The pseudo DUBs KIAA and Abraxas are
The pseudo DUBs KIAA0157 and Abraxas are required for the catalytic function of BRCC36. Comparison of the BRCC36–KIAA0157 heterodimer structure with an inactive BRCC36 homodimer structure provides a model for understanding how this functional interplay is achieved. As shown in schematic form in Figure 5B, the binding of KIAA0157 to BRCC36 is required for stabilization of (1) the CCHB, (2) the E-loop of BRCC36 and optimal positioning of the catalytic glutamate side chain for catalysis, and (3) the Ins-1 loop of BRCC36 and its proper positioning for substrate binding. Stabilization of all three elements is supported through direct interactions between KIAA0157 and BRCC36 within a single heterodimer as well as through interactions across BRCC36–KIAA0157 heterodimers within a super dimer. Despite the modeled vicinity of the proximal ubiquitin moiety to the opposing BRCC36 protomer within the super dimer (Figure 7A), our data demonstrate that super-dimerization does not contribute positively to substrate recognition (Figure 7C). In addition to supporting the catalytic function of BRCC36, KIAA0157 and Abraxas also serve scaffolding functions to recruit the BRCC45–MERIT40 heterodimer and the targeting subunits RAP80 or SHMT2. Surprisingly, super dimerization is not only required for the catalytic function of BRCC36 but it Thienoguanosine mg is also required for the binding of SHMT2 and RAP80/BRCA1 to BRISC and ARISC. It is also evident from modeling and mutational studies that super dimerization of BRISC/ARISC may impinge on enzyme function by restricting substrate binding to only one BRCC36 protomer in the super dimer at one time (Figure 7E). These inter-dependencies of structure and function provide intriguing potential for enzyme regulation and we speculate they may be responsible, in part, for the pronounced sensitivity to substrate inhibition displayed by the Ala205Asp and Ile212Asp single site dimer breaker mutants (Figure 6B). The apparent obligate folding dependency of KIAA0157 and Abraxas on BRCC36 that we observed in insect cells hints that the heterodimer interaction is constitutive and may not require post-translational modifications. However, the super dimerization of BRISC (and by extension ARISC) may be regulated in principle, since mutations that drive a monomer state are well tolerated in our insect cell expression studies with no evidence of protein aggregation. If BRISC transitions between an inactive monomer and active dimer states, and only the dimer state can recognize the targeting elements SHMT2 and RAP80, this would provide a tight mechanism to couple enzyme activation with the recruitment to biologically relevant substrates. How might super dimerization be regulated? Others have shown that residues in the BRCC36 E-loop and in the KIAA0157 E-loop supporting element are phosphorylated in cells (Phosida database [Gnad et al., 2011]). Phosphorylation at these sites is well placed to affect catalytic function directly or indirectly by influencing super dimerization. Since super dimerization involves contributions of both KIAA0157/Abraxas and BRCC36 subunits, it may be possible to differentially regulate ARISC and BRISC enzyme activity with small molecules by specifically targeting the MPN– subunit. This would permit selective intervention on ARISC biological function in the DNA damage response or BRISC involvement in inflammatory cytokine signaling, even though the two holoenzymes share the same catalytic subunits. Additionally, although both complexes share a common architecture, the relative contribution and strength of the component interactions within ARISC and BRISC differ to some degree, as evidenced by our mutational studies in cells. Moreover, KIAA0157 is sufficient to impart DUB activity to BRCC36, while Abraxas and BRCC45 are both necessary for minimal DUB activity in ARISC (Cooper et al., 2010, Feng et al., 2010, Patterson-Fortin et al., 2010). These differences in function might also be exploited for selective intervention by small molecules.