Binding Specificity of SH2 Domains Revealed by a Combinatorial Peptide Library

Thesis

Presented in Partial Fulfillment of the Requirements for the Master’s Degree in the Graduate School of The Ohio State University

By

Andrew Kunys, B.S. Graduate Program in Chemistry

The Ohio State University 2013

Thesis Committee: Dehua Pei, Advisor Ross Dalbey

ii

Copyright by

Andrew Kunys

2013

iii

Abstract

Src homology 2 Domains (SH2 domains) are modular protein binding domains which recognize a phosphotyrosine and the residues adjacent to the phosphotyrosine. Although they are known to bind to specific phosphopeptides motifs, the exact peptide binding motifs for each of the 120 human SH2 domains are not known. These proteins are associated mainly with cellular signaling affecting processes as disparate as apoptosis and proliferation.

In this work, high quality specificity data was obtained from 18 different SH2 domains by screening them individually against a One-Bead-One-Peptide library. Interestingly, no two domains have the same specificity, suggesting that subtle differences mark each domain as unique. Further, most of the domains studied recognize multiple different classes of binding peptides, which implies their role in binding may be more complicated that previously realized. The data obtained is unique, in that no other method can produce data of high enough quality to unambiguously establish the subtle specificity patterns of each domain or decipher the data obtained from a domain which can recognize multiple peptide ligand classes. This data will provide insight into the basic biology of cell signaling and should prove to be a valuable resource to researchers studying in this area.

ii

Acknowledgements

I would like to thank, first and foremost, my mother and father for their love and support. Second, I would like to thank my colleagues and friends in Dr. Pei’s research group.

I thank Yanyan Zhang and Pauline Tan for their encouragement, friendship and mentorship.

I thank Nicholas Selner and Tao Liu for their encouragement and the many helpful discussions we have shared. I thank Tiffany Waller, Xianwen Chen, and Rinrada

Luechapanichkul for their friendship and perspective. Finally, I thank Dr. Dehua Pei for his mentorship and support.

iii

Vita

2008 ………………………...... B.S. Chemistry, University of Michigan

2008 to present ...... Graduate Research Associate, Department of

Chemistry, The Ohio State University

2008-2009...... Graduate Teaching Assistant, Department of

Chemistry, The Ohio State University

2012-2013...... Graduate Teaching Assistant, Center for Life

Science Education, The Ohio State University

Publications

Kunys, A.R., Lian, W., and Pei, D. 2012. Specificity profiling of protein-binding domains using one-bead-one-compound Peptide libraries. Curr Protoc Chem Biol 4:331-355.

Zhang, Y., Wavreille, A.S., Kunys, A.R., and Pei, D. 2009. The SH2 domains of inositol polyphosphate 5-phosphatases SHIP1 and SHIP2 have similar ligand specificity but different binding kinetics. Biochemistry 48:11075-11083.

Fields of Study

Major Field: Biological Chemistry

iv

Table of Contents

Abstract ………………………………………………………………………………………….….……………………………………...ii

Acknowledgements ……………………………………………………………………….………………….………………………iii

Vita ……………………………………………………………………….……………………………………….….………………………iv

List of Tables …………………………………………………………………………………………………………………………….vii

List of Figures ……………………………………………………………………………………………………………………………..x

Introduction

SH2 domains discovery and function ………………………………………………………....…………………1

SH2 Structure and specificity determining elements ……………………………………………………3

Traditional methods of determining binding specificities .…………………………………….……….5

Determination of SH2 Specificity via One Bead One Compound libraries …..…………….……8

Experimental methods

Materials ……………………………………………………….………………...... 13

Expression of the SH2 Domains ……………………….…………………………………….……………...... 14

Purification and labeling of the SH2 Domains ………………………….……….………….…….………14

Covalent Labeling of SH2 Domains …………………………………………………….……..…………….....15

Library Synthesis …………………………………………………………………………………..……...... 15

Screening SH2 Domains against the pY library ……………………………………..………………….....16

Partial Edman Degradation …………………………………………………………………..……………………..17

v

MALDI-TOF analysis …………………………………………………….…………………………………...... 18

Results

Library Used ….…………………………………………………………………………………………………………….19

Kinase SH2 Domains (Itk, Brk, SRMS, Frk, Btk) ………………………………….……………..……….….19

Adapter SH2 Domains (Gads, GRAP) ……………………………………………………………………………37

Shc Family SH2 Domains (Shc1, Shc2, Shc3, Shc4) ………………………………………….……….….44

Vav3 and SHE ………………………………………………………………………………………………………………59

(PLCG1-N, DAPP1, Blnk, SLA2, SH2D3A ……………………………….……………………………………..71

Discussion ……………………………………………………………………………………………………………………….……….94

Bibliography ………………………………………………………………………………………………………………….…………96

vi

List of Tables

Table 1. SH2 domains Examined in this study ………………………………………………………………………….23

Table 2. Peptide sequences selected from the Itk-SH2 Domain ………………………………………….…..24

Table 3. Peptide sequences selected from the Brk-SH2 Domain ………………………………….…….……25

Table 4. Peptide sequences selected from the SRMS-SH2 Domain ……………………………….…………26

Table 5. Peptide sequences selected from the Frk-SH2 Domain ………………………………….….……….27

Table6. Peptide sequences selected from the Btk-SH2 Domain ………………………………….….….……28

Table 7. Predicted Binding Partners of the Itk-SH2 Domain (Class 1) ………………….…………..….……31

Table 8. Predicted Binding Partners of the Itk-SH2 Domain (Class 2) ………………………………….……32

Table 9. Predicted Binding Partners of the Brk-SH2 Domain ……………………………………..……….……33

Table 10. Predicted Binding Partners of the SRMS-SH2 Domain ……………………………..………….……34

Table 11. Predicted Binding Partners of the Frk-SH2 Domain …………………………………………….……35

Table 12. Predicted Binding Partners of the Btk-SH2 Domain ………………………………………...….……36

Table13. Peptide sequences selected from the Gads-SH2 Domain …………………………….……………39

Table14. Peptide sequences selected from the GRAP-SH2 Domain …………………………………………40

Table 15. Predicted Binding Partners of the Gads-SH2 Domain ………………………………………..….….42

Table 16. Predicted Binding Partners of the GRAP-SH2 Domain ……………………………………..….……43

Table 17. Peptide sequences selected from the Shc1-SH2 Domain ………………………………………….47

Table 18. Peptide sequences selected from the Shc2-SH2 Domain …………………………..……………..48

Table 19. Peptide sequences selected from the Shc3-SH2 Domain ………………………………………….49

Table 20. Peptide sequences selected from the Shc4-SH2 Domain ………………………………….………50

vii

Table 21. Predicted Binding Partners of the Shc1-SH2 Domain (Class 1) ………………………………….53

Table 22. Predicted Binding Partners of the Shc1-SH2 Domain (Class 2) ………………………………….54

Table 23. Predicted Binding Partners of the Shc2-SH2 Domain …………….………………………………….55

Table 24. Predicted Binding Partners of the Shc3-SH2 Domain (Class 1) ………………………………….56

Table 25. Predicted Binding Partners of the Shc3-SH2 Domain (Class 2) ………………………………….57

Table 26. Predicted Binding Partners of the Shc4-SH2 Domain ………………………….…………………….58

Table 27. Peptide sequences selected from the Vav3-SH2 Domain ………………….………………………61

Table 28. Peptide sequences selected from the SHE-SH2 Domain …………………………………..….…..62

Table 29. Predicted Binding Partners of the Vav3-SH2 Domain (Class 1) ………………………………….65

Table 30. Predicted Binding Partners of the Vav3-SH2 Domain (Class 2) ………………………………….66

Table 31. Predicted Binding Partners of the Vav3-SH2 Domain (Class 3) ………………………………….67

Table 32. Predicted Binding Partners of the Vav3-SH2 Domain (Class 4) ………………………………….68

Table 33. Predicted Binding Partners of the SHE-SH2 Domain (Class 1) ………….……………………….69

Table 34. Predicted Binding Partners of the SHE-SH2 Domain (Class 2) …………….…………………….70

Table 35. Peptide sequences selected from the PLCG1-N-SH2 Domain ………………….………………..75

Table 36. Peptide sequences selected from the DAPP1-SH2 Domain ………………….…………………..76

Table 37. Peptide sequences selected from the Blnk-SH2 Domain ………………….……………………….77

Table 38. Peptide sequences selected from the SLA2-SH2 Domain ………………….………………………78

Table 39. Peptide sequences selected from the SH2D3A-SH2 Domain ………………….…………………79

Table 40. Predicted Binding Partners of the PLCG1-N-SH2 Domain (Class 1) ……………………..…….83

Table 41. Predicted Binding Partners of the PLCG1-N-SH2 Domain (Class 2) …………..……………….84

Table 42. Predicted Binding Partners of the DAPP1-SH2 Domain (Class 1) ……………………………….85

Table 43. Predicted Binding Partners of the DAPP1-SH2 Domain (Class 2) ……………………………….86

Table 44. Predicted Binding Partners of the Blnk-SH2 Domain (Class 1) ………………………….……….87

viii

Table 45. Predicted Binding Partners of the Blnk-SH2 Domain (Class 2) ………………….……………….88

Table 46. Predicted Binding Partners of the Blnk-SH2 Domain (Class 3) ………………………….……….89

Table 47. Predicted Binding Partners of the SLA2-SH2 Domain (Class 1) ………………………………….90

Table 48. Predicted Binding Partners of the SLA2-SH2 Domain (Class 2) ………………………………….91

Table 49. Predicted Binding Partners of the SH2D3A-SH2 Domain (Class 1) …………………………….92

