T-Cell Intracellular Antigens and Hu Antigen R Antagonistically

RESEARCH ARTICLE crossm

T-Cell Intracellular Antigens and Hu

Antigen R Antagonistically Modulate Downloaded from Mitochondrial Activity and Dynamics by Regulating Optic Atrophy 1 Gene Expression

Isabel Carrascoso, José Alcalde, Carmen Sánchez-Jiménez, http://mcb.asm.org/ Paloma González-Sánchez, José M. Izquierdo Centro de Biología Molecular Severo Ochoa, Consejo Superior de Investigaciones Cientíﬁcas, Universidad Autónoma de Madrid, Madrid, Spain

ABSTRACT Mitochondria undergo frequent morphological changes to control their function. We show here that T-cell intracellular antigens (TIA1b/TIARb) and Hu anti- Received 5 April 2017 Returned for modification 26 April 2017 Accepted 6 June gen R (HuR) have antagonistic roles in mitochondrial function by modulating the ex- 2017 pression of mitochondrial shaping proteins. Expression of TIA1b/TIARb alters the mito- Accepted manuscript posted online 19 on December 20, 2017 by Red de Bibliotecas del CSIC chondrial dynamic network by enhancing fission and clustering, which is accompanied June 2017 by a decrease in respiration. In contrast, HuR expression promotes fusion and cristae re- Citation Carrascoso I, Alcalde J, Sánchez- Jiménez C, González-Sánchez P, Izquierdo JM. modeling and increases respiratory activity. Mechanistically, TIA proteins downregulate 2017. T-cell intracellular antigens and Hu the expression of optic atrophy 1 (OPA1) protein via switching of the splicing patterns antigen R antagonistically modulate of OPA1 to facilitate the production of OPA1 variant 5 (OPA1v5). Conversely, HuR en- mitochondrial activity and dynamics by regulating optic atrophy 1 gene expression. hances the expression of OPA1 mRNA isoforms through increasing steady-state levels Mol Cell Biol 37:e00174-17. https://doi.org/10 and targeting translational efficiency at the 3= untranslated region. Knockdown of TIA1/ .1128/MCB.00174-17. TIAR or HuR partially reversed the expression profile of OPA1, whereas knockdown of Copyright © 2017 American Society for Microbiology. All Rights Reserved. OPA1 or overexpression of OPA1v5 provoked mitochondrial clustering. Middle-term ex- Address correspondence to José M. Izquierdo, pression of TIA1b/TIARb triggers reactive oxygen species production and mitochondrial [email protected]. DNA damage, which is accompanied by mitophagy, autophagy, and apoptosis. In contrast, HuR expression promotes mitochondrion-dependent cell proliferation. Collectively, these results provide molecular insights into the antagonistic functions of TIA1b/TIARb and HuR in mitochondrial activity dynamics and suggest that their balance might contribute to mitochondrial physiopathology.

KEYWORDS TIA1, TIAR, HuR, OPA1, mitochondrial spatial dynamics, mitophagy, apoptosis, autophagy

itochondria function as cellular “powerhouses” to efﬁciently generate metabolic Menergy in the form of ATP. They also play critical roles in many cellular functions, including intermediary metabolism, calcium signaling, reactive oxygen species (ROS) defense and production, and cell survival and death, and their activity and function are optimized and adapted to the metabolic demands of different cell and tissue types (1). Mitochondria form interconnected networks that continually fuse and divide. Be- yond de novo mitochondrial biogenesis, the dynamic equilibrium between fusion and ﬁssion events and associated inner membrane remodeling are fundamental for mitochondrial morphology and function in eukaryotic cells (2). The inner mitochondrial membrane is extensively folded, producing invaginations called cristae, which control oxidative phosphorylation (OXPHOS), electron and metabolite transport, and cell death and survival (3). Consequently, mitochondrial network dysregulation is linked to several

September 2017 Volume 37 Issue 17 e00174-17 Molecular and Cellular Biology mcb.asm.org 1 Carrascoso et al. Molecular and Cellular Biology pathophysiological conditions, including cancer, diabetes, and neurodegenerative dis- eases (4–8). Mitochondrial biogenesis is controlled by speciﬁc transcription factors, such as peroxisome proliferator-activated receptor ␥ coactivator 1␣ (PGC-1␣)(9), which activates mitochondrial DNA (mtDNA) replication, and by RNA binding proteins (RBPs), such as, for example, Y-box-binding protein 1 (YB-1) and clustered mitochondria (cluA/CLU1) homolog (CLUH), which exert posttranscriptional control (10, 11).

Conversely, mitochondrial dynamics is controlled by mitochondrion-shaping proteins Downloaded from that regulate fusion and fission events. Core components of the mitochondrial fusion/ fission machinery include mitofusin 1 (MFN1), mitofusin 2 (MFN2), and optic atrophy 1 (OPA1), which promote fusion, whereas fission is governed by dynamin-related protein 1 (DRP1) and by adaptor proteins such as mitochondrial fission factor (MFF), mitochondrial dynamics proteins (MiD49 and MiD51), and fission 1 (FIS1) (1–8). T-cell intracellular antigen 1 (TIA1), TIA1-like/related protein (TIAL1/TIAR), and Hu antigen R (HuR/ELAVL1) are RBPs that exert transcriptional and/or posttranscriptional control of gene expression (12–23). Their regulatory roles are directed at specific sites http://mcb.asm.org/ within the transcriptome through association with specific RNA sequence motifs (13– 20). TIA1 and TIAR have two main isoforms generated by alternative splicing of their pre-mRNAs. The TIA1a isoform (43 kDa) differs from isoform TIA1b (40 kDa) by inclusion of an 11-amino-acid sequence, encoded by exon 5 (12). Isoform TIARa (50 kDa) differs from isoform TIARb (42 kDa) in that it contains a sequence of 17 amino acids, encoded by the last 51 nucleotides of exon 3 (12). In the nucleus, TIA and HuR proteins modulate DNA-dependent transcription and processing of the precursor RNAs (i.e., constitutive and alternative splicing), and in the cytoplasm, they regulate localization, stability, on December 20, 2017 by Red de Bibliotecas del CSIC and/or translation of mRNAs. Their roles are critical for cell homeostasis, since they control the expression of essential genes in biological programs such as survival and death, proliferation and differentiation, inflammation, environmental stress, viral infec- tions, embryogenesis, tumorigenesis, and aging, with an impact in physiopathology (12–18, 20, 21). Further, they seem to have an important role during embryogenesis, since mice deficient for TIA1, TIAR, or HuR (as well as ectopic overexpression of TIAR) have high rates of embryonic and postnatal lethality (24–27). Here we show that ectopic expression of TIA1b, TIARb, and HuR alters mitochondrial morphology and function by targeting, differentially and antagonistically, regulatory events associated with transcriptional and/or posttranscriptional control of OPA1 gene expression.

RESULTS TIA1b, TIARb, and HuR proteins have opposing effects in mitochondrial dynamics. Inducible HEK293 cells expressing green ﬂuorescent protein (GFP), GFP-TIA1b, GFP-TIARb, or GFP-HuR (FT293 cells) were generated using the Flp-In T-REx platform (28) and analyzed by Western blotting (Fig. 1A). Ectopic expression of TIA1b or TIARb isoforms for 24 to 48 h resulted in a disorganization of the mitochondrial network as assessed by confocal microscopy using an antibody against Tom20 (Fig. 1A) and by Mitotracker staining (data not shown). Compared with control FT293 cells, expression of either TIA1b or TIARb led to changes in mitochondrial spatial dynamics without observable changes in mitochondrial mass (Fig. 1A and B). In contrast, cells expressing HuR presented enlarged mitochondria that were distributed throughout the cell cytoplasm (Fig. 1A and B). Detailed analysis by transmission electron microscopy (TEM) revealed that whereas TIA1b and TIARb expression promoted mitochondrial clustering and ﬁssion, HuR expression provoked mitochondrial fusion (Fig. 1B). Closer inspection of mitochondria revealed evident changes in crista organization in TIA1b- and TIARb- expressing cells, with many cristae having a slightly wider and more loosely organized intermembrane space than those of control cells (Fig. 1B and C). These phenotypes were not apparent in FT293 cells overexpressing truncated TIA1(ΔQ) and TIAR(ΔQ) proteins lacking the C-terminal glutamine (Q)-rich domain (28) involved in protein- protein interactions (data not shown). Conversely, HuR-expressing cells showed more

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 2 TIA and HuR Are Drivers of Mitochondrial Dynamics Molecular and Cellular Biology Downloaded from http://mcb.asm.org/ on December 20, 2017 by Red de Bibliotecas del CSIC FIG 1 Mitochondrial architecture and morphology in FT293 cells. (A) Mitochondrial spatial morphology in FT293 cells cultured for 24 to 48 h visualized by confocal fluorescence microscopy. Shown are the time course and expression (24 to 48 h) of ectopic GFP, GFP-TIA1b, GFP-TIARb, and GFP-HuR proteins and endogenous ␣-tubulin (TUBA) in FT293 cells by Western blotting. The asterisk indicates a nonspecific protein band. Molecular weight markers (in thousands) and the identities of protein bands are shown. At the bottom are fluorescence images from FT293 cells expressing GFP-tagged proteins (green) and immunolabeled for mitochondrial Tom20 (red). Boxes in the merge row (white) are enlarged in the detail row. Bars, 10 ␮m. The detail row shows 2ϫ zoomed images. (B) Details of mitochondrial morphology and crista architecture in FT293 cells visualized by transmission electron microscopy. Bars, 2 ␮m (left images), 500 nm (middle images), and 200 nm (right images). (C) Estimation of mitochondrial dynamics in FT293 cells. Data represent means Ϯ SEM (n ϭ 200 to 400 cells and 150 to 200 mitochondria per condition, respectively). *, P Ͻ 0.05; **, P Ͻ 0.01. tightly packed cristae together with structural changes affecting crista organization and architecture, becoming more electron dense (Fig. 1B and C). Overall, these observations suggest that TIA1b, TIARb, and HuR are antagonistic modulators of mitochondrial dynamics by promoting mitochondrial crista remodeling, clustering, and fission/fusion. Ectopic expression of TIA1b, TIARb, and HuR alters cell respiratory rates. Because changes in mitochondrial dynamic networks often go hand in hand with alterations in mtDNA copy number (1, 2), we quantified the ratio of mtDNA and nuclear DNA (nDNA) in cell lines using specific primers against two mtDNA genes (ND1 and Dloop) and two nDNA genes (ACTB and 18S). The mtDNA/nDNA ratios were similar between control and FT293 cells (Fig. 2A), suggesting that the changes to mitochondria are linked to reorganization/dynamics rather than de novo mitochondrial biogenesis. We hypothesized that the differences in mitochondrial morphology observed between FT293 cells would result in distinct metabolic phenotypes. Specifically, we expected that the mitochondrial cristae and spatial remodeling in TIA-expressing cells would lead to a dissociation of the electron transport chain (ETC) and a less efficient oxidative phosphorylation (OXPHOS) capacity. The opposite effect would be expected in HuR-expressing cells. We therefore measured mitochondrial oxygen consumption rate (OCR) in FT293 cells using the Seahorse Bioscience XF analyzer. The mitochondrial respiratory response was measured before and after stress tests by sequential addition of oligomycin, 2,4-dinitrophenol, and rotenone and antimycin A to determine ATP- linked, maximal, and nonmitochondrial respirations, respectively (Fig. 2B). Basal mitochondrial respiration was higher in HuR-expressing cells than in GFP-, TIA1b-, and TIARb-expressing cells (Fig. 2B and C). Moreover, maximal mitochondrial respiration, ATP production, and spare mitochondrial respiration capacity were also significantly higher in HuR-expressing cells than in control GFP cells, whereas these metrics were

