Fig 2: The BioID interactome analysis and its subsequent validation show that both BMP2K splicing variants can associate with SEC16A protein and localize to the early secretory compartment.(A) Dot plots showing BioID-MS detection scores (log scale) of proteins found as proximal to both N- and C-terminally tagged BMP2K-S or -L variants in HEK293 cells. (B) Levels of BMP2K-L, BMP2K-S, EGFP or EGFP-tagged SEC16A in whole cell lysates (INPUTS) or in immunoprecipitates (IP) using anti-GFP antibodies from HEK293 cells. Different combinations of simultaneous ectopic expression of the analyzed proteins are indicated above the images. Graph shows non-normalized densitometric analysis of western blotting bands expressed in arbitrary units (a.u.; n = 3 +/- SEM). (C) Maximum intensity projection images from confocal microscope showing localization of ectopically expressed EGFP-tagged BMP2K-L or -S with respect to the indicated proteins and cell nuclei marked with DAPI stain (blue) in K562 cells. Insets: Magnified views of boxed regions in the main images. Scale bars, 10 µm. *p<0.05, **p<0.01.Figure 2—source data 1.Numerical data for graphs in Figure 2.

Fig 3: BMP2K splicing variants differentially regulate SEC16A protein levels and distribution and control the abundance of SEC31A-positive structures.(A) Western blots showing the effect of shRNA-mediated depletion of single (shBMP2K-L or shBMP2K-S) or all BMP2K splicing variants (shBMP2K) on the levels of SEC16A and SEC24B proteins, as compared to empty pLKO vector or non-targeting shRNA construct, shCtr. Graphs show densitometric analysis of western blotting bands for the indicated proteins using tubulin abundance for normalization (n = 5 +/- SEM). (B) SEC16A mRNA fold levels in control cells or in cells with shRNA-mediated depletion of BMP2K variants (n = 5 +/- SEM). (C) Representative maximum intensity projection images from confocal microscope, showing the effect of shRNA-mediated depletion of BMP2K variants on immunolocalization of SEC16A and SEC31A proteins in K562 cells. Cell nuclei marked with DAPI stain (blue). Insets: Magnified views of boxed regions in the main images. Scale bar, 10 µm. (D and E) The number of juxtanuclear and dispersed SEC31A-positive vesicular structures per cell (D) or SEC31A mean fluorescence intensity per vesicular structure (SEC31A protein load) presented in arbitrary units (a.u. in E) in control cells or cells lacking BMP2K variants. Quantification from images represented by those in C (n = 5 +/- SEM). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.Figure 3—source data 1.Numerical data for graphs in Figure 3.

Fig 4: (A) Graphical comparison of amino acid sequences adjacent to Thr156 in µ2 and Thr102 in NUMB (both known to be phosphorylated in human cells by AAK1, a BMP2K homologue) and Thr1045 in SEC16A, that we identify as a candidate BMP2K phosphorylation site.The color code indicates that two upstream residues of SEC16A Thr1045 are identical to the respective residues in µ2, while two downstream amino acids are identical to those in NUMB. (B) Alignment of ~60 amino acids surrounding Thr1045 in human SEC16A and in its homologues from selected vertebrates. Green color indicates residues shown in A, grey color indicates a hydrophobic residue (f) five amino acids upstream of the putative phosphorylation site. (C) Western blots showing the effect of increasing concentrations of LP935509, AAK1/BMP2K inhibitor, on the levels of endogenous SEC16A and phosphorylated µ2 in K562 cells. (D) Western blots showing the effects of control shRNA (Ctr), shBMP2K-L (L) or shBMP2K-S (S) on the levels of ectopically expressed EGFP-SEC16A either WT or with Thr1045 mutated to Ala (TA) or Glu (TE) in K562 cells. (E) Western blots showing the effects of the LP inhibitor on the levels of ectopically expressed WT or mutant EGFP-SEC16A in K562 cells.

Fig 5: (A) Representative western blot showing the efficiency of SEC16A depletion using two single shRNAs.(B) Percentage of benzidine-positive control K562 cells or cells depleted of SEC16A, under basal growth conditions or after stimulation for 48 h with 20 µM hemin (n=4 for basal or n=3 for hemin +/- SEM). (C) Fold changes in mRNA levels of the indicated erythroid markers in control cells or in cells depleted of SEC16A (n=3 +/- SEM).