Search

Microtubules are one of the major types of cytoskeletal filaments in cells. They are very dynamic polymers composed of αβ-tubulin dimers arranged longitudinally in head-to-tail fashion as well as laterally to assemble 13-protofilament hollow cylindrical tubes. The incorporation of GTP-bound αβ-tubulin dimers generates a fast growing plus end exposing β-tubulin and a slow growing minus end exposing α-tubulin. In cells, microtubules are assembled de novo from a template, called γ-TuRC, which interacts with α-tubulin. Microtubules can either remain capped by γ-TuRC and anchored to the microtubule-organizing centers (MTOCs) or be released if they are cut by microtubule severing enzymes like katanin. The release of microtubules from MTOC generates free minus ends, which are then stabilized by minus-end binding proteins called CAMSAPs. However, the plus ends remain very dynamic and undergo transitions from growth to shrinkage, termed “catastrophes”, and the opposite transitions termed “rescues”. Numerous microtubule regulatory proteins act at the plus ends, minus ends and the microtubule shafts connecting the two ends to control the organization and density of cellular microtubule networks. In this thesis, we focused on each of these aspects and explored the dynamic life of microtubules by reconstituting these processes in vitro using purified proteins. We first focused on the birth and growth of microtubules. We reconstituted microtubule nucleation using purified γ-TuRC and microtubule regulatory proteins and showed that CDK5RAP2, CLASP2 and chTOG promoted microtubule nucleation from γ-TuRC. We discovered that CAMSAPs can bind to γ-TuRC-capped microtubule minus ends and displace γ-TuRC from these ends, generating free and stable microtubule minus ends. Furthermore, we found out that CDK5RAP2, but not CLASP2 or chTOG, can inhibit CAMSAP binding and microtubule release. We propose that the destiny of a microtubule depends on the type of protein complex that activates its nucleation. We then described a mechanism for stabilization of microtubule lattice by TRIM46, a neuronal protein, which can bundle parallel microtubules and promote microtubule rescues within these bundles. We also revealed that Ankyrin-G, a scaffold protein, can recruit TRIM46-stabilized microtubule bundles to the axonal membrane to drive the assembly of the axon initial segment in neurons. We also uncovered a new role of CLASP2 as a microtubule repair factor participating in microtubule maintenance. We demonstrated that CLASP2, an anti-catastrophe factor, can promote complete repair of damaged microtubule lattices by inhibiting microtubule depolymerization and promoting tube closure at the damage sites, causing lattice renewal. Finally, we described a three-protein module involving katanin, CAMSAPs, and WDR47 that can regulate microtubule polymer mass and minus-end stability. We showed that katanin can cut and amplify CAMSAP2/3-stabilized microtubule minus ends. WDR47 can inhibit the binding of katanin to CAMSAP2/3-stabilized minus ends and protect them from severing. The presence of WDR47 shifts the balance from microtubule amplification to minus-end growth regulation. To conclude, we obtained mechanistic insights into the regulation of microtubule nucleation, minus-end dynamics, lattice stabilization and maintenance, microtubule number and the interplay between microtubule regulatory proteins. These insights will help to understand how microtubule arrays are organized in cells.

