Glycosidases hydrolyse the glycosidic bond in carbohydrates. Structural studies of three glycosidases with different substrate specificities are presented in this work. Dextranase catalyzes the hydrolysis of α-1,6-glycosidic linkage in dextran polymers. The structure of dextranase, Dex49A, from Penicillium minioluteum was solved in the apo-enzyme (1.8 Å resolution) and product-bound (1.65 Å resolution) forms. The main domain of the enzyme is a right-handed β-helix, which is connected to a β-sandwich domain at the N-terminus. Using NMR spectroscopy the reaction course was shown to occur with net inversion at the anomeric carbon. A new clan is suggested that links glycoside hydrolase (GH) families 28 and 49. Endo-β-1,4-D-mannanase catalyzes the depolymerization of β-1,4-mannan polymers. The structure of endo-1,4-β-mannanase Man5A from blue mussel Mytilus edulis has been determined at 1.6 Å resolution. Kinetic analysis of Man5A revealed that the enzyme requires at least 6 subsites for efficient hydrolysis. The architecture of the catalytic cleft differs significantly from other GH 5 enzyme structures. We therefore suggest that Man5A represents a new subfamily in GH 5. Both the Dex49A and the Man5A structures were determined by multiple-wavelength anomalous diffraction using the selenium K-edge with selenomethionyl enzymes expressed in the yeast Pichia pastoris. Endoglucanase Cel6A from Thermobifida fusca hydrolyzes the β-1,4 linkages in cellulose. The structure of the catalytic domain of Cel6A from T. fusca in complex with a non-hydrolysable substrate analogue has been determined to 1.5 Å resolution. The glycosyl unit in subsite –1 was sterically hindered by Tyr73 and forced into a distorted 2SO conformation. In the enzyme where Tyr73 was mutated to a serine residue the hindrance was removed and the glycosyl unit in subsite –1 had a relaxed 4C1 chair conformation.

Adenovirus E1A is needed to activate early viral genes and induce cell cycle progression to optimise the conditions for viral replication. This is mostly achieved through interactions between the first exon of E1A and cellular transcriptional regulatory proteins. The carboxy terminus of E1A binds the cellular corepressor of transcription C-terminal Binding Protein (CtBP), resulting in derepression of CtBP target genes. Inducible stable U2OS cell lines were established, expressing wild type E1A (E1Awt) and a mutant unable to bind CtBP (E1A∆CID). Low inducible levels and loss of protein expression after prolonged induction together with induction of apoptosis were consistent with the fact that wild type E1A is a cytotoxic protein and correlated with the ability of CtBP to repress proapoptotic genes. E1A∆CID did not induce apoptosis and could be expressed at high levels for prolonged time periods. Moreover, the binding of CtBP contributed to E1A-induced activation of viral E1B and E4 genes, through possible targeting of Sp1 and ATF transcription factors. In a micorarray study on mRNA levels in E1A-expressing cells, several genes consistent with the tumour suppressive and apoptotic properties of E1Awt were identified as differentially expressed. Furthermore, the differences between the two cell lines correlated with the presence of binding sites for CtBP-interacting transcription factors in the promoters of regulated genes, enabling the possible identification of new CtBP target genes. Finally, a molecular characterisation of the CtBP mechanism of repression revealed that positioning proximal to the basal promoter element was required for efficient repression, suggesting that CtBP interferes with the basal transcriptional machinery. Two separate domains were identified in CtBP, conferring transcriptional repression and activation when expressed alone, achieved through their interaction with HDACs and HATs, respectively. However, together they cooperate to ensure maximal repression through recruitment of histone deacetylase and inhibition of histone acetyl transferase activity. Together, these data shows important modulation of E1A activities by the binding of CtBP and suggests the involvement of acetylation/deacetylation complexes for the regulation of E1A function.

