Senecavirus A induces mitophagy to promote self-replication through direct interaction of 2C protein with K27-linked ubiquitinated TUFM catalyzed by RNF185

ABSTRACT

Senecavirus A (SVA) is a newly emerging picornavirus associated with swine vesicular lesions and neonatal mortality, threatening the global pig industry. Despite sustained efforts, the molecular mechanisms of SVA pathogenesis have not yet been fully elucidated. Here, we demonstrate for the first time that SVA infection can induce complete mitophagy in host cells, which depends on SVA replication. Mitophagy has been subsequently proven to promote SVA replication in host cells. Genome-wide screening of SVA proteins involved in inducing mitophagy showed that although VP2, VP3, 2C, and 3A proteins can independently induce mitophagy, only the 2C protein mediates mitophagy through direct interaction with TUFM (Tu translation elongation factor, mitochondrial). The glutamic acids at positions 196 and 211 of TUFM were shown to be two key sites for its interaction with 2C protein. Moreover, TUFM was discovered to interact directly with BECN1 and indirectly with the ATG12–ATG5 conjugate. Further experiments revealed that TUFM needs to undergo ubiquitination modification before being recognized by the macroautophagy/autophagy receptor protein SQSTM1/p62, and E3 ubiquitin ligase RNF185 catalyzes K27-linked polyubiquitination of TUFM through the interaction between RNF185’s transmembrane domain 1 and TUFM to initiate SVA-induced mitophagy. The ubiquitinated TUFM is recognized and bound by SQSTM1, which in turn interacts with MAP1LC3/LC3, thereby linking the 2C-anchored mitochondria to the phagophore for sequestration into mitophagosomes, which ultimately fuse with lysosomes to achieve complete mitophagy. Overall, our results elucidated the molecular mechanism by which SVA induces mitophagy to promote self-replication and provide new insights into SVA pathogenesis.Abbreviations: aa: amino acid; Baf A1: bafilomycin A₁; BHK-21: baby hamster kidney-21; CCCP: carbonyl cyanide m-chlorophenyl hydrazone; co-IP: co-immunoprecipitation; CQ: chloroquine; DAPI: 4’,6-diamidino-2’-phenylindole; DMSO: dimethyl sulfoxide; EGFP: enhanced green fluorescent protein; ER: endoplasmic reticulum; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; GFP: green fluorescent protein; GST: glutathione S-transferase; HA: hemagglutinin; hpi: hours post-infection; hpt: hours post-transfection; IPTG: isopropyl β-D-1-thiogalactopyranoside; mAb: monoclonal antibody; MAP1LC3/LC3: microtubule associated protein 1 light chain 3; MAVS: mitochondrial antiviral signaling protein; Mdivi-1: mitochondrial division inhibitor-1; MOI: multiplicity of infection; mRFP: monomeric red fluorescent protein; MS: mass spectrometry; ORF: open reading frame; PBS: phosphate-buffered saline; SD: standard deviation; SQSTM1/p62: sequestosome 1; ST: swine testis; SVA: Senecavirus A; TCID₅₀: 50% tissue culture infectious dose; TIMM23: translocase of inner mitochondrial membrane 23; TM: transmembrane; TOMM20: translocase of outer mitochondrial membrane 20; TUFM: Tu translation elongation factor, mitochondrial; Ub: ubiquitin; UV: ultraviolet; VDAC1: voltage dependent anion channel 1; WT: wild-type; μg: microgram; μm: micrometer; μM: micromole.

Introduction

Senecavirus A (SVA) is a newly emerging causative agent of vesicular disease in pigs and acute death of newborn piglets [1]. The characteristic clinical signs include lameness and vesicles on the oral cavity, snout, nares, and hoof coronary bands [2,3]. Taxonomically, SVA belongs to the genus Senecavirus within the family Picornaviridae [4]. The virus has a linear, non-segmented, single-stranded, positive-sense RNA genome of approximately 7.3 kb that consists of a 5’-untranslated region, a large open reading frame/ORF, and a 3’-untranslated region, followed by a poly(A) tail [4]. The large open reading frame encodes a single polyprotein that is initially processed into a leader (L) protein and three precursor proteins (P1, P2, and P3), the latter three of which are subsequently cleaved into four structural proteins (VP1, VP2, VP3, VP4), three non-structural proteins (2A, 2B, 2C), and an additional four non-structural proteins (3A, 3B, 3C, 3D), respectively [1,4]. SVA was inadvertently isolated from immortalized human embryonic retinal cells during the cultivation of adenovirus-5 in the US and was presumed to be a contaminant of the cell culture [4]. The importance of SVA in the veterinary field was not well recognized until 2014 when its complete genome was identified in the vesicular fluids and sera of pigs suffering from a serious vesicular disease in Brazil [3].

Existing studies have shown that SVA not only has a wide range of tissue tropisms for various organs in pigs, including kidneys, tonsils, spleen, lungs, and liver [5], but also has a broad cell tropism and is capable of infecting a variety of cell types of diverse animal origin, such as porcine kidney epithelial (PK-15, SK6, SK-RST, and IBRS-2) cells, swine testis (ST) cells, baby hamster kidney-21 (BHK-21) cells, and tumor cells [6,7]. Notably, there is an obvious difference in the pathogenicity of historical and contemporary SVA isolates from pigs, as evidenced by the fact that the majority of contemporary isolates can cause vesicular disease and neonatal deaths in pigs, while historical isolates usually fail to reproduce the disease in experimentally infected pigs [8]. These findings indicate that SVA is evolving toward a more virulent phenotype over time, thus posing a greater threat to the global swine industry. However, what causes the increased pathogenicity of SVA in pigs and its pathogenesis have not yet been fully clarified and warrant further exploration. Screening and identifying key host and viral factors related to SVA’s pathogenicity, replication, and virulence can provide valuable clues for designing novel antiviral strategies [2,6].

Accumulating evidence suggests that extremely complex networks of crosstalk exist between SVA and host cells [9]. Multiple types of programmed cell death, including apoptosis, pyroptosis, and autophagy, have been shown to be triggered during SVA infection of host cells and to participate in regulating SVA replication [9]. Apoptosis was discovered to be activated in in vitro infected primary swine turbinate cells in late infection stages and in vivo infected skin tissues of the natural swine host, which is deemed to be a mechanism that achieves effective virus release and transmission from SVA-infected cells [10]. The SVA 3C protein itself is capable of inducing apoptosis by means of its protease activity [10]. Further studies showed that SVA can induce apoptosis in human embryonic kidney 293T cells through both extrinsic and intrinsic pathways, and both 2C and 3C proteins are involved in SVA-induced apoptosis [11]. Notably, the SVA 2C protein was found to be localized in mitochondria and interact with the anti-apoptotic protein BCL2L1/Bcl-xL [11]. Moreover, pyroptosis was demonstrated to occur in SVA-infected SK6 cells in a caspase-dependent and -independent manner, the latter of which was caused by the direct cleavage of porcine GSDMD (gasdermin D) at glutamine 277 via the protease activity of 3C protein [12]. However, although the biological significance of pyroptosis in response to SVA infection is still unclear, researchers have speculated that SVA-induced pyroptosis may contribute to its pathogenicity in pigs [12]. Compared with SVA-induced apoptosis and pyroptosis, the reciprocal regulation between SVA and autophagy seems more complex. SVA infection has been found to activate autophagy in PK-15 and BHK-21 cells through the EIF2AK3/PERK and ATF6 unfolded protein response pathways in favor of productive SVA replication [13]. Subsequent research has revealed that SVA induces autophagy by activating the AKT-AMPK-MAPK-MTOR signaling pathway through the synergistic contributions of VP1, VP3, and 3C proteins [14]. Another study has shown that autophagy induced by SVA in the early stage of infection plays an antiviral role by degrading the 3C protein, whereas the 2AB protein functions to inhibit autophagy through degradation of the MAP1LC3/LC3 (microtubule associated protein1 light chain 3) in the late stage of infection to promote SVA replication in PK-15 cells [15]. Similarly, SVA 3C has been shown to be able to antagonize the SQSTM1/p62-mediated selective autophagic degradation of VP1 and VP3 proteins through the cleavage of SQSTM1 to facilitate viral replication in SK6 and ST cells [16]. Interestingly, the autophagy induced by SVA infection has been demonstrated to promote SVA replication in porcine SK6 and ST cells but inhibit SVA replication in human 293T and H1299 cells, demonstrating the species-specific effect of cellular autophagy on SVA replication [16]. The aforementioned findings suggest that SVA has evolved diverse strategies to combat autophagy, but whether other underlying antagonistic mechanisms exist has yet to be clarified.

Mitophagy is a selective form of autophagy by which damaged, abnormal, or superfluous mitochondria are selectively sequestered by autophagosomes and then delivered to lysosomes for hydrolytic degradation, thereby achieving the quality and quantity control of mitochondria [17]. Mitophagy plays a vital role in maintaining homeostasis, regulating innate immunity, influencing various diseases, and combating microbial infections [18–20]. Although current research on the interplay between mitophagy and viral infections is still in the initial stage, a few viruses have been reported to subvert and benefit from mitophagy through different strategies [18]. For example, influenza A virus can induce mitophagy to degrade MAVS (mitochondrial antiviral signaling protein) through the interaction of viral PB1-F2 protein with TUFM (Tu translation elongation factor, mitochondrial), thereby inhibiting type I interferon production [21]. Similarly, the newly emergent severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) can also antagonize innate immunity by inducing mitophagy to degrade MAVS, which is activated by the interaction of ORF10 with BNIP3L/NIX [22]. A recent study showed that SVA infection can cause mitochondrial damage and dysfunction in PK-15 cells [23]. Our previous proteomic data also found that SVA infection caused significant changes to the abundance of several mitochondrial membrane proteins in PK-15 cells [7]. On the basis of these findings, we inferred that SVA infection can activate mitophagy in host cells.

The main goal of the present study was thus to clarify whether SVA infection can induce mitophagy, and if so, the role it has in this process and the molecular mechanism involved.

Results

SVA infection activates complete mitophagy in host cells

To analyze the ultrastructural changes of host cells in response to SVA infection, transmission electron microscopy analyses were performed on mock- and SVA-infected BHK-21 and PK-15 cells. Cells treated with carbonyl cyanide m-chlorophenyl hydrazone (CCCP) were used as a positive control for mitophagy activation [24–26]. As shown in Figure 1A, similar to the ultrastructural changes of CCCP-treated cells, SVA infection led to the formation of a large number of single- and double-membrane vesicles in the cytoplasm of BHK-21 and PK-15 cells; a decrease in the number of mitochondria; the appearance of numerous damaged mitochondria, which manifested swelling, rounding, and vacuolation; and the disappearance or rupture of mitochondrial cristae. More importantly, many recognizable mitochondria with abnormal morphology were sequestered in the mitophagic vacuoles that were generated, and some of them underwent varying degrees of fragmentation or degradation, which is typical of mitophagy activation [19]. In contrast, the cytoplasm of mock-infected BHK-21 and PK-15 cells was very dense and contained sparse vesicular structures. The morphology of most mitochondria inside these cells was normal, and the mitochondria exhibited a nearly rod-shaped morphotype with clear and distinguishable mitochondrial cristae. Further quantitative analysis showed that compared to mock-infected cells, the number of mitophagosome-like vesicles in the cytoplasm of SVA-infected cells and CCCP-treated cells significantly increased, accompanied by a significant decrease in the number of mitochondria (Figure 1A, right panels)

Figure 1.

SVA infection causes the formation of autophagosome-like vesicles enveloping damaged mitochondria in host cells. (A) transmission electron microscopic analysis of mock-, SVA-infected or CCCP-treated BHK-21 and PK-15 cells (MOI = 1, 12 h). Blue arrows indicate damaged mitochondria, while yellow arrows indicate mitophagosome-like vesicles. Representative profiles are shown and similar results were obtained in three independent experiments. The right panels show quantification of the number of mitochondria and mitophagosome-like vesicles per cell profile in mock-, SVA-infected or CCCP-treated cells. Error bars, mean ± SD for 5 randomly selected cells per experimental condition of three independent experiments (two-way ANOVA; ***P < 0.001). (B) BHK-21 and PK-15 cells were transfected with plasmid pEGFP-N2-LC3 for 12 h, and then mock infected or infected with SVA (MOI = 1) for additional 12 h. After staining with MitoTracker red CMXRos, the cells were analyzed by confocal immunofluorescence analysis using anti-VP2 mAb and Alexa Fluor plus 405-conjugated goat anti-mouse IgG as the primary and secondary antibodies, respectively. The white arrows indicate the positions where there are green puncta but no red puncta. The right panels show the fluorescence intensity profile of EGFP-LC3 (green) and MitoTracker (red) measured along the line drawn by ImageJ. Representative profiles are shown and similar results were obtained in three independent experiments. Scale bar: 10 μm. (C) BHK-21 and PK-15 cells were transfected with pmCherry-GFP-LC3 for 12 h, and then mock infected or infected with SVA (MOI = 1) for additional 12 h. The cells were analyzed by confocal immunofluorescence microscopy. The white arrows indicate the positions where there are red puncta but no green puncta. The right panels show the fluorescence intensity profile of GFP-LC3 (green) and mCherry-LC3 (red) measured along the line drawn by ImageJ. Representative profiles are shown and similar results were obtained in three independent experiments. Scale bar: 10 μm. (D) BHK-21 and PK-15 cells were transfected with pmRFP-GFP-Mito for 12 h, and then mock infected or infected with SVA (MOI = 1) for an additional 12 h. The cells were analyzed by confocal immunofluorescence microscopy. The white arrows indicate the positions where there are red puncta but no green puncta. The right panels show the fluorescence intensity profile of mito-GFP (green) and mito-RFP (red) measured along the line drawn by ImageJ. Representative profiles are shown and similar results were obtained in three independent experiments. Scale bar: 10 μm.

