Grb2 binding induces phosphorylation-independent activation of Shp2

The regulation of phosphatase activity is fundamental to the control of intracellular signalling and in particular the tyrosine kinase-mediated mitogen-activated protein kinase (MAPK) pathway. Shp2 is a ubiquitously expressed protein tyrosine phosphatase and its kinase-induced hyperactivity is associated with many cancer types. In non-stimulated cells we find that binding of the adaptor protein Grb2, in its monomeric state, initiates Shp2 activity independent of phosphatase phosphorylation. Grb2 forms a bidentate interaction with both the N-terminal SH2 and the catalytic domains of Shp2, releasing the phosphatase from its auto-inhibited conformation. Grb2 typically exists as a dimer in the cytoplasm. However, its monomeric state prevails under basal conditions when it is expressed at low concentration, or when it is constitutively phosphorylated on a specific tyrosine residue (Y160). Thus, Grb2 can activate Shp2 and downstream signal transduction, in the absence of extracellular growth factor stimulation or kinase-activating mutations, in response to defined cellular conditions. Therefore, direct binding of Grb2 activates Shp2 phosphatase in the absence of receptor tyrosine kinase up-regulation.


Introduction
The reciprocal process of phosphorylation by kinases, and dephosphorylation by phosphatases, of selected residues regulates the intensity and longevity of intracellular tyrosine kinase-mediated signal transduction. The SH2 domaincontaining tyrosine phosphatase 2, Shp2, (aka. PTPN11) plays a prominent role in this process in a multitude of receptor tyrosine kinase (RTK)-mediated signalling pathways, including activation of the extracellular signal-regulated kinase Erk1/2 (aka. mitogen-activated protein kinase, MAPK) pathway (Neel et al., 1997;Van Vactor et al., 1998;Feng et al., 1999;Tonks & Neel 2001;Araki et al., 2003). Shp2 is ubiquitously expressed in vertebrate cells and consists largely of, in sequential order: two Src homology 2 (SH2) domains (NSH2 and CSH2 respectively); a protein tyrosine phosphatase (PTP) domain; and a C-terminal tail with two tyrosyl phosphorylation sites (Y542 and Y580) and a proline-rich sequence (residues 559 PLPPCTPTPP 568 ).
Shp2 was the first phosphatase to be identified as a human oncoprotein Feng, 2007 andChan et al., 2008), and a large body of experimental and clinical studies has indicated that the hyperactivation of Shp2 contributes to tumour progression in, for example, breast cancer (Bentires-Alj et al., 2004;Bentires-Alj et al., 2006;Bocanegra et al., 2010;Aceto et al., 2012). A great deal of interest has been shown in anti-cancer therapeutic approaches involving down-regulation of Shp2. However, despite extensive investigation, a lack of a mechanistic understanding of its up-regulation in solid cancers has hindered pharmaceutical development. Existing inhibitors targeting phosphatase activity show low selectivity because of the highly conserved amino acid sequences of phosphatase domains.
Crystal structural detail revealed that Shp2 utilises an auto-inhibitory mechanism which prevails under basal conditions (Hof et al., 1998). NSH2 forms an intramolecular interaction with the PTP domain, directly blocking access to the catalytic site, and resulting in a 'closed' state. In this state the NSH2 domain adopts a conformation which contorts its phosphopeptide binding cleft. Gain-of-function somatic mutations which result in the abrogation of the interaction between NSH2 and the PTP domain have been shown to be activating (O'Reilly et al., 2000;Chan et al., 2008). Auto-inhibition is released through one of two mechanisms involving the SH2 domains of Shp2. In the first, both N-and CSH2 interact with a binding partner including a phosphorylated bisphosphoryl tyrosine-based activation motif (BTAM) (Lechleider et al., 1993;Sugimoto et al., 1994;Pluskey et al., 1995;Eck et al., 1996;Cunnick et al., 2001). The second, more controversial mechanism occurs under conditions where Shp2 has been phosphorylated, typically by an RTK (Lu et al., 2001;Araki et al., 2003;Keilhack et al., 2005). This induces an intramolecular, bidentate interaction between the two phosphorylated tyrosine residues in the Cterminus of Shp2 and both NSH2 and CSH2 (Lu et al., 2001;Neel et al., 2003;Sun et al., 2013).
We have previously investigated the constitutive control that the adaptor protein growth factor receptor-bound protein 2, Grb2, exerts over Shp2 in non-stimulated cells (Ahmed et al., 2010;Ahmed et al., 2013). Grb2 consists of an SH2 domain sandwiched between two SH3 domains and is integral to several RTK-mediated signalling pathways. Non-phosphorylated Grb2 exists in a concentration-dependent dimer-monomer equilibrium (K d = 0.8 µM; Ahmed et al., 2015). In addition to the result of depletion of cellular concentration (Timsah et al, 2014), monomeric Grb2 (mGrb2) will also prevail when it is phosphorylated on tyrosine 160 (Y160) in the dimer interface (Ahmed et al., 2015). In the absence of growth factor stimulation Grb2 cycles between the phosphorylated mGrb2, and the non-phosphorylated, typically dimeric state. The former is dependent on constitutive, background RTK activity, e.g. fibroblast growth factor receptor 2 (FGFR2; Lin et al., 2012), whilst the latter results from concomitant Shp2 activity (Ahmed et al., 2015).
In this work we provide molecular mechanistic detail on the activation of Shp2 in the absence of the two phosphorylation-dependent mechanisms highlighted above. We show that in non-stimulated cells mGrb2 is able to greatly enhance Shp2 phosphatase activity via a bidentate interaction involving to two discrete interfaces; 1) between the NSH2 domain of Shp2 and the SH2 of Grb2, and 2) between the PTP domain of Shp2 and the CSH3 of Grb2. The binding of mGrb2 releases Shp2 from its auto-inhibited state and results in an increase in the phosphatase activity independent of kinase-induced stimulation. To endorse our mechanistic model we were able to down-regulate Shp2 activity both in vitro and in a triple negative breast cancer cell line using a polypeptide that blocks the Grb2 SH2 binding region on the Shp2 NSH2 domain.

