Figure 1｜ Structure function analysis of SARS-Cov2 proteins.
The SARS-CoV2 genome generates 13–15 open reading frames (Orf) that get cleaved into several segments coding for structural proteins (Sp), nonstructural proteins (Nsp), and accessory proteins (Figure 1A). To study the role of SARS-CoV2 protein in cellular processes like cell death, two model systems were utilized: Drosophila eye and murine neuroblastoma, Neuro-2a, and cell line (Figure 1B). Using a forward genetic screen strategy, we tested the effect of targeted misexpression of individual SARS-CoV2 proteins in a fly eye model employing various transgenic fly lines (Table 1). We used the GMR-Gal4 (Moses and Rubin, 1991) line to misexpress SARS-CoV2 protein(s) in differentiating neurons and assay the effect of these proteins on the survival of the retinal neurons. The strategy was to screen the adult eye phenotype and check the retinal cell fate in the developing larval eye imaginal disc using an anti-ELAV antibody (Singh et al., 2002; Gogia et al., 2020; Irwin et al., 2020). For this study, we utilized the UAS-transgenes from two sources: one from the collection with UAS-SARS-CoV2 transgene (Zhu et al., 2021) and the other is HA-tagged UAS- SARS-CoV2 transgene (Kanca et al., 2022; Guichard et al., 2023). The rationale for using these two sets of transgenic lines was that HA-tagged lines allow detection of transgene expression in tissue using anti- HA antibody whereas the other set only has the SARS-CoV2 protein coding region for targeted misexpression studies. Using this Gal4-UAS system (Brand and Perrimon, 1993), we identified Nsp2, Nsp3, and Orf6 as candidates which resulted in neurodegenerative phenotypes in the eye (Figure 1D–F). The wild-type Canton-S fly eye is a well-organized structure, containing 600–800 ommatidia arranged in a hexagonal array with even pigmentation (Figure 1C). Comparatively, all three SARS-CoV2 proteins upon misexpression in differentiating retinal neurons displayed a rough eye phenotype. Among the three SARS-CoV2 proteins (Figure 1D–F), Nsp3 displayed the most severe rough eye phenotype characterized as disorganized and fused ommatidia with necrotic spots, a sign of neurodegeneration (Figure 1F).
We also verified the role of Nsp3 in neurodegeneration using another set of responder lines where SARS-CoV2 protein(s) are tagged with HA epitope (Kanca et al., 2022; Guichard et al., 2023). This HA-tag also allows spatiotemporal localization of the protein(s) via immunohistochemistry (IHC) approaches (anti-HA antibody) in tissue. Targeted misexpression of UAS-Nsp3-HA (GMR>Nsp3-HA) exhibits a strong neurodegenerative phenotype which is similar to the other UAS-Nsp3 transgene (Additional Figure 1). We confirmed Nsp3 expression using immunohistochemistry with Nsp3 antibody staining in GMR>Nsp3 third instar eye-antennal imaginal discs that show Nsp3 protein localization within differentiating retinal photoreceptor neurons in the GMR domain (Figure 1H and H’). Whereas the control GMR-Gal4 did not show any Nsp3 protein expression (Figure 1G and G’). We also employed HA staining to validate the localization of Nsp3 protein in the retinal neurons (Additional Figure 1). We also analyzed the sub-cellular localization of Nsp3 protein in the eye imaginal disc using pan-neuronal marker ELAV that marks the retinal neuron nuclei and PSD-95 homolog, Disclarge (Dlg) as membrane fate marker. In high-magnification images, we found that Nsp3 colocalizes with Dlg on the membrane as well as intracellularly. No extra-cellular expression was seen (Additional Figure 2).
In Neuro-2a cells, successful pEGFPN1-Nsp3 plasmid transfection was confirmed by GFP-positive Neuro-2a cells (Figure 1J) that displayed fluorescence at a 57.86% average transfection efficiency in pEGFPN1-Nsp3 transfected Neuro-2a cells (Figure 1K) as compared to control mock transfected Neuro-2a cells (Figure 1I) with no Nsp3 (P < 0.001, Figure 1K). Overall, our studies using these two models were responsive and could be used to further assess the role of Nsp3. The rough eye phenotype observed in GMR>Nsp3 indicated neurodegeneration, which is often characterized by induction of cell death. Therefore, we tested the effect of Nsp3 misexpression on cell death markers.
