microRNA-222 Attenuates Mitochondrial Dysfunction During Transmissible Gastroenteritis Virus Infection*

Zhao et al. identify microRNA-222 as a mediator of mitochondrial dysfunction, THBS1, and CD47 induced during TGEV infection and showed that microRNA-222 inhibits TGEV-induced mitochondrial dysfunction via targeting and downregulating CD47. Graphical Abstract Highlights microRNA-222 attenuates TGEV-induced mitochondrial dysfunction. microRNA-222 downregulates THBS1 and CD47. THBS1 is the target of microRNA-222 during TGEV infection. THBS1 and CD47 increase mitochondrial Ca2+ level and reduced mitochondrial membrane potential (MMP). Transmissible gastroenteritis virus (TGEV) is a member of Coronaviridae family. Our previous research showed that TGEV infection could induce mitochondrial dysfunction and upregulate miR-222 level. Therefore, we presumed that miR-222 might be implicated in regulating mitochondrial dysfunction induced by TGEV infection. To verify the hypothesis, the effect of miR-222 on mitochondrial dysfunction was tested and we showed that miR-222 attenuated TGEV-induced mitochondrial dysfunction. To investigate the underlying molecular mechanism of miR-222 in TGEV-induced mitochondrial dysfunction, a quantitative proteomic analysis of PK-15 cells that were transfected with miR-222 mimics and infected with TGEV was performed. In total, 4151 proteins were quantified and 100 differentially expressed proteins were obtained (57 upregulated, 43 downregulated), among which thrombospondin-1 (THBS1) and cluster of differentiation 47 (CD47) were downregulated. THBS1 was identified as the target of miR-222. Knockdown of THBS1 and CD47 decreased mitochondrial Ca2+ level and increased mitochondrial membrane potential (MMP) level. Reversely, overexpression of THBS1 and CD47 elevated mitochondrial Ca2+ level and reduced mitochondrial membrane potential (MMP) level. Together, our data establish a significant role of miR-222 in regulating mitochondrial dysfunction in response to TGEV infection.

Transmissible gastroenteritis virus (TGEV) 1 , an enteropathogenic coronavirus, is an enveloped virus with a positivesense single-stranded RNA genome (1). TGEV infection can cause apoptosis, inflammation, immune response, autophagy (2)(3)(4). We reported that TGEV infection could induce cell apoptosis via mitochondria-mediated pathway and caused a decrease of mitochondrial membrane potential, implying that mitochondrial dysfunction occurs (5,6). Mitochondrion not only plays a vital role in energy production, but also mediates many pathologic processes, including viral infection, inflam-mation, apoptosis (7). Therefore, TGEV-induced pathologic processes, including apoptosis, inflammation, immune response, autophagy, might be the consequences of mitochondrial dysfunction. microRNAs (miRNAs) are a class of small non-coding RNA and plays a crucial role in regulating gene expression at post-transcriptional level by specifically binding to 3Ј UTR of target mRNA (8). miRNAs are involved in regulating many biological processes, including viral infection (9,10). Moreover, miRNAs are also involved in modulating the mitochondrial functions (11). miR-222 is a miRNA that is overexpressed in several types of cancers and related to mitochondrial function (12). In our previous study, TGEV infection was found to elevate miR-222 level and to cause mitochondrial dysfunction (13,14). Therefore, we speculate that miR-222 might have an association with TGEV-induced mitochondrial dysfunction.
To verify the hypothesis, Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of miR-222 targets were conducted and showed that miR-222 might be implicated in regulation of mitochondrial function. The effect of miR-222 on mitochondrial dysfunction during TGEV infection was measured. Subsequently, relative protein quantification analysis in response to miR-222 was performed and revealed that THBS1 and CD47 were downregulated by miR-222 and related to mitochondria function. THBS1 was identified as the target of miR-222. We examined the effects of THBS1 and CD47 on mitochondrial Ca 2ϩ and MMP and found that THBS1 and CD47 were activators of mitochondrial Ca 2ϩ levels and inhibitors of MMP. The data show miR-222 prevent TGEV-induced mitochondrial dysfunction via targeting THBS1 and restraining CD47.
Experimental Design and Statistical Rationale for Proteomics-To investigate the effect of miR-222 on TGEV-induced mitochondrial dysfunction, PK-15 cells were transfected with miR-222 mimics (experimental samples) or mimics control (control samples) using Lipo-fectamine 3000 and infected with TGEV at an MOI of 1.0 at 24 h post-transfection (hpt). The cells were collected for quantitative proteomic analysis at 24 h postinfection (hpi). Two biological replicates preparations labeled with TMT were analyzed. The two control samples, which were transfected with miRNA mimics control and infected with TGEV, were respectively labeled with TMT-126 and TMT-129. The two experimental samples, which were transfected with miR-222 mimics and infected with TGEV, were respectively labeled with TMT-128 and TMT-131. To assess the transfection 1 The abbreviations used are: TGEV, transmissible gastroenteritis virus; PK-15 cells, porcine kidney (PK) 15 cells; PCR, polymerase chain reaction; miRNA, microRNA; THBS1, thrombospondin-1; CD47, cluster of differentiation 47; MMP, mitochondrial membrane potential; qRT-PCR, quantitative reverse transcription PCR; UTR, untranslated region; GO, gene ontology; KEGG, Kyoto encyclopedia of genes and genomes; MOI, multiplicity of infection; TMT, tandem mass tags; AGC, automatic gain control; hpi, hours post-infection; hpt, hours post-transfection; HPLC, high performance liquid chromatography; MS, mass spectrometry; LC/MS, liquid chromatographymass spectrometry; LC-MS/MS, liquid chromatography-tandem mass spectrometry.

