293T & THP1 cells were obtained from the American Type Culture Collection (ATCC). Cells were grown at 37°C with an atmosphere of 98% humidity and 5% CO₂. 293T cells were maintained in Dulbecco’s modified Eagle’s medium + GlutaMAX™ supplement with pyruvate (GIBCO), 1% non-essential amino acids (SIGMA), and 10% fetal bovine serum (FBS) (Takara Bio). 293T-ρ⁰ were generate as previously described in (23). 293T-ρ⁰ cells were maintained in the same media used to maintain 293T cells with the addition of 100 ug/mL of Uridine (SIGMA). THP1 cells were grown in RPMI 1640 Medium (GIBCO) supplemented with 10% FBS (Takara Bio). 293T and 293T-ρ⁰ cells were infected with lenti viruses (MOI 4) as previously described (24). THP1 cells were infected with lenti viruses (MOI 8) with the addition of 10 μg/ml polybrene (VectorBuilder) and spin-inoculated at 700×g for 25min.

293T cells were plated at a density of 45 × 10³ in 0.2% gelatin-coated (ScienCell) 96-well glass-bottom plates with high-performance #1.5 mm cover glass (Cellvis). Twenty-four hours post-plating, sub-confluent monolayers of 293T cells were transduced with LV-EV or LV-E3a. Twenty-four hours post-transduction, cells were washed two times with phosphate-buffered saline (PBS), then stained. For determination of mROS levels, cells were co-stained with 3 μM MitoSOX™ Red (MitoSOX, mitochondrial superoxide indicator) and 50 nM MitoTracker™ Deep Red FM (MTDR) for 30 min at 37°C. For assaying mROS levels after treatment with Thapsigargin (TG) using the plate reader, cells were stained with 3 μM MitoSOX for 30 min at 37°C.To quantify mROS after staining, cells were washed three times in PBS, maintained in FluoroBrite™ DMEM (GIBCO) supplemented with 12.5 mM HEPES (SIGMA) and 1% non-essential amino acids (SIGMA), and the fluorescence measured.

To determine cytosolic Ca⁺⁺ levels, cells were washed three times with Tyrode’s Salts (Sigma-Aldrich), stained for 40 min with 2 μM Fura Red™, acetoxymethyl ester (AM), cell-permeant (Fura-Red) in 0.02% pluronic F127 (Pluronic^® F-127) detergent. To determine mitochondrial Ca⁺⁺ levels, cells were washed three times with Tyrode’s Salts, stained for 40 min with 10 μM Rhod-2, AM, cell-permeant (Rhod2) and 50 nM MTDR in 0.02% pluronic F127. After staining cells were washed three times, maintained in Tyrode’s Salts, and the florescence measured. To determine mitochondrial Ca⁺⁺ levels after treating with TG, cells were stained for 40 min with 10 μM Rhod2, washed three times, maintained in Tyrode’s Salts, and the florescence measured.

Live-cell fluorescence imaging was performed using a Zeiss 710 LSM confocal microscope with an environmental chamber maintained at 37°C and 5% CO2. Laser lines used: diode 405 nm, Argon 514 nm, HeNe lasers 543 nm and 633 nm excitation wavelengths. Fluorescence was analyzed using ImageJ. Fluorescence intensities of stained cells were normalized to the unstained negative cells. A minimum of 60 images was taken for each condition across at least three independent experiments.

Fluorescence lifetime imaging microscopy (FLIM) of NADH autofluorescence was performed according to Schaefer et al. (26). Using FLIM, NADH was measured on a Zeiss LSM 710 Laser Scanning Microscope equipped with a pulsed, two-photon titanium-sapphire laser (80 MHz, 100 fs pulse width). NADH was excited at 730 nm and emission was recorded through a 460/60 nm bandpass filter using the hybrid detector HPM-100-40, mounted on the NDD port of the LSM710. Time-correlated single photon counting (TCSPC) was performed with a temporal resolution of 256 time channels within a pulse period of 12.5 ns, resulting in FLIM images of 512 × 512 pixel. The final measurement settings were 60 sec collection time; ≈ 15 μsec pixel dwell time; 135 × 135 μm² scanning area using a Plan-Apochromat 63x/1.40 Oil DIC M27 lens. SPCM 9.8 was used to record the data, which were subsequently analyzed using SPCImage 8.0 by fitting (WLS method) of a biexponential decay with lifetime components fixed to 400 ps and 2500 ps for free and protein-bound NADH, respectively. Final analysis settings were square binning of 2, peak threshold adapted to background, and shift fixed at a pixel with clear NADH signal. The mean NADH lifetime (t_mean) was calculated and is depicted in false-color coding with red indicating a shorter and blue a longer NADH lifetime. For subcellular analysis, region of interests (ROIs) of similar size were drawn for five nuclei across each image. For the mitochondria-rich region, the peak threshold was increased to remove nuclear and cytosolic NADH signal, averaging only the t_mean of the mitochondrial NADH.