Understanding protein-ligand interactions is highly important in drug development. In the present work the objective is to comprehend the link between structure and function using molecular modelling. Specifically, this thesis has been focused on implementation of receptor flexibility in molecular docking and studying structure-activity relationships of potassium ion channels and their blockers. In ligand docking simulations protein motion and heterogeneity of structural waters are approximated using an ensemble of protein structures. Four methods of combining multiple target structures within a single grid-based lookup table of interaction energies are tested. Two weighted average methods permit consistent and accurate ligand docking using a single grid representation of the target protein structures. Quaternary ammonium ions (QAIs) are well known K+ channel blockers. Conformations around C–N bonds at the quaternary centre in tetraalkylammonium ions in water solution are investigated using quantum mechanical methods. Relative solvation free energies of QAIs are further estimated from molecular dynamics simulations. The torsion barrier for a two-step interconversion between the conformations D2d and S4 is calculated to be 9.5 kcal mol–1. Furthermore D2d is found to be more stable than the S4 conformation which is in agreement with experimental studies. External QAI binding to the K+ channel KcsA is also studied. Computer simulations and relative binding free energies of the KcsA complexes with QAIs are calculated. This is done with the molecular dynamics free energy perturbation approach together with automated ligand docking. In agreement with experiment, the Et4N+ blocker in D2d symmetry has better binding than the other QAIs. Binding of blockers to the human cardiac hERG potassium channel is studied using a combination of homology modelling, automated docking and molecular dynamics simulations. The calculations reproduce the relative binding affinities of a set of drug derivatives very well and indicate that both polar interactions near the intracellular opening of the selectivity filter as well as hydrophobic complementarity in the region around F656 are important for blocker binding. Hence, the derived model of hERG should be useful for further interpretations of structure-activity relationships.

Francisella tularensis, the causative agent of tularemia, is a potent human and animal pathogen. Initially upon infection of the host, intramacrophage proliferation of F. tularensis occurs but after activation of the acquired host immunity, the phagocytes become activated to kill the bacterium. In my thesis, I focused on mechanisms utilized by F. tularensis to survive intracellularly and on host mechanisms responsible for macrophage-mediated killing and control of infection. The F. tularensis-specific protein IglC has been previously shown to be essential to the intramacrophage proliferation and virulence of the bacterium in mice. By electron microscopy of macrophages infected with either the live vaccine strain of F. tularensis or an isogenic mutant, denoted ∆iglC, expression of IglC was found to be necessary for the bacterium to escape from the phagosome. IFN-g-activated macrophages significantly inhibited the escape of the live vaccine strain of F. tularensis from the phagosome. iNOS and phox generate NO and O2-, respectively. These molecules and their reaction products possess both bactericidal and immunoregulatory properties. We investigated the capability of IFN-g-activated peritoneal exudate cells from gene deficient iNOS-/- or p47phox-/- mice to control an intracellular F. tularensis LVS infection. iNOS was found to contribute significantly to the IFN-g induced killing, while phox contributed only to a minor extent. Unexpectedly, bacteria were eradicated even in the absence of both a functional phox and an active iNOS. The eradication was found to depend on ONOO-, the reaction product of NO and O2-, because addition of a decomposition catalyst of ONOO- completely inhibited the killing. Studies on iNOS-/- or p47phox-/- mice infected with F. tularensis LVS showed phox to be important during the first days of infection, a stage when iNOS seemed dispensable. Eventually, iNOS-/- mice died of the infection, suggesting a role of iNOS later in the course of infection. iNOS-/- mice exhibited elevated IFN-g serum levels and severe liver damage suggesting that the outcome of infection was at least in part related to an uncontrolled immune response. Several pathogenic bacteria express Cu,Zn-SOD, which in combination with other enzymes detoxifies reactive oxygen species produced by the host. A deletion mutant of F. tularensis LVS lacking the gene encoding Cu,Zn-SOD was attenuated at least 100-fold compared to LVS in mice. In peritoneal exudate cells from mice, Cu,Zn-SOD was found to be required for effective intramacrophage proliferation and, in mice, important for bacterial replication at the very early phase of infection. In summary, the most conspicuous findings were a capability of IFN-g activated macrophages to retain F. tularensis LVS in the phagosome, an essential role of ONOO- in intracellular killing of F. tularensis, and an importance of Cu,Zn-SOD to the virulence of F. tularensis LVS.

