1 Introduction

Cell–cell adhesion junctions are intricate assemblies of specialized molecular complexes that facilitate the adhesion of adjacent cells and the linkage of their actin cytoskeleton, orchestrating various cellular processes such as cell polarity, proliferation, differentiation, and the entry of cells into a quiescent state in the epithelial tissue [1,2,3,4]. Specifically, the adherens junction protein, E-cadherin, forms clusters [5], a critical step for maintaining tissue cohesion and orchestrating cellular rearrangement during growth and development [5,6,7,8,9]. The downstream effectors of E-cadherin signaling, the β-catenin-α-catenin complex, play a pivotal role in adherens junction formation. Here, β-catenin binds to the intracellular domain of E-cadherin, while α-catenin acts as a molecular bridge, binding to β-catenin on one side and filamentous actin (F-actin) on the other [10]. Importantly, the conformational state of α-catenin dynamically regulates its interaction with F-actin [11]. A force-dependent conformational change shifts α-catenin from a “closed” to an “open” state, enhancing its affinity for F-actin and vinculin—an additional adaptor protein binding F-actin [12,13,14,15]. Consequently, α-catenin facilitates a force-dependent increase in the strength of epithelial cell adhesion, emphasizing its pivotal role in maintaining tissue integrity and function.

Structural studies have shown that the interaction between β-catenin and α-catenin involves the N-terminal region of β-catenin spanning amino acid residues 57–143 of α-catenin and the N-terminal amino acid residues 118–149 of β-catenin [16,17,18]. In both proteins, these regions adopt α-helical structures, leading to the assembly of a four-helix bundle. This bundle comprises one helix from α-catenin and three helices from β-catenin [19, 20]. Additionally, it was shown that residue Y142, located immediately after the N-terminal region of β-catenin, is critical for α-catenin binding through an interaction between the hydroxyl group of residue Y142 with residue D144 in β-catenin via a water molecule. It also forms a van der Waals interaction with residue E147 in β-catenin [20]. Thus, residue Y142 in α-catenin stabilizes helix-helix interactions between α- and β-catenin [18].

Importantly, the α- and β-catenin complex along with E-cadherin has been shown to play a critical role in tumor suppression and cancer metastasis [21,22,23,24,25,26]. Numerous studies have suggested the potential of the cadherin/catenin complex as a prognostic marker for diagnosing tumor stage, metastasis, and patient survival across various cancers [9]. Also, it has been reported that changes in the interaction between β- and α-catenin either by phosphorylation or competitive binding of other proteins affect E-cadherin-based cell–cell adhesion [27]. Hence, targeting the β-catenin-α-catenin complex could be a good choice for inhibiting the formation or maintenance of E-cadherin-based cell–cell adhesion.

In this study, we attempted to develop β-catenin-derived peptides containing one or more mutations through structural analysis of the β- and α-catenin complex with the ability to bind α-catenin. Peptides are a type of biomaterial consisting of 2–50 amino acids that forms a polymer chain with different structures such as α-helix or β-sheet [28, 29]. Due to their higher efficacy and specificity and lower adverse effect and toxicity, peptides are emerging as a promising therapeutic option in drug discovery and development [30,31,32,33,34,35,36,37]. Computational modeling along with conventional as well as steered molecular dynamics (MD) simulations were used to determine the stability and the interaction strength of the wild-type (WT) and several β-catenin-derived mutant peptides with α-catenin. These revealed an enhanced interaction of the P128E/M131E/L132D mutant along with P128E, M131E, and P128E/M131E mutant peptides. We then developed a Bioluminescence Resonance Energy Transfer (BRET)-based assay [38,39,40,41,42,43,44,45,46] to test the binding of the β-catenin-derived peptides with α-catenin. BRET depends on non-radiative energy transfer between a luciferase energy donor and a fluorescent protein acceptor upon oxidation of the luciferase substrate. Therefore, for this study, we have engineered β-catenin-derived peptides fused with the mScarlet (mSca; acceptor) fluorescent protein and an α-catenin construct containing amino acid residues 57–262 of α-catenin fused with the NanoLuc luciferase (NLuc; donor). Live cell as well as in vitro BRET assays, however, revealed similar levels of interaction of the β-catenin-derived mutant and the WT peptides with α-catenin.

