Figure 1. Biosynthetic profile of transient versus stable FGFR3 mutants. (a) and (b) HEK293 cells were transiently and stably transfected with the HA-tagged wt and SADDAN FGFR3. 48 h later the cells were metabolically labeled with [³⁵S]Met/Cys for 2 h, lysed and immunoprecipitated (IP) with anti-HA antibodies, followed by autoradiography. Samples were treated (+) or not (−) with EndoH. The different receptor glyco-isoforms are indicated. (c) TDII and SADDAN receptors show a restored biosynthesis in HEK293 stable cell lines. Cell lysates were immunoprecipitated with anti-HA antibodies and analyzed by immunoblot (IB) with anti-HA and anti-PhTyr antibodies. (d) Cell surface biotinylation of the SADDAN–FGFR3. wt and SADDAN cell clones were biotinylated for 10 min. The biotinylated proteins were recovered by streptavidin-agarose while the intracellular isoforms (un-biotinylated) were recovered by subsequent immunoprecipitation with anti-HA antibodies. The mature 130 kDa forms carrying the SADDAN mutation were exposed on the cell surface.

Figure 2. The TDII and SADDAN receptors constitutively activate the FRS2α-dependent pathway from the cell surface. HEK293 cells transiently and stably transfected with the TDII and SADDAN mutants were lysed and immunoprecipitated with antibodies recognizing FRS2α, SHP-2 (a) and Grb2 (c). The indicated antibodies were used to immunoprecipitate a higher amount (2 mg) of cell lysate from clone TDII cl77 to better visualize the FRS2/SHP-2/Grb2 complex (b). The same antibodies were used in the immunoblot, by stripping and re-probing the same filters. Activated FRS2α was detected with anti-pTyr antibodies or anti-Phospho FRS2. The mutant receptors were capable to activate the ternary complex FRS2α/SHP-2/Grb2 only in stable cell lines. (d) RCS and HEK293 cells treated with FGF1 for 10 min show the typical shift in molecular mass that characterizes FRS2α activation.

Figure 3. Erks can be activated from both the ER and the cell surface. (a) wt15, TDII77 and SADDAN3 stable clones were analyzed for Erks activation in the presence (+) or absence (−) of 3 μM monensin for 5 h. 20 μg of whole cell extract were loaded in each lane and an immunoblot with anti-Ph-Erk was performed. Monensin treatment abolished Erks activation by mutant FGFR3. (b) The SADDAN receptor from the stable cell clone 3 treated or not with monensin was immunoprecipitated and analyzed by immunoblot following treatment or not with EndoH and PNGase F. Monensin causes accumulation of intracellular receptor forms. (c) Erks activation in HEK293 cells treated with FGF1. (d) Erks activation by different FGFR3 mutants transiently transfected in HEK293 cells. 1 mg of total protein was immunoprecipitated with anti-Erk antibodies from the same lysates used in the experiment of Figure 5(b). The SADDAN-754F mutant retains the ability to activate the Erks from the ER. (e) and (f) Erks activation by the SADDAN FGFR3 (transient expression) is abrogated by a 2.5 h treatment with PD98059, PP2 and SU5402 chemical inhibitors.

Figure 4. Transfection kinetic time course. Erk1/2 phosphorylation over a basal level is detected at 10 h following transfection. A similar level of activated Erks is present at 24 and 48 h. The empty pRcCMV vector is transfected as a negative control. A similar amount of total Erks protein is shown. In the lower panel an immunoprecipitate with HA shows an increasing amount of expressed FGFR3.

Figure 5. TDII and SADDAN receptors recruit PLCγ and Pyk2 from the ER. (a) HEK293 transfected with the wt, wt-754F, SADDAN and SADDAN-754F were lysed and immunoprecipitated with anti-PLCγ antibodies. The immunocomplexes were analyzed with anti-PLCγ and anti-HA antibodies as indicated. PLCγ can directly bind the SADDAN but not the SADDAN-754F receptor. (b) PLCγ is detected performing immunoprecipitation with HA followed by immunoblot with PLCγ. (c) Whole cell extracts prepared from the same set of transfections were analyzed with anti-Ph-PLCγ antibodies. (d) The indicated lysates were immunoprecipitated with anti-Pyk2 antibodies. The immunocomplexes were deglycosylated with PNGaseF and analyzed by immunoblot with antibodies detecting Pyk2, HA and Ph.Tyr (4G10) by stripping and re-probing the same filter. The same amount of cell lysates were immunoprecipitatedwith anti-HA antibodies followed by immunoblot with HA antibodies to show the level of expression of the different receptors (bottom panel).

Figure 6. The KD mutation restores the SADDAN receptor biosynthesis and abolishes signaling from the ER. HEK293 cells were transfected with wt-KD and SADDAN-KD receptors and metabolically labeled. (a) The receptors were immunoprecipitated with anti-HA antibodies and treated with EndoH and PNGaseF. (b) Pulse–chase experiment showing that the wt-KD and SADDAN-KD present a similar turnover.

Figure 7. Cartoon model describing signaling by mutant FGFR3. Left, the highly activated K644E/M FGFR3 mutants are trapped in the ER in their immature phosphorylated forms in transient transfection. Mutants FGFR3 recruit PLCγ, Pyk2 and JAK1. The Erk1/2 activation occurs through a Src family member through an FRS2α and PLCγ-independent pathway. Right, stable clones present a different scenario: both immature and mature phosphorylated FGFR3 are present in the ER and cell surface, respectively. The mature receptors signal from the cell surface in the absence of ligand through the canonical FRS2α-dependent pathway, ending in Erk1/2 activation. Erk1/2 and STAT1 are not activated from the ER. Transient and stable cells are colored differently because the cellular environment is different under the two conditions.