Previous work has shown that the Afadin residue Y1230 is hyperphosphorylated in PTPRK KO epithelial cells, suggesting that this is a tyrosine targeted for dephosphorylation by PTPRK (Fearnley et al., 2019). To test this, we generated an antibody against a synthetic tyrosine phosphorylated peptide corresponding to amino acids 1224–1235 of Afadin (Figure 1—figure supplement 2A). Consistent with our previous tyrosine phosphoproteomics data, a band at the predicted molecular weight of the longest isoform of Afadin is detected at higher levels in PTPRK KO compared to WT MCF10A cells (Figure 1—figure supplement 2B). This band is increased by pervanadate (Huyer et al., 1997), a cell treatment that globally increases protein tyrosine phosphorylation, and is decreased by knockdown of Afadin using siRNA (Figure 1—figure supplement 2C). It should be noted that phosphorylation appears to interfere with the signal when using an anti-Afadin (epitope: 1091–1233) antibody from BD Transduction Laboratories on immunoblots, resulting in decreased intensity when Afadin is phosphorylated (Figure 1—figure supplement 2A and C). To further validate the site specificity of the Afadin-pY1230 antibody, we generated tyrosine to phenylalanine (YF) mutants of mScarlet-tagged Afadin. The antigenic phosphopeptide used as an epitope possesses two tyrosine residues, Y1226 and Y1230, which we mutated alone or in combination. mScarlet-Afadin phosphorylation, as detected by our antibody, increased with pervanadate treatment and was completely reversed by calf intestinal phosphatase treatment (Figure 1—figure supplement 2D). However, there was no detectable increase in signal upon pervanadate treatment of mScarlet-Afadin-Y1230F mutants, supporting that this is the predominant phosphosite recognized by our Afadin phosphoantibody. We previously demonstrated that the PTPRK ICD, but not the PTPRK D1 domain alone, dephosphorylates Afadin (Fearnley et al., 2019). This is recapitulated using the pY1230 Afadin antibody after incubation of recombinant domains with pervanadate-treated, DTT-quenched, MCF10A lysates (Figure 1B, Figure 1—figure supplement 2E). Over a time course, PTPRK-ICD shows greater activity against Afadin pY1230 than PTPRK-D1. In contrast, p120^Cat was dephosphorylated equally by the two domains, with minimal dephosphorylation of the non-substrate control, paxillin-pY118. Finally, the observed hyperphosphorylation of Afadin Y1230 in PTPRK KO MCF10A cell lysates could be reduced by doxycycline-induced expression of PTPRK to wildtype levels, but not by a catalytically inactive mutant (Figure 1—figure supplement 2F and G). Together, these data confirm that this new antibody is specific to Afadin phosphorylated at Y1230 and that pY1230 is targeted for dephosphorylation by PTPRK in a D2 domain-dependent manner. This pTyr residue is located in the C-terminal half of Afadin within an extended region of predicted disorder (Figure 1A). We next sought to determine how Afadin is recognized by the PTPRK D2.

The large size and predicted disorder of Afadin make it poorly suited to recombinant expression in Escherichia coli. Therefore, to assay direct binding of Afadin to PTPRK, recombinant full-length Afadin was expressed in vitro using a wheat germ-based cell-free translation system. Bacterially expressed and biotinylated PTPRK intracellular domain (ICD, Figure 1—figure supplement 3) was immobilized on streptavidin beads for pull-downs from wheat germ lysate containing full-length Afadin. We observed clear binding of full-length Afadin to the PTPRK-ICD (Figure 1C), consistent with a direct interaction. To determine whether PTPRK binds a region of Afadin incorporating tyrosine 1230, pull-downs were performed using C-terminal truncations of Afadin produced in wheat germ lysate. Deletion of the Afadin C-terminal disordered tail was sufficient to abolish binding to PTPRK (Figure 1D, aa. 1–1097). However, additional truncations showed substantial nonspecific binding to the streptavidin resin hindering further interaction mapping using this strategy (e.g., aa. 1–972, Figure 1D).

We next mapped PTPRK binding to the Afadin C-terminus in more detail using GST-fusion proteins (Figure 1E). A GST-fusion construct encompassing the entire C-terminal disordered region of Afadin, shown to be important in the wheat germ system, also showed direct binding to PTPRK, supporting the identification of this region as sufficient for binding (Figure 1E). Several truncation constructs within this region were tested for binding to PTPRK, clearly identifying that only constructs containing a short, 63-residue predicted coiled-coil (CC) region spanning residues 1393–1455 was able to bind PTPRK (Figure 1E). Therefore, this CC region is both necessary and sufficient for binding to the PTPRK-ICD. The sequence composition of this region of Afadin (Afadin-CC) includes a high proportion of charged residues (Figure 1F), suggesting that the PTPRK–Afadin interaction could be electrostatically mediated. In support of this, the PTPRK–Afadin interaction was sensitive to salt concentration, with less GST-Afadin-CC present in PTPRK pull-downs at increasing NaCl concentrations (Figure 1F). Further, PTPRK-ICD and Afadin-CC have complementary predicted isoelectric points (pI = 5.3 and 10.2, respectively), suggesting that these fragments would retain charge complementarity at physiological pH. Together, these data show that the PTPRK-ICD interacts directly and specifically with the C-terminal Afadin-CC region.