Catalase activity was assessed through cleavage of H₂O₂ in 293T cell lysates collected twenty-four hours post-transfection with p-mCAT or p-EV, using a Catalysis Activity Kit (Abcam, ab83464).

Twenty-four hours post-transduction, cells were washed with cold PBS and then lysed with 2.5% n-Dodecyl-B-D-Maltoside in 20 mM HEPES (pH 7.4), 50 mM β-glycerophosphate, 2 mM EGTA, 10% (v/v) glycerol, and 0.01% Bromophenol blue. Lysates were electrophoresed on 4 to 12%, Bis-Tris gels (Invitrogen) SDS-polyacrylamide NuPAGE™ gels. Gels were transferred to a nitrocellulose membrane by the iBlot Gel Transfer System (Invitrogen), membranes blocked for one hour in 5% nonfat milk in 150 mM NaCl and 25 mM Tris-buffered saline in 0.1% Tween^® 20 Detergent (TBST) then incubated overnight at 4°C with shaking. The primary antibody was diluted 1:1000 in TBST. Membranes were then washed three times with TBST and incubated with Alexa Fluor-conjugated secondary antibodies for one hour at room temperature. Protein levels were quantified using the Odyssey imaging system (LiCOR Biosciences). GAPDH as a loading control.

293T cells were plated at a density of 100 × 10³ in 24-well plates. Twenty-four hours post-plating sub-confluent monolayers of 293T cells were infected with LV-EV or LV-E3a. Six hours post-infection cells were co-transfected using the TransIT-X2^® Dynamic Delivery System (Mirus Bio) with the plasmids encoding the components of the NLRP3-I (25), followed by p-mCAT or p-EV transduction. Twenty-four hours post-transfection, cells were washed two times with PBS, then cultured with or without the addition of 50 uM MnTBAP. Cell lysates and culture supernatants were collected twelve hours post-treatment and centrifuged to remove cell debris. Supernatant IL-1β was quantified by ELISA (Abcam, ab197742).

THP-1 cells were plated at a density of 100 × 10³ in 96-well plates. THP1 cells were infected with LV-EV or LV-E3a. Six hours post-infection, THP1 cells were differentiated into macrophages with 50 ng/ml Phorbol 12-myristate 13-acetate (PMA) overnight. After differentiation, cells were washed two times with PBS, and fresh media was added with the addition of 100 μM MnTBAP, 5 μM MCC950, or 10 μM N-methyl-4-isoleucine-cyclosporin (NIM811). Nine hours post-treatment with MnTBAP or MCC950 or one-hour post-treatment with NIM811, THP1 macrophages were stimulated with 100 ng/ml Lipopolysaccharides (LPS) and 2.5 μM of nigericin for nine hours. Supernatants were collected, centrifuged, and the amount of IL-1β in the supernatants measured by ELISA (Abcam, ab46052). THP1 cells stably expressing LV-mCAT or control vector, were infected with LV-EV or LV-E3a, and 6 hours post-infection the THP1 cells were differentiated into macrophages, and supernatant IL-1β quantified via ELISA.

Twenty-four hours post-plating, sub-confluent monolayers of 293T cells were infected with LV-EV or LV-E3a. Six hours post-infection, cells were transfected with the plasmids encoding p-mCAT or p-EV, or twenty-four hours post-infection cells were treated with or without the addition of 50 uM MnTBAP or 10 uM NIM811. Twenty-four hours post-transfection or ten hours post-treatment, cells were washed with PBS, then dissociated from the plate with Trypsin. The dissociated cells were then pelleted via centrifugation and lysed with hypertonic buffer consisting of 20 mM Tris-HCl (pH 7.4), 10 mM NaCl, and 3 mM MgCl2 for fifteen minutes on ice, then treated with Non-Ionic Detergent P-40 (NP-40) 0.05% (v/v) and vortexed at max speed. Cell lysates were centrifuged for 15 minutes at 1,000 g, and the resulting supernatants were centrifuged for 1 hour at 200,000 g to isolate a cytosolic enriched fraction. After centrifugation, protein concentrations were determined from the resulting supernatants by Bradford assay using the Bio-Rad protein assay (Bio-Rad, Hercules, CA). DNA was isolated from equal amounts of protein using the DNA Clean & Concentrator-5 (Zymo Research, 11-303). Levels of mitochondrial DNA in the cytosolic enriched fraction were then determined via TaqMan Real-Time PCR assays to detect mitochondrial NADH dehydrogenase subunit 5 (MT-ND5) using the TaqMan probe (MT-ND5 Hs02596878_g1).

