Targeting Oncogene Promoters and Ribosomal RNA Biogenesis by G-Quadruplex Binding Ligands Translate to Anticancer Activity

Abstract

G-Quadruplex (GQ) nucleic acids are promising therapeutic targets in anticancer research due to their structural robustness, polymorphism, and gene-regulatory functions. Here, we presented the structure-activity relationship of carbazole-based monocyanine ligands using region-specific functionalization with benzothiazole (TCA and TCZ), lepidine (LCA and LCZ), and quinaldine (QCA and QCZ) acceptor moieties and evaluated their binding profiles with different oncogenic GQs. Their differential turn-on fluorescence emission upon GQ binding confirmed the GQ-to-duplex selectivity of all carbazole ligands, while the isothermal titration calorimetry results showed selective interactions of TCZ and TCA to c-MYC and BCL-2 GQs, respectively. The aldehyde group in TCA favors stacking interactions with the tetrad of BCL-2 GQ, whereas TCZ provides selective groove interactions with c-MYC GQ. Dual-luciferase assay and chromatin immunoprecipitation (ChIP) showed that these molecules interfere with the recruitment of specific transcription factors at c-MYC and BCL-2 promoters and stabilize the promoter GQ structures to inhibit their constitutive transcription in cancer cells. Their intrinsic turn-on fluorescence response with longer lifetimes upon GQ binding allowed real-time visualization of GQ structures at subcellular compartments. Confocal microscopy revealed the uptake of these ligands in the nucleoli, resulting in nucleolar stress. ChIP studies further confirmed the inhibition of Nucleolin occupancy at multiple GQ-enriched regions of ribosomal DNA (rDNA) promoters, which arrested rRNA biogenesis. Therefore, carbazole ligands act as the "double-edged swords" to arrest c-MYC and BCL-2 overexpression as well as rRNA biogenesis, triggering synergistic inhibition of multiple oncogenic pathways and apoptosis in cancer cells.

Figure 1

Design of carbazole derivatives. (A) Chemical structures of carbazole-based monocyanine derivatives with a donor D carbazole moiety appended to various acceptor A moieties resulting in (B) TCA, LCA, QCA, TCZ, LCZ, and QCZ. Left panel: Normalized fluorescence intensity (NFI) of carbazole derivatives measured in 20 mM PBS, 100 mM KCl, pH 7.4, at 24 °C. λ_em, maximum fluorescence emission wavelength of carbazole ligands.

Figure 2

Selective recognition of GQ with turn-on fluorescence response of carbazole ligands. Relative fluorescence intensities (RFIs) of carbazole ligands (A) TCA, (B) LCA, (C) QCA, (D) TCZ, (E) LCZ, and (F) QCZ (5 μM) titrated with varying concentrations of GQ or duplex (0–30 μM) forming DNA sequences measured in phosphate buffer (20 mM PBS, 100 mM KCl, pH 7.4, at 24 °C). Error bars present mean ± SE (N = 3).

Figure 3

Apoptosis induction by carbazole ligands. (A) Flow cytometric analysis of an Annexin V-FITC/PI dual staining assay after treatment with TCA/LCA/QCA (1 and 10 μM) and TCZ/LCZ/QCZ (500 nM and 1 μM) for 24 h in MDAMB-231 cancer cells; LL, LR, UR, and UL indicate the live cells and early, late apoptotic, and necrotic cells, respectively. (B) Bar plots denote the estimation of live, early apoptotic, and necrotic or dead cells. FITC: fluorescein isothiocyanate and PI: propidium iodide. (C) Western blot images showing expression of different apoptotic markers under increasing concentrations of TCA and TCZ in MDAMB-231 cells at 24 h. Actin was used as the housekeeping protein. (D) Bar plots determine the fold change expression of apoptotic markers compared to the control based on densitometric analyses. Error bars present means ± SE (N = 3).

