Participation of both Gab1 and Gab2 in the activation of the ERK/MAPK pathway by epidermal growth factor

Abstract

Three members of Gab family docking proteins, Gab1, Gab2 and Gab3, have been identified in humans. Previous studies have found that the hepatocyte growth factor preferentially utilizes Gab1 for signalling, whereas Bcr-Abl selectively signals through Gab2. Gab1-SHP2 interaction has been shown to mediate ERK (extracellular-signal-regulated kinase) activation by EGF (epidermal growth factor). However, it was unclear whether EGF selectively utilizes Gab1 for signalling to ERK and whether Gab2 is dispensable in cells where Gab1 and Gab2 are co-expressed. Using T47D and MCF-7 human breast carcinoma cells that express endogenous Gab1 and Gab2, we examined the role of these docking proteins in EGF-induced ERK activation. It was found that EGF induced a similar amount of SHP2-Gab1 and SHP2-Gab2 complexes. Expression of either SHP2-binding defective Gab1 or Gab2 mutant blocked EGF-induced ERK activation. Down-regulation of either Gab1 or Gab2 by siRNAs (small interfering RNAs) effectively inhibited the EGF-stimulated ERK activation pathway and cell migration. Interestingly, the inhibitory effect of Gab1 siRNA could be rescued not only by expression of an exogenous mouse Gab1 but also by an exogenous human Gab2 and vice versa, but not by IRS1 (insulin receptor substrate 1). These results reveal that Gab2 plays a pivotal role in the EGF-induced ERK activation pathway and that it can complement the function of Gab1 in the EGF signalling pathway. Furthermore, Gab1 and Gab2 are critical signalling threshold proteins for ERK activation by EGF.

Figure 1. Formation of Gab1–SHP2 and Gab2–SHP2 complexes in T47D cells

Near-confluent T47D cells in 100 mm plates were serum-starved and stimulated with EGF or mock-treated with BSA for 10 min. Gab1 and Gab2 were immunoprecipitated with antibodies to Gab1 and Gab2. Immunoprecipitates were analysed by immunoblotting with antibodies to phosphotyrosine, SHP2, and Gab1 or Gab2. IP, immunoprecipitation; IB, immunoblotting; α, antibody.

Figure 2. Effects of SHP2-binding defective Gab1 and Gab2 mutants on ERK activation

(A) T47D cells in 60 mm plates were co-transfected with HA–ERK2 (0.4 μg) and Gab1FF, Gab2F or an empty vector (3.6 μg). Transfected cells were serum-starved and stimulated with EGF (5 ng/ml, 5 min) or mock-treated. An aliquot of each cell lysate supernatant (10 μg) was analysed for expression of Gab1FF and Gab2F by immunoblotting with an anti-FLAG antibody (lower panel). HA–ERK2 kinase activity was determined by an immunocomplex kinase assay using MBP as substrate. Reaction mixtures were separated by SDS/polyacrylamide gel and transferred on to a nitrocellulose membrane. After detection of MBP phosphorylation, the membrane was used for immunoblot analysis to determine the relative amount of HA–ERK2 protein in each sample. (B) COS-7 cells in 60 mm plates were co-transfected with HA–ERK2 (0.2 μg) and Gab1FF, Gab2F or an empty vector (1.8 μg) as indicated. Transfected cells were serum-starved and stimulated with EGF (0.5 ng/ml, 5 min) or mock-treated. Gab1FF and Gab2F expression was examined and ERK2 kinase activities were determined as above. Graphs represent averages and ranges of the ERK2 kinase activity from two independent experiments.

Figure 3. Specificities of Gab1 and Gab2 siRNAs

(A) Alignment of human Gab1 and Gab2 siRNA targeting nucleotide sequences with the corresponding sequences in mouse Gab1 and Gab2. (B, C) COS-7 cells were co-transfected with pGab1siRNA, pGab2siRNA or empty vector (1.8 μg) and FLAG-tagged human Gab1 (hGab1), human Gab2 (hGab2), mouse Gab1 (mGab1) or mouse Gab2 (mGab2; 0.2 μg) as indicated. Cell lysates (30 μg) were subjected to immunoblot analyses with antibodies to the FLAG-tag and β-actin after 48 h transfection.

