Mitochondrial E3 ubiquitin ligase 1 promotes brain injury by disturbing mitochondrial dynamics in a rat model of ischemic stroke

European Journal of Pharmacology 861 (2019) 172617

Contents lists available at ScienceDirect

European Journal of Pharmacology journal homepage: www.elsevier.com/locate/ejphar

Full length article

Mitochondrial E3 ubiquitin ligase 1 promotes brain injury by disturbing mitochondrial dynamics in a rat model of ischemic stroke

T

Kai-Di Rena,b,1, Wei-Ning Liub,1, Jing Tianb, Yi-Yue Zhangb, Jing-Jie Penga,b, Di Zhanga, Nian-Sheng Lib,c, Jie Yangd, Jun Pengb,c, Xiu-Ju Luoa,∗ a

Department of Laboratory Medicine, The Third Xiangya Hospital of Central South University, Changsha, 410013, China Department of Pharmacology, Xiangya School of Pharmaceutical Sciences, Central South University, Changsha, 410078, China c Hunan Provincial Key Laboratory of Cardiovascular Research, Xiangya School of Pharmaceutical Sciences, Central South University, Changsha, 410078, China d Department of Neurology, Xiangya Hospital, Central South University, Changsha, 410008, China b

A R T I C LE I N FO

A B S T R A C T

Keywords: Mitochondrial E3 ubiquitin ligase 1 Ischemic stroke Mitochondrial fission Dynamin-related protein 1 mitofusin2

Mitochondrial dysfunctions contribute to brain injury in ischemic stroke while disturbance of mitochondrial dynamics results in mitochondrial dysfunction. Mitochondrial E3 ubiquitin ligase 1 (Mul1) involves in regulation of mitochondrial fission and fusion. This study aims to explore whether Mul1 contributes to brain injury in ischemic stroke and the underlying mechanisms. First, a rat ischemic stroke model was established by middle cerebral artery occlusion (MCAO), which showed ischemic injuries (increase in neurological deficit score and infarct volume) and upregulation of Mul1 in brain tissues. Next, Mul1 siRNAs were injected intracerebroventricularly to knockdown Mul1 expression, which evidently attenuated brain injuries (decrease in neurological deficit score, infarct volume and caspase-3 activity), restored mitochondrial dynamics and functions (decreases in mitochondrial fission and cytochrome c release while increase in ATP production), and restored protein levels of dynamin-related protein 1 (Drp1, a mitochondrial fission protein) and mitofusin2 (Mfn2, a mitochondrial fusion protein) through suppressing their sumoylation and ubiquitination, respectively. Finally, PC12 cells were cultured under hypoxic condition to mimic the ischemic stroke. Consistently, knockdown of Mul1 significantly reduced hypoxic injuries (decrease in apoptosis and LDH release), restored protein levels of Drp1 and Mfn2, recovered mitochondrial dynamics and functions (decreases in mitochondrial fission, mitochondrial membrane potential, reactive oxygen species production and cytochrome c release while increase in ATP production). Based on these observations, we conclude that upregulation of Mul1 contributes to brain injury in ischemic stroke rats and disturbs mitochondrial dynamics through sumoylation of Drp1 and ubiquitination of Mfn2.

1. Introduction Stroke is one of the leading causes for death and disability around the world, and a large majority of strokes (about 80%) are ischemic strokes caused by embolic or thrombotic events that interrupt the blood supply to the brain. Depending on the loss of brain cells caused by the ischemia, ischemic stroke may show different symptoms, such as sudden numbness of face, leg and/or arm and difficulty in speaking, understanding and/or walking. Multiple mechanisms, such as oxidative stress (Bavarsad et al., 2018; Zhang et al., 2015), energy metabolic disorders and calcium overload (Narne et al., 2017), are reported to contribute to brain injury in ischemic stroke, all of which are closely related to alterations in mitochondrial function (Liu et al., 2018; Narne

et al., 2017; Yang et al., 2018). Mitochondria are highly dynamic and they maintain a balance between fusion and fission. The balance between these two processes is essential for mitochondrial homeostasis, cell stability and cell survival (Anzell et al., 2018). A set of specific mitochondrial fission proteins (such as dynamin-related protein 1, Drp1) and fusion proteins (such as mitofusin 2, Mfn2) play a key role in maintaining the balance between fusion and fission (Peng et al., 2016). It has been shown that upregulation of Drp1 or downregulation of Mfn2 can lead to increase of the fragmented mitochondria concomitant with mitochondrial dysfunction and cell death (Martorell-Riera et al., 2014; Reddy et al., 2011), and Drp1 inhibitor can reduce brain injury in rats suffered ischemic stroke (Flippo et al., 2018; Zuo et al., 2016).

∗

Corresponding author. E-mail address: [email protected] (X.-J. Luo). 1 These authors contributed equally to this work. https://doi.org/10.1016/j.ejphar.2019.172617 Received 9 May 2019; Received in revised form 15 August 2019; Accepted 16 August 2019 Available online 17 August 2019 0014-2999/ © 2019 Published by Elsevier B.V.

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

into 4 groups (n = 12 per group): (1) the sham group; (2) the stroke group; (3) the stroke + Mul1 siRNA group, animals received the injection of Mul1 siRNA 24 h before ischemia; and (4) the stroke + siRNA negative control (siRNA NC) group, animals received the injection of scrambled siRNAs 24 h before ischemia. After 24 h of reperfusion, neurological deficit score was assessed, and then the brain tissues of 6 rats from each group were saved for infarct volume measurement, whereas the brain tissues (dissected from ischemic boundary area) of the remaining 6 rats from each group were collected for morphological, biochemical and molecular analysis.

Mitochondrial E3 ubiquitin ligase 1 (Mul1) is a multifunctional mitochondrial membrane protein, and it possesses dual functions of ubiquitination and sumoylation (Braschi et al., 2009; Sugiura et al., 2014). Mul1 can ubiquitinate a variety of signal molecules through its RING finger domain as an ubiquitin-ligase, such as mitofusin2 (Mfn2), p53, Akt and ULK1 (Bae et al., 2012; Jung et al., 2011; Li et al., 2015; Peng et al., 2016). More specifically, Mul1-mediated ubiquitination of Mfn2 can cause degradation of Mfn2 and enhance mitochondrial fission (Yun et al., 2014). Meanwhile, as a small ubiquitin-like modifiers (SUMO) E3 ligase, Mul1 can sumoylate Drp1 and stabilize it, subsequently strengthening mitochondrial fission (Prudent et al., 2015). Through the dual functions of ubiquitination and sumoylation, Mul1 participates in regulation of multiple pathophysiological processes, such as mitochondrial dynamics, apoptosis, and mitophagy (Peng et al., 2016). Based on the key role of Drp1 and Mfn2 in controlling the balance of mitochondrial dynamics as well as the dual functions of Mul1, we hypothesize that Mul1 contributes to brain injury in ischemic stroke through targeting Drp1 (via sumoylation) and/or Mfn2 (via ubiquitination). The aims of this study were to: (1) evaluate the relationship between Mul1 and brain injury by using a rat model of ischemic stroke or a cell model of hypoxic injury in vitro; (2) explore whether Mul1 contributes to disturbance of mitochondrial dynamics and mitochondrial dysfunction; and (3) investigate whether Mul1 enhances mitochondrial fission through sumoylation of Drp1 and/or ubiquitination of Mfn2.