Table 50. Predicted Binding Partners of the SH2D3A-SH2 Domain (Class 2) …………………………….93

ix

List of Figures

Figure 1. Src-SH2 domain bound to pYEEI ……………………………………………….………………..………………4

Figure 2. Library used for screening ……………………………………………………….……………………………….9

Figure 3. Screening Scheme ………………………………………………………………….…………………….…………..10

Figure 4. Partial Edman Degradation Scheme …………………………………….……………………….…………..11

Figure 5. Sequence specificities of the Itk, Brk and SRMS SH2 Domains. ………………………………….29

Figure 6. Sequence specificities of the Frk and Btk SH2 Domains ……………………………………….……30

Figure 7. Sequence specificities of the Gads and GRAP SH2 Domains ..…………………………………….41

Figure 8. Sequence specificities of the shc1 SH2 Domain …………………………………………………………51

Figure 9. Sequence specificities of the Shc2, Shc3 and Shc4 SH2 Domains ………………………………52

Figure 10. Sequence specificity of the Vav3 SH2 domain …………………………………………………………63

Figure 11. Sequence specificity of the SHE SH2 domain ………………………………………………….……….64

Figure 12. Sequence specificity of the PLCG1-N and DAPP1 SH2 domains .………………………………80

Figure 13. Sequence specificity of the SLA2 and SH2D3A SH2 domains ……………………………………81

Figure 14. Sequence specificity of the Blnk SH2 domain ………………………….………………………………82

x

Introduction

SH2 domains Discovery and Function

Cellular signaling and regulation has been studied extensively by biologists and chemists for nearly fifty years.1 Phosphorylation is a post-translational modification in which a phosphate group is added to one of several different residues including tyrosine, threonine, serine, histidine and aspartic acid.2 This modification is one of the most significant and most studied mechanisms involved in cellular signaling pathways. Although protein phosphorylation has been reported for more than a century, its role in signaling was not fully elucidated until the mid-

1960s.2-3 The basic mechanism of protein-phoshorylation mediated cellular signaling involves an extracellular hormone binding to the cell surface triggering a conformational change in the protein resulting in intracellular phosphorylation. This initial phosphorylation event triggers a cascade of subsequent phosphorylation events via intracellular activation of protein kinases and binding of phosphoprotein binding domains. Eventually this cascade causes a change in gene regulation which can lead to various different outcomes ranging from cell proliferation and differentiation to apoptosis.3

Although it only account for about 2% of total protein phosphorylation in eukaryotes, a great deal of research has been focused towards phosphotyrosine.4 This is due to several reasons, the first being that phosphotyrosine is relatively stable and easy to study. Second, this seemingly minor class of protein phosphorylation has proven to be extremely important in understanding human cell signaling. As a result of the study invested into tyrosine 1 phosphorylated proteins, we have acquired a wealth of knowledge concerning the three main classes of protein domains associated with them. The first groups of protein domains, kinases, add a phosphate group to other proteins and affect a wide range of cellular responses. 5 Kinases are regulated by a number of factors, including phosphorylation, thus making them the targets of other kinases.6 This fact is consistent with the role of kinases in signal transduction where one protein activates the next down the line. Mutations and misregulation of these proteins typically have overtly harmful effects, often leading to cancer or cell death.7 Second, phosphatases, which remove a phosphate group from a protein. Initially, these enzymes were thought to be simple ”housekeeping enzymes,” but are now known to act in concert with kinases to effect influence cellular signaling.8 The final type of protein domain, Phosphoprotein binding proteins, link cellular response to protein phosphorylation. Like kinases and phosphatases, there are several different types of phosphoprotein binding proteins. First, the eponymously named Phoshotyrosine-binding domains (PTB) tend to recognize residues N- terminal relative to the phosphotyrosine. These proteins all share a similar fold, and are found in approximately 60 different proteins.9

A second and far more studied class of phosphotyrosine binding protein domain is the

Src homology 2 domain (SH2). As the name suggests, Src homology 2 domains were discovered in the protein Src. Whereas the Src homology 1 domain was later identified as a kinase, the Src homology 2 domain’s function was suggested as a possible Phosphotyrosine binding domain based off its proximity to the Src kinase and other mutational studies.10 When the first crystal structure of the Src SH2 domain was solved, it was co-crystallized with a phosphopeptide thus establishing SH2 domains as phosphotyrosine binding domains. Further, the crystal structure elucidated the general binding mode between an SH2 domain and a phosphopeptide, wherein 2 the phosphotyrosine is buried in an internal pocket with the rest of the peptide laying laterally across the SH2 domain (Figure 1).11

SH2 Structure and specificity determining elements

The initial publication of the SH2 domain’s structure revealed the general fold which all

SH2 domains exhibit. In this fold, there are three anti-parallel β-sheets flanked on either side by

2 alpha helices. There are four smaller β -sheets on the periphery of the protein, although many crystal structures do not mark them owing to their small size. Within the binding cleft of every

SH2 domain, there is a conserved arginine residue (βB5). The positively charged arginine is able to coordinate with the negatively charged phosphate and is necessarily for the binding between the phosphoprotein and the SH2 domain (Figure 1). Mutation of this arginine to any other residue confirms this necessity and results in loss of function of the SH2 domain.4

3

Figure 1. The Src SH2 domain bound to the peptide pYEEI. The phosotyrosine and isoleucine residues, important in the protein-peptide interaction, are colored in yellow. The βB5 arginine residue, which is required for SH2 domain binding, is colored in dark blue. The figure is adapted from the crystal structure (code 1shb) and was rendered using VMD software.

Thus far, 120 SH2 domains have been discovered in 115 different proteins.12 Owing to the sheer number of domains in this class, it has been postulated that each domain has a unique specificity. This postulate was confirmed in 1993 with a seminal publication which showed that several different SH2 domains bound to specific phosphopeptides.13 In this publication, the gross specificity of SH2 domains was established. Essentially, SH2 domains recognize residues

4 on either side of the phosphotyrosine. However, the residues immediately C-terminal, particularly at the +3 position, tend to be the most important. Further this paper suggests that individual SH2 domains fall into broad categories of similar binding. Finally, because this paper used only phosphopeptides, and not proteins, it can be inferred that SH2 domains do not recognize any secondary structures, and are able to recognize and differentiate simple disordered peptides.

Recently, there have been some advances in understanding SH2 domain binding specificity. First, in 2012, it was shown that mutation of three specific residues in an SH2 domain resulted in a modified domain which was able to bind phosphoproteins with 1000 fold greater affinity.14 The fact that “super-binders” could be made with the mutation of three residues suggests that these domains are not evolved for tight binding, but rather for reversible binding. This is in line with the proposed biological function of these molecules in signal transduction. In a different paper, a correlation was found between the presence of auxiliary loops and the general specificity of the SH2 domain.15 Again, these results are in line with what was previously established, in that each SH2 domain falls into a broad category based off of general binding.

Traditional methods of determining binding specificities

Although research suggests that individual SH2 domains can be grouped by a general specificity, there is evidence to suggest that each domain has a more narrow and unique specificity. First, if there are only a few different classes of SH2 binding, there would not need to be 120 SH2 domains in the human genome. It is possible that simply looking at the broad 5 specificity of SH2 domains could cause one to overlook subtle, but physiologically important differences in each SH2 domain’s specificity. Second, it has been established that an individual domain is capable of binding multiple different classes of peptide ligands.16 Finally, we cannot be certain that each domain binds only in a canonical fashion, as some domains have been shown to bind differently. 17 As a result, it is highly likely that each SH2 domain has a unique specificity designed to fit a biochemical niche. The only way to evaluate this possibility is to generate high-quality specificity information for each SH2 domain in the human genome.

There are multiple ways in which specificity could be mapped, each with their advantages and disadvantages. Biological methods such as yeast-two-hybrids, coimmunoprecipitation, and pull-down experiments are the traditional methods researchers have used to verify protein-protein interactions. However, these methods would hardly be suited for an endeavor such as this. First, these methods are not high-throughput, meaning that substantial time and resources would need to be dedicated to complete these experiments.

Second, these methods reveal nothing into the molecular interactions occurring. To acquire that data, these methods would rely on crystal structures, which are even more time-consuming and costly. Third, these methods require that the researcher has a general idea of the binding interaction which is going to occur. Aside from introducing bias into the experiment, this disadvantage reveals a far deeper problem. The number of phosphopeptides discovered is increasing every day. If information was published from one of these methods today, the data would be incomplete in just a few months. As such, these methods cannot be reliably used to map the specificity of each SH2 domain.

6

Combinatorial methods address many of the caveats which the above methods have.

First, they provide individual binding peptide sequences, which would paint a more complete picture of the binding profile of each individual domain. This profile, in turn, can be used to approximate the molecular interactions which are occurring, suggesting which residues are selected for and against. Second, combinatorial methods are by their nature high-throughput, making them ideal for large projects such as this. The two combinatorial methods which could be most easily used to solve this problem are phage display and One-Bead-One-Peptide chemical libraries. Phage display, as well as the other display technologies, is high throughput and can deliver the results required but has its disadvantages. First, any display technology besides phage display is notoriously difficult. Further, because SH2 domains require a phosphotyrosine for binding, the method would have to incorporate a phosphotyrosine onto a phage. This can be done, but would introduce another layer of complexity and difficulty to the experiment. Second, phage display can have trouble with avidity and non-specific interactions.

Third, generating the lists of data via phage display may prove to be prohibitively expensive, as each phage would take $5 to $10 to sequence. Fourth, phage display may lose low-abundance high-affinity ligands due to the screening process. Finally, phage display will select for only the best binders, and may miss lower-affinity classes of ligands altogether. Taken together, these caveats effectively discount the possibility of using these methods to solve our problem.

Chemical methods, such as one-bead-one-compound libraries and Oriented peptide

Array libraries address many of the problems inherent in phage display.18 Although incorporation of non-natural amino acids is trivial in peptide synthesis, these methods are not without their caveats. First, some sort of screen must be devised to select for the peptide in question. Second, there must be a method in place to identify each peptide on the surface of 7 the bead. Ideally, this method would ideally also be cheap and able to select for peptides missed by phage display. Oriented peptide array libraries also have an additional caveat, as they are unable to distinguish multiple classes of peptide ligands an individual domain may have.