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 3 Carrascoso et al. Molecular and Cellular Biology Downloaded from http://mcb.asm.org/

FIG 2 Ectopic expression of TIA1b, TIARb, or HuR modulates mitochondrial respiration. (A) Mitochondrial DNA copy number was estimated by qPCR. Data were normalized and expressed relative to GFP and represent means Ϯ SEM (n ϭ 3). (B) Real-time changes in oxygen consumption rate (OCR) using cells in basal on December 20, 2017 by Red de Bibliotecas del CSIC condition (glucose 10 mM) and after sequential injection of oligomycin (Oligo), 2,4-dinitrophenol (DNP), and rotenone plus antimycin A (RotϩAA). Values were normalized to milligram of protein. (C) Mitochondrial parameters of nonmitochondrial, basal, and maximal respiration, proton (Hϩ) leak, ATP production, and spare respiratory capacity. (D) Real-time changes in ECAR, an indicator of lactic acid production or glycolysis. (E) Determination of the aerobic and glycolytic components of cellular bioenergetics. Values are means Ϯ SEM (n ϭ 4). *, P Ͻ 0.05. modestly lower in TIA1b- and TIARb-expressing cells (Fig. 2B and C and data not shown). The extracellular acidification rate (ECAR), an index of glycolysis, was also higher in HuR-expressing cells than in controls, whereas it was lower in TIA1b- and TIARb-expressing cells (Fig. 2D), indicating that anaerobic glucose oxidation is increased in HuR-expressing cells and decreased in TIA1b- and TIARb-expressing cells (Fig. 2E and data not shown). Ectopic expression of TIA1b or TIARb thus provokes respiratory deficiency and compromises mitochondrial function, but without significant impact on short-term cell survival (28). These findings indicate that expression of TIA1b and TIARb has a negative effect on ETC function, whereas HuR expression has a positive effect. TIA1b, TIARb, and HuR differentially regulate OPA1 gene expression. A key player of both mitochondrial architecture and dynamics is the mitochondrial inner membrane-distributed protein dynamin-related GTPase OPA1, which promotes inner membrane fusion and crista remodeling (29–31). Given the major changes observed in crista architecture and morphology in FT293 cells, we tested whether this reflected differential expression of OPA1. The human OPA1 gene contains 29 exons (Fig. 3A)(29, 30). Alternative splicing of OPA1 exons 4, 4b, and 5b generates eight different RNA isoforms (Fig. 3B), resulting in eight protein variants with diverse effects on mitochondrial architecture and dynamics (29–31). We first established the expression profile of OPA1 RNA isoforms in FT293 cells by reverse transcription-PCR (RT-PCR) using primer pairs designed for constitutive exon 3 as well as constitutive exons 9 and 10 (Fig. 3B). We identified three amplified products of 762, 708, and 600/597 bp in GFP- and HuR-expressing cells (Fig. 3C), which may correspond to OPA1 isoforms 8, 7, and 4/1, respectively (32). The DNA fragments of 708 and 600/597 bp were robustly amplified. In contrast, we observed a reproducible splicing switch in TIA1b- and TIARb-expressing cells, with a decrease in the levels of the 762- and 708-bp fragments and the amplification of a fragment of 654/651 bp, which may correspond to OPA1 isoforms 6 and/or 5, respectively, containing the 4b exon (Fig. 3C). This splicing switch was specific for

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 4 TIA and HuR Are Drivers of Mitochondrial Dynamics Molecular and Cellular Biology Downloaded from http://mcb.asm.org/ on December 20, 2017 by Red de Bibliotecas del CSIC FIG 3 OPA1 gene expression is antagonistically modulated by TIA1b, TIARb, and HuR proteins. (A) Schematic representation of the human OPA1 gene. Boxes below the gene map show location of protein domains with functional importance. The protein domains indicated correspond to basic domain, alternative splicing domain (AS), coiled-coil domain (CC), GTPase domain, dynamin domain, and GTPase effector domain (GED). (B) Schematic map of the main eight OPA1 mRNA spliced variants. Cleavage of mitochondrial targeting sequence (MTS) by mitochondrial processing peptidase (MPP) drives the long isoforms of OPA1. Additional cleavage at sites S1 (exon 5) or S2 (exon 5b) leads to the short isoforms of OPA1. TM, TM1, and TM2a/b are transmembrane domains. MIS, mitochondrial import sequence (29, 30). (C) Patterns of OPA1 RNA isoforms in FT293 cells by elongation RT-PCR analysis. The oligonucleotide pairs used were as follows: Exon 3.forward, 5=-GGATTGTGCCTGACATTGTG-3=, and Exons 9/10.reverse, 5=-CACTCAGAGTCACCTTAACTGG-3=. Amplification products are assigned to OPA1 isoforms according to expected sizes of DNA fragments. Western blot analysis of the expression levels of GFP-tagged fusion proteins in FT293 cells was performed. (D) Truncated versions of TIA1 and TIAR proteins without C-terminal glutamine (Q)-rich domains are refractory for regulation of OPA1 exon 4b splicing. RT-PCR and Western blotting were carried out as for panel C. (E) qPCR analysis of human OPA1 spliced mRNAs in FT293 cells. Shown is a schematic representation of isoform 8 of OPA1 mRNA and oligonucleotide pairs used in qPCR analysis (29). Values were normalized and expressed relative to GFP for each of the detected OPA1 isoforms. Data were plotted and represent means Ϯ SEM (n ϭ 2or3).*, P Ͻ 0.05; **, P Ͻ 0.01. (F) Proteolytic cleavage of OPA1 produces long protein isoforms (L-OPA1) a and b from eight OPA1 mRNA variants shown in panel B. Additional cleavages at protease sites S1 by OMA1 and S2 by YME1L generate short protein isoforms (S-OPA1) indicated as c to e. (G) Western blot analysis of the key players in fusion and fission events in FT293 cells. G-TIA1/R and G-HuR represent GFP-TIA1b/GFP-TIARb and GFP-HuR, respectively. (H) Time course of OPA1 spliced variants in FT293 cells after 24 to 48 h. RT-PCR analysis was performed as described for panel C. For panels C, D, and H, molecular size markers for DNA (base pairs) and molecular mass markers for protein (kilodaltons) as well as the identities of amplified DNA fragments and proteins bands are shown.

TIA1b and TIARb, because cells expressing truncated versions of TIA1b and TIARb without the C-terminal domain (28) failed to regulate OPA1 exon 4b inclusion (Fig. 3D), and the splicing patterns were identical to those of control and HuR-expressing cells (Fig. 3C and D). To characterize the amplified DNA fragments from endogenous OPA1 RNA isoforms, we performed an analysis with the restriction endonucleases Eco 81l and Bsh 1236ll and DNA sequencing (32) (data not shown). OPA1 mRNA isoforms 1 and 7 were predominantly expressed in GFP- and HuR-expressing cells, whereas OPA1 isoform 5 was expressed in TIA1b- and TIARb-expressing cells. We confirmed this by quantitative PCR (qPCR) analysis using specific primer pairs against exons 3, 4, 4b, 5, and 5b (to quantify the combination of alternative OPA1 exons generated by alternative splicing) and against constitutive exon 21 (to quantify the total amount of OPA1 mRNA) (Fig. 3E)(30, 32). The results showed that OPA1 isoforms 1, 2 (this isoform must be underrepresented, because it was not significantly detected by RT-PCR [Fig. 3C to E]), and 5 were expressed, whereas OPA1 isoforms 4, 7, and 8 were downregulated. OPA1 isoforms 3 and 6 were not detected. Additionally, HuR-expressing cells had approximately twice the steady-state levels of OPA1 mRNA as did GFP-, TIA1b-, or TIARb-expressing cells (Fig. 3E). These findings indicate that TIA1 and TIAR proteins modulate alternative splicing of OPA1 pre-mRNA, particularly exon 4b inclusion and exon 5b skipping, whereas HuR increases the steady-state levels of OPA1 RNA isoforms.

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 5 Carrascoso et al. Molecular and Cellular Biology

To investigate whether these changes were reﬂected at the protein level, we examined the relative protein expression of OPA1 in FT293 cells by Western blotting. The steady-state level of OPA1 protein is controlled by de novo synthesis and also by sequential proteolytic cleavage by posttranslational processing for import into mitochondria via a mitochondrial targeting sequence by mitochondrial processing peptidase (Fig. 3B)(29, 30). Its fate in the inner membrane is dependent on cleavage by mitochondrial OMA1 and YME1L proteases at S1 and S2 sequences, respectively (Fig.