2.4. Assessment of neurological deficit score and infarct volume Rat focal cerebral ischemic injury was evaluated by an investigator blinded to the experimental groups according to a five-point neurological deficit score (0 = no deficit, 1 = failed to extent the left forepaw, 2 = decreased grip strength of left forepaw, 3 = circling to left by pulling the tail, 4 = spontaneous circling) (Schabitz et al., 2004). The infarct volume was evaluated by 2, 3, 5-triphenyltetrazolium chloride (TTC) staining. After neurological function assessment, the animals were euthanized under anesthesia. Brains were rapidly removed and sliced into coronal sections (0.2–0.3 cm thickness) with the aid of a brain matrix. Sections were incubated with 2% TTC for 10 min at 37 °C, fixed in 4% paraformaldehyde overnight, scanned into a computer and analyzed with an imaging software (Image J, NIH, USA). The absence or presence of infarction was determined by examining TTC stain. The infarct volume (in cm3) of each section was equal to infarct area (in mm2) multiply by the section thickness (0.2–0.3 cm). The total infarct volume of each brain was then calculated by summing up the infarct volumes of all sections. To eliminate the effect of edema on the accuracy of infarct volume assay, the final infarct volume was corrected by following equation: corrected infarct volume = total infarct volume (right hemisphere volume/left hemisphere volume). Left hemisphere refers to no-ischemic hemisphere of brain while right hemisphere refers to ischemic ipsilateral side (see definition in Fig. 1B).

2. Materials and methods 2.1. Animals All experimental procedures for animals were carried out in accordance with the Guidelines for the Care and Use of Laboratory Animals, published by the National Institutes of Health (NIH Publication, 8th edition, 2011), and the ARRIVE guidelines (Animal Research: Reporting In Vivo Experiments). The animal experiments were approved by the Central South University Veterinary Medicine Animal Care and Use Committee.

2.5. Protocols for cell experiments 2.2. Establishment of rat ischemic stroke model The well-differentiated PC12 cells (Chinese Academy of Sciences, Shanghai, China) were seeded at a density of 1 × 104 cells/cm2 and cultured in DMEM with 10% FBS. According to the instruction offered by the cell supplier, the differentiation of PC12 cells was induced by nerve growth factor. The cultures were maintained at 37 °C in 95% air/ 5% CO2 in a humidified incubator. Under such conditions, cells grew well and exhibited spindle-shaped cell morphology similar to neuronal cells, which was consistent with morphological characteristics of welldifferentiated PC12 cells. To render cell quiescent, cells were kept in serum-free DMEM for 16 h before the experiments. To verify the role of Mul1 in ischemic stroke, the PC12 cells from passages 5 to 8 were cultured in a medium without glucose under hypoxic condition (N2/CO2, 95:5), which mimics the pathological conditions of cerebral ischemia and is widely used as a neuronal cell model of ischemic stroke (Cui et al., 2017). After 8 h of hypoxia-treatment (Huang et al., 2014), the cells were switched to normoxic condition for 24 h of reoxygenation in a medium containing glucose. Cells were divided into 4 groups (6 individual experiments per group): (1) The control group, cells were cultured under normoxic condition; (2) The hypoxia/reoxygenation (H/R) group, cells were cultured under hypoxic condition for 8 h followed by 24 h reoxygenation; (3) The hypoxia/reoxygenation plus Mul1 siRNA group, cells transfected with Mul1 siRNAs and subjected to hypoxia/reoxygenation; and (4) The hypoxia/ reoxygenation plus siRNA NC group, cells transfected with scrambled siRNAs and subjected to H/R. At the end of the experiments, cells and culture mediums were collected for analysis of cell apoptosis, LDH release, reactive oxygen

Male Sprague-Dawley (SD) rats weighing 230–280 g were fasted for 24 h before the experiments, and they were free access to tap water. A rat model of ischemic stroke was established by middle cerebral artery occlusion (MCAO), as we described previously (Fu et al., 2014). Briefly, under condition of anesthesia (sodium pentobarbital, 60 mg/kg, i.p.), the left common carotid artery (CCA) and the external carotid artery (ECA) were surgically exposed. CCA was clipped with artery clamp, while ECA was ligatured. A nylon suture with a blunted tip (0.40 mm diameter) was inserted into internal carotid artery (ICA) through a tiny incision in ECA. The nylon suture occluded the middle cerebral artery (MCA) at the position of 18–20 mm distal from a carotid bifurcation. After occlusion of MCA (ischemia) for 2 h, the nylon suture was removed for reperfusion. Rats from the sham group were subjected to the same procedure except that the nylon suture was not inserted. 2.3. Protocols for animal experiments The first set of animal experiments was utilized to investigate the expression of Mul1 after stroke. The animals were randomly divided into 3 groups (n = 12 per group): (1) the control group, no surgery for rats; (2) the sham group, rats with surgical procedures but without ischemic insult; and (3) the stroke group, the rats were subjected to 2 h of ischemia followed by 24 h of reperfusion. In the second set of animal experiments, Mul1 expression was knocked down by siRNA to evaluate the relationship between the upregulation of Mul1 and brain injury in stroke rats. Animals were divided 2

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

Fig. 1. Upregulation of Mul1 in ischemic rat brains. A shows significantly increased neurological deficit score in the stroke group compared with sham group (n = 12 per group). B presents the representative images of triphenyltetrazolium chloride-stained brain tissues (left panel), showing increased infarct volume in the stroke group compared with sham group (right panel, n = 6 per group). C indicates remarkably increased Mul1 mRNA level in brain tissues in the stroke group compared with sham group (n = 6 per group). D reveals evidently increased Mul1 protein level in brain tissues in the stroke group compared with sham group. Top: optical density for the protein blot of Mul1 vs β-actin, which are further normalized by the control; bottom, representative images of Western blots (n = 3 per group). All values are expressed as means ± S.E.M. **P < 0.01 vs Sham.

follow-up experiments. The specificity or efficiency of gene knockdown was evaluated by real-time PCR and Western blot (Fig. S1B). The negative control cells were transfected with scrambled siRNAs.

species level, ATP production, mitochondrial membrane potential (ΔΨm) and mitochondrial fission.

2.6. Mul1 knockdown in vivo and in vitro 2.7. Observation of mitochondrial morphology RiboBio genOFF™ siRNAs against rat Mul1 were purchased from Guangzhou RiboBio Co., Ltd (Guangzhou, China) for in vivo or in vitro studies. RiboBio genOFF™ in vivo siRNAs possess advantages of high stability, high efficiency, low toxicity and easy utility due to the chemically modified duplexes. The siRNAs were delivered into the brain tissues according to the method described in the literatures with slight modification (Chen et al., 2009; Yan et al., 2017). The stereotaxic coordinates were 1.5 mm lateral, 0.8 mm posterior to the bregma over the left hemisphere, and 4.5 mm ventral to the surface of the skull. Ten μg/ 10 μL Mul1siRNA or negative control siRNA was mixed gently with 5 μL of EntransterTM-in vivo transfection reagent (Engreen, Bei-Jing, China). The mixture was injected intracerebroventricularly (i.c.v.) under anesthesia by using a microsyringe (Hamilton, Nevada, USA) under the guidance of the stereotaxy instrument (Kent Scientific Corporation, CT, USA). The specificity and efficiency of gene knockdown was evaluated by real-time PCR and Western blot (Fig. S1A). The negative control rats were treated with the same quantity of scrambled siRNAs. To knockdown Mul1 gene expression in vitro, PC12 cells were transfected with siRNA following the manufacturer's instruction. Briefly, the mixture of 100 pmol of siRNAs, 5.0 μl of Lipofectamine™ 2000(Invitrogen, Carlsbad, CA) and 500 μl of DMEM were incubated at room temperature for 15–20 min. Then the mixture was transferred to the plates for incubation at 37 °C and the culture medium was replaced with fresh growth medium containing fetal bovine serum 6 h later. The cells were harvested 24 h after the transfection and subjected to the

