Abstract

Of all ubiquitin-like small protein modifiers, Rub1/NEDD8 is the closest kin of ubiquitin in sequence and in structure. Despite their profound similarities, prevalence of ubiquitin and of Rub1 is starkly different: targets of ubiquitin modification reach into the thousands, whereas unique targets of Rub1/NEDD8 appear limited to one family of proteins, Cullins. This distinction is likely due to dedicated E1 activating enzymes that select either one or the other and relay the modifier until it is covalently attached to a target. To convert typical neddylation targets for modification by ubiquitin, and vice versa, we designed reciprocal substitutions at position 72 of Rub1 and of ubiquitin to render them substrates for activation by their non-cognate E1 activating enzymes. We found that this single amino acid is sufficient to distinguish between Ub and Rub1 in living cells, and determine their targets. Thus, modification of Cullins by UbR72T could compensate for loss of Rub1, even as it maintained its ability to polymerize and direct conjugates for degradation. Conversely, Rub1T72R activated by ubiquitin-activating enzyme entered into the ubiquitination cascade, however was not efficiently polymerized, essentially capping polyubiquitin chains. Upon shortage of free ubiquitin under stress, even native Rub1 spilled-over into the ubiquitinome suppressing polyubiquitination. By contrast, the need to maintain monomeric modifications on unique targets is a likely explanation for why the Rub1-activating enzyme strictly discriminates against ubiquitin. Swapping Rub1 and ubiquitin signals uncovered a reason for maintaining two separate pathways across eukaryotic kingdom.