To confirm the activation of mitophagy in BHK-21 and PK-15 cells upon infection by SVA, confocal immunofluorescence analyses were carried out to observe the colocalization of mitochondria with autophagosomes using an enhanced green fluorescent protein (EGFP)-LC3 (for autophagosome labeling) and MitoTracker Red CMXRos co-labeling approach, which is commonly used for proving the occurrence of mitophagy [19,26]. To ensure successful viral infection during each experiment, a monoclonal antibody (mAb; 2F5) that specifically recognizes the SVA VP2 protein was used to indicate the SVA-infected cells [7]. As shown in Figure 1B, the green fluorescent signals of EGFP-LC3 were diffusely distributed throughout the entire cytoplasm of mock-infected BHK-21 and PK-15 cells. The red mitochondrial staining signals of these mock cells, though punctate, were evenly dispersed throughout the cytoplasm. However, after SVA infection, the EGFP-LC3 fluorescent signals developed a distinctly punctate perinuclear distribution, suggestive of autophagosome formation [19,26], while the mitochondria congregated around nuclei and formed larger puncta. Of note, a subset of green EGFP-LC3 puncta did not colocalize with the red fluorescent puncta of mitochondria in SVA-infected BHK-21 and PK-15 cells. Specifically, at the location where the EGFP-LC3 puncta appeared, the corresponding red mitochondrial puncta disappeared, indicative of the autophagic removal of mitochondria. These results reveal that mitochondrial loss occurred concurrently with or very rapidly after the onset of autophagy, indicating that SVA infection activated mitophagy in host cells.

To address whether SVA infection induces complete mitophagy, a pH-sensitive tandem fluorescent reporter construct – pmCherry-GFP-LC3 [21]—was used to detect changes in autophagic flux upon SVA infection. This dual-color fluorescent probe uses the different stability of mCherry and GFP (green fluorescent protein) in acidic lysosomal environments to discriminate between autophagosomes and autolysosomes. The GFP is quenched in the acidic autolysosome, while mCherry remains fluorescent. Therefore, the yellow puncta formed by the colocalization of red mCherry signal with green GFP signal usually indicate that the generated autophagosome has not fused with the lysosome, whereas the appearance of red mCherry-only puncta without green GFP puncta is considered to indicate autolysosome formation [26]. As shown in Figure 1C, SVA infection caused the transformation of the dual-fluorescence signal of LC3 from a diffuse cytoplasmic distribution to a discrete punctate cytoplasmic distribution. Although the majority of the red LC3 fluorescent puncta colocalized well with the corresponding green LC3 fluorescent puncta in SVA-infected BHK-21 and PK-15 cells, numerous red LC3 fluorescent puncta still existed alone without colocalizing with green puncta. This reveals that at least a portion of autophagosomes fused with lysosomes to form autolysosomes, which caused the fluorescence quenching of GFP. Furthermore, another pH-sensitive dual fluorescent reporter construct, pmRFP-GFP-Mito [21], expressing tandem monomeric red fluorescent protein (mRFP) and GFP that were fused with mitochondrial outer membrane FIS1 (fission, mitochondrial 1) protein, was used to analyze whether the mitophagosomes produced in SVA-infected cells eventually fused with lysosomes. Under steady-state conditions, mRFP-GFP-Mito can simultaneously emit red and green fluorescence; however, once complete mitophagy occurs, the green GFP signal is quenched when the mitochondria are delivered to the lysosome, which then emit only red mRFP fluorescence signals [19]. As shown in Figure 1D, the dual fluorescence signals of the mitochondria were evenly distributed in a discrete punctate formation in the cytoplasm of mock-infected BHK-21 and PK-15 cells and appeared as yellow puncta after merging. In contrast, SVA infection not only led to a decrease in the number of mitochondria but also altered the distribution of mitochondria, which congregated around nuclei and formed larger fluorescent puncta. Moreover, SVA infection also resulted in the appearance of numerous mCherry-only red fluorescent puncta, suggesting that some mitophagosomes completed fusion with lysosomes to form mitolysosomes in SVA-infected cells. These results demonstrate that SVA infection can induce a complete mitophagic process.

To further confirm mitophagy activation was triggered by SVA infection, western blot analyses were conducted to detect the conversion of LC3-I to LC3-II and the autophagic degradation of SQSTM1, TOMM20 (translocase of outer mitochondrial membrane 20), and TIMM23 (translocase of inner mitochondrial membrane 23), all of which reflect the activation of mitophagy [26]. As shown in Figure 2A–D, an obvious conversion of LC3 occurred in both BHK-21 and PK-15 cells as SVA infection progressed, which manifested as the band intensity of LC3-I gradually decreasing while that of LC3-II gradually increased, a typical hallmark of autophagy activation [26]. With the activation of autophagy, the expression levels of SQSTM1, TIMM23, and TOMM20 proteins gradually decreased in SVA-infected cells. By quantifying the optical density of protein bands and conducting statistical analysis on the obtained values, the densitometric ratios of LC3-II to ACTB/β-actin significantly increased, whereas those of SQSTM1, TIMM23, and TOMM20 to ACTB significantly decreased in SVA-infected BHK-21 and PK-15 cells from 12 hours post-infection (hpi) onward compared with those of mock-infected cells (Figure 2B–D). Next, in order to verify that the degradation of mitochondrial markers was indeed related to autophagy, we disrupted the autophagy-lysosomal degradation activity of BHK-21 and PK-15 cells using bafilomycin A₁ (Baf A1) and chloroquine (CQ) [26], and then detected the expression of TIMM23 and TOMM20 proteins by western blot. Compared with the cells treated with solvent dimethyl sulfoxide (DMSO) or solvent-free control (SFC), Baf A1 and CQ treatments significantly rescued the expression of TIMM23 and TOMM20 proteins in BHK-21 and PK-15 cells, regardless of whether the two cell lines were infected with SVA or not (Fig. S1A-D). Protein band optical density analyses further confirmed that the densitometric ratios of LC3-II, TIMM23, TOMM20 and SQSTM1 to ACTB significantly increased, whereas those of VP2 to ACTB significantly decreased upon treatment with Baf A1 or CQ (Fig. S1B and D). These results not only confirm that the degradation of mitochondrial markers was indeed related to mitophagy, but also suggest that mitophagy might promote SVA replication. Moreover, to further verify whether SVA infection can also induce mitophagy in other permissive cells, the response of another two porcine cell lines, IBRS-2 and ST, to SVA infection was also analyzed by western blot. Similar results were also obtained in SVA-infected IBRS-2 and ST cells (Fig. S2). Overall, these results indicate that mitophagy was successfully activated in SVA-infected host cells. Since LC3 is recruited to mitochondria through the interaction with autophagy receptor proteins, such as SQSTM1, when mitophagy occurs [20], and therefore we employed immunoblotting to detect the translocation of LC3 from the cytoplasm to mitochondria using separated mitochondrial and cytoplasmic fractions. GAPDH (glyceraldehyde-3-phosphate dehydrogenase) and mitochondrial outer membrane protein VDAC1 (voltage dependent anion channel 1) were used as loading controls for the cytoplasmic and mitochondrial fractions, respectively [26]. Cells treated with CCCP were used as a positive control for mitophagy activation [24–26]. Compared with the mock-infected or DMSO-treated BHK-21 cells, either CCCP treatment or SVA infection significantly increased the amount of LC3-I to LC3-II conversion, not only in the whole cell lysates but also in their mitochondrial and cytoplasmic fractions (Figure 2E). Most importantly, the degree of LC3 conversion in the mitochondrial fractions was significantly higher than that in the corresponding cytoplasmic fractions, which was confirmed by an analysis of the relative densitometric ratios of LC3-II and the corresponding internal references (Figure 2F). These results demonstrate that LC3-II was recruited to the mitochondria after CCCP treatment or SVA infection. Interestingly, we also observed the VP2 protein of SVA is highly associated with the mitochondrial fraction, suggesting that VP2 May interact with mitochondria. A similar phenomenon was observed in SVA-infected PK-15 cells (Figure 2G,H). Taken together, these results clearly demonstrate that SVA infection induces complete mitophagy in host cells.

Figure 2.

SVA infection promotes LC3 conversion accompanied by degradation of mitochondrial membrane proteins. (A) BHK-21 cells were mock infected or infected with SVA (MOI = 0.1). At 0, 12, 18, 24, 30, or 36 hpi, the cells were analyzed by western blot. Representative results are shown and similar results were obtained in three independent experiments. (B) Densitometry of the target protein bands shown in (A) and calculation of their relative densitometric ratios to ACTB/β-actin. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; **P < 0.01; ***P < 0.001). (C) PK-15 cells were infected and analyzed as described in (A). (D) Densitometric analysis of the target protein bands shown in (C). (E) BHK-21 cells were mock infected or infected with SVA (MOI = 0.1) or treated with 10 μM of CCCP or DMSO solvent control for 24 h. The cells were then subjected to isolate mitochondria and cytoplasmic fractions. The expression of LC3, VP2, VDAC1, and GAPDH in the whole-cell lysates (WCL), cytoplasmic (Cyto) and mitochondrial (Mito) fractions was detected by western blot. Representative results are shown and similar results were obtained in three independent experiments. (F) Densitometric analysis of the target protein bands shown in (E). Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; ***P < 0.001). (G) PK-15 cells were infected and analyzed as described in (E). (H) densitometric analysis of the target protein bands shown in (G). Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; ***P < 0.001).

The replication of SVA is indispensable for the induction of mitophagy

To address whether the SVA inducement of mitophagy depends on viral replication, replication-competent native SVA was inactivated by ultraviolet (UV) irradiation as previously described [27], and the ability of inactivated SVA to induce mitophagy was evaluated. Prior to the formal experiments, an indirect immunofluorescence assay was conducted to demonstrate the loss of infectivity of SVA after UV treatment, and to determine whether the inactivated SVA can enter the cells. As shown in Figure 3A, compared with the native SVA-infected BHK-21 and PK-15 cells, no VP2-staining fluorescent signal was observed in the mock-infected cells or cells inoculated with UV-inactivated SVA, indicating that UV-inactivated SVA completely lost the ability to infect and enter host cells. Figure 3B indicates that only the LC3 proteins in replication-competent native SVA-infected BHK-21 cells underwent obvious conversion from LC3-I to LC3-II, and the degree of LC3 conversion was enhanced with an increase in SVA infectious dose. Notably, as the SVA infectious dose increased, the degradation of SQSTM1, TIMM23, and TOMM20 proteins also increased. In contrast, the respective amounts of LC3-I, LC3-II, SQSTM1, TIMM23, and TOMM20 proteins in BHK-21 cells inoculated with low, medium, and high doses of UV-inactivated SVA were comparable to those in mock-infected cells, and no significant conversion from LC3-I to LC3-II and no VP2 protein synthesis were detected. Relative densitometric analyses of the protein bands further confirmed that it was the native SVA, rather than the UV-inactivated SVA, that caused a statistically significant conversion of LC3 and degradation of SQSTM1, TIMM23, and TOMM20 proteins in a dose-dependent manner (Figure 3C). The inactivated SVA produced similar experimental results in PK-15 cells (Figure 3D, E). Furthermore, confocal immunofluorescence analyses using the EGFP-LC3 and MitoTracker Red CMXRos co-labeling approach were also employed to evaluate the ability of UV-inactivated SVA to induce mitophagy. As shown in Figure 3F, G, the distribution of EGFP-LC3 and mitochondrial staining signals in BHK-21 and PK-15 cells inoculated with UV-inactivated SVA resembled those of their respective mock-infected cells. Only the native SVA was able to cause the redistribution of EGFP-LC3 and mitochondrial staining signals, which were mainly distributed around the nuclei as larger puncta. Altogether, these results indicate that the UV-inactivated SVA lost its ability to induce mitophagy due to its inability to enter the cells, meaning that the replication of SVA is required for the induction of mitophagy.

Figure 3.

