The plasmid pHO-1-Luc was provided by J. Alam and contained a 15- kb murine HO-1 promoter upstream of a luciferase gene (reporter gene vector pSK-luc) as described previously (48). pCEFL-AU5-vGPCR was constructed in Dr. Gutkind’s laboratory introducing vGPCR cDNA as a Bgl2/Not1 fragment in pCEFLAU5 (28). The reporter construct pGL3_Nrf2-3xARE-Luc (minimum promoter containing three ARE sites in tandem, T/CGCTGAGTCA) was a gift of M. Marinissen, and the pG5 Luc (minimum promoter containing 5 sites for GAL4 binding) was purchased from Promega. The expression plasmid for Nrf2 WT, pCDNA3 His B – V5 Nrf2 (Wild Type), was a kind gift from M. McMahon y J. Hayes and previously described (49). Briefly, the PCR amplification of the murine Nrf2 coding sequence and 50 nucleotides of the 5′-untranslated region was ligated into EcoRV-digested pcDNA3.1/V5HisB (Invitrogen), allowing expression of mNrf2-V5-his fusion protein from a CMV promoter. The expression plasmid for Nrf2 Dominant Negative (Nrf2 DN) (reporter gene vector pEF) was provided by J. Alam and described previously (48). The expression plasmids pCDNA3_Gα12-QL and pCEFL_HA_Gα13-QL were obtained by cloning the constitutively active forms of both cDNAs and cloned as Bgl2/Not1 fragments into pCDNA3 and pCEFL-HA, respectively. pCEFL_AU5_RhoA-QL and pCEFL_Rho-N19 Dominant Negative were cloned as BamH1/EcoR1 fragments and introduced in the respective vectors in Dr. Gutkind’s laboratory who provided them as a gift (40, 50, 51). The expression vectors for the MEKs were previously described (52).

NIH3T3 fibroblasts were received from Dr. S. Gutkind’s laboratory and maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen) supplemented with 10% calf serum. Stable transfections of NIH3T3 for vGPCR were received from Dr. S. Gutkind’s laboratory and described previously (40). Stable transfections of NIH3T3_vGPCR for different shRNAs were performed using the Lipofectamine Plus Reagent (Invitrogen). NIH3T3_vGPCR cells were plated at 60% confluence in 10 cm plates and transfected with 2ug of shRNA (shScramble and shRNAs targeted against Nrf2 - GIPZ Nfe2l2 shRNA Thermo RMM4532-EG18024). Transfected cells were selected with 750 ug/ml G418 (Promega Corp., Madrid, Spain) and 0.4ug/ml Puromycin (Invitrogen). mECs were obtained from Balb/C mice (NCI, Bethesda, MD) as previously described (33). By transfection with KSHVBac36, the vector containing the insert with the genome of KSHV in Bacterial Artificial Chromosome (KSHVBac36), mECK36 cells were generated. mECKnull is a variant of the latter that lost the plasmid and was subsequently used as the source for the generation of mECK16-delta-Revertant or mECK16-delta–vGPCR cell lines, which express reinstalled KSHV genomes in its complete or devoid of vGPCR versions, respectively (53).

Cells were transfected with different expression plasmids together with 1ug of the indicated reporter plasmid per well in 6-well plates. In all cases, the total amount of plasmid DNA was adjusted with pcDNA3 empty and 0.2ug of pCDNA3-b- galactosidase. Firefly luciferase activity in cellular lysates was assayed using the luciferase reporter system catalog number E1500 (Promega Corp.), and light emission was quantified using a luminometer (Junior Berlthold). To study transcriptional activation domains of transcription factors, we used the Luciferase GAL4 reporter system, where a Gal4 DBD – Nrf2 TAD vector expresses a fusion protein containing the DNA binding domain of GAL4 and the transactivation domain of Nrf2, is co-transfected with the pG5 Luc. When co-transfected in the same cell, these two constructs allow for the assessment of the regulation of gene expression by signal transduction pathways that impinge on the transcription factor Nrf2 using luciferase as a reporter gene. All assays were performed in three biological replicates and measured in triplicates for quantification.

