Insights into NEDD8 function and the regulation of its conjugation system

Dissertation

zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften (Dr. rer. nat.)

vorgelegt von

Dana Pagliarini

an der

Mathematisch‐Naturwissenschaftliche Sektion Fachbereich Biologie

Tag der mündlichen Prüfung: 07. Juni 2013 1. Referent: Prof. Dr. Martin Scheffner 2. Referent: Prof. Dr. Thomas U. Mayer

Für meine Eltern

Table of contents

Abbreviations i Abstract ii Zusammenfassung iii 1. Introduction 1 1.1 Ubiquitin‐proteasome system 1 1.1.1 Ubiquitin‐conjugation cascade 2 1.1.2 Modes of ubiquitination 4 1.1.3 Ubiquitin recycling and the proteasome 6 1.2 Ubiquitin‐like proteins (UBLs) 6 1.2.1 SUMO 7 1.2.2 Other UBLs 8 1.3 NEDD8 9 1.3.1 Substrates and functions of NEDD8 10 1.3.1.1 Cullins 10 1.3.1.2 NEDD8 and transcriptional regulation 12 1.3.1.3 Further substrates and functions of NEDD8 13 1.3.2 NEDD8‐conjugation cascade 14 1.3.2.1 APPBP1/UBA3, the NEDD8‐activating enzyme 16 1.3.2.2 NEDD8‐conjugating enzymes 18 1.4 Aim of the studies 21 2. Material and Methods 22 2.1 Material 22 2.1.1 Solutions and media 22 2.1.2 Chemicals and Reagents 24 2.1.3 Bacterial strains 25 2.1.4 Mammalian cell lines 26 2.1.5 Antibodies 26 2.1.6 Primers 27 2.1.7 Plasmids constructed and used in this study 28 2.1.8 Other plasmids used in this study 29 2.1.9 DNA‐ and protein markers 30 2.2 Methods 30 2.2.1 PCR and cloning 30

2.2.1.1 Polymerase chain reaction (PCR) 30 2.2.1.2 Gene synthesis 30 2.2.1.3 Site directed mutagenesis 30 2.2.1.4 Restriction digest 31 2.2.1.5 Agarose gel electrophoresis 31 2.2.1.6 Purification of DNA from agarose gels 31 2.2.1.7 Ligation 31 2.2.1.8 Transformation of DNA into chemical competent E. coli 31 2.2.1.9 Preparation of DNA in low and high scale 31 2.2.1.10 Measurement of DNA and RNA concentrations 32 2.2.1.11 DNA sequencing 32 2.2.2 Maintenance of bacterial cultures and mammalian cell lines 32 2.2.2.1 Bacterial cultivation and preparation of glycerol stocks 32 2.2.2.2 Maintenance of mammalian cell lines 32 2.2.2.3 Freezing of cells in liquid nitrogen 32 2.2.3 Protein expression and ‐purification 32 2.2.3.1 Expression and purification of GST‐fusion proteins in E. coli 32 2.2.3.2 Expression and purification of His‐tagged proteins in E. coli 33 2.2.3.3 Expression and purification of NEDD8 and ubiquitin for click reaction 33 2.2.3.4 Expression and purification of PCNA for click reaction 33 2.2.3.5 In vitro translation 34 2.2.4 Protein analysis 34 2.2.4.1 Bradford assay 34 2.2.4.2 SDS‐polyacrylamide gel electrophoresis (SDS‐PAGE) 35 2.2.4.3 Coomassie Blue and colloidal Coomassie staining 35 2.2.4.4 Fluorography 35 2.2.4.5 Western Blot 35 2.2.5 In vitro assays 36 2.2.5.1 Methanol‐Chloroform precipitation of proteins 36 2.2.5.2 GST‐pulldown assay 36 2.2.5.3 Affinity chromatography 36 2.2.5.4 In vitro NEDDylation assay 37 2.2.5.5 Thioester assay 37 2.2.5.6 Cu(I)‐catalyzed Huisgen azide‐alkyne cycloaddition 37

2.2.5.7 Transient transfection 37 2.2.5.8 TNN cell lysis and ß‐galactosidase assay 38 2.2.5.9 Immunoprecipitation 38 2.2.5.10 Cycloheximide chase 38 2.2.5.11 Cellular fractionation 39 2.2.5.12 Preparation of total RNA 39 2.2.5.13 Reverse transcription 39 2.2.5.14 Antibody purification 39 2.2.6 In cellulo assays 40 2.2.6.1 In cellulo ubiquitination and NEDDylation assays 40 2.2.6.2 Immunofluorescence 40 3. Results 41 3.1 AutoNEDDylation of NEDD8‐conjugating enzymes increases the affinity to their cognate E1 41 3.1.1 Ubc12 and Nce2 form a thioester bond with NEDD8 but not with ubiquitin 42 3.1.2 NEDD8‐conjugating enzymes are autoNEDDylated 43 3.1.3 HPNI mutants of the NEDD8‐conjugating enzymes are not impaired in thioester formation but in autoNEDDylation 45 3.1.4 AutoNEDDylation of the NEDD8 E2 enzymes predominantly occurs in their N terminus 46 3.1.5 AutoNEDDylation enhances the affinity of Ubc12 and Nce2 to the NEDD8 E1 enzyme APPBP1/UBA3 50 3.1.6 Fusion of NEDD8 to its E2s leads to an enhanced localization in the nucleus 51 3.2 PCNA as a new substrate for the NEDD8‐conjugation pathway 54 3.2.1 PCNA interacts with NEDD8 and ubiquitin 54 3.2.2 PCNA is NEDDylated in cells being dependent on K164 56 3.2.3 PCNA NEDDylation depends on the activity of APPBP1/UBA3 in cells 58 3.2.4 NEDDylation of PCNA is enhanced by the E3 ligase Rad18 60 3.2.5 NEDD8‐conjugating enzymes do not bind to PCNA in cells 60 3.2.6 PCNA Y211F, a phosphorylation deficient mutant, is NEDDylated in cells 61 3.2.7 Cu(I)‐catalyzed Huisgen azide‐alkyne cycloaddition as a tool to study functions of NEDDylated PCNA 62 3.3 A new isoform of the NEDD8‐conjugating enzyme Nce2 65 3.3.1 mRNA encoding Nce2 isoform 2 is expressed in HEK293T cells 67

3.3.2 Tertiary structure prediction for Nce2 isoform 2 reveals an unstructured, flexible C terminus 67 3.3.3 Nce2 isoform 2 forms a thioester bond with NEDD8 but not with ubiquitin and is NEDDylated in vitro 68 3.3.4 Nce2 isoform 2 is NEDDylated and ubiquitinated in cells 70 3.3.5 Nce2 isoform 2 has a shorter half‐life than isoform1 and is degraded by the proteasome 72 3.3.6 Nce2 isoform 1 and 2 differ in their subcellular localization 74 3.3.7 Development of an antibody specifically recognizing Nce2 isoform 2 75 4. Discussion 77 4.1 AutoNEDDylation as a regulatory mechanism of NEDD8‐conjugating enzymes 77 4.1.1 NEDD8‐conjugating enzymes are autoNEDDylated in their unique N terminus 77 4.1.1.1 Indications for endogenous autoNEDDylation of Ubc12 and Nce2 77 4.1.1.2 The HPNI‐motif of Ubc12 and Nce2 is important for isopeptide bond formation 78 4.1.1.3 The N termini of Ubc12 and Nce2 are crucial for an efficient autoNEDDylation 79 4.1.2 Possible functions of autoNEDDylation 81 4.1.2.1 AutoNEDDylation as regulator of the subcellular localization of Ubc12 and Nce2 82 4.1.2.2 Acetylation of NEDD8 E2 enzymes as a competitive modification to NEDDylation 82 4.1.2.3 AutoNEDDylation enhances the affinity of NEDD8 E2s to APPBP1/UBA3 83 4.1.2.4 Further possible functions and impacts of autoNEDDylation 85 4.2 Interplay between PCNA and the NEDD8 system 88 4.2.1 PCNA as an interaction partner of NEDD8 88 4.2.2 PCNA as a substrate for NEDD8 89 4.2.2.1 Evidence for NEDDylation of PCNA in vitro and in cellulo 89 4.2.2.2 Effects of MLN4924 on the NEDDylation of PCNA 91 4.2.2.3 Hints for possible functions of NEDDylated PCNA 92 4.2.2.4 Click reaction as tool to identify functions of monoNEDDylated PCNA 94 4.3 A second isoform of Nce2 with individual properties 96 4.3.1 Nce2 isoform 2 as a splice variant with an extended C terminus 96

4.3.2 mRNA of the second isoform of Nce2 is present in HEK293T cells 96 4.3.3 Nce2 isoform 2 is active in vitro and can be NEDDylated 97 4.3.4 The cellular localization distinguishes Nce2 isoform 2 from isoform 1 99 4.3.5 A C‐terminal unstructured region promotes insolubility in bacteria and rapid degradation in cells 100 4.3.6 Possible functions of Nce2 variants 101 5. References 105 Eidesstattliche Erklärung 124 Danksagung 125

Abbreviations

aa amino acid(s) bp base pairs cDNA complementary DNA DMSO Dimethylsulfoxide dNTP Deoxynucleoside triphosphate DTT Dithiothreitol EDTA Ethylenediaminetetraacetic acid E1 UBL‐activating enzyme E2 UBL‐ conjugating enzyme E3 UBL ligase FCS Fetal calf serum GST Glutathione‐S‐transferase HA‐tag Hemagglutinin‐tag His‐tag 6x Histidin‐tag IP Immunoprecipitation IPTG Isopropyl‐β‐D‐thiogalactopyranoside kDa kilo Dalton mRNA messenger RNA Ni‐NTA Nickel‐nitrilotriacetic acid OD optical density ONPG Ortho‐nitrophenyl‐β‐galactoside PBS Phosphate buffered saline PCR polymerase chain reaction rpm revolutions per minute RT reverse transcription SDS Sodium dodecyl sulfate UBL ubiquitin‐like protein wt wild‐type

i

Abstract

NEDD8 belongs to the family of ubiquitin‐like proteins which share a common basic structure. In an enzymatic cascade, NEDD8 can be attached to other proteins via its C terminus (“NEDDylation”). NEDDylation plays an important role in various cellular processes including cell cycle, transcription or the regulation of protein stability. Previous studies revealed NEDD8 to have many yet uncharacterized substrates. During this work, the sliding clamp PCNA, which functions in replication and DNA damage repair, was identified as a new substrate for NEDDylation. PCNA is modified with NEDD8 at the same lysine (K164) residue that can be targeted for SUMOylation and ubiquitination. Rad18, the E3 ligase for ubiquitination of PCNA upon DNA damage, enhances monoNEDDylation of PCNA. We furthermore expressed NEDD8 and PCNA containing non‐natural amino acids that can be used for the formation of a triazole linkage via click reaction. This “monoNEDDylated” PCNA provides a tool for a functional characterization in the future. In a further part of this study, regulation and function of the NEDD8‐conjugating enzymes Ubc12 and Nce2 were investigated. To date, not much is known about if and how the E2 enzymes of the NEDD8‐conjugation cascade are regulated. We here provide evidence that these enzymes are autoNEDDylated in the N‐terminal extension of the catalytic core domain, which is not present in most other E2 enzymes. This autoNEDDylation of Ubc12 and Nce2 appears to enhance their affinity to APPBP1/UBA3, the E1 enzyme of the NEDD8‐conjugation cascade. Using NEDD8‐E2 fusion proteins an increased efficiency in NEDD8 transfer from the E1 to the E2 enzyme compared to wt Ubc12 and Nce2 was observed. However, the exact function of autoNEDDylation in cells, which might be a change in substrate specificity, enhanced substrate NEDDylation or a switch‐off mechanism for the whole cascade, still needs to be determined. To gain further insights into function and regulation of NEDD8‐conjugating enzymes, a second isoform of Nce2 was characterized which has recently been identified in a screen for splice variants. Nce2 isoform 2 mRNA was verified to be present in cells. To prove the existence of endogenous protein, an antibody specifically recognizing the second, but not the first isoform of Nce2 was developed. Tertiary structure prediction of the new isoform revealed an unstructured, flexible C terminus. Using overexpressed isoform 2, a difference in the subcellular localization of both isoforms was demonstrated. In contrast to Nce2 isoform 1, isoform 2 is probably not autoNEDDylated, but regulated via ubiquitination leading to a shorter half‐life. In future studies, the functionality and the role of Nce2 isoform 2 in cells need to be investigated. In conclusion, identification of PCNA as a new substrate of NEDD8, evidence for a regulation of NEDD8‐conjugating enzymes by autoNEDDylation as well as the new isoform of Nce2 provide insights into yet unknown physiological functions of NEDD8.

ii

Zusammenfassung

NEDD8 gehört zur Familie der Ubiquitin‐ähnlichen Proteine, die eine gemeinsame Grundstruktur aufweisen. In einer enzymatischen Kaskade kann NEDD8 über seinen C‐ Terminus an andere Proteine konjugiert werden („NEDDylierung“). Die NEDDylierung spielt eine wichtige Rolle in verschiedenen zellulären Prozessen wie z.B. dem Zellzyklus, der Transkription oder der Regulation der Proteinstabilität. Frühere Studien wiesen bereits auf die Existenz einiger weiterer, bisher noch nicht charakterisierter Substrate von NEDD8 hin. Im Rahmen dieser Arbeit konnte PCNA, das sowohl für die Replikation als auch für die DNA‐ Reparatur benötigt wird, als neues Substrat für NEDD8 identifiziert werden. PCNA wird an demselben Lysinrest (K164) monoNEDDyliert, der auch SUMOyliert und ubiquitiniert werden kann. Diese MonoNEDDylierung von PCNA wird durch die E3 Ligase Rad18 verstärkt, die auch die Ubiquitinierung von PCNA nach DNA‐Schäden vermittelt. Während dieser Studien wurden zudem NEDD8 und PCNA mit nicht‐natürlichen Aminosäuren exprimiert, die in der sog. „Click Reaktion“ eine Triazolbindung eingehen können. Das so gebildete „monoNEDDylierte“ PCNA kann zukünftig für eine funktionelle Charakterisierung verwendet werden. Ein weiterer Teil dieser Arbeit beinhaltet die Untersuchung der Regulation und der Funktion der NEDD8‐konjugierenden Enzyme Ubc12 und Nce2. Bisher war nicht sehr viel darüber bekannt, ob und wie die E2‐Enzyme der NEDDylierungskaskade reguliert werden. Im Zuge dieser Arbeit konnte der Beweis erbracht werden, dass beide Enzyme autoNEDDyliert werden können. Diese AutoNEDDylierung findet dabei in der N‐terminalen Verlängerung der katalytischen Kerndomäne der E2s statt, die in den meisten anderen bekannten E2‐Enzymen nicht zu finden ist. Die AutoNEDDylierung von Ubc12 und Nce2 scheint die Affinität zu APPBP1/UBA3, dem E1‐ Enzym der NEDDylierungskaskade, zu verstärken. Unter Zuhilfenahme von NEDD8‐E2 Fusionsproteinen konnte gezeigt werden, dass der Transfer von NEDD8 von dem E1‐ zum E2‐ Enzym im Vergleich zu Ubc12 und Nce2 wt effizienter ist. Die genaue Funktion der AutoNEDDylierung der NEDD8 E2s in Zellen gilt es allerdings noch zu untersuchen. Möglicherweise resultiert sie in einer Änderung der Substratspezifität, einer erhöhten SubstratNEDDylierung oder einem Abschaltmechanismus für die gesamte Kaskade. Um weitere Kenntnisse über Funktion und Regulation von NEDD8‐konjugierenden Enzymen zu erlangen, wurde eine zweite Isoform von Nce2 charakterisiert, die kürzlich in einem Screen nach Splice‐Varianten entdeckt wurde. Die mRNA dieser Isoform konnte tatsächlich in Zellen nachgewiesen werden. Zwecks Detektion des endogenen Proteins wurde ein Antikörper entwickelt, der spezifisch die zweite, aber nicht die erste Isoform von Nce2 erkennt. Das erstellte Modell der Tertiärstruktur der Isoform 2 lässt einen unstrukturierten und flexiblen C‐Terminus erkennen. Mittels Überexpression von Isoform 2 konnte ein Unterschied in der subzellulären

iii

Lokalisation beider Isoformen nachgewiesen werden. Überdies wird Isoform 2 im Gegensatz zur ersten Isoform wahrscheinlich nicht autoNEDDyliert, aber über Ubiquitinierung reguliert, die eine verkürzte Halbwertszeit des Proteins zur Folge hat. In zukünftigen Studien muss die Funktionalität und die Rolle der zweiten Isoform von Nce2 in der Zelle untersucht werden. Die Identifikation von PCNA als neues Substrat von NEDD8, der Beweis einer Regulation der E2‐ Enzyme mittels AutoNEDDylierung sowie der Nachweis einer neuen Isoform von Nce2 ermöglichen bisher unbekannte Einblicke in die physiologische Rolle von NEDD8.

iv

1. Introduction

Protein function, localization and stability are regulated by posttranslational modifications which involve for example small chemical groups such as acetyl‐, methyl‐ or phosphate moieties, or larger groups serving as membrane anchors such as palmitoyl‐ or myristoyl groups (Han and Martinage, 1992; Magee and Courtneidge, 1985). With the discovery of the protein ubiquitin and its ability to serve as posttranslational modification in the 1980s, a new chapter of the functions of these modifications has begun. Ubiquitin was found to target proteins for proteasomal degradation thereby offering the cell a second major pathway to regulate protein levels in addition to lysosomal degradation. In the meantime, several proteins with high structural similarity to ubiquitin were identified most of which act as posttranslational modifiers controlling various cellular functions (Hershko and Ciechanover, 1998; Kerscher et al., 2006).

1.1 Ubiquitin‐proteasome system

Ubiquitin is a small protein of 76 aa and 8.6 kDa which is highly conserved among eukaryotic organisms (Hershko and Ciechanover, 1998). Structurally, ubiquitin is characterized by a globular domain with a β‐grasp fold and a flexible C terminus (Figures 1 and 4B) (Vijay‐Kumar et al., 1987). C

N

Figure 1. Crystal structure of ubiquitin revealing the characteristic β­grasp fold Ribbon representation of human ubiquitin (protein data base 1UBQ; modeled with Pymol). The structure of ubiquitin is characterized by four antiparallel β‐sheets grasping an α‐helix. The C terminus of ubiquitin protrudes from the globular domain (Vijay‐Kumar et al., 1987).

- 1 - Introduction

1.1.1 Ubiquitin‐conjugation cascade

In humans, expression of ubiquitin from one of the four different genes first leads to the formation of an inactive precursor protein (Baker and Board, 1987; Lund et al., 1985; Wiborg et al., 1985). Being rapidly processed by ubiquitin‐specific proteases, the functionally important C‐ terminal double glycine motif of ubiquitin is exposed (Wilkinson, 1997). In an enzymatic cascade, ubiquitin is covalently attached to other proteins via its C terminus (“ubiquitination”). Ubiquitination occurs through several consecutive steps catalyzed by three (or four) different classes of enzymes: ubiquitin‐activating enzymes (E1), ubiquitin‐conjugating enzymes (E2) and ubiquitin ligases (E3) (Figure 2). In some cases, an E4 enzyme may be required for the formation of ubiquitin chains (Hershko and Ciechanover, 1998; Hoppe, 2005). In a first step, the carboxyl group of the C‐terminal glycine of ubiquitin forms a high energy thioester linkage with an active site cysteine residue of one of two E1 enzymes, UBA1 or UBA6. This step is ATP‐dependent and involves the formation of a ubiquitin adenylate intermediate (Jin et al., 2007; Pelzer et al., 2007; Pickart, 2001). The E2 is then able to accept activated ubiquitin from its cognate E1, forming a thioester linkage. In the last step, which in most cases requires the presence of an E3, ubiquitin is transferred to the lysine residue of a substrate protein by the formation of an isopeptide bond (Pickart, 2001) (Figure 2). An asparagine residue in the conserved HPNI/V motif next to the catalytic cysteine of the E2 enzyme plays a crucial role for accomplishing this transfer, as it might be important for oxyanion intermediate stabilization during lysine attack (Wu et al., 2003b). Structural and functional analyses indicate that E3s consist of at least two functional domains: one domain (i.e. RING/RING‐like or HECT) is interacting with the cognate E2 while the other mediates the specific interaction with the substrate, thereby conferring substrate specificity on the whole conjugation cascade. Depending on their mode of action, E3 ligases can be divided into two major classes: HECT and RING/RING‐like ligases (Metzger et al., 2012). HECT ligases contain a HECT (Homologous to E6AP Carboxyl Terminus) domain which consists of an N‐terminal and a C‐terminal lobe. The N‐terminal lobe is necessary for the interaction with the E2 enzyme, whereas the C‐terminal lobe contains a catalytic cysteine which forms a thioester linkage with ubiquitin before transferring it to the substrate (Huang et al., 1999).

- 2 - Introduction

ATP AMP+PPi

Figure 2. Ubiquitin­conjugation cascade Ubiquitin is first activated by the E1 and then transferred to one of a number of E2 enzymes. RING E3 ligases facilitate the conjugation of ubiquitin to the substrate by acting as adaptors between E2s and substrates. In contrast, HECT E3 ligases form thioester bonds with ubiquitin and subsequently transfer it to the substrate. By repeating these steps, the substrate can be modified with several ubiquitin moieties (modified from (Di Fiore et al., 2003)).

Being encoded by more than 600 genes in mammals, RING ligases represent the largest family of E3 ligases (Li et al., 2008). This type of ligase is characterized by a RING (Really Interesting New Gene) domain containing the consensus motif C‐X2‐C‐X(9‐39)‐C‐X(1‐3)‐H‐X(2‐3)‐C/H‐X2‐C‐X(4‐ 48)‐C‐X2–C (X stands for any aa), which is stabilized by two zinc ions (Lovering et al., 1993). By bringing the E2 enzyme and the substrate into close proximity, RING ligases allow a direct transfer of ubiquitin from the E2 to the substrate. The family of RING ligases comprises ligases that function as monomer, homo‐ or heterodimer or as large multi‐subunit complexes. Most of the known RING ligases possess intrinsic E3 ligase activity, but there are also RING domain containing proteins that do not show activity by themselves, e.g. Bard1, Bmi1 and HdmX. Heterodimerization of these E3 ligases with another RING ligase (Brca1, Ring1B and Hdm2, respectively), which involves the interaction of the RING domains, leads to the stimulation of ligase activity of the latter (Hashizume et al., 2001; Linares et al., 2003; Wang et al., 2004). Moreover, other RING E3 ligases like RNF4 or TRAF6 form

- 3 - Introduction homodimers to execute their function (Liew et al., 2010; Yin et al., 2009). Recent data about the mechanism of ubiquitin transfer by RNF4 sheds light on why dimerization is necessary for RING ubiquitin ligase activity. Dimerization of the RING domains facilitates the interaction with both components of the ubiquitin‐charged E2 at the same time: one monomer binds to ubiquitin and the other one to the E2. By altering the conformation of the active site of the E2, the E2‐ubiquitin thioester bond is then activated and ubiquitin can be transferred to the substrate (Plechanovova et al., 2011; Plechanovova et al., 2012). Being composed of multiple subunits, the APC/C complex and cullin‐RING ligases such as SCF form an additional type of RING E3 ligases. APC/C, a complex consisting of at least twelve subunits including the RING domain containing protein APC11, plays a major role in cell cycle regulation, especially in mitotic progression. Dependent on the substrate adaptor bound, a specific set of substrates is recognized, ubiquitinated and degraded by the proteasome (reviewed in (Peters, 2006)). The SCF complex belongs to the largest family of ubiquitin E3 ligases known, the cullin‐RING ligases (see chapter 1.3.1.1). It consists of Cullin1 as scaffold protein, the RING ligase RBX1, Skp1 as adaptor protein and an exchangeable F‐box protein recognizing a specific substrate. Not only cell cycle inhibitors like p21 or p27, but also oncogenic proteins such as Cyclin E or c‐Myc turned out to be substrates for SCF complexes, underlining its important role in controlling the cell cycle (summarized in (Kitagawa et al., 2009)).

1.1.2 Modes of ubiquitination

In most cases, ubiquitination of substrates occurs via isopeptide bond formation between the C‐ terminal glycine residue of ubiquitin and the ‐amino group of a lysine residue in the substrate. Furthermore, there are few publications showing that serine, threonine and cysteine residues as well as the N‐terminal amino group of some proteins are used for the attachment of ubiquitin (Cadwell and Coscoy, 2005; Ciechanover and Ben‐Saadon, 2004; Shimizu et al., 2010; Wang et al., 2007b). In addition to mono‐ or multiubiquitination where single ubiquitin moieties are conjugated to one or several distinct residues in the substrate protein, respectively, ubiquitin is also capable of forming “chains” (polyubiquitination) (Figure 3). Monoubiquitination was found to be involved in DNA damage response and endocytosis (reviewed in (Hicke, 2001)). For instance, Rad6‐ and Rad18‐dependent ubiquitination of the sliding clamp PCNA leads to the recruitment of translesion synthesis polymerases that are crucial for inducing the DNA damage tolerance pathway (Hoege et al., 2002; Kannouche et al., 2004). Moreover, multiubiquitination plays a particular role in the internalization and lysosomal degradation of plasma membrane receptors, e.g. receptor tyrosine kinases (Haglund et al., 2003).

- 4 - Introduction

Each of the seven internal lysine residues of ubiquitin (K6, K11, K27, K29, K33, K48 and K63) can be used for isopeptide bond formation with the C‐terminal carboxyl group of another ubiquitin moiety and hence, for the formation of polyubiquitin chains (Figure 3). Best characterized and most abundant are K11‐, K48‐ and K63‐linked chains (Ye and Rape, 2009). As an example, the E3 ligase complex APC/C triggers degradation of its mitotic substrates via formation of K11‐linked ubiquitin chains (Jin et al., 2008). Furthermore, the ubiquitin‐ conjugating enzyme Ubc6 which is involved in ER‐associated degradation (ERAD), a protein quality control system mainly localized in the ER membrane, is ubiquitinated and degraded via K11‐linkage of ubiquitin in yeast (Xu et al., 2009). Nonetheless, K48‐linked chains are a more common signal to target proteins for proteasomal degradation (Chau et al., 1989). For a long time it was believed that a minimal chain length of four ubiquitin moieties linked via K48 is absolutely required for the recognition by the proteasome (Thrower et al., 2000). This hypothesis is contradicted by recent studies indicating that monoubiquitinated or multiubiquitinated proteins like PAX3 or p105, respectively, can also be recognized and degraded by the proteasome (Boutet et al., 2007; Kravtsova‐Ivantsiv et al., 2009; Shabek et al., 2012). In contrast to K11‐ and K48‐chains, K63‐linked chains are relevant for a variety of non‐ proteolytic cellular processes like endocytosis, DNA repair or activation of kinases (Deng et al., 2000; Duncan et al., 2006; Spence et al., 1995). Mixed and forked chains, as well as chains formed on internal lysine residues other than K11, K48 and K63 of ubiquitin are still under intense investigation (Figure 3).

Figure 3. Modes of ubiquitination Ubiquitin is primarily attached to the ‐amino group of one or more lysine residues of a substrate (mono‐ and multiubiquitination, respectively) or, in rare cases, to the N terminus of a substrate (not shown). It contains seven internal lysine residues all of which can serve as an acceptor for another ubiquitin moiety, thereby forming ubiquitin chains (polyubiquitination). K11‐, K48‐, K63‐linked and linear chains, which are linked via C‐ and N terminus of two ubiquitin moieties, function in proteasomal degradation, DNA‐repair or intracellular signaling, whereas the function of the other lysine linked chains and forked chains remains elusive (Ye and Rape, 2009). - 5 - Introduction

1.1.3 Ubiquitin recycling and the proteasome

Processing of ubiquitin precursor proteins as well as recycling of ubiquitin is carried out by deubiquitinating enzymes (DUBs) which mainly exhibit cysteine protease activity. The two major classes of DUBs are UCHs (Ubiquitin COOH‐terminal Hydrolases) that preferentially cleave ubiquitin from substrates and USPs (Ubiquitin‐Specific Proteases) that additionally hydrolyze isopeptide bonds between two ubiquitin moieties. In addition, OUTs (otubain proteases), JAMM metalloproteases and MJDs were found to function as DUBs (summarized in (Sorokin et al., 2009)). Some DUBs are associated with or even part of the 26S proteasome which degrades 80‐90 % of intracellular proteins (Rock et al., 1994). The 26S proteasome consists of the 20S core particle and two 19S regulatory particles. The core proteasome is composed of 14 α‐ and 14 β‐subunits forming four heptameric rings which build a channel in whose inner part the target protein is hydrolyzed. Three of the β‐subunits possess proteolytic activities ensuring efficient cleavage of the target: subunit β1 exhibits caspase‐like activity, β2 has trypsin‐like and β5 chymotrypsin‐ like activity. The regulatory particles contain subunits that fulfill three important functions: interaction with the ubiquitinated substrate, cleaving ubiquitin from the substrate (by isopeptidases) and unfolding the target protein (by ATPases) (reviewed in (Murata et al., 2009; Sorokin et al., 2009)).

1.2 Ubiquitin‐like proteins (UBLs)

The family of ubiquitin‐like proteins (UBLs) consists of more than a dozen members in mammals that are structurally characterized by the β‐grasp fold (Figures 1 and 4). UBLs use a similar conjugation mechanism as ubiquitin which involves the subsequent action of activating and conjugating enzymes and, in most cases, that of ligases. Functions of UBLs vary from regulation of protein stability, interactions or localization, autophagy and pre‐mRNA splicing to regulation of inflammation and development (see table 1) (summarized in (Herrmann et al., 2007; Kerscher et al., 2006; Schulman and Harper, 2009). UBLs are conserved among eukaryotes. Nevertheless, it is speculated that they have arisen from a common ancestor in prokaryotes which plays a role in sulfur metabolism. The reason for this hypothesis lies in the existence of proteins in E. coli that reveal a ubiquitin‐like structure and exhibit parallels to the ubiquitin‐conjugation cascade. One of those proteins is MoaD, a sulfur carrier protein being involved in the biosynthesis of the molybdenum cofactor Moco (Rivers et al., 1993). The double glycine motif at the C terminus of MoaD is adenylated by MoeB, which shows similarities to the ubiquitin E1 enzyme. Adenylation of MoaD leads to its activation, subsequent formation of a thiocarboxylate at its C terminus and the insertion of sulfur into the

- 6 - Introduction

Moco precursor protein. Interestingly, the eukaryotic ubiquitin‐like protein URM1 might represent the connection to the prokaryotic system since it functions in sulfur transfer, too (reviewed in (Hochstrasser, 2009)).

Table 1. Ubiquitin­like proteins Summary of the known ubiquitin‐like proteins in humans, their functions and enzymes involved in their conjugation; NA: data not available; NC: no conjugation to other proteins (adapted from (Herrmann et al., 2007)).

Ubiquitin­like protein E1, E2, E3 enzymes Functions

NEDD8 E1: APPBP1/UBA3 Transcriptional regulation, E2: Ubc12, Nce2 protein stability, E3: e.g. Hdm2, RBX1/2 proteasomal degradation… SUMO1‐3 E1: SAE1/SAE2 Transcriptional regulation, E2: Ubc9 protein stability, protein E3: RanBP2, Pc2, PIAS localization… ISG15 E1: UBE1L IFN‐induced immune E2: UbcH8 response

FAT10 E1: UBA6 Proteasomal degradation, E2: USE1 apoptosis LC3/GABARAP E1: ATG7 Autophagy E2: ATG3 ATG12 E1: ATG7 Autophagy E2: ATG10 FUB1 NA Regulation of leukocyte functions URM1 E1: Uba4 Oxidative stress response UFM1 E1: Uba5 Endoplasmic stress E2: Ufc1 response

HUB1 NC Pre‐mRNA splicing

1.2.1 SUMO

The small ubiquitin‐related modifier (SUMO), which is 11.5 kDa in size, is one of the best characterized ubiquitin‐like proteins. In vertebrates, there are four SUMO paralogues revealing the typical β‐grasp fold with a short N‐terminal extension (Bayer et al., 1998; Bohren et al., 2004; Kamitani et al., 1998). In contrast to SUMO‐2 and SUMO‐3, the elongated N terminus of SUMO‐1 does not contain a SUMOylation site and thus, it is not able to form chains (Saitoh and Hinchey, 2000; Tatham et al., 2001). All SUMO family members need to be processed in order to expose their double glycine motif at the C terminus (Johnson et al., 1997). However, SUMO‐4 is the only

- 7 - Introduction member which is probably not conjugated to substrates (Owerbach et al., 2005). According to their subcellular localization, SUMO‐1, ‐2 and ‐3 are able to carry out distinct functions: SUMO‐1 is mostly found at the nuclear membrane whereas SUMO‐2 is localized in nuclear bodies and nucleoli and SUMO‐3 in the cytoplasm and in nucleoli (reviewed in (Hay, 2005)). The SUMO‐conjugation cascade starts with its activation by the heterodimeric E1 enzyme SAE1/SAE2 with SAE2 containing the active site cysteine. Ubc9, the SUMO conjugating enzyme which is particularly found in the nucleus, is able to directly transfer SUMO to substrates without involvement of an E3 enzyme (Desterro et al., 1999; Schwarz et al., 1998). Most SUMO substrates contain a special binding motif for Ubc9 consisting of the four amino acids [I/V/L]‐K‐X‐[D/E] (X stands for any aa), which also contains the lysine residue that in turn is SUMOylated (Bernier‐ Villamor et al., 2002; Johnson, 2004). SUMO E3 ligases belong to the family of RanBP2‐, Pc2‐ or PIAS E3 ligases that have specific substrates and differ in their localization in the cell (Hay, 2005; Johnson and Gupta, 2001; Kagey et al., 2003; Pichler et al., 2002). For SUMOylation, more than 100 substrates have been identified so far that are partially overlapping for SUMO‐1 and SUMO‐ 2/‐3. These substrates can mainly be grouped into transcriptional regulators, nuclear envelope proteins, signaling proteins and cell membrane proteins (Rosas‐Acosta et al., 2005; Vertegaal et al., 2004). For instance, SUMOylation of RanGAP1, a factor which is necessary for nucleo‐ cytoplasmic transport, regulates its localization (Matunis et al., 1996). Another substrate of SUMO is PML which needs to be SUMOylated to initiate the formation of PML nuclear bodies that are involved in important cellular functions like transcriptional regulation (Shen et al., 2006; Sternsdorf et al., 1997). A unique feature of the SUMOylation cascade is the regulation of substrate specificity by autoSUMOylation of the E2 enzyme Ubc9. SUMOylation of Ubc9 occurs in its N‐terminal α‐helix and does not affect thioester bond formation with SUMO. Indeed, autoSUMOylation of Ubc9 leads to inhibited SUMOylation of RanGAP1 and enhanced SUMOylation of Sp100, which is at least partially due to a SUMO‐interaction motif (SIM) in Sp100 (Knipscheer et al., 2008).

1.2.2 Other UBLs

In the 1980s, ISG15 (Interferon‐Stimulated Gene product of 15 kDa) was the first ubiquitin‐like protein to be identified. ISG15 exhibits a dimeric ubiquitin structure. As the name already indicates, ISG15 expression is interferon‐inducible which is in line with its role in the induction of inflammation (D'Cunha et al., 1996; Haas et al., 1987; Korant et al., 1984). FAT10 is characterized by two head‐to‐tail linked ubiquitin‐like domains which were shown to be able to directly interact with the proteasome. Accordingly, FAT10 offers a ubiquitin‐ independent way of delivering proteins for proteasomal degradation (Fan et al., 1996; Hipp et al., 2005). The double glycine motif at the C terminus of FAT10 is activated by UBA6, transferred

- 8 - Introduction to USE1 and attached to other proteins as shown by the presence of FAT10 conjugates in cells (Aichem et al., 2010; Jin et al., 2007; Pelzer et al., 2007; Raasi et al., 2001). However, only four substrates of FAT10 have been identified to date, among these are USE1 which is autoFAT10ylated in a cis‐mechanism, the ubiquitin‐activating enzyme UBA1, p53, and the autophagosomal receptor p62 (Aichem et al., 2012; Aichem et al., 2010; Li et al., 2011; Rani et al., 2012). Two UBLs, the LC3/GABARAP family and ATG12, are involved in autophagy, a conserved mechanism in eukaryotes to degrade macromolecules and organelles. Although showing no obvious homology to ubiquitin, they exhibit a similar structure and conjugating system (Ichimura et al., 2000; Sugawara et al., 2004; Suzuki et al., 2005). The LC3/GABARAP family (in mammals; Atg8 in yeast) is conjugated to phosphatidylethanolamine thereby initiating autophagosome formation. Autophagosomes are double‐membrane vesicles which deliver their content to the lysosome (Geng and Klionsky, 2008; Ichimura et al., 2000; Tanida et al., 2004). Interestingly, aggregated ubiquitinated proteins were found to be targets for so called “selective” autophagy. Some adaptor proteins are associated with the autophagosome via LC3 and interact with ubiquitinated proteins at the same time, thereby leading to lysosomal degradation of the respective protein (reviewed in (Kirkin et al., 2009)). Another UBL, ATG12, is conjugated constitutively to ATG5 which is necessary for completion of autophagosome formation (Geng and Klionsky, 2008; Ichimura et al., 2000).

1.3 NEDD8

NEDD8 (Neural precursor cell‐Expressed Developmentally Down‐regulated 8) was first discovered in a subset of genes that are down‐regulated during mouse brain development and is highly conserved from yeast to humans. With 60 % identity and 80 % homology to ubiquitin, NEDD8 forms its closest relative among the family of ubiquitin‐like proteins (Kumar et al., 1992; Kumar et al., 1993). Human NEDD8 is expressed as an inactive, 81 aa precursor protein which needs to be processed in order to be conjugated to substrates (Kamitani et al., 1997). Similar to ubiquitin, the conjugation of NEDD8 involves three consecutive steps being catalyzed by the heterodimeric NEDD8 E1 enzyme APPBP1/UBA3, one of the two E2 enzymes Ubc12 or Nce2, and an E3 enzyme (Gong and Yeh, 1999; Huang et al., 2009). This process called “NEDDylation” will be described in detail in section 1.3.2. NEDD8 is a 9.1 kDa protein with a high structural similarity to ubiquitin. Its globular β‐grasp fold includes four antiparallel β‐sheets and one helix on top of them (Figure 4). In addition, surface charge distribution resembles that of ubiquitin: one face contains hydrophobic residues whereas the other one is acidic (Whitby et al., 1998).

- 9 - Introduction

A B C C

N N

Figure 4. Crystal structure of NEDD8 (A) and overlay with ubiquitin and SUMO­1 (B) (A) The globular NEDD8 structure reveals a β‐grasp fold with four antiparallel β‐sheets and an α‐helix on top of them, being characteristic for ubiquitin and all UBLs (protein data base 1NDD; modeled using Pymol). The C‐terminal double glycine motif is marked in green. (B) Overlay of ubiquitin (blue), SUMO‐1 (green) and NEDD8 (red). Ubiquitin and the UBLs SUMO‐1 and NEDD8 exhibit a highly similar structure characterized by the β‐grasp fold (Welchman et al., 2005).

Investigation of NEDD8 expression in different tissues using northern blot and immuno‐ cytochemical analysis revealed NEDD8 mRNA is enriched in brain and skeletal muscle, and that NEDD8 protein is predominantly found in the nucleus whereas ubiquitin is equally distributed in the cell (Kamitani et al., 1997). In mice, knockout of the catalytic subunit of the NEDD8 E1 enzyme, UBA3, was shown to be lethal in utero at the periimplantation state. The reason for this phenomenon lies in the involvement of the NEDD8 pathway in cell cycle progression and morphogenesis, underlining its indispensability for early development (Tateishi et al., 2001). In S. cerevisiae, however, depletion of the NEDD8 homologue Rub1, as well as the respective E1 or the E2 enzymes does not affect normal cell growth (Liakopoulos et al., 1998).

1.3.1 Substrates and functions of NEDD8

1.3.1.1 Cullins In 1998, Cullin4a was described as the first substrate of the NEDD8‐conjugation system (Osaka et al., 1998). In the meantime, it turned out that modification of all canonical cullins with NEDD8 is crucial to execute their function (Hori et al., 1999; Ohh et al., 2002). Cullins are components of multi‐protein complexes, the cullin‐RING ligases, which play important roles in cell growth, development, signal transduction, transcriptional control, genomic integrity and tumor suppression (see also 1.1.1). In mammals, there are six canonical cullins (Cullin1, Cullin2, Cullin3, Cullin4a, Cullin4b, Cullin5) and three atypical cullins (APC2, Cullin7 and PARC) known

- 10 - Introduction which together build more than 500 distinct multi‐subunit complexes (Petroski and Deshaies, 2005a; Skaar et al., 2007; Zachariae et al., 1998). Cullin‐RING ligases usually consist of cullins serving as scaffold proteins, substrate adaptors, which recognize and bind the respective substrate, and a RING domain protein with the catalytic activity to ubiquitinate the substrate. NEDDylation of the cullin subunit at a conserved lysine residue leads to a conformational change which in turn recruits the ubiquitin‐loaded E2 enzyme (Duda et al., 2008; Kawakami et al., 2001). In addition, NEDDylation seems to promote dimerization of cullin‐RING ligases through their substrate recognition subunit (Wimuttisuk and Singer, 2007). Recently, it was suggested that rotation of the RING domain of the E3 also plays a crucial role in the activation of the cullin‐RING complex (Calabrese et al., 2011). Both the NEDD8‐ and the ubiquitin E3 ligase activity of the complex are carried out by either RBX1 or RBX2 (Petroski and Deshaies, 2005a). Moreover, Dcn1 in yeast and DCNL proteins in humans serve as a scaffold‐type E3 ligases which interact with the respective cullin and the NEDD8 E2 at the same time thereby enhancing NEDDylation of cullins and thus, the ubiquitination activity of cullin‐RING complexes (Kurz et al., 2008; Monda et al., 2013; Yang et al., 2007). As already mentioned, cullin complexes share a general composition. Dependent on the cullin, however, several different substrate adaptors can be used. So called SCF complexes consist of Cullin1, RBX1, the adaptor protein Skp1 and an F‐box protein which specifically interacts with the substrate to be degraded (Lyapina et al., 1998). One of the substrates for the SCF complex with Skp2 as F‐box protein is p21, an important regulator of cell proliferation and differentiation. Phosphorylation of p21 by Cdk2‐Cyclin E enhances its recognition and ubiquitination by SCFSkp2, leading to its rapid degradation (Bornstein et al., 2003). Additionally, not only cell cycle inhibitors but also cell cycle activators like Cyclin E are substrates for SCFSkp2 (Nakayama et al., 2000). In contrast to Cullin1, Cullin2 and Cullin5 assemble with elongin B/C as adaptor and a SOCS‐box containing protein as substrate receptor. A well‐studied substrate of the Cullin2‐RING ligase is HIFα, an important player in oxygen metabolism. A complex formed by HIFα and HIFβ under hypoxic conditions controls the expression of several genes like VEGF or erythropoietin (Maxwell, 2003). Oxygen‐dependent hydroxylation of HIFα leads to its recognition by the tumor suppressor protein pVHL, a substrate receptor of the Cullin2‐complex, and its subsequent ubiquitination and degradation (Ivan et al., 2001; Jaakkola et al., 2001; Petroski and Deshaies, 2005a; Yu et al., 2001). Deactivation of cullin‐RING complexes is achieved through removal of NEDD8 by the CSN5 subunit of the COP9 signalosome which belongs to the family of JAMM metalloproteases (Cope et al., 2002; Lyapina et al., 2001). Another mechanism to silence cullin‐RING complexes is the binding of Cand1 (Cullin‐Associated and NEDDylation‐Dissociated 1) which inhibits NEDDylation and the assembly of the whole complex by specifically binding to the region of the NEDD8 acceptor lysine within the cullin (Zheng et al., 2002). However, it was recently shown

- 11 - Introduction that Cand1 also serves as an important exchange factor for cullin‐RING ligase adaptors (Pierce et al., 2013).

