The UBE2L3 Ubiquitin Conjugating Enzyme: Interplay with Inflammasome Signalling and Bacterial Ubiquitin Ligases

The UBE2L3 ubiquitin conjugating enzyme: interplay with inflammasome signalling and bacterial ubiquitin ligases

Matthew James George Eldridge

2018

Imperial College London Department of Medicine

Submitted to Imperial College London for the degree of Doctor of Philosophy

Abstract Inflammasome-controlled immune responses such as IL-1β release and pyroptosis play key roles in antimicrobial immunity and are heavily implicated in multiple hereditary autoimmune diseases. Despite extensive knowledge of the mechanisms regulating inflammasome activation, many downstream responses remain poorly understood or uncharacterised. The cysteine protease caspase-1 is the executor of inflammasome responses, therefore identifying and characterising its substrates is vital for better understanding of inflammasome-mediated effector mechanisms. Using unbiased proteomics, the Shenoy grouped identified the ubiquitin conjugating enzyme UBE2L3 as a target of caspase-1. In this work, I have confirmed

UBE2L3 as an indirect target of caspase-1 and characterised its role in inflammasomes-mediated immune responses. I show that UBE2L3 functions in the negative regulation of cellular pro-IL-1 via the ubiquitin- proteasome system.

Following inflammatory stimuli, UBE2L3 assists in the ubiquitylation and degradation of newly produced pro-IL-1. However, in response to caspase-1 activation, UBE2L3 is itself targeted for degradation by the proteasome in a caspase-1-dependent manner, thereby liberating an additional pool of IL-1 which may be processed and released.

UBE2L3 therefore acts a molecular rheostat, conferring caspase-1 an additional level of control over this potent cytokine, ensuring that it is efficiently secreted only in appropriate circumstances. These findings on UBE2L3 have implications for IL-1- driven pathology in hereditary fever syndromes, and autoinflammatory conditions associated with UBE2L3 polymorphisms.

Acknowledgments First and foremost, I would like to extend my most sincere thanks to my supervisor Avinash. Having the opportunity to join a brand-new lab has taught me so many things at and away from the bench and being able to learn from you first hand has been invaluable. Not every student gets the opportunity to work so closely with their supervisor and I will be forever grateful for your mentorship.

I would like to thank my lab mates Julia and Pippa, you two have been the best lab mates ever. Thank you for always being around, not only have you been great colleagues, you have been fantastic and supportive friends. Without the two of you, the 4th floor would have been a very different place, it was a pleasure working with you both. Julia, thank you for all your contributions to the paper, it would have taken three times as long without you.

Thank you to the other inhabitants of the 4th floor, old and new Dan, Amy,

Sophie, and Nancy, you couldn’t ask for better people to share a lab space with. Thank you to my supervisees who helped me with my bacteria side projects Anja, Youssef, and Razia. I would also like to thank all the members of the CMBI who I have crossed paths with over the years in particular Vani, Miles, Robbie, Sina, Agnes, Angie thank you for making it such a great place to work.

Thank you Sanika, who entered the fray a little later but made the last year and half of my PhD that little bit easier.

Finally, thank you to my parents, my sister and the rest of my family whose continued support has made my academic studies possible. Mum and Dad thank you for giving me the opportunity to pursue what I love to do.

Declaration of originality I hereby declare that the work presented in this thesis is my own, any work carried out by anyone other than myself be it published and unpublished has been acknowledged and clearly credited in the relevant figure legends. Any resources or ideas that are from the work of others has been reference and a full bibliography can be found at the end of this thesis.

Creative Commons Attribution Non-Commercial No Derivatives licence. Researchers are free to copy, distribute or transmit the thesis on the condition that they attribute it, that they do not use it for commercial purposes and that they do not alter, transform or build upon it. For any reuse or redistribution, researchers must make clear to others the licence terms of this work.

Table of contents Abstract ...... 2 Acknowledgments...... 3 Declaration of originality ...... 4 Copyright declaration ...... 4 Tables and figures ...... 10 Abbreviations ...... 15 Chapter 1 – Introduction ...... 18 1.1 Introduction to pattern recognition ...... 19 1.2 Toll-like receptor signalling ...... 20 1.2.1 MyD88 dependent activation of NF-κB ...... 22 1.2.2 TRIF dependent signalling ...... 24 1.3 Inflammasomes ...... 25 1.3.1 Principles of inflammasome activation...... 25 1.3.3 NLRP3 inflammasome ...... 28 1.3.4 NLRC4 inflammasome ...... 31 1.3.5 AIM2 inflammasome ...... 33 1.3.6 Non-canonical inflammasome activation ...... 34 1.3.7 PYRIN inflammasome ...... 35 1.3.8 NLRP1 inflammasome ...... 37 1.3.9 Other inflammasomes ...... 38 1.4 Caspase-1 Substrates and Targets ...... 39 1.4.1 pro-IL-1β ...... 40 1.4.2 pro-IL-18 ...... 42 1.4.3 IL-1α ...... 43 1.4.4 Gasdermin D ...... 44 1.5 Inflammasome signalling during bacterial infection ...... 46 1.5.1 Pathogen sensing by inflammasomes ...... 46 1.5.2 Bacterial circumvention of inflammasome sensing and mechanisms ...... 48 1.5.3 Host-protective properties of inflammasomes during bacterial infection ...... 49 1.6 Role of inflammasomes in inheritable illnesses ...... 51 1.7 Novel downstream targets of inflammasomes ...... 53 1.8 Enzymes of the ubiquitin cascade ...... 55

1.9 UBE2L3: an atypical E2 conjugating enzyme ...... 58 1.10 Aims and Objectives ...... 63 Chapter 2 – Materials and Methods ...... 64 2.1 Eukaryotic and bacterial cell growth conditions...... 65 2.1.1 Eukaryotic cell culture ...... 65 2.1.2 Bacterial growth conditions for infection ...... 66 2.2 Bacterial strains, plasmids, and primers ...... 69 2.2.1 Bacterial strains ...... 69 2.2.2 Bacterial and eukaryotic expression plasmids ...... 70 2.2.4 Primers used during this study ...... 74 2.3 Molecular Biology techniques ...... 76 2.3.1 Preparation of chemically component bacteria and heat shock transformation. . 76 2.3.2 Preparation of electrocompetent cells and transformation by electroporation ... 77 2.3.3 Cloning and site-directed mutagenesis ...... 77 2.3.4 Retroviral packaging and transduction ...... 79 2.3.5 Isolation of bacterial genomic DNA ...... 79 2.3.6 Lambda (λ) red recombinase gene disruption of nleL ...... 80 2.4 Cell biology treatments and analyses ...... 81 2.4.1 Inflammasome activation assays ...... 81 2.4.2 Cell priming assays ...... 83 2.4.3 Sample preparation for SDS-polyacrylamide gel electrophoresis and protein visualisation ...... 83 2.4.4 Cytokine ELISA, cell death, and membrane permeability assays ...... 84 2.4.5 Quantitative reverse transcription PCR ...... 85 2.4.6 Immunofluorescence analysis ...... 86 2.4.7 Pro-IL-1β immunoprecipitation (IP) ...... 87 2.4.8 Growth-curve analysis of wildtype and ΔnleL EHEC ...... 87 2.5 Recombinant protein purification ...... 88 2.5.1 GST-tag based protein purification...... 88 2.5.2 Nickel (Ni) Affinity purification of His-tagged UBE1 protein ...... 89 2.5.3 Bradford Protein Assay...... 90 2.6 In vitro Ubiquitylation assays ...... 90 2.6.1 General in vitro assay ...... 90 2.6.2 Trans ubiquitylation assay ...... 91 6

2.6.3 Unanchored ubiquitin chain formation assay ...... 91 2.6.4 N-terminal ubiquitylation assay ...... 91 2.6.5 Recombinant protein production and caspase-1 cleavage assays ...... 92 2.7 Antibodies and Reagents ...... 93 2.7.1 List of antibodies ...... 93 2.7.2 List of treatment reagents and commercial recombinant protein ...... 94 Chapter 3 – Depletion of UBE2L3 is a common consequence of inflammasome activation 95 3.1 Introduction ...... 96 3.2 Results ...... 100 3.2.1 Depletion of UBE2L3 is triggered by NLRP3 inflammasome agonists ...... 100 3.2.2 UBE2L3 depletion depends on potassium efflux in response to NLRP3 inflammasome agonists ...... 103 3.2.3 Depletion of UBE2L3 temporally correlates with caspase-1 activation ...... 105 3.2.4 NLRP3 activation by sterile particulate agonists triggers UBE2L3 depletion in primary cells ...... 107 3.2.5 UBE2L3 depletion is dependent on core canonical inflammasome components 108 3.2.6 Inflammasome dependent UBE2L3 depletion occurs in human macrophage-like THP1 cells ...... 111 3.2.7 Activation of caspase-4 is required but not sufficient for UBE2L3 depletion induced by non-canonical activation of NLRP3 ...... 113 3.2.8 Activation of NLRP1 and PYRIN inflammasomes triggers UBE2L3 depletion ...... 115 3.2.9 Activation of AIM2 inflammasomes triggers UBE2L3 depletion ...... 118 3.2.10 NLRC4 activation mediates depletion of UBE2L3 and does not require the ASC adaptor ...... 119 3.2.11 Listeria monocytogenes infection activates caspase-1 and triggers UBE2L3 depletion ...... 122 3.2.12 UBE2L3 depletion in human and mouse macrophages is dependent on caspase-1 activity ...... 125 3.3 Discussion ...... 127 3.4 Conclusions ...... 131 Chapter 4 – Elucidating the mechanism of caspase-1-mediated depletion of UBE2L3 ...... 132 4.1 Introduction ...... 133 4.2 Results ...... 137 4.2.1 UBE2L3 is not cleaved by caspase-1 in vitro or in cells...... 137

4.2.2 Depletion of UBE2L3 is not mediated through caspase-1-mediated autocrine or paracrine signalling ...... 144 4.2.3 UBE2L3 is not lost via GSDMD pore formation or lysis ...... 146 4.2.4 Caspase-1 mediates cell-intrinsic, proteasome-dependent depletion of UBE2L3...... 150 4.2.5 Lysine residues are required for UBE2L3 protein stability but are dispensable for caspase-1-mediated degradation...... 154 4.3 Discussion ...... 159 4.4 Conclusions ...... 167 Chapter 5 – UBE2L3 regulates IL-1β production by inflammasomes by regulating pro-IL-1β stability ...... 168 5.1 Introduction ...... 169 5.1.2 Functions of ubiquitin chain linkages ...... 169 5.1.3 Regulation of ubiquitylation via UBE2L3 ...... 171 5.1.4 Role of ubiquitin in inflammasome signalling ...... 172 5.1.5 Chapter Aims ...... 175 5.2 Results ...... 176 5.2.1 IL-1β secretion is reduced by sustained UBE2L3 expression ...... 176 5.2.2 Reduced IL-1β release is not due to priming defects ...... 180 5.2.3 UBE2L3 controls pro-IL-1β protein levels ...... 182 5.2.4 UBE2L3 functions in the proteasome-dependent turnover of pro-IL-1β ...... 188 5.2.5 UBE2L3 controls pro-IL-1β levels during priming by bacteria ...... 194 5.2.6 UBE2L3 increases the ubiquitylation of pro-IL-1β ...... 197 5.2.7 UBE2L3 silencing enhances pro-IL-1β protein levels ...... 200 5.2.8 UBE2L3 depletion enhances IL-1β secretion by inflammasomes ...... 204 5.2.9 UBE2L3 modulates IL-1β release in response to bacterial infection or priming .. 207 5.3 Discussion ...... 209 5.4 Summary ...... 218 Chapter 6 – UBE2L3 is polyubiquitylated by the enterohemorrhagic E. coli effector NleL 219 6.1 Introduction ...... 220 6.2 Results ...... 226 6.2.1 UBE2L3 is rapidly ubiquitylated by NleL ...... 226 6.2.2 UBE2L3 is trans-ubiquitylated by NleL ...... 232 6.2.3 UBE2L3 is ubiquitylated on five residues ...... 235

6.2.4 UBE2L3 is predominantly modified on one lysine with multiple linkage types ... 237 6.2.5 NleL modifies lysine 82 with K6 and K48 polyubiquitin chains ...... 241 6.2.6 NleL modifies multiple lysine residues with K29-linked ubiquitin ...... 245 6.2.7 UBE2L3 is not modified with linear ubiquitin chains ...... 250 6.2.8 The N-terminus of UBE2L3 is modified by NleL ...... 253 6.2.9 Lysine availability affects free chain formation ...... 257 6.3 Discussion ...... 260 6.4 Summary ...... 266 Chapter 7 – Final summary and future perspectives ...... 267 7.1 Introduction ...... 268 7.2 Final summary ...... 268 7.3 Future perspectives ...... 270 7.3.1 Elucidate how UBE2L3 is targeted to the proteasome ...... 270 7.3.2 Further characterisation of pro-IL-1β ubiquitylation ...... 272 7.3.3 Physiological relevance and wider in vivo implications ...... 274 References ...... 267 Appendix ...... 336

Tables and figures Chapter 1

Table 1.1: TLR receptor ligands and signalling pathways. Figure 1.1: Pro-inflammatory signalling pathways downstream of MYD88 Figure 1.2: Schematic of TRIF-dependent TLR signalling Figure 1.3: Schematic of inflammasome activation by signal 1 and signal 2 Figure 1.4: Domain organisation of key inflammasome receptor proteins Figure 1.5: Domain organisation and cleavage sites of Caspase-1 Figure 1.6 Known and novel functions of inflammasomes Figure 1.7: The ubiquitin transfer cascade Figure 1.8: Structure of UBE2L3 Figure 1.9: Phylogenetic and sequence analysis of UBE2L3 protein sequence across species

Chapter 2

Table 2.1. Bacterial strains Table 2.2. Plasmids for the expression of recombinant proteins in E. coli Table 2.3. Plasmids for expression of proteins and genetic constructs in eukaryotic cells Table 2.4. Sequencing primers Table 2.6 Primers used for cloning Table 2.7 Primers used in the generation of ΔnleL EHEC Table 2.8. Primers used in qPCR analyses Table 2.9. Inflammasome agonists used in these studies Table 2.10. Additional treatments used alongside inflammasome activation Table 2.11: List of antibodies used during this study Table 2.12: List of reagents used in cell treatments and commercial recombinant proteins

Chapter 3

Figure 3.1: Identification of caspase-1 substrates and targets by comparative PROTOMAP proteomics. Figure 3.2: NLRP3 activation triggers UBE2L3 depletion independently of cell lysis.

Figure 3.3: UBE2L3 does not deplete following separate addition of signal 1 and signal 2 agonists Figure 3.4: Depletion of UBE2L3 is dependent on potassium efflux Figure 3.5: UBE2L3 depletion temporally correlates with caspase-1 activation Figure 3.6: Alum induced UBE2L3 depletion in primary immune cells Figure 3.7: UBE2L3 depletion is dependent on core components of NLRP3 inflammasome Figure 3.8: NLRP3 mediated loss of UBE2L3 is strictly caspase-1 dependent Figure 3.9: K+ efflux and NLRP3 dependent UBE2L3 depletion in human macrophage-like THP-1 cells 3.2.7 Activation of caspase-4 is required but not sufficient for UBE2L3 depletion induced by non-canonical activation of NLRP3 Figure 3.10: Caspase-4 is required but not sufficient for UBE2L3 depletion Figure 3.11: Activation of NLRP1b inflammasome triggers UBE2L3 depletion Figure 3.12: Pyrin inflammasome activation triggers UBE2L3 depletion Figure 3.13: AIM2 inflammasome mediates UBE2L3 depletion Figure 3.14: SPI-1 and NLRC4 dependent depletion of UBE2L3 during Salmonella infection. Figure 3.15: Listeria induced UBE2L3 depletion by ASC dependent inflammasomes. Figure 3.16: UBE2L3 depletion by Listeria infection is dependent on caspase-1 activity and listeriolysin. Figure 3.17: Loss of UBE2L3 is dependent on caspase-1 activity in human and mouse cells.

Chapter 4

Figure 4.1: Recombinant UBE2L3 is not cleaved by caspase-1 in vitro. Figure 4.2: Caspase-1 cleaves HA-tag from UBE2L3 in cells. Figure 4.3: UBE2L3 is not cleaved by caspase-1 in cells Figure 4.4: UBE2L3 does not localise to the inflammasome speck during inflammasome activation. Figure 4.5: UBE2L3 depletes in Il1r1-/- and Il18r1-/- BMDMs, and its depletion is independent of inflammasome-dependent secreted factors Figure 4.6: Depletion of UBE2L3 is independent of GSDMD-mediated pore formation and lysis. Figure 4.7: UBE2L3 is common to multiple inflammasomes in GSDMD deficient cells

Figure 4.8: UBE2L3 loss in human cells is independent of GSDMD Figure 4.9: UBE2L3 is stable under non-inflammasome activating conditions. Figure 4.10: UBE2L3 depletes in Atg7 silenced cells and is independent of autophagy Figure 4.11: Caspase-1-dependent UBE2L3 depletion requires proteasomal activity. Figure 4.12: Lysineless UBE2L3 is degraded by the proteasome under resting conditions. Figure 4.13: Lysine residues, catalytic activity and E3-binding residues impact UBE2L3 stability in macrophages Figure 4.14: UBE2L3 depleted independently of lysine ubiquitylation and catalytic activity. Table 4.1: Summary of UBE2L3 variant stability under resting conditions Figure 4.15: Potential ubiquitin dependent an independent mechanism which may cause caspase-1 dependent, proteasomal targeting of UBE2L3.

Chapter 5

Table 5.1: Regulation of cellular pathways by ubiquitin chain linkage. Figure 5.1: N-terminal tagging of UBE2L3 does not affect ubiquitin charging. Figure 5.2: Sustained UBE2L3 expression reduces IL-1β secretion independently of cell death Figure 5.3: Sustained UBE2L3 expression reduces IL-1β secretion Figure 5.4: UBE2L3 overexpression does not alter signal 1 priming Figure 5.5: Sustained UBE2L3 expression increases pro-IL-1β turnover Figure 5.6: Reduced IL-1β stability is independent of transcription Figure 5.7: UBE2L3 controls IL-1β and IL-1α protein levels Figure 5.8: Sustained UBE2L3 expression increases proteasomal turnover of pro-IL-1β Figure 5.9: Proteasome inhibition by multiple inhibitors block IL-1β depletion. Figure 5.10: IL-1β p28 requires caspase-1 activity and is not required for proteasomal turnover. Figure 5.11: Caspase-1-dependent cytokines are differentially regulated by the proteasome Figure 5.12 UBE2L3 controls pro-IL-1β levels during priming with bacterial mutants in murine cells. Figure 5.13 UBE2L3 controls pro-IL-1β levels during priming and infection with bacterial mutants in human cells Figure 5.14: UBE2L3 promotes K48-linked ubiquitylation of pro-IL-1β Figure 5.15: Silencing UBE2L3 expression increases pro-IL-1β production

Figure 5.16 UBE2L3 controls pro-IL-1β levels during priming and infection with bacterial mutants that that cannot activate inflammasomes Figure 5.17 UBE2L3 depletion increases has cells specific roles in priming Figure 5.18 UBE2L3 depletion increases mature IL-1β production by inflammasomes. Figure 5.19 UBE2L3 depletion increases mature IL-1β production by AIM2 and non-canonical inflammasomes. Figure 5.20 UBE2L3 status controls mature IL-1β production Figure 5.21 UBE2L3 silencing increases spontaneous processing of pro-IL-1β Figure 5.22 Working model pro-IL-1β regulation by UBE2L3

Chapter 6

Figure 6.1 Crystal structure and domain organisation NleL Figure 6.2 NleL makes UBE2L3 exists as multiple species of increasing molecular weight Figure 6.3: UBE2L3 is rapidly ubiquitylated by NleL in vitro. Figure 6.4: UBE2L3 is not ubiquitylated by SopA in vitro. Figure 6.5: NleL mediates trans ubiquitylation of UBE2L3 lysine residues in vitro. Figure 6.6: UBE2L3 can be ubiquitylated on 5 residues Figure 6.7: NleL modifies lysine 82 of UBE2L3 with polyubiquitin Figure 6.8: Lysine 82 does not contribute to ligase activity Figure 6.9: Full length NleL ubiquitylates UBE2L3 at lysine 82 Figure 6.10: NleL modifies UBE2L3 with K48 ubiquitin chains Figure 6.11: NleL modifies UBE2L3 with K6 ubiquitin chains Figure 6.12: NleL modifies UBE2L3 on multiple with K29 ubiquitin Figure 6.13: Full-length NleL is better able to carry out K29-ubiquitylation of UBE2L3 as compared with NleL170-782 Figure 6.14: NleL modifies lysine 73 with K29 di- ubiquitin Figure 6.15: Lysine 82 is not modified with linear polyubiquitin Figure 6.16: NleL ubiquitylates UBE2L3 on a non-lysine residue Figure 6.17: UBE2L3 is not modified with thioester or oxyester linked ubiquitin. Figure 6.18: NleL modifies the N-terminus of UBE2L3 with ubiquitin which can be blocked by an N-terminal GST-tag.

Figure 6.19: Availability of lysine reactive UBE2L3 affects the formation of free Ub chains by NleL.

Appendix

Appendix Figure 1 – Amino acid sequence of recombinant UBE2L3 used for in vitro ubiquitylation assays Appendix Table 1 – Lysine residues present in UBE2L3 and mutant variants

Abbreviations Activator protein-1 AP-1 AIM2-like receptors ALRs Apoptosis-associated speck-like protein containing a ASC CARD Baculovirus inhibitor repeat BIR Cardiolipin CL Caspase activation and recruitment domain CARD CpG-DNA C-phosphate-G DNA C-type lectin receptors CLRs Damage-associated molecular patterns DAMPs DNA Deoxyribonucleic acid dsRNA Double-stranded RNA Glyceraldehyde-3-phosphate dehydrogenase GAPDH Homologous to the E6AP carboxyl terminus HECT IL-1 receptor type I IL-1R1 IL-1 receptor type II IL-1RII IL-18 receptor α chain IL-18Rα IL-1R accessory protein IL-1RAP IL-1R antagonists IL-1RA Interferon-regulatory factor 3 IRF 3 Interleukin-1 receptor 1 IL-1R1 Legionella-containing vacuole LCV Leucine-rich repeat LLR Liquid chromatography-tandem mass spectrometry LC-MS LPS Lipopolysaccharide Mitogen-activated protein kinase MAPK Mouse embryonic fibroblasts ( MEFs NLR and PYD domains-containing NLRP

NLR family apoptosis inhibitory proteins NAIP NLR family CARD domain-containing NLRC NOD-leucine rich repeat domain containing NLRs receptors Novel e3 ligases NELs Nucleotide-binding domain NBD Pathogen-associated molecular patterns PAMPs Phosphatidylinositol 4,5-bisphosphate PI(4,5)P Phosphatidylinositol 4-phosphate (PI(4)P) Potassium ion K+) Pro-interleukin-18 (pro-IL-18), Pro-interleukin-1β ( pro-IL-1β) PRRs Pattern recognition receptors Pyrin domain PYD Reactive oxygen species ROS Really interesting new gene RING RIGI-like receptors RLRs RING-in-between-RING RBR RNA Ribonucleic acid rRNA Ribosomal RNA Salmonella enterica Typhimurium STm Salmonella-containing vacuole SCV Sequence and ligation independent cloning SLIC Sequestosome1-like receptors SLRs Single-nucleotide polymorphisms SNPs ssRNA Single-stranded RNA Sterol regulatory element-binding proteins SREBPs TNF receptor-associated factor 6 TRAF6 Toll/interleukin-1 receptor TIR Toll-like receptors TLRs

Transforming growth factor-β-activated kinase-1 TAK1 TIR-domain-containing adapter-inducing interferon- TRIF β Tumor necrosis factor TNF Type three secretion system T3ss Ubiquitin-activating enzyme E1 Ubiquitin-binding domains UBDs Ubiquitin-conjugating enzyme E2 Ubiquitin-conjugating enzyme E2 L3 UBE2L3 Ubiquitin-protein ligase E3 Winged-helix domain WHD

Chapter 1 – Introduction

1.1 Introduction to pattern recognition The immune system of higher vertebrates possesses two arms: innate, and adaptive immunity. Both systems are comprised of bone marrow-derived leukocyte cells which serve to enact and regulate many immune responses towards pathogens.

Tightly regulated processes of differentiation and maturation control these cells development into distinct classes which carry out specific immune functions (Murphy et al., 2012; Parkin and Cohen, 2001). Phagocytic cells of the innate immune system are derived from myeloid progenitor cells which migrate into the blood and can mature into dendritic cells, granulocytes (neutrophil, eosinophil, and basophil cell types) and monocytes. When fully matured, these cells can interact directly with invading pathogens and engage in antimicrobial responses such as phagocytosis, cytokine release and antigen presentation (Mogensen, 2009; Murphy et al., 2012). Innate immunity plays a key role in generating rapid, first-line antimicrobial responses, consisting of a multitude of different cell-types, receptors, and an array of intricate, interconnected signalling pathways. It is because of this inherent breadth that the innate immune system can generate appropriate protective responses despite its genetic invariability.

Since Charles Janeway first proposed the pattern recognition theory in 1989 the innate immune system has been shown to, not only stimulate adaptive immunity but produce its own robust and protective immune responses (Janeway, 1989). This is achieved largely through the action of numerous, well conserved germline-encoded proteins called pattern recognition receptors (PRRs), which recognise and respond to pathogen-associated molecular patterns (PAMPs) and host-derived damage- associated molecular patterns (DAMPs) (Takeuchi and Akira, 2010).

The most extensively studied family of PRRs are the Toll-like receptors (TLRs), which trigger the transcriptional upregulation and production of numerous antimicrobial and pro-inflammatory molecules. Other PRRs include the C-type lectin receptors (CLRs) as well as intracellular receptors such sequestosome1-like receptors

(SLRs), RIGI-like receptors (RLRs), AIM2-like receptors (ALRs), and NOD-leucine rich repeat domain containing receptors (NLRs) (Newton and Dixit, 2012). Some notable members of the NLR and ALR family of receptors do not control transcription but are able to trigger robust pro-inflammatory responses by assembling large signalling complexes called inflammasomes. PRRs, therefore, play a vital role in triggering and modulating our body’s first-line defence against a wide array of invading pathogenic organisms (Mogensen, 2009; Newton and Dixit, 2012; Zhang and Liang, 2016).

1.2 Toll-like receptor signalling TLRs are a family of transmembrane PRRs which decorate various extracellular and intracellular membranes (Kawasaki and Kawai, 2014). These receptors consist of extracellular or luminal leucine-rich repeat (LRR) domains which bind PAMPs, and a conserved cytosolic toll/interleukin-1 receptor (TIR) domain which transmits these signals to the cytosol via different adapter proteins (Botos et al., 2011; Kawasaki and

Kawai, 2014). TLRs 1, 2, 4, 6 and 10 are found on the cytoplasmic membrane whereas

TLRs 2, 7, 8 and 9 are found at intracellular membrane-bound compartments such as endosomes (Kawasaki and Kawai, 2014; Newton and Dixit, 2012; O'Neill et al., 2013;

Takeuchi and Akira, 2010). TLRs can detect diverse microbe-derived molecules

(Table 1.1) including bacterial lipopolysaccharide (LPS), peptidoglycan, DNA and RNA

(Takeuchi et al., 1999). Activation of TLRs triggers pro-inflammatory responses which are in large part mediated by nuclear factor-κB (NF-κB) which is responsible for the transcription of many pro-inflammatory genes (Kawai and Akira, 2007). The active NF-

κB complex consists of the p50 and RelA subunits. Under resting conditions, these

components are held in the cytoplasm by the inhibitory protein IκBα which prevents

pro-inflammatory signalling (Huxford et al., 1998; Jacobs and Harrison, 1998).

Recognition of their respective ligand induces receptor dimerization and the

recruitment of cytosolic TIR domain-containing adaptors which propagate the signal

through the cell (Nyman et al., 2008). Recruitment of the adaptors MyD88 and Mal or

Receptor Localisation Pathway Ligand Reference TLR1 Membrane MyD88 Triacyl lipoproteins (Jin et al., 2007b) TLR2 Membrane MyD88 Lipoteichoic acid, (Schwandner et al., peptidoglycan 1999)

TLR3 Endosomal TRIF dsRNA (Alexopoulou et al., 2001)

TLR4 Membrane/ MyD88/ LPS (Poltorak et al., 1998) Endosomal TRIF

TLR5 Membrane MyD88 Flagellin (Hayashi et al., 2001)

TLR6 Membrane MyD88 Diacyl lipoproteins (Takeuchi et al., 2001)

TLR7 Endosomal MyD88 ssRNA (Heil et al., 2004)

TLR8 Endosomal MyD88 ssRNA (Heil et al., 2004)

TLR9 Endosomal MyD88 CpG-DNA (Bauer et al., 2001)

TLR10 Membrane MyD88 Lipoproteins (Guan et al., 2010)

TLR11 Endosomal MyD88 Flagellin/Profilin (Hatai et al., 2016) TLR12 Endosomal MyD88 Profilin (Raetz et al., 2013)

TLR13 Endosomal MyD88 23S rRNA (Li and Chen, 2012)

Table 1.1: TLR receptor ligands and signalling pathways. Table describes TLR localisation, adaptor signalling pathways and common stimulatory ligands/agonist. ds, double stranded; ss, single stranded. Table adapted from (Savva and Roger, 2013).

TRIF and TRAM trigger divergent signalling cascades to activate NF-KB.

Except for TLR3 which signals only via TIR-domain-containing adapter-inducing interferon-β (TRIF), all TLRs can signal through the adaptor MyD88. TLR4 can sequentially utilise both pathways subject to its localization; signalling via MyD88 when at the cell membrane and via TRIF during receptor endocytosis (Newton and Dixit,

2012).

1.2.1 MyD88 dependent activation of NF-κB

Following their oligomerisation, TLRs bind MyD88 via TIR-TIR domain interactions, either directly (TLR 5, 7, 8, 9, 10, 11, 12 and 13) or via the adaptor protein

MAL/TIRAP (TLR1, 2, 4 and 6) (Figure 1.1). MyD88 triggers the sequential recruitment and activation of Ser/Thr kinases IRAK-4, IRAK-2, and IRAK-1 (Lin et al., 2010). IRAK-

1 becomes autophosphorylated and interacts with the E3 ligase TNF receptor- associated factor 6 (TRAF6). Once recruited, TRAF6 utilises the E2 conjugating enzyme Ubc13 to modify itself and IRAK1 with K63-linked polyubiquitin chains (Tseng et al., 2010). K63-linked polyubiquitin forms a platform for the recruitment and activation of the transforming growth factor-β-activated kinase-1 (TAK1) and the IKK complex which is composed of the regulatory component NEMO and catalytic subunits

IKKα and IKKβ (Laplantine et al., 2009). TAK1 then activates two diverging pathways:

(1) the mitogen-activated protein kinase (MAPK) cascade which culminates in the activation of the activator protein-1 (AP-1) transcription factor (Clark, 2014); (2) TAK1 phosphorylates and activates IKKβ. IKKβ phosphorylates the NF-κB inhibitory protein

IκBα, which is then ubiquitylated degraded at the proteasome (Newton and Dixit,

2012). NF-κB is then translocated to the nucleus where it mediates the transcription of pro-inflammatory genes. More recently, the linear ubiquitin chain assembly complex

(LUBAC) has been shown to be heavily involved in many aspects of innate immune

22 receptor-signalling (Rittinger and Ikeda, 2017). LUBAC is made up of three subunits

HOIP, HOIL-1L and SHARPIN (Ikeda et al., 2011; Rittinger and Ikeda, 2017). Linear polyubiquitin has important scaffolding roles in TLR receptor signalling, particularly by supporting the recruitment of IKK via interactions with NEMO (Emmerich et al., 2013;

Ikeda et al., 2011). In agreement with this, macrophages from SHARPIN-deﬁcient mice have defective activation of NF-κB following treatment with LPS (Ikeda et al.,

Figure 1.1: Pro-inflammatory signalling pathways downstream of MYD88 Following ligand-induced receptor dimerization MyD88 is recruited to the cytoplasmic TIR domains of TLRs either directly or via an additional adaptor called MAL. MyD88 sequentially recruits IRAK4/2/1 which subsequently recruit the E3 ligase TRAF6. TRAF6 generates K63-polyubiquitin which forms a recruitment platform for TAK1 which subsequently activates both the MAPK cascade and the IKK complex leading the activation of AP-1 and NF-κB which begin transcription of pro-inflammatory genes.

2011) Likewise, SHARPIN-deﬁcient mice show reduced secretion of NF-kB regulated cytokines in response to TLR stimulation with PAMCSK4 (TLR2), LPS (TLR4), R848

(TLR7), CpG-B (TLR9), and infection with the bacteria Listeria monocytogenes (Lm)

(TLR2 and TLR5) (Zak et al., 2011).

1.2.2 TRIF dependent signalling

TLR3, in contrast to other TLRs, requires the TRIF adaptor to mediate its signalling (Figure 1.2). Similarly, TLR4 can also utilise TRIF once it has been internalised by endocytosis. TRIF-dependent signalling can activate late phase NF-kB

Figure 1.2: Schematic of TRIF-dependent TLR signalling Following the dimerization of receptors TLR3 or endosomally localised TLR4, TRIF is recruited. TRIF then forms a multiprotein complex consisting of TRAF3, TRAF6 and RIPK1 which can activate two diverging pathways. In one RIPK1 is modified with K63- linked polyubiquitin which in turn recruits TAK1 leading to NF-κB activation via IKK complex. The second pathway leads to the transcription of Type-1 interferons via the activation of IRF3 which requires TRAF3, TBK1 and IKKi.

24 activation and type 1 interferon signalling via the transcription factor IRF3. TLR3 can directly bind TRIF whereas TLR4 requires another protein adapter called TRAM. TRIF, in turn, forms a multiprotein signalling complex consisting of TRAF3, TRAF6, RIPK1 and the E3 ligase Pellino1. Pellino1 modifies RIPK1 with K63-linked polyubiquitin which is believed to facilitate the recruitment of TAK1, which in turn leads to the activation of NF-κB (Moynagh, 2014). TRIF-dependent signalling also activates type

1 interferon gene expression by activating interferon-regulatory factor 3 (IRF3) (Hacker and Karin, 2006). TLR signalling generates initial pro-inflammatory cytokine release, but also serves to ready cells for the activation of more robust inflammatory responses should they be required. This includes signalling by multi-protein inflammasomes, as described below, which trigger the maturation of the potent cytokines, pro-interleukin-

1β (pro-IL-1β) and pro-interleukin-18 (pro-IL-18), and initiate pro-inflammatory cell death (von Moltke et al., 2013).

1.3 Inflammasomes 1.3.1 Principles of inflammasome activation

The inflammasome was first described in 2002 by Jurg Tschopp and colleagues and is a large, multiprotein signalling hub which mediates robust posttranslational, pro- inflammatory responses, most notably, in immune cells (Martinon et al., 2002). These complexes serve as a molecular platform for the recruitment and activation of the pro- inflammatory cysteine protease caspase-1 (Martinon et al., 2002). It is now known that specific subfamilies of NLR proteins such as the NLR and PYD domains-containing

(NLRP), NLR family CARD domain-containing (NLRC) proteins, AIM2-like receptors

(ALRs) and PYRIN act as intracellular molecular sensors which drive the formation of inflammasome complexes (Broz and Dixit, 2016; Eldridge and Shenoy, 2015; Latz et al., 2013). The active caspase-1 enzyme serves as the executioner of inflammasome-

dependent signalling, and directly mediates key processes such as the proteolytic

activation of pro-IL-1β and pro-IL-18, and the initiation of pyroptosis by inducing the

formation of the Gasdermin D (GSDMD) pores in the cytoplasmic membrane (Fantuzzi

et al., 1998; Howard et al., 1991a; Howard et al., 1991b). In healthy individuals,

inflammasomes are under the control of multiple regulatory mechanisms as their

dysregulation or aberrant activation can cause debilitating and persistent

hyperinflammatory disorders (de Torre-Minguela et al., 2017).

To trigger robust activation of inflammasomes, two independent signals are

usually required, these are called signal 1 and signal 2 (Figure 1.3). Signal 1 refers to

the preparation or “priming” of cells for inflammasome signalling and can be triggered

Figure 1.3: Schematic of inflammasome activation by signal 1 and signal 2 (1) Signal 1 is often provided by external pro-inflammatory agonists, such as LPS. Signal 1 prepares cells for inflammasome activation by promoting transcription of inflammasome related genes and/or initiating post-transcriptional priming of NLR/ALR sensors. (2) Signal 2 provides the activating signal which triggers NLR/ALR oligomerisation and the formation of the inflammasome complex. A typical inflammasome consists of a sensor protein, the ASC adaptor and caspase-1.

26 by exogenous pro-inflammatory stimuli or host-derived cytokines such as tumor necrosis factor (TNF). The underlying mechanisms that mediate priming are specific to different inflammasome sensors and are addressed in more detail in subsequent sections. Broadly speaking, priming can occur in three ways: (1) transcriptional upregulation of the inflammasome sensors themselves, e.g. NLRP3 and AIM2; (2) post-translational priming of inflammasome sensors; and (3) transcriptional upregulation of key substrates of caspase-1, such as pro-IL-1β, which are not expressed in naïve phagocytes (von Moltke et al., 2013). The TLR4 agonist LPS is a commonly used priming agent which causes the transcriptional upregulation of inflammasome-related genes by transcription factors such as NF-κB and AP-1 (Latz et al., 2013). Signal 2 refers to the “activation” and assembly of the NLR/ALR sensors into the multiprotein inflammasome complex (Chu et al., 2001; Fernandes-Alnemri et

Figure 1.4: Domain organisation of key inflammasome receptor proteins The NLR family members NLRP1, NLRP3, NLRC4 all possess C-terminal leucine-rich repeat (LRR) domains, and a central a nucleotide-binding domain (NBD). At the N- terminus they either have a pyrin domain (PYD) or caspase activation and recruitment domain (CARD). NLRP1 also has a CARD at its C-terminus. NAIP has similar LRR and NBD domains and three baculovirus inhibitor (BIR) domains. AIM2 has a PYD and a HIN200 domain. PYRIN has a N-terminal PYD central B-box zinc finger domain (B-box) and coil-coiled domains and a C-terminal B30.2 domain.

27 al., 2009; He et al., 2016b). Following their activation NLR/ALR sensors oligomerise and recruit the apoptosis-associated speck-like protein containing a CARD (ASC) adaptor protein (Abderrazak et al., 2015). ASC then undergoes prion-like polymerisation into cross-linked filaments which are typical of inflammasomes and referred to as the ASC “speck” (Cai et al., 2014; Lu et al., 2014; Sborgi et al., 2015).

Because ASC contains both a caspase activation and recruitment domain (CARD) and a pyrin domain (PYD) it can function as an adapter for the recruitment of pro-caspase-

1 to the inflammasome complex via CARD-CARD interactions (Proell et al., 2013).

Pro-caspase-1 is then able to undergo proximity induced auto-proteolytic processing and activation (Yang et al., 1998). In addition to the requirement for signals 1 and 2, inflammasomes are further controlled by additional regulatory pathways and processes, such as ubiquitylation, which serve to fine-tune inflammasome activation and are discussed in more detail below.

1.3.3 NLRP3 inflammasome

Also known as Cryopyrin or NALP3, NLRP3 is perhaps the most extensively studied and best understood NLR. NLRP3 consists of an N-terminal PYD, an internal nucleotide-binding domain (NBD) and a C-terminal LLR domain (Figure 1.4). Because of its PYD, it has an absolute requirement for the ASC adaptor to activate caspase-1.

NLRP3 is poorly expressed in resting cells and requires both transcriptional and post- translational-mediated priming (Mehta et al., 2001) (He et al., 2016a). NLRP3 transcription can be driven by c-Myc, AP-1 and NF-κB, transcription factors triggered by either MyD88 or TRIF-dependent signalling pathways (Anderson et al., 2008;

Bauernfeind et al., 2012; Bauernfeind et al., 2009). In myeloid cells, NLRP3 transcripts are also negatively regulated by the microRNA miR-233 (Bauernfeind et al., 2012).

In an experimental context, it is common to prime cells for several hours to drive the production of more NLRP3 and thereby induce stronger inflammasome activation.

However, short signal 1 treatments or the simultaneous addition of signal 1 and signal

2 agonists can cause rapid NLRP3-inflammasome activation which is independent of new transcription (Schroder et al., 2012) but dependent on MyD88 and IRAK1 (Lin et al., 2014). The exact pathways which control non-transcriptional priming are not yet fully understood. However, they do appear to be temporally regulated as early phase non-transcriptional priming (within 10 mins) is MyD88 dependent and intermediate phase non-transcriptional priming (30-60 mins) is TRIF dependent (Fernandes-

Alnemri et al., 2013). Post-translational modifications also play a role in NLRP3 priming. Studies have shown that newly synthesised NLRP3 is held in an inactive state by ubiquitylation, which must be removed for it to be activated (Yang et al., 2017). One known regulator of this process is the deubiquitinating enzyme (DUB) BRCC3 which removes K63-linked polyubiquitin from NLRP3 following LPS stimulation, thereby promoting inflammasome activation (Py et al., 2013).

Once primed, NLRP3 is set to be activated by signal 2. NLRP3 is a particularly versatile sensor and has not been shown to recognise a specific ligand. Instead, it can be activated by a host of diverse agonists including adenosine triphosphate (ATP), ionophores such as the Streptomyces toxin nigericin (Mariathasan et al., 2006;

Perregaux and Gabel, 1994), sterile particulates (e.g. silica, alum and MSU) (Hornung et al., 2008; Martinon et al., 2006) and pore-forming toxins from pathogenic bacteria

(Munoz-Planillo et al., 2013). Given the breadth of its activating compounds, it is believed that all defined NLRP3 agonists converge upon a single signal-mediator which in turn activates NLRP3 (He et al., 2016a; Muñoz-Planillo et al.). Several mediators have been proposed such as potassium ion (K+) efflux, calcium ion

29 signalling, lysosome disruption and mitochondrial reactive oxygen species (ROS) (He et al., 2016a). As of now, K+ efflux is thought to be the most likely common and unifying signal for NLRP3 activation (He et al., 2016a). Indeed, activation of NLRP3 by ATP, nigericin or pore-forming toxins, which directly induce K+ efflux, can be blocked by the addition of extracellular potassium (Munoz-Planillo et al., 2013; Walev et al., 1995).

Likewise, NLRP3 activation by other signals such as increased intracellular calcium or particulate matter, such as silica, both require K+ efflux and can be blocked by adding extracellular potassium (Hornung et al., 2008; Munoz-Planillo et al., 2013).

How NLRP3 detects K+ efflux is not yet fully understood. In an attempt to identify novel regulators of NLRP3 activation three independent studies showed that the serine/threonine kinase NEK7 is an essential mediator of NLRP3 inflammasome activation (He et al., 2016b; Schmid-Burgk et al., 2016; Shi et al., 2016b). Independent from its kinase activity, NEK7 interacts with the LRR domain of NLRP3 in response to

K+ efflux (He et al., 2016b; Shi et al., 2016b), and assists in NLRP3 oligomerisation and ASC protein nucleation (He et al., 2016b). NEK7 appears to be a common activator of NLRP3 as cells lacking NEK7 do not respond to treatment with ATP, nigericin, MSU crystals and alum (He et al., 2016b). However, it still remains unclear exactly how NEK7 links K+ efflux to NLRP3 activation.

Recently, an “alternative” NLRP3 inflammasome pathway, which activated is independently of K+ efflux, was described in human monocytes (Gaidt et al., 2016).

The alternative inflammasome can be activated by LPS alone, and mediates pro-IL-

1β activation and secretion, but not pyroptosis (Gaidt et al., 2016). A single signal was thus able to activate inflammasomes via this ‘alternative’ pathway. Mechanistically, this pathway was dependent on TLR4, TRIF, RIPK1, FADD, caspase-8 and the inflammasome components NLRP3, ASC and caspase-1 (Gaidt et al., 2016).

However, despite requiring the canonical inflammasome components and triggering caspase-1 autoproteolytic processing, the inflammasome-speck does not form (Gaidt et al., 2016).

In addition to the factors regulating its priming, the activity of NLRP3 is also modulated (Eldridge and Shenoy, 2015; He et al., 2016a). Guanylate-binding proteins

(GBPs) are a large family of IFN-y inducible GTPases implicated in many antimicrobial functions such as trafficking bactericidal peptides to autophagosomes and enhancing

ROS killing (Kim et al., 2011a). GBP5 was found to be an inducible and specific regulator of NLRP3-activation. It was shown to enhance NLRP3-mediated ASC oligomerisation in IFN-γ stimulated macrophages in response to ATP and bacterial infection but not particulate matter such as alum (Shenoy et al., 2012). Ubiquitylation also plays important roles in both promoting and suppressing NLRP3 activity which is discussed further in Chapter 5.

1.3.4 NLRC4 inflammasome

NLRC4 consists of N-terminal CARD, an internal NBD and C-terminal LLR domain (Figure 1.4), meaning that it can recruit caspase-1 with or without the ASC adaptor (Broz et al., 2010b). Under resting conditions, NLRC4 is held in an autoinhibited state by the steric-occlusion of its NBD by an internal winged-helix domain (WHD) and C-terminal LRR domain. Mutational disruption of these interactions causes constitutive activation (Hu et al., 2013; Tenthorey et al., 2014), as does deletion of the LRR domain (Poyet et al., 2001). Unlike NLRP3, NLRC4 is constitutively expressed in cells under resting conditions however, its expression can be induced further by TNF stimulation (Gutierrez et al., 2004). Despite not requiring transcriptional priming, NLRC4 is post-translationally primed by phosphorylation

(Duncan and Canna, 2018). Flagellin or the type three secretion system (T3SS) rod

31 protein PrgJ can induce phosphorylation of Ser533 by PKCδ which is essential for

NLRC4 activation (Matusiak et al., 2015; Qu et al., 2012).

Early studies showed that overexpression of NLRC4 triggered caspase-1 activation (Geddes et al., 2001; Poyet et al., 2001). It was later shown by Mariathasan et al. that NLRC4 could be specifically activated by Salmonella enterica Typhimurium

(STm) infection which could induce caspase-1-mediated cell death independently of the ASC adaptor, which is required for most inflammasomes (Mariathasan et al.,

2004).

The NLRC4 inflammasome is now known to be activated by multiple bacterially derived PAMPs, including flagellin and various protein components of the Gram- negative T3SSs (Duncan and Canna, 2018; Eldridge and Shenoy, 2015). However,

NLRC4 is not the receptor for these ligands which are in fact detected by NLR family apoptosis inhibitory proteins (NAIPs). NAIPs are NLR proteins which have three tandem baculovirus inhibitor repeat (BIR) domains at their N-terminus instead of a

CARD or PYD (Figure 1.4). Direct binding of a single ligand molecule is sufficient to activate a NAIP and likewise, a single ligand-bound NAIP is sufficient to trigger NLRC4 oligomerisation into ring-like inflammasomes (Hu et al., 2015; Zhang et al., 2015).

Ligand-binding induces conformational changes in NAIPs which expose a single nucleation surface which interacts with NLRC4. NLRC4 then undergoes similar structural rearrangements, causing the sequential recruitment of additional NLRC4 proteins into a multimeric ring-shaped complex which nucleates ASC thereby forming the inflammasome (Diebolder et al., 2015; Hu et al., 2015; Zhang et al., 2015).

Humans possess a single NAIP which can bind to bacterial flagellin, T3SS needle and T3SS rod proteins (Grandjean et al., 2017; Reyes Ruiz et al., 2017; Yang et al., 2013a). In mice, however, Naip genes have undergone functional and genetic

32 divergence. The genome of C57BL/6 (B6) mice encodes four functional Naip genes;

Naip1, Naip2, Naip5 and Naip6 (Endrizzi et al., 2000; Kong et al., 2011) and Naip3 has been pseudogenized (Growney and Dietrich, 2000). Murine Naip1 binds T3SS needle proteins (Yang et al., 2013a), Naip2 binds T3SS rod proteins (Zhao et al., 2011) and Naip5 and 6 bind flagellin; all of which converge on NLRC4 (Kofoed and Vance,

2011).

Because it has a CARD domain NLRC4 can directly bind to and activate caspase-1 in the absence of ASC. Therefore, NLRC4 can assemble two distinct inflammasome structures, one with ASC and one without, which cause different downstream effector mechanisms. The NLRC4-ASC complex triggers caspase-1 auto-processing into its p10 and p20 subunits, mediates cell death and efficient proteolytic activation of pro-IL-1β. The NLRC4-caspase-1 complex (also called the

‘death complex’) does not facilitate caspase-1 auto-processing. Instead, this complex produces a proteolytically activate, full-length p45 pro-caspase-1 enzyme which is able to activate pyroptotic cell death but is a weak processor of pro-IL-1β (Broz et al.,

2010b; Yang et al., 1998). NLRC4 can have distinct roles in different cell types. For example, neutrophils challenged with STm to activate the NLRC4 inflammasome do not undergo pyroptosis enabling them to remain viable for longer periods of time and secrete more bioactive IL-1β (Chen et al., 2014).

1.3.5 AIM2 inflammasome

AIM2 does not possess an LRR domain, and instead has a C-terminal HIN200 domain (Figure 1.4) which binds cytosolic double-stranded DNA (dsDNA)

(Burckstummer et al., 2009; Fernandes-Alnemri et al., 2009; Hornung et al., 2009;

Roberts et al., 2009) which can be derived from pathogens (Rathinam et al., 2010) or the host (Di Micco et al., 2016; Hu et al., 2016).

AIM2 is an IFN-inducible protein and can be primed by transcriptional upregulation in response to IFN-γ and IFN-β (Burckstummer et al., 2009; Fernandes-

Alnemri et al., 2009) or more typical priming agents such as LPS (Hornung et al.,

2009). Under resting conditions, AIM2 is autoinhibited through intramolecular interactions between its HIN200 and PYD domains (Jin et al., 2013). The HIN200 domain binds DNA independently of its sequence by recognising the sugar-phosphate backbone of dsDNA. It has been shown that dsDNA of a minimum of ~80 bp is sufficient to activate AIM2 (Jin et al., 2012). dsDNA acts as an oligomerization platform for multiple AIM2 molecules, this interaction displaces the PYD domain which mediates ASC polymerisation and caspase-1 recruitment (Lu et al., 2015). AIM2 has also been reported to trigger NLRP3 activation in response to infection with Legionella pneumophila (Cunha et al., 2017). Following infection, activated AIM2 yielded an unprocessed but activate caspase-1, similar to NLRC4 (Broz et al., 2010b), though this mechanism required the ASC adaptor protein (Cunha et al., 2017). Active p45 caspase-1 then triggered pore formation and K+ efflux which in turn activated the

NLRP3 inflammasome, caspase-1 processing and cytokine processing and release

(Cunha et al., 2017).

1.3.6 Non-canonical inflammasome activation

The non-canonical pathway refers to the activation of the NLRP3-ASC- caspase-1 inflammasome via upstream actions of caspase-4 in response to cytosolic

LPS (Kayagaki et al., 2011) or endogenous host-derived oxidized phospholipids such as 1-palmitoyl2-arachidonoyl-sn-glycerol-3-phosphorylcholine (oxPAPC) (Zanoni et al., 2016). In addition to its downstream role in activating NLRP3, Caspase-4 is also the receptor for the pathway (Shi et al., 2014). Caspase-4 directly binds the penta- and hexa-acylated lipid A portions of LPS via its CARD domain, (Hagar et al., 2013;

Kayagaki et al., 2013; Shi et al., 2014) or endogenous oxPAPC at its catalytic domain

(Zanoni et al., 2016) which triggers its oligomerisation and autoproteolytic activation.

Like other inflammasome sensors caspase-4 can be primed, particularly in response to type I interferons which induce caspase-4 expression and enhance its activation (Broz et al., 2012; Rathinam et al., 2012). In addition to activating NLRP3 caspase-4 can also directly cleave and activate the pyroptosis executioner protein

GSDMD triggering cell-death independently of caspase-1 (Kayagaki et al., 2015; Shi et al., 2015).

How caspase-4 mediates the activation of NLRP3 remains unclear, but it appears to be dependent on K+ efflux (Ruhl and Broz, 2015). Caspase-4 is proposed to mediate K+ efflux by two mechanisms which may trigger cell-intrinsic activation of

NLRP3. One mechanism proposes that caspase-4-dependent activation of GSDMD leads to K+ efflux and NLRP3 activation as Gsdmd-deficient macrophages do not activate the non-canonical NLRP3 inflammasome despite retaining caspase-4 activation (Kayagaki et al., 2015). Alternatively, a study by Yang et al. showed that non-canonical caspase-1 activation required pannexin-1 channels which mediate the releases of ATP (Yang et al., 2015). The authors went on to show that caspase-4 cleaves the pannexin-1 channel and proposed that this triggers ATP release causing autocrine activation P2X7channels causing mediate K+ efflux and NLRP3 activation

(Yang et al., 2015).

1.3.7 PYRIN inflammasome

Unlike the inflammasome sensors discussed thus far, PYRIN is a non-ALR/NLR protein. It consists of an N-terminal PYD domain, central B-box zinc finger domain (B- box) and coiled-coil domains, and a C-terminal B30.2 domain (Figure 1.4) (D'Cruz et al., 2013). PYRIN (encoded by MEFV) is predominantly expressed in immune cells

35 such as neutrophils, monocytes, and dendritic cells and to a lesser extent in macrophages (Seshadri et al., 2007; Tidow et al., 2000). Transcription of the MEFV gene has been shown to be triggered by a number of pro-inflammatory stimuli including LPS, IFN-γ, TNF and IL-1β (Abedat et al., 2002; Centola et al., 2000). Under resting conditions, PYRIN is held in a closed inactive state. This is mediated by the serine-threonine kinases PKN1 and PKN2 which phosphorylate PYRIN at Ser205 and

Ser241 (Ser208 and Ser242 in human) (Park et al., 2016). Phosphorylation recruits the phosphoserine-binding protein 14-3-3 which holds PYRIN in an autoinhibited conformation (Gao et al., 2016; Park et al., 2016).

PYRIN is also distinct from other inflammasome sensors in that its activation follows a guard-hypothesis model more akin to the NLRs found in plants (Jones and

Dangl, 2006). Specifically, PYRIN has been shown to sense the inactivation of the

RhoA GTPase (Xu et al., 2014). Thus, PYRIN can be activated by multiple bacterial toxins which target Rho GTPases such as TecA from Burkholderia (Aubert et al.,

2016), VopS from Vibrio parahaemolyticus, the C3 toxin of Clostridium botulinum and

TcdB toxin from Clostridium difficile (Xu et al., 2014). Interestingly, microtubules have also been shown to affect PYRIN activation, however, the underlying mechanisms of this remain elusive (Xu et al., 2014).

Rather than mediating the direct activation of PYRIN, RhoA inactivation leads to the loss of PYRIN inhibition. The activity of PKN1/2 kinases is regulated by RhoA.

Therefore, inactivation of RhoA for example by the C3 toxin which ADP-ribosylates

Asn41, causes the subsequent inactivation of PKN1/2, leading to reduced phosphorylation of PYRIN and its eventual activation (Gao et al., 2016; Park et al.,

2016). Activation of PYRIN causes it to oligomerise via its coiled-coil domain (Yu et al., 2007), recruit ASC and form the inflammasome (Vajjhala et al., 2014).

1.3.8 NLRP1 inflammasome

NLRP1 was the first NLR shown to form the multiprotein complexes we now call inflammasomes (Chavarria-Smith and Vance, 2015; Martinon et al., 2002).

Humans possess a single NLRP1 gene, however, in other mammalian species,

NLRP1 has undergone considerable diversification (mice, for instance, have three).

Because of these variations, NLRP1 has been reported to respond to different stimuli depending on the allele and deriving species (Chavarria-Smith and Vance, 2015). For example, human NLRP1 is thought to be activated by muramyl dipeptide (Faustin et al., 2007), Nlrp1b from BALB/c mice is responsive to the lethal toxin (LT) of Bacillus anthracis. The Nlrp1b from C57BL/6 is not (Levinsohn et al., 2012), despite maintaining the propensity to form active inflammasomes (Chavarría-Smith and

Vance, 2013). Compared with other inflammasome-NLRs, the NLRP1 family have a distinct domain organisation. All possess an N-terminal NBD and a central LRR domain, they also have a function to ﬁnd domain (FIIND), followed by the CARD

(Figure 1.4) (Hlaing et al., 2001). Human and mouse NLRP1 proteins are also distinct as they require autoproteolytic processing of the FIIND to become active (Finger et al., 2012; Frew et al., 2012).

Anthrax toxin is comprised of three proteins: the membrane binding protective antigen (PA); and the enzymatic lethal (LF) and edema (EF) factors. The PA mediates the cytoplasmic delivery of the EF and LF. The EF possesses adenylyl cyclase activity whereas the LF is a metalloprotease (Liu et al., 2014b). Nlrp1b activation requires the metalloprotease activity of LF (Chavarría-Smith and Vance, 2013; Levinsohn et al.,

2012) however, the precise mechanism is not known. Two models have been proposed. Studies have shown that LF can directly cleave the N-terminus of Nlrp1b which is required for its inflammasome assembly and activation of caspase-1

(Chavarría-Smith and Vance, 2013; Levinsohn et al., 2012) however other factors may be involved however variants of Nlrp1b which are not activated by anthrax toxin are also cleaved by the LF (Hellmich et al., 2012). The indirect model suggests that Nrp1b is inhibited by an unknown factor which is cleaved and destabilised by LF leading to inflammasome activation (Squires et al., 2007).

1.3.9 Other inflammasomes

Despite there being extensive work on a select few inflammasome sensors many NLRs remain poorly characterised. The human-specific NLRP7 was described to mediate ASC-dependent activation of caspase-1 and IL-1β and IL-18 secretion.

NLRP7 is activated by bacterial-derived acylated lipopeptides and synthetic lipoproteins FSL-1 or PAM3CSK4 (Khare et al., 2012). Interestingly, NLRP7 activation does not induce pyroptotic cell lysis, suggesting it only mediates caspase-1 dependent cytokine secretion (Khare et al., 2012).

Other inflammasome sensors such as NLRP12 and NLRP6 play important roles in regulating the environment of the host gut. NLRP12 has been described to do this via inflammasome-dependent and independent roles. Multiple reports suggest that

NLRP12 is important for the maintenance of colonic homeostasis by supressing inflammation triggered by NF-κB, ERK and AKT signalling pathways (Allen et al., 2012;

Zaki et al., 2011). In vivo NLRP12 can suppress inflammation induced by dextran sodium sulfate during experimental colitis in mice (Shi et al., 2016a). NLRP12 can also mediate inflammasome-dependent protection against Yersinia pestis by promoting IL-

18 production and IFN-γ induction to limit infection; the precise ligand or activator of

NLRP12, however, is unknown (Vladimer et al., 2012). Similarly, NLRP6 can regulate colonic homeostasis and mediate anti-inflammatory responses (Levy et al., 2017).

Nlrp6-/- mice had heightened resistance to infection and could control the replication

38 of intracellular pathogens Lm and STm. Much like the roles described for NLRP12,

NLRP6 can negatively regulated MAPK and NF-κB signal transduction after TLR activation (Anand et al., 2012). It has also been suggested that NLRP6 is involved in the maintenance of healthy gut flora (Levy et al., 2015). Microbiota-derived metabolites have been shown to activate NLPR6, triggering the release of IL-18.

This in turn, causes the secretion of antimicrobial peptides which help to maintain a normal microbiota (Levy et al., 2015).

1.4 Caspase-1 Substrates and Targets Inflammasome assembly culminates in the activation of caspase-1 which is the executioner of inflammasome-dependent signalling (Franchi et al., 2009). Following its recruitment to the inflammasome complex pro-caspase1 undergoes autoproteolytic activation (Figure 1.5) into its p20 and p10 subunits which is a hallmark feature of caspase-1 activation (Yang et al., 1998). Interestingly, different caspase-1 effector mechanisms have differential requirements for autoprocessing (Broz et al.,

2010b). For example, pro-IL-1β maturation requires autoprocessing whereas the induction of pyroptosis does not (Broz et al., 2010b).

A recent study showed that active caspase-1 exists as multiple cleaved forms which are generated by sequential self-processing and eventual inactivation (Boucher et al., 2018). It is suggested that inflammasome-associated unprocessed p45 and an intermediate p33/p10 form of caspase-1 are proteolytically active, while the generation of p20/p10 subunits triggers deactivation and dissociation from the inflammasome complex (Boucher et al., 2018). The active forms of caspase-1 were shown to interact with oligomeric ASC suggesting that active caspase-1 is associated with the inflammasome speck. The sequential processing is proposed to control caspase-1 activity and substrate specificity which develops over time (Boucher et al., 2018).

Figure 1.5: Domain organisation and cleavage sites of Caspase-1 Schematic representation of mouse caspase-1. The N-terminal CARD facilitates recruitment to the inflammasomes via CARD-CARD interactions with ASC or NLRC4. The catalytic cysteine 284 (red) is located in the p20 domain which along with the p10 subunit forms the active enzyme. Caspase-1 is activated by autoproteolysis at multiple sites which are the domain linker regions (black).

While in its p45 form, caspase-1 mediates selective substrate processing of GSDMD, p33/p10 mediates general substrate processing which includes pro-IL-1β, pro-IL-18 and GSDMD, and autoprocessing into p20/p10 causes inactivation (Boucher et al.,

2018). In vitro recombinant caspase-1 has been demonstrated to have concentration- dependent substrate promiscuity (Walsh et al., 2011). This mechanism is therefore thought to be important in regulating caspase-1 effector mechanisms by controlling its ability to process key substrates such as pro-IL-1β and GSDMD (Boucher et al., 2018).

1.4.1 pro-IL-1β

Pro-IL-1β is the original substrate of caspase-1 and is predominantly expressed in hematopoietic-derived cell types such as macrophages, monocytes, and dendritic cells (Dinarello, 1996). It is also one of the most potent inflammatory cytokines which mammals produce and is therefore tightly regulated, requiring transcriptional induction

(Cogswell et al., 1994) and subsequent activation by caspase-1 (Dinarello, 1998). In addition, it is regulated post-transcriptionally via the AU-rich elements in the 3’ UTR of

40 the IL1B transcript which mediate its degradation (Kastelic et al., 1996). Co-treatment of cells with LPS and interferon-γ increases the IL1B transcript accumulation by enhancing both its transcription and the stability of its transcript. These mechanisms prevent the production of pro-IL-1β in response to erroneous stimuli (Arend et al.,

1989). Pro-IL-1β is further regulated at the post-translational level. Pro-IL-1β is an inactive precursor protein which requires proteolytic activation and conversion into its bioactive p17 in order to be recognised by the interleukin-1 receptor 1 (IL-1R1) (Black et al., 1988; Mosley et al., 1987b). Caspase-1 sequentially cleaves pro-IL-1β at two aspartic acid residues Asp26 and Asp116 (Howard et al., 1991a). Cleavage at Asp26 first produces an intermediate p28 fragment, which exposes Asp116 for cleavage releasing the bioactive p17 portion of IL-1β (Swaan et al., 2001).

Following its maturation IL-1β, is released from the cell by the unconventional secretion pathway, rather than by conventional ER-Golgi-mediated secretion

(Rubartelli et al., 1990). The precise mechanisms behind the release of IL-1β have been hotly debated with multiple mechanisms being put forward as follows: release via microvesicles and exosomes (Qu et al., 2007), via caspase-1 induced pore formation (Fink and Cookson, 2006), and passive release by cell lysis. Because of the speed at which inflammasome activation and cell death occur, it has been difficult to uncouple IL-1β secretion from cell lysis. Studies using glycine as an osmoprotectant to inhibit cell lysis in macrophages showed that caspase-1-dependent IL-1β secretion can occur independently of cell-death. Importantly, glycine treatments do no block caspase-mediated pore formation suggesting that these may serve as conduits for IL-

1β release (DiPeso et al., 2017; Fink and Cookson, 2006). More recent studies have shown that ectopic expression of inflammasome components in mouse embryonic fibroblasts (MEFs) causes IL-1β secretion independently of cell death (Conos et al.,

2016). Similarly, neutrophils infected with STm and monocytes activated by the alternative-NLRP3 pathways do not undergo cell death by pyroptosis yet still release

IL-1β (Chen et al., 2014; Gaidt et al., 2016). Not until very recently has cell death and

IL-1β been temporally separated in macrophages, with studies showing that pore- formation induced by GSDMD (following its cleavage by caspase-1) facilitates the release of IL-1β from living macrophage cells (Evavold et al., 2017; Heilig et al., 2017).

Once activated and released, IL-1β has potent pro-inflammatory properties and can stimulate fever, neutrophil influx and the acute phase response (Cybulsky et al.,

1986; Dinarello, 1996). Bioactive p17 IL-1β binds its primary receptor subunit IL‑1 receptor type I (IL‑1R1), this facilitates its Tir-domain mediated dimerization with the

IL‑1R accessory protein (IL‑1RAP) and leading to MyD88 dependent signalling (Sims and Smith, 2010). Signalling via IL-1R is also extensively regulated. A decoy receptor called IL‑1 receptor type II (IL-1RII) which lacks a cytoplasmic Tir domain can interfere with sequester IL-1RAP from forming its functional dimer (Malinowsky et al., 1998).

Cells also produce IL-1R antagonists (IL-1RA) which compete with mature IL-1β for binding to the IL-1R (Dinarello and Thompson, 1991).

1.4.2 pro-IL-18

Pro-IL-18 is another common substrate of caspase-1 and is also synthesised as an inactive precursor which requires proteolytic activation by caspase-1 to release its bioactive p18 component (Ghayur et al., 1997; Gu et al., 1997). IL-18, however, has some distinctive features. IL-18 is more widely expressed, and is produced by multiple tissues, including the gastrointestinal tract and spleen, and can be secreted by epithelial and immune cell types (Puren et al., 1999). Unlike pro-IL-1β, IL-18 is constitutively expressed (Puren et al., 1999) however, stimulation with stimuli such as

LPS can increase mRNA levels (Marshall et al., 1999) which may be caused by to the

42 enhanced stabilisation of the IL18 transcript (Bufler et al., 2004). IL-18 is cleaved by caspase-1 at Asp36 (Gu et al., 1997) and like IL-1β does not possess a signal peptide and is therefore likely secreted similarly by the GSDMD pore.

Once secreted, mature IL-18 serves as the ligand for the IL-18 receptor α chain

(IL-18Rα), which subsequently recruits the accessory IL-18 receptor β (IL-18Rβ) for signal transduction via MyD88 in a manner similar to induced by IL-1 (Debets et al.,

2000; Gerdes et al., 2002). Despite IL-18 and IL-1β signalling via similar pathways

(Tsutsumi et al., 2014) IL-18 mediated signalling is less potent than IL-1β and requires

~4-fold higher concentrations of IL-18 to achieve comparable activation of NF-κB or

AP-1 transcription factors (Lee et al., 2004).

1.4.3 IL-1α

Despite also stimulating the IL-1R, IL-1α differs in that it is expressed in a wider range of cell types, including epithelial cells, endothelial cells, and keratinocytes, several of which lack caspase-1 (Hurgin et al., 2007). Despite this, IL-1α is able to induce inflammatory responses as it does not require caspase-1-dependent proteolysis to become active. The pro-form of IL-1α can be cleaved by calpains.

(Afonina et al., 2015; Kobayashi et al., 1990) which increases its activity and binding affinity for the IL-1R (Mosley et al., 1987a). Though IL-1α is not cleaved by caspase-1

(Howard et al., 1991b), its release from macrophages is caspase-1-dependent, it is therefore considered to be an inflammasome-dependent ‘alarmin’ (Groß et al., 2012).

The precise regulatory mechanisms underlying IL-1α secretion remain elusive, with studies suggesting both caspase-1-dependent and -independent mechanisms being involved (Groß et al., 2012). Phagocytosed particles such as MSU or silica were shown to induce caspase-1-independent IL-1α release, while stimulation with nigericin or ATP uses the inflammasome pathway (Groß et al., 2012). Interestingly, both

43 pathways generated cleaved IL-1α which was processed by the calpain proteases.

Much like IL-1β, the secretion of IL-1α has been proposed to be dependent on pyroptosis (Eigenbrod et al., 2008). IL-1α release has been suggested to be dependent on IL-1β and is blocked in Il1b-/- dendritic cells, suggesting their secretion may require a co-dependent mechanism (Fettelschoss et al., 2011). More recently, the secretion of IL-1α in response to inflammasome activation was also been shown to be dependent on GSDMD pore formation, but not on cell death (Evavold et al.,

2017). However, it was not made clear whether the secreted IL-1α was processed or not. These studies along with others suggest that caspase-1 mediated IL-1α secretion requires the formation of GSDMD pores which form non-specific channels allow bioactive IL-1α to pass through (Evavold et al., 2017; Heilig et al., 2017).

1.4.4 Gasdermin D

Until recently how caspase-1 caused pyroptosis was unknown. GSDMD had been identified previously as a substrate of caspase-1 (Agard et al., 2010) but was only recently discovered to be the executioner of pyroptotic cell death. Multiple studies showed that GSDMD can be cleaved at Asp276 in the mouse (Asp275 in human) by both caspase-1 and caspase-4 and does not require caspase-1 autoproteolysis (He et al., 2015; Kayagaki et al., 2015; Shi et al., 2015). Under resting conditions, the full- length GSDMD protein is held in an autoinhibited conformation (Ding et al., 2016;

Kuang et al., 2017).Cleavage is required for the induction pyroptosis which is specifically mediated by the free p30 N-terminal fragment of GSDMD, which alone is sufficient to cause cell death (He et al., 2015; Kayagaki et al., 2015; Shi et al.,

2015). Following its release, the p30 of GSDMD localises to the plasma membrane where it oligomerises and forms pores approximately 10–16 nm in diameter (Ding et al., 2016). Studies showed that Ile104 is key residue in mediating homomeric

44 oligomerising as I104N mutation blocks pore formation (Aglietti et al., 2016). Binding to the plasma membrane is mediated by specific recognition of cardiolipin (CL), phosphatidylinositol 4-phosphate (PI(4)P) and phosphatidylinositol 4,5-bisphosphate

(PI(4,5)P) lipid types which are abundant in the inner plasma membrane (Aglietti et al., 2016; Ding et al., 2016). This degree of specificity also limits aberrant lytic effects on neighbouring cells following lysis as it prevents GSDMD form attaching to the outer leaflet of plasma membranes (Liu et al., 2016b). Interestingly, these lipids are also found in the membranes of Gram-positive and Gram-negative bacteria.

Supernatants from pyroptotic cells have been shown to have bactericidal properties, which led some to suggest that GSDMD may also harbour direct antimicrobial activity during infection (Liu et al., 2016b). GSDMD pores have also been shown to be required for IL-1β release independently of their role in cell lysis (Evavold et al.,

2017; Heilig et al., 2017).

