WRN depletion results in decreased levels of de-novo protein synthesis
Our previous work suggested that the metabolic changes induced by WRN depletion in cancer cells are caused by a posttranscriptional mechanism of deregulation(25), and we confirmed by RT-qPCR that mRNA levels of the metabolic enzymes G6PD and IDH1 are not downregulated after shWRN induction in HeLa cancer cells (Supplementary Fig. 1A). To gain an understanding of the processes that are responsible for the metabolic changes observed in cancer cells upon WRN depletion(25), we measured de-novo protein synthesis by metabolic pulse labeling experiments with [35S]-methionine/cysteine (35S-met/cys) in HeLa cells after the induction of shWRN or control shRNAs. We analyzed the cells after three days of shRNA induction since we have previously shown that this is the earliest time point with more than 80% depletion of WRN. The extracted proteins were resolved by polyacrylamide-gel electrophoresis (SDS-PAGE) and visualized by Coomassie stain. The dried gels were then exposed to a phosphorimaging screen to visualize the de-novo synthesized radiolabeled proteins. The results of these experiments show a significant decline in radiolabeled proteins in WRN-depleted cells (~43%; p<0.001) as compared to the control cells (Fig. 1A and B), indicating that loss of WRN negatively affects de-novo protein synthesis.
To specifically determine whether the decrease in the levels of the metabolic enzyme G6PD after WRN depletion is the result of reduced rates of protein synthesis, the extracts prepared from pulse-labeled WRN-depleted and control cells were subjected to immunoprecipitation (IP) with G6PD antibody. The immunoprecipitated products were resolved on an SDS-PAGE and Coomassie stain was used to visualize the immunoglobulin heavy chain, as a means to assure that equal amounts of antibody were added to each reaction. The level of radiolabeled G6PD was measured by exposing the gel to a phosphorimaging screen. The radiolabeled met/cys incorporated in de-novo synthesized G6PD in WRN-depleted HeLa cells was plotted relative to control cells and normalized to Coomassie signal intensity of the antibody heavy chain. The results of these experiments show a significant decrease in the levels of radiolabeled G6PD in WRN depleted cells as compared to the control cells (Fig. 1C), indicating that reduced translation of G6PD mRNA is likely contributing to the reduced levels of G6PD protein observed in these cells(25).
Since the results shown in Fig. 1A suggest that WRN depletion has a more general effect on de-novo protein synthesis, we then performed immunoprecipitation reactions using antibodies against two arbitrarily selected proteins, tubulin and Ku70. These experiments show reduced levels of de-novo synthesized tubulin and Ku70 in WRN depleted cells (Fig. 1C and D), a result that is consistent with WRN depletion having a rather general effect on protein biosynthesis. To rule out that reduced levels of de-novo protein synthesis is the result of a stress response, we analyzed stress granules (SGs) formation in WRN-depleted and control cells by immunofluorescence microscopy using an antibody against the RNA-binding protein TIAR. In response to environmental stress, including oxidative conditions, TIAR accumulates into the cytoplasm and aggregates at SGs(36-38). The formation of SGs is closely linked to the inhibition of translation initiation(39). The results show that TIAR aggregation is not detected after WRN depletion (Supplementary Fig. 1B and C). Importantly, the formation of SGs is not inherently inhibited by WRN depletion, since discrete cytoplasmic foci are formed in WRN depleted treated with arsenite, a known inducer of SGs(40) (Supplementary Fig. 1C).
We have previously shown that DNA damage and the ensuing response occur at later time points (approximately 5 days) after WRN depletion than the downregulation of metabolic enzymes(25, 41) and confirmed that three days after WRN knockdown there is no significant increase in the levels of the phosphorylated form of g-H2AX, an established marker of DNA breaks (Supplementary Fig. 1D and E)(17, 42), in HeLa cells. Collectively, these results rule out that activation of a stress response pathways is the cause of attenuated translation after WRN depletion.