2 Materials and methods

2.1 MD simulations

MD simulations were performed using NAMD 12.4 software [47, 48]. The initial structure of the β-catenin-derived peptide bound to α-catenin (PDB ID: 1DOW [49], α-catenin: chain A, aa 57–262; β-catenin: chain B, aa 118–149) was obtained from the RCSB PDB database. Mutations in the β-catenin peptide were introduced using the MODELLER 10.1 suite [50], and the model with the lowest energy was selected for further simulation. The resulting β-catenin-derived peptides, both the wild type (WT) and mutants, complexed with α-catenin, were used as input to prepare the biomolecular simulation systems through the QwikMD toolkit integrated into the Visual Molecular Dynamics (VMD) software [51, 52]. The complex structures were solvated with a rectangular, explicit aqueous solvent model with a NaCl concentration of 0.15 M. The simulation systems underwent energy minimization and thermal equilibration as described previously [53]. This was followed by 25 ns of production simulation runs where conventional and steered MD simulations were performed in triplicates and quadruplicates, respectively. An integration timestep of 2-fs was chosen for all trajectories and trajectory frames were exported every 10,000 steps. For the conventional molecular dynamics, all production runs were performed at constant temperature of 310 K and at pressure of 1 atm using Langevin temperature control and Nose–Hoover Langevin piston control, respectively. Short-range non-bonded interactions were managed by employing a 12 Å cutoff alongside a 10 Å switching distance, while the Particle-mesh Ewald (PME) method was employed to manage long-range electrostatic interactions with a PME grid spacing of 1 Å. For the SMD, production runs were performed by applying a constant pulling force of 5 Å/ns. The pulling force was applied only on the residues of β-catenin (aa 118–149), while the β-catenin binding domain of α-catenin (aa 57–149) was kept dynamic and rest of residues after K149 were set as the static reference throughout the simulation runs.

2.2 MD simulation trajectory analysis

Analysis of MD simulation trajectories was performed using the available tools in the Visual Molecular Dynamics (VMD) software [52] as well as the MDAnalysis package in Python [54]. The analysis encompassed various parameters, including root-mean-squared deviation (RMSD), root-mean-squared fluctuation (RMSF), inter-chain center-of-mass (Come) distance, solvent accessible surface area (SASA), and radius of gyration (RoG). Hydrogen bond analysis in this study involved a switching angle of 200 and a cut-off distance of 3.5 Å, and it was performed using the “Hydrogen Bonds” plugin in the VMD software [52]. Additionally, salt bridge analysis was carried out with an oxygen–nitrogen distance cut-off of 3.2 Å, employing the “Salt bridges” plugin in the VMD software [52].

2.3 Plasmid design

To develop the BRET-based assay for monitoring the interaction between β-catenin-derived peptides and α-catenin, plasmid constructs expressing WT and various mutant β-catenin-derived peptides were generated. These constructs featured a C-terminally fused mScarlet (mSca) red fluorescent protein as the BRET acceptor and an α-catenin protein fragment spanning residues 57 to 261 with a C-terminal nanoluciferase (NLuc) protein as the BRET donor. For generating the BRET acceptor plasmids, the WT β-catenin peptide DNA sequence was synthesized with NheI and BamHI restriction enzyme sites at the 5′ and 3′ ends, respectively (Integrated DNA Technologies, IDT; Iowa, USA). This sequence was then inserted into the pIDTSmart (Kan) vector, resulting in the pIDT-β-catenin-WT plasmid construct. Plasmid constructs expressing mutant β-catenin-derived peptides were generated through PCR amplification using specific primers (Supporting Information) and subsequent insertion into the pIDTSmart vector. All vectors underwent transformation into E. coli for amplification and were purified using the Qiagen mini-prep kit. Then, restriction enzymes NheI and BamHI were used to excise the DNA fragments of interests from entry clones pIDT-β-catenin-peptide plasmids (both WT and mutants) and ligated into similarly digested destination plasmid pLifeAct_mScarlet_N1 [Addgene: 85054] [55]. β-Catenin-mSca plasmid constructs were confirmed by restriction digestion using BstXI and SalI to cut parenteral plasmid pLifeact-mScarlet-N1 as control along with the β-catenin-WT and mutants constructs. The BRET donor plasmid, generated by GenScript (Singapore), included an N-terminal mNeonGreen (mNG) green fluorescent protein with an α-helical linker to monitor protein levels through fluorescence measurements. α-Catenin plasmid construct was digested with EcoRI and XhoI releasing a fragment of 1175 bp confirmed by GenScript.

2.4 Cell culture and transfection

HEK 293T cells were used for in vitro and live cell assays. Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin and incubated at 37 °C in 5% CO2 [11, 39, 44, 56, 57]. For transfections for live cell assays, the following protocol was followed: HEK 293T cells were seeded in 96-well white plates and incubated for 24 h before transfection. The transfection cocktail was prepared using Opti-MEM (Invitrogen; 31,985,088), 1.25 μg/well of polyethyleneimine (PEI) lipid (Sigma-Aldrich; 408,727–100 mL), and the plasmid DNA (α-catenin plasmid construct and β-catenin-derived peptides with mSca). The DNA-PEI complex was incubated for 30 min at room temperature before being added to the cells. A 2 mg/mL stock solution of PEI was prepared by diluting it in sterile Milli-Q water and stored at − 80 °C.

2.5 Cell lysate preparation for in vitro BRET assays

HEK 293 T cells were transfected with α-catenin, pmScarlet-C1 (Addgene # 85042), and either WT or mutant β-cat-mSca expressing plasmid DNA separately in 6 cm dishes. After 48 h post-transfection, cells were washed with cold Dulbecco’s Phosphate-Buffered Saline (DPBS). Subsequently, cells were lysed using a lysis buffer containing 50 mM HEPES (pH 7.5), 50 mM NaCl, 0.1% Triton-X 100, 1 mM dithiothreitol (DTT), 1 mM ethylenediaminetetraacetic acid (EDTA), and a protease-inhibitor cocktail on ice. After scraping the cells, the lysates were collected in 1.5-mL Eppendorf tubes and subjected to sonication for lysis. Following sonication, lysates were centrifuged at 4 °C for 1 h at 13,000 rotations per min (RPM), and the resulting supernatants were collected and stored at − 80 °C until further usage.

