity, activation of the mtUPR may also have implications for metabolic diseases such as T2DM. Specifically, some evidence exists that members of the mtUPR gene set are deregulated F I G U R E 5 MiR-382-5p silencing in C2C12 myotubes reduces basal oxygen consumption without affecting mitochondrial content.
(a) Basal oxygen consumption of the C2C12 myotubes after 48 hr of miR-382-5p silencing was significantly reduced (n = 6). (b,c) Enzymatic activity of citrate synthase (n = 3) and mitochondrial DNA copy number (n = 3) did not change in C2C12 myotubes after 48 hr of miR-382-5p silencing, with * representing p < 0.05 in both animal models for obesity/diabetes and human obese and type T2DM subjects (Bruce, Carey, Hawley, & Febbraio, 2003;Chung et al., 2008;Drew et al., 2014;Henstridge et al., 2014;Imatoh et al., 2009;Rong et al., 2007). However, most studies have been conducted in PBMCs or adipose tissue and relatively few studies have used skeletal muscle. Reports in skeletal muscle, however, do show that the mtUPR is impaired in mouse models for obesity/ diabetes and that there is a close link between components of the mtUPR and mitochondrial function and morphology Gariani et al., 2016;Henstridge et al., 2014;Rong et al., 2007;Zhang et al., 2016). Furthermore, reduced activity of heat shock proteins was not only observed in skeletal muscle of obese and T2DM human subjects compared with controls (Bruce et al., 2003;Chung et al., 2008), but also in plasma of T2DM subjects compared with lean controls (Imatoh et al., 2009). These animal and human studies hence suggest the importance of a functional mtUPR, and underscore the relevance of modulators of the mtUPR such as miR-382-5p.
In conclusion, we previously identified specific miRNAs as modulators of skeletal muscle mitochondrial metabolism. Here, we singled out miR-382-5p and showed that silencing of this miRNA leads to a collective downregulation of mitochondrial ribosomal protein expression, induces a mitonuclear protein imbalance, and activates the mtUPR, previously associated with improved longevity.