Survivin can alter mitochondrial architecture by regulating phosphatidylethanolamine synthesis

Abstract

Survivin (encoded by BIRC5) is an essential protein with established roles in mitosis and the inhibition of apoptosis. It is overexpressed in cancers, its abundance correlating with resistance to radiotherapies and chemotherapies. Survivin expression is normally limited to G2 and M phases; however, in cancer cells, it is also present during interphase and gains access to the mitochondria. Phosphatidylethanolamine (PE) is a phospholipid that facilitates negative curvature of membranes. It is enriched in the cytokinetic furrow and mitochondria, where it enables tight packing of the cristae and the increased accommodation of proteins. Here, we report the remarkable discovery that mitochondrial survivin regulates phosphatidylserine decarboxylase activity, thereby affecting PE availability. This novel molecular insight suggests that some apparently disparate roles of this 'multitasking' protein might be fundamentally linked to membrane architecture, and offers a new perspective on its contribution to cancer and potentially other metabolic disorders.

Keywords: Cancer; Mitochondria; Phosphatidylethanolamine; Phosphatidylserine decarboxylase; Phospholipid; Survivin.

Fig. 1.

Survivin interacts directly with active PSD. (A) Immunoprecipitation of endogenous survivin (SVN) using IgG control and anti-survivin antibodies and sepharose (protein A/G) or dynabeads, followed by immunoblotting with anti-PSD and anti-survivin antibodies. N=2, reproduced from Dunajova, 2016. (B) Immunoprecipitation of FLAG and FLAG–SVN from HEK293T cells transiently transfected with the relevant FLAG construct and untagged wild-type PSD (WT) or catalytically inactive PSD (S387A). The immunoblot was probed with anti-PSD and anti-FLAG antibodies. Inputs are 10% total (20 µg) of the whole-cell extract (WCE) and are indicated in left lanes. Representative of N=2. (C) In vitro pulldown of ³⁵S-labelled IVT PSD, expressed from pcDNA3.1 PSD, with GST control, GST-tagged full length survivin (SVN), the N-terminus and BIR domain of survivin (SVN_1-90) and the C-terminus (SVN_90-142). Equal loading of GST bait proteins is shown by Coomassie stain. N=2, reproduced from Dunajova, 2016. (D,E) Stick models of PSD and SVN highlighting the most salient features. MTS, mitochondrial targeting sequence; IM, inner membrane signal; LGST, autocatalytic motif. Residue numbers are indicated, T34 of SVN is phosphorylated by Cdk1. BIR domain and C-terminus are indicated on crystal structure (taken from PDB 1F3H, and constructed using UCSF-Chimera).

Fig. 2.

Expression of survivin mutants causes mitochondrial abnormalities. (A) Immunoblot of mitochondria isolated from HeLa cells in a protease protection assay treated with trypsin and/or hypotonic shock to investigate whether survivin and PSD are found in the same intramitochondrial compartment(s). N=1. Cox II was used to indicate the IMM and IMS, and Hsp60, the mitochondrial matrix. See Fig. S5B for source files. (B) Electron micrographs of mitochondria in HeLa cells expressing GFP or SVN variants. Scale bar: 5 µm. Lower panels show representative magnified images. EM preparations were made two independent times. See Fig. S2A for cell numbers. (C) Representative 2D projected images of deconvolved optical fluorescence z-stacks of SVN–GFP (green) variant-expressing cells with mitochondria visualised live using Mitotracker^® (red). Images representative of two independent repeats. Scale bar: 10 µm. (D) Respiration was measured for 50,000 cells of each line using the resazurin assay and the mean of N=2 independent experiments expressed as a percentage of the GFP control value. *P<0.05; **P<0.005; ns, not significant (unpaired two-tailed t-test against the GFP control, r² indicates the magnitude of difference).

Fig. 3.

Survivin regulates PE availability. (A) Abundance of PS and PE in the stable cell lines indicated was determined by thin layer chromatography (TLC) and plotted as the ratio of PE to PS. Mean of 5 independent repeats, error bars show 95% confidence intervals. A multiple comparison one-way ANOVA was applied. (B) Sensitivity of cell lines to duramycin (48 h). Cell number was determined using a resazurin assay and expressed as a percentage of the untreated control population and a multiple t-test was applied (N=3); P<0.05 is considered significant (*), P<0.005 (**); ns, not significant.