Figure 2｜ Nsp3 misexpression induces apoptosis.
We first used hid5’F-WT-GFP, a transcriptional reporter for the hid gene as an indicator of apoptotic cell death (Tanaka-Matakatsu et al., 2009). hid is highly expressed in cells undergoing apoptosis. A basal level of hid expression is seen in the control GMR-Gal4 third instar eye-antennal imaginal disc (Figure 2A and A’). In GMR>Nsp3 third instar eye-antennal imaginal discs, we observed a significant increase in hid expression in the GMR domain (Figure 2B and B’). To quantify hid expression, we measured the GFP fluorescence intensity in the GMR domain of the imaginal disc using an ROI approach (Chimata et al., 2022b). We found an almost two-fold increase in hid expression in GMR>Nsp3 discs compared to control, GMR-Gal4, discs (P < 0.05; Figure 2C). We also tested the effect of Nsp3 on apoptotic cell death using Dcp1, or death caspase-1, which encodes an effector caspase that cleaves specific proteins during apoptosis (Figure 2D, D’, E, E’). Very few Dcp1 positive cells are seen in the control GMR-Gal4 third instar eye-antennal imaginal discs (Figure 2D and D’) whereas in GMR>Nsp3 discs, we observed a significant increase in Dcp1 expression (Figure 2E, E’, P < 0.05, Figure 2F). We also performed a TUNEL assay (McCall and Peterson, 2004; Chimata et al., 2022b) that detects and labels fragmented DNA as seen in early apoptosis in control and GMR>Nsp3 backgrounds (Figure 2G, G’, H, and H’). GMR-Gal4 third instar larval eye disc displayed a basal level of apoptosis and had few TUNEL-positive cells (Figure 2G and G’). In contrast, GMR>Nsp3 discs showed a significantly higher number of TUNEL-positive cells in the GMR domain (Figure 2H and H’). TUNEL expression was quantified by counting the number of TUNEL-positive cells in the imaginal discs in the GMR domain. We found an almost three-fold increase in TUNEL-positive cells in GMR>Nsp3 discs compared to GMR-Gal4 control discs (P < 0.001; Figure 2I).
We further assayed the effects of Nsp3 using a cell line model. We first checked the mitochondrial/metabolic activity in Nsp3 transfected cells using the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide) assay (van Meerloo et al., 2011). In this study, we conducted the MTT assay in either mock or Nsp3 transfected Neuro-2a cells. Our results showed that Nsp3-transfected Neuro-2a cells showed significantly reduced metabolic activity compared to mock-transfected control cells (P < 0.05; Figure 2J). After 24 hours, the metabolic activity of Nsp3 transfected Neuro-2a cells reduced (0.67 ± 0.08 fold) as compared to mock-transfected control Neuro-2a cells (1.00 ± 0.07 fold). To test if cell viability is compromised due to an increase in cell death, we determined the effect of SARS-CoV2 Nsp3 overexpression on Neuro-2a cells using PI staining (Stiefel et al., 2015; Hussain et al., 2019; Lima-Silveira et al., 2019). Our results indicated that SARS-CoV2 Nsp3 expression significantly reduced Neuro-2a viability (P < 0.05; Figure 2K). After 24 hours of SARS-CoV2 Nsp3 expression, PI fluorescence intensity, an indicator of cell death, was increased (63.68 ± 1.20) compared to mock-transfected control Neuro-2a cells (52.06 ± 0.90). Taken together, data from flies and Neuro-2a cells showed increased cell death upon expression of Nsp3. Since during development and disease, several mechanisms of cell death (e.g., apoptosis, autophagy) are described, we tested these cell death mechanisms in our models.
Figure 3｜Nsp3 misexpression results in induction of autophagy and necrosis.