Determination of SH2 Specificity via One Bead One Compound libraries

The method we have chosen to determine the specificity of the human complement of

SH2 domains relies on a One-Bead-One-Peptide chemical library.19 To address the problems caused by avidity, the concentration of peptide on the surface of the bead has been reduced. In order to accomplish this, we mixed 4:1 boc-Met-OSU:Fmoc-Met-OSU amounts of a differentially labeled amino acid as the first step. We then deprotected and blocked the boc labeled amino acid, resulting in a 10-fold reduction in surface peptide concentration (Figure 2). At the same time, we are able to keep the concentration of the peptide inside the interior of the bead high to facilitate identification of each peptide.20

8

Figure 2. Library used for screening. In this case, X is any of the naturally occurring amino acids minus methionine and cysteine. Methionine has been replaced with Norleucine, while cysteine has been replaced with α-aminobutyric acid or left out. Reagents and Conditions: (a) Soak in water overnight; (b) 0.4 equivBoc-Phe-OSu, 0.1 equiv Fmoc-Met-OSu, and 0.5 equiv DIPEA in 55:45Et2O/DCM, 30 min; (c) 4 equiv Fmoc-Met-OH, HBTU; (d) Solid phase organic synthesis HBTU/HobT; (e) split-and-pool synthesis by Fmoc/HBTU chemistry

Because this library is probing only direct interactions between the SH2 domain in question and the peptide library, a 5-Bromo-4-chloro-3-indolyl phosphate (BCIP) based screening method has been employed. Briefly, the SH2 domain is covalently labeled with Biotin, incubated with the beads, then with Streptavidin-Alkaline Phosphatase, which is then allowed to react with BCIP (Figure 3). This screening method works well for a number of reasons. First, it is fast and easy. It can be done in a single day with just a dissecting microscope and without much difficulty. Second, the amount of washing can be varied at each step, meaning results can easily be optimized. Finally, the screen generates colored beads of different intensity. Generally, it has been shown that more intensely colored beads correlate with tighter binding, although beads of all colors are valuable. 9

Figure 3. Screening Scheme. Binding of an SH2 domain to the peptide ligand on the outside of the bead results in individual beads turning turquoise/blue. These beads can then easily be picked manually.

The final caveat of a one-bead-one-peptide library method lies in how to determine the sequence of individual peptides on the surface of beads. In order to obtain usable data, we employ several different techniques and technologies. The overall goal of these methods is to generate peptide ladders, which can be read quickly and inexpensively using MALDI-TOF. To do this, we employ a special technology developed in this laboratory called Partial Edman

Degradation. Briefly, the technique involves reacting the free N-terminus with both a capping reactant (Fmoc-OSU) and a cleaving reactant (phenylisothiocynate, PITC) once for each random

10 position in the library. Next, the peptide ladder can be removed from the bead via a methionine residue and reaction to cyanogen bromide. Finally, the molecular weight of the cleaved peptide can be determined by MALDI-TOF and used to identify the random positions of the peptide.

Isobaric amino acids can be differentiated at this step via a capping peak, which is introduced during the synthesis of the library (Figure 4). This method, although seemingly complicated, provides a quick and inexpensive strategy to unambiguously determine the peptide sequence on the surface of selected beads.21

11

Figure 4. Partial Edman Degradation Scheme. The difference in mass between each peak of the peptide ladder corresponds to each individual residue. The overall sequence can be determined by determining the mass distance between each peak, and correlating that to an amino acid. In the above spectrum the leucine reside, which is isobaric with both isoleucine and norleucine, is unambiguously identified through the use of a +45 capping peak.

12

Experimental Protocols

Materials

The pGEX expression vectors used in this study were obtained as a gift from Professor

Shawn Li from the University of Western Ontario (London Ontario). The pET-32c expression vectors used were a gift from Dr. Gavin Macbeath from Harvard University (Cambridge,

Massachusetts). The competent BL21 cells used for expression in these experiments were purchased from Invitrogen (Carlsbad, California). The isopropyl β-D-1-thiogalactopyranoside

(IPTG) used for induction was purchased from Affymetrix (Santa Clara, California). All biochemical used to prepare buffers were purchased from Sigma-Aldrich (St. Louis,Missouri) unless stated otherwise. The trypsin inhibitor used during protein purification was purchased from Amresco. The resins and Amicron concentrators used for protein purification was purchased from Millipore (Darmstadt, Germany). The library was synthesized on 90 micrometer resin purchased from RAPP Polymere (Tubingen, Germany). Dichloromethane (DCM), triflouroacetic acid (TFA), N,N-dimethylformamide (DMF) and all other chemicals used in synthesis and Partial Edman Degradation were purchased from Sigma-Aldrich (St.

Louis,Missouri) unless stated otherwise. All Fmoc-protected amino acids used in the synthesis of the libraries along with the coupling reagents O-(benzotriazol-1-yl)-N,N,N,N- tetramethyluronium hexafluorophosphate (HBTU), (O-(7-azabenzotriazol-1-yl)-N,N,N,N- tetramethyluronium hexafluorophosphate) (HATU), hydroxybenzotriazole (HOBt) were purchased from Aapptec (Louisville, Kentucky). The streptavidin-alkaline phosphatase used in 13 screening was purchased from Promega (Fitchburg, Wisconsin). The Fmoc-OSU used in partial

Edman Degradation was purchased from Advanced Chemtech (Louisville, Kentucky). Finally, the

α-cyano-4-hydroxycinnamic acid (α-cca) used to perform MALDI-TOF analysis was purchased from Sigma-

Aldrich (St. Louis,Missouri).

Expression of the SH2 Domains

The plasmids coding for the SH2 domains used in this study were obtained as pGEX and pEt-32c vectors as gifts from Professor Shawn Li (University of Western Ontario) and Dr. Gavin Macbeath

(Harvard). The vectors were transformed into chemically competent BL21-DE3 Rosetta cells and relative expression was verified through the addition of isopropyl β-D-1-thiogalactopyranoside

(IPTG). Cells were grown in one liter of Lysogeny Broth (LB) to an OD600 of between 0.4 to 0.6 before being induced with IPTG. The cells were harvested by centrifugation at 5000 RPM for 15 minutes. The cell pellets were stored at -80 oC for no longer than 1 week before purification.

Purification and labeling of the SH2 Domains

The cell pellets were resuspended in 50 mL of Lysis buffer containing 150 mM Sodium

Chloride, 50 nM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), and 0.01% trition

X-1000 at a pH of 7.4 at room tempertature. Several protease inhibitors including trypsin inhibitor, phenylmethanesulfonylfluoride (PMSF) and aprotonin were added along with lysozyme, and the mixture was allowed to mix at 4oC with mixing for 20 minutes. Next, the resuspended pellet was subjected to ultra-sonicatiation to further lyse and membranate the solution. The solution was then centrifuged at 15000 RPM for 15 minutes. The resultant supernatant was then purified through the appropriate affinity column (GST-columns for proteins purified from pGEX vectors and Nickel TALON columns for the proteins purified from 14 pET-32c vectors). The presence or absence of protein was confirmed using SDS-PAGE. Typical yield varied based on the domain, but was generally 15 – 50 mg of protein per liter of cells.

Covalent Labeling of SH2 Domains

Purified proteins were labeled on their free lysine redisues using Biotin-NHS (obtained from New England Biolabs). Purified proteins were buffer exchanged using either 15 kDa

Amicron concentrators or G25 Columns such that the final buffer concentration was 150 mM sodium Chloride, 50 nM HEPES with a pH of 8.0. Biotin-NHS was then added in 2.5 molar excess and allowed to react for 1 hour at room temperature or overnight at 4oC. After the reaction is completed, the excess biotin was removed using a 15kDa Amicron concentrator. Protein was stored at -20oC for no longer than 1 week before use.

Library Synthesis

The library used in these studies was synthesized on 5 g of 90 µm Tentagel NH2 resin.

Because this library was reduced density, the beads needed to be segregated. To accomplish this, the beads were soaked overnight in water, drained, then quickly resuspended in a 55:45

(v/v) (DCM)/diethyl ether solution in containing 0.05 equivalents of Fmoc-Met-NHS and 0.45 equivalents of Boc-Met-NHS. The reaction was allowed to proceed for 30 minutes at room temperature on a rotary wheel. The beads were then washed with 55:45 DCM: diethyl ether and N,N-dimethylformamide (DMF). The resin was then reacted with 5 equivalents of Fmoc-

Met-OH with HBTU/HOBt/NMM in DMF for 1 hour. The Boc group was then removed with 70%

TFA in water (v/v) for 2 hours. The free amines were then capped with acetic anhydride catalyzed with DMAP/NMM for 30 minutes.

15

The linker region LNBBRM (where B is Beta-alanine) was synthesized with 4 equivalents of the corresponding fmoc-amino acids, with four equivalents of HBTU/HOBt and eight equivalents NMM in DMF. The reaction was allowed to proceed at room temperature in a closed vessel on a rotatory wheel for a period of 1 hour. Afterwards, the resin was washed three times with DMF, dichloromethane (DCM), and three more times with DMF. The Fmoc group was removed with 20% piperidine in DMF (5 minutes followed by 15 minutes). The beads were then washed again as described above. Random positions were generated by splitting the library into nineteen equal portions containing each of the amino acids to be synthesized. Each vessel was coupled with the corresponding amino acid, HBTU/HOBt, and NMM as described above. Phosphotyrosines were coupled using 2.5 equivalents of the Fmoc-pY-OH for 2 hours with HATU. Leucine and lysine, which are isobaric with isoleucine and glutamine (respectively) were additionally reacted with 3% deuterated acetic acid during coupling to generate capping peaks. Additionally, Norleucine was reacted with 3% deuterated proprionic acid during coupling for the same reason. Once the library was completed, the N-terminal Fmoc group was removed using 20% piperidine as described above. Next, the side-chain protecting groups were removed using Reagent K (TFA with 7.5% Phenol, 5% thioanisole, 5% water, 2.5% ethanedithiol, and 1% anisole) for 2 hours at room temperature. The resin was then washed with TFA, DCM, and finally DCM containing 5% DIPEA. After deprotection, the library was desiccated and stored at -

20oC.