Microtubules are one of the major types of cytoskeletal filaments in cells. They are very dynamic polymers composed of αβ-tubulin dimers arranged longitudinally in head-to-tail fashion as well as laterally to assemble 13-protofilament hollow cylindrical tubes. The incorporation of GTP-bound αβ-tubulin dimers generates a fast growing plus end exposing β-tubulin and a slow growing minus end exposing α-tubulin. In cells, microtubules are assembled de novo from a template, called γ-TuRC, which interacts with α-tubulin. Microtubules can either remain capped by γ-TuRC and anchored to the microtubule-organizing centers (MTOCs) or be released if they are cut by microtubule severing enzymes like katanin. The release of microtubules from MTOC generates free minus ends, which are then stabilized by minus-end binding proteins called CAMSAPs. However, the plus ends remain very dynamic and undergo transitions from growth to shrinkage, termed “catastrophes”, and the opposite transitions termed “rescues”. Numerous microtubule regulatory proteins act at the plus ends, minus ends and the microtubule shafts connecting the two ends to control the organization and density of cellular microtubule networks. In this thesis, we focused on each of these aspects and explored the dynamic life of microtubules by reconstituting these processes in vitro using purified proteins. We first focused on the birth and growth of microtubules. We reconstituted microtubule nucleation using purified γ-TuRC and microtubule regulatory proteins and showed that CDK5RAP2, CLASP2 and chTOG promoted microtubule nucleation from γ-TuRC. We discovered that CAMSAPs can bind to γ-TuRC-capped microtubule minus ends and displace γ-TuRC from these ends, generating free and stable microtubule minus ends. Furthermore, we found out that CDK5RAP2, but not CLASP2 or chTOG, can inhibit CAMSAP binding and microtubule release. We propose that the destiny of a microtubule depends on the type of protein complex that activates its nucleation. We then described a mechanism for stabilization of microtubule lattice by TRIM46, a neuronal protein, which can bundle parallel microtubules and promote microtubule rescues within these bundles. We also revealed that Ankyrin-G, a scaffold protein, can recruit TRIM46-stabilized microtubule bundles to the axonal membrane to drive the assembly of the axon initial segment in neurons. We also uncovered a new role of CLASP2 as a microtubule repair factor participating in microtubule maintenance. We demonstrated that CLASP2, an anti-catastrophe factor, can promote complete repair of damaged microtubule lattices by inhibiting microtubule depolymerization and promoting tube closure at the damage sites, causing lattice renewal. Finally, we described a three-protein module involving katanin, CAMSAPs, and WDR47 that can regulate microtubule polymer mass and minus-end stability. We showed that katanin can cut and amplify CAMSAP2/3-stabilized microtubule minus ends. WDR47 can inhibit the binding of katanin to CAMSAP2/3-stabilized minus ends and protect them from severing. The presence of WDR47 shifts the balance from microtubule amplification to minus-end growth regulation. To conclude, we obtained mechanistic insights into the regulation of microtubule nucleation, minus-end dynamics, lattice stabilization and maintenance, microtubule number and the interplay between microtubule regulatory proteins. These insights will help to understand how microtubule arrays are organized in cells.

The centrosome is the main microtubule organizing center (MTOC) in animal cells. It consists of a pair of centrioles surrounded by pericentriolar material (PCM) that nucleates and anchors microtubules. The binding of PCM to centrioles, PCM self-assembly and PCM transport by the microtubule-based motor dynein are the prerequisites for centrosome assembly, but the mechanisms of PCM self-assembly in interphase cells are poorly understood. Moreover, the relative importance of different molecular pathways of PCM assembly varies between different cell systems and different phases of the cell cycle. Previous studies have mostly focused on mitotic cells, but systematic investigations have not been performed in interphase cells. In addition, γ-tubulin ring complex (γ-TuRC) as template for microtubule assembly largely determines the function of MTOC, but the relative importance of different centrosomal components in microtubule nucleation and maintenance of microtubule networks has not been systematically investigated. Moreover, it is unclear whether γ-TuRC can nucleate microtubules and maintain microtubule arrays in mammalian cells in the absence of γ-TuRC adaptors. In this thesis, we address these questions. In addition to the PCM, the other major component of the centrosome are centrioles, microtubule-based organelles whose formation and elongation are regulated by a large number of specific proteins. In vitro reconstitution experiments showed that two centriole biogenesis factors, CPAP and CP110, cooperatively regulate microtubule plus-end growth in vitro, but the importance of their interaction for centriole formation has not been examined on cells. In this thesis, we investigated the role of the CPAP-CP110 interaction in regulating centriole elongation. In Chapter 2, we reported that when centrioles are lost due to either depletion or inhibition of PLK4, self-clustering of PCM proteins is sufficient to form a compact acentriolar MTOC (caMTOC) and thus organize a dense radial microtubule array in interphase cells. We showed that such self-clustering of PCM is dynein-dependent and requires pericentrin, CDK5RAP2, ninein and γ-tubulin, but not CEP192, NEDD1 and CEP152. We found that the formation of caMTOC is sensitive to the non-centrosomal MTOC pathways such as AKAP450-dependent PCM recruitment to the Golgi apparatus and CAMSAP2-mediated stabilization of free microtubule minus ends. We also reproduced the findings of our cellular experiments using computer simulations. In Chapter 3, we describe tools for comprehensive and systematic functional analysis of the complex regulatory pathways of microtubule nucleation and anchoring at mammalian centrosomes using a cellular model that simultaneously lacks multiple PCM components. We found that in cells lacking the non-centrosomal microtubule minus-end stabilization pathways, different γ-TuRC adaptors exert additive effects on microtubule nucleation and attachment at the centrosome, and knocking them out in interphase cells prevents γ-TuRC from generating a dense microtubule array, even though all positive regulators of microtubule dynamics, are still present. In Chapter 4, we examined the importance of CPAP-CP110 interaction in cells by replacing wild-type CPAP with CPAP mutants that cannot bind CP110. Our data indicate that centrioles, albeit shorter ones, could still form in the presence of the CPAP mutant, demonstrating that the CPAP-CP110 interaction is not essential for centriole formation, but promotes centriole elongation.

