Tandem engagement of phosphotyrosines by the dual SH2 domains of p120RasGAP

SUMMARY

p120RasGAP is a multidomain GTPase-activating protein for Ras. The presence of two Src homology 2 domains in an SH2-SH3-SH2 module raises the possibility that p120RasGAP simultaneously binds dual phosphotyrosine residues in target proteins. One known binding partner with two proximal phosphotyrosines is p190RhoGAP, a GTPase-activating protein for Rho GTPases. Here, we present the crystal structure of the p120RasGAP SH2-SH3-SH2 module bound to a doubly tyrosine-phosphorylated p190RhoGAP peptide, revealing simultaneous phosphotyrosine recognition by the SH2 domains. The compact arrangement places the SH2 domains in close proximity resembling an SH2 domain tandem and exposed SH3 domain. Affinity measurements support synergistic binding, while solution scattering reveals that dual phosphotyrosine binding induces compaction of this region. Our studies reflect a binding mode that limits conformational flexibility within the SH2-SH3-SH2 cassette and relies on the spacing and sequence surrounding the two phosphotyrosines, potentially representing a selectivity mechanism for downstream signaling events.

eTOC Blurb:

In the current work, Stiegler et al examine the binding of the two SH2 domains in p120RasGAP to dual phosphotyrosine residues in p190RhoGAP. The co-crystal structure, solution scattering data, and affinity measurements together support synergistic phosphopeptide binding that promotes a compact arrangement of the domains.

Graphical Abstract

INTRODUCTION

p120RasGAP (RASA1, RasGAP, GTPase-activating protein 1) is a key regulatory protein of the Ras subfamily of small guanine-nucleotide binding GTPases. The small GTPase proteins act as binary molecular switches whose signaling state is determined by their GTP-bound (on) or GDP-bound (off) conformations. Ras GAP proteins such as p120RasGAP enhance the weak intrinsic catalytic activity of Ras GTPases by binding directly and contributing a catalytic arginine residue, the so-called ‘arginine finger’, to the active site to stabilize the reaction intermediate and thus turn off Ras signaling ^1,2. p120RasGAP has been shown to enhance the GTPase activity of Ras by two to five orders of magnitude ^3,4. In contrast to GAPs, which promote the off-state of GTPases, the Ras guanine nucleotide exchange factor (GEF) proteins including son of sevenless (Sos), RasGRF1 and RasGRP, stimulate the Ras on-state by enhancing the dissociation of GDP and re-association of GTP, thus completing the GTPase cycle ⁵.

p120RasGAP was the first GAP identified for the Ras GTPases ^3,6. Its absence in mice is embryonically lethal due to defective cardiovascular development ⁷, and the human RASA1 gene has been linked with both cancer ⁸ and vascular diseases including capillary malformation-arteriovenous malformation syndromes and vein of Galen malformations ^9–11. p120RasGAP is a multidomain protein (Figure 1A) with a C-terminal Ras-binding GAP domain ¹² preceded by plekstrin homology (PH) domain and a putative calcium-dependent phospholipid-binding (CaLB or C2) domain ^13,14. At the N-terminus, a roughly 170 residue predicted disordered region is followed by a compact cassette containing two Src homology 2 domains (N-SH2 and C-SH2) flanking a Src homology 3 (SH3) domain ^3,6,15.

Figure 1. Solution studies of the complex between p120RasGAP and bis-phosphorylated p190RhoGAP peptide.

A) Domain architecture of p120RasGAP (Uniprot ID: P20936. Domain IDs: SH2: Src Homology 2, SH3: Src Homology 3, PH: Pleckstrin Homology, C2: Protein Kinase C Conserved domain 2; RasGAP: Ras GTPase Activating Protein) and p190RhoGAP (Uniprot ID: Q9NRY4. Domain IDs: N-pseudoGTPase, FF, pseudoGTPase, RhoGAP). Residue numbers are indicated. Domains from each protein used in this study are highlighted in color. B) Amino acid sequence of p190RhoGAP residues 1083–1111, with pTyr-1087 and pTyr-1105 indicated. C) SAXS scattering profiles for apo (blue) and peptide-bound (red) SH2-SH3-SH2 samples, averaged across the protein elution peaks. D) Guinier analysis (top) and the plotted residuals between the data and fit (bottom) are shown. There are no systematic deviations from linearization for either sample indicating high quality data. E) Kratky analysis. The p190-bound sample decays to zero faster than the apo sample. F) Porod-Debeye analysis. The lack of a plateau demonstrates flexibility is present in both samples. G) Pair Distribution Functions. Both proteins are globular, and the extendend tail in the apo sample indicates flexibility. For all panels blue indicates apo curve and red indicates p190-bound curve. See also Supplementary Figure 2 and Supplementary Figure 3.

p120RasGAP is the only member of the Ras GAP family to contain SH2 domains ⁵. The SH2 domain is a long-studied signaling module ¹⁶ that transmits and integrates intracellular signals from activated nonreceptor and receptor tyrosine kinases (RTKs) by binding directly to phosphotyrosine residues in substrate proteins including RTKs themselves ^17–19. This recruitment allows the SH2 domain-containing proteins to transduce downstream functions including regulation of cell growth, differentiation, migration and immune cell response (reviewed in ²⁰). The SH2 domain was first discovered as a conserved noncatalytic region of approximately 100 residues in the oncogenic retroviral kinase v-Fps and related Src family nonreceptor cytoplasmic tyrosine kinases like Src and Abl ²¹. Soon afterwards, this domain was identified in other proteins including p120RasGAP, phospholipase C gamma (PLC γ), phosphoinositide-3 kinase (PI3-K), and the adaptor protein Crk, ^6,15,22,23, with the human proteome ultimately found to contain approximately 120 SH2 domains in 111 SH2 domain-containing proteins ²⁴. Notably several of these, including p120RasGAP, contain two SH2 domains.

The SH2 fold is well-characterized, and many structural studies have assessed the interactions of these domains with phosphotyrosine partners to reveal a highly conserved domain fold with a common peptide binding mechanism ^25–27. This “two-prong” binding mode has two binding pockets, one required for interaction with phosphotyrosines termed the ‘phosphotyrosine binding pocket’ and one to facilitiate selectivity, termed the ‘specifity pocket’ ^28,29. The phosphotyrosine binding pocket contains a highly conserved amino acid sequence -Phe-Leu-Val-Arg-, or FLVR, motif whose arginine makes a salt bridge with phosphotyrosine. The specificity pocket allows binding parter recognition by selectively interacting with the residue located +3 positions from phosphotyrosine ^25–27. These conserved features are almost completely ubiquitous in SH2 domains, but p120RasGAP contains an exception. Although the N-terminal of its two SH2 domains (N-SH2) is a canonical SH2 ³⁰, its C-terminal SH2 domain (C-SH2) uses a unique mode of phosphotyrosine coordination distinct from all analyzed SH2 domains because it is not mediated by the FLVR arginine residue ³¹. We therefore proposed the C-SH2 domain of p120RasGAP to be ‘FLVR-unique’ ³¹. The consequences of this unique binding mechanism are not fully understood, but both SH2 domains are known to engage phosphotyrosine motifs ^17,19, suggesting that understanding the structure of multidomain arrangements may provide insight into specificity and signaling from this protein.