Screening SH2 Domains against the pY library

Although the amount of resin used in each screen varied, the standard screen was performed by weighing 30 mg of deprotected and dried library in a 2 mL Biorad column. The

16 resin was then resuspended in 2 mL blocking buffer (30 mM HEPES, 150 mM sodium chloride,

0.1% gelatin, 0.05% sodium azide, 0.01% TWEEN, at a pH of 7.4) for 1 hour before use at 4oC.

Next, 250 nM of the biotin labeled protein was added to the solution and the mixture was allowed to reach equilibrium overnight. The next morning, the solution was drained, washed with blocking buffer and then re-suspended in 1mL of streptavidin-akalaline-phosphatase- binding-buffer (10 mM magnesium chloride, 0.07 mM zinc chloride, 30 mM Tris, and 250 mM sodium chloride at a pH of 7.4). To this solution, 1 µL of streptavidin-akaline phosphatase was added, and the mixture was allowed to incubate at 4oC for 20 minutes. The solution was then drained, washed, and finally resuspended in 1mL of staining buffer (30 mM Tris, 100 mM sodium chloride, 5mM magnesium chloride, 0.02 mM zinc chloride at a pH of 8.4). Using a micropipette, the beads were resuspended and placed in a 3.5 cm petri dish, to which 100 µL of 5 mg/mL BCIP was added. The reaction was allowed to proceed until the appearance of blue beads, at which time the reaction was quenched with addition of 1 M HCl and the beads were picked using a micropipette under a dissecting microscope. Successful screens typically take between 20 minutes and 1 hour for adequate staining to occur, but longer times were noted.

Partial Edman Degradation

Beads from the previous step were sorted based on color, placed into specialized vessels, and washed three times with water, pyridine, then 2:1:0.1% pyridine:water:triethylamine. After washing, the beads were reacted with 80:1 Fmoc-

OSU:phenylisothiocynate in 2:1:0.1% pyridine:water:triethylamine for 6 minutes at room temperature. Afterwards, the vessels were drained, and the beads washed once with pyridine, then three times with DCM. The beads were then reacted with TFA for 6 minutes, drained, and

17 reacted with TFA again for 6 minutes. Afterwards, the beads were washed three times with

DCM, and reacted again with 80:1 Fmoc-OSU:PITC. This cycle of reactions was repeated eight times, one for every residue that needed to be identified.

After all of the required rounds of Partial Edman degradation, the beads were washed three times with DCM and five times with DMF. Next, the N-terminal Fmoc groups were removed by reaction with a 20% piperidine solution in DMF twice, for 5 then 15 minutes. Next, the beads were washed three times with DMF and five times with DCM. The beads where then reduced by reaction with a spatula of ammonium iodide and 50 µL methyl sulfide in TFA for 20 minutes at room temperature. The beads were then drained, and washed 10 times with water before being picked out of the vessel and placed into individual Eppendorf tubes. The beads were then reacted with 40 mg/mL cyanogen bromide in 70% TFA overnight (12-14 hours). The next morning, the cyanogen bromide and 70% TFA were removed using a speedvac. The cleaved peptides were stored dry, at 4oC for no more than 1 week before sequencing via MALDI-

TOF.

MALDI-TOF analysis

The dried peptides were dissolved in 5 µL of a 1:1:0.1% mixture of acetonitrile:water:TFA and vortexed three times. Next, 1 µL of this solution was mixed with 2 µL of a solution of 10 mg/mL α-cca dissolved in a 1:1:0.1% mixture of qcetonitrile:water:TFA. Next,

1 µL of this solution of this solution was spotted onto a MALDI plate, and analyzed via MALDI-

TOF.

18

Results

Library Used

The library used for these studies was designed specifically to bind to SH2 domains, that is, it contained a central pY flanked on either side by random positions. The overall design of this library was NH2-XXpYXXXXLNBBRM-bead or NH2-XXpYXXXLNBBRM-bead, where X is any one of the naturally occurring amino acids excluding cysteine and methionine and including norleucine and alpha-aminobutyric acid (as structural surrogates for methionine and cysteine respectively). The surface density of the peptide in this library was reduced by 90%, as described in the methods section, to reduce effects from avidity. In the linker region, the methionine has been strategically placed to allow for cleavage of the peptide from the bead, the arginine is present to provide a fixed positive charge for the peptide (which aids in MALDI-TOF analysis) and the beta-alanines have been included to increase the overall flexibility of the peptide (Figure 2).

Kinase SH2 Domains (Itk, Brk, SRMS, Frk, Btk)

The library NH2-XXpYXXXLNBBRM-bead was screened against the kinase SH2 domains

Itk, Brk, SRMS, Frk, and Btk.22 With the exception of one class of peptides discovered from the

Itk SH2 domain, the specificities of these SH2 domains are all like that of Src in that the +3 position (relative to the pY) is hydrophobic with less specificity in other positions. The first SH2 domain listed, IL2-inducible T-cell Kinase (Itk) , was screened against 100 mg of dry resin and

19 produced 117 beads. From the results of these screens, two distinct families of peptides were discovered. The first class of peptide binders contained 51 hits and had the general trend of

XXpYEN(API), where letters in brackets designate different amino acids at the same position and upper case letters indicate a higher abundance than lower case letters. The second class of peptides had 66 hits, and followed the general trend XXpY(Eda)N(Ed). Interestingly, for both classes of peptides, Itk selects an N at the +2 position, making it similar to the Grb family discussed below. Also of note is the fact that the first family selects proline at the +3 position as much as it selects for alanine or leucine, making this protein distinct among the ones discussed here. The Itk-SH2 domain is known to bind to Vav1, seemingly through its SH2 domain.23 It appears that this interaction would fall into the second class of the Itk-SH2 domain’s specificity

(Table 8).

The next SH2 domain, Breast Tumor Kinase (Brk), was screened against approximately

100 mg of dried resin and produced 157 colored beads. Fortunately, these hits all fell into the general family XXpY(DE)i(VCy). There are a few things to note about the specificity of this protein. First, there is a minor preference for isoleucine at the +2 position, which makes this protein unique amongst the kinases discussed here. Also, at the +3 position, there is a selection for both valine and cysteine. When this library was constructed, it contained alpha- aminobutyric acid (Abu, sometimes abbreviated U or C) as a replacement for cysteine. This unnatural amino acid was used owing to the fact that it cannot form disulfide bonds. We have noted that Abu is often selected with valine, alanine, and isoleucine, suggesting that its selection simply means that a small aliphatic residue is selected at the position. Further, because Abu does not exist in nature and cysteines are unlikely to be situated on a disordered loop on the

20 surface of a protein, the selection of Abu does not mean that cysteine is selected for by the domain. As such, its use in this project has been discontinued.

The Btk protein has been studied rather extensively given its link to breast cancer. As such, many of its binding partners and phosphorylation targets are know. In particular, the Brk

SH2 domain is known to have two known binders. The first binder is the Insulin Receptor

Substrate 4 (IRS-4) protein, which is involved in numerous pathways.24 The second is Sam68, an adapter protein known to bind to RNA and affect cell cycle progression.25 Both of these proteins have regions which match the specificity of the Brk SH2 domain (Table 9).

The Third kinase SH2 domain which was screened in these experiments was the SRMS

SH2 domain. About 150 mg of resin was used to screen this domain, and only about 116 sequences were obtained. The results for this domain were messier than normal, and followed the overall trend was HXpY(EH)X(ILVMCF). Like Brk, SRMS also selected for Abu at the +3 position, but also selected for essentially all other small aliphatic amino acids as well, suggesting a weak specificity at this position. Inspection of the selected peptides may lead one to believe that histidine was selected at the -2 and +1 positions. Unfortunately, these moieties may also be due to non-specific interactions, and may not be real. Further, evidence for this protein’s existence is only at the transcript level, meaning it may or may not actually be expressed. Taken together, these factors suggest the data for this domain should not be taken as infallible. As this protein has not been observed outside of the genomic level, there are no known binders.

The fourth kinase SH2 domain studied was the Fyn-Related Kinase, or Frk.

Approximately 100 mg of resin was used to screen this domain, and 114 sequences were obtained. All 114 sequences followed the general trend of HXpYEX(IL). The first aspect of the

21 specificity to note is the presence of H at -2 position. Like the SRMS SH2 domain, this may or may not be real. The second thing to note is the selection of predominately isoleucine or leucine at the +3 position. Although the specificity of this domain is fairly similar to other kinases, it is not identical to them. The Frk SH2 domain is known to have at least one physiological binder in the Retinoblastoma-associated protein (Rb).26 This protein is a tumor suppressor which has a tyrosine phosphorylation motif which fits the Frk-SH2 domain’s specificity profile (Table 11).

The final kinase SH2 domain screened in this study was from Burton’s Tyrosine Kinase, or Btk. Roughly 100 mg of resin was screened against the domain resulting in 88 positive hits.

