H1299 cells (ATCC CRL-5803) and A549 cells (ATCC CCL-185) were grown in Dulbecco’s modified Eagle’s medium (DMEM; Sigma) supplemented with 10% fetal calf serum (PAN Biotech) and 100U of penicillin/100 µg of streptomycin per ml (PAN Biotech) in a 5% CO₂ atmosphere at 37°C. IKKα knock-down cell lines were generated by lentiviral transduction using recombinant lentiviral particles expressing shRNA targeted to IKKα (5’-ACA GCG TGC CATTGA TCT ATA-3’; Mission RNA Sigma-Aldrich). All cell lines were regularly tested for mycoplasma contamination.

For transient transfections, subconfluent cells were treated with a mixture of DNA and 25 kDa linear polyethylenimine (PEI; Polysciences), following a recently described protocol (21). Please refer to Table 1 for information about the plasmids.

H5pg4100 was used as the wild-type (wt) virus with a cell culture dispensable deletion (nt 28593-30471) in the E3 region (22). The mutant virus H5pm4149 contains stop codons in the E1B-55K open reading frame, preventing its expression (23). All viruses were propagated and titrated as previously described (24). For virus yield experiments, infected cells were harvested 24 and 48 hours post-infection (h p.i.), lysed by three freeze-thaw cycles, and then reinfected into cells for 24 h. Virus growth was assessed by immunofluorescence staining for the adenoviral DNA-binding protein DBP/E2A.

H1299 cells were co-transfected with a Firefly luciferase plasmid, controlled by either an NF-κB-responsive promoter (5x NF-κB-ELAM-promoter) or viral promoters from early transcription units (E1A, E1B, E2E, E2L, and E3), along with effector plasmids for 24 h ( Table 1 ). To control for transfection efficiency, a Renilla luciferase plasmid was co-transfected and used for normalization. For NF-κB promoter activation following TNFα treatment, H1299 cells were first transfected with the Firefly luciferase plasmid under the control of the NF-κB promoter. 22 h after transfection, the cells were treated with TNFα (PeproTech) for 2 h, then immediately infected at the indicated multiplicity of infection (MOI, defined as ffu/cell) for up to 24 h. For measurements, the cells were lysed and analyzed using the Dual-Luciferase Reporter Assay System (Promega), with sequential measurement of Firefly and Renilla luciferase activity.

All protein extracts were prepared using radioimmunoprecipitation assay (RIPA) lysis buffer, as previously described and protein concentration was determined (25). For immunoprecipitation, protein A-sepharose beads (Sigma-Aldrich Inc.) were coupled with 1 µg of monoclonal antibody for 1 h at 4°C (3 mg per immunoprecipitation). The antibody-coupled protein A-sepharose was then added to precleared pansorbin-sepharose extracts (50 µl per lysate; Calbiochem) and rotated for 2 h at 4°C. Proteins bound to the antibody-coupled protein A-sepharose were precipitated by centrifugation, washed three times. Immunoprecipitation and input control samples, with the input controls adjusted to the same protein concentration, were boiled in Laemmli buffer at 95°C for 3 minutes and then analyzed by immunoblotting as described (26). Please refer to Table 2 for all primary antibodies. Secondary antibodies conjugated to horseradish peroxidase (HRP) for protein detection via immunoblotting included anti-rabbit IgG, anti-mouse IgG, and anti-rat IgG (Jackson/Dianova).

