Vacuolar-type ATPases (V-ATPases) are ubiquitous membrane-bound protein complexes present in the endo-membrane system of all eukaryotic cells. In eukaryotic cells, the reversible dissociation of the V1 and Vo regions is an essential mechanism for regulating V-ATPase activity. Therefore, knowledge of the structure of the dissociated V1-ATPase is necessary for understanding the regulation of V-ATPase activity. In this thesis, I showed that by introducing a 3xFLAG tag at the C terminus of different V1-ATPase subunits, highly purified V1-ATPase complex could be isolated. Electron cryomicroscopy (cryo-EM) was used for initial analysis of the intact V1-ATPase. In addition to the intact complex, partial V1-ATPase subcomplexes with different subunit compositions were isolated from yeast cells in late log phase. All of the isolated subcomplexes were found to contain the major V1-ATPase subunits A and B, but differed in the peripheral stalk subunit composition.

Vacuolar-type ATPases (V-ATPases) are ubiquitous membrane-bound protein complexes present in the endo-membrane system of all eukaryotic cells. In eukaryotic cells, the reversible dissociation of the V1 and Vo regions is an essential mechanism for regulating V-ATPase activity. Therefore, knowledge of the structure of the dissociated V1-ATPase is necessary for understanding the regulation of V-ATPase activity. In this thesis, I showed that by introducing a 3xFLAG tag at the C terminus of different V1-ATPase subunits, highly purified V1-ATPase complex could be isolated. Electron cryomicroscopy (cryo-EM) was used for initial analysis of the intact V1-ATPase. In addition to the intact complex, partial V1-ATPase subcomplexes with different subunit compositions were isolated from yeast cells in late log phase. All of the isolated subcomplexes were found to contain the major V1-ATPase subunits A and B, but differed in the peripheral stalk subunit composition.

In eukaryotic translation initiation, eIF2B, a 295 kDa multisubunit (from α to ε) complex,is the guanine nucleotide exchange factor (GEF) of eIF2, a GTP binding protein, and hasmultiple roles in regulating the level of active eIF2-GTP-Met-tRNAi ternary complexes inthe cytoplasm. Mutations in eIF2B subunits affect global protein synthesis and, in human,are responsible to cause a genetically inherited lethal childhood brain disease calledLeukoencephalopathy with Vanishing White Matter (VWM). Although the genetic aspectseIF2B have been widely studied over decades, detailed structural knowledge only becameavailable in recent years but is still limited. This study aims to gain structural insights intoyeast eIF2B by a range of electron microscopy techniques to improve our understandingtowards its GEF activity with eIF2 and regulatory response. By performing size-exclusion chromatography and multi-angle static light scattering (SECMALS), it was found that eIF2B is a stable dimer of pentamers (~600 kDa). Negativestaining (25.8 Å) and cryo-EM (12.1 Å) eIF2B decamer models that showed 2-foldrotational symmetry were generated by single particle reconstruction. Homology modelingof yeast eIF2B subunits revealed an eIF2B(αβδ)2 hexameric core and two separate arm-likeeIF2Bγε catalytic domains with potential flexibility. To constrain subunit position in thearm structure, Ni-NTA-Nanogold labeling against the multihistidine tag of eIF2Bγ wasperformed. In addition, genetic approaches were applied to eliminate synthesis of eIF2Bα(34 kDa) and eIF2B(βγδε)2 octamer complexes (532 kDa) were purified by SEC-MALSand analysed by negative staining single particle reconstruction. It was speculated thatdeletion of eIF2Bα might have triggered significant conformational rearrangement that ledto high uniformity in the 2D class averages. A hypothetical model was thus proposed forthe octamer where the two arm-like domains clamp together to form a compact structure.

Antibiotic resistance in bacteria is becoming a major threat to public health. Many of the antibiotics used today in the clinic target the process of protein synthesis performed by the ribosome. Recent prospects for blocking ribosome function are increasingly focusing on preventing the assembly of bacterial ribosomes. A number of ribosome assembly factors are emerging as attractive targets for novel antibiotics that work in new ways. YphC and YsxC are essential GTPases in Bacillus subtilis that facilitate the assembly of the 50S ribosomal subunit; however, their roles in this process are still uncharacterized. To explore their function, we biochemically and structurally characterized the 45SYphC and 44.5SYsxC precursor particles accumulated from strains depleted of YphC and YsxC, respectively. Quantitative mass spectrometry analysis and 5-6 Å resolution cryo-EM maps of the 45SYphC and 44.5SYsxC particles revealed that the two GTPases participate in maturation of functional sites of the 50S subunit. We also observed that YphC and YsxC bind specifically to the two immature particles. In addition, we characterized the structure of the 50S subunits in complex with the RbgA protein. The preliminary 3D structure shows that the RbgA protein binds to the P site of the 50S subunit and displaces h69. There are also missing densities in the structure for h68 and the uL16 ribosomal protein. We expect that the atomic resolution structure of the 50S.RbgA complex will reveal the function and molecular mechanisms of this assembly factor. The deep structural understanding of protein synthesis process done by the ribosome led to the optimization of over a hundred antibiotics that are currently used in thev clinic. In the same manner, work described in this thesis provides novel insights into understanding the maturation of the large ribosomal subunit, and is paving the way to use the bacterial ribosome biogenesis pathway as a target for the development of new antimicrobials. / Thesis / Doctor of Philosophy (PhD)