Shp2 interacts with monomeric Grb2 in the absence of growth factor
Mutation of Y160 to glutamate (a phosphotyrosine charge mimetic) in the Grb2 dimer interface abrogates self-association of the adaptor protein (Ahmed et al., 2015). To characterise the interaction between full-length Shp2 (Shp2 FL ) and monomeric Grb2 (including Y160E mutation, Grb2 Y160E ; Ahmed et al., 2015) we initially used microscale thermophoresis (MST) to measure the affinity of the interaction of full length proteins (K D = 0.33 ± 0.04 µM; Fig 1A). We demonstrated that Grb2 Y160E can bind at two discrete sites on Shp2 using bio-layer interferometry (BLI) on the following four GST-tagged phosphatase constructs: Shp2 FL , the tandem SH2 domains (residues 1-220: Shp2 2SH2 ), the PTP domain (221-524: Shp2 PTP ) and a peptide corresponding to the C-terminal 69 amino acids (525-593: Shp2 C69 ). Both Shp2 FL as well as Shp2 2SH2 polypeptides were able to interact with Grb2 Y160E . The truncated Shp2 PTP also bound, whilst the C-terminal tail failed to interact ( Fig 1B).
These two binding sites were confirmed in an in vitro pull down experiment in which Grb2 Y160E was precipitated by both GST-Shp2 2SH2 and GST-Shp2 PTP (Fig 1C). The interaction with GST-Shp2 PTP was less pronounced. Interaction between Grb2 WT and the Shp2 constructs appears to be negligible suggesting that, under the experimental conditions, the prevailing dimeric Grb2 is unable to interact. The limited complex formation seen with an extended exposure of the blot is again presumed to be with the low population of monomeric protein at equilibrium (Fig 1C inset).
To measure intracellular binding of Shp2 Δ69 (deleted of residues 525-593) to Grb2 Y160E we used fluorescence lifetime imaging microscopy (FLIM) to detect stable complexes through fluorescence resonance energy transfer (FRET) between fluorophore-tagged proteins transfected into HEK293T cells under serum-starved conditions. FRET was confirmed by the left-shift of the fluorescence lifetime from the control lifetime measurements of donor  in the presence of non-specific RFP-alone acceptor (2.2 nsec; (Fig 1D). No interaction between CFP-Grb2 WT and RFP-Shp2 Δ69 was observed since the fluorescent lifetime remains largely unaffected.
Grb2 WT exists in a monomer-dimer equilibrium and CFP-Grb2 WT overexpression would favour dimer at the elevated adaptor protein concentration, hence limiting the availability of mGrb2 to interact with Shp2 Δ69 . To circumvent concentrationdependent dimerization, we used the monomeric CFP-Grb2 Y160E in the FLIM binding assay. FRET between CFP-Grb2 Y160E and Shp2 Δ69 shows a left-shifted fluorescence lifetime by 100 psec shorter than the RFP-alone control or CFP-Grb2 WT (Fig 1D lower panel). This clearly demonstrates that only the monomeric Grb2 Y160E interacts with Shp2. Immunoprecipitation further revealed that Grb2 Y160E constitutively associates with Shp2 FL in the absence of ligand stimulation ( Fig EV1).