3F). The action of these proteases on the long form of OPA1 (93 kDa) generates a short Downloaded from form (88 kDa) (Fig. 3F)(29–31). Both forms are essential for the function of OPA1 on mitochondrial architecture and dynamics (29–31). As shown in Fig. 3G, the relative level of short-form OPA1 was slightly lower in TIA1b- and TIARb-expressing cells than in controls, whereas the level of both long and short forms in HuR-expressing cells was higher. These observations suggest that OPA1 expression is modulated by TIA1b, TIARb, and HuR, determining the balance between fission and fusion mediated by additional regulatory genes such as those for proteases OMA1, YME1L, MFN1, and http://mcb.asm.org/ MFN2, involved in fusion, and DRP1, MFF, and FIS1, implicated in fission (2, 31). Accordingly, molecular components associated with fusion (OMA1, YMEL1, and MFN2) were modestly downregulated in TIA1b- and/or TIARb-expressing cells, whereas the fission-associated protein MFF was slightly upregulated (Fig. 3G), as recently reported (33). Conversely, the effects of HuR expression were associated exclusively with the robust upregulation of long and short forms of OPA1 protein. The splicing switch of OPA1 was also visualized in TIA1b- and TIARb-expressing cells for 24 to 48 h postin-

duction (Fig. 3H). As OPA1 protein was the only shared target modulated by TIA1b, on December 20, 2017 by Red de Bibliotecas del CSIC TIARb, and HuR expression, we focused on the potential molular mechanisms through which these proteins regulate its expression. TIA proteins regulate OPA1 exon 4b splicing. Data from individual-nucleotide resolution UV cross-linking and immunoprecipitation (iCLIP) analysis of TIA1 and TIAR (18) and HuR (20, 21) in human cells point to human OPA1 as a posttranscriptional target of TIA1, TIAR, and HuR (Fig. 4A and B). To assess whether TIA1b and TIARb directly regulate alternative splicing of OPA1 exon 4b, we generated a human OPA1 minigene containing exons 4, 4b, and 5 (Fig. 4C) driven by the cytomegalovirus (CMV) promoter and performed a cotransfection study to determine the endogenous responses of these introns to TIA1b, TIARb, and HuR. TIA1b and TIARb could efficiently promote exon 4b inclusion (Fig. 4D and E). Additionally, we performed a competitive- splicing analysis by mixing OPA1 with SMN1/SMN2 and NF1 minigenes (which are targeted by TIA1 and TIAR to promote the inclusion of exons 7 and 23a, respectively [34, 35]) in cells cotransfected with these proteins, to gauge the yield of OPA1 exon 4b splicing. The pattern of the alternatively spliced products was monitored by RT-PCR, and expression levels of the GFP-fused proteins were assessed by Western blotting. The results showed that the same degree of overexpression of TIA1b and TIARb promoted also efficient inclusion of OPA1 exon 4b (Fig. 4E). However, overexpression of HuR alone did not appreciably affect exon 4b splicing (Fig. 4E). To strengthen these observations, we generated a mutant OPA1 minigene harboring a substitution at the uridine-rich sequence (UUUUU to UAGAA) of intron 4b close to the 5= splice site (Fig. 4C). Cotransfection analysis of mutant versus wild-type (WT) OPA1 minigenes with the individual TIA1b, TIARb, and HuR proteins showed that the mutant minigene was resistant to exon 4b inclusion by ectopic TIA1b and TIARb (Fig. 4F). Likewise, a 32P-uridine-labeled RNA probe containing the same mutation bound with less efficiency (2-fold) to recombinant glutathione S-transferase (GST)–TIA1b and GST-TIARb proteins as evaluated by UV cross-linking (Fig. 4G). Collectively, these results are consistent with a splicing enhancing activity of TIA1b and TIARb for exon 4b inclusion in OPA1 pre-mRNA. HuR regulates OPA1 mRNA translation via the 3= UTR. The RNA maps of TIA1 and TIAR (18) and HuR (20, 21) protein for OPA1 pre-mRNA suggest that the likely region of binding is at the 3= untranslated region (UTR) (Fig. 4A and B). To test this, we generated

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 6 TIA and HuR Are Drivers of Mitochondrial Dynamics Molecular and Cellular Biology Downloaded from http://mcb.asm.org/ on December 20, 2017 by Red de Bibliotecas del CSIC

FIG 4 TIA1b, TIARb, and HuR proteins regulate alternative splicing of OPA1 pre-mRNA and translation of OPA1 mRNA in a 3=-UTR-dependent manner. (A) RNA map of TIA proteins on the human OPA1 gene. Shown are profiles of experimental cross-link sites of TIA proteins on human OPA1 gene (18). (B) Summary of HuR-iCLIP analysis of human OPA1 gene. Data were adapted from HuR-iCLIP analysis (20). (C) Schematic representation of the human OPA1 minigene. Exons 4, 4b, and 5 are represented by colored boxes, introns by thick lines, and spliced variants by dashed lines. The sizes of exons and introns are indicated in nucleotides. Arrows illustrate the location of primers used in RT-PCR analysis. Also indicated is the RNA sequence (wild-type [WT] versus mutated [MUT] version) on intron 4b containing a putative functional U-rich element to bind TIA proteins. (D) Analysis of endogenous and ectopic OPA1 alternative splicing in HEK293 cells cotransfected with OPA1 minigene and GFP-, GFP-TIA1b-, GFP-TIARb-, and GFP-HuR-expressing plasmids. Western blots of the expression levels of GFP-tagged fusion proteins transfected in HEK93 cells are shown. (E) Analysis of alternative splicing of TIA-targeted minigenes for SMN1/SMN2, NF1, and OPA1 in HEK293 cells. The indicated minigenes were cotransfected and analyzed as for panel D. (F) Analysis of alternative splicing of a mutated OPA1 minigene (C). RT-PCR and Western blotting were carried out as for panel D. (G) UV-cross-linking of GST-TIA1, GST-TIAR, and MBP-HuR recombinant proteins to U-rich and mutated sequences derived from WT and MUT minigenes. (H) Chimeric GFP constructs containing human OPA1 and OMA1 3= UTRs were generated and transfected in HEK293 cells. The relative expression levels of GFP versus ectopic GFP-tagged fusion proteins and endogenous ␣-tubulin levels were analyzed by Western blotting. (I) Relative expression levels of reporter chimeric mRNAs and endogenous mRNAs were quantified by qPCR analysis. The data were plotted and represent means Ϯ SEM (n ϭ 2or3).*, P Ͻ 0.05; **, P Ͻ 0.01. Molecular size markers for DNA (base pairs), molecular mass markers for protein (kilodaltons), and the identities of amplified DNA fragments and proteins bands are shown. G, 1b, Rb, and H stand for GFP, TIA1b, TIARb, and HuR samples, respectively.

a reporter construct with the entire human OPA1 3= UTR inserted downstream of the GFP coding region in plasmid pEGFP-C1 (Fig. 4H). GFP expression was assessed by Western blotting following cotransfection of the reporter construct and GFP-TIA1b, GFP-TIARb, and GFP-HuR. As a control, the cotransfection of empty pEGFP-C1 with TIA1, TIAR, and HuR vectors showed that ectopic expression of these proteins had no effect on endogenous GFP expression (Fig. 4H). When TIA1b, TIARb, and HuR vectors were contransfected with the GFP-OPA1 3= UTR, the level of GFP was lower in TIARb- expressing cells than in control (empty plasmid) and TIA1b-expressing cells, whereas it was higher in HuR-expressing cells (Fig. 4H). A pEGFP-C1-OMA1 3=-UTR reporter construct was used as an additional putative target; however, GFP reporter expression was not affected by TIA1b, TIARb, or HuR expression (Fig. 4H). Thus, the enhanced expression of OPA1 protein in HuR-expressing cells is due to activity at the 3= UTR and to the increased state-steady levels of OPA1 mRNA (Fig. 4H and I), whereas the reduced expression of OMA1 can be due to its reduction at the mRNA level (Fig. 4I). In addition, the increase in GFP did not result from changes in state-steady levels of chimeric GFP mRNAs, since a similar abundance of GFP mRNAs was detected by qPCR analysis (Fig. 4I). Collectively, these results show that HuR (functioning as a translational activator) and TIARb (operating as a translational repressor) can regulate translation of OPA1

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 7 Carrascoso et al. Molecular and Cellular Biology Downloaded from http://mcb.asm.org/ on December 20, 2017 by Red de Bibliotecas del CSIC