Interactions of p120RasGAP with phosphotyrosine containing binding partners have been shown to be mediated by the SH2 domains. These phosphorylated partner proteins include the Rho GTPase activating protein p190RhoGAP ^32,33, the adaptor proteins p62Dok (Dok-1) ¹⁹, p56Dok (Dok-2 or Dok-R) ³⁴, SH2 domain-containing adapter protein B (SHB) ³⁵, and receptor tyrosine kinases including the platelet derived growth factor (PDGF) receptor ^17,36 and the ephrin type B (EphB) receptors ³⁷. Many of these binding partners contain two closely spaced phosphotyrosine sites, raising the possibility of simultaneous engagement of the two p120RasGAP SH2 domains with a single protein target. The best described example is p190RhoGAP-A (ARHGAP35, p190RhoGAP), which contains two sites of tyrosine phosphorylation in its unstructured middle domain: Tyr-1087 and Tyr-1105 (Figures 1A and 1B) ^38,39. p190RhoGAP is a major regulator of Rho activity in the cell, helping to control cell migration, adhesion, and polarity among other functions ⁴⁰, and its recruitment to p120RasGAP is observed to be important for downregulation of Rho signaling ⁴¹. Crosstalk between Ras and Rho pathways may be impacted by this p120RasGAP-p190RhoGAP interaction ⁴² raising the potential that bidentate SH2-phosphotyrosine interactions act as a mechanism to achieve intersection of these signaling pathways.

The potential that p120RasGAP can simultaneously engage tandem phosphotyrosine sites has not been explored at the molecular level. Therefore in this study, we examine the molecular basis for binding of a doubly phosphorylated p190RhoGAP peptide to the SH2-SH3-SH2 region of p120RasGAP. We find that the bidentate interaction is extremely tight compared to single SH2-phosphotyrosine interactions, observe by small angle X-ray scattering that the interaction stabilizes a conformational compaction, and find in the X-ray crystal structure a unique orientation of the SH2 domains that is facilitated by the engagement of the doubly phosphorylated p190RhoGAP peptide. Our findings provide a detailed examination of how p120RasGAP binds a key doubly-phosphorylated binding partner, p190RhoGAP.

RESULTS

Determination of affinity of p120RasGAP for p190RhoGAP phosphopeptide

To probe the ability of p120RasGAP to interact with doubly-phosphorylated binding partners, we assessed its interaction with p190RhoGAP. This is the best studied of the documented doubly-phosphorylated binding partners, but the literature remains unclear about the ability of p120RasGAP to engage p190RhoGAP in a bidentate manner. To resolve this question, we began by expressing and purifying the SH2-SH3-SH2 region (residues 174–444) of human p120RasGAP and synthesizing a bis-phosphorylated peptide derived from the sequence of human p190RhoGAP-A (residues 1083–1111, phosphorylated on pY1087 and pY1105) (Figures 1A and 1B). We directly measured the affinity of the interaction using isothermal titration calorimetry (ITC) and find that the doubly-phosphoylated p190RhoGAP binds the SH2-SH3-SH2 region of p120RasGAP with a dissociation constant (K_d) of 10 ± 6 nM (Table 1, Supplementary Figure 1 and Supplementary Table 1). This high-affinity binding is approximately 15- and 30-fold tighter than the affinities we measured previously for the individual SH2 domains with singly phosphorylated p190RhoGAP peptides: the N-terminal SH2 domain (N-SH2) binds a peptide encompassing pY1105 with an affinity of 300 ± 100 nM, and the C-terminal SH2 (C-SH2) binds a peptide encompassing pY1087 with an affinity of 150 ± 40 nM ^30,31 (Table 1). These measurements support the hypothesis that p120RasGAP engages the two phosphorylated tyrosine sites in p190RhoGAP in a bivalent manner, since a similar enhancement of affinity has been observed in other tandem SH2 systems ^43–45.

Table 1. K_d values from ITC of p120RasGAP with p190Rho-GAP derived phosphopeptides.

See also Supplementary Table 1, Supplementary Table 2 and Supplementary Figure 1.

p120RasGAP/p190RhoGAP	Kd (nM)	n	ΔH (kJ/mol)	ΔS (J/mol*K)
SH2-SH3-SH2/pYpY	10 ± 6	0.6 ± 0.1	−130 ± 20	−280 ± 80
N-SH2/pY-1105 31	300 ± 100	0.68 ± 0.01	−70 ± 10	−95 ± 40
C-SH2/pY-1087 30	150 ± 40	0.91 ± 0.08	−60 ± 7	−70 ± 30

Solution scattering studies of p120RasGAP SH2-SH3-SH2 region reveals a compact module in presence of phosphopeptide

When other tandem SH2 domains engage doubly-phosphorylated partners they can achieve conformational rearrangements important for functional regulation which is driven by significantly tighter interactions with the bis-phosphorylated partner than a single phosphotyrosine target ^43–45. Our affinity measurements therefore lead us to ask whether the engagement of the doubly-phosphorylated peptide might coordinate a conformational response in p120RasGAP. To investigate this we employed small-angle X-ray scattering (SAXS). SAXS can provide a solution state assessment of protein size, shape and flexibility, and when performed in series with size exclusion chromatography (SEC) and multi-angle light scattering (MALS) can provide a comprehensive conformational analysis. We investigated two samples by SEC-MALS-SAXS: the apo p120RasGAP SH2-SH3-SH2 protein, and p120RasGAP SH2-SH3-SH2 bound to the doubly-phosphorylated p190RhoGAP peptide. Both samples are found to be monodisperse by SEC and MALS and have molecular weights consistent with monomers, indicating that the phosphopeptide does not bridge two SH2-SH3-SH2 protomers to dimerize them (Table 2 and Supplementary Figure 2). Likewise, scattering intensity and radius of gyration (R_g) across the SEC elution peaks are consistent for both, indicating homogeneous protein samples (Table 2 and Supplementary Figure 2). We therefore were in a position to conduct a more thorough SAXS analysis and comparison (Figure 1C thru 1G).

Table 2.

SAXS Measurements of p120RasGAP SH2-SH3-SH2 apo and in complex with p190Rho-GAP derived phosphopeptide

	Apo SH2-SH3-SH2	p190-bound SH2-SH3-SH2
Rg (Å) (from Guinier analysis)	26.1 ± 0.1	24.2 ± 0.1
Dmax (Å) (from P(R))	103	79
Volume of Correlation (Vc) MW (kDa)	30.8	31.1
MALS MW (kDa)	34.7 ± 0.5	35.2 ± 0.1
Theoretical MW (kDa)	31.4	34.9

We first assessed the samples by binning the SEC elution profiles and observe no concentration dependency for shape or flexibility (Supplementary Figure 3). This allowed us to average the scattering frames of each protein peak and to calculate scattering profiles for the apo and phosphopeptide-bound samples (Figure 1C). Although similar volume of correlation (Vc) and MALS molecular weight estimates are observed for both samples and consistent with monomers (Table 2), Guinier and radius of gyration (R_g) analyses reveal the apo protein to be conformationally more extended than the peptide-bound p120RasGAP SH2-SH3-SH2 protein (Figure 1D and Supplementary Figure 3B and 3G). Similarly, dimensionless Kratky analysis demonstrates a slower decay to zero at high q for the apo protein compared to the peptide-bound, suggesting an overall compaction of the SH2-SH3-SH2 domains upon peptide binding (Figure 1E). Nonetheless, when we conduct Porod-Debye analysis to assess flexibility ⁴⁶ we find a failure to plateau within the Porod-Debye q range for both samples, suggesting both contain some inherent flexibility (Figure 1F). These observations are supported by the pair distribution function (P(R)) curves in which the apo p120RasGAP sample displays a larger D_max value than the peptide-bound sample (Figure 1G and Table 2). Although the overall shape of the P(R) distribution indicates that both samples exist as globular proteins, the extended tail in the apo P(R) curve agrees with an increased flexibility as observed in the Kratky and Porod-Debeye analyses. Furthermore, since the two samples have similar mass, altered rigidity is likely the cause of the D_max difference. Taken together, the results of our SAXS analysis reveals a conformational compaction and a lessening in flexibility of the SH2-SH3-SH2 protein upon engagement of the p190RhoGAP bisphosphorylated peptide.