Viral replication is required for SVA to induce mitophagy. (A) BHK-21 and PK-15 cells were mock infected or inoculated with native SVA or UV-inactivated SVA (MOI = 1) for 24 h. The cells were analyzed by indirect immunofluorescence assays using anti-VP2 mAb and Alexa Fluor 488-conjugated goat anti-mouse IgG as the primary and secondary antibodies, respectively. Cell nuclei were counterstained with DAPI. Scale bar: 50 μm. Representative results are shown and similar results were obtained in three independent experiments. (B) BHK-21 cells were mock infected or inoculated with native SVA or UV-inactivated SVA (MOI = 0.1, 1, or 5) for 24 h. The cells were analyzed by western blot. Representative results are shown and similar results were obtained in three independent experiments. (C) Densitometry of the target protein bands shown in (B) and calculation of their relative densitometric ratios to ACTB/β-actin. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; ***P < 0.001). (D) PK-15 cells were infected and analyzed as described in (B). (E) Densitometric analysis of the target protein bands shown in (D). (F) BHK-21 cells were transfected with plasmid pEGFP-N2-LC3 for 12 h, and then mock infected or inoculated with native SVA or UV-inactivated SVA (MOI = 1) for additional 12 h. After staining with MitoTracker red CMXRos, the cells were analyzed by confocal immunofluorescence assay using anti-VP2 mAb and Alexa Fluor plus 405-conjugated goat anti-mouse IgG as the primary and secondary antibodies, respectively. The right panels show the fluorescence intensity profile of EGFP-LC3 (green) and MitoTracker (red) measured along the line drawn by ImageJ. Representative results are shown and similar results were obtained in three independent experiments. Scale bar: 10 μm. (G) PK-15 cells were infected and analyzed as described in (F).

Mitophagy promotes the replication of SVA in host cells

To explore what role mitophagy plays in SVA replication, we first activated mitophagy using CCCP prior to SVA infection, and then analyzed its effect on VP2 protein expression and viral progeny yields using western blot and 50% tissue culture infectious dose (TCID₅₀) assays, respectively. Compared with the mock-infected BHK-21 and PK-15 cells, cells with SVA infection showed not only LC3 conversion but also degradation of SQSTM1, TIMM23, and TOMM20 proteins at the indicated times post-infection (Figure 4A–F), indicating the successful activation of mitophagy. Of note, CCCP treatment further exacerbated the degree of conversion from LC3-I to LC3-II and the degradation of SQSTM1, TIMM23, and TOMM20 proteins in SVA-infected cells, which were accompanied by an obvious increase in the expression of SVA VP2 protein (Figure 4A–F). Protein band optical density analyses further confirmed that the densitometric ratios of LC3-II and VP2 to ACTB significantly increased, whereas those of TIMM23 and TOMM20 to ACTB significantly decreased in CCCP-pretreated and SVA-infected BHK-21 and PK-15 cells at most of the indicated times post-infection, compared with mock-infected cells or cells only infected with SVA (Figure 4B–F). Through virus titrations, we further discovered that CCCP treatment significantly enhanced progeny virus yields at the indicated times post-infection, with peak titers appearing at 48 and 36 hpi, while a maximum titer difference of up to approximately 19-fold appeared at 72 and 60 hpi in SVA-infected BHK-21 and PK-15 cells, respectively (Figure 4C,G). In addition, we further analyzed the effect of CCCP treatment on both intracellular and extracellular SVA titers. As shown in Figure 4D,H, CCCP treatment significantly increased intracellular virus yields in SVA-infected BHK-21 and PK-15 cells, but had no significant effect on extracellular virus yields. These results demonstrate that CCCP treatment further enhanced the level of mitophagy in SVA-infected BHK-21 and PK-15 cells and that mitophagy promoted the replication of SVA in host cells.

Figure 4.

Induction of mitophagy promotes SVA replication. (A) BHK-21 cells were pretreated with 10 μM of CCCP or DMSO solvent control for 4 h, and then mock infected or infected with SVA (MOI = 0.1). The cells were further cultured in fresh medium supplemented with or without 10 μM of CCCP. At 0, 12, 18, and 24 hpi, the cells were analyzed by western blot. Representative results are shown and similar results were obtained in three independent experiments. (B) Densitometry of the target protein bands shown in (A) and calculation of their relative densitometric ratios to ACTB/β-actin. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; *P < 0.05; **P < 0.01; ***P < 0.001). (C) BHK-21 cells were pretreated and infected as described in (A). At 12, 24, 36, 48, 60, and 72 hpi, both cells and supernatants were harvested and freeze-thawed three times to release intracellular viruses. After centrifugation, the viral titer of the supernatants was detected by TCID₅₀ assay on BHK-21 cells. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; **P < 0.01; ***P < 0.001). (D) BHK-21 cells were pretreated and infected as described in (A). At 12, 24, 36 and 48 hpi, the supernatants were harvested as extracellular viruses, and the remaining cells were washed thrice with PBS and then freeze-thawed three times in 1 mL of DMEM. After centrifugation, the resultant supernatants were harvested as intracellular viruses. Both extracellular and intracellular viral titers were determined by TCID₅₀ assay on BHK-21 cells. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; **P < 0.01; ***P < 0.001). (E) PK-15 cells were pretreated, infected, and analyzed as described in (A). (F) Densitometric analysis of the target protein bands shown in (E). (G) PK-15 cells were pretreated, infected, and analyzed as described in (C). (H) PK-15 cells were pretreated, infected, and analyzed as described in (D).

As a reverse validation, we used mitochondrial division inhibitor-1 (Mdivi-1), a widely used mitophagy inhibitor [28,29], to downregulate mitophagy and analyzed its impact on SVA replication. The results show that Mdivi-1 treatment significantly alleviated the level of mitophagy in BHK-21 and PK-15 cells caused by SVA infection (Figure 5A–F), which was reflected in the significant decline in the amount of LC3-II and its relative ratio to ACTB, as well as a significant increase in the amount of SQSTM1, TIMM23, TOMM20 and their relative ratios to ACTB (Figure 5B, E). These results indicate that the SVA-induced activation of mitophagy was partially inhibited by Mdivi-1 treatment. The inhibition of mitophagy ultimately led to a significant decrease in the expression of SVA VP2 protein and the SVA progeny yield at the indicated time points (Figure 5B, C, E and F). Interestingly, the effect of Mdivi-1 treatment on SVA virus titer seemed stronger than that of CCCP, with peak titers appearing at 48 and 36 hpi, while maximum titer differences of up to approximately 40.8-fold and 35.9-fold appeared at 48 and 12 hpi in SVA-infected BHK-21 and PK-15 cells, respectively. Collectively, these results confirm that SVA-induced mitophagy benefits SVA replication in host cells.

Figure 5.

Inhibition of mitophagy suppresses SVA replication. (A) BHK-21 cells were pretreated with 10 μM of Mdivi-1 or DMSO solvent control, and then mock infected or infected with SVA (MOI = 0.1). The cells were further cultured in fresh medium supplemented with or without 10 μM of Mdivi-1. At 0, 12, 18, and 24 hpi, the cells were analyzed by western blot. Representative results are shown and similar results were obtained in three independent experiments. (B) Densitometry of the target protein bands shown in (A) and calculation of their relative densitometric ratios to ACTB/β-actin. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; *P < 0.05; **P < 0.01; ***P < 0.001). (C) BHK-21 cells were pretreated and infected as described in (A). At 12, 24, 36, 48, 60, and 72 hpi, both cells and supernatants were harvested and freeze-thawed three times to release intracellular viruses. After centrifugation, the viral titer of the supernatants was detected by TCID₅₀ assay on BHK-21 cells. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; **P < 0.01; ***P < 0.001). (D) PK-15 cells were pretreated, infected, and analyzed as described in (A). (E) Densitometric analysis of the target protein bands shown in (D). (F) PK-15 cells were pretreated, infected, and analyzed as described in (C).

Individual protein of VP2, VP3, 2C, and 3A functions to induce mitophagy

To screen which SVA protein(s) is (are) involved in the induction of mitophagy at the whole-genome level, a series of eukaryotic plasmids expressing hemagglutinin (HA)-tagged proteins of SVA were individually transfected into BHK-21 cells to induce the transient expression of the related proteins, and the cells were then harvested for the detection of LC3 conversion. As shown in Figure 6A, B, LC3 conversion from LC3-I to LC3-II only occurred in the cells transfected with plasmids expressing VP2, VP3, 2AB, 2B, 2C, 3A, or 3AB but not in the mock-transfected cells, cells transfected with the empty vector (pCAGGS-HA), or cells with plasmids expressing L, VP4, VP1, 3C, or 3D. For further confirmation, confocal immunofluorescence analyses were also performed to analyze the changes in EGFP-LC3 distribution signals in BHK-21 cells co-transfected with pEGFP-N2-LC3 and each of the plasmids expressing SVA proteins. As shown in Figure 6C, only the cells co-transfected with pEGFP-N2-LC3 and plasmids expressing VP2, VP3, 2AB, 2B, 2C, 3A, or 3AB caused the formation of a large number of EGFP-LC3 puncta with a perinuclear distribution, which resembled those caused by SVA infection (Figure 1B). In contrast, the mock-transfected cells or cells co-transfected with pEGFP-N2-LC3 or plasmids expressing L, VP4, VP1, 3C, or 3D exhibited a diffuse cytoplasmic distribution of EGFP-LC3 signal. After combining the results of western blot and confocal immunofluorescence analyses, we concluded that the VP2, VP3, 2B, 2C, and 3A proteins of SVA can induce autophagy. To further determine which of these protein(s) induces mitophagy, three doses of eukaryotic plasmids expressing HA-tagged VP2, VP3, 2B, 2C, or 3A were separately transfected into BHK-21 cells. The resultant total cell proteins were subjected to western blot analyses alongside with densitometric analyses of protein bands. Compared with the mock- or empty vector (pCAGGS-HA)-transfected cells, cells transfected with plasmids expressing VP2, VP3, 2C, or 3A proteins displayed significant amounts of LC3 conversion from LC3-I to LC3-II and degradation of SQSTM1, TIMM23, and TOMM20 proteins in a dose-dependent manner (Figure 6D and Fig. S3) and resembled cells with SVA infection (Figure 2A). These results reveal that the SVA VP2, VP3, 2C, and 3A proteins can induce mitophagy. Interestingly, transfection of BHK-21 cells with a 2B protein-expressing plasmid only caused LC3 conversion but did not result in the degradation of SQSTM1, TIMM23, and TOMM20 proteins (Figure 6D, middle panel), indicating that the SVA 2B protein can only induce autophagy, not mitophagy. The ability of VP2, VP3, 2C, and 3A proteins to induce mitophagy was further confirmed in PK-15 cells (Fig. S4 and S5). Taken together, these data suggest that the SVA VP2, VP3, 2C, and 3A proteins can induce mitophagy in both BHK-21 and PK-15 cells.

Figure 6.

Single proteins of VP2, VP3, 2C, and 3A can independently induce mitophagy in BHK-21 cells. (A) BHK-21 cells were individually transfected with recombinant pCAGGS-HA plasmids expressing HA-tagged SVA proteins. At 36 hpt, the cells were analyzed by western blot. The target protein bands are indicated by red triangles. (B) Densitometry of the target protein bands shown in (A) and calculation of their relative densitometric ratios to ACTB/β-actin. Representative results are shown, and similar results were obtained in three independent experiments. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ***P < 0.001). (C) BHK-21 cells were co-transfected with pEGFP-N2-LC3 and each of the recombinant pCAGGS-HA plasmids expressing HA-tagged SVA proteins. At 24 hpt, the cells were analyzed by confocal immunofluorescence assays using anti-HA mAb and Alexa Fluor 568-conjugated goat anti-mouse IgG as the primary and secondary antibodies, respectively. Cell nuclei were counterstained with DAPI. The right panels show the fluorescence intensity profile of EGFP-LC3 (green) and HA-tagged SVA proteins (red) measured along the line drawn by ImageJ. Representative results are shown, and similar results were obtained in three independent experiments. Scale bar: 10 μm. (D) BHK-21 cells were mock transfected or transfected with empty vector (pCAGGS-HA) or with 1, 2, and 3 µg/well of each of the recombinant plasmids pCAGGS-HA-VP2, pCAGGS-HA-VP3, pCAGGS-HA-2B, pCAGGS-HA-2C, and pCAGGS-HA-3A. At 36 hpt, the cells were analyzed by western blot. Representative results are shown, and similar results were obtained in three independent experiments.

