dsRNA is the intermediate products of + RNA virus replication (Knoops et al., 2008, 2012). To find some clues for the role of mitochondria in SARS-CoV-2 replication, we immunofluorescently stained SARS-CoV-2-infected cells using antibodies specific for dsRNA and the mitochondrial markers, including the mitochondrial matrix protein HSP60 and the outer mitochondrial membrane protein Tom20 (Gupta and Knowlton, 2005). In both Vero E6 and Huh-7 cells, a merge of signals from Hsp60 and dsRNA were observed at 12 h post SARS-CoV-2 infection (Figures 1A,B). Similarly, colocalization of Tom20 and dsRNA was also apparent at 6 and 12 h post infection (Figures 1C,D). We next performed immunoelectron microscopy experiments using Vero E6 cells to further confirm the mitochondrial localization of SARS-CoV-2 dsRNA directly. In mitochondria, small numbers of gold particle of dsRNA signals were visualized at 12 h post infection (Figure 1E), suggesting dsRNA accumulation and possible virus replication in mitochondria.

Moreover, we evaluated the dsRNA localization in other primary subcellular organelles, including endoplasmic reticulum and lysosomes. No overlapping signals of dsRNA and the endoplasmic reticulum marker Calnexin, as well as dsRNA and the lysosome marker Lamp2b, were observed in Vero E6 cells at 12 h post SARS-CoV-2 infection (Figure 1F).

To examine the effect of SARS-CoV-2 on mitochondrial function, we performed JC-1 staining to detect mitochondrial membrane potential (ΔΨm) on Vero E6 and Huh-7 cells. JC-1 dye aggregates in healthy mitochondria and fluoresces red. Upon the loss of ΔΨm, JC-1 can only exists as monomers and fluoresces green (Koppikar et al., 2010). In Vero E6 and Huh-7 cells, we found that SARS-CoV-2 disrupted the ΔΨm as observed by decrease in the intensity of red fluorescence, as well as increase in JC-1 monomers (green fluorescence) along with time (Figures 2A,B).

Depolarization of ΔΨm will lead to the opening of mitochondrial permeability transition pore (MPTP) and the change of mitochondrial permeability. Next, we used TMRM dye to detect MPTP opening. TMRM dye is lipophilic and has strong red fluorescence in the inner membrane of mitochondria. Upon MPTP opening, the dye enters the cytoplasm resulting into decrease in red fluorescence (Javed et al., 2012; Boyman et al., 2019). In accordance with this, the red fluorescence of TMRM gradually decreased at 3, 6, and 12 h after SARS-CoV-2 infection in Vero E6 and Huh-7 cells, suggesting the MPTP opening and the release of mitochondrial contents into the cytoplasm (Figure 2C).

Mitochondrial respiratory chain and mitochondrial lipid peroxidation are two main sources of ROS production in animal cells. Excessive release of ROS leads to oxidative damage and abnormal energy metabolism (Zorov et al., 2014). We used carboxy-H2DCFDA probe which releases green fluorescence in response to esterase oxidation and can reflect ROS levels by fluorescence intensity (Wu and Yotnda, 2011). By fluorescence microscopy, we observed that the cells infected with SARS-CoV-2 exhibited a time dependent increase in green fluorescence compared to mock-infected cells (Figures 2D,E). This indicates increased ROS release caused by SARS-CoV-2 which might be associated with mitochondrial dysfunction.

Next, we visualized mitochondrial morphology by transmission electron microscopy and found mitochondrial swelling and vacuolization at 24 h in SARS-CoV-2-infected Vero E6 cells. In addition, we observed the virus replication complexes of DMV structures around mitochondria (Figure 2F). Taken together, all these results suggested that SARS-CoV-2 induced mitochondrial dysfunction, including the loss of ΔΨm, MPTP opening and increased ROS release.

To investigate the role of mitochondria in SARS-CoV-2 replication, we pretreated cells with mitochondrial membrane stabilizer prior to SARS-CoV-2 infection. Mdivi-1, an inhibitor of mitochondrial fission protein dynamin-related protein-1, can mitigate mitochondrial depolarization and improve the morphological abnormality of mitochondria. First, we analyzed the cytotoxic effects of the two inhibitors in Vero E6 and Huh-7 cells, and found that mdivi-1 (5 μM) and cyclosporin A (40 μM) did not affect cell viability (Figure 3A). In both Vero E6 and Huh-7 cells, mdivi-1 (5 μM) reduced SARS-CoV-2 gene copies at 24 and 48 h post infection. Similarly, cyclosporin A (40 μM), a cyclophilin D inhibitor known to restrain MPTP opening, also decreased SARS-CoV-2 viral load (Figures 3B,C). These results indicates that the loss of ΔΨm and MPTP opening are beneficial to SARS-CoV-2 replication.