Co-crystal structure of p120RasGAP SH2-SH3-SH2 with p190RhoGAP phosphopeptide

Our SAXS analysis suggests conformational changes in the SH2-SH3-SH2 region of p120RasGAP upon binding the doubly-phosphorylated p190RhoGAP peptide. We therefore sought a crystallographic analysis of the protein to provide insights into the compaction of p120RasGAP upon its recruitment to p190RhoGAP. We conducted crystallization trials for both apo and peptide-bound p120RasGAP SH2-SH3-SH2 protein. Interestingly, while the complex with the doubly-phosphorylated p190RhoGAP peptide readily crystallized in a number of conditions, we were unable to obtain crystals for apo p120RasGAP SH2-SH3-SH2. We inferred this to be a consequence of the increased conformational flexibility of the apo sample compared to peptide-bound, consistent with our SAXS analysis, and continued our assessment by focusing on the co-crystals. We obtained a 1.95 Å co-crystal structure which we determined by molecular replacement and contains two copies of both p120RasGAP SH2-SH3-SH2 and doubly-phosphorylated p190RhoGAP peptide per asymmetric unit (Table 3). Both copies show strong electron density for the bound phosphopeptide, and since there is high similarity (RMSD 0.5 Å over 269 Cα) we focus our analysis on the copy with complete peptide connectivity (Supplementary Figure 4).

Table 3.

X-ray diffraction data collection and refinement statistics.

Data Collection	p120RasGAP SH2-SH3-SH2 + p190RhoGAP pYpY peptide
PDB accession code	8DGQ
Wavelength (Å)	0.97920
Resolution range (Å)	50 – 1.95 (2.02 – 1.95)
Space group	P 21 21 21
Cell dimensions a, b, c (Å)	55.2 113.5 119.1
α, β, γ (°)	90, 90, 90
Unique reflections	55390
Multiplicity	25.5 (26.0)
Completeness (%)	100 (100)
Mean I/σI	12.4 (1.9)
Wilson B factor (Å2)	38.3
Rpim (%)	3.4 (46.2)
CC½ (%)	99.8 (71.2)
Refinement
Resolution range (Å)*	49.65 – 1.95 (2.00 – 1.95)
Reflections used in refinement	55376
Reflections used for Rfree	2010
% Reflections used for Rfree	3.6
Rwork (%)*	21.4 (37.3)
Rfree (%)*	25.6 (43.8)
No. of non-hydrogen atoms
All	5079
Protein, peptide, solvent	4453, 372, 254
Residue numbers modeled:
p120RasGAP chains A, B	172–443, 173–443
p190RhoGAP chains U, V	1085–1109, 1084–1094/1103–1111
Buried surface area (Å2)‡
Chains A/U	1332, 1139
Chains B/V	1344, 1123
RMSD
Bond lengths (Å)	0.006
Bond angles (°)	0.8
Ramachandran plot (%)
Favored, allowed, outliers	98.2, 1.8, 0
Rotamer outliers (%)	1.0
MolProbity† clashscore	4.1 (99th percentile)
Average B factor (Å2)
Overall	50.6
Protein overall	50.6
Chains A, B, U, V	48.7, 50.1, 72.7, 56.2
Water, other solvent	49.5, 56.8

^*Values in parenthesis represent the highest resolution shell (2.0 – 1.95 Å)

^†Molprobity v4.5.1 ⁷⁴

^‡As determined in Pisa Server v1.52 ⁴⁸

In the co-crystal structure of p120RasGAP and p190RhoGAP we find a compact arrangement where the two SH2 domains directly contact one another and the SH3 domain stacks against the C-terminal SH2 domain (Figure 2A). In this arrangement, the phosphotyrosine binding pockets of both SH2 domains are located on the same face of the compact surface, and this allows the phosphopeptide to bind in a bi-dentate fashion to allow a stoichiometric 1:1 interaction between p120RasGAP and p190RhoGAP. The interactions between p120RasGAP and p190RhoGAP occur in a head-to-tail orientation where the N-terminal SH2 domain binds the C-terminal phosphotyrosine (pY1105) and the C-terminal SH2 domain binds to N-terminal phosphotyrosine (pY1087). The SH3 domain does not contact the phosphopeptide. Overall, this arrangement presents an extended surface in p120RasGAP with dual SH2 binding sites engaging the doubly-phosphorylated p190RhoGAP peptide.

Figure 2. Structure of p120RasGAP SH2-SH3-SH2 bound to doubly-phosphorylated peptide of p190RhoGAP.

A) Ribbon diagram of the co-crystal structure of p120RasGAP SH2-SH3-SH2 domain (purple, green, cyan, also shown as transparent surface) in complex with p190RhoGAP phosphopeptide (yellow). Labelled are the individual domains of p120RasGAP, and pY-1087 and pY-1105 of p190RhoGAP. The amino- and carboxy-termini of the peptide are also labeled. The two views are related by a 90 degree rotation as indicated. B) Comparison of theoretical and experimental scattering curves (top panel) and residuals of the fit (bottom panel). Theoretical scattering curve (red line) generated in FoXS ⁶⁹ of p120RasGAP chain A in complex with p190RhoGAP chain U overlayed with the experimental scattering data from the p190-bound SH2-SH3-SH2 sample (grey dots), χ² = 1.52. C) Electron density calculated in DENSS ⁴⁷ using the experimental scattering data, superposed with the cocrystal structure of p120RasGAP with p190RhoGAP phosphopeptide (superposition performed in BioXTAS RAW ⁷⁰). Map is contoured at 2σ, illustrated in CCP4mg ⁷¹. See also Supplementary Figure 4.

Comparison of the co-crystal structure with solution scattering data

Upon observing the compact arrangement of the domains in the co-crystal structure, we evaluated whether this is consistent with our SAXS data. We generated a theoretical scattering plot from our co-crystal structure and compared it to the experimental SAXS data of the peptide-bound SH2-SH3-SH2 protein by overlaying the curves and examining residuals, and find that the theoretical scattering curve agrees well with our experimental values (χ² = 1.52) (Figure 2B). Using DENSS ⁴⁷ we calculated a three dimensional particle electron density map from from the experimental solution scattering and observe that when superposed with the co-crystal structure the electron density envelope fits well, with each domain situated inside each of the three lobes of the envelope (Figure 2C). Therefore, the co-crystal structure represents a similar size and shape as measured in our solution studies.