The SVA 2C protein induces mitophagy via direct interaction with TUFM protein

To elucidate the mechanism by which SVA induces mitophagy, co-immunoprecipitation (co-IP) assay was performed using transiently expressed HA-tagged VP2, VP3, 2C, and 3A proteins as bait, along with anti-HA magnetic beads. The obtained immune complexes were first separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (Figure 7A) and subsequently identified by mass spectrometry (MS) analysis. We screened the MS data, and all mitochondrial proteins that potentially interact with the mitophagy-inducing SVA proteins and have been shown to be related to autophagy are listed in Figure 7B. Among these, the autophagy-promoting protein TUFM, which resides in mitochondria and contains three key structural domains (I – III), has been proven to regulate not only autophagy but also mitophagy [21,30]. It was therefore selected as a target for subsequent experiments. A series of co-IP assays were performed to validate the potential interactions between the TUFM protein and the identified VP2, VP3, 2C, or 3A proteins. As shown in Figure 7C, endogenous TUFM coimmunoprecipitated with HA-VP2, HA-VP3, and HA-2C proteins but not with HA-3A in BHK-21 cells. For further verification, both Flag-tagged TUFM and HA-tagged VP2, VP3, or 2C were ectopically expressed in 293T cells and then analyzed by reciprocal co-IP assays (Figure 7D,E). The results indicate that TUFM indeed interacted with VP2, VP3, and 2C proteins. To further determine whether TUFM interacts directly or indirectly with VP2, VP3, and 2C, we performed an in vitro Strep II tag affinity-isolation assay using a prokaryotically expressed Strep II-TUFM fusion protein with a purity of approximately 94.63% (Fig. S6) and eukaryotically expressed HA-tagged VP2, VP3, or 2C protein in 293T cells. As shown in Figure 7F, Strep II-TUFM only pulled down 2C and did not pull down VP2 or VP3, indicating a direct interaction between 2C and TUFM. To further verify this, another in vitro glutathione S-transferase (GST) affinity-isolation assay was conducted using two prokaryotically expressed proteins, Strep II-TUFM (~94.63% purity) and GST-2C (~90.33% purity). The results showed that it was the GST-2C fusion protein, rather than the GST tag, that pulled down Strep II-TUFM in a dose-dependent manner, which confirms that 2C directly interacted with TUFM (Figure 7G). In addition, we additionally selected two other porcine cell lines for ectopic expression of 2C protein to further validate the role of 2C protein in inducing mitophagy. To do this, high, medium, and low dose of plasmid pCAGGS-HA-2C expressing HA-tagged full-length 2C protein were transfected into IBRS-2 and ST cells. Their lysates were detected by western blot to observe changes in mitochondrial markers. The results showed that the ectopic expression of 2C protein in IBRS-2 and ST cells caused a significant LC3 conversion and degradation of SQSTM1, TIMM23, and TOMM20 proteins in a dose-dependent manner (Fig. S7A and S7B), which resembled those observed in BHK-21 and PK-15 cells overexpressing 2C protein (Figure 6D and Fig. S4B). Moreover, we further evaluated the effect of 2C protein on mitophagy by transfecting BHK-21 or PK-15 cells with a recombinant plasmid pCAGGS-HA-2C expressing 2C protein before infection with SVA. Our data showed that compared with the cells only transfected with 2C protein, SVA infection further exacerbated the degree of LC3 conversion and degradation of SQSTM1, TIMM23, and TOMM20 proteins (Fig. S7C and S7D). Collectively, our results clearly indicate that the 2C protein of SVA can indeed induce mitophagy in host cells.

Figure 7.

SVA 2C protein induces mitophagy via direct interaction with TUFM protein. (A) BHK-21 cells were individually transfected with plasmids pCAGGS-HA-VP2, pCAGGS-HA-VP3, pCAGGS-HA-2C, pCAGGS-HA-3A, or pCAGGS-HA for 36 h. The cells were analyzed by co-IP assays using anti-HA magnetic beads. The obtained immune complexes were separated by SDS-PAGE, and proteins were visualized by silver staining and identified by mass spectrometry. (B) List of cellular proteins identified by mass spectrometry that potentially interact with VP2, VP3, 2C, and 3A and participate in inducing mitophagy. Percentage (%) refers to the ratio of the number of amino acids identified in the peptide segment to the full length of the indicated proteins. (C) BHK-21 cells were individually transfected with plasmids pCAGGS-HA-VP2, pCAGGS-HA-VP3, pCAGGS-HA-2C, pCAGGS-HA-3A, or pCAGGS-HA for 36 h. The cells were analyzed by co-IP assays using rabbit anti-TUFM polyclonal antibody plus protein A/G magnetic beads, followed by western blot analyses. (D) 293T cells were co-transfected with p3×Flag-CMV-TUFM and one of the following plasmids pCAGGS-HA-VP2, pCAGGS-HA-VP3, pCAGGS-HA-2C, or empty vector pCAGGS-HA for 36 h. The cells were analyzed by co-IP assays using anti-HA magnetic beads, followed by western blot analyses. (E) 293T cells were transfected and analyzed by co-IP assay as described in (D) but using anti-Flag magnetic beads. (F) 293T cells were individually transfected with plasmids pCAGGS-HA-VP2, pCAGGS-HA-VP3, pCAGGS-HA-2C, or pCAGGS-HA for 36 h. The cells were harvested, lysed, and supernatants of the cell lysates were incubated with the prokaryotically expressed Strep II-TUFM protein bound to Streptactin beads 4FF. The formed complexes were analyzed by western blot analyses. (G) Ten micrograms of purified prokaryotically expressed GST-2C fusion protein or GST tag were bound to glutathione Sepharose 4B, which were then incubated with 5, 10, or 20 μg of purified Strep II-TUFM protein. The formed complexes were analyzed by western blot analyses. (H) BHK-21 cells were mock transfected (MT) or transfected with 40 pmol/well of siTUFM-hamster or siNC for 24 h. Then the cells were mock infected or infected with SVA (MOI = 0.1) for an additional 24 h. The cells were analyzed by western blot analyses. (I) Densitometry of the target protein bands shown in (H) and calculation of their relative densitometric ratios to ACTB/β-actin. (J) BHK-21 cells were mock transfected (MT) or transfected with p3×Flag-CMV-10 or 1, 2, or 3 μg/well of p3×Flag-CMV-TUFM. At 24 hpt, the cells were mock infected or infected with SVA (MOI = 0.1) for 24 h. The cells were analyzed by western blot analyses. (K) Densitometric analysis of the target proteins shown in (J). (L) BHK-21 cells were transfected and then infected as described in (H). The silencing of TUFM was confirmed by western blot prior to virus titration. At 12, 24, 36, and 48 hpi, viral samples were harvested, and viral titers were detected by TCID₅₀ assay on BHK-21 cells. (M) BHK-21 cells were transfected and then infected as described in (J). Viral titrations were performed and analyzed as described in (L). All experiments related to this figure were conducted in three independent biological replicates. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; *P < 0.05; **P < 0.01; ***P < 0.001).

To explore the role of TUFM protein in SVA-induced mitophagy, we evaluated its impact on mitophagy and SVA replication by knocking down or overexpressing TUFM in BHK-21 cells. As shown in Figure 7H, I, the expression of endogenous TUFM was significantly knocked down by the designed TUFM-specific siRNA (siTUFM-hamster), rather than the scrambled siRNA (siNC), and this was accompanied by a significant decrease in LC3-II and VP2 and a significant increase in SQSTM1, TIMM23, and TOMM20. These results reveal that TUFM is positively correlated with SVA-induced mitophagy. In contrast, overexpression of TUFM in BHK-21 cells caused a statistically significant conversion of LC3-I to LC3-II and degradation of SQSTM1, TIMM23, and TOMM20 proteins in a dose-dependent manner, which were accompanied by a significant increase in VP2 expression following SVA infection (Figure 7J, K). Moreover, virus titration further confirmed that knockdown of TUFM by siRNA significantly decreased SVA yields in BHK-21 cells from 12 hpi onward (Figure 7L), while overexpression of TUFM significantly increased SVA yields from 24 hpi onward (Figure 7M) compared to the mock-, siNC-, or empty vector-transfected cells. Notably, the effect of interfering with TUFM expression through siTUFM-pig or TUFM overexpression on mitophagy and SVA replication in PK-15 cells is completely consistent with the results obtained in BHK-21 cells (Fig. S8). Overall, our results indicate that, although the VP2, VP3, and 2C proteins of SVA can independently induce mitophagy, only the 2C protein mediates mitophagy through direct interaction with the TUFM protein, which acts as a positive regulatory factor for SVA-induced mitophagy.

The amino acids 56–252 of TUFM and 231–321 of 2C are key interaction regions, with E196 and E211 of TUFM being two key sites mediating interaction with 2C

To identify the key regions or amino acid (aa) sites essential for the interaction between TUFM and 2C proteins, we first constructed nine TUFM truncation mutants—∆I, ∆II, ∆III, aa1–252, aa1–55, aa56–252, aa275–452, aa275–344, and aa345–452—the details of which are illustrated in Figure 8A. These constructs, together with the wild-type (WT) TUFM construct, were used to transfect 293T cells, the lysates of which were analyzed by co-IP assays. As shown in Figure 8B, only those constructs containing aa56–252 were able to coimmunoprecipitate with the 2C protein, while those lacking this domain failed to coimmunoprecipitate, suggesting that TUFM interacts with 2C through its amino acid regions 56–252. Similarly, we first constructed six truncated mutants of 2C protein – aa1–175, aa1–69, aa70–175, aa176–321, aa212–321, and aa231–321 (Figure 8C) – and analyzed their ability to interact with TUFM using co-IP assays. The results showed that the constructs WT 2C, aa176–321, aa212–321, and aa231–321 coimmunoprecipitated with TUFM, whereas aa1–175, aa1–69, and aa70–175 did not (Figure 8D), indicating that aa231–321 of the 2C protein is a key domain for interaction with the TUFM protein. To further validate the interaction between TUFM and 2C proteins, confocal immunofluorescence assay was performed to analyze the co-localization of full-length TUFM and 2C proteins as well as their respective truncated mutants aa56–252 and aa231–321 in co-transfected BHK-21 cells. As shown in Fig. S9A, not only the full-length TUFM and 2C proteins, but also their respective truncated mutants aa56–252 and aa231–321, were highly colocalized in the co-transfected BHK-21 cells, further confirming the interaction between TUFM and 2C. Next, we made an attempt to further shorten the 231–321 amino acid region of the 2C protein interacting with TUFM by constructing two additional truncations aa256–321 and aa285–321; however, neither truncation could coimmunoprecipitate TUFM protein (Figure 8E). In view of this, we took a step further to analyze the aa56–252 of TUFM and aa231–321 of 2C using the Alphafold2 online server (https://alphafold.ebi.ac.uk/) with the aim of predicting the potential interaction interfaces and residues of TUFM and 2C proteins. A total of five predicted structural models of the TUFM-2C complex were obtained with the optimal model shown in Figure 8F and the other four models shown in Fig. S9B-E. The optimal model was selected for subsequent analysis according to the pLDDT value. The predicted results indicate that salt bridges and hydrogen bonds rather than hydrophobic bonds are involved in the interaction between TUFM and 2C (Figure 8F). Specifically, the glutamic acid (abbreviated as E or Glu) at positions 193 (E193), 196 (E196), and 211 (E211) of the TUFM protein forms a salt bridge with arginine residues at positions 286 (R286) and 308 (R308) of the 2C protein, respectively. Whereas the arginine at position 200 (R200) and the glutamic acid residues at positions 196 (E196) and 212 (E212) of the TUFM protein form a hydrogen bond with the serine at position 293 (S293) and asparagine at position 240 (N240) of the 2C protein, respectively. To identify which of the five predicted amino acid residues of TUFM determines its interaction with 2C, five TUFM mutants were constructed by introducing a single amino acid mutation at position 200 (R→A), 193 (E→A), 196 (E→A), 212 (E→A), and 211 (E→A), and the ability of the mutants to interact with 2C protein was analyzed using co-IP assays. As shown in Figure 8G, the WT TUFM and its derivative mutants R200A, E193A, and E212A coimmunoprecipitated with 2C, while the mutants E196A and E211A did not. This result reveals that the E196 and E211 amino acid residues of TUFM are key binding sites for interactions with the 2C protein. Similarly, four 2C mutants with a single-site mutation of N240A, R286A, S293A, and R308A were also constructed and analyzed by co-IP assays. The results show that any single mutation at the four predicted amino acid residues of 2C protein did not affect its interaction with TUFM (Figure 8H). For further exploration, we constructed additional eleven single-site mutants of 2C, each of which contained one of the potential key sites predicted in the other four models (Fig. S9B-E), and analyzed their interaction with TUFM by co-IP assays. Figure 8I shows that any single mutation of the eleven predicted amino acid residues in the aa231–321 region of 2C protein did not affect its interaction with TUFM. Overall, despite our tremendous efforts, the potential key sites of 2C protein function to interact with TUFM have not yet been identified. Taken together, the 56–252 amino acids of TUFM and the 231–321 amino acids of 2C are critical regions mediating their interaction, and the E196 and E211 amino acid residues of TUFM are two key sites for the interaction with 2C.

Figure 8.

The amino acids 56–252 of TUFM protein and 231–321 of 2C protein are the key regions for their interaction. (A) Schematic diagram of domain structure of TUFM and construction strategy of Flag-tagged wild-type (WT) TUFM and its truncation mutants. (B) 293T cells were co-transfected with pCAGGS-HA-2C and p3×Flag-CMV-TUFM WT or any one of its mutants for 36 h. The cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. Representative results are shown and similar results were obtained in three independent experiments. (C) Construction strategy of HA-tagged WT 2C and its truncation mutants. (D) 293T cells were co-transfected with p3×Flag-CMV-TUFM and pCAGGS-HA-2C WT or any one of its mutants for 36 h. The cells were analyzed by co-IP assays using anti-HA magnetic beads, followed by western blot analyses. Representative results are shown and similar results were obtained in three independent experiments. (E) Two additional truncation mutants aa256–321 and aa285–321 of 2C protein were constructed, transfected and analyzed as described in (D). (F) The predicted model of the TUFM/2C interaction complex was obtained using the Alphafold2 online server. The 2C (left) and TUFM (right) proteins are marked in pink and green, respectively. The amino acid residues predicted to be possible interaction sites are labeled with one-letter type name in stick representation. (G) 293T cells were co-transfected with pCAGGS-HA-2C and p3×Flag-CMV-TUFM WT or any one of the five mutants with single mutation sites of R200A, E193A, E196A, E212A, or E211A. At 36 hpt, the cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. Representative results are shown and similar results were obtained in three independent experiments. (H) 293T cells were co-transfected with p3×Flag-CMV-TUFM and pCAGGS-HA-2C WT or any one of the four mutants with a single mutation site of N240A, R286A, S293A, or R308A. At 36 hpt, the cells were analyzed by co-IP assays using anti-HA magnetic beads, followed by western blot analyses. Representative results are shown and similar results were obtained in three independent experiments. (I) Ddditional eleven single-site mutants of 2C protein each containing one of the potential key sites predicted in the other four TUFM/2C interaction complex models as shown in Fig. S9B-E. These mutants were transfected and analyzed as described in (H).