While GSDMD cleavage and pyroptosis are now well studied in macrophages, recent studies have highlighted that pyroptosis might not be universal in all cell types that undergo inflammasome activation. For example, a recent study revealed that dendritic cells, stimulated with host-derived oxPAPC activated the non-canonical inflammasome, secreted IL-1β secretion but did not undergo pyroptotic cell death

(Zanoni et al., 2016). A so-called “hyperactive” state can also be induced by treating macrophages with PGPC (1-palmitoyl-2-glutaryl-sn-glycero-3-phosphocholine),

POVPC (1-palmitoyl-2-(5′-oxo-valeroyl)-sn-glycero-3-phosphocholine) or infecting them with S. aureus OatA-/- which lack the O-acetyltransferase A (Evavold et al., 2017).

Using these inflammasome stimuli, Evavold et al. were able to uncouple cell-death and cytokine secretion and discover that IL-1β release occurred via GSDMD pores under conditions in which pyroptosis remained undetectable.

It remains to be shown how cells control the degree to which GSDMD is cleaved and therefore the extent of membrane permeabilization and cell death. A recent study showed that caspase-3 and -7 can cleave GSDMD at Asp87 and inactivate it, thereby preventing GSDMD pore-formation and pyroptosis (Taabazuing et al., 2017).

Caspase-3 and -7 are also substrates of caspase-1 (Lamkanfi et al., 2008), and their activation by either caspase-1 or by apoptotic stimuli could cause GSDMD inactivation.

It is tempting to speculate whether cleavage at Asp87 plays a role in suppressing pyroptosis during pyroptosis-independent inflammasome responses, for instance by the alternative inflammasome in monocytes or STm induced NLRC4 activation in neutrophils (Chen et al., 2014; Gaidt et al., 2016).

1.5 Inflammasome signalling during bacterial infection 1.5.1 Pathogen sensing by inflammasomes

Despite residing in the cytosol, inflammasomes can detect pathogens that inhabit a variety of cellular milieus. Infection by diverse bacterial species can trigger inflammasome activation (von Moltke et al., 2013). Because of the array of signals which can be collectively detected by host NLR/ALRs, a bacterial infection can be sensed either directly via detection of bacterial proteins and structural components or indirectly via the actions of bacterial toxins (Storek and Monack, 2015).

The intracellular pathogen STm is a good example of how inflammasomes can detect infection as it can activate multiple inflammasome sensors (Storek and Monack,

2015). The detection of flagellin by NAIP/NLRC4 plays an important role in mediating protective inflammasome responses as mutant bacteria which do not express flagella components (FliC and FljB in STm) do not activate inflammasomes (Franchi et al.,

2006). By contrast mutant strains which constitutively express flagellin induce stronger

46 inflammasome-dependent response and cause attenuated infection in vivo (Miao et al., 2010a). STm usually reside in modified membrane-bound compartments called

Salmonella-containing vacuoles (SCVs). How flagellin and components of the T3SS, such as the inner rod protein is PrgJ and the needle protein is PrgI, gain access to the cytosol to be detected by NAIP/NLRC4 (Miao et al., 2010b; Zhao et al., 2011) is not known but it is thought that they may leak from the SCV into the cytoplasm (Broz et al., 2010a; Zhao et al., 2011). When STm does enter the cytosol, they can also activate the non-canonical NLRP3 inflammasome via caspase-4 mediated detection of LPS

(Aachoui et al., 2013). As a result, STm can trigger dual activation of NLRC4 and

NLRP3, and interestingly both NLRs can be recruited to the same inflammasome complex (Broz et al., 2010a; Man et al., 2014).

The causative agent of Legionnaires disease, L. pneumphilia, also reside in modified vesicles called Legionella-containing vacuoles (LCVs) (Isberg et al., 2009).

L. pneumphilia induces the NLRC4 inflammasome via NAIP5 in a T4SS-dependent manner (Jamilloux et al., 2013). As with STm how the cytosolic inflammasomes can detect vacuolar bacteria is not fully understood. Over recent years a growing body of evidence shows that host proteins can mediate the release of bacteria into the cytosol.

GBPs have been shown to localise to bacterial-containing vacuoles and lyse them

(Kim et al., 2011a; Meunier et al., 2014). This serves to expose bacteria and bacterial ligands to cytosolic receptor and can enhance non-canonical inflammasome activation via caspase-4 (Meunier et al., 2014).

Other intracellular bacteria do not reside in vacuoles, for example, Lm invade cells by pathogen-induced endocytosis (Ireton, 2007) and rapidly enter the cytosol by lysing the entry vacuole with its pore-forming toxin listeriolysin O (LLO) and phospholipases (Hybiske and Stephens, 2008). Once in the cytosol, Lm can activate

47 multiple inflammasome receptors including NLRC4, AIM2 and NLRP3 which cooperatively trigger proinflammatory responses (Kim et al., 2010; Wu et al., 2010).

The intracellular pathogen Francisella tularensis has a similar invasion process, initially residing in a vacuole before escaping into the cytoplasm (Chong et al., 2008) where they activate the AIM2 inflammasome (Fernandes-Alnemri et al., 2010; Jones et al., 2010; Rathinam et al., 2010). Both Lm and F. tularensis have been shown to undergo bacteriolysis while in the host cytosol which enables AIM2 to access bacterial

DNA for activation (Broz and Monack, 2011; Jones et al., 2010; Warren et al., 2010).

A recent study showed that host proteins GBP2 and GBP5 promote the activation of

AIM2 in response to Francisella novicida by binding to and lysing the bacteria (Man et al., 2015; Meunier et al., 2015). Inflammasomes can also become activated by indirectly detecting the actions of bacterial toxins. This is the case for the PYRIN and

NLRP1 inflammasomes which can detect the catalytic activity of bacterial toxins

(Levinsohn et al., 2012; Xu et al., 2014) (See Sections 1.3.7 and 1.3.8)

Similarly, the NLRP3 inflammasome can detect diverse pore-forming toxin- producing pathogens which induced K+ efflux (Munoz-Planillo et al., 2013). These can include both Gram-negative and Gram-positive bacteria including S. aureus (Munoz-

Planillo et al., 2009), Clostridium difficile (Ng et al., 2010), Vibrio cholerae (Toma et al., 2010), Streptococcus pneumoniae (Karmakar et al., 2015) and enterohaemorrhagic E. coli (Zhang et al., 2012).

1.5.2 Bacterial circumvention of inflammasome sensing and mechanisms

Because inflammasomes play such important roles in antimicrobial immunity, it is unsurprising that bacteria have evolved mechanisms to subvert their activation and downstream responses. The YopM effector from Yersinia can inhibit the PYRIN inflammasome by hijacking the hosts own regulatory mechanisms (Chung et al.,

2016). YopM lacks any catalytic activity but has been shown to bind to host kinases

PKN1 and 2 which are responsible for maintaining the inhibition of PYRIN (Park et al.,

2016). In doing so YopM enhances PKN-mediated phosphorylation of PYRIN thus maintaining it in an inhibited state (Chung et al., 2016). Effectors from enteropathogenic and enterohemorrhagic Escherichia coli can also disrupt caspase- mediated responses. For example, the effector NleF has been shown to block caspase-4 activation and IL-18 release in epithelial cells (Pallett et al., 2016) and NleA blocks caspase-1 activation and secretion of IL-1β in macrophages (Yen et al., 2015).

Interestingly, NleA also takes advantage of native regulatory mechanisms by directly interacting with NLRP3 and preventing its deubiquitylation which is essential for its activation and inflammasomes complex formation (Py et al., 2013; Yen et al., 2015).

Other effectors function via indirect means, the Legionella pneumophila effector SdhA reduces activation of AIM2 by maintaining the stability of the bacterial-containing vacuole thus reducing the escape of bacterial dsDNA into the host cytoplasm (Creasey and Isberg, 2012; Ge et al., 2012).

1.5.3 Host-protective properties of inflammasomes during bacterial infection

Inflammasomes still play a vital role in generating protective antimicrobial responses (von Moltke et al., 2013). The use of gene knockout mice has been paramount in determining the specific antimicrobial functions of the inflammasomes.

In the context of bacterial infection, loss of inflammasomes usually causes increased bacterial replication and host susceptibility, and decreased survival of mice (Broz et al., 2010a). Interestingly, inflammasome related protective phenotypes can often be linked to a specific inflammasome effector mechanism which can vary depending on the pathogen in question (von Moltke et al., 2013).

During STm infection, the majority of caspase-1-mediated protection can be attributed to IL-1β and IL-18 signalling as Il1b−/− and Il18−/− mice are more susceptible to oral challenge with STm compared to wildtype mice (Raupach et al., 2006). In agreement with this caspase-1 deficient mice are also more permissive to bacterial growth following oral infection in vivo (Broz et al., 2010a). The role that caspase-4 plays during in vivo STm infection is less clear and may vary depending on the site of infection. One study has shown that in the absence of caspase-1, caspase-4 increased susceptibility to infection and permitted increased bacterial growth in the spleen and liver (Broz et al., 2012). Other studies have reported that caspase-4 and non-canonical signalling play a protective role during STm infection in the gut (Knodler et al., 2014;

Sellin et al., 2014). Casp1−/− mice are also permissive to bacterial growth and have reduced survival during infection with Shigella flexneri (Sansonetti et al., 2000) and

Lm (Tsuji et al., 2004) both of which are at least partially dependent on IL-18 inflammatory responses. By contrast, resistance to Burkholderia thailandensis infection in mice is independent of both IL-1β or IL-18 but required caspase-4 (Aachoui et al., 2013).

Inflammasomes can also mediate enhanced cell-autonomous bacterial kill via methods that are independent of cytokine signalling. Caspase-4 has been shown to restrict the intracellular growth of L. pneumophila by promoting the fusion of LCVs with lysosomes (Akhter et al., 2012). Likewise, caspase-1 can also modulate bactericidal effects of lysosomes. During S. aureus infection caspase-1 enhances the acidification of phagosomes, thereby increasing bacterial killing (Sokolovska et al., 2013). This is achieved by cleavage and inactivation of the phagosome-associated NOX2 complex which neutralises the phagosome by producing reactive oxygen species (Sokolovska et al., 2013). Another more recent study showed that caspase-1 and caspase-4 restrict

50 the growth of cytosolic STm independently of either cytokine signalling and GSDMD- induced cell death, suggesting that additional caspase-1 substrates may have roles in controlling intracellular-bacterial growth (Thurston et al., 2016).

1.6 Role of inflammasomes in inheritable illnesses In previous sections, I discussed the intricacies of the regulatory mechanisms which underlie inflammasome activation which ensure they only become active under appropriate circumstances. Many naturally occurring single-nucleotide polymorphisms

(SNPs) exist within key inflammasome genes which cause aberrant activation of inflammasomes (Wilson and Cassel, 2010). The resulting uncontrolled activation of caspase-1 and release of pro-inflammatory cytokines are directly responsible for the pathologies of multiple inheritable autoinflammatory diseases.

Over 100 inflammatory disorder-related mutations have been reported within the NLRP3 gene (Sarrauste de Menthiere et al., 2003). The majority of these are gain- of-function mutations which result in constitutive activation of the NLRP3 inflammasome. This results in persistent release of pro-inflammatory cytokines IL-1β and IL-18 resulting in chronic inflammation which causes recurrent fevers, urticaria

(skin rashes) and joint and ocular inflammation (de Torre-Minguela et al., 2017). Such mutations are linked to three autosomal dominant inherited diseases; familial cold autoinflammatory syndrome (FCAS), Muckle-Wells syndrome (MWS), and neonatal- onset multisystem inflammatory disease (NOMID), which are collectively referred to as cryopyrin-associated periodic syndromes (CAPs) (Yu and Leslie, 2011). Gain-of- function SNPs are most commonly located within the NBD of NLRP3, R260W is one of the most frequently occurring mutations and causes milder symptoms associated with MWS (Neven et al., 2004). Other SNPs such as T348M and D303N, occur at a

51 lower frequency but are more commonly associated with the more severe pathological phenotypes of NOMID (Granel et al., 2003; Rosen-Wolff et al., 2003).

Familial Mediterranean fever (FMF) is another common periodic fever syndrome associated with PYRIN and causes recurring fever, dermal and rheumatoid inflammation (Alghamdi, 2017). There are over 300 different mutations associated with

FMF which can cause gain-of-function mutations in MEFV gene which encodes for

PYRIN (Sarrauste de Menthiere et al., 2003). M694V and V726A mutations being amongst the most common (Moradian et al., 2017) and have been shown to disrupt the interactions between PYRIN, PKN and 14-3-3 causing constitutive activation of the inflammasome (Park et al., 2016). Similar gain-of-function SNPs exist within NLRC4

(Duncan and Canna, 2018) which are linked to two illnesses; macrophage activation syndrome (MAS) (Canna et al., 2014) and infantile enterocolitis (Romberg et al.,

2014).

Despite the complexity surrounding inflammasome responses, the symptoms associated with these illnesses can be effectively treated with biopharmaceuticals.

Specifically, biologic therapies that block IL-1β signalling are extremely effective in treating these disorders (Yu and Leslie, 2011). These include Anakinra, a recombinant form of IL-1RA; Canakinumab, a neutralising monoclonal antibody to IL-1β and

Rilonacept, a recombinant fusion protein of IL-1R, IL-1RAP and IgG Fc subunit

(Hoffman and Broderick, 2016). These treatments are incredibly fast acting and can alleviate symptoms within hours of being administered (Hawkins et al., 2003; Sibley et al., 2015). However, these treatments are expensive and short lasting (Anakinra requires daily dosing), and though safe these treatments may also confer an enhanced risk to infection requiring frequent or prophylactic antibiotic treatment (Cabral et al.,

2016; Neven et al., 2010). Small molecule drugs which target inflammasome

52 signalling, such as the NLRP3 inhibitor MCC950 (Coll et al., 2015) are now being to be considered as potential, more cost-effective treatments.

1.7 Novel downstream targets of inflammasomes Caspase-1 is the key executioner of inflammasome-dependent immune responses. In order to understand caspase-1 effector mechanisms, we need to know the identity of its substrates and what effect caspase-1 cleavage has on that substrate

(Figure 1.6). In the study of apoptosis, multiple proteomics-based studies have led to significant advancements in the field through the identification and characterisation of many novel substrates (Dix et al., 2014; Mahrus et al., 2008; Van Damme et al., 2005).

Similar approaches have been used to identify caspase-1 substrates following inflammasome activation (Agard et al., 2010; Lamkanfi et al., 2008). Indeed, in recent years key downstream effector mechanisms of caspase-1 have been elucidated, in particular, the landmark discovery of GSDMD as the mediator of pyroptosis (He et al.,

2015; Kayagaki et al., 2015; Shi et al., 2015). However, despite this progress there are many features of caspase-1 activation which remain poorly understood.

For instance, caspase-1 can mediate the biogenesis of lipids in CHO and HeLa cells by activating sterol regulatory element-binding proteins (SREBPs) (Gurcel et al.,

2006). During infection, this process promotes cell survival and membrane repair in response bacterial pore-forming toxins (Gurcel et al., 2006) and while caspase-1 activity is required, the precise mechanism is unknown (Gurcel et al., 2006). Likewise, there are many inflammasome-regulated antimicrobial responses which require caspase-1 or -4 but are independent of known substrates such as the in vivo cytokine- independent, immune protection against Burkholderia (Aachoui et al., 2013) and cytokine and pyroptosis independent STm restriction (Thurston et al., 2016).

In addition, there are also substrates whose roles are not clearly defined. The glycolytic enzymes glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and aldolase which are processed by caspase-1 in STm infected macrophages (Shao et al., 2007). It is believed that cleavage of these proteins contribute to the reduction in glycolysis which is observed during inflammasome activation, however, this has yet to be thoroughly examined, and the physiological relevance remains to be determined.

The current list of caspase-1 substrates stands at ~100 proteins, however for many of them, the functional consequences of their cleavage are not known (Denes et al., 2012). Characterisation of these known substrates, as well as the identification

Figure 1.6 Known and novel functions of inflammasomes Inflammasome activation triggers many immune related functions via the catalytic activity of caspas-1. Many processes (inflammation, pyroptosis and phagosome acidification) are well defined in terms of their physiological relevance and the caspase-1 substrates which mediate them. Other known caspase-1 dependent process such (STm growth restriction and lipid biogenesis) are mediated by unknown substrates. Identification and characterisation of substrates (known or unknown) can help identify new caspase-1 mediated response. Adapted from Rathinam, V. A., & Fitzgerald, K. A. (2016)

54 of new caspase-1 targets, can help shed light on poorly understood inflammasome- mediated processes. In turn, this could advance our understanding of the interactions between inflammasomes and bacterial pathogens as well highlight new potential therapeutic targets for the treatment of inflammasome-associated autoimmune diseases. Using shotgun proteomics, the Shenoy group identified many putative targets and substrates of caspase-1 (discussed further in Chapter 3). Among these was a protein called ubiquitin-conjugating enzyme E2 L3 (UBE2L3, also known as

UBCH7). Ubiquitylation is involved in the regulation of many aspects of immune signalling including TLR signalling (Chuang and Ulevitch, 2004), NF-κB activation

(Iwai, 2010), autophagy/mitophagy (Kuang et al., 2013) and the regulation of inflammasomes (Rodgers et al., 2014). Targeting of UBE2L3 by caspase-1 could have the potential to impact a number of cellular processes. UBE2L3 was therefore selected as the key focus of these studies.

1.8 Enzymes of the ubiquitin cascade Ubiquitylation refers to the post-translational addition of the small protein modifier ubiquitin to a target protein. The attachment of ubiquitin to a substrate protein requires three enzymes: an ubiquitin-activating enzyme (E1), a ubiquitin-conjugating enzyme (E2) and ubiquitin-protein ligase (E3) (Ciechanover, 1994; Hershko and

Ciechanover, 1992, 1998) (Figure 1.7). The ubiquitylation cascade follows three distinct biochemical steps: (1) initiation by the E1 enzyme by activating a bound ubiquitin molecule. This involves the ATP-dependent adenylation of the ubiquitin C- terminal, and subsequent attachment to the active cysteine of the E1 enzyme by thioesterification (Ciechanover et al., 1981; Haas et al., 1983); (2) activated ubiquitin is transferred by transthiolation to the catalytic cysteine of an E2 enzyme (Hershko et al., 1983; Pickart and Rose, 1985); (3) either an E2 or E3 (the latter requiring additional

55 transthiolation transfer) mediates the formation of an isopeptide bond by catalysis of an aminolysis reaction between the ubiquitin C-terminal glycine and the ε-amino group of a substrate lysine residue (Hershko and Heller, 1985; Hershko et al., 1983).

Ubiquitin modifications can be attached in a variety of forms, most notably long polyubiquitin chains which have different signalling properties owing to which lysine residue is used to connect adjacent ubiquitin molecules.

The human genome encodes two ubiquitin-activating enzymes UBA1 and

UBA6 (Chiu et al., 2007; Jin et al., 2007a; Pelzer et al., 2007). Both enzymes are ubiquitously expressed across cellular tissues (Jin et al., 2007a) however, UBE1 is the principal enzyme as its expression is 10-fold greater than that of UBA6 in cells (Yang et al., 2013b). Functionally, UBA1 and UBA6 are both able to operate with a shared set of E2 conjugating enzymes which includes UBE2L3 and the UBE2D1-4 enzymes

(Jin et al., 2007a). However, they also have clear, and distinct specificities for other

E2s. For example, UBE2S and UBE2K which only function with UBA1, and UBE2Z

Figure 1.7: The ubiquitin transfer cascade Schematic of the sequential transfer of ubiquitin: (1) Ubiquitin activating enzyme (E1) mediated ATP-dependent adenylation of the ubiquitin C-terminal, and subsequent attachment to E1 the active cysteine by thioesterification; (2) transfer of ubiquitin by transthiolation to the catalytic cysteine of a ubiquitin conjugating enzyme (E2); (3) ligation of ubiquitin to substrate by ubiquitin ligase (E3).

(aka USE1) is specific to UBA6 (Jin et al., 2007a) and can conjugate both ubiquitin and the ubiquitin-like modifier FAT10 (Aichem et al., 2010; Chiu et al., 2007). These specificities create distinctive ubiquitinomes consisting of hundreds of specific targets

(Liu et al., 2017).

Humans have approximately 35 E2 ubiquitin-conjugating enzymes which are characterised by a conserved ubiquitin-conjugating catalytic (UBC) domain. In addition to the UBC, some E2 enzymes possess N- or C-terminal extensions which can add varying functional roles such as regulating enzymatic activity and interactions with E3 and E1 enzymes (Stewart et al., 2016). Once an E2 conjugating enzyme has received a ubiquitin molecule by transthiolation it is said to be “charged” (E2~Ub). Flexibility within the C-terminal residues of conjugated ubiquitin enables it to take on distinctive

“open” and “closed” conformations which alter its reactivity towards target residues

(Pruneda et al., 2011). Once charged, the E2 can be utilised by an E3 ligase to transfer the ubiquitin onto a substrate residue. In vitro analysis has shown that E2 enzymes possess weak intrinsic ubiquitin transfer activity, however, in cells this reactivity is significantly enhanced through the interaction with an E3 ligase (Wenzel et al., 2011).

Estimations of the number of E3 ligases present in mammalian genomes range from between approximately 700 to 1000 enzymes which fall into three families: really interesting new gene (RING); homologous to the E6AP carboxyl terminus (HECT) and

RING-in-between-RING (RBR) (Zheng and Shabek, 2017). The RING E3 ligases constitute the largest family (>600). These enzymes lack intrinsic reactivity towards lysine residues and require the activity of E2 enzymes to directly mediate ubiquitin transfer. This is achieved by promoting a “closed” E2~Ub conformation which is more favourable for aminolysis (Plechanovova et al., 2012; Pruneda et al., 2012) and

57 facilitating the interaction between a charged E2 and substrate onto which it may discharge ubiquitin (Metzger et al., 2014).

In the context of RING ligases, the E2 enzyme is thought to be the key determining factor of the type of ubiquitin linkage which is formed. Indeed, certain E2 conjugating enzymes have been described to only mediate the formation of a single type of linkage, such as UBE2S which specifically forms K11-linked ubiquitin chains

(Wickliffe et al., 2011). It has also been shown that specificity for a particular linkage requires ubiquitin “priming” by the initial ligation of a single monoubiquitin molecule which is then more easily extended with a homogeneously linked ubiquitin chain

(Christensen et al., 2007; Windheim et al., 2008; Ye and Rape, 2009). Other E2s such as the UBE2D family are more promiscuous in their choice of linkage type and have been shown to produce ubiquitin chains of varying linkage topologies (Kirkpatrick et al., 2006; Stewart et al., 2016).

By contrast, because RBR and HECT ligases possess their own aminolysis reactivity, the requirement for a closed E2~Ub state is not essential; indeed an “open” state is favoured as this promotes transthiolation from the E2 catalytic cysteine to E3 catalytic cysteine (Dove et al., 2017; Kamadurai et al., 2009; Lechtenberg et al., 2016).

Likewise, because the E3 themselves are responsible for transferring ubiquitin onto substrate proteins they have more control over chain linkage specificity. One example is the LUBAC enzyme HOIP, which orientates acceptor ubiquitin molecules in such a way that its N-terminal α-amino group is proximal to its active site favouring the formation of linear, M1 linked ubiquitin chains (Stieglitz et al., 2013).

1.9 UBE2L3: an atypical E2 conjugating enzyme PROTOMAP-based analysis identified the E2 conjugating enzyme UBE2L3 as a putative target of caspase-1 which depleted from the cytoplasm of inflammasome

58 activated macrophages. Structurally, UBE2L3 consists only of a UBC domain (Class

I) and does not possess N- or C-terminal extensions (Class II and III respectively) which have been implicated in facilitating additional regulatory and/or functional roles

(Figure 1.8) (Huang et al., 1999). In this regard UBE2L3, at first glance, seems like a typical E2 enzyme, however despite its characteristic structure UBE2L3 has several unique attributes.

UBE2L3 is one of approximately 35 E2 conjugating enzymes which are encoded within the human genome (Stewart et al., 2016). Because of the intermediary role E2 enzymes play during ubiquitylation it is often assumed that a large degree of functional redundancy exists between them, particularly within Class I enzymes.

Indeed, E3 ligases can function with multiple E2s in vitro and in vivo (Geisler et al.,

2014; Wenzel et al., 2011), for instance UBE2L3, UBE2D2 and UBE2D3 have been

Figure 1.8: Structure of UBE2L3 Crystal structure of human UBE2L3. Protein surfaces involved in E1 (purple) and E3 (blue) binding and the catalytic cleft (orange) are highlight. The catalytic cysteine 86 is shown. Residues in the α-helix and loop 4 and 7 are particularly important for E3 interactions. Ribbon representation of UBE2L3 PDB ID: 3SY2 (Lin et al., 2012)

59 shown to function redundantly with the E3 ligase PARKIN (Fiesel et al., 2014).

However, despite its typical structure UBE2L3 has been shown to be an essential gene as its deletion leads to embryonic lethality in mice (Harbers et al., 1996) and is thus far the only E2 enzyme shown to possess this quality. Interestingly UBE2L3 is also very well conserved amongst higher vertabrates, for example human UBE2L3 shares

100% amino acid sequence identity with distinct classes of animals including Mus musculus, Lonchura striata domestica (striated finch) and Anolis carolinensis (anole lizard), 97% identity with Xenopus tropicalis (western clawed frog) and 86% with

Cryptotermes secundus (drywood termite) (Figure 1.9 A + B). This coupled with the ubiquity of UBE2L3 expression across different tissues and its comparatively high abundance in cells points to UBE2L3 playing key house-keeping roles within the cell

(Clague et al., 2015).

Another defining feature of UBE2L3 is its lack of lysine reactivity, unlike all other

E2 enzymes UBE2L3 is unable to off-load charged ubiquitin to lysine residues, and can only discharge onto a cysteine (Wenzel et al., 2011). As a consequence, UBE2L3 is unable to function with RING E3 ligases (Moynihan et al., 1999; Zheng et al., 2000) and is therefore functionally limited to HECT and RBR families of E3 ligase (Wenzel et al., 2011). Sequence alignment analysis with UBE2D3 (which is reactive to both lysine and cysteine residues) revealed that the catalytic cleft of UBE2L3 has a proline

(Pro88) and a histidine (His119) which are absent in UBE2D3, which has aspartic acid residues at the same positions. Interestingly, the introduction of Pro88 and His119 into

UBE2D3 impaired its ability to transfer ubiquitin to lysine residues, however, introducing the reverse mutations (P88>D88 and H119>D119) to UBE2L3 did not confer lysine reactivity (Wenzel et al., 2011). Interestingly, despite being unable to

60 function RING-type E3 ligases UBE2L3 is still able to interact with them, in fact

UBE2L3 was often used in early structural studies and has been co-crystallised with several RING E3 ligases such as c-CBL (Zheng et al., 2000). The formation of these

A Homo sapiens Mus musculus

93 Bos taurus

Lonchura striata domestica

Anolis carolinensis

Xenopus tropicalis

Cryptotermes secundus B 0.020 Homo 1 MAASRRLMKELEEIRKCGMKNFRNIQVDEANLLTWQGLIVPDNPPYDKGAFRIEINFPAE Mus 1 MAASRRLMKELEEIRKCGMKNFRNIQVDEANLLTWQGLIVPDNPPYDKGAFRIEINFPAE Bos 1 MAASRRLMKELEEIRKCGMKNFRNIQVDEANLLTWQGLIVPDNPPYDKGAFRIEINFPAE Lonchura 1 MAASRRLMKELEEIRKCGMKNFRNIQVDEANLLTWQGLIVPDNPPYDKGAFRIEINFPAE Anolis 1 MAASRRLMKELEEIRKCGMKNFRNIQVDEANLLTWQGLIVPDNPPYDKGAFRIEINFPAE Xenopus 1 MAASRRLMKELEEIRKTGMKNFRNIQVEDSNLLTWQGLIVPDNPPYDKGAFRIEINFPAE Cryptotermes 1 MAATRRLQKELADIRTSGMKSFRDIQVDESNILTWQGLIVPDNPPYNKGAFRIEINFPAE

Catalytic Cys86 Homo 61 YPFKPPKITFKTKIYHPNIDEKGQVCLPVISAENWKPATKTDQVIQSLIALVNDPQPEHP Mus 61 YPFKPPKITFKTKIYHPNIDEKGQVCLPVISAENWKPATKTDQVIQSLIALVNDPQPEHP Bos 61 YPFKPPKITFKTKIYHPNIDEKGQVCLPVISAENWKPATKTDQVIQSLIALVNDPQPEHP Lonchura 61 YPFKPPKITFKTKIYHPNIDEKGQVCLPVISAENWKPATKTDQVIQSLIALVNDPQPEHP Anolis 61 YPFKPPKITFKTKIYHPNIDEKGQVCLPVISAENWKPATKTDQVIQSLIALVNDPQPEHP Xenopus 61 YPFKPPKITFKTKIYHPNIDEKGQVCLPVISAENWKPATKTDQVIQSLIALVNDPQPEHP Cryptotermes 61 YPFKPPKINFKTKIYHPNIDEKGQVCLPIISAENWKPATKTDQVIQALVALVNDPEPEHP

Homo 121 LRADLAEEYSKDRKKFCKNAEEFTKKYGEKRPVD Mus 121 LRADLAEEYSKDRKKFCKNAEEFTKKYGEKRPVD Bos 121 LRADLAEEYSKDRKKFCKNAEEFTKKYGEKRPVD Lonchura 121 LRADLAEEYSKDRKKFCKNAEEFTKKYGEKRPVD Anolis 121 LRADLAEEYSKDRKKFCKNAEEFTKKYGEKRPVD Xenopus 121 LRADLAEEYSKDRKKFCKNAEEFTKKYGEKRPVD Cryptotermes 121 LRADLAEEYLKDRKKFVKNAEEFTKKHSEKRPSD

Figure 1.9: Phylogenetic and sequence analysis of UBE2L3 protein sequence across species (A) Phylogenetic tree of UBE2L3 was generated using the Maximum Likelihood method based on the JTT matrix-based model. The tree with the highest log likelihood (-577.64) is shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 7 amino acid sequences. Evolutionary analyses were conducted in MEGA7. (B) Amino acid sequence alignment of UBE2L3 proteins from stated animal species. Identical residues are shown in red and the catalytic cysteine is highlighted.

61 non-functional complexes with RING E3 ligases has also been shown in cells by affinity-capture and two-hybrid methods . Whether such interactions serve some function, such as to negatively regulate RING E3 activity, remains to be examined.

Because UBE2L3 is functionally restricted to HECT and RBR ligases it is known to function in many key cellular pathways. Notably, UBE2L3 is the preferred E2 of

LUBAC and has a high affinity for HOIP (Martino et al., 2018) and plays a key role in

NF-kB activation (Lewis et al., 2015). Recent structural studies revealed that UBE2L3 has a unique propensity to interact with RBR ligases such as HOIP to take on distinct structural architectures owing to the identity of their bound E2 (Martino et al., 2018).

Human UBE2L3 has also been linked to multiple autoimmune diseases by genome-wide association studies (GWAS) (Franke et al., 2010; Fransen et al., 2010;

Han et al., 2009). In particular, SNPs which cause heightened expression of UBE2L3 have been linked to Crohn’s disease (Franke et al., 2010), rheumatoid arthritis (Stahl et al., 2010) and systemic lupus erythematosus (Orozco et al., 2011). UBE2L3 has been shown to be a binding partner for a number of bacterial effectors, most notably those which display E3 ligase activity, such as the STm SPI-1 effector SopA and EHEC effector NleL, whih were shown to bind and utilise UBE2L3 as their cognate E2 enzyme (Diao et al., 2008; Fiskin et al., 2017; Kamanova et al., 2016; Lin et al., 2011;

Sheng et al., 2017). Additionally, a Shigella effector called OspG has been shown to bind and inhibit UBE2L3 (Grishin et al., 2014; Kim et al., 2005; Pruneda et al., 2014;

Zhou et al., 2013).

1.10 Aims and Objectives Shotgun proteomics identified the E2 ubiquitin-conjugating enzyme UBE2L3 as a novel target of caspase-1. Previous studies suggest the UBE2L3 is central to many key immune signalling processes and given the functional specificity of UBE2L3 for

HECT and RBR E3 ligases, its targeting by caspase-1 has the potential to impact many cellular functions. In my PhD, I decided to further characterise the targeting of

UBE2L3 by inflammasomes. The specific objectives therefore were:

1. Characterise UBE2L3 targeting in the wider context of inflammasome

signalling by various agonists and bacterial infectious agents.

2. Determine the mechanisms of UBE2L3 targeting by inflammasomes.

3. Elucidate the role of UBE2L3 in inflammasome and caspase-1 mediated

responses.

4. Better characterise UBE2L3 interactions with bacterial effectors in the

context of its role in immune-signalling.

Chapter 2 – Materials and Methods

2.1 Eukaryotic and bacterial cell growth conditions 2.1.1 Eukaryotic cell culture

Mouse macrophage cell lines were isolated from Casp1/4-/- and Nlrp3-/- mice

(Sutterwala et al., 2006) provided by Richard Flavell (Yale University and Millenium

Pharmaceuticals Inc.), Asc-/- and Nlrc4-/- mice (Mariathasan et al., 2004) provided by

Vishwa Dixit (Genentech), and Casp4-/- mice (Wang et al.) provided by Junying Yuan

(Harvard University, USA). Il1r1-/- and Il18r1-/- cells were obtained from Jackson labs, and iGsdmd-/- cells (Shi et al., 2015) were provided by Feng Shao (National Institute of Biological Sciences, Beijing, China). Casp1/4-/- cells stably expressing Flag-tagged mouse caspase-4 (Thurston et al., 2016) were provided Teresa Thurston (Imperial

College London, UK).

High-glucose Dulbecco's Modified Eagle's Medium (DMEM) supplemented with

1 mM sodium pyruvate, 10% heat-inactivated fetal calf serum (HI-FCS) is called complete DMEM henceforth, and DMEM-PS when also containing penicillin- streptomycin. Immortalized bone-marrow derived macrophages (iBMDMs) (wild-type cells were from the C57Bl/6N strain of mice) were routinely cultured in DMEM-PS containing 20% L929 culture spent supernatant (which contained MCSF). L929-spent medium was not included when seeding cells prior to experimentation. Cells were plated 24 h prior to treatment. Primary bone-marrow derived macrophages were grown in the same medium as above but differentiated in non-tissue culture treated 10 cm

Petri-plates and used within 7-15 days. Primary bone-marrow derived dendritic cells

(BMDCs) were prepared in Roswell Park Memorial Institute medium (RPMI) supplemented with 10 mM HEPES, 10 mM sodium pyruvate and 10% HI-FCS and penicillin-streptomycin (RPMI-PS) supplemented with GM-CSF (a gift from Gyorgy

Fejer, University of Plymouth). For detachment and plating, primary BMDMs and

BMDCs were incubated ice-cold PBS containing 2 mM EDTA for 10-15 mins at 4°C, centrifuged at 100 xg, resuspended in relevant medium and plated for experimentation.

Human THP-1 cells were routinely cultured in Roswell Park Memorial Institute medium (RPMI) supplemented with 10 mM HEPES, 10 mM sodium pyruvate and 10%

HI-FCS and penicillin-streptomycin (RPMI-PS). Cells were differentiated for 2 days in

100 ng.mL-1 phorbol myristoyl acetate (PMA). Cells were incubated in fresh medium without PMA 24 h prior to treatments.

HeLa, HEK293E, and HEK293T were routinely cultured culture in DMEM-PS.

Retrovirally transduced cell lines stably expressing plasmids which contain puromycin resistance were routinely cultured in puromycin; either 2 μg.mL-1 (THP1, HeLa,

HEK293T, HEK293E) or -6-8 μg.mL-1 (iBMDMs). Puromycin was not included when cells were seeded for experiments. For use in bacterial infection studies, cells were seeded in antibiotic-free medium.

2.1.2 Bacterial growth conditions for infection

Prior to infection with either STm or Lm for inflammasome activation, cells were primed with 250 ng/mL-1 LPS for 3 h in serum-containing medium. Following the addition of bacteria, infections were synchronised by centrifugation at 750 xg for 10 mins. For all infections, extracellular bacteria were killed by the addition of 100 µg.mL-

1 gentamycin 30 min post-infection (1 h in THP-1 infections) and infections allowed to proceed for the stated lengths of time.

Salmonella enterica Typhimurium (STm) NTCC 12023 (ATCC 14028) wildtype and ΔprgH strains (from David Holden (Beuzon et al., 1999)) were routinely maintained in Luria-Bertani (LB) broth or LB agar. For infection and inflammasome activation, STm were grown to induce high expression of the SPI-1 T3SS. For this, STm were grown

66 overnight at 37°C without shaking in LB containing 300 mM NaCl (LB-salt). The following day, overnight cultures were diluted 1:50 in fresh, pre-warmed LB-salt medium and grown with shaking until an to an optical density measured at a

wavelength of 600 nm (OD600) of 0.9-1.2 was reached. Bacterial density in colony forming units per mL (CFU/mL) was estimated by measuring OD600 (OD600 of 0.1 =

~1.2x108 CFU/mL) and bacteria were washed twice in sterile PBS and diluted accordingly for the required multiplicity of infection (MOI) in OptiMEM. After the addition of 100 µg.mL-1 gentamycin infections were allowed to proceed for a further 2-

3 hours.

Listeria monocytogenes (Lm) 14043S and Δhly (from Dan Portnoy(Jones and

Portnoy, 1994)) were routinely maintained in Brain Heart Infusion (BHI) broth or agar.

For infection, Lm were grown overnight in BHI medium at 37°C with shaking. Prior to infection, bacteria were washed three times with sterile PBS to remove excess listeriolysin toxin and other secreted proteins. Bacterial CFU/mL was estimated by

8 OD600 (OD600 of 0.1 = ~1x10 CFU/mL) and bacteria were diluted for infections to the required MOI indicated which are indicated in figure legends. After the addition of 100

µg.mL-1 gentamycin infections were allowed to proceed for a further 1-2 hours.

For priming experiments with STm ΔprgH and Lm Δhly, host cells were infected at an MOI of 5. After the addition of 100 µg.mL-1 gentamycin infections were allowed to proceed for increasing lengths of time for up to 24 hours.

For priming infections with commensal bacteria, E. coli ATCC 11775 was grown in LB overnight at 37°C with shaking, Bacillus subtilis strain 168 (from Angelika

Grundling) was grown overnight at 37°C in 2x YT broth with shaking and

Streptococcus gordonii strain Challis (from Andrew Edwards) was grown overnight in

Tryptic Soy Broth in a humidified CO2 incubator at 37 °C and without shaking. Bacteria

67 were washed once in PBS and cells were infected at an MOI of 5 for a total of 18 hours. At 1 h post-infection, 100 µg.mL-1 gentamycin was added to all wells and incubated for 1 h after which cells were washed once in warm DMEM and medium replaced with complete DMEM containing 50 µg.mL-1 gentamycin for the remaining 16 hours.

Escherichia coli O157:H7 EDL933 (EHEC, from Gad Frankel (Berger et al.,

2012)) infection of HeLa cells were performed as previously described (Sheng et al.,

2017) with minor alterations. Briefly, EHEC strains were grown overnight at 37°C in 3 mL 2x YT medium without shaking. The following day, expression of the type III secretion system was induced by sub-culturing (1:40) in 5 mL pre-warmed low-glucose serum-free DMEM medium. Inducing bacteria were incubated without shaking for 4 h at 37°C in 5% CO2. Bacterial cell density was estimated by measuring culture optical density at a wavelength of 600 nm (OD600) and added to HeLa cells at an MOI of 10.

Infections proceed for 2.5 hours, after which cells were washed 3 times in 500 μL warm serum-free DMEM and used for immunofluorescence analysis or bacterial attachment assays by CFU.

For attachment CFU experiments, infections were performed in triplicate and cells were infected at an MOI of 5. After 2.5 h of infection and 3x washes in serum- free DMEM, cells were washed once in warm sterile PBS and lysed using 100µl of sterile PBS containing 0.5% Triton-X100 incubated for 5 min. Lysates were diluted 1:5- fold serially (1/625, 1/3125 and 1/15625) which were plated for CFU counts. Each dilution was drop plated (10µL) onto LB agar in triplicate. Plates were incubated at

37°C overnight and colonies were enumerated and used to calculate CFU counts which are finally expressed as CFU/mL. Uninfected wells underwent the same washes, medium changes, and incubations in antibiotic as infected wells.

2.2 Bacterial strains, plasmids, and primers 2.2.1 Bacterial strains

Strain Description Source MAX Efficiency™ Stbl2™ High-efficiency chemically competent cells for cloning. Contains recA1 and other genetic markers which reduce Thermo Fisher Competent E. coli recombination of retroviral sequences and tandem repeat genes. This strain was used for all molecular cloning. BL21-CodonPlus (DE3)- Contain additional copies of E. coli argU, ileY, leuW, proL tRNA genes correcting for codon bias and improving expression of Agilent Technologies RIPL E. coli sequences from other organisms. This strain was used for expression of recombinant protein. Salmonella enterica Wildtype Salmonella enterica Typhimurium 12023. NTCC, David Holden, Imperial Typhimurium NTCC 12023 College London Salmonella enterica Salmonella enterica Typhimurium 12023 lacking ring component of the SPI-1 type III secretion system. SPI-1 mutant which (Beuzon et al., 1999) David Holden, Typhimurium NTCC 12023 fails to activate the inflammasome. Imperial College London ΔprgH Listeria monocytogenes Wildtype Listeria monocytogenes 10403S. Angelika Grundling, Imperial 10403S College London Listeria monocytogenes Listeria monocytogenes 10403S lacking hly gene which encodes listeriolysin O pore-forming toxin. Angelika Grundling, Imperial 10403S Δhly Infection mutant which fails to activate inflammasomes. College London Escherichia coli (ATCC® Non-pathogenic strain of Escherichia coli. ATCC 11775™) Bacillus subtilis strain 168 Wildtype Bacillus subtilis strain 168, a non-pathogenic soil bacterium. (Jones and Portnoy, 1994) Angelika Grundling, Imperial College London Streptococcus gordonii Wildtype Streptococcus gordonii strain Challis, non-pathogenic, commensal bacterium. Andrew Edwards, Imperial College strain Challis London Escherichia coli O157:H7 Wildtype pathogenic strain of enterohemorrhagic E. coli (EHEC) (Berger et al., 2012) EDL933 Gad Frankel, Imperial College London Escherichia coli O157:H7 EHEC expressing λ-red recombinase plasmid pKD46 This Study EDL933 pKD46 Escherichia coli O157:H7 EHEC with a chromosomal deletion of nleL E3 ligase effector. nleL was deleted by λ-red mutagenesis and replaced with This Study EDL933 ΔnleL #3 pKD4 kanamycin resistance cassette, flanked by FRT sequences.

Table 2.1. Bacterial strains

2.2.2 Bacterial and eukaryotic expression plasmids

Sequence insert Vector Description Tag Source GST pGEX-6P1 For cloning and expression of recombinant protein. Target gene is N-terminally tagged with GST linked by a N-term, GST Avinash Shenoy rhinovirus C3 protease site. Target gene is expressed from a tac promoter. Expression is inducible with IPTG. UBE2L3 pGEX-6P1 Expression of recombinant UBE2L3WT N-term, GST Arno Alpi, University of Dundee UBE2L3D124N pGEX-6P1 Expression of recombinant UBE2L3D124N Caspase-1 uncleavable UBE2L3, mutation introduced by single N-term, GST This study oligonucleotide-based linear PCR UBE2L3K20 only pGEX-6P1 Expression of recombinant UBE2L3K20 only UBE2L3 with single Lys20, mutation introduced by single N-term, GST This study oligonucleotide-based linear PCR on UBE2L318R Mohini Kalyan UBE2L3K64 only pGEX-6P1 Expression of recombinant UBE2L3K64 only UBE2L3 with single Lys64, mutation introduced by single N-term, GST This study oligonucleotide-based linear PCR on UBE2L318R Mohini Kalyan UBE2L3K73 only pGEX-6P1 Expression of recombinant UBE2L3K73 only UBE2L3 with single Lys73, mutation introduced by single N-term, GST This study oligonucleotide-based linear PCR on UBE2L318R Mohini Kalyan UBE2L3K82 only pGEX-6P1 Expression of recombinant UBE2L3K82 only UBE2L3 with single Lys82, mutation introduced by single N-term, GST This study oligonucleotide-based linear PCR on UBE2L318R Mohini Kalyan UBE2L3K131 only pGEX-6P1 Expression of recombinant UBE2L3K131 only UBE2L3 with single Lys131, mutation introduced by single N-term, GST This study oligonucleotide-based linear PCR on UBE2L318R Mohini Kalyan UBE2L3K138 only pGEX-6P1 Expression of recombinant UBE2L3K138 only UBE2L3 with single Lys138, mutation introduced by single N-term, GST This study oligonucleotide-based linear PCR on UBE2L318R Mohini Kalyan UBE2L3K150 only pGEX-6P1 Expression of recombinant UBE2L3K150 only UBE2L3 with single Lys150, mutation introduced by single N-term, GST This study oligonucleotide-based linear PCR on UBE2L318R Mohini Kalyan UBE2L3K64R pGEX-6P1 Expression of recombinant UBE2L3K64R UBE2L3 with Lys64 to Arg64 mutation, mutation introduced by N-term, GST This study single oligonucleotide-based linear PCR on UBE2L3WT UBE2L3K82R pGEX-6P1 Expression of recombinant UBE2L3K82R UBE2L3 with Lys82 to Arg82 mutation, mutation introduced by N-term, GST This study single oligonucleotide-based linear PCR on UBE2L3WT UBE2L3K131R pGEX-6P1 Expression of recombinant UBE2L3K131R UBE2L3 with Lys131 to Arg131 mutation, mutation introduced by N-term, GST This study single oligonucleotide-based linear PCR on UBE2L3WT UBE2L3K138R pGEX-6P1 Expression of recombinant UBE2L3K138R UBE2L3 with Lys138 to Arg138 mutation, mutation introduced by N-term, GST This study single oligonucleotide-based linear PCR on UBE2L3WT

UBE2L3K150R pGEX-6P1 Expression of recombinant UBE2L3K150R UBE2L3 with Lys150 to Arg150 mutation, mutation introduced by N-term, GST This study single oligonucleotide-based linear PCR on UBE2L3WT UBE2L39R pGEX-6P1 Expression of recombinant UBE2L39R UBE2L3 with K9R, K20R, K64R, K67R, K73R, K82R, K100R, N-term, GST This study K131R, K145R mutations, cloned from pMAT UBE2L39R UBE2L318R pGEX-6P1 Expression of recombinant UBE2L318R UBE2L3 with all Lys residues mutated to Arg, cloned from N-term, GST This study pMAT UBE2L318R UBE2L3C86A pGEX-6P1 Expression of recombinant UBE2L3C86A Inactive UBE2L3 with catalytic Cys82 mutated to Ala, mutation N-term, GST This study introduced by single oligonucleotide-based linear PCR on UBE2L3WT UBE2L3P62A/F63A pGEX-6P1 Expression of recombinant UBE2L3P62A/F63A E3 binding mutant UBE2L3 with Pro62 and Phe63 mutated to Ala, N-term, GST This study mutation introduced by single oligonucleotide-based linear PCR on UBE2L3WT NleLFL pGEX-6P1 Expression of recombinant NleLFL Full-length wildtype NleL, cloned from genomic DNA derived from wildtype N-term, GST This study EHEC O157:H7 strain EDL933 NleL170-782 pGEX-6P1 Expression of recombinant NleL170-782 Truncated wildtype NleL residues 170-782, cloned from genomic N-term, GST This study DNA derived from wildtype EHEC O157:H7 strain EDL934 SopA163-783 pGEX-6P1 Expression of recombinant SopA163-783 Truncated wildtype SopA residues 163-783, cloned from N-term, GST This study genomic DNA derived from wildtype Stm 12023 UBE1 pET21D Expression of recombinant human UBE1. Target gene has N-terminal hexahistidine tag and is IPTG inducible. N-term, 6x His Cynthia Wolberger (Addgene plasmid # 34965) CAS1 p10 pProExHT Expression of recombinant human caspase-1 p10 (residues 317-404). Target gene has N-terminal N-term, 6x His Avinash Shenoy hexahistidine tag and is IPTG inducible CAS1 p20 pProExHT Expression of recombinant human caspase-1 p20 (residues 120-297). Target gene has N-terminal N-term, 6x His Avinash Shenoy hexahistidine tag and is IPTG inducible hIL-1β110-269 pGEX-6P1 Expression of recombinant human IL-1β (amino acids 110-269). N-term, GST Avinash Shenoy

Table 2.2. Plasmids for the expression of recombinant proteins in E. coli

Sequence insert Vector Description Tag Source Empty plasmid pMXsIP Retroviral plasmid used for the stable introduction of genes by retroviral transduction. N/A Lab stock YFP-Ctrl-miR30E pMxCMV CMV promoter and enhanced YFP form pEYFP-C1 (Clontech) with miRNA30E backbone cloned into pMXsIP. for N/A This study subsequent generation of N-terminally YFP-tagged proteins and used as a control for shRNA-based silencing. Julia Sanchez YFP-UBE2L3 pMxCMV Wildtype UBE2L3 cloned from pGEX-UBE2L3 into YFP-pMxCMV generating N-terminally YFP-tagged UBE2L3WT. N-term, YFP This study

YFP-stop pMxCMV Expression of YFP under CMV promotor. YFP PCR amplified from YFP-Ctrl-miR30E to cloned into pMxCMV N/A This study Julia Sanchez YFP-UBE2L3C86S pMxCMV UBE2L3 with catalytic Cys86 mutated to a serine, the mutation was introduced by single oligonucleotide-based N-term, YFP This study linear PCR on YFP-pMxCMV UBE2L3WT. 3F-UBE2L3-HA pMPP 3xFLAG at N-terminus (from p3Tag1 vector, Agilent) was introduced by SLIC based cloning. Wildtype UBE2L3 was N-term 3x This study then 2xHA at C-terminus (YPYDVPDYA introduced via two PCRs) were added to generate pMXCMV-FLAG FLAG. C-term, Julia Sanchez UBE2L3-2HA plasmid. The PKG from pRetro plasmid was subsequently introduced to yield pMPP. 2x HA 3F-UBE2L3-Step pMPP FLAG-UBE2L3 was PCR amplified to introduce single C-terminal StrepII tag (WSHPQFEK) N-term 3x This study FLAG. C-term, 1x StrepII 3F-UBE2L3C86AStrep pMPP Catalytically inactive UBE2L3 with catalytic Cys86 mutated to alanine, the mutation was introduced by single N-term 3x This study oligonucleotide-based linear PCR on 3F-UBE2L3-Step. FLAG. C-term, 1x StrepII UBE2L39R pMA-T The synthetic gene UBE2L3-9KR was assembled from synthetic oligonucleotides N/A Invitrogen and/or PCR products. The fragment was inserted into pMA-T. Made by Invitrogen GeneArt™ gene synthesis. GeneArt™ UBE2L318R pMA-T The synthetic gene UBE2L3-18KR was assembled from synthetic oligonucleotides N/A Invitrogen and/or PCR products. The fragment was inserted into pMA-T. Made by Invitrogen GeneArt™ gene synthesis. GeneArt™ UBE2L314R pMX The synthetic gene 3F-UBE2L3-14R-2Strepwas assembled from synthetic oligonucleotides N-term 3x Invitrogen and/or PCR products. The fragment was inserted into pMX. Made by Invitrogen GeneArt™ gene synthesis. FLAG. C-term, GeneArt™ 2x StrepII 3F-UBE2L39RStrep pMPP Expression of tagged UBE2L39R in eukaryotic cells – UBE2L3 with K9R, K20R, K64R, K67R, K73R, K82R, N-term 3x This study K100R, K131R, K145R mutations, cloned from pMA-T UBE2L39R FLAG. C-term, 2x StrepII 3F-UBE2L318RStrep pMPP Expression of tagged UBE2L318R in eukaryotic cells – UBE2L3 with all Lys residues mutated to Arg, cloned from N-term 3x This study pMA-T UBE2L318R FLAG. C-term, 1x StrepII

3F-UBE2L314RStrep pMPP Expression of tagged UBE2L314R in eukaryotic cells – UBE2L3 with 14 Lys residues mutated to Arg, cloned from N-term 3x This study pMX F-UBE2L314R-Strep FLAG. C-term, 1x StrepII 3F-UBE2L3 pMPP Expression of tagged UBE2L3 P62A/F63A in eukaryotic cells – cloned from pGEX UBE2L3 P62A/F63A N-term 3x This study P62A/F63AStrep FLAG. C-term, 1x StrepII 3F-UBE2L3 pMPP Expression of tagged UBE2L3 18R/C86A in eukaryotic cells N-term 3x This study 18R/C86AStrep FLAG. C-term, 1x StrepII UBE2L3-YFP pMxCMV Eukaryotic expression of UBE2L3-YFP. UBE2L3 was PCR amplified from YFP-UBE2L3 to remove the stop codon C-term, YFP Julia Sanchez and cloned upstream of YFP in YFP-stop to generate C-terminally tagged UBE2L3-YFP. UBE2L3miR30E1 (Spr) pMxCMV Expression of antisense 22- mer sequences for silencing of UBE2L3 expression N/A This study 5’ TTTCTTTGTAAACTCTTCAGCA 3’ UBE2L3miR30E2 (Vys) pMxCMV Expression of antisense 22- mer sequences for silencing of UBE2L3 expression N/A This study 5’ TTTCATCCCACATTTGCGGATT 3’ NLRP3miR30E pMxCMV Expression of antisense 22- mer sequences for silencing of human NLRP3. 5’AAATTGCGACTCCTGAGTCTCT3’ N/A This study (adapted from TRCN0000431574) Julia Sanchez GSDMDmiR30E pMxCMV Expression of antisense 22- mer sequences for silencing of human GSDMD. 5’CAGCACCTCAATGAATGTGTA3’ N/A This study (TRCN0000179101). Julia Sanchez Atg7 shRNA pLKO.1 Expression of Atg7 targeting shRNA (TRCN0000092167) N/A Teresa Thurston

LacZmiR30E pLXCMV Used as a non-targeting control for Atg7 shRNA N/A Teresa Thurston

mCas1 All A pMxCMV Caspase-1(p45) with D296N, D308N, D313A and D314A mutations to abolish self-cleavage N/A This study Julia Sanchez mCas1 C284A pMxCMV Caspase-1(p45) with C284A inactivating mutation N/A This study Julia Sanchez

Table 2.3. Plasmids for expression of proteins and genetic constructs in eukaryotic cells

2.2.4 Primers used during this study

Primer Sequence 5’ to 3’ M13F_seq TGTAAAACGACGGCCAGT M13R_seq CAGGAAACAGCTATGAC pGEX6P_seqF GGGCTGGCAAGCCACGTTTGGTG pGEX6P_seqR CACCGAAACGCGCGAGGCAGATC pMXs-seqF GCTTGGATACACGCCGC pMXs-seqR GAGGCCTGCAGGAATTCC CMV_seqF CGCAAATGGGCGGTAGGCGTG T7pro_seqF TAATACGACTCACTATAGGG T7term_seqR GCTAGTTATTGCTCAGCGG NleL_seqF TCTAATTTCAGCAATTCC Table 2.4. Sequencing primers

Primer Sequence 5’ to 3’

Primer_NleL_C753A_PstI-fwd ACAAGGAAGAAATAATGTcTTcACTgctActgcagTGCTGACTGATATGCTAA Primer_UBE2L3_P62F63_AA_BlpI CCAGCAGAGTACgCAgctAAgCCACCGAAGATC -fwd Primer_hUBE2L3_D124N_BbvCI- CCAGCCTGAGCACCCGCTgaGGGCTaACCTAGCTGAAGAATACTCTAAGG fwd Primer_UBE2L3_C86A_PsiI-fwd CGACGAAAAGGGGCAGGTCgctCTGCCAGTtATaAGTGCCGAAAACTGGAAGC

Primer_UBE2L3_C86S_HindIII- CATCGACGAAAAGGGGCAGGTaagcttGCCAGTAATTAGTGCCGAAAA fwd Primer_UBE2L3_K20only_ApoI- ATGTGGGATGAaGAAtTTCCGTAACATCC fwd Primer_UBE2L3_K82only_ClaI CCCAAACATCGAtGAAAagGGGCAGGTCTGT

Primer_UBE2L3_K73only_BstBI AGGATCACATTTcGAACAAaGATCTATCACCCA

Primer_UBE2L3_K131only_-XbaI- AAGAATACTCTAaAGACCGTAGGAG fwd Primer_UBE2L3_K138only_ApoI GAGATTCTGTAaGAATGCTGAAGAaTTTACAAGGAG

Primer_UBE2L3-9KR-XhoI-6P1-rv gtcacgatgcggccgctcgagttaGTCCACAGGTCGCTTTTCCC

Primer_UBE2L318R_C86A_PsiI- CGACGAAAGAGGGCAGGTCgctCTGCCAGTtATaAGTGCCGAAAACTGGAGAC fwd Primer_UBE2L3_K82R-ClaI-fwd CCCAAACATCGAtGAAagaGGGCAGGTCTGT

Primer_UBE2L3_K64R-6P1-R atcttcggtggtCTGAATGGGTACTCTGCTGGAAAGT

Primer_UBE2L3_K131R-XbaI-rvs tacagaattttttacggtctCTAGAGTATTCTTCAGCTAGGTCAG

Primer_UBE2L3-K138R-SphI-fwd gaccgtaaaaaattctgtaCGcATGCTGAAGAGTTTACAAAG

Primer_UBE2L3-18KR-6P1Xho-R gtcacgatgcggccgctcgagttaGTCCACAGGTCGCCTTTCCC

Table 2.5. Mutagenic primers – Mismatch mutations are presented in lowercase lettering.

Primer Sequence 5’ to 3’

Primer_UBE2L3_MxYBam-slic-F GTCGACGGTACCGCGGGCCCGatggcggccagcaggagg Primer_UBE2L3_MxYBam-slic-R GTACCACCACACTGGGATCCttagtccacaggtcgcttttc

Primer_3xF_UBE-fwd gataaagcccgggcggacgcgtcgATGGCGGCCAGCAGGAGG

Primer_2xHA_UBE-rvs aacatcgtatgggtacgcataatctggaacatcgtatgggtacgcgtaGTCCA CAGGTCGCTTTTCCC

Primer_UBE2L3_Strep-EcoRI_rvs gaggcctgcaggaattcatttttcgaattgaggatgactccaacgcgtGTCCA CAGGTCGCTTTTCCC

Primer_UBE2L3_NT-Mfe_Kz-fwd agcgctaccggtcgccaccaattgccaccATGGCGGCCAGCAGGAGG

Primer_UBE2L3-ΔStop_Mfe-rvs tgctcaccatggtggcaatGTCCACAGGTCGCTTTTCCC

Primer_UBE2L3-18KR_Strep-rvs ttgaggatgactccaacgcgtGTCCACAGGTCGCCTTTCCC

Primer_UBE2L3-9KR_Strep- ttgaggatgactccaacgcgtGTCCACAGGTCGCTTTTCCC EcoRI_rvs Primer_UBE2L3-BamHI-6P1-F ttccaggggcccctgggatccATGGCGGCCAGCAGGAGG

Primer_UBE2L3-9KR-XhoI-6P1-R gtcacgatgcggccgctcgagttaGTCCACAGGTCGCTTTTCCC

Primer_UBE2L3-18KR-6P1Xho-R gtcacgatgcggccgctcgagttaGTCCACAGGTCGCCTTTCCC

Primer_NleL_FL-Bam-6P1-fwd ctgttccaggggcccctgATGCTGCCCACTACAAATAT

Primer_NleL-782stop-Xho-6P1-rvs gtcacgatgcggccgctcgagttaTCAACGCCACGCAACAGGATAATA

Primer_NleL_S170-Bam-6P1-fwd ctgttccaggggcccctgAGTCAAGGCCGTGCCTGCCT

Primer_SopA-782stop-Xho-6P1-rvs gtcacgatgcggccgctcgagttaCGCCCAGGCCAGTGGCAGGA

Primer_SopA-163_BamHI-6P1-fwd ctgttccaggggcccctgGCTACCTCTTCCCCGTCATC Table 2.6 Primers used for cloning – Uppercase letters are homologous to target sequence, lowercase letters are homologous to plasmid sequence, epitope tag sequences are underlined.

Primer Sequence 5’ to 3’

Lambda-P1-fwd TGTAGGCTGGAGCTGCTTCG

Lambda-P2-rvs CATATGAATATCCTCCTTAG

NleL_KO_pkd4_50mer_Fw tgtgttttctaattttactttagtggactaagtaaaaaggagtgagataaTGTAGGCTG GAGCTGCTTCG

NleL_KO_pkd4_50mer_Rv gcaaacctaaaaagacagtgttacaataatattcatccatggtctctaaaATATGAATA TCCTCCTTAG Table 2.7 Primers used in the generation of ΔnleL EHEC – Lowercase underlined sequence is homologous to 50 bps directly upstream and downstream of nleL in EDL993

Target Forward primer Reverse Primer hGAPDH TGCCATCAATGACCCCTTC CTGGAAGATGGTGATGGGATT hIL1B GACAAAATACCTGTGGCCTTG AGACAAATCGCTTTTCCATCTTC hTNF ACTTTGGAGTGATCGGCC GCTTGAGGGTTTGCTACAAC mGapdh AATGGTGAAGGTCGGTGTG GTGGAGTCATACTGGAACAT mIl1b CTACCTGTGTCTTTCCCGTG TGCAGTTGTCTAATGGGAACG mTnf AGACCCTCACACTCAGATCA TGTCTTTGAGATCCATGCCG mIl6 CAAAGCCAGAGTCCTTCAGAG GTCCTTAGCCACTCCTTCTG

Table 2.8. Primers used in qPCR analyses

2.3 Molecular Biology techniques 2.3.1 Preparation of chemically component bacteria and heat shock transformation.

E. coli (Stbl2™ or BL21-CodonPlus (DE3)-RIPL) were grown overnight at 37°C in (LB) broth with shaking. The following day bacteria were subcultured (1:100 dilution)

into 250 mL of fresh LB. Cells were grown at 37°C to an optical density OD600 of ~0.4

– 0.5 at which point the culture was chilled on ice for 20 min. Once chilled the bacterial culture was aliquoted into pre-cooled 50 mL centrifuge tubes and centrifuged for 5 min at 4,000 xg at 4°C The cleared medium was removed, and the resulting bacterial pellets were resuspended in a final volume of 2 mL of sterile transformation and storage solution (TSS) containing 10% PEG 4000, 5% DMSO, 5 mM MgCl2 and 2%

LB broth. Chemically competent bacteria were kept as 75 µL aliquots, flash frozen with liquid nitrogen and stored at -80°C.

For transformation, chemically competent bacteria were thawed on ice and incubated with ~5 μL of plasmid for 30 min. Bacteria were then heat-shocked at 42°C for 45 seconds and incubated on ice for a further 2 min. Fresh pre-warmed LB was

76 then added (300 μL) and transformed bacteria were grown for 1 h at 37°C followed by overnight growth on LB agar plates with relevant selective antibiotics.

2.3.2 Preparation of electrocompetent cells and transformation by electroporation

Bacteria to be transformed were grown overnight at 37°C in LB broth with shaking. The following day the bacteria were sub-cultured into 5 mL of fresh LB broth and grown with shaking at 37 °C to an OD600 of 0.6-0.7. Bacterial cultures were then chilled on ice for 30 min. Cultures were subsequently centrifugation at 1,500 xg for 10 min at 4°C. All subsequent steps were also carried out at 4 °C or on ice. The culture medium was decanted, and the bacterial pellet was washed by being resuspended in

1 mL ice-cold, sterile 15% glycerol and centrifuged again at 1500 x g for 10 min. The wash steps were repeated twice more, and the bacterial pellet was resuspended in a final volume of 100 μL 15% glycerol. For electroporation 100 ng of plasmid (maximum

5 µL) was added to 50 µL electrocompetent cells and incubated on ice for 2 min. Cells were transferred to a pre-chilled 0.2 cm electroporation cuvette and pulsed using a

BioRad MicroPulser™ electroporator using the EC2 (2.5 kV) program. Immediately after electroporation, 400 µL of pre-warmed LB broth was added to the bacteria and the transformed culture was incubated at 37 °C for 1 h with shaking. Positive transformants were selected by overnight growth at 37 °C on LB agar plates containing relevant antibiotics.

2.3.3 Cloning and site-directed mutagenesis

Routine cloning used the Phusion polymerase for PCR followed by sequence and ligation independent cloning (SLIC) as previously described (Jeong et al., 2012).

Plasmids were linearized by restriction enzyme digestion of the appropriate site(s).

Genes of interest were amplified by PCR using primers which had 5’ extensions (at

77 least 18 base pairs) identical to the digested ends of the linearized vector. Digested plasmids and PCR amplicons were purified by gel extraction prior to further use.

Vector (100 ng) and insert were mixed at a 1:4 molar ratio and incubated with

T4 polymerase (NEB) for 2.5 min to generate compatible 5' overhangs via its 3' to 5' exonuclease activity. Reactions were stopped by incubating on ice for 10 min and subsequently transformed into chemically competent E. coli Stbl2 by the heat shock method described in section 2.3.1. Site-directed mutagenesis was carried out using a single mutagenic oligonucleotide primer based linear-PCR approach as previously described (Shenoy and Visweswariah, 2003). Mutagenic primers contained a minimum of 12 bases of sequence homology either side of the introduced base-pair mismatch. Mutagenesis was performed using Phusion® High-Fidelity DNA

Polymerase. Typical PCR reaction was 25 μL and contained ~100 ng of template DNA,

20 pmol of the mutagenic oligonucleotide, 200 µM dNTPs, 0.5 U of Phusion polymerase in 1X Phusion HF Buffer. A typical cycle was; denaturation at 98 °C for 3 min and then 20-30 cycles of denaturation at 98°C for 30 secs, annealing at temperatures ranging from 50-60°C according to the primer, and an extension step at

72°C for times that would enable one full extension along the length of the plasmid (30 s per kb). A final extension at 72°C for 8 min was included at the end of the cycle.

Reactions were incubated with 20 U of DpnI overnight at 37°C. Half the reaction would then be transformed into chemically competent bacteria by the method described in section 2.3.1. Plasmids were verified by sequencing (GATC Biotech) before use.

For gene silencing of UBE2L3, two 22 mer oligonucleotides were cloned into pMXsCMV miRNA30E backbone plasmids (Chang et al., 2013; Fellmann et al., 2013).

UBE2L3 targeting sequences: UBE2L3miR30E1 (5’TTTCTTTGTAAACTCTTCAGCA3’)

78 and UBE2L3miR30E2 (5’TTTCATCCCACATTTGCGGATT3’) were adapted from previous reports (Fiesel et al., 2014; 1060 Lewis et al., 2015).

2.3.4 Retroviral packaging and transduction

HEK293E or HEK293T cells were seeded in complete DMEM media supplemented with 20 mM HEPES, in 12-well plates at a density 2x105 cells per well.

The following day, cells were transfected with retroviral vector pMXsIP containing gene of interest, pCMV-MMLV-pack plasmid expressing the Moloney murine leukemia virus

(MMLV) capsid, pVSVG expressing Gag-Pol and VSV-G (both were obtained from

Walther Mothes, Yale University). Respective plasmids were used at a ratio of 5:4:1 in a total of 1 μg of DNA to be transfected. Transfections were performed using

Lipofectamine® 2000 under there recommended conditions. Briefly, DNA and

Lipofectamine® 2000 were diluted in Opti-MEM™ I Reduced Serum Medium, mixed at a 1:1 ratio and incubated for 20 min to allow DNA: lipofectamine complexes to form.

Complexes were then added, dropwise, onto cells which were incubated for 48 hours.

Medium/virus was collected and filtered using a 0.45 μm PES filter. Cells to be transduced were seed 24 h prior and received 300 µL of the virus-containing medium.

Cells were incubated with virus for 48 hours, successfully transduced cells were selected with the relevant antibiotic for 7-14 days. All cells used in this study were selected using puromycin (2 µg.mL-1 for THP-1 or HeLa and 6-8 µg.mL-1 for murine

BMDMs).

2.3.5 Isolation of bacterial genomic DNA

Overnight bacterial cultures were centrifuged at full speed for 5 min and resuspended in 500 µl TES buffer (20 mM ethylenediaminetetraacetic acid (EDTA), 100 mM Tris (pH 8), 0.8% Sodium Dodecyl Sulfate (SDS) and incubated at 55°C for 30 min. The lysate was centrifuged at 17,000 xg for 10 minutes, the supernatant was

79 mixed 1:1 with chloroform and vortexed then centrifuged at 17,000 xg for a further 10 min. The upper aqueous layer was collected and mixed 1:1 with isopropanol before centrifugation at 17,000 xg for 10 min at 4°C. The resulting pellet was washed in 70% ethanol and centrifuged again. The DNA pellet was allowed to air dry for 10 min at room temperature and finally resuspended in 100 µL autoclaved Milli-Q water.

2.3.6 Lambda (λ) red recombinase gene disruption of nleL

The lambda red system (Datsenko and Wanner, 2000) was used to generate a genomic deletion of nleL in enterohaemorrhagic Escherichia coli O157:H7 str. EDL933

(EHEC EDL933). Electrocompetent EHEC EDL933 were prepared as described in

Methods 2.3.2 and transformed with the λ-red recombinase pKD46 plasmid. pKD46 is temperature labile, so following transformation all subsequent growth steps were carried out at 30°C and under ampicillin selection. The PCR product used for the deletion of nleL was generated using primers with 50-nucleotide extensions which are homologous to the regions directly adjacent to the nleL gene ORF (See primer table

2.7). These primers were used in PCR reactions with KOD Hot Start DNA Polymerase.

The pKD4 plasmid acted as a template, generating a fragment containing an FRT- flanked kanamycin resistance cassette with 50-nucleotide extensions. The PCR reaction (50 μL) contained 50-100 ng pKD4, 1x KOD polymerase buffer, 1.5 mM

MgSO4, 0.2 mM dNTPs, 0.3 μM of each primer, 1 U KOD Hot Start DNA Polymerase and 2% DMSO. Cycling conditions were denaturation at 95°C for 2 min, followed by

30 cycles of denaturation at 95°C for 20 secs, annealing at 50°C for 30 secs, and an extension step at 70°C for 40. PCR yielded a ~1.5 kb fragment which was purified using a Sigma GenElute Gel Extraction Kit. EHEC EDL933 expressing pKD46 were grown overnight at 30°C in LB broth containing 100 µg.mL-1 ampicillin. The following day the bacteria were subcultured (1:100) in 50 mL of fresh LB broth and grown for 2

– 2.5 hours. To induce expression of the λ-red recombinase gene 1 mM arabinose was added and the culture was grown for a further 30 min to an OD600 of ~0.6 after which the culture was chilled on ice for 30 min. The culture was then centrifuged at

4°C and the bacterial pellet was washed twice with 15% ice-cold glycerol; once in 25 mL and once in 12.5 mL. The bacterial pellet and was finally resuspended in 80 µL cold 15% glycerol, 1-2 µg of purified PCR product was added and bacteria were electroporated and grown as described above on LB agar plates containing 30 µg.mL-

1 kanamycin (Methods section 2.3.2). Positive colonies were patched onto a fresh kanamycin LB agar plates (30 µg.mL-1), colonies that displayed regrowth were verified for gene deletion of nleL by colony PCR. Positive ΔnleL mutants were cured of pKD46 by overnight growth at 42°C and tested for AmpR sensitivity by colony patching on antibiotic containing plates.

