Fluorescence Turn-Off Ligand for Parallel G-Quadruplexes

Parallel-stranded G-quadruplex structures are found to be common in the human promoter sequences. We tested highly fluorescent 9-methoxyluminarine ligand (9-MeLM) binding interactions with different parallel G-quadruplexes DNA by spectroscopic methods such as fluorescence and circular dichroism (CD) titration as well as UV melting profiles. The results showed that the studied 9-MeLM ligand interacted with the intramolecular parallel G-quadruplexes (G4s) with similar affinity. The binding constants of 9-methoxyluminarine with different parallel G4s were determined. The studies upon oligonucleotides with different flanking sequences on c-MYC G-quadruplex suggest that 9-methoxyluminarine may preferentially interact with 3′end of the c-MYC promoter. The high decrease in 9-MeLM ligand fluorescence upon binding to all tested G4s indicates that 9-methoxyluminarine molecule can be used as a selective fluorescence turn-off probe for parallel G-quadruplexes.


Introduction
Guanine quadruplexes (G-quadruplex DNA, G4 DNA) are forms of DNA, important in the research on cancer and telomerase, as well as in the research on transcriptional changes by the formation of G-quadruplexes in the promoter region of genes [1,2].These non-canonical single-or multistranded nucleic acid structures are formed by co-planar arrangements of four guanines (G-tetrads), stabilized by Hoogsteen-type hydrogen bonds in the presence of selected metal cations (Na + and K + ) and can be stabilized by small organic ligands.Until now, many organic compounds that induce, stabilize or disrupt the structure of G4 have been obtained and are collected in the updated G4 and i-motifs ligand database, G4LDB 2.2 [3].Ligands with negligible affinity for DNA duplexes and ligands with specificity for some G4s over others are still being sought [4][5][6].However, some selective ligands have been recently identified such as a thiazole peptide [7], pyridinium salt [8], a coumarin-quinazolinone [9], a quinazoline-quinazolinone [10], a coreextended naphthalene diimide [11], squaraine dyes [12,13], and anthracene-based [14], BODIPY-based [15], triarylimidazole-based [16], cyanine dye-based [17] and bis(quinolinium) pyridodicarboxamide-based [18] compounds.
Fluorescent ligands interacting with G4 can be classified as follows: (1) "light-up" probes that display a strong enhancement upon G4 binding, (2) "light-off" probes that display a decreased fluorescence upon G4 binding and (3) permanent probes ("tagged" G4-binders) that exhibit no changes in fluorescence signals but possess high quadruplex binding specificity [19].For instance, "light-off" probes are in general less suitable for imaging in vivo, whereas they may be very useful for monitoring interactions and evaluating binding affinity and selectivity for G-quadruplexes.
Previously, we have shown that the fluorescence of the 9-methoxyluminarine (9-MeLM) ligand is selectively and effectively quenched by the c-MYC G-quadruplex over nonparallel G4s.We proved that observed quenching occurs in a static mode, most probably due to a π-π interaction between 9-MeLM and the exposed parallel G4 guanine tetrad (Figure 1) [21].In the case of parallel G-quadruplexes, it was shown that the terminalstacking binding mode of selective ligands is the main kind of interaction for them, which was supported by molecular docking and mechanistic studies [22].Lately, Deiana et al. developed a fluorescent probe with a minimalistic amidinocoumarin-based molecular scaffold that selectively recognizes parallel G4 structures, which is manifested by turn-on emission [23].In their case, titration of the probe with c-MYC sG4 (16-mer being only c-MYC G4 core) induced 6-fold higher enhancement in fluorescence signals than for complexation with 22-mer c-MYC (possessing TGA at the 5 ′ end and TAA at the 3 ′ end).Hence, they indicated that its light-up ability was higher and correlated with better π-stacking possibilities for the planar conformation of the coumarin core [23].On the other hand, the studies by Gai et al. indicated that the flanking sequence might form a binding cavity above the terminal G-quartet for ligands and interact with ligands through π-π stacking or, in the opposite manner, exhibit steric hindrance effects to hamper ligand/G4 binding [24].Their results revealed that the flanking sequences provide a positive facilitating effect when the flanking chain is extended to two to five residues, whereas a negative hindrance effect to ligand binding was caused by one or six residues (the overlong flanking sequence) [24].solubility (is positively charged), a high fluorescence quantum yield (0.99), a lifetime of 10 ns and photostability [20].