FIG 5 Loss and gain of function of OPA1 influence mitochondrial spatial dynamics in HEK293 cells. (A) Protein extracts from control (C) and TIA1, TIAR, and HuR knockdown HeLa cells were analyzed by Western blotting as indicated for Fig. 3G. Patterns of OPA1 RNA isoforms in above HeLa cells were analyzed by RT-PCR as described for Fig. 3C (size markers are in base pairs). The oligonucleotide pairs for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) used were as follows: GAPDH.forward, 5=-GATCATCAGCAATGCCTCCT-3=, and GAPDH.reverse, 5=-TTCAGCTCAGGGATGACCTT-3=. (B) Transmission electron micrographs of mitochondria populations in control and knockdown HeLa cells. Bars, 10 ␮m. (C) Transient depletion of OPA1 in HEK293 cells. Protein extracts {5 ␮g [C(1/4)] or 20 ␮g [C and lanes 1 to 3] of protein} were prepared 72 h after transfection with siRNAs against control (C) and OPA1 (samples 1 to 3) and analyzed by Western blotting. (D) Analysis of OPA1 expression and mitochondrial spatial morphology in FT293 cells transfected transiently with 1 and 3 siRNAs against OPA1 mRNA for 72 h and visualized by confocal fluorescence microscopy. GFP-, GFP-TIA1b-, and GFP-HuR-expressing FT293 cells (green) were immunolabeled (red) for mitochondrial OPA1 and Tom20. Nuclei were stained with DAPI (4=,6-diamidino-2-phenylindole; blue). Bars, 10 ␮m. (E) Transient transfection of human OPA1 variant 5 (OPA1v5) and mt-Keima in HEK293 cells. The blot was probed with the antibodies indicated. (F) Effect of OPA1v5 expression on mitochondrial spatial morphology in HEK293 cells visualized by a mt-Keima probe using confocal fluorescence microscopy. Bars, 10 ␮m. (G) Limited proteolytic processing of OMA1-mediated OPA1 protein in TIA1b-expressing cells. GFP-, GFP-TIA1b-, and GFP-HuR-expressing FT293 cells were transfected with OPA1v5, and the ratio of OPA1 short and long forms was examined by Western blotting using the antibodies indicated. (H) Effect of carbonyl cyanide3-chlorophenylhydrazone (CCCP)-induced mitochondrial uncoupling on OPA1 proteolytic processing in GFP-, TIA1b-, and HuR-expressing FT293 cells. Proteolytic cleavage of mitochondrial membrane potential-dependent OPA1 was analyzed after treating GFP-, TIA1b-, and HuR-expressing FT293 cells with CCCP (10 ␮M) for 0, 5, and 10 min. Molecular weight markers (in thousands) and the identities of protein bands are shown. mRNA in a 3=-UTR-dependent manner. Thus, HuR increases both mRNA and protein levels of OPA1. Loss of function of TIA1, TIAR, and HuR alters OPA1 gene expression and mitochondrial dynamics. We used different loss-of-function cell models of TIA1, TIAR, or HuR to analyze the protein expression of OPA1 and other molecular components involved in mitochondrial dynamics. Ablation of TIA1 or TIAR expression in mouse embryonic fibroblasts (MEFs) favored the appearance of short forms of OPA1, which was more evident in TIAR knockout (KO) MEFs (data not shown). Additionally, other molecular components of mitochondrial fission and fusion were downregulated in TIA1 KO MEFs, including MFF (33), and both DRP1 and MFN2, albeit modestly (data not shown). In agreement with previous findings (33, 36), the dominant mitochondrial phenotype in TIA1 and TIAR KO MEFs was the appearance of elongated mitochondria more compatible with fusion phenotypes (data not shown). We extended the analysis to TIA1, TIAR, or HuR knockdown HeLa cells. The results showed a modest increase of short protein forms of OPA1 in TIA1, TIAR, and HuR knockdown HeLa cells compared to those in control cells (Fig. 5A), in agreement with previous observations in TIA1 and TIAR KO MEFs. However, we observed that OPA1 RNA isoform 7 in TIA1 and TIAR knockdown HeLa cells and OPA1 RNA isoform 1 in HuR

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 8 TIA and HuR Are Drivers of Mitochondrial Dynamics Molecular and Cellular Biology knockdown HeLa cells were slightly favored compared to control cells (Fig. 5A). These observations were also validated by transient RNA interference analysis in HEK293 cells (data not shown). Other components involved in mitochondrial fusion and ﬁssion were mostly not altered, although a slight increase in OMA1 and MFN2, as well as a decrease in one of the two protein forms assigned to MFF, was detected. Finally, the mitochondrial phenotypes observed by TEM (Fig. 5B) were very mild and more heterogeneous than those found in TIA1 and TIAR KO MEFs (30, 33), and no major differences are found

between HuR knockdown and control HeLa cells (Fig. 5B). Downloaded from Loss and gain of function of OPA1 inﬂuence mitochondrial spatial dynamics. To gain direct insight into the relevance of OPA1 expression and its isoform 5 on mitochondrial dynamics, we performed loss- and gain-of-function experiments with FT293 cells. Transfection of three commercial small interfering RNA (siRNAs) targeting all human OPA1 mRNA isoforms reduced OPA1 protein levels by 50 to 70% (Fig. 5C). OPA1 knockdown using siRNA 1 or 3 was visualized by OPA1 immunostaining (Fig. 5D, left side). The reduction of OPA1 expression resulted in partial mitochondrial clustering http://mcb.asm.org/ in GFP- and HuR-expressing cells as evaluated by Tom20 immunostaining, whereas the opposite was modestly seen in TIA1-expressing cells, which showed partial mitochondrial scattering (Fig. 5D, right side). Based on this observation, we tested the impact of OPA1 variant 5 (OPA1v5) expression on mitochondrial spatial dynamics. OPA1v5 expression was detected by Western blotting as an increase (2.2-fold) in the OPA1 short form (83 kDa) (Fig. 5E), consistent with previous results (29). We assessed mitochondrial dynamics in OPA1v5-expressing cells using the ﬂuorescence marker mitochondrial

Keima (mt-Keima) (37, 38). OPA1v5 expression promoted mitochondrial clustering on December 20, 2017 by Red de Bibliotecas del CSIC (clustering, 80%, versus network, 20%) compared to control cells (clustering, 5%, versus network, 95%) and remodeling around the perinuclear region (Fig. 5F), in agreement with previous findings (29). We also questioned whether the diminished expression of OMA1 observed in TIA-expressing cells (Fig. 3G) would result in a partial reduction in the proteolytic processing of OPA1 short forms. To verify this observation of OPA1 proteolytic processing-associated dysfunction (39), transient overexpression of OPA1v5 (Fig. 5G) or treatment with carbonyl cyanide 3-chlorophenylhydrazone (CCCP) (39)to induce mitochondrial uncoupling (Fig. 5H) in GFP-, TIA1b-, and HuR-expressing FT293 cells was carried out. The results showed that TIA1b-expressing cells had a modestly lower proteolysis-associated yield to generate OPA1 short forms than did GFP- or HuR-expressing cells (Fig. 5G and H). Collectively, these findings suggest that OPA1 expression contributes significantly to mitochondrial spatial dynamics observed in TIA1b-, TIARb-, and HuR-expressing FT293 cells. Ectopic expression of TIA proteins triggers ROS production and oxidative damage to mitochondrial DNA. Reactive oxygen species (ROS) and mitochondrial membrane potential (Δ⌿) are two parameters that are closely integrated with mitochondrial bioenergetics. To further characterize the mechanisms underlying the effects of TIA1b, TIARb, and HuR expression on mitochondria, we determined Δ⌿ using the fluorescence probe tetramethylrhodamine methylester (TMRM) (36). Whereas Δ⌿ in TIA1b- and TIARb-expressing cells was similar to that in control cells (Fig. 6A), HuR- expressing cells showed a 2-fold increase in Δ⌿ (Fig. 6A). As controls for these observations, we treated the same cells with oligomycin (an uncoupler of ATP synthase and an inhibitor of oxidative phosphorylation) and CCCP (a dissipator of Δ⌿). Increased oxidative phosphorylation is often associated with increased levels of superoxide due to electron leakage from the ETC. TIA and HuR expression, particularly HuR expression, significantly increased superoxide levels compared with those in control cells (Fig. 6B) and also significantly augmented Δ⌿ (Fig. 6A). Both parameters can be either increased or buffered with hydrogen peroxide and/or MitoQ (MQ, an antioxidant). However, TIA1b and TIARb expression moderately increased ROS levels under conditions of nonefficient mitochondrial metabolism (Fig. 2), suggesting that ROS could be a deleterious subproduct in this experimental setting. Excessive ROS leads to oxidative damage of nuclear and especially mitochondrial DNA (mtDNA), with the progressive

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 9 Carrascoso et al. Molecular and Cellular Biology Downloaded from http://mcb.asm.org/

FIG 6 High rates of reactive oxygen species production and mitochondrial DNA damage in TIA-expressing FT293 cells. (A) Estimation of mitochondrial membrane potential with 100 nM tetramethylrhodamine methylester (TMRM) and flow cytometry. Values represent means Ϯ SEM (n ϭ 3). *, P Ͻ 0.001. (B) Estimation of reactive oxygen species (ROS) with 5 ␮M CellROX Deep Red and flow cytometry. Values represent means Ϯ ϭ Ͻ SEM (n 3). *, P 0.001. In panels A and B, the final concentrations of H2O2 and MitoQ (MQ) were 1 mM, and the final concentrations of oligomycin and CCCP were 10 ␮M. In all the cases, incubation was carried out for 16 h. (C) Visualization of oxidized mitochondrial DNA by indirect ␮ immunofluorescence using an anti-8-oxo-dG antibody. Control cells were treated with 300 MH2O2 for 20 min to promote cell DNA oxidation as a positive control. Bars, 20 ␮m. on December 20, 2017 by Red de Bibliotecas del CSIC

accumulation of mutations leading to cell and mitochondrial dysfunction. To evaluate the potential deleterious effects of excess ROS on the integrity of cellular DNA in FT293cells, we used an anti-8-oxo-2=-deoxyguanosine (anti-8-oxo-dG) antibody to detect DNA damage caused by oxidative radicals (36). As a control, we induced oxidative

DNA damage with H2O2 in control cells. As shown in Fig. 6C, anti-8-oxo-dG labeling of mitochondria was readily apparent in TIA1b- and TIARb-expressing cells but not in GFP- or HuR-expressing cells (Fig. 6C). These observations suggest that the mitochondrial DNA of cells expressing either TIA1b or TIARb is extensively damaged by ROS and/or there is an inactivation of antioxidant defense; either possibility suggests a malfunction of mitochondrial metabolism. Collectively, these observations demonstrate that TIA1b- and TIARb-expressing cells present several abnormal mitochondrial parameters involv- ing excessive ROS production and extensively damaged mitochondrial DNA. Ectopic TIA expression drives mitophagy, autophagy, and apoptosis. Mitophagy selectively eliminates damaged mitochondria (40) and can be activated during starva- tion or in response to mitochondrial dysfunction (40, 41). We therefore explored whether the mitochondrial dysfunction in TIA-expressing cells triggered mitophagy. To evaluate mitophagy flux, we used the mt-Keima fluorescence probe (37, 38), which emits green fluorescence at 458 nm and at pH 7.0 in mitochondria and red fluorescence at 561 nm and at pH 4.0 in mitochondrion-lysosome fusions (37). TIA1b- and TIARb- expressing cells had a higher rate of mitophagy than GFP- and HuR-expressing cells, estimated as normalized red/green mt-Keima signals (Fig. 7A). Increasing evidence indicates that autophagy is activated in situations where cellular survival is seriously compromised (40). Therefore, we next investigated whether the ectopic expression of TIA1b, TIARb, or HuR protein triggers autophagy. During autophagy, the microtubule-associated protein light chain 3-I (LC3-I) is converted to phosphatidylethanolamine-conjugated LC3-II, which associates with autophagic vesi- cles. As shown in Fig. 7B, LC3B-II expression was higher in TIA1b- and TIARb-expressing cells than in GFP- and HuR-expressing cells. Western blotting for autophagic markers p62/SQSTM1 and ATG16␤/␣ also illustrated the activation of the autophagic response (Fig. 7B). The autophagic flux can also be measured by assessing the fusion of autophagosomes with lysosomes, classical hallmarks of autophagy, using the fluorescence probe GPF-LC3B-red fluorescent protein (RFP) (41–43). The increased expression