E1B-55K fragments tagged with glutathione-S-transferase (GST) in pGEX vectors (27, 28) were cultured in LB medium at 37°C until the OD600 reached 0.6. Protein expression was induced by adding 0.5 mM isopropylthio-β-D-galactoside (IPTG) for 4 h. The bacterial pellets were frozen overnight at -80°C to enhance cell lysis. After thawing, the pellets were lysed in MTTB buffer (50 mM Tris, 150 mM NaCl, 1% (w/v) Triton X-100, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), aprotinin (5 mg/ml), leupeptin (10 mg/ml) and pepstatin A (1 mg/ml)). The protein lysates were incubated with GST beads for 4 h at 4°C. An aliquot of the E1B-55K-bound GST beads was boiled in 2x Laemmli loading buffer. To standardize protein concentration for the GST pulldown assay, an aliquot from each sample was analyzed via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using a bovine serum albumin (BSA) standard curve to determine protein concentration. For the pulldown assay, cell lysates were prepared with RIPA buffer as described previously. 1 mg of protein lysate was incubated with amounts of GST-protein coupled beads, adjusted based on protein concentration, overnight followed by immunoblot analysis after boiling in 2x Laemmli loading buffer. Overnight incubation with GST-protein coupled beads also occurred at 4°C.

For indirect immunofluorescence, H1299 cells were seeded on glass coverslips at a density of 1.5 × 10⁵ cells per well. At various time points, the cells were fixed with 4% paraformaldehyde (PFA) for 20 minutes at 4°C and then permeabilized with PBS containing 0.5% Triton X-100 for 5 minutes at room temperature. After a 15-minute blocking step in Tris-buffered saline-BG (TBS-BG; BG consists of 5% (w/v) BSA and 5% (w/v) glycine), the coverslips were incubated with the indicated primary antibody diluted in PBS for 30 minutes ( Table 2 ). The coverslips were then washed three times with TBS-BG, followed by a 20-minute incubation with Alexa 488 (Invitrogen) or Cy3 (Dianova)-conjugated secondary antibodies. Afterward, the coverslips were washed twice with TBS-BG and once with PBS. Finally, the coverslips were mounted using Glow medium (Energene), and digital images were captured using a confocal laser scanning microscope (Nikon). The images were cropped using Adobe Photoshop and assembled in Adobe Illustrator.

All statistical analyses were conducted using GraphPad Prism version 9. Details on the specific statistical tests used are provided in the corresponding figure legends. A p value of ≤0.05 was considered statistically significant.

It has been shown that proteins from the HAdV-C5 E3 region have immunomodulatory functions (29). Therefore, we used HAdV-C5, which does not encode the viral E3 region, to elucidate the role of HAdV-C5 proteins on NF-κB regulation during viral infection. We first investigated whether the NF-κB promoter is activated during a productive infection by HAdV-C5, comparing the activation levels to those triggered by TNFα, a representative of well-known activators of NF-κB signaling (30). The rationale behind this comparison is that TNFα, which is frequently secreted by macrophages or infected cells, binds to TNF receptors on nearby cells, thereby activating an immune response and preparing these cells to defend against potential infections. This approach allows us to assess NF-κB promoter activation in inflamed or primed cells ( Figure 1 ). To control luciferase expression of the (5x) NF-κB-ELAM-promoter, the construct was transfected, and cells were treated for 24 h with 20 ng/ml TNFα 6 h post-transfection and harvested after indicated time points before cell lysis. H1299 cells treated with 20 ng/ml TNFα induce threefold luciferase expression after 4 h, raising up to 4.5 fold 24 h post-treatment ( Figure 1A ). However, H1299 cells infected with HAdV-C5 with an MOI of 20 or 100 did not show activation at any time point after infection. NF-κB activity is even slightly reduced at 24 h p.i. at MOI 20 and 100 compared to non-infected cells ( Figures 1B, C ). These results suggest that HAdV-C5 does not activate NF-κB in a manner similar to TNFα and it can be assumed that HAdV-C5 encodes for a protein that counteracts NF-κB activity. Thus, in the early phase of infection, the virus does not significantly disrupt the TNFα-induced activation of NF-κB. However, in the later phase of infection, the virus may interfere with TNFα activity.

In order to further investigate TNFα-mediated NF-κB activation, we combined infection with TNFα treatment. As shown, TNFα activates NF-κB 4 h p.i. The activity is not affected due to adenoviral infection at different virus concentrations ( Figures 1D, E ). Intriguingly, TNFα-mediated activation of NF-κB is decreased to the level of non-treated cells 24 h p.i. ( Figures 1D, E ). Therefore, we assume that HAdV-C5 encodes for a protein that counteracts this host defense mechanism.