1.3.1.2 NEDD8 and transcriptional regulation In the last decade, several new substrates of NEDD8 were described, giving insights into the role of NEDD8 in various cellular functions. Interestingly, a majority of these substrates is involved in transcriptional regulation. The tumor suppressor protein p53, which plays a major role in the regulation of cell cycle arrest and apoptosis, is not only a substrate for ubiquitin but also for ubiquitin‐like proteins such as SUMO or NEDD8 (Kruse and Gu, 2009; Rodriguez et al., 1999; Scheffner et al., 1993; Xirodimas et al., 2004). NEDDylation of p53 by the RING ligase Hdm2 was shown to inhibit its transcriptional activity. In the case of p53, Hdm2 displays a dual specificity since it had also been described as an E3 ligase for the ubiquitination of p53 (Honda et al., 1997; Xirodimas et al., 2004). Modification of p53 with NEDD8 is also promoted by FBXO11 and specifically inhibited by the histone acetyltransferase Tip60 (Abida et al., 2007; Dohmesen et al., 2008). Interestingly, C‐ terminal fusions of p53 with ubiquitin and NEDD8 to mimic its modification with these proteins showed that p53‐ubiquitin is rather found in the cytoplasm, whereas fusions of p53 with NEDD8 localize in the nucleus (Carter and Vousden, 2008). Another member of the p53 family, TAp73, also serves as a substrate for Hdm2‐dependent NEDDylation. Modification of TAp73 with NEDD8 inhibits its transcriptional activity which can in part be explained by its localization to the cytoplasm (Watson et al., 2006). In 2008, some ribosomal proteins were found to be modified with NEDD8 causing enhanced stability (Xirodimas et al., 2008). Further investigation of the ribosomal protein L11 revealed that its NEDDylation leads to a localization to the nucleolus (Sundqvist et al., 2009). Upon nucleolar stress, L11 is deNEDDylated and recruited to promoters of p53 regulated genes where it interacts with several co‐activators. In addition, binding of L11 to Hdm2 at these promoter sites inhibits interaction of p53 with Hdm2, thereby promoting transactivation of p53 target genes. Interestingly, ribosome biogenesis is not affected under deNEDDylation conditions in spite of reduced L11 levels (Mahata et al., 2011). NEDDylation also seems to play an important role in the regulation of NFκB activity. On the one hand, suppression of the transcriptional activity of NFκB by BCA3 is dependent on its modification with NEDD8 (Gao et al., 2006). On the other hand, TRIM40‐catalyzed NEDDylation of IKKγ, an inhibitor of NFκB signaling, enhances the repression of NFκB (Noguchi et al., 2011). Another protein which is regulated by NEDDylation is AICD, the intracellular domain of the amyloid precursor protein (APP). APP is found in plaques that accumulate in brains of Alzheimer patients. Cleaving of APP by secretases leads to the formation of AICD amongst others, whose role in Alzheimer development and progression is only poorly understood. Modification of AICD

- 12 - Introduction with NEDD8 prevents the interaction with its co‐activator Fe65 and the histone acetyltransferase Tip60, resulting in an inhibition of the transactivator function for genes involved in e.g. cell growth and motility (Lee et al., 2008; Muller et al., 2008).

1.3.1.3 Further substrates and functions of NEDD8 Modification with NEDD8 does not only play roles in transcriptional regulation but also in the regulation of protein stability. The ribosomal protein L11 and the E3 ligase Hdm2 as well as PINK1, a protein involved in Parkinson´s disease, show an enhanced stability upon modification with NEDD8 (Choo et al., 2012; Xirodimas et al., 2004; Xirodimas et al., 2008). In addition, NEDDylation regulates the stability of distinct receptors. Ubiquitination of EGFR is mediated by the RING‐ligase c‐Cbl, leading to its internalization and lysosomal degradation. C‐Cbl is also capable of NEDDylating EGFR and therefore displays a dual specificity as Hdm2 does for modification of p53 (Oved et al., 2006; Xirodimas et al., 2004). Modification of EGFR with NEDD8 leads to an increased turnover rate caused by intensified ubiquitination (Oved et al., 2006). In the case of steroid hormone receptors, NEDD8 was even found to be required for their ubiquitination and degradation. Therefore, inactivation of the NEDD8 pathway might be involved in the development of steroid hormone dependent tumors (Fan et al., 2003; Fan et al., 2002). Parkin, a RING‐type E3 ligase which is frequently mutated in patients suffering from Parkinson´s disease, reveals an enhanced activity upon NEDDylation. Substrates of parkin take part in diverse cellular functions like transcription, neurotransmission, synaptic function or cell cycle control (Choo et al., 2012; Walden and Martinez‐Torres, 2012). As a component of the Cullin2/elongin B/C complex, pVHL is involved in the regulation of oxygen‐dependent ubiquitination of HIFα (see 1.3.1.1). Interestingly, pVHL is also required for fibronectin matrix assembly, independent of Cullin2 (Stickle et al., 2004). Toggling the binding of pVHL to the cullin complex is achieved by its modification with NEDD8: NEDDylation leads to its interaction with fibronectin whereas its association with Cullin2 is inhibited (Russell and Ohh, 2008). IAPs (Inhibitor of Apoptosis) are often found to be overexpressed in cancer, thereby contributing to cell proliferation and survival. IAPs are RING ubiquitin E3 ligases that negatively regulate caspase activity and additionally have an influence on cellular survival functions. In Drosophila and humans, effector caspases were identified to act as substrates for NEDDylation by IAPs leading to their inactivation (Broemer et al., 2010). However, it was also supposed that IAPs themselves, rather than caspases, might be substrates for NEDD8 (Nagano et al., 2012). Very recently, both reduced levels of Ubc12 and inhibition of APPBP1/UBA3 were discovered to impair T‐cell proliferation and cytokine production. Thereupon, NEDDylation of Shc, an adapter

- 13 - Introduction protein between the antigen receptor of T‐cells and the Erk‐pathway, was identified as an important event in T‐cell receptor signaling (Jin et al., 2013). In proteomic analyses, many further potential substrates for NEDD8 were identified which are mainly involved in mRNA splicing, DNA replication and repair, chromatin remodeling and proteasomal degradation (Jones et al., 2008; Xirodimas et al., 2008). Finally, one well‐studied interaction partner of NEDD8 which targets NEDD8 and its conjugates for proteasomal degradation is NUB1 (NEDD8 Ultimate Buster1) (Kamitani et al., 2001). By interacting with the S5a subunit of the proteasome, NUB1 not only delivers NEDD8 but also the UBL FAT10 for degradation (Hipp et al., 2004; Tanji et al., 2005). Interaction of NEDD8 with NUB1 additionally results in inhibition of NEDDylation and enhances ubiquitination of p53, leading to its cytoplasmic localization (Liu and Xirodimas, 2010).

1.3.2 NEDD8‐conjugation cascade

The NEDD8‐conjugation cascade is very similar to the one of ubiquitin, involving three specialized types of enzymes which activate NEDD8 and transfer it to a substrate (Figure 5). In humans, NEDD8 is expressed as a precursor protein of 81 aa. The last five amino acids need to be cleaved off by the isopeptidase NEDP1/Den1/SENP8 in order to expose the C‐terminal double glycine motif which is crucial for NEDD8 attachment to substrates (Gan‐Erdene et al., 2003; Kamitani et al., 1997; Wu et al., 2003a). Alternatively, UCHL3, a deubiquitinating enzyme, is capable of processing both NEDD8 and ubiquitin (Wada et al., 1998). In a first step of the conjugation pathway, NEDD8 is activated by the heterodimeric E1 enzyme APPBP1/UBA3 with UBA3 comprising the catalytically active cysteine residue. Activation includes the formation of a NEDD8 adenylate under consumption of ATP, and the subsequent covalent attachment of NEDD8 C‐terminal glycine to the catalytic cysteine of UBA3, generating a high‐energy thioester bond (Gong and Yeh, 1999). As soon as one NEDD8 is bound in a thioester, a second NEDD8 interacts with the adenylation site of APPBP1/UBA3 and can be activated. Transfer of NEDD8 to the catalytic cysteine of one of the two conjugating enzymes Ubc12 or Nce2 is only accomplished if the E1 binds two NEDD8 proteins at the same time. In this so called transthiolation reaction, the E2 enzyme takes over NEDD8 from the E1 by the formation of a thioester linkage (Huang et al., 2009; Huang et al., 2007). Finally, an isopeptide bond is formed between the C‐terminal glycine of NEDD8 and a lysine residue of its substrate (Oved et al., 2006; Wada et al., 1999; Xirodimas et al., 2004). This last step involves the action of an E3 ligase catalyzing the transfer of NEDD8. In contrast to ubiquitin of which two classes of E3 ligases are known, RING ligases seem to be the major protein family playing a role in the NEDDylation cascade (see 1.1.1) (reviewed in (Ardley and Robinson, 2005; Watson et al., 2011)). This type of ligase brings the E2 enzyme and the substrate into proximity thereby allowing the transfer of

- 14 - Introduction

NEDD8 to its substrate. However, a direct interaction between E2 and the RING domain of an E3 has only been shown for Ubc12 and the E3 ligases RBX1 and RNF 111 so far (Calabrese et al., 2011; Ma et al., 2013). In contrast to NEDD8 E1 and E2 enzymes, which are highly specific for NEDD8, all RING E3 ligases involved in NEDDylation display a dual specificity for ubiquitin and NEDD8 (Broemer et al., 2010; Huang et al., 2008; Morimoto et al., 2003; Oved et al., 2006; Walden et al., 2003a; Xirodimas et al., 2004). Nevertheless, there is one unique E3 enzyme for NEDD8 without any known ligase domain: Dcn1/DCNL (yeast/humans), which acts as a scaffold‐ type E3 ligase enhancing NEDDylation of cullins (see 1.3.1.1) (Kurz et al., 2008; Monda et al., 2013).

Figure 5. NEDD8­conjugation cascade In a first step, NEDD8 is activated by its heterodimeric E1 enzyme APPBP1/UBA3 thereby forming a thioester bond. Subsequently, NEDD8 is transferred to the catalytic cysteine of one of its E2 enzymes, Ubc12 or Nce2. Finally, NEDD8 is conjugated to a lysine residue of a substrate which involves the action of a RING E3 ligase. - 15 - Introduction

DeNEDDylation of substrates is predominantly carried out by NEDP1 which cleaves NEDD8 both from cullins and other substrates (Chan et al., 2008; Mendoza et al., 2003). A specific deNEDDylating enzyme complex for cullins is the COP9 signalosome consisting of eight subunits including CSN5 as the catalytically active one (Cope et al., 2002; Lyapina et al., 2001). Moreover, USP21 is able to remove both NEDD8 and ubiquitin from substrates (Gong et al., 2000). NEDD8 contains nine internal lysine residues at positions 4, 6, 11, 22, 27, 33, 48, 54 and 60, five of which are conserved in ubiquitin. In vitro, mainly K22, K48 and K54 are used for NEDD8 chain formation. K27 and K33 were also found to be modified with NEDD8 although they occurred with a lower abundance (Jeram et al., 2010; Jones et al., 2008). However, an existence of NEDD8 chains under normal cellular conditions still needs to be verified as proteomic analyses were performed using tryptic digest prior to mass spectrometry. Although a double glycine motif was identified at K11, K22, K48 and K60 of NEDD8, it is not clear whether NEDD8 or ubiquitin is attached since they are not distinguishable in this approach (Jones et al., 2008). Additionally, in human cells and in vitro NEDD8 can be attached to ubiquitin but seems to be a bad acceptor for itself or ubiquitin, arguing against the existence of NEDD8 chains (Singh et al., 2012; Whitby et al., 1998).

1.3.2.1 APPBP1/UBA3, the NEDD8‐activating enzyme

In eukaryotes, E1 enzymes consist of three conserved domains: an adenylation domain which is already present in a common E1 ancestor, MoeB, and two eukaryotic specific domains (reviewed in (Hochstrasser, 2000)). One of those evolutionary newer domains contains the catalytic cysteine and the other one, which is located at the C terminus, is involved in transferring the UBL to the E2. This so called “ubiquitin fold domain” (UFD) adopts a ubiquitin‐like structure and plays a crucial role in the interaction with the E2. Comparison of the heterodimeric APPBP1/UBA3 with the ubiquitin E1 UBA1 reveals sequence and structural homology of APPBP1 to the N‐terminal part of UBA1 and of UBA3 to its C‐terminal part (Hochstrasser, 2000; Lake et al., 2001; Walden et al., 2003b). The overall structure of APPBP1/UBA3 can be described as a canyon with a groove in the middle. A crossover loop connects the adenylation domain with the catalytic cysteine domain and divides the canyon into two clefts (Walden et al., 2003b). In the structure of APPBP1/UBA3 bound to NEDD8 and ATP, the globular domain of NEDD8 occupies one cleft and its flexible C terminus points towards ATP (Figure 6). Furthermore, there is a bipartite interaction between APPBP1/UBA3 and NEDD8. On the one hand, the conserved acidic surface of NEDD8 contacts a part of the catalytic cysteine domain of APPBP1 forming a polar interface. On the other hand, the hydrophobic surface of NEDD8 interacts with a conserved part of the adenylation domain in UBA3. Interestingly, all residues involved in these hydrophobic interactions are also present in

- 16 - Introduction ubiquitin. Nonetheless, discrimination between NEDD8 and ubiquitin can be achieved by the C terminus of NEDD8, which binds to the adenylation domain and the crossover loop of UBA3. An alanine at position 72 of NEDD8 interacts with L206 and Y207 in the crossover loop. Because of repulsion from R190 in UBA3, the corresponding arginine at position 72 of ubiquitin prevents binding to the NEDD8 E1. R190 equates to a conserved glutamine residue in UBA1 which allows binding of ubiquitin but not of NEDD8. Along these lines, an A72R mutant of NEDD8 can be activated by the ubiquitin E1 enzyme and vice versa (Walden et al., 2003a; Whitby et al., 1998).

APPBP1

UBA3

Figure 6. Structure of the APPBP1/UBA3­NEDD8­ATP complex The overall APPBP1/UBA3 structure (blue/red) can be divided into an adenylation domain and a catalytic cysteine domain (catalytic cysteine in green). NEDD8 (yellow) occupies one cleft in the major groove of APPBP1/UBA3 while its C terminus points towards the ATP‐binding site. A crossover loop connecting adenylation and catalytic cysteine domain fastens NEDD8 in this position (modified from (Walden et al., 2003a)).

The structure of the complex between APPBP1/UBA3, NEDD8 and ATP also shows that ATP interacts with a conserved ATP‐binding motif in the adenylation domain of APPBP1/UBA3 next to the C terminus of NEDD8, thereby enabling an attack of NEDD8 on the α‐phosphate of ATP (Figure 6). Two sites of interaction between the E1 and ATP were identified: the adenine ring binds to a hydrophobic patch in UBA3 and the phosphate groups form hydrogen bonds and salt bridges with several residues of both APPBP1 and UBA3. Mg2+, which is necessary for the neutralization of the negative charge of phosphate within ATP, is probably coordinated by D146 in UBA3 (Walden et al., 2003a). To facilitate binding and activation of NEDD8, conformational changes of APPBP1/UBA3 and NEDD8 need to occur. By moving the catalytic cysteine domain away from the adenylation domain, NEDD8 is able to bind the E1 and subsequently fastened by the crossover loop. In addition, rotation of the C terminus of NEDD8 allows docking into the ATP binding site of UBA3

- 17 - Introduction

(Walden et al., 2003a). In a next step, NEDD8 is attacked by the catalytic cysteine at position 216 in UBA3, leading to the formation of a thioester linkage. However, because of a large gap between the C terminus of NEDD8 and C216 as shown in figure 6, an additional rotation of NEDD8 and/or conformational changes in APPBP1/UBA3 might be required for the formation of this linkage (Walden et al., 2003a). In 2009, the first inhibitor of APPBP1/UBA3 was described as a potential anti‐cancer drug (Soucy et al., 2009). MLN4924 was shown to specifically inhibit the NEDD8‐activating enzyme by the formation of a covalent NEDD8‐MLN4924 adduct which mimics the NEDD8‐adenylate. The mechanism of action of MLN4924 was proposed as follows: NEDD8 is first activated and transferred to the catalytic cysteine of APPBP1/UBA3 forming a thioester bond. In a next step, MLN4924 interacts with the adenylation site of the E1 enzyme and is finally covalently attached to the C terminus of NEDD8. Because of a stable binding to APPBP1/UBA3, this adduct blocks its activity but is not further used in the NEDDylation cascade (Brownell et al., 2010). At the moment, MLN4924 is in clinical trials for the therapy of several types of cancer such as acute myeloid leukemia and lymphoma (Milhollen et al., 2010; Swords et al., 2010).

1.3.2.2 NEDD8‐conjugating enzymes

Conjugating enzymes of the ubiquitin and UBL pathways are generally characterized by a catalytic core domain of 150‐200 aa, which in some cases is extended at the N‐ or the C terminus. As the evolutionary highest conserved region among E2 enzymes, the N‐terminal α‐helix of the core domain was found to mediate interaction with the E1 (Huang et al., 2005; Winn et al., 2004). In contrast to ubiquitin, which holds about 40 conjugating enzymes in mammals, there are only two conjugating enzymes known for NEDD8: Ubc12 and Nce2 (Huang et al., 2009; Ye and Rape, 2009). Interestingly, Nce2 is only found in higher eukaryotes, whereas Ubc12 is conserved from yeast to human. Ubc12 and Nce2 are ~40 % identical at amino acid level. On the structural level, both NEDD8‐conjugating enzymes are also highly similar with the exception of a loop insertion next to the catalytic cysteine of Nce2 consisting of six amino acids (aa 124‐129) (Figure 7). Comparing substrate specificity of the NEDD8 E2s, it was found that Ubc12 rather NEDDylates Cullins1‐4 using RBX1 as a ligase, whereas Nce2 prefers RBX2 as ligase and Cullin5 as substrate (Huang et al., 2009). Like Nce2, Cullin5 and RBX2 are only present in metazoans being involved in cytokine signaling, neuronal migration or myogenesis (Feng et al., 2007; Huang et al., 2009; Kile et al., 2002; Nastasi et al., 2004). For most other NEDD8 substrates, it is not clear which E2 catalyzes NEDD8 transfer. However, Ubc12 seems to be predominantly used since its knockdown leads to severe effects like G2/M‐arrest, aneuploidy and apoptosis. In contrast, cells

- 18 - Introduction comprising a stable knockdown of Nce2 expression reveal a normal phenotype. Nonetheless, Nce2 overexpression partially rescues knockdown effects of Ubc12 (Huang et al., 2009).

A B

C C N

N

Figure 7. Crystal structures of the core domains of Ubc12 (A) and Nce2 (B) Ribbon representations of NEDD8‐conjugating enzymes Ubc12 (A; protein data base 1Y8X; modeled with Pymol) and Nce2 (B; protein data base 2EDI; modeled using Pymol) show a high structural similarity of the enzymes. The catalytic cysteine is displayed as magenta sticks. Ubc12 and Nce2 only differ in the region next to the catalytic cysteine where a loop insertion is present in Nce2 (arrow).

Both NEDD8‐conjugating enzymes contain an N‐terminal extension being unique to the NEDD8 cascade (Huang et al., 2009; Huang et al., 2004). The first 13 residues of the N‐terminal extension of Ubc12 dock in a groove of the adenylation domain of UBA3 under the formation of hydrogen bonds and hydrophobic contacts, mainly mediated by F5 and L7 (Figure 8A). As expected, the hydrophobic motif ‐‐X‐ (‐hydrophobic aa; X‐any aa) in the N terminus of Ubc12, which plays a major role in the interaction with APPBP1/UBA3, is also present in the N‐terminal extension of Nce2 (Huang et al., 2009; Huang et al., 2004). Moreover, the binding groove within APPBP1/UBA3 is conserved through different species but not in the E1 enzymes of other UBLs. Thus, the N termini of Ubc12 and Nce2 bind directly and selectively to APPBP1/UBA3. Accordingly, deletion of the N‐terminal 26 aa of Ubc12 decreases the affinity to APPBP1/UBA3 around 25‐fold. Nonetheless, thioester formation with NEDD8 and transfer of NEDD8 to substrates can still proceed (Huang et al., 2004). An additional interaction takes place between the core domain of Ubc12, which is conserved in all ubiquitin and UBL conjugating enzymes, and the ubiquitin fold domain (UFD) of UBA3 forming two sets of mainly hydrophobic contacts (Figure 8A). Binding occurs at the opposite site of the catalytic cysteine of Ubc12 so that a rotation is required to facilitate the transesterification reaction (Huang et al., 2005). Structure determination of Ubc12 C111A bound to NEDD8‐loaded APPBP1/UBA3 revealed that this rotation is performed by the UFD domain, leading to a new surface next to the ATP‐binding site that in turn serves as an interaction site for Ubc12 (Figure

- 19 - Introduction

8B). Furthermore, rotation of the UFD allows the thioester‐bound NEDD8 to occupy the central groove of the E1, thereby providing a further binding site for Ubc12. Such an interaction seems to be conserved among ubiquitin and UBLs since thioester‐bound ubiquitin and SUMO also interact with their cognate E2 (Huang et al., 2007; Miura et al., 1999; Reverter and Lima, 2005).

A B

Figure 8. Structures of APPBP1/UBA3 in complex with Ubc12 (A) or NEDD8 and Ubc12 (B) (A) Superposition of Ubc12 core (cyan) ‐ UFD (red) complex and APPBP1/UBA3 (blue/pink) in complex with the N‐ terminal 13 aa of Ubc12 (cyan) (Huang et al., 2005). The catalytic cysteine of Ubc12 points away from the catalytic cysteine of UBA3. (B) Structure of the APPBP1/UBA3 (blue/pink and red)–NEDD8 (thioester‐bound; yellow)‐NEDD8 (bound to adenylation site; lime)‐MgATP‐Ubc12 C111A (cyan) complex. To enable the transesterification reaction, UFD rotates around 120° thereby bringing the catalytic cysteines of Ubc12 and UBA3 into close proximity (Huang et al., 2007).

Once NEDD8 is transferred to the catalytic cysteine of the E2, the interaction between E1 and E2 is disrupted as flipping back of the UFD would clash with the E2. Hence, conformational changes of APPBP1/UBA3 are important to ensure the progression of the NEDDylation cascade (Huang et al., 2007). Correspondingly, as already shown for the ubiquitin system, a simultaneous binding of the E1 and the E3 to the E2 enzyme is not possible since the interaction sites are overlapping (Wenzel et al., 2011). In addition, the E3 shows a higher affinity to the thioester‐loaded E2 than to the E2 alone (Kawakami et al., 2001; Siepmann et al., 2003). An identity of around 60 % between NEDD8 and ubiquitin raises the question as to how the enzymes involved in their conjugation distinguish which one to use. It was found that both the E1 and the E2 enzyme contribute to the specificity of the particular conjugation pathway. On the one hand, R190 of UBA3 strongly repulses the basic side chain of R72 within ubiquitin whereas A72 of NEDD8 allows binding. In addition, several other residues within APPBP1/UBA3 are involved in preventing binding of ubiquitin (Souphron et al., 2008). On the other hand, E2 enzymes also confer specificity by a unique interaction with the respective E1 enzyme. In case of the NEDD8 E2s, there is a bipartite binding to APPBP1/UBA3: their N terminus docks in a

- 20 - Introduction specific binding site within the E1 and specific residues within their core domain bind to the UFD of UBA3. Moreover, K147 and E151 of Ubc12 point towards the catalytic cysteine of UBA3 where unique insertions are found in the ubiquitin and SUMO E1 enzymes conferring specificity to the respective conjugation cascade (Huang et al., 2008; Lois and Lima, 2005; Szczepanowski et al., 2005; Wang et al., 2007a).

1.4 Aim of the studies

NEDD8 (Neural precursor cell‐Expressed Developmentally Down‐regulated 8) was originally identified as a developmentally down‐regulated ubiquitin‐like protein. Moreover, it was found to be involved in cell cycle regulation by being covalently conjugated to cullins. In the meantime, several additional substrates for NEDD8 have been identified being involved in various cellular functions like e.g. transcriptional regulation, cell growth or apoptosis. Similar to ubiquitin, covalent attachment of NEDD8 to substrates involves three consecutive steps catalyzed by activating enzymes E1, conjugating enzymes E2, and ligases E3. For NEDD8, one E1 enzyme (APPBP1/UBA3 heterodimer), two E2 enzymes (Ubc12 and Nce2), and several substrate specific E3 enzymes have been identified so far. However, only little is known about if and how these enzymes are regulated. Furthermore, the specificity of Ubc12 and Nce2 for most of the NEDD8 substrates remains elusive. Especially, functions of Nce2 other than the NEDDylation of Cullin5 have not been identified yet. Recent proteomics data reveal a variety of further possible substrates for NEDDylation, playing roles in mRNA splicing, DNA replication and repair, chromatin remodeling and proteasomal degradation. Thus, it seems that there are many yet unknown and uncharacterized substrates for the NEDDylation pathway in the cell.

This study therefore aimed at:

1. Identifying new substrates and interaction partners of NEDD8 2. Investigating the regulation of the NEDD8‐conjugation cascade 3. Characterizing a second isoform of Nce2

In conclusion, this work shall not only contribute to a better understanding of physiological functions of NEDD8 but also of the NEDDylation pathway in general. Currently, this field is of special interest since MLN4924, an inhibitor of the NEDD8‐activating enzyme, is tested in clinical trials for the treatment of several types of cancer such as acute myeloid leukemia and lymphoma.

- 21 -

2. Material and Methods

2.1 Material 2.1.1 Solutions and media

Name Composition Laemmli loading buffer (2x) 125 mM Tris‐HCl pH 6.8, 200 mM DTT, 4 % SDS, 0.001 % Bromphenol‐blue

10x Ergänzungspuffer 7.15 M β‐Mercaptoethanol, 40 % (v/v) Glycerol, 100 mM Tris‐HCl, Orange G, pH 7.5

Urea loading buffer (2x) 6 M Urea, 4 % (v/v) SDS, 125 mM Tris‐HCl pH 6.8, 20 % (v/v) Glycerol

Laemmli running buffer (10x) 250 mM Tris‐HCl, 2 M Glycin, 1 % SDS, pH 8.4

Stacking gel buffer 0.5 M Tris‐HCl pH 6.8, 0.4 % SDS Separating gel buffer 1.5 M Tris‐HCl pH 8.8, 0.4 % SDS

Transfer buffer (20x) 12.5 mM Tris‐HCl, 100 mM Glycine, pH 8.3

TNE ‐T 10 mM Tris‐HCl, 2.5 mM EDTA, 50 mM NaCl, 0.1 % Tween 20, pH 7.5

Stripping buffer for western blots 62.5 mM Tris‐HCl pH 6.8, 2 % SDS, 100 mM β‐Mercaptoethanol

Coomassie Blue staining solution 2 g/L Coomassie Brilliant Blue R250 in Coomassie Destain Solution

SDS gel fixation solution / Destain solution 40 % Methanol, 10 % Acetic Acid

DNA loading buffer (10x) 60 % Saccharose, 0.25 M EDTA, Bromphenol‐blue

- 22 - Material and Methods

TAE buffer (50x) 2 M Tris‐HCl, 950 mM Acetic Acid, 50 mM EDTA

Affinity pulldown buffer 20 mM TEA, 50 mM NaCl, 1 mM EDTA, 0,5 % NP40, 10 % glycerol, pH 7.4

TNN lysis buffer 100 mM Tris‐HCl, 100 mM NaCl, 1 % NP‐40, 1 mM Pefabloc, 1 μg/mL Aprotinin/Leupeptin, 1 mM DTT, pH 8.0

RIPA lysis buffer 25 mM Tris‐HCl, 50 mM NaCl, 0.5 % NP40, 0.5% Deoxycholate, 0.1 % SDS, 1 mM Pefabloc, 1 μg/mL Aprotinin/Leupeptin, pH 8.8

Guanidinium lysis buffer 100 mM Na2HP04/NaH2PO4, 6 M Guanidinium‐Hydrochloride, 10 mM Imidazole, 10 mM β‐Mercaptoethanol, pH 8.0

HP buffer (cellular fractionation) 10 mM HEPES‐NaOH, 20 mM NaCl, 10 mM MgCl2, 0.5 mM ATP, pH 7.4

2x KE buffer (cellular fractionation) 4 g Saccharose, 40 mM HEPES‐NaOH, 1 mM MgCl2, 1 mM ATP, pH 7.4

Buffer Z (β‐gal assay) 100 mM NaH2PO4, 10 mM KCl, 1 mM MgS04, 50 mM β−Μercaptoethanol, pH 7.0

ONPG 4 mg/mL in 100mM Na2HPO4, pH 7.0

Luria Broth medium (LB) 10 g/L NaCl, 5g/L yeast extract 10g/L Bacto‐Tryptone, pH 7.5 (complemented with 100 μg/mL ampicillin, 25 μg/mL kanamycin or 170 μg/mL chloramphenicol final concentrations)

SOC medium 20 g/L Tryptone, 5 g/L yeast Extract, 0.5 g/L NaCl, 20 mM Glucose, pH 7

NMM medium 7.5 mM (NH4)2 SO4, 8.5 mM NaCl, 22 mM KH2PO4, 50 mM K2HPO4, 1 mM MgSO4, 1 mg/L CaCl2, 1 mg/L FeCl2, 1 μg/L CuCl2, 1 μg/L MnCl2, 1μg/L ZnCl2, 20 mM Glucose, 10 mg/L thiamine hydrochloride, 10 mg/L d‐biotin, 50 mg/L of all amino acids except methionine, 100 mg/mL ampicillin, pH 7.0 (+0.05 mM methionine or 0.5 mM azidohomoalanine)

S1 50 mM Tris‐HCl, 10 mM EDTA, 100 μg/mL RNaseA, pH 8.0

S2 200 mM NaOH, 1 % SDS

- 23 - Material and Methods

S3 2.8 M KAc, pH 5.1

T25N50 25 mM Tris‐HCl, 50m M NaCl, pH 8.0

Washing buffer (antibody purification) 20 mM Tris‐HCl, 500mM NaCl, 0.2 % Triton‐X‐ 100, pH 7.5

Elution buffer (antibody purification) 150 mM NaCl, 200mM Glycin, pH 2.3

2.1.2 Chemicals and Reagents

Name Company Amplifyer GE Healthcare Aprotinin/Leupeptin SIGMA Ammonium persulfate ROTH ATP SIGMA β‐Mercaptoethanol Merck BIO‐RAD protein assay BIORAD Cycloheximide SIGMA DAPI SIGMA DTT ROTH EDTA ROTH Ethidiumbromide ROTH Glutathione Sepharose 4B GE Healthcare Glycine ROTH Guanidinium Hydrochloride ROTH HA‐beads SIGMA (monoclonal anti‐HA agarose conjugate) HEPES (1M) Gibco Imidazole USB IPTG ROTH Lipofectamine 2000 Invitrogen MG132 SIGMA

MgCl2 ACROS organics Milk powder ROTH MLN4924 Active Biochem

Na2HPO4 SIGMA

NaH2PO4 Merck

- 24 - Material and Methods

Ni‐NTA‐Agarose Qiagen NP‐40 MP Biomedicals NP‐40 for cellular fractionation Fluka ONPG SIGMA PBS Gibco PefaBloc Boehringer Ingelheim Protein A‐Sepharose GE Healthcare Q Sepharose Fast Flow GE Healthcare 5x Roti Blue ROTH Rotiphorese Gel 30 ROTH SDS ROTH Sepharose CL‐4B GE Healthcare SulfoLink Coupling Gel Pierce TEMED ROTH Triton‐X‐100 ROTH Trizma Base (Tris) SIGMA TRIZOL Reagent Invitrogen Turbofect Fermentas Tween‐20 ROTH Western Lightning ECL Perkin Elmer

2.1.3 Bacterial strains

E. coli DH5α: F‐ endA1 glnV44 thi‐1 recA1 relA1 gyrA96 deoR nupG Φ80dlacZΔM15 Δ(lacZYA­argF)U169, ‐ + hsdR17(rK mK ), λ–

E. coli BL21‐CodonPlus‐RIL competent cells: – – – + r r B F ompT hsdS(rB mB ) dcm Tet gal endA Hte [argU ileY leuW Cam ] (Stratagene)

E. coli BL21 (DE3): – ‐ ‐ F ompT gal dcm lon hsdSB(rB mB ) λ(DE3 [lacI lacUV5‐T7 gene 1 ind1 sam7 nin5])

E. coli B834 (DE3): ‐ ‐ ‐ F ompT hsdSB(rB mB ) gal dcm met (DE3)

E. coli XL10‐Gold: Tetr D(mcrA)183 D(mcrCB­hsdSMR­mrr)173 endA1 supE44 thi­1 recA1 gyrA96 relA1 lac Hte [F¢ proAB lacI qZDM15 Tn10 (Tetr) Amy Camr] (Stratagene)

- 25 - Material and Methods

2.1.4 Mammalian cell lines

H1299: non‐small cell lung carcinoma, p53‐/‐ ; HEK293T: human embryonal kidney cells; Cells were cultured in DMEM (Gibco) containing 10 % FCS (Gibco) and 50 µg/mL Normocin (Invivogen) or Penicillin/Streptomycin (1:100, Gibco) on CellStar plates (Greiner).

2.1.5 Antibodies

Primary antibodies protein name species & type company dilution HA‐tag α HA 1.1 mouse, Covance 1:2500 monoclonal 6x His‐tag α His mouse, Qiagen 1:1000 monoclonal Myc‐tag α c‐Myc 9E10 rabbit, Abcam 1:1000 monoclonal Nce2 α Nce2 C‐ rabbit, Abgent 1:100 (IF) isoform 1 term polyclonal Nce2 α Nce2 rabbit, University of 1:200 (WB) isoform 2 isoform 2 polyclonal Konstanz 1:50 (IF) NEDD8 α NEDD8 rabbit, Alexis 1:1000 polyclonal NEDD8 α NEDD8 rabbit, Epitomics 1:1000 monoclonal PCNA α PC10 mouse, Abcam 1:5000 monoclonal

Secondary Antibodies name species company dilution HRP‐coupled α mouse goat Dianova 1:20000 HRP‐coupled α rabbit goat Dianova 1:20000 AlexaFluor568 α mouse donkey Invitrogen 1:1000 AlexaFluor488 α mouse goat Invitrogen 1:1000

- 26 - Material and Methods

2.1.6 Primers

Name Sequence AR28 ggggatccttacccacctctgagacggag AR31 ggcagggatccttatcctcctctaagagccaacaccaggtg AR64 gcggtcgacttatttcaggcagcgctc AR84 gtcggatccatgctaacgctagcaagtaaactg AR89 gtccatatgctaacgctagcaagtaaactg AR95 cacctcgagtcatctggcataacgtttgat AR109 cccggatcctcagtgatggtgatggtgatgtctggcataacgtttgatgtagt AR111 gacaggggaaatatCtctgagtttattgag AR112 ctcaataaactcagagatatttcccctgtc AR113 gacaggggaaataGCtctgagtttattgag AR114 ctcaataaactcagagctatttcccctgtc AR121 gtcggatccatgactcggagggtttctgtgagag AR130 GagatatacatATGcatcaccatcaccatcacatgctaattaaagtgaagacg AR131 GagatatacatATGcatcaccatcaccatcacATGCAGATCTTCGTCAAGACC AR190 ggcGGATCCatgctaattaaagtgaagacg DP7 CAAGATCTGGCACCCCGCCATCACAGAGACAGG DP8 CCTGTCTCTGTGATGGCGGGGTGCCAGATCTTG DP9 CAATGGTCTATCACCCCGCCATTGACCTCGAGGGC DP10 GCCCTCGAGGTCAATGGCGGGGTGATAGACCATTG DP24 TAACGCTAGCAAGTAGACTGAGGCGTGACGATGGTCT DP25 AGACCATCGTCACGCCTCAGTCTACTTGCTAGCGTTA DP26 ATGATCAGGCTGTTCTCGCTGAGGCAGCAGAGGAAGGAGGAGGAGTC DP27 GACTCCTCCTCCTTCCTCTGCTGCCTCAGCGAGAACAGCCTGATCAT DP28 GGCAGATCTATGCTAATTAAAGTGAAGACGC DP31 TTGGCTCTGAGAGGAGTAATGATCAAGCTGTTCTCG DP32 CGAGAACAGCTTGATCATTACTCCTCTCAGAGCCAA DP46 GTCGGATCCATGCTAACGCTAGCAAGTAG DP59 CTTTTGCACTGAGGTTCCTGAACTTCTTTAC DP60 GTAAAGAAGTTCAGGAACCTCAGTGCAAAAG DP68 GACTTGTCCCGCAAATGATG DP69 CATCATTTGCGGGACAAGTCCCCAATGCTGTTACTCCACAGGAG DP70 CATCCACTTTATTCCGGAAGTCCTCCTGTGGAGTAACAGCATT DP71 GACTTCCGGAATAAAGTGGATGACTACATCAAACGTTATGCCAG

- 27 - Material and Methods

Name Sequence DP72 CTGCAATCGTCCCCTTTTATTATCTGGCATAACGTTTGATGTAGT DP73 ATAATAAAAGGGGACGATTGCAGGCCCATGGACTGTGTTACA DP74 AACTCGAGTCATGTTAGAGACAAACTGTAACACAGTCCATGGGC DP75 CCGGATCCTCAGTGATGGTGATGGTGATGTGTTAGAGACAAACTGTAACAC DP76 GGCCTCGAGTTATCTTAGTCTTAAGACAAG DP77 CTGTGGAGTAACAGCATTGG

2.1.7 Plasmids constructed and used in this study

Name Insert Vector Primers Dana1 His‐NEDD8 pET3a AR130/AR31 Dana4 His‐Ubiquitin pET3a AR131/AR28 Dana7 Nce2 wt‐His pET3a AR89/AR109 Dana19 His‐Nce2 N108A pET3a Mutagenesis (DP7/DP8) Dana20 His‐Ubc12 N103A pET3a Mutagenesis (DP9/DP10) Dana45 Ubc12 K3R/K8R/K11R pcDNA 3HA Mutagenesis (DP 26/DP27) Dana46 Ubc12 K3R/K8R/K11R‐ pET3a Mutagenesis His (DP 26/DP27) Dana48 Nce2 K7R/K9R pcDNA 3HA DP46/AR95 (from Dana52) Dana49 Nce2 K7R/K9R‐His pET3a Mutagenesis (DP24/DP25) Dana51 Fusion Nedd8 G76V ‐ pcDNA 3HA DP31/DP32 Ubc12 DP28/AR64 Dana54 Fusion Nedd8 G76V ‐ Nce2 pcDNA 3HA AR190/AR95 (from Dana‐58) Dana77 His‐PCNA Y211F pcDNA Mutagenesis 3.1/Genestorm DP59/DP60 Dana95 His‐Nedd8 M50A G76M pGDR11 subcloned from pMA‐T (for click reaction) (Geneart) EcoRI/BamHI Dana97 HA‐ Nce2 isoform 2 pcDNA3HA subcloned from Dana99 (pGEX2TK‐ Nce2 isoform2) BamHI/XhoI Dana98 Nce2 isoform 2‐His pET3a AR89/DP75 Dana104 ΔN26 Nce2 isoform2 pcDNA3HA AR121/DP74 Dana105 Nce2 isoform2 C116S pcDNA3HA Mutagenesis AR111/AR112 Dana106 Nce2 isoform2 C116A pcDNA3HA Mutagenesis AR113/AR114 Dana107 Nce2 isoform 2 C116S‐His pET3a Mutagenesis AR111/AR112 Dana108 Nce2 isoform 2 C111A‐His pET3a Mutagenesis AR113/AR114

- 28 - Material and Methods

2.1.8 Other plasmids used in this study

Name Insert Vector Reference ARF41 NEDD8 pcDNA3HA M. Scheffner ARF56 Ubc12 ΔN26‐His pET3a M. Scheffner ARF123 Nce2 wt pcDNA3HA M. Scheffner ARF129 Nce2 C116A pcDNA3HA M. Scheffner ARF134 Ubc12 wt pcDNA3HA M. Scheffner ARF135 Ubc12 C111S pcDNA3HA M. Scheffner ARF136 Ubc12 C111A pcDNA3HA M. Scheffner ARF137 Ubc12 ΔN26 pcDNA3HA M. Scheffner ARF139 Nce2 C116S pcDNA3HA M. Scheffner ARF140 Nce2 ΔN26‐His pET3a M. Scheffner ARF144 Ubc12‐His pET3a M. Scheffner ARF145 UbcH5b‐His pET3a M. Scheffner ARF162 Nce2 ΔN26 pcDNA3HA M. Scheffner ARF175 Nce2 pcDNA4TOmycHisB M. Scheffner ARF181 Ubc12 pcDNA4TOmycHisB M. Scheffner β‐gal β‐galactosidase pRcCMV M. Scheffner Click NEDD8 His‐NEDD8 M50A pMA‐T ordered from Geneart G76M (Invitrogen) Click Ubiquitin His‐Ubiquitin pGDR11 S. Eger (AG Marx, G76M University of Konstanz) GST ‐ pGEX2TK GE Healthcare GST‐NEDD8 NEDD8 pGEX2TK M. Scheffner GST‐Ubiquitin Ubiquitin pGEX2TK M. Scheffner HA‐Ubiquitin Ubiquitin pcDNA3HA M. Scheffner His‐Nedp1 Nedp1 pcDNA3.1/V5‐His R. Hay (SCILLS, University of Dundee) His‐SUMO1 His‐SUMO1 pSG5.0‐Spl M. Scheffner His‐Ub His‐myc‐ubiquitin pcDNA3.1 M. Scheffner Myc‐Rad18 Myc‐Rad18 pCAGGS Satoshi Tateishi (human) (Kumamoto University, Japan) N31,N32 His‐NEDD8 pSG5.0‐Spl M. Scheffner

PCNA for click reaction (164TAG)PCNA pET11a M. Rubini (AG Marx, (yeast), University of tRNA (M. barkeri) Konstanz) PCNA‐SV5‐His wt PCNA‐SV5‐HispcDNA3.1 Genestorm S. Jentsch (human) (MPI,Martinsried) PCNA K164R‐SV5‐His PCNA K164R‐ pcDNA3.1 Genestorm S. Jentsch SV5‐His (human) (MPI,Martinsried) pGSTHsAPPBP1rbsUBA3 GST‐ pGST (Huang and Schulman, APPBP1/UBA3 2005) PylS Pyrrolysine pRSFDuet M. Rubini (AG Marx, Synthetase (Kanamycin) University of Konstanz) pRcCMV ‐ pRcCMV Invitrogen

- 29 - Material and Methods

2.1.9 DNA‐ and protein markers

‐GeneRulerTM 1kb Plus DNA Ladder (MBI Fermentas): 20000, 10000, 7000, 5000, 4000, 3000, 2000, 1500, 1000, 700, 400, 300, 200, 75 [bp] ‐ PageRulerTM Prestained Protein Ladder (MBI Fermentas): 170, 130, 95, 72, 55, 43, 34, 26, 17, 10 [kDa] ‐ PageRulerTM Unstained Protein Ladder (MBI Fermentas): 200, 150, 120, 100, 85, 70, 60, 50, 40, 30, 25, 20, 15, 10 [kDa]

2.2 Methods 2.2.1 PCR and cloning

2.2.1.1 Polymerase chain reaction (PCR) For cloning, Phusion Hot Start II High Fidelity DNA Polymerase (Finnzymes) was used according to manufacturer´s instructions. For colony PCR, Taq polymerase (AG Scheffner) and Thermo Pol 10x buffer (New England Biolabs) were employed. 10 µM primers and 0.2 mM of all dNTPs were used for each reaction.

