Theses - MRC Laboratory of Molecular Biology
Permanent URI for this collection
Browse
Recent Submissions
Item Embargo Structural insights into the GABAA receptor gating cycleMihaylov, DanielGABAA receptors are pentameric ligand-gated ion channels (pLGICs) that control multiple aspects of animal physiology, including the rapid inhibitory neurotransmission within the vertebrate central nervous system. They activate and desensitize in a few milliseconds, which is a crucial requirement for accurate synaptic signalling. Such fast conformational transitions, however, have proven impervious to investigation via conventional structural biology methods. Moreover, GABAA receptors operate in cellular membranes with diverse lipid compositions and biophysical properties. These are thought to play essential modulatory functions, however, a molecular understanding of such impacts is lacking. As a result, the mechanisms of GABA-ergic signalling, in physiologically meaningful environments and time scales, remain among the most important unknowns in molecular neuroscience. To address these issues, I first sought to identify optimal reconstitution environments for recombinantly expressed β3 homomeric and synaptic α1β3γ2L GABAA receptors by testing multiple nanodisc scaffolds and lipid compositions. Subsequently, with the optimal condition established, I proceeded to investigate a series of GABAA receptor structures bound to β-lactam antibiotics that are postulated to act as open channel pore blockers. Surprisingly, complexes with the penicillin derivative nafcillin suggested that the GABAA receptor gating motions might include conformations that diverge considerably from the pseudo-symmetric arrangements observed previously. To test whether the asymmetry observed is physiologically meaningful, and to achieve a comprehensive characterisation of the GABAA receptor gating cycle, I employed a time-resolved cryo-EM strategy that enables millisecond-scale time delays between agonist application and sample freezing. This approach allowed the visualization of novel and complex asymmetric gating motions involved in the activation and desensitization of the receptor, ranging from side chain to major secondary structure rearrangements. Remarkably, the asymmetric gating cycle is conserved among homomeric and heteromeric GABAA receptors, with lipids such as cholesterol playing a crucial role in stabilizing the newly formed interfaces. Moreover, these results have further revealed a novel inhibitory mechanism driven by the lipid PIP2, which directly controls the opening of α1β3 and α1β3γ2L receptors. Lastly, I will also demonstrate how positive allosteric modulators alter the gating landscape and how the general anaesthetic propofol can directly activate the receptor. This work provides a new structural and mechanistic framework to explore and interpret the action of pharmacological modulators and to design novel molecules with improved therapeutic properties.Item Embargo Deciphering the regulation of perforin-2-mediated endocytic escape in dendritic cellsLaub, MarcoDendritic cells (DCs) are important players in the initiation of cytotoxic CD8+ T cell-mediated immune responses against pathogens and tumours. DCs continuously sample their environment for antigens and load them on their MHC class I molecules to prime naïve CD8+ T cells in a process known as cross-presentation. Unlike most cell types, cross-presenting DCs have leaky endocytic compartments from where internalised proteins can escape into the cytosol for proteasome-mediated generation of MHC-I-binding peptides. Prior to my project, the Kozik lab had identified the pore-forming protein perforin-2 as a dedicated effector of endocytic escape that allows cytosolic entry of exogenous antigen during cross-presentation. Perforin-2 is a member of the membrane attack complex and perforin superfamily (MACPF) and, similar to perforin and complement, can form oligomeric pores on liposomes *in vitro*. However, the regulation of perforin-2 pore formation in the context of endocytic escape *in vivo* is still poorly understood. The aim of my PhD project was to decipher the mechanisms that enable DCs to assemble perforin-2 pores on their own membranes for endocytic escape without causing uncontrolled lysis of intracellular compartments. Analysis of its steady-state distribution showed that perforin-2 was proteolytically processed with full-length and cleaved forms present in early and late endocytic compartments, respectively. Since perforin-2 was recruited to antigen-containing phagosomes, I studied its stepwise proteolytic processing during phagosome maturation. Flow cytometry analysis of isolated phagosomes revealed that, in line with previous *in vitro* studies, perforin-2 processing was dependent on acidification. However, by developing an assay to monitor perforin-2 activity at the phagosome, I discovered that inhibiting acidification had no impact on the efficiency of perforin-2-mediated escape. In addition, perforin-2 activity had no detectable effect on the luminal pH or degradative potential of phagosomes. To better understand the role of pH-dependent processing of perforin-2 during phagosome maturation, I studied the proteases involved. The analysis showed that the pore-forming ectodomain was cleaved off the transmembrane domain by cysteine proteases and further processed by asparagine endopeptidase while the remaining membrane stub was removed by γ-secretase. Specific inhibition of these proteases had no impact on perforin-2-mediated endocytic escape suggesting that they are not directly required for pore formation. However, analysis of a perforin-2 trafficking mutant suggested that a currently unknown processing step may be necessary for efficient perforin-2 activity at the phagosome. In summary, this thesis offers novel insights into the regulation of perforin-2 which potentially releases antigens for cross-presentation early during phagosome maturation.Item Embargo Biochemical and structural studies of human telomerase regulationSekne, ZalaTelomerase is a ribonucleoprotein (RNP) enzyme that extends telomeric repeats at eukaryotic chromosome ends (Blackburn and Collins, 2011; Blackburn et al., 2006; Greider et al., 1985; Morin, 1989; Nguyen, 2021). This counterbalances telomere loss caused by incomplete genome replication (Blackburn and Collins, 2011; Greider et al., 1985). Human telomerase has two distinct functional lobes, termed the catalytic core and the Hinge and ACA (H/ACA) box RNP, which are tethered together by telomerase RNA (hTR) (Nguyen et al., 2018). The H/ACA RNP lobe is essential for telomerase biogenesis and hTR accumulation, and it harbours many disease mutation residues (Egan and Collins, 2010, 2012b; Ghanim et al., 2021; Nguyen et al., 2018). H/ACA RNPs also have a general role in pseudouridylation of spliceosomal and ribosomal RNAs (Charette and Gray, 2000). The telomerase catalytic core, on the other hand, is responsible for telomeric DNA extension and telomerase recruitment to telomeres (Egan and Collins, 2010; Nandakumar et al., 2012; Nguyen et al., 2018; Zhong et al., 2012). In mammals, telomerase is recruited to telomeres by the shelterin component TPP1 (Nandakumar et al., 2012; Sexton et al., 2012; Zhong et al., 