Mesenchymal-epithelial interactions are pivotal for proper pancreatic growth and development. The pancreatic progenitor cells present in the early pancreatic anlagen proliferate and eventually give rise to all pancreatic cell types. The Fibroblast Growth Factor 2b (FGFR2b) high-affinity ligand Fibroblast Growth Factor 10 (FGF10) has been linked to pancreatic epithelial cell proliferation and we have previously shown that Notch signalling controls pancreatic cell differentiation via lateral inhibition. By overexpressing FGF10 under the control of the Ipf1/Pdx1 promoter in mice, we have shown that persistent FGF10 activation in the embryonic pancreas of transgenic mice perturbs pancreatic epithelial cell proliferation and also inhibits pancreatic cell differentiation by maintaining Notch activation. In the Ipf1/Fgf10 transgenic mice, the pancreatic epithelial cells are ‘locked’ in an undifferentiated progenitor-like state with sustained proliferative capacity. Collectively, our data suggest a key role for FGFR2b/FGF10 signalling in the regulation of pancreatic growth and differentiation and that FGFR2b/FGF10 signalling interact with the Notch signalling pathway. Glucose homeostasis in mammals is critically dependent on co-ordinated glucose uptake, oxidative metabolism and insulin secretion in β-cells. Although, several key genes controlling various aspects of glucose sensing, glucose metabolism, insulin expression and secretion have been identified, we know relatively little about the molecular mechanisms that induce and maintain the expression of genes required for glucose-stimulated insulin secretion (GSIS) in β-cells. Attenuation of FGFR1c signalling leads to diabetes in mice. Overexpression of FGF2, a high-affinity FGFR1c ligand, under the control of the Ipf1/Pdx1 promoter also leads to diabetes in mice. The Ipf1/Fgf2 mice present with normal endocrine and exocrine differentiation but display impaired glucose-stimulated insulin secretion (GSIS), perturbed expression of genes required for glucose sensing uptake together with oxidative metabolism and increased expression of the FGF-signalling inhibitors Spry-2 and Pyst1/MKP3 in β-cells. Thus, stringent control of FGF signalling activation appears crucial for the maintenance of the regulatory circuit that ensures proper GSIS in pancreatic β-cells and hence normoglycaemia. The Wnt family of ligands via their receptors Frizzled (Frz) have been shown to mediate mesenchymal-epithelial interactions and cell proliferation in a variety of different systems. Expression of a plethora of Wnt ligands and Frz receptors has been previously reported in the pancreas and mice missexpressing Wnt1 and Wnt5a under the Ipf1/Pdx1 promoter display severely perturbed development. Here, we show the temporal and spatial expression of Wnt4, Wnt7b and Frz3 at different stages of pancreas development. To elucidate the role of Wnt signalling in the pancreas, we overexpressed a dominant negative form of mouse Frz8 under the Ipf1/Pdx1 promoter in mice. The Ipf1/Frz8CRD mice display severe pancreatic hypoplasia demonstrating that attenuation of Wnt signalling in the pancreas leads to perturbed pancreatic growth. Nevertheless, the transgenic mice present with normal endocrine and exocrine differentiation and remain normoglycaemic. The maintenance of normoglycaemia in these mice appears to be the consequence of a relative increase in endocrine cell number per pancreatic area combined with enhanced insulin biosynthesis and insulin secretion. Collectively our data provide evidence that Wnt signalling is required pancreatic growth but not adult β-cell function.