The mitochondrial outer membrane protein Tom20 is a subunit of Tom complexes and serves as the central entry gate for nuclear-encoded mitochondrial precursor proteins. Additionally, Tom20 is known to interact with the mitochondrial targeting sequence of many mRNA, and its deletion results in decreases in the localization of the mitochondrial targeting sequence-GFP reporter mRNA (Eliyahu et al., 2010; Lesnik et al., 2015). In Vero E6 cells post SARS-CoV-2 infection, the expression of Tom20 increased gradually along with time (Figure 4A). In order to explore whether Tom20 protein is involved in the mitochondrial entry process of SARS-CoV-2 dsRNA, we knocked down the expression of Tom20 protein by siRNA (Figure 4B). TMRM staining showed that Tom20 knockdown improved mitochondrial permeability at 6 and 12 h post SARS-CoV-2 infection (Figure 4C). This may be related to decreased mitochondrial localization of SARS-CoV-2 dsRNA in mitochondrial as observed by reduced overlapping signals from dsRNA and Hsp60 (Figure 4D), indicating a role of Tom20 in facilitating the mitochondrial entry of SARS-CoV-2 dsRNA. Importantly, this mitochondrial localization is associated to SARS-CoV-2 replication and Tom20 knockdown reduced viral load at 48 h post SARS-CoV-2 infection (Figure 4E). Then, after infection of the tom20-silenced cells with the SARS-CoV-2, it was found that the level of viral N protein (nucleocapsid protein) in the mitochondria was significantly reduced (Figure 4F), and the release of ROS was also significantly reduced (Figure 4G).

In defense against SARS-CoV-2, mitophagy represents a powerful tool in mitochondrial quality control and the clearance of viral dsRNA. To further investigate whether the mitochondrial damage induced by SARS-CoV-2 would subsequently initiate mitophagy, we examined the Pink1/Parkin pathway. Under healthy and steady state, Pink1, a mitochondrial serine/threonine kinase, can be rapidly degraded by ΔΨm-dependent proteolytic enzymes. When mitochondrial damage and depolarization occurs, Pink1 accumulates in the outer membrane of mitochondria and recruits Parkin protein (Ashrafi and Schwarz, 2013; Matsuda et al., 2013). Parkin is a cytoplasmic protein with E3 ubiquitin ligase activity. The polyubiquitination of mitochondrial substrate mediated by Parkin is considered to be an important step in inducing mitophagy (Geisler et al., 2010). In SARS-CoV-2-infected Vero E6 cells, the expression of Pink and Parkin increased gradually along with time (Figure 5A). Additionally, by immunofluorescence microscopy, we observed a significant increase in a merge of signals from Hsp60 and Pink1, as well as Hsp60 and Parkin, in SARS-CoV-2 infected Vero E6 cells at 12 h compared to mock-infected cells (Figures 5B,C). These results indicate the activation of Pink1/Parkin pathway and accumulation of Pink1 and Parkin protein in mitochondria by SARS-CoV-2 infection.

Parkin mediates polyubiquitination of mitochondrial substrates which can be recognized by the ubiquitin adaptor protein P62. The recruitment of p62 to mitochondria labeled damaged mitochondria as a selective autophagy substrate (Geisler et al., 2010). In Vero E6 cells, the accumulation of P62 to mitochondria by SARS-CoV-2 infection was confirmed as observed by a substantial increase in the overlapping signals from Hsp60 and P62 (Figure 5D). Next, to explore whether the mitochondrial recruitment of P62 is mediated by Pink1/Parkin pathway, we knocked down the expression of Pink1 protein by siRNA (Figure 5E). In SARS-CoV-2-infected cells, we found decreased mitochondrial localization of P62 by Pink1 knock down (Figure 5D). Taken together, these results suggest the activation of Pink1-Parkin-P62 pathway known for mitophagy in SASRS-CoV-2 infected cells.

Although autophagy aims at clearance, in an ongoing arm race, viruses have evolved the potent strategy to escape and inhibit autophagy for their benefit (Choi et al., 2018; Mao et al., 2019). To further investigate whether SARS-CoV-2 interferes with host cell mitophagy, we examined Hsp60 expression to reflect mitochondrial number. In Vero E6 cells, the abundance of Hsp60 remains constant post SARS-CoV-2 infection, indicative of unimplemented mitochondrial clearance (Figure 6A).

In mitophagy, flawed mitochondria were selectively sequestered into autophagosomes and subsequently degraded in lysosomes for mitochondria quality control (Lemasters, 2005; Ashrafi and Schwarz, 2013). We next performed immunofluorescent staining using antibodies against Hsp60, LC3 and Lamp2b, to explore the different stages in SARS-CoV-2-induced mitophagy. No apparent overlapping signals of Hsp60 and the autophagosome marker LC3, as well as Hsp60 and Lamp2b, were observed in Vero E6 cells at 12 h post SARS-CoV-2 infection (Figure 6B). These results suggesting that the flawed mitochondria after SARS-CoV-2 infection were not encapsulated by autophagosomes, as well as subsequently transported into lysosomes.