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 10 TIA and HuR Are Drivers of Mitochondrial Dynamics Molecular and Cellular Biology Downloaded from http://mcb.asm.org/ on December 20, 2017 by Red de Bibliotecas del CSIC

FIG 7 Sustained expression of TIA1b and TIARb triggers mitophagy, autophagy and apoptosis. (A) Mitophagic flux analysis using mitochondrial Keima probe. FT293 cells were transiently transfected with mt-Keima-expressing plasmid and visualized 4 days postinduction. Histogram shows normalized red/green mt-Keima signal (n ϭ 4). *, P Ͻ 0.05. Bars, 10 ␮m. (B) Analysis of molecular markers for autophagy by Western blotting using specific antibodies against the indicated proteins. (C) Autophagic flux analysis. FT293 cells were transiently transfected with GFP-LC3B-RFP plasmid and visualized by confocal microscopy 4 days postinduction. The histogram shows the ratio of autophagosomes and autolysosomes, estimated as yellow dots per cell and free red dots per cell, respectively (n ϭ 4). **, P Ͻ 0.01. (D) Estimation of apoptotic rates measured by flow cytometry. Values represent means Ϯ SEM (n ϭ 3). **, P Ͻ 0.01. (E) Rates of cell growth. FT293 cells were grown in culture medium containing glucose (10 mM), galactose (10 mM), or AlbuMAX I (1 mg/ml) and counted (n ϭ 3). *, P Ͻ 0.05. (F) Analysis of cell cycle phases by flow cytometry. The percentage of cells was quantified in every cell cycle phase for each experimental condition indicated. Values represent means Ϯ SEM (n ϭ 3). *, P Ͻ 0.05; **, P Ͻ 0.01. (G) Summary of cellular, mitochondrial, and molecular events associated with the ectopic expression of TIA1b, TIARb, and HuR proteins.

of LC3B-II correlated with the formation of autophagosomes (yellow dots), as shown by colocalization between autophagosomes (positive for GFP labeling, green dots) and lysosomes (positive for RFP labeling, red dots) (Fig. 7C). The autophagic flux is visualized by the fusion between autophagosomes and lysosomes to generate autolysosomes (free red dots). The formation of autophagosomes and autolysosomes was found to be higher in TIA1b- and TIARb-expressing cells than in GFP- and HuR-expressing cells (Fig. 7C). We confirmed this result using the GFP-LC3-RFP-LC3ΔG fluorescent probe (44). This probe is cleaved by endogenous ATG4 protease into equimolar amounts of GFP-LC3 and RFP-LC3ΔG; GFP-LC3 is degraded by autophagy, whereas RFP-LC3ΔG remains in the cytosol, serving as an internal control. Thus, autophagic flux can be estimated by calculating the GFP/RFP signal ratio. The results validated the increased flux in TIA1b- and TIARb-expressing cells (data not shown). We next measured cell death by flow cytometry. Early apoptosis was significantly higher in TIA1b- and TIARb-expressing cells than in GFP- and HuR-expressing cells (Fig. 7D). Analysis of cleaved poly (ADP-ribose) polymerase (PARP1) confirmed this result (Fig. 7B). Overall, these observations reveal that sustained expression of TIA proteins for 4 days drives the progressive development of apoptosis and autophagy, with mitophagy-associated phenotypes. TIA1b, TIARb, and HuR have mitochondrion-dependent opposing effects on cell proliferation. Finally, to determine the middle-term impact of TIA1b, TIARb, and HuR expression on the cell cycle, the percentage of cells in each cell cycle phase was quantified by flow cytometry. TIA1b and TIARb expression had a significant impact on cell cycle progression by delaying G1/S transition (Fig. 7E), which is consistent with a

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 11 Carrascoso et al. Molecular and Cellular Biology previous report for short-term expression (28). In contrast, sustained expression of HuR signiﬁcantly increased the transition of S and G2/M cell cycle phases (Fig. 7E). To conﬁrm this, we analyzed cell proliferation under stringent experimental conditions to determine the degree of mitochondrial dependency, by replacing glucose with either galactose or fatty acids to promote a switch from glycolysis to OXPHOS. The results showed antagonist behaviors between TIA1b, TIARb, and HuR proteins (Fig. 7F). Accordingly, HuR-expressing cells proliferated more quickly under all tested

conditions, which was signiﬁcant for mitochondrial substrates, whereas TIA1b- and Downloaded from TIARb-expressing cells were consistently compromised.

DISCUSSION We provide the first characterization of the functional antagonism between TIA1/ TIAR and HuR proteins as master regulators of mitochondrial dynamics. Expression of TIA1b or TIARb triggers significant changes to mitochondrial spatial dynamics via mitochondrial fission and clustering concomitant with the modulation of key genes involved in mitochondrial architecture, including OPA1 and MFN2, which are down- http://mcb.asm.org/ regulated, and OPA1v5 and MFF, which are upregulated. These regulatory events are accompanied by a reduction in mitochondrial respiration and the increase in mitochondrial dysfunction through elevation of ROS and oxidized mtDNA, triggering mitophagy, autophagy, and apoptosis (Fig. 7G). In contrast, ectopic expression of HuR promotes mitochondrial cristae remodeling and fusion and increases OPA1 expression and mitochondrial respiration. HuR overexpression also accelerates progression through the G2/S cell cycle phase and proliferation in a mitochondrion-dependent manner (Fig. 7G). These findings thus suggest that TIA1/TIAR and HuR proteins play antagonistic roles in the on December 20, 2017 by Red de Bibliotecas del CSIC modulation of mitochondrial activity and dynamics. Mitochondria are highly dynamic structures and frequently undergo fusion and fission (1–9, 31), which are events fundamental for the function and quality control of these organelles. Mitochondrial respiration is recognized as an indicator of mitochondrial health (2, 38). We show that TIA1b and TIARb expression is associated with a deficiency in respiratory activity, whereas HuR enhances OXPHOS. These functional alterations correlate strongly with proliferative versus quiescent phenotypes found in HuR- and TIA1/TIAR-expressing cells, respectively (Fig. 7G). OPA1 plays a key role in mitochondrial remodeling and fusion (29–31), and OPA1 protein isoforms relay instructions that help determine the morphology of mitochondria (29–31). OPA1 regulates crista junction number and stability and is the sole regulator of cristae junction width (45). Our findings show that TIA proteins posttranscriptionally downregulate OPA1 gene expression through direct regulatory events linked to splicing and translation of OPA1 and through the functional forms of OPA1 associated with proteolytic processing mediated by OMA1. In contrast, HuR upregulates OPA1 expression without affecting OMA1 expression. This opposing behavior likely contributes to the mitochondrial phenotypes of mitochondrial fission and clustering versus mitochondrial fusion and cristae remodeling observed in TIA1/TIAR- and HuR- expressing cells, respectively. Mitochondrial clustering has been observed upon mitochondrial fragmentation, loss of mitochondrial membrane potential preceding mitophagy, disruption of anterograde movement of mitochondria, and mitochondrial biogenesis (46–51). However, these events are relatively understudied. Mitochondria also respond to key stresses, including hypoxia and calcium flux, by changing their subcellular distribution and localization, including coalescing/clustering around the nucleus and altering their proximity to the endoplasmic reticulum. Mitochondrial migration in cells is modulated by events that are poorly understood (51). Our findings suggest that TIA1b- and TIARb-expressing cells may be interesting cellular models to study the molecular mechanisms triggering the spatial remodeling of mitochondria, which connect these organelles with environmental stress-associated regulatory proteins. TIA1 and TIAR facilitate splicing of pre-mRNAs by improving the selection of constitutive and atypical 5= splice sites (52–54). Many genes whose splicing is regulated

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 12 TIA and HuR Are Drivers of Mitochondrial Dynamics Molecular and Cellular Biology by TIA1/TIAR proteins contain auxiliary speciﬁc sequences close to the intron 5= splice sites composed mainly by U-rich stretches (52, 53). Our results illustrate that TIA proteins regulate alternative pre-mRNA splicing of the human OPA1 exon 4b through the gain of function of a cryptic site located on the 4b intron and close to an atypical 5= splice site characterized by three tandem GU dinucleotides, a very rare and atypical 5= splice site. We suggest that binding of TIA proteins to the identiﬁed U-rich sequence on human OPA1 intron 4b promotes the recruitment of U1 small ribonucleoparticle (U1