CC domains consist of heptad repeats (designated a–g), in which residues ‘a’ and ‘d’ are generally nonpolar, forming a hydrophobic seam that facilitates oligomerization (Truebestein and Leonard, 2016). The primary sequence of the Afadin-CC region is highly charged and does not maintain this pattern of hydrophobic residues (Figure 1F), suggesting that it may not be a bona fide CC. To test whether the Afadin-CC forms higher oligomers, the GST tag was removed (Figure 2—figure supplement 1) and the Afadin-CC alone was analyzed by size-exclusion chromatography coupled to multi-angle light scattering (SEC-MALS). This technique allows for the direct measurement of the molar mass of a particle in solution and demonstrated that the Afadin-CC is a monomer, with a molecular mass of ~8 kDa (predicted mass = 8.3 kDa, Figure 2A). The chromatography trace also revealed a small amount of truncated Afadin-CC, suggesting that following removal of the GST tag this region may be prone to some degradation. To further characterize the interaction of the PTPRK-ICD with Afadin-CC, isothermal titration calorimetry (ITC) experiments were performed to measure the binding affinity and stoichiometry. Afadin-CC (without the GST tag) forms an equimolar complex with PTPRK-ICD with a binding affinity in the low micromolar range (K_D = 3.2 ± 0.6 µM, Figure 2B and E). This binding affinity is consistent with a reversible, transient protein–protein interaction as expected for an enzyme-substrate complex. The high ΔH of this interaction (–17.6 ± 0.7 kcal/mol) indicates it is enthalpically driven, consistent with the highly charged nature of the Afadin-CC region.

Figure 2 with 1 supplement see all

Download asset Open asset

Figure 2—source data 1

Uncropped, unedited gels for Figure 2C.

https://cdn.elifesciences.org/articles/79855/elife-79855-fig2-data1-v2.zip

Download elife-79855-fig2-data1-v2.zip

Previous work highlights a critical role for the D2 pseudophosphatase domain in the specific interaction of Afadin with R2B family receptors (Fearnley et al., 2019; Hay et al., 2020). To investigate the contribution of the D2 domain to Afadin binding, pull-downs against Afadin-CC were performed using biotinylated PTPRK ICD, D1 and D2 domains. The PTPRK-D2 domain alone was sufficient to bind Afadin-CC to similar levels observed for the full ICD (Figure 2C). No interaction of Afadin-CC with the PTPRK-D1 domain was observed, highlighting that the D2 domain is necessary and sufficient for the interaction with this part of Afadin. To confirm this, ITC experiments were performed monitoring the binding of the Afadin-CC to the PTPRK-D2 domain. These data produced very similar results to those for the full ICD, specifically the Afadin-CC forms an equimolar complex with the PTPRK-D2 domain with a binding affinity in the low micromolar range (K_D = 6.3 ± 0.3 µM, Figure 2D and E). These data suggest there is only a very modest (twofold) reduction in binding affinity for the D2 domain when compared with the ICD, consistent with the D2 domain being the primary Afadin binding surface.

Having mapped the regions of Afadin and PTPRK required for binding, we sought to understand the structural basis of the interaction, but extensive efforts to crystallize the Afadin-CC:PTPRK-D2 complex were unsuccessful. Therefore, we exploited recent advances in deep learning neural network techniques by using AlphaFold2 (AF2)-Multimer to model the complex based on the truncation mapping data. Initial structural predictions were performed using the PTPRK-D2 domain (aa. 1154–1446) with Afadin-CC (aa. 1393–1455) as inputs. The AF2 model of PTPRK-D2 in complex with the full Afadin-CC fragment used for biochemical characterization highlighted that many residues at both the N- and C-termini of the Afadin-CC may not be contributing to the putative interaction interface (Figure 3—figure supplement 1). Therefore, further mapping experiments to refine the binding region of the Afadin-CC were conducted. Both N- and C-terminal truncations of GST-Afadin-CC were expressed, purified, and tested in pull-downs for their ability to bind PTPRK (Figure 3A). For N-terminal truncations of Afadin-CC, residues 1393–1400, which are predicted to be unstructured and outside the core helix (Figure 3—figure supplement 1), were replaced with an equivalent glycine/serine linker sequence to avoid the possibility of steric hindrance with the N-terminal GST-tag during mapping experiments (Figure 3A). In agreement with the initial AF2 model, regions at both the N- and C-termini of Afadin-CC were dispensable for the interaction, while disruption of the core sequence encompassing the charged residues (aa. 1408–1448) resulted in a lack of detectable binding (Figure 3A). Interestingly, this core charged region of Afadin is almost completely conserved across vertebrates and diverges in Drosophila melanogaster, which lacks a full-length PTPRK ortholog (Figure 3—figure supplement 2; Hatzihristidis et al., 2015).

Figure 3 with 3 supplements see all

Download asset Open asset

Figure 3—source data 1

Uncropped, unedited gels for Figure 3A.

https://cdn.elifesciences.org/articles/79855/elife-79855-fig3-data1-v2.zip

Download elife-79855-fig3-data1-v2.zip

Figure 3—source data 2

Uncropped, unedited gel images for Figure 3D.