Peptide binding conformation of SH2-SH3-SH2 module

The phosphotyrosine residues in p190RhoGAP are separated by 18 residues (Figure 1B), and of the 25 peptide residues modeled in our structure (1085–1109), only 4 (Lys-1096, Asn-1099, Glu-1101 and Glu-1102) do not contact p120RasGAP (Pisa Server ⁴⁸). Accordingly, these residues have higher average B factors compared to the remainder of the phosphopeptide which contacts the p120RasGAP protein directly. From the N-terminus of the peptide, the pY-1087 region binds to the C-terminal SH2 domain and then sharply makes a β-turn (Figure 3A). A short flexible region follows, then a 3/10 helix precedes the canonical SH2-phosphopeptide interactions between pY-1105 and the N-terminal SH2 domain (Figure 3A and 3B). Interestingly, the positively charged pTyr-binding pockets are closely spaced (the phosphotyrosine phosphates are ~18Å distal) (Figure 3C) but because the linear motifs around pY-1087 and pY1105 are almost perpendicular to one another the peptide requires the sharp β-turn to bind both sites simultaneously (Figure 3A). The interaction interfaces between p120RasGAP and p190RhoGAP are highly conserved over evolution (Figure 3D). Likewise, the sequence surrounding the two phosphotyrosine sites in p190RhoGAP is extremely well conserved over both isoforms (p190RhoGAP-A and p190RhoGAP-B) and through evolution (Supplementary Figure 5).

Figure 3. Orientation of tandem SH2 domains.

In all panels, the SH3 domain is located behind C-SH2 and is not visible. A) Schematic of SH2 domain arrangement. Two pronged binding mode of SH2 domains is indicated by two grey spheres corresponding to phosphotyrosine and specificity pockets. B) Surface representation with the atoms contacting the phosphopeptide colored lilac (N-SH2) and teal (C-SH2). C) Surface electrostatics (computed in CCP4mg ⁷¹). D) Surface conservation mapped onto the structure (Consurf ⁷²). See also Supplementary Figure 4.

Comparison of N-SH2 structures; role of FLVR-unique binding mode in C-SH2

Inspection of the individual SH2 domains in the current structure has the potential to reveal new insights into p120RasGAP binding to phosphotyrosine peptides. We previously reported the cocrystal structures of the isolated p120RasGAP SH2 domains in complex with short p190RhoGAP-derived phosphopeptides: N-terminal SH2 domain in complex with a pY-1105 phosphopeptide ³⁰ and C-terminal SH2 domain in complex with a pY-1087 phosphopeptide ³¹. Upon superposition we observe that the overall fold of the N-terminal SH2 domain is experimentally identical in the two structures (RMSD 0.6 Å over 103 Cαs) (Supplementary Figure 6A). Consistent with our previous study, N-SH2 possesses the two-pronged binding mode for the phosphopeptide (Figure 4A). The phosphotyrosine binding site in N-SH2 in our current structure is again dominated by the canonical SH2 domain ‘FLVR’ Arginine Arg-207 (residue βB5), Arg-188 (αA2) and Ser-209 (βB7) (Figure 4B), and the specificity pocket by Phe-230, Ile-241, Tyr-256, and Leu-262, which form a shallow hydrophobic pocket to recognize p190RhoGAP’s +3 Pro-1108 (Figure 4C). In contrast, differences in peptide binding are observed at the newly observed 3/10 helix (residues 1100–1102), which was not present in the individual N-SH2 domain structure and allows the contribution of additional hydrogen bonds from Glu-1100 to Arg-231 which is near the phosphotyrosine binding site (Figure 4B).

Figure 4. Details of phosphotyrosine binding to N-SH2 and C-SH2.

A) Ribbon diagram of p120RasGAP N-SH2 (purple) bound to p190RhoGAP peptide (yellow sticks). Peptide residues 1098–1109 are included. The locations of the phosphotyrosine and specificity pocket are indicated by dashed circles and labelled. B) Details of the phosphotyrosine binding site interactions. C) Details of the specificity site interactions. The C-terminus of the peptide is labelled. D) Ribbon diagram of C-SH2 with p190RhoGAP phosphotyrosine peptide (showing residues 1085–1095 only) in yellow sticks. The location of the phosphotyrosine binding and specificity pockets are circled. E) Details of the phosphotyrosine interaction site. F) Details of the specificity pocket and extended interface sites. Salt bridges and hydrogen bonds are depicted as dashed lines. See also Supplementary Figure 5 and Supplementary Figure 6.

Like the N-terminal SH2 domain, we oberve that the C-terminal SH2 domain folds similarly to the crystal structure of the isolated C-terminal SH2 domain in complex with a pY-1087 phosphopeptide (RMSD 0.8 Å over 105 Cαs) (Supplementary Figure 6B) ³¹. It also utilizes the two-pronged binding mode (Figure 4D) and importantly possesses the same novel ‘FLVR unique’ phosphotyrosine binding site as observed prior, where the FLVR arginine residue Arg-377 does not coordinate the phosphotyrosine directly but rather makes an intramolecular salt bridge with Asp-380 in the BC loop (at the BC1 position) (Figure 4E). However, when comparing structures we find that in the current structure the phosphate of pTyr-1087 binds approximately ~2 Å deeper into the pocket via rotation of the phosphate about the tyrosine Cζ to OH bond and a translation of the Cα position by 0.6 Å; subsequently, additional H-bonds are made by Lys-358 and Arg-398 (Supplementary Figure 6C). Interestingly, when we compare the binding site to our previous apo C-SH2 crystal structure in the absence of phosphopeptide ³¹, we find that a sulfate ion contributed by the crystallization solution is present at the phosphotyrosine binding site (Supplementary Figure 6D) and that the location of this sulfate ion is similar (distance ~0.9 Å) compared to the current phosphotyrosine phosphate; also, the neighboring sidechains of Lys-358 and Arg-398 bind the sulfate directly and are in similar rotamers as in the current structure. These differences may suggest that the C-terminal SH2 domain can bind phosphates in a variety of positions within the phosphotyrosine binding pocket. We also find the C-SH2 domain specificity pocket is further defined by three major features: (i) The shallow hydrophobic pocket formed by the sidechains of Phe-399, Met-411, Tyr-426 and Ile-431 to recognize proline at the peptide +3 position (Figure 4F), (ii) the formation of a salt bridge with His-425 and a hydrogen bond with Tyr-426 in αB by the Asp-1092 at the +5 position of the peptide (Figure 4F), and (iii) the β-turn between Pro-1090 (at the +3 position) and Ala-1093 which seems to require a small residue, in this case Ala-1093, at the +6 position of the peptide (Figure 4F). This peptide β-turn resembles the conformation of phosphopeptides bound to the Grb2 SH2 domain, which instead form a β-turn between the phosphotyrosine and the +3 residue (Ogura et al., 1999; Rahuel et al., 1996). This turn forms because of steric clashes in the specificity pocket by residue Trp-121 which blocks binding of the +3 residue to the specificity site and instead prefers Asn at the +2 position (Supplementary Figure 6E and 6F). Notably, all three features of the specificity pocket are conserved in our previous structure of the isolated C-SH2 domain in complex with a pTyr-1087 phosphopeptide ³¹.