For quantitative analyses, a minimum of 3 biological replicates were analyzed. Western blot data are from the respective experiment, processed in parallel, and represent at least three independent experiments. One-way ANOVA was performed for statistical differences between three or more groups, followed by a post hoc Tukey’s HSD test to test for statistical differences. For studies that require a quantitative evaluation between two groups, statistical significance was determined using unpaired two-tail Student’s t-test. All data bar graphs, data are reported as mean ± standard error of the mean (SEM). All statistical analysis was done on GraphPad Prism 9.01. For student’s t-test * = p < value 0.05, ** = p < value 0.01, *** = p < value 0.001, **** = p < value 0.0001).

A major unanswered question is how 2-E and 2-3a activate the NLRP3-I. We hypothesized that expression of 2-E and 2-3a results in increased Ca⁺⁺ flux into the cytosol where it is taken up by the mitochondrion through the MCU. Within the mitochondrion, the Ca⁺⁺ activates the pyruvate and α-ketoglutarate dehydrogenases to generate excessive NADH. The increased NADH overloads the electron transport chain producing increased mROS. The mROS oxidizes the mtDNA which is released through the mtPTP to bind to the NLRP3-I. This activates caspase 1 to cleave pro-IL-1β and pro-IL-18 resulting in the secretion of active IL-1β and IL-18 (10, 22).

To test this hypothesis, we constructed a polycistronic expression vector combining 2-E+2-3a (LV-E3a) to determine how these viroporins function synergistically to engage the NLRP3-I ( Figures 1A, B ). In this vector the 2-E+2-3a sequences were separated by the self-cleaving 2A peptide site (27) to allow co-expression from a single transcript. The expression of 2-E+2-3a in LV-E3a transduced 293T cells was confirmed by Western blot ( Figure 1C ). We then demonstrated that LV-E3a transduced 293T cells experience increased cytosolic Ca⁺⁺ with Fura-Red by plate reader assay ( Figures 1D, J ) and confocal microscopy ( Figure 1E ). Additionally, we demonstrated that LV-E3a transduced 293T cells experience increased mitochondrial Ca++ with Rhod2 by plate reader assay ( Figures 1F, J ) and confocal microscopy ( Figure 1G ). The LV-E3a transduced 293T cells had elevated levels of both cytosolic and mitochondrial Ca⁺⁺.

Treating with thapsigargin (TG) triggers Ca⁺⁺ flux into the cytosol resulting in increased mitochondrial Ca⁺⁺ uptake (28, 29). Treatment with the mitochondrial calcium uniporter inhibitor 11 (MCUi11) blocks Ca⁺⁺ entry into the mitochondrion (30, 31). Treatment with MCUi11 abolished mitochondrial calcium uptake in LV-E3a transduced cells ( Figure 1H ) following TG treatment. Thus, expression of 2-E+2-3a in 293T cells results in increased Ca⁺⁺ flux into the cytosol resulting in an elevation of cytosolic Ca⁺⁺. The cytosolic Ca⁺⁺ is then taken up by the mitochondrial calcium uniporter resulting in elevated mitochondrial Ca⁺⁺ ( Figure 1H ).

Elevated mitochondrial Ca⁺⁺ activates the tricarboxylic acid cycle dehydrogenases generating excess NADH (21). Expression of 2-E+2-3a in 293T cells resulted in increased mitochondrial NADH levels detected via NADH fluorescence lifetime imaging (FLIM) (NAD⁺/NADH) mean (ps) ( Figures 1I, J ).

Excessive NADH levels can overload the electron transport chain producing increased mROS. The increased NDAH in 2-E+2-3a expressing 293T cells was associated with increased mROS production detected by staining transduced cells with MitoSOX, which detects mitochondrial superoxide anion production ( Figures 2A–C ). To determine if the increased mROS was due to the entry of Ca⁺⁺ into the mitochondrion, we treated the cells with MCUi11 which blocked the increased mROS production ( Figure 2D ). As predicted MCUi11 blocked the increased mROS production. Hence the increased mROS production in 2-E+2-3a cells was dependent on elevated mitochondrial Ca⁺⁺.

To confirm that the 2-E+2-3a induced ROS production was mROS, we transfected the 2-E+2-3a expressing 293T cells with a vector expressing mitochondrially-targeted catalase (mCAT) ( Figures 2E–I, L ) which removes mitochondrial H₂O₂ (32). We also treated the cells with the mitochondrially targeted catalytic metalloporphyrin anti-oxidant, MnTBAP ( Figures 2J, K, M ) (33). Treatment with either MnTBAP or mCAT extinguished the 2-E+2-3a activated ROS product, confirming that the ROS was generated by the mitochondrion.

