Complex Biophysical and Computational Analyses of G‐Quadruplex Ligands: The Porphyrin Stacks Back Authors: Giuseppe Satta, Marko Trajkovski, Alessio Cantara, Monica Mura, Claudia Meloni, Giulia Olla, Michaela Dobrovolná, Luisa Pisano, Silvia Gaspa, Andrea Salis, Lidia De Luca, Francesca Mocci, Vaclav Brazda, Janez Plavec, Massimo Carraro Published: 2024 ABSTRACT AbstractG‐quadruplexes (G4 s), as non‐canonical DNA structures, attract a great deal of research interest in the molecular biology as well as in the material science fields. The use of small molecules as ligands for G‐quadruplexes has emerged as a tool to regulate gene expression and telomeres maintenance. Meso‐tetrakis‐(N‐methyl‐4‐pyridyl) porphyrin (TMPyP4) was shown as one of the first ligands for G‐quadruplexes and it is still widely used. We report an investigation comprising molecular docking and dynamics, synthesis and multiple spectroscopic and spectrometric determinations on simple cationic porphyrins and their interaction with different DNA sequences. This study enabled the synthesis of tetracationic porphyrin derivatives that exhibited binding and stabilizing capacity against G‐quadruplex structures; the detailed characterization has shown that the presence of amide groups at the periphery improves selectivity for parallel G4 s binding over other structures. Taking into account the ease of synthesis, 5,10,15,20‐tetrakis‐(1‐acetamido‐4‐pyridyl) porphyrin bromide could be considered a better alternative to TMPyP4 in studies involving G4 binding. Introduction Several genomic regions are characterized by a relative abundance of guanosine. The large amount of guanosine units, in the presence of monovalent cations such as K+ and Na+, allows their folding and leads to the formation of planar tetrads called G‐quartets, together by a network of hydrogen bonds through Hoogsteen base pairing.[ 1 , 2 ] Stacking of G‐quartets, stabilized by the coordination of the monovalent metal ion with the O6 oxygen of guanines, leads to the formation of “non‐canonical” structures, characterized by a four chains helical structure, called G‐quadruplexes (G4) (Figure 1a). The presence of G4 in telomeres and several other non‐codifying regions of genome have been associated with a variety of regulatory functions. The human telomere is approximately 5,000 to 8,000 base pairs in length and features a single‐stranded 3′ overhang ranging from 100 to 200 bases. [1] This overhang is primarily composed of the repetitive TTAGGG sequence. Most healthy cells possess a limited capacity for division, as was demonstrated by Hayflick and Moorhead in 1961. [3] It is currently known that this effect is related to telomere length, which act as biological clocks that, after reaching a limiting length, trigger the senescence process. [4] A problem often correlated with the presence of tumor activity is abnormal cellular proliferation, which in most cases, is associated with the overexpression of telomerase activity. [5] This enzyme, a reverse cellular transcriptase, regulates telomere elongation, thereby preserving its integrity. A correlation has been assessed between telomere length maintenance and the cellular ability, typical of cancer cells, to escape replicative senescence, and in fact more than 90 percent of cancer cells show telomerase enzyme activity. [6] The fact that most healthy cells are telomerase‐silent while there is its overexpression in cancer cells, makes this enzyme an attractive target for post‐diagnosis treatments. The inhibition of telomerase can be achieved either through small molecules that bind directly to the enzyme[ 7 , 8 ] or by stabilizing G4 structures. The latter method involves the use of ligands that bind to and stabilize G4 structures, which act as physiological blocks to the telomerase's ability to code and extend telomeres. [9] G4 can also be found in oncogene promoters, replication initiation sites and untranslated regions, showing their biological relevance.[ 10 , 11 , 12 ] For example, the c‐Myc gene plays an essential role in the regulation of cell growth, proliferation, and apoptosis. When over‐expressed or mutated, as in cancer cells, this gene can drive cells toward uncontrolled proliferation and thus contribute to the formation of various types of cancer. [13] Within its sequence, the c‐Myc proto‐oncogene possesses the nuclear hypersensitivity element III1 (NHEIII1) region which has been shown to be highly influential in the regulation of this gene. [14] Stabilizing this structure with ligands was noted to suppress further transcriptional activation of the c‐Myc gene. [15] G4 can be unimolecular or intermolecular. Depending on the orientation of the chains, they can adopt different topologies [16] (Figure 1b), influenced by factors such as molecular crowding[ 17 , 18 ] and nature of monovalent cation. [19] Furthermore, the structural and biochemical features of G4 s, prompted their use in biosensors [20] and nanomaterials.