2.4 Cell biology treatments and analyses 2.4.1 Inflammasome activation assays

Cells were primed with ultra-pure E. coli O111:B5 LPS (250 ng.mL-1) for 3 h or

PAM3CSK4 (1 μg.mL-1) for 2 h in serum-containing medium. After priming, the culture medium was removed cells and replaced with serum-free OptiMEM supplemented with 1 mM sodium pyruvate. Cells were then treated with inflammasome-specific agonists see table 2.9. For experiments with additional treatments, these were added as described in table 2.10 relative to inflammasome activation.

Inflammasome Activators Concentration Length of treatment NLRP3 ATP 5 mM Up to 1 h NLRP3 Nigericin 10 µM/25 µM Up to 1 h AIM2 poly(dA:dT) transfection 2 µg.ml-1 5 h with Lipofecatimine® 2000

Non-canonical Cholera toxin B (CTB) CTB, 20 μg.mL-1 8 h caspase-4-NLRP3 and LPS and LPS, 5 μg.mL-1

NLRP1 Anthrax toxin: Lethal LF, 2.5-10 6 h factor (LF) Protective μg.mL-1 and PA, antigen (PA) 10 μg.mL-1

PYRIN C3 exotoxin 0.5-2.5 μg.mL-1 4 h NLRP3 Imject Alum 200 µg.mL-1 5 h

Table 2.9. Inflammasome agonists used in these studies Concentrations and timescale of inflammasome agonist treatments used throughout these studies.

Treatment Concentration Time of addition

Potasium Chloride (KCl) 50 mM 20 min prior

Glycine 5 mM 20 min prior ac-YVAD-fmk 50 µM 20 min prior

MG132 20 µM 5 min post

Epoxomycin 5 µM 5 min post

Table 2.10. Additional treatments used alongside inflammasome activation Concentrations of additional chemicals and inhibitors used along with inflammasome activating agents. The times presented are relative to the addition of the inflammasome agonist and incubations were continued in the presence of above agents for various durations required for activation of various inflammasomes.

2.4.2 Cell priming assays

Priming time course assays were performed for time points ranging between

0.5 – 24 hours. Cells are primed with LPS (250 ng.mL-1), PAM3CSK4 (1 μg.mL-1), human recombinant TNF (2.5 ng.mL-1) or bacterial mutants as described above in

Methods section 2.2.2. Inhibitors included at various time points and are described in the respective results sections as they appear. Inhibitors used included: MG132 (20

μM), epoxomicin (5 μM), Bafilomycin A (20 nM), pepstatin A (10 μg.mL-1), E64D (10

μg.mL-1) and zVADfmk (20 μM). Treated cells were analysed by, western blot, ELISA or qPCR as stated (Methods sections; 2.4.3, 2.4.4 and 2.4.5)

2.4.3 Sample preparation for SDS-polyacrylamide gel electrophoresis and protein visualisation

For detection of proteins in cell lysates cells were washed once in PBS and lysed in hot 2x Laemmli buffer (4% SDS, 20% glycerol, 120 mM Tris-Cl (pH 6.8), 20 mM EDTA, 0.2% bromophenol blue) containing; 2-mercaptoethanol (5%),

Phenylmethylsulfonyl fluoride (PMSF, 2 mM), 1x protease inhibitor cocktail, MG132

(10 µM) and PR619 (25 µM). For detection of secreted proteins in serum-free

OptiMEM-based inflammasome-activation assays, cell culture supernatants were collected and precipitated overnight in three volumes of ice-cold acetone at -20°C, followed by centrifugation at 15,000 xg for 15 min. Air dried pellets were resuspended in 2x Laemmli buffer containing 2-mercaptoethanol (5%).

Prior to gel loading, samples were boiled, vortexed two times and centrifuged at 17,000 xg for 3 min. Proteins were separated by SDS polyacrylamide gel electrophoresis (SDS-PAGE) using either Tris-Glycine running buffer (25 mM Tris-

HCl, 192 mM glycine and 0.1% (w/v) SDS) or Tris-tricine buffers (Cathode buffer: 100

83 mM Tris-HCl, 100 mM tricine, 1%SDS; Anode buffer: 100 mM Tris-HCl adjusted to pH

8.9).

For western blot analysis SDS-PAGE gels were soaked in transfer buffer (10 mM Tris-HCl, 100 mM glycine and 10% (v/v) methanol) for 10- 20 min. Immun-Blot® polyvinyl difluoride (PVDF, 0.2 μm, Biorad) was activated in 100% methanol and soaked in transfer buffer for 10 min. Proteins were transferred in a BioRad Trans-Blot®

Turbo™ Transfer System (up to 1.0 A; 25 V, 30 min.). Membranes were washed once in PBS-Tween (PBST, 1x PBS, 0.1% (v/v) Tween-20) then subsequently incubated in blocking buffer (10% milk (w/v) or 5% BSA (w/v) in PBST) for 1 hour to block non- specific interactions. Primary antibodies were diluted as specified by the manufacturer and incubated overnight with membranes at 4°C with gentle agitation. The following day, membranes were washed three times in PBST (5 min per wash) and incubated in 5% milk (w/v) PBST with relevant HRP-conjugated secondary antibodies for 1 h at room temperature. Membranes were washed in PBST as before and incubated with enhanced chemiluminescence (ECL) reagent for 5 min before being imaged using

BioRad ChemiDoc Imaging System. For Coomassie staining, gels were incubated with

Coomassie Brilliant Blue staining solution (0.1% (w/v) Coomassie Brilliant Blue, 50%

(v/v) Methanol and 10% (v/v) acetic acid) for 1 h, excess stain was removed using detaining solution (40% (v/v) Methanol and 10% (v/v) acetic acid) overnight.

2.4.4 Cytokine ELISA, cell death, and membrane permeability assays

For these analyses, experiments were performed in 96-well plates and a minimum of two technical replicates were performed for each treatment condition. The following commercial ELISA kits were used: mouse IL-1β (eBioscience; 88-7013) mouse TNF (88-7324; eBioscience), mouse IL-6 (88-7064; eBioscience), human IL-

1β (DY201; R&D Systems or eBioscience;88-7261), human IL-6 (88-7066;

84 eBioscience), human TNF (88-7346; eBioscience). Standard curves were made by serial 2-fold dilution of standard solution as per the specification of each kit.

Cell death was measured by quantifying the release of LDH from cells using the CytoTox 96® Non-Radioactive Cytotoxicity Assay (Promega; #G1780). Assays were performed according to the manufacturer’s specification. Briefly, 40 µL of cell culture supernatant was mixed with 40 µL of assay substrate mix and incubated in the dark for ~5 – 10 min. Reactions were stopped by the addition of 40 µL of stop solution.

The assay generates a red formazan product that can be measured by the absorbance at 490 nm. Results were normalised to a blank medium control and were presented as a percentage of 100%-lysed cell control (untreated cells lysed with 1% Triton X-

100).

Cell permeability was measured propidium iodide (PI) uptake. Following cell treatment, cell medium was removed and replaced with sterile PBS containing 5

µg.mL-1 PI. Cells were incubated for 15 min and the fluorescence was measured at

620 nm. Measurements were normalised to empty wells containing PBS with PI and expressed as a percentage of a 100% permeabilised control (Untreated cells incubate in PBS, with PI and 0.05% Triton-X100)

2.4.5 Quantitative reverse transcription PCR

Prior to RNA extraction surfaces and equipment were cleaned using 70% ethanol and RNase AWAY™ surface decontaminant spray. RNA was prepared from cells (two experimental replicates) using the RNeasy mini kit (Qiagen) or PureLink

RNA mini kit (Thermo Fisher Scientific) according to the manufacturer's specifications.

0.5 – 1 µg of purified RNA was used for reverse transcription (RT) using the ProtoScript

First Strand cDNA Synthesis Kit (NEB) or TaqMan Reverse Transcription Reagents

(Thermo Fisher Scientific) using random hexamer primers and under the conditions

85 specified by the respective manufacturers. Quantitative PCR (qPCR) was carried out using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad Laboratories) on a

StepOnePlus Real-Time PCR System (Thermo Fisher Scientific). A typical reaction of

10 μL contained 5 μL SsoAdvanced™ Universal SYBR® Green Supermix (2x), 200-

500 ng of cDNA and 500 nM of each primer (see Table 2.8). qPCR was performed using the following parameters for amplification: polymerase activation and DNA denaturation for 30 s at 95°C; 40 cycles of 3 s denaturation at 95 °C, 30 s annealing/extension at 60 °C. Reactions were performed in duplicate including negative control lacking cDNA or primer. Data were expressed as fold-change in mRNA between treated and untreated calculated by Ct method normalised to

GAPDH.

2.4.6 Immunofluorescence analysis

Cells were plated on coverslips in 24-well plates 24 h before experimental use.

After treatment/infection cells were washed twice in PBS and fixed with 300 µL of cold

4% paraformaldehyde in PBS for 20 min. Cells were then washed 3x in PBS and permeabilized for 3 min in PBS with 0.1 % saponin. Coverslips were washed another

3 times with PBS for 5 min each. Non-specific binding was blocked by incubating cells in PBS containing 5% BSA and 10 % donkey serum for 1 h at room temperature. Cells were then washed in PBS and incubated for 1 h with primary antibody in 5% BSA-PBS with 0.1% saponin. The antibody solution was aspirated, and coverslips were washed

3 times in PBS. Donkey anti-rabbit antibody-Alexa 647 was used as secondary antibody and was incubated with coverslips for 30 min. Finally, nuclei were stained with 0.1 µg.mL-1 Hoechst dye for 5 min. Cells were rinsed with PBS mounted on slides using Prolong Gold anti-fade mounting medium (ThermoFisher Scientific). Fluorescent microscopy was performed using Zeiss AxioObserver Z1 microscope.

2.4.7 Pro-IL-1β immunoprecipitation (IP)

Cells were seeded at a density of 3x106 cells/well in a 6-well plate. Treatments were carried out in triplicate. Cells were treated with 250 ng.mL-1 LPS for 6 hours, or treated with LPS were MG132 (20 μM) was added for the last hour. Each well was lysed in IP lysis buffer consisting of 10 mM Tris, 150 mM NaCl, 1% NP-40, 1x protease inhibitor cocktail, 1x phosphatase inhibitor cocktail, 25 mM N-ethylmaleimide (NEM),

10 μM MG132 and 25 μM PR619. Similarly treated lysates were collected and pooled in chilled 1.5 mL sample and incubated with 4° with agitation for 1 h. Lysates were centrifuged at 17,000 xg for 15 min. The soluble fraction was removed and 40 μL of each lysate was kept aside as “input” sample. The remaining lysate was incubated overnight with agitation at 4°C with 2 μg of anti-mouse IL-1β antibody (B122,

BioLegend). As a negative control for antigen specificity, the lysate was incubated with IgG1 isotype control (G235-2356, BioLegend). The following day, IP samples were incubated with 40 μL of SureBeads™ Protein G Magnetic Beads 1 hour. Bound

IgG-protein complexes were washed times in IP lysis buffer without inhibitors and proteins were eluted by being resuspended in 2x Laemmli buffer with 5% (v/v) 2- mercaptoethanol. Samples were then boiled for 5 min., centrifuged and analysed by

SDS-PAGE and western blot.

2.4.8 Growth-curve analysis of wildtype and ΔnleL EHEC

Wildtype and ΔnleL were grown overnight with shaking at 37°C in LB broth. The following day each overnight was subscultured into fresh LB in three new cultures at a starting OD600 of ~0.1 which acted as technical replicates. Cultures were then grown at shaking at 37°C and bacterial growth was determined by measuring OD600 every 30 mins.

2.5 Recombinant protein purification 2.5.1 GST-tag based protein purification.

BL21-CodonPlus (DE3)-RIPL cells were transformed with pGEX-6P1 encoding the gene of interest for expression of GST-tagged proteins. For purification of NleLFL and NleL170-782, single colonies were inoculated into 5.5 mL LB containing ampicillin

(100 μg.mL-1) and grown overnight with shaking at 37°C. The following day overnights were sub-cultured 1:100 into 500 mL of fresh LB containing ampicillin (100 μg.mL-1) in a 2 L flask. The cells were grown at to an OD600 of 0.4 – 0.5 at which point protein expression was induced with Isopropyl β-D-1-thiogalactopyranoside (IPTG) at a final concentration of 100 µM. Cultures were then incubated for 18 h at 20°C with agitation.

For purification of UBE2L3, 50 mL cultures were used and were induced with IPTG for

3 h at 37°C.

Bacterial pellets were collected by centrifugation at 2,500 xg for 10 min at 4°C.

The growth medium was decanted, and the bacteria were washed three times with ice-cold PBS. Washed bacterial pellets were then frozen at -80°C. Frozen bacterial pellets were resuspended in 5-20 mL of cold lysis buffer 1 (50 mM Tris pH 8, 1 mM dithiothreitol (DTT), 100 mM NaCl, 2 mM EDTA, 0.1% Triton-X) and PMSF was added to a final concentration of 2 mM to reduce proteolysis. The bacterial pellet was rapidly thawed at 37°C for 5 min and thoroughly mixed by vortexing. All subsequent steps were performed at 4°C or on ice. The bacterial suspension was sonicated for 5-10 min at 50% amplitude with 10-second pulses and 10 second rests. Bacterial lysates were cleared by centrifugation at 30,000 xg for 25 min. Protein was collected by column purification; cleared bacterial lysates were passed through glutathione sepharose resin packed columns twice and washed three times with three column volumes of lysis buffer. For purification of GST-tagged proteins, proteins were eluted from the

88 column with elution buffer (100 mM Tris pH 8, 1 mM DTT, 100 mM NaCl, 10% glycerol and 10 mM reduced glutathione). To remove reduced glutathione eluted proteins were desalted into desalting buffer using an ÄKTA Start FPLC system and GE HiTrap™

Desalting columns. For removal of GST-tags, column-bound proteins were incubated overnight at 4°C in desalting buffer (25 mM Tris pH 8, 1 mM DTT, 100 mM NaCl and

10% glycerol) containing GST-rhinovirus C3 protease. The C3 protease cleaves at

LEVLFQ↓GP and removes the GST-tag, releasing purified protein was collected by gravity elution. Protein purity was assessed by Coomassie staining and protein concentration was determined by Bradford assay (Ernst and Zor, 2010)

2.5.2 Nickel (Ni) Affinity purification of His-tagged UBE1 protein

His-tagged proteins were purified similarly to a previously described protocol

(Carvalho et al., 2012). BL21-CodonPlus (DE3)-RIPL cells expressing pET-21d UBE1 were grown to an OD600 of 0.4-0.5 and induced with IPTG. Cultures were then incubated for 18 h at 20°C with shaking. For purification of UBE1, 1 L cultures were induced with 100 µM IPTG for 3 h at 37°C. Bacteria were collected and lysed as before in lysis buffer 2 (50 mM Tris pH 8, 500 mM NaCl, 0.1% Triton-X, 25 mM imidazole and

5 mM 2-mercaptoethanol). Bacterial lysates were cleared as before and His-UBE1 was purified with the ÄKTA Start system using a 5 mL GE HisTrap™ column containing Ni Sepharose High-Performance resin. The column was equilibrated with five column volumes of lysis buffer 2. Cleared lysate was loaded onto the column at a rate of 1 mL/min, the resulting flow-through was again loaded on to the column to ensure complete binding of the protein. Unbound protein was removed by washing with 25 column volumes of lysis buffer 2 applied at a rate of 2.5 mL/min. Bound protein was eluted with Ni elution buffer (50 mM Tris pH 8, 500 mM NaCl, 0.1% Triton-X, 500 mM imidazole and 5 mM 2-mercaptoethanol) and collected in 1 mL fractions. Fractions

89 containing His-UBE1 were pooled and imidazole was removed by desalting into Ni desalting buffer (10 mMTris pH 8 and 5 mM 2-mercaptoethanol) using the ÄKTA Start system and GE HiTrap™ Desalting columns. His-UBE1 was then purified further by anion exchange and eluted with a 0-0.5 M NaCl gradient. Fractions containing UBE1 were pooled and concentrated using centrifugal concentrators. Further purification involved size exclusion chromatography using a HiPrep™ 16/60 Sephacryl™ S-200

HR using gel filtration buffer (10 mM Tris pH 8, 100 mM NaCl and 5 mM 2- mercaptoethanol). Protein purity was assessed by Coomassie staining and protein concentration was determined by Bradford assay.

2.5.3 Bradford Protein Assay.

Bovine serum albumin (BSA) was serially diluted to produce a standard curve with a concentration range from 5 µg to 0.0049 µg. The purified recombinant protein was diluted in 50 µL Milli-Q. 200 µL of Bradford dye reagent was added to each sample and absorbance was measured using at a ratio of 595 nm over 450 nm. Final protein concentrations were determined using the standard curve.

2.6 In vitro Ubiquitylation assays 2.6.1 General in vitro assay

Ubiquitylation assays were carried out as previously described (Hospenthal et al., 2013) with alterations. Assays were performed in 10 μL reactions in ubiquitylation assay reaction buffer consisting of 40 mM Tris, 10 mM MgCl2 40 mM Tris (pH 7.5), 10 mM MgCl2 and 0.6 mM DTT. Proteins were incubated at the following concentrations: wildtype and mutant variants of ubiquitin (Ub, 57 μM), commercial or lab-made UBE1

(E1, 125 nM), wildtype or mutant variants of UBE2L3 (E2, 5 μM) and NleL or SopA variants (E3, 1.5 μM). Reactions were started with the addition of 2.5 mM ATP and were incubated at 37°C for 20 min (unless other stated). Reactions were stopped by

90 adding 10 μL 2x Laemmli dye containing 200 mM DTT. Samples were boiled for 5 min and centrifuged at 17,000 xg for 3 min prior to SDS-PAGE.

2.6.2 Trans ubiquitylation assay

Assays were performed as described above with the following changes.

UBE2L318R was used as a ubiquitin donor and UBE2L3C86A or UBE2L3P62A/F63A were used as ubiquitin receivers. UBE2L3 variants were either incubated alone with NleL or as a mixture of donor and receiver E2. Reactions which required the co-incubation of two UBE2L3 variants (UBE2L318R and UBE2L3C86A or UBE2L3P62A/F63A) had 5 μM of each.

2.6.3 Unanchored ubiquitin chain formation assay

Assays were performed as described above with the following alterations.

Reactions were performed with GST-UBE2L3 variants and an untagged wildtype

UBE2L3 as a control. After a 20 min reaction the assay was stopped by the addition of 10 μL 2x termination buffer (100 mM Tris (pH 8), 2 mM DTT, 200 mM NaCl, 40 mM

EDTA and 0.2% (v/v) triton-x100). Ubiquitin modified UBE2L3 was removed from samples by incubating them with 10 µL (bed volume) of glutathione sepharose resin for 45 min at 4°C with agitation. Reaction supernatant contained unanchored ubiquitin was removed mixed with 2x Laemmli dye and analysed by SDS-PAGE and western blot.

2.6.4 N-terminal ubiquitylation assay

Untagged and GST-tagged wildtype and K82 only UBE2L3 proteins were used together with Ubiquitin His-NoK in the ubiquitylation assay and halted with termination buffer as indicated above. Resin bound protein was washed three times with 1x termination buffer and resuspended in a final volume of 30 µL. A portion (15 uL) was removed to act as an uncleaved fraction and the rest was incubated at 4°C overnight

91 with GST-rhinovirus C3 protease to remove GST-tags. Samples were centrifuged and the supernatant containing cleaved UBE2L3 protein was removed. Uncleaved and cleaved samples were analysed by SDS-PAGE and western blot.

2.6.5 Recombinant protein production and caspase-1 cleavage assays

Human IL-1β (amino acids 1090 110-269) was expressed from pGEX-6P1 to add a GST tag, and human caspase-1 p20 (amino acids 120-297) and 1091 p10

(amino acids 317-404) were expressed from pProExHT adding N-terminal 6xHis tags.

GST-fusion proteins were purified as described in section 2.5.1. Recombinant caspase-1 p20 and p10 were purified from protein inclusion bodies renatured without malonate (Scheer et al., 2005). Briefly, protein expression was induced with 100 μM

IPTG for 3 h at 37°C. Bacterial pellets were collected as described in 2.5.1. Bacteria were lysed by sonication in lysis buffer 1 (50 mM Tris–HCl, pH 8, 150 mM NaCl, and

0.1% Triton X-100, 5 mM DTT). Inclusion bodies were isolated by centrifugation at

3,000 xg for 15 min and were washed 2x in lysis buffer 1 and renatured overnight in lysis buffer containing 6 M GnHCl. Proteins were dialysed overnight and p20 and p10 were purified by Ni-affinity purification. Proteins were refolded by mixing 1 mg of each caspase-1 subunit in 100 mL of buffer containing 50 mM HEPES, pH 8, 100 mM NaCl,

1 M non-detergent sulfobetaine 201 (NDSB 201) and 10 mM DTT. Precipitates were removed by centrifugation at 18,000 xg, proteins were then concentrated using centrifugal concentrators (Millipore), and dialysed overnight against 100 mM HEPES,

10 mM DTT and 10% sucrose. Caspase-1 assay cleavage assays were performed for

1 h at 37°C in buffer containing 100 mM HEPES, 0.1 % CHAPS, and 10 mM DTT.

2.7 Antibodies and Reagents 2.7.1 List of antibodies

Primary Application Species Source Anti-NLRP3 (Cryo-2) WB Mouse Adipogen Anti-ASC (AL177) WB/IF Rabbit Adipogen Anti-UBE2L3 WB Rabbit GeneTex Anti-mCaspase-1 (p45/p10) WB Rabbit SCBT Anti-mCaspase-1 (Casper-1, WB Mouse Adipogen p45/p20) Anti-hCaspase-1 (p45/p20) WB Rabbit Cell Signaling Anti-mIL-1β WB Goat R&D Systems Anti-hIL-1β WB Goat R&D Systems Anti-hIL-18 WB Rabbit MBL Anti-mIL-18 WB Rabbit Biovision Anti-mIL-1α WB Goat Biolegend Anti-mIL-1β IP Hamster Biolegend IgG1 isotype control IP Hamster Biolegend Anti-GSDMD WB Mouse SCBT Anti-HMGB1 WB Rabbit GeneTex Anti-β-Actin, Peroxidase WB Mouse Sigma Anti-Caspase-11 (mCas4) WB Rat Bioscience Anti-ATG7 WB Rabbit CST Anti-FLAG® M2 WB Mouse Sigma Anti-HA WB Mouse Biolegend Anti-GFP WB Mouse Roche Anti-ubiquitin(K48-linked)-HRP WB Rabbite CST Anti-ubiquitin(K63-linked)-HRP WB Rabbit CST Anti-ubiquitin(P4D1)-HRP WB Mouse SCBT Anti-O157 IF Goat Fitzgerald Secondary Application Species Source Amersham anti-Mouse IgG WB Donkey GE HRP LifeSciences Amersham anti-Rabbit IgG WB Donkey GE HRP LifeSciences Anti-Goat IgG HRP WB Donkey SCBT Phalloidin Alexa Fluor™ 488 IF N/A Invitrogen anti-Rabbit Alexa Fluor™ 647 IF Goat Invitrogen anti-Goat Alexa Fluor™ 647 IF Donkey Invitrogen Table 2.11: List of antibodies used in this study IF = immunofluorescence WB = Western blot

2.7.2 List of treatment reagents and commercial recombinant protein

Reagent Source Adenosine 5′-triphosphate (ATP) disodium salt Sigma Nigericin sodium salt Sigma Glycine Sigma Cholera Toxin B subunit (CTB) Sigma MG132 Sigma Pepstatin-A Sigma Puromycin Sigma Propidium iodide Sigma Ac-YVAD-cmk Sigma Epoxomycin SCBT zVAD-fmk R&D E64D Calbiochem PR619 Calbiochem poly(dA:dT) Invivogen PAM3CSK4 Invivogen Ultrapure E. coli O111:B4 Invivogen Imject Alum Adjuvant thermo Recombinant mouse TNF eBioscience Recombinant human IL-1 R&D C3 Exoenzyme Cytoskeleton Anthrax lethal factor List Laboratories Activated protective antigen List Laboratories Recombinant Protein Source Ubiquitin Boston Biochem His-NoK Ubiquitin Boston Biochem Ubiquitin Mutant No K Boston Biochem Ubiquitin Mutant with K6 only Boston Biochem Ubiquitin Mutant with K48 only Boston Biochem Ubiquitin Mutant with K29 only Boston Biochem His6-Ubiquitin Mutant No K Boston Biochem Table 2.12: List of reagents used in cell treatments and commercial recombinant proteins

Chapter 3 – Depletion of UBE2L3 is a common consequence of inflammasome activation

3.1 Introduction Over the last decade, the signalling pathways which mediate the activation of inflammasomes and caspase-1 have been extensively studied. Recently, research efforts have moved to identifying new substrates and targets of caspase-1 as a means of elucidating the mechanisms behind caspase-1-dependent processes such as pyroptosis and unconventional cytokine secretion.

Most notably, genome-wide based screens from multiple studies identified gasdermin D (encoded by GSDMD) as the executioner of pyroptotic cell death

(Kayagaki et al., 2015; Shi et al., 2015). Following inflammasome activation, GSDMD is cleaved by caspase-1 yielding an N-terminal p30 fragment which forms membrane pores and induces cell death (Aglietti et al., 2016). In addition to acting directly on proteins which serve as substrates, such as gasdermin D and pro-IL-1β, caspase-1 also regulates many non-substrate targets. These include proteins such as pro-IL-1α

(Groß et al., 2012) and HMGB1 (Lamkanfi et al., 2010), the secretion of which is mediated by caspase-1 but neither protein is proteolytically processed and are therefore not a caspase-1 substrate.

Despite these recent advancements, there are many caspase-1-mediated responses which remain poorly understood. These processes are likely regulated by uncharacterised caspase-1 substrates and targets. Many large-scale proteomics- based studies have been performed to identify new substrates of caspase enzymes.

In fact, prior to the discovery of its role in pyroptosis, gasdermin D was first identified as a caspase-1 substrate in a proteomics study that used quantitative stable-isotope labelling with amino acids in cell culture (SILAC) and mass spectroscopy (Agard et al.,

2010). Many other caspase-1 substrates have been similarly identified and can be found in the MEROPS database (https://www.ebi.ac.uk/merops/) (Rawlings et al.,

2016), an online repository of peptidase substrates from multiple proteomic studies.

Many of these studies, however, focused on the identification of proteins which are cleaved by caspase-1. These so-called degradomics approaches require the tagging and/or enrichment of newly generated protein termini prior to peptide identification by mass spectroscopy (Agard et al., 2010; Lamkanfi et al., 2008; Mahrus et al., 2008).

However, such techniques may not be able to identify cleaved proteins whose stability is altered, proteins which have terminal modifications or non-substrate proteins which are targeted in a caspase-dependent manner (i.e. pro-IL-1α and HMGB1). Other methods such as protein topography and migration analysis platforms (PROTOMAP) have been used to identify proteolytic events during apoptosis (Dix et al., 2008) and

Plasmodium infection (Bowyer et al., 2011). PROTOMAP detects changes in protein molecular weight and peptide migration which are taken as an indication of proteolytic cleavage (Dix et al., 2014).

PROTOMAP analysis compares the proteomes of control (untreated) and experimental (e.g. caspase-1 activated) cells. Cell lysates are then separated by one- dimensional SDS- polyacrylamide gel electrophoresis (PAGE) and each gel lane is cut at exactly 5 mm intervals to produce ~16 pieces per lane. Gel pieces are then processed for in-gel tryptic digestion and peptide mass fingerprinting. Comparative mass spectrometric analyses of gel pieces are performed, and the resulting proteomics data organised into a peptograph which integrates peptide sequence coverage and relative gel migration of a given protein. Therefore, if peptides corresponding to a protein in treated cell lysates are found in lower portions of the gel as compared to that in control samples (indicating lower molecular weight (Mw) than that of full-length protein) it is assumed to have been proteolytically cleaved.

Alternatively, if none or fewer peptides corresponding to a given protein are found in

samples from caspase-1 activated cells, the protein may have been depleted in cell

lysates, for example by secretion/release from cells. It should be noted, that

considerations must be made during the interpretation of these analyses. For example,

pyroptosis and cell death are hallmark features caspase-1 activation, the loss of cells

and cellular material as a result of lysis should be considered and accounted for when

determining the specific requirement for caspase-1. Furthermore, this method can only

detect ‘stable’ fragments which result from proteolysis. Additional cleavage into

Figure 3.1: Identification of caspase-1 substrates and targets by comparative PROTOMAP proteomics. Schematic of PROTOMAP experiment and analysis. Lysates from control (LPS treated) and experimental (Exp; LPS+nigericin treated) immortalised bone marrow-derived macrophages (iBMDMs) were separated by one- dimensional SDS-PAGE. The gel lanes were cut into bands at 5 mm intervals to yield approximately 16 gel pieces per lane. Gel pieces were then processed for in-gel tryptic digestion and peptide mass fingerprinting. A putative substrate of caspase-1 (green band) would appear as multiple peptides of lower molecular weights (Mw) in Exp sample as compared with those in Ctrl samples. Putative indirect targets of caspase-1 (orange band) would, however, be absent or depleted from Exp samples versus Ctrl, for example due to their release/secretion (Schematic adapted from Dix et al., 2014)

98 smaller peptides or subsequent protein degradation would be detected as a loss of protein in treated samples.

To identify proteins which are either cleaved (substrate) or depleted (indirect target) in a caspase-1-dependent manner, shotgun proteomic analyses similar to

PROTOMAP was performed (Dix et al., 2014) by the Shenoy group. Unbiased comparative analyses of murine macrophages before and after caspase-1 activation by nigericin helped identify potential substrates and indirect targets of caspase-1 (Fig

3.1). Such analyses were performed (by Dr Shenoy) on LPS-treated immortalised bone marrow-derived macrophages (iBMDMs) and LPS plus nigericin treated iBMDMs. This revealed known substrates and targets such as pro-IL-1β, pro-IL-18 and HMGB1, as well novel, substrates and targets of caspase-1 in murine macrophages.

One newly identified target was the ubiquitin-conjugating enzyme UBE2L3

(Eldridge et al., 2017). Peptides corresponding to UBE2L3 were missing from samples prepared from murine macrophages activated with LPS and nigericin to activate the

NLRP3 inflammasome. This suggested that UBE2L3 could be a target of caspase-1.

I decided to validate this rigorously in human and mouse macrophages, and in conditions where other inflammasomes were activated. This chapter shows that the activation of multiple inflammasomes can trigger rapid loss of cellular UBE2L3 protein in mouse and human macrophages independently of cell death. Depletion of UBE2L3 is genetically and functionally dependent on inflammasome components, and ultimately requires the activation of the cysteine protease caspase-1. UBE2L3 is, therefore, a new indirect target of inflammasomes and caspase-1

3.2 Results 3.2.1 Depletion of UBE2L3 is triggered by NLRP3 inflammasome agonists

Previous comparative mass spectrometric analyses performed by the Shenoy group, revealed that peptides corresponding to UBE2L3 are lost from murine macrophages following robust activation of NLRP3 following LPS-priming and subsequent activation by K+ efflux induced by the bacterial ionophore nigericin.

To confirm these initial observations obtained through mass spectrometry, murine iBMDMs were stimulated to activate the NLRP3 inflammasome followed by immunoblot-based analyses. Production of NLRP3 and related pro-inflammatory cytokines (i.e pro-IL-1β) were induced by stimulating iBMDMs with 250 ng.mL-1 LPS for 3 hours in full medium. LPS containing medium was then removed and replaced with serum-free OptiMEM. NLRP3 inflammasome assembly and activation was then triggered by pulsing cells with 10 μM nigericin for 1 h leading to K+ efflux. The culture supernatant was removed and precipitated overnight in ice-cold acetone. The resulting precipitate was resuspended in Laemmli loading dye and analysed by SDS-PAGE and western blotting for detection of active caspase-1, and cellular lysates were prepared for the detection of UBE2L3. Western blot analyses confirmed that UBE2L3 was not detected in iBMDMs where NLRP3 had been activated (Fig 3.2A, lane 2). NLRP3 activation was evident by the appearance of the active p10 caspase-1 subunit in cell supernatants (Fig 3.2A, lane 2).

Membrane rupture leading to cell lysis is a hallmark feature of pyroptotic cell death and permits the release of intracellular components. Others have shown that glycine may act as an osmoprotectant, and prevents osmotic stress-induced lysis

(DiPeso et al., 2017; Fink and Cookson, 2006; Wellington et al., 2014). To test

100

Figure 3.2: NLRP3 activation triggers UBE2L3 depletion independently of cell lysis. Wildtype iBMDMs primed with LPS for 3 hours, were left untreated, treated with 10 µM nigericin for 1 hour with or without pre-incubation with 5 mM glycine as indicated. Cell lysates or supernatants were used for immunoblots as labelled. (A) Immunoblots for UBE2L3 and β-actin in cell lysate, and caspase-1 in culture supernatants. Data are representative of 4 independent experiments. (B) Cytotoxicity of treatments measured by lactate dehydrogenase (LDH) release. Mean±S.E.M. from four independent experiments are plotted. * P<0.001 by unpaired two- tailed Student’s t-test. whether UBE2L3 was passively lost during pyroptosis in cells stimulated with 10 μM nigericin in the presence of 5 mM glycine.

The addition of glycine reduced cell death, as determined by LDH release, by approximately 40% without affecting caspase-1 activation (Fig 3.2A and B). Despite this reduction in cell death, UBE2L3 depletion was still observed. This suggested that

UBE2L3 was not lost passively during cell lysis (Fig 3.2A). Both signal 1 and signal 2 are required to trigger NLRP3 inflammasome assembly and activation. However, common signal 1 agonists also activate a plethora of non-inflammasome related cell signalling pathways (i.e. LPS triggering TLR4 dependent NF-κB activation and

101 inflammation) (Newton and Dixit, 2012). Likewise, potassium efflux is also associated with non-inflammasome related signalling (Urrego et al., 2014). To test whether either signal 1 or signal 2 agonists were alone responsible for UBE2L3 depletion, iBMDMs were treated with either a signal 1 or signal 2 agonists. iBMDMs pulsed with 250 ng.mL-1 LPS for 3 hours showed no change in levels of UBE2L3 when compared with unstimulated cells (Fig 3.3A). Similarly, cells treated with NLRP3 agonists nigericin or the endogenous danger signal ATP retained UBE2L3, which suggested that K efflux, via the nigericin toxin or P2X7 receptor activation were, alone, not sufficient to cause

UBE2L3 depletion (Fig 3.3B).

Figure 3.3: UBE2L3 does not deplete following separate addition of signal 1 and signal 2 agonists (A) Immunoblots of UBE2L3 and β-actin in cell lysates of wildtype iBMDMs left untreated or treated with LPS for 3 hours. Data are representative of 2 independent experiments. (B) iBMDMs were left untreated or treated with either nigericin or ATP for 1 hour. Immunoblots for UBE2L3 and β-actin in cell lysate, and caspase-1 in culture supernatants. Data are representative of 2 independent experiments.

102

3.2.2 UBE2L3 depletion depends on potassium efflux in response to

NLRP3 inflammasome agonists

To better explore a direct causative link between NLRP3 activation and

UBE2L3 depletion I tested whether UBE2L3 depletion was prevented when NLRP3 activation by the endogenous agonist ATP or the bacterial toxin nigericin were inhibited. The NLRP3 inflammasome assembles in response to a multitude of diverse and unrelated signals (Eldridge and Shenoy, 2015). The most widely accepted view is that NLRP3 activation is itself triggered by potassium ion efflux, which serves as a common convergent signal for most NLRP3 agonists (Muñoz-Planillo et al.). Studies have established that the addition of potassium to culture medium inhibits the activation of the NLRP3 inflammasome by preventing potassium ion efflux (Muñoz-

Planillo et al.).

Dendritic cells were prepared by culturing in medium containing G-CSF and macrophages similarly prepared by using M-CSF from L929 cell spent medium.

Primary BMDMs and BMDCs were primed with 250 ng.mL-1 LPS for 3 hours. Cells were then either treated with ATP alone for 1 hour, or with an additional pre-treatment with either 50 mM KCl or 5 mM glycine 20 mins prior to and during treatment with ATP

(Fig 3.4). Western blot analyses revealed that primary BMDCs and BMDMs show reduced cellular UBE2L3 following NLRP3 and caspase-1 activation (Fig 3.4, lanes

2+4 and 6+8). As with ATP treatment in iBMDMs, the addition of glycine prior to stimulation does not affect either caspase-1 activation, IL-1β release or UBE2L3 depletion.

However, pre-incubation with KCl, inhibited caspase-1 activation in response to ATP treatment and prevented the loss of UBE2L3 (Fig 3.4, lane 3 and 7).

103

These data showed that in response to NLRP3 agonists, K+ efflux leading to

inflammasome activation was required for UBE2L3 depletion. These data additionally

show that like immortalised cells, primary BMDMs and BMDCs lose their UBE2L3

upon caspase-1 activation in response to NLRP3-activating signals.

Figure 3.4: Depletion of UBE2L3 is dependent on potassium efflux LPS primed pri-BMDCs (left) and pri-BMDMs (right) either left untreated or pulsed with 10 µM nigericin for 1 h. Cells were also pre-treated with 50 mM KCl or 5 mM glycine (as indicated) 20 minutes prior to nigericin and during nigericin treatment. Immunoblots of UBE2L3 and β-actin in cell lysate, and caspase-1 and IL-1β in culture supernatant. Data are representative of two independent experiments.

104

3.2.3 Depletion of UBE2L3 temporally correlates with caspase-1 activation

Inflammasomes regulate the autoproteolysis and activation of caspase-1 into its p10 and p20 fragments. Active caspase-1 in turn mediates the release of pro- inflammatory cytokines (e.g. IL-1β and IL-18) and other pro-inflammatory DAMPs (e.g.

HMGB1). To examine whether the dynamics of UBE2L3 depletion correlates with other inflammasome dependent processes, time-course analyses of UBE2L3 depletion following NLRP3 activation with ATP was performed. Following priming with

250 ng.mL-1 LPS for 3 hours, iBMDMs were pulsed with ATP for increasing lengths of time. Samples were prepared at 20-minute intervals for 80 minutes. Immunoblot analysis shows that UBE2L3 was rapidly depleted from cell lysates following treatment with ATP (Fig 3.5A). UBE2L3 can be seen depleting from cells within 20 mins of ATP treatment, with >90% of UBE2L3 lost within 80 minutes of ATP treatment.

Interestingly, UBE2L3 depletion temporally correlated with the appearance of active cleaved p20 caspase-1 and the secretion of known caspsase-1 substrate IL-1β, both of which begin to appear at 20 mins post ATP-treatment. HMGB1 appears to be secreted later during inflammasome activation which was consistent with previous studies (Lotze and Tracey, 2005). As before, the addition of glycine during inflammasome activation reduced cell death by pyroptosis over time as measured by

LDH release (Fig 3.5B). Despite the reduction in cell death following seen with glycine, the rate of UBE2L3 depletion, as well as IL-1β and HMGB1 secretion remained unaffected, which suggested that cell lysis was not a contributing factor to UBE2L3 depletion (Fig 3.5A). These data showed that UBE2L3 depletion correlated with caspase-1 activation. Further, UBE2L3 loss from cells appeared to be uncoupled from inflammasome mediated cell lysis.

105

Figure 3.5: UBE2L3 depletion temporally correlates with caspase-1 activation iBMDMs primed with 250 ng.ml-1 LPS for 3 hours and subsequently left untreated or pulsed with 5 mM ATP for stated periods of time with or without pre-treatment with 5 mM glycine. (A) Immunoblot analysis of lysate UBE2L3 and β-actin and supernatant caspase-1, IL-1β and HMGB1. Data are representative of three independent experiments. (B) Percentage lactate dehydrogenase (LDH) release. Plot of mean±S.D from two independent experiments. * BH corrected P<0.05 by two-way ANOVA.

106

3.2.4 NLRP3 activation by sterile particulate agonists triggers UBE2L3

depletion in primary cells

In addition to molecules such as nigericin and ATP which trigger K+ efflux

directly, phagocytosed particulate matter such as alum activate NLRP3 following the

disruption of phagosomal membranes (Franchi and Núñez, 2008; Hornung et al.,

2008). Such stimuli are slower acting and less potent activators of the inflammasome

(Li et al., 2008; Shenoy et al., 2012), but commonly associated with chronic

inflammatory syndromes such as siliciosis and gout (Hornung et al., 2008; Martinon et

al., 2006).

Figure 3.6: Alum induced UBE2L3 depletion in primary immune cells LPS-primed pri-BMDCs (left) and pri-BMDMs (right) untreated (UT) or treated with alum adjuvant for 5 h show caspase-1 activation and UBE2L3 depletion. Immunoblots for UBE2L3 and β-actin in cell lysate, and caspase-1 in culture supernatants. Data are representative of two independent experiments.

107

To test whether sterile particulates stimulate depletion of UBE2L3, primary bone marrow derived dendritic cells and macrophages were used. pri-BMDCs and pri-

BMDMs were primed for 3 hours with 250 ng.mL-1 LP,S followed by 5 hours of treatment with 200 µg.mL-1 Imject Alum Adjuvant (alum). Following alum treatment both cell lysates and supernatant samples were analyased by western blotting.

Western blot analysis shows that primary BMDCs and BMDMs both respond to alum and display active caspase-1 (Fig 3.6). Like previous observations with iBMDMs, activation of the NLRP3 inflammasome led to the depletion of UBE2L3 from the cell lysate. These data show that multiple immune cell types lose cellular UBE2L3 in response to NLRP3-inflammasome activating stimuli.

3.2.5 UBE2L3 depletion is dependent on core canonical inflammasome components

Having demonstrated a temporal correlation between inflammasome activation and the loss of UBE2L3, I wanted to identify the genes involved in UBE2L3 depletion by caspase-1 and the steps involved. Results from various NLRP3-activatotors suggest that caspase-1 could be the common mediator of UBE2L3 depletion, possibly along with ASC. To test this formally, gene-knockouts were used. Importantly, under resting conditions or following treatment with LPS, UBE2L3 expression was unaffected by the deletion of inflammasome- related genes which were used throughout these studies (Fig 3.7A). Using the NLRP3 agonists ATP or nigericin, wildtype iBMDMs were tested alongside cells with chromosomal deletions of Nlrp3, Asc and Casp1/4.

In wildtype cells, activation of NLRP3 by LPS priming and either ATP (Fig 3.7B) or nigericin (Fig 3.7C) triggered the depletion of UBE2L3. Caspase-1 activation and the processing of IL-1β also occurred as expected in wildtype iBMDMs. Interestingly, loss of any component of the NLRP3 inflammasome protected UBE2L3 from being

108 depleted, as Nlrp3 -/-, Asc -/- and Casp1/4-/- cells all failed to deplete UBE2L3, activate caspase-1 and secrete IL-1β (Fig 3.7B and C, lanes 2 to 4). This suggested that, in the context of canonical NLRP3 activation, UBE2L3 depletion depends on the NLRP3 sensor, ASC adaptor protein and ultimately caspase-1.

Figure 3.7: UBE2L3 depletion is dependent on core components of NLRP3 inflammasome (A) iBMDMs with chromosomal deletions of stated genes were left untreated or treated with 250 ng.mL-1 for 3 hours (A) or primed with 250 ng.mL-1 LPS for 3 hours followed by treatment with 5 mM ATP (B) or 10 µM nigericin (C) for 1 h. Immunoblots of lysate UBE2L3 and β-actin, and supernatant caspase-1 and IL-1β. Schematics in (A) and (B)show NLRP3 inflammasome activation pathways. Data in A and B represent experiments repeated three times.

109

As caspase-4 plays no role in the activation or responses of canonical NLRP3 inflammasomes it can be assumed that it plays no role in UBE2L3 depletion in response to ATP or nigericin. Indeed, only wildtype cells were capable of UBE2L3 depletion (Fig 3.8, lane 3), as double knockout Casp1/4-/- cells which are complemented with flag-tagged mouse caspase-4 (C1/4-/-+Casp4) behaved identically to Casp1/4-/- and retain UBE2L3 following stimulation with LPS and nigericin (Fig 3.8, lane 4). These data showed that UBE2L3 depletion in response to NLRP3 activation by relevant signal 1 and signal 2 agonists ultimately required the presence of caspase-

1. Caspase-1 is the common executor of inflammasome mediated response downstream of multiple NLRs/ALRs. The next questions were whether UBE2L3 was depleted in response to the activation of other common inflammasomes, and which

Figure 3.8: NLRP3 mediated loss of UBE2L3 is strictly caspase-1 dependent Wildtype or Casp1/4-/-+Casp4 iBMDMs primed with 250 ng.mL-1 LPS for 3 hours and left untreated (lanes 1 and 2) or pulsed with 10 µM nigericin for 1 hour (lanes 3 and 4). Immunoblot analysis of lysate UBE2L3, caspase-4, caspase-1 and β-actin and supernatant caspase-1. Data are representative of two experiments.

110 specific genes were required. Another question that arose was whether UBE2L3 was depleted upon inflammasome activation in human cells.

3.2.6 Inflammasome dependent UBE2L3 depletion occurs in human macrophage-like THP1 cells

UBE2L3 is well conserved amongst mammalian species. The mRNA open reading frame of mouse and human UBE2L3 share 96% sequence identity and their protein sequences share 100% identity. It was therefore hypothesised that human

UBE2L3 would also deplete from human macrophage cells following inflammasome activation. To test this, the human leukaemia-derived monocyte cell line THP-1 was used. THP-1 cells can be differentiated into a macrophage-like state when treated with phorbol 12-myristate 13-acetate (PMA). The following experiments were conducted using THP-1 cells differentiated for 3 days, 48 hours in 100 ng.mL-1 PMA, followed by

24 hours culture in fresh medium without PMA. Differentiated THP-1 cells were primed with 250 ng.mL-1 LPS for 3 hours and pulsed with nigericin for 1 hour (Fig 3.9A). As seen with mouse macrophages (Fig 3.2), nigericin treated THP-1 cells lost UBE2L3 following the activation of caspase-1 (Fig 3.9A, lane 2). Consistent with previous data in mouse cells, K+ pre-treatment blocked UBE2L3 depletion, demonstrating that in human cells this process also required NLRP3 activation (Fig 3.9A, lane 3).

To verify the genetic basis of UBE2L3 depletion in THP-1 cells, NLRP3 expression was silenced using shRNA. THP-1 cells underwent viral transduction to stably express a plasmid containing NLRP3 targeting shRNA in an extended micorRNA30 (miR30E) backbone (THP1NLRP3miR) or non-targeting control shRNA

(THP1Ctrl#1).

Western blot analysis showed that THP1 NLRP3miR cells had significantly reduced NLRP3 expression (Fig 3.9B). Similar to results obtained with Nlrp3-/-

111 iBMDMs, THP-1 cells with reduced NLRP3 did not lose UBE2L3 in response to

LPS/nigericin treatment (Fig 3.9B, lane 4).

Figure 3.9: K+ efflux and NLRP3 dependent UBE2L3 depletion in human macrophage- like THP-1 cells Immunoblots of parental (A) or Control (Ctrl#1) and NLRP3 (NLRP3miR) shRNA expressing (B) PMA-differentiated THP-1 cells primed in 250 µg.ml-1 LPS for 3 hours followed by stated treatments. (A) Cells left untreated or pulsed with 10 µM nigericin for 1 hour, in the absences or presence of 50 mM KCl, followed by specified immunoblot analysis. (B) Ctrl#1 and NLRP3miR cells left untreated (left) and or pulsed with 10 µM for one hour treated. Lysates immunoblotted for expression of UBE2L3, NLRP3 and β-actin. Data are representative of four (A) and two (B) experiments.

112

3.2.7 Activation of caspase-4 is required but not sufficient for UBE2L3 depletion induced by non-canonical activation of NLRP3

Inflammasome activation pathways which depend upon a caspase (e.g. caspase-4 or caspase-8) to become activate are referred to as non-canonical mechanisms of inflammasome activation. Most commonly the term is used to refer to the activation of NLRP3 by caspase-4 (also known as caspase-11 in the mouse, but called caspase-4 by the HUGO Gene Nomenclature Committee (HGNC) and therefore called caspase-4 throughout this thesis) in response to cytosolic LPS (Kayagaki et al.,

2011; Shi et al., 2014). Because this pathway requires two similar and functionally related proteases, it created an opportunity to test multiple lines of enquiry. First, does cytosolic LPS trigger UBE2L3 depletion and second, can caspase-4 itself meditate the loss of UBE2L3, or was it simply required for its sensory role in the context of the non- canonical inflammasome pathway.

These questions were addressed relying on genetic knockouts of genes involved in the non-canonical inflammasome activation pathway. iBMDMs which lack individual components of the non-canonical inflammasome were first primed with 1

μg.mL-1 PAM3CSK4 for 2 hours, and subsequently co-treated with cholera toxin B

(CTB, 20 μg.mL-1) and LPS (5 μg.mL-1) for 8 hours. The native function of CTB is to translocate active cholera toxin A component into host cells. Others have since shown that it is also capable of binding to, and translocating other molecules, across the host cell membrane (Hagar et al., 2013; Kayagaki et al., 2011; Kayagaki et al., 2013). These include the O-polysaccharide of specific types of LPS molecules such as LPS

O111:B4 (Kayagaki et al., 2013). LPS O11:B4 was co-treated with CTB to enable the delivery of LPS into the cytoplasm so that it may be detected by caspase-4.

113

Figure 3.10: Caspase-4 is required but not sufficient for UBE2L3 depletion Specified iBMDMs cells were primed with 1 μg.mL-1 PAM3CSK4 for 2 h followed by co- treated cholera toxin B (CTB, 20 μg.mL-1) and LPS (5 μg.mL-1) for 8 hours. Immunoblot analysis of lysate UBE2L3 and β-actin and stated proteins in culture supernatant of (A) iBMDMs lacking non-canonical inflammasome genes. Schematic (centre) shows LPS- dependent non-canonical activation of caspase-1. (B) Asc-/- iBMDMs with or without complementation with ASC-RFP. Data are representative of experiments repeated three times. *non-specific bands.

Notably, UBE2L3 was depleted from wildtype iBMDMs following inflammasome activation by LPS+CTB, as determined by caspase-1 autoproteolysis and IL-1β secretion (Fig 3.10A, lane 1). Akin to previous results Casp4-/-, Casp1/4-/- or Nlrp3-/- which lack individual components of the non-canonical inflammasome fail to deplete

UBE2L3 after treatment (Fig 3.10A, lanes 2 to 4). However, the retention of UBE2L3 in Nlrp3-/- cells (Fig 3.10A, lane 4) was particularly noteworthy because caspase-4 acts upstream of NLRP3 in the non-canonical pathway. Nlrp3-/- cells treated with LPS+CTB would have active caspase-4, but not caspase-1 (Fig.3.10A) (Kayagaki et al., 2011).

114

The presence of UBE2L3 in these cells suggests that caspase-4 cannot trigger

UBE2L3 depletion in the absence of caspase-1.

This was further explored using Asc-/- iBMDMs which do not activate caspase-

1 in response to LPS+CTB (Fig 3.10B), but caspase-4 is activated in these cells in response to cytosolic LPS. As expected, active caspase-1 is not detected in Asc-/- cells and UBE2L3 is not depleted, despite robust activation of caspase-4 (Fig 3.10B, lane

3). Only when Asc-/- iBMDMs are complemented with RFP-tagged ASC protein (ASC-

RFP), which enables caspase-1 to be activated, is UBE2L3 depletion restored (Fig

3.10B, lane 4). These data showed that during non-canonical inflammasome signalling, caspase-4 catalytic activity was required but not sufficient to trigger the loss of UBE2L3, and caspase-4 only functioned in its role as an activator of non-canonical signalling.

3.2.8 Activation of NLRP1 and PYRIN inflammasomes triggers UBE2L3 depletion

The data presented shows that activation of the NLRP3 inflammasome causes the rapid loss of cellular UBE2L3. Immune cells such as macrophages possess a multitude of cytosolic sensors which can assemble into inflammasomes. I therefore wanted to explore whether UBE2L3 depletion was specific to NLRP3 or could result from the activation of other inflammasomes.

For the purpose of this study I decided to focus on the Nlrp1b allele- 1 from

BALB/c mouse strains which is responsive to LT (Chavarria-Smith et al., 2016; Van

Opdenbosch et al., 2014). Primary BMDMs from BALB/c mice were first primed with

250 ng.mL-1 LPS for 3 hours. Following this, cells were treated with a fixed concentration of PA (10 μg.mL-1) and increasing concentrations of LF (2.5-10 μg.mL-

1) for 6 hours (Fig 3.11) . LF concentrations of 2.5 and 5 μg.mL-1 were insufficient to

115 trigger inflammasome and caspase-1 activation, presumably due to low delivery into cells, and under these conditions UBE2L3 expression was similar to that in untreated cells. However, increasing the LF concentration to 10 μg.mL-1 led to robust caspase-

1 activation and UBE2L3 depletion form the cell lysate (Fig 3.11, lane 4).

I next explored whether the non-NLR PYRIN inflammasome could similarly mediate UBE2L3 depletion. The PYRIN inflammasome detects and responds to covalent modifications of host Rho-GTPases, most notably by bacterial toxins (Xu et al., 2014). One such example is the ADP-ribosolyating C3 toxin from Clostridium botulinum. Interestingly, PYRIN expression is lost in iBMDMs and is only seen in primary cells (Xu et al., 2014). Therefore, primary cells were used here. BMDMs from

C57BL/6 mice were primed with LPS for 3 hours. Primed cells were then treated with

Figure 3.11: Activation of NLRP1b inflammasome triggers UBE2L3 depletion Primary BMDMs (BALB/c) primed with 250 ng.ml-1 LPS for 3 hours followed by a 6-hour treatment of a fixed concentration of recombinant protective antigen (PA) and increasing concentrations of recombinant lethal factor (LF). Immunoblot analysis (left) of lysate UBE2L3 and β-actin and supernatant caspase-1. Schematic (right) of Nlrp1b activation pathway by lethal toxin (LF). Experiments were repeated twice. (Western blots are courtesy of Julia Sanchez)

116

a commercially available C3 exotoxin that is modified to be readily cell permeable

(concentrations ranging 0.5-2.5 μg.mL-1) for 4 hours. Concentrations as low as 0.5

μg.mL-1 were sufficient to trigger caspase-1 activation and the secretion of IL-1β.

Likewise UBE2L3 depletion was observed at all concentrations of C3 toxin that induced caspase-1 and IL-1β processing (Fig 3.12). These data showed that UBE2L3 depletion was not only observed when NLRP3 inflammasomes were activated, but also when other inflammasomes, such as NLRP1 and PYRIN, were activated. Further, in both cases, it correlated with caspase-1 activation.

Figure 3.12: Pyrin inflammasome activation triggers UBE2L3 depletion Primary BMDMs (C57BL/B6) primed with 250 ng.ml-1 LPS for 3 hours followed by a 4- hour treatment with stated concentrations of recombinant C3 toxin. Immunoblots (left) of lysate UBE2L3 and β-actin, and supernatant caspase-1 and IL-1β. Schematic (right) of Pyrin activation pathway by C3 toxin. Experiments were repeated twice. (Western blots are courtesy of Julia Sanchez)

117

3.2.9 Activation of AIM2 inflammasomes triggers UBE2L3 depletion

The AIM2 inflammasome is activated by cytoplasmic dsDNA, which can be achieved by transfecting DNA into cells using Lipofectamine® 2000. To avoid translocating contaminants which may be bound to cellular derived DNA, synthetic repeats of double stranded poly(dA:dT) DNA was used. Similarly, to avoid LPS being translocated into cells which activates caspase-4, iBMDMs were primed with 1 μg.mL-

1 of the TLR2 agonist PAM3CSK4 for 2 hours. Poly(dA:dT) was then transfected into cells using Lipofectamine® 2000 and incubated for 5 hours (Fig 3.13). Analogous to

Figure 3.13: AIM2 inflammasome mediates UBE2L3 depletion Indicated iBMDMs were primed with 1 μg.mL-1 PAM3CSK4 for 2 h followed by transfection of 1 µg poly (dA:dT) for 5 hours. Immunoblot analysis (left) of UBE2L3 and β-actin in cell lysates plus caspase-1 and IL-1β in culture supernatants and schematic of AIM2 activation (right) by poly(dA:dT). Data are representative of three independent experiments.

118

ATP and nigericin treatments, poly(dA:T) stimulated caspase-1 activation and depletion of UBE2L3 in wildtype iBMDMs (Fig 3.10, lane 1). However, Asc -/- and

Casp1/4-/- cells which either lack ASC, a vital component of the AIM2 inflammasome, or caspase-1, display no active caspase-1 and retain UBE2L3 (Fig 3.13, lanes 2 and

3).

3.2.10 NLRC4 activation mediates depletion of UBE2L3 and does not require the ASC adaptor

The NLRC4 inflammasome responds to a number of bacterial molecules via the NLR family apoptosis inhibitory proteins (NAIPs) which bind ligands and facilitate the oligomerisation of NLRC4 (Tenthorey et al., 2017; Zhang et al., 2015). Humans have a single NAIP gene which is able to broadly detect bacterial T3SS inner rod and needle proteins, and bacterial flagellin (Reyes Ruiz et al., 2017; Yang et al., 2013a).

Mice possess four NAIP genes which display ligand specificity: NAIP1 binds to T3SS needles; NAIP2 to T3SS rod proteins; and NAIP5 and 6 which bind flagellin (Kofoed and Vance, 2011; Yang et al., 2013a; Zhao et al., 2011). All NAIPs bind and activated the NLRC4 inflammasome (Duncan and Canna, 2018).

Because NAIPs directly sense bacterial components the NLRC4 inflammasome plays an important role in antimicrobial immunity, particularly against intracellular bacterial pathogens. Attempts to stimulate NLRC4 biochemically by transfection of flagellin yielded poor caspase-1 activation and unreliable results.

Salmonella enterica serovar Tyhphimurium (STm) has been shown to activate

NLRP3 and NLRC4 inflammasomes (Broz et al., 2010a; Man et al., 2014; Qu et al.,

2016). However, STm grown under high SPI-1 inducing conditions has been shown to rapidly and specifically activate NLRC4 in mouse macrophages (Mariathasan et al.,

2004; Miao et al., 2010a; Shenoy et al., 2012). By using STm, not only was I able to

119 study bacterial induced UBE2L3 depletion, but also investigate its requirement for components of the NLRC4 inflammasome. To induce high SPI-1 expression

Salmonella strains were grown overnight at 37°C without shaking in LB containing 300 mM NaCl. Bacterial grown overnight were subcultured into fresh LB-salt medium and grown with shaking to an OD600 of 0.9-1.2.

Prior to infection, iBMDMs were primed with LPS for 3 hours. Wildtype iBMDMs were infected with wildtype STm at an increasing multiplicity of infection (MOI) (5-50) and a ΔprgH strain which lacks PrgH, a core component of the SPI-1 secretion apparatus and fails to activate the inflammasomes (Fig 3.14A). Immunoblot analysis shows that infection at MOIs of 10, 25 and 50 triggers caspase-1 activation (Fig 3.14A, lanes 3 to 5)). However, despite there being active caspase-1, infection at MOI 10 was insufficient to trigger UBE2L3 depletion (Fig 3.14A, lane 3). Infection with MOIs of 25 and 50 triggered more robust caspase-1 activation which was enough to cause the depletion of UBE2L3 (Fig 3.14A, lanes 4 and 5). Inflammasome activation, and notably, the loss of UBE2L3 occurred in a SPI-1 dependent manner as the ΔprgH STm was unable to elicit either (Fig 3.14A, lane 6).

To explore the host requirements for mediating STm-induced UBE2L3 depletion I utilised iBMDMs lacking components of the NLRC4 inflammasome. STm were grown as before to induce high SPI-1 expression which specifically activate

NLRC4 inflammasomes as confirmed by loss of caspase-1 activation in LPS-primed

Nlrc4-/- cells (Fig 3.14B, lane 2). Following LPS priming, wildtype, Nlrc4 -/-, Asc-/- and

Casp1/4-/- iBMDMs were infected with a MOI of 40 for 3 hours (Fig 3.14B). Considering previous results, UBE2L3 depletion was expected to be dependent on the inflammasome and each of its component genes. As expected, wildtype iBMDM cells lost UBE2L3 following caspase-1 activation whilst Nlrc4 -/-, Casp1/4-/- cells, which have

120

Figure 3.14: SPI-1 and NLRC4 dependent depletion of UBE2L3 during Salmonella infection. (A) Wildtype iBMDMs primed with 250 ng.mL-1 LPS for 3 hours. Cells were left uninfected (UI) or infected with wildtype STm at stated MOIs for 3 hours, an MOI of 50 was used for infection with ΔprgH STm. Immunoblotting as indicated. (B) Specified iBMDMs primed with 250 ng.mL-1 LPS for 3 hours followed by infection with wildtype STm at MOI 40. Indicated immunoblot analysis (left) and schematic (right) of NLRC4 inflammasome activation pathways. Data are representative of experiments repeated three (A) and two (B) times.

121

no active caspase-1, retained UBE2L3 (Fig 3.14B, lanes 2 and 4). Interestingly,

Asc-/- iBMDMs were still capable of depleting UBE2L3 (Fig 3.14B, lane 3) despite there being no autoprocessing of caspase-1 as no active p10 was detected in the culture supernatant. This is likely because the CARD domain of NLRC4 can directly recruit caspase-1, yielding an unprocessed but catalytically active p45 enzyme (Broz et al.,

2010b). Asc-/- cells also secreted significantly less bioactive, p17 IL-1β compared to wildtype iBMDMs. This is in keeping with the hypothesis that p45 caspase-1 is present as it has been shown to be defective in cleaving pro-IL-1β, whilst still able to induce pyroptosis to the same extent as the conventional p20/10 enzyme (Broz et al., 2010b).

These data suggested that the active p45 variant of caspase-1 is capable and sufficient to induce UBE2L3 depletion. In summary, NLRC4 inflammasome activation resulted in UBE2L3 depletion via caspase-1, and in this context, catalytically active p45 caspase-1 was sufficient to trigger the loss of UBE2L3.

3.2.11 Listeria monocytogenes infection activates caspase-1 and triggers

UBE2L3 depletion

So far results showed that bacterial derived molecules such as LPS, and the

C3 and LT exotoxins and whole bacteria (e.g. STm) can trigger the depletion of

UBE2L3 in a manner dependent on specific inflammasomes. Inflammasome mediated immune responses such as pyroptosis and IL-1β release play important roles in host defence however, many bacteria are capable of subverting these processes (Cunha and Zamboni, 2013). Having examined that activation of specific inflammasomes is sufficient to trigger depletion of UBE2L3 I wanted to examine whether inflammasome activation in response to bacteria which stimulated multiple inflammasomes, such as L. monocytogenes would produce the same outcome.

122

Unlike high SPI-1 expressing STm, infection with the Gram-positive Listeria monocytogenes (Lm) leads to the activation of multiple inflammasomes. Lm is detected by the NLRP3 via listeriolysin induced K+ efflux, NLRC4 via flagellin and AIM2 by release of DNA (Eitel et al., 2010; Kim et al., 2010; Warren et al., 2008; Wu et al.,

2010). As specific activation of NLRP3, NLR4 and AIM2 triggered UBE2L3 depletion it was hypothesised that Lm infection would produce the same outcome.

iBMDMs were first LPS primed (250 ng.mL-1) and infected with Lm at a MOI of

40 for three hours (Fig 3.15). As expected Lm infection of wildtype cells stimulated robust caspase-1 activation and elicited UBE2L3 depletion (Fig 3.15, lane 1). Similar to previous results, Asc-/-, Casp1/4-/- iBMDMs did not lose UBE2L3 expression and no

Figure 3.15: Listeria induced UBE2L3 depletion by ASC dependent inflammasomes. Specified iBMDMs were primed with 250 ng.mL-1 LPS for 3 hours followed by a hour infection with wildtype L. monocytogenes (Lm) at MOI 40. Immunoblot analysis (left) of UBE2L3 and β-actin in cell lysates plus caspase-1 in culture supernatants and schematic (right) of Lm sensing by inflammasomes. Data is representative of three independent experiments.

123

active caspase-1 could be detected in these cells upon Lm infection (Fig 3.15, lanes

2 and 3). Together, these data showed that inflammasome activation by complex living agonists like bacterial pathogens also elicited UBE2L3 depletion, and further suggested that UBE2L3 loss was therefore not specific to sterile inflammasome activators.

The response of human THP1 cells to bacterial infection was also tested.

Following LPS-priming, THP-1 cells were infected for 3 hours with L. monocytogenes at a range of MOIs (Fig 3.16). UBE2L3 depletion in response to Lm showed a

Figure 3.16: UBE2L3 depletion by Listeria infection is dependent on caspase-1 activity and listeriolysin. PMA-differentiated human THP-1 cells were primed with 250 ng.mL-1 LPS for 3 hours. Cells were left uninfected or infected with wildtype L. monocytogenes at stated MOIs for 3 hours. Ac-YVAD-fmk (50 µM) was added 20 minutes prior to infection with wildtype Lm at an MOI 80, Δhly Lm were used at an MOI of 80. Immunoblot analysis of UBE2L3 and β-actin in cell lysates plus caspase-1 in culture supernatant. Experiments were repeated three times

124 dependence on the strength of caspase-1 activation with only 50% depletion occurring following infection at an MOI of 20 and complete depletion occurring when infected with MOIs of 40 and 80. Previous studies have shown that Listeria mediated inflammasome activation requires the pore forming toxin listeriolysin (encoded by hlyA) which mediates potassium efflux (and therefore NLRP3 activation) and phagosomal escape which is require for AIM2 activation (Kim et al., 2010). Thus, bacteria lacking hlyA fail to activate inflammasomes upon infection of macrophages

(Kim et al., 2010).

Consistent with these reports, ΔhlyA Lm, used at am MOI of 80, failed to activate caspase-1 or induce UBE2L3 depletion (Fig 3.16, lane 6). To ensure that in human macrophages this process was indeed dependent on caspase-1 (as with mouse macrophages) cells were also treated with the caspase-1 specific inhibitor ac-

YVAD-fmk prior to infection. Addition of ac-YVAD-fmk (50 μM) 20 minutes prior to infection strongly inhibited caspase-1 activation and likewise prevented the loss of

UBE2L3 (Fig 3.16, lane 5). Analogous to murine iBMDMs, in human THP-1 cells,

UBE2L3 depletion induced by Lm infection was dependent on caspase-1 activity.

3.2.12 UBE2L3 depletion in human and mouse macrophages is dependent on caspase-1 activity

The results so far showed that UBE2L3 depletion in response to inflammasome agonists and displayed an absolute requirement for the sensory, structural, and responsive components of inflammasomes. Genetic studies in mouse iBMDMs revealed that UBE2L3 loss was strictly dependent on caspase-1 as Casp1/4-/- cells failed to mediate UBE2L3 depletion in response to multiple inflammasome agonists.

To confirm that loss of UBE2L3 specifically required caspase-1 catalytic activity and not a scaffolding function of caspase-1, mouse iBMDMs (Fig 3.17A) and

125 differentiated human THP-1 cells (Fig 3.17B) were treated with ATP or nigericin, respectively, in the presence of the irreversible capsae-1 inhibitor ac-YVAD-fmk. As before stimulation of NLRP3 in both cell types triggered the loss of cellular UBE2L3 following the activation of caspase-1 (Fig 3.17A and B, lane 2).

By using the caspase-1 specific inhibitor ac-YVAD-fmk I was able to corroborate that blocking caspase-1 was protective for UBE2L3 (Fig 3.17A and B, lane

3). The presence of caspase-1 protein alone therefore was not sufficient to mediate the loss of UBE2L3, and caspase-1 catalytic activity was specifically required to trigger its depletion.

Figure 3.17: Loss of UBE2L3 is dependent on caspase-1 activity in human and mouse cells. (A) Wildtype mouse iBMDMs were primed with 250 ng.mL-1 LPS for 3 hours and subsequently left untreated or pulsed with 5 mM ATP for 1 hour in the absence or presence of 10 µM Ac-YVAD-fmk. Data represent experiments repeated three times. (B) PM- differentiated human THP-1 cells were primed in 250 µg.mL-1 LPS for 3 hours and subsequently left untreated or pulsed with 10 µM Nigericin 1 hour in the absence or presence of 50 µM Ac-YVAD-fmk. Data represent experiments repeated twice. Ac-YVAD-fmk was added 20 minutes prior to ATP/nigericin treatment. Immunoblot analysis (A and B) of UBE2L3 and β-actin in cell lysates and caspase-1 in culture supernatants.

126

3.3 Discussion The results presented in this chapter identified UBE2L3 as a novel target of the cysteine protease caspase-1 which was lost from human and mouse macrophages following inflammasome and caspases-1 activation. PROTOMAP based shotgun proteomics was used to obtain the cellular proteomes of murine macrophages prior to and following caspase-1 activation providing us with a comprehensive profile of putative caspase-1 targets. An advantage to PROTOMAP analyses is that it can detect and report on changes in proteins regardless of whether they are proteolytically cleaved. Indeed, this analysis revealed that UBE2L3 was lost from cell lysates following inflammasome activation. Western blot corroborated this and showed rapid depletion of UBE2L3 protein, with no apparent cleavage products.

This may indicate several potential scenarios. Much like pro-IL-1β or pro-IL-1α

UBE2L3 may be secreted from cells following inflammasome activation either with or without direct cleavage by caspase-1. Alternatively, UBE2L3 may be directly cleaved caspase-1 and rapidly degraded. Examples of this scenario exist for the apoptotic caspases 3 and 7 which have multiple substrates which are degraded by the proteasome (Demontis et al., 2006; Semple et al., 2007) and autophagy (Dolan and

Johnson, 2010) following their cleavage. Further, the caspase-1 substrate aldolase has also been shown to be degraded following inflammasome activation in a caspase-

1 dependent manner (Shao et al., 2007).

Alternatively, the depletion of UBE2L3 could be a result of passive loss due to cell lysis during pyroptosis. Results suggested that this is not the case as inhibiting cell lysis with glycine did not block or affect the depletion of UBE2L3. This could be explored further in the context of GSDMD, the executioner of pyroptosis, to test

127 whether the loss of UBE2L3 displays a dependence on GSDMD induced pore formation and cell death.

Finally, the loss of UBE2L3 may result from being indirectly targeted by another caspase-1 substrate, which in turn prompts loss of UBE2L3, for example via ubiquitylation and subsequent degradation by the proteasome. Indeed, E3 ligases such as; NEDD4 (Harvey et al., 1998), PARKIN (Kahns et al., 2003), RNF4 and

HUWE1(Lamkanfi et al., 2008) have been reported to be substrates of caspase-1.

