Prediction of short linear protein binding regions

Short linear motifs in proteins (typically 3-12 residues in length) play key roles in protein-protein interactions by frequently binding specifically to peptide binding domains within interacting proteins. Their tendency to be found in disordered segments of proteins has meant that they have often been overlooked. Here we present SLiMPred (short linear motif predictor), the first general de novo method designed to computationally predict such regions in protein primary sequences independent of experimentally defined homologs and interactors. The method applies machine learning techniques to predict new motifs based on annotated instances from the Eukaryotic Linear Motif database, as well as structural, biophysical, and biochemical features derived from the protein primary sequence. We have integrated these data sources and benchmarked the predictive accuracy of the method, and found that it performs equivalently to a predictor of protein binding regions in disordered regions, in addition to having predictive power for other classes of motif sites such as polyproline II helix motifs and short linear motifs lying in ordered regions. It will be useful in predicting peptides involved in potential protein associations and will aid in the functional characterization of proteins, especially of proteins lacking experimental information on structures and interactions. We conclude that, despite the diversity of motif sequences and structures, SLiMPred is a valuable tool for prioritizing potential interaction motifs in proteins.     

On the origin of the stereoselective affinity of Nutlin-3 geometrical isomers for the MDM2 protein

The stereoselective affinity of small-molecule binding to proteins is typically broadly explained in terms of the thermodynamics of the final bound complex. Using Brownian dynamics simulations, we show that the preferential binding of the MDM2 protein to the geometrical isomers of Nutlin-3, an effective anticancer lead that works by inhibiting the interaction between the proteins p53 and MDM2, can be explained by kinetic arguments related to the formation of the MDM2:Nutlin-3 encounter complex. This is a diffusively bound state that forms prior to the final bound complex. We find that the MDM2 protein stereoselectivity for the Nutlin-3a enantiomer stems largely from the destabilization of the encounter complex of its mirror image enantiomer Nutlin-3b, by the K70 residue that is located away from the binding site. On the other hand, the trans-Nutlin-3a diastereoisomer exhibits a shorter residence time in the vicinity of MDM2 compared with Nutlin-3a due to destabilization of its encounter complex by the collective interaction of pairs of charged residues on either side of the binding site: Glu25 and Lys51 on one side, and Lys94 and Arg97 on the other side. This destabilization is largely due to the electrostatic potential of the trans-Nutlin-3a isomer being largely positive over extended continuous regions around its structure, which are otherwise well-identified into positive and negative regions in the case of the Nutlin-3a isomer. Such rich insight into the binding processes underlying biological selectivity complements the static view derived from the traditional thermodynamic analysis of the final bound complex. This approach, based on an explicit consideration of the dynamics of molecular association, suggests new avenues for kinetics-based anticancer drug development and discovery.     

Phage-peptide display identifies the interferon-responsive, death-activated protein kinase family as a novel modifier of MDM2 and p21WAF1

Phage-peptide display is a versatile tool for identifying novel protein-protein interfaces. Our previous work highlighted the selection of phage-peptides that bind to specific isoforms of MDM2 protein and in this work we subjected the putative MDM2-binding proteins to phage-peptide display to expand further on putative protein interaction maps. One peptide that bound MDM2 had significant homology to members of the death-activated protein kinase (DAPK) family, an enzyme family of no known direct link to the p53 pathway. We examined whether a nuclear member of the DAPK family named DAPK3 or ZIP kinase had direct links to the p53 pathway. ZIP kinase was cloned, purified, and the enzyme was able to phosphorylate MDM2 at Ser166, a site previously reported to be modified by Akt kinase, thus demonstrating that ZIP kinase is a bona fide MDM2-binding protein. Native ZIP kinase fractions were then subjected to phage-peptide display and one ZIP kinase consensus peptide motif was identified in p21(WAF1). ZIP kinase phosphorylates p21(WAF1) at Thr145 and alanine-substituted mutations in the p21(WAF1) phosphorylation site alter its ability to be phosphorylated by ZIP kinase. Thus, although ZIP kinase consensus sites were then defined as containing a minimal RKKx(T/S) consensus motif, alternate contacts in ZIP kinase binding are implicated, since amino acid residues surrounding the phospho-acceptor site can effect the specific activity of the kinase. Transfected ZIPK can promote the phosphorylation of p21(WAF1) at Thr145 in vivo and can increase the half-life of p21(WAF1), while the half-life of p21(WAF1[T145A]) is not effected by ZIP kinase. Thus, phage-peptide display identified an interferon-responsive protein kinase family as a novel modifier of two components of the p53 pathway, MDM2 and p21(WAF1), and underscores the utility of phage-peptide display for gaining novel insights into biochemical pathways.     