2.6 Western blot analysis

An anti-His-tag antibody and an anti-mouse Ig:HRP secondary antibody were purchased from R&D systems (catalogue # MAB050) and ECM Bioscience (catalogue # 3001), respectively. HEK 293T cells lysates were collected as mentioned above and incubated with 5× Laemmli buffer (50 mM Tris–Cl pH 6.8, 1.6% SDS, 8% glycerol, 4% β-mercaptoethanol, and 0.04% bromophenol blue) at 95 °C for 10 min. Lysate was loaded and separated by 10% SDS-PAGE. Polyvinylidene fluoride (PVDF) membrane was used to transfer the proteins followed by blocking the membrane using 5% bovine serum albumin (BSA) for 1 h at room temperature. Blot was incubated with the anti-His antibody (1:2500 dilution in blocking buffer) at 4 °C overnight. Secondary anti-mouse IgG HRP (anti-Mouse Ig:HRP Donkey pAb, 1:5000 dilution in 5% BSA) was used to detect the His-tagged mNG-α-cat-NLuc protein.

2.7 In vitro BRET-based β-catenin-derived peptide binding assay

BRET-based peptide binding affinity in vitro assays were conducted by incubating the cell lysate containing the α-catenin plasmid construct with cell lysate containing either β-catenin-derived peptides with mSca or control (mScarlet) at a 1:10 ratio in lysis buffer in the presence of 5% polyethylene glycol (PEG) 8000 Da. BRET measurements were performed at 37 °C by adding the luciferase substrate at a dilution of 1:10,000. The Tecan SPARK multimode microplate reader was utilized to record bioluminescence intensities at 467 and 615 nm and calculate BRET ratios (615/467 nm). Total mNG and mSca fluorescence in cell lysates containing the α-catenin plasmid construct and β-catenin-derived peptides with mSca was measured by exciting the samples at 480 nm and acquiring emission at a wavelength of 530 nm and by exciting the samples at 560 nm and acquiring emission at a wavelength of 600 nm, respectively, before the addition of the luciferase substrate.

2.8 BRET-based β-catenin-derived peptide binding assay in live cells

Live cell BRET-based peptide binding affinity assays were performed by co-transfecting HEK 293 T cells in a white 96-well plate with α-catenin plasmid construct and β-catenin-derived peptides with mSca in various ratios including 0.25, 0.5, 1, 2, 3, 4, 5 as per the protocol described above. A control, pcDNA3.1-based plasmid, was also co-transfected to keep the amount of plasmid DNA in transfection constant. BRET measurement was performed at 37 °C by the addition of luciferase substrate at a dilution of 1:10,000 using Tecan SPARK® multimode microplate reader 24 h of post transfection. Spectral scanning for bioluminescence was acquired from 380 to 664 nm wavelengths with an acquisition time of 200 ms for each wavelength to determine relative emissions from mSca (acceptor) and quantify BRET as a ratio of 615/467 nm. Changes in BRET (ΔBRET) were determined after subtracting from initial value. Total mNG and mSca fluorescence in cell lysates containing the α-catenin-NLuc (fused with mNG) and β-cat-mSca peptides was measured by exciting the samples at 480 nm and emission acquired at a wavelength of 530 nm and by exciting the samples at 560 nm and emission acquired at a wavelength of 600 nm, respectively, prior to luciferase substrate addition.

2.9 Cell imaging and aspect ratio quantification

A549 cells were seeded at high density on an 8-well glass bottom plate (μ-slide 8-well, Ibidi GMBH) and allowed to form a cell monolayer [11, 39, 44, 56, 57]. Post 24 h of seeding, cells were transfected with either the pmScarlet-C1 (control) and various β-cat-mSca (WT and mutants) plasmids using Lipofectamine™ 2000 (Thermo Fisher; catalogue # 11,668,027) according to the manufacturer’s protocol. Post 24 h of transfection, cells were washed with DPBS and incubated with pre-warmed (37 °C) imaging buffer (25 mM Tris, 140 mM NaCl, 3 mM KCl, 2 M CaCl2, 1 mM MgCl2, and 5.5 mM D-glucose) for 30 min in 5% CO2. A Nikon Eclipse Ti inverted microscope was used for imaging cell monolayers using a 20 × objective. Brightfield and mScarlet (red fluorescence) images were acquired and analyzed with the Fiji/ImageJ software [58, 59]. Cells expressing the β-catenin-derived peptides were identified based on the mScarlet fluorescence, and cell peripheries were annotated manually to determine the major and minor axes and were used to calculate aspect ratio of the cells in the monolayers [60,61,62].

