G-quadruplex structural transition driven by a platinum compound.

G-quadruplex (G4) transitions play integral roles in regulating biological functions and can be modified by ligands. However, little is known about G4 transitions. Herein, we reveal distinct pathways of a platinum(II) compound Pt-phen converting parallel-stranded MYC G4 to a hybrid-type structure. Three NMR structures, 1:1 5'-end binding, 1:1 3'-end binding and 2:1 Pt-phen-MYC G4 complexes, were determined by NMR. We find that Pt-phen drives G4 transition at a low ratio. Under physiological 100 mM K+ conditions, a significant stable hydrogen-bonded T:T:A triad is formed at 3'-end of hybrid-type Myc1234, and consequently, Pt-phen first binds the 5'-end to form a 1:1 5'-end binding complex and then disrupts the 3' T:T:A triad and binds 3'-end to form a 2:1 complex with more Pt-phen. Remarkably, the G4 transition pathway is different in 5 mM K+ with Pt-phen first binding the 3'-end and then the 5'-end. 'Edgewise-loop and flanking/ligand/G-tetrad' sandwich structure formation and terminal T:T:A triad stabilization play decisive roles in advancing and altering transition pathways. Our work is the first to elucidate the molecular structures of G4 transitions driven by a small molecule. The ligand-driven G4 transition is a dynamic process that includes a quick G4 transition and multiple complexes formation.

INTRODUCTION

Structural regulation of nucleic acid secondary structure is a thriving area of research, and many DNA-based transitions have been studied to date (1–3). G-quadruplexes (G4s) are noncanonical nucleic acid secondary structures formed by stacked G-tetrads with Hoogsteen hydrogen bonding and cations (4). G-rich sequences able to form G4 structures are prevalent in biologically significant regions, such as human telomeres and oncogene promoter regions (5–6). Their formation and transition are involved in many important physiological and pathological processes and the development of cancer or other diseases (7–9), receiving great attention as potential biomedical targets for drug development (10). The MYC proto-oncogene, as one of the research hotspots in molecular biology in recent years, has been proven to be important in the malignant progression of many human tumors (11). The MYC promoter nuclease hypersensitive element (NHE) III1 region controls 80–95% of MYC expression and can form DNA G4 structures (12). As a result, inhibition of MYC transcription by G4 structures is an attractive strategy for targeting the MYC gene for anticancer therapy (13).

Of great interest is that the G4 has structural diversity, which points to a variety of potential applications in nanodevices (14–15) and controlling G4 functions in biological settings (16–17). Sequencing of G4 indicates ∼700 000 DNA sequences in the human genome that can form various G4 structures (18), which challenges our understanding of G4s. However, the conformational reaction pathways and structural information of G4 transitions by ligands are not well understood. It is crucial to elucidate the structural information on the dynamic transitions of G4 structures to provide a structural and scientific basis for revealing G4 formation and evolution, designing G4-targeting agents, and understanding the relationship between G4 and diseases. In fact, reported studies have demonstrated that the structures of G4 are closely related to their biological functions (16,19–20). Helicases such as DHX36 display conformation-selective binding with a preference for parallel-stranded G4s (19–20).

Studies have shown that the G4 structure can be converted from one folding pattern to another by cations, molecular crowding conditions or small molecule modifications (21–25). To date, various small molecules have been developed and shown to have the ability to convert the conformation of biologically significant DNA G4s (26–29). The structural transition of G4 induced by ligands is complicated. Clearly, to the best of our knowledge, accurate structures of G4 transitions with ligands are challenging to obtain and have not been reported. Previous studies have demonstrated that NMR technology is a powerful tool for exploring and characterizing the binding of G4s with small molecules at the structural level (30–39).

Herein, we report that the platinum(II) compound Pt(1,10-phenanthroline)2 (Pt-phen) (40) binds the MYC promotor G4 (41) and drives its structural transition from a parallel-stranded to hybrid-type conformation (Figure 1). Using the NMR method, various pathways of structural transition of Myc1234 G4 bound to Pt-phen are unambiguously characterized under physiological 100 mM K+ and 5 mM K+ conditions. The molecular structures of Pt-phen–Myc1234 G4 complexes show that the hybrid-type Myc1234 G4 allows a greater stacking of loop and flanking residues, and this highly extensive π–π stacking interaction in the Pt-phen sandwiched multilayer complex and the formation of terminal hydrogen bonding T:T:A non-G-triad dependent on K+ concentrations likely drive the transition of Myc1234 G4. This paper presents an integrated work showing not only the structural transition of G4 by ligand modifications but also the binding with multiple transition pathways, providing structural basis for G4 polymorphism.

Figure 1.

The conformational transition of MYC G4 from parallel-stranded to hybrid-type structure induced by Pt-phen. The chemical structure of Pt-phen was shown. Deep pink box = (syn) guanine, light pink box = (anti) guanine, green ball = adenine, blue ball = thymine.

MATERIALS AND METHODS

Sample preparation

DNA sequences were obtained from Sangon Biotech Co., Ltd. (Shanghai) and purified by HPLC. The 8%-enrichment site-specific 15N-labeled oligonucleotides were synthesized using 15N-labeled dG-phosphoramidite from Cambridge Isotope Laboratories (USA). DNA oligonucleotides in buffer were annealed to form the G4 structures by heating at 95 °C for 5 min and then slowly cooling to room temperature overnight. The concentration of DNA was tested by Nanodrop 200/200c (Thermo Science). The Pt-phen ligand was synthesized as described in our previous report (40).