The SH2 domain of Grb2 binds to the N-terminal SH2 domain of Shp2
The affinity of Shp2 2SH2 binding to monomeric Grb2 Y160E was determined by MST (K d = 0.28 ± 0.03 µM; Fig 2A). The binding of the pY-mimetic glutamate of Grb2 Y160E to either of the SH2 domains of Shp2 2SH2 was ruled out because on mutating arginine residues normally required for pY binding in the respective SH2 domains of both proteins (R32A/R138A; Shp2 2SH2 R32A/R138A and R86A; Grb2 R86A/Y160E ), Grb2 R86A/Y160E was still capable of binding to Shp2 2SH2 R32A/R138A with similar affinity to the wild type Shp2 2SH2 construct (K d = 0.15 ± 0.01 µM; Fig 2B). Having observed the Shp2 2SH2 interaction with Grb2, we sought to identify whether an individual domain was responsible for the critical contact. GST-tagged Shp2 2SH2 R32A/R138A as well as the individual SH2 domains (N-terminal SH2 domain: Shp2 NSH2 and C-terminal SH2 domain: Shp2 CSH2 ) were used in pull down experiments in which the NSH2, and not the CSH2, domain of Shp2 was shown to be sufficient for Grb2 binding (Fig 2C). Reprobing the blot with an anti-pY antibody revealed that the interaction was not mediated through phosphorylated tyrosine(s) or glutamate on Grb2 Y160E . This was further confirmed through in vitro binding assays using two phosphopeptides corresponding to the phosphorylatable tyrosine residues on Grb2 (pY160 and pY209) which show negligible interaction with Shp2 2SH2 . (Fig EV2A). Our data therefore are consistent with a non-canonical, phosphorylation-independent interaction between Shp2 2SH2 and mGrb2.
Having established the role of Shp2 NSH2 in binding to unphosphorylated Grb2 Y160E , we attempted to establish which domain(s) of Grb2 was required for recognition of the phosphatase. GST-tagged Shp2 NSH2 was captured on a GST BLI sensor and was exposed to the isolated SH2, NSH3 and CSH3 domains of Grb2 (Grb2 SH2 , Grb2 NSH3 and Grb2 CSH3 respectively). The data indicate that Grb2 SH2 binds to Shp2 NSH2 (Fig EV2B), accompanied by a weaker interaction with Grb2 CSH3 . MST experiments on the same isolated domains of Grb2 binding to Shp2 NSH2 confirmed that the interaction with Grb2 SH2 was dominant (K d = 33.8 ± 4.5 µM). Interactions with both the SH3 domains of Grb2 were substantially weaker (Grb2 NSH3 ; K d = 422 ± 21.5 µM, and Grb2 CSH3 ; K d = 207 ± 16.8 µM ( Fig 2D).  Fig EV3). Addition of Grb2 SH2 led to chemical shift perturbations (CSPs) for several resonances, indicating disruption of local chemical environments of specific amino acids as a result of complex formation ( Fig 2E). Comparison of the spectra of the free Shp2 NSH2 , and the Shp2 NSH2 -Grb2 SH2 complex showed that the average CSPs are relatively small.
However, a limited number of residues show pronounced changes ( Fig 2F). The CSPs (>0.0075 ppm) were mapped onto a crystal structural representation of Shp2 NSH2 (PDB code: 2SHP; Figs 2G and 2H). From this it is possible to interpret a mechanism for Grb2 SH2 domain binding. The reported crystal structural data of Shp2 (PDB: 2SHP) reveal a non-phosphorylated auto-inhibited structure in which NSH2 directly interacts with the PTP domain, blocking the access of phosphorylated substrates. We see that a number of residues which are not within the Shp2 NSH2-PTP interface are perturbed by binding Grb2 SH2 (T12, V25, Q57 and Y100; Fig 2H) and are likely to represent a discrete and extended interface formed between Grb2 and Shp2. In addition to this we observe that a subset of residues, that are within the intramolecular auto-inhibited interface also show CSPs (G39, F41, G60, Y62 and K70; Figs 2F, and 2G). For instance, G60 and Y62 are located on the DE loop, which inserts into the PTP domain catalytic site interacting with residues of the catalytic loop of Shp2 (Hof et al., 1998). This suggests that binding of Grb2 SH2 is also able to induce a distal conformational change in Shp2 NSH2 capable of disrupting the autoinhibited intramolecular interface releasing the 'closed' structure and hence activating the phosphatase.

