Structural mapping of PEAK pseudokinase interactions identifies 14-3-3 as a molecular switch for PEAK3 signaling

Abstract

PEAK pseudokinases regulate cell migration, invasion and proliferation by recruiting key signaling proteins to the cytoskeleton. Despite lacking catalytic activity, alteration in their expression level is associated with several aggressive cancers. Here, we elucidate the molecular details of key PEAK signaling interactions with the adapter proteins CrkII and Grb2 and the scaffold protein 14-3-3. Our findings rationalize why the dimerization of PEAK proteins has a crucial function in signal transduction and provide biophysical and structural data to unravel binding specificity within the PEAK interactome. We identify a conserved high affinity 14-3-3 motif on PEAK3 and demonstrate its role as a molecular switch to regulate CrkII binding and signaling via Grb2. Together, our studies provide a detailed structural snapshot of PEAK interaction networks and further elucidate how PEAK proteins, especially PEAK3, act as dynamic scaffolds that exploit adapter proteins to control signal transduction in cell growth/motility and cancer.

Fig. 1. PEAK domain organization and interaction motifs in N-terminal IDR.

a Domain organization of the PEAK family and diagram of PEAK homodimer arrangement showing the SHED, PsK and IDR motifs identified (boxed). b Sequence alignment of the N-terminal IDR of human PEAK3 with the corresponding region of PEAK1 and PEAK2. c Multiple sequence alignment (MSA) of PEAK vertebrate orthologues highlighting short linear interaction motifs (SLiMs) identified in regions of high sequence conservation, including a pY/SH2 motif (Grb2^SH2/CrkII^SH2) and the tandem site encompassing a proline-rich motif (CrkII^NSH3) and conserved putative 14-3-3 motif. d Schematic showing overall domain organization of Grb2, CrkII and 14-3-3, highlighting sequence motifs.

Fig. 2. Structural and biophysical analysis of PEAK/Grb2^SH2 interaction.

a TYSNL site of PEAK3^FL (Y²⁴) and PEAK1^IDR1 (Y¹¹⁰⁷) can be phosphorylated by Src (MS/MS analysis of tryptic peptides following in vitro kinase assay of PEAK3^FL and PEAK1^IDR1 are available in source data). Both Abl and Src can phosphorylate CrkII^Y221, a known Abl phosphorylation site, demonstrating that both Abl and Src kinases are active. b Zoom in structure of the PEAK SH2-pY peptide bound to the Grb2^FL dimer, with the PEAK phosphopeptide shown in yellow and Grb2^FL Chain B (peptide bound chain) shown in surface representation, colored by electrostatic surface potential calculated using the APBS Electrostatics plugin for PyMOL (Schroedinger) (blue = positive, white = hydrophobic, red = negative). Inset shows a 2Fo-Fc electron density map (blue mesh, contoured at 1.0σ) showing the final modeled PEAK3/PEAK1 consensus phosphopeptide (TpYSNLGQ, yellow sticks, underlined residues modeled). See also Supplementary Fig. 2. c, d PEAK3^23–38 and PEAK1^1106–1121 regions alongside full MSA/Web Logo of orthologs for the corresponding sequence, showing a section of high sequence conservation (pale yellow box) adjacent to the TYSNL SH2 site (yellow box, underlined) including a region predicted in AF2 to adopt helical secondary structure (red box). e Tabulated SPR-determined mean steady-state binding affinity values (K_D) for PEAK phosphotyrosine peptides (consensus pY-SH2 motif and extended peptides PEAK3^23-38pY²⁴ and PEAK1^1106–1121-pY¹¹⁰⁷) binding to CrkII^SH2, full-length CrkII (CrkII^FL) monomer or dimer, Grb2^SH2 or full length Grb2 (Grb2^FL). Values are colored as a heat map to illustrate trends in affinity. Data represent mean values from n = 3 independent titrations. See Supplementary Fig. 2d for representative SPR sensorgrams and Supplementary Data 1 for full SPR tabulated data and sensorgrams.