https://cdn.elifesciences.org/articles/79855/elife-79855-fig3-data2-v2.zip

Download elife-79855-fig3-data2-v2.zip

Based on these additional mapping data, a new AF2 model was generated using the PTPRK-D2 domain with Afadin-CC minimal region (aa. 1408–1448) as inputs. Of the five models calculated, the top four had the Afadin-CC placed in the same position on the surface of the PTPRK-D2 domain (Figure 3—figure supplement 3), strongly suggesting that the CC domain adopts a linear alpha-helical structure that stretches across one face of the D2 domain (Figure 3B). The top AF2 model (ipTM + pTM score = 0.58) predicts the overall fold of each domain with high confidence and their relative orientations with moderate to high confidence (Figure 3C). To test the validity of this model and further explore the residues involved in Afadin binding to PTPRK, a series of single and double mutations were introduced into this region of the Afadin-CC domain. Mutations of Afadin used in these pull-down assays included residues predicted to be at the interaction interface as well as residues away from the interface (Figure 3D and E). Double alanine mutation at Afadin residues W1418-Y1419 or R1432-K1433, which contribute large sidechains to the interaction interface, abolishes binding to PTPRK, while E1429-R1430 to alanine reduces but does not abolish binding. Importantly, additional mutations to residues distal from the interface (R1410, Q1443, T1446) do not interfere with binding to PTPRK. These mutational data strongly support the validity of the predicted structure for how Afadin binds to the PTPRK D2 domain.

Our previous work demonstrated that both PTPRK and PTPRU bind Afadin while PTPRM cannot (Fearnley et al., 2019; Hay et al., 2020). Therefore, residues on the D2 domain that are critical for Afadin binding should be conserved in PTPRK and PTPRU but divergent in PTPRM. Based on a multiple-sequence alignment of PTPRK, PTPRU, and PTPRM (Figure 4—figure supplement 1), the positions of unique PTPRM residues were mapped onto the AF2 model of PTPRK-D2 (Figure 4A, the predicted position of the D1 domain is illustrated for clarity and orientation). Interestingly, a cluster of four residues mapped to a region of the PTPRK-D2 surface that is predicted to be the binding site for Afadin-CC (Figure 4A) and are located adjacent to several Afadin-CC residues that we have shown to be necessary for complex formation (Figure 3D and E). This binding site is located on the opposite face relative to the cysteine of the would-be ‘active’ site in the PTPRK-D2 domain (Figure 4A). Although these residues of PTPRK (Y1270, G1273, E1332, and L1335) are not all highly charged, within the context of the PTPRK model they form an acidic pocket into which the Afadin-CC is predicted to bind (Figure 4B and C). To investigate potential molecular differences at this site, a PTPRM-D2 (aa. 1160–1452) AF2 model was generated to allow comparison to PTPRK-D2. Strikingly, in a model of the PTPRM-D2 the altered residues in this region result in a different surface topology and charge, having lost the acidic pocket observed for PTPRK-D2 (Figure 4B and C). These changes in surface properties may therefore account for the differential binding to Afadin.

Figure 4 with 4 supplements see all

Download asset Open asset

Figure 4—source data 1

Uncropped, unedited gel for Figure 4D.

https://cdn.elifesciences.org/articles/79855/elife-79855-fig4-data1-v2.zip

Download elife-79855-fig4-data1-v2.zip

Figure 4—source data 2

Uncropped, unedited gel for Figure 4E.

https://cdn.elifesciences.org/articles/79855/elife-79855-fig4-data2-v2.zip

Download elife-79855-fig4-data2-v2.zip

Figure 4—source data 3

Uncropped, unedited blot for Figure 4G.

https://cdn.elifesciences.org/articles/79855/elife-79855-fig4-data3-v2.zip

Download elife-79855-fig4-data3-v2.zip

Figure 4—source data 4

Uncropped, unedited gel images for Figure 4H.

https://cdn.elifesciences.org/articles/79855/elife-79855-fig4-data4-v2.zip

Download elife-79855-fig4-data4-v2.zip

To further test the validity of the predicted PTPRK-D2:Afadin-CC complex, and to better understand the binding specificity, residues in the PTPRK-D2 acidic pocket were mutated to their PTPRM equivalents. Specifically, PTPRK-G1273 was mutated to histidine (G1273H) and L1335 to arginine (L1335R). These changes to PTPRM-equivalent residues are predicted to cause steric hindrance (G1273H) or introduce a basic residue into the acidic pocket (L1335R). Each of these single-point mutations were sufficient to abolish the interaction of the PTPRK-D2 with the GST-Afadin-CC (Figure 4D). Furthermore, the introduction of both mutations (double mutant [DM]) into the PTPRK-D2 or full PTPRK-ICD prevents binding to Afadin-CC (Figure 4E). To quantify the loss in affinity due to these point mutations, ITC binding experiments were repeated using the PTPRK-D2-DM with Afadin-CC and no binding was detected, confirming loss of the interaction (Figure 4F). To further confirm that these mutations could interfere with Afadin binding in the context of the full-length protein, wheat germ lysate pull-downs were repeated with WT and DM PTPRK constructs. While PTPRK-ICD and PTPRK-D2 bind Afadin, introduction of these two mutations into PTPRK reduces Afadin binding to background levels (Figure 4G). Two additional mutations were introduced into the PTPRK-D2 domain to test the predicted complex model. The first, F1225A, alters the PTPRK residue that is predicted to interact with the W1418-Y1419 residues in Afadin that were essential for binding (Figure 3D, Figure 4—figure supplement 2A). This single-point mutation (F1225A) in PTPRK-D2 vastly reduces binding to the Afadin-CC domain (Figure 4H). This residue is conserved across the R2B family so although it contributes to the binding affinity and validates the complex model, it does not contribute to specificity. Another mutation (labeled the ‘M-loop’) was introduced at the edge of the interface, changing residues CEE in PTPRK (aa. 1372–1374) to the PTPRM-equivalent YNG (aa. 1378–1380) (Figure 4—figure supplement 2B). This mutation retains binding, suggesting that this loop does not contribute to specificity of Afadin binding to PTPRK versus PTPRM (Figure 4H). The cysteine residue of the M-loop is distal to Afadin in the predicted PTPRK complex and the EE to NG substitution would not introduce significant steric hinderance or charge repulsion, consistent with retained binding. For all mutant forms of the PTPRK-D2 domain, correct folding was confirmed by monitoring their thermal stability and comparing melting temperatures (T_m) to wildtype PTPRK-D2 (Figure 4—figure supplement 3). Wildtype and mutant PTPRK-D2 domains possessed T_m values between 39 and 41°C, supporting that any loss of binding was not driven by misfolding. These data strongly validate the predicted model for the PTPRK-D2:Afadin-CC complex and identify two key residues within PTPRK D2 that are crucial for binding to full-length Afadin and, based on their lack of conservation in PTPRM, are likely to be important for determining binding specificity.