Helical assemblies of proteins are ubiquitous in nature and they perform vital functions in a wide range of organisms. The recent development of direct electron detectors and other imaging techniques in cryo-electron microscopy (cryo-EM) has opened new possibilities in solving helical structures at atomic resolution. Existing software packages for helical processing often require experience in tuning many ad hoc parameters to achieve optimal reconstruction results. REgularised LIkelihood OptimisatioN (RELION), an open-source single-particle analysis package, reduces the need for user expertise by the formulation of an empirical Bayesian framework, and has yielded some of the highest resolution density maps in recent years. Prior information about the helical assemblies can be conveniently incorporated into the statistical framework of RELION and thereby improves the helical reconstructions. This PhD thesis describes the development of a helical processing computation workflow with reduced user intervention in RELION. Chapter 1 introduces the theoretical basis of cryo-EM data acquisition and single-particle data processing, the concepts of helical symmetry, and a previously described method for iterative real-space reconstruction of helical assemblies, to which the RELION implementation bears resemblance. Chapter 2 discusses multiple adaptations to RELION that are necessary for helical processing. Key elements include the imposition and local refinement of helical symmetry, masks on helical segments and references, expressions of angular and translational prior information, manual and automated segment picking as well as initial model generation for helices. Calculations have been performed on four test data sets showing that the developed methods in RELION yield results that are as good as or better than alternative approaches for the tests performed. Chapter 3 describes the same methodology adapted to helical sub-tomogram averaging in RELION. Chapter 4 introduces the local symmetry option developed for special types of filaments with pseudo-helical symmetry. The concept can be extended to general single-particle analysis as well. Chapter 5 describes four helical structures determined in collaboration with other research groups using helical RELION for data processing. Chapter 6 concludes the thesis with a brief summary and future prospects.

The main objective of this work is to determine the atomic structure of mammalian respiratory complex I. Mitochondrial complex I (also known as NADH:ubiquinone oxidoreductase) is one of the central enzymes in the oxidative phosphorylation pathway. It couples electron transfer between NADH and ubiquinone to proton translocation across the inner mitochondrial membrane, contributing to cellular energy production. Complex I is the largest and most elaborate protein assembly of the respiratory chain with a total mass of 970 kilodaltons. It consists of 14 conserved ‘core subunits’ and 31 mitochondria-specific ‘supernumerary subunits’. Together they form a giant, Lshaped molecule, with one arm buried in the mitochondrial membrane and another protruding into the mitochondrial matrix. Here, a novel method for the purification of ovine (Ovis aries) complex I was developed and suitable conditions for cryo-EM imaging established, after extensive screening of detergents and additives. Cryo-EM images were acquired with the recently developed direct electron detector and processed using the latest software. This allowed the solution of the nearly complete atomic model of mitochondrial complex I at 3.9 Å resolution. The membrane part of the complex contains 78 transmembrane helices, mostly contributed by conserved antiporter-like subunits responsible for proton translocation. These helices are stabilized by tightly bound lipids (including cardiolipins). The hydrophilic arm harbours flavin mononucleotide and 8 iron–sulfur clusters involved in electron transfer. Supernumerary subunits build a scaffold around the conserved core, strongly stabilizing the complex. Additionally, subunits containing cofactors (NADPH, zinc ion and phosphopantetheine) may play a regulatory role. Two distinct conformations of the complex are observed, which may describe the active and deactive states or reflect conformations occurring during the catalytic cycle of the enzyme. Currently this is the most detailed model of this molecular machine, providing insight into the mechanism, assembly and dysfunction of mitochondrial complex I. It also allows molecular analysis of numerous disease-causing mutations, and so the structure may serve as a stepping-stone for future medical developments.

