SUMO and Ubiquitin;The Yin and Yang of IGF-1R Function

Bita Sehat

Thesis for doctoral degree (Ph.D.)Stockholm 2007Bita Sehat

SUM

O and U

biquitin; The Yin and Yang of IG

F-1R Function

Popular scientific summery

Cancer develops when cells in the body begin to grow uncontrollably. Normalbody cells grow, divide, and die in an orderly fashion. During the early yearsof a person’s life, normal cells divide rapidly until the person becomes anadult. After that, cells in most parts of the body divide only to replaceworn-out or dying cells and to repair injuries. Cancer cells, on the otherhand, continue to grow and divide and instead of dying, they outlive normalcells and continue to form new abnormal cells. Often, cancer cells travel toother parts of the body where they begin to grow. This process is calledmetastasis. Regardless of where a cancer may spread to, it is always namedafter the place it began. For instance, breast cancer that spreads to the liveris still called breast cancer, not liver cancer. Not all tumors are cancerous.Benign, non-cancerous tumors do not spread (metastasize) to other partsof the body, and with very rare exceptions, are not life threatening. Thereare many reasons why cancer cells grow faster than normal cells. Usuallythe mechanism used by cancer cells for abnormal growth already exists innormal cells, but they have lost control over them. On the surface of normalcells there are proteins called receptors. Their job is to communicate withneighboring cells, and if needed, to send a message (signal) into the celltelling molecules that are responsible for growth that the cell must grow.The receptors on cancer cells, on the other hand, send false signals into thecell, and as a result, the cells grow more than they should and they do not die.I have worked with one of these receptors, called IGF-1R, and investigatedits relationship with two other proteins, called ubiquitin and SUMO. One ofthe many tasks of ubiquitin is to control the receptors, and if necessary, todestroy them. However, under some conditions not only will ubiquitin stopdestroying the receptors but it will start helping them to send growth signalsinto the cell. We know much less about SUMO. We believe it regulatesgrowth and survival of tumor cells. Better understanding of the mechanismsbehind IGF-1R and its interactions with ubiquitin and SUMO will hopefullyhelp us find some of the key molecules important for uncontrolled cell growthand prolonged cell survival of cancer cells.

i

ii

Abstract

Tumorigenesis is a multistep process involving genetic and epigenetic alter-ations that drive the progressive transformation of normal human cells intohighly malignant derivatives. The insulin-like growth factor 1 receptor (IGF-1R) plays pivotal roles in cancer. Based on the traditional view, IGF-1Rmediates its biological responses primarily through receptor phosphorylationand signaling. The subsequent termination of signaling occurs via receptorendocytosis and degradation. Over the last years we have begun to appreci-ate the biological roles of IGF-1R ubiquitination.

In this thesis we established that IGF-1R ubiquitination is dependent onreceptor phosphorylation and integrity of the C-terminal domain. We identi-fied β-arrestin as an adapter protein essential for congregating IGF-1R and itsligase MDM2. The β-arrestin/MDM2-mediated IGF-1R ubiquitination wascrucial for receptor degradation and also for IGF-1-mediated ERK activationand cell cycle progression. Further, Cbl was recognized as a new ligase forIGF-1R inducing Lys48 poly-ubiquitination of the receptor, whereas MDM2modified the receptor with Lys63 ubiquitin chain. Cbl-mediated ubiquitina-tion caused caveolin lipid-raft internalization of the receptor, while MDM2-mediated ubiquitination caused clathrin dependent internalization.

Moreover, we identified IGF-1R as a novel target for SUMOylation. Lig-and dependent SUMO-modification of IGF-1R promoted nuclear transloca-tion of the receptor. Electrophoretic mobility shift assay and chromatinimmunoprecipitation-based cloning studies demonstrated binding of IGF-1Rto DNA and identified DNA binding regions. By cloning the obtained se-quences upstream of a luciferase reporter, we found an IGF-1R induced en-hancement of gene transcription. Altogether, Our data suggests that nuclearsequestered IGF-1R activates transcription by binding to enhancer regions.

In conclusion, this thesis emphasizes the mechanisms by which ubiquiti-nation of IGF-1R affects signaling and degradation. Furthermore, this thesisshows the importance of various ligases, orchestrating trafficking of the re-ceptor. We also describe the novel involvement of IGF-1R SUMOylation inpromoting the translocation of cell surface IGF-1R to the nucleus. Detailedunderstanding of how these processes are controlled under physiological andpathological conditions may be important for future therapeutic approaches.

iii

iv

List of Publications

I. Girnita L, Shenoy SK, Sehat B, Vasilcanu R, Girnita A, LefkowitzRJ, and Larsson O. β-Arrestin is crucial for ubiquitination and down-regulation of the insulin-like growth factor-1 receptor by acting asadapter for the MDM2 E3 ligase. J Biol Chem, 280(26):24412-9, 2005.

II. Girnita L, Shenoy SK, Sehat B, Vasilcanu R, Vasilcanu D, Girnita A,Lefkowitz RJ, and Larsson O. β-arrestin and MDM2 mediate IGF-1receptor-stimulated ERK activation and cell cycle progression. J BiolChem, 282(15):11329-38, 2007.

III. Sehat B, Andersson S, Vasilcanu R, Girnita L, and Larsson O. Roleof ubiquitination in IGF-1 receptor signaling and degradation. PLoSONE, 2(4):e340, 2007.

IV. Sehat B, Andersson S, Girnita L, and Larsson O. Identification of c-Cbl as a new ligase for IGF-1R with distinct roles to Mdm2 in receptorubiquitination and endocytosis. Submitted 2007.

V. Sehat B, Tofigh A, Girnita L, Lagergren J, and Larsson O. SUMOy-lation, nuclear translocation and transcriptional activity of IGF-1 re-ceptor. Submitted 2007.

v

vi

Contents

List of Abbreviations 1

I Introduction 5

1 Cancer 71.1 History of cancer . . . . . . . . . . . . . . . . . . . . . . . . . 71.2 Mutations in cancer . . . . . . . . . . . . . . . . . . . . . . . . 91.3 Cancer is a multistep process . . . . . . . . . . . . . . . . . . 10

2 Cell signaling 132.1 General principles of cell signaling . . . . . . . . . . . . . . . . 132.2 Overview of receptors involved

in signal transduction . . . . . . . . . . . . . . . . . . . . . . . 142.3 Receptor-mediated endocytosis . . . . . . . . . . . . . . . . . 15

2.3.1 Clathrin-dependent receptor endocytosis . . . . . . . . 152.3.2 Raft-dependent endocytosis . . . . . . . . . . . . . . . 162.3.3 Endocytosis more than signaling

attenuation mechanism . . . . . . . . . . . . . . . . . . 18

3 What’s RTK got to do with it? 213.1 The role of RTKs in cancer . . . . . . . . . . . . . . . . . . . . 213.2 So similar yet so different . . . . . . . . . . . . . . . . . . . . . 23

4 The IGF family 274.1 Insulin like Growth Factors . . . . . . . . . . . . . . . . . . . 27

4.1.1 IGF-1 peptide . . . . . . . . . . . . . . . . . . . . . . . 284.1.2 IGF-2 peptide . . . . . . . . . . . . . . . . . . . . . . . 29

4.2 Insulin like Growth Factor bindingproteins (IGFBPs) and IGFBP proteases . . . . . . . . . . . . 29

4.3 Insulin like Growth Factor Receptors . . . . . . . . . . . . . . 30

vii

4.3.1 IGF-2R . . . . . . . . . . . . . . . . . . . . . . . . . . 304.3.2 IGF-1R . . . . . . . . . . . . . . . . . . . . . . . . . . 30

5 IGF-1R 315.1 Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315.2 IGF-1R activation

and signal transduction . . . . . . . . . . . . . . . . . . . . . . 325.2.1 PI3K pathway . . . . . . . . . . . . . . . . . . . . . . . 325.2.2 MAPK pathway . . . . . . . . . . . . . . . . . . . . . . 355.2.3 14-3-3 pathway . . . . . . . . . . . . . . . . . . . . . . 385.2.4 JNK pathway . . . . . . . . . . . . . . . . . . . . . . . 385.2.5 Other interesting molecules in IGF-1R signaling . . . . 38

5.3 The Role of IGF-1R in Cancer . . . . . . . . . . . . . . . . . . 39

6 Ubiquitin system,general introduction 416.1 Ubiquitin and SUMO . . . . . . . . . . . . . . . . . . . . . . . 426.2 Ubiquitination and SUMOylation,

as easy as 1, 2, 3 . . . . . . . . . . . . . . . . . . . . . . . . . 436.2.1 Ubiquitination . . . . . . . . . . . . . . . . . . . . . . . 436.2.2 SUMOylation . . . . . . . . . . . . . . . . . . . . . . . 46

6.3 Ubiquitin/SUMO modificationand function . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496.3.1 Mono-ubiquitination . . . . . . . . . . . . . . . . . . . 496.3.2 Multiple-mono-ubiquitination . . . . . . . . . . . . . . 496.3.3 Poly-ubiquitination . . . . . . . . . . . . . . . . . . . . 496.3.4 N-terminal ubiquitination . . . . . . . . . . . . . . . . 516.3.5 Mono- and poly-SUMOylation . . . . . . . . . . . . . . 52

6.4 Deubiquitination and deSUMOylation . . . . . . . . . . . . . 526.4.1 Deubiquitination . . . . . . . . . . . . . . . . . . . . . 526.4.2 DeSUMOylation . . . . . . . . . . . . . . . . . . . . . 53

6.5 Ubiquitin dependent degradation . . . . . . . . . . . . . . . . 546.6 IGF-1R ubiquitination . . . . . . . . . . . . . . . . . . . . . . 56

6.6.1 MDM2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 566.6.2 β-arrestin . . . . . . . . . . . . . . . . . . . . . . . . . 576.6.3 Nedd4 . . . . . . . . . . . . . . . . . . . . . . . . . . . 586.6.4 Grb10 . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

6.7 IGF-1R SUMOylation andnuclear translocation . . . . . . . . . . . . . . . . . . . . . . . 596.7.1 Nuclear RTKs . . . . . . . . . . . . . . . . . . . . . . . 59

6.8 Enhancers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

viii

II Results and Discussion 65

7 Results and Discussion 677.1 Paper I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 677.2 Paper II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687.3 Paper III . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 707.4 Paper IV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717.5 Paper V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

8 Concluding remarks 77

Acknowledgements 79

Bibliography 82

ix

x

List of Abbreviations

4E-BP-1 eukaryotic translation initiation factor 4E binding protein 17TMR seven transmembrane receptorsA-loop activation loopAkt/PKB/RAC-PK active human protein kinase/protein kinase B/

A and C protein kinasesAPP1 amyloid precursor protein 1ASK-1 signal regulating kinase-1ATP adenosine triphosphateBARD1 BRCA1-associated RING domainBRACA1 breast cancer 1, early onsetBad Bcl associated death promotorBax Bcl-2 associated X proteinBcl-2 B-cell lymphoma 2Bcl-x Bcl-2-related protein XCbl casitas B-lineage lymphomaCCV clathrin coated vesiclesChIP chromatin immunoprecipitationCREB yclic AMP response element binding factorCRM cis-regulating modulesCtBP C-terminal binding proteinCys/C cystineDMEM Dulbecco’s modified essential mediumDNA deoxyribonucleic acidDUB deubiquitination enzymeE1 ubiquitin activating enzymeE2 ubiquitin conjugating enzymeE3 ubiquitin ligaseECM extracellular matrixEEA1 early endosomal antigen 1EGF epidermal growth factorEGFR epidermal growth factor receptoreIF4E eukaryotic initiation factor 4EEMSA electrophoretic mobility shift assayEps15 epidermal growth factor receptor substrate 15Erb erythroblastosis virusERK extracellular signal-regulated kinaseFAK focal adhesion kinaseFas-L Fas ligandFKHR forkhead-related transcription factor

1

GDP guanosine diphosphateGEF guanine nucleotid exchangerGF growth factorGly/G glycineGPCR G protein-coupled receptorsGrb growth factor receptor-bound proteinGSK-3 glycogen synthase kinas-3 betaGST glutathione S-transferaseGTP guanosin triphosphateHECT homology to E6 associated protein C-terminusHEK human embryonic kidney cellsHGF enhanced hepatocyte growth factorHis/H histidineIκB inhibitor of NFκBIGF insulin-like growth factorIGFBP insulin-like growth factor binding proteinIGFR insulin-like growth factor receptorIlk integrin-linked kinaseIR insulin receptorIRS insulin-receptor substrateJAK/STAT Janus kinase/signal transducers and activators of transcriptionJAMM Jab1/Pad1/MPN-domain metallo-enzymeJNK Jun N-terminal KinaseJNK c-jun NH2-terminal protein kinaseLDL low density lipoproteinLyI lysosome inhibitorLys/K lysineM6P mannos-6-phosphateMAPK mitogen-activated protein kinaseMDM2 murine double minute 2MEF mouse embryo fibroblastMEK MAPK or ERK kinaseMet tyrosine kinase encoded by the MET oncogenemRNA messenger ribonucleic acidmTOR mammalian target of rapamycinMVB multi-vesicular bodyMyc myelocytomatosisMyc myelocytomatosis oncogeneNCP nuclear core complexNFκB nuclear factor-kappa BNGF nerve growth factor

2

Nedd4 neural precursor cells-expressed developmentally down-regulated 4OTU ovarian tumor related proteasesp110 the 110 kDa regulatory subunit of PI 3 kinasep85 the 85 kDa regulatory subunit of PI 3 kinasePDGF platelet-derived growth factorPDGFR platelet-derived growth factor receptorPDK 3-phosphoinositide-dependent protein kinasePI proteasome inhibitorPI3K phosphatidylinositol-3’-kinasePIAS protein inhibitor of activated STATsPIP3 phosphatidylinositol 3,4,5-trisphosphatePLC-γ phospholipase C-gammaPP2A protein phosphatase 2APP2A protein phosphatase 2APTEN phosphatase and tensin homolog deleted on chromosome 10PTP protein tyrosine phosphatasesRACK1 receptor for activated C kinaseRaf retrovirus associated sequencesRanBP Ran-binding proteinRAS human homologue of rat sarcomaRING really interesting new geneRTK Receptor tyrosine kinaseSCF stem cell factorSDS-PAGE sodium dodecyl sulfate-polyacrylamide gel electrophoresisSEA SUMO-1 activating enzymeSENP sentrin-specific proteaseSer/S serineSERCA sarcoplasmic/endoplasmic reticulum Ca2+ ATPasesSH src homologySH-PTP SH2 containing phosphotyrosine phosphataseShc Src homology and collagenSmad human homologue of drosophila mad

(mothers against decapentaplegic)SOS son of sevenlessSPAK stress-activated protein kinasesSrc protein encoded by src proto-oncogeneSUMO small ubiquitin-related modifierSV40 simian vacuolating virus 40TF transcription factorTFBS transcription factor binding siteTGF transformin growth factor

3

TNF tumor necrosis factorTRAF TNF receptor associated factorTrk tyrosine kinase receptor or tropomyosin-related kinaseTRR transcription regulatory regionsTSS transcription start siteTyr/Y tyrosineUb ubiquitinUBB+1 product of ubiquitin B transcript with +1 frame shiftUBD ubiquitin binding domainUBP ubiquitin-specific processing proteaseUCH ubiquitin C-terminal hydrolaseUEV ubiquitin conjugating enzyme variantUFD ubiquitin fusion degradationUlp ubiquitin-like protease 1UTR untranslated regionVEGF vascular endothelial growth factor

4

Part I

Introduction

5

Chapter 1

Cancer

Cancer is a common term referring to a large number of diseases character-ized by uncontrolled growth that progresses toward limitless expansion withability to infiltrate and destroy normal body tissue.

1.1 History of cancer

Cancer has a long history and can be dated back to as long as 150 millionyears ago based on the finding of tumor in the fossil record of Jurassic di-nosaur bone. The first documented human cancer hails from ancient Egypt,in 1500 BC. The origin of the word cancer, however, is credited to the Greekphysician Hippocrates (460-370 BC) who described cancer in detail and usedthe term “carsinos”, meaning crab. Later on Celsus, a roman doctor, trans-lated the greek word carsinos into the Latin word “cancer”. There are spec-ulations claiming that the form of the cancerous lesions first observed byHippocrates recalled the form of a crab and thereby the name “cancer”.

People have wondered about the cause of cancer for centuries. Withthe work of Hooke in the 1600s, and then Virchow in the 1800s, came theunderstanding that living tissues are composed of cells, and that all cellsarise as direct descendants of other cells. Yet, this understanding raisedmore questions about cancer than it answered. One of the most importantearly observations was that the incidence of cancer varies between differentpopulations. Noting the fact that snuff and cigars users, and pitchblendeminers were more prone to cancer, suggested that the origin or causes ofcancer may lie outside the body. These ideas led to a widespread search foragents that might cause cancer. This view was strongly supported by thediscovery that exposure to X-ray could cause cancer.

The observation that cancer sometimes ran in families, however, suggested

7

that cancer is hereditary. Some key discoveries led to the current knowledgethat cancer is a genetic disease harboring from normal cells. The discovery,in 1910, showing that a submicroscopic agent isolated from a chicken tumorcould induce new tumors in healthy chickens, evoked the idea that a tumorcould be traced back to a single cause. Today, this agent is known as Roussarcoma virus – one of several viruses that can act as causative factors in thedevelopment of cancer. Likewise, the discovery by Herman Muller in 1927,that X-ray-irradiation of fruit flies often resulted in mutant offspring, raisedone of the key questions: Can the two known effects of X-rays – promotionof cancer and genetic mutation – be related to one another?

Support for this idea came from the work of Bruce Ames and others whoshowed in 1975 that compounds known to be potent carcinogens were gen-erally also potent mutagens. Further, a genetic model, proposed by AlfredKnudson in 1971, explaining the differences between sporadic retinoblastomaand familial retinoblastoma influenced the development of a unified view ofthe origin and development of cancer. His model states that children withsporadic retinoblastoma are genetically normal at the moment of conception.However, experiencing two somatic mutations will lead to the developmentof an eye tumor. Children with familial retinoblastoma, on the other hand,already carry one mutation. In effect, in familial retinoblastoma, each retinalcell is already primed for tumor development needing only a second muta-tional event to trigger the cancerous state. The difference in probabilitesbetween acquiring one or two mutational events at random explains whychildren affected by sporadic retinoblastoma have only one tumor in one eye,while in familial retinoblastoma the affected children usually have multipletumor foci growing in both eyes.

An important piece of the puzzle in identifying the origin of cancer cellscame from transplantation studies. One observed that transplanting tissuesbetween unrelated or distantly related people was not as successful as trans-planting tissues between identical twins. This phenomenon was explained bythe ability of the immune system to distinguish between self and foreign cells.This discovery led to the invension of tissue-typing, which made it possibleto distinguish cells from different individuals. The application of this methodon tumor cells revealed that tumors arise from the patient’s own tissues, andnot from cells introduced into the body for example by infection as previ-ously believed. The next important clue was the discovery of cancer cellsbeing monoclonal (derived from a single cell) rather than polyclonal (manyindividual cells become cancerous). The support for this view came fromstudying the inactivated X-chromosome of human females. X-chromosomeinactivation occurs randomly in all cells during female embryonic develop-ment. Because the inactivation is random, the female is like a mosaic in

8

terms of the X chromosome, with different copies of the X turned on or off indifferent cells of the body. Once inactivation occurs in a cell, all generationsof cells coming from that cell have the same chromosome inactivated (eitherthe maternal or the paternal X). The observation that all the cells within agiven tumor invariably have the same X-chromosome inactivated, suggestedthat all cells in the tumor must have descended from a single ancestral cell.

1.2 Mutations in cancer

By the mid-1970s, a unified view started to develop the basis of our modernmolecular understanding of cancer. In particular, findings establishing therelationship of mutagenesity and cancerogenecity led to a straightforwardmodel: Carcinogens induce mutations in critical genes, and these mutationscause an abnormal growth which is discovered years later as a tumor. Thismodel could even explain how cancer is inherited. The next step was findingthe genes involved.

The main difference between normal and cancerous cells are that cancercells have lost the restraints on growth that characterize normal cells. Theyare genetically unstable and prone to rearrangements, duplications, and dele-tions that cause their progeny to display unusual traits.

Two categories of genes play major roles in triggering cancer. One cat-egory, called proto-oncogenes, encourages cell division. The other category,called tumor-suppressor genes, inhibits it. Together, proto-oncogenes andtumor-suppressor genes regulate growth that normally ensures that each tis-sue and organ in the body maintains a size and structure that meets thebody’s needs.

Mutated proto-oncogenes become oncogenes, i.e., genes that stimulateexcessive division. Mutations in tumor-suppressor genes, on the other hand,eliminate the critical inhibition of cell division that normally prevents exces-sive growth. Collectively, mutations in these two categories of genes accountfor much of the uncontrolled cell division that occurs in human cancers.

In addition to proliferation control regulated by proto-oncogenes and tu-mor suppressor genes, cells have at least three other mechanisms which con-tribute to avoiding uncontrolled cell division. The first of these mechanismsis the DNA repair system, which detects and corrects errors in DNA. Errorsoccur either during DNA replication or DNA damage caused by attack ofcarcinogens. In most cases, such errors are rapidly corrected by the cell’sDNA repair system. Should the system fail, however, the error, now a muta-tion, becomes a permanent feature in that cell and in all of its descendants.The high efficiency of the normal DNA repair system makes mutations a rare

9

event and an inefficient cancer progression process. Mutations in DNA re-pair genes themselves, however, remarkably decreases a cell’s ability to repairerrors in its DNA. As a result, mutations appear in the cell much more fre-quently leading to accumulation of mutation needed for cancer progression.

The second system is the ability of cells to undergo programmed celldeath, apoptosis, which occurs when essential components are damaged orits control system is deregulated. Acquired ability to resist apoptosis is ahallmark of many types of cancers. Ability to evade apoptosis not onlycontributes to the growth of tumors but also makes cancer cells resistant tothe cell suicide inducing effects of radiation and chemotherapy.

The third backup system, known as the Hayflick limit, is limiting thenumber of times a cell can divide, thereby ensuring that cells do not reproduceendlessly. The cell is then referred to as senescent. Senescent cells deteriorateand die, which also leads to the aging process. This system is governed bytelomers – DNA segments at the ends of chromosomes – which become shortereach time a chromosome replicates and eventually can no longer protect thechromosomes. This leads to DNA damage which inevitably causes cell death.Cancer cells, on the other hand, are “immortal” due to the activity of anenzyme called telomerase, which rebuilds the telomers, allowing division tocontinue indefinitely.

1.3 Cancer is a multistep process

Taken all the above mentioned discoveries together, one could summarize thatcancer is a disease in which a single normal body cell undergoes a genetictransformation into a cancer cell. This cell and its descendants divide acrossmany years, producing the population of cells that we recognize as a tumor.Tumors produce the symptoms that the patient experiences as cancer.

This picture, although accurate in its essence, dose not represent a com-plete description of the events involved in tumor formation. Cancer is, inessence, a genetic disease. However, there are two key differences betweencancer and most other genetic diseases. First, cancer could also occur dueto somatic mutations, whereas all other genetic diseases of mammals (ex-cluding those involving mitochondrial genes) are caused solely by germ-linemutations. Second, each individual cancer arises, not from a single mutation,but from the accumulation of several mutations, each of which are impor-tant for one of the attributes needed for cancer cells. Cancer cells exhibit awide range of important differences from normal cells. As suggested by R.Wineberg [81] cancer cell formation depends on six essential alterations thatcollectively dictate malignant growth:

10

1. self-sufficiency in growth signals,

2. insensitivity to growth-inhibitory signals,

3. evasion of apoptosis,

4. limitless replicative potential,

5. sustained angiogenesis, and

6. tissue invasion and metastasis.

Insights into these six hallmarks of cancer has provided a better understand-ing of the details in cancer cell biology. The main focus of this thesis isthe first mentioned feature – self-sufficiency in growth signals – and will bediscussed in further detail in the coming chapters.

11

12

Chapter 2

Cell signaling

2.1 General principles of cell signaling

Cell signaling is part of a complex system of communication that governs ba-sic cellular activities and coordinates cell actions. Mechanisms which makeone cell influence the behavior of another cells has existed even in unicellu-lar organisms. Although these organisms live an independent life, they havethe ability to communicate with each other. For example, when an indi-vidual of the budding yeast Saccharomyces cerevisiae is ready to mate, itsecretes a peptide (mating factor) into its environment. The mating factorbinds and signals to the cells of opposite mating type to stop proliferatingand start conjugating, which is the first step in the mating process [131].Further, in the human gastrointestinal tract, bacteria exchange signals witheach other and with human epithelial and immune system cells [30]. How-ever, the cell signaling in higher animals have become much more complex,involving hundreds of kinds of signaling molecules including proteins, smallpeptides, amino acids, nucleotides, steroids, fatty acid derivates, and evendissolved gases such as nitric oxide and carbon monoxide. Whereas most ofthese signals (ligands) are secreted from the signaling cell, others are releaseby diffusion through the plasma membrane (e.g. signaling via gap junctions),while some remain tightly bound to the cell surface of the signaling cell andonly influence cells that come in direct contact with it (e.g. signaling duringdevelopment and immune responses). Regardless of the nature of the ligand,a complementary set of receptor proteins on the target cell, bind specificallyto the ligand and get activated so as to generate a cascade of intracellularsignals that alter the behavior of the cell.