Details of the interdomain interfaces

These insights into the SH2 domains of p120RasGAP are futher enhanced by analysis of the conformational arrangement of the SH2-SH3-SH2 module (Figure 5A). As discussed above, the N-terminal SH2 domain contacts only C-SH2, via N-SH2 βA (residues 182–186) and αA (residues 188–198) interacting with C-SH2 BC loop (residues 381–385) and DE loop (residues 402–406) (Figure 5B). The involvement of the C-SH2 BC loop in this interaction is notable, as it also binds the pTyr-1087. Therefore, we postulate that the position of the BC loop may be dictated by the FLVR-unique Arg-377 binding to Asp-380 and by the extensive contact of both the mainchain and sidechain of Asn-381 with the phosphotyrosine (Figure 5B); in turn, hydrogen bonds between Asn-381, Thr-382 and Lys-185, and between Ser-402/Pro-403 with Arg-194 help stabilize the SH2-SH2 interface. Thus, the conformation of the BC loop in the FLVR-unique phosphotyrosine binding pocket may help facilitate the interdomain interaction with the N-terminal SH2 domain.

Figure 5. Interdomain contacts in p120RasGAP.

A) Overview of structure with interfaces detailed in parts (B) and (C) boxed and labelled. The SH3 domain N-Src and RT loops, which together compromise the typical polyproline binding, are labelled. A putative PxxP peptide at the typical binding site is depicted in orange, based on superposition with the Fyn SH3 domain bound to a PxxP ligand of its binding partner 3BP2 (PDB ID: 1FYN ⁷³), RMSD 1.1 Å over 56 equivalent Cα positions. B) Details of interface between N-SH2 and C-SH2. C) Details of interface between SH3 and C-SH2 involving interdomain linker residues (in blue). See also Supplementary Figure 7.

Additionally, we find that the SH3 domain and the C-terminal SH2 domain engage in extensive contacts, largely mediated by the linker (residues Val-339 to Lys-349) that connects these domains (Figure 5C). An extensive hydrogen bonding and van der Waals network is formed by the linker to both C-SH2 and the SH3 domain. Interestingly, although the linker is flexible in previously determined NMR structures of the isolated SH3 domain (PDB IDs: 2M51, 2GQI) and C-SH2 domain (PDB ID: 2GSB) (Supplementary Figure 7A and 7B), the conformation of the linker and the SH3/C-SH2 orientation are consistent in both copies in our crystal structure (Supplementary Figure 7C); furthermore, this relative conformation of these domains is also consistently predicted by all six Alphafold models of p120RasGAP to date across different species (Supplementary Figure 7D) ⁴⁹ with RMSD over the SH3/C-SH2 region (residues 282–442) of 1.5 ± 0.1 Å over 160 equivalent Cα positions. In contrast, these current Alphafold models do not accurately predict the orientation of the N-SH2 compared to our structure (Supplementary Figure 7D), potentially suggesting that domain rearrangement of the N-SH2 position occurs in response to phosphopeptide binding.

Since prototypical SH3 domains contain a binding site for polyproline containing PxxP motifs comprised of the n-Src and RT loops, we identified this putative binding site in the SH3 domain in our structure and find that it is exposed (Figure 5A and Supplementary Figure 7D). This raises the possibility that the extensive contacts of the linker between the SH3 and C-SH2 domain serve as an anchor to stabilize the relative orientations of these domains, so that all three canonical sites of the SH2-SH3-SH2 module are poised to bind partner peptides/proteins simultaneously (Figure 5A).

DISCUSSION

In this study we have conducted an extensive analysis of the interactions between p120RasGAP and p190RhoGAP. These proteins are essential regulators of their respective GTPase cascades, and their interaction is well documented ^38,50, but how this is achieved has remained relatively poorly understood. Our study suggests that engagement of doubly-phosphorylated p190RhoGAP with the SH2-SH3-SH2 region of p120RasGAP is an extremely tight stoichiometric interaction which may provide stringent selectivity for RasGAP interaction partners; furthermore, the interaction induces a conformational change in the arrangement of the domains which exposes all three Src homology domain binding sites for partner interactions.

The interactions we observe with the intact p120RasGAP SH2-SH3-SH2 module with a doubly-phosphorylated peptide are significantly tighter than those observed in the previous studies of the individual SH2 domains with singly phosphorylated binding partners that showed affinities between 15- and 30-fold weaker ^30,31 than the bidentate SH2-bis-phosphopeptide interaction. This non-linear increase in affinity is consistent with other tandem SH2 domain interactions with doubly-phosphorylated peptides where affinity enhancements have been reported including ZAP-70, Syk, SHP-2, PLC- γ1 and PI3K p85 ^43–45. Additionally, a major observation in other tandem SH2-pTyr interactions is an enhancement in selectivity by as much as 10,000-fold (compared to 20–50 fold for individual SH2 domains) due to combinatorial effects of specificity for the individual phosphotyrosine sequences, spacing between phosphotyrosines, and interaction orientation of the SH2 domains ^44,51} which can lead to increases in dwell time of the partners due to rebinding effects ⁵². Thus, we propose that the recognition of doubly phosphorylated p190RhoGAP by the tandem SH2 domains of p120RasGAP may therefore represent a highly selective signaling gate with stringent requirements, potentially to facilitate temporal control of complex formation and/or signaling strength.

Interestingly, although tyrosine phosphorylation of p190RhoGAP is well documented ^{38,50,53–58}, only Tyr-1105 has been shown to be phosphorylated by Src, Arg, or Btk tyrosine kinases in vitro ^53–55. Along these lines, it has been suggested previously that this single tyrosine phosphorylation event is sufficient for recruitment of p120RasGAP to p190RhoGAP ⁵⁴. Nevertheless, there are over 1000 records of tyrosine phosphorylation at Tyr-1087 and over 3000 for Tyr-1105 reported in Phosphosite ³⁹, and phosphorylation of both sites is reported in PeptideAtlas ⁵⁹, raising the possibility that phosphorylation at both sites can occur simultaneously. However, it remains difficult to observe both sites simultaneously by standard trypsin/mass spectrometry approaches since the two tyrosine residues reside on separate trypsin peptides. Thus, the particular kinase and upstream signaling required to trigger phosphorylation at both Tyr-1087 and Tyr-1105 remain to be fully established. We propose that under certain cellular conditions, double phosphorylation of p190RhoGAP may allow tight recruitment to p120RasGAP.

Concommitent with this proposition, we observe conformational changes in the relative arrangement of the three Src homology domains of p120RasGAP upon binding to the doubly phosphorylated peptide. When other tandem SH2 domain proteins bind doubly phosphorylated peptides, conformational changes can alter enzymatic activity ⁶⁰ or impart enhanced affinity and specificity between binding partners ^43,44. What the effect of binding the doubly phosphorylated peptide has on the interactions of cellular p120RasGAP remains to be studied, but the binding could induce alterations in enzymatic activity or changes in binding partners, exposure/shielding of posttranslational modification sites, or targeted degradation. Future studies are needed to reveal the potential effects.

