Two SKAP2 modular organizations differently recognize SRC kinases depending on their activation status and localization

doi:10.21203/rs.3.rs-1430558/v1

Download PDF

Research Article

Two SKAP2 modular organizations differently recognize SRC kinases depending on their activation status and localization

https://doi.org/10.21203/rs.3.rs-1430558/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: The activation of SRC kinases by trans-phosphorylation raises the question of their interplay. This work addresses this question by studying the interactions of SRC kinases with the SRC Kinase Adaptor Phosphoprotein 2 (SKAP2) by combining a luciferase complementation assay and extensive site-directed mutagenesis.

Results: SKAP2 interacts with SRC kinases through two different module organizations that have different affinities for the activated and non-activated kinase conformations. The DIM domain of SKAP2 is necessary and sufficient to bind to most activated SRC kinases. A second SKAP2 module organization for binding to SRC kinases at their steady state is composed of the DIM and SH3 domains as well as the linker located between the PH and the SH3 domains. Classical binding sites of the PH and SH3 domains and the tyrosines located inside linkers modulate these interactions. Analysis of HCK mutants supports this model and shows the role of two residues, Y390 and K7, on its degradation following activation.

Conclusion: In this article, we show that SKAP2 have a modular architecture composed of 2 domains and one interdomain to interact with SRC kinases, which binding capacity of each module depends on their localization and activation. This work opens new perspectives on how activation of SRC kinases occurs, which could have new and interesting therapeutic consequences.

SKAP2

SRC kinase family

Protein-protein interaction

Modular architecture

Luciferase Complementation Assay

Mutagenesis

The SRC kinase family is composed of the SrcA subfamily with FGR, FYN, SRC, and YES, the SrcB subfamily with BLK, HCK, LCK, and LYN, and finally FRK [1]. An SRC-related kinase, SRMS, also exists lacking a C-terminal tyrosine which phosphorylation status regulates kinase activity, and N-terminal myristylation sites. SRC kinases phosphorylate tyrosine residues, and exhibit a modular architecture composed of three domains, a SH3 domain, a SH2 domain, and a tyrosine kinase catalytic domain (Figure S1B). SRC kinases adopt either a closed, inactive conformation provided by intramolecular interactions between (i) the SH3 domain and a proline-rich motif located between the SH2 and kinase domains, (ii) the SH2 domain and a phosphorylated tyrosine motif in the C terminus of the protein or an open active conformation without these intra-molecular interactions (Figure 1A). SRC kinase activity is modulated by a fine allosteric control that modulates the passage from closed to open conformations [2-4]. The active conformation is locked by trans-phosphorylation of a tyrosine positioned in the active-loop of the catalytic site [5]. This phosphorylation also induces the appearance of a SH2 binding motif for Cbl ubiquitin ligases that controls SRC kinase degradation [6]. Interestingly, the kinase domain spontaneously adopts an active conformation in presence of its N-terminal region containing tryptophan 234, which induces rearrangement of two hydrophobic structures, the catalytic (CS) and the regulatory (RS) spines [7-10]. The localization of SRC kinases in different membranous subcellular compartments is finely regulated by N-terminal myristylation and palmitoylation with some differences between each SRC kinase [11, 12].

The SRC Kinase Adaptor Phosphoprotein 2 (SKAP2) is a broadly expressed assembling platform for multiple proteins such as SRC kinases with diverse subcellular localizations [13-15]. SKAP2 plays multiple functions, from control of actin-polymerization [16, 17] to podosome stabilization [18], during cellular activation by integrin clustering [19] or tyrosine kinase receptors [20]. Its role in cellular differentiation and cell migration explains that this protein is associated to multiple human diseases from metastatic progression [21] to sensibility to autoimmune diseases [22-24]. SKAP2^-/- mouse model [25] allows to study the functions of this protein in different cell types such as macrophages [18, 26, 27] and neutrophils [28-31]. SKAP2 is composed of three domains: an N-terminal dimerization (DIM) domain, a Pleckstrin homology (PH) domain, and a SRC homology 3 (SH3) domain (Figure S1A). Multiple sites of tyrosine phosphorylation are found across the entire protein from the PH domain probably modulating its capacity to bind to membranes, to linkers and their SRC homology 2 (SH2) binding motifs [32, 33]. SKAP2 interacts directly through its SH3 domain to proline-rich motifs found in many proteins such as FYB, PRAM1, PTK2B, and FAM102A [14, 34-36]. Crystallographic structure of the N-terminal region of the murine Skap2 protein shows that the DIM domain, interacts with the PH domain of the same molecule, blocking access to its PIP3 (phosphatidylinositol [3-5] P3) binding pocket [37]. DIM also dimerizes by forming a four-helix bundle, this dimerization was recently proposed to induce for most SRC kinases a fine-tuning between activation and an increase of their binding capacity to SKAP2 [34].

We previously analyzed how the dimerization of SKAP2 induces, for most SRC kinases, a fine-tuning between activation and an increase of their binding capacity to SKAP2 [34]. In this work, we wished to better qualify the underlying mechanism of this regulation. We show that the interaction of SKAP2 with SRC kinases depends on three modules with different properties associated with the activation and/or the localization of SRC kinases. Interestingly, this modular architecture of SKAP1, another member of the same family, is different. To study SRC kinase mutations, we choose to use HCK, which was the only human SRC kinase in which the complete 3D structure was determined [38, 39]. Our study suggests that the loss of signal for some HCK mutants with SKAP2 is due to a loss of protein stability, probably linked to proteasome-mediated degradation in which K7, play a major role. In conclusion, this work provided insight into the complex interaction between SKAP2 and SRC kinases. Such properties have been essential to understand the activation of other kinases with new and interesting therapeutic possibilities.

We finely analyzed the interactions between the SRC Kinase Adaptor Phosphoprotein 2 (SKAP2) and nine SRC kinases. The aim was (i) to define SRC binding sites on SKAP2, and (ii) to clarify how SKAP2/SRC interactions impacted on the SRC activation and sub-cellular localization of SRC kinases are involved their binding to SKAP2. For that purpose, we used a luciferase complementation assay where the C-terminal hemi-gaussia2 luciferase (Gluc2) is fused at the N-terminal of SKAP2 (N2-SKAP2) and the N-terminal hemi-gaussia1 luciferase (Gluc1) is fused either to the N-terminal (N1) or C-terminal (C1) of the SRC kinases [34].