The CSH3 domain of Grb2 interacts with Shp2 PTP
Data shown in Fig 1B and 1C revealed that, as well as binding to Shp2 2SH2 , Grb2 Y160E also formed an interaction with Shp2 PTP . We explored this further using MST and found that the isolated Shp2 PTP binds Grb2 Y160E with a similar affinity to the interaction with Shp2 2SH2 (K d = 0.30 ± 0.03 µM; Fig 3A). To identify which domain of  Table). This interaction is approximately 60-fold weaker than the interaction between the intact PTP domain and Grb2 Y160E suggesting that the peptides do not entirely represent the full site of interaction.
Since Grb2 binds to Shp2 through two apparently discrete interactions largely represented by, 1) Grb2 SH2 binding to Shp2 NSH2 , and 2) Grb2 CSH3 binding to Shp2 PTP , we used isothermal titration calorimetry (ITC) to qualitatively corroborate these two binding events. The binding isotherm shows a biphasic profile resulting from an initial binding event with a stoichiometry of 2:1 Shp2 Δ69 :Grb2 Y160E , which was succeeded by a second binding event with a stoichiometry of unity as more Grb2 Y160E was titrated ( Fig 3D). These data can be reconciled by a model in which in the initial injections Grb2 Y160E is saturated by excess Shp2 Δ69 . Under these conditions Shp2 Δ69 will occupy both binding sites on Grb2 (i.e. Grb2 SH2 binding to Shp2 NSH2 on one Shp2 Δ69 molecule, and Grb2 CSH3 binding to Shp2 PTP on another Shp2 Δ69 molecule).
On further addition of Grb2 Y160E , based on the final stoichiometry of 1:1, the complex involves a bidentate interaction in which both the SH2 and CSH3 domain of a single Grb2 concomitantly bind to the NSH2 and PTP domains of a single Shp2 respectively.

Grb2 increases activity independent of Shp2 phosphorylation
The bidentate binding of Grb2 which includes an interaction with both the NSH2 and the PTP domains of Shp2 implies that complex formation could influence phosphatase activity. From our accumulated data we speculate that the binding of monomeric Grb2 promotes a conformational change releasing Shp2 from the autoinhibited state. To investigate this, we tagged Shp2 Δ69 at both the N-and C-termini with blue fluorescent protein (BFP) and green fluorescent protein (GFP) respectively.  4B). A significant increase in phosphatase activity was measured with the increasing presence of Grb2 Y160E . Importantly, a similar level of turnover of the peptide was observed on comparing the monomeric Grb2-induced phosphatase activity to that observed for phosphorylated Shp2 in the absence of the adaptor protein (pShp2; Fig 4C). These in vitro experiments suggest that under basal conditions Shp2 activity can be up-regulated through binding to monomeric Grb2 alone, and that this activity is at least as high as occurs when the phosphatase is phosphorylated as seen in growth factor-stimulated cells.