2012), which forms a heterodimer with another shelterin component POT1 (Xin et al., 2007). After synthesizing approximately 60 telomeric repeats, telomerase is terminated by a CST (CTC1-STN1-TEN1) complex (Feng et al., 2018). In my thesis, I describe the biochemical and structural work aimed at understanding the molecular basis of human telomerase and its interaction with regulatory factors. In Chapter 2, I reconstruct telomerase complexes with shelterin regulatory factors. This yields the first structures of the human telomerase catalytic core with TPP1 and POT1, resolved at 3.2 and 3.9 Å resolution. The structures elucidate specific interactions between telomerase and TPP1-POT1 that are essential for telomerase recruitment to telomeres and rationalise prior genetic and biochemical findings. Together with mutagenesis studies, the structures reveal an unexpected path of telomeric DNA and provide insights into a DNA anchor site on telomerase, which is important for telomerase processivity. The same cryo-EM dataset was used to obtain a structure of the telomerase H/ACA RNP at 2.7 Å resolution, which is presented in Chapter 3. This structure reveals unprecedented insights into a region critical for eukaryotic pseudouridylation in canonical H/ACA RNPs. Additionally, we find that many disease-associated residues of the H/ACA RNP proteins form interactions critical for the H/ACA RNP stability. The cryo-EM structures obtained in Chapter 2 and 3 represent a monomeric configuration of human telomerase with one catalytic core and one H/ACA RNP. In collaboration with S. Balch, who collected a large cryo-EM dataset of human telomerase with various DNA substrates, we resolved the first high-resolution structure of the telomerase dimer. In Chapter 4, preliminary structural and biochemical analyses of the telomerase dimer reveal the unexpected inter-protomer RNA-protein and protein-protein interactions. Moreover, regions of hTR implicated in telomerase dimerization as well as assembly and biogenesis are resolved for the first time. Finally, to investigate the interplay between telomerase, shelterin and CST, I conducted biochemical and structural analyses of the CST complex in conjunction with TPP1-POT1-TIN2 and telomerase. As presented in Chapter 5, a newly-observed conformation of CST, together with the pull-down experiments, sheds light on CST interaction with shelterin components. Furthermore, preliminary data on the ternary telomerase-TPP1-POT1-TIN2-CST complex provide an invaluable platform for further studying telomerase termination. Overall, the work presented in my thesis provides a framework for future mechanistic studies on telomerase regulation at telomeres.Item Embargo Biogenesis of Multipass Membrane ProteinsSmalinskaite, LukaRoughly one-fourth of all genes code for integral membrane proteins. Most membrane proteins pass back and forth across the lipid bilayer multiple times. Multipass membrane proteins, such as ion channels and receptors, are essential for nearly all transport and communication across the plasma membrane. In eukaryotes, nascent multipass membrane proteins are initially inserted into the endoplasmic reticulum (ER) membrane, where they fold and assemble before being trafficked to their final destination. Membrane proteins typically insert co-translationally as the nascent multipass protein emerges from an ER-associated ribosome. The long-standing model has been that each transmembrane domain (TMD) inserts into the membrane sequentially via the Sec61 translocation channel. This framework was primarily extrapolated from mechanistic studies of a few single-pass membrane proteins, but direct analyses of multipass membrane proteins were limited. This thesis reports efforts to understand the sequence of events, machinery, and mechanisms of multipass membrane protein biogenesis. Biochemical reconstitution combined with genetic and chemical perturbations showed that the first two TMDs of most multipass proteins with a cytosol-facing N-terminus are inserted by a Sec61-independent mechanism. Instead, this insertion relies, at least in part, on the ER membrane complex (EMC), a TMD insertase recently appreciated to be a member of the universally conserved Oxa1 family of insertases. Strict dependence on Sec61 was only seen for TMD pairs separated by a translocated region longer than ≈100 amino acids. Purification and analysis of later multipass insertion intermediates revealed a ribosome-associated machinery comprised of four complexes: the Sec61 translocation channel, an intramembrane chaperone called PAT, an Oxa1 insertase family member called GEL, and a complex of unknown function termed BOS. Structural and mutational analysis showed that the PAT complex contains a membrane-exposed amphipathic surface that binds nascent TMDs and a ‘latch’ domain that prevents opening of the Sec61 channel. These activities re-direct the nascent protein away from Sec61 and towards GEL, which is seen adjacent to subsequently inserted TMD pairs. Crosslinking experiments and structural work show that newly inserted TMDs accumulate in a semi-circular lipid-filled cavity formed by Sec61, PAT, GEL and BOS, known as the multipass translocon. These findings lead to a model for multipass protein biogenesis in which most TMDs are inserted by Oxa1 family members; only TMDs flanked by long translocated domains rely strictly on the Sec61 channel. The nascent TMDs of a multipass protein progressively accumulate in the semi-protected environment provided by the multipass translocon, where they can fold to complete their biogenesis.Item Embargo Integrin signalling in pluripotent cells acts as a gatekeeper of mouse germline entryMakhlouf, Aly; Makhlouf, Aly [0000-0003-4025-0576]Primordial germ cells (PGCs) are the undifferentiated precursors of gametes and the sole mechanism by which genetic information is transmitted across generations through the germline. They appear at the proximal posterior region of the mouse embryo, where the somatic germ layers begin to emerge and a series of dramatic morphological changes occur at the onset of gastrulation. While the transcriptional and epigenetic mechanisms that drive entry into the germline have been extensively characterised, whether there is a morphogenetic control of the germline-soma bifurcation has never been explored. Here, we uncover a novel and critical role for integrin signalling in regulating germline entry. *In vivo*, we show that mouse PGCs have diminished integrin signalling interactions with the underlying basement membrane and specify in a region that is physically segregated from it. In line with this, the addition of an exogenous extracellular matrix (ECM) inhibits PGC specification in *in vitro*-cultured mouse embryos, as well as during PGC-like cell (PGCLC) specification in both mouse and human *in vitro* stem cell models. Mechanistically, we demonstrate that β1 integrin signalling, in response to the ECM protein, laminin, blocks mouse PGCLC specification by inhibiting Wnt signalling, which leads to failed downregulation of the PGC transcriptional repressor, Otx2, during a critical window of competence. Therefore, integrin signalling restricts mouse germline entry, pointing to a mechanism through which the loss of basement