UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Identifying the domains involved in the regulation of RasGRP1 in response to BCR ligation Beaulieu, Nadine 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2007-266786.pdf [ 12.44MB ]
Metadata
JSON: 831-1.0100321.json
JSON-LD: 831-1.0100321-ld.json
RDF/XML (Pretty): 831-1.0100321-rdf.xml
RDF/JSON: 831-1.0100321-rdf.json
Turtle: 831-1.0100321-turtle.txt
N-Triples: 831-1.0100321-rdf-ntriples.txt
Original Record: 831-1.0100321-source.json
Full Text
831-1.0100321-fulltext.txt
Citation
831-1.0100321.ris

Full Text

IDENTIFYING THE DOMAINS INVOLVED IN T H E REGULATION OF R A S G R P I IN RESPONSE TO B C R LIGATION by NADINE BEAULIEU B.Sc, McGill University, 1999 A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F PHILOSOPHY in T H E F A C U L T Y O F G R A D U A T E STUDIES (Genetics) The University of British Columbia October 2006 © Nadine Beaulieu, 2006 ABSTRACT Guanine nucleotide exchange factors (GEFs) are proteins that act as activators for the Ras family of small GTPases. In order to become activated, they typically need to localize in membrane structures, where their substrates are located. RasGRPI is a guanine nucleotide exchange factor whose activation is associated with translocation to membranes. Previous studies have indicated that translocation of RasGRPI is driven solely by binding o f its C I domain to diacylglycerol ( D A G ) generated in membranes following receptor-coupled phospholipase C ' s activation. In DT40 B cells, ligation o f the B cell receptor induces RasGRPI activation, and translocation of RasGRPI occurs exclusively to the plasma membrane. M y thesis focused on identifying the domains involved in RasGRPI translocation and activation. M y results show that the C I domain o f RasGRPI is essential for efficient BCR-induced plasma membrane translocation and activation, but on its own the C I domain is unable to drive translocation to the plasma membrane. I identified a domain of RasGRPI , termed the K W E N loop, which acts in conjunction with an adjacent leucine zipper (LZ) to confer BCR-induced translocation to the plasma membrane, either autonomously or in cooperation with the C I domain. In addition, a newly identified repressor region was shown to dampen KWEN/LZ-media ted translocation and activation. I propose a revised model for RasGRP 1 translocation and activation: plasma membrane localization in response to B C R ligation is provided by the K W E N loop and L Z . Interaction between D A G and the C I domain enhances K W E N -mediated binding to the plasma membrane, overriding the suppressive effect of the repressor region and enabling DAG-dependent activation of R a s G R P I . TABLE OF CONTENTS A B S T R A C T . . ii T A B L E OF C O N T E N T S iii LIST O F FIGURES vii LIST O F T A B L E S x LIST OF ABBREVIATIONS xi A C K N O W L E D G E M E N T S xiii DEDICATION xiv 1.1 Ras family of GTPases 1 1.2 Ras in lymphocyte development and differentiation 2 1.2.1 Ras involvement in thymocyte development 2 1.2.2 Ras involvement in B lymphocyte development 6 1.3 Regulation of Ras GTPases by G E F s 9 1.3.1 Three different families of RasGEFs 10 1.4 The RasGRP family of proteins 12 1.4.1 Four members of the RasGRP family 12 1.4.2 The role of RasGRPI in lymphocyte development 14 1.5 Regulation of RasGRPI following antigen receptor activation 18 1.5.1 Structure of RasGRPI 19 1.5.2 DAG and C1 domains as signal transducers 20 1.5.3 Regulation of RasGRPI by its C1 domain 22 1.5.4 Potential roles of other domains in the regulation of RasGRPI 30 1.6 Thesis Objectives 33 C H A P T E R 2: MATERIALS AND METHODS 35 C H A P T E R 3: MECHANISMS OF RASGRP1 TRANSLOCATION T O THE P L A S M A M E M B R A N E IN R E S P O N S E TO BCR LIGATION 40 3.1 Experimental System 40 3.1.1 DT40 B cells as a model system 40 3.1.2 Generation of DT40 cells expressing GFP-tagged RasGRPI 40 3.2 Cellular Localization of RasGRPI in DT40 cells 44 3.2.1 RasGRPI translocates to the plasma membrane in response to BCR ligation 44 3.2.2 RasGRPI translocation can be induced by DAG or phorbol esters. 45 3.3 The C1-DAG interaction is not solely responsible for RasGRPI translocation to the plasma membrane in response to B C R ligation 48 3.3.1 The C1 domain of RasGRPI does not localize to the plasma membrane in response to BCR ligation 48 3.3.2 A C1 domain/DAG interaction is necessary for optimum translocation of RasGRPI to the plasma membrane following BCR ligation 50 3.3.3 The C1 -domain provides specificity for DAG binding 53 3.3.4 Revision of the model for BCR-mediated translocation of RasGRPI 54 3.4 The C-terminal region of RasGRPI is responsible for mediating C1-cooperative and C1-independent mechanisms of plasma membrane translocation in response to BCR ligation 56 3.4.1 The region C-terminal to the C1 domain is essential for BCR-mediated translocation of RasGRI to the plasma membrane 56 iv 3.4.2 The Sp-LZ region confers plasma membrane translocation in response to BCR ligation via a PLCy2-independent mechanism 58 3.4.3 The C1 domain cooperates with the Sp-LZ region to increase the efficiency of BCR-induced translocation to the plasma membrane 60 3.4.4 The Leucine Zipper motif contributes to the C1 domain-independent and C1 domain-cooperative translocation mechanisms of RasGRPI 61 3.4.5 The KWEN loop is responsible for C1 domain-independent and leucine zipper-cooperative BCR-mediated translocation of RasGRPI 65 3.5 Conclusion for the Mechanisms of RasGRPI Translocation to the Plasma Membrane in Response to BCR activation 72 C H A P T E R 4: MECHANISMS OF RASGRP1 ACTIVATION IN R E S P O N S E TO B C R LIGATION 73 4.1 RasGRPI Activation in DT40 cells 73 4.1.1 RasGRPI gets activated in response to BCR ligation 76 4.1.2 RasGRPI gets activated in response to DAG analogs 78 4.2 The C1 domain-DAG interaction is necessary for RasGRPI activation following BCR ligation 79 4.2.1 The C1 domain is required for BCR-induced activation of RasGRPI 79 4.2.2 RasGRPI does not get activated in the absence of DAG 81 4.2.3 Artificial targeting of RasGRPI to the plasma membrane can bypass the need for the C1 domain in BCR-mediated activation of RasGRPI 81 4.3 KWEN/LZ can mediate RasGRPI activation 83 4.3.1 KWEN/LZ is necessary for efficient C1/DAG-mediated activation of RasGRP I in response to BCR ligation 84 4.3.2 KWEN/LZ-driven translocation can confer BCR-mediated activation of RasGRP I in the absence o f theC I domain 85 4.4 The KWEN loop partially contributes to the activation of RasGRPI downstream of the BCR 86 4.5 The leucine zipper motif partially contributes to the activation of RasGRPI downstream of the BCR 87 4.6 Conclusions on the contribution of different domains of RasGRPI in plasma membrane translocation and activation 88 C H A P T E R 5: DISCUSSION 90 5.1 The Initial Model for the Regulation of RasGRPI 90 5.2 New Insights in the Mechanism of RasGRPI Translocation to the Plasma Membrane in Response to BCR Ligation 91 5.2.1 RasGRPI Localization in Resting DT40 Cells 91 5.2.2 RasGRPI Localization in Response to BCR Ligation 92 5.3 Implications for the Regulation of RasGRPI 96 5.3.1 Regulation of RasGRPI via a Single versus Multiple Signals 96 5.3.2 The Role of DAG in the Activation of RasGRPI 97 5.3.3 The Role of Membrane Localization in the Activation of RasGRPI . 98 5.4 Implications for the Regulation of other RasGRP Family Members. 99 BIBLIOGRAPHY 102 vi LIST OF FIGURES Figure 1.1 Schematic representation of the involvement of Ras in thymocyte development 5 Figure 1.2 Schematic representation of the involvement of Ras in B lymphocyte development 8 Figure 1.3 Schematic representation of the regulation of Ras GTPases 10 Figure 1.4 Schematic representation of the three different Ras G E F s families 11 Figure 1.5 Schematic representation of the murine RasGRP family members 13 Figure 1.6 Schematic representation of the involvement of RasGRPI in thymocyte development 17 Figure 1.7 Schematic representation of the interaction between the C I domain and D A G at the plasma membrane 21 Figure 1.8 Schematic representation o f t h e D A G - C l hypothesis 29 Figure 1.9 Schematic representation of leucine zipper dimerization 32 Figure 3.1 Structures of GFP-tagged RasGRPI proteins 42 Figure 3.2 Expression data of different RasGRPI protein mutants 43 Figure 3.3 Schematic representation of GFP-tagged protein localization 45 Figure 3.4 RasGRPI translocates to the plasma membrane in response to a-IgM and D A G 47 Figure 3.5 The C I domain is insufficient for B C R - or DAG-induced translocation to the plasma membrane 50 Figure 3.6 C I domain and D A G are required for efficient translocation of RasGRPI . . . . 51 Figure 3.7 Revised model for the mechanism of RasGRPI translocation to the plasma membrane in response to B C R ligation 55 Figure 3.8 The region C-terminal to the C I domain is essential for B C R and D A G -mediated translocation to the plasma membrane 57 Figure 3.9 The region C-terminal to the C I domain translocates to the plasma membrane in response to anti-IgM in a P L C y 2 - and DAG-independent manner 59 vi i Figure 3.10 The C-terminus construct containing the C I domain and the region C -terminal to it translocates to the plasma membrane in response to anti-IgM and P M A with high efficiency 61 Figure 3.11 Requirement of the leucine zipper for efficient translocation of RasGRPI to the plasma membrane 63 Figure 3.12 The leucine zipper is not responsible for but contributes to CI-independent translocation of RasGRPI 64 Figure 3.13 The spacer is responsible for CI-independent translocation of RasGRPI to the plasma membrane 66 Figure 3.14 Defining the region within the spacer which modulates BCR-induced translocation to the plasma membrane 68 Figure 3.15 Deletion of the repressor region enhances KWEN-mediated CI-independent translocation of RasGRPI in response to B C R ligation 70 Figure 3.16 The K W E N loop is necessary for B C R - and DAG-mediated translocation o f RasGRPI to the plasma membrane 71 Figure 4.1 Schematic representation of the activated Ras affinity precipitation 75 Figure 4.2 Representation of the M A P K pathway activated downstream of B C R 76 Figure 4.3 Increased activity of RasGRPI causes an increase in Ras and E R K activation in response to B C R ligation and D A G 78 Figure 4.4 The C I domain and D A G are required for efficient activation of R a s G R P I . 80 Figure 4.5 B C R ligation is required to activate RasGRPI even with artificial targeting o f RasGRPI to the plasma membrane 83 Figure 4.6 The region C-terminal to the C I domain is necessary for optimum activation of RasGRPI in response to B C R or D A G 85 Figure 4.7 Deletion of the repressor region enhances KWEN-mediated CI-independent activation of RasGRP 1 in response to B C R ligation 86 Figure 4.8 The K W E N loop is necessary for optimum activation of RasGRPI in response to B C R ligation, but is dispensable for a response to D A G treatment 87 Figure 4.9 The leucine zipper is necessary for optimum activation of RasGRPI in response to B C R ligation and D A G treatment ....88 via Figure 5.1 Model for the mechanism of RasGRPI translocation to the plasma membrane in response to B C R ligation 95 LIST OF TABLES Table 3.1 Proportion of cells demonstrating plasma membrane localization following B C R stimulation 52 LIST OF ABBREVIATIONS A I C D : activation-induced cell death Ave : average B C R : B cell receptor D A G : diacylglycerol D G K : D A G kinase D N : double negative D P : double positive E R K : extracellular signal-regulated kinase E T P : early T-cell progenitor G A P : GTPase activating protein G D P : guanine nucleotide diphosphate G E F : guanine nucleotide exchange factor <e)GFP: (enhanced) green fluorescent protein G S T : glutathione-s-transferase G T P : guanine nucleotide triphosphate GTPase: guanosine triphosphatase lg : immunoglobulin I M : internal or inner membranes IP3: inositol triphosphate K L : K W E N ligand L Z : leucine zipper M g : magnesium M y r : myristylation P K C : protein kinase C P L C : phospholipase C P M : plasma membrane P M A : phorbol 12-myristate 13-acetate Pren: prenylation R A G : recombinase-activating gene RasGRF: Ras guanosine nucleotide releasing factor RasGRP: Ras guanine nucleotide releasing protein RBD: Ras-binding domain R E M : Ras exchange motif RP: repressor region SP: single positive Sp: spacer region Sos: son-of-sevenless TBS-T: Tris-buffered saline containing Tween-20 TCR: T cell receptor ACKNOWLEDGEMENTS I would first like to thank my supervisor, Rob Kay, for the incredible support that you gave me all along my PhD. The completion of my thesis would not have been possible without your support. But beyond being a supervisor, you are a mentor to me. You have taught me so much; about science and teaching, about fairness, trust and respect, about balance, and about always being yourself. You have had a very positive influence on my life. Thank you. I also want to thank my husband, Colin. 1 truly believe that you make this world a better place; you make my world a much better one. Without your love and support, the completion of my thesis would not have been possible. It's funny how happiness can convert the biggest mountain into a hill... If only everyone had a Colin in their lives... Thank you. Et finalement, le plus grand des merci a mes parents, Roch et Suzanne Beaulieu, a qui je dedie ma these. Vous m'avez toujours entouree d'amour, vous m'avez toujours encouragee et supportee - merci. Le secret de ma reussite, c'est vous; sans votre amour et votre support, je ne serais pas la ou je suis. Je vous aime de tout mon coeur. Nadine xi i i Ames parents xiv Chap te r 1: Introduct ion In order to fulfill their role within multicellular organisms, cells need to be able to communicate with their local environment. In many instances, the transmission of signals from the outside to the inside of the cell, referred to as signal transduction, involves the engagement of a cell surface receptor. These receptors are capable of sensing a stimulus that occurs outside the cell, and transmit a signal to their intracellular machinery for a response. The engagement of a receptor subsequently leads to the activation of multiple signaling pathways, resulting in the induction of genes involved in diverse cellular responses. To understand the regulation of these cellular responses, it is important to identify the components of each of the pathways involved, and furthermore to identify the exact mechanism of regulation for each component. My thesis discusses the analysis of the regulation of one component of a signal transduction pathway initiated by the ligation of the B cell receptor (BCR) and leading to Ras activation. 1.1 R A S FAMILY OF G T P A S E S Ras proteins are molecular switches that exist in the cell in equilibrium between GDP- and GTP-bound forms [reviewed by (Colicelli 2004; Mor and Philips 2006)]. In the GTP-bound activated state, they display a binding surface with high affinity for downstream effector proteins. One of the well-characterized Ras downstream effector pathways is the Raf pathway, which involves a series of phosphorylation events that lead to the activation of the MAPK proteins ERK1 and ERK2. The Ras family of small GTPases is known to participate in the transduction of extracellular signals essential for cell growth, survival and differentiation. In mammals, the Ras family contains four highly related members: H-Ras, N-Ras, and K-RasA and B, which are generated by alternative splicing of the K-ras gene. The N-terminal part of the protein is 100% identical among the 4 Ras proteins, and comprises the guanine nucleotide binding region, the Ras downstream effector binding sequence and the region involved in binding a nucleotide-associated M g 2 + ion. The C-terminal region on the other hand is highly 1 variable, except for a conserved CAAX motif at the very C-terminal end, which directs posttranslational modification events, including prenylation, proteolysis and carboxyl methylesterification. These modifications result in a highly hydrophobic C-terminal region which is necessary to localize Ras proteins to membranes, where they can interact with their upstream activators and downstream targets. The type of posttranslational modification contributes to determining the engagement of different downstream targets, possibly due to the fact that the biochemical nature of the membrane anchor could dictate the efficiency of interaction with regulator molecules. Alternatively, the differential signaling could be due to distinctive sites of localization within the membrane, impacting the availability of regulators and effectors which might themselves be localized in distinct subdomains within various cell membranes. Other subgroups of the Ras family of GTPases include R-Ras (R-Rasl, TC21 and M-Ras), mainly involved in mitogenesis and cytoskeleton control, Rap (Rapl A, IB, 2A, 2B), involved in cell adhesion and cell spreading and Ral (RalA, B), involved in several processes including mitogenic responses, differentiation, protein trafficking and cytoskeleton dynamics. They share about 40-50% amino acid identity with the classical Rases, and their requirement for GAPs and GEFs overlaps with the classical Rases [reviewed by (Colicelli 2004; Mor and Philips 2006)]. 1.2 R A S IN LYMPHOCYTE DEVELOPMENT AND DIFFERENTIATION The proper development, differentiation and function of lymphocytes require Ras proteins to be appropriately activated following engagement of the B cell receptor (BCR) and T cell receptor (TCR), expressed in B and T lymphocytes respectively. My thesis focuses on one component of a pathway leading to Ras activation following BCR engagement, but I will also briefly review its role in T lymphocyte development. 1.2.1 Ras involvement in thymocyte development 2 Thymocyte development takes place in the thymus, where early T-cell progenitors (ETP) first receive signals for T cell lineage commitment [reviewed by (Bhandoola and Sambandam 2006; Takahama 2006)]. At the initial stage of development, thymocytes are referred to as double negative cells (DN), based on the absence of expression of the TCR coreceptors CD4 and CD8 (Figure 1.1a). One of the first important checkpoints in the development of thymocytes happens during the DN stage, where the TCRp locus undergoes its first gene rearrangement. Failure to express a functional pre-TCR leads to thymocyte deletion. By contrast, a productively rearranged TCRP will be expressed and pair with pTa and the CD3 complex, forming a functional pre-TCR, which drives the progression from the DN to the double positive stage (DP), where cells become CD4+CD8+. At the DP stage, further gene rearrangement events occur, leading to the expression of TCRa which associates with the pre-existing TCRp, forming the TCR. Whereas failure to express a functional TCR leads to the elimination of thymocytes, cells expressing a signaling-competent TCR will be facing three possible outcomes: 1) high-avidity interactions between the TCR and peptide-MHC ligands expressed in the thymus cortex will cause the cells to undergo negative selection, a process that serves to eliminate cells that present a TCR whose specificity is against a self-antigen; 2) a lack of interaction between the TCR and a ligand will subject the cells to death by neglect; 3) low-avidity interactions between the TCR and peptide-MHC ligands in the thymus cortex will lead to positive selection, where the cells migrate to the thymus medulla and mature into CD4+CD8" or CD4"CD8+ single positive (SP) thymocytes. Once in the medulla, SP thymocytes mature into naive T cells, which are competent to mount an immune response. Gene rearrangement at the TCRP and a loci is a process dependent on the recombinase-activating gene (RAG)l and RAG2, and thymocytes deficient in these genes have a complete arrest at the DN stage of development (Mombaerts, Iacomini et al. 1992; Shinkai, Rathbun et al. 1992). The involvement of Ras proteins in thymocyte development was demonstrated by several independent experiments (Figure 1.1). Mice transgenic for dominant negative forms of Raf or MEK1, two kinases activated downstream of Ras, show a block in thymocyte development at the DN stage (Crompton, Gilmour et al. 1996; O'Shea, 3 Crompton et al. 1996). Conversely, the DN block encountered by Rag2~'~ thymocytes can be overridden by expression of a transgene encoding a constitutively active form of Ras or a constitutively active Raf (Swat, Shinkai et al. 1996; Iritani, Alberola-Ila et al. 1999). Furthermore, the involvement of Ras-specific GEFs (Ras activators described in section 1.3) in thymocyte development was demonstrated when transgenic expression of a dominant-negative form of Ras (RasN17) caused a complete block in the DN to DP progression (Swan, Alberola-Ila et al. 1995). This Ras mutant sequesters GEFs by failing to release them after binding. The progression from DN to DP thymocyte is therefore dependent, at least in part, on the proper signaling of the Ras-Raf-MEKl pathway, and the activity of Ras-specific GEFs. 4 Figure 1.1 Schematic representation of the involvement of Ras in thymocyte development, a Bone Marrow ELP ETP Thymus CD4 CD8" Pre-1 TCR T R A G A Mice "wild Type Mice + Dominant Negative Ras or M E K or Rat CD4" CD8 + y CD4 + TCR ( S P ) C D 8 + ^ <3 CD4 >^ CD8" y R A G Mice + constitutively active Ras orRaf Figure 1.1. Schematic representation of the involvement of Ras in thymocyte development. Representation of thymocyte development in wild type (A) and Rag-deficient (B) mice. Different developmental stages are indicated inside the appropriate circles representing cells. Expression of specific proteins are indicated above each cell. A block in development is represented by a black "T" bar. Bypass of a developmental block is represented by a lined arrow overriding the "T" bar . Cells present following the bypass of a developmental block are represented by dotted grey cells. ELP, early lymphoid progenitor; ETP, early T-cell progenitor; DN, double-negative thymocytes; DP, double-positive thymocytes; SP, single-positive thymocytes; TCR, T-cell receptor. 