To determine whether the PTPRK-D2:Afadin-CC interaction is necessary for PTPRK-mediated dephosphorylation of Afadin, dephosphorylation assays were performed using both WT and DM PTPRK-ICDs. Since we and others have observed allosteric modulation of RPTP D1 domains by D2 domains (Fearnley et al., 2019; Wen et al., 2020), we tested whether introducing D2 domain mutations affects PTPRK-ICD phosphatase activity. The WT and DM PTPRK-ICDs were compared in activity assays using the generic phosphatase substrate 4-nitrophenyl phosphate (pNPP), which confirmed that both proteins have identical activity profiles (Figure 5A). Dephosphorylation assays were then performed using PTPRK-ICD WT and DM proteins incubated with quenched, pervanadate-treated lysates. Afadin was dephosphorylated at Y1230 by WT PTPRK-ICD, but not PTPRK-ICD DM (Figure 5B, Figure 5—figure supplement 1). Comparatively, another PTPRK substrate, p120-Catenin, which has been shown previously to be dephosphorylated efficiently by the PTPRK-D1 domain alone, was dephosphorylated by both proteins – independent of D2 domain mutation status. Results using our Afadin-pY1230 antibody were also confirmed by immunoprecipitation (IP) of pTyr proteins from WT and DM dephosphorylation assays (Figure 5C). Afadin was depleted in pTyr-IPs from lysates treated with WT PTPRK-ICD, showing dephosphorylation had occurred, while the levels of phosphorylated Afadin remained unchanged in PTPRK-ICD DM-treated samples (Figure 5C). Again, both WT and DM PTPRK were equivalently able to dephosphorylate p120-Catenin. These data show that D2 domain recruitment of Afadin is critical for efficient dephosphorylation by the D1 domain, and that PTPRK uses distinct surfaces to recruit different substrates for dephosphorylation.

Figure 5 with 1 supplement see all

Download asset Open asset

Figure 5—source data 1

Uncropped, unedited blots for Figure 5B.

https://cdn.elifesciences.org/articles/79855/elife-79855-fig5-data1-v2.zip

Download elife-79855-fig5-data1-v2.zip

Figure 5—source data 2

Uncropped, unedited blots for Figure 5C.

https://cdn.elifesciences.org/articles/79855/elife-79855-fig5-data2-v2.zip

Download elife-79855-fig5-data2-v2.zip

Amino acid (aa) numbering used throughout is based on the following sequences: PTPRK, UniProt ID: Q15262-3; PTPRM, UniProt ID: P28827-1; Afadin, UniProt ID: P55196-4. All point mutations were generated by site-directed mutagenesis using polymerase chain reaction with Phusion Hot Start II DNA polymerase (Thermo Fisher Scientific). The pET-15b.His.TEV.AviTag expression vectors used for the generation of biotinylated PTPRK-ICD (aa. 864–1439), D1 (aa. 865–1157), and D2 (aa. 1154–1446) domains were generated in a previous study (Fearnley et al., 2019). PTPRK-ICD and D2 G1273/L1335R single and double mutants (DM) were generated in these His.TEV.AviTag vectors. For ITC experiments, PTPRK-ICD, D2, and D2-DM were subcloned into a pET-15b expression vector in frame with an N-terminal His6-tag only. For wheat germ cell-free expression, the full-length human Afadin cDNA and truncations (outlined in Figure 1C) were subcloned into the pF3A WG (BYDV) vector (Promega) with a modified multiple cloning site (SpeI-AvrII-EcoRI-NheI-SacI-KpnI downstream of an N-terminal Myc-epitope; pF3A-WG-Myc). For bacterial expression, Afadin C-terminal fragments (outlined in Figure 1D) were subcloned into the pGEX-6P-1 vector with an in-frame N-terminal GST tag followed by the human rhinovirus 3C protease recognition sequence. For transient overexpression experiments, full-length Afadin was subcloned into pmScarlet-C1 (gift from Z. Kadlecova) and mutated to generate Y1226F and Y1230F single and double (YFYF) mutants.

All antibodies, except Afadin-pY1230, were used at a 1:1000 dilution in TBS-T (20 mM Tris–HCl, pH 7.4, 137 mM NaCl, 0.2% [v/v] Tween-20) with 3% (w/v) bovine serum albumin (BSA) unless otherwise noted. Rabbit anti-Myc, rabbit anti-Afadin, mouse anti-His-tag, rabbit anti-pTyr, rabbit anti-phospho-p120-Catenin (Y228 and Y904), rabbit anti-paxillin, and rabbit anti-phospho-paxillin (Y118) primary antibodies were purchased from Cell Signaling Technologies. Mouse anti-p120-Catenin and mouse anti-Afadin primary antibodies were purchased from BD Transduction Laboratories. Mouse anti-α-tubulin primary antibody was purchased from Sigma-Aldrich. Mouse anti-RFP antibody was purchased from Thermo Fisher Scientific. Anti-mouse, anti-goat, and anti-rabbit HRP-conjugated secondary antibodies (1:5000 in TBS-T) were purchased from Jackson ImmunoResearch. The rabbit anti-PTPRK antibody was generated previously (Fearnley et al., 2019).