[ 21 , 22 ] (A) structure of the G‐quartet formed by Hoogsteen base pairing of four guanine residues and piling up to give G4. (B) Schematic intramolecular topologies of G4 structures. Blue and orange rectangles depict guanines exhibiting syn and anti conformation of glycosidic torsion angle, respectively. In the case of antiparallel topology, a and b denote the orientations of the G‐tracts, corresponding to upward and downward directions, respectively, starting from the 5′ end in a clockwise manner. The study of G4 ligands has thus become a productive field of research.[ 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 ] Basically, there are 4 ways in which a ligand can interact with a G4: stacking on top or bottom G‐quartets, intercalation between them, interaction with loops, or a combination of these ways. Electron‐deficient cores promote interactions through π‐π stacking, while cationic portions on substituents determine electrostatic interactions with negative charged phosphodiester groups, and so on. [33] These considerations led to the discovery of various types of ligands, among others porphyrins: natural derivatives include Fe(III)‐protoporphyrin IX, [34] while the first synthetic derivative is 5,10,15,20‐tetrakis‐(N‐methyl‐4‐pyridyl)porphyrin (TMPyP4), whose activity as telomerase inhibitor was demonstrated by Wheelhouse et al. in 1998. [35] TMPyP4 has been shown also to reduce the expression of the proto‐oncogene c‐Myc and several c‐Myc‐regulated genes that contain G4‐forming sequences. This modulation led to in vivo antitumor effects in various models, including the ability to inhibit tumor growth and prolong survival. [36] TMPyP4, has become a staple between ligands used in G4 studies but displayed a limited selectivity, [37] prompting to the development of various derivatives with demanding synthesis.[ 38 , 39 , 40 , 41 ] In this work, several simple porphyrin derivatives with potential binding and stabilizing activity toward G4 were screened through a molecular modelling procedure, which included docking and molecular dynamics. Simulations were conducted on different G4 s comprising one from the human telomeric sequence with a parallel topology (PDB ID:1KF1 [42] ) and another present in the NHEIII1 region of the c‐Myc proto‐oncogene, also known to have a parallel propeller type topology (PDB ID: 1XAV [43] ). The results of the docking procedure guided in the choice of the derivatives to be synthesized and subjected to various stability and binding studies. Specifically, a derivative of 5,10,15,20‐tetrakis‐4‐pyridylporphyrin (TPyP) with amidomethyl substituent was found to be the most promising, and its properties are here compared with those of well‐known porphyrin derivative, TMPyP4. One of the focuses of the work was to assess whether substituting the methyl groups of TMPyP4 affect its interaction with G4 from telomere and c‐Myc. This comparison was made by evaluating the results obtained for the substituted derivatives alongside those obtained for TMPyP4 using NMR spectroscopy and mass spectrometry. NMR spectroscopy has proven to be a very versatile technique for assessing the formation of a complex, and providing crucial insights into the mode of ligand/G4 interaction. In particular NOESY and ROESY experiments,[ 43 , 44 , 45 ] have been employed for structural assessment of the complexes and identification of the corresponding G4‐ligand interfaces. The stability of the ligand/G4 complexes was evaluated by mass spectrometry,[ 46 , 47 , 48 ] specifically, with MS/MS technique: the intensity of isolated molecular ion generated by the porphyrins/G4 complex was measured as a function of increasing collision energy in the analyzer. Porphyrin derivatives were then subjected to some preliminary tests to evaluate their cytotoxicity on breast cancer (MCF‐7). Results and Discussion In this work, the interaction between tetracationic porphyrins derivatives and different DNA G4 s was evaluated. Proposed ligand structures, based on the porphyrin core (Figure 2) with different meso substituents, are reported in Table 1. TMPyP4, chosen as literature reference ligand, and designed ligands, PL1 to PL7, were first subjected to molecular docking calculations on different G4 structures. Most promising ligands were then synthetized, and their G4 binding abilities studied by UV and NMR spectroscopies and mass spectrometry. Chosen sequences were