2.2.1.2 Gene synthesis To synthesize cDNA of the specific part of Nce2 Isoform2, 3 µl of primers DP69‐DP74 (10 µM each) were pooled and 1 µl of the oligo pool was used for touch down PCR (7 cycles starting from 60 °C annealing temperature decreasing to 57.9 °C; 3 cycles with 57.9 °C) with Phusion polymerase (Finnzymes). For gene synthesis PCR, 1 µl of the assembly PCR was applied as template and DP69 and DP74 as primers (27 cycles touch down PCR starting from 60 °C annealing temperature and decreasing in 0.3 °C steps each cycle; 3 cycles with 52 °C annealing temperature). In a PCR using AR84 and DP74, the obtained product was fused to the unspecific part of Nce2 which was amplified before with primers AR84 and DP68.

2.2.1.3 Site directed mutagenesis Point mutations were introduced using QuickChange Site‐Directed Mutagenesis (Stratagene) according to manufacturer´s instructions. Complementary primers containing mutations were used for performing PCR. 10 µl of PCR product were digested with DpnI (NEB) and one fifth of the digest was transformed into supercompetent XL10 gold E. coli cells.

- 30 - Material and Methods

2.2.1.4 Restriction digest All restriction enzymes were obtained from NEB and incubated with DNA in the recommended buffer in a volume of 50 µl for 2 h at 37 °C. Vectors cut with only one enzyme were additionally treated for 1 h with 1 µl Antarctic Phosphatase (NEB) in the appropriate buffer. Before ligation, DNA was purified using NucleoSpin Extract II kit (Clontech).

2.2.1.5 Agarose gel electrophoresis To prepare 1 % or 2 % (w/v) agarose gels, agarose was boiled in TAE buffer. Subsequently, ethidium bromide (ROTH) was added to a final concentration of 0.5 µg/mL. Gels were submerged in TAE buffer in a horizontal gel electrophoresis chamber and run at 3 V/cm. Agarose gels were analyzed with a UV transilluminator and photographed using the LAS‐3000 imaging system (Fujifilm).

2.2.1.6 Purification of DNA from agarose gels After gel electrophoresis, desired DNA fragments were excised from the agarose gel under a UV transilluminator and purified using NucleoSpin Extract II kit (Clontech).

2.2.1.7 Ligation Purified vector and insert were ligated using T4 ligase and either 5x rapid ligation buffer or 10x ligation buffer (Fermentas). The whole reaction was transformed into supercompetent E. coli XL10 gold cells.

2.2.1.8 Transformation of DNA into chemical competent E. coli Plasmid DNA (whole ligation reaction or 100 ng of purified DNA) was mixed with 100 µl chemical competent bacteria and incubated on ice for 30 min. After a heat shock for 45 sec at 42 °C, cells were cooled down for 10 min on ice. Transformants containing Ampicillin resistance were directly plated on LB agar plates. Transformants carrying Chloramphenicol, Kanamycin or Zeocin resistance were incubated for 1 h in SOC medium at 37 °C to allow expression of the resistance genes. After plating, selection plates were incubated over night at 37 °C.

2.2.1.9 Preparation of DNA in low and high scale Small scale plasmid purification was performed using alkaline lysis (Birnboim and Doly, 1979). Preparation of higher amounts of DNA was carried out using 100 mL overnight culture and PureYield Plasmid Midiprep System (Promega).

- 31 - Material and Methods

2.2.1.10 Measurement of DNA and RNA concentrations DNA and RNA concentration was measured using a nanophotometer (NorthStar Scientific) according to manufacturer’s instructions.

2.2.1.11 DNA sequencing All sequencing reactions were performed by GATC (Konstanz/Köln).

2.2.2 Maintenance of bacterial cultures and mammalian cell lines

2.2.2.1 Bacterial cultivation and preparation of glycerol stocks Glycerol stocks of BL21 RIL bacteria containing plasmids for protein expression were prepared by mixing 600 µl sterile glycerol and 1.4 mL overnight culture. Cryovials containing glycerol stocks were stored at ‐80 °C.

2.2.2.2 Maintenance of mammalian cell lines

Mammalian cells were kept in the Heraeus CO2 incubator BBD 6220 (Thermo Fisher Scientific) at 37 °C, 95 % humidity and 5 % CO2. All cell lines were cultured with DMEM (Gibco) supplemented with 10 % FCS (v/v) (Gibco) and 50 µg/mL Normocin (Invivogen) or Penicillin/Streptomycin (1:100, Gibco). To split cells, 90 % confluent dishes were washed with PBS and trypsinized with 0.05 % trypsin‐ EDTA (Gibco). After a short incubation at 37 °C, detached cells were collected with DMEM containing 10 % FCS to stop the trypsin reaction, and transferred into a Falcon tube. Cells were spun down at 1000 rpm for 1 min. The obtained cell pellet was resuspended and distributed in new dishes according to requirements.

2.2.2.3 Freezing of cells in liquid nitrogen 90 % confluent cells were trypsinized and collected by centrifugation. The cell pellet was resuspended in 1 mL FCS containing 10 % DMSO (v/v) and transferred into a cryovial. The vial was kept in the freezing container Cryo 1°C (Nalgene) at ‐80 °C overnight and finally stored in liquid nitrogen.

2.2.3. Protein expression and ‐purification

2.2.3.1 Expression and purification of GST­fusion proteins in E. coli A distinct volume of LB medium containing 100 mg/mL ampicillin and 170 µg/mL chloramphenicol was inoculated with bacteria from a LB agar plate or a glycerol stock and kept overnight at 37 °C and 200 rpm. The overnight culture was diluted to OD600nm=0.1 the next

- 32 - Material and Methods

morning. After growing to an OD600nm=0.6‐0.8, protein expression was induced by the addition of IPTG to a final concentration of 0.4 mM. The bacterial culture was incubated at 37 °C and 200 rpm for 4‐6 h and then centrifuged at 5000 rpm for 15 min. The obtained pellet was resuspended with PBS/1 % Triton‐X‐100 (v/v). Bacterial lysate was then sonicated 2x with 40 pulses (duty cycle: 30‐40 %, output control: 3‐4) using BRANSON sonifier 250 and centrifuged at 13000 rpm for 15 min. Supernatant was transferred into a fresh tube and incubated with 200 µl Glutathione Sepharose 4B (GE Healthcare) in PBS/1 % Triton‐X‐100 per 200 mL diluted overnight culture for 2 h at 4 °C. Beads were washed three times with 5 mL PBS/1 % Triton‐X‐ 100 and once with T25N50. Proteins were eluted from the beads using 10 mM Glutathione in 50 mM Tris‐HCl pH 8 or used bound to beads for GST‐pulldown (2.2.5.2). GST‐APPBP1/UBA3 was prepared according to (Huang and Schulman, 2005).

2.2.3.2 Expression and purification of His­tagged proteins in E. coli Bacterial lysate containing the desired protein was prepared as described in 2.2.3.1. Lysate was incubated with 10 mM Imidazole and 100 µl Ni‐NTA‐Agarose (Qiagen) in PBS/1 % Triton‐X‐100 per 200 mL diluted overnight culture for 4 h or overnight at 4 °C. Beads were washed three times with 5 mL PBS/1 % Triton‐X‐100 containing 20 mM Imidazole and once with T25N50. Elution was performed with 300 mM Imidazole in T25N50.

2.2.3.3 Expression and purification of NEDD8 and Ubiquitin for click reactions Methionine auxotrophic E. coli B834 (DE3) containing His‐Ubiquitin G76M in pGDR11 were kindly provided by Silvia Eger (AG Marx, University of Konstanz). His‐NEDD8 M50A G76M was cloned into pGDR11 and transformed into chemical competent methionine auxotrophic E. coli B834 (DE3) provided by AG Marx, University of Konstanz. To express His‐Ubiquitin or His‐ NEDD8, 10 mL LB medium containing 100 mg/mL ampicillin were inoculated with the respective bacteria and incubated overnight. 500 µl overnight culture were added to 500 mL new minimal media (NMM) containing 0.4 mM methionine and grown to an OD600nm of 0.8. Medium was then changed to NMM with 0.5 mM azidohomoalanine (provided by AG Marx, University of Konstanz). After 30 min, protein expression was induced with 1 mM IPTG. After growing for 4‐6 h, bacteria were centrifuged and the pellet was stored at ‐80 °C until further use. His‐Ubiquitin G76M (Aha) and His‐NEDD8 M50A G76M (Aha) were purified as described in 2.2.3.2 and dialyzed using Slide‐A‐Lyzer MINI Dialysis Devices (3.5 MWCO) to exchange the buffer to 20 mM Tris‐HCl (pH 7.5), 20 mM NaCl, 5 % glycerol.

2.2.3.4 Expression and purification of PCNA for click reaction Plasmids containing tRNA synthetase PylS, pyrrolysine tRNA and (164TAG) PCNA were kindly provided by Marina Rubini (AG Marx, University of Konstanz). After co‐transformation of the

- 33 - Material and Methods plasmids into E. coli BL21 (DE3), several colonies were tested for expression of PylS and truncated PCNA. The colony with the best expression was used to create a glycerol stock. 10 mL LB medium containing 10 mg/mL ampicillin and 25 µg/mL kanamycin were inoculated with the before mentioned glycerol stock and incubated at 37 °C and 200 rpm overnight. Cells were diluted 1:100 the next morning and at OD600nm= 0.3, Plk (provided by AG Marx, University of Konstanz) was added to a final concentration of 1 mM. Protein expression was induced with

1 mM IPTG at OD600nm= 0.8. After 6 h, cells were harvested by centrifugation. The pellet was resuspended in lysis buffer (50 mM Tris‐HCl, 1 mM EDTA, 0.2 mg/mL lysozyme, 1 µg/mL aprotinin, 1 µg/mL leupeptin, 1 mM PefaBlock, 5 % glycerol, pH 7.5) and incubated on ice for 30 min. Afterwards, 0.02 % IGEPAL was added and the lysate was sonicated 2x 40 pulses (duty cycle: 30‐40 %, output control: 3‐4) using BRANSON sonifier 250. After centrifugation (13000 rpm, 15 min), 150 mM ammonium sulfate and polyethyleneimine (400 µl 5 % PEI for 10 mL lysate) were added to the supernatant and stirred for 10 min at room temperature. In a next step, lysate was centrifuged at 15000 rpm for 40 min. Additional ammonium sulfate (0,23 g/mL) was added to the supernatant, stirred for 1 h at room temperature and then centrifuged at 15000 rpm for 30 min. The same procedure was repeated with 0,24 g/mL ammonium sulfate and the obtained pellet was resuspended in buffer A (20 mM Tris/HCl; 20 mM NaCl; 1 mM EDTA; 1 mM DTT; 5 % glycerol, pH 7.5). The protein was the loaded on a Q sepharose column (3 mL beads/1L culture) (Q sepharose Fast Flow, GE Healthcare) and eluted with a NaCl gradient in buffer A from 100 mM‐1 M NaCl (elution of PCNA at 300‐500 mM). Fractions containing pure protein were pooled and dialyzed (Spectra/Por MWCO 6‐8kDa, 14.6 mm, Serva) against buffer A containing 0.5 mM EDTA.

2.2.3.5 In vitro translation In vitro translation was performed using TNT coupled Reticulocyte Lysate System (Promega) and S35‐labeled methionine (Perkin Elmer) according to manufacturer´s instructions.

2.2.4 Protein analysis

2.2.4.1 Bradford assay Bio‐Rad Protein Assay (Bio‐Rad) was used to normalize total protein amounts according to manufacturer´s instructions. 200 µl Bradford reagent were mixed with 1 µl cell lysate and absorption was measured at 595 nm using a micro plate reader (Wallac 1420 VICTOR Multilabel Plate Reader).

- 34 - Material and Methods

2.2.4.2 SDS­polyacrylamide gel electrophoresis (SDS­PAGE) SDS‐PAGE was performed according to the protocol of (Laemmli, 1970). Dependent on the size of proteins, 12.5 %, 15 % or 5‐15 % (gradient gels) final concentration of acrylamide in separating gels was used. For cellular assays, protein amounts were equalized either by β‐ galactosidase or Bradford assay. Samples were incubated with 5x or 2x loading buffer (reducing conditions) for 5 min at 95 °C or 2x Urea buffer (non‐reducing conditions) for 15 min at RT. Electrophoresis was performed at a constant current of 50 mA (or 25 mA for thioester assays). Gels were analyzed by Coomassie blue staining, western blot or fluorography.

2.2.4.3 Coomassie Blue and colloidal Coomassie staining After SDS‐PAGE, gels were stained with Coomassie for 15 min at room temperature and destained with destain solution for several hours. For Colloidal Coomassie staining, 5x Roti‐Blue (ROTH) was used according to manufacturer´s manual.

2.2.4.4 Fluorography To detect S35‐labeled proteins, gels were fixed for 15‐30 min in 40 % methanol and 10 % acetic acid. Signals were enhanced by incubating gels for 15‐30 min in Amplify (Amersham Biosciences). After drying for 2 h at 80 °C, gels were applied to an Imaging plate BASIIIs (Fuji) and signals were detected using BASReader (Raytest).

2.2.4.5 Western Blot Protein samples were separated by SDS‐PAGE and gels were equilibrated in transfer buffer. PVDF membrane (Millipore) was activated in methanol and consecutively washed in transfer buffer. Western blotting was performed in a wet transfer apparatus (BIO‐RAD) for 90 min at 60 V. Afterwards, membranes were incubated in 5 % milk (w/v) in TNE‐T for 30 min and washed with TNE‐T two times for 10 min. After incubation with the primary antibody for 1 h, membranes were washed three times for 10 min. Secondary antibody coupled to horse radish peroxidase was applied for 45 min. Blots were washed three times and developed using Western Lightning‐Plus ECL (Perkin Elmer) in the imaging system LAS‐3000 (Fujifilm). When necessary, blots were stripped using stripping buffer for 30 min at room temperature. Afterwards, blots were washed several times with TNE‐T, blocked again and incubated with another antibody.

- 35 - Material and Methods

2.2.5 In vitro assays

2.2.5.1 Methanol­Chloroform precipitation of proteins Proteins were precipitated according to (Wessel and Flugge, 1984). The protein solution was mixed with the same volume Methanol: Chloroform (4:1) by vortexing. After centrifugation at 13200 rpm for 5 min at room temperature, supernatant was removed and Methanol was added (same volume as before with Methanol and Chloroform). The mixture was vortexed and centrifuged again. Supernatant was completely removed and the protein pellet was dried in a water bath at 55 °C. Afterwards, the pellet was resuspended in 2 % SDS and 10x Ergänzungspuffer was added. Proteins were boiled and separated using SDS‐PAGE.

2.2.5.2 GST­pulldown assay GST or GST‐APPBP1/UBA3 were expressed and bound to beads as described in 2.2.3.1. C‐ terminally His‐tagged E2s were purified according to 2.2.3.2. GST‐protein coupled GSH‐beads were incubated with unmodified or autoNEDDylated E2s (2.2.5.4; stopped for 15 min with 10 mM EDTA) in T25N50 containing BSA for at least 4 h at 4 °C. Afterwards, beads were washed 5x with T25N50 and boiled with 2x Laemmli loading buffer. Inputs for His‐tagged proteins and pulled‐down proteins were detected via western blot using α His‐antibody. Input gels for GST‐ proteins were subjected to Coomassie staining. GST, GST‐ubiquitin and GST‐NEDD8 were expressed and bound to beads as described in 2.2.3.1. Beads were incubated with bacterially expressed His‐PCNA (purified according to 2.2.3.2). GST‐ protein coupled GSH‐beads were incubated with His‐PCNA in T25N50 containing BSA for 4 h at 4 °C. Afterwards, beads were washed 5x with T25N50 and boiled with 2x Laemmli loading buffer. Inputs for His‐tagged proteins and pulled‐down proteins were detected via western blot using α His‐antibody. Input gels for GST‐proteins were subjected to Coomassie staining.

2.2.5.3 Affinity chromatography Bacterially expressed His‐NEDD8, His‐ubiquitin or BSA were coupled to activated NHS‐ Sepharose (Hi‐Trap NHS‐activated HP, GE Healthcare) according to manufacturer´s instructions. Columns were afterwards equilibrated with affinity pulldown buffer and blocked with 5 mg/mL BSA for 2 h at 4 °C. Cell lysate was prepared from thirty 15 cm plates HEK293T cells per pulldown. Cells were lysed in affinity pulldown buffer (500 µl / plate), sonicated and centrifuged for 1 h at 20000 rpm and 4 °C. Lysate was precleared with Sepharose beads (Sepharose CL‐4B, GE Healthcare) for 3 h at 4 °C. Cell lysate was applied to the affinity columns (flow ~0.5 mL/min) over night. After this, columns were washed using 20 mL affinity pulldown buffer and proteins were eluted with 3 mL

- 36 - Material and Methods affinity pulldown buffer including 400 mM NaCl, 1 M NaCl and 6 M Urea. Elutions were precipitated using Methanol/Chloroform and resuspended in 2 % SDS. To load the samples on a SDS gel, 10x Ergänzungspuffer was added. Finally, SDS‐PAGE was performed using a 5 % ‐ 15 % gradient gel and proteins were stained with Colloidal Coomassie (Roti Blue, ROTH). Protein bands of the 1M NaCl elution were excised and further examined by the mass spectrometry facility of the University of Konstanz. Upon tryptic digest, samples were measured on an LTQ Orbitrap using LC‐MS/MS (LTQ Orbitrap Discovery with Eksigent 2D‐nano HPLC, Thermo).

2.2.5.4 In vitro NEDDylation assay His‐NEDD8, GST‐APPBP1/UBA3 and Ubc12‐His or Nce2‐His were expressed in bacteria (2.2.3.1. and 2.2.3.2). HA‐tagged E2s or NEDD8‐E2 fusions and Nce2 isoform2‐His were in vitro translated (2.2.3.5). Proteins were mixed and incubated in T25N50 containing 1 mM DTT, 2 mM

MgCl2 and 2 mM ATP. After incubation for 2 h at 30 °C (ubiquitination) or 35 °C (NEDDylation), 5x Laemmli loading buffer was added and samples were heated for 5 min at 95 °C prior to loading.

2.2.5.5 Thioester assay Bacterially expressed His‐NEDD8 or GST‐NEDD8, GST‐Ubiquitin, GST‐APPBP1/UBA3, Ubc12‐ His, Nce2‐His, UbcH5b‐His, in vitro translated Nce2 Isoform2‐His, baculovirally expressed UBA1 and Ubiquitin (SIGMA) were used for thioester assays. Proteins were mixed and incubated in

T25N50 under thioester conditions (1.2 mM DTT, 6.25 mM MgCl2, 3.75 mM ATP) for 30 sec to 2 min. Samples were divided and 2x Urea sample buffer or 2x Laemmli buffer was added. Only samples containing Laemmli buffer were boiled prior to loading.

2.2.5.6 Cu(I)­catalyzed Huisgen azide­alkyne cycloaddition His‐Ub(Aha) or His‐NEDD8(Aha) and PCNA(Plk) in 20 mM Tris‐HCl (pH 7.5), 20 mM NaCl, 5 % glycerol (and 0.5 mM EDTA for PCNA) were mixed with 1 mM TCEP, 10 µM TBTA and 1 mM

CuSO4. Reaction vessels were flushed with argon to prevent Cu‐induced protein oxidation (Eger et al., 2011). After incubation at room temperature for 1 h the reaction was stopped by addition of 5x Laemmli loading buffer.

2.2.5.7 Transient transfection For transient transfection, H1299 cells were transfected with Lipofectamine 2000 (Invitrogen) and HEK293T cells with Turbofect (Fermentas) according to manufacturer´s instructions and a ratio of transfection reagent:DNA of 2 µl:1 µg.

- 37 - Material and Methods

2.2.5.8 TNN cell lysis and ß­galactosidase assay Cell pellets were lysed in TNN lysis buffer containing 1 mM DTT, 1 µg/mL aprotinin, 1 µg/mL leupeptin and 1 mM PefaBlock in a volume dependent on the size of the pellet (30 ‐ 100 µl) (for lysis procedure, see 2.2.5.4). 3 ‐ 5 µl lysate were pipetted into a 96 well plate and mixed with 120 µl buffer Z and 5 µl ONPG. The plate was incubated at 37 °C and absorbance was measured at 405 nm in a micro plate reader (Wallac 1420 VICTOR Multilabel Plate Reader).

2.2.5.9 Immunoprecipitation For immunoprecipitation of HA‐tagged NEDD8 E2 enzymes from cytosolic fractions (2.2.5.11), 20 µl HA‐beads were added to 2 mL of H1299 lysate and incubated for 4 h at 4 °C. Afterwards, beads were washed 4x with 1 mL HP buffer. Proteins were eluted by boiling in 30 µl 2x Laemmli loading buffer. For immunoprecipitation of endogenously NEDDylated Ubc12, one well of a 6 well plate untransfected or transfected HEK293T cells (HA‐Ubc12; Turbofect) was treated with DMSO or 1 µM MLN4924 for 3 h. Cells were lysed according to 2.2.5.11 to obtain cytosolic fractions and the lysate was incubated with 1 µl α NEDD8‐antibody (Epitomics) for 1 h at 4 °C. Afterwards, 30 µl protein A‐sepharose beads were added and incubated for 4 h at 4 °C. Beads were then washed 4x with 1 mL HP buffer and boiled with 30 µl 2x Laemmli loading buffer. To immunoprecipitate PCNA, H1299 cells were divided in a Triton‐soluble and a triton‐insoluble fraction according to (Kannouche et al., 2004). Lysates (1 mL each; 2 % used as input) were precleared for 1 h using 30 µl protein A‐sepharose beads. Thereafter, 1 µg α PCNA‐antibody (PC10, Abcam) and 50 µl protein A‐sepharose beads were added to the supernatant and incubated for 4 h at 4 °C. Beads were washed 3x with 1 mL of the respective buffer used for lysis and boiled in 2x Laemmli loading buffer. HA‐IP and Nce2‐isoform 2‐IP were conducted as follows: two 10 cm plates of HEK293T cells (untransfected, transfected with 4 µg HA‐Nce2 isoform 1 or 8 µg isoform 2 each) were lysed in 1 mL TNN each. 2 % of the lysate were used as input. 1 mL lysate was incubated with either 2 µl α HA‐antibody (α HA 1.1, Covance) or 10 µl α Nce2 isoform 2‐antibody (TFA, University of Konstanz; purified (2.2.5.14)) for 1 h at 4 °C. Afterwards, 50 µl Protein A‐sepharose beads were added and mixture was incubated for 4 h at 4 °C. IPs were washed 5x with 1 mL TNN and proteins were eluted with boiling in 50 µl 2x Laemmli loading buffer.

2.2.5.10 Cycloheximide chase HA‐Nce2 isoform1 and HA‐Nce2 isoform2 constructs were transfected in H1299 cells (6 cm dishes). After 4 h, cells were trypsinized and seed onto a 6 well plate. At the next day, cells were treated with 60 µg/mL cycloheximide for 0 h, 1 h, 2 h and 4 h or for 4 h with cycloheximide and 10 µM MG132, harvested and lysed in TNN buffer.

- 38 - Material and Methods

2.2.5.11 Cellular fractionation To divide cellular proteins in a cytosolic and several nuclear fractions, cell pellets (10 cm dish) were resuspended in 2.5 mL HP buffer. After 15 min incubation on ice, cells were dounced 50 times. 125 µl of this lysate were mixed with 50 µl 10 % SDS as input sample. In a next step, the lysate was centrifuged at 600 g for 5 min at 4 °C. The supernatant now contained only cytosolic proteins. The pellet was resuspended again in 2.5 mL HP buffer containing 10 % NP40 (Fluka). After 15 min swelling on ice, lysate was centrifuged at 1000 rpm for 5 min at 4 °C. Supernatant contains soluble nucleosolic proteins. The pellet was resuspended in 1x KE buffer containing 50 mM NaCl, incubated on ice for 15 min and centrifuged again. To obtain different nuclear fractions, this procedure was repeated using 200 mM, 350 mM and 450 mM NaCl. The remaining pellet was then resuspended in 2.5 mL 2 % SDS and sonicated with 20 pulses (duty cycle: 30‐40 %, output control: 3‐4) using BRANSON sonifier 250. Input and 250 µl of every fraction were precipitated with Methanol and Chloroform (2.2.5.1).

2.2.5.12 Preparation of total RNA Total RNA of HEK293T cells was obtained from one confluent 10 cm cell culture dish using TRIzol Reagent (Invitrogen) according to manufacturer´s instructions.

2.2.5.13 Reverse transcription Reverse transcription was performed using SuperScript III Reverse Transcriptase Kit (Invitrogen) according to manufacturer´s instructions. 2 µg total RNA of HEK293T cells were applied. Primers DP74 or DP77 were used for reverse transcription and the following PCR, where AR89 was included as forward primer.

2.2.5.14 Antibody purification 1.5 mg of bacterially expressed and purified Nce2 isoform 2‐His were coupled to 1 mL SulfoLink resin (Pierce) according to manufacturer´s instructions. Afterwards, beads were washed with 10 mL PBS, 10 mL elution buffer and 30 mL PBS. 5 mL filtered (0.45 µM filter, ROTH) serum were applied to the column overnight. In the morning, serum was incubated with the upright standing column for 30 min more. In a next step, beads were washed with 30 mL PBS, with 30 mL washing buffer and again with 30 mL PBS. Antibody was eluted with 10x 500 µl elution buffer into tubes already provided with 1 M Tris‐HCl pH 8.5 to finally reach pH 7.5.

- 39 - Material and Methods

2.2.6 In cellulo assays

2.2.6.1 In cellulo ubiquitination and NEDDylation assays 90 % confluent cells were transfected with His‐myc‐ubiquitin (pcDNA3.1), His‐NEDD8 (pSG5.0‐ Spl) or His‐tagged substrate using Lipofectamine 2000 (Invitrogen). After 24 h, cells were harvested and one third was used for input lysis, two thirds for Ni‐NTA‐pulldown. For endogenous PCNA as substrate, input lysis was performed using 50 µl RIPA buffer containing 1 µg/mL aprotinin, 1 µg/mL leupeptin and 1 mM PefaBlock, including sonication. For overexpressed substrates, 30 µl TNN containing 1 mM DTT, 1 µg/mL aprotinin, 1µg/mL leupeptin and 1 mM PefaBlock were used for lysis. After resuspension of cell pellets, they were incubated for 30 min on ice and centrifuged for 30 min at 13200 rpm and 4 °C. Supernatant was used for β‐galactosidase assay to equalize transfection efficiency. For pulldowns, cells were lysed in 500 µl Guanidinium‐HCl buffer and incubated on ice for 30 min. 100 µl sepharose beads (sepharose 4B, GE Healthcare) were added to preclear the lysate for 2 h at 4 °C. After centrifuging, supernatant was incubated for 4 h with 50 µl Ni‐NTA‐agarose (Qiagen). Beads were washed two times with Guanidinium‐HCl buffer, two times with a mixture of 1 part Guanidinium‐HCl buffer and 4 parts 50 mM Tris‐HCl containing 20 mM Imidazole, and three times with 50 mM Tris‐HCl containing 20 mM Imidazole. After removal of the supernatant, beads were boiled using a 1:1 mixture of 1 M Imidazole and 5x Laemmli loading buffer.

2.2.6.2 Immunofluorescence Cells were seeded on 12 mm cover slips in a 24 well plate (Greiner) and transfected using Lipofectamine 2000 (Invitrogen). 24 h after transfection, cells were washed with PBS and fixed with 500 µl paraformaldehyde in PBS for 15 min at room temperature. Afterwards, cover slips were washed twice with PBS. Cells were permeabilized using 500 µl 0.1 % Triton‐X‐100 (v/v) in PBS for 5 min on ice and washed twice with PBS. Unspecific binding sites were blocked with PBS containing 2 % (v/v) FCS and 0.5 % (v/v) Triton‐X‐100 for 1 h at room temperature. Cells were then incubated with primary antibody in PBS containing 2 % (v/v) FCS and 0.5 % (v/v) Triton‐ X‐100 for 1 h at room temperature on a shaker. After washing three times with PBS containing 2 % (v/v) FCS and 0.5 % (v/v) Triton‐X‐100, cells were incubated with the secondary antibody for 1 h at room temperature on a shaker in the dark. Afterwards, cells were washed again three times with PBS containing 2 % (v/v) FCS and 0.5 % (v/v) Triton‐X‐100. To stain nuclei, cells were then incubated with DAPI (1:20.000) for 10 min at room temperature. Cells were mounted on glass slides (Thermo Fisher Scientific) using cover slips (Roth) and Mowiol 4‐88 (Roth). Pictures were taken with LSM 510 Meta (Zeiss) at the Bioimaging Center (University of Konstanz).

- 40 -

3. Results

3.1 AutoNEDDylation of NEDD8‐conjugating enzymes increases the affinity to their cognate E1

Conjugation of UBLs to substrates is of major importance for the regulation of various cellular functions. However, only little is known about how the enzymatic cascades of UBL conjugation are regulated. Only few modes of regulation at the E2 level of the ubiquitination, SUMOylation and NEDDylation cascade have been described so far, some of which involve the posttranslational modification of the E2 with a UBL protein. Dependent on binding to APC/C, its E2 enzyme UbcH10 is autoubiquitinated and degraded when cells exit from mitosis (Yamanaka et al., 2000). Inhibition of activity by autoubiquitination was also shown for Ube2t, a conjugating enzyme involved in the Fanconi anemia DNA repair pathway (Machida et al., 2006). In both cases, ubiquitin is conjugated to a conserved lysine residue next to the catalytic cysteine (Lin et al., 2002; Machida et al., 2006). However, as shown for Ubc7, ubiquitin chains are also formed directly on the catalytic cysteine (Ravid and Hochstrasser, 2007). Additionally, modification of Ubc13 with ISG15 as well as SUMOylation of Ube2k/E2‐25k/Hip2 in the E1 binding interface suppresses thioester formation with ubiquitin (Pichler et al., 2005; Zou et al., 2005). The SUMO conjugating enzyme Ubc9 uses autoSUMOylation to regulate substrate specificity. Modification of Ubc9 with SUMO enhances the affinity for its substrate Sp100 which contains an SUMO interaction motif, whereas SUMOylation of other substrates like RanGAP1 is reduced (Knipscheer et al., 2008). For NEDD8‐conjugating enzymes, the only known regulatory mechanism is acetylation of Ubc12 and Nce2 at the N‐terminal methionine which facilitates binding to a hydrophobic pocket of DCNL, a scaffold‐type E3 ligase involved in the NEDDylation of cullins, thereby enhancing the efficiency of cullin modification (Monda et al., 2013; Scott et al., 2011). However, up to now it is not known whether there are additional mechanisms to regulate NEDD8 conjugation and how NEDD8 substrate specificity is achieved.

- 41 - Results

3.1.1 Ubc12 and Nce2 form a thioester bond with NEDD8 but not with ubiquitin

To confirm activity and specificity of the NEDD8‐conjugating enzymes, we performed thioester assays with bacterially expressed Ubc12‐His or Nce2‐His and His‐NEDD8 or His‐ubiquitin (see 2.2.5.5). The ubiquitin E2 UbcH5b was used as specificity control. In the presence of the NEDD8 E1 enzyme APPBP1/UBA3, both Ubc12 and Nce2 form a thioester linkage with NEDD8 whereas UbcH5b does not (Figure 9A, ‐DTT). In reduced samples, higher migrating bands also appear for both NEDD8‐conjugating enzymes, representing a stable covalent modification with NEDD8 (Figure 9A, +DTT). Moreover, NEDD8 E2 enzymes are not capable of accepting ubiquitin from the ubiquitin E1 UBA1 indicating that, under the conditions used, enzymes of the NEDD8 and the ubiquitin‐conjugation system are specific (Figure 9B, ‐DTT).

A

B

Figure 9. Ubc12 and Nce2 specifically form thioester bonds with NEDD8 Bacterially expressed Ubc12‐His, Nce2‐His, UbcH5b‐His, GST‐APPBP1/UBA3 and His‐NEDD8 (A) or bacterially expressed Ubc12‐His, Nce2‐His, UbcH5b‐His, His‐ubiquitin and baculovirally expressed UBA1 (B) were incubated as indicated under thioester conditions. Samples were divided and afterwards reduced (+DTT) or not (‐DTT). After SDS‐ PAGE, proteins were detected by western blot using α His‐antibody. Ubc12 and Nce2 form thioester linkage with NEDD8 but not with ubiquitin.

- 42 - Results

3.1.2 NEDD8‐conjugating enzymes are autoNEDDylated

In figure 9A, a non‐reducible modification of Ubc12 and Nce2 with NEDD8 was observed in thioester assays, providing a hint that E2 enzymes might act as NEDD8 substrates. Thus, we studied whether NEDD8‐conjugating enzymes are indeed capable of NEDDylating themselves in vitro and in cells. In in vitro NEDDylation assays (see 2.2.5.4), we detected higher migrating bands of both Ubc12 and Nce2 after incubation with APPBP1/UBA3 and NEDD8, corresponding to mono‐ and multi‐ or polyNEDDylated E2 enzymes (Figure 10A, asterisks). In the upper part of the gel, unspecific bands appeared upon incubation (Figure 10A). NEDDylation of the E2s was also found upon addition of NEDD8 only, presumably due to the presence of APPBP1/UBA3 in the reticulocyte lysate used for in vitro translation (Figure 10A). To test whether NEDD8‐ conjugating enzymes are NEDDylated in cellulo, HA‐tagged wt E2s or catalytically inactive mutants (Ubc12 C111S and C111A; Nce2 C116S and C116A) were co‐transfected with His‐ tagged NEDD8 into H1299 cells. After 24 h, cells were harvested and one third was used for TNN lysis (Input). Remaining cells were lysed under denaturing conditions to avoid deNEDDylation during lysis. Subsequently, a Ni‐pulldown was performed to enrich for NEDDylated proteins (see 2.2.6.1). Upon pulldown, we found HA‐Ubc12 and HA‐Nce2 to be NEDDylated in cellulo under overexpression conditions (Figure 10B). Interestingly, NEDDylation of the E2 enzymes was dependent on the catalytic cysteine, suggesting an autoNEDDylation mechanism. Notably, cysteine to serine mutants of both conjugating enzymes (Ubc12 C111S and Nce2 C116S) formed an oxyester bond with His‐NEDD8 (Figure 10B). When we performed a cellular fractionation assay to investigate the subcellular localization of NEDD8‐conjugating enzymes, we identified higher migrating bands of HA‐Ubc12 and HA‐Nce2 in the cytosolic fraction of H1299 cells which could correspond to NEDDylated E2 enzymes (see figure 19). To confirm this presumption, we performed an α HA‐IP from the cytosolic fraction of H1299 cells upon overexpression of HA‐Ubc12 or HA‐Nce2 (Figure 11A). By using an α NEDD8‐ antibody for detection, we indeed found Ubc12 and Nce2 to be modified with endogenous NEDD8 (Figure 11A, arrows in left panel). While HA‐Ubc12 was monoNEDDylated, at least two NEDD8 moieties were attached to HA‐Nce2. Moreover, treating the cells with the NEDD8 E1 inhibitor MLN4924 completely abolished NEDDylation of the E2s (Figure 11A). In the same experiment, we tested whether modification of the NEDD8‐conjugating enzymes with endogenous NEDD8 occurs in an autoNEDDylation reaction. Therefore, we included catalytically inactive mutants of the Ubc12 and Nce2. Upon α HA‐IP, we found that mutation of the catalytic cysteine residue to an alanine inhibits NEDDylation of the E2 enzymes (Figure 11A). In order to prove the existence of NEDDylated endogenous Ubc12, we additionally conducted an α NEDD8‐ IP from cytosolic fractions of untransfected HEK293T cells and cells transfected with HA‐Ubc12 (Figure 11B).

- 43 - Results

A

unspecific bands * * * * *

B

Figure 10. Ubc12 and Nce2 are autoNEDDylated in vitro and in cells (A) In vitro translated (Ivt), radiolabeled HA‐Ubc12 and HA‐Nce2 were incubated with bacterially expressed His‐ NEDD8 and/or GST‐APPBP1/UBA3 (as indicated) under NEDDylation conditions. E2 enzymes were detected using fluorography. Both NEDD8‐conjugating enzymes are modified with NEDD8 (asterisks). (B) H1299 cells were transfected with the indicated DNA and harvested after 24 h. One third of the cells were used for the preparation of whole cell lysate (input). With the remaining cells, Ni‐pulldown was performed under denaturing conditions. Western blot using α HA‐antibody shows that both HA‐Ubc12 wt (left panel) and HA‐Nce2 wt (right panel) are NEDDylated in cells whereas catalytically inactive mutants (Ubc12 C111S and C111A or Nce2 C116S and C111A)are not.

Upon IP and western blot using α NEDD8‐antibody, we identified a band at a size of ~30 kDa in untransfected cells which slightly shifted up in cells transfected with HA‐Ubc12 (NEDD8: 9 kDa, Ubc12: 21 kDa, HA‐tag: 1.5 kDa) (Figure 11B, upper panel). Our results indicate that the detected bands correspond to Ubc12 and HA‐Ubc12, respectively, being covalently modified with endogenous NEDD8. Moreover, these bands disappeared upon treatment with the APPBP1/UBA3‐inhibitor MLN4924 (Figure 11B, upper panel). Due to non‐reducing conditions used for lysis and IP, unmodified HA‐Ubc12, which formed a thioester bond with NEDD8 during lysis, was also immunoprecipitated (Figure 11B, lower panel). Notably, we were not able to directly show a modification of Ubc12 or Nce2 with NEDD8 in cellulo because of a low detection efficiency of the available antibodies (data not shown). In conclusion, our data demonstrate that NEDD8‐conjugating enzymes are capable of autoNEDDylating in vitro and in cellulo using endogenous NEDD8.

- 44 - Results

A

* *

B

Figure 11. Overexpressed Ubc12 or Nce2 and endogenous Ubc12 are modified with endogenous NEDD8 in the cytosol of H1299 or HEK293T cells (A) HA‐Ubc12 wt, HA‐Nce2 wt or catalytically inactive mutants (HA‐Ubc12 C111A/HA‐Nce2 C116A) were overexpressed in H1299 cells. Cells were treated with DMSO or 1 µM MLN4924 (MLN) for 3 h. Upon hypotonic lysis, 5 % of cytosolic lysate were used as input (In) and an α HA‐IP was performed with the cytosolic fraction. Proteins were first detected using western blot α NEDD8 (Alexis; left panel) and in a second step with α HA‐antibody (right panel). Asterisks mark the size of unmodified HA‐Ubc12 and HA‐Nce2 (right panel). Overexpressed NEDD8‐ conjugating enzymes are modified with endogenous NEDD8 (arrows), being dependent on the catalytic cysteine and on the activity of APPBP1/UBA3. (B) Untransfected HEK293T cells or cells transfected with HA‐Ubc12 were treated with DMSO or 1 µM MLN4924 for 3 h before harvest. Upon hypotonic lysis, 5 % of cytosolic lysate were used as input (In) and an α NEDD8‐IP was performed with the cytosolic fraction of the lysate. Western Blot was performed with an α NEDD8‐ (Epitomics) and α HA‐antibody.

