TXNIP (thioredoxin-interacting protein) negatively regulates the antioxidative activity of thioredoxin and participates in pleiotropic cellular processes. Its deregulation is linked to various human diseases, including diabetes, acute myeloid leukaemia and cardiovascular diseases. The E3 ubiquitin ligase Itch (Itchy homologue) polyubiquitinates TXNIP to promote its degradation via the ubiquitin–proteasome pathway, and this Itch-mediated polyubiquitination of TXNIP is dependent on the interaction of the four WW domains of Itch with the two PPxY motifs of TXNIP. However, the molecular mechanism of this interaction of TXNIP with Itch remains elusive. In the present study, we found that each of the four WW domains of Itch exhibited different binding affinities for TXNIP, whereas multivalent engagement between the four WW domains of Itch and the two PPxY motifs of TXNIP resulted in their strong binding avidity. Our structural analyses demonstrated that the third and fourth WW domains of Itch were able to recognize both PPxY motifs of TXNIP simultaneously, supporting a multivalent binding mode between Itch and TXNIP. Interestingly, the phosphorylation status on the tyrosine residue of the PPxY motifs of TXNIP serves as a molecular switch in its choice of binding partners and thereby downstream biological signalling outcomes. Phosphorylation of this tyrosine residue of TXNIP diminished the binding capability of PPxY motifs of TXNIP to Itch, whereas this phosphorylation is a prerequisite to the binding activity of TXNIP to SHP2 [SH2 (Src homology 2) domain-containing protein tyrosine phosphatase 2] and their roles in stabilizing the phosphorylation and activation of CSK (c-Src tyrosine kinase).
- PPxY motifs
TXNIP (thioredoxin-interacting protein), together with ARRDC1–ARRDC5 (α-arrestin domain-containing protein 1–5) constitute the α-arrestin family. A shared characteristic feature of ARRDC2, ARRDC3, ARRDC4 and TXNIP is that they all possess two highly conserved C-terminal PPxY motifs [1–3] (Figure 1A). TXNIP is initially recognized as an important modulator of the redox system . In the cytoplasm, TXNIP inhibits the endogenous antioxidative function of thioredoxin-1, allowing oxidative stress to accumulate in the cell [4,5]. In the mitochondria, TXNIP binds to thioredoxin-2, and promotes phosphorylation and activation of ASK1 (apoptosis signal-regulating kinase 1), leading to the release of cytochrome c from the mitochondria and eventual apoptosis . TXNIP also plays a critical role in glucose and lipid metabolism and is an important tumour suppressor in various cancers [5,7–10]. Consistent with its critical roles in diverse cellular processes, TXNIP is tightly regulated at multiple levels [3,11]. Intracellular stability of TXNIP is controlled via the ubiquitin–proteasome pathway . The E3 ubiquitin protein ligase Itch (Itchy homologue) negatively regulates TXNIP, as Itch overexpression increases the rates of polyubiquitination and proteasomal degradation of TXNIP. Correspondingly, intracellular TXNIP levels increase when Itch is knocked down . Therefore Itch-mediated degradation of TXNIP serves as an important means to modulate TXNIP protein levels, and thus fine-tunes the activity of TXNIP in oxidative stress, metabolism and apoptosis . Given its involvement in multiple important signalling pathways, dysregulation of TXNIP has been linked to several human diseases, including cardiovascular diseases  and cancers, especially acute myeloid leukaemia .