Cleavage of a E3 ligase could alter ligase activity, substrate preference or stability which could lead to destabilisation of UBE2L3.

Regulation of E2 conjugating enzyme stability by an E3 ligase has been previously described. The ubiquitin-editing enzyme A20 is known to regulate the stability of the E2 conjugating enzymes UBE2N and UBE2D1 by disrupting their interactions with cognate E3 ligases TRAF6, TRAF2 and cIAP1. This causes ubiquitin to accumulate on the E2 enzymes which targets them to the proteasome for degradation (Shembade et al., 2010). Such a mechanism could be postulated for

UBE2L3 where by caspase-1 disrupts its interaction with a cognate E3 ligase leading to its destabilisation. Indeed, UBE2L3 has been shown to be depleted from HeLa cells during S phase of the cell cycle. This process is also dependent on the proteasome though the precise mechanism remains unclear (Whitcomb et al., 2009).

Cleavage of an E3 ligase could drastically affect its activity. As previously mentioned the E3 ligase PARKIN has been shown to be cleaved by caspase-1 at aspartic acid residue D126, leading to its inactivation and the disruption of mitophagy

(Yu et al., 2014). Interestingly cleavage at this residue would remove the ubiquitin-like

(UBL) domain of PARKIN whilst leaving the catalytic domains intact.

128

The intramolecular interactions between the UBL and catalytic domains of

PARKIN keep it in an autoinhibited state which is regulated by addition or removal of posttranslational modifications (Kondapalli et al., 2012; Shiba-Fukushima et al., 2012).

Studies have shown that mutations within or removal of the UBL domain disrupts PARKIN autoinhibition making the enzyme hyperactive and unstable due to excessive autoubiquitylation (Chaugule et al., 2011; Trempe et al., 2013).

Though caspase-1 has not been shown to regulate PARKIN via this process, it could be hypothesized based on literature how the cleavage and destabilisation of an

E3 ligase could drastically alter its activity which, in turn, could affect the stability of an interacting E2 enzyme, akin to the role of A20 (Duong et al., 2015). Similarly, in the widely used model plant Arabidopsis thaliana a peptidase called DA1 regulates its own activation via two of its substrates creating a negative feedback loop (Dong et al.,

2016). DA1 becomes active following ubiquitylation by two RING E3 ligases, Big

Brother and DA2. Big Brother and DA2 are then themselves cleaved by DA1 leading to their destabilisation and thereby serving to negative regulate its own activation

(Dong et al., 2016). Results also showed that UBE2L3 depletion occurred in a strictly caspase-1 dependent manner as UBE2L3 was preserved following inflammasome activation in Casp-1/4-/- murine macrophages or by the addition of the caspase-1 inhibitor Ac-YVAD-cmk.

In macrophage cells, UBE2L3 was a common target of different inflammasomes, much like the recently identified GSDMD. However, studies have shown that pyroptosis is not an absolute outcome of inflammasome activation in certain immune cell types. For example, neutrophils have been shown not to undergo pyroptosis following activation of NLRC4, AIM2 (Chen et al., 2014), and NLRP3 (Heilig et al.; Karmakar et al., 2015; Karmakar et al., 2016; Martin-Sanchez et al., 2016)

129 despite inducing caspase-1 activation, IL-1β processing and secretion. Dendritic cells have likewise been shown not to undergo pyroptosis in response to non-canonical inflammasome activation with LPS and host derived oxidized phospholipids (oxPAPC)

(Zanoni et al., 2016) as have human monocytes following activation of an alternative

NLRP3 pathway activated by TRIF, RIPK1, FADD and caspase-8 (Gaidt et al., 2016).

In mouse macrophages, GSDMD processing has recently been shown to be required for lysis-independent release of IL-1(Evavold et al., 2017; Heilig et al., 2017).

It would be interesting to explore whether the depletion of UBE2L3 occurs in theses cell types under non-pyroptosis inducing conditions.

Results showed that UBE2L3 was completely lost from the cell following inflammasome activation, one question that follows therefore was why is UBE2L3 depleted? It could be that UBE2L3 may function in a process which negatively regulates inflammasomes or caspase-1-mediated processes. Depletion by caspase-1 would therefore serve to elevate this negative regulation and further propagate caspase-1 effector mechanism. Indeed, ubiquitylation plays important roles in positively and negatively regulating inflammasomes at all levels of activation (see

Chapter 6, section 6.1), including NLRP3 and ASC assembly, caspase-1 auto processing and IL-1β processing (Bednash and Mallampalli, 2016; Kattah et al., 2017;

Yang et al., 2017). Depleting a single E2 conjugating enzyme could potentially affect the activity of hundreds of E3 ligases affording caspase-1 a significant amount of control over the ubiquitylation system.

As UBE2L3 depletion showed no specificity to a particular inflammasome or agonist it would stand to reason that any role it plays would be broad, perhaps downstream of caspase-1, and therefore is more likely to be involved in mediating key caspase-1 effector functions like cytokine secretion or pyroptosis. Some of these

130 aspects on how and why UBE2L3 may deplete have been addressed in experiments described in Chapter 4 and Chapter 5 respectively.

3.4 Conclusions In this chapter, I explored the rapid loss of the E2 conjugating enzyme UBE2L3 following inflammasome activation. Immunoblot analysis confirmed previous proteomics data which showed that UBE2L3 was depleted in mouse macrophages treated with LPS and nigericin. Different NLRP3 agonists such as ATP, nigericin and alum are capable of triggering UBE2L3 depletion. This process was dependent on

NLRP3 activation by signal 1 and signal 2 and could be blocked by inhibiting NLRP3 with KCl. UBE2L3 depletion appeared to be independent of pyroptotic cell death as blocking cell lysis with glycine did not affect the dynamics or extent of UBE2L3 loss.

Multiple immune cell types such as murine primary BMDMs and BMDCs and human

THP-1 cells can deplete caspase-1. Activation of the NLRP1b, PYRIN, NLRC4, AIM2 and the non-canonical NLRP3 inflammasomes by various agonists, including bacterial infection, all led to UBE2L3 depletion. Genetically this process was dependent on core inflammasome components, including the adaptor protein ASC and caspase-1 as

UBE2L3 was retained in cells which lacked caspase-1 or had caspase-1 inhibited by

Ac-YVAD-fmk. Taken together, UBE2L3 depletion was a common occurrence following the activation of multiple inflammasomes. This process was not a passive consequence of cell lysis during pyroptosis but appeared to be an active process that required the activity of caspase-1.

131

Chapter 4 – Elucidating the mechanism of caspase-1-mediated depletion of UBE2L3

132

4.1 Introduction The previous chapter described UBE2L3 as a novel target of caspase-1 which was rapidly depleted from phagocytic cells following inflammasome activation. The loss of UBE2L3 was common to multiple inflammasomes and agonists and therefore a general feature of inflammasome activation by diverse agonists including bacterial pathogens. This process specifically required caspase-1 catalytic activity, therefore a non-proteolytic scaffolding function of caspase-1 was not sufficient to deplete cellular

UBE2L3 protein. As caspase-1 is best known for its protease activity, it was hypothesised that it directly influenced a protein through proteolytic cleavage.

Caspase-1 dependent cellular processes are therefore, in the first instance, triggered or mediated via the direct cleavage of a substrate protein such for example GSDMD and the execution of pyroptosis (Aglietti et al., 2016; Ding et al., 2016; Kayagaki et al.,

2015; Shi et al., 2015). Alternatively, there are also caspase-1 targets, proteins which are affected in a caspase-1 activity-dependent manner but not via direct cleavage.

These include alarmins such as pro-IL-1α and HMGB1 which are secreted in a caspase-1-dependent manner but are not directly proteolysed (Afonina et al., 2015;

Lotze and Tracey, 2005). The depletion of UBE2L3 could therefore be a result of direct or indirect actions of caspase-1.

The list of caspase-1 substrates stands at over 100 proteins (Denes et al.,

2012) and includes diverse proteins with varying functional roles within the cell. This is believed to be due to concentration-dependent promiscuity which caspase-1 displays (Walsh et al., 2011). Broadly speaking proteolytic cleavage of a protein can produce two outcomes; gain of function or loss of function.

The classical substrates of caspase- 1 are the pro-forms of IL-1β and IL-18.

Pro-IL-1β is cleaved at two sites, Asp26 and Asp116, while IL-18 is cleaved solely at

133

Asp36 (Afonina et al., 2015). Cleavage at these sites removes the pro-domain and releases the bioactive portion of each cytokine (Gu et al., 1997; Mosley et al., 1987b)

More recently, the newly characterised substrate GSDMD was discovered to mediate pyroptotic cell death (Kayagaki et al., 2015; Shi et al., 2015). Under resting conditions it exists in an autoinhibited conformation via interactions between its N- and

C-terminal domains (Ding et al., 2016). Cleavage by caspase-1 at Asp275 removes the

C-terminus and alleviates autoinhibition. This frees the N-terminal region of the protein to oligomerise, form membrane pores and induce cell death (Kayagaki et al., 2015;

Shi et al., 2015).

Alternatively, cleavage by caspase-1 can cause the loss of function of a protein.

In response to infection with Staphylococcus aureus caspase-1 associates with bacterial containing phagosomes to enhance bacterial killing. It does this by cleaving phagosome-associated proteins Rac1 and gp91 phox which limits the activity of the

NADPH oxidase (NOX2). This results in heightened acidification of the phagosome and enhanced bactericidal activity (Sokolovska et al., 2013). Likewise, the DNA damage-sensor and repair enzyme poly(ADP-ribose) polymerase 1 (PARP1) is known to be cleaved and inactivated by multiple caspase enzymes, including caspase-1

(Chaitanya et al., 2010; Lazebnik et al., 1994; Malireddi et al., 2010). This is thought to help pressure cellular energy stores during pyroptosis and enhance cell death induction by promoting DNA damage (Malireddi et al., 2010). The peroxisome proliferator-activated receptor gamma (PPARγ) is also cleaved by caspase-1 in tumour associated macrophages. Cleavage at Asp64 generates a truncated protein which is preferentially translocated to cell mitochondria. Here it can interact and thereby attenuate the activity of medium-chain acyl-CoA dehydrogenase (MCAD) enzyme which promotes macrophages differentiation (Niu et al., 2017). In these

134 examples, targeting by caspase-1 results in cleavage and the generation of clearly defined fragments some of which are retained in the cell and remain detectable in cell lysates.

Caspsae-1 can, however, also mediate the loss of proteins in a fashion more akin to the observed effects on UBE2L3. This is most commonly seen with indirect targets whose release via unconventional secretion is mediated by caspase-1. This is the case for the non-substrate alarmins HMGB1 and pro-IL-1α which are released from cells in a caspase-1-dependent manner. The underlying mechanisms that control these events are controversial and poorly characterised. Multiple studies have suggested different potential mechanisms including cell death (Cullen et al., 2015;

Martin, 2016), autophagy (Harris et al., 2011b), microvesicle release (Qu et al., 2007), and caspase-1-induced pore formation (Fink and Cookson, 2006). GSDMD pores are now believed to act as conduits for the release of bioactive IL-1β and IL-1α. However, it remains unclear whether IL-1α requires cleavage in order to pass through the pore

(Evavold et al., 2017; Heilig et al., 2017).

Alternatively, caspase-1 dependent processes could control the overall stability of a protein. Though such processes have not been described for caspase-1, other caspase enzymes have been shown to trigger substrate degradation via various processes following their cleavage. The transcription factor Twist is ubiquitylated and degraded by the proteasome following cleavage by caspase-3/7 (Demontis et al.,

2006). Similarly, the microtubule-associated protein Tau is cleaved by caspases (de

Calignon et al., 2010). Cleaved Tau forms large aggregates which are then subsequently ubiquitylated and degraded by autophagy (Dolan and Johnson, 2010).

The Drosophila protein Diap1 is cleaved by the drICE and Dcp-1 caspases, which exposes an unstable N-terminal sequence, known as a N-degron, which is

135 ubiquitylated and targeted to the proteasome for degradation (Ditzel et al., 2003).

Similar observations have been made of PKCδ and others during apoptosis

(Varshavsky, 2011). Ubiquitin-independent degrons such as those found p21 can be directly recognised by the 20S core component of the proteasome independently of ubiquitin (Chen et al., 2007). Others such as the PEST-motif have also been described to mediate proteasomal degradation after being liberated following caspase cleavage

(Belizario et al., 2008).

Additionally, caspase-1-dependent cytokines can also induce protein degradation. Much like TLR signalling, stimulation of the IL-1 receptor (IL-1R) or IL-

18 receptor (IL-18R) induces the degradation of proteins such as IκBα (Newton and

Dixit, 2012). Interestingly, IL-1 and TNF signalling have been shown to induce the proteasomal degradation of the E2 conjugating enzymes UBE2N and UBE2D3 via interactions with the ubiquitin-editing enzyme A20 (Shembade et al., 2010). This suggests that cell-extrinsic indirect effects may affect protein degradation after caspase-1 activation.

Many caspase-1 substrates possess their own function or catalytic activity; therefore, their cleavage could potentially have knock-on effects which impact many cellular processes. Caspase-1 therefore has the potential to indirectly impact the fate of many proteins such as UBE2L3 independently of cell death or cytokine signalling. As an example, the E3 ligase PARKIN has also been shown to be cleaved by caspase-1 at Asp126 (Kahns et al., 2003; Yu et al., 2014). Widescale proteomic analysis identified more than 200 proteins which were ubiquitylated by PARKIN following mitochondrial depolarisation (Sarraf et al., 2013). These were mostly mitochondrial associated proteins, but also included cytosolic, membrane bound and

136 nucleolar proteins. Thus caspase-1-mediated inactivation of PARKIN could indirectly impact the degradtion and/or functionality of hundreds of proteins (Sarraf et al., 2013).

The aim of this chapter was to determine the mechanism underlying UBE2L3 depletion by caspase-1 by exploring well understood and common mechanisms, including cleavage and secretion, as well as more poorly defined secondary effects which may be impacted or controlled by caspase-1.

4.2 Results 4.2.1 UBE2L3 is not cleaved by caspase-1 in vitro or in cells.

UBE2L3 depletion following inflammasome activation was dependent on caspase-1. Specifically, caspase-1 catalytic activity was needed as addition of the caspase-1 specific inhibitor Ac-YVAD-fmk prevented UBE2L3 depletion. The requirement for caspase-1 was therefore as a protease and not just as a signalling scaffold. It was hypothesised that UBE2L3 may be a proteolytic substrate of caspase-

1 which then led to its depletion. In silico prediction using multiple online resources

(e.g. ExPASy PeptideCutter (https://web.expasy.org/peptide_cutter/), Protter

(http://wlab.ethz.ch/protter/start/) and CasCleave (http://sunflower.kuicr.kyoto- u.ac.jp/~sjn/Cascleave/index.html)) (Gasteiger et al., 2003; Omasits et al., 2014; Song et al., 2010) suggested a putative cleavage site at aspartic acid residue 124 within a four amino acid caspase-recognition sequence LRAD124. Because UBE2L3 was rapidly depleted, any apparent cleavage event or the appearance of cleaved products were difficult to examine. Further, it was not possible to determine the antigenic region of commercially available polyclonal antibodies (GeneTex – GTX104717, SCBT – sc-

47549) used to detect endogenous UBE2L3. To overcome this limitation, cleavage of

UBE2L3 by caspase-1 was first assessed in vitro. Recombinant human caspase-1 p20

(aa 120-297) and p10 (aa 317-404) components were purified and combined to yield

137 active caspase-1 (provided by Dr Avinash Shenoy and Gil Ferreira Hoben). The resulting enzyme was then incubated with GST- tagged wildtype UBE2L3, or a cleavage resistant mutant variant where the putatively cleaved aspartic acid 124 had been mutated to an asparagine (UBE2L3D124N) residue (Figure 4.1A). Concurrently, to confirm the activity of the purified enzyme recombinant caspase-1, it was incubated with GST-tagged human IL-1 (aa 110- 269) to serve as a positive control (Figure

4.1B). GST-UBE2L3 variants migrated at an approximate molecular mass of 45 kDa.

Incubation with active caspase-1 caused no shift in molecular mass, which indicated that UBE2L3 had not been cleaved (Figure 4.1A, lane 3); reactions with the addition

Figure 4.1: Recombinant UBE2L3 is not cleaved by caspase-1 in vitro. Coomassie gels of in vitro proteolysis assays using purified recombinant caspase-1 p20 and p10 subunits with GST-UBE2L3 or GST-UBE2L3D124N (~46 kDa) (A) and GST-IL-1β (45 kDa) (B) as substrates. Caspase-1 assays were performed for at 37 °C for 1 hour. Pan-caspase inhibitor zVAD-fmk was added as indicated (20 μM). Schematics of GST-fusion proteins and positions D116 and D124 cleavage sites. were used for or proteins as substrates are shown below. * indicates caspase-1 p20 which lost hexa-histidine tag. Cleavage of GST-UBE2L3 at D124 would produce a ~42 kDa protein. Cleavage GST-IL- 1β is at D116 releases GST (26 kDa) and p17 IL-1β (red arrows). Assays were performed simultaneously. Experiments were repeated twice. (Coomassie images are courtesy of Avinash Shenoy and Gil Ferreira Hoben)

138

of the pan-caspase inhibitor z-VAD-fmk or the cleavage mutant showed no difference

(Figure 4.1A, lane 4 and 5). Contrary to this, incubation with caspase-1 enzyme led to processing of recombinant IL-1β (approximately 46 kDa) generating free GST (25 kDa) and the cleaved bioactive p17 portion of the protein (Figure 4.1B, lane 2). This process could be inhibited by the addition of the pan-caspase inhibitor z-VAD-fmk as only the uncleaved protein was observed. This data showed that UBE2L3 cannot be cleaved in vitro by recombinant caspase-1 enzyme that was competent to process IL-1β.

To test whether other cellular proteins or components were required to facilitate cleavage, further experiments were carried out in HEK293 cells. HEK293 cells lack major components of the inflammasome, including caspase-1. Cells were transfected with plasmids encoding caspase-1 and various UBE2L3 proteins as described next.

Because the hypothesized cleaved C-terminal fragment of UBE2L3 would be small

(approximately 3.5 kDa), N-terminal triple FLAG and C-terminal double HA epitope tags were added (flagUBE2L3HA; schematic in Figure 4.2B). The addition of these tags would make it easier to detect cleaved fragments as they would increase the size of the fragment to 6 kDa, cause a detectable shift in the migration of UBE2L3 (decreasing from 24 to 17 kDa) and allow the use of reliable commercially available anti-tag antibodies. Additionally, removal of the C-terminus could be detected via the loss or reduction in detectable HA-tag when analysed by western blot. flagUBE2L3HA was co- expressed with mouse caspase-1 (Casp1) containing mutations which prevented self- cleavage (D296N, D308N, D313A and D314A) but retained catalytic capability as it is spontaneously active when overexpressed in cells. An inactive variant which has the catalytic cysteine mutated to alanine (Casp1C284A) served as a negative control.

Interestingly, co-transfection of Casp1 and flagUBE2L3HA generated an additional fragment which was detected using anti-FLAG antibody. This shorter,

139 presumably cleaved, protein migrated 1-2 kDa below the full length FlagUBE2L3HA

(Figure 4.2A, lane 2). This smaller fragment was absent in cells which were not transfected with Casp1 or were transfected with catalytically inactive Casp1C284A

(Figure 4.2A, lanes 1 and 3). As the shorter protein was detected with anti-FLAG antibody, the N-terminal regions were intact and suggested cleavage of C-terminal

Figure 4.2: Caspase-1 cleaves HA-tag from UBE2L3 in cells. flag HA (A) Immunoblot analysis of HEK293E expressing tagged UBE2L3 and indicated caspase-1 variants. Cleavage was detected using anti-FLAG and anti-HA antibodies. flag HA HEK293 cells were co-transfected with equal amounts of plasmids encoding UBE2L3 C284A and either constitutively active (Casp1) or catalytically inactive (Casp1 ) caspase-1. Cells were incubated for 24 hours prior to cell lysis and sample preparation. Schematic (below) of tagged UBE2L3 protein. Data is representive of two independent experiments. flag HA (B) UBE2L3 protein primary sequence is indicated in bold lettering. Triple FLAG tag (purple) and double HA tag (blue) are indicated. Possible cleavage sites are underlined, with the region around D124 shown in orange. Note the potential caspase-1-cleavage sites in the HA-tag.

140 regions. This was confirmed with anti-HA antibody immunoblots. A reduction in HA- tag was observed in cells which expr essed active caspase-1 but not inactive caspase-

1 (Figure 4.2A). Despite these observations the reduction in molecular mass seen with the “cleaved” flagUBE2L3HA was less than would be expected from its processing at the presumed D124 site. Examining the peptide sequence of flagUBE2L3HA highlighted that the final four residues of the UBE2L3 protein, RVPD and the internal DVPD residues of the HA tag resemble caspase-cleavage motifs (Figure 4.2B, underlined). Indeed, examining the literature revealed that the HA-tag has been shown to act as a substrate of apoptotic caspase-3/7 during apoptosis (Schembri et al., 2007). It was highly likely that the whole, or at least part, of the C-terminal HA tags were being lost due to cleavage by either caspase-1 or caspase-3/7, which can be activated by caspase-1 in cells (Akhter et al., 2009; Lamkanfi et al., 2008; Malireddi et al., 2010; Sagulenko et al., 2018), and not due to cleavage of the UBE2L3 protein at D124.

To avoid this potential confounding factor due to caspase-substrate-like sites in the HA-tag sequence, additional variants of UBE2L3 with other tags were tested to confirm whether UBE2L3 was cleaved by caspase-1 in cells. UBE2L3, either with a N-

(YFP-UBE2L3) or C- (UBE2L3-YFP) terminal YFP tag, or with N-terminal triple FLAG and C-terminal single StrepTag II epitope tags (WSHPQFEK; flagUBE2L3strep) were used (Figure 4.3). Transfected caspase-1 was active as co-expressed IL-1β was proteolytically processed causing the generation of IL-1β p17 (Figure 4.3C, lane 2).

Co-transfection of YFP-tagged UBE2L3 and caspase-1 showed no apparent cleavage of UBE2L3 (Figure 4.3A) when analysed by western blot with anti-GFP antibody

(Figure 4.3A).

Likewise, flagUBE2L3strep was not cleaved when co-expressed with caspase-1

(Figure 4.3B). Taken together, these experiments led to the conclusion that UBE2L3

141

Figure 4.3: UBE2L3 is not cleaved by caspase-1 in cells ® HEK293 cells were transfected using Lipofectamine 2000 with equal amounts of tagged

UBE2L3 plasmids and indicated caspase-1 plasmids. Cells were incubated for 24 hours prior to cell lysis and sample preparation. (A) Immunoblot analysis of HEK293E expressing UBE2L3 with either a N-terminal (YFP- UBE2L3) or C-terminal (UBE2L3-YFP) YFP-tag and either constitutively active (Casp1) C284A or catalytically inactive (Casp1 ). Schematic (below) show tagged UBE2L3 proteins. flag strep (B) Immunoblot analysis of HEK293E expressing UBE2L3 (left) or untagged human IL-1β (right) without or with co-transfection of caspase-1 (Casp1). Schematic (below) of tagged UBE2L3 protein .Data are representative of three experiments. (Western blots in A and B are courtesy of Julia Sanchez)

142

was not proteolytically cleaved by caspase-1 in cells. Caspase-1 is activated following its recruitment to the inflammasome speck formed by inflammasome sensors and the ASC adaptor, where it undergoes autoproteolytic activation. Microscopy- based studies have shown that pro-IL-1β is also recruited to the inflammasome speck where it may be processed into its bioactive form (Man et al., 2014; Man et al., 2013).

I therefore wanted to investigate whether UBE2L3 was recruited to the inflammasome speck following caspase-1 activation. To do this mouse macrophages were transduced using the Moloney murine leukemia virus (M-MLV) based retroviral vector pMXsIP expressing UBE2L3 with an N-terminal YFP-tag. Mouse macrophages

Figure 4.4: UBE2L3 does not localise to the inflammasome speck during inflammasome activation. Immunofluorescence analyses of UBE2L3 and ASC protein localisation in iBMDMs. YFP- UBE2L3 expressing iBMDMs were primed with LPS followed by treatment with nigericin to activate the NLRP3 inflammasome. After 30 minutes of treatment cells were fixed in 4% PFA. Cells were stained with anti-ASC antibody and anti-rabbit-Alexa-647 and nuclei were stained with Hoechst stain. Image is represented of at least 50 cells across two independent experiments. Scale bar, 10 μm.

143 stably expressing YFP-UBE2L3 were stimulated where LPS primed and treated with nigericin for 30 minutes and examined by immunofluorescence microscopy. UBE2L3 showed no co-localisation or recruitment to the ASC-speck (Figure 4.4). This data supported the notion that UBE2L3 was not a direct substrate of caspase-1 as its cleavage by caspase-1 or co-localisation with caspase-1 could not be detected.

4.2.2 Depletion of UBE2L3 is not mediated through caspase-1-mediated autocrine or paracrine signalling

In the previous chapter, results showed that the depletion of UBE2L3 was dependent on capase-1. Both genetic deletion and biochemical inhibition of caspase-

1 caused retention of UBE2L3 in response to inflammasome activation. This dependency was not direct as UBE2L3 was not a substrate of caspase-1, but was likely mediated by another caspase-1-controlled protein or process.

Shembade et al. (2010) showed that IL-1 and TNF treatment in mouse embryonic fibroblasts (MEFs), BMDMs and BDMDCs triggers the proteasomal degradation of the E2 conjugating enzymes UBE2N and UBE2D3 after 4-6 h of treatment. They showed that following treatments with these cytokines, the ubiquitin- editing complex A20 disrupts the interactions between these E2 enzymes and there cognate E3 ligases, TRAF6, TRAF2 and cIAP1. This in turn causes UBE2N and

UBE2D3 to become ubiquitylated and degraded at the proteasome (Shembade et al.,

2010). My previous observations revealed that UBE2L3 depletion temporally correlated with IL-1β secretion (Figure 3.5). It was therefore hypothesised that a similar process may occur following inflammasome activation, resulting from autocrine and/or paracrine signalling of caspase-1 dependent cytokines such as IL-1 or IL-18, explaining its indirect dependence on caspase-1.

144

To this end, primary BMDMs from IL-1R knockout (Il1r1-/-) and IL-18R knockout

(Il18r1-/-) mice were used to investigate a potential causative link between caspase-1- dependent cytokine signalling and loss of UBE2L3, as they are unable to detect IL-

-/- -/- Figure 4.5: UBE2L3 depletes in Il1r1 and Il18r1 BMDMs, and its depletion is independent of inflammasome-dependent secreted factors -/- -/- (A) UBE2L3 depletes in Il1r1 or Il18r1 pri-BMDMs primed with LPS and pulsed with ATP for 1 hour. Cells were incubated with 20 mM KCl or 5 mM glycine (as indicated) in the presence of ATP. Immunoblots of UBE2L3 and β-actin in cell lysate, and caspase-1 and IL-1β in culture supernatant are shown. (B) Wildtype iBMDMs were primed with LPS and left untreated (UT) or treated with nigericin for 1 hour. As expected, nigericin treatment induced UBE2L3-depletion. Supernatants from cells having undergone indicated treatments were transferred to LPS- -/- primed Casp1/4 cells and incubated for 1 hour. Immunoblots for UBE2L3 in cell lysates in supernatant-donor and -receiver cells are shown. Schematic (right) of experiment in B). Data are representative of two experiments.

145

1α/β and IL18 respectively. Il1r1-/- and Il18r -/- BMDMs were primed with LPS and subsequently pulsed with ATP (Figure 4.5A). Interestingly, cells of both genotypes behaved like wildtype macrophages, and displayed robust depletion of UBE2L3 following caspase-1 activation and secretion of IL-1β.

These data suggested that neither IL-1R1 nor IL-18R1 individually contributed to the loss of UBE2L3. This observation may, however, be due to redundancy between the two receptors, or the result of unknown secreted factor(s). To ensure the inclusion of all potential caspase-1-dependent secreted factors, culture supernatant from inflammasome activated BMDMs was added to LPS-primed Casp1/4-/- BMDMs. This would allow testing whether a caspase-1-dependent secreted factor, which would be present in the supernatant of wild-type BMDMs, could induce the loss of UBE2L3 from

Casp1/4-/- cells. Following a one-hour treatment with nigericin, wildtype cells displayed strong depletion of UBE2L3 (Figure 4.5B, top). However, Casp1/4-/- cells failed to deplete UBE2L3 even after incubation in spent medium from inflammasome-activated wildtype iBMDMs. These data showed that the loss of UBE2L3 was not the result of secondary signalling cascades induced by caspase-1-dependent cytokines, alarmins, lipid mediators or others.

4.2.3 UBE2L3 is not lost via GSDMD pore formation or lysis

Previous results showed that preventing pyroptotic cell lysis with glycine did not affect the depletion of UBE2L3, which suggested that the loss did not occur passively during pyroptotic cell death. Cells lacking GSDMD do not undergo pyroptosis or release processed IL-1β despite maintaining caspase-1 activity. GSDMD is now known to be the executor of pyroptosis due to its pore-forming activity following cleavage by caspase-1. GSDMD pores have been proposed to also mediate the release of cleaved IL-1β (He et al., 2015; Heilig et al., 2017; Kayagaki et al., 2015; Shi

146 et al., 2015) and IL-18 (Heilig et al., 2017), as well as the non-substrate caspase-1 target IL-1α (Evavold et al., 2017). With increasing evidence suggesting that that lysis and IL-1 secretion are uncoupled from cell death, yet still dependent on GSDMD pore formation, it was thought that perhaps UBE2L3 depletion was dependent on caspase-

1-induced GSDMD pores. To explore this, immortalised Gsdmd-/- mouse

Figure 4.6: Depletion of UBE2L3 is independent of GSDMD mediated pore formation and lysis. (A) Immunoblot analyses of UBE2L3 protein expression in wildtype and Gsdmd-/- BMDMs. (B) Gsdmd-/- BMDMs primed with 250 ng.ml-1 LPS for 3 hours and subsequently left untreated or pulsed with 10 μM nigericin for stated periods of time. Immunoblot analysis of UBE2L3, caspase-1 and IL-1β in cell lysates and supernatant. Data are representative of two experiments. (C) Graph of cell lysis measured by percentage lactate dehydrogenase (LDH) release from similarly nigericin treated cells. Plot of mean±S.D from three experiments.

147 macrophages were tested for their ability to deplete UBE2L3. Under resting conditions, the basal protein expression of UBE2L3 was unaltered in Gsdmd-/- iBMDMs and equivalent to expression in wildtype cells (Figure 4.6A). Gsdmd-/- cells were primed with LPS for 3 hours followed by treatment with 10 μM nigericin for 30 and 60-minute timepoints. Caspase-1 activation and IL-1β processing was observed within 30 minutes of treatment (Figure 4.6B).

Figure 4.7: UBE2L3 is common to multiple inflammasomes in GSDMD deficient cells. -/- -1 (A) Gsdmd iBMDMs were primed with PAM3CSK4 (1 μg.mL ) for 2 hours followed by ® transfection of 1 µg poly (dA:dT) using lipofectamine 2000 for 5 hours. Immunoblot analysis of UBE2L3, β-actin, caspase-1 and IL-1β in cell lysates. Data are representative of 2 independent experiments. -/- -1 (B) Gsdmd iBMDMs primed with 250 ng.mL LPS for 3 hours. Cells were left uninfected (UI) or infected with wildtype STm at stated MOIs for 3 hours. Immunoblot analysis of UBE2L3, β-actin, caspase-1 and IL-1β in cell lysates. Data are representative of 2 independent experiments. (Western blots and LDH analysis in A and B are courtesy of Julia Sanchez)

148

As previously described, cell death was significantly reduced in Gsdmd-/- cells as they do not undergo pyroptosis (Figure 4.6C additionally both processed caspase-

1 and IL-1β were retained in the cell cytoplasm with none detected in the culture supernatant (Figure 4.6A) (He et al., 2015; Kayagaki et al., 2015; Shi et al., 2015).

However, UBE2L3 depletion was still observed in iGsdmd-/- cells following caspase-1 activation and was not detected in the cell culture supernatant (Figure 4.6C, lanes 2 and 3). This suggested that UBE2L3 depletion was independent of GSDMD and

Figure 4.8: UBE2L3 loss in human cells is independent of GSDMD (A) PMA differentiated THP-1 cells stably transduced with GSDMD-targeting shRNA miR -1 (GSDMD ), primed with 250 µg.ml LPS for 3 hours followed treatment with nigericin for 1 hour. Immunoblot analysis of UBE2L3, β-actin, caspase-1 in cell lysates. Data are representative of two experiments. GSDMD blot needs Mw markers miR -1 (B) PMA differentiated control, or GSDMD THP-1 cells were primed with 250 µg.ml LPS for 3 hours followed by treatment with nigericin for 1 hour. Graph of cell lysis as determined by release of lactate dehydrogenase (LDH). Mean±S.E.M from three independent experiments are plotted. *P <0.05 by unpaired, two-tailed Student’s t-test. Blots miR (below) show GSDMD expression in control or GSDMD THP-1 cells. (Western blots are courtesy of Julia Sanchez)

149

GSDMD-induced cell death and pore formation. UBE2L3 was likewise depleted in

Gsdmd-/- macrophages following AIM2 activation by Pam3CSK4 priming and subsequent cytosolic deliver of poly(dA:dT) by transfection (Figure 4.7A). Similarly, depletion of UBE2L3 was observed in LPS-primed Gsdmd-/- cells infected with high-

SPI-1 induced STm which activated NLRC4 inflammasomes (Figure 4.6B).

I then decided to test whether UBE2L3 depletion in human THP1 cells was also independent of GSDMD-driven pyroptosis. Silencing of GSDMD expression in THP1 cells was achieved using a GSDMD-specific miR30E plasmid. As expected, GSDMD- silencing (THP1GSDMDmiR) efficiently blocked nigericin induced pyroptotic cell lysis

(Figure 4.7B). Importantly, GSDMD was also dispensable for UBE2L3 loss. Thus, as with mouse iGsdmd-/- cells, differentiated THP1GSDMDmiR cells depleted UBE2L3 following caspsase-1 activation, which pointed to pyroptosis-independent and cell- intrinsic mechanisms underlying its loss from cells in human and mouse cells (Figure

4.7A.

4.2.4 Caspase-1 mediates cell-intrinsic, proteasome-dependent depletion of UBE2L3.

Cleavage or secretion are the typical consequences of caspase-1 targeted proteins. The depletion of UBE2L3, though dependant on caspase-1 activity, was not due to its proteolytic cleavage, caspase-1 dependent autocrine/paracrine signalling or

GSDMD mediated secretion. As a ubiquitin conjugating enzyme, UBE2L3 functions in autophagy and the ubiquitin-proteasome system (UPS), two major pathways which regulate protein degradation and turnover in cells. Therefore, I considered whether

UBE2L3 was being degraded by a ubiquitylation-dependent process following caspase-1 activation. The depletion of UBE2L3 was a rapid process occurring within one hour of inflammasome activation with robust agonists such as ATP or nigericin. I

150 therefore wanted to examine the native stability of UBE2L3 to determine if it was degraded under resting conditions.

THP1 cells were treated with cycloheximide over increasing periods of time.

UBE2L3 was shown to be a stable protein with a half-life greater than 8 hours in resting cells (Figure 4.9). Further, the addition of the pro-inflammatory stimulus LPS did not markedly affect UBE2L3 stability (Figure 4.9). The fact that UBE2L3 was inherently stable suggested that UBE2L3 depletion was actively triggered following caspase-1 activation and did not merely result from a shut-off or reduction in the production of new UBE2L3 protein.

Autophagy is a process which mediates protein turnover, requiring the encapsulation of proteins and cargo in autophagosomes before subsequent degradation. The protein ATG7 is an essential component of the autophagy pathway by functioning in the elongation of autophagosomal membranes (Mizushima et al.,

2011). Previous studies have shown that cells possessing genetic deletions of Atg7 are deficient in autophagy (Komatsu et al., 2005). To test whether UBE2L3 depletion required autophagy, the expression of Atg7 was silenced in iBMDMs by stably

Figure 4.9: UBE2L3 is stable under non-inflammasome activating conditions. Differentiated THP1 cells were left unprimed or primed with LPS for 3 hours. Cells were -1 subsequently treated with 20 µg.mL cycloheximide (CHX) for indicated time. Immunoblots show lysate UBE2L3 and β-actin. Data is presentative of 3 experiments.

151 expressing Atg7-targeting miRNA30E-based shRNA (Figure 4.10). Immunoblots confirmed that the expression of Atg7 was significantly reduced via stable silencing

(Figure 4.10B). However, UBE2L3 depletion proceeded similarly in control and Atg7- silenced cells treated with LPS and nigericin (Figure 4.10A); Atg7-silencing did not affect nigericin-induced caspase-1 activation (Figure 4.10A). This indicated that autophagy was dispensable for caspase-1-dependent loss of cellular UBE2L3 protein.Next, the proteasome was tested using MG-132 and epoxomicin to inhibit the

β-subunits of the proteasome and assess the effect on UBE2L3 loss (Figure 4.11A).

Differentiated THP-1 cells were primed with LPS, followed by a 1-hour treatment with nigericin without and with prior treatment with inhibitors. Five minutes

Figure 4.10: UBE2L3 depletes in Atg7 silenced cells and is independent of autophagy. (A) iBMDMs stably expressing non-targeting control (Ctrl) or Atg7-specific shRNA were LPS primed for 3 hours followed by treatment with nigericin for 1 hour. Immunoblots show UBE2L3 and β-actin in cell lysate, and caspase-1 culture supernatant. (B) Immunoblots of Atg7 protein expression in iBMDMs stably expressing non-targeting control (NT) or Atg7-specific shRNA. Data are representative of two experiments.

152 prior to the addition of nigericin, cells were either left untreated or treated with 20 μM

MG132 or 5 μM epoxomicin. As seen previously, THP-1 cells with active caspase-1 effciently depleted UBE2L3 (Figure 4.11A, lane 2). Remarkably, the addition of either

MG132 or epoxomicin prevented UBE2L3 depletion (Figure 4.11A, lane 3 and 4). This effect was independent of secondary effects on caspase-1 activation as the level of active p20 detected was equivalent in control and inhibitor-treated cells. These data showed that proteasomal activity was required for caspase-1-dependent loss of

UBE2L3. To ensure that this mechansim was common across different cell types, iBMDMs and primary BMDCs tested for caspase-1-mediated UBE2L3 depletion in the presence of proteasome inhibitors. As with THP1 cells, the addition of MG132 five- minutes before ATP prevented UBE2L3 depletion despite robust activation of

Figure 4.11: Caspase-1-dependent UBE2L3 depletion requires proteasomal activity. (A) Differentiated THP1 cells were primed with LPS for 3 hours and left untreated or pulsed with nigericin for 1 hour. After 5 minutes of nigericin treatment; DMSO, or proteasome inhibitors MG132 (20 µM) or Epoxomicin (5 µM) were added to the cells for the remaining period of treatment. (B) iBMDMs (left) and pri-BMDCs (right) were LPS primed and subsequently remained untreated or treated with 5 μM ATP. DMSO, or MG132 (20 µM) was added 5 min after ATP addition. Experiments (A and B) were repeated 3 times

153 caspase- 1 (Figure 4.11B). The proteasome was therefore required for UBE2L3 depletion following inflammasome and caspase-1 activation.

4.2.5 Lysine residues are required for UBE2L3 protein stability but are dispensable for caspase-1-mediated degradation.

The proteasomal degradation of cellular proteins typically requires them to be polyubiquitylated, which serves as a degradative tag that targets the protein to the proteasome. Degradative ubiquitylation is most commonly ligated to lysine the residues of a protein, and typically as K48-polyubiquitin chains. UBE2L3 has 18 lysine residues, 14 of which have previously been shown to be ubiquitylated in non- phagocytic cells and cancer cell lines through the use of di-glycine IP-based proteomic analyses (Hornbeck et al., 2015). This process uses liquid chromatography-tandem mass spectrometry (LC-MS) to identify ubiquitin modified residues based on the presence of a residual di-glycine tag which is left on lysine residues following tryptic digestion of ubiquitylated peptide. This gives modified residue additional mass (114

Da) which can be detected by LC-MS (Na and Peng, 2012).

I anticipated that UBE2L3 was ubiquitylated prior to its proteasomal degradation. To explore whether ubiquitylation of UBE2L3 lysine residues was required for caspase-1-triggered degradation, a variant of UBE2L3 lacking all lysines was generated and stably expressed in mouse macrophages. Such a protein lacked all 18 lysine residues which were mutated to arginine residues to generate UBE2L318R protein, which presumably cannot be ubiquitylated. UBE2L318R was cloned with an N- terminal 3xFLAG and C-terminal 2xStrep-II tags to enable it to be detected and distinguished from endogenous UBE2L3. Tagged UBE2L318R was retrovirally transduced for stable expression in iGsdmd-/- mouse macrophages and experiments.

154

I first observed that under resting conditions the expression of UBE2L318R was significantly lower than that of tagged wildtype UBE2L3 (Figure 4.12, lanes 1 and 3).

This difference was due to the increased degradation of UBE2L318R as addition of

MG132 for 6 hours increased UBE2L318R protein to a level equivalent to that of wildtype UBE2L3 (Figure 4.12, lane 4). Additionally, though not as marked, wildtype tagged UBE2L3 levels also increased upon MG132 treatment. However, MG132 did not appear to affect the levels of endogenous UBE2L3 during the 6 h treatment time- frame (Figure 4.12). This suggested that one or more lysine residues were important for mediating the turnover of UBE2L3. This data suggested that it was plausible that basal ubiquitylation of UBE2L3 may have a stabilising effect in cells. It is worth noting that UBE2L318R was natively folded and fully catalytically active as reported previously

Figure 4.12: Lysineless UBE2L3 is degraded by the proteasome under resting conditions. -/- flag strep iGsdmd cells stably expressing either wildtype (WT) or lysineless (18R) UBE2L3 variants were left untreated or treated with 10 µM MG132 for 6 hours. Immunoblots show flag strep UBE2L3 detected with anti-Flag antibody, endogenous UBE2L3 and β-actin. Data is representative of three independent experiments.

155

(Hospenthal et al., 2013) and from my own experiments (Chapter 6), which suggested that the reduced stability of UBE2L318R was less likely to be due to protein misfolding.

To further characterise UBE2L318R degradation in cells, additional variants were generated to assess the contribution of the catalytic activity to UBE2L3 turnover.

In addition to wildtype and lysine-less UBE2L3 proteins, the catalytically inactive C86A and E3-binding impaired P62A/F63A mutants were tested. As before, iGsdmd-/- cells stably expressing each of these variants were left untreated, treated alone with MG132 or the autophagy inhibitor bafilomycin A1 (Figure 4.13A). As before, UBE2L318R was less stable than wildtype UBE2L3 in untreated cells but accumulated in the presence of MG132 (Figure 4.13A, lane 2). Interestingly, bafilomycin A1 did not increase

UBE2L318R levels (Figure 4.13A, lane 3), which suggested that the proteasome predominantly mediated UBE2L3 turnover. Similarly, immunoblotting showed that

UBE2L3P62A/F63A was significantly poorly expressed as compared to wildtype UBE2L3

(Figure 4.13A, lane 10).

This was again due to increased protein turnover via the proteasome as the addition of MG132, but not bafilomycin, increased UBE2L3P62A/F63A protein levels

(Figure 4.13A, lanes 11 and 12). However, UBE2L3C86A levels were similar to wildtype

UBE2L3 but, its levels did not increase following MG132 treatment (Figure 4.13A, lanes 7 to 9). Catalytic C86 residue or E3-binding P62/F63 residues, and the 18 lysine residues (which did not affect catalytic activity) had significant and opposing influences on the stability of UBE2L3. This suggested that catalytic activity and E3 binding contributed to UBE2L3 turnover. I next tested whether the introduction of the C86A mutation in the UBE2L318R protein would increase its stability. Indeed, western blot analysis showed that UBE2L318R/C86A was more stable than UBE2L318R (Figure 4.13B, lanes 7 and 8). I also tested the stability of a UBE2L3 variant lacking the 14 lysines

156 which have been previously shown to carry ubiquitin in cells (UBE2L314R; see

Appendix table 1 for residue numbers). Expression of UBE2L314R was higher compared to UBE2L318R (Figure 4.13B, lane 3).

However, UBE2L314R levels were comparable to the wildtype protein only upon the addition of MG132 (Figure 4.13B, lane 6). These data showed that under resting conditions, the proteasomal-turnover of UBE2L3 was controlled by many factors including lysine availability, catalytic activity and E3-binding which may have opposing

Figure 4.13: Lysine residues, catalytic activity and E3-binding residues impact UBE2L3 stability in macrophages flag strep Immunoblots show UBE2L3 detected with anti-Flag antibody and β-actin. -/- flag strep (A) iGsdmd cells stably expressing indicated UBE2L3 variants were left untreated or treated with MG132 (10 µM) or bafilomycin A1 (20 nM) for 6 hours. -/- flag strep (B) iGsdmd cells stably expressing stated UBE2L3 variants were left untreated, MG132 treated (10 µM) for 6 hours. Data is representative of three independent experiments.

157

contributions to the post-translation regulation of protein abundance. These highlight

multiple aspects that underlie the regulation of UBE2L3 which caspase-1 may exploit

to trigger its depletion following inflammasome activation.

Given the effect that these mutations have on the stability of UBE2L3 it was

hypothesized that they may also impact UBE2L3 depletion by caspase-1. To examine

Figure 4.14: UBE2L3 depleted independently of lysine ubiquitylation and catalytic activity. -/- flag strep (A/B) iGsdmd stably expressing indicated UBE2L3 variants were LPS-primed for 3 flag strep hours followed by treatment with nigericin for 1 hour. Immunoblots show UBE2L3 detected with anti-Flag antibody and β-actin in cell lysate, and caspase-1 culture supernatant. Anti-Flag immunoblot of UBE2L318R (B) is a taken at a higher exposure relative WT. Data are representative of three (A) and two (B) experiments.

158 this, iGsdmd-/- macrophages expressing either wildtype tagged UBE2L3 or non- ubiquitylatable UBE2L318R with LPS and nigericin. Western blot analyses showed that, like wildtype protein, UBE2L318R depleted upon caspase-1 activation (Figure 4.14A).

Depletion was therefore independent of ability to ubiquitylate UBE2L3 on lysine residues. To further test this, studies were extended to include UBE2L314R and

UBE2L318R/C86A mutant variants. Both variants of UBE2L3 also depleted following caspase-1 activation (Figure 4.14B). These data showed that despite being important in regulating the basal stability of UBE2L3, lysine residues and the catalytic cysteine are dispensable for caspase-1 triggered depletion. This suggested that the loss of

UBE2L3 following inflammasome activation whilst dependent on the proteasome, occurs via atypical degradative signals.

4.3 Discussion The results presented in this chapter showed that UBE2L3 was not proteolytically processed by caspase-1 and was therefore not a direct substrate.

Rather, UBE2L3 was an indirect target, and somewhat analogous to pro-IL-1α and

HMGB1. However, unlike these proteins UBE2L3 was not secreted from the cell nor was it lost because of cell lysis. For example, loss of GSDMD resulted in the retainment of processed IL-1β, IL-18 and IL-1α and HMGB1 in cells, whereas UBE2L3 was still degraded. UBE2L3 was degraded by the proteasome following caspase-1 activation, and this could be prevented using proteasome inhibitors. Under resting conditions, and even with the addition of a pro-inflammatory stimuli like LPS, UBE2L3 was a stable protein with a half-life of more than 8 hours. This was far longer than the average time required for UBE2L3 depletion to occur following inflammasome stimulation, during which it had an approximate half-life of 30 minutes following treatments with ATP or nigericin. It is therefore unlikely that this effect was due to the

159 basal turnover of UBE2L3 which may only be overcome by the production of new protein. This suggested that the destabilisation and subsequent degradation of

UBE2L3 was an active process which is in some way mediated or triggered by the proteolytic activity of caspase-1.

Destabilisation of caspase substrates following cleavage is well document and has been reported for multiple proteins. The proteins Twist (Demontis et al., 2006) and

SSRP1 (Landais et al., 2006) cleaved by caspase-3 and IRF-3 cleaved by caspase-8

(Sears et al., 2011) are destabilised and subsequently ubiquitylated and degraded at by the proteasome. Similarly, there are examples of ubiquitin-independent degradation following caspase cleavage. Likewise, caspase-1 has been reported to be required for the proteasomal degradation of c-AIP in a degron-dependent manner following treatment of macrophages with anthrax toxin (Wickliffe et al., 2008). Inversely in Drosophila, drICE can also stabilise proteins, for example the protein Grim, by removing a lysine residue which is critical for degradative ubiquitylation (Yeh and

Bratton, 2013) . However, despite there being precedence for caspase-dependent destabilisation and degradation of cellular proteins, I found that UBE2L3 was not proteolytically processed by caspase-1 in vitro or in cells. Likewise, preventing depletion in the presence of MG132 and proteasome inhibition did not lead to the appearance of a cleaved fragment of UBE2L3. This would suggest that a different and more indirect mechanism mediated UBE2L3 destabilisation, perhaps via ubiquitylation or regulation of a degron.

As previously stated, UBE2L3 was a stable protein under resting conditions.

However, the mutation of 18 lysine residues in UBE2L318R drastically decreased its stability, which was restored by proteasome-inhibition. The presence of lysine residues appears to have a stabilising function as UBE2L314R, which has only four lysine

160 residues, displayed slightly increased basal protein levels as compared with

UBE2L318R. Further, UBE2L3P62A/F63A, which cannot bind most E3 ligases, was also unstable in a proteasome dependent manner. Inversely, introduction of the inactivating

C86A mutation stabilised UBE2L3. Further, introducing this mutation in UBE2L318R partially rescued protein expression, these results are summarised in Table 4.1.

These results revealed multiple factors that regulate UBE2L3 stability in cells.

Ubiquitylation is most commonly associated with targeting proteins for degradation however, the degradation of lysine-less UBE2L318R suggested that lysine ubiquitylation of UBE2L3 was required for maintaining protein stability. Many non- canonical degradative signals have been described in mediating protein turnover at the proteasome (Kravtsova-Ivantsiv and Ciechanover, 2012). These include ubiquitin- dependent and -independent signalling events which may be responsible for the degradation of lysine-less UBE2L3 (Figure 4.15).

Stability UBE2L3 variant (-) MG132 (+) MG132 Wildtype +++ ++++ 18R + +++ C86A ++++ ++++ P62A/F63A - + 14R ++ +++ 18R/C86A ++ ++

Table 4.1: Summary of UBE2L3 variant stability under resting conditions Table shows summary of the stability of tagged-UBE2L3 variants used in this study. (+) refers to estimated protein abundance and has been judged relatively to the amount of wildtype UBE2L3 in the absence of MG132.

161

Ubiquitin-dependent, lysine independent, signalling involves the ligation of ubiquitin to non-canonical serine, threonine, cysteine and N-terminal methionine residues (Kravtsova-Ivantsiv and Ciechanover, 2012), UBE2L3 has four serine, six threonine, and three cysteines. Multiple proteins have been described to be degraded via modification of non-canonical internal residues (McDowell et al., 2010; Tait et al.,

2007) and N-terminal methionine (Bloom et al., 2003; Coulombe et al., 2004; Kuo et al., 2004; Noy et al., 2012) residues. Indeed, such non-canonical ubiquitylation has been described to regulate the activity of E2 conjugating enzymes such as Ubc7 in yeast, which undergoes degradation in conditions which limit the availability of its cognate E3, leading to an accumulation of ubiquitin on its catalytic cysteine which leads to its targeting to the proteasome (Ravid and Hochstrasser, 2007). Indeed, it is of note that the introduction of an alanine in place of the catalytic Cys86 has a partial stabilising effect for UBE2L3C86A and UBE2L318R/C86A in resting conditions. It would be informative to test the abundance of ubiquitin-charged UBE2L3 variants, including the

P62A/F63A.

Ubiquitin-independent proteasomal degradation has also been described for many proteins via various mechanisms. One such example is the direct recognition of proteins by the 20S proteasome via ubiquitin-independent degrons, unstructured regions of a protein which can be directly detected by the 20S proteasome (Baugh et al., 2009; Fortmann et al., 2015; Schrader et al., 2009) . Alternatively, p21 for example, not only can be N-terminally ubiquitylated (Bloom et al., 2003; Coulombe et al., 2004), but can be targeted for degradation through direct interactions with the REGγ complex which forms an alternative lid for 20S proteasomes (Chen et al., 2007). p21 acts as an inhibitor of the cell cycle and, in addition to the above N-terminus mechanism, can be degraded following ubiquitylation by the SCF complex which allows for progression

162 during the G1/S transition and S phase (Abbas and Dutta, 2009). Interestingly,

UBE2L3 protein levels have been shown to decrease upon entry into S phase in a proteasome dependent manner (Whitcomb et al., 2009). Its levels are restored during

G2 phase and it is thought that depletion of UBE2L3 is important for regulating the timing of S phase progression, which double in UBE2L3 silenced HEK293, HeLa and

HLE cell lines (Whitcomb et al., 2009).

Alternatively, caspase-1-mediated depletion of UBE2L3 could occur independently of ubiquitin. It is possible that a portion of UBE2L3 acts as a degron which mediates proteasomal targeting. This could occur due to disruption of direct interactions with an E3 ligase or an adapter protein such as Smad7 (Ogunjimi et al.,

2005) or NDFIP1 (Kathania et al., 2015) interactions, which as shown in Figure 4.12A also heavily contribute to E2 stability. Caspase-1 has been reported to cleave several

E3 ligases this could subsequently lead to their destabilisation or inactivation which, in turn, could block UBE2L3 interactions leading to destabilisation either by a degron or unconventional ubiquitylation (Chaugule et al., 2011; Shembade et al., 2010).

Finally, the activity of a proteasome adaptor, such REGγ complex, may be altered by caspase-1 causing it to shuttle UBE2L3 to the proteasome (Chen et al., 2007).

As lysine-less UBE2L3 was less stable than the wildtype protein, it would be safe to hypothesise that lysine residues or indeed their ubiquitylation, have a stabilising effect. Such a process has been described for proteins such as EGFR (Ray et al., 2011), MAD2 (Osmundson et al., 2008) and NEDD9/HEF1 (Moore et al., 2010) which are unstable and degraded by the proteasome. These proteins were found to be substrates of the E3 ligase SMURF2 which mediated the addition of K63 ubiquitin chains and increase their stability (Moore et al., 2010; Osmundson et al., 2008; Ray et al., 2011). Such a process could be involved in regulating UBE2L3; indeed,

163

SMURF2 is known to function with UBE2L3 and requires an adapter protein called

SMAD7 which facilitates the interaction between the E3 and E2 pairs (Ogunjimi et al.,

2005).

Similarly, the introduction of E3 binding mutations in UBE2L3P62/F62 also destabilises UBE2L3. This too has been previously described for E2s, where limited

E3 availability (Ravid and Hochstrasser, 2007) or disruption of E2-E3 interactions

(Shembade et al., 2010) led to proteasomal degradation of the E2 conjugating enzymes UBE2N and UBE2D3. The stability of UBE2L3 appeared to be regulated by multiple factors under resting conditions. However, the depletion of UBE2L3 in response to caspase-1 activation appeared to be independent of lysine ubiquitylation, as despite alterations in basal stability, UBE2L318R and UBE2L314R were both depleted following caspase-1 activation. Likewise, neither catalytic activity nor the catalytic cysteine were required for depletion. Because depletion proceeded regardless of lysine residues, these results suggest that the addition or removal of ubiquitin by a caspase-1 regulated E3 ligase or DUB was unlikely. Although depletion was not dependent on these factors, the underlying mechanisms regulating basal UBE2L3 stability could be positively or negatively impacted by caspase-1 activity.

Another possible mechanism could be via N-terminal ubiquitylation which has been shown to cause proteasomal degradation. Indeed, the E3 ligase HUWE1 which is a known substrate of caspase-1 (Lamkanfi et al., 2008) has been shown to ubiquitylate MyoD on its N-terminus, thus targeting it for degradation (Noy et al., 2012).

Interestingly, cleavage by capase-1 at the proposed Asp2384 residue removes the large

N-terminus but leaves the catalytic HECT domain intact. Cleavage of HUWE1 by caspase-1 could have the potential to drastically alter HUWE1 activity, localisation or substrate preference as a means of mediating N-terminal ubiquitylation of UBE2L3.

164

Interestingly a recent SILAC based proteomics study noted that in modest 1.25 fold increase in UBE2L3 protein abundance in HUWE1-null HEK293 cells (Xu et al., 2016).

Further, my experiments with UBE2L3 and the bacterial NleL E3 ligase (Chapter 6) suggest that UBE2L3 can function in its own N-terminal ubiquitylation by NleL in vitro.

Further analysis will be required to determine which mechanism(s) mediate UBE2L3 depletion. It should be noted that the introduction of epitope tags to the UBE2L3 as described, introduced several lysine residues. Despite this, the observation that the

C86A mutation or differing stability of UBE2L318R and UBE2L314R suggested that the lack of lysine residues in the core UBE2L3 sequence are key. Nonetheless, further studies with untagged variants would be required to examine these more accurately.

During studies on N-terminal ubiquitylation others have shown that the addition of a large N-terminal tags inhibits N-terminal ubiquitylation (Coulombe et al., 2004).

Indeed, in this study proteins with N-terminal tags e.g. 3xFLAG, were used. However,

Coulombe et al. (2004) stated that tags larger than 5 kDa were required to inhibit N- terminal ubiquitylation, and the 3x FLAG tag used here has an estimated molecular mass of 3.6 kDa (25 kDa amino acids) and should remain permissive to N-terminal modification. This is also supported by the apparent responsiveness of tagged

UBE2L3 variants to proteasome inhibition. Further studies to detect ubiquitin on

UBE2L3, and its protein interactions which occur during caspase-1 activation need to be performed.

These results raise the question of what is the purpose of UBE2L3 depletion?

Given that UBE2L3 was degraded by the proteasome it would make sense that this would be phenotypically the same as its inhibition. There are approximately 40 E2

165 conjugating enzymes which may function with hundreds E3 ligases (Stewart et al.,

2016). UBE2L3 is unique in that it is functionally limited to RBR and HECT E3 ligases

Figure 4.15: Potential ubiquitin-dependent an independent mechanism which may cause caspase-1 dependent, proteasomal targeting of UBE2L3. (Top) Ubiquitin-dependent mechanisms of depletion: An E3 ligase activated by caspase-1 ligates ubiquitin onto non-lysine residues either serine, threonine, cysteine or the N-terminal methionine. Polyubiquitin on these residues form a degradative signal which targets UBE2L3 to the 26S proteasome. (Bottom) Ubiquitin-independent mechanisms of depletion: active caspase-1 destabilises or inactivates a cognate E3 ligase, disrupting UBE2L3/E3 interactions causing a loss of stability due to the presence of degron or similar signal, causing degradation at the open 20S proteasome. An proteasomal adaptor is altered by caspsase-1 and shuttles UBE2L3 to the proteasome.

166

(Wenzel et al., 2011), it has also been shown to be the preferred E2 enzyme of the LUBAC complex (Lewis et al., 2015). Removal of UBE2L3 could have the potential to affect the activities of multiple E3 ligases and by extension thousands of substrates.

What impact this may have will need to be examined. However, given that UBE2L3 is depleted from phagocytes this may suggest that it plays a role in negatively regulation of inflammasome mechanisms. These possibilities are explored to the next chapter.

4.4 Conclusions The results presented here demonstrate that UBE2L3 was not a direct substrate of capase-1 and is therefore an indirect target. UBE2L3 was not secreted from macrophages following inflammasome activation. Protein samples of combined lysate and supernatant from nigericin treated, wildtype THP1 cells show loss of

UBE2L3 suggesting that UBE2L3 was not lost to the extracellular space but turned over internally. Work in Chapter 3 showed that cell lysis does not contribute to UBE2L3 depletion. Here this was examined further in iGsdmd-/- macrophages which also depleted UBE2L3 following caspase-1 activation. This showed that UBE2L3 depletion occurs via an active intracellular process and is not secreted from cells. This process was dependent on the proteasome which mediated the degradation of UBE2L3 following caspase-1 activation. Under resting conditions, UBE2L3 stability was shown to be regulated by lysine availability, catalytic activity and E3 binding. Together, these data show that UBE2L3 was subjected to non-typical proteasomal targeting and degradation in a caspase-1-dependent manner.

167

Chapter 5 – UBE2L3 regulates IL-1β production by inflammasomes by regulating pro-IL-1β stability

168

5.1 Introduction The previous chapters identified the ubiquitin conjugating enzyme UBE2L3 as a novel indirect target of caspase-1 which is targeted to the proteasome following inflammasome activation. As an E2 enzyme, UBE2L3 functions in the ubiquitylation of numerous proteins with wide ranging consequences, including altering protein function, activity or stability (Komander and Rape, 2012). Ubiquitylation is known to regulate many aspects of the immune system, such as TLR signalling (Chuang and

Ulevitch, 2004), NF-κB activation (Iwai, 2010), autophagy, mitophagy (Kuang et al.,

2013) and the even regulation of inflammasomes (Bednash and Mallampalli, 2016;

Rodgers et al., 2014). The caspase-1-mediated loss of UBE2L3 could therefore have the potential to impact many cellular processes.

5.1.2 Functions of ubiquitin chain linkages

Over recent years it has become evident that ubiquitin-based modifications are incredibly diverse and can be attached to a substrate protein in varying forms such as monoubiquitylation, multi-monoubiquitylation and polyubiquitylation. The latter is generated by covalently linking ubiquitin molecules together, which can form either continuous or branched ubiquitin chains (Komander and Rape, 2012). The building of ubiquitin chains requires isopeptide bond formation between the C-terminal glycine of a donor ubiquitin and one of seven lysine residues (K6, K11, K27, K29, K33, K48 and

K63) of an acceptor ubiquitin. Alternatively, linear chains can be formed by the attachment of one ubiquitin to the N-terminus of another by a peptide bond (Stieglitz et al., 2013). Chains of differing linkage type have distinct structural topologies which have specific effects on the fate of the substrate protein. The variety that exists within ubiquitylation makes up the so called “ubiquitin code” which can be read by specialised ubiquitin-binding domains (UBDs) that recognise specific linkage types. (Komander

169

and Rape, 2012; Yau and Rape, 2016). Similarly, de-ubiquitylating enzymes which

remove ubiquitin modifications or ubiquitin-editing enzymes such as A20 (Vereecke et

al., 2009) can have distinct linkage specificities and provide an added layer of

regulatory and functional complexity to the code (Mevissen and Komander, 2017).

Ubiquitin can therefore impact many cellular processes inside a cell depending

on the type of linkage (Table 5.1). The most common consequence of ubiquitylation is

the targeting of proteins for degradation by the proteasome. This is most commonly

carried out by K48-linked polyubiquitin chains which is also the most abundant form of

polyubiquitin found in cells. Other forms such as K63-linked chains are non-

degradative and instead aid in the formation of and function of signalling complexes

by promoting recruitment of ubiquitylated proteins. K63-linked chains are therefore

important for processes such as autophagy where they act as a signal for the

Cellular Pathway Ubiquitin linkages References Inflammasome signalling K48, K63, M1 (Bednash and Mallampalli, 2016; Yang et al., 2017) Endocytosis MultiMono, K63 (Haglund et al., 2003; Rajalingam and Dikic, 2016) Xenophagy K63 (Manzanillo et al., 2013; Rajalingam and Dikic, 2016) Mitophagy K6, K11, pS65-K63 (Cunningham et al., 2015; Kondapalli et al., 2012; Rajalingam and Dikic, 2016) Post-golgi traficking K33 (Yuan et al., 2014) NF-κB signalling K63, M1, K27 (Liu et al., 2014a; Rittinger and Ikeda, 2017; Wang et al., 2014) Proteasomal degradation K48, K11, K29, K63 (Grice et al., 2015; Grice and Nathan, 2016; Saeki et al., 2009; Xu et al., 2009) Endoplasmic-reticulum- K11, K48 (Locke et al., 2014; Sommer associated protein and Jentsch, 1993) degradation Cell-cycle regulation K11, K48 (Matsumoto et al., 2010; Yau et al., 2017)

Table 5.1: Regulation of cellular pathways by ubiquitin chain linkage. Examples of diverse cellular functions regulated by ubiquitylation. Ubiquitin chain linkage topology produces distinctive outcomes in regulating protein activity and fate.

170 recruitment of autophagy adaptors such as sequestosome-1/p62 (Cohen-Kaplan et al., 2016). Similarly, linear ubiquitin chains have emerged as key components in the formation signalling hubs within the cell, notably during NF-κB activation (Rittinger and

Ikeda, 2017). Protein regulation by the ubiquitin code is therefore a multidimensional process involving the initial determination of linkage-type by the E2 and E3 ubiquitylation enzymes and subsequent adjustment by DUBs or editing enzymes.

5.1.3 Regulation of ubiquitylation via UBE2L3

In the context of RING E3 ligases, which lack intrinsic ubiquitin transfer reactivity, it is the E2 which acts as the deciding factor in determining ubiquitin linkage specificity (Stewart et al., 2016). For example, UBE2N specifically forms K63-linked chains (Geisler et al., 2014) and UBE2K forms K48-linked chains (Middleton and Day,

2015) via secondary interactions between ubiquitin and the E2 (Stewart et al., 2016).

When an E2 binds to a RING E3 ligase, it promotes a so called “closed state” whereby ubiquitin is folded back onto the E2, priming it for aminolysis (Plechanovova et al.,

2012). These combined factors likely mediate linkage specificity depending on which of ubiquitin’s lysine residues are preferentially exposed (Stewart et al., 2016). UBE2L3, however, is functionally limited to HECT and RBR E3 ligases (Wenzel et al., 2011) which are capable of controlling linkage specificity themselves (Kim and Huibregtse,

2009; Kobayashi et al., 2018). The LUBAC enzyme HOIP for example specifically orientates ubiquitin so that its N-terminal amino group is exposed and ready for aminolysis (Stieglitz et al., 2012; Stieglitz et al., 2013). Because UBE2L3 is functionally limited to HECT and RBR ligases it is unlikely to control linkage specificity. Recent studies have however shown that UBE2L3 has quicker chain formation with HOIP than the structurally similar UBE2D1 (Dove et al., 2016). This is in agreement with reports that LUBAC preferentially utilises UBE2L3 in NF-κB signalling (Lewis et al., 2015) and

171 in TNF-α induced linear ubiquitylation (Fu et al., 2014). This could be due to a distinctive closed conformation which UBE2L3 undertakes when conjugated with ubiquitin, whereas others such as UBE2D1 adopt for an open conformation (Dove et al., 2016). Another recent study showed that RBR/UBE2L3 interactions are directly mediated by UBE2L3 whereas UBE2D1 must be charged in order to support its interaction via additional non-covalent interactions with ubiquitin (Martino et al., 2018).

RBR/UBE2L3 complexes have also been shown to take on a so called extended- complex conformation which could be responsible for the observed increase

HOIP/UBE2L3 reactivity and may confer a similar effect on other RBR E3 ligases

(Dove et al., 2016; Martino et al., 2018). Given the observed preference in binding and activity with UBE2L3, caspase-1 mediated loss of UBE2L3 could hamper the activity of multiple HECT and/or RBR E3 ligases to exert cellular effects.

5.1.4 Role of ubiquitin in inflammasome signalling

Ubiquitylation has roles in regulating many aspects of immune signalling such as TLR signalling, NF-κB activation and autophagy (Deretic et al., 2013; Hu and Sun,

2016). More recently, ubiquitylation has also been shown to function in both positively and negatively regulating inflammasome activation (Yang et al., 2017). Studies have shown that ubiquitin can regulate almost every inflammasome component ranging from NLR/ALRs sensors to caspase-1 and IL-1β. The ASC adapter protein which is a central component to most inflammasomes, has been shown to be decorated with K63 and linear ubiquitin chain types (Guan et al., 2015; Rodgers et al., 2014; Shi et al.,

2012). Shi et al. (2012) reported that inflammasomes were negatively regulated by the ubiquitylation of ASC (Shi et al., 2012). They showed that stimulation of either NLRP3 or AIM2 inflammasomes promoted ASC K63-linked polyubiquitylation and subsequent autophagic degradation, thereby reducing excessive inflammasome activation (Shi et

172 al., 2012). On the other hand, K63-linked polyubiquitylation of Lys174 of ASC by the E3 ligase TRAF3 has been described to enhance inflammasome activation in response to RNA viruses (Guan et al., 2015). Similarity, ASC modification with linear ubiquitin by the LUBAC enzyme HOIL-1 (encoded by Rbck1 gene) was shown to be essential for the formation and activity of the NLRP3 inflammasome. The role of HOIL-1 was essential for NLRP3 activation and its absence in Rbck1-/- mice was protective against endotoxic shock (Rodgers et al., 2014). Inflammasome sensors themselves are also regulated by ubiquitylation. AIM2 is targeted for autophagic degradation by the E3 ligase TRIM11 (Liu et al., 2016a). Interestingly this is not by the direct modification of the AIM2 protein. In response to cytosolic DNA, TRIM11 undergoes autoubiquitylation on Lys458 and binds AIM2. The ubiquitin chains act as a scaffold and recruits the autophagy adaptor p62 resulting in them all being targeted for degradation by autophagy (Liu et al., 2016a).

NLRP3 is also regulated by ubiquitylation. Under resting conditions, in the absence of signal 1 priming, basal ubiquitylation of NLRP3 blocks its activation

(Juliana et al., 2012; Lopez-Castejon et al., 2013). Signal 1 agonists initiate non- transcriptional priming which triggers the removal of ubiquitin by DUBs (Juliana et al.,

2012) for example, BRCC3 (Py et al., 2013). Under resting conditions, NLRP3 is also modified with K48-linked polyubiquitin by TRIM31 (Song et al., 2016) and FBXL2 (Han et al., 2015) E3 ligases which target it for degradation by the proteasome. More recently, NLRP3 has also been shown to be modified with K48- and K63-linked ubiquitin chains which are added by the RBR E3 ligase ARIH2 (aka TRIAD1). Some of these modifications also serve to negatively regulate NLRP3 activation independently of the proteasome (Kawashima et al., 2017).

173

Caspase-1 is also regulated by K63-linked polyubiquitylation by the E3 enzymes cIAP1 and cIAP2. This modification is not an absolute requirement for caspase-1 activation but significantly enhances caspase-1 activity and downstream events (Labbe et al., 2011). Targets of inflammasomes are also regulated by ubiquitylation. Ainscough et al. (2014) showed that pro-IL-1β and pro- IL-1α in human macrophage and dendritic cell are basally ubiquitylated and turned over by the proteasome however, they did not identify which lysine residues were modified or which E3 ligase was responsible. Duong et al. (2015) have shown that the ubiquitin- editing protein A20, which possess both ubiquitin ligase and de-ubiquitylating activity, specifically removes K63-linked ubiquitin from Lys133 of pro-IL-1β in mouse macrophages. Ubiquitylation at Lys133 of mouse pro-IL1 was found to promote its proteolytic processing by caspase-1; however, the identity of the E3 ligase or ligases which were responsible for modifying IL-1β remains to be determined. The only E3 ligase which has been described to modify IL-1β is the HECT E3 ligase E6-AP (aka

UBE3A). Niebler et al. (2013) showed that when activated by the HPV virus E6 protein, E6-AP specifically targeted pro-IL-1β for proteasomal degradation (Niebler et al., 2013).