Figure 4

Selectivity of carbazole derivatives for different oncogenic GQs in MDAMB-231 cells. (A) Schematic diagrams of oncogene promoter sequences with upstream GQ-forming elements, cloned into a promoter-less pGL4.72[hRlucCP] luciferase vector. hRluc, Renilla luciferase gene; P1 and P2, promoters of oncogenes. Evaluation of c-MYC, BCL-2, KRAS, and VEGF-A promoter activities under 24 h of treatment of (B) TCA and (C) TCZ using the reporter plasmids with or without the respective wild-type GQ-forming sequences. Relative promoter activity determined by normalizing the Rluc/Fluc values to that of the cells transfected with oncogene promoter constructs (GQ-null), having no GQ-forming motif. Error bars represent mean ± SE (N = 5). (D) Semiquantitative RT-PCR analysis and (E) RT-PCR showing the expression profiles of c-MYC, BCL-2, KRAS, and VEGF-A transcription upon 24 h of treatment of TCA/LCA/QCA (1 and 10 μM) and TCZ/LCZ/QCZ (500 nM and 1 μM) in MDAMB-231 cancer cells. M, 100 bp ladder; Con, control. Quantification of the transcripts’ level relative to the control (GAPDH). Error bars represent mean ± SE (N = 3). Statistical differences determined compared to the control by two-tailed Student’s t test (*P < 0.05, **P < 0.01, ***P < 0.001, n, statistically nonsignificant). (F) ITC binding profiles of TCZ and TCA with c-MYC and BCL-2 promoter GQ sequences, respectively, in 10 mM potassium phosphate buffer (pH 7.0), 100 mM KCl at 25 °C. The solid line represents the best fit data using “one site binding model”. (G) Temperature dependence of heat capacity changes (ΔC_p) at constant pressure measured for TCA-GQ or TCZ-GQ complex formation.

Figure 5

Role of TCA and TCZ to regulate GQ-mediated transcription of BCL-2 and c-MYC. (A) Schematic of the c-MYC promoter with the GQ-forming region (Pu27) in the NHE III₁ region. ChIP primers were designed at the −453F to −215R region. F, forward and R, reverse primers. (B) ChIP experiments using NM23-H2-, Sp1-, and Nucleolin (NCL)-specific antibodies to evaluate their occupancy at the c-MYC promoter in MDAMB-231 cells under 1 μM TCZ treatment for 24 h. The input fraction indicates total DNA; negative control immunoprecipitation uses rabbit IgG showing no signal. (C) Quantification of Sp1, NCL, and NM23-H2 occupancy at the c-MYC promoter by the percentage of the input method based on quantitative real-time PCR under 1 μM TCZ treatment. (D) ChIP experiments using NCL-specific antibodies and quantification of NCL occupancy at the c-MYC promoter in MDAMB-231 cells. rhNM23-H2, recombinant NM23-H2 treatment for 24 h. (E) ChIP experiments using NM23-H2-specific antibodies and quantification of NM23-H2 occupancy at the c-MYC promoter in MDAMB-231 cells. Statistical difference for with or without TCZ in rhNM23-H2- or siNCL-treated cells calculated by Student’s t test. (##P < 0.01) (F) Schematic representation of the BCL-2 promoter having GQ-forming motif (Pu39) upstream P₁ and P₂ promoters. Target-specific primers for ChIP experiments were designated at the −1550F to −1441R region. (G) ChIP experiments using WT1-specific antibodies to evaluate its occupancy at the BCL-2 promoter in MDAMB-231 cells under 1 μM TCA treatment for 24 h. (H) Quantification of WT1 at the BCL-2 promoter by the percentage of the input method based on quantitative real-time PCR. Error bars represent mean ± SE (N = 3). Statistical experiments for ChIP studies were performed by a one-way ANOVA Tukey–Kramer test. (*P < 0.05, **P < 0.01, ***P < 0.001).

Figure 6

Cellular localization and lifetime decay of carbazole ligands. (A) Localization of TCA and TCZ (green) at different concentrations (0.5 and 1 μM) in MDAMB-231 cells by colocalizing with the Alexa 647-conjugated Nucleolin (red) antibody at 1 h timepoint. (B) Confocal images showing colocalization of TCA and TCZ (0.5 μM; green) with GQ selective antibody BG4 (red). Nuclei were stained with DAPI (blue). Scale bar: 10 μm. (C) Fluorescence lifetime decay profiles of TCA and TCZ in the presence of GQ and duplex DNA at a 1:5 stoichiometric ratio.

Figure 7

Ribosome biogenesis inhibition in MDAMB-231 cells. Confocal images showing (A) Nucleolin distortion and (B) nucleolar displacement upon treatment with TCA/TCZ (1 μM) at 12 and 24 h. Scale bar: 10 μm. (C) ChIP results using Nucleolin-specific antibodies. Quantification of Nucleolin occupancy (% of input) at different rDNA regions (−2905, +5605, +48, and +12855) upon treatment with increasing concentration of TCA and TCZ for 12 h. Error bars represent mean ± SE (N = 3).