Itch belongs to the Nedd4 (neural-precursor-cell-expressed developmentally down-regulated 4)-like family of E3 ubiquitin ligases, which all share three conserved regions: a Ca2+/lipid-binding (C2) domain in the N-terminus, two to four WW domains in the middle and a HECT (homologous with E6-associated protein C-terminus)-type catalytic ligase domain in the C-terminus  (Figure 1A). The C2 domain targets membranes and membrane proteins , the HECT domain accepts ubiquitin from E2 ubiquitin-conjugating enzymes , and the WW domains determine substrate selectivity through specific interactions with amino acid sequence motifs in target proteins . The WW domain is a ubiquitous structural and functional unit found in a large number of otherwise unrelated proteins . The WW domains present in E3 ligases typically belong to group I, which is defined by the ability to target ligands containing a PPxY motif [19,21]. The four WW domains of Itch bind to the two PPxY motifs of TXNIP. These structural domains and motifs do not contribute equally to the interaction and the subsequent Itch-mediated degradation of TXNIP . Details of the molecular mechanism remain unclear. The tyrosine residue of the PPxY motif may undergo phosphorylation, an important post-translational modification for regulating cellular signalling pathways . This tyrosine phosphorylation mark can be read by a large number of protein domains, such as the SH2 (Src homology 2) and BRCT (breast cancer early-onset 1 C-terminal) domains , but how it can modulate the interaction and function of TXNIP has not yet been reported. A recent study showed that TXNIP's PPTY motif binds to SH2 domains of SHP2 (SH2 domain-containing protein tyrosine phosphatase 2), which prevents the dephosphorylation and deactivation of CSK (c-Src tyrosine kinase), whereas mutation from PPTY to PPTA disrupts this interaction . However, whether this interaction depends on tyrosine phosphorylation is not clear.
In the present study, we show that, although all four WW domains of Itch displayed binding activities to both PPxY motifs of TXNIP, their respective binding affinities varied drastically, from the micromolar to millimolar range. The simultaneous engagement of multiple domains significantly enhanced the binding between Itch and TXNIP, suggesting multivalent binding as a mechanism governing selectivity and affinity. We also solved the high-resolution crystal structure of the complex formed by the Itch WW3–WW4 tandem domains and the PPCY motif of TXNIP. Moreover, we found that tyrosine phosphorylation of PPxY motifs in TXNIP abolished its binding activity to Itch, but promoted its interaction with SHP2, demonstrating that tyrosine phosphorylation of PPxY motifs of TXNIP plays a key role in dictating the choices of binding partners of TXNIP and thereby the downstream biological signalling outcomes.
MATERIALS AND METHODS
Protein expression and purification
The DNA fragments corresponding to four WW domains (WW1, residues 324–362; WW2, residues 356–394; WW3, residues 436–474; WW4, residues 475–514) of Itch and tandem WW1–WW2 domains (residues 324–394), tandem WW3–WW4 domains (residues 433–521) and tandem WW1–WW4 domains (residues 324–521) were subcloned into a modified pET28-GST vector to generate N-terminally GST/His-tagged fusion proteins. The DNA fragment corresponding to native TXNIP peptide with two PPxY motifs (residues 327–382) was subcloned into a modified pET28-MHL vector to generate N-terminally His-tagged fusion proteins. The plasmids of N-terminal (residues 1–104), C-terminal (residues 101–222) and tandem (residues 1–222) SH2 domains of SHP2 were generously provided by Dr Karen Colwill (Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, New York, NY, U.S.A.). Recombinant proteins were overexpressed in Escherichia coli BL21(DE3) Codon plus RIL cells (Stratagene) at 15°C and purified by affinity chromatography on Ni-NTA (Ni2+-nitrilotriacetate) resin (Qiagen) followed by thrombin or TEV (tobacco etch virus) protease treatment to remove the tag. The protein was further purified on a Superdex 75 gel-filtration column (GE Healthcare). For crystallization experiments, purified protein was concentrated to 10 mg/ml in a buffer containing 20 mM Tris/HCl, pH 7.5, 150 mM NaCl and 1 mM DTT.
ITC (isothermal titration calorimetry)
All TXNIP peptides, except for native TXNIP peptide with two PPxY motifs, were synthesized by Peptide 2.0 Inc. For ITC measurements, the concentrated proteins were diluted in 20 mM Tris/HCl, pH 7.5, and 150 mM NaCl. Likewise, the freeze-dried peptides were dissolved in the same buffer and the pH was adjusted by adding NaOH. Tyrosine-containing peptide concentrations were estimated with absorbance spectroscopy using the molar absorption coefficient, ε280=1280 M−1·cm−1. All measurements except those of binding between tandem WW1–WW2 and 2_PPxY_short peptide were performed at 25°C, using a VP-ITC microcalorimeter (GE Healthcare). Protein with a concentration of 50–100 μM was placed in the cell chamber, and the peptides with a concentration of 1–3 mM in syringe were injected in 25 successive injections with a spacing of 180 s and a reference power of 13 mcal/s (1 mcal=4.184 mJ). Control experiments were performed under identical conditions to determine the heat signals that arise from injection of the peptides into the buffer. Data were fitted using the single-site binding model within the Origin software package (MicroCal).