13

2.2 Overview of receptors involved

in signal transduction

Upon activation, cell surface receptors initiate a response referred to as signaltransduction. These receptors are divided into four general classes:

Enzyme-linked receptors, when activated, either function as an enzymeor are associates with enzymes. The majority of receptors in this classare receptor tyrosine kinases (RTKs), which are the most clearly un-derstood form of receptor involved in cancerogenesis and of the maininterest of this thesis. Receptors with intrinsic tyrosine kinase activityare capable of autophosphorylation as well as phosphorylation of othersubstrates. For detailed discussion about RTKs see section 3. Tyrosinephosphatases (e.g. CD45 (cluster determinant-45) protein of T cellsand macrophages), guanylate cyclases (e.g. natriuretic peptide recep-tors) and serine/threonine kinases (e.g. activin and TGF-β receptors)are other members of this receptor category.

G-protein coupled receptors (GPCRs) are receptors that are associ-ated intracellularly to GTP-binding and hydrolyzing proteins (termedG-proteins). GPCRs all have a structure that is characterized by 7transmembrane-spanning domains and are therefore termed serpentineor 7TM receptors. Examples of this class are the adrenergic recep-tors, odorant receptors, and certain hormone receptors (e.g. glucagon,angiotensin, vasopressin and bradykinin).

Intracellular hormone receptors are found intracellularly, which uponligand binding migrate to the nucleus where the ligand-receptor com-plex directly affects gene transcription.

Integrins are a specialized class of cell surface receptors whose extracellu-lar domains bind to components of the extracellular matrix (ECM) andthe intracellular domains bind to components of the cytoskeleton. Theyare a large family of α and β heterodimeric transmembrane cell surfacereceptors. In mamals, 19 α and 8 β subunit genes encode polypeptidesthat combine to form 25 different receptors with specificity for differ-ent ECM components. In addition to performing a structural role,integrins are also important sites of signal initiation and modulationand play a signaling role during tumor angiogenesis. Integrins andRTKs form a cooperative signaling network to regulate various biolog-ical processes [100, 152, 153]. For instance, integrin adhesion promotes

14

activation of several RTKs, including the epidermal growth factor re-ceptor (EGFR). Conversely, RTKs affect cell adhesion, spreading, andmigration by regulating integrin localization and activation. Evidenceprovided by Wang et al., 1998 [222] shows that in a human tumorigenicmammary epithelial cell line that expresses high levels of EGFR andβ1 integrin, downregulation of either β1 integrin or EGFR results indownregulation of both. This downregulation leads to a reversion ofthe transformed phenotype, suggesting that the integrin-RTK interac-tion is required for maintaining a transformed state and that integrinsmay function to amplify oncogenic signaling from RTKs. Further, thework of Giancotti and Tarone suggests that integrin-specific signals im-part a stringent control to the action of RTKs, determining whethercells proliferate or undergo growth arrest, migrate or remain station-ary, live or undergo apoptosis when adhering to a specific matrix [66].In accordance with this hypothesis, studies on signaling molecules thatfunction downstream of integrins and RTKs, such as focal adhesionkinase (FAK), Src, Shc, and ILK, have documented a general role forjoint integrin-RTK signaling [117, 210, 98].

2.3 Receptor-mediated endocytosis

Receptor-mediated endocytosis is the most understood form of endocytosis.It selectively internalizes specific ligands together with their receptors. It in-volves constitutive formation of small vesicles, which is usually preceded bythe formation of clathrin-coated vesicles (CCVs). In addition, there exists an-other type of internalization which is clathrin independent, raft-dependentendocytosis [173]. After internalization, the receptors are either detachedfrom their ligand and recycled back to the cell surface or are transferredto their destruction and degradation. While some receptors (e.g. Transfer-rin) are constitutively internalized, others (e.g. RTKs and GPCRs) undergoligand-dependent internalization.

2.3.1 Clathrin-dependent receptor endocytosis

Once ligands bind to their receptors, the receptors move within the plasmamembrane and become concentrated in small depressions called clathrin-coated pits. Clathrin-coated pits are formed when many clathrin proteinmolecules interact with each other to form a convex, basket like structure(the so called polyhedral lattice) on the inside of the plasma membranethat molds the membrane into a pit. The coated pits then progressively

15

invaginate, or form inward, to form clathrin-coated vesicles that pinch offthe plasma membrane into the cytoplasm after which the clathrin coat isremoved. The uncoated vesicle, still carrying receptor proteins and theirbound cargo molecules, fuses with another membranous organelle called en-dosome. Endosomes function as sorting stations. Although it remains dis-puted whether trafficking from the early endosome to the next compartmentin the endocytic pathway involves vesicular transport or maturation, thecargo proteins proceed into a compartment referred to as either the late en-dosome or the multi-vesicular body (MVB). This compartment is character-ized by different protein composition, low luminal pH, and multiple internalvesicles. Receptors brought into the cell by receptor-mediated endocytosishave one of two fates after unloading their cargo and leaving the endosome:(1) The ligand is disassociated from the receptor due to low pH environmentof MVB and the receptor is recycled back to the plasma membrane; or (2)they can be transported to the lysosome where they are degraded (Figure2.1A).

2.3.2 Raft-dependent endocytosis

A less defined endocytic pathway is raft dependent endocytosis. Lipid raftsare low-density, detergent-resistant microdomains of plasma membrane thatare enriched in cholesterol and glycosphingolipids. Nearly all molecules thatare known to be internalized independently of clathrin are found in biochem-ically defined lipid rafts, but archetypical substrates for uptake by clathrin-coated pits, such as the low-density lipoprotein (LDL) and transferrin re-ceptors, are not [165]. Caveolae, a subclass of rafts, are characterized byflask-like invaginations of the plasma membrane that are distinguished frombulk lipid rafts by the presence of caveolin-1. Rafts-caveolae are known to beabundant in various signaling molecules, such as cell surface receptors andintracellular signaling molecules. Further, these microdomains have been in-volved in many cellular functions, including the regulation of apoptosis andcell proliferation [62]. RTKs have been shown to interact with rafts/caveolae,and after the activation of RTKs some intracellular signaling molecules areredistributed to rafts/caveolae [225, 26]. Moreover, this redistribution playsan important role in the regulation of receptor-mediated cellular function,and accordingly, the disruption of rafts/caveolae results in the impairmentof signaling events and receptor function. Therefore, it has been proposedthat rafts/caveolae serve as molecular platforms that spatially organize ap-propriate molecules for specific signaling pathways [242] (Figure 2.1).

16

P

P

P

P

MVB

AdaptorsKinase substrates

B

Early endosome

CaveolinClathrin

Cav+, EEA1-

Caveosome

Early endosomeCav-, EEA1+

Recyclingendosome

Late endosome MVB Lysosome

A Clathrin dependant endocytosis

Caveolin-raft dependant endocytosis

Figure 2.1: Clathrin- and caveolin/raft-internalization. A. This car-toon summarizes the process of clathrin- and caveolin/raft-dependent endo-cytosis. B. During the passage through the endosomal compartments, RTKsbecome increasingly concentrated in the internal membranes of MVBs. Theseinternal structures might represent isolated vesicles or deep invaginations ofthe limiting endosomal membrane. Receptors that are incorporated intovesicles are not accessible for interactions with adapter proteins and cannotphosphorylate cytoplasmic signaling proteins. So, sequestration of active re-ceptors inside MVBs might terminate RTK signaling before their degradationin lysosomes.

17

2.3.3 Endocytosis more than signalingattenuation mechanism

Endocytosis of RTKs has long been recognized as a means to terminate sig-naling by clearing activated receptors from the plasma membrane and routingthem through the endosomal system and on to lysosomes for degradation.In this way the cell can safely and quickly abrogate mitogenic signals thatwould otherwise prove hyperproliferative and ultimately deleterious. How-ever, it has become clear that the output of a signaling process depends notonly on activation of a particular set of signaling molecules but also on whereand for how long the signal is emitted. Endosomes appear to be ideally suitedfor such regulation. Intriguing studies suggest that receptor trafficking playsan important role in regulation of receptor signaling, going beyond its con-ventional role in cargo degradation [9]. The predominant view that signalingoccurs only at the plasma membrane was challenged by Bergeron [8], Posner[9], and colleagues, who observed that shortly after ligand addition the ma-jority of activated EGFRs were found not on the plasma membrane but onearly endosomes [46].

In many cell types the rate of the internalization upon ligand stimulationis significantly higher than the rate of degradation, which results in pro-longed presence of signaling receptors in early endosomes. Furthermore, itwas found that EGF, the main ligand for EGFR, does not significantly disas-sociate at endosomal pH [204, 203, 243]. Consequently, a substantial amountof ligand-bound receptors remain in endosomes. Interestingly, the same re-ceptor may vary in its ability to release different ligands in endosomes. Forexample, binding of transforming growth factor-α (TGF-α) to the EGFR ismore pH sensitive than is EGF. Consequently, TGF-α is disassociated fromthe receptor in the endosomes favoring recycling of the internalized recep-tor [59], while EGF favors receptor sorting to the lysosomes. Finally, theopportunity for RTKs to signal during endocytosis is also determined bythe rate of their sorting to the internal membrane of multi vesicular bod-ies (MVBs) (Figure 2.2). Whereas receptors that are associated with theouter membrane of MVBs are accessible for functional interaction with cy-toplasmic proteins, sequestration inside MVBs makes receptors inaccessibleto the signal-transduction machinery present in the cytplasm, and thereforeterminates signaling [61, 134] (Figure 2.1B).

One question that arises from these observations is why endosomes shouldbe involved in the propagation of signal from cell membrane throughout thecytoplasm. One obvious role for endocytosis in signaling is to control themagnitude of the response as the duration of signaling is an important pa-rameter determining the biological outcome. Another reason could be that

18

Blume-Jensen and Hunter 2001

Figure 2.2: Human receptor protein-tyrosine kinases. The prototypicreceptor for each family is indicated above the receptor, and the known mem-bers are listed below. Abbreviations of the prototypic receptors: EGFR,epidermal growth factor receptor; InsR, insulin receptor; PDGFR, platelet-derived growth factor receptor; VEGFR; vascular endothelial growth fac-tor receptor; FGFR, fibroblast growth factor receptor; KLG/CCK, coloncarcinoma kinase; NGFR, nerve growth factor receptor; HGFR, hepatocytegrowth factor receptor, EphR, ephrin receptor; Axl, a Tyro3 PTK; TIE, ty-rosine kinase receptor in endothelial cells; RYK, receptor related to tyrosinekinases; DDR, discoidin domain receptor; Ret, rearranged during transfec-tion; ROS, RPTK expressed in some epithelial cell types; LTK, leukocytetyrosine kinase; ROR, receptor orphan; MuSK, muscle-specific kinase; LMR,Lemur. Other abbreviations: AB, acidic box; CadhD, cadherin-like domain;CRD, cysteine-rich domain; DiscD, discoidin-like domain; EGFD, epider-mal growth factor-like domain; FNIII, fibronectin type III-like domain; IgD,immunoglobulin-like domain; KrinD, kringle-like domain; LRD, leucine-richdomain. In bold and italic type are receptors implicated in human malig-nancies. An asterisk indicates that the member is devoid of intrinsic kinaseactivity.

19

the movement of endosomes can direct activated signaling molecules to theirtarget site [220]. A third speculation is that given the fact that several cyto-plasmic signaling molecules are shared between different pathways, unwantedinteractions may be avoided by restricting signaling cascades in space. Be-sides acting as transport carriers over long distances, endosomes may providetwo additional interconnected levels of control: selective targeting of signal-ing molecules to specific organelles, and their segregation to sub-domainswithin a given organelle.

Clearly, the next steps that need to be taken are to understand the mech-anistic basis of selective endosomal signaling and its biological consequence.For example, what is the relative contribution of signal duration and specificsubstrate assembly? As previously mentioned, receptors can be internalizedvia several routes (e.g. clathrin-dependent and caveolin-dependent), and intodifferent types of endosomes (e.g. EEA1- or Caveolin-positive ones). As weshow in Paper IV, IGF-1R is internalized through both internalization path-ways. Could we then have specialized endosomal compartments devoted tosignaling?

20

Chapter 3

What’s RTK got to do with it?

Receptor tyrosine kinases (RTKs) transmit signals that regulate cell prolifer-ation and differentiation, promote cell migration and survival, and modulatecellular metabolism. They play a critical role in a wide range of biologicalprocesses, including embryonic development, growth of an organism, angio-genesis, synaptic plasticity, and oncogenesis.

3.1 The role of RTKs in cancer

When the sequence of EGFR, the first RTK to be discovered, was recognizedin 1984, it surprised everyone by being closely related to a potent retroviral-encoded oncoprotein ErbB which was known for rapidly inducing erythro-leukimia (leukemia of red blood cells). Insights into EGFR function, andlater on function of other RTKs, provided an explanation of one of the long-lasting question marks in cancer cell biology.

Normal cells had long been known to require stimulating mitogenic growthsignals for their proliferation. Typically, such signaling begins with the pro-duction of a growth factor (GF), a protein that ultimately stimulates celldivision, by the signaling cells. GFs can be “local mediators” secreted byneighboring cells affecting the immediate environment (paracrine signaling)or secreted into the bloodstream affecting the whole organism (endocrine sig-naling). The released GFs will attach to specific receptor proteins located onthe surfaces of target cells. As mentioned earlier, the receptor conveys a stim-ulatory signal to proteins in the cytoplasm which emit stimulatory signals toother proteins in the cell until the division-promoting message reaches thecell’s nucleus and activates a set of genes that help move the cell through itsgrowth cycle. In this manner growth factors make individual cells dependantof each other as members of a community and through this communication,

21

which benefits the entire organism rather than the single cell, the tissue archi-tecture and the warfare of the organism is maintained. However, this depen-dency on environment in cancer cells is strongly reduced. The explanation ofhow this was possible come from the discovery of the ErbB-EGFR connec-tion. The ErbB oncoprotein lacks sequences of the N-terminal ectodomainof EGFR responsible for ligand binding. Consequently, ErbB cannot recog-nize EGF and sends a constant growth signal into the cell persuading thecell that EGF is present in the surrounding. This was the first discovery ex-plaining how cancer cells are self-sufficient. Further discoveries revealed thatmany proto-oncogens code for proteins involved in growth-promoting path-ways. Gain-of-function mutations of these proto-oncogens usually mimic thenormal growth signaling causing the proteins involved to be overactive, andthus the cell proliferates much faster.

In fact, there are four common molecular strategies by which self-sufficien-cy is maintained, and all these mechanisms are, directly or indirectly, RTK-signaling dependent:

Ligand independent signaling: The first pathway by which cells becomeself-sufficient, as already mentioned, is expression of mutated form ofthe growth factor receptor which can convey ligand independent sig-naling [45]. Truncated version of EGFR, lacking the ectodomain, arefound in a number of human tumor cell types. Further, truncation ofectodomains of Met receptor (HFG family) and TrkA receptor (Trkfamily) and fusion of the remaining portion of the protein with otherproteins prone to dimerization results in an oncoprotein, constitutivelydimerized, which leads to ligand independent signaling.

Autocrine signaling: The second explanation for self-sufficiency came onceagain from the connection between a member of the RTK system,namely PDGF, and an oncoprotein, the PDGF-like sis-encoded onco-protein. Sis-protein is a product of sis-oncogen synthetized by Simiansarcoma virus. Infected cells produce sis-protein in large quantities andthey mimic the PDGF by binding and activating PDGFR in the samecells and thereby provide cells with a constant flux of growth stimu-latory signals. Here the signaling and targeting cell is the same. Theability of cancer cells to produce GFs to which they are responsive isreferred to as autocrine signaling and contributes to increased auton-omy of cancer cells, thus the uncontrolled growth. A variety of humancancers are known to overexpress growth factors and/or their receptors.As examples gliobalstomas producing PDGF (platelet-derived growthfactor) and sarcomas producing TGFα (tumor growth factor α) couldbe mentioned [2]. Some human cancers, such as certain lung cancers,

22

produce as many as three different growth factors – TGFα, stem cellfactor (SCF), and insulin like growth factor (IGF) – and at the sametime expresse the receptors of these ligands establishing three differentautocrine signaling loops.

Mutation of signaling molecule RAS: A third mechanism by which self-sufficiency is maintained is altering downstream elements in the sig-naling pathway. The discovery of ErbB and sis-oncoproteins, theirinvolvement in the RTK signaling and their role in cancerogenecity,increased the interest in the other oncogens involved in the signalingmachinery, particularly the RAS oncogen. Wt RAS binds to GDP in itsinactive form. RTKs upon activation transduce a guanine nucleotideexchanger factor (GEF) which helps RAS to shed its GDP and bindto GTP and thereby become active. RAS has a GTPase activity bywhich it hydrolysis GTP to GDP. This controlled on-off system is lostin oncoprotein-RAS resulting in signal transmission for a long, possiblyindefinite period of time. An example is activation of oncogene-RASin bladder carcinoma [65, 4].

Alteration in integrin heterodimers: As mentioned earlier, different in-tegrin heterodimers affect signaling either positively or negatively. Can-cer cells can alter integrin heterodimer expression in a way to promotethe ones that favors growth signals [65] and thereby contribute to theself-sufficiency of cancer cells.

3.2 So similar yet so different

The amino acid sequences of RTKs are highly conserved. The proteins encod-ing RTKs contain four major domains: extracellular ligand binding domain,transmembrane domain, intracellular tyrosine kinase domain, and intracel-lular regulatory domain. RTK proteins are classified into families based onstructural features in their extracellular portions which include the cysteinerich domains, immunoglobulin-like domains, leucine-rich domains, Kringledomains, cadherin domains, fibronectin type III repeats, discoidin I-like do-mains, acidic domains, and EGF-like domains. Based upon the presence ofthese various extracellular domains, the RTKs have been sub-divided intoat least 20 different families (Figure 2.2). In spite of the great similaritybetween family members, there is tissue and ligand specificity (see Table 3.1and Figure 2.2 for details).

Signaling through RTKs is mainly accomplished via mitogenic signalingpathways. In order to convey their message, RTKs cause tyrosine phospho-

23

Rece

pto

rFam

ilyC

onta

ins

Rece

pto

rL

igand

Resp

ondin

gC

ells

EG

F

Erb

B-1

=E

GF

RT

GF

a,E

GF

,H

B-E

GF

,ep

igen,

Epith

elialcells,

epiregu

linam

phiregu

lin,b

atacelluli

some

mesen

chym

alcells

Erb

B-2

=H

ER

2N

oligan

ds

Erb

B-3

=H

ER

3neu

regulin

1an

d2

Erb

B-4

=H

ER

4E

GF

,H

B-E

GF

,neu

regulin

1-4,batacellu

lin

Insu

linIR

Insu

linW

ide

varietyof

cellty

pes

IGF

-1RIG

F-1,

Insu

lin,

IGF

-2

PD

GF

PD

GF

RP

DG

FE

ndoth

elial,fibrob

lasts,oth

erm

esench

ym

alC

-Kit

CSF

glialan

dhem

atopoetic

cellsF

GF

FG

FR

1,2,

3,4,

6P

DG

F-A

,-B

,-C

,an

d-D

Endoth

elial,fibrob

lasts,oth

erm

esench

ym

alan

dneu

roecto

derm

alcells

VE

GF

VE

GF

RV

EG

F-A

,-B

,-C

,an

d-D

Endoth

elialcells

incap

illaris,ly

mph

ducts

HG

Fc-M

etH

GF

=SF

Variou

sep

ithelial

cells

Trk

Trk

AN

GF

,N

T-3

Neu

rons

Trk

BB

DN

F,

NT

-3T

rkC

NT

-3N

GF

RN

GF

Tab

le3.1:

Rece

pto

rs,lig

an

ds,

an

dre

spondin

gce

llsof

the

most

com

mon

hum

an

RT

Ks.

24

rylation, whereas other signaling processes involved in the cell are almostexclusively serine and threonine phosphorylation. Most RTKs consist of sin-gle polypeptides, and they dimerize in response to ligand binding. The twoexceptions are the insulin receptor (IR) and insulin-like growth factor 1 recep-tor (IGF-1R) that exist as a tetramer in the absence of its ligand. After theligand has bound to and activated its receptor, the tyrosine kinase activityof the cystolic domain of the RTKs are activated. Each kinase phospho-rylates specific tyrosine residues in the cytosolic domain of its dimer part-ner, a process termed trans-autophosphorylation. The receptor kinase thenphosphorylates other tyrosine residues outside the catalytic domain, creatingphosphotyrosine residues that serve as docking ports for additional proteinsthat transmit intracellular signals downstream of the activated receptors. Es-sential for most RTK-mediated signaling is the engagement and activationof MAPK- (mitogen-activated protein kinase) , PI3K- (Phosphatidylinositol-3’-kinase), JNK- (Jun N-terminal Kinase), PLC-γ- (phospholipase C-γ), andJAK/STAT- (Janus kinase/signal transducers and activators of transcrip-tion) pathways in a phosphotyrosine-dependent manner. Together, thesefactors contribute to the diversity of biological responses generated by RTKsignaling. Different RTKs stimulate similar collections of intracellular signal-ing pathways. Given this similarity of structure and response, an importantquestion about the role of RTKs is why, upon activation of a particular RTK,a particular fate is chosen [211]. The two proposed models are described be-low.

1. The first model suggests that RTKs generally send signals in the sameway, but cells interpret these signals differently. For example, differentcell types could exhibit distinct responses to RTK activation becausethey contain different sets of transcription factors waiting to interpretthe RTK signal. Support for this model comes from studies show-ing that the expression of a constitutively activated RAS protein cancompensate for the absence of RTK function in many developmentalsettings [211]. In addition, other studies have shown that misexpressionof Homeobox-containing transcription factors can alter the outcome ofRTK signals [111, 14, 139]. Together these studies suggest that a majorrole for RTKs during cell fate decisions is to provide a go-signal thatstimulates the receiving cells to execute the next stage of their devel-opment. However, this model does not easily explain how multipotentstem cells differentially respond to different RTK-binding growth fac-tors, a common theme in mammalian systems.

2. The second model postulates that there are intrinsic differences in theintracellular signaling pathways activated by various RTKs. These dif-

25

ferences could either be quantitative (strength or duration of the signal)or qualitative (a different combination of intracellular pathways beingactivated). Several studies have supported the idea that either quanti-tative or qualitative differences between RTK signaling pathways canspecify the nature of the cellular response to RTK activation. For ex-ample, studies of PC12 cells have shown that NGF, which generatesa long-lasting activation of MAPK, induces neuronal differentiation inthese cells, while EGF, which induces a much shorter MAPK activa-tion, induces proliferation [140]. Consistent with this idea, experimen-tal manipulations that lengthen the response to EGF can cause EGFsignaling to induce neuronal differentiation [140].

The reminder of this thesis will be concerned with the receptor tyrosinekinase IGF-1R and will focus on ubiquitin modification of the receptor andit’s role in signaling, degradation, and internalization of the receptor. Inaddition, evidence for SUMOylation of IGF-1R will be provided.

26

Chapter 4

The IGF family

The molecular era of the IGF system started in the 1970s with the purificationand amino acid sequencing of the two ligands IGF-1 and IGF-2. During thepast 30 years, the number of members of the IGF system has increased.The IGF system consists of three ligands which includes the polypeptideligands insulin, IGF-1, and IGF-2, three types of cell membrane receptorsIGF-1R, IGF-2R, and insulin receptor, and six binding proteins, IGFBP-1-6.In addition, a large group of IGFBP proteases may also be regarded as partof the IGF family. They hydrolyze IGFBPs resulting in the release of boundIGFs which then resume their ability to interact with IGF-1R.

Insulin is generally regarded as being mainly involved in metabolic home-ostasis, whereas the IGFs are regulators of cellular proliferation. Therefore,I will focus on IGFs rather than insulin. Nevertheless, one should mentionthat the involvement of insulin and insulin receptor in cancer is an area ofextensive research.

4.1 Insulin like Growth Factors

IGFs are potent mitogens for many different cell types and play a central rolein growth and development. In addition to their function in general growth,locally produced IGFs have important roles in tissue regulation and woundhealing. All these processes involve mitogenic stimulation and cell divisionand need to be controlled accurately. The elevated concentration of IGFs iscorrelated with cancer.

27

4.1.1 IGF-1 peptide

IGF-1 is both a systematic growth factor, produced primarily by liver, and alocal growth factor. The mature IGF-1 protein has 70 amino acids, its genespans 9 kb and contains two promoters and six exons [236]. Although IGF-1 transcripts are not exclusively tissue-restricted, those initiated at Exon 2(class 2) are predominate in the liver, and are highly growth hormone respon-sive and as such are major endocrine effectors of growth hormone (GH) [234].By contrast, transcripts initiating at Exon 1 (class 1) are widely expressedin all tissues, and are less affected by circulating growth hormone levels, pre-sumably performing autocrine or paracrine functions. The significance of theproteins encoded by these different transcripts is widely debated, and untilrecently, little attention has been paid to their potential functional diversity.

The IGF-1 mRNA level increases in embryos between the eight-cells andblactocyte stages and decreases during the later stages of embryogenesis.During this time the IGF-1 transcription is occurring in many different celltypes and tissues, liver not being the main producer, and the mode of actionof IGF-1 is regarded as autocrine/paracrine. However, hepatocytic IGF-1mRNA levels increases markedly during birth and the liver becomes the mainproducer of endocrine IGF-1. At the peak of postnatal growth the levels ofcirculating IGF-1 reaches its maximum, and declines with age. In humans, asecond peak is observed at the so called growth spurt during puberty. Serumlevels of IGF-1 in humans are relatively low in the fetal stage, ranging from20ng/mL in week 30 to over 100ng/mL at full term pregnancy. In adults,concentration of IGF-1 in serum ranges between 100 and 200 ng/mL.