The Afadin pTyr-1230 antibody was generated by MRCPPU Reagents and Services, University of Dundee (https://mrcppureagents.dundee.ac.uk/) using a phosphorylated Afadin-Y1230 peptide (CTYTRE[pY]FTFPA) as the antigen. Briefly, the N-terminal Cys residue was used to conjugate the peptide to both keyhole limpet hemocyanin (KLH) and BSA. Sheep were immunized with the antigen followed by four further injections 28 days apart, with bleeds performed 7 days after each injection. Antibodies were affinity purified from serum using the Afadin-Y1230 phospho-peptide. For Western blotting, the antibody was used at 1 μg/ml in TBS-T with 3% milk, after pre-incubation with 10 μg/ml non-phosphorylated peptide (TYTREYFTFPA) for 30 min at room temperature (RT) or 4°C.

Expression plasmids were transformed into competent E. coli BL21 (DE3) Rosetta cells and cultured in 2× TY medium supplemented with 50 µg/ml ampicillin and 34 µg/ml chloramphenicol at 37°C and 220 RPM to an optical density at 600 nm of 0.6. Cultures were then equilibrated to 20°C and protein expression induced with addition of 1 mM isopropyl-thio-β-d-galactopyranoside for 18 hr. For in vivo biotinylation of Avi-tagged proteins, cultures were supplemented with 200 µM D-biotin (Sigma-Aldrich) at the point of induction. Bacterial cultures were then harvested by centrifugation (4000 × g, 20 min) and cell pellets stored at –20°C. Bacterial cell pellets were thawed and resuspended in ice-cold lysis buffer (50 mM Tris–HCl, pH 7.4, 500 mM NaCl, 0.5 mM TCEP, 5% [v/v] glycerol, 1× EDTA-free protease inhibitors) prior to lysis by high-pressure disruption at 25 kpsi (Constant Systems Ltd). Cell lysates were then clarified by centrifugation (40,000 × g, 4°C, 30 min).

For GST-Afadin constructs, clarified cell lysates were incubated with washed glutathione sepharose 4B beads (Cytiva) for 1 hr at 4°C. Beads were packed into a gravity column and washed with 20 ml of GST wash buffer (50 mM Tris–HCl, pH 7.4, 500 mM NaCl, 0.5 mM TCEP) and protein eluted in 5 × 1 ml fractions of GST wash buffer supplemented with 50 mM reduced glutathione. Eluted proteins were further purified by SEC or dialysis. SEC was performed using a HiLoad Superdex 200 pg 16/600 column equilibrated in 50 mM Tris–HCl, 150 mM NaCl, 5% [v/v] glycerol, 5 mM DTT. Dialysis was performed using a Slide-A-Lyzer 20K MWCO cassette (Thermo Fisher Scientific) in SEC buffer (50 mM Tris–HCl pH 7.4, 150 mM NaCl, 5% [v/v] glycerol, 5 mM DTT) for 2 hr at RT, followed by overnight at 4°C, and another 2 hr with fresh SEC buffer to remove GSH. The purified protein was recovered from the cassettes and immediately used for streptavidin pull-downs.

For His6-tagged PTP constructs, clarified cell lysates were incubated with washed Ni-NTA agarose beads (QIAGEN) for 1 hr at 4°C. Beads were packed into a gravity column and washed with 10 ml of Ni-NTA wash buffer (50 mM Tris–HCl, pH 7.4, 500 mM NaCl, 5% [v/v] glycerol, 0.5 mM TCEP) with 10 mM imidazole, followed by 10 ml of Ni-NTA wash buffer with 20 mM imidazole. Protein was eluted in 6 × 0.5 ml fractions of elution buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5% [v/v] glycerol, 0.5 mM TCEP, 200 mM imidazole). For His6-tag-only constructs used in ITC experiments, proteins were further purified by anion-exchange chromatography (AEX). For PTPRK-D2, eluted Ni-NTA fractions were pooled and then diluted to low-salt/imidazole concentration by 1:10 dilution with dilution buffer (50 mM Tris–HCl, pH 7.4, 5% [v/v] glycerol, 5 mM DTT). The diluted sample was applied to a MonoQ 5/50 GL AEX column (Cytiva), equilibrated in low-salt buffer (50 mM Tris–HCl, pH 7.4, 10 mM NaCl, 5% [v/v] glycerol, 5 mM DTT) and bound protein eluted using a linear gradient with high-salt buffer (50 mM Tris–HCl, pH 7.4, 1 M NaCl, 5% [v/v] glycerol, 5 mM DTT). For PTPRK-ICD, AEX purification was performed as above, but with all buffers adjusted to pH 7.0.

For biotinylated Avi-tag constructs for use in pull-down assays, eluted Ni-NTA fractions were purified by SEC on a HiLoad Superdex 75 (D1 and D2 domains) or 200 (ICDs) pg 16/600 column equilibrated in 50 mM Tris, pH 7.4, 150 mM NaCl, 5% (v/v) glycerol, 5 mM DTT. Routinely, purity of all constructs was assessed by SDS-PAGE followed by staining with Coomassie.