Monomeric Grb2 is associated with the up-regulation of Shp2 in cancer
Shp2 activity has been shown to play a key role in cancer progression. Since mGrb2 can promote phosphatase activity in the absence of Shp2 phosphorylation, this could To ascertain whether the elevated pErk1/2 was accrued from increased phosphatase activity resulting from the interaction of Grb2 with Shp2 we identified a polypeptide inhibitor (PI; Appendix S2A), which bound to Shp2 and abrogated Grb2 binding. PI binds to Shp2 NSH2 (K d = 7.24 ± 0.52 µM) and was shown not to bind to either fulllength Grb2 or Shp2 PTP (Fig 4E and Table). Pre-incubating Shp2 NSH2 with PI inhibits the mGrb2-Shp2 NSH2 interaction reducing the affinity by approximately 50-fold (K d = 101 ± 6 µM; compared to K d = 2.3 ± 0.2 µM; Fig 4E). The inhibitory effect of PI keeps Shp2 in its auto-regulated state because its presence does not impact on the binding of Shp2 PTP to Shp2 NSH2 in the closed state (Shp2 PTP + Shp2 NSH2 , K d = 0.25 µM; Shp2 PTP + Shp2 NSH2 + PI, K d = 0.55 µM; Fig 4E and Table). NMR CSP mapping revealed that the binding interface between Shp2 NSH2 and PI overlapped with that of Grb2 SH2 (Fig 4F and 4G). Importantly, the binding of PI does not seem to affect the residues in the interface of the intramolecular NSH2-PTP domain interaction, hence Shp2 is able to maintain its closed, auto-inhibited state (Fig 4H and 4I). The ability of PI to inhibit up-regulation of Shp2 by blocking the interaction with Grb2, but maintaining the intramolecular auto-inhibition interaction is exemplified by showing its ability to reduce phosphatase activity ( Fig 4J). Therefore, the PI provides a useful 'bio-tool' to support the mechanistic detail, as well as enabling us to probe the effect of inhibiting activation of Shp2 by Grb2.
Using our novel PI bio-tool we can demonstrate that inhibition of the mGrb2-Shp2 interaction through increased doses of PI, resulted in a reduction in Erk1/2 signalling in MDA-MB-468 breast cancer cells (Fig 4D). Stimulation of the cells with EGF negates the effect of inhibition of the mGrb2 activation of Shp2. This results from upregulation of kinase activity in the cells as is clearly seen by the appearance of pShp2 and the significant appearance pErk1/2 even in the presence of PI. After 15 minutes in the phosphatase assay the activity is reduced by approximately one third in the presence of the PI. These data provide compelling evidence for the ability of the Grb2-Shp2 driven up-regulation of Erk1/2 signalling to contribute to the cancer pathology represented by this cell line.

Discussion
Kinase and phosphatase activity has to be precisely controlled to limit aberrant signal transduction in cells. It has previously been demonstrated that Shp2 function is dependent on RTK activity, and is accompanied by phosphorylation of tyrosine residues Y542 and Y580 which appears to facilitate the conformational change that

Plasmids
For recombinant protein production in E coli, genes of interest were amplified using standard PCR methods. Following designated restriction enzyme digestions, fragments were ligated into pET28b or pGEX4T1. For cell-based strep-tag pulldown experiments, Grb2 genes (wild type or Y160E) were amplified with an N-terminal strep-tag and cloned in a pcDNA6 vector. For the cellular FLIM study, Grb2 genes were cloned in pECFP, and pcDNA-RFP vectors.