Owing to limited protein yield, the WW1–WW2 ITC was performed using a Nano-ITC microcalorimeter (TA Instruments). The binding data should be consistent with those from the regular ITC instrument, confirmed by ITC results of WW3–WW4 using both regular and Nano-ITC instruments. GST/His-tagged fusion protein with a concentration of 50 μM was placed in the cell chamber, and 2_PPxY_short peptide (lacking the linker sequence between the two PPxY motifs) with a concentration of 0.5 mM was injected in 25 successive injections with a spacing of 120 s at 25°C. Control experiments were performed under identical conditions to determine the heat signals that arise from injections of the peptide into the buffer or into the GST/His-tag-only protein. Data were fitted using the independent model within the NanoAnalyze software package (TA Instruments).
NMR sample preparation and NMR spectroscopy
Uniformly 15N-labelled proteins were prepared by growing bacteria in M9 medium using 15NH4Cl (0.5 g/l) as a stable isotope source. The purified 15N-labelled tandem WW3–WW4 domain proteins were dissolved to a final concentration of 0.25 mM in a buffer containing 20 mM Tris/HCl, pH 7.5, 150 mM NaCl and 10% 2H2O. All NMR experiments were performed at 298 K on a Bruker DMX500 spectrometer. 1H-15N HSQC spectra of Itch protein in the free state or in the presence of TXNIP peptides were recorded. The assignment of Gln465 was extracted from a previous study .
Purified protein was mixed with TXNIP peptides at a 1:3 molar ratio and crystallized using the sitting-drop vapour-diffusion method at 20°C after mixing 0.5 μl of the protein solution with 0.5 μl of the reservoir solution. Crystals were obtained with a reservoir solution containing 30% PEG 4000, 0.2 M MgCl2 and 0.1 M Tris/HCl, pH 8.5, for tandem (WW3–WW4)–PPCY or 20% PEG 3350, 0.2 M ammonium formate for SHP2 SH2 with PPTpY peptide. Before flash-freezing crystals in liquid nitrogen, crystals were soaked in a cryoprotectant consisting of 85% reservoir solution and 15% glycerol.
Data collection and structure determination
Diffraction data were collected at the Advanced Photon Source beam line 19ID. Intensities of symmetry-related reflections were merged with AIMLESS . PHASER  was used for molecular replacement. ARP/WARP  was used for automated model building. COOT , JLIGAND , REFMAC  and PHENIX/MOLPROBITY [32,33] were used for molecular model building, restraint preparation, restrained model refinement and geometry validation, respectively. For Itch-WW3–WW4, initially, diffraction intensities were measured on a rotating copper anode and integrated and scaled with DENZO and SCALEPACK , respectively. The structure was solved by molecular replacement using preliminary co-ordinates of the WW3 crystal structure from this laboratory (PDB code 5DWS). ARP/WARP was used for map improvement  and automated model building. Synchrotron diffraction data of an isomorphous crystal were reduced with XDS  and used for further refinement of the crystallographic model. For SHP2 SH2, synchrotron diffraction data were reduced with DENZO and SCALEPACK. Models for molecular replacement based on PDB codes 3TKZ  and 4XZ0  were prepared using the FFAS03 server  and SCWRL . ARP/wARP was used for phase improvement and automated model building. JLIGAND , the GRADE server (http://grade.globalphasing.org) and MOGUL  were used for restraint preparation.
PDB_EXTRACT  was used to prepare the models for PDB deposition. The IOTBX library , PHENIX and CCP4  programs were used to summarize information for Table 1. Figures of molecular models were generated with PyMOL (Schrödinger; http://www.pymol.org). Potential surfaces were calculated using PyMOL's built-in vacuum electrostatics function.