Cdc48A and its human homologue P97 are from ATPase family, which play a variety of roles in cellular activates and it has a crucial involvement in protein quality control pathways. It is best known for its involvement in endoplasmic reticulum associated protein degradation (ERAD), where it mediates the degradation of the aggerated or misfolded proteins by the proteasome. Considering the multiple functions of Cdc48A in many protein regulatory processes, it is a potential therapeutic target for neurogenerative diseases and cancer. Cdc48A polypeptide comprises N domain, followed by D1 and D2 domains respectively that are joined by linkers, whereas functionally it forms a homo hexameric complex. Since Cdc48A is from the ATPase family, it uses the ATP hydrolysis to generate a mechanical force with its co-factors to perform its functions. There are many cofactors that interact with Cdc48A and two of them are Ufd1-NpI4 which in turn interact with ubiquitinated proteins from the ER membrane. The mechanism linking the conversion of the energy of ATP hydrolysis into mechanical force and unfolding the substrate is vague. My aim is to reconstruct a first 3D- model of plant Cdc48A using single particle cryo-EM, which serves the basis to conduct more detailed mechanistic studies towards substrate unfolding and threading/unfolding in the future. In general, results showed one defined structure of cdc48A at ~ 9.8 Å, which is the ADP-ATP conformation. Although another other structure was also resolved at ~ 8.9 Å, it was hard to characterize due to its dissimilarity with known structures of Cdc48A homologues and thus requires further characterization.

Legionella pneumophila (Lp) is a gram-negative, intracellular parasite. These bacteria evade the host response with the help of a Type IVb secretion system (T4bSS), composed of Defective in organelle trafficking (Dot) and Intracellular multiplication (Icm) proteins. This secretion system delivers over 300 effectors into the host, and a large number of these molecules are dependent on IcmS and IcmW. These effectors are essential for the bacterium’s survival in the host. This work and previous studies have shown that IcmS and IcmW interact to stabilize each other and the C-terminal "tail" of DotL (residues 655-783), a coupling protein in the T4bSS, binds to the IcmSW complex to further increase its stability. All three components are α-helical, making the complex amenable to structural studies by X-ray crystallography and cryo-electron microscopy. Three maps of the IcmSW-DotL-tail complex (~42 kDa) were generated from cryo-EM images recorded with a Volta phase plate and K2 Summit direct electron detector at 500-1000 nm under-defocus. The final maps were processed with RELION-2 and resolved to 5.5-6.5 Å resolution using 57k, 60k and 80k particles, respectively. Concurrent with work in this thesis, a crystal structure of IcmSW-DotL(656-783) was solved by Dr. Byung-Ha Oh’s group at KAIST. This structure was used as a comparative model for our cryo-EM 3D reconstructions which were determined to evaluate size limits imposed on single particle methods with current technology and to provide snapshots of the complex in solution. Comparisons between the crystal structure and cryo-EM maps show that the overall structure is similar in solution, but there is significant flexibility within each subunit with a repositioning of some α-helices and surface loops. Flexibility in the absence of a central subunit (LvgA), and a low number of good particles may have limited the final resolution. Although the current maps were determined at α-helical resolution, this work provides a road map for solving near atomic structures at or near the size of IcmSW-DotL-tail. This structural technology will provide a means to probe the solution structure and function of biological machines in a large range of sizes and conformations.

Characterization of Rubisco plays a crucial role when it comes to the development and understanding of carbon sequestration in plants. This project took place at BMC in Uppsala, in the Gunn lab, and aimed to structure three Rubisco structures and analyze these with regard to the assembly pathway of the biogenesis of Rubisco but also how fast the reaction of binding of atmospheric carbon dioxide takes place with regard to different isoforms of the small subunit. The structural regulations led to the conclusion that an additional step in the assembly pathway would be added when one side of Rubisco had the chaperone BSD2 bound while the other side of Rubisco had the small subunit bound.The different subunits are believed to effect the structure of the LSu. The result also indicate that when the SSu are binding to the LSu octomer the interactions between the BSD2 and the LSu changes. This indicats that the SSu could indirectly facilitate the binding of the SSu on the other side by affecting the interactions of the LSu and the BSD2. Therefore the cooperative binding of the different subunits would be interesting to further evaluate. The NtL8B4(S-T1)4 , which is the first model for this structure to be determined, and therefore extended the assembly pathway for the biogenesis of higher plants, had the CABP bound, indicating that this intermediate structure could be analytically competent. This hypothesis is only based on the analyses of the structural determination, therefore further studies are needed to determine whether this is legitimate. Teknisk-naturvetenskapliga fakulteten, Uppsala universitet. Utgivni