Total cellular RNA was isolated using TRIReagent (Genbiotech-MRC) using the manufacturer’s protocols. RNA concentration was measured by Nanodrop™ (Thermo Fisher), and 2ug of RNA was treated with RNase-free DNase I (Invitrogen). DNase treatment was performed for 15 min at room temperature, followed by EDTA treatment for 10 minutes at 65 ° to stop the DNase. After DNase treatment, the samples were incubated for 5 minutes at 65 ° in the presence of 500ng OligodT (Genbiotech). cDNA synthesis was performed using MMLV (Moloney murine leukemia virus) reverse transcriptase (RT) (Promega) with RNaseOUT (Invitrogen) for 50 minutes at 37°. qPCR was performed using 1ul of cDNA in a BioRad CFX96 instrument with 9ul of the Master Mix qPCR 2.0 (PB-L, Productos Bio-Logicos), including SYBR Green detection. qPCR cycles: STEP 1 95° 3min; STEP 2 (x40) 95° 20seg, 62° 20 seg and 72° 30 seg with reading; STEP 3 95° 1 min; STEP 4 Melting curve 60° 30seg, 95° 30seg (temperature of the sample is then increased incrementally as the instrument continues to measure fluorescence). Luciferase mRNA was quantified by qPCR using the following primer set: forward (5-CCGCCGTTGTTGTTTTG- 3) and reverse (5-ACACAACTCCTCCGCGC-3). For vGPCR, the primer sequences were: forward (5-AGGAGCGATAGATATACTG-3) and reverse (5-CAACACTTCTGCCAATAG-3). Luciferase and vGPCR mRNA expression were normalized against the Beta-galactosidase mRNA, in which primer sequences were: forward (5-CCACGGAGAATCCGACG-3) and reverse (5-GCGAGGCGGTTTTCTCC-3). In every run, melting curve analysis was performed to verify the specificity of products and water and No–RT controls. Three biological replicates were used for each sample and measured in triplicates in the qPCR analysis. Data were analyzed using the ΔΔCT method previously described (54); we used this method because the genes were amplified with comparable efficiencies. Target gene expression was normalized to Beta-galactosidase by taking the difference between CT values for target genes and Beta-galactosidase (ΔCT value). These values were then calibrated to the control sample to give the ΔΔCT value. The fold target gene expression is given by the formula: 2–ΔΔCT. We used Beta-galactosidase as a reference gene to normalize transfected cells.

Nrf2 and phospho-Nrf2 were detected by Western blotting with anti-Nrf2 (Abcam ab89443) and anti-phospho-Nrf2 (Abcam ab76026). For kinase analysis, we used anti-ERK2 (Santa Cruz: sc-154), Anti-Phospho–ERK1/2 (Santa Cruz sc-7383), Anti-p38 (Santa Cruz sc-535-G), Anti-Phospho-p38 (Cell Signaling 9211), Anti-JNK (Santa Cruz sc- 474-G), Anti-Phospho–JNK (Cell Signaling 9255S), Anti-Akt1 (Santa Cruz sc-1618), Anti-Phospho-Akt1/2/3 (Santa Cruz sc-16646-R), Anti-Keap1 (Cell Signaling 8047) and anti-HO-1 (Cell Signaling 43966). Proteins were visualized by enhanced chemiluminescence detection (Amersham Biosciences) using secondary antibodies coupled to horseradish peroxidase or secondary antibodies coupled to fluorophores and detected using an Odyssey System (Li-cor). The MEK inhibitor PD98059 (catalog number 513000) was purchased from Calbiochem (San Diego, California).

