Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Oct 19;171(3):696-709.e23.
doi: 10.1016/j.cell.2017.08.051. Epub 2017 Sep 28.

Chemical Proteomics Identifies Druggable Vulnerabilities in a Genetically Defined Cancer

Free PMC article

Chemical Proteomics Identifies Druggable Vulnerabilities in a Genetically Defined Cancer

Liron Bar-Peled et al. Cell. .
Free PMC article


The transcription factor NRF2 is a master regulator of the cellular antioxidant response, and it is often genetically activated in non-small-cell lung cancers (NSCLCs) by, for instance, mutations in the negative regulator KEAP1. While direct pharmacological inhibition of NRF2 has proven challenging, its aberrant activation rewires biochemical networks in cancer cells that may create special vulnerabilities. Here, we use chemical proteomics to map druggable proteins that are selectively expressed in KEAP1-mutant NSCLC cells. Principal among these is NR0B1, an atypical orphan nuclear receptor that we show engages in a multimeric protein complex to regulate the transcriptional output of KEAP1-mutant NSCLC cells. We further identify small molecules that covalently target a conserved cysteine within the NR0B1 protein interaction domain, and we demonstrate that these compounds disrupt NR0B1 complexes and impair the anchorage-independent growth of KEAP1-mutant cancer cells. Our findings designate NR0B1 as a druggable transcriptional regulator that supports NRF2-dependent lung cancers.

Keywords: KEAP1; NR0B1; NRF2; activity-based profiling; chemical proteomics; covalent; fragment; lung cancer; mass spectrometry.

Conflict of interest statement

Competing Financial Interests

The authors declare competing financial interests. B.F.C. is a founder and advisor to Vividion Therapeutics, a biotechnology company interested in using chemical proteomic methods to develop small-molecule drugs to treat human disease. T.A.P., N.I., P.J., Z.Z. and M.M.H. are employees of Pfizer.


