Thesis Prospectus O. Fisher Structural And Functional .

7m ago
1.33 MB
10 Pages

Thesis ProspectusO. FisherStructural and Functional Studies of Cerebral Cavernous Malformations ProteinsSPECIFIC AIMS:Cerebral cavernous malformation (CCM) is a relatively common condition characterized by vascular lesions inthe brain caused by thinning of the endothelial lining of blood vessels. It affects approximately 0.5% of thepopulation and is caused by mutations within three genes: CCM1, CCM2, and CCM3. The mechanisms bywhich mutations in the CCM genes lead to disease are not completely determined, partly because their normalphysiologic roles have yet to be fully described. Although CCM1, CCM2, and CCM3 interact to form a tripartitecomplex1, each of the three individual CCM proteins also function in various cellular pathways. CCM2 consistsof an N-terminal phosphotyrosine binding (PTB) domain and a C-terminal domain of unknown structure andfunction. CCM2 has been implicated in a variety of signaling pathways, including TrkA-mediated apoptoticsignaling2,3, regulation of osmolarity through MAP kinase cascades4,5, and degradation of the GTPaseRhoA3,4,6. The precise mechanisms by which CCM2 functions within these roles are not fully understood. Inanother example of the involvement of CCM proteins in signaling pathways, recent studies have indicated thatCCM3 binds directly to STK25, STK24, and MST4, the three kinases in the germinal center kinase III (GCKIII)subfamily of sterile-20 kinases7,8. Although CCM3 and GCKIII proteins have been observed to heterodimerizewith one another, the structural details of this interaction and its effect on the enzymatic activity of the GCKIIIshave not been entirely characterized8. I hypothesize that the CCM proteins function as adaptors thatregulate signaling pathways through specific binding interactions. I plan to address this hypothesisthrough two complementary approaches: determining the structure and function of the C-terminal domain ofCCM2 and elucidating the structural and biochemical nature of the CCM3-GCKIII interaction.Aim 1: Determine the structure and function of the C-terminal domain of CCM2.CCM2 is the hub of the CCM complex. It is believed to contain a phosphotyrosine binding (PTB) domain at itsN-terminus and a C-terminal domain (CTD) of potentially novel fold. The PTB domain of CCM2 might directlyinteract with NPXY/F motifs on CCM19, while an LD-like repeat region C-terminal to the PTB domain interactswith CCM3 (unpublished data). Other binding partners of CCM2 outside of the CCM complex have also beenidentified. These include the MAP kinase MEKK34, the receptor tyrosine kinase TrkA2, the E3 ubiquitin ligaseSmurf13, and the small GTPase RhoA3. For many of these proteins, the region of CCM2 with which theyinteract has not been characterized, nor have the effects of these protein-protein interactions on theirrespective cellular processes been studied. In particular, the function of the CCM2 CTD remains almostentirely unknown, although it has been hinted that it might mediate CCM2’s interaction with MEKK34. Theexperiments described in this aim will seek to elucidate the structure and function of the C-terminal region ofCCM2. A structurally stable construct of the CCM2 CTD will be identified using bioinformatics and limitedproteolysis. This construct will then be purified and crystallized to determine its structure using X-raycrystallography. Then, known CCM2 binding partners will be tested for their ability to bind to this region of theprotein. Finally, the structural and biochemical effects of the CCM2 CTD binding to its interaction partners willbe determined.Aim 2: Elucidate the structural and biochemical nature of the CCM3-GCKIII interaction.CCM3 consists of two domains: an N-terminal domain that mediates homodimerization and a C-terminal focaladhesion targeting (FAT) homology domain10. In addition to its role in the CCM complex, CCM3 has also beenshown to interact with members of the GCKIII family of serine/threonine kinases8. GCKIII kinases arecomprised of an N-terminal kinase domain followed by a C-terminal regulatory region. The dimerizationdomain of CCM3 and the C-terminus of the GCKIII proteins interact with one another to form heterodimers8.Although all three members of the GCKIII family bind to CCM3, only STK25 has been shown to phosphorylateit11. The function of this interaction, however, has not been addressed, though CCM3 might regulate theactivity of the kinase7. Therefore, the experiments in this aim are designed to determine whether and howCCM3 regulates the activity of STK25 through structural and biochemical studies. To this end, a co-crystalstructure of CCM3 in complex with a GCKIII protein will be determined. Additionally, kinase activity assays willbe carried out in vitro using a GCKIII consensus substrate peptide to determine whether its interaction withCCM3 has a direct effect on STK25 kinase activity. Finally, the complex will be studied in cellular assays todetermine whether the activity of the kinase, the phosphorylation of CCM3, or the formation of the CCM3STK25 complex is required for cellular localization.1

