Evolution of condensin and cohesin complexes driven by replacement of Kite by Hawk proteins

Abstract

Mitotic chromosome condensation, sister chromatid cohesion, and higher order folding of interphase chromatin are mediated by condensin and cohesin, eukaryotic members of the SMC (structural maintenance of chromosomes)-kleisin protein family. Other members facilitate chromosome segregation in bacteria [1]. A hallmark of these complexes is the binding of the two ends of a kleisin subunit to the apices of V-shaped Smc dimers, creating a tripartite ring capable of entrapping DNA (Figure 1A). In addition to creating rings, kleisins recruit regulatory subunits. One family of regulators, namely Kite dimers (Kleisin interacting winged-helix tandem elements), interact with Smc-kleisin rings from bacteria, archaea and the eukaryotic Smc5-6 complex, but not with either condensin or cohesin [2]. These instead possess proteins containing HEAT (Huntingtin/EF3/PP2A/Tor1) repeat domains whose origin and distribution have not yet been characterized. Using a combination of profile Hidden Markov Model (HMM)-based homology searches, network analysis and structural alignments, we identify a common origin for these regulators, for which we propose the name Hawks, i.e. HEAT proteins associated with kleisins.

Figure 1

Hawk proteins form an evolutionarily related cluster and have displaced Kites in condensin and cohesin. (A) The amino- and carboxy-terminal domains of the Smc polypeptides together form the globular nucleotide-binding domain (NBD). Kleisin subunits (yellow) then close the ring, topologically entrapping DNA in the process. In Smc5-6, Kite proteins interact with kleisins. In cohesin, Scc2 competes with Pds5 for its binding site on the kleisin. (B) Cohesin (Scc3, Scc2, Pds5) and condensin (Ycs4, Ycg1) HEAT regulators form a compact Hawk cluster (circled). Each cluster represented by a single colour. For clarity, only edges with a mean probability ≥99.0% are shown. Disconnected sub-graphs are hidden — the exception to this is Scc3, which is the weakest member of the hawk cluster and has been manually added (dashed edges). Above this threshold, the Hawk cluster has strong links to TPD3 (Protein phosphatase PP2A regulatory subunit A) and APL2 (Clathrin assembly protein large beta-1 chain). Members of the latter family (white labels) retain some of the strongest links to both hawks and lokiarchaeal proteins. (C) Despite a pairwise sequence identity of ∼15%, SA2 and Pds5B (4PJU and 5HDT, respectively) are similar, with a TM-score of 0.44 (scores lower than 0.3 are spurious and alignment significance increases rapidly above 0.5). The top scoring local sequence alignment between the two HMM profiles was between Pds5B residues 311–418 and SA2 residues 285–400. Using TMalign to perform a pairwise structural alignment between these resulted in a fit with a TM-score of 0.73. The alignment is disrupted by a large indel in Pds5B — realigning SA2 to the region directly after this produces an improved TM-score of 0.64. The amino-terminal region of Pds5B has been truncated for clarity. (D) In the LECA Smc–kleisin ancestor, the Kite dimer was presumably flanked by the ancestral HEAT-protein/Hawk. Successive duplications of Hawks led to the Kites being displaced. The lack of Hawks in Smc5-6 suggests that it diverged earlier from the cohesins and condensins, whose specialised functions were facilitated by the recruitment of the Hawk family.