microRNA-222 Attenuates Mitochondrial Dysfunction
Molecular & Cellular Proteomics 18.1 53 efficiency, miR-222 level was measured by qRT-PCR. The sequences of miR-222 mimics, mimics control, and primers are shown in supplemental S1. Quantification of miRNA and mRNA by qRT-PCR-Total RNA was isolated using Trizol reagent (Invitrogen, California) from PK-15 cells. qRT-PCR was performed as previously described (13). 2 g of total RNA was treated with DNase I (Fermentas, Germany) for 30 min at 37°C and reversely transcribed using the First-strand cDNA synthesis kit (Invitrogen). The cDNA was used as the template for qRT-PCR. qRT-PCR was performed using the AccuPower 2ϫGreenstar qPCR Master mix (Bioneer, Korea) on Bio-Rad iQ5 Real-Time PCR System (Bio-RadA). The relative fold changes of miRNA and mRNA were calculated using two-ddCt method (respectively normalized to U6 and ␤-actin) (16). The primers are shown in supplemental S1.
Protein Isolation and Labeling with TMT Reagents-Samples were sonicated three times on ice in lysis buffer (8 M urea, 1% Triton-100, 65 mM DTT and 0.1% Protease Inhibitor Mixture) and centrifuged at 20,000 ϫ g at 4°C for 20 min. The supernatant was treated with cold 15% trichloroacetic acid (TCA) for 2 h at Ϫ20°C and centrifuged for 10 min at 20,000 ϫ g at 4°C. The precipitate was washed with cold acetone for three times. The protein was dissolved in buffer (8 M urea, 100 mM TEAB, pH 8.0) and concentration was determined with 2-D Quant kit according to the manufacturer's instructions. 100 g protein for each sample were digested with trypsin for following experiments. The two control samples transfected with miRNA mimics control was respectively labeled with TMT-126 and TMT-129. The two experimental samples treated with miR-222 mimics were respectively labeled with TMT-128 and TMT-131.
LC-MS/MS Analysis-The samples were fractionated by high pH reverse-phase High Performance Liquid Chromatography (HPLC). Briefly, peptides were dried by vacuum centrifugation, dissolved in 0.1% formic acid, and then directly loaded onto a reversed-phase pre-column (Acclaim PepMap 100), (Thermo Fisher Scientific, Waltham, MA) Subsequently, peptides were separated using a reversedphase analytical column (Acclaim PepMap RSLC) (Thermo Fisher Scientific) and analyzed by Q Exactive TM hybrid quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific). The peptides were subjected to NSI source followed by tandem mass spectrometry (MS/MS) in Q Exactive TM (Thermo Fisher Scientific) coupled online to the HPLC. Intact peptides were detected in the Orbitrap at a resolution of 70,000. Peptides were selected for MS/MS using NCE setting as 32. Ion fragments were detected in the Orbitrap at a resolution of 17,500. A data-dependent procedure that alternates between one MS scan followed by 20 MS/MS scans was applied for the top 20 precursor ions above a threshold ion count of 2E4 in the MS survey scan with 30.0 s dynamic exclusion. The electrospray voltage was 2.0 kV. Automatic gain control (AGC) was set at 5E4. For MS scans, the m/z scan range was 350 to 1800. Fixed first mass was set as 100 m/z.
Analysis of LC-MS/MS Data-MS/MS data were analyzed using Mascot search engine (version 2.3.0) following a previously described protocol (17). Briefly, raw MS data files were processed using the LC/MS software Proteome Discoverer (version 1.3.0.339) (Thermo Fisher Scientific) and converted into the Mascot generic format (mgf)

