Supplementary Figure S1. The co-chaperones Aha1 and Hsp40 are required for SRPK1 protein folding. (A) HeLa cells treated with control siRNA or siRNA against AHA1 or HSP40 were analyzed by anti-SRPK1 immuniprecipitation followed by Western blotting. Equal amounts of SRPK1 were brought down by anti-SRPK1. Consistent with the data shown in Fig. 2, cells treated with siAHA1 and siHSP40 (Aha1 knockdown was verified by Western blotting, but the data was not shown) prevented the association of SRPK1 with Hsp90 and Hsp70, respectively. The treatment with siAHA1 also reduced SRPK1 association with Hsp70, indicating a crosstalk between the Hsp70 and Hsp90 machineries as also shown in Fig. 2. However, the effect of HSP40 knockdown on SRPK1 association with Hsp90 was not evident in this particular experiment. (B) The kinase activity immunprecipitated from the cells treated with control siRNA and siRNAs against AHA1 and HSP40. The results showed that depletion of the co-chaperones reduced the kinase activity, consistent with misfolding of the kinase in chaperone impaired cells. (C) SRPK1 distribution in different siRNA-treated cells (panel a, c, e). Nuclei were marked by DAPI staining (panel b, d, f). A minor shift of SRPK1 to the nucleus (panel c and d) was observed in siAHA1-treated cells relative to control siRNA treated cells (panel a and b) whereas little difference was observed in siHSP40-treated cells (panel e and f). This observation suggests that long-term depletion of the co-chaperones was ineffective in inducing SRPK1 nuclear translocation.

Supplementary Figure S2. Characterization of SR protein phosphorylation by phosphatases treatment. Total HeLa cell lysate was directly probed with mAb104 (lane 1) or after incubation in the CIP buffer for 60 min (lane 2). The phosphoepitope in SR proteins was diminished to some extent due to the activity of endogenous phosphatases during the incubation. Inclusion of CIP further reduced the mAb104 epitope in SR proteins, which was accompanying with an increased gel mobility shift (lane 3). Dephosphorylation was completely prevented by the inhibitor to CIP (lane 4). Total cell lysate from sorbitol-treated HeLa cells was similarly analyzed (lane 5 to 8). Key is the observation that the up-shifted SRp55 was lost upon the CIP treatment, indicating that SRp55 was hyper-phosphorylated in sorbitol-treated cells.

Supplementary Figure S3. Role of SRPKs in the control of E1A alternative splicing. RT-PCR analysis of E1A splicing is shown in the top panel and quantification of E1A pre-mRNA and individual E1A isoforms is presented in the bottom panel. Treatment of the transfected with sorbitol induced the accumulation of the E1A pre-mRNA; the 9S isoform was also induced relative to the level of 12S and 13S isoforms as shown in Fig. 6A (lane 2 and 3). Depletion of SRPK1 or SRPK2 had little effect on E1A splicing (lane 4 to 7), but depletion of both kinases caused the accumulation of the E1A pre-mRNA in cells without the sorbitol treatment (lane 8). Importantly, depletion of both kinases prevented the elevation of the 9S isoform in response to the sorbitol treatment (lane 9). Consistent data based on three replicates of the experiment in sorbitol-treated cells are presented in Fig. 6E and 6F.