The significance of ubiquitylation-mediated regulation of inflammasomes is further underscored by its targeting by bacterial effectors. The EPEC effector NleA has been shown to inhibit IL-1β secretion from THP1 cells. NleA binds to ubiquitylated

NLRP3 in its LRR and PYD domains, and by doing so blocks NLRP3 de-ubiquitylation which is required for NLRP3 activation (Yen et al., 2015). The Yersinia pestis E3 ligase effector YopM also manipulates inflammasome activation and has been described to have both inhibitory and stimulatory effects on inflammasome signalling (Wei et al.,

2016). YopM has been shown to bind caspase-1 thereby antagonising its activation

174

(Chung et al., 2014; LaRock and Cookson, 2012), and prevent the activation of the

PYRIN inflammasome by other Y. pestis effectors(Chung et al., 2016; Ratner et al.,

2016). These inhibitory effects, however, were not examined in the context of its ubiquitin ligase activity. On the contrary, YopM can also activate the NLRP3 inflammasome. YopM binds and ubiquitylates NLRP3 with K63-linked polyubiquitin which leads to enhanced cell death (Wei et al., 2016).

5.1.5 Chapter Aims

Ubiquitylation has diverse cellular effects and its role in inflammasome activation and functionality are a clear reflection of this. As UBE2L3 is functionally limited to HECT and RBR E3 ligases (Wenzel et al., 2011), which include key ligases such as HOIP, its depletion by caspase-1 could have a significant impact on signalling.

Because UBE2L3 is depleted, I hypothesised that it may be involved in negatively regulating inflammasomes and was therefore removed from cells. The aim of this chapter therefore was to determine the consequences of UBE2L3 depletion.

175

5.2 Results 5.2.1 IL-1β secretion is reduced by sustained UBE2L3 expression

In previous chapters I have shown that casapase-1 mediated the rapid proteasomal degradation of UBE2L3 following inflammasome activation. It was hypothesised that UBE2L3 may function to negatively regulate caspase-1 mediated processes, and therefore must be depleted to facilitate optimal inflammasome responses. To this end, I wanted to examine the consequences of sustained UBE2L3 protein expression in the presence of active caspase-1. To do this, human THP1 cells were transduced to stably express wildtype UBE2L3 by retroviral transduction using

CMV promoter-driven high expression of an N-terminally YFP-tagged UBE2L3

(THP1YFP-UBE2L3 cells). The hope was that the exogenous expression of YFP-UBE2L3 would exceed the rate of depletion and maintain higher total UBE2L3 even in the presence of active caspase-1.

βME _ + _ + Ub~ * 50

1 2 3 4

GFP

Figure 5.1: N-terminal tagging of UBE2L3 does not affect ubiquitin charging. C86S PMA-differentiated THP1 cells stably expressing either YFP-UBE2L3 or YFP-UBE2L3 were lysed in 2x Laemmli loading dye without (-) or with (+) 5% v/v 2-meracaptoethanol (βME) to examine the charging and reduction of the catalytic residue of UBE2L3. Ub~ * refers to UBE2L3 charged with ubiquitin. YFP-UBE2L3 was detected with anti-GFP antibody. Data is representative of two experiments.

176

To ensure the resulting protein retained catalytic activity I tested whether it was capable of being charged with ubiquitin in cells. Immunoblot analysis with anti-GFP antibody revealed an additional band above YFP-UBE2L3, which migrated at ~50 kDa in non-reducing conditions (Figure 5.1, lane 1). This would correspond to the cumulative molecular weight of YFP-UBE2L3 (~43 kDa) charged with a ubiquitin molecule (~8.5 kDa) on its catalytic cysteine (Cys86), via a thioester bond. The additional band was lost following the addition of 2-mercaptoethanol (5% v/v), confirming that it was a thioester-linked modification as thioester bonds are labile under reducing conditions (Figure 5.1, lane 2). To confirm that the modification was on the catalytic C86 residue, I tested a mutant which had a serine residue in place of the catalytic cysteine (YFP-UBE2L3C86S). The substituted serine would force the formation of an oxyester bond that, unlike the thioester bond, is stable under reducing conditions. Therefore, if the ubiquitin were indeed attached at residue 86 the modification would be more stable and would not disappear in the presence of 2- mercaptoethanol. Consistent with this, the addition of 2-mercaptoethanol did not affect the charged ubiquitin species (Ub~) from lysates of cells expressing YFP-UBE2L3C86S

(Figure 5.1, lanes 3 and 4). This data confirmed that YFP-UBE2L3 could be charged with ubiquitin via a thioester bond (Figure 5.1).

To test whether the presence of YFP-UBE2L3 impacted caspase-1 responses,

THP1 cells stably expressing YFP (THP1Ctrl#1) or YFP-UBE2L3 (THP1YFPUBE2L3) were primed with LPS and pulsed with nigericin for 1 hour (Figure 5.3A). Secreted IL-1β was detected by ELISA and showed that nigericin-treated THP1YFPUBE2L3 cells secreted ~2 fold less IL-1β compared to THP1Ctrl#1 (Figure 5.2A). Despite the reduced cytokine secretion, cell death was unaffected as THP1Ctrl#1 and THP1YFPUBE2L3 cells displayed similar levels of pyroptosis (measured by LDH release) and plasma

177 membrane permeabilization (measured by PI uptake) (Figure 5.2A). These data showed that the difference in IL-1β release was not due to an increase in the release of cellular contents by cell lysis. They also suggested that caspase-1 activation was similar in the two cell lines, as an effect on caspase-1-activation would have influenced the extent of cell death. As previous data showed that UBE2L3 depletion was common to multiple inflammasomes and agonists, I next tested whether persistent expression of YFP-UBE2L3 effected IL-1β secretion in response to inflammasome activation by bacterial infection. Differentiated THP1Ctrl#1 and THP1YFPUBE2L3 were primed with LPS for 3 hours and infected with high SPI-1-induced STm at an MOI of 40 for three hours.

The supernatants were then tested for IL-1β secretion, cell death, and cell membrane permeabilization. As with nigericin, THP1YFPUBE2L3 had a ~5-fold reduction in IL-1β

Figure 5.2: Sustained UBE2L3 expression reduces IL-1β secretion independently of cell death Ctrl#1 YFPUBE2L3 -1 PMA-differentiated THP1 and THP1 cells were primed with 250 ng.ml LPS for 3 hours and pulsed with 10 µM nigericin (A) or infected with high SPI1-expressing STm (MOI 40) for 3 hours (B). Graphs show ELISA quantification of secreted IL-1β (left Y-axis), cell death determined by release of LDH or uptake of propidium iodide (PI; both plotted on right Y-axis). Mean±S.E.M. from three independent experiments are shown. * Benjamini- Hochberg (BH) corrected P<0.05 by two-way ANOVA; ns not significant.

178 secretion as compared with THP1Ctrl#1, but similar levels of cell death in both cell types

(Figure 5.2B).

It is of note that ELISA analyses were unable to distinguish mature and pro- forms of IL-1β. To ascertain whether the observed differences were due to alterations in the release of either mature p17 IL-1β or pro-IL-1β western blot analysis was performed. Such analysis was also used to confirm whether caspase-1 activation was similar in both THP1Ctrl#1 and THP1YFPUBE2L3 cells. Treatment with LPS followed by nigericin for 30 and 60 minutes showed that both cell types activated caspase-1 to comparable levels across multiple experiments, as they showed similar amounts of active p10 caspase-1 in their respective cell supernatants (Figure 5.3A, Casp1 p20).

Endogenous UBE2L3 depleted in both cell types as has been shown previously.

Figure 5.3: Sustained UBE2L3 expression reduces IL-1β secretion Ctrl#1 YFPUBE2L3 -1 Differentiated THP1 and THP1 cells were primed with 250 ng.ml LPS for 3 hours and pulsed with 10 µM nigericin for indicated times. (A) Indicated proteins were immunoblotted in cell lysates and supernatants. (B) Graph shows quantification of mature p17 IL-1β in supernatant relative to lysate β-actin from replicate experiments similar to (A) (Mean±S.D.). Images are from the same blot with intervening lanes removed. Data are representative of experiments three experiments.

179

However, YFP-UBE2L3 displayed no discernible depletion following caspase-

1 activation and appeared to be stable throughout the treatment (Figure 5.3A, YFP-

UBE2L3). Interestingly, western blot analyses also showed that pro-IL-1β depleted from the cell lysates more rapidly in THP1YFPUBE2L3 cells as compared with THP1Ctrl#1

(Figure 5.3A, pro-IL-1β). However, the secretion of mature p17 IL-1β was reduced in

THP1YFP-UBE2L3 cells as compared with THP1Ctrl#1 (Figure 5.3A, IL-1β p17) confirming that the reduction of total IL-1β secretion previously observed was specifically due to a decrease in mature IL-1β (Figure 5.2A).

In agreement with the ELISA data, quantification of immunoblot band intensity of mature IL-1β relative to lysate β-actin across biologically independent experiments confirmed that THP1YFPUBE2L3 cells secreted less mature IL-1β; at 60 minutes post- nigericin treatment, there was ~50% less mature IL-1β released by THP1YFPUBE2L3 than

THP1Ctrl#1 cells (Figure 5.3B). These data established that (i) caspase-1 activation was similar between both cell types, and (ii) THP1YFPUBE2L3 cells were specifically secreting less mature p17 IL-1β, due to the reduced depletion of YFP-UBE2L3, (iii)

THP1YFPUBE2L3 cells allowed the effect of UBE2L3 in the presence of active caspase-

1 to be examined.

5.2.2 Reduced IL-1β release is not due to priming defects

UBE2L3 has been shown to be essential for NF-kB activation in HeLa and HEK cells (Fu et al., 2014; Lewis et al., 2015). If this were also the case in macrophages, there would be an effect on NLRP3-inflammasome priming. I therefore decided to examine its role in NF-kB signalling upon LPS and PAM priming of human and mouse macrophages. This was relevant because pro-IL-1β is not expressed in unprimed

(naïve) cells and is transcriptionally upregulated by NF-kB-dependent mechanisms in response to microbial molecules such as LPS.

180

Contrary to these reports, in THP1 cells the overexpression of UBE2L3 had anti-inflammatory effect because of its effect on reducing IL-1β. In addition, despite there being similar induction of pro-IL-1β and NLRP3 proteins in THP1Ctrl#1 and

THP1YFPUBE2L3 cells (Figure 5.4A), an effect on their transcriptional had not been ruled out. I therefore wanted to examine whether the observed reduction in IL-1β release could be due to a disruption in pro-inflammatory gene expression in response to LPS.

To this end THP1Ctrl#1 and THP1YFPUBE2L3 left untreated or stimulated with LPS for 3 hours were tested for transcriptional induction of IL1B and TNF (Figure 5.4B). Both

THP1Ctrl#1 and THP1YFPUBE2L3 cells had similar fold-change of IL1B and TNF transcripts after a 3-hour treatment with LPS (Figure 5.4B), showing that transcriptional induction

Figure 5.4: UBE2L3 overexpression does not alter signal 1 priming Ctrl#1 YFP-UBE2L3 PMA-differentiated THP1 and THP1 expressing were primed with 250 -1 ng.mL LPS for 3 hours. (A) Immunoblot analyses of indicated proteins in cell lysates. Data are representative of 3 experiments. (B) Fold change of IL1B and TNF mRNA following LPS-treatment. Fold-change calculated by ΔΔCt method normalised to GAPDH control. Graph shows Mean±S.E.M. from two independent experiments. ns, not significant.

181 of common NF-B-dependent pro-inflammatory genes was unaffected by UBE2L3 over-expression. This suggested that the reduced production of mature IL-1β by

THPYFPUBE2L3 cells is not due to transcriptional defects in response to LPS.

5.2.3 UBE2L3 controls pro-IL-1β protein levels

So far data showed that the sustained expression of UBE2L3 reduced

IL-1β secretion. This occurred independently of cell death, changes in caspase-1 activation and transcription of IL1B. I decided to examine the stability of pro-IL-1β as it had previously been described to be ubiquitylated and regulated by both the proteasome and the autophagy pathway (Ainscough et al., 2014; Duong et al., 2015;

Harris et al., 2011a; Moors and Mizel, 2000). To examine the role of UBE2L3 in mouse cells, time course analysis of pro-IL-1β protein expression was carried out using mouse iBMDMs expressing YFP (iWTCtrl#1) or YFP-UBE2L3 (iWTYFPUBE2L3). Cells were primed with 250 ng.mL-1 LPS for up to18 hours with samples taken at regular intervals

(Figure 5.5A). Immunoblot analyses showed that the induction of pro-IL-1β protein was similar in both cell types up to 6 hours post-LPS treatment (Figure 5.5A, lanes 1 to 7).

However, after 6 hours, pro-IL-1β levels in iWTYFPUBE2L3 started to reduce and protein was undetectable by 18 hours. iWTCtrl#1 cells, however, had retained continuous expression of pro-IL-1β with only a slight drop at 15 hours (Figure 5.5B, lanes 8 to 11).

Quantification of pro-IL-1β band intensity (relative to β-actin) across multiple independent experiments confirmed that iWTYFPUBE2L3 cells underwent a significant and sustained reduction in pro-IL-1β protein levels after 6 hours as compared with iWTCtrl#1 (Figure 5.5B). Detection of Nlrp3 protein, which is also induced by LPS in a

NF-B-dependent manner, showed no reduction at later time points in both cell types

(Figure 5.5A and B).

182

A LPS

Time (h) 0 .5 1 2 3 4 6 9 12 15 18 Ctrl#1 pro-IL-1 YFP-UBE2L3

Ctrl#1 NLRP3 YFP-UBE2L3

Ctrl#1 -actin YFP-UBE2L3 1 2 3 4 5 6 7 8 9 10 11 B

C LPS (15 h)

37 pro-IL-1β

25 Anti-Flag

-actin iGsdmd-/-

Figure 5.5: Sustained UBE2L3 expression increases pro-IL-1β turnover -1 (A) YFP (Ctrl#1) or YFP-UBE2L3 expressing iBMDMs were treated with 250 ng.mL LPS for indicated times. Immunoblots show UBE2L3, NLRP3 and β-actin in cell lysates. Data are representative of 4 independent experiments. (B) Quantification of pro-IL-1β and NLRP3 protein band intensity relative to β-actin from experiments as described in (A). Graph shows mean pro-IL-1β/β-actin ratio plotted as percentage of 6 h timepoint. Mean±S.E.M. from three independent experiments are shown. * BH corrected P<0.05 , ** BH corrected P<0.01 by two-way ANOVA. -/- (C) Indicated Gsdmd iBMDMs cell lines were treated with LPS for 15 h. Immunoblots for pro-IL-1β, anti-FLAG and β-actin in cell lysates are shown. Data are representative of three experiments.

183

This suggested that the loss of protein was specific to pro-IL-1β and that a general shutdown of pro-inflammatory transcription was unlikely to be responsible for these observations. To ensure that the impact of UBE2L3 on pro-IL-1β was not due to the N-terminal YFP tag, mouse iGsdmd-/- expressing UBE2L3 with a triple N-terminal flag tag and C-terminal StrepII-tag were also tested. Immunoblot detection of pro-IL-

1β showed that at 15 hours post-LPS treatment, UBE2L3-expressing cells had much lower pro-IL-1β protein (Figure 5.5C). To confirm that transcription of Il1b remained unaffected at later timepoints when no pro-IL-1β protein could be detected, qPCR analysis was performed on cells stimulated with LPS for 3, 6 and 15 hours.

Transcription of Il1b in iWTCtrl#1 and iWTYFPUBE2L3 was similarly induced and maintained at all tested time points; even though a slight decrease was observed at

15 h this was not statistically significant (Figure 5.6A). The same samples were also tested for the transcriptional expression of NF-κB dependent pro-inflammatory cytokine genes Tnf and Il6. Both genes were similarly induced at 3, 6 and 15 h and displayed no significant difference between iWTCtrl#1 and iWTYFPUBE2L3 cells (Figure

5.6A). Secretion of TNF and IL-6 proteins was also measured by ELISA (Figure 5.6B).

The secretion of TNF was similar across both iWTCtrl#1 and iWTYFPUBE2L3 cells, which was consistent with the observed reduction in transcription. IL-6 secretion, however, continued to accumulate over time and YFP-UBE2L3 expressing cells exhibited less secretion at 15 h (Figure 5.6B), contrary to the observed pattern of transcriptional induction which showed no change at 15 h (Figure 5.6A).

LPS is one of many inflammatory signals which can stimulate pro-inflammatory gene expression. To test whether other pro-inflammatory stimuli produced a similar phenotype, the two cell lines were treated with either TNF (2.5 ng.mL-1) or the TLR2 agonist PAM3CSK4 (1 µg.mL-1) for up to 24 hours. Immunoblot analysis showed that

184

the induction of pro-IL-1β protein by PAM3CSK4 was similar in both iWTCtrl#1 and

iWTYFPUBE2L3 cell types at 3 hours post-treatment (Figure 5.7A, lane 2 and 7). In

iWTCtrl#1 cells, pro-IL-1β levels increased further reaching its maximum protein at 6

hours post-treatment after which pro-IL-1β levels reduced. In comparison, pro-IL-1β

Figure 5.6: Reduced IL-1β stability is independent of transcription -1 YFP (Ctrl#1) or YFP-UBE2L3 expressing iBMDMs were treated with 250 ng.mL LPS for indicated times. (A) Fold change of Il1b, Tnf and Il6 mRNA in response to LPS fold change calculated by ΔΔCt method using GAPDH as a normalising control. (B) Graphs show ELISA quantification of secreted TNF and IL-6 by from indicated iBMDMs left untreated (UT) or treated with LPS for times as shown. Mean±S.E.M from three independent experiments are plotted. ns, not significant by two- way ANOVA.

185 levels did not increase after 3 hours and depleted almost entirely by 24 hours in iWTYFPUBE2L3 cells (Figure 5.7A).

Likewise, iWTYFPUBE2L3 cells also displayed more rapid turnover of pro-IL-1β produced in response to TNF treatment (Figure 5.7A). In this instance, iWTYFPUBE2L3 began to lose pro-IL-1β protein after 3 hours as compared with iWTCtrl#1 which had detectable pro-IL-1β beyond 6 hours of TNF treatment (Figure 5.7A). These data showed that newly induced pro-IL-1β protein was being turned over more rapidly in cells which expressed YFP-UBE2L3, and that the kinetics of turnover were different based on the stimulus. This reduction in pro-IL-1β protein was independent of transcriptional effects and therefore likely due to a post-translational regulatory mechanism.

I next wanted to examine whether other caspase-1-regulated cytokines were similarly affected by UBE2L3. I therefore tested the production of pro-IL-1α and pro-

IL-18 in response to LPS. IL-1α and IL-1β belong to the same family of cytokines and much like IL-1β, the expression IL-1α is inducible and is trigged by NF-κB-dependent signalling (Cogswell et al., 1994). However, despite activating the same receptor they only share ~26% primary sequence similarity (March et al., 1985). Interestingly,

UBE2L3 exerted a similar effect on pro-IL-1α and reduced its levels in cells (Figure

5.7B, top). Much like the observed effect on IL-1β, the induction of IL-1α protein was similar in iWTCtrl#1 and iWTYFPUBE2L3 cells at early timepoints (up to 15 h).

Subsequently, however, protein levels were depleted more rapidly in YFP-UBE2L3- expressing cells as compared with YFP-expressing iWTCtrl#1 cells which maintained of pro-IL-1α protein for up to 24 h-post LPS treatment (Figure 5.8B, top). The same samples were also tested for pro-IL-18 levels, which unlike IL-1, is constitutively expressed in cells (Afonina et al., 2015). Notably, pro-IL-18 showed no significant

186 change in response to LPS and no reduction in its levels in iWTYFPUBE2L3 cells (Figure

5.7B, middle). The effect of UBE2L3 therefore seemed to be specific to pro-IL-1α and pro-IL-1β and not general on all inflammasome-related cytokines.

Figure 5.7: UBE2L3 controls IL-1β and IL-1α protein levels -1 -1 (A) Indicated iBMDMs were treated with 1 µg.mL PAM3CSK4 (PAM) or 2.5 ng.mL TNF for stated times. Immunoblots show pro-IL-1β and β-actin in cell lysates. -1 (B) Indicated iBMDMs were treated with 250 ng.mL LPS for times as shown and pro-IL- 1α, pro-IL-18 and β-actin were detected in cell lysates by immunoblots. Data are representative of two experiments.

187

5.2.4 UBE2L3 functions in the proteasome-dependent turnover of pro-IL-

1β

Because similar levels of pro-IL-1β transcript and protein were induced at early time points, it was hypothesised that a UBE2L3-driven post-translational mechanism could be responsible for the more rapid turnover of pro-IL-1β in iWTYFPUBE2L3 cells.

Considering the role of UBE2L3 in ubiquitylation and previous reports of pro-IL-1β regulation by the proteasome and autophagy (Ainscough et al., 2014; Harris et al.,

2011a), I tested whether blocking either process abolished the effect of UBE2L3 on pro-IL-1β protein. Previous data suggested that iWTYFPUBE2L3 cells began to induce IL-

1β turnover approximately 6 hours after the addition of LPS. iWTCtrl#1 and iWTYFPUBE2L3 cells were therefore primed with 250 ng.mL-1 for 6 h followed by further incubation without or with specific inhibitors. Without changing culture medium post-LPS addition, cells were exposed to inhibitors which blocked either the proteasome (MG132, 10 μM), autophagosome/lysosome acidification (bafilomycin A, 20 nM), or autophagic proteases (pepstatin-A,10 μg.mL-1 and E64D, 10 μg.mL-1). Cells were incubated for up to 14 hours after the addition of inhibitors (Figure 5.8A). Cells which received DMSO

(vehicle control) behaved as before i.e. pro-IL-1β displayed an increased rate of turnover in iWTYFPUBE2L3 but not iWTCtrl#1 cells (Figure 5.8B). The inhibition of autophagy, either by blocking autophagosome acidification (Figure 5.8D) or proteases

(Figure 5.8E), had no impact on the faster rate of pro-IL-1β turnover in iWTYFPUBE2L3.

However, treatment with MG132 prevented pro-IL-1β turnover in iWTYFPUBE2L3 cells

(Figure 5.8C). Interestingly, MG132 treatment led to the appearance of an ~28 kDa form of IL-1β in both cell types (called p28 IL-1β henceforth to distinguish it from the p30 pro-IL-1β) (Figure 5.8C). The molecular weight of the p28 IL-1β corresponded to pro-IL-1β after cleavage at Asp26 and which precedes the cleavage of Asp116 by

188 caspase-1 (Davaro et al., 2014). Although the population of pro-IL-1β was split between p30 and p28 forms, analysis of the cumulative intensity of both bands

(relative to β-actin) showed that MG132 stabilised pro-IL-1β in iWTYFPUBE2L3 cells and restored protein levels similar to that of DMSO-treated iWTCtrl#1 cells (Figure 5.8F).

Figure 5.8: Sustained UBE2L3 expression increases proteasomal turnover of pro-IL- 1β Ctrl#1 UBE2L3 Schematic of experiment design (A). iWT and iWTYFP cells were primed with -1 LPS (250 ng.mL ) for 6 h and then treated with DMSO (solvent) (B), MG132 (C), Bafilomycin A (BafA) (D), or pepstatin-A plus E64D (E). Data are representative of 2 independent experiments. Quantification of pro-IL-1β band intensity (F) relative to β-actin from experiments similar to (B) and (C). Mean pro-IL-1β/β-actin ratios are normalised to the start of inhibitor treatment (0 h) and Mean±S.E.M. from two independent experiments are plotted. * Benjamini-Hochberg (BH) corrected P<0.05 by two-way ANOVA for indicated comparisons.

189

This suggested that the proteasome was responsible for the loss of pro-IL-1β in iWTYFPUBE2L3 cells.

MG132 has been shown to have some off-target effects and although the concentration used here was low I decided to use a more specific inhibitor (Kisselev and Goldberg, 2001). Epoxomicin has no reported off-target effects (Kisselev et al.,

2012; Meng et al., 1999) owing to highly specific interactions with the β5/Pre2 subunit of the proteasome (Groll et al., 2000). To ensure that the observed reduction in pro-

IL-1β stability was specific to the proteasome, experiments were repeated using epoxomicin. As before iWTCtrl#1 and iWTYFPUBE2L3 were treated with LPS for 6 hours followed by incubation with either MG132 (10 μM) or epoxomicin (5 μM) for a further

9 hours. The addition of epoxomicin blocked the reduction in pro-IL-1β protein levels similarly to MG132 (Figure 5.9, lanes 5 and 6). These data further confirmed that pro-

IL-1β was being degraded specifically by the proteasome in iWTYFPUBE2L3 cells.

LPS (6 h) + 9 h:

pro-IL-1β p30

-actin

1 2 3 4 5 6 Ctrl#1 YFP-UBE2L3

Figure 5.9: Proteasome inhibition by multiple inhibitors block IL-1β depletion. -1 Indicated iBMDMs were treated with LPS (250 ng.mL ) for a total of 15 h in the presence of solvent (DMSO), MG132 (10 µM) or epoxomicin (Epox; 5 µM) as indicated to accumulate pro-IL-1β over an additional 9 h period with inhibitors. Lysates were immunoblotted for pro-IL-1β and β-actin. Data are representative of at least four independent experiments.

190

The appearance of p28 IL-1β was a common feature of MG132 treatment. The strict requirement for proteasome-inhibition for p28 IL-1β to be detectable suggested that the p28 fragment was inherently unstable; p28 may be an intermediate generated before the complete degradation of pro-IL-1β; and p28 was produced by an MG132- and epoxomicin-insensitive protease. I therefore asked whether cleavage of pro-IL-1β at Asp26 was required for mediating its proteasomal turnover and whether a caspase was involved. Previous studies have reported that generation of p28 IL-1β in response A B MG132 Epox WT C1/11-/- DMSO DMSO IL-1β IL-1β

Actin Actin 1 2 3 4 5 6 7 8 1 2 3 4 5 6 iBMDMs iBMDMs

IL-1β

β-Actin 1 2 3 4 iBMDMs

Figure 5.10: IL-1β p28 requires caspase-1 activity and is not required for proteasomal turnover. -1 Indicated iBMDMs were treated with LPS (250 ng.mL ) for a total of 15 hours (A) Over the last 9 hours cells were treated with either MG132 (10 µM) or epoxomicin -1 (Epox; 5 µM) and DMSO, Bafilomycin A (BafA; 20 nM) or E64D (10 μg.mL )as indicated. (B) Over the last 9 hours either DMSO, MG132 (10 µM) or epoxomicin (Epox; 5 µM) were added. (C) Over the last 9 hours either DMSO, MG132 (10 µM), zVADfmk (20 µM) or a combination of both were added. Lysates were immunoblotted for pro-IL-1β and β -actin. Data are representative of at two independent experiments.

191 to treatment with statins required caspase activity but was independent of caspases -

1, -6 and -8 (Davaro et al., 2014). IL-1β has been reported to localise to autophagosomes (Harris et al., 2011a). To test whether autophagic proteases were required, I primed cells with LPS for 6 hours followed by a 9 hour treatment with combination of proteasome inhibitors (MG132 or epoxomicin) and either bafilomycin

A or E64D. Immunoblot analyses showed that neither bafilomycin A or E64D blocked p28 formation (Figure 5.10A) suggesting that autophagy plays no role in the formation of p28 IL-1β. Next, I investigated whether either inflammasome related caspases

(caspase-1 and caspase-4) were required. Immunoblot analysis showed that p28 formation was induced in both wildtype and Casp1/4-/- mouse macrophages (Figure

5.10B) which ruled out a role for either caspase-1 or -4. Despite this, the process was dependent on caspase activity as cells treated with both MG132 and zVADfmk blocked p28 formation (Figure 5.10C, lane 4). Interestingly, treatment with zVADfmk alone did not protect pro-IL-1β from degradation (Figure 5.10C, lane 1 to 3). These results suggested that pro-IL-1β cleavage into p28 – by an unknown caspase or zVADfmk- sensitive protease – is not essential for the proteasomal turnover of pro-IL-1β.

Previous data showed that pro-IL-1α was similarly turned over at an increased rate in iWTYFPUBE2L3 cells (Figure 5.7). I wanted to test whether this UBE2L3-dependent effect required the proteasome which has been reported to mediate pro-IL-1α turnover

(Ainscough et al., 2014). iWTCtrl#1 and iWTYFPUBE2L3 cells were LPS primed for 6 h followed by a 9 h treatment with proteasome inhibitors MG132 or epoxomicin or autophagy inhibitors bafilomycin A or pepstatin A plus E64D (Figure 5.11). As before, pro-IL-1β levels were lower at 15 h in iWTYFPUBE2L3 cells but was increased with the addition of proteasome inhibitors (Figure 5.11, right). Also consistent with previous

192

LPS (6 h) + 9 h IL-1β

Actin

IL-1α

Actin

IL-18

Actin 1 2 3 4 5 6 7 8 9 10 Ctrl#1 YFP-UBE2L3

Figure 5.11: Caspase-1-dependent cytokines are differentially regulated by the proteasome -1 Indicated iBMDMs were treated with LPS (250 ng.mL ) for a total 15 hours in the presence of solvent (DMSO), MG132 (10 µM), bafilomycin A (BafA; 20 nM) or epoxomicin (Epox; 5 µM) as indicated to accumulate pro-IL-1β over a short (4 h) or long (9 h) period with inhibitors. Lysates were immunoblotted for pro-IL-1β, IL-1α, IL-18 and β-actin. Data are representative of 3 independent experiments. data was the observation that WTYFPUBE2L3 cells had reduced levels of pro-IL-1α at 15 h as compared to iWTCtrl#1 (Figure 5.11, IL-1α).

Interestingly, the addition of either epoxomicin or MG132 proteasome inhibitors rescued pro-IL-1α protein levels (Figure 5.11, IL-1α) though not to the levels seen in iWTCtrl#1 and not to the same extent seen with pro-IL-1β. Autophagy-inhibition, however, had no effect on pro-IL-1α levels (Figure 5.11, IL-1α). This suggested that the contribution of the proteasome to the regulation pro-IL-1α may be minor. However, it was clear from previous priming experiments (Figure 5.7) that the dynamics of pro-

193

IL-1α induction and turnover followed a different temporal trend as compared with pro-

IL-1β. The timings used in this experiment may not have been appropriate to examine pro-IL-1α targeting to the proteasome. I additionally tested pro-IL-18 protein expression which, as before, showed no difference in protein stability between cell types. Interestingly, however, the addition of proteasome inhibitors reduced the levels of pro-IL-18 protein in both iWTCtrl#1 and iWTYFPUBE2L3 cells (Figure 5.11, IL-18), the proteasome may therefore be a positive regulator pro-IL18, perhaps via a degradative effect on an unknown negative regulatory factor.

5.2.5 UBE2L3 controls pro-IL-1β levels during priming by bacteria

To examine the role of UBE2L3 under more dynamic pro-inflammatory circumstances I asked whether pro-IL-1β levels would be affected by UBE2L3 in response to bacterial infection. To avoid activating inflammasomes, which would lead to pro-IL-1β processing and pyroptosis, bacterial mutants which do not activate

NLR/ALR sensors were used (Figure 5.16). To this end, ΔprgH STm and ΔhlyA Lm were used to infect cells at an MOI of 5 for increasing periods of time. Induction of pro-

IL-1β by bacteria mirrored previous results with LPS or PAM3CSK4 (Figures 5.5A and

5.87). Western blot analysis showed that at 3 hours post-infection pro-IL-1β levels were similar between iWTCtrl#1 and iWTYFPUBE2L3 cells which were infected with either

ΔprgH STm and ΔhlyA Lm (Figure 5.12ABC). However, beyond 3 hours iWTYFPUBE2L3 infected cells failed to accumulate more pro-IL-1β and, overall, produced less protein over time as compared with pro-IL-1β than iWTCtrl#1 cells (Figure 5.12ABC).

194

Figure 5.12 UBE2L3 controls pro-IL-1β levels during priming with bacterial mutants in murine cells. Murine iBMDMs expressing YFP (Ctlr#1) or YFP-UBE2L3 were infected with STm ΔprgH (A) or Lm ΔhlyA (B) at MOI 5 and relative pro-IL-1β expression quantified by immunoblots at indicated times. Data are representative of three experiments. (A-B) Representative immunoblots and (C) quantification of relative pro-IL-1β (ratio of pro- IL-1β/β-actin) from western blotting, normalised to levels in control cells at 10 h (which were taken as 100 percent. Mean±S.E.M. from 3 independent experiments are plotted. * BH corrected P<0.05 by two-way ANOVA for indicated comparisons.

195

Figure 5.13 UBE2L3 controls pro-IL-1β levels during priming and infection with bacterial mutants in human cells THP1 cells expressing YFP (Ctlr#3) or YFP-UBE2L3 were infected with STm ΔprgH (A) or Lm ΔhlyA (B) at MOI 5 and relative pro-IL-1β expression quantified by immunoblots at indicated times. Data are representative of four experiments. (A-B) Representative immunoblots and (C) quantification of relative pro-IL-1β (ratio of pro- IL-1β/β-actin) from western blotting, normalised to levels in control cells at 10 h (which were taken as 100 percent. Mean±S.E.M. from three (ΔprgH) and two (ΔhlyA) independent experiments are plotted. * BH corrected P<0.05 by two-way ANOVA for indicated comparisons.

196

Human THP1Ctrl#1 and THP1YFPUBE2L3 also responded in a similar manner. In response to infection with ΔprgH STm THP1Ctrl#3 produced more pro-IL-1β overtime which persisted up to 24 hours post infection (Figure 5.13A). By contrast

THP1YFPUBE2L3 produced less pro-IL-1β which began to deplete after 10 hours post infection (Figure 5.13A). Similar differences were observed with ΔhlyA Lm infection

(Figure 5.13B). However, in these instances there was no depletion observed in either cell line, only a net reduction in pro-IL-1β in THP1YFPUBE2L3 over time (Figure 5.13B).

Despite the kinetics of pro-IL-1β production differing between iBMDMs and THP1 cells, the expression of YFP-UBE2L3 in human THP1 cells also reduced pro-IL-1β protein levels.

5.2.6 UBE2L3 increases the ubiquitylation of pro-IL-1β

Data so far has shown that UBE2L3 has a negative impact on pro-IL-1β stability and promotes its turnover by the proteasome. The resulting hypothesis was that UBE2L3 functioned in degradative ubiquitylation of pro-IL-1β. To examine this, I immunoprecipitated pro-IL-1β from LPS treated iWTCtrl#1 and iWTYFPUBE2L3 cells to test for differences in the extent and/or type of ubiquitylation. iWTCtrl#1 and iWTYFPUBE2L3 cells were primed with LPS for 5 hours an either left untreated or treated with 25 μM

MG132 for 1 hour. Lysates were used for immunoprecipitation using goat anti-mouse

IL-1β IgG or goat IgG as a negative control (Figure 5.14). At 6 h post-LPS-treatment, pro-IL-1β enriched from iWTCtrl#1 cells displayed no detectable modification with ubiquitin (Figure 5.13, lane 2). However, with the addition of MG132 there was an increase in polyubiquitin which co-precipitated with IL-1β (Figure 5.14, lane 3). These modifications were detectible with both anti-ubiquitin (P4D1) antibody which detects all ubiquitin types and an antibody that specifically detects K48 linked polyubiquitin.

This showed that in iWTCtrl#1 cells, pro-IL-1β was basally ubiquitylated with K48-linked

197 ubiquitin chains within 6 hours of protein induction with LPS. By contrast, pro-IL-1β immunoprecipitated from iWTYFPUBE2L3 cells displayed significantly greater levels of basal ubiquitylation (Figure 5.14, lane 4), which even exceed the levels produced in proteasome inhibited iWTCtrl#1 cells; the ubiquitin modifications were detectable with total and K48-linkage specific antibodies. The addition of MG132 further elevated the amount of detectable polyubiquitin (Figure 5.14, lane 5), to the extent that anti-IL-1β antibody detected higher molecular weight streaks by western blot, which is consistent with polyubiquitylation. UBE2L3 therefore seemed to be involved in the generation of

K48 polyubiquitin chains on newly produced pro-IL-1β, which would explain the enhanced proteasomal degradation which was observed in iWTYFPUBE2L3 cells.

198

Ctrl#1 YFP-UBE2L3

IP antibody IgG aIL-1 IgG _ _ _ MG132 + + 150 IP blot: anti-K48 Ubi 100 75 * 50 IP blot: anti-total Ubi 150 100 75 * IP blot: anti-IL-1 150 (high-exposure) 100 75 * * 50 37 IP blot: anti-IL-1 37 (low-exposure) Lysate blot: anti-IL-1 1 2 3 4 5

Figure 5.14: UBE2L3 promotes K48-linked ubiquitylation of pro-IL-1β -1 Indicated iBMDMs were primed with LPS (250 ng.mL ) for 5 h and left untreated or treated with MG132 (25 μM) for an additional 1 h, followed by immunoprecipitation (IP) using a normal goat IgG as negative control or goat anti-mouse IL-1β IgG. Cell lysates and IP fractions were immunoblotted with antibodies against K48 polyubiquitin (Ubi), total Ubi or IL-1β. * IgG heavy chain. Data are representative of 3 independent experiments.

199

5.2.7 UBE2L3 silencing enhances pro-IL-1β protein levels

The data presented thus far suggested that expression of UBE2L3 negatively regulates pro-IL-1β protein by enhancing its turnover by the proteasome. I therefore wanted to test if silencing UBE2L3 expression would produce the inverse effect. This was achieved using a miRNA30E based shRNA which targeted UBE2L3 transcripts.

Human THP1 cells were stably transduced to express two UBE2L3-targeting shRNAs

(THP1UBE2L3miR), and THP1 cells transduced to express non-targeting shRNA

Figure 5.15: Silencing UBE2L3 expression increases pro-IL-1β production (A) LPS-primed THP1 cells from THP1 cells expressing non-targeting control (Ctrl#2) or UBE2L3 specific miRNA (miR). Immunoblots show UBE2L3, NLRP3 and β-actin in lysates. (B) THP1 cells expressing non-targeting control (Ctrl#2) or UBE2L3 specific miRNA30E (miR) were treated with LPS for indicated times. Immunoblots show pro-IL-1β and β-actin in lysates. (C) Graph shows quantification of pro-IL-1β accumulation from experiments similar to those described in (C). Mean±S.E.M. of ratios of pro-IL-1β to β- actin in cell lysate from two independent experiments are plotted. * BH corrected P<0.05 by two-way ANOVA.

200

(THP1Ctrl#2) were used as negative-controls. Western blot analysis confirmed that

UBE2L3 expression was drastically reduced in THP1UBE2L3miR (Figure 5.15A).

Following stimulation with LPS, THP1UBE2L3miR accumulated more pro-IL-1β over time as compared with THP1Ctrl#2 (Figure 5.15B). Quantification of band

Figure 5.16 UBE2L3 controls pro-IL-1β levels during priming and infection with bacterial mutants that that cannot activate inflammasomes Non-targeting (Ctrl#2) or UBE2L3 miRNA expressing THP1 cells were infected with STm ΔprgH (A) or Lm ΔhlyA (B) at MOI 5 for indicated times. (A-B) Representative immunoblots show pro-IL-1β and β-actin in lysates. (C) Graph shows quantification of pro-IL-1β from (A) and (B). Data is the ratio of pro-IL- 1β / β -actin from western blotting normalised to levels in control cells at 21 h which were taken as 100 percent. Data is presented as percentage pro-IL-1β. Mean±S.D. from two experiments are plotted. * BH corrected P<0.05 by two-way ANOVA for indicated comparisons.

201 intensities confirmed this observation and highlighted that THP1UBE2L3miR cells expressed pro-IL-1β without LPS treatment (Figure 5.15C).

To examine this effect further I utilised the same non-death inducing infection conditions tested previously using ΔprgH STm and Δhly Lm. Like observations made with LPS, UBE2L3-silenced cells accumulated more pro-IL-1β protein (Figure 5.16AB) in response to infection with either bacteria. Like before THP1UBE2L3miR cells had heightened basal expression of pro-IL-1β protein (Figure 5.16C). Contrary to these observations UBE2L3 silencing in HEK293 and HeLa cells has been reported to impair

NF-kB activation thereby reducing the expression of pro-inflammatory genes (Fu et al., 2014; Lewis et al., 2015). Considering this I wanted to examine whether this response was due heightened transcriptional upregulation of pro-IL-1β. qPCR analysis was performed on cells stimulated with LPS for 3, 9 and 15 h. Interestingly, the induction of IL1B transcription in THP1UBE2L3miR was reduced compared with

THP1Ctrl#2, particularly at 9 hours post (Figure 5.17A, left). Similarly, the induction of

TNF was marginally lower in THP1UBE2L3miR cells (Figure 5.17A, right). However, despite there being comparable transcriptional induction of TNF between the cell types, more

TNF protein was secreted by THP1UBE2L3miR cells as measured by ELISA (Figure

5.18B, left). However, this was not limited to TNF as THP1UBE2L3miR cells also secreted more IL-6 in response to LPS treatment (Figure 5.17B, right). These data strongly suggested that the observed increase in pro-IL-1β protein in UBE2L3-silenced cells was not due to changes in transcription, particularly as IL1B transcriptional induction was slightly reduced in THP1UBE2L3miR cells. These data suggested that the effect of

UBE2L3 silencing in macrophages is somewhat opposite to what has been described fibroblasts and epithelial cells (Lewis et al., 2015; Fu et al., 2014) .

202

To investigate the role of UBE2L3 in cell-type specific NF-κB-dependent cytokine production further, I used the same shRNA strategy to stably silence UBE2L3 expression in HeLa cells. HeLa cells responded poorly to TNF or LPS; therefore, IL-

1β treatment was tested as it induces NF-κB-dependent IL-6 production in these cells.

Figure 5.17 UBE2L3 depletion increases has cells specific roles in priming (A-B) IL1B and TNF mRNA fold-change relative to GAPDH (B) and quantification of secreted TNF and IL-6 by ELISA (C) from indicated THP1 cells left untreated (UT) or treated with LPS (250 ng.mL-1) for times as shown. Mean±S.E.M from three independent experiments are plotted. * BH corrected P<0.05 by two-way ANOVA; ns, not significant. -1 (C) IL-6 by ELISA (left) from indicated HeLa cells treated with 2 ng.mL for 6 hours. Mean±S.E.M. from three independent experiments are shown. * BH corrected P<0.05 by two-way ANOVA. (right) Immunoblot of UBE2L3 in stated HeLa cells.

203

Indeed, silencing of UBE2L3 expression in HeLa cells caused a 3-fold reduction in IL-

6 secretion in response to stimulation with IL-1 (2 ng.mL-1) for 6 hours (Figure 5.17C).

This indicated that reduced UBE2L3 expression in fibroblast and macrophages results in markedly different effects on proinflammatory cytokine production by these cell types.

5.2.8 UBE2L3 depletion enhances IL-1β secretion by inflammasomes

It light of the opposing effects on pro-IL-1β stability caused by UBE2L3 silencing and overexpression, I asked whether UBE2L3 depletion could also impact the release of mature IL-1β. PMA-differentiated THP1Ctrl#2 and THP1UBE2L3miR cells were primed with LPS and treated with nigericin at 15-minute intervals for up to 1 h (Figure 5.19A).

Nigericin treatment induced similar levels of caspase-1 activation in both cell lines, and as expected UBE2L3 depleted from in THP1Ctrl#2 Figure 5.19A). Importantly, despite having similar caspase-1 activation, THP1UBE2L3miR secreted significantly more bioactive p17 IL-1β (Figure 5.18A, IL-1β p17). Quantification of the band intensity of mature p17 IL-1β (relative to β-actin) across multiple experiments showed that secreted mature IL-1β was approximately 4-fold higher in THP1UBE2L3miR (Figure

5.18B). Furthermore, the release of mature IL-1β was independent of cell death which was comparable in both cell types which showed similar levels of cell lysis and membrane permeabilization (Figure 5.18 C and D). Interestingly western blot analysis also showed that while pro-IL-1β was rapidly depleted from THP1Ctrl#2 cells it did not deplete from THP1UBE2L3miR despite these cells producing significantly more mature IL-

1β (Figure 5.18A, pro-IL-1β). This is in direct contrast to observations made with

THP1YFPUBE2L3 cells which depleted pro-IL-1β more rapidly despite releasing less mature cytokine (Figure 5.3). The same cell supernatants were also tested for IL-18

204 secretion. Remarkably, UBE2L3 silencing also affected the maturation and release of

IL-18 as THP1UBE2L3miR released less bioactive IL-18 protein.

As with previous experiments, I wanted to test whether the observed increase in the secretion of bioactive IL-1β was common to other inflammasomes. Activation of the AIM2 or non-canonical inflammasomes with poly(dA:dT) or LPS transfection respectively, also showed that THP1UBE2L3miR cells secreted more bioactive IL-1β as

Figure 5.18 UBE2L3 depletion increases mature IL-1β production by inflammasomes. (A) LPS-primed PMA-differentiated THP1 cells stably expressing control (Ctrl#2) or UBE2L3-specific miRNA30E-based shRNA were treated with nigericin for indicated times. Immunoblots for pro-IL-1β and β-actin (cell lysates), and caspase-1, IL-1β and IL-18 (culture supernatants) are shown. Data are representative of four experiments. (B) Quantification of mature IL-1β secretion from experiments similar to those described in (A). Mean±S.E.M. of ratios of secreted mature IL-1β to β-actin in cell lysate from four independent experiments are plotted. * BH corrected P<0.05 by two-way ANOVA. Cell lysis measured by release of LDH assays and uptake of propidium iodide (PI) in experiments similar to (A). Mean±S.E.M. from three independent experiments are shown.

205 compared with THP1Ctrl#2 cells (Figure 5.19A and B). This suggested that UBE2L3- mediated regulation of mature IL-1β release was independent of the inflammasome signalling pathway.

A B Ctrl#2 UBE2L3miR Ctrl#2 UBE2L3miR _ _ iLPS + + poly(dA:dT) _ + _ + β-actin β-actin Casp1 50 Casp1 50 37 37

p20 p20 20 20 IL-1β

IL-1β Supernatant Supernatant p17 15 p17 15

Figure 5.19 UBE2L3 depletion increases mature IL-1β production by AIM2 and non- canonical inflammasomes. Representative western blots used for IL-1β and caspase-1 in supernatants from indicated -1 -1 THP1 cells transfected with (A) poly(dA:dT) (5 μg.mL ) or (B) LPS (iLPS, 5 μg.mL ) are -1 shown. Cells were primed with 1μg.mL PAM3CSK4 for 2 hours before transfections. Data in (A) and (B) are representative of two independent experiments.

206

5.2.9 UBE2L3 modulates IL-1β release in response to bacterial infection or priming

The presence or absence of cellular UBE2L3 can alter the release of mature

IL-1β in response to robust inflammasome activation (Figure 5.3 and 5.18). The implication would be that UBE2L3 depletion by caspase-1 serves to promote IL-1β secretion. Humans are constantly exposed to non-pathogenic and commensal bacteria which despite being unable to establish an infection in healthy individuals, retain the propensity to induce inflammatory responses. Therefore, I wanted to examine the effect of UBE2L3 on inflammatory responses to physiologically relevant commensal bacteria which are weak inflammasomes activators but capable of producing a strong priming signal (Chen et al., 2016; Seo et al., 2015). Naïve PMA- differentiated THP1YFPUBE2L3 and THP1UBE2L3miR were infected with either E. coli,

Bacillus subtilis or Streptococcus gordonii at MOIs of 5 for 18 hours. Interestingly,

Figure 5.20 UBE2L3 status controls mature IL-1β production Indicated PMA-differentiated THP1 cells were infected with E. coli (Ec), B. subtilis (Bs) or S. gordonii (Sg) for 18 hours at MOI 5 and IL-1β secretion measured by ELISA. Mean±S.E.M from three experiments plotted; * P<0.05 by paired Student’s t test.

207 expression of YFP-UBE2L3 caused a noticeable reduction in the release of IL-1β following exposure to all bacterial species that were tested (Figure 5.20, left). In contrast THP1UBE2L3miR released more IL-1β upon exposure to these commensal bacteria (Figure 5.20, right). In these experiments, the bacteria themselves served as the source of signals 1 and 2 as THP1 cells were not pre-treated with priming agents like LPS or PAM3CSK4; this allowed the testing of the impact of UBE2L3 under more physiological priming conditions in response to commensals. Together, these data showed that UBE2L3 can drastically impact, not only pro-IL-1β accumulation by inflammatory MAMPs (LPS, PAM3CSK4) or proinflammatory cytokines (TNF), but also aberrant IL-1β release in response to weak inflammasome activating stimuli such as commensals.

LPS Untreated 250 ng.mL-1

pro-IL-1β 37 (low-exposure)

pro-IL-1β 37 (high-exposure)

p17 15 β-Actin 1 2 3 4

Figure 5.21 UBE2L3 silencing increases spontaneous processing of pro-IL-1β -1 Indicated THP1 cells were left untreated or treated with 250 ng.mL for 3 hours. Immunoblots show pro-IL-1β and β-actin in lysates. The same pro-IL-1β blot is shown twice at representative low and high exposures. Data are representative of two experiments.

208

Similar observations of enhanced IL-1β release have been described in macrophages deficient in the E3 ligase ITCH (Kathania et al., 2015) or the A20 ubiquitin-editing complex (Duong et al., 2015). In these cases, the authors were able to detect spontaneous release of mature p17 IL-1β.

To examine whether UBE2L3-silececing caused spontaneously release of mature IL-1β untreated and LPS-primed cells were immunoblotted using more sensitive reagents so small amounts of mature IL-1β could be detected. In untreated cells, pro-IL-1β became detectable at low exposure times first in THP1UBE2L3miR cells

(Figure 5.21); this is most small amounts of residual pro-IL-1β induced by PMA- differentiation, which are better retained in UBE2L3-silenced THP1. At the highest exposure, mature IL-1β could be detected in UBE2L3-silenced cells but not control cells. These effects were enhanced by the addition of LPS (Figure 5.16B). This suggested that loss of UBE2L3 might promote spontaneous maturation of pro-IL-1β.

5.3 Discussion This chapter has identified UBE2L3 as a regulator of pro-IL-1β stability.

Alongside an unknown E3 ligase, UBE2L3 functions in the K48-linked polyubiquitylation of pro-IL-1β protein which is subsequently targeted for degradation at the proteasome. As a result, the abundance of UBE2L3 can negatively impact IL-

1β secretion by macrophages in response to inflammasomes activation. The current working model (Figure 5.22) is as follows: inflammatory stimuli trigger the production of large quantities of pro-IL-1β cytokine and other inflammasome related proteins priming cells for inflammasome activation should the need arise. In the absence of inflammasome activation, UBE2L3 present in the cell assists in the ubiquitylation and degradation of newly produced pro-IL-1β thereby turning off a potentially dangerous inflammatory signal and returning cells to a resting non-inflammatory state. However,

209 following caspase-1 activation, UBE2L3 is itself degraded by the proteasome thereby liberating pro-IL-1 from potential degradation and instead promoting its maturation and secretion.

The precise mechanism of pro-IL-1β turnover remains elusive, particularly with regards to which E3 ligase functions with UBE2L3 to generate the K48-linked

(i) Signal 1 stimuli

Pro-IL-1β UBE2L3 ubiquitin E2 conjugating enzyme Pro-IL-1β K48-ubiquitylation & turnover of pro-IL-1β

(ii) Signal 1 + UBE2L3 degradation Signal 2 via proteasomes UBE2L3 Increased Pro-IL-1β Caspase-1

Heightened mature IL-1β release

Figure 5.22 Working model pro-IL-1β regulation by UBE2L3 (i) In the presence of Signal 1 priming or non-inflammasome activating infection UBE2L3 participates in the ubiquitylation of pro-IL-1β mediating its proteasomal thereby helping return the cell to resting non-inflammatory state. (ii) In the presence of Signal 1 and Signal 2 caspase-1 is activated and rapidly depletes UBE2L3. This serves to shut down the proteasomal turnover of pro-IL-1β increasing the available pool which may be proteolytically activated and secreted.

210 polyubiquitin on pro-IL-1β. Ubiquitylation of IL-1β has been previously reported, and reportedly consists of both anchored and unanchored K48-linked chains (Ainscough et al., 2014; Duong et al., 2015; Niebler et al., 2013) and anchored K63 chains (Duong et al., 2015). Pro-IL-1β modification with K63-linked polyubiquitin at Lys133 increases its processing by caspase-1 (Duong et al., 2015). It was shown that in response to

LPS K63-linked ubiquitylation mediated pre-inflammasome (assembly) associations, between pro-IL-1β, caspase-1, RIPK1, RIPK3 and caspase-8 (Duong et al., 2015).

The ubiquitin-editing enzyme A20 supresses pro-IL-1β processing and prevents its spurious release by specifically removing the K63-linked ubiquitin chains (Duong et al., 2015). Ubiquitylation therefore appears to play a key role in regulating pro- inflammatory activation by caspase-1 and anti-inflammatory degradation of inflammasomes and its substrates. Considering these reports and the data presented in this chapter, it could be proposed that following its synthesis cellular pro-IL-1β can exists in one of three distinct populations. Those being: unmodified pro-IL-1β; K63- linked ubiquitylated pro-IL-1β; or 48-linked ubiquitylated pro-IL-1β.

Co-IP and western blot analyse of pro-IL-1β from iWTCtrl#1 and iWTYFPUBE2L3 clearly showed that UBE2L3 overexpression caused more pro-IL-1β to be K48- polyubiquitylated. This increases the pool of pro-IL-1β which gets targeted to the proteasome. Depletion of UBE2L3 using targeted gene silencing would lessen this effect, therefore allowing pro-IL-1β to be in either an unmodified state or be modified by different linkages, for instance K63-linked chains. This would be consistent with the reduction in pro-IL-1β protein observed in YFP-UBE2L3 expressing cells and the increased protein seen in UBE2L3-silenced cells at late timepoints without caspase-1 activation.

211

Because UBE2L3 is functionally limited to HECT and RBR E3 ligases (Wenzel et al., 2011) the regulation of pro-IL-1β turnover is limited to ~28 members of the HECT and ~14 members of the RBR ligase families (Buetow and Huang, 2016). So far, the only E3 ligase described to ubiquitylate pro-IL-1β is the HECT E3 ligase E6-AP which only does so when activated by the HPV E6 protein (Niebler et al., 2013). As for A20, even though it possesses ubiquitin E3 ligase activity, it has only been shown to function as a DUB in the context of pro-IL-1β modification and regulation (Duong et al., 2015). It is of note that the phenotypes associated with UBE2L3-silencing, such as increased pro-IL-1β protein abundance and spontaneous generation of bioactive p17, are similar to that observed with the genetic deletion of A20 (which is encoded by

Tnfiaip3) (Duong et al., 2015). This could be suggestive of a degree of synergy between two. It is tempting to speculate that UBE3L3 and A20 may function together in regulating pro-IL-1β, whereby A20 removes K63-linked polyubiquitin from Lys133 and it is replaced with UBE2L3 mediated K48-linked ubiquitin chains. Indeed, UBE2L3 has been shown to function with A20 in vitro and A20 is able to generate K48-linked ubiquitin chains (Bellail et al., 2012; Bosanac et al., 2010; Wertz et al., 2004).

An interesting observation throughout these studies was the generation of p28

IL-1β as a common consequence of proteasome inhibition. The molecular weight of this fragment corresponds to pro-IL-1β cleaved at Asp27 (Asp26 in mice). This site is recognised by caspase-1, and initial cleavage at this site increases caspase-1 accessibility to Asp116 which generates the bioactive p17 form (Swaan et al., 2001).

The significance of the p28 IL-1β in the context of pro-IL-1β turnover is yet to be determined. A previous study has shown that statin-induced p28 formation is dependent on an as of yet unknown caspase but is independent of caspases -1, -6 and -8 (Davaro et al., 2014). My results confirm the involvement of a caspase and

212 show that p28 generation is also independent of caspase-4. Co-treatments with caspase and proteasome inhibitors stabilised pro-IL-1β and blocked the formation of p28. Interestingly the formation of p28 IL-1β does not seem to be required for mediating pro-IL-1β turnover, as the addition of ZVAD alone did not block its degradation. However, this would also suggest that p28 IL-1β is inherently unstable and that its formation could therefore represent yet another level of regulation. For instance, the pre-inflammasomes processing at Asp27 could “prime” IL-1β for maturation by caspase-1 at the second D116 site, and if this does not occur, the p28 is turned over at the proteasome. I further propose that the mechanisms leading to p28 formation here and the study by Davaro et al (2014) are the same. The study by

Davaro et al investigated the anti-inflammatory roles of statins which act as 3-hydroxyl-

3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors and noted that statin- treatment of macrophages induced the formation of p28 IL-1β. This had anti- inflammatory effects due to interference with IL-1R-mediated detection of bioactive p17 IL-1β.

Interestingly, statins, such as lavostatin which was one used by Davaro et al.

(2014), have been described to have cellular effects like those triggered by proteasome inhibitors, such as protein stabilization, with some arguing that this is due off target inhibition the of proteasome (Rao et al., 1999). Further, Davaro et al. (2014) noted in their study that p28 IL-1β formation was independent of HMG-CoA reductase- inhibitor function of statins (Davaro et al., 2014), which further supports my hypothesis that ability of statins to produce p28 IL-1β may be due to their inhibition of proteasomes. It is therefore plausible that these observations of p28 IL-1β may be due to its stabilisation by non-specific inhibition of the proteasome by statins. It should be

213 noted, however, that other studies have not observed proteasome-inhibition by statins

(Ludwig et al., 2005; Murray et al., 2002).

In addition to this newly described role in pro-IL-1β regulation, human

UBE2L3 has also been linked with several autoinflammatory and autoimmune diseases (Alpi et al., 2016). Genome-wide association studies (GWAS) identified multiple single nucleotide polymorphisms (SNPs) within the UBE2L3 gene which are associated with inflammatory bowel disease (Jostins et al., 2012), Crohn’s disease

(Franke et al., 2010), rheumatoid arthritis (Stahl et al., 2010) and most notable systemic lupus erythematosus (SLE) (Han et al., 2009; Lewis et al., 2015). The strongest disease-associated SNPs appear to be located in the promoter region of

UBE2L3 and lead to heighten expression of UBE2L3 and consequently elevated levels of protein (Fransen et al., 2010; Lewis et al., 2015). Lewis et al. (2015) showed that heightened expression of UBE2L3 from the SNP-related genotype (rs140490) increased NF-kB nuclear translocation in primary human B Cells and monocytes. In vitro analysis in HEK293 cells suggested this was due to enhanced LUBAC signalling, as co-expression of UBE2L3 along with either HOIL-1 and HOIP or Sharpin and HOIP caused overactive NF-κB nuclear translocation. Similarly they showed that siRNA disruption of either UBE2L3 or HOIP reduced NF-κB activation (Lewis et al., 2015).

Likewise, in this study, UBE2L3-silenced HeLa fibroblast cells showed significantly reduced IL-6 induction, an NF-B-dependent cytokine, in response recombinant IL-1β

(Figure 5.18C). HeLa cells are not known to express any inflammasome components other than caspase-4. UBE2L3 depletion is therefore unlikely to occur in these cells as previous results showed that caspase-4 alone cannot mediate this process

(Chapter 3.2.7). However, UBE2L3-silencing seemed to have the opposite effect in macrophages, which suggested a cell type-specific functionality of UBE2L3. This is

214 similar to cell type-specific roles of the LUBAC components HOIL-1 and Sharpin

(Rodgers et al., 2014). These opposing functionalities of UBE2L3 are surprising considering that it is an intermediary component of the ubiquitylation cascade.

However, it is plausible that UBE2L3 may operate with specific ligases which vary in different cell types. Inflammasomes, particularly NLRC4, also operate in epithelial cells

(Sellin et al., 2014), in which UBE2L3 is essential for NF-κB signalling, whether

UBE2L3 is similarly regulated in these cells remains unknown.

Given its important role in mediating inflammatory responses, UBE2L3 could be a potential therapeutic target. Nonetheless, such cell-specific functionality should be considered as interfering with UBE2L3 activity, stability or abundance could have opposing effects in myeloid and non-myeloid cells.

These results have established that UBE2L3 depletion promotes efficient IL-1β secretion by inflammasomes in macrophages. UBE2L3 could be thought as a molecular rheostat whereby in the presence of proinflammatory signals or infection which do not lead to inflammasome activation, UBE2L3 mediates the proteasomal degradation of newly synthesised pro-IL-1β and attempts to restore cells to a non- primed non-inflammatory state (Figure 5.22).

However, the difference in protein abundance and stability do not fully explain the disparity in IL-1β secretion from UBE2L3 -overexpressing versus -silenced cells following caspase-1 activation. Pro-IL-1β protein levels were largely similar when inflammasomes were activated 3 hours post LPS-treatment. However, what is noteworthy is that upon caspase-1 activation, pro-IL-1β depleted more rapidly from

THP1YFPUBE2L3 and then less rapidly from THP1UBE2L3miR as compared with their respective controls. This was in stark contrast to the amount of mature IL-1β they each produced and was suggestive of a dynamic and competitive ubiquitylation which

215 favours either proteasomal degradation or proteolytic processing by caspase-1.

Interestingly, in UBE2L3 silenced cells, the enhanced release of IL-1β correlated with a reduction in IL-18 processing and secretion. This could also be indicative of preferential processing of pro-IL-1β which in turn displaces pro-IL-18 from being cleaved by caspase-1. Whether these opposing ubiquitylation signals could be attached onto multiple lysine residues or solely on Lys133 remains unknown. Mass spectroscopy from Duong et al. (2015) only showed Lys133 to be modified in mouse pro-IL-1β however, they also showed that while a K→R mutation of Lys133 reduced ubiquitylation it was not eliminated completely, which suggested additional ubiquitylation sites exist (Duong et al., 2015). Likewise, whether pro-IL-1β ubiquitylation is mediated by single E3 which utilises different E2s, for example

UBE2L3 for K48 and another for K63, or multiple E3 ligases with distinct linkage species also needs to be determined. UBE2L3 has been shown to mediate both of these type of ubiquitin chain and could, in principle operate in both pathways (Sheng et al., 2012). Because UBE2L3 can only function with HECT and RBR E3 ligases it begs the question of why altering its expression produces such pronounced affects.

For RBR E3 ligases such as HOIP UBE2L3 is the preferred E2 which could be explained by its distinctive binding affinity and conformation (Dove et al., 2017; Dove et al., 2016; Lewis et al., 2015; Martino et al., 2018). If this is true for multiple RBR E3 ligases the loss of UBE2L3 could have a significant impact on their activity. Therefore, if pro-IL-1β was ubiquitylation by an RBR E3, the loss of UBE2L3 by caspase-1 could inhibit this process.

Alternatively, UBE2L3 is might not impact linkage specificity a broad range of

E3 ligases. Some enzymes have been described to mediate so called “en bloc” ubiquitin transfer whereby the polyubiquitin chain is formed on the catalytic cysteine

216 of an E3 ligase and subsequently transfer as whole to the substrate (Ronchi et al.,

2013; Wang et al., 2006). E6-AP has been described to do this and interestingly has been shown to simultaneously interact with two E2 enzymes via two distinct binding sites (Ronchi et al., 2017; Ronchi et al., 2013). In this context because the ubiquitin chain is being built on the E3 cysteine it could mean that the decision of linkage type is relinquished to the E2 enzymes. Therefore, switching of the E2 enzyme, for example after UBE2L3 depletion, could alter the functionality of an E3 ligase like E6-

AP. This would be an interesting avenue to explore, particularly as E6-AP has been described to ubiquitylate pro-IL-1β (Niebler et al., 2013).

Whilst these studies have showed that UBE2L3 depletion can impact inflammasome related mature IL-1β release, cell death due to pyroptosis would inherently limit the time within UBE2L3 depletion can have an effect. Recent studies have shown that certain cell types do not undergo pyroptosis despite the activation of inflammatory and caspase-1 (Chen et al., 2014; Gaidt et al., 2016; Zanoni et al., 2016).

Such cells are therefore better able to release more bioactive IL-1β over longer periods of time. It would be interesting to examine whether UBE2L3 is similarly regulated by caspase-1 in these cells and if it has a more sustained effect on IL-1β release in the absence of cell death, much like I observed with commensal infection.

UBE2L3 has been shown to be a binding partner for a number of bacterial effectors, most notably those which display E3-ligase activity, such as the Salmonella

SPI-1 effector SopA, EHEC effector NleL (Diao et al., 2008; Lin et al., 2011; Piscatelli et al., 2011; Sheng et al., 2017), and Legionella SidC (Hsu et al., 2014) which have been shown to bind and utilise UBE2L3 as their cognate E2. The non-E3 ligases effector OspG from Shigella spp. also interacts with E2 enzymes, and has been shown to bind and inhibit ubiquitin-conjugated UBE2L3 in cells (Grishin et al., 2014; Kim et

217 al., 2005; Pruneda et al., 2014; Zhou et al., 2013). An interesting dynamic may therefore exist whereby both host and pathogen are at odds over which can reach

UBE2L3 first, the host via caspase-1 to eliminate it from the cell or the pathogen via an effector to utilise UBE2L3 for its own means.

5.4 Summary In summary, the work presented in this chapter has shown that UBE2L3 acts as a molecular rheostat which regulates IL-1β. Sustained expression of UBE2L3 in the presence of active caspase-1 reduced the secretion of bioactive IL-1β independently of reductions in caspase-1 activity, transcriptional priming and cell death. UBE2L3 mediated pro-IL-1β K48-linked polyubiquitylation as this was increased in YFP-

UBE2L3 expressing cells and led to increased pro-IL-1β degradation by the proteasome. Silencing UBE2L3 expression had the inverse effect. UBE2L3 appeared to influence pro-IL-1β levels independently of transcription, and its loss enhanced the secretion of mature IL-1β without impacting caspase-1 activation or cell death.

Alteration of cellular UBE2L3 had a significant effect on macrophage responses to commensal bacteria, which pointed to a role as a modulator of inflammation. These findings also showed that UBE2L3 possesses distinct cell type-specific functions as seen from opposing roles in different cells from different lineages.

218

Chapter 6 – UBE2L3 is polyubiquitylated by the enterohemorrhagic E. coli effector NleL

219

6.1 Introduction Ubiquitylation is a common process within eukaryotic cells and is involved in the regulation of diverse molecular pathways ranging from cell cycle progression to immune signalling (Komander and Rape, 2012). Because ubiquitylation is central to so many cellular processes it is unsurprising that it is often targeted by pathogens as a means of hijacking and subverting host cell responses to promote their own survival and replication (Ashida and Sasakawa, 2017; Li et al., 2016; Maculins et al., 2016).

Though they lack a ubiquitin system of their own, many bacterial species produce diverse effector proteins which mimic eukaryotic ubiquitylation enzymes including

DUBs and E3 ligases (Maculins et al., 2016). These bacterial E3 ligases include enzymes which are structurally and functionally similar to eukaryotic RING/UBox or

HECT E3 ligases. Some bacterial pathogens possess a third class of E3 ubiquitin ligase which is absent from eukaryotic species called Novel E3 Ligases (NELs)

(Ashida and Sasakawa, 2017). Bacteria do not produce other ubiquitylation enzymes such as E1s or E2s and must hijack the host ubiquitylation machinery to function. I the context of UBE2L3 bacterial HECT E3 ligases are of particular interest as these can form a functional partnership. To date two HECT type bacterial E3 ligases have been characterised; SopA from STm, and NleL from EHEC both of which have been structurally characterised with UBE2L3 acting as there cognate E2 (Lin et al., 2012;

Lin et al., 2011).

Like there eukaryotic counterparts the HECT domains of bacterial ligases have a bi-lobal structure (Figure 6.1) which consists of a C-terminal lobe that contains the catalytic cysteine (Cys753 in both SopA and NleL) and a N-terminal lobe which regulates E2 enzyme binding (Diao et al., 2008; Huang et al., 1999; Rotin and Kumar,

2009). These two structures are linked by a flexible hinge region which enables correct

220 positioning of the catalytic domain so it may perform transthioesterification of ubiquitin from the E2 to the E3 (Diao et al., 2008; Huang et al., 1999; Rotin and Kumar, 2009).

In addition to the core HECT domain eukaryotic these ligases have various N-terminal domains which confer substrate specificity and mediate additional regulatory roles

(Buetow and Huang, 2016). Similarly, bacterial HECTs have an N-terminal β-helix domain which can mediate interactions with substrate proteins (Fiskin et al., 2017;

Sheng et al., 2017).

A C-lobe

β-Helix

N-lobe UBE2L3 B

Cys753 1 170 370 593 615 782

β-helix N-lobe Hinge C-lobe

HECT Domain

Figure 6.1 Crystal structure and domain organisation NleL (A) Crystal structure of UBE2L3 and NleL from the E. coli O157:H7 str. Sakai. Ribbon representation of UBE2L3 (blue) complex with NleL (green). PDB ID: 3SY2 (Lin et al., 2012). (B) Schematic representation of NleL domain organisation. Key domains are coloured and labelled. The catalytic cysteine (Cys753) responsible for E3 ligase activity is shown in red (Adapted from (Sheng et al., 2017)).

221

Despite being structurally similar to eukaryotic enzymes SopA and NleL interact with E2 enzymes via a different region of their N-lobe domain (Lin et al., 2012).

However, these regions make contact with the canonical E3 binding surface of E2 enzymes which typically consists of the α-helix1 and loop 4 and 7 (Chapter 1, Figure

1.6) (Diao et al., 2008; Huang et al., 1999; Lin et al., 2012). Specifically, key interactions for SopA/UBE2L3 include Arg5 and Arg6 of UBE2L3 which form hydrogen bonds with SopA Gly515 and Tyr566 respectively (Lin et al., 2012). UBE2L3 Phe63 also forms van der Waals interactions with SopA Glu564 and Tyr566 (Diao et al., 2008; Lin et al., 2012). Mutating either one of these residues in UBE2L3 significantly reduces its affinity for SopA binding (Diao et al., 2008; Lin et al., 2012). For UBE2L3/NleL binding, the same hydrogen bond forming resides are not required for interactions with UBE2L3

(Lin et al., 2012). However, Phe63 remains an important residue and forms close associations with Phe569 which, if mutated drastically reduces NleL activity (Lin et al.,

2012). Both SopA and NleL have broad E2 enzyme specificity in vitro. SopA has been shown to function with D1, D3, E1 and L3 E2 enzymes (Zhang et al., 2006) and NleL can utilise D1, D2, D3, L6 and L3 (Lin et al., 2011).