5 1.2.2 Ras involvement in B lymphocyte development As for thymocytes, the developmental pattern of B lymphocytes is defined by the proper rearrangement of the antigen receptor genes, leading to expression of a functional B cell receptor containing IgH and IgL chains [reviewed by (Ollila and Vihinen 2005; Hagman and Lukin 2006; Pelanda and Torres 2006)] (Figure 1.2a). The first major checkpoint encountered is when pro-B cells undergo rearrangement at the lg heavy chain (HC) locus. While failure to express a functionally rearranged p HC results in cell deletion, a successful rearrangement results in the expression of a functional p HC, which associates with the surrogate light chain (LC) proteins (Vp r eB and A,5) to form the pre-B cell receptor (pre-BCR). Expression of the pre-BCR leads to the differentiation of pro-B cells to pre-B cells, in which the rearrangement at the LC locus takes place. Once again, failure to successfully rearrange the LC locus results in cell deletion. In contrast, a functionally rearranged LC pairs up with the pre-existing HC to form the BCR, and its expression can lead to different outcomes. On the one hand, a cell expressing a receptor with specificity for a self-antigen will either: 1) undergo receptor editing, wherein secondary V(D)J recombination events take place with the possibility of expressing a novel non-autoreactive antigen receptor, or 2) will turn anergic, where the cell becomes non-reactive to antigens. Self-reactivity can also lead to cell deletion. On the other hand, cells expressing a functional BCR with no specificity against self-antigens will be positively selected and differentiate from pre-B cells into immature B cells. Because of their failure to rearrange immunoglobulin genes, Rag2_/" mice also present a defect in B lymphocyte development, with a block at the pro-B cell stage (Mombaerts, Iacomini et al. 1992; Shinkai, Rathbun et al. 1992). In addition to its role in thymocyte development, the proper function of Ras has proven critical for B lymphogenesis (Figure 1.2). In Rag2"/"mice, expression of a constitutively active form of Ras (Shaw, Swat et al. 1999) or a constitutively active Raf (Iritani, Alberola-Ila et al. 1999) bypasses the need for the pre-BCR and allows progression beyond the pro-B cell stage. The involvement of Ras-specific GEFs was also demonstrated in B lymphocyte development; expression of the dominant negative 6 RasN17 was shown to cause a block at the pro-B cell stage, which could be over ridden by activated Raf (Iritani, Forbush et al. 1997). Thus, as is the case for thymocytes, the development of B lymphocytes is dependent, at least in part, on the proper function of Ras and Raf, as well as Ras-activating GEFs. A tightly controlled regulation of Ras proteins is therefore critical for the proper development and function of lymphocytes. 7 Figure 1.2 Schematic representation of the involvement of Ras in B lymphocyte development. Bone Marrow Pre-B C R (Pre-Bj B C R Periphery B C R R A G ' Mice Wild Type Mice + Dominant Negative Ras Bone Marrow CLP Pro-B Pre-B .'Immature '>. B cell/ Periphery .' Mature-B cell R A G ' Mice + constitutively active Ras or Raf Wild Type Mice + Dominant Negative Ras + constitutively active Raf Figure 1.2. Schematic representation of the involvement of Ras in B lymphocyte development. Representation of B lymphocyte development in wild type (A) and Rag-deficient (B) mice. Different developmental stages are indicated inside the appropriate circles representing cells. Expression of specific proteins are indicated above each cell. A block in development is represented by a black "T" bar. Bypass of a developmental block is represented by a lines arrow overriding the "T" bar. Cells present following the bypass of a developmental block are represented by dotted grey cells. CLP, common lymphoid progenitor; BCR, B-cell receptor. 8 1.3 REGULATION OF RAS GTPASES BY G E F S Ras proteins are generally activated following the engagement of cell surface receptors, and a tight control of their activation state is crucial to maintain normal cellular functions [reviewed by (Paduch, Jelen et al. 2001)] (Figure 1.3). Upon receptor activation, the GDP molecule bound to Ras is displaced by members of the guanine nucleotide exchange factors (GEF) family. GEF proteins interact with the nucleotide/Mg2+-binding region of Ras, and insert residues (Leu and Glu for Sosl) at the binding site creating structural changes that are inhibitory for the binding of phosphates and Mg . The nucleotide-free complex can then bind a nucleotide molecule, predominantly GTP because of its higher concentration in the cell. Ras proteins contain an intrinsic GTPase activity that is typically low. The presence of GTPase activating proteins (GAP) greatly enhances the hydrolysis of GTP by Ras, contributing to the negative regulation of Ras-induced signal transduction. GAP proteins facilitate the hydrolysis of GTP by supplying a so-called arginine finger that acts as a catalytic residue, as well as by stabilizing the GDP/Ras complex through interaction with the nucleotide/Mg2+-binding region of Ras. 9 Figure 1.3 Schematic representation of the regulation of Ras GTPases. GAP Inactive Ras Active Ras j GEF Figure 1.3. Schematic representation of the regulation of Ras GTPases. GEF, catalytic domain; GAP, GTPase activating protein; GDP, guanine nucleotide diphosphate; GTP, guanine nucleotide triphosphate. 1.3.1 Three different families of R a s G E F s All guanine exchange factor proteins acting on Ras or Ras-related GTPases share in common a catalytic domain homologous to the exchange domain of Cdc25, the first GEF identified in Saccharomyces cerevisiae (Broek, Toda et al. 1987). There are three different families of GEFs that are known to catalyze the GDP/GTP exchange for Ras GTPases: Son-of-sevenless proteins (Sosl and Sos2), Ras guanosine nucleotide releasing factors (RasGRFl and RasGRF2), and Ras guanine nucleotide releasing proteins (RasGRPI through 4) (Quilliam, Rebhun et al. 2002) (Figure 1.4). The Sos family of GEFs typically gets activated following the engagement of protein tyrosine kinase receptors. Upon activation, receptor phosphorylation events take place, providing docking sites for the adaptor protein Grb2, which also binds Sos. Sos proteins are thereby recruited to the plasma membrane where they can activate Ras. The RasGRF proteins are also activated via recruitment to the plasma membrane, where they can activate Ras. They are regulated by calcium, and several of their domains have to cooperate for 10 calcium responses, membrane translocation and protein activation. In addition, RasGRF activity is further controlled by phosphorylation. RasGRPs are expressed in lymphocytes, and can be activated by antigen receptors (Ebinu, Stang et al. 2000; Bivona, Perez De Castro et al. 2003; Caloca, Zugaza et al. 2003; Sanjuan, Pradet-Balade et al. 2003; Aiba , Oh-hora et al. 2004; Ehrhardt, David et al. 2004; Guilbault and Kay 2004; Perez de Castro, Bivona et al. 2004). L ike the Sos and G R F families, the regulation of RasGRP family of proteins involves membrane translocation. This family of GEFs , particularly R a s G R P I , w i l l be discussed in detail in later sections. Figure 1.4 Schematic representation of the three different Ras GEFs families. S o s DH PH PxxP. R a s G R F - f PH CcllQ DH PH PxxP-R a s G R P Figure 1.4. Schematic representation of the three different Ras GEFs families. GEF, guanine exchange factor catalytic domain; DH; Dbl homology domain; PH, pleckstrin homology domain; PxxP, proline-rich region; CC, coiled-coil region; IQ, calmodulin binding domain; EF, calcium-binding EF hands; CI, DAG-binding CI domain. 11 1.4 THE R A S G R P FAMILY OF PROTEINS 1.4.1 Four members of the R a s G R P family The RasGRP family of proteins is composed of 4 members: RasGRPI through 4 (Figure 1.5). Several reports have described this family of GEFs as being predominantly expressed in brain and hematopoietic tissues, although expression of different family members has been shown in other tissues (Yamashita, Mochizuki et al. 2000; Dupuy, Morgan et al. 2001; Li, Yang et al. 2002). RasGRPI and RasGRP4 are activators for Ras proteins exclusively, with RasGRPI having shown specificity for H-, N-, K-, R-, M-Ras and TC21 (Ebinu, Bottorff et al. 1998; Kawasaki, Springett et al. 1998; Tognon, Kirk et al. 1998; Ohba, Mochizuki et al. 2000) and RasGRP4 having shown specificity for H-Ras (Reuther, Lambert et al. 2002; Yang, Li et al. 2002). RasGRP3 has exchange activity for H-Ras, R-Ras, M-Ras, TC21, RaplA and Rap2A (Ohba, Mochizuki et al. 2000; Yamashita, Mochizuki et al. 2000), and RasGRP2 mainly catalyzes GTP exchange for Rap (Kawasaki, Springett et al. 1998), although exchange activity for N-, K-, R-Ras and TC21 has also been reported (Clyde-Smith, Silins et al. 2000; Ohba, Mochizuki et al. 2000; Yamashita, Mochizuki et al. 2000). The highest similarity exists between RasGRP 1 and RasGRP3 proteins, but the four family members share several common domains: a Ras exchange motif, a GEF catalytic domain, calcium-binding EF hands and a DAG-binding CI domain. There are however several features differing among the RasGRP family members. The C-terminus of RasGRPI protein contains a leucine-zipper motif that is not found in any other family member. The N-terminus of RasGRP3 reveals a sequence that could potentially subject the protein to lipid modifications and thus membrane targeting. In contrast to other RasGRPs, the CI domain of RasGRP2 does not bind phorbol esters and DAG. RasGRP4 differs from the other family members in that it only has one EF hand, and it may be incapable of binding calcium. Furthermore, there exist at least 3 different splice variants of mouse RasGRP4, one of which contains an insertion of 5 amino acids in the CI domain that prevents DAG binding, and another one 12 which does not contain any CI domain at all. The role and characteristics of the RasGRPI domains will be treated in details in subsequent sections. Figure 1.5 Schematic representation of the murine RasGRP family members. R a s G R P I R E M T184 _GEF_ C a 2 + t DAG t Ras J-Z, R a s G R P 2 R a s G R P 3 R E M Myr R E M R a s G R P 4 { R E M h R E M Rap DAG t GEFW _j—gJg—CZ C1 Z> D o e . D o n Ras, Rap GEF Ras 3 E-<£lt> 5 a.a. insertion Figure 1.5. Schematic representation of the murine RasGRP family members. REM, Ras exchange motif; GEF, guanine exchange factor catalytic domain; EF, EF hand motif; CI, CI domain; LZ, leucine zipper motif; DAG, diacylglycerol; Myr, potential myristylation signal; a.a., amino acid. T184 shows the Tyrosine phosphorylation site. 1 3 1.4.2 The role of R a s G R P I in lymphocyte development 1.4.2.1 RasGRPI-deficient lymphocytes The involvement of RasGRPI in thymocyte development was first established by studying RasGRPI-deficient mice (Dower, Stang et al. 2000; Priatel, Teh et al. 2002). These mice presented a defect in thymocyte development and maturation, with substantially fewer single positive cells in the RasGRPI7" mice compared to RasGRP 1+ / + mice; that is, a 90% reduction in the CD4+CD8" population, and a 76% reduction in the CD4"CD8+ population. Positive selection was impaired in RasGRPI-deficient mice, as demonstrated by Priatel and colleagues (Priatel, Teh et al. 2002). In this study, the absence of RasGRPI greatly affected the selection of T cells expressing the HY TCR, a TCR that has a low affinity for the male-specific HY antigen, with a 4-fold reduction in CD8+ cells and a delayed maturation of the positively selected cells. On the other hand, the selection of T cells expressing the TCR 2C, which harbours high affinity for its ligand, was not significantly affected by the absence of RasGRPI. RasGRP 1-deficient T cells were defective in Ras and ERK activation and proliferation in response to TCR activation or to treatment with DAG analogs. This demonstrated the dependence on RasGRPI for TCR-induced Ras activation and cell proliferation. The fact that some CD8 SP cells were being produced in RasGRPI-deficient mice indicated a RasGRP 1-independent mechanism in positive selection. The selected CD8 SP cells in the absence of RasGRPI were suspected to represent high-affinity TCR expressing cells, since strongly activating TCR were less dependent on RasGRPI. RasGRPI-deficient mice also exhibited lymphoproliferative autoimmune syndrome with features of systemic lupus erythematosus (Layer, Lin et al. 2003). The lymphoproliferative disorder is thought to arise from the selection of high-affinity TCR-expressing T cells which are defective in activation-induced cell death (AICD). In contrast to T cells, B lymphocytes in RasGRP 1 -deficient mice displayed numbers of mature B cells elevated 2- to 3-fold, albeit without any apparent defect in B 14 lymphocyte development, activation or apoptosis (Dower, Stang et al. 2000; Layer, Lin et al. 2003). The accumulation of B cells may be due to the ability of activated T cells to kill autoreactive B lymphocytes, leading to the autoimmunity observed in these mice. The role of RasGRPI seems therefore more prominent in T cell than in B cell development and selection. Coughlin et al. however demonstrated that RasGRPI7" /RasGRP3_/" splenic B cells were defective in proliferation, whereas RasGRPI"7" and RasGRP3"7" splenic B cells only showed partial defects in proliferation (Coughlin, Stang et al. 2005). The function of RasGRPI in B cell development and activation is therefore likely to present some redundancy with the function of RasGRP3. 1.4.2.2 RasGRPI-transgenic mice Bevan's group along with our lab (Norment, Bogatzki et al. 2003) determined that RasGRPI was expressed at very low levels in DN cells, and was up-regulated when DN cells underwent differentiation induced by pre-TCR signaling in wild-type animals. Cross-linking of the TCR co-receptor CD3 in Rag2"/_ mice mimicked pre-TCR signaling, and also induced DN differentiation and up-regulation of RasGRPI expression (Figure 1.6a). DN differentiation and RasGRPI up-regulation were accompanied by an increase in ERK activation. In this study, overexpression of RasGRPI by a transgenic cDNA revealed a role for RasGRPI in T lymphocytes. When RasGRPl tg and Rag2_/" mice were crossed, the DN block imposed by the failure to form the TCRp component of the pre-TCR in Rag2/_ mice was overridden, which led to the production of DP cells in RasGRP 1 tg/Rag2/~ mice (Figure 1.6b). These results demonstrated that expression of RasGRPI is induced by pre-TCR signaling, and increased levels of RasGRPI can compensate for the need of a pre-TCR in the differentiation of DN to DP thymocytes. Increased levels of RasGRPI in wild-type mice induced maturation of DN cells and caused expansion and hyperresponsiveness of CD8 SP cells, despite their low TCR levels, which would normally not allow the progression of positive selection. RasGRPI might therefore lower the threshold of TCR signaling necessary for maturation and 15 expansion of thymocytes. Using a different line of RasGRP l t g mice, our lab established that deregulated and elevated levels of RasGRPI contributed to the initiation of T lymphomagenesis (Klinger, Guilbault et al. 2005). Evidence from this study demonstrated that RasGRPI initiated lymphomas prior to positive selection and independently of the presence of pre-TCR or TCR expression. This was first achieved by demonstrating the induction of lymphomas in Rag2-/" mice, and second, by showing the presence of lymphomas too immature to have undergone positive selection in RasGRP l t g/HY-TCR positively selecting females and in negatively selecting males. Other studies have linked deregulated expression of RasGRPI to lymphomagenesis: bioinformatic analysis demonstrated that most T acute lymphoblastic leukemias, which originate predominantly from transformed thymocytes, harbour high levels of RasGRPI expression (Yeoh, Ross et al. 2002; Asnafi, Beldjord et al. 2003). Moreover, the vicinity of the RasGRPI gene has been identified as a common site for proviral integrations leading to mouse thymic lymphomas (Kim, Trubetskoy et al. 2003). RasGRPI transgenic mice display reduced levels of pre-B cells, though this defect does not alter the number of peripheral B cells (B. Guilbault and R. Kay, unpublished results). Once again, it is apparent that the role of RasGRPI downstream of the TCR is different from its role in B lymphocytes. Given that RasGRPI gets activated downstream of the BCR, it is intriguing to find out what its function might be. 16 Figure 1.6 Schematic representation of the involvement of RasGRPI in thymocyte development, a R A G Mice + Cross-linking with anti-CD3 R A G 7 / RasGRPI^ mice Figure 1.6. Schematic representation of the involvement of RasGRP 1 in thymocyte development. Representation of thymocyte development in Rag-deficient mice that have been injected with anti-CD3 antibodies (A) or crossed with RasGRP 1-trangenic mice (B). Different developmental stages are indicated inside the appropriate circles representing cells. The expression level of RasGRP 1 at the different developmental stages is indicated above each cell in (A). A block in development is represented by a grey "T" bar. Bypass of a developmental block is represented by a lined arrow overriding the "T" bar . Cells present following the bypass of a developmental block are represented by dotted grey cells. DN, double-negative thymocytes; DP, double-positive thymocytes; tg; transgene. 17 1.4.2.3 Conclusion on the role of RasGRPI in lymphocyte development The fate of a thymocyte depends on the strength of signaling transmitted through the TCR complex. Such strength is interpreted by proteins downstream of the TCR, including RasGRPI and Ras, which take part in the transmission of the signal and direct the cell towards survival and differentiation, or death. The proper regulation of RasGRPI expression and activity towards its substrate Ras is therefore crucial for the development of T lymphocytes. In fact, deregulated expression of RasGRPI has been linked to T lymphomagenesis (Yeoh, Ross et al. 2002; Asnafi, Beldjord et al. 2003). Many signaling events downstream of the TCR share similarities with BCR signaling, including the activation of Ras, which requires PLCy activation following receptor ligation. It is however apparent that the signaling events determining T cell and B cell fate are different. In contrast to the developmental block in T cells, the absence of RasGRPI in mice causes elevated levels of B lymphocytes, with minor activation defect. Conversely, whereas RasGRPI transgenic mice develop thymic lymphomas, they have reduced levels of pre-B cells, with no defect in peripheral B cell numbers (Guilbault and Kay, unpublished results). It is therefore possible that the role of RasGRPI downstream of the BCR is different from its role in TCR signaling. In fact, RasGRP3 may play a more potent role in the development of B lymphocytes, as suggested by findings from Kurosaki's group (Oh-hora, Johmura et al. 2003). 1.5 REGULATION OF RASGRPI FOLLOWING ANTIGEN RECEPTOR ACTIVATION Studies of mice deficient in or overexpressing RasGRPI have clearly demonstrated the crucial role for RasGRPI in the development of T lymphocytes and lymphomagenesis. Earlier, studies in cell lines had revealed that RasGRPI was activated downstream of the TCR; Stone's group was the first one to demonstrate that TCR-induced Ras activation in the Jurkat T cell line was blocked by treating the cell lysates with an antibody against RasGRPI (Ebinu, Stang et al. 2000). This study also showed 18 that overexpression of RasGRPI in Jurkat cells caused an increase in Ras and ERK activation in response to TCR stimulation. Besides being activated downstream of the TCR, RasGRPI has also been shown to be activated downstream of the BCR in the DT40 and WEHI 231 murine immature B cell lines (Caloca, Zugaza et al. 2003; Guilbault and Kay 2004). Overexpression of RasGRPI in DT40 cells resulted in an increase in Ras and ERK2 activation (DT40 cells only express ERK2) following BCR ligation. Oh-hora et al. subsequently showed that BCR-mediated Ras and ERK activation were almost completely abolished in RasGRP3/RasGRPl double-deficient DT40s (Oh-hora, Johmura et al. 2003). RasGRP3 plus RasGRPI are therefore required for BCR-mediated Ras activation in DT40 cells. The involvement of RasGRPI in the activation of Ras downstream of the BCR was also demonstrated in a study using the WEHI 231 immature B cell line. This cell line is a model system for the regulatory mechanisms relevant to negative selection of primary B cells, and prolonged ligation of the BCR in these cells causes cell cycle arrest and apoptosis. Guilbault and Kay (Guilbault and Kay 2004) showed that expression of RasGRPI in these cells led to hypersensitization to BCR-induced Ras activation and apoptosis, and expression of a catalytically inactive, potentially dominant-negative form of RasGRPI reduced BCR-mediated apoptosis. My research interests were focused on determining the mechanism of RasGRPI activation in response to BCR ligation. The first requirement was to examine the structure of the RasGRPI protein and evaluate the potential role of different domains in the function of the protein. 1.5.1 Structure of R a s G R P I The first step in understanding the mechanism of RasGRP 1 activation following BCR ligation was to examine the structure of the protein and identify domains that could be involved in this process. RasGRPI is a 795 amino-acid protein with a molecular weight of-90 kDa (Ebinu, Bottorff et al. 1998; Tognon, Kirk et al. 1998); the structure of the protein is illustrated in Figure 1.5. The N-terminal region contains the REM motif typical of GEFs which interact with Ras and its closest relatives. Adjacent to the REM motif is 19 the catalytic GEF domain, immediately followed by a phosphorylation site at residue T184 (Zheng, Liu et al. 2005). The C-terminal portion of the protein covers the regulatory region, which classified RasGRPI as a third type of mammalian Ras GEF in the CDC25 family. It includes a pair of calcium-binding EF-hands, which by binding calcium can lead to a change in protein conformation, a DAG-binding CI domain, which by binding DAG can lead to membrane localization, and a leucine zipper motif that can potentially mediate protein-protein interactions. 1.5.2 D A G and C1 domains as signal t ransducers The enzyme phospholipase C (PLC) is typically coupled to tyrosine kinase receptors, and gets recruited to membranes and activated upon receptor engagement. It has for substrate phosphatidyl-4,5-bisphosphate, which PLC hydrolyses to DAG and inositol-1,4,5-triphosphate (IP3). These second messengers initiate two events: IP3 binds to intracellular receptors located on the endoplasmic reticulum, which causes the release of intracellular calcium stores, while DAG recruits to membranes and often activates signaling proteins that contain CI domains. CI domains were first identified in protein kinase C (PKC) and are defined as cysteine-rich sequences consisting of 50 conserved amino acids bearing the motif HXn_ 1 2 C X 2 C X 1 2 . 1 4 C X 2 C X 4 H X 2 C X 6 . 7 C (H, histidine; C, cysteine; X, any amino acid). The ternary structure forms two cavities, each binding a Zn 2 + ion. Some but not all CI domains have the capacity of binding DAG and the DAG analogs phorbol esters with high affinity. CI domains fold independently of adjacent amino acid sequences, generating a well-characterized structure with a hydrophobic base that interacts with the membrane. This base forms two parallel loops that generate a groove between them, in which DAG binds, allowing CI domain insertion in the membrane (Figure 1.7) (Zhang, Kazanietz et al. 1995; Kang, Benzaria et al. 2006). CI domains are classified according to residues implicated in DAG interaction, which are conserved in typical CI domains and are not found in atypical CI domains (Hurley, Newton et al. 1997). Although many 20 proteins contain CI domains, P K C s were long thought to be the only receptors for D A G or D A G analogs, due to the presence of atypical CI domains in non-PKC CI-containing proteins such as Raf-1 and Vav. However, non-PKC proteins containing a typical C I domain have been identified, including certain isoforms of D A G kinases (DGKs) , P-chimaerins and RasGRPs [reviewed by (Yang and Kazanietz 2003)]. Figure 1.7 Schematic representation of the interaction between the CI domain and DAG at the plasma membrane. first loop hydrophobic core of the plasma membrane Hydrophilic lipid headgroups of the p lasma membrane C1 domain Figure 1.7. Schematic representation of the interaction between the CI domain and D A G at the plasma membrane. D A G is shown here anchored in the plasma membrane, and inserts between the two loops of the folded CI domain. D A G , diacyl glycerol. 