membrane contact by epiblast cells *in vivo* acts as a morphogenetic fate switch during the germline-soma bifurcation.Item Open Access Hardware developments to improve image quality for cryoEM of biological specimensDickerson, JoshuaSeveral advances in cryoEM hardware have been integral in transforming the technique into a standard technique for high-resolution macromolecular structure determination. We are now entering an era where we can experimentally determine or predict the structure of almost the entire proteome. Efforts are naturally moving from solving purified structures to visualising macromolecules inside cells. However, because these specimens are an order of magnitude thicker than purified specimens, cryoEM inside cells is restricted to large complexes, such as ribosomes. Here, I investigate the use of two hardware developments, liquid helium cooling of the specimen during imaging and chromatic aberration (*Cc*) correction, to improve the quality of cryoEM images, the latter in particular for thick specimens. I demonstrate that *Cc* correction can be used to correctly focus inelastically scattered electrons, enhancing the signal significantly for thick specimens. I also demonstrate that reducing the specimen temperature from 81 K to 13 K reduces the rate of radiation damage in single-particle cryoEM, and I present progress in understanding other changes that occur when irradiating specimens at these temperatures. Lastly, I calculate the impact that these hardware developments could have on the minimum molecular mass of protein that we can visualise *in situ* by cryoEM. Taken together with other hardware and software developments, these technologies could dramatically alter the landscape of *in situ* cryoEM, with proteins as small as 100 kDa potentially being within reach.Item Open Access Regulation of polarised trafficking by Klp98ABittleston, AliceDirectional trafficking of cargo, a hallmark of living cells, relies on a polarised landscape of microtubule tracks. Here, I studied how the regulation of an endosomal motor affects polarised trafficking, using the asymmetric division of *Drosophila* Sensory Organ Precursor (SOP) cells as a model system. In this pathway, the PI(3)P-binding kinesin-3, Klp98A biases the motility of signalling endosomes towards the posterior daughter cell. Since these endosomes contain the cell fate determinants Notch and Delta, this contributes to an asymmetric cell fate. Mechanistically, Klp98A-mediated transport is achieved along the microtubule-based anaphase-specific spindle midzone, termed the central spindle. During SOP division, the central spindle becomes asymmetric, with a higher density of microtubules on the anterior side. Since Klp98A is a plus-end directed motor, this biases the motility of early endosomes into the posterior daughter cell, towards which more microtubule plus-ends point. In this work, I focused on how the modulation of Klp98A activity can control polarised trafficking on three levels: i) via post-translational modifications, ii) through changes in microtubule binding, and iii) by polarising the microtubules Klp98A walks on. By first using an *in vitro* reconstitution with purified motors on micropatterned supported lipid bilayers containing PI(3)P, I investigated how Klp98A is activated. Lipid-binding was sufficient to induce activation of Klp98A, so I proceeded to use both enzymatic-induced proximity labelling and a candidate-based approach to identify possible regulators of Klp98A activity. *In vivo*, modulation of two enzymes, p-21 activated kinase 1 (Pak1) and Sirtuin 2 (Sirt2), resulted in symmetric endosome segregation despite an asymmetric landscape of microtubule tracks. In the case of Pak1, this endosomal phenotype appeared to arise due to a delay in the targeting of endosomes to the central spindle, either directly due to modulation of post-translational modifications on the motor itself or through an intermediary. Next, I focused on characterising how the loop 8 (L8) region of the Klp98A motor domain modulates its activity. This well-conserved loop within the kinesin-3 family is thought to be involved in microtubule binding and could thus be a central way by which the activity of this motor is regulated *in vivo*. To understand how L8 controls motor activity, I studied a previously isolated Klp98A mutant, Klp98A Δ6, which consists of a two amino-acid deletion and one amino-acid substitution within the L8 site. Strikingly, *in vivo*, Klp98A Δ6 localises to endosomes like the *wild-type* but leads to symmetric segregation of cell fate determinants in dividing SOP cells, phenocopying a complete loss of function of Klp98A. A combination of *in vitro* reconstitution, single-molecule assays and crystal-structure determination showed that, unexpectedly, Klp98A Δ6 retains wild-type microtubule binding affinity but is unable to move processively. The cause of this transport defect by Klp98A Δ6 is likely a consequence of a long-range structural rearrangement from the L8 site to the catalytic domain, preventing ATP hydrolysis independently of microtubule binding. Thus, it seems in *wild-type* Klp98A, L8 functions as a sensor, coupling microtubule binding to ATP hydrolysis. Lastly, I investigated how the asymmetry of the central spindle, which polarises Klp98A motility, is established and maintained during SOP cell division. I demonstrated that at the cell cortex, the Par complex promotes central spindle symmetry breaking. Downstream of that, I found that Elongator, originally identified as a regulator of translation, promotes the formation of an asymmetric central spindle. Here, Elongator is enriched on the anterior side of the spindle, where it stabilises microtubules. Altogether, this data sheds light on how polarised trafficking emerges from the synergistic effects of both the polarity within the network of microtubules and the local control of the activity of motors at the surface of endosomes.Item Open Access Studies of the MYM-type zinc finger protein ZMYM3Shapson-Coe, AlexanderRNase H2 is a trimeric ribonuclease that specifically degrades the RNA moiety of DNARNA hybrids. Mutations in RNase H2 result in Aicardi–Goutières syndrome (AGS), a rare inflammatory disorder notable for being a ‘mendelian mimic’ of congenital viral brain infection. Some AGS-associated mutations in the RNase H2B subunit do not affect the catalytic activity of the RNase H2 *in vitro* and are clustered together on the surface of the complex, suggesting a possible role for these residues in mediating important interactions of RNase H2. I set out to identify binding partner(s) of RNase H2B whose binding is impaired by this cluster of mutations, by screening for interactors of FLAG-tagged wild-type and mutant RNase H2B in HEK293T cells. In this work, I identify several members of a putative chromatin-silencing complex as novel RNase H2B binding partners whose binding is impaired by each mutation in this cluster, including: HDAC2, a histone deacetylase, LSD1, a histone demethylase, CoREST, a co-repressor of transcription, TFII-I, a transcription factor, and ZMYM3, an MYM-type zinc finger protein of unknown function. By making several truncation mutants of ZMYM3, I show that a C-terminal region containing a ‘PXP’ repeat motif