Research ArticleCELL BIOLOGY

BNIP-2 retards breast cancer cell migration by coupling microtubule-mediated GEF-H1 and RhoA activation

See allHide authors and affiliations

Science Advances  31 Jul 2020:
Vol. 6, no. 31, eaaz1534
DOI: 10.1126/sciadv.aaz1534
  • Fig. 1 BNIP-2 interacts with RhoA via BCH domain.

    (A) Schematic representation of expression plasmids encoding BNIP-2-FL and its truncation and deletion mutants. In the BNIP-2-ΔRBD construct, 45 amino acids (167 to 211) are deleted. (B) BNIP-2 specifically binds to RhoA. Lysates of human embryonic kidney (HEK) 293T cells transiently expressing hemagglutinin (HA)–tagged BNIP-2 with FLAG-tagged empty vector, RhoA, Rac1, or Cdc42 were immunoprecipitated (IP) with anti-FLAG beads and then Western blotted (WB) with FLAG or HA antibodies. Blue arrows denote heavy chains and light chains from M2 beads that are present in all samples. (C) BNIP-2 uses its BCH domain to interact with RhoA. Lysates of HEK293T cells transiently expressing green fluorescent protein (GFP)–tagged BNIP-2, BNIP-2-ΔBCH, or BNIP-2-CBCH with FLAG-tagged RhoA were immunoprecipitated with anti-FLAG beads and then Western blotted with FLAG or GFP antibodies. Blue arrow denotes heavy chains from FLAG beads. (D) BNIP-2 uses the RBD-like region in the BCH domain to interact with RhoA. Lysates of HEK293T cells transiently expressing FLAG-tagged BNIP-2-FLand BNIP-2-ΔRBD with HA-tagged empty vector or RhoA were immunoprecipitated with anti-HA beads and then Western blotted with FLAG or HA antibodies. Blue arrow denotes heavy chains from HA beads that are present across all samples. BNIP-2 signal is shown just under the heavy chain in lane 3, denoted by the blue asterisk.

  • Fig. 2 BNIP-2 knockdown suppresses RhoA activity and retards cell polarization during early cell spreading.

    (A) Endogenous RhoA activity of MDA-MB-231 cells was examined in the presence of different amounts of BNIP-2. Lysates of MDA-MB-231 cells transfected with BNIP-2–targeting siRNA (siBNIP-2) or transiently expressing gradually increasing amount of GFP-BNIP-2 (illustrated by blue triangle) were used for pulldown with immobilized glutathione S-transferase (GST)–RBD (GST-RBD) of rhotekin and then Western blotted with RhoA, BNIP-2, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies. Blue arrow denotes the GFP-BNIP-2, while purple arrow denotes the endogenous BNIP-2. The ratio of active RhoA to total RhoA is normalized to untransfected in lane 2 and labeled at the bottom. (B) Snapshots of shVector control and BNIP-2 knockdown cells during spreading on collagen I–coated plastic-bottom plates. Representative cells spreading at 0, 20, 40, and 54 min are displayed. (C) Cell aspect ratios are quantified and plotted for control and BNIP-2 knockdown cells during 60-min spreading. Data are means ± SEM of two independent experiments.

  • Fig. 3 BNIP-2 knockdown promotes MDA-MB-231 cell migration.

    (A and B) BNIP-2 expression level is down-regulated in patient breast cancer samples in comparison to normal tissues. Microarray data were obtained from National Center for Biotechnology Information Gene Expression Omnibus database GDS3853 and GDS3139. Two-tailed t test was conducted. (A) In GDS3853, P < 0.01. (B) In GDS3139, P < 0.01. (C to F) BNIP-2 overexpression suppresses MDA-MB-231 cell transwell migration, while BNIP-2 knockdown promotes this process. (C) Representative images of transwell migration assay on MDA-MB-231 control and BNIP-2 overexpression cells. Transwell migration analysis was conducted using 10% fetal bovine serum–containing medium as a chemoattractant. Cells were fixed by 4% paraformaldehyde (PFA) after 6-hour migration. Cells migrated through and localized at the bottom side of the insert were stained with crystal violet for cell counting. (D) Statistical analysis between control and overexpression cells in transwell migration assay. Cell number per area was counted form randomly choosing sites and averaged. Final results presented here were normalized to the number of control cells (equals 1). Data are means ± SEM of four independent experiments, P < 0.05. (E) Representative images of transwell migration assay on MDA-MB-231 control and BNIP-2 knockdown cells. (F) Statistical analysis between control and knockdown cells in transwell migration assay. Data are means ± SEM of four independent experiments, P < 0.01. Scale bars in C and E, 100 μm. (G to J) BNIP-2 retards collective cell migration in MDA-MB-231 cell. (G) Stable BNIP-2–expressing MDA-MB-231 cells retard collective migration than control cells. (H) Statistical analysis for (G). Data are means ± SEM of five independent experiments, P < 0.05. (I) Knockdown of BNIP-2 increased the speed of collective cell migration. Red dashed rectangles denote gap area when stencile was removed (0 hour), and white dashed rectangles denote gap area after cells migrate collectively after 6 hours. (J) Statistical analysis for (I). Data are means ± SEM of four independent experiments, P < 0.05.