21 1.5.3 Regulation of R a s G R P I by its C1 domain Translocation of GEF proteins to membrane structures following receptor ligation is a common mechanism of activation, since it allows the GEFs to interact with their membrane-anchored substrates. After stimulation of a receptor coupled to PLC, DAG is generated and accumulates in membranes. Proteins that contain a typical CI domain can consequently be recruited to membranes by binding to DAG. This mechanism of membrane translocation is indeed responsible for the regulation of several proteins such as PK.Cs and chimaerins, which contain typical CI domains. In addition to the 8 cysteine and histidine residues necessary for Zn binding, the CI domain of RasGRPI contains all nine conserved residues typically found in CI domains capable of binding DAG. Several studies demonstrated binding of the RasGRPI CI domain to DAG or to phorbol esters, including in vitro binding studies of natural (Lorenzo, Beheshti et al. 2000; Madani, Hichami et al. 2004) and synthetic (Irie, Masuda et al. 2004) CI domains, molecular modeling studies (Rong, Enyedy et al. 2002) and in vivo studies (Carrasco and Merida 2004). The RasGRPI CI domain could serve as a means to bring the protein to DAG-enriched membranes following activation of PLC-coupled receptors, allowing RasGRPI to come in proximity of its membrane-bound substrate, Ras. 1.5.3.1 RasGRPI-mediated Ras activation of fibroblasts is induced by phorbol esters and requires the C1 domain One system that has been used to assay the activation of RasGRPI is fibroblast transformation; when RasGRPI is expressed in serum-grown fibroblasts, the cells become transformed, a phenomenon that involves the protein Ras (Ebinu, Bottorff et al. 1998; Tognon, Kirk et al. 1998). The first indication for a role played by the CI domain in the activation of RasGRPI was obtained by studying the transformation-inducing properties of RasGRPI. In the absence of serum, treatment with the phorbol ester PMA readily caused the transformation of RasGRPI-expressing cells, a phenomenon that did not occur in the absence of the CI domain (Tognon, Kirk et al. 1998). 22 1.5.3.2 The C1 domain is essential for RasGRPI-mediated Ras activation in response to serum, phorbol ester or DAG Activation of Ras and the MAPK pathway by RasGRPI in non-haematopoietic cells RasGRPI activation can be indirectly detected by measuring GTP loading of one of its Ras substrates, or by measuring phosphorylation of ERK1 and ERK2, kinases which are activated downstream of Ras. In fibroblasts, RasGRPI overexpression caused an increase in Ras GTP loading and ERK phosphorylation, (Ebinu, Bottorff et al. 1998; Tognon, Kirk et al. 1998), which was enhanced with serum or PMA treatment. Several other studies have indicated a role for RasGRP 1 in Ras activation via phorbol esters or DAG. Lorenzo's group determined that stimulation of keratinocytes with PMA caused an increase in Ras activation, which was enhanced in the presence of RasGRPI (Rambaratsingh, Stone et al. 2003). Prescott's group investigated the effect on RasGRPI of diacylglycerol kinase L, (DGKQ, which eliminates DAG by converting it to phosphatidic acid. When co-expressed with H-Ras, RasGRPI could cause Ras activation in HEK293 cells (Topham and Prescott 2001), but the presence of D G K i ^ blocked RasGRPI-mediated activation of Ras. Inhibition of RasGRPI activity did not occur with the expression of a kinase-dead DGK^, indicating that the inhibition of RasGRPI by DGK^ was most likely due to increased negative regulation of DAG. Furthermore, DGK^ did not inhibit PMA-induced activation of RasGRPI, likely due to the fact that this phorbol ester cannot be metabolized by DAG kinases. RasGRPI was also shown to induce the activation of Ras and ERK following stimulation of G protein-coupled receptors, which are coupled to PLCPs (Ebinu et al., 1998; Keiper et al., 2004). All these studies demonstrated that RasGRPI activation in non-haematopoietic cells can be dependent on DAG and is inducible by the DAG-mimetic phorbol esters. Role of membrane localization in RasGRPI activation GEFs need to co-localize with their membrane-bound substrates in order to be active. This suggests that the role of the CI domain and DAG in RasGRPI activation might be 23 to provide a mechanism for membrane localization. In fibroblasts, the content of RasGRPI in the membrane fraction increased after treatment with PMA (Ebinu, Bottorff et al. 1998). In the absence of the CI domain however, the accumulation of RasGRPI in the membrane fraction was not observed, indicating the requirement for the CI domain in PMA-induced membrane translocation. The use of fluorescence microscopy allowed determining more precisely the subcellular localization of RasGRPI after stimulation, since it is possible to distinguish among the different membrane structures by microscopy. Our group demonstrated that under low-serum conditions, RasGRPI was predominantly in internal membranes of 3T3 fibroblasts, and relocalized to the plasma membrane in response to PMA (Tognon, Kirk et al. 1998). Deletion of the CI domain eliminated all localization patterns, indicating that the localization of RasGRPI to membranes was mediated by its CI domain. Studies in keratinocytes showed that RasGRPI localized to the plasma membrane in response to PMA (Rambaratsingh, Stone et al. 2003), and studies in COS-1 cells showed that treatment with PMA caused the localization of RasGRPI to the plasma membrane, followed by Golgi localization (Caloca, Zugaza et al. 2003). These results support the hypothesis that RasGRPI activation is dependent on CI-mediated binding to membranes enriched in DAG or phorbol esters. Accordingly, studies in COS-1 cells showed that Ras activation by RasGRPI happened early on at the plasma membrane, and was sustained at the Golgi apparatus in these cells, and RasGRPI was the only GEF capable of inducing Ras activation on the Golgi apparatus (Bivona, Perez De Castro et al. 2003; Caloca, Zugaza et al. 2003). 24 1.5.3.3 The role of the C1 domain and DAG in the activation of RasGRPI in T lymphocytes Although they established a correlation between RasGRPI activation and its localization in membranes, all the aforementioned studies used serum and phorbol esters as activators, and do not address how specific receptors can activate RasGRPI. My interests lie in the activation of RasGRPI following antigen receptor activation. Since these receptors are coupled to PLCy proteins, they could activate RasGRPI through the production of DAG, leading to the recruitment of CI domain-containing proteins to the plasma membrane. This hypothesis is supported by studies which show that the increase in Ras and ERK activation following TCR stimulation can be blocked by PLCy inhibitors (Vassilopoulos, Kovacs et al. 1995). In a different set of studies, Topham et al. demonstrated that by overexpressing DGKc^ , TCR-induced activation of Ras was abolished, indicating a downregulation of Ras activity through DAG phosphorylation, thus suggesting a downregulation of RasGRPI (Topham and Prescott 2001). Conversely, using D G K c 7 / _ splenic T cells, Zhong et al. observed elevated Ras and ERK activation in response to TCR stimulation, presumably due to increased activation of RasGRPI (Zhong, Hainey et al. 2002). Studies involving the diacylglycerol kinase alpha family member (DGKa) also revealed a role for DAG in the activation of Ras and RasGRPI (Jones, Sanjuan et al. 2002; Sanjuan, Pradet-Balade et al. 2003) following TCR stimulation. These studies are in accordance with the fact that RasGRPI can activate Ras in response to TCR ligation in a DAG-dependent manner. Membrane translocation of RasGRPI in T cells In Jurkat cells, RasGRPI accumulates in the membrane fraction following ligation of the TCR (Ebinu, Stang et al. 2000; Jones, Sanjuan et al. 2002). Microscopy experiments more specifically showed that RasGRPI was localizing to the plasma membrane in response to TCR ligation (Sanjuan, Pradet-Balade et al. 2003). Conversely, the work of Bustelo's and Phillips' groups showed that in Jurkat cells, RasGRPI did not translocate 25 to the plasma membrane in response to TCR ligation, but rather went to the Golgi apparatus (Bivona, Perez De Castro et al. 2003; Perez de Castro, Bivona et al. 2004). Interestingly, Ras activation in response to TCR ligation was observed to take place at the Golgi apparatus, a status which was abolished in the presence of a dominant negative form of RasGRPI, and in the absence of PLCyl. The discrepancy between the different results obtained from different groups could be attributed to clonal variation within the Jurkat cell line, or to variations in the experimental procedures. In primary T cells, Merida's group used fluorescence microscopy to show that stimulation of the TCR caused RasGRPI to transiently translocate from the cytoplasm to the plasma membrane, a phenomenon that was sustained in the presence of a dominant-negative DGKct, and reduced in the presence of a constitutively active form of DGKa (Jones, Sanjuan et al. 2002). The protein Vav, which is necessary for the activation of PLCyl and thus for DAG generation in response to TCR ligation, was also implicated in the membrane localization of RasGRPI following TCR ligation. In primary T cells, RasGRPI was found to localize in membrane fractions in response to TCR ligation of wild-type cells, but not in Vav-deficient cells (Reynolds, de Bettignies et al. 2004). All these investigations demonstrated that RasGRPI translocates to membrane structures in response to TCR ligation in a DAG-dependent manner. However, although it is assumed that the translocation of RasGRPI happens through binding of the CI domain to DAG, it is notable that the role of the CI domain in this process has not been addressed in any of these studies. 1.5.3.4 The role of the C1 domain and DAG in the activation of RasGRPI in B lymphocytes As is the case for T cells, Caloca and colleagues showed that BCR-mediated Ras and ERK activation were defective in the absence of Vav3 and in the presence of a PLCy2 inhibitor (Caloca, Zugaza et al. 2003). The activation of Ras and ERK has also been shown to be defective in PLCy2~/~ DT40s (Takata, Homma et al. 1995; Oh-hora, Johmura 26 et al. 2003; Ehrhardt, David et al. 2004). These groups demonstrated that the use of phorbol esters in Vav3~'~ and PLCy2_/" DT40s could overcome the failure to activate Ras and ERK in these cells. Thus, as for T cell activation, RasGRPI seems to link the B cell receptor to Ras and ERK through the activation of Vav and PLCy proteins. Notably, in WEHI-231 B lymphocytes, RasGRPI-mediated apoptosis in response to BCR ligation required the CI domain. This study was the first one to show the involvement of the CI domain in an event happening downstream of the BCR. Thus, RasGRPI is responsible for the activation of Ras in response to BCR ligation, and its activity is dependent on the generation of DAG. Membrane translocation of RasGRPI in B cells Bustelo's group demonstrated that treatment of DT40 cells with PMA could lead to the translocation of RasGRPI to the plasma membrane as well as to inner membranes (Caloca, Zugaza et al. 2003). When the BCR of these cells was ligated, RasGRPI was recruited to the plasma membrane, a process that required Vav3. The lack of RasGRPI translocation in the absence of Vav3 was attributed to the failure of these cells to activate PLCy2 and generate DAG. Thus, ligation of the BCR in DT40 cells can lead to the translocation of RasGRPI to the plasma membrane, a process that is dependent on the production of DAG. 1.5.3.5 Conclusion - model for RasGRPI activation The activation of membrane-bound Ras commonly involves the recruitment of GEFs to cell membranes. The recruitment of RasGRPI to membrane structures following BCR and TCR activation has been shown to be dependent on Vav and PLCy proteins, which are required for receptor-mediated generation of DAG. On the other hand, membrane localization of RasGRPI can be induced by DAG or phorbol esters. These observations together with the presence of a DAG-binding CI domain in RasGRPI led to the DAG-C1 hypothesis for RasGRPI activation. This hypothesis states that upon ligation of the BCR 27 or TCR, PLCy is activated and DAG is generated in cell membranes. The CI domain of RasGRPI binds to DAG, resulting in translocation of RasGRPI to the DAG-enriched membranes. Once RasGRP 1 gets to the membranes, it can activate its substrate Ras (Figure 1.8). Concordant with the DAG-C1 hypothesis is the fact that as for translocation, the activation of RasGRPI occurs in the presence of DAG and phorbol esters. However, antigen receptors initiate many other signaling events that could equally contribute to the activation of RasGRPI. For example, the activation of RasGRPI through DAG was shown to be induced not solely by means of membrane translocation, but also via phosphorylation by protein kinase Cs (PKCs) (Aiba, Oh-hora et al. 2004; Roose, Mollenauer et al. 2005; Zheng, Liu et al. 2005). 28 Figure 1.8 Schematic representation of the DAG-C1 hypothesis. a BCR plasma membrane — Inactive GDP-bound Ras Figure 1.8. Schematic representation of the DAG-C1 hypothesis. (A), In resting cells, RasGRPI is cytosolic, kept away from the plasma membrane due to the lack of D A G present in membranes. Ras is located at the plasma membrane and is inactive. (B), Upon ligation of the BCR, PLCy2 gets activated, which causes the accumulation of D A G at the plasma membrane. The CI domain of RasGRPI can bind D A G at the plasma membrane, and RasGRPI can now physically interact with Ras and activate it. 29 1.5.4 Potential roles of other domains in the regulation of R a s G R P I Is Cl/DAG-mediated translocation to membranes sufficient for RasGRPI activation? For other proteins like PKC and P-chimaerin, the CI domain is necessary but not sufficient for activation, suggesting that the CI domain of RasGRPI might not be the only regulator of translocation and activation [reviewed by (Cho 2001; Hall, Lim et al. 2005)]. Results from Carrasco and Merida (Carrasco and Merida 2004) demonstrated that the isolated CI domain of RasGRPI was not sufficient for translocation to the plasma membrane following stimulation of the TCR in Jurkat cells. Under conditions where full-length RasGRPI translocated to the plasma membrane, the isolated CI domain concentrated in the cytoplasm and the perinuclear region, but did not localize at the plasma membrane. This result indicated that in the case of TCR-induced RasGRPI translocation, the CI domain is not solely responsible for the localization of RasGRPI to the plasma membrane, suggesting the collaboration of other regions with the CI domain to confer translocation of RasGRPI to cellular membranes. The existence of other potential regulatory domains in RasGRPI, namely the EF hands and the leucine zipper motif, raises the question of whether these domains have a role in RasGRPI translocation and activation in response to receptor activation. 1.5.4.1 Regulation of RasGRPI by the leucine zipper motif Leucine zipper (LZ) domains are defined as two a-helical turns of seven amino acids, in which a leucine or other aliphatic residue occupies every seventh position [reviewed by (Vinson, Myakishev et al. 2002)]. LZ configurations come together to form a coiled-coil structure that consists of four to five heptads, forming homo or heterodimers (Figure 1.9). The amino acids within each heptad are labelled a, b, c, d, e, f, and g, and amino acids in the a, d, e and g positions regulate dimerization stability (dimer vs oligomer) and dimerization specificity (homo vs heterodimer). Amino acids a and d are on the same surface of the a-helix and are typically hydrophobic (V, L, I, M or F), with a leucine 30 residue in 84% of the d positions; they interact with amino acids d' and a' from the second monomer, respectively, creating a hydrophobic core. The a position is involved in dimer specificity, meaning that it determines the homo- vs heterodimer status, whereas the d position is involved in dimer stability, ensuring the formation of a dimer rather than an oligomer. Amino acids g and e are frequently charged amino acids (typically E, R, K or Q) and interact with amino acids e' and g' from the other monomer, respectively, which are oppositely charged. Those amino acids can determine dimer specificity based on their charges: an attractive interaction between amino acids of different charges would lead to homodimer formation, whereas a repulsive interaction between amino acids of similar charges would prevent homodimerization, but could lead to attraction between different monomers that present compatible amino acids, forming then heterodimerization. The sequence of the RasGRPI leucine zipper is concordant with the consensus sequence of leucine zipper motifs described above: RasGRPI LZ reveals 5 heptads which contain leucine residues in 4 of 5 d positions, hydrophobic residues in 3 of 4 a positions and 9 of 10 charged residues in e and g positions. By dimerizing with another protein, the LZ motif could get RasGRPI to interact at the plasma membrane with a membrane-localized protein. Alternatively, RasGRPI could homodimerize and increase the avidity of the protein for its ligand at the plasma membrane. 31 Figure 1.9 Schematic representation of leucine zipper dimerization. Figure 1.9. Schematic representation of leucine zipper dimerization. Ffeptads of two leucine zippers motifs (LZ) dimerize to form a coiled-coil structure. The amino acids within each heptad are labelled a, b, c, d, e, f, and g, with amino acids in the a, d, e and g positions (in bold) regulating dimerization stability and specificity. The dotted lines indicate interaction between respective amino acids. 32 1.5.4.2 PKC-mediated phosphorylation of RasGRPI The activation of RasGRPI by DAG or phorbol esters could reflect a requirement for PKC in addition to or instead of CI-mediated translocation. Three recent reports demonstrated that phosphorylation of threonine 184 located a few amino acids upstream of the GEF domain in RasGRPI was mediated by PKC (Aiba, Oh-hora et al. 2004; Roose, Mollenauer et al. 2005; Zheng, Liu et al. 2005). PKC-driven phosphorylation had to occur in order for RasGRPI to be activated in response to TCR ligation, a process dependent on DAG. PKC-dependent phosphorylation was also necessary for the activation of RasGRP3 following BCR ligation (Aiba, Oh-hora et al. 2004; Zheng, Liu et al. 2005). The regulation of RasGRPI might therefore not be as straightforward as the simple DAG-C1 hypothesis stipulates. 1.6 THESIS OBJECTIVES My thesis is focused on understanding the mechanisms by which BCR ligation leads to RasGRP 1 activation. Work by other groups before and during my thesis research supported the hypothesis that RasGRPI activation in response to BCR ligation occurs via Cl/DAG-mediated recruitment to membranes. My initial research objectives were to address the following outstanding questions: 1. Does RasGRPI translocate to the plasma membrane in response to BCR ligation? 2. Is the CI domain of RasGRPI sufficient for RasGRPI translocation in response to BCR ligation? 3. Is the C1 domain necessary for BCR-mediated translocation of RasGRP 1 ? 4. Is the translocation of RasGRPI dependent on PLCy2? 5. What other region of RasGRPI is involved in the translocation of RasGRPI to the plasma membrane in response to BCR ligation? 33 M y discovery of additional domains required for BCR-induced translocation o f RasGRPI to the plasma membrane led me to formulate a new model of RasGRPI activation via B C R . Accordingly, I subsequently addressed the following questions: 6. Are the additional domains responsible for BCR-induced translocation o f RasGRP 1 necessary for its activation in response to B C R ligation? 7. Can these domains provide plasma membrane localization independently o f the C I domain, or in collaboration with the C I domain? The translocation of RasGRPI to the plasma membrane in response to B C R ligation is assumed to lead to activation of the protein. I therefore addressed the following questions: 8. Are all the domains involved in the translocation of RasGRPI to the plasma membrane also involved in the activation of RasGRPI following B C R ligation? 9. Is translocation of RasGRPI to the plasma membrane necessary or sufficient for the activation of RasGRPI? 34 CHAPTER 2: MATERIALS AND METHODS Cells and Reagents. Wild-type and PLCy2-deficient DT40 cells were obtained from Mike Gold (University of British Columbia) and were originally from T. Kurosaki (RIKEN Research Centre, Yokohama). All DT40 cells used in this study were transfected with an expression plasmid expressing the ecotropic receptor, to make them permissible for infection with murine retroviral vectors. Cells were cultured in RPMI 1640 (StemCell Technologies) supplemented with 10% fetal bovine serum (HyClone), 1% chicken serum (Invitrogen), 50 uM 2-mercaptoethanol (Sigma-Aldrich), 2 mM L-glutamine (StemCell Technologies). Anti-chicken IgM polyclonal antibody was from Bethyl Laboratories. PMA was from Sigma-Aldrich, and DAG (1,2-dioctanoyl-sn-glycerol) was from Sigma-Aldrich or Calbiochem. Anti-ERKl/2 and anti-phospho-specific ERK1/2 antibodies were from Cell Signaling Technology, anti-GFP was from AbCam, anti-HA was from Babco/Covance and anti-pan Ras (detecting all Ras GTPases) was from Oncogene Research Products. Alexa fluor 488-conjugated anti-GFP polyclonal antibody was from Molecular Probes. HRP-conjugated secondary antibodies were from Jackson Immunoresearch. Construction of modified forms of RasGRPI. The N-terminally eGFP-tagged (GFP) form of full length murine RasGRPI (RG1) was derived from the XFL construct previously described (Tognon, Kirk et al. 1998), with the eGFP coding sequences from pEGFP-Cl (Clontech) fused to amino acid 2 of RasGRPI (GenBank accession NP 035376). The sequence at the fusion site is: ...DELYY^SGLRSSAgSJFGTLGKAR..., with italicization of the sequences that don't naturally occur in eGFP or RasGRPI. The following constructs are modifications of this RG1 construct with the indicated changes (sequences that don't naturally occur in RasGRPI are italicized, C-terminus indicated by star): RG1AC1 = deletion of amino acids 538 to 595; sequence around deletion = ...YSKLG67KSPAIS... RG1ALZ - deletion of amino acids 745 to C-terminus; sequence around deletion = ...LHLRLRST*. RG1 ASp-LZ = deletion of amino acids 597 to C-terminus; sequence around deletion = ...FECKKRIKPJE4*. RG1AK = deletion of amino acids 714 to 733; sequence around deletion = .. .CPSPASTKPKPEL... 