Synthesis and Spectroscopic Investigation of a Hexaaza Lanthanum(III) Macrocycle with a Hybrid-Type G4 DNA Stabilizing Effect.

Herein we present a mononuclear lanthanum(III) complex obtained in a template cyclocondensation reaction of lanthanum(III) nitrate salt, 1,2-propanediamine, and 2,6-diacetylpyridine (LaPA complex). A preliminary investigation of the biological potential of this compound was conducted using a biomedically relevant target Tel26. We found that, different from parallel G4, antiparallel G4, and duplex DNA, only a hybrid-type G4 structure of Tel26 in a K+ solution was significantly stabilized by ≥7 °C, which emerged in our UV melting studies. Moreover, LaPA induced structural changes in the Tel26 structure in a K+-deprived solution, suggesting that it may also lead to conformational changes in "non-G4" telomeric DNA.

The G-quadruplex (G4) is a noncanonical DNA structure formed in the guanine-rich regions and preserved by Hoogsteen interactions.[1] Only in 2013 were G4s shown to exist in human genome by visualization using a immunofluorescence staining method with a specific BG4 antibody.[2] They are mostly present in the key regulatory regions of the human genome such as promoters and gene bodies.[3] Moreover, they were revealed to form at the ends of the chromosomes in the regions called telomeres. Hence, they are expected to be involved in maintaining the chromosome stability by preventing the accidental damage of genetic material.[4]

Telomeric DNA, when structured in the G4 form, cannot be recognized by telomerase. This inhibits the telomere lengthening, notably responsible for cancer cell immortality. Thus, directing ligands to stabilize specific telomeric G4-structured DNA constitutes a valuable anticancer strategy.[5] We have recently reported on the telomeric DNA binding capacity of phenylalanine, as verified with Tel26 and Pu22.[6]

Several examples of G4 targeting derivatives are based on planar organic heteroaromatic systems and are able to interact through π–π-stacking, hydrogen-bonding, and electrostatic interactions at the apex and with the loops of G4s, as well as intercalate between G quartets.[7] So far, mostly derivatives of pyridine, acridine, porphyrin, phenanthrene, and naphthalenediimide are known as G4 DNA targeting agents and stabilizers.[8] However, there are also some reports on macrocyclic entities such as hexaoxazoles, which are able to selectively induce the antiparallel G4 DNA structure of Tel24.[9] The studies available on this topic also show that metal complexes can play an important role as G4 DNA binders.[10] Although the actual function of G4 DNA folds is still under investigation, their relevance is widely recognized, justifying the current interest in the development of new molecules able to stabilize the G4 structures of interest.[11]

An encouraging pathway to build the macrocyclic planar complexes of interest arises from the core concept of supramolecular chemistry: self-assembly. A properly designed macrocycle can be formed in a one-pot synthesis in high yields.[12] The molecular ring size and metal–ligand interactions are relevant in the context of application as biomolecule targeting systems. Therefore, flat macrocycles are able to selectively interact with specific structures of nucleic acids.[13]

In the present work, we report a mononuclear lanthanum(III) complex with an 18-membered hexaaza macrocyclic ligand. The described compound acts as a parallel/antiparallel hybrid-type G4 DNA stabilizer, while it destabilized parallel G4 and duplex DNA and had no influence on the antiparallel DNA stability, as shown by UV melting studies. Additionally, circular dichroism (CD), UV, fluorescence, and scanning electron microscopy (SEM) studies were performed to better elucidate the observed phenomena.

Herein, we report on the formation of a mononuclear lanthanum(III) nitrate complex of a hexaaza macrocyclic ligand of the formula [La(C24H30N6)(NO3)2CH3OH]NO3, which is soluble and stable in water (Figure S1). A template reaction of lanthanum(III) nitrate salt, 1,2-propanediamine, and 2,6-diacetylpyridine allows one to obtain the LaPA complex of an 18-membered macrocyclic ligand as a result of [2 + 2] cyclocondensation (Scheme ). The compound was characterized by electrospray ionization mass spectrometry (Figure S2), IR (Figure S3), and elemental analysis (see the Supporting Information).

