To predict levels of SCF assembly Liu et al first
To predict levels of SCF assembly, Liu et al. (2018) first quantified the remaining necessary kinetic parameters and derived a model in combination with known functions and concentrations of components. The model agrees astonishingly well with many experimental observables. In terms of SCF assembly, the model correlates with the ratios of SKP1-FBP-associated CUL1 upon inhibiting neddylation or doubly knocking out (DKO) CAND1 and CAND2. The model also predicted how the perturbed system in DKO Isochlorogenic acid C affects levels of the SCFβ-TrCP substrate phosphorylated IκBα (Liu et al., 2018). A perplexing result not explored is that the half-life of phosphorylated IκBα is 15 times shorter than the time it would take for half the cellular complexes between this substrate and SCFβ-TrCP to dissociate. It is possible that the dissociation rate for this substrate-FBP complex is accelerated by ubiquitylation, as this value has not yet been measured. Alternatively, this may imply that an unknown exchange factor actively removes substrate from FBPs or that the proteasome potentially degrades ubiquitylated substrates bound to a CUL1-SKP1-FBP complex. Unexpectedly, Liu et al. (2018) uncovered that the CAND1-bound population of CUL1 is dramatically biased for binding to DCN1, a neddylation co-E3 that helps recruit the NEDD8 conjugating enzyme UBE2M (Kurz et al., 2005, Scott et al., 2014). While the underlying mechanism is unclear, the authors speculate that this primes CUL1 to be neddylated concurrently with binding a SKP1-FBP module and displacement of CAND1. It will be interesting to see how inhibition of UBE2M binding to DCN1 could alter the landscape of SCF assembly. One extraordinary parameter that emerged from the modeling is that in the absence of bound substrate, CUL1 undergoes an entire exchange cycle in less than 1 min. Even more astonishingly, if all FPBs have equal access to CUL1, the entire pool of FPBs would sample CUL1 in less than 4 min in 293T cells (Liu et al., 2018). Such rapid and indiscriminate cycling could safeguard the SCF system from bias against FBPs that are expressed at low levels or that display weak affinity for CUL1. In a grander sense, the implications of these numbers are profound and suggest that CUL1-RBX1 and, by inference, other cullin-RING complexes are on an endless search-and-rescue mission continuously on the hunt for substrate-bound FBPs, ensuring active SCF assembly only upon increased substrate demand. In order to gauge the model’s predictive strength, Liu et al. (2018) simulated effects of varying the concentrations of SCF components and predicted that overexpression of CUL1, but not the FBP that targets phosphorylated IκBα, would rescue defects in the rate of its degradation in DKO cells. Experimental validation of these predictions presented a new paradox: if CUL1 upregulation can obviate the need for CAND exchange, why does such a complex system exist in the first place? A potential answer came from calculating a matrix of response coefficients, which suggested that increasing the total FBP concentration would delay substrate degradation specifically in DKO cells. Indeed, gross overexpression of an FBP in DKO cells that lack dynamic CAND-mediated exchange clogs the system: this restricts CUL1 from accessing other FBPs, thereby stabilizing their ubiquitylation substrates (Liu et al., 2018). The authors conclude that CAND-driven exchange permits the SCF system to tolerate changes in FBP expression associated with development, without requiring CUL1 levels to change in diverse regulatory settings. Nonetheless, some SCF substrates (p27, CyclinE) are stabilized in DKO cells only when total FBP levels are increased by overexpression, implying that these substrates can be efficiently degraded independently of CAND exchange. Why some SCF substrates do not require CAND-dependent FBP exchange is unclear, but could reflect variations in the levels of their cognate FBPs, or differences in dissociation rates of these substrates versus phosphorylated IκBα. Whatever the case, it will be interesting to see whether this model can predict threshold conditions for substrates that require CAND-dependent exchange.