SARS-CoV-2 viral protein Nsp2 stimulates translation under normal and hypoxic conditions

To investigate the interaction between Nsp2 and eIF4E2 we first generated a cell line expressing SARS-CoV-2 protein Nsp2 containing an N-terminal Streptavidin tag. As a control cell line, we used HEK293T cells transfected with empty vector. We selected puromycin resistant Nsp2- and control cells in a similar manner. Western blot analysis confirmed expression of Strep-Nsp2 protein in the SARS-CoV-2 Nsp2 cells (Fig. 1A). Next, we evaluated interaction of Nsp2 and eIF4E2. Cell lysates containing equal amount of total protein from control and Nsp2-transfected cells were incubated with either Biotin-Agarose (Fig. 1B) or with m⁷GTP-Sepharose (Fig. 1C). With both affinity matrices, we could confirm the interaction between Nsp2 and eIF4E2 as the two proteins were enriched upon copurification only from extracts of cells expressing Nsp2.

Previous studies have shown that Nsp2 viral protein localizes in the cytoplasm of cells [44]. To investigate whether Nsp2 co-localizes with eIF4E2 we performed immune-fluorescent staining of controls and Nsp2-expressing cells grown under normal or hypoxic conditions (Fig. 2A). We observed colocalization of Nsp2 and eIF4E2 in cell cytoplasm under both conditions. Next, we investigated whether Nsp2 present at the sites of protein synthesis. We observed colocalization of Nsp2 with the rough endoplasmic reticulum marker, calnexin (Fig. 2B) and with a small ribosomal subunit protein S3 (RbS3) (Fig. 2C). These data indicate that Nsp2 co-localizes with eIF4E2 at dense granular cytoplasmic structures that are also sites of active protein synthesis, and thus presumed to be RER. Interestingly, we observed increased co-localization of Nsp2 with eIF4E2, calnexin, and RbS3 under hypoxic conditions. It was reported that the synthesis of proteins targeted for export/secretion (RER-bound transcripts) are highly increased upon hypoxic conditions [45].

To investigate whether Nsp2 affect protein synthesis, we monitored expression of Renilla and Firefly luciferases in cells grown under normal and hypoxic conditions (1% oxygen) for 12–18 h. Control and Nsp2-cell were transiently transfected with a [pCMV Luc Renilla-HCV IRES-Firefly luciferase] plasmid that expresses Renilla in a cap-dependent manner and Firefly luciferase in an IRES-dependent manner (cap-Ren-HCV-FF). We observed that Nsp2 cells have higher expression levels of both Renilla and Firefly luciferases under normoxia and hypoxia (Fig. 3 and Additional file 1: Fig. S2). To evaluate whether observed increase in translation in Nsp2-cells is due to more efficient transfection or due to stabilization of mRNA we performed real-time PCR to measure the levels of GAPDH, Renilla and Firefly mRNAs. We found that GAPDH mRNA levels were similar between control and Nsp2-cells in both normoxia and hypoxia. Ratio of GAPDH mRNA in control versus Nsp2-cells was 0.98 ± 0.02 (n = 6) under normoxia and 0.99 ± 0.01 (n = 6) under hypoxia. To analyze the level of Renilla and Firefly mRNA, we first normalized luciferase signal to the GAPDH in the corresponding sample. Then we have measured the ratio of luciferase mRNA (normalized to GAPDH) to the average of corresponding luciferase mRNA in control cells. We found that the expression level of Renilla mRNA was similar between control and Nsp2 cells in normoxia and hypoxia: 1 ± 0.13 in control cells (n = 3) and 0.9 ± 0.32 in Nsp2-cells (n = 3) under normoxia; and 1 ± 0.26 in control cells (n = 3); and 1 ± 0.37 in Nsp2-cells (n = 3) under hypoxia. We have also confirmed that the ratio of Renilla mRNA to Firefly mRNA is similar in control and Nsp2-cells cells in normoxia and hypoxia: 1.1 ± 0.11 in control cells (n = 3) and 0.9 ± 0.12 in Nsp2 cells (n = 3) under normoxia; and 1.2 ± 0.06 in control cells (n = 3) and 0.99 ± 0.098 in Nsp2-cells (n = 3) under hypoxia. These data suggest that Nsp2 expression does not affect transfection efficiency, level of expression, or stability of mRNAs driven from the [pCMV Luc Renilla-HCV IRES-Firefly luciferase] plasmid under normoxia or hypoxia.