Finally, in order to find out the way in which SARS-CoV-2 inhibits mitophagy, we detected the binding ability of p62 to LC3 protein by immunoprecipitation. P62 contains many functional domains, of which UBA domain can bind to ubiquitinated substrate, while LIR domain can bind to LC3 protein, which provides a link for autophagosomes to selectively encapsulate ubiquitinated substrates (Ashrafi and Schwarz, 2013). We found that the binding ability of p62 to LC3 protein in Vero E6 cells decreased at 12 h post SARS-CoV-2 infection (Figure 6C), suggesting that SARS-CoV-2 hindered the binding of p62 to LC3 protein, thus inhibiting the p62-labeled mitochondria to be encapsulated by autophagosomes.

African green monkey kidney cell Vero E6 (ATCC, CRL-1586) and human hepatocarcinoma cell Huh-7 (JCRB, 0403) were maintained in Dulbecco’s modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS), supplemented with 50 U ml–1 of penicillin and 50 μg ml–1 of streptomycin.

The SARS-CoV-2 strain BetaCoV/Beijing/IME-BJ01/2020 was originally isolated by CFQ’s laboratory as described previously (Gordon et al., 2020; Shang et al., 2021a,b). SARS-CoV-2 stocks were propagated and titered on Vero E6 cells. All work with live virus was performed in ABSL-3 in Changchun Veterinary Research Institute. In our experiments, cells were infected with SARS-CoV-2 at a MOI of 0.008.

Mdivi-1 (HY-15886) and cyclosporin A (HY-B0579) were purchased from MedChemExpress (New Jersey, United States).

The rabbit or mouse antibodies against GAPDH (5174), Hsp60 (12165), Calnexin (2679) and LC3 (83506) were obtained from Cell Signaling Technology (Danvers, MA, United States). The rabbit antibody against LC3 (L8918) was obtained from Sigma-Aldrich. The rabbit antibody against Tom20 (ab186735) and Anti-SARS-CoV-2 nucleocapsid protein antibody (ab271180) were purchased from Abcam (Cambridge, United Kingdom). The rabbit or mouse antibodies against Pink1 (23274-1-AP), Parkin (66674-1-Ig), HSP60 (66041-1-Ig), Lamp2b (27823-1-AP or 66301-1-Ig) and P62 (66184-1-Ig) were obtained from Proteintech (Chicago, United States). The mouse anti-dsRNA (10010200) antibody was obtained from SCICONS (Szirák, Hungary). The Alexa Fluor^® 647-conjugated anti-mouse secondary antibody (4410), Alexa Fluor^® 488-conjugated anti-rabbit secondary antibody (4412) and horseradish peroxidase (HRP)–conjugated secondary antibodies (7074 or 7076) were purchased from Cell Signaling Technology. The FITC-conjugated anti-mouse secondary antibody (A0568) and Cy3-conjugated anti-rabbit antibody were purchased from Beyotime Biotechnology (Shanghai, China). The10-nm colloidal gold-conjugated secondary antibodies (G7777) was purchased from Sigma-Aldrich.

The control siRNA (sc-37007), siRNA against Tom20 (sc-36691) and siRNA against Pink1 (sc-44598) were purchased from Santa Cruz Biotechnology (California, United States). Twenty-four hours after seeding, the Vero E6 cells were transfected with 50 nM of siRNA using Lipofectamine™ RNAiMAX reagent (Thermo Fisher Scientific, 13778150) and experiments were conducted 48 h post transfection.

SARS-CoV-2-infected Vero E6 or Huh-7 cells were cultured for an indicated time on coverslips. Then, cells were fixed with 4% paraformaldehyde for 1 h and blocked with 0.3% Triton X-100 and 5% Bovine Serum Albumin (BSA) for 1 h. After washing for three times with PBS, specific primary antibodies were added and incubated overnight at 4°C. Cells were then washed and incubated with the anti-rabbit or anti-mouse fluorescent secondary antibodies for 1 h. Images were captured using a fluorescence microscope (Olympus BX43).

SARS-CoV-2-infected Vero E6 cells were fixed with ice-cold 4% paraformaldehyde containing 0.5% glutaraldehyde for 4 h. The fixed cells were washed with PBS for three times and passed through the following procedures: agarose solidification, dehydrate, resin penetration, embedding and polymerization. The blocks were cut to 70–80 nm ultrathin sections and immunolabeled with anti-dsRNA antibodies on the nickel grids, followed by incubation with 10-nm colloidal gold-conjugated goat anti-mouse antibodies. Images were captured under a transmission electron microscope (TEM).