Figure 1C shows the initial study of the interaction between N2-fused SKAP2 deletion mutants (Figure 1B) with N1 or C1-fused HCK SRC kinase. The signal is greater or equal for most C1-fused mutants than N1-fused ones but greatly varies depending on mutants. Due to these variations, both orientations of hemi-gaussia1 luciferase fused to SRC kinases were therefore studied in this paper, even if classical views of protein-protein interaction (PPI) using Luciferase Complementation Assay support to study first C1-fused SRC kinases.

To assess the effect of SKAP2 deletions or mutations, the Normalized Luciferase Ratio (NLR) [40] generated by their binding to SRC kinase has been divided by their binding to the full length, wild-type SKAP2. This generated a relative NLR enabling to standardize the interaction of mutated SKAP2 relative to the wild-type protein for the different SRC kinases. The interaction interface of SKAP2 with SRC kinases was studied with 21 SKAP2 mutants, eight of which are deletion mutants (Figure 1B and the others contain non-synonymous mutations (Figure S1A). Next, the role of sub-cellular localization and activation of SRC kinases was assessed by studying the interaction of wild-type SKAP2 with 24 HCK kinase mutants from 11 mutations (Figure S1B).

The DIM domain of SKAP2 is necessary and sufficient to bind C1-fused SRC kinases

A previous study supports that the DIMPH SKAP2 mutant (Figure 1B), which contains its two N-terminal domains, a dimerization domain (DIM) and a Pleckstrin Homology domain (PH), is able to bind most C1-fused SRC kinases but not N1-fused ones [34]. To confirm and understand the structural basis of this result, the eight SKAP2 deletion mutants (Figure 1B), were studied with the nine C1-fused SRC kinases (Figure 2). Only the two SKAP2 mutants lacking the DIM domain, Link2SH3 and SH3, did not bind to all the C1-fused SRC kinases supporting, a necessary role of this domain. Interestingly, the strength of the interaction for FRK but also YES is greater in three SKAP2 mutants, DomDIM, DIMPH and DSH3 than in the wild-type SKAP2 suggesting that the DIM domain is sufficient. However, both the PH and the SH3 domains as the interdomain Link2 modulate the PPI.

Furthermore, the study of SKAP1, a paralog of SKAP2, supports a more complex binding of SRC kinases (Figure 2). The high homology between SKAP1 and SKAP2 suggests some similarities for the binding of SRC kinases. In contrast to previous results with deletion SKAP2 mutants, the PPI signal of the N1-fused SRC kinases with SKAP1 was higher than that of the C1-fused SRC kinases. To reconcile these data, we took account that SRC kinases adopt different conformations affecting their kinase activity [4, 41] and that PPI signal between SKAP2 and SRC kinases was higher for activated than inactivated SRC kinases in our initial work. We now suppose that first, the orientation of the fused luciferase differently affects the conformations and activation of SRC kinases, and second, that SKAP2 has at least two binding sites for SRC kinases with different binding properties depending on their conformations, one of these sites was absent from SKAP1. This hypothesis has three consequences: First, the DIM domain of SKAP1 binds less efficiently C1-fused SRC kinases than that of SKAP2; Second, the DIM domain of SKAP2 binds more efficiently activated HCK mutants than the DIM domain of SKAP1, and the wild-type SKAP2, independently of the orientation for the hemi-luciferase; Third, different SKAP2 binding sites will be defined by studying N1-fused and C1-fused SRC kinases with SKAP2 deletion mutants.

Functional differences between the DIM domain of SKAP1 and SKAP2

A phylogeny analysis of each SKAP1 and SKAP2 protein allows to define the amino acid conservation in both structures using CONSURF website [42]. A stronger conservation of the DIM domain of SKAP2 than that of SKAP1 was detected (Figure S2). This loss of SKAP1 conservation during evolution suggests that the binding capacity to SRC kinases by its DIM domain has changed over time compared to the DIM domain of SKAP2. This gives a molecular support that the DIM domain of SKAP1 has a lower binding capacity for C1-fused SRC kinases than that of SKAP2. A differential interaction scatterplot of data from DomDIM1 SKAP1 mutant (y-axis) versus those from DomDIM2 SKAP2 mutant (x-axis) supports also the first consequence of the hypothesis (Figure 3). The data of N1-fused SRC kinases were aligned along the diagonal supporting that their capacity of binding is similar for both SKAP1 and SKAP2 proteins. In contrast, most data of C1-fused SRC kinases are in the lower right-hand quadrant in agreement with a higher binding capacity for DomDIM2 SKAP2 mutant than for DomDIM1 SKAP1 mutant.

Study of HCK mutants agrees with the second consequence (Figure 3). The PPI signal strength of the HCK mutants and C1-fused HCK was stronger for the DIM domain of SKAP2 than for either the DIM domain of SKAP1 or wild-type SKAP2 itself. These results support that the DIM domain of SKAP2 binds more efficiently activated HCK mutants and C1-fused HCK than the DIM domain of SKAP1 and that SKAP2 itself. These results agree with the two first consequences of our hypothesis. We cannot exclude that the difference of localizations of SRC kinases induced by the N1-fused or C1-fused hemi luciferase plays a role.

Two different modular organizations for the binding to SRC kinases exist on SKAP2

When the Gluc1 fragment is fused at the N terminus of the SRC kinases, only two SKAP2 mutants, deleted of either the PH domain alone or with the linker 1 (Link1) have similar signal strength for SRC kinases than for wild-type SKAP2 (Figure 4). There data are supporting an important role for the rest of the SKAP2 domains, i.e. the DIM and SH3 domains as well as the linker 2. The low signals of SKAP2 Link2SH3 and SH3 mutants for N1-fused SRC kinases support that the SH3 domain with the second linker are not sufficient for the binding. As previous results have shown that the DIM domain alone did not bind to N1-fused SRC kinases, this suggests that both the DIM and SH3 domains as well as the linker 2 are necessary. Furthermore, differential interaction scatterplots, comparing the mean of luminescence induced by the interaction of each SKAP2 mutant to that induced by the interaction of SKAP2 wild type, confirmed the major effect of the Gluc1 orientation on DomDIM2, DIMPH, and DSH3 SKAP2 mutants (Figure S3). These results show that two different modular organizations for the binding to SRC kinases exist on SKAP2: one comprising mainly the DIM domain, a second one that would be located at the interface between the DIM and SH3 domains and the linker between the PH and the SH3 domain. The module comprising mainly the domain DIM is sufficient to bind activated SRC kinases located at the membrane.