To evaluate the rate of protein synthesis (PS) we monitored the rate of H³-Leu incorporation in the control and Nsp2-cell lines grown under normoxic conditions. We determined that there is statistically significant greater PS rate in the Nsp2-expressing cells compared to that in control cells (Fig. 3C, D). We did not measure the rate of PS under hypoxic conditions due to the technical difficulties to do such experiment in the hypoxic chamber because of the difficulty of adding the label after hypoxia is reached, and then for removing the time aliquots. Increased rate of PS in Nsp2-cells indicate that Nsp2 expression affects the rate of but does not distinguish between cap- and IRES-dependent translation.

To investigate how expression of Nsp2 affects translation of endogenous mRNAs we performed polysomal analysis of cells grown under normal and hypoxic conditions. First, we observed increase in 80S peak in control cells grown under hypoxia compared to that in normoxia confirming inhibition of initiation step of translation (Fig. 4A). At the same time, under both conditions, Nsp2 cells had lower 80S ribosomal peak compared to that in control cells, suggesting more efficient initiation step of translation. To evaluate the effect of hypoxic conditions on the distribution of ribosomes along sucrose density gradients, we isolated RNA from each fraction of the sucrose density gradients and performed real-time PCR assay using specific primers for 18S ribosomal RNA. We then calculated a ratio between 18S RNA co-sedimented with polyribosomes (fractions #7–11) and monosome (fractions #4 and 5) (Fig. 4A, Ps/Ms). We observed that while in control cells Ps/Ms ratio decreased from 3.8 ± 0.2 under normoxia to 2.3 ± 0.3 under hypoxia, in Nsp2-cells the ratio Ps/Ms stayed higher under both conditions: 4.2 ± 0.4 and 4.3 ± 0.3 under normoxia and hypoxia, respectively. We also calculated a ratio between 18S RNA co-sedimented with heavy polyribosomes, containing efficiently translated mRNAs (fractions #7–11), and light polyribosomes, containing inefficiently translated mRNAs (fractions #1–6) (Fig. 4A, H/L). We observed that in control cells, the H/L number decreased from 0.7 ± 0.1 observed under normal conditions to 0.5 ± 0.1 under hypoxia suggesting enrichment of 18S in light polysomal fractions, in agreement with inhibition of initiation step of translation. In contrast, in Nsp2 cells, the 18S H/L number did not differ between normoxia and hypoxia, 1.0 ± 0.2 and 1.1 ± 0.1, respectively. Higher Ps/Ms and H/L ratios in Nsp2-cells suggest a more efficient translation under both normal conditions and hypoxia compared to the control cells.