Many T=3 plant and insect viruses use a molecular switch in form of order/disorder of a segment of the polypeptide chain to regulate the quasi-equivalent contacts. The structure of a mutant of the T=3 capsid of bacteriophage fr confirms that this virus and other members of the Leviviridae family lack a switch mechanism. The geometric principles underlying the construction of spherical virus capsid do not allow more than 60 protein monomers to from a capsid while maintaining an identical chemical environment. Most virus capsid, however, contain many more protein subunits. Quasi-equivalence explains how the capsid proteins can have slightly different interactions in the virus shell. Quasi-equivalence requires the capsids to be constructed from multiples of 60 subunits, where the T number denotes the multiplicity. The structure of the T=4 Nudaurelia capensis ω Virus shows a molecular switch in form of a C-terminal helix inserted in some contacts between protein dimers. This virus is very similar in structure to the T=3 nodaviruses. In the nodaviruses a five-membered helix bundle, formed by cleaved peptides around the five-fold axes on the inside of the shell, are suggested to aid in membrane translocation of the genomic RNA. In Nudaurelia capensis ω Virus the helix bundle is formed by 10 helices, of which 5 are still covalently attached to the capsid proteins. Bacteriophage HK97 has T=7 quasi-symmetry. A domain that is degraded during maturation and is not present in the structure of the mature virion controls the quasi-equivalence. During maturation covalent bonds are formed between the protein subunits, producing a set of interlocking covalently bound rings, resembling chainmail. Structural studies of complexes between the bacteriophage MS2 and variants of its translational operator are also included in this work. A dimer of the MS2 coat protein binds with sequence specificity to an operator in its genomic RNA, and causes translational repression. Structures of multiple RNA segments with altered sequence at some positions which are required for binding to the capsid protein, has been determined.

Integrins are a large family of membrane-spanning heterodimeric (αβ) receptors that bind to ligands on other cells or to extracellular matrix (ECM) proteins. These receptors mediate bidirectional signaling over the cell membrane to induce signaling cascades mediating functions as cell adhesion, spreading and migration. This signaling takes place at cell-matrix adhesions, which are sites where clustered and ligand-bound integrins connect to and mediate stabilization of the actin cytoskeleton, and induce signaling cascades. Integrins have a short cytoplasmic tail that is crucial for the bidirectional signaling, and the β1-integrin subunit exists in five splice variants only differing in the membrane-distal part of the cytoplasmic tail. This region of the almost ubiquitously expressed β1-integrin, β1A, contains two protein tyrosine motifs (NPXYs) interspaced with a threonine-rich region, while this region of the β1B splice variant is completely different and lacks known motifs. In contrast to the β1A-integrin, the β1B variant cannot mediate cell-matrix adhesion formation following binding to ECM ligands. The enteropathogenic bacterium Yersinia pseudotuberculosis binds to β1-integrins on the host cell with invasin, and this stimulates uptake of the bacterium. However, upon binding to the host cell, pathogenic Yersinia strains inject virulence effectors that block uptake. One effector responsible for the blocking is a tyrosine phosphatase, YopH. We identified the targets for this effector in the macrophage-like cell line J774A.1, which represent a professional phagocyte and thus is the likely target cell for the antiphagocytic effect of Yersinia. Two YopH target proteins were p130Cas and ADAP, of which the latter interestingly is an adapter protein specifically expressed in hematopoietic cells. ADAP has previously been implicated to participate in Fc-receptor-mediated phagocytosis and in communication between T-cell receptors and integrins. We also studied the importance of the cytoplasmic tail of β1-integrin for uptake of Yersinia. The GD25 cell line, which is a fibroblast-like cell line that lacks endogenous β1-integrins, was used together with GD25 cells transfected with β1B, β1Α or cytoplasmic tail mutants of β1A. These studies revealed that β1B-integrins could bind to invasin but not mediate uptake of Yersinia, while β1A both bound to invasin and mediated uptake. The first NPXY motif (unphosphorylated) and the double-threonines of the unique part of β1A were important for the ability of integrin to mediate uptake of Yersinia. These studies lead to the interesting finding that, when these cells were allowed to spread on invasin, those that expressed β1A spread as normal fibroblasts while for β1B-integrin-expressing cells, only finger-like protrusions of filopodia were formed. This provided us with a tool to study formation of filopodia without interference of the tightly linked process of lamellipodia formation. Initially, proteins that localized to the tip complex of these filopodia were identified. These were talin, VASP and interestingly the p130Cas-Crk-DOCK180 scaffold, while FAK, paxillin and vinculin were absent. In addition, VASP, p130Cas and Crk were shown to be important for the filopodia formation in GD25β1B. Further, the role of the actin motor myosin X, which previously has been implicated in formation of filopodia, was studied in the GD25Β1B cells and it was shown that myosin X not was important for filopodia formation, but that it recruited FAK and vinculin to the tip complexes of filopodia.