The mitochondria in brain cells were observed by transmission electron microscope (TEM). Brain tissues (ischemic penumbra) were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) for 2 h at 4 °C and post-fixed with 1% osmium tetroxide on ice for 1 h. Tissues were then embedded in Epon 12 after dehydration in a graded acetone series (up to 100%). The sections were obtained with a Reichert ultramicrotome (Leica, Wetzlar, Germany) and stained with 3% uranyl acetate and lead citrate. Then, the sections were visualized under transmission electron microcopy (TEM) (Technai G2 Spirit TWIN, FEI, USA) at 80 kV. Images (7 per animal) were randomly taken at a magnification of 10,000 × , which were chosen for quantification of mitochondrial number and size. Mitochondrial numbers, total or average mitochondrial area per cell were analyzed blindly using Image J as previously described (Cui et al., 2017). The mitochondrial fission in PC12 cells was detected by MitoTracker Green assay (Beyotime Co. LTD, Jiangsu, China) according to the manufacturers’ instructions. Briefly, cells were loaded with Mitotracker Green and imaged with fluorescence microscopy (Olympus IX71, Tokyo, Japan) to assess the mitochondrial fission rate with excitation set at 490 nm and emission set at 516 nm. The images were taken under a 40 × objective lens. The extent of mitochondrial fission was analyzed on a cell-to-cell basis. A punctiform mitochondrial phenotype was scored as fragmented mitochondrion when at least 90% of its tubular mitochondria were disintegrated. At least 100 randomly selected cells in multiple fields were assessed and scored. The rate of 3

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

Table 1 Primers for real-time PCR. Gene

Forward Primer

Reverse Primer

Product Size (bp)

Mul1 Drp1 Mfn2 β-actin

5′-TCACTCGGTCAGGTCATCCT-3′ 5′- CGCTGATCCCGGTCATCAAT-3′ 5’ -GACTGGATTGTGCCGATGAC-3’ 5’-CCCATCTATGAGGGTTACGC-3’

5′-CCACAAACTGGCTGTTGAGC-3′ 5′- TCAACTCCATTTTCTTCTCCTGT-3′ 5’-CAGAAGAGGAGGAGGCTTGA-3′, 5′-TTTAATGTCACGCACGATTTC-3’

232 250 100 150

Mul1: mitochondrial E3 ubiquitin protein ligase 1; Drp1: dynamin 1-like; Mfn2: mitofusion 2.

mitochondrial fission was expressed as the percentage of fragmented mitochondria to the total number of cells (Aung et al., 2017).

2.11. Western blot analysis Brain tissue (ischemic penumbra) or PC12 cells were homogenized in ice-cold lysis buffer and the protein concentration in homogenate was measured by using a BCA Protein Assay kit (Beyotime, Nanjing, China). Western blot was performed according to standard procedures. In brief, samples containing ~40 μg of proteins proceeded 10% SDSPAGE and then the proteins were transferred to polyvinylidene fluoride membranes, which were incubated with primary antibodies against Mul1 (Proteintech, Chicago, USA), Drp1(Abcam, Cambridge, USA), Mfn2(Abcam, Cambridge, USA), Sumo1(Proteintech, Chicago, USA), Ub (Boster, Wuhan, China), CytC (Boster, Wuhan, China), VDAC1(Proteintech, Chicago, USA) or β-actin (Boster, Wuhan, China), followed by HRP-conjugated secondary antibodies. The signals were detected by Luminata Creseendo Western HRP substrate through Molecular Imager ChemiDoc XRS System (Bio-Rad, Philadelphia, PA). Densitometric quantification was performed with Image J (NIH, USA).β-actin (Beyotime, Shanghai, China) served as a loading control for total proteins while VDAC1 served as a loading control for mitochondrial proteins. Arbitrary optical density units of the targeting protein were normalized against control and expressed as fold change.

2.8. Measurement of mRNA expression Real-time PCR was utilized to quantify Mul1, Drp1 and Mfn2 mRNA expression [in brain tissue (ischemic penumbra) and PC12 cell]. Total RNA was isolated using TRIzol reagent (TakaRa, Dalian, China). The concentration and purity of RNA were determined spectrophotometrically. Five hundred ng of RNA was subjected to reverse transcription reaction by a transcription Kit according to the manufacturer's instructions (TaKaRa, Dalian, China). Quantitative PCR was carried out using SYBR Premix Ex Taq (TaKaRa Dalian, China) and the ABI 7300 Real-time PCR system. Briefly, a 10-μl reaction mixture containing 2 μl cDNA template, 5 μl SYBR Master mix, 0.20 μl ROX, 2.4 H2O, and 0.20 μl of each primer was amplified by the following thermal parameters: an initial incubation at 95 °C for 15 s, followed by 40 cycles of denaturation at 95 °C for 5 s, annealing and extension at 60 °C for 31 s. The PCR primers for Mul1, DRP1, Mfn2 and β-actin were listed in Table 1. Data analysis was performed by comparative Ct method using the ABI software. β-Actin served as an internal control. The mRNA levels of Mul1, DRP1 and Mfn2 were normalized against β-actin and expressed as fold of control.

2.12. Measurement of cellular apoptosis Flow cytometry was used to evaluate apoptosis in cell experiments. Briefly, PC12 cells were trypsinized and then centrifuged at 800 g for 5 min. After washing with PBS, the cells were re-suspended in 195 μl of binding buffer. Next 5 μl of FITC-conjugated AnnexinV and 10 μl of PI (propidium iodide) were added to the cells and cells were incubated at room temperature for 15 min in the dark. The cells were then analyzed for apoptosis rate by flow cytometry.

2.9. Co-immunoprecipitation (Co-IP) assay To explore the levels of ubiquitination or sumoylation, Co-IP of Mfn2 or DRP1 with ubiquitin (Mfn2) or sumo1(Drp1) was performed according to the instruction provided by the supplier (Thermo Fisher, Waltham, MA, USA). Brain tissues were rinsed once with ice-cold phosphate-buffered saline (PBS) and then lysed in IP buffer supplemented with protease inhibitors. Briefly, the supernatants of brain tissue were immunoprecipitated with anti-RIP-DRP1 antibody (Abcam, Cambridge, USA) or anti-Mfn2antibody (Abcam, Cambridge, USA) or IgG antibody (Beyotime, Shanghai, China) overnight at 4 °C, and then incubated with Protein A/G Agarose beads (Fast Flow, Beyotime alloqute) for another 3 h at 4 °C. Agarose beads were washed with ice-cold PBS followed by elution of bound proteins. Precipitated proteins were determined by Western bolt with anti-Ub (Boster, Wuhan, China) or anti-sumo1 (Proteintech, Chicago, USA) antibody. Rat IgG was used as negative control. The samples in the input group only proceeded Western blot and served as positive controls.

2.13. Measurement of caspase-3 activity and lactate dehydrogenase (LDH) release The lysates from brain tissues or PC12 cells were incubated with caspase-3 substrate (Ac-DEVD-pNA) according to the manufacturer's instruction (Beyotime, Shanghai, China) and the absorbance was determined at 405 nm. One unit of the enzyme activity was defined as the amount of enzyme required to cleave 1.0 nmol of Ac-DEVD-pNA per hour at 37 °C. The caspase-3 activities were presented as % of control, which was set at 100%. Lactate dehydrogenase (LDH) is an enzyme with stable activity in cells, and can be released to extracellular matrix when cell membrane is damaged. Thus, the release of LDH is commonly used as an important indicator for cellular necrosis. Culture medium was collected for analysis of LDH release (an indicator of cellular necrosis) by using a colorimetric assay kit (Beyotime, Jiangsu, China) following the manufacturerʼs instructions. Released LDH was measured with a coupled enzymatic reaction that resulted in the conversion of a tetrazolium salt into a red color formazan by diaphorase. In brief, 120 μl of culture medium was mixed with 60 μl of LDH work solution and they were incubated at 25–30 °C for 30 min. The absorbance was measured at 490 nm. The percentage of LDH release was calculated following a formula provided by the manufacturer.

2.10. Subcellular fractionation of brain tissues and PC12 cells The cytosolic and mitochondrial fractions of brain tissues or PC12 cells were separated by commercial kits following the instruction (Beyotime, Nanjing, China). Briefly, the brain tissues or PC12 cells were homogenized in cold cytosol extraction buffer mix with a homogenizer for 10 min on ice. The homogenate was centrifuged at 700 g for 10 min at 4 °C. Then the supernatant was collected followed by a centrifugation at 15000g for 15 min (4 °C). The supernatant was saved as cytosolic fraction while the pellet was re-suspended with the mitochondrial extraction buffer mix and saved as mitochondrial fraction after vortexing for 10s. 4

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

rats compared to the sham (0 Vs 2) (Fig. 1A). Compared with the shamoperated rats, the infarct volume of ischemic brain was notably enhanced in the stroke rats (0 Vs 0.35 ± 0.03 cm3) concomitant with upregulation of Mul1 [mRNA (1.00 ± 0.07 Vs 1.69 ± 0.12) and protein (1.00 ± 0.03 Vs 1.87 ± 0.17)] in the brain tissue (Fig. 1, B-D), while there were no significant differences in both infarct volume and Mul1 expression between the control rats and the sham-operated rats. In the following experiments, we thus kept only the sham-operated rats as the control.