Functionally, IGF-1 not only stimulates cell proliferation but also inhibitsapoptosis. It has now been recognized that the combination of these mito-genic and antiapoptotic effects has a profound impact on tumor growth [35].High serum concentrations of IGF-1 are associated with an increased risk ofmany cancers such as breast [178], prostate [83], colorectal [137], and lungcancer [240].

IGF-1 exerts its mitogenic action by increasing DNA synthesis and bystimulating the expression of cyclin D1, which accelerates progression of thecell cycle from G1 to S phase [60, 50]. Further, IGF-1 inhibits apoptosisby stimulating the expression of Bcl proteins, suppressing the expression ofBax, which results in an increase of the relative amount of the Bcl/Baxheterodimer, thereby blocking initiation of the apoptotic pathway [151, 172,245].

28

4.1.2 IGF-2 peptide

IGF-2 is a small single-chain peptide, 67 amino acids in length, and has fourpromoter sites, P1-P4, of which promoters P2-P4 are genomically imprinted.Disruption of imprinting and the resulting increase of gene dosage have beenshown to be implicated in tumor progression in a variety of human tumors.IGF-2 transcription is initiated from the P1 promoter in adult tissue, whereastranscription of promoters P3 and P4 is often seen in fetal tissues [245].Further, increased transcription from P2-P4 promoter sites has been observedin certain cancers [245, 184, 214, 69, 150, 56, 116].

The concentration of serum IGF-2 is around 100ng/mL at week 30 toalmost 300ng/mL at term. After birth, the level of IGF-2 rises to 700ng/mL.In adults IGF-2 levels vary between 400 and 600 ng/mL which is higherthan IGF-1 levels (100- to 200 ng/mL). Moreover, IGF-2 levels in serum arerelatively stable after puberty and GH has no or little effect on that. Despiteits presence at higher concentration than IGF-1, IGF-2 is believed to play aless important role in postnatal growth than does IGF-1. This conclusion isbased on (1) the impact of these IGFs on body growth, (2) their regulationby GH, and (3) their relative binding affinities to IGF-1R and IGFBPs. Thecombination of higher affinity of the IGF-1 to the receptor and lower affinityto the binding proteins, results in more IGF-1 than IGF-2 interacting withIGF-1R. Like IGF-1, IGF-2 has mitogenic and antiapoptotic actions andregulates cell proliferation and differentiation.

4.2 Insulin like Growth Factor binding

proteins (IGFBPs) and IGFBP proteases

The major part of circulating IGFs (∼98%) is bound to one of the sixIGF Binding Proteins (IGFBPs 1-6). Approximately 80% of IGFs bindsto IGFBP-3, which makes this protein the major carrier of IGFs in plasma.IGFBPs have multiple and complex functions. They are able to inhibit orto enhance the action of IGFs, resulting in either suppression or stimulationof cell proliferation [108, 32]. They do so by binding to IGFs with higheraffinity than IGF-1R does; and by doing so block the interaction betweenIGFs and IGF-1R causing suppression of IGF action [108, 32]. However,binding of IGFBPs to IGFs also protects IGFs from degradation, and thatprotection can enhance the action of IGFs by increasing their bioavailabilityin local tissues [108, 32]. One can say that IGFBPs, by binding to IGFs,play three major roles:

1. transporting IGFs from circulation to peripheral tissues (e.g. IGFBP-1,

29

2, and 4),

2. protecting IGFs from degradation (IGFBP-3), and

3. regulating the interaction between IGFs and IGF-1R (IGFBP-1-6).

Since the actions of IGFs can be either suppressed or enhanced by IGF-BPs, proteolysis of IGFBPs by IGFBP proteases is an important factor indetermining the regulatory impact of IGFBPs on IGF action. Regulation ofIGFBP proteolysis is complex and remains poorly understood.

4.3 Insulin like Growth Factor Receptors

The biological action of IGFs are mediated by binding to cell surface recep-tor IGF-1R. As mentioned there are three receptors in IGF system: insulinreceptor, IGF-1R, and IGF-2R. Although there is a certain degree of cross-interaction between the various ligands and their receptor, most of the effectsof both IGF-1 and IGF-2 on growth and differentiation are elicited by theligand-dependent activation of IGF-1R. Both IGF-1R and IGF-2R are glyco-proteins and are located on the cell membrane. The two receptors, however,differ completely in structure and function [127, 208].

4.3.1 IGF-2R

IGF-2R is monomeric [208, 167] and does not belong to the tyrosine kinasefamily. The IGF-2R is a bifunctional binding protein that beside IGF-2,binds to proteins containing mannose-6-phosphate (M6P) and therefore isalso called the IGF-2/M6P receptor. The IGF-2 receptor has no signalingcapacity and is believed to play essential roles in (1) the intracellular traf-ficking of lysosomal enzymes by binding M6P containing ligands in the Golginetwork and delivering them to the prelysosomal compartment [87], (2) acti-vation of TGFβ [28], and (3) reducing IGF-2’s biological activity by binding,internalizing, and degrading the ligand [126].

4.3.2 IGF-1R

As mentioned earlier, IGF-1R is a member of the RTK family, and being themain protein of interest in this thesis, will be discussed in detail in the nextchapter.

30

Chapter 5

IGF-1R

5.1 Structure

Abbott et al. [1] determined that the human IGF-1R is a product of asingle gene that spans about 100 kilobases (kb) and contains 21 exons atthe distal end of chromosome 15 (bands q15-26). The 5’ untranslated region(UTR), the signal peptide, and the N-terminal non-cystine rich domain ofthe α-subunit are encoded by exon 1-3. Exons 4-10 encode the rest of theα-subunit. Exon 11 contains the proteolytic cleavage site (Arg-Lys-Arg-Arg)that generate mature α- and β-subunit. The β-subunit is encoded by exons12-21, with exon 14 encoding the transmembrane domain and exons 16-20encoding the tyrosine kinase (TK) domain [215]. Cooke et al. [33] analyzedthe promoter region and found that the 5’UTR is GC-rich and containsnumerous potential SP1 and AP2 binding sites as well as a thyroid responseelement, but no TATA or CCAAT elements.

Unlike many other family members that are activated by dimerizationupon ligand binding, IGF-1R is a heterotetrameric plasma membrane con-sisting of two identical α-subunits and two identical β-subunits, held togetherby disulfide bonds. The α-chains contribute to the formation of the cystinerich region comprising the ligand binding domain, while the β-chains havea transmembrane domain, an intracellular tyrosine kinase domain, and C-terminal domain [215].

31

5.2 IGF-1R activation

and signal transduction

Ligand binding to the α-subunit of the receptor triggers autophosphoryla-tion of the three tyrosine residues, Tyr1131, Tyr1135, Tyr1136 in the activa-tion loop (a-loop) within the kinase domain of the β-subunit of the receptor[107, 160]. Tyr1135 is the first tyrosine to be phosphorylated followed byTyr1131 and Tyr1136. These three tyrosines are transphosphorylated bythe dimeric subunit partner, causing a major conformational change in thea-loop resulting in unrestricted access of ATP and substrate to the kinaseactive site [54]. Phosphorylation of these three tyrosines increases the intrin-sic tyrosine kinase activity, which allows phosphorylation of other residuesin the β-subunit and creates high affinity binding sites for signaling proteins.Other known residues important for the function of IGF-1R are:

1. Tyr950 at the proximal juxtamembrane domain which is the main bind-ing site of signaling molecules IRS-1-4 (insulin receptor substrate 1-4),and Shc (Src homology and collagen).

2. Lys1003 which is the ATP binding site. Mutation of the ATP bindingsite or the IRS/Sch binding site abolish both proliferation and trans-formation capacities of IGF-1R [74, 34, 129, 44].

3. Tyr1250, Tyr1251, His1293, and Lys1294 in the C-terminal, althoughnot essential for signaling and mitogenic responses of IGF-1R, are im-portant for antiapoptotic functions of IGF-1R [168].

4. Ser1280 and 1283 which are the binding site of 14-3-3 proteins andimportant for IGF-1R-mediated anti-apoptotic activity [176]. Figure5.1 shows a schematic picture of IGF-1R.

In the prelapsarian days of signal transduction, pathways were linear andsimple. However, the more we learn about these pathways the more complexthey get. Their paths cross at multiple levels, each enhancing activation ofthe other. Here, by simplifying, I try to reduce the complexity to a man-ageable size, but it is important to keep in mind that there is extensiveinteraction between discussed pathways. The interaction points are criticalnodes, important for IGF-1R function.

5.2.1 PI3K pathway

One such node is IRS proteins (IRS-1-4) which are recruited to the phos-phorylated Tyr950 of the IGF-1R in a matter of minutes. IRSs have the

32

S1280-1283

aa1245C

-tail

Cystine richdomain

Y950

K1003 TK dom

ain

Y 1131Y1135Y1136

IRS-1, Shcbinding domain

ATPbinding domain

Activation loop

ß

α

Y1250-1251

H1293K1294

Antiapoptoticfunction

14-3-3binding site

Figure 5.1: Cartoon of IGF-1R with important residues for tyrosinekinase activation and signal transduction denoted. Abbreviations: Y,Tyrosine; K, Lysine; S, Serine; H, Histidine; aa, amino acid.

33

capacity to bind many adapter proteins such as Grb2, Nck, the p85 regula-tory subunit of PI3 kinase, the tyrosine phosphates SH-PTP2, the Src-likekinase Fyn, and Ca2+-ATPase SERCA1-2 (for review see [23]). The PI3 ki-nase, the p85/p110 heterodimer, is activated by tyrosine phosphorylation onthe p85 regulatory subunit. The binding of p85 to IGF-1R, either directlyor indirectly through IRS, links the IGF-1R to PI3 kinase pathway, whichleads to cellular processes such as mitosis, differentiation, and antiapopto-sis. Activated PI3 kinase induces production of second messengers, PIP2and PIP3 (phosphatidylinositol 3,4,5-trisphosphate 1 and 2), that activateskinases PDK1 and 2 (3-phosphoinositide-dependent protein kinase-1 and 2),which in turn activates Akt by phosphorylating it at Thr308 and Ser473. Aktis the second important node in this pathway, also known as PKB (proteinkinase B) or RAC-PK (A and C protein kinases). Akt is a serine/threoninekinase which in turn phosphorylates and inhibits several proapoptotic pro-teins such as the proapoptotic member of Bcl-2 family BAD, caspace 9,survival transcription factor CREB (Cyclic AMP Response Element BindingFactor), proapoptotic protein GSK-3β (glycogen synthase kinas-2), as well asmembers of the FKHR (Forkhead-related transcription factor) family [218].

Bcl-2 family proteins may be either pro-apoptotic (e.g. BAD) or anti-apoptotic (e.g. Bcl-X). BAD is known to be a heterodimeric partner forboth Bcl-X and Bcl-2, neutralizing their protective effect and promoting celldeath. By phosphorylating BAD on Ser136, Akt makes it dissociate fromthe Bcl-2/Bcl-X complex, after which BAD is sequestered into the cytosol,is bound to 14-3-3, and loses the pro-apoptotic function (see Figure 5.2).

Phosphorylation of FKHR1 by Akt leads to binding of 14-3-3 which re-stricts nuclear localization of FKHR1. [161, 21]. This impedes transcrip-tional activation of specific Forkhead target genes, such as Fas ligand (FasL),an inducer of apoptosis [22], p27KIP, an inhibitor of cell cycle progression[31], and insulin-like growth factor binding protein-1 (IGFBP-1), a presumedinhibitor of IGF-1 [115, 75].

Akt could also activate NF-κB via regulating IκB kinase (IκK), thusresult in transcription of pro-survival genes.

Akt can also stimulate mTOR, a serine/threonine protein kinase, to phos-phorylate kinase p70S6K which in turn can stimulate the initiation of proteinsynthesis through activation of S6 Ribosomal protein and other componentsof the translational machinery [177]. Another substrate for mTOR is 4E-BP1(binding protein 1). Non-phosphorylated 4E-BP1 binds tightly to the trans-lation initiation factor eIF4E (eukaryotic initiation factor 4E), preventingit from binding to 5’-capped mRNAs and recruiting them to the ribosomalinitiation complex [174]. Upon phosphorylation by mTOR, 4E-BP1 releaseseIF4E allowing it to perform its function, which is increasing translation of,

34

amongst others, Cyclin D1 [18].

Another target of Akt is MDM2 which gets phosphorylated at Ser166 andSer186, leading to the translocation of MDM2 from the cytoplasm to the cellnucleus where MDM2 decreases p53 levels [246, 143]. Schematic cascade isshown in Figure 5.2.

5.2.2 MAPK pathway

Following stimulation with IGFs, Shc is recruited to phosphorylated ty-rosine residue Tyr950, and transduces differentiation signals through theGrb2/RAS/ RAF/MAPK pathway. Since IRS-1 binds to the same phos-phorylated residue and leads to proliferative phenotype, it is possible thatcommitment to proliferation or differentiation depends on competition be-tween these two adapter proteins for access to this binding site on the IGF-1R. Another interaction between these two major signaling pathways is Grb2which, as mentioned above, can also directly bind to IRS-1.

Grb2 has one SH2 (Src homology 2) domain and two SH3 domains. Thetwo SH3 domains bind to the proline-rich regions of the guanine nucleotidereleasing factor SOS (son of sevenless) protein. The Grb2-SOS complex pre-exists in the cytoplasm of resting cells. Recruitment of Grb2 from the cytosol,where it is already bound to SOS, brings SOS in close proximity to RAS atthe plasma membrane. RAS, a small GTPase in the GDP-bound inactivestate in quiescent cells, then undergoes nucleotide exchange of GDP for GTP,which facilitates binding of the serine/threonine protein kinase Raf-1 (alsoknown as c-Raf) and its subsequent activation. This initiates a cascade of ki-nase activation: activated Raf-1 phosphorylates and activates MEK1/MEK2,which in turn phosphorylate and stimulate the MAP kinases ERK1/ERK2.Activated ERKs translocate to the nucleus and phosphorylate transcriptionfactors such as Elk-1, STAT1, STAT3, and Myc, activating gene expression(Figure 5.3).

By modifying the expression of its target genes, Myc activation results innumerous biological effects. The first to be discovered was its capability todrive cell proliferation (upregulates cyclins, downregulates p21), but it alsoplays a very important role in regulating cell growth (upregulates ribosomalRNA and proteins), apoptosis (upregulates Bcl-2), differentiation, and stemcell self-renewal. Myc is a very strong proto-oncogene and it is very oftenfound to be upregulated in many types of cancers.

35

14-3-3

Apoptosis

IRSP P

p85

p110PI3K

PIP2PIP2PIP3

mTOR

P

PDK1

Akt

elF4E

Protein synthesis

Cell Growth

GSK3β

BAD

P

Caspase 9 Caspase 9P

Survival

Akt

FKHR FKHR

Nucleus

Fas-L

4E-BP1 p70S6K

BAD

IGFBP-1

Bcl-2

PTEN

14-3-3

MDM2

MDM2P

p53

Bcl-X

Cyt-C

14-3-3P

FKHR

Figure 5.2: Schematic representation of PI3K-pathway signaling.

36

14-3-3

G1/S phase progression

ShcGrb2

RAS

ProliferationCell GrowthSurvivalMigrationDiffrenciation

MAX Nucleus

Raf

STAT1/3

SOS

GDP GTP

ERK1/2

c-Fos

CREB

c-Myc

p90RSK

P PPc-Jun P

EtsP E2FRB E2F

RBP

Cyclin D

SOSP

PSRF

Elk PP

14-3-3

BAD

P

Survival

BAD

14-3-3RAS

14-3-3MEK1/2

Figure 5.3: Schematic summary of MAPK signal transduction.

37

5.2.3 14-3-3 pathway

14-3-3 proteins are a family of conserved regulatory molecules expressed inall eukaryotic cells. 14-3-3 have the ability to bind a multitude of functionallydiverse signaling proteins, including kinases, phosphatases, and transmem-brane receptors. More than 100 signaling proteins have been reported as14-3-3 ligands.

The name 14-3-3 refers to the characteristic migration pattern of theseproteins on electrophoretic gels. In the IGF-1R context beside the two majorsignaling pathways, PI3K/Akt and the MAPK/ERK signaling pathways, athird one that results in the mitochondrial translocation of Raf-1 upon IGFsstimulation is referred to as the 14-3-3 pathway. 14-3-3 can either be activateddirectly through binding to IGF-1R and depends on the integrity of a Ser1272and Ser1283 in the C terminus of the receptor [36], or indirectly through IRS-1 [49]. Upon activation, 14-3-3 associates and inhibits proapoptotic proteinssuch as BAD and ASK1 (signal regulating kinase-1) prohibiting apoptosis.

Notably, all these three pathways result in BAD phosphorylation. Thepresence of multiple antiapoptotic pathways may explain the remarkable ef-ficacy of the IGF-1R in protecting cells from apoptosis.

5.2.4 JNK pathway

JNKs (N-terminal kinases) belong to a family of stress-activated protein ki-nases (SAPKs). The role of the SAPKs, in particular JNK, in IGF-1R sig-naling and also in tumorigenesis has been controversial. JNK is activatedin response to different stress stimuli, and signaling effects of JNK seemsto depend on whether it is transiently or persistently activated [186]. Thetransient activation, prevents apoptosis whereas the persistant activation ispro-apoptotic in a caspase-dependent manner [186]. JNK is transiently acti-vated by IGF-1 in tumor cell lines [155], and inhibition of JNK activity cansuppress IGF-1-mediated antiapoptotic activity.

5.2.5 Other interesting molecules in IGF-1R signaling

Grb10 binds to IGF-1R and many other RTKs (e.g. IR, c-Kit, VEGF).Whereas this interaction is a positive mediator of IR, VEGF, and c-Kit sig-naling through the PI3K pathway, Grb10 interaction acts as a negative mod-ulator of IGF-1R mitogenic function (will be discussed in further detail).

RACK1 is a scaffolding protein that plays an important role in the regu-lation of IGF-1R and other tyrosine kinase signaling in tumor cells. RACK1has been found to be upregulated in angiogenesis and several human carci-

38

nomas [141]. RACK1 can interact with many proteins that regulate IGF-1Rsignaling such as PKCs (protein kinase C), Src, p85, and PTP (protein ty-rosine phosphatases). Further, RACK1 directly binds to IGF-1R and tointegrin β, and modulates IGF-1R and integrin signaling [144].

One should not forget the involvement of protein tyrosine phosphatases(PTPs) in IGF-1R signal transduction pathways. The serine/threonine phos-phatases such as PP2A (Protein phosphates 2A), and the inositol phos-phatases, in particular PTEN (Phosphates and Tensin homologue deletedon chromosome 10), have also been shown to have essential functions in theattenuation of IGF-1R signaling.

5.3 The Role of IGF-1R in Cancer

The importance of IGF-1R in normal mammalian development is clear fromstudies in mice lacking functional receptors. IGF-1R null mice are 45% ofthe size of wild-type (wt) littermates at birth, and die shortly due to se-vere organ hypoplasia [133]. Mouse embryonic fibroblasts cultured fromIGF-1R null mice (R- cells) grow more slowly than wild-type fibroblasts,and are unable to proliferate under anchorage-independent conditions [193].Anchorage-independence in vitro is indicative of transformation potentialand tumorigenicity in vivo. Further, it has been known for some time thatR- cells are refractory to transformation by many oncogenes, including SV40large T and activated RAS [194, 193, 13]. There is abundant data showingthat IGF-1R regulates cancer cell proliferation, survival, and metastasis inthe transformed cell. Recent studies indicate that the requirement of theIGF-1R for transformation is related to its ability to tyrosine phosphorylateIRS-1 [41].

IGF-1R has been implicated in the initiation and development of manydifferent human cancers [148]. For example ovarian carcinoma [185], rhab-domyosarcoma [195], glial tumors [63, 145], adenocarcinoma of the colon [58],pancreatic cancer [149], and progression of colorectal adenoma to carcinoma[79] are associated with increased expression of IGF-1R. There are indica-tions pointing to IGF-1R as a promising drug target because many tumorcells undergo apoptosis when the IGF-1R is downregulated [12]. Use of adominant negative mutant receptor in tumors, to inhibit IGF-1R signaling,has been shown to cause massive apoptosis [39]. Further, there are studiesshowing inhibition of IGF-1 by blocking the action of the ligand by usingantibodies against IGF-1R will inhibit cell growth of breast cancer cell linesand Wilms’ tumor cells [7, 64]. Furthermore, reduction in IGF-1R levels us-ing antisense technologies inhibit tumor growth [196, 162]. All these findings

39

suggest the importance of IGF-1R in tumor formation and progression.Although a large body of information has been obtained over the past

decade, the mechanism by which IGF-1R mediates tumor formation andgrowth remains uncertain. Two of the most important aspects of IGF-1Rinvolvement in cancer are, firstly, its overexpression, and secondly, how thereceptor regulates its downstream signaling. One common example is loss offunction of the tumor-suppresser gene PTEN which encodes a phosphatasethat normally attenuates IGF-1R signals (see Figure 5.2).

Phosphorylation is considered to be the primary mechanism in the regu-lation of IGF-1R signaling. However, we and others have shown involvementof another post-translational modification, namely ubiquitination, of IGF-1R. The mechanisms by which ubiquitination affects the function of IGF-1Rare still poorly understood. The involvement of oncoproteins such as MDM2,Nedd4, and c-Cbl in the ubiquitination of IGF-1R suggests the importanceof IGF-1R ubiquitination for understanding the role of the receptor in cancerdevelopment.

40

Chapter 6

Ubiquitin system,general introduction

Cells must be able to respond to rapid changes in both their internal and ex-ternal environments. A particularly sensitive, rapid, and reversible responseis the post-translational modification of specific proteins. Small-moleculemodifications, such as phosphorylation, methylation, and acetylation arewell-characterized examples of post-translational events that modulate pro-tein function, but there is also a class of larger modifications where the mod-ifier itself is a small polypeptide [93]. The best-known example of this typeof modification is ubiquitination. Ubiquitination was first discovered in 1987and has since been extensively studied [20, 89]. Even though ubiquitinationwas discovered as a degradation tool which increases protein turn over in acontrolled, ATP-dependent manner, we know today that ubiquitin modifi-cations has many other functions such as cell cycle regulation, endocytosis,viral budding, transcriptional regulation, DNA repair, and the list is growing[88, 92].

Over the past 20 years, modification by multiple polypeptides distinctfrom, but related to, ubiquitin (Ub) has been discovered. This group ofproteins are collectively called ubiquitin-like modifiers (Ubls). Among UblsSUMO (small ubiquitin modifier) is the most well characterized. However,unlike ubiquitin which targets proteins for degradation the known func-tions of SUMO-modification are nuclear transport, transcriptional regulation,apoptosis, and protein stability.

41

6.1 Ubiquitin and SUMO

Ubiquitin is a highly conserved, 76 amino acids long protein, and is presentuniversally in eukaryotic cells. It is found throughout the cell, thus, givingrise to its name. Ub is transcribed as a precursor protein which is later cleavedby ubiquitin C-terminal hydrolyzes (UCHs) into single Ub moieties. Thereare several Ub genes available, which together with the recycling mechanismof conjugated Ubs, ensure high levels of ubiquitin that are required for theubiquitination system function. Ub is a remarkably stable protein due toits compact structure and abundant hydrogen bonds, and is resistant toextremes of pH and temperature. The majority of ubiquitin is present inconjugated form while only a small fraction is free. However, these pools (freeand conjugated Ub) are not static and ubiquitin cycles dynamically betweenthese pools mediated by ubiquitination and deubiquitination enzymes [78,84]. The conjugated Ub pool is not homogenous and includes mono-, multi-and several poly-ubiquitination forms.

Ubiquitin-like modifiers (Ubls) are proteins that, despite low sequencesimilarity with Ub, have almost identical tertiary structure: 5 β-strandswrapped around an α-helix [191] (Figure 6.1). To date, there are about 15known Ubls that have been shown to couple to other target protein molecules[192]. Ubl conjugation resembles Ub conjugation and each Ubl modificationhas distinct cellular functions. Among the most frequent and best studiedUbls is SUMO that is a family of small proteins which are around 100 aminoacids in length and 12 kDa in mass.

There are at least four SUMO isoforms in mammalian cells, namelySUMO-1, 2, 3, and the most recently identified SUMO-4 [19, 42]. SUMO-1 shows 47% similarity at the protein level with SUMO-2 and 3, whereasSUMO-2 and 3 are 95% identical [15]. SUMO-4 shows 87% similarity withSUMO-2 and 3, and there are indications that it is involved in diabetes type1 and 2. Within a few months different groups independently identified thefirst three members of the SUMO family. Consequently, SUMO-1 is knownunder many names, for example PIC1, Ubl1, sentrin, GMP1, Smt3c, andhSmt3, whereas its relatives SUMO-2 and SUMO-3 are also referred to assentrin2 and sentrin3, or Smt3a and Smt3b. SUMO-1 exists mainly in con-jugated form, whereas the SUMO-2 and SUMO-3 isoforms are primarily freeproteins that are subject to rapid conjugation after cellular stress. This in-dicates that the different SUMO isoforms might perform distinct functions.

42

Garce Gill. Genes and Dev. 2004

SUMO-1 Ubiquitin

Figure 6.1: tertiary structure of SUMO and ubiquitin. Ribbon dia-grams showing the similarity of the three-dimensional structures of SUMO-1and ubiquitin. Secondary structure elements are indicated: β-sheets aregreen and α-helices are red [67].

