Systematic discovery of linear binding motifs targeting an ancient protein interaction surface on MAP kinases
2015; Springer Nature; Volume: 11; Issue: 11 Linguagem: Inglês
10.15252/msb.20156269
ISSN1744-4292
AutoresAndrás Zeke, Tomas Bastys, Anita Alexa, Ágnes Garai, Bálint Mészáros, Klára Kirsch, Zsuzsanna Dosztányi, Olga V. Kalinina, Attila Reményi,
Tópico(s)Protein Structure and Dynamics
ResumoArticle4 November 2015Open Access Source Data Systematic discovery of linear binding motifs targeting an ancient protein interaction surface on MAP kinases András Zeke András Zeke Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Tomas Bastys Tomas Bastys Max Planck Institute for Informatics, Saarbrücken, Germany Graduate School of Computer Science, Saarland University, Saarbrücken, Germany Search for more papers by this author Anita Alexa Anita Alexa Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Ágnes Garai Ágnes Garai Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Bálint Mészáros Bálint Mészáros Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Klára Kirsch Klára Kirsch Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Zsuzsanna Dosztányi Zsuzsanna Dosztányi MTA-ELTE Lendület Bioinformatics Research Group, Department of Biochemistry, Eötvös Loránd University, Budapest, Hungary Search for more papers by this author Olga V Kalinina Olga V Kalinina Max Planck Institute for Informatics, Saarbrücken, Germany Search for more papers by this author Attila Reményi Corresponding Author Attila Reményi Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author András Zeke András Zeke Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Tomas Bastys Tomas Bastys Max Planck Institute for Informatics, Saarbrücken, Germany Graduate School of Computer Science, Saarland University, Saarbrücken, Germany Search for more papers by this author Anita Alexa Anita Alexa Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Ágnes Garai Ágnes Garai Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Bálint Mészáros Bálint Mészáros Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Klára Kirsch Klára Kirsch Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Zsuzsanna Dosztányi Zsuzsanna Dosztányi MTA-ELTE Lendület Bioinformatics Research Group, Department of Biochemistry, Eötvös Loránd University, Budapest, Hungary Search for more papers by this author Olga V Kalinina Olga V Kalinina Max Planck Institute for Informatics, Saarbrücken, Germany Search for more papers by this author Attila Reményi Corresponding Author Attila Reményi Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary Search for more papers by this author Author Information András Zeke1, Tomas Bastys2,3, Anita Alexa1, Ágnes Garai1, Bálint Mészáros4, Klára Kirsch1, Zsuzsanna Dosztányi5, Olga V Kalinina2 and Attila Reményi 1 1Lendület Protein Interaction Group, Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary 2Max Planck Institute for Informatics, Saarbrücken, Germany 3Graduate School of Computer Science, Saarland University, Saarbrücken, Germany 4Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary 5MTA-ELTE Lendület Bioinformatics Research Group, Department of Biochemistry, Eötvös Loránd University, Budapest, Hungary *Corresponding author. Tel: +36 1 3826613; E-mail: [email protected] Molecular Systems Biology (2015)11:837https://doi.org/10.15252/msb.20156269 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Abstract Mitogen-activated protein kinases (MAPK) are broadly used regulators of cellular signaling. However, how these enzymes can be involved in such a broad spectrum of physiological functions is not understood. Systematic discovery of MAPK networks both experimentally and in silico has been hindered because MAPKs bind to other proteins with low affinity and mostly in less-characterized disordered regions. We used a structurally consistent model on kinase-docking motif interactions to facilitate the discovery of short functional sites in the structurally flexible and functionally under-explored part of the human proteome and applied experimental tools specifically tailored to detect low-affinity protein–protein interactions for their validation in vitro and in cell-based assays. The combined computational and experimental approach enabled the identification of many novel MAPK-docking motifs that were elusive for other large-scale protein–protein interaction screens. The analysis produced an extensive list of independently evolved linear binding motifs from a functionally diverse set of proteins. These all target, with characteristic binding specificity, an ancient protein interaction surface on evolutionarily related but