Ubiquitination of TUFM is essential for SVA-induced mitophagy

To elucidate the mechanism underlying how TUFM mediates SVA-induced mitophagy, co-IP assays were performed to screen the potential molecules that interact with TUFM, especially those related to autophagic signaling pathways. As shown in Figure 9A, endogenous TUFM coimmunoprecipitated with BECN1, ATG12–ATG5, ATG12, and SQSTM1 but not LC3-I, LC3-II, ATG5, ATG7, or ATG16L1 from both mock- and SVA-infected BHK-21 cells. Moreover, in in vitro affinity-isolation assays, the prokaryotically expressed Strep II-TUFM fusion protein (~94.63% purity) pulled down HA-tagged 2C and endogenous BECN1, but it did not pull down ATG12, SQSTM1, or VP2 (Figure 9B). More importantly, in in vitro affinity-isolation assays, the binding of BECN1 (~97.25% purity) with Strep II-TUFM showed a dose-dependent manner (Figure 9C), both of which were prokaryotically expressed. These results reveal that TUFM directly interacts with BECN1. Interestingly, we found that the endogenous TUFM protein can interact with endogenous SQSTM1 (Figure 9A), while the prokaryotically expressed TUFM cannot interact with endogenous SQSTM1 (Figure 9B). To further validate the interaction between TUFM and SQSTM1 proteins, we co-transfected eukaryotic expression plasmids p3×Flag-CMV-TUFM and pCAGGS-HA-SQSTM1 into 293T cells to prepare Flag-tagged TUFM and HA-tagged SQSTM1 fusion proteins, which were used for reciprocal co-IP assays. As shown in Figure 9D, E, the eukaryotically expressed fusion proteins TUFM and SQSTM1 pulled each other down. These results indicate that TUFM may need some modification in eukaryotic cells to enable its interaction with SQSTM1. As one of the important cargo receptors, SQSTM1 is usually recruited to damaged mitochondria by binding to the ubiquitin (Ub) chains on mitochondrial outer membrane proteins, thereby mediating the occurrence of mitophagy [20,31]. Accordingly, we speculated that TUFM needs to undergo ubiquitination modification before being recognized by SQSTM1 during SVA-induced mitophagy. To verify our speculation, we first detected whether TUFM undergoes ubiquitination modification in eukaryotic cells, and then analyzed the effect of SVA infection on TUFM ubiquitination. We co-transfected 293T cells with the plasmids p3×Flag-CMV-TUFM and pcDNA3.1-HA-Ub or the corresponding empty vectors. The resultant cell lysates were subjected to co-IP assays. As shown in Figure 9F, despite the amount of Flag-TUFM pulled down by anti-Flag mAb from the cells co-transfected with or without pcDNA3.1-HA-Ub being almost equivalent, co-transfection with pcDNA3.1-HA-Ub significantly increased the ubiquitination level of Flag-TUFM. This result indicates that TUFM does undergo ubiquitination modification in host cells. In addition, to analyze whether SVA infection affects the ubiquitination modification of TUFM, BHK-21 and PK-15 cells were co-transfected with p3×Flag-CMV-TUFM and pcDNA3.1-HA-Ub, followed by mock infection or infection with low or high doses of SVA. The results showed that SVA infection increased the ubiquitination levels of Flag-TUFM in BHK-21 cells in a dose-dependent manner, which was confirmed on PK-15 cells (Figure 9G, H). Next, we analyzed whether an increase in the TUFM ubiquitination level enhances its ability to interact with SQSTM1 using co-IP assays. As shown in Figure 9I, although the amounts of input proteins from BHK-21 cells co-transfected with p3×Flag-CMV-SQSTM1 and pEGFP-N2-TUFM were the same, more SQSTM1 was pulled down by the EGFP antibody from the BHK-21 cells expressing HA-tagged ubiquitin (HA-Ub) than those not expressing HA-Ub. This confirms that the SQSTM1 protein interacts with ubiquitinated TUFM protein. Finally, we carried out an endogenous co-IP assay using a rabbit anti-SQSTM1 polyclonal antibody to pull down its interacting proteins and confirmed that SQSTM1 interacts with LC3 (Figure 9J), as reported previously [20]. Taken together, our results demonstrate that the ubiquitination of TUFM is crucial for SVA-induced mitophagy, and that ubiquitinated TUFM mediates mitophagy through interaction with SQSTM1.

Figure 9.

SVA-induced mitophagy relies on TUFM ubiquitination modification. (A) BHK-21 cells were mock infected or infected with SVA (MOI = 0.1) for 24 h. The cells were analyzed by co-IP assays using rabbit anti-TUFM polyclonal antibody plus protein A/G magnetic beads, followed by western blot analyses. (B) BHK-21 cells were transfected with pCAGGS-HA-2C or pCAGGS-HA for 24 h, and then mock infected or infected with SVA (MOI = 0.1) for additional 24 h. The cells were harvested and lysed, and supernatants of the cell lysates were incubated with the prokaryotically expressed Strep II-TUFM protein bound to Streptactin beads 4FF. The formed complexes were analyzed by western blot analyses. (C) Ten micrograms of purified prokaryotically expressed Strep II-TUFM protein bound to Streptactin beads 4FF were incubated with 5, 10, or 20 μg of purified prokaryotically expressed His-BECN1 protein. The formed complexes were analyzed by western blot analyses. (D) 293T cells were co-transfected with p3×Flag-CMV-TUFM and pCAGGS-HA-SQSTM1 or their respective empty vectors for 36 h. The cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. (E) 293T cells were transfected and analyzed as described in (D) but using anti-HA magnetic beads. (F) 293T cells were co-transfected with p3×Flag-CMV-TUFM and pcDNA3.1-HA-Ub or their respective empty vectors for 36 h. The cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. (G) BHK-21 cells were co-transfected with p3×Flag-CMV-TUFM and pcDNA3.1-HA-Ub for 36 h, and then mock infected or infected with SVA (MOI = 0.1 and 1) for 24 h. The cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. (H) PK-15 cells were transfected, infected, and analyzed as described in (G). (I) BHK-21 cells were co-transfected with pcDNA3.1-HA-Ub, p3×Flag-CMV-SQSTM1, and pEGFP-N2-TUFM or their respective empty vectors for 36 h. The cells were analyzed by co-IP assays using anti-EGFP antibody plus protein A/G magnetic beads, followed by western blot analyses. (J) BHK-21 cells were mock infected or infected with SVA (MOI = 0.1) for 24 h. The cells were analyzed by co-IP assays using anti-SQSTM1 antibody plus protein A/G magnetic beads, followed by western blot analyses. All experiments related to this figure were conducted in three independent biological replicates with representative results are shown.

E3 ubiquitin ligase RNF185 catalyzes polyubiquitination of TUFM and 2C proteins by interacting with TUFM and 2C, respectively

On the basis of demonstrating that SVA-induced mitophagy relies on the ubiquitination of TUFM protein, we proceeded to identify the critical E3 ubiquitin ligase that mediates TUFM ubiquitination. The cellular proteins that potentially interact with endogenous TUFM were pulled down by co-IP assays and identified by MS. After analyzing the MS data, a total of six ubiquitination-related proteins (SMURF1, RPS27A, UBC, UBB, MARCHF5/MARCH5, and RNF185) that potentially interact with TUFM were identified and are listed in Figure 10A. To verify which of them mediates TUFM ubiquitination, each of the recombinant plasmids expressing MYC-tagged SMURF1, RPS27A, UBC, UBB, MARCHF5, or RNF185 was transfected into 293T cells together with the p3×Flag-CMV-TUFM plasmid. As shown in Figure 10B, only overexpression of RNF185, not the other five proteins, strongly decreased TUFM expression, suggesting that RNF185 might mediate the ubiquitination and degradation of TUFM. Subsequently, we explored whether RNF185 interacts with TUFM and what role RNF185 plays in SVA-induced mitophagy. We co-transfected 293T cells with three plasmids expressing MYC-tagged RNF185, Flag-tagged TUFM, or HA-tagged Ub, and the cell lysates produced were analyzed by co-IP using anti-Flag mAb. As shown in Figure 10C, MYC-RNF185 coimmunoprecipitated with Flag-TUFM, suggesting that RNF185 interacts with TUFM. Notably, the amount of ubiquitinated TUFM in the cells co-expressing MYC-RNF185, Flag-TUFM, and HA-Ub was significantly higher than that in the cells only co-expressing Flag-TUFM and HA-Ub, indicating that RNF185 mediates ubiquitination of TUFM. Next, we conducted co-IP assays to analyze whether RNF185 interacts with 2C and SQSTM1, and discovered that RNF185 and 2C could pull each other down (Figure 10D, E), but RNF185 couldn’t pull down SQSTM1 (Figure 10F). These data indicate that RNF185 interacts with 2C but not with SQSTM1. On this basis, we further explored whether RNF185 also catalyzes the polyubiquitination of 2C protein by co-transfecting 293T cells with plasmids pCMV-MYC-RNF185, pCAGGS-HA-2C, and pCMV-MYC-Ub or the corresponding empty vectors, and conducting co-IP assays. The results showed that RNF185 not only interacts with 2C, but also catalyzes the polyubiquitination of 2C protein (Figure 10G). Subsequently, we proceeded to investigate the role of RNF185 in SVA-induced mitophagy. To this end, the expression of endogenous RNF185 in BHK-21 cells was knocked down using a siRNF185-hamster prior to SVA infection. The results showed that knockdown of RNF185 significantly increased TUFM expression in both mock- and SVA-infected cells, which was accompanied by significantly decreased LC3-II and significantly increased SQSTM1, TIMM23, TOMM20 and their relative ratios to ACTB (Figure 10H,I). Of note, knockdown of RNF185 also decreased VP2 expression in SVA-infected cells (Figure 10H). Conversely, regardless of whether BHK-21 cells were infected with SVA or not, overexpression of MYC-tagged RNF185 significantly reduced the expression of TUFM, SQSTM1, TIMM23, and TOMM20, whereas it markedly increased LC3-II expression. Most importantly, the degree of change in the expression of these proteins in SVA-infected cells was much higher than that in the mock-infected cells (Figure 10J, K). These results suggest that RNF185 is involved in SVA-induced mitophagy as well. Additionally, virus titration experiments further confirmed that knockdown of RNF185 by siRNA significantly decreased SVA yields in BHK-21 cells from 12 hpi onward (Figure 10L), while overexpression of RNF185 significantly increased SVA yields from 24 hpi onward (Figure 10M), in comparison with the corresponding controls. In addition, the effect of interfering with RNF185 expression using siRNF185-pig or of RNF185 overexpression on mitophagy and SVA replication in PK-15 cells is consistent with the results obtained in BHK-21 cells (Fig. S10). Taken together, our results demonstrate that RNF185 catalyzes polyubiquitination of TUFM and 2C by interacting with both proteins, which is crucial for SVA-induced mitophagy.

Figure 10.

