During the process of cell division, it is essential that daughter cells receive the correct complement of chromosomes. An incorrect number of chromosomes (aneuploidy) is associated with varying pathological outcomes. Tumour development has been strongly linked to aneuploidy, while many genetic diseases such as Down's syndrome are caused by loss or gain of single chromosomes. To prevent these conditions, cells have developed a sophisticated molecular machine which as well as segregating chromosomes during mitosis, has a surveillance role to ensure correct distribution.
Our laboratory is trying to understand the mechanism of this machine at a molecular level. In particular, we are interested in the kinetochore, a large multi-protein complex that attaches the condensed chromatids to the mitotic spindle, and provides the origin for the checkpoint signal. Many kinetochore proteins are subject to very rapid evolution, so it is often difficult to identify clear sequence motifs that may enable prediction of function, or to identify orthologs in different species. By using a combination of structural biology and in vitro biochemistry, we hope to clarify the functions of individual components of the kinetochore, and build up a model for the overall assembly.
Locating the centromere
In budding yeast, the chromosomal location of each centromere is determined by a tripartite DNA sequence motif, that contains an invariant binding site (CDEIII) for the Cep3 protein. This is a constituent of the CBF3 complex, a 480 kDa hetero-hexamer which is necessary for localisation and assembly of the rest of the kinetochore. Current evidence suggests that the purpose of CBF3 is to locate a centromeric nucleosome containing the histone H3 variant, CenH3 to the proximal CDEII sequence element. In addition to sequence-specific binding by Cep3, CBF3 contains the Ndc10 protein, which also binds the centromere in a CBF3-independent manner, and is required for completion of cytokinesis. Despite the importance of these proteins in multiple aspects of cell division, little is know about their structure or mechanisms. We have determined the structure of the Cep3 protein, and are now turning our attention to how the intact CBF3 complex is formed, and binds centromeric DNA. This has been complicated by the difficulty in producing sufficient quantities of recombinant protein. In collaboration with the Protein Production Facility at the LRI, we have overcome these problems, and currently re-constituting the intact CBF3 complex for further study.
The chromatin connection
It seems likely that the main mechanical connection to chromosomes occurs through sections of chromatin containing modified nucleosomes containing the histone H3 variant, CenH3. A large number of inner kinetochore proteins associated with centromeric chromatin have been identified, but it is not currently known which of these are responsible for directly binding the nucleosomes. In vertebrates, it is thought that the so-called CCAN proteins (constitutive centromere-associated network) contain multiple chromatin-binding activities that bind both CenH3 and conventional nucleosomes. Currently, about 17 distinct proteins have been implicated in the CCAN. The roles of the individual proteins, assembly hierarchy and relationships between them are unclear and somewhat controversial. Without such data, it has proved extremely difficult to carry out in vitro experiments on the proteins. However, several orthologs of the CCAN proteins have been identified in the simpler budding yeast inner kinetochore, and we are now attempting to map their interactions with the centromeric nucleosome.
The core of the kinetochore
Recent studies from various labs have identified a conserved super-complex of proteins, which appears to represent a central scaffold that directly links the inner layers of the kinetochore to the microtubule plus-tips. This is the so-called KMN complex, after the constituent sub-complexes, KNL1 (Spc105 in budding yeast), Mtw1 and Ndc80. Our understanding of the function of this complex has substantially advanced with the demonstration that Ndc80 can directly bind microtubules and forms a stable complex with Mtw1 and KNL1, an interaction that enhances the microtubule affinity (Figure 1a). Exactly what the functions of Mtw1 and Knl1 are is unclear. KNL1 alone has a weak affinity for microtubules but there are no obvious sequence motifs that can be directly related to this activity. The Mtw1 complex is equally mysterious. It has no intrinsic MT affinity, but has been suggested that it acts as a scaffold, to anchor the KMN complex and correctly orient the Ndc80 and KNL1 proteins for MT attachment. Importantly, all three sub-complexes are targets of the IPL1/Aurora B kinase that is essential for the recognition and removal of incorrect MT-chromosome attachments. It is probable that phosphorylation of the Ndc80 MT-binding domains directly reduces affinity for tubulin subunits, but the function of the phosphorylation in the other proteins is unknown. We have successfully expressed intact, full-length versions of both the Ndc80 and Mtw1 complexes. The recombinant complexes can interact, and may be isolated as a single stable species (Figure 1b). We are now attempting to reconstitute the entire complex including KNL1 for structural and biochemical analysis.
Figure 1. a) Schematic representation of the KMN complex. The coiled-coil containing Ndc80 complex can directly bind to microtubules via EB1-like domains (light and dark blue). KNL1 can also bind to MTs, and the Mtw1 complex appears to act as a scaffold, and contact the inner kinetochore. b) The Mtw1 and Ndc80 tetramers associate to form a 310 kDa octameric complex that may be purified as a discrete species in large quantities. The Coomassie-stained gel shows all eight individual proteins.
For a list of refereed research papers, see Publications (in navigation on left).