physiologically clearly distinct three MAPKs (JNK, ERK, and p38). This inventory of human protein kinase binding sites was compared with that of other organisms to examine how kinase-mediated partnerships evolved over time. The analysis suggests that most human MAPK-binding motifs are surprisingly new evolutionarily inventions and newly found links highlight (previously hidden) roles of MAPKs. We propose that short MAPK-binding stretches are created in disordered protein segments through a variety of ways and they represent a major resource for ancient signaling enzymes to acquire new regulatory roles. Synopsis The disordered part of the human proteome contains a large number of short linear motif occurrences that can bind to MAP kinases. These simple protein-protein recruitment sites represent a major resource for ancient signaling enzymes to acquire new regulatory roles. A combined computational and experimental approach identifies hundreds of putative MAP kinase (MAPK) binding linear motifs—referred to as D(ocking)-motifs—in the human proteome. The wide distribution of D-motifs in functionally diverse set of proteins explains how MAPKs can regulate a broad spectrum of physiological processes. MAPK based interactomes changed fast over time as D-motif composition of vertebrate proteomes is remarkably different. D-motifs mostly emerge by random mutations in disordered protein regions. Introduction Protein–protein interactions influence all aspects of cellular life and the most direct mechanism through which proteins can influence each other is by physical interaction. This brings them into proximity to exert control on activity or to create opportunities for posttranslational modification. Protein–protein associations often involve so-called linear binding motifs which are short (5–20 amino acid long) protein regions lacking autonomous tertiary structure. These functional sites reside in intrinsically disordered protein regions and adopt stable conformation only upon binding. Currently, we can only guess how abundant linear motif-based interactions are; nevertheless, it was recently estimated that there are ~100,000 linear binding motifs targeting dedicated protein surfaces in the human proteome (Tompa et al, 2014). As an example relevant to cellular signaling, mitogen-activated protein kinases (MAPKs) are prototypical enzymes that depend on short segments from partner proteins and on their dedicated protein–protein interaction hot spots. They mainly recognize their substrates not with the catalytic site but with auxiliary docking surfaces on their kinase domains (Tanoue et al, 2000; Biondi & Nebreda, 2003). The most important of these docking sites consists of a hydrophobic docking groove and the negatively charged CD (common docking) region (Chang et al, 2002) (Fig 1A). Together, they can bind the so-called D(ocking)-motifs of the target proteins. D-motifs are short linear motifs ranging from 7 to 18 amino acids in length and are typically found in intrinsically disordered segments—potentially far away from target phosphorylation sites (Garai et al, 2012). Such docking elements are not only restricted to substrates: They are also found in MAPK-activating kinases (MAP2Ks), in MAPK-inactivating phosphatases (MKPs), and in a variety of scaffold proteins. While extracellularly regulated kinases (ERK1-2), c-Jun N-terminal kinases (JNK1-3), and the 38-kDa protein kinases (p38α-δ) control diverse physiological processes, they phosphorylate most of their substrates at Ser-Pro or Thr-Pro (S/TP) sequence motifs. Naturally, this weak consensus provided by their catalytic site is insufficient for selective target recognition, and additional linear binding motifs provide specificity (Johnson & Lapadat, 2002; Bardwell, 2006). Therefore, the MAPK D-motif protein–protein interaction system is an ideal test bed for linear binding motif discovery. Figure 1. Structural classification of MAPK-docking motifs The MAPK-docking groove comprises two distinct regions: the common docking (CD) and the hydrophobic region. These are colored in blue and light brown, respectively, and are shown on the JNK1 surface from the JNK1-NFAT4 protein–peptide complex crystal structure (Garai et al, 2012). (The CD groove is extended by the ED region, extra negatively charged residues for ERK and p38; see (C); Tanoue et al, 2001.) Different binding modes of D-motifs. The hydrophobic docking groove binds three hydrophobic amino acids in a row, while admitting two different spacing schemes. At the same time, θ to φ linker length determines the MAPK specificity of a given motif. These two features can combine freely with each other, resulting