Figure 4. Inhibition of EGF-stimulated Mek1 activation in T47D and MCF-7 cells by Gab1siRNA and Gab2siRNA

(A, D) T47D cells or MCF-7 cells in 60 mm plates were co-transfected with HA–Mek1 (0.4 μg) and pGab1siRNA, pGab2siRNA, pGab1/pGab2siRNA, or the empty vector (3.6 μg). Cells were serum-starved for 20 h after 24 h transfection, and then stimulated with EGF (5 ng/ml, 5 min) or mock-treated with BSA. HA–Mek1 was immunoprecipitated and Mek1 kinase activity was determined by phosphorylation of a kinase-defective ERK2 (KR) in the presence of [γ-³²P]ATP. After autoradiography and phosphoimaging analysis, membranes were subjected to immunoblotting analysis of HA–Mek1. The histogram in (A) represents the averages and ranges from two independent experiments. (B, C) T47D cells in 60 mm plates were transfected with HA–Mek1 (0.4 μg)+siRNA and mouse Gab expression plasmids (1.8 μg/each) or control vector (−) as indicated. EGF-induced HA–Mek1 activation was measured as in (A and D).

Figure 5. Gab1 and Gab2 are functionally equivalent in the EGF-induced ERK activation pathway

T47D cells in 60 mm plates were co-transfected with HA–Mek1 (0.4 μg)+pGab1siRNA, pGab2siRNA, human Gab1 or Gab2 expression vector, or a control plasmid as indicated. The ratio of siRNA to Gab expression plasmid was 1:1 (1.8 μg/each). Transfected cells were processed for the determination of EGF-induced HA–Mek1 activation as described in the legend to Figure 4. (A) Co-expression of human Gab2 (hGab2) with Gab1siRNA. (B) Co-expression of human Gab1 with Gab2siRNA.

Figure 6. IRS1 cannot complement Gab1 in the EGF-induced ERK activation pathway

T47D cells were co-transfected with HA–Mek1+pGab1siRNA and expression vectors for human IRS1 or mouse Gab1 as indicated. EGF-induced HA–Mek1 activation was determined as in the legend to Figure 4. The averages and ranges of Mek1 activity from two independent experiments are shown (A). (B) Autoradiograph (upper panel) and immunoblot (lower panel) of a representative experiment. (C) Cell lysates were analysed by immunoblotting with an anti-FLAG antibody for expression of IRS1 and mGab1.

Figure 7. Effects of adenoviral-based siRNAs on endogenous Gab protein levels and Mek1 activity

T47D cells were incubated with recombinant adenoviruses as indicated for 16 h and then with fresh medium for 48 h followed by serum starvation and EGF stimulation. AdGab1+2siRNA indicates a 1:1 mixture of AdGab1siRNA and AdGab2siRNA viruses (MOI 125). Aliquots of cell lysates were analysed by immunoblotting with antibodies to Gab1, Gab2 and β-actin (A). Mek1 was immunoprecipitated from cell lysates and Mek1 kinase activity was determined (B). The histogram represents data from two independent experiments performed in duplicates (n=4).

Figure 8. Effects of down-regulation of Gab1 and Gab2 on EGF-stimulated cell migration

T47D cells were infected with recombinant adenoviruses and serum-starved as mentioned in the legend to Figure 7 and used for the Transwell cell-migration assay in DMEM with EGF as the chemoattractant in the lower chamber. Migrated cells on the lower membrane surface were enumerated under a microscope (10×20 lens) in five randomly chosen fields and the average numbers of cells per field are shown. The data were from four independent experiments (n=6).

Figure 9. Comparison of total Gab proteins in adenovirus-based rescue experiment

T47D cells were incubated with recombinant adenoviruses as indicated (1=42 MOI) for 18 h. After the infection period, cells were cultured in fresh DMEM/10% FBS for another 10 h before serum-starvation in DMEM/0.1% BSA for 18 h, and then stimulated with EGF (5 ng/ml, 5 min) or mock-treated. Portions of cell lysates were subjected to immunoprecipitation with an anti-Mek1 antibody followed by the immunocomplex kinase assay. One half of each reaction mixture was used for autoradiography/phosphoimaging analysis and the other half for immunoblotting analysis of Mek1 after electrophoresis on SDS/polyacrylamide gels (B, upper panels). Aliquots of cell lysates (20 μg/each) were also used for immunoblottting analyses of Gab1, Gab2 or β-actin (A and lower panels of B). Relative activity was obtained from phosphoimager data. Relative levels of immunoreactive band intensities were measured from scanned images of X-ray films using the ImageQuant program.