3.1.3 HPNI mutants of the NEDD8‐conjugating enzymes are not impaired in thioester formation but in autoNEDDylation

In proximity to the catalytic cysteine, E2 conjugating enzymes contain a conserved HPNI/V motif. The side chain of the asparagine residue in this motif has been proposed to stabilize the oxyanion intermediate formed during lysine attack on the ubiquitin‐loaded E2, therefore being crucial for isopeptide bond formation (Wu et al., 2003b). Furthermore, the asparagine is also indispensable for transferring SUMO from its conjugating enzyme Ubc9 to substrates (Wu et al.,

- 45 - Results

2003b). To investigate whether mutation of the conserved asparagine in the HPNI motif of Ubc12 or Nce2 affects the NEDDylation cascade, we tested HPNI mutants of the NEDD8‐ conjugating enzymes regarding their ability to form thioester with NEDD8 and to autoNEDDylate. In a time course experiment, we found that HPNI mutants (Ubc12 N103A; Nce2 N108A) form thioesters as rapidly as wt E2 enzymes (Figure 12). However, autoNEDDylation, which is very efficient for Ubc12 wt and Nce2 wt after 10 min, is completely abrogated in case of the HPNI mutants (Figure 12). Thus, the asparagine residue in the HPNI motif of conjugating enzymes plays an important and conserved role in facilitating isopeptide bond formation between the C terminus of UBLs and a lysine residue of their substrate.

Figure 12. Ubc12 and Nce2 HPNI mutants are not capable of autoNEDDylation in vitro Bacterially expressed Ubc12‐His wt or HPNI mutant (N103A; HPNI mut), Nce2‐His wt or HPNI mutant (N108A; HPNI mut), His‐NEDD8 and GST‐APPBP1/UBA3 were incubated under thioester conditions. Samples were taken at the indicated time points and divided in an afterwards reduced (+DTT) and not reduced part (‐DTT). Upon SDS‐PAGE and western blot, proteins were detected with α His‐antibody. HPNI mutants of both conjugating enzymes form a thioester bond with NEDD8, but are impaired in autoNEDDylation.

3.1.4 AutoNEDDylation of the NEDD8 E2 enzymes predominantly occurs in their N terminus

To gain further insights into the mechanism of autoNEDDylation, we next aimed to identify NEDDylated lysines within Ubc12 and Nce2. Therefore, in vitro autoNEDDylated E2 enzymes were analyzed by mass spectrometry. Lysines 3, 8, 11, 36 and 45 of Ubc12 and lysine 7 and 9 of Nce2 were found to be modified with NEDD8 (data not shown). Interestingly, most of the NEDDylated lysines are located in the N‐terminal extension of Ubc12 and Nce2 which is not found in most other UBL‐conjugating enzymes including the ubiquitin‐conjugating enzyme UbcH5b (Figure 13). However, after mutating K3, K8 and K11 of Ubc12, and K7 and K9 of Nce2

- 46 - Results to arginine, both conjugating enzymes were still capable of autoNEDDylation in vitro and in cells (see figures 15A and 15B). Nonetheless, NEDDylation efficiency of the E2 lysine mutants in cells was slightly decreased in comparison to wt E2s (see figure 15B). These data suggest that additional lysines or the very N terminus are used for NEDDylation. In mass spectrometric analyses of the autoNEDDylated lysine mutants of the E2 enzymes, we indeed identified lysine 15 as an additional modified lysine in Nce2, but no additional lysines were detected for Ubc12. We also did not find evidence for the presence of NEDD8 chains, indicating that NEDD8‐ conjugating enzymes are rather multi‐ than polyNEDDylated (data not shown).

Figure 13. NEDD8­conjugating enzymes are predominantly NEDDylated in their unique N terminus Alignment of the NEDD8‐conjugating enzymes Nce2 and Ubc12 and the ubiquitin E2 enzyme UbcH5b was carried out using ClustalW (www.ebi.ac.uk/Tools/msa/clustalw2/). Lysines identified to be NEDDylated are marked in red. NEDD8 E2 enzymes are NEDDylated in their unique N‐terminal extension.

We showed that both Ubc12 and Nce2 are autoNEDDylated at different lysine residues in vitro, which are almost exclusively located in the N termini being unique to NEDD8‐conjugating enzymes. Since the N terminus is involved in the interaction with the E1 enzyme and its modification might play a role in the regulation of this interaction, we wanted to confirm that it is indeed the major NEDDylation site of the E2s. In NEDDylation and thioester experiments, we found that mutants of Ubc12 and Nce2 lacking the N terminus are not NEDDylated anymore in vitro and their NEDDylation was strongly impaired in cells (Figure 14A and C). However, they are not impaired in thioester formation under the conditions used (Figure 14B).

- 47 - Results

A

B

C

Figure 14. Ubc12 and Nce2 mutants lacking their N­termini are impaired in autoNEDDylation but not in thioester formation (A) In vitro translated HA‐Ubc12 wt, HA‐Ubc12 ΔN26, HA‐Nce2 wt, HA‐Nce2 ΔN26, bacterially expressed His‐NEDD8 and GST‐APPBP1/UBA3 were incubated under NEDDylation conditions. Samples were taken after the indicated time points. SDS‐PAGE was performed and proteins were detected using fluorography (Johanna Bialas, under supervision). Deletion of the N‐terminal 26 aa of the E2s inhibits autoNEDDylation in vitro. (B) Bacterially expressed Ubc12‐His wt, ΔN26, Nce2‐His wt, ΔN26, His‐NEDD8 and GST‐APPBP1/UBA3 were incubated under thioester conditions. Samples were taken after the indicated time points and divided in an afterwards reduced (+DTT) and not reduced (‐DTT) part. SDS‐PAGE was performed and proteins were detected with α His‐antibody solution. ΔN26 mutants of Ubc12 and Nce2 are not impaired in thioester linkage formation (asterisk) with NEDD8 in vitro. (C) H1299 cells were transfected with the indicated DNA and harvested after 24 h. One third of the cells were used for the preparation of whole cell lysate (input). With the remaining cells, Ni‐pulldown was performed under denaturing conditions. Western blots using α HA‐antibody reveal NEDDylation of E2 ΔN26 mutants to be impaired in cells.

- 48 - Results

To gain first hints about the function of autoNEDDylation, fusion proteins of NEDD8 and the E2 enzymes were generated in order to mimic a modification with NEDD8 within the N terminus. E2s fused to the C terminus of NEDD8 showed a remarkably enhanced modification with NEDD8 compared to wt E2 enzymes in cells (Figure 15B). Concordantly, these NEDD8‐E2 fusion proteins were more efficiently modified with NEDD8 than wt E2s in autoNEDDylation kinetics in vitro (Figure 15C). To sum up, NEDD8‐conjugating enzymes are capable of autoNEDDylating their extended N termini that are not present in most other conjugating enzymes. AutoNEDDylation might therefore interfere with the interaction with the NEDD8 E1 enzyme APPBP1/UBA3 and have an influence on the processivity of the NEDDylation cascade.

A B

C

Figure 15. Ubc12 and Nce2 lysine mutants are still modified with NEDD8 and NEDDylation of the E2s is strongly enhanced upon fusion to NEDD8 (A) Bacterially expressed Ubc12‐His wt, K3R/K8R/K11R (K mut), NEDD8‐Ubc12‐His, Nce2‐His wt, K7R/K9R (K mut), NEDD8‐Nce2‐His, His‐NEDD8 and GST‐APPBP1/UBA3 were incubated under NEDDylation conditions. SDS‐PAGE was performed and gel was stained with Coomassie stain solution. Lysine mutants of both NEDD8‐conjugating enzymes are still NEDDylated in vitro. (B) H1299 cells were transfected with the indicated DNA and harvested after 24 h. One third of the cells were used to prepare whole cell lysate (input). With the remaining cells, Ni‐pulldown was performed under denaturing conditions. Western blots using α HA‐antibody show that NEDDylation of the lysine mutants (K mut) is decreased compared to wt E2s and NEDD8‐E2 fusions (N8 fus.) are strongly modified with NEDD8. (C)In vitro translated (Ivt), radiolabeled HA‐Ubc12, HA‐NEDD8‐Ubc12, HA‐Nce2 or HA‐NEDD8‐Nce2 were incubated with bacterially expressed His‐NEDD8 and GST‐APPBP1/UBA3 for the indicated time under NEDDylation conditions. NEDD8‐E2 fusion proteins show enhanced autoNEDDylation kinetics compared to wt E2 enzymes.

- 49 - Results

3.1.5 AutoNEDDylation enhances the affinity of Ubc12 and Nce2 to the NEDD8 E1 enzyme APPBP1/UBA3

Since we found that NEDD8 is conjugated to lysines in the N terminus of its E2 enzymes which is involved in the interaction with APPBP1/UBA3, we next tested whether autoNEDDylation of Ubc12 and Nce2 might affect this interaction. To do so, we performed a GST‐pulldown assay (see 2.2.5.2) using GST‐tagged APPBP1/UBA3 and autoNEDDylated Ubc12 or Nce2 (note that autoNEDDylation reaction was stopped with EDTA before the pulldown) (Figure 16). The results clearly showed that autoNEDDylated E2s possess a higher affinity to the E1 than unmodified ones. Strikingly, an interaction between Nce2 and APPBP1/UBA3 was only detected upon NEDDylation (Figure 16).

Figure 16. Affinity of Ubc12 and Nce2 to APPBP1/UBA3 is enhanced upon autoNEDDylation Bacterially expressed Ubc12‐His or Nce2‐His was incubated with His‐NEDD8 and GST‐APPBP1/UBA3 under NEDDylation conditions. Reaction was stopped with 10 mM EDTA. Afterwards, a GST‐pulldown was performed using GST and GST‐APPBP1/UBA3 and the autoNEDDylated E2s. 10 % of beads were loaded on an SDS‐PAGE and stained with Coomassie (input; left panel). 20 % of autoNEDDylated E2s and pulldown samples were subjected to SDS‐PAGE and western blot using α His‐antibody (right panel). APPBP1/UBA3 interacts more strongly with NEDDylated than with unmodified Ubc12 or Nce2.

The finding that modification of Ubc12 and Nce2 with NEDD8 enhances their affinity to the E1 enzyme is supported by the increased NEDDylation of NEDD8‐E2 fusion proteins in cells and in vitro (Figures 15B and 15C). In addition, it is supported by an observation we made after in vitro translation of the E2s and NEDD8‐E2 fusion proteins containing a mutation of the active site cysteine to serine (Figure 17). During in vitro translation, a ~10 kDa higher migrating band appeared both for Ubc12 C111S and NEDD8‐Ubc12 C111S. Since this band disappeared in the presence of the APPBP1/UBA3 inhibitor MLN4924 and because it was not detected for Ubc12 wt and NEDD8‐Ubc12 wt, we concluded that it corresponds to the E2 forming an oxyester bond with NEDD8 present in the reticulocyte lysate (Figure 17). Moreover, adding GST‐NEDD8 to the

- 50 - Results in vitro translation reaction also lead to modification of Ubc12 and NEDD8‐Ubc12 C111S which was inhibited by MLN4924. Interestingly, in the case of NEDD8‐Ubc12 C111S oxyester formation was much more efficient compared to Ubc12 C111S. Therefore, the higher affinity of APPBP1/UBA3 to NEDDylated than to unmodified E2s might lead to an enhanced modification of NEDD8‐Ubc12 during in vitro translation (Figure 16 and 17). Similar results were obtained using Nce2 C116S and NEDD8‐Nce2 C116S (data not shown).

Figure 17. Ubc12 C111S and NEDD8­Ubc12 C111S are modified with NEDD8 during in vitro translation HA‐Ubc12 C111S and HA‐NEDD8‐Ubc12 C111S were in vitro translated according to manufacturer´s instructions. To one fourth of the reaction, 100 µM MLN4924 was added, to another fourth GST‐NEDD8 and to one fourth, GST‐NEDD8 and 100 µM MLN4924 were added. Samples were loaded on an SDS‐PAGE and radiolabeled proteins were detected using fluorography. NEDD8‐Ubc12 C111S is modified more strongly with NEDD8 (asterisks) or GST‐NEDD8 (arrows) than Ubc12 C111S.

3.1.6 Fusion of NEDD8 to its E2s leads to an enhanced localization in the nucleus

Fusions of NEDD8 or ubiquitin to substrates are widely used to study the function of these modifications. For instance, C‐terminal fusions of NEDD8, SUMO‐1 or ubiquitin to the tumor suppressor p53 turned out to be suitable tools to study the localization of modified p53. The p53‐ubiquitin fusion localizes to the cytoplasm, resembling the localization of monoubiquitinated p53. In contrast, the p53‐NEDD8 fusion is found in the nucleus, and p53‐ SUMO fusions reside in nuclear bodies (Carter and Vousden, 2008). Hence, to further investigate autoNEDDylation of NEDD8‐conjugating enzymes, we tested whether the fusion of NEDD8 to the N terminus of its E2 enzymes affects their subcellular localization. In immunofluorescence experiments using overexpressed HA‐Ubc12 or HA‐Nce2, the NEDD8‐conjugating enzymes were found mainly in the cytoplasm and to a minor extent in

- 51 - Results the nucleus (Figure 18). Additionally, HA‐NEDD8‐Ubc12 and HA‐NEDD8‐Nce2 were present in the same compartments, indicating that fusing NEDD8 to the E2 enzymes has no major effect on their localization in H1299 cells (Figure 18).

Figure 18. NEDD8­E2 fusions localize to the cytoplasm and the nucleus of H1299 cells HA‐Ubc12, HA‐NEDD8‐Ubc12, HA‐Nce2 or HA‐NEDD8‐Nce2 were transfected into H1299 cells and stained with α HA‐ antibody (red) and DAPI (blue; scale bar 20 µm). Both NEDD8‐conjugating enzymes and fusions to NEDD8 localize to the cytoplasm and to the nucleus.

To gain deeper insights into the localization of Ubc12, Nce2 and the NEDD8‐E2 fusion proteins, we performed a cellular fractionation assay with overexpressed E2s in H1299 cells (Figure 19). Upon fractionation, we found the majority of HA‐Ubc12 being present in the cytosol. Nonetheless, in all of the nuclear fractions and the pellet, minor amounts were detected. In contrast, a higher percentage of HA‐NEDD8‐Ubc12 was present in the nuclear fractions compared to the cytoplasmic fraction (Figure 19, upper two panels). We only found Nce2 in the cytoplasm and the soluble nuclear fraction (nucleosol), but the NEDD8‐Nce2 fusion was present in all of the nuclear fractions (Figure 19, lower two panels). Hence, fusion of NEDD8 to the E2 enzymes leads to an enhanced localization in the nucleus. Additionally, we detected a higher migrating band for all of the proteins in the cytosolic fraction which we identified as their monoNEDDylated forms (see chapter 3.1.2). To conclude, NEDDylation of the NEDD8‐conjugating enzymes might at least partially influence their subcellular localization.

- 52 - Results

Figure 19. NEDD8­E2 fusions localize stronger to the nucleus than wt E2s 10 cm plates H1299 cells were transfected with HA‐Ubc12, HA‐NEDD8‐Ubc12, HA‐Nce2 or HA‐NEDD8‐Nce2. Cells were lysed by douncing in hypotonic buffer and nuclear fractions were obtained using increasing NaCl concentrations. 5 % of input (total cell lysate) and 10 % of all fractions were subjected to SDS‐PAGE and subsequent western blot. Proteins were detected using α HA‐antibody. In comparison to wt E2s, NEDD8‐E2 fusions are rather found in nuclear fractions.

- 53 - Results

3.2 PCNA as a new substrate for the NEDD8‐ conjugation pathway

PCNA (Proliferating Cell Nuclear Antigen) has first been described as an autoantigen in patients suffering from systemic lupus erythematosus (Miyachi et al., 1978). It adopts a homotrimeric ring structure with each monomer including two similar domains. The inner part of the ring predominantly comprises basic residues which allow binding to DNA (Krishna et al., 1994). Upon loading onto DNA by the clamp loader complex RFC, PCNA is capable of acting as a sliding clamp and thus, as processivity factor for replicative polymerases (Kelman and O'Donnell, 1995; Lee and Hurwitz, 1990; Mossi and Hubscher, 1998). In 2002, PCNA was identified as a substrate both for ubiquitin and SUMO in yeast (Hoege et al., 2002). Ubiquitination and SUMOylation occur at the same lysine residue (K164) in the S phase of the cell cycle. PCNA is monoubiquitinated by the conjugating enzyme Rad6 and the E3 ligase Rad18 which is recruited to sites of stalled replication by RPA, a protein coating single stranded DNA (Davies et al., 2008; Hoege et al., 2002). In contrast to PCNA monoubiquitination, K63‐ linked polyubiquitination is catalyzed by Ubc13/Mms2 and Rad5. Ubiquitinated PCNA functions in bypassing DNA damages using either translesion synthesis or error‐free pathways, and the mechanism of PCNA ubiquitination as well as its function seem to be conserved from yeast to humans (Bergink and Jentsch, 2009; Hoege et al., 2002; Kannouche et al., 2004; Watanabe et al., 2004). In addition to Rad18, the Cullin4A‐RING complex is capable of monoubiquitinating PCNA at K164 during replication in unperturbed cells, thereby promoting translesion synthesis. However, the same ligase also regulates PCNA degradation via polyubiquitination of the same residue (Lo et al., 2012; Terai et al., 2010). SUMOylation at K164 is thought to enhance binding of PCNA interaction partners such as the Srs2 helicase in yeast or PARI in humans, respectively, which are involved in controlling recombination (Moldovan et al., 2012; Pfander et al., 2005). It was additionally shown that modification of PCNA with SUMO leads to a decrease in the formation of double strand breaks and inhibits recombination in case of stalled replication forks in human cells (Gali et al., 2012). In conclusion, modification of PCNA with ubiquitin or the ubiquitin‐like protein SUMO plays an important role in maintaining the integrity of DNA.

3.2.1 PCNA interacts with NEDD8 and ubiquitin

Modification of proteins with ubiquitin plays a role in various cellular processes (see 1.1.2) and most of the outcomes of ubiquitination are mediated by proteins containing ubiquitin‐binding domains (reviewed in (Hurley et al., 2006)). The identification of new NEDD8‐interaction partners might therefore contribute to a better understanding of physiological functions of NEDD8. Thus, an affinity pulldown was performed in duplicate using His‐NEDD8 or His‐ - 54 - Results ubiquitin, and BSA as a control in one of the pulldowns (Figure 20; see 2.2.5.3). These proteins were covalently coupled to NHS‐sepharose. Afterwards, HEK293T cell lysate was applied and interacting proteins were eluted using 1 M NaCl (Figure 20). The identification of eluted proteins was conducted using tryptic digest followed by mass spectrometry (Table 2).

A B

Figure 20. Affinity pulldown with His­NEDD8, His­ubiquitin or BSA and HEK293T lysate His‐NEDD8, His‐ubiquitin or BSA (A) and His‐NEDD8 or His‐ubiquitin (B) were covalently coupled to NHS‐sepharose via primary amino groups. HEK293T cell lysate was applied to identify interacting proteins that were finally eluted with 1 M NaCl. Elutions were loaded on a gradient SDS‐gel which was stained using colloidal Coomassie.

Table 2. Proteins identified as potential interaction partners for NEDD8 and ubiquitin Bands were excised from Coomassie stained SDS‐gels (Figure 20) and potential interacting proteins were identified using tryptic digest followed by mass spectrometry (Proteomics Facility, University of Konstanz).

His­NEDD8 His­ubiquitin Proteins identified in both pulldowns APPBP1 UCHL1, UCHL3, OTUB1 (DUBs) NUB1L Exosome component 2 NEDD8 DCN1 Ubiquitin Ubiquitin PCNA Elongation initiation factor 4A RuvB like 1 and 2 RuvB like 1 and 2 Programmed cell death 8 MAD2L1 Proteins identified in one pulldown (examples) Skp1 UBA1 (E1) NEDP1 UbcH5A, UbcH7 (E2s) COP9 signalosome subunit 4 RBBP6, CHIP (E3 ligases) p97 p97 Prp4 Proteasome subunits DNA polymerase δ interacting DNA polymerase δ interacting protein 2 protein 2 Programmed cell death 6 Programmed cell death 8 Histone H4 PCNA Proteasome 26S ATPase subunit 5 SMAC3 Primase MAD2L1

In the affinity chromatography using His‐NEDD8, NUB1L as an already known interacting protein was found, showing the applicability of this approach (Tanaka et al., 2003). In addition,

- 55 - Results the E1 enzyme APPBP1/UBA3 was identified in both pulldowns. Using His‐ubiquitin for affinity chromatography, several DUBs as well as the ubiquitin‐activating enzyme UBA1, the conjugating enzymes UbcH5a and UbcH7, and E3 ligases were identified. Interestingly, many of the eluted proteins were found both for NEDD8 and ubiquitin including p97, PCNA and MAD2L1, which might be due to the high similarity of NEDD8 and ubiquitin. For the following studies, we decided to further investigate PCNA as interaction partner of NEDD8 and ubiquitin. To verify the interaction between PCNA and NEDD8 or ubiquitin, respectively, we performed a GST‐pulldown assay (Figure 21). Indeed, we found that GST‐NEDD8 and GST‐ubiquitin directly interact with bacterially expressed His‐PCNA. Furthermore, we could confirm this interaction using C‐terminally tagged NEDD8 and ubiquitin (data not shown). These results suggest that PCNA binds to ubiquitinated or NEDDylated proteins in the cell.

Figure 21. GST­NEDD8 and GST­ubiquitin bind to PCNA GST‐pulldown was performed using bacterially expressed GST, GST‐NEDD8, GST‐ubiquitin and His‐PCNA. 10 % input of PCNA and pulldown samples were subjected to SDS‐PAGE and stained with Coomassie (input GST‐beads; left panel). PCNA was detected by western blot using α PCNA‐antibody (right panel). Both ubiquitin and NEDD8 interact with PCNA.

3.2.2 PCNA is NEDDylated in cells being dependent on K164

Since several proteins including PCNA, p53 or EGFR are known to be modified both with ubiquitin and ubiquitin‐like proteins, PCNA could not only be a substrate for ubiquitin and SUMO‐1 but also for NEDD8 (Hoege et al., 2002; Oved et al., 2006; Rodriguez et al., 1999; Scheffner et al., 1993; Xirodimas et al., 2004). This hypothesis is affirmed by the identification of PCNA as a possible NEDD8 substrate upon affinity pulldown from HeLa cells stably expressing TAP‐NEDD8 (Xirodimas et al., 2008). To investigate whether PCNA is indeed a substrate for NEDD8, we performed an in cellulo NEDDylation assay using overexpressed His‐NEDD8 and included His‐SUMO‐1 and His‐myc‐ubiquitin (Figure 22A). We indeed found PCNA to be monoNEDDylated in H1299 cells. As mentioned before, ubiquitination of PCNA plays a role in translesion synthesis, whereas SUMOylation is rather involved in controlling recombination

- 56 - Results both in yeast and in human cells (Hoege et al., 2002; Moldovan et al., 2012; Pfander et al., 2005; Stelter and Ulrich, 2003). We identified monoubiquitinated PCNA upon overexpression of His‐ myc‐ubiquitin in H1299 cells, but we did not detect modification of PCNA with overexpressed His‐SUMO‐1 (Figure 22A).

A

B

C

Figure 22. PCNA is monoNEDDylated in cellulo and in vitro (A) H1299 cells were transfected with the indicated DNA and harvested after 24 h. One third of the cells were used to prepare whole cell lysate (input). With the remaining cells, Ni‐pulldown was performed under denaturing conditions. Afterwards, proteins were detected using western blot and α PCNA‐antibody. Endogenous PCNA is both monoNEDDylated and –ubiquitinated but not modified with SUMO‐1. (B) In vitro translated PCNA‐His‐SV5 wt or K164R (arrow) was incubated with the indicated bacterially expressed and purified proteins under NEDDylation conditions. Ubc12 and Nce2 are capable of transferring NEDD8 onto both PCNA wt and PCNA K164R in vitro (asterisks). (C) H1299 cells were transfected with the indicated DNA and harvested after 24 h. One third of the cells were used for the preparation of whole cell lysate (input). With the remaining cells, Ni‐pulldown was performed under denaturing conditions. Western blots using α HA‐antibody show that PCNA wt is both monoNEDDylated and ‐ubiquitinated whereas PCNA K164R is not. - 57 - Results

In a next step, we conducted an in vitro NEDDylation assay using in vitro translated PCNA (Figure 22B). In the presence of NEDD8, APPBP1/UBA3 and one of the conjugating enzymes Ubc12 or Nce2, two additional bands of PCNA were detected presumably representing monoNEDDylated PCNA (Figure 22B, lanes 1 and 4). The appearance of two bands might be due to the usage of two different lysines on PCNA or to the modification of a truncated form of PCNA which is produced by usage of an internal start codon (Figure 22B, weak band at the bottom). In the absence of bacterially expressed GST‐APPBP1/UBA3, NEDDylation of PCNA might be accomplished by APPBP1/UBA3 which is present in the reticulocyte lysate used for in vitro translation (Figure 22B, lanes 2 and 5). In S phase of the cell cycle, PCNA can be SUMOylated or ubiquitinated at lysine 164 in order to avoid recombination or to facilitate translesion synthesis upon DNA damage or stresses accompanying replication (Arakawa et al., 2006; Hoege et al., 2002; Kannouche et al., 2004; Terai et al., 2010). Since for example the same lysine residues in the C terminus of p53 are used for the conjugation to both ubiquitin and NEDD8, we hypothesized that this might also be the case for PCNA (Rodriguez et al., 2000; Xirodimas et al., 2004). To test whether NEDD8 is conjugated to K164, we performed in vitro and in cellulo NEDDylation assays using a K164R mutant of PCNA. In the in vitro assay, we found the mutant to be as efficiently monoNEDDylated as wt PCNA suggesting that K164 is not required for NEDDylation of PCNA (Figure 22B, lanes 7 and 10). However, upon transfection of PCNA‐His‐SV5 wt or PCNA‐His‐SV5 K164R and HA‐ tagged NEDD8 or ubiquitin into H1299 cells, PCNA wt is both monoubiquitinated and ‐NEDDylated whereas PCNA K164R is not (Figure 22C). These data indicate that lysine residue 164 is at least necessary for NEDDylation of PCNA under overexpression conditions.

3.2.3 PCNA NEDDylation depends on the activity of APPBP1/UBA3 in cells

In 1998, NEDD8 was described to be activated by the ubiquitin‐activating enzyme UBA1 in vitro although around 100‐fold less efficiently than ubiquitin (Whitby et al., 1998). Some years later, this result was supported by the finding that mutation of R190 in UBA3 to glutamine, the corresponding residue in UBA1, does not significantly inhibit the activation of NEDD8. In contrast, ubiquitin cannot be used by the NEDD8 E1 enzyme because of clashing with R190 in UBA3 (Souphron et al., 2008). These data suggest that there is a strict limitation of the activation and transfer of ubiquitin to ubiquitin enzymes whereas NEDD8 might be used by ubiquitin enzymes under certain conditions. In 2012, this hypothesis was proven correct by two studies showing that a shift in the ratio of free ubiquitin and NEDD8 towards NEDD8 triggers its usage by UBA1 in cells (Hjerpe et al., 2012; Leidecker et al., 2012). Under diverse stress conditions such as inhibition of the proteasome or oxidative stress, free ubiquitin is depleted. At the same

- 58 - Results time, the amount of NEDDylated proteins increases being reversed by the knockdown of UBA1 but not by treatment with the NEDD8 E1 inhibitor MLN4924. In conclusion, an overexpression of NEDD8 at least partially leads to its activation by UBA1 thereby mimicking stress conditions in the cell (Hjerpe et al., 2012; Leidecker et al., 2012). To address whether monoNEDDylation of PCNA upon overexpression of His‐NEDD8 is mediated by the NEDD8‐ or the ubiquitin‐ conjugation system, we treated cells with the APPBP1/UBA3 inhibitor MLN4924 for 20 h (Figure 23). As functionality control for the inhibitor we tested its effect on the autoNEDDylation of HA‐ Ubc12. We found that treatment with MLN4924 diminished autoNEDDylation of overexpressed Ubc12 and that co‐expression of the deNEDDylating enzyme NEDP1 completely abolished its NEDDylation (Figure 23A). These data suggest autoNEDDylation to be mediated by APPBP1/UBA3 despite overexpression of His‐NEDD8. Concordantly, MLN4924 inhibited monoNEDDylation of PCNA and overexpression of NEDP1 further reduced the levels of NEDDylated PCNA (Figure 23B, asterisks). Interestingly, upon treatment with MLN4924 a faster migrating band than the one of the monoNEDDylated PCNA appeared in the input which might correspond to PCNA modified with endogenous ubiquitin (Figure 23B, circle). This result indicates that inhibiting APPBP1/UBA3 leads to a switch from NEDDylation to ubiquitination of PCNA.

A B

* * °

Figure 23. NEDDylation of PCNA is mediated by the NEDD8­conjugation system H1299 cells were transfected as indicated and treated with 1 µM MLN4924 for 20 h. Cells were harvested after 24 h and divided to prepare whole cell lysate (one third of the cells) and to conduct a Ni‐pulldown under denaturing conditions (A,B). PCNA presumably modified with endogenous ubiquitin is marked with a circle. NEDDylation of both HA‐Ubc12 (A) and endogenous PCNA (B; asterisk) is inhibited upon treatment with MLN4924.

- 59 - Results

3.2.4 NEDDylation of PCNA is enhanced by the E3 ligase Rad18

In the previous chapter, we showed that PCNA is NEDDylated in vitro and in cells and may therefore represent a new substrate for NEDD8 (Figure 22). In addition, it is assumed that transfer of NEDD8 to substrates is exclusively catalyzed by RING E3 ligases, and all of the so far identified NEDD8 E3 ligases are also capable of conjugating ubiquitin (Broemer et al., 2010; Huang et al., 2008; Morimoto et al., 2003; Oved et al., 2006; Xirodimas et al., 2004). Therefore, we tested whether Rad18, the E3 ligase for monoubiquitination of PCNA, might possess a dual specificity and affected PCNA NEDDylation in cells (Figure 24) (Hoege et al., 2002). Indeed, monoNEDDylation and monoubiquitination of PCNA were strongly enhanced upon overexpression of myc‐Rad18. Furthermore, co‐expression of NEDP1, a deNEDDylating enzyme, weakens modification of PCNA with NEDD8, but not with ubiquitin (Figure 24). These results suggest that Rad18 might act as E3 ligase for the monoNEDDylation of PCNA in cells.

Figure 24. Overexpression of Rad18 enhances NEDDylation of PCNA in cells H1299 cells were transfected with the indicated DNA and harvested after 24 h. One third of the cells were used to prepare whole cell lysate (input). With the remaining cells, Ni‐pulldown was performed under denaturing conditions. Western blot was conducted using α Myc‐ and α PCNA‐antibodies. Both monoNEDDylation and ‐ubiquitination of PCNA are enhanced upon overexpression of myc‐Rad18 but only NEDDylation is diminished after co‐transfection of His‐NEDP1.

3.2.5 NEDD8‐conjugating enzymes do not bind to PCNA in cells

PCNA has been described as a substrate for the ubiquitin‐like protein SUMO. Along these lines, the SUMO conjugating enzyme Ubc9, which is able to directly bind to several SUMO substrates, was found to interact with PCNA in a yeast two‐hybrid approach (Bernier‐Villamor et al., 2002; Gali et al., 2012; Hoege et al., 2002). To investigate whether PCNA also binds to the NEDD8‐conjugating enzymes Ubc12 or Nce2, we transfected myc‐tagged E2s into H1299 cells and performed an α PCNA‐IP. Myc‐Rad18 served as positive control (Figure 25). Due to the fact that there is a free pool of PCNA and a chromatin‐

- 60 - Results bound fraction, we divided cell lysate into a triton‐soluble and triton‐insoluble fraction. PCNA was found to interact with myc‐Rad18 in both fractions but neither Ubc12 nor Nce2 was co‐ immunoprecipitated in detectable amounts (Figure 25).

Figure 25. NEDD8 E2s presumably do not interact with PCNA in cells H1299 cells were transfected with the indicated DNA and harvested after 24 h. After dividing the cells into a Triton‐ soluble and a Triton‐insoluble fraction, 2 % of the cells were used as input. With the remaining cells, an α PCNA‐IP was performed. Western blots using α Myc antibody and α PCNA‐antibody show that PCNA binds to myc‐Rad18 but not to Ubc12‐myc‐His or Nce2‐myc‐His.

3.2.6 PCNA Y211F, a phosphorylation deficient mutant, is NEDDylated in cells

Taking primarily place in S phase and in the chromatin‐bound fraction, PCNA is a target for phosphorylation by the EGFR tyrosine kinase (Wang et al., 2006). Mutation of the tyrosine that is subjected to phosphorylation (Y211) leads to lower PCNA levels compared to wt PCNA which is due to increased polyubiquitination and degradation. Furthermore, polyubiquitination of PCNA Y211F is independent of the known ubiquitination machinery of PCNA (i.e. Rad6//Rad18 and Ubc13/Mms2//Rad5). Thus, phosphorylation of PCNA by EGFR is crucial for its stability and the execution of its functions, e.g. in DNA replication (Wang et al., 2006). To find out whether chromatin‐bound or free PCNA is NEDDylated, we performed an in cellulo NEDDylation and ubiquitination assay using PCNA Y211F (Figure 26). Our data show that the PCNA mutant is as efficiently monoNEDDylated and ‐ubiquitinated as PCNA wt, being enhanced by overexpression of myc‐Rad18. In addition, mutation of Y211 did not lead to polyubiquitination, which is confirmed by comparable input levels of PCNA wt‐SV5‐His and PCNA Y211F‐SV5‐His (Figure 26). These results indicate that free, chromatin‐unbound PCNA can be modified with ubiquitin and NEDD8. However, a conclusion about chromatin‐bound PCNA cannot be drawn.

- 61 - Results

Figure 26. PCNA Y211F is both NEDDylated and ubiquitinated H1299 cells were transfected with the indicated DNA and harvested after 24 h. One third of the cells were used to prepare whole cell lysate (input). With the remaining cells, Ni‐pulldown was performed under denaturing conditions. Western blots using α Myc‐, α PCNA‐ and α HA‐antibody show that both PCNA wt and Y211F are monoNEDDylated and monoubiquitinated.

3.2.7 Cu(I)‐catalyzed Huisgen azide‐alkyne cycloaddition as a tool to study functions of NEDDylated PCNA

Due to a very inefficient NEDDylation of PCNA under in vitro conditions, we decided to site‐ specifically link NEDD8 and PCNA using Cu(I)‐catalyzed Huisgen azide‐alkyne cycloaddition, abbreviated as “Click Reaction” (Figure 27) (reviewed in (Li, 2011)). This 1, 3‐dipolar cycloaddition between an alkyne‐ and an azide function results in the formation of a stable 1, 4‐disubstituted triazole. Hence, “monoNEDDylated PCNA” generated via click reaction provides a tool to investigate its functions. In order to perform the click reaction, non‐natural amino acids were incorporated into PCNA on the one hand, and into NEDD8 and ubiquitin on the other hand. We aimed at linking PCNA containing the propargyl‐protected lysine derivative Plk including the alkyne function to NEDD8 or ubiquitin containing azidohomoalanine with the azide function. Since we found that K164 of PCNA might be the lysine residue which is NEDDylated, we expressed PCNA comprising the artificial amino acid at this position. PCNA containing an amber stop codon at position 164 and the tRNA synthetase from Methanosarcina barkeri (pyrrolysine tRNA synthetase), which is able to recognize the amber stop codon, were therefore co‐transformed into E. coli BL21 (DE3). Upon induction of protein expression, pyrrolysine tRNA synthetase permits the incorporation of Plk at amino acid position 164 of PCNA. PlkPCNA was successfully expressed and purified using Q‐ sepharose (Figure 28A)(Eger et al., 2011).

- 62 - Results

Figure 27. Schematic representation of the click reaction between PCNA and NEDD8 or ubiquitin PCNA containing the propargyl‐protected lysine derivative Plk at amino acid position 164 is clicked with its alkyne function to the azide‐functionalized AhaNEDD8 or AhaUbiquitin in a Cu(I)‐catalyzed reaction, thereby forming a triazole linkage (modified from (Eger et al., 2010; Stoimenov and Helleday, 2009). For simplification, click reaction is shown for one monomer of PCNA.

Due to the usage of the methionine analogue azidohomoalanine (Aha) by the methionyl tRNA synthetase, NEDD8 and ubiquitin were expressed in methionine auxotrophic E. coli (Eger et al., 2011). To achieve site specific incorporation of Aha into the C termini of NEDD8 and ubiquitin, their Gly76 codon was mutated to a methionine codon. Moreover, the Met50 codon of NEDD8 was mutated to an alanine codon. To make sure that click reaction occurs exclusively at position 76, we inserted codons for small amino acids after the initial methionine codon so that Aha at the N terminus should be co‐translationally removed by the endogenous methionine aminopeptidase. Upon expression of His‐tagged AhaNEDD8 and ‐AhaUbiquitin, proteins were purified using Ni‐NTA‐agarose (Figure 28B).

A B

Coomassie Coomassie stain stain

Figure 28. Functionalized PlkPCNA, AhaUbiquitin and AhaNEDD8 are expressed in bacteria (A) PlkPCNA and pyrrolysine tRNA were co‐expressed in E. coli BL21(DE3). After lysis and ammonium sulfate precipitation, PlkPCNA (arrow) was purified using Q‐sepharose. After elution with different salt concentrations, the fraction with the highest purity and highest protein amount was dialyzed and loaded on SDS‐gel which was stained with Coomassie (Justus Hülse, under supervision). (B)AhaUbiquitin and AhaNEDD8 (arrow) were expressed in methionine auxotrophic bacteria using NMM media containing 0.5 mM azidohomoalanine (Aha). Proteins were purified with Ni‐NTA‐agarose. After elution, SDS‐PAGE was performed followed by Coomassie staining.

- 63 - Results

In a next step, the purified components were used in a click reaction (Figure 29). To find suitable conditions, increasing amounts of AhaUbiquitin or AhaNEDD8 were incubated with decreasing amounts of PlkPCNA in the presence of Cu(I) as catalyst, tris[(1‐benzyl‐1,2,3‐triazol‐4‐ yl)methyl]amine (TBTA) as Cu(I) stabilizing agent and tris(2‐carboxyethyl)phosphine (TCEP) as reducing agent (Eger et al., 2011). AhaUbiquitin was successfully linked to PlkPCNA as can be seen in lanes 3 and 4 (Figure 29). However, because of too many contaminating proteins in the AhaNEDD8 preparation (see “Input NEDD8”) it remains unclear whether it was linked to PlkPCNA (Figure 29, right part). Hence, the purification of AhaNEDD8 still needs to be improved to facilitate an efficient click reaction (see 4.2.2.4).

Figure 29. Click reaction between PlkPCNA and AhaNEDD8 or AhaUbiquitin Click reaction was performed using increasing amounts of PlkPCNA and AhaUbiquitin or AhaNEDD8. The highest amount used was loaded as input. After 1 h, reaction was stopped with 5x Laemmli loading buffer and loaded on SDS‐ gel which was finally stained with Coomassie. AhaUbiquitin was successfully clicked to PlkPCNA, whereas it is not clear whether a linkage was achieved between PlkPCNA and AhaNEDD8.

- 64 - Results

3.3 A new isoform of the NEDD8‐conjugating enzyme Nce2

In 2009, Nce2 was described as a second conjugating enzyme for NEDD8 in addition to Ubc12 (Huang et al., 2009). It shares around 40 % sequence identity with Ubc12, and contains a loop insertion next to the catalytic cysteine. Such an insertion of a flexible loop in proximity to the active site is also present in some ubiquitin E2 enzymes like Ubc7, for the conjugating activity of which two acidic residues in the loop are crucial (Ju et al., 2010). Nce2 prefers Cullin5 as substrate that is involved in cytokine signaling, neuronal migration or myogenesis, whereas Ubc12 rather NEDDylates Cullin1‐4 (Feng et al., 2007; Huang et al., 2009; Kile et al., 2002; Nastasi et al., 2004). However, up to now the specificity of Ubc12 and Nce2 for other NEDD8 substrates remains elusive. Thus, investigation of Nce2 functions is of considerable importance to gain further insights into the NEDD8‐conjugation system. Recently, the NEDO (New Energy and Industrial Technology Development Organization) human cDNA sequencing project, which focuses on splice variants of mRNAs, identified four new splice variants of Nce2 in glioma cells (www.uniprot.org). Except isoform 2, all other isoforms encode a shorter protein than the canonical isoform, and presumably do not display E2 activity (Figure 30A). The cDNA sequence of isoform 2 contains an insertion which is not present in isoform 1 and results from the usage of TG instead of an AG as alternative 3´ splicing site for the last exon (Figure 30B). The insertion leads to a frame shift and a 26 aa elongation of the translation product compared to isoform 1 (Figure 30C). Because of the extended C terminus of isoform 2, which might provide an additional interaction site for other cellular proteins, we decided to further investigate the properties of this isoform.