To assess biotinylation, 10 µg of in vivo biotinylated protein was solubilized in an appropriate volume of 5× SDS loading buffer and heated to 95°C for 5 min. Samples were allowed to cool to RT prior to addition of a threefold molar excess of streptavidin (IBA Lifesciences). Samples were incubated for 10 min at RT, then immediately analyzed by SDS-PAGE followed by Coomassie staining. Approximate levels of biotinylation were calculated by 2D densitometry of protein band depletion upon addition of streptavidin using Fiji (Schindelin et al., 2012).

Recombinant PTP domains were made up to 500 µl in assay buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5% [v/v] glycerol, 5 mM DTT) at 0.6 µM. PTP domains and 20 mM pNPP substrate (in assay buffer) were equilibrated to 30°C for 15 min in an orbital shaking heat block at 500 RPM. To initiate reactions, 500 µl of 20 mM pNPP substrate was added to PTP containing tubes (0.3 µM PTP and 10 mM pNPP final concentrations). Reactions were carried out for 30 min at 30°C in an orbital shaking heat block at 500 RPM. At each time point (0, 2, 5, 10, 20, and 30 min), 100 µl of the total reaction was transferred to a 96-well microplate well containing 50 µl 0.58 M NaOH, terminating the reaction. After the final time point, absorbance of each sample was measured at 405 nm in a Spectramax M5 plate reader (Molecular Devices). Product formation was calculated by interpolation of absorbance values using a 4-nitrophenol standard curve of known concentration.

Wheat germ reactions were performed using the TnT SP6 High-Yield Wheat Germ Protein Expression System (Promega). On ice, 3 µg of the relevant WG expression plasmid in a total volume of 20 µl ddH₂O was combined with 30 µl of freshly thawed wheat germ mastermix. Reactions were carried out for 2 hr at 25°C in a thermocycler.

For wheat germ pull-down assays, 0.5 nmol recombinant in vivo biotinylated PTP domain (36 or 18 µg for ICD and D1/D2 domains, respectively) was bound to 125 µl streptavidin-coated magnetic beads in a total volume of 1 ml of purification buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5% [v/v] glycerol) for 1 hr at 4°C with rotation. Beads were collected on a magnetic stand and washed twice with purification buffer, followed by two washes in pull-down buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 10% [v/v] glycerol, 1% [v/v] Triton X-100). Wheat germ reactions were mixed thoroughly and diluted to a total volume of 200 µl in pull-down buffer. Pull-downs were carried out with 100 µl diluted wheat germ reaction combined with protein-conjugated streptavidin beads in a total volume of 200 µl (approximately 2.5 µM final bait concentration) pull-down buffer in 96-well microplate wells, for 1 hr at RT on a high-speed orbital shaking platform, 700 rpm. Magnetic beads were collected on a 96-well magnetic stand and supernatants removed. To wash, beads were resuspended in 200 µl of pull-down buffer containing 250 mM NaCl followed by shaking for 1 min at 700 rpm. This step was repeated for a total of four washes. After the final wash, samples were transferred to 1.5 ml tubes on a vertical magnetic stand. Beads were rinsed once with 1 ml TBS, then resuspended in 80 µl TBS with 20 µl 5× SDS sample buffer supplemented with 7 mM biotin. Pull-downs were eluted by incubation at 95°C for 10 min, beads separated on a magnetic stand and supernatants collected. Routinely, 50 µl of the final 100 µl sample were analyzed by SDS-PAGE followed by immunoblotting.

For PTPRK pull-downs from GST-fusion preys, recombinant in vivo biotinylated PTP domains were bound to streptavidin magnetic beads as described for wheat germ-based pull-downs (see above). Pull-downs were performed using 0.5 nmol PTPRK bait (1 µM) combined with 1 nmol (2 µM) Afadin GST-fusions in a total volume of 500 µl of pull-down buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 10% [v/v] glycerol, 1% [v/v] Triton X-100), incubated for 1 hr with end-over-end rotation at 4°C. Beads were collected on a magnetic stand and supernatants removed. Pull-downs were washed by thorough resuspension in 500 µl pull-down buffer for a total of four washes. Beads were rinsed once with 1 ml TBS, followed by resuspension in 80 µl TBS with 20 µl 5× SDS sample buffer supplemented with 7 mM biotin. Pull-downs were eluted by incubation at 95°C for 10 min, beads separated on a magnetic stand and supernatants collected. Routinely, 50 µl of the final 100 µl sample were analyzed and visualized by Coomassie staining.

For GST-Afadin pull-downs from PTPRK preys, 1 nmol of GST-Afadin-CC was bound to 30 µl glutathione sepharose beads for 1 hr at 4°C with rotation. Bead-bound GST-Afadin-CC was combined with 0.5 nmol PTPRK proteins in a total volume of 500 µl of pull-down buffer (2 µM GST-Afadin, 1 µM PTPRK final concentration) and pull-downs incubated for 1 hr at 4°C with end-over-end rotation. Beads were collected by brief centrifugation and pull-downs washed by thorough resuspension in 500 µl of pull-down buffer. After a total of four washes, pull-downs were eluted by boiling at 95°C in 1× SDS sample buffer in TBS.

All AF2 models were generated using default parameters and run via a locally installed version of AF2-Multimer (version 2.1.0; Evans et al., 2022; Jumper et al., 2021). All models and associated statistics have been deposited in the University of Cambridge Data Repository (https://doi.org/10.17863/CAM.82741). Graphical figures were rendered in PyMOL (Schrödinger LLC).