2.14. Assay for mitochondrial membrane potential Mitochondrial membrane potential (ΔΨm) was assayed by using a JC-1 kit (Beyotime, Shanghai, China). The membrane-permeant JC-1 dye exhibits potential-dependent accumulation in mitochondria. JC-1 dyes gather in the mitochondrial matrix in the formation of polymer and produce red fluorescence when ΔΨm is high, but they cannot accumulate in the mitochondrial matrix when ΔΨm is low and exhibit green fluorescence. Therefore, the potential-sensitive color shift is due to concentration-dependent formation of red fluorescent J-aggregates. To measure ΔΨm, PC12 cells were seeded in six-well plates and incubated with 2 mL of culture medium overnight, 1 mL JC-1 dye and 1 mL culture medium were mixed, and then the plates were incubated at 37 °C for 20 min. The cells were washed twice and visualized by a fluorescence microscope (Olympus IX71, Tokyo, Japan) with excitation set at 514 nm and emission set at 529 nm. The images were taken under a 20 × objective lens. Healthy cells emitted red-orange fluorescence in JC-1 aggregates while the injured cells emitted green fluorescence due to the presence of JC-1 monomers in the cytoplasm.

3.2. Knockdown of Mul1 attenuated brain injury in stroke rats To explore the correlation between the up-regulation of Mul1 and brain injury in stroke rats, Mul1 was knocked down by siRNA in vivo. As shown in Fig. S1, injection of Mul1 siRNA into left lateral ventricle significantly suppressed Mul1mRNA (left panel) (1.00 ± 0.09 Vs 0.43 ± 0.05) and protein (right panel) (1.00 Vs 0.48 ± 0.07) expression in the brain tissue 48 h after injection. The scrambled siRNAs (siRNA negative control) demonstrated negative results as expected, indicating that Mul1 expression was successfully knocked down in vivo. Twenty 4 h after injection of Mul1 siRNA or the scrambled siRNAs, the rats were subjected to MCAO and 24 h of reperfusion. Compared to the stroke rats, the neurological function was markedly restored (2 Vs 1), the infarct volume (0.35 ± 0.03 cm3 Vs 0.04 ± 0.01 cm3) and caspase-3 activity (212.64 ± 7.63 Vs 156.32 ± 20.32) in the brain tissue were obviously decreased in the stroke rats received Mul1 siRNAs, but those received scrambled siRNAs did not show such effects (Fig. 2).

2.15. Measurement of ATP production ATP production was determined by using a firefly luciferase ATP assay kit (Beyotime, Shanghai, China) according to the manufacturer's instructions. Briefly, PC12 cells were schizolysised and centrifuged at 12,000×g for 5 min, and then 100 μl of supernatant was mixed with 100 μl of ATP detection working dilution. The luminescent signal was recorded by a high sensitivity luminometer (Sirius L Tube Luminometer, Pforzheim, Germany). A standard curve was used to calculate the cellular ATP production, which was normalized by protein content in each sample (nmol/mg protein).

3.3. Knockdown of Mul1 restored mitochondrial dynamics and functions in stroke rats

2.16. Determination of reactive oxygen species levels

To determine whether Mul1 involves in regulation of mitochondrial dynamics in stroke rats, the ultrastructure of mitochondria in cortex neuron was observed via electron microscopy (EM). As shown in Fig. 3A and B, compared with the sham-operated rats, the mitochondrial number was increased (11.75 ± 0.85 Vs 20.25 ± 1.18) while the average mitochondrial area was decreased in the stroke rats (0.24 ± 0.01 nm2 Vs 0.11 ± 0.01 nm2) though the total mitochondrial area was not changed (2.82 ± 0.15 nm2Vs 3.17 ± 0.16 nm2), indicating an enhancement of mitochondrial fragmentation; these phenomena were reversed in the presence Mul1 siRNAs. Consistent with the disturbance of mitochondrial dynamics, cytochrome c (Cyt c) release from mitochondria (1.00 Vs 0.48 ± 0.02) to cytosol (1.00 Vs 2.25 ± 0.12) was increased while the ATP production was suppressed (6.14 ± 1.21 nmol/mg.protein Vs 1.06 ± 0.16 nmol/mg.protein) in the stroke rat brain compared to those in the control (Fig. 3C and D), suggesting a dysfunction of mitochondria. Knockdown of Mul1significantly restored the mitochondrial functions. The siRNA negative control did not affect mitochondrial dynamics and mitochondrial function in the stroke rat brains.

The measurement of intracellular reactive oxygen species levels was based on the fluorescent signal of 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA), a cell-permeable indicator of reactive oxygen species (Beyotime, Jiangsu, China). DCFH-DA is nonfluorescent until the acetate groups are removed by intracellular reactive oxygen species. Briefly, PC12 cells were washed with PBS and incubated with DCFH-DA (10 μM) at 37 °C for 20 min. Then, the reactive oxygen species-mediated fluorescence was observed under a fluorescent microscope with excitation set at 502 nm and emission set at 523 nm. The images were taken under a 20 × objective lens. Arbitrary fluorescent units were normalized against control and expressed as fold change. 2.17. Statistical analysis SPSS software (Version 20.0) was used for statistical analysis. The results were presented as mean ± S.E.M.. Differences in measured values among multiple groups were analyzed by the analysis of variance with One-way ANOVA and Student-Newman-Keuls multiple comparison tests. The neurological deficit scores were analyzed by Kruscal-Wallis H and Wilcoxon tests. Differences were considered as significant when P < 0.05.

3.4. Knockdown of Mul1 restored the protein levels of Drp1 and Mfn2 in stroke rats As shown in Fig. 4A, injection of Mul1 siRNA (but not the scrambled siRNAs) into left lateral ventricle successfully suppressed Mul1 protein expression in the stroke rats compared with the sham-operated rats (1.00 Vs 1.69 ± 0.12). Since Drp1 plays a key role in regulation of mitochondrial fission, we therefore first evaluated the relationship between Mul1 and Drp1. As shown in Fig. 4, B-D, both total (in the whole cell lysates) (1.00 Vs 1.67 ± 0.04) and mitochondrial (in mitochondria-enriched fractions) (1.00 Vs 2.68 ± 0.30) Drp1 protein levels were significantly elevated in the stroke rats compared with sham-operated rats concomitant with a decrease in cytoplasmic Drp1 (1.00 Vs 0.55 ± 0.05); these phenomena were reversed in Mul1 siRNA-treated

3. Results 3.1. Up-regulation of Mul1 in rat brains following ischemic stroke A rat model of ischemic stroke was established to investigate the expression of Mul1 in brain tissue after ischemia. Based on a 5-point rating scale of neurological deficit score, a method for evaluating the neurological function, the disturbance of neurological function was not seen in both the sham-operated and the normal control rats, while the neurological deficit scores were dramatically increased in the stroke 5

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

Fig. 2. Knockdown of Mul1 attenuated brain injury in stroke rats. A shows significantly decreased neurological deficit score in the Mul1 knockdown group compared with the stroke group (n = 12 per group). B presents the representative images of triphenyltetrazolium chloride-stained brain tissues, showing evidently decreased infarct volume in the Mul1 knockdown group compared with the stroke group (C, n = 6 per group). D reveals remarkedly decreased caspase-3 activity in the Mul1 knockdown group compared with the stroke group (n = 6 per group). All values are expressed as means ± S.E.M. +Mul1 siRNA: Stroke + Mul1 siRNA; +siRNA NC: Stroke + siRNA negative control. **P < 0.01 vs Sham; ##P < 0.01 vs Stroke.