snRNP), to recognize and promote the processing of the first GU to properly facilitate Downloaded from the inclusion of OPA1 exon 4b to generate OPA1 variant 5 (54). This capacity is lost in a mutated version of TIA1 or TIAR lacking the Q-rich carboxy-terminal domain and in a mutant OPA1 minigene with a modified TIA-binding site, in agreement with previous results (28, 52–54). This mechanism also operates in human genes regulated by TIA proteins at the alternative splicing level, as, for example, NF1 (34) and SMN1/SMN2 (35). Our observations underscore the critical involvement of alternative splicing in regulating the expression of OPA1 RNA variants. To our knowledge, this is the first description of specific splicing factors that regulate OPA1 exon 4b splicing, possibly also exons 4 http://mcb.asm.org/ and 5b, which may help to elucidate the molecular mechanisms underlying mitochondrial dynamics and genome maintenance by providing candidate genes for mitochondrial physiopathology. Why should many mitochondrial dynamics-associated genes be regulated by TIA and/or HuR? The most likely answer is that their mRNAs or pre-mRNAs are targeted by TIA1/TIAR and HuR through one or multiple layers to exert control of their gene expression since these regulators act as multifunctional proteins (12–20). Indeed, the numbers of transcripts targeted by TIA1, TIAR, and HuR in human cells have been on December 20, 2017 by Red de Bibliotecas del CSIC estimated at 11,977, 23,854, and 55,432 transcripts, respectively, based on databases of iCLIP analysis (18, 20, 21, 23). These transcripts constitute 5 to 25% of the human genome. The relevance of this estimation is that the binding sites on human pre- mRNAs are found both in up- and downregulated genes and are located commonly in introns and in the last exons of these pre-mRNAs, particularly on the sequences located at 3= UTRs. We have identified several mitochondrial genes involved in mitochondrial dynamics that may be regulated by TIA1/TIAR and/or HuR protein (18, 20, 21, 23). OPA1 (this study), MFN1, MFN2, OMA1, YME1Ll, and MFF (33) are potential targets of TIA proteins, and OPA1, MFN1, MFN2, MFF, and FIS1 are also potential targets of HuR. Thus, TIA1/TIAR and HuR could orchestrate biological responses and programs whose targets may be linked to mitochondrial activity and dynamics. Accordingly, two mitochondrial components linked to survival/cell death and/or electron transport chain, cytochrome c and coenzyme Q, are posttranscriptionally regulated by TIA1/TIAR and/or HuR (55, 56). TIA1b- and TIARb-expressing cells present mitochondrial dysfunction associated with oxidative stress. The high production of ROS and/or failure of antioxidant defense may be the cause of the massive oxidative damage to mtDNA of TIA-expressing cells. OPA1 protein isoforms containing exon 4b (i.e., isoforms 3, 5, 6, and 8) are known to be involved in the maintenance of mitochondrial genome integrity (57). We did not detect OPA1 isoforms 3 and 6 in TIA-expressing cells, whereas the expression of isoform 8 was unchanged and isoform 5 was upregulated but apparently underprocessed by OMA1. Thus, we suggest that the downregulation of OPA1 exon 4b isoforms could impact the maintenance of mtDNA integrity and perhaps contribute to the high rates of oxidized mtDNA found in TIA-expressing cells. Mitophagy selectively degrades damaged mitochondria, which is critical for main- taining cellular homeostasis and functions (1, 40, 41). It is feasible that increasing the removal of defective mitochondria by mitophagy can lower the levels of ROS and the resultant cellular damage and its consequences in proliferating cells driven to quies- cence (58). Also, in response to mitochondrial DNA damage, mitophagy is increased in a positive loop, and we found that increasing mitophagy via TIA1 and TIAR overexpression could decrease the rates of mtDNA instability. These findings underline the relevance of mitophagy and/or autophagy to control the survival of cells with unstable mitochondrial DNA. On the other hand, we found that HuR-expressing cells are more

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 13 Carrascoso et al. Molecular and Cellular Biology efﬁcient energy producers than corresponding control cells. Recent studies with mouse models have begun to link HuR to mitochondrion-regulated processes (59–61). In summary, our data shed light on the mechanisms of TIA and HuR regulation of OPA1 gene expression to tailor mitochondrial function and dynamics, in agreement with the pivotal role of OPA1 for mitochondrial homeostasis (62–64).

MATERIALS AND METHODS Cell cultures. FT293 cell lines expressing GFP-tagged proteins, wild-type and TIA1- and TIAR- knocked-out mouse embryonic fibroblasts (MEFs), and control and TIA1-, TIAR-, and HuR-downregulated Downloaded from HeLa cells by RNA interference were generated and grown as described previously (28, 36, 65). Immunofluorescence and electron microscopy analysis. Cells were processed for immunofluorescence and transmission electron microscopy analysis as described previously (28, 36, 65). Estimation of mitochondrial DNA copy number. We performed quantitative PCR by using SYBR green and total DNA template. We used the following primers (Sigma): mtND1primers, 5=-CCCTAAAAC CCGCCACATCT-3= and 5=-GAGCGATGGTGAGAGCTAAGCT-3=; mtDloop primers, 5=-CTGTCTTTGATTCCTG CCTC-3= and 5=-TTGAGGAGGTAAGCTACAT-3=; n18S primers, 5=-ATCCATTGGAGGG CAAGTC-3= and 5=-G CTCCCAAGATCCAACTACG-3=; and nACTB primers, 5=-ATCATGTTT GAGACCTTCAAC-3= and 5=-CATCTCTT

GCTCGAAGTCCA-3=. We normalized the amount of mtDNA to the amount of the nuclear DNA (nDNA). http://mcb.asm.org/ RNA isolation, RT-PCR, qPCR, and Western blot analysis. RNAs were isolated and purified using the RNeasy kit (Qiagen). RNAs were treated with RNase-free DNase I (Promega). cDNA was synthesized with reverse transcriptase (Promega) and amplified by elongation PCR (RT-PCR) and/or quantified by real-time quantitative PCR (qPCR) analysis as described previously (28). The oligonucleotide pairs used to analyze the splicing patterns of human OPA1 by RT-PCR and qPCR have been described previously (29, 32). To identify OPA1 isoforms from RT-PCR analysis, the amplified DNA fragments were digested with appropriate restriction enzymes (32). Protein extracts from FT293 cells, MEFs, and HeLa cells were processed for Western blotting as described previously (28, 36, 65). Mitochondrial activity assay and determination of mitochondrial oxygen consumption rate.

Cellular OCR was determined using the XF24 extracellular flux analyzer (Seahorse Bioscience). Cells were on December 20, 2017 by Red de Bibliotecas del CSIC plated on XF24 microplates at 15,000/well in supplemented medium and incubated at 37°C and 5% CO2 for 24 h. After basal respiration was measured, 6 ␮M oligomycin was injected to inhibit complex V, and then 0.75 mM 2,4-dinitrophenol was injected to uncouple respiration. Finally, complexes I and III were inhibited by injection of 1 ␮M rotenone and 1 ␮M antimycin A, respectively. OCR was determined by subtracting the nonmitochondrial OCR after treatment with rotenone plus actinomycin A, whereas mitochondrial basal respiration was determined from mitochondrial OCR before administration of oligomycin. Mitochondrial maximal respiration was defined as OCR after administration of 2,4- dinitrophenol. Spare respiration capacity was defined as maximal respiration minus basal respiration. The cells shift to an almost exclusive aerobic phenotype as indicated by a low extracellular acidification rate (ECAR), and the cells shift to a more glycolytic phenotype with an average OCR equal to 20 pmol/min and an average ECAR equal to 75 mpH/min. Plasmids, transfections, and cross-linking analysis. The OPA1 minigene used for transient transfections was generated by cloning a PCR product of human OPA1 genomic DNA from exon 4 to the end of exon 5. HindIII and BamHI sequences were added to oligonucleotides to allow subcloning into pcDNA3.1 (ϩ) (Invitrogen). Specific primers for T7 promoter and OPA1 exons 4 and 5 were used for RT-PCR analysis, specifically of transcripts derived from the expression vector. Mutant OPA1 minigene was generated by site-directed PCR-based mutagenesis and confirmed by sequencing. Chimeric GFP reporters containing human OPA1 and OMA1 3= UTRs were generated by cloning PCR fragments into pEGFP-C1 vector (Clontech). Sequences were verified by sequencing. Cells were seeded in Dulbecco modified Eagle medium (DMEM) supplemented with 10% serum and antibiotics 24 h before transfection with TurboFect (Thermo Scientific). The reporter plasmids were diluted in 150 mM NaCl and incubated for 10 min at room temperature with 4 ␮l of TurboFect reagent. After incubation for 24 h, protein and RNA samples were prepared and analyzed as reported previously (28). RNA interference of human OPA1 was performed using 27-mer siRNA duplexes (SR303287; Origene) and Lipofectin reagent (Invitrogen) as reported previously (28). UV cross-linking of GST-TIA1, GST-TIAR, and maltose binding protein (MBP)-HuR recombinant proteins to U-rich and mutated sequences derived from WT and MUT minigenes was performed using RNAs uniformly labeled with 32P-uridine corresponding to final nucleotides from exon 4b and 4b intron 5= splice sites containing U-rich and mutated sequences. Cross-linking assays were carried out in the absence or presence of 50 ng/␮l of added GST-TIA1b, GST-TIARb, MBP-HuR, GST, and MBP, and the products of cross-linking were fractionated by electrophoresis on 10% SDS-PAGE gels and detected by autoradiography. Measurement of mitophagic and autophagic flux. To monitor and quantify mitophagy and autophagy, we used fluorescence probes mt-Keima (37), GFP-LC3-RFP (41–43), and GFP-LC3B-RFP- LC3BΔG (44), respectively. Molecular markers of mitophagy, autophagy, and apoptosis were validated by Western blotting using specific antibodies against LC3, p62, ATG16␤, and PARP1 (BioNova). Cell proliferation, cell cycle, and cell death analysis. Analysis of cell proliferation, cell cycle, and cell death was carried out as described previously (28, 36, 65). Statistical analysis. All data were expressed as means Ϯ standard errors of the means (SEM). Student’s t test (paired 2-tailed) was applied to determine statistical significance between 2 groups. P values of Ͻ0.05 were considered statistically significant.

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 14 TIA and HuR Are Drivers of Mitochondrial Dynamics Molecular and Cellular Biology

ACKNOWLEDGMENTS We thank J. M. Sierra and J. Satrústegui for critical reading of the manuscript and encouragement. We are indebted to the generosity and invaluable support with reagents and tips from the following researchers: D. C. Chan, T. Finkel, M. Guerra, S. C. Janga, T. Johansen, H. Lou, A. Miyawaki, N. Mizushima, M. Murphy, M. Rejas, J. Satrústegui, R. N. Singh, and J. Ule. This work was supported by grants from the Spanish Ministry of Economic Affairs

and Competitiveness (BFU2011-29653 and BFU2014-57735R) to José M. Izquierdo. The Downloaded from CBMSO receives an institutional grant from Fundación Ramón Areces. We declare that we have no competing interests. J.M.I. conceived the research. I.C. and J.M.I. designed all the experiments. I.C., J.A., C.S.-J., P.G.-S., and J.M.I. performed the experiments. J.M.I. wrote the paper. All authors provided feedback and approved the ﬁnal manuscript.

