The intracellular domains of the EphB6 and EphA10 receptor tyrosine pseudokinases function as dynamic signalling hubs

Abstract

EphB6 and EphA10 are two poorly characterised pseudokinase members of the Eph receptor family, which collectively serves as mediators of contact-dependent cell-cell communication to transmit extracellular cues into intracellular signals. As per their active counterparts, EphB6 and EphA10 deregulation is strongly linked to proliferative diseases. However, unlike active Eph receptors, whose catalytic activities are thought to initiate an intracellular signalling cascade, EphB6 and EphA10 are classified as catalytically dead, raising the question of how non-catalytic functions contribute to Eph receptor signalling homeostasis. In this study, we have characterised the biochemical properties and topology of the EphB6 and EphA10 intracellular regions comprising the juxtamembrane (JM) region, pseudokinase and SAM domains. Using small-angle X-ray scattering and cross-linking-mass spectrometry, we observed high flexibility within their intracellular regions in solution and a propensity for interaction between the component domains. We identified tyrosine residues in the JM region of EphB6 as EphB4 substrates, which can bind the SH2 domains of signalling effectors, including Abl, Src and Vav3, consistent with cellular roles in recruiting these proteins for downstream signalling. Furthermore, our finding that EphB6 and EphA10 can bind ATP and ATP-competitive small molecules raises the prospect that these pseudokinase domains could be pharmacologically targeted to counter oncogenic signalling.

Keywords: Eph receptors; kinase inhibitors; molecular scaffolds; molecular switches; pseudokinases; signalling.

Figure 1.. EphB6 and EphA10 pseudokinase domains bind nucleotides and protein kinase inhibitors.

(A) An exemplar crystal structure of an Eph receptor kinase domain (EphA3; PDB: 2QO2 [48]), with the conserved motifs (residues) responsible for kinase activity (as tabulated in panel (B)) highlighted. (B) A schematic diagram showing the intracellular domain/motifs of EphB6/EphA10, which contains the juxtamembrane (JM) region, the pseudokinase domain (PsKD), the sterile-α-motif (SAM) domain and the C-terminal PDZ domain-binding motif (PBM). The sequences of the N-JM and partial C-JM of EphB6, EphA10 and EphB4 are highlighted, with the JX1 and JX2 positions coloured in brown. The underlined tyrosine residues in the N-JM region are conserved in all and only Type B Eph receptors. In the table, the conserved catalytic motifs (residues) required for protein kinase activity, and corresponding residues in the pseudokinase domains of EphB6 and EphA10 are listed. (C) The estimated K_D for ATP binding to EphB6 is 94 μM. (D) The estimated K_D for ATP binding to ΔN-EphA10 is 95 μM. Titrations of type I (Dasatinib) kinase inhibitor for binding to EphB6 (E), ΔN-EphA10 (F). Titrations of type II (AMG-Tie 2-1) kinase inhibitors for binding to EphB6 (G), ΔN-EphA10 (H). Titrations were done in triplicate and error bars represent the standard error of the mean (SEM). At least two independent experiments were carried out for each titration, and representatives are shown here.

Figure 2.. The EphB6 and EphA10 intracellular domains are monomeric and exhibit elongated conformations in solution.

(A) Recombinant EphB6 (residues 625–1021), ΔN-EphB6 (residues 643–1021), EphA10 (residues 600–1008) and ΔN-EphA10 (residues 618–1008) elution profiles from an S200 Increase 10/300 GL gel filtration column. The following molecular mass markers, Thyroglobulin (670 kDa), bovine γ-globulin (158 kDa), chicken ovalbumin (44 kDa), equine myoglobin (17 kDa), and vitamin B12 (1.35 kDa) (Bio-Rad) were used to estimate the molecular mass of the eluted proteins. (B) The ΔN-EphB6 SAXS profile compared with the theoretical scatter of the AlphaFold EphB6 structure co-ordinates (EphB6 entry: O15197, residues 643–1021) by CRYSOL (χ² = 0.129). The Guinier plot (inset) shows monodispersed proteins in solution, free from aggregation and inter-particle interaction. (C) The ΔN-EphB6 + AMPPNP SAXS profile compared with the theoretical scatter of the EphB6 structure co-ordinates predicted by AlphaFold (EphB6 entry: O15197, residues 643–1021) by CRYSOL (χ² = 0.130). The Guinier plot (inset) shows monodispersed proteins in solution, free from aggregation and inter-particle interaction. (D) The EphB6 SAXS profile fitted to the theoretical scatter of the AlphaFold EphB6 structure co-ordinates (EphB6 entry: O15197, residues 625–1021) by CRYSOL (χ² = 0.309). The Guinier plot (inset) shows monodispersed proteins in solution, free from aggregation and inter-particle interaction. (E) The EphA10 SAXS profile fitted to the theoretical scatter of the AlphaFold EphA10 structure co-ordinates (EphA10 entry: Q5JZY3, residues 600–1008) by CRYSOL (χ² = 0.220). The Guinier plot (inset) shows monodispersed proteins in solution, free from aggregation and inter-particle interaction. Pairwise distribution analysis is shown for ΔN-EphB6 (F), ΔN-EphB6 + AMPPNP (G), EphB6 (H) and EphA10 (I), with the corresponding structures of EphB6 and EphA10 from AlphaFold presented.