Modulation of the PPI between SKAP2 and SRC kinases

To exclude an effect of dimerization specifically on the binding of activated C1-fused SRC kinases to DIM domain, a SKAP2 DDIM mutant bearing the V24D, F27G, and V28E mutations (Figure S1A), which lost the capacity to dimerize was studied (Figure 5 and Figure S4). A similar decrease of the signal of both N1-fused and C1-fused SRC kinases to SKAP2 DDIM mutant was detected. This result strongly suggests that dimerization has an effect on the binding but different to that of the DomDIM2 SKAP2 mutant.

Classical views suggest that the binding of SRC kinases to SKAP2 mainly occurs by their SH2 domain and one of the phosphorylated tyrosines Y237 or Y261 of SKAP2 [13, 15, 20]. Our results of the SKAP2 deletion mutants do not support this hypothesis. To directly test it, a SKAP2 3YF mutant bearing three non-synonymous mutations, Y75F, Y237F, and Y261F, was studied (Figure S1A). Only a slight decrease of the signal for N1-fused SRC kinases was detected (Figures 5 and S4). This result confirms a limited role of these tyrosines.

The PH domain of SKAP2 is known to bind to the phosphatidyl inositol 3, 4, 5 tri-phosphate (PI[3,4,5]P₃). It also interacts with the DIM domain of the same molecule, which prevents the binding to PI[3,4,5]P₃[37]. Three mutations in the PH domain known to affect the interaction with the DIM domain, D129K [16, 17, 37], R140M [37], and K161E [43] and four mutation new ones located in the DIM domain E15A, R18A, or in the PH domain Y151F and Y197F (Figure S1A) were studied alone or in combination. The R140M and K161E mutations are known to affect the binding of PI[3,4,5]P₃ and D129K to prevent the intra-molecular interaction with DIM domain. Crystallographic data on the murine Skap2 support that E15A and R18A mutations located in the DIM domain prevent its interaction with the PH domain. Literature curation supports the view that Y151F and Y197F mutations affect the binding to PI[3,4,5]P₃ and possibly that to the DIM domain [37]. As shown in Figures 5 and S3, these mutations have a limited effect on the interaction with SRC kinases, except D129K, which surprisingly increased the signal mainly with C1-fused SRC kinases.

The SH3 domain of SKAP2 interacts with proline-rich helix of its partners in which its tryptophan 336 is required [14]. Interestingly, the study of W336K SKAP2 mutant (Figure S1A) shows an increased signal for both N1-fused and C1-fused SRC kinases (Figure 5).

In conclusion in agreement with a modular organization, the data presented in this paragraph show that these non-synonymous mutations alone or in combination have a lower effect on the binding of SRC kinases than the deletion of the corresponding domain or linker region.

Statistical validations

Previous hypotheses were first statistically tested using univariate analyses. For each mutant, three effects were tested on N1-fused SRC kinases, and on C1-fused ones comparing to theorical mean, and on the difference between the signal of the C1-fused SRC kinases and that of N1-fused ones. Student’s t tests were performed on the log-normal transformed mean luminescence ratio since data are linearized after log-log transformation in differential interaction scatterplots (see Figures S3, S4). Results were considered statistically significant only after Bonferroni’s correction (Table 1).

Comparison for the deletion mutants shows that (i) for DSH3, DIMPH, and DomDIM2, the signal was higher for C1-fused than N1-fused SRC kinases and decreases for C1-fused SRC kinases only for DomDIM2; (ii) for only one mutant, DLINK1PH SKAP2, the signal decreases specifically for C1-fused SRC kinases. For all other deletion mutants, the signal decreases for both N1-fused and C1-fused SRC kinases. These results strongly support two classes of sites for SRC kinases, the first one containing the DIM domain and suggest that the second one is composed of the DIM and the SH3 domains as wells as the linker 2 (Link2). The SH3 domain and the Link2 region altogether are not sufficient for the second site (see Link1SH3 and SH3 SKAP2 mutants). The fact that the signal of SKAP1 decreases specifically for C1-fused SRC kinases agrees with a different modular organization. Other statistical analyses for single point mutants show that for W336K and D129K SKAP2 mutants, the signal significantly increases while for DDIM SKAP2 mutant, it significantly decreases independently of the orientation of the fused Gluc1. For 3YF SKAP2 mutant, only the signal of N1-fused SRC kinases decreases. For D129K-3YF mutant, the signal significantly increases and for DDIM-D129K, significantly decreases only for C1-fused SRC kinases; For DDIM-E15AR18 SKAP2 mutant, the signal significantly decreases for both orientation of the fused luciferase. In conclusion, these results give a statistical support to our previous analyses.

Second, a statistical modeling was performed taking account both literature curation and our results. Nine variables were generated, describing function possibly affected by mutations and a status for each variable was assigned for each mutant. The two variables describing how SKAP2 protein interacts with either N1-fused or C1-fused SRC kinases were generated according to the following protocol: first which of the five domains or regions decreases significantly the signal; second what are the association between significant regions which decreases significantly the signal. The study of C1-fused SRC kinases data shows that the binding capacity of SKAP2 depends mainly on the presence of the DIM domain but is modulated by the presence of the SH3 domain and of the Link2 interdomain. The analysis of the N1-fused SRC kinases data shows that the second binding site of SKAP2 protein depends on the presence of three regions, DIM, Link2, and SH3 altogether. A two-way ANOVA analysis was carried out after lognormal transformation of the dependent variable on these nine variables with either N1-fused or C1-fused SRC kinase data. An explicative model was generated from of a full model by a step-by-step procedure that removed at each step the lowest non-significant variable until no non-significant variables were present (Table 2).