The general specificity for this domain was XXpYDX(VC). Like Brk, Btk prefers valine and Abu at the +3 position. However, unlike any other SH2 domain studied in this group, Brk prefers aspartic acid at the +1 position rather than glutamic acid. Interestingly, the known physiological substrates of this domain have an aspartic acid at the +1 position (Table 12). The first of these is caveolin-1, a scaffolding protein known to be involved in T-cell activation.27 The second known binder is the Toll-like receptor 8, which is involved in various pathways related to immune response.28

22

Table 1. SH2 domains Examined in this study

23

Table 2. Peptides selected against the Itk-SH2 Domain

24

Table 3. Peptides selected against the Brk-SH2 Domain

25

Table 4. Peptides selected against the SRMS-SH2 Domain

26

Table 5. Peptides selected against the Frk-SH2 Domain

27

Table 6. Peptides selected against the Btk-SH2 Domain

28

Figure 5. Sequence specificities of the Itk, Brk and SRMS SH2 Domains

29

Figure 6. Sequence specificities of the Frk and Btk SH2 Domains

30

Table 7. Predicted binding partners of the Itk-SH2 Domain (Class 1)

Table 7. Predicted binding partners of the Itk-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

31

Table 8. Predicted binding partners of the Itk-SH2 Domain (Class 2)

Table 8. Predicted binding partners of the Itk-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Itk-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.23

32

Table 9. Predicted binding partners of the Brk-SH2 Domain

Table 9. Predicted binding partners of the Brk-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Brk-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.24 25

33

Table 10. Predicted binding partners of the SRMS-SH2 Domain

Table 10. Predicted binding partners of the SRMS-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

34

Table 11. Predicted binding partners of the Frk-SH2 Domain

Table 11. Predicted binding partners of the Frk-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Frk-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.26

35

Table 12. Predicted binding partners of the Btk-SH2 Domain

Table 12. Predicted binding partners of the Btk-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Btk-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.27 28

36

Adapter SH2 Domains

Two adapter SH2 domains were successfully screened and studied. Both of these SH2 domains belong to the Grb2 family of SH2 domains, which is a widely studied cancer drug.29

These SH2 domains are similar in that they require binding peptides to have an asparagine at the +2 position. This specificity arises due to an enlarged loop in the SH2 domain, leading to the phosphopeptides being bent and a special pocket for the asparagines.15 About 100 mg of the

NH2-XXpYXXXLNBBRM-bead library was screened against both the GRB2-related adapter protein

SH2 Domain (GRAP) and the GRB2-related adapter protein 2 SH2 Domain (GRAP 2 or Gads). The

GRAP SH2 domain selected 114 different peptides which all followed the general trend of

(lv)XpY(iav)N(LV). The first thing to note about this SH2 domain’s specificity is the weak specificity on the N-terminal side of the phosphotyrosine. Weak specificity on the N-terminal side is relatively uncommon, and likely real. The GRAP SH2 domain also selects for small aliphatic residues at the +1 and +3 position. At the +1 position, it appears as though alanine may be the most selected suggesting a pocket just the right size for alanine (as no Abu was selected). At the +3 position, leucine and valine are predominately selected. The Gads SH2 domain selected 98 distinct peptides following the general trend of XXpY(Ca)N(CI). In this case the selection of Abu at both the +1 and +3 position suggests a slightly less specific binding interaction which can accommodate the slightly larger residue. Interestingly, valine is selected rarely at either position, suggesting this residue may in fact be selected against.

These proteins each contain 1 SH2 domain and 2 SH3 domains, each of which can bind to different classes of peptide ligands. The inclusion of these domains is fundamental to the function of these proteins as adaptors. For the purposes of this study, proteins that were

37 known to specifically bind to the SH2 domains of these proteins were examined. The Gads SH2 domain is known to bind to two physiological substrates. The first of these is the Lck-interacting transmembrane adapter 1 (LIME1) which is involved in multiple pathways in relating to immune cell signaling.30 The Second of these is the Linker for activation of T-cells family member 1 (LAT), which is involved in immune cell maturation.30 Interestingly, both of these proteins have a hydrophilic residue at the +1 position relative to the phosphotyrosine to which they are thought to bind. This is slightly at odds with the specificity data generated, and may suggest the +1 position is not as important for binding as the +2 and +3 positions (Table 15). The GRAP SH2 domain is known to bind to three different proteins. Each of these proteins has multiple sites at which GRAP could associate, but the exact sites are not known. The first of these proteins is

Mast/stem cell growth factor receptor Kit (Kit), which plays a role in numerous different pathways related to a multitude of cellular functions.31 The second of these is the

Erythropoietin receptor (EpoR), which regulates the differentiation and proliferation of erythroblasts.31 The third is the Linker for activation of T-cells family member 1 (LAT), which also binds to Gads.32

38

Table 13. Peptides selected against the Gads-SH2 Domain

39

Table 14. Peptides selected against the GRAP-SH2 Domain

40

Figure 7. Sequence specificities of the Gads and GRAP SH2 Domains

41

Table 15. Predicted binding partners of the Gads-SH2 Domain

Table 15. Predicted binding partners of the Gads-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Gads-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.30

42

Table 16. Predicted binding partners of the GRAP-SH2 Domain

Table 16. Predicted binding partners of the GRAP-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the GRAP-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.31 32

43

Shc Family SH2 domains

The Shc family SH2 domains are all found in the Shc family of proteins and consist of

Shc1, Shc2, Shc3, and Shc4. Shc1 has attracted the most attention, and is involved with the regulation of apoptosis.33 The other members of the family are not expressed outside of neural tissue, and shc4 is found overexpressed in melanomas.34 A greater understanding of these proteins, and particularly their SH2 domains, may aid in understanding neurological development.

The SH2 domains of all members of this family were screened against approximately

100 mg of the NH2-XXpYXXXXLNBBRM-bead resin. The Shc1 SH2 domain selected three related classes of peptide binders. The first class contained 68 peptides and had the general specificity of (I/L)pY(HTSq)X(ILyw)(WY). The second class contained only 14 peptides and had the general trend XXpY(IY)X(TSn)(WY). The third and final class of peptides consisted of only 9 peptides, and followed a general specificity (va)(if)pY(nqt)iH(iva). The first two classes are related in that they contain an aromatic residue at the +4 position. Both of these classes tend to not select phenylalanine, meaning that a slightly more hydrophilic aromatic residue may be required. The two classes seem to be inverted at the +1 and +3 positions, as 1 class prefers a hydrophilic residue where the other prefers a hydrophobic one. This trend may suggest the peptide is capable of binding in multiple orientations. The final class is defined by the Histidine at the +3 position, as well as an aliphatic residue at the +4.

As Shc1 is the most studied of the Shc1 family, it is not surprising that there are many proteins known to associate with it. As before, the interactions that are examined are those known to be mediated through its SH2 domain. Interestingly, the binding partners of the Shc1

44

SH2 domain can be divided into the specificity classes discovered through screening. The first specificity class matches phosphopeptides from four different proteins (Table 21). The first,

Discoidin domain-containing receptor 2 (DDR2), is a tyrosine kinase involved in regulating cellular differentiation.35 The second protein ALK tyrosine kinase receptor (ALK) plays a role in neuronal development.36 The third known binder of the Shc1 SH2 domain is the Ephrin type-B receptor 1 (EphB1) protein, which is involved in intercellular signaling.37 The final protein that fits this class of ligands is Platelet-derived growth factor receptor beta (PDGFRb), which is involved in the regulation of a milieu of disparate cellular functions.38 The second specificity class matches peptides from three different proteins, Kit,39 EphB1, and PDGFRb; all of which have been discussed previously.

The second member of the Shc family, shc2, proved to be rather unstable, and seemed to lose its function after being frozen. Further, it seemed to last considerably shorter at 4oC or -

20oC than other SH2 domains, lasting at most a night. To overcome this problem, the protein was expressed, purified, labeled, and screened on the same day. Although it was screened against 100 mg, only 51 peptides were obtained. However, as all the peptides seemed to belong to a single class, X(eg)pY(edi)i(IY)X, we have reached the conclusion that the screens were successful. The specificity of Shc2 is vastly different from any of the classes generated from shc1. The domain has a preference for negatively charged residues at the +1 position, and tyrosine at the +3 position, making it unique among the Shc family proteins. The Shc2 SH2 domain is known to bind to Grb2, an adaptor protein whose misregulation leads to multiple types of cancer (Table 23).40

45

The third member of the Shc family, Shc3, selected 2 classes of peptides. The first class contained 106 and the second contained 26, having the specificities XXpY(Qnedhr)X(IMFV)X and

XXpYXX(Stqn)(YW) respectively. These two specificities are similar to those found in shc1, in that the positions the hydrophilic residues seem to be inverted. One main difference here is the position of the hydrophobic residues. The first class prefers mainly aliphatic residues at the +3 position, whereas the second class prefers mainly aromatic residues at the +4 position. It is believed that the Shc3 SH2 domain binds to Grb2, though the interaction does not seem to fit the specificity profile (Table 24).40

The final member of the Shc family, Shc4, selected 70 peptides of the similar class

(IQ)XpY(IMFY)(IMQN)(IMQN)X. This SH2 domain appears to bind to aliphatic residues at positions +1, +2, and +3. This is seen in only a few other SH2 domains, and in none of the other

Shc family domains. Interestingly, expression of this SH2 domain is linked with cancer, suggesting that non-specific interactions of this domain may be oncogenic 34. Although Shc4 seems to be an interesting cancer target, there are no known binders for this protein domain.

46

Table 17. Peptides selected against the Shc1-SH2 Domain

47

Table 18. Peptides selected against the Shc2-SH2 Domain

48

Table 19. Peptides selected against the Shc3-SH2 Domain

49

Table 20. Peptides selected against the Shc4-SH2 Domain

50

Figure 8. Sequence specificities of the shc1 SH2 Domain

51

Figure 9. Sequence specificities of the Shc2, Shc3 and Shc4 SH2 Domains

52

Table 21. Predicted binding partners of the Shc1-SH2 Domain (Class 1)

Table 21. Predicted binding partners of the Shc1-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Shc1-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.35 36 37 38

53

Table 22. Predicted binding partners of the Shc1-SH2 Domain (Class 2)

Table 22. Predicted binding partners of the Shc1-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Shc1-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.