  • Fig. 4 BNIP-2 binds GEF-H1 and scaffolds for RhoA/GEF-H1 signaling.

    (A) MDA-MB-231 cells stained for BNIP-2 and α-tubulin. Boxes A and B highlight the regions where BNIP-2 colocalizes with microtubules. (B) BNIP-2 interacts with GEF-H1 under physiological conditions. HEK293T (top) or MDA-MB-231 (bottom) cell lysates were incubated with BNIP-2 antibody or immunoglobulin G (IgG) control for immunoprecipitation and then Western blotted with GEF-H1 and BNIP-2 antibodies. The blue arrow denotes heavy chains. (C) BNIP-2 uses the C terminus to interact with GEF-H1. Lysates of HEK293T cells transiently expressing HA-GEF-H1 and GFP-vector, BNIP-2-FL, or different mutants were immunoprecipitated with anti-HA beads and then Western blotted with GFP or HA antibodies. The blue arrow denotes heavy chains. (D) BNIP-2 uses the RBD-like region in BCH domain to interact with GEF-H1. Lysates of HEK293T cells transiently expressing HA-tagged BNIP-2 or BNIP-2-ΔRBD with FLAG-tagged GEF-H1 were immunoprecipitated with anti-HA beads and then Western blotted with FLAG or HA antibodies. (E) Knockdown of BNIP-2 reduces interaction between RhoA and GEF-H1 in HEK293T cells. Lysates of shVector control and BNIP-2 knockdown HEK293T cells transiently expressing HA-GEF-H1 and FLAG-RhoA were immunoprecipitated with anti-FLAG beads and then Western blotted with FLAG, HA, BNIP-2, or α-tubulin antibodies. (F) Knockdown of BNIP-2 reduces interaction between RhoA and GEF-H1 in MDA-MB-231 cells. Lysates of control and knockdown MDA-MB-231 cells were incubated with GEF-H1 antibody or IgG control for immunoprecipitation and then Western blotted with GEF-H1, RhoA, BNIP-2, or GAPDH antibodies. (G) BNIP-2 has a scaffolding effect for RhoA and GEF-H1 interaction. Lysates of knockdown HEK293T cells transiently expressing HA-tagged GEF-H1, FLAG-tagged RhoA, and increasing amount of HA-BNIP-2 (illustrated by blue triangle) were immunoprecipitated with anti-FLAG beads and Western blotted with FLAG, HA, or BNIP-2 antibodies. The blue asterisk denotes the endogenous BNIP-2, and the purple asterisk denotes the HA-tagged BNIP-2.

  • Fig. 5 BNIP-2 association with KLC-1 is required for its GEF-H1 binding and migration control.

    (A) BNIP-2 associates with KLC-1. Schematic representation of BNIP-2-WT with the kinesin-binding motif and BNIP-2-5A mutant, which is unable to bind KLC-1, where WE-WED motif in the N terminus is mutated to five alanines. Bottom: MDA-MB-231 cells were transfected with mCherry-KLC-1 and GFP-BNIP-2-WT, GFP-BNIP-2-5A, or GFP-vector for live imaging. Maximum Z-projected images are presented. Scale bars, 10 μm. (B) KLC-1 binding motif of BNIP-2 is required for interaction with GEF-H1. Lysates of HEK293T cells transiently expressing FLAG-tagged BNIP-2-WT, BNIP-2-5A, or empty vector with HA-tagged GEF-H1 were immunoprecipitated with anti-FLAG beads and then Western blotted with FLAG or HA antibodies. FLAG-V stands for FLAG-vector. (C) BNIP-2-5A mutant cannot rescue cell motility increased by BNIP-2 knockdown. Representative images of transwell migration assay on MDA-MB-231 shVector control cells, shBNIP-2 cells, and BNIP-2 knockdown cells rescued with BNIP-2-5A mutant. Scale bars, 100 μm. Right: Statistical analysis for shVector control cells, shBNIP-2 cells, and BNIP-2 knockdown cells rescued with BNIP-2-WT or BNIP-2-5A mutant. Cell number per area was counted from randomly choosing sites and averaged. Final results presented here were normalized to the number of control cells (equals 1). Data are means ± SEM of three independent experiments, P < 0.01. n.s., not significant.

  • Fig. 6 BNIP-2 is required for GEF-H1–driven RhoA activation upon microtubule disassembly.