- 65 - Results

A

B

>Nce2 isoform2, cDNA ATGCTAACGCTAGCAAGTAAACTGAAGCGTGACGATGGTCTCAAAGGGTCCCGGACGGCAGCCAC AGCGTCCGACTCGACTCGGAGGGTTTCTGTGAGAGACAAATTGCTTGTTAAAGAGGTTGCAGAAC TTGAAGCTAATTTACCTTGTACATGTAAAGTGCATTTTCCTGATCCAAACAAGCTTCATTGTTTT CAGCTAACAGTAACCCCAGATGAGGGTTACTACCAGGGTGGAAAATTTCAGTTTGAAACTGAAGT TCCCGATGCGTACAACATGGTGCCTCCCAAAGTGAAATGCCTGACCAAGATCTGGCACCCCAACA TCACAGAGACAGGGGAGATATGTCTGAGTTTATTGAGAGAACATTCAATTGATGGCACTGGCTGG GCTCCCACAAGAACATTAAAGGATGTCGTTTGGGGATTAAACTCTTTGTTTACTGATCTTTTGAA TTTTGATGATCCACTGAATATTGAAGCTGCAGAACATCATTTGCGGGACAAGTCCCCAATGCTGT TACTCCACAGGAGGACTTCCGGAATAAAGTGGATGACTACATCAAACGTTATGCCAGATAATAAA AGGGGACGATTGCAGGCCCATGGACTGTGTTACAGTTTGTCTCTAACATGA

C

Figure 30. Isoforms of Nce2 (A) Schematic representation of predicted isoforms of Nce2. The NEDO sequencing project identified five isoforms of Nce2 that differ in length and partially in their amino acid sequence. Compared to isoform 1, isoform 2 contains a different and elongated C terminus, isoform 3 lacks a part of the E1 binding site and isoforms 4 and 5 are much shorter and differ in their C termini, too (www.uniprot.org). Catalytic cysteine (C116) is marked with a dark red line. (B) cDNA sequence of Nce2 isoform 2. In comparison to isoform 1, isoform 2 contains an insertion and an elongation at the 3´ end (both marked in blue) resulting in an extended translation product. (C) Alignment of Nce2 isoform 1 (Isoform 1) and Nce2 isoform 2 (Isoform 2) was conducted using ClustalW (www.ebi.ac.uk/Tools/msa/clustalw2/). The last common lysine residue is marked in red and the catalytic cysteine residue is marked in green. Nce2 isoform 2 contains an elongated C terminus which completely differs from isoform 1.

- 66 - Results

3.3.1 mRNA encoding Nce2 isoform 2 is expressed in HEK293T cells

To confirm the existence of Nce2 isoform 2 mRNA in cells, an RT‐PCR was performed (Figure 31). For this purpose, two different reverse primers were used: one primer anneals in the 3´ end of both the mRNA of isoform 1 and 2, the other only anneals in the unique insertion of isoform 2 (Figure 31 A). As shown in figure 31B, a PCR product was obtained with both primers (lanes 1 and 2) but not in the absence of reverse transcriptase or primers (lanes 3 and 4). PCR controls using vectors containing the respective cDNA as template prove the specificity of the primer used for isoform 2 amplification (lanes 5 and 6). Thus, mRNA of Nce2 isoform 2 is indeed present in HEK293T cells.

A B

* *

*

*

Figure 31. Isoform 2­specific mRNA is found in HEK293T cells (A) Schematic representation of primers used for reverse transcription of mRNAs of isoform 1 and 2 (grey arrows). One of the primers anneals in the 3´ UTR of isoform 1 and the 3´ end of isoform 2 mRNA, the other one specifically anneals in the insertion being only present in isoform 2. Expected sizes of PCR products are indicated above brackets. (B) 1 % agarose gel showing PCR products after reverse transcription with isoform 1‐ and 2‐specific primer (lane 1), isoform 2‐specific primer (lane 2), isoform 2‐specific primer without reverse transcriptase (lane 3) and without any primer (lane 4). PCR using cDNA of isoform 2 or isoform 1 as template and the isoform 2‐specific reverse primer was performed as control (lanes 5 and 6). Asterisks mark unspecific products in the PCR control. An RT‐PCR product was obtained with isoform 2‐specific primers confirming the existence of mRNA of Nce2 isoform 2 in cells.

3.3.2 Tertiary structure prediction for Nce2 isoform 2 reveals an unstructured, flexible C terminus

The structure of Nce2 isoform 1 is characterized by the catalytic core domain that is present in all conjugating enzymes of the ubiquitin and UBL field (Figure 32A) (Winn et al., 2004). Like Ubc12, Nce2 additionally contains an extension at the N terminus which is unique for NEDD8‐ conjugating enzymes and necessary for an efficient interaction with the E1 enzyme APPBP1/UBA3 (Huang et al., 2004; Huang et al., 2005). As already mentioned, the only major

- 67 - Results structural difference between Nce2 and Ubc12 is a loop insertion next to the catalytic cysteine of Nce2, whose function remains elusive (Huang et al., 2009). The N‐terminal 169 aa of the second isoform of Nce2 correlate to the first isoform and thus, they most likely adopt the same structure. The C terminus of Nce2 isoform 1, which differs from isoform 2 in the amino acid sequence, forms an α‐helix including four turns (Figure 32A). To gain an idea how the tertiary structure of the C‐terminal part of isoform 2 might look like, a prediction program was used which models the sequence of the protein with the unknown structure on an already known structure of a protein with a similar amino acid sequence (http://toolkit.tuebingen.mpg.de/hhpred). Hence, the tertiary structure of aa 170‐211 of isoform 2 was predicted as an α‐helix containing four turns which correlates to that of isoform 1, followed by a flexible, unstructured region and an additional small α‐helix (in collaboration with Prof. Dr. T. Frickey, University of Konstanz) (Figure 32B). A B

N

C

C

Figure 32. Structure of Nce2 isoform 1 (A) and structure prediction for the C terminus of isoform 2 (B) (A) Ribbon representation of the structure of Nce2 isoform 1 (aa 26‐185). Amino acids present in both isoforms are colored in green, the last common residue (K169) is shown as red sticks and the unique C‐terminal α‐helix of isoform 1 is marked in yellow (protein data base 2EDI, modeled using Pymol). (B) Ribbon representation of tertiary structure prediction of aa 159‐211 (C terminus) of Nce2 isoform 2. Colors as in (A) but the unique C terminus of isoform 2 (aa 170‐211) is depicted in yellow (structure prediction by http://toolkit.tuebingen.mpg.de/hhpred; modeled with Pymol). The C terminus of isoform 2 contains an α‐helix which is connected to a second, shorter α‐helix by a flexible, unstructured linker (in collaboration with Prof. Dr. T. Frickey, University of Konstanz).

3.3.3 Nce2 isoform 2 forms a thioester bond with NEDD8 but not with ubiquitin and is NEDDylated in vitro

Nce2 isoform 1 acts as a NEDD8‐conjugating enzyme which forms a thioester bond with NEDD8 and transfers it to a substrate (Huang et al., 2009). Since the catalytic cysteine at position 116 of isoform 1 is also present in isoform 2, its ability to form a thioester linkage with NEDD8 was tested. An in vitro thioester assay showed that C‐terminally His‐tagged Nce2 isoform 2 indeed is capable of forming a thioester bond with GST‐NEDD8 but not with GST‐ubiquitin, suggesting a role for Nce2 isoform 2 in the NEDD8‐conjugation cascade (Figure 33A, asterisk).

- 68 - Results

Because both Ubc12 and Nce2 are autoNEDDylated thereby enhancing their affinity to APPBP1/UBA3 (see chapter 3.1), the question arose whether the second isoform of Nce2 behaves the same. In an in vitro NEDDylation assay, Nce2 isoform 2 wt was monoNEDDylated in the presence of APPBP1/UBA3 and NEDD8 (Figure 33B). However, the catalytically inactive mutant C116A was still NEDDylated, indicating that NEDDylation of isoform 2 occurs at least in part not via autoNEDDylation. The modified C116S mutant of Nce2 isoform 2 displays a different migration behavior in the SDS‐gel, pointing out that an oxyester was formed with NEDD8, again showing Nce2 isoform 2 to be catalytically active (Figure 33B; own observations).

A

B

*

Figure 33. Nce2 forms thioester linkage with NEDD8 (A) and is NEDDylated in vitro (B) (A) Radiolabeled Nce2 isoform 2‐His (Nce2 Iso2‐His) was incubated with indicated bacterially (GST‐tagged ubiquitin, NEDD8 and APPBP1/UBA3) or baculovirally (UBA1) expressed proteins under thioester conditions. Samples were divided in an afterwards reduced (+DTT) and not reduced part (‐DTT). A thioester linkage is formed with GST‐NEDD8 (asterisk) but not with GST‐ubiquitin. (B) Radiolabeled Nce2 isoform 2‐His (Nce2 Iso2‐His) wt, C116S or C116A were incubated with indicated bacterially expressed proteins under NEDDylation conditions. Isoform 2 wt and C116A are modified with NEDD8 (asterisk) whereas isoform 2 C116S forms an oxyester with NEDD8.

- 69 - Results

3.3.4 Nce2 isoform 2 is NEDDylated and ubiquitinated in cells

To gain insights into the function and regulation of Nce2 isoform 2 in cells, NEDDylation and ubiquitination assays were performed using overexpressed HA‐tagged Ubc12, HA‐Nce2 isoform 1, HA‐Nce2 isoform 2 and His‐NEDD8 (Figure 34). As shown in chapter 3.1, Ubc12 and Nce2 isoform 1 were found to be modified with NEDD8. Interestingly, neither Ubc12 nor Nce2 isoform 1 were ubiquitinated whereas Nce2 isoform 2 was both ubiquitinated and NEDDylated (Figure 34). Hence, the additional C terminus of isoform 2 which is not present in isoform 1 might provide the ubiquitination site.

Figure 34. Nce2 isoform 2 is modified with NEDD8 and ubiquitin in cellulo H1299 cells were transfected with the indicated DNA and harvested after 24 h. One third of the cells were used for the preparation of whole cell lysate (input). With the remaining cells, Ni‐pulldown was performed under denaturing conditions. Western blots using α HA‐antibody reveal that Ubc12 and Nce2 isoform 1 are NEDDylated and Nce2 isoform 2 is both NEDDylated and ubiquitinated.

To further investigate Nce2 isoform 2 in comparison to isoform 1, we tested whether Nce2 isoform 2 is autoNEDDylated in cellulo. Therefore, catalytically inactive mutants of isoform 2 (C116S and C116A) were transfected into H1299 cells (Figure 35A). His‐NEDD8 was conjugated to these mutants with the same efficiency as to wt isoform 2, confirming the in vitro result that the second isoform of Nce2 is NEDDylated but probably not capable of autoNEDDylation (Figure 35A). The N termini of Ubc12 and Nce2 are involved in the interaction with the E1 enzyme (Huang et al., 2009; Huang et al., 2004). In addition, we could show that it comprises the lysine residues which are NEDDylated in an autoNEDDylation reaction (see chapter 3.1). Hence, an isoform 2 deletion mutant lacking the N‐terminal 26 aa was included in the NEDDylation assay in cellulo (Figure 35A). In contrast to Ubc12 and Nce2 isoform 1, deletion of the N terminus had a minor influence on the modification of isoform 2 with NEDD8 (compare figures 15C and 35A). However, the double bands appearing for Nce2 isoform 2 wt, C116S and C116A, which might

- 70 - Results correspond to two different lysine residues that are NEDDylated, are not detected for the ΔN26 mutant, supporting the idea that Nce2 isoform 2 may in part be NEDDylated in the N terminus (Figure 35A, asterisks). In figure 11, we showed NEDDylated Ubc12 and Nce2 isoform 1 to be present in the cytosol of H1299 cells. To find out whether isoform 2 is also modified with endogenous NEDD8, an α HA‐IP was performed using cytosolic fractions of H1299 cells after transfection of HA‐tagged Nce2 isoform 1 or isoform 2 (Figure 35B). Both isoforms were efficiently immunoprecipitated as shown in the western blot with α HA‐antibody (Figure 35B, circles left panel), where monoNEDDylated Nce2 isoform 1 was also detected (Figure 35B, arrow left panel; note: Nce2 isoform 1 runs at the same height as the light chain of the α HA‐antibody). Western blot analysis with α NEDD8‐antibody also revealed the first but not the second isoform of Nce2 to be modified with endogenous NEDD8, at least not with the same efficiency (Figure 35B, arrows right panel). A

* * *

B

° ° *

Figure 35. Presumably, Nce2 isoform 2 is not capable of autoNEDDylation (A) and not modified with endogenous NEDD8 (B) (A) H1299 cells were transfected with the indicated DNA and harvested after 24 h. One third of the cells were used to prepare whole cell lysate (input). With the remaining cells, Ni‐pulldown was performed under denaturing conditions. Western blots using α HA‐antibody shows catalytically inactive mutants of Nce2 isoform 2 to be NEDDylated in cells. Moreover, Nce2 isoform 2 lacking its N‐terminal 26 aa is still modified with NEDD8. Asterisks mark double bands presumably corresponding to different lysine residues that are NEDDylated. (B) HA‐Nce2 isoform 1 or HA‐Nce2 isoform 2 were overexpressed in H1299 cells. Upon hypotonic lysis, 5 % of lysate were used as input and an α HA‐IP was performed with the cytosolic fraction. Proteins were first detected using western blot with α NEDD8‐ (Alexis) and in a second step with α HA‐antibody. Asterisk marks light chain of the antibody used for IP. Circles mark unmodified HA‐Nce2 isoform 1 and 2. While Nce2 isoform 1 is modified with endogenous NEDD8 (arrows), NEDDylation of Nce2 isoform 2 was not detected.

- 71 - Results

In conclusion, the alteration in the C‐terminal amino acid sequence seems to completely change the properties of Nce2 isoform 2 in comparison to isoform 1 regarding NEDDylation and ubiquitination. The second isoform is probably not autoNEDDylated and also not predominantly NEDDylated in its N terminus. Instead, it is ubiquitinated which might have an influence on its half‐life.

3.3.5 Nce2 isoform 2 has a shorter half‐life than isoform1 and is degraded by the proteasome

In the experiment to figure 34, we showed that the second isoform of Nce2 can be modified with ubiquitin in overexpression experiments. To determine whether the half‐life of isoform 2 may be affected by ubiquitination, we performed a cycloheximide chase in the presence or absence of the proteasome inhibitor MG132 (Figure 36). Nce2 isoform 1 remained relatively stable after 4 h of cycloheximide treatment and the addition of MG132 had no observable effect on its protein levels (Figure 36A). In contrast, protein levels of isoform 2 rapidly decreased and MG132 partially restored them (Figure 36A). Thus, the half‐life of Nce2 isoform 2 was calculated to be approximately 2.7 h. Mutation of the catalytic cysteine of isoform 2 had no effect on its half‐life (Figure 36B), indicating that its catalytic activity is not required for degradation. Supporting this finding, we did not see a change in ubiquitination efficiency when comparing Nce2 isoform 2 wt and its cysteine mutants in H1299 cells (data not shown). Since Nce2 isoform 2 contains two lysine residues in its unique C terminus, we proposed that one of them might be modified with ubiquitin leading to proteasomal degradation. However, neither mutation of one lysine residue nor of both residues to arginine had a major influence on the modification with ubiquitin or the half‐life of isoform 2 (data not shown), suggesting that other lysines, which are also present in isoform 1, are used for the attachment of ubiquitin.

- 72 - Results

A

B

Figure 36. Nce2 isoform 2 has a shorter half­life than isoform 1 that is not affected by mutation of the catalytic cysteine (A) H1299 cells were transfected with HA‐Nce2 isoform 1 or HA‐Nce2 isoform 2. Upon treatment with 60 µg/mL cycloheximide (CHX), cells were harvested after 0, 1, 2 or 4 h. In addition, cells were subjected to CHX and 10 µM MG132 and harvested after 4 h. After TNN lysis, samples were loaded on SDS‐gel. Proteins were detected using western blot and α HA‐antibody (left panel). Isoform 2 is rapidly degraded and protein levels are restored by MG132. To calculate half‐lives, protein levels of HA‐Nce2 isoform 1 and isoform 2 were quantified (right panel). Isoform 2 has a shorter half‐life than isoform 1. (B) Quantification of protein levels of HA‐Nce2 isoform 2 wt, C116S and C116A. H1299 cells were transfected with HA‐Nce2 isoform 2 wt, C116S or C116A. Upon CHX and MG132 treatment, cells were harvested at indicated time points and lysed with TNN. Samples were subjected to SDS‐PAGE and western blot α HA and protein levels were quantified. Mutation of the catalytic cysteine does not affect half‐life of isoform 2. Note that (A) and (B) were done in triplicates and representative results are shown.

- 73 - Results

3.3.6 Nce2 isoform 1 and 2 differ in their subcellular localization

To shed light on possible functions of the second isoform of Nce2, we examined the cellular localization of the protein using immunofluorescence as well as cellular fractionation assays (Figure 37). In immunofluorescence experiments with overexpressed HA‐tagged Nce2, we found isoform 1 to be localized mainly in the cytoplasm but also in the nucleus (Figure 37A). For isoform 2, a similar distribution was found in most cases (Figure 37B, middle and right panel). However, a strong nuclear staining and aggregates (arrow) were visible in a minor percentage of H1299 cells (Figure 37B, left and right panel).

A B

C

Figure 37. Overexpressed HA­Nce2 isoform 2 localizes more to the nucleus than isoform 1 (A) Immunofluorescence of HA‐tagged Nce2 isoform 1 (red or green) in H1299 cells using α HA‐antibody. Nuclei were stained with DAPI (blue; scale bar 20 µm). (B) Immunofluorescence of HA‐tagged Nce2 isoform 2 (red or green) in H1299 cells using α HA‐antibody. Nuclei were stained with DAPI (blue; scale bar 20 µm). (C) Cellular fractionation of H1299 cells. HA‐tagged isoform 1 or HA‐tagged isoform 2 were transfected into H1299 cells. Cells were divided into a cytosolic fraction, a nuclear fraction containing soluble proteins (nucleosol) and several fractions containing structure‐bound nuclear proteins. Input corresponds to whole cell lysate. Proteins were detected using western blot and α HA‐antibody. In contrast to Nce2 isoform 1, isoform 2 is found in several nuclear fractions of H1299 cells.

- 74 - Results

These findings were confirmed in cellular fractionation assays using the same cell line (Figure 37C). HA‐tagged Nce2 isoform 1 was mainly detected in the cytosolic and the soluble nuclear fraction (nucleosol) whereas the second isoform was additionally present in several nuclear fractions containing structure‐bound proteins, and the pellet fraction (Figure 37C). These results indicate that the two isoforms are present in partially distinct cellular compartments and thus, most likely execute different functions. Additionally, we detected NEDDylated Nce2 isoform 1 but no modified isoform 2 in the cytosolic fraction of H1299 cells providing a hint that Nce2 isoform 2 behaves different from the first isoform and might be differentially regulated (Figure 37C, compare Figure 35B).

3.3.7 Development of an antibody specifically recognizing Nce2 isoform 2

To prove the existence of endogenous Nce2 isoform 2 in cells, a specific antibody was developed. Since isoform 1 and 2 only differ in their C terminus, a KLH (Keyhole Limpet Hemocyanin)‐ coupled peptide corresponding to aa 187‐200 of isoform 2 was ordered from Genescript (Figure 38A). After testing preimmune sera of several rabbits, the one giving the lowest background in a western blot was chosen for four immunizations with the peptide (performed by TFA, University of Konstanz). As shown in figure 38B, the obtained purified antibody is specific for bacterially expressed Nce2 isoform 2 and does not recognize isoform 1. Moreover, around 5 ng of protein were necessary for detection by the antibody, as determined by Bradford measurement (data not shown). In order to test the suitability of the antibody for IP, we overexpressed HA‐Nce2 isoform 1 and isoform 2 in HEK293T cells and performed an IP using the α isoform 2‐antibody as well as the α HA‐antibody as a control (Figure 38C). After immunoprecipitation, we divided the samples and subjected them to western blots using the α HA‐ as well as the α Nce2 isoform 2‐antibody. With the α HA‐antibody, we successfully immunoprecipitated both overexpressed isoforms (Figure 38C, left panel; note that isoform 2 runs exactly at the same size as the light chain of the α HA‐antibody‐the difference in running behavior compared to figure 35B is explained by different lysis methods). In addition, HA‐Nce2 isoform 2 but not isoform 1 was detected with the Nce2 isoform 2‐specific antibody (Figure 38C, arrow right panel). Using this antibody for IP, we specifically immunoprecipitated the overexpressed HA‐Nce2 isoform 2 (Figure 38C, arrow left panel). Although we treated the cells used for IP with the proteasome inhibitor MG132 to enrich for Nce2 isoform 2, we did not detect the endogenous second isoform of Nce2 upon α isoform 2‐IP and subsequent western blot using the same antibody (Figure 38C, right panel). As determined in a further western blot, the three weak bands visible in this blot at the expected size of Nce2 isoform 2 (~24 kDa) correspond to the light chains of the antibody used for IP (Figure 38C, data not shown). In summary, we developed an antibody specifically

- 75 - Results recognizing the second but not the first isoform of Nce2. However, this antibody is functional but not well suitable for IP experiments. Thus, we were not able to prove the existence of endogenous Nce2 isoform 2 in cells, probably being due to a low abundance of the protein in the cell.

A B

C

*

Figure 38. An antibody raised against aa 187­200 of Nce2 isoform 2 specifically recognizes isoform 2 in vitro and in cells (A) Ribbon representation of predicted tertiary structure of aa 159‐211 of Nce2 isoform 2 including the peptide sequence used for immunization. Common sequence of isoform 1 and 2 is shown in green, the last common lysine residue as red sticks, the unique C terminus of isoform 2 is depicted in yellow and the peptide used for immunization in cyan. (B) Increasing amounts of bacterially expressed Nce2 isoform 2‐His and Nce2 isoform 1‐His were subjected to SDS‐ PAGE followed by colloidal Coomassie staining (upper panel) or western blot using α isoform 2‐antibody (lower panel). The antibody raised against isoform 2 specifically recognizes Nce2 isoform 2. (C) HA‐Nce2 isoform 1 and HA‐Nce2 isoform 2 were transfected into HEK293T cells. After 20 h, cells were treated with 10 µM MG132 for 4 h. After TNN lysis, lysate was divided and subjected to an α HA‐IP or to an α Nce2 isoform 2‐ IP (=iso2‐IP). 2 % lysate were loaded as input. HA‐Nce2 isoform 2 runs at the same height as the light chain of the α HA‐antibody (left panel, asterisk). The α isoform 2‐antibody specifically immunoprecipitates overexpressed HA‐Nce2 isoform 2 (arrow left panel) and specifically detects overexpressed isoform 2 (arrow right panel).

- 76 -

4. Discussion

4.1 AutoNEDDylation as a regulatory mechanism of NEDD8‐conjugating enzymes

Until to date, not much is known about if and how the enzymes involved in the NEDDylation cascade are regulated. In contrast, for the ubiquitination‐, the SUMOylation‐ as well as the FAT10ylation system, an autoregulatory mechanism on the level of the E2 enzymes has been described (Aichem et al., 2012; Knipscheer et al., 2008; Machida et al., 2006; Yamanaka et al., 2000). During this study, we found the NEDD8‐conjugating enzymes Ubc12 and Nce2 to be autoNEDDylated and thereby provide potential insights into the regulation of the NEDDylation machinery.

4.1.1 NEDD8‐conjugating enzymes are autoNEDDylated in their unique N terminus

4.1.1.1 Indications for endogenous autoNEDDylation of Ubc12 and Nce2 Initially, we found NEDD8‐conjugating enzymes to be capable of NEDDylating themselves in vitro and in cells being dependent on the catalytic cysteine residue and the activity of APPBP1/UBA3 (Figures 9, 10 and 11). These results show that it is indeed an autoNEDDylation reaction which is catalyzed by enzymes of the NEDD8‐ and not the ubiquitin cascade (see 3.2.3). AutoNEDDylation occurs in a cis‐mechanism, as shown in previous studies (Ph.D. thesis A. Rojas Fernández, University of Konstanz: “Regulation of Hdm2/HdmX‐mediated ubiquitination and neddylation”). Overexpressed HA‐tagged E2 enzymes are both modified with overexpressed His‐ NEDD8 as well as with endogenous NEDD8, indicating that the NEDDylation of E2s is not an artifact of NEDD8 overexpression (Figures 10C and 11A). Moreover, stably overexpressed HA‐ Nce2 is also NEDDylated with endogenous NEDD8 (data not shown). Notably, the HA‐tag at the

- 77 - Discussion

N terminus of Ubc12 and Nce2 inhibits acetylation of the N‐terminal methionine which otherwise might affect autoNEDDylation (see 4.1.1.5)(Monda et al., 2013). As antibodies with high sensitivity against for Ubc12 or Nce2 are not available, we confirmed the existence of NEDDylated endogenous Ubc12 by Ni‐NTA pulldown from MCF‐7 cells stably expressing His‐NEDD8 (data not shown). Moreover, by using an α NEDD8‐IP we identified a band of ~30 kDa in an α NEDD8‐western blot, which disappeared upon addition of the APPBP1/UBA3‐inhibitor MLN4924 (Figure 11B). This band slightly shifted up when HA‐Ubc12 was transfected. Therefore, we conclude that the observed band corresponds to monoNEDDylated endogenous Ubc12 (NEDD8: 9 kDa; Ubc12: ~21 kDa), and that the shift in the presence of HA‐Ubc12 results from the additional 1.5 kDa of the HA‐tag. Interestingly, we found E2 enzymes modified with endogenous NEDD8 only in the cytoplasm of either H1299 or HEK293T cells under native lysis conditions (Figure 11). The fact that Ubc12 and Nce2 are not deNEDDylated in spite of native lysis conditions and the presence of the deNEDDylating enzyme NEDP1 in the cytoplasm points to a stabilization of NEDDylated E2 enzymes by a yet unknown factor (Sundqvist et al., 2009). Additionally, NEDDylation of the E2 enzymes itself might be a signal for cytoplasmic localization. However, fusions of NEDD8 to the N terminus of Ubc12 and Nce2, which were established to mimic NEDDylation of the E2 enzymes, are found both in the cytoplasm and in nuclear fractions (see 4.1.1.4; figure 19). To investigate whether NEDD8‐conjugating enzymes are NEDDylated in the cytoplasm or in the nucleus and subsequently transported to the cytoplasm, a nuclear export inhibitor such as Leptomycin B could be used. In principle, both scenarios are possible since the majority of NEDD8 is located in the nucleus and a smaller fraction is present in the cytoplasm (Kamitani et al., 1997). Alternatively, autoNEDDylation of Ubc12 and Nce2 could be a co‐translational event. This hypothesis is underlined by the finding that we detected strongly NEDDylated E2 enzymes upon overexpression of HA‐tagged E2s, but we hardly found NEDDylated E2s in cell lines stably expressing HA‐tagged Ubc12 or Nce2 (Figure 11A, data not shown). Even upon overexpression of His‐NEDD8, we could not enrich for NEDDylated forms of stably expressed HA‐tagged E2 enzymes (data not shown). In addition, transfection of HA‐Ubc12 completely inhibited NEDDylation of endogenous Ubc12, as there is only one band detected upon α NEDD8‐IP (Figure 11B). To further investigate this issue, it would be worthwhile to check for levels of NEDDylated E2s upon inhibition of translation and protein degradation at the same time.

4.1.1.2 The HPNI­motif of Ubc12 and Nce2 is important for isopeptide bond formation In 2003, it was shown that an asparagine in a conserved HPNI/V motif in proximity to the catalytic cysteine of conjugating enzymes is crucial for the formation of isopeptide bonds between ubiquitin or SUMO and their respective substrates. In contrast, the asparagine is

- 78 - Discussion dispensable for thioester formation and proper folding of the E2 enzymes. It is speculated that this residue stabilizes the oxyanion intermediate formed during lysine attack on the UBL‐loaded E2 (Wu et al., 2003b). In the present studies, we found that the asparagine is also conserved in NEDD8 E2 enzymes. We additionally confirmed that it is necessary for the conjugation of NEDD8 to Ubc12 and Nce2, but its mutation does not affect thioester formation (Figure 12). Hence, the conserved asparagine residue in the HPNI/V motif of conjugating enzymes very likely plays an important general role in facilitating isopeptide bond formation between UBLs and their substrates. Importantly, with the identification of Ubc12 and Nce2 as NEDD8 substrates in these and former studies, we provide an easy system to study effects of mutations in NEDD8 and its activating or conjugating enzymes in vitro (Ph.D. thesis A. Rojas Fernández, University of Konstanz: “Regulation of Hdm2/HdmX‐mediated ubiquitination and neddylation”). Related to this topic, the function of a flexible loop insertion next to the catalytic cysteine of Nce2 which is not present in Ubc12 should be examined. Such a loop is present in several other conjugating enzymes such as UbcH3 and yeast Ubc7 (Ju et al., 2010; Ptak et al., 1994). The acidic residues in the flexible loop of Ubc7 were shown to be crucial for its activity as they might play a role in the polarization of the histidine in the HPNI/V motif or to guide the lysine to be modified into the active site (Ju et al., 2010). The acidic loop of UbcH3, however, is involved in the formation of K48‐chains: the interaction of the loop with a ubiquitin moiety which is attached to a substrate orientates it in a way to accept another ubiquitin (Gazdoiu et al., 2007; Petroski and Deshaies, 2005b). Since Nce2 also contains an acidic residue in its loop, it would be interesting to elucidate whether a mutation of the aspartate (D126) in the loop insertion or a deletion of the whole loop influences the conjugating activity of Nce2.

4.1.1.3 The N termini of Ubc12 and Nce2 are crucial for an efficient autoNEDDylation To gain a first hint about the function of autoNEDDylation, we aimed at identifying the NEDDylated lysine residue(s). In vitro NEDDylation assays and subsequent mass spectrometric analyses showed lysine 3, 8, 11, 36 and 45 of Ubc12 to be modified and lysine 7, 9 and 15 of Nce2 (Figure 13). When overexpressing HA‐Ubc12 K3R/K8R/K11R and HA‐Nce2 K7R/K9R together with His‐NEDD8 in cells, NEDDylation of these E2 mutants was slightly diminished compared to wt enzymes (Figure 15B). These results are in line with our in vitro data that the mutated lysines are targets for NEDDylation but that there are additional lysines modified with NEDD8 (Figure 13). Strikingly, most of the lysine residues we identified to be NEDDylated, are located in the unique N‐terminal extension of NEDD8‐conjugating enzymes which is not present in most other E2 enzymes and which is involved in the interaction with APPBP1/UBA3 (Figure 13) (Huang et al., 2004). Our findings regarding the NEDDylation sites within Ubc12 are in line with a recent publication, in which lysine 8, 11, 12, 21, 25 and 26 were identified to be modified with NEDD8 in vitro.

- 79 - Discussion

However, these data suggest that Ubc12 is modified by NEDD8 chains for which our data adduces no evidence (Jeram et al., 2010). Our results rather indicate that NEDD8 E2 enzymes might be modified with single NEDD8 moieties at multiple lysine residues (Figure 13, data not shown). The difference in the identified lysine residues between our data and the publication might be due to different digestion methods of the proteins prior to mass spectrometric analysis. We digested proteins with trypsin, which cleaves C‐terminal of lysines and arginines. Because of many lysines in the N terminus of Ubc12, some peptides might be too small to be detected and therefore, we might have missed some modification sites. To confirm the NEDDylation sites in the N termini of the NEDD8 E2 enzymes in cells, further mass spectrometric analyses need to be conducted. Upon IP of NEDDylated Ubc12 and Nce2 from cells, immunoprecipitated proteins should be digested using Lys‐C instead of trypsin. Tryptic digest generates a double glycine attached to a lysine residue both for ubiquitin and NEDD8 and therefore, it is not suitable to discriminate between a modification with NEDD8 or ubiquitin. To further investigate the N‐terminal extension of Ubc12 and Nce2 with respect to an allocation of NEDDylation sites, we performed assays with deletion mutants lacking the first 26 aa. In vitro studies showed that these mutants are still capable of forming a thioester bond with NEDD8 (Figure 14B), but they are impaired in NEDDylating themselves (Figure 14A). In cells, NEDDylation of HA‐Ubc12 and HA‐Nce2 was significantly decreased upon deletion of the N terminus. Nonetheless, there was some residual transfer of His‐NEDD8 to the E2 mutants lacking the N‐terminal extension (Figure 14C). NEDDylation of Ubc12 despite the lack of the first 26 aa can be explained by the finding that lysines 36 and 45 were also identified as conjugation sites for NEDD8 in vitro (Figure 13). For Nce2 ΔN26 yet uncovered NEDDylation sites must exist in the core domain since it was still NEDDylated in cells (Figure 14C). In 2004, Huang et al. reported that the N terminus of Ubc12 directly and selectively binds to

APPBP1/UBA3. Comparison of Ubc12 and Ubc12 ΔN26 revealed a similar kcat for the formation of a thioester bond with NEDD8 in vitro, but a ~25 fold higher Km for Ubc12 ΔN26. However, Ubc12 ΔN26 was not impaired in the transthiolation reaction and the transfer of NEDD8 to substrates under in vitro conditions (Huang et al., 2004). Under the conditions we used Ubc12 ΔN26 and Nce2 ΔN26 were not impaired in thioester formation compared to wt enzymes (Figure 14B). This result might be due to a high level of APPBP1/UBA3 applied in the assay which probably compensates the defect in affinity (Huang et al., 2004). Thus, the defect in autoNEDDylation of the E2s ΔN26 seen in figure 14A does not result from the partial lack of the interaction site with UBA3. Nonetheless, the question arises whether the autoNEDDylation defect of Ubc12 ΔN26 and Nce2 ΔN26 in cells results from a limited amount of APPBP1/UBA3 or indeed from the deletion of the NEDDylation sites (Figure 14C). To investigate this issue, all lysines in the N termini of Ubc12 and Nce2 could be mutated, and the resulting mutants could be tested for their ability to be modified with NEDD8 within cells. If they would still be NEDDylated

- 80 - Discussion with a comparable efficiency as wt E2 enzymes, it would indicate that the N terminus is necessary for mediating the NEDDylation rather than to provide the NEDDylation sites. Along these lines, we fused the N terminus of either Ubc12 or Nce2 to a ubiquitin‐conjugating enzyme (UbcH5b). In a ubiquitination assay, these fusion proteins were indeed efficiently ubiquitinated whereas UbcH5b wt was not (Bachelor thesis S. Kabelitz, University of Konstanz: “Charakterisierung von NEDD8‐ und Ubiquitin‐konjugierenden Enzymen”). Thus, the N‐terminal extension of NEDD8‐conjugating enzymes generally seems to be a good target for NEDDylation or ubiquitination, which may be due to its flexibility and its high content of lysine residues.

4.1.2 Possible functions of autoNEDDylation

As mentioned above, automodifications of conjugating enzymes within the ubiquitin‐, SUMO‐ and FAT10 system have already been described before. Inhibition of activity by autoubiquitination of a lysine residue in proximity to the catalytic cysteine was shown both for UbcH10, an E2 enzyme important for cell cycle regulation, and Ube2t, which is involved in DNA repair (Lin et al., 2002; Machida et al., 2006; Yamanaka et al., 2000). Furthermore, modification of the SUMO conjugating enzyme Ubc9 at lysine 14 in its N terminus has an influence on its substrate specificity (Knipscheer et al., 2008). The exact function of USE1 autoFAT10ylation, however, remains unclear. It is speculated that USE1 FAT10ylates itself in the absence of an E3 ligase, thereby changing its activity, function or interaction partners (Aichem et al., 2012). Thus, one could hypothesize similar outcomes of autoNEDDylation of NEDD8‐conjugating enzymes, which we investigated during these studies (Figure 39).

Figure 39. Possible impacts and functions of E2 autoNEDDylation The modification of either Ubc12 or Nce2 (E2) with NEDD8 at a lysine residue (Lys) in the N terminus might affect their localization, interactions with E1 enzymes or E3 ligases, substrate specificity or NEDD8 conjugation in general.

- 81 - Discussion

4.1.2.1 AutoNEDDylation as regulator of the subcellular localization of Ubc12 and Nce2 Fusions of UBLs to their substrate are widely used as a tool to study function and localization of modified substrates. For instance, the localization of p53, which can be modified by ubiquitin, NEDD8 and SUMO‐1, has been investigated using fusion proteins (Carter and Vousden, 2008). In general, NEDDylated substrates are not only found in the nuclear portion of the cell but also in the cytoplasm (Furukawa et al., 2000; Meyer‐Schaller et al., 2009; Sundqvist et al., 2009). In order to find out whether autoNEDDylation of Ubc12 and Nce2 has an influence on their subcellular localization, we investigated the localization of NEDD8‐E2 fusion proteins using immunofluorescence and cellular fractionation assays. In immunofluorescence experiments, there was no observable difference in the localization of the fusion proteins and the wt conjugating enzymes: both were mainly present in the cytoplasm under overexpression conditions (Figure 18). In the cellular fractionation assay, overexpressed HA‐Ubc12 mainly localized to the cytoplasm and to a minor percentage to the nucleus (Figure 19). In contrast, HA‐Nce2 was only detected in the cytoplasmic fraction. Upon fusion of NEDD8 to the N termini of the conjugating enzymes, both Ubc12 and Nce2 showed an enhanced localization to the nucleus (Figure 19). NEDD8 itself localizes mainly in the nucleus and might therefore contribute to a change in localization of its conjugating enzymes (Kamitani et al., 1997). Such a change in localization might point to a different function that NEDD8‐E2 fusions and thus, maybe also autoNEDDylated E2 enzymes may have in the nucleus. Along these lines, we tested whether DNA damage leads to a localization of Ubc12, Nce2 or the NEDD8‐E2 fusion proteins to the nucleus. Upon cisplatin treatment, none of these enzymes changed its localization, suggesting that autoNEDDylation is not involved in the DNA damage response (data not shown). Taken into consideration that we identified endogenously NEDDylated E2 enzymes only in the cytoplasm, overexpression experiments might not perfectly reflect the endogenous situation in this case (Figure 11).

4.1.2.2 Acetylation of NEDD8 E2 enzymes as a competitive modification to NEDDylation Mass spectrometric data about post‐translational modifications of Ubc12 and Nce2 in the AML cell line MV4‐11 revealed K3 of Ubc12 and K7 of Nce2 to be acetylated (Choudhary et al., 2009). These two lysine residues overlap with the ones we found to be monoNEDDylated in vitro, suggesting that the function of NEDD8‐conjugating enzymes is regulated by different modifications which at least in part compete for the same lysine residues (Figure 13). Furthermore, the N‐terminal methionine of Ubc12 and Nce2 can be co‐translationally acetylated thereby increasing the affinity to DCN‐like (DCNL) E3 ligases which serve as scaffold for the NEDDylation of cullins. Since the unique N terminus of NEDD8‐conjugating enzymes contributes to the interaction with the PONY domain of DCNL, autoNEDDylation of N‐terminal lysine residues most likely interferes with DCNL binding (Kurz et al., 2008; Monda et al., 2013). To

- 82 - Discussion obtain evidence about the function of autoNEDDylation, it should be tested whether it indeed affects DCNL‐mediated NEDDylation of cullins and whether N‐terminal acetylation and NEDDylation in the respective N terminus of Ubc12 or Nce2 can occur simultaneously or are mutually exclusive. There are already some indications that methionine acetylation does not generally interfere with autoNEDDylation: upon overexpression of C‐terminally tagged Ubc12 and Nce2, we still detected autoNEDDylation with overexpressed NEDD8 (data not shown). Furthermore, a publication about a screen for ubiquitinated proteins revealed both an acetyl group at the N‐terminal methionine of Ubc12 and a diglycine at lysine 3. Nonetheless, it is not clear which UBL was attached to K3 since tryptic digest of both ubiquitin and NEDD8 leads to a double glycine motif attached to a lysine residue (Wagner et al., 2011).

4.1.2.3 AutoNEDDylation enhances the affinity of NEDD8 E2s to APPBP1/UBA3 The structure of the N terminus of Ubc12 bound to APPBP1/UBA3 demonstrates that the N‐terminal extension, especially the hydrophobic residues F5 and L7, interacts with the adenylation domain of UBA3 (Huang et al., 2004). Moreover, there is evidence that mutation of K12 within Ubc12 decreases the relative binding affinity to the E1 enzyme (Huang et al., 2004). Hence, a modification of this or the adjacent lysine residue with NEDD8 might impair the interaction with APPBP1/UBA3 (Figure 13) (Jeram et al., 2010). To investigate whether autoNEDDylation in the N terminus of Ubc12 and Nce2 has an effect on the affinity to APPBP1/UBA3, we performed a GST‐pulldown assay. Surprisingly, autoNEDDylated NEDD8‐ conjugating enzymes showed higher affinity to the NEDD8 E1 enzyme (Figure 16). This finding needs to be confirmed using other approaches such as size exclusion chromatography. However, the pulldown data were supported by results obtained using NEDD8‐E2 fusion proteins. In vitro translation reactions showed NEDD8‐Ubc12 C111S and NEDD8‐Nce2 C116S to form an oxyester bond with NEDD8 more efficiently than wt E2s bearing the same mutation (Figure 17 and data not shown). This result points to an increased reaction rate which might be due to an enhanced affinity of the NEDD8‐E2 fusions to APPBP1/UBA3 present in the reticulocyte lysate. Moreover, NEDD8‐E2 fusions were more efficiently modified with His‐NEDD8 than wt E2 enzymes both in vitro and in cellulo (Figures 15B and 15C). However, the enhanced NEDDylation in cells was not completely dependent on the catalytic cysteine as found by the use of cysteine mutants (data not shown). Thus, NEDD8 attached to a lysine in the N terminus of E2 enzymes itself might be a good target for the conjugation of further NEDD8 moieties although it is not clear yet, whether NEDD8 indeed forms chains under normal cellular conditions. So far, there is only evidence for the formation of NEDD8 chains and mixed chains consisting of NEDD8 and ubiquitin moieties upon cellular stress such as heat shock or oxidative stress (Leidecker et al., 2012). Consistently, we identified only monoNEDDylated Ubc12 under endogenous conditions and monoNEDDylated, stably expressed HA‐Nce2 (Figure 11B, data not shown).