RNF185 catalyzes polyubiquitination of TUFM and 2C to initiate SVA-induced mitophagy. (A) BHK-21 cells were lysed, and supernatants of the cell lysates were analyzed by co-IP assays using rabbit anti-TUFM polyclonal antibody plus protein A/G magnetic beads, followed by mass spectrometry identification. The identified cellular proteins that potentially interact with TUFM and mediate its ubiquitination are listed. (B) 293T cells were mock transfected (MT) or co-transfected with p3×Flag-CMV-TUFM and pCMV-MYC-SMURF1, pCMV-MYC-RPS27A, pCMV-MYC-UBC, pCMV-MYC-UBB, pCMV-MYC-MARCHF5, pCMV-MYC-RNF185, or pCMV-MYC for 36 h. The cells were analyzed by western blot analyses. (C) 293T cells were co-transfected with pCMV-MYC-RNF185, p3×Flag-CMV-TUFM, and pcDNA3.1-HA-Ub or the corresponding empty vectors for 36 h. The cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. (D) 293T cells were co-transfected with pCMV-MYC-RNF185 and pCAGGS-HA-2C or the corresponding empty vectors for 36 h. The cells were analyzed by co-IP assays using anti-HA magnetic beads, followed by western blot analyses. (E) 293T cells were transfected and analyzed as described in (D) but using anti-MYC magnetic beads. (F) 293T cells were co-transfected with pCAGGS-HA-RNF185 and p3×Flag-CMV-SQSTM1 or the corresponding empty vectors for 36 h. The cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. (G) 293T cells were co-transfected with pCMV-MYC-RNF185, pCAGGS-HA-2C, and pCMV-MYC-Ub or the corresponding empty vectors for 36 h. The cells were analyzed by co-IP assays using anti-HA magnetic beads, followed by western blot analyses. (H) BHK-21 cells were mock transfected (MT) or transfected with 40 pmol/well of siRNF185-hamster or siNC for 24 h. Then the cells were mock infected or infected with SVA (MOI = 0.1) for an additional 24 h. The cells were analyzed by western blot analyses. (I) Densitometry of the target protein bands shown in (H) and calculation of their relative densitometric ratios to ACTB/β-actin. (J) BHK-21 cells were mock transfected (MT) or transfected with pCMV-MYC or 0.5, 1, or 2 μg/well of pCMV-MYC-RNF185 for 24 h, and then mock infected or infected with SVA (MOI = 0.1) for additional 24 h. The cells were analyzed by western blot analyses. (K) Densitometric analysis of the target proteins shown in (J). (L) BHK-21 cells were transfected and then infected as described in (H). The silencing of TUFM was confirmed by western blot prior to virus titration. At 12, 24, 36, and 48 hpi, viral samples were harvested, and viral titers were detected by TCID₅₀ assay on BHK-21 cells. (M) BHK-21 cells were transfected and then infected as described in (J). Viral titrations were performed and analyzed as described in (L). All experiments related to this figure were conducted in three independent biological replicates. Data are shown as mean ± SD of three independent experiments (two-way ANOVA; ns, no significance; *P < 0.05; **P < 0.01; ***P < 0.001).

The TM1 domain of RNF185 interacts with TUFM, polyubiquitinating TUFM via ubiquitin K27-linkage

The RNF185 protein, composed of 192 amino acids, mainly contains three functional domains: one N-terminal RING-finger and two C-terminal transmembrane (TM) domains TM1 and TM2 (Figure 11A) [32]. To identify the interacting domain of RNF185 with TUFM, we constructed HA-tagged WT RNF185 and its seven deletion mutants, ∆2–18, ∆19–38, ∆39–80, ∆81–132, ∆133–191, ∆155–191, and ∆172–191 (Figure 11A), which were then individually transfected into 293T cells along with Flag-tagged TUFM, followed by co-IP assays. As shown in Figure 11B, except for ∆133–191, the other six mutants and the WT RNF185 all pulled down TUFM. These data indicate that the TM1 domain corresponding to the amino acids 133–155 of RNF185 is responsible for the interaction with TUFM. Next, we also made an attempt to identify the interacting domain of TUFM with RNF185 using the constructed Flag-tagged WT TUFM and its nine deletion mutants, which were previously described in Figure 8A. When HA-tagged RNF185 was co-expressed with WT TUFM or each of the nine mutants in 293T cells, WT TUFM and its mutants ∆I, ∆II, ∆III, aa1–252, aa56–252, and aa275–452 pulled down RNF185, whereas mutants aa1–55, aa275–344, and aa345–452 did not (Figure 11C). On the basis of these results, we concluded that there may be two key regions in the TUFM protein, one located within aa56–252 and the other located across the 344^th/345^th amino acid, either of which can independently mediate its interaction with RNF185. To verify this discovery, we further analyzed the TUFM-RNF185 interaction complex by means of homology modeling and subsequent rigid protein – protein docking using I-TASSER (https://zhanggroup.org/I-TASSER/) and GRAMM-X online software (http://gramm.compbio.ku.edu/). The docking results showed that there are seven pairs of amino acid residues forming hydrogen bonds at the interaction interface between RNF185 and TUFM, which are crucial for mediating the interaction between RNF185 and TUFM. These residue pairs include PHE133-GLU149, MET135-GLU149, ALA141-GLY155, ILE146-ILE119, THR149-ASP201, GLY145-LYS204, and ILE146-ARG369 (Figure 11D). Among them, the PHE133, MET135, ALA141, ILE146, THR149, GLY145, and ILE146 amino acid residues of RNF185 are all located in its TM1 domain (aa133–155), while the GLU149, GLY155, ILE119, ASP201, LYS204, and ARG369 amino acid residues of TUFM are all located in the two interaction regions identified. Specifically, GLU149, GLY155, ILE119, ASP201, and LYS204 are all located in the aa56–252 region of TUFM, whereas ARG369 is located in the region spanning 344^th/345^th amino acids. Obviously, the protein docking results further confirmed the reliability of the identification of the interaction region between RNF185 and TUFM.

Figure 11.

RNF185 interacts with TUFM via its TM1 domain and polyubiquitinates TUFM via ubiquitin K27-linkage. (A) Schematic diagram of domain structure of RNF185 and construction strategy of HA-tagged wild-type (WT) RNF185 and its truncation mutants. (B) 293T cells were co-transfected with p3×Flag-CMV-TUFM and pCAGGS-HA-RNF185 WT or any one of its mutants for 36 h. The cells were analyzed by co-IP assays using anti-HA magnetic beads, followed by western blot analyses. (C) 293T cells were co-transfected with pCAGGS-HA-RNF185 and p3×Flag-CMV-TUFM WT or any one of its mutants for 36 h. The cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. (D) Surface diagram of the molecular docking model and their interfacing residues between the full-length TUFM protein and the TM1 domain of RNF185 protein (TUFM, blue; RNF185, green; hydrogen bond interaction, dotted line). (E) 293T cells were co-transfected with p3×Flag-CMV-TUFM, pCAGGS-HA-RNF185 and pCMV-MYC-Ub WT or any one of the seven mutants with a single-site mutation of K6R, K11R, K27R, K29R, K33R, K48R, or K63R. At 36 hpt, the cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. (F) 293T cells were co-transfected with pCMV-MYC-RNF185, pcDNA3.1-HA-Ub, and p3×Flag-CMV-TUFM WT or any one of its mutants for 36 h. The cells were analyzed by co-IP assays using anti-Flag magnetic beads, followed by western blot analyses. All experiments related to this figure were conducted in three independent biological replicates with representative results are shown.

Ubiquitin has seven key lysine residues (K6, K11, K27, K29, K33, K48, and K63), all of which can be ubiquitinated, but distinct linkage types of ubiquitin chains guide the modified proteins toward different cellular fates [33]. To clarify how TUFM is polyubiquitinated, we constructed seven MYC-tagged Ub mutants, K6R, K11R, K27R, K29R, K33R, K48R, and K63R, by mutating the corresponding seven lysines in WT Ub into arginine. Flag-TUFM were co-expressed with WT Ub or each of the Ub mutants in 293T cells in the presence of HA-tagged RNF185, and the polyubiquitination of TUFM was analyzed by co-IP assays. As shown in Figure 11E, only K27R out of the seven Ub mutants resulted in a significant decrease in the polyubiquitination level of TUFM protein, indicating that RNF185 polyubiquitinates TUFM via K27-linked Ub chain.

To further clarify which region of TUFM is involved in RNF185-mediated polyubiquitination, the constructed Flag-tagged WT TUFM and its nine truncation mutants (Figure 8A) were individually co-expressed with MYC-RNF185 in 293T cells in the presence of HA-Ub. As shown in Figure 11F, WT TUFM and its mutants—∆I, ∆II, ∆III, aa1–252, aa56–252, and aa275–452—clearly underwent ubiquitination, whereas mutants aa1–55, aa275–344, and aa345–452 did not. These data are completely consistent with the identified regions in TUFM responsible for interaction with RNF185 (Figure 11C). Specifically, two regions might exist in TUFM that can undergo RNF185-catalyzed ubiquitination: one located within aa56–252 and the other located across the 344^th/345^th amino acid. Notably, although we made an attempt to identify the lysine residues in TUFM that might act as RNF185-mediated ubiquitination sites by introducing single-lysine mutations, such sites were not ultimately identified. Taken together, our data demonstrate that RNF185 interacts with TUFM via its TM1 domain and catalyzes the K27-linked polyubiquitination of TUFM.

Discussion

Although our knowledge of the etiology and epidemiology of SVA has greatly expanded since its frequent outbreaks, we still lack a comprehensive understanding of its pathogenesis, especially its molecular pathogenesis. To explore potential therapeutic targets for SVA, the underlying mechanisms involved in the interactions between SVA and host cells need to be investigated in more detail. Conducting such research not only helps us discover unexpected clues to combat SVA infection, but also provides a reference for research on other animal pathogens. In the present study, four types of cells, BHK-21, PK-15, IBRS-2, and ST, were used as in vitro infection models to investigate the crosstalk between SVA infection and mitophagy. Our results clearly indicated that SVA infection could trigger complete mitophagy in host cells, which is beneficial for the replication of SVA. Mechanistically, after SVA invaded the host cells, the 2C protein of SVA was anchored to the mitochondria through direct interaction with the mitochondrial TUFM protein. Subsequently, the E3 ubiquitin ligase RNF185 catalyzed the polyubiquitination of TUFM and 2C proteins by interacting with TUFM and 2C, respectively, where the TUFM protein underwent K27-linked ubiquitination modification. Then the ubiquitinated TUFM was recognized and bound by SQSTM1, which in turn interacted with LC3, thereby sequestrating 2C-anchored mitochondria into the phagophore to form mitophagosomes, which ultimately fused with lysosomes to achieve complete mitophagy. A proposed model for the molecular mechanism by which SVA induces mitophagy to promote self-replication is shown in Figure 12.

Figure 12.

A proposed model for mitophagy induced by the SVA 2C protein. After SVA invades the host cells, the 2C protein is anchored to the mitochondria through direct interaction with TUFM. Then, the E3 ubiquitin ligase RNF185 catalyzes the polyubiquitination of TUFM and 2C proteins by interacting with TUFM and 2C, respectively, where TUFM undergoes K27-linked ubiquitination modification to enable it to interact with SQSTM1, which in turn interacts with LC3 to recruit the 2C-anchored mitochondria to the phagophore, forming mitophagosomes that eventually fuse with lysosomes to achieve complete mitophagy.

Previous studies have shown that SVA can activate autophagy during the infection of host cells [13,14], and SVA uses different viral proteins to manipulate autophagy at different stages of infection, aiming to benefit the replication of SVA itself [15,16]. However, it was not clear whether mitophagy is involved in SVA infection before this study was launched. To our knowledge, our current study is the first to systematically demonstrate the activation of mitophagy in host cells upon infection with SVA. We also demonstrated that SVA-induced mitophagy is beneficial for SVA replication in host cells, although the cell lines we used come from different species. Our data are consistent with those of a previous study demonstrating that SVA infection activated autophagy in BHK-21 and PK-15 cells to facilitate virus replication [13]. However, another study showed that the induction of autophagy promoted SVA replication in pig-derived cells SK6 and ST, but inhibited SVA replication in human-derived cells H1299 and 293T [16]. Although the root causes of this discrepancy are currently unknown, the impact of autophagy on SVA replication may be cell-species specific [16]. Notwithstanding, our results indicate that the role of mitophagy in SVA replication appears to be unrelated to the species specificity of the host cells. Undoubtedly, further studies are needed to conclusively address this issue by using more cell types of different species that are permissive to SVA infection to detect the impact of mitophagy on SVA replication.

TUFM is an important multifunctional mitochondrial protein with RNA binding and GTPase activity [30,34], and it is mainly related, inter alia, to mitochondrial protein translation elongation and biosynthesis, oxidative phosphorylation, immune modulation, and protein quality control [21,30]. More importantly, TUFM is also involved in the struggle between host cells and certain invading viruses [18]. For example, the matrix protein of human parainfluenza virus functioned as a tether to simultaneously interact with TUFM and LC3 to activate mitophagy and thereby attenuate the type I interferon response in virus-infected cells [35]. A similar phenomenon was also discovered in influenza A virus-infected cells, in which the viral PB1-F2 protein acted as a bridge to concurrently interact with TUFM and LC3 to trigger mitophagy, which caused the degradation of MAVS, thereby attenuating innate immunity [21]. However, our current study suggested that, although the SVA 2C protein also induces mitophagy through its interaction with TUFM, the molecular details involved are different from those of the influenza virus PB1-F2 protein and the parainfluenza virus matrix protein because the 2C protein did not directly interact with LC3; instead, it used a complex formed by the interaction between ubiquitinated TUFM and SQSTM1 as a linker to bridge the phagophore and the 2C-anchored mitochondria. Obviously, the way 2C protein mediated mitophagy seems more complex. Moreover, our study further found that TUFM interacts directly with BECN1 but indirectly with the ATG12–ATG5 conjugate, which might mediate the extension of the phagophore to engulf mitochondria and the consequent formation of mitophagosomes. Our data are consistent with those of a previous study, which demonstrated that TUFM potentiated vesicular stomatitis virus-induced autophagy and attenuated type I interferon activation through simultaneous interactions with ATG12–ATG5 and NLRX1 [30]. In addition, we used ectopic expression and RNA interference approaches to investigate the role of TUFM in SVA-induced mitophagy and found that TUFM functions as a positive regulator in SVA-induced mitophagy. Frankly, we made an attempt to further verify the impact of TUFM on SVA-induced mitophagy by constructing TUFM-knockout BHK-21 and PK-15 cells using the CRISPR/Cas9 system, as described previously [36]. Unfortunately, despite our best efforts in screening over 100 monoclonal cells, these cells were confirmed to be all heterozygous (data not shown). However, another study reported the successful construction of TUFM-knockout human lung adenocarcinoma A549 cells [21]. Although the reasons for this inconsistency remain unclear, we speculated that TUFM can be knocked out without killing cancer cells, while the same operation is fatal for normal cells like BHK-21 and PK-15.