Previously, we have shown that the fluorescence of the 9-methoxylu MeLM) ligand is selectively and effectively quenched by the c-MYC G-qua nonparallel G4s.We proved that observed quenching occurs in a static mod ably due to a π-π interaction between 9-MeLM and the exposed parallel G4 g (Figure 1) [21].In the case of parallel G-quadruplexes, it was shown that stacking binding mode of selective ligands is the main kind of interaction for was supported by molecular docking and mechanistic studies [22].Lately, developed a fluorescent probe with a minimalistic amidinocoumarin-bas scaffold that selectively recognizes parallel G4 structures, which is manifeste emission [23].In their case, titration of the probe with c-MYC sG4 (16-mer MYC G4 core) induced 6-fold higher enhancement in fluorescence signals t plexation with 22-mer c-MYC (possessing TGA at the 5′ end and TAA at the 3 they indicated that its light-up ability was higher and correlated with bett possibilities for the planar conformation of the coumarin core [23].On the ot studies by Gai et al. indicated that the flanking sequence might form a b above the terminal G-quartet for ligands and interact with ligands through or, in the opposite manner, exhibit steric hindrance effects to hamper ligan [24].Their results revealed that the flanking sequences provide a positive fac when the flanking chain is extended to two to five residues, whereas a negat effect to ligand binding was caused by one or six residues (the overlong flank [24].Since our previous work suggested selectivity of 9-MeLM for parallel G parallel ones, we decided to check the binding affinity of highly fluorescen other parallel G4s.For this purpose, we have examined interaction of the 9-m narine ligand with G-quadruplexes formed by sequences that correspond proto-oncogenes c-KIT and RET, as well as hemin aptamer catG4, well kn parallel topology.The c-KIT proto-oncogene encodes a receptor tyrosine kin volved in the regulation of a number of physiological processes such as prol ferentiation, migration, maturation and cell survival and the overexpression of the c-KIT proto-oncogene that is identified in many diseases [25][26][27].As a plate, we chose a well-characterized 22-mer G-rich sequence derived from forms unique parallel-stranded intramolecular G-quadruplex (PDB ID 2O other proto-oncogene that also encodes a receptor tyrosine kinase is the pr Since our previous work suggested selectivity of 9-MeLM for parallel G4s over nonparallel ones, we decided to check the binding affinity of highly fluorescent 9-MeLM to other parallel G4s.For this purpose, we have examined interaction of the 9-methoxyluminarine ligand with G-quadruplexes formed by sequences that correspond to the human protooncogenes c-KIT and RET, as well as hemin aptamer catG4, well known to form parallel topology.The c-KIT proto-oncogene encodes a receptor tyrosine kinase that is involved in the regulation of a number of physiological processes such as proliferation, differentiation, migration, maturation and cell survival and the overexpression or mutation of the c-KIT proto-oncogene that is identified in many diseases [25][26][27].As a research template, we chose a well-characterized 22-mer G-rich sequence derived from c-KIT1 that forms unique parallel-stranded intramolecular G-quadruplex (PDB ID 2O3M) [28].Another proto-oncogene that also encodes a receptor tyrosine kinase is the proto-oncogene RET, which is expressed in tissues and tumors, and its activation contributes to the development of human cancers [29,30].We chose a 20-mer sequence derived from RET as the molecular target in our studies (PDB ID 2L88) [31].The hemin-binding DNA aptamer catG4 is commonly used to obtain DNAzymes [32,33].In our experiments, intramolecular parallel G4-forming sequence of catG4 was used as model target [32].
Therefore, we present the stability and interaction of 9-methoxyluminarine ligand (9-MeLM) with various parallel G-quadruplexes by spectroscopic methods such as fluorescence and circular dichroism (CD) titration and UV melting profiles.Moreover, a few sequences without 3 ′ or 5 ′ flanking sequence extension based on c-MYC were designed to gain insight into the (c-MYC G-quadruplex)/9-MeLM ligand interaction.We demonstrate the possibility of using the 9-methoxyluminarine molecule and the common fluorescence spectroscopy technique as a preliminary screening allowing the recognition of parallel G-quadruplex topology.