Both have also been shown to produce different linkage types of polyubiquitin;

SopA was recently shown to generate K11- and K48-linked polyubiquitin on its substrate proteins (Fiskin et al., 2017), while NleL is able to synthesise unanchored

K48- and K6-linked polyubiquitin (Hospenthal et al., 2013; Lin et al., 2011) and anchored K27-, K29- and K33- polyubiquitin chains (Sheng et al., 2017). Whether or not NleL produces unanchored ubiquitin chains in host cells or in the presence of a substrate remains unknown. However, unanchored polyubiquitin has been reported to act as a second messenger in cells and can specifically activate cellular kinases such as TAK1 (Xia et al., 2009) and IKKε (Rajsbaum et al., 2014) which function in the

222 activation interferon and antiviral immune responses (Rajsbaum and Bharaj, 2016;

Rajsbaum et al., 2014; Zeng et al., 2010).

Until recently the cellular targets and actions of SopA and NleL were unknown.

SopA had been shown to contribute to enteritis and diarrhoea during infection (Wood et al., 2000; Zhang et al., 2002). More recently two host E3 ligases, TRIM56 and

TRIM65 were identified as its substrates (Kamanova et al., 2016). How their targeting by SopA results in inflammation is not yet fully understood; one study has suggested that SopA mediated ubiquitylation of TRIM56 and TRIM65 increased their activity which causes elevated inflammatory gene expression via IRF3 (Kamanova et al.,

2016), however, has also been shown that SopA attaches K11- and K48-linked polyubiquitin on to TRIM56 and TRIM65 which targets them for degradation at the proteasome (Fiskin et al., 2017).

The EHEC effector NleL was first identified in Citrobacter rodentium (also called

EspX7) (Tobe et al., 2006) and has been shown to affect infection of host epithelial cells by regulating the formation of attaching and effacing lesions (A/E lesions, also referred to as actin pedestals) which are vital for bacterial colonisation of the host by certain pathogenic species (Frankel and Phillips, 2008; Piscatelli et al., 2011; Sheng et al., 2017). One study by Piscatelli et al. showed that EHEC expressing catalytically- dead mutant NleLC753A on the chromosome had increased cell attachment and pedestal formation (Piscatelli et al., 2011). However, a recent study by Sheng et al. showed that NleL had the inverse effect, as EHEC which lacked nleL displayed reduced cell attachment and pedestal formation (Sheng et al., 2017). They showed that NleL modified the host kinase JNK1 with mono-ubiquitin, K29-linked and K33- linked polyubiquitin chains (Sheng et al., 2017). Ubiquitylation of JNK1 interfered with its ability to interact with other proteins, thereby blocking its activation by the upstream

223 kinase MKK7 and suppressing the activation of the AP-1 transcription factor (Sheng et al., 2017). They went on to show that either knockdown or inhibition of host JNK1 complemented the attachment defect of associated with NleL deficiency suggesting that JNK1 activity was required for the formation of actin pedestals (Sheng et al.,

2017).

Multiple in vitro studies have used UBE2L3 to examine the catalytic activity of multiple bacterial E3 ligases (Fiskin et al., 2017; Kamanova et al., 2016; Sheng et al.,

2017). These include SopA, NleL (Hospenthal et al., 2013; Lin et al., 2012) and SidC

(Legionella pneumophila) (Hsu et al., 2014; Luo et al., 2015). Interestingly, the non- ligase effector OspG from Shigella flexneri has also been shown to bind UBE2L3

(Grishin et al., 2014; Kim et al., 2005; Pruneda et al., 2014). OspG is in fact a kinase, which has been reported to interfere with innate immune signalling by obstructing the proteasomal degradation of IκB thereby blocking the activation and nuclear translocation of NF-κB (Kim et al., 2005). Consequently, S. flexneri which lack OspG trigger a more robust inflammatory response (Kim et al., 2005). OspG was found to bind multiple E2 conjugating enzymes, (Kim et al., 2005), specifically those charged with ubiquitin (Grishin et al., 2014; Pruneda et al., 2014). However, E2 enzymes do not function as substrates for its kinase activity (Zhou et al., 2013). Rather the binding of ubiquitin-charged E2 (E2~Ub) enzymes activates the kinase activity of OspG

(Grishin et al., 2014; Pruneda et al., 2014). In addition to its role as an activating co- factor, E2~Ub binding protects OspG from proteasomal degradation (Grishin et al.,

2014). OspG appears to have a broad specificity for different E2~Ub enzymes however, proteomic studies showed that it preferentially interacts with UBE2L3 which does so via the canonical E3 binding surface (Grishin et al., 2014; Pruneda et al.,

2014). This led multiple investigators to postulate whether OspG binding interfered

224 with E2 enzyme activity by competing with E3 ligases. Indeed, multiple studies have found that OspG can block E2 discharging (Pruneda et al., 2014). Grishin et al. showed that the OspG and UBE2L3~Ub interacted with a high affinity which was greater than

UBE2L3~Ub interactions with E3 ligases PARKIN and E6AP (Grishin et al., 2014).

This tight association was able to inhibit PARKIN meditated polyubiquitination in vitro when OspG and UBE2L3 were added at a 1:1 molar ratio (Grishin et al., 2014).}.

Recently OspG has also been shown to block LUBAC mediated NF-κB activation (de

Jong et al., 2016). This effect was independent of its kinase activity and is believed to be due to the inhibition of UBE2L3 (de Jong et al., 2016) which is consistent with previous reports of the preferential interaction between UBE2L3 and LUBAC (Lewis et al., 2015). These studies with OspG highlight that, in the context of bacterial effectors, UBE2L3 can also act as a specific effector target.

Such a dynamic could also be true of E3 ligase effectors. Indeed, within the literature, there are several reports of UBE2L3 being ubiquitylated by bacterial E3 ligases. The L. pneumophila effector SidC is a member of a new and structurally distinct family of E3 ligases (Hsu et al., 2014) and is involved in the maturation of LCVs and pathogen survival (Hsu et al., 2014). SidC lacks any sequence or structural similarity to other E3 ligase families, but like RBR and HECT enzymes it possesses its own catalytic activity (Hsu et al., 2014). In the same study while examining the catalytic activity and requirements of SidC the investigators also noted that UBE2L3 was modified with either multi-mono or polyubiquitin, however, the reasons and consequences behind this were not explored (Hsu et al., 2014). NleL has also been shown to ubiquitylate UBE2L3 specifically on lysine residues (Hospenthal et al., 2013;

Lin et al., 2011).

225

Ubiquitylation of E2 conjugating enzymes is not a novel phenomenon (Stewart et al., 2016) and has been seen to occur within large N- and C-terminal domain extensions which some E2 enzymes, but not UBE2L3, possess (Banka et al., 2015;

Schumacher et al., 2013). Multiple proteomics-based studies examining cell ubiquitylomes have also suggested that UBE2L3 is innately ubiquitylated in epithelial cells (Kim et al., 2011b; Mertins et al., 2013; Wagner et al., 2011). Though the purpose of these modifications is yet to be determined a recent study revealed that ubiquitylation of UBE2L3 decreases in HCT116 cells infected with STm in a SPI-1 dependent manner (Fiskin et al., 2016). Modulation of UBE2L3 ubiquitylation might, therefore, have important roles during bacterial infection.

With this concept in mind, I set out to further characterise the interactions and modification of UBE2L3 by bacterial E3 ligase. For these studies, I focused on bacterial HECT ligases. Particularly I have focused on NleL as its propensity for free polyubiquitin made its activity easy to study and because of it both functions with and ubiquitylates UBE2L3.

6.2 Results 6.2.1 UBE2L3 is rapidly ubiquitylated by NleL

While examining free K6-chain formation by NleL, Hospenthal et al. noted that

UBE2L3 was ubiquitylated during in vitro ubiquitylation assays with NleL. They went on to show that this occurred on lysine residues as a lysine-less UBE2L3 protein could not be ubiquitylated (Hospenthal et al., 2013). Previous studies had used NleL protein lacking the first 169 residues which were believed to be unstructured and contributed to protein instability during crystallisation (Diao et al., 2008; Lin et al., 2011). To remain

226 consistent with previously published work which showed UBE2L3 ubiquitylation, the same region of NleL was used here, consisting of amino acid residues 170-782

(NleL170-782) and similar reaction conditions were used throughout (Hospenthal et al.,

2013). NleL170-782 was cloned into the pGEX6P1 expression plasmid, to add an N- terminal GST-tag. The protein was subsequently purified using glutathione sepharose resin, and the GST-tag was removed by cleavage with rhinovirus C3 protease.

Firstly, I wanted to confirm that UBE2L3 was ubiquitylated by NleL. Reactions containing NleL170-782, UBE2L3, recombinant human E1, and wildtype ubiquitin were activated using ATP and stopped by adding 2x Laemmli buffer containing DTT.

Western blot analyses confirmed that UBE2L3 is modified by NleL in vitro which creates multiple species of ubiquitylated UBE2L3. The predicted molecular weights

(Mw) of these proteins (Figure 6.2A) corresponded to the observed Mw of UBE2L3 plus sequentially added ubiquitin molecules as determined by SDS-PAGE (Figure

6.2B).

Figure 6.2 NleL makes UBE2L3 exists as multiple species of increasing molecular weight (A)Table of the approximate Mw of unmodified and modified UBE2L3 species and (B) 170-782 immunoblot (anti-UBE2L3) of ubiquitylated UBE2L3 from in vitro assay with NleL and wildtype ubiquitin.

227

To examine the kinetics of these modifications similar in vitro assays were performed and reactions were halted at varying timepoints, allowing for time course analysis of NleL activity to be observed (Figure 6.3A). Because NleL makes unanchored free-ubiquitin chains its activity can be determined by western blotting with anti-ubiquitin antibodies. Likewise, by immunoblotting with a UBE2L3-specific antibody, the extent of UBE2L3 modifications by NleL can also be determined.

Reactions lacking ATP or NleL were included as negative controls. Prior to use 2 μg of each protein was analysed by SDS-PAGE and Coomassie staining to assess the purity of recombinant protein. Within this amount, no impurities could be detected

(Figure 6.3C)

Western blot analysis of UBE2L3 showed that its ubiquitylation by NleL is a rapid process. Within the few seconds between activating and halting the reaction, an additional ~26 kDa band is observed in the 0 min sample (Figure 6.3A, lane 1). This corresponded to UBE2L3 (18 kDa) plus a single ubiquitin molecule (8.5 kDa) (Figure

6.1). Interestingly, this preceded the formation of unanchored ubiquitin chains or autoubiquitylation of NleL as no polyubiquitin could detected in the same sample when immunoblotted with anti-ubiquitin antibody (Figure 6.3A, lane 1). However, unanchored ubiquitin chains were readily detected from 5 mins onwards (Figure 6.3A).

Within 5 minutes, four distinct bands of modified UBE2L3 (ranging 25 – 50 kDa) and high Mw streaks are observed (Figure 6.3A, lane 2). Within 10 minutes a fifth band at ~60.5 kDa appeared (Figure 6.3A, lane 3), the Mw of these modified proteins also correspond to that of UBE2L3 plus two, three, four and five ubiquitin molecules

(Figure 6.2). UBE2L3 modification was dependent on the catalytic activity of E1 enzyme and NleL as reactions lacking either ATP or NleL showed no modification of

UBE2L3 (Figure 6.3A, lanes 7 and 8). Interestingly, after 15 minutes the amount of

228 modified UBE2L3 appeared to reduce as fewer bands corresponding to modified and polyubiquitylated UBE2L3 were seen (Figure 6.3A, top, lanes 4 to 6). This same pattern was not observed when examining total free ubiquitin chains, which continued

229

◄Figure 6.3: UBE2L3 is rapidly ubiquitylated by NleL in vitro. 170-782 FL Purified recombinant (A) NleL or (B) NleL were incubated with wildtype UBE2L3 for stated times. (A) Reactions without ATP or E3 ligase were included as negative controls. UBE2L3 modification was detected using anti-UBE2L3 antibody (top), ligase activity was determined by detection of polyubiquitin by western blot using anti-ubiquitin HRP antibody (bottom). Reactions were stopped using 2x Laemmli dye containing 100 µM DTT. Data are representative of two independent experiments. * non-specific bands. (C) Assessment of recombinant protein purity. ~2 μg of purified wildtype human UBE2L3, FL 170-782 NleL , NleL and human UBE1 were separated by SDS-PAGE and stained with Coomassie dye. to increase over time (Figure 6.3A, bottom, lanes 4 to 6). It is possible that highly modified UBE2L3 did not transfer from the gels however, this is unlikely as high molecular weight polyubiquitin chains were observed. Most subsequent assays were therefore performed for 20 mins.

Many E3 ligases regulate their activity through intramolecular interactions, for example by interactions between the N-terminus and catalytic domains (Buetow and

Huang, 2016). To ensure that NleL mediated modification of UBE2L3 was not an artefactual feature of NleL170-782 due to N-terminal truncation, full length NleL (NleLFL) was also tested. As before, NleLFL was used for time course analysis which showed that it facilitated rapid ubiquitylation of UBE2L3 like NleL170-782 enzyme (Figure 6.3B).

Similarly, at later timepoints the amount of modified UBE2L3 appeared to decrease while the total polyubiquitin continued to increase (Figure 6.3B, lanes 3 to 5).

Truncated and full-length NleL proteins were therefore both capable of modifying

UBE2L3 in vitro which suggested this property resides within the 170-782 region of

Nlel. To test whether this was common amongst bacterial HECT ligases SopA was also tested in vitro. In contrast to NleL, SopA does not form free ubiquitin chains, and

230 its E3 ligase activity was therefore determined by the presence of SopA autoubiquitylation (Lin et al., 2011). Similar time-course analysis was performed using the HECT and N-terminal portions of SopA (SopA168-782) which has previously been co-crystallised with UBE2L3 (Lin et al., 2012). Surprisingly, UBE2L3 is not modified by

SopA as no additional bands were detected despite both enzymes being catalytically active (Figure 6.4) These data confirmed previous observations that UBE2L3 is ubiquitylated by NleL in vitro. However, SopA, which is of the same family of proteins

Figure 6.4: UBE2L3 is not ubiquitylated by SopA in vitro. 163-782 Purified recombinant SopA was incubated with wildtype UBE2L3 for stated times. Reactions without ATP or E3 ligase were included as negative controls. UBE2L3 was detected using anti-UBE2L3 antibody (top), ligase activity was determined by detection of SopA autoubiquitylation by western blot using anti-ubiquitin HRP antibody (bottom). Reactions were stopped using 2x Laemmli dye containing 100 µM DTT. Data are representative of two independent experiments.

231 fails to cause UBE2L3 ubiquitylation. I therefore proceeded to further assess the mechanisms of UBE2L3 ubiquitylation by NleL biochemically.

6.2.2 UBE2L3 is trans-ubiquitylated by NleL

Previous studies that noted UBE2L3 ubiquitylation by bacterial E3 ligases did not go on to determine the underlying mechanisms in greater detail. As I had found important roles for UBE2L3 (Chapter 5) in controlling inflammatory responses in multiple cell types, I decided to investigate its interactions with a bacterial E3 ligase.

Because UBE2L3 is modified I hypothesised that, it may also be a specific target of

NleL during EHEC infection, as well as acting as an E2 enzyme. A key question was how, in the context of a ubiquitylation reaction, is UBE2L3 modified at the molecular level. Given the components of the in vitro assays two potential scenarios seemed likely: first, UBE2L3 ubiquitylation occurred on the same molecule that also served as a ubiquitin-donor E2 (cis ubiquitylation), or second, UBE2L3 was ubiquitylated by a separate E3-E2 pair (trans ubiquitylation) (Figure 6.5A).

To test these possibilities, it was important to be able to distinguish between different molecules of UBE2L3. Different variants of UBE2L3 were used to serve either as a ubiquitin donor or acceptor in assays. First, a lysine-less UBE2L3 (UBE2L318KR) was used as the ubiquitin donor as it was unable to be ubiquitylated but, retained full activity (Figure 6.5B) (Hospenthal et al., 2013). Two variants were used as ubiquitin acceptors, neither of which could function in the catalysis of ubiquitin transfer. These included a catalytically inactive UBE2L3 mutant where the catalytic cysteine residue

86 was mutated to alanine (UBE2L3C86A), or a variant of UBE2L3 where two key residues required for E3 binding (Pro62 and Phe63) were mutated to alanine

(UBE2L3P62A, F63A) (Figure 6.5B). Mutating all 18 lysine’s in UBE2L3 prevents its own

232 ubiquitylation without affecting its activity as lysine-less UBE2L318KR facilitated the formation of unanchored ubiquitin chains (Figure 6.5C, lane 1).

233

◄ Figure 6.5: NleL mediates trans ubiquitylation of UBE2L3 lysine residues in vitro. (A) Schematic representation of cis and trans ubiquitylation of UBE2L3 by NleL. For sake of clarity, only a single lysine acceptor site on UBE2L3 is depicted. (B) Coomassie of recombinant UBE2L3 variants, ~2 μg of stated UBE2L3 variants were separated by SDS-PAGE and stained with Coomassie dye to assess the purity of the prep. 170-782 FL Purified recombinant NleL (C), or NleL (D), were incubated with UBE2L3 variants; 18KR, C86A and P62A F63A for in vitro ubiquitylation assays with wildtype ubiquitin and UBE1. UBE2L3C86A or UBE2L3P62A F63A proteins were incubated with NleL variants without or with the addition of UBE2L318KR for 20 mins. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped using 2x Laemmli dye containing 100 mM DTT. Activity was determined by detection of polyubiquitin by western blot using anti-ubiquitin HRP antibody (bottom panel), and ubiquitylated UBE2L3 was detected using anti-UBE2L3 antibody (top panel). Western blots are representative of 3 independent experiments.

As expected, in assays with UBE2L3C86A alone neither unanchored chains or

UBE2L3 ubiquitylation were observed as this variant lacked catalytic activity (Figure

6.5C, lane 2). Likewise, the E3 binding mutant UBE2L3P62AF63A had significantly reduced activity and was poorly ubiquitylated (Figure 6.5C, lane 4.) Importantly, however, when both the ubiquitin donor (UBE2L318KR) and receiver (UBE2L3C86A or

UBE2L3P62AF63A) variants were combined in a single reaction UBE2L3 ubiquitylation was restored (Figure 6.5C, Lanes 3 and 5) to levels similar to wildtype UBE2L3 (Figure

6.5C, Lane 6). This suggested that UBE2L318KR, along with NleL, trans-ubiquitylated

UBE2L3C86A or UBE2L3P62AF63A variants in vitro. To confirm that full-length NleL enzyme behaved in a similar manner, NleLFL was also used in these assays (Figure

6.5D). Full-length protein gave results identical to those with NleL170-782 (Figure 6.5C).

Together these results established that ubiquitylation of UBE2L3 by NleL occurred in trans.

234

6.2.3 UBE2L3 is ubiquitylated on five residues

Ubiquitylated proteins can exist in many forms, for example carrying mono or polyubiquitin chains on individual or multiple residues. In addition, chain-lengths on each acceptor site may vary and include different kinds of chain type or branches

(Swatek and Komander, 2016; Yau and Rape, 2016). Previous results showed that multiple species of ubiquitylated UBE2L3 were formed when combined with NleL in vitro. These included high Mw polyubiquitylated UBE2L3, and distinct populations of mono, di, tri, tetra and penta-ubiquitylated UBE2L3 which could be seen more clearly by immunoblotting. These populations could represent varying degrees of multi- monoubiquitylation and/or chains of increasing length on a single residue on UBE2L3

(Figure 6.6A).

To explore the ubiquitin architecture of NleL-modified UBE2L3, ubiquitylation assays were performed using His-tagged lysine-less ubiquitin (His-NoK; Mw 9.3 kDa)

His-NoK ubiquitin is unique in that it can only be used in mono-ubiquitylation. This is because once it has been ligated via its C-terminal glycine residue it has no lysine residues with which it can form chains and the N-terminal His-tag blocks methionine- linked linear chain formation. Therefore, depending on the number of modified

UBE2L3 variants which are formed with His-NoK it can be determined whether

UBE2L3 modified on one or multiple residues. Western blotting of in vitro ubiquitylation assays showed that His-NoK ubiquitin was successively utilised by NleL in an ATP dependent manner. This was evident as ubiquitin was detected above the molecular weight of NleL using anti-ubiquitin antibody (Figure 6.6B, right panel), indicating that

NleL-autoubiquitylation had occurred. In reactions containing wildtype UBE2L3, multiple bands of ubiquitin were observed ranging from 10-250 kDa (Figure 6.6B, right panel, lane 2). Blotting the same sample with anti-UBE2L3 antibody revealed that the

235 bands between 25-75 kDa corresponded to ubiquitin-modified UBE2L3 (Figure 6.6B, left panel, lane 2). Not including unmodified UBE2L3 (~17 kDa), five additional bands were observed, suggesting that UBE2L3 was modified on five separate residues

(Figure 6.3A). These residues appeared to be lysine’s as the lysine-less UBE2L3

(UBE2L318KR) was not modified despite retaining its catalytic activity as evident by the continued autoubiquitylation of NleL (Figure 6.6B, lane 3).

Figure 6.6: UBE2L3 can be ubiquitylated on 5 residues (A) Schematic representation of possible ubiquitylated species of UBE2L3 170-782 (B) Purified recombinant WT UBE2L3 or 18KR UBE2L3 were incubated with NleL and His-NoK Ubiquitin for 20 mins. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped using 2x laemmli dye containing 100 mM DTT. Total ubiquitin and UBE2L3 modification were detected by western blotting using anti-UBE2L3 antibody (top panel), and anti-ubiquitin HRP antibody (bottom panel). Western blots are representative of 2 independent experiments.

236

6.2.4 UBE2L3 is predominantly modified on one lysine with multiple linkage types

UBE2L3 contains eighteen lysine residues, five of which appear to be capable of being ubiquitylated by NleL. The next question that arose was, which lysine residues were modified? To begin to narrow the potential candidate residues an online protein modification database PhosphoSitePlus® (Hornbeck et al., 2015) was used to investigate whether any residues of UBE2L3 had previously been reported to be ubiquitylated. Fourteen lysine’s had been reported to be ubiquitylated on UBE2L3, primarily from mass spectrometry data sets. Out of these, 5 where chosen which included; K64, K82, and K131 which had greater than five references asserting ubiquitylation, and K138 and K150. To study whether these residues were ubiquitylated by NleL, they were individually reintroduced into UBE2L318KR generating variants which contained a single lysine residue. In parallel, the same lysine residues in wildtype UBE2L3 were mutated to arginine thereby blocking ubiquitylation at these specific residues. Each UBE2L3 variant was tested in assays with NleL170-782 and wildtype ubiquitin to determine if any of these five lysine residues were the target of ubiquitylation (Figure 6.7).

Alongside the lysine only variants, wildtype UBE2L3, and UBE2L318KR were included as positive and negative controls respectively. Interestingly, four out of the five single-lysine UBE2L3 variants (UBE2L3K64only, UBE2L3K131only, UBE2L3K138only,

UBE2L3K150only) were not ubiquitylated despite retaining catalytic activity as judged by the formation of unanchored ubiquitin chains (Figure 6.7A). These variants thus behaved like UBE2L318KR (Figure 6.7A). Notably, only K82 was ubiquitylated to levels comparable to wildtype UBE2L3 (Figure 6.7A, lane 5). Western blot analysis with anti-

UBE2L3 antibody showed five clearly distinct ubiquitylated forms plus high molecular

237

Figure 6.7: NleL modifies lysine 82 of UBE2L3 with polyubiquitin WT, 18KR and stated lysine-only (A) or lysine-to-arginine (B) variants of UBE2L3 were 170-782 incubated with NleL and K48 only ubiquitin for 20 mins. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped with 2x laemmli dye containing 100 mM DTT. Total ubiquitin and UBE2L3 modification were detected by western blotting using anti-UBE2L3 antibody (top panel), and anti-ubiquitin HRP antibody (bottom panel). Western blots are representative of two independent experiments. (C+D) Coomassie of recombinant UBE2L3 variants, ~2 μg of stated UBE2L3 variants were separated by SDS-PAGE and stained with Coomassie dye to assess the purity of the prep.

238

polyubiquitylated UBE2L3 as was observed with wildtype UBE2L3 (Figure

6.7A, lane five). From this it can be concluded that out of the five lysine residues tested,

K82 can be polyubiquitylated by NleL even when presented alone and to an extent comparable to the wildtype protein. However, because it produced an immunoblotting pattern similar to the wildtype protein it remained unclear if other lysine residues which were not mutated in our initial screen of five could also be ubiquitylated.

Figure 6.8: Lysine 82 does not contribute to ligase activity 170-782 Stated variants of UBE2L3 were incubated with NleL and wildtype ubiquitin for indicated periods of time (min) following the addition of ATP. Reactions were stopped using 2x Laemmli dye containing 100 µM DTT. Total ubiquitin and UBE2L3 modification were detected by western blotting using anti-UBE2L3 antibody (top panel), and anti-ubiquitin HRP antibody (bottom panel). Western blots are representative of 3 independent experiments

239

To test this further the same five lysine residues were mutated to arginine to generate UBE2L3K64R, UBE2L3K131R, UBE2L3K138R and UBE2L3K150R variants (Figure

6.7B). Mirroring the results from the single-lysine UBE2L3 variants, K→R mutation of the same four residues (K64R, K131R, K138R, K150R) did not affect the ubiquitylation of UBE2L3 or its ability to form unanchored polyubiquitin chains by NleL (Figure 6.7B).

However. the introduction of the K82R mutation drastically reduced UBE2L3

Figure 6.9: Full length NleL ubiquitylates UBE2L3 at lysine 82 170-782 FL Stated variants of UBE2L3 were incubated with either NleL or NleL with wildtype ubiquitin for 20 mins. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped using 2x Laemmli dye containing 100 μM DTT. UBE2L3 and total ubiquitin modification were detected by western blotting using anti-UBE2L3 (top panel) and anti-ubiquitin HRP antibodies (bottom panel). Experiments were repeated twice.

240 ubiquitylation which suggested that K82 is the principal site ubiquitylated by NleL

(Figure 6.7B, lane five). However, UBE2L3K82R was still weakly ubiquitylated by NleL

(Figure 6.7B, lane five), it was possible that these modifications could represent additional lysine residues which were modified with His-NoK ubiquitin (Figure 6.6).

However, these findings suggested that the majority of UBE2L3 ubiquitylation was attached to K82. It is possible that in lieu of making ubiquitin chains in assays with His-

NoK ubiquitin, NleL ubiquitylates weaker acceptor sites.

Interestingly, ubiquitylation of K82 does not appear to affect the activity of

UBE2L3 with NleL. Side-by-side comparative assays with UBE2L3K82R, UBE2L3K82only and UBE2L318KR proteins at multiple time points showed that all UBE2L3 variants produced similar amounts of wildtype-polyubiquitin over time (Figure 6.8, bottom).

Again, the ubiquitylation of UBE2L3 decreased over time (Figure 6.8, top), and was observed with wildtype and UBE2L3K82only protein (Figure 6.8, top). To corroborate that the observed preference for K82 was not specific to the shorter NleL170-782 protein,

UBE2L3K82only and UBE2L3K82R variants were also tested with NleLFL (Figure 6.9). Both forms of NleL behaved similarly and only ubiquitylated wildtype and UBE2L3K82only proteins. However, NleLFL appeared to produce greater amounts of modified UBE2L3

(Figure 6.9, top).

6.2.5 NleL modifies lysine 82 with K6 and K48 polyubiquitin chains

Previous studies have shown that NleL makes unanchored “free” K6 and K48 ubiquitin chains and branched chains (Hospenthal et al., 2013; Lin et al., 2011). More recently NleL was shown to ubiquitylate the newly described substrate JNK1 with K27,

K29 and K33 linked chains (Sheng et al., 2017). My results suggested that UBE2L3 was primarily modified on a single lysine (K82) with a polyubiquitin chain as the K82R mutation almost entirely abolished UBE2L3 modification by NleL. Ubiquitin moieties

241 can be connected via seven different lysine residues or its N-terminal methionine to form chains of varying linkage architecture. Recombinant ubiquitin proteins with a single lysine residue produce homotypic chains and can be used to determine the linkage specificities or preferences of E3 ligases. Recombinant K48-only ubiquitin, which exclusively forms K48-linked chains, was used in assays with NleL and various lysine-only or lysine-to-arginine variants of UBE2L3.

Akin to previous assays, K48-only ubiquitin could be used by all UBE2L3 variants, which showed no differences in their activity and induced similar levels of total polyubiquitin (Figure 6.10AB, bottom panel). Interestingly, the populations of total

K48-linked polyubiquitin from different lysine-only UBE2L3 variants produced distinct migration and banding pattern profiles (Figure 6.10A and B, bottom panel) when detected with an anti-ubiquitin antibody. By comparing lysine only variants it was clear that the total K48-linked polyubiquitin produced by wildtype and UBE2L3K82only were similar (Figure 6.10A, bottom, lanes 2 and 5), while the remaining lysine-only proteins had profiles similar to UBE2L318KR (Figure 6.10A, bottom panel, lanes 3, 4, 6, 7 and

8). The lysine-to-arginine mutants had similar variability in their K48-linked polyubiquitin profiles (Figure 6.10B, bottom panel). Here however, it was the total K48- linked polyubiquitin profiles of UBE2L318KR and UBE2L3K82R which were similar and distinct from the other (Figure 6.10B, bottom panel, lanes 2 and 5). However, these differences in K48-linked polyubiquitin were likely due to the detection of ubiquitylated-

UBE2L3, which contributes additional bands to the total ubiquitin detected on western blots. Indeed, when compared to anti-UBE2L3 western blots of the same samples

(Figure6.7AB, top panels), it can be postulated that the lack of bands corresponding

242

Figure 6.10: NleL modifies UBE2L3 with K48 ubiquitin chains UBE2L3 WT, UBE2L318KR and stated lysine-only (A) or lysine-to-arginine (B) variants of 170-782 UBE2L3 were incubated with NleL and K48-only ubiquitin for 20 mins following reaction initiation with ATP. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped by adding 2x Laemmli dye containing 100 µM DTT. UBE2L3 and total ubiquitin modification were detected by western blotting using anti-UBE2L3 (top panel) and anti-ubiquitin HRP antibodies (bottom panel). Western blots are representative of two independent experiments.

to ubiquitylated UBE2L3 (Figure 6.107A lanes 3, 4, 6, 7, 8 and Figure 6.10B, lanes 2 and 5) are responsible for the differences seen in the total ubiquitin blots (bottom panels). As with wildtype ubiquitin, UBE2L3K82only protein could be modified with K48- polyubiquitin chains (Figure 6.10A, lane 5) and mutation of UBE2L3K82R protein almost entirely lost its ability to be ubiquitylated (Figure 6.10B, lane 5). In addition, five distinct populations of modified UBE2L3 and high-molecular-weight streaks could be

243 observed, which suggested that multiple K48-modified species of UBE2L3 can be formed again ranging from mono to polyubiquitin (Figure 6.10A, lane 2).

Next, K6-polyubiquitin chain formation was examined, using a K6-only ubiquitin and the various lysine-only and lysine-to-arginine UBE2L3 variants (Figure 6.11). K6- only ubiquitin produced results like wildtype and K48-only ubiquitin. Only

UBE2L3K82only could get K6-polyubiquitylated to an extent similar to the wildtype

Figure 6.11: NleL modifies UBE2L3 with K6 ubiquitin chains UBE2L3 WT, UBE2L318KR and stated lysine-only (A) or lysine-to-arginine (B) variants of 170-782 UBE2L3 were incubated with NleL and K6-only ubiquitin for 20 mins following reaction initiation with ATP. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped by adding 2x Laemmli dye containing 100 μM DTT. UBE2L3 and total ubiquitin were detected by western blotting using anti-UBE2L3 (top panel), and anti- ubiquitin HRP antibodies (bottom panel). Western blots are representative of two independent experiments.

244

UBE2L3 (Figure 6.11A, lane 5). Mutation of K82 to an arginine almost entirely eradicated UBE2L3 modification, which again suggested that K82 is the predominant carrier of the K6 ubiquitin chain (Figure 6.11B, lane 5).

Interestingly, different UBE2L3 variants generated distinct K6-linked polyubiquitin banding pattern profiles and exhibited ranging proficiencies for generating K6-linked chains as determined by western blotting with anti-ubiquitin antibody (Figure 6.11, bottom panel). Proteins which were not ubiquitylated (18KR,

K64, K131, K138, K150) showed reduced levels of total unanchored polyubiquitin chains (Figure 6.11A). Likewise, the lysine-to-arginine mutants showed a similar trend: that is UBE2L318KR and to a lesser extent UBE2L3K82R produced reduced polyubiquitin as detected by western blot (Figure 6.11B).

6.2.6 NleL modifies multiple lysine residues with K29-linked ubiquitin

Recently NleL was shown to ubiquitylate the host kinase JNK1 with predominately K29 linked chains (Sheng et al., 2017). As JNK1 is the first substrate of

NleL to be reported it is possible that K29 polyubiquitin may be the preferred linkage used during substrate modification, despite NleL being unable to generate unanchored-K29 linked chains (Hospenthal et al., 2013). K29-only ubiquitin was therefore tested with UBE2L3, which acts as both a substrate and a ubiquitin donor to

UBE2L3 in vitro (Figure 6.12). When used with recombinant K29-only ubiquitin, wildtype UBE2L3 protein was ubiquitylated by NleL170-782 (Figure 6.12A, lanes 2 and

5). Similar to previous assays, ubiquitylation of wildtype UBE2L3 with K29-only ubiquitin produced multiple and distinct populations of modified UBE2L3 (Figure

6.12A, lane 2, top panel). However, K29-only ubiquitin could not be used to synthesise polyubiquitin chains on the UBE2L3K82only protein and instead formed two modified species which would relate to mono- and di-ubiquitin modifications (Figure 6.9A, lane

245

5). Unlike wildtype, K6-only or K48-only ubiquitin proteins, this suggested that NleL cannot attach long K29-ubiquitin chains to K82 of UBE2L3 and that additional lysine residues must be ubiquitylated in wildtype UBE2L3.

I then tested the K→R variants of UBE2L3. Intriguingly, results with K29- ubiquitin with UBE2L3K82R were different from those obtained with previously tested ubiquitin molecules. As expected the mutation of K82 to arginine did not abolish K29-

Figure 6.12: NleL modifies UBE2L3 on multiple with K29 ubiquitin UBE2L3 WT, UBE2L318KR and stated lysine-only (A) or lysine-to-arginine (B) variants of 170-782 UBE2L3 were incubated with NleL and K29-only ubiquitin for 20 mins following reaction initiation with ATP. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped by adding 2x Laemmli dye containing 100 μM DTT. UBE2L3 and total ubiquitin modification were detected by western blotting using anti-UBE2L3 (top), and anti-ubiquitin HRP antibodies (bottom). Western blots are representative of two (A) or one (B) experiments.

246 modification of UBE2L3, supporting the notion that K29-ubiquitin can be attached to other lysine residues, though there was one fewer band detected at approximately 30 kDa. (Figure 6.12B, lane 2). As with previous experiments UBE2L3K64R, UBE2L3K131R,

UBE2L3K138R, UBE2L3K150R proteins all behaved like wildtype UBE2L3 which indicated that these residues were modified with K29-ubiquitin.

All variants of UBE2L3 were functionally active with K29-only ubiquitin protein as the formation of free di-ubiquitin (~17 kDa) and NleL autoubiquitylation was observed in all cases (Figure 6.12, bottom). This suggested that NleL may be able to attach ubiquitin onto K82 in the form of a di-ubiquitin chain, which would explain the appearance of two modified species of UBE2L3.

In their recent article Sheng et al. (2017), showed that full-length NleL exhibited greater activity towards its substrate JNK1 when compared to the shorter NleL170-782 which had been used in most previous studies (Sheng et al., 2017). As JNK1 was shown to be modified with mostly K29 chains it was suggested that perhaps that

NleL170-782 may be less able to form K29 chains, which may explain the differences observed UBE2L3 K29 ubiquitylation.

To test this NleLFL and NleL170-782 were used side-by-side in ubiquitylation assays with wildtype, UBE2L3K82only, UBE2L3K82R and UBE2L318KR versions of

UBE2L3 and processed side-by-side to allow a direct comparison of their activity and

UBE2L3 modification (Figure 6.13). These assays confirmed that full-length NleL was better able to produce K29-ubiquitin chains compared with the shorter NleL170-782. This was evident by greater NleL autoubiquitylation and UBE2L3 modification, which showed higher molecular weight species (Figure 6.13, lanes 2 and 7). However, full- length NleL remained unable to produce longer polyubiquitin chains on UBE2L3K82only even though one additional band was visible at ~50 kDa which was absent in reactions

247 with NleL170-782. (Figure 6.13, lanes 3 and 8) These immunoblots showed that NleL can readily produce unanchored di-K29 ubiquitin but no longer chains (Figure 6.13, bottom panel). Therefore, I hypothesized that mono- or di-K29 ubiquitin could be ligated to UBE2L3 producing species that migrate at the observed sizes corresponding to mono- or di-ubiquitylated UBE2L3 (Figure 6.10, top panel). To explore this

Figure 6.13: Full-length NleL is better able to carry out K29-ubiquitylation of UBE2L3 as compared with NleL170-782 FL 170-782 Stated variants of UBE2L3 were incubated with either NleL or NleL with and K29- only ubiquitin for 20 mins following reaction initiation with ATP. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped by adding 2x Laemmli dye containing 100 μM DTT. UBE2L3 and total ubiquitin modification were detected by western blotting using anti-UBE2L3 (top) and anti-ubiquitin HRP antibodies (bottom). Western blots are representative of two independent experiments.

248 possibility, additional variants of UBE2L3 were tested to identify sites which may be modified with K29-ubiquitin.

UBE2L3 proteins containing only K20 (UBE2L3K20only) or K73 (UBE2L3K20only) were tested alongside K82 only, lysine-less proteins. A UBE2L3 variant lacking 9 lysine residues (UBE2L39KR) (see Appendix Table 1 for details) was also included. As

Figure 6.14: NleL modifies lysine 73 with K29 di- ubiquitin FL Stated variants of UBE2L3 were incubated with NleL and K29 only ubiquitin for 20 mins following reaction initiation with ATP. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped by adding 2x Laemmli dye containing 100 μM DTT. Total ubiquitin and UBE2L3 modification were detected by western blotting using anti- UBE2L3 antibody (top panel), and anti-ubiquitin HRP antibody (bottom panel). Western blots are representative of two experiments.

249 before, all versions could facilitate the formation of free di-K29 ubiquitin in assays with

NleLFL (Figure 6.14, bottom panel). However, only the wildtype UBE2L3 was modified with multiple K29-ubiquitin molecules (Figure 6.14, top panel, lane 2). Interestingly, like K82 the K73 of UBE2L3 was modified with two ubiquitin molecules (Figure 6.14, top, lanes 4 and 5) and UBE2L3K20only appeared to be weakly K29-ubiquitylated. By contrasted, UBE2L39KR was also modified multiple times, though to a reduced extent as compared to wildtype UBE2L3 (Figure 6.14, top, lane 6). Indeed, the reduction in ubiquitylation could suggest that some of the lysine residues which are mutated in

UBE2L39KR (Appendix Table 1) which includes lysine 73 and 82, may be recipients of

K29-ubiquitin. Therefore, the lysine residues which still may be modified include K16,

K71, K96, K134, K135 and K146 as these are the only remaining lysines in UBE2L39KR which have not been examined as individual lysine variants. However, K16, K135, and

K146 have never been reported to be modified in human or mouse cells. It is possible that K48, K71, K96 may receive K29-ubiquitin via NleL. In summary, UBE2L3 is modified by K29-ubiquitin and the precise single or multiple residues that receive this chain-type remain to be identified.

6.2.7 UBE2L3 is not modified with linear ubiquitin chains

The final ubiquitin linkage type tested was linear methionine-linked (M1) chains.

UBE2L3 is commonly associated with LUBAC and functions in the generation of linear ubiquitin in host cells (Lewis et al., 2015). The formation of homogeneous linear ubiquitin chains can be achieved by using lysine-less (No-K) ubiquitin during in vitro assays as these can only form chains via the free N-terminal methionine residue.

To this end, No-K ubiquitin was used with wildtype UBE2L3, UBE2L318KR and single lysine variants of UBE2L3. As with previous ubiquitin types only wildtype, and

UBE2L3K82only were modified with No-K ubiquitin (Figure 6.15, top panel, lanes 2 and

250

5). However, UBE2L3K82only was ubiquitylated with two ubiquitin molecules (Figure

6.15, lane 5). Note that a similar band was observed in reactions with K29 only ubiquitin and UBE2L3K82only or UBE2L3K73only (Figure 6.14 for k29, top panel, lane 4 and 5).

Immunoblotting for ubiquitin showed that NleL was not capable of generating linear chains as the no reactions showed higher molecular weight species of ubiquitin

Figure 6.15: Lysine 82 is not modified with linear polyubiquitin 170-782 WT, 18KR and stated lysine-only variants of UBE2L3 were incubated with NleL and No-K ubiquitin for 20 mins following reaction initiation with ATP. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped by adding Laemmli dye containing 100 μM. Total ubiquitin and UBE2L3 modification were detected by western blotting using anti-UBE2L3 antibody (top), and anti-ubiquitin HRP antibody (bottom). Western blots are representative of two experiments

251 such as di- ubiquitin (Figure 6.15, bottom panel). Instead, the only ubiquitin that could be detected corresponded to modified UBE2L3 or bands migrating at ~75 kDa which were likely to be autoubiquitylated NleL170-782 (NleL170-782 is ~70 kDa). It was therefore unclear whether UBE2L3 was being modified with di-ubiquitin or another non-lysine residue was being modified.

Figure 6.16: NleL ubiquitylates UBE2L3 on a non-lysine residue 170-782 Stated variants of UBE2L3 were incubated with NleL and His-NoK only ubiquitin for 20 mins following reaction initiation with ATP. WT UBE2L3 was incubated with and without ATP as a control. Reactions were stopped by adding Laemmli dye containing 100 μM. Total ubiquitin and UBE2L3 modification were detected by western blotting using anti- UBE2L3 antibody (top), and anti-ubiquitin HRP antibody (bottom). Western blots are representative of two independent experiments

252

To examine this, these experiments were repeated using His-NoK ubiquitin which can only be attached as monoubiquitin. Therefore, if UBE2L3K82only was modified with two molecules of ubiquitin it would indicate that NleL was indeed ubiquitylating

UBE2L3 on a non-lysine residue.

In assays with His-NoK and wildtype UBE2L3, five modified species were observed similar to that with No-K-ubiquitin (Figure 6.16, lane 1). As might be expected

UBE2L3K82R was only modified at four sites, as evident by the loss of one of the modified bands of UBE2L3 (Figure 6.16, lane 3). Interestingly, however, this did not cause the loss of the highest molecular weight species as might be expected; instead, a specific band at ~37 kDa was lost (Figure 6.16, lane 3). In addition, the second modified species which appears at ~30 kDa was altered and migrated as three species which migrate close to each other. Most interestingly, UBE2L3K82only formed two modified species (Figure 6.16, lane 2). This suggested that NleL was able to ubiquitylate UBE2L3 on a non-lysine residue (Figure 6.16, lane 2).

6.2.8 The N-terminus of UBE2L3 is modified by NleL

Ubiquitin can be ligated onto several different amino acid residues via specific chemical bonds: lysine via an isopeptide bond, N-terminal residues (i.e. methionine) via a peptide bond, cysteine via a thioester bond and serine, threonine, and tyrosine residues via an oxyester bond. The prospect of there being an additional isopeptide bond was disregarded as there are no other lysine residues in the UBE2L3K82only protein. Other possibilities included cysteine, by a thioester bond, or serine/threonine/tyrosine residues via an oxyester bond, both of which are labile in reducing and alkaline conditions respectively. It should be noted that thioester bonds are highly labile under reducing conditions, and all assays were stopped by the

253 addition of 2x Laemmli dye and DTT at a final concentration of 100 mM. However, it remained possible that the reducing conditions or DTT that was used were insufficient.

To test whether either of these two chemical bonds were responsible for the second modified species of UBE2L3, wildtype and UBE2L3K82only proteins were assayed with His-NoK ubiquitin and terminated in the presence or absence of DTT (to reduce thioester-linked ubiquitin) or additionally incubated in NaOH (to remove the labile oxyester bonds) (Figure 6.17).

In the absence of NleL, a band (~25 kDa) corresponding to the ubiquitin- charged form (i.e. thioester-linked ubiquitin at the C86 active site) UBE2L3 was observed with both variants (Figure 6.17, lanes 1 and 7). The addition of 100 mM DTT or 100 mM DTT along with 100 mM NaOH was sufficient to reduce the sample and

Figure 6.17: UBE2L3 is not modified with thioester or oxyester linked ubiquitin. 170-782 WT and K82 only UBE2L3 were incubated with or without NleL and His-NoK ubiquitin for 20 mins following reaction initiation with ATP. Reactions were stopped by adding 2x Laemmli dye and boiled for 3 mins. Samples were then left untreated, treated with DTT, or DTT and NaOH for 1 hour at 37°C. UBE2L3 modification was detected by western blotting using anti-UBE2L3 antibody. Experiments were repeat twice.

254 remove ubiquitin ((Figure 6.17, lanes 2, 3 and 8, 9) from the catalytic cysteine of

UBE2L3 (Lazarou et al., 2013; Wang et al., 2012). These were repeated in the presence of NleL which, as expected, generated five ubiquitylated species of wildtype

UBE2L3 and two species of UBE2L3K82only (Figure 6.17).

Interestingly, one-hour incubation with DTT or DTT plus NaOH did not alter the ubiquitylation of UBE2L3 as the number of modified bands remained the same regardless of treatments (Figure 6.17 lanes, 4 to 6 and 10 to 12). This suggests that the additional ubiquitin modification on K82 only is not attached via thioester or oxyester bonds on a non-lysine residue.

The final possibility was that the N-terminus of UBE2L3 was modified by NleL via a peptide bond. To test this, untagged and GST-tagged wildtype and UBE2L3K82only proteins (Figure 6.18A) were used in ubiquitylation assays with His-NoK ubiquitin. It was hypothesised that the large GST-tag (25 kDa) would preclude the N-terminus of

UBE2L3 from being available for ubiquitylation. Ubiquitylated GST-UBE2L3 was then incubated with rhinovirus C3 protease (Figure 6.18A) to remove the GST tag and assess the ubiquitylation status of UBE2L3 by western blotting. The cleavage of GST- tag was required because GST itself was modified by NleL and interfered with the immunoblot-based assay. GSH-resin was used to remove cleaved GST from reactions prior to SDS-PAGE and immunoblotting.

As before, 20 min reactions were performed and stopped by the addition of

EDTA. Reactions were split, and half was incubated with rhinovirus C3 protease for overnight at 4°C and the resulting protein mixtures were analysed by western blotting for UBE2L3. Western blot analyses of the untreated samples showed that both tagged and untagged proteins were modified (Figure 6.18B, bottom). Interestingly, GST-

UBE2L3K82only showed multiple modified species suggesting that NleL can also modify

255

GST (Figure 6.18, bottom, lane 3). Untagged proteins behaved similarly as before.

Examination of the cleaved samples, however, showed that post-GST-cleavage,

UBE2L3K82only showed a single modified species as compared with the untagged protein which had two modified species (Figure 6.18, top, lanes 3 and 6).

Figure 6.18: NleL modifies the N-terminus of UBE2L3 with ubiquitin which can be blocked by an N-terminal GST-tag. (A) Coomassie gel of recombinant GST-tagged UBE2L3 variants (left) and rhinovirus C3 protease (right). (B) Untagged or N-terminal GST-tagged variants of WT and K82 only UBE2L3 protein were incubated with NleL170-782 and His-NoK ubiquitin for 20 mins. Reactions were stop with the addition of 20 mM EDTA, split into fresh tubes either left untreated (top panel) or incubated overnight with rhinovirus C3 protease (bottom). UBE2L3 modification was detected by western blotting using anti-ubiquitin HRP antibody. *Loss of N-terminal modification. Data are representative of two experiments.

256

GST- cleaved wildtype protein showed no obvious loss of a modified band, however, because residual uncleaved GST-UBE2L3 was present in the sample it was difficult to accurately determine. These data strongly suggested that UBE2L3 is ubiquitylated at its N-terminus by NleL and that the addition of large N-terminal tags such as GST is capable of blocking this.

It is of note that the recombinant UBE2L3 generated and used for these studies possesses a short residual linker sequence (figure) at the N-terminus which is left following the removal GST. Recombinant UBE2L3, therefore, starts with a glycine as opposed to a methionine residue, these experiments should be repeated with protein that has its native N-terminus to ensure that ubiquitylation still occurs.

6.2.9 Lysine availability affects free chain formation

Because of the robust activity of NleL most of the available UBE2L3 was modified with polyubiquitin after 20 mins. An unfortunate consequence of this was that it was difficult to accurately determine how much free unanchored-polyubiquitin was being produced as anchored and unanchored chains were indiscernible from one another when reactions were immunoblotted for total ubiquitin. Results from previous assays which appeared to show that the presence of certain UBE2L3 lysine residues could alter NleL activity were therefore difficult to interpret as it was impossible to ascertain whether, or to what extent UBE2L3-anchored polyubiquitin was contributing to the total polyubiquitin detected by western blot.

I wanted to assess the extent of unanchored chain formation by NleL using various ubiquitin’s (e.g. K6, K48, and K29) without interference from the UBE2L3 signal. To do this wildtype GST-UBE2L3, GST-UBE2L3K82only and GST-UBE2L3K82R variants were used, for in vitro ubiquitylation assays with ubiquitin variants (Figure

6.19B). Following the 20 min reaction, the assays were terminated by the addition of

257

20 mM EDTA and GST-UBE2L3 was removed by incubating with glutathione sepharose resin. The resin-bound and supernatant fractions were separately analysed by western blot and compared to ubiquitylation reactions containing untagged wildtype

UBE2L3.

Following the removal of GST-UBE2L3 western blot analysis showed that different UBE2L3 variants produced different amounts of unanchored polyubiquitin. In reactions using wildtype ubiquitin GST-UBE2L3K82R synthesised noticeable higher amounts of free-polyubiquitin compared with wildtype and GST-UBE2L3K82only (Figure

6.19A). Similar results were obtained with K48 only ubiquitin which (Figure 6.19B). By contrast in vitro assays using K6 only ubiquitin produced different and very distinct banding patterns (Figure 6.19C). Contrary to mixtures containing both free and

UBE2L3-anchored polyubiquitin (Figure 6.19C, lane 5) samples without GST-UBE2L3 had fewer species of ubiquitin chains (Figure 6.19C, lanes 2 to 4). This suggested that most (if not all) of the K6 polyubiquitin produced by NleL is fact anchored to UBE2L3.

By contrasted NleL could only produce short, unanchored K6-linked chains (Figure

6.19C, lanes 2 to 4). Interestingly the availability ofUBE2L3 lysine also affected free

K6-chain formation as GST-UBE2L3K82R produced more K6-tri-ubiquitin (Figure 6.19C, lane 3) These results help explain previous observations that UBE2L318KR and

UBE2L3K82R appeared to synthesise less K6-linked polyubiquitin (Figure 6.11). In fact, the opposing scenario seems to be true, as the common trend with all three ubiquitin types was that UBE2L3K82R produced more free-unanchored polyubiquitin (Figure

6.19ABC, lane 3). These data seem to suggest that the availability of UBE2L3 lysine can alter the formation of anchored free polyubiquitin.

258

Figure 6.19: Availability of lysine reactive UBE2L3 affects the formation of free Ub chains by NleL. 170- GST-WT, GST-K82 only, GST-K82R or WT UBE2L3 proteins were incubated with NleL 782 and either wildtype (A), K48-only (B), or K6-only (C) ubiquitin for 20 mins. Reactions were stop with the addition of 20 mM EDTA and incubated overnight with glutathione agarose to remove modified UBE2L3. Total free ubiquitin and was detected by western blotting using anti-ubiquitin HRP antibody. The presented data are representative of two independent experiments.

259

6.3 Discussion Covalent modification of E2 conjugating enzymes is not an uncommon phenomenon, there are in fact multiple examples of E2 enzymes being regulated by ubiquitylation (Stewart et al., 2016). However, in most cases, such regulatory mechanisms are autoregulated by the E2 enzyme (Stewart et al., 2016).

The data presented here showed that NleL is capable of ubiquitylating UBE2L3 in trans. In other words, the modification of UBE2L3 can occur independently of its native, and functional interaction with NleL. Experiments with UBE2L3P62A, F63 suggested that NleL can modify UBE2L3 via interactions that are distinct from those required for catalytic E2-E3 interaction. Therefore, NleL and UBE2L3 may interact at additional surfaces which remain to be identified. The trans ubiquitylation of UBE2L3 by NleL was further supported by the observation that the catalytically inactive

UBE2L3C86A protein was also modified in the presence of lysine-less UBE2L318KR. It should be noted that this data does not conclusively show that UBE2L3 modification cannot also occur in cis. To address this, further studies should be carried out with

UBE2L3 variants which are dually catalytically inactive and unable to bind NleL via its typical E2 binding surface.

The Salmonella E3 ligase effector SspH2 has been shown to interact with the

E2 conjugating enzyme UBE2D1 via its alanine 96 residue which constitutes a non- typical binding surface (Levin et al., 2010). UBE2L3 has a corresponding alanine at position 98 which perhaps could be involved in secondary interactions with NleL.

However, it should be noted that NleL and SspH2 share little sequence or structural similarity and belong to different families of convergently evolved bacterial E3 ligases.

Thorough mutational analyses, or crystallographic analyses with full-length NleL, might reveal additional sites of NleL-UBE2L3 interactions.

260

The targeted ubiquitylation of UBE2L3 by NleL could have the potential to drastically alter its cellular activity. UBE2L3 has been shown to be ubiquitylated on multiple lysine residues (Kim et al., 2011b; Mertins et al., 2013; Wagner et al., 2011), and my own data, presented in Chapter 4, hints to a potential role in ubiquitin and proteasomal dependent regulation of UBE2L3 stability.

Previous reports have suggested that this kind of proteasomal regulation often results from the disruption of E2-protein interactions (Stewart et al., 2016). For example, when the yeast E2 enzyme Ubc7 is unable to interact with its binding partner

Cue1, polyubiquitin chains form on its catalytic cysteine which targets it for proteasomal degradation (Ravid and Hochstrasser, 2007). A similar mechanism also exists in eukaryotic cells, for the enzymes UBE2D3 and UBE2N which are both degraded by the proteasome when their interaction with E3 ligases is disrupted

(Shembade et al., 2010). In vitro assays with K48-ubiquitin showed that NleL can form

K48-linked polyubiquitin chains on UBE2L3 (Figure 6.10) which is a degradative signal. NleL could, therefore, regulate UBE2L3 by targeting it to the proteasome, however experiments by Sheng at al., saw no effect on UBE2L3 protein expression when NleL was ectopically expressed HEK293T cells (Sheng et al., 2017).

Alternatively, ubiquitylation by NleL could alter the activity of UBE2L3, particularly as the predominant site of ubiquitylation (Lys82) is proximal to the catalytic cysteine (Cys86) of UBE2L3. It is feasible that the presence of a polyubiquitin chain near the catalytic cleft could disrupt transthiolation. UBE2E1 has been shown to undergo autoregulatory ubiquitylation in its extended N-terminal region which enables it to switch its activity from mono- to polyubiquitylation in vitro (Banka et al., 2015;

Schumacher et al., 2013). It was also shown to autoubiquitylate itself near its catalytic

261 cysteine which was reported to reduce its activity (Banka et al., 2015). Similarly, the

E2 enzyme UBE2T has also been described to undergo autoubiquitylation on Lys91 close to its catalytic Cys86 residue (Machida et al., 2006). Ubiquitylation of this site was reported to significantly reduce the catalytic activity of UBE2T (Machida et al., 2006).

Autoubiquitylation of UBE2E1 and UBE2T requires their respective N- and C-terminal extension domains and is enhanced by interactions with their cognate RING E3 ligases (Banka et al., 2015; Machida et al., 2006). These types of self-regulating modifications are believed to limit non-specific lysine ubiquitylation by E2 conjugating enzymes. However, UBE2L3 lacks similar terminal end domains and cannot mediate the transfer of ubiquitin to lysine residues (Wenzel et al., 2011). Additionally, UBE2L3 does not possess an equivalent lysine residue at position 91, therefore Lys82 could serve a similar regulatory purpose and be exploited by NleL to negatively impact

UBE2L3 activity. Such targeted regulation has been shown with the E2 enzyme

UBE2N which is modified on the conserved Lys92 residue with the ubiquitin-like protein

ISG15 (Takeuchi and Yokosawa, 2005; Zou et al., 2005). ISGylation of UBE2N appeared to block thioester formation on its own catalytic cysteine but did not affect its interaction with its binding partner MMS2 (Zou et al., 2005). Though this modification was not explicitly shown to occur in trans, it did require dedicated

ISGylation enzymes (E1-UBE1L and E2-UbcM8) to be ectopically expressed for it to occur suggesting that it UBE2N was targeted as a substrate (Zou et al., 2005).

In the study presented above no obvious activity defect was observed between variants of UBE2L3 which could not be ubiquitylated by NleL. UBE2L3K82R was seen to produce more wildtype and K48 anchored polyubiquitin chains (Figure 6.19) however this could be due to an inability to ligate ubiquitin onto UBE2L3. Additional studies with purified ubiquitylated UBE2L3 would need to be performed to better

262 understand the impact that the modification may have on its E2 activity with both host and bacterial E3 ligases.

Interestingly, the ubiquitylation of Lys82 could potentially impact UBE2L3 interactions with eukaryotic HECT ligase E6-AP. E6-AP has been shown to bind E2 enzymes at two separate sites in order to facilitate ubiquitin chain formation (Ronchi et al., 2017; Ronchi et al., 2013). A recent study carried out in silico modelling to characterise these interactions and predicted that Lys82 would interaction with Glu539 of E6-AP via a hydrogen bond (Ronchi et al., 2017). Early studies have shown that mutation of Glu539 reduced E6-AP thioester formation by 90% (Huang et al., 1999).

Whether the reduction in activity is due to a disrupted interaction with UBE2L3 has not been shown, however, this does highlight that ubiquitylation of UBE2L3 could have the potential to disrupt interactions with host E3 ligases.

Using linkage-specific ubiquitin I showed that NleL can generate K6- and K48- polyubiquitin chains on UBE2L3 (Figures 6.10 and 6.11). However, the precise linkage architecture of wildtype polyubiquitin anchored onto UBE2L3 remains unknown. K6- and K48-linked polyubiquitylation predominantly occurred on Lys82, and mutation of this residue to arginine almost entirely abolished self-ubiquitylation. However, NleL could also utilise K29-only ubiquitin to modify UBE2L3 (Figure 6.12) which is consistent with its known substrate JNK1(Sheng et al., 2017). Importantly, long K29- linked polyubiquitin chains were not formed on the K82 residue (Figure 6.12). This suggested that NleL was capable of selectively modifying different UBE2L3 lysine residues with specific linkage types of ubiquitin. In fact, UBE2L3K82only was only modified with two ubiquitin molecules when assayed. However, wildtype UBE2L3, according to its molecular weight, was modified with at least six ubiquitin molecules.

This together with the observation of di- ubiquitin formation and the similar

263 ubiquitylation of Lys73, I propose that UBE2L3 could be multimono- or di-ubiquitylated with K29-ubiquitin chains. However, these results may also be due to N-terminal modification as seen with No-K- and His-NoK-ubiquitin. Therefore, further experimentation is required to conclusively answer which residue or residues are modified with K29-ubiquitin by NleL.

Additional observations showed that other residues could be modified as wildtype UBE2L3 was multimono-ubiquitylated with His-NoK ubiquitin. Such assays would create five distinct populations of modified UBE2L3 with increasing numbers of ubiquitin attached to different lysine residues. The identity and significance of these residues remain to be determined. However, as UBE2L3K82R is poorly ubiquitylated, I concluded that when NleL can extend ubiquitin chains on Lys82, it does so more readily than multimono-ubiquitylate UBE2L3 on less preferred sites.

Data also suggested that NleL ubiquitylated the N-terminus of UBE2L3 (Figure

6.18). To date only the E3 ligase HUWE1 (Scaglione et al., 2013) and E2 enzyme

UBE2W (Scaglione et al., 2013) have been shown to facilitate the N-terminal ubiquitylation of substrate proteins (Varland et al., 2015). NleL is, therefore, the first bacterial E3 ligases to be shown to be shown to do so. It should be noted that the recombinant UBE2L3 proteins used in this study process a short linker peptide which remains at the N-terminus of the protein following removal of the GST-tag (Appendix

Figure 1). Though the linker does not contain lysine residues (which may be ubiquitylated via an isopeptide bond), these experiments should be repeated with protein that has the native N-terminal sequence, including the native methionine residue, to ensure this phenomenon is not an artifact of the residual linker. N-terminal ubiquitylation is commonly associated with protein destabilisation and degradation

(Bloom et al., 2003; Coulombe et al., 2004; Noy et al., 2012). As with the potential for

264 the addition of K48-polyubiquitin chains to Lys82 UBE2L3 stability should be assessed during infection with EHEC.

Initial attempts to examine the role of NleL inside cells were unsuccessful.

Ectopically expressed NleL was unstable and difficult to detect consistently. Similarly, attempts to create a HeLa cell line which stably expressed NleL were unsuccessful.

The recent article from Sheng et al. showed that JNK1 ubiquitylation by NleL was largely dependent on UBE2L3. The conclusion of the study was that JNK1 ubiquitylation blocked its activation, which in turn was important for bacterial attachment and pedestal formation (Sheng et al., 2017). They also showed that chemical inhibition or silencing of JNK1 compensate for the deletion of nleL and permitted ΔnleL EHEC to infected to the same extent as wildtype bacteria (Sheng et al., 2017). Though they did not explicitly show that UBE2L3 silencing affected attachment, it could be inferred from these results that silencing UBE2L3 would phenocopy nleL deletion and reduce cell attachment of wildtype EHEC. I was able to replicate the same infection defect using an independently generate ΔnleL mutant strain (data not shown). However, despite this UBE2L3 was not required for NleL- mediated attachment as would have been expected (data not shown). It should be noted that the ΔnleL strain used by Sheng et al. was made in a different strain background (E. coli O157:H7 Sakai). It is, therefore, possible that differences between the Sakai and EDL933 strains may have contributed to these apparent discrepancies.

The reasoning behind this discrepancy requires further investigation. UBE2L3 ubiquitylation by NleL remains to be shown in cells during infection. However, if this process does occur NleL would have the potential to drastically alter the activity of

UBE2L3 in cells. Like the Shigella effector OspG, NleL could in turn affect many

265 aspects of cell signalling such as LUBAC signalling or PARKIN mediated mitophagy

(de Jong et al., 2016; Grishin et al., 2014; Kim et al., 2005)

6.4 Summary The results presented in this chapter confirm that NleL ubiquitylates UBE2L3 on lysine residues in vitro. Ubiquitylation of UBE2L3 occurred in trans and did not require the typical catalytically pertinent E2-E3 interaction residues on the acceptor molecule. Lysine 82 is the favored residue to receive K6- and K48-polyubiquitin chains and mutation of this residue almost entirely abolished UBE2L3 modification. However, results with non-chain forming His-NoK ubiquitin showed that UBE2L3 can also be modified at other less preferred residues in vitro. Similarly, UBE2L3 may also be multimono- and di-ubiquitylated with K29-ubiquitin. Data also showed that NleL was able to ligate one ubiquitin molecule to the N-terminus of UBE2L3, which makes it the first bacterial ligase identified to do so. A precise role of NleL ubiquitylation of UBE2L3 during infection should be determined in future experiments.

266

Chapter 7 – Final summary and future perspectives

267

7.1 Introduction Inflammasome and caspase-1 mediated immune responses have important roles in antimicrobial immunity. However, despite an ever-increasing body of literature describing the mechanisms leading to the activation of caspase-1, downstream effector mechanisms remain poorly characterised. Such processes are likely controlled by additional, yet uncharacterised caspase-1 substrates and indirect targets. Using shotgun proteomics, the Shenoy lab identified a number novel candidate targets of caspase-1. One such protein identified is an E2 Ubiquitin- conjugating enzyme called UBE2L3. The principal aim of this thesis was to investigate the role of UBE2L3 in inflammasome regulation and responses. To explore this four research objectives were set out: (1) Characterise UBE2L3 targeting in the wider context of inflammasome signalling by various agonists and bacterial infectious agents; (2) Determine the mechanisms of UBE2L3 targeting by inflammasomes; (3)

Elucidate the role of UBE2L3 in inflammasome and caspase-1 mediated responses;

(4) Better characterise UBE2L3 interactions with bacterial effectors in the context of its role in immune-signalling.

7.2 Final summary UBE2L3 was identified as an indirect target of caspase-1 which is depleted from macrophage and dendritic cells in response to inflammasome activation. Like cytokine secretion and pyroptosis, the loss of UBE2L3 was common to caspase-1 activation by

NLRP3, NLRC4, AIM2, NLRP1b, PYRIN and non-canonical inflammasomes following exposure to bacteria or sterile danger signals. Mechanistically, UBE2L3 depletion required cell-autonomous activity of caspase-1 and proteasomes and occurred independently of cell death and cytokine signalling. The underlying molecular pathways linking caspase-1 activity to UBE2L3-proteasomal degradation are yet to be

268 determined but was independent of lysine ubiquitylation. Several potential mechanisms of non-canonical proteasome degradation could be responsible (Figure

4.15) and will require further study.

UBE2L3 was found to aid in the K48-linked polyubiquitylation of pro-IL-1β and enhanced its degradation by the proteasome. This agreed with previous reports that have demonstrated pro-IL-1β ubiquitylation and degradation. Though the identity of the E3 ligase is still unknown, UBE2L3 is the first protein identified to function in this pathway. This process also negatively impacted the secretion of mature IL-1β. In the context of inflammasome responses UBE2L3 protein levels inversely correlated with

IL-1β stability and secretion. UBE2L3 must, therefore, be depleted by caspase-1 for efficient secretion of IL-1β to proceed. This could significantly impact IL-1β in response to weak activators of inflammasomes such as commensal bacteria.

UBE2L3 can also be targeted by multiple bacterial effectors including the EHEC effector NleL. The new role for UBE2L3 in regulating pro-IL-1β, coupled with previously reported roles in mediating linear polyubiquitylation, suggest that bacteria could impact many cellular processes which could vary depending on what cell type is being infected. Here UBE2L3 was confirmed to be ubiquitylated by NleL in agreement with previous reports. Several novel observations were also made. Using different mutants, I was able to show that Nlel can ubiquitylate UBE2L3 in trans suggesting that

UBE2L3 may, in fact, be specifically targeted. Lys82 was also identified as the predominant carrier of polyubiquitin chains though other residues could also be weakly modified.

Additionally, NleL could also attach a single ubiquitin to the N-terminus of

UBE2L3, NleL is the first bacterial ligase reported to mediated N-terminal ubiquitylation. The effects NleL may have on UBE2L3 during infection are yet to be

269 determined. However, UBE2L3 is not required for NleL-mediated bacterial attachment to HeLa cells. NleL may, therefore, have other roles during infection.

These studies identify new functions for UBE2L3 in regulating inflammatory responses in macrophages which were independent of pro-inflammatory transcription.

Several key questions remain pertaining to the mechanism by which UBE2L3 and pro-

IL-1β are degraded at the proteasome. These results also highlight a number of interesting research avenues which could be pursued.

7.3 Future perspectives 7.3.1 Elucidate how UBE2L3 is targeted to the proteasome

As discussed in Chapter 4 there are multiple potential mechanisms which could mediate the proteasomal degradation of UBE2L3. Experiments examining UBE2L3 stability under resting conditions highlighted that lysine-less UBE2L3 was unstable and degraded by the proteasome. This suggested that lysine residues, and perhaps lysine ubiquitylation, have a stabilising effect on UBE2L3. Importantly these lysine residues were not required for caspase-1-dependent depletion of UBE2L3 (Figure 4.14). As outlined in Figure 4.15 this suggests that UBE2L3 degradation is mediated by non- canonical ubiquitylation or ubiquitin-independent degradation signals. These include

N-terminal ubiquitylation, degron signalling or chaperone-mediated shuttling to the proteasome.

Over the course of these studies, several lines of evidence have emerged which suggested that N-terminal ubiquitylation of UBE2L3 could be responsible for mediating its degradation. Firstly, lysine-less UBE2L3 was still depleted by caspase-1 and was intrinsically turned over by the proteasome. The proteasomal degradation of lysine- less proteins is often used to infer the occurrence of N-terminal ubiquitin-mediated degradation (Kravtsova-Ivantsiv and Ciechanover, 2012). Secondly, the addition of a

270 small 3x Flag-tag to the N-terminus did not block UBE2L3 depletion. As discussed in

Chapter 4 the size of this tag may not be sufficient to block N-terminal ubiquitylation.

By contrast, the attachment of a larger YFP N-terminal tag appeared to prevent

UBE2L3 depletion following the activation of caspase-1 (Breitschopf et al., 1998). This suggests that a free N-terminus is required for caspase-1 proteasomal-mediated depletion. Thirdly, UBE2L3 was found to be N-terminally ubiquitylated by NleL and therefore may be also be modified in this fashion by a host E3 ligase. Interestingly,

NleL appears to modify UBE2L3 with an N-terminal monoubiquitin which has been previously reported to be sufficient to mediate proteasomal degradation (Shabek et al., 2009). Previous studies have shown that tags above a certain size can block N- terminal ubiquitylation (Breitschopf et al., 1998). By repeating experiments similar to those outlined in in Chapter 4, the effect of a larger N-terminal tag on the stability of mutant UBE2L3 variants can be assessed further. For instance, would the presence of a large N-terminal tag to lysine-less UBE2L3 stabilise it? And would this also block caspase-1 mediated depletion? The presence of an N-terminal modification could then be directly confirmed by western blot or LC-MS based analyses of lysine-less UBE2L3 extracted from cells.

Alternatively, UBE2L3 degradation may be mediated by an intrinsic degron sequence. Because these signals are within the protein itself previous studies aimed at identifying internal degron sequences have done so by tagging a stable protein

(such as GFP) with different segments of the hypothesised degron containing protein.

This approach was taken to identify an N-terminal degron in IκBα (Fortmann et al.,

2015). Likewise, portions of UBE2L3 could be attached to GFP protein to see it affects its stability.

271

Finally, UBE2L3 could be degraded by the 20S proteasome-mediated by a chaperone such as the REGγ complex which forms an alternative lid for 20S proteasomes (Chen et al., 2007). This can be tested by assessing whether UBE2L3 can be degraded in vitro by the 20S proteasome and REGγ which do not require target proteins to be ubiquitylated for degradation (Chen et al., 2007). Following the identification of the mechanism which mediates the loss UBE2L3 stability, the following question would be how does caspase-1 regulate this? As discussed previously this too could be done in multiple ways which is dependent on the how UBE2L3 is targeted to the proteasome. In brief, caspase-1 could either actively stimulate the addition of a degradative signal, for example activating an E3 ligase or chaperone. Alternatively, caspase-1 could inhibit a stabiliser of UBE2L3 such as an E3 interaction which allows for a native degradation signal to take effect. These would have to be examined in the context of a specific degradation signal.