Only 3 variables have similar effect on both N1-fused and C1-fused SRC kinase, the gain of function associated to D129K and W336K mutations, which increase the signal of interaction, the DDIM mutation, which has an opposite effect. The 3YF mutation, Y75F, Y237F, Y261F, specifically decreases the signal of interaction with N1-fused SRC kinases. Mutations located in the PI[3,4,5]P_3-binding pocket of the PH domain and SKAP1 protein specifically decrease the signal of interaction with C1-fused SRC kinases, and the loss of association between the DIM and the PH domain, which has an opposite effect. In conclusion, this analysis strongly supports two modular organizations of SKAP2 for binding to SRC kinases, which differ by the binding capacity of each module depending on the activation and/or the localization of SRC kinases and are differently modulated. Also, note that SKAP1, another member of the same family, has lost the capacity of binding of its DIM domain alone with C1-fused SRC kinases.

Analysis of HCK mutants

Two non-exclusive hypotheses from literature curation might explain differences between N-terminal and C-terminal fused SRC kinases: a different sub-cellular localization and their level of activation. First, most SRC kinases such as HCK, are located to membranes by N-terminal myristylation and/or palmitoylation [11, 12]. This process should be blocked by the N-terminal fusion of a hemi-luciferase. Second, a major process of activation is the dephosphorylation of Tyr501, which suppresses its intramolecular binding to the SH2 domain [3, 4]. C-terminal fusion of a hemi-luciferase might decrease the phosphorylation of Tyr501 by CSK kinase by masking a C-terminal binding site. In our previous paper [34], we have shown that the binding capacity of SKAP2 for SRC kinases specifically increases with their activation. These two non-exclusive hypotheses were investigated using HCK mutants. Twenty-two HCK mutants generated from 11 mutations (Figure S1B) and fused with hemi-luciferase1 at each of the two orientations were tested for their binding to SKAP2 protein (Figure S5). HCK mutants bearing known mutations that activate kinase activity have a higher signal of interaction with SKAP2 than wild-type HCK independently of the Gluc1 orientation. N1-fused W93F HCK mutant is the only discordant result. Interestingly, the signal of C1-fused SRC kinases is significatively higher than that of N1-fused SRC kinases (Figure S6) in agreement with its higher activation, which was just shown to induce a greater strength for PPI signal.

Four mutations, Y30F, W93F, W234A, and RS, induced differential signals of HCK kinase depending on the orientation of the fused luciferase fragment. Also, the signal completely disappeared for C1-fused G2A HCK mutant. However, these effects are not mainly due to a difference of binding capacity but to a difference of protein steady state level (Figure 6). N2-SKAP2 and N1 or C1-HCK mutants and wild-type were detected by a polyclonal antibody against Gaussia princeps luciferase. The degradation of HCK is followed by the disappearance of the upper band (HCK) relative to the lower band (SKAP2) as well as by the appearance of small-sized degradation products. The HCK stability and the binding capacity were recovered by adding a second Y390F mutation to the four mutants of HCK C1-Y30F, N1-W93K, N1-W234A, C1-RS, and only protein stability was rescued for C1-G2A (Figures 6A, 6B, and S4). Interestingly, small-sized degradation products are detected with C1-fused wild type and mutant HCK but not with N1-fused wild type and mutant HCK suggesting an N-terminal degradation. To support a role of ubiquitination, we studied K7R mutation in which degradation possibly does not occur. For all the mutants tested, C1-fused Y30F HCK, N1-fused W93K HCK, N1-fused W234A HCK, a full binding capacity was recovered when K7R mutation was present as well as protein stability for two of them (Figures 6B and S7). In conclusion, our present work did not confirm a role of W93K HCK mutation on SKAP2 binding as suggested in our original work but suggested that different HCK mutations have an effect on protein stability. The following model supported by literature suggests that ubiquitination of Lysine 7 after the binding of Cbl ubiquitin ligase by its SH2 domain on phosphorylated tyrosine 390 activates proteasome degradation [6, 33]. Interestingly, the binding capacity to SKAP2 is poorly recovered for G2A HCK mutant suggesting that membranous localization of SRC kinases is necessary to bind SKAP2 at least for this HCK isoform.

In a previous paper [34], protein-protein interactions (PPIs) between SKAP2 and most of its partners were finely analyzed. However, for SRC kinases we were only able to town down the role of the SH2 domain of SRC kinases, at least for HCK and those of three SKAP2 phosphorylated tyrosines, Y75, Y237, Y26. Interestingly, DIMPH, a deletion mutant of SKAP2, has completely lost its capacity to bind N1-fused SRC kinases but has retained it for most C1-fused SRC kinases. The purpose of the present paper is to solve how SKAP2 and SRC kinases interact. One major difficulty to study SRC kinases comes from their activation, which seems to increase the binding to SKAP2 for the most part independently of the classical binding sites inside their SH2 and SH3 domains (Figure 6). In particular, the PPI signal of SKAP2 for wild-type HCK is lower than that for W93K R150M HCK double mutant, in which each classical binding site for both the SH3 and SH2 domain were inactivated. So, we decided to concentrate our efforts on SKAP2 binding sites. The interactions of 21 SKAP2 mutants, eight of which are deletion mutants with nine SRC kinases fused with hemi-gaussia luciferase at either N-terminal or C-terminal part were studied. Study of SKAP2 deletion mutants shows that the DIM domain is necessary and sufficient to bind most C1-fused SRC kinases. However, this result does not explain how SKAP1 protein, a paralog of SKAP2, binds SRC kinases. We were thus supposing first that the orientation of the fused luciferase differently affects the conformations and activation of SRC kinases, and second, that SKAP2 has different modular organizations for binding to SRC kinases depending on their conformations and localizations. Three different results support this hypothesis: First the strength of interaction of most SRC kinases for the DIM domain of SKAP2 is higher than for the DIM domain of SKAP1; Second the strength of interaction of activated HCK mutants is higher for the DIM domain of SKAP2 than for the DIM domain of SKAP1 and wild-type SKAP2; Third a second modular organization containing the SH3 domain, the second linker, and the DIM domain was defined by studying deletion SKAP2 mutants and N1-fused SRC kinases. These three independent confirmations make the risk of these protein-protein interactions being affected by the spatial position of the hemi-luciferase negligible.