Circular Dichroism Spectroscopy
Circular dichroism spectroscopy (CD) has become a very useful technique in the study of G-quadruplex structure and of G-quadruplex/ligand interactions.By analyzing CD spectra, we can study the influence of various factors such as, e.g., the pH, temperature, type and concentration of ions or the presence of ligands on conformational changes in DNA.In the short-wavelength region where DNA possess absorption bands, the CD bands appear, whose position and magnitude are commonly assigned to specific G-quadruplex conformations.A typical parallel G-quadruplex structure exhibit a positive CD signal at about 265 nm and a negative CD signal at 240 nm in the presence of 100 mM KCl [34].Thus, CD spectroscopy can be used to effectively distinguish between parallel and antiparallel Gquadruplex structures [35].Circular dichroism spectra were recorded to assess the effect of the 9-methoxyluminarine ligand on the conformation of all tested parallel G4s, to monitor stability and to verify the binding mode in the ligand/G4 complex.Upon addition of the ligand, the intensity of peaks changed slightly (between 15% and 25% at DNA/ligand molar ratio, 1:5), but their position was stable, suggesting that the general topology remains intact (Figure 2).The corresponding UV-vis spectra are shown in Supplementary Materials (Figure S1).The induced signals (ICD) did not occur in the long-wavelength range, where absorption band with maximum at 390 nm of achiral 9-methoxyluminarine is present.Based on this observation, we can exclude the groove binding to G-quadruplex structures, because a positive ICD, signal as well as the induced exciton CD signal have been taken as an indicator of groove binding to G-quadruplex structures [36].These results indicated that the G-quadruplexes formed by c-KIT1, RET and catG4 bind tested ligand similarly to c-MYC G4 [21].The slight decrease in the intensity of the CD at 265 nm may be caused by the interaction of chromoforic group of ligand and outer G-tetrads in the parallel Gquadruplexes [37][38][39].Finally, the most likely binding mode of 9-methoxyluminarine to the parallel G-quadruplexes' DNA is via end-stacking interactions with external G-tetrads, which are easily accessible (tested G4s possess only sidewise loops).
Moreover, we used this analytical technique to check whether 9-methoxyluminarine has the ability to induce the G-quadruplex folding without the addition of potassium ions.The 9-methoxyluminarine ligand is not able to induce the folding of single-stranded G-rich DNA into quadruplex structures in the absence of K + even after 24 h (Figure 3) as we expected on the basis on our previous studies [21].DNA oligonucleotide conformational changes were clearly observed after the addition of 20, 50 and 100 mM KCl, and even after 24 h from these additions.The smallest effect was observed in the case of the RET oligonucleotide after the addition of 100 mM KCl; however, after 24 h, a slight increase in the intensity of the positive signal at 265 nm was noticed.On the other hand, all CD spectra of RET oligonucleotide (regardless of the compounds added) exhibit the presence of a very weak positive band at 295 nm, which could indicate the existence of a small portion of the antiparallel structure in the tested mixture.In the case of catG4 oligonucleotide, the addition of 20 mM KCl caused the formation of a mixed parallel-/antiparallel-stranded structure, predominantly as an antiparallel structure as evidenced by the major positive signal which appeared at 295 nm.The same addition of potassium ions resulted in a mixture of parallel and antiparallel G4s for the c-KIT1 oligonucleotide.After the addition of 100 mM KCl to c-KIT1 or catG4 oligonucleotides, their CD spectra were dominated by the parallel over the antiparallel G4 form.After 24 h, only parallel G-quadruplexes were observed (Figure 3).These results suggest that the formation of parallel G-quadruplexes in the presence of the 9-MeLM ligand, but induced by potassium ions, occurs with the participation of the antiparallel form.Moreover, we could observe that the kinetics of final folding into parallel G4s is different for each tested sequence.Moreover, we used this analytical technique to check whether 9-methoxyluminarine has the ability to induce the G-quadruplex folding without the addition of potassium ions.The 9-methoxyluminarine ligand is not able to induce the folding of single-stranded Grich DNA into quadruplex structures in the absence of K + even after 24 hours (Figure 3) as we expected on the basis on our previous studies [21].DNA oligonucleotide conformational changes were clearly observed after the addition of 20, 50 and 100 mM KCl, and even after 24 h from these additions.The smallest effect was observed in the case of the RET oligonucleotide after the addition of 100 mM KCl; however, after 24 h, a slight increase in the intensity of the positive signal at 265 nm was noticed.On the other hand, all CD spectra of RET oligonucleotide (regardless of the compounds added) exhibit the presence of a very weak positive band at 295 nm, which could indicate the existence of a small portion of the antiparallel structure in the tested mixture.In the case of catG4 oligonucle- dition of 100 mM KCl to c-KIT1 or catG4 oligonucleotides, their CD spectra were dominated by the parallel over the antiparallel G4 form.After 24 h, only parallel G-quadruplexes were observed (Figure 3).These results suggest that the formation of parallel Gquadruplexes in the presence of the 9-MeLM ligand, but induced by potassium ions, occurs with the participation of the antiparallel form.Moreover, we could observe that the kinetics of final folding into parallel G4s is different for each tested sequence.