7.3.2 Further characterisation of pro-IL-1β ubiquitylation

Pro-IL-1β has been previously described to be ubiquitylated and turned over by the proteasome. During these studies, UBE2L3 was found to function in the K48 polyubiquitylation of pro-IL-1β. The E3 ligase which specifically mediates this process is yet to be identified. In most cases studies aim to identify substrates of an E3 rather than the E3 of a specific substrate which an E2 uses. In this respect current methods which are used to examine ligase specific ubiquitylation may not be appropriate to identify the E3 ligases which UBE2L3 functions with. Fortuitously, because UBE2L3 lacks lysine reactivity the possible candidates are limited to members of the HECT and

RBR families which collectively make ~45 potential ligases. In the lab, initial attempts were carried out using a siRNA-based screening process to examine whether the silencing of individual E3 ligases could block the enhanced pro-IL-1β turnover

272 observed in YFP-UBE2L3 expressing cells. However, because many of these ligases

(particularly the components of LUBAC) are required for transcriptional induction of pro-IL-1β it was difficult to accurately interpret variation in pro-IL-1β levels. This approach could still be utilised but would require a TLR-induction free system. This could be achieved by using an artificially inducible pro-IL-1β expression system. This would allow all potential candidate E3 ligases to be assessed purely on their ability to mediate pro-IL-1β turnover. This system would also be able to inform on whether the degradation of pro-IL-1β itself is inducible, for example by LPS.

Alternatively, this could be tested in a cell-free system. Over recent years multiple in vitro tools for studying ubiquitylation have been developed. These include activity-based probes (ABPs) which can be used to study the enzymatic activity of ubiquitin-enzymes (Witting et al., 2017). One recently developed system enables the identification of specific ubiquitylating enzymes which function together during the ubiquitin transfer cascade (Mulder et al., 2016). This ABP uses an electrophilic dehydroalanine (Dha) moiety which is attached to the C-terminus of the ubiquitin protein. The addition of the Dha enables ubiquitin to undergo two different chemical reactions when it is transferred between the catalytic cysteine residues of ubiquitin enzymes. In one reaction the Dha can form the native thioester bond allowing it to be transferred to the next member of the cascade. In the other, it forms a stable thioether bond and becomes covalently trapped to the catalytic cysteine (Mulder et al., 2016).

The Dha probe and any stably modify enzymes can be purified and examined.

Using this probe UBE2L3 can be artificially “charged” in vitro and be applied to cell lysates from LPS treated cells. Any E3 ligase which utilised the Dha charged

UBE2L3 may then form stable thioether bonds with Ub Dha which can be used to purify and identify the specific E3 ligases UBE2L3 uses. Any potential hits could then

273 be examined in more detail on an individual basis. The identified E3 ligase should then be examined using methods similar to those used in Chapters 3, 4 and 5 to determine whether it is also regulated by the inflammasomes and whether its continued expression or silencing phenocopies the observed effects of UBE2L3 (Mulder et al.,

2016).

Furthermore, the polyubiquitylation pro-IL-1β may be examined further. As discussed in Chapter 5, pro-IL-1β has been shown to be modified with K63-linked polyubiquitin on Lys133, which enhances its processing by caspase-1 (Duong et al.,

2015). Mutational analysis can be used to examine whether the same lysine is responsible for carrying UBE2L3-generated K48-linked chains. In doing so it can be determined whether pro-IL-1β is competitively ubiquitylated with different types of ubiquitin chain. This can help explain whether the increase or decrease in mature IL-

1β secretion which has been observed in response to UBE2L3 overexpression or silencing respectively, is due to disruptive ubiquitylation or simply because of the differences in pro-IL-1β stability.

7.3.3 Physiological relevance and wider in vivo implications

These studies have highlighted a novel cell-specific role of UBE2L3. Previous studies examining UBE2L3 have mostly been done in epithelial or fibroblast cells where it has a key role in stimulating pro-inflammatory gene expression in partnership with LUBAC (Fu et al., 2014; Lewis et al., 2015). By contrast in macrophages, UBE2L3 has an anti-inflammatory role through its mediation of pro-IL-1β degradation. In this sense, the cell-type-specific roles of UBE2L3 seem to mirror that of LUBAC, as in macrophages LUBAC does not affect transcriptional induction but is important for inflammasome activation (Rodgers et al., 2014).

274

However, in these studies, UBE3L3 overexpression or silencing did not affect inflammasome activation or caspase-1-induced cell death suggesting that the functional dependence that LUBAC has for UBE2L3 in epithelial cells (Lewis et al.,

2015) may not hold true in macrophages. The wider implications of UBE2L3 in different cell types or even in living tissue where multiple cell types co-habit would be interesting to explore.

Of particular interest, would be determining the role of UBE2L3 in inflammasome activated cells which do not undergo pyroptosis. Despite UBE2L3 depletion having a role in promoting IL-1β release in macrophages, these responses would ultimately be limited by cell death. It would be interesting to examine whether

UBE2L3 is similarity regulated during activation of the alternative inflammasome in monocytes or in neutrophils during STm infection (Gaidt et al., 2016) (Chen et al.,

2014) or in macrophages in a “hyperactivation” state (Evavold et al., 2017) as the loss or presence of UBE2L3 could have a longer lasting impact. This could be examined by testing these cells for UBE2L3 depletion in response to their specific non-pyroptotic inflammasome-activating treatments.

The role of UBE2L3 modification by bacterial E3 effectors remains to be examined during infection. The interaction and functional consequences of UBE2L3 binding by OspG are probably the best understood and leads to UBE2L3 inhibition and blocking of NF-κB signalling (de Jong et al., 2016; Grishin et al., 2014). NleL- mediated ubiquitylation of UBE2L3 is yet to be observed in host cells during infection.

Because UBE2L3 has been reported to be ubiquitylated in the host it may be difficult to accurately observe NleL-dependent alterations to UBE2L3. However, in these studies Lys82 was identified as the predominant site of ubiquitylation by NleL. Using an K82 only mutant variant of UBE2L3 may help reduce background ubiquitylation by

275 host E3 ligases. Additionally, by using cells which only express UBE2L3K82only or

UBE2L3K82R. We may better investigate the consequences of UBE2L3 ubiquitylation by NleL and whether its modification has inhibitory or activatory roles.

Unfortunately examining the role of UBE2L3 in an in vivo setting is not currently possible as deletion of UBE2L3 is embryonic lethal (Harbers et al., 1996). It is currently unknown what the basis of this lethality is, however using a conditional and inducible knockout system like the Cre-loxP system (Sharma and Zhu, 2014) could allow for the role of UBE2L3 in immune cells to be examined in vivo.

276

References Aachoui, Y., Leaf, I.A., Hagar, J.A., Fontana, M.F., Campos, C.G., Zak, D.E., Tan, M.H., Cotter, P.A., Vance, R.E., Aderem, A., et al. (2013). Caspase-11 Protects Against Bacteria That Escape the Vacuole. Science 339, 975-978.

Abbas, T., and Dutta, A. (2009). p21 in cancer: intricate networks and multiple activities. Nature reviews. Cancer 9, 400-414.

Abderrazak, A., Syrovets, T., Couchie, D., El Hadri, K., Friguet, B., Simmet, T., and Rouis, M. (2015). NLRP3 inflammasome: From a danger signal sensor to a regulatory node of oxidative stress and inflammatory diseases. Redox Biol 4, 296- 307.

Abedat, S., Urieli-Shoval, S., Shapira, E., Calko, S., Ben-Chetrit, E., and Matzner, Y. (2002). Effect of colchicine and cytokines on MEFV expression and C5a inhibitor activity in human primary fibroblast cultures. Isr Med Assoc J 4, 7-12.

Afonina, I.S., Muller, C., Martin, S.J., and Beyaert, R. (2015). Proteolytic Processing of Interleukin-1 Family Cytokines: Variations on a Common Theme. Immunity 42, 991-1004.

Agard, N.J., Maltby, D., and Wells, J.A. (2010). Inflammatory Stimuli Regulate Caspase Substrate Profiles. Mol Cell Proteomics 9, 880-893.

Aglietti, R.A., Estevez, A., Gupta, A., Ramirez, M.G., Liu, P.S., Kayagaki, N., Ciferri, C., Dixit, V.M., and Dueber, E.C. (2016). GsdmD p30 elicited by caspase-11 during pyroptosis forms pores in membranes. Proceedings of the National Academy of Sciences of the United States of America 113, 7858-7863.

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.

Ainscough, J.S., Frank Gerberick, G., Zahedi-Nejad, M., Lopez-Castejon, G., Brough, D., Kimber, I., and Dearman, R.J. (2014). Dendritic cell IL-1alpha and IL-

277

1beta are polyubiquitinated and degraded by the proteasome. The Journal of biological chemistry 289, 35582-35592.

Akhter, A., Caution, K., Abu Khweek, A., Tazi, M., Abdulrahman, B.A., Abdelaziz, D.H.A., Voss, O.H., Doseff, A.I., Hassan, H., Azad, A.K., et al. (2012). Caspase-11 Promotes the Fusion of Phagosomes Harboring Pathogenic Bacteria with Lysosomes by Modulating Actin Polymerization. Immunity 37, 35-47.

Akhter, A., Gavrilin, M.A., Frantz, L., Washington, S., Ditty, C., Limoli, D., Day, C., Sarkar, A., Newland, C., Butchar, J., et al. (2009). Caspase-7 activation by the Nlrc4/Ipaf inflammasome restricts Legionella pneumophila infection. PLoS Pathog 5, e1000361.

Alexopoulou, L., Holt, A.C., Medzhitov, R., and Flavell, R.A. (2001). Recognition of double-stranded RNA and activation of NF-kappa B by Toll-like receptor 3. Nature 413, 732-738.

Alghamdi, M. (2017). Familial Mediterranean fever, review of the literature. Clinical rheumatology 36, 1707-1713.

Allen, I.C., Wilson, J.E., Schneider, M., Lich, J.D., Roberts, R.A., Arthur, J.C., Woodford, R.M.T., Davis, B.K., Uronis, J.M., Herfarth, H.H., et al. (2012). NLRP12 Suppresses Colon Inflammation and Tumorigenesis through the Negative Regulation of Noncanonical NF-kappa B Signaling. Immunity 36, 742-754.

Alpi, A.F., Chaugule, V., and Walden, H. (2016). Mechanism and disease association of E2-conjugating enzymes: lessons from UBE2T and UBE2L3. The Biochemical journal 473, 3401-3419.

Anand, P.K., Malireddi, R.K.S., Lukens, J.R., Vogel, P., Bertin, J., Lamkanfi, M., and Kanneganti, T.D. (2012). NLRP6 negatively regulates innate immunity and host defence against bacterial pathogens. Nature 488, 389-+.

Anderson, J.P., Mueller, J.L., Misaghi, A., Anderson, S., Sivagnanam, M., Kolodner, R.D., and Hoffman, H.M. (2008). Initial description of the human NLRP3 promoter. Genes and immunity 9, 721-726.

278

Arend, W.P., Gordon, D.F., Wood, W.M., Janson, R.W., Joslin, F.G., and Jameel, S. (1989). Il-1-Beta Production in Cultured Human-Monocytes Is Regulated at Multiple Levels. J Immunol 143, 118-126.

Ashida, H., and Sasakawa, C. (2017). Bacterial E3 ligase effectors exploit host ubiquitin systems. Current opinion in microbiology 35, 16-22.

Aubert, D.F., Xu, H., Yang, J.L., Shi, X.Y., Gao, W.Q., Li, L., Bisaro, F., Chen, S., Valvano, M.A., and Shao, F. (2016). A Burkholderia Type VI Effector Deamidates Rho GTPases to Activate the Pyrin Inflammasome and Trigger Inflammation. Cell host & microbe 19, 664-674.

Banka, P.A., Behera, A.P., Sarkar, S., and Datta, A.B. (2015). RING E3-Catalyzed E2 Self-Ubiquitination Attenuates the Activity of Ube2E Ubiquitin-Conjugating Enzymes. J Mol Biol 427, 2290-2304.

Bauer, S., Kirschning, C.J., Hacker, H., Redecke, V., Hausmann, S., Akira, S., Wagner, H., and Lipford, G.B. (2001). Human TLR9 confers responsiveness to bacterial DNA via species-specific CpG motif recognition. Proceedings of the National Academy of Sciences of the United States of America 98, 9237-9242.

Bauernfeind, F., Rieger, A., Schildberg, F.A., Knolle, P.A., Schmid-Burgk, J.L., and Hornung, V. (2012). NLRP3 Inflammasome Activity Is Negatively Controlled by miR- 223. J Immunol 189, 4175-4181.

Bauernfeind, F.G., Horvath, G., Stutz, A., Alnemri, E.S., MacDonald, K., Speert, D., Fernandes-Alnemri, T., Wu, J.H., Monks, B.G., Fitzgerald, K.A., et al. (2009). Cutting Edge: NF-kappa B Activating Pattern Recognition and Cytokine Receptors License NLRP3 Inflammasome Activation by Regulating NLRP3 Expression. J Immunol 183, 787-791.

Baugh, J.M., Viktorova, E.G., and Pilipenko, E.V. (2009). Proteasomes Can Degrade a Significant Proportion of Cellular Proteins Independent of Ubiquitination. J Mol Biol 386, 814-827.

Bednash, J.S., and Mallampalli, R.K. (2016). Regulation of inflammasomes by ubiquitination. Cellular & molecular immunology 13, 722-728.

279

Belizario, J.E., Alves, J., Garay-Malpartida, M., and Occhiucci, J.M. (2008). Coupling caspase cleavage and proteasomal degradation of proteins carrying PEST motif. Current protein & peptide science 9, 210-220.

Bellail, A.C., Olson, J.J., Yang, X., Chen, Z.J., and Hao, C. (2012). A20 ubiquitin ligase-mediated polyubiquitination of RIP1 inhibits caspase-8 cleavage and TRAIL- induced apoptosis in glioblastoma. Cancer discovery 2, 140-155.

Berger, C.N., Crepin, V.F., Baruch, K., Mousnier, A., Rosenshine, I., and Frankel, G. (2012). EspZ of Enteropathogenic and Enterohemorrhagic Escherichia coli Regulates Type III Secretion System Protein Translocation. mBio 3.

Beuzon, C.R., Banks, G., Deiwick, J., Hensel, M., and Holden, D.W. (1999). pH- dependent secretion of SseB, a product of the SPI-2 type III secretion system of Salmonella typhimurium. Mol Microbiol 33, 806-816.

Black, R.A., Kronheim, S.R., Cantrell, M., Deeley, M.C., March, C.J., Prickett, K.S., Wignall, J., Conlon, P.J., Cosman, D., Hopp, T.P., et al. (1988). Generation of Biologically-Active Interleukin-1-Beta by Proteolytic Cleavage of the Inactive Precursor. Journal of Biological Chemistry 263, 9437-9442.

Bloom, J., Amador, V., Bartolini, F., DeMartino, G., and Pagano, M. (2003). Proteasome-mediated degradation of p21 via N-terminal ubiquitinylation. Cell 115, 71-82.

Bosanac, I., Wertz, I.E., Pan, B., Yu, C., Kusam, S., Lam, C., Phu, L., Phung, Q., Maurer, B., Arnott, D., et al. (2010). Ubiquitin Binding to A20 ZnF4 Is Required for Modulation of NF-kappa B Signaling. Mol Cell 40, 548-557.

Botos, I., Segal, D.M., and Davies, D.R. (2011). The structural biology of Toll-like receptors. Structure 19, 447-459.

Boucher, D., Monteleone, M., Coll, R.C., Chen, K.W., Ross, C.M., Teo, J.L., Gomez, G.A., Holley, C.L., Bierschenk, D., Stacey, K.J., et al. (2018). Caspase-1 self- cleavage is an intrinsic mechanism to terminate inflammasome activity. The Journal of experimental medicine.

280

Bowyer, P.W., Simon, G.M., Cravatt, B.F., and Bogyo, M. (2011). Global Profiling of Proteolysis during Rupture of Plasmodium falciparum from the Host Erythrocyte. Mol Cell Proteomics 10.

Breitschopf, K., Bengal, E., Ziv, T., Admon, A., and Ciechanover, A. (1998). A novel site for ubiquitination: the N‑terminal residue, and not internal lysines of MyoD, is essential for conjugation and degradation of the protein. The EMBO journal 17, 5964-5973.

Broz, P., and Dixit, V.M. (2016). Inflammasomes: mechanism of assembly, regulation and signalling. Nature Reviews Immunology 16, 407-420.

Broz, P., and Monack, D.M. (2011). Molecular mechanisms of inflammasome activation during microbial infections. Immunol Rev 243, 174-190.

Broz, P., Newton, K., Lamkanfi, M., Mariathasan, S., Dixit, V.M., and Monack, D.M. (2010a). Redundant roles for inflammasome receptors NLRP3 and NLRC4 in host defense against Salmonella. J Exp Med 207, 1745-1755.

Broz, P., Ruby, T., Belhocine, K., Bouley, D.M., Kayagaki, N., Dixit, V.M., and Monack, D.M. (2012). Caspase-11 increases susceptibility to Salmonella infection in the absence of caspase-1. Nature 490, 288-+.

Broz, P., von Moltke, J., Jones, J.W., Vance, R.E., and Monack, D.M. (2010b). Differential Requirement for Caspase-1 Autoproteolysis in Pathogen-Induced Cell Death and Cytokine Processing. Cell host & microbe 8, 471-483.

Buetow, L., and Huang, D.T. (2016). Structural insights into the catalysis and regulation of E3 ubiquitin ligases. Nat Rev Mol Cell Bio 17, 626-642.

Bufler, P., Gamboni-Robertson, F., Azam, T., Kim, S.H., and Dinarello, C.A. (2004). Interleukin-1 homologues lL-1F7b and IL-18 contain functional mRNA instability elements within the coding region responsive to lipopolysaccharide. Biochemical Journal 381, 503-510.

Burckstummer, T., Baumann, C., Bluml, S., Dixit, E., Durnberger, G., Jahn, H., Planyavsky, M., Bilban, M., Colinge, J., Bennett, K.L., et al. (2009). An orthogonal

281 proteomic-genomic screen identifies AIM2 as a cytoplasmic DNA sensor for the inflammasome. Nature immunology 10, 266-272.

Cabral, V.P., Andrade, C.A., Passos, S.R., Martins, M.F., and Hokerberg, Y.H. (2016). Severe infection in patients with rheumatoid arthritis taking anakinra, rituximab, or abatacept: a systematic review of observational studies. Revista brasileira de reumatologia 56, 543-550.

Cai, X., Chen, J.Q., Xu, H., Liu, S.Q., Jiang, Q.X., Halfmann, R., and Chen, Z.J.J. (2014). Prion-like Polymerization Underlies Signal Transduction in Antiviral Immune Defense and Inflammasome Activation. Cell 156, 1207-1222.

Canna, S.W., de Jesus, A.A., Gouni, S., Brooks, S.R., Marrero, B., Liu, Y., DiMattia, M.A., Zaal, K.J., Sanchez, G.A., Kim, H., et al. (2014). An activating NLRC4 inflammasome mutation causes autoinflammation with recurrent macrophage activation syndrome. Nat Genet 46, 1140-1146.

Carvalho, A.F., Pinto, M.P., Grou, C.P., Vitorino, R., Domingues, P., Yamao, F., Sa- Miranda, C., and Azevedo, J.E. (2012). High-yield expression in Escherichia coli and purification of mouse ubiquitin-activating enzyme E1. Molecular biotechnology 51, 254-261.

Centola, M., Wood, G., Frucht, D.M., Galon, J., Aringer, M., Farrell, C., Kingma, D.W., Horwitz, M.E., Mansfield, E., Holland, S.M., et al. (2000). The gene for familial Mediterranean fever, MEFV, is expressed in early leukocyte development and is regulated in response to inflammatory mediators. Blood 95, 3223-3231.

Chaitanya, G.V., Steven, A.J., and Babu, P.P. (2010). PARP-1 cleavage fragments: signatures of cell-death proteases in neurodegeneration. Cell Commun Signal 8.

Chang, K., Marran, K., Valentine, A., and Hannon, G.J. (2013). Creating an miR30- based shRNA vector. Cold Spring Harbor protocols 2013, 631-635.

Chaugule, V.K., Burchell, L., Barber, K.R., Sidhu, A., Leslie, S.J., Shaw, G.S., and Walden, H. (2011). Autoregulation of Parkin activity through its ubiquitin‑like domain. The EMBO journal 30, 2853-2867.

282

Chavarria-Smith, J., Mitchell, P.S., Ho, A.M., Daugherty, M.D., and Vance, R.E. (2016). Functional and Evolutionary Analyses Identify Proteolysis as a General Mechanism for NLRP1 Inflammasome Activation. PLOS Pathogens 12.

Chavarria-Smith, J., and Vance, R.E. (2015). The NLRP1 inflammasomes. Immunol Rev 265, 22-34.

Chavarría-Smith, J., and Vance, R.E. (2013). Direct Proteolytic Cleavage of NLRP1B Is Necessary and Sufficient for Inflammasome Activation by Anthrax Lethal Factor. PLOS Pathogens 9, e1003452.

Chen, K.J., Shanmugam, N.K.N., Pazos, M.A., Hurley, B.P., and Cherayil, B.J. (2016). Commensal Bacteria-Induced Inflammasome Activation in Mouse and Human Macrophages Is Dependent on Potassium Efflux but Does Not Require Phagocytosis or Bacterial Viability. PloS one 11.

Chen, K.W., Gross, C.J., Sotomayor, F.V., Stacey, K.J., Tschopp, J., Sweet, M.J., and Schroder, K. (2014). The neutrophil NLRC4 inflammasome selectively promotes IL-1beta maturation without pyroptosis during acute Salmonella challenge. Cell Rep 8, 570-582.

Chen, X., Barton, L.F., Chi, Y., Clurman, B.E., and Roberts, J.M. (2007). Ubiquitin- independent degradation of cell-cycle inhibitors by the REGgamma proteasome. Mol Cell 26, 843-852.

Chiu, Y.H., Sun, Q., and Chen, Z.J.J. (2007). E1-L2 activates both ubiquitin and FAT10. Mol Cell 27, 1014-1023.

Chong, A., Wehrly, T.D., Nair, V., Fischer, E.R., Barker, J.R., Klose, K.E., and Celli, J. (2008). The Early Phagosomal Stage of Francisella tularensis Determines Optimal Phagosomal Escape and Francisella Pathogenicity Island Protein Expression. Infection and immunity 76, 5488-5499.

Christensen, D.E., Brzovic, P.S., and Klevit, R.E. (2007). E2-BRCA1 RING interactions dictate synthesis of mono- or specific polyubiquitin chain linkages. Nature structural & molecular biology 14, 941-948.

283

Chu, Z.L., Pio, F., Xie, Z., Welsh, K., Krajewska, M., Krajewski, S., Godzik, A., and Reed, J.C. (2001). A novel enhancer of the Apaf1 apoptosome involved in cytochrome c-dependent caspase activation and apoptosis. The Journal of biological chemistry 276, 9239-9245.

Chuang, T.H., and Ulevitch, R.J. (2004). Triad3A, an E3 ubiquitin-protein ligase regulating Toll-like receptors. Nature immunology 5, 495-502.

Chung, L.K., Park, Y.H., Zheng, Y., Brodsky, I.E., Hearing, P., Kastner, D.L., Chae, J.J., and Bliska, J.B. (2016). The Yersinia Virulence Factor YopM Hijacks Host Kinases to Inhibit Type III Effector-Triggered Activation of the Pyrin Inflammasome. Cell host & microbe 20, 296-306.

Chung, L.K., Philip, N.H., Schmidt, V.A., Koller, A., Strowig, T., Flavell, R.A., Brodsky, I.E., and Bliska, J.B. (2014). IQGAP1 is important for activation of caspase- 1 in macrophages and is targeted by Yersinia pestis type III effector YopM. mBio 5, e01402-01414.

Ciechanover, A. (1994). The ubiquitin-proteasome proteolytic pathway. Cell 79, 13- 21.

Ciechanover, A., Heller, H., Katz-Etzion, R., and Hershko, A. (1981). Activation of the heat-stable polypeptide of the ATP-dependent proteolytic system. Proceedings of the National Academy of Sciences of the United States of America 78, 761-765.

Clague, M.J., Heride, C., and Urbe, S. (2015). The demographics of the ubiquitin system. Trends in cell biology 25, 417-426.

Clark, K. (2014). Protein kinase networks that limit TLR signalling. Biochemical Society transactions 42, 11-24.

Cogswell, J.P., Godlevski, M.M., Wisely, G.B., Clay, W.C., Leesnitzer, L.M., Ways, J.P., and Gray, J.G. (1994). Nf-Kappa-B Regulates Il-1-Beta Transcription through a Consensus Nf-Kappa-B Binding-Site and a Nonconsensus Cre-Like Site. J Immunol 153, 712-723.

284

Cohen-Kaplan, V., Ciechanover, A., and Livneh, I. (2016). p62 at the crossroad of the ubiquitin-proteasome system and autophagy. Oncotarget 7, 83833-83834.

Coll, R.C., Robertson, A.A.B., Chae, J.J., Higgins, S.C., Munoz-Planillo, R., Inserra, M.C., Vetter, I., Dungan, L.S., Monks, B.G., Stutz, A., et al. (2015). A small-molecule inhibitor of the NLRP3 inflammasome for the treatment of inflammatory diseases. Nature medicine 21, 248-+.

Conos, S.A., Lawlor, K.E., Vaux, D.L., Vince, J.E., and Lindqvist, L.M. (2016). Cell death is not essential for caspase-1-mediated interleukin-1beta activation and secretion. Cell death and differentiation 23, 1827-1838.

Coulombe, P., Rodier, G., Bonneil, E., Thibault, P., and Meloche, S. (2004). N- terminal ubiquitination of extracellular signal-regulated kinase 3 and p21 directs their degradation by the proteasome. Mol Cell Biol 24, 6140-6150.

Creasey, E.A., and Isberg, R.R. (2012). The protein SdhA maintains the integrity of the Legionella-containing vacuole. Proceedings of the National Academy of Sciences of the United States of America 109, 3481-3486.

Cullen, S.P., Kearney, C.J., Clancy, D.M., and Martin, S.J. (2015). Diverse Activators of the NLRP3 Inflammasome Promote IL-1beta Secretion by Triggering Necrosis. Cell Rep 11, 1535-1548.

Cunha, L.D., Silva, A.L.N., Ribeiro, J.M., Mascarenhas, D.P.A., Quirino, G.F.S., Santos, L.L., Flavell, R.A., and Zamboni, D.S. (2017). AIM2 Engages Active but Unprocessed Caspase-1 to Induce Noncanonical Activation of the NLRP3 Inflammasome. Cell Reports 20, 794-805.

Cunha, L.D., and Zamboni, D.S. (2013). Subversion of inflammasome activation and pyroptosis by pathogenic bacteria. Frontiers in cellular and infection microbiology 3, 76.

Cunningham, C.N., Baughman, J.M., Phu, L., Tea, J.S., Yu, C., Coons, M., Kirkpatrick, D.S., Bingol, B., and Corn, J.E. (2015). USP30 and parkin homeostatically regulate atypical ubiquitin chains on mitochondria. Nat Cell Biol 17, 160-169.

285

Cybulsky, M.I., Colditz, I.G., and Movat, H.Z. (1986). The Role of Interleukin-1 in Neutrophil Leukocyte Emigration Induced by Endotoxin. Am J Pathol 124, 367-372.

D'Cruz, A.A., Babon, J.J., Norton, R.S., Nicola, N.A., and Nicholson, S.E. (2013). Structure and function of the SPRY/B30.2 domain proteins involved in innate immunity. Protein science : a publication of the Protein Society 22, 1-10.

Datsenko, K.A., and Wanner, B.L. (2000). One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proceedings of the National Academy of Sciences of the United States of America 97, 6640-6645.

Davaro, F., Forde, S.D., Garfield, M., Jiang, Z., Halmen, K., Tamburro, N.D., Kurt- Jones, E., Fitzgerald, K.A., Golenbock, D.T., and Wang, D. (2014). 3-Hydroxyl-3- methylglutaryl coenzyme A (HMG-CoA) reductase inhibitor (statin)-induced 28-kDa interleukin-1beta interferes with mature IL-1beta signaling. The Journal of biological chemistry 289, 16214-16222. de Calignon, A., Fox, L.M., Pitstick, R., Carlson, G.A., Bacskai, B.J., Spires-Jones, T.L., and Hyman, B.T. (2010). Caspase activation precedes and leads to tangles. Nature 464, 1201-1204. de Jong, M.F., Liu, Z.X., Chen, D.D., and Alto, N.M. (2016). Shigella flexneri suppresses NF-kappa B activation by inhibiting linear ubiquitin chain ligation. Nat Microbiol 1. de Torre-Minguela, C., del Castillo, P.M., and Pelegrin, P. (2017). The NLRP3 and Pyrin Inflammasomes: Implications in the Pathophysiology of Autoinflammatory Diseases. Frontiers in immunology 8.

Debets, R., Timans, J.C., Churakowa, T., Zurawski, S., Malefyt, R.D., Moore, K.W., Abrams, J.S., O'Garra, A., Bazan, J.F., and Kastelein, R.A. (2000). IL-18 receptors, their role in ligand binding and function: Anti-IL-1RAcPL antibody, a potent antagonist of IL-18. J Immunol 165, 4950-4956.

Demontis, S., Rigo, C., Piccinin, S., Mizzau, M., Sonego, M., Fabris, M., Brancolini, C., and Maestro, R. (2006). Twist is substrate for caspase cleavage and proteasome-mediated degradation. Cell death and differentiation 13, 335-345.

286

Denes, A., Lopez-Castejon, G., and Brough, D. (2012). Caspase-1: is IL-1 just the tip of the ICEberg? Cell death & disease 3.

Deretic, V., Saitoh, T., and Akira, S. (2013). Autophagy in infection, inflammation and immunity. Nature Reviews Immunology 13, 722-737.

Di Micco, A., Frera, G., Lugrin, J., Jamilloux, Y., Hsu, E.T., Tardivel, A., De Gassart, A., Zaffalon, L., Bujisic, B., Siegert, S., et al. (2016). AIM2 inflammasome is activated by pharmacological disruption of nuclear envelope integrity. Proceedings of the National Academy of Sciences of the United States of America 113, E4671-E4680.

Diao, J., Zhang, Y., Huibregtse, J.M., Zhou, D., and Chen, J. (2008). Crystal structure of SopA, a Salmonella effector protein mimicking a eukaryotic ubiquitin ligase. Nature structural & molecular biology 15, 65-70.

Diebolder, C.A., Halff, E.F., Koster, A.J., Huizinga, E.G., and Koning, R.I. (2015). Cryoelectron Tomography of the NAIP5/NLRC4 Inflammasome: Implications for NLR Activation. Structure 23, 2349-2357.

Dinarello, C.A. (1996). Biologic basis for interleukin-1 in disease. Blood 87, 2095- 2147.

Dinarello, C.A. (1998). Interleukin-1β, Interleukin-18, and the Interleukin-1β Converting Enzymea. Annals of the New York Academy of Sciences 856, 1-11.

Dinarello, C.A., and Thompson, R.C. (1991). Blocking IL-1: interleukin 1 receptor antagonist in vivo and in vitro. Immunology today 12, 404-410.

Ding, J., Wang, K., Liu, W., She, Y., Sun, Q., Shi, J., Sun, H., Wang, D.C., and Shao, F. (2016). Pore-forming activity and structural autoinhibition of the gasdermin family. Nature 535, 111-116.

DiPeso, L., Ji, D.X., Vance, R.E., and Price, J.V. (2017). Cell death and cell lysis are separable events during pyroptosis. Cell death discovery 3, 17070.

Dix, M.M., Simon, G.M., and Cravatt, B.F. (2008). Global mapping of the topography and magnitude of proteolytic events in apoptosis. Cell 134, 679-691.

287

Dix, M.M., Simon, G.M., and Cravatt, B.F. (2014). Global Identification of Caspase Substrates Using PROTOMAP (Protein Topography and Migration Analysis Platform). In Caspases,Paracaspases, and Metacaspases: Methods and Protocols, P. V. Bozhkov, and G. Salvesen, eds. (New York, NY: Springer New York), pp. 61- 70.

Dolan, P.J., and Johnson, G.V.W. (2010). A Caspase Cleaved Form of Tau Is Preferentially Degraded through the Autophagy Pathway. Journal of Biological Chemistry 285, 21978-21987.

Dong, H., Dumenil, J., Lu, F.-H., Na, L., Vanhaeren, H., Naumann, C., Klecker, M., Prior, R., Smith, C., McKenzie, N., et al. (2016). Ubiquitylation activates a peptidase that promotes cleavage and destabilization of its activating E3 ligases and diverse growth regulatory proteins to limit cell proliferation in Arabidopsis. bioRxiv.

Dove, K.K., Olszewski, J.L., Martino, L., Duda, D.M., Wu, X.S., Miller, D.J., Reiter, K.H., Rittinger, K., Schulman, B.A., and Klevit, R.E. (2017). Structural Studies of HHARI/UbcH7 approximately Ub Reveal Unique E2 approximately Ub Conformational Restriction by RBR RING1. Structure 25, 890-900 e895.

Dove, K.K., Stieglitz, B., Duncan, E.D., Rittinger, K., and Klevit, R.E. (2016). Molecular insights into RBR E3 ligase ubiquitin transfer mechanisms. Embo Rep 17, 1221-1235.

Duncan, J.A., and Canna, S.W. (2018). The NLRC4 Inflammasome. Immunol Rev 281, 115-123.

Duong, B.H., Onizawa, M., Oses-Prieto, J.A., Advincula, R., Burlingame, A., Malynn, B.A., and Ma, A. (2015). A20 restricts ubiquitination of pro-interleukin-1beta protein complexes and suppresses NLRP3 inflammasome activity. Immunity 42, 55-67.

Eigenbrod, T., Park, J.H., Harder, J., Iwakura, Y., and Nunez, G. (2008). Cutting Edge: Critical Role for Mesothelial Cells in Necrosis-Induced Inflammation through the Recognition of IL-1 alpha Released from Dying Cells. J Immunol 181, 8194- 8198.

288

Eitel, J., Suttorp, N., and Opitz, B. (2010). Innate immune recognition and inflammasome activation in listeria monocytogenes infection. Frontiers in microbiology 1, 149.

Eldridge, M.J., and Shenoy, A.R. (2015). Antimicrobial inflammasomes: unified signalling against diverse bacterial pathogens. Current opinion in microbiology 23, 32-41.

Eldridge, M.J.G., Sanchez-Garrido, J., Hoben, G.F., Goddard, P.J., and Shenoy, A.R. (2017). The Atypical Ubiquitin E2 Conjugase UBE2L3 Is an Indirect Caspase-1 Target and Controls IL-1β Secretion by Inflammasomes. Cell Reports 18, 1285-1297.

Emmerich, C.H., Ordureau, A., Strickson, S., Arthur, J.S., Pedrioli, P.G., Komander, D., and Cohen, P. (2013). Activation of the canonical IKK complex by K63/M1-linked hybrid ubiquitin chains. Proceedings of the National Academy of Sciences of the United States of America 110, 15247-15252.

Endrizzi, M.G., Hadinoto, V., Growney, J.D., Miller, W., and Dietrich, W.F. (2000). Genomic sequence analysis of the mouse Naip gene array. Genome Res 10, 1095- 1102.

Ernst, O., and Zor, T. (2010). Linearization of the bradford protein assay. Journal of visualized experiments : JoVE.

Evavold, C.L., Ruan, J., Tan, Y., Xia, S., Wu, H., and Kagan, J.C. (2017). The Pore- Forming Protein Gasdermin D Regulates Interleukin-1 Secretion from Living Macrophages. Immunity.

Fantuzzi, G., Puren, A.J., Harding, M.W., Livingston, D.J., and Dinarello, C.A. (1998). Interleukin-18 regulation of interferon gamma production and cell proliferation as shown in interleukin-1beta-converting enzyme (caspase-1)-deficient mice. Blood 91, 2118-2125.

Faustin, B., Lartigue, L., Bruey, J.M., Luciano, F., Sergienko, E., Bailly-Maitre, B., Volkmann, N., Hanein, D., Rouiller, I., and Reed, J.C. (2007). Reconstituted NALP1

289 inflammasome reveals two-step mechanism of caspase-1 activation. Mol Cell 25, 713-724.

Fellmann, C., Hoffmann, T., Sridhar, V., Hopfgartner, B., Muhar, M., Roth, M., Lai, D.Y., Barbosa, I.A., Kwon, J.S., Guan, Y., et al. (2013). An optimized microRNA backbone for effective single-copy RNAi. Cell Rep 5, 1704-1713.

Fernandes-Alnemri, T., Kang, S., Anderson, C., Sagara, J., Fitzgerald, K.A., and Alnemri, E.S. (2013). Cutting edge: TLR signaling licenses IRAK1 for rapid activation of the NLRP3 inflammasome. J Immunol 191, 3995-3999.

Fernandes-Alnemri, T., Yu, J.W., Datta, P., Wu, J., and Alnemri, E.S. (2009). AIM2 activates the inflammasome and cell death in response to cytoplasmic DNA. Nature 458, 509-513.

Fernandes-Alnemri, T., Yu, J.W., Juliana, C., Solorzano, L., Kang, S., Wu, J.H., Datta, P., McCormick, M., Huang, L., McDermott, E., et al. (2010). The AIM2 inflammasome is critical for innate immunity to Francisella tularensis. Nature immunology 11, 385-394.

Fettelschoss, A., Kistowska, M., LeibundGut-Landmann, S., Beer, H.D., Johansen, P., Senti, G., Contassot, E., Bachmann, M.F., French, L.E., Oxenius, A., et al. (2011). Inflammasome activation and IL-1 beta target IL-1 alpha for secretion as opposed to surface expression. Proceedings of the National Academy of Sciences of the United States of America 108, 18055-18060.

Finger, J.N., Lich, J.D., Dare, L.C., Cook, M.N., Brown, K.K., Duraiswami, C., Bertin, J.J., and Gough, P.J. (2012). Autolytic Proteolysis within the Function to Find Domain (FIIND) Is Required for NLRP1 Inflammasome Activity. Journal of Biological Chemistry 287, 25030-25037.

Fink, S.L., and Cookson, B.T. (2006). Caspase-1-dependent pore formation during pyroptosis leads to osmotic lysis of infected host macrophages. Cell Microbiol 8, 1812-1825.

290

Fiskin, E., Bhogaraju, S., Herhaus, L., Kalayil, S., Hahn, M., and Dikic, I. (2017). Structural basis for the recognition and degradation of host TRIM proteins by Salmonella effector SopA. Nature communications 8.

Fiskin, E., Bionda, T., Dikic, I., and Behrends, C. (2016). Global Analysis of Host and Bacterial Ubiquitinome in Response to Salmonella Typhimurium Infection. Mol Cell 62, 967-981.

Fortmann, K.T., Lewis, R.D., Ngo, K.A., Fagerlund, R., and Hoffmann, A. (2015). A Regulated, Ubiquitin-Independent Degron in IkappaBalpha. J Mol Biol 427, 2748- 2756.

Franchi, L., Amer, A., Body-Malapel, M., Kanneganti, T.D., Ozoren, N., Jagirdar, R., Inohara, N., Vandenabeele, P., Bertin, J., Coyle, A., et al. (2006). Cytosolic flagellin requires Ipaf for activation of caspase-1 and interleukin 1 beta in salmonella-infected macrophages. Nature immunology 7, 576-582.

Franchi, L., Eigenbrod, T., Munoz-Planillo, R., and Nunez, G. (2009). The inflammasome: a caspase-1-activation platform that regulates immune responses and disease pathogenesis. Nature immunology 10, 241-247.

Franchi, L., and Núñez, G. (2008). The Nlrp3 inflammasome is critical for aluminium hydroxide-mediated IL-1β secretion but dispensable for adjuvant activity. European Journal of Immunology 38, 2085-2089.

Franke, A., McGovern, D.P.B., Barrett, J.C., Wang, K., Radford-Smith, G.L., Ahmad, T., Lees, C.W., Balschun, T., Lee, J., Roberts, R., et al. (2010). Genome-wide meta- analysis increases to 71 the number of confirmed Crohn's disease susceptibility loci. Nat Genet 42, 1118-+.

Frankel, G., and Phillips, A.D. (2008). Attaching effacing Escherichia coli and paradigms of Tir-triggered actin polymerization: getting off the pedestal. Cell Microbiol 10, 549-556.

Fransen, K., Visschedijk, M.C., van Sommeren, S., Fu, J.Y.Y., Franke, L., Festen, E.A.M., Stokkers, P.C.F., van Bodegraven, A.A., Crusius, J.B.A., Hommes, D.W., et al. (2010). Analysis of SNPs with an effect on gene expression identifies UBE2L3

291 and BCL3 as potential new risk genes for Crohn's disease. Hum Mol Genet 19, 3482-3488.

Frew, B.C., Joag, V.R., and Mogridge, J. (2012). Proteolytic Processing of Nlrp1b Is Required for Inflammasome Activity. PLOS Pathogens 8.

Fu, B., Li, S., Wang, L., Berman, M.A., and Dorf, M.E. (2014). The ubiquitin conjugating enzyme UBE2L3 regulates TNFalpha-induced linear ubiquitination. Cell research 24, 376-379.

Gaidt, M.M., Ebert, T.S., Chauhan, D., Schmidt, T., Schmid-Burgk, J.L., Rapino, F., Robertson, A.A., Cooper, M.A., Graf, T., and Hornung, V. (2016). Human Monocytes Engage an Alternative Inflammasome Pathway. Immunity 44, 833-846.

Gao, W.Q., Yang, J.L., Liu, W., Wang, Y.P., and Shao, F. (2016). Site-specific phosphorylation and microtubule dynamics control Pyrin inflammasome activation. Proceedings of the National Academy of Sciences of the United States of America 113, E4857-E4866.

Gasteiger, E., Gattiker, A., Hoogland, C., Ivanyi, I., Appel, R.D., and Bairoch, A. (2003). ExPASy: The proteomics server for in-depth protein knowledge and analysis. Nucleic Acids Res 31, 3784-3788.

Ge, J., Gong, Y.N., Xu, Y., and Shao, F. (2012). Preventing bacterial DNA release and absent in melanoma 2 inflammasome activation by a Legionella effector functioning in membrane trafficking. Proceedings of the National Academy of Sciences of the United States of America 109, 6193-6198.

Geddes, B.J., Wang, L., Huang, W.J., Lavellee, M., Manji, G.A., Brown, M., Jurman, M., Cao, J., Morgenstern, J., Merriam, S., et al. (2001). Human CARD12 is a novel CED4/Apaf-1 family member that induces apoptosis. Biochem Bioph Res Co 284, 77-82.

Geisler, S., Vollmer, S., Golombek, S., and Kahle, P.J. (2014). The ubiquitin- conjugating enzymes UBE2N, UBE2L3 and UBE2D2/3 are essential for Parkin- dependent mitophagy. Journal of cell science 127, 3280-3293.

292

Gerdes, N., Sukhova, G.K., Libby, P., Reynolds, R.S., Young, J.L., and Schonbeck, U. (2002). Expression of interleukin (IL)-18 and functional IL-18 receptor on human vascular endothelial cells, smooth muscle cells, and macrophages: Implications for atherogenesis. The Journal of experimental medicine 195, 245-257.

Ghayur, T., Banerjee, S., Hugunin, M., Butler, D., Herzog, L., Carter, A., Quintal, L., Sekut, L., Talanian, R., Paskind, M., et al. (1997). Caspase-1 processes IFN- gamma-inducing factor and regulates LPS-induced IFN-gamma production. Nature 386, 619-623.

Grandjean, T., Boucher, A., Thepaut, M., Monlezun, L., Guery, B., Faudry, E., Kipnis, E., and Dessein, R. (2017). The human NAIP-NLRC4-inflammasome senses the Pseudomonas aeruginosa T3SS inner-rod protein. Int Immunol 29, 377-384.

Granel, B., Philip, N., Serratrice, J., Ene, N., Grateau, G., Dode, C., Cuisset, L., Disdier, P., Berbis, P., Delpech, M., et al. (2003). CIAS1 mutation in a patient with overlap between Muckle-Wells and chronic infantile neurological cutaneous and articular syndromes. Dermatology 206, 257-259.

Grice, G.L., Lobb, I.T., Weekes, M.P., Gygi, S.P., Antrobus, R., and Nathan, J.A. (2015). The Proteasome Distinguishes between Heterotypic and Homotypic Lysine- 11-Linked Polyubiquitin Chains. Cell Reports 12, 545-553.

Grice, G.L., and Nathan, J.A. (2016). The recognition of ubiquitinated proteins by the proteasome. Cell Mol Life Sci 73, 3497-3506.

Grishin, A.M., Condos, T.E., Barber, K.R., Campbell-Valois, F.X., Parsot, C., Shaw, G.S., and Cygler, M. (2014). Structural basis for the inhibition of host protein ubiquitination by Shigella effector kinase OspG. Structure 22, 878-888.

Groll, M., Kim, K.B., Kairies, N., Huber, R., and Crews, C.M. (2000). Crystal Structure of Epoxomicin:20S Proteasome Reveals a Molecular Basis for Selectivity of α‘,β‘-Epoxyketone Proteasome Inhibitors. Journal of the American Chemical Society 122, 1237-1238.

Groß, O., Yazdi, Amir S., Thomas, Christina J., Masin, M., Heinz, Leonhard X., Guarda, G., Quadroni, M., Drexler, Stefan K., and Tschopp, J. (2012).

293

Inflammasome Activators Induce Interleukin-1α Secretion via Distinct Pathways with Differential Requirement for the Protease Function of Caspase-1. Immunity 36, 388- 400.

Growney, J.D., and Dietrich, W.F. (2000). High-resolution genetic and physical map of the Lgnl interval in C57BL/6J implicates Naip2 or Naip5 in Legionella pneumophila pathogenesis. Genome Res 10, 1158-1171.

Gu, Y., Kuida, K., Tsutsui, H., Ku, G., Hsiao, K., Fleming, M.A., Hayashi, N., Higashino, K., Okamura, H., Nakanishi, K., et al. (1997). Activation of interferon- gamma inducing factor mediated by interleukin-1beta converting enzyme. Science 275, 206-209.

Guan, K., Wei, C.W., Zheng, Z.R., Song, T., Wu, F.X., Zhang, Y.H., Cao, Y., Ma, S.L., Chen, W., Xu, Q.B., et al. (2015). MAVS Promotes Inflammasome Activation by Targeting ASC for K63-Linked Ubiquitination via the E3 Ligase TRAF3. J Immunol 194, 4880-4890.

Guan, Y., Ranoa, D.R., Jiang, S., Mutha, S.K., Li, X., Baudry, J., and Tapping, R.I. (2010). Human TLRs 10 and 1 share common mechanisms of innate immune sensing but not signaling. J Immunol 184, 5094-5103.

Gurcel, L., Abrami, L., Girardin, S., Tschopp, J., and van der Goot, F.G. (2006). Caspase-1 activation of lipid metabolic pathways in response to bacterial pore- forming toxins promotes cell survival. Cell 126, 1135-1145.

Gutierrez, O., Pipaon, C., and Fernandez-Luna, J.L. (2004). Ipaf is upregulated by tumor necrosis factor-alpha in human leukemia cells. FEBS letters 568, 79-82.

Haas, A.L., Warms, J.V.B., and Rose, I.A. (1983). Ubiquitin Adenylate - Structure and Role in Ubiquitin Activation. Biochemistry-Us 22, 4388-4394.

Hacker, H., and Karin, M. (2006). Regulation and function of IKK and IKK-related kinases. Science's STKE : signal transduction knowledge environment 2006, re13.

294

Hagar, J.A., Powell, D.A., Aachoui, Y., Ernst, R.K., and Miao, E.A. (2013). Cytoplasmic LPS Activates Caspase-11: Implications in TLR4-Independent Endotoxic Shock. Science 341, 1250-1253.

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. Nat Cell Biol 5, 461-466.

Han, J.W., Zheng, H.F., Cui, Y., Sun, L.D., Ye, D.Q., Hu, Z., Xu, J.H., Cai, Z.M., Huang, W., Zhao, G.P., et al. (2009). Genome-wide association study in a Chinese Han population identifies nine new susceptibility loci for systemic lupus erythematosus. Nat Genet 41, 1234-U1298.

Han, S., Lear, T.B., Jerome, J.A., Rajbhandari, S., Snavely, C.A., Gulick, D.L., Gibson, K.F., Zou, C., Chen, B.B., and Mallampalli, R.K. (2015). Lipopolysaccharide Primes the NALP3 Inflammasome by Inhibiting Its Ubiquitination and Degradation Mediated by the SCFFBXL2 E3 Ligase. The Journal of biological chemistry 290, 18124-18133.

Harbers, K., Muller, U., Grams, A., Li, E., Jaenisch, R., and Franz, T. (1996). Provirus integration into a gene encoding a ubiquitin-conjugating enzyme results in a placental defect and embryonic lethality. Proceedings of the National Academy of Sciences of the United States of America 93, 12412-12417.

Harris, J., Hartman, M., Roche, C., Zeng, S.G., O'Shea, A., Sharp, F.A., Lambe, E.M., Creagh, E.M., Golenbock, D.T., Tschopp, J., et al. (2011a). Autophagy controls IL-1beta secretion by targeting pro-IL-1beta for degradation. The Journal of biological chemistry 286, 9587-9597.

Harris, J., Hartman, M., Roche, C., Zeng, S.J.G., O'Shea, A., Sharp, F.A., Lambe, E.M., Creagh, E.M., Golenbock, D.T., Tschopp, J., et al. (2011b). Autophagy Controls IL-1 beta Secretion by Targeting Pro-IL-1 beta for Degradation. Journal of Biological Chemistry 286, 9587-9597.

Harvey, K.F., Harvey, N.L., Michael, J.M., Parasivam, G., Waterhouse, N., Alnemri, E.S., Watters, D., and Kumar, S. (1998). Caspase-mediated cleavage of the

295 ubiquitin-protein ligase Nedd4 during apoptosis. Journal of Biological Chemistry 273, 13524-13530.

Hatai, H., Lepelley, A., Zeng, W.Y., Hayden, M.S., and Ghosh, S. (2016). Toll-Like Receptor 11 (TLR11) Interacts with Flagellin and Profilin through Disparate Mechanisms. PloS one 11.

Hawkins, P.N., Lachmann, H.J., and McDermott, M.F. (2003). Interleukin-1-receptor antagonist in the Muckle-Wells syndrome. The New England journal of medicine 348, 2583-2584.

Hayashi, F., Smith, K.D., Ozinsky, A., Hawn, T.R., Yi, E.C., Goodlett, D.R., Eng, J.K., Akira, S., Underhill, D.M., and Aderem, A. (2001). The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5. Nature 410, 1099-1103.

He, W.T., Wan, H.Q., Hu, L.C., Chen, P.D., Wang, X., Huang, Z., Yang, Z.H., Zhong, C.Q., and Han, J.H. (2015). Gasdermin D is an executor of pyroptosis and required for interleukin-1 beta secretion. Cell research 25, 1285-1298.

He, Y., Hara, H., and Nunez, G. (2016a). Mechanism and Regulation of NLRP3 Inflammasome Activation. Trends in biochemical sciences 41, 1012-1021.

He, Y., Zeng, M.Y., Yang, D.H., Metro, B., and Nunez, G. (2016b). NEK7 is an essential mediator of NLRP3 activation downstream of potassium efflux. Nature 530, 354-+.

Heil, F., Hemmi, H., Hochrein, H., Ampenberger, F., Kirschning, C., Akira, S., Lipford, G., Wagner, H., and Bauer, S. (2004). Species-specific recognition of single- stranded RNA via toll-like receptor 7 and 8. Science 303, 1526-1529.

Heilig, R., Dick, M.S., Sborgi, L., Meunier, E., Hiller, S., and Broz, P. (2017). The Gasdermin-D pore acts as a conduit for IL-1β secretion in mice. European Journal of Immunology, n/a-n/a.

Hellmich, K.A., Levinsohn, J.L., Fattah, R., Newman, Z.L., Maier, N., Sastalla, I., Liu, S.H., Leppla, S.H., and Moayeri, M. (2012). Anthrax Lethal Factor Cleaves Mouse Nlrp1b in Both Toxin-Sensitive and Toxin-Resistant Macrophages. PloS one 7.

296

Hershko, A., and Ciechanover, A. (1992). The ubiquitin system for protein degradation. Annual review of biochemistry 61, 761-807.

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

Hershko, A., and Heller, H. (1985). Occurrence of a Polyubiquitin Structure in Ubiquitin-Protein Conjugates. Biochem Bioph Res Co 128, 1079-1086.

Hershko, A., Heller, H., Elias, S., and Ciechanover, A. (1983). Components of ubiquitin-protein ligase system. Resolution, affinity purification, and role in protein breakdown. The Journal of biological chemistry 258, 8206-8214.

Hlaing, T., Guo, R.F., Dilley, K.A., Loussia, J.M., Morrish, T.A., Shi, M.M., Vincenz, C., and Ward, P.A. (2001). Molecular cloning and characterization of DEFCAP-L and -S, two isoforms of a novel member of the mammalian Ced-4 family of apoptosis proteins. The Journal of biological chemistry 276, 9230-9238.

Hoffman, H.M., and Broderick, L. (2016). The role of the inflammasome in patients with autoinflammatory diseases. The Journal of allergy and clinical immunology 138, 3-14.

Hornbeck, P.V., Zhang, B., Murray, B., Kornhauser, J.M., Latham, V., and Skrzypek, E. (2015). PhosphoSitePlus, 2014: mutations, PTMs and recalibrations. Nucleic Acids Res 43, D512-D520.

Hornung, V., Ablasser, A., Charrel-Dennis, M., Bauernfeind, F., Horvath, G., Caffrey, D.R., Latz, E., and Fitzgerald, K.A. (2009). AIM2 recognizes cytosolic dsDNA and forms a caspase-1-activating inflammasome with ASC. Nature 458, 514-518.

Hornung, V., Bauernfeind, F., Halle, A., Samstad, E.O., Kono, H., Rock, K.L., Fitzgerald, K.A., and Latz, E. (2008). Silica crystals and aluminum salts activate the NALP3 inflammasome through phagosomal destabilization. Nature immunology 9, 847-856.

297

Hospenthal, M.K., Freund, S.M., and Komander, D. (2013). Assembly, analysis and architecture of atypical ubiquitin chains. Nature structural & molecular biology 20, 555-565.

Howard, A.D., Chartrain, N., Ding, G.F., Kostura, M.J., Limjuco, G., Schmidt, J.A., and Tocci, M.J. (1991a). Probing the Role of Interleukin-1-Beta Convertase in Interleukin-1-Beta Secretion. Agent Action Suppl 35, 77-83.

Howard, A.D., Kostura, M.J., Thornberry, N., Ding, G.J., Limjuco, G., Weidner, J., Salley, J.P., Hogquist, K.A., Chaplin, D.D., Mumford, R.A., et al. (1991b). IL-1- converting enzyme requires aspartic acid residues for processing of the IL-1 beta precursor at two distinct sites and does not cleave 31-kDa IL-1 alpha. J Immunol 147, 2964-2969.

Hsu, F., Luo, X., Qiu, J., Teng, Y.B., Jin, J., Smolka, M.B., Luo, Z.Q., and Mao, Y. (2014). The Legionella effector SidC defines a unique family of ubiquitin ligases important for bacterial phagosomal remodeling. Proceedings of the National Academy of Sciences of the United States of America 111, 10538-10543.

Hu, B., Jin, C.C., Li, H.B., Tong, J.Y., Ouyang, X.S., Cetinbas, N.M., Zhu, S., Strowig, T., Lam, F.C., Zhao, C., et al. (2016). The DNA-sensing AIM2 inflammasome controls radiation-induced cell death and tissue injury. Science 354, 765-768.

Hu, H., and Sun, S.C. (2016). Ubiquitin signaling in immune responses. Cell research 26, 457-483.

Hu, Z.H., Yan, C.Y., Liu, P.Y., Huang, Z.W., Ma, R., Zhang, C.L., Wang, R.Y., Zhang, Y.T., Martinon, F., Miao, D., et al. (2013). Crystal Structure of NLRC4 Reveals Its Autoinhibition Mechanism. Science 341, 172-175.

Hu, Z.H., Zhou, Q., Zhang, C.L., Fan, S.L., Cheng, W., Zhao, Y., Shao, F., Wang, H.W., Sui, S.F., and Chai, J.J. (2015). Structural and biochemical basis for induced self-propagation of NLRC4. Science 350, 399-404.

298

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.

Hurgin, V., Novick, D., Werman, A., Dinarello, C.A., and Rubinstein, M. (2007). Antiviral and immunoregulatory activities of IFN-gamma depend on constitutively expressed IL-1alpha. Proceedings of the National Academy of Sciences of the United States of America 104, 5044-5049.

Huxford, T., Huang, D.B., Malek, S., and Ghosh, G. (1998). The crystal structure of the IkappaBalpha/NF-kappaB complex reveals mechanisms of NF-kappaB inactivation. Cell 95, 759-770.

Hybiske, K., and Stephens, R.S. (2008). Exit strategies of intracellular pathogens. Nat Rev Microbiol 6, 99-110.

Ikeda, F., Deribe, Y.L., Skanland, S.S., Stieglitz, B., Grabbe, C., Franz-Wachtel, M., van Wijk, S.J., Goswami, P., Nagy, V., Terzic, J., et al. (2011). SHARPIN forms a linear ubiquitin ligase complex regulating NF-kappaB activity and apoptosis. Nature 471, 637-641.

Ireton, K. (2007). Entry of the bacterial pathogen Listeria monocytogenes into mammalian cells. Cell Microbiol 9, 1365-1375.

Isberg, R.R., O'Connor, T.J., and Heidtman, M. (2009). The Legionella pneumophila replication vacuole: making a cosy niche inside host cells. Nat Rev Microbiol 7, 12- 24.

Iwai, K. (2010). Functions of Linear Ubiquitin Chains in the NF-kappaB Pathway : Linear Polyubiquitin in NF-kappaB Signaling. Sub-cellular biochemistry 54, 100-106.

Jacobs, M.D., and Harrison, S.C. (1998). Structure of an IkappaBalpha/NF-kappaB complex. Cell 95, 749-758.

Jamilloux, Y., Pierini, R., Querenet, M., Juruj, C., Fauchais, A.L., Jauberteau, M.O., Jarraud, S., Lina, G., Etienne, J., Roy, C.R., et al. (2013). Inflammasome activation

299 restricts Legionella pneumophila replication in primary microglial cells through flagellin detection. Glia 61, 539-549.

Janeway, C.A., Jr. (1989). Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harbor symposia on quantitative biology 54 Pt 1, 1-13.

Jeong, J.Y., Yim, H.S., Ryu, J.Y., Lee, H.S., Lee, J.H., Seen, D.S., and Kang, S.G. (2012). One-step sequence- and ligation-independent cloning as a rapid and versatile cloning method for functional genomics studies. Applied and environmental microbiology 78, 5440-5443.

Jin, J.P., Li, X., Gygi, S.P., and Harper, J.W. (2007a). Dual E1 activation systems for ubiquitin differentially regulate E2 enzyme charging. Nature 447, 1135-U1117.

Jin, M.S., Kim, S.E., Heo, J.Y., Lee, M.E., Kim, H.M., Paik, S.G., Lee, H., and Lee, J.O. (2007b). Crystal structure of the TLR1-TLR2 heterodimer induced by binding of a tri-acylated lipopeptide. Cell 130, 1071-1082.

Jin, T.C., Perry, A., Jiang, J.S., Smith, P., Curry, J.A., Unterholzner, L., Jiang, Z.Z., Horvath, G., Rathinam, V.A., Johnstone, R.W., et al. (2012). Structures of the HIN Domain: DNA Complexes Reveal Ligand Binding and Activation Mechanisms of the AIM2 Inflammasome and IFI16 Receptor. Immunity 36, 561-571.

Jin, T.C., Perry, A., Smith, P., Jiang, J.S., and Xiao, T.S. (2013). Structure of the Absent in Melanoma 2 (AIM2) Pyrin Domain Provides Insights into the Mechanisms of AIM2 Autoinhibition and Inflammasome Assembly. Journal of Biological Chemistry 288, 13225-13235.

Jones, J.D.G., and Dangl, J.L. (2006). The plant immune system. Nature 444, 323- 329.

Jones, J.W., Kayagaki, N., Broz, P., Henry, T., Newton, K., O'Rourke, K., Chan, S., Dong, J., Qu, Y., Roose-Girma, M., et al. (2010). Absent in melanoma 2 is required for innate immune recognition of Francisella tularensis. Proceedings of the National Academy of Sciences of the United States of America 107, 9771-9776.

300

Jones, S., and Portnoy, D.A. (1994). Characterization of Listeria monocytogenes pathogenesis in a strain expressing perfringolysin O in place of listeriolysin O. Infection and immunity 62, 5608-5613.

Jostins, L., Ripke, S., Weersma, R.K., Duerr, R.H., McGovern, D.P., Hui, K.Y., Lee, J.C., Schumm, L.P., Sharma, Y., Anderson, C.A., et al. (2012). Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Nature 491, 119-124.

Juliana, C., Fernandes-Alnemri, T., Kang, S., Farias, A., Qin, F.S., and Alnemri, E.S. (2012). Non-transcriptional Priming and Deubiquitination Regulate NLRP3 Inflammasome Activation. Journal of Biological Chemistry 287, 36617-36622.

Kahns, S., Kalai, M., Jakobsen, L.D., Clark, B.F.C., Vandenabeele, P., and Jensen, P.H. (2003). Caspase-1 and caspase-8 cleave and inactivate cellular parkin. Journal of Biological Chemistry 278, 23376-23380.

Kamadurai, H.B., Souphron, J., Scott, D.C., Duda, D.M., Miller, D.J., Stringer, D., Piper, R.C., and Schulman, B.A. (2009). Insights into ubiquitin transfer cascades from a structure of a UbcH5B approximately ubiquitin-HECT(NEDD4L) complex. Mol Cell 36, 1095-1102.

Kamanova, J., Sun, H., Lara-Tejero, M., and Galan, J.E. (2016). The Salmonella Effector Protein SopA Modulates Innate Immune Responses by Targeting TRIM E3 Ligase Family Members. PLOS Pathogens 12.

Karmakar, M., Katsnelson, M., Malak, H.A., Greene, N.G., Howell, S.J., Hise, A.G., Camilli, A., Kadioglu, A., Dubyak, G.R., and Pearlman, E. (2015). Neutrophil IL-1β Processing Induced by Pneumolysin Is Mediated by the NLRP3/ASC Inflammasome and Caspase-1 Activation and Is Dependent on K⁺ Efflux. The Journal of Immunology 194, 1763-1775.

Karmakar, M., Katsnelson, M.A., Dubyak, G.R., and Pearlman, E. (2016). Neutrophil P2X7 receptors mediate NLRP3 inflammasome-dependent IL-1beta secretion in response to ATP. Nature communications 7, 10555.

301

Kastelic, T., Schnyder, J., Leutwiler, A., Traber, R., Streit, B., Niggli, H., MacKenzie, A., and Cheneval, D. (1996). Induction of rapid IL-1 beta mRNA degradation in THP- 1 cells mediated through the Au-rich region in the 3'UTR by a radicicol analogue. Cytokine 8, 751-761.

Kathania, M., Zeng, M.H., Yadav, V.N., Moghaddam, S.J., Yang, B.L., and Venuprasad, K. (2015). Ndfip1 Regulates Itch Ligase Activity and Airway Inflammation via UbcH7. J Immunol 194, 2160-2167.

Kattah, M.G., Malynn, B.A., and Ma, A. (2017). Ubiquitin-Modifying Enzymes and Regulation of the Inflammasome. J Mol Biol 429, 3471-3485.

Kawai, T., and Akira, S. (2007). Signaling to NF-kappaB by Toll-like receptors. Trends in molecular medicine 13, 460-469.

Kawasaki, T., and Kawai, T. (2014). Toll-like receptor signaling pathways. Frontiers in immunology 5, 461.

Kawashima, A., Karasawa, T., Tago, K., Kimura, H., Kamata, R., Usui-Kawanishi, F., Watanabe, S., Ohta, S., Funakoshi-Tago, M., Yanagisawa, K., et al. (2017). ARIH2 Ubiquitinates NLRP3 and Negatively Regulates NLRP3 Inflammasome Activation in Macrophages. J Immunol 199, 3614-3622.

Kayagaki, N., Stowe, I.B., Lee, B.L., O'Rourke, K., Anderson, K., Warming, S., Cuellar, T., Haley, B., Roose-Girma, M., Phung, Q.T., et al. (2015). Caspase-11 cleaves gasdermin D for non-canonical inflammasome signalling. Nature 526, 666- 671.

Kayagaki, N., Warming, S., Lamkanfi, M., Vande Walle, L., Louie, S., Dong, J., Newton, K., Qu, Y., Liu, J., Heldens, S., et al. (2011). Non-canonical inflammasome activation targets caspase-11. Nature 479, 117-121.

Kayagaki, N., Wong, M.T., Stowe, I.B., Ramani, S.R., Gonzalez, L.C., Akashi- Takamura, S., Miyake, K., Zhang, J., Lee, W.P., Muszynski, A., et al. (2013). Noncanonical Inflammasome Activation by Intracellular LPS Independent of TLR4. Science 341, 1246-1249.

302

Khare, S., Dorfleutner, A., Bryan, N.B., Yun, C., Radian, A.D., de Almeida, L., Rojanasakul, Y., and Stehlik, C. (2012). An NLRP7-Containing Inflammasome Mediates Recognition of Microbial Lipopeptides in Human Macrophages. Immunity 36, 464-476.

Kim, B.H., Shenoy, A.R., Kumar, P., Das, R., Tiwari, S., and MacMicking, J.D. (2011a). A Family of IFN-gamma-Inducible 65-kD GTPases Protects Against Bacterial Infection. Science 332, 717-721.

Kim, D.W., Lenzen, G., Page, A.L., Legrain, P., Sansonetti, P.J., and Parsot, C. (2005). The Shigella flexneri effector OspG interferes with innate immune responses by targeting ubiquitin-conjugating enzymes. Proceedings of the National Academy of Sciences of the United States of America 102, 14046-14051.

Kim, H.C., and Huibregtse, J.M. (2009). Polyubiquitination by HECT E3s and the determinants of chain type specificity. Mol Cell Biol 29, 3307-3318.

Kim, S., Bauernfeind, F., Ablasser, A., Hartmann, G., Fitzgerald, K.A., Latz, E., and Hornung, V. (2010). Listeria monocytogenes is sensed by the NLRP3 and AIM2 inflammasome. European Journal of Immunology 40, 1545-1551.

Kim, W., Bennett, E.J., Huttlin, E.L., Guo, A., Li, J., Possemato, A., Sowa, M.E., Rad, R., Rush, J., Comb, M.J., et al. (2011b). Systematic and quantitative assessment of the ubiquitin-modified proteome. Mol Cell 44, 325-340.

Kirkpatrick, D.S., Hathaway, N.A., Hanna, J., Elsasser, S., Rush, J., Finley, D., King, R.W., and Gygi, S.P. (2006). Quantitative analysis of in vitro ubiquitinated cyclin B1 reveals complex chain topology (vol 8, pg 700, 2006). Nat Cell Biol 8, 896-896.

Kisselev, A.F., and Goldberg, A.L. (2001). Proteasome inhibitors: from research tools to drug candidates. Chem Biol 8, 739-758.

Kisselev, A.F., van der Linden, W.A., and Overkleeft, H.S. (2012). Proteasome Inhibitors: An Expanding Army Attacking a Unique Target. Chem Biol 19, 99-115.

Knodler, L.A., Crowley, S.M., Sham, H.P., Yang, H.J., Wrande, M., Ma, C.X., Ernst, R.K., Steele-Mortimer, O., Celli, J., and Vallance, B.A. (2014). Noncanonical

303

Inflammasome Activation of Caspase-4/Caspase-11 Mediates Epithelial Defenses against Enteric Bacterial Pathogens. Cell host & microbe 16, 249-256.

Kobayashi, F., Nishiuchi, T., Takaki, K., and Konno, H. (2018). Ubiquitin chain specificities of E6AP E3 ligase and its HECT domain. Biochem Biophys Res Commun 496, 686-692.

Kobayashi, Y., Yamamoto, K., Saido, T., Kawasaki, H., Oppenheim, J.J., and Matsushima, K. (1990). Identification of Calcium-Activated Neutral Protease as a Processing Enzyme of Human Interleukin-1-Alpha. Proceedings of the National Academy of Sciences of the United States of America 87, 5548-5552.

Kofoed, E.M., and Vance, R.E. (2011). Innate immune recognition of bacterial ligands by NAIPs determines inflammasome specificity. Nature 477, 592-595.

Komander, D., and Rape, M. (2012). The ubiquitin code. Annual review of biochemistry 81, 203-229.

Komatsu, M., Waguri, S., Ueno, T., Iwata, J., Murata, S., Tanida, I., Ezaki, J., Mizushima, N., Ohsumi, Y., Uchiyama, Y., et al. (2005). Impairment of starvation- induced and constitutive autophagy in Atg7-deficient mice. The Journal of cell biology 169, 425-434.

Kondapalli, C., Kazlauskaite, A., Zhang, N., Woodroof, H.I., Campbell, D.G., Gourlay, R., Burchell, L., Walden, H., Macartney, T.J., Deak, M., et al. (2012). PINK1 is activated by mitochondrial membrane potential depolarization and stimulates Parkin E3 ligase activity by phosphorylating Serine 65. Open Biology 2.

Kong, F.Z., Su, Z.L., Zhou, C.L., Sun, C.X., Liu, Y.F., Zheng, D., Yuan, H.Y., Yin, J.P., Fang, J., Wang, S.J., et al. (2011). Role of Positive Selection in Functional Divergence of Mammalian Neuronal Apoptosis Inhibitor Proteins during Evolution. J Biomed Biotechnol.

Kravtsova-Ivantsiv, Y., and Ciechanover, A. (2012). Non-canonical ubiquitin-based signals for proteasomal degradation. Journal of cell science 125, 539-548.

304

Kuang, E., Qi, J., and Ronai, Z. (2013). Emerging roles of E3 ubiquitin ligases in autophagy. Trends in biochemical sciences 38, 453-460.

Kuang, S., Zheng, J., Yang, H., Li, S., Duan, S., Shen, Y., Ji, C., Gan, J., Xu, X.W., and Li, J. (2017). Structure insight of GSDMD reveals the basis of GSDMD autoinhibition in cell pyroptosis. Proceedings of the National Academy of Sciences of the United States of America 114, 10642-10647.

Kuo, M.L., den Besten, W., Bertwistle, D., Roussel, M.F., and Sherr, C.J. (2004). N- terminal polyubiquitination and degradation of the Arf tumor suppressor. Genes & development 18, 1862-1874.

Labbe, K., McIntire, C.R., Doiron, K., Leblanc, P.M., and Saleh, M. (2011). Cellular inhibitors of apoptosis proteins cIAP1 and cIAP2 are required for efficient caspase-1 activation by the inflammasome. Immunity 35, 897-907.