Another important finding of our current study is that the ubiquitination of TUFM is crucial for SVA-induced mitophagy and enables TUFM to be recognized by SQSTM1. Moreover, we further identified that the E3 ubiquitin ligase RNF185 catalyzed the ubiquitination of TUFM via ubiquitin K27-linkage, demonstrating for the first time that TUFM is an important substrate for RNF185. Intriguingly, RNF185 resides in both mitochondria and endoplasmic reticulum (ER). The mitochondria-resident RNF185 was found to regulate mitophagy by polyubiquitinating BNIP1 through K63-linked ubiquitin chains [37], while the ER-resident RNF185 was demonstrated to polyubiquitinate the misfolded proteins that had accumulated in ER for proteasome-mediated degradation [32]. Our results revealed that RNF185 regulates SVA-induced mitophagy through the interaction between its TM1 domain and TUFM (Figure 11A, B), and catalyzes the polyubiquitination of TUFM and 2C proteins Figure 10C,G). Similarly, RNF185 was shown to be able to polyubiquitinate glycoprotein GP_1,2 of Ebolavirus on lysine 673 through ubiquitin K27-linkage, ultimately leading to the degradation of the GP_1,2 protein by reticulophagy [32]. RNF185 was further demonstrated to regulate Ebolavirus-induced reticulophagy through interactions between its TM1 and TM2 domains and GP_1,2 [32]. In addition, we discovered for the first time that the SVA VP2 protein is highly associated with the mitochondrial fraction due to its interaction with TUFM (Figures 2E,G, and 7C–E). However, as our study mainly focused on elucidating the molecular mechanism of 2C-mediated mitophagy, we did not explore the biological significance of VP2 protein targeting mitochondria. Nevertheless, an increasing number of viruses having been reported to manipulate cell death, metabolic status, and evade host cell immune responses by specifically targeting mitochondria with their own encoded proteins, thereby facilitating the survival of the virus [38].

Existing studies have shown that the main physiological functions of mitophagy include controlling the quality and quantity of mitochondria [39], maintaining cellular homeostasis [20], regulating innate and adaptive immune responses [20,40], modulating cytokine production [40], managing autoimmune diseases [40], and others [31]. Hence, steady-state mitophagy can enhance the defense ability of host cells; however, excessive mitophagy or defects in mitophagy usually lead to cellular dysfunction, which has been found to be associated with various diseases, such as cancer, lung diseases, neurodegenerative diseases [31,40]. In addition to the physiological and pathophysiological functions of mitophagy mentioned above, accumulating evidence suggests that mitophagy is also involved in the infection processes of various pathogens [20,41]. Thus far, a few viruses have been demonstrated to have evolved diverse strategies to subvert and benefit from mitophagy [18], such as the aforementioned influenza A virus, human parainfluenza virus, and SARS-CoV-2. A common feature involving the induction of mitophagy by these viruses is that the virus can independently induce mitophagy by means of an individual protein. Similarly, our results indicate that the 2C protein of SVA can independently induce mitophagy, although its molecular mechanism is different from that of the aforementioned viruses. Specifically, the 2C protein does not directly act as a mitophagy receptor like other viral proteins, but interacts directly with TUFM and activates mitophagy under the combined action of RNF185, SQSTM1, and LC3. In contrast, other viruses mainly use a certain protein to simultaneously interact with mitochondrial proteins and LC3, bridging mitochondria and phagophores and thereby activating mitophagy [18].

The 2C protein of picornaviruses is a relatively conservative and multifunctional non-structural protein with helicase activity that plays a complex but yet poorly understood role in viral life cycle [4]. Wen and colleagues showed that the 2C protein of SVA inhibited the production of type I interferon by degrading RIG-I and MAVS through the caspase signaling pathway [42]. It is evident that their results are consistent with our study that 2C protein-induced mitophagy can lead to mitochondrial clearance. Given that mitochondria are the signal transduction platform for innate immunity [17], the clearance of mitochondria along with MAVS will inevitably damage the host’s antiviral ability, thereby facilitating SVA survival. Frankly, research on SVA antagonizing innate immunity by degradating MAVS protein through mitophagy is currently underway in our laboratory. Furthermore, the SVA 2C protein was demonstrated to induce the activation of CASP3 (caspase 3) and CASP9, thereby activating apoptosis through a mitochondrion-mediated intrinsic pathway [11]. Amino acids 1–102 at the N terminus of the 2C protein are crucial for 2C-induced apoptosis [11]. Notably, amino acids 231–321 of the 2C protein were demonstrated to play a vital role in SVA-induced mitophagy in our study. These findings indicate that different regions of the 2C protein may play different biological functions. A recent study demonstrated that the SVA 2C protein was able to degrade the CGAS (cyclic GMP-AMP synthase) sensor, thereby antagonizing the host innate antiviral response [23]. Both the N-terminal region and the RNA helicase domains of 2C were found to be responsible for 2C-induced CGAS degradation and autophagy activation [23]. This study further demonstrated that overexpression of SVA 2C significantly decreased the expression of the CGAS protein while enhancing SVA replication [23]. This finding is completely consistent with our research results, which showed that 2C-medicated mitophagy functions to benefit SVA replication in host cells. Moreover, our finding of a direct interaction between SVA 2C and TUFM also confirms the results of another study showing that the 2C protein is located in mitochondria in SVA-infected cells [11]. Undoubtedly, our results further enrich our understanding of the function of the SVA 2C protein. On the basis of the above evidence, the 2C protein can be considered an important virulence factor of SVA. In addition to the 2C protein, our results revealed that the VP2, VP3, and 3A proteins of SVA also exhibited phenotypes that induce mitophagy. Interestingly, although our data clearly indicated that the SVA 2C protein and even VP2, VP3, and 3A proteins could independently induce mitophagy (Figures 6 and 7), the UV-inactivated SVA couldn’t (Figure 3). To explain this phenomenon, we demonstrated through indirect immunofluorescence assay that the inactivated SVA lost the ability to enter the host cells (Figure 3A). This means that although the inactivated SVA virions still contained viral proteins, they lost their ability to induce mitophagy due to their inability to enter the cells. Therefore, only 2C protein transfected into host cells could induce mitophagy, while the inactivated SVA couldn’t induce mitophagy. Furthermore, it is worth noting that since VP2, VP3, and 3A proteins did not directly interact with TUFM, we speculate that their molecular mechanism of inducing mitophagy should be different from that of 2C. Obviously, further studies are needed to elucidate the respective molecular mechanisms for inducing mitophagy.

It is noteworthy that although the mitochondrial uncoupler CCCP is not a specific inducer of mitophagy, it is still currently recognized and widely used as an inducer of mitophagy, which activates mitophagy by causing acute dissipation of mitochondrial membrane potential [24–26]. A recent study demonstrated that low concentrations of CCCP (<1 μM) mainly cause instant complete mitochondrial depolarization, and only high concentrations of CCCP (usually >10 μM for several hours) can induce mitophagy, especially when the cell culture medium contains high concentrations of fetal bovine serum or bovine serum albumin [43]. Based on these findings and combined with screening of safe concentrations, we ultimately chose to use 10 μM and 15 μM of CCCP to treat BHK-21 and PK-15 cells, respectively, to ensure successful activation of mitophagy. Our subsequent results showed that under the treatment of CCCP at the working concentrations we used, mitophagy in BHK-21 and PK-15 cells was successfully activated, as manifested by the formation of mitophagosome-like vesicles (Figure 1A), and significant LC3 conversion accompanied by degradation of SQSTM1, TIMM23 and TOMM20 (Figure 4). As for the phenotype of increased SVA production caused by CCCP treatment, in addition to the role played by mitophagy, it is not clear in our current study whether other stress responses that might be related to CCCP treatment were also involved. For example, a previous study showed that CCCP treatment could inhibit the activation of STING1 and its downstream signaling molecules TBK1 and IRF3 through dynamin-related protein 1-mediated mitochondrial fission, thereby inhibiting the production of type I interferon [44]. Furthermore, we also observed that the degree of increase in SVA titer caused by CCCP treatment is apparently higher than that of VP2 protein expression (Figure 4). Actually, this is not unanticipated phenomenon. Similar to other picornaviruses, SVA is a cell lytic virus that can efficiently replicate in susceptible host cells and cause cell disintegration in a relatively short infection time. As SVA infection progressed, the degree of cell disintegration gradually increased, leading to the release of newly synthesized soluble proteins, including VP2 protein, into the cell supernatants. This ultimately caused a relatively lower VP2 protein content in the total cellular protein samples used for immunoblotting detection than its true value. In contrast, the viral samples used for SVA titration were obtained by freezing and thawing the SVA-infected cells together with the supernatants three times. Therefore, cell disintegration did not affect the measurement of SVA titers. Interestingly, a similar phenomenon was also observed in our previous study on autophagy induced by another important picornavirus, the encephalomyocarditis virus [27]. In addition, we analyzed the effect of CCCP treatment on both intracellular and extracellular SVA titers, and discovered that the effect of CCCP treatment on intracellular virus production appeared to be significantly higher than that of extracellular viruses (Figure 4D,H). However, our current study is not yet clear whether the mechanism by which coxsackievirus B escapes the infected cells through mitophagosomes sequestrating extracellular microvesicles that contain infectious virions also exists in SVA-infected cells [45], which deserves further research.

In conclusion, our study demonstrated for the first time that SVA infection can induce mitophagy in host cells, which in turn promotes SVA replication. Mechanistically, the 2C protein interacts directly with TUFM, which further undergoes K27-linked ubiquitination under catalysis by RNF185. The ubiquitinated TUFM is recognized by SQSTM1, which in turn interacts with LC3, ultimately leading to mitochondria being enveloped in the phagophore, thus triggering mitophagy. Our results clearly explain how mitochondria were tagged for autophagosomal recognition in SVA-infected host cells at the molecular level. Undoubtedly, clarification of the reciprocal regulation mechanisms linking SVA and mitophagy will be important to improve our understanding of SVA pathogenesis, which, hopefully, will lead to the development of more effective therapeutic strategies against SVA infection.

Materials and methods

Virus and cells

The SVA SDta/2018 strain (GenBank accession no. MN433300.1) was isolated from a piglet suffering from vesicular disease in our laboratory [7,46]. BHK-21 (ATCC, CCL-10), PK-15 (ATCC, CCL-33), IBRS-2, ST (ATCC, CRL-1746) and 293T (ATCC, CRL-11268) cells were all cultured with Dulbecco’s modified Eagle’s medium (DMEM; Gibco 12,491,015) supplemented with 10% fetal bovine serum (Gibco 16,140,071) and 1% penicillin-streptomycin (Beyotime, C0222) at 37°C with 5% CO₂ in a humidified incubator.

Antibodies and reagents

The VP2-specific mAb 2F5 used for tracking SVA infection was prepared in our laboratory [7]. Rabbit anti-LC3 (ab192890), anti-RNF185 (ab181999), and anti-SQSTM1/p62 (ab109012) mAbs were purchased from Abcam. Rabbit anti-TIMM23 (11123–1-AP), anti-TOMM20 (11802–1-AP), anti-BECN1 (11306–1-AP), anti-ATG5 (10181–2-AP), anti-ATG12 (11122–1-AP), anti-ATG7 (10088–2-AP), and anti-MYC (16286–1-AP) polyclonal antibodies and mouse anti-ACTB/β-actin (66009–1-Ig) mAb were purchased from Proteintech Group, Inc. Rabbit anti-TUFM (A6423) and anti-ATG16L1 (A1871) polyclonal antibodies and mouse anti-Strep II-Tag mAb (AE066) were purchased from ABclonal, Inc. Mouse anti-HA (M180), anti-Flag (M185), and anti-GFP (D153–3) mAbs were purchased from MBL Beijing Biotech Co., Ltd. Horseradish peroxidase (HRP)-conjugated goat anti-rabbit (ZB-2301) and anti-mouse (ZB-2305) IgG secondary antibodies were purchased from ZSGB-BIO. 4’,6-diamidino-2’-phenylindole (DAPI), MitoTracker Red CMXRos (M7512), Alexa Fluor Plus 405-conjugated goat anti-mouse IgG (H+L) secondary antibody (A48255), Alexa Fluor 488-conjugated goat anti-mouse IgG (H+L) secondary antibody (A-11001), and Alexa Fluor 568-conjugated goat anti-mouse IgG (H+L) F(ab’)2 fragment (A-11004) were purchased from Thermo Fisher Scientific. CCCP (HY-100941), Mdivi-1 (HY-15886), Baf A1 (HY-100558), CQ (HY-17589A), protein A/G magnetic beads (HY-K0202A), anti-HA magnetic beads (HY-K0201A), and anti-Flag magnetic beads (HY-K0207) were all purchased from MedChem Express.