Figure 1
Figure 1. Chemical proteomic map of NRF2-regulated cysteines in NSCLC cells
A) Proliferation of KEAP1-mutant (H2122) and KEAP1-WT (H1975) cells expressing shRNAs targeting NRF2 (shNRF2) or a control (shGFP), as determined by measuring intracellular ATP concentrations. Data represent mean values ± SD (n = 6/group). B) Immunoblot of NRF2 in shNRF2- or shGFP-H2122 cells. C) isoTOP-ABPP (R) ratios for cysteines in shNRF2- or shGFP-H2122 of -H1975 cells. Red data points mark R values ≥ 2.5, which was used as a cutoff for NRF2-dependent changes in cysteine reactivity. Average R values from n = 4–5 biological replicates per group are shown. See also Table S1. D) Distribution of proteins harboring NRF2-regulated cysteines by functional class. E) Distribution of NRF2-regulated cysteines reflecting changes in reactivity versus protein expression. See also Table S1. F) Representative proteins with NRF2-regulated changes in cysteine reactivity. Representative parent mass (MS1) profiles for tryptic peptides with IA-alkyne-reactive cysteines in shNRF2- (red) and shGFP- (blue) H2122 cells. Two cysteines are shown per protein, one with altered and the other with unaltered reactivity between shNRF2- and shGFP-H2122 cells. G) Representative MS1 profiles for cysteine-containing tryptic peptides in SQSTM1 in shNRF2- (red) and shGFP- (blue) H2122 cells (F). H) Immunoblot of GAPDH and PDIA3 expression in shNRF2- and shGFP-H1975 and H2122 cells. I) GAPDH activity in shNRF2- and shGFP-H2122 and -H1975 cells. Data represent mean values ± SD (n =16/group). ****p < 0.0001 for shNRF2 versus shGFP groups. J) Glycolytic flux is impaired in shNRF2-H2122 cells. ECAR = extracellular acidification rate. Data represent mean values ± SD (n = 20–26/group) from three biological replicates. ***p < 0.001, *p < 0.05 for shNRF2 versus shGFP groups.
Figure 2
Figure 2. Cysteine ligandability mapping of KEAP1-mutant and KEAP1-WT NSCLC cells
A) isoTOP-ABPP ratios (R values; DMSO/compound) for cysteines in H2122 cell (KEAP1-mutant) and H358 cell (KEAP1-WT) proteomes treated with DMSO or ‘scout’ fragments 2 or 3 (500 μM, 1 h). Red data points mark R values ≥ 5, which was used as a cutoff for defining liganded cysteines. Average R values from n = 3 biological replicates per group are shown. See also Table S3. B) Pie chart of NRF2-regulated genes/proteins in NSCLC cell lines denoting the subset that contain liganded cysteines (red). C) Cysteine ligandability map for representative NRF2 pathways. Blue marks proteins with liganded cysteines in NSCLC cells. ND, not detected. D) Circos plot showing the overlap in liganded cysteines between KEAP1-mutant (red) and KEAP1-WT (black) NSCLC cells. Gray and blue chords represent liganded cysteines found in both KEAP1-WT and KEAP1-mutant cell lines and selectively in KEAP1-mutant cell lines, respectively. Numbers in parenthesis indicate total liganded cysteines per cell line. E) Immunoblot of AKR1B10, CYP4F11 and NR0B1in shNRF2- and shGFP- H2122 cells.
Figure 3
Figure 3. Characterization of liganded proteins selectively expressed in KEAP1-mutant NSCLC cells
A) Heat map depicting RNAseq data in KEAP1-WT and KEAP1-mutant NSCLC cell lines for genes encoding NRF2-regulated proteins with liganded cysteines. RNAseq data obtained from (Klijn et al., 2015) (also see Figure S3A). B) NR0B1, AKR1B10, and CYP4F11 expression in lung adenocarcinoma (LUAD) tumors grouped by NRF2/KEAP1 mutational status. Data obtained from TCGA. C) Effect of knockdown of ARK1B10, CYP4F11, and NR0B1 on the anchorage-independent growth of H460 cells. Left: Immunoblot of the indicated proteins in H460 cells expressing the luciferase protein and indicated shRNAs. Dashed line represents a lane that was cropped from an immunoblot. Middle and right: Representative brightfield images (middle) and quantification (right) of growth of indicated cell variants in soft agar. Data represent mean values ± SD (n = 4–10/group) from at least two biological replicates. ****p < 0.0001, ***p < 0.001 for shRNAs targeting indicated genes vs shGFP. D) and E) Effect of CRISPR-generated knockout of NR0B1 (D) or CYP4F11 (E) on the anchorage-independent growth of H460 cells. Left: Immunoblot of NR0B1 or CYP4F11 in null clones (_1 and _2) or H460 cells expressing sgRNA CRISPR-CTRL and/or luciferase (denoted as H460-luc). Middle and right: Representative brightfield images (middle) and quantification of growth of indicated cell variants. Data represent mean values ± SD (n = 6–12/group) from at least two biological replicates. ****p < 0.0001, for NR0B1- or CYP4F11-null H460 clones vs sgRNA CRISPR-CTRL H460 cells. Scale bar equals 0.75 mm.
Figure 4
Figure 4. NR0B1 nucleates a transcriptional complex that supports the NRF2 gene-expression program
