Our group studies the molecular mechanisms of eukaryotic chromosome segregation, with a particular interest in the macromolecular complexes that attach replicated sister chromatids to each other, and to spindle microtubules.
We are interested in determining the atomic structures of these complexes, and understanding how their overall architecture and function aids the cell in bringing about the rapid and accurate dissemination of the genome.
We employ a combination of biophysical and biochemical techniques to address these questions, primarily those of x-ray crystallography for high-resolution studies and electron microscopy to analyse larger complexes.
After DNA replication, sister chromatids are held together by the ring-shaped cohesin complex, which opposes spindle forces until proteolytic cleavage of the Scc1 subunit at anaphase onset. Cohesin is initially loaded onto chromosomes in the late G1 phase of the cell cycle. However, at this point the association between cohesin and chromatin remains reversible and is not capable of generating cohesion.
The switch to a persistent cohesive state requires the acetylation of the ATPase head of Smc3 by a replication-fork associated acetyltransferase, Eco1. In some unknown manner, this modification substantially alters the dynamics of the cohesin-chromatin interaction, irreversibly locking the chromatids together until anaphase.
Studies by several laboratories have implicated two 'cohesin accessory' proteins, Pds5 and Wapl (Wpl1 in budding yeast) as being mediators of the acetylation-induced switch. According to one model, these proteins form a dimeric complex that directly binds cohesin and inhibits formation of the cohesion-competent state. Acetylation of the Smc3 ATPase head is proposed to counteract the activity of the Pds5-Wpl1 complex, allowing the relevant switch to occur in the cohesin complex. In addition to its role in mitosis, Wapl has been implicated in the global regulation of chromatin structure during interphase and in transcriptional regulation of certain genes.
It has been known for some time that cohesin can directly influence gene expression, possibly by looping enhancer elements to promoters. It is not unreasonable, therefore to imagine that Wapl-mediated loading or unloading of cohesin might be used to direct transcription from associated loci. Given the multiple roles of Wapl, and its conservation throughout evolution, we thought it represented an interesting structural and biochemical target.
At the level of primary structure, Wapl orthologs are characterised by an extended N-terminal domain that is highly divergent in both size and amino acid composition, and a highly conserved C-terminal domain, known as the Wapl domain (Figure 1A). As the N-terminal is predicted to be unstructured, we decided to crystallise the Wapl domain alone. This was successful and we determined the structure of the domain at 2 .1 Å resolution (Figure 1B).
The crescent-shaped structure is comprised of a core of HEAT repeats, flanked by small helical bundles. A long arginine-lysine rich loop and helix are inserted into one of the canonical HEAT repeats and help form a highly positively charged patch on the N-terminal face of the protein. This appears to be a highly conserved feature of the Wapl domain and lies close to two absolutely conserved acidic residues that form a smaller negatively charged patch (Figure 1C).
Interestingly, a mutation in one of these acidic residues had previously been identified as a suppressor of Wpl1 activity, and we subsequently demonstrated that mutations that eliminated the charge of either the acidic or basic patch would similarly inactivate the protein.
Figure 1. A. Primary structure of Wapl orthologs. The large divergence in the size of the N-terminal is apparent. The conserved C-terminal Wapl domain is depicted as a purple bar. B. Ribbon diagram of the Wapl domain from A. gossypii, colour-coded from the amino (blue) to carboxy (red)terminal. C. Mutants in Wpl1 that supress lethality associated with Eco1 inactivation are located in the charged patches at the N-terminal of the protein. D. Diagram of the cohesin complex showing a speculative model for Wapl activity.
Exactly what the Wapl domain does has been somewhat harder to determine. It has been shown that the N-terminal tail is, at least in part, responsible for binding to Pds5, that can itself bind cohesin. We reasoned that this must place the Wapl domain in close proximity to the rest of the cohesin complex.
We carried out a candidate-based peptide array screen, which strongly suggested the domain directly interacted with the Smc3 ATPase head. Using purified recombinant proteins, we were able to quantify the interaction and, importantly, show that the in vivo suppressor mutations in the acidic and basic patches substantially reduced the affinity of the Wapl-Smc3 binding.
If binding of the Pds5-Wapl complex to the cohesin heads is indeed responsible for preventing stable cohesion establishment, an attractive idea is that acetylation of Smc3 may disrupt the interaction leading to release or re-organisation of the complex (Figure 1D).
Experiments with the Wapl domain alone showed a slight decrease in binding when acetyl-mimicking mutations were introduced into Smc3, but it is not clear if these are representative of those that would occur in the context of the intact Pds5-Wapl heterodimer.
Further studies, both structural and biochemical, will be required to fully dissect what is undoubtedly a highly complex mechanism.