Insulin-like growth factor I controls a mutually exclusive association of RACK1 with protein phosphatase 2A and beta1 integrin to promote cell migration

Abstract

The WD repeat scaffolding protein RACK1 can mediate integration of the insulin-like growth factor I receptor (IGF-IR) and integrin signaling in transformed cells. To address the mechanism of RACK1 function, we searched for regulatory proteins that associate with RACK1 in an IGF-I-dependent manner. The serine threonine phosphatase protein phosphatase 2A (PP2A) was found associated with RACK1 in serum-starved cells, and it dissociated immediately upon stimulation with IGF-I. This dissociation of PP2A from RACK1 and an IGF-I-mediated decrease in cellular PP2A activity did not occur in cells expressing either the serine 1248 or tyrosine 1250/1251 mutants of the IGF-IR that do not interact with RACK1. Recombinant RACK1 could bind to PP2A in vitro and restore phosphatase activity to PP2A from IGF-I-stimulated cells. Ligation of integrins with fibronectin or Matrigel was sufficient to facilitate IGF-I-mediated dissociation of PP2A from RACK1 and also to recruit beta1 integrin as PP2A dissociated. By using TAT-fused N-terminal and C-terminal deletion mutants of RACK1, we determined that both PP2A and beta1 integrin interact in the C terminus of RACK1 within WD repeats 4 to 7. This suggests that integrin ligation displaces PP2A from RACK1. MCF-7 cells overexpressing RACK1 exhibited enhanced motility, which could be reversed by the PP2A inhibitor okadaic acid. Small interfering RNA-mediated suppression of RACK1 also decreased the migratory capacity of DU145 cells. Taken together, our findings indicate that RACK1 enhances IGF-I-mediated cell migration through its ability to exclusively associate with either beta1 integrin or PP2A in a complex at the IGF-IR.

FIG. 1.

RACK1 interacts with PP2A and dissociates upon IGF stimulation. MCF-7, PC-3, or DU145 cells were starved and stimulated with IGF-I for the indicated times. Cell lysates were immunoprecipitated with anti-RACK1 antibody, followed by Western blotting with anti-PP2A C antibody. The presence of RACK1 in the immunoprecipitate was confirmed using anti-RACK1 antibody on the same blot. (A) MCF-7 cells. (B) PC-3 and DU145 cells. (C) Purified PP2A (0.5 μg) was incubated with column-bound MBP or MBP-RACK1. After the proteins were washed extensively, they were eluted with maltose and assessed by Western blotting for associated PP2A. MCF-7 cell lysate was a control for Western blotting. The results from two separate preparations of MBP-RACK1 are shown along with a Coomassie-stained gel of the bacterial crude lysate from which MBP and MBP-RACK1 were affinity purified.

FIG. 2.

The association of RACK1 with IGF-IR regulates the association of RACK1 with PP2A. (A) R⁻ cells were transiently transfected with plasmids encoding full-length WT IGF-IR or the Y1250/1251F and S1248A mutants. IGF-IR was immunoprecipitated, followed by Western blotting with anti-RACK1 antibody to detect associated endogenous RACK1. The blot was reprobed with anti-IGF-IR antibody to demonstrate expression of IGF-IR. (B) WT, Y1250/1251F, or S1248A IGF-IR was transiently expressed into R⁻ cells. Endogenous RACK1 was immunoprecipitated with anti-RACK1 antibody, followed by Western blotting with anti-PP2A C antibody. The membrane was reprobed with anti-RACK1 antibody to confirm that comparable levels of RACK1 were immunoprecipitated in each sample. A lysate control from the experiment was included, showing expression of each transfected IGF-IR. (C, left panel) MCF-7 cells were starved for 4 h and stimulated with IGF-I. Cell lysates were prepared and PP2A immunoprecipitated prior to measurement of PP2A activity, using p-NPP as a substrate. Okadaic acid (OA) (1 nM) was added where indicated, 1 h prior to cell lysis. Data are presented from triplicate samples, where PP2A activity in cells not stimulated with IGF-I is represented as 100%. (C, right panel) R⁻ cells transiently transfected with plasmids encoding full-length IGF-IR or the Y1250/1251F mutant IGF-IR were serum starved for 4 h and stimulated with IGF-I for the indicated times. Cell lysates were prepared and PP2A immunoprecipitated prior to assay of PP2A activity. Data from triplicate samples are presented, with 100% activity representing that measured in cells expressing WT IGF-IR not stimulated with IGF-I.

FIG. 3.