Our present results did not support a role for the SH3 domain of HCK in initiating the interaction with SKAP2 protein. In our initial study [34], the stability of HCK mutants was not evaluated. The present data support that at least five mutations affect protein stability depending on the orientation of hemi gaussia1 fusion (Figure 6). Previous articles have reported that SRC kinase activation induces proteasome degradation after the binding of Cbl ubiquitin ligase by its SH2 domain on phosphorylated tyrosine 390 [6]. Our data agree with this model and suggest a degradation where ubiquitination of lysine7 plays a central role.

Separate analyses of interactions for SKAP2 deletion mutants with either N1-fused or C1-fused SRC kinases support two different SKAP2 modular organizations. Further analyses using HCK mutants also suggest that localization and activation of kinases differently affect their binding capacity for SKAP2. C1-fused SRC kinases are located to membranes and are probably more activated compared to N1-fused SRC kinases as suggested by the study of HCK mutants (Figures S5). C1-fused SRC kinases bind to the DIM SKAP2 domain, which is positively modulated by the presence of the SH3 domain and that of the second interdomain, Link2 (Figures 2 and S2, Table 1). N1-fused SRC kinases bind only to SKAP2 mutants containing both the SH3 domain with the second interdomain, Link2, and the DIM domain (Figures 5 and S2, Table 1). Interestingly, the DIM domain of the SKAP1 paralog, which is less conserved in the evolution than the DIM domain of SKAP2 protein, lost this binding capacity for C1-fused SRC kinases. However, in agreement with this hypothesis, the PPI signal of SKAP2 for C1-fused SRC kinases is higher than that for N1-fused ones (Figure S6).

Our results confirmed and extended the limited role of the three SKAP2 tyrosines, Y75, Y237, and Y261 (Figure 5 and S3). However, we are not able to conclude if this limited effect affects directly or indirectly the binding to SRC kinases. Interestingly inactivation of SH3 binding capacity by W336K mutation increases the signal of interaction with SRC kinases, independently of the orientation of hemi gaussia1 fusion (Table 1). It suggests a new hypothesis where SKAP2 would be maintained away from SRC kinases through interaction of its SH3 domain with other interactors. The study of mutations inside the PI[3,4,5]P₃-binding pocket of the PH domain strongly suggests that mutations of this site modulate SKAP2 interaction with SRC kinases by decreasing its localization close to SRC kinases (Table 2). Interestingly, D129K SKAP2 mutation possibly increases the strength of the interaction with SRC kinases by two different mechanisms: it makes the PI[3,4,5]P₃-binding pocket accessible to PI[3,4,5]P₃and/or induces another function, possibly a new site for positively charge lipids [37, 44, 45]. Our statistical modeling suggests that both mechanisms affect this binding (Table 2). Study of the DDIM SKAP2 mutant shows the positive role of dimerization in the PPI strength of both N1-fused and C1-fused SRC kinases even if the mechanism is less clear. One interesting hypothesis is a positive allosteric interaction between the two SKAP2 binding sites.

A similar structure to the SKAP2 dimer consisting in a four-helixes bundle, has been already reported to have a central role in kinase activation through allosteric control [46]. Future work will determine whether this is a new example of a protein-protein interaction that allosterically modulates kinase activation with interesting new therapeutic possibilities [47-49].

In this article, we show that the modular architecture of SKAP2 for binding to SRC kinase family involves its dimerization and SH3 domains as the second interdomain. The binding capacity of each module to SRC kinases differs depending on the localization and the activation of the kinases. Also, we propose that the stability of HCK SRC kinase family member may be affected by proteasome-mediated degradation. These data are necessary to understand how SKAP2, which is associated from autoimmune disorders to some cancer prognostics, affects SRC kinase functions and what is its relationship with its own functions. Modulating specifically this interaction might have interesting therapeutic consequences.

Plasmids, mutagenesis, BP and LR cloning, and luciferase complementation assay

Previously used plasmids and a more detailed methodology will be found elsewhere [34]. The ORF of the new SKAP2 deletion mutants flanked by two gateway sites were amplified by PCR, cloned into pCR-II TOPO vector using TOPO TA cloning kit (Thermo Fisher Scientific, Waltham) and transferred into pDONR207 vector. Mutagenesis of SKAP2, and HCK ORFs were performed using QuickChange Lightening Site-Directed Mutagenesis kit (Agilent Technologies, Les Ulis) according to the manufacturer’s protocol. Supplementary Table S1 shows the new primers used for all these purposes. After each mutagenesis and PCR amplification, the insert was completely sequenced at GATC services using LightRun Barcodes (Eurofins Genomics, Ebersberg, Germany). Sequence alignments were performed using DNA Strider [50]. The ORF of SKAP2 mutants were transferred into pSPICA-N2 vector and those of HCK mutant into pSPICA-N1 and pSPICA-C1 ones. The pSPICA vectors are mammalian expressing vectors designed for luciferase complementation assay. They expressed Gaussia princeps Luciferase fragment 1 (amino acid residues from 18 to 109) or 2 (amino acid residues from 110 to 185) at the N or C-terminal part of the fusion protein. The luciferase complementation assay was performed as described [40]. Transfections were performed at least in triplicate. Protein-protein interactions were monitored by measuring interaction-mediated normalized luminescence ratio (NLR) [40]. For each experiment, a threshold was defined as NLR mean + SEM of sample FAM102B.

