Proteomic analysis reveals APC-dependent post-translational modifications and identifies a novel regulator of β-catenin

Abstract

Wnt signaling generates patterns in all embryos, from flies to humans, and controls cell fate, proliferation and metabolic homeostasis. Inappropriate Wnt pathway activation results in diseases, including colorectal cancer. The adenomatous polyposis coli (APC) tumor suppressor gene encodes a multifunctional protein that is an essential regulator of Wnt signaling and cytoskeletal organization. Although progress has been made in defining the role of APC in a normal cellular context, there are still significant gaps in our understanding of APC-dependent cellular function and dysfunction. We expanded the APC-associated protein network using a combination of genetics and a proteomic technique called two-dimensional difference gel electrophoresis (2D-DIGE). We show that loss of Drosophila Apc2 causes protein isoform changes reflecting misregulation of post-translational modifications (PTMs), which are not dependent on β-catenin transcriptional activity. Mass spectrometry revealed that proteins involved in metabolic and biosynthetic pathways, protein synthesis and degradation, and cell signaling are affected by Apc2 loss. We demonstrate that changes in phosphorylation partially account for the altered PTMs in APC mutants, suggesting that APC mutants affect other types of PTM. Finally, through this approach Aminopeptidase P was identified as a new regulator of β-catenin abundance in Drosophila embryos. This study provides new perspectives on the cellular effects of APC that might lead to a deeper understanding of its role in development.

Keywords: 2D-DIGE; APC; ApepP; Post-translational modification; Wnt signaling.

Fig. 1.

Summary of the Wnt signaling pathway and 2D-DIGE workflow. (A) The Wnt signaling pathway. When no Wnt ligand is present (left) the destructosome promotes cytoplasmic β-cat degradation. With a Wnt ligand (right), the destructosome is deactivated and cytoplasmic β-cat accumulates, enters the nucleus and activates Wnt target genes. (B) Principal steps in the 2D-DIGE workflow: (1) covalent labeling of protein lysates with either propyl-Cy3 or methyl-Cy5; (2) combine labeled protein lysates; (3) co-electrophorese on a 2DE gel; (4) fluorescent gel imaging of Cy3 (green) and Cy5 (red) (common spots are yellow, whereas difference-proteins appear as green or red); (5) selected spots shown in the inset are quantified using SourceExtractor; (6) concurrently genetically validate the difference-proteins (6a) and identify the proteins by LC-MS/MS (6b).

Fig. 2.

2D-DIGE of Apc2 null versus wild-type Drosophila embryos. A tiled array of a whole gel comparing Apc2^g10 (green) and wild-type (red) embryos. The 16 Apc2-dependent difference-protein pairs are demarcated by white boxes. Contrast and brightness were manipulated on the whole-gel images using ImageJ (full details of image analysis are provided in the supplementary Materials and Methods).

Fig. 3.

Quantification of Apc2-dependent difference-proteins. (A) The fold change for each difference-protein pair as calculated using the Cy3 and Cy5 raw fluorescence intensities. Cy3/Cy5 ratios <1 were converted to negative fold change. (B) Total protein abundance, defined as the sum of the fluorescence intensities of putative difference-protein isoforms, was calculated for each protein region in wild-type and Apc2^g10 embryos. The percentage difference in total protein isoform fluorescence relative to wild type was then calculated. Difference-proteins that display a >10% change in total abundance are marked with an asterisk.

Fig. 4.

Immunoblot confirmation that CaBP1 is an APC2-dependent difference-protein. (A) 2D-DIGE image of the region containing CaBP1 isoforms. (B) Fluorescent images of wild-type and Apc2^g10 embryo lysates that were separately labeled with Cy3 DIGE dye, resolved on different 2DE gels and transferred to nitrocellulose. (C) Immunoblot images of proteins stained with anti-CaBP1 antibody. (D) Superimposed images of total Cy3-labeled protein (red) and the CaBP1 2D immunoblot (green) confirm protein identification by LC-MS/MS. The predominant isoform found in wild-type lysate is the left isoform, whereas the right isoform is predominant in Apc2 null embryos.

Fig. 5.

APC2-dependent isoform changes are due to phosphorylation and other PTMs. (A) Model for the PTM responsible for Dp1 isoforms. (A′,A″) Dp1 isoforms are affected by phosphorylation, as the left isoform collapses to the right isoform after λPP treatment for both wild-type and Apc2^g10 embryo lysates. (B) Model for the PTM(s) responsible for ApepP isoforms. (B′,B″) ApepP isoforms were not affected by λPP treatment, as there were no isoform shifts observed after phosphatase treatment. Difference-protein isoforms (L, left; M, middle; R, right) are circled.

Fig. 6.

Validating the APC-dependent difference-proteins. (A) A decision tree showing the genetic approach taken to validate the APC-dependent difference-proteins. Six of the 16 difference-proteins were concluded to be APC dependent and to be independent of wild-type variation. (B) 2D-DIGE comparisons of FL-GFP-Apc2;Apc2^g10 versus Apc2^g10 or wild-type embryo lysates were performed to test whether ectopic APC2 could reverse the observed Apc2^g10 versus wild-type proteome changes. Shown here are sub-images cropped from tiled array images of whole 2D-DIGE gels. ApepP and Irp-1B exemplify reversal and failure to compensate, respectively.

Fig. 7.

ApepP^EY embryos display Wnt activation phenotypes similar to Apc2 mutant embryos. (A-F) Anterior is to the left and posterior is to the right. (A-C′) Wild-type (A) and ApepP^EY mutant (B,C) cuticles. ApepP^EY cuticles exhibit head defects, fusion of the first and second denticle rows (compare A′ with B′, arrowhead) and denticle loss (B′, circled). ApepP^EY cuticles also have body holes (C,C′, circled). (D) Dose reduction of Apc2 in ApepP^EY embryos modifies the ApepP^EY phenotype. This example has head defects and denticle loss. (E,F) Wild-type (E) Arm protein localization and accumulation reflect patterned Wnt pathway activation, whereas ApepP^EY mutant embryos (F) exhibit a more uniform accumulation of Arm protein, as exemplified in the traces beneath. (G) 2D-DIGE comparison of wild-type and ApepP^EY embryos showing that total ApepP protein is decreased in mutant embryos and that there is a shift in isoform distribution toward the left isoform in ApepP^EY embryos. (H-J) Anti-Arm immunoblots of embryo lysates at 0-2 h (I) and 4-6 h (J) AEL. At 0-2 h AEL Arm significantly accumulates in both mutants, consistent with reduced Arm degradation, whereas at 4-6 h AEL Arm levels are decreased in both mutants (H; n=3 replicates).