Les bactéries doivent s'adapter rapidement aux modifications de leur environnement en ajustant leur modèle d'expression génétique et leurs activités enzymatiques. Dans la plupart des cas, les variations de leur habitat impliquent de petites molécules que les bactéries peuvent détecter et auxquelles elles peuvent réagir. Le ribosome, la machinerie de la cellule qui catalyse la formation de la liaison peptidique, est capable de détecter les métabolites ou les antibiotiques afin de réguler l'expression des gènes, où le peptide naissant au sein du ribosome est capable d’induire l’arrêt de la traduction. Dans ce mécanisme, le peptide en cours de traduction (peptide d'arrêt) bloque le ribosome en interagissant avec les parois du tunnel ribosomal correspondant à la cavité par laquelle le peptide atteint le cytoplasme. L'arrêt peut dépendre uniquement de la séquence du peptide ou bien nécessiter la liaison d’une petite molécule. L’arrêt du ribosome en cours de traduction contrôle à son tour l'expression sur le même ARNm d'un gène situé en aval. Malgré plusieurs études biochimiques et structurales antérieures, le mécanisme exact de détection de ces petits métabolites par le peptide d’arrêt est encore inconnu. Mon travail de doctorat a porté sur : (1) comprendre comment de petites molécules sont détectées par les peptides d'arrêt ribosomaux, et (2) un cas particulier d'arrêt de la traduction dépendant du ligand : la détection des antibiotiques par des peptides d'arrêt courts.Pour répondre au premier problème, j'ai étudié biochimiquement et structurellement un nouveau peptide d'arrêt (appelé SpeFL) qui détecte l’ornithine (un petit métabolite) et qui est codé en amont de l'opéron speF chez Escherichia coli. La structure cryo-EM que j'ai résolue a révélé comment l’ornithine est détectée de manière très spécifique par un complexe ribosomal en cours de traduction. De plus, j'ai montré que le mécanisme d'induction du gène en aval speF implique un arrêt du ribosome au niveau de speFL empêchant ainsi une terminaison prématurée de la transcription Rho-dépendante.Dans la deuxième partie de ma thèse, je me suis concentrée sur la façon dont un antibiotique ciblant les ribosomes, l'érythromycine, est détecté par un peptide d'arrêt court. L'érythromycine est capable de bloquer la traduction de manière séquence-dépendante, où le motif (+)X(+) est le motif principal de blocage. Des données biochimiques publiées antérieurement suggèrent que l'encombrement stérique et électrostatique causé par le premier acide aminé chargé positivement (+) empêche l'addition du second, arrêtant ainsi le ribosome en cours de traduction. La résolution de la structure cryo-EM d'un ribosome arrêté par un peptide MKFR en présence d'érythromycine suggère le contraire, ce qui ouvre la voie à d'autres recherches sur le sujet. / Bacteria need to rapidly adapt to the changing environment by adjusting their gene expression patterns and enzymatic activities. In most cases, the variations in their habitat involve small molecules that bacteria are able to sense and respond to. The ribosome, the machinery of the cell that catalyzes peptide bond formation, is able to detect metabolites or antibiotics to regulate gene expression via nascent-chain mediated translational arrest. In this mechanism, the peptide that is being translated (arrest peptide) stalls the ribosome by interacting with the walls of the ribosomal tunnel, the cavity through which it reaches the cytoplasm. The arrest may depend solely on the sequence of the peptide or need a small molecule to be triggered. Ribosomal stalling in turn, controls the expression of a gene that is located downstream on the same mRNA. Despite previous biochemical and structural studies, the exact mechanism of sensing of small metabolites by the nascent chain is still unknown. My PhD work focused on: (1) understanding how small molecules are sensed by ribosomal arrest peptides, and (2) a special case of ligand-dependent translational arrest: drug sensing by short arrest peptides.To address the first issue, I studied biochemically and structurally a novel L-ornithine sensing arrest peptide (SpeFL) encoded upstream the speF operon in Escherichia coli. The cryo-EM structure that I solved revealed how a small molecule is sensed by a ribosome nascent chain complex in a highly specific manner. Besides, I showed that the mechanism of induction of the downstream gene speF involves ribosomal arrest at speFL preventing premature Rho-dependent transcriptional termination.On the second part of my thesis, I focused on how a ribosome-targeting antibiotic, erythromycin, is sensed by a short arrest peptide. Erythromycin is able to block translation in a sequence dependent manner, with the (+)X(+) motif being the main stalling motif. Previously published biochemical data suggest that steric and static hindrance caused by the first positively charged amino acid prevents the addition of the second one arresting the ribosome. I solved the cryo-EM structure of a ribosome arrested by an MKFR peptide in the presence of erythromycin that shows otherwise and opens up further investigation on the matter.