For SEC-MALS analysis, the GST-tag was removed from Afadin-CC by PreScission (3C) protease cleavage, using 30 µg GST-3C per 1 mg GST-Afadin-CC in 50 mM Tris, pH 7.4, 150 mM NaCl, 5% (v/v) glycerol, 5 mM DTT, 0.5 mM EDTA. Cleavage reactions were performed for 16 hr at 4°C, followed by removal of GST-3C and cleaved GST by reabsorption onto glutathione sepharose beads for 1 hr at 4°C. Cleavage reactions were then passed through a gravity column to separate sepharose beads from eluate containing the cleaved Afadin-CC. Afadin-CC was immediately loaded onto a HiLoad Superdex 75 pg 16/600 column (Cytiva) equilibrated in SEC buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5% [v/v] glycerol, 5 mM DTT) and peak fractions concentrated in 3K MWCO centrifugal filter unit (Merck Millipore) to 1 mg/ml (120 µM).

A 100 µl sample at 1 mg/ml was injected onto a Superdex 75 Increase 10/300 GL column (Cytiva) equilibrated in SEC buffer at a flow rate of 0.5 ml/min. Data for static light scattering and differential refractive index were measured in-line using DAWN 8+ and Optilab T-rEX detectors, respectively (both Wyatt Technology). The absolute molar masses of the elution peaks were calculated in ASTRA 6 (Wyatt Technology) using a protein dn/dc value of 0.185 ml/g.

For ITC experiments, the GST-tag was removed from Afadin-CC as described for SEC-MALS experiments (see above), with final SEC purification into ITC buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 0.5 mM TCEP). PTP domains were exchanged into ITC buffer by SEC using a HiLoad Superdex 200 pg 16/600 (PTPRK-ICD) or Superdex 75 Increase 10/300 (PTPRK-D2) column. ITC experiments were performed on an automated MicroCal PEAQ-ITC (Malvern Panalytical) at 25°C. Syringe titrant (Afadin-CC) at 260 µM was titrated into 30 µM cell titrate (PTPRK-ICD and D2) as 13 × 3 µl injections. ITC data were analyzed and fit to a one-site binding model using the MicroCal PEAQ-ITC analysis software (Malvern Panalytical).

DSF was performed using Sypro Orange dye (Invitrogen) as per manufacturer’s protocol in a Bio-Rad Real-Time PCR system. Reaction mixes consisting of 5 μg recombinant protein and 1× Sypro Orange dye were made up to a total volume of 20 μl in 1× PBS. Samples were then heated on a 1°C per 10 s gradient from 25 to 95°C and protein unfolding at each temperature monitored by measurement of fluorescence using the FRET channel. Fluorescent signal vs. temperature was fitted to a nonlinear Boltzmann-sigmoidal regression in GraphPad Prism, with the T_m calculated from the inflection point of the fitted curve.

MCF10A cells were obtained from the American Type Culture Collection (ATCC, CRL-10317). HEK293T cells were obtained from D Ron. Cells were maintained in a 37°C humidified 5% CO₂ ventilator and passaged every 2–4 days depending on the cell line. MCF10A cells were grown in MCF10A culture medium (1:1 DMEM:Ham’s F12, supplemented with 5% [v/v] horse serum, 20 ng/ml EGF, 0.5 mg/ml hydrocortisone, 100 ng/ml cholera toxin, and 10 mg/ml insulin) (Debnath et al., 2003). PTPRK-KO MCF10A cells and doxycycline-inducible stable cell lines were generated in a previous study (Fearnley et al., 2019) and cultured as for WT cells. HEK293T cells were cultured in DMEM containing 10% (v/v) FBS (Sigma-Aldrich), 2 mM l-glutamine (Sigma-Aldrich). For doxycycline induction of MCF10A cell lines, 2 × 10⁶ were seeded per 10 cm dish and cultured for 6 days, with media changes on days 2 and 4. Doxycycline (Sigma-Aldrich) was included on day 4, 48 hr prior to lysis. Cell lines that were not obtained from commercial sources were subjected to Mycoplasma testing using MycoAlert PLUS (Lonza) or MycoProbe (R&D Systems) Mycoplasma Detection Kits.

6 × 10⁵ HEK293T cells were reverse transfected with mScarlet or mScarlet-Afadin constructs using GeneJuice (Merck Millipore #70967-3). For each well of a 6-well plate, 6 µl GeneJuice transfection reagent was incubated with 200 µl serum/antibiotic-free OptiMEM (Thermo Fisher Scientific) for 5 min at RT, then combined with 2 µg of plasmid DNA for a further 15 min at RT prior to addition to cells in full growth media. After 4 hr, cells underwent a media change.

HEK293T cells were transiently transfected with indicated mScarlet constructs, pervanadate treated and lysed in RIPA buffer (50 mM Tris–HCl pH 7.5, 150 mM NaCl, 1% [v/v] NP-40, 0.5% [w/v] sodium deoxycholate, 1 mM EDTA, 0.2% [w/v] SDS), cOmplete Protease Inhibitor Cocktail (Roche) and phosphatase inhibitor phosSTOP (Roche). Lysates were cleared by centrifugation at 13,000 × g for 15 min at 4°C. Supernatant protein concentration was determined by BCA assay. 50 µg protein was resuspended with 50 U calf intestinal alkaline phosphatase (CIP) and 1× dephosphorylation buffer (Invivogen #18009-019). Samples were incubated for 40 min at 37°C with gentle agitation. The temperature was increased to 65°C for 10 min to inactivate CIP. 1× sample buffer was added to samples prior to SDS-PAGE and immunoblotting.