Fig. 3. Structural and biophysical analysis of PEAK/CrkII^NSH3 interaction.

a SPR data summarizing measured steady-state binding affinity (K_D) of synthetic PEAK PRM peptides toward immobilized CrkII^NSH3 domain. b Depiction of the published CrkII^NSH3:Abl⁷⁵⁸ structure (PDB: 5IH2), showing key residues within the CrkII^NSH3 consensus motif in Abl⁷⁵⁸ (peptide shown in gold stick representation) crucial for high affinity binding to CrkII^NSH3 (main chain light blue; CrkII^NSH3 surface is colored by electrostatic surface potential calculated using UCSF Chimera v 1.16; blue = positive, white = hydrophobic, red = negative). Aligned is the sequence of PEAK3^54–66, showing key conserved residues of the CrkII^NSH3 motif. c, d Interaction studies with recombinant PEAK1^IDR1/PEAK2^IDR1 and CrkII^FL underscore the role of avidity for high affinity binding of CrkII^FL to PEAK dimers; c Incubation of PEAK2^IDR1 with CrkII^FL dimer at a 1:1.3 ratio results in a complex formation while (d) incubation with CrkII^FL monomer does not result in a complex, as confirmed by SDS-PAGE analysis of SEC eluted fractions (n = 3, independent experiments). Plots of the CrkII^FL dimer and monomer are shown in light blue; PEAK2^IDR1 in pink; and the complex of PEAK2^IDR1 dimer:CrkII^FL dimer in black (see Supplementary Fig. 3d for PEAK1^IDR1 dimer:CrkII^FL dimer complex).

Fig. 4. PEAK3:14-3-3 forms a stable high affinity heterocomplex.

a Recombinant PEAK3^FL from insect cells elutes on size exclusion chromatography (SEC; S200 10/300) as a high affinity stoichiometric complex of dimeric PEAK3 with a 14-3-3ε,ζ heterodimer, results supported by SDS-PAGE analysis of eluted fractions (n = 3, independent experiments) (see also Supplementary Fig. 4a, source data). b Immunoprecipitation of PEAK3 WT and PEAK3 S69A showing S69 is required for 14-3-3 co-immunoprecipitation from cells (n = 3 biologically independent samples, see Supplementary Fig. 4a for biological repeats). c SEC-MALS analysis of PEAK3^FL−14-3-3ε,ζ heterodimer complex confirming the experimentally determined mass closely matches the mass of 155 kDa expected for a stoichiometric dimer:dimer complex (n = 3, independent experiments). d SEC profile (S200 10/300) of recombinant PEAK3^FL−14-3-3ε,ζ heterodimer pre-incubated with CrkII^FL dimer in a 1:1.2 molar ratio showing no complex formation (n = 3, independent experiments). Source data are provided as a source data file.

Fig. 5. Structural and biophysical analysis of PEAK3/14-3-3 interaction.

a Binding of PEAK3^tandem-pS69 (left) and PEAK3^tandem-S69 peptides (right) to 14-3-3ɣ as measured by SPR. b Overall structure of 14-3-3ε: PEAK3^tandem-pS69 peptide showing 14-3-3 dimer antiparallel arrangement with each monomer (pink cartoon/surface) bound to a single copy of the PEAK3^tandem-pS69 peptide (yellow sticks). Chain A of 14-3-3ε is in dark pink and chain B is in light pink. c Zoom in highlighting peptide groove and showing the central pS69 residue of the PEAK3 phosphopeptide interacting with K50, R57, Y131 and R130. d Unbiased Fo-Fc omit map (green mesh, contoured at 3.0σ) showing peptide density prior to modeling and final modeled PEAK3^tandem-pS69 phosphopeptide (yellow sticks). e Superposition of the peptide modeled in each 14-3-3 monomer (chains A–D, sticks) in the asymmetric unit.

Fig. 6. 14-3-3ɣ binding to PEAK3^tandem peptide reduces the affinity of CrkII^NSH3 binding to the adjacent CrkII motif.

a Sequential ITC binding studies using purified 14-3-3ɣ and purified CrkII^NSH3 domain and PEAK3^tandem-S69 non-phosphorylated and b PEAK3^tandem-pS69 phosphorylated peptides. b High affinity binding of 14-3-3ɣ to PEAK3^tandem-pS69 peptide reduces the affinity of CrkII^NSH3 binding to the adjacent CrkII motif at the tandem site (negative cooperativity). See Supplementary Fig. 6a–d and Supplementary Data 5 for all ITC data.