in the four basic arrangements shown here. Distinct binding conformations at the CD groove. N-termini of longer D-motifs are variable and ERK2- or p38α-binding peptides may take a variety of conformations—ranging from fully linear (e.g., MEF2A, green) to highly alpha-helical (e.g., HePTP, magenta). Structural heterogeneity of D-motifs. The combinations of the three variable features yield structurally well-defined, distinct classes of D-motifs. Many of these models also define separate types of linear motifs, but their consensus sequences are not always exclusive. JNK kinases only admit two major types of motifs, the NFAT4 class (1-spacing, short linker) and the JIP1 class (2-spacing, short linker). On the other hand, known ERK1/2 and p38 binder peptides may belong to the greater MEF2A class (1-spacing, longer linker, linear end), the greater HePTP class (1-spacing, longer linker, helical end), or the greater DCC class (2-spacing, longer linker, linear end). A sixth class of ERK or p38 interactors is theoretically also possible (2-spacing, longer linker, helical end), but this combination can only be observed in long reverse (revD) motifs (Garai et al, 2012), and no classical motif of this type is known up to date. Subtypes and other variants within a given greater class are also featured wherever applicable. These are shown based on structures of MAPK-docking motif complexes. Dashed lines indicate N-terminal peptide regions that are usually not visible in the crystal structures, albeit indispensable for binding. Consensus motif of each subtype is shown below, where φU, φL, φA, and φB letters denote positions that are filled by hydrophobic amino acids—L, A, and B refer to the lower pocket, and pockets A and B, respectively—while the θ positions are positively charged (Arg or Lys) while letter "x" denotes arbitrary amino acids. Download figure Download PowerPoint Several previous attempts were aimed at predicting MAPK-binding proteins from full proteomes by using a generic consensus of D-motifs as it had been established more than a decade ago (Sharrocks et al, 2000). This consensus was derived from an observation that D-motifs almost always include at least a single positively charged residue (termed the θ position: arginine or lysine) and a series of alternating hydrophobic residues (φ positions: frequently leucine), connected by a linker of a variable length (Dinkel et al, 2014). But despite the use of extensive multiple alignments and sophisticated algorithms, predictions had only low success rates and large-scale assessment of predicted hits was not performed (Whisenant et al, 2010; Garai et al, 2012; Gordon et al, 2013). Regarding experimental MAPK network discovery, ERK2 has been the most widely explored. For example, several different methods were utilized to identify ERK2 substrates by large-scale phosphoproteomics (Kosako et al, 2009; Carlson et al, 2011; Courcelles et al, 2013). Unfortunately, pairwise overlaps between the lists of substrates are low across studies (e.g., around ~10%), with not a single overlap between five different studies that aimed to find ERK2-phosphorylated substrates (Courcelles et al, 2013), suggesting great dependence on the experimental conditions used. It was noted that D-motif-like sequences are enriched in experimentally detected ERK2 substrates (Carlson et al, 2011), yet detection or verification of direct physical association was not performed. In addition, studies that used a high-throughput approach to identify partners of JNK1 (Chen et al, 2014) or p38α (Bandyopadhyay et al, 2010) based on direct physical interaction resulted in low number of hits. Thus, it is likely that a protein–protein interaction-based MAPK network discovery could greatly benefit from a target tailored approach, which takes into account—and possibly capitalizes on—specific biochemical and biophysical knowledge already available on known MAPK–partner protein interactions. In recent years, the number of experimentally determined MAPK–partner protein complex structures increased considerably (Garai et al, 2012). This development made it possible to amend the definition of the underlying sequence motifs and it became clear that D-motifs encompass multiple classes of similarly built, but structurally distinct linear motifs (similar to SH3- or PDZ-domain-binding sequences, for example) (Lim et al, 1994; Tonikian et al, 2008). In the current study, we show that by building a strategy that can handle this conformational diversity in binding, and using structural compatibility with specific interaction surface topography as an additional criterion for prediction, the identification of novel