54

Table 23. Predicted binding partners of the Shc2-SH2 Domain

Table 23. Predicted binding partners of the Shc2-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Shc2-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.40

55

Table 24. Predicted binding partners of the Shc3-SH2 Domain (Class 1)

Table 24. Predicted binding partners of the Shc3-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

56

Table 25. Predicted binding partners of the Shc3-SH2 Domain (Class 2)

Table 25. Predicted binding partners of the Shc3-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

57

Table 26. Predicted binding partners of the Shc4-SH2 Domain

Table 26. Predicted binding partners of the Shc4-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. 58

Vav3 and SHE SH2 Domains

The Guanine nucleotide change factor Vav3 SH2 domain contained mutations on the

SH2 domain and was thus overlooked. Upon inspection of the structure of vav1 and vav2, it was estimated that the mutations would fall on regions distil to the phosphopeptide binding pocket of the SH2 domain. Thus, these mutations would probably not affect the binding interaction.

Indeed, the specificity of vav3 is similar to the other members of the family it is from (vav1 and vav2, data not shown). The Vav3 SH2 domain selected 4 different classes of peptides. The first class contained 38 peptides following the general trend X(pg)pY(IM)(Nqts)(ed)X. The first class is interesting in that it resembles the Grb2 family on the C-terminal side, while containing a small residue, glycine or proline, at the -1 position. The second class consisted of 31 peptides with the trend (wfqn)(Ed)pY(IMV)(FYivg)XX. The presence of a glutamic acid at the -1 position on these peptides made this family unique as well. The third family contained 20 peptides following the trend (IAt)(Iqat)pY(IM)(DEIM)(Pg)X. The final family contained 15 peptides each following the trend (Vig)(hin)pY(IM)GXX. The presence of small residues at the +3 and +2 positions for the third and fourth family (respectively) are unique among other Vav family binding peptides.

Further, the selection of a glycine at the +2 position is unique among all SH2 domains. The Vav3

SH2 domain is known to bind to the Ephrin type-A receptor 2 (EphA2), which is known to influence intercellular interactions.41 Interestingly, EphA2 contains peptides which fall into three of the four classes of peptides selected by Vav3 (Tables 29, 31 and 32). Although, the literature suggests that the interactions described as Class 3 (Table 31) are physiologically relevant, this pattern may suggest Vav3 has evolved multiple ways to bind to its partners.41

59

Like the Vav3 SH2 domain, the SHE SH2 domain was started but abandoned. Unlike the

Vav3 domain, the reason for this was failure at the screening step. As such, we feel the poor results of this domain can be attributed to lower than average stability. To circumvent this obstacle, we expressed, purified, labeled and screened this domain in a single day. As a result, screening 100 mg of dry resin revealed two distinct families of peptide ligands. The first family contains 86 peptides, which follow the overall specificity (ia)(ig)pYE(eni)X(iy). The most striking feature of this family is the strong selection for only glutamic acid at the +1 position. Further, the domain appears to have weaker specificity at the -1, -2, and +4 positions, selecting hydrophobic and aliphatic residues at those positions. The second class of ligand followed the more canonical specificity (ige)XpY(IMV)e(Psa)(de). In this class, hydrophobic residues are selected at the +1 position, while smaller residues, particularly proline, are selected at the +3 position. Because the SHE protein has not been demonstrated to be expressed, it has not been implicated in any protein-protein interactions.

60

Table 27. Peptides selected against the Vav3-SH2 Domain

61

Table 28. Peptides selected against the SHE-SH2 Domain

62

Figure 10: Sequence specificity of the Vav3 SH2 domain

63

Figure 11. Sequence specificity of the SHE SH2 domain

64

Table 29. Predicted binding partners of the Vav3-SH2 Domain (Class 1)

Table 29. Predicted binding partners of the Vav3-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Vav3-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.41 65

Table 30. Predicted binding partners of the Vav3-SH2 Domain (Class 2)

Table 30. Predicted binding partners of the Vav3-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Vav3-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.41

66

Table 31. Predicted binding partners of the Vav3-SH2 Domain (Class 3)

Table 31. Predicted binding partners of the Vav3-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Vav3-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.41

67

Table 32. Predicted binding partners of the Vav3-SH2 Domain (Class 4)

Table 32. Predicted binding partners of the Vav3-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Vav3-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.41

68

Table 33. Predicted binding partners of the SHE-SH2 Domain (Class 1)

Table 33. Predicted binding partners of the SHE-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. 69

Table 34. Predicted binding partners of the SHE-SH2 Domain (Class 2)

Table 34. Predicted binding partners of the SHE-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

70

Blnk, SLA2, DAPP1, PLCG1, SH2D3A SH2 Domains

Through the course of screening the SH2 domains sent from Dr. Shawn Li, it was determined that several of the SH2 domains would not express or were otherwise nonfunctional. Rather than re-cloning the domains which did not work, we opted instead to request the domains from a verified source.42 The domains were cloned into a different construct, containing a six-histidine tag and thioredoxin domain. Because labeling typically occurs on a GST domain, which contains 20 lysine residues, there was initial concern that labeling may be a problem. However, neither labeling nor purification of these domains proved to be difficult. Each of these domains was screened against approximately 100 mg of the NH2-

XXpYXXXXLNBBRM-bead library.

The first SH2 domain to be studied from this group was the N-terminal SH2 domain from the protein 1-phosphatidylinositol 4,5-bisphosphate phosphodiesterase gamma-1 (PLGC1). This

SH2 domain selected 2 similar, but distinct classes of phosphopeptide binders. The first class contained 51 peptides and followed the general trend YXpY(Ivm)(Ivmd)(ivmt)X, while the second contained 25 sequences and followed the trend (hy)(Y)pY(HT)(ivfm)(ivf)(Yvs). Although both classes seemed to select aliphatic residues at the +2 and +3 positions as well as a tyrosine at the

-2 position, a covariance was found between the +1 and the +4 and -1 positions. If the +1 position was hydrophobic, the +4 and -1 positions were typically not required for binding. If the

+1 position was hydrophilic, typically histidine or threonine, then the +4 and -1 positions were typically aromatic. Like the Shc1 SH2 domain, the different specificity classes of PLCG1-N correlate to different physiological ligands (Table 40). The first class of peptidyl binders identified through screening recognized three different proteins, being LAT, PDGFRa, and

71

PDGFRb.32, 43 The second class of peptides correlated to five known physiological substrates

(Table 41). The first three of these are Fibroblast growth factor receptors (FGFR2, FGFR3, and

FGFR4) each of which regulate a plethora of cellular processes which for brevity concerns, will not be elaborated on here.44 The second is villin, which plays a role in actin-remodeling during mitosis.45 The third protein known to bind to PLCG1-N following this trend is PDGFRa, which also bound through the first motif.

The second SH2 domain to be studied in this group was the DAPP1 SH2 domain, which could not be expressed using the old constructs. This domain selected two different classes of peptides, each selecting for arginine at a specific position. The first class contained 67 peptides and followed the trend hXpY(Enqs)(FYHwi)(Iv)R, selecting an arginine at the +4 position. The second class contained 21 sequences and followed the trend (Hnq)(ivrk)pYX(MI)(RK)Y, selecting an arginine at the +3 position. It should also be noted that both classes selected for an aliphatic residue on the position immediately N-terminal of the arginine. Although a fair amount of research has been focused on DAPP1, none has been done concerning the binding partners of the protein’s SH2 domain.

The third domain to be studied from this group was the B-cell linker SH2, or Blnk. This domain was attempted, but ultimately unsuccessful using the plasmids from Dr. Shawn Li’s group. With the new plasmid however, the Blnk SH2 domain’s specificity was shown to fall into three separate classes when screened. The first class contained 52 peptides and followed the general trend ifpYD(ivfy)I(fy). This class seemed to specifically select aspartic acid at the +1 position, and Isoleucine at the +3 position, and did not select even close structural analogs. The second class of peptides contained 29 peptides, and had the general trend (yi)XpY(Yfw)(YI)X(iv),

72 selecting primarily aromatic residues at the +1 position and hydrophobic residues at the +2 position. The final class of peptides selected against the Blnk-SH2 domain contained 31 peptides and followed the trend XXpY(nqhsy)(De)(Iv)(ivyf). This class of peptides seems close to the first, with aspartic acid being selected predominately over glutamic acid, but at the +2 position rather than at the +1 position. This coincidence may be due to the peptide shifting one residue over.

The Blnk-SH2 domain selected three distinct classes of peptide binders, two of which seemed to correlate to physiological binders. The first class of peptides selected by Blnk matched three physiological ligands (Table 44). The first of these is the kinase Syk, which regulates numerous different cellular processes.46 The second is B-cell antigen receptor complex-associated protein alpha chain (CD79A), which is involved in immune response.47 The final protein is SLP adapter and CSK-interacting membrane protein (SCIMP), which mediates

MHC-II antigen presentation.48 The second class of Blnk binding peptides correlates to some of those found in the Syk kinase (Table 45).

The fourth domain to be screened was the Src-Like Adapter 2, or SLA2. Unlike Blnk, which seemed to fail at the screening stage with the old construct, SLA2 was unable to be successfully expressed in the old construct. Luckily, the protein showed no problems in the new construct, and was successfully screened, yielding 2 peptide classes. The first contained 38 peptides following the general trend (hya)(ei)pY(eqhts)(Enyi)(Imp)(ry), selecting a glutamic acid at the +2 position and an isoleucine at the +3. The second class consisted of 25 peptides and followed the trend X(Iv)pY(yva)(Im)RN. This class is unique among SH2 domains for selecting both an arginine at the +3 position and an asparagine at the +4. Although there are listed

73 binders of this SH2 domain, it should be noted that these are done by similiarity, meaning that that are actually SLA1 binders. As such, these binders have not been verified.49

The final SH2 domain examined in these studies was the SH2 domain of the protein

SH2D3A. This SH2 domain selected two binding motifs, but was different from every other SH2 domain studied in that it did not seem to select aliphatic residues at all. The first class of peptides which bound to SH2D3A contained 47 sequences and followed the general trend

XXpY(Wyf)XX(Fwy). This class of peptide binder seemed to contain only aromatic residues at the

+1 and +4 positions, suggesting 2 deep binding pockets in which these residues could be wedged. The second class of peptide contained 37 peptides following the general trend yfXpYPXnX. This second class of peptides is unique as well, as no other SH2 domains selected a proline at the +1 position or an asparagine at the +3 position. Unfortunately, no crystal structure of this domain exists, limiting the possibilities for further studies. There are currently no known binders to the SH2D3A SH2 domain.