To date, the only structurally-defined SH2 domain tandems bound to a double phosphotyrosine-containing peptide are ZAP-70 ⁶¹ and Syk ⁶² bound to peptides from immunoreceptor tyrosine-based activation motifs (ITAM), and now p120RasGAP bound to p190RhoGAP. Upon comparison, we find that the orientation of the p120RasGAP SH2 domains is different to the arrangements observed for Syk and ZAP-70 bound to doubly phosphorylated ITAM peptides (Supplementary Figure 8), suggestive of a broad conformational landscape for tandem SH2 domain arrangement. The uncovering of structural details of other tandem SH2 pairs in complex with their phosphotyrosine targets may highlight an even wider range of SH2 domain orientations.

Our analysis of the orientation of the Src homology domains may reveal further insights into p120RasGAP’s interactions. Our data reveals that the canonical SH3 domain binding site is in an exposed position, this suggests that p120RasGAP may act as a node or hub to localize phosphorylated parters and SH3-binding partners together. Interestingly, however, the SH3 domain was previously shown to contain an unusual PxxP binding site ⁶³ and this has remained consistent in our structure and the five previous single domain depositions (Supplementary Figure 7E). Futhermore multiple peptide screening attempts have failed to identify a consensus PxxP binding motif ^64,65, suggesting that it may not interact with polyproline peptides but instead with folded proteins. One possibility is DLC1 RhoGAP which has been identified as a binding partner of the p120RasGAP SH3 domain ^66,67 in a complex that we have characterized structurally and biochemically ⁶⁸. The orientation of the SH2 and SH3 domains in p120RasGAP may therefore allow simulataneous use of all three Src homology binding sites for recruitment of protein partners. Taken together, our analyses therefore provide insights into the interactions of two major Ras and Rho regulating GAP proteins.

STAR Methods

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Titus Boggon (titus.boggon@yale.edu).

Materials availability

Plasmids generated in this study may be available upon request to the Lead Contact.

Data and code availability

• Coordinates and structure factors have been deposited at the Protein Data Bank under accession nymber 8DGQ and are publicly available as of the date of publication. X-ray diffraction images have been deposited at SBGrid Data Bank: https://data.sbgrid.org/dataset/879/ and are publicly available as of the date of publication. Accession numbers are listed in the key resources table. This paper analyzes existing, publicly available data. These accession numbers for the datasets are listed in the key resources table.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Key resources table.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Bacterial and virus strains
Escherichia coli (E. coli) Rosetta(DE3) competent cells	Millipore Sigma (Novagen)	Cat#70954
Chemicals, peptides, and recombinant proteins
p190RhoGAP-A (1083 to 1111) phosphopeptide	Genscript	Uniprot ID: Q9NRY4
His6-p120RasGAP SH2-SH3-SH2 (174–444) protein	This paper	UniProt ID: P20936
Ni-NTA Agarose resin	Qiagen	Cat#30210
HiLoad 16/600 Superdex 75 pg	Cytiva	Cat#28989333
MonoQ 5/50 GL	Cytiva	Cat#17516601
Superdex 75 Increase 10/300 GL	Cytiva	Cat#29148721
Superdex 200 Increase 10/300 GL	Cytiva	Cat# 28990944
Critical commercial assays
QuikChange Multi Site-Directed Mutagenesis Kit	Agilent	Cat#200514
QuikChange Site-Directed Mutagenesis Kit	Agilent	Cat#200519
Deposited data
p120RasGAP SH2-SH3-SH2 (apo) SAXS data	This work	SASBDB ID: SASDPM4
p120RasGAP SH2-SH3-SH2 bound to p190RhoGAP phosphopeptide SAXS data	This work	SASBDB ID: SASDPN4
p120RasGAP SH2-SH3-SH2 bound to p190RhoGAP phosphopeptide atomic coordinates and structure factors	This work	PDB: 8DGQ
p120RasGAP SH2-SH3-SH2 bound to p190RhoGAP phosphopeptide X-ray diffraction images	This work	https://data.sbgrid.org/dataset/879/
NMR structure of the SH3 domain of human RAS p21 protein activator (GTPase activating protein) 1	Joint Center for Structural Genomics (JCSG), Partnership for T-Cell Biology (TCELL)	PDB: 2M51
NMR Solution structure of the SH3 domain of human Ras GTPase-activating protein 1	RIKEN Structural Genomics/Proteomics Initiative (RSGI)	PDB: 2GQI
NMR Solution structure of the second SH2 domain of human Ras GTPase-activating protein 1	RIKEN Structural Genomics/Proteomics Initiative (RSGI)	PDB: 2GSB
Crystal structure of the RasGAP SH3 domain at 1.5 Angstrom resolution	(Ross, et al., 2007)	PDB: 2J05
Crystal structure of the RasGAP SH3 domain at 1.8 Angstrom resolution	(Ross, et al., 2007)	PDB: 2J06
Crystal structure of Fyn SH3 domain bound to a 3BP2 PxxP ligand	(Musacchio, et al., 1994)	PDB: 1FYN
Crystal structure of N-Terminal SH2 domain of the p120RasGAP bound to a p190RhoGAP phosphotyrosine peptide	(Jaber Chehayeb, et al., 2020)	PDB: 6PXC
Crystal structure of C-terminal SH2 domain of p120RasGAP in complex with p190RhoGAP phosphotyrosine peptide	(Jaber Chehayeb, et al., 2020)	PDB: 6WAY
Crystal structure of C-terminal SH2 domain of p120RasGAP	(Jaber Chehayeb, et al., 2020)	PDB: 6WAX
Crystal structure of a RAS p21 protein activator (RASA1) SH2 domain (C-terminal)	Joint Center for Structural Genomics (JCSG), Partnership for T-Cell Biology (TCELL)	PDB: 4FSS
Crystal structure of the tandem SH2 domain of Syk bound to a dually Tyrosine-phosphorylated ITAM peptide	(Futterer, et al., 1998)	PDB: 1A81
Crystal structure of the tamden SH2 domains of ZAP-70 bound to ITAM derived peptide from T cell receptor	(Hatada, et al., 1995)	PDB: 2OQ1
Oligonucleotides
p120RasGAP SH2-SH3-SH2 PCR Forward GTAGGATCCACCGCTCCTCCAACT	This paper	N/A
p120RasGAP SH2-SH3-SH2 PCR Reverse CTCGAGTTACTGCATTGGTACAGG	This paper	N/A
C236S mutagenesis oligo CCATTTTAGGATTATTGCTATGAGTGGAGATTACTACATTGGTGG	This paper	N/A
C261S 5’ mutagenesis oligo CTAATAGGTTATTACAGTCATGTTTCTAGTTTGCTTAAAGGAGAAAAATTACTTT	This paper	N/A
C261S 3’ mutagenesis oligo AAAGTAATTTTTCTCCTTTAAGCAAACTAGAAACATGACTGTAATAACCTATTAG	This paper	N/A
C372S mutagenesis oligo TGACAGTTGGTCAAGTCAGCAGTTTTCTTGTGAGG	This paper	N/A
C402S mutagenesis oligo ACCAATGAAAATATTCAGCGATTTAAAATAAGTCCAACGCCAAACAA	This paper	N/A
Recombinant DNA
p120RasGAP (human)	This paper	Uniprot ID: P20936
modified pET vector	This paper	N/A
Software and algorithms
XDS	(Kabsch, et al., 2010)	https://xds.mr.mpg.de/
XSCALE	(Kabsch, et al., 2010)	https://xds.mr.mpg.de/
Phenix	(Adams, et al., 2010)	https://phenix-online.org/
Phaser	(McCoy, et al., 2007)	https://phenix-online.org/
Phenix AutoBuild	(Terwilliger, et al., 2008)	https://phenix-online.org/
SBGrid	(Morin, et al., 2013)	https://sbgrid.org/
Coot	(Emsley, et al., 2010)	https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/
Phenix Refine	(Afonine, et al., 2012)	https://phenix-online.org/
Molprobity v. 4.5.1	(Williams, et al., 2018)	http://molprobity.biochem.duke.edu/index.php
Pisa server	(Krissinel, et al., 2007)	https://www.ebi.ac.uk/pdbe/pisa/
Consurf	(Ashkenazy, et al., 2016)	https://consurf.tau.ac.il/consurf_index.php
CCP4mg v. 2.11.0	(McNicholas, et al., 2011)	https://www.ccp4.ac.uk/ccp4-legacy/MG/download/
BioXTAS RAW	(Hopkins, et al., 2017)	https://bioxtas-raw.readthedocs.io/en/latest/index.html#
GNOM	(Svergun, et al., 1992)	https://www.embl-hamburg.de/biosaxs/gnom.html
FoXS	(Schneidman-Duhovny, et al., 2013)	https://modbase.compbio.ucsf.edu/foxs/
DENSS	(Grant, et al., 2018)	https://denss.ccr.buffalo.edu/