Defining the specificity space of the human SRC homology 2 domain

Src homology 2 (SH2) domains are the largest family of interaction modules encoded by the human genome to recognize tyrosine-phosphorylated sequences and thereby play pivotal roles in transducing and controlling cellular signals emanating from protein-tyrosine kinases. Different SH2 domains select for distinct phosphopeptides, and the function of a given SH2 domain is often dictated by the specific motifs that it recognizes. Therefore, deciphering the phosphotyrosyl peptide motif recognized by an SH2 domain is the key to understanding its cellular function. Here we cloned all 120 SH2 domains identified in the human genome and determined the phosphotyrosyl peptide binding properties of 76 SH2 domains by screening an oriented peptide array library. Of these 76, we defined the selectivity for 43 SH2 domains and refined the binding motifs for another 33 SH2 domains. We identified a number of novel binding motifs, which are exemplified by the BRDG1 SH2 domain that selects specifically for a bulky, hydrophobic residue at P + 4 relative to the Tyr(P) residue. Based on the oriented peptide array library data, we developed scoring matrix-assisted ligand identification (or SMALI), a Web-based program for predicting binding partners for SH2-containing proteins. When applied to SH2D1A/SAP (SLAM-associated protein), a protein whose mutation or deletion underlies the X-linked lymphoproliferative syndrome, SMALI not only recapitulated known interactions but also identified a number of novel interacting proteins for this disease-associated protein. SMALI also identified a number of potential interactors for BRDG1, a protein whose function is largely unknown. Peptide in-solution binding analysis demonstrated that a SMALI score correlates well with the binding energy of a peptide to a given SH2 domain. The definition of the specificity space of the human SH2 domain provides both the necessary molecular basis and a platform for future exploration of the functions for SH2-containing proteins in cells.     

The conformationally flexible S9-S10 linker region in the core domain of p53 contains a novel MDM2 binding site whose mutation increases ubiquitination of p53 in vivo

Although the N-terminal BOX-I domain of the tumor suppressor protein p53 contains the primary docking site for MDM2, previous studies demonstrated that RNA stabilizes the MDM2.p53 complex using a p53 mutant lacking the BOX-I motif. In vitro assays measuring the specific activity of MDM2 in the ligand-free and RNA-bound state identified a novel MDM2 interaction site in the core domain of p53. As defined using phage-peptide display, the RNA.MDM2 isoform exhibited a notable switch in peptide binding specificity, with enhanced affinity for novel peptide sequences in either p53 or small nuclear ribonucleoprotein-U (snRNP-U) and substantially reduced affinity for the primary p53 binding site in the BOX-I domain. The consensus binding site for the RNA.MDM2 complex within p53 is SGXLLGESXF, which links the S9-S10 beta-sheets flanking the BOX-IV and BOX-V motifs in the core domain and which is a site of reversible conformational flexibility in p53. Mutation of conserved amino acids in the linker at Ser(261) and Leu(264), which bridges the S9-S10 beta-sheets, stimulated p53 activity from reporter templates and increased MDM2-dependent ubiquitination of p53. Furthermore, mutation of the conserved Phe(270) within the S10 beta-sheet resulted in a mutant p53, which binds more stably to RNA.MDM2 complexes in vitro and which is strikingly hyper-ubiquitinated in vivo. Introducing an Ala(19) mutation into the p53(F270A) protein abolished both RNA.MDM2 complex binding and hyper-ubiquitination in vivo, thus indicating that p53(F270A) protein hyper-ubiquitination depends upon MDM2 binding to its primary site in the BOX-I domain. Together, these data identify a novel MDM2 binding interface within the S9-S10 beta-sheet region of p53 that plays a regulatory role in modulating the rate of MDM2-dependent ubiquitination of p53 in cells.     