The centrosome is the main microtubule organizing center (MTOC) in animal cells. It consists of a pair of centrioles surrounded by pericentriolar material (PCM) that nucleates and anchors microtubules. The binding of PCM to centrioles, PCM self-assembly and PCM transport by the microtubule-based motor dynein are the prerequisites for centrosome assembly, but the mechanisms of PCM self-assembly in interphase cells are poorly understood. Moreover, the relative importance of different molecular pathways of PCM assembly varies between different cell systems and different phases of the cell cycle. Previous studies have mostly focused on mitotic cells, but systematic investigations have not been performed in interphase cells. In addition, γ-tubulin ring complex (γ-TuRC) as template for microtubule assembly largely determines the function of MTOC, but the relative importance of different centrosomal components in microtubule nucleation and maintenance of microtubule networks has not been systematically investigated. Moreover, it is unclear whether γ-TuRC can nucleate microtubules and maintain microtubule arrays in mammalian cells in the absence of γ-TuRC adaptors. In this thesis, we address these questions. In addition to the PCM, the other major component of the centrosome are centrioles, microtubule-based organelles whose formation and elongation are regulated by a large number of specific proteins. In vitro reconstitution experiments showed that two centriole biogenesis factors, CPAP and CP110, cooperatively regulate microtubule plus-end growth in vitro, but the importance of their interaction for centriole formation has not been examined on cells. In this thesis, we investigated the role of the CPAP-CP110 interaction in regulating centriole elongation. In Chapter 2, we reported that when centrioles are lost due to either depletion or inhibition of PLK4, self-clustering of PCM proteins is sufficient to form a compact acentriolar MTOC (caMTOC) and thus organize a dense radial microtubule array in interphase cells. We showed that such self-clustering of PCM is dynein-dependent and requires pericentrin, CDK5RAP2, ninein and γ-tubulin, but not CEP192, NEDD1 and CEP152. We found that the formation of caMTOC is sensitive to the non-centrosomal MTOC pathways such as AKAP450-dependent PCM recruitment to the Golgi apparatus and CAMSAP2-mediated stabilization of free microtubule minus ends. We also reproduced the findings of our cellular experiments using computer simulations. In Chapter 3, we describe tools for comprehensive and systematic functional analysis of the complex regulatory pathways of microtubule nucleation and anchoring at mammalian centrosomes using a cellular model that simultaneously lacks multiple PCM components. We found that in cells lacking the non-centrosomal microtubule minus-end stabilization pathways, different γ-TuRC adaptors exert additive effects on microtubule nucleation and attachment at the centrosome, and knocking them out in interphase cells prevents γ-TuRC from generating a dense microtubule array, even though all positive regulators of microtubule dynamics, are still present. In Chapter 4, we examined the importance of CPAP-CP110 interaction in cells by replacing wild-type CPAP with CPAP mutants that cannot bind CP110. Our data indicate that centrioles, albeit shorter ones, could still form in the presence of the CPAP mutant, demonstrating that the CPAP-CP110 interaction is not essential for centriole formation, but promotes centriole elongation.