mediates its interaction with RNase H2, and that the PXP-containing proteins ZMYM2 and ZMYM4 interact with RNase H2, whereas the PXP-null proteins ZMYM1 and ZMYM6 do not. By systematic truncation of the zinc fingers of ZMYM3 I also demonstrate that the first zinc finger mediates its interaction with TFII-I, while the eighth and ninth zinc fingers are each sufficient for its interaction with LDS1, HDAC2 and CoREST. These interaction sites implicate ZMYM3 as a novel type of scaffold protein mediating interactions between deacetylase, demethylase and RNase H-type enzymes. To better understand the function of ZMYM3, I generate a ZMYM3 knock-out mouse embryonic stem cell line, and show that ZMYM3 is not required for cell viability or DNA repair, and is not cell-cycle regulated. Given the association of ZMYM3 with various regulators of transcription, I hypothesised that ZMYM3 might be involved in regulating a set of transcripts through these associations. To test this, I compare the transcriptomes of ZMYM3 knock-out and wild-type mouse embryonic stem cell lines by RNA-Seq. While deletion of ZMYM3 has no effect on transcription globally in this context, ZMYM3-null cells express lower levels of the primary transcript of microRNA 142, a specific target of the AGS-associated dsRNA deaminase ADAR1. To better understand the *in vivo* function of ZMYM3, an ES cell line with a ZMYM3 exon flanked by loxP sites is generated, to allow the conditional deletion of ZMYM3 in a whole organism. The association of deacetylase, demethylase and RNase H-type enzymes raises the possibility that histone modification and degradation of DNA-RNA hybrids may be coordinated, and is therefore of potential importance for the field of chromatin biology. As the roles of these associations remain unknown, several functional models of these interactions are proposed, with a discussion of the possible reasons why disruption of these interactions may result in Aicardi–Goutières syndrome.Item Open Access The Logic of Innately Aversive Olfactory PathwaysMyers, PhilipAnimals must locate food to survive and appropriate sexual partners with whom to reproduce. At the same time they must avoid predators and pathogens that threaten their survival, or the survival of their offspring. Recognising salient environmental cues is therefore critical to fitness-related behaviours. In many cases, certain stimuli evoke stereotypical behavioural responses. These innate behaviours, in contrast to learned behaviours, are thought to be hardwired in the neural circuitry of an animal. How are these innate behavioural responses instantiated in the circuitry of the brain? We used the olfactory sensory circuit of a microbial odorant, geosmin, as a model to understand how an innately aversive olfactory signal is processed in the brain in order to produce an appropriate behavioural response. We used connectomics, functional imaging, and behavioural assays to determine the anatomy and function of this geosmin processing circuitry. Using behavioural assays, we show that the aversive response to geosmin is due to chemotaxis, and provide evidence of a post-mating switch in behavioural response to geosmin. We then show that geosmin is detected via a functionally segregated pathway in the sensory periphery until it reaches the central brain, where the signal diverges significantly, and use connectomics to show that this divergent connectivity is largely stereotyped. We then imaged the activity of one previously described third-order neuron downstream of the geosmin pathway to show that it integrates olfactory signals of similar ethological significance at its dendrites. We used connectomics to follow this circuit deeper into the brain, and discovered a previously unknown descending neuron that senses geosmin, and is required for the innate aversion to geosmin exhibited during egg-laying. Connectomics analysis showed that this cell converges with other descending neurons in the ventral nerve cord of the fly that are known to induce escape behaviours. We further find evidence of a learned innate interaction via memory input to this descending neuron, implicating this circuit in aversive memory recall. These findings extend our understanding of how aversive signals are processed by the brain by providing a functional and connectomic characterisation of an olfactory circuit all the way from the brain to the nerve cord.Item Embargo Structural and Biochemical Investigation of Fanconi Anemia Pathway ActivationSijacki, TamaraDNA interstrand-crosslinks (ICLs) are covalent links between complementary DNA strands that prevent their separation, interfering with fundamental cellular processes such as DNA replication and transcription. Consequently, these DNA lesions are highly toxic and the inability to repair them results in a severe genetic disease, Fanconi Anemia (FA), manifested by developmental impairment, bone marrow failure, and predisposition to various types of cancers. In healthy cells, a specialized cascade of DNA repair proteins comes together to establish the FA pathway that specifically recognizes and repairs DNA ICLs. Ubiquitination of the FANCD2-FANCI (D2-I) complex by a multi-subunit ubiquitin E3 ligase, the FA core complex, is a key step of the FA pathway. D2-I ubiquitination initiates ICL repair by recruiting endonucleases to remove the DNA lesion, which is subsequently repaired via homologous recombination and trans-lesion synthesis. Due to its immense importance in the FA pathway, D2-I ubiquitination has been thoroughly studied using genetic, cellular, biochemical, and structural tools. FANCI phosphorylation by the ATR DNA damage kinase stimulates D2-I ubiquitination in cells, which enables timely and coordinated FA pathway activation to ensure DNA repair fidelity. In this dissertation, I investigated the mechanism of this process by employing a triple-phosphomimetic FANCI. First, I confirmed the stimulatory effect of phosphomimetic FANCI on D2-I ubiquitination through in vitro ubiquitination assays using purified proteins. To elucidate the structural basis of this stimulation, I visualized different states of phosphomimetic D2-I using cryo-electron microscopy (cryo-EM). This revealed that phosphomimetic D2-I securely closes around double stranded DNA. Upon closure, target ubiquitination sites on D2-I become accessible, explaining how FANCI phosphorylation promotes D2-I ubiquitination. I further explored how phosphomimetic mutations affected biophysical properties and dynamics of the D2-I complex. Strikingly, while phosphomimetic mutations did not significantly alter DNA binding, in vitro ubiquitination assays suggested that phosphomimetic D2-I closes more readily in solution even in the absence of DNA. This is not due to increased binding of the phosphomimetic D2-I to the FA core complex. Instead, using quantitative crosslinking mass spectrometry, I discovered that phosphomimetic mutations locally alter the D2-I dimerization interface to prime the complex for closure. Thus, phosphomimetic mutations shift the conformational equilibrium of D2-I towards a closed state. Moreover, I observed that in the presence of DNA, phosphomimetic and wild-type D2-I exist in both open and closed states, however phosphomimetic D2-I has a higher preference for the closed conformation. Overall, the work presented in this thesis provides mechanistic insight into how an interplay between DNA binding and FANCI phosphorylation by the ATR kinase alters the conformational dynamics of the D2-I complex, promoting its closure and ubiquitination to activate the FA pathway.Item Open Access Encoded synthesis and evolution of clinically approved 2'-modified ribonucleic acidsFreund, NiklasXeno-nucleic acids (XNAs) are unnatural nucleic acids with altered sugar, phosphodiester backbone, or nucleobase components. 