Energy stress regulates hippo-YAP signaling involving AMPK-mediated regulation of angiomotin-like 1 protein

Abstract

Hippo signaling is a tumor-suppressor pathway involved in organ size control and tumorigenesis through the inhibition of YAP and TAZ. Here, we show that energy stress induces YAP cytoplasmic retention and S127 phosphorylation and inhibits YAP transcriptional activity and YAP-dependent transformation. These effects require the central metabolic sensor AMP-activated protein kinase (AMPK) and the upstream Hippo pathway components Lats1/Lats2 and angiomotin-like 1 (AMOTL1). Furthermore, we show that AMPK directly phosphorylates S793 of AMOTL1. AMPK activation stabilizes and increases AMOTL1 steady-state protein levels, contributing to YAP inhibition. The phosphorylation-deficient S793Ala mutant of AMOTL1 showed a shorter half-life and conferred resistance to energy-stress-induced YAP inhibition. Our findings link energy sensing to the Hippo-YAP pathway and suggest that YAP may integrate spatial (contact inhibition), mechanical, and metabolic signals to control cellular proliferation and survival.

Figure 1. Energy stress inhibits YAP

A, Confluent and serum stimulated HEK293A cells were treated with DMSO control, AICAR (1mM), metformin (10mM) or phenformin (1mM), with or without Compound C (10 μM) for 3h. YAP nuclear localization is quantified by determining the Pearson’s correlation coefficient of nuclear staining. Data are represented as mean, S.D. n=3. (*, p<0.03; **, p<0.001). B, Small molecule energy stressors induce YAP S127 and TAZ Ser89 phosphorylation in HaCaT cells. HaCaT cells were treated with DMSO (control), 1 mM phenformin, 1 mM AICAR or 25 mM 2-DG for 6h. C, Inhibition of YAP-reporter activity. HEK293 YAP reporter cells were treated with DMSO control, 2-DG (8.3 or 25 mM), AICAR (1 or 3 mM) or Metformin (10 mM). (data are represented as mean, S.D. n=3). D, Energy stressors suppress YAP target gene expression. HEK293A cells were treated with DMSO (ctrl), 1mM phenformin, or 1mM AICAR for 16 hours. Relative mRNA levels are normalized to DMSO (ctrl). E, MCF10A cells transduced with a vector control, YAP wild type or YAP S127A mutant were grown on Matrigel without EGF. Cells were treated with metformin (2mM or 10mM) or DMSO control. Scale bar: 100 μm. F, The colony numbers per field are quantified by Image J, and shown in bar graph (data are represented as mean, SD, n=4; *, p<0.05; **, p<0.0005 by comparing to the S127A mutant samples with the same treatment). G, Metformin dose dependently inhibits the proliferation of primary mouse hepatocellular carcinoma (HCC) cell line JF001 (Mst1^−/−Mst2^−/−). JF001 cells transduced with vector, wild-type YAP, or YAP S127A mutant were treated as indicated. (Data are represented as mean, SD, n=3; *, p<0.05; **, p<0.001 by comparing to the DMSO control). See also Figure S1.

Figure 2. Energy stress inhibits YAP through AMP-activated protein kinase (AMPK) and Lats1/2

A, AMPK is required for YAP inhibition. Wild type or AMPK^−/− (AMPKα1α2 double knockout) MEFs were treated with DMSO control, 2-DG (25 mM) or phenformin (1 mM) for 12h. Cells were then harvested for western blot analysis. B, HaCaT cells were transfected with siRNAs targeting AMPKα1 and α2, and then treated for 16h with DMSO or phenformin (1mM). C, siRNAs targeting AMPKα1 and α2 blocks YAP cytoplasmic retention. Cells were transfected with siRNAs targeting AMPK, and then treated for 16h with DMSO, metformin (10mM) or phenformin (1mM). YAP nuclear localization is quantified by determining the Pearson’s correlation coefficient between YAP positive areas with the nuclear staining. Data are represented as mean, S.D. n=3. (*, p<0.05; **, p<0.001 by comparing to the DMSO control). D, HCCs were isolated from tumors derived from liver specific Mst1^−/−Mst2^−/− mice. Cells are treated with metformin (10mM) or co-treated with Compound C (10 μM) for 8h. p-YAP levels were evaluated by western blot. E, HEK293A cells were transfected with siRNA control or siRNA targeting Lats1 and Lats2 (#5+8, or #5+10). Cells were treated with AICAR (1mM) or Metformin (10mM) for 8h, serum free media (SFM, 3h) or 2-DG (25mM, 3h). Cells were then harvested for western blot analysis of p-YAP (Ser127). See also Figure S2.