NMR experiments

The NMR experiments were performed on Bruker AVIII 600/700 MHz spectrometers equipped with a cryoprobe. DNA samples were prepared in D2O/H2O (10%/90%) or D2O (99.8%). The concentration of DNA samples was 0.1–1.0 mM. In the 1D NMR titration experiments, after each addition of Pt-phen, the solution was stirred and allowed to equilibrate for 10 min. The imino H1 protons of G-tetrad guanines and H3 protons of hydrogen bonding thymines were assigned by the 1H-15N HMQC experiments. The 2D COSY, TOCSY and NOESY experiments were collected at different temperatures, 5, 15 and 25 °C, for the Pt-phen−Myc1234 complexes. The NOESY mixing times were 50−300 ms, and the TOCSY mixing times were 40 ms. Water suppression techniques (WATERGATE or presaturation) were used for H2O NMR samples. Peak assignments and integrations were performed by Sparky (UCSF).

NOE-distance restrained molecular dynamics simulation

Proton distances were obtained based on the NOE integrated at NOESY spectra with ± 20% variance. The distance of Me-H6 (2.99 Å) in thymine was set as the reference. The X-PLOR program and Accelrys Discovery Studio 2.5.5 were used to perform the structure calculations. The Pt-phen molecule was optimized using Gaussian 03, and the Pt-phen topology and parameters were generated by hand. The starting models of the 1:1 and 2:1 Pt-phen–Myc1234 complexes were constructed in the Discovery Studio program based on the NOE data. In Accelrys Discovery Studio, these models were first minimized and equilibrated and then soaked into a 15 Å water layer, and a dynamic simulation was run with the CHARMM force field. NOE-restrained simulation annealing refinement calculations were conducted in the XPLOR program. The whole system was put into energy minimization of 1000 steps and then equilibrated at 1000 K for 10 ps. Then, the molecular dynamics and cooling simulation were started with a 25 K reduction in temperature, each cycle with 1000 time-steps of 2 fs, until the temperature reached 300 K. Finally, the structure systems were subjected to energy minimization of 1000 steps. The best structures were selected based on the minimum energy and number of NOE violations.

CD spectroscopy

CD spectra were measured in a J-810 spectropolarimeter (JASCO, Japan) at room temperature. The wavelength range of 220−360 nm, 1 cm optical path length, and 200 nm/min scan speed were set. The buffer used to prepare the DNA G4 samples was 10 mM Tris–HCl in the presence of different concentrations of KCl and NaCl (pH 7.4). The DNA concentration was 3 μM. After each addition of Pt-phen, the samples were stirred and equilibrated for 30 min. Data analysis was performed using Origin 8.5 (OriginLab Corp.).

Native polyacrylamide gel electrophoresis (PAGE)

The native gel contained 18% acrylamide with 1× TBE buffer (pH 8.0). DNA G4 samples were prepared in a buffer comprising 10 mM Tris-HCl and different concentrations of KCl (100, 50 and 10 mM) with a pH of 7.4. Different ratios of Pt-phen were added to DNA G4 samples. Each sample included 0.3 nM DNA. GelRed was used for DNA band staining. The gels were photographed in a FluorChem M imager (ProteinSimple).

Fluorescence lifetime imaging microscopy (FLIM)

FLIM imaging of NBTE (4,4',4''-(nitrilotris(benzene-4,1-diyl))tris(1-ethylpyridin-1-ium)iodide) (37) and Pt-phen was performed using Zeiss LSM 880 NLO multiphoton microscopes equipped with a bh TCSPC FLIM system (Becker & Hickl GmbH, Berlin, Germany). The two-photo excitation of samples was performed by a femtosecond Ti:sapphire laser (Coherent Chameleon), and the emission was collected using a Zeiss BiG-2 GaAsP detector. The TCSPC module type is an SPC-150 module, and the resolution of the images is 512 × 512 pixels. For FLIM imaging, HeLa cells were incubated with 20 μM NBTE and Pt-phen (or PDS) at 37 °C for 24 h, and the control group was incubated with only 20 μM NBTE. Then, the dish was washed three times with ice-cold PBS and detected by the FLIM system. The wavelength for two-photon excitation of NBTE was 810 nm, and the emission was recorded from 545 to 590 nm. The acquisition time was sufficient to obtain a better SNR. Lifetime data were fitted to a tri-exponential function for each pixel using SPC-Image software.

Cell viability assay

The cytotoxicity of Pt-phen toward different cell lines was determined by MTT assay. The cells were seeded in 96-well plates (Corning) and grown overnight. Then, the cells were treated with a series of concentrations of Pt-phen and cisplatin. After incubation for 44 h, 20 μl MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide, Sigma) solution (5 mg/ml in PBS) was added to each well. The plates were incubated for an additional 4 h before the media was carefully removed, and DMSO was added (150 μl per well). After shaking for 5 min, the absorbance at 595 nm was measured using a microplate reader (Infinite M 200 Pro, Tecan, Männedorf, Switzerland). Each experiment was repeated at least three times to obtain the mean values.

Western blotting

The cells (HeLa and LO2) were seeded into 60 mm tissue culture dishes and incubated under a 5% CO2 atmosphere at 25 °C for 24 h. Cells were treated with 5.0 μM and 10.0 μM Pt-phen for different times. After pelleting, cells (HeLa and LO2) were solubilized in 1× RIPA buffer supplemented with 1× protease inhibitor cocktail (Roche, USA). The nucleus and intracellular organelles were disrupted by sonication five times for 10 s each. After incubation for 20 min at 4 °C, the suspensions were separated by centrifugation at 13 000 rpm for 10 min at 4 °C, and the supernatant was collected. The protein bands were detected using anti-MYC (D84C12, Cell Signaling Technology, USA) and anti-GAPDH (FL-335, Santa Cruz Biotechnology, USA) antibodies. An equal amount of protein (20 μg) was electrophoresed on a 12% SDS-PAGE gel and transferred to a 0.45 μm PVDF membrane (Millipore, USA) at 200 V for 2 h. The membrane was blocked with QuickBlock Western buffer (Biyuntian, China) for 1 h at room temperature. The membrane was incubated overnight at 4 °C with the primary antibodies anti-MYC (9E10, Santa Cruz Biotechnology, USA) and anti-GAPDH (FL-335, Santa Cruz Biotechnology, USA). After three washes in TBST, the membrane was incubated with the appropriate HRP-conjugated secondary antibody at room temperature for 2 h. After another three washes in TBST, developing reagent (Biyuntian, China) was added to the membrane. The Western photo was taken by GeneGnomeXRQ (Syngene, UK).