Adhesion signals are required for IGF-I-mediated dissociation of PP2A from RACK1. (A) Cell lysates were prepared from MCF-7 cells that were serum starved under adherent conditions or while maintained in nonadherent cultures for 4 h. The cells were then stimulated with IGF-I, and RACK1 was immunoprecipitated, followed by Western blotting with anti-PP2A C antibody. The membrane was reprobed with anti-RACK1 antibody. (B) Cell lysates were prepared from adherent MCF-7 cells that had been serum starved for 4 h and from nonadherent MCF-7 cells that had been starved and stimulated with IGF-I. PP2A was immunoprecipitated and assayed for activity, using p-NPP as a substrate. Data are presented for triplicate samples, where 100% activity represents that measured in adherent cells not stimulated with IGF-I. (C and D) Cell lysates were prepared from MCF-7 cells that were serum starved for 4 h or that were serum starved in nonadherent cultures for 4 h in the presence of (C) fibronectin (5 μg/ml) or (D) Matrigel (5%) prior to stimulation with IGF-I. RACK1 was immunoprecipitated, followed by Western blot analysis to detect associated β1 integrin or PP2A. The blot was also probed with anti-RACK1 antibody.

FIG. 4.

Both PP2A and β1 integrin interact with WD repeats 4 to 7 of RACK1. (A) Model illustrating the N-terminal fragment of RACK1, amino acids 1 to 180 (RACK1ΔC), or the C-terminal fragment of RACK1, amino acids 138 to 317 (RACK1ΔN), both of which have WD4 and can bind endogenous RACK1. (B) MCF-7 cells were transduced with the TAT fusion proteins TAT-RACK1, RACK1ΔC, and RACK1ΔN for 2 h at 37°C. TAT-RACK1 proteins were immunoprecipitated from cells by using anti-HA antibodies and probed by Western blotting for associated endogenous RACK1. (C) Immunoprecipitated TAT-RACK1, RACK1ΔC, and RACK1ΔN were investigated for association with PP2A and β1 integrin by Western blotting. (D) MCF-7 cells were transduced with the TAT fusion proteins, serum starved for 3 h, and stimulated with IGF-I for the indicated times. TAT-RACK1, RACK1ΔC, and RACK1ΔN were immunoprecipitated from cell lysates by using anti-HA antibodies and then investigated for associated PP2A and β1 integrin by Western blotting. FL, full length.

FIG. 5.

RACK1 restores PP2A activity from IGF-I-stimulated cells in vitro. RACK1 was cleaved from MBP, using factor Xa. PP2A was immunoprecipitated from serum-starved or IGF-I-stimulated MCF-7 cells and either cleaved RACK1 or MBP was incubated with bead-bound immune complexes or with purified PP2A for 45 min prior to assay of PP2A activity. Data are presented from triplicate samples, where PP2A activity in MCF-7 cells not stimulated with IGF-I is represented as 100%. A sample from MCF-7 cells where okadaic acid (OA) was added prior to cell lysis was also included as a control.

FIG. 6.

RACK1 enhances MCF-7 cell motility and migration toward IGF-I. (A) MCF-7/Neo and MCF-7/HARACK1 cells were grown to 60% confluence in medium supplemented with 10% fetal calf serum. A wound was scored in each culture and cells allowed to migrate for 24 h. Cells were then stained with Giemsa and photographed at a magnification of ×10. A representative of multiple similar fields is presented for each condition. (B) The migratory potentials of MCF-7/ Neo and MCF-7/HARACK1 cells were compared in Transwell assays. Cells were plated at 50,000 cells/chamber and allowed to migrate toward IGF-I for 24 h. Cells were fixed by using methanol before they were stained with crystal violet and examined under a magnification of ×10. Expression of RACK1 in each cell line tested compared with that in MCF-7 cells is shown in Western blots in right-hand panels. (C) MCF-7/Neo and MCF-7/HARACK1 cells were starved and stimulated with IGF-I. Cell lysates were immunoprecipitated with anti-β1 integrin antibody, followed by Western blotting with anti-RACK1 antibody. The presence of IGF-IR in the immunoprecipitate was confirmed by using anti-IGF-IR antibody on the same blot. A lysate control is included to compare levels of RACK1 in the cells. (D) IGF-IR was immunoprecipitated from MCF-7, MCF-7/Neo, and MCF-7/HARACK1 cells, and RACK1 was detected in the immunoprecipitate by probing the blot with anti-RACK1 antibody. (E) Lysates from MCF-7/Neo or MCF-7/RACK1 cells were assessed by Western blotting for FAK phosphorylation by using anti-phospho-FAK (Y397) antibody. The blot was stripped and reprobed for FAK and actin as loading controls.