6.2 Ubiquitination and SUMOylation,

as easy as 1, 2, 3

6.2.1 Ubiquitination

The covalent linkage of Ub to target proteins requires the action of threegroups of enzymes, called E1 (Ub-activating enzymes), E2 (Ub-conjugatingenzymes), and E3 (Ub ligases), which work sequentially in a cascade. Thepoly-ubiquitination occurs through sequential E1-E2 and E2-E3 interactions.Because the El- and E3-binding sites on E2 overlap, E2 cannot be reactivatedby an E1 while it is bound to E3, and therefore poly-ubiquitination requiresmultiple E2-E3 binding events [51].

E1

The initial step in the Ub cascade is the activation of Ub by the ubiquitin-activating enzymes (E1). E1 hydrolyzes ATP to AMP and pyrophosphate togenerate a thioester bond between the active site Cysteine of the E1 and theC-terminal Glycine of the Ub.

43

E2

The activated Ub is then transferred to one of several different ubiquitin-conjugating enzymes (E2) in an ATP-independent manner. The E2 enzymesare catalytically similar to E1 in that a thioester bond is formed with Ub.The different E2 enzymes are able to interact with overlapping sets of E3ligases. All the known E2 enzymes have the same conserved globular domainwith the active Cys positioned at the highly conserved sequence. E2 enzymesshare striking similarity with another enzyme important for some Ub modifi-cations, the so called UEVs (ubiquitin conjugating enzyme variants). UEVsare similar to E2 enzymes in structure and sequence but lack an active Cyssite [217, 156]. In many cases, for the synthesis of Lys63-linked polyubiquitinchains, UEV forms a heterodimer with the E2, Ubc13, and acts as a cofactorto form an active complex. [209, 96].

E3

The ubiquitin-conjugated E2 functions in combination with an E3 ligaseto transfer Ub to the target protein. The final Ub transfer results in anisopeptide bond between the C-terminal Gly of Ub and the ε-amino group ofa Lys residue on the target protein. The E3s represent the largest and mostdiverse group of enzymes in the ubiquitination process. The E3 ligases canbe divided into two major families: the HECT (homology to E6 associatedprotein C-terminus) domain E3 enzymes and the RING (really interestingnew gene) domain E3 enzymes (Figure 6.2A).

HECT domain E3s are the first characterized ligases and they directlyconjugate ubiquitin, forming an E3-Ub thioester, and then transfer it to thetarget substrate. RING E3s, on the other hand, facilitate ubiquitinationnot by directly conjugating ubiquitin, but by acting as an adapter protein,bringing together the E2 and the substrate which allows the transfer of theUb from the E2-Ub thioester to the target protein. RING E3s are defined bycontaining an octet of cysteine and histidine residues, Cys1-X2-Cys2-X(9-39)-Cys3-X(1-3)-His-X(2-3)-Cys4-X2-Cys5-X(4-48)-Cys6-X2-Cys7, where X canbe any amino acid. It is characterized by a highly conserved spacing thatbinds two zinc ions in a cross-brace structure for conformational stability.The RING finger family has two subclasses: RING-HC and RING-H2 (whereHis replaces Cys4). These two RING domain types were found to be essentialfor catalyzing E3 ligase activity of RING-containing proteins [136]. Further,RING finger E3s can auto-ubiquitinate themselves in order to regulate theiractivity or stability. In some cases of Ub chain elongation, another enzyme,namely E4, is required [113]. The enzymes in the E4 family share a conserved

44

Gly

Isopeptidebond

ATP

Ub

AMP+PPi

E1 Ub

E1E2

E2 Ub

E3 HECT Ub E3 Ring Ub

Targetprotein

UbLys―N―C―

H

O

DUBUb

UbiquitinationA

Cys

IsopeptidebondOSUMO

ATPAMP+PPi

E1E2

Ubc9

SUMO

SEA1

SEA2

E3 ?

Lys―N―C―

Hx

ψ

E

SENP

SUMOylationB

Cys

SUMO

SUMO

SUMO

Figure 6.2: Protein modification pathways for ubiquitin (A) andSUMO (B). Both pathways require activating (E1 or SEA1/SEA2) andconjugating enzyme (E2 or Ubc9). In the ubiquitin pathway, E3 ubiquitinligases (HECT and Ring) are required for substrate selection and transfer ofubiquitin to its substrate. Recently, E3-like SUMO ligases (PIAS proteins,RanBP2 and hPC2) have been identified. Ubiquitination and SUMOylationare reversible and the Ub or SUMO modification detachment is mediated byDUBs and SENPs, respectively.

45

domain designated the ’U-box’ [86]. The U-box is a modified version of theRING-finger domain that lacks the hallmark metal-chelating residues of thelatter [57, 189] and can mediate ubiquitin-conjugation either independentfrom a classical E3 [86], or in some cases together with an E3 [113].

Ubiquitin ligation without a ligase

Exciting new data published by Hoeller et al. [95] demonstrate that E2enzymes can interact with and ubiquitinate some substrates without theparticipation of E3 ligases, both in vitro and in intact cells. Common forall these substrates are that they contain various types of UBDs (ubiquitin-binding domains). UBDs are a collection of modular protein domains thatnon-covalently bind to ubiquitin, a process termed ”coupled monoubiquiti-nation” [233]. Hoeller et al. [95] Showed that chimeric proteins containing aUBD fused to an unrelated polypeptide, such as glutathione S-transferase(GST), can also be monoubiquitinated in an E3-independent reaction invitro. Whereas these data indicate a low specificity of the E3-independentubiquitination, there was certain selectivity in the potency among various E2enzymes to catalyze ubiquitination of UBD-containing proteins. Ubiquitin-UBD interaction was necessary for the coupled ubiquitination. Mutation inthe UBD that disrupted its binding to ubiquitin, as well as the mutation ofubiquitin Ile44 disrupting its binding to a UBD, prevented E3-independentubiquitination. The authors hypothesize that the UBD proteins themselvesdo not function as ordinary E3 ligases. They promote their own ubiquitina-tion but they do not appear able to promote transfer of ubiquitin to anotherprotein (Figure 6.3).

6.2.2 SUMOylation

Similar to ubiquitination, covalent attachment of SUMO to its target proteinsrequires three or four enzymes, namely (1) SUMO proteases, (2) the SUMOE1 SAE1/SAE2 heterodimer, (3) the SUMO E2 Ubc9, and in some cases (4)the SUMO E3s PIAS, RanBP2, and CtBP2.

SUMO proteases

At least seven genes in mammalian genomes encode proteins with the Ulp(ubiquitin-like protease) domain which constitute the so-called Sentrin-spe-cific protease (SENP) family [239]. They all possess specific subcellular lo-calization properties that restricts substrate availability and could be anexplanation for distinct functions described for different SUMO proteases.

46

UBD

Reconstructed from, molecular cell, Alexandra Sorkin, 2007

KUb

S

(1)

SH

K

Ub

(2)

UBD

KUb

UBD

UBD*

K

Ub

UBD(4) (3)

E2Cys E2

Cys

Figure 6.3: Hypothetical mechanism and consequences of the UBD-coupled, E3-independent monoubiquitination. (1) A UBD-containingprotein binds to an E2-conjugating enzyme by means of the interaction ofthe UBD with ubiquitin linked to E2 via a thioester bond. (2) The ubiquitinis then transferred to a ε-amino group of the lysine residue in the UBD-containing protein. (3) The ubiquitin moiety of the UBD-containing proteincan interact either with the UBD of the same substrate molecule or (4) theUBD (UBD*) of another protein.

47

The SUMO pre-protein is not active until the last four amino acids of theC-terminus have been cleaved off to make the C-terminal glycine 97 residueaccessible, the so called SUMO maturation. SUMO proteases, such as SENP-1 and 2, possess two enzymatic activities and can function both as carboxyterminal hydrolyzes providing mature SUMO, or as isopeptidases removingSUMO from SUMO-conjugated substrates.

SUMO E1

SUMO E1 activating enzyme is a heterodimer (SAE1/SAE2 also known asAos1/Uba2) with two separate enzymatic activities: while the SAE1 (orAos1) is responsible for adenylation, the SAE2 (or Uba2) makes the thioesterbond. First SAE1 utilizes ATP to adenylate the C-terminal glycine of SUMO-1. Formation of a thioester bond between the C-terminal glycine of SUMO-1and a cysteine residue in SAE2 is accompanied by the release of AMP.

SUMO E2

This activation step is followed by conjugation of the activated SUMO toE2-conjugating enzyme Ubc9 at cysteine residue 93 through the formationof yet another thioester bond (transesterification).

SUMO E3

Earlier in vitro studies showed that SUMO E1 and E2 were biochemicallysufficient for transfer of SUMO to specific target proteins [147]. For SUMOmodification, consensus motif ΨKXE (where Ψ represents a large hydropho-bic amino acid, and X represents any amino acid) is preferred. Many knownSUMO conjugation sites occur within this motif, but SUMOylation alsooccurs on lysine residues located within non-consensus regions. Primarily,SUMO-1 conjugation have been found outside the consensus region. Ubc9has the ability to indirectly bind to the ΨKXE motif in vitro and in vivoand SUMOylate the substrate. These facts led to the suspicion that theSUMOylation might not require E3s. But this suspicion was erased by thediscovery of E3 ligases. Based on current understanding there are three typesof SUMO E3s – PIAS proteins, RanBP2, and Pc2 – with important functionin specifying the SUMO substrate. In the last step, an isopeptide bond isformed between SUMO proteins and substrates through the mutual actionof Ubc9 and SUMO E3 ligases. (Fig 6.2B)

48

6.3 Ubiquitin/SUMO modification

and function

As already mentioned, there are different types of Ub modification which canconsist of covalent attachment of a single ubiquitin (mono-ubiquitination),several ubiquitin monomers (multiple-mono-ubiquitination), or a polymericchain of ubiquitin composed of ubiquitin monomers joined by covalent bondsbetween the C-terminus of one ubiquitin and one of the seven lysine sidechains on the next ubiquitin (poly-ubiquitination). Different types of ubiq-uitination lead to different cellular outcomes.

6.3.1 Mono-ubiquitination

Modification with only one ubiquitin molecule, where the Gly at the C-terminus of Ub binds to the internal Lys of the target protein, regulatesseveral process such as Histone regulation, endocytosis, and virus budding[92]. The mono-Ub modification mediate its function due to the fact that themono-ubiquitinated proteins can be recognized by a multitude of ubiquitin-binding proteins that enable the modified proteins to interact and form com-plexes with other proteins, change the subcellular localization, and alter cer-tain structural and targeting properties (Figure 6.4A).

6.3.2 Multiple-mono-ubiquitination

Multiple-mono-ubiquitination, where mono-Ub is bound at multiple Lys resi-dues of the target protein, has recently been discovered. Multiple-mono-ubiquitination of RTKs, EGFR, and PDGFR is important for internalization,signaling, and lysosomal degradation of the receptors (Fig 6.4A).

6.3.3 Poly-ubiquitination

Poly ubiquitination occurs due to the fact that ubiquitin contains sevenlysines (Lys6, Lys11, Lys27, Lys29, Lys33, Lys48, and Lys63) that can actas sites of self-conjugation. It has been shown that Ub is modified at all theseven lysine residues in vitro, providing a grand diversity in polyubiquitinchains [175].

Early on, attention focused on poly-Ub modification Lys48 linked chainsthat primarily target proteins for proteasomal degradation. Lys48 ubiqui-tination has also been reported to have other functions such as enhancing

49

K

Ub

K K K

Ub Ub Ub

K48K48

K

K48K48

K

K63K63K63K63

K29K29

K

K29K29

K K

K6K6 K6

K6K11

K11 K11

K11

Mono-ubiquitination• Endocytosis• Membrane trafficking• DNA repair• Histone regulation

Multiple-mono-ubiquitination• Endocytosis• Lysosomal degradation

Lys48-poly-ubiquitination• Proteasomal degradation

Lys63-poly-ubiquitination• Endocytosis• DNA repair

Lys29-poly-ubiquitination• Proteasomal degradation

Lys6/11-poly-ubiquitination• Unknown

A

K

NH2 K48K48

K48K48

N-terminal-ubiquitination• Degradation

NH

KNH2 COOH

C O

NHC O

G76K48

G76K48

Lys48-poly-ubiquitination

NHC O

G76K48

KNH COOHC

O

G76K48

N-terminal-ubiquitination

B

UbUb

Ub

UbUb

Ub

Figure 6.4: Ubiquitin modification. A. Schematic representation of thedifferent Ub modifications with their functional roles. B. Ubiquitination onthe internal Lysine and on the N-terminal residue of the target substrate.Abbreviations: Ub, Ubiquitin; K, Lysine residue; G, Glycine residue.

50

endosomal trafficking and proteasomal independent degradation of Met re-ceptors [27]. Four ubiquitins connected together through Lys48 (Ub4-Lys48),is a minimum signal for proteasomal recognition [213].

Poly-ubiquitination through Lys63 depends on the engagement of an E2different from those used for Lys48 polyubiquitin chains: a dimeric enzymecomposed of Mms2(Uev1A)/Ubc13 [96, 97]. The overall topology of Lys63polyubiquitin chains differ from that of chains formed at Lys48 in that Lys63chains are more elongated and extended, whereas Lys48 chains are morecompact and globular [212]. This difference in topology may explain whyLys63 ubiquitination does not target its substrate for degradation and insteadplays a role in cellular signaling [118], DNA repair [205, 94], and proteinlocalization [94].

Lys29 polyubiqutination, though rare, has been detected in vivo. Theknown substrates are non cleavable UFD (Ub fusion degradation) substrates[104]. Lidsten et al. identified UBB+1 (also a UFD substrate) as a substratefor Lys29 ubiquitination. Interestingly, in order for UBB+1 to be directed toproteasomal degradation it must be modified by both Lys48 and Lys29.

Lys6-modified ubiquitin is a potent and specific inhibitor of ubiquitin-mediated protein degradation [37]. The tumor suppresser important for ovar-ian and breast cancer, BRCA1, when heterodimerized with BARD1 is ableto catalyze the production of lys6-linked ubiquitin polymers, and it has beenhypothesized that lys6 ubiquitination play a role in DNA repair [166].

Lys11 polyubiquitination has been shown, in vitro, to target proteins forproteasomal degradation although their biological function is still unknown(Figure 6.4A).

6.3.4 N-terminal ubiquitination

While the majority of polyubiquitinations take place at the ε-NH2 group of aninternal lysine residue, in some cases the first ubiquitin moiety is conjugatedto a free α-NH2 group of the N-terminal residue refereed to as the N-terminalubiquitination leading to proteasomal-mediated degradation. Thirteen pro-teins have been identified to undergo N-terminal ubiquitination.

The different functions of various Ub-chains could be due to the factthat combinations of different topologies in different ratios, each assemblesdifferent structures, and thus might serve as a potential docking site fora distinct protein interaction and therefore distinct functions [180] (Figure6.4B).

51

6.3.5 Mono- and poly-SUMOylation

Initially, it was doubtful if SUMO could form polymerization in the same wayas Ub, as the lysines 29, 48, and 63 of Ub are absent in SUMO. However, itwas shown that SUMO-2 and SUMO-3 can form poly-SUMO chains by lys11conjugation, which is a perfect ΨKXE consensus motif. More recently it wasshown that SUMO-1 can perform poly-SUMOylation in vitro predominantlyvia Lys7, Lys16, and Lys17. SUMO-1 is able to SUMOylate lysine residuesoutside the consensus motif even though it is rare.

The function of SUMO chains has not yet been fully elucidated, butSUMO modification has been implicated in nuclear transport, chromosomesegregation, transcriptional regulation, apoptosis, and protein stability.

There are many questions that needs to be answered in this rapidly ex-panding field of research. For example, how is the selectivity between theSUMO isoforms achieved? Do some proteins have preferable affinity for aparticular SUMO isoform, and if so, what are the structural determinantsfor this selectivity? Does SUMO function as a mere tag that recruits someinteraction partners that the unSUMOylated target protein couldn’t recruit,or does SUMO mediate its function by preventing access to some part ofthe target protein’s surface? Moreover, we need to identify other proteinsthat are involved in the SUMOylation process and work out how mono- orpoly-SUMOylation affect the function and location.

6.4 Deubiquitination and deSUMOylation

Protein ubiquitination and SUMOylation are dynamic processes involvingconjugation enzymes as described above and also deconjugating enzymesknown as deubiquitination enzymes (DUBs) and deSUMOylation enzymes

6.4.1 Deubiquitination

The human genome project has led to identification of nearly 100 poten-tial DUBs [29, 230]. DUBs are a large group of proteases that could bedivided into four classes based on their similarity and mechanism of action:UBP (ubiquitin-specific processing protease) family, UCH (Ub C-terminalhydrolase) family, OTU (Ovarian tumor related protease) family, and JAMM(Jab1/Pad1/MPN-domain metallo-enzyme) family. UBP and UCH are thelargest families.

Deubiquitinating enzymes act at multiple points in the ubiquitin Path-way:

52

1. These enzymes are responsible for the processing of linear polyubiquitinchains to generate free ubiquitin from precursor fusion proteins [231].

2. these enzymes impact on free ubiquitin pools by recycling ubiquitinfrom branched-chain polyubiquitin [55, 229].

3. Deubiquitinating enzymes also remove ubiquitin from Ub-conjugatedtarget proteins, thereby regulating the localization or the activity ofthe targets.

4. Another biochemical role of deubiquitinating enzymes is to removeubiquitin from a ubiquitinated target protein and to rescue the pro-tein from degradation by the 26S proteasome. This could occur eitherbefore or after the ubiquitinated substrate interacts with the protea-some [120, 119].

5. Finally, deubiquitinating enzymes play an important role in clearingthe proteasome of peptide remnants conjugated to ubiquitin chains.Removal of ubiquitin from these protein remnants may “unclog” thebarrel-shaped proteasome and thereby maintain or increase proteaso-mal activity (for review see [110]).

6.4.2 DeSUMOylation

SUMO-specific proteases participate in both processing of immature SUMOand removal of SUMO conjugates from substrates. There are at least sevendeSUMOylases known as SENPs (SUMO/sentrin-specific proteases) encodedby six genes in mammalian cells. Although both the known SUMO-specificproteases and the ubiquitin-specific proteases are cysteine proteases, these en-zymes share no sequence similarity. The mammalian SUMO proteases havebeen found to have distinct subcellular localization. SENP1 is a nuclear pro-tease that deconjugates a large number of SUMOylated proteins [70]. SENP2is a nuclear envelope-associated protease that, when overexpressed, appearsto have an activity similar to that of SENP1 [70, 82, 244, 101]. SENP5 con-stitute a family of nucleolar SENPs with preference for SUMO-2/3 [71]. Thethird SENP family consists of nuclear proteases SENP6 and SENP7 withunknown function. Although the ability of SENPs to reverse SUMOylationis well established, the specificity of each SENP and the overall biologic rel-evance of the SENP family remains undefined. Undoubtedly, future studieswill reveal more about the functional specificity of the different SUMO pro-teases.

53

6.5 Ubiquitin dependent degradation

Proteins in cells are in a dynamic state, being continuously synthesized anddegraded. For a long time, lysosomes were known as the only place wereproteins could be degraded in the cell where most of protein degradationoccurred after pinocytosis. Brial Poole in 1978 [182] could demonstrate in-volvement of another degradation pathway for endogenous proteins whichhe summarized: “The exogenous proteins will be broken down in the lyso-somes, while the endogenous proteins will be broken down wherever it is thatendogenous proteins are broken down”, explicitly predicting existence of anon-lysosomal proteolytic system [182, 16].

For a long time the fact that this non-lysosomal proteolytic system re-quired ATP energy in order to degrade energy-rich proteins puzzled manyuntil the discovery of the ubiquitin-proteasome system resolved the enigma.Hough et al. discovered the 26S proteasome by its ability to degrade Ub-lysozyme conjugates [99]. Based on current understanding, the 26S protea-some has several functions dependent on its ability to degrade lys48 ubiq-uitinated substrates: (1) eliminate misfolded or aberrant proteins [3, 5], (2)maintain a free amino acid pool [77, 216], and (3) regulate the half-time ofproteins whose concentrations must vary over time in the cell [90].

The 26S proteasome is a large multisubunit complex composed of a coreproteinase (CP), known as the 20S proteasome, and a pair of symmetri-cally disposed 19S regulatory particles (RP). The 20S CP is a barrel-shapedstructure composed of four stacked rings, two identical outer α-rings andtwo identical inner β-rings. The eukaryotic α and β-rings are composed eachof a set of seven distinct subunits, giving the 20S complex a pseudo-sevenfold symmetry and a general structure of α1-7β1-7β1-7α1-7 [72, 73]. Thecatalytic sites are localized to some of the β-subunits. Each extremity ofthe 20S barrel can be capped by a 19S RP which has many functions inregulating proteasomal activity. It activates the 20S proteasome, serves asa docking site for poly-ubiquitinated proteins, and unfolds and translocatespolypeptides into the 20S catalytic core because a folded protein would notbe able to fit through the narrow proteasomal channel. The 19S RP containssix different ATPase subunits. After degradation of the substrate, short pep-tides derived from the substrate, as well as reusable ubiquitin, are released(Figure 6.5).

54

Peptides

Lys48 tagged protein K48

K48K48

K48

Ub

Ub

UbUb

Figure 6.5: Schematic illustration of the proteasome. Following sub-strate degradation, the ubiquitin tags and peptides are recycled for futureuse.

55

6.6 IGF-1R ubiquitination

Two distinct E3 ubiquitin ligases, MDM2 and Nedd4, have been shown tobe involved in the ubiquitination of IGF-1R. MDM2 is a member of theRING finger domain E3 class while Nedd4 is a member of HECT domain E3ubiquitin ligases. In MDM2-mediated ubiquitination, β-arrestin functions asa molecular scaffold in bridging the ligase to the receptor (Paper I). Similarly,Nedd4-mediated IGF-1R ubiquitination requires Grb10 as an adapter protein[219].

6.6.1 MDM2

The oncoprotein RING finger domain E3 MDM2 (Murine double minute 2)is a highly regulated protein, primarily known as a p53-specific E3 ubiquitinligase, and acts as a suppressor of p53 activity. The p53 protein, which in-hibits cellular proliferation and induces cell death, lies at the heart of stressresponse pathways that prevent growth and survival of potentially malig-nant cells. In unstressed cells, MDM2 constantly monoubiquitinates p53leading to export from the nucleus, which is the critical step in mediating itsdegradation by proteasomes. Because MDM2 is also a transcriptional targetof p53, an autoregulatory feedback loop exists in which increased activityof p53 leads to increased expression of its own negative regulator. Loss ofMDM2 function (such as by mutations or knockout) results in p53-drivenapoptosis early in embryogenesis [40], whereas overexpression leads to p53loss and tumor development. With stress or DNA damage, p53 and MDM2are phosphorylated and acetylated, and they dissociate [146]. Subsequently,p53 forms a tetramer that acts as transcriptional activator leading eitherto cell cycle arrest through the increased expression of the tumor suppres-sor protein ARF [226], or apoptosis through the increased expression of theproapoptotic Bcl-2 family member Bax [154] (Figure 2.1). Furthermore, asmentioned earlier, after growth factor stimulation, MDM2 is phosphorylatedby Akt and enters the nucleus. This leads to reduction of both p53 levelsand transactivation activity [143].

Previous work in our group identified MDM2 as an E3 ligase for IGF-1R.While the major amount of MDM2 is nuclear, there is a pool of cytoplas-mic MDM2 as well. The relationship between MDM2 in the nucleus andMDM2 in the cytosol remains unclear. Undoubtedly, the major role of nu-clear MDM2 is to regulate p53 function by protein-protein interaction, de-pending on MDM2 concentration. The concentration of MDM2 in the cytosoldepends on its interaction with p53 in the nucleus, the activity of the PI3K-Akt pathway, and its degradation rate. The balance among these processes

56

determines how much MDM2 is cytosolic and available for ubiquitinationof IGF-1R. Intriguingly, another connection between the IGF-1R pathwayand the MDM2-p53 pathway is that overexpression of p53 represses IGF-1Rtranscription [68]. Based on current knowledge, cytoplasmic MDM2 inducesubiquitination and internalization of IGF-1R. An interesting and unansweredquestion is whether MDM2 is degraded together with IGF-1R. This wouldaffect the duration of the change in p53 activity. A massive p53 expressionwould deplete the cytoplasmic MDM2 pool, and as a consequence of thisincrease, the expression of IGF-1R would be enhanced, and simultaneously,the receptor degradation would decrease, favoring survival.

Another intriguing discovery by our group was identifying β-arrestin asan adapter protein which brings MDM2 to the IGF-1R and make MDM2-mediated ubiquitination possible (Paper I).