Scheme 1

Template Synthesis of the LaPA Macrocycle

As revealed by X-ray diffraction measurement (Figures and S4 and Table S1), the coordination sphere of lanthanum(III) is filled by six nitrogen atoms (four nitrogen atoms from azomethine groups and two nitrogen atoms from pyridines) and five oxygen atoms (four oxygen atoms from two nitrate groups behaving as bidentate chelators and one oxygen atom from a methanol molecule). Hence, the coordination number of lanthanum(III) is formally equal to 11. A similar coordination was previously observed in a number of lanthanide complexes with hexaaza macrocycles.[14] It might be noted that such a large coordination number is relatively rare. In the Cambridge Structural Database,[15] there are 14063 structures of lanthanide complexes with coordination number 9, while only 514 have coordination number 11 or higher. Among those 514 complexes, 346 have NO3– ligands (in the chelating disposition). The close vicinity of available oxygen atoms seems to prefer the high coordination numbers: 67.3% of all complexes compared with 16.4% for coordination number 9. The macrocycle, with a roughly planar skeleton, is close to mirror C symmetry (cf. torsion angles). Its planarity and the presence of the central cavity filled with positively charged lanthanum(III) became the foundation for the evaluation of LaPA as a G4 DNA binder. The macrocycle is 12.77 Å long and 9.76 Å wide, and the thickness between the two most distant oxygen atoms from nitrates is 8.52 Å. In the G4 DNA hybrid-1 structure,[16] the minor and major grooves are 6.16 and 9.62 Å wide, respectively, and the distance between the stacked tetrads is ca. 3.18 Å. Because of this size mismatch, it is envisaged that the binding and assembly of the LaPA–Tel26 complex could occur at the apex of the stacked tetrads through multiple interactions (π–π-stacking, hydrogen-bonding, and electrostatic interactions) stabilizing the loop and flanking regions.[10b]

Figure 1

Perspective view of the macrocycle LaPA in the top-to-bottom (A) and side (B) views.

In order to explore the potential of LaPA to interact with G4 DNA structures, we have selected three G-rich strands of DNA forming various G4 structures: Pu22 forming parallel G4,[17] Tel22 forming antiparallel G4 in the presence of Na+,[18] and Tel26 folding into a hybrid-type parallel–antiparallel G4[19] in K+ solutions. Our main goal was to obtain a macrocycle that selectively targets the G4s; therefore, the duplex DNA was also tested because binding to it would be clearly undesired as it could diminish the potential therapeutic effects of a candidate as a G4 DNA binder. In this study, the buffers were chosen so as to obtain distinct G4s that are not too stable for the binding experiments (Table S2). Because the experimental conditions are quite different for each oligonucleotide, one has to be aware that no quantitative comparison between various DNA strands can be made here.

In an initial UV-based thermal stabilization screening, we noticed that LaPA provokes a ∼2–4 °C structural destabilization of both duplex and parallel G4 (Pu22) DNA, while it causes no significant effect on antiparallel DNA, as verified using Tel22 in a Na+ solution (Figure S5 and Table ). Conversely, in the case of the hybrid-type parallel–antiparallel G4 structure of Tel26 in a K+ solution, LaPA led to a ∼7 °C stabilization effect (Figure A and Table ), a feature that attracted our interest because of the above-mentioned potential of Tel26 stabilizers.[20]

Table 1

ΔTm (°C) for Complexes of LaPA with Duplex DNA, Pu22, Tel22, and Tel26 (First Derivatives Shown in Figures S9–S12)a

sample	1 equiv of LaPA	2 equiv of LaPA
duplex DNA	–1.9 ± 0.4	–3.6 ± 0.4
Pu22	–1.6 ± 0.4	–3.6 ± 0.6
Tel22	+0.3 ± 0.4	–0.4 ± 0.4
Tel26	+7.9 ± 0.4	+6.4 ± 0.6

The results are reported as mean ± standard deviation values. 1 equiv = 2.5 μM.

Figure 2

UV melting curves of Tel26 alone and with LaPA (A) and fluorescence quenching of Tel26 with the incremental addition (1–8 equiv) of LaPA (B). Stern–Volmer plot I0/I – 1 versus [LaPA] at the fixed wavelength λ = 312 nm: ●, experimental data points; solid line, linear fitting of the data (inset in part B). Buffer: 10 mM Tris, 100 mM KCl, pH = 7.4. 1 equiv = 2.5 μM.