NIH3T3, NIH3T3_vGPCR, NIH3T3_Gα12-QL, NIH3T3 Gα13-QL, and NIH3T3 RhoA-QL cells were seeded on glass coverslips. Cells were serum-starved for 24 h, washed twice with 1ml PBS, and then fixed and permeabilized with 4% formaldehyde and 0.05% Triton X-100 in 1ml PBS for 10 min. After washing with PBS, cells were blocked with 1% bovine serum albumin and incubated with anti Nrf2 (Abcam) as primary antibody O.N. at 4C°. Following incubation, cells were washed three times with 1ml PBS and then incubated for an additional hour with the corresponding secondary antibody (1:1000) conjugated with fluorescein isothiocyanate (Molecular Probes). Cells were washed three times with 1ml PBS and stained with Propidium Iodide (1 ug/ml) (Molecular Probes) in the last wash. Coverslips were mounted in Mowiol mounting medium (Sigma) and viewed using a confocal microscope (Fluoview FV300, Olympus, Japan). For nucleus/cytoplasm ratio quantification of protein localization, we used ImageJ and the results were displayed in the graph as fluorescence intensity ratio. Values near one indicate homogenous protein localization, while values over one indicate mainly nuclear localization, and values below one indicate localization mainly cytoplasmically. For mouse-KS tumor immunofluorescence analysis, we used frozen mouse KSHV (+) and KSHV (–) tumor samples obtained as previously described (33, 55). Briefly, mECK36 KSHV (+) cells injected subcutaneously into the flanks of nude mice form mouse KSHV (+) tumors 5 weeks after injection. KSHV (–) tumor cells injected subcutaneously into the flanks of nude mice form KSHV (–) tumors 3 weeks after injection. Immunofluorescence assay (IFA) of mECK16-deltavGPCR, mECK16Δ-Revertant, Mouse KSHV-positive, and KSHV-negative tumors was performed as previously described (56). Briefly, cells were fixed in 4% paraformaldehyde for 10 min and washed with PBS. Cells were permeabilized in 0.2% Triton-X/PBS for 20 min at 4°C. After blocking with 3% of BSA in PBS and 0.1% Tween 20 for 60 min, samples were incubated with primary antibodies overnight at 4C. After PBS washing, samples were incubated with fluorescent secondary antibodies for 1 hour (Molecular Probes), washed, and mounted with ProLong Gold antifade reagent with DAPI (Molecular Probes).

NIH3T3, NIH3T3_vGPCR, NIH3T3_vGPCR_shScramble, NIH3T3_vGPCR_shNrf2 stable cell lines were used to induce tumor allografts in 7-week athymic (nu/nu) nude mice. Cells were harvested, washed, counted, and resuspended in PBS. 1 X 10⁶ NIH3T3_vGPCR (control) or NIH3T3_vGPCR_shNrf2 cells were injected subcutaneously in the right flank of six and five nude mice, respectively. Mice were monitored twice weekly until each animal developed one tumor in the area of the cell injection. Tumor volume and body weight were measured every other day during the investigation period. Tumor volumes (V) were determined by the formula V=LxW ²x0.5, with L being the longest cross-section and W the shortest. Data are mean ± S.E.M expressed as tumor volume (cm³), calculated as described above. Animals from each group were euthanized for tissue retrieval at the various time points indicated in the study. Using standard protocols, the tissues were fixed in 4% buffered paraformaldehyde overnight, dehydrated, and embedded in paraffin. H&E-stained sections were used for diagnostic purposes.

All animal studies were carried out according to the Institutional Animal Care and Use Committee of the Facultad de Ciencias Exactas y Naturales (FCEN) the University of Buenos Aires approved protocol, and local government regulations (Servicio Nacional de Sanidad y Calidad Agroalimentaria, RS617/2002, Argentina). All mice were maintained under a 12:12 h dark/light cycle with food and water ad libitum. They were grouped-housed (3-5 mice per cage). Male and female mice were used and randomly assigned to the different experimental groups. All mice were euthanized by Carbon Dioxide inhalation using especially appropriate equipment of the Animal Facility.

Sample sizes were estimated based on variance obtained from previous studies, an alpha level of 0.05, and statistical power >0.8. To minimize any possible bias during tumor measurements, mice from both experimental groups were housed together and were evaluated in random order by experimenters blinded to the treatment conditions. No exclusion criteria were a priori established, and all injected mice were included in the analysis. Normality of data distributions was estimated with Shapiro-Wilk tests, two-tailed parametric statistics were used in all cases, and the threshold for significance was set at p = 0.05. The p values are indicated along with statistical parameters, such as the number of cells and animals used in each experiment in the results and the figure legends. Statistical analysis was performed with GraphPad Prism 9 software. Group differences were analyzed by student t-test or ANOVA followed by Dunnet Test. Experimental groups are compared to the control. Two-way repeated measures ANOVA was used for tumor volume analysis. A p-value <0.05 (*) was considered statistically significant.