To investigate the effect of Nsp2 expression on translation of the mRNAs that are specifically regulated at the initiation step of translation, we monitored the distribution of FGF2 and VEGF-C mRNAs in the sucrose density gradients. The FGF2 mRNA has a G-C rich 5′ highly structured region with a calculated stability of 50 kcal/mole and is known to be translationally stimulated by high level of eIF4E and eIF4F [46]. Similarly, VEGF-C mRNA was shown to be translationally stimulated in the presence of high eIF4E level [47]; [48]. As a control, we also monitored distribution of GAPDH mRNA along sucrose density gradient, since it encodes a house-keeping protein and is not translationally regulated under hypoxia. First, we confirmed that neither Nsp2 expression nor hypoxia affect the total amount of GAPDH, FGF2, and VEGF-C mRNAs (data not shown). In Fig. 4B, graphs on the left, the amount of individual mRNA in each fraction is presented as a percentage of the amount of this mRNA in all 11 fractions. To analyze the amount of individual mRNA in the light (L, fractions #1–6) and heavy (H, fractions #7–11) polysomal fractions, we expressed it as a percentage of mRNA sedimented in all fractions set as 100% (Fig. 4B, bars on the right). We did not observe any changes in GAPDH mRNA distribution along the sucrose density gradient in both, control and Nsp2 cells under normal and hypoxic conditions (Fig. 4B, upper panels). Statistical analysis showed that about 90% of the GAPDH mRNA co-sedimented with polyribosomal complexes (fractions #7–11) in all samples. FGF2 mRNA has similar polyribosomal distribution between control and Nsp2 cells under normal conditions (58 ± 6% vs 66 ± 6% in H fractions, respectively). However, under hypoxia, the percent of FGF2 mRNA in Nsp2 cells slightly increased in heavy polysomal fractions (75 ± 8%) compared to that in control cells (61 ± 10%) (Fig. 4B, middle panels) suggesting translational upregulation. Hypoxia increased the amount of VEGF-C mRNA in heavy polysomal fractions in both, control (75 ± 6% in hypoxia vs 53 ± 5% in normoxia) and Nsp2 (83 ± 11% in hypoxia vs 60 ± 5% in normoxia) cells (Fig. 4B, lower panels). These data suggest that VEGF-C mRNA is more efficiently translated under hypoxic conditions in control and Nsp2 cells. To confirm translational activation of FGF2 and VEGF-C mRNAs we monitored the FGF2 and VEGF-C protein expression and compared it to the GAPDH level (Fig. 5). The FGF2 mRNA has a G-C rich 5′ highly structured region containing at least two translation initiation sites. Translation from the cap-proximal CUG1 of the mammalian FGF2 mRNA was shown to be stimulated by excess eIF4E/eIF4F in vitro and in vivo [46, 49]. In contrast, translation from the internal AUG (~ 400 nt downstream from it) is believed to occur via an IRES-driven mechanism [50], particularly under hypoxic conditions, although there is also some evidence that under presence of excess eIF4F, it may be translated via leaky-scanning upon melting of the highly G/C-rich secondary structure of the ~ 500 nt-long 5′UTR [46, 49]. Although most evidence for this complex mechanism of translation initiation at alternative CUG and AUG start codons is attributed to features in the 5′UTR, one should not disregard the demonstrated potential regulation via the ~ 6000 nt long 3′UTR [49, 51]. We found that in hypoxic conditions, the translation from CUG1 was increased in both control and Nsp2 expressing cells (Fig. 5A, left image). Under milder hypoxia, translation from the CUG start codon was highly stimulated in the Nsp2 but not in control cells (Fig. 5A, right image and quantified from independent experiments in Fig. 5B). In agreement with the VEGF-C mRNA shift to the heavy polysomes under hypoxia, we observed increased VEGF-C protein levels in control and Nsp2 cells under hypoxic conditions (Fig. 5C).

To investigate whether Nsp2 expression affects composition of translation initiation complexes we first analyzed proteins co-sedimented with 40S ribosomal peak in the sucrose density gradients. According to the polysomal profiles, fractions #2 contain 40S (Fig. 4A) but lack a biomarker of large 60S ribosomal subunit, RsL7a (Fig. 6). This suggest that proteins in fractions #2 may be a part of 48S initiation complexes. Western blot analysis revealed that under normal conditions, both control and Nsp2 cells contain eIF4G and eIF4A in their fractions #2 of polysomal gradient confirming the presence of initiation complex eIF4F (Fig. 6A). However, control cells also had eIF4E2 and 4EBP in fraction #2 suggesting the presence of some inhibitory complexes. Hypoxia increased the 4EBP signal in fractions #2 in both control and Nsp2 cells but at a lesser extent in the Nsp2 cells (Fig. 6B). These data suggest that Nsp2 expression reduces the inhibitory effects of eIF4E2 and/or 4EBP on translation. Indeed, Nsp2 cells exhibit stronger eIF4G signal along polysomal fractions in both normal and hypoxic conditions compared to control cells suggesting more efficient initiation of translation (Fig. 6, eIF4G). Interestingly, we detected Nsp2 signal in the fractions #2 under normal and at a lesser amount in hypoxic conditions. This result suggests that Nsp2 protein may be a part of complexes that regulate initiation step of translation.

First, we confirmed that our hypoxic conditions induce the biomarkers of hypoxia by probing lysates with anti-Hif-1α antibodies. Indeed, we observed increase in Hif-1α expression in both control and Nsp2-cells under hypoxia (Fig. 7A). Hypoxia is also known to downregulate protein synthesis largely through changes in mTOR and eIF2α activities [43, 45]. Indeed, mTOR phosphorylation at S2448 (a marker of its activity) was suppressed under hypoxia, in both controls and Nsp2-expressing cells (Fig. 7A). Consequently, there was a loss of 4EBP phosphorylation, both seen by P-S65 and increase of the high mobility form after hypoxic exposure. Although the ratio of P-4EBP to 4EBP was decreased in both control and Nsp2 cells, overall Nsp2 cells maintained a higher ratio of P-4EBP to 4EBP (Fig. 7A, graph on a right). This is expected to be reflected in the association of 4EBP with eIF4E, thereby releasing eIF4E for its recruitment in the eIF4F complex.