As expected, the percentage of apoptotic cells in the H/R group was dramatically increased compared with that in the control group (6.42 ± 0.70 Vs 35.45 ± 1.13), which was abrogated by Mul1 siRNAs (35.45 ± 1.13 Vs 13.86 ± 0.95) (Fig. 5A and B). Consistent with the results of flow cytometry, both caspase-3 activity (100.0 Vs 330.4 ± 64.2) and lactate dehydrogenase (LDH) release (8.92 ± 0.14 Vs 37.79 ± 2.27) were also obviously increased compared with the control, and these increases were abolished in the presence of Mul1 siRNA (Fig. 5C and D). The scrambled siRNAs did not exhibit such effects on apoptosis, caspase-3 activity and LDH release.

stroke rats but not in scrambled siRNAs-treated stroke rats, suggesting that Mul1 might be able to stabilize Drp1 and recruit it from the cytoplasm to the mitochondria. Since Mul1 can stabilize certain proteins through its function of sumoylation and Drp1 was reported as a substrate for Mul1, we examined the sumoylation of Drp1. As displayed in Fig. 4E, Drp1 sumoylation was enhanced in the stroke rat brain, while this phenomenon was attenuated in the presence of Mul1 siRNA but not the scrambled siRNAs. Since Mfn2 plays a critical role in regulation of mitochondrial fusion, next we examined the relationship between Mul1 and Mfn2. As shown in Fig. 4F and G, there was an evident decrease in total (1.00 Vs 0.46 ± 0.11) and mitochondrial(1.00 Vs 0.47 ± 0.07) Mfn2 protein levels in the stroke rats compared with the sham-operated rats; these decreases were blocked in Mul1 siRNA-treated rats but not in scrambled siRNAs-treated rats. Since Mul1 can also destabilize certain proteins through its function of ubiquitination and Mfn2 was a reported substrate for Mul1, we also examined the ubiquitination of Mfn2. As shown in Fig. 4H, Mfn2 ubiquitination was evidently strengthened in the stroke rat brain, which was attenuated in the presence of Mul1 siRNA. The scrambled siRNAs had no such effect.

3.6. Knockdown of Mul1 restored the protein levels of Drp1 and Mfn2 in hypoxia-treated PC12 cells As shown in Fig. 6A, transfection of Mul1 siRNA (but not the scrambled siRNAs) into PC12 cells successfully suppressed hypoxia-induced Mul1 protein expression compared with the control (1.00 Vs 1.58 ± 0.05). In agreement with the results from the in vivo experiments, total Drp1 protein levels (1.00 Vs 2.43 ± 0.21) as well as mitochondrial Drp1 protein levels (1.00 Vs 2.20 ± 0.29) were significantly elevated in the hypoxia-treated PC12 cells compared with the control concomitant with a decrease in cytoplasmic Drp1 (1.00 Vs 0.52 ± 0.07) (Fig. 6B–D); these phenomena were reversed in the presence of Mul1 siRNAs but not scrambled siRNAs. As expected, the total (1.00 Vs 0.37 ± 0.04) and mitochondrial (1.00 Vs 0.29 ± 0.09) Mfn2 protein levels in hypoxia-treated PC12 cells were markedly decreased compared with the control (Fig. 6E and F); these decreases were blocked in the presence of Mul1 siRNA. The scrambled siRNAs did not show such effect.

3.5. Knockdown of Mul1 reduced hypoxic injury in PC12 cells To verify our findings in the rat model of ischemic stroke, PC12 cells were subjected to hypoxia treatment to mimic the ischemia in vivo. As shown in Fig. S1B, transfection of Mul1 siRNA (but not the scrambled siRNAs) into PC12 cells successfully suppressed Mul1 mRNA (1.00 ± 0.06 Vs 0.45 ± 0.04) and protein (1.00 Vs 0.38 ± 0.97) expressioncompared with the control. 6

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

Fig. 3. Knockdown of Mul1 restored mitochondrial dynamics and functions in stroke rats. A presents the representative images of transmission electron microscope (mitochondria were indicated by white arrow). B shows a significant decrease in the mitochondrial number (top), no change in total mitochondrial area (middle) and an evident increase in average mitochondrial area per cell (bottom, n = 3 per group) in the Mul1 knockdown group compared with the stroke group. C reveals an evident decrease in cytochrome c release from mitochondria (left panel) to cytoplasma (right panel) in the Mul1 knockdown group compared with the stroke group. Top: optical density for the protein blot of Cyt c vs VDAC1 (mitochondrial protein marker) or β-actin, which are further normalized by the sham; bottom, representative images for Western blot (n = 6 per group). D shows significantly increased ATP levels in the Mul1 knockdown group compared with the stroke group (n = 6 per group). All values are expressed as means ± S.E.M. +Mul1 siRNA: Stroke + Mul1 siRNA; +siRNA NC: Stroke + siRNA negative control. **P < 0.01 vs Sham; ##P < 0.01 vs Stroke.

increase in reactive oxygen species level (1.00 Vs 4.07 ± 0.11) (Fig. 7C), while a decrease in ATP production (5.46 ± 0.10 nmol/ mg.protein Vs 3.23 ± 0.07 nmol/mg.protein) (Fig. 7D). Moreover, the release of Cyt c from mitochondria (1.00 Vs 0.30 ± 0.05) to cytoplasma (1.00 Vs 2.25 ± 0.12) was also enhanced compared with the control (Fig. 7E and F), confirming the dysfunctions of mitochondria. The mitochondrial functions were restored in the presence of Mul1 siRNAs. The scrambled siRNAs did not exhibit such effect. To the best of our knowledge, this is the first study to provide evidence that upregulation of Mul1 contributes to cerebral I/R (H/R) injury.

3.7. Knockdown of Mul1 restored mitochondrial dynamics and functions in hypoxia-treated PC12 cells Consistent with EM results, Mito-Tracker Green staining showed that mitochondrial fission in hypoxia-treated PC12 cells was significantly enhanced compared with that in the control cells due to the increase in the number of fragmented mitochondria (8.95 ± 1.03 Vs 51.31 ± 1.85); these phenomena were reversed in the presence of Mul1 siRNAs but not the scrambled siRNAs (Fig. 7A). In agreement with the disturbance of mitochondrial dynamics, the green signal in hypoxiatreated cell was much stronger than that in the control cells (Fig. 7B), indicating that ΔΨm was markedly reduced, accompanied by an 7

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

Fig. 4. Knockdown of Mul1 restored the protein levels of Drp1 and Mfn2 in stroke rats. A-D show significantly decreased Mul1 or Drp1 protein levels in whole lysates, mitochondria or cytoplasma of brain tissues in the Mul1 knockdown group compared with the stroke group. Top: optical density for the protein blot of Mul1 or Drp1 vs β-actin or VDAC1 (mitochondrial protein marker), which are further normalized by the sham; bottom, the representative images of Western blots (n = 6 per group). E. Representative images of Drp1 sumoylation assay, showing the sumoylation of Drp1 is decreased in the Mul1 knockdown group compared with the stroke group. F and G indicate evidently increased Mfn2 protein levels in whole lysates or mitochondria of brain tissues in the Mul1 knockdown group compared with the stroke group. Top: optical density for the protein blot of Mfn2 vs β-actin or VDAC1, which are further normalized by the sham; bottom, representative images for Western blot (n = 6 per group). H presents the representative images of Mfn2 ubiquitination assay, showing the ubiquitination of Mfn2 is decreased in the Mul1 knockdown group compared with the stroke group. All values are expressed as means ± S.E.M. +Mul1 siRNA: Stroke + Mul1 siRNA; +siRNA NC: Stroke + siRNA negative control. **P < 0.01 vs Sham; #P < 0.05, ##P < 0.01 vs Stroke.

