Microtubule nucleation by γ-tubulin complexes and beyond

Abstract

In this short review, we give an overview of microtubule nucleation within cells. It is nearly 30 years since the discovery of γ-tubulin, a member of the tubulin superfamily essential for proper microtubule nucleation in all eukaryotes. γ-tubulin associates with other proteins to form multiprotein γ-tubulin ring complexes (γ-TuRCs) that template and catalyse the otherwise kinetically unfavourable assembly of microtubule filaments. These filaments can be dynamic or stable and they perform diverse functions, such as chromosome separation during mitosis and intracellular transport in neurons. The field has come a long way in understanding γ-TuRC biology but several important and unanswered questions remain, and we are still far from understanding the regulation of microtubule nucleation in a multicellular context. Here, we review the current literature on γ-TuRC assembly, recruitment, and activation and discuss the potential importance of γ-TuRC heterogeneity, the role of non-γ-TuRC proteins in microtubule nucleation, and whether γ-TuRCs could serve as good drug targets for cancer therapy.

Keywords: MTOC; centrosome; g-TuRC; gamma-tubulin ring complex; microtubule.

Figure 1. Templated microtubule nucleation

(A) γ-tubulin molecules (yellow) within the γ-TuRC are positioned in a single-turn helix via their binding to GCP proteins (blue). (B) γ-tubulin molecules bind to incoming α/β-tubulin dimers from the cytosol and this is thought to promote the lateral interaction between the α/β-tubulin dimers as they grow into protofilaments (a protofilament is a single end-to-end chain of tubulin dimers). (C) Microtubule assembly progresses slowly through an unstable stage where disassembly is more likely than continued assembly (as indicated by the thickness of the two-way arrows). (D) Assembly is thought to reach a stable stage, where a microtubule seed containing sufficient tubulin dimers has formed (although the size of this stable seed remains unclear). (E) Once the stable seed has formed, microtubule polymerisation is favoured and can progress rapidly. Abbreviation: GCP, γ-tubulin complex protein.

Figure 2. Non-core γ-TuRC components

(A) The canonical γ-TuSC comprises two molecules of γ-tubulin and one each of GCP2 and GCP3, but alternative γ-TuSCs may exist in which GCP2, GCP3, or both are replaced by either GCP4, 5, or 6. (B) GCP2–6 all contain a Grip1 and a Grip2 domain. The Grip1 domain mediates interactions between GCP proteins, while the Grip2 domain mediates interactions with γ-tubulin. In addition, GCP6 contains an expanded central region that includes nine 27-aa repeats of unknown function. (C) GCP4, 5 and 6 (depicted here in purple) are predicted to replace some of the GCP2 and GCP3 molecules within the ring, but their exact positions remain unknown. MZT1 binds to the N-terminal regions of GCP proteins and acts as an adapter protein for the binding of the tethering protein NEDD1 (left) and tethering proteins that contain an N-terminal CM1-domain, such as CDK5RAP2 (right). NEDD1 contains putative WD40 repeats that form a β-propeller structure known to mediate protein–protein interactions (presumably with proteins at MTOCs). The structure of the C-terminus of NEDD1 is currently unknown but is required for binding to the γ-TuRC. The positions of NME7 kinase, MZT2 and LGALS3BP remain unknown. (D) The function of the extra sequence in GCP6 is unknown, but may provide a binding site for a GCP6-specific tethering protein (top) or may function in ring assembly by forming interactions with other γ-TuRC components (bottom). Abbreviations: CM1, centrosomin motif 1; MZT1, MOZART1; MZT2, MOZART 2.

Figure 3. Known and potential forms of γ-TuRC heterogeneity

(A) In mouse keratinocytes, γ-TuRCs are bound mutually exclusively by either NEDD1 or CDK5RAP2, suggesting that both tethering proteins bind to a similar region of the γ-TuRC. NEDD1-bound γ-TuRCs serve to anchor microtubules while CDK5RAP2-bound γ-TuRCs nucleate microtubules. The position of GCP4, 5 and 6 within γ-TuRCs in these cells remains unknown. (B) In Drosophila, most γ-TuRCs do not contain MZT1 (red), which is expressed only in the testes. Within the testes, MZT1 is predominantly expressed in sperm cells but in early elongating sperm is only present in γ-TuRCs that are recruited to the basal body, and is not present in γ-TuRCs recruited to mitochondria. The position of GCP4, 5 and 6 within Drosophila γ-TuRCs remains unknown. (C) While the position of GCP4, 5 and 6 within γ-TuRCs remains unknown, positive FRET data in HeLa cells suggest that GCP4 and GCP5 are adjacent to each other within the ring, while negative FRET data suggest GCP6 is not adjacent to either GCP4 or GCP5 (left) [29]. Stoichiometry measurements from HEK293T cells and immunoprecipitation experiments from HeLa cells suggest that some complexes do not contain GCP6 (middle left) [27] and that some complexes do not contain GCP4 (middle right) [58] respectively. GCP6 can still associate with γ-TuRCs in the absence of GCP4 or GCP5, suggesting that some complexes can form with only GCP6 (right) [28].

Figure 4. Different modes of microtubule nucleation

(A) The γ-TuRC templates the addition of tubulin dimers to form a microtubule, but is predicted to promote only lateral and not longitudinal interactions between tubulin dimers. (B) Under certain conditions, the combination of tubulin dimers and microtubule-associated proteins (MAPs) is sufficient to promote microtubule nucleation. Tog domain proteins help polymerise the microtubule by promoting the longitudinal addition of tubulin dimers. TPX2 homologues bind across tubulin dimers within the lattice and help prevent catastrophe of the nascent microtubule seed. (C) In vivo, it is likely that a combination of a γ-TuRC and these MAPs drive highly efficient microtubule nucleation. This presumably occurs at centrosomes, where all of these proteins concentrate, but whether other MTOCs (that are less-efficient microtubule nucleators) use specific mechanisms remains unknown.