Thesis ProspectusO. FisherBACKGROUND AND SIGNIFICANCE:Cerebral Cavernous Malformations: Disease and GeneticsMutations within the CCM proteins have been linked to cerebral cavernous malformations, a vascular disorderthat affects approximately 1 in 200 individuals, with a higher prevalence in the Hispanic population due to afounder mutation12. CCM lesions predominantly occur in the brain, but can also occur in the retina, skin, andspinal cord13. They are characterized by several cavernous channels forming within one endothelial layer,resulting in a structure resembling a mulberry14. Although these lesions are often asymptomatic, if theyrupture, patients can experience hemorrhage, stroke, and other neurological effects.Acquisition of cerebral cavernous malformations can be either genetic or sporadic. Patients with the geneticform usually have several lesions, while in sporadic cases there typically is only one15, suggesting that thegenetic form of CCM is caused by a two-hit mechanism16. Analysis of CCM patients has identified three geneswhose mutation causes the disease. These are CCM1 (Krit 1)17, CCM2 (OSM; malcavernin)18,19, and CCM3(PDCD10)20. Most of the mutations that have been identified within these genes are loss of function, thoughthere is a point mutation in the PTB domain of CCM2 and a deletion mutation in CCM3 that are clinicallysignificant11,21.The three CCM proteins are well conserved through evolution, and both mouse and zebrafish models haveproven to be tractable systems for studying their function. Although the animal models do not all show cerebraleffects, knocking out and knocking down these proteins results in vascular phenotypes that have similarendothelial effects to those observed in human patients. For example, loss of function in santa and valentine,the zebrafish homologs of CCM1 and CCM2, cause the heart walls of the animals to have a thin single layer ofcells reminiscent of the walls of the brain lesions seen in human patients22. Additionally, knockdown of thezebrafish ccm3 homolog showed a similar phenotype11. Both CCM1 and CCM2 knockout mice die as embryosdue to problems in arterial development6, and endothelial-specific CCM2 knockout mice die about 11 days intoembryonic development23. Mouse models of CCM3 have a slightly different phenotype, as these animalsexhibit enlarged veins rather than arterial developmental abnormalities24. This suggests that CCM2 and CCM1might function through different mechanisms than CCM3. The current understanding of the biological roles ofthe CCM proteins is incomplete, but recently some of this information has begun to be elucidated.Architecture of the CCM Proteins and their Interactions with One AnotherThe three CCM proteins have been observed to interact with oneanother to form a complex, with CCM2 forming the hub1 (Fig. 1). CCM1contains three NPXY/F motifs at its N-terminus, an ankyrin repeatdomain, and a 4.1 protein/ezrin/radixin/moesin (FERM) domain at its Cterminus25. The CCM2 protein consists of a putative phosphotyrosinebinding (PTB) domain at its N-terminus18 and a C-terminal region thathas been suggested to adopt a novel three-dimensional fold2. Thecrystal structure of CCM3 revealed that it consists of an N-terminaldimerization domain and C-terminal focal adhesion targeting (FAT)homology domain10.The PTB domain of CCM2 is believed to interact with the N-terminus ofCCM1, likely by binding to the latter two of CCM1’s three NPXY/F,motifs9 26. A CCM3 disease mutant lacking part of its dimerizationdomain can bind to CCM2, indicating that CCM3’s FAT homologyFigure 1: The CCM Complex. CCM1domain mediates the CCM3-CCM2 interaction11. Additionally, an LDinteracts with the PTB domain of CCM2 viaits NPXF motifs while the LD-like motif inlike motif slightly C-terminal to the PTB domain of CCM2 appears toCCM2 interacts with the FAT-homologybind to the CCM3 FAT homology domain (unpublished data).domain of CCM3.The Role of the CCM Proteins in Cellular Signaling PathwaysAll three CCM proteins have been found to function in signaling pathways, likely mediated through scaffoldinginteractions as they lack enzymatic activity. CCM1 interacts with several proteins involved in cytoskeletalregulation such as ICAP-1 and Rap126,27. The functions of CCM2 include binding to proteins in the MAPkinase pathway4, playing a role in apoptotic signaling of the receptor tyrosine kinase TrkA2, and promoting thedegradation of the small GTPase RhoA3 (Table 1). Some examples of the role of CCM3 in signaling include itsability to bind to the LD repeats of paxillin via its FAT homology domain28, function as part of the STRIPAKprotein kinase and phosphatase complex29, and interact with and potentially regulate members of the GCKIIIfamily of kinases7.CCM2 functions in degradation of RhoA, TrkA-mediated cell death, and osmoregulation.2