74

Table 35. Peptides selected against the PLCG1-N-SH2 Domain

75

Table 36. Peptides selected against the DAPP1-SH2 Domain

76

Table 37. Peptides selected against the Blnk-SH2 Domain

77

Table 38. Peptides selected against the SLA2-SH2 Domain

78

Table 39. Peptides selected against the SH2D3A-SH2 Domain

79

Figure 12: Sequence specificity of the PLCG1-N and DAPP1 SH2 domains

80

Figure 13. Sequence specificity of the SLA2 and SH2D3A SH2 domains

81

Figure 14. Sequence specificity of the Blnk SH2 domain

82

Table 40. Predicted binding partners of the PLCG1-N-SH2 Domain (Class 1)

Table 40. Predicted binding partners of the PLCG1-N-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the PLCG1-N-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.32, 43

83

Table 41. Predicted binding partners of the PLCG1-N-SH2 Domain (Class 2)

Table 41. Predicted binding partners of the PLCG1-N-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the PLCG1-N-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.44 45

84

Table 42. Predicted binding partners of the DAPP1-SH2 Domain (Class 1)

Table 42. Predicted binding partners of the DAPP1-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

85

Table 43. Predicted binding partners of the DAPP1-SH2 Domain (Class 2)

Table 43. Predicted binding partners of the DAPP1-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

86

Table 44. Predicted binding partners of the Blnk-SH2 Domain (Class 1)

Table 44. Predicted binding partners of the Blnk-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Blnk-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.46 47 48

87

Table 45. Predicted binding partners of the Blnk-SH2 Domain (Class 2)

Table 45. Predicted binding partners of the Blnk-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. Known binders of the Blnk-SH2 domain were found at Uniprot.org, the binding motifs were estimated based on the specificity of each domain.46 47 48

88

Table 46. Predicted binding partners of the Blnk-SH2 Domain (Class 3)

Table 46. Predicted binding partners of the Blnk-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

89

Table 47. Predicted binding partners of the SLA2-SH2 Domain (Class 1)

Table 47. Predicted binding partners of the SLA2-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above. No known binders of the SLA2-SH2 domain are known, the ones recorded are based off SLA2 similiarity to SLA.49

90

Table 48. Predicted binding partners of the SLA2-SH2 Domain (Class 2)

Table 48. Predicted binding partners of the SLA2-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

91

Table 49. Predicted binding partners of the SH2D3A-SH2 Domain (Class 1)

Table 49 Predicted binding partners of the SH2D3A-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

92

Table 50. Predicted binding partners of the SH2D3A-SH2 Domain (Class 2)

Table 50. Predicted binding partners of the SH2D3A-SH2 domain, derived from the specificity profile presented in this document. Phosphoproteins, taken from phosphosite plus (phosphositeplus.org) which match the specificity profile are recorded above.

93

Discussion

Using a One-bead-one-peptide chemical library, the specificity of 18 SH2 domains were determined. The information obtained from these studies is unique not only in content, but also in quality. Many of the domains used in this study have received next to no research attention, despite their roles in signaling pathways. Further, the information obtained is truly one of a kind as no specificity study conducted on SH2 domains has been able to determine, with single residue accuracy, the specificity profile of these domains. The data from these studies has been summed up along with the rest of the group’s data to produce a comprehensive list of all successfully screen SH2 domains along with their specificity profile. We hope that by making this list accessible to others, new avenues of research can be opened, and new ideas can be explored.

Although these studies were successful in producing specificity profiles, it should be noted that individual peptide sequences from these families have not been confirmed by binding studies. Traditionally, this would be done via either Surface Plasmon Resonance or

Fluorescence Anisotropy. In either case, previous studies have confirmed there is a direct correlation between color formation on the surface of the bead during screening and binding affinity between a protein and the peptide on the surface of the bead.16 Further, these verified binding results have been shown to correlate to bona fide binding proteins. As a result, a hit generated by this method is a very valuable piece of data, in that our method has been shown to be predictive of physiologically relevant interactions. 94

Another discovery made from this study is that, while related SH2 domains have similar specificity profiles, no two SH2 domains are exactly the same. This observation supports the notion that each SH2 domain has a specific niche which that only it can fill. As a result, further investigation of the subtle differences in specificity of related domains may reveal hitherto undiscovered and profound implications. Further, our method is the only one capable of revealing whether a protein is binding to multiple classes of peptide ligands. This work suggests, particularly in the cases of Shc1 and PLCG1-N, that multiple classes of peptide ligands correlate to multiple physiological ligands. Other methods would either become hopelessly convoluted by such a pattern or only show the tighter interaction. Further, this work suggests that some of these SH2 domains may have evolved to bind to their ligands using multiple modes. This type of interaction is apparent in the Vav3 SH2 domain as well as the Blnk SH2 domain. There is currently no other data which could facilitate this hypothesis. In conclusion, the data generated from our studies should prove to be a valuable tool for researchers studying SH2 domain mediated cell signaling.

95

Bibliography

1. Burnett, G.; Kennedy, E. P., The enzymatic phosphorylation of proteins. The Journal of biological chemistry 1954, 211 (2), 969-80. 2. Levene PA, A. C., The cleavage products of vitellin. J. Biol. Chem 1906, 2 ((1)), 127–133. 3. Pawson, T.; Scott, J. D., Protein phosphorylation in signaling--50 years and counting. Trends Biochem Sci 2005, 30 (6), 286-90. 4. Hunter, T.; Pawson, T., The evolution of protein phosphorylation. Preface. Philos Trans R Soc Lond B Biol Sci 2012, 367 (1602), 2512. 5. Hubbard, S. R.; Till, J. H., Protein tyrosine kinase structure and function. Annu Rev Biochem 2000, 69, 373-98. 6. Nolen, B.; Taylor, S.; Ghosh, G., Regulation of protein kinases; controlling activity through activation segment conformation. Mol Cell 2004, 15 (5), 661-75. 7. Tsatsanis, C.; Spandidos, D. A., The role of oncogenic kinases in human cancer (Review). Int J Mol Med 2000, 5 (6), 583-90. 8. Tonks, N. K., Protein tyrosine phosphatases: from genes, to function, to disease. Nat Rev Mol Cell Biol 2006, 7 (11), 833-46. 9. Kaneko, T.; Joshi, R.; Feller, S. M.; Li, S. S., Phosphotyrosine recognition domains: the typical, the atypical and the versatile. Cell Commun Signal 2012, 10 (1), 32. 10. Sadowski, I.; Stone, J. C.; Pawson, T., A noncatalytic domain conserved among cytoplasmic protein-tyrosine kinases modifies the kinase function and transforming activity of Fujinami sarcoma virus P130gag-fps. Mol Cell Biol 1986, 6 (12), 4396-408. 11. Waksman, G.; Kominos, D.; Robertson, S. C.; Pant, N.; Baltimore, D.; Birge, R. B.; Cowburn, D.; Hanafusa, H.; Mayer, B. J.; Overduin, M.; Resh, M. D.; Rios, C. B.; Silverman, L.; Kuriyan, J., Crystal structure of the phosphotyrosine recognition domain SH2 of v-src complexed with tyrosine-phosphorylated peptides. Nature 1992, 358 (6388), 646-53. 12. Liu, B. A.; Engelmann, B. W.; Nash, P. D., The language of SH2 domain interactions defines phosphotyrosine-mediated signal transduction. FEBS Lett 2012, 586 (17), 2597-605. 13. Songyang, Z.; Shoelson, S. E.; Chaudhuri, M.; Gish, G.; Pawson, T.; Haser, W. G.; King, F.; Roberts, T.; Ratnofsky, S.; Lechleider, R. J.; et al., SH2 domains recognize specific phosphopeptide sequences. Cell 1993, 72 (5), 767-78. 14. Kaneko, T.; Huang, H.; Cao, X.; Li, X.; Li, C.; Voss, C.; Sidhu, S. S.; Li, S. S., Superbinder SH2 domains act as antagonists of cell signaling. Sci Signal 2012, 5 (243), ra68. 15. Kaneko, T.; Huang, H.; Zhao, B.; Li, L.; Liu, H.; Voss, C. K.; Wu, C.; Schiller, M. R.; Li, S. S., Loops govern SH2 domain specificity by controlling access to binding pockets. Sci Signal 2010, 3 (120), ra34. 16. Sweeney, M. C.; Wavreille, A. S.; Park, J.; Butchar, J. P.; Tridandapani, S.; Pei, D., Decoding protein-protein interactions through combinatorial chemistry: sequence specificity of SHP-1, SHP-2, and SHIP SH2 domains. Biochemistry 2005, 44 (45), 14932-47. 17. (a) Zhang, Y.; Zhang, J.; Yuan, C.; Hard, R. L.; Park, I. H.; Li, C.; Bell, C.; Pei, D., Simultaneous binding of two peptidyl ligands by a SRC homology 2 domain. Biochemistry 96