DNA Melting Studies
We performed UV-monitored melting experiments to investigate the thermodynamic properties of G-quadruplexes and their complexes with the small molecule of 9-methoxyluminarine.Our thermal denaturation studies were carried out in 10 mM Tris-HCl buffer (pH 7.2) containing the stabilizing potassium cation in a concentration reduced to 10 mM and with the addition of 90 mM LiCl.Thanks to this approach, the proper melting range

DNA Melting Studies
We performed UV-monitored melting experiments to investigate the thermodynamic properties of G-quadruplexes and their complexes with the small molecule of 9-methoxyluminarine.Our thermal denaturation studies were carried out in 10 mM Tris-HCl buffer (pH 7.2) containing the stabilizing potassium cation in a concentration reduced to 10 mM and with the addition of 90 mM LiCl.Thanks to this approach, the proper melting range could be observed, and the effect of the ligand could be evaluated.The melting curves of Gquadruplexes were found to be monophasic, and the same effect was observed for curves obtained for G4 with the presence of three equiv. of the ligand (Figure 4).Temperatures at half transition for the studied complexes can be found in Table 1.The obtained T m indicates that the 9-MeLM ligand does not significantly affect the stability of the G-quadruplexes with parallel topology.The hysteresis was observed between the melting and annealing curves (temperature change rate of 1 • C per minute), which indicates the slow kinetics of the processes (Figure S2) [40,41].