Real-time quantitative PCT (RT–qPCR)

RT–qPCR experiments were performed in the HeLa and LO2 cell lines. RNA was extracted with RNAiso Plus (TAKARA) and converted to complementary DNA with an Eco M-MLV RT Kit with gDNA Clean for Qpcr II (Accurate Biotechnology (Hunan) Co. Ltd). Real-time quantitative PCR (qPCR) was conducted in triplicate using the LightCycler 480 system (Roche Diagnostics, IN, USA). The following primers were used: c-myc (forward primer: 5’-GCTGCTTAGACGCTGGATT-3’; reverse primer: 5’-TCCTCCTCGTCGCAGTAGA-3’) and GAPDH (forward primer: 5’-GGTGGTCTCCTCTGACTTCAACA-3’, reverse primer: 5’-GTTGCTGTAGCCAAATTCGTTGT-3’). The specificity of each PCR product was measured and controlled using the melting curve. The 2–ΔΔCt method was used to calculate the relative gene expression levels in which the amount of MYC mRNA was normalized to an endogenous reference (GAPDH), where Ct represents the threshold cycle.

RESULTS

The transition of MYC G4 from a parallel-stranded to hybrid-type structure driven by Pt-phen

CD is a common method for the structural characterization of G4 DNA in solution. The CD spectrum of free Pt-phen shows no signal (Supplementary Figure S1A). Myc1234 G4 forms a parallel-stranded structure under physiological 100 mM K+ with 10 mM Na+ conditions (Figure 2A), as indicated by the characteristic maximum at 260 nm and minimum at 240 nm in the CD spectra. Interestingly, induced by Pt-phen, a new maximum at 290 nm appeared, and the peaks at 240 nm and 260 nm still existed. These are characteristics of the hybrid-type G4 structure, illustrating the fast transition of Myc1234 G4 from the parallel-stranded conformation to the hybrid-type conformation. However, the CD spectra of duplex DNA or single-stranded DNA with Pt-Phen show no structural changes after binding to Pt-phen (Supplementary Figure S1B, C and Supplementary Table S1). Clearly, CD spectra cannot provide more accurate structural details for DNA G4 transitions. Therefore, we conducted NMR studies to confirm the structural change and determine the transition mechanism of Myc1234 G4.

Figure 2.

(A) The CD spectra of Myc1234 G4 (3 μM) titrated by Pt-phen at different Pt-phen/Myc1234 ratios in 100 mM K+, 10 mM Na+, pH 7.4 buffer, 30 min equilibration time. (B) Imino proton region of the 1D 1H NMR titration spectra of Myc1234 with Pt-phen in 100 mM K+, 10 mM Na+, pH 7.0, 25 °C. The DNA concentration is 150 μM. The ratios of Pt-phen/Myc1234 are shown. New peaks arising for Pt-phen–Myc1234 complexes are marked by asterisks with different colour.

1D 1H NMR titration experiments of Myc1234 G4 and Pt-phen were first performed (Figure 2B and Supplementary Figure S2A). Under physiological 100 mM K+ and 10 mM Na+ conditions, upon titration of Pt-phen, a new set of imino proton signals appeared (red asterisks) at a low 0.5:1 Pt-phen/Myc1234 ratio, suggesting the presence of a first well-defined Pt-phen–Myc1234 complex. After adding more Pt-phen, the spectra broadened at 1:1 and 2:1 ratios. At a 3:1 ratio, peaks for the second Pt-phen–Myc1234 complex were observed (green asterisks), coexisting with other conformations. Upon reaching the 4:1, 5:1 and 6:1 ratio (Supplementary Figure S2A), the major well-resolved peaks from the second complex were observed, illustrating the formation of a single dominating complex structure. As presented by the NMR titrations, Pt-phen shows multiple binding with Myc1234 G4 and forms two stable complex structures during binding. Pt-phen-bound MYC G4 was also visualized by PAGE under native conditions (Supplementary Figure S3). In the gel, the free Myc1234 G4 DNA formed a major monomeric structure in 100 mM K+ conditions and the Pt-phen binding keeps the MYC G4 structure in the monomeric state.

2D NMR proton assignment of the complexes revealing MYC G4 adopting hybrid-type folding topology converted by Pt-phen

As the 1D NMR spectra were suitable for NMR structural analysis, we decided to solve these two Pt-phen−Myc1234 complex structures to elucidate how Pt-phen binds and converts the Myc1234 conformation. We chose the spectra at 0.5:1 and 4:1 Pt-phen/Myc1234 ratios in physiological 100 mM K+ solutions for NMR analysis of the two complex structures. Complete sets of 2D NOESY, TOCSY and COSY NMR spectra of these complexes in both water and D2O at temperatures of 5, 15 and 25 °C were collected.

NMR spectral assignments of the Pt-phen−Myc1234 complex at a 0.5:1 ratio were made in the copresence of two molecular species, as the free and bound Myc1234 G4 coexisted at a 0.5:1 ratio. The imino protons of Myc1234 G4 at a 0.5:1 ratio were unambiguously assigned by 1H–15N HMQC experiments (Supplementary Figures S4 and S5). The bound Myc1234 G4 included two thymine imino peaks (Figure 2B and Supplementary Figure S5), which suggests that the complex formation involved a hydrogen-bonded structure with two thymines. Based on the imino proton assignments, we assigned other protons of bound Myc1234 G4 through NOE connections using NOESY, TOCSY and COSY (Figure 3A, B, Supplementary Figures S6 and S7 and Supplementary Tables S2 and S3). The protons of free and bound Pt-phen molecules were also unambiguously assigned by COSY experiments (Supplementary Figures S8 and S9). As shown in the COSY results, a weak H3/H1 cross-peak was shown in the free Pt-phen but was lost in the complex with G4 DNA, which was probably because of the decrease in the Pt-phen concentration from 10 mM to 0.4 mM. After obtaining the complete assignment of this complex at a 0.5:1 ratio, we found intermolecular NOEs between the protons of Pt-phen and the Myc1234 5′-end (Supplementary Table S3), which implies specific binding of Pt-phen at the 5′-end of Myc1234.