Mitochondrial membrane potential (ΔΨm) was detected by JC-1 staining. Vero E6 and Huh-7 cells were seeded on coverslips at a density of 1 × 10⁵ cells/well in 12-well plates. Mock- or SARS-CoV-2-infected Vero E6 and Huh-7 cells were cultured for an indicated time. At 3, 6, and 12 h post infection, cells were washed with PBS for three times. 10 μg/ml JC-1 fluorescent dye (Sigma Aldrich, T4069) was added and incubated for 15 min. Images were captured by fluorescence microscopy.

Mitochondria permeability transition pore (MPTP) opening was assessed by the tetramethylrhodamine methyl ester (TMRM) probe (Sigma Aldrich; T5428). At 3, 6, and 12 h post infection, Vero E6 and Huh-7 cells were washed with PBS for three times. 300 nM TMRM was added and incubated for 15 min. Images were captured by fluorescence microscopy.

Reactive Oxygen Species (ROS) release was detected by the carboxy-H2DCFDA probe (Invitrogen, I36007). At 3, 6, and 12 h post infection, Vero E6 and Huh-7 cells were washed with PBS for three times. 25 μM carboxy-H2DCFDA working solution were added and incubated for 30 min. Cells were then returned to pre-warmed growth medium and incubate at 37°C for another 1 h. Images were captured by fluorescence microscopy. After the above steps, flow cytometry can also be performed to observe the changes in ROS levels.

Vero E6 cells were seed to a 6-well plate and cultured for 24 h. Then the SARS-CoV-2 infection was performed. After 48 h, the supernatant was discarded, and washed with PBS for three times. The crystal violet staining solution was added 600 μL to each well, and the cells were stained for 10 min at room temperature. Finally, the cells were washed with PBS for three times, and images were captured for analysis.

Vero E6 cells were mock-infected or SARS-CoV-2-infected for 24 h and then fixed with 2.5% glutaraldehyde. Areas containing cells were block mounted and sliced. Ultrathin sections were contrasted with uranyl acetate and lead hydroxide, and subsequently viewed under a TEM (Hitachi, HT7800).

Cells were seeded in 6-well plates and infected with SARS-CoV-2 after incubation for 24 h. At various time points as indicated, cells were collected, lysed and protein concentration was determined by BCA protein assay kit (Beyotime, P0010). After mixed with loading buffer and boiled for denaturing, equal amounts (30 μg) of proteins were electrophoresed, transferred into a membrane and then incubated overnight at 4°C with specific primary antibodies. After further incubation with HRP-conjugated secondary antibodies, the specific bands were visualized with the Pierce ECL Immunoblotting Substrate (Thermo Fisher Scientific, 32106) using the Amersham Imager 600 system. In siRNA gene silencing experiments, cells were collected without SARS-CoV-2 infection and subjected to the immunoblotting procedure described above.

Vero E6 cells were seeded in 12-well plates at a density of 2 × 10⁵ cells/well. After SARS-CoV-2 infection, cells were incubated for an indicated time. Viral RNA in supernatant was isolated with the QIAamp Viral RNA Kit (Qiagen, 52906) following the manufacturer’s instructions. Virus copies were then detected by RT-qPCR methods with the HiScript II One Step RT-qPCR SYBR Green Kit (Vazyme Biotech, Nanjing, China) using an ABI 7500 real time PCR system (Applied Biosystems, CA, United States). The protocol for RT-qPCR was as follows: 50°C for 15 min, 95°C for 30 s, followed by 45 cycles at 95°C for 10 s and 63°C for 35 s. The specific primers used to detect the SARS-CoV-2 N gene were as follows: Forward: 5′-GGGGAACTTCTCCTGCTAGAAT-3′; Reverse: 5′-CAGACATTTTGCTCTCAAGCT-3′. The sequence of Taqman probe was as follows: 5′-FAM-TTGCTGCTGCTTGACAGATT-TAMRA-3′.

Mock- or SARS-CoV-2-infected Vero E6 cells were cultured for 12 h. At least 2 × 10⁷ cells were collected, lysed and protein concentration was determined by BCA protein assay kit (Beyotime, P0010). Some samples were denatured and subjected to immunoblotting procedure for input experiments. The rest of the samples were incubated with anti-P62 or IgG for 4 h at 4°C, followed by incubation with protein A-agarose for another 2 h at 4°C. The bound proteins were eluted with 1 × SDS loading buffer, denatured and analyzed by immunoblotting with anti-LC3 antibodies.

Statistical significance was determined using GraphPad Prism, Version 8.0 (GraphPad Software, San Diego, CA, United States). Data were presented as mean ± SEM in all experiments and analyzed with t test. p < 0.05 was considered statistically significant.