Sensory-motile cells fulfill various biological functions ranging from immune activity or wound healing to the formation of the highly complex nervous systems of vertebrates. In the case of neurons, a dynamic structure at the tip of outgrowing processes navigates towards target cells or areas during the generation of neural networks. These fan shaped growth cones are equipped with a highly complex molecular machinery able to detect various external stimuli and to translate them into directed motion. Receptor and adhesion molecules trigger signaling cascades that regulate the dynamics of an internal polymeric scaffold, the cytoskeleton. It plays a crucial role in morphology maintenance as well as in the generation and distribution of growth cone forces. The two major components, actin and microtubules (MTs) connect on multiple levels through interwoven biochemical and mechanical interactions. Actin monomers assemble into semiflexible filaments (F-actin) which in turn are either arranged in entangled networks in the flat outer region of the growth cone (lamellipodium) or in radial bundles termed filopodia. The dynamic network of actin filaments extends through polymerization at the front edge of the lamellipodium and is simultaneously moving towards the center (C-domain) of the growth cone. This retrograde flow (RF) of the actin network is driven by the polymerizing filaments themselves pushing against the cell membrane and the contractile activity of motor proteins (myosins), mainly in the more central transition zone (T-zone). Through transmembrane adhesion molecules, a fraction of the retrograde flow forces is mechanically transmitted to the cellular substrate in a clutch-like mechanism generating traction and moving the GC forward. MTs are tubular polymeric structures assembled from two types of tubulin protein subunits. They are densely bundled in the neurite and at the growth cone “neck” (where the neurite opens out into the growth cone) they splay apart entering the C-domain and more peripheral regions (P-domain). Their advancement is driven by polymerization and dynein motor protein activity. The two subsystems, an extending array of MTs and the centripetal moving actin network are antagonistic players regulating GC morphology and motility. Numerous experimental findings suggest that MTs pushing from the rear interact with actin structures and contribute to GC advancement. Nevertheless, the amount of force generated or transmitted through these rigid structures has not been investigated yet. In the present dissertation, the deformation of MTs under the influence of intracellular load is analyzed with fluorescence microscopy techniques to estimate these forces. RF mechanically couples to MTs in the GC periphery through friction and molecular cross-linkers. This leads to MT buckling which in turn allows the calculation of the underlying force. It turns out that forces of at least act on individual MT filaments in the GC periphery. Compared to the relatively low overall protrusion force of neuronal GCs, this is a substantial contribution. Interestingly, two populations of MTs buckle under different loads suggesting different buckling conditions. These could be ascribed to either the length-dependent flexural rigidity of MTs or local variations in the mechanical properties of the lamellipodial actin network. Furthermore, the relation between MT deformation levels and GC morphology and advancement was investigated. A clear trend evolves that links higher MT deformation in certain areas to their advancement. Interactions between RF and MTs also influence flow velocity and MT deformation. It is shown that transient RF bursts are related to higher MT deformation in the same region. An internal molecular clutch mechanism is proposed that links MT deformation to GC advancement. When focusing on GC dynamics it is often neglected that the retraction of neurites and the controlled collapse of GCs are as important for proper neural network formation as oriented outgrowth. Since erroneous connections can cause equally severe malfunctions as missing ones, the pruning of aberrant processes or the transient stalling of outgrowth at pivotal locations are common events in neuronal growth. To date, mainly short term pausing with minor cytoskeletal rearrangements or the full detachment and retraction of neurite segments were described. It is likely that these two variants do not cover the full range of possible events during neuronal pathfinding and that pausing on intermediate time scales is an appropriate means to avoid the misdetection of faint or ambiguous external signals. In the NG108-15 neuroblastoma cells investigated here, a novel type of collapse was observed. It is characterized by the degradation of actin network structures in the periphery while radial filopodia and the C-domain persist. Actin bundles in filopodia are segmented at one or multiple breaking points and subsequently fold onto the edge of the C-domain where they form an actin-rich barrier blocking MT extension. Due to this characteristic, this type of collapse was termed fold collapse. Possible molecular players responsible for this remarkable process are discussed. Throughout fold collapse, GC C-domain area and position remain stable and only the turnover of peripheral actin structures is abolished. At the same time, MT driven neurite elongation is hindered, causing the GC to stall on a time scale of several to tens of minutes. In many cases, new lamellipodial structures emerge after some time, indicating the transient nature of this collapse variant. From the detailed description of the cytoskeletal dynamics during collapse a working model including substrate contacts and contractile actin-myosin activity is derived. Within this model, the known and newly found types of GC collapse and retraction can be reduced to variations in local adhesion and motor protein activity. Altogether the results of this work indicate a more prominent role of forward directed MT-based forces in neuronal growth than previously assumed. Their regulation and distribution during outgrowth has significant impact on neurite orientation and advancement. The deformation of MT filaments is closely related to retrograde actin flow which in turn is a regulator of edge protrusion. For the stalling of GCs it is not only required that actin dynamics are decoupled from the environment but also that MT pushing is suppressed. In the case of fold collapse, this is achieved through a robust barrier assembled from filopodial actin bundles.