Figure 3.

The 2D NMR spectra and the solution structure of the 1:1 5′-end binding Pt-phen−hybrid-type MYC G4 complex (PDB code: 7DJU) during G4 transition. (A) The expanded H8/H6-H1′ region of the NOESY spectra at 0.5:1 Pt-phen/Myc1234 ratio, 100 mM K+, 10 mM Na+, pH 7.0, 25 °C. The H8/H6-H1′ NOEs of the nucleotides in syn conformation are labelled in italic and bold. (B) The expanded H8/H6/H2-H1 region of the NOESY spectrum of the same ion condition at 5 °C. The DNA concentration is 800 μM. The G-tetrad GxH1-GyH8 cross-peaks of free Myc1234 are coloured in black, and that of the 1:1 5′-end binding complex are coloured in red. Inter-residue cross-peaks of the G-tetrad and T10:T19:A11 triad are framed in red box. Intermolecular cross-peaks of the Pt-phen and 5′-G-tetrad are coloured in purple. (C) A representative structure of the 1:1 5′-end binding complex. (D, E) Two different views of the Pt-phen-induced binding pockets at the 5′-end. (F, G) Two different views of the hydrogen-bonded T10:T19:A11 triad at the 3′-end, with residues labelled and coloured in cyan. In the structures, guanines, adenines and thymines are coloured in pink, green and blue, respectively.

Importantly, the bound Myc1234 G4 in this complex was determined to be in a hybrid-type mixed parallel/antiparallel-stranded G4 conformation, consistent with the CD results (Figure 2A). With the NMR assignment, five guanine residues of the core G-tetrads of bound Myc1234, including G3, G7, G12, G13 and G16, were observed to adopt a syn glycosidic conformation (Figure 3A and Supplementary Figure S7), unlike the free Myc1234 G4 in which all residues adopt an anti glycosidic conformation. The anti-to-syn conversion illustrates important conformational rearrangements that occurred in Myc1234 G4 induced by Pt-phen binding. The assignment of the guanine imino H1 and aromatic H8 protons leads to the direct determination of the folding topology of Myc1234 G4 in the complex with Pt-phen. According to the NOE connections in the assignments, the first, second and fourth G-strands of this hybrid-type G4 are parallel with each other and antiparallel with the third G-strand (Figure 1). The first two G-strands (from the 5′-end) are linked with an A6 double-chain-reversal side loop, the second and third strands are linked with the T10-A11 edgewise loop, and the third and fourth G-strands are linked with the A15 edgewise loop. In addition, the first G-tetrad (G3-G7-G14-G16) has a (syn:syn:anti:syn) G-arrangement, while the middle (G4-G8-G13-G17) and bottom G-tetrads (G5-G9-G12-G18) have the same reversed (anti:anti:syn:anti) G-arrangement.

Solution structure of the 1:1 5′-end binding Pt-phen–hybrid-type MYC G4 complex in 100 mM K+ showing a significant hydrogen-bonded T:T:A triad formed at the 3′-end, which facilitates the G4 transition

We calculated the solution structure of this 1:1 Pt-phen–hybrid-type Myc1234 complex at a 0.5:1 ratio in physiological 100 mM K+ conditions based on the NOE data (Figure 3C and Supplementary Table S4) and selected 10 lowest-energy refined complex structures (Supplementary Figure S10). As shown in the 1:1 5′-end binding complex structure, due to Pt-phen binding, the structure of MYC G4 is disrupted and converted from the parallel-stranded conformation into the hybrid-type conformation in which the Pt-phen binds at the 5′-end of hybrid-type MYC G4 to form a well-defined binding pocket. The Pt-phen stacks over four bases, G3, G7, G14 and G16, of the 5′-G-tetrad for a highly extensive π–π interaction (Figure 3D). One phenanthroline ring of Pt-phen covers the bases of G3 and G7, and another covers the bases of G14 and G16. The flanking and loop residues of MYC G4 were also rearranged due to conformational transition and Pt-phen binding. At the 5′-end binding pocket, the A2 at the 5′-flanking stacks on top of one phenanthroline ring of Pt-phen, and the T1 folds back and covers another phenanthroline ring to stabilize the complex structure, while the A15 at the 5′-edgewise loop partly stacks on the Pt-phen molecule (Figure 3E).

Interestingly, a significant hydrogen-bonded base pair structure is observed at the 3′-end of Myc1234 G4 during conformational transition. Based on the interresidue NOEs (Figure 3B), we found a stable T10:T19:A11 hydrogen-bonded triad formation (Figure 3F, G). The T10:T19:A11 triad stacked very well on top of the 3′-G-tetrad, with T10 covering G5 and G9, A11 covering G12, and T19 covering G18. Formation of the T10:T19:A11 triad by hydrogen bonding makes the 3′-flanking and loop residues almost fix on the 3′-end and further stabilizes the whole hybrid structure of MYC G4, which probably facilitates the G4 conformational transition. As the T:T:A triad is forming and the 3′-end is not easy to bind, the Pt-phen interacted with the hybrid MYC G4 at the 5′-end (Figure 3C). Therefore, under physiological 100 mM K+ conditions, Pt-phen drives parallel-stranded MYC G4 switching to a hybrid-type structure with a T:T:A triad formation by hydrogen bonding at the 3′-end, and Pt-phen first binds at the 5′-end of hybrid MYC G4.