2'-modified nucleic acid chemistries, such as 2'-O-methyl-RNA (2'OMe-RNA) and 2'-O-2-methoxyethyl-RNA (MOE-RNA), have shown promise in clinical applications due to their enhanced properties like high antisense-binding affinity and increased biostability. However, conventional phosphor-amidite synthesis for these 2'-modified nucleic acids is limited in length, prevents parallel exploration of different sequences and modifications, and precludes evolution. This study aims to develop an encoded enzymatic synthesis approach using polymerase engineering to enable the efficient synthesis of 2'OMe-RNA and MOE-RNA. This approach opens up possibilities for the evolution and selection of modified aptamers and nucleic acid enzymes with therapeutic potential. First, I describe how structure-guided engineering on a thermophilic archaeal polymerase led to the identification and mutation of a two-residue nascent-strand steric gate near the active site. This modification alleviated steric clashes within the polymerase, enabling processive synthesis of 2'-modified nucleic acids, including mixed-chemistry 2'OMe-/MOE-RNA aptamers, and unlocking their evolution. Furthermore, a reverse transcriptase-free selection procedure was established for MOE-RNA aptamer selections, enabling re-selection of a 2'OMe-RNA VEGF aptamer in the MOE-RNA chemistry. Additionally, procedures for the de novo selection of Tau-binding RNA, 2'OMe-RNA, and MOE-RNA aptamers were established. These selection procedures involved pre-enrichment over four rounds, followed by deep screening using an Illumina HiSeq instrument. By advancing these aptamer selections further, this study lays the groundwork for the development of the first MOE-RNA aptamer and demonstrate the potential of encoded synthesis of 2'-modified nucleic acids in creating biostable aptamers using nucleic acid chemistries approved for human use.Item Embargo Mechanisms underlying substrate engagement by TRIM7 and TRIM21 and strategies for re-directing their E3 ligase activityLuptak, Jakub; Luptak, Jakub [0000-0002-9527-8755]The work presented in this thesis focuses on two members of the TRIM protein family: TRIM7 and TRIM21. Both proteins are ubiquitin E3 ligases that ubiquitinate and degrade substrates. My work on TRIM7 has uncovered a unique molecular mechanism of substrate selection and explains why TRIM7 is reported to interact with unrelated and diverse substrates, ranging from cellular proteins involved in metabolism to viral proteins. In summary, TRIM7 recognises a dipeptide motif at the C terminus of substrates, so if this sequence is exposed in the folded protein and the substrate is oligomeric, this results in substrate engagement in cells and subsequent degradation. The function of TRIM21 as a cytosolic antibody receptor has been firmly established – it can recognise antibody coated viruses that have managed to enter the cytoplasm and mediate their degradation. This feature of TRIM21 has also been exploited to develop a technology called Trim-Away, where antibodies against endogenous proteins can be introduced to the cytosol of cells, resulting in targeted degradation of these proteins. While it is understood that the RING-E3 domain of TRIM21 is essential for this degradation, along with a number of E2 enzymes, how and why the formation of a TRIM21:antibody:substrate ternary complex results in the degradation of each component is incompletely understood. I have developed and applied *in vitro* assays to monitor ubiquitination of proteins involved in physiological TRIM21 function and in Trim-Away experiments. In both cases, the role of N terminal ubiquitination is explored. Finally, I have developed degrader molecules based on the mechanistic rules underpinning TRIM21 activation. These molecules can be further used as tools for studying endogenous TRIM21 function and potentially developed into therapeutics.Item Embargo Characterization of endocytic escape pathway and its role in adaptive immunityKrawczyk, Patrycja AnnaIn dendritic cells, endocytic escape appears critical for the induction of cytotoxic T-cell responses against viruses and tumors, in a process termed cross-presentation. During cross-presentation, dendritic cells present exogenous antigens on MHC class I, and the predominant molecular model for it is one where exogenous antigens are first released from endosomes into the cytosol, and further processed like endogenous proteins for presentation on MHC class I. Endocytic escape of antigens appears to be the rate-limiting step of cross-presentation, yet we know very little about the molecular mechanisms involved. To identify the machinery involved in endocytic escape of proteins, the Kozik lab has established a genetic screening strategy where escape of a ribosome-inactivating protein, called saporin, is being followed. When saporin enters the cytosol, it causes translational arrest quantified using puromycin labelling of newly synthesized polypeptides. Using the saporin assay, the Kozik lab performed CRISPR/Cas9 genetic screen targeting 200 candidate genes. The screen yielded one strong hit: pore-forming protein, perforin-2 (*Mpeg1*; Macrophage-expressed gene 1) and we proposed perforin-2 forms pores in endolysosomal compartments, allowing for endocytic escape of saporin. The main aim of my PhD was to extend our understanding of mechanisms governing endocytic escape through perforin-2, to identify other members of endocytic escape pathway, and to study their role in cross-presentation. We demonstrated that perforin-2 activity might be regulated by proteolytic cleavage and set out to identify the protease involved. The cleavage depends on low pH and can be stimulated by a range of pathogen-associated molecular patterns. To identify the new players of the endocytic escape pathway, we employed a genome-wide CRISPR/Cas9 genetic screening approach, where we followed endocytic escape of either saporin or of another translation inhibitor, a small peptide AGAAMSH (here named Pep). The screen confirmed that *Mpeg1* is a main regulator of endocytic escape of saporin in MutuDCs, and we identified and validated multiple potential *Mpeg1* regulators, including proteins involved in trafficking and posttranslational modifications. Interestingly, the screen with use of Pep as a translation inhibitor identified largely different hits. Many of the negative regulation hits were mitochondrial proteins, especially complex I related, which might indicate the involvement of mitochondrial ROS in endocytic escape of Pep. Finally, we went to address whether any of the validated hits indeed play a role in cross-presentation. Using the B3Z T cell hybridoma-based cross-presentation assay, we did not observe