Activation of NF-κB in response to viral pathogens is associated with the establishment of protective immunity, however the NF-κB pathway provides an attractive target to viruses as the rapid, immediate early (IE) event results in a strong transcriptional stimulation not only for cellular, but also for several early viral genes to enhance viral replication as they harbor NF-κB binding sites in their promoter sequences (3). NF-κB-binding sites have already been revealed within the promoter regions of HAdV-C5 E2 and E3 transcriptional units (31). This is associated with activation of the NF-κB pathway early after infection, which is beneficial for HAdV-C5 gene expression and replication. Here, we also investigated whether the early promoter regions of HAdV-C5 can be activated by the NF-κB transcription factor, with the ultimate goal of determining whether E1B-55K, as an adenoviral regulatory protein, interferes with this immune pathway – similar to the regulatory proteins of other viruses. H1299 cells were transfected with constructs encoding HAdV-C5 promoter sequences downstream of a reporter luciferase gene and co-transfected with a plasmid encoding p65, a subunit of NF-κB. As illustrated, p65 activates the E1A and the E2E promoter up to 14-fold, and the E2L promoter showed a 17-fold induction ( Figure 2A ). In contrast to the E1A, E2L and E3 promoter sequences, E2E has two NF-κB binding sites as shown by Machitani et al. (31). However, we observe strong E1A and E2L activation by p65 in contrast to E1B or E3 ( Figure 2A ). The lack of E3 promoter activation in our assay is notable, as it contradicts the current literature. However, these results suggest a highly regulated association between the cellular NF-κB pathway and the onset of adenoviral gene expression.

Viruses tightly regulate the immediate early step of immune activation in order to benefit from activation or inhibition of immune determinants. To elucidate the role of adenoviral proteins on NF-κB pathway regulation, H1299 cells were co-transfected with the NF-κB promoter construct and either early or late adenoviral gene encoding constructs ( Figure 2B ). The reporter gene expression construct used in this assay encodes a luciferase gene under the control of a (5x) NF-κB-ELAM-promoter. Firefly luciferase expression directly correlates with transcriptional activation by NF-κB. Renilla-luciferase was co-transfected to determine the transfection efficiency for normalization. This experiment shows that the early protein E1A increases the NF-κB promoter to 7-fold, while E2A, E4orf6, and the late protein L4-100K activate the NF-κB promoter 2- to 4-fold compared to the control. E1B-55K, however, showed no influence on the promoter activity. Thus, E1A appears to strongly activate the NFkB promoter, in contrast to E1B-55K ( Figure 2B ).

Regulation of the NF-κB pathway by HAdVs could be maintained at several steps after infection at multiple NF-κB signaling molecules. Many viruses encode for viral proteins to target the NF-κB pathway. It has been reported that E1B-55K controls the expression of genes, which are involved in the NF-κB pathway and the regulation of interferon response genes (32). Building on these and our previous observations, we next examined the impact of HAdV-C5 infection on the formation of the IKK complex. H1299 cells were transfected with human Flag-tagged IKKα, IKKβ, NEMO plasmids and superinfected with HAdV-C5 wt at 8 h post-transfection. The western blot results show a higher migrating NEMO-specific band upon overexpression, indicating dimerization of the NEMO protein ( Figure 2C , lane 5). Our results reveal interaction between E1B-55K with Flag-tagged IKKα and Flag-tagged NEMO ( Figure 2C , lanes 3 and 5). These findings suggest a novel mechanism of NF-κB pathway modulation by the virus.