In order to investigate the importance of the triad in G4 transition, we conducted DNA sequence mutation studies. First, concerning the cellular long MYC sequence, we added 1 or 2 extra bases at the 5′ and 3′ ends of Myc1234 according to the MYC gene sequence. The results of CD spectra revealed that Pt-phen can still convert the longer Myc1234 mutants from the parallel-stranded conformation to the hybrid-type conformation (Supplementary Figure S11A–D). To test if the T10:T19:A11 triad is forming in the longer Myc1234 mutants during structural transition, we performed 1D NMR titration experiments (Supplementary Figure S12). As shown by the spectra, upon Pt-phen binding, two characteristic imino signals of T10 and T19 (red asterisks) for the T10:T19:A11 triad were still present, indicating that the T:T:A triad is still formed. Moreover, the NMR titration spectra pattern of these Myc1234 mutants were observed to be similar with that of the original Myc1234 (Figure 2B), implying the Pt-phen has similar binding modes. Thus, the Pt-phen probably still binds preferably at the 5′ end of the longer Myc1234 mutants with T:T:A triad forming.

Next, we mutated the T for A of the T10:T19:A11 triad of Myc1234 and performed CD and 1D NMR titration experiments. As shown by the CD results, after mutating one of the T bases for A (T10 to A10, or T19 to A19), the structural conversion was still observed upon Pt-phen binding (Supplementary Figure S11E, F). In the NMR titration spectra of these two Myc1234 mutants (Supplementary Figure S13A, B), one characteristic imino signal (red asterisks) of T was present, which might be due to the formation of T:A:A triad. Compared with the NMR spectra of the original Myc1234 (Figure 2B), the titration spectra pattern were similar, which further revealed that Pt-phen binds these two Myc1234 mutants with similar modes and the triad is also probably forming. However, in the case of a sequence unable to form a triad (mutated both of the T10 and T19 for A), the structural conversion was not able to observe upon Pt-phen binding (Supplementary Figure S11G). Moreover, the NMR titration spectra of this Myc1234 mutant were totally different from the original Myc1234. The imino proton signals were gradually broadened and poorly resolved after adding Pt-phen and no well-defined complex structure was formed (Supplementary Figure S13C), indicating less-defined binding of the Pt-phen to this Myc1234 mutant and no structural conversion. Thus, after mutating the T10 and T19 for A, the triad (A:A:A) is not able to form and the G4 structural conversion cannot happen. These results demonstrated that the formation of 3′-end hydrogen-bonded triad is significant for the advancing of G4 transition.

Solution structure of the 2:1 Pt-phen–hybrid-type MYC G4 complex

Next, we unambiguously assigned all protons of the NMR spectra at a 4:1 Pt-phen/Myc1234 ratio in physiological 100 mM K+, 10 mM Na+ solution (Figure 4A, B, Supplementary Figures S14-S17 and Supplementary Tables S5 and S6). The complex was determined to be a 2:1 Pt-phen–hybrid-type Myc1234 G4 complex. The Myc1234 G4 in the 2:1 complex adopted the same hybrid-type folding as the above 1:1 5′-end binding complex. Intermolecular NOEs between the protons of Pt-phen with the 5′-end and 3′-end of Myc1234 were observed, revealing the second Pt-phen molecule interacting with the 3′-end of Myc1234.

Figure 4.

The 2D NMR spectra and the solution structure of the 2:1 Pt-phen−hybrid-type MYC G4 complex (PDB Code: 7DJV). (A) The expanded H8/H6-H1′ region of the NOESY spectra of 2:1 complex at 4:1 Pt-phen/Myc1234 ratio, 100 mM K+, 10 mM Na+, pH 7.0, 25 °C. The H8/H6-H1′ NOEs of the nucleotides in syn conformation are labelled in italic and bold. (B) The expanded H8/H6/H2-H1 region of the NOESY spectrum of 2:1 complex at 5 °C. The DNA concentration is 800 μM. The G-tetrad GxH1-GyH8 cross-peaks of the 2:1 complex are shown. Intermolecular cross-peaks of the Pt-phen and G-tetrad are coloured in purple. (C) A representative structure of the 2:1 complex. (D, E) Two different views of the Pt-phen-induced binding pockets at the 5′-end. (F, G) Two different views of the Pt-phen-induced binding pockets at the 3′-end. In the structures, guanines, adenines and thymines are coloured in pink, green and blue, respectively.

We calculated the 2:1 complex structure based on the NOE data (Supplementary Table S7) and presented the ensemble of 10 lowest-energy structures (Supplementary Figure S18). As shown by the representative structure (Figure 4C), two Pt-phen molecules bind the hybrid-type MYC G4, with one Pt-phen stacked on the 5′-G-tetrad and the other stacked on the 3′-G-tetrad. The 5′-end binding pocket is the same as that in the 1:1 5′-end binding complex above (Figure 4D, E). At the 3′-end of MYC G4, the T10:T19:A11 hydrogen-bonded triad in the 1:1 5′-end binding complex is disrupted upon the second Pt-phen binding. The Pt-phen molecule is well defined and matches the 3′-G-tetrad with the Pt atom aligning with the center of the G-tetrad (Figure 4F). One phenanthroline ring of Pt-phen covers the base of G5 and G18, and another covers the bases of G9 and G12. T10 at the 3′-edgewise loop and T19 at the 3′-flanking further stack on top of two phenanthroline rings to stabilize the complex (Figure 4G). The A11 at the 3′-edgewise loop partly stacks on the Pt-phen.

Therefore, in physiological 100 mM K+ solution, the Pt-phen switches the MYC G4 from parallel-stranded into hybrid-type structure upon binding. Pt-phen first binds at the 5′-end of the hybrid MYC G4 with higher affinity to form a 1:1 complex as the hydrogen-bonded T:T:A triad at the 3′-end is stably formed and hard to disrupt. Further adding more Pt-phen, the Pt-phen molecule disrupts the T:T:A triad and binds the 3′-end to form the 2:1 complex.