Thesis ProspectusO. FisherBindingRegion ofCCM2 facilitates the degradation of the RhoAFunctionTypePartnerCCM2GTPase. Pull-down and co-IP experiments26CCM1ScaffoldingPTB domainCCM Complexidentified RhoA, Rac1, and Smurf1 as direct306,4FormationCCM3ScaffoldingLD motifbinding partners for CCM2 . Because Smurf1 is6RhoASmallGTPaseUnknownan E3 ubiquitin ligase that targets RhoA, it isperhaps through this mechanism that CCM26RhoA DegradationRac1Small GTPaseUnknownpromotes RhoA degradation3. This pathway might3Smurf1E3 ligasePTB domainalso require an interaction between CCM1 and31MEKK3MAP kinaseC-terminal?OsmosensingCCM2, as cells expressing only a mutant of CCM2MKK3MAP kinaseUnknownwith reduced CCM1 binding do not exhibit an effectTrkA-induced cellReceptor tyrosine322on RhoA activity .TrkAPTB domaindeathkinase30CCM2 has also been identified as an interactionSTK25Ste-20 kinaseUnknown31partner for TrkA, a receptor tyrosine kinase thatEF1A1Elongation factorPTBcauses prosurvival signaling of neurotrophins in the Other31ActinCytoskeletonUnknownnervous system and induces tumor cell death in31PtdInsPhospholipidUnknownneuroblastoma2. The CCM2 PTB domain bindsdirectly to the juxtamembrane region of TrkA, and Table 1: CCM2 interacting proteins. Proteins that have beenshown to interact directly with CCM2 are grouped by function andits C-terminal domain is required for signaling tothe CCM2 binding region, if known, is listed.downstream cell death pathways through anunknown mechanism2.One of the most extensively characterized roles of CCM2 is that of its function in osmoregulation. A yeast twohybrid screen and subsequent confirmation by co-IP and micro-FRET identified CCM2 as an interactionpartner of MEKK3, a kinase that is part of the mammalian osmolarity regulation pathway4. CCM2 is believed torecruit MEKK3 to membrane ruffles where it induces the activation of p38 in response to osmotic stress4.CCM3 Interacts with GCKIII KinasesThe germinal center kinase III (GCKIII) proteins are a family of Sterile 20-like kinases that consist of an Nterminal kinase domain followed by a C-terminal regulatory region. There are three proteins within the family:STK25 (Ysk1; Sok1), STK24 (MST3), and MST4 (MASK). These kinases are activated throughautophosphorylation33,34 of a critical threonine residue within the activation loop of the kinase domain35.All three members of the GCKIII family have been shown to interact directly with CCM3. In vitro bindingstudies showed the C-terminal region of the kinase and the N-terminal dimerization domain of CCM3 mediatethe formation of GCKIII-CCM3 heterodimers8,36 (Fig. 2). The respective affinities of the CCM3 homodimer, theMST4 homodimer, and the heterodimer are 2.7µM, 2.5µM, and 0.1µM as measured by analytical ultracentrifugation8. Though all three GCKIII kinases can interact with CCM3,STK25 alone appears to be able to phosphorylate it at Ser39 and Thr4311,30.The CCM3-GCKIII interaction appears to be physiologically relevant, becausea CCM3 disease mutant cannot bind either STK25 or MST411. The kinasesappear to be activated by oxidative stress, and STK25 in particular has beenshown to play a dual role depending on the cellular environment. Undernormal conditions, it is localized to the Golgi where it interacts with GM130, aGolgi matrix protein35. Under conditions of oxidative stress, caspases cleaveFigure 2: CCM3-GCKIIIthe C-terminal region from the kinase domain, allowing the kinase domain toInteractions. GCKIII kinases are37translocate to the nucleus . It appears that the kinase activity of STK25 is not believed to form heterodimers withnecessary for nuclear localization37. Like STK25, CCM3 has also been shown CCM3, although the precise natureto localize to the Golgi of cells36, so perhaps CCM3 plays a role in regulatingof this binding is unknown.the cellular localization of STK25.Significance:The biological importance of the CCM proteins has been demonstrated through their role in disease and thephenotypes of transgenic animal models. Despite the currently known information regardin