Different band intensities corresponding to Western blot detection of protein samples were quantified using the ImageJ software.

We performed Luciferase Reporter Assays to identify the elements within the HO-1 promoter responsible for vGPCR activation. For this aim, we used a reporter construct containing 4.9Kb of the HO-1 promoter (HO-1 4.9 Luc) and serial deletions (HO-1 3.8 Luc, HO-1 1.4 Luc, and HO-1 0.3 Luc) ( Figure 1A ). As shown in Figure 1B , the loss of the proximal ARE reduces the transcriptional activity of the HO-1 (4.9Kb) promoter. This result suggests that the ARE sequence might be responsible for HO-1 promoter activation by vGPCR. Next, we used a minimum promoter containing three ARE sites in tandem (3xARE Luc) to determine the role of the HO-1 ARE sites in activating the HO-1 promoter by vGPCR. We co-transfected this reporter with expression plasmids for Nrf2 WT and Nrf2 Dominant Negative (DN), lacking the Nrf2 transactivation domain as control of responsiveness of the reporter ( Supplementary Figure 1 ). To evaluate the effect of over-expression of vGPCR, we co-transfected the 3xARE Luc reporter with expression plasmids for vGPCR. As seen in Figure 1C , vGPCR produced an increment in the reporter activity, indicating that vGPCR may activate ARE sites in target genes. To determine if the transcription factor Nrf2 mediated this effect, we co-transfected vGPCR with plasmids that express Nrf2 WT or Nrf2 DN. Interestingly co-transfection with vGPCR and Nrf2 WT produce a higher induction than the observed for vGPCR alone, suggesting that vGPCR and Nrf2 may be linked by a common signaling pathway. The induction produced by vGPCR was abolished by Nrf2 DN. Altogether, these results ( Figure 1C ) suggest that vGPCR activates ARE sites through Nrf2. These results were confirmed by RT-qPCR analysis of Luciferase mRNA expression ( Figure 1D ). vGPCR over-expression was confirmed by RT-qPCR ( Figure 1E ).

Like many GPCRs, vGPCR activates downstream effectors like Gα12, Gα13, and RhoA. To determine if those proteins were involved as mediators in the activation mechanism of the ARE sites by vGPCR, we performed luciferase assays with the 3xARE Luc construct co-transfected with expression vectors for vGPCR and downstream effectors. As shown in Figure 1F , vGPCR and its downstream effectors activate transcriptional response at the ARE sites. Moreover, the activation of ARE sites by vGPCR, Gα12, and Gα13 are RhoA dependent, as the co-transfection with RhoA Dominant Negative (RhoAN19) produces a decrease in the activation of the reporter when compared with the effect produced by vGPCR, Gα12 and Gα13 alone ( Figure 1G ). These results suggest that vGPCR, Gα12, and Gα13 activate ARE sites through RhoA.

To evaluate the role of Nrf2 in the induction of the HO-1 promoter by vGPCR, we performed Luciferase Reporter Assays. We used a reporter with the luciferase gene upstream of the murine HO-1 promoter. As seen in Figure 2 , vGPCR induces an increment in the activity of the HO-1 promoter when compared to cells transfected with a control plasmid (pCDNA3.1 empty). To demonstrate that this effect was mediated through Nrf2, we co-transfected vGPCR with Nrf2 WT or with Nrf2 DN. The effect of co-transfection of vGPCR and Nrf2 WT in HO-1 promoter was bigger than the effect of vGPCR alone. Interestingly, the co-transfection of vGPCR with Nrf2 DN produced a decrease of about 30% respective to the effect of vGPCR alone. These results suggest that Nrf2 plays a key role in activating the HO-1 promoter by vGPCR ( Figure 2A ). These results were confirmed by RT-qPCR analysis of Luciferase mRNA expression ( Figure 2B ).