REFERENCES http://mcb.asm.org/ 1. Kauppila TE, Kauppila JH, Larsson NG. 2017. Mammalian mitochondria of a C-rich signature motif in target mRNAs of RNA-binding protein TIAR. and aging: an update. Cell Metab 25:57–71. https://doi.org/10.1016/j Mol Cell Biol 27:6806–6817. https://doi.org/10.1128/MCB.01036-07. .cmet.2016.09.017. 17. Abdelmohsen K, Lal A, Kim HH, Gorospe M. 2007. Posttranscriptional 2. Sebastián D, Palacín M, Zorzano A. 2017. Mitochondrial dynamics: cou- orchestration of an anti-apoptotic program by HuR. Cell Cycle pling mitochondrial ﬁtness with healthy aging. Trends Mol Med 23: 6:1288–1292. https://doi.org/10.4161/cc.6.11.4299. 201–215. https://doi.org/10.1016/j.molmed.2017.01.003. 18. Wang Z, Kayikci M, Briese M, Zarnack K, Luscombe NM, Rot G, Zupan B, 3. Lee H, Yoon Y. 2016. Mitochondrial ﬁssion and fusion. Biochem Soc Trans Curk T, Ule J. 2010. iCLIP predicts the dual splicing effects of TIA-RNA 44:1725–1735. https://doi.org/10.1042/BST20160129. interactions. PLoS Biol 8:e1000530. https://doi.org/10.1371/journal.pbio 4. López-Lluch G. 2017. Mitochondrial activity and dynamics changes re- .1000530.

garding metabolism in ageing and obesity. Mech Ageing Dev 162: 19. Kishore S, Jaskiewicz L, Burger L, Hausser J, Khorshid M, Zavolan M. 2011. on December 20, 2017 by Red de Bibliotecas del CSIC 108–121. https://doi.org/10.1016/j.mad.2016.12.005. A quantitative analysis of CLIP methods for identifying binding sites of 5. Flippo KH, Strack S. 2017. Mitochondrial dynamics in neuronal injury, RNA-binding proteins. Nat Methods 8:559–564. https://doi.org/10.1038/ development and plasticity. J Cell Sci 130:671–681. https://doi.org/10 nmeth.1608. .1242/jcs.171017. 20. Mukherjee N, Corcoran DL, Nusbaum JD, Reid DW, Georgiev S, Hafner M, 6. Trotta AP, Chipuk JE. 2017. Mitochondrial dynamics as regulators of Ascano M, Jr, Tuschl T, Ohler U, Keene JD. 2011. Integrative regulatory cancer biology. Cell Mol Life Sci 74:1999–2017. https://doi.org/10.1007/ mapping indicates that the RNA-binding protein HuR couples pre-mRNA s00018-016-2451-3. processing and mRNA stability. Mol Cell 43:327–339. https://doi.org/10 7. Srinivasan S, Guha M, Kashina A, Avadhani NG. 2017. Mitochondrial .1016/j.molcel.2011.06.007. dysfunction and mitochondrial dynamics. The cancer connection. 21. Lebedeva S, Jens M, Theil K, Schwanhäusser B, Selbach M, Landthaler M, Biochim Biophys Acta 2017:S0005-2728(17)30005-1. https://doi.org/10 Rajewsky N. 2011. Transcriptome-wide analysis of regulatory interactions .1016/j.bbabio.2017.01.004. of the RNA-binding protein HuR. Mol Cell 43:340–352. https://doi.org/ 8. Rovira-Llopis S, Bañuls C, Díaz-Morales N, Hernández-Mijares A, Rocha M, 10.1016/j.molcel.2011.06.008. Víctor VM. 2017. Mitochondrial dynamics in type 2 diabetes: pathophys- 22. Friedersdorf MB, Keene JD. 2014. Advancing the functional utility of iological implications. Redox Biol 11:637–645. https://doi.org/10.1016/j PAR-CLIP by quantifying background binding to mRNAs and lncRNAs. .redox.2017.01.013. Genome Biol 15:R2. https://doi.org/10.1186/gb-2014-15-1-r2. 9. Ryan MT, Hoogenraad NJ. 2007. Mitochondrial-nuclear communications. 23. Nishtala S, Neelamraju Y, Janga SC. 2016. Dissecting the expression Annu Rev Biochem 76:701–722. https://doi.org/10.1146/annurev relationships between RNA-binding proteins and their cognate targets .biochem.76.052305.091720. in eukaryotic post-transcriptional regulatory networks. Sci Rep 6:25711. 10. Matsumoto S, Uchiumi T, Tanamachi H, Saito T, Yagi M, Takazaki S, Kanki https://doi.org/10.1038/srep25711. T, Kang D. 2012. Ribonucleoprotein Y-box-binding protein-1 regulates 24. Piecyk M, Wax S, Beck AR, Kedersha N, Gupta M, Maritim B, Chen S, mitochondrial oxidative phosphorylation (OXPHOS) protein expression Gueydan C, Kruys V, Streuli M, Anderson P. 2000. TIA-1 is a translational after serum stimulation through binding to OXPHOS mRNA. Biochem J silencer that selectively regulates the expression of TNF-alpha. EMBO J 443:573–584. https://doi.org/10.1042/BJ20111728. 19:4154–4163. https://doi.org/10.1093/emboj/19.15.4154. 11. Schatton D, Pla-Martin D, Marx MC, Hansen H, Mourier A, Nemazanyy I, 25. Beck AR, Miller IJ, Anderson P, Streuli M. 1998. RNA-binding protein TIAR Pessia A, Zentis P, Corona T, Kondylis V, Barth E, Schauss AC, Velagapudi is essential for primordial germ cell development. Proc Natl Acad Sci V, Rugarli EI. 2017. CLUH regulates mitochondrial metabolism by con- U S A 95:2331–2336. https://doi.org/10.1073/pnas.95.5.2331. trolling translation and decay of target mRNAs. J Cell Biol 216:675–693. 26. Katsanou V, Milatos S, Yiakouvaki A, Sgantzis N, Kotsoni A, Alexiou M, https://doi.org/10.1083/jcb.201607019. Harokopos V, Aidinis V, Hemberger M, Kontoyiannis DL. 2009. The 12. Sánchez-Jiménez C, Izquierdo JM. 2015. T-cell intracellular antigens in RNA-binding protein Elavl1/HuR is essential for placental branching health and disease. Cell Cycle 14:2033–2043. https://doi.org/10.1080/ morphogenesis and embryonic development. Mol Cell Biol 29: 15384101.2015.1053668. 2762–2776. https://doi.org/10.1128/MCB.01393-08. 13. López de Silanes I, Zhan M, Lal A, Yang X, Gorospe M. 2004. Identiﬁcation 27. Kharraz Y, Salmand PA, Camus A, Auriol J, Gueydan C, Kruys V, Morello of a target RNA motif for RNA-binding protein HuR. Proc Natl Acad Sci D. 2010. Impaired embryonic development in mice overexpressing the U S A 101:2987–2292. https://doi.org/10.1073/pnas.0306453101. RNA-binding protein TIAR. PLoS One 5:e11352. https://doi.org/10.1371/ 14. López de Silanes I, Galbán S, Martindale JL, Yang X, Mazan-Mamczarz K, journal.pone.0011352. Indig FE, Falco G, Zhan M, Gorospe M. 2005. Identiﬁcation and functional 28. Sánchez-Jiménez C, Ludeña MD, Izquierdo JM. 2015. T-cell intracellular outcome of mRNAs associated with RNA-binding protein TIA-1. Mol Cell antigens function as tumor suppressor genes. Cell Death Dis 6:e1669. Biol 25:9520–9531. https://doi.org/10.1128/MCB.25.21.9520-9531.2005. https://doi.org/10.1038/cddis.2015.43. 15. Mazan-Mamczarz K, Lal A, Martindale JL, Kawai T, Gorospe M. 2006. 29. Song Z, Chen H, Fiket M, Alexander C, Chan DC. 2007. OPA1 processing Translational repression by RNA-binding protein TIAR. Mol Cell Biol controls mitochondrial fusion and is regulated by mRNA splicing, mem- 26:2716–2727. https://doi.org/10.1128/MCB.26.7.2716-2727.2006. brane potential, and Yme1L. J Cell Biol 178:749–755. https://doi.org/10 16. Kim HS, Kuwano Y, Zhan M, Pullmann R, Jr, Mazan-Mamczarz K, Li H, .1083/jcb.200704110. Kedersha N, Anderson P, Wilce MC, Gorospe M, Wilce JA. 2007. Elucidation 30. Olichon A, Elachouri G, Baricault L, Delettre C, Belenguer P, Lenaers G.

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 15 Carrascoso et al. Molecular and Cellular Biology