WRN depletion does not impact the levels of ribosomal proteins (RPs) nor ribosomal RNA (rRNA)
Ribosomes are complex structures composed of two main components, the 40S, and 60S subunits. Both subunits are organized by ribosomal RNA (rRNA), proteins, and accessories factors(43). The large subunit (60S) consists of the 28S, 5S and 5.8S rRNAs and 47 proteins, while the small subunit (40S) has a single 18S rRNA and 33 proteins (44-46). It has been reported that WS fibroblasts show decreased levels of rRNA transcription compared with wild-type cells, which was reversed by ectopic expression of wild-type WRN(47). Since this defect could directly affect protein synthesis, we wanted to determine if attenuated translation is the result of changes in the abundance of rRNAs in mature ribosomes after WRN depletion in HeLa cells. Polysome extracts from WRN depleted and control cells were prepared by sucrose gradient differential centrifugation of purified cytoplasmic fractions (Fig. 2A and B, and Supplementary Fig. 2A and B). After RNAs purification, quantitative RT-qPCR analysis was performed using primers sets for the 28S, 18S, and 5.8S rRNAs. The results of this experiment do not show any significant difference in rRNA abundance between WRN depleted and control HeLa cells (Fig. 2C). This result suggests that attenuated protein synthesis during the early phase of WRN depletion is not the result of altered rRNA levels in mature ribosomes of HeLa cells.
Next, we analyzed the levels of a subset of ribosomal proteins (RPs) and their cognate mRNAs in WRN depleted and control HeLa whole cell lysates. We performed semi-quantitative analysis of a set of ribosomal proteins using Lamin A/C as a normalizing protein(48) and did not observed any detectable decrease in the expression levels of the analyzed RPs (Fig. 2D). Rather, WRN-depleted cells show a significant increase in the signal intensity of these RPs (Fig. 2E), possibly indicating a compensation mechanism in response to decreased translation. Analysis by qPCR did not show any difference in the steady-state levels of RPs mRNAs between WRN depleted and control samples (Fig. 2F), suggesting that the observed increase in protein levels might be the result of increased proteins stability. Taken together, these data indicate that the decrease in de-novo protein synthesis observed in WRN depleted cells is not due to lower abundance of ribosomal components of the translational machinery.
WRN depletion affects the nuclear/cytoplasmic distribution of mRNAs
The aforementioned experiments led us to then test whether the dysfunction in protein synthesis may be caused by alteration in mRNA nucleocytoplasmic transport. mRNA export is a critical step in promoting gene expression and upregulation of the mRNA export factors has been observed in many types of cancer(22, 23, 49). We isolated nuclear and cytoplasmic fractions from equal number of WRN depleted and control HeLa cells and measured the mRNA present in each fraction by RT-qPCR. Prior to running the qPCR reactions, each fraction was analyzed by SDS-PAGE and immunoblotted using nuclear and cytoplasmic protein markers to ensure no cross-contamination between these two compartments (Fig. 3A). We performed qPCR reactions using primer sets for actin, tubulin, G6PD and IDH1 and observed altered nuclear to cytoplasmic mRNA ratio for all these transcripts in WRN depleted cells when compared to the control (Fig. 3B). To assess the generality of our observation and rule out potential effects caused by the shRNA silencing system used in this analysis, we compared nuclear/cytoplasmic ratios of the same set of mRNAs between parental WS fibroblasts and WS fibroblasts reconstituted with Flag-tagged WRN (F-WRN). The results of this experiment show that the cells reconstituted with Flag-WRN display a lower nuclear/cytoplasmic mRNA ratio compared to the parental cells (Fig. 3C and D), indicating that this process is also affected in normal cells. However, because of the lower proliferation rates of fibroblasts, the effects on these cells are likely to manifest after many more cell divisions. Consistent with these data, the alterations in the nuclear/cytoplasmic mRNA ratio associated with WRN depletion in HeLa cells were in part reversed after removal of doxycycline, which allows re-expression of WRN (Fig. 3E and F).
To confirm the alteration in the mRNA nuclear/cytoplasmic distribution, control and WRN-depleted HeLa cells were analyzed by mRNA Fluorescent In Situ Hybridization (FISH) using a conjugated-Cy3 Oligo(dT) probe. Three days after doxycycline treatment the cells were fixed, incubated with the Oligo(dT) probe and analyzed by confocal microscopy. A fraction of the cells was used to confirm WRN depletion by immunoblot (Fig. 4A). The signal intensity of the Oligo(dT) Cy3-conjugated probe was measured in the cell-fixed images by tracing a pixel-fixed line across the entire cell and the nuclear/cytoplasmic boundary was determined at the intersection of the DAPI and Cy3 channel (Fig. 4B and C). After measuring the individual intensities in the Cy3 channel, the nuclear/cytoplasmic ratio was calculated for the WRN-depleted and control cells. This analysis shows a significant increase in the nuclear/cytoplasmic ratio of the poly (A)+ signal in WRN depleted cells when compared to the controls (Fig. 4D), reinforcing the conclusion that WRN depletion affects the subcellular distribution of mRNAs in HeLa cells.