Like many transcription factors, Nrf2 has to be activated on its Transcriptional Activated Domain (TAD) to recruit the transcriptional machinery needed to activate the transcription of target genes. To determine if vGPCR affects TAD activation, we performed luciferase assays. We co-expressed a Gal4DBD–Nrf2TAD fusion protein and a Gal4 Binding Element upstream luciferase and evaluated reporter activity in control cells and cells overexpressing vGPCR, Gα12-QL, Gα13-QL, or RhoAQL. As shown in Figure 3A , vGPCR and the downstream signaling components can activate this promoter, suggesting that Nrf2-TAD targets vGCPR-triggered signaling. In basal conditions, Nrf2 localizes in the cytoplasm, and Keap1 drives it to proteasomal degradation, whereas in stimulated conditions, the complex Nrf2-Keap1 is dissociated, and Nrf2 translocated to the nucleus. To determine if vGPCR affects Nrf2 nuclear translocation, we performed immunofluorescence assays using a specific Nrf2 antibody on NIH3T3, NIH3T3_vGPCR, NIH3T3_Gα12-QL, NIH3T3_Ga13-QL and NIH3T3_RhoA-QL (all stable cell lines). As shown in Figure 3B , in NIH3T3 cells, Nrf2 has a predominantly cytoplasmic localization whereas, in the four stable cell lines that express vGPCR, Gα12-QL, Gα13-QL, or RhoA-QL, Nrf2 is found predominantly in the nucleus ( Figures 3B–C ). These results suggest that vGPCR and its effectors downstream have the same effect on Nrf2 nuclear translocation.

Given that it has been reported that Nrf2 is directed to proteasomal degradation and that in-stimulated conditions, it can be phosphorylated; we wondered if overexpression of vGPCR could stabilize the protein levels of Nrf2 and affect its phosphorylation levels. For this purpose, we performed Western Blots using lysates of NIH3T3 and NIH3T3_vGPCR cells and evaluated the levels of total and phosphorylated Nrf2 proteins in both cell lines. As seen in Figure 4A , vGPCR augments the level of total and phosphorylated Nrf2. These results suggest that vGPCR not only stabilizes Nrf2, but it also induces Nrf2 phosphorylation levels.

It is known that different post-translational modifications such as phosphorylation, ubiquitination, sumoylation, and many others can modify Nrf2 and Keap1. We evaluated different kinase activation pathways in our model to determine if phosphorylation on the Nrf2-Keap1 complex is a mediator of the vGPCR effect. We performed Western Blot assays with specific antibody pairs (phospho-protein and total protein) for four different relevant kinases, ERK1/2 ( Figure 4B ), p38 ( Figure 4C ), AKT ( Figure 4D ), and JNK ( Figure 4E ). In all cases, we evaluated NIH3T3, NIH3T3_vGPCR, and an appropriate positive control (see Figures 4B–E ). As seen in Figure 4 , vGPCR can activate ERK1/2 and p38 but not AKT and JNK in our model.

As we showed, vGPCR affects Nrf2 transcriptional activity, so we wanted to know if the ERK1/2 pathway mediated this effect. For this aim, we performed Luciferase assays with the murine HO-1 promoter ( Figure 5A ), the 3xARE Luc ( Figure 5B ), and the GAL4–Nrf2TAD reporter system ( Figure 5C ). We transfected NIH3T3_vGPCR cells with the reporter construct. We activated the ERK1/2 pathway co- transfecting with expression plasmids for MEK-EE (constitutive activated) or inhibited the pathway co-transfecting with MEK-AA (dominant negative). Figure 5A shows that co-transfection of vGPCR and MEK-EE did not produce a higher effect in HO-1 promoter activation compared to vGPCR alone. This might be due to a saturation of the system when activation of the signaling axis to this particular reporter construct is already ignited by vGPCR. On the other hand, co-transfection of vGPCR with MEK-AA diminished the effect observed with vGPCR alone. Figures 5B, C showed a higher effect when co-transfecting vGPCR with MEK-EE than vGCPR alone. According to this result, co-transfection with vGPCR and MEK-AA produced a decrease in reporter activation concerning vGPCR alone. These results suggest that the ERK1/2 pathway mediates the vGCPR effect on Nrf2 transcriptional activity, which might have participated in HO-1 promoter activation.