8

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

Fig. 5. Knockdown of Mul1 reduced hypoxic injury in PC12 cells. A presents the representative images of flow cytometry for apoptosis assay, showing significantly decreased percentage of apoptotic cells in the Mul1 knockdown group compared with the H/R group (B). C indicates significantly decreased caspase-3 activity in the Mul1 knockdown group compared with the H/R group. D shows remarkably decreased LDH release in the Mul1 knockdown group compared with the H/R group. All values are expressed as means ± S.E.M., n = 3 per group. +Mul1 siRNA: H/R + Mul1 siRNA; +siRNA NC: H/R+ siRNA negative control. **P < 0.01 vs Control; ##P < 0.01 vs H/R.

4. Discussion

ubiquitination, which subsequently results in disturbance of mitochondrial dynamics and functions. It is well established that the mechanisms for brain injury caused by ischemic stroke involve energy metabolic disorders, calcium overload and oxidative stress, all of which reflect the change in mitochondrial functions. For example, the mitochondrial dysfunction is not only caused by mtDNA mutation but also by the dysfunction of metabolism, which leads to cell injury or death (Bozza et al., 2013; Evstafieva et al., 2014). Numerous reports have demonstrated that mitochondrial functions in brain cells were compromised in ischemic stroke or in hypoxiatreated nerve cells (Liu et al., 2018; Sun et al., 2014). Furthermore, mitochondrial I ~ IV complexes and ATPase reaction spontaneously create reactive oxygen species. In the present study, we have found that the ATP production was evidently decreased while the reactive oxygen species production and Cyt c release were significantly increased in the

In this study, we explored the role of Mul1 in ischemic injury of the stroke rats or in hypoxic injury of PC12 cells and the underlying mechanisms. Our results showed an obvious upregulation of Mul1 and Drp1 while a significant downregulation of Mfn2 in ischemic stroke rat brain in hypoxia-treated PC12 cells, concomitant with the disturbed mitochondrial dynamics and functions. Knockdown of Mul1 in vivo or in vitro restored the levels of Drp1 and Mfn2 as well as the mitochondrial homeostasis and functions, accompanied by alleviation of ischemic injury in rat brains or hypoxic injury in PC12 cells, and decrease of Drp1 sumoylation or Mfn2 ubiquitination levels. To the best of our knowledge, this is the first study to demonstrate that upregulation of Mul1 contributes to rat brain injury following ischemic stroke through mechanisms involving in Drp1 sumoylation and/or Mfn2 9

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

Fig. 6. Knockdown of Mul1 restored the protein levels of Drp1 and Mfn2 in hypoxia-treated PC12 cells. A-D show significantly decreased Mul1 or Drp1 protein levels in whole lysates, mitochondria or cytoplasma of PC12 cells in the Mul1 knockdown group compared with the H/R group. Top: optical density for the protein blot of Mul1 or Drp1 vs β-actin or VDAC1 (mitochondrial protein marker), which are further normalized by the control; bottom, representative images of Western blots. E and F indicate evidently increased Mfn2 protein levels in whole lysates or mitochondria of PC12 cells in the Mul1 knockdown group compared with the H/R group. Top: optical density for the protein blot of Mfn2 vs β-actin or VDAC1, which are further normalized by the control; bottom, representative images of Western blots. All values are expressed as means ± S.E.M., n = 3 per group. +Mul1 siRNA: H/ R + Mul1 siRNA; +siRNA NC: H/ R + siRNA negative control. **P < 0.01 vs Control; ##P < 0.01 vs H/R.

are also critical for maintaining the balance between mitochondrial fission and fusion. Theoretically, the enhanced mitochondrial fission in rat brains suffered ischemic stroke could be due to activation of fission proteins and/or deactivation of fusion proteins. To date, at least three fusion proteins, including Mfn1, Mfn2 and OPA1, are thought to be critical for mitochondrial fusion (Peng et al., 2016; Tilokani et al., 2018). According to a recent report, it is Mfn2 but not Mfn1 or OPA1 that contributes to mitochondrial dysfunction in the stroke rat brain (Martorell-Riera et al., 2014). We thus focused on Mfn2 in this study. As expected, there was an obvious decrease in total and mitochondrial Mfn2 protein levels in the stroke rat brain or hypoxia-treated PC12 cells, supporting a weakened mitochondrial fusion. Although Drp1 protein levels were altered in the stroke rat brain or hypoxia-treated PC12 cells, there was no significant change at mRNA levels for Drp1 (Fig. S2A), suggesting that the alterations in Drp1 protein levels occur at the post-translational levels. There are many types of post-translational modifications, such as phosphorylation, methylation, glycosylation, ubiquitination and sumoylation, which are mostly catalyzed by enzymes at specific sequences in target proteins. Among them, ubiquitination and sumoylation attract our special attention because both of them are crucial for protein and cellular “destiny”. Ubiquitination (or ubiquitylation) involves the binding of ubiquitin to target proteins which were usually labeled for degradation by the proteasome system, controlling a protein's half-life and expression levels. Similarly, sumoylation involves the binding of small ubiquitin-like modifier (SUMO) to target proteins. To date, at least four different SUMO isoforms including SUMO1, SUMO2, SUMO3 and SUMO4 have been identified in mammals (Zhao, 2018). Different from ubiquitination, sumoylation could increase protein stability under various conditions (Bao et al., 2018; Wen et al., 2017). Based on these reports, it is reasonable to speculate that the elevation of Drp1protein levels might be related to increase in sumoylation of Drp1. There were reports that Mul1 possesses function of sumoylation and Drp1 is a substrate of Mul1 (Prudent et al., 2015). Thus, we postulated

rat brains suffered ischemic stroke or in hypoxia-treated PC12 cells, supporting the occurrence of mitochondrial dysfunctions. Accordingly, cellular apoptosis and necrosis were increased reflected by an elevation in caspase-3 activity and/or LDH release. Mitochondria are highly dynamic organelles, and uphold their shape and morphology through fission and fusion. The balance between fission and fusion processes is critical for maintaining mitochondrial function and energy homeostasis. The status of mitochondrial dynamics is critical for cell survival and death. It has been shown that the disturbance of mitochondrial homeostasis is an early upstream event in neuronal death after ischemic stroke (Flippo et al., 2018; Wang et al., 2018). As expected, in this study, the numbers of small and punctate mitochondria in brain cells were increased in the stroke rats or in hypoxia-treated PC12 cells, which indicated an enhancement of mitochondrial fragmentation and reflected the disturbance of mitochondrial dynamics. It is well known that a number of proteins play important roles in regulation of mitochondrial fusion and fission. Mfn1/ 2 and Opa1 are responsible for the fusion of the outer and inner mitochondrial membranes, respectively, while fission is controlled by Drp1, a cytosolic GTPase which is recruited to mitochondria by four mitochondrial receptors (Peng et al., 2016; Tilokani et al., 2018). Recently, growing evidence revealed that Drp1-dependent excessive mitochondrial fission occurred in ischemic stroke animal models(Flippo et al., 2018; Wu et al., 2017; Xu et al., 2017), indicating that activation of Drp1 leading to enhancement of mitochondrial fission contributed to ischemic injury in the brain because pharmacological inhibition of Drp1 or genetic knockout of Drp1 could mitigate brain injury in ischemic stroke rats or mice (Cui et al., 2016; Liu et al., 2018). Consistent with these reports, in this study, we have found that both total Drp1 protein levels and mitochondrial Drp1 protein levels were significantly elevated in the stroke rat brain or in hypoxia-treated PC12 cells concomitant with a decrease in cytoplasmic Drp1, suggesting a translocation of Drp1 from cytoplasm to mitochondria, a step for Drp1 activation. In addition to mitochondrial fission protein (Drp1), fusion proteins 10

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

(caption on next page)

