Search

Axons are slender, cable-like projections of neurons which electrically wire our nervous system and bodies. They are the longest cellular processes and this unique architecture has to be maintained for an organism's lifetime. Notably, axons contain dense bundles of parallel microtubules (MTs) which form the structural backbones as well as the tracks for long-distance transport between cell bodies and synaptic endings. We lose about 50% of axons towards high age, and this is often accompanied by disorganisation of MT bundles in axon swellings. My host group proposes the model of 'local homeostasis' to explain how MT bundles are maintained. The model proposes that (a) MT bundles are highly dynamic with constant turnover to prevent senescence caused by wear and tear; (b) in the axonal environment the dynamic MTs tend to go off-track into disorganisation; (c) different mechanisms mediated by MT regulators 'tame' MTs into axonal bundles. The aim of my project was to identify further molecular mechanisms contributing to MT bundle formation and maintenance, focussing on candidate MT regulators which are all linked to neurodegeneration. These proteins were (1) Futsch/MAP1B (expected to act as a MT cross-linker, potentially in conjunction with the spectraplakin Short stop/Shot), (2) Spastin (expected to sustain MT turnover through its MT severing activity), and (3) the motor protein kinesin-1/Kinesin heavy chain (Khc) (expected to be involved in the even distribution of newly forming MTs through its MT sliding activity). For my studies, I capitalised on the versatile Drosophila primary neurons. I used combinatorial genetics and pharmacological approaches, combined with informative subcellular readouts including MT organisation, axon length, mitochondrial shape and distribution, presynaptic proteins distribution and endoplasmic reticulum. Detailed studies of all three regulators failed to confirm the above proposed working hypotheses, with loss of Futsch, of Spastin or of the Khc-driven MT sliding function failing to generate any MT disorganisation even in long-term studies or when challenged in different ways. Nevertheless, my work on all three candidates produced important new data and concepts. First, my results shifted the focus from Futsch to Shot, in particular its plakin-repeat region (PRR; specific to one isoform abundant in neurons). My work provided first analysis on a new CRISPR-mediated shotPRR allele lacking the PRR, which causes MT disorganisation, providing indication of the importance of PRR in MT bundling. Second, my results shifted the focus from Spastin to Katanin 60, another MT severer which I found to cause strong MT disorganisation when functionally depleted; future work will show whether it performs the roles I proposed for Spastin. Third, my main focus lay on Khc which causes strong MT disorganisation when depleted. However, this was not caused by absence of Khc-mediated MT sliding, but linked to its roles in cargo/vesicle transport (absence of the Kinesin light chain/Klc linker) and mitochondrial dynamics (absence of Miro or Milton linkers). My experiments aiming to explain how these sub-functions of Khc relate to MT bundle organisation led into new ways of thinking about kinesin-1-linked neurodegeneration. So far, the best supported model involves oxidative stress downstream of mitochondrial dysfunction. The second model involves force imbalance caused by loss of specific motor proteins from specific MT bundle subtracts; it is experimentally less well supported, but consistent with my experimental results so far. Although my experiments failed to support my original hypotheses for roles of Futsch, Spastin and Khc within the model of local axon homeostasis, the work I present opens up new concepts for thinking about axonal cell biology and provides new working hypotheses, candidate molecules and molecular tools readily available for future studies.

