Fig 2: NEDD8 is present on the UBE1 catalytic cysteine in vivo, and NEDD8 overexpression attenuates proteasomal degradation of a reporter substrate(A) Untagged mature NEDD8 was co-expressed with WT HA–UBE1 as well as with HA–UBE1(C632S; OXY), which forms a non-reducible oxyester with the UBL. Immunoprecipitation reveals reducible thioester (WT) and non-reducible oxyester (OXY) adducts with NEDD8. In reducing conditions, free NEDD8 can be seen to fall off the WT HA–UBE1, but not the oxyester mutant, demonstrating that NEDD8 resides on the catalytic cysteine residue of this enzyme. Higher molecular mass forms of NEDD8 are also apparent after reduction from UBE1 WT, possibly corresponding to forms of NEDD8 more efficiently activated by UBE1. (B) The same experiment as described in (A) was performed for WT HA–UBE1, but without simultaneous NEDD8 overexpression, to investigate if endogenous NEDD8 would be available for activation upon UBE overexpression. A reducible thioester adduct with endogenous NEDD8 is apparent after immunopreciptiation (IP), demonstrating that endogenous NEDD8 can be activated when UBE1 is available. (C) scNEDD8 overexpression attenuates degradation of a Trp1 reporter substrate. Serial dilution of Trp1 reporter yeast carrying galactose-inducible HA–scNEDD8 or HA–scNEDD8?GG. Induction of HA–scNEDD8 restores growth on -TRP plates, indicating stabilization of the Trp1 reporter. Deletion of the yeast N-end rule E3 ligase Ubr1 stabilizes Trp1 and results in loss of tryptophan auxotrophy. (D) Bortezomib treatment induces atypical NEDDylation. U2OS cells were treated with 3 µM bortezomib or DMSO (vehicle) for 4 h, and with siRNA to UBE1 and MLN4924 as indicated. Widespread UBE1-dependent NEDDylation is apparent after bortezomib treatment. Molecular masses in kDa are shown to the left-hand side of the Western blots. CTRL, control.

Fig 3: Genetic alleles of the NEDD8 pathway do not affect atypical NEDDylation(A) Mature NEDD8 and NEDD8?GG were overexpressed in CHO TS41 cells harbouring a temperature-sensitive allele of NAE, and incubated at the permissive (32°C) and restrictive (39°C) temperature, as indicated. A shift to the restrictive temperature does not affect atypical NEDDylation, suggesting that NEDD8 conjugates are independent of NAE. Importantly, cullin NEDDylation is strongly reduced at the restrictive temperature. Expression of NEDD8?GG does not induce atypical NEDDylation, indicating that smears observed upon mature NEDD8 expression correspond to conjugates to proteins. (B) Atypical NEDDylation by UBE1 is conserved in S. cerevisiae. A Western blot using an anti-HA antibody of cells expressing endogenous levels of untagged or HA-tagged scNEDD8 is shown. The bands detected correspond (by molecular mass) to NEDDylated cullin proteins. The asterisk denotes the antibody cross-reacting band. (C) HA–scNEDD8 forms ectopic conjugates upon overexpression (GAL promoter), whereas the ?GG version does not. (D) Deletions of the scNEDD8 E2 Ubc12 and E1 Uba3 affect cullin NEDDylation but not atypical NEDDylation, showing that atypical NEDDylation is strictly independent of the NEDD8 enzymes. (E) Overexpression of scNEDD8 in a temperature-sensitive UBA1 mutant (uba1–204) at the restrictive temperature diminishes atypical NEDDylation, indicating that the ubiquitin E1 enzyme is mediating the formation of atypical NEDD8 conjugates. Molecular masses are shown in kDa to the left-hand side of the Western blots.

Fig 4: The neddylation pathway is overactivated in patient-derived glioblastoma specimens. (A) Gene expression levels that related neddylation pathway across normal brain and glioblastoma specimens on TCGA and Rembrandt data set. (B) Kaplan-Meier analysis for survival on a Rembrandt dataset of glioblastoma patients according to their NEDD8 levels. Data analyzed using log-rank (Mantel-Cox) test. (C) Immunohistochemical staining of various glioblastoma patient-derived xenograft model specimens using NEDD8-specific antibody. Negative control was stained without primary antibody. Bars represent 100 µm.

Fig 5: Perturbing the NEDD8 to ubiquitin ratio leads to UBE1-dependent NEDDylation(A) Ratios of free NEDD8 to ubiquitin under the indicated conditions, and after expression of different forms of NEDD8. Error bars represent ± S.D. (B) Pattern of endogenous NEDD8 conjugates in U2OS cells under non-reducing [-BME (2-mercaptoethanol)] and reducing conditions (+BME). Main bands correspond by molecular mass to neddylated cullins and thioesters with the NEDD8 E2s and E1. (C) NEDD8 overexpression induces strong ectopic NEDDylation (lanes 3 and 4). This type of NEDDylation is resistant to MLN4924 treatment, but strongly reduced by down-regulation of UBE1 by siRNA. UBE1 is detected in the anti-NEDD8 antibody blot due to a prior Western blot using UBE1 antibody. (D) MG132 treatment induces atypical NEDDylation. U2OS cells were treated with 30 µM MG132 for 4 h, and with siRNA to UBE1 and MLN4924 as indicated. Widespread UBE1-dependent NEDDylation is apparent after MG132 treatment. Molecular masses in kDa are shown to the left-hand side of the Western blots. NT, non-transfected.