Atypical recognition consensus of CIN85/SETA/Ruk SH3 domains revealed by target-assisted iterative screening

Target-assisted iterative screening applied to random peptide libraries unveiled a novel and atypical recognition consensus shared by CIN85/SETA/Ruk SH3 domains, PX(P/A)XXR. Confirmed by mutagenesis and in vitro binding experiments, the novel consensus allowed for the accurate mapping of CIN85 SH3 binding sites within known CIN85 interactors, c-Cbl, BLNK, Cbl-b, AIP1/Alix, SB1, and CD2 proteins, as well as the prediction of CIN85 novel-interacting partners in protein databases. Synaptojanin 1, PAK2, ZO-2, and TAFII70, which contain CIN85 SH3 recognition consensus sites, were selectively precipitated from mouse brain lysates by CIN85 SH3 domains in glutathione S-transferase pull-down experiments. A direct interaction of synaptojanin 1 and PAK2 with CIN85 SH3 domains was confirmed by Far Western blotting.     

Differential Recognition Preferences of the Three Src Homology 3 (SH3) Domains from the Adaptor CD2-associated Protein (CD2AP) and Direct Association with Ras and Rab Interactor 3 (RIN3)

CD2AP is an adaptor protein involved in membrane trafficking, with essential roles in maintaining podocyte function within the kidney glomerulus. CD2AP contains three Src homology 3 (SH3) domains that mediate multiple protein-protein interactions. However, a detailed comparison of the molecular binding preferences of each SH3 remained unexplored, as well as the discovery of novel interactors. Thus, we studied the binding properties of each SH3 domain to the known interactor Casitas B-lineage lymphoma protein (c-CBL), conducted a peptide array screen based on the recognition motif PxPxPR and identified 40 known or novel candidate binding proteins, such as RIN3, a RAB5-activating guanine nucleotide exchange factor. CD2AP SH3 domains 1 and 2 generally bound with similar characteristics and specificities, whereas the SH3-3 domain bound more weakly to most peptide ligands tested yet recognized an unusually extended sequence in ALG-2-interacting protein X (ALIX). RIN3 peptide scanning arrays revealed two CD2AP binding sites, recognized by all three SH3 domains, but SH3-3 appeared non-functional in precipitation experiments. RIN3 recruited CD2AP to RAB5a-positive early endosomes via these interaction sites. Permutation arrays and isothermal titration calorimetry data showed that the preferred binding motif is Px(P/A)xPR. Two high-resolution crystal structures (1.65 and 1.11 Å) of CD2AP SH3-1 and SH3-2 solved in complex with RIN3 epitopes 1 and 2, respectively, indicated that another extended motif is relevant in epitope 2. In conclusion, we have discovered novel interaction candidates for CD2AP and characterized subtle yet significant differences in the recognition preferences of its three SH3 domains for c-CBL, ALIX, and RIN3.     

Identifying true protein complex constituents in interaction proteomics: The example of the DMXL2 protein complex

A typical high-sensitivity antibody affinity purification-mass spectrometry experiment easily identifies hundreds of protein interactors. However, most of these are non-valid resulting from multiple causes other than interaction with the bait protein. To discriminate true interactors from off-target recognition, we propose to differentially include an (peptide) antigen during the antibody incubation in the immuno-precipitation experiment. This contrasts the specific antibody-bait protein interactions, versus all other off-target protein interactions. To exemplify the power of the approach, we studied the DMXL2 interactome. From the initial six immuno-precipitations, we identified about 600 proteins. When filtering for interactors present in all anti-DMXL2 antibody immuno-precipitation experiments, absent in the bead controls, and competed off by the peptide antigen, this hit list is reduced to ten proteins, including known and novel interactors of DMXL2. Together, our approach enables the use of a wide range of available antibodies in large-scale protein interaction proteomics, while gaining specificity of the interactions.     