To investigate whether E1B-55K exerts an immune suppressive function by counteracting the NF-κB pathway through targeting the IKK proteins, we performed an immunofluorescence analysis. A549 cells were infected with HAdV-C5 and co-stained for E1B-55K and IKKα or NEMO. Consistent with previous publications, E1B-55K localizes within the cytoplasm and nucleus during infection (33, 34). Staining of E1B-55K reveals its nuclear localization with a few cytoplasmic nuclear membrane adjacent intensely stained bodies (perinuclear bodies) that is diffusely distributed ( Figure 2D ). Within the nucleus, E1B-55K shows mostly a granular diffuse distribution ( Figure 2D ). However, we also observed E1B-55K localization in globular structures, mostly associated globular ring-like structures ( Figure 2D ). Endogenous IKKα shows mainly cytoplasmic localization as it acts as member of the NF-κB signaling pathway within the cytoplasm, but a small proportion of IKKα is also detectable within the nucleus, likely to exert its NF-κB dependent as well as independent functions. However, IKKα shows a relocalization into the nucleus in infected cells ( Figure 2D ).

To elucidate the role of E1B-55K on IKKα localization in putative replication centers, H1299 cells were infected with an E1B-55K null mutant virus H5pm4149 and co-stained for the viral DNA binding protein (E2A) and IKKα. The early adenoviral protein E2A is a scaffold protein viral replication centers (35, 36) and serves as a control for infection ( Figure 2D ). Co-staining of IKKα and E2A reveals different localization pattern of IKKα early upon infection before constitution of viral replication centers in comparison to later time ( Figure 2D ). Early upon infection, which is indicated by a diffuse staining pattern of E2A, IKKα loses its preferred nuclear lamina localization and seems to be dispersed throughout the whole cell ( Figure 2D ). However, upon constitution of viral replication centers later upon infection, IKKα is mainly relocalized into the inter-viral replication center space of the nucleus ( Figure 2D ). Importantly, NEMO is partially relocalized to virus-induced nuclear globular compartments following adenovirus infection ( Figure 2D ). In the absence of infection, endogenous NEMO predominantly localizes in the cytoplasm. However, upon infection with HAdV-C5 wt, NEMO is partially relocalized to the nucleus, where it colocalizes with E1B-55K within the viral replication centers (as indicated by colocalization with DBP (36); Figure 2D ). In summary, infection with an E1B-55K null mutant virus reveals E1B-55K independent relocalization of IKKα into the nucleus. This result does not only indicate that further viral factors could induce relocalization of IKKα, but also that there are further regulatory roles of E1B-55K on IKKα. Since E1B-55K null mutant virus infection also induces IKK relocalization, it suggests that other HAdV proteins may play a role in this effect, which warrants further investigation.

Next, we aimed to further analyze the interaction between E1B-55K and IKK proteins through transfection experiments to determine whether this interaction occurs independently of other viral proteins. We repeated immunoprecipitation experiments with E1B-55K, co-transfecting it with each of the Flag-tagged IKK proteins in H1299 cells ( Figure 3A ). Input samples for direct immunoprecipitation of E1B-55K showed sufficient and comparable levels of both transfected and endogenous proteins. Our findings indicate that E1B-55K may interact with IKKα but not with NEMO, unlike in a full viral infection ( Figures 2C , 3A ). This suggests that additional viral proteins may be required for interaction with NEMO.

To investigate these interactions in more detail, we conducted immunofluorescence analysis. For this, H1299 cells were co-transfected similarly to the immunoprecipitation experiments, using Flag-tagged IKKα, IKKβ, NEMO, and E1B-55K. Cells were then fixed with PFA 48 hours post-transfection and visualized through double-labeled immunofluorescence microscopy ( Figure 3B ). In cells transfected only with Flag-tagged E1B-55K or IKKα, both proteins predominantly showed cytoplasmic localization. However, upon co-transfection of both proteins, E1B-55K was almost entirely relocalized diffusely within the nucleus, while Flag-tagged IKKα lost its cytoplasmic localization and became diffusely distributed throughout the entire cell ( Figure 3B ). Interestingly, upon co-transfection with E1B-55K, both IKKβ and NEMO showed complete relocalization into the nucleus and perinuclear bodies where they colocalized with E1B-55K ( Figure 3B ).