    (A) Nocodazole treatment increases BNIP-2 and GEF-H1 interaction. MDA-MB-231 cells were treated with dimethyl sulfoxide (DMSO) (D) or 1 μM nocodazole (N) for 30 min and lysed. Lysates were incubated with BNIP-2 antibody or IgG control, incubated with Protein A/G PLUS-Agarose beads for immunoprecipitation, and then Western blotted with GEF-H1, BNIP-2, or GAPDH antibodies. (B) shVector control and BNIP-2 knockdown MDA-MB-231 cells were treated with DMSO or 1 μM nocodazole for 30 min, followed by immunostaining for α-tubulin and GEF-H1. Maximum Z-projected images are presented. (C) BNIP-2 is required for nocodazole-induced RhoA activation. MDA-MB-231 cells transfected with siControl and siBNIP-2 were treated with DMSO or 1 μM nocodazole for 30 min, lysed for RBD assay, and then Western blotted with RhoA, BNIP-2, or GAPDH antibodies. Quantification of Rho activity is displayed at the bottom with nocodazole normalized against DMSO for each cell line. (D) BNIP-2 and ROCK are required for nocodazole-induced cell rounding. shVector control and BNIP-2 knockdown MDA-MB-231 cells with or without 2-hour pretreatment with 20 μM Y-27632 (Y) were treated with DMSO or 1 μM nocodazole for 30 min and fixed for phalloidin stainning. Maximum Z-projected images are presented. Right: Statistical analysis for cell area. Data presented are means ± SEM, P < 0.0001. n.s., not significant.

  • Fig. 7 BNIP-2 phenocopies GEF-H1 effects in cell rounding upon nocodazole treatment and focal adhesion dynamics.

    (A) Nocodazole treatment causes dissolution of myosin filaments and focal adhesions. Top: shVector control MDA-MB-231 cells were treated with DMSO (D) or 1 μM nocodazole (N) for 30 min, followed by immunostaining for myosinIIA and paxillin. Bottom: BNIP-2 knockdown (KD) MDA-MB-231 cells were treated with DMSO or 1 μM nocodazole for 30 min, followed by immunostaining for myosinIIA and paxillin. Single-plane images are presented. Scale bars, 20 μm. (B) GEF-H1 knockdown cells have reduced cell rounding upon nocodazole treatment. Top: MDA-MB-231 cells transfected with control siRNA were treated with DMSO or 1 μM nocodazole for 30 min, followed by immunostaining for phalloidin and paxillin. Bottom: MDA-MB-231 cells transfected with GEF-H1 siRNA were treated with DMSO or 1 μM nocodazole for 30 min, followed by immunostaining for phalloidin and paxillin. Single-plane images are presented. Scale bars, 20 μm. (C to E) Focal adhesion turnover is affected by BNIP-2 knockdown, GEF-H1 knockdown, or Rho inhibition. (C) FRAP images of shVector control, BNIP-2 knockdown, GEF-H1 knockdown, and Rho-inhibited MDA-MB-231 cells expressing paxillin-mApple. Scale bars, 5 μm. (D) FRAP-relative fluorescence intensity curves of shVector control, BNIP-2 knockdown, GEF-H1 knockdown, and Rho-inhibited MDA-MB-231 cells expressing paxillin-mApple. The fluorescence intensity is normalized by the average prebleach intensity. (E) The mean ± SD half-life time (T1/2) of FRAP for each focal adhesion in shVector control, BNIP-2 knockdown, GEF-H1 knockdown, and Rho-inhibited MDA-MB-231 cells expressing paxillin-mApple. The half-life time is calculated from FRAP-relative fluorescence intensity (D). **P < 0.01 and ***P < 0.001.

  • Fig. 8 BNIP-2 as a scaffold for GEF-H1/RhoA activation in cancer cell migration.

    GEF-H1 plays an important role in the interplay between microtubules, actomyosin, and focal adhesions. After GEF-H1 is released from microtubules, it still requires BNIP-2 to act as a scaffold to activate RhoA via GEF-H1. When the levels of BNIP-2 are too low or too high, the interaction between GEF-H1 and RhoA and the levels of Rho–guanosine triphosphate are reduced. An optimal level of BNIP-2 promotes RhoA activation and inhibits cell migration. On the other hand, although microtubules sequester GEF-H1 from release, the binding of BNIP-2 to the microtubule motor kinesin-1 is required for the interaction between BNIP-2 and GEF-H1. This suggests that the BNIP-2/kinesin-1 trafficking pathway may regulate the dynamics of GEF-H1 on microtubules and release GEF-H1 in a spatially and temporally specific manner.

Supplementary Materials

  • Supplementary Materials

    BNIP-2 retards breast cancer cell migration by coupling microtubule-mediated GEF-H1 and RhoA activation

    Meng Pan, Ti Weng Chew, Darren Chen Pei Wong, Jingwei Xiao, Hui Ting Ong, Jasmine Fei Li Chin, Boon Chuan Low

    Download Supplement

    The PDF file includes:

    • Figs. S1 to S6
    • Legends for movies S1 and S2

    Other Supplementary Material for this manuscript includes the following:

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article