Protein expression and purification
Proteins were expressed in BL21 (DE3) cells. 20 ml of cells grown overnight were used to inoculate 1 litre of LB media with antibiotic (50 µg/ml kanamycin for pET28bbackboned plasmids or 50 µg/ml ampicillin for pGEX4T1-backboned plasmids). The culture was grown at 37 °C with constant shaking (200 rpm) until the OD 600 = 0.7. At this point the culture was cooled down to 18 °C and 0.5 mM of IPTG was added to induce protein expression for 12 h before harvesting. Harvested cells were suspended in Talon buffer A (20mM Tris, 150mM NaCl, and 1mM β-ME, at pH 8.0) and lysed by sonication. Cell debris was removed by centrifugation (20,000 rpm at 4 °C for 1 h). The soluble fraction was applied to an Akta Purifer System for protein purification. Elution was performed using Talon buffer B (20mM Tris, 150mM NaCl, 200mM imidazole and 1mM β-ME at pH 8.0). Proteins were concentrated to 2 ml and applied to a Superdex SD75 column using a HEPES buffer at pH 7.5 (20 mM HEPES, 150mM NaCl, and 1mM TCEP, at pH 7.5).

Pulldown/Precipitation
Purified protein or total HEK293T cell lysate were prepared in 1 ml volume for GST pulldown. 50 μl of 50% glutathione beads were added and incubated for overnight. The beads were spun down at 5,000 rpm for 3 minutes, supernatant was removed and the beads were washed with 1 ml lysis buffer. This washing procedure was repeated five times in order to remove non-specific binding. After the last wash, 50 μl of 2x Laemmli sample buffer were added, the samples were boiled and subjected to SDS-PAGE and western blot assay.
After 24 h cells were seeded onto glass coverslips and allowed to grow for a further 24 h, fixed by the addition of 4% (w/vol) paraformaldehyde (PFA) pH 8.0. Following 20 min incubation at room temperature cells were washed 6-7 times with PBS, pH 8.0. Coverslips were mounted onto a slide with mounting medium (0.1% pphenylenediamine and 75% glycerol in PBS at pH 7.5-8.0). FLIM images were captured using a Leica SP5 II confocal microscope with an internal FLIM detector. CFP was excited at 860 nm with titanium-sapphire pumped laser (Mai Tai BB, Spectral Physics) with 710-920 nm tunability and 70 femtosecond pulse width. A Becker & Hickl SPC830 data and image acquisition card was used for timecorrelated single-photon counting (TCSPC); electrical time resolution was 8 picoseconds with a pixel resolution of 256 x 256. Data processing and analysis were performed using a B&H SPC FLIM analysis software. The fluorescence decays were fitted to a single exponential decay model.

Isothermal Titration Calorimetry, ITC
ITC experiments were carried out using a MicroCal iTC200 instrument (Malvern), and data were analysed using ORIGIN7 software. To avoid heats associated with protein dissociation Grb2 Y160E was titrated into Shp2 Δ69 at 25°C. The heat per injection was determined and subtracted from the binding data. Data was analysed using a single independent site model using the Origin software.

Microscale Thermophoresis, MST
The Grb2 and Shp2 interactions were measured using the Monolith NT.115 MST instrument from Nanotemper Technologies. Proteins were fluorescently labelled with Atto 488 NHS ester (Sigma) according to the manufacturer's protocol. Labelling efficiency was determined to be 1:1 (protein to dye) by measuring the absorbance at 280 nm and 488 nm. A solution of unlabelled protein was serially diluted in the presence of 100 nM labelled protein. The samples were loaded into capillaries (Nanotemper Technologies). Measurements were performed at 25°C in 20 mM HEPES buffer, pH 7.5, with 150 mM NaCl, 0.5 mM TCEP, and 0.005% Tween 20, Data analyses were performed using Nanotemper Analysis software, v.1.5.41 and plotted using OriginPro 9.1.

Biolayer Interferometry, BLI
BLI experiments were performed using a FortéBio Octet Red 384 using Anti-GST sensors. Assays were done in 384 well plates at 25 °C. Association was measured by dipping sensors into solutions of analyte protein for 120 seconds and was followed by moving sensors to wash buffer for 120 seconds to monitor the dissociation process. Raw data shows a rise in signal associated with binding followed by a diminished signal after application of wash buffer.