microRNA-222 Attenuates Mitochondrial Dysfunction
GO and KEGG Analysis of Differential-Expressed Proteins-GO annotations of proteome were derived from the UniProt-GOsA database (http://www.ebi.ac.uk/GOA). KEGG database was used to annotate proteins pathways. The Wolf Psort database (https:// wolfpsort.hgc.jp/) was used to predict subcellular localization.
Protein-Protein Interaction Analysis-All differentially expressed protein identifiers (Uniprot accession) were searched against STRING database (version 10.0) for protein-protein interactions (19).
Vector Construction-To overexpress THBS1 and CD47, the fulllength sequences of THBS1 and CD47 were obtained by PCR from PK-15 cells and cloned into plasmid pCI-neo (Promega). The constructions were respectively named pCI-neo-THBS1 and pCI-neo-CD47. The primer sequences are shown in supplemental S1.
Measurement of Mitochondrial Ca 2ϩ Level-Mitochondrial Ca 2ϩ level was assessed using Rhod-2 kit (GENMED, China) following the manufacturer's instruction. Briefly, after being washed, the PK-15 cells were dyed using dye reagent containing Rhod-2 for 30 min at room temperature and then incubated for 30 min at 37°C in dark places. The absorbance was measured at 550 ex/590 em using fluorescence spectrophotometer. The relative fluorescence unit (RFU) was calculated according the manufacturer's instruction.
Evaluation of MMP-MMP of PK-15 cells was assessed using JC-1 kit (GENMED, China) following the manufacturer's protocol. PK-15 cells were treated with dye reagent containing JC-1 and incubated at room temperature for 10 min. The absorbance was measured at 490 ex/590 em using fluorescence spectrophotometer. The relative fluorescence unit (RFU) was calculated according the manufacturer's instruction.
Western Blot Analysis-Cells were treated with RIPA lysis buffer containing phenylmethyl sulfonylfluoride (PMSF). Total proteins were separated on a 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequently transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore). The membrane was blocked with 5% BSA for 2 h at room temperature and successively incubated with primary antibody overnight at 4°C and with HRPconjugated secondary antibody at room temperature for 1 h. The signal was detected using enhanced chemiluminescence (ECL) kit (Bio-Rad).