derived from the NHEIII1 region of the c‐Myc proto‐oncogene, and the human telomeres. Porphyrin general core. R groups indicate meso substituents. Table 1Proposed porphyrin ligands’ meso substituents. Ligands Ligands’ meso substituents (R) MW TMPyP4 678.84 PL1 943.07 PL2 943.07 PL3 1115.75 PL4 911.16 PL5 850.83 PL6 907.05 PL7 850.94 PL7‐Me 907.05 PL7‐2Me 963.16 Wiley‐VCH GmbH Proposed porphyrin ligands’ meso substituents. Ligands Ligands’ meso substituents (R) MW TMPyP4 678.84 PL1 943.07 PL2 943.07 PL3 1115.75 PL4 911.16 PL5 850.83 PL6 907.05 PL7 850.94 PL7‐Me 907.05 PL7‐2Me 963.16 Molecular Docking The GScore values for the best poses obtained for the studied ligands on the studied G4 s are reported in Table 2. According to the docking results obtained on all‐parallel G4 1XAV and 1KF1, candidates PL3, PL6 and PL7 were the most promising. Unfortunately, the synthesis of PL3 proved to be challenging. Given that PL6 and PL7 exhibited a comparable G4 affinity to that showed by PL3, the synthesis of the latter was not pursued further. It was observed that both PL6 and PL7 demonstrated binding ability to other topologies in addition to the all‐parallel ones. However, given the absence of stereocenters and the lower molecular weight, PL7 was selected for further computational studies. Table 2GScore results from molecular docking experiments. Ligands 1XAV[a] 1KF1[b,f] 2JPZ[c] 2HY9[d] 143D[e] TMPyP4 −13,50 np[g] np np np PL1 −14,11 −14,63 ‐[h] ‐ np PL2 −14,77 −14,16 ‐ ‐ −12,79 PL3 −17,84 −19,14 ‐ ‐ −14,70 PL4 −9,89 −15,77 ‐ ‐ np PL5 −14,07 −14,28 np −14,09 ‐ PL6[j] −17,74 −19,12 −15,29 −15,36 ‐ PL7 −18,69 −18,66 −18,21 −16,79 −14,81 G4 names correspond to PDB ID. [a] Main G4 forms in the c‐MYC promoter gene. [b] Human telomeric parallel G4. [c] Human telomeric hybrid‐form 1 G4. [d] Human telomeric hybrid‐form 2 G4. [e] Human telomeric antiparallel G4. [f] All G4 structures were determined in solution by NMR except for 1KF1 (X‐Ray solid structure). [g] np: no poses were found. [h] the dash symbol indicates that the docking was not performed. [j] Configuration of the chiral centers of ligand PL6 in the best pose: SRSR, SSSS, SRRS, SSSS.Wiley‐VCH GmbH GScore results from molecular docking experiments. Ligands 1XAV[a] 1KF1[b,f] 2JPZ[c] 2HY9[d] 143D[e] TMPyP4 −13,50 np[g] np np np PL1 −14,11 −14,63 ‐[h] ‐ np PL2 −14,77 −14,16 ‐ ‐ −12,79 PL3 −17,84 −19,14 ‐ ‐ −14,70 PL4 −9,89 −15,77 ‐ ‐ np PL5 −14,07 −14,28 np −14,09 ‐ PL6[j] −17,74 −19,12 −15,29 −15,36 ‐ PL7 −18,69 −18,66 −18,21 −16,79 −14,81 G4 names correspond to PDB ID. [a] Main G4 forms in the c‐MYC promoter gene. [b] Human telomeric parallel G4. [c] Human telomeric hybrid‐form 1 G4. [d] Human telomeric hybrid‐form 2 G4. [e] Human telomeric antiparallel G4. [f] All G4 structures were determined in solution by NMR except for 1KF1 (X‐Ray solid structure). [g] np: no poses were found. [h] the dash symbol indicates that the docking was not performed. [j] Configuration of the chiral centers of ligand PL6 in the best pose: SRSR, SSSS, SRRS, SSSS. Molecular Dynamics Simulations To investigate the impact of PL7 on the stability of G4, a series of molecular dynamics (MD) simulations were conducted. The system was simulated in a water solution using 1KF1 as G4, both in the presence and absence of the ligand. For the complex PL7/G4, the structure obtained from molecular docking studies was used. Simulations were conducted at temperatures of 300, 500, 525 and 550 K, to induce denaturation. Experimentally, the process of thermal denaturation is observed to occur over a timescale that is too long to be replicated within the constraints of reasonable timescales in silico. The use of higher temperatures than those measured in real laboratory experiments has been found to accelerate the denaturation process and make it occur within accessible timescales for MD simulations.[ 49 , 50 , 51 ] It is important to note that this approach is feasible due to the harmonic potentials of the force field, which prevent significant deviations in bond length or bond breakage even at elevated temperatures. Furthermore, previous studies have demonstrated that increasing temperature does not alter the denaturation pathway.