D-motifs can be dramatically improved. This analysis in combination with tailored experimental techniques for the validation of low-affinity (1–20 μM) protein–protein interactions produced unique, molecular-level insight into physiological roles and evolution of MAPK-based protein networks. Results Structure-guided prediction of MAPK-binding linear motifs MAPK–D-peptide complex structures revealed distinct D-motif binding modes in the MAPK-docking groove (Fig 1). For example, D-motifs from the JNK-binding scaffold protein JIP1 and from the JNK-regulated transcription factor NFAT4 bind to the same docking surface differently (Fig 1A and B) (Heo et al, 2004; Garai et al, 2012; Laughlin et al, 2012). Similarly, ERK- and p38-binding D-motifs may also be structurally distinct; nonetheless, D-motifs could be described with a common loosely defined consensus [θ1,2-x(0-5)-φL-x(1,2)-φA-x-φB; φL, φA, and φB denote positions that are typically filled by hydrophobic amino acids—L, A, and B refer to the lower pocket, and pockets A and B, respectively—while the θ denotes positively charged (arginine or lysine) and "x" denotes any amino acid]. However, the rules are much stricter for sequences that are compatible with a given MAPK-docking surface in a given binding mode. Interestingly, D-motifs and their binding modes may be conserved from yeast to human as the docking surface is ancient and well conserved across all eukaryotes (Reményi et al, 2005; Grewal et al, 2006). Because the CD region of ERK and p38 is wider compared to that of JNK, the N-termini of motifs binding to the two former kinases have larger conformational freedom (Fig 1C) (Garai et al, 2012). These can be classified as MEF2A- and DCC-type motifs named after the proteins in which they were first identified. Some motifs with longer intervening regions also exists (HePTP) (Zhou et al, 2006). Interestingly, the typical helical conformation at the N-terminus of HePTP-type docking motif is also characteristic to some MAPK interactors from yeast (Ste7) and peptides with such motifs are known to bind human ERK2 with high affinity (Fernandes et al, 2007). Therefore, we also set up a hypothetical subclass of Ste7-type motifs, hitherto unknown in humans (Fig 1D). Simple pattern matching of motifs normally produces a large number of false positives, because motif-matching sequences may occur simply by chance. In order to drastically reduce the number of false hits, an in silico filtering procedure was implemented to search for putative linear motifs (Fig 2). The first step was to screen for motifs in regions with intrinsic disorder but with propensity for disorder-to-order transition (ANCHOR) (Dosztányi et al, 2009; Mészáros et al, 2009). This procedure was used in order to eliminate those consensus motif occurrences that would either be buried in folded domains or permanently locked in an unfavorable conformation. Importantly, it also removed initial hits with an inappropriate amino acid composition, not being able to adopt a stable structure upon binding to a protein surface. Motifs were then filtered for MAPK accessibility: Motifs that were predicted to lie in extracellular protein segments or in other, kinase-inaccessible compartments (e.g., ER, Golgi) were discarded. In addition, an auxiliary check was performed against structured Pfam domains. This was applied to remove all spurious motifs in ordered regions which had been retained after ANCHOR filtering. Since almost all the known motifs passed these filters (45 out of 47, with enrichment ratios between 4.1 to 6.6 depending on motif type), we concluded that the dataset was of sufficient quality for further testing. Figure 2. Motif finding work flowTo find novel MAPK-docking motifs, the primary motif-matching step (on UniProt KB sequences) was followed by a series of filters. Valid motifs had to pass through an ANCHOR filter, a localization filter (combined from SignalP and Phobius) and an auxiliary Pfam filter, in order to be scored by FoldX homology models. ANCHOR (middle panel) had the most important role in selecting segments that can potentially act as linear motifs, while FoldX gave motif-specific binding energy estimates (see Source data). Predicted motifs were subsequently tested as short fragments and (in selected cases) as full-length proteins. Finally, an automated evolutionary analysis was performed to give information on motif conservation trends. Source data is available online for this figure. Source Data for Figure 2 [msb156269-sup-0007-SDataFig2.xls] Download figure Download PowerPoint The classification of D-motifs based on a coherent structural model enabled us to make use of