In Vitro Phosphatase Assay
In vitro Shp2 activity assays were carried out using the Tyrosine Phosphatase Assay System (Promega) according to the manufacturer's manual. In brief, recombinant Shp2 Δ69 was mixed with the phospho-peptide substrate in the presence or absence of different concentrations of Grb2 (Grb2 WT or Grb2 Y160E ). The method is based on measuring the absorbent change generated after formation of a reaction mixture of molybdate:malachite green-phosphate complex of the free phosphate.

Shp2 interacts with monomeric Grb2 in a phosphorylationindependent manner.
A MST measurement of full length Shp2 binding to fluorescent labelled monomeric Grb2. For further details, see Table 1 and Materials and Methods.

B
BLI characterisation of individual Shp2 domains binding to GST-Grb2 Y160E immobilised on a GST sensor. GST-Grb2 Y160E was captured and the sensor and 10µM of each Shp2 protein was used to test the interaction. Black: Shp2 WT , Red: Shp2 2SH2 , Green: Shp2 PTP , Blue: Shp2 C69 . The BLI sensorgram indicates that the Shp2 2SH2 and Shp2 PTP mediate the interaction with Grb2.

C
GST pulldown experiment using GST tagged Shp2 2SH2 and GST tagged Shp2 PTP to precipitate monomeric or dimeric Grb2 (Grb2 Y160E or Grb2 WT respectively). The pulldown results clearly demonstrate the interaction of Shp2 with monomeric Grb2.

D
Fluorescence lifetime imaging microscopy (FLIM) displaying fluorescence resonance energy transfer (FRET) between CFP-Grb2 WT and RFP alone control (top); CFP-Grb2 WT and RFP-Shp2 Δ69 (middle); Grb2 Y160E and RFP-Shp2 Δ69 (bottom). The lifetime-image was generated using false colour range pixel-by-pixel lifetime value corresponding to the average lifetime shown in the histograms.

Figure 2.
Identification of binding domain on Grb2 and Shp2 2SH2 domains.
A MST measurement of Shp2 2SH2 binding to fluorescent labelled Grb2 Y160E . For further details, see Table 1 and Materials and Methods.

C
GST pulldown experiment using GST tagged Shp2 NSH2 , Shp2 CSH2 , Shp2 2SH2 and its SH2 domain deficient mutants. The pulldown results indicate that SH2 domain deficient mutants have no effect on precipitating Grb2 and identify that the NSH2 of Shp2 mediates the interaction with Grb2. Together with the negative result from the phosphotyrosine immunoblotting of precipitated Grb2, it is further confirmed that the Shp2 2SH2 -Grb2 interaction is phosphorylation-independent.

D
MST characterisation of Grb2 individual domains binding to fluorescent labelled Shp2 NSH2 . The result indicate that the SH2 domain of Grb2 predominately binding to Shp2 NSH2 .

E
Interaction of 15 N-labelled Shp2 NSH2 and Grb2 SH2 determined by NMR chemical shift perturbation (CSP) plot for the interaction of Grb2 SH2 to 15 N-labelled Shp2 NSH2 . Minimal chemical shift perturbation upon Grb2 SH2 mapped on Shp2 NSH2 sequence.

F
Comparison of Shp2 NSH2 chemical shifts in the presence and absence of Grb2 SH2 . Overlay of a region of the 15 N-1 H HSQC (heteronuclear single-quantum coherence) spectra of Shp2 NSH2 and Grb2 SH2 complex (red) and of Shp2 NSH2 alone (black).

G
Mapping on the Shp2 N-SH2 domain of the (PDB: 2SHP) of the consensus residues (painted light yellow to red) exhibiting strong CSPs upon Grb2 SH2 binding.

H
Mapping on the surface of Shp2 (PDB: 2SHP) of the consensus residues (painted light yellow to red) exhibiting strong CSPs upon Grb2 SH2 binding. The Shp2 NSH2 domain is in grey, Shp2 CSH2 domain is in blue, and the SHp2 PTP domain is in purple. This identify the potential binding sites on Shp2 NSH2 for Grb2 SH2 binding.   Grb2 Y160E upregulates Shp2 activity in a phosphorylationindependent manner.