To evaluate the ability of vGPCR to induce Nrf2 nuclear translocation by ERK1/2 signaling, we performed immunofluorescence assays using an Nrf2-specific antibody treating NIH3T3_ vGPCR cells with or without the MEK inhibitor PD98059 (20uM) for 2hs. Figure 6 shows that the inhibition of the ERK1/2 pathway by PD98059 impairs the nuclear translocation of Nrf2. The ratio of Nuclear Fluorescence/Cytoplasm Fluorescence drops from 2.5 in NIH3T3_vGCPR cells, which indicates nuclear localization to 0.5, indicative of cytoplasmic localization. This result suggests that the effect of vGPCR on Nrf2 nuclear localization is mediated by ERK1/2. Whereas vGPCR increases Nrf2 stability and phosphorylation and considering that vGCPR activates the ERK1/2 pathway, we wanted to test if the ERK1/2 pathway has a role in Nrf2 phosphorylation. As seen in Supplementary Figure 2 , the inhibition of the ERK1/2 pathway did not produce a decrease in the phosphorylation levels of Nrf2 induced by vGPCR, suggesting that the effect of ERK1/2 may not be through direct phosphorylation of Nrf2 ( Supplementary Figure 2 ).

Whereas parental NIH3T3 cells are non-transformed, they acquire the capability to form foci in cell culture models and to induce tumors in nude mice when transfected by a plasmid construct expressing an oncogene. Thus, vGPCR-overexpressing NIH3T3 cells, but not the parental cells, have been reported to induce tumors when injected into nude mice (57). Prompted by our findings, we used these models to investigate whether silencing Nrf2 could affect vGPCR-induced tumorigenesis in vivo. For this aim, we transfected the NIH3T3_vGPCR cell line with different shRNAs directed against Nrf2 (NIH3T3_vGCPR–shNrf2) and with a control shRNA (NIH3T3_vGPCR–shScramble). We generated stable cell lines and evaluated them for Nrf2 expression levels. As shown in Figure 7A , stable cell lines named NIH3T3_vGPCR-shNrf2-2 and NIH3T3_vGPCR-shNrf2-2.1 showed lower levels of Nrf2 expression when compared with NIH3T3_vGPCR. We injected 1x10⁶ NIH3T3_vGPCR (control) or NIH3T3_vGPCR–shNrf2-2.1 cells into the right flank of six and five nude mice, respectively, and observed the mice twice a week. In concordance with the in vitro described effects upon HO-1 expression, silencing Nrf2 by shRNA strongly impacts the tumorigenic behavior of transformed cells in vivo. Tumors produced by NIH3T3_vGPCR–shNrf2_2.1 tend to be smaller than the ones observed in the NIH3T3_vGPCR group ( Figures 7B–C ); probably due to the significant delay in tumor onset found for Nrf2 silenced cells ( Figure 7D ).

To determine the specific contribution of vGPCR signaling to the subcellular localization of Nrf2 in the context of full KSHV genome-bearing cells, we used a previously described Bac16-delta vGPCR mutant or its revertant in mECK36 cells that have lost the Bac36 episome by lack of antibiotic selection, mECK16-deltavGPCR, and mECK16-revertant, respectively (53). In Figure 8A , top panel, Nrf2 subcellular localization is mainly nuclear in cells infected with the complete KSHV genome (mECK16-revertant cells), but in cells infected with a KSHV virus lacking vGPCR (mECK16-deltavGPCR cells) the Nrf2 subcellular localization shift to a more cytoplasmic signal ( Figure 8A bottom panel). Moreover, western blot analysis of these cells showed that the cytoplasmic subcellular localization of Nrf2 in vGPCR mutant cells correlates with increased expression of Keap1 and decreased expression of HO-1, Figure 8B . Finally, we were able to show that in KSHV-positive [KSHV (+)] mouse KS-like tumors, Nrf2 activation (Nrf2 phosphorylation) and subcellular localization is mainly nuclear in contrast with that observed in KSHV-negative [KSHV (–)] mouse tumors, Figure 8C .