35 RGlKp = ...RKRAsadtENKESSTTCPKPEL... In addition, amino acids 538 and 539, immediately N-terminal to the CI domain, are modified from LG to ST in the RG1 ASp-LZ construct. RG1AC1ARP = deletion of amino acids 538 to 715; sequence around deletion = ... YSKLGSTALVRKR... RGl/Pren = replacement of RasGRPI amino acids 538 to C-terminus with an HA epitope tag plus the prenylation signal of K-Ras; sequence from fusion to C-terminus = ...YSKLGSTEA YPYDYASGSRKHKEKMSKDGKKKKKKSKTKCVIM*. The isolated CI domain construct consisted of amino acids 540 to 596 of RasGRPI, fused at the end terminus to eGFP (sequence around fusion = ...DELYKSGLRSLKSTF?HNF...) and with the C-terminal sequence ...FECKKRIKETEA*. For all of the other GFP fusions of the C-terminal domains of RasGRPI, the sequence at the N-terminal (non-RasGRPl) side of the fusion had the sequence ...YKSGLRSLKST... The sequence at the RasGRPI side of the fusion is indicated in Figure 3.14a (e.g. ...KSPAIST... for Sp-LZ). Transduction and establishment ofDT40 cell lines. Transfection of BOSC 23 ecotropic packaging cells with retroviral vector plasmid DNA was performed as described (Pear, Nolan et al. 1993). All DT40 cells used in this study were electroporated with 20 pg of DNA encoding the ecotropic receptor to make them permissive for infection with ecotropic retroviral vectors. 1 x 107 cells were resuspended in 800 pl of serum-free DT40 medium and electroporated at 550V, 25pF. DT40 cells were then resuspended in DT40 medium and plated at a density of 5 x 105 cells/ml. In order to establish DT40 cells permanently permissive for infection with ecotropic retroviral vectors, the electroporated cells were infected with viruses carrying the ecotropic receptor 16-24 hours after electroporation. Virus-containing medium was used to infect 0.5-1.0 x 106 DT40 cells/ml at a ratio of 1:1 viral supernatant:DT40 medium and supplemented with 10 pg/ml polybrene. 2-3 volumes of DT40 medium was added to the cells 5 to 10 hours after the start of infection. Infected cells were then subjected to selection with bleocin (10 pg/ml) until the untransduced cells were dead (approximately 5 days). Stable levels of GFP-RasGRPl protein expression were achieved by infecting the cells with viruses carrying the desired mutant of RasGRPI, followed by selection with puromycin (0.25 pg/ml) until the untransduced cells were dead (approximately 5 days). Following antibiotic selection, 36 cells expressing high levels of GFP compared to untransduced counterparts were sorted by flow cytometry using the FACSVantage SE cell sorter (Becton Dickinson); these cells represented about 10-25% of the total population prior to sorting. For each experiment, 3 populations of RasGRP 1-expressing cells were sorted, according to their level of expression (high, medium or low). After sorting, cells were kept on ice, spun down and resuspended in DT40 medium at a density of 5 x 105 cells/ml or lower. Expression levels of the different constructs were assessed by FACS analysis for all cell lines. In addition, Western blot analysis was performed on truncation mutants, as indicated in the appropriate figures. Stimulation of cells for fluorescence microscopy. DT40 cells expressing different mutants of RasGRPI were plated on poly-L-lysine-coated glass coverslips and starved in serum-free medium for 3-4 hours prior to stimulation. Cells were stimulated in Hank's buffer (StemCell Technologies) with 5 ug/ml anti-chicken IgM, 100 uM DAG or 500 ng/ml PMA, followed by fixing with 4% formaldehyde in PBS. The cells were then permeabilized with 0.2% triton-XlOO in PBS and stained with alexa fluor 488-conjugated anti-GFP antibody. This increases sensitivity of detection of GFP, compared to GFP fluorescence on its own. Fluorescence microscopy was performed using the Carl Zeiss Axioplan 2 Imaging Universal Research Microscope (Zeiss). A 450-485 nm excitation filter and a 500-545 emission filter were used. Images were captured using the Retiga EX mono 12 bit cooled camera (Quorum Technologies) and the OpenLab Imaging software. Stimulation and lysis of cells for electrophoresis. 2.5 X 107 DT40 cells were incubated in activation buffer (Saxton, van Oostveen et al. 1994) (25 mM Hepes pH 7.2, 125 mM NaCl, 5 mM KC1, 1 mM CaCl2, 1 mM Na2HP04, 0.5 mM MgS04, 2 mM L-glutamine, 1 mM sodium pyruvate, 0.1% glucose, 0.1% BSA, 50 uM 2-mercaptoethanol) for 10 min at 37°C prior to stimulation. Anti-chicken IgM or DAG were then added to a concentration of 5 ug/ml and 100 uM respectively, for the indicated times. Cells were immediately lysed with 1.5 volume of ice cold lysis buffer (25 mM Hepes pH 7.5, 150 mM NaCl, 1% NP-40, 0.25% sodium deoxycholate, 10% glycerol, 25 mM NaF, 10 mM MgCb, 1 mM EDTA, 1 mM NaMoCu) containing 2 pg/ml aprotinin, 2 pg/ml leupeptin, 37 0.5 mM PMSF and 1 mM activated Na3VC>4. Proteins were then denatured using BioRad XT sample buffer and heated prior to gel electrophoresis on 12% XT-Criterion acrylamide gels (Bio-Rad). For each experiment, equal numbers of cells were loaded in each well (between 50 000 and 100 000 cells/well for 26-well gels, and between 125 000 and 250 000 cells/well for 18-well gels). Affinity precipitation of activated Ras proteins. DT40 cell lysates were prepared as described above and 900 pg of proteins were incubated for 30 min with GST-RafRBD fusion proteins prebound to glutathione-agarose beads, as derived from (Taylor, Resnick et al. 2001). After washing, proteins were denatured and subjected to gel electrophoresis as described above. Western Blot analysis. After electrophoresis, proteins were transferred to polyvinylidene difluoride membranes (Millipore). After blocking the membranes overnight at 4°C in TBS-T (25 mM Tris-HCl pH 7.4, 3 mM KC1, 150 mM NaCl, 0.05% Tween-20) containing 5% bovine serum albumin (BSA), primary antibodies were applied to the membrane for 90 min at room temperature in 2% BSA TBS-T, and horseradish peroxidase-conjugated secondary antibodies were applied for 45 min at room temperature in TBS-T containing 1% BSA. Membranes were exposed to substrate/ECL (Santa Cruz) and chemiluminescence was detected using the VersaDoc 5000 imaging system (Bio-Rad). Band volume analysis used Quantity One software (Bio-Rad). Cell profile drawing and analysis. The intensity and intracellular localization of GFP-fusion proteins were demonstrated graphically using the Scion Image program (version 4.0, Scion Corp. Frederick MD USA). Each profile was generated by drawing a straight line from the centre of the cell of interest, through and beyond the plasma membrane, as shown in the Figures. Analysis of fluorescence microscopy results and selection of representative cells for figures. For each figure, a minimum of three different experiments were conducted. For each experiment, two slides were looked at for every construct, at any condition. General 38 observations were made by eye, comparing the different mutants to the appropriate control (e.g. GFP and RasGRPI). For each experiment, between 200 (for extreme phenotypes) and 800 (for intermediate phenotypes) cells were scored (table 3.1). Representative cells were chosen for the figures to illustrate the phenotype of the majority of the cells. In some cases, only a minority of cells showed a gain-of-function phenotype relative to controls; in these cases, the proportion of cells represented by the figure is indicated within the figure. 39 CHAPTER 3: MECHANISMS OF R A S G R P I TRANSLOCATION TO T H E PLASMA M E M B R A N E IN RESPONSE TO B C R LIGATION 3.1 EXPERIMENTAL SYSTEM 3.1.1 DT40 B cells as a model system The experimental system that I chose to study the function of RasGRPI is the chicken immature B cell line DT40. DT40 cells express a functional IgM receptor of unknown specificity, and it is possible to engage this receptor by cross-linking with an antibody raised against the variable part of the IgM (a-IgM). IgM cross-linking in these cells triggers a series of phosphorylation events, leading to the activation of PLCy2, which results in the activation of Ras and ERK. In addition, this cell line harbours unusually high levels of homologous recombination efficiency. A variety of knockout DT40 cell lines have consequently been generated, targeting proteins involved in various biological events including DNA repair, somatic hypermutation, cell cycle regulation, apoptosis, and most extensively, events involved in B cell receptor signalling. DT40 cells express RasGRPI and are, therefore, a useful model system for studying the contribution of RasGRPI in Ras activation following BCR ligation. 3.1.2 Generation of DT40 cells expressing GFP-taqqed R a s G R P I In order to determine the intracellular localization of RasGRPI, I generated DT40 cells expressing the full-length murine RasGRPI cDNA fused to the enhanced green fluorescent protein (GFP) at its N-terminus (Figure 3.1). Due to the fusion, the first amino acid of the RasGRPI cDNA was deleted. The cDNAs were cloned in a pCTV211 40 retroviral vector derived from pCTV3 (Tognon et al., 1998), which contains a puromycin resistance gene for drug selection of transduced cells. The vectors were transfected into the ecotropic viral packaging cell line BOSC 23, using the calcium-phosphate precipitation technique (see material and methods). All DT40 cells used in this study were transfected with an expression plasmid expressing the ecotropic receptor to make them permissive for infection with ecotropic retroviral vectors. Stable levels of protein expression were achieved by infecting the cells with viruses, followed by selection with puromycin until the untransduced cells were dead (approximately 5 days). Following antibiotic selection, cells expressing high levels of GFP compared to untransduced counterparts were sorted by flow cytometry: these cells represented about 10-25% of the total population prior to sorting. For each experiment, 3 populations of RasGRP 1-expressing cells were sorted, according to their level of expression (high, medium or low) (Figure 3.2a). This insured that cells expressing different mutants of RasGRPI could be compared to cells expressing similar levels of RasGRPI. Because the available antibodies raised against RasGRPI do not allow the detection of endogenous chicken RasGRPI in DT40 cells, I was unable to determine the fold increase of RasGRPI protein after RasGRPI transduction. Using the same expression system, our lab obtained a three-fold increase in RasGRPI levels compared to endogenous RasGRPI found in the murine immature B cell line WEHI 231. The levels of expression attained in DT40 cells were comparable to levels achieved in WEHI 231 cells; I therefore assumed that overexpression of murine RasGRPI was achieved at relatively high levels in DT40 cells. The level of expression and size of protein were characterized by Western blot analysis using an anti-GFP antibody, as shown in Figure 3.2b. 41 Figure 3.1 Structures of GFP-tagged RasGRPI proteins. GFP GEF O KWEN LZ RG1 RGlGEFu RG1AC1 RG1AC1ALZ RGlAClARp RG1ALZ RG1AK ® 0 RGI/K41 RGlASp-LZ RGl/Pren Figure 3.1. Structures of GFP-tagged RasGRPI proteins. G E F , guanine nucleotide exchange factor domain; C I , C I domain; Sp, spacer region; K W E N , K W E N loop; L Z , leucine zipper; Pren, artificial prenylation signal; Rp, translocational repressor region. Figure 3.2 Expression data of different RasGRPI protein mutants. non-expressing cells i 1 r 10° 101 102 1 0 3 1 0 4 GFP high RasGRPI expression medium RasGRPI expression low RasGRPI expression n r 10° 101 102 103 10" G F P RG1 G F P RG1 RG1ALZ RG1AC1 AC1ALZ RGIGEFJJ GFP-tagged RasGRPI Figure 3.2. Expression data of different RasGRPI protein mutants. A, DT40 cells expressing GFP-fusion proteins were sorted to obtain high-expressing cell populations - figure shows RasGRPI as an example. For each experiment, 3 populations of RasGRPI-expressing cells were sorted, according to their levels of expression (high, medium or low). This insured that mutant-expressing lines could be compared to cells expressing similar levels of full-length RasGRPI. B, Unstimulated and stimulated cells expressing different RasGRP 1 mutants constructs were lysed and subjected to Western Blot analysis as described in the Materials and Methods. An anti-GFP antibody was used to detect expression levels and the size of different RasGRPI proteins. FSC, Forward Side Scatter. 43 3.2 CELLULAR LOCALIZATION OF R A S G R P I IN DT40 CELLS 3.2.1 RasGRPI translocates to the plasma membrane in response to BCR ligation Ras proteins are localized in membrane domains, and it is consequently required for their exchange factors to localize at membranes so as to activate the small GTPases. I was first interested in determining the localization of RasGRPI following BCR ligation. The localization of GFP-RasGRPl (RasGRPI) was addressed by fluorescence microscopy, an effective technique for determining membrane specificity of protein localization (see Materials and Methods for result analysis and selection of representative cells). Cellular structures were described as follows - a schematic representation is shown in Figure 3.3: localization of GFP-fusion proteins at the plasma membrane was characterized by a well-defined bright rim at the cell periphery (a), whereas localization at the nuclear membrane was characterized by a well-defined bright rim surrounding the nucleus (b). Protein localization at the endoplasmic reticulum and at the Golgi apparatus (referred to as inner membranes) were represented as a diffused yet structured concentration of signal around the nucleus for the endoplasmic reticulum, and as a bright signal concentration at a specific site of the nuclear membrane periphery for the Golgi apparatus (c). Cytoplasmic localization was regarded as an evenly distributed signal localized outside of the nucleus (d), and nuclear localization was seen as a signal found in the nucleus (e). In resting cells, RasGRPI localized mainly outside of the nucleus, with a tendency to concentrate in inner membranes (Figure 3.4a). No apparent localization at the plasma membrane was observed. Upon ligation of the B cell receptor with a-IgM, RasGRPI strongly localized to the plasma membrane, as determined by the defined signal at the periphery of the cells and the failure to detect the protein at any other cellular structure. The localization pattern of RasGRPI was distinct from the GFP control which is evenly distributed throughout the cell (Figure 3.4b). Thus, RasGRPI translocates to the plasma membrane in response to BCR ligation. 44 Figure 3.3 Schematic representation of GFP-tagged protein localization. a b c plasma membrane localization nuclear membrane localization inner membrane localization o o Golgi Endoplasmic reticulum d e cytoplasmic localization nuclear localization Figure 3.3. Schematic representation of GFP-tagged protein localization. In this schematic representation, the highlighted portion represents the described structure. The localization of GFP-fusion proteins was referred to as follows (a), plasma membrane localization was characterized by a well-defined bright rim at the cell periphery (b), localization at the nuclear membrane was characterized by a well-defined bright rim surrounding the nucleus (c), localization at the endoplasmic reticulum and at the Golgi apparatus were defined as a concentration of signal around the nucleus and a bright signal concentration at a specific site of the nuclear membrane periphery respectively (d), cytoplasmic localization was defined as an evenly distributed signal localized outside of the nucleus (e), and nuclear localization was seen as a signal found in the nucleus. 3.2.2 RasGRPI translocation can be induced by DAG or phorbol esters Ligation of the B cell receptor is coupled to the activation of PLCy2, leading to the production of DAG at the plasma membrane as well as to the activation of Ras. According to the DAG-C1 hypothesis, the DAG generated following BCR ligation acts as anchor for the CI domain of RasGRPI, leading to the recruitment of RasGRPI to the plasma membrane. My observation that RasGRPI translocates to the plasma membrane 45 in response to BCR ligation fit the CI-DAG hypothesis, and implies that this translocation occurs through the interaction between the RasGRPI CI domain and DAG. A prediction of this model is that RasGRPI translocation would be induced by DAG. To address this prediction, I stimulated RasGRPI-transduced DT40 cells with the DAG 1,2-Dicapryloyl-s«-glycerol. As seen in Figure 3.4, treatment of RasGRPI-expressing cells with DAG induced the translocation of the protein to the plasma membrane in a manner similar to BCR-induced translocation. Thus, in accordance with the DAG-C1 hypothesis, these results demonstrate that translocation of RasGRPI to the plasma membrane can be induced by DAG. 46 Figure 3.4 RasGRPI translocates to the plasma membrane in response to a-IgM and DAG. RG1 GFP Figure 3.4. RasGRPI translocates to the plasma membrane in response to a-IgM and DAG. DT40 cells expressing RG1 (A) or GFP (B) were unstimulated (nil) or stimulated with anti-IgM (5 ug/ml) or DAG (100 uM) for 15 minutes. Cells were then fixed, stained with Alexa fluor 488-conjugated anti-GFP to enhance fluorescence, and photographed by fluorescence microscopy. On the right of each cell is a graphic representation of the GFP-protein intensity across the cell. Each graph is representative of the section marked by a white line on the corresponding cell, and a line analogous to the white section can be found above each graph. The vertical mark above the line represents the boundary of the cell. The Y axis shows the relative intensity of the fluorescence. Details regarding the phenotype of RasGRPI-expressing cells stimulated with a-IgM are presented in table 3.1. 47 3.3 THE C1-DAG INTERACTION IS NOT SOLELY RESPONSIBLE FOR RASGRPI TRANSLOCATION TO THE PLASMA MEMBRANE IN RESPONSE TO BCR LIGATION 3.3.1 The C1 domain of R a s G R P I does not localize to the plasma membrane in response to B C R ligation The DAG-C1 hypothesis predicts that translocation of RasGRPI to the plasma membrane following a-IgM stimulation happens solely through binding of the CI domain to DAG. This hypothesis assumes that DAG is predominantly located at the plasma membrane. If the DAG-C1 hypothesis is correct, the isolated CI domain of RasGRPI should translocate in the same manner as RasGRPI after stimulation with cx-IgM. To test this hypothesis, I used a GFP-C1 domain fusion protein as a probe for DAG and determined the localization pattern of the protein following BCR stimulation. In resting cells, the CI domain was distributed more or less ubiquitously throughout the cell, with subtle inner membrane localization in some cells. Surprisingly, under conditions where a-IgM provided strong signals for RasGRPI translocation to the plasma membrane, the isolated CI domain did not localize to the plasma membrane (Figure 3.5). These results either suggest that the production of DAG in cell membranes following BCR ligation is not abundant enough to cause the localization of the CI domain to cell membranes, or that the interaction between the CI domain and DAG is not sustained or strong enough to cause the localization of the CI domain to cell membranes. Alternatively, the presence of the CI domain at the plasma membrane could be repressed by an unknown mechanism. The translocation of RasGRPI to the plasma membrane in response to BCR ligation might therefore not be driven only, if at all, by the interaction between DAG and the CI domain. I used the CI-domain probe to test whether DAG-induced translocation of RasGRPI to the plasma membrane occurs via binding of the CI domain to this ligand. When treated with DAG, the CI domain did not relocalize at the plasma membrane, but instead 4 8 weakly concentrated in inner membranes (Figure 3.5), suggesting that the translocation of RasGRPI to the plasma membrane via DAG stimulation is not exclusively due to binding of the CI domain to DAG. One question that emerges is whether the CI domain of RasGRPI is capable of binding ligand, or whether it is prevented from doing so at the plasma membrane. The phorbol ester phorbol 12-myristate 13-acetate (PMA) binds CI domains with a higher affinity than DAG, and it does not get metabolized by DAG kinases (DGKs). If the CI domain is capable of binding ligand at the plasma membrane, treatment with PMA will cause the CI domain to localize where the ligand is located, that is, in membranes. When the cells were stimulated with PMA, the CI domain strongly translocated to the plasma membrane as well as to inner membranes, indicating that the CI domain is competent for binding ligand and is not excluded from the plasma membrane. The differential localization of the CI domain after DAG or PMA treatment could reflect the fact that these molecules localize at different membrane structures, or that they get metabolized differently. The absence of the CI domain at the plasma membrane in response to BCR ligation and to DAG treatment might also be an indication that, given the moderate affinity of the CI domain for DAG, the concentration of DAG generated at the plasma membrane is insufficient to cause a stable accumulation of the CI domain at the plasma membrane. 49 Figure 3.5 The C I domain is insufficient for B C R - or DAG-induced translocation to the plasma membrane. RG1 C1 of RG1 Figure 3.5. The C1 domain is insufficient for BCR- or DAG-induced translocation to the plasma membrane. DT40 cells expressing RG1 or the isolated CI domain with an N-terminal GFP tag (CI of RG1) were unstimulated (nil) or stimulated with anti-IgM or DAG for 15 minutes, or with PMA (500 ng/ml) for 5 minutes. After stimulation, cells were prepared for fluorescence microscopy and analyzed as described for figure 3.4. 3.3.2 A C1 domain/DAG interaction is necessary for optimum translocation of RasGRPI to the plasma membrane following BCR ligation The fact that the CI domain is not sufficient for localizing RasGRPI to the plasma membrane in response to BCR ligation raises two possibilities: the CI domain might simply not be involved in BCR-mediated translocation of RasGRPI to the plasma 50 membrane, or other domains must cooperate with the CI domain to achieve BCR-induced translocation of RasGRPI to the plasma membrane. To distinguish between these two possibilities, a mutant form of RasGRPI with a precise deletion of the CI domain was expressed in DT40 cells (RG1AC1), and its localization following BCR ligation was addressed. Like RasGRPI, RG1AC1 localized mainly outside of the nucleus in resting cells, with a slight tendency to concentrate in inner membranes (Figure 3.6 a). However, unlike RasGRPI, BCR-mediated translocation of RG1AC1 was severely affected, although not completely eliminated. These results demonstrate that the C1 domain is involved in BCR-mediated translocation, and requires the cooperation of other domains to mediate the translocation of RasGRPI to the plasma membrane. Furthermore, the fact that translocation is not completely abolished in the absence of the CI domain indicates the existence of a CI-independent mechanism of RasGRPI translocation to the plasma membrane in response to BCR ligation. Figure 3.6 C I domain and D A G are required for efficient translocation of RasGRPI. b PLCy2 -A DT40 RG1AC1 RG1 Figure 3.6. The CI domain and DAG are required for efficient translocation of RasGRPI. A, DT40 cells expressing RG1AC1 were unstimulated (nil) or stimulated with anti-IgM (5 pg/ml) or DAG (100 uM) for 15 minutes. After stimulation, cells were prepared for fluorescence microscopy and analyzed as described for figure 3.4. B, PLCy2-deficient DT40 cells expressing RG1 were treated as described for panel (A). Details regarding the phenotype of cells stimulated with a-IgM are presented in table 3.1. 