6.6.2 β-arrestin

There are four known mammalian arrestins. Visual and cone arrestins areexpressed in photoreceptors, regulating rhodopsin and cone opsins, respec-tively, whereas ubiquitously expressed β-arrestins 1 and 2 regulate manyGPCRs. The amino acid sequences of β-arrestin 1 and 2 isoforms are 78%identical with most of the coding differences appearing in the C-terminal.Knockout studies show that mice lacking either β-arrestin 1 or 2 are viable,whereas the double-knockout phenotype is embryonic lethal [43], implyingthat each β-arrestin functionally substitutes for the other isoform to somedegree. They were first characterized as blocking signal transduction by pre-venting interaction between receptor cytoplasmic domains and heterotrimericG proteins. Arrestins are now known to participate in endocytosis and sig-naling of GPCRs. By acting as an adapter they bring activated receptors toclathrin-coated pits for endocytosis, a process critical for receptor recyclingand degradation. β-arrestins also bind to various other proteins implicated inreceptor internalization [199, 124]. Importantly, both ubiquitination and de-phosphorylation of β-arrestins appear essential to agonist-dependent receptorendocytosis [130, 200]. Mammalian arrestins exist in a constitutively phos-phorylated state in the cytosol and are dephosphorylated upon binding toactivated GPCRs at the plasma membrane. Dephosphorylation of β-arrestinsat the plasma membrane is necessary for engaging endocytic partners, suchas clathrin, but does not seem to affect receptor binding or desensitization[130]. Furthermore, β-arrestin ubiquitination, which is mediated by MDM2,is required for rapid receptor internalization [200]. In addition to their clas-sic roles in desensitization and internalization, β-arrestins can also act assignaling scaffolds for many pathways, in particular those of the MAPKs

57

(see [43] for review). Despite these suggested roles, β-arrestins and primarilyβ-arrestin 1 act as an adapter protein important for MDM2-mediated ubiq-uitination of IGF-1R and this discovery will be discussed in detail in theresults section (paper I and II).

6.6.3 Nedd4

Since the original discovery of Nedd4 (neural precursor cells-expressed devel-opmentally down-regulated 4) gene, nine related proteins (Nedd4, Nedd4-2/Nedd4L, WWP1/Tiul1, WWP2, AIP4/Itch, Smurf1, Smurf2, HecW1/NEDL1, and HecW2/NEDL2) have been identified.

The assertion that Nedd4-like E3s play a role in cancer is supported bythe overexpression of Smurf2 in esophageal squamous cell carcinoma, WWP1in prostate and breast cancer, Nedd4 in prostate and bladder cancer, andSmurf1 in pancreatic cancer. Several Nedd4-like E3s directly regulate vari-ous cancer related transcription factors from the Smad, p53, KLF, RUNX,and Jun families. Additionally, Nedd4-like E3s regulate ubiquitin-mediatedtrafficking, lysosomal or proteasomal degradation, and nuclear translocationof multiple membrane receptors, e.g. TGFβ receptor, Notch receptor, EGFR,VEGFR2, Chemokine (C-X-C motif) receptor 4, and IGF-1R. Further, theymodulate important signaling pathways involved in tumorigenesis like TGF-β, EGF, IGF, VEGF, SDF-1, and TNFα.

Vecchione and colleagues found that Nedd4 forms a complex with Grb10and IGF-1R in mouse embryo fibroblasts (MEF). This interaction is mediatedby the SH2 domain of Grb10 and the C2 domain of Nedd4 [158]. Interestingly,IGF-1R, but not Grb10, was shown to be ubiquitinated, suggesting thatGrb10 serves as an adapter to mediate the interaction between Nedd4 andIGF-1R. In support of this notion, overexpression of a catalytically inactiveform of Nedd4 decreased IGF-1R ubiquitination and degradation upon IGF-1stimulation [219].

6.6.4 Grb10

Grb10 is a member of the Grb proteins (growth factor receptor-bound pro-teins) which includes at least seven isoforms of structurally related mul-tidomain adapters with diverse cellular functions, e.g. Grb7 appears pre-dominantly to be involved in focal adhesion kinase-mediated cell migration,whereas Grb10 and Grb14, in particular, have been implicated in the regula-tion of IGF-1R and IR signaling. Grb10 was originally identified as a bindingpartner of tyrosine-phosphorylated EGFR [170]. Grb10 contain a conservedSH2 (Src homology 2) domain which occurs in the C-terminal, a pleckstrin

58

homology (PH) domain in the central region, and a less conserved N-terminuscontaining proline-rich sequences which may provide binding sites for SH3domain-containing proteins, as well as a BPS domain (between the PH andSH2 domains), which mediates binding to activated IGF-1R [157].

6.7 IGF-1R SUMOylation and

nuclear translocation

We discovered that IGF-1R can also be SUMOylated (Paper V) and trans-ported to the nucleus. Being at the early stage of this discovery the functionof this modification still remains to be elucidated. Intriguingly, SUMOylatedIGF-1R was exclusively found in the nucleus and when SUMOylation wasblocked upon SENP-2 overexpression the nuclear IGF-1R could no longerbe detected. These two observations suggest a role for SUMOylation of thereceptor in its nuclear import. Further, the nuclear transport of IGF-1R wasligand dependent (For more detail see under Results and Discussion, PaperV).

6.7.1 Nuclear RTKs

Several other RTKs, including EGFR, mouse ErbB1, HER-2, rat p185neu,HER-3, truncated C-terminal HER-4, FGFR, TrkA/B, FGFR, and VEGFR-2, undergo nuclear transport [183, 138, 197, 179, 142, 237, 132, 164, 169, 190,223], but the functional importance of this process remains unclear.

Nuclear import of cell-surface receptors occurs in both ligand-dependentand ligand-independent manner. Cell surface receptors which are translo-cated to the nucleus in a ligand-dependent manner include EGFR, FGFR,IFN-γR, IL-1R, type I TGF-β receptor, and IL-5R [38, 102, 121, 247, 132].In the case of HER-2 which lacks the ligand-binding domain, its kinase ac-tivity is required for nuclear entry [223]. Nuclear import of EGFR and typeI TGF-β receptor can also occur in a ligand-independent manner [247, 48].

In the light of the observation that many cell-surface receptors appear toexist as non-membrane-bound receptors in the nucleus, it is thus speculatedthat cells may utilize a general mechanism for passage through the nuclearpore complex (NPC). However, such mechanism has not yet been identified.It remains to be answered if SUMOylation could be a general mechanism bywhich nuclear import of cell surface receptors occur.

The affect of RTKs on gene transcription, mediated through their canon-ical signaling as discussed earlier, is a confirmed mechanism. Beyond thiswell-established function of RTK signaling, direct communication between

59

the RTK and transcription has also been reported. Transcriptional targetsof nuclear EGFR and HER-2, those identified so far, are closely involved intumorigenesis and tumor proliferation and progression [132, 223, 80, 135].The traditional and nuclear-signaling of EGFR is summarized in Figure6.6. Both cyclin D1 and B-Myb are positive regulators of G1/S progres-sion [132, 103]. COX-2 and iNOS are enzymes that produce prostaglandinsand nitric oxide, respectively, and emerge as major targets for chemopreven-tion and chemotherapy [76, 238]. Consistently, a positive correlation betweennuclear EGFR and cyclin D1/iNOS in a cohort of breast carcinomas has beenfound [135]. Further, a correlation of nuclear accumulation of EGFR/mouseErbB1 and their ligands, EGF/TGF-β, with cell proliferation/DNA synthe-sis has been reported by several studies [142, 190]. In agreement with itsrole in proliferation/DNA synthesis, EGFR undergoes nuclear translocaliza-tion in regenerating livers [142], pregnant uterus, and proliferative basal cellswithin normal mouth mucosa [132]. A potential role of nuclear EGFR inDNA damage/repair in response to irradiation/oxidative stress has also beensuggested [48].

Given the notion that the EGFR receptor family lacks a putative DNA-binding domain, it is suspected that these receptors first associate withDNA-binding transcription factors and then enhance target gene transcrip-tion via their intrinsic transactivational activity. In this regard, nuclearEGFR interacts with STAT3 and co-regulates iNOS expression [135]. Nu-clear EGFR/E2F1 complex activates expression of B-Myb, a positive regula-tor of G1/S cell cycle progression [80]. Nuclear HER-4 forms a complex withSTAT5a and co-activates β-casein gene promoter [232]. However, it is sug-gested that HER-4 may have a weak transactivational activity and requiresinteraction with a strong transcription co-activator YAP for its gene regula-tory function [114]. In line with this observation, nuclear FGFR associateswith and activates transcription co-activator CBP to upregulated gene pro-moters [52]. Known targets of nuclear FGFR include FGF-2, neurofilament-L, and tyrosine hydroxylase [159, 207]. In support of this notion, EGF,EGFR, and TrkA and its ligand NGF have been shown to bind to chromatin[183, 105] and Schwannoma-derived growth factor, a ligand for EGFR, boundto AT-rich DNA sequences [112]. Together, these data suggest an emergingrole that nuclear receptors play in transcriptional regulation.

Using chromatin immunoprecipitation (ChIP) based cloning strategy, weidentified DNA sequences to which IGF-1R binds. The characteristics ofthese sequences resembled those of transcriptional regulatory elements, distalenhancers, rather than promoter regions.

60

B Nuclear EGFR pathway

E2F1

Reconstructed from, H-W Lo and M-C Hung, 2006

PLC-γ PI3K RAS STATs

Nuclear targets

TumorigenesisProliferationMetastasis

Chemoresistanceradioresistance

Cyclin D1

STAT3 DNA-PK

B-Myb iNOS DNA repaire

G1/S progressionproliferation

ProliferationMetastasis

radioresistance

A Cytoplasmic EGFR pathway

EGFRCytoplasm

Nucleus

EGFVitamin D EGF

E2F

Radaiation, cisplatin

Heat, H2O2

Figure 6.6: Cytoplasmic and nuclear modes of the EGFR signalingpathway. The EGFR signaling pathway exerts its biological effects via twomajor modes of actions: (A) cytopslamic/traditional and (B) nuclear modes.

61

6.8 Enhancers

Eukaryotic genomes contain both protein coding and non-coding DNA. Non-coding DNA harbors a variety of regulatory elements called transcriptionregulatory regions (TRR), also referred to as transcription factor bindingsites (TFBSs), promoter modules, or cis-regulatory modules (CRMs). Otherterms used for TRRs include booster, activator, locus control region, up-stream activating sequence, and upstream repressing sequence, which areoften collectively called enhancers in the literature [171].

Eukaryotic genomes contain both protein coding and non-coding DNA.Non-coding DNA harbors a variety of regulatory elements called transcrip-tion regulatory regions (TRR), transcription factor binding sites (TFBSs),promoter modules, or cis-regulatory modules (CRMs). The term enhancerwas originally defined as a regulatory element that elevates transcription ina position- and orientation-independent manner. The term has since beenused in a broader sense to refer to any region of DNA that produces a spe-cific aspect of a transcription profile, sometimes even referring to regions thatsuppress transcription [235].

In DNA sequences, the base where transcription starts is called the tran-scription the transcription start site (TSS). The TSS of a transcription unitis conventionally numbered +1. Bases extending in the direction of tran-scription (downstream) are assigned positive numbers and those extendingin the opposite direction (upstream) are assigned negative numbers. A DNAsequence region upstream of the TSS are usually called the promoter. A core(or basal) promoter is located approximately between -40 and +35 relativeto the TSS of the genes [202].

Although necessary for transcription, the core promoter is not a commonpoint of gene regulation. It cannot by itself generate functionally significantlevels of mRNA. Most proteins that bind to the core promoter are ubiqui-tously expressed and provide little regulatory specificity. The production offunctionally significant levels of mRNA is controlled through sequence spe-cific binding of transcription factors (TFs) to DNA sequences outside of thecore promoter [235, 25], the so called enhancers.

An examination of well-characterized eukaryotic promoters suggests thatit is not unusual to have 10-50 enhancers affecting a single promoter [235, 6].

The positions of enhancers relative to TSSs differ enormously amonggenes. Often, they are located within a few kilo-bases upstream of the TSS[235], but they can be found at >30 kb upstream or downstream of the TSS[201, 106, 47]. One extreme example is an enhancer of the Shh locus in bothhumans and mice that lies ∼800 kb from the TSS [128]. Further, they can belocated within the 5’ UTR [24], within introns [241, 11], and in rare instances,

62

even in coding exons [163, 188]. Some enhancers lie on the far side of an ad-jacent locus [235]. The diversity of enhancer positions is possible because ofthe ability of DNA to loop and bend, allowing interaction between proteinsbinding on DNA at distant sites with regard to the primary structure [235].

63

64

Part II

Results and Discussion

65

Chapter 7

Results and Discussion

7.1 Paper I

β-Arrestin Is Crucial for Ubiquitination and Down-regulation ofthe Insulin-like Growth Factor-1 Receptor by Acting as Adapterfor the MDM2 E3 Ligase

By the time I joined the group, MDM2 had recently been identified asan E3 ligase for IGF-1R. Unexpected new data by Lin et al. introduced β-arrestins (β-arresin 1 and 2) as IGF-1R binding proteins with the capacity toenhance the receptor internalization. Additionally, previous work by Shenoyet al. had demonstrated β-arrestins functioning as adapter proteins regulat-ing the ubiquitination of β-adrenergic receptors. Considering these findings,it was tempting to hypothesize that β-arrestins could bind to IGF-1R as anadapter protein and utilize receptor ubiquitination.

With the hypothesis in mind we investigated whether mutated MDM2with deletion at the β-arrestin binding site (MDM21−400) could bind andubiquitinate IGF-1R. Wt MDM2 (MDM21−491), and mutated MDM2 withdeletion of the both ligase domain and β-arrestin binding sites (MDM21−161)were used as positive and negative control, respectively. As expected, WtMDM2 associated and increased ubiquitination and degradation of IGF-1R,while mutated MDM21−161 (with double deletion of ligase- and β-arrestinbinding domains) did not affect the ubiquitination or degradation of the re-ceptor. However, MDM21−400 (lacking the β -arrestin binding domain) inhib-ited the ubiquitination of the receptor. Performing co-immunoprecipitation,we could observe that β-arrestin 1 and 2 (β1 and β2), MDM2, and IGF-1Rare associated in one big complex. Using in vitro (cell free) ubiquitinationassay it was demonstrated that β-arrestins are essential for MDM2-mediatedubiquitination of IGF-1R. Further, the role of β1 and β2 in inducing ubiquiti-

67

nation and degradation of IGF-1R was studied in cell system by both overex-pression and downregulation of these two genes. Consistent with in vitro re-sults, over expression of β1 and β2 increased ubiquitination and degradationof the receptor while downregulation had the opposite affect. Furthermore,β1 was more potent than β2 in mediating ubiquitination and increasing theturn over of the IGF-1R. Altogether, in this paper we identified β-arrestinsas adapter proteins which brings MDM2 and IGF-1R together, accentuatingubiquitination and degradation of the receptor. In contrast to GPCRs, inwhich β2 is more important for ubiquitination and degradation of the recep-tor, β1 is the primarily adapter protein involved in MDM2-mediated IGF-1Rubiquitination. Further studies was needed in order to elucidate the role ofβ-arrestin/MDM2-mediated regulation of IGF-1R.

7.2 Paper II

β-arrestin and MDM2 Mediated IGF-1R Stimulated ERK Activa-tion and Cell Cycle Progression

Having establish the role of β-arrestins in internalization and degradationof IGF-1R our attention was drawn towards their role in IGF-1R signaling.β-arrestins have a dual role in GPCR signaling. The classical role where β-arrestins function as negative regulators occurs due to the fact that bindingof β-arrestin to activated GPCR uncouples the receptor from cognate G-proteins leading to internalization and signal attenuation [123]. However,β-arrestins also function as a positive signaling regulator, acting as scaffoldproteins that interact with several cytoplasmic proteins and link GPCRs tointracellular signaling pathways such as MAPK cascades [125, 124].

First, we investigated the requirement of β1, being the most potent ofthe two arrestins in IGF-1R binding, in ERK phosphorylation as a measure-ment for MAPK activation. By comparing ERK activation of β1 knock out(KO) mouse fibroblasts (MEFs) with MEFs having endogenous wt β1 andβ1-KO-MEF with β1 reintroduced, we could conclude that β1 is not only im-portant for the onset of ERK activation upon IGF-1 stimulation but also forits duration. Consistent with this data, when β1 was down regulated (usingsiRNA against β1) in human melanoma BE cells, the ERK phosphorylationwas significantly reduced compared with mock transfected BE suggesting theimportance of β1 for IGF-1R ERK signaling. However, as a consequence ofβ2 downregulation, ERK phosphorylation was sustained. Intriguingly, IGF-1stimulation addressed β-arrestins into endocytic vesicles. Further, upon β2knock down using siRNA, β1 was constitutively localized in endocytic vesi-cles even in unstimulated cells. Recently, it was shown that one important

68

consequence of β-arrestin dependent MAPK activation is the compartmen-talization of the signals [198]. The endosomal MAPK activity is a probableexplanation for sustained ERK signaling when endocytosis is enhanced uponβ2 downregulation.

As evidenced by above mentioned studies, upon IGF-1 stimulation β-arrestin functions as both an endocytic and a signaling adapter for IGF-1R and bears a functional analogy to adapter proteins such as Eps15 andTRAF6. Mono-ubiquitination of Eps15 and poly-ubiquitination of TRAF6have great impact on EGFR internalization and NFκB signaling, respectively[181, 221]. Further, β2-Ub chimera was shown to enhance β2-adrenergicreceptor internalization. Could β1 ubiquitination play a role in increasedinternalization and signaling of IGF-1R? First we showed that β1 was ubiq-uitinated upon IGF-1 stimulation. Then, by comparing β1 ubiquitination inMEF cells and MDM2-KO-MEFs, we could show that MDM2, the E3 ligasefor IGF-1R, is also a ligase for β1. Further, MDM2 overexpression in BEcells revealed that MDM2 increases the endocytosis of β-arrestin as well asERK activation giving a greater understanding in the integration role of β1ubiquitination in trafficking and signaling.

Using wt and various IGF-1R mutants (C-terminal truncated IGF-1Rdenoted 56; mutated IGF-1R at the Shc and IRS binding site denoted 46;double mutate IGF-1R lacking both C-terminal and Shc/IRS binding sitedenoted 96) we could show that the C-terminal domain of IGF-1R is impor-tant for β1 recruitment, ubiquitination, and endocytosis, and additionally,determined that the C-terminal is crucial for ERK activation upon IGF-1stimulation.

MAPK/ERK is the major signaling pathway that is required for G1-phasecell cycle progression after mitogenic stimulation. While the transient earlyERK onset triggers immediate early response, the sustained ERK signal playsa pivotal role in enabling cells to re-enter the cell cycle and progress into Sphase mainly by cyclin D1 expression [122, 227]. Based on the findings thatβ1 plays an important role in both transient and sustained ERK activation,we sough to examine the consequences of MDM2/β-arrestin-mediated ERKsignaling in cell cycle progression. Upon transaction of wt MDM2 and domi-nant negative (DN) MDM2 in human melanoma cell line BE and R+ (MEFsstably overexpressing IGF-1R), the DNA content were analyzed using FACSafter propidium iodide staining. These data demonstrated that MDM2 isimportant for G1/S phase progression upon IGF-1 stimulation. DN MDM2caused G1 accumulation similar to MEK1 inhibitor PD98059. Using IGF-1R negative mouse cell line (R-) we could demonstrate that the effect wasthrough IGF-1R.

By summarizing paper I and II I now explain the scenario as we know it

69

today. Upon IGF-1 stimulation,β1 is recruited and binds to the C-terminaldomain of the receptor working as an adapter protein making it possiblefor MDM2 to bind and ubiquitinate both IGF-1R and β1. In which orderthis happens is still an open question. Further, the ubiquitination of β1regulates the endosomal trafficking and ERK activation which is importantfor cell cycle progression upon IGF-1 stimulation. We still don’t know ifβ-arrestin serves as a locus of control deciding MAPK/ERK activation andduration, and consequently deciding how cells should proliferate, or wheterβ-arrestin is just a “glue” that makes communication between regulatingproteins possible.

7.3 Paper III

Role of Ubiquitination in IGF-1 Receptor Signaling and Degrada-tion

IGF-1R is undergoing several types of post-translational modifications,including phosphorylation and ubiquitination. These modifications usuallyfunction as molecular switches that control the activity and function of pro-teins. The aim of this paper was to investigate the relationship betweenphosphorylation and ubiquitination and their effects on biological role ofIGF-1R. Using cells expressing IGF-1R with targeted mutations, we coulddemonstrate that IGF-1R ubiquitination is dependent on the status of IGF-1R phosphorylation and that tyrosine kinase activity is prerequisite for theubiquitination to happen.

Further, we could show that the C-terminal tail of the receptor is criticalfor IGF-1R ubiquitination. There are 29 lysine residues in the β subunit ofIGF-1R of which only 3 exist in the C-terminal domain. The depletion ofIGF-1R ubiquitination in C-terminal truncated mutant could either be dueto that these 3 lysine residues are the main ubiquitination sites or that C-terminal is an important binding site of other regulating proteins involved inthis process. As shown in paper II, C-terminal truncation inhibits β-arrestinbinding to the receptor. The reason why the C-terminal of the receptor iscrucial for so many functions still remains unclear.

Signaling through IGF-1R is an area of intense investigation. Althoughthere is a huge body of literature focusing on tyrosine kinase activity of thereceptor and its role in cell signaling, there is only a limited understanding ofhow the ligand dependent ubiquitination is regulated and how this postmod-ification affects the cell signaling through IGF-1R. In order to investigatethis, we studied ERK and Akt activation in cells expressing wt and variousmutant IGF-1R with different phosphorylation and ubiquitination patterns.

70

We could conclude that in order to activate both ERK and Akt pathways,the receptor must be phosphorylated and ubiquitinated. Further, while ERKactivation seemed to be ubiquitin dependent, the Akt activation required acomplete IGF-1R kinase activity. This is based on the fact that two mutants(point mutated IGF-1R at Tyr 1136 denoted Tyr1136F and mutated IGF-1R at the Tyr 950 denoted Tyr950F) impaired in phosphorylation could notactivate Akt.

As an ubiquitinated protein, IGF-1R could be a potential target for pro-teasomal degradation. On the other hand, EGFR ubiquitination results inendocytosis and mainly, if not entirely, undergoes lysosomal degradation. Inorder to determine if degradation of IGF-1R occurs in proteasomes and/orlysosomes, specific proteasome- and lysosome-inhibitors (PI and LyI) wereused. Wt IGF-1R was mainly degraded by lysosomes, but nevertheless, theproteasome inhibitor epoxomicin inhibited the degradation suggesting theinvolvement of proteasome in IGF-1R degradation. It could be argued thatthe non-physiological conditions caused by PI, leading to elimination of thefree pool of Ub, could alter ubiquitination and lysosomal degradation of thereceptor. Taking advantage of C-terminal truncated IGF-1R, unable to beubiquitinated, we could show that this receptor was entirely degraded in lyso-somes with no effect after PI treatment. Unexpectedly, two of the mutantswith impaired phosphorylation capacities and unable to activate Akt, weremainly degraded through proteasomes. The reason for this phenomenon isstill unclear. There are other studies showing that activated Akt inhibitsprotein degradation [53]. Further, activated Akt leads to phosphorylationand translocation of MDM-2 into the nucleus [85]. By inhibiting MDM2phosphorylation, one might increase the cytoplasmic pool of MDM2 empha-sizing its role in ubiquitination and proteasomal degradation of IGF-1R. Thisexplanation is merely speculative and further investigations are needed.

In summery, our results demonstrate a dual function for IGF-1R ubiqui-tination: receptor degradation and ERK signaling activation.

7.4 Paper IV

Identification of c-Cbl as a New Ligase for IGF-1R with DistinctRoles to MDM2 in Receptor Ubiquitination and Endocytosis

This paper started with us being curious about details in MDM2-mediatedIGF-1R ubiquitination. Studies by Haglund et al. demonstrating involve-ment of multi-ubiquitination of EGFR using Ub-specific antibodies, andHuang et al. showing poly-ubiquitination of EGFR using mass spectrom-etry caught out attention. The type of Ub modification, mono-, multi-, and

71

the different lysine residue poly-ubiquitination, influences the fate and thefunction of the ubiquitinated protein [228], and is therefore of great inter-est. In order to investigate the type of IGF-1R Ub modification mediated byMDM2, we used two human osteosarcoma cell lines, U2OS and SAOS2, cho-sen due to the difference in their MDM2 expression. While U2OS expresseshigh levels of MDM2, SAOS-2 is p53 negative and therefore exhibits very lowMDM2 expression.

The first interesting observation was that while Ub-IGF-1R was seen atlow doses (5-50ng/mL) in U2OS it was seen at high doses (50-100ng/mL)in SAOS2 cells. The fact that the MDM2 negative cell line SAOS2 alsomediated IGF-1R ubiquitination, suggested involvement of another ligase.

Surprisingly, using specific poly-Ub antibody (FK1) we could detect poly-ubiquitination of IGF-1R in both cell lines. Knowing that the other estab-lished ligase for IGF-1R Nedd4 preferably multi-ubiquitinates the receptor,we assumed a third ligase must be involved. We started our search in thejungle of E3 ligases with the most probable one. c-Cbl being a knownligasefor many RTKs was the best candidate. We detected IGF-1R/c-Cbl complexin HEK293 cells under basal (DMEM+10% serum) conditions. Moreover,MDM2 associated with IGF-1R when cells were stimulated with low doseIGF-1 (5ng/mL) while IGF-1R/c-Cbl complex were seen mainly after highdose IGF-1 (100ng/mL). By overexpressing MDM2 and c-Cbl in HEK293cells, we could demonstrate that MDM2-mediated IGF-1R ubiquitinationoccurs after low dose stimulation, whereas c-Cbl-mediated ubiquitinationhappen after high dose stimulation.

With the knowledge that both ligases are capable of poly-ubiquitinatingIGF-1R, the next interesting question was if there was a preference for MDM2and c-Cbl-mediated poly-ubiquitination when it came to the choice of theLysine residue used to form the ubiquitin chain. Using various mutated Ubin an in vitro ubiquitination assay, we demonstrated that MDM2 modifiesIGF-1R with Lys-63 chain, whereas c-Cbl causes lys-48 ubiquitination of thereceptor.