Identification of miR-222
Target-3Ј UTRs of 9 candidate target genes, which are related to mitochondria and contain the binding site of miR-222, were respectively amplified by PCR and cloned into dual-luciferase reporter vector, psiCHECK-2 (Promega), to obtain the recombinant plasmids for wildtype of targets. The binding sites of miR-222 in 3Ј UTR of targets were mutated following a mutagenesis protocol to obtain the mutants of targets. The primer sequences are shown in supplemental S1. miR-222 mimics, miRNA mimics control, miR-222 inhibitors, and miRNA inhibitors control were synthesized by Ribo Biotech (RiboBio, China) (The sequences are shown in upplemental S1). PK-15 cells were grown in 24-well plate and then cotransfected with recombinant plasmids for wildtype (or mutant) and miR-222 mimics (or miR-222 inhibitors) using Lipofectamine 3000 (Invitrogen). Luciferase activities were measured on luminometer using Dual-Glo Luciferase Assay System (Promega) following the manufacturer's manual.

Gene Ontology Enrichment Analysis of miR-222 Targets-
TargetScan and miRanda were used to predict miR-222 targets (20). Three hundred ten targets were obtained (sup-

microRNA-222 Attenuates Mitochondrial Dysfunction
plemental S2) and searched against GO database to provide enrichment information on Biological Process, Molecular Function, and Cellular Component. GO enrichment of the 310 targets of miR-222 showed that 48.5% of targets were enriched in metabolic process and 24.2% of targets were involved in mitochondria in Cellular component (Fig. 1).
KEGG Pathway Enrichment Analysis of Differentially Expressed Proteins-KAAS, an online service tool of KEGG, was used to annotate the 310 targets of miR-222. Annotations were mapped on KEGG pathway database using KEGG mapper. KEGG enrichment analysis revealed the targets of miR-222 were primarily enriched in Peroxisome, PPAR signaling pathway, Biosynthesis of unsaturated fatty acids, and Fatty acid degradation (Fig. 2A). The mitochondria-related targets of miR-222 were mainly enriched in Fatty acid degradation, Peroxisome, PPAR signaling pathway, and Metabolic pathway (Fig. 2B).
miR-222 Decreased Mitochondrial Ca 2ϩ Level and Increased MMP Level-To investigate the impact of miR-222 on mitochondrial dysfunction during TGEV infection, PK-15 cells were transfected with miR-222 mimics/miR-222 inhibitors and subsequently infected with TGEV or treated with medium instead of TGEV, named MOCK infection. Mitochondrial Ca 2ϩ level and MMP level were measured at 24 hpi. Overexpression and silence of miR-222 were successfully achieved by miR-222 mimics (Fig. 3A) and miRNA-222 inhibitors (Fig. 3B). Mitochondrial Ca 2ϩ level was downregulated by miR-222 mimics and was upregulated by miR-222 inhibitors (Fig. 3C). MMP level was increased by miR-222 mimics and was decreased by miR-222 inhibitors (Fig. 3D). The results suggest that miR-222 counteracts TGEV-induced mitochondrial dysfunction in PK-15 cells.
Identification of 100 Differentially Expressed Proteins in TGEV-infected PK-15 Cells-To analyze the molecular mechanism of miR-222 in TGEV-induced mitochondrial dysfunction, relative protein quantification was performed. A total of 4209 proteins were detected and quantified (supplemental Fig. S3). When setting quantification ratio of Ͼ1.2 (p Ͻ 0.05) as upregulated threshold and Ͻ0.83 (p Ͻ 0.05) as downregulated threshold, 100 differentially expressed proteins were obtained, including 57 upregulated proteins and 43 downregulated proteins (Table I)
KEGG Pathway Enrichment Analysis of Differentially Expressed Proteins-The 100 differentially expressed proteins were searched for functional enrichment by a KEGG database search. The results showed that these proteins were mainly enriched in Ribosome, ECM-receptor interaction, Oxidative phosphorylation, and Focal adhesion (Fig. 5A). The mitochondria-related proteins were prevailingly enriched in Oxidative phosphorylation and Mitochondrial ribosome (Fig. 5B). The differentially expressed proteins were used as an input list for the web-tool STRING to generate protein-protein interaction networks (Fig. 6). It is interesting that THBS1 interacts with CD47. THBS1 is predicted as one of the targets of miR-222.
Verification of Differentially Expressed Proteins by Western blotting and qRT-PCR-Relative quantitative proteomic analysis showed that miR-222 mimics downregulated expression of THBS1 and CD47. According to Protein-Protein interaction analysis of miR-222 targets, THBS1 is predicted the target of miR-222 and interacts with CD47. Moreover, THBS1 is related to regulate mitochondrial function (21,22). Therefore, expression of THBS1 and CD47 were verified. The mRNA level of THBS1 was not influenced by miR-222 (Fig. 7A). The mRNA level of CD47 was reduced by miR-222 mimics and elevated by miR-222 inhibitors (Fig. 7C). Expression levels of THBS1 and CD47 were inhibited by miR-222 mimics and promoted by miR-222 inhibitors (Fig. 7B and 7D).