Lamkanfi, M., Kanneganti, T.D., Van Damme, P., Vanden Berghe, T., Vanoverberghe, I., Vandekerckhove, J., Vandenabeele, P., Gevaert, K., and Nunez, G. (2008). Targeted peptidecentric proteomics reveals caspase-7 as a substrate of the caspase-1 inflammasomes. Mol Cell Proteomics 7, 2350-2363.

Lamkanfi, M., Sarkar, A., Vande Walle, L., Vitari, A.C., Amer, A.O., Wewers, M.D., Tracey, K.J., Kanneganti, T.D., and Dixit, V.M. (2010). Inflammasome-dependent release of the alarmin HMGB1 in endotoxemia. J Immunol 185, 4385-4392.

Landais, I., Lee, H., and Lu, H. (2006). Coupling caspase cleavage and ubiquitin- proteasome-dependent degradation of SSRP1 during apoptosis. Cell death and differentiation 13, 1866-1878.

Laplantine, E., Fontan, E., Chiaravalli, J., Lopez, T., Lakisic, G., Veron, M., Agou, F., and Israel, A. (2009). NEMO specifically recognizes K63-linked poly-ubiquitin chains through a new bipartite ubiquitin-binding domain. The EMBO journal 28, 2885-2895.

LaRock, C.N., and Cookson, B.T. (2012). The Yersinia virulence effector YopM binds caspase-1 to arrest inflammasome assembly and processing. Cell host & microbe 12, 799-805.

305

Latz, E., Xiao, T.S., and Stutz, A. (2013). Activation and regulation of the inflammasomes. Nature Reviews Immunology 13, 397-411.

Lazarou, M., Narendra, D.P., Jin, S.M., Tekle, E., Banerjee, S., and Youle, R.J. (2013). PINK1 drives Parkin self-association and HECT-like E3 activity upstream of mitochondrial binding. The Journal of cell biology 200, 163-172.

Lazebnik, Y.A., Kaufmann, S.H., Desnoyers, S., Poirier, G.G., and Earnshaw, W.C. (1994). Cleavage of Poly(Adp-Ribose) Polymerase by a Proteinase with Properties Like Ice. Nature 371, 346-347.

Lechtenberg, B.C., Rajput, A., Sanishvili, R., Dobaczewska, M.K., Ware, C.F., Mace, P.D., and Riedl, S.J. (2016). Structure of a HOIP/E2~ubiquitin complex reveals RBR E3 ligase mechanism and regulation. Nature 529, 546-550.

Lee, J.K., Kim, S.H., Lewis, E.C., Azam, T., Reznikov, L.L., and Dinarello, C.A. (2004). Differences in signaling pathways by IL-1 beta and IL-18. Proceedings of the National Academy of Sciences of the United States of America 101, 8815-8820.

Levin, I., Eakin, C., Blanc, M.P., Klevit, R.E., Miller, S.I., and Brzovic, P.S. (2010). Identification of an unconventional E3 binding surface on the UbcH5 ~ Ub conjugate recognized by a pathogenic bacterial E3 ligase. Proceedings of the National Academy of Sciences of the United States of America 107, 2848-2853.

Levinsohn, J.L., Newman, Z.L., Hellmich, K.A., Fattah, R., Getz, M.A., Liu, S., Sastalla, I., Leppla, S.H., and Moayeri, M. (2012). Anthrax Lethal Factor Cleavage of Nlrp1 Is Required for Activation of the Inflammasome. PLOS Pathogens 8, e1002638.

Levy, M., Shapiro, H., Thaiss, C.A., and Elinav, E. (2017). NLRP6: A Multifaceted Innate Immune Sensor. Trends in immunology 38, 248-260.

Levy, M., Thaiss, C.A., Zeevi, D., Dohnalova, L., Zilberman-Schapira, G., Mahdi, J.A., David, E., Savidor, A., Korem, T., Herzig, Y., et al. (2015). Microbiota- Modulated Metabolites Shape the Intestinal Microenvironment by Regulating NLRP6 Inflammasome Signaling. Cell 163, 1428-1443.

306

Lewis, M.J., Vyse, S., Shields, A.M., Boeltz, S., Gordon, P.A., Spector, T.D., Lehner, P.J., Walczak, H., and Vyse, T.J. (2015). UBE2L3 polymorphism amplifies NF- kappaB activation and promotes plasma cell development, linking linear ubiquitination to multiple autoimmune diseases. American journal of human genetics 96, 221-234.

Li, H., Willingham, S.B., Ting, J.P., and Re, F. (2008). Cutting edge: inflammasome activation by alum and alum's adjuvant effect are mediated by NLRP3. J Immunol 181, 17-21.

Li, J., Chai, Q.Y., and Liu, C.H. (2016). The ubiquitin system: a critical regulator of innate immunity and pathogen-host interactions. Cellular & molecular immunology 13, 560-576.

Li, X.D., and Chen, Z.J.J. (2012). Sequence specific detection of bacterial 23S ribosomal RNA by TLR13. Elife 1.

Lin, D.Y., Diao, J., and Chen, J. (2012). Crystal structures of two bacterial HECT-like E3 ligases in complex with a human E2 reveal atomic details of pathogen-host interactions. Proceedings of the National Academy of Sciences of the United States of America 109, 1925-1930.

Lin, D.Y., Diao, J., Zhou, D., and Chen, J. (2011). Biochemical and structural studies of a HECT-like ubiquitin ligase from Escherichia coli O157:H7. The Journal of biological chemistry 286, 441-449.

Lin, K.M., Hu, W., Troutman, T.D., Jennings, M., Brewer, T., Li, X.X., Nanda, S., Cohen, P., Thomas, J.A., and Pasare, C. (2014). IRAK-1 bypasses priming and directly links TLRs to rapid NLRP3 inflammasome activation Proceedings of the National Academy of Sciences of the United States of America 111, 3195-3195.

Lin, S.C., Lo, Y.C., and Wu, H. (2010). Helical assembly in the MyD88-IRAK4-IRAK2 complex in TLR/IL-1R signalling. Nature 465, 885-890.

Liu, J., Han, C., Xie, B., Wu, Y., Liu, S., Chen, K., Xia, M., Zhang, Y., Song, L., Li, Z., et al. (2014a). Rhbdd3 controls autoimmunity by suppressing the production of IL-6

307 by dendritic cells via K27-linked ubiquitination of the regulator NEMO. Nat Immunol 15, 612-622.

Liu, S.H., Moayeri, M., and Leppla, S.H. (2014b). Anthrax lethal and edema toxins in anthrax pathogenesis. Trends Microbiol 22, 317-325.

Liu, T., Tang, Q., Liu, K.P., Xie, W.H., Liu, X., Wang, H.S., Wang, R.F., and Cui, J. (2016a). TRIM11 Suppresses AIM2 Inflammasome by Degrading AIM2 via p62- Dependent Selective Autophagy. Cell Reports 16, 1988-2002.

Liu, X., Zhang, Z., Ruan, J., Pan, Y., Magupalli, V.G., Wu, H., and Lieberman, J. (2016b). Inflammasome-activated gasdermin D causes pyroptosis by forming membrane pores. Nature 535, 153-158.

Liu, X., Zhao, B., Sun, L., Bhuripanyo, K., Wang, Y., Bi, Y., Davuluri, R.V., Duong, D.M., Nanavati, D., Yin, J., et al. (2017). Orthogonal ubiquitin transfer identifies ubiquitination substrates under differential control by the two ubiquitin activating enzymes. Nature communications 8, 14286.

Locke, M., Toth, J.I., and Petroski, M.D. (2014). Lys11- and Lys48-linked ubiquitin chains interact with p97 during endoplasmic-reticulum-associated degradation. The Biochemical journal 459, 205-216.

Lopez-Castejon, G., Luheshi, N.M., Compan, V., High, S., Whitehead, R.C., Flitsch, S., Kirov, A., Prudovsky, I., Swanton, E., and Brough, D. (2013). Deubiquitinases Regulate the Activity of Caspase-1 and Interleukin-1 beta Secretion via Assembly of the Inflammasome. Journal of Biological Chemistry 288, 2721-2733.

Lotze, M.T., and Tracey, K.J. (2005). High-mobility group box 1 protein (HMGB1): nuclear weapon in the immune arsenal. Nature reviews. Immunology 5, 331-342.

Lu, A., Li, Y., Yin, Q., Ruan, J.B., Yu, X., Egelman, E., and Wu, H. (2015). Plasticity in PYD assembly revealed by cryo-EM structure of the PYD filament of AIM2. Cell Discov 1.

Lu, A., Magupalli, V.G., Ruan, J., Yin, Q., Atianand, M.K., Vos, M.R., Schroder, G.F., Fitzgerald, K.A., Wu, H., and Egelman, E.H. (2014). Unified polymerization

308 mechanism for the assembly of ASC-dependent inflammasomes. Cell 156, 1193- 1206.

Ludwig, A., Friedel, B., Metzkow, S., Meiners, S., Stangl, V., Baumann, G., and Stangl, K. (2005). Effect of statins on the proteasomal activity in mammalian endothelial and vascular smooth muscle cells. Biochem Pharmacol 70, 520-526.

Luo, X., Wasilko, D.J., Liu, Y., Sun, J., Wu, X., Luo, Z.-Q., and Mao, Y. (2015). Structure of the Legionella Virulence Factor, SidC Reveals a Unique PI(4)P-Specific Binding Domain Essential for Its Targeting to the Bacterial Phagosome. PLOS Pathogens 11, e1004965.

Machida, Y.J., Machida, Y., Chen, Y.F., 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. Mol Cell 23, 589-596.

Maculins, T., Fiskin, E., Bhogaraju, S., and Dikic, I. (2016). Bacteria-host relationship: ubiquitin ligases as weapons of invasion. Cell research 26, 499-510.

Mahrus, S., Trinidad, J.C., Barkan, D.T., Sali, A., Burlingame, A.L., and Wells, J.A. (2008). Global sequencing of proteolytic cleavage sites in apoptosis by specific labeling of protein N termini. Cell 134, 866-876.

Malinowsky, D., Lundkvist, J., Laye, S., and Bartfai, T. (1998). Interleukin-1 receptor accessory protein interacts with the type II interleukin-1 receptor. FEBS letters 429, 299-302.

Malireddi, R.K., Ippagunta, S., Lamkanfi, M., and Kanneganti, T.D. (2010). Cutting edge: proteolytic inactivation of poly(ADP-ribose) polymerase 1 by the Nlrp3 and Nlrc4 inflammasomes. J Immunol 185, 3127-3130.

Man, S.M., Hopkins, L.J., Nugent, E., Cox, S., Gluck, I.M., Tourlomousis, P., Wright, J.A., Cicuta, P., Monie, T.P., and Bryant, C.E. (2014). Inflammasome activation causes dual recruitment of NLRC4 and NLRP3 to the same macromolecular complex. Proceedings of the National Academy of Sciences of the United States of America 111, 7403-7408.

309

Man, S.M., Karki, R., Malireddi, R.K.S., Neale, G., Vogel, P., Yamamoto, M., Lamkanfi, M., and Kanneganti, T.D. (2015). The transcription factor IRF1 and guanylate-binding proteins target activation of the AIM2 inflammasome by Francisella infection. Nature immunology 16, 467-U174.

Man, S.M., Tourlomousis, P., Hopkins, L., Monie, T.P., Fitzgerald, K.A., and Bryant, C.E. (2013). Salmonella Infection Induces Recruitment of Caspase-8 to the Inflammasome To Modulate IL-1 beta Production. J Immunol 191, 5239-5246.

Manzanillo, P.S., Ayres, J.S., Watson, R.O., Collins, A.C., Souza, G., Rae, C.S., Schneider, D.S., Nakamura, K., Shiloh, M.U., and Cox, J.S. (2013). The ubiquitin ligase parkin mediates resistance to intracellular pathogens. Nature 501, 512-+.

March, C.J., Mosley, B., Larsen, A., Cerretti, D.P., Braedt, G., Price, V., Gillis, S., Henney, C.S., Kronheim, S.R., Grabstein, K., et al. (1985). Cloning, sequence and expression of two distinct human interleukin-1 complementary DNAs. Nature 315, 641-647.

Mariathasan, S., Newton, K., Monack, D.M., Vucic, D., French, D.M., Lee, W.P., Roose-Girma, M., Erickson, S., and Dixit, V.M. (2004). Differential activation of the inflammasome by caspase-1 adaptors ASC and Ipaf. Nature 430, 213-218.

Mariathasan, S., Weiss, D.S., Newton, K., McBride, J., O'Rourke, K., Roose-Girma, M., Lee, W.P., Weinrauch, Y., Monack, D.M., and Dixit, V.M. (2006). Cryopyrin activates the inflammasome in response to toxins and ATP. Nature 440, 228-232.

Marshall, J.D., Aste-Amezaga, M., Chehimi, S.S., Murphy, M., Olsen, H., and Trinchieri, G. (1999). Regulation of human IL-18 mRNA expression. Clin Immunol 90, 15-21.

Martin-Sanchez, F., Diamond, C., Zeitler, M., Gomez, A.I., Baroja-Mazo, A., Bagnall, J., Spiller, D., White, M., Daniels, M.J.D., Mortellaro, A., et al. (2016). Inflammasome-dependent IL-1 beta release depends upon membrane permeabilisation. Cell death and differentiation 23, 1219-1231.

Martin, S.J. (2016). Cell death and inflammation: the case for IL-1 family cytokines as the canonical DAMPs of the immune system. Febs J 283, 2599-2615.

310

Martino, L., Brown, N.R., Masino, L., Esposito, D., and Rittinger, K. (2018). Determinants of E2-ubiquitin conjugate recognition by RBR E3 ligases. Scientific reports 8, 68.

Martinon, F., Burns, K., and Tschopp, J. (2002). The inflammasome: a molecular platform triggering activation of inflammatory caspases and processing of proIL-beta. Mol Cell 10, 417-426.

Martinon, F., Petrilli, V., Mayor, A., Tardivel, A., and Tschopp, J. (2006). Gout- associated uric acid crystals activate the NALP3 inflammasome. Nature 440, 237- 241.

Matsumoto, M.L., Wickliffe, K.E., Dong, K.C., Yu, C., Bosanac, I., Bustos, D., Phu, L., Kirkpatrick, D.S., Hymowitz, S.G., Rape, M., et al. (2010). K11-linked polyubiquitination in cell cycle control revealed by a K11 linkage-specific antibody. Mol Cell 39, 477-484.

Matusiak, M., Van Opdenbosch, N., Vande Walle, L., Sirard, J.C., Kanneganti, T.D., and Lamkanfi, M. (2015). Flagellin-induced NLRC4 phosphorylation primes the inflammasome for activation by NAIP5. Proceedings of the National Academy of Sciences of the United States of America 112, 1541-1546.

McDowell, G.S., Kucerova, R., and Philpott, A. (2010). Non-canonical ubiquitylation of the proneural protein Ngn2 occurs in both Xenopus embryos and mammalian cells. Biochem Biophys Res Commun 400, 655-660.

Mehta, V.B., Hart, J., and Wewers, M.D. (2001). ATP-stimulated release of interleukin (IL)-1 beta and IL-18 requires priming by lipopolysaccharide and is independent of caspase-1 cleavage. Journal of Biological Chemistry 276, 3820- 3826.

Meng, L.H., Mohan, R., Kwok, B.H.B., Elofsson, M., Sin, N., and Crews, C.M. (1999). Epoxomicin, a potent and selective proteasome inhibitor, exhibits in vivo antiinflammatory activity. Proceedings of the National Academy of Sciences of the United States of America 96, 10403-10408.

311

Mertins, P., Qiao, J.W., Patel, J., Udeshi, N.D., Clauser, K.R., Mani, D.R., Burgess, M.W., Gillette, M.A., Jaffe, J.D., and Carr, S.A. (2013). Integrated proteomic analysis of post-translational modifications by serial enrichment. Nat Methods 10, 634-637.

Metzger, M.B., Pruneda, J.N., Klevit, R.E., and Weissman, A.M. (2014). RING-type E3 ligases: master manipulators of E2 ubiquitin-conjugating enzymes and ubiquitination. Biochimica et biophysica acta 1843, 47-60.

Meunier, E., Dick, M.S., Dreier, R.F., Schurmann, N., Broz, D.K., Warming, S., Roose-Girma, M., Bumann, D., Kayagaki, N., Takeda, K., et al. (2014). Caspase-11 activation requires lysis of pathogen-containing vacuoles by IFN-induced GTPases. Nature 509, 366-+.

Meunier, E., Wallet, P., Dreier, R.F., Costanzo, S., Anton, L., Ruhl, S., Dussurgey, S., Dick, M.S., Kistner, A., Rigard, M., et al. (2015). Guanylate-binding proteins promote activation of the AIM2 inflammasome during infection with Francisella novicida. Nature immunology 16, 476-U185.

Mevissen, T.E.T., and Komander, D. (2017). Mechanisms of Deubiquitinase Specificity and Regulation. Annual Review of Biochemistry, Vol 86 86, 159-192.

Miao, E.A., Leaf, I.A., Treuting, P.M., Mao, D.P., Dors, M., Sarkar, A., Warren, S.E., Wewers, M.D., and Aderem, A. (2010a). Caspase-1-induced pyroptosis is an innate immune effector mechanism against intracellular bacteria. Nature immunology 11, 1136-U1194.

Miao, E.A., Mao, D.P., Yudkovsky, N., Bonneau, R., Lorang, C.G., Warren, S.E., Leaf, I.A., and Aderem, A. (2010b). Innate immune detection of the type III secretion apparatus through the NLRC4 inflammasome. Proceedings of the National Academy of Sciences of the United States of America 107, 3076-3080.

Middleton, A.J., and Day, C.L. (2015). The molecular basis of lysine 48 ubiquitin chain synthesis by Ube2K. Scientific reports 5, 16793.

Mizushima, N., Yoshimori, T., and Ohsumi, Y. (2011). The role of Atg proteins in autophagosome formation. Annual review of cell and developmental biology 27, 107- 132.

312

Mogensen, T.H. (2009). Pathogen recognition and inflammatory signaling in innate immune defenses. Clinical microbiology reviews 22, 240-273, Table of Contents.

Moore, F.E., Osmundson, E.C., Koblinski, J., Pugacheva, E., Golemis, E.A., Ray, D., and Kiyokawa, H. (2010). The WW-HECT protein Smurf2 interacts with the Docking Protein NEDD9/HEF1 for Aurora A activation. Cell Div 5.

Moors, M.A., and Mizel, S.B. (2000). Proteasome-mediated regulation of interleukin- 1beta turnover and export in human monocytes. Journal of leukocyte biology 68, 131-136.

Moradian, M.M., Babikyan, D., Banoian, D., Hayrapetyan, H., Manvelyan, H., Avanesian, N., and Sarkisian, T. (2017). Comprehensive analysis of mutations in the MEFV gene reveal that the location and not the substitution type determines symptom severity in FMF. Mol Genet Genom Med 5, 742-750.

Mosley, B., Dower, S.K., Gillis, S., and Cosman, D. (1987a). Determination of the minimum polypeptide lengths of the functionally active sites of human interleukins 1 alpha and 1 beta. Proceedings of the National Academy of Sciences of the United States of America 84, 4572-4576.

Mosley, B., Urdal, D.L., Prickett, K.S., Larsen, A., Cosman, D., Conlon, P.J., Gillis, S., and Dower, S.K. (1987b). The interleukin-1 receptor binds the human interleukin- 1 alpha precursor but not the interleukin-1 beta precursor. The Journal of biological chemistry 262, 2941-2944.

Moynagh, P.N. (2014). The roles of Pellino E3 ubiquitin ligases in immunity. Nature Reviews Immunology 14, 122-131.

Moynihan, T.P., Ardley, H.C., Nuber, U., Rose, S.A., Jones, P.F., Markham, A.F., Scheffner, M., and Robinson, P.A. (1999). The ubiquitin-conjugating enzymes UbcH7 and UbcH8 interact with RING finger/IBR motif-containing domains of HHARI and H7-AP1. The Journal of biological chemistry 274, 30963-30968.

Mulder, M.P.C., Witting, K., Berlin, I., Pruneda, J.N., Wu, K.P., Chang, J.G., Merkx, R., Bialas, J., Groettrup, M., Vertegaal, A.C.O., et al. (2016). A cascading activity-

313 based probe sequentially targets E1-E2-E3 ubiquitin enzymes. Nat Chem Biol 12, 523-+.

Munoz-Planillo, R., Franchi, L., Miller, L.S., and Nunez, G. (2009). A Critical Role for Hemolysins and Bacterial Lipoproteins in Staphylococcus aureus-Induced Activation of the Nlrp3 Inflammasome. J Immunol 183, 3942-3948.

Munoz-Planillo, R., Kuffa, P., Martinez-Colon, G., Smith, B.L., Rajendiran, T.M., and Nunez, G. (2013). K+ Efflux Is the Common Trigger of NLRP3 Inflammasome Activation by Bacterial Toxins and Particulate Matter. Immunity 38, 1142-1153.

Muñoz-Planillo, R., Kuffa, P., Martínez-Colón, G., Smith, Brenna L., Rajendiran, Thekkelnaycke M., and Núñez, G. K+ Efflux Is the Common Trigger of NLRP3 Inflammasome Activation by Bacterial Toxins and Particulate Matter. Immunity 38, 1142-1153.

Murphy, K., Travers, P., Walport, M., and Janeway, C. (2012). Janeway's immunobiology, 8th edn (New York: Garland Science).

Murray, S.S., Tu, K.N., Young, K.L., and Murray, E.J.B. (2002). The effects of Lovastatin on proteasome activities in highly purified rabbit 20 S proteasome preparations and mouse MC3T3-E1 osteoblastic cells. Metabolism 51, 1153-1160.

Na, C.H., and Peng, J. (2012). Analysis of ubiquitinated proteome by quantitative mass spectrometry. Methods in molecular biology 893, 417-429.

Neven, B., Callebaut, I., Prieur, A.M., Feldmann, J., Bodemer, C., Lepore, L., Derfalvi, B., Benjaponpitak, S., Vesely, R., Sauvain, M.J., et al. (2004). Molecular basis of the spectral expression of CIAS1 mutations associated with phagocytic cell- mediated autoinflammatory disorders CINCA/NOMID, MWS, and FCU. Blood 103, 2809-2815.

Neven, B., Marvillet, I., Terrada, C., Ferster, A., Boddaert, N., Couloignier, V., Pinto, G., Pagnier, A., Bodemer, C., Bodaghi, B., et al. (2010). Long-term efficacy of the interleukin-1 receptor antagonist anakinra in ten patients with neonatal-onset multisystem inflammatory disease/chronic infantile neurologic, cutaneous, articular syndrome. Arthritis and rheumatism 62, 258-267.

314

Newton, K., and Dixit, V.M. (2012). Signaling in Innate Immunity and Inflammation. Csh Perspect Biol 4.

Ng, J., Hirota, S.A., Gross, O., Li, Y., Ulke-Lemee, A., Potentier, M.S., Schenck, L.P., Vilaysane, A., Seamone, M.E., Feng, H.P., et al. (2010). Clostridium difficile Toxin- Induced Inflammation and Intestinal Injury Are Mediated by the Inflammasome. Gastroenterology 139, 542-552.

Niebler, M., Qian, X., Hofler, D., Kogosov, V., Kaewprag, J., Kaufmann, A.M., Ly, R., Bohmer, G., Zawatzky, R., Rosl, F., et al. (2013). Post-translational control of IL- 1beta via the human papillomavirus type 16 E6 oncoprotein: a novel mechanism of innate immune escape mediated by the E3-ubiquitin ligase E6-AP and p53. PLoS Pathog 9, e1003536.

Niu, Z., Shi, Q., Zhang, W., Shu, Y., Yang, N., Chen, B., Wang, Q., Zhao, X., Chen, J., Cheng, N., et al. (2017). Caspase-1 cleaves PPARgamma for potentiating the pro-tumor action of TAMs. Nature communications 8, 766.

Noy, T., Suad, O., Taglicht, D., and Ciechanover, A. (2012). HUWE1 ubiquitinates MyoD and targets it for proteasomal degradation. Biochem Bioph Res Co 418, 408- 413.

Nyman, T., Stenmark, P., Flodin, S., Johansson, I., Hammarstrom, M., and Nordlund, P. (2008). The crystal structure of the human toll-like receptor 10 cytoplasmic domain reveals a putative signaling dimer. The Journal of biological chemistry 283, 11861-11865.

O'Neill, L.A., Golenbock, D., and Bowie, A.G. (2013). The history of Toll-like receptors - redefining innate immunity. Nature reviews. Immunology 13, 453-460.

Ogunjimi, A.A., Briant, D.J., Pece-Barbara, N., Le Roy, C., Di Guglielmo, G.M., Kavsak, P., Rasmussen, R.K., Seet, B.T., Sicheri, F., and Wrana, J.L. (2005). Regulation of Smurf2 ubiquitin ligase activity by anchoring the E2 to the HECT domain. Mol Cell 19, 297-308.

315

Omasits, U., Ahrens, C.H., Muller, S., and Wollscheid, B. (2014). Protter: interactive protein feature visualization and integration with experimental proteomic data. Bioinformatics 30, 884-886.

Orozco, G., Eyre, S., Hinks, A., Bowes, J., Morgan, A.W., Wilson, A.G., Wordsworth, P., Steer, S., Hocking, L., consortium, U., et al. (2011). Study of the common genetic background for rheumatoid arthritis and systemic lupus erythematosus. Annals of the rheumatic diseases 70, 463-468.

Osmundson, E.C., Ray, D., Moore, F.E., Gao, Q.S., Thomsen, G.H., and Kiyokawa, H. (2008). The HECT E3 ligase Smurf2 is required for Mad2-dependent spindle assembly checkpoint. Journal of Cell Biology 183, 267-277.

Pallett, M.A., Crepin, V.F., Serafini, N., Habibzay, M., Kotik, O., Sanchez-Garrido, J., Di Santo, J.P., Shenoy, A.R., Berger, C.N., and Frankel, G. (2016). Bacterial virulence factor inhibits caspase-4/11 activation in intestinal epithelial cells. Mucosal immunology.

Park, Y.H., Wood, G., Kastner, D.L., and Chae, J.J. (2016). Pyrin inflammasome activation and RhoA signaling in the autoinflammatory diseases FMF and HIDS. Nature immunology 17, 914-+.

Parkin, J., and Cohen, B. (2001). An overview of the immune system. Lancet 357, 1777-1789.

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. Journal of Biological Chemistry 282, 23010-23014.

Perregaux, D., and Gabel, C.A. (1994). Interleukin-1-Beta Maturation and Release in Response to Atp and Nigericin - Evidence That Potassium-Depletion Mediated by These Agents Is a Necessary and Common Feature of Their Activity. Journal of Biological Chemistry 269, 15195-15203.

Pickart, C.M., and Rose, I.A. (1985). Functional-Heterogeneity of Ubiquitin Carrier Proteins. Journal of Biological Chemistry 260, 1573-1581.

316

Piscatelli, H., Kotkar, S.A., McBee, M.E., Muthupalani, S., Schauer, D.B., Mandrell, R.E., Leong, J.M., and Zhou, D.G. (2011). The EHEC Type III Effector NleL Is an E3 Ubiquitin Ligase That Modulates Pedestal Formation. PloS one 6.

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.

Poltorak, A., He, X.L., Smirnova, I., Liu, M.Y., Van Huffel, C., Du, X., Birdwell, D., Alejos, E., Silva, M., Galanos, C., et al. (1998). Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: Mutations in Tlr4 gene. Science 282, 2085-2088.

Poyet, J.L., Srinivasula, S.M., Tnani, M., Razmara, M., Fernandes-Alnemri, T., and Alnemri, E.S. (2001). Identification of Ipaf, a human caspase-1-activating protein related to Apaf-1. Journal of Biological Chemistry 276, 28309-28313.

Proell, M., Gerlic, M., Mace, P.D., Reed, J.C., and Riedl, S.J. (2013). The CARD plays a critical role in ASC foci formation and inflammasome signalling. The Biochemical journal 449, 613-621.

Pruneda, J.N., Littlefield, P.J., Soss, S.E., Nordquist, K.A., Chazin, W.J., Brzovic, P.S., and Klevit, R.E. (2012). Structure of an E3:E2~Ub complex reveals an allosteric mechanism shared among RING/U-box ligases. Mol Cell 47, 933-942.

Pruneda, J.N., Smith, F.D., Daurie, A., Swaney, D.L., Villen, J., Scott, J.D., Stadnyk, A.W., Le Trong, I., Stenkamp, R.E., Klevit, R.E., et al. (2014). E2~Ub conjugates regulate the kinase activity of Shigella effector OspG during pathogenesis. The EMBO journal 33, 437-449.

Pruneda, J.N., Stoll, K.E., Bolton, L.J., Brzovic, P.S., and Klevit, R.E. (2011). Ubiquitin in motion: structural studies of the ubiquitin-conjugating enzyme approximately ubiquitin conjugate. Biochemistry-Us 50, 1624-1633.

Puren, A.J., Fantuzzi, G., and Dinarello, C.A. (1999). Gene expression, synthesis, and secretion of interleukin 18 and interleukin 1 beta are differentially regulated in human blood mononuclear cells and mouse spleen cells. Proceedings of the National Academy of Sciences of the United States of America 96, 2256-2261.

317

Py, B.F., Kim, M.S., Vakifahmetoglu-Norberg, H., and Yuan, J. (2013). Deubiquitination of NLRP3 by BRCC3 critically regulates inflammasome activity. Mol Cell 49, 331-338.

Qu, Y., Franchi, L., Nunez, G., and Dubyak, G.R. (2007). Nonclassical IL-1 beta secretion stimulated by P2X7 receptors is dependent on inflammasome activation and correlated with exosome release in murine macrophages. J Immunol 179, 1913- 1925.

Qu, Y., Misaghi, S., Izrael-Tomasevic, A., Newton, K., Gilmour, L.L., Lamkanfi, M., Louie, S., Kayagaki, N., Liu, J.F., Komuves, L., et al. (2012). Phosphorylation of NLRC4 is critical for inflammasome activation. Nature 490, 539-+.

Qu, Y., Misaghi, S., Newton, K., Maltzman, A., Izrael-Tomasevic, A., Arnott, D., and Dixit, V.M. (2016). NLRP3 recruitment by NLRC4 during Salmonella infection. J Exp Med 213, 877-885.

Raetz, M., Kibardin, A., Sturge, C.R., Pifer, R., Li, H.Y., Burstein, E., Ozato, K., Larin, S., and Yarovinsky, F. (2013). Cooperation of TLR12 and TLR11 in the IRF8- Dependent IL-12 Response to Toxoplasma gondii Profilin. J Immunol 191, 4818- 4827.

Rajalingam, K., and Dikic, I. (2016). SnapShot: Expanding the Ubiquitin Code. Cell 164, 1074-1074 e1071.

Rajsbaum, R., and Bharaj, P. (2016). Unanchored Polyubiquitin Chains and Innate Immune Sensing of Influenza Virus Infection. J Immunol 196.

Rajsbaum, R., Versteeg, G.A., Schmid, S., Maestre, A.M., Belicha-Villanueva, A., Martinez-Romero, C., Patel, J.R., Morrison, J., Pisanelli, G., Miorin, L., et al. (2014). Unanchored K48-Linked Polyubiquitin Synthesized by the E3-Ubiquitin Ligase TRIM6 Stimulates the Interferon-IKK epsilon Kinase-Mediated Antiviral Response. Immunity 40, 880-895.

Rao, S., Porter, D.C., Chen, X., Herliczek, T., Lowe, M., and Keyomarsi, K. (1999). Lovastatin-mediated G1 arrest is through inhibition of the proteasome, independent

318 of hydroxymethyl glutaryl-CoA reductase. Proceedings of the National Academy of Sciences 96, 7797-7802.

Rathinam, V.A.K., Jiang, Z.Z., Waggoner, S.N., Sharma, S., Cole, L.E., Waggoner, L., Vanaja, S.K., Monks, B.G., Ganesan, S., Latz, E., et al. (2010). The AIM2 inflammasome is essential for host defense against cytosolic bacteria and DNA viruses. Nature immunology 11, 395-403.

Rathinam, V.A.K., Vanaja, S.K., Waggoner, L., Sokolovska, A., Becker, C., Stuart, L.M., Leong, J.M., and Fitzgerald, K.A. (2012). TRIF Licenses Caspase-11- Dependent NLRP3 Inflammasome Activation by Gram-Negative Bacteria. Cell 150, 606-619.

Ratner, D., Orning, M.P., Proulx, M.K., Wang, D., Gavrilin, M.A., Wewers, M.D., Alnemri, E.S., Johnson, P.F., Lee, B., Mecsas, J., et al. (2016). The Yersinia pestis Effector YopM Inhibits Pyrin Inflammasome Activation. PLoS Pathog 12, e1006035.

Raupach, B., Peuschel, S.K., Monack, D.M., and Zychlinsky, A. (2006). Caspase-1- mediated activation of interleukin-1 beta (IL-1 beta) and IL-18 contributes to innate immune defenses against Salmonella enterica serovar typhimurium infection. Infection and immunity 74, 4922-4926.

Ravid, T., and Hochstrasser, M. (2007). Autoregulation of an E2 enzyme by ubiquitin-chain assembly on its catalytic residue. Nat Cell Biol 9, 422-427.

Rawlings, N.D., Barrett, A.J., and Finn, R. (2016). Twenty years of the MEROPS database of proteolytic enzymes, their substrates and inhibitors. Nucleic Acids Res 44, D343-D350.

Ray, D., Ahsan, A., Helman, A., Chen, G.A., Hegde, A., Gurjar, S.R., Zhao, L.L., Kiyokawa, H., Beer, D.G., Lawrence, T.S., et al. (2011). Regulation of EGFR Protein Stability by the HECT-type Ubiquitin Ligase SMURF2. Neoplasia 13, 570-U512.

Reyes Ruiz, V.M., Ramirez, J., Naseer, N., Palacio, N.M., Siddarthan, I.J., Yan, B.M., Boyer, M.A., Pensinger, D.A., Sauer, J.D., and Shin, S. (2017). Broad detection of bacterial type III secretion system and flagellin proteins by the human

319

NAIP/NLRC4 inflammasome. Proceedings of the National Academy of Sciences of the United States of America 114, 13242-13247.

Rittinger, K., and Ikeda, F. (2017). Linear ubiquitin chains: enzymes, mechanisms and biology. Open Biol 7.

Roberts, T.L., Idris, A., Dunn, J.A., Kelly, G.M., Burnton, C.M., Hodgson, S., Hardy, L.L., Garceau, V., Sweet, M.J., Ross, I.L., et al. (2009). HIN-200 Proteins Regulate Caspase Activation in Response to Foreign Cytoplasmic DNA. Science 323, 1057- 1060.

Rodgers, M.A., Bowman, J.W., Fujita, H., Orazio, N., Shi, M., Liang, Q., Amatya, R., Kelly, T.J., Iwai, K., Ting, J., et al. (2014). The linear ubiquitin assembly complex (LUBAC) is essential for NLRP3 inflammasome activation. The Journal of experimental medicine 211, 1333-1347.

Romberg, N., Al Moussawi, K., Nelson-Williams, C., Stiegler, A.L., Loring, E., Choi, M., Overton, J., Meffre, E., Khokha, M.K., Huttner, A.J., et al. (2014). Mutation of NLRC4 causes a syndrome of enterocolitis and autoinflammation. Nat Genet 46, 1135-1139.

Ronchi, V.P., Kim, E.D., Summa, C.M., Klein, J.M., and Haas, A.L. (2017). In silico modeling of the cryptic E2 approximately ubiquitin-binding site of E6-associated protein (E6AP)/UBE3A reveals the mechanism of polyubiquitin chain assembly. The Journal of biological chemistry 292, 18006-18023.

Ronchi, V.P., Klein, J.M., and Haas, A.L. (2013). E6AP/UBE3A ubiquitin ligase harbors two E2~ubiquitin binding sites. The Journal of biological chemistry 288, 10349-10360.

Rosen-Wolff, A., Quietzsch, J., Schroder, H., Lehmann, R., Gahr, M., and Roesler, J. (2003). Two German CINCA (NOMID) patients with different clinical severity and response to anti-inflammatory treatment. European journal of haematology 71, 215- 219.

Rotin, D., and Kumar, S. (2009). Physiological functions of the HECT family of ubiquitin ligases. Nat Rev Mol Cell Bio 10, 398-409.

320

Rubartelli, A., Cozzolino, F., Talio, M., and Sitia, R. (1990). A novel secretory pathway for interleukin-1 beta, a protein lacking a signal sequence. The EMBO journal 9, 1503-1510.

Ruhl, S., and Broz, P. (2015). Caspase-11 activates a canonical NLRP3 inflammasome by promoting K+ efflux. European Journal of Immunology 45, 2927- 2936.

Saeki, Y., Kudo, T., Sone, T., Kikuchi, Y., Yokosawa, H., Toh-e, A., and Tanaka, K. (2009). Lysine 63-linked polyubiquitin chain may serve as a targeting signal for the 26S proteasome. The EMBO journal 28, 359-371.

Sagulenko, V., Vitak, N., Vajjhala, P.R., Vince, J.E., and Stacey, K.J. (2018). Caspase-1 Is an Apical Caspase Leading to Caspase-3 Cleavage in the AIM2 Inflammasome Response, Independent of Caspase-8. J Mol Biol 430, 238-247.

Sansonetti, P.J., Phalipon, A., Arondel, J., Thirumalai, K., Banerjee, S., Akira, S., Takeda, K., and Zychlinsky, A. (2000). Caspase-1 activation of IL-1 beta and IL-18 are essential for Shigella flexneri-induced inflammation. Immunity 12, 581-590.

Sarraf, S.A., Raman, M., Guarani-Pereira, V., Sowa, M.E., Huttlin, E.L., Gygi, S.P., and Harper, J.W. (2013). Landscape of the PARKIN-dependent ubiquitylome in response to mitochondrial depolarization. Nature 496, 372-+.

Sarrauste de Menthiere, C., Terriere, S., Pugnere, D., Ruiz, M., Demaille, J., and Touitou, I. (2003). INFEVERS: the Registry for FMF and hereditary inflammatory disorders mutations. Nucleic Acids Res 31, 282-285.

Savva, A., and Roger, T. (2013). Targeting toll-like receptors: promising therapeutic strategies for the management of sepsis-associated pathology and infectious diseases. Frontiers in immunology 4, 387.

Sborgi, L., Ravotti, F., Dandey, V.P., Dick, M.S., Mazur, A., Reckel, S., Chami, M., Scherer, S., Huber, M., Bockmann, A., et al. (2015). Structure and assembly of the mouse ASC inflammasome by combined NMR spectroscopy and cryo-electron microscopy. Proceedings of the National Academy of Sciences of the United States of America 112, 13237-13242.

321

Scaglione, K.M., Basrur, V., Ashraf, N.S., Konen, J.R., Elenitoba-Johnson, K.S., Todi, S.V., and Paulson, H.L. (2013). The ubiquitin-conjugating enzyme (E2) Ube2w ubiquitinates the N terminus of substrates. The Journal of biological chemistry 288, 18784-18788.

Scheer, J.M., Wells, J.A., and Romanowski, M.J. (2005). Malonate-assisted purification of human caspases. Protein Expres Purif 41, 148-153.

Schembri, L., Dalibart, R., Tomasello, F., Legembre, P., Ichas, F., and De Giorgi, F. (2007). The HA tag is cleaved and loses immunoreactivity during apoptosis. Nat Methods 4, 107-108.

Schmid-Burgk, J.L., Chauhan, D., Schmidt, T., Ebert, T.S., Reinhardt, J., Endl, E., and Hornung, V. (2016). A Genome-wide CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) Screen Identifies NEK7 as an Essential Component of NLRP3 Inflammasome Activation. Journal of Biological Chemistry 291, 103-109.

Schrader, E.K., Harstad, K.G., and Matouschek, A. (2009). Targeting proteins for degradation. Nat Chem Biol 5, 815-822.

Schroder, K., Sagulenko, V., Zamoshnikova, A., Richards, A.A., Cridland, J.A., Irvine, K.M., Stacey, K.J., and Sweet, M.J. (2012). Acute lipopolysaccharide priming boosts inflammasome activation independently of inflammasome sensor induction. Immunobiology 217, 1325-1329.

Schumacher, F.R., Wilson, G., and Day, C.L. (2013). The N-Terminal Extension of UBE2E Ubiquitin-Conjugating Enzymes Limits Chain Assembly. J Mol Biol 425, 4099-4111.

Schwandner, R., Dziarski, R., Wesche, H., Rothe, M., and Kirschning, C.J. (1999). Peptidoglycan- and lipoteichoic acid-induced cell activation is mediated by toll-like receptor 2. Journal of Biological Chemistry 274, 17406-17409.

Sears, N., Sen, G.C., Stark, G.R., and Chattopadhyay, S. (2011). Caspase-8- mediated Cleavage Inhibits IRF-3 Protein by Facilitating Its Proteasome-mediated Degradation. Journal of Biological Chemistry 286, 33037-33044.

322

Sellin, M.E., Muller, A.A., Felmy, B., Dolowschiak, T., Diard, M., Tardivel, A., Maslowski, K.M., and Hardt, W.D. (2014). Epithelium-Intrinsic NAIP/NLRC4 Inflammasome Drives Infected Enterocyte Expulsion to Restrict Salmonella Replication in the Intestinal Mucosa. Cell host & microbe 16, 237-248.

Semple, J.I., Smits, V.A.J., Fernaud, J.R., Mamely, I., and Freire, R. (2007). Cleavage and degradation of Claspin during apoptosis by caspases and the proteasome. Cell death and differentiation 14, 1433-1442.

Seo, S.U., Kamada, N., Munoz-Planillo, R., Kim, Y.G., Kim, D., Koizumi, Y., Hasegawa, M., Himpsl, S.D., Browne, H.P., Lawley, T.D., et al. (2015). Distinct Commensals Induce Interleukin-1 beta via NLRP3 Inflammasome in Inflammatory Monocytes to Promote Intestinal Inflammation in Response to Injury. Immunity 42, 744-755.

Seshadri, S., Duncan, M.D., Hart, J.M., Gavrilin, M.A., and Wewers, M.D. (2007). Pyrin levels in human monocytes and monocyte-derived macrophages regulate IL- 1beta processing and release. J Immunol 179, 1274-1281.

Shabek, N., Herman-Bachinsky, Y., and Ciechanover, A. (2009). Ubiquitin degradation with its substrate, or as a monomer in a ubiquitination-independent mode, provides clues to proteasome regulation. Proceedings of the National Academy of Sciences of the United States of America 106, 11907-11912.

Shao, W., Yeretssian, G., Doiron, K., Hussain, S.N., and Saleh, M. (2007). The caspase-1 digestome identifies the glycolysis pathway as a target during infection and septic shock. The Journal of biological chemistry 282, 36321-36329.

Sharma, S., and Zhu, J. (2014). Immunologic Applications of Conditional Gene Modification Technology in the Mouse. Current protocols in immunology / edited by John E. Coligan ... [et al.] 105, 10.34.11-10.34.13.

Shembade, N., Ma, A., and Harhaj, E.W. (2010). Inhibition of NF-kappaB signaling by A20 through disruption of ubiquitin enzyme complexes. Science 327, 1135-1139.

Sheng, X.P., You, Q., Zhu, H.N., Chang, Z.N., Li, Q.R., Wang, H.F., Wang, C., Wang, H.Y., Hui, L.J., Du, C.T., et al. (2017). Bacterial effector NleL promotes

323 enterohemorrhagic E. coli-induced attaching and effacing lesions by ubiquitylating and inactivating JNK. PLOS Pathogens 13.

Sheng, Y., Hong, J.H., Doherty, R., Srikumar, T., Shloush, J., Avvakumov, G.V., Walker, J.R., Xue, S., Neculai, D., Wan, J.W., et al. (2012). A human ubiquitin conjugating enzyme (E2)-HECT E3 ligase structure-function screen. Mol Cell Proteomics 11, 329-341.

Shenoy, A.R., and Visweswariah, S.S. (2003). Site-directed mutagenesis using a single mutagenic oligonucleotide and DpnI digestion of template DNA. Analytical biochemistry 319, 335-336.

Shenoy, A.R., Wellington, D.A., Kumar, P., Kassa, H., Booth, C.J., Cresswell, P., and MacMicking, J.D. (2012). GBP5 promotes NLRP3 inflammasome assembly and immunity in mammals. Science 336, 481-485.

Shi, C.S., Shenderov, K., Huang, N.N., Kabat, J., Abu-Asab, M., Fitzgerald, K.A., Sher, A., and Kehrl, J.H. (2012). Activation of autophagy by inflammatory signals limits IL-1 beta production by targeting ubiquitinated inflammasomes for destruction. Nature immunology 13, 255-U274.

Shi, F.S., Yang, Y., Kouadir, M., Xu, W., Hu, S.H., and Wang, T.C. (2016a). Inflammasome-independent role of NLRP12 in suppressing colonic inflammation regulated by Blimp-1. Oncotarget 7, 30575-30584.

Shi, H.X., Wang, Y., Li, X.H., Zhan, X.M., Tang, M., Fina, M., Su, L.J., Pratt, D., Bu, C.H., Hildebrand, S., et al. (2016b). NLRP3 activation and mitosis are mutually exclusive events coordinated by NEK7, a new inflammasome component. Nature immunology 17, 250-258.

Shi, J., Zhao, Y., Wang, K., Shi, X., Wang, Y., Huang, H., Zhuang, Y., Cai, T., Wang, F., and Shao, F. (2015). Cleavage of GSDMD by inflammatory caspases determines pyroptotic cell death. Nature 526, 660-665.

Shi, J., Zhao, Y., Wang, Y., Gao, W., Ding, J., Li, P., Hu, L., and Shao, F. (2014). Inflammatory caspases are innate immune receptors for intracellular LPS. Nature 514, 187-192.

324

Shiba-Fukushima, K., Imai, Y., Yoshida, S., Ishihama, Y., Kanao, T., Sato, S., and Hattori, N. (2012). PINK1-mediated phosphorylation of the Parkin ubiquitin-like domain primes mitochondrial translocation of Parkin and regulates mitophagy. Scientific reports 2, 1002.

Sibley, C.H., Chioato, A., Felix, S., Colin, L., Chakraborty, A., Plass, N., Rodriguez- Smith, J., Brewer, C., King, K., Zalewski, C., et al. (2015). A 24-month open-label study of canakinumab in neonatal-onset multisystem inflammatory disease. Annals of the rheumatic diseases 74, 1714-1719.

Sims, J.E., and Smith, D.E. (2010). The IL-1 family: regulators of immunity. Nature reviews. Immunology 10, 89-102.

Sokolovska, A., Becker, C.E., Ip, W.K., Rathinam, V.A., Brudner, M., Paquette, N., Tanne, A., Vanaja, S.K., Moore, K.J., Fitzgerald, K.A., et al. (2013). Activation of caspase-1 by the NLRP3 inflammasome regulates the NADPH oxidase NOX2 to control phagosome function. Nat Immunol 14, 543-553.

Sommer, T., and Jentsch, S. (1993). A protein translocation defect linked to ubiquitin conjugation at the endoplasmic reticulum. Nature 365, 176-179.

Song, H., Liu, B., Huai, W., Yu, Z., Wang, W., Zhao, J., Han, L., Jiang, G., Zhang, L., Gao, C., et al. (2016). The E3 ubiquitin ligase TRIM31 attenuates NLRP3 inflammasome activation by promoting proteasomal degradation of NLRP3. Nature communications 7, 13727.

Song, J., Tan, H., Shen, H., Mahmood, K., Boyd, S.E., Webb, G.I., Akutsu, T., and Whisstock, J.C. (2010). Cascleave: towards more accurate prediction of caspase substrate cleavage sites. Bioinformatics 26, 752-760.

Squires, R.C., Muehlbauer, S.M., and Brojatsch, J. (2007). Proteasomes control caspase-1 activation in anthrax lethal toxin-mediated cell killing. The Journal of biological chemistry 282, 34260-34267.

Stahl, E.A., Raychaudhuri, S., Remmers, E.F., Xie, G., Eyre, S., Thomson, B.P., Li, Y.H., Kurreeman, F.A.S., Zhernakova, A., Hinks, A., et al. (2010). Genome-wide

325 association study meta-analysis identifies seven new rheumatoid arthritis risk loci. Nat Genet 42, 508-U556.

Stewart, M.D., Ritterhoff, T., Klevit, R.E., and Brzovic, P.S. (2016). E2 enzymes: more than just middle men. Cell research 26, 423-440.

Stieglitz, B., Morris-Davies, A.C., Koliopoulos, M.G., Christodoulou, E., and Rittinger, K. (2012). LUBAC synthesizes linear ubiquitin chains via a thioester intermediate. Embo Rep 13, 840-846.

Stieglitz, B., Rana, R.R., Koliopoulos, M.G., Morris-Davies, A.C., Schaeffer, V., Christodoulou, E., Howell, S., Brown, N.R., Dikic, I., and Rittinger, K. (2013). Structural basis for ligase-specific conjugation of linear ubiquitin chains by HOIP. Nature 503, 422-+.

Storek, K.M., and Monack, D.M. (2015). Bacterial recognition pathways that lead to inflammasome activation. Immunol Rev 265, 112-129.

Sutterwala, F.S., Ogura, Y., Szczepanik, M., Lara-Tejero, M., Lichtenberger, G.S., Grant, E.P., Bertin, J., Coyle, A.J., Galan, J.E., Askenase, P.W., et al. (2006). Critical role for NALP3/CIAS1/Cryopyrin in innate and adaptive immunity through its regulation of caspase-1. Immunity 24, 317-327.

Swaan, P.W., Knoell, D.L., Helsper, F., and Wewers, M.D. (2001). Sequential processing of human ProIL-1beta by caspase-1 and subsequent folding determined by a combined in vitro and in silico approach. Pharmaceutical research 18, 1083- 1090.

Swatek, K.N., and Komander, D. (2016). Ubiquitin modifications. Cell research 26, 399-422.

Taabazuing, C.Y., Okondo, M.C., and Bachovchin, D.A. (2017). Pyroptosis and Apoptosis Pathways Engage in Bidirectional Crosstalk in Monocytes and Macrophages. Cell chemical biology 24, 507-514 e504.

326

Tait, S.W., de Vries, E., Maas, C., Keller, A.M., D'Santos, C.S., and Borst, J. (2007). Apoptosis induction by Bid requires unconventional ubiquitination and degradation of its N-terminal fragment. The Journal of cell biology 179, 1453-1466.

Takeuchi, O., and Akira, S. (2010). Pattern recognition receptors and inflammation. Cell 140, 805-820.

Takeuchi, O., Hoshino, K., Kawai, T., Sanjo, H., Takada, H., Ogawa, T., Takeda, K., and Akira, S. (1999). Differential roles of TLR2 and TLR4 in recognition of gram- negative and gram-positive bacterial cell wall components. Immunity 11, 443-451.

Takeuchi, O., Kawai, T., Muhlradt, P.F., Morr, M., Radolf, J.D., Zychlinsky, A., Takeda, K., and Akira, S. (2001). Discrimination of bacterial lipoproteins by Toll-like receptor 6. Int Immunol 13, 933-940.

Takeuchi, T., and Yokosawa, H. (2005). ISG15 modification of Ubc13 suppresses its ubiquitin-conjugating activity. Biochem Biophys Res Commun 336, 9-13.

Tenthorey, J.L., Haloupek, N., Lopez-Blanco, J.R., Grob, P., Adamson, E., Hartenian, E., Lind, N.A., Bourgeois, N.M., Chacon, P., Nogales, E., et al. (2017). The structural basis of flagellin detection by NAIP5: A strategy to limit pathogen immune evasion. Science 358, 888-893.

Tenthorey, J.L., Kofoed, E.M., Daugherty, M.D., Malik, H.S., and Vance, R.E. (2014). Molecular Basis for Specific Recognition of Bacterial Ligands by NAIP/NLRC4 Inflammasomes. Mol Cell 54, 17-29.

Thurston, T.L., Matthews, S.A., Jennings, E., Alix, E., Shao, F., Shenoy, A.R., Birrell, M.A., and Holden, D.W. (2016). Growth inhibition of cytosolic Salmonella by caspase-1 and caspase-11 precedes host cell death. Nature communications 7, 13292.

Tidow, N., Chen, X., Muller, C., Kawano, S., Gombart, A.F., Fischel-Ghodsian, N., and Koeffler, H.P. (2000). Hematopoietic-specific expression of MEFV, the gene mutated in familial Mediterranean fever, and subcellular localization of its corresponding protein, pyrin. Blood 95, 1451-1455.

327

Tobe, T., Beatson, S.A., Taniguchi, H., Abe, H., Bailey, C.M., Fivian, A., Younis, R., Matthews, S., Marches, O., Frankel, G., et al. (2006). An extensive repertoire of type III secretion effectors in Escherichia coli O157 and the role of lambdoid phages in their dissemination. Proceedings of the National Academy of Sciences of the United States of America 103, 14941-14946.

Toma, C., Higa, N., Koizumi, Y., Nakasone, N., Ogura, Y., McCoy, A.J., Franchi, L., Uematsu, S., Sagara, J., Taniguchi, S., et al. (2010). Pathogenic Vibrio Activate NLRP3 Inflammasome via Cytotoxins and TLR/Nucleotide-Binding Oligomerization Domain-Mediated NF-kappa B Signaling. J Immunol 184, 5287-5297.

Trempe, J.-F., Sauvé, V., Grenier, K., Seirafi, M., Tang, M.Y., Ménade, M., Al-Abdul- Wahid, S., Krett, J., Wong, K., Kozlov, G., et al. (2013). Structure of Parkin Reveals Mechanisms for Ubiquitin Ligase Activation. Science 340, 1451-1455.

Tseng, P.H., Matsuzawa, A., Zhang, W., Mino, T., Vignali, D.A., and Karin, M. (2010). Different modes of ubiquitination of the adaptor TRAF3 selectively activate the expression of type I interferons and proinflammatory cytokines. Nat Immunol 11, 70-75.

Tsuji, N.M., Tsutsui, H., Seki, E., Kuida, K., Okamura, H., Nakanishi, K., and Flavell, R.A. (2004). Roles of caspase-1 in Listeria infection in mice. Int Immunol 16, 335- 343.

Tsutsumi, N., Kimura, T., Arita, K., Ariyoshi, M., Ohnishi, H., Yamamoto, T., Zuo, X.B., Maenaka, K., Park, E.Y., Kondo, N., et al. (2014). The structural basis for receptor recognition of human interleukin-18. Nature communications 5.

Urrego, D., Tomczak, A.P., Zahed, F., Stuhmer, W., and Pardo, L.A. (2014). Potassium channels in cell cycle and cell proliferation. Philos T R Soc B 369.

Vajjhala, P.R., Kaiser, S., Smith, S.J., Ong, Q.R., Soh, S.L., Stacey, K.J., and Hill, J.M. (2014). Identification of multifaceted binding modes for pyrin and ASC pyrin domains gives insights into pyrin inflammasome assembly. The Journal of biological chemistry 289, 23504-23519.

328

Van Damme, P., Martens, L., Van Damme, J., Hugelier, K., Staes, A., Vandekerckhove, J., and Gevaert, K. (2005). Caspase-specific and nonspecific in vivo protein processing during Fas-induced apoptosis. Nat Methods 2, 771-777.

Van Opdenbosch, N., Gurung, P., Vande Walle, L., Fossoul, A., Kanneganti, T.D., and Lamkanfi, M. (2014). Activation of the NLRP1b inflammasome independently of ASC-mediated caspase-1 autoproteolysis and speck formation. Nature communications 5, 3209.

Varland, S., Osberg, C., and Arnesen, T. (2015). N-terminal modifications of cellular proteins: The enzymes involved, their substrate specificities and biological effects. Proteomics 15, 2385-2401.

Varshavsky, A. (2011). The N-end rule pathway and regulation by proteolysis. Protein science : a publication of the Protein Society 20, 1298-1345.

Vereecke, L., Beyaert, R., and van Loo, G. (2009). The ubiquitin-editing enzyme A20 (TNFAIP3) is a central regulator of immunopathology. Trends in immunology 30, 383-391.

Vladimer, G.I., Weng, D., Paquette, S.W.M., Vanaja, S.K., Rathinam, V.A.K., Aune, M.H., Conlon, J.E., Burbage, J.J., Proulx, M.K., Liu, Q., et al. (2012). The NLRP12 Inflammasome Recognizes Yersinia pestis. Immunity 37, 96-107. von Moltke, J., Ayres, J.S., Kofoed, E.M., Chavarria-Smith, J., and Vance, R.E. (2013). Recognition of bacteria by inflammasomes. Annual review of immunology 31, 73-106.

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. Mol Cell Proteomics 10.

Walev, I., Reske, K., Palmer, M., Valeva, A., and Bhakdi, S. (1995). Potassium- Inhibited Processing of Il-1-Beta in Human Monocytes. The EMBO journal 14, 1607- 1614.

329

Walsh, J.G., Logue, S.E., Luthi, A.U., and Martin, S.J. (2011). Caspase-1 Promiscuity Is Counterbalanced by Rapid Inactivation of Processed Enzyme. Journal of Biological Chemistry 286, 32513-32524.

Wang, M., Cheng, D., Peng, J., and Pickart, C.M. (2006). Molecular determinants of polyubiquitin linkage selection by an HECT ubiquitin ligase. EMBO J 25, 1710-1719.

Wang, Q., Liu, X., Cui, Y., Tang, Y., Chen, W., Li, S., Yu, H., Pan, Y., and Wang, C. (2014). The E3 ubiquitin ligase AMFR and INSIG1 bridge the activation of TBK1 kinase by modifying the adaptor STING. Immunity 41, 919-933.

Wang, S., Miura, M., Jung, Y.-k., Zhu, H., Li, E., and Yuan, J. (1998). Murine Caspase-11, an ICE-Interacting Protease, Is Essential for the Activation of ICE. Cell 92, 501-509.

Wang, X., Herr, R.A., and Hansen, T.H. (2012). Ubiquitination of substrates by esterification. Traffic 13, 19-24.

Warren, S.E., Armstrong, A., Hamilton, M.K., Mao, D.P., Leaf, I.A., Miao, E.A., and Aderem, A. (2010). Cutting Edge: Cytosolic Bacterial DNA Activates the Inflammasome via Aim2. J Immunol 185, 818-821.

Warren, S.E., Mao, D.P., Rodriguez, A.E., Miao, E.A., and Aderem, A. (2008). Multiple Nod-like receptors activate caspase 1 during Listeria monocytogenes infection. J Immunol 180, 7558-7564.

Wei, C., Wang, Y., Du, Z., Guan, K., Cao, Y., Yang, H., Zhou, P., Wu, F., Chen, J., Wang, P., et al. (2016). The Yersinia Type III secretion effector YopM Is an E3 ubiquitin ligase that induced necrotic cell death by targeting NLRP3. Cell death & disease 7, e2519.

Wellington, M., Koselny, K., Sutterwala, F.S., and Krysan, D.J. (2014). Candida albicans triggers NLRP3-mediated pyroptosis in macrophages. Eukaryotic cell 13, 329-340.

330

Wenzel, D.M., Lissounov, A., Brzovic, P.S., and Klevit, R.E. (2011). UBCH7 reactivity profile reveals parkin and HHARI to be RING/HECT hybrids. Nature 474, 105-108.

Wertz, I.E., O'Rourke, K.M., Zhou, H., Eby, M., Aravind, L., Seshagiri, S., Wu, P., Wiesmann, C., Baker, R., Boone, D.L., et al. (2004). De-ubiquitination and ubiquitin ligase domains of A20 downregulate NF-kappaB signalling. Nature 430, 694-699.

Whitcomb, E.A., Dudek, E.J., Liu, Q., and Taylor, A. (2009). Novel control of S phase of the cell cycle by ubiquitin-conjugating enzyme H7. Molecular biology of the cell 20, 1-9.

Wickliffe, K.E., Leppla, S.H., and Moayeri, M. (2008). Killing of macrophages by anthrax lethal toxin: involvement of the N-end rule pathway. Cell Microbiol 10, 1352- 1362.

Wickliffe, K.E., Lorenz, S., Wemmer, D.E., Kuriyan, J., and Rape, M. (2011). The Mechanism of Linkage-Specific Ubiquitin Chain Elongation by a Single-Subunit E2. Cell 144, 769-781.

Wilson, S.P., and Cassel, S.L. (2010). Inflammasome-Mediated Autoinflammatory Disorders. Postgrad Med 122, 125-133.

Windheim, M., Peggle, M., and Cohen, P. (2008). Two different classes of E2 ubiquitin-conjugating enzymes are required for the mono-ubiquitination of proteins and elongation by polyubiquitin chains with a specific topology. Biochemical Journal 409, 723-729.

Witting, K.F., Mulder, M.P.C., and Ovaa, H. (2017). Advancing our Understanding of Ubiquitination Using the Ub-Toolkit. J Mol Biol 429, 3388-3394.

Wood, M.W., Jones, M.A., Watson, P.R., Siber, A.M., McCormick, B.A., Hedges, S., Rosqvist, R., Wallis, T.S., and Galyov, E.E. (2000). The secreted effector protein of Salmonella dublin, SopA, is translocated into eukaryotic cells and influences the induction of enteritis. Cell Microbiol 2, 293-303.

331

Wu, J., Fernandes-Alnemri, T., and Alnemri, E.S. (2010). Involvement of the AIM2, NLRC4, and NLRP3 inflammasomes in caspase-1 activation by Listeria monocytogenes. Journal of clinical immunology 30, 693-702.

Xia, Z.P., Sun, L.J., Chen, X., Pineda, G., Jiang, X.M., Adhikari, A., Zeng, W.W., and Chen, Z.J. (2009). Direct activation of protein kinases by unanchored polyubiquitin chains. Nature 461, 114-U125.

Xu, H., Yang, J.L., Gao, W.Q., Li, L., Li, P., Zhang, L., Gong, Y.N., Peng, X.L., Xi, J.Z.J., Chen, S., et al. (2014). Innate immune sensing of bacterial modifications of Rho GTPases by the Pyrin inflammasome. Nature 513, 237-+.

Xu, P., Duong, D.M., Seyfried, N.T., Cheng, D.M., 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.

Xu, Y., Anderson, D.E., and Ye, Y.H. (2016). The HECT domain ubiquitin ligase HUWE1 targets unassembled soluble proteins for degradation. Cell Discov 2.

Yang, D., He, Y., Munoz-Planillo, R., Liu, Q., and Nunez, G. (2015). Caspase-11 Requires the Pannexin-1 Channel and the Purinergic P2X7 Pore to Mediate Pyroptosis and Endotoxic Shock. Immunity 43, 923-932.

Yang, J., Liu, Z.H., and Xiao, T.S. (2017). Post-translational regulation of inflammasomes. Cellular & molecular immunology 14, 65-79.

Yang, J., Zhao, Y., Shi, J., and Shao, F. (2013a). Human NAIP and mouse NAIP1 recognize bacterial type III secretion needle protein for inflammasome activation. Proceedings of the National Academy of Sciences of the United States of America 110, 14408-14413.

Yang, X.F., Brownell, J.E., Xu, Q., Zhu, F.Y., Ma, J.Y., Loke, H.K., Rollins, N., Soucy, T.A., Minissale, J.J., Thomas, M.P., et al. (2013b). Absolute Quantification of E1, Ubiquitin-Like Proteins and Nedd8-MLN4924 Adduct by Mass Spectrometry. Cell Biochem Biophys 67, 139-147.

332

Yang, X.L., Chang, H.Y., and Baltimore, D. (1998). Autoproteolytic activation of pro- caspases by oligomerization. Mol Cell 1, 319-325.

Yau, R., and Rape, M. (2016). The increasing complexity of the ubiquitin code. Nat Cell Biol 18, 579-586.

Yau, R.G., Doerner, K., Castellanos, E.R., Haakonsen, D.L., Werner, A., Wang, N., Yang, X.W., Martinez-Martin, N., Matsumoto, M.L., Dixit, V.M., et al. (2017). Assembly and Function of Heterotypic Ubiquitin Chains in Cell-Cycle and Protein Quality Control. Cell 171, 918-933 e920.

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

Yeh, T.C., and Bratton, S.B. (2013). Caspase-dependent regulation of the ubiquitin- proteasome system through direct substrate targeting. Proceedings of the National Academy of Sciences of the United States of America 110, 14284-14289.

Yen, H., Sugimoto, N., and Tobe, T. (2015). Enteropathogenic Escherichia coli Uses NleA to Inhibit NLRP3 Inflammasome Activation. PLoS Pathog 11, e1005121.

Yu, J., Nagasu, H., Murakami, T., Hoang, H., Broderick, L., Hoffman, H.M., and Horng, T. (2014). Inflammasome activation leads to Caspase-1-dependent mitochondrial damage and block of mitophagy. Proceedings of the National Academy of Sciences of the United States of America 111, 15514-15519.

Yu, J.R., and Leslie, K.S. (2011). Cryopyrin-Associated Periodic Syndrome: An Update on Diagnosis and Treatment Response. Curr Allergy Asthm R 11, 12-20.

Yu, J.W., Fernandes-Alnemri, T., Datta, P., Wu, J., Juliana, C., Solorzano, L., McCormick, M., Zhang, Z., and Alnemri, E.S. (2007). Pyrin activates the ASC pyroptosome in response to engagement by autoinflammatory PSTPIP1 mutants. Mol Cell 28, 214-227.

Yuan, W.C., Lee, Y.R., Lin, S.Y., Chang, L.Y., Tan, Y.P., Hung, C.C., Kuo, J.C., Liu, C.H., Lin, M.Y., Xu, M., et al. (2014). K33-Linked Polyubiquitination of Coronin 7 by

333

Cul3-KLHL20 Ubiquitin E3 Ligase Regulates Protein Trafficking. Mol Cell 54, 586- 600.

Zak, D.E., Schmitz, F., Gold, E.S., Diercks, A.H., Peschon, J.J., Valvo, J.S., Niemisto, A., Podolsky, I., Fallen, S.G., Suen, R., et al. (2011). Systems analysis identifies an essential role for SHANK-associated RH domain-interacting protein (SHARPIN) in macrophage Toll-like receptor 2 (TLR2) responses. Proceedings of the National Academy of Sciences of the United States of America 108, 11536- 11541.

Zaki, M.H., Vogel, P., Malireddi, R.K.S., Body-Malapel, M., Anand, P.K., Bertin, J., Green, D.R., Lamkanfi, M., and Kanneganti, T.D. (2011). The NOD-Like Receptor NLRP12 Attenuates Colon Inflammation and Tumorigenesis. Cancer Cell 20, 649- 660.

Zanoni, I., Tan, Y., Di Gioia, M., Broggi, A., Ruan, J., Shi, J., Donado, C.A., Shao, F., Wu, H., Springstead, J.R., et al. (2016). An endogenous caspase-11 ligand elicits interleukin-1 release from living dendritic cells. Science 352, 1232-1236.

Zeng, W.W., Sun, L.J., Jiang, X.M., Chen, X., Hou, F.J., Adhikari, A., Xu, M., and Chen, Z.J.J. (2010). Reconstitution of the RIG-I Pathway Reveals a Signaling Role of Unanchored Polyubiquitin Chains in Innate Immunity. Cell 141, 315-330.

Zhang, L., Chen, S., Ruan, J., Wu, J., Tong, A.B., Yin, Q., Li, Y., David, L., Lu, A., Wang, W.L., et al. (2015). Cryo-EM structure of the activated NAIP2-NLRC4 inflammasome reveals nucleated polymerization. Science 350, 404-409.

Zhang, S.P., Santos, R.L., Tsolis, R.M., Stender, S., Hardt, W.D., Baumler, A.J., and Adams, L.G. (2002). The Salmonella enterica serotype typhimurium effector proteins SipA, SopA, SopB, SopD, and SopE2 act in concert to induce diarrhea in calves. Infection and immunity 70, 3843-3855.

Zhang, X.A., Cheng, Y.L., Xiong, Y.W., Ye, C.Y., Zheng, H., Sun, H., Zhao, H.Q., Ren, Z.H., and Xu, J.G. (2012). Enterohemorrhagic Escherichia coli Specific Enterohemolysin Induced IL-1 beta in Human Macrophages and EHEC-Induced IL-1 beta Required Activation of NLRP3 Inflammasome. PloS one 7.

334

Zhang, Y., Higashide, W.M., McCormick, B.A., Chen, J., and Zhou, D.G. (2006). The inflammation-associated Salmonella SopA is a HECT-like E3 ubiquitin ligase. Molecular Microbiology 62, 786-793.

Zhang, Y., and Liang, C.L. (2016). Innate recognition of microbial-derived signals in immunity and inflammation. Sci China Life Sci 59, 1210-1217.

Zhao, Y., Yang, J.L., Shi, J.J., Gong, Y.N., Lu, Q.H., Xu, H., Liu, L.P., and Shao, F. (2011). The NLRC4 inflammasome receptors for bacterial flagellin and type III secretion apparatus. Nature 477, 596-U257.

Zheng, N., and Shabek, N. (2017). Ubiquitin Ligases: Structure, Function, and Regulation. Annual review of biochemistry 86, 129-157.

Zheng, N., Wang, P., Jeffrey, P.D., and Pavletich, N.P. (2000). Structure of a c-Cbl- UbcH7 complex: RING domain function in ubiquitin-protein ligases. Cell 102, 533- 539.

Zhou, Y., Dong, N., Hu, L., and Shao, F. (2013). The Shigella type three secretion system effector OspG directly and specifically binds to host ubiquitin for activation. PloS one 8, e57558.

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. Biochem Biophys Res Commun 336, 61-68.

335

Appendix

UBE2L3 estimated Mw 18.23 kDa

1 11 21 31 41 51 61 GPLGS MAASRRLMKE LEEIRKCGMK NFRNIQVDEA NLLTWQGLIV PDNPPYDKGA FRIEINFPAE YPFKPPKITF 71 81 91 101 111 121 131 KTKIYHPNID EKGQVCLPVI SAENWKPATK TDQVIQSLIA LVNDPQPEHP LRADLAEEYS KDRKKFCKNA 141 151 EEFTKKYGEK RPVD

Appendix Figure 1 – Amino acid sequence of recombinant UBE2L3 used for in vitro ubiquitylation assays. Lysine residues are in bold and underlined. Residual linker sequence after GST-removal is in green.

Lysine Residue Lys only Lys > Arg Lys present in 9R Lys present in 14R Lysine 9 Lysine 16 Y Y Lysine 20 Y Lysine 48 Y Lysine 64 Y Y Lysine 67 Lysine 71 Y Y Lysine 73 Lysine 82 Y Y Lysine 96 Y Lysine 100 Lysine 131 Y Y Lysine 134 Y Y Lysine 135 Y Y Lysine 138 Y Y Y Lysine 145 Lysine 146 Y Y Lysine 150 Y Y Y

Appendix Table 1 – Lysine residues present in UBE2L3 and mutant variants. Lysine residues which have previously been reported to be ubiquitylated are in bold

336