RESULTS AND DISCUSSION
WW domains of Itch bind to PPxY motifs of TXNIP with various affinities
Previous studies have shown that the WW domains of Itch interact with the PPxY motifs in the C-terminus of TXNIP . To elucidate specific determinants of this interaction, we performed ITC assays using synthetic peptides and recombinant proteins of WW domains (Figure 1). All four WW domains of Itch bound to both PPxY motifs of TXNIP. However, their binding affinities varied, depending on both the WW domains and PPxY motifs (Figure 1B, and Supplementary Figures S1A and S1B). The first two WW domains (WW1 and WW2) bound to either PPxY motif of TXNIP more strongly than the last two WW domains (WW3 and WW4) (Figure 1B, and Supplementary Figures S1A and S1B).
Multivalent engagement of different WW domains and PPxY motifs produce a strong binding avidity
Many proteins contain multiple WW domains that may increase selectivity and affinity for ligands or enhance the proteins' functional diversity by their ability to bind peptide sequences from different proteins. To study the effects of multivalent interactions of the tandem WW domains of Itch on the binding affinity, we performed a series of ITC assays (Figure 1B and Supplementary Figure S1C). We were able to express stable proteins for both WW1–WW2 and WW3–WW4 constructs, but for the WW1–WW2 construct, we were unable to remove the GST tag and the protein yield was low. Nevertheless, the GST tag should not affect the WW-binding ability because, in control experiments, the GST tag did not bind to any peptides used in the study (results not shown). Owing to the limitations of our WW1–WW2 protein preparation, we resorted to use tandem WW3–WW4 protein for most of our assays. A TXNIP peptide containing both PPCY and PPTY motifs (2_PPxY_short, tandem PPxY peptide that lacks the linker sequence between these two PPxY motifs) bound to the tandem WW1–WW2 or WW3–WW4 domains significantly more tightly than any single PPxY peptide bound to either individual Itch WW domain (Figure 1B and Supplementary Figure S1C). Although single WW1 and WW2 showed higher binding affinities than single WW3 and WW4, tandem WW domains (WW1–WW2 or WW3–WW4) bound to peptides containing two PPxY motifs with similar affinities, suggesting that either tandem may play a role in the interaction between Itch and TXNIP. As a control, we measured the binding affinity of the recombinant TXNIP fragment containing both PPCY and PPTY motifs with the linker sequence (2_PPxY_long) and performed ITC against the tandem WW3–WW4 domains (Figure 1B and Supplementary Figure S1C). The observed affinity was comparable with that of the short TXNIP peptide, suggesting that the linker sequence between the motifs was not crucial for multivalent interactions, similar to a previous study on FBP21 (formin-binding protein 21) . Taken together, these results suggest that multivalent engagement of WW domains and PPxY motifs generates a strong binding avidity between TXNIP and Itch.
This multivalent binding mode is consistent with previous findings that binding of full-length ARRDC3 (a homologue of TXNIP) and Nedd4 (another member of the E3 ubiquitin ligases) is mediated by interactions between both of ARRDC3’s PPxY motifs and Nedd4’s WW2–WW3 or WW3–WW4 tandem WW domains . Accordingly, although a single mutation in WW3 weakens the interaction, a double or triple mutation in tandem WW2–WW3, WW3–WW4 or WW2–WW3–WW4 almost abrogates the interaction, also confirmed by co-immunoprecipitation assay in HEK (human embryonic kidney)-293 cells , underlining the importance of multivalent interactions between TXNIP and Itch. The strength of the interaction is also similar to Nedd4’s, which also showed submicromolar binding affinities to ENaC (epithelial Na+ channel) or ARRDC3 [46,47]. The presence of multiple binding sites in these proteins balances between the requirements of dynamic regulation and stabilization of a specific interaction at a locus after initial recruitment. In conclusion, combinations of multiple interacting domains can greatly increase substrate selectivity, binding strength and adaptability of the interaction. Importantly, the multivalent engagement may explain why both PPxY motifs are required for Itch to polyubiquitinate TXNIP, as deletion of either motif abrogates TXNIP proteasomal degradation due to decreased selectivity and affinity .