11

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

Fig. 7. Knockdown of Mul1 restored mitochondrial dynamics and functions in hypoxia-treated PC12 cells. A presents the representative images for detection of mitochondrial fission in PC12 cells by Mito-Tracker Green assay (left panel), showing an evident decrease in percentage of mitochondrial fission (right panel) in the Mul1 knockdown group compared with the H/R group (n = 3 per group). B presents the representative images for assay of mitochondrial membrane potential (ΔΨm), showing the ΔΨm is increased in the Mul1 knockdown group compared with the H/R group. C presents the representative images for assay of reactive oxygen species (left panel), showing evidently decreased reactive oxygen species production in the Mul1 knockdown group compared with the H/R group (right panel, n = 3 per group). D reveals remarkably increased ATP levels in the Mul1 knockdown group compared with the H/R group (n = 3 per group). E and F indicate an evident decrease in cytochrome c release from mitochondria (E) to cytoplasma (F) in the Mul1 knockdown group compared with the H/R group. Top: optical density for the protein blot of Cyt c vs VDAC1 (mitochondrial protein marker) or β-actin, which are further normalized by the control; bottom, representative images of Western blots (n = 3 per group). All values are expressed as means ± S.E.M.. +Mul1 siRNA: H/R + Mul1 siRNA; +siRNA NC: H/R + siRNA negative control. **P < 0.01 vs Control; ##P < 0.01 vs H/R.

that Mul1may account for the increase of Drp1 protein levels in the stroke rat brain or hypoxia-treated PC12 cells. In the present study, we actually found that Mul1 protein levels were obviously elevated in the stroke rat brain or hypoxia-treated PC12 cells concomitant with upregulation of Drp1. Knockdown of Mul1 in vivo or in vitro could partially restore Drp1 protein levels, confirming the link between Mul1 and Drp1. Further experiments showed that Drp1 sumoylation was obviously enhanced in the stroke rat brain, which were attenuated in the presence of Mul1 siRNAs. These findings support our hypothesis. Since Mul1 also possesses function of ubiquitination and Mfn2 is a substrate of Mul1(Farmer et al., 2017), we thus assume that Mul1 may account for, at least partially, the decrease of Mfn2 protein levels in the stroke rat brain or hypoxia-treated PC12 cells though its mRNA levels were downregulated (Fig. S2B). In the present study, the results really showed that ubiquitination of Mfn2 was significantly enhanced in the stroke rat brain, which was attenuated in the presence of Mul1 siRNAs. Furthermore, knockdown of Mul1 in vivo or in vitro could restore Mfn2 protein levels, confirming the link between Mul1 and Mfn2. In summary, in the present study, we have demonstrated for the first time that upregulation of Mul1 contributes to brain injury in ischemic stroke rats. The upregulation of Mul1 leads to disturbance of mitochondrial dynamics and functions through sumoylation of Drp1 and ubiquitination of Mfn2. Thus, Mul1 could serve as a novel therapeutic target for ischemic stroke.

Bioenergetics, mitochondrial dysfunction, and oxidative stress in the pathophysiology of septic encephalopathy. Shock 39 (Suppl. 1), 10–16. https://doi.org/10. 1097/SHK.0b013e31828fade1. Braschi, E., Zunino, R., McBride, H.M., 2009. MAPL is a new mitochondrial SUMO E3 ligase that regulates mitochondrial fission. EMBO Rep. 10, 748–754. https://doi.org/ 10.1038/embor.2009.86. Chen, C., Hu, Q., Yan, J., Yang, X., Shi, X., Lei, J., Chen, L., Huang, H., Han, J., Zhang, J.H., Zhou, C., 2009. Early inhibition of HIF-1alpha with small interfering RNA reduces ischemic-reperfused brain injury in rats. Neurobiol. Dis. 33, 509–517. https:// doi.org/10.1016/j.nbd.2008.12.010. Cui, M., Ding, H., Chen, F., Zhao, Y., Yang, Q., Dong, Q., 2016. Mdivi-1 protects against ischemic brain injury via elevating extracellular adenosine in a cAMP/CREB-CD39dependent manner. Mol. Neurobiol. 53, 240–253. https://doi.org/10.1007/s12035014-9002-4. Cui, N., Lu, H., Li, M., Yan, Q., 2017. PTPN21 protects PC12 cell against oxygen-glucose deprivation by activating cdk5 through ERK1/2 signaling pathway. Eur. J. Pharmacol. 814, 226–231. https://doi.org/10.1016/j.ejphar.2017.08.021. Evstafieva, A.G., Garaeva, A.A., Khutornenko, A.A., Klepikova, A.V., Logacheva, M.D., Penin, A.A., Novakovsky, G.E., Kovaleva, I.E., Chumakov, P.M., 2014. A sustained deficiency of mitochondrial respiratory complex III induces an apoptotic cell death through the p53-mediated inhibition of pro-survival activities of the activating transcription factor 4. Cell Death Dis. 5, e1511. https://doi.org/10.1038/cddis.2014. 469. Farmer, T., Reinecke, J.B., Xie, S., Bahl, K., Naslavsky, N., Caplan, S., 2017. Control of mitochondrial homeostasis by endocytic regulatory proteins. J. Cell Sci. 130, 2359–2370. https://doi.org/10.1242/jcs.204537. Flippo, K.H., Gnanasekaran, A., Perkins, G.A., Ajmal, A., Merrill, R.A., Dickey, A.S., Taylor, S.S., McKnight, G.S., Chauhan, A.K., Usachev, Y.M., Strack, S., 2018. AKAP1 protects from cerebral ischemic stroke by inhibiting Drp1-dependent mitochondrial fission. J. Neurosci. 38, 8233–8242. https://doi.org/10.1523/JNEUROSCI.0649-18. 2018. Fu, S.H., Zhang, H.F., Yang, Z.B., Li, T.B., Liu, B., Lou, Z., Ma, Q.L., Luo, X.J., Peng, J., 2014. Alda-1 reduces cerebral ischemia/reperfusion injury in rat through clearance of reactive aldehydes. Naunyn Schmiedebergs Arch Pharmacol 387, 87–94. https://doi. org/10.1007/s00210-013-0922-8. Huang, Y., Yu, J., Wan, F., Zhang, W., Yang, H., Wang, L., Qi, H., Wu, C., 2014. Panaxatriol saponins attenuated oxygen-glucose deprivation injury in PC12 cells via activation of PI3K/Akt and Nrf2 signaling pathway. Oxid Med Cell Longev 978034. https://doi.org/10.1155/2014/978034. 2014. Jung, J.H., Bae, S., Lee, J.Y., Woo, S.R., Cha, H.J., Yoon, Y., Suh, K.S., Lee, S.J., Park, I.C., Jin, Y.W., Lee, K.H., An, S., Lee, J.H., 2011. E3 ubiquitin ligase Hades negatively regulates the exonuclear function of p53. Cell Death Differ. 18, 1865–1875. https:// doi.org/10.1038/cdd.2011.57. Li, J., Qi, W., Chen, G., Feng, D., Liu, J., Ma, B., Zhou, C., Mu, C., Zhang, W., Chen, Q., Zhu, Y., 2015. Mitochondrial outer-membrane E3 ligase MUL1 ubiquitinates ULK1 and regulates selenite-induced mitophagy. Autophagy 11, 1216–1229. https://doi. org/10.1080/15548627.2015.1017180. Liu, F., Lu, J., Manaenko, A., Tang, J., Hu, Q., 2018. Mitochondria in ischemic stroke: new insight and implications. Aging Dis 9, 924–937. https://doi.org/10.14336/AD.2017. 1126. Martorell-Riera, A., Segarra-Mondejar, M., Munoz, J.P., Ginet, V., Olloquequi, J., PerezClausell, J., Palacin, M., Reina, M., Puyal, J., Zorzano, A., Soriano, F.X., 2014. Mfn2 downregulation in excitotoxicity causes mitochondrial dysfunction and delayed neuronal death. EMBO J. 33, 2388–2407. https://doi.org/10.15252/embj. 201488327. Narne, P., Pandey, V., Phanithi, P.B., 2017. Interplay between mitochondrial metabolism and oxidative stress in ischemic stroke: an epigenetic connection. Mol. Cell. Neurosci. 82, 176–194. https://doi.org/10.1016/j.mcn.2017.05.008. Peng, J., Ren, K.D., Yang, J., Luo, X.J., 2016. Mitochondrial E3 ubiquitin ligase 1: a key enzyme in regulation of mitochondrial dynamics and functions. Mitochondrion 28, 49–53. https://doi.org/10.1016/j.mito.2016.03.007. Prudent, J., Zunino, R., Sugiura, A., Mattie, S., Shore, G.C., McBride, H.M., 2015. MAPL SUMOylation of Drp1 stabilizes an ER/mitochondrial platform required for cell death. Mol. Cell 59, 941–955. https://doi.org/10.1016/j.molcel.2015.08.001. Reddy, P.H., Reddy, T.P., Manczak, M., Calkins, M.J., Shirendeb, U., Mao, P., 2011. Dynamin-related protein 1 and mitochondrial fragmentation in neurodegenerative diseases. Brain Res. Rev. 67, 103–118. https://doi.org/10.1016/j.brainresrev.2010. 11.004. Schabitz, W.R., Schade, H., Heiland, S., Kollmar, R., Bardutzky, J., Henninger, N., Muller, H., Carl, U., Toyokuni, S., Sommer, C., Schwab, S., 2004. Neuroprotection by hyperbaric oxygenation after experimental focal cerebral ischemia monitored by MRI. Stroke 35, 1175–1179. https://doi.org/10.1161/01.STR.0000125868.86298.8e.