A
The effect of mGrb2 binding on Shp2 conformation was studied using steadystate FRET. 0.1µM of Shp2 Δ69 was N-terminally tagged with a BFP and C-terminally tagged with a GFP. Upon Grb2 Y160E titration, the emission of both FRET donor (BFP) and acceptor (GFP) was recorded.

B
In vitro phosphatase assay using recombinant-unphosphorylated Shp2 Δ69 (0.1 µM) and Grb2 (WT or Y160E). Free phosphate generated from hydrolysis of the pY from the substrate peptide (ENDpYINASL) was measured by the absorbance of a malachite green molybdate phosphate complex. The increasing concentration of Grb2 Y160E gradually enhances Shp2 activity while the dimeric Grb2 has no effect on Shp2 activity.

C
In vitro phosphatase assay using both recombinant-unphosphorylated (0.1 µM) and phosphorylated (0.1 µM) full length Shp2 to compare the degree of enhanced phosphatase activity mediated by mGrb2 (10 µM) binding. This assay demonstrates that the mGrb2 binding -induced enhancement of Shp2 activity is comparable to the phosphorylated Shp2.

F
The NMR HSQC titration results indicate that the predominant CSPs on Shp2 NSH2 occur on those residues involved in the auto-inhibition interface. This strongly suggest that binding of mGrb2 results in the conformational change of Shp2 and releases it form it auto-inhibition state (left panel). However, binding of Peptide Inhibitor (PI) results in the less CSPs within the auto-inhibition interface (right panel).

G
Mapping on the surface of Shp2 (PDB: 2SHP) of the consensus residues (painted light yellow to red) exhibiting strong CSPs upon PI binding. The Shp2 NSH2 domain is in grey, Shp2 CSH2 domain is in blue, and the SHp2 PTP domain is in purple. This identify the potential binding sites on Shp2 NSH2 for PI binding.

H
Comparison of residues G60, Y62, and K70 chemical shifts in the presence of Grb2 SH2 (left panel) or PI (right panel). Overlay of a region of the 15 N-1 H HSQC (heteronuclear single-quantum coherence) spectra of Shp2 NSH2 alone (black) and of Shp2 NSH2 and its complex (red). Binding of PI induces less CSPs on residues G60, Y62, and K70.

I
Interaction of 15 N-labelled Shp2 NSH2 and peptide inhibitor (PI) determined by NMR chemical shift perturbation (CSP) plot for the interaction of PI to 15 N-labelled Shp2 NSH2 . Minimal chemical shift perturbation upon PI mapped on Shp2 NSH2 sequence.

J
Using an in vitro phosphatase assay it is demonstrated that the PI inhibits mGrb2-induced upregulation of Shp2 (0.1 µM) phosphatase activity. Black: Shp2 activity towards a substrate peptide in the presence of mGrb2 (10 µM). Red: Shp2 activity towards a substrate peptide in the presence of mGrb2 (10 µM) and PI (10 µM).  Figure EV1: Interaction of Shp2 with monomeric Grb2 at basal state.
(A) Phosphorylation of Grb2 Y160 results in the dimer dissociation. Although there is no trace of phosphorylation on the phosphorylation mimic Grb2 Y160E , we performed experiments to confirm the interaction is phosphorylation-independent. Y160 and Y209 are two major phosphorylation sites on Grb2; synthesised phospho-peptides that bear PY160 or PY209 were used for biophysical characterisation. The PY 160 peptide, shows an affinity of 20 µM in ITC measurement and 206 µM in MST measurement. Bot determined affinity are at least 70-fold weaker than the phosphorylation-independent Shp2 2SH2 and Grb2 Y160E interaction (0.28 ± 0.03 µM). Similar to the PY160 peptide, PY209 peptide has a 180-fold weaker affinity comparing to Shp2 2SH2 and Grb2 Y160E interaction in the MST measurement. (B) In order to determine the domain in Grb2 that is responsible for Shp2 NSH2 binding, GSTtagged Shp2 NSH2 was immobilised on a BLI sensor and individual Grb2 domains (Black: NSH3, Red: SH2, and Blue: CSH3) were used to characterise the interaction. The BLI screen results clearly indicates that the SH2 domain of Grb2 is required for binding to Shp2 NSH2 .