Western blot

Pooled triplicate samples from luciferase complementation assay diluted in Laemmli x1 buffer were passed through Qiashredder column (Qiagen, Hiden, Germany) before eating at 95°C during 5 min. After 2h migration on a 4-12% NuPAGE Trisminigel (ThermoFisher Scientific, Waltham) at 120V at room temperature in x1 MES buffer, proteins were transferred to a PVDF membrane at 5°C. After saturating the membrane with 1x PBS, Tween 0.25% 5% milk during 30 min at room temperature, an overnight hybridization at 4°C with a primary antibody in 1x PBS, tween 0.25%, BSA 1% buffer was performed followed by 1h hybridization at room temperature with a secondary antibody in the same buffer. Revelation was performed on G:Box system (Syngene, Cambridge, UK) after incubation with Pierce ECL2 Western Blotting substrate (ThermoFisher Scientific, Waltham)). Table S2 shows the different antibodies used.

Graphic presentation and statistics

Approximation of the mean and the variance of a ratio were performed using Taylor expansions with a null covariance between the numerator and the denominator. Two-sample z test and two-way ANOVA after log-normal transformation were used to compare these ratios. Differential interaction scatterplots and PPI-mutation plots were performed as described in [34]. Three levels of significance were used for the P-value lower than 0.05 (*), 0.01 (**), and 0.001 (***). Bonferroni correction was used for multiple testing with an initial threshold p = 0.05. To compare different experiments, we used modified NLR for each SRC kinases as the ratio between NLR of the SKAP2 mutant and NLR of wild-type SKAP2 protein. Each mutant experiment was repeated at least three times. Statistical analyses were performed using Stata/IC version 10.1 (Stata Corporation, College Station, USA). Comparison of mean to theorical value (i.e. Mean ratio to 1 and its log-normal transformation to 0) and between two means was performed by using Student t test and unpaired Student t test, respectively. Comparison of mean to theorical value was also studied using binomial distribution after calculating the number of samples upper/lower to theoretical value. Comparison of two means was also performed by Kruskal-Wallis test. Parametric and non-parametric tests giving similar results, only results of the parametric are shown in Table1. Two-way ANOVA were performed after a log-normal transformation of the modified NLR and using analytic weights. For modeling the data, the full model contains 9 explanatory variables and N1-fused and C1-fused SRC kinase data were separately analyzed. For generating the variable explaining which domains are necessary to bind to either N1-fused SRC kinases or C1-fused ones, the following protocol was used. Two-way ANOVA was performed on deletion SKAP2 mutant data using five variables, each explaining the effect of a SKAP2 region. Significant variables and all their possible region combinations were used in a second two-way ANOVA. Significant variables of the second analysis were used to generate the definitive variable describing which regions are necessary for defining the binding site. An explicative model was generated from of this full model by a step-by-step procedure that removed at each step the lowest non-significant variable until no non-significant variables were present.

Conservation of the DIM domain of SKAP1 and SKAP2

The N-terminal fragment of human SKAP1 and SKAP2 protein structures were predicted from I-TASSER server [51] using two crystal structures of a N-terminal fragment of murine SKAP2 containing both the helical dimerization domain and the PH domain (ID PDB: 2OTX and 1U5E). Clustalw program [52] was used to align respectively 37 and 38 protein sequences of SKAP1 and SKAP2 from the Coelacanth to mammals. ConSurf server was used to visualize evolutionary conservation in these 2 macromolecules from their respective predict structure and aligned sequences using a maximum likelihood procedure. UCSF Chimera version 1.13.1 was used to visualize the conservation of these structures using a script from ConSurf server [42, 53].

DIM domain domain of dimerization

DIMPH dimerization and Pleckstrin homology domains

SKAP1 SRC Kinase Adaptor Phosphoprotein 1

SKAP2 SRC Kinase Adaptor Phosphoprotein 2

BLK B Lymphocyte Kinase

FGR Feline Gardner-Rasheed sarcoma viral homolog

FRK Fyn Related src family tyrosine Kinase

FYN FGR-YES1-related Novel protein

HCK Hematopoietic Cell Kinase

LCK Lymphocyte-Specific Kinase

LYN LCK/Yes-related Novel protein tyrosine kinase

PH domain Pleckstrin Homology domain

SRMS Src-Related kinase lacking C-terminal regulatory tyrosine and N-terminal Myristylation Sites

SH2 domain Src Homology 2 domain

SH3 domain Src Homology 3 domain

NLR Normalized Luciferase Ratio

Ethics approval and consent to participate

This research has received an OGM agreement from the Ministère de l’Education Nationale (Dossier DUO n°9216).

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Funding

Funding was provided by Institut Pasteur (Paris), Centre National de la Recherche Scientifique, Université Paris Diderot, Sorbonne Paris Cité.

Author contributions

JFB, and YJ designed research; JFB, LL, MD, MDD, LL, and KD performed research; PC, MD, and MDD contributed new reagents or analytical tools; JFB, LL, LL, and KD analyzed data; JFB, CD, AS, and YJ wrote the paper.

Acknowledgement

We thank Chen Kuang-Yu, Tim Krischuns and Guillaume Dugied for regularly providing HEK 293T cell culture, and the Dana-Farber Cancer Institute (Boston, MA, USA) for providing us with ORFs.

1. Ingley E: Src family kinases: regulation of their activities, levels and identification of new pathways. Biochim Biophys Acta 2008, 1784(1):56-65.

2. Bradshaw JM: The Src, Syk, and Tec family kinases: distinct types of molecular switches. Cell Signal 2010, 22(8):1175-1184.

3. Engen JR, Wales TE, Hochrein JM, Meyn MA, 3rd, Banu Ozkan S, Bahar I, Smithgall TE: Structure and dynamic regulation of Src-family kinases. Cell Mol Life Sci 2008, 65(19):3058-3073.

4. Roskoski R, Jr.: Src protein-tyrosine kinase structure, mechanism, and small molecule inhibitors. Pharmacological research 2015, 94:9-25.

5. Cooper JA, MacAuley A: Potential positive and negative autoregulation of p60c-src by intermolecular autophosphorylation. Proc Natl Acad Sci U S A 1988, 85(12):4232-4236.

6. Mohapatra B, Ahmad G, Nadeau S, Zutshi N, An W, Scheffe S, Dong L, Feng D, Goetz B, Arya P et al: Protein tyrosine kinase regulation by ubiquitination: critical roles of Cbl-family ubiquitin ligases. Biochim Biophys Acta 2013, 1833(1):122-139.

