A Gatekeeper Residue for NEDD8-Activating Enzyme Inhibition by MLN4924 Julia I. Toth, Li Yang, Russell Dahl, Matthew D. Petroski  Cell Reports  Volume 1, Issue 4, Pages 309-316 (April 2012) DOI: 10.1016/j.celrep.2012.02.006 Copyright © 2012 The Authors Terms and Conditions

Cell Reports 2012 1, 309-316DOI: (10.1016/j.celrep.2012.02.006) Copyright © 2012 The Authors Terms and Conditions

Figure 1 HCT116 Cells Lose Sensitivity to the NAE Inhibitor MLN4924 upon Prolonged Exposure and Serial Culturing (A–C) Growth of HCT116 cells (HMS versus HMR cells) in the presence or absence of 3 μM MLN4924 (A). The viability of HMS and HMR cells treated for 48 hr with MLN4924 (B) or bortezomib (C) was assessed by quantifying cellular ATP levels. Error bars (n = 3) represent the SE. (D) HMS and HMR cells were treated with 3 μM bortezomib (B) or 3 μM MLN4924 (M) for the indicated times, and extracts were analyzed under nonreducing conditions by immunoblotting for NEDD8. NEDD8-modified cullins, NEDD8-activated UBC12, and NEDD8 are indicated. α-Tubulin was used as a loading control. See Figure S1 for immunoblotting for ubiquitin. (E) The steady-state abundance of CRL substrates NRF2, p27, CDT1, cyclin E, and p21 and the cell death marker cleaved PARP (clPARP) were examined after 24 hr of treatment with the indicated concentration of MLN4924. GAPDH was used as a loading control. Cell Reports 2012 1, 309-316DOI: (10.1016/j.celrep.2012.02.006) Copyright © 2012 The Authors Terms and Conditions

Figure 2 HMR Cells Are Heterozygous for a Single-Nucleotide Transition within the Coding Sequence of the UBA3 Nucleotide Binding Pocket (A) Domain structure of NAE subunits, UBA3 (UBE1C) and NAE1 (ULA1, APPBP1). (B) Chromatograms from sequencing reactions analyzing genomic DNA from HMS and HMR cells determined that HMR cells are heterozygous for a G to A transition (indicated by arrow) that alters an amino acid within the UBA3 nucleotide binding pocket. This transition results in the conversion of alanine 171 to threonine (A171T) in the UBA3 open reading frame. See Figure S2 for sequences of cDNAs and extended genomic DNA sequence. (C–G) A sequence alignment of canonical E1s indicates similarly positioned alanine residues (indicated by arrow) found in other E1s (C). The catalytic cysteines are highlighted in yellow. Molecular models of the NAE nucleotide binding pocket with ATP bound to UBA3 (D) or UBA3 A171T (E) and MLN4924-NEDD8 bound to UBA3 (F) or UBA3 A171T (G). The solvent-accessible nucleotide binding pocket is indicated as are side chains important for MLN4924 binding. Dot-surface representation for A171 and computer-modeled A171T is shown with only the rotamer with the highest frequency of occurrence in proteins. Computer-modeled A171T rotamers in (E) had significant van der Waals overlaps with other residues in UBA3, suggesting that it could alter the nucleotide binding pocket. Interestingly, the two major A171T rotamers modeled in (G) had significant van der Waals overlap with the aminoindane of MLN4924, but only minor contacts with other UBA3 residues. These models are based on PDB accession codes 1R4N (D and E) and 3GZN (F and G). Cell Reports 2012 1, 309-316DOI: (10.1016/j.celrep.2012.02.006) Copyright © 2012 The Authors Terms and Conditions

Figure 3 NAE (UBA3 A171T) Is Less Sensitive to MLN4924 and Activates NEDD8 More Efficiently In Vitro MLN4924 titrations (2-fold dilutions) in NAE ATP:PPi exchange assay using 100 μM ATP (A) or 1 mM ATP (B). Kinetic measurements of ATP:PPi exchange by NAE and NAE (UBA3 A171T) in ATP (C) and PPi (D) titrations. ATP titrations used 1 mM PPi with the indicated concentrations of ATP. PPi titrations used 2 mM ATP with the indicated concentrations of PPi. All reactions included 50 cpm/pmol of [32P]PPi, 100 μM NEDD8, and 10 nM of NAE or NAE (A171T), and were incubated at 37°C for 30 min. See Figure S3 for raw data. Error bars (n = 3) represent the SE. Cell Reports 2012 1, 309-316DOI: (10.1016/j.celrep.2012.02.006) Copyright © 2012 The Authors Terms and Conditions

Figure 4 Expression of UBA3 A171T Is Sufficient to Confer MLN4924 Resistance (A) HCT116 cells transfected with the indicated plasmids were either treated with DMSO (−MLN4924) or 3 μM MLN4924 for 72 hr prior to crystal violet staining. (B) The expression of FLAG-tagged UBA3 was examined by immunoblot of UBA3 and FLAG in transfected HCT116 cells. GAPDH was used as a loading control. (C) Expression of UBA3 A171T increases HCT116 cell survival in the presence of increasing concentrations of MLN4924. Cell viability measurements for HMS and HMR cells treated with MLN4924 are shown for comparison. Error bars (n = 3) represent the SE. Cell Reports 2012 1, 309-316DOI: (10.1016/j.celrep.2012.02.006) Copyright © 2012 The Authors Terms and Conditions