Cytoplasmic dynein 1 is a large multi-subunit complex that has a multitude of functions and therefore, many interactions. Dynein has two light intermediate chains (LICs) that bind in a mutually exclusive manner to the heavy chain and are proposed to form two distinct populations of dynein that could bind to different cargoes. Both unique and shared interactions have been discovered for the two LICs. Identifying the interactions of the LICs is hindered by the large size of the dynein complex, and if immunoprecipitation or pull-down approaches are used, by the need to retain the interactor-complex link through subsequent cell lysis and purification. We used the novel BioID method in an attempt to identify potential new LIC interactors, be they unique or shared. The BioID method utilises a mutated biotin ligase, BirA*, that generates activated biotin that then labels proteins in its vicinity. These biotinylated proteins can then be purified using streptavidin. The LICs were tagged with Myc BirA* on both termini and we confirmed that this tag had no dominant negative effect on dynein functions, such as positioning the Golgi apparatus. The tagged LICs were able to rescue the effects caused by a double endogenous LIC knockdown, therefore suggesting that they are functional and part of the dynein complex. Stable cell lines of Myc BirA* LIC1, Myc BirA* LIC2, LIC2 Myc BirA* and Myc BirA* GFP were generated. The tagged LICs in the stable cell lines could still rescue a double endogenous LIC knockdown whereas the Myc BirA* GFP did not. We tested the amount of incorporation of the tagged LICs into the dynein complex by sucrose density gradient centrifugation to try and limit unspecific biotin labelling. Our BioID mass spectrometry data for the tagged LICs revealed several dynein and dynactin subunits, demonstrating successful incorporation of the LICs into intact dynein complexes that could interact with dynactin. In addition, we identified the known dynein-dynactin adaptor protein, BicD2. Several other proteins of potential interest were identified, such as Importin 7 and Mre11. Further experiments are needed to confirm these as LIC near-neighbours and to then validate them as direct interactors.

Cytoplasmic dynein is a minus end directed microtubule motor protein. It is a multi-subunit complex that has many functions in eukaryotic cells, ranging from the transport of membrane cargoes during interphase to essential roles in spindle function during mitosis. Recently, dynein has been shown to participate in a critical process during meiosis, by driving telomere-led chromosome movements that are a prelude to chromosome pairing and recombination. Dynein interacts with the LINC complex component and outer nuclear membrane (ONM) protein, KASH5. The inner nuclear membrane (INM) partner of KASH5, SUN1, associates with the telomeres of meiotic chromosomes, providing a connection between the meiotic nucleus and cytoskeleton. By manipulating the subunit composition and activity of dynein, dynactin and associated proteins, the means by which KASH5 recruits dynein is becoming clearer. Dynein interacts directly with KASH5 via LICs 1 and 2, which function redundantly in recruitment of KASH5 to dynein. In contrast to the generally accepted dogma, dynactin, the major dynein accessory protein, is not required for recruitment of dynein to KASH5 and is instead recruited downstream of dynein. The accessory protein LIS1 may play a role in regulating the interaction between KASH5 and dynactin but is not required for dynein recruitment. The data presented in this thesis suggests that KASH5 could be a new member of the adaptor protein family that assembles with dynein and dynactin in tripartite complexes. Further work has addressed the functional significance of a KASH5 mutation that causes infertility in human patients. This mutation caused KASH5 to relocalise from the NE/ER membrane network to the mitochondria in cultured cells. Such mislocalisation in sperm would be predicted to prevent chromosome movement in prophase I, leading to meiotic failure and infertility.