EXPERIMENTAL MODEL AND SUBJECT DETAILS

All recombinant proteins were expressed in Rosetta(DE3) cells (Millipore Sigma) grown in Miller’s Luria Broth base (Life Technologies) at 37°C shaking at 200 rpm in a Forma Orbital Shaker (ThermoFisher), with protein expression induced at OD₆₀₀ = 0.6 with 0.2 mM isopropyl β-d-thiogalactopyranoside (IPTG) overnight at 16°C with shaking.

METHOD DETAILS

Protein expression and purification

cDNA encoding the (N)SH2-SH3-(C)SH2 domains of human p120RasGAP (residues 174–444 Uniprot ID: P20936), was amplified by PCR and ligated into a modified pET vector containing an N-terminal hexahistidine (His₆−) tag and TEV protease recognition site. Codons encoding four native Cysteine residues – Cys-236, Cys-261, Cys-372 and Cys-402 - were mutated to serine codons by QuikChange Multi-Site Mutagenesis kit (Agilent) to prevent formation of inter-molecular disulfide bonds in the expressed protein. Protein expression was performed in Rosetta(DE3) cells, which were cultured in 1 L Luria Broth at 37 °C to an OD 600 of 0.6 – 0.8, then cooled to 18°C and protein expression induced with 0.1 mM isopropyl β-d-thiogalactopyranoside (IPTG). Cells were harvested by centrifugation, resuspended in lysis buffer (50 mM HEPES pH 7.3 and 500 mM NaCl), and lysed by freeze-thaw cycles in the presence of lysozyme followed by sonication. Total cellular lysate was clarified by centrifugation at 5000 × g at 4°C, and applied to Ni-NTA agarose resin (Qiagen) for 1 h at 4 °C to capture His₆-tagged protein. Beads with bound protein were washed with 20 column volumes of wash buffer (lysis buffer supplemented with 20 mM imidazole). Protein was then eluted with step-wise gradient of lysis buffer containing increasing concentrations of imidazole (100 mM, 250 mM, 500 mM). Elution fractions containing p120RasGAP protein were pooled, mixed with His₆-tagged TEV protease, and dialyzed overnight against wash buffer stirring at 4°C. Proteins were then recovered from the dialysis tubing and reapplied to Ni-NTA resin to caputure the cleaved His₆-tag, uncleaved His₆-p120 protein, and His₆-TEV protease; flow-through containing untagged p120 protein was collected and applied to size exclusion chromatography (HiLoad 16/600 Superdex 75, Cytiva) in buffer containing 20 mM Tris pH 7.4, 250 mM NaCl. Purified protein was finally concentrated in a centrifugal filter with MWCO 10,000 Da (Amicon Ultra, Millipore Sigma). 1 L of Rosetta(DE)3 cell culture resulted in a final purified protein yield of approximately 10 mg. For small angle X-ray studies, p120RasGAP SH2-SH3-SH2 protein was purified by anion exchange chromatography (MonoQ 5/50 GL, Cytiva) in Tris pH 8.0 buffer, and size exclusion chromatography (HiLoad 16/600 Superdex 75, Cytiva) in buffer containing 20 mM Tris pH 8.0 and 350 mM NaCl. Protein concentrations were determined by Nanodrop absorbance at 280 nm using the calculated extinction coefficient of 45840 M⁻¹cm⁻¹ based on the primary sequence.

Peptide synthesis

A synthetic 29 amino acid peptide of sequence DPSDpY(1087)AEPMDAVVKPRNEEENIpY(1105)SVPHDS native to p190RhoGAP residues 1083 to 1111 (Uniprot Q9NRY4) phosphorylated at Tyr-1087 and Tyr-1105, with N-terminal acetylation and C-terminal amidation, were commercially synthesized (GenScript). The lyophilized peptide was reconstituted in ultrapure water to 10 mg/ml (2.8 mM).

Crystallization, data collection, structure determination and refinement:

For co-crystallization with the p190RhoGAP pY-1087 phosphopeptide, p120RasGAP (N)SH2-SH3-(C)SH2 protein at 0.3 mM (10 mg/ml) was premixed with phosphopeptide at 0.6 mM final concentrations for a 2:1 final peptide:protein molar ratio. Initial crystal screening was conducted with Index HT and PEGRx HT kits (Hampton Research) using a TTP Labtech Mosquito in sitting drop vapor diffusion trays at room temperature. Clusters of thin plate crystals were observed in PEG Rx HT position H10, containing 1.0 M Sodium malonate pH 5.0, 0.1 M Sodium acetate trihydrate pH 4.5, 2% w/v Polyethylene glycol 20,000, with a 1:1 (v:v) protein:reservoir solution ratio. Optimization of this crystallization condition was achieved in hanging drop vapor diffusion plates at room temperature, with a reservoir buffer containing 0.5 M Sodium Malonate pH 5.0, 0.1 M sodium acetate trihydrate salt, 6% PEG 20,000 (w/v) and ratio of 1:1 (v:v) protein:reservoir solution. Single crystals with approximate dimensions 300 × 200 × 50 μm were harvested from the drop and cryopreserved in reservoir solution supplemented with 25% ethylene glycol and flash cooled in liquid nitrogen.