During embryonic development, generation of functional organs is dependent on proper interactions between different cell types. Elucidation of the mechanisms operating during organ formation might provide insights into the origin of many pathological disorders in the adult. Gene inactivation studies in mice have provided invaluable tool to study the function of genes critical for morphogenesis of distinct organs. A LIM-homeodomain transcription factor Lhx2 has previously been reported to play a role in fetal liver development and hematopoiesis, as its inactivation leads to lethal anemia due to underdeveloped liver. This thesis focuses on the function of Lhx2 in the development of these two organ systems. Reciprocal signaling between ventral foregut endoderm and mesenchyme of the septum transversum regulates the liver formation, expansion and differentiation. A fully formed liver is composed of endoderm-derived hepatocytes and cholangiocytes and a variety of mesenchyme-derived cell types, such as endothelial cells and hepatic stellate cells. In early stages of liver development Lhx2 is expressed in the liver-associated septum transversum mesenchyme, a part of which becomes integrated into the liver organ and develops into hepatic stellate cells. Functional Lhx2 expression in the hepatic mesenchyme is necessary for normal liver outgrowth and differentiation. Loss of Lhx2 from developing hepatic stellate cells leads to their activation and excessive deposition of collagen fibres, resulting in hepatic fibrosis and severely distorted liver architecture. Transfection of Lhx2 to human stellate cell line downregulates genes associated with stellate cell activation and fibrogenesis. Thus, Lhx2 is the first gene identified to negatively regulate events leading to hepatic fibrosis. Elucidation of the molecular mechanisms involved in this process might therefore be instrumental for the development of novel therapies useful in treatment of this disorder. Fetal liver is also a major site of hematopoiesis in the embryo and provides physiological conditions necessary for the efficient expansion of hematopoietic stem cells (HSCs). The hematopoietic defect observed in Lhx2-deficient embryos is cell-nonautonomous, indicating that Lhx2 might control secreted factors involved in the self-renewal of HSCs. This putative second role of Lhx2 has been investigated by analyzing the mechanism whereby Lhx2 expression generates in vitro self-renewing HSC-like cell lines. Interestingly, in agreement with the cell nonautonomous phenotype of the lethal anemia in Lhx2-/- embryos, the mechanism of self-renewal is dependent on Lhx2 expression and occurs via secreted factor(s). Identification of these factor(s) might potentially allow ex vivo expansion of HSCs for therapeutic purposes. The Lhx2-immortalized HSC-like cell lines share many basic features with HSCs and self-renew in vitro in presence of Steel factor (SF). SF/c-Kit signaling mediates a wide variety of biological activities in cells at many different levels in the hematopoietic hierarchy. We used the HSC-like cell lines as an in vitro model system to compare signal transduction pathways from c-Kit receptor in stem cells versus differentiated hematopoietic cells. HSCs require PI-3K dependent activation of Raf1-Mek-Erk cascade for their survival and self-renewal in response to SF, whereas activation of Erk is PI-3K independent in committed myeloid and mast cells. Thus, the mode of SF/c-Kit signaling is dependent on the differentiation status of the cells.