A) Intersection between NR0B1-regulated genes and transcriptional start sites (TSSs) bound by NR0B1. Outer circle: Chromosomes with cytogenetic bands. Middle circle: Whole genome plot of mapped NR0B1 reads (black) determined by ChIP-Seq corresponding to the transcriptional start sites (TSSs) of genes differentially expressed (up- (blue) or down- (red) regulated > 1.5-fold) in shNR0B1-H460 cells compared to shGFP-H460 cells (inner circle). See also Table S2. B) Overlap (left) and correlation (right) between genes up- (red) or down- (blue) regulated (> 1.5-fold) in shNR0B1- and shNRF2-H460 cells compared to shGFP-H460 control cells. r and p values were determined by Pearson correlation analysis. C) Heat map depicting RNAseq data for the indicated genes in shNR0B1-, shNRF2-, or shGFP-H460 cells. Expression was normalized by row. See also Table S2. D) Heat map representing NR0B1-interacting proteins in NSCLC cells. E) Endogenous NR0B1 co-immunoprecipitates with FLAG-RBM45 and FLAG-SNW1, but not control protein FLAG-RAP2A, in H460 cells, as determined by immunoblotting (left); right: schematic of NR0B1 protein interactions.
Figure 5
Figure 5. A covalent ligand targeting C274 disrupts NR0B1 protein complexes
A) Co-crystal structure of mouse NR0B1 (white) and LRH1 (burnt orange) from (Sablin et al., 2008) highlighting the location of C274 (orange) at the protein interaction interface that is also flanked by AHC mutations: R267, V269 and L278 (red). B) Left: Schematic for an NR0B1-SNW1 in vitro-binding assay. Right: immunoblot showing that NR0B1 interacts with SNW1, but not a control (METAP2) protein. C) Small molecule screen of electrophilic compounds (50 μM) for disruption of binding of FLAG-SNW1 to NR0B1 as shown in (B). Percentage of NR0B1 bound to SNW1 was normalized to vehicle (DMSO). A hit compound BPK-26 is marked in red. D) Structures of NR0B1 ligands (BPK-26 and BPK-29), clickable probe (BPK-29yne), and inactive control compounds (BPK-9 and BPK-27). E) BPK-26 and BPK-29, but not BPK-9 and BPK-27, disrupt the in vitro interaction of FLAG-SWN1 with NR0B1. F) BPK-29yne labels WT-NR0B1, but not an NR0B1-C274V mutant. HEK293T cells expressing the indicated proteins were treated with BPK-29 or vehicle (3 h) prior to treatment with BPK-29yne (30 min). Immunoprecipiated proteins were analyzed by in-gel fluorescence-scanning and immunoblotting. G) BPK-29 disrupts protein interactions for NR0B1-WT, but not a NR0B1-C274V mutant. HEK293T cells expressing HA-NR0B1-WT or HA-NR0B1-C274V proteins were treated with DMSO or BPK-29, after which lysates were generated and evaluated for binding to FLAG-SNW1, as shown in (B).
Figure 6
Figure 6. Characterization of NR0B1 ligands in KEAP1-mutant NSCLC cells
A) isoTOP-ABPP of H460 cells treated with NR0B1 ligands and control compounds (40 μM, 3 h). Dashed lines designate R values ≥ 3 (DMSO/compound), which was used as a cutoff to define cysteines liganded by the indicated compounds. Insets show MS1 profiles for C274 in NR0B1 for DMSO (blue) versus compound (red) treatment. Data are from individual experiments representative of at least three biological replicates. B) Venn diagram comparing the proteome-wide selectivity of NR0B1 ligands BPK-29 and BPK-26 and control compounds BPK-9 and BPK-27 in H460 cells as determined in (A). (See also Tables 1 and S3). C) BPK-29 and BPK-26 block the RBM45-NR0B1 interaction in H460 cells. H460 cells stably expressing FLAG-RBM45 were incubated with indicated compounds for 3 h, whereupon FLAG immunoprecipitates were performed and analyzed by immunoblotting. D) Concentration-dependent blockade of NR0B1 binding to FLAG-RBM45 by BPK-29 (left) and BPK-26 (right) in H460 cells. Experiments performed as described in (C). E) Effect of BPK-29 and control compounds (5 μM) on anchorage-independent growth of KEAP1-mutant H460 cells. Representative brightfield images (top) and quantification (bottom) of cell growth. Data represent mean values ± SD (n = 4–8 per group) and are representative of two to four biological replicates; ****p < 0.0001 for BPK-29 vs BPK-9 or BPK-27. Scale bar equals 0.75 mm. F) Ectopic expression of NR0B1-WT or NR0B1-C274V, but not control protein (RAP2A) partially and significantly rescues the impairment in anchorage-independent growth produced by BPK-29 (5.7 μM) in H460 cells expressing luciferase. Proliferation was determined by colony counting and data represent mean values ± SD (n = 3/group) and representative of at least two biological replicates; ****p < 0.0001 for FLAG-Rap2a, BPK-29 vs FLAG-NR0B1-C274, BPK-29; ***p < 0.001 for FLAG-RAP2A, BPK-29 vs FLAG-NR0B1-WT, BPK-29. Bottom: Immunoblot showing expressed NR0B1 (or RAP2A control) proteins. G) SILAC ratio plots for light amino acid-labeled cells (pulse phase) switched into media containing heavy amino acids for 3 h (chase phase) followed by proteomic analysis. Dashed line designates R values (light/heavy) of < 8, which was used as a cutoff for fast-turnover proteins. Inset shows MS1 peak ratio for NR0B1, which is among the top 5% of fast-turnover proteins. See also Table S1. H) Proteins regulated by NRF2 in NSCLC cells are enriched in fast-turnover proteins. Charts comparing fraction of NRF2-regulated genes (as determined by RNAseq) for which the corresponding proteins are designated as fast or slow turnover (as determined in G) further divided into groups showing reduced expression (left) or not (right) on day 1 following NRF2 knockdown (as determined by isoTOP-ABPP). See also Table S1.

Similar articles

See all similar articles

Cited by 31 articles

See all "Cited by" articles

MeSH terms