It is still not clear how the four WW domains of Itch interact with two PPxY motifs of TXNIP, since we could not generate protein containing all four WW domains of Itch, preventing us from performing corresponding binding and structural studies. Considering that tandem WW domains WW1–WW2 and WW3–WW4 exhibited similar binding affinities for peptides containing both PPxY motifs, it is conceivable that the full-length Itch protein should have binding affinity for TXNIP comparable with that of the tandem WW domains, and any two WW domains from the four WW domains might recognize the two PPxY motifs simultaneously, although we favour a model in which four WW domains formed two functional units, such as WW1–WW2 and WW3–WW4, and each of them could recognize both PPxY motifs in vivo, which would double the possibility of the initial recruitment of TXNIP by Itch. We observed a similar phenomenon for the methylarginine Tudor-binding protein SND1 (staphylococcal nuclease and Tudor domain-containing 1), in which multiple methylarginine sites do not significantly enhance its binding to SND1, but would increase the possibility of initial recruitment of the Tudor protein SND1 .
The presence of two binding sites between the tandem WW domains and the bivalent PPxY peptide raised the question of which relative orientation between the bivalent TXNIP peptide and the bivalent Itch WW domains was preferred. To address this question, we performed NMR chemical shift perturbation experiments with tandem WW3–WW4, which we could isotopically label for NMR study. The NMR titration revealed that a peak corresponding to the WW3 residue Gln465, a residue in the XP groove (see below), moved in different directions upon addition of PPCY or PPTY peptides (Supplementary Figure S2A). One possible interpretation is that WW3 domain of Itch bound to the PPCY motif and the WW4 domain bound to the PPTY motif in the context of these tandem WW domains (Supplementary Figure S2B).
Structural basis for the PPxY motif recognition by Itch WW domains
To uncover the molecular basis of the interaction between the Itch WW domains and the TXNIP PPxY motifs at atomic resolution, we determined the crystal structure of tandem WW3–WW4 domains (residues 433–521) in complex with the TXNIP PPCY peptide at 1.40 Å (1 Å=0.1 nm) resolution (Table 1). The WW domains adopted the canonical conformation of a twisted triple-stranded antiparallel β-sheet, with the PPCY peptide packing against the concave side of the β-sheet and aligning approximately in parallel with the middle β-strand of the WW domain, similar to observations in other WW–PPxY complex structures [46,49] (Figure 2). Since the single WW domains show strong sequence similarity (Figure 2A) and both tandem WW domains (WW1–WW2 and WW3–WW4) showed similar binding affinities for TXNIP (Figure 1B), the structure should be representative for the interaction between Itch and TXNIP.
The PPCY peptide has the sequence TPEAPPCYMDVI, corresponding to residues 327′–338′ of TXNIP [hereinafter, TXNIP residue numbers are marked with a prime (′) to distinguish them from the Itch residue numbers] (Figure 1A). In the (WW3–WW4)–PPCY complex, the substructures of WW3–PPCY and WW4–PPCY are similar; we refer to the WW3–PPCY component of (WW3–WW4)–PPCY complex in the following discussion. The residues Pro331′ and Pro332′ of the PPCY peptide form a polyproline type II helix , whereas residues 334′–337′ adopt a helical fold (Figure 2B). The binding site of the peptide is located on one face of the WW sheet and consists of two canonical grooves, namely the XP groove and tyrosine-binding groove (Figure 2). Pro331′ and Pro332′ insert into the XP groove by stacking against the conserved Trp466 and Tyr455 residues of WW3 respectively (Figures 2A and 2B). The side chain of Tyr334′ is accommodated in the tyrosine-binding groove formed by the conserved residues Val457, His459 and Arg462 of WW3 (Figure 2B). In addition to the hydrophobic packing, some hydrogen bonds between the peptide and WW domains contribute to the stabilization of the complex, for example between the carbonyl of Pro332′ and the hydroxy group of Thr464, along with the phenolic hydroxy group of Tyr334′ and Nπ of His459 (Figure 2B). The interaction between Pro332′ and Thr464 is indispensable, since phosphorylation of Thr464 destroys the binding between Itch WW3 domain and LMP2A PPxY peptide .