2.10 Data analysis and figure preparation

GraphPad Prism (Version 9 for macOS, GraphPad Software, La Jolla California USA; 141 www.graphpad.com) and Microsoft Excel were used for data analysis and preparations of the graphs and tables. Matplotlib was used to prepare MD simulation graphs [63]. Either Inkscape or Adobe Illustrator software was used for assembling the figures.

3 Results and discussion

3.1 Structural analysis revealed the possibility of increased binding affinity of mutant β-catenin-derived peptides for α-catenin

To engineer β-catenin-derived peptides that bind α-catenin with a higher affinity, we analyzed the available 3-D structure of the α- and β-catenin complex (PDB ID: 1DOW) [49] to select amino acid residue in the β-catenin peptide that do not form any interaction in the WT complex and therefore can be mutated to favorable residue that can enhance the interaction between the two proteins. α- and β-catenin interaction is mediated by amino acid residues 118–146 from the N-terminal region of β-catenin and amino acid residues 57–143 from the N-terminal region of α-catenin (Fig. 1A) [16]. Structural analysis of the complex using the PDBsum (https://www.ebi.ac.uk/thornton-srv/databases/pdbsum/) (Fig. 1B) [64, 65] revealed the presence of a total of 9 hydrogen bond interactions (formed by residues H118, I140, Y142, V122, T120, N121, and R124 from the β-catenin peptide) and a number of non-bonded interactions [20]. However, no salt bridge interactions were detected between the two proteins.

Fig. 1
figure 1

Structural analysis of β-catenin-derived peptide in complex with α-catenin and proposed mutations to enhance interaction. A Schematic showing the β-catenin-derived peptide in complex with α-catenin [PDB: 1DOW] [49]. B Interfacial interactions formed by residues in α-catenin and β-catenin. C Cartoon representations of α- and β-catenin showing additional interactions formed by indicated mutations in the β-catenin peptide

Full size image

Importantly, a region of no interaction was identified between residues 125 and 140 in β-catenin, thereby presenting a promising opportunity to introduce mutations for increasing the binding affinity between the β-catenin-derived peptides and α-catenin. Specifically, we selected residues L125, P128, M131, and L132 in the β-catenin peptide for mutagenesis into negatively charged residues (either aspartic acid or glutamic acid) to establish new polar interactions with the positively charged residues in α-catenin. Structural modeling of the mutant β-catenin-derived peptides in complex with α-catenin revealed the possibility of an additional salt bridge between E128 (P128E mutation) in β-catenin and K75 in α-catenin (Fig. 1C). Similarly, E131 (M131E mutation) and D132 (L132D mutation) in β-catenin could individually form hydrogen bond interaction with Q67 and T69 in α-catenin, respectively, which could potentially result in a stronger interaction (Fig. 1C). Together, these results highlight the possibility of generating β-catenin-derived peptides that could show a higher affinity for binding α-catenin.

3.2 Conventional MD simulations showed higher stability of P128E and P128E/M131E/L132D mutant β-catenin-derived peptide and α-catenin complex

To investigate the conformational dynamics and structural stability of β-catenin-derived peptides in complex with α-catenin, we performed conventional molecular dynamics (MD) simulations [43, 44, 46, 53, 66,67,68,69] using the available crystal structure of the α- and β-catenin complex (PDB ID: 1DOW) [49]. Previous studies have indicated that α-catenin amino acid residues 57–143 are crucial for binding to β-catenin [18]. Therefore, α-catenin residue positions 57–149 in the PDB structure, along with β-catenin residue positions 118–149, were selected for performing the MD simulations. Based on the structural analysis of interfacial residues reported above, structural models of seven different β-catenin-derived mutant peptides with one or more mutations were generated for the simulations. Importantly, the β-catenin-derived peptide forms an α-helical structure, which is essential to interact with α-catenin [70]. Therefore, to ensure that the mutations did not result in loss of the α-helical secondary structure of the peptide, secondary structure of the peptides was predicted using the online bioinformatics tool, PSIPRED, and the results showed that all mutant peptides could potentially maintain the α-helical structure (Supporting Fig. 1). All-atom, explicit solvent MD simulations were performed for a total 25 ns in triplicates at 37 °C using NAMD, and trajectories were analyzed for various features using either VMD [52] or MDAnalysis package in Python [54].

Visual inspection of the MD simulation trajectory of the complexes revealed that the peptides remained bound to α-catenin over the 25 ns of simulation with some loss or “kinking” of the α-helical structure of the peptides, especially in the case of L132D, P128E/M131E/L132D, and L125D/M131E/L132D mutant peptides (Fig. 2) [10, 17, 18]. We then determined the root-mean-square-deviation (RMSD) values of the Cα atoms of the peptides in each complex. This analysis revealed a reduced RMSD for the P128E, M131E, and P128E/M131E mutants compared to the WT peptide (Fig. 3A, B), while the L132D mutant showed an increased RMSD value compared to the WT peptide (Fig. 3A, B). However, the differences were not found to be significant (p < 0.05). Examining the root-mean-square-fluctuations (RMSF) of individual amino acid residues of the peptide in each complex showed similar residue-level fluctuations in the mutant peptides compared to the WT peptide, with most of the mutant peptides showing a reduction in the fluctuations at the C-terminal (Fig. 3C, D). Similarly, P128E, M131E, and P128E/M131E mutants showed lesser fluctuations except the residues from 145 to 149 at C-terminal for the β-catenin-derived peptides (Fig. 3D). Moreover, for the α-catenin chain, L132D and L125D/M131E/L132D mutants showed the highest fluctuations whereas P128E, M131E, and P128E/M131E mutants showed reduced fluctuations (Fig. 3E, F).