NMR studies showed different transition complexes formed under lower K+ conditions

We further tested the interaction of Myc1234 G4 with Pt-phen in lower K+ conditions. In the CD spectra (Figure 5A, B), similar changes were observed as those in 100 mM K+, revealing that Myc1234 G4 is converted from a parallel-stranded to a hybrid-type structure in both 50 mM and 5 mM K+. The native PAGE data showed that Pt-phen binding keeps the MYC G4 structure in the monomeric state (Supplementary Figure S3). Then, we conducted NMR studies (Figure 5C, D and Supplementary Figure S2B, C). Excitingly, we found different transition complexes forming.

Figure 5.

(A, B) The CD spectra of Myc1234 G4 (3 μM) titrated by Pt-phen at different Pt-phen/Myc1234 ratios in 50 or 5 mM K+, 10 mM Na+, pH 7.4, 30 min equilibration time. (C, D) Imino proton region of the 1D 1H NMR titration spectra of Myc1234 with Pt-phen in 50 or 5 mM K+, 10 mM Na+, pH 7.0, 25 °C. The DNA concentration is 150 μM. The ratios of Pt-phen/Myc1234 are shown. New peaks arising for Pt-phen–Myc1234 complexes are marked by asterisks with different colour.

In the 50 mM K+ and 10 mM Na+ solution (Figure 5C), after adding Pt-phen, two sets of new peaks emerged at a 0.5:1 ratio (red and orange asterisks), implying the formation of two complexes at low Pt-phen equivalents. One set of the complex spectrum (red asterisks) is shown to be the same as the 1:1 5′-end binding complex in 100 mM K+ (Supplementary Figure S19), while the other set of complex spectra is very different. With more Pt-phen, the spectra were broadened at 1:1 and 2:1 ratios. At 3:1 and 4:1 ratios, only peaks of the same final 2:1 complex emerged (green asterisks).

In contrast, in the 5 mM K+ and 10 mM Na+ solution (Figure 5D), the first complex (blue asterisks) was observed at a 0.5:1 ratio upon Pt-phen titration, and its spectrum was completely distinct from the complexes formed at a 0.5:1 ratio in 100 mM (Figure 2B) and 50 mM K+ (Figure 5C). At a 1:1 ratio, the imino peaks of the same 2:1 complex emerged (green asterisks). Upon further addition of more Pt-phen, only peaks for the final 2:1 complex were observed. As shown by the NMR results, multiple Pt-phen−MYC G4 complexes were formed during the Pt-phen-modified G4 transition, which fully demonstrates various G4 transition pathways.

Solution structure of the 1:1 3′-end binding Pt-phen–hybrid-type MYC G4 complex formed in 5 mM K+

We decided to elucidate the NMR structure of the Pt-phen–Myc1234 complex at a 0.5:1 Pt-phen/Myc1234 ratio in 5 mM K+ and 10 mM Na+. Using the NOE connections and sequential assignment method, we were able to assign all Myc1234 protons of the complex (Figure 6A, B, Supplementary Figures S20–S23 and Supplementary Table S8). The complex adopted the same hybrid-type folding as those above two complexes.

Figure 6.

The 2D NMR spectra and the solution structure of the 1:1 3′-end binding Pt-phen−hybrid-type MYC G4 complex (PDB Code: 7DJW) during G4 transition. (A) The expanded H8/H6-H1′ region of the NOESY spectrum at 0.5:1 Pt-phen/Myc1234 ratio in 5 mM K+, 10 mM Na+, pH 7.0, 25 °C. The H8/H6-H1′ NOEs of the nucleotides in syn conformation are labelled in italic and bold. (B) The expanded H8/H6/H2-H1 region of the NOESY spectrum at the same condition. The DNA concentration is 800 μM. The G-tetrad GxH1-GyH8 cross-peaks of free Myc1234 are coloured in black, and that of the 1:1 3′-end binding complex are coloured in blue. Intermolecular cross-peaks of the Pt-phen and 3′-G-tetrad are coloured in purple. (C) A representative structure of the 1:1 3′-end binding complex. (D, E) Two different views of the 5′-end of the complex structure. (F, G) Two different views of the Pt-phen-induced binding pockets at the 3′-end. In the structures, guanines, adenines and thymines are coloured in pink, green and blue, respectively.

Intermolecular NOEs between the protons of Pt-phen and Myc1234 3′-end were observed, revealing a first specific binding of Pt-phen at the 3′-end of hybrid-type Myc1234 G4, rather than forming a T:T:A triad at the 3′-end as that in 100 mM K+. This suggests that the formation of the T:T:A triad depends on K+ (42); that is, a higher K+ concentration can stabilize the T:T:A triad. We calculated the solution structure of the 1:1 3′-end binding complex using NMR restraints (Figure 6C-G and Supplementary Table S9). The 10 lowest-energy refined complex structures were selected and are presented in Supplementary Figure S24. The 3′-end binding pocket of this complex is the same as that of the 2:1 binding complex above (Figure 6F, G). At the 5′-end, A2 and A15 rearrange and stack over the 5′-G-tetrad, and T1 is more flexible and exposed to the solvent (Figure 6D, E).

We further performed the NMR spectra showing the exchanges of different complex structure species through K+ titration at 0.5:1 Pt-phen/Myc1234 ratio. As shown by the spectra (Supplementary Figure S25), in 5 mM K+, the 1:1 3′-end binding Pt-phen−MYC G4 complex (blue asterisks) appeared and is in slow exchange with the free Myc1234. After adding the K+ concentration (20 mM K+), 1:1 5′-end binding Pt-phen–MYC G4 complex (red asterisks) is gradually appeared and shows slow exchange with the 3′-end complex (blue asterisks) and free Myc1234. Continuously adding K+ to 40–60 mM, the 3′-end complex (blue asterisks) is disappeared, and another complex III (orange asterisks, not elucidated here) is observed and is in slow exchange with the 5′-end complex (red asterisks) and free Myc1234. In 80–100 mM K+ conditions, the complex III (orange asterisks) is disappeared, and only the 5′-end complex (red asterisks) existed and is in slow exchange with free Myc1234. The results revealed that in different K+ concentration, different complexes are forming and in slow exchange with free Myc1234.