Systematic discovery of new recognition peptides mediating protein interaction networks

Many aspects of cell signalling, trafficking, and targeting are governed by interactions between globular protein domains and short peptide segments. These domains often bind multiple peptides that share a common sequence pattern, or “linear motif” (e.g., SH3 binding to PxxP). Many domains are known, though comparatively few linear motifs have been discovered. Their short length (three to eight residues), and the fact that they often reside in disordered regions in proteins makes them difficult to detect through sequence comparison or experiment. Nevertheless, each new motif provides critical molecular details of how interaction networks are constructed, and can explain how one protein is able to bind to very different partners. Here we show that binding motifs can be detected using data from genome-scale interaction studies, and thus avoid the normally slow discovery process. Our approach based on motif over-representation in non-homologous sequences, rediscovers known motifs and predicts dozens of others. Direct binding experiments reveal that two predicted motifs are indeed protein-binding modules: a DxxDxxxD protein phosphatase 1 binding motif with a K_D of 22 μM and a VxxxRxYS motif that binds Translin with a K_D of 43 μM. We estimate that there are dozens or even hundreds of linear motifs yet to be discovered that will give molecular insight into protein networks and greatly illuminate cellular processes.     

Identification of a novel calcium binding motif based on the detection of sequence insertions in the animal peroxidase domain of bacterial proteins

Proteins of the animal heme peroxidase (ANP) superfamily differ greatly in size since they have either one or two catalytic domains that match profile PS50292. The orf PP_2561 of Pseudomonas putida KT2440 that we have called PepA encodes a two-domain ANP. The alignment of these domains with those of PepA homologues revealed a variable number of insertions with the consensus G-x-D-G-x-x-[GN]-[TN]-x-D-D. This motif has also been detected in the structure of pseudopilin (pdb 3G20), where it was found to be involved in Ca(2+) coordination although a sequence analysis did not reveal the presence of any known calcium binding motifs in this protein. Isothermal titration calorimetry revealed that a peptide containing this consensus motif bound specifically calcium ions with affinities ranging between 33-79 μM depending on the pH. Microcalorimetric titrations of the purified N-terminal ANP-like domain of PepA revealed Ca(2+) binding with a K(D) of 12 μM and stoichiometry of 1.25 calcium ions per protein monomer. This domain exhibited peroxidase activity after its reconstitution with heme. These data led to the definition of a novel calcium binding motif that we have termed PERCAL and which was abundantly present in animal peroxidase-like domains of bacterial proteins. Bacterial heme peroxidases thus possess two different types of calcium binding motifs, namely PERCAL and the related hemolysin type calcium binding motif, with the latter being located outside the catalytic domains and in their C-terminal end. A phylogenetic tree of ANP-like catalytic domains of bacterial proteins with PERCAL motifs, including single domain peroxidases, was divided into two major clusters, representing domains with and without PERCAL motif containing insertions. We have verified that the recently reported classification of bacterial heme peroxidases in two families (cd09819 and cd09821) is unrelated to these insertions. Sequences matching PERCAL were detected in all kingdoms of life.     

Peptide array X-linking (PAX): a new peptide-protein identification approach

Many protein interaction domains bind short peptides based on canonical sequence consensus motifs. Here we report the development of a peptide array-based proteomics tool to identify proteins directly interacting with ligand peptides from cell lysates. Array-formatted bait peptides containing an amino acid-derived cross-linker are photo-induced to crosslink with interacting proteins from lysates of interest. Indirect associations are removed by high stringency washes under denaturing conditions. Covalently trapped proteins are subsequently identified by LC-MS/MS and screened by cluster analysis and domain scanning. We apply this methodology to peptides with different proline-containing consensus sequences and show successful identifications from brain lysates of known and novel proteins containing polyproline motif-binding domains such as EH, EVH1, SH3, WW domains. These results suggest the capacity of arrayed peptide ligands to capture and subsequently identify proteins by mass spectrometry is relatively broad and robust. Additionally, the approach is rapid and applicable to cell or tissue fractions from any source, making the approach a flexible tool for initial protein-protein interaction discovery.     