To investigate the composition of proteins associated with the cap-binding complexes, we performed a pull-down experiment using m⁷GTP-Sepharose. To avoid contribution of RNA into the complexes retained on m⁷GTP-Sepharose we incubated some samples with RNase A. Graphs on the Fig. 7 represents results of three independent experiments. As we expected, we observed lower level of 4EBP associated with m⁷GTP-Sepharose (meaning decreased binding with eIF4E [1] that would correspond in its greater association with eIF4G as seen from their ratio) in Nsp2-cells compared to that in control cells under both, normoxic and hypoxic conditions regardless of RNase A treatment (Fig. 7B; normalization for proteins input in Additional file 1: Fig. S3). This observation can explain in part the retention of active, cap-dependent, protein synthesis under hypoxia in cells expressing Nsp2. The alternative explanation involving the activity of eIF2α seems less likely since it was hyperphosphorylated equally in control and Nsp2 cells after hypoxia (Fig. 7A). Instead, we noticed that eIF4E2 had a lower retention on m⁷GTP-Sepharose when Nsp2 is expressed, suggesting that Nsp2 interferes with its cap-binding capacity (Figs. 7B and 1C). This is an intriguing observation, suggesting that even though Nsp2 would have no direct effect on the eIF4F complex formation and retention on m⁷GTP-Sepharose, under conditions where eIF4F and eIF4E2/RBM4 compete for binding to the pool of capped-mRNAs, the presence of Nsp2 could shift the overall equilibrium towards eIF4F binding to mRNAs and promote translation. We should add that we have no evidence confirming the Uniacke’s reports that eIF4E2 promotes translation of select mRNAs containing RBM4 binding sites under hypoxic conditions, leading for instance to increased EGFR expression [42] (see Fig. 5C). Our evidence is rather consistent with the previous reports that eIF4E2 acts as a competitive inhibitor of eIF4E and thus translation. In fact, depletion of eIF4E2 with siRNA in Hela cells was reported to result in a 30% increase in protein synthesis rates [33], and we obtained similar results with the lung epithelial carcinoma cells A549 (Additional file 1: Fig. S1) confirming its general inhibitory mechanism even when applied to the more relevant lung tissue in the context of SARS-CoV-2 NSP2-eIF4E2 interaction.

We are aware that our study has some limitations, as it's hard to conduct Virology work without viruses and extrapolate results from overexpression studies. Here at LSUHS, we are still several months away from having our certified BSL3 unit ready, but our work can possibly help similar current projects in this area, and at least provide a rationale for using Ribavirin in clinical trials. Certainly, a 25% increase in protein synthesis by Nsp2 is far from insignificant, since typically protein synthesis consumes over 40% of the ATP in the cells, so, such an increase (particularly during the hypoxia deficit) could make a huge difference for the virus replication strategy.

Human embryonic kidney 293T (HEK293T) cells were cultured in Dulbecco’s modified eagle medium (DMEM) supplemented with 10% FBS and 1% antibiotic–antimycotic and maintained in humidified incubator at 37 °C with 5% CO₂. To generate Nsp2-cell line, HEK293T cells were transfected with pLVX-EF1α-IRES-Nsp2 mammalian construct (a kind gift from Dr. David Gordon, University of California, San Francisco) by Lipofectamine LTX and Plus™ reagent following the manufacturer’s protocol. To generate control cells line, HEK293T cells were transfected with pDRGFP plasmid in which GFP is translated out of frame and does not produce a functional protein. Both control and Nsp2-cells were selected by 1 µg/ml of puromycin treatment for at least seven days. After the selection, cells were passaged, and early passages of cells were stored at − 80 °C. Cells underwent to selection by 0.1 µg/ml of puromycin after thawing for one more time and then grown without puromycin prior to the experiments.

