The mechanism of full activation of tumor suppressor PTEN at the phosphoinositide-enriched membrane

Abstract

Tumor suppressor PTEN, the second most highly mutated protein in cancer, dephosphorylates signaling lipid PIP₃ produced by PI3Ks. Excess PIP₃ promotes cell proliferation. The mechanism at the membrane of this pivotal phosphatase is unknown hindering drug discovery. Exploiting explicit solvent simulations, we tracked full-length PTEN trafficking from the cytosol to the membrane. We observed its interaction with membranes composed of zwitterionic phosphatidylcholine, anionic phosphatidylserine, and phosphoinositides, including signaling lipids PIP₂ and PIP₃. We tracked its moving away from the zwitterionic and getting absorbed onto anionic membrane that harbors PIP₃. We followed it localizing on microdomains enriched in signaling lipids, as PI3K does, and observed PIP₃ allosterically unfolding the N-terminal PIP₂ binding domain, positioning it favorably for the polybasic motif interaction with PIP₂. Finally, we determined PTEN catalytic action at the membrane, all in line with experimental observations, deciphering the mechanisms of how PTEN anchors to the membrane and restrains cancer.

Keywords: Cancer; In Silico Biology; Structural Biology.

Graphical abstract

Figure 1

PTEN sequence and structure (A) The sequence and domain structure of PTEN. In the sequence, the underlined residues highlight the PIP₂ binding domain (PBD) at the N-terminal and the PDZ-binding motif at the C-terminal regions. In the domain structure, the phosphorylated serine-threonine cluster (residues 380–385) in the C-terminal tail is marked. (B) The crystal structures of PTEN (PDB: 1D5R). The marked residues denote the missing regions. (C) In silico model of the full-length PTEN structure. The missing regions in PTEN, the PBD (residues 1–15), the intrinsically disordered region (IDR, residues 282–312), and the C-terminal tail (residues 351–403) were constructed by the I-TASSER program (Roy et al., 2010; Yang et al., 2015; Yang and Zhang, 2015).

Figure 2

Membrane interaction of PTEN (A) Snapshot representing the final PTEN conformation in the lipid bilayer composed of DOPC:DOPS:PIP₂:PIP₃ = 32:6:1:1, in molar ratio. (B) The root-mean-squared-fluctuations (RMSFs) of PTEN residues. (C) Time series of the deviations of the center of mass of individual PTEN domains from the bilayer surface. (D) Highlight showing the interaction of unfolded PBD with the lipids. Key basic residues, Lys6, Arg11, Lys13, Arg14, and Arg15, are marked.

Figure 3

Lipid contact probability The probability of lipid contacts for PTEN residues for different bilayer systems, including PC (pure DOPC), PS (DOPC:DOPS = 4:1, molar ratio), P2 (DOPC:DOPS:PIP₂ = 16:3:1), P3 (DOPC:DOPS:PIP₃ = 16:3:1), and PP (DOPC:DOPS:PIP₂:PIP₃ = 32:6:1:1) systems.

Figure 4

Salt bridge interactions of PTEN (A) Mapping of the basic residues on the membrane-binding surface of the phosphatase and C2 domains. PIP₃-favored residues in the phosphatase domain are colored red and those in the C2 domain are colored blue. (B) Contour map of the probabilities of salt bridge formations between the key basic residues at the membrane-binding interface and the lipids. Salt bridge is calculated between the nitrogen atoms in the side chains of basic residues and the oxygen atoms in the phosphate group of all lipids and the inositol head group of phosphoinositides with the cutoff distance of 3.2 Å. (C) Averaged deviations of the amide nitrogen in the side chains of Arg and Lys residues from the bilayer surface for the PIP₃-favored residues in the phosphatase (left panel) and C2 (right panel) domains. Error bars denote standard deviation.

Figure 5

PIP₃ coordination induces the P loop conformational change (A) A snapshot highlighting P loop and protruded PIP₃ from the bilayer surface at the active site. (B) Deviations of the phosphate atoms of PIP₂ and PIP₃ from the averaged position of the phosphate atoms of DOPC and DOPS for the apo and holo bilayer leaflets in the absence and presence of the protein, respectively. The red and black lines in the box graphs indicate the mean and median values, respectively, and whiskers above and below the box indicate the 90th and 10th percentiles. (C) The open conformation of P loop with the coordination of PIP₃ at the active site.

Figure 6

High density of waters at the active site for catalysis (A) Highlight showing the tunnel-like, water-filled space formed by the inner surfaces of phosphatase and C2 domains and the bilayer surface denoting as the aqueous canal. Along the water-filled interface between the phosphatase and C2 domains, PIP₃-favored distal basic residues are marked. (B) The structure and diameter of aqueous canal calculated with HOLE program (Smart et al., 1993). For the canal structure, green denotes the diameter in the range 4 ≤ d ≤ 10 Å and blue denotes the diameter of d > 10 Å. (C) Three-dimensional water density map with probabilities p = 0.5 (yellow surface) and p = 0.4 (blue mesh). (D) Interdomain residue-residue contacts for PTEN. For two intramolecular residues i and j, the probability of contact for the distance between the $C_{\beta }^i-C_{\beta }^j$ (C_α is used for Gly residue) with cutoff 10 Å was calculated. Residue j with |j – i| < 4 was omitted, and p < 0.1 was neglected in the calculation.

Figure 7

Schematic diagram illustrating membrane localization of PTEN In the cytosol, PTEN is autoinhibited by the phosphorylated CTT, through the “closed-closed” conformation. Dephosphorylation on the CTT removes the autoinhibition, shifting the populations toward the “open-closed” conformation. With the exposed membrane-binding surface, PTEN can localize at the anionic membrane. Full activation of PTEN requires the membrane anchorage of PBD at the membrane microdomain enriched by both phosphoinositides, PIP₂ and PIP₃. In the cartoon, PTP denotes the phosphatase domain.