Fig. 2
figure 2

Structural evolution of β-catenin-derived peptides in complex with α-catenin. Cartoon β-catenin-derived peptides (blue) and surface α-catenin (gray) representation of β-catenin-derived peptides in complex with α-catenin showing structural evolution of the complexes in representative 25 ns of conventional MD simulations. Snapshots were captured every 5 ns

Full size image
Fig. 3
figure 3

Structural dynamics of β-catenin-derived peptides in complex with α-catenin. A, B Graphs showing backbone (Cα) root-mean-square deviation (RMSD) values (A) and bar graph showing RMSD values (mean ± S.D.) (B) of the WT and mutant β-catenin-derived peptides in complex with α-catenin from three independent MD simulation runs. p values shown were obtained from unpaired Student’s t-test against the WT β-catenin-derived peptide. C–F Graphs showing backbone (Cα) root-mean-square fluctuation (RMSF) values (C) and changes in RMSF (ΔRMSF) (D) of β-catenin-derived peptides, RMSF values (E), and changes in RMSF (ΔRMSF) (F) of α-catenin. All values were determined from 3 independent, all-atom 25 ns of MD simulations. Gray lines in the graphs show the mean of 3 runs

Full size image

We then analyzed the trajectories for the number of hydrogen bonds and salt bridge interactions as well as radius of gyration (RoG) and solvent accessible surface area (SASA). Determination of the number of hydrogen bonds formed by the various β-catenin-derived peptide and α-catenin complexes over time revealed a significantly increased number of hydrogen bonds formed by the P128E, L132D, P128E/M131E, and P128E/M131E/L132D mutant peptides suggesting an enhanced interaction between these mutant peptides and α-catenin (Fig. 4A, B). Although the number of hydrogen bonds formed by other mutants was not significantly different (p < 0.05), all of them showed a general increase in the number of hydrogen bonds. In contrast, none of the mutant β-catenin-derived peptides, except the P128E/M131E/L132D mutant, in complex with α-catenin showed a significant change (Fig. 4C, D). Analysis of the trajectories for RoG revealed a significant reduction for the L125D and P128E/M131E mutant peptides whereas P128E, M131E, and L132D showed a non-significant, but decreased RoG while the triple mutants, P128E/M131E/L132D and L125D/M131E/L132D, showed a non-significant increase in RoG (Fig. 4E, F). Further, analysis of the trajectories for SASA revealed a significant decrease in the case of P128E and M131E and an increase in the case of L132D and the triple mutants, P128E/M131E/L132D and L125D/M131E/L132D (Fig. 4G, H) [71], suggesting an enhancement in the interaction between β-catenin-derived peptides and α-catenin through the introduction of the P128E and M131E single point mutations.

Fig. 4
figure 4

Structural dynamics of β-catenin-derived peptides in complex with α-catenin. A, B Graph and bar graph showing number of hydrogen bonds (A) and the mean of number of hydrogen bond (B) formation for 3 MD simulation runs of WT and mutant β-catenin-derived peptides in complex with α-catenin. C, D Graph and bar graph showing number of salt bridges (C) and the mean of number of salt bridge (D) formation for 3 MD simulation runs of WT and mutant β-catenin-derived peptides in complex with α-catenin. E, F Graphs and bar graph showing radius of gyration (RoG) and the mean of RoG for 3 MD simulation runs of WT and mutant β-catenin-derived peptides in complex with α-catenin. G, H Graph and bar graph showing the solvent accessible surface area (SASA) and the mean of SASA for 3 MD simulation runs of WT and mutant β-catenin-derived peptides in complex with α-catenin. I Graph showing the fractional occupancy of secondary structure elements of β-catenin-derived peptides in complex with α-catenin. All values were determined from 3 independent, all-atom 25 ns of MD simulations. Gray lines in the graphs show the mean of three independent MD simulation runs. p values shown were obtained from unpaired Student’s t-test against the WT β-catenin-derived peptide

Full size image

Furthermore, the structural stability of the protein relies on secondary structural elements, including α-helix, β-sheets, and coils or loops, which play a pivotal role in determining the overall folding and stability of a protein. We, therefore, performed secondary structure analysis of the trajectories and determined the fractional occupancy of each peptide residue in any of the secondary structural states indicated a general maintenance of the α-helical structure of the WT as well as the mutant β-catenin-derived peptides over the period of the MD simulation (Fig. 4I). Finally, we analyzed the conventional MD simulation trajectories for the free energy change of binding of the β-catenin-derived peptides and α-catenin through the molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) method [72] using the CaFE 1.0 tool [73] available as a plugin in VMD [52] in order to determine affinity of the peptides for α-catenin and if there are any changes in the binding affinity of the mutant peptides in comparison to the WT β-catenin-derived peptide. This analysis revealed the lowest free energy change of binding for the P128E mutant peptide whereas most of the other peptides showed a general decrease in the free energy change of binding while the triple mutant, L125D/M131E/L132D, showed an increase in the free energy change of binding (Fig. 5). These results suggest an increased binding affinity of the P128E mutant β-catenin-derived peptide with α-catenin.