Protein interactions with the platelet integrin αIIb regulatory motif

Integrins are transmembrane proteins regulating cellular shape, mobility and the cell cycle. A highly conserved signature motif in the cytoplasmic tail of the integrin α‐subunit, KXGFFKR, plays a critical role in regulating integrin function. To date, six proteins have been identified that target this motif of the platelet‐specific integrin α_IIbβ₃. We employ peptide‐affinity chromatography followed‐up with LC‐MS/MS analysis as well as protein chips to identify new potential regulators of integrin function in platelets and put them into their biological context using information from protein:protein interaction (PPI) databases. Totally, 44 platelet proteins bind with high affinity to an immobilized LAMWKVGFFKR‐peptide. Of these, seven have been reported in the PPI literature as interactors with integrin α‐subunits. 68 recombinant human proteins expressed on the protein chip specifically bind with high affinity to biotin‐tagged α‐integrin cytoplasmic peptides. Two of these proteins are also identified in the peptide‐affinity experiments, one is also found in the PPI databases and a further one is present in the data to all three approaches. Finally, novel short linear interaction motifs are common to a number of proteins identified.     

Stabilization of the GDP-bound conformation of Gialpha by a peptide derived from the G-protein regulatory motif of AGS3

The G-protein regulatory (GPR) motif in AGS3 was recently identified as a region for protein binding to heterotrimeric G-protein alpha subunits. To define the properties of this approximately 20-amino acid motif, we designed a GPR consensus peptide and determined its influence on the activation state of G-protein and receptor coupling to G-protein. The GPR peptide sequence (28 amino acids) encompassed the consensus sequence defined by the four GPR motifs conserved in the family of AGS3 proteins. The GPR consensus peptide effectively prevented the binding of AGS3 to Gialpha1,2 in protein interaction assays, inhibited guanosine 5'-O-(3-thiotriphosphate) binding to Gialpha, and stabilized the GDP-bound conformation of Gialpha. The GPR peptide had little effect on nucleotide binding to Goalpha and brain G-protein indicating selective regulation of Gialpha. Thus, the GPR peptide functions as a guanine nucleotide dissociation inhibitor for Gialpha. The GPR consensus peptide also blocked receptor coupling to Gialphabetagamma indicating that although the AGS3-GPR peptide stabilized the GDP-bound conformation of Gialpha, this conformation of Gialpha(GDP) was not recognized by a G-protein coupled receptor. The AGS3-GPR motif presents an opportunity for selective control of Gialpha- and Gbetagamma-regulated effector systems, and the GPR motif allows for alternative modes of signal input to G-protein signaling systems.     

Analysis of the key elements of FFAT-like motifs identifies new proteins that potentially bind VAP on the ER, including two AKAPs and FAPP2

Background

Two phenylalanines (FF) in an acidic tract (FFAT)-motifs were originally described as having seven elements: an acidic flanking region followed by 6 residues (EFFDA–E). Such motifs are found in several lipid transfer protein (LTP) families, and they interact with a protein on the cytosolic face of the ER called vesicle-associated membrane protein-associated protein (VAP). Mutation of which causes ER stress and motor neuron disease, making it important to determine which proteins bind VAP. Among other proteins that bind VAP, some contain FFAT-like motifs that are missing one or more of the seven elements. Defining how much variation is tolerated in FFAT-like motifs is a preliminary step prior to the identification of the full range of VAP interactors.

Results

We used a quantifiable in vivo system that measured ER targeting in a reporter yeast strain that over-expressed VAP to study the effect of substituting different elements of FFAT-like motifs in turn. By defining FFAT-like motifs more widely than before, we found them in novel proteins the functions of which had not previously been directly linked to the ER, including: two PKA anchoring proteins, AKAP220 and AKAP110; a family of plant LTPs; and the glycolipid LTP phosphatidylinositol-four-phosphate adaptor-protein-2 (FAPP-2).

Conclusion

All of the seven essential elements of a FFAT motif tolerate variation, and weak targeting to the ER via VAP is still detected if two elements are substituted. In addition to the strong FFAT motifs already known, there are additional proteins with weaker FFAT-like motifs, which might be functionally important VAP interactors.     