For immunofluorescence microscopy, sterile coverslips were placed into the wells of a 6-well plate and 0.3 × 10⁶ Nsp2 cells/well seeded 24 h prior to hypoxia (1%O₂) or normoxia treatment in 5%CO₂ incubator at 37 °C. Cells treated for 12 h, and media removed. PBS wash twice. Cells fixed afterwards with 4%PFA in PBS at room temperature (RT) for 10 min. PBS wash done twice. Permeabilization was done using 0.5% Triton in PBS at room temperature for 10 min. PBS wash done twice. Blocking was done using PBS + 1%BSA + 0.1%Tween20 for 1 h at RT. Primary antibody (anti-strep, 1:500; anti-calnexin, 1:500; anti-eIF4E2, 1:100; anti-RbS3, 1:100) diluted in blocking buffer and added to samples. Incubated for overnight at 4 °C. After one wash with PBS, secondary antibody (anti-rabbit and anti-mouse) against primary antibody was used for 1 h at RT. DAPI in anti-fade mounting media added on pre-cleaned slides and coverslips mounted. Fluorescence was detected and imaged using Nikon A1R-Super Resolution microscope equipped with a 60× oil objective lens (Apo 60x/1.40NA) with a numerical aperture of 1.4, two GaAsP detectors for laser 488 nm and 561 nm and a standard detector for 405 nm (DAPI). Immunofluorescence images were taken using Nikon NIS-Elements C software at room temperature with 55 optical sections (n = 3) separated by 0.1 μm step-size. For representation of eIF4E2 and Nsp2 (main targets) colocalization was obtained for max intensity projection of 10 optical sections. For Calnexin and RbS3, best focal plane has been shown for representation.

Control and Nsp2-expressing HEK293T cells were lysed in 0.3 ml of lysis buffer containing 50 mM Tris–Cl (pH 7.5), 100 mM KCl, 0.5% NP-), 2 mM DTT, and EDTA free Halt™ Protease and Phosphatase Inhibitor Cocktail (100X—ThermoFisher Scientific). After sonication, cells were incubated on ice for 10 min and centrifuged at 14,000×g for 10 min at 4 °C followed by the collection of supernatants and measurement of protein concentration by BCA Protein Assay Kit ThermoFisher Scientific. 200 µg of lysate was used for both m⁷GTP-Sepharose and Biotin-Agarose (Pharamcia Biotech) pull down experiments, which were diluted with equal volume of dilution buffer containing 50 mM Tris–Cl (pH 7.5), 2 mM DTT, and 5% glycerol. 200 µl of m⁷GTP-Sepharose and Biotin-Agarose (i.e., 50% slurry) were equilibrated in one ml of wash buffer containing 50 mM Tris–Cl (pH 7.5), 100 mM KCl, 2 mM DTT, and 5% glycerol for three times. Afterwards, all wash buffer was removed, and equal volume of dilution buffer was added to the resin. The diluted lysates from each experimental group were mixed with 50 µl of resin and incubated for 2 h at 4 °C with rotation. After the incubation, the resin was spun down and the supernatant was removed and saved as unbound fraction. After washing three times with 150 µl of wash buffer, proteins were eluted by resuspending the resins in 40 µl of 2X Laemmli (for m⁷GTP-Sepharose) or 4X Laemmli sample buffer (for Biotin-Agarose). The samples were heated at boiling temperature for 6 min and further analyzed by western blotting.

One hundred μg of lysates were prepared as described in the “Ribosome Fractionation” section and diluted in equal volume of buffer containing 50 mM Tris–HCl (pH 7.5) and 2 mM DTT. Then samples were mixed with 50 μl m⁷GTP-Sepharose, 50% slurry in buffer containing 20 mM Tris–HCl (pH 7.5), 100 mM KCl, 1 mM DTT, and 10% (v/v) glycerol. The resins were incubated for 2 h at 4 °C with rotation. Duplicate samples were treated with RNase A for the last 40 min of incubation. Then, the resin was washed three times with 150-μl aliquots of the same buffer. Proteins were eluted in 20 μl 2 × SDS-electrophoresis buffer and analyzed by Western blotting as it is described above. This experiment was repeated three times.