2-E+2-3a induced mROS is involved in NLRP3-activation and IL-1β production. SARS-CoV-1, activation of the NLRP3-I and associated pathogenicity were observed for both viroporins 1-E+1-3a (11, 14, 34, 35). To determine if this is the case for SARS-CoV-2, we used two model systems to determine if 2-E+2-3a expression activates the NLRP3-I via the mitochondrion. In the first, we reconstituted the inflammasome in 293T cells by transforming with plasmids encoding the components of the NLRP3-I ( Figure 3A ) (25). In the second, we employed the human acute monocytic leukemia derived cell line, THP1, which can be converted to macrophages by treatment with phorbol ester (PMA) and the macrophages treated with LPS + nigericin ( Figure 3B ) (36). The reconstituted NLRP3-I 293T cells were transduced with the 2-E+2-3a expression vector ( Figure 3A ). The THP-1 cells were first transduced with the 2-E+2-3a expression vector and then treated with PMA and LPS + nigericin ( Figure 3B ). The expression of 2-E+2-3a in both cell systems, 293T ( Figure 3C ) and THP1 macrophages ( Figures 3D–F ) resulted in enhanced secretion of NLRP3-activated IL-1β in 293T-NLRP3-I ( Figure 3E ) and THP1 macrophages ( Figures 3D, E ).

We then confirmed that 2-E+2-3a expression activates the NLRP3-I and IL-1β secretion via increased mROS production. 293T cells expressing the inflammasome proteins ( Figure 3C ) and LPS-nigericin treated THP1 macrophages ( Figures 3D, E ) were treated with mitochondrially targeted antioxidants: transformation with mCAT or treatment with MnTBAP. Both mCAT expression and MnTBAP treatment impaired IL-1β secretion.

We then determined if mROS activation of the NLRP3-I was mediated by release of a mitochondrial component via the mtPTP, which has been conjectured but not proven. We treated 2-E+2-3a transduced THP1 macrophages with the specific mtPTP inhibitor N-methyl-4-isoleucine-cyclosporin (NIM811). NIM811 blocks the mtPTP by binding to cyclophilin D, analogous to cyclosporin A (CsA), but without calcineurin inactivation (37–39). NIM811 treatment suppressed the secretion of IL-1β following LPS + nigericin activation of the THP1 macrophages ( Figure 3F ) demonstrating for the first time the mtPTP is the route by which a mitochondrial factor, presumably mtDNA, reaches the NRLP3-I. Finally, to confirm that 2-E+2-3a secretion of IL-1β occurred through NRLP3-I, we treated THP1 macrophages with MCC950, a selective small molecule inhibitor that binds to the NACHT domain of NLRP3 (40). MCC950 blocked IL-1β secretion in 2-E+2-3a expressing THP1 macrophages, demonstrating that E3a secretion of IL-1β, was mainly through the activation of the NLRP3-I ( Figure 3G ).

We next demonstrated that LV-E3a transduced 293T cells increased levels of cytosolic mtDNA via Real-Time PCR quantification of mitochondrial NADH dehydrogenase subunit 5 (MT-ND5) in cytosolic-enriched supernatants ( Figures 4A–C ). To confirm that 2-E+2-3a increased the levels of cytosolic mtDNA via efflux through the mtPTP, we treated 2-E+2-3a expressing 293T cells with NIM811, which extinguished the 2-E+2-3a elevation of cytosolic mtDNA. This confirms that mtPTP opening contributes to 2-E+2-3a elevation of cytosolic mtDNA ( Figure 4B ). To determine if mROS contributed to 2-E+2-3a release of mtDNA we transduced 293T cell expressing 2-E+2-3a cells with mCAT or treated the cells with MnTBAP. Both mitochondrial catalytic antioxidants impaired 2-E+2-3a elevation of cytosolic mtDNA ( Figures 4B, C ). Finally, to confirm that 2-E+2-3a activation of the NRLP3-I, was dependent on mtDNA we showed that secretion of the IL-1β was lost in 293T cells with a reconstituted inflammasome ( Figure 4D ) that had been cured of their resident mtDNA (ρ⁰ cells). The presence of mtDNA is thus an absolute requirement for 2-E+2-3a activation of the NRLP3-I.

Thus, we have demonstrated that co-expression of 2-E+2-3a enhances Ca⁺⁺ leakage into the cytosol, increasing levels of cytosolic and mitochondrial Ca⁺⁺. This 2-E+2-3a mediated increase in mitochondrial Ca⁺⁺ increases mitochondrial NADH in association with increased mROS. mtDNA is then released into the cytosol via the mtPTP which activates the NLRP3-I stimulating the secretion of IL-1β. Increasing mitochondrial antioxidant defenses through treatment with the pharmacological mROS scavenger MnTBAP or genetic expression of mCAT blocks 2-E+2-3a induced activation of the NLRP3-I, which is absolutely dependent on the presence of mtDNA. Together these findings reveal that the mechanism by which 2-E+2-3a engage the NLRP3-I is via viroporin manipulation of mitochondrial physiology causing release of mtDNA to bind to the NLRP3-I.