Plasmids and mutagenesis

Plasmids of pEGFP-N2-LC3, pcDNA3.1-HA-Ub, p3×Flag-CMV-SQSTM1, pCAGGS-HA-SQSTM1, pEGFP-N2, pCAGGS-HA, pCMV-MYC, and p3×Flag-CMV-10 were preserved in our laboratory. Eukaryotic plasmids pmCherry-GFP-LC3 and pmRFP-GFP-Mito were kindly provided by the College of Veterinary Medicine of Huazhong Agricultural University, Wuhan, China [21]. Recombinant plasmids pCAGGS-HA-L, pCAGGS-HA-VP4, pCAGGS-HA-VP2, pCAGGS-HA-VP3, pCAGGS-HA-VP1, pCAGGS-HA-2AB, pCAGGS-HA-2B, pCAGGS-HA-2C, pCAGGS-HA-3A, pCAGGS-HA-3AB, pCAGGS-HA-3C, and pCAGGS-HA-3D were constructed by RT-PCR amplification of the corresponding fragments of the SVA SDta/2018 strain using the primers listed in Table S1, followed by ligating each fragment into the indicated restriction enzyme-linearized pCAGGS-HA vector using Hi-T4 DNA ligase (NEB, M2622S). Recombinant plasmids p3×Flag-CMV-TUFM, pEGFP-N2-TUFM, pET-28a-Strep II-TUFM, pCMV-MYC-SMURF1, pCMV-MYC-RPS27A, pCMV-MYC-UBC, pCMV-MYC-UBB, pCMV-MYC-MARCHF5, pCMV-MYC-RNF185, and pCAGGS-HA-RNF185 were constructed by RT-PCR amplification of each gene encoding TUFM, SMURF1, RPS27A, UBC, UBB, MARCHF5, and RNF185 proteins using total RNA from BHK-21 cells as the template together with the primers listed in Table S2, followed by ligating each fragment into the corresponding restriction enzyme-linearized vectors p3×Flag-CMV-10, pET-28a, pEGFP-N2, pCMV-MYC, or pCAGGS-HA using Hi-T4 DNA ligase. Furthermore, nine TUFM deletion mutants (∆I, ∆II, ∆III, aa1–252, aa1–55, aa56–252, aa275–452, aa275–344, and aa345–452), eight 2C truncated mutants (aa1–175, aa1–69, aa70–175, aa176–321, aa212–321, aa231–321, aa256–321, and aa285–321), and seven RNF185 deletion mutants (∆2–18, ∆19–38, ∆39–80, ∆81–132, ∆133–191, ∆155–191, and ∆172–191) were constructed using p3×Flag-CMV-TUFM (Table S2), pCAGGS-HA-2C (Table S1), and pCAGGS-HA-RNF185 (Table S2) as the backbone plasmids, respectively, together with the respective primer pairs listed in Table S3. The resulting PCR fragments were individually cloned into the corresponding Hind III/BamH I-linearized p3×Flag-CMV-10 or EcoR I/Kpn I-linearized pCAGGS-HA via homologous recombination using the ClonExpress II One Step Cloning Kit (Vazyme, C112). A series of recombinant plasmids expressing TUFM, 2C, and Ub proteins, each with a single-site mutation, were constructed using p3×Flag-CMV-TUFM, pCAGGS-HA-2C, and pCMV-MYC-Ub as the backbone plasmids, respectively, together with the respective mutagenic primer pairs listed in Table S4. Recombinant plasmids pET-28a-Strep II, pET-30a-GST-2C, pET-30a-GST, pET-30a-His-BECN1, and pET-30a-His were constructed using the synthesized target fragments from Beijing Protein Innovation, Co., Ltd. All the constructed plasmids were verified by Sanger sequencing to ensure their accuracy.

Cell transfection and drug treatment

Cells grown to 50–60% confluence were transfected with the specified plasmids or small interfering RNAs (siRNAs) in the relevant figures using Lipofectamine 2000 reagent (Invitrogen 11,668,019) according to the manufacturer’s instructions. For mitophagy induction, BHK-21 and PK-15 cells were pretreated with 10 and 15 μM of CCCP for 4 h, respectively, prior to SVA infection. For mitophagy inhibition, both BHK-21 and PK-15 cells were pretreated with 10 μM of Mdivi-1 for 4 h prior to SVA infection. For autophagic degradation suppression, both mock- and SVA-infected BHK-21 and PK-15 cells were cultured in fresh medium supplemented with 100 μM of CQ or 100 nM of Baf A1 for 24 h.

Western blot and co-IP assay

Cells grown in six-well cell plates were harvested and lysed with 200 μL/well of Western & IP cell lysis buffer (Beyotime, P0013) containing 1 mg/mL protease inhibitor cocktail (Sigma-Aldrich, P8340) for 30 min on ice. After centrifugation for 30 min at 12,000 g at 4°C, the protein concentration in the supernatant of each sample was determined using a BCA protein assay kit (Thermo Scientific 23,225). Twenty micrograms of protein from each sample were separated by SDS-PAGE gels and electro-transferred onto polyvinylidene fluoride membranes (Millipore, ISEQ00010). After blocking with 5% nonfat milk in phosphate-buffered saline (PBS; Solarbio, P1020) containing 0.1% Tween 20 (Sigma-Aldrich, P1379) for 2 h at room temperature, the membranes were probed with the relevant primary antibodies overnight at 4°C. After thoroughly washing, the membranes were incubated with the corresponding HRP-conjugated secondary antibodies for 1 h at 37°C. Following a final washing step, the target protein signals on the membranes were developed using an Enlight-plus enhanced chemiluminescent reagent kit (Engreen 17,500), and images were taken with a GelDoc Go Image Lab Touch system (Bio-Rad). Densitometry of the target protein band intensities was conducted using ImageJ software (National Institutes of Health, Bethesda, MD, USA). For the co-IP assays, the precleared supernatants of cell lysates were incubated with anti-HA or anti-Flag magnetic beads, or with anti-TUFM, anti-GFP, or anti-SQSTM1 antibody plus protein A/G magnetic beads overnight at 4°C with slow rotation, depending on the specific situations. The formed immunoprecipitation pellets were washed five times with Tris-buffered saline (Solarbio, T1150) containing 0.1% Tween 20 with the aid of a DynaMag-2 magnetic separation device (Invitrogen), and then subjected to western blot analyses.

Mitochondria isolation

BHK-21 and PK-15 cells grown in 9-cm² cell dishes were mock infected or infected with SVA at a multiplicity of infection (MOI) of 0.1 for 24 h. After washing three times with precooled PBS, the cells were detached with a cell scraper and centrifuged at 500 g for 5 min. The cell pellets were resuspended in 800 μL mitochondria isolation buffer (Beyotime, C3601) supplemented with protease inhibitor cocktail and then homogenized using a Dounce tissue grinder (Sigma-Aldrich, D9063) for 20 min at 4°C. After centrifugating at 500 g for 10 min at 4°C, the supernatants were transferred to a new tube and centrifuged at 5,000 g for 10 min to obtain the cytosolic fraction. The precipitates were resuspended in 1 mL mitochondria extraction buffer and separated by 20%–40% double-gradient sucrose centrifugation at 95,000 g for 30 min at 4°C. The white bands between the 20% and 40% sucrose gradients were the purified mitochondrial fractions.

Confocal immunofluorescence microscopy

When BHK-21 and PK-15 cells grown on coverslips in 12-well plates approached 60% confluence, they were transfected with 1 μg of plasmid/well using Lipofectamine 2000 reagent for 12 h and then infected with SVA at an MOI of 1. At 24 hpi, the cells were stained with MitoTracker Red CMXRos for 30 min at 37°C and then fixed, permeabilized, and incubated with the appropriate primary antibodies, followed by incubation with the corresponding fluorescent secondary antibodies. Images were taken using a Nikon A1 confocal microscope (Nikon Instruments Inc., Tokyo, Japan) and then processed with the NIS-Elements Viewer. The fluorescence intensity profile of images was quantified using ImageJ software as described previously [47].

Transmission electron microscopy

BHK-21 and PK-15 cells were treated with 10 and 15 μM CCCP for 12 h, respectively, or mock infected or infected with SVA at an MOI of 1 for 12 h. After washing three times with precooled PBS, the cells were scraped and centrifuged at 800 g for 10 min. The cell pellets were fixed in 2.5% glutaraldehyde overnight at 4°C, postfixed in 1% osmium tetroxide for 1 h at room temperature, and thoroughly rinsed with distilled H₂O prior to staining with 1% aqueous uranyl acetate for 1 h. Subsequently, the cells were dehydrated in a graded ethanol series and embedded in Eponate 12 resin (Ted Pella 18,010). Ultrathin sections (70–80 nm in thickness) were cut with an ultramicrotome. After staining with uranyl acetate and lead citrate, the sections were observed with an H-7500 transmission electron microscope (Hitachi, Japan).

RNA interference

Two sets of siRNAs were designed for specific silencing of TUFM and RNF185 using the BLOCK-iT RNAi Designer tool (https://rnaidesigner.thermofisher.com/) and synthesized by GenePharma Co., Ltd (Suzhou, China). The designed siRNAs were as follows: siTUFM-hamster (Gene ID: 101822521): (sense, 5’-GGACUUAAAGCUCAGCCUAGU-3’;v anti-sense, 5’-UAGGCUGAGCUUUAAGUCCUC-3’), siTUFM-pig (Gene ID: 100516488): (sense, 5’-GCAGAUUACGUUAAGAAUAUG-3’; anti-sense, 5’-UAUUCUUAACGUAAUCUGCGU-3’), siRNF185-hamster (Gene ID: 101835149): (sense, 5’-UCCAAAAGACAUCUGAAAGCC-3’; anti-sense, 5’- CUUUCAGAUGUCUUUUGGAAU-3’), siRNF185-pig (Gene ID: 100157569) (sense, 5’-GGUUGGAGACCAGACCUAACA-3’; anti-sense, 5’-UUAGGUCUGGUCUCCAACCAC-3’), and the scrambled siRNA (siNC) (sense, 5”-UUCUCCGAACGUGUCACGUTT-3‘; anti-sense 5’-ACGUGACACGUUCGGAGAATT-3”). BHK-21 or PK-15 cells grown to ~50% confluence in six-well cell plates were transfected with the designed species-specific siRNAs (40 pmol/well) using Lipofectamine 2000 reagent. At 36 h post-transfection (hpt), the transfected cells were mock infected or infected with SVA at an MOI of 0.1 or 0.01. At the specified time points, the cells and progeny viruses were harvested for western blot analysis and viral yield titration, respectively.

Strep II or GST affinity-isolation assays

Recombinant prokaryotic plasmids pET-28a-Strep II-TUFM, pET-30a-GST-2C, pET-30a-His-BECN1, and pET-30a-GST were individually transformed into E. coli BL21 competent cells. Each transformant was grown in 100 mL of Luria – Bertani medium containing 100 μg/mL kanamycin at 37°C with shaking at 180 rpm. When the optical density at 600 nm for the culture reached 0.6, a final concentration of 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG; Sigma-Aldrich, I6758) was added to induce protein expression. Cultures expressing Strep II-TUFM, GST-2C, and His-BECN1 were further cultivated for 12 h, 24 h, and 4 h at 25°C, 16°C, and 37°C, respectively. Bacteria were harvested by centrifugation at 6,000 g for 5 min, and pellets were resuspended in PBS containing 1 mM PMSF (Merck 52,332). After ultrasonication and lysis with 1% Triton X-100 (Merck, T8787) for 30 min on ice, the cell lysates were centrifugated at 12,000 g for 30 min at 4°C. The supernatants containing Strep II-TUFM, GST-2C, and His-BECN1 proteins were bound to Streptactin Beads 4FF (Smart-Lifesciences, SA053005), Glutathione-Sepharose 4B bead column (GE HealthCare, 17-0756-05), and Ni-NTA (Qiagen 30,210), respectively, and subsequently purified according to the manufacturers’ instructions. For Strep II affinity-isolation assay, 10 μg of the purified Strep II-TUFM protein per sample were bound to 50 μL of Streptactin Beads 4FF for 2 h at room temperature. The cell lysates specified in the corresponding figures and His-BECN1 protein were added and then incubated overnight at 4°C. After washing five times with TBST (Solarbio, T1085), the mixtures were resuspended in PBS and subjected to western blot analysis. For GST affinity-isolation assay, 10 μg of the purified GST-2C protein or GST tag protein was first bound to 50 μL of glutathione-Sepharose 4B beads and then incubated with Strep II-TUFM protein overnight at 4°C. After thoroughly washing with TBST, the formed complexes were analyzed by western blot analysis.

Statistical analysis

Experimental data were presented as means ± standard deviation (SD). Statistical significance was analyzed by two-way ANOVA multiple comparisons using GraphPad Prism software version 8.0 (La Jolla, CA, USA). Differences were considered statistically significant at P values of < 0.05 (*), < 0.01 (**), and < 0.001 (***).

Funding Statement

This study was funded by the National Natural Science Foundation of China (32072840), the Beijing Natural Science Foundation (6212013), and the earmarked fund for China Agriculture Research System (CARS-35).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2023.2293442