Phosphorylation inhibits the binding of TXNIP to Itch, but enables recruitment of SH2 domain-containing proteins
The PPxY motif is often subject to phenol phosphorylation, a ubiquitous mechanism that regulates activity and function of proteins. Indeed, Tyr378′ of the TXNIP PPTY motif can be phosphorylated in vivo [51,52]. How would phosphorylation affect TXNIP binding to WW domains? In our ITC assays, no binding was observed between the phosphorylated PPCY (PPCpY) peptide and tandem WW3–WW4 domains (Figure 3A and Supplementary Figure S3), indicating that phosphorylation of the PPxY motif would abolish the interaction between TXNIP and Itch. On the basis of the crystal structure of the WW domains and PPCY peptide complex, phosphorylation would disrupt the hydrogen bond between Tyr334′ and His459/His499 and cause steric hindrance.
Does phosphorylation of the PPxY motif enable recognition by other proteins? There are a number of phosphorylation readers, including the SH2 domain, BRCT domain and 14-3-3 proteins. The SH2 domain is a sequence-specific phosphotyrosine-binding module present in many signalling molecules. Significantly, the TXNIP's PPTY motif binds to SH2 domains of SHP2, which prevents the dephosphorylation and deactivation of CSK . According to the PhosphositePlus server, PPTY's tyrosine residue may be subject to phosphorylation [51,52], but whether this interaction between SHP2 and TXNIP requires tyrosine phosphorylation is not clear. Tyrosine phosphorylation within the PPxY motifs of β-dystroglycan has been shown to block interaction with the WW domains of dystrophin and utrophin, while promoting the recruitment of SH2 domain-containing proteins . Thus we performed ITC assays for the N-terminal, C-terminal and tandem SH2 domains of SHP2 to different tyrosine-phosphorylated or non-phosphorylated PPxY peptides. All of these SH2 domains of SHP2 showed moderate binding to the tyrosine-phosphorylated TXNIP peptides, and the binding depended on the phosphorylation mark (Figure 3A). It remains unclear whether the observed binding between tandem SH2 domains and tandem PPxpY motifs represents a multivalent interaction and whether a putative multivalent interaction would be physiologically relevant.
We crystallized the tandem SH2 domains of SHP2 (residues 1–222) in complex with the PPTpY peptide (Figures 3B) and determined the crystal structure. In the crystallographic model, each SH2 domain binds one PPTpY peptide using similar binding modes (Supplementary Figure S3A), thus the C-terminal SH2–PPTpY complex will be used in the following discussion. The SH2 domain of SHP2 adopts the characteristic SH2 fold, which consists of a core antiparallel β-sheet (βB–βD), two α-helices (αA and αB), which pack on opposite faces of the sheet respectively, and a cap element formed by another small antiparallel β-sheet (C-terminal βD and βE, and a loop instead of βE in our SHP2 model). To facilitate the comparison, we used the secondary-structure notation introduced previously for Src and Lck [53,54]. We did not observe the N-terminal βA, βE and βF and C-terminal βG strands seen in other SH2 domains [53,54]. The peptide is bound in an extended conformation roughly perpendicular to the central β-strands of the SH2 domain (Figure 3B and Supplementary Figure S4B). This binding mode is highly conserved in other SH2 domains [37,53–57].
According to our complex structure and ITC results, the phosphorylation of the tyrosine residue in the PPxY motif appears to be indispensable for complex formation as the unmodified PPxY peptide showed no detectable binding to the SH2 domain of SHP2 (Figure 3A). Taken together, our study suggests a novel regulatory role for the PPxY motifs of TXNIP: the phosphorylation mark on the tyrosine residue not only turns off the signal for Itch-mediated degradation, but also turns on a signal to recruit SH2 domain-containing proteins (Figures 3A and 3C). Phosphorylated PPxY motifs of TXNIP bind to SH2 domains of SHP2, which prevents the dephosphorylation and deactivation of CSK , although it remains to be investigated whether any other SH2 domain-containing proteins in addition to SHP2 are physiological targets of TXNIP.