Cyclophilin A binds to linear peptide motifs containing a consensus that is present in many human proteins

Cyclophilin A (CypA) is a peptidyl-prolyl cis/trans-isomerase that is involved in multiple signaling events of eukaryotic cells. It might either act as a catalyst for prolyl bond isomerization, or it can form stoichiometric complexes with target proteins. We have investigated the linear sequence recognition code for CypA by phage display and found the consensus motif FGPXLp to be selected after five rounds of panning. The peptide FGPDLPAGD showed inhibition of the isomerase reaction and NMR chemical shift mapping experiments highlight the CypA interaction epitope. Ligand docking suggests that the peptide was able to bind to CypA in the cis- and trans-conformation. Protein Data Bank searches reveal that many human proteins contain the consensus motif, and several of these protein motifs are shown to interact with CypA in vitro. These sequences represent putative target sites for binding of CypA to intracellular proteins.     

C-terminal recognition by 14-3-3 proteins for surface expression of membrane receptors

Diverse functions of 14-3-3 proteins are directly coupled to their ability to interact with targeted peptide substrates. RSX(pS/pT)XP and RXPhiX(pS/pT)XP are two canonical consensus binding motifs for 14-3-3 proteins representing the two common binding modes, modes I and II, between 14-3-3 and internal peptides. Using a genetic selection, we have screened a random peptide library and identified a group of C-terminal motifs, termed SWTY, capable of overriding an endoplasmic reticulum localization signal and redirecting membrane proteins to cell surface. Here we report that the C-terminal SWTY motif, although different from mode I and II consensus, binds tightly to 14-3-3 proteins with a dissociation constant (K(D)) of 0.17 microM, comparable with that of internal canonical binding peptides. We show that all residues but proline in -SWTX-COOH are compatible for the interaction and surface expression. Because SWTY-like sequences have been found in native proteins, these results support a broad significance of 14-3-3 interaction with protein C termini. The C-terminal binding consensus, mode III, represents an expansion of the repertoire of 14-3-3-targeted sequences.     

The Effects of Phosphomimetic Lid Mutation on the Thermostability of the N-terminal Domain of MDM2

The multidomain E3 ubiquitin ligase MDM2 catalyzes p53 ubiquitination by a “dual-site” docking mechanism whereby MDM2 binding to at least two distinct peptide motifs on p53 promotes ubiquitination. One protein-protein interaction occurs between the N-terminal hydrophobic pocket of MDM2 and the transactivation motif of p53, and the second interaction occurs between the acidic domain of MDM2 and a motif in the DNA-binding domain of p53. A flexible N-terminal pseudo-substrate or “lid” adjacent to the N-terminal hydrophobic pocket of MDM2 has a phosphorylation site, and there are distinct models proposed on how the phosphorylated lid could affect MDM2 function. Biochemical studies have predicted that phosphomimetic mutation will stabilize the lid on the surface of MDM2 and will “open” the hydrophobic pocket and stabilize the MDM2-p53 complex, while NMR studies proposed that phosphomimetic mutation “closes” the lid over the MDM2 pocket and inhibits MDM2-p53 complex formation. To resolve these discrepancies, we utilized a quantitative fluorescence-based dye binding assay to measure the thermal unfolding of wild-type (wt), ΔLid, and S17D N-terminal domains of MDM2 as a function of increasing ligand concentration. Our data reveal that S17D lid mutation increases, rather than decreases, the thermostability of the N-terminal domain of MDM2 in the absence or in the presence of ligand. ΔLid mutation, by contrast, increases MDM2 thermoinstability. This is consistent with biochemical data, using full-length MDM2, showing that the S17D mutation stabilizes the MDM2-p53 complex and increases the specific activity of the E3 ubiquitin ligase function of MDM2. These data indicate that phosphomimetic lid mutation results in an “opening,” rather than a “closing,” of the pocket of MDM2 and highlight the ability of small intrinsically disordered or unstructured peptide motifs to regulate the specific activity of a protein.