7. Kornev AP, Taylor SS: Defining the conserved internal architecture of a protein kinase. Biochim Biophys Acta 2010, 1804(3):440-444.

8. Kornev AP, Taylor SS: Dynamics-Driven Allostery in Protein Kinases. Trends Biochem Sci 2015, 40(11):628-647.

9. Meng Y, Roux B: Computational study of the W260A activating mutant of Src tyrosine kinase. Protein Sci 2016, 25(1):219-230.

10. Gonfloni S, Frischknecht F, Way M, Superti-Furga G: Leucine 255 of Src couples intramolecular interactions to inhibition of catalysis. Nat Struct Biol 1999, 6(8):760-764.

11. Patwardhan P, Resh MD: Myristoylation and membrane binding regulate c-Src stability and kinase activity. Mol Cell Biol 2010, 30(17):4094-4107.

12. Sato I, Obata Y, Kasahara K, Nakayama Y, Fukumoto Y, Yamasaki T, Yokoyama KK, Saito T, Yamaguchi N: Differential trafficking of Src, Lyn, Yes and Fyn is specified by the state of palmitoylation in the SH4 domain. J Cell Sci 2009, 122(Pt 7):965-975.

13. Kouroku Y, Soyama A, Fujita E, Urase K, Tsukahara T, Momoi T: RA70 is a src kinase-associated protein expressed ubiquitously. Biochem Biophys Res Commun 1998, 252(3):738-742.

14. Liu J, Kang H, Raab M, da Silva AJ, Kraeft SK, Rudd CE: FYB (FYN binding protein) serves as a binding partner for lymphoid protein and FYN kinase substrate SKAP55 and a SKAP55-related protein in T cells. Proc Natl Acad Sci U S A 1998, 95(15):8779-8784.

15. Marie-Cardine A, Verhagen AM, Eckerskorn C, Schraven B: SKAP-HOM, a novel adaptor protein homologous to the FYN-associated protein SKAP55. FEBS Lett 1998, 435(1):55-60.

16. Alenghat FJ, Baca QJ, Rubin NT, Pao LI, Matozaki T, Lowell CA, Golan DE, Neel BG, Swanson KD: Macrophages require Skap2 and Sirpalpha for integrin-stimulated cytoskeletal rearrangement. J Cell Sci 2012, 125(Pt 22):5535-5545.

17. Bourette RP, Therier J, Mouchiroud G: Macrophage colony-stimulating factor receptor induces tyrosine phosphorylation of SKAP55R adaptor and its association with actin. Cell Signal 2005, 17(8):941-949.

18. Tanaka M, Shimamura S, Kuriyama S, Maeda D, Goto A, Aiba N: SKAP2 Promotes Podosome Formation to Facilitate Tumor-Associated Macrophage Infiltration and Metastatic Progression. Cancer Res 2016, 76(2):358-369.

19. Wang H, Rudd CE: SKAP-55, SKAP-55-related and ADAP adaptors modulate integrin-mediated immune-cell adhesion. Trends Cell Biol 2008, 18(10):486-493.

20. Curtis DJ, Jane SM, Hilton DJ, Dougherty L, Bodine DM, Begley CG: Adaptor protein SKAP55R is associated with myeloid differentiation and growth arrest. Exp Hematol 2000, 28(11):1250-1259.

21. Kuranami S, Yokobori T, Mogi A, Altan B, Yajima T, Onozato R, Azuma Y, Iijima M, Kosaka T, Kuwano H: Src kinase-associated phosphoprotein2 expression is associated with poor prognosis in non-small cell lung cancer. Anticancer Res 2015, 35(4):2411-2415.

22. Liu JZ, van Sommeren S, Huang H, Ng SC, Alberts R, Takahashi A, Ripke S, Lee JC, Jostins L, Shah T et al: Association analyses identify 38 susceptibility loci for inflammatory bowel disease and highlight shared genetic risk across populations. Nat Genet 2015, 47(9):979-986.

23. Barrett JC, Clayton DG, Concannon P, Akolkar B, Cooper JD, Erlich HA, Julier C, Morahan G, Nerup J, Nierras C et al: Genome-wide association study and meta-analysis find that over 40 loci affect risk of type 1 diabetes. Nat Genet 2009, 41(6):703-707.

24. International Multiple Sclerosis Genetics C, Beecham AH, Patsopoulos NA, Xifara DK, Davis MF, Kemppinen A, Cotsapas C, Shah TS, Spencer C, Booth D et al: Analysis of immune-related loci identifies 48 new susceptibility variants for multiple sclerosis. Nat Genet 2013, 45(11):1353-1360.

25. Togni M, Swanson KD, Reimann S, Kliche S, Pearce AC, Simeoni L, Reinhold D, Wienands J, Neel BG, Schraven B et al: Regulation of in vitro and in vivo immune functions by the cytosolic adaptor protein SKAP-HOM. Mol Cell Biol 2005, 25(18):8052-8063.

26. Reinhold A, Reimann S, Reinhold D, Schraven B, Togni M: Expression of SKAP-HOM in DCs is required for an optimal immune response in vivo. J Leukoc Biol 2009, 86(1):61-71.

27. Takagane K, Umakoshi M, Itoh G, Kuriyama S, Goto A, Tanaka M: SKAP2 suppresses inflammation-mediated tumorigenesis by regulating SHP-1 and SHP-2. Oncogene 2022, 41(8):1087-1099.

28. Shaban L, Nguyen GT, Mecsas-Faxon BD, Swanson KD, Tan S, Mecsas J: Yersinia pseudotuberculosis YopH targets SKAP2-dependent and independent signaling pathways to block neutrophil antimicrobial mechanisms during infection. PLoS Pathog 2020, 16(5):e1008576.

29. Nguyen GT, Shaban L, Mack M, Swanson KD, Bunnell SC, Sykes DB, Mecsas J: SKAP2 is required for defense against K. pneumoniae infection and neutrophil respiratory burst. Elife 2020, 9.

30. Boras M, Volmering S, Bokemeyer A, Rossaint J, Block H, Bardel B, Van Marck V, Heitplatz B, Kliche S, Reinhold A et al: Skap2 is required for beta2 integrin-mediated neutrophil recruitment and functions. J Exp Med 2017, 214(3):851-874.