To pinpoint which regions of E1B-55K are responsible for interacting with IKKα, we conducted GST pulldown assays. Therefore, bacterial expressed GST-tagged E1B-55K wt, truncation mutants as well as isoforms of E1B-55K ( Figure 4A ) were purified and conjugated to glutathione sepharose beads. The E1B-55K mutants we used have already been extensively studied, allowing us to explore potential functions or mechanisms. Cell lysates were incubated with purified GST-tagged protein conjugated beads and an interaction was analyzed by SDS-PAGE and Western blotting. GST pulldown experiments confirm the interaction between IKKα and E1B-55K. Additionally, IKKα interacts with the isoform 156R of E1B-55K ( Figure 4A ), which shares the same N- and C-terminus of E1B-55K but misses the central part. This isoform lacks both the NES and the SUMO conjugation motif (SCM), resulting in its localization to the cytoplasm. However, like wt E1B-55K, 156R is phosphorylated at its C-terminus. These findings suggest that the central region of E1B-55K is not essential for its interaction with IKKα, but that protein folding or phosphorylation of E1B-55K may play a crucial role in enabling this interaction.

In the past, numerous efforts were made to characterize E1B-55K by generating various single-amino-acid substitution mutants, as well as multiple amino-acid substitutions, to disrupt specific motifs or PTMs such as SUMOylation or phosphorylation, resulting in the regulation of protein-protein interactions (19, 20, 26). Consequently, we investigated whether these PTMs of E1B-55K are essential for its interaction with IKKα by co-immunoprecipitation experiments. Our findings strongly indicated that all E1B-55K PTM mutants tested still bind IKKα ( Figure 4B ). However, we also we found that IKKα interaction was impaired in specific E1B-55K mutants (RF6 and E2, that have previously been reported to show defects in binding to the PML-NB factors Mre11 (E1B-55K-RF6) and Daxx (E1B-55K-E2); (37); Figure 4C ), indicating that point mutations or structural differences in these mutants, but not PTMs disrupt the interaction and suggesting a degree of flexibility in the interaction.

Viruses modulate their hosts in order to achieve a microenvironment, which promote productive viral infection. Therefore, cells have to counteract anti-viral defense measurements by either being targeted to protein degradation, subcellular localization changes or they are inactivated by modulation of their PTMs. To analyze the effect of IKKα on HAdV-C5 progeny production, lentiviral particles harboring shRNA (shIKKα) was transduced to deplete endogenous IKKα from H1299 cells prior to H5pg4100 virus infection. IKKα was efficiently depleted as shown with western blot analysis (see below). Growth curve analysis indicates, that the knockdown of IKKα did not affect the growth rate of the cells ( Figure 5A ). In comparison to the parental cell lines, which were transduced with lentiviral particles harboring scrambled shRNA, knockdown of IKKα decreased the production of infectious virus particles at 8.62-fold at 48 h p.i. and 2.76-fold at 72 h p.i. ( Figure 5B ). Our data indicate that IKKα acts as a negative regulator of HAdV-C5 progeny production.

Next, the role of IKKα on transcriptional regulation of adenoviral genes was investigated. In order to analyze the role of IKKα on viral early and late protein expression in H1299 cells, western blot analysis was made to monitor viral protein expression levels at different time points after infection. Consistent with affected HAdV-C5 progeny production, expression of particular HAdV-C5 early proteins E1A and E1B-55K are decreased in IKKα depleted cells compared to the parental cells and protein steady state levels of E1B-19K and adenoviral capsids are similarly reduced at 24 hpi, comparable to the reduction observed in E1A and E1B-55K. Moreover, the Mre11 steady state levels are higher in the parental cells compared to the shIKKα knockdown cells throughout the course of infection ( Figure 5C ). Taken together, these data indicate that IKKα is a positive regulator of HAdV-C5 replication during infection. Additionally, the data suggest that IKKα likely influences the DNA damage response through Mre11.