Fig 2: Model. Following cleavage and polyadenylation of the transcript, the CID of PCF11 is transiently phosphorylated by WNK1. This weakens the interaction between the PCF11 CID and the Pol II CTD, leading to the dissociation and release of mRNP from transcription loci. As a result, nuclear export of mRNA is promoted.

Fig 3: WNK1 participates in mRNA export. (A) Fluorescent oligo-dT FISH for the detection of mRNA levels in control and WNK1-depleted cells. Alexa fluor 488-dT(23) was used as a probe, and DAPI was used for DNA staining. (B) Quantitation of the percentage of nuclear-localized mRNA over total. The integrated intensity of the nuclear fluorescent signal and the total fluorescent signal were calculated for every cell (n = 30; data are from two biological repeats), and the ratio between the two is presented graphically. Error bars represent SEM. (C) Heat map representing the log2 ratio of nuclear over cytoplasmic (Nuc/Cyt) mRNA in control and WNK1-depleted cells for the top 20% of expressed genes. The darker shade indicates an increased Nuc/Cyt ratio. (D) Quantification of the fold change of Nuc/Cyt mRNA levels between siLUC and siWNK1 cells in the top 20% of expressed genes. Most genes in this set show an increase of Nuc/Cyt ratio upon WNK1 depletion.

Fig 4: WNK1 is present in the nucleus and interacts with PCF11. (A) Co-IP experiments of PCF11 and WNK1. Immunoprecipitation of PCF11 (top panel) and WNK1 (bottom panel). Lane 1 corresponds to 5% input. Lanes 2–4 show mock immunoprecipitation using rabbit IgG (negative control). Lanes 2 and 5 show immunoprecipitation with 150 mM NaCl, and lanes 3 and 6 show immunoprecipitation with 250 mM NaCl. Lane 7 shows immunoprecipitation with 150 Mm NaCl after incubation with benzonase to test for nucleic acid dependence. (B) Superresolution microscopy and detection of WNK1 kinase. Pol II staining was used as a control for nuclear localization. The same pattern was observed in all imaged cells. n = 20. Data are from two biological repeats. The same laser intensity was used for all images. (C) Cytoplasmic and nuclear levels of MYC and TBP mRNA in control and WNK1-depleted cells. Data are from biological repeats. Error bars represent standard error of mean (SEM). Data were normalized to siRNA targeting luciferase (siLUC). (D) U6 RNA levels measured with RT-qPCR in nuclear and cytoplasmic extracts.

Fig 5: WNK1 phosphorylates the PCF11 CID. (A, top) Schematic representation of endogenous PCF11 phosphorylation sites identified in this study (Supplemental Table 2). (Bottom) Schematic representation of the recombinant GST-PCF11 CID. (B) Autoradiography of a recombinant WNK1 kinase domain in vitro kinase assay on the GST-PCF11 CID. MBP was used as a positive control for WNK1 phosphorylation. The upper band corresponds to autophosphorylated WNK1. The lower bands likely correspond to truncated products of the GST-PCF11 CID. (C) Autoradiography of a WNK1 in vitro kinase assay on the wild-type CID and S120A–T121A CID. Coomassie staining shows equal levels of GST-CID loading for wild type and the mutant. (D) Quantitation of kinase assays with the wild-type and S120A–T121A CID. Phosphorylation levels were measured in four independent repeats of the assay. Error bars represent SEM. (E) PhosTag Western blot for the detection of endogenous PCF11 migration upon WNK1 depletion. The same cell extracts treated with λ phosphatase were used as controls. (F) PhosTag Western blots comparing migration of wild-type PCF11-HA and S120A–T121A PCF11-HA in LUC and WNK1-depleted cells. The control lane shows the migration of an unspecific band recognized by the αHA antibody, which provides a positional and loading control.