microRNA-222 Attenuates Mitochondrial Dysfunction
dria-related targets, including THBS1, were respectively amplified and cloned into 3Ј UTR of Renilla luciferase of dualluciferase reporter plasmid psiCHECK-2. The recombinants and miR-222 mimics/inhibitors were co-transfected into PK-15 cells. The luciferase activities were detected. The results showed that the Renilla luciferase activity of plasmid containing the 3Ј UTR of THBS1 was markedly lower than that of the control (normalized to Firefly luciferase activity) (Fig.  8A). To further confirm the direct binding between miR-222 and 3Ј UTR of THBS1, the mutated 3Ј UTR sequence of THBS1, in which miR-222 seed sequence was mutated, was cloned into the 3Ј UTR of Renilla luciferase to generate construction THBS1-mut. The constructs were respectively cotransfected into PK-15 cells with either miR-222 mimics or inhibitors. The Renilla luciferase activity of THBS1-WT was reduced by miR-222 mimics and potentiated by miR-222 inhibitors (normalized to Firefly luciferase activity) ( Fig. 8B and  8C). The Renilla luciferase activity of THBS1-mut was not affected by mimics and inhibitors of miR-222 ( Fig. 8D and 8E). Moreover, the mRNA level of THBS1 was not changed by miR-222 mimics and miR-222 inhibitors (Fig. 7A). However, the protein level of THBS1 was reduced by miR-222 mimics and increased by miR-222 inhibitors (Fig. 7B). Taken together, our data identified that THBS1 was a direct target of miR-222.
Effects of THBS1 on Mitochondrial Ca 2ϩ and MMP Levels-We demonstrated that miR-222 targeted THBS1 and reduced TGEV-induced mitochondrial dysfunction. So, we speculate that miR-222 might suppress TGEV-induced mito-
Effects of CD47 on Mitochondrial Ca 2ϩ and MMP Levels-pCI-neo-CD47 was constructed to overexpress CD47. siRNAs of CD47 were design and synthetized (RiboBio, China). PK-15 cells were respectively transfected with pCIneo-CD47 and siCD47 and subsequently infected with TGEV. Mitochondrial Ca 2ϩ and MMP were tested. Mitochondrial Ca 2ϩ was decreased by siCD47 and increased by pCI-neo-CD47 (Fig. 10A). MMP level was facilitated by siCD47 and reduced by pCI-neo-CD47 (Fig. 10B).