structure-based scoring. As a motif occurrence could always be unambiguously matched with its corresponding MAPK-docking peptide structural model, we used FoldX, which had been validated on protein–peptide complexes, to estimate the change of the protein–peptide binding energy with respect to the energy of an experimentally resolved complex (Appendix Table S1) (Schymkowitz et al, 2005). This allowed the scoring of motifs according to their structural compatibility to the MAPK-docking groove. FoldX-derived binding energy estimates were also used as a guide when motifs were being chosen for later experimental screening. Experimental screening After completion of initial lists, we chose a number of candidate proteins from each motif type to test. Expression of full-length human proteins of large size (> 1,000 amino acids) in recombinant systems can be a limiting step in experimental validation; therefore, first we opted for a fragment-based approach. Former experiments showed that simple binding assays (such as pull-downs with recombinant D-motif-containing proteins or immobilized solid-phase peptide arrays) lack the robustness to reliably detect low-affinity (1–20 μM) protein–peptide interactions. Therefore, we developed a different assay which was based on substrate phosphorylation enhancement on a solid-phase support (Fig 3A). As the majority of known D-motif-containing proteins are MAPK substrates, this adequately captures the original function of these motifs. An artificial substrate was constructed containing the D-motifs as well as the Thr71 phosphorylation site from ATF2, which is a well-known MAPK target site (Livingstone et al, 1995) (Fig 3B). As linkers and substrate sites in the recombinant proteins were identical, the "docking efficiency" of the given motifs could be directly compared to each other. Phosphorylation of this reporter solely depended on the presence or absence of specific docking motifs, and phosphorylation of the target site was low without a functional D-motif (Fig 3C and D). Figure 3. Dot-blot arrays of D-motifs The principle of the phosphorylation enhancement dot-blot arrays. Protein constructs are immobilized onto a solid-phase support where phosphorylation takes place. Afterward, the phosphorylated epitopes are detected through standard Western blot procedures using a phosphorylation-sensitive antibody. The schematic structure of the artificial substrates utilized in the dot-blot arrays. All constructs share the same tags, substrate sites, and linkers: Only the docking motif-containing fragments differ. A sample dot-blot array for detecting JNK-binding docking motifs. This specific array contains 48 of the 70 motifs tested in total, and was incubated against activated JNK1. Quantitative analysis of the sample dot-blot array. All intensities are relative to that of the NFAT4 motif (positive control), error bars were derived from three parallel samples on the same membrane and show standard deviation from the mean (N = 3). "+" denotes additional, non-overlapping motifs tested from the same protein. "m" refers to murine (non-human) sequence. (The corresponding ERK2 and p38α 48 motif arrays and the 70 motif arrays for all three MAPKs are shown on Appendix Fig S1 or on Fig EV1). Download figure Download PowerPoint In the final panels, we included 70 different constructs: 63 of these were directly selected from the lists ranked by the predicted binding energy (Fig EV1 and Appendix Fig S1). We also included additional seven motifs based on sequence similarity to known motifs. This was done in order to test whether some other similar motifs not conforming to the formerly defined sequence patterns have the capacity to bind MAPKs. Out of 70, a total of 52 motifs were found to interact with at least one of the three MAPKs (ERK2, JNK1, or p38α). In particular, we were able to detect several novel interactors based on the JIP1, NFAT4, MEF2A, MKK6, and DCC models. As for our hypothetic Ste7 model, we also found a novel hit: a motif from RHDF1 that is also found in the related RHDF2 protein. Such a high number of hits suggest that D-motifs are in fact quite widespread in the human proteome (Table 1, Fig EV2, and Appendix Fig S2). Click here to expand this figure. Figure EV1. Results of dot-blot experimentsPanels show dot blots of all 71 constructs (70 D-motif-containing substrates + 1 negative control) against three MAPKs and the epitope density control (see alsoAppendix Fig S1). Download figure Download PowerPoint Table 1. Validated sequences grouped by D-Motif class Greater MEF2A class (phosphorylation by p38α) Greater DCC class (phosphorylation by ERK2) Greater HePTP class (phosphorylation by ERK2) MEF2A-type MKK6-type Misc. DCC-type Far1-type Ste7-like HePTP-like AAKG2/PRKAG2 (28–37) CCNT2 (498–509) AMPD1 (109–120) DCC (1,144–1,155) CBLB (489–500) RHDF1/iRhom1 (11–24) ZEP1/HIVEP1 (1,422–1,438) JAZF1 (77–86) GAB3 (363–374) AMPD3 (79–90) CACNA1G (1,030–1,041) ELMSAN1 (601–612) RHDF2/iRhom2 (18–31) INO80a (1,318–1,327) INO80a (1,316–1,327) TRERF1 (653–664) MEF2A (268–277) KSR2 (330–341) GAB1 (526–536) KLF3 (88–97) KMT2C/MLL3a (1,195–1,206) KMT2C/MLL3a (1,197–1,206) RIPK2a (326–335) TSHZ3a (321–330) JIP1 class (phosphorylation by JNK1) APBA2/MINT2 (279–285) ATF2 (164–170) ATF7 (162–168) APC2 (962–968) BMPR2 (753–759) DOCK5 (1,762–1,768) DOCK7 (884–890) DUSP10/MKP5 (18–24) ELK1 (314–320) IRS1 (856–862) JIP3 (203–209) M3K10/MLK2 (876–882) MADD (809–815) MCL1 (136–142) MYO9B (1,249–1,255) PDE4B (72–78) PRGC1/PPARGC1A (253–259) SAC2/INSPP5 (1,009–1,015) NFAT4 class (phosphorylation by JNK1) AKAP6/mAKAP (433–440) CCSER1 (573–580) DYH12/DNAH12 (12–19) FMN1 (672–679) FHOD3 (506–513) JUND (46–53) KANK2 (244–251) M3K10/MEKK1 (1,077–1,084) MABP1 (1,292–1,299) NFATC3/NFAT4 (145–152) RIPK2a (327–334) TSHZ3a (322–329) The table lists motifs that tested as positive ("hits") in the dot-blot arrays and it shows the most commonly used names of proteins (when necessary, two variants), together with the position of the core motif—based on the reference isoform featured in UniProt. Names in bold type denote previously known docking motifs, while the names in normal type indicate novel interactors. a Denotes motifs that appear under more than one class as they satisfy multiple consensus sequences. Click here to expand this figure. Figure EV2. Constructs used in dot-blot experiments and grouped according to motif classAmino acid sequences of D-motif-containing short segments ligated into the pAZAD vector are displayed in a tabulated format. Tested sequences are grouped according to D-motif types. In addition, the conformity of the motif to a loose or tight motif definition (see Appendix Fig S2) is also indicated: Motifs whose sequence fits to its tight D-motif class definition is written in bold type, while those that only correspond to a looser definition is shown in normal type. The outcome of dot-blot experiments is also indicated: A motif is regarded as a "strong binder" (unequivocally positive) if it performed above the cutoff (non-cognate D-motif control) in all experiments. In case if it performed below this cutoff in at least one round of experiments, but still consistently above the level of zero control (substrate with no D-motif), it is considered a "weak binder". Otherwise, the motif is classified as a "non-binder". Only "strong binders" were considered as "positives" for the purpose of sequence and evolutionary analysis. +: These proteins had more than one, non-overlapping MAPK-binding elements tested (names with/without "+" refer to different sequences). *: The sequence of these constructs satisfies at least two different, overlapping consensus motifs (thus, they are featured under more than one motif class/subclass). For the sake of simplicity, not all combinations are shown. ∼: This novel isoform is found in NCBI and other databases but not in UniProt. m: This sequence refers to the murine instead of the human protein (Kim et al, 2012). Download figure Download PowerPoint To show that the phosphorylation enhancements were indeed due to the presence of canonical MAPK D-motifs binding into the MAPK-docking groove, a set of 16 chemically synthesized peptides were titrated against fluorescently labeled control peptides known to bind at the MAPK-docking groove (Fig EV3 and Appendix Fig S3). In addition to confirming binding in the MAPK-docking groove, the dissociation constant (Kd) of unlabeled test peptides could also be calculated. Click here to expand this figure. Figure EV3. Protein–peptide binding affinity assaysThe results of fluorescence polarization (FP) titrations are displayed in a tabulated format, showing the sequence of synthetic peptide used together with the dissociation constants (Kd) obtained for each MAPK (see also Appendix Fig S3). Dashes indicate cases where the curves could not be fitted (the Kd is above the limit of quantitation of this assay, approximately ~100 μM). Peptides are grouped in two clusters, the first being the strong JNK1 binders (frequently with little or no ERK2/p38α-binding ability). The second group (below the magenta line) contains peptides with strong ERK2- and p38α-binding capacity (but often little or no affinity toward JNK1). Coloring of the peptide residues displays the critical charged θ residues in blue and the φ hydrophobic
Referência(s)