31. Nguyen GT, Xu S, Adams W, Leong JM, Bunnell SC, Mansour MK, Sykes DB, Mecsas J: Neutrophils require SKAP2 for reactive oxygen species production following C-type lectin and Candida stimulation. iScience 2021, 24(8):102871.

32. Dinkel H, Michael S, Weatheritt RJ, Davey NE, Van Roey K, Altenberg B, Toedt G, Uyar B, Seiler M, Budd A et al: ELM--the database of eukaryotic linear motifs. Nucleic Acids Res 2012, 40(Database issue):D242-251.

33. Hornbeck PV, Zhang B, Murray B, Kornhauser JM, Latham V, Skrzypek E: PhosphoSitePlus, 2014: mutations, PTMs and recalibrations. Nucleic Acids Res 2015, 43(Database issue):D512-520.

34. Bureau JF, Cassonnet P, Grange L, Dessapt J, Jones L, Demeret C, Sakuntabhai A, Jacob Y: The SRC-family tyrosine kinase HCK shapes the landscape of SKAP2 interactome. Oncotarget 2018, 9(17):13102-13115.

35. Takahashi T, Yamashita H, Nagano Y, Nakamura T, Ohmori H, Avraham H, Avraham S, Yasuda M, Matsumoto M: Identification and characterization of a novel Pyk2/related adhesion focal tyrosine kinase-associated protein that inhibits alpha-synuclein phosphorylation. J Biol Chem 2003, 278(43):42225-42233.

36. Moog-Lutz C, Peterson EJ, Lutz PG, Eliason S, Cave-Riant F, Singer A, Di Gioia Y, Dmowski S, Kamens J, Cayre YE et al: PRAM-1 is a novel adaptor protein regulated by retinoic acid (RA) and promyelocytic leukemia (PML)-RA receptor alpha in acute promyelocytic leukemia cells. J Biol Chem 2001, 276(25):22375-22381.

37. Swanson KD, Tang Y, Ceccarelli DF, Poy F, Sliwa JP, Neel BG, Eck MJ: The Skap-hom dimerization and PH domains comprise a 3'-phosphoinositide-gated molecular switch. Mol Cell 2008, 32(4):564-575.

38. Schindler T, Sicheri F, Pico A, Gazit A, Levitzki A, Kuriyan J: Crystal structure of Hck in complex with a Src family-selective tyrosine kinase inhibitor. Mol Cell 1999, 3(5):639-648.

39. Sicheri F, Moarefi I, Kuriyan J: Crystal structure of the Src family tyrosine kinase Hck. Nature 1997, 385(6617):602-609.

40. Cassonnet P, Rolloy C, Neveu G, Vidalain PO, Chantier T, Pellet J, Jones L, Muller M, Demeret C, Gaud G et al: Benchmarking a luciferase complementation assay for detecting protein complexes. Nat Methods 2011, 8(12):990-992.

41. Yang S, Blachowicz L, Makowski L, Roux B: Multidomain assembled states of Hck tyrosine kinase in solution. Proc Natl Acad Sci U S A 2010, 107(36):15757-15762.

42. Ashkenazy H, Abadi S, Martz E, Chay O, Mayrose I, Pupko T, Ben-Tal N: ConSurf 2016: an improved methodology to estimate and visualize evolutionary conservation in macromolecules. Nucleic Acids Res 2016, 44(W1):W344-350.

43. Witte A, Meineke B, Sticht J, Philipsen L, Kuropka B, Muller AJ, Freund C, Schraven B, Kliche S: D120 and K152 within the PH Domain of T Cell Adapter SKAP55 Regulate Plasma Membrane Targeting of SKAP55 and LFA-1 Affinity Modulation in Human T Lymphocytes. Mol Cell Biol 2017, 37(7).

44. Vonkova I, Saliba AE, Deghou S, Anand K, Ceschia S, Doerks T, Galih A, Kugler KG, Maeda K, Rybin V et al: Lipid Cooperativity as a General Membrane-Recruitment Principle for PH Domains. Cell Rep 2015, 12(9):1519-1530.

45. Yamamoto E, Domanski J, Naughton FB, Best RB, Kalli AC, Stansfeld PJ, Sansom MSP: Multiple lipid binding sites determine the affinity of PH domains for phosphoinositide-containing membranes. Sci Adv 2020, 6(8):eaay5736.

46. Lavoie H, Sahmi M, Maisonneuve P, Marullo SA, Thevakumaran N, Jin T, Kurinov I, Sicheri F, Therrien M: MEK drives BRAF activation through allosteric control of KSR proteins. Nature 2018, 554(7693):549-553.

47. Kung JE, Jura N: Structural Basis for the Non-catalytic Functions of Protein Kinases. Structure (London, England : 1993) 2016, 24(1):7-24.

48. Paul MD, Hristova K: The RTK Interactome: Overview and Perspective on RTK Heterointeractions. Chem Rev 2019, 119(9):5881-5921.

49. Wacker D, Stevens RC, Roth BL: How Ligands Illuminate GPCR Molecular Pharmacology. Cell 2017, 170(3):414-427.

50. Marck C: 'DNA Strider': a 'C' program for the fast analysis of DNA and protein sequences on the Apple Macintosh family of computers. Nucleic Acids Res 1988, 16(5):1829-1836.

51. Yang J, Yan R, Roy A, Xu D, Poisson J, Zhang Y: The I-TASSER Suite: protein structure and function prediction. Nat Methods 2015, 12(1):7-8.

52. Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R et al: Clustal W and Clustal X version 2.0. Bioinformatics 2007, 23(21):2947-2948.

53. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, Ferrin TE: UCSF Chimera--a visualization system for exploratory research and analysis. J Comput Chem 2004, 25(13):1605-1612.

Tables 1-2 are available in the Supplementary Files section.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Two SKAP2 modular organizations differently recognize SRC kinases depending on their activation status and localization

Status:

Version 1

Abstract

Figures

Background

Results

Discussion

Conclusions

Methods

Abbreviations

Declarations

References

Tables 1-2

Additional Declarations

Supplementary Files

Status:

Version 1