51 Table 3.1 Proportion of cells demonstrating plasma membrane localization following B C R ligation. RG1 # of cells counted ave % of cells with translocation at P M Description unstimulated > 1500 - 0 % nuclear exclusion a-IgM >2800 > 95 % strong localization exclusively at P M RG1AC1 # of cells counted ave % of cells with translocation at P M Description unstimulated >900 - 0 % nuclear exclusion a-IgM 1599 49.58% weak localization at P M R G l A S p L Z # of cells counted ave % of cells with translocation at P M Description unstimulated >700 - 0 % ubiquitous - very weak localization in I M a-IgM > 1300 - 0 % ubiquitous - very weak localization in IM R G 1 A L Z # of cells counted ave % of cells with translocation at P M Description unstimulated >700 - 0 % ubiquitous a-IgM 1672 36.00% weak localization at P M RG1AC1ARP # of cells counted ave % of cells with translocation at P M Description unstimulated 1659 8.79% weak localization at P M in a few cells a-IgM 2887 74.60% medium translocation at P M R G 1 A K # of cells counted ave % of cells with translocation at P M Description unstimulated >900 - 0 % nuclear exclusion a-IgM 2220 36.00% weak localization at P M RGIK41 # of cells counted ave % of cells with translocation at P M Description unstimulated >800 - 0 % nuclear exclusion a-IgM 2153 44.50% weak localization at P M P L C y 2 R G 1 # of cells counted ave % of cells with translocation at P M Description unstimulated >600 - 0 % nuclear exclusion a-IgM 822 42.64% weak localization at P M Table 3.1: Proportion o f cells demonstrating plasma membrane localization following B C R ligation. Column 1, shows the different constructs, with numbers for unstimulated and B C R -stimulated cells. Column 2, # of cells counted represents three different experiments pooled together. Column 3, ave % of cells with translocation at PM, shows the percentage o f cells in which the GFP-protein harbours plasma membrane localization, as an average among the 3 experiments. Column 4, description, describes the strength o f the plasma membrane localization when observed, ave, average; P M , plasma membrane; I M , internal membranes. 52 3.3.3 The C1-domain provides specificity for D A G binding The CI domain is necessary for achieving optimum BCR-mediated translocation of RasGRP 1 to the plasma membrane, and does so most likely by binding DAG. To test this possibility, I evaluated whether the CI domain is necessary for DAG-mediated translocation of RasGRPI. If the CI domain functions by binding DAG, RasGRPI will not translocate to the plasma membrane in response to DAG in the absence of the CI domain. Under conditions where DAG treatment caused the translocation of RasGRPI to the plasma membrane, RG1AC1 did not show any translocation (Figure 3.6 a), indicating that the CI domain is necessary for DAG-mediated translocation to the plasma membrane. The plasma membrane translocation of RasGRPI occurring in response to BCR ligation is partially dependent on the CI domain, whereas the DAG-mediated translocation of RasGRPI is entirely dependent on the CI domain. The CI-dependent component of BCR-mediated translocation might therefore reflect the requirement for DAG generation in RasGRPI translocation. To test if RasGRPI translocation is also dependent on DAG production, I expressed full-length RasGRPI in PLCy2"/" DT40s. As mentioned previously, the generation of DAG following BCR ligation is dependent on PLCy2. If the defect in RasGRPI translocation in the absence of the CI domain is due to the lack of interaction between the CI domain and DAG, BCR-mediated translocation in the absence of DAG should be equivalently deficient. In PLCY2"7" cells, translocation of RasGRPI was severely reduced following BCR activation (Figure 3.6 b), indicating that the absence of DAG in those cells impairs BCR-mediated translocation of RasGRPI. Notably, in the absence of PLCy2, weak localization to the plasma membrane was observed in some cells, demonstrating a DAG-independent mechanism of translocation. Altogether, these results are concordant with the requirement for the CI domain to bind DAG to achieve RasGRPI translocation, and they demonstrate the existence of a CI/DAG-independent mechanism of translocation in response to BCR ligation. 53 3.3.4 Revis ion of the model for BCR-mediated translocation of R a s G R P I The model for the mechanism of RasGRPI activation downstream of the BCR involves the recruitment of RasGRPI to membranes where it can activate Ras. In this regard, the DAG/C1 hypothesis stipulates that the translocation of RasGRPI occurs through binding of the CI domain to DAG, which is generated by PLCy2 upon receptor activation. I determined that the interaction between the CI domain and DAG is crucial to achieve optimum translocation of RasGRPI to the plasma membrane. I also made two new findings which compelled us to revise our previous model: first, the CI domain does not provide a sufficient means for BCR-mediated translocation to the plasma membrane, and second, some translocation can occur at the plasma membrane in response to BCR ligation independently of DAG or the CI domain. I therefore suggest an alternative model for the recruitment of RasGRPI to the plasma membrane, in which domains of RasGRPI other than the CI domain are involved in the translocation of the protein following BCR ligation (Figure 3.7). These domains are presumably responsible for the DAG/C1 -independent mechanism of RasGRPI translocation, and cooperate with the CI domain to enhance translocation to the plasma membrane. This presented a new goal for my thesis: to identify domains of RasGRPI that confer DAG-independent translocation to the plasma membrane and to identify domains of RasGRPI that cooperate with the CI domain to achieve high-efficiency translocation to the plasma membrane. 5 4 Figure 3.7 Revised model for the mechanism of RasGRPI translocation to the plasma membrane in response to B C R ligation. DAG-C1 model BCR b Revised model BCR Figure 3.7. Revised model for the mechanism of RasGRPI translocation to the plasma membrane in response to BCR ligation. A, The DAG/C1 hypothesis attributes the translocation of RasGRPI solely to binding of the CI domain to DAG, which is generated following BCR ligation. B, The revised model for BCR-mediated translocation of RasGRPI to the plasma membrane involves the binding of the CI domain to DAG, as well as the binding of another RasGRP 1 domain to an unknown (X) BCR-induced ligand at the plasma membrane. 5 5 3.4 T H E C-TERMINAL REGION OF R A S G R P I IS RESPONSIBLE FOR MEDIATING C1 -COOPERATIVE AND C1-INDEPENDENT MECHANISMS OF PLASMA MEMBRANE TRANSLOCATION IN RESPONSE TO B C R LIGATION 3.4.1 The region C-terminal to the C1 domain is essential for B C R -mediated translocation of RasGRI to the plasma membrane As stated in section 1.4, the structure of RasGRPI comprises the REM and GEF domains at the N-terminus, followed by the EF hands, the CI domain and a leucine zipper motif at the C-terminus (Figure 1.5). In order to identify the domains that cooperate with the CI domain to induce BCR-mediated translocation of RasGRPI, I started by assessing the requirement for the region C-terminal to the CI domain. This region of RasGRPI contains a stretch of 36 amino acids that form a leucine zipper (LZ), which could potentially contribute to translocation through protein-protein interactions. The sequence located between the CI domain and the LZ motif, namely the spacer region (Sp), does not share extensive homology with any known sequence, based on sequence alignments generated using the NCBI BLAST program. In the absence of Sp and LZ (RG1 ASpLZ), the protein adopted a ubiquitous localization pattern investing cells, and no membrane translocation was observed in response to BCR ligation (Figure 3.8). These results are concordant with Sp-LZ being essential for the CI domain-independent and the CI domain-cooperative mechanism of plasma membrane translocation in response to BCR ligation. The fact that RG1 ASpLZ does not translocate to the plasma membrane in response to BCR ligation despite the presence of the CI domain, is in agreement with the theory that Sp-LZ cooperates with the CI domain to achieve DAG-induced translocation. If this is true, RG1 ASpLZ should not be able to translocate to the plasma membrane in response to DAG treatment, since the presence of Sp-LZ would be required for Cl/DAG interaction at the plasma membrane. As with BCR ligation, translocation of RG1 ASpLZ 56 to the plasma membrane was not observed when cells were treated with DAG; instead, a strong signal was detected at the nuclear membrane (Figure 3.8). These results suggest that Sp-LZ cooperates with the CI domain to achieve translocation at the plasma membrane in response to DAG treatment. Figure 3.8 The region C-terminal to the CI domain is essential for B C R and DAG-mediated translocation to the plasma membrane. Figure 3.8. The region C-terminal to the C1 domain is essential for BCR and DAG-mediated translocation to the plasma membrane. DT40 cells expressing RG1 or RG1 ASp-LZ were unstimulated (nil) or stimulated with anti-IgM for 15 minutes. Cells were then prepared for fluorescence microscopy and analyzed as described for figure 3.4. Details regarding the phenotype of cells stimulated with a-IgM are presented in table 3.1. 57 3.4.2 The Sp-LZ region confers plasma membrane translocation in response to B C R ligation via a PLCy2-independent mechanism RGl ASpLZ cannot translocate to the plasma membrane in response to BCR ligation, indicating that the region N-terminal to the CI domain does not confer translocation on its own. These results suggest that the Sp-LZ region contains the sequence which provides RasGRPI with a mechanism of translocation that is independent of the CI domain. If this is the case, Sp-LZ should be able to translocate to the plasma membrane following BCR activation, despite the absence of the CI domain. As shown in Figure 3.9a, Sp-LZ mainly localizes in the nucleus in resting cells and is able to translocate to the plasma membrane in response to BCR ligation, indicating that the Sp-LZ region is responsible for the CI-independent mechanism of translocation to the plasma membrane. This mechanism of translocation is suspected to function independently of DAG. To verify this possibility, the localization of Sp-LZ was evaluated in response to DAG, as well as in response to BCR ligation of PLCy2"A cells. Following treatment with DAG or PMA, Sp-LZ did not translocate to cell membranes (Figure 3.9a), demonstrating that DAG does not induce the translocation of Sp-LZ to membranes, and suggesting that the BCR-mediated translocation of Sp-LZ to the plasma membrane happens independently of DAG generation. If this is the case, Sp-LZ should be able to translocate to the plasma membrane of PLCy2_/" cells in response to BCR ligation, despite the defect in DAG generation in these cells. As predicted, ligation of the BCR in PLCy2_/" cells resulted in the translocation of Sp-LZ to the plasma membrane to the same extent as Sp-LZ translocation observed in wild-type cells (Figure 3.9b), indicating that DAG is not required for this process. Altogether, these results point towards the existence of a mechanism of RasGRPI translocation that is distinct from CI-mediated translocation. Furthermore, this is the first indication that RasGRPI is regulated through a mechanism that involves more than just PLCy2 activation. 5 8 Figure 3.9 The region C-terminal to the CI domain translocates to the plasma membrane in response to anti-IgM in a PLCy2- and DAG-independent manner. Figure 3.9. The region C-terminal to the CI domain translocates to the plasma membrane in response to anti-IgM in a PLCy2- and DAG-independent manner. Wild-type DT40 cells (A) and PLCy2-deficient DT40 cells (B) expressing the illustrated C-terminal domains of RasGRPI, each with an N-terminus GFP tag were unstimulated (nil) or stimulated with anti-IgM for 15 minutes or PMA for 5 minutes. Cells were then prepared for fluorescence microscopy and analyzed as described for figure 3.4. 59 3.4.3 The C1 domain cooperates with the Sp-LZ region to increase the efficiency of BCR-induced translocation to the plasma membrane The presence of the CI domain in RG1 ASpLZ is not sufficient to mediate translocation of RasGRPI to the plasma membrane in response to BCR ligation or to DAG treatment. Moreover, the isolated CI domain is not sufficient for plasma membrane localization following BCR ligation and DAG treatment (Figure 3.5). Sp-LZ therefore seems to play a major role in BCR-mediated membrane localization of RasGRPI. Yet, although Sp-LZ provides translocation independently of DAG or the CI domain, the translocation efficiency of Sp-LZ is low compared to the full-length protein (Figure 3.9a). This suggests that Sp-LZ cooperates with the CI domain to provide efficient plasma membrane targeting. Accordingly, my results show that CI-Sp-LZ localizes in the cytoplasm with a slight concentration in inner membranes of resting cells, a pattern similar to the full-length protein (Figure 3.10). Following BCR ligation or DAG treatment, CI-Sp-LZ translocates to the plasma membrane with high efficiency, although slightly lower than the full-length protein. This indicates that Sp-LZ cooperates with the CI domain to achieve plasma membrane localization in response to BCR or DAG stimulation. 60 Figure 3 . 10 The C-terminus construct containing the CI domain and the region C-terminal to it translocates to the plasma membrane in response to anti-IgM and P M A with high efficiency. Figure 3.10. The C-terminus construct containing the CI domain and the region C-terminal to it translocates to the plasma membrane in response to anti-IgM and PMA with high efficiency. DT40 cells expressing the illustrated C-terminal domains of RasGRPI were unstimulated (nil) or stimulated with anti-IgM for 15 minutes or with PMA for 5 minutes. Cells were then prepared for fluorescence microscopy and analyzed as described for figure 3.4. 61 3.4.4 The Leucine Zipper motif contributes to the C1 domain-independent and C1 domain-cooperative translocation mechanisms of R a s G R P I 3.4.4.1 C1-cooperative function of the leucine zipper The Sp-LZ region provides a mechanism for RasGRPI translocation that is independent of the CI domain and DAG, and yet cooperates with the CI domain for BCR- and DAG-induced translocation at the plasma membrane. My next interest was to determine what part of the Sp-LZ region was responsible for CI-independent and cooperative translocation. The Sp-LZ region of RasGRPI contains the leucine zipper (LZ) at its C-terminus end, and the spacer region (Sp) located just N terminal to the LZ. Of these two regions, the leucine zipper motif is the only one that can form a domain of known characteristics. LZ could heterodimerize with a protein that localizes at the plasma membrane. If this is the case, LZ will be essential for translocation in response to BCR ligation. On the other hand, LZ could homodimerize and create more binding opportunities for the regions involved in membrane interaction, increasing membrane avidity. In such case, LZ could enhance or be essential for plasma membrane translocation. Deletion of the leucine zipper from the full-length protein caused a great reduction in BCR- and DAG-mediated translocation of RasGRPI (Figure 3.11), demonstrating the need for the LZ motif in BCR- and DAG-mediated translocation of RasGRPI. The defect harboured by RG1ALZ in BCR-mediated translocation could indicate either that the leucine zipper cooperates with the CI domain to achieve translocation, or that LZ acts independently of the CI domain. The fact that RG1ALZ also harbours a defect in DAG-mediated translocation suggests a role for the leucine zipper in enhancing CI domain functions in translocation. Although there is the possibility for the LZ motif to interact with a DAG-regulated protein, this situation is unlikely, based on the fact that Sp-LZ can translocate to the plasma membrane in a DAG-independent manner. 62 Figure 3 .11 Requirement of the leucine zipper for efficient translocation of RasGRPI to the plasma membrane. RG1 RG1ALZ nil a-IgM D A G Figure 3.11. Requirement of the leucine zipper for efficient translocation of RasGRPI to the plasma membrane. DT40 cells expressing RG1 or RG1ALZ were unstimulated (nil) or stimulated with anti-IgM or DAG for 15 minutes. Cells were prepared for fluorescence microscopy and analyzed as described for figure 3.4. Details regarding the phenotype of RasGRPI-expressing cells stimulated with a-IgM are presented in table 3.1. 3.4.4.2 C1-independent function of the leucine zipper The region C-terminal to the CI domain, Sp-LZ, can translocate to the plasma membrane in response to BCR ligation in a CI - and DAG-independent manner. Sp-LZ also cooperates with the CI domain to confer translocation to the plasma membrane in response to BCR ligation and to DAG. The leucine zipper motif plays a role in enhancing C1 domain-driven translocation in response to BCR ligation. Nonetheless, the cooperativity between LZ and the CI domain in mediating translocation does not exclude the possibility of a role for the LZ motif in the CI-independent translocation mechanism of RasGRPI. To evaluate if the leucine zipper motif can confer membrane translocation in response to BCR ligation, I expressed the LZ motif on its own. The LZ motif showed a localization pattern similar to the full-length protein in resting cells (Figure 3.12a). LZ did not relocalize to any membrane in response to BCR ligation, indicating that the 63 leucine zipper motif is not sufficient to localize to the plasma membrane following BCR ligation (Figure 3.12a). However, the BCR-mediated CI-independent translocation observed in the absence of the CI domain was eliminated when the LZ motif was deleted (Figure 3.12b, RG1 AC 1 vs RG1AC1ALZ), indicating that the leucine zipper is responsible, at least in part, for the CI /DAG-independent translocation mechanism of RasGRPI in response to BCR ligation. Altogether, these results indicate that although the LZ motif is not sufficient to confer CI-independent translocation, it contributes to the CI-independent mechanism of translocation in response to BCR ligation, possibly in cooperation with other domains. Figure 3.12 The leucine zipper is not responsible for but contributes to Cl-independent translocation of RasGRPI. Figure 3.12. The leucine zipper is not responsible for but contributes to CI -independent translocation of RasGRPI. A, DT40 cells expressing the isolated leucine zipper were unstimulated (nil) or stimulated with anti-IgM for 15 minutes. Cells were prepared for fluorescence microscopy and analyzed as described for figure 3.4. B, DT40 cells expressing RG1, RG1AC1 or RG1AC1ALZ were unstimulated (nil) or stimulated with anti-IgM or DAG for 15 minutes. Cells were prepared for fluorescence microscopy and analyzed as described for figure 3.4. 64 3.4.5 The K W E N loop is responsible for C1 domain-independent and leucine zipper-cooperative BCR-mediated translocation of R a s G R P I 3.4.5.1 The spacer region shows C1-independent translocation capacity The LZ motif was shown to cooperate with the CI domain for BCR-induced translocation, and was shown to be in part responsible, but not sufficient, for the C l -independent mechanism of RasGRPI translocation in response to BCR ligation. The fact that RG1 ALZ, but not RG1 ASpLZ, retains some plasma membrane translocation capability suggests a role for Sp in the cooperativity with the CI domain in BCR-mediated translocation. Since LZ is not sufficient for BCR-mediated membrane localization, I evaluated the possibility for a role of the Sp in BCR-mediated C l -independent mechanism of translocation. I expressed the spacer region on its own and determined that in response to BCR ligation, Sp was able to translocate to the plasma membrane, albeit with very low efficiency compared to Sp-LZ (Figure 3.13). Such translocation was found to be DAG-independent, as the plasma membrane localization of Sp did not occur with PMA treatment. Thus, contrarily to the LZ motif, Sp is sufficient to provide translocation to the plasma membrane in response to BCR ligation, and is consequently the part of the C-terminus that provides the CI /DAG-independent mechanism of translocation, which gets greatly enhanced in the presence of the LZ motif. Sp is therefore responsible for the CI-cooperative and the CI-independent functions of RasGRPI, both of which get enhanced by the presence of the leucine zipper motif. 65 Figure 3.13 The spacer is responsible for Cl-independent translocation of RasGRPI to the plasma membrane. Figure 3.13. The spacer is responsible for Cl-independent translocation of RasGRPI to the plasma membrane. DT40 cells expressing the illustrated C-terminal domains of RasGRPI were unstimulated (nil) or stimulated with anti-IgM for 15 minutes or with PMA for 5 minutes. Cells were then prepared for fluorescence microscopy and analyzed as described for figure 3.4. 3.4.5.2 The spacer contains a positive and a negative regulatory sequence for translocation The spacer region, which does not share any sequence homology with a known protein domain or structure, provides the CI/DAG-independent and the CI-cooperative mechanisms of BCR-mediated translocation to the plasma membrane. To achieve a more precise definition of the sequence responsible for this process, a series of increasing deletions were introduced from the N-terminus of the Sp-LZ construct (Figure 3.14a, SpAl-LZ to SpA4-LZ); the choice of this construct was justified by the stronger translocation detected with Sp-LZ in comparison with the translocation detected with the spacer region alone. Deletion up to amino acid 643 did not cause any change in the protein translocation efficiency in response to BCR ligation compared to Sp-LZ (Figure 66 3.14b, SpAl-LZ and SpA2-LZ). Surprisingly, deletion up to amino acid 694 caused high constitutive plasma membrane localization in resting cells, which was further increased in response to BCR ligation (SpA3-LZ). A minimal 18 amino acid segment was identified within the spacer region as sufficient to confer high efficiency translocation in response to BCR ligation (SpA4-LZ). This region was referred to as the KWEN loop, due to the presence of 4 amino acids sharing homology with RasGRP3 and with chicken RasGRPI (Figure 3.14a). The spacer region therefore contains a stretch of 18 amino acids, the KWEN loop, within which a sequence holds the capability for high efficiency plasma membrane targeting. Conversely, the part of the spacer N-terminal to the KWEN loop seems to encompass a negative regulatory sequence (RP), since the presence of this sequence greatly reduces the plasma membrane translocation efficiency of the KWEN loop in conjunction with the LZ motif. Since it has been determined that Sp-LZ translocates to the plasma membrane in response to BCR ligation but in a DAG-independent manner, it was predicted that KWEN-mediated translocation in response to BCR ligation would also be DAG-independent. When PLCy2"A DT40s expressing SpA3-LZ were subjected to BCR ligation, SpA3-LZ translocated to the plasma membrane with the same efficiency as in wild-type cells, demonstrating that DAG generation is not necessary for KWEN-mediated translocation to the plasma membrane (Figure 3.14c). 