Fig. 5
figure 5

Increased binding energies of mutant β-catenin-derived peptides in complex with α-catenin. Graph showing free energy change (left panel) of β-catenin-derived peptide (WT and mutants) binding to α-catenin determined from three independent 25 ns MD simulation runs. p values shown were obtained from unpaired Student’s t-test against the WT β-catenin-derived peptide. Right panel—graph showing changes in the free energy change of β-catenin-derived mutant peptides binding to α-catenin (compared to the WT complex)

Full size image

3.3 Steered MD shows enhanced binding of the M131E and P128E/M131E/L132D mutant β-catenin-derived peptides with α-catenin

To further assess the binding of the WT β-catenin-derived peptide with α-catenin and the impact of the mutations on the binding, we performed steered MD simulations using the same complex structures as used for the conventional MD simulation reported above. This involved application of a constant pulling force on the β-catenin-derived peptides while anchoring residues in α-catenin. This well-established technique induces receptor-ligand unbinding on a specific time scale, allowing for the examination of protein–protein interactions. Steered MD simulation, thus, provides an opportunity to measure the unbinding force required to detect changes in the interaction between peptides [74]. For this, four independent, 20-ns long, all-atom simulations were performed at a constant pulling rate of 5 Å/ns for both the WT and the mutant β-catenin-derived peptides in complex with α-catenin (Fig. 6A). We then analyzed the trajectories for the center of mass distance between β-catenin-derived peptides and α-catenin over time individually for each complex (Fig. 6B). This approach aligns with the principles of single molecule Atomic Force Microscopy (AFM) experiments, which have previously demonstrated a force-dependent extension of the α- and β-catenin complex [75]. It enables a more accurate detection of alterations in interaction upon mutation under force, better mimicking the in vivo interaction between β-catenin and α-catenin under actomyosin tension [1, 76, 77]. Distance measurement revealed a general trend of a substantial increase between 10 and 15 ns of the simulations (Fig. 6B), suggesting a stable interaction between α-catenin and all β-catenin-derived peptides. Importantly, a statistically significant reduction in the distance was observed for the P128E/M131E/L132D compared to the WT at 15 ns of the simulation (p = 0.043), while none of the other mutants exhibited a statistically significant difference in the distance (Fig. 6C), suggesting the possibility of an increased affinity of the P128E/M131E/L132D triple mutant for α-catenin.

Fig. 6
figure 6

Mutant β-catenin-derived peptides show a prolonged interaction with α-catenin in SMD simulations. A Snapshots obtained from a representative 20 ns SMD simulation at a constant pulling force of 5 Å/ns showing unbinding of the WT β-catenin-derived peptide from α-catenin (cartoon representation). B Graphs showing center-of-mass distances of WT and mutant β-catenin-derived peptides and α-catenin obtained from 4 independent, all-atom 20 ns SMD simulations at a constant pulling force of 5 Å/ns. C Graph showing the distances between β-catenin-derived peptides (WT and various mutants) and α-catenin after 15 ns of SMD simulations. p values shown were obtained from unpaired Student’s t-test against the WT β-catenin-derived peptide

Full size image

3.4 In vitro BRET assays revealed similar binding of WT and mutant β-catenin-derived peptides

Based on the results obtained from the conventional and steered MD simulations, we attempted to experimentally test the binding of the WT and P128E, M131E, P128E/M131E, and P128E/M131E/L132D mutant β-catenin-derived peptides with α-catenin. For this, we selected a region of α-catenin spanning amino acid residues 57–261, as this region is known to bind with β-catenin and exhibit a binding pattern comparable to full-length α-catenin [18]. Given that Bioluminescence Resonance Energy Transfer (BRET) assays are well established for protein–protein interaction studies and for analyzing protein conformational changes [36, 38] [78], we developed a BRET-based assay to monitor interaction between β-catenin-derived peptides and α-catenin [38,39,40, 42, 44,45,46, 68]. Specifically, we generated a fusion protein containing the β-catenin binding region of α-catenin (spanning amino acid residues 57–261) and the nanoluciferase (NLuc) bioluminescent protein at the C-terminal as the BRET donor [79]. Additionally, we included mNeonGreen (mNG) fluorescent protein at the N-terminal linked through a rigid α-helical linker for monitoring the expression of the protein through fluorescence measurements (Supporting Fig. 2A, B). Further, we generated fusion proteins containing either the WT or the mutant β-catenin-derived peptides fused to the mScarlet (mSca) fluorescent protein at the C-terminal as the BRET acceptor (Supporting Fig. 2C, D, E).