Developmental biology adresses how cells are organised into functional structures and eventually into a whole organism. It is crucial to understand the molecular basis for processes in development, by studying the expression and function of relevant genes and their relationship to each other. A gene function can be studied be creating loss-of-function situations, in which the change in developmental processes is examined in the absense of a functional gene product, or in gain-of-function studies, where a gene product is either intrinsically overproduced or ectopically upregulated. One approach for a loss-of-function situation is the creation of specific mutants in single genes, and the zebrafish (Danio rerio) has proven to be an excellent model organism for this purpose. In this thesis, I report on two forward genetic screens performed to find new mutants affecting brain development, in particular mutants defective in development and function of the midbrain-hindbrain boundary (MHB), an organiser region that patterns the adjacent brain regions of the midbrain and the hindbrain. In the first screen, I could identify 10 specific mutants based on morphology and the analysis of the expression patterns of lim1 and fgf8, genes functioning as early neuronal markers and as a patterning gene, respectively. Three of these mutants lacked an MHB, and by complementation studies, I identified these mutants as being defective in the spg locus. The second screen produced 35 new mutants by screening morphologically and with antibodies against acetylated Tubulin, which marks all axonal scaffolds, and anti-Opsin, which is a marker for photoreceptors in the pineal gland. According to their phenotype, I distributed the mutant lines into 4 phenotypic subgroups, of which the brain morphology group with 18 mutant lines was studied most intensively. In the last part of my thesis, I characterise one of these brain morphology mutants, broken heart. This mutant is defective in axonal outgrowth and locomotion, and shows a striking reduction of serotonergic neurons in the epiphysis and in the raphe nuclei in the hindbrain, structures involved in serotonin and melatonin production. Studies in other model organisms suggested a role of factors from the floor plate and the MHB in induction of the serotonergic neurons in the hindbrain, and using broken heart, I show that Fgf molecules such as Fgf4 and Fgf8 can restore partially the loss of serotonergic neurons in the mutant. I conclude that forward genetic screens are an invaluable tool to generate a pool of mutations in specific genes, which can be used to dissect complex processes in development such as brain development.

Developmental biology adresses how cells are organised into functional structures and eventually into a whole organism. It is crucial to understand the molecular basis for processes in development, by studying the expression and function of relevant genes and their relationship to each other. A gene function can be studied be creating loss-of-function situations, in which the change in developmental processes is examined in the absense of a functional gene product, or in gain-of-function studies, where a gene product is either intrinsically overproduced or ectopically upregulated. One approach for a loss-of-function situation is the creation of specific mutants in single genes, and the zebrafish (Danio rerio) has proven to be an excellent model organism for this purpose. In this thesis, I report on two forward genetic screens performed to find new mutants affecting brain development, in particular mutants defective in development and function of the midbrain-hindbrain boundary (MHB), an organiser region that patterns the adjacent brain regions of the midbrain and the hindbrain. In the first screen, I could identify 10 specific mutants based on morphology and the analysis of the expression patterns of lim1 and fgf8, genes functioning as early neuronal markers and as a patterning gene, respectively. Three of these mutants lacked an MHB, and by complementation studies, I identified these mutants as being defective in the spg locus. The second screen produced 35 new mutants by screening morphologically and with antibodies against acetylated Tubulin, which marks all axonal scaffolds, and anti-Opsin, which is a marker for photoreceptors in the pineal gland. According to their phenotype, I distributed the mutant lines into 4 phenotypic subgroups, of which the brain morphology group with 18 mutant lines was studied most intensively. In the last part of my thesis, I characterise one of these brain morphology mutants, broken heart. This mutant is defective in axonal outgrowth and locomotion, and shows a striking reduction of serotonergic neurons in the epiphysis and in the raphe nuclei in the hindbrain, structures involved in serotonin and melatonin production. Studies in other model organisms suggested a role of factors from the floor plate and the MHB in induction of the serotonergic neurons in the hindbrain, and using broken heart, I show that Fgf molecules such as Fgf4 and Fgf8 can restore partially the loss of serotonergic neurons in the mutant. I conclude that forward genetic screens are an invaluable tool to generate a pool of mutations in specific genes, which can be used to dissect complex processes in development such as brain development.