- 83 - Discussion

The enhanced affinity of NEDDylated Ubc12 and Nce2 to the E1 enzyme points to an interaction of the attached NEDD8 with APPBP1/UBA3. This heterodimeric E1 possesses two known NEDD8‐binding sites: a non‐covalent one in the cleft between the catalytic cysteine– and the adenylation domain, and the catalytic cysteine residue which covalently binds NEDD8 at its C‐terminal glycine residue (Walden et al., 2003a; Walden et al., 2003b). Hence, one could speculate that the autoNEDDylated E2 enzymes have a higher affinity to the E1 enzyme as the conjugated NEDD8 in their N terminus is non‐covalently bound to the adenylation site. However, our data shows NEDD8‐E2 fusion proteins to be NEDDylated at multiple lysine residues and more active in oxyester formation as well as autoNEDDylation than wt E2 enzymes (Figures 13, 15B and 17; data not shown). Thus, the NEDD8 E1 enzyme is still capable of transferring NEDD8 to autoNEDDylated E2 enzymes or NEDD8‐E2 fusions indicating that its adenylation site can be occupied by free NEDD8, and that there might be a yet uncharacterized interaction site for NEDD8 on the E1 enzyme. Nonetheless, Ubc12 and Nce2 are only monoNEDDylated under endogenous/stable expression conditions within cells, indicating that the autoNEDDylation reaction is stopped or at least slowed down after the transfer of one NEDD8 moiety to the E2 enzymes (Figure 11B, data not shown). In conclusion, an identification of the interaction site between NEDD8 attached to the conjugating enzyme and APPBP1/UBA3 by mutating surface residues or NMR measurements would provide further insights into the function of autoNEDDylation. Alternatively, an already described mechanism could account for the enhanced affinity of autoNEDDylated E2 enzymes to their E1. In addition to a covalent interaction, ubiquitin and SUMO were found to non‐covalently interact with their E2 enzymes UbcH5 and Ubc9, respectively, at the opposite surface of the catalytic cysteine, thereby enhancing the processivity of conjugation (Brzovic et al., 2006; Knipscheer et al., 2007; Sakata et al., 2010). Hence, the NEDD8 moiety attached to the flexible N terminus of Ubc12 and Nce2 might bind to the backside of the conjugating enzyme in a similar manner as these “free” UBLs. This non‐covalent interaction of the attached NEDD8 with the E2 enzyme might then lead to a conformational change which enhances the affinity to APPBP1/UBA3. However, as shown in chemical shift perturbation experiments, free NEDD8 only non‐covalently interacts with the ubiquitin E2 enzyme UbcH5b through its hydrophobic patch around I44, but not with Ubc12. This interaction is speculated to enable a recruitment of ubiquitin‐loaded E2s to NEDDylated cullin‐RING complexes (Sakata et al., 2007).

- 84 - Discussion

4.1.2.4 Further possible functions and impacts of autoNEDDylation As binding of E1 enzymes and E3 ligases to conjugating enzymes is mutually exclusive, the enhanced interaction of autoNEDDylated NEDD8 E2 enzymes with APPBP1/UBA3 may suggest the function of a switch‐off mechanism for the whole NEDDylation cascade (Wenzel et al., 2011). Usually, E2 enzymes only bind their cognate E1 enzyme with significant affinity if it is loaded with the UBL, and APPBP1/UBA3 readily releases the E2 as soon as the latter is loaded with NEDD8 (Haas et al., 1988; Huang et al., 2007; Schulman and Harper, 2009). Hence, if no substrate is present, the E2 enzymes might NEDDylate themselves thereby sequestering the E1 enzyme and inhibiting an interaction with an E3 ligase, whose binding site on the E2 overlaps with the one of the E1 enzyme (Wenzel et al., 2011). Additionally, the interaction between the respective E1 and the E2 enzyme was shown to inhibit further adenylation of SUMO, ubiquitin and NEDD8 (Siepmann et al., 2003; Wang et al., 2010; Wee et al., 2000). Nonetheless, it is very likely that the autoNEDDylated forms of Ubc12 and Nce2 still form thioester linkages with NEDD8 as we identified Ubc12 and Nce2 to be modified with more than one NEDD8 in vitro (Figures 15A and C). Along these lines, the question arises whether autoNEDDylated E2 enzymes still bind to APPBP1/UBA3 if they carry a NEDD8 at their catalytic cysteine residue. In conclusion, a possible involvement of autoNEDDylation in substrate NEDDylation and the effect of acetylation in the N termini of Ubc12 and Nce2 require further investigation. Overexpression of Ubc12 and Nce2 mutants lacking all lysine residues in the respective knockdown cells would be a suitable tool to gain first hints about the effect of autoNEDDylation on substrate NEDDylation under cellular conditions. In addition, the activity of NEDDylated and unmodified NEDD8‐conjugating enzymes towards substrates could be analyzed by in vitro NEDDylation experiments of cullins (Morimoto et al., 2003). On the one hand, switching off the NEDDylation cascade by autoNEDDylation seems to be likely because of an enhanced interaction of APPBP1/UBA3 and a NEDD8‐conjugating enzyme. On the other hand, it is not clear why such a mechanism should exist. However, autoNEDDylation might be necessary to maintain a distinct pool of free NEDD8 to avoid its usage be the ubiquitination machinery being triggered by a change in the ratio of free NEDD8 to free ubiquitin (Hjerpe et al., 2012; Leidecker et al., 2012). As our findings indicate that autoNEDDylation of Ubc12 or Nce2 leads to a higher affinity to the E1 enzyme, one can speculate about several further possibilities for the function of this modification apart from a switch‐off mechanism (Figures 15, 16 and 17). Since NEDD8‐E2 fusions show an enhanced efficiency compared to wt E2 enzymes in oxyester formation and autoNEDDylation reactions, NEDDylation of Ubc12 and Nce2 in their N termini might increase the efficiency of substrate NEDDylation either in general or dependent on the substrate (Figures 15C and 17). Such a function of autoNEDDylation would be in agreement with the one of autoSUMOylation of Ubc9 which enhances SUMOylation for selective substrates (Knipscheer et

- 85 - Discussion al., 2008). To gain first insights into this issue, one could compare the total amount of NEDDylated substrates within E2 knockdown cells stably expressing the E2 or the NEDD8‐E2 fusion protein. However, it must be considered that the acetylation of the N‐terminal methionine of Ubc12 and Nce2 is inhibited by N‐terminal fusion to NEDD8, which might affect substrate NEDDylation (Monda et al., 2013). Furthermore, the attached NEDD8 in the N terminus of Ubc12 and Nce2 might provide an interaction site for various proteins in the cell. Therefore, binding of E3 ligases to the NEDD8‐ conjugating enzymes and hence, substrate specificity may be regulated by autoNEDDylation. Additionally, autoNEDDylation of Ubc12 and Nce2 might enable these conjugating enzymes to transfer NEDD8 without the action of an E3 ligase by direct interaction with certain substrates. Most SUMO substrates contain an Ubc9‐interaction motif enabling a direct binding of the SUMO conjugating enzyme Ubc9 (Bernier‐Villamor et al., 2002). Thus, an interaction motif for autoNEDDylated NEDD8 E2 enzymes might exist within some NEDD8 substrates. Alternatively, the attached NEDD8 might also interact with another NEDD8‐conjugating enzyme, thereby leading to dimer formation which in turn increases E1 binding. Dimer formation has already been shown for Ubc13 and the inactive E2 variant Mms2 which orientates ubiquitin in a way to facilitate the formation of a K63‐linked chain (Hofmann and Pickart, 1999; VanDemark et al., 2001). However, we could not detect any binding of NEDD8‐E2 fusions to wt E2 enzymes in GST‐pulldown assays (data not shown). Nonetheless, these results need to be confirmed to exclude a function of conjugated NEDD8 in the N‐termini of E2 enzymes as an intensifier of processivity on the level of the E2. In 2012, NEDD8 has been shown to be activated by the ubiquitin E1 enzyme under cellular stress conditions (Hjerpe et al., 2012; Leidecker et al., 2012). In these publications, nothing was mentioned about to which E2 enzyme NEDD8 is transferred upon activation. Hence, it is worthwhile to test whether autoNEDDylated Ubc12 and Nce2 are capable of interacting with and accepting NEDD8 from UBA1. Binding of E1 to their cognate E2 enzymes differs between the NEDD8‐ and the ubiquitin‐conjugation system. In the NEDDylation cascade, there is a mainly hydrophobic interaction between the UFD of APPBP1/UBA3 and Ubc12 (Huang et al., 2005). In the ubiquitin system, however, this interaction is organized by polar contacts (Lee and Schindelin, 2008). Thus, these findings argue against binding of autoNEDDylated E2s to UBA1. Nonetheless, NEDD8 attached to the N termini of NEDD8‐conjugating enzymes might lead to binding of UBA1 by changing the interaction surface. UBA1 might therefore be able to transfer ubiquitin to autoNEDDylated E2 enzymes. Concordantly, Ubc12 thioester formation with ubiquitin via UBA1 can be forced by high concentration of UBA1 and a long incubation time (Huang et al., 2008). In addition, deletion of the N terminus of Ubc12 facilitates an interaction with the ubiquitin‐activating enzyme (Huang et al., 2008). To address whether autoNEDDylated

- 86 - Discussion

NEDD8‐conjugating enzymes indeed interact with UBA1, thioester assays including NEDD8 or ubiquitin, UBA1 and NEDD8‐E2 fusion proteins would be suitable.

In conclusion, our data show that autoNEDDylation of Ubc12 and Nce2 in their unique N terminus leads to an enhanced affinity to APPBP1/UBA3. Moreover, we provide evidence that autoNEDDylated E2 enzymes are present in the cytoplasm. The outcome of the enhanced interaction between NEDD8 E1 and E2 enzymes upon autoNEDDylation which might switch off the NEDDylation cascade, induce a change in substrate specificity or lead to enhanced substrate NEDDylation still needs to be elucidated.

- 87 - Discussion

4.2 Interplay between PCNA and the NEDD8 system

The sliding clamp PCNA is not only of big importance for proper replication of DNA but also for appropriate repair upon DNA damage, hence playing a major role in maintaining the integrity of DNA (Hubscher and Maga, 2011). In accordance to results obtained during these studies, NEDD8 may be an additional posttranslational modification regulating PCNA functions.

4.2.1 PCNA as an interaction partner of NEDD8

In an affinity pulldown, we identified PCNA as a possible interaction partner of NEDD8 and ubiquitin (Table 2). Subsequently, we confirmed a direct interaction in a GST‐pulldown assay using N‐ or C‐terminally GST‐tagged NEDD8 and ‐ubiquitin, the latter of which should mimic NEDDylated or ubiquitinated proteins (Figure 21, data not shown). These data indicate that PCNA can interact with NEDDylated or ubiquitinated proteins in the cell. Notably, the position of the tag had no influence on binding to PCNA (data not shown). As PCNA plays a role in DNA replication and repair, binding of polymerases to PCNA may be regulated by ubiquitination or NEDDylation. It is already known that upon DNA damage, the family of Y‐polymerases is recruited to ubiquitinated PCNA via its ubiquitin‐binding domain (UBD). Y‐polymerases themselves can also be monoubiquitinated leading to a preferred binding of the UBD to this ubiquitin so that the polymerase can no longer interact with PCNA (Bienko et al., 2005; Bienko et al., 2010). In addition, the interaction of replicative polymerases with PCNA may be regulated by ubiquitination and/or NEDDylation. Along these lines, the polymerase δ subunits p12 and p66 were identified to be targets for ubiquitination (Liu and Warbrick, 2006). However, whether replicative polymerases are indeed modified with NEDD8 and what would be the effect on the interaction with PCNA still needs to be investigated. Interestingly, in the affinity pulldown with NEDD8 and ubiquitin we also found DNA polymerase δ‐interacting protein 2 (POLDIP2), an interaction partner of PCNA, polymerase δ and the Y‐polymerase η (Table 2) (Liu et al., 2003; Tissier et al., 2010). This result may simply be explained by complex formation with PCNA, thus confirming the interaction of PCNA with NEDD8 and ubiquitin. Moreover, although POLDIP2 was already found to interact with PCNA independently of NEDD8 and ubiquitin, there might be a regulatory role for both UBLs in this context (Liu et al., 2003). To gain further insights into the function of PCNA binding to NEDD8 or ubiquitin, the interaction sites should be mapped. Furthermore, it would be worthwhile to test whether DNA‐bound, trimeric PCNA still binds to NEDD8 or ubiquitin. PCNA is a substrate for SUMOylation and the SUMO conjugating enzyme Ubc9 interacts with PCNA in a two‐hybrid assay (Gali et al., 2012; Hoege et al., 2002; Moldovan et al., 2012; Pfander

- 88 - Discussion et al., 2005). As our data indicate PCNA to be a substrate for NEDDylation, we speculated that PCNA interacts with NEDD8‐loaded E2s before being modified with NEDD8 (Figure 22). So far, nothing is known about whether there is a direct interaction of Ubc12 and Nce2 with their substrates. An interaction with an E3 ligase, however, has only been shown for Ubc12 and the E3 ligases RBX1 and RNF111 yet (Calabrese et al., 2011; Ma et al., 2013). Thus, it seems possible that NEDD8 E2s interact directly with their substrate and that this interaction is stimulated by a NEDD8 moiety conjugated to the E2 enzyme via thioester bond. This hypothesis is supported by the identification of Histone H4 as interaction partner of His‐NEDD8 which was recently shown to act as a substrate for NEDD8 (Table 2) (Ma et al., 2013). To investigate this issue, we performed an IP of overexpressed E2s and the E3 ligase Rad18 as a control and tested whether PCNA is bound (Figure 25). In order to distinguish between free and chromatin‐bound PCNA, we divided cell lysate into a triton‐soluble and ‐insoluble fraction. Myc‐Rad18 interacted with PCNA in both fractions. Ubc12 and Nce2, however, were only found in the triton‐soluble fraction where they did not interact with PCNA in detectable amounts (Figure 25). Nonetheless, there might be a binding to PCNA in a NEDD8‐dependent manner which could be tested in an in vitro pulldown assay using NEDD8‐conjugating enzymes bearing an oxyester‐bound NEDD8. For this approach, NEDD8 could also be linked to the active site of the E2 via an isopeptide bond by mutating the catalytic cysteine to a lysine residue (Plechanovova et al., 2012). Since an interaction between PCNA and the E2s might only take place under certain cellular conditions such as a distinct cell cycle phase, one should repeat the IP with synchronized cells that are harvested in different stages of cell cycle. Additionally, it is important to perform the IP under non‐reducing conditions or in the presence of an alkylating agent so that the E2 enzymes are still loaded with NEDD8. In conclusion, one should elucidate whether free NEDD8 and ubiquitin bind to PCNA in cells or whether they need to be attached to a substrate. To this end, a combination of PCNA‐pulldown and mass spectrometric analysis of interacting proteins would be useful.

4.2.2 PCNA as a substrate for NEDD8

4.2.2.1 Evidence for NEDDylation of PCNA in vitro and in cellulo In vitro and in cellulo NEDDylation assays indicated that PCNA can be monoNEDDylated (Figure 22). In vitro, both NEDD8‐conjugating enzymes were capable of transferring NEDD8 to PCNA. The double band appearing upon addition of the NEDD8‐conjugating system may result from modification at two different lysine residues within PCNA (Figure 22B). Alternatively, a truncated translation product generated by using an internal methionine as start codon may be modified with NEDD8. The E3 ligase needed for catalysis of the NEDD8 transfer from the E2 to PCNA must be present in the reticulocyte lysate used for in vitro translation of PCNA, especially because NEDDylation of bacterially expressed PCNA was not detectable without addition of an

- 89 - Discussion

E3 ligase (data not shown). Since the efficiency of NEDDylation in the in vitro NEDDylation assay was very low, one should investigate whether DNA binding of PCNA increases its modification with NEDD8 (Figure 22B). Moreover, the addition of an E3 ligase mediating the NEDDylation of PCNA might enhance its modification in vitro, too. As all known NEDD8 E3 ligases are RING‐ ligases showing dual specificity for NEDD8 and ubiquitin, a possible candidate is Rad18, an E3 ligase responsible for the monoubiquitination of PCNA (Broemer et al., 2010; Hoege et al., 2002; Huang et al., 2008; Morimoto et al., 2003; Oved et al., 2006; Xirodimas et al., 2004). Interestingly, co‐expression of Rad18 and His‐NEDD8 strongly enhanced the conjugation of NEDD8 to endogenous PCNA in cells. Rad18 therefore very likely represents the E3 ligase not only for ubiquitination but also for NEDDylation of PCNA (Figure 24). However, we cannot exclude that Rad18 indirectly affects monoNEDDylation of PCNA in cellulo. To elucidate this issue, an E2 binding mutant or catalytically inactive mutant of Rad18 should be compared to the wild‐type (Hibbert et al., 2011). To confirm the E3 ligase activity of Rad18 towards NEDD8, in vitro NEDDylation assays and pulldown assays with Rad18 and the NEDD8‐conjugating enzymes should be performed. As already mentioned, our results also showed endogenous PCNA to be monoNEDDylated upon overexpression of His‐NEDD8 (Figure 22A). In coincidence with our in vitro results, we always detected a double band at the size of monoNEDDylated PCNA (Figure 24). Thus, two different lysines might be modified with His‐NEDD8 or another small post‐translation modification such as a phosphate group might in addition be attached to one monoNEDDylated form of PCNA (see 4.2.2.2) (Wang et al., 2006). Unfortunately, we were not able to show PCNA NEDDylation with endogenous NEDD8. A Ni‐pulldown from MCF‐7 cells stably expressing His‐NEDD8 also did not yield NEDDylated PCNA (data not shown) (Xirodimas et al., 2004). Although we found modified PCNA with a size of ~40 kDa upon IP, we could not definitely prove whether it corresponds to a monoNEDDylated or a monoubiquitinated form of PCNA (data not shown). Treatment of this modified PCNA with either a deNEDDylating‐ or a deubiquitinating enzyme reduced the modification but both enzymes were not specific for either cleaving ubiquitin or NEDD8 under the conditions used. Moreover, probably due to low detection efficiencies, NEDD8‐ and ubiquitin‐specific antibodies did not detect modified PCNA (data not shown). Hence, NEDDylated PCNA might be present at very low levels under normal cellular conditions and the conjugation of NEDD8 to PCNA might depend on a special trigger such as stress conditions or a distinct cell cycle phase. Furthermore, the existence of NEDDylated PCNA at endogenous NEDD8 levels is supported by the fact that it was found in HeLa cells stably expressing TAP‐tagged NEDD8 (Xirodimas et al., 2008). As both PCNA ubiquitination and SUMOylation with SUMO‐1 at lysine 164 has already been described, we also had a look at these two modifications (Gali et al., 2012; Hoege et al., 2002; Kannouche et al., 2004). Upon overexpression of His‐ubiquitin and subsequent Ni‐pulldown, we

- 90 - Discussion found PCNA to be monoubiquitinated which is a marker for ongoing translesion synthesis (Figure 22A) (Hoege et al., 2002; Kannouche et al., 2004; Terai et al., 2010). As DNA preparations usually contain a portion of nicked DNA, transient overexpression conditions might cause DNA repair. Alternatively, monoubiquitination might have a yet unknown, additional function on PCNA. Overexpression of Rad18 enhanced monoubiquitination of PCNA thereby confirming its function as an E3 ligase (Figure 24) (Hoege et al., 2002; Watanabe et al., 2004). Although Hdm2 was published to act as ubiquitin E3 ligase for PCNA (Groehler and Lannigan, 2010), its overexpression did neither have an effect on NEDDylation nor on ubiquitination of PCNA in our hands (data not shown). Moreover, we never detected polyubiquitinated PCNA in our assays, which is involved in error‐free DNA repair pathways (K63‐linked chain) or serves as degradation signal (Bergink and Jentsch, 2009; Lo et al., 2012). In contrast, SUMOylation of PCNA inhibits recombination during S phase of the cell cycle and is important for the prevention of double strand breaks (Gali et al., 2012; Moldovan et al., 2012). In overexpression experiments, we did not find PCNA to be modified with His‐SUMO‐1 whereas we were able to pull down SUMOylated p53 as positive control (Figure 22A, data not shown). This result might be due to a too low amount of cells in S phase. Therefore, a synchronization of cells and harvest in S phase might reveal SUMOylation of PCNA.

4.2.2.2 Effects of MLN4924 on the NEDDylation of PCNA Recently, the activation of NEDD8 by the ubiquitin E1 enzyme UBA1 upon overexpression of NEDD8 or under diverse stress conditions in the cell has been described. Transfer of NEDD8 to any of several tested ubiquitin E2 enzymes in turn leads to NEDDylation of otherwise ubiquitinated substrates. However, it remains to be verified that this mechanism is physiologically relevant (Hjerpe et al., 2012; Leidecker et al., 2012). Because of these data we tested whether NEDDylation of PCNA is indeed mediated by enzymes of the NEDDylation cascade in cellulo. We could show that treating cells with the APPBP1/UBA3 inhibitor MLN4924 for 20 h significantly diminished monoNEDDylation of PCNA. However, neither overexpression of the deNEDDylating enzyme NEDP1 nor treatment with MLN4924 completely abolished monoNEDDylation of PCNA, suggesting NEDDylated PCNA to be very stable (Figure 23B). This finding is supported by the facts that we did not see an effect of MLN4924 when treating cells only for 3 h and that NEDP1 predominantly localizes to the cytoplasm (data not shown) (Sundqvist et al., 2009). Hence, mainly the nuclear fraction of PCNA might be NEDDylated via the NEDD8 cascade. Moreover, a minor amount of NEDD8 might indeed be activated by UBA1 and transferred to PCNA via the ubiquitin system. However, it is not clear yet whether only mixed chains consisting of ubiquitin and NEDD8 are formed upon overexpression of NEDD8 or whether monoNEDDylation, which we found for PCNA, can also be mediated by UBA1 (Leidecker et al., 2012).

- 91 - Discussion

The NEDD8 E1‐inhibitor MLN4924 forms an NEDD8‐AMP mimetic which occupies the nucleotide binding pocket within APPBP1/UBA3, thereby interfering with NEDD8 activation. The complex of APPBP1/UBA3 and the NEDD8‐MLN4924 adduct is very stable as demonstrated by a dissociation constant of less than 1 nM (Brownell et al., 2010). Treatment of HCT‐116 cells with 0.3 µM MLN4924 was shown to result in S phase arrest after 8 h, and DNA damage as well as apoptosis after 24 h (Soucy et al., 2009). In addition, re‐replication and checkpoint activation are induced by MLN4924 because of a stabilization of the DNA replication factor Cdt1 (Lin et al., 2010). These findings complicate the interpretation of our data obtained by using MLN4924. In immunofluorescence experiments, we found an accumulation of PCNA in large S phase foci already after 3 h of treatment with 1 µM MLN4924, indicating an arrest in S phase (data not shown) (Essers et al., 2005). In addition, 20 h of treatment probably resulted in DNA damage as seen by the appearance of a band at the size of monoubiquitinated PCNA, which gives notice of ongoing repair mechanisms (Figure 23B). Thus, data obtained by using MLN4924 to prove NEDDylation by APPBP1/UBA3 should be interpreted carefully with respect to secondary effects because of cell cycle arrest. For example, as shown in figure 23B NEDDylation of PCNA might only decrease upon addition of MLN4924 as DNA damage induces ubiquitination of the same lysine and not because of the inhibition of APPBP1/UBA3. Modification of endogenous PCNA with overexpressed His‐NEDD8 might therefore be catalyzed by enzymes of the ubiquitination cascade, the physiological relevance of which must be proven (Hjerpe et al., 2012; Leidecker et al., 2012). However, the fact that PCNA is NEDDylated by APPBP1/UBA3 and Ubc12 or Nce2 in vitro, and the finding of NEDDylated PCNA in a cell line stably expressing NEDD8 point to PCNA to be a new substrate of the NEDD8 system (Xirodimas et al., 2008). Nonetheless, conditions under which PCNA is modified with endogenous NEDD8 as well as the physiological function of this modification remain to be determined.

4.2.2.3 Hints for possible functions of NEDDylated PCNA Both ubiquitin and SUMO‐1 are attached to lysine residue 164 of PCNA (Hoege et al., 2002; Kannouche et al., 2004). In addition, SUMOylation can occur at K254 (Gali et al., 2012). To test whether NEDD8 is also conjugated to K164, we included a K164R mutant in our assays. In vitro, we found K164R mutant to be as efficiently NEDDylated as wt PCNA (Figure 22B). In cells, however, mutation of K164 completely abolished monoNEDDylation (Figure 22C). This result demonstrates that K164 is at least necessary for NEDD8 attachment in cellulo by serving either as NEDDylation site or by recruiting NEDDylating enzymes. In contrast, under in vitro conditions NEDD8 might either be generally conjugated to another lysine residue or mutation of K164 forces the usage of an alternative lysine residue for the attachment of NEDD8. If K164 was indeed the conjugation site for NEDD8, this would lead to a competition of three different modifications namely SUMOylation, ubiquitination and NEDDylation, for one single lysine

- 92 - Discussion residue, probably depending on different cellular conditions. A similar situation is found in the regulation of p53. Its lysine residues in the C‐terminal domain are targets for methylation, acetylation, NEDDylation, ubiquitination and SUMOylation. NEDD8 and ubiquitin both can be conjugated to K370, K372 and K373 whereas SUMO‐1 and ubiquitin can be attached to K386. However, defined function and exact regulation of these modifications remain to be determined (summarized in (Kruse and Gu, 2009). PCNA is constitutively expressed but resides in the nucleus only in S phase of the cell cycle or upon DNA damage (Bravo and Macdonald‐Bravo, 1985; Celis and Madsen, 1986; Takasaki et al., 1981; Toschi and Bravo, 1988). To address whether free or chromatin‐bound PCNA is NEDDylated, we used PCNA Y211F which cannot be phosphorylated anymore, leading to its inability to bind to DNA, its polyubiquitination and subsequent proteasomal degradation (Lo et al., 2012; Wang et al., 2006). In our hands, PCNA Y211F levels were comparable to wt and polyubiquitination was not detected. Therefore, additional experiments including a proteasome inhibitor need to be performed. Furthermore, our data demonstrate that PCNA Y211F is monoNEDDylated with the same efficiency as PCNA wt, suggesting that free PCNA serves as NEDD8 substrate (Figure 26). Since there is free and chromatin‐bound wt PCNA, we cannot draw a conclusion about whether chromatin‐bound PCNA can also be NEDDylated. Therefore, dividing cell lysate in a triton‐soluble and a triton‐insoluble fraction might give further information about this issue. Recently, it was shown that some ubiquitin‐binding domains such as the UBA domain of Rad23 bind both to ubiquitin and Rub1 (yeast NEDD8) (Singh et al., 2012). Therefore, some polymerases might recognize NEDDylated and/or ubiquitinated PCNA dependent on cellular conditions. Since ubiquitinated PCNA recruits polymerases mediating translesion synthesis and a similar electrostatic distribution is found on the surface of ubiquitin and NEDD8, NEDDylated PCNA might also be involved in DNA repair mechanisms (Davies et al., 2008; Hoege et al., 2002; Whitby et al., 1998). Alternatively, PCNA NEDDylation might inhibit its ubiquitination if no DNA damage or replication stress is present. However, it is not clear when SUMOylation comes into play and how the switch between these three different modifications is controlled. In budding yeast, there is evidence that SUMOylated PCNA recruits the ubiquitin E3 ligase Rad18 via its SUMO‐interaction motif. Furthermore, the activity of Rad18 is stimulated by PCNA SUMOylation but not completely dependent on SUMO. These data also suggest that different subunits of the PCNA trimer are SUMOylated and ubiquitinated at the same time. Moreover, probably neither ubiquitination nor SUMOylation occurs at all three subunits of the PCNA trimer simultaneously (Parker and Ulrich, 2012). In addition to SUMOylation and the Rad18/Rad5‐mediated ubiquitination involved in DNA repair mechanisms, K164 of PCNA has been described as target of the Cullin4A‐RING complex (CRL4). Like Rad18, CRL4 catalyzes the monoubiquitination of PCNA in cells during replication,

- 93 - Discussion thereby promoting DNA repair. A major difference between Rad18 and CRL4, however, is the trigger for ubiquitination: Rad18 acts upon externally induced DNA damage, whereas CRL4 regulates monoubiquitination of PCNA in unperturbed cells (Terai et al., 2010). Furthermore, polyubiquitination of PCNA by CRL4 leads to its proteasomal degradation (Lo et al., 2012). Thus, NEDDylation of PCNA at lysine 164 might protect PCNA from being degraded. A similar antagonistic effect of NEDDylation and ubiquitination has been published for the TGF‐β type II receptor which, when NEDDylated, is protected from ubiquitination and degradation (Zuo et al., 2013). Moreover, CRL4(Cdt2)‐mediated degradation of several substrates such as p21 or Cdt1 in S phase or upon DNA damage depends on PCNA. A PCNA‐binding motif (PIP‐box) within these substrates interacts with PCNA, thereby creating a degron recruiting CRL4(Cdt2) (Abbas et al., 2008; Havens and Walter, 2009, 2011; Senga et al., 2006). Therefore, it is worthwhile to investigate whether NEDDylated PCNA plays a role in the regulation of substrate degradation by CRL4 and whether it has a general influence on the interaction with other proteins. However, as shown by the tertiary structure of a PCNA monomer, the interaction site of the PIP‐box on PCNA is not located in very close proximity to lysine 164 (Bruning and Shamoo, 2004).

4.2.2.4 Click reaction as tool to identify functions of monoNEDDylated PCNA To shed light on the function of monoNEDDylated PCNA, we aimed for site‐specifically linking the C terminus of NEDD8 to PCNA at amino acid position 164 using Cu(I)‐catalyzed Huisgen azide‐alkyne cycloaddition (Eger et al., 2011). In a second step, we intended to use this “NEDDylated” PCNA as bait to identify interacting proteins and to test whether it affects the activity of DNA polymerases. After expression and purification, we measured the mass of AhaNEDD8 to confirm the incorporation of azidohomoalanine. It turned out that the cleavage of the N‐terminal methionine inside the bacteria was not efficient since we found NEDD8 proteins with different masses (data not shown). Moreover, purification of AhaNEDD8 was not successful, as the preparation contained many contaminating proteins that unspecifically bound to the Ni‐beads (Figure 28B). Anyway, we performed the click reaction with AhaNEDD8 and PlkPCNA. Unfortunately, it was not clear whether a triazole linkage was formed between both proteins because of the high background in the Coomassie gel (Figure 29). In contrast, we successfully linked PCNA to AhaUbiquitin which should serve as a control in future experiments, and which can be used for further investigations on the function of ubiquitinated PCNA (Figure 29). To improve the production of AhaNEDD8, it would be worthwhile to clone a GST‐tagged NEDD8 whose tag, including the initial methionine, can be cleaved off. By doing so, the preparation would be much more pure and only contained the desired protein with the C‐terminal azidohomoalanine (Schneider et al., 2013).

- 94 - Discussion

As determined by the autoNEDDylation of Ubc12 and Nce2 in vitro and the NEDDylation of p53 in cellulo, NEDD8 M50A, the mutant used for click reaction, was still utilized by the NEDD8‐ conjugation system, although with a slightly lower efficiency (data not shown). These data indicate that artificially assembled “monoNEDDylated PCNA” is suitable to study its functions.

To sum up, we identified PCNA to serve as an interaction partner of NEDD8 and as a substrate of the NEDD8‐conjugation cascade, too. Modification of PCNA by endogenous NEDD8, however, still needs to be demonstrated and the conditions triggering this event need to be identified. It is especially of interest how SUMOylation, ubiquitination and NEDDylation of the same lysine residue within PCNA are regulated. First hints about the function of NEDDylated PCNA can be obtained in the future by interaction studies using “monoNEDDylated” PCNA generated via click reaction.

- 95 - Discussion

4.3 A second isoform of Nce2 with individual properties

Nce2 has been shown to mediate the NEDDylation of Cullin5 playing roles in neuronal migration, chondrocyte differentiation, cytokine signaling as well as myogenesis (Dentice et al., 2005; Feng et al., 2007; Kile et al., 2002; Nastasi et al., 2004). In addition, Nce2 seems to be involved in conjugating NEDD8 to the transcription factor E2F1 since a dominant negative mutant of Nce2 abolishes its NEDDylation (Aoki et al., 2012). However, the role of Nce2 in the cell is only poorly understood and requires further investigation. Therefore, during this study we characterized a new isoform of Nce2 which might be involved in the NEDDylation cascade.

4.3.1 Nce2 isoform 2 as a splice variant with an extended C terminus

In 2012, five isoforms of Nce2 have been identified in glioma cells by the NEDO (New Energy and Industrial Technology Development Organization) human cDNA sequencing project (see 3.3). As a result of alternative splicing, isoform 2 contains a 26 aa long C‐terminal extension compared to isoform 1. The usage of TG instead of AG as an alternative 3´ splice site in intron 9 leads to an insertion of 23 bases at position +507 of the mRNA which itself gives rise to a frame shift and an elongated coding sequence (Figure 30). The last 55 bases in the coding sequence of isoform 2 mRNA therefore correspond to a part of the 3´ UTR of isoform 1. “TG” has first been reported as an acceptor splice site in U2‐dependent introns of eukaryotic pre‐mRNAs in 2007 (Szafranski et al., 2007). Interestingly, the rare TG splicing sites are always associated with the canonical AG 3´ splice site, thus only serving as alternative acceptor site. Moreover, ratios of the alternative splice variants were found to be tissue‐specific for some introns (Szafranski et al., 2007). Hence, comparison of mRNA and protein levels of Nce2 isoforms in different tissues and investigation of the cis‐regulatory elements of the splicing mechanism would shed light on the functions and differences of Nce2 isoform 1 and 2.

4.3.2 mRNA of the second isoform of Nce2 is present in HEK293T cells

In a first experiment, we verified the presence of Nce2 isoform 2 mRNA in HEK293T cells (Figure 31). However, a conclusion about the amounts of isoform 1 and isoform 2 mRNA in these cells cannot be drawn from that particular RT‐PCR experiment since we did not digest the

- 96 - Discussion genomic DNA after reverse transcription. There are four pseudogenes of UBE2F (gene name of Nce2) present on human chromosomes 1, 3, 16 and 20 which were amplified with primers annealing in the mRNA of both isoforms (National Center for Biotechnology Information). Nonetheless, as UBE2F pseudogenes do not contain the isoform 2‐specific insertion, only mRNA of Nce2 isoform 2 was amplified with the specific primer. In order to confirm the existence of Nce2 isoform 2 on protein level in cells, we developed a polyclonal antibody recognizing aa 187‐200 of isoform 2. We were able to verify the specificity of this antibody for isoform 2 in vitro and in cells (Figure 38). Although the antibody was functional and specific in IP experiments and only 5 ng of Nce2 isoform 2 are required for detection, we could not verify the existence of endogenous protein (Figure 38C, data not shown). An IP using an antibody against the N terminus of Nce2 also did not yield enough isoform 2 to detect it with the specific antibody (data not shown). Possible reasons are the low steady state levels and the short half‐life of Nce2 isoform 2 (Figure 36). Because of toxic effects after a long treatment, we added MG132 for only 4 h to enrich for Nce2 isoform 2 prior to IP (Figure 38C). However, it was recently shown that the toxicity of proteasome inhibitors is due to an amino acid deficit and can be rescued by providing additional cysteine to the cells (Suraweera et al., 2012). Thus, longer treatment with MG132 and supplementation of cysteine to the cells might facilitate a detection of endogenous Nce2 isoform 2. In immunofluorescence experiments using the specific antibody raised against Nce2 isoform 2, we detected a juxtanuclear staining in H1299 cells which might point to the existence of the protein (see figure 40). However, to exclude an unspecific staining, this experiment needs to be repeated with an Nce2 knockdown cell line. Finally, to investigate whether Nce2 isoform 2 is indeed expressed, another specific antibody recognizing a presumably better accessible region would be useful. In order to characterize the second isoform of Nce2, we therefore used overexpressed Nce2 isoform 2 for further experiments. In summary, although we detected mRNA of isoform 2 it is not clear whether a functional protein is translated. Noteworthy, the alternative splicing site, as well as the 3´ UTR of UBE2F are only conserved in Pan troglodytes but not in Macaca mulatta or Mus musculus (data not shown) (National Center for Biotechnology Information).

4.3.3 Nce2 isoform 2 is active in vitro and can be NEDDylated

Since the amino acid sequences of Nce2 isoform 1 and 2 are identical from position 1 to 169, the catalytic cysteine residue at position 116 is present in both proteins. Therefore, we tested whether isoform 2 is active as a NEDD8‐conjugating enzyme. In an in vitro thioester assay using GST‐NEDD8, we confirmed the activity of a C‐terminally His‐tagged Nce2 isoform 2 as a NEDD8 E2 enzyme. The formation of a thioester linkage between isoform 2 and NEDD8 was verified by

- 97 - Discussion

DTT‐sensitivity of the additional band. Noteworthy, Nce2 isoform 2 was not capable of forming a thioester with GST‐ubiquitin under the conditions used, underlining its specificity for the NEDD8 system (Figure 33A). Furthermore, thioester formation between NEDD8 and Nce2 isoform 2 was inhibited by an N‐terminal tag of the E2 as found for several E2 enzymes before (data not shown; AG Scheffner, unpublished data). The ability of Nce2 isoform 2 to accept NEDD8 from the NEDD8 E1 enzyme was further affirmed by oxyester formation between the catalytically inactive mutant C116S and NEDD8 in vitro, which is usually characterized by a different running behavior compared to Nce2 isoform 2 containing a monoNEDDylated lysine residue (own observations; figure 33B; see also figure 10). Nonetheless, so far we are lacking an evidence for the activity of Nce2 isoform 2 as a NEDD8‐conjugating enzyme in cells. Upon overexpression of Nce2 isoform 1, isoform 2 or Ubc12 (wt or catalytically inactive) and NEDD8, we did not see an effect on the NEDDylation of different NEDD8 substrates such as Cullin1, Cullin5 or p53, indicating that endogenous E2 enzymes are already sufficient for the NEDDylation of those proteins (data not shown). Hence, indications for the involvement of Nce2 isoform 2 in the NEDDylation of known NEDD8 substrates could be obtained with a specific knockdown of isoform 2. To examine whether Nce2 isoform 2 is capable of autoNEDDylation, we performed in vitro and in cellulo NEDDylation assays using the catalytically inactive mutant C116A and a mutant lacking the first 26aa. We found that the C116A mutant is still NEDDylated with the same efficiency as wt isoform 2 (Figures 33B and 35A). Deletion of the N‐terminal 26 aa of isoform 2 also only had a minor effect on its NEDDylation in cells, being in stark contrast to the other NEDD8 E2 enzymes where deletion of the N terminus almost completely abolished their NEDDylation (Figures 35A and 15). Moreover, in an IP of overexpressed HA‐tagged Nce2 isoform 1 and 2 from the cytosolic fraction of H1299 cells, we only detected isoform 1 but not 2 to be modified with endogenous NEDD8 (Figure 35B). These data indicate that, in contrast to isoform 1 and Ubc12, Nce2 isoform 2 is probably not capable of autoNEDDylation, and the N terminus is not the predominant NEDDylation site as it is the case for Ubc12 and Nce2 isoform 1 (Figures 13 and 35A). Thus, the C terminus of isoform 2 seems to impair autoNEDDylation within cells, which could be due to an interaction with a NEDD8 substrate or an E3 ligase enhancing the transfer of NEDD8 to a substrate. Since we did not see an effect on the efficiency of NEDDylation of isoform 2 upon addition of Ubc12 or Nce2 isoform 1 in vitro and in cells, we cannot draw a conclusion about which E2 is involved in the NEDDylation of Nce2 isoform 2 (data not shown). In any case, the enzymes present in the reticulocyte lysate and E2 enzymes present in cells seem to be sufficient to mediate NEDDylation of Nce2 isoform 2. Alternatively, NEDD8 could be directly transferred from the E1 enzyme to a lysine residue of isoform 2.

- 98 - Discussion

4.3.4 The cellular localization distinguishes Nce2 isoform 2 from isoform 1

To identify further differences between Nce2 isoform 1 and 2 we investigated the localization of both proteins in overexpression experiments. As shown in a cellular fractionation assay, Nce2 isoform 1 was mainly present in the cytosol whereas isoform 2 was also detected in nuclear fractions and the pellet, which should only contain chromatin (Figure 37C). When we performed this assay using tubulin α, which should only be found in the cytosol, we also detected a small amount in the pellet fraction (data not shown) which complicates the interpretation of the data for Nce2. In immunofluorescence experiments, we did not see a striking difference in the localization of isoform 1 and 2. Both proteins mainly localized in the cytosol but we also found cells with a stronger staining of isoform 2 in the nucleus (Figures 37A and B). To elucidate whether there is indeed a difference in the localization of both isoforms, differently tagged Nce2 isoform 1 and 2 should be co‐transfected for immunofluorescence. Using the specific Nce2 isoform 2‐antibody, we detected a juxtanuclear staining (Figure 40B). In contrast, isoform 1 was distributed everywhere in the cell as found by using an antibody specifically recognizing its very C terminus (Figure 40A). A potential difference in localization between the endogenous and the overexpressed Nce2 isoform 2 might be explained by different protein levels. However, it is of huge importance to prove that the staining obtained with both antibodies is specific. Since the subcellular distribution of Nce2 isoform 2 slightly differed from isoform 1 in our experiments, we speculate that the different C termini have an influence on the regulation of localization. In the literature, there are already examples for different compartmentalization of splice variants. One of them is survivin, an inhibitor of apoptosis, which has two functionally divergent splice variants (Mahotka et al., 2002). Survivin and one of its splice variants, survivin‐ 2B, mainly localize to the cytosol whereas survivin‐ΔEx3, which contains a nuclear localization signal (NLS), accumulates in the nucleus in a cell cycle‐dependent manner (Mahotka et al., 2002). Hence, an investigation of the localization of Nce2 isoforms in different cell cycle phases would provide first hints about their regulation and their function.

- 99 - Discussion

A

B

Figure 40. Nce2 isoform 2 localizes next to the nucleus in H1299 cells (A) H1299 cells were stained with α Nce2 C‐term antibody (red; Abgent) and DAPI (blue; scale bar 20 µm). Nce2 isoform 1 is distributed all over the cell. (B) H1299 cells were stained with α Nce2 isoform 2 antibody (red; University of Konstanz) and DAPI (blue; scale bar 20 µm). Nce2 isoform 2 accumulates in a juxtanuclear region.