DISCUSSION
In this study, relative protein quantification was performed using TMT MS/MS to quantify proteome in presence of miR-222 mimics treatment. 100 differently expressed proteins were obtained, including 57 upregulated and 43 downregulated proteins. Based on bioinformatics analysis, miR-222 was linked to mitochondria, so we presumed that miR-222 might be involved in regulating TGEV-induced mitochondrial dysfunction. Our results indicate that miR-222 can attenuate TGEV-induced mitochondrial dysfunction by targeting THBS1 and downregulating CD47 (Fig. 11).
By quantitative proteomic analysis, 100 differentially expressed proteins were obtained. Subcellular localization and GO analysis of the 100 differentially expressed proteins showed that 18.0% proteins are located at mitochondria and prevailingly enriched in metabolic process. Among the 100 differentially expressed proteins, many mitochondrial proteins are regulated by miR-222 during TGEV infection, including MRPL11, MRPL24, MRPS26, MRPL38, MRPL46, MRPL47, NDUFV1, NDUFS1, NDUFA13, NDUFC2, and NDUFB5. MRPL11, MRPL24, MRPS26, MRPL38, MRPL46, and MRPL47 are mammalian mitochondrial ribosomal proteins and contribute to synthesis of essential mitochondrial pro-teins (23,24). NDUFV1, NDUFS1, NDUFA13, NDUFC2, and NDUFB5 are NADH-dehydrogenases of mitochondria (25). NADH-dehydrogenase is one of the most important enzyme of complex I (26). Results of KEGG pathway analysis reveal that these differentially proteins are involved in ribosome, ECM-receptor interaction, oxidative phosphorylation, pyruvate metabolism, glucagon signaling pathway, and fatty acid metabolism. Results of GO and KEGG analysis show that miR-222 might be primarily involved in regulation of mitochondrial function during TGEV infection.
Mitochondria are eukaryotic organelle and mainly involved in many metabolic processes, including metabolism of glucose, fat, protein, and nucleic acid (27)(28)(29)(30). Mitochondrial dysfunction is a key reason in many pathogenic processes, including metabolism, cell death, autophagy, and viral infection (31,32). It was reported that depletion of miR-222 damaged mitochondrial membrane potential, indicating that miR-222 functions as an inhibitor of mitochondrial dysfunction (33). In this present study, we found that miR-222 resulted in an increase of mitochondrial membrane potential. Mitochondria have ability to take up cytoplasmic Ca 2ϩ to maintain calcium homeostasis. But, once excessive Ca 2ϩ enters mitochondria, mitochondria will be damaged (34). Here, we observed that mitochondrial Ca 2ϩ level was augmented by miR-222. Our data imply miR-222 might counteract TGEV-induced mitochondrial dysfunction. These results are consistent with the previous findings.
THBS1 is a secretive glycoprotein and involved in regulation of mitochondrial function. Knockout of THBS1 results in mitochondrial dysfunction in mice (21). THBS1 can prevent mitochondrial pathway of apoptosis from being triggered (35) and inhibits palmitate-induced release of mitochondrial cytochrome c, cleavage of capase-3, and apoptosis (36), indicat-

microRNA-222 Attenuates Mitochondrial Dysfunction
ing THBS1 plays a role in mitochondrial function. We demonstrated that THBS1 is a regulator of mitochondrial dysfunction during TGEV infection, being consistent with the previous reports.
CD47 is a cell-surface receptor and can bind to THBS1 to form a complex to modulate mitochondrial function (22,37). THBS1 interacting with CD47 leads to mitochondrial dysfunction by disruption of the MMP (22). We found that CD47 and THBS1 promoted mitochondrial dysfunction during TGEV infection, indicating that interaction between CD47 and THBS1 play an important role in TGEV-induced mitochondrial dysfunction.
In conclusion, we found that proteome was changed by miR-222 during TGEV infection by quantitative proteomic analysis. Moreover, miR-222 played a role in decreasing mitochondrial Ca 2ϩ level and increasing TGEV-reduced MMP in PK-15 cells by targeting THBS1 and suppressing CD47.

DATA AVAILABILITY
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD010927 (http:// www.ebi.ac.uk/pride; Project accession: PXD010927).