Fig. 7. S69A mutant increases PEAK3 p-Tyr and cell motility.

a S69A mutant increases PEAK3 p-Tyr. MCF-10A expressing WT and S69A PEAK3 cells were starved in –EGF medium overnight. Anti-HA IPs were prepared from denatured cell lysates and Western blotted as indicated. Data represent the mean ± S.E.M. of n = 3 independent experiments. **p < 0.01 (p = 0.0078), by ratio paired t-test, two tailed (Repeats shown in source data). b Effect of S69A PEAK3 on random cell motility in MCF-10A cells. Vector, HA-tagged WT or S69A PEAK3 were stably overexpressed in MCF-10A cells and random cell motility was determined by live cell tracking. Representative images are shown. Scale bar, 200 μm. The histogram indicates the average maximum displacement distance from the origin. ~40 cells were analyzed in each experiment. Data represent the mean ± S.E.M. of N = 3 independent experiments. ****p < 0.0001 by one-way ANOVA with Dunnett’s multiple comparisons test (Vector vs. WT, p = 0.00090; WT vs. S69A, p = 0.000003), (Repeats shown in source data). Source data are provided as a Source Data File.

Fig. 8. Unified SPR analysis of PEAK protein interactions and proposed model for PEAK3/Grb2 and PEAK3/CrkII regulation via the tandem site 14-3-3 motif.

a SPR binding contributions were measured for PEAK interactors (14-3-3ɣ, Grb^FL and CrkII^FL and sub-domains) toward immobilized PEAK proteins with or without phosphorylation by Src. Tabulated mean SPR affinity values (K_D) are colored as a heat map to illustrate trends in affinity. Gray cells represent K_D > 50 µM or no detectable binding. b Representative SPR sensorgrams for key interactions of PEAK3^FL. c A unifying model of PEAK3 interactions. Avidity stabilizes PEAK3/CrkII interaction with clustered CrkII (interacting via the CrkII^NSH3 and tandem motif PRM of the PEAK3 dimer). CrkII is likely to be clustered by interaction with other adapters/scaffolds, such as SH2 domain-mediated binding to partner proteins containing multiple adjacent pYxxP motifs (e.g., present in known Src substrates p130cas/NEDD9, paxillin), but possibly also via dimerization (e.g., SH2-mediated as we identify for recombinant protein). Phosphorylation of PEAK3 at S69 creates a high-affinity binding site for 14-3-3, leading to the formation of a highly stable PEAK:14-3-3 dimer:dimer. Phosphorylation/binding of 14-3-3 to PEAK3 destabilizes CrkII binding in the adjacent tandem site (negative cooperativity), disfavoring PEAK3/CrkII complexes. Possible outcomes are termination of PEAK3/CrkII signaling and/or alter PEAK3 sub-cellular localization. CrkII may facilitate colocalization of PEAK3 to regions with active Src kinase (e.g., p130cas at focal adhesions). Phosphorylation of PEAK3 at Y24 (e.g., by Src kinase) recruits Grb2 via its SH2 domain leading to downstream signaling that alters cell motility (e.g., via association with ASAP1 and PYK2 and potential PYK2/ASAP1-mediated activation of the PI3K/AKT pathway). Thus, PEAK3 S69A mutation abrogates 14-3-3 binding and promotes CrkII association and PEAK Y24 phosphorylation/Grb2 signaling, culminating in enhanced in cell motility. d Illustrative representation of a potential dimeric human PEAK3^FL:14-3-3ε complex to show relative scale and key sites of interaction. Depicted is the 14-3-3ε:PEAK3^tandem-pS69 crystal structure (PDB 8DGP) and a structural model of the PEAK3^SHED-PsK dimer prepared from AlphaFold2 multimer modeling of a human PEAK3^FL dimer^,. Unmodelled IDR linker regions are depicted as dotted lines.