Moreover, the fluorescence titration experiments performed on different oligonucleotides showed that LaPA quenches the intrinsic emission[21] of Tel26 (Figure B). Conversely, it has a negligible impact on the intrinsic emission of other oligonucleotides (Figure S6). The observed hypochromicity without any significant shift in the emission spectrum of Tel26 suggests that the macrocycle, in fact, may directly interact with the terminal G-tetrads, which found confirmation in the CD intensity increase in the spectrum of LaPA–Tel26 complex (Figure S7A). Furthermore, to obtain more quantitative information, the quenching constant KSV was calculated from the Stern–Volmer equation (see the Supporting Information)[22] and is equal to 1.5 × 104 M–1 (Figure B, inset). Knowing that the fluorescence lifetime τ of biomolecules is ∼10–8 s, the quenching rate constant kq of Tel26 was determined from the equation KSV = kqτ0[23] as 1.5 × 1012 M–1·s–1. Similar values were obtained in the case of allocryptopine (KSV = 1.0 × 105 M–1; kq = 4.2 × 1013 M–1·s–1), which binds partially in the loops and through end stacking to the hybrid-type G4. Higher values of KSV and kq would suggest that groove binding of the compounds may occur.[24]

It needs to be noted that UV titration and CD studies of Pu22 and Tel22 with LaPA do not reveal any significant structural changes in the spectra of oligonucleotides (Figures S7B,C and S8).[25]

The CD spectrum of the folded hybrid-type parallel–antiparallel telomeric G4 is characterized by a negative band at 240 nm, and the two positive bands at ca. 265 and 290 nm (Figure S7A),[26] while a parallel Pu22 G4 lacks the long-wavelength band (Figure S7B). An antiparallel telomeric G4 exhibits one negative band at 265 nm and two positive bands at 245 and 295 nm (Figure S7C).[26] In a K+-deprived solution, we obtained a CD spectrum for Tel26 exhibiting two strong bands—a negative band at 240 nm and a positive band at 260 nm—and a bump at ca. 292 nm, which was previously associated with the “non-G4” form of this G-rich DNA[5c] (Figure A). Upon ligand addition, CD spectral bands underwent significant changes, as reported in Figure A, clearly indicating that LaPA causes conformational changes in the “non-G4” Tel26.[27] Interestingly, when the “non-G4” Tel26 was annealed in the presence of LaPA, a melting sigmoid with a high Tm of 62.0 ± 0.6 °C in the UV melting experiment was observed, testifying to the formation of a “non-G4” Tel26–LaPA stable complex (Figure B).

Figure 3

“Non-G4” Tel26 in the absence of K+ and with LaPA: CD studies (A) and UV–vis melting experiments (B; the first derivative is shown in Figure S13). 1 equiv = 2.5 μM.

In order to better characterize the complex formation occurring between “non-G4” Tel26 and LaPA, we performed also a SEM analysis of the morphologies adopted by the “non-G4” Tel26–LaPA complex. SEM images of this complex shown in the micrographs (Figure ) revealed spherical particles of width ranging from 3 μm up to 8 μm able to further agglomerate. It is analogous to other literature reports on nucleic acid complexes, which induce the clustering of several nanofibers into larger aggregates (Figure B).[6a,28]

Figure 4

SEM micrographs of Tel26 at concentration 2.5 μM folded in the presence of 1 equiv of LaPA in a Tris buffer (10 mM Tris, pH = 7.4).

Herein, we described the synthesis and characterization of the lanthanum(III) complex [La(C24H30N6)(NO3)2(CH3OH)]NO3, indicated as LaPA, whose potential in biomedicine was explored by our preliminary studies. We used UV, CD, and fluorescence spectroscopies to investigate the ability of LaPA to interact with different DNA models, including duplex DNA and oligonucleotides, forming various G4 structures: parallel Pu22, antiparallel Tel22, and hybrid-type parallel–antiparallel Tel26. We found that LaPA was able to stabilize hybrid-type G4 but not parallel and antiparallel G4 or double-helical DNA. The LaPA–Tel26 complex was endowed with good stability (Tm ≥ 60 °C), even in the absence of K+, suggesting that the lanthanum(III) derivative is intrinsically able to form complexes with “non-G4” Tel26. A tendency of Tel26 to form spherical entities in the presence of a LaPA binder was also evidenced by SEM. We hypothesize that LaPA is an active agent binding probably by π–π-stacking and electrostatic interactions to the apex of the folded hybrid-type G4 DNA, thus stabilizing it.