For Fig. 1, samples from the pull-down experiments or cell lysates were run in either 7.5% or 12% SDS-PAGE gel and transferred to a PVDF) membrane. After blocking in 5% nonfat dry milk, the membranes were incubated in primary antibodies overnight at 4 °C followed by the incubation in HRP-conjugated secondary antibodies for 1 h at room temperature. The blots were developed using ECL substrates and imaged in ChemiDoc Imaging System (Bio-Rad). Densitometric quantification of the blots were done using ImageJ software. For Figs. 5C and 7, to analyze the expression of proteins in cell lysates, equal amounts of total protein were loaded on a 12% or 4–12% NuPAGE® Novex® Bis–Tris Gel. The FullRange Rainbow protein molecular weight marker was loaded on the same gel to identify the position of specific proteins. Proteins were separated by SDS-PAGE gel and then transferred to a Nitrocellulose membrane using a Mini Trans-Blot cell. Expression of specific proteins in total lysates were determined by probing the membrane with antibodies against P-mTOR (Ser2448) (dilution 1:5000), mTOR (dilution 1:5000), P-4EBP (Ser65) (dilution 1:5000), P-eIF2α (Ser51) (dilution 1:2000), and eIF2α (D7D3) XP (dilution 1:4000), all from Cell Signaling; 4EBP (dilution 1:4000), eIF4E2 (dilution 1:1000), and Strep-Nsp2 (dilution 1:2000) all from Invitrogen; actin (dilution 1:4000) from Sigma; VEGF-C (dilution 1:1000) from R & D systems; and GAPDH (dilution 1:5000) from Fitzgerald. The membranes were incubated with primary antibodies in 5% BSA in buffer TBS-T (20 mM Tris–HCl, 150 mM NaCl, and 0.1% Tween 20, pH 7.5) overnight at 4 °C, washed three times for 15 min with TBS-T, and incubated for 1 h at room temperature with anti-mouse secondary antibodies or anti-rabbit secondary antibodies conjugated with horse-peroxidase in 5% non-fat dry milk in TBS-T. Blots were developed with the Western Lightning ECL Pro development kit and exposed to HyBlot CL autoradiography film. Quantitative analysis of Western blot images was performed using the ImageJ software. Results are presented as the mean of two independent treatments using different cell passages. Student’s t test was applied to the data to determine statistical significance, and data with p value lower than 0.05 was considered to be statistically different.

The control and Nsp2 cells were cultured in DMEM supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. Cells were plated in to a 96-well white plate at 1 × 10⁴ cells per well and allowed to adhere overnight. Next day, cells were transfected with 0.1 µg DNA, pCMV Luc Renilla-HCV IRES-Firefly Luciferase, Lipofectamine LTX and Plus™ reagent following the manufacturer’s protocol. After 9 h of transfection, cells were washed, supplemented with fresh media and grown for another 12–18 h either under normoxia or hypoxia (1% O₂, 5% CO₂, 37 °C). Expression of Renilla and Firefly luciferases was detected using Dual-Glo® kit Luciferase Assay system according to the manufacture protocol. All Renilla luciferase signals (from capped mRNA) have been normalized to the mean value of Renilla in the control cells grown under normal conditions, set as 1. The experiment was repeated two times. All Firefly signals (from HCV-IRES) have been normalized to the mean value of Firefly luciferase in the control cells grown under normal conditions, set as 1. Student’s t test was applied to the data to determine statistical significance, and data with p value lower than 0.05 was considered to be statistically different.

10⁵ cells (control and Nsp2) were each plated in 12 wells of a 24-well plate with 1 ml complete D-MEM and 10% FCS (duplicate samples). The next day the medium was replaced with medium containing 5 µCi/ml L-[3,4,5-3H(N)]-Leucine (150 Ci/mmol—NEN), and sequential aliquots were removed at 1/2 h intervals starting 15 min later. After solubilization with 0.5 ml 1%SDS, 10% TCA insoluble material (0.1 mg Protein) was collected on 2.5 cm GFA filter. Following washing with 90% EtOH, the filters were air-dried and placed in scintillation vials for counting with OptiScint LLT NPE-Free Scintillation cocktail in a Beckman LS6500 counter.