2007. OPA1 alternate splicing uncouples an evolutionary conserved and HDAC6-dependent mitophagy. J Cell Biol 189:671–679. https://doi function in mitochondrial fusion from a vertebrate restricted function in .org/10.1083/jcb.201001039. apoptosis. Cell Death Differ 14:682–692. https://doi.org/10.1038/sj.cdd 50. Gao J, Schatton D, Martinelli P, Hansen H, Pla-Martin D, Barth E, Becker .4402048. C, Altmueller J, Frommolt P, Sardiello M, Rugarli EI. 2014. CLUH regulates 31. MacVicar T, Langer T. 2016. OPA1 processing in cell death and disease— mitochondrial biogenesis by binding mRNAs of nuclear-encoded mito- the long and short of it. J Cell Sci 129:2297–2306. https://doi.org/10 chondrial proteins. J Cell Biol 207:213–223. https://doi.org/10.1083/jcb .1242/jcs.159186. .201403129. 32. Satoh M, Hamamoto T, Seo N, Kagawa Y, Endo H. 2003. Differential 51. Boland ML, Chourasia AH, MacLeod K. 2013. Mitochondrial dysfunction sublocalization of the dynamin-related protein OPA1 isoforms in mito- in cancer. Front Oncol 3:292. https://doi.org/10.3389/fonc.2013.00292. chondria. Biochem Biophys Res Commun 300:482–493. https://doi.org/ 52. Del Gatto-Konczak F, Bourgeois CF, Le Guiner C, Kister L, Gesnel MC, 10.1016/S0006-291X(02)02874-7. Stévenin J, Breathnach R. 2000. The RNA-binding protein TIA1 is a novel Downloaded from 33. Tak H, Eun JW, Kim J, Park SJ, Kim C, Ji E, Lee H, Kang H, Cho DH, Lee K, mammalian splicing regulator acting through intron sequences adjacent Kim W, Nam SW, Lee EK. 2017. T-cell-restricted intracellular antigen 1 toa5= splice site. Mol Cell Biol 20:6287–6299. https://doi.org/10.1128/ facilitates mitochondrial fragmentation by enhancing the expression of MCB.20.17.6287-6299.2000. mitochondrial fission factor. Cell Death Differ 24:49–58. https://doi.org/ 53. Förch P, Puig O, Kedersha N, Martínez C, Granneman S, Séraphin B, 10.1038/cdd.2016.90. Anderson P, Valcárcel J. 2000. The apoptosis-promoting factor TIA-1 is a 34. Zhu H, Hinman MN, Hasman RA, Mehta P, Lou H. 2008. Regulation of regulator of alternative pre-mRNA splicing. Mol Cell 6:1089–1098. neuron-specific alternative splicing of neurofibromatosis type 1 pre-mRNA. https://doi.org/10.1016/S1097-2765(00)00107-6. Mol Cell Biol 28:1240–1251. https://doi.org/10.1128/MCB.01509-07. 54. Förch P, Puig O, Martínez C, Séraphin B, Valcárcel J. 2002. The splicing 35. Singh NN, Se J, Ottesen EW, Shishimorova M, Bhattacharya D, Singh RN. regulator TIA-1 interacts with U1-C to promote U1 snRNP recruitment to 2011. TIA1 prevents skipping of a critical exon associated with spinal 5= splice sites. EMBO J 21:6882–6892. https://doi.org/10.1093/emboj/ http://mcb.asm.org/ muscular atrophy. Mol Cell Biol 31:935–954. https://doi.org/10.1128/ cdf668. MCB.00945-10. 55. Kawai T, Lal A, Yang X, Galban S, Mazan-Mamczarz K, Gorospe M. 2006. 36. Sánchez-Jiménez C, Izquierdo JM. 2013. T-cell intracellular antigen (TIA)- Translational control of cytochrome c by RNA-binding proteins TIA-1 proteins deficiency in murine embryonic fibroblasts alters cell cycle and HuR. Mol Cell Biol 26:3295–3307. https://doi.org/10.1128/MCB.26.8 progression and induces autophagy. PLoS One 8:e75127. https://doi .3295-3307.2006. .org/10.1371/journal.pone.0075127. 56. Cascajo MV, Abdelmohsen K, Noh JH, Fernández-Ayala DJ, Willers IM, Brea 37. Sun N, Yun J, Liu J, Malide D, Liu C, Rovira II, Holmström KM, Fergusson G, López-Lluch G, Valenzuela-Villatoro M, Cuezva JM, Gorospe M, Siendones MM, Yoo YH, Combs CA, Finkel T. 2015. Measuring in vivo mitophagy. E, Navas P. 2016. RNA-binding proteins regulate cell respiration and coen- Mol Cell 60:685–696. https://doi.org/10.1016/j.molcel.2015.10.009. zyme Q biosynthesis by post-transcriptional regulation of COQ7. RNA Biol

38. Williams JA, Zhao K, Jin S, Ding WX. 2017. New methods for monitoring 13:622–634. https://doi.org/10.1080/15476286.2015.1119366. on December 20, 2017 by Red de Bibliotecas del CSIC mitochondrial biogenesis and mitophagy in vitro and in vivo. Exp Biol 57. Elachouri G, Vidoni S, Zanna C, Pattyn A, Boukhaddaoui H, Gaget K, Med 242:781–787. https://doi.org/10.1177/1535370216688802. Yu-Wai-Man P, Gasparre G, Sarzi E, Delettre C, Olichon A, Loiseau D, 39. Griparic L, Kanazawa T, van der Bliek AM. 2007. Regulation of the Reynier P, Chinnery PF, Rotig A, Carelli V, Hamel CP, Rugolo M, Lenaers mitochondrial dynamin-like protein Opa1 by proteolytic cleavage. J Cell G. 2011. OPA1 links human mitochondrial genome maintenance to Biol 178:757–764. https://doi.org/10.1083/jcb.200704112. mtDNA replication and distribution. Genome Res 21:12–20. https://doi 40. Klionsky DJ, et al. 2016. Guidelines for the use and interpretation of .org/10.1101/gr.108696.110. assays for monitoring autophagy (3rd edition). Autophagy 12:1–222. 58. Genova ML, Lenaz G. 2015. The interplay between respiratory super- https://doi.org/10.1080/15548627.2015.1100356. complexes and ROS in aging. Antioxid Redox Signal 23:208–238. https:// 41. Lefebvre V, Du Q, Baird S, Ng AC, Nascimento M, Campanella M, McBride doi.org/10.1089/ars.2014.6214. HM, Screaton RA. 2013. Genome-wide RNAi screen identifies ATPase inhib- 59. Díaz-Muñoz MD, Bell SE, Fairfax K, Monzon-Casanova E, Cunningham AF, itory factor 1 (ATPIF1) as essential for PARK2 recruitment and mitophagy. Gonzalez-Porta M, Andrews SR, Bunik VI, Zarnack K, Curk T, Heggermont Autophagy 9:1770–1779. https://doi.org/10.4161/auto.25413. WA, Heymans S, Gibson GE, Kontoyiannis DL, Ule J, Turner M. 2015. The 42. Kimura S, Noda T, Yoshimori T. 2007. Dissection of the autophagosome RNA-binding protein HuR is essential for the B cell antibody response. maturation process by a novel reporter protein, tandem fluorescent- Nat Immunol 16:415–425. https://doi.org/10.1038/ni.3115. tagged LC3. Autophagy 3:452–460. https://doi.org/10.4161/auto.4451. 60. Skliris A, Papadaki O, Kafasla P, Karakasiliotis I, Hazapis O, Reczko M, 43. Pankiv S, Clausen TH, Lamark T, Brech A, Bruun JA, Outzen H, Øvervatn Grammenoudi S, Bauer J, Kontoyiannis DL. 2015. Neuroprotection re- A, Bjørkøy G, Johansen T. 2007. p62/SQSTM1 binds directly to Atg8/LC3 quires the functions of the RNA-binding protein HuR. Cell Death Differ to facilitate degradation of ubiquitinated protein aggregates by au- 22:703–718. https://doi.org/10.1038/cdd.2014.158. tophagy. J Biol Chem 282:24131–24145. https://doi.org/10.1074/jbc 61. Lu L, Wang S, Zheng L, Li X, Suswam EA, Zhang X, Wheeler CG, Nabors .M702824200. LB, Filippova N, King PH. 2009. Amyotrophic lateral sclerosis-linked 44. Kaizuka T, Morishita H, Hama Y, Tsukamoto S, Matsui T, Toyota Y, mutant SOD1 sequesters Hu antigen R (HuR) and TIA-1-related protein Kodama A, Ishihara T, Mizushima T, Mizushima N. 2016. An autophagic (TIAR): implications for impaired post-transcriptional regulation of vas- flux probe that releases an internal control. Mol Cell 64:835–849. https:// cular endothelial growth factor. J Biol Chem 284:33989–33998. https:// doi.org/10.1016/j.molcel.2016.09.037. doi.org/10.1074/jbc.M109.067918. 45. Glytsou C, Calvo E, Cogliati S, Mehrotra A, Anastasia I, Rigoni G, Raimondi 62. Civiletto G, Varanita T, Cerutti R, Gorletta T, Barbaro S, Marchet S, A, Shintani N, Loureiro M, Vazquez J, Pellegrini L, Enriquez JA, Scorrano Lamperti C, Viscomi C, Scorrano L, Zeviani M. 2015. Opa1 overexpression L, Soriano ME. 2016. Optic atrophy 1 is epistatic to the core MICOS ameliorates the phenotype of two mitochondrial disease mouse models. component MIC60 in mitochondrial cristae shape control. Cell Rep Cell Metab 21:845–854. https://doi.org/10.1016/j.cmet.2015.04.016. 17:3024–3034. https://doi.org/10.1016/j.celrep.2016.11.049. 63. Varanita T, Soriano ME, Romanello V, Zaglia T, Quintana-Cabrera R, 46. Tanaka Y, Kanai Y, Okada Y, Nonaka S, Takeda S, Harada A, Hirokawa N. Semenzato M, Menabò R, Costa V, Civiletto G, Pesce P, Viscomi C, Zeviani 1998. Targeted disruption of mouse conventional kinesin heavy chain, M, Di Lisa F, Mongillo M, Sandri M, Scorrano L. 2015. The Opa1- kif5B, results in abnormal perinuclear clustering of mitochondria. Cell dependent mitochondrial cristae remodeling pathway controls atrophic, 93:1147–1158. https://doi.org/10.1016/S0092-8674(00)81459-2. apoptotic, and ischemic tissue damage. Cell Metab 21:834–844. https:// 47. Smirnova E, Griparic L, Shurland DL, van der Bliek AM. 2001. Dynamin- doi.org/10.1016/j.cmet.2015.05.007. related protein Drp1 is required for mitochondrial division in mamma- 64. Buck MD, O’Sullivan D, Klein Geltink RI, Curtis JD, Chang CH, Sanin DE, lian cells. Mol Biol Cell 12:2245–2256. https://doi.org/10.1091/mbc.12.8 Qiu J, Kretz O, Braas D, van der Windt GJ, Chen Q, Huang SC, O’Neill CM, .2245. Edelson BT, Pearce EJ, Sesaki H, Huber TB, Rambold AS, Pearce EL. 2016. 48. Cho KI, Cai Y, Yi H, Yeh A, Aslanukov A, Ferreira PA. 2007. Association of Mitochondrial dynamics controls T cell fate through metabolic program- the kinesin-binding domain of RanBP2 to KIF5B and KIF5C determines ming. Cell 166:63–76. https://doi.org/10.1016/j.cell.2016.05.035. mitochondria localization and function. Traffic 8:1722–1735. https://doi 65. Izquierdo JM, Alcalde J, Carrascoso I, Reyes R, Ludeña MD. 2011. Knock- .org/10.1111/j.1600-0854.2007.00647.x. down of T-cell intracellular antigens triggers cell proliferation, invasion 49. Lee JY, Nagano Y, Taylor JP, Lim KL, Yao TP. 2010. Disease-causing and tumour growth. Biochem J 435:337–344. https://doi.org/10.1042/ mutations in parkin impair mitochondrial ubiquitination, aggregation, BJ20101030.

September 2017 Volume 37 Issue 17 e00174-17 mcb.asm.org 16