defects in cross-presentation in *Mpeg1* knockouts, however, knockout of a negative regulator of saporin escape, *Sppl2a*, seemed to increase cross-presentation of soluble ovalbumin, whereas knockout of a negative regulator of Pep escape, *Chchd5*, might enhance cross-presentation of immune complexes of ovalbumin. We hope that further characterization and cross-presentation studies will shed more light on the role of validated hits in the endocytic escape pathways, and on how different endocytic escape pathways contribute to adaptive immunity.Item Embargo Quality control of protein complex assemblyCarrillo Roas, SaraAll cells contain numerous multi-subunit protein complexes of exceptional abundance and physiological importance. Prominent examples include the ribosome, proteasome, chaperonin, tubulin, haemoglobin, and many others. A major challenge for the cell is to synthesise individual subunits in the correct stoichiometry. Any imbalance in subunit synthesis results in excess unassembled “orphan” subunits lacking their interaction partner(s). Orphans are accentuated in disease states such as cancer where altered gene expression caused by aneuploidy is common. Acute aneuploidy is known to cause proteotoxicity, probably as a consequence of high orphan burden. How orphan subunits are targeted for degradation to maintain protein homeostasis remains an unanswered question for most of the cell’s major protein complexes. The work described in this thesis investigated how the seven beta subunits of the 20S core proteasome (named PSMB1-7) are targeted for degradation when produced in excess. I used a mammalian *in vitro* translation system to identify candidate interactors of nascent PSMB subunits and found several proteins related to the ubiquitin-proteasome system. Using fluorescent reporters of PSMB subunit degradation in cells depleted of individual candidates, I found that the E3 ubiquitin ligases KCMF1 and UBR4 are required for efficient degradation of unassembled PSMB subunits. Biochemical assays showed that KCMF1 and UBR4 form a stable complex that physically interacts with nascent PSMB subunits. Moreover, using a cell-based flow cytometry assay, I determined which domains of KCMF1 and UBR4 are required for effective degradation of PSMB subunits. Analysis of orphans from other protein complexes showed that the KCMF1-UBR4 complex was also required for efficient degradation of PSMC5 (a 19S proteasomal subunit) and CCT4 (a chaperonin subunit). This was unexpected because orphaned PSMC5 and CCT4 were recently shown to be multiply mono-ubiquitinated by the E3 ubiquitin ligases HERC1 and HERC2, respectively. This led to the hypothesis that the KCMF1-UBR4 complex might build ubiquitin chains on pre-ubiquitinated orphans, thereby improving their recognition by the proteasome. *In vitro* reconstitution of orphan ubiquitination by the KCMF1-UBR4 complex showed that substrate pre-ubiquitination is both necessary and sufficient to create a KCMF1-UBR4 substrate. This reconstitution system revealed that the KCMF1-UBR4 complex builds K48-linked polyubiquitin chains and helped to define the key domains on both ubiquitin and UBR4 involved in substrate recognition. UBR4 mutants impaired in substrate recognition *in vitro* were also impaired in substrate degradation in cell-based assays. These results support a model in which the KCMF1-UBR4 E3 ligase complex acts as a chain elongator for orphaned subunits that are mono-ubiquitinated by a priming ligase. Elongation by the KCMF1-UBR4 complex generates K48 polyubiquitin chains that are ultimately required for efficient degradation by the proteasome. The KCMF1-UBR4 complex may therefore represent a general quality control pathway for the degradation of many types of orphan subunits. Consistent with this model, a wide range of cancer cells are observed to be strongly reliant on KCMF1 and UBR4 for their fitness.Item Embargo Crosstalk between natural killer cells and cancer associated fibroblasts in the tumour microenvironmentNunes Rodrigues, LeonorCancer is a deadly disease where malignant cells proliferate and avoid elimination by the immune system. Although natural killer (NK) cells, of the innate immune response, recognise and eliminate cancer cells, their cytotoxic function is often impaired in the tumour microenvironment. Indeed, difficult to treat solid cancers, like pancreatic cancer, are rich in cancer associated fibroblasts (CAF) which support tumour growth and suppress immune cell functions. Overcoming barriers like CAF-derived immune suppression in the solid tumour microenvironment remains a challenge. Successful restraint of CAF has proven challenging due to their complex intratumoral heterogeneity and plasticity. However, the potential exploitation of NK cell cytotoxicity remains attractive in the field of cancer immunotherapy. This work aimed to investigate the crosstalk between CAF and NK cells and to identify potential targets in solid tumours, thus improving NK cell cytotoxicity and reducing tumour growth. Initial flow cytometry and imaging analysis of mouse models of solid tumours revealed NK cell dysfunction within the tumours, characterised by an immature and less cytotoxic profile as compared to more mature blood circulating NK cells. Moreover, within the tumours, NK cells interacted closely with CAF defined by expression of platelet derived growth factor receptor alpha (PDGFR-⍺), while being distant from myofibroblastic CAF. Investigation of the roles of distinct CAF populations *in vitro* and *ex vivo* by flow cytometry and quantitative PCR, revealed that TGF-β-driven CAF (myofibroblastic) were strong NK cell modulators, in contrast to inflammatory CAF and healthy fibroblasts. Furthermore, *in vitro* assays revealed that pancreatic CAF-derived factors suppressed NK cell functions, in contrast to healthy fibroblasts. Comparative mass spectrometry and ELISA revealed CAF increased secretion of extracellular matrix proteins osteopontin, laminin and perlecan, but also the lipid metabolite prostaglandin E2 (PGE2). Blocking CAF-derived PGE2 *in vitro* restored NK cell functions such as tumour killing capacity and granzyme B secretion. *In vivo*, targeting this crosstalk by blocking PGE2 receptors EP2 and EP4 with antagonists reduced tumour growth and restored NK and T cell derived granzyme B. This work revealed unknown and contrasting roles for CAF populations in NK cell anti-tumour function, suggesting that targeting myofibroblastic CAF and derived factors like PGE2 could be relevant to improving intratumoral NK cell cytotoxicity in difficult to treat cancers such as pancreatic cancer.Item Open Access Interaction between form and fate during human neural developmentChiaradia, IlariaGenetics instructs on cell differentiation. Controlled cell differentiation over time and space determines the formation of complex tissue structures. Tissue architecture is permissive to tissue function. Little is known about whether tissue architecture influences cell fate acquisition. We sought to investigate the interaction between shape and fate during neural development. We relied on cerebral organoids as a tractable *in vitro* system where tissue shape can be manipulated following several complementary approaches and tissue identity