The control and Nsp2 cells were cultured in DMEM supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. Cells were plated in to a 10 cm plates and allowed to adhere overnight. Next day, cells were either continued to grow under normal conditions or transferred to a hypoxic chamber (1% O₂, 5% CO₂, 37 °C) for another 12–18 h. Prior harvesting, cells were incubated with 50 μg/ml Cycloheximide for 10 min and then washed two times with warm PBS and one time with warm Ps wash buffer [50 mM Tris–HCl, pH 8.0; 140 mM NaCl; 5 mM KCl, 6 mM MgCl₂], both buffers supplemented 50 μg/ml Cycloheximide. Cells were lysed with 0.3 ml per plate ice cold Lysis buffer [100 mM Tris–HCl, pH 8.0; 150 mM KCl, 10 mM MgCl₂, 200 mM Sucrose; 5 mM DTT, 0.1 mg/ml Cycloheximide, 0.5 mg/ml Heparin, 0.5% Triton x-100, 0.5% NP-40, 0.5% Deoxycholate, 1 × protease inhibitor EDTA-free, 1 × phosphatase inhibitor 2, and 1 × phosphatase inhibitor 3]. Lysates were centrifugated at 10,000×g for 10 min at 4 °C. Supernatants were aliquoted for polysomal and western blot analyses and stored at − 80 °C until further use. Lysates of control and Nsp2 cells grown under normal or hypoxic conditions were analyzed by the polysomal ultracentrifugation. An equal amount of A₂₅₄ optical units was layered onto a 15–45% (w/v) sucrose gradient containing 50 μg/ml cycloheximide and 1 mM DTT and centrifuged in a Beckman SW41Ti rotor at 38,000 rpm at 4 °C for 2 h. Gradients were collected in 1-ml fractions with continuous monitoring of absorbance at 254 nm using an Isco syringe pump with UV-6 detector (Teledyne Isco Inc.). Samples were stored at − 80 °C until further use. The polysomal analysis was repeated three times using lysates from different cellular passages. To analyze proteins along sucrose density gradients the 10 μl aliquots from each fraction were loaded on 4–12% NuPAGE® Novex® Bis–Tris Gel, and analyzed by the Western Blotting as it is described above.

Before RNA isolation, 300 μl aliquots from each fraction after polysomal ultracentrifugation in sucrose density gradients were spiked with 100 pg of Luciferase mRNA (internal control, Promega). Then, RNA was purified with TRIzol®-LS reagent according to the manufacturer’s protocol. The RNA was further precipitated with 0.8 M Na-acetate and 1.2 M NaCl, re-suspended in RNase-free water and precipitated again with 2 M LiCl overnight at − 20 °C. Amplification and detection were performed using the iCycler IQ Real-time PCR detection system with Luna® universal One-Step RT-qPCR kit. Quantitative real-time PCR was used to measure the Firefly luciferase, 18S, GAPDH, FGF2, and VEGF-C RNAs level in each fraction. The 18S, GAPDH, FGF2, and VEGF-C RNAs levels were normalized with the luciferase internal control. Relative amount of individual RNA in each fraction (after normalization to Luciferase signal) was expressed as a percentage of the sum of this RNA in all 11 fractions set as 100%. To assist statistical significance of the changes in the RNA redistribution along the sucrose density gradients, the percentage of individual RNA co-sedimented with light polyribosomes, containing inefficiently translated mRNAs (fractions #1–6) and heavy polyribosomes, containing efficiently translated mRNAs (fractions #7–11), was calculated as a percentage of the total mRNA. The percentage of individual mRNA in polysomal fractions was investigated in three polysomal analyses using lysates from different cellular passages.

Expression of specific proteins in polysomal profiles (Fig. 6) was determined by probing the membrane with antibodies against RsL7a (dilution 1:4000) from Cell Signaling; 4EBP (dilution 1:4000), eIF4E2 (dilution 1:1000), and Strep-Nsp2 (dilution 1:2000) all from Invitrogen; mouse monoclonal anti-eIF4A antibody (dilution 1:8000) was a gift from Dr. Hans Trachsel, Bern, Switzerland; and rabbit anti-eIF4G antibody (dilution 1:20.000) was a gift from Dr. Robert E. Rhoads, LSUHSC. The 10 μl aliquots from each fraction after polysomal ultracentrifugation in sucrose density gradients were loaded on 4–12% NuPAGE® Novex® Bis–Tris Gel, and analyzed by the Western Blotting as it is described above.

Statistical analyses were performed using GRAPH-PAD PRISM 9 and MICROSOFT EXCEL software. Data are presented as mean ± standard error of the mean (SEM). Statistical significance was determined by 2-tailed Student’s t-test when comparing the mean between two groups, or by one-way ANOVA followed by Tukey’s post hoc analysis when comparing more than two groups. P-values < 0.05 were considered significant (Additional file 2: Table S1).

Key Resources Table