The RNA-binding domain and PLD of HNRNPD are required for SNB formation. (A) Schematics of all the HNRNPD constructs used in the rest of the figure. Four isoforms (p37, p40, p42, and p45) are shown above. p45 is the standard isoform construct corresponding to HNRNPD shown in Fig. 2 A. The alternatively selected exons 2 and 7 are shown in blue and red, respectively. Schematics of two RRM mutants (RRM1-M and RRM2-M), in which two phenylalanine residues in each of RRM1 and RRM2 (pale green and green boxes, respectively) were substituted with aspartic acid (F140D and F142D for RRM1-M and F225D and F227D for RRM2-M), are shown in the middle. The schematics of PLD mutants are shown at the bottom. Nine or 18 of the 18 tyrosine residues present in the PLD (yellow box) were mutated to serine residues (Y-S 9/18 and Y-S 18/18) to generate PLD-M1 and PLD-M2, respectively. The PLD mutation resulted in the cytoplasmic mislocalization of the mutant HNRNPDs; therefore, a SV40-NLS was inserted at the N terminus (orange box). (B) Isoform-specific localization of HNRNPD in SNBs. The Venus-tagged HNRNPD isoforms shown on the left were transfected and their localizations were detected. Sam68 was detected as the marker of SNBs. (C) RNA binding of HNRNPD through RRM1 and RRM2 is required for its SNB localization. Venus-tagged p45 and two RRM mutants were transfected, and their localizations were detected as in B. (D) The PLD is required for the SNB localization of HNRNPD. The FLAG-tagged HNRNPD isoforms were transfected, and their localizations were detected as in B. The cell populations (%) in which Venus or FLAG and Sam68 signals overlap are shown in the merge panels in B, C, and D (>100 cells, n = 3). (E) The RRMs and PLD are required for SNB formation. Rescue of the defect in SNB formation by the HNRNPD mutant constructs shown in A. The HNRNPD constructs were transfected into HeLa cells in which endogenous HNRNPD had been depleted by RNAi, and then SNB-positive cells (Sam68 foci–positive cells) were counted (>100 cells, ±SD, n = 3). As a negative control, the FLAG-EGFP plasmid was transfected (EGFP). (F) The PLD is required for protein–protein interactions of HNRNPD. The FLAG-tagged PLD mutants used in D were immunoprecipitated to detect the interaction with Sam68 and cotransfected Venus-tagged HNRNPD (p45). GAPDH was used as the input control. The molecular mass marker (kD) is shown on the left. Bars, 10 µm.