assessed by transcriptomics. We first discovered key protocol variables that contribute to differences in organoid morphology. We designed a pipeline to quality- screen organoids based on their morphological features. Morphology alone can predict organoid cytoarchitecture and transcriptional proximity to the *in vivo* developing brain. Organoids with complex morphology display more and larger ventricles with thicker progenitor layer. They better resemble the transcriptome of foetal brain. Organoids with poor morphology and cytoarchitecture display an aberrant cell repertoire with fewer basal progenitors and upper-layer neurons. They fail to undergo proper developmental trajectories of events. Neurons maintain the signature of progenitors. Progenitors express neuronal genes early on. We observed a similar dysmaturity as in cortical dysplasia. We perturbed cytoarchitecture by mechanical dissociation-reaggregation of the organoid. We achieved a controlled morphology by encapsulating the organoid in hydrogel. Both perturbation paradigms resulted in aberrant cytoarchitecture and cells with a poised neuron/progenitor identity. We demonstrated that cell positioning influences cell fate. Next, we showed neural network dynamics in mature organoids. Organoids with poor morphology display isotropic neural networks. Overall, we demonstrated how shape influences fate during neural development.Item Open Access Mechanisms of TRIM21-directed intracellular degradationRhinesmith, TylerThe cytosolic antibody receptor TRIM21 has been studied for more than a decade for its role in mediating antibody-dependent neutralization of nonenveloped viruses. The presence of antibodies in the cytosol is abnormal—by binding them and initiating inflammatory signaling cascades, TRIM21 acts as a danger sensor and intracellular effector of humoral immunity. In addition to antibody binding, TRIM21 is a ubiquitin E3 ligase, an enzyme which catalyzes the formation of polymers of the protein ubiquitin around target substrates. When viruses coated in non-neutralizing antibodies escape into the cytosol, TRIM21 binds and decorates the complex with these polyubiquitin chains. Polyubiquitin is a highly-conserved protein post- translational modification with a number of important roles, including the initiation of degradative processes and antiviral genetic programs. Polyubiquitin-coated TRIM21:antibody:virus complexes are degraded by the cell, thus neutralizing viral infection. Its ability to bind generic immunoglobulins has allowed TRIM21 to be developed into the targeted protein degradation technology Trim-Away, whose intracellular mechanism is thought to mirror that of antibody-coated virions. Prior to the work presented here, the specific mechanistic steps following TRIM21-antibody binding had been established in some detail, and were believed to involve the sequential action of ubiquitin E2 conjugating enzymes UBE2W and UBE2N, which cooperated with TRIM21 to catalyze the formation of a specific type of linear polyubiquitin chain linked through lysine 63 on ubiquitin. This was thought to lead to the recruitment of the segregase p97/VCP and the 26S proteosome which together disassembled and degraded the encapsidated virion. In this thesis, I describe my efforts to define the sequential enzymatic mechanism of TRIM21 degradative signaling in greater detail. I report data from a series of biochemical and in-cell experiments which are not consistent with a role for UBE2W in polyubiquitin chain priming on TRIM21 during adenoviral neutralization. Then, I detail my development of two proximity-proteomics screens and a whole-genome CRISPR/Cas9 knockout screen for TRIM21 interacting partners, in both Trim-Away and adenovirus neutralization contexts. Collectively, the data from these experiments led me to hypothesize that the TRIM21 mechanism is plastic, and sensitive to substrate features and intracellular context. The whole-genome screen and follow up validatory experiments were consistent with a new model for TRIM21-adenoviral neutralization—rather than processing through the proteasome, intracellular adenoviruses coated with antibodies are degraded by selective autophagy following polyubiquitin deposition by TRIM21. Taken together, this work represents a clear advancement in understanding how TRIM21 targets substrates for proteolysis, and opens several new lines of inquiry into how the protein affects adaptive immunity at the cellular level.Item Embargo Genome-Wide and Multi-Scale Mapping of DNA Supercoiling in Differentiating Human Stem CellsPerez, ConsueloThe molecular programme of every cell is encoded in its genome. Human cells must organise and compact their genetic material at multiple levels to accommodate the two-meter-long DNA within the confined nuclear space of six microns. Chromatin folding is not only a cellular need, but is also crucial for gene regulation, and coordinates the precise rewiring of gene expression programmes in every cell type during developmental transitions. Therefore, our genome works in a dynamic rather than a static manner, allowing feedback between function and structure. However, how exactly the genome reaches a functional equilibrium in a particular cell type and the mechanisms for maintaining such balance remain incompletely understood. Increasing understanding of the genome topology derived from mathematical descriptions, *in silico* and *in vitro* studies, to single-molecule approaches and functional experiments in cells, has revealed that the free energy stored as twist in the DNA molecule can be used to facilitate key biological processes. Accordingly, changes in the winding of DNA, known as supercoiling, were shown to impact gene expression and chromosome architecture. Nonetheless, its intrinsic dynamic and multifaceted nature, together with experimental limitations, have hindered the study of the timing and factors orchestrating supercoiling formation and dissolution across human chromosomes. This thesis explores the cellular processes responsible for introducing DNA supercoiling, and its remodelling by topoisomerase enzymes, and addresses how supercoiling underlies chromatin contacts and genome organisation during cell fate transitions. Initially, we mapped DNA supercoiling dynamics genome wide by adapting bTMP-seq (Naughton et al., 2013, Achar et al., 2020) to high-throughput sequencing. This approach offered increased resolution compared to previous studies and allowed us to interrogate the entire human genome. By then integrating this dataset with a chromatin conformation capture Hi-C map, we studied DNA supercoiling across multiple scales: from kilobase genes, through larger-scale topologically associating domains, to megabase epigenomic compartments. Crucially, we observed the formation of supercoiling domains throughout the genome and revealed their temporal remodelling due to transcriptional activity and topoisomerase-mediated relaxation across cell differentiation. In summary, our multi-omic study revealed widespread supercoiling dynamics in differentiating human stem cells, which likely contribute to the chromatin environment for gene regulation during developmental transitions, in conjunction with previously well-described chromatin features.Item Embargo