Figure 3. Endogenous angiomotin like-1 (AMOTL1) is stabilized in response to energy stress and is required to mediate AMPK-induced YAP-cytoplasmic retention and Ser127 phosphorylation

A, HEK293A cells were treated with DMSO control or phenformin (1 mM) for 16h. Protein levels of Hippo pathway components (AMOTL1, AMOT, AMOTL2, p-YAP, YAP, p-Lats1, Lats1, MST2, and NF2) were analyzed by western blots. B, Cells were treated with metformin (10mM) for 4h and with cycloheximide (CHX) with indicated time. Protein levels of endogenous AMOTL1 were analyzed by western blots. C, Protein levels from B were quantified by densitometry of the bands. Projected degradation curve of AMOTL1 was plotted and fitted using Prism software. D, HEK293A cells were transfected with siRNAs targeting AMOT and AMOTL1, with GFP to mark transfected cells. Cells were treated with phenformin (1mM) or DMSO. YAP nuclear localization is quantified by determining the Pearson’s correlation coefficient, in GFP positive (transfected) or GFP negative (non-transfected) cells. Data are represented as mean, S.D. n=3. (*, p<0.02; **, p<0.001 by comparing to the DMSO controls). E. HaCaT cells were transfected with shRNA targeting AMOTL1. Cells were treated with DMSO control or phenformin (1mM) and p-YAP level is analyzed. F. Cells were transfected shRNA targeting AMOTL1, and then treated with DMSO, metformin (10mM) or phenformin (1mM) for 48 hours. The cell proliferation is determined by measuring the cell viability (*, p<0.01). See also Figure S3.

Figure 4. AMPK directly phosphorylates AMOTL1 Ser793 leading to its stabilization and YAP inhibition

A, Alignment of the conserved AMPK substrate motifs in angiomotin family proteins (AMOTL1, AMOT and AMOTL2) of different species. B, LC-MS/MS studies of phosphorylation of AMOTL1 Ser793. Flag-AMOTL1 was purified from HEK293A cells treated with metformin (10mM) or metformin and C ompound C (10μM). Extracted ion chromatographs of peptides containing Ser793 were shown. Peak A (unphosphoryated peptide), Peak B (peptide with p-Ser805), Peak C (peptide with p-Ser793), Peak D (peptide with p-Ser793 and p-Ser805). C, The abundance of the peptides was quantified by measuring the area under the curve (AUC) in the spectra. D, Recombinant AMPK phosphorylates AMOTL1 wild type, but not S793A mutant in vitro. E, HEK239T cells were treated for 8 hours with DMSO (control), 1μM A-769662, or 1mM Phenformin. Protein lysates were then analyzed by Phos-Tag SDS-PAGE followed by Western blotting with an anti-AMOTL1 antibody. F, Phenformin treatment increased AMOTL1 S793 phosphorylation. An antibody recognizing p-AMOTL1 (Ser793) was used in western blot. HaCaT cells were transfected with Flag-AMOTL1, and then treated with DMSO, or phenformin (1mM) for 1hr or 2hr. G, Knock-down of AMPK decreased the levels of p-AMOTL1. Cells were transfected with siRNAs targeting AMPKα1 and α2, and Flag-AMOTL1. Cells were then treated with metformin (10mM) or DMSO control. p-AMOTL1 levels were determined by using a phosphospecific antibody recognizing AMOTL1 Ser793. H, AMOTL1 S793A has a shorter half-life than wild type AMOTL1. HEK293T cells transfected with either wild type or S793A mutant AMOTL1 were treated for 16 hours with A-769662 then treated for the indicated time with cycloheximide (CHX). I. Cells were transfected with AMOTL1 wild type or S793A mutant, and then treated with DMSO control, or phenformin (1mM) for 48 hours. The cell viability is determined by CellTiter Glo (*, p<0.01). J. A proposed model of energy stress-mediated YAP inhibition. See also Figure S4.