Various pathways of structural transition of MYC G4 bound to Pt-phen

Solving the NMR structures of Myc1234 G4 interacting with Pt-phen allows us to clarify the conformational transition of G4 from parallel-stranded to hybrid-type folding and elucidate the dynamic binding mechanism of the Pt-phen with Myc1234 G4. As shown by the NMR data, there are significant differences in the structural conversion and binding pathways of Myc1234 bound to Pt-phen under different K+ concentrations (Figure 7).

Figure 7.

The structural transition of MYC G4 driven by Pt-phen modification under different ion conditions. In the folding topology of Myc1234 G4, deep pink box = (syn) guanine, light pink box = (anti) guanine, green ball = adenine, blue ball = thymine.

At the physiological 100 mM K+ concentrations, induced by Pt-phen, a stable hydrogen-bonded T:T:A triad was newly induced and covered the 3′-G-tetrad of hybrid-type Myc1234. As a result, Pt-phen first bound the 5′-end of Myc1234 to form the 1:1 5′-end binding Pt-phen–hybrid-type Myc1234 complex at low Pt-phen/Myc1234 ratios (Figure 7 step (i)) and then disrupted the 3′ T:T:A triad and binds the 3′-end at higher ratios to form the 2:1 complex (Figure 7 step (II)). At a moderate 50 mM K+ concentration, the binding of Pt-phen formed a 1:1 5′-end binding complex and another complex (not elucidated here) at low ratios and continued to form the same 2:1 complex at high ratios. However, at a 5 mM K+ concentration, the structures indicate that Pt-phen first bound at the 3′-end of the hybrid-type Myc1234 to form the 1:1 3′-end binding complex at low ratios (Figure 7 step (ii)) and then bound at the 5′-end at a higher equivalence of Pt-phen to form the same final 2:1 complex (Figure 7, step (II)), indicating various transition pathways.

Pt-phen targets G4 in living cells and represses MYC gene expression in cancer cells

MYC transcription inhibition can be achieved by targeting the MYC promoter DNA G4 structures (13). To explore the cellular effects of Pt-phen on MYC gene expression, we performed several experiments. The fluorescence lifetime imaging microscopy (FLIM) experiment was performed in the HeLa cell line to understand the cellular targeting of Pt-phen (Figure 8A). We used our recently reported fluorescence lifetime probe, NBTE (37), which can detect the G4 binding ligand in living cells through FLIM competitive experiments, as proven by the results of canonical G4 binder PDS (Supplementary Figure S26). In the FLIM, the ratio value of DNA peak area to total area counting from the lifetime distribution reflects the relative percentage of the photon counts contributed by different DNA. As we can see from FLIM competitive results, the ratio of G4 peak area decreased from 41 ± 2% to 30 ± 2% after we added Pt-phen, implying that Pt-phen can partially replace NBTE and bind G4 DNAs. Therefore, we demonstrated that Pt-phen can target G4 structures in living cells.

Figure 8.

Pt-phen represses MYC expressions. (A) FLIM experiments of Hela cells with G4-probe (NBTE, 20 μM) and Pt-phen (20 μM) for 24 h. The decrease of the ratio of G4 peak area proved that Pt-phen enters the nucleus and targets G4. (B) The mRNA transcription levels of MYC were inhibited after Pt-phen treatment at 5 and 10 μM as shown by RT–qPCR in the HeLa and LO2 cell lines. (C) The MYC protein levels were reduced after Pt-phen treatment at 5 and 10 μM as shown by western blotting in the HeLa and LO2 cell lines. GAPDH was used as internal control.

Next, we conducted MTT, RT–qPCR and Western blotting assays to investigate G4-targeted MYC gene repression by Pt-phen. The MTT experiments showed that Pt-phen has moderate cytotoxicity against various cancer and normal cell lines (Supplementary Table S10). Then, we performed RT–qPCR and Western blotting assays in the HeLa cancer cell line. We chose two Pt-phen concentrations, 5 and 10 μM, which are lower than the IC50 value of Pt-phen in HeLa cells (17.33 ± 1.04 μM). We found that Pt-phen could effectively inhibit MYC expression. The MYC mRNA levels were significantly reduced below 40% after 48 h Pt-phen incubation at both 5 and 10 μM in a time-dependent manner, as shown by the RT–qPCR results (Figure 8B). Consistently, MYC protein expression levels were decreased significantly after 48 h of Pt-phen treatment at 5 μM and 36 h at 10 μM (Figure 8C). Considerably lower MYC gene expression was observed with increased Pt-phen treatment time. These results demonstrated that Pt-phen functions as an effective MYC repressor in cancer cells, likely by binding and stabilizing the MYC promoter G4 structures to inhibit transcription. In the LO2 normal cell line (Figure 8B, C), inhibition of MYC expression was also observed and shown to be weaker than that in the HeLa cancer cell line, which might be because the amount of G4 structures in normal cells is less than in cancer cells (37). The results indicated that Pt-phen presents a little bit cancer selectivity.

DISCUSSION

Common CD spectroscopy is insufficient for describing the structural features and transitions of nucleic acid secondary structures. As shown in the CD spectra, at a low 0.5:1 Pt-phen/Myc1234 ratio of all ion conditions, the CD signal changes only slightly and looks nearly the same as free Myc1234 G4. However, as elucidated by the NMR method, Pt-phen rapidly changes Myc1234 G4 from parallel to hybrid folding at a low Pt-phen ratio. Moreover, the CD spectra could not offer accurate structural information.