[ 49 , 50 ] This strategy has recently been employed to assess the stabilization of the G4 structure following interaction with ligands, with results that are consistent with experimental data. [52] Figure 3 presents the root‐mean‐square deviation (RMSD) values calculated for the entire G4 structure, its complex with PL7, and selected portions during the MD simulation performed at 300 K. To ascertain the stability of the PL7/1KF1 complex, both the DNA atoms (shown in blue) and the ligand atoms (shown in yellow) were considered in RMSD simulations. The matching of the two RMSD values indicated that the ligand, once it was bounded, did not move significantly from its initial position, remaining stably linked to the G4. Moreover, the comparison of the RMSD between the free 1KF1 and PL7/1KF1 indicated that the ligand induced stabilization, as evidenced by the slightly smaller RMSD values in its presence. RMSD Results. (A)Color code used for quartets (B) Color code used for loops. These colors are the same used to indicate RMSD values. (C)RMSD calculated on: (top row) 1KF1 atoms except the first base in the 5’ direction (blue), and PL7/1KF1 atoms (yellow); (middle row) on each of the three G‐quartets; (bottom row) on each of the three loops. Simulations conducted at 300 K. (D) RMSDs calculated on three separate simulations at 525 K of free 1KF1. (E) RMSDs calculated on three separate simulations at 525 K of PL7/1KF1. This stabilization could be primarily attributed to the enhanced stability of the loops. The simulations at higher temperatures confirmed the stabilization effect of the ligand, as shown by the RMSD of the simulations performed at 525 K, shown in Figures 3D and E. As expected, the RMSDs of the system at 525 K were in general higher than those at 300 K, however the behavior with and without the ligand was qualitatively very different. The RMSD of the loops revealed that these flexible portions did not keep their original organization at high temperature, either in absence or in presence of the ligand. However, the ligand had a strong effect on the stability of the G‐quartets: in free 1KF1, denaturation occurred within the first 50 ns. In the presence of PL7, denaturation was slower or not reached at all (Figure 3E ‐ middle row), keeping the G4 structure stable. The simulations performed at 500 K and 550 K confirmed the stabilizing effect of the ligand, as shown by the RMSD reported Figure S3. Synthetized Oligonucleotides The sequences used in the docking studies were taken as a starting point for choosing those to be used in the spectroscopic and spectrometric studies (Table 3). The sequence indicated by the acronym 23TAG (PDB ID: 2JSK [53] ) has been employed, corresponding to tandem repeats of the human telomeric region. This sequence in K+ solution leads to the folding of a form1/form2 hybrid topology G4, with a 70 : 30 ratio respectively. [54] On the other hand, the modified sequences reported as CMA (PU19‐A2 A11, [55] PDB ID:2LBY) and CMT (PU19‐T2 A11 [55] ) are derived from the first 4 of 5 guanosine domains present in the NHEIII1 region of the c‐Myc gene. [44] CMA and CMT tend to fold in parallel topology G4 in a K+ solution. This enabled the assessment of both the binding capacity and the selectivity of the proposed ligands toward G4 structures with different topologies. Table 3DNA oligonucleotide sequences used in this study. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 23TAG 5’‐ T A G G G T T A G G G T T A G G G T T A G G G ‐3’ CMA 5’‐ T A G G G A G G G T A G G G A G G G T ‐3’ CMT 5’‐ T T G G G A G G G T A G G G A G G G T ‐3’ DNA oligonucleotides were synthesized on DNA/RNA H‐8 Synthesizer using standard phosphoramidite chemistry with DMT protecting groupWiley‐VCH GmbH DNA oligonucleotide sequences used in this study. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 23TAG 5’‐ T A G G G T T A G G G T T A G G G T T A G G G ‐3’ CMA 5’‐ T A G G G A G G G T A G G G A G G G T ‐3’ CMT 5’‐ T T G G G A G G G T A G G G A G G G T ‐3’ DNA oligonucleotides were synthesized on DNA/RNA H‐8 Synthesizer using standard phosphoramidite chemistry with DMT protecting group Binding Constant UV‐Visible Evaluation UV‐visible spectra of solutions of TMPyP4 and PL7 (Figure S8) were recorded upon addition of calf thymus DNA, 23TAG, and CMA, at 25 °C within the wavelength range of 200–800 nm. The porphyrins spectra typically exhibited a Soret band around 428 nm. Upon addition of DNA, in the TMPyP4 solution, a red shift of the maximum absorption band was observed for the duplex, 23TAG, and CMA, with shifts of 2.3 nm, 2.1 nm, and 2.9 nm, respectively. Similarly, the red shifts observed for the PL7 spectra amounted to 0.8 nm with duplex, 9.7 nm with 23TAG, and 5.8 nm with CMA. This red shift phenomenon can be associated with a decrease in the energy of the π→π* transition due to the interaction between the π‐bonding orbital of the DNA base pairs and the empty π*‐antibonding orbital of the ligand. The hypochromic effect, determined by comparing absorbance maxima, was evidence of the interaction occurring between the nucleotides and porphyrins. The hypochromic effect (Table 4) increased in the order duplex<23TAG