6 × 10⁵ HEK293T cells were reverse transfected with nontargeting or AFADIN siRNA duplexes (ON-TARGETplus SMARTpool, Dharmacon Horizon Discovery) using lipofectamine RNAiMAX as per the manufacturer’s instructions (Thermo Fisher Scientific # 13778030). Briefly, for each well of a 6-well plate, 6 µl RNAiMAX was used to transfect 10 nM siRNA duplexes. After 24 hr, media was replaced with complete growth medium and left to recover for another 24 hr prior to cell treatments and processing for analysis.

Pervanadate was generated based on Huyer et al., 1997. To avoid decomposition, it was prepared immediately prior to use. In a 1.5 ml tube, 100 µl of 100 mM sodium orthovanadate (Alfa Aesar) was combined with 103 µl 0.49 M H₂O₂ in 20 mM HEPES, pH 7.3, mixed by gentle inversion and incubated at RT for 5 min. Excess H₂O₂ was quenched by addition of 23 µl of 0.5 mg/ml catalase in 50 mM potassium phosphate, followed by mixing by gentle inversion. This yields a 44.2 mM pervanadate stock solution.

To generate pervanadate-treated lysates, MCF10A cells were seeded at 2 × 10⁶ in a 10 cm dish and cultured until confluent. Culture medium was then aspirated and cells treated with 8 ml of fresh MCF10A growth medium (see ‘Cells and cell culture’) supplemented with 100 µM pervanadate. Cells were treated for 30 min in a 37°C, 5% CO₂ incubator, then moved to ice and washed twice with ice-cold PBS. Each dish was lysed in 600 µl lysis buffer (50  mM Tris–HCl, pH 7.4, 150  mM NaCl, 10% [v/v] glycerol, 1% [v/v] Triton X-100, 1 mM EDTA, 5 mM iodoacetamide [IAA], 1 mM sodium orthovanadate, 10 mM NaF, 1× EDTA-free protease inhibitor) on ice, with periodic agitation, in the dark at 4°C for 30 min. Lysates were scraped, collected into 1.5 ml tubes, and treated with 10 mM DTT on ice for 15 min. Lysates were cleared by centrifugation (14,000 × g, 15 min, 4°C). Supernatants were then collected, snap frozen in liquid nitrogen, and stored at –80°C until use.

For Coomassie-stained gels, proteins were resolved on 4–12% Bis-Tris NuPAGE, and for Western blotting samples were run on 8 or 10% SDS-PAGE. Protein was transferred to 0.2 µm nitrocellulose membranes by wet transfer at 70 V for 2 hr at 4°C. Membranes were blocked in 5% (w/v) skimmed-milk in TBS-T (20 mM Tris–HCl, pH 7.4, 137 mM NaCl, 0.2% [v/v] Tween-20), prior to incubation with primary antibody overnight at 4°C followed by incubation with HRP-conjugated secondary antibody for 1 hr at RT. Blots were developed using EZ-ECL solution (Geneflow) and imaged using a Bio-Rad ChemiDoc MP imaging system.

Recombinant PTP domains (300 nM final concentration) were added to 200 µg of pervanadate-treated MCF10A lysate in a total volume of 400 µl wash buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 10% [v/v] glycerol, 1% [v/v] Triton X-100), supplemented with or without 5 mM DTT. Reactions were incubated with end-over-end rotation for indicated times (0.75–1.5 hr) at 4°C, then either directly analyzed by immunoblotting or subjected to pTyr immunoprecipitation.

For pTyr IP experiments, 400 µl dephosphorylation assays were made up to 0.4% (w/v) SDS, vortexed, and incubated on ice for 5 min to terminate reactions. Reactions were then made up to a total volume of 800 µl (0.2% SDS) in wash buffer and 6 µl of rabbit anti-pTyr antibody added to each sample. Samples were then incubated with end-over-end rotation at 4°C for 3 hr, followed by addition of 40 µl washed protein A agarose. IPs were carried out overnight (>16 hr) with end-over-end rotation at 4°C. Agarose beads were collected by centrifugation (15,000 × g, 30 s) and washed by resuspension in wash buffer. This wash step was repeated for a total of four washes. To elute IPs, beads were resuspended in 80 µl of 2.5× SDS loading buffer (diluted in wash buffer) and boiled for 10 min at 95°C. Supernatants were then collected and immediately analyzed by SDS-PAGE followed by immunoblot.

MSAs were generated using Clustal Omega (Sievers et al., 2011) and edited using Jalview (Waterhouse et al., 2009).

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. For the purpose of Open Access, the authors have applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission.

We thank Benjamin Butt for assistance with installing AlphaFold and Leah Wynn for assistance with Afadin mutagenesis. This study was supported by a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society awarded to HJS (109407/Z/15/Z), which also supports KEM. JED was supported by a Royal Society University Research Fellowship (UF100371) and a Wellcome Trust Senior Research Fellowship (219447/Z/19/Z). IMH was funded by a CIMR PhD studentship and Babraham Institutional bridging funds. TL was funded by a PhD jointly funded by Trinity College, Cambridge, and the BBSRC Cambridge SBS-DTP. HJS is an EMBO Young Investigator and Lister Institute Prize Fellow. A Titan V graphics card used for this research was donated to SCG by the NVIDIA Corporation. For the purpose of open access, the author has applied a Creative Commons Attribution (CC BY) license to any Author Accepted Manuscript version arising from this submission.

© 2022, Hay et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.