Moreover, MDM2-mediated IGF-1R ubiquitination addressed the recep-tor to endosomes, whereas IGF-1R ubiquitinated by c-Cbl was internalizedvia caveolin/lipid raft route. With this new insight, it is interesting to studyhow these different Ub modifications would affect signaling and degradationof the receptor.

72

7.5 Paper V

SUMOylation, Nuclear Translocation, and Transcriptional Activityof IGF-1 Receptor

The idea for this paper started with the two SUMOylation site predic-tion tools, SUMOplot and SUMOsp, indicating 11 high-probability sites forSUMO modification in the IGF-1R, three of which were overlapping betweenpredictions. Using different cell lines we could demonstrate that IGF-1R wasindeed SUMOylated, and that the SUMO-IGF-1R bands intensified afterUbc9 overexpression. The observation that three distinct bands appearedsuggested that IGF-1R is SUMO-1-conjugated at three different sites. Thekinetic of SUMOylation after IGF-1 stimulation was relatively long (6-8h)compared to ubiquitination which is a much more rapid modification startingafter 1 min stimulation and declining after 20 min.

Because SUMOylation is primarily regarded as a nuclear or perinuclearphenomenon, we got interested in the localization of the SUMOylated IGF-1R. Using subcellular fractionalization we could demonstrate that, underbasal conditions, a fraction of IGF-1R was nuclear and we could concludethat the SUMO-conjugated IGF-1R is exclusively in the nucleus. Consistantwith the SUMOylation kinetics, we could see that IGF-1R was imported intothe nucleus after 6-8h IGF-1 stimulation.

The next question was what would a ∼200kDa membrane bound proteindo in the nucleus? Does IGF-1R have any DNA binding capabilities? Aftersynthesizing random, double-stranded, 40bp long DNA sequences, we per-formed EMSA (electrophoretic mobility shift assay) and could observe thatIGF-1R could bind to 4 out of the 17 probes tested indicating specificityin its binding. Further, using chromatin immunoprecipitation (ChIP) basedcloning strategy we could identify some of the genomic DNA sequences towhich IGF-1R binds. The obtained sequences were analyzed using the BLATtool of the UCSC genome browser [109]. They were located in different chro-mosomes, 4 of which were intronic. The distances to closest genes were ratherlarge (18-475 kb). These features suggest that the sequences represent en-hancers [187, 206].

To investigate if these sequences would affect transcription, they werecloned into a luciferase vector with a weak promoter. If our sequences wereenhancers, we should see an increase in the transcription of the luciferasegene. After transfecting HEK293 cells and HEK293 cells overexpressingIGF-1R with our reporters, we could show that five of the seven sequencesincreased the transcriptional activity between 200%-1500%.

Furthermore, the fact that IGF-1R overexpression markedly enhancedthis activity, suggests a role of the IGF-1R in transcription activation by

73

binding to enhancer regions.Enhancers usually exhibit specific binding motifs to which transcription

factors or activating proteins bind [17]. It is a challenging task to find thesebinding motifs. We used several motif finding programs, such as MEME [10]and CONSENSUS [91], in order to find a motif in the random sequencesand the ChIP sequences for which the luciferase experiment yielded positiveresults. IGF-1R binds to 4 of the 17 randomly generated DNA sequences.Since the sum of the lengths of the sequences is 17 · 40 = 680 base-pairs, weestimate that, on average, IGF-1R binds to at least one position every 170base-pairs. More accurately, assume that l is the length of the DNA sequencemotif to which IGF-1R binds. Let p be the probability of IGF-1R bindingto a random DNA sequence of length l. The probability that IGF-1R bindsto a random DNA sequence of length 40 is then approximately

1− (1− p)40−l+1

Our best estimate of this probability from data is 417

. Equating these twoexpressions and rearranging, we get

p = 1−(

13

17

) 11−40+1

.

For l between 4 and 15, the expected number of base-pairs between hits variesbetween 138 and 97. We can contrast this with, for example, data from [224]where the estimated frequency of potential binding sites for known transcrip-tion factors varies between 190 and 7100 base-pairs between hits. Also notethat the binding specificity depends on the actual experiment and might notreflect the specificity within the environment of the cell. The consensus se-quence of the best motif found was GTCCTGCCT. However, the expectedfrequency computed for the motif was 2.25 which is not statistically signifi-cant (values below 1 are desired). Expected frequency of a motif is defined asthe number of motifs, with the same or better statistical properties (p-valueand length), that we would expect to find by chance [91]. Simulations withrandom sequences showed that 35 to more than 100 sequences of length 190bp are needed to have a 50significant motif for a transcription factor withthe same binding specificity as that of IGF-1R.

The indirect role of IGF-1R in gene transcription through receptor sig-naling is well established. Based on our findings, we believe that IGF-1Rmight have a more direct role in gene transcription by binding to enhancerregions. In order to get any further in identifying gene(s) directly affectedby nuclear IGF-1R, there are several pieces of the puzzle that must be put inplace. Finding the SUMO-modification sites of IGF-1R and mutating them

74

will give us a tool by which we can more closely study the affect of nuclearIGF-1R. Further, by analyzing all the sequences from the ChIP cloning weshould have enough material to find a putative consensus region to whichIGF-1R binds. Having a consensus region would be a valuable tool in iden-tification of genes affected by nuclear IGF-1R.

75

76

Chapter 8

Concluding remarks

IGF-1R is required for establishment of the transformed state, tumor cellinvasion, and metastasis. Epidemiological evidence links elevated expressionlevels of IGFs or IGF-1R with the pathogenesis of several cancers, e.g. breast,prostate, and colon cancer, with resistance to radiotherapy. Dramatic anti-tumor responses have been observed in a variety of experimental systemswhen the function of the IGF-1R has been impaired. There is a major needfor highly specific, less toxic anti-cancer therapies, and IGF-1R is a wellvalidated target. A better understanding of the IGF-1R mechanism wouldprovide us with the tools needed for targeting key molecules essential forIGF-1R function.

The work presented in this thesis has been an attempt at understandingthe mechanisms by which IGF-1R is regulated upon the posttranslationalmodifications ubiquitination and SUMOylation. I would like to summarizethis thesis in five concluding remarks:

First, β-arrestins are adapter proteins essential for MDM2/IGF-1R bind-ing which makes the MDM2-mediated ubiquitination of IGF-1R possible.Further, β1 binds to IGF-1R with a higher affinity than β2. β1/MDM2-mediated IGF-1R ubiquitination has a dual role: on the one hand it enhancesERK activation and cell cycle progression upon IGF-1 stimulation, and onthe other hand it increases degradation of the receptor.

Second, the C-terminal domain of IGF-1R has a functional role beingimportant for β-arrestin binding and the overall ubiquitination and ERKactivation of the receptor.

Third, IGF-1R is degraded mainly by the lysosomal pathway, but nev-ertheless, the proteasomal pathway is also important for degradation of Ub-modified IGF-1R.

Fourth, the E3 ligase c-Cbl lys48 ubiquitinates the receptor, while the E3ligase MDM2 lys63 ubiquitinates IGF-1R. Further, c-Cbl ubiquitinates IGF-

77

1R in the presence of high dose IGF-1 causing caveolin/raft internalization,while MDM2 ubiquitinates IGF-1R at low dose IGF-1 stimulation resultingin clathrin-dependent internalization.

And fifth, IGF-1R is SUMOylated, translocated to the nucleus, and bybinding to enhancer regions might affect the gene transcription.

In summary, ubiquitination of IGF-1R is much more involved in the re-ceptor function than we believed when I started. I have mentioned the dualrole of ubiquitination in IGF-1R function several times. The interestingquestion is, if not ubiquitination itself, what is the “switch” deciding whenthe receptor should increase signaling and cell cycle progression, and whenit should be degraded leading to signal attenuation? Could the E3-specificreceptor ubiquitination be part of the answer? Could the two degradationpathways involved in IGF-1R degradation solve the mystery? Could theadapter protein be the regulator? Could a variety of adapter proteins me-diate ubiquitination through one and the same ligase and somehow causedifferent functional outcomes? Another interesting discovery in this thesiswas involvement of SUMO-modification. How does SUMOylation affect thenuclear translocation? By inhibiting overall SUMOylation we could detectthat nuclear IGF-1R is eliminated. Is SUMOylation of the receptor directlyinvolved in the the nuclear import or is another protein responsible? Whichgenes are directly affected by nuclear pool of IGF-1R? Are there any nuclearproteins phosphorylated by nuclear IGF-1R and what would the outcomebe? As usually is the case in research, as soon as you answer one questionyou create many more. I hope this work has contributed to an increased un-derstanding of the complex function of IGF-1R, as well as creating questionsinteresting enough to be answered.

78

Acknowledgements

What a ride! The work on this thesis has been an inspiring, often exciting,sometimes challenging, but always interesting experience whereby I have beenaccompanied and supported by many people to whom I owe many thanks.

I am grateful to my supervisor, Olle Larsson, for letting me mature froma student into an independent researcher. For your continual encouragementand for always letting me try things my way. For your “Laissez-Faire” styleof leadership and your constant open door. For being open minded, evenwhen the hypotheses tested were not in line with the scientific dogmas. Foryour great sense of humor which takes some getting used to. I still can’t besure when you’re joking and when you’re not, even when it comes to seriousmatters! (and I know that’s what makes you enjoy them so much :o))

To my co-supervisor, Leonard Girnita, for introducing me to the mostinteresting field of science, ubiquitination, and for mentoring my introductionto the lab.

A special thanks to my best friend in the lab, Paddy the dark, for yourout of proportion intelligence, scientific creativity, your absolutely fantasticdark sense of humor, for the stress ball!!, for sharing your music even thougheverything is not possible to listen to!, but most of all for your friendship.

I would like to thank my labmates; Sandra, for all the laughter andhelp. Your honesty and sincerity is truly inspiring. Thank you for bringingGerman organizational skills in my otherwise chaotic way of work! Thankyou for your generosity, for encouraging me to give sushi a chance and mostof all for sharing “signaling” frustration with me :o). I will always cherish thewonderful time spent together. Linda, I thank you for “sharing the burden”with me as we wrote on our thesis together. When we were fighting to finishour “last” experiments, if we had measured our adrenaline, I think we wouldhave broken records! It was good to have a comrade like you during thishassling period!

I would also like to acknowledge past and present members of Larssonslab: Maria, for your great advice and support in the beginning of my Ph.D.period. Lotta, for a fantastic time in Sicily and for “feeding my ego” with

79

your kind words whenever we meet. Mario, for being the coolest Ph.D. Iknow. I was sincerely impressed by your calmness during the last periodbefore your defence. Radu, for you laid-back attitude and for having per-spective in life. Diana, for you humbleness and for your passion in science.Natalia, for your independent personality and for thinking outside the box,and Ada.

I am also grateful to the members of Klas Wiman’s, Olle Sangfeldt’s, andAnders Zetterberg’s group for making the 4:th floor a friendly place to workin. A special thanks goes to Salah, Jinfeng, Jermy, Nina Rokaeus for be-ing dear colleagues. Nader, for your infinite positive energy, and Vladimir,for your never fading benevolent attitude.

During this work I have been blessed with the help of many friends andcolleagues for whom I have great regard, and I wish to extend my warmestthanks to: Mikael Lerner, for all your scientific and unscientific advice.CCK feels so much safer knowing that you are always on the third floorready to help. Thank you! Florian Salomons, for all the help with con-focal microscopy. You are a truly skilled researcher. It is a Ph.D. student’sdream come true to be tutored by you. Nico, for your great knowledge inthe ubiquitination field and for your generosity in sharing it. Malihe, foryour warm personality. Goncalo Castelo-Branco, for all the help withthe ChIP. Paola, for your scientific enthusiasm and inspirational attitude.Jens Lagergren, Sudha Shenoy, and Prof Lefkowitz for fruitful col-laborations. Prof Baserga for providing the overall wisdom and a broaderview on science. Klas Wiman, your door has always been open to discussscientific issues. Thank you for your valuable help. Hodjatt Rabbani, foryour vast knowledge in molecular biology and all the help with cloning andprimer design. Ali moshfegh, for listening to my scientific dilemmas andfor your valuable suggestions. Linn, for being a great listener and support.Katja Pokrovskaja, for all the help with Luciferase assay. I would alsolike to thank Arne Ostman, for your advice, Anders Zetterberg, for youdedication to the scientific world, and Olle Sangfeldt for your generousinputs.

Ann-Britt and Evi, for all the help with administrative business duringmy Ph.D. studies. Hating paper work as much as I do, I would be lostwithout your help.

I am in dept to many friends that made my life so rich.My best friend Nina, for being the great person that you are. For con-

stantly challenging, stimulating, and supporting me. I feel so privileged tohave you as my friend. Thank you for always being there during the goodtimes and the bad.

Rahele, for being a proof that friendship has no time or geographic

80

boundaries. For the wonderful time in Iran and India. Amir, for your jokesthat made me cry of laughter, and for all the contemplation about to be or notto be (in an academic world after defence!). Elahe, for you calm and soothingattitude, a wisdom that does not come with age, and for all the laughter.Jonas, for being the master of argumentation, the only person able to makeme speechless! We all love you the way you are, cocky, talkative, being able tospice up any environment and for being absolutely wonderful. Niloofar, foryour golden heart. Nicole, for being the exotic mixture of soccer enthusiast,HARD Rock music lover, cleaning pedant, and the most lovely, cozy person Iknow with a laughter that makes sadness vanish. Marjan, for your unselfishfriendship. Mikael Hansson, for being the most mature person I know andfor always being ready to help. Alexandra Ekman, for your genuinity andwarmth. Theresa Vincent, for being so contagiously energetic, and foryour high morals. The research world needs you.

I had the advantage of having many teachers at KI, who offered me theirtrust and assistance in many aspects of my professional life. I would liketo thank Carl-Johan Sundberg, the illuminating star with a tremendouscharisma. Thank you for your positive and helpful attitude. You are areal inspiration source. Stefan Einhorn, for your mentorship. Jan-olovHoog, for your support during all these years. You are the reason I stayed inresearch. Your first lecture in DNA replication mesmerized me with the worldof science and has not stopped amazing me to this day. Tomas Cronholm,for the great biomedicine education that I have put to use on an everydaybasis during my Ph.D. period. Ewa Ehrenborg, for your encouragementduring the many years at KI. Klas Karre, for being a role model. MariaMasucci, for your great knowledge and inspirational courses throughout theyears.

I would also like to thank the present and former memberes of our smalldance group. My dance teacher, Ulrika Larssen, for your invigoratingdance lessons. No matter how stressed and scattered I came to your lessonsyou made me centered and focused. Thank you for all the happiness thatyou brought into my life with dancing. Edith, thank you for your sweetpersonality, your amazing capacity to remember all the steps in choreographyand being my private “teacher” whenever needed, and Petra for being awonderful person.

I consider myself fortunate to have met you, Tomita Sensei. Thank youfor your guidance as a real master with deep knowledge, not only in Budoculture, but life in general. It has been a great honor to be your student.

I owe my loving thanks to may family: My Mom, for all the love, and forteaching me the right values in life. For equipping me with all the resourcesany young woman needs, self-confidence and emotional security. My Dad,

81

for initiating and cherishing my curiosity as the only biologist in the family!and for teaching me to ask why. You are truly missed. My closest sister,Anita, for introducing me to your way of thinking, and challenging me witha new view of reality. For being my greatest supporter and harshest critic.For knowing me better than myself and still liking me ;o). Dansita-Clara,for being the greatest source of joy, helping to put things in perspective. Ishould also thank you for being the most cheerful person in the lab duringyour many visits. My brother, Benjamin-Hossein, for all your help withthe design of the front page of this thesis, and also all the help with paperIII. For your Carpe-Diem style of living, sarcastic sense of humor, and mostof all for being you. By the way, my book came out before yours ;o). Tomy youngest sister, Mahdis, for all your positive energy, for embracing lifeto the fullest, and for making me proud so many times. My oldest sister,Rozita, for your calm and composed attitude in life. Parnia, for your greatpersonality. Parmida, for being the most determined young lady I know.

My husband Ali: without exaggeration I could write a book as thickas this thesis just to thank you. Let me just summerize it by saying thatwithout your unconditional love, support, and help in all aspects of life itwould have been impossible for me to finish this work.

This thesis has been supported by grants from the Swedish Cancer So-ciety, the Cancer Society in Stockholm, the Swedish Research Council, theSwedish Children Cancer Society, Ingabritt and Arne Lundberg’s researchfoundation, the King Gustaf V’s research foundation, Alex and Eva Wall-strom’s foundation and the Karolinska Institute.

82

Bibliography

[1] AM Abbott, R. Bueno, MT Pedrini, JM Murray, and RJ Smith.Insulin-like growth factor I receptor gene structure. Journal of Bio-logical Chemistry, 267(15):10759–10763, 1992.

[2] M. Alimandi, L.M. Wang, D. Bottaro, C.C. Lee, A. Kuo, M. Frankel,P. Fedi, C. Tang, M. Lippman, and J.H. Pierce. Epidermal growth fac-tor and betacellulin mediate signal transduction through co-expressedErbB2 and ErbB3 receptors. The EMBO Journal, 16:5608–5617, 1997.

[3] A.Y. Amerik and M. Hochstrasser. Mechanism and function of deubiq-uitinating enzymes. Biochimica et Biophysica Acta (BBA)-MolecularCell Research, 1695(1-3):189–207, 2004.

[4] AE Aplin, A. Howe, SK Alahari, and RL Juliano. Signal Transduc-tion and Signal Modulation by Cell Adhesion Receptors: The Role ofIntegrins, Cadherins, Immunoglobulin-Cell Adhesion Molecules, andSelectins. Pharmacological Reviews, 50(2):197–264, 1998.

[5] HC Ardley and PA Robinson. E3 ubiquitin ligases. Essays Biochem,41(1):15–30, 2005.

[6] MI Arnone. The hardwiring of development: organization and functionof genomic regulatory systems, 1997.

[7] C.L. Arteaga, LJ Kitten, EB Coronado, S. Jacobs, FC Kull Jr,DC Allred, and CK Osborne. Blockade of the type I somatomedinreceptor inhibits growth of human breast cancer cells in athymic mice.J Clin Invest, 84(5):1418–1423, 1989.

[8] F. Authier, RA Rachubinski, BI Posner, and JJ Bergeron. Endosomalproteolysis of insulin by an acidic thiol metalloprotease unrelated to in-sulin degrading enzyme. Journal of Biological Chemistry, 269(4):3010–3016, 1994.

83

[9] PC Baass, GM Di Guglielmo, F. Authier, BI Posner, and JJ Bergeron.Compartmentalized signal transduction by receptor tyrosine kinases.Trends Cell Biol, 5(12):465–70, 1995.

[10] TL Bailey and C. Elkan. Fitting a mixture model by expectation max-imization to discover motifs in biopolymers. Proc Int Conf Intell SystMol Biol, 2:28–36, 1994.

[11] M.J. Bamshad, S. Mummidi, E. Gonzalez, S.S. Ahuja, D.M. Dunn,W.S. Watkins, S. Wooding, A.C. Stone, L.B. Jorde, R.B. Weiss, et al. Astrong signature of balancing selection in the 5’cis-regulatory region ofCCR5. Proceedings of the National Academy of Sciences, 99(16):10539–10544, 2002.

[12] R. Baserga. The insulin-like growth factor-I receptor as a target forcancer therapy. Expert Opin Ther Targets, 9(4):753–68, 2005.

[13] R. Baserga, A. Hongo, M. Rubini, M. Prisco, and B. Valentinis. TheIGF-I receptor in cell growth, transformation and apoptosis. BiochimBiophys Acta, 1332(3):F105–26, 1997.

[14] K. Basler, D. Yen, A. Tomlinson, and E. Hafen. Reprogramming cellfate in the developing Drosophila retina: transformation of R7 cellsby ectopic expression of rough. Genes & Development, 4(5):728–739,1990.

[15] P. Bayer, A. Arndt, S. Metzger, R. Mahajan, F. Melchior, R. Jaenicke,and J. Becker. Structure Determination of the Small Ubiquitin-relatedModifier SUMO-1. Journal of Molecular Biology, 280(2):275–286, 1998.

[16] JWC Bird, AM Spanier, WN Schwartz, H. Segal, and D. Doyle. ProteinTurnover and Lysosome Function, 1978.

[17] Elizabeth M. Blackwood and James T. Kadonaga. Going the Distance:A Current View of Enhancer Action. Science, 281(5373):60–63, 1998.

[18] P. Blume-Jensen and T. Hunter. Oncogenic kinase signalling. Nature,411:355–365, 2001.

[19] K.M. Bohren, V. Nadkarni, J.H. Song, K.H. Gabbay, and D. Owerbach.A M55V Polymorphism in a Novel SUMO Gene (SUMO-4) Differen-tially Activates Heat Shock Transcription Factors and Is Associatedwith Susceptibility to Type I Diabetes Mellitus. Journal of BiologicalChemistry, 279(26):27233, 2004.

84

[20] J.S. Bonifacino and A.M. Weissman. Ubiquitin and the control ofprotein fate in the secetory and endocytic pathways 1. Annual Reviewof Cell and Developmental Biology, 14(1):19–57, 1998.

[21] A.M. Brownawell, G.J.P.L. Kops, I.G. Macara, and B.M.T. Burgering.Inhibition of Nuclear Import by Protein Kinase B (Akt) Regulates theSubcellular Distribution and Activity of the Forkhead TranscriptionFactor AFX. Molecular and Cellular Biology, 21(10):3534–3546, 2001.

[22] A. Brunet. Transcription factors of the Forkhead family are targetproteins for Akt involved in apoptosis. MS-PARIS-, 15:897–898, 1999.

[23] AA Butler, S. Yakar, IH Gewolb, M. Karas, Y. Okubo, and D. LeRoith.Insulin-like growth factor-I receptor signal transduction: at the inter-face between physiology and cell biology. Comp Biochem Physiol BBiochem Mol Biol, 121(1):19–26, 1998.

[24] V.C. Calhoun, A. Stathopoulos, and M. Levine. Promoter-proximal tethering elements regulate enhancer-promoter specificity inthe DrosophilaAntennapedia complex. Proceedings of the NationalAcademy of Sciences, 99(14):9243–9247, 2002.

[25] M.F. Carey and S.T. Smale. Transcriptional Regulation in Eukaryotes:Concepts, Strategies, and Techniques. Cold Spring Harbor LaboratoryPr, 1999.

[26] G. Carpenter. The EGF receptor: a nexus for trafficking and signaling.BioEssays, 22(8):697–707, 2000.

[27] S. Carter, S. Urbe, and M.J. Clague. The Met Receptor DegradationPathway: Requirement for lys48-linked polyubiquitin independent ofproteasome activity. Journal of Biological Chemistry, 279(51):52835,2004.

[28] A. Chen, B.H. Davis, M.D. Sitrin, T.A. Brasitus, and M. Bisson-nette. Transforming growth factor-β1 signaling contributes to Caco-2cell growth inhibition induced by 1, 25 (OH) 2D3. American Journalof Physiology- Gastrointestinal and Liver Physiology, 283(4):864–874,2002.

[29] C.H. Chung and S.H. Baek. Deubiquitinating Enzymes: Their Di-versity and Emerging Roles. Biochemical and Biophysical ResearchCommunications, 266(3):633–640, 1999.

85

[30] M.B. Clarke and V. Sperandio. Events at the Host-Microbial Interfaceof the Gastrointestinal Tract III. Cell-to-cell signaling among microbialflora, host, and pathogens: there is a whole lot of talking going on, 2005.

[31] M. Collado, R.H. Medema, I. Garcia-Cao, M.L.N. Dubuisson, M. Bar-radas, J. Glassford, C. Rivas, B.M.T. Burgering, M. Serrano, andE.W.F. Lam. Inhibition of the Phosphoinositide 3-Kinase PathwayInduces a Senescence-like Arrest Mediated by p27Kip1. Journal ofBiological Chemistry, 275(29):21960–21968, 2000.

[32] PF Collett-Solberg and P. Cohen. The role of the insulin-like growthfactor binding proteins and the IGFBP proteases in modulating IGFaction. Endocrinol Metab Clin North Am, 25(3):591–614, 1996.

[33] R.M. Cooke, T.S. Harvey, and I.D. Campbell. Solution structure ofhuman insulin-like growth factor 1: a nuclear magnetic resonance andrestrained molecular dynamics study. Biochemistry, 30(22):5484–5491,1991.

[34] D. Coppola, A. Ferber, M. Miura, C. Sell, C. D’Ambrosio, R. Rubin,and R. Baserga. A functional insulin-like growth factor I receptor isrequired for the mitogenic and transforming activities of the epidermalgrowth factor receptor. Molecular and Cellular Biology, 14(7):4588–4595, 1994.

[35] S. Cory, DL Vaux, A. Strasser, AW Harris, and JM Adams. Insightsfrom Bcl-2 and Myc: malignancy involves abrogation of apoptosis aswell as sustained proliferation. Cancer Res, 59(7 Suppl):1685s–1692s,1999.

[36] A. Craparo, R. Freund, and TA Gustafson. 14-3-3 epsilon interacts withthe insulin-like growth factor 1 receptor and insulin receptor substrateI in phosphotyrosine-independent a manner. J Biol Chem, 272:11663–11670, 1997.

[37] D. Cripps, S. Thomas, Y. Jeng, F. Yang, P. Davies, and A. Yang.Alzheimer’s disease-specific conformation of hyperphosphorylatedPHF-tau is polyubiquitinated through lys-48, lys-11, and lys-6 ubiq-uitin conjugation. J Biol Chem, 281:10825–10838, 2006.