In-depth knowledge of the NMR complex structures of Pt-phen and Myc1234 G4 provides accurate binding details for explaining the G4 structural change. Myc1234 G4 in parallel-stranded folding has short flanking segments and stretched-out side loop residues, whose ligand binding pockets can only form by the short 5′ or 3′ flanking segments. For example, DAOTA-M2 binds the 5′- and 3′-G-tetrads of Myc1234 and is covered by only one residue (33). In contrast, hybrid-type G4s contain edgewise loops that can contribute to more extensive binding pocket formation (34,35). As shown by our complex structures (Figures 3-4, 6), Myc1234 converts its conformation from parallel-stranded to hybrid-type to achieve stronger binding by Pt-phen with flanking and edgewise loop residues. Pt-phen possesses a large planar surface for extensive π–π stacking. The structural changes allow Myc1234 G4 to rearrange both flanking residues and loop residues to form stable multilayer π–π stacking structures. These results suggest that (a) the formation of a stable ‘edgewise loop/flanking-ligand-G-tetrad’ sandwich stacking structure and (b) the large surface of the ligand for enhanced extensive π–π stacking interactions likely motivates the conformational transition of G4. In particular, (c) the stabilization of the external hydrogen-bonded T:T:A non-G-triad also plays an important role in the MYC G4 transition.

The results demonstrate that Pt-phen can convert the structural change of Myc G4 at both high and low K+ concentrations and form identical end complex structures. The parallel Myc1234 G4 was converted to the same hybrid-type conformation by Pt-phen, which has two external G-tetrad binding sites. Pt-phen is a planar compound, and its size is suitable for external G-tetrad binding. As a result, Pt-phen successively binds the two external G-tetrads (5′ and 3′) in the hybrid-type MYC G4 structure. Our results have implied that the 3′-end hydrogen bonding triad is of special importance in the altering of transition pathway (43). More importantly, the binding affinity of G4 3′-end is negatively correlated with the stability of 3′ T:T:A triad (Figure 7). Under higher 100 mM K+ conditions, the T10:T19:A11 triad is stably formed at the 3′-end of hybrid-type G4 induced by Pt-phen. Because of the hindrance of the triad, the 3′-end is less available for binding than the 5′-end. Therefore, Pt-phen first binds at the 5′-end and then binds at the 3′-end of G4 and finally forms a 2:1 Pt-phen–Myc1234 complex. However, in lower K+ concentration, the 3′-end has higher affinity for Pt-phen binding when 3′ T:T:A triad is not formed. Under the 5 mM K+ condition, the triad is not formed and both of 5′ and 3′-ends are open to ligands. Then, Pt-phen first binds at the 3′-end and then binds at the 5′-end, finally forming the 2:1 Pt-phen−Myc1234 complex. Thus, the end form of the 2:1 Pt-phen–Myc1234 G4 complex is identical in the high and low K+ concentrations.

Structural differences between the parallel and hybrid-type folding adopted by Myc1234 G4 inspired us to propose possible structural intermediates in the transition induced by Pt-phen modification. Previous studies suggest that folding of the intramolecular G4s may progress through various intermediate structures, such as hairpins and G-triplexes (44–46). Compared with the G-tetrads of the parallel-stranded and hybrid-type Myc1234, we found that the positions of the G12 and G14 residues were exchanged, while the positions of other guanines remained unchanged. Therefore, we hypothesized that Myc1234 successfully converts its structure by the stabilization of the external G-triad by Pt-phen and the rotation of the G12 and G14 residues around the G13 residue (Supplementary Figure S27). As induced by Pt-phen, G12 and G14 first dissociate from the G-tetrads and rotate around G13, and then some guanines change from anti- to syn-glycosidic conformations. At the same time, the 5′-G-triad and 3′-G-triad are stabilized by Pt-phen or the capping residues, as shown by the intermediates in the illustration.

Previous studies showed that the MYC NHE III1 sequence, which controls 80–95% of MYC transcription, coexists with different parallel G4 conformations (13) and can be converted to hybrid structures by ligands in Na+ or in the absence of ion conditions (16,26). The X-ray complex structure of DHX36 helicase and a parallel-stranded MYC G4 together with other studies revealed that DHX36 helicase specifically recognizes parallel-stranded G4s and unwinds them (19). As a result, switchable control of G4 structures could offer exciting opportunities to alter G4 recognition by helicases and further regulate gene expression. In this study (Figure 8), we demonstrated through FLIM experiments that Pt-phen can target G4 structures in living cells. Furthermore, we found that the MYC mRNA levels and protein expression levels were significantly reduced after Pt-phen incubation, as shown by the RT-qPCR and western blotting results. Studies have shown that the G4 content in cancer cells is 4-fold that in normal cells (37). Here, the Pt-phen presents a little bit cancer selectivity. The inhibition of MYC expression in the cancer cell line was shown to be stronger than that in the normal cell line, which might be because the more of G4 structures in the cancer cells. These results demonstrated that Pt-phen is an effective MYC repressor, likely by binding and stabilizing the MYC promoter G4 structures to inhibit the transcription.

CONCLUSION

In summary, we found that the platinum(II) compound Pt-phen can drive the structural transition of Myc1234 G4 from parallel-stranded to hybrid-type folding, which illustrated G4 structure polymorphism. After adding a low ratio of Pt-phen, Myc1234 can convert to a hybrid structure. Various transition pathways are shown under different K+ concentrations. The structure determination of the NMR complex revealed that the G4 structural transition is mainly through (a) the formation of an ‘edgewise loop/flanking-ligand-G-tetrad’ sandwich structure, (b) the matched surface of Pt-phen with G-tetrad for extensive π–π stacking interaction mode and thus high-affinity binding and (c) the hydrogen-bonded terminal T:T:A non-G-triad stabilization. This work is the first example reported in a complete pathway map and molecular structures of the G4 conformational transition driven by a small molecule.

DATA AVAILABILITY

The coordinates for structures of the 1:1 5′-end binding Pt-phen–Myc1234 G4 complex (PDB code: 7DJU), 1:1 3′-end binding Pt-phen–Myc1234 G4 complex (PDB code: 7DJW) and 2:1 Pt-phen–Myc1234 G4 complex (PDB code: 7DJV) have been deposited in the Protein Data Bank.

Click here for additional data file.