X-ray data collection on a single crystal was performed at Northeastern Collaborative Access Team (NE-CAT) beamline 24-ID-C at Argonne National Laboratory Advanced Photon Source. Four individual 180° sweeps of X-ray data were collected at different positions along the length of the crystal, X-ray data were integrated separately in XDS ⁷⁵, and scaled together in XSCALE ⁷⁵. Data were processed in spacegroup P212121 with unit cell dimensions a=55.2, b=113.5, 119.1 α = β = γ = 90° to 1.95 Å resolution. Matthews probability calculator predicts two copies of N-SH2-SH3-C-SH2 per asymmetric unit. Xtriage detected the presence of translational non crystallographic (pseudo) symmetry (tNCS), with an off-origin Patterson peak of height ~ 45%. A molecular replacement solution was found in Phaser ⁷⁶, which utilized the tNCS vector, in spacegroup P2₁2₁2₁, using the solution structure of the p120RasGAP C-SH2 domain (PDB ID: 2GSB, unpublished) and crystal structure of the SH3 domain (PDB ID: 2J05 ⁶³) as search models. The correct solution contained 4 copies of SH2 and 2 copies of SH3, with a final translation function (TFZ) score of 20.2. Autobuilding was performed in Phenix AuoBuild ⁷⁷, which built 536 total residues and correctly placed 500 residues into the sequence (chain A: residues 176–255, 270–275, 281–442; chain B: residues 176–258, 270–275, 281–442). The p120RasGAP SH2-SH3-SH2 proteins were then manually built using Coot ⁷⁸ and refinement in Phenix ^79,80. The connectivity of the individual domains is unambiguous. For model building of the phosphopeptides, Autobuild in Phenix successfully built the backbone trace of residues 1087–1094 in chain U (bound to protein chain A) and residues 1087–1094 and 1105–1108 of chain V (bound to protein chain B). After refinement in Phenix ⁸⁰, the resultant 2Fo-Fc electron density map contoured at 1 sigma and Fo-Fc difference density electron density map contoured at 3 sigma were used to guide manual building in iterative cycles in Coot ⁷⁸ and refinement in Phenix ^79,80. Amino acids were manually added in Coot when the 2Fo-Fc maps showed signal at 1 sigma. For peptide chain U, a complete chain was modeled, with 1 sigma 2Fo-Fc backbone density evident for all but Arg-1098, which refines to a high B factor (125 Ų) to reflect the low-density signal. Crystallography software was compiled by SBGrid ⁸¹. All structure figures were generated in CCP4mg ⁷¹.

SEC-MALS-SAXS

The reconstituted p190RhoGAP phosphopeptide was added to purified p120RasGAP SH2-SH3-SH2 protein to a 2:1 peptide:protein molar ratio (approximately 60 μM peptide to 30 μM protein in 500 μl) and incubated on ice for 30 minutes. Apo SH2-SH3-SH2 or peptide-bound samples were resolved by size exclusion chromatography (SEC, Cytiva Superdex 75 Increase 10/300 GL) in 1X SEC buffer containing 20 mM Tris pH 8, 350 mM NaCl, 1 mM DTT. Matched 1X SEC buffer was used for subsequent SEC-MALS-SAXS. Native PAGE was performed to verify that the protein remained fully bound to peptide following SEC. Samples were concentrated to approximately 10 mg/mL using the same centrifugal filter used for the crystallographic studies and diluted to 7.6 mg/mL with fresh 1x SEC buffer. Protein samples and SEC buffer were flash frozen in liquid nitrogen and shipped overnight to Argonne National Laboratory Advanced Photon Source (APS) for SEC-MALS-SAXS experiments. On the day of data collection the samples were thawed and centrifuged to pellet any precipitates. The samples were injected at room temperature onto a SEC column (Cytiva Superdex 200 Increase 10/300 GL) in 1X SEC buffer at a flow rate of 0.5 mL/min using a 1260 Infinity II HPLC (Agilent). The sample elution was first detected by UV (Agilent UV monitor), followed by Wyatt DAWN HELEOS II MALS+DLS and Wyatt Optilab T-rEX dRI detectors for measuring molecular weight through MALS-DLS-RI analysis. Following MALS-DLS, the sample flowed to the SAXS sample chamber where X-ray scattering data were collected at room temperature by a Pilatus3 X 1M detector at a wavelength of 1.033 Å, camera length of 3.628 m, an exposure time of 0.5 s, and a q-measurement range of 0.0045 to 0.35 Å⁻¹. Data were normalized using an active beamstop containing a silicon PIN diode. All data were collected at APS BioCAT Beamline18ID using synchrotron radiation from Undulator A.

SAXS Analysis

Image files were reduced using BioXTAS RAW (v 2.0.3) ⁷⁰ and total intensity per frame calculated to produce a scattergram. From this scattergram, buffer subtraction was performed by binning and averaging approximately 100 frames of the eluate before elution of the protein from the column, also in BioXTAS RAW. Frames were chosen from the sample peak and averaged based on initial R_g calculations. The averaged frames were saved as a single intensity profile and used for further data analysis. Using BioXTAS RAW as a GUI interface ⁷⁰, Guinier analysis was performed on each intensity profile to determine data quality. Molecular weight estimations were performed in RAW ^82,83. Pair Distribution functions were determined by using GNOM ⁸⁴ in RAW using default parameters. To compare the solution scattering data with the crystal structure, we generated a theoretical scattering profile from the crystal structure coordinates containing chain A/chain U and fit it to the scattering data in FoXS ⁶⁹. DENSS models were created using default parameters in RAW which generated 20 initial models which were then aligned, averaged, and refined ⁴⁷. Superposition of the DENSS electron density envelope with the structure was performed in BioXTAS RAW and the figure was made in CCP4mg ⁷¹. Detailed SAXS information is provided in Supplementary Table 2.

Isothermal Titration Calorimetry

p120RasGAP SH2-SH3-SH2 protein and p190RhoGAP bisphosphorylated peptide were dialyzed overnight against a common buffer containing 20 mM Tris pH 8.0 and 250 mM NaCl in dialysis cartridges (Thermo Scientific Slide-A-Lyzer Dialysis Cassette G2 10,000 MWCO for protein and Spectra/Por Micro Float-A-Lyzer MWCO 100–500 D for peptide). Protein concentration was measured as A280 value (ε = 45840 M⁻¹cm⁻¹) on a Nanodrop Lite (Thermo) instrument. Peptide concentration was determined by Amino Acid Analysis (performed at UC Davis Molecular Structure Facility) and by A280 measurement using an extinction coefficient of 0.404 mM⁻¹cm⁻¹. After degassing using TA instrument’s Degassing Station for 7 minutes, the 190 μL sample cell was filled with 5–10 μM protein and the syringe with 23–50 μM peptide, and the experiment was performed using TA Instrument’s nanoITC by collecting a 300 second baseline before titrating 20 2.5 μL injections during a 300 second injection time. Data were analyzed in TA Instrument’s NanoAnlyze by subtracting the blank (constant) model from the independent model. Final reported values are those averaged by model outputs of four titrations performed over 4 days. To ensure there was no heat released due to the protein interacting with itself, we titrated ITC buffer into SH2-SH3-SH2 protein, and to ensure there was no heat released due to p190 peptide interacting with itself, we titrated p190 peptide into ITC buffer.

Bioinformatics and structure analysis

Structural biology applications used in this project were compiled and configured by SBGrid ⁸¹. Sequence conservation was performed with SHOOT.bio ⁸⁵ and Consurf ⁷², and sequence alignments in ClustalO ⁸⁶. Figures were generated in CCP4mg ⁷¹.

QUANTIFICATION AND STATISTICAL ANALYSIS

Mean ±s.d. and N are indicated in figure and table legends. Reproducibility of results and sample sizes are discussed in the figure and table legends. Data from all ITC experiments are shown in supplemental material.

Highlights:

• Two SH2 domains in p120RasGAP bind dual phosphotyrosine residues in p190RhoGAP

• Solution scattering supports compaction of the domains upon phosphopeptide binding

• Crystal structure reveals a compact arrangement that resembles a SH2 domain tandem

• Affinity measurements support synergistic binding