[38] BM Curtis. IL-1 and its receptor are translocated to the nucleus. TheJournal of Immunology, 144(4):1295–1303, 1990.

86

[39] C. D’Ambrosio, A. Ferber, M. Resnicoff, and R. Baserga. A solubleinsulin-like growth factor I receptor that induces apoptosis of tumorcells in vivo and inhibits tumorigenesis cancer. Res, 56:4013–4020,1996.

[40] S. de Rozieres, R. Maya, M. Oren, and G. Lozano. The loss of mdm2induces p53 mediated apoptosis. Oncogene, 19(13):1691–1697, 2000.

[41] T. DeAngelis, J. Chen, A. Wu, M. Prisco, and R. Baserga. Transfor-mation by the simian virus 40 T antigen is regulated by IGF-I receptorand IRS-1 signaling. Oncogene, 25:32–42, 2006.

[42] J.M.P. Desterro, M.S. Rodriguez, G.D. Kemp, and R.T. Hay. Identifi-cation of the Enzyme Required for Activation of the Small Ubiquitin-like Protein SUMO-1. Journal of Biological Chemistry, 274(15):10618–10624, 1999.

[43] S.M. Dewire, S. Ahn, R.J. Lefkowitz, and S.K. Shenoy. Beta-arrestinsand cell signaling. Annu Rev Physiol, 69:483–510, 2007.

[44] M. Dews, M. Prisco, F. Peruzzi, G. Romano, A. Morrione, andR. Baserga. Domains of the Insulin-Like Growth Factor I ReceptorRequired for the Activation of Extracellular Signal-Regulated Kinases1. Endocrinology, 141(4):1289–1300, 2000.

[45] PP Di Fiore, JH Pierce, MH Kraus, O. Segatto, CR King, andSA Aaronson. erbB-2 is a potent oncogene when overexpressed inNIH/3T3 cells. Science, 237(4811):178, 1987.

[46] GM Di Guglielmo, PC Baass, WJ Ou, BI Posner, and JJ Bergeron.Compartmentalization of SHC, GRB2 and mSOS, and hyperphospho-rylation of Raf-1 by EGF but not insulin in liver parenchyma. EMBOJ, 13(18):4269–77, 1994.

[47] R.J. DiLeone, L.B. Russell, and D.M. Kingsley. An Extensive 3’Reg-ulatory Region Controls Expression of Bmp5 in Specific AnatomicalStructures of the Mouse Embryo. Genetics, 148(1):401–408, 1998.

[48] K. Dittmann, C. Mayer, and H.P. Rodemann. Inhibition of radiation-induced EGFR nuclear import by C225 (Cetuximab) suppresses DNA-PK activity. Radiother Oncol, 76(2):157–61, 2005.

[49] H. Dudek, SR Datta, TF Franke, MJ Birbaum, R. Yao, GM Cooper,RA Segal, DR Kaplan, and ME Greenberg. Akt signaling: linking

87

membrane events to life and death decisions. Science, 275:628–630,1997.

[50] B. Dufourny, J. Alblas, H.A.A.M. van Teeffelen, F.M.A. van Schaik,B. van der Burg, P.H. Steenbergh, and J.S. Sussenbach. Mitogenic Sig-naling of Insulin-like Growth Factor I in MCF-7 Human Breast Can-cer Cells Requires Phosphatidylinositol 3-Kinase and Is Independentof Mitogen-activated Protein Kinase. Journal of Biological Chemistry,272(49):31163–31171, 1997.

[51] Z.M. Eletr, D.T. Huang, D.M. Duda, B.A. Schulman, and B. Kuhlman.E2 conjugating enzymes must disengage from their E1 enzymes beforeE3-dependent ubiquitin and ubiquitin-like transfer. Nature Struct. Mol.Biol, 12:933–934, 2005.

[52] X. Fang, E.K. Stachowiak, S.M. Dunham-Ems, I. Klejbor, and M.K.Stachowiak. Control of CREB-binding Protein Signaling by NuclearFibroblast Growth Factor Receptor-1: A novel mechanism of gene reg-ulation. Journal of Biological Chemistry, 280(31):28451, 2005.

[53] J. Faridi, J. Fawcett, L. Wang, and R.A. Roth. Akt promotes increasedmammalian cell size by stimulating protein synthesis and inhibitingprotein degradation. American Journal of Physiology- EndocrinologyAnd Metabolism, 285(5):964–972, 2003.

[54] S. Favelyukis, J.H. Till, S.R. Hubbard, and W.T. Miller. Structure andautoregulation of the insulin-like growth factor 1 receptor kinase. NatStruct Biol, 8(12):1058–63, 2001.

[55] D. Finley, B. Bartel, and A. Varshavsky. The tails of ubiquitin precur-sors are ribosomal proteins whose fusion to ubiquitin facilitates ribo-some biogenesis. Nature, 338(6214):394–401, 1989.

[56] M. Fiorentino, WF Grigioni, P. Baccarini, A. D’Errico, MS De Mitri,E. Pisi, and AM Mancini. Different in situ expression of insulin-likegrowth factor type II in hepatocellular carcinoma. An in situ hybridiza-tion and immunohistochemical study. Diagn Mol Pathol, 3(1):59–65,1994.

[57] P.S. Freemont et al. The RING Finger. Annals of the New YorkAcademy of Sciences, 684:174–192, 1993.

[58] S. Freier, O. Weiss, M. Eran, A. Flyvbjerg, R. Dahan, I. Nephesh,T. Safra, E. Shiloni, and I. Raz. Expression of the insulin-like growth

88

factors and their receptors in adenocarcinoma of the colon. Gut,44(5):704–708, 1999.

[59] A.R. French, D.K. Tadaki, S.K. Niyogi, and D.A. Lauffenburger. In-tracellular Trafficking of Epidermal Growth Factor Family Ligands IsDirectly Influenced by the pH Sensitivity of the Receptor/Ligand In-teraction. Journal of Biological Chemistry, 270(9):4334, 1995.

[60] RW Furlanetto. Insulin-like growth factor-I induces cyclin-D1 expres-sion in MG63 human osteosarcoma cells in vitro. Molecular Endocrinol-ogy, 8(4):510–517, 1994.

[61] CE Futter, LM Collinson, JM Backer, and CR Hopkins. Human VPS34is required for internal vesicle formation within multivesicular endo-somes. The Journal of Cell Biology, 155(7):1251–1264, 2001.

[62] F. Galbiati, B. Razani, and M.P. Lisanti. Emerging Themes in LipidRafts and Caveolae. Cell, 106(4):403–411, 2001.

[63] S. Gammeltoft. Expression of two types of receptor for insulin-like growth factors in human malignant glioma. Cancer Research,48(5):1233–1237, 1988.

[64] T. Gansler, R. Furlanetto, TS Gramling, KA Robinson, N. Blocker,MG Buse, DA Sens, and AJ Garvin. Antibody to type I insulinlikegrowth factor receptor inhibits growth of Wilms’ tumor in culture andin athymic mice. American Journal of Pathology, 135(6):961–966, 1989.

[65] F.G. Giancotti et al. Integrin Signaling. Science, 285(5430):1028–1033,1999.

[66] F.G. Giancotti and G. Tarone. Positional control of cell fate throughjoint integrin/receptor protein kinase signaling. Annual Review of Celland Developmental Biology, 19(1):173–206, 2003.

[67] G. Gill. SUMO and ubiquitin in the nucleus: different functions, similarmechanisms? Genes & Development, 18(17):2046–2059, 2004.

[68] L. Girnita, A. Girnita, and O. Larsson. Mdm2-dependent ubiquiti-nation and degradation of the insulin-like growth factor 1 receptor.Proceedings of the National Academy of Sciences, 100(14):8247–8252,2003.

[69] T. Gloudemans. Insulin-like growth factor gene expression in humansmooth muscle tumors. Cancer Research, 50(20):6689–6695, 1990.

89

[70] L. Gong, S. Millas, G.G. Maul, and E.T.H. Yeh. Differential Regulationof Sentrinized Proteins by a Novel Sentrin-specific Protease. Journalof Biological Chemistry, 275(5):3355–3359, 2000.

[71] L. Gong and E.T.H. Yeh. Characterization of a Family of Nucleo-lar SUMO-specific Proteases with Preference for SUMO-2 or SUMO-3.Journal of Biological Chemistry, 281(23):15869, 2006.

[72] M. Groll, W. Heinemeyer, S. Jager, T. Ullrich, M. Bochtler, D.H. Wolf,and R. Huber. The catalytic sites of 20S proteasomes and their role insubunit maturation: A mutational and crystallographic study, 1999.

[73] M. Groll, Y. Koguchi, R. Huber, and J. Kohno. Crystal Structure ofthe 20 S Proteasome: TMC-95A Complex: A Non-covalent ProteasomeInhibitor. Journal of Molecular Biology, 311(3):543–548, 2001.

[74] M. Gronborg, BS Wulff, JS Rasmussen, T. Kjeldsen, and S. Gam-meltoft. Structure-function relationship of the insulin-like growthfactor-I receptor tyrosine kinase. Journal of Biological Chemistry,268(31):23435–23440, 1993.

[75] LC Guidice. IGFs in the female reproductive system, 2000.

[76] R.A. Gupta and R.N. DuBois. Colorectal cancer prevention and treat-ment by inhibition of cyclooxygenase-2. Cancer Metastasis Rev, 23:63–75, 2001.

[77] A. Guterman and M.H. Glickman. Complementary Roles for Rpn11and Ubp6 in Deubiquitination and Proteolysis by the Proteasome.Journal of Biological Chemistry, 279(3):1729, 2004.

[78] AL Haas and PM Bright. The dynamics of ubiquitin pools withincultured human lung fibroblasts. Journal of Biological Chemistry,262(1):345–351, 1987.

[79] A. Hakam, T.J. Yeatman, L. Lu, L. Mora, G. Marcet, S.V. Nicosia,R.C. Karl, and D. Coppola. Expression of insulin-like growth factor-1receptor in human colorectal cancer. Human Pathology, 30(10):1128–1133, 1999.

[80] N. Hanada, H.W. Lo, C.P. Day, Y. Pan, Y. Nakajima, and M.C. Hung.Co-regulation of B-Myb expression by E2F1 and EGF receptor. MolCarcinog, 45(1):10–7, 2006.

90

[81] D. Hanahan and R.A. Weinberg. The Hallmarks of Cancer. Cell,100(1):57–70, 2000.

[82] J. Hang and M. Dasso. Association of the Human SUMO-1 Pro-tease SENP2 with the Nuclear Pore. Journal of Biological Chemistry,277(22):19961–19966, 2002.

[83] S.E. Hankinson, W.C. Willett, G.A. Colditz, D.J. Hunter, D.S.Michaud, B. Deroo, B. Rosner, F.E. Speizer, and M. Pollak. Circu-lating concentrations of insulin-like growth factor-I and risk of breastcancer. Lancet, 351(9113):1393–1396, 1998.

[84] J. Hanna, D.S. Leggett, and D. Finley. Ubiquitin Depletion as a KeyMediator of Toxicity by Translational Inhibitors. Molecular and Cellu-lar Biology, 23(24):9251–9261, 2003.

[85] S.L. Harris and A.J. Levine. The p53 pathway: positive and negativefeedback loops. Oncogene, 24:2899–2908, 2005.

[86] S. Hatakeyama, M. Yada, M. Matsumoto, N. Ishida, and K.I.Nakayama. U Box Proteins as a New Family of Ubiquitin-ProteinLigases. Journal of Biological Chemistry, 276(35):33111–33120, 2001.

[87] C. Hawkes and S. Kar. The insulin-like growth factor-II/mannose-6-phosphate receptor: structure, distribution and function in the centralnervous system. Brain Res Rev, 44:117–140, 2004.

[88] A. Hershko and A. Ciechanover. The ubiquitin-proteasome pathway.Annu. Rev. Biochem, pages 425–479, 1998.

[89] A. Hershko, A. Ciechanover, and A. Varshavsky. Basic Medical Re-search Award. The ubiquitin system. Nat Med, 6(10):1073–81, 2000.

[90] Avram Hershko and Aaron Ciechanover. The ubiquitin system. AnnualReview of Biochemistry, 67(1):425–479, 1998.

[91] GZ Hertz and GD Stormo. Identifying DNA and protein patterns withstatistically significant alignments of multiple sequences. Bioinformat-ics, 15:563–577, 1999.

[92] L. Hicke. A New Ticket for Entry into Budding Vesicles-Ubiquitin.Cell, 106(5):527–530, 2001.

[93] M. Hochstrasser. Evolution and function of ubiquitin-like protein-conjugation systems. Nature Cell Biology, 2:E153–E157, 2000.

91

[94] C. Hoege, B. Pfander, G.L. Moldovan, G. Pyrowolakis, and S. Jentsch.RAD6-dependent DNA repair is linked to modification of PCNA byubiquitin and SUMO. Nature, 419:135–141, 2002.

[95] D. Hoeller, C.M. Hecker, S. Wagner, V. Rogov, V. Dotsch, and I. Di-kic. E3-Independent Monoubiquitination of Ubiquitin-Binding Pro-teins. Molecular Cell, 26(6):891–898, 2007.

[96] R.M. Hofmann and C.M. Pickart. Noncanonical MMS2-encodedubiquitin-conjugating enzyme functions in assembly of novel polyu-biquitin chains for DNA repair. Cell, 96(5):645–653, 1999.

[97] R.M. Hofmann and C.M. Pickart. In Vitro Assembly and Recogni-tion of Lys-63 Polyubiquitin Chains. Journal of Biological Chemistry,276(30):27936–27943, 2001.

[98] JL Hood, BB Logan, AP Sinai, WH Brooks, and TL Roszman. Asso-ciation of the calpain/calpastatin network with subcellular organelles.Biochem Biophys Res Commun, 310(4):1200–12, 2003.

[99] R. Hough, G. Pratt, and M. Rechsteiner. Ubiquitin-lysozyme conju-gates. Identification and characterization of an ATP-dependent pro-tease from rabbit reticulocyte lysates. Journal of Biological Chemistry,261(5):2400–2408, 1986.

[100] R.O. Hynes. Integrins Bidirectional, Allosteric Signaling Machines.Cell, 110(6):673–687, 2002.

[101] Y. Itahana, E.T.H. Yeh, and Y. Zhang. Nucleocytoplasmic ShuttlingModulates Activity and Ubiquitination-Dependent Turnover of SUMO-Specific Protease 2. Molecular and Cellular Biology, 26(12):4675–4689,2006.

[102] D.A. Jans and G. Hassan. Nuclear targeting by growth factors, cy-tokines, and their receptors: a role in signaling? BioEssays, 20(5):400–411, 1998.

[103] M. Joaquin and RJ Watson. Cell cycle regulation by the B-Mybtranscription factor. Cellular and Molecular Life Sciences (CMLS),60(11):2389–2401, 2003.

[104] J.A. Johnston, E.S. Johnson, P.R.H. Waller, and A. Varshavsky.Methotrexate Inhibits Proteolysis of Dihydrofolate Reductase by theN-end Rule Pathway. Journal of Biological Chemistry, 270(14):8172,1995.

92

[105] T. Kamio, K. Shigematsu, H. Sou, K. Kawai, and H. Tsuchiyama.Immunohistochemical expression of epidermal growth factor receptorsin human adrenocortical carcinoma. Hum Pathol, 21(3):277–82, 1990.

[106] B. Kammandel, K. Chowdhury, A. Stoykova, S. Aparicio, S. Brenner,and P. Gruss. Distinct cis-essential modules direct the time-space pat-tern of the Pax6 gene activity. Dev. Biol, 205(1):79–97, 1999.

[107] H. Kato, TN Faria, B. Stannard, CT Roberts, and D. LeRoith. Roleof tyrosine kinase activity in signal transduction by the insulin-likegrowth factor-I (IGF-I) receptor. Characterization of kinase-deficientIGF-I receptors and the action of an IGF-I-mimetic antibody (alphaIR-3). Journal of Biological Chemistry, 268(4):2655–2661, 1993.

[108] KM Kelley, Y. Oh, SE Gargosky, Z. Gucev, T. Matsumoto, V. Hwa,L. Ng, DM Simpson, and RG Rosenfeld. Insulin-like growth factor-binding proteins (IGFBPs) and their regulatory dynamics. Int JBiochem Cell Biol, 28(6):619–37, 1996.

[109] W.J. Kent. BLAT-The BLAST-Like Alignment Tool, 2002.

[110] J.H. Kim, K.C. Park, S.S. Chung, O. Bang, and C.H. Chung. Deubiq-uitinating Enzymes as Cellular Regulators. Journal of Biochemistry,134(1):9–18, 2003.

[111] BE Kimmel, U. Heberlein, and GM Rubin. The homeo domain proteinrough is expressed in a subset of cells in the developing Drosophila eyewhere it can specify photoreceptor cell subtype. Genes & Development,4(5):712–727, 1990.

[112] H. Kimura. Schwannoma-Derived Growth Factor must be Transportedinto the Nucleus to Exert its Mitogenic Activity. Proceedings of theNational Academy of Sciences, 90(6):2165–2169, 1993.

[113] M. Koegl, T. Hoppe, S. Schlenker, HD Ulrich, TU Mayer, andS. Jentsch. A novel ubiquitination factor. E4, is involved in mul-tiubiquitin chain assembly+ Cell, 96:635–644, 1999.

[114] A. Komuro, M. Nagai, N.E. Navin, and M. Sudol. WW Domain-containing Protein YAP Associates with ErbB-4 and Acts as a Co-transcriptional Activator for the Carboxyl-terminal Fragment of ErbB-4 That Translocates to the Nucleus*. Journal of Biological Chemistry,278(35):33334–33341, 2003.

93

[115] G.J.P.L. Kops and B.M.T. Burgering. Forkhead transcription factors:new insights into protein kinase B (c-akt) signaling. Journal of Molec-ular Medicine, 77(9):656–665, 1999.

[116] H. Lahm, P. Amstad, J. Wyniger, A. Yilmaz, JR Fischer, M. Schreyer,and JC Givel. Blockade of the insulin-like growth-factor-I receptor in-hibits growth of human colorectal cancer cells: evidence of a functionalIGF-II-mediated autocrine loop. Int J Cancer, 58(3):452–9, 1994.

[117] K.M.V. Lai and T. Pawson. The ShcA phosphotyrosine docking pro-tein sensitizes cardiovascular signaling in the mouse embryo. Genes &Development, 14(9):1132–1145, 2000.

[118] A. Laine and Z. Ronai. Ubiquitin Chains in the Ladder of MAPKSignaling, 2005.

[119] Y.A. Lam, G.N. DeMartino, C.M. Pickart, and R.E. Cohen. Specificityof the Ubiquitin Isopeptidase in the PA700 Regulatory Complex of 26S Proteasomes. Journal of Biological Chemistry, 272(45):28438–28446,1997.

[120] Y.A. Lam, W. Xu, G.N. DeMartino, and R.E. Cohen. Editing of ubiq-uitin conjugates by an isopeptidase in the 26 S proteasome. Nature,385(6618):737–740, 1997.

[121] J. Larkin, H.M. Johnson, and P.S. Subramaniam. Differential NuclearLocalization of the IFNGR-1 and IFNGR-2 Subunits of the IFN- Re-ceptor Complex Following Activation by IFN. Journal of Interferon &Cytokine Research, 20(6):565–576, 2000.

[122] JN Lavoie. l’Allemain G, Brunet A, Muller R, Pouyssegur J: CyclinD1 expression is regulated positively by the p42/p44MAPK and neg-atively by the p38/HOGMAPK pathway. J Biol Chem, 271:20608–20616, 1996.

[123] R.J. Lefkowitz. G Protein-coupled Receptors III. New roles for receptorkinases and β-arrestins in receptor signaling and desensitization, 1998.

[124] R.J. Lefkowitz and S.K. Shenoy. Transduction of Receptor Signals byβ-Arrestins. Science, 308(5721):512–517, 2005.

[125] R.J. Lefkowitz and E.J. Whalen. Beta-arrestins: traffic cops of cellsignaling. Curr Opin Cell Biol, 16(2):162–8, 2004.

94

[126] D. LeRoith, H. Werner, D. Beitner-Johnson, and CT Roberts. Molecu-lar and cellular aspects of the IGF-I receptor. Endocr Rev, 16:143–153,1995.

[127] D. LeRoith, H. Werner, D. Beitner-Johnson, and CT Roberts Jr.Molecular and cellular aspects of the insulin-like growth factor I re-ceptor. Endocr Rev, 16(2):143–63, 1995.

[128] L.A. Lettice, T. Horikoshi, S.J.H. Heaney, M.J. van Baren, H.C. van derLinde, G.J. Breedveld, M. Joosse, N. Akarsu, B.A. Oostra, N. Endo,et al. Disruption of a long-range cis-acting regulator for Shh causespreaxial polydactyly. Proceedings of the National Academy of Sciences,99(11):7548, 2002.

[129] S. Li, A. Ferber, M. Miura, and R. Baserga. Mitogenicity and trans-forming activity of the insulin-like growth factor-I receptor with muta-tions in the tyrosine kinase domain. Journal of Biological Chemistry,269(51):32558–32564, 1994.

[130] F.T. Lin, K.M. Krueger, H.E. Kendall, Y. Daaka, Z.L. Freder-icks, J.A. Pitcher, and R.J. Lefkowitz. Clathrin-mediated Endo-cytosis of the β-Adrenergic Receptor Is Regulated by Phosphoryla-tion/Dephosphorylation of β-Arrestin1. Journal of Biological Chem-istry, 272(49):31051–31057, 1997.

[131] J.C. Lin, K. Duell, and J.B. Konopka. A Microdomain Formed bythe Extracellular Ends of the Transmembrane Domains Promotes Ac-tivation of the G Protein-Coupled α-Factor Receptor. Molecular andCellular Biology, 24(5):2041–2051, 2004.

[132] S.Y. Lin, K. Makino, W. Xia, A. Matin, Y. Wen, K.Y. Kwong, L. Bour-guignon, and M.C. Hung. Nuclear localization of EGF receptor andits potential new role as a transcription factor. Nature Cell Biology,3:802–808, 2001.

[133] J.P. Liu, J. Baker, A.S. Perkins, E.J. Robertson, and A. Efstratiadis.Mice carrying null mutations of the genes encoding insulin-like growthfactor I (Igf-1) and type 1 IGF receptor (Igf1r). Cell, 75(1):59–72, 1993.

[134] T.E. Lloyd, R. Atkinson, M.N. Wu, Y. Zhou, G. Pennetta, and H.J.Bellen. Hrs Regulates Endosome Membrane Invagination and TyrosineKinase Receptor Signaling in Drosophila. Cell, 108(2):261–269, 2002.

95

[135] H.W. Lo, S.C. Hsu, M. Ali-Seyed, M. Gunduz, W. Xia, Y. Wei,G. Bartholomeusz, J.Y. Shih, and M.C. Hung. Nuclear interaction ofEGFR and STAT3 in the activation of the iNOS/NO pathway. CancerCell, 7(6):575–89, 2005.

[136] K.L. Lorick, J.P. Jensen, S. Fang, A.M. Ong, S. Hatakeyama, andA.M. Weissman. RING fingers mediate ubiquitin-conjugating enzyme(E2)-dependent ubiquitination, 1999.

[137] J. Ma, M.N. Pollak, E. Giovannucci, J.M. Chan, Y. Tao, C.H. Hen-nekens, and M.J. Stampfer. Prospective Study of Colorectal CancerRisk in Men and Plasma Levels of Insulin-Like Growth Factor (IGF)-Iand IGF-Binding Protein-3, 1999.

[138] PA Maher. Nuclear Translocation of fibroblast growth factor (FGF) re-ceptors in response to FGF-2. The Journal of Cell Biology, 134(2):529–536, 1996.

[139] JN Maloof and C. Kenyon. The Hox gene li n-39 i s required during.C. elegans, 1998.

[140] CJ Marshall et al. Specificity of receptor tyrosine kinase signaling:transient versus sustained extracellular signal-regulated kinase activa-tion. Cell, 80(2):179–85, 1995.

[141] A.M. Martelli, R. Bortul, G. Tabellini, I. Faenza, A. Cappellini,R. Bareggi, L. Manzoli, and L. Cocco. Molecular characterization ofprotein kinase ca binding to lamin a. Journal of Cellular Biochemistry,86:320–330, 2002.

[142] U. Marti, C. Ruchti, J. Kampf, G.A. Thomas, E.D. Williams, H.J.Peter, H. Gerber, and U. Burgi. Nuclear Localization of EpidermalGrowth Factor and Epidermal Growth Factor Receptors in HumanThyroid Tissues. Thyroid, 11(2):137–145, 2001.

[143] L.D. Mayo and D.B. Donner. A phosphatidylinositol 3-kinase/Aktpathway promotes translocation of Mdm2 from the cytoplasm tothe nucleus. Proceedings of the National Academy of Sciences, page181181198, 2001.

[144] A. McCahill, J. Warwicker, G.B. Bolger, M.D. Houslay, and S.J.Yarwood. The RACK1 Scaffold Protein: A Dynamic Cog in Cell Re-sponse Mechanisms, 2002.

96

[145] RH McCusker, C. Camacho-Hubner, DR Clemmons, ML Bayne, andMA Cascieri. Insulin-like growth factor(IGF) binding to human fibrob-last and glioblastoma cells: The modulating effect of cell released IGFbinding proteins(IGFBPs). Journal of Cellular Physiology, 144(2):244–253, 1990.

[146] DW Meek. Mechanisms of switching on p53: a role for covalent modi-fication? Oncogene, 18(53):7666–75, 1999.