FIG. 6.

RACK1 enhances MCF-7 cell motility and migration toward IGF-I. (A) MCF-7/Neo and MCF-7/HARACK1 cells were grown to 60% confluence in medium supplemented with 10% fetal calf serum. A wound was scored in each culture and cells allowed to migrate for 24 h. Cells were then stained with Giemsa and photographed at a magnification of ×10. A representative of multiple similar fields is presented for each condition. (B) The migratory potentials of MCF-7/ Neo and MCF-7/HARACK1 cells were compared in Transwell assays. Cells were plated at 50,000 cells/chamber and allowed to migrate toward IGF-I for 24 h. Cells were fixed by using methanol before they were stained with crystal violet and examined under a magnification of ×10. Expression of RACK1 in each cell line tested compared with that in MCF-7 cells is shown in Western blots in right-hand panels. (C) MCF-7/Neo and MCF-7/HARACK1 cells were starved and stimulated with IGF-I. Cell lysates were immunoprecipitated with anti-β1 integrin antibody, followed by Western blotting with anti-RACK1 antibody. The presence of IGF-IR in the immunoprecipitate was confirmed by using anti-IGF-IR antibody on the same blot. A lysate control is included to compare levels of RACK1 in the cells. (D) IGF-IR was immunoprecipitated from MCF-7, MCF-7/Neo, and MCF-7/HARACK1 cells, and RACK1 was detected in the immunoprecipitate by probing the blot with anti-RACK1 antibody. (E) Lysates from MCF-7/Neo or MCF-7/RACK1 cells were assessed by Western blotting for FAK phosphorylation by using anti-phospho-FAK (Y397) antibody. The blot was stripped and reprobed for FAK and actin as loading controls.

FIG. 7.

Okadaic acid reverses the enhanced motility of MCF-7/RACK1 cells. (A) MCF-7/Neo and MCF-7/HARACK1 cells to IGF-I were compared for migratory potential in Transwell assays. Cells were starved and incubated with or without okadaic acid (OA) at 1 nM, seeded at 50,000 cells/chamber, and allowed to migrate toward IGF-I for 16 h. Cells were fixed by using methanol before they were stained with crystal violet and examined under a magnification of ×10. The data are presented as the number of cells/well for triplicate wells of each culture condition. MCF-7 cells are included as a control. (B) MCF-7/Neo and MCF-7/HARACK1 cells were starved and stimulated with IGF-I for the indicated times. Cell lysates were immunoprecipitated (IP) with anti-RACK1 antibody, followed by Western blotting with anti-PP2A C antibody. The presence of RACK1 in the immunoprecipitates was confirmed by using anti-RACK1 antibody on the same blot. Levels of each protein in cell lysates are shown in the right-hand panel. (C) PP2A activities in MCF-7/Neo (Neo) and MCF-7/HARACK1 (HARACK1) cells were assayed, using p-NPP as a substrate. The data are presented from triplicate samples, where 100% activity is represented by Neo cells not stimulated with IGF-I.

FIG. 8.

RACK1 siRNA ablates IGF-I-mediated regulation of PP2A activity and cell migration. (A) MCF-7 cells transfected with two different siRNA oligonucleotides directed to RACK1 (siRNA1 and siRNA2) or a scrambled oligonucleotide (control) were assessed for RACK1 expression by Western blotting (upper panels). PP2A was immunoprecipitated from cells stimulated or not stimulated with IGF-I and was assayed for phosphatase activity. Data from triplicate samples are presented, with 100% activity representing that measured in MCF-7 cells transfected with a negative control. (B) To investigate the effects of siRNA RACK1 on cell migration, DU145 cells were transfected with siRNA1 or a control oligonucleotide (10 μM). Expression of RACK1 was assessed by Western blotting (upper panel). Cells were assessed for migration toward serum in Transwell assays at 24 h posttransfection. Cells were fixed by using methanol before they were stained with crystal violet and counted under a magnification of ×10.

FIG. 9.

Model to illustrate actions of RACK1 in regulating recruitment of β1 integrin and PP2A to control cell migration. The left panel represents serum-starved cells in which RACK1 is associated with the IGF-IR and PP2A. FAK phosphorylation and PP2A activity are high, and there is no cell migration. The right panel represents cells stimulated with IGF-I. PP2A is rapidly released from RACK1, and β1 integrin is recruited to RACK1, resulting in a complex that includes the IGF-IR and β1 integrin. PP2A activity is transiently decreased, FAK phosphorylation is decreased, and cell migration is stimulated.