2011, 50 (35), 7637-46; (b) Narula, S. S.; Yuan, R. W.; Adams, S. E.; Green, O. M.; Green, J.; Philips, T. B.; Zydowsky, L. D.; Botfield, M. C.; Hatada, M.; Laird, E. R.; et al., Solution structure of the C-terminal SH2 domain of the human tyrosine kinase Syk complexed with a phosphotyrosine pentapeptide. Structure 1995, 3 (10), 1061-73; (c) Zhang, Y.; Wavreille, A. S.; Kunys, A. R.; Pei, D., The SH2 domains of inositol polyphosphate 5-phosphatases SHIP1 and SHIP2 have similar ligand specificity but different binding kinetics. Biochemistry 2009, 48 (46), 11075-83. 18. Rodriguez, M.; Li, S. S.; Harper, J. W.; Songyang, Z., An oriented peptide array library (OPAL) strategy to study protein-protein interactions. The Journal of biological chemistry 2004, 279 (10), 8802-7. 19. Kunys, A. R.; Lian, W.; Pei, D., Specificity profiling of protein-binding domains using one-bead-one-compound Peptide libraries. Curr Protoc Chem Biol 2012, 4 (4), 331-55. 20. Chen, X.; Tan, P. H.; Zhang, Y.; Pei, D., On-bead screening of combinatorial libraries: reduction of nonspecific binding by decreasing surface ligand density. J Comb Chem 2009, 11 (4), 604-11. 21. (a) Thakkar, A.; Cohen, A. S.; Connolly, M. D.; Zuckermann, R. N.; Pei, D., High- throughput sequencing of peptoids and peptide-peptoid hybrids by partial edman degradation and mass spectrometry. J Comb Chem 2009, 11 (2), 294-302; (b) Thakkar, A.; Wavreille, A. S.; Pei, D., Traceless capping agent for peptide sequencing by partial edman degradation and mass spectrometry. Anal Chem 2006, 78 (16), 5935-9. 22. Zhao, B.; Tan, P. H.; Li, S. S.; Pei, D., Systematic characterization of the specificity of the SH2 domains of cytoplasmic tyrosine kinases. J Proteomics 2013, 81, 56-69. 23. Dombroski, D.; Houghtling, R. A.; Labno, C. M.; Precht, P.; Takesono, A.; Caplen, N. J.; Billadeau, D. D.; Wange, R. L.; Burkhardt, J. K.; Schwartzberg, P. L., Kinase-independent functions for Itk in TCR-induced regulation of Vav and the actin cytoskeleton. J Immunol 2005, 174 (3), 1385-92. 24. Qiu, H.; Zappacosta, F.; Su, W.; Annan, R. S.; Miller, W. T., Interaction between Brk kinase and insulin receptor substrate-4. Oncogene 2005, 24 (36), 5656-64. 25. Derry, J. J.; Richard, S.; Valderrama Carvajal, H.; Ye, X.; Vasioukhin, V.; Cochrane, A. W.; Chen, T.; Tyner, A. L., Sik (BRK) phosphorylates Sam68 in the nucleus and negatively regulates its RNA binding ability. Mol Cell Biol 2000, 20 (16), 6114-26. 26. Craven, R. J.; Cance, W. G.; Liu, E. T., The nuclear tyrosine kinase Rak associates with the retinoblastoma protein pRb. Cancer Res 1995, 55 (18), 3969-72. 27. Vargas, L.; Nore, B. F.; Berglof, A.; Heinonen, J. E.; Mattsson, P. T.; Smith, C. I.; Mohamed, A. J., Functional interaction of caveolin-1 with Bruton's tyrosine kinase and Bmx. The Journal of biological chemistry 2002, 277 (11), 9351-7. 28. Doyle, S. L.; Jefferies, C. A.; Feighery, C.; O'Neill, L. A., Signaling by Toll-like receptors 8 and 9 requires Bruton's tyrosine kinase. The Journal of biological chemistry 2007, 282 (51), 36953-60. 29. Giubellino, A.; Burke, T. R., Jr.; Bottaro, D. P., Grb2 signaling in cell motility and cancer. Expert Opin Ther Targets 2008, 12 (8), 1021-33. 30. Zhang, W.; Trible, R. P.; Zhu, M.; Liu, S. K.; McGlade, C. J.; Samelson, L. E., Association of Grb2, Gads, and phospholipase C-gamma 1 with phosphorylated LAT tyrosine residues. Effect of LAT tyrosine mutations on T cell angigen receptor-mediated signaling. The Journal of biological chemistry 2000, 275 (30), 23355-61.

97

31. Feng, G. S.; Ouyang, Y. B.; Hu, D. P.; Shi, Z. Q.; Gentz, R.; Ni, J., Grap is a novel SH3-SH2- SH3 adaptor protein that couples tyrosine kinases to the Ras pathway. The Journal of biological chemistry 1996, 271 (21), 12129-32. 32. Zhang, W.; Sloan-Lancaster, J.; Kitchen, J.; Trible, R. P.; Samelson, L. E., LAT: the ZAP-70 tyrosine kinase substrate that links T cell receptor to cellular activation. Cell 1998, 92 (1), 83- 92. 33. Audero, E.; Cascone, I.; Maniero, F.; Napione, L.; Arese, M.; Lanfrancone, L.; Bussolino, F., Adaptor ShcA protein binds tyrosine kinase Tie2 receptor and regulates migration and sprouting but not survival of endothelial cells. The Journal of biological chemistry 2004, 279 (13), 13224-33. 34. Fagiani, E.; Giardina, G.; Luzi, L.; Cesaroni, M.; Quarto, M.; Capra, M.; Germano, G.; Bono, M.; Capillo, M.; Pelicci, P.; Lanfrancone, L., RaLP, a new member of the Src homology and collagen family, regulates cell migration and tumor growth of metastatic melanomas. Cancer Res 2007, 67 (7), 3064-73. 35. Ichikawa, O.; Osawa, M.; Nishida, N.; Goshima, N.; Nomura, N.; Shimada, I., Structural basis of the collagen-binding mode of discoidin domain receptor 2. Embo J 2007, 26 (18), 4168- 76. 36. Degoutin, J.; Vigny, M.; Gouzi, J. Y., ALK activation induces Shc and FRS2 recruitment: Signaling and phenotypic outcomes in PC12 cells differentiation. FEBS Lett 2007, 581 (4), 727- 34. 37. Vindis, C.; Cerretti, D. P.; Daniel, T. O.; Huynh-Do, U., EphB1 recruits c-Src and p52Shc to activate MAPK/ERK and promote chemotaxis. J Cell Biol 2003, 162 (4), 661-71. 38. Wang, J. F.; Zhang, X.; Groopman, J. E., Activation of vascular endothelial growth factor receptor-3 and its downstream signaling promote cell survival under oxidative stress. The Journal of biological chemistry 2004, 279 (26), 27088-97. 39. Ronnstrand, L., Signal transduction via the stem cell factor receptor/c-Kit. Cell Mol Life Sci 2004, 61 (19-20), 2535-48. 40. Liu, H. Y.; Meakin, S. O., ShcB and ShcC activation by the Trk family of receptor tyrosine kinases. The Journal of biological chemistry 2002, 277 (29), 26046-56. 41. Hunter, S. G.; Zhuang, G.; Brantley-Sieders, D.; Swat, W.; Cowan, C. W.; Chen, J., Essential role of Vav family guanine nucleotide exchange factors in EphA receptor-mediated angiogenesis. Mol Cell Biol 2006, 26 (13), 4830-42. 42. Jones, R. B.; Gordus, A.; Krall, J. A.; MacBeath, G., A quantitative protein interaction network for the ErbB receptors using protein microarrays. Nature 2006, 439 (7073), 168-74. 43. Reddi, A. L.; Ying, G.; Duan, L.; Chen, G.; Dimri, M.; Douillard, P.; Druker, B. J.; Naramura, M.; Band, V.; Band, H., Binding of Cbl to a phospholipase Cgamma1-docking site on platelet-derived growth factor receptor beta provides a dual mechanism of negative regulation. The Journal of biological chemistry 2007, 282 (40), 29336-47. 44. Mohammadi, M.; Honegger, A. M.; Rotin, D.; Fischer, R.; Bellot, F.; Li, W.; Dionne, C. A.; Jaye, M.; Rubinstein, M.; Schlessinger, J., A tyrosine-phosphorylated carboxy-terminal peptide of the fibroblast growth factor receptor (Flg) is a binding site for the SH2 domain of phospholipase C-gamma 1. Mol Cell Biol 1991, 11 (10), 5068-78. 45. Wang, Y.; Tomar, A.; George, S. P.; Khurana, S., Obligatory role for phospholipase C- gamma(1) in villin-induced epithelial cell migration. Am J Physiol Cell Physiol 2007, 292 (5), C1775-86.

98

46. Kulathu, Y.; Hobeika, E.; Turchinovich, G.; Reth, M., The kinase Syk as an adaptor controlling sustained calcium signalling and B-cell development. Embo J 2008, 27 (9), 1333-44. 47. Taguchi, T.; Kiyokawa, N.; Takenouch, H.; Matsui, J.; Tang, W. R.; Nakajima, H.; Suzuki, K.; Shiozawa, Y.; Saito, M.; Katagiri, Y. U.; Takahashi, T.; Karasuyama, H.; Matsuo, Y.; Okita, H.; Fujimoto, J., Deficiency of BLNK hampers PLC-gamma2 phosphorylation and Ca2+ influx induced by the pre-B-cell receptor in human pre-B cells. Immunology 2004, 112 (4), 575-82. 48. Draber, P.; Vonkova, I.; Stepanek, O.; Hrdinka, M.; Kucova, M.; Skopcova, T.; Otahal, P.; Angelisova, P.; Horejsi, V.; Yeung, M.; Weiss, A.; Brdicka, T., SCIMP, a transmembrane adaptor protein involved in major histocompatibility complex class II signaling. Mol Cell Biol 2011, 31 (22), 4550-62. 49. Holland, S. J.; Liao, X. C.; Mendenhall, M. K.; Zhou, X.; Pardo, J.; Chu, P.; Spencer, C.; Fu, A.; Sheng, N.; Yu, P.; Pali, E.; Nagin, A.; Shen, M.; Yu, S.; Chan, E.; Wu, X.; Li, C.; Woisetschlager, M.; Aversa, G.; Kolbinger, F.; Bennett, M. K.; Molineaux, S.; Luo, Y.; Payan, D. G.; Mancebo, H. S.; Wu, J., Functional cloning of Src-like adapter protein-2 (SLAP-2), a novel inhibitor of antigen receptor signaling. J Exp Med 2001, 194 (9), 1263-76.

99