Investigating tau propagation in situBehr, Tiana SophiaThe abnormal assembly of tau protein in neurons is the pathological hallmark of multiple neurodegenerative diseases, including Alzheimer’s disease (AD). In AD, tau assembly begins in the locus coeruleus and transentorhinal cortex, and progresses over years to connected regions in the limbic system and neocortex. Cellular and in vivo models of tau assembly suggest that prion-like propagation may underlie this progression, including the intracellular trafficking, intercellular transfer and seeded aggregation of assembled tau. However, the underlying molecular mechanisms are poorly understood. The goal of my thesis was to investigate tau propagation in situ. I focussed on characterising the ultrastructural context of assembled tau filaments at different stages of propagation, including their subcellular localisation and molecular interactions. To achieve this, I employed electron cryo-tomography (cryo-ET), a powerful technique that allows the visualization of biological samples at molecular resolution in near-native conditions. I wanted to study tau propagation in a disease relevant context. I first focused on human tissue and then developed a neuronal cell model of propagation to enable further investigation. In the first chapter of my thesis, I investigated the association of assembled tau with brain extracellular vesicles (EVs). This association has been documented in AD and has been linked to the clearance and propagation of assembled tau. However, the molecular species of assembled tau and how they associate are not known. By performing cryo-ET on EVs isolated from AD patient brain, I discovered tau filaments enclosed within the lumen of large EVs. I also observed multiple novel molecular interactions of filaments, including molecules that tether filaments to the EV limiting membrane. These findings suggest that tau filaments are selectively packaged within EVs in AD. Using single particle electron cryo-microscopy (cryo-EM), I found that the filaments contained distinct anionic molecules compared to tau filaments within intracellular inclusions. These results reveal the molecular species and structures of assembled tau associated with brain EVs in AD, as well as how they associated with EVs. This will guide future investigations of the mechanisms of assembled tau secretion, and inform biomarker and therapeutic strategies targeting extracellular tau. In the second chapter, I investigated tau assembly at the synapse in AD. Synaptic dysfunction is the earliest manifestation of neurodegeneration. Evidence suggests that this may be driven by tau assembly directly at the synapse. How assembled tau contributes to synaptic dysfunction is not known. In addition, synapses may mediate the intercellular transfer of assembled tau during propagation. I analysed synaptosomes isolated from AD patient brain. Using cryo-ET, I found tau filaments in the cytoplasm of pre-synapses. The filaments associated with one another, as well as with vesicles. The filaments were also tethered to the plasma membrane and were decorated with globular densities, similar to the interactions that I observed in brain EVs. These results provide insights into disease mechanisms of tau assembly and guide future studies to ultimately understand synapse dysfunction in AD. In the third chapter of my thesis, I established a neuronal cell culture model of tau propagation compatible with in situ studies using cryo-ET. Observations made in AD patient brain tissue relate directly to disease. However, they only provide static snap shots of tau assembly and propagation, whereas these are dynamic processes. Therefore, a model system that recapitulates tau propagation is important to further investigate observations made in human brain. I induced tau propagation in primary mouse hippocampal neurons by incubation with purified tau filaments from AD patient brain. I introduced strategies for the fluorescent labelling of tau filaments and for neuronal culture on EM grids in order to make this model system compatible with fluorescent live cell imaging, correlative light and electron cryomicroscopy and cryo-ET. These results lay the foundation for future in situ studies of dynamic propagation events. During the course of this thesis, I established cryo-ET of human brain tissue to investigate the subcellular localisation of assembled tau and its molecular interactions at different stages of propagation. This revealed tau filaments within EVs and pre-synapses, and identified their tethering to the limiting membranes of these compartments. I then developed a model system to enable in situ studies of dynamic tau propagation events. This work sets the foundations for future studies of the molecular mechanisms of tau assembly and will guide therapeutic strategies and biomarker development to target assembled tau in AD.Item Embargo Identifying regulators of natural killer cell development through CRISPR screening of common lymphoid progenitorsMurphy, JaneNatural Killer (NK) cells are cytotoxic lymphocytes that perform a key role in the innate immune response, lysing abnormal host cells such as those that are cancerous or virally infected. They are the prototypic member of the innate lymphoid cell (ILC) family, which also comprises the functionally divergent ILC1s, ILC2s, ILC3s and lymphoid tissue inducer (LTi) cells. Most NK cells arise from progenitors in the bone marrow (BM), from which other ILCs are also derived. However, the precise signals that specify the commitment of ILC progenitors to the NK cell lineage are poorly characterised. The aim of this thesis was to study the transcriptional networks and BM niche factors that promote NK cell development from uncommitted BM progenitors, using the mouse as a model organism. Pooled CRISPR screens were performed by culturing CRISPR-edited primary common lymphoid progenitors in vitro to generate NK cells and ILC2s, so that the regulators unique to each lineage could be discovered. Reassuringly, this approach identified previously studied regulators of NK cell development such as *Nfil3*, *Runx3* and *Eomes*. Additional putative regulators, such as *Tcf3* (encoding E2A), *Rere*, *Bach1* and *Zfp652* were identified and validated using arrayed CRISPR screens. Due to the known role of *Tcf3*/E2A in the decision between the innate and adaptive lymphocyte lineages, this was an attractive target for further study. Using a novel *Tcf3FLAG* mouse strain, E2A protein was detectable in progenitors but not mature ILC subtypes. However, transcriptional analysis revealed that *Tcf3* was expressed by mature NK cells so a conditional knockout of *Tcf3* in NK cells/ILC1s was generated and characterised. Fewer NK cells in these mice produced effector molecules such as granzyme B, and mice challenged with B16F10 melanoma cells had more lung metastasis. Finally, the developmental niche of ILCs was studied using immunohistochemistry of bone cryosections, and the role of Notch signalling as a potential regulator of E2A was studied. Overall, the work presented in this thesis adds to our knowledge of innate lymphopoiesis, with potential implications for how this might be distorted in systemic diseases such as cancer or targeted therapeutically.