The diffraction limit restricts conventional light microscopes to approximately 250 nm laterally and 500 nm axially, these limits being first proposed by Abbe in 1873. Despite this, optical microscopes have found many applications in biological research and single cells that are 10 - 100 um in size. Furthermore by coupling the non-invasive nature of a light microscope with highly sensitive fluorescent probes, fluorescence microscopy has also become a standard imaging technique. Recent advances in fluorescence microscopy now provide a number of methods to circumvent the Abbe diffraction limit, with many techniques becoming prevalent over the last 10 years including direct Stochastic Optical Reconstruction Microscopy (dSTORM). A dSTORM system has been constructed and calibrated using a commercially available inverted florescence microscope and total internal reflection florescence (TIRF) imaging. dSTORM relies on the ability to switch sparse subsets of fluorophores and temporally separate them. Provided the spatial separation is sufficient between any member of a subset, the average error with which the emission can be localized is much less than size of the emission profile itself. The underlying mechanism for this switching is detailed based on the principle of photoinduced electron transfer (PET). The switching characteristics of the common florescent dye Alexa Fluor 568 are investigated and shown to be controlled by a number of factors including the excitation intensity and concentration of the primary thiol cysteamine beta-MEA. A number of parameters are defined, including the dye switching rate, for a given set of physical parameters. U2OS cells are labelled for the microtubule protein Tubulin using immunofluorescent labelling strategies. A direct comparison is made between diffraction limited TIRF images and dSTORM reconstructed images, with an average width for microtubules determined to (58.2 ± 8.1) nm. Further measurements are made by labelling the Rab5 effector Early Endosome Antigen 1 (EEA1). From this the aspect ratio for early endosomes is determined to be 1.68 ± 0.7 with an average radius of (45.8 ± 18.8) nm. The point spatial distribution of EEA1 is investigated by using the linearised form of Ripley's K-function H(r) and the null hypothesis of complete spatial randomness tested. EEA1 is shown to cluster at radius of 58.7 nm on individual endosomes, thought to be due to the well defined binding domains present on early endosomes for EEA1. Further evidence suggests that clustering is also exhibited at another maximum of approximately 500 nm when looking at an ensemble of EEA1 and early endosomes.

Cytoplasmic dynein 1 (dynein) is a multi-subunit complex that transports cargo along microtubules towards their minus ends. These microtubule minus ends are normally located toward the centre of the cell. Dynein is involved in transport of endocytic and autophagic membranes and is tightly regulated by interactions between dynein subunits and by dynein-accessory proteins. Dynein accessory proteins that are involved in a wide range of dynein-driven transport events include dynactin, Lis1 and the paralogues Nde1 and Ndel1. Lis1 and Nde1/Ndel1 interact with each other and are involved in the recruitment of dynein to cargo and in regulating dynein activity. Although much is known about the specific interactions of dynein and accessory proteins, the interplay between dynein and its network of regulators in living cells is not well defined. This project used RNAi to investigate how the dynein subunits light intermediate chain (LIC) and intermediate chain (IC) as well as Lis1 and Nde1/Ndel1 influence the endocytic pathway, autophagy and cargo recruitment. Biochemical analysis of bulk membrane preparations showed that IC is important for dynein and dynactin association with intracellular membranes. In addition, dynein and dynactin recruitment to Rab interacting lysosomal protein (RILP)-positive membranes was shown to require LIC and there was redundancy between LIC1 and LIC2 in this role. Lis1 was also needed for dynactin-dynein recruitment to these membranes, in a context that was Nde1/Ndel1-independent. Loss of LIC, IC, Lis1 and Nde1 had differing effects on endocytic compartment size and distribution, but they all led to mislocalisation of early endosomes and lysosomes and caused lysosomes to become enlarged. Loss of LIC led to a specific phenotype whereby cells formed lamellipodia-like regions in which early endosomes and lysosomes accumulated. Loss of Lis1 prevented traffic from the early endosome to late endosomes and caused a striking enlargement of late endosomes and lysosomes. These enlarged lysosomes were LC3-positive, indicating that they were autophagic. In addition, loss of IC and LIC also led to an increase in LC3 puncta, but the LC3 did not colocalise specifically with lysosomes. In summary, the results from this project show that i) dynactin recruitment to intracellular membranes, including RILP-postivie membranes, requires dynein, ii) Lis1 and LIC1 or LIC2 are necessary but not sufficient, individually, to recruit dynein and dynactin to RILP-positive membranes iii) LIC, IC, Lis1 and Nde1/Ndel1 influence endocytic progression in specific ways, which may in turn affect autophagic flux.

