Binding partner- and force-promoted changes in αE-catenin conformation probed by native cysteine labeling

Adherens Junctions (AJs) are cell-cell adhesion complexes that sense and propagate mechanical forces by coupling cadherins to the actin cytoskeleton via β-catenin and the F-actin binding protein αE-catenin. When subjected to mechanical force, the cadherin•catenin complex can tightly link to F-actin through αE-catenin, and also recruits the F-actin-binding protein vinculin. In this study, labeling of native cysteines combined with mass spectrometry revealed conformational changes in αE-catenin upon binding to the E-cadherin•β-catenin complex, vinculin and F-actin. A method to apply physiologically meaningful forces in solution revealed force-induced conformational changes in αE-catenin when bound to F-actin. Comparisons of wild-type αE-catenin and a mutant with enhanced vinculin affinity using cysteine labeling and isothermal titration calorimetry provide evidence for allosteric coupling of the N-terminal β-catenin-binding and the middle (M) vinculin-binding domain of αE-catenin. Cysteine labeling also revealed possible crosstalk between the actin-binding domain and the rest of the protein. The data provide insight into how binding partners and mechanical stress can regulate the conformation of full-length αE-catenin, and identify the M domain as a key transmitter of conformational changes.

In order to compare solution structural data to the 4IGG crystal structure, we measured small angle x-ray scattering (SAXS) data from the 82-883 dimer construct. Comparison with the scattering curve calculated from the crystal structure shows very poor agreement (Fig. 2C). The radius of gyration (shape Rg) calculated from the crystal structure is 39 Å, much smaller than the measured Rg of 56 Å and the distance distribution function P(r) derived from the SAXS data shows a more extended structure (Fig. 2D). Similar findings were reported in SAXS studies of full-length αE-catenin 1,4 . The SAXS measurements and the lack of electron density for the ABD in the present crystal structure and 4K1N 1 indicate that full-length αE-catenin contains a flexibly linked ABD and does not form a compact structure in solution. Whether the compact structure with ordered ABDs visualized in the 4IGG structure represents a minor population of the solution ensemble remains to be determined; evidence presented below is consistent with communication between the ABD and the N-M region of αE-catenin. Conformational heterogeneity of the MI bundle (see also below) may also contribute to the larger Rg compared to the crystal structure; previously published modeling of αE-catenin SAXS data treated the M domain as a rigid unit as observed in the crystal structures 4 .

Supplemental Methods
Small Angle X-ray Scattering (SAXS) SAXS data were collected at beamline 4-2 at the Stanford Synchrotron Radiation Lightsource. Scattering data for αE-catenin 82-883 were obtained at 5 different concentrations (10, 5, 2.5, 1.25, and 0.6 mg ml -1 ) in 20 mM HEPES, pH 8.0, 150 mM NaCl and 1mM DTT buffer. 1 s exposures were recorded on a Pilatus 3XIM detector with a 0.1 x 0.1 mm beam and a 655 mm sample -detector distance. Scattering curves from 5 concentrations were merged and the radius of gyration Rg was determined by Guinier analysis using autoRg in the program Primus 5 . The P(r) function was calculated with GNOM 6 using an s range extended to 0.3 Å -1 .
To obtain data from monodisperse samples of full-length wild type and R551A αE-catenin monomer and dimer, the SEC-coupled SAXS setup at the beamline was utilized. 100 μl of 10 mg ml -1 wild type or R551A αE-catenin were injected onto a small-volume size exclusion column (Superdex PC3.2/200) equilibrated with PBS, 1mM DTT and 1% glycerol at a flow rate of 0.05 ml/min. The column eluate was directed through a quartz capillary and was exposed to the beam in 1 s intervals. Data were recorded on a Pilatus 3XIM detector with a 0.3 x 0.3 mm beam and a 1.7 m sample-detector distance, and processed with the program SasTool. The first 100 images were averaged and used as a buffer profile, which was then subtracted from the subsequent images. Rg is automatically calculated for each frame and then plotted together with the extrapolated I(0) value. Inspection of frames across the elution peak allows to select and average datasets with minimal contamination of either monomer or dimer. Frames 380-399 and 380-389 were averaged for wild-type and R551A monomer and frames 290-299 were averaged for wild-type and R551A dimer to determine the Rg and the P(r) function. The programs Primus and GNOM were used for analysis. The Rg was determined from the linear region of the Guinier plot with an sRg limit < 1.3. The program Crysol 7 was used to compute the scattering curve of the αE-catenin crystal structure (PDB 4IGG).

Crystallographic analysis of αE-catenin 82-883
Hexagonal crystals of αE-catenin 82-883 were grown at 22 °C by hanging drop vapor diffusion with a well solution containing 50 mM Tris, pH 8.5, 60 mM LiSO4, 23.5 % PEG 400. Typical crystal dimensions were 100 x 100 x 100 μm 3 . Crystals were frozen in liquid nitrogen without additional cryoprotectant. Data were collected in 0.5° rotation frames on a Pilatus 6M detector at BL 11-1 of the Stanford Synchrotron Radiation Laboratory. Data were integrated using XDS 8 and scaled using the program Aimless in the CCP4 program package 9 . The data were anisotropic, extending to 4.0 Å in the strongest (hk plane) direction, and 4.5 Å in the weaker (l axis) direction, as assessed by CC1/2 > 0. 5 (Ref. 9 ). Statistics are shown in Table S1.
The structure of αE-catenin 82-883 was solved by molecular replacement using copy A of the αE-catenin 82-906 dimer structure (PDB code: 4IGG) as a search model. Two copies of the model could be placed using the program Phaser 10 . Model refinement was carried out with the program Phenix 11 . The first round of rigid body refinement allowing the two copies to move individually resulted in Rwork = 45.8% and Rfree = 46.5%. After several rounds of rigid body refinement where the molecule was broken up into individual domains the R-factors dropped to Rwork = 31.0% and Rfree = 34.6%. Further refinement of the structure used interactive cycles of manual building with positional and grouped temperature factor refinement with secondary structure restraints. The refined model statistics are given in Table S1, and a representative portion of the electron density map is shown in Supplemental Figure 6. The coordinates and structure factors for αE-catenin (83-882) are available from the Protein Data Bank under accession code 6O3E.

Circular Dichroism measurements
CD melting curves of wild-type and R551A mutant αE-catenin monomer were recorded with an Aviv 202-01 (AVIV Biomedical, Inc.) or a Jasco J-815 (Jasco Analytical Instruments) CD spectrometer. Measurements were performed at 2.5 μM concentration in phosphate buffered saline with 1 mM DTT. CD melting curves were measured from 10 -95 °C at 222 nm at 1 °C intervals. For comparison of wild-type and R551A mutant melting curves, the CD signal was normalized. Melting curves were fitted to a two-state unfolding model 12 using an analysis program written in Python.