Clathrin-mediated endocytosis is a process by which cells can internalise diverse molecules such as nutrients, antigens and signalling-surface receptors. The creation of clathrin-coated vesicles demands interplay between the plasma membrane lipids, cargo molecules and the proteins that build up the coat. This thesis deals with the identification and characterisation of sorting nexin 9 (SNX9) as a novel component of the endocytic machinery. SNX9 belongs to a large family of proteins based on the presence of a PX domain. In addition, SNX9 harbours an SH3 domain followed by a region with predicted low-complexity and a C-terminal BAR homology domain. Binding studies demonstrated that SNX9 interacted with the endocytic core components clathrin and AP-2 and dynamin-2, a GTPase known to be crucial for vesicle scission. The C-terminal region bound to phosphatidylinositols and targeted SNX9 to artificial liposomes and cellular membranes. Consistent with a role in endocytosis, a large portion of SNX9 co-localised with AP-2 and dynamin-2 but not with markers for early endosomes, Golgi. Over-expression of truncated variants of SNX9 in K562 and HeLa cells interfered with the uptake of transferrin. SNX9 recycles between a membrane-bound and a cytosolic pool. In cytosol, SNX9 formed a resting complex together with dynamin-2 and the metabolic enzyme aldolase. Activation for membrane binding involved ATP hydrolysis and correlated with phosphorylation of SNX9 and the release of aldolase. Aldolase bound to a tryptophan-containing acidic region near the clathrin and AP-2 motifs and blocked lipid binding of purified SNX9 derivatives. SNX9 was required for membrane targeting of dynamin2 in vitro and knockdown of SNX9 in HeLa cells by RNAi resulted in impaired membrane localisation. Together these results argue strongly for a role of SNX9 in recruiting and linking of dynamin-2 to sites of vesicle creation.

The fruit fly Drosophila melanogaster is an excellent model organism to study the innate immune response, because insects and mammals share conserved features regarding the recognition and destruction of microorganisms and Drosophila is easily accessible to genetic manipulation. In my present study, I identified a new family of pattern recognition molecules for bacterial peptidoglycan in Drosophila, the Peptidoglycan Recognition Proteins (PGRP). This family of proteins is widespread in the animal kingdom, for instance there are 4 PGRP genes in humans with unknown function. So far, all tested PGRPs (from insects and mammals) have been shown to bind peptidoglycan. In Drosophila, we found and characterized 13 PGRP genes, which fall into two classes: Short PGRPs and Long PGRPs. To the short group belong PGRP-SA, SB1, SB2, SC1A, SC1B, SC2, and SD with short transcripts and predicted extracellular proteins. The long members are PGRP-LA, LB, LC, LD, LE, and LF with long transcripts and predicted intracellular and membrane spanning proteins. Transcripts from the 13 different PGRP genes are present in immune competent organs, and the majority are inducible by infection. The transcriptional regulation of the inducible PGRP genes occurs either via the imd/Relish or in some cases Toll/Dif pathway. My RNAi experiments in mbn-2 cells revealed that the peptidoglycan recognition protein PGRP-LC is a major activator of the imd/Relish pathway. In PGRP-LC deficient mbn-2 cells, Relish signalling is almost entirely blocked. However, the complex PGRP-LC gene generates three alternative splice forms, each of them carrying one of three possible PGRP domains, LCx, LCy, and LCa. I found that in the tissue culture system PGRP-LCa plays a specific role in the recognition of Gram-negative bacteria, while PGRP-LCx is crucial for the recognition of Gram-positive and Gram-negative bacteria, and peptidoglycan. Targeted mutagenesis of the PGRP-LCa isoform in vivo shows that the situation is more complicated than in the cell culture experiments. In conclusion, PGRPs constitute a highly diversified family of proteins, including key players of the innate immune response.