Membrane association of the PTEN tumor suppressor: molecular details of the protein-membrane complex from SPR binding studies and neutron reflection

Abstract

The structure and function of the PTEN phosphatase is investigated by studying its membrane affinity and localization on in-plane fluid, thermally disordered synthetic membrane models. The membrane association of the protein depends strongly on membrane composition, where phosphatidylserine (PS) and phosphatidylinositol diphosphate (PI(4,5)P(2)) act pronouncedly synergistic in pulling the enzyme to the membrane surface. The equilibrium dissociation constants for the binding of wild type (wt) PTEN to PS and PI(4,5)P(2) were determined to be K(d)∼12 µM and 0.4 µM, respectively, and K(d)∼50 nM if both lipids are present. Membrane affinities depend critically on membrane fluidity, which suggests multiple binding sites on the protein for PI(4,5)P(2). The PTEN mutations C124S and H93R show binding affinities that deviate strongly from those measured for the wt protein. Both mutants bind PS more strongly than wt PTEN. While C124S PTEN has at least the same affinity to PI(4,5)P(2) and an increased apparent affinity to PI(3,4,5)P(3), due to its lack of catalytic activity, H93R PTEN shows a decreased affinity to PI(4,5)P(2) and no synergy in its binding with PS and PI(4,5)P(2). Neutron reflection measurements show that the PTEN phosphatase "scoots" along the membrane surface (penetration <5 Å) but binds the membrane tightly with its two major domains, the C2 and phosphatase domains, as suggested by the crystal structure. The regulatory C-terminal tail is most likely displaced from the membrane and organized on the far side of the protein, ∼60 Å away from the bilayer surface, in a rather compact structure. The combination of binding studies and neutron reflection allows us to distinguish between PTEN mutant proteins and ultimately may identify the structural features required for membrane binding and activation of PTEN.

Figure 1. Quantification of PTEN Binding to stBLMs by Surface Plasmon Resonance.

Exemplary SPR results showing a raw data set (binding of wt PTEN to an stBLM composed of DOPC and ∼2 mol% PI(4,5)P₂; inset) and two binding isotherms (main panel). The green isotherms shows a fit to the data in the inset, which are well described by assuming a simple Langmuir adsorption (Eq. 1) with K_d∼300 nM. In distinction, the isotherm shown in red (binding of wt PTEN to an stBLM composed of DOPC, ∼28 mol% DOPS and ∼2 mol% PI(4,5)P₂ with chol) shows bimodal binding behavior of the protein to the two anionic components in the bilayer in which one component has a K_d of ∼40 nM and the second has a K_d of ∼10 µM. While the first of these values is smaller than the K_d of PTEN binding to PI(4,5)P₂ alone (see cross-over of the two curves in the low-concentration regime), the second K_d coincides approximately with that measured for wt PTEN binding to stBLMs that contain only DOPS as a single anionic component. A conversion of the raw data into equivalent protein mass per unit area is given in Information S1.

Figure 2. Quantification of PTEN Binding to stBLMs by Neutron Reflection.

Exemplary NR data set showing the changes of the neutron reflection upon wt PTEN association with a preformed stBLM composed of DOPC∶DOPS 70∶30, the residuals which emphasize these changes (bottom), and the corresponding nSLD profiles (inset). The NR spectra (main panel) for the stBLM before (black) and after incubation with 20 µM wt PTEN (red) are normalized to the Fresnel reflectivity—i.e., the reflectivity of a neat Si/buffer interface without interfacial roughness—in order to emphasize the interference patterns due to the interfacial structures. Changes of the spectra upon PTEN association with the membrane are shown as residuals, normalized to the magnitude of the experimental errors, at the bottom. Lines in the main panel show the computed NR of the nSLD profiles shown in the inset. Note that these profiles are derived by fitting multiple data sets simultaneously (neat stBLM and stBLM with PTEN, each measured at different isotopic buffer contrasts) by sharing model parameters as appropriate. A dashed box in the inset indicates the region of the distal lipid headgroups and associated PTEN protein, shown in close-up view in Fig. 4. The signal-to-noise in these measurements is comparable to that in similar studies on the incorporation of α-hemolysin into stBLMs and the binding of the HIV-1 matrix protein to bilayer surfaces .

Figure 3. Neutron Scattering Length Density Profile of PTEN Protein Bound to a Lipid Bilayer in an stBLM.

nSLD distribution of wt PTEN bound to a thermally disordered lipid membrane composed of DOPC with 30 mol% DOPS (+chol), as determined in a Continuous Distribution (CD) model. The sample is the same as that shown in Fig. 2. The red and black lines show those profiles for the as-prepared stBLM and the same sample after protein adsorption from a 20 µM PTEN solution, respectively, that derive from the most likely parameter set (Table 3). For the sample with PTEN protein, blue shaded contours visualize the bandwidth (1σ) of nSLD profiles that are also consistent with the experimental results, as determined by Monte-Carlo resampling of the data. The intensity of the shading corresponds to the probability of the model to be a true representation of the underlying structure.

Figure 4. Comparison of nSLD profiles of PTEN Interfacially Bound to Lipid Bilayers and Their Correspondence with the Crystal Structure of the Truncated PTEN.

The protein contributions (d = 60–120 Å from the Au surface) to the overall nSLD profile correspond to wt PTEN adsorbed to stBLMs (A) with 30 mol% DOPS (+chol), (B) with 28.6 mol% DOPS and 2.6 mol% PI(4,5)P₂ (+chol), and to H93R PTEN adsorbed to an stBLM with 30 mol% DOPS (+chol). The peak near d = 50 Å originates from the lipid headgroups. The color coding is the same as in Fig. 3, and panel (A) is a close-up view of the results shown there. Dashes black lines represent the best-fit overlay of a scaled nSLD distribution calculated from the crystal structure of the truncated PTEN variant. Because different amounts of protein was adsorbed in the individual experiments, the ratios between the protein and lipid headgroup peaks varies from sample to sample.

Figure 5. Schematic Depiction of the PTEN Phosphatase on the Surface of a Thermally Disordered stBLM.

Peptide backbone representation of (A) wt PTEN and (B) H93R PTEN positioned at a DOPC/DOPS (7∶3) membrane surface as deducted from the NR results. The membrane-associated protein penetrates the lipid headgroups (PC: violet, PS: orange) only barely. The PTEN core domains (PD: magenta, C2: grey) are shown in a conformation and membrane orientation deduced from the crystal structure . The close correspondence, observed in Fig. 4, between the nSLD distribution across the interface determined in this work and the nSLD distribution of the truncated PTEN computed from the crystal structure suggests that this is a good approximation. Moreover, about 20% of the protein mass have been deleted in the truncated protein, and ∼20% of the nSLD remains unaccounted for in the overall nSLD distribution for wt PTEN in Fig. 4A, if we position the PTEN core domains at the membrane as shown here. The C-terminal tail, which forms the bulk of the deleted peptide, is apparently quite different in its organization in wt and H93R PTEN at the membrane. Shown here in red, yellow and green are three distinct conformations, obtained from Monte-Carlo simulations , on each PTEN protein core that are consistent with the observed nSLD distributions shown in Figs. 4A and C.