4.3.5 A C‐terminal unstructured region promotes insolubility in bacteria and rapid degradation in cells

When we compared Ubc12 and the two isoforms of Nce2 concerning their ability to be ubiquitinated, we found Nce2 isoform 2 to be targeted for ubiquitination and proteasomal degradation (Figures 34 and 36). The different C terminus of the second isoform diminishes the half‐life of Nce2 to less than 3 h (Figure 36). These findings explain the low steady state levels of Nce2 isoform 2 upon overexpression in H1299 cells and provide an explanation for detection problems of the endogenous protein (Figures 34 and 38C). Furthermore, catalytic activity of Nce2 isoform 2 does not play a role for its degradation since mutating the catalytic cysteine does not influence the degradation rate (Figure 36B). Interestingly, mutation of one or both C‐ terminal lysine(s) of Nce2 isoform 2 also did not change its half‐life, suggesting that the ubiquitinated lysine residue(s) is located in the common sequence of Nce2 isoform 1 and 2 but only the C terminus of isoform 2 attracts the ubiquitination machinery (data not shown). By means of the tertiary structure prediction program HHpred and Pymol, we were able to model the structure of Nce2 isoform 2. The unique C terminus of the second isoform of Nce2

- 100 - Discussion forms an α‐helix containing four turns followed by an unstructured, flexible region and an additional α‐helix being composed of only one turn (Figure 32). In general, unstructured proteins and proteins containing disordered regions seem to be more susceptible to proteasomal degradation being independent of ubiquitination (Baugh et al., 2009; Gsponer et al., 2008). Thus, these regions might serve as a universal signal for ubiquitin‐independent degradation by the proteasome. For example, ODC was shown to comprise an unstructured region in its N terminus serving as degradation signal. Interestingly, transplantation of this ~45‐residue domain to other proteins only resulted in their instability when fused to an α‐helix but not to a β‐sheet (Godderz et al., 2011). Nonetheless, it is also widely accepted that ubiquitinated substrates require a disordered region to be efficiently degraded by the proteasome (Prakash et al., 2004; Takeuchi et al., 2007). Therefore, the C terminus of Nce2 isoform 2 containing an α‐helix and an unstructured region might serve as a signal for ubiquitination and proteasomal degradation. To ascertain this hypothesis, one could fuse it to an otherwise stable protein and determine its half‐life. Upon expression of Nce2 isoform 2 in bacteria, the majority of protein was found to be insoluble (data not shown). Since isoform 1 is highly soluble, we concluded that the C terminus of isoform 2 may lead to folding problems. Therefore, we generated Kyte‐Doolittle plots using the amino acid sequences of Nce2 isoform 1 and 2 (see figure 41). The graph of the second isoform shows only the very C‐terminal residues to be above the line, i.e. hydrophobic, but their hydrophobicity does not exceed the one of other parts of the protein. Nonetheless, folding problems might occur because the C terminus is exposed. Thus, bacterial expression of Nce2 isoform 2 could be improved by fusion of a long, soluble tag such as GST to its C terminus. Another hint for an insolubility of Nce2 isoform 2 was discovered in immunofluorescence experiments, where overexpression of isoform 2 led to the formation of aggregates in rare cases (Figure 37B). However, upon TNN lysis only a negligible amount of protein was present in the pellet (data not shown). These results suggest that under cellular conditions, folding problems and subsequent aggregate formation of Nce2 isoform 2 are usually prevented by interacting proteins.

4.3.6 Possible functions of Nce2 variants

The mRNAs of all five splice variants of Nce2 were identified in glioma cells by the NEDO human cDNA sequencing project that focused on splice variants. In a first step, one should elucidate whether the corresponding proteins are indeed translated and functional in cells. If these proteins are present in cells, a dominant negative effect on the activity of Nce2 isoform 1 should be considered. Moreover, it would be worthwhile to test whether mRNA‐ and protein levels of

- 101 - Discussion the five isoforms vary in different tissues/cell lines and whether their expression levels are differentially regulated dependent on cellular conditions. To prove the existence and to identify functions of Nce2 isoform 2 in cells, most importantly another antibody with a higher sensitivity to isoform 2 must be developed. With this antibody in hands, the localization‐, half‐life‐ and functional studies should be carried out for the endogenous protein. Nonetheless, the different subcellular localization of both overexpressed and endogenous Nce2 isoform 1 and 2 already points to a divergent function of the two isoforms (Figures 37 and 39). Since we have evidence for NEDDylation and ubiquitination but not for autoNEDDylation of Nce2 isoform 2, it is also differentially regulated than the first isoform. Its half‐life is significantly shorter than that of isoform 1, suggesting an important function which has to be strictly controlled (Figures 34 and 36). However, expression and/or stability of Nce2 isoform 2 might be enhanced upon certain signals. A

B

Figure 41. The very C terminus of Nce2 isoform 2 shows slight hydrophobicity Kyte‐Doolittle plots for Nce2 isoform 1 (A) and isoform 2 (B). Hydrophilic residues are below, hydrophobic residues above the line (http://fasta.bioch.virginia.edu/fasta_www2/fasta_www.cgi?rm=misc1). The red line marks the beginning of the isoform 2‐specific C terminus. The very C‐terminal residues of Nce2 isoform 2 are hydrophobic.

- 102 - Discussion

As Nce2 was identified as NEDD8‐conjugating enzyme particularly mediating the transfer of NEDD8 to Cullin5, it would be worthwhile to check the effect of isoform 2 on the NEDDylation of Cullin5 (Huang et al., 2009). Although we found Nce2 isoform 2 to be capable of forming a thioester bond with NEDD8, it is not clear whether it possesses the ability to transfer NEDD8 to substrates (Figure 33A). Thus, the substrate specificity of Nce2 isoform 1 and 2 should be investigated. Another possible function of isoform 2 is inhibition of isoform 1 functions either by a dominant‐ negative effect or by forming a heterodimer. Hence, the affinity of both isoforms to APPBP1/UBA3 should be addressed e.g. by isothermal titration calorimetry. Furthermore, interaction studies between Nce2 isoform 1 and 2 should be performed. For example, in case of the yeast ubiquitin‐conjugating enzyme Ubc13 and the catalytically inactive E2 enzyme Mms2 heterodimerization is crucial for mediating the extension of K63‐linked ubiquitin chains (Eddins et al., 2006; McKenna et al., 2001). Therefore, Nce2 isoform 2 might interact with isoform 1 to modulate its function. Another hint for physiological functions of Nce2 and its variants is provided by transcription factors binding to the UBE2F promoter. Although Nce2 isoform 1 is expressed in most adult tissues, binding sites for GATA‐6, FOXD1 and FOXA2 indicate Nce2 and its variants to have preferential roles in embryonic development and cellular differentiation (www.genecards.org/cgi­bin/carddisp.pl?gene=UBE2F) (Carreres et al., 2011; Hatini et al., 1996; Huang et al., 2009; Kaestner, 2010; Maeda et al., 2005). In addition, SILAC experiments comparing wt cells and Nce2 isoform 2 knockdown cells would be suitable to elucidate physiological functions of isoform 2. Since the unique, exposed and hydrophobic C terminus of Nce2 isoform 2 very likely represents an interaction site for various proteins in the cell, binding studies could be conducted using either full length isoform 1 and 2 or only their very C‐termini. We already performed a GST‐ pulldown assay using overexpressed HA‐GST‐Nce2 isoform 1 and 2 and HA‐GST as a control. Although the assay conditions still need to be improved because of a high background binding to the GSH‐beads, we identified three potential interaction partners which were not pulled down with HA‐GST‐Nce2 isoform 1 or HA‐GST: 60S ribosomal protein L4, SAMM50 and the catalytic subunit of the DNA‐dependent protein kinase (data not shown). By affinity chromatography and proteomic analysis, ribosomal proteins have already been shown to serve as targets for the NEDD8 pathway. However, ribosomal protein L4 was not identified in that particular approach (Xirodimas et al., 2008). Nonetheless, L4 might be a substrate for NEDDylation mediated by Nce2 isoform 2. SAMM50 plays a role in the assembly of the outer mitochondrial membrane whereas the DNA‐dependent protein kinase is a sensor for DNA damage and mediates DNA repair (Dip and Naegeli, 2005; Kozjak et al., 2003). Additionally, it is found at telomeres where it inhibits the fusion of chromosomal ends (Dip and Naegeli, 2005). Although a direct or indirect

- 103 - Discussion interaction of Nce2 isoform 2 with these proteins needs to be verified, they provide first hints about possible functions of this isoform. For isoform 1, we identified desmoplakin and galectin‐7 as potential interaction partners in the GST‐pulldown (data not shown). These two proteins amongst others are involved in mediating cell‐cell contacts (Delva et al., 2009; Nakahara and Raz, 2006). Moreover, the heavy chain of clathrin, an important coat protein of vesicles (Brodsky et al., 2001), was identified in a yeast‐two hybrid screen that we conducted using Nce2 isoform 1 (data not shown). Hence, these findings may point towards yet unknown functions of Nce2 isoform 1 in cell‐cell interactions and vesicle transport. To gain insights into possible functions of Nce2 variants, we also investigated the third isoform of Nce2. This isoform lacks a part of the APPBP1/UBA3 binding site. We found that overexpressed HA‐Nce2 isoform 3 is hardly detectable in H1299 cells. Thus, one should examine whether this is due to a low expression level or a fast degradation rate. Moreover, Nce2 isoform 3 did not form a thioester bond with NEDD8 in vitro (Bachelor thesis S. Kabelitz, University of Konstanz: “Charakterisierung von NEDD8‐ und Ubiquitin‐konjugierenden Enzymen”). This result met the expectation that Nce2 isoform 3 must be at least impaired in accepting NEDD8 from APPBP1/UBA3 since isoform 3 lacks aa 40‐71, a region containing important residues for the interaction with the E1 enzyme (Huang et al., 2005). In conclusion, isoform 3 may have dominant negative effects on the activity of the first isoform.

In summary, we provide first hints about the existence and the regulation of a second isoform of Nce2 which increases the complexity of the NEDD8‐conjugation cascade and may be of importance to understand the control of NEDDylation in general.

- 104 -

5. References

Abbas, T., Sivaprasad, U., Terai, K., Amador, V., Pagano, M., and Dutta, A. (2008). PCNA‐dependent regulation of p21 ubiquitylation and degradation via the CRL4Cdt2 ubiquitin ligase complex. Genes & development 22, 2496‐2506.

Abida, W.M., Nikolaev, A., Zhao, W., Zhang, W., and Gu, W. (2007). FBXO11 promotes the Neddylation of p53 and inhibits its transcriptional activity. The Journal of biological chemistry 282, 1797‐1804.

Aichem, A., Kalveram, B., Spinnenhirn, V., Kluge, K., Catone, N., Johansen, T., and Groettrup, M. (2012). The proteomic analysis of endogenous FAT10 substrates identifies p62/SQSTM1 as a substrate of FAT10ylation. Journal of cell science.

Aichem, A., Pelzer, C., Lukasiak, S., Kalveram, B., Sheppard, P.W., Rani, N., Schmidtke, G., and Groettrup, M. (2010). USE1 is a bispecific conjugating enzyme for ubiquitin and FAT10, which FAT10ylates itself in cis. Nature communications 1, 13.

Aoki, I., Higuchi, M., and Gotoh, Y. (2012). NEDDylation controls the target specificity of E2F1 and apoptosis induction. Oncogene.

Arakawa, H., Moldovan, G.L., Saribasak, H., Saribasak, N.N., Jentsch, S., and Buerstedde, J.M. (2006). A role for PCNA ubiquitination in immunoglobulin hypermutation. PLoS biology 4, e366.

Ardley, H.C., and Robinson, P.A. (2005). E3 ubiquitin ligases. Essays in biochemistry 41, 15‐30.

Baker, R.T., and Board, P.G. (1987). The human ubiquitin gene family: structure of a gene and pseudogenes from the Ub B subfamily. Nucleic acids research 15, 443‐463.

Baugh, J.M., Viktorova, E.G., and Pilipenko, E.V. (2009). Proteasomes can degrade a significant proportion of cellular proteins independent of ubiquitination. Journal of molecular biology 386, 814‐827.

Bayer, P., Arndt, A., Metzger, S., Mahajan, R., Melchior, F., Jaenicke, R., and Becker, J. (1998). Structure determination of the small ubiquitin‐related modifier SUMO‐1. Journal of molecular biology 280, 275‐286.

Bergink, S., and Jentsch, S. (2009). Principles of ubiquitin and SUMO modifications in DNA repair. Nature 458, 461‐467.

Bernier‐Villamor, V., Sampson, D.A., Matunis, M.J., and Lima, C.D. (2002). Structural basis for E2‐ mediated SUMO conjugation revealed by a complex between ubiquitin‐conjugating enzyme Ubc9 and RanGAP1. Cell 108, 345‐356.

- 105 - References

Bienko, M., Green, C.M., Crosetto, N., Rudolf, F., Zapart, G., Coull, B., Kannouche, P., Wider, G., Peter, M., Lehmann, A.R., et al. (2005). Ubiquitin‐binding domains in Y‐family polymerases regulate translesion synthesis. Science 310, 1821‐1824.

Bienko, M., Green, C.M., Sabbioneda, S., Crosetto, N., Matic, I., Hibbert, R.G., Begovic, T., Niimi, A., Mann, M., Lehmann, A.R., et al. (2010). Regulation of translesion synthesis DNA polymerase eta by monoubiquitination. Molecular cell 37, 396‐407.

Birnboim, H.C., and Doly, J. (1979). A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic acids research 7, 1513‐1523.

Bohren, K.M., Nadkarni, V., Song, J.H., Gabbay, K.H., and Owerbach, D. (2004). A M55V polymorphism in a novel SUMO gene (SUMO‐4) differentially activates heat shock transcription factors and is associated with susceptibility to type I diabetes mellitus. The Journal of biological chemistry 279, 27233‐27238.

Bornstein, G., Bloom, J., Sitry‐Shevah, D., Nakayama, K., Pagano, M., and Hershko, A. (2003). Role of the SCFSkp2 ubiquitin ligase in the degradation of p21Cip1 in S phase. The Journal of biological chemistry 278, 25752‐25757.

Boutet, S.C., Disatnik, M.H., Chan, L.S., Iori, K., and Rando, T.A. (2007). Regulation of Pax3 by proteasomal degradation of monoubiquitinated protein in skeletal muscle progenitors. Cell 130, 349‐362.

Bravo, R., and Macdonald‐Bravo, H. (1985). Changes in the nuclear distribution of cyclin (PCNA) but not its synthesis depend on DNA replication. The EMBO journal 4, 655‐661.

Brodsky, F.M., Chen, C.Y., Knuehl, C., Towler, M.C., and Wakeham, D.E. (2001). Biological basket weaving: formation and function of clathrin‐coated vesicles. Annual review of cell and developmental biology 17, 517‐568.

Broemer, M., Tenev, T., Rigbolt, K.T., Hempel, S., Blagoev, B., Silke, J., Ditzel, M., and Meier, P. (2010). Systematic in vivo RNAi analysis identifies IAPs as NEDD8‐E3 ligases. Molecular cell 40, 810‐822.

Brownell, J.E., Sintchak, M.D., Gavin, J.M., Liao, H., Bruzzese, F.J., Bump, N.J., Soucy, T.A., Milhollen, M.A., Yang, X., Burkhardt, A.L., et al. (2010). Substrate‐assisted inhibition of ubiquitin‐like protein‐activating enzymes: the NEDD8 E1 inhibitor MLN4924 forms a NEDD8‐AMP mimetic in situ. Molecular cell 37, 102‐111.

Bruning, J.B., and Shamoo, Y. (2004). Structural and thermodynamic analysis of human PCNA with peptides derived from DNA polymerase‐delta p66 subunit and flap endonuclease‐1. Structure 12, 2209‐2219.

Brzovic, P.S., Lissounov, A., Christensen, D.E., Hoyt, D.W., and Klevit, R.E. (2006). A UbcH5/ubiquitin noncovalent complex is required for processive BRCA1‐directed ubiquitination. Molecular cell 21, 873‐880.

Cadwell, K., and Coscoy, L. (2005). Ubiquitination on nonlysine residues by a viral E3 ubiquitin ligase. Science 309, 127‐130.

Calabrese, M.F., Scott, D.C., Duda, D.M., Grace, C.R., Kurinov, I., Kriwacki, R.W., and Schulman, B.A. (2011). A RING E3‐substrate complex poised for ubiquitin‐like protein transfer: structural insights into cullin‐RING ligases. Nature structural & molecular biology 18, 947‐949.

- 106 - References

Carreres, M.I., Escalante, A., Murillo, B., Chauvin, G., Gaspar, P., Vegar, C., and Herrera, E. (2011). Transcription factor Foxd1 is required for the specification of the temporal retina in mammals. The Journal of neuroscience : the official journal of the Society for Neuroscience 31, 5673‐5681.

Carter, S., and Vousden, K.H. (2008). p53‐Ubl fusions as models of ubiquitination, sumoylation and neddylation of p53. Cell Cycle 7, 2519‐2528.

Celis, J.E., and Madsen, P. (1986). Increased nuclear cyclin/PCNA antigen staining of non S‐phase transformed human amnion cells engaged in nucleotide excision DNA repair. FEBS letters 209, 277‐283.

Chan, Y., Yoon, J., Wu, J.T., Kim, H.J., Pan, K.T., Yim, J., and Chien, C.T. (2008). DEN1 deneddylates non‐cullin proteins in vivo. Journal of cell science 121, 3218‐3223.

Chau, V., Tobias, J.W., Bachmair, A., Marriott, D., Ecker, D.J., Gonda, D.K., and Varshavsky, A. (1989). A multiubiquitin chain is confined to specific lysine in a targeted short‐lived protein. Science 243, 1576‐1583.

Choo, Y.S., Vogler, G., Wang, D., Kalvakuri, S., Iliuk, A., Tao, W.A., Bodmer, R., and Zhang, Z. (2012). Regulation of parkin and PINK1 by neddylation. Human molecular genetics 21, 2514‐2523.

Choudhary, C., Kumar, C., Gnad, F., Nielsen, M.L., Rehman, M., Walther, T.C., Olsen, J.V., and Mann, M. (2009). Lysine acetylation targets protein complexes and co‐regulates major cellular functions. Science 325, 834‐840.

Ciechanover, A., and Ben‐Saadon, R. (2004). N‐terminal ubiquitination: more protein substrates join in. Trends in cell biology 14, 103‐106.

Cope, G.A., Suh, G.S., Aravind, L., Schwarz, S.E., Zipursky, S.L., Koonin, E.V., and Deshaies, R.J. (2002). Role of predicted metalloprotease motif of Jab1/Csn5 in cleavage of Nedd8 from Cul1. Science 298, 608‐611.

D'Cunha, J., Knight, E., Jr., Haas, A.L., Truitt, R.L., and Borden, E.C. (1996). Immunoregulatory properties of ISG15, an interferon‐induced cytokine. Proceedings of the National Academy of Sciences of the United States of America 93, 211‐215.

Davies, A.A., Huttner, D., Daigaku, Y., Chen, S., and Ulrich, H.D. (2008). Activation of ubiquitin‐ dependent DNA damage bypass is mediated by replication protein a. Molecular cell 29, 625‐636.

Delva, E., Tucker, D.K., and Kowalczyk, A.P. (2009). The desmosome. Cold Spring Harbor perspectives in biology 1, a002543.

Deng, L., Wang, C., Spencer, E., Yang, L., Braun, A., You, J., Slaughter, C., Pickart, C., and Chen, Z.J. (2000). Activation of the IkappaB kinase complex by TRAF6 requires a dimeric ubiquitin‐ conjugating enzyme complex and a unique polyubiquitin chain. Cell 103, 351‐361.

Dentice, M., Bandyopadhyay, A., Gereben, B., Callebaut, I., Christoffolete, M.A., Kim, B.W., Nissim, S., Mornon, J.P., Zavacki, A.M., Zeold, A., et al. (2005). The Hedgehog‐inducible ubiquitin ligase subunit WSB‐1 modulates thyroid hormone activation and PTHrP secretion in the developing growth plate. Nature cell biology 7, 698‐705.

Desterro, J.M., Rodriguez, M.S., Kemp, G.D., and Hay, R.T. (1999). Identification of the enzyme required for activation of the small ubiquitin‐like protein SUMO‐1. The Journal of biological chemistry 274, 10618‐10624.

- 107 - References

Di Fiore, P.P., Polo, S., and Hofmann, K. (2003). When ubiquitin meets ubiquitin receptors: a signalling connection. Nature reviews Molecular cell biology 4, 491‐497.

Dip, R., and Naegeli, H. (2005). More than just strand breaks: the recognition of structural DNA discontinuities by DNA‐dependent protein kinase catalytic subunit. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 19, 704‐715.

Dohmesen, C., Koeppel, M., and Dobbelstein, M. (2008). Specific inhibition of Mdm2‐mediated neddylation by Tip60. Cell Cycle 7, 222‐231.

Duda, D.M., Borg, L.A., Scott, D.C., Hunt, H.W., Hammel, M., and Schulman, B.A. (2008). Structural insights into NEDD8 activation of cullin‐RING ligases: conformational control of conjugation. Cell 134, 995‐1006.

Duncan, L.M., Piper, S., Dodd, R.B., Saville, M.K., Sanderson, C.M., Luzio, J.P., and Lehner, P.J. (2006). Lysine‐63‐linked ubiquitination is required for endolysosomal degradation of class I molecules. The EMBO journal 25, 1635‐1645.

Eddins, M.J., Carlile, C.M., Gomez, K.M., Pickart, C.M., and Wolberger, C. (2006). Mms2‐Ubc13 covalently bound to ubiquitin reveals the structural basis of linkage‐specific polyubiquitin chain formation. Nature structural & molecular biology 13, 915‐920.

Eger, S., Castrec, B., Hubscher, U., Scheffner, M., Rubini, M., and Marx, A. (2011). Generation of a mono‐ubiquitinated PCNA mimic by click chemistry. Chembiochem : a European journal of chemical biology 12, 2807‐2812.

Eger, S., Scheffner, M., Marx, A., and Rubini, M. (2010). Synthesis of defined ubiquitin dimers. Journal of the American Chemical Society 132, 16337‐16339.

Essers, J., Theil, A.F., Baldeyron, C., van Cappellen, W.A., Houtsmuller, A.B., Kanaar, R., and Vermeulen, W. (2005). Nuclear dynamics of PCNA in DNA replication and repair. Molecular and cellular biology 25, 9350‐9359.

Fan, M., Bigsby, R.M., and Nephew, K.P. (2003). The NEDD8 pathway is required for proteasome‐ mediated degradation of human estrogen receptor (ER)‐alpha and essential for the antiproliferative activity of ICI 182,780 in ERalpha‐positive breast cancer cells. Mol Endocrinol 17, 356‐365.

Fan, M., Long, X., Bailey, J.A., Reed, C.A., Osborne, E., Gize, E.A., Kirk, E.A., Bigsby, R.M., and Nephew, K.P. (2002). The activating enzyme of NEDD8 inhibits steroid receptor function. Mol Endocrinol 16, 315‐330.

Fan, W., Cai, W., Parimoo, S., Schwarz, D.C., Lennon, G.G., and Weissman, S.M. (1996). Identification of seven new human MHC class I region genes around the HLA‐F locus. Immunogenetics 44, 97‐103.

Feng, L., Allen, N.S., Simo, S., and Cooper, J.A. (2007). Cullin 5 regulates Dab1 protein levels and neuron positioning during cortical development. Genes & development 21, 2717‐2730.

Furukawa, M., Zhang, Y., McCarville, J., Ohta, T., and Xiong, Y. (2000). The CUL1 C‐terminal sequence and ROC1 are required for efficient nuclear accumulation, NEDD8 modification, and ubiquitin ligase activity of CUL1. Molecular and cellular biology 20, 8185‐8197.

Gali, H., Juhasz, S., Morocz, M., Hajdu, I., Fatyol, K., Szukacsov, V., Burkovics, P., and Haracska, L. (2012). Role of SUMO modification of human PCNA at stalled replication fork. Nucleic acids research.

- 108 - References

Gan‐Erdene, T., Nagamalleswari, K., Yin, L., Wu, K., Pan, Z.Q., and Wilkinson, K.D. (2003). Identification and characterization of DEN1, a deneddylase of the ULP family. The Journal of biological chemistry 278, 28892‐28900.

Gao, F., Cheng, J., Shi, T., and Yeh, E.T. (2006). Neddylation of a breast cancer‐associated protein recruits a class III histone deacetylase that represses NFkappaB‐dependent transcription. Nature cell biology 8, 1171‐1177.

Gazdoiu, S., Yamoah, K., Wu, K., and Pan, Z.Q. (2007). Human Cdc34 employs distinct sites to coordinate attachment of ubiquitin to a substrate and assembly of polyubiquitin chains. Molecular and cellular biology 27, 7041‐7052.

Geng, J., and Klionsky, D.J. (2008). The Atg8 and Atg12 ubiquitin‐like conjugation systems in macroautophagy. 'Protein modifications: beyond the usual suspects' review series. EMBO reports 9, 859‐864.

Godderz, D., Schafer, E., Palanimurugan, R., and Dohmen, R.J. (2011). The N‐terminal unstructured domain of yeast ODC functions as a transplantable and replaceable ubiquitin‐ independent degron. Journal of molecular biology 407, 354‐367.

Gong, L., Kamitani, T., Millas, S., and Yeh, E.T. (2000). Identification of a novel isopeptidase with dual specificity for ubiquitin‐ and NEDD8‐conjugated proteins. The Journal of biological chemistry 275, 14212‐14216.

Gong, L., and Yeh, E.T. (1999). Identification of the activating and conjugating enzymes of the NEDD8 conjugation pathway. The Journal of biological chemistry 274, 12036‐12042.

Groehler, A.L., and Lannigan, D.A. (2010). A chromatin‐bound kinase, ERK8, protects genomic integrity by inhibiting HDM2‐mediated degradation of the DNA clamp PCNA. The Journal of cell biology 190, 575‐586.

Gsponer, J., Futschik, M.E., Teichmann, S.A., and Babu, M.M. (2008). Tight regulation of unstructured proteins: from transcript synthesis to protein degradation. Science 322, 1365‐ 1368.

Haas, A.L., Ahrens, P., Bright, P.M., and Ankel, H. (1987). Interferon induces a 15‐kilodalton protein exhibiting marked homology to ubiquitin. The Journal of biological chemistry 262, 11315‐11323.

Haas, A.L., Bright, P.M., and Jackson, V.E. (1988). Functional diversity among putative E2 isozymes in the mechanism of ubiquitin‐histone ligation. The Journal of biological chemistry 263, 13268‐13275.

Haglund, K., Sigismund, S., Polo, S., Szymkiewicz, I., Di Fiore, P.P., and Dikic, I. (2003). Multiple monoubiquitination of RTKs is sufficient for their endocytosis and degradation. Nature cell biology 5, 461‐466.

Han, K.K., and Martinage, A. (1992). Post‐translational chemical modification(s) of proteins. The International journal of biochemistry 24, 19‐28.

Hashizume, R., Fukuda, M., Maeda, I., Nishikawa, H., Oyake, D., Yabuki, Y., Ogata, H., and Ohta, T. (2001). The RING heterodimer BRCA1‐BARD1 is a ubiquitin ligase inactivated by a breast cancer‐derived mutation. The Journal of biological chemistry 276, 14537‐14540.

- 109 - References

Hatini, V., Huh, S.O., Herzlinger, D., Soares, V.C., and Lai, E. (1996). Essential role of stromal mesenchyme in kidney morphogenesis revealed by targeted disruption of Winged Helix transcription factor BF‐2. Genes & development 10, 1467‐1478.

Havens, C.G., and Walter, J.C. (2009). Docking of a specialized PIP Box onto chromatin‐bound PCNA creates a degron for the ubiquitin ligase CRL4Cdt2. Molecular cell 35, 93‐104.

Havens, C.G., and Walter, J.C. (2011). Mechanism of CRL4(Cdt2), a PCNA‐dependent E3 ubiquitin ligase. Genes & development 25, 1568‐1582.

Hay, R.T. (2005). SUMO: a history of modification. Molecular cell 18, 1‐12.

Herrmann, J., Lerman, L.O., and Lerman, A. (2007). Ubiquitin and ubiquitin‐like proteins in protein regulation. Circulation research 100, 1276‐1291.

Hershko, A., and Ciechanover, A. (1998). The ubiquitin system. Annual review of biochemistry 67, 425‐479.

Hibbert, R.G., Huang, A., Boelens, R., and Sixma, T.K. (2011). E3 ligase Rad18 promotes monoubiquitination rather than ubiquitin chain formation by E2 enzyme Rad6. Proceedings of the National Academy of Sciences of the United States of America 108, 5590‐5595.

Hicke, L. (2001). Protein regulation by monoubiquitin. Nature reviews Molecular cell biology 2, 195‐201.

Hipp, M.S., Kalveram, B., Raasi, S., Groettrup, M., and Schmidtke, G. (2005). FAT10, a ubiquitin‐ independent signal for proteasomal degradation. Molecular and cellular biology 25, 3483‐3491.

Hipp, M.S., Raasi, S., Groettrup, M., and Schmidtke, G. (2004). NEDD8 ultimate buster‐1L interacts with the ubiquitin‐like protein FAT10 and accelerates its degradation. The Journal of biological chemistry 279, 16503‐16510.

Hjerpe, R., Thomas, Y., Chen, J., Zemla, A., Curran, S., Shpiro, N., Dick, L.R., and Kurz, T. (2012). Changes in the ratio of free NEDD8 to ubiquitin triggers NEDDylation by ubiquitin enzymes. The Biochemical journal 441, 927‐936.

Hochstrasser, M. (2000). Evolution and function of ubiquitin‐like protein‐conjugation systems. Nature cell biology 2, E153‐157.

Hochstrasser, M. (2009). Origin and function of ubiquitin‐like proteins. Nature 458, 422‐429.

Hoege, C., Pfander, B., Moldovan, G.L., Pyrowolakis, G., and Jentsch, S. (2002). RAD6‐dependent DNA repair is linked to modification of PCNA by ubiquitin and SUMO. Nature 419, 135‐141.

Hofmann, R.M., and Pickart, C.M. (1999). Noncanonical MMS2‐encoded ubiquitin‐conjugating enzyme functions in assembly of novel polyubiquitin chains for DNA repair. Cell 96, 645‐653.

Honda, R., Tanaka, H., and Yasuda, H. (1997). Oncoprotein MDM2 is a ubiquitin ligase E3 for tumor suppressor p53. FEBS letters 420, 25‐27.

Hoppe, T. (2005). Multiubiquitylation by E4 enzymes: 'one size' doesn't fit all. Trends in biochemical sciences 30, 183‐187.

Hori, T., Osaka, F., Chiba, T., Miyamoto, C., Okabayashi, K., Shimbara, N., Kato, S., and Tanaka, K. (1999). Covalent modification of all members of human cullin family proteins by NEDD8. Oncogene 18, 6829‐6834.

- 110 - References

Huang, D.T., Ayrault, O., Hunt, H.W., Taherbhoy, A.M., Duda, D.M., Scott, D.C., Borg, L.A., Neale, G., Murray, P.J., Roussel, M.F., et al. (2009). E2‐RING expansion of the NEDD8 cascade confers specificity to cullin modification. Molecular cell 33, 483‐495.

Huang, D.T., Hunt, H.W., Zhuang, M., Ohi, M.D., Holton, J.M., and Schulman, B.A. (2007). Basis for a ubiquitin‐like protein thioester switch toggling E1‐E2 affinity. Nature 445, 394‐398.

Huang, D.T., Miller, D.W., Mathew, R., Cassell, R., Holton, J.M., Roussel, M.F., and Schulman, B.A. (2004). A unique E1‐E2 interaction required for optimal conjugation of the ubiquitin‐like protein NEDD8. Nature structural & molecular biology 11, 927‐935.

Huang, D.T., Paydar, A., Zhuang, M., Waddell, M.B., Holton, J.M., and Schulman, B.A. (2005). Structural basis for recruitment of Ubc12 by an E2 binding domain in NEDD8's E1. Molecular cell 17, 341‐350.

Huang, D.T., and Schulman, B.A. (2005). Expression, purification, and characterization of the E1 for human NEDD8, the heterodimeric APPBP1‐UBA3 complex. Methods in enzymology 398, 9‐ 20.

Huang, D.T., Zhuang, M., Ayrault, O., and Schulman, B.A. (2008). Identification of conjugation specificity determinants unmasks vestigial preference for ubiquitin within the NEDD8 E2. Nature structural & molecular biology 15, 280‐287.

Huang, L., Kinnucan, E., Wang, G., Beaudenon, S., Howley, P.M., Huibregtse, J.M., and Pavletich, N.P. (1999). Structure of an E6AP‐UbcH7 complex: insights into ubiquitination by the E2‐E3 enzyme cascade. Science 286, 1321‐1326.

Hubscher, U., and Maga, G. (2011). DNA replication and repair bypass machines. Current opinion in chemical biology 15, 627‐635.

Hurley, J.H., Lee, S., and Prag, G. (2006). Ubiquitin‐binding domains. The Biochemical journal 399, 361‐372.

Ichimura, Y., Kirisako, T., Takao, T., Satomi, Y., Shimonishi, Y., Ishihara, N., Mizushima, N., Tanida, I., Kominami, E., Ohsumi, M., et al. (2000). A ubiquitin‐like system mediates protein lipidation. Nature 408, 488‐492.

Ivan, M., Kondo, K., Yang, H., Kim, W., Valiando, J., Ohh, M., Salic, A., Asara, J.M., Lane, W.S., and Kaelin, W.G., Jr. (2001). HIFalpha targeted for VHL‐mediated destruction by proline hydroxylation: implications for O2 sensing. Science 292, 464‐468.

Jaakkola, P., Mole, D.R., Tian, Y.M., Wilson, M.I., Gielbert, J., Gaskell, S.J., Kriegsheim, A., Hebestreit, H.F., Mukherji, M., Schofield, C.J., et al. (2001). Targeting of HIF‐alpha to the von Hippel‐Lindau ubiquitylation complex by O2‐regulated prolyl hydroxylation. Science 292, 468‐472.

Jeram, S.M., Srikumar, T., Zhang, X.D., Anne Eisenhauer, H., Rogers, R., Pedrioli, P.G., Matunis, M., and Raught, B. (2010). An improved SUMmOn‐based methodology for the identification of ubiquitin and ubiquitin‐like protein conjugation sites identifies novel ubiquitin‐like protein chain linkages. Proteomics 10, 254‐265.

Jin, H.S., Liao, L., Park, Y., and Liu, Y.C. (2013). Neddylation pathway regulates T‐cell function by targeting an adaptor protein Shc and a protein kinase Erk signaling. Proceedings of the National Academy of Sciences of the United States of America 110, 624‐629.

Jin, J., Li, X., Gygi, S.P., and Harper, J.W. (2007). Dual E1 activation systems for ubiquitin differentially regulate E2 enzyme charging. Nature 447, 1135‐1138.

- 111 - References

Jin, L., Williamson, A., Banerjee, S., Philipp, I., and Rape, M. (2008). Mechanism of ubiquitin‐chain formation by the human anaphase‐promoting complex. Cell 133, 653‐665.

Johnson, E.S. (2004). Protein modification by SUMO. Annual review of biochemistry 73, 355‐382.

Johnson, E.S., and Gupta, A.A. (2001). An E3‐like factor that promotes SUMO conjugation to the yeast septins. Cell 106, 735‐744.

Johnson, E.S., Schwienhorst, I., Dohmen, R.J., and Blobel, G. (1997). The ubiquitin‐like protein Smt3p is activated for conjugation to other proteins by an Aos1p/Uba2p heterodimer. The EMBO journal 16, 5509‐5519.

Jones, J., Wu, K., Yang, Y., Guerrero, C., Nillegoda, N., Pan, Z.Q., and Huang, L. (2008). A targeted proteomic analysis of the ubiquitin‐like modifier nedd8 and associated proteins. Journal of proteome research 7, 1274‐1287.

Ju, T., Bocik, W., Majumdar, A., and Tolman, J.R. (2010). Solution structure and dynamics of human ubiquitin conjugating enzyme Ube2g2. Proteins 78, 1291‐1301.

Kaestner, K.H. (2010). The FoxA factors in organogenesis and differentiation. Current opinion in genetics & development 20, 527‐532.

Kagey, M.H., Melhuish, T.A., and Wotton, D. (2003). The polycomb protein Pc2 is a SUMO E3. Cell 113, 127‐137.

Kamitani, T., Kito, K., Fukuda‐Kamitani, T., and Yeh, E.T. (2001). Targeting of NEDD8 and its conjugates for proteasomal degradation by NUB1. The Journal of biological chemistry 276, 46655‐46660.

Kamitani, T., Kito, K., Nguyen, H.P., and Yeh, E.T. (1997). Characterization of NEDD8, a developmentally down‐regulated ubiquitin‐like protein. The Journal of biological chemistry 272, 28557‐28562.

Kamitani, T., Nguyen, H.P., Kito, K., Fukuda‐Kamitani, T., and Yeh, E.T. (1998). Covalent modification of PML by the sentrin family of ubiquitin‐like proteins. The Journal of biological chemistry 273, 3117‐3120.

Kannouche, P.L., Wing, J., and Lehmann, A.R. (2004). Interaction of human DNA polymerase eta with monoubiquitinated PCNA: a possible mechanism for the polymerase switch in response to DNA damage. Molecular cell 14, 491‐500.

Kawakami, T., Chiba, T., Suzuki, T., Iwai, K., Yamanaka, K., Minato, N., Suzuki, H., Shimbara, N., Hidaka, Y., Osaka, F., et al. (2001). NEDD8 recruits E2‐ubiquitin to SCF E3 ligase. The EMBO journal 20, 4003‐4012.

Kelman, Z., and O'Donnell, M. (1995). Structural and functional similarities of prokaryotic and eukaryotic DNA polymerase sliding clamps. Nucleic acids research 23, 3613‐3620.

Kerscher, O., Felberbaum, R., and Hochstrasser, M. (2006). Modification of proteins by ubiquitin and ubiquitin‐like proteins. Annual review of cell and developmental biology 22, 159‐180.

Kile, B.T., Schulman, B.A., Alexander, W.S., Nicola, N.A., Martin, H.M., and Hilton, D.J. (2002). The SOCS box: a tale of destruction and degradation. Trends in biochemical sciences 27, 235‐241.

Kirkin, V., McEwan, D.G., Novak, I., and Dikic, I. (2009). A role for ubiquitin in selective autophagy. Molecular cell 34, 259‐269.

- 112 - References

Kitagawa, K., Kotake, Y., and Kitagawa, M. (2009). Ubiquitin‐mediated control of oncogene and tumor suppressor gene products. Cancer science 100, 1374‐1381.

Knipscheer, P., Flotho, A., Klug, H., Olsen, J.V., van Dijk, W.J., Fish, A., Johnson, E.S., Mann, M., Sixma, T.K., and Pichler, A. (2008). Ubc9 sumoylation regulates SUMO target discrimination. Molecular cell 31, 371‐382.

Knipscheer, P., van Dijk, W.J., Olsen, J.V., Mann, M., and Sixma, T.K. (2007). Noncovalent interaction between Ubc9 and SUMO promotes SUMO chain formation. The EMBO journal 26, 2797‐2807.

Korant, B.D., Blomstrom, D.C., Jonak, G.J., and Knight, E., Jr. (1984). Interferon‐induced proteins. Purification and characterization of a 15,000‐dalton protein from human and bovine cells induced by interferon. The Journal of biological chemistry 259, 14835‐14839.

Kozjak, V., Wiedemann, N., Milenkovic, D., Lohaus, C., Meyer, H.E., Guiard, B., Meisinger, C., and Pfanner, N. (2003). An essential role of Sam50 in the protein sorting and assembly machinery of the mitochondrial outer membrane. The Journal of biological chemistry 278, 48520‐48523.

Kravtsova‐Ivantsiv, Y., Cohen, S., and Ciechanover, A. (2009). Modification by single ubiquitin moieties rather than polyubiquitination is sufficient for proteasomal processing of the p105 NF‐ kappaB precursor. Molecular cell 33, 496‐504.

Krishna, T.S., Kong, X.P., Gary, S., Burgers, P.M., and Kuriyan, J. (1994). Crystal structure of the eukaryotic DNA polymerase processivity factor PCNA. Cell 79, 1233‐1243.

Kruse, J.P., and Gu, W. (2009). Modes of p53 regulation. Cell 137, 609‐622.

Kumar, S., Tomooka, Y., and Noda, M. (1992). Identification of a set of genes with developmentally down‐regulated expression in the mouse brain. Biochemical and biophysical research communications 185, 1155‐1161.

Kumar, S., Yoshida, Y., and Noda, M. (1993). Cloning of a cDNA which encodes a novel ubiquitin‐ like protein. Biochemical and biophysical research communications 195, 393‐399.

Kurz, T., Chou, Y.C., Willems, A.R., Meyer‐Schaller, N., Hecht, M.L., Tyers, M., Peter, M., and Sicheri, F. (2008). Dcn1 functions as a scaffold‐type E3 ligase for cullin neddylation. Molecular cell 29, 23‐35.

Laemmli, U.K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680‐685.

Lake, M.W., Wuebbens, M.M., Rajagopalan, K.V., and Schindelin, H. (2001). Mechanism of ubiquitin activation revealed by the structure of a bacterial MoeB‐MoaD complex. Nature 414, 325‐329.

Lee, I., and Schindelin, H. (2008). Structural insights into E1‐catalyzed ubiquitin activation and transfer to conjugating enzymes. Cell 134, 268‐278.

Lee, M.R., Lee, D., Shin, S.K., Kim, Y.H., and Choi, C.Y. (2008). Inhibition of APP intracellular domain (AICD) transcriptional activity via covalent conjugation with Nedd8. Biochemical and biophysical research communications 366, 976‐981.

Lee, S.H., and Hurwitz, J. (1990). Mechanism of elongation of primed DNA by DNA polymerase delta, proliferating cell nuclear antigen, and activator 1. Proceedings of the National Academy of Sciences of the United States of America 87, 5672‐5676.