HEK 293T cells were transfected with the α-catenin (residues 57–261) plasmid, and its cytosolic expression was observed using epifluorescence microscopy (Supporting Fig. 3A). Moreover, we confirmed mNG-α-catenin-NLuc protein expression by western blotting analysis using an anti-His tag antibody as the protein construct was tagged with 6 × His-tag (Supporting Fig. 3B). We also confirmed the β-catenin-derived peptide expression plasmids through restriction digestion (Supporting Fig. 2E) and the expression of β-catenin-derived peptides fused with mSca in HEK 293T cells using epifluorescence microscopy (Supporting Fig. 4). Following transfection and expression for 48 h, cell lysates were prepared from cells that were individually transfected with either the α-catenin plasmid DNA or various β-catenin-derived peptide plasmids DNA and were mixed at a ratio of 1:10 to monitor binding through the measurement of BRET over time. The mNG fluorescence (Fig. 7A) and bioluminescence (Fig. 7B) indicated the presence of similar amounts of the α-catenin (residues 57–261) protein. In the same way, the mSca fluorescence arising from the β-catenin-derived peptides fused with mSca was found to be similar (Fig. 7C), suggesting similar levels of expression of the β-catenin derived peptides in the lysates. Incubation of the WT β-catenin-mSca expressing cell lysates with α-catenin-NLuc expressing cell lysates resulted in a time-dependent increase in BRET signal, which was not seen with the control (mSca alone) expressing cell lysates (Fig. 7D, E), suggesting a specific binding of the WT β-catenin peptide with α-catenin. Additionally, the P128E, M131E, and P128E/M131E double mutant showed similar increases in BRET (Fig. 7D, E). Importantly, however, the P128E/M131E/L132D triple mutant failed to show any increases in BRET (Fig. 7D, E). We then analyzed the BRET kinetic profiles to determine the maximum change in BRET, rate of change in BRET (k), and half time (t1/2) required for the BRET change (Fig. 7F–H). These analyses revealed similar changes in the BRET (maximum BRET change) (Fig. 7F), rate of BRET change (Fig. 7G), and the t1/2 of the BRET change (Fig. 7H), suggesting a similar affinity of the WT and P128E, M131E, and P128E/M131E mutant β-catenin derived peptides for α-catenin.

Fig. 7
figure 7

Monitoring the interaction of α-catenin with β-catenin-derived peptides using a BRET-based assay in vitro. A–C Graphs showing the total mNG fluorescence (A) and bioluminescence (B) of α-catenin plasmid construct and mSca fluorescence of β-catenin-derived peptides (C). D Graph showing the BRET ratio (ratio emission at 533 and 467 nm) of the interaction of α-catenin plasmid construct with β-catenin-derived WT and mutant peptides. E Graph showing the change in BRET over time as compared to the 0 s value for each sample. F–H Graphs representing the maximum change in BRET (F), rate of the reaction (G), and t1/2 for each sample. ND—not determined. Data shown are mean ± S.D. obtained from three independent experiments. p values shown were obtained from unpaired Student’s t-test against the WT β-catenin-derived peptide. ND, not determined

Full size image

3.5 Live cell BRET assay to monitor binding of β-catenin-derived peptides with α-catenin

Though in vitro methods are valuable to investigate protein–protein interaction, they often overlook the intrinsic complexity of living cells [80]. To address this limitation, we sought to detect interactions between the WT and mutant β-catenin-derived peptides and α-catenin in live cells. For this, we co-transfected HEK 293T cells with a constant amount of the α-catenin plasmid DNA and increasing amounts of the mSca (control) and WT and mutant β-catenin-derived peptide-mSca plasmid DNA. Post 24 h of transfection, we determined mNG and mSca fluorescence, corresponding to levels of the α-catenin-NLuc and various β-catenin-derived peptides in the cells, respectively. BRET, which corresponds to the binding of the peptides with α-catenin, was determined from the bioluminescence spectra obtained from the same cells upon addition of the luciferase substrate. We then analyzed the live cell data by plotting BRET against the ratio of mSca:mNG fluorescence at each β-catenin-derived peptide-mSca plasmid DNA amount for all peptides. This analysis revealed a high affinity interaction between the WT β-catenin-derived peptide and α-catenin with a rate constant (k) of 0.20 ± 0.20 (unit: BRET increase per mSca:mNG ratio) while no large changes in BRET were observed for the mSca control construct (Fig. 8A, B). Additionally, both P128E and M131E single mutants showed large increases in BRET with similar rate constants of 0.16 ± 0.05 and 0.12 ± 0.03, respectively (Fig. 8A, B), suggesting an enhanced interaction between these peptides and α-catenin. On the other hand, the double mutant, P128E/M131E, showed a somewhat reduced BRET at similar mSca/mNG ratios as well as reduced rate constant of 0.05 ± 0.03 (Fig. 8A, B), suggesting a reduced affinity of this mutant peptide for α-catenin compared to the WT peptide. None of these mutants have any statistically significant difference in rate constant (k) compared to WT (Fig. 8A, B). Finally, no significant changes were observed for the triple mutant P128E/M131E/L132D peptide (Fig. 8A, B), suggesting a loss of binding of this peptide with α-catenin.