67 Figure 3.14 Defining the region within the spacer which modulates BCR-induced translocation to the plasma membrane. 596 619 644 4 C1 domain { 'y*. Sp-LZ \-*SpA1-LZ \-*-SpA2-LZ mRasGRPI C K D C Q M H C H K Q C K D L W F E C K K E I K S P A I S T E N I S S V V P M S T L C P L G T K D L L H A P E E G S F I F Q N G E I V D H S E E S K D R T I M L cRasGRPI C K D C G M N C H K Q C K D L V E C K R K S S L C K H E E G F F N G H E S K D R T I M L mRasGRP3 C K D C G N C H K Q C K D L V C P S N S L P F E I L "\^%pA3-LZ \fffiSpA4-LZ A K mRasGRPI L G V S S Q K I S V R L K R W A H K S T f f T E S F P W V G G E T T P G H F V L S S P R K S A Q G A L Y V H S P A S P C P S P A L V R K R A F V K W K M K E S L I CRasGRPI G 3 Q K I S V R L K V H K Q T S G H K Y P S P P S P L RK A V K W E N K S mRasGRP3 ss K I S V R L R T Q T E WV G F P H A F K W E N sadt st \*\.Z , ALZ , mRasGRPI K P K P E L H L R L R T Y O E L E C B I W r L X A D N D A L K I O L i C y A O K K I B S L t ^ G K S N H V L A Q M D H G D S A cRasGRPI K K E H YQELE E N LKADN LK OL A K IESL NHV H D mRasGRP3 a-IgM Figure 3.14. Defining the region within the spacer which modulates BCR-induced translocation to the plasma membrane. A, Sequence of the C-terminal region of murine RasGRPI, from amino acid 571 to the C-terminus (GenBank accession NP 035376). The C-terminal boundary of the CI domain is indicated, as is the potential leucine zipper near the C-terminus. The arrows and the numbered residues show the N-terminal boundaries of the RasGRPI portion of the various proteins whose localizations are shown in panels (B). Below the sequence are the residues in chicken RasGRPI (GenBank accession XP421210) and murine RasGRP3 (GenBank accession NP 997129) which are identical to murine RasGRPI, with holding of the residues which are identical in all three. B, DT40 cells expressing the illustrated C-terminal domains of RasGRPI, were either unstimulated (nil) or stimulated with anti-IgM for 15 minutes. The cells were then prepared for fluorescence microscopy as described for figure 3.4. C, PLCy2-deficient DT40 cells expressing the illustrated RasGRPI domains were unstimulated (nil) or stimulated with anti-IgM for 15 minutes. The cells were then prepared for fluorescence microscopy as described for figure 3.4. 3.4.5.3 The KWEN loop contributes to BCR-mediated plasma membrane localization of RasGRPI The spacer region contains the KWEN loop that provides strong CI - and DAG-independent translocation capabilities, and the RP sequence that dampens this effect. The BCR-mediated CI-independent translocation occurring in the absence of the CI domain is weak, despite the presence of the KWEN loop. This weakness is possibly due to the presence of the RP sequence, which dampens KWEN-mediated plasma membrane translocation. To test this possibility, I expressed a mutant of RasGRPI harbouring a deletion of the CI domain and the RP region, but retaining the KWEN loop and the LZ motif (RG1AC1ARP). If the repressor region prevents the KWEN loop from providing high CI-independent plasma membrane localization, its deletion should increase the efficiency of translocation in response to BCR ligation. In resting cells, the localization pattern of RG1AC1ARP was similar to full-length RasGRPI, except for the fact that some spontaneous localization at the plasma membrane was observed. Following BCR ligation, RG1AC1 ARP translocated to the plasma membrane with high efficiency despite the absence of the CI domain (Figure 3.15). These results demonstrate that the KWEN loop can target RasGRPI to the plasma membrane in response to BCR ligation with high efficiency and in a CI-independent manner, a mechanism that is strongly dampened by the repressor sequence found in the N-terminus of the spacer region. 69 Figure 3.15 Deletion of the repressor region enhances KWEN-mediated Cl-independent translocation of RasGRPI in response to BCR ligation. RGl RG1AC1 RG1AC1ARP Figure 3.15. Deletion of the repressor region enhances KWEN-mediated Cl-independent translocation of RasGRPI in response to BCR ligation. DT40 cells expressing R G l , RG1AC1 or R G l AC1 ARP were unstimulated (nil) or stimulated with anti-IgM or D A G for 15 minutes. Cells were prepared for fluorescence microscopy and analyzed as described for figure 3.4. Details regarding the phenotype of cells stimulated with a-IgM are presented in table 3.1. Although the leucine zipper was found to cooperate with the CI domain to induce RasGRPI translocation to the plasma membrane, it is apparent that additional domains within the C-terminal region are involved. To test if the KWEN loop was responsible for CI cooperativity, a deletion of 20 amino acids spanning the KWEN loop (AK - Figure 3.14a) was introduced in RasGRPI. In resting cells, the localization pattern of RG1AK was similar to the full-length protein (Figure 3.16). Following BCR ligation, RG1AK translocated to the plasma membrane, albeit with lower efficiency compared to full-length RasGRPI. In response to DAG, translocation occurred more efficiently in inner membranes than at the plasma membrane. In addition to testing the effect of a KWEN loop deletion on RasGRPI function, we introduced point mutations in conserved residues of the KWEN loop (Figure 3.14a - italicized residues) (RGlKp). Upon stimulation of the BCR, RGlKp showed a defect in RasGRPI translocation as pronounced as RG1AK, demonstrating that the mutations introduced were potent disruptors of the KWEN function. Again, in response to DAG treatment, these mutations affected plasma 70 membrane translocation of the protein, and caused localization in inner membranes. These results demonstrate the need for a functional KWEN loop in the efficient translocation of RasGRPI to the plasma membrane in response to BCR ligation. The fact that the KWEN loop does not directly localize to the plasma membrane in response to DAG, but is necessary for this process, supports the idea that the CI domain and the KWEN loop cooperate to achieve plasma membrane localization. Binding of the KWEN loop to its ligand could be necessary for stabilizing or increasing CI domain-DAG binding at the plasma membrane, or binding of the CI domain to DAG could act to enhance KWEN-driven translocation. Figure 3.16 The KWEN loop is necessary for BCR- and DAG-mediated translocation of RasGRPI to the plasma membrane. RG1 RG1AK RG1Ku Figure 3.16. The KWEN loop is necessary for BCR- and DAG-mediated translocation of RasGRPI to the plasma membrane. DT40 cells expressing RG1, RG1AK or RGlKp were either unstimulated (nil) or stimulated with anti-IgM or DAG for 15 minutes. Cells were prepared for fluorescence microscopy and analyzed as described for figure 3.4. Details regarding the phenotype of cells stimulated with a-IgM are presented in table 3.1. 71 3.5 CONCLUSION FOR THE MECHANISMS OF R A S G R P I TRANSLOCATION TO THE PLASMA MEMBRANE IN RESPONSE TO B C R ACTIVATION The DAG/C1 hypothesis suggests that in response to BCR ligation, RasGRPI gets recruited to the plasma membrane where it can activate Ras, a process happening solely through the interaction between the CI domain and DAG. Studies of CI-containing proteins have shown that the regulatory function of CI domains can be more complex than a simple CI-DAG interaction, and often involves the contribution from other domains of the protein [reviewed by (Cho 2001; Hall, Lim et al. 2005)]. I demonstrated that the translocation of RasGRPI to the plasma membrane in response to BCR ligation involves more than what the DAG/C1 hypothesis presents. I showed that although the CI domain plays a major role in BCR-mediated translocation of RasGRPI, it is not sufficient for this process, and other domains also play a role in BCR-mediated translocation of RasGRPI. I demonstrated that the region C-terminal to the CI domain, Sp-LZ, cooperates with the CI domain to confer CI/DAG-mediated translocation to the plasma membrane in response to BCR ligation. This region contains a newly identified 18 amino acid sequence responsible for CI-cooperative translocation, called the KWEN loop. In addition to its cooperative function, the KWEN loop is responsible for a BCR-mediated mechanism of translocation that is independent of the CI domain and DAG. A leucine zipper motif that enhances KWEN functions, and a negative regulatory sequence (RP) that dampens KWEN functions, are also found within the region C-terminus to the CI domain. The leucine zipper motif could involve RasGRPI in protein-protein interactions through homo- or heterodimerization, a process that could contribute to translocation in response to BCR ligation. The leucine zipper was shown to enhance the CI-cooperative and the Cl-independent functions of the KWEN loop in BCR-mediated plasma membrane translocation, but dimerization of LZ did not seem to largely contribute to this process. Thus, the results show that the translocation of RasGRPI involves several domains of the protein. Since translocation of GEFs to membrane structures generally lead to their activation, it was then of interest to evaluate whether the contribution of the different domains of RasGRPI to BCR-mediated translocation results in the activation of the protein. 72 CHAPTER 4: MECHANISMS OF R A S G R P I ACTIVATION IN RESPONSE TO BCR LIGATION 4.1 R A S G R P I ACTIVATION IN DT40 CELLS RasGRPI is mainly not associated with membranes in resting DT40 cells, and translocates to the plasma membrane in response to BCR ligation. It is believed that in order to catalyze the GTP exchange on Ras proteins, RasGRP 1 must associate with the plasma membrane or other membrane structures where Ras is located. But is the translocation of RasGRPI to the plasma membrane necessary to lead to its activation? Is plasma membrane localization the only requirement for activation? The DAG-C1 hypothesis states that since the CI domain is the sole regulatory domain responsible for translocation in response to BCR ligation, this same domain is the only regulatory event involved in the activation of RasGRPI in response to BCR ligation. Recently, results have shown that the phosphorylation of RasGRPI by PKCs is also an essential factor to the activation of the protein (Aiba, Oh-hora et al. 2004; Zheng, Liu et al. 2005). My findings revealed that a complex interaction between different domains of RasGRPI, including the CI domain, the KWEN loop and the leucine zipper, regulates the translocation of RasGRPI to the plasma membrane in response to BCR ligation. These findings raise the question whether the contribution of these domains in translocation is also necessary for the activation of RasGRPI at the plasma membrane. If RasGRPI needs to get to the plasma membrane to activate Ras, one would predict that the domains involved in the recruitment of RasGRP 1 to the plasma membrane in response to BCR ligation are also involved in its activation. I assessed the level of Ras activation in the cells using a Ras pull-down assay, a method that takes advantage of the property of Ras to bind its downstream targets only when in the active GTP-bound state. In this assay, the Ras binding domain (RBD) of the effector Raf is fused to a GST protein and used for affinity precipitation of RBD-bound activated Ras, which can then be detected with standard Western blotting techniques (see 73 Figure 4.1). One of the pathways activated downstream of Ras is the MAPK pathway, where Ras binds to the kinase Raf, which activates the kinase MEK, in turn activating the kinase ERK through phosphorylation (Figure 4.2). ERK2 activation has been shown to happen following BCR ligation in DT40 cells (Takata and Kurosaki 1996), and the absence of RasGRPI and RasGRP3 in these cells causes an almost complete elimination of Ras and ERK activation in response to BCR ligation (Oh-hora, Johmura et al. 2003). It has been established by other groups and confirmed with my experiments that in DT40 cells, the kinetics of Ras and ERK activations downstream of the B cell receptor closely correlate, allowing the measurement of ERK phosphorylation as a read out for Ras activation, and thereby RasGRP 1 activation. 74 Figure 4.1 Schematic representation of the activated Ras affinity precipitation. Inactive Ras in fusion protein only binds to activated Ras I The amount of activated Ras retrieved after the pull-down can be detected by Western blot Active Ras Figure 4.1. Schematic representation of the activated Ras affinity precipitation. The Ras binding domain (RBD) of the effector Raf is fused to a GST protein. Raf has the property to bind to activated Ras and not to inactive Ras. Affinity precipitation using Glutathione Agarose beads concentrates the fusion protein as well as its binding partners, which can then be detected by standard Western Blot techniques. 75 Figure 4.2 Representation of the M A P K pathway activated downstream of BCR. Ras-GDP MEK ERK ® - M E K ® - E R K Detection of Ras-GTP and p-ERK as a measure for RasGRPI activation Figure 4.2. Representation of the MAPK pathway activated downstream of BCR. Following BCR ligation, RasGRPI gets activated and leads to GTP loading of Ras. Ras-GTP binds to the kinase Raf, which phosphorylates and activates the kinase MEK, in turn activating the kinase ERK through phosphorylation. The dotted lines represent the possibility of detecting the level of activation of Ras and ERK as a measure of RasGRPI activation. 4.1.1 R a s G R P I gets activated in response to B C R ligation Since the overexpressed RasGRPI protein gets to the plasma membrane in response to BCR ligation, I hypothesized that this translocation would correlate with an increase in Ras and ERK activation following BCR ligation. DT40 cells overexpressing RasGRPI were stimulated with a-IgM and assayed for Ras activation using the pull-down assay 76 and for ERK activation using an a-phospho-ERK antibody. As shown in Figure 4.3a, untransduced cells show a slight increase in the level of Ras activation after BCR stimulation, and overexpression of RasGRPI enhances BCR-induced activation of Ras. Likewise, the level of ERK activation increases following BCR ligation of control cells, and overexpression of RasGRPI enhances this activation approximately 2- to 10-fold depending on the experiment and conditions (Figure 4.3b). When a catalytically inactive form of RasGRPI (RGlGEFu) was expressed in the same way, a further increase in Ras and ERK activation compared to control cells was not observed (Figure 4.3b), indicating that the additional BCR-induced increase in Ras and ERK activation in RasGRP 1-transduced cells is due to the presence of a functional RasGRPI. Figure 4.3c demonstrates that the level of ERK proteins does not vary upon BCR stimulation. These experiments indicate that RasGRPI gets activated downstream of the B cell receptor, and can then activate Ras and ERK. 77 Figure 4 .3 Increased activity of RasGRPI causes an increase in Ras and ERK activation in response to BCR ligation and DAG. a GFP RG1 RG1-GEFu _nil_ RG1 a-IgM 0' 3' 10' 0' 3' 10' transduced N-Ras — endogenous K-Ras endogenous H-Ras "* transduced RasGRPI Figure 4.3. Increased activity of RasGRP 1 causes an increase in Ras and ERK activation in response to BCR ligation and DAG. A, DT40 cells transduced with a retroviral vector expressing HA epitope-tagged wild-type N-Ras, and untransduced (nil) or transduced with RG1, were stimulated with anti-IgM (5 ug/ml) for the indicated times. Activated Ras GTPases were purified by Raf-RBD chromatography and detected by Western blot, using anti-HA to detect the transduced N-Ras and an anti-Ras antibody to detect endogenous K-Ras and H-Ras. B, DT40 cells expressing either GFP as a control, RG1 or RG1 with a mutation inactivating Ras binding (RGl-GEFu) were stimulated with anti-IgM (5 ug/ml) or DAG (100 uM) for the indicated times, and levels of activated ERK were quantified by Western blot. The numbers below the P-ERK blot are the relative quantities of each band. The expression levels of transduced RasGRPI protein in each sample, determined by Western blot with an anti-GFP antibody, are shown in the lower blot. C, An anti-ERK antibody was used to demonstrate that the level of ERK protein does not vary upon stimulation. 4.1.2 R a s G R P I gets activated in response to D A G analogs Because transduced RasGRPI is predominantly at the plasma membrane after BCR ligation, Ras activation via RasGRPI is presumably occurring at the plasma membrane. Thus, plasma membrane translocation of RasGRPI is predicted to be essential for RasGRPI activation. If this is the case, treatment of RasGRPI-expressing cells with DAG should induce the activation of RasGRPI, since it is sufficient for its translocation to the plasma membrane (section 3.2.2). Cells transduced with GFP as a control vector 78 showed that ERK activation was induced by DAG (Figure 4.3b). Upon the same treatment, RasGRPI-expressing cells showed a greater increase in ERK activation compared to control cells, indicating that RasGRPI gets activated downstream of DAG alone. Again, this increase required a functional RasGRPI, as demonstrated by the failure of RGlGEFu, to cause further increase in ERK activation (Figure 4.3b). These observations are in accordance with the model wherein the translocation of RasGRPI to the plasma membrane leads to its activation. 4.2 T H E C 1 DOMAIN-DAG INTERACTION IS NECESSARY FOR R A S G R P I ACTIVATION FOLLOWING B C R LIGATION RasGRPI translocates to the plasma membrane and gets activated in response to BCR ligation. In the same way, treatment of the cells with DAG leads to RasGRPI translocation to the plasma membrane, and to its activation. These results suggest that by binding to DAG, the CI domain is responsible for the translocation of the protein, which leads to the activation of RasGRPI. I demonstrated that in the absence of the CI domain or DAG production, most of RasGRPI translocation to the plasma membrane is eliminated. Although recruitment to the plasma membrane via DAG/C1 domain interaction was shown to require the contribution of other RasGRPI domains, DAG/C1 binding was demonstrated to play a major role in BCR-mediated translocation to the plasma membrane. It is therefore of interest to evaluate the contribution of the CI domain in BCR-mediated activation of RasGRP 1. 4.2.1 The C1 domain is required for BCR-induced activation of R a s G R P I In order to address the role of the CI domain in BCR-mediated activation of RasGRPI, I evaluated the effect of the absence of the CI domain on RasGRPI activation. Since translocation is greatly affected in this mutant, it is expected that the 79 absence of the CI domain will also severely affect the activation of RasGRPI. On the other hand, the remaining translocation that occurs in the absence of the CI domain may be sufficient to activate RasGRPI. In contrast to the full-length protein, overexpression of RGl AC 1 did not cause any increase in BCR-mediated Ras activation compared to untransduced cells (Figure 4.4a), indicating that the CI domain is required for BCR-mediated activation of RasGRPI. Similarly, the RasGRPI-mediated increase in ERK activation in response to BCR ligation was eliminated in the absence of the CI domain (Figure 4.4b). These results suggest that the defect in translocation observed in the absence of the CI domain results in a defect in RasGRPI activation, due to the lack of interaction between the CI domain and DAG. Furthermore, the results indicate that the slight amount of translocation occurring in the absence of the CI domain is not sufficient to provide a detectable level of RasGRPI activation in response to BCR ligation. Figure 4.4 The CI domain and DAG are required for efficient activation of RasGRPI. a b GFP RG1 RG1AC1 oil RG1 RG1AC1 a-IgM 0' 3' 10' 0' 3' 10' 0' 3' 10' transduced N-Ras -» > endogenous K-Ras-» endogenous H-Ras-*' o a-IgM < o a a-IgM < a-IgM < 0' 2' 5' 15' 5' 0' O Q 2' 5' 15' 5' 0' 2' 5' 15' 5' Mfc^^feam^k w i " D M * 31162104197 3 64 38 24 30 1 H 43 23 ^  31 •»-P-ERK transduced RasGRPI PLCy2 DT40 GFP RG1 a-lqM < a a-IgM < 0' 5' 15' 5' 0' 5' 15' 5' -P-ERK Figure 4.4. The CI domain and DAG are required for efficient activation of RasGRPI. A, DT40 cells transduced with a retroviral vector expressing HA epitope-tagged wild-type N-Ras, and untransduced (nil), transduced with RGl or with RG1AC1 were stimulated and detected as described in figure 4.3 (A). B, DT40 cells expressing GFP, RGl or RG1AC1 were stimulated and detected as described in figure 4.3 (B). C, PLCy2-deficient DT40 cells expressing either GFP or RGl were stimulated and analyzed as described for panel (B). 80 4.2.2 R a s G R P I does not get activated in the absence of D A G The requirement for the CI domain in the activation of RasGRPI following BCR ligation is likely due to the fact that the protein cannot interact with DAG in the absence of the CI domain. If this is true, the CI domain should also be required for the activation of RasGRPI following DAG treatment. As shown in Figure 4.4b, deletion of the CI domain completely abrogated RasGRPI activation (RGl A C 1 vs RGl), indicating that the CI domain is necessary for D AG-induced activation of RasGRPI. BCR ligation in P L C Y 2 " / _ cells, which are defective in DAG generation, leads to partial recruitment of RasGRPI to the plasma membrane compared to wild-type cells. Correspondingly, if the requirement for the CI domain in the activation of RasGRPI following BCR ligation is due to the fact that the protein cannot interact with DAG, the absence of DAG should equally impair the capacity for RasGRPI to be activated following BCR ligation. In the absence of PLCy2, ERK activation following BCR ligation was completely eliminated, and overexpression of RasGRPI failed to induce an increase in ERK activation following BCR ligation (Figure 4.4c). The generation of DAG through PLCy2 activation following BCR ligation is therefore necessary to activate RasGRPI. 