[147] F. Melchior. SUMO–nonclassical ubiquitin. Annual Review of Cell andDevelopmental Biology, 16(1):591–626, 2000.

[148] B.S. Miller and D. Yee. Type I Insulin-like Growth Factor Receptor asa Therapeutic Target in Cancer. Cancer Research, 65(22):10123–10127,2005.

[149] Y. Min, Y. Adachi, H. Yamamoto, H. Ito, F. Itoh, C.T. Lee, S. Nadaf,D.P. Carbone, and K. Imai. Genetic Blockade of the Insulin-likeGrowth Factor-I Receptor A Promising Strategy for Human Pancre-atic Cancer 1, 2003.

[150] CP Minniti, M. Tsokos, WA Newton Jr, and LJ Helman. Specificexpression of insulin-like growth factor-II in rhabdomyosarcoma tumorcells. Am J Clin Pathol, 101(2):198–203, 1994.

[151] C. Minshall. IL-4 and insulin-like growth factor-I inhibit the decline inBcl-2 and promote the survival of IL-3-deprived myeloid progenitors.The Journal of Immunology, 159(3):1225–1232, 1997.

[152] C.K. Miranti and J.S. Brugge. Integrins: Signalling and Disease. Na-ture Cell Biology, 4:E83–E90, 2002.

[153] C.K. Miranti and J.S. Brugge. Sensing the environment: a historicalperspective on integrin signal transduction. Nat Cell Biol, 4(4):E83–90,2002.

[154] T. Miyashita and JC Reed. Tumor suppressor p53 is a direct transcrip-tional activator of the human bax gene. Cell, 80(2):293–9, 1995.

[155] S. Monno, M.V. Newman, M. Cook, and W.L. Lowe. Insulin-LikeGrowth Factor I Activates c-Jun N-Terminal Kinase in MCF-7 BreastCancer Cells 1. Endocrinology, 141(2):544–550, 2000.

97

[156] T.F. Moraes, R.A. Edwards, S. McKenna, L. Pastushok, W. Xiao,J.N.M. Glover, and M.J. Ellison. Crystal structure of the human ubiq-uitin conjugating enzyme complex, hMms2- hUbc13. Nature StructuralBiology, 8:669–673, 2001.

[157] A. Morrione. Grb 10 adapter protein as regulator of insulin-like growthfactor receptor signaling. Journal of Cellular Physiology, 197(3):307–311, 2003.

[158] A. Morrione, P. Plant, B. Valentinis, O. Staub, S. Kumar, D. Rotin,and R. Baserga. mGrb10 Interacts with Nedd4. Journal of BiologicalChemistry, 274(34):24094–24099, 1999.

[159] M. Mumby, H. Peng, J. Moffett, J. Myers, X. Fang, E.K. Stachowiak,P. Maher, E. Kratz, J. Hines, S.J. Fluharty, et al. Novel Nuclear Signal-ing Pathway Mediates Activation of Fibroblast Growth Factor-2 Geneby Type 1 and Type 2 Angiotensin II Receptors. Molecular Biology ofthe Cell, 12(2):449–462, 2001.

[160] MS Murakami and OM Rosen. The role of insulin receptor autophos-phorylation in signal transduction. Journal of Biological Chemistry,266(33):22653–22660, 1991.

[161] J. Nakae, V. Barr, and D. Accili. Differential regulation of gene expres-sion by insulin and IGF-1 receptors correlates with phosphorylation ofa single amino acid residue in the forkhead transcription factor FKHR.The EMBO Journal, 19:989–996, 2000.

[162] S. Neuenschwander, A. Schwartz, T.L. Wood, C.T. Roberts, Jr, L. Hen-ninghausen, and D. LeRoith. Involution of the Lactating MammaryGland Is Inhibited by the IGF System in a Transgenic Mouse Model.Journal of Clinical Investigation, 97(10):2225–2232, 1996.

[163] N. Neznanov, A. Umezawa, and R.G. Oshima. A Regulatory Ele-ment within a Coding Exon Modulates Keratin 18 Gene Expressionin Transgenic Mice. Journal of Biological Chemistry, 272(44):27549–27557, 1997.

[164] C.Y. Ni, H. Yuan, and G. Carpenter. Role of the ErbB-4 CarboxylTerminus in γ-Secretase Cleavage. Journal of Biological Chemistry,278(7):4561–4565, 2003.

[165] B.J. Nichols and J. Lippincott-Schwartz. Endocytosis without clathrincoats. Trends Cell Biol, 11(10):406–412, 2001.

98

[166] H. Nishikawa, S. Ooka, K. Sato, K. Arima, J. Okamoto, R.E. Kle-vit, M. Fukuda, and T. Ohta. Mass Spectrometric and Muta-tional Analyses Reveal Lys-6-linked Polyubiquitin Chains Catalyzedby BRCA1-BARD1 Ubiquitin Ligase. Journal of Biological Chemistry,279(6):3916–3924, 2004.

[167] A.J. Oates, L.M. Schumaker, S.B. Jenkins, A.A. Pearce, S.A. DaCosta,B. Arun, and M.J.C. Ellis. The mannose 6-phosphate/insulin-likegrowth factor 2 receptor (M6P/IGF2R), a putative breast tumor sup-pressor gene. Breast Cancer Research and Treatment, 47(3):269–281,1998.

[168] R. O’Connor, A. Kauffmann-Zeh, Y. Liu, S. Lehar, GI Evan,R. Baserga, and WA Blattler. Identification of domains of the insulin-like growth factor I receptor that are required for protection from apop-tosis. Molecular and Cellular Biology, 17(1):427–435, 1997.

[169] M. Offterdinger, C. Schofer, K. Weipoltshammer, and T.W. Grunt. c-erbB-3: A Nuclear Protein in Mammary Epithelial Cells. The Journalof Cell Biology, 157(6):929–939, 2002.

[170] J. Ooi, V. Yajnik, D. Immanuel, M. Gordon, JJ Moskow, AM Buch-berg, and B. Margolis. The cloning of Grb10 reveals a new family ofSH2 domain proteins. Oncogene, 10(8):1621–30, 1995.

[171] Y. Pan. Advances in the Discovery of cis-Regulatory Elements. CurrentBioinformatics, 1:321–336, 2006.

[172] M. Parrizas and D. LeRoith. Insulin-Like Growth Factor-1 Inhibition ofApoptosis Is Associated with Increased Expression of the bcl-xL GeneProduct. Endocrinology, 138(3):1355–1358, 1997.

[173] R.G. Parton and A.A. Richards. Lipid Rafts and Caveolae as Por-tals for Endocytosis: New Insights and Common Mechanisms. Traffic,4(11):724–738, 2003.

[174] A. Pause, G.J. Belsham, A.C. Gingras, O. Donze, T.A. Lin, J.C.Lawrence, and N. Sonenberg. Insulin-dependent stimulation of proteinsynthesis by phosphorylation of a regulator of 5’-cap function. Nature,371(6500):762–767, 1994.

[175] J. Peng, D. Schwartz, J.E. Elias, C.C. Thoreen, D. Cheng, G. Mar-sischky, J. Roelofs, D. Finley, and S.P. Gygi. A proteomics ap-proach to understanding protein ubiquitination. Nature Biotechnology,21(8):921–926, 2003.

99

[176] F. Peruzzi, M. Prisco, M. Dews, P. Salomoni, E. Grassilli, G. Romano,B. Calabretta, and R. Baserga. Multiple Signaling Pathways of theInsulin-Like Growth Factor 1 Receptor in Protection from Apoptosis.Molecular and Cellular Biology, 19(10):7203–7215, 1999.

[177] R.T. Peterson and S.L. Schreiber. Translation control: connectingmitogens and the ribosome. Curr Biol, 8(7):R248–50, 1998.

[178] MC Pike, RK Ross, PA Jones, and BE Henderson. Increased celldivision as a cause of human cancer. Cancer Res., 50:7415–7421, 1990.

[179] G. Pillai, N. Cook, H. Turley, RD Leek, C. Blasquez, F. Pezzella,AL Harris, and KC Gatter. The expression and cellular localiza-tion of phosphorylated VEGFR 2 in lymphoma and non-neoplasticlymphadenopathy: an immunohistochemical study. Histopathology,46(2):209–216, 2005.

[180] J. Piotrowski, R. Beal, L. Hoffman, K.D. Wilkinson, R.E. Cohen, andC.M. Pickart. Inhibition of the 26 S Proteasome by PolyubiquitinChains Synthesized to Have Defined Lengths. Journal of BiologicalChemistry, 272(38):23712–23721, 1997.

[181] S. Polo, S. Sigismund, M. Faretta, M. Guidi, M.R. Capua, G. Bossi,H. Chen, P. De Camilli, and P.P. Di Fiore. A single motif responsiblefor ubiquitin recognition and monoubiquitination in endocytic proteins.Nature, 416(6879):451–455, 2002.

[182] B. Poole, S. Ohkuma, and M. Warburton. Some aspects of the in-tracellular breakdown of exogenous and endogenous proteins. ProteinTurnover and Lysosome Function. Academic Press, New York, NYUSA, pages 43–58, 1978.

[183] EM Rakowicz-Szulczynska. Nerve growth factor receptors in chromatinof melanoma cells, proliferating melanocytes, and colorectal carcinomacells in vitro. Cancer Research, 48(24):7200–7206, 1988.

[184] A.E. Reeve, M.R. Eccles, R.J. Wilkins, G.I. Bell, and L.J. Millow. Ex-pression of insulin-like growth factor-II transcripts in Wilms’ tumour.Nature, 317(6034):258–260, 1985.

[185] M. Resnicoff, D. Ambrose, D. Coppola, and R. Rubin. Insulin-likegrowth factor-1 and its receptor mediate the autocrine proliferation ofhuman ovarian carcinoma cell lines. Lab Invest, 69(6):756–60, 1993.

100

[186] A. Roulston, C. Reinhard, P. Amiri, and L.T. Williams. Early Activa-tion of c-Jun N-terminal Kinase and p38 Kinase Regulate Cell Survivalin Response to Tumor Necrosis Factor α. Journal of Biological Chem-istry, 273(17):10232–10239, 1998.

[187] T. Sandmann, L.J. Jensen, J.S. Jakobsen, M.M. Karzynski, M.P.Eichenlaub, P. Bork, and EE Furlong. A temporal map of transcrip-tion factor activity: mef2 directly regulates target genes at all stagesof muscle development. Dev. Cell, 10:797–807, 2006.

[188] F. Sandrelli, S. Campesan, M.G. Rossetto, C. Benna, E. Zieger,A. Megighian, M. Couchman, C.P. Kyriacou, and R. Costa. Molec-ular Dissection of the 5’Region of no-on-transientA of Drosophilamelanogaster Reveals cis-Regulation by Adjacent dGpi1 Sequences.Genetics, 157(2):765–775, 2001.

[189] AJ Saurin, KLB Borden, MN Boddy, and PS Freemont. Does thishave a familiar RING? Trends in Biochemical Sciences, 21(6):208–214,1996.

[190] E. Schausberger, R. Eferl, W. Parzefall, M. Chabikovsky, P. Breit, E.F.Wagner, R. Schulte-Hermann, and B. Grasl-Kraupp. Induction of DNAsynthesis in primary mouse hepatocytes is associated with nuclear pro-transforming growth factor {alpha} and erbb-1 and is independent ofc-jun. Carcinogenesis, 24(5):835–841, 2003.

[191] D.C. Schwartz and M. Hochstrasser. A superfamily of proteintags: ubiquitin, SUMO and related modifiers. Trends Biochem. Sci,28(6):321–328, 2003.

[192] JS Seeler and A. Dejean. Nuclear and unclear functions of SUMO. NatRev Mol Cell Biol, 4(9):690–9, 2003.

[193] C. Sell, G. Dumenil, C. Deveaud, M. Miura, D. Coppola, T. DeAngelis,R. Rubin, A. Efstratiadis, and R. Baserga. Effect of a null mutation ofthe type 1 IGF receptor gene on growth and transformation of mouseembryo fibroblasts. Mol Cell Biol, 14:3604–3612, 1994.

[194] C. Sell, M. Rubini, R. Rubin, J. Liu, A. Efstratiadis, and R. Baserga.Simian Virus 40 Large Tumor Antigen is Unable to Transform MouseEmbryonic Fibroblasts Lacking Type 1 Insulin-Like Growth Factor Re-ceptor. Proceedings of the National Academy of Sciences, 90(23):11217–11221, 1993.

101

[195] DN Shapiro, BG Jones, LH Shapiro, P. Dias, and PJ Houghton.Antisense-mediated reduction in insulin-like growth factor-I receptorexpression suppresses the malignant phenotype of a human alveolarrhabdomyosarcoma. J Clin Invest, 94(3):1235–42, 1994.

[196] DN Shapiro, BG Jones, LH Shapiro, P. Dias, and PJ Houghton.Antisense-mediated reduction in insulin-like growth factor-I receptorexpression suppresses the malignant phenotype of a human alveolarrhabdomyosarcoma. J Clin Invest, 94(3):1235–42, 1994.

[197] A. Shay-Salit, M. Shushy, E. Wolfovitz, H. Yahav, F. Breviario, E. De-jana, and N. Resnick. VEGF receptor 2 and the adherens junction as amechanical transducer in vascular endothelial cells. Proceedings of theNational Academy of Sciences, 99(14):9462, 2002.

[198] S.K. Shenoy. Seven-Transmembrane Receptors and Ubiquitination.Circulation Research, 100(8):1142, 2007.

[199] S.K. Shenoy and R.J. Lefkowitz. Multifaceted roles of-arrestins in theregulation of seven-membrane-spanning receptor trafficking and signal-ing. Biochem. J, 375:503–515, 2003.

[200] S.K. Shenoy, P.H. McDonald, T.A. Kohout, and R.J. Lefkowitz.Regulation of Receptor Fate by Ubiquitination of Activatedbeta 2-Adrenergic Receptor and beta-Arrestin. Science’s STKE,294(5545):1307, 2001.

[201] J. Simon, M. Peifer, W. Bender, and M. O’Connor. Regulatory el-ements of the bithorax complex that control expression along theanterior-posterior axis. EMBO J, 9(12):3945–56, 1990.

[202] S.T. Smale and J.T. Kadonaga. The rna polymerase ii core promoter.Annual Review of Biochemistry, 72(1):449–479, 2003.

[203] A. Sorkin, A. Eriksson, C.H. Heldin, B. Westermark, and L. Claesson-Welsh. Pool of Ligand-bound Platelet-derived Growth Factor a-Receptors Remain Activated and Tyrosine-Phosphorylated After In-ternalization. J. Cell Physiol, 156:373–82, 1993.

[204] AD Sorkin, LV Teslenko, and NN Nikolsky. The endocytosis of epider-mal growth factor in A431 cells: a pH of microenvironment and thedynamics of receptor complex dissociation. Exp Cell Res, 175(1):192–205, 1988.

102

[205] J. Spence, S. Sadis, AL Haas, and D. Finley. A ubiquitin mutant withspecific defects in DNA repair and multiubiquitination. Molecular andCellular Biology, 15(3):1265–1273, 1995.

[206] C.G. Spilianakis, M.D. Lalioti, T. Town, G.R. Lee, and R.A. Flavell. In-terchromosomal associations between alternatively expressed loci. Na-ture, 435:637–645, 2005.

[207] EK Stachowiak, PA Maher, J. Tucholski, E. Mordechai, A. Joy, J. Mof-fett, S. Coons, and MK Stachowiak. Nuclear accumulation of fibroblastgrowth factor receptors in human glial cells-association with cell pro-liferation. Oncogene, 14(18):2201–2211, 1997.

[208] CE Stewart and P. Rotwein. Growth, differentiation, and survival:multiple physiological functions for insulin-like growth factors, 1996.

[209] G. Suzuki, Y. Yanagawa, S.F. Kwok, M. Matsui, and X.W. Deng. Ara-bidopsis COP10 is a ubiquitin-conjugating enzyme variant that actstogether with COP1 and the COP9 signalosome in repressing photo-morphogenesis, 2002.

[210] C. Tan, S. Cruet-Hennequart, A. Troussard, L. Fazli, P. Costello,K. Sutton, J. Wheeler, M. Gleave, J. Sanghera, and S. Dedhar. Reg-ulation of tumor angiogenesis by integrin-linked kinase (ILK). CancerCell, 5(1):79–90, 2004.

[211] PB Tan and SK Kim. Signaling specificity: the RTK/RAS/MAP kinasepathway in metazoans. Trends Genet, 15(4):145–9, 1999.

[212] T. Tenno, K. Fujiwara, H. Tochio, K. Iwai, E.H. Morita, H. Hayashi,S. Murata, H. Hiroaki, M. Sato, K. Tanaka, et al. Structural basis fordistinct roles of Lys63-and Lys48-linked polyubiquitin chains. Genesto Cells, 9(10):865, 2004.

[213] J.S. Thrower, L. Hoffman, M. Rechsteiner, and C.M. Pickart. Recog-nition of the polyubiquitin proteolytic signal. The EMBO Journal,19:94–102, 2000.

[214] JV Tricoli. Enhanced levels of insulin-like growth factor messengerRNA in human colon carcinomas and liposarcomas. Cancer Research,46(12):6169–6173, 1986.

103

[215] A. Ullrich, A. Gray, AW Tam, T. Yang-Feng, M. Tsubokawa,C. Collins, W. Henzel, T. Le Bon, S. Kathuria, E. Chen, et al. Insulin-like growth factor I receptor primary structure: comparison with in-sulin receptor suggests structural determinants that define functionalspecificity. EMBO J, 5(10):2503–2512, 1986.

[216] R.M. Vabulas and F.U. Hartl. Protein Synthesis upon Acute NutrientRestriction Relies on Proteasome Function, 2005.

[217] A.P. VanDemark, R.M. Hofmann, C. Tsui, C.M. Pickart, and C. Wol-berger. Molecular Insights into Polyubiquitin Chain Assembly Crys-tal Structure of the Mms2/Ubc13 Heterodimer. Cell, 105(6):711–720,2001.

[218] B. Vanhaesebroeck and D.R. Alessi. The PI3K-PDK1 connection: morethan just a road to PKB. Biochem J, 346(Pt 3):561–576, 2000.

[219] A. Vecchione, A. Marchese, P. Henry, D. Rotin, and A. Morrione. TheGrb10/Nedd4 Complex Regulates Ligand-Induced Ubiquitination andStability of the Insulin-Like Growth Factor I Receptor. Molecular andCellular Biology, 23(9):3363–3372, 2003.

[220] K.J. Verhey, D. Meyer, R. Deehan, J. Blenis, B.J. Schnapp, T.A.Rapoport, and B. Margolis. Cargo of Kinesin Identified as JIP Scaf-folding Proteins and Associated Signaling Molecules. The Journal ofCell Biology, 152(5):959–970, 2001.

[221] C. Wang, L. Deng, M. Hong, GR Akkaraju, J. Inoue, and ZJ Chen.TAK1 is a ubiquitin-dependent kinase of MKK and IKK. Nature,412(6844):283–285, 2001.

[222] F. Wang, V.M. Weaver, O.W. Petersen, C.A. Larabell, S. Dedhar,P. Briand, R. Lupu, and M.J. Bissell. Reciprocal interactions betweenβ1-integrin and epidermal growth factor receptor in three-dimensionalbasement membrane breast cultures: A different perspective in epithe-lial biology, 1998.

[223] S.C. Wang, H.C. Lien, W. Xia, I.F. Chen, H.W. Lo, Z. Wang, M. Ali-Seyed, D.F. Lee, G. Bartholomeusz, F. Ou-Yang, et al. Binding atand transactivation of the COX-2 promoter by nuclear tyrosine kinasereceptor ErbB-2. Cancer Cell, 6(3):251–61, 2004.

104

[224] W.W. Wasserman and J.W. Fickett. Identification of regulatory regionswhich confer muscle-specific gene expression. J. Mol. Biol, 278(1):167–181, 1998.

[225] MG Waugh, S. Minogue, JS Anderson, M. dos Santos, and JJ Hsuan.Signalling and non-caveolar rafts. Biochem. Soc. Trans, 29:509–511,2001.

[226] J.D. Weber, J.R. Jeffers, J.E. Rehg, D.H. Randle, G. Lozano, M.F.Roussel, C.J. Sherr, and G.P. Zambetti. p53-independent functions ofthe p19ARF tumor suppressor. Genes & Development, 14(18):2358–2365, 2000.

[227] J.D. Weber, D.M. Raben, P.J. Phillips, and J.J. Baldassare. Sustainedactivation of extracellular-signal-regulated kinase 1 (ERK1) is requiredfor the continued expression of cyclin D1 in G1 phase. Biochem J,326(Pt 1):61–8, 1997.

[228] A.M. Weissman. Themes and variations on ubiquitylation. Nat RevMol Cell Biol, 2(3):169–178, 2001.

[229] O. Wiborg, MS Pedersen, A. Wind, LE Berglund, KA Marcker, andJ. Vuust. The human ubiquitin multigene family: some genes con-tain multiple directly repeated ubiquitin coding sequences. EMBO J,4(3):755–759, 1985.

[230] KD Wilkinson. Regulation of ubiquitin-dependent processes by deu-biquitinating enzymes. The FASEB Journal, 11(14):1245–1256, 1997.

[231] KD Wilkinson and M. Hochstrasser. The deubiquitinating enzymes.Ubiquitin and the biology of the cell, pages 99–126, 1998.

[232] C.C. Williams, J.G. Allison, G.A. Vidal, M.E. Burow, B.S. Beckman,L. Marrero, and F.E. Jones. The ERBB4/HER4 receptor tyrosine ki-nase regulates gene expression by functioning as a STAT5A nuclearchaperone. J Cell Biol, 167:469–478, 2004.

[233] T. Woelk, B. Oldrini, E. Maspero, S. Confalonieri, E. Cavallaro, P.P.Di Fiore, and S. Polo. Molecular mechanisms of coupled monoubiqui-tination. Nat Cell Biol, 8:1246–1254, 2006.

[234] TL Wood, M. Berelowitz, MC Gelato, CT Roberts Jr, D. LeRoith,WJ Millard, and JF McKelvy. Hormonal regulation of rat hypotha-lamic neuropeptide mRNAs: effect of hypophysectomy and hormone

105

replacement on growth-hormone-releasing factor, somatostatin and theinsulin-like growth factors. Neuroendocrinology, 53(3):298–305, 1991.

[235] G.A. Wray, M.W. Hahn, E. Abouheif, J.P. Balhoff, M. Pizer, M.V.Rockman, and L.A. Romano. The Evolution of Transcriptional Regula-tion in Eukaryotes. Molecular Biology and Evolution, 20(9):1377–1419,2003.

[236] C. Xian, Y. Wang, B. Zhang, L. Tang, J. Pan, Y. Luo, P. Jiang, andD. Li. Alternative splicing and expression of the insulin-like growthfactor (IGF-1) gene in osteoblasts under mechanical stretch. ChineseScience Bulletin, 51(22):2731–2736, 2006.

[237] Y. Xie and MC Hung. Nuclear localization of p185neu tyrosine ki-nase and its association with transcriptional transactivation. BiochemBiophys Res Commun, 203(3):1589–98, 1994.

[238] W. Xu, LZ Liu, M. Loizidou, M. Ahmed, and IG Charles. The role ofnitric oxide in cancer. Cell Res, 12(5-6):311–20, 2002.

[239] ET Yeh, L. Gong, and T. Kamitani. Ubiquitin-like proteins: new winesin new bottles. Gene, 248(1-2):1–14, 2000.

[240] H. Yu, M.R. Spitz, J. Mistry, J. Gu, W.K. Hong, and X. Wu. PlasmaLevels of Insulin-Like Growth Factor-I and Lung Cancer Risk: a Case-Control Analysis. Journal of the National Cancer Institute, 91(2):151–156, 1999.

[241] CH Yuh, CT Brown, CB Livi, L. Rowen, PJC Clarke, and EH David-son. Patchy Interspecific Sequence Similarities Efficiently Identify Posi-tive cis-Regulatory Elements in the Sea Urchin. Developmental Biology,246(1):148–161, 2002.

[242] L.D. Zajchowski and S.M. Robbins. Lipid rafts and little caves Com-partmentalized signalling in membrane microdomains. FEBS Journal,269(3):737–752, 2002.

[243] A. Zapf-Colby and J.M. Olefsky. Nerve Growth Factor Processingand Trafficking Events Following TrkA-Mediated Endocytosis 1. En-docrinology, 139(7):3232–3240, 1998.

[244] H. Zhang, H. Saitoh, and M.J. Matunis. Enzymes of the SUMO Mod-ification Pathway Localize to Filaments of the Nuclear Pore Complex.Molecular and Cellular Biology, 22(18):6498–6508, 2002.

106

[245] L. Zhang, Q. Zhan, S. Zhan, F. Kashanchi, AJ Fornace Jr, P. Seth, andLJ Helman. p53 regulates human insulin-like growth factor II gene ex-pression through active P4 promoter in rhabdomyosarcoma cells. DNACell Biol, 17(2):125–31, 1998.

[246] B.P. Zhou, Y. Liao, W. Xia, Y. Zou, B. Spohn, and M.C. Hung. HER-2/neu induces p53 ubiquitination via Akt-mediated MDM2 phospho-rylation. Nature Cell Biology, 3:973–982, 2001.

[247] JC Zwaagstra, A. Guimond, and MD O’Connor-McCourt. Pre-dominant Intracellular Localization of the Type I TransformingGrowth Factor-beta Receptor and Increased Nuclear Accumulation af-ter Growth Arrest. Experimental Cell Research, 258(1):121–134, 2000.

107