Figure 3.. EphB6 intracellular domains exhibit dynamic inter-domain interaction by chemical cross-linking-mass spectrometry experiments.

(A) Reducing SDS–PAGE of the EphB6 recombinant proteins cross-linked by a concentration series of DMTMM and BS³ stained with Simply Blue SafeStain. (B) The amino acid residues cross-linked by DMTMM (0 Å spacer) are highlighted in cyan, and the amino acid residues cross-linked by BS³ (11.4 Å spacer) are highlighted in magenta. The distance of the backbone carbon and the secondary structures in the AlphaFold EphB6 structure (EphB6 entry: O15197, residues 625–1021) are highlighted in light yellow. C′: the carboxylic end of proteins. (C) The cross-linked amino acid residues are connected by cyan lines (cross-linked by DMTMM), and by magenta lines (cross-linked by BS³) in the Alphafold EphB6 structure (EphB6 entry: O15197, residues 625–1021).

Figure 4.. The EphB6 C-terminal JM region is the site of phosphorylation by EphB4.

(A) EphB6 recombinant proteins (WT and the YY → FF (JX1 and JX2) mutant) were phosphorylated by recombinant EphB4 (WT and the YY → FF (JX1 and JX2) mutant) in vitro. The reaction mixture was resolved by reducing SDS–PAGE and probed using an anti-phosphotyrosine antibody (Clone: 4G10). (B) Different phosphorylated EphB6 recombinant proteins: ΔN-EphB6 (P_L: low level of phosphorylation, P_H: high level of phosphorylation) were separated by anion-exchange chromatography (MonoQ column). (C) The phosphorylation level of a selected phosphopeptide including pY651 (JX2) and a phosphopeptide including pY669 identified is shown by the relative abundance of phosphorylation in the non-phosphorylated, P_L- and P_H-EphB6 measured by mass spectrometry. (D) SEC profiles from an S75 10/300 GL gel filtration column of phosphorylated ΔN-EphB6 recombinant proteins demonstrating that they remain monomeric in solution.

Figure 5.. Binding of SH2 domains to the phosphorylated EphB6 (P_H-ΔN-EphB6) measured by surface plasmon resonance (SPR).

Binding of P_H-ΔN-EphB6 to the SH2 domains of Abl (A), Vav2 (B), Vav3 (C), Nck1 (D), Nck2 (E), CrkII (F), CrkL (G), Grb7 (H) and Grb10 (I). The immobilisation level of P_H-ΔN-EphB6 on the SA sensor chip was 1705.1 RU.

Figure 6.. Binding of Abl SH2 domain to different phosphorylated EphB6 peptides measured by surface plasmon resonance (SPR).

Binding of Abl SH2 domain to pJX2 EphB6 peptide (A), pJX1 EphB6 peptide (B) and p644, pJX1 and pJX2 tri-phospho EphB6 peptide (C). (D) No binding was observed for the non-phosphorylated EphB6 peptide and Abl SH2 domain. (E) p669 EphB6 peptide binding to Abl SH2 domain. (F) No binding was observed for the non-phosphorylated EphB6 peptide and Abl SH2 domain. The immobilisation level of Abl SH2 domain on the CM5 chip was 1599.5 RU.

Figure 7.. Proposed model for signalling by the EphB6 pseudokinase.

The ectodomains of EphB6 and EphB4 can potentially dimerise and oligomerise following the binding of ephrin ligands. Upon dimerisation, the EphB4 kinase domain undergoes autophosphorylation, and then phosphorylates the JM region of EphB6. The phosphorylated EphB6 JM region can then serve as a docking site for downstream SH2 domain-containing signalling adaptors, such as Src, Abl and Vav3. Therefore, phosphorylated EphB6 can promote the assembly of specific signalling hub depending on the SH2 domain-containing proteins recruited. However, whether EphB6 can also potentially homodimerise/oligomerise upon binding to ephrin ligands and amplify the recruitment of downstream interactors to elicit EphB6 specific driven signalling output remains to be investigated. The crystal structure of the EphB6 ectodomains (PDB: 7K7J [69]) was used to generate the schematic illustration of the EphB6 ectodomains. The crystal structure of the EphA2 ectodomains (PDB: 3FL7 [51]) was used to generate the schematic illustration of the EphB4 ectodomains. SH2: SH2 domain-containing proteins. JM: juxtamembrane region. PsKD: pseudokinase domain. KD: kinase domain. SAM: sterile-α-motif domain. PBM: PDZ domain-binding motif.