Fig. 8
figure 8

BRET-based monitoring of the interaction of α-catenin with β-catenin-derived peptides in living cells. A Graph showing the BRET ratio of α-catenin plasmid construct upon increasing expression of β-catenin-derived peptides in living cells. Changes in BRET are plotted against the ratio of mSca (from β-catenin-derived peptides) to mNG (from α-catenin plasmid construct) fluorescence. B Graphs representing the rate of the reaction for each sample. ND, not determined. Data shown are mean ± S.D. obtained from three independent experiments with each experiment performed in triplicates

Full size image

3.6 Expression of β-catenin-derived peptides alters cell aspect ratio in a cell monolayer assay

Based on the results reported above on the binding of the β-catenin-derived peptides with α-catenin, we investigated if the expression of the peptides can cause a change in the morphology of the cells in a monolayer. This is because the interaction between β-catenin and α-catenin is required for the assembly and maintenance of cell–cell adhesion. Changes in intercellular adhesiveness result in morphological alterations that serve as a hallmark in various malignancies [81]. We posited that the interaction of β-catenin and α-catenin will allow the assembly and maintenance of cell–cell adhesion between cells in the monolayer resulting in a more “rounded” shape with a smaller aspect ratio (ratio of the long and the short axis determined from the cell shape) of the cells [60,61,62]. On the other hand, a disruption of the interaction between β-catenin and α-catenin due to the binding of the β-catenin-derived peptides to α-catenin will result in a disruption of cell–cell adhesion leading to a more “elongated” shape with a higher aspect ratio of the cells [60,61,62]. For this, we seeded A549 cells at high density to form a monolayer and transfected the cells with either mScarlet or β-catenin-derived peptide fused to mScarlet (WT or mutant) expressing plasmid DNA and determined the aspect ratio of the cells in the monolayer post 24 h of transfection from fluorescent images of the cell monolayers (Fig. 9A). Analysis of the fluorescent images to determine the aspect ratio of cells expressing the fluorescent proteins revealed a significant difference between the control (mScarlet expressing cells) and the WT β-catenin-derived peptide (1.859 ± 0.5914 vs. 1.959 ± 0.7879, respectively; p = 0.005) (Fig. 9B). Similarly, expression of the M131E mutant peptide resulted in a significantly higher aspect ratio (1.960 ± 0.9446; p = 0.012) compared to the control cells (Fig. 9B). However, expression of the P128E, P128E/M131E, and P128E/M131E/L132D mutant peptide did not result in a significant change in the aspect ratio of the cells (1.798 ± 0.6845, 1.902 ± 0.8237, 1.861 ± 0.9312, respectively; p = 0.061, 0.242, and 0.962) compared to the control cells (Fig. 9B). These results are largely in agreement with the BRET results presented above and suggest that the expression of the WT and the M131E mutant β-catenin-derived peptides can bind to endogenously expressed α-catenin in the cells and potentially interfere with β-catenin-α-catenin interaction and, in turn, can impact cell–cell adhesion in a cell monolayer.

Fig. 9
figure 9

Effects of β-cat-derived peptides on the aspect ratio of cells in a monolayer. A Brightfield and epi-fluorescence images of A549 cell monolayers expressing either the mSca (control) or β-catenin-derived peptides (WT and mutants) fused with mSca. Scale bar, 10 μm. B Graph showing the aspect ratio of cells. Data shown are mean ± S.D. obtained from data obtained from three independent experiments. p values shown were obtained from unpaired Student’s t-test against the control

Full size image

4 Conclusion

In conclusion, the analysis of the β-catenin-α-catenin complex three-dimensional structure allowed us to generate a high-affinity α-catenin binding peptide (WT) and generate mutations that could potentially increase the affinity of the WT β-catenin-derived peptides with increased interaction between the peptide and α-catenin. A combination of conventional and steered MD simulations of the β-catenin-derived peptide and α-catenin complexes revealed a differential structural dynamics and inter-molecular interactions for the WT and the mutant peptides. In vitro and live cell BRET experiments validated the binding of the WT β-catenin-derived peptide and the P128E and M131E mutants showed an enhanced interaction with α-catenin. Indeed, expression of the WT and M131E mutant β-catenin-derived peptides resulted in a change in the aspect ratio of the cells in monolayers, suggesting that these peptides can be utilized for disrupting cell–cell adhesion. We believe that the β-catenin-derived peptides reported in this manuscript can be utilized to further develop inhibitors for blocking β-catenin and α-catenin interaction and engineered to develop peptide-based α-catenin degraders for inhibiting E-cadherin-mediated cell–cell adhesion with potential applications towards abrogating metastasis of the subset of breast cancer cells that require E-cadherin-mediated adhesion for metastasis.