The accurate regulation of cytoplasmic dynein-1 (dynein) is very important since dynein performs multiple functions in cells. In interphase, dynein is responsible for the correct positioning of membrane organelles, such as the Golgi complex and lysosomes. Previous work suggests that dynein's accessory proteins NudEL/Nde1/LIS1¬ may be involved in regulating dynein-dependent organelle movement. This study focuses on how NudEL regulates dynein-driven membrane movement. By using various NudEL fragments, this work presents the first evidence that NudEL is involved in the regulation of dynein-driven ER movement in vitro. Moreover, the in vivo organelle positioning assays also indicate additional regulatory function of NudEL.NudEL fragment (1-157 aa) which contains both the dynein and LIS1 binding domains is sufficient to activate dynein-driven membrane movement, since NudEL1-157 aa activates ER motility in vitro and enhances clustering of the Golgi complex and lysosomes in the peri-nuclear region in vivo. On the other hand, NudEL 96-206 aa containing the LIS1 binding domain alone inhibits ER motility in vitro and causes scattering of the Golgi complex and lysosomes in vivo, indicating an inhibition of dynein-dependent organelle movement. The activation of dynein activity requires the recruitment of LIS1 to the dynein complex by NudEL, since NudEL 1-157 aa has strong binding affinity to both LIS1 and dynein whereas NudEL 96-206 aa binds to LIS1 but not dynein which suggests the sequestering of LIS1 from the dynein complex. Interestingly, NudEL 1-206 aa, which also contains both the dynein and LIS1 binding domains, causes the dispersal of the Golgi complex and lysosomes in vivo, but to a lesser extent than NudEL 96-206 aa. The putative NudEL regulatory domain (157 -242 aa, which contains various phosphorylation sits and is less conserved between NudEL and Nde1) in NudEL 1-206 aa may regulate the interaction of LIS1 and the dynein complex, since NudEL 1-206 aa has strong binding affinity to LIS1 and weak binding affinity to dynein. However, further work is needed to understand the exact mechanism by which this putative NudEL domain regulates dynein activity.

In this work the functions of the microtubule motors kinesin-1, -2 and -3 have been analysed in various settings. The reconstitution of microtubule-dependent motor activity in vitro has been primarily used to dissect the contributions of individual motors to cargo motility in two specific scenarios. Initially the regulation of kinesin-1 in a cell cycle-dependent manner has been examined by studying the ability of rat liver endoplasmic reticulum (ER) tubules to move in cytosols prepared from Xenopus laevis egg extracts arrested in interphase, meiosis or mitosis. It was found that kinesin-1-driven ER motility is significantly disrupted during metaphase in vitro. This is likely due to the recruitment or loss of binding partners which has a concomitant influence upon kinesin-1 activity. This work presents the first evidence that kinesin-1-driven ER movement, and not simply network morphology, varies during cell division. Furthermore, it is postulated that the replication of such regulation of kinesin-1 activity in vivo may contribute to the well documented changes in organelle positioning and cargo transit through membrane trafficking pathways which occur during cell division.The fungal metabolite brefeldin A (BFA) induces tubulation of several compartments located within the secretory and endocytic pathways in a microtubule-dependent fashion. The identity of the motor(s) responsible for this motility remains unconfirmed and controversial since several reports with conflicting data have been published. The contributions of kinesin-1, -2 and -3 to these processes have been investigated using in vitro motility assays in which rat liver Golgi membranes were combined with Xenopus laevis egg extract cytosol in the presence of BFA. Function blocking antibodies and dominant negative proteins were used to perturb the activities of various kinesin motors. This data indicates a particular isoform of kinesin-3, KIF1C, is solely responsible for the movement of BFA-induced tubules in vitro. This work was complemented by in vivo immunofluorescence studies using the HeLaM cultured cell line. Transient transfections of dominant negative proteins, or siRNA-mediated depletion, were used to disrupt the activities of various kinesin motors, either in isolation or in combination with each other. This approach revealed a contribution of KIF1C and kinesin-1 to the movement of early endosomal BFA-induced tubules in vivo.