Binding Parameters via Fluorescence Spectroscopy
9-Methoxyluminarine (9-MeLM) emits bright and stable fluorescence in the visible region with its maximum at 495 nm.We used fluorescence titration method to obtain the binding characterization of 9-MeLM to the studied G4s, such as RET, c-MYC, c-KIT1, catG4, c-MYC1, c-MYC2, and c-MYC3.Compared with c-MYC1 (being indeed core of c-MYC), c-MYC2 has a 5 ′ -end flanking sequence TAA, c-MYC3 has a 3 ′ -end flanking sequence TAA while c-MYC has two flanking sequences on both ends (TGA at 5 ′ -end and TAA at 3 ′ -end).As in our previous work [21], the quenching of 9-MeLM was observed during the addition of chosen G-quadruplexes.A decrease in the fluorescence intensity at 495 nm with increasing concentration of G4 (1-10 µM) indicates its binding to the G-quadruplex structure (Figures 5 and S3).As shown in Figure 5B, the changes in 9-MeLM fluorescence intensity under the same c-MYC G4 to 9-MeLM ratio, indicating the order of binding affinity to 9-MeLM, is c-MYC3 (with 3 ′ -end flanking sequence) < c-MYC (with both 5 ′ -end and 3 ′ -end flanking sequences) < c-MYC2 (with the 5 ′ -end flanking sequence) < c-MYC1 (without flanking sequences at both ends).
region with its maximum at 495 nm.We used fluorescence titration method to obtain the binding characterization of 9-MeLM to the studied G4s, such as RET, c-MYC, c-KIT1, catG4, c-MYC1, c-MYC2, and c-MYC3.Compared with c-MYC1 (being indeed core of c-MYC), c-MYC2 has a 5′-end flanking sequence TAA, c-MYC3 has a 3′-end flanking sequence TAA while c-MYC has two flanking sequences on both ends (TGA at 5′-end and TAA at 3′-end).As in our previous work [21], the quenching of 9-MeLM was observed during the addition of chosen G-quadruplexes.A decrease in the fluorescence intensity at 495 nm with increasing concentration of G4 (1-10 µM) indicates its binding to the G-quadruplex structure (Figures 5 and S3).As shown in Figure 5B, the changes in 9-MeLM fluorescence intensity under the same c-MYC G4 to 9-MeLM ratio, indicating the order of binding affinity to 9-MeLM, is c-MYC3 (with 3′-end flanking sequence) < c-MYC (with both 5′-end and 3′-end flanking sequences) < c-MYC2 (with the 5′-end flanking sequence) < c-MYC1 (without flanking sequences at both ends).The fluorescence data were analyzed using the Stern-Volmer equation [42]: where K SV is the Stern-Volmer constant, Q is the total quencher concentration, and F and F 0 are the fluorescence intensities in the presence and absence of the quencher (here, G4), respectively.In all cases, we observed a linear relationship between 9-MeLM fluorescence and G4 concentration, with a correlation coefficient higher than 0.990 (Figures S4 and S5).It is worth mentioning that we have recently proved that the fluorescence of the 9-methoxyluminarine ligand is selectively and effectively quenched by the parallel c-MYC G-quadruplex via static mode due to the formation of a 1:1 stoichiometry of the 9-MeLM/c-MYC complex.The fluorescence lifetimes calculated for 9-methoxyluminarine (9-MeLM) and its complexes with DNA (9-MeLM/DNA, 1:10 or 1:20) clearly indicated static quenching as there is no apparent change in the fluorescence lifetime [21].Thus, we have assumed that all the tested parallel G4s quenched 9-methoxyluminarine fluorescence by static interactions forming non-fluorescent complexes.The obtained values of the correlation coefficient indicate that the static quenching fits in the Stern-Volmer model as assumed (Table 2).
Table 2. Parameters for the interaction of 9-methoxyluminarine with parallel G-quadruplexes determined using the S-V model in fluorescence titration experiments (K SV -Stern-Volmer quenching constant is interpreted as K-binding constant, k q -quenching rate constant, R-correlation coefficient, λex = 390 nm).The formation of complex was further confirmed from the values of the quenching rate constant k q , which was estimated from the values of K SV and the value of the fluorescence lifetime of 9-MeLM in Tris buffer (τ 0 = 9.98 ns) [21] in the absence of G4, using the following equation:

G4 K (×10
The obtained k q values (collected in Table 2) are higher than the previously reported values for various quenchers in the presence of the biopolymer of 2 × 10 10 M −1 s −1 (maximum value for scatter collision quenching constant) [42].Thus, the obtained results indicate that the quenching of 9-methoxyluminarine by c-MYC is caused by formation of the complex and is not induced by collision-mediated quenching.
Using the definition of the association constant and the fact that the fluorescence intensity in the presence of a quencher is proportional to the concentration of uncomplexed fluorophores ([L] = [L 0 ] − [LQ]), one can obtain an expression describing the ratio of fluorescence intensities without and in the presence of a quencher: In the case of static quenching, the Stern-Volmer quenching constant (K SV ) from Equation (1) has been replaced by the association constant K in Equation ( 3).However, one should remember that [Q] in Equation ( 3) is not the total but a free amount of quencher.Thus, for small [Q] values in the case of complex formation, the Stern-Volmer quenching constant is not equal to K, but rather where [L 0 ] is initial concentration of fluorophore [43].Thus, the experimental quenching constant is equal to the equilibrium one (complexing) only if K[L 0 ] << 1.In our case, the starting concentration of 9-MeLM was only 1 µM, so we decided to interpret K SV as the association constant, K.The estimated values of the K data along with correlation coefficient are given in Table 2.The calculated K values are a few times higher for c-MYC1 and c-MYC2 in comparison with other parallel G4s.
Particularly, while analyzing the order of obtained binding constants, K for c-MYC with and without flanking sequences, we noticed that the lack of flanking sequences at the 3 ′ -end of c-MYC G4 is correlated with a higher value of K. Thus, these results suggest that 9-MeLM preferentially binds to the 3 ′ end of the c-MYC promoter.

Materials
9-Methoxyluminarine (9-MeLM) was obtained as previously performed [21].The stock solutions of 9-MeLM were prepared in H 2 O at a concentration of 1.33 mM and was stored at 4 • C. Tris Base (CAS Number 77-86-1) and Tris HCl (CAS Number 1185-53-1) were obtained from Aldrich Chemical Co.(Pozna ń, Poland) and used as received.All experiments were carried out using quartz cuvettes with a 10 mm optical path, which in the case of fluorescence measurements was for the excitation direction and a 4 mm path length in the emission direction.

Circular Dichroism Spectroscopy (CD)
Circular dichroism spectra were recorded at 25 • C on a Jasco J-810 spectropolarimeter (Jasco, Tokyo, Japan), equipped with a Peltier accessory (Jasco, Kyoto, Japan).Each spectrum represented an average of three scans accumulated with a scan speed of 200 nm/min from 220 to 450 nm.The corresponding buffer (10 mM Tris-HCl with or w/o 100 mM KCl) was used as a blank solution.A spectrum of buffer solution was subtracted from the spectra of the investigated probes.CD spectra were recorded at a strand concentration of 5 µM.The ligand was added to G4 DNA or unfolded oligonucleotide solutions at increasing concentrations from 0.5 to 5 molar equivalents.After each addition of the ligand, the titrated solution was incubated for 3 min prior the CD spectrum measurement.The total volume of added ligand was kept below 2% to avoid dilution.The corresponding UV-vis spectra were collected simultaneously in channel 3, whereas channels 1 and 2 monitored the CD signal and HT (voltage), respectively.

Figure 1 .
Figure 1.A schematic illustration of binding mode between 9-MeLM ligand and c-M plex.

Figure 1 .
Figure 1.A schematic illustration of binding mode between 9-MeLM ligand and c-MYC G-quadruplex.

30 a
Tm of G4s DNA in the presence of 10 mM KCl and 90 mM LiCl in 10 mM Tris-HCl buffer (pH 7.2).b Tm of G4s DNA incubated with 3 equiv.of ligand.Data were collected at 295 nm.Typically, three replicate experiments were performed.

Table 3 .
The oligonucleotides used in this study and their PDB ID.The guanines engaged in G-tetrad formation are shown in bold, whereas bases involved in sidewide loops are shown in italics.