The protein levels of components of the mRNA export pathway are not affected by WRN depletion in HeLa cells
Our data suggested that WRN depletion affects the mRNA export pathway. Therefore, we determined whether changes in the levels of the export receptors could be responsible for these alterations. We first measured the levels of the two major nuclear export factors NXF1 and CRM1 using semi-quantitative Western blot analysis. The results of this experiment show no significant difference in the levels of these two proteins between control and WRN-knockdown HeLa cells, suggesting that alterations in mRNA export after WRN depletion are unlikely caused by lower levels of either one of these proteins (Fig. 5A). Next, we examined the Transcription and Export 1 (TREX-1) factor. TREX-1 is a conserved multiprotein complex that plays a critical role in mRNPs biogenesis and maturation in eukaryotes(50-55). This large complex links processing and export of mRNAs and provides a surveillance platform for maintaining the high-fidelity of the gene expression(56). Several studies have shown that inhibition of TREX-1 components results in the accumulation of mRNPs in the nucleus(51, 57). To determine whether alteration in the mRNA spatial distribution resulting from WRN depletion may be caused by reduced levels of the TREX-1 complex, we performed immunoblotting using antibodies against factors within the THO complex subcomplexes THOC1 and THOC2, including the adaptor protein ALYREF. In this analysis we also examined the levels of other factors linked to mRNA metabolism, including UAP56 and CBP80, two proteins implicated in the splicing and capping of the mRNAs, respectively, GANP (Germinal center–Associated Nuclear Protein), a protein that is actively involved in recruitment and transport of mRNPs and whose depletion causes nuclear accumulation of Poly (A)+ RNA(58), and eIF4E, a eukaryotic translation initiation factor that has been shown to facilitate nuclear export of specific transcripts(59, 60). The results of this experiment show that depletion of WRN does not result in a decrease in the levels of any of these proteins as compared to the control cells (Fig. 5B-D). Interestingly, as we observed for the RPs, some of these adaptor proteins are slightly upregulated, possibly as a compensatory response to the reduced nuclear export of mRNA. Taken together, these results rule out deficiencies in components of export machinery or mRNA processing as an early response to WRN depletion and suggest that a different mechanism is likely responsible for reduced levels of de-novo protein synthesis in WRN depleted HeLa cells.
WRN associates with mRNA through direct interaction with the mRNA export receptor NXF1
Since the mRNA export receptors NXF1 and CRM1 play key roles in nuclear export of RNA, we reasoned that WRN might aid these nuclear receptors during the export of mRNPs by directly interacting with mRNA and/or the export receptor. To test this hypothesis, we first examined whether WRN associates with mRNAs using Oligo(dT) pull-down. For this purpose, we isolated whole cell extracts from HeLa cells and divided them into two aliquots. One aliquot was treated with RNase A (DNase-Free) and the other with Ribonucleoside Vanadyl Complex (RVC) to protect from endogenous ribonucleases. Each fraction was incubated with Oligo(dT) beads and after extensive washes the proteins were eluted from the beads, resolved on SDS-PAGE and analyzed by immunoblotting using antibodies against the nuclear export receptor NXF1, WRN and the RNA-binding protein CBP80 as a control. An antibody against tubulin was used as a negative control. Both NXF1 and the mRNA binding proteins CBP80, but not tubulin, were efficiently pulled down only in the extracts that were not treated with RNase A (Fig. 6A and B), confirming the specificity of the assay to capture RNA-binding proteins. Remarkably, WRN was detected in the Oligo(dT) bound material from the RVC treated extract, but absent from the pull-down of the RNase A treated sample, indicating that pull-down of WRN is dependent on the presence of undamaged mRNA (Fig. 6B). This finding demonstrates a novel link between WRN and mRNA. Next, we examined whether WRN binds directly to the export receptors. For this purpose, whole cell extracts from HeLa cells were subjected to co-immunoprecipitation (Co-IPs) assays using antibodies against NXF1 or CRM1. The results of this experiment show that WRN co-precipitates with NXF1 but not CRM1 (Fig. 6C). Significantly, this interaction is not mediated by RNA, since it is resistant to Benzonase, which degrades nucleic acids (Fig. 6D), suggesting a direct and specific protein-protein interaction between WRN and NXF1.