We assessed the effect of Nsp3 in other cell death modes like autophagy and necrosis. Autophagy is marked by phagophore formation, autophagosome formation, and autophagolysosomal degradation that are genetically controlled by evolutionarily conserved genes first identified in yeast (He and Klionsky, 2009; Shpilka et al., 2011). We used atg8a-mCherry, an autophagic marker to assess the levels of autophagy in discs (Chang and Neufeld, 2009; Nagy et al., 2015). Atg8a is an important member of the genetic machinery for autophagy. It is involved in autophagosome formation and used to degrade cargo and recycle cell components in autophagy. In control, GMR-Gal4 third instar eye-antennal discs, we observed a basal autophagy level (Figure 3A and A’). However, in GMR>Nsp3 third instar eye-antennal discs, we observed a significant increase in atg8a reporter expression in the GMR domain (Figure 3B and 3B’). Upon quantification of atg8a reporter expression using an ROI approach, we found an almost two-fold increase when compared to GMR-Gal4 (P < 0.05; Figure 3C).
Next, we wanted to assess the levels of reactive oxygen species (ROS) that increasess in stress and necrosis. We used a superoxide indicator called dihydroethidium (DHE) (Deshpande et al., 2021). In the presence of increased ROS, DHE gets oxidized to form 2-hydroxyethidium (2-OH-E+) that emits a red fluorescent signal upon intercalating with nucleic acid. In control GMR-Gal4 third instar eye-antennal imaginal discs, we observe basal levels of ROS as evident from DHE staining (Figure 3D and D’). Comparatively, in GMR>Nsp3, ROS levels are increased in the GMR domain where Nsp3 is misexpressed (Figure 3E and E’). We quantified ROS expression using an ROI approach and found a significant nearly two-fold increase in GMR>Nsp3 discs compared to control discs (P < 0.001; Figure 3F). These results clearly show that SARS-CoV2 protein Nsp3 induces multiple modes of cell death that contribute to the neurodegeneration observed in adults.
Figure 6｜Misexpression of Nsp3 in AD model exacerbates cell death and reactive oxygen species.
We next wanted to understand if the Nsp3-mediated worsening of GMR>Aβ42 phenotype and necrosis spots are a result of increased cell death. Therefore, we assayed the cell death markers in this background. Misexpression of Nsp3 in the AD model (GMR>Aβ42+Nsp3) (Figure 6B and B’) resulted in increased expression of hid5’F-WT-GFP reporter when compared to GMR>Aβ42 discs (Figure 6A and A’). Quantification of these discs revealed a significant increase of more than 1.5 folds in GMR>Aβ42+Nsp3 discs (Figure 6C’). When Nsp3 is misexpressed in the AD background (GMR>Aβ42+Nsp3), we observed significantly increased Dcp1 expression in the eye-antennal imaginal disc (Figure 6E and E’) when compared to the GMR>Aβ42 discs (Figure 6D and D’). Quantification via an ROI approach shows a consistently increased Dcp1 expression in GMR>Aβ42+Nsp3 compared to GMR>Aβ42 alone (P < 0.01, Figure 6F). Furthermore, using a TUNEL assay, we show that misexpression of Nsp3 in AD flies, GMR>Aβ42+Nsp3 (Figure 6H and H’) results in an almost two-fold increase (P < 0.01, Figure 6I) in TUNEL-positive cells when compared to GMR>Aβ42 third instar eye-antennal imaginal discs (Figure 6G and G’). We also checked ROS levels using DHE staining which reveals that GMR>Aβ42+Nsp3 (Figure 6K and K’) have a significant increase in DHE puncta when compared to GMR>Aβ42 (Figure 6J and J’) discs (P < 0.05; Figure 6L). Taken together, these results show that Nsp3 is highly pathogenic and induces high levels of apoptosis, autophagy, and necrosis in fly and cell culture models. Furthermore, misexpression of Nsp3 in the AD fly model worsens the neurodegenerative phenotype and exacerbates the cell death markers. Thus, our results can be summarized in a model, which suggests that Nsp3 can potentiate neurodegeneration observed in AD as seen in Drosophila eye model (Figure 7).