Investigation of targets and anticancer mechanisms of covalently acting natural products by functional proteomics

Abstract

Eriocalyxin B (EB), 17-hydroxy-jolkinolide B (HJB), parthenolide (PN), xanthatin (XT) and andrographolide (AG) are terpenoid natural products with a variety of promising antitumor activities, which commonly bear electrophilic groups (α,β-unsaturated carbonyl groups and/or epoxides) capable of covalently modifying protein cysteine residues. However, their direct targets and underlying molecular mechanisms are still largely unclear, which limits the development of these compounds. In this study, we integrated activity-based protein profiling (ABPP) and quantitative proteomics approach to systematically characterize the covalent targets of these natural products and their involved cellular pathways. We first demonstrated the anti-proliferation activities of these five compounds in triple-negative breast cancer cell MDA-MB-231. Tandem mass tag (TMT)-based quantitative proteomics showed all five compounds commonly affected the ubiquitin mediated proteolysis pathways. ABPP platform identified the preferentially modified targets of EB and PN, two natural products with high anti-proliferation activity. Biochemical experiments showed that PN inhibited the cell proliferation through targeting ubiquitin carboxyl-terminal hydrolase 10 (USP10). Together, this study uncovered the covalently modified targets of these natural products and potential molecular mechanisms of their antitumor activities.

Keywords: ABPP; eriocalyxin B; natural products; parthenolide; proteomics.

Fig. 1. EB, HJB, PN, XT, AG inhibited breast cancer cell proliferation.

a Structure of EB, HJB, PN, XT and AG. b Dose-dependent effects on MDA-MB-231 cells evaluated at 72 h of treatment in three biological replicates. Data were presented as mean ± SEM. c IC₅₀ value of five natural products. Data were presented as mean ± SD.

Fig. 2. Characterization of global protein expression for five natural products.

(a) The workflow of the TMT-based quantitative proteomic profiling of MDA-MB-231 cells with treatment of DMSO, EB, HJB, PN, XT and AG. (b) The shared GO Biological Process enrichment results of five natural products. (c) KEGG pathway enrichment analysis of significantly differentiated proteins (P value < 0.05, fold change > 1.2) of five natural products. The global heatmap showing the enriched pathways with each natural products compared to DMSO treated in MDA-MB-231 cells. The color is according to FDR value, and the darkest blue represents N/A. (d and e) Protein-protein interaction network of significantly differentially expressed proteins of EB and PN, respectively. The top 3 clusters of highly interconnected networks were exhibited. Interaction networks of EB/PN-regulated proteins were listed in gene names, based on STRING database. Cluster 1 of EB: MCODE score = 30.5, nodes = 41, edges = 610. Cluster 2 of EB: MCODE score = 17, nodes = 17, edges = 136. Cluster 3 of EB: MCODE score = 16, nodes = 16, edges = 120. Cluster 1 of PN: MCODE score = 17, nodes = 17, edges = 136. Cluster 2 of PN: MCODE score = 16, nodes = 16, edges = 120. Cluster 3 of PN: MCODE score = 14, nodes = 14, edges = 91. GO Biological Process enrichment analysis was performed in each cluster.

Fig. 3. ABPP to map EB and PN targets in MDA-MB-231 cells.

a Competitive ABPP for quantitative mapping of potential covalent binding sites of targets. Treatment of cells with DMSO or EB/PN in situ, proteome labeling with an IA probe, Cu-catalyzed based click chemistry to incorporate biotin group, trypsin digestion, enrichment of probe-modified peptides with streptavidin, followed by sodium dithionite cleavage for LC-MS/MS analysis. b Distribution of competitive ABPP ratios (R values). Probe-modified peptides only occurred in heavy samples with good reproducibility, for which R values were assigned to maximum value. R value greater than 5 was shown by dashed lines to mark potential binding sites that exhibited high sensitivity to EB/PN. (c) Percentage of potential targets of EB and PN found in DrugBank proteins or Non-DrugBank proteins, and functional classification. d Overlap of potential targets with tractability group1/2/3 reported by other researchers. e Protein-protein interaction networks of the potential targets of EB and PN were analyzed respectively. The top 4 clusters of highly interconnected networks were performed GO Biological Process enrichment analysis. f The Venn diagram of shared binding sites and potential targets between EB and PN.

Fig. 4. Biochemical validation of USP10 as a target of PN.

a Western blotting analyses of the whole cell lysate using anti-ubiquitin (pan) antibody were performed in control and PN-treated samples. Coomassie blue staining was used as the loading control. b The hypothesis that PN increased ubiquitin signals. c The critical regions for USP10 and sequence alignment of the potential binding site of Cys40 on USP10 among different species. d PN covalently bound to Cys40 on USP10. Purified USP10 (100 nM) was incubated with PN (10 μM) at room temperature for 30 min prior to LC-MS/MS analysis. e IA-Fluorescein labeling of WT and C40A USP10. Labeled proteins were scanned with Typhoon FLA 9500 scanner and silver staining was used as the loading control. f DUB activity assay kit was applied to measure USP10 activity. (g) USP10-knockdown MDA-MB-231 cells were constructed by lentiviral particles containing USP10 shRNA. USP10 expressions were validated by Western blotting of USP10-knockdown (KD) cells compared to negative control (NC) cells. Actin expressions were shown as loading control. The shown gel was from three biological replicates. Then, PN effects were assessed on USP10-knockdown cells and negative control cells. The shown data were a representative one from biological replicates. Data was shown as mean ± SEM. Statistical significance was calculated with unpaired two-tailed Student’s t-tests. One asterisk indicates P < 0.05, two asterisks indicate P < 0.01 and three asterisks indicate P < 0.001.