Figure S1 MLN4924 Uptake/Retention, Cytochrome P450 Family Gene Expression, and Evaluation of Ubiquitin System, Related to Figure 1 (A) We measured the intracellular levels of MLN4924 using liquid chromatography tandem mass spectrometry (LC-MS/MS). Extracts from HMS and HMR cells were analyzed at various times after addition of 3 μM MLN4924 using indomethacin as an internal standard. The peak area (PA) of MLN4924 and indomethacin were quantified and expressed as a ratio of MLN4924 PA:Standard PA. Data points represent the mean (n = 3) with SEM error bars. (B) Transcripts from HMS (DMSO treated or treated with 3 μM MLN494 for 8 or 24 hr) and HMR (grown constitutively with 3 μM MLN4924) cells were profiled using an Illumina Human HT-12 v4 Exprssion BeadChip microarray. After standardization, data sets were filtered for cytochrome P450 genes and clustered to evaluate changes in expression between the different cell lines and treatments. A heat map summarizing the experiment is shown with 3 replicates. Hierarchical clustering of standardized data did not identify any significant differences in expression of these genes in HMR cells relative to untreated HMS cells or HMS cells treated for 8 or 24 hr with 3 μM MLN4924. (C) Cell extracts from HMS and HMR cells treated with bortezomib (B) and MLN4924 (M) for the indicated times were analyzed by immunoblot for ubiquitin conjugates. GAPDH was used to verify normalized protein loading. High molecular weight ubiquitin conjugates are indicated. Cell Reports 2012 1, 309-316DOI: (10.1016/j.celrep.2012.02.006) Copyright © 2012 The Authors Terms and Conditions

Figure S2 Sequences of UBA3 cDNAs Cloned from HMR Cells and Extended Genomic DNA Sequences, Related to Figure 2 (A) Chromatograms of UBA3 cDNAs cloned from HMS and HMR cells identify a single nucleotide transition found in HMR cDNAs. Sequences from nucleotide 463 to 561 are shown, with the identified G to A transition at 511 found in 3 out of 6 cDNAs from HMR cells indicated. All other sequences between HMS and HMR cDNAs were identical. The resulting amino acid sequences with the corresponding change in amino acid 171 from alanine to threonine found in HMR cells are shown. Numbering is based on UBA3 isoform 1. (B) PCR amplified genomic DNAs corresponding to 69,112,979 to 69,112,115 of chromosome 3 from HMS and HMR cells were sequenced. The shown chromatograms (69,112,664 to 69,112,380) are an extended version of Figure 2B and include exon 8 (boxed) and flanking intron sequences. The heterozygous purine transition (R, guanine and adenine) at 69,112,616 in HMR cells, but not HMS cells, is indicated. Cell Reports 2012 1, 309-316DOI: (10.1016/j.celrep.2012.02.006) Copyright © 2012 The Authors Terms and Conditions

Figure S3 Raw Data from ATP:PPi Exchange Assays, Related to Figure 3 Phosphor screen scans show the NEDD8-dependent ATP generated by NAE and NAE (UBA3 A171T) in MLN4924 titrations using either 100 μM (A) or 1 mM ATP (B). These experiments used 10 nM NAE or NAE (UBA3 A171T), 100 μM (top panel) or 1 mM (bottom panel) ATP, 100 μM PPi (with 50 cpm/pmol [32P] PPi, and 1 μM NEDD8. MLN4924 dilutions (11 2-fold serial dilutions in 0.025% final DMSO concentration or DMSO only) were added to reactions containing NAE, ATP, and PPi with NEDD8 added to start the reaction. Reactions were incubated for 30 min at 37°C prior to ATP purification on activated charcoal filter paper and filter processing. Known ATP concentrations (containing 50 cpm/pmol [α-32P] ATP) in duplicate per filter (3-fold serial dilutions from 617 pmol to 2.54 pmol) were used to generate an ATP standard curve for quantification of NAE- and NEDD8-dependent ATP generation. (C) NEDD8-dependent NAE and NAE (UBA3 A171T) activity was measured in the presence of increasing concentrations of ATP in the ATP:PPi exchange assay. These experiments were performed as above, but using 1 mM PPi and 2-fold serial dilutions of ATP from 2 mM. (D) PPi titrations were performed to examine the NEDD8-dependent activities of NAE and NAE (UBA3 A171T) using 11 2-fold serial dilutions from 1 mM PPi and 10 nM or NAE (UBA3 A171T), 2 mM ATP, and 1 μM NEDD8. Cell Reports 2012 1, 309-316DOI: (10.1016/j.celrep.2012.02.006) Copyright © 2012 The Authors Terms and Conditions