Fig 2: PABP redistributes to the cell membrane after re-adhesion in dependency of ADAM15.(A) Immunoblot of enriched plasma membrane fractions isolated by sequential (ultra-) centrifugation, showing ADAM15 and integrin α5 as membrane marker proteins, and also detection of PABP and tubulin as well as calnexin, a marker for ER membranes; myc, marker for nucleus, was negative. (B) Immunoblot of PFA-fixed (5% for 5 min), cell surface biotinylated plasma membrane proteins isolated using streptavidin-conjugated magnetic beads, detecting ADAM15 and integrin α5, but not calnexin and PABP. WCL = whole cell lysate. (C, D) Cells were re-adhered to a chamberslide for various time points, fixed with PFA (1% for 5 min), and cell surface biotinylated membrane proteins isolated by streptavidin magnetic beads as well as whole cell lysates prior to purification on the magnetic beads were subjected to immunoblotting using anti-PABP antibodies. Blots were reprobed using tubulin and ADAM15 antibodies as loading controls. (C) In isolated cell membranes of synovial fibroblasts and HeLa cells, detection of increased amounts of PABP at the plasma membrane upon adhesion for 30–120 min and (D, left panel) increased amount of PABP at the plasma membrane after 60 min adhesion in chondrocytes transfected with full-length ADAM15 (A15) in comparison to vector control-transfected cells (-) or (D, right panel) cells transfected with a deletion mutant of ADAM15 lacking the cytoplasmic tail (ΔC). Contr: non biotinylated cell lysates enriched on streptavidin magnetic beads served as a background control.

Fig 3: ADAM15 binds to the proline-rich domain of PABP.(A) Domain structure of PABP. (B) left, SDS/PAGE (10%) of recombinant expressed GST-tagged PABP domains and the cytoplasmic domain of ADAM15; right, SDS/PAGE (14%) of the purified cytoplasmic domain of ADAM15 and GST protein. (C) binding of the various GST-tagged PABP domains to cytoplasmic domain of ADAM15 immobilized to a 96-well, showing strongest binding of the proline-rich domain 392–553, and no binding to the RRM domain 1–192 and the C-terminal PABC domain 553–636. (D) binding of PABP fragment 392–553 in concentrations from 0–500 nM to immobilized ADAM15 revealed still binding above an OD of 0.2 in low nM concentration (100nM), as compared to the binding of PABP 553–663, which showed no binding at a concentration of 500 nM.

Fig 4: Puromycin-terminated proteins are detected at the cell membrane in the presence of ADAM15.(A) chondrocytes transfected with full-length ADAM15 (+) and the deletion mutant of ADAM15 lacking the cytoplasmic domain (ΔC) were detached and re-adhered for 45 and 60 min, fed with puromycin, crosslinked with PFA and cell surface biotinylated. The streptavidin purified membrane proteins as well as whole cell lysates were analyzed by immunoblotting using anti-puromycin IgG, showing markedly more puromycin-conjugated protein fragments in the plasma membrane of ADAM15-transfected cells after 45 and 60 minutes of adhesion (marked by arrows). Additional co-incubation with cycloheximide (CHX), showing nearly complete inhibition of protein translation at the plasma membrane. The blots were cut in half and reprobed using either ADAM15 or tubulin antibodies to check for the expression of full-length and deletion mutant ADAM15 or for equal loading, respectively. (B) HeLa cells were subjected to the same experimental procedures as described in (A), but prior to that the ADAM15 protein expression was downregulated by two specific siRNAs (I and II) and compared to a nonspecific control siRNA (N) or transfection reagent (T), showing stronger staining of puromycin protein conjugates in ADAM15 expressing HeLa cells after 60 min adhesion when compared to ADAM15 silenced cells or normally grown cells (0 min). One half of the membrane was reprobed using anti ADAM15 IgG showing a silencing efficiency of ~80%, the other half using tubulin IgG as a loading control.

Fig 5: ADAM15 co-immunoprecipitates with poly(A) binding protein 1 (PABP).(A) Co-immunoprecipitations (IP) of chondrocyte lysates transfected with a vector control (V) or full-length ADAM15 (A15) using anti-ADAM15 antibodies were separated by SDS/PAGE. A specifically precipitated protein at 75 kDa in ADAM15- cells was identified as poly(A) binding protein 1 (arrow). (B, C) IPs using anti-ADAM15 or anti-PABP antibodies of V- and ADAM15-transfected cells and subsequent immunoblot (IB) using anti-PABP or vice versa ADAM15 antibodies confirmed a specific precipitation of PABP in the ADAM15-transfected cells only. As controls, the blots were reprobed with ADAM15 or PABP antibodies, respectively. (B) IPs of cell lysates incubated with in lysis buffer ± RNase I (5 U/ml) overnight using anti-ADAM15 antibodies, showing ADAM15/PABP binding in the ADAM15-transfected cells only, which is independent of mRNA potentially bound to PABP. (D) IPs of A15-lysates and a deletion mutant of ADAM15 lacking the cytoplasmic tail (A15 ?cyto) using 2 different anti-ADAM15 antibodies (a, c, mouse and b, d, goat) specifically precipitates PABP in lysates that expressed full-length ADAM15 only. As control, the blot was reprobed using anti-ADAM15 antibodies, showing full-length ADAM15 (~100 kDa) and A15 ?cyto (~ 90 kDa).