- 113 - References

Leidecker, O., Matic, I., Mahata, B., Pion, E., and Xirodimas, D.P. (2012). The ubiquitin E1 enzyme Ube1 mediates NEDD8 activation under diverse stress conditions. Cell Cycle 11.

Li, T., Santockyte, R., Yu, S., Shen, R.F., Tekle, E., Lee, C.G., Yang, D.C., and Chock, P.B. (2011). FAT10 modifies p53 and upregulates its transcriptional activity. Archives of biochemistry and biophysics 509, 164‐169.

Li, W., Bengtson, M.H., Ulbrich, A., Matsuda, A., Reddy, V.A., Orth, A., Chanda, S.K., Batalov, S., and Joazeiro, C.A. (2008). Genome‐wide and functional annotation of human E3 ubiquitin ligases identifies MULAN, a mitochondrial E3 that regulates the organelle's dynamics and signaling. PloS one 3, e1487.

Li, X. (2011). Click to join peptides/proteins together. Chemistry, an Asian journal 6, 2606‐2616.

Liakopoulos, D., Doenges, G., Matuschewski, K., and Jentsch, S. (1998). A novel protein modification pathway related to the ubiquitin system. The EMBO journal 17, 2208‐2214.

Liew, C.W., Sun, H., Hunter, T., and Day, C.L. (2010). RING domain dimerization is essential for RNF4 function. The Biochemical journal 431, 23‐29.

Lin, J.J., Milhollen, M.A., Smith, P.G., Narayanan, U., and Dutta, A. (2010). NEDD8‐targeting drug MLN4924 elicits DNA rereplication by stabilizing Cdt1 in S phase, triggering checkpoint activation, apoptosis, and senescence in cancer cells. Cancer research 70, 10310‐10320.

Lin, Y., Hwang, W.C., and Basavappa, R. (2002). Structural and functional analysis of the human mitotic‐specific ubiquitin‐conjugating enzyme, UbcH10. The Journal of biological chemistry 277, 21913‐21921.

Linares, L.K., Hengstermann, A., Ciechanover, A., Muller, S., and Scheffner, M. (2003). HdmX stimulates Hdm2‐mediated ubiquitination and degradation of p53. Proceedings of the National Academy of Sciences of the United States of America 100, 12009‐12014.

Liu, G., and Warbrick, E. (2006). The p66 and p12 subunits of DNA polymerase delta are modified by ubiquitin and ubiquitin‐like proteins. Biochemical and biophysical research communications 349, 360‐366.

Liu, G., and Xirodimas, D.P. (2010). NUB1 promotes cytoplasmic localization of p53 through cooperation of the NEDD8 and ubiquitin pathways. Oncogene 29, 2252‐2261.

Liu, L., Rodriguez‐Belmonte, E.M., Mazloum, N., Xie, B., and Lee, M.Y. (2003). Identification of a novel protein, PDIP38, that interacts with the p50 subunit of DNA polymerase delta and proliferating cell nuclear antigen. The Journal of biological chemistry 278, 10041‐10047.

Lo, Y.H., Ho, P.C., and Wang, S.C. (2012). Epidermal growth factor receptor protects proliferating cell nuclear antigen from cullin 4A protein‐mediated proteolysis. The Journal of biological chemistry 287, 27148‐27157.

Lois, L.M., and Lima, C.D. (2005). Structures of the SUMO E1 provide mechanistic insights into SUMO activation and E2 recruitment to E1. The EMBO journal 24, 439‐451.

Lovering, R., Hanson, I.M., Borden, K.L., Martin, S., O'Reilly, N.J., Evan, G.I., Rahman, D., Pappin, D.J., Trowsdale, J., and Freemont, P.S. (1993). Identification and preliminary characterization of a protein motif related to the zinc finger. Proceedings of the National Academy of Sciences of the United States of America 90, 2112‐2116.

- 114 - References

Lund, P.K., Moats‐Staats, B.M., Simmons, J.G., Hoyt, E., D'Ercole, A.J., Martin, F., and Van Wyk, J.J. (1985). Nucleotide sequence analysis of a cDNA encoding human ubiquitin reveals that ubiquitin is synthesized as a precursor. The Journal of biological chemistry 260, 7609‐7613.

Lyapina, S., Cope, G., Shevchenko, A., Serino, G., Tsuge, T., Zhou, C., Wolf, D.A., Wei, N., and Deshaies, R.J. (2001). Promotion of NEDD‐CUL1 conjugate cleavage by COP9 signalosome. Science 292, 1382‐1385.

Lyapina, S.A., Correll, C.C., Kipreos, E.T., and Deshaies, R.J. (1998). Human CUL1 forms an evolutionarily conserved ubiquitin ligase complex (SCF) with SKP1 and an F‐box protein. Proceedings of the National Academy of Sciences of the United States of America 95, 7451‐7456.

Ma, T., Chen, Y., Zhang, F., Yang, C.Y., Wang, S., and Yu, X. (2013). RNF111‐Dependent Neddylation Activates DNA Damage‐Induced Ubiquitination. Molecular cell.

Machida, Y.J., Machida, Y., Chen, Y., Gurtan, A.M., Kupfer, G.M., D'Andrea, A.D., and Dutta, A. (2006). UBE2T is the E2 in the Fanconi anemia pathway and undergoes negative autoregulation. Molecular cell 23, 589‐596.

Maeda, M., Ohashi, K., and Ohashi‐Kobayashi, A. (2005). Further extension of mammalian GATA‐ 6. Development, growth & differentiation 47, 591‐600.

Magee, A.I., and Courtneidge, S.A. (1985). Two classes of fatty acid acylated proteins exist in eukaryotic cells. The EMBO journal 4, 1137‐1144.

Mahata, B., Sundqvist, A., and Xirodimas, D.P. (2011). Recruitment of RPL11 at promoter sites of p53‐regulated genes upon nucleolar stress through NEDD8 and in an Mdm2‐dependent manner. Oncogene.

Mahotka, C., Liebmann, J., Wenzel, M., Suschek, C.V., Schmitt, M., Gabbert, H.E., and Gerharz, C.D. (2002). Differential subcellular localization of functionally divergent survivin splice variants. Cell death and differentiation 9, 1334‐1342.

Matunis, M.J., Coutavas, E., and Blobel, G. (1996). A novel ubiquitin‐like modification modulates the partitioning of the Ran‐GTPase‐activating protein RanGAP1 between the cytosol and the nuclear pore complex. The Journal of cell biology 135, 1457‐1470.

Maxwell, P. (2003). HIF‐1: an oxygen response system with special relevance to the kidney. Journal of the American Society of Nephrology : JASN 14, 2712‐2722.

McKenna, S., Spyracopoulos, L., Moraes, T., Pastushok, L., Ptak, C., Xiao, W., and Ellison, M.J. (2001). Noncovalent interaction between ubiquitin and the human DNA repair protein Mms2 is required for Ubc13‐mediated polyubiquitination. The Journal of biological chemistry 276, 40120‐40126.

Mendoza, H.M., Shen, L.N., Botting, C., Lewis, A., Chen, J., Ink, B., and Hay, R.T. (2003). NEDP1, a highly conserved cysteine protease that deNEDDylates Cullins. The Journal of biological chemistry 278, 25637‐25643.

Metzger, M.B., Hristova, V.A., and Weissman, A.M. (2012). HECT and RING finger families of E3 ubiquitin ligases at a glance. Journal of cell science 125, 531‐537.

Meyer‐Schaller, N., Chou, Y.C., Sumara, I., Martin, D.D., Kurz, T., Katheder, N., Hofmann, K., Berthiaume, L.G., Sicheri, F., and Peter, M. (2009). The human Dcn1‐like protein DCNL3 promotes Cul3 neddylation at membranes. Proceedings of the National Academy of Sciences of the United States of America 106, 12365‐12370.

- 115 - References

Milhollen, M.A., Traore, T., Adams‐Duffy, J., Thomas, M.P., Berger, A.J., Dang, L., Dick, L.R., Garnsey, J.J., Koenig, E., Langston, S.P., et al. (2010). MLN4924, a NEDD8‐activating enzyme inhibitor, is active in diffuse large B‐cell lymphoma models: rationale for treatment of NF‐ {kappa}B‐dependent lymphoma. Blood 116, 1515‐1523.

Miura, T., Klaus, W., Gsell, B., Miyamoto, C., and Senn, H. (1999). Characterization of the binding interface between ubiquitin and class I human ubiquitin‐conjugating enzyme 2b by multidimensional heteronuclear NMR spectroscopy in solution. Journal of molecular biology 290, 213‐228.

Miyachi, K., Fritzler, M.J., and Tan, E.M. (1978). Autoantibody to a nuclear antigen in proliferating cells. J Immunol 121, 2228‐2234.

Moldovan, G.L., Dejsuphong, D., Petalcorin, M.I., Hofmann, K., Takeda, S., Boulton, S.J., and D'Andrea, A.D. (2012). Inhibition of homologous recombination by the PCNA‐interacting protein PARI. Molecular cell 45, 75‐86.

Monda, J.K., Scott, D.C., Miller, D.J., Lydeard, J., King, D., Harper, J.W., Bennett, E.J., and Schulman, B.A. (2013). Structural conservation of distinctive N‐terminal acetylation‐dependent interactions across a family of mammalian NEDD8 ligation enzymes. Structure 21, 42‐53.

Morimoto, M., Nishida, T., Nagayama, Y., and Yasuda, H. (2003). Nedd8‐modification of Cul1 is promoted by Roc1 as a Nedd8‐E3 ligase and regulates its stability. Biochemical and biophysical research communications 301, 392‐398.

Mossi, R., and Hubscher, U. (1998). Clamping down on clamps and clamp loaders‐‐the eukaryotic replication factor C. European journal of biochemistry / FEBS 254, 209‐216.

Muller, T., Meyer, H.E., Egensperger, R., and Marcus, K. (2008). The amyloid precursor protein intracellular domain (AICD) as modulator of gene expression, apoptosis, and cytoskeletal dynamics‐relevance for Alzheimer's disease. Progress in neurobiology 85, 393‐406.

Murata, S., Yashiroda, H., and Tanaka, K. (2009). Molecular mechanisms of proteasome assembly. Nature reviews Molecular cell biology 10, 104‐115.

Nagano, T., Hashimoto, T., Nakashima, A., Kikkawa, U., and Kamada, S. (2012). X‐linked inhibitor of apoptosis protein mediates neddylation by itself but does not function as a NEDD8‐E3 ligase for caspase‐7. FEBS letters.

Nakahara, S., and Raz, A. (2006). On the role of galectins in signal transduction. Methods in enzymology 417, 273‐289.

Nakayama, K., Nagahama, H., Minamishima, Y.A., Matsumoto, M., Nakamichi, I., Kitagawa, K., Shirane, M., Tsunematsu, R., Tsukiyama, T., Ishida, N., et al. (2000). Targeted disruption of Skp2 results in accumulation of cyclin E and p27(Kip1), polyploidy and centrosome overduplication. The EMBO journal 19, 2069‐2081.

Nastasi, T., Bongiovanni, A., Campos, Y., Mann, L., Toy, J.N., Bostrom, J., Rottier, R., Hahn, C., Conaway, J.W., Harris, A.J., et al. (2004). Ozz‐E3, a muscle‐specific ubiquitin ligase, regulates beta‐catenin degradation during myogenesis. Developmental cell 6, 269‐282.

Noguchi, K., Okumura, F., Takahashi, N., Kataoka, A., Kamiyama, T., Todo, S., and Hatakeyama, S. (2011). TRIM40 promotes neddylation of IKKgamma and is downregulated in gastrointestinal cancers. Carcinogenesis 32, 995‐1004.

- 116 - References

Ohh, M., Kim, W.Y., Moslehi, J.J., Chen, Y., Chau, V., Read, M.A., and Kaelin, W.G., Jr. (2002). An intact NEDD8 pathway is required for Cullin‐dependent ubiquitylation in mammalian cells. EMBO reports 3, 177‐182.

Osaka, F., Kawasaki, H., Aida, N., Saeki, M., Chiba, T., Kawashima, S., Tanaka, K., and Kato, S. (1998). A new NEDD8‐ligating system for cullin‐4A. Genes & development 12, 2263‐2268.

Oved, S., Mosesson, Y., Zwang, Y., Santonico, E., Shtiegman, K., Marmor, M.D., Kochupurakkal, B.S., Katz, M., Lavi, S., Cesareni, G., et al. (2006). Conjugation to Nedd8 instigates ubiquitylation and down‐regulation of activated receptor tyrosine kinases. The Journal of biological chemistry 281, 21640‐21651.

Owerbach, D., McKay, E.M., Yeh, E.T., Gabbay, K.H., and Bohren, K.M. (2005). A proline‐90 residue unique to SUMO‐4 prevents maturation and sumoylation. Biochemical and biophysical research communications 337, 517‐520.

Parker, J.L., and Ulrich, H.D. (2012). A SUMO‐interacting motif activates budding yeast ubiquitin ligase Rad18 towards SUMO‐modified PCNA. Nucleic acids research 40, 11380‐11388.

Pelzer, C., Kassner, I., Matentzoglu, K., Singh, R.K., Wollscheid, H.P., Scheffner, M., Schmidtke, G., and Groettrup, M. (2007). UBE1L2, a novel E1 enzyme specific for ubiquitin. The Journal of biological chemistry 282, 23010‐23014.

Peters, J.M. (2006). The anaphase promoting complex/cyclosome: a machine designed to destroy. Nature reviews Molecular cell biology 7, 644‐656.

Petroski, M.D., and Deshaies, R.J. (2005a). Function and regulation of cullin‐RING ubiquitin ligases. Nature reviews Molecular cell biology 6, 9‐20.

Petroski, M.D., and Deshaies, R.J. (2005b). Mechanism of lysine 48‐linked ubiquitin‐chain synthesis by the cullin‐RING ubiquitin‐ligase complex SCF‐Cdc34. Cell 123, 1107‐1120.

Pfander, B., Moldovan, G.L., Sacher, M., Hoege, C., and Jentsch, S. (2005). SUMO‐modified PCNA recruits Srs2 to prevent recombination during S phase. Nature 436, 428‐433.

Pichler, A., Gast, A., Seeler, J.S., Dejean, A., and Melchior, F. (2002). The nucleoporin RanBP2 has SUMO1 E3 ligase activity. Cell 108, 109‐120.

Pichler, A., Knipscheer, P., Oberhofer, E., van Dijk, W.J., Korner, R., Olsen, J.V., Jentsch, S., Melchior, F., and Sixma, T.K. (2005). SUMO modification of the ubiquitin‐conjugating enzyme E2‐25K. Nature structural & molecular biology 12, 264‐269.

Pickart, C.M. (2001). Mechanisms underlying ubiquitination. Annual review of biochemistry 70, 503‐533.

Pierce, N.W., Lee, J.E., Liu, X., Sweredoski, M.J., Graham, R.L., Larimore, E.A., Rome, M., Zheng, N., Clurman, B.E., Hess, S., et al. (2013). Cand1 Promotes Assembly of New SCF Complexes through Dynamic Exchange of F Box Proteins. Cell.

Plechanovova, A., Jaffray, E.G., McMahon, S.A., Johnson, K.A., Navratilova, I., Naismith, J.H., and Hay, R.T. (2011). Mechanism of ubiquitylation by dimeric RING ligase RNF4. Nature structural & molecular biology 18, 1052‐1059.

Plechanovova, A., Jaffray, E.G., Tatham, M.H., Naismith, J.H., and Hay, R.T. (2012). Structure of a RING E3 ligase and ubiquitin‐loaded E2 primed for catalysis. Nature 489, 115‐120.

- 117 - References

Prakash, S., Tian, L., Ratliff, K.S., Lehotzky, R.E., and Matouschek, A. (2004). An unstructured initiation site is required for efficient proteasome‐mediated degradation. Nature structural & molecular biology 11, 830‐837.

Ptak, C., Prendergast, J.A., Hodgins, R., Kay, C.M., Chau, V., and Ellison, M.J. (1994). Functional and physical characterization of the cell cycle ubiquitin‐conjugating enzyme CDC34 (UBC3). Identification of a functional determinant within the tail that facilitates CDC34 self‐association. The Journal of biological chemistry 269, 26539‐26545.

Raasi, S., Schmidtke, G., and Groettrup, M. (2001). The ubiquitin‐like protein FAT10 forms covalent conjugates and induces apoptosis. The Journal of biological chemistry 276, 35334‐ 35343.

Rani, N., Aichem, A., Schmidtke, G., Kreft, S.G., and Groettrup, M. (2012). FAT10 and NUB1L bind to the VWA domain of Rpn10 and Rpn1 to enable proteasome‐mediated proteolysis. Nature communications 3, 749.

Ravid, T., and Hochstrasser, M. (2007). Autoregulation of an E2 enzyme by ubiquitin‐chain assembly on its catalytic residue. Nature cell biology 9, 422‐427.

Reverter, D., and Lima, C.D. (2005). Insights into E3 ligase activity revealed by a SUMO‐ RanGAP1‐Ubc9‐Nup358 complex. Nature 435, 687‐692.

Rivers, S.L., McNairn, E., Blasco, F., Giordano, G., and Boxer, D.H. (1993). Molecular genetic analysis of the moa operon of Escherichia coli K‐12 required for molybdenum cofactor biosynthesis. Molecular microbiology 8, 1071‐1081.

Rock, K.L., Gramm, C., Rothstein, L., Clark, K., Stein, R., Dick, L., Hwang, D., and Goldberg, A.L. (1994). Inhibitors of the proteasome block the degradation of most cell proteins and the generation of peptides presented on MHC class I molecules. Cell 78, 761‐771.

Rodriguez, M.S., Desterro, J.M., Lain, S., Lane, D.P., and Hay, R.T. (2000). Multiple C‐terminal lysine residues target p53 for ubiquitin‐proteasome‐mediated degradation. Molecular and cellular biology 20, 8458‐8467.

Rodriguez, M.S., Desterro, J.M., Lain, S., Midgley, C.A., Lane, D.P., and Hay, R.T. (1999). SUMO‐1 modification activates the transcriptional response of p53. The EMBO journal 18, 6455‐6461.

Rosas‐Acosta, G., Russell, W.K., Deyrieux, A., Russell, D.H., and Wilson, V.G. (2005). A universal strategy for proteomic studies of SUMO and other ubiquitin‐like modifiers. Molecular & cellular proteomics : MCP 4, 56‐72.

Russell, R.C., and Ohh, M. (2008). NEDD8 acts as a 'molecular switch' defining the functional selectivity of VHL. EMBO reports 9, 486‐491.

Saitoh, H., and Hinchey, J. (2000). Functional heterogeneity of small ubiquitin‐related protein modifiers SUMO‐1 versus SUMO‐2/3. The Journal of biological chemistry 275, 6252‐6258.

Sakata, E., Satoh, T., Yamamoto, S., Yamaguchi, Y., Yagi‐Utsumi, M., Kurimoto, E., Tanaka, K., Wakatsuki, S., and Kato, K. (2010). Crystal structure of UbcH5b~ubiquitin intermediate: insight into the formation of the self‐assembled E2~Ub conjugates. Structure 18, 138‐147.

Sakata, E., Yamaguchi, Y., Miyauchi, Y., Iwai, K., Chiba, T., Saeki, Y., Matsuda, N., Tanaka, K., and Kato, K. (2007). Direct interactions between NEDD8 and ubiquitin E2 conjugating enzymes upregulate cullin‐based E3 ligase activity. Nature structural & molecular biology 14, 167‐168.

- 118 - References

Scheffner, M., Huibregtse, J.M., Vierstra, R.D., and Howley, P.M. (1993). The HPV‐16 E6 and E6‐AP complex functions as a ubiquitin‐protein ligase in the ubiquitination of p53. Cell 75, 495‐505.

Schneider, D., Schneider, T., Rösner, D., Scheffner, M., and Marx, A. (2013). Improving bioorthogonal protein ubiquitylation by Click reaction. Bioorganic & Medicinal Chemistry in press.

Schulman, B.A., and Harper, J.W. (2009). Ubiquitin‐like protein activation by E1 enzymes: the apex for downstream signalling pathways. Nature reviews Molecular cell biology 10, 319‐331.

Schwarz, S.E., Matuschewski, K., Liakopoulos, D., Scheffner, M., and Jentsch, S. (1998). The ubiquitin‐like proteins SMT3 and SUMO‐1 are conjugated by the UBC9 E2 enzyme. Proceedings of the National Academy of Sciences of the United States of America 95, 560‐564.

Scott, D.C., Monda, J.K., Bennett, E.J., Harper, J.W., and Schulman, B.A. (2011). N‐terminal acetylation acts as an avidity enhancer within an interconnected multiprotein complex. Science 334, 674‐678.

Senga, T., Sivaprasad, U., Zhu, W., Park, J.H., Arias, E.E., Walter, J.C., and Dutta, A. (2006). PCNA is a cofactor for Cdt1 degradation by CUL4/DDB1‐mediated N‐terminal ubiquitination. The Journal of biological chemistry 281, 6246‐6252.

Shabek, N., Herman‐Bachinsky, Y., Buchsbaum, S., Lewinson, O., Haj‐Yahya, M., Hejjaoui, M., Lashuel, H.A., Sommer, T., Brik, A., and Ciechanover, A. (2012). The size of the proteasomal substrate determines whether its degradation will be mediated by mono‐ or polyubiquitylation. Molecular cell 48, 87‐97.

Shen, T.H., Lin, H.K., Scaglioni, P.P., Yung, T.M., and Pandolfi, P.P. (2006). The mechanisms of PML‐nuclear body formation. Molecular cell 24, 331‐339.

Shimizu, Y., Okuda‐Shimizu, Y., and Hendershot, L.M. (2010). Ubiquitylation of an ERAD substrate occurs on multiple types of amino acids. Molecular cell 40, 917‐926.

Siepmann, T.J., Bohnsack, R.N., Tokgoz, Z., Baboshina, O.V., and Haas, A.L. (2003). Protein interactions within the N‐end rule ubiquitin ligation pathway. The Journal of biological chemistry 278, 9448‐9457.

Singh, R.K., Zerath, S., Kleifeld, O., Scheffner, M., Glickman, M.H., and Fushman, D. (2012). Recognition and cleavage of related to ubiquitin 1 (Rub1) and Rub1‐ubiquitin chains by components of the ubiquitin‐proteasome system. Molecular & cellular proteomics : MCP 11, 1595‐1611.

Skaar, J.R., Florens, L., Tsutsumi, T., Arai, T., Tron, A., Swanson, S.K., Washburn, M.P., and DeCaprio, J.A. (2007). PARC and CUL7 form atypical cullin RING ligase complexes. Cancer research 67, 2006‐2014.

Sorokin, A.V., Kim, E.R., and Ovchinnikov, L.P. (2009). Proteasome system of protein degradation and processing. Biochemistry Biokhimiia 74, 1411‐1442.

Soucy, T.A., Smith, P.G., Milhollen, M.A., Berger, A.J., Gavin, J.M., Adhikari, S., Brownell, J.E., Burke, K.E., Cardin, D.P., Critchley, S., et al. (2009). An inhibitor of NEDD8‐activating enzyme as a new approach to treat cancer. Nature 458, 732‐736.

Souphron, J., Waddell, M.B., Paydar, A., Tokgoz‐Gromley, Z., Roussel, M.F., and Schulman, B.A. (2008). Structural dissection of a gating mechanism preventing misactivation of ubiquitin by NEDD8's E1. Biochemistry 47, 8961‐8969.

- 119 - References

Spence, J., Sadis, S., Haas, A.L., and Finley, D. (1995). A ubiquitin mutant with specific defects in DNA repair and multiubiquitination. Molecular and cellular biology 15, 1265‐1273.

Stelter, P., and Ulrich, H.D. (2003). Control of spontaneous and damage‐induced mutagenesis by SUMO and ubiquitin conjugation. Nature 425, 188‐191.

Sternsdorf, T., Jensen, K., and Will, H. (1997). Evidence for covalent modification of the nuclear dot‐associated proteins PML and Sp100 by PIC1/SUMO‐1. The Journal of cell biology 139, 1621‐ 1634.

Stickle, N.H., Chung, J., Klco, J.M., Hill, R.P., Kaelin, W.G., Jr., and Ohh, M. (2004). pVHL modification by NEDD8 is required for fibronectin matrix assembly and suppression of tumor development. Molecular and cellular biology 24, 3251‐3261.

Stoimenov, I., and Helleday, T. (2009). PCNA on the crossroad of cancer. Biochemical Society transactions 37, 605‐613.

Sugawara, K., Suzuki, N.N., Fujioka, Y., Mizushima, N., Ohsumi, Y., and Inagaki, F. (2004). The crystal structure of microtubule‐associated protein light chain 3, a mammalian homologue of Saccharomyces cerevisiae Atg8. Genes to cells : devoted to molecular & cellular mechanisms 9, 611‐618.

Sundqvist, A., Liu, G., Mirsaliotis, A., and Xirodimas, D.P. (2009). Regulation of nucleolar signalling to p53 through NEDDylation of L11. EMBO reports 10, 1132‐1139.

Suraweera, A., Munch, C., Hanssum, A., and Bertolotti, A. (2012). Failure of amino acid homeostasis causes cell death following proteasome inhibition. Molecular cell 48, 242‐253.

Suzuki, N.N., Yoshimoto, K., Fujioka, Y., Ohsumi, Y., and Inagaki, F. (2005). The crystal structure of plant ATG12 and its biological implication in autophagy. Autophagy 1, 119‐126.

Swords, R.T., Kelly, K.R., Smith, P.G., Garnsey, J.J., Mahalingam, D., Medina, E., Oberheu, K., Padmanabhan, S., O'Dwyer, M., Nawrocki, S.T., et al. (2010). Inhibition of NEDD8‐activating enzyme: a novel approach for the treatment of acute myeloid leukemia. Blood 115, 3796‐3800.

Szafranski, K., Schindler, S., Taudien, S., Hiller, M., Huse, K., Jahn, N., Schreiber, S., Backofen, R., and Platzer, M. (2007). Violating the splicing rules: TG dinucleotides function as alternative 3' splice sites in U2‐dependent introns. Genome biology 8, R154.

Szczepanowski, R.H., Filipek, R., and Bochtler, M. (2005). Crystal structure of a fragment of mouse ubiquitin‐activating enzyme. The Journal of biological chemistry 280, 22006‐22011.

Takasaki, Y., Deng, J.S., and Tan, E.M. (1981). A nuclear antigen associated with cell proliferation and blast transformation. The Journal of experimental medicine 154, 1899‐1909.

Takeuchi, J., Chen, H., and Coffino, P. (2007). Proteasome substrate degradation requires association plus extended peptide. The EMBO journal 26, 123‐131.

Tanaka, T., Kawashima, H., Yeh, E.T., and Kamitani, T. (2003). Regulation of the NEDD8 conjugation system by a splicing variant, NUB1L. The Journal of biological chemistry 278, 32905‐ 32913.

Tanida, I., Ueno, T., and Kominami, E. (2004). LC3 conjugation system in mammalian autophagy. The international journal of biochemistry & cell biology 36, 2503‐2518.

- 120 - References

Tanji, K., Tanaka, T., and Kamitani, T. (2005). Interaction of NUB1 with the proteasome subunit S5a. Biochemical and biophysical research communications 337, 116‐120.

Tateishi, K., Omata, M., Tanaka, K., and Chiba, T. (2001). The NEDD8 system is essential for cell cycle progression and morphogenetic pathway in mice. The Journal of cell biology 155, 571‐579.

Tatham, M.H., Jaffray, E., Vaughan, O.A., Desterro, J.M., Botting, C.H., Naismith, J.H., and Hay, R.T. (2001). Polymeric chains of SUMO‐2 and SUMO‐3 are conjugated to protein substrates by SAE1/SAE2 and Ubc9. The Journal of biological chemistry 276, 35368‐35374.

Terai, K., Abbas, T., Jazaeri, A.A., and Dutta, A. (2010). CRL4(Cdt2) E3 ubiquitin ligase monoubiquitinates PCNA to promote translesion DNA synthesis. Molecular cell 37, 143‐149.

Thrower, J.S., Hoffman, L., Rechsteiner, M., and Pickart, C.M. (2000). Recognition of the polyubiquitin proteolytic signal. The EMBO journal 19, 94‐102.

Tissier, A., Janel‐Bintz, R., Coulon, S., Klaile, E., Kannouche, P., Fuchs, R.P., and Cordonnier, A.M. (2010). Crosstalk between replicative and translesional DNA polymerases: PDIP38 interacts directly with Poleta. DNA repair 9, 922‐928.

Toschi, L., and Bravo, R. (1988). Changes in cyclin/proliferating cell nuclear antigen distribution during DNA repair synthesis. The Journal of cell biology 107, 1623‐1628.

VanDemark, A.P., Hofmann, R.M., Tsui, C., Pickart, C.M., and Wolberger, C. (2001). Molecular insights into polyubiquitin chain assembly: crystal structure of the Mms2/Ubc13 heterodimer. Cell 105, 711‐720.

Vertegaal, A.C., Ogg, S.C., Jaffray, E., Rodriguez, M.S., Hay, R.T., Andersen, J.S., Mann, M., and Lamond, A.I. (2004). A proteomic study of SUMO‐2 target proteins. The Journal of biological chemistry 279, 33791‐33798.

Vijay‐Kumar, S., Bugg, C.E., and Cook, W.J. (1987). Structure of ubiquitin refined at 1.8 A resolution. Journal of molecular biology 194, 531‐544.

Wada, H., Kito, K., Caskey, L.S., Yeh, E.T., and Kamitani, T. (1998). Cleavage of the C‐terminus of NEDD8 by UCH‐L3. Biochemical and biophysical research communications 251, 688‐692.

Wada, H., Yeh, E.T., and Kamitani, T. (1999). Identification of NEDD8‐conjugation site in human cullin‐2. Biochemical and biophysical research communications 257, 100‐105.

Wagner, S.A., Beli, P., Weinert, B.T., Nielsen, M.L., Cox, J., Mann, M., and Choudhary, C. (2011). A proteome‐wide, quantitative survey of in vivo ubiquitylation sites reveals widespread regulatory roles. Molecular & cellular proteomics : MCP 10, M111 013284.

Walden, H., and Martinez‐Torres, R.J. (2012). Regulation of Parkin E3 ubiquitin ligase activity. Cellular and molecular life sciences : CMLS.

Walden, H., Podgorski, M.S., Huang, D.T., Miller, D.W., Howard, R.J., Minor, D.L., Jr., Holton, J.M., and Schulman, B.A. (2003a). The structure of the APPBP1‐UBA3‐NEDD8‐ATP complex reveals the basis for selective ubiquitin‐like protein activation by an E1. Molecular cell 12, 1427‐1437.

Walden, H., Podgorski, M.S., and Schulman, B.A. (2003b). Insights into the ubiquitin transfer cascade from the structure of the activating enzyme for NEDD8. Nature 422, 330‐334.

Wang, H., Wang, L., Erdjument‐Bromage, H., Vidal, M., Tempst, P., Jones, R.S., and Zhang, Y. (2004). Role of histone H2A ubiquitination in Polycomb silencing. Nature 431, 873‐878.

- 121 - References

Wang, J., Cai, S., and Chen, Y. (2010). Mechanism of E1‐E2 interaction for the inhibition of Ubl adenylation. The Journal of biological chemistry 285, 33457‐33462.

Wang, J., Hu, W., Cai, S., Lee, B., Song, J., and Chen, Y. (2007a). The intrinsic affinity between E2 and the Cys domain of E1 in ubiquitin‐like modifications. Molecular cell 27, 228‐237.

Wang, S.C., Nakajima, Y., Yu, Y.L., Xia, W., Chen, C.T., Yang, C.C., McIntush, E.W., Li, L.Y., Hawke, D.H., Kobayashi, R., et al. (2006). Tyrosine phosphorylation controls PCNA function through protein stability. Nature cell biology 8, 1359‐1368.

Wang, X., Herr, R.A., Chua, W.J., Lybarger, L., Wiertz, E.J., and Hansen, T.H. (2007b). Ubiquitination of serine, threonine, or lysine residues on the cytoplasmic tail can induce ERAD of MHC‐I by viral E3 ligase mK3. The Journal of cell biology 177, 613‐624.

Watanabe, K., Tateishi, S., Kawasuji, M., Tsurimoto, T., Inoue, H., and Yamaizumi, M. (2004). Rad18 guides poleta to replication stalling sites through physical interaction and PCNA monoubiquitination. The EMBO journal 23, 3886‐3896.

Watson, I.R., Blanch, A., Lin, D.C., Ohh, M., and Irwin, M.S. (2006). Mdm2‐mediated NEDD8 modification of TAp73 regulates its transactivation function. The Journal of biological chemistry 281, 34096‐34103.

Watson, I.R., Irwin, M.S., and Ohh, M. (2011). NEDD8 pathways in cancer, Sine Quibus Non. Cancer cell 19, 168‐176.

Wee, K.E., Lai, Z., Auger, K.R., Ma, J., Horiuchi, K.Y., Dowling, R.L., Dougherty, C.S., Corman, J.I., Wynn, R., and Copeland, R.A. (2000). Steady‐state kinetic analysis of human ubiquitin‐activating enzyme (E1) using a fluorescently labeled ubiquitin substrate. Journal of protein chemistry 19, 489‐498.

Welchman, R.L., Gordon, C., and Mayer, R.J. (2005). Ubiquitin and ubiquitin‐like proteins as multifunctional signals. Nature reviews Molecular cell biology 6, 599‐609.

Wenzel, D.M., Stoll, K.E., and Klevit, R.E. (2011). E2s: structurally economical and functionally replete. The Biochemical journal 433, 31‐42.

Wessel, D., and Flugge, U.I. (1984). A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids. Analytical biochemistry 138, 141‐143.

Whitby, F.G., Xia, G., Pickart, C.M., and Hill, C.P. (1998). Crystal structure of the human ubiquitin‐ like protein NEDD8 and interactions with ubiquitin pathway enzymes. The Journal of biological chemistry 273, 34983‐34991.

Wiborg, O., Pedersen, M.S., Wind, A., Berglund, L.E., Marcker, K.A., and Vuust, J. (1985). The human ubiquitin multigene family: some genes contain multiple directly repeated ubiquitin coding sequences. The EMBO journal 4, 755‐759.

Wilkinson, K.D. (1997). Regulation of ubiquitin‐dependent processes by deubiquitinating enzymes. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 11, 1245‐1256.

Wimuttisuk, W., and Singer, J.D. (2007). The Cullin3 ubiquitin ligase functions as a Nedd8‐bound heterodimer. Molecular biology of the cell 18, 899‐909.

Winn, P.J., Religa, T.L., Battey, J.N., Banerjee, A., and Wade, R.C. (2004). Determinants of functionality in the ubiquitin conjugating enzyme family. Structure 12, 1563‐1574.

- 122 - References

Wu, K., Yamoah, K., Dolios, G., Gan‐Erdene, T., Tan, P., Chen, A., Lee, C.G., Wei, N., Wilkinson, K.D., Wang, R., et al. (2003a). DEN1 is a dual function protease capable of processing the C terminus of Nedd8 and deconjugating hyper‐neddylated CUL1. The Journal of biological chemistry 278, 28882‐28891.

Wu, P.Y., Hanlon, M., Eddins, M., Tsui, C., Rogers, R.S., Jensen, J.P., Matunis, M.J., Weissman, A.M., Wolberger, C., and Pickart, C.M. (2003b). A conserved catalytic residue in the ubiquitin‐ conjugating enzyme family. The EMBO journal 22, 5241‐5250.

Xirodimas, D.P., Saville, M.K., Bourdon, J.C., Hay, R.T., and Lane, D.P. (2004). Mdm2‐mediated NEDD8 conjugation of p53 inhibits its transcriptional activity. Cell 118, 83‐97.

Xirodimas, D.P., Sundqvist, A., Nakamura, A., Shen, L., Botting, C., and Hay, R.T. (2008). Ribosomal proteins are targets for the NEDD8 pathway. EMBO reports 9, 280‐286.

Xu, P., Duong, D.M., Seyfried, N.T., Cheng, D., Xie, Y., Robert, J., Rush, J., Hochstrasser, M., Finley, D., and Peng, J. (2009). Quantitative proteomics reveals the function of unconventional ubiquitin chains in proteasomal degradation. Cell 137, 133‐145.

Yamanaka, A., Hatakeyama, S., Kominami, K., Kitagawa, M., Matsumoto, M., and Nakayama, K. (2000). Cell cycle‐dependent expression of mammalian E2‐C regulated by the anaphase‐ promoting complex/cyclosome. Molecular biology of the cell 11, 2821‐2831.

Yang, X., Zhou, J., Sun, L., Wei, Z., Gao, J., Gong, W., Xu, R.M., Rao, Z., and Liu, Y. (2007). Structural basis for the function of DCN‐1 in protein Neddylation. The Journal of biological chemistry 282, 24490‐24494.

Ye, Y., and Rape, M. (2009). Building ubiquitin chains: E2 enzymes at work. Nature reviews Molecular cell biology 10, 755‐764.

Yin, Q., Lin, S.C., Lamothe, B., Lu, M., Lo, Y.C., Hura, G., Zheng, L., Rich, R.L., Campos, A.D., Myszka, D.G., et al. (2009). E2 interaction and dimerization in the crystal structure of TRAF6. Nature structural & molecular biology 16, 658‐666.

Yu, F., White, S.B., Zhao, Q., and Lee, F.S. (2001). HIF‐1alpha binding to VHL is regulated by stimulus‐sensitive proline hydroxylation. Proceedings of the National Academy of Sciences of the United States of America 98, 9630‐9635.

Zachariae, W., Shevchenko, A., Andrews, P.D., Ciosk, R., Galova, M., Stark, M.J., Mann, M., and Nasmyth, K. (1998). Mass spectrometric analysis of the anaphase‐promoting complex from yeast: identification of a subunit related to cullins. Science 279, 1216‐1219.

Zheng, J., Yang, X., Harrell, J.M., Ryzhikov, S., Shim, E.H., Lykke‐Andersen, K., Wei, N., Sun, H., Kobayashi, R., and Zhang, H. (2002). CAND1 binds to unneddylated CUL1 and regulates the formation of SCF ubiquitin E3 ligase complex. Molecular cell 10, 1519‐1526.

Zou, W., Papov, V., Malakhova, O., Kim, K.I., Dao, C., Li, J., and Zhang, D.E. (2005). ISG15 modification of ubiquitin E2 Ubc13 disrupts its ability to form thioester bond with ubiquitin. Biochemical and biophysical research communications 336, 61‐68.

Zuo, W., Huang, F., Chiang, Y.J., Li, M., Du, J., Ding, Y., Zhang, T., Lee, H.W., Jeong, L.S., Chen, Y., et al. (2013). c‐Cbl‐Mediated Neddylation Antagonizes Ubiquitination and Degradation of the TGF‐ beta Type II Receptor. Molecular cell.

- 123 -

Eidesstattliche Erklärung

Ich erkläre hiermit, dass ich die vorliegende Arbeit ohne unzulässige Hilfe Dritter und ohne Benutzung anderer als der angegebenen Hilfsmittel angefertigt habe. Die aus anderen Quellen direkt oder indirekt übernommenen Daten und Konzepte sind unter Angabe der Quelle gekennzeichnet. Weitere Personen, insbesondere Promotionsberater, waren an der inhaltlich materiellen Erstellung dieser Arbeit nicht beteiligt. Die Arbeit wurde bisher weder im In‐ noch im Ausland in gleicher oder ähnlicher Form einer anderen Prüfungsbehörde vorgelegt.

Dana Pagliarini

Konstanz, März 2013

- 124 -

Danksagung

Zunächst möchte ich mich ganz herzlich bei Professor Dr. Martin Scheffner bedanken. Danke, dass ich meine Doktorarbeit in Deinem Labor anfertigen durfte. Danke für Deine uneingeschränkte Hilfsbereitschaft, für Diskussionen und Kritik. Herrn Professor Dr. Thomas U. Mayer danke ich für die Begutachtung dieser Arbeit. Darüber hinaus möchte ich mich bei der Konstanz Research School Chemical Biology für die interessanten Seminare und Kurse, den konstruktiven wissenschaftlichen Austausch und die schönen Retreats bedanken. Vielen Dank an Dr. Andreas Marquardt (Proteomics Facility, Universität Konstanz) und Anna Sladewska für die Messung der vielen Proben und die Hilfe bei der Auswertung. Mein weiterer Dank gilt Professor Dr. Tancred Frickey für seine Hilfe bei der Tertiärstrukturvorhersage. Außerdem möchte ich mich bei dem Bioimaging Center der Universität Konstanz für die Möglichkeit der Nutzung des Konfokalmikroskops bedanken. Danke auch an Dr. Silvia Eger und Daniel Rösner für die Hilfe bei der Click Reaktion und vielen Dank an Stephan Hacker und Professor Dr. Andreas Marx für die nette und erfolgreiche Kollaboration. Ein ganz großes Dankeschön außerdem an alle ehemaligen und derzeitigen Mitglieder der AG Scheffner ‐ vor allem an Alejandro, Simone, Hans‐Peter, Nadine, Gregor, Meike, Thomas, Nicole, Toto, Franzi, Myriam, Daniel, Stefan und Elli ‐ für die sehr nette Arbeitsatmosphäre, für Diskussionen, Anregungen und die großartige Hilfsbereitschaft. Danke an alle, die mir mit der Korrektur der Doktorarbeit geholfen haben, vor allem an Elli, Alejandro und Daniel. Danke auch an unsere lustige Pokerrunde, die mir sicherlich sehr fehlen wird. Besonders bedanken möchte ich mich bei Nicole Richter‐Müller für ihre tolle Unterstützung bei vielen Experimenten und bei Elisabeth Stürner für ihre Hilfe in allen Lebenslagen. Danke für die vielen schönen gemeinsamen Mittagspausen und danke für Eure Freundschaft! Special thanks to Dr. Alejandro Rojas Fernández for your optimism, your never ending ideas, your help with corrections of my thesis and for being a good friend. You were the best supervisor in the world! Das allergrößte Dankeschön gilt abschließend meiner Familie, besonders aber meinen Eltern und meinem Freund Mike. Vielen Dank für Eure uneingeschränkte Unterstützung während meiner gesamten Ausbildung und danke für Eure Liebe!

- 125 -