Acknowledgements This work was supported by the National Natural Science Foundation of China,China (No. 81573430 to Xiu-Ju Luo, No. 81872873 to Jun Peng), Natural Science Foundation of Hunan Province, China (No. 2015JJ2156 to Xiu-Ju Luo) and the Fundamental Research Funds For the Central Universities of Central South University (No. 2019zzts330 to Jing Tian). Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ejphar.2019.172617. References Anzell, A.R., Maizy, R., Przyklenk, K., Sanderson, T.H., 2018. Mitochondrial quality control and disease: insights into ischemia-reperfusion injury. Mol. Neurobiol. 55, 2547–2564. https://doi.org/10.1007/s12035-017-0503-9. Aung, L.H.H., Li, R., Prabhakar, B.S., Li, P., 2017. Knockdown of Mtfp1 can minimize doxorubicin cardiotoxicity by inhibiting Dnm1l-mediated mitochondrial fission. J. Cell Mol. Med. 21, 3394–3404. https://doi.org/10.1111/jcmm.13250. Bae, S., Kim, S.Y., Jung, J.H., Yoon, Y., Cha, H.J., Lee, H., Kim, K., Kim, J., An, I.S., Kim, J., Um, H.D., Park, I.C., Lee, S.J., Nam, S.Y., Jin, Y.W., Lee, J.H., An, S., 2012. Akt is negatively regulated by the MULAN E3 ligase. Cell Res. 22, 873–885. https://doi.org/ 10.1038/cr.2012.38. Bao, J., Qin, M., Mahaman, Y.A.R., Zhang, B., Huang, F., Zeng, K., Xia, Y., Ke, D., Wang, Q., Liu, R., Wang, J.Z., Ye, K., Wang, X., 2018. BACE1 SUMOylation increases its stability and escalates the protease activity in Alzheimer's disease. Proc. Natl. Acad. Sci. U. S. A. 115, 3954–3959. https://doi.org/10.1073/pnas.1800498115. Bavarsad, K., Barreto, G.E., Hadjzadeh, M.A., Sahebkar, A., 2018. Protective effects of curcumin against ischemia-reperfusion injury in the nervous system. Mol. Neurobiol. https://doi.org/10.1007/s12035-018-1169-7. Bozza, F.A., D'Avila, J.C., Ritter, C., Sonneville, R., Sharshar, T., Dal-Pizzol, F., 2013.

12

European Journal of Pharmacology 861 (2019) 172617

K.-D. Ren, et al.

inhibiting neuronal apoptosis and PKCdelta/Drp1-mediated excessive mitochondrial fission. Oxid Med Cell Longev 2017, 1832093. https://doi.org/10.1155/2017/ 1832093. Yan, H., Rao, J., Yuan, J., Gao, L., Huang, W., Zhao, L., Ren, J., 2017. Long non-coding RNA MEG3 functions as a competing endogenous RNA to regulate ischemic neuronal death by targeting miR-21/PDCD4 signaling pathway. Cell Death Dis. 8, 3211. https://doi.org/10.1038/s41419-017-0047-y. Yang, J.L., Mukda, S., Chen, S.D., 2018. Diverse roles of mitochondria in ischemic stroke. Redox Biol 16, 263–275. https://doi.org/10.1016/j.redox.2018.03.002. Yun, J., Puri, R., Yang, H., Lizzio, M.A., Wu, C., Sheng, Z.H., Guo, M., 2014. MUL1 acts in parallel to the PINK1/parkin pathway in regulating mitofusin and compensates for loss of PINK1/parkin. Elife 3, e01958. https://doi.org/10.7554/eLife.01958. Zhang, H.F., Li, T.B., Liu, B., Lou, Z., Zhang, J.J., Peng, J.J., Zhang, X.J., Ma, Q.L., Peng, J., Luo, X.J., 2015. Inhibition of myosin light chain kinase reduces NADPH oxidasemediated oxidative injury in rat brain following cerebral ischemia/reperfusion. Naunyn Schmiedebergs Arch Pharmacol 388, 953–963. https://doi.org/10.1007/ s00210-015-1125-2. Zhao, X., 2018. SUMO-mediated regulation of nuclear functions and signaling processes. Mol. Cell 71, 409–418. https://doi.org/10.1016/j.molcel.2018.07.027. Zuo, W., Yang, P.F., Chen, J., Zhang, Z., Chen, N.H., 2016. Drp-1, a potential therapeutic target for brain ischaemic stroke. Br. J. Pharmacol. 173, 1665–1677. https://doi.org/ 10.1111/bph.13468.

Sugiura, A., McLelland, G.L., Fon, E.A., McBride, H.M., 2014. A new pathway for mitochondrial quality control: mitochondrial-derived vesicles. EMBO J. 33, 2142–2156. https://doi.org/10.15252/embj.201488104. Sun, J., Li, Y.Z., Ding, Y.H., Wang, J., Geng, J., Yang, H., Ren, J., Tang, J.Y., Gao, J., 2014. Neuroprotective effects of gallic acid against hypoxia/reoxygenation-induced mitochondrial dysfunctions in vitro and cerebral ischemia/reperfusion injury in vivo. Brain Res. 1589, 126–139. https://doi.org/10.1016/j.brainres.2014.09.039. Tilokani, L., Nagashima, S., Paupe, V., Prudent, J., 2018. Mitochondrial dynamics: overview of molecular mechanisms. Essays Biochem. 62, 341–360. https://doi.org/ 10.1042/EBC20170104. Wang, R., Dong, Y., Lu, Y., Zhang, W., Brann, D.W., Zhang, Q., 2018. Photobiomodulation for global cerebral ischemia: targeting mitochondrial dynamics and functions. Mol. Neurobiol. https://doi.org/10.1007/s12035-018-1191-9. Wen, D., Wu, J., Wang, L., Fu, Z., 2017. SUMOylation promotes nuclear import and stabilization of polo-like kinase 1 to support its mitotic function. Cell Rep. 21, 2147–2159. https://doi.org/10.1016/j.celrep.2017.10.085. Wu, B., Luo, H., Zhou, X., Cheng, C.Y., Lin, L., Liu, B.L., Liu, K., Li, P., Yang, H., 2017. Succinate-induced neuronal mitochondrial fission and hexokinase II malfunction in ischemic stroke: therapeutical effects of kaempferol. Biochim. Biophys. Acta (BBA) Mol. Basis Dis. 1863, 2307–2318. https://doi.org/10.1016/j.bbadis.2017.06.011. Xu, Y., Wang, Y., Wang, G., Ye, X., Zhang, J., Cao, G., Zhao, Y., Gao, Z., Zhang, Y., Yu, B., Kou, J., 2017. YiQiFuMai powder injection protects against ischemic stroke via

13