These new data on the function of PPxY motifs have implications for the regulation of cell signalling pathways in which TXNIP participates. For the interaction between Itch and TXNIP, which regulates the degradation of TXNIP, we propose a multivalent binding model. The model requires further refinement and validation, including the identification of specific pairs of WW domains and PPxY motifs that interact in vivo. Phosphorylation of tyrosine in PPxY motifs of TXNIP would block TXNIP's interaction with Itch, but promote recruitment of SH2-containing proteins such as SHP2. This phosphorylation-dependent switch is reminiscent of our previous study, in which phosphorylation of Ser350 of the PxLPxI/L motif of HDAC4 reduced the binding of ANKRA2 (ankyrin repeat family A2) but recruited 14-3-3 proteins for binding with a strong affinity . Furthermore, this phosphorylation-status-dependent recognition of PPxY motifs of TXNIP by the WW domains of Itch probably plays important functional roles in cellular signalling pathways in vivo, similar to the importance of phosphorylation of Smad that modulates its binding affinities for the WW domains of transcriptional partners such as YAP (Yes-associated protein) and Pin1 (peptidyl-prolyl cis-trans isomerase NIMA-interacting 1), consequently affecting the BMP (bone morphogenetic protein) and TGFβ (transforming growth factor β) signalling pathways . If one considers TXNIP as a node in the regulatory network of interacting proteins, interaction with SH2 domains should define a significant portion TXNIP's connectivity within that network. Whereas the number of potential direct or indirect effects on cellular processes is vast, TXNIP's SH2-containing ‘neighbours’ in the network context have yet to be exhaustively identified and characterized, and the associated WW–SH2 affinity switch's function clarified in the contexts of cellular homoeostasis and pathogenesis.
Yanli Liu, Johnathan Lau and Weiguo Li purified and crystallized the protein. Yanli Liu, Weiguo Li and Ashrut Narula conducted the ITC assays. Wolfram Tempel collected diffraction data and determined the Itch crystal structure. Aiping Dong determined the SHP2 crystal structure. Jinrong Min conceived of and designed the study. Yanli Liu, Johnathan Lau, Li Li, Su Qin and Jinrong Min wrote the paper. All authors approved the final version of the paper.
Co-ordinates and structure factors of Itch WW3–WW4–PPCY and SHP2 SH2–PPTpY were deposited under PDB codes 5CQ2 and 5DF6, respectively.
The Structural Genomics Consortium (SGC) is a registered charity (no. 1097737) that receives funds from AbbVie, Boehringer Ingelheim, the Canadian Institutes of Health Research (CIHR), Genome Canada, Ontario Genomics Institute [grant number OGI-055], GlaxoSmithKline, Janssen, Lilly Canada, the Novartis Research Foundation, the Ontario Ministry of Economic Development and Innovation, Pfizer, Takeda and Wellcome Trust [grant number 092809/Z/10/Z]. This study was also supported by the project of Hubei Key Laboratory of Genetic Regulation and Integrative Biology [grant number GRIB201403], research funds of Central China Normal University (CCNU) from the college's basic research and operation of Ministry of Education (MOE) [grant number CCNU15A05036] and National Natural Science Foundation of China [grant numbers 31500613 and 21272090].
We are grateful to Dr Sachdev Sidhu and Dr Karen Colwill for the gift of constructs of single WW domains of Itch and SH2 domains of SHP2. We thank Dr Scott Houliston for performing the NMR experiment and Dr John Walker for suggestions on crystallographic model refinement. Some results shown in this paper are derived from work performed at Argonne National Laboratory, Structural Biology Center, at the Advanced Photon Source. Argonne is operated by UChicago Argonne, LLC, for the U.S. Department of Energy, Office of Biological and Environmental Research under contract DE-AC02-06CH11357.
Abbreviations: ARRDC, α-arrestin domain-containing protein; BRCT, breast cancer early-onset 1 C-terminal; CSK, c-Src tyrosine kinase; HECT, homologous with E6-associated protein C-terminus; ITC, isothermal titration calorimetry; Itch, Itchy homologue; Nedd4, neural-precursor-cell-expressed developmentally down-regulated 4; SH2, Src homology 2; SHP2, SH2 domain-containing protein tyrosine phosphatase 2; SND1, staphylococcal nuclease and Tudor domain-containing 1; TXNIP, thioredoxin-interacting protein
- © 2016 Authors; published by Portland Press Limited