4.2.3 Artificial targeting of R a s G R P I to the plasma membrane can bypass the need for the C1 domain in BCR-mediated activation of R a s G R P I The failure to activate RasGRPI in the absence of the CI domain or in the absence of PLCy2 despite the occurrence of some plasma membrane translocation could be due to two effects: the low level of translocation could result in an increase in Ras activation that is below the detectable limits of my assays, or RasGRPI could be inactive in the absence of the CI domain, even when localized at the plasma membrane. To test the latter possibility, I expressed a form of RasGRPI from which the entire C-terminal 81 region (CI domain, spacer and LZ motif) was replaced by the K-Ras prenylation signal (RGl/Pren), which constitutively localizes RasGRPI at the plasma membrane. If membrane localization represents the only requirement for RasGRPI activation, RGl/Pren should show constitutive membrane localization and activation. Alternatively, if the CI domain and/or the KWEN loop provide a post-translocation regulatory function in addition to its membrane-targeting role, RGl/Pren should not be sufficient for activation in response to BCR ligation. The results demonstrate that although resting cells showed constitutive membrane localization of RGl/Pren (Figure 4.5a), the level of ERK activation was only slightly elevated (Figure 4.5b), demonstrating that membrane localization of RasGRPI is not the only requirement for its activation, and suggesting the involvement of other signals for RasGRPI activation. When the BCR was stimulated however, RasGRPI activation was induced, leading to the conclusion that the CI domain and the KWEN loop are not needed for the activation of RasGRPI if membrane localization is conferred by a prenylation signal. Failure to activate RasGRPI in resting cells despite the constitutive membrane localization is likely due in part to the requirement for the phosphorylation of RasGRPI by BCR-induced PKC activation. 82 Figure 4.5 BCR ligation is required to activate RasGRPI even with artificial targeting of RasGRPI to the plasma membrane. b GFP RG1 RG1/Pren a-IgM 0' 1' 3' 10' 0' 1' 3' 10' 0' 1' 3' 10' : :, =SSw8'« P~ERK 1 2 21 16 3 19 138136 9 32 113101 " " RasGRPI Figure 4.5. BCR ligation is required to activate RasGRPI even with artificial targeting of RasGRPI to the plasma membrane. A, DT40 cells expressing a prenylated form of RasGRPI (RGl/Pren) were treated as described in figure 3.4. B, DT40 cells expressing GFP, RG1 or RGl/Pren were stimulated with anti-IgM and analyzed as described in figure 4.3 (B). 4.3 K W E N / L Z CAN MEDIATE RASG R P I ACTIVATION The interaction between the CI domain and DAG is necessary for the activation of RasGRPI following BCR ligation. To achieve efficient DAG/C1-driven translocation at the plasma membrane, however, RasGRPI requires the presence of other domains, suggesting that the domains involved in RasGRPI translocation through cooperation with the CI domain could also be required to cooperate with the CI domain for RasGRPI activation. Furthermore, the translocation of RasGRPI to the plasma membrane can be induced efficiently in the absence of the CI domain and the repressor region (RG1AC1ARP). This Cl-independent mechanism of translocation could also represent a Cl/DAG independent mechanism of activation following BCR ligation. 83 4.3.1 K W E N / L Z is necessary for efficient C1/DAG-mediated activation of R a s G R P I in response to B C R ligation The activation of RasGRPI in response to BCR ligation seems to require interaction between the CI domain and DAG at the plasma membrane. Despite the presence of the CI domain in RG1 ASpLZ, this mutant does not relocalize to the plasma membrane following BCR ligation. This raises the question whether the cooperativity provided by the Sp-LZ region in plasma membrane translocation is also necessary for RasGRPI activation, or whether the CI domain is sufficient for the activation of RasGRPI following BCR ligation. Upon BCR ligation, RG1 ASpLZ showed a slight increase in ERK activation compared to control cells (Figure 4.6a), indicating that the C-terminal region is required for efficient BCR-mediated activation of RasGRPI. With DAG treatment, the RG1 ASpLZ mutant localizes at the nuclear membrane, indicating that the CI domain of RasGRPI is capable of interacting with DAG in the absence of Sp-LZ, but that it does not do so at the plasma membrane in response to BCR or DAG stimulation. Treatment with DAG caused a slight increase in ERK activation of RG1 ASpLZ-transduced cells compared to control cells (Figure 4.6b). These results could indicate that the interaction between DAG and the CI domain happening at the nuclear membrane in the absence of Sp-LZ contributes to, but is not entirely responsible for, the activation of RasGRPI. Alternatively, the increase in ERK activation in the absence of Sp-LZ could be attributable to a specific localization of RasGRPI that is not detectable within the limits of my assays. In both possibilities, my results indicate that domains from the C-terminal region cooperate with the CI domain to achieve efficient DAG-mediated RasGRPI activation at the plasma membrane. 84 Figure 4.6 The region C-terminal to the CI domain is necessary for optimum activation of RasGRPI in response to BCR or DAG. GFP RG1 RGIASpLZ O O O q-lgMgj g-lgMci a-IgM g 0' 3' 10' 5' 0' 3' 10' 5' 0' 3' 10' 5' 1 13 8 45 2.36861 1182.42314 74 transduced RasGRPI Figure 4.6. The region C-terminal to the CI domain is necessary for optimum activation of RasGRPI in response to BCR or DAG. DT40 cells expressing GFP, RGl or RGIASpLZ were stimulated and analyzed as described in figure 4.3 (B). 4.3.2 KWEN/LZ-dr iven translocation can confer BCR-mediated activation of R a s G R P I in the absence of the C1 domain While BCR ligation causes RasGRPI to weakly translocate to the plasma membrane in the absence of the CI domain, deletion of the CI domain plus the repressor sequence causes high plasma membrane localization in response to BCR ligation. This raises the possibility that BCR-mediated targeting of RasGRPI to the plasma membrane via KWEN/LZ is sufficient to induce RasGRPI activation. Although RGl AC 1 ARP did not get activated to the same extent as the full-length protein, some activation was observed in response to BCR ligation (Figure 4.7). These results indicate that Cl-independent KWEN/LZ-driven translocation to the plasma membrane can result in RasGRPI activation. However, in the context of wild-type RasGRPI, KWEN/LZ requires the presence of the C1 domain to counteract the effect of the repressor region. 8 5 Figure 4.7 Deletion of the repressor region enhances KWEN-mediated Cl-independent activation of RasGRPI in response to BCR ligation. GFP RG1 RG1AC1ARP a-IgM 0' 3' 10' 0' 3' 10' 0' 3' 10' 1 * 16 2 * 63 2 3 26 — — *m»4 M transduced RasGRPI Figure 4.7. Deletion of the repressor region enhances KWEN-mediated Cl-independent activation of RasGRPI in response to BCR ligation. DT40 cells expressing GFP, RG1 or RG1AC1ARP were stimulated and analyzed as described in figure 4.3 (B). 4.4 THE K W E N LOOP PARTIALLY CONTRIBUTES TO THE ACTIVATION OF R A S G R P I DOWNSTREAM OF THE B C R In further assessing the function of KWEN/LZ, I determined that specific deletion of or point mutations in the KWEN loop severely affected BCR- and DAG-mediated translocation of RasGRPI. The KWEN loop is therefore predicted to be necessary for the efficient activation of RasGRPI in response to BCR ligation or DAG treatment. In response to BCR ligation, RG1AK and RGIKu both showed a partial reduction in ERK activation (Figure 4.8). However, treatment of RG1AK- and RGlKu-expressing cells with DAG led to the complete activation of the mutants. These results indicate that the KWEN loop is necessary for RasGRPI activation in response to BCR ligation. The fact that the KWEN loop is dispensable for DAG-mediated activation but is necessary for plasma membrane localization in response to DAG (Figure 3.16) could indicate that DAG-mediated activation of RG1AK and RGlKp can take place in internal membranes, since treatment with DAG caused RG1 AK and RGIKu to localize in inner membranes. 86 Figure 4.8 The KWEN loop is necessary for optimum activation of RasGRPI in response to BCR ligation, but is dispensable for a response to DAG treatment. a 1 0' 3' 10' 0' 3' 10' 0' 3' 10' 0' 3' 10' M M mm — 1 16 13 2 56 54 2 37 31 2 31 28 -P-ERK transduced GFP RG1 RG1AK RG1KM2 0' 5' 0' 5' 0' 5' 0' 5' U | «IMr -rnmrn «•*» •«-P-ERK 1 30 2 47 2.6 48 2.5 51 K3SoKr 1 Figure 4.8. The KWEN loop is necessary for optimum activation of RasGRPI in response to BCR ligation, but is dispensable for a response to DAG treatment. DT40 cells expressing GFP, RGl, RGl AK or RGlKp were stimulated with anti-IgM (A) and DAG (B) and analyzed as described in figure 4.3 (B). 4.5 THE LEUCINE ZIPPER MOTIF PARTIALLY CONTRIBUTES TO THE ACTIVATION OF RASGRPI DOWNSTREAM OF THE BCR Cl/DAG-driven activation of RasGRPI requires the presence of Sp-LZ to achieve optimum activation at the plasma membrane. Sp-LZ is also required for Cl/D AG-driven translocation of RasGRPI to the plasma membrane, and the leucine zipper motif plays a crucial role in this process by enhancing KWEN-mediated cooperation with the CI domain in translocation. LZ is therefore predicted to be required for optimum activation of RasGRPI in response to BCR ligation. The absence of the LZ motif (RG1ALZ) partially reduced the capacity for RasGRPI to get activated in response to BCR ligation 87 (Figure 4.9), indicating that the LZ motif enhances the KWEN-C1 cooperativity in BCR-mediated activation of RasGRPI. Interestingly, DAG-mediated activation of RG1 was also reduced in the absence of the leucine zipper, supporting a role for CI-domain enhancing functions. Figure 4.9 The leucine zipper is necessary for optimum activation of RasGRPI in response to BCR ligation and DAG treatment. GFP RG1 RG1ALZ CD CD CD q-IgM Q a-IgM Q a-IgM ^ 0' 2' 5' 15' 5' 0' 2' 5' 15' 5' 0' 2' 5' 15' 5' i » « l « » ^ P - E R K 1 31 20 10 18 1.4 62 51 30 53 5 32 38 29 47 H H P M transduced ^ RasGRPI Figure 4.9. The leucine zipper is necessary for optimum activation of RasGRP 1 in response to BCR ligation and DAG treatment. DT40 cells expressing GFP, RG1 or RG1 ALZ were stimulated with anti-IgM or DAG and analyzed as described in figure 4.3 (B). 4.6 CONCLUSIONS ON THE CONTRIBUTION OF DIFFERENT DOMAINS OF R A S G R P I IN PLASMA MEMBRANE TRANSLOCATION AND ACTIVATION The CI-domain and the region C-terminal to the CI domain cooperate to induce plasma membrane translocation and activation of RasGRPI in response to BCR ligation. The CI domain can provide partial activation of RasGRPI in the absence of KWEN/LZ (RG1 ASpLZ). However, unlike the full-length protein, RG1 ASpLZ localizes in internal membranes. It is therefore possible that the mechanism of activation adopted by RG1 ASpLZ differs from the mechanism of activation of RasGRPI, for which the natural 88 site of activation is the plasma membrane. KWEN/LZ is also capable of conferring RasGRPI activation independently of the CI domain. This function however seems to be normally prevented by the repressor region in the full-length protein. The nature of the cooperativity between KWEN/LZ and the CI domain is unknown, but could occur through stabilizing the interaction between the CI domain and its ligand DAG. Additionally or alternatively, the leucine zipper motif could promote homodimerization, resulting in increased avidity between the CI domain and DAG. 89 5.1 C H A P T E R 5: DISCUSSION THE INITIAL MODEL FOR THE REGULATION OF R A S G R P I At the beginning of my PhD, RasGRPI was the only Ras activator of this family that had been identified, and it was shown to be expressed in T and B lymphocytes. RasGRPI was consequently suspected to play an equivalent role in TCR and BCR signaling. My thesis focused on determining the role of RasGRPI in B cell receptor signaling. Several questions were to be answered: "Does RasGRPI get activated in response to BCR ligation? Does BCR-mediated activation of RasGRPI involve membrane localization? Is membrane localization occurring solely through the interaction of the CI domain and DAG?" Several studies suggested that an interaction between the CI domain and membrane-bound DAG molecules was essential for membrane localization, and this interaction was assumed to be the only mechanism of membrane targeting for RasGRPI. My initial results revealed that upon BCR ligation, RasGRP 1 was activated and localized exclusively at the plasma membrane. My results were concordant with the membrane localization and activation being mediated by the interaction between the CI domain and DAG, since BCR-induced plasma membrane localization and activation were prevented either by deleting the CI domain, or by expressing RasGRPI in PLCy2-deficient cells. Furthermore, the CI domain and the portion C-terminal to it could be replaced by a membrane-localizing prenylation signal, and still efficiently lead to the activation of RasGRPI in response to BCR ligation. All these results were in accordance with the current model of RasGRPI activation which links BCR to the activation of PLCy2, leading to the accumulation of DAG at the plasma membrane, followed by the recruitment of RasGRPI to the plasma membrane via its CI domain, thereby leading to RasGRPI activation (Figure 3.7a). 90 5.2 NEW INSIGHTS IN THE MECHANISM OF R A S G R P I TRANSLOCATION TO THE PLASMA MEMBRANE IN RESPONSE TO B C R LIGATION Further examination of the mechanism of plasma membrane translocation and activation of RasGRPI following BCR ligation revealed that the aforementioned model is incomplete. Although the interaction between the CI domain and DAG is necessary in the regulation of RasGRPI, other domains also play major roles in the plasma membrane translocation and activation of RasGRPI. I present here a more extensive alternative model for the mechanism of RasGRPI translocation following BCR ligation. 5.2.1 R a s G R P I Localization in Resting DT40 Cells One limitation of working with mutant proteins and fusion proteins is the fact that improper folding could occur as a result of the modification. In assessing the localization of the RasGRPI protein variants, I assumed that proper protein folding was achieved. This assumption was based on the fact that the localization adopted by the different proteins was different from GFP alone. In addition, some of the proteins demonstrated activity, indicating functionality. It is however important to keep in mind that whereas some of the mutations do not impact the proper protein folding, other mutations may cause improper folding and lead to a misleading phenotype. For example, this could be the case for RG1AC1 ALZ, since this mutant adopts a ubiquitous localization pattern regardless of the condition. On the other hand, the isolated LZ and isolated CI are thought to fold properly, since their localization patterns are different from GFP (in resting cells for LZ, and in PMA-treated cells for CI). My observations show that in resting DT40 cells, RasGRPI has a mainly cytoplasmic localization, with some concentration in internal membranes. The idea that the leucine zipper dictates this localization is supported by the fact that the localization of the isolated leucine zipper is identical to full-length RasGRPI (Figure 3.12), and by the fact that the LZ-deletion mutant loses its nuclear excluded pattern in resting cells (Figure 91 3.11). The involvement of the CI domain in the cytoplasmic and internal membrane localization of RasGRPI is also a possibility. The CI domain however seems to play a minor role in the localization of RasGRPI in resting cells. My results show that deletion of the CI domain from the full-length protein does not alter the nuclear exclusion localization pattern of the protein (Figure 3.6). However, the isolated CI domain weakly localizes in internal membranes of resting cells. Thus in resting DT40s, RasGRPI is kept away from the plasma membrane, apparently via the leucine zipper and the CI domain. 5.2.2 R a s G R P I Localization in Response to B C R Ligation 5.2.2.1 Involvement of DAG and the C1 domain in RasGRPI translocation My results show that the isolated CI domain is not sufficient to localize at the plasma membrane following BCR ligation or DAG treatment, but rather weakly localizes to internal membranes. This is supported by Merida's group, who showed that the CI domain of RasGRPI does not translocate to the plasma membrane in response to TCR ligation (Carrasco and Merida 2004). Although it has been assumed that the site of DAG accumulation following PLCy2 activation is the plasma membrane, it is not clear where DAG accumulates in the cell following BCR activation. If the CI domain represents a probe for the site of DAG accumulation, why isn't the CI domain accumulating at the plasma membrane following BCR ligation? Several potential explanations exist. First, it is possible that the amount of DAG generated at the plasma membrane following BCR ligation is not sufficient to cause the relocalization of the CI domain to the plasma membrane. This, however, is unlikely since treating cells with a high concentration of exogenous DAG does not lead to the localization of the CI domain to the plasma membrane. A second possibility is that the DAG generated in response to BCR ligation may not accumulate at the plasma membrane due to the presence of DAG kinases (DGKs). The rapid conversion of DAG by DGKs could prevent the accumulation of the CI domain at the plasma membrane. In support of this, I showed that in contrast to 92 treatment with DAG, the CI domain is able to localize at the plasma membrane following treatment of cells with PMA, which is not metabolized by DAG kinases. Despite the fact that the CI domain is not sufficient for BCR-mediated translocation to the plasma membrane, my results demonstrate that DAG is involved in the localization of RasGRP 1 at the plasma membrane in response to BCR ligation. First, treating cells with DAG causes RasGRPI to translocate to the plasma membrane in a very similar manner as BCR stimulation. Second, BCR-mediated plasma membrane localization of RasGRPI is not observed following BCR ligation of PLCy2-deficient cells. My results show that DAG binding to the CI domain is unlikely to be the only determinant of plasma membrane localization of RasGRPI in response to BCR ligation. The alternative model for the regulation of RasGRPI proposes that in the context of the full-length protein, binding of the CI domain to DAG provides weak plasma membrane binding, which is insufficient, by itself, to stabilize the RasGRPI/plasma membrane interaction, but can provide a significant enhancement of binding through other domains of RasGRPI. 5.2.2.2 The mechanism of plasma membrane localization of RasGRPI involves the KWEN loop My results demonstrate that the KWEN loop and the leucine zipper are essential to achieve efficient translocation of RasGRPI to the plasma membrane. When the KWEN loop or the leucine zipper are deleted separately (RGl AK or RGl ALZ), the translocation efficiency to the plasma membrane is reduced, and when the KWEN loop and the leucine zipper are both deleted (RGIASpLZ), the translocation of RasGRPI to the plasma membrane is lost. My results point toward the involvement of the leucine zipper and the CI domain in enhancing the stability of the KWEN loop at the plasma membrane. Evidence for this includes the fact that in isolation the KWEN loop, but not the leucine zipper, can achieve BCR-mediated plasma membrane translocation. Nevertheless, KWEN-driven translocation is greatly enhanced in the presence of the leucine zipper, 93 and even further with the CI domain (Sp vs Sp-LZ vs CI-Sp-LZ). BCR-induced KWEN-mediated translocation of RasGRPI to the plasma membrane could be due to an increase in KWEN ligand (KL) at the plasma membrane, or to some modification in RasGRPI or in KL which would allow more efficient binding of KWEN to its ligand. KWEN-mediated membrane translocation seems to be reduced in the presence of the repressor region, based on the fact that in the absence of the repressor, KWEN-mediated translocation to the plasma membrane is increased. This could be due to enhanced binding of KWEN to its ligand in the absence of the repressor. Thus, the stability of the interaction between KWEN and its ligand is reduced by the action of the repressor, but is enhanced by the binding of the leucine zipper at the membrane, as well as by the interaction between DAG and the CI domain. In this model, it is not the binding of one domain to its ligand, but high avidity and multivalency, that are responsible for efficient plasma membrane localization of RasGRPI. Such stable membrane localization can then allow proper phosphorylation of RasGRPI by PKC, and lead to the activation of Ras (Figure 5.1). 94 Figure 5.1 Model for the mechanism of RasGRPI translocation to the plasma membrane in response to B C R ligation. BCR activation BCR interaction between DAG and c1 domain plasma membrane interaction between KWEN and KL possible interaction with protein or lipids internal -membranes Figure 5.1. Model for the mechanism of RasGRPI translocation to the plasma membrane in response to BCR ligation. A, In resting DT40 cells, RasGRPI is kept away from the plasma membrane, presumably through binding for the leucine zipper ( L Z ) to internal membranes, and binding of the C1 domain (C1) to basal levels of DAG in internal membranes. B, Upon BCR ligation, the generation of DAG at the plasma membrane attracts the CI domain to the plasma membrane. Stable membrane localization is however provided by binding of the K W E N loop (K) to its plasma membrane localized ligand (KL). 95 5.3 IMPLICATIONS FOR THE REGULATION OF R A S G R P I 5.3.1 Regulation of RasGRPI via a Single versus Multiple Signals The above model of regulation for RasGRPI implies that in DT40 cells, full activation of RasGRPI requires 2 independent signals generated by BCR: the production of DAG in membranes, and the accessibility of KWEN ligand at the plasma membrane. My results demonstrate this by the fact that neither RG1 ASpLZ which retains the CI domain, nor RG1AC1ARP which retains the KWEN loop, get activated to the same extent as RasGRPI. One possible consequence of the requirement for 2 separate signals in the full activation of RasGRPI is the increase in stringency for the regulation of RasGRP 1. That is, RasGRP 1 could be efficiently activated only by receptors capable of providing these two signals, and not by receptors leading to the generation of DAG only, or to the activation of KL only. Several other receptors can lead to the activation of phospholipases and to the accumulation of DAG in membranes, including G protein-coupled receptors. Thus, in order to provide specificity for BCR signaling, RasGRPI requires additional features that will allow the protein to respond to BCR ligation exclusively. One of these signals is the presence of KL at the plasma membrane, which seems to be triggered by BCR ligation. Because the spectrum of receptors regulating the accumulation of KL at the plasma membrane is not known, I cannot conclude that KWEN-mediated activation of RasGRPI is exclusive to BCR. RasGRPI was also shown to induce the activation of Ras and ERK following TCR activation (Ebinu, Stang et al. 2000), as well as following the stimulation of G protein-coupled receptors, which are coupled to PLCps (Ebinu, Bottorff et al. 1998; Keiper, Stope et al. 2004). Whether RasGRP 1 requires the presence of the KWEN loop to get activated by these receptors remains undetermined. One question that comes to mind is whether under some circumstances only one of these signals is sufficient to trigger the activation of RasGRPI. My data demonstrated that RasGRPI could be activated by BCR in a CI/DAG-independent manner, albeit to a 96 much lesser extent than with dual signals. Interestingly, this KWEN-mediated activation was only allowed in the absence of the repressor region, suggesting a role for the repressor in enforcing the dependence on 2 separate signals. A natural mechanism that inactivates the repressor region could therefore allow RasGRPI to be activated solely through the K W E N / K L signal, in a DAG/C1-independent manner. M y data also demonstrated partial activation of RasGRPI in the absence of the K W E N loop, again to a lesser extent than with multiple signals. DAG /C1 can therefore provide a signal for RasGRPI activation independently of K W E N / K L . The requirement for single versus dual signal in the regulation of RasGRPI could be cell-specific. For example, activity o f RasGRPI in NIH3T3 cells requires only the C I domain, and not the K W E N loop. The involvement of RasGRPI in T C R signaling also involves D A G and the C I domain, but the requirement for other domains has not been fully explored. Whether the involvement of two different signals in the activation of RasGRPI is restricted to B C R signaling has yet to be determined. 5.3.2 The Role of D A G in the Activation of R a s G R P I According to the above model, D A G generated in response to B C R ligation does not provide a signal sufficient to promote the activation of R a s G R P I . However, my results demonstrate that treatment with exogenous D A G leads to the strong translocation and activation of R a s G R P I . Interestingly, the regulation of RasGRPI through exogenous D A G is dependent not only on the C I domain, but also on the presence of K W E N / L Z . There are several signalling events triggered by D A G , but my results do not suggest a role for D A G in KWEN-mediated translocation of RasGRP 1. The requirement for K W E N / L Z in DAG-mediated translocation may reflect the need for this domain in the proper binding of the C I domain to D A G . Evidence from my data suggests that there is a basal level o f K L at the plasma membrane, which increases upon B C R ligation in a PLCy2/D AG-independent manner. The basal level of membrane-bound K L may not be 97 sufficient for stable binding to the plasma membrane when minimal amounts o f D A G are present, but with high amounts of D A G obtained by exogenous D A G treatment, the basal level of K L may be sufficient to provide stable binding of RasGRPI to the plasma membrane. In interpreting my results, I am assuming that the C l / D A G interaction contributes directly to the membrane localization of RasGRPI by providing an additional membrane binding site. However, the possibility of alternative mechanisms should be considered. For example, D A G binding to the C I domain could trigger a conformational change which could cause the release of the repressor function, allowing K W E N to bind K L more efficiently than under repressed state. 5.3.3 The Role of Membrane Localization in the Activation of R a s G R P I The activation of RasGRP 1 involves localizing the protein in membranes where its substrate is available. Which precise membrane structure RasGRPI has to be targeted to is however not clear. M y observations suggest that the natural site of BCR-mediated RasGRPI activation is the plasma membrane, since the full-length protein translocates from the cytoplasm to the plasma membrane in response to B C R ligation. Is the plasma membrane the only site for RasGRPI activation, or can the protein also be activated in internal membranes? M y results demonstrate that the activation of RasGRPI can happen in the absence o f plasma membrane localization. In the absence of K W E N / L Z (RG1 A S p L Z ) , RG1 can be partially activated in response to B C R and D A G stimulations, and this appears to involve localization at internal membranes. This suggests that the amount o f RasGRPI localized in inner membranes can get activated in response to B C R ligation. However, full-length RG1 is depleted from internal membranes, and is almost entirely localized to the plasma membrane following D A G treatments and B C R ligation. 9 8 Therefore, CI/DAG-mediated localization of RasGRPI to internal membranes in BCR stimulated DT40 cells may be physiologically irrelevant. Activation of RasGRPI in sites different from the plasma membrane seems to be possible in other cell types. In Jurkat T cells for example, RasGRPI translocation and Ras activation following TCR engagement were shown to occur at the Golgi apparatus (Bivona, Perez De Castro et al. 2003; Caloca, Zugaza et al. 2003), although other groups reported RasGRPI localization at the plasma membrane in Jurkat cells (Sanjuan, Pradet-Balade et al. 2003). RasGRPI localization and Ras activation were also reported to occur at the plasma membrane and Golgi apparatus in COS-1 cells (Bivona, Perez De Castro et al. 2003; Caloca, Zugaza et al. 2003). When NIH 3T3 cells are grown in serum, RasGRPI localizes in inner membranes. In these cells, RasGRPI causes ERK activation and cell transformation, indicating that RasGRPI can be activated and mediate its functions when localized at inner membranes (Tognon, Kirk et al. 1998). Membrane localization of RasGRPI may not be sufficient for activation. In support of this, I demonstrated that targeting RasGRPI to the plasma membrane with a K-Ras prenylation signal still required BCR stimulation to induce RasGRPI activation, despite the constitutive plasma membrane localization. As an example of additional events required for the activation of RasGRPI, it has previously been demonstrated that phosphorylation of RasGRPI is required for activation (Aiba, Oh-hora et al. 2004; Roose, Mollenauer et al. 2005; Zheng, Liu et al. 2005). Thus, although localizing RasGRPI to membrane structures may be important, other events are required for the activation of RasGRPI. 5.4 IMPLICATIONS FOR THE REGULATION OF OTHER R A S G R P FAMILY MEMBERS Among the other RasGRP family members, RasGRPI shares the highest sequence similarity with RasGRP3. It is, however, not well determined whether these two family 99 members carry out the same functions. RasGRP3 is regulated by phorbol esters (Lorenzo, Kung et al. 2001) and its CI domain is necessary for BCR-mediated activation (Oh-hora, Johmura et al. 2003). In addition, like RasGRPI, the phosphorylation of RasGRP3 by PK.Cs is required for its activation (Teixeira, Stang et al. 2003; Aiba, Oh-hora et al. 2004). Because of these similarities and particularly due to the functional equivalence of their CI domains, the DAG-C1 hypothesis predicts a very similar mechanism of regulation and activation for RasGRPI and RasGRP3. Conversely, the multiple-signal hypothesis that I established does not predict a straightforward parallel between the regulation of the two proteins. RasGRP3 has a KWEN loop, and it is therefore possible that RasGRP3 be regulated via a KWEN-mediated mechanism of translocation. On the other hand, a major difference between RasGRPI and RasGRP3 is the leucine zipper motif, which is present in RasGRPI but not in RasGRP3. Since the leucine zipper plays a major role in the activation of RasGRP 1, it is reasonable to expect a more efficient activation of RasGRPI in response to BCR ligation compared to RasGRP3. In DT40 cells however, Ras activation in response to BCR ligation is mostly attributed to the activity of RasGRP3, and only marginally caused by the action of RasGRPI (Oh-hora, Johmura et al. 2003). This could simply be due to the fact that RasGRP3 is expressed at a higher level compared to RasGRPI. It is also possible that the mechanisms of regulation of RasGRPI and RasGRP3 following BCR ligation are quantitatively different. For example, the KWEN loop of RasGRP3 could bind KL more efficiently than RasGRPI, so that RasGRP3 would have no need for the leucine zipper. Or, RasGRP3 might have a domain capable of compensating for the lack of leucine zipper. One possibility for such a domain lies within the N-terminal region, which contains a consensus sequence that could potentially lead to protein myristylation, a site that is not present in RasGRPI. Covalent attachment of the myristyl lipid group could increase the affinity for membranes and thus provide an alternative means of attaining efficient membrane targeting. Since RasGRP4 lacks both the KWEN homology and the leucine zipper, and also lacks an N-terminal myristylation signal, its regulatory mechanisms must be even more divergent from those of RasGRPI. One splice variant of RasGRP 4 does have a DAG-100 binding CI domain; although the CI-DAG hypothesis could suggest a similar role for RasGRPI and 4 in BCR signaling, the multiple-signal hypothesis would suggest otherwise, due to the crucial role for the LZ and KWEN domains in the regulation of RasGRPI. Our lab is currently investigating the mechanism of RasGRP4 activation. RasGRP2 is the family member that shares the least homology with RasGRPI. In addition to the lack of the LZ and the KWEN loop, RasGRP2 contains a CI domain which differs considerably from the RasGRPI CI domain, in that it binds DAG with very low affinity, if any. The mechanisms of RasGRPI and RasGRP2 regulation are consequently predicted to vary to a large extent. Our lab is also investigating the mechanism of regulation of RasGRP2. When I started my thesis research, all members of the RasGRPI family were believed to be activated solely via a mechanism that involves membrane localization through binding of the CI to domain to DAG. My results indicated that RasGRPI needs the CI domain, the KWEN loop and the leucine zipper motif to get properly activated. The other three RasGRP family members lack one or more of these domains. Their regulation will consequently be distinct from the regulation of RasGRPI activation. They will either be non-activatable by stimuli that activate RasGRPI, or they will utilize unique membrane localizing domains that compensate for their deficiencies in KWEN loop, leucine zipper and/or CI domain. 101 B I B L I O G R A P H Y Aiba, Y., M. Oh-hora, et al. (2004). "Activation of RasGRP3 by phosphorylation of Thr-133 is required for B cell receptor-mediated Ras activation." Proc Natl Acad Sci U S A 101(47): 16612-7. Asnafi, V., K. Beldjord, et al. (2003). "Analysis of TCR, pT alpha, and RAG-1 in T-acute lymphoblastic leukemias improves understanding of early human T-lymphoid lineage commitment." Blood 101(7): 2693-703. Bhandoola, A. and A. Sambandam (2006). "From stem cell to T cell: one route or many?" Nat Rev Immunol 6(2): 117-26. Bivona, T. G., I. Perez De Castro, et al. (2003). "Phospholipase Cgamma activates Ras on the Golgi apparatus by means of RasGRPI." Nature 424(6949): 694-8. Broek, D., T. Toda, et al. (1987). "The S. cerevisiae CDC25 gene product regulates the RAS/adenylate cyclase pathway." Cell 48(5): 789-99. Caloca, M. J., J. L. Zugaza, et al. (2003). "Exchange factors of the RasGRP family mediate Ras activation in the Golgi." J Biol Chem 278(35): 33465-73. Caloca, M. J., J. L. Zugaza, et al. (2003). "Vav mediates Ras stimulation by direct activation of the GDP/GTP exchange factor Ras GRP1." Embo J 22(13): 3326-36. Carrasco, S. and I. Merida (2004). "Diacylglycerol-dependent binding recruits PKCtheta and RasGRPI CI domains to specific subcellular localizations in living T lymphocytes." Mol Biol Cell 15(6): 2932-42. Cho, W. (2001). "Membrane targeting by CI and C2 domains." J Biol Chem 276(35): 32407-10. Clyde-Smith, J., G. Silins, et al. (2000). "Characterization of RasGRP2, a plasma membrane-targeted, dual specificity Ras/Rap exchange factor." J Biol Chem 275(41): 32260-7. Colicelli, J. (2004). "Human RAS superfamily proteins and related GTPases." Sci STKE 2004(250): RE 13. Coughlin, J. J., S. L. Stang, et al. (2005). "RasGRPI and RasGRP3 regulate B cell proliferation by facilitating B cell receptor-Ras signaling." J Immunol 175(11): 7179-84. Crompton, T., K. C. Gilmour, et al. (1996). "The MAP kinase pathway controls differentiation from double-negative to double-positive thymocyte." Cell 86(2): 243-51. Dower, N. A., S. L. Stang, et al. (2000). "RasGRP is essential for mouse thymocyte differentiation and TCR signaling." Nat Immunol 1(4): 317-21. Dupuy, A. J., K. Morgan, et al. (2001). "Activation of the Rapl guanine nucleotide exchange gene, CalDAG-GEF I, in BXH-2 murine myeloid leukemia." J Biol Chem 276(15): 11804-11. Ebinu, J. O., D. A. Bottorff, et al. (1998). "RasGRP, a Ras guanyl nucleotide- releasing protein with calcium- and diacylglycerol-binding motifs." Science 280(5366): 1082-6. Ebinu, J. O., S. L. Stang, et al. (2000). "RasGRP links T-cell receptor signaling to Ras." Blood 95(10): 3199-203. 102 Ehrhardt, A., M. D. David, et al. (2004). "Distinct mechanisms determine the patterns of differential activation of H-Ras, N-Ras, K-Ras 4B, and M-Ras by receptors for growth factors or antigen." Mol Cell Biol 24(14): 6311-23. Guilbault, B. and R. J. Kay (2004). "RasGRPI sensitizes an immature B cell line to antigen receptor-induced apoptosis." J Biol Chem 279(19): 19523-30. Hagman, J. and K. Lukin (2006). "Transcription factors drive B cell development." Curr Opin Immunol 18(2): 127-34. Hall, C , L. Lim, et al. (2005). "CI, see them all." Trends Biochem Sci 30(4): 169-71. Hurley, J. H., A. C. Newton, et al. (1997). "Taxonomy and function of CI protein kinase C homology domains." Protein Sci 6(2): 477-80. Irie, K., A. Masuda, et al. (2004). "Tumor promoter binding of the protein kinase C CI homology domain peptides of RasGRPs, chimaerins, and Uncl3s." Bioorg Med Chem 12(17): 4575-83. Iritani, B. M., J. Alberola-Ila, et al. (1999). "Distinct signals mediate maturation and allelic exclusion in lymphocyte progenitors." Immunity 10(6): 713-22. Iritani, B. M., K. A. Forbush, et al. (1997). "Control of B cell development by Ras-mediated activation of Raf." Embo J 16(23): 7019-31. Jones, D. R., M. A. Sanjuan, et al. (2002). "Expression of a catalytically inactive form of diacylglycerol kinase alpha induces sustained signaling through RasGRP." Faseb J 16(6): 595-7. Kang, J. H., S. Benzaria, et al. (2006). "Conformationally constrained analogues of diacylglycerol. 26. Exploring the chemical space surrounding the CI domain of protein kinase C with DAG-lactones containing aryl groups at the sn-1 and sn-2 positions." J Med Chem 49(11): 3185-203. Kawasaki, H., G. M. Springett, et al. (1998). "A Rap guanine nucleotide exchange factor enriched highly in the basal ganglia." Proc Natl Acad Sci U S A 95(22): 13278-83. Keiper, M., M. B. Stope, et al. (2004). "Epac- and Ca2+ -controlled activation of Ras and extracellular signal-regulated kinases by Gs-coupled receptors." J Biol Chem 279(45): 46497-508. Kim, R., A. Trubetskoy, et al. (2003). "Genome-based identification of cancer genes by proviral tagging in mouse retrovirus-induced T-cell lymphomas." J Virol 77(3): 2056-62. Klinger, M. B., B. Guilbault, et al. (2005). "Deregulated expression of RasGRPI initiates thymic lymphomagenesis independently of T-cell receptors." Oncogene 24(16): 2695-704. Layer, K., G. Lin, et al. (2003). "Autoimmunity as the consequence of a spontaneous mutation in Rasgrpl." Immunity 19(2): 243-55. Li, L., Y. Yang, et al. (2002). "Cloning of rat Ras guanine nucleotide releasing protein 4, and evaluation of its expression in rat mast cells and their bone marrow progenitors." Mol Immunol 38(16-18): 1283-8. Lorenzo, P. S., M. Beheshti, et al. (2000). "The guanine nucleotide exchange factor RasGRP is a high -affinity target for diacylglycerol and phorbol esters." Mol Pharmacol 57(5): 840-6. Lorenzo, P. S., J. W. Kung, et al. (2001). "Phorbol esters modulate the Ras exchange factor RasGRP3." Cancer Res 61(3): 943-9. 103 Madani, S., A. Hichami, et al. (2004). "Diacylglycerols containing Omega 3 and Omega 6 fatty acids bind to RasGRP and modulate MAP kinase activation." J Biol Chem 279(2): 1176-83. Mombaerts, P., J. Iacomini, et al. (1992). "RAG-1-deficient mice have no mature B and T lymphocytes." Cell 68(5): 869-77. Mor, A. and M. R. Philips (2006). "Compartmentalized Ras/MAPK signaling." Annu Rev Immunol 24: 771-800. Norment, A. M., L. Y. Bogatzki, et al. (2003). "Transgenic expression of RasGRPI induces the maturation of double-negative thymocytes and enhances the production of CD8 single-positive thymocytes." J Immunol 170(3): 1141-9. O'Shea, C. C , T. Crompton, et al. (1996). "Raf regulates positive selection." Eur J Immunol 26(10): 2350-5. Oh-hora, M., S. Johmura, et al. (2003). "Requirement for Ras guanine nucleotide releasing protein 3 in coupling phospholipase C-gamma2 to Ras in B cell receptor signaling." J Exp Med 198(12): 1841-51. Ohba, Y., N. Mochizuki, et al. (2000). "Regulatory proteins of R-Ras, TC21/R-Ras2, and M-Ras/R-Ras3." J Biol Chem 275(26): 20020-6. Ollila, J. and M. Vihinen (2005). "B cells." Int J Biochem Cell Biol 37(3): 518-23. Paduch, M., F. Jelen, et al. (2001). "Structure of small G proteins and their regulators." Acta Biochim Pol 48(4): 829-50. Pear, W. S., G. P. Nolan, et al. (1993). "Production of high-titer helper-free retroviruses by transient transfection." Proc Natl Acad Sci U S A 90(18): 8392-6. Pelanda, R. and R. M. Torres (2006). "Receptor editing for better or for worse." Curr Opin Immunol 18(2): 184-90. Perez de Castro, I., T. G. Bivona, et al. (2004). "Ras activation in Jurkat T cells following low-grade stimulation of the T-cell receptor is specific to N-Ras and occurs only on the Golgi apparatus." Mol Cell Biol 24(8): 3485-96. Priatel, J. J., S. J. Teh, et al. (2002). "RasGRPI transduces low-grade TCR signals which are critical for T cell development, homeostasis, and differentiation." Immunity 17(5): 617-27. Quilliam, L. A., J. F. Rebhun, et al. (2002). "A growing family of guanine nucleotide exchange factors is responsible for activation of Ras-family GTPases." Prog Nucleic Acid Res Mol Biol 71: 391-444. Rambaratsingh, R. A., J. C. Stone, et al. (2003). "RasGRPI represents a novel non-protein kinase C phorbol ester signaling pathway in mouse epidermal keratinocytes." J Biol Chem 278(52): 52792-801. Reuther, G. W., Q. T. Lambert, et al. (2002). "RasGRP4 is a novel Ras activator isolated from acute myeloid leukemia." J Biol Chem 277(34): 30508-14. Reynolds, L. F., C. de Bettignies, et al. (2004). "Vavl transduces T cell receptor signals to the activation of the Ras/ERK pathway via LAT, Sos, and RasGRPI." J Biol Chem 279(18): 18239-46. Rong, S. B., I. J. Enyedy, et al. (2002). "Structural basis of RasGRP binding to high-affinity PKC ligands." J Med Chem 45(4): 853-60. Roose, J. P., M. Mollenauer, et al. (2005). "A diacylglycerol-protein kinase C-RasGRPl pathway directs Ras activation upon antigen receptor stimulation of T cells." Mol Cell Biol 25(11): 4426-41. 104 Sanjuan, M. A., B. Pradet-Balade, et al. (2003). "T cell activation in vivo targets diacylglycerol kinase alpha to the membrane: a novel mechanism for Ras attenuation." J Immunol 170(6): 2877-83. Saxton, T. M., I. van Oostveen, et al. (1994). "B cell antigen receptor cross-linking induces phosphorylation of the p21ras oncoprotein activators SHC and mSOSl as well as assembly of complexes containing SHC, GRB-2, mSOSl, and a 145-kDa tyrosine-phosphorylated protein." J Immunol 153(2): 623-36. Shaw, A. C , W. Swat, et al. (1999). "Activated Ras signals developmental progression of recombinase-activating gene (RAG)-deficient pro-B lymphocytes." J Exp Med 189(1): 123-9. Shinkai, Y., G. Rathbun, et al. (1992). "RAG-2-deficient mice lack mature lymphocytes owing to inability to initiate V(D)J rearrangement." Cell 68(5): 855-67. Swan, K. A., J. Alberola-Ila, et al. (1995). "Involvement of p21ras distinguishes positive and negative selection in thymocytes." Embo J 14(2): 276-85. Swat, W., Y. Shinkai, et al. (1996). "Activated Ras signals differentiation and expansion of CD4+8+ thymocytes." Proc Natl Acad Sci U S A 93(10): 4683-7. Takahama, Y. (2006). "Journey through the thymus: stromal guides for T-cell development and selection." Nat Rev Immunol 6(2): 127-35. Takata, M., Y. Homma, et al. (1995). "Requirement of phospholipase C-gamma 2 activation in surface immunoglobulin M-induced B cell apoptosis." J Exp Med 182(4): 907-14. Takata, M. and T. Kurosaki (1996). "A role for Bruton's tyrosine kinase in B cell antigen receptor-mediated activation of phospholipase C-gamma 2." J Exp Med 184(1): 31-40. Taylor, S. J., R. J. Resnick, et al. (2001). "Nonradioactive determination of Ras-GTP levels using activated ras interaction assay." Methods Enzvmol 3 3 3 : 333-42. Teixeira, C , S. L. Stang, et al. (2003). "Integration of DAG signaling systems mediated by PKC-dependent phosphorylation of RasGRP3." Blood 102(4): 1414-20. Tognon, C. E., H. E. Kirk, et al. (1998). "Regulation of RasGRP via a phorbol ester-responsive CI domain." Mol Cell Biol 18(12): 6995-7008. Topham, M. K. and S. M. Prescott (2001). "Diacylglycerol kinase zeta regulates Ras activation by a novel mechanism." J Cell Biol 152(6): 1135-43. Vassilopoulos, D., B. Kovacs, et al. (1995). "TCR/CD3 complex-mediated signal transduction pathway in T cells and T cell lines from patients with systemic lupus erythematosus." J Immunol 155(4): 2269-81. Vinson, C , M. Myakishev, et al. (2002). "Classification of human B-ZIP proteins based on dimerization properties." Mol Cell Biol 22(18): 6321-35. Yamashita, S., N. Mochizuki, et al. (2000). "CalDAG-GEFIII activation of Ras, R-ras, and Rapl." J Biol Chem 275(33): 25488-93. Yang, C. and M. G. Kazanietz (2003). "Divergence and complexities in DAG signaling: looking beyond PKC." Trends Pharmacol Sci 24(11): 602-8. Yang, Y., L. Li, et al. (2002). "RasGRP4, a new mast cell-restricted Ras guanine nucleotide-releasing protein with calcium- and diacylglycerol-binding motifs. Identification of defective variants of this signaling protein in asthma, mastocytosis, and mast cell leukemia patients and demonstration of the 105 importance of RasGRJM in mast cell development and function." J Biol Chem 277(28): 25756-74. Yeoh, E. J., M. E. Ross, et al. (2002). "Classification, subtype discovery, and prediction of outcome in pediatric acute lymphoblastic leukemia by gene expression profiling." Cancer Cell 1(2): 133-43. Zhang, G., M. G. Kazanietz, et al. (1995). "Crystal structure of the cys2 activator-binding domain of protein kinase C delta in complex with phorbol ester." Cell 81(6): 917-24. Zheng, Y., H. Liu, et al. (2005). "Phosphorylation of RasGRP3 on threonine 133 provides a mechanistic link between PKC and Ras signaling systems in B cells." Blood 105(9): 3648-54. Zhong, X. P., E. A. Hainey, et al. (2002). "Regulation of T cell receptor-induced